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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
cba897af-afa7-4783-9bce-eae3c98748ff | a-model-to-measure-the-spread-power-of-rumors | 2002.07563 | null | https://arxiv.org/abs/2002.07563v5 | https://arxiv.org/pdf/2002.07563v5.pdf | A Model to Measure the Spread Power of Rumors | With technologies that have democratized the production and reproduction of information, a significant portion of daily interacted posts in social media has been infected by rumors. Despite the extensive research on rumor detection and verification, so far, the problem of calculating the spread power of rumors has not been considered. To address this research gap, the present study seeks a model to calculate the Spread Power of Rumor (SPR) as the function of content-based features in two categories: False Rumor (FR) and True Rumor (TR). For this purpose, the theory of Allport and Postman will be adopted, which it claims that importance and ambiguity are the key variables in rumor-mongering and the power of rumor. Totally 42 content features in two categories "importance" (28 features) and "ambiguity" (14 features) are introduced to compute SPR. The proposed model is evaluated on two datasets, Twitter and Telegram. The results showed that (i) the spread power of False Rumor documents is rarely more than True Rumors. (ii) there is a significant difference between the SPR means of two groups False Rumor and True Rumor. (iii) SPR as a criterion can have a positive impact on distinguishing False Rumors and True Rumors. | ['Taymaz Akan', 'Mohammad-Ali Balafar', 'Elnaz Zafarani-Moattar', 'Mehrdad Ranjbar-Khadivi', 'Ali-Reza Feizi-Derakhshi', 'Narjes Nikzad-Khasmakhi', 'Meysam Asgari-Chenaghlu', 'Mohammad-Reza Feizi-Derakhshi', 'Zoleikha Jahanbakhsh-Nagadeh', 'Majid Ramezani'] | 2020-02-18 | null | null | null | null | ['rumour-detection'] | ['natural-language-processing'] | [-5.35087585e-01 -2.49724067e-03 -3.02454531e-01 -1.40497699e-01
3.55391741e-01 -1.60562366e-01 1.05813909e+00 2.98959345e-01
-7.17199296e-02 8.87730598e-01 6.06960952e-01 -3.12144011e-01
-3.71454619e-02 -8.10247958e-01 -1.80295199e-01 -4.25068051e-01
-2.72002220e-01 2.03641862e-01 6.36132285e-02 -6.20342970e-01
1.08036804e+00 3.37701678e-01 -1.53039968e+00 3.89526427e-01
6.89500570e-01 6.23423517e-01 2.47318387e-01 1.89279467e-01
-4.22833949e-01 1.52740490e+00 -9.68328953e-01 -1.69731174e-02
-2.12817237e-01 -7.50119388e-01 -1.08593464e+00 1.11494884e-01
-2.15753704e-01 -5.89643717e-01 -2.29558468e-01 1.17232883e+00
1.13715336e-01 -1.48711190e-01 5.52838862e-01 -1.42778957e+00
-8.60146165e-01 1.02396727e+00 -7.26420224e-01 6.33433521e-01
6.89874649e-01 -3.20007920e-01 6.28946006e-01 -8.08469117e-01
5.30672610e-01 1.50985253e+00 7.81750619e-01 -1.06312810e-02
-6.55714154e-01 -8.07017267e-01 -3.37054819e-01 4.72621858e-01
-1.30671585e+00 -1.01333670e-03 6.48477852e-01 -6.70181513e-01
5.17216921e-01 6.37409985e-01 8.13533247e-01 8.65129113e-01
7.51873195e-01 3.72233182e-01 1.76624429e+00 -9.82524753e-02
-6.77540824e-02 8.63252699e-01 6.15223825e-01 4.61125225e-01
3.60192060e-01 4.71493192e-02 -4.99757141e-01 -5.56859910e-01
6.12097025e-01 1.89661354e-01 -4.51685458e-01 6.21470511e-01
-9.02527452e-01 1.39671934e+00 4.05548453e-01 5.62539458e-01
-6.51669562e-01 -7.97484398e-01 3.83623809e-01 8.70238900e-01
7.29909360e-01 4.23337847e-01 6.27304986e-02 -3.15566286e-02
-9.83891606e-01 3.03658068e-01 1.13373911e+00 2.44452819e-01
6.52093410e-01 -7.11339116e-02 1.60090148e-01 7.96025753e-01
3.58414322e-01 5.36399961e-01 9.37238514e-01 -1.80976421e-01
-8.96896329e-03 7.23619282e-01 4.40933943e-01 -1.74609601e+00
-5.42185843e-01 -4.99352396e-01 -7.23653018e-01 9.28951278e-02
7.11406991e-02 -1.93818107e-01 -3.58755738e-01 1.21516371e+00
2.54387677e-01 -1.04223423e-01 -2.99665302e-01 1.17517328e+00
9.75257337e-01 9.19813991e-01 -4.83779579e-01 -7.69266546e-01
1.60694957e+00 -6.10678792e-01 -1.34398270e+00 6.52626678e-02
1.01589188e-01 -1.17653847e+00 6.87562883e-01 4.16913480e-01
-9.37903225e-01 -1.80624813e-01 -9.88051414e-01 4.26353097e-01
-2.69976854e-01 -5.42221904e-01 4.75743353e-01 6.84471726e-01
-7.23426163e-01 4.92735863e-01 -8.86860415e-02 -4.56869066e-01
-3.79121006e-02 -3.09925616e-01 5.16819619e-02 1.78130716e-01
-1.71574628e+00 1.16850567e+00 3.25631469e-01 -4.34297025e-02
-9.59987938e-01 -3.22709531e-01 4.70166430e-02 -1.58154771e-01
8.40579495e-02 -4.90220517e-01 9.40900147e-01 -1.11021030e+00
-1.11168039e+00 5.53041339e-01 9.13143679e-02 -4.20006126e-01
7.58395433e-01 -2.97120005e-01 -6.65674210e-01 8.72033238e-02
2.35257342e-01 -5.07038653e-01 1.08573866e+00 -1.35950434e+00
-7.15523660e-01 -5.93094110e-01 -2.26487473e-01 1.71050534e-01
-2.14778469e-03 6.76334321e-01 5.23361981e-01 -3.92604023e-01
5.16619265e-01 -5.68644822e-01 3.30324262e-01 -6.23166978e-01
-5.11144280e-01 -1.45646602e-01 9.74740803e-01 -8.51217866e-01
1.35465860e+00 -1.78121090e+00 -5.85193217e-01 1.05274506e-01
5.72753251e-01 1.83657601e-01 7.13814437e-01 8.62539589e-01
1.24820650e-01 2.20546722e-01 2.35132083e-01 3.74791235e-01
-3.74220759e-01 6.58054203e-02 -7.03821719e-01 8.45791996e-01
-3.15306932e-01 2.24376738e-01 -9.22137678e-01 -3.79123479e-01
-3.55373062e-02 2.39028826e-01 -1.21124163e-01 4.30178672e-01
3.90442550e-01 1.85687259e-01 -4.57310349e-01 4.75227058e-01
1.12433851e+00 -3.34785759e-01 1.49717391e-01 7.88858607e-02
-6.28550649e-01 5.85432768e-01 -7.87390292e-01 3.86242904e-02
-1.07192494e-01 6.69543207e-01 2.41691858e-01 -4.90043581e-01
1.37133467e+00 3.90765041e-01 9.33246017e-02 -3.29322278e-01
5.11079013e-01 2.04649389e-01 -1.64744392e-01 -6.63003087e-01
1.01407826e+00 -5.84776044e-01 3.04049104e-01 8.58039558e-01
-4.89204139e-01 1.70150608e-01 -3.27273101e-01 3.87703061e-01
9.30330694e-01 -8.37879002e-01 7.34679699e-01 -4.77528602e-01
3.93156946e-01 1.09785080e-01 2.29911983e-01 7.11930931e-01
-2.22120285e-01 1.79294378e-01 8.63614500e-01 -7.97797069e-02
-7.87955344e-01 -9.03422177e-01 -3.58966410e-01 9.80642855e-01
4.17262912e-01 2.21942868e-02 -5.75309992e-01 -4.03330177e-01
2.02171847e-01 9.55873549e-01 -7.33749032e-01 1.17713548e-01
-1.09057784e-01 -1.08390343e+00 6.25749528e-01 -2.07767934e-01
7.16761172e-01 -1.03546607e+00 -3.07322085e-01 -4.88758646e-03
-4.30862427e-01 -5.37226975e-01 -4.70985174e-02 -3.29864740e-01
-8.85910630e-01 -1.10903156e+00 -3.09766680e-01 -4.65773821e-01
6.84977472e-01 8.62902462e-01 7.31821358e-01 5.20157099e-01
2.80797005e-01 -3.02729458e-01 -7.40539551e-01 -3.48215133e-01
-9.13257658e-01 -4.80614036e-01 1.00971363e-01 2.70180497e-02
2.50465959e-01 -4.42336261e-01 -2.70756871e-01 7.59771347e-01
-6.77505612e-01 -5.57824783e-02 5.40141404e-01 7.31841326e-01
-3.51424932e-01 2.40677461e-01 9.76437211e-01 -9.53904867e-01
1.43287325e+00 -1.29750824e+00 -1.17563523e-01 -5.33326387e-01
-1.08711433e+00 -4.54550415e-01 3.63257438e-01 -2.51741469e-01
-1.18344975e+00 -1.18704474e+00 1.01945877e-01 -3.32502204e-05
9.75332558e-02 9.72727597e-01 5.22211373e-01 1.65336341e-01
7.32257664e-01 3.10105771e-01 6.95007682e-01 -6.65668190e-01
-2.62981299e-02 1.35372794e+00 8.55584964e-02 2.28639185e-01
4.64403242e-01 5.03359318e-01 -4.80209112e-01 -1.21852279e+00
-7.35212922e-01 -7.06767738e-01 -4.13439758e-02 -4.08774823e-01
3.41446698e-01 -5.57321370e-01 -7.82406032e-01 6.11189902e-01
-1.23989356e+00 5.23722947e-01 3.38855386e-01 7.29349136e-01
-1.06742889e-01 6.08490407e-01 -1.05753624e+00 -1.14145589e+00
-5.29737949e-01 -6.02446556e-01 -2.43236646e-01 7.91434795e-02
-4.10073727e-01 -9.00451183e-01 1.19541846e-01 5.96968234e-01
7.68643916e-01 1.06400728e-01 7.06606448e-01 -1.00521088e+00
-1.62191689e-01 -3.51543605e-01 -6.11736953e-01 2.50648737e-01
2.62599349e-01 -5.94698973e-02 -6.90822661e-01 -1.51659399e-01
9.44069743e-01 -5.74168824e-02 5.13207197e-01 -7.13583501e-03
2.18521506e-01 -1.03019404e+00 -1.64452389e-01 -1.28075406e-01
1.05279100e+00 -5.07218055e-02 7.11847603e-01 5.63113391e-01
2.28646040e-01 6.59854770e-01 9.22691107e-01 9.65307593e-01
2.10514382e-01 2.37635687e-01 8.04324687e-01 3.39996219e-01
1.02078885e-01 -4.19896454e-01 6.58133268e-01 1.34015083e+00
-2.71098763e-01 -1.06589636e-02 -7.88585305e-01 5.19497871e-01
-1.43596578e+00 -1.41090310e+00 -8.84438932e-01 2.08918381e+00
7.51804888e-01 5.60930893e-02 1.61220253e-01 2.10261822e-01
1.09989643e+00 4.14602697e-01 1.24568634e-01 -6.37574315e-01
-3.23907919e-02 -7.10819006e-01 3.73916149e-01 6.84119701e-01
-4.77946937e-01 3.29891145e-01 6.66035461e+00 3.24596286e-01
-1.37939060e+00 2.60072768e-01 2.14216456e-01 2.49990135e-01
-2.53027916e-01 9.55263898e-02 -6.40069723e-01 7.38741159e-01
9.05177891e-01 -5.68243444e-01 3.24719995e-01 6.46363914e-01
8.18959773e-01 -3.18306059e-01 -2.99999237e-01 5.40214360e-01
2.81880945e-01 -9.59318817e-01 1.81868419e-01 1.90755695e-01
5.70914149e-01 9.96119976e-02 -7.73774981e-02 2.04356000e-01
1.41004920e-01 -9.44406807e-01 6.82764232e-01 5.27545989e-01
-9.11037531e-03 -5.78863263e-01 1.33602738e+00 8.08486581e-01
-2.58433521e-01 -7.61679113e-02 -6.36089385e-01 -6.99509740e-01
3.20151806e-01 9.76267695e-01 -1.56812930e+00 4.43724126e-01
3.97263736e-01 5.35004616e-01 -2.90381104e-01 8.68526280e-01
-1.29367173e-01 9.84037042e-01 4.48614033e-03 -4.06309634e-01
1.55809194e-01 -1.48000941e-01 8.29066038e-01 1.14323509e+00
1.60030454e-01 -5.73492497e-02 -1.57953292e-01 1.12177896e+00
9.24430192e-02 4.28390384e-01 -3.65405262e-01 -1.80416703e-01
8.79446983e-01 1.01831460e+00 -5.18938839e-01 -2.97604084e-01
-1.20879322e-01 4.42538440e-01 -1.40193507e-01 -1.63015649e-01
-9.10729885e-01 -8.39757100e-02 2.47024983e-01 7.42922246e-01
-3.58198695e-02 1.71410203e-01 -5.05992532e-01 -7.67336071e-01
-5.64608216e-01 -9.02124166e-01 6.82352930e-02 -6.52449965e-01
-1.56577158e+00 5.39232492e-01 1.48223983e-02 -1.03176200e+00
-5.14776967e-02 6.12165732e-03 -7.58792937e-01 1.06703055e+00
-1.26147759e+00 -6.62937820e-01 -6.47396803e-01 5.61467707e-01
2.36263886e-01 -1.08412676e-01 4.82927293e-01 1.80704638e-01
-3.47471476e-01 1.86132461e-01 9.92228314e-02 -7.30033442e-02
5.32987297e-01 -7.24787772e-01 -1.94558099e-01 2.57045925e-01
-7.15592086e-01 9.92788792e-01 1.42604208e+00 -1.04489601e+00
-1.10233688e+00 -6.29588306e-01 1.06683576e+00 -2.49454916e-01
9.91118908e-01 3.50731492e-01 -1.15682292e+00 5.36242902e-01
2.74030954e-01 -5.85247576e-01 3.93942088e-01 3.24677497e-01
-3.19932431e-01 3.63937795e-01 -1.54022717e+00 2.34720618e-01
2.42224187e-01 -4.24894482e-01 -1.20388591e+00 2.91459292e-01
2.95096934e-01 2.19869494e-01 -8.46039295e-01 -3.62493843e-02
6.27203882e-01 -1.58845544e+00 5.01750410e-01 -3.78614128e-01
7.38017976e-01 -4.23540622e-02 1.46907970e-01 -1.17028666e+00
-5.21114111e-01 -3.66674036e-01 8.11366513e-02 1.12961841e+00
1.66125163e-01 -1.10620213e+00 2.76132196e-01 -3.18428874e-01
1.31066516e-01 -5.23337305e-01 -6.17774844e-01 -5.89853227e-01
-5.84766567e-02 1.57066435e-01 5.53917527e-01 1.66511023e+00
5.39614260e-01 4.81680065e-01 -7.23598540e-01 6.12498000e-02
5.14805555e-01 1.03336692e-01 7.46357799e-01 -1.26398551e+00
4.24004719e-02 -4.17380333e-01 -2.93557465e-01 -8.48566055e-01
-4.86313611e-01 -6.08112037e-01 -1.84009507e-01 -1.32597542e+00
6.17088199e-01 -2.76160181e-01 -1.71817541e-01 -6.02080207e-03
-1.92218199e-01 -1.59428909e-01 -5.83280763e-03 1.08660388e+00
9.12026912e-02 4.62207675e-01 1.27979827e+00 2.91472912e-01
-1.93051666e-01 1.50617495e-01 -6.82281971e-01 1.01532292e+00
1.05086601e+00 -6.22042239e-01 -2.51695246e-01 4.30505663e-01
4.37171578e-01 5.30973971e-01 3.78254563e-01 -3.91266733e-01
-6.02862425e-02 -5.65290451e-01 -2.99262583e-01 -7.88149059e-01
6.75225779e-02 -3.20531398e-01 1.19817719e-01 6.16575837e-01
-1.34929016e-01 2.03876257e-01 -3.09877038e-01 4.82118338e-01
-2.83172280e-01 -5.03464162e-01 7.44203925e-01 -2.46204495e-01
-2.54199028e-01 -3.83644581e-01 -7.74304748e-01 -1.02324747e-01
7.27261543e-01 -7.90292770e-02 -8.77859890e-01 -8.19343984e-01
-3.52984369e-01 -3.26256394e-01 4.62305218e-01 4.42928910e-01
7.99062788e-01 -9.87598240e-01 -1.07716274e+00 6.26270100e-02
-2.95189142e-01 -8.57431054e-01 2.44953826e-01 1.56112838e+00
-6.23136520e-01 2.46533200e-01 -5.98518491e-01 -1.44848466e-01
-1.36058402e+00 4.09048051e-01 4.19851318e-02 -1.80693008e-02
-3.67069066e-01 3.63056391e-01 -5.74621521e-02 -2.16653556e-01
-3.00056040e-01 2.78391898e-01 -6.28404558e-01 4.47711974e-01
1.00544214e+00 9.80865121e-01 -2.56519794e-01 -9.37156081e-01
-5.31541109e-02 -4.01175439e-01 -5.20907164e-01 1.32829100e-01
1.00902176e+00 -6.50401175e-01 -9.84508574e-01 8.04154992e-01
9.20661092e-01 5.07323742e-01 -2.28747204e-01 -5.10337986e-02
-2.25413255e-02 -8.76777232e-01 3.59239429e-01 -9.14641142e-01
-5.45636237e-01 3.36811602e-01 1.95348904e-01 1.14488947e+00
4.99699980e-01 -6.43280447e-02 6.63748264e-01 -1.66719228e-01
3.82550329e-01 -7.97861397e-01 -3.32965516e-02 6.28072202e-01
1.13737166e+00 -9.96874273e-01 3.85532349e-01 -6.13035440e-01
-6.88861907e-01 7.51276016e-01 2.10265373e-03 -2.71558672e-01
7.52911985e-01 -8.92046690e-02 1.72680020e-02 -6.74281120e-01
-6.53262436e-01 3.61958504e-01 -4.06245887e-02 2.01742813e-01
7.74035692e-01 3.92605513e-01 -1.34548116e+00 7.24687576e-01
-6.30604327e-01 2.23860666e-01 1.13651657e+00 8.30203593e-01
-1.36904442e+00 -2.08146110e-01 -1.02200913e+00 7.01632082e-01
-7.87750423e-01 1.88985560e-02 -4.85849172e-01 8.31131160e-01
-1.49320751e-01 1.34939909e+00 -1.29416287e-01 -7.72122920e-01
-2.88225353e-01 -2.06482992e-01 -6.66865632e-02 -3.63084048e-01
-6.79999828e-01 -3.68513256e-01 5.45473635e-01 -2.21106589e-01
-2.43364319e-01 -5.30056655e-01 -1.12819815e+00 -1.34387636e+00
-8.65152121e-01 6.58618927e-01 7.83584476e-01 1.04752254e+00
-6.82067275e-02 2.21008912e-01 1.08840525e+00 -2.74692833e-01
-1.19273663e+00 -1.51093268e+00 -1.13519549e+00 5.91967225e-01
3.28590333e-01 -8.33984017e-01 -9.97281969e-01 -4.99906868e-01] | [8.241090774536133, 10.15260124206543] |
67e23011-1320-43ac-8dcc-459eb3f90937 | dereverberation-in-acoustic-sensor-networks | 2301.07649 | null | https://arxiv.org/abs/2301.07649v1 | https://arxiv.org/pdf/2301.07649v1.pdf | Dereverberation in Acoustic Sensor Networks Using Weighted Prediction Error With Microphone-dependent Prediction Delays | In the last decades several multi-microphone speech dereverberation algorithms have been proposed, among which the weighted prediction error (WPE) algorithm. In the WPE algorithm, a prediction delay is required to reduce the correlation between the prediction signals and the direct component in the reference microphone signal. In compact arrays with closely-spaced microphones, the prediction delay is often chosen microphone-independent. In acoustic sensor networks with spatially distributed microphones, large time-differences-of-arrival (TDOAs) of the speech source between the reference microphone and other microphones typically occur. Hence, when using a microphone-independent prediction delay the reference and prediction signals may still be significantly correlated, leading to distortion in the dereverberated output signal. In order to decorrelate the signals, in this paper we propose to apply TDOA compensation with respect to the reference microphone, resulting in microphone-dependent prediction delays for the WPE algorithm. We consider both optimal TDOA compensation using crossband filtering in the short-time Fourier transform domain as well as band-to-band and integer delay approximations. Simulation results for different reverberation times using oracle as well as estimated TDOAs clearly show the benefit of using microphone-dependent prediction delays. | ['Simon Doclo', 'Joerg Bitzer', 'Toon van Waterschoot', 'Anselm Lohmann'] | 2023-01-18 | null | null | null | null | ['speech-dereverberation'] | ['speech'] | [ 3.26696247e-01 -2.38018334e-01 6.93073809e-01 9.28558186e-02
-8.12438786e-01 -5.08935630e-01 4.01450172e-02 2.44853079e-01
-2.99637675e-01 4.81066585e-01 3.52776617e-01 -1.14970222e-01
-4.87286001e-01 -4.61349726e-01 -4.85810310e-01 -8.58102918e-01
-3.20778161e-01 -3.66802871e-01 1.66825548e-01 2.51933962e-01
-1.78648546e-01 2.61765420e-01 -1.48694050e+00 -1.14224009e-01
8.89564335e-01 1.27980864e+00 3.70249301e-01 8.85636687e-01
1.52710184e-01 2.82713473e-01 -1.33674741e+00 1.81312747e-02
5.06542981e-01 -4.71226782e-01 5.52005708e-01 -2.24182740e-01
6.65985346e-02 -7.01334849e-02 -2.60952741e-01 1.04067314e+00
8.88361037e-01 4.37583178e-01 3.66343975e-01 -7.62037218e-01
-7.02460930e-02 5.29009640e-01 -4.89838183e-01 9.27827582e-02
2.65939415e-01 -3.37608874e-01 6.27564669e-01 -1.11218750e+00
-1.46226048e-01 1.01914716e+00 9.09903526e-01 1.16129167e-01
-1.08146906e+00 -8.30290496e-01 -3.68647307e-01 3.29586565e-01
-1.57570541e+00 -7.48983920e-01 1.11076665e+00 -2.86555082e-01
6.04064167e-01 4.47588891e-01 3.48539919e-01 5.84393442e-01
1.12158507e-02 1.96600288e-01 1.07652986e+00 -8.00530910e-01
2.75926858e-01 -1.95933968e-01 -8.93554240e-02 -9.51909646e-02
6.76708817e-02 5.49375653e-01 -5.40771604e-01 -4.43752855e-01
4.13623720e-01 -2.90958762e-01 -7.27897346e-01 1.09669708e-01
-1.16868806e+00 2.42232844e-01 2.47408673e-01 4.68208730e-01
-5.47723413e-01 1.08657070e-01 -2.26999409e-02 4.30957168e-01
6.16425216e-01 3.13911378e-01 -1.72022924e-01 1.28577158e-01
-1.12073386e+00 1.24967601e-02 8.55959117e-01 5.84978402e-01
3.91020447e-01 3.59896839e-01 -1.65791526e-01 1.17771673e+00
5.43982029e-01 9.33473766e-01 7.05949962e-02 -8.65360141e-01
4.15704519e-01 -6.51915073e-01 5.01740396e-01 -1.15737963e+00
-1.90673620e-01 -1.10674727e+00 -9.21008527e-01 1.10820271e-01
7.69834697e-01 -5.72754025e-01 -3.92394900e-01 1.61018324e+00
4.65189308e-01 5.54560244e-01 2.51932025e-01 8.93600345e-01
8.22728723e-02 9.85557318e-01 -4.76497769e-01 -9.32658851e-01
8.30963492e-01 -9.77264345e-01 -1.16488469e+00 -1.63290411e-01
-4.30031158e-02 -1.26410520e+00 2.70883709e-01 8.58187914e-01
-1.11465776e+00 -6.76654398e-01 -1.17323291e+00 5.32549858e-01
1.07689776e-01 -9.58567709e-02 -5.55299640e-01 9.17353690e-01
-9.18287873e-01 2.97777563e-01 -6.63915455e-01 3.36696684e-01
-4.81336772e-01 -5.61387576e-02 1.48140728e-01 -8.64890739e-02
-9.05771136e-01 5.14005601e-01 -4.68458802e-01 4.29929405e-01
-5.45232892e-01 -1.02932787e+00 -2.46504977e-01 9.10615772e-02
2.38994986e-01 -2.08372697e-01 1.27088511e+00 -1.06267726e+00
-1.41688240e+00 -2.10743621e-01 -6.63932085e-01 -4.82720524e-01
2.60499716e-01 -1.29813462e-01 -1.21602833e+00 3.11387032e-02
-3.23092751e-03 -3.23041320e-01 1.13105774e+00 -1.16346443e+00
-6.34460390e-01 -1.72128066e-01 -5.53253233e-01 2.24317014e-01
-1.76360235e-01 -2.56735861e-01 1.06156036e-01 -1.06480336e+00
5.78123927e-01 -6.02032721e-01 -4.00622606e-01 -7.08239600e-02
-3.12503278e-01 2.40101665e-01 5.45794070e-01 -7.84399271e-01
1.37363386e+00 -2.81007862e+00 -2.25750104e-01 2.90297002e-01
-6.78473189e-02 3.37750196e-01 -2.15118691e-01 7.55550921e-01
-9.19670090e-02 -5.50801873e-01 4.48976047e-02 -3.69322389e-01
-3.45810324e-01 -1.88734569e-02 -3.15948337e-01 5.99821031e-01
-2.49313816e-01 -2.35603169e-01 -9.42178309e-01 5.67836463e-02
1.77267924e-01 6.78609431e-01 -3.31814349e-01 6.13032520e-01
3.12054664e-01 6.47872865e-01 -3.57475178e-03 2.84111142e-01
9.53185976e-01 5.88548541e-01 -2.01469690e-01 -4.23886418e-01
-6.92911923e-01 2.10558236e-01 -1.44166636e+00 1.13559043e+00
-9.62723911e-01 7.13112235e-01 8.86175871e-01 -7.61807501e-01
1.12178791e+00 7.28301644e-01 2.84497917e-01 -6.29267871e-01
-3.86894524e-01 8.80467474e-01 2.81871676e-01 -3.13775539e-01
6.11744784e-02 -1.68223605e-01 5.94648659e-01 1.09237805e-01
-2.08728462e-01 1.03433639e-01 -1.72397103e-02 -6.05250299e-01
1.01330256e+00 -5.68123102e-01 3.94544423e-01 -2.02603266e-01
7.81893790e-01 -6.45518243e-01 7.23200262e-01 4.93010640e-01
-4.28471193e-02 6.42704129e-01 -2.31680334e-01 2.17046559e-01
-7.02253044e-01 -1.03854251e+00 1.04167178e-01 8.05229068e-01
2.14343332e-02 -8.91300589e-02 -6.72987580e-01 1.39726028e-01
-2.21782938e-01 9.48213995e-01 1.51105314e-01 1.33026704e-01
-6.40735269e-01 -6.30429089e-02 4.77837831e-01 4.08362448e-02
3.52247208e-01 -2.62160804e-02 -5.27652085e-01 8.53720903e-01
-2.58990843e-02 -8.88523817e-01 -7.41009295e-01 5.04456878e-01
-6.10655308e-01 -7.36469686e-01 -1.22804189e+00 -4.86168891e-01
6.75309420e-01 7.84289420e-01 3.21806818e-01 -3.38986307e-01
3.02194804e-01 6.54985189e-01 -3.51295739e-01 -6.45064116e-01
-2.44962692e-01 -8.28708410e-01 2.16371611e-01 4.86018032e-01
-1.80843666e-01 -9.41679239e-01 -7.60317028e-01 7.13337660e-01
-5.69610953e-01 -2.09685698e-01 3.93971890e-01 7.50909328e-01
3.43476862e-01 6.88835382e-01 5.81755400e-01 6.98264539e-02
8.39364052e-01 -3.88627559e-01 -8.00517619e-01 7.77599886e-02
-4.62360859e-01 -4.81249899e-01 1.06386364e+00 -6.97264135e-01
-1.22848094e+00 -2.31174886e-01 -3.86237562e-01 -3.71824026e-01
-5.68644851e-02 5.40002704e-01 -2.49605119e-01 -3.42376605e-02
5.75541556e-01 3.51201892e-01 -2.22733423e-01 -6.41745567e-01
4.03822288e-02 7.64133692e-01 5.86111009e-01 -4.28907387e-02
8.72480333e-01 1.84394494e-01 1.38065726e-01 -1.34975374e+00
-2.48830497e-01 -6.04632795e-01 7.40173385e-02 -2.94000000e-01
3.54243040e-01 -7.87734389e-01 -1.64751083e-01 7.43594229e-01
-1.51793218e+00 -4.18784432e-02 -1.82365239e-01 1.08693659e+00
3.72841470e-02 2.66852826e-01 -1.44913644e-01 -1.34310091e+00
-3.17848384e-01 -9.49114919e-01 4.22775030e-01 2.57145822e-01
-2.85735369e-01 -7.89903522e-01 6.71383515e-02 -1.87141940e-01
7.74766326e-01 5.13123255e-03 6.98617816e-01 -5.41973054e-01
-2.84133613e-01 -3.27914476e-01 3.50792438e-01 6.69802010e-01
2.86710948e-01 -2.19773844e-01 -9.98376548e-01 -7.55950212e-02
6.27590179e-01 7.46148467e-01 1.85970694e-01 1.03497553e+00
4.11939174e-01 -4.70043480e-01 -3.17360967e-01 3.08002591e-01
1.32931828e+00 7.19812632e-01 3.27528238e-01 -2.69528478e-01
1.67026401e-01 6.45717204e-01 7.22987235e-01 4.54904884e-01
-1.14289463e-01 5.42790473e-01 1.63845450e-01 -5.54081239e-02
-3.10091436e-01 -5.73911406e-02 3.18042994e-01 1.24791753e+00
3.22169155e-01 -7.15168417e-01 -4.91505861e-01 6.84090734e-01
-1.16232872e+00 -5.87936759e-01 -5.79869092e-01 2.66232705e+00
6.04158580e-01 -3.31554949e-01 -1.51290193e-01 6.66312695e-01
8.08023989e-01 2.95650542e-01 -5.06596565e-01 -3.48706037e-01
-1.03050508e-01 3.67398024e-01 8.08857739e-01 8.74603093e-01
-4.31552023e-01 -2.65284687e-01 5.22822475e+00 1.07312202e+00
-1.30854213e+00 4.94882494e-01 1.33653790e-01 -5.77266775e-02
-3.66143852e-01 -3.44132543e-01 -3.61376971e-01 4.13566828e-01
1.03119838e+00 -5.24474800e-01 6.67232573e-01 3.41140598e-01
7.74300992e-01 -3.93566370e-01 -9.99406636e-01 1.02049255e+00
-1.78248659e-01 -5.90391040e-01 -9.04992104e-01 -1.05683081e-01
4.65946525e-01 -2.37804741e-01 2.07932711e-01 -2.82549322e-01
-4.70844716e-01 -7.60619104e-01 8.96499097e-01 4.93994743e-01
4.11580950e-01 -5.91682911e-01 3.70057344e-01 4.98581558e-01
-1.34225225e+00 -1.88451126e-01 -1.74496755e-01 -1.52744412e-01
2.92869449e-01 1.58896542e+00 -8.87447536e-01 5.12707531e-01
4.11279291e-01 8.30374807e-02 4.55368131e-01 1.65427148e+00
-3.73284400e-01 9.34107840e-01 -6.96014047e-01 -1.06266193e-01
-1.78492799e-01 -2.35077709e-01 1.10044885e+00 9.23950076e-01
1.18154490e+00 3.20669800e-01 -2.95297623e-01 6.09732628e-01
1.82314828e-01 -9.81520712e-02 -1.75586313e-01 2.78358400e-01
1.02308774e+00 8.30170512e-01 -2.90468425e-01 5.17054461e-02
-4.16417509e-01 6.70249104e-01 -8.34098220e-01 9.12732661e-01
-6.15002096e-01 -7.48155236e-01 8.25678349e-01 1.52791217e-01
3.65136951e-01 -5.30397892e-01 -2.36038398e-02 -2.79674470e-01
2.82274216e-01 -6.71307147e-01 -3.11499715e-01 -7.11071432e-01
-1.08207369e+00 4.91650432e-01 -4.21872854e-01 -1.61627591e+00
-5.39252646e-02 -2.07821697e-01 -6.97580874e-01 1.38568103e+00
-1.30905592e+00 -4.39644635e-01 2.81665504e-01 4.91207361e-01
5.48425615e-01 2.72384226e-01 9.50215280e-01 7.42267430e-01
-1.10566951e-01 4.90756571e-01 8.14407051e-01 -3.59513193e-01
7.78824329e-01 -8.57595921e-01 -1.10793538e-01 9.63604748e-01
-1.18527070e-01 6.97802961e-01 1.40281856e+00 -3.80671650e-01
-1.25266111e+00 -9.04405534e-01 9.78415787e-01 4.66981024e-01
4.53236967e-01 -3.49289179e-01 -7.40889788e-01 -1.22664966e-01
4.64890122e-01 1.37072533e-01 1.04956615e+00 -3.77766341e-01
-1.36074394e-01 -8.14667583e-01 -1.01951385e+00 4.57201779e-01
4.77832347e-01 -4.00964797e-01 -4.69901055e-01 1.11379206e-01
5.91004312e-01 -3.17184985e-01 -5.99036694e-01 1.98522359e-01
6.95396781e-01 -1.11031640e+00 9.67515528e-01 7.70944655e-01
-1.61621988e-01 -7.28710771e-01 -4.67475504e-01 -1.84493172e+00
-6.22773655e-02 -9.58842337e-01 -2.77799200e-02 1.47008336e+00
4.63671863e-01 -9.78848398e-01 1.96255639e-01 1.35499969e-01
-1.14784680e-01 -4.17399585e-01 -1.16072249e+00 -1.17098522e+00
-3.42607409e-01 -6.76364541e-01 3.09383184e-01 2.84347445e-01
-1.39301000e-02 1.75766230e-01 -4.86376613e-01 1.05836272e+00
8.90724421e-01 -2.16629907e-01 4.26603347e-01 -9.71177697e-01
-6.50481224e-01 -1.70608371e-01 2.38854155e-01 -1.79180300e+00
-4.34850127e-01 -8.44511464e-02 4.61661607e-01 -1.36840332e+00
-9.92751122e-01 -4.32357550e-01 -3.04059893e-01 -4.13610458e-01
-3.03411465e-02 -1.39567897e-01 5.46578765e-02 -2.36721650e-01
1.08237304e-01 4.90192145e-01 8.90829027e-01 -6.50549531e-02
-3.05728257e-01 8.08906376e-01 -1.20662533e-01 7.28150010e-01
3.07521790e-01 -6.83378875e-01 -5.01195908e-01 -6.73726678e-01
-2.58513186e-02 5.94430625e-01 9.55245122e-02 -1.47489893e+00
7.40596473e-01 2.65944839e-01 -5.22116348e-02 -8.91280651e-01
6.54601812e-01 -1.44366920e+00 8.72229397e-01 5.17876744e-01
-2.58317381e-01 -5.41315615e-01 9.85569041e-03 8.96101773e-01
-5.23905873e-01 -1.35515273e-01 6.18721604e-01 2.64709294e-01
-8.69432762e-02 -2.98856407e-01 -8.71468186e-01 -4.68466043e-01
7.53140390e-01 -1.96174502e-01 -2.78322082e-02 -7.62996078e-01
-5.08356392e-01 -5.30539900e-02 -2.85193205e-01 -1.55999642e-02
5.50930798e-01 -1.07427132e+00 -5.69301605e-01 2.08708540e-01
-3.70304644e-01 -1.67316109e-01 5.34533203e-01 1.11399281e+00
-5.54358624e-02 5.48104882e-01 2.60117441e-01 -5.67863762e-01
-1.41646051e+00 2.55830318e-01 5.15249610e-01 2.05399960e-01
1.91972032e-01 1.20834160e+00 1.39017940e-01 1.13794357e-01
5.12941241e-01 -4.22798336e-01 7.97629170e-03 -3.84285003e-02
8.35518718e-01 8.79266083e-01 1.44762486e-01 -4.24313933e-01
-1.61935523e-01 8.31038594e-01 8.15119684e-01 -6.31991029e-01
9.31393206e-01 -7.18256891e-01 -1.91890195e-01 5.70282578e-01
1.34935498e+00 1.06434822e+00 -9.22730565e-01 -3.89259338e-01
-9.18089077e-02 -6.98555827e-01 2.15566978e-01 -8.38580310e-01
-8.29566002e-01 9.55249727e-01 8.95775437e-01 5.99769831e-01
1.70776546e+00 -4.78874803e-01 9.26078796e-01 1.22556994e-02
6.25663221e-01 -8.13803792e-01 -1.37910545e-01 2.91633606e-01
7.96115458e-01 -3.79208446e-01 -3.46713543e-01 -4.64224547e-01
1.85071334e-01 1.18568480e+00 8.61759707e-02 -6.18864782e-02
7.96761155e-01 3.50927681e-01 4.09097761e-01 5.95558286e-01
-3.38017225e-01 1.53452173e-01 1.14352256e-01 7.33158827e-01
4.91100371e-01 1.61801219e-01 -5.78339875e-01 8.09037566e-01
-1.95309103e-01 -3.87939513e-01 3.97984952e-01 3.65203977e-01
-4.87198770e-01 -1.27928567e+00 -1.23379183e+00 1.72973752e-01
-5.60211897e-01 -2.85726458e-01 1.25696603e-02 -1.66876331e-01
1.99443653e-01 1.80877542e+00 8.76193792e-02 -3.27236593e-01
9.89882708e-01 -6.08398691e-02 1.52687490e-01 -2.85645008e-01
-4.92543817e-01 1.02388263e+00 2.15522736e-01 -2.05110356e-01
-2.74551213e-01 -5.42109549e-01 -1.13696694e+00 -1.18962355e-01
-5.38056016e-01 5.20730019e-01 8.23273420e-01 5.88800848e-01
1.88178927e-01 8.10288310e-01 1.06053936e+00 -7.47600019e-01
-5.82342982e-01 -8.78506899e-01 -9.71595585e-01 -3.49046081e-01
9.32766080e-01 -1.91594630e-01 -7.89138794e-01 -2.34123051e-01] | [15.157439231872559, 5.740784645080566] |
d39480e6-be2c-4594-af9b-578d27e77e20 | high-resolution-photorealistic-image | 2105.09188 | null | https://arxiv.org/abs/2105.09188v1 | https://arxiv.org/pdf/2105.09188v1.pdf | High-Resolution Photorealistic Image Translation in Real-Time: A Laplacian Pyramid Translation Network | Existing image-to-image translation (I2IT) methods are either constrained to low-resolution images or long inference time due to their heavy computational burden on the convolution of high-resolution feature maps. In this paper, we focus on speeding-up the high-resolution photorealistic I2IT tasks based on closed-form Laplacian pyramid decomposition and reconstruction. Specifically, we reveal that the attribute transformations, such as illumination and color manipulation, relate more to the low-frequency component, while the content details can be adaptively refined on high-frequency components. We consequently propose a Laplacian Pyramid Translation Network (LPTN) to simultaneously perform these two tasks, where we design a lightweight network for translating the low-frequency component with reduced resolution and a progressive masking strategy to efficiently refine the high-frequency ones. Our model avoids most of the heavy computation consumed by processing high-resolution feature maps and faithfully preserves the image details. Extensive experimental results on various tasks demonstrate that the proposed method can translate 4K images in real-time using one normal GPU while achieving comparable transformation performance against existing methods. Datasets and codes are available: https://github.com/csjliang/LPTN. | ['Lei Zhang', 'Hui Zeng', 'Jie Liang'] | 2021-05-19 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Liang_High-Resolution_Photorealistic_Image_Translation_in_Real-Time_A_Laplacian_Pyramid_Translation_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Liang_High-Resolution_Photorealistic_Image_Translation_in_Real-Time_A_Laplacian_Pyramid_Translation_CVPR_2021_paper.pdf | cvpr-2021-1 | ['color-manipulation', 'photo-retouching'] | ['computer-vision', 'computer-vision'] | [ 4.68831390e-01 -3.98549557e-01 4.12690565e-02 5.43961441e-03
-8.08028519e-01 -4.05198395e-01 4.56893474e-01 -4.69615728e-01
-3.87163013e-01 4.55557406e-01 3.29942346e-01 -9.12466180e-03
-1.02817535e-01 -8.56079221e-01 -8.67506683e-01 -7.68365681e-01
4.57117766e-01 -2.21898165e-02 2.78244764e-01 -1.26715869e-01
2.59581238e-01 4.68863428e-01 -1.38249183e+00 3.69195610e-01
8.81152093e-01 1.00226450e+00 2.86566615e-01 5.81480920e-01
1.35798514e-01 5.87771893e-01 -3.14863548e-02 -2.47604400e-01
4.84802067e-01 -2.90515870e-01 -5.88093281e-01 1.19461276e-01
6.68605983e-01 -4.40378338e-01 -5.09936869e-01 1.31593537e+00
6.31196916e-01 1.01935156e-01 3.81130308e-01 -1.00741792e+00
-1.11163211e+00 1.46895409e-01 -1.00521672e+00 1.24223456e-01
1.99420109e-01 1.64753944e-01 6.01675212e-01 -1.34385431e+00
5.78961194e-01 1.27532780e+00 7.08250940e-01 1.65744841e-01
-1.45281208e+00 -6.97512805e-01 -5.29720001e-02 3.27919245e-01
-1.59892321e+00 -5.17541528e-01 9.95149851e-01 -2.68401235e-01
6.01908267e-01 3.94180357e-01 4.45327193e-01 6.80625975e-01
2.49235258e-01 3.40068847e-01 1.30303645e+00 -3.51673722e-01
-2.29001880e-01 -1.37923390e-01 -3.67072999e-01 8.27332199e-01
1.15252584e-01 1.05560243e-01 -7.45807767e-01 7.02561960e-02
1.42944503e+00 1.92721277e-01 -6.47136807e-01 -1.14152804e-01
-1.43899131e+00 5.26579320e-01 6.31667793e-01 3.34974766e-01
-4.94546324e-01 3.29213530e-01 1.02883630e-01 2.47436985e-01
6.42757893e-01 4.24479507e-02 -2.94277579e-01 1.46618843e-01
-7.84733951e-01 -8.68149996e-02 2.61366189e-01 9.66406643e-01
1.14171040e+00 -1.11069232e-01 -3.76774311e-01 8.48982096e-01
-3.45909111e-02 5.17351270e-01 4.03129905e-01 -1.13059807e+00
4.87771422e-01 3.26190859e-01 2.20380262e-01 -1.26866221e+00
-2.05209568e-01 -2.31145531e-01 -1.47500241e+00 1.60020202e-01
9.70900804e-02 9.36204866e-02 -8.02643001e-01 1.56642866e+00
4.61266845e-01 1.84442729e-01 -8.37482512e-02 1.03089988e+00
6.97440326e-01 8.61900628e-01 -2.88095623e-01 -1.97158009e-01
1.58609819e+00 -1.14133370e+00 -7.99393296e-01 1.02173597e-01
6.23613708e-02 -1.20724082e+00 1.42685914e+00 -4.79369238e-02
-1.28006864e+00 -7.75069416e-01 -8.39819670e-01 -7.39392102e-01
-3.03739049e-02 3.89189094e-01 4.72889125e-01 1.42664924e-01
-1.15722704e+00 6.83204710e-01 -7.67277658e-01 -1.48147225e-01
4.27670956e-01 3.36965829e-01 -3.66280764e-01 -3.03732425e-01
-1.00952470e+00 5.87291241e-01 1.06999345e-01 3.35460782e-01
-3.44073802e-01 -1.04498684e+00 -7.21446037e-01 5.85839711e-02
2.30235234e-01 -7.74250984e-01 8.33936155e-01 -7.14856207e-01
-1.71011579e+00 6.71501219e-01 -3.57334256e-01 6.92584291e-02
5.83177567e-01 -1.32488117e-01 -1.04189865e-01 2.81578213e-01
1.56311974e-01 6.83781981e-01 1.14869761e+00 -1.09680271e+00
-6.64580822e-01 -2.77833700e-01 -2.07819104e-01 4.28059727e-01
-4.96403843e-01 -3.00690476e-02 -8.01886559e-01 -8.11825871e-01
3.53346229e-01 -8.86817098e-01 6.30364716e-02 4.60419089e-01
-2.53495902e-01 1.27229139e-01 8.17125857e-01 -8.54074299e-01
9.90512133e-01 -2.33641815e+00 2.93911368e-01 -5.26820943e-02
4.11890835e-01 1.50020689e-01 -1.92556858e-01 2.76554465e-01
-2.73188837e-02 -1.11193225e-01 -1.84980705e-01 -2.39147529e-01
-1.52415648e-01 -5.97007386e-02 -4.43372965e-01 6.15077555e-01
1.38841435e-01 1.11112881e+00 -6.79207563e-01 -4.68259543e-01
3.07055652e-01 1.03865969e+00 -5.98672926e-01 1.05599865e-01
3.05861264e-01 7.16613173e-01 -3.06474358e-01 4.01916385e-01
9.93143022e-01 -3.47960502e-01 -3.02124411e-01 -9.44388509e-01
-4.92973357e-01 -5.75480759e-02 -8.60637069e-01 1.80359316e+00
-7.43207932e-01 5.33246934e-01 2.46606529e-01 -6.03836596e-01
7.94920504e-01 8.99753720e-02 3.80647689e-01 -9.08774197e-01
2.04795990e-02 1.86570346e-01 -2.63804644e-01 -2.87243068e-01
3.95490497e-01 -7.80590028e-02 1.06100336e-01 3.24931949e-01
-3.61185044e-01 -3.37534517e-01 2.76513603e-02 -5.74634336e-02
7.60023594e-01 2.37854689e-01 1.86960936e-01 -2.46779606e-01
7.72167623e-01 -2.24378765e-01 5.45970559e-01 3.68537992e-01
3.59914936e-02 8.30207586e-01 1.45489320e-01 -5.51901340e-01
-1.27253342e+00 -9.53128338e-01 2.53817476e-02 1.11868012e+00
5.07396519e-01 -1.98725909e-01 -7.20720410e-01 6.93459213e-02
-3.43694419e-01 3.04651946e-01 -5.23351133e-01 -1.19852372e-01
-7.44097650e-01 -6.96479857e-01 3.94874781e-01 2.91536599e-01
1.10018182e+00 -9.21562850e-01 -5.23660719e-01 5.18793575e-02
-4.77719337e-01 -1.30292106e+00 -1.17006624e+00 -2.07795739e-01
-7.51600385e-01 -7.35544384e-01 -7.98362672e-01 -1.04566133e+00
9.24297571e-01 6.36231244e-01 7.52011001e-01 1.78421214e-02
-4.21178520e-01 1.43978626e-01 2.21231617e-02 1.90502703e-01
7.35331625e-02 -1.36057109e-01 -8.63378793e-02 3.29990715e-01
-8.83173570e-02 -9.39937890e-01 -7.94759452e-01 3.75491917e-01
-1.00341749e+00 6.76027119e-01 8.73778164e-01 9.58236754e-01
1.03980958e+00 3.59661311e-01 -8.94217417e-02 -6.45396233e-01
5.19327343e-01 1.97543338e-01 -7.10867643e-01 2.86468327e-01
-3.81252497e-01 2.28075925e-02 8.79643679e-01 -7.35979795e-01
-1.29046726e+00 1.33026347e-01 9.55422297e-02 -6.21876895e-01
2.28242248e-01 9.49947909e-02 -6.27939776e-02 -6.12437069e-01
4.65418488e-01 6.77729666e-01 -1.67105541e-01 -5.00048280e-01
4.55042750e-01 2.47410819e-01 8.04619372e-01 -6.45862937e-01
1.30455327e+00 7.79414475e-01 1.39548138e-01 -7.66852856e-01
-7.81358123e-01 -8.42459649e-02 -6.83698297e-01 -5.70406802e-02
8.45043957e-01 -1.07234037e+00 -8.09653640e-01 5.76520801e-01
-1.02558279e+00 -3.55756611e-01 -2.77368665e-01 3.86795908e-01
-5.60069978e-01 4.03334290e-01 -8.67151141e-01 -2.36611426e-01
-6.82932734e-01 -1.22958815e+00 1.20227182e+00 2.98150212e-01
2.96326399e-01 -6.69289887e-01 -1.34418130e-01 3.10161680e-01
7.99513578e-01 4.88034636e-02 9.35826421e-01 4.48409289e-01
-9.60904241e-01 9.44683179e-02 -9.03573751e-01 1.09096669e-01
2.61602968e-01 -2.68015981e-01 -7.45593131e-01 -4.61482257e-01
1.11343227e-01 -2.28460729e-01 7.40463078e-01 2.86973119e-01
1.40469646e+00 -4.91745114e-01 4.43593338e-02 1.22868764e+00
1.52611923e+00 -2.99320012e-01 7.17269778e-01 1.69962168e-01
1.16332424e+00 3.01424801e-01 4.02638286e-01 2.30031803e-01
4.84536648e-01 9.84676301e-01 -7.46728033e-02 -5.59377432e-01
-6.92498446e-01 -3.31845671e-01 2.69380778e-01 1.05171084e+00
-3.56180042e-01 2.87109613e-01 -5.52765191e-01 2.47484446e-01
-1.75625026e+00 -7.47054636e-01 -5.09177595e-02 2.05990577e+00
1.14967537e+00 -2.71352082e-01 -4.22991514e-01 -3.94414663e-02
8.14148724e-01 2.22441852e-01 -5.89331985e-01 -2.18730550e-02
-1.39531434e-01 2.19359353e-01 6.21992052e-01 7.60972917e-01
-9.79581475e-01 9.80386615e-01 5.27831078e+00 1.11096263e+00
-1.24697888e+00 2.83742279e-01 7.11710632e-01 -1.72533870e-01
-2.74887323e-01 -2.12373048e-01 -5.01879811e-01 3.07465404e-01
3.87188435e-01 -4.10862535e-01 8.68306816e-01 4.09398586e-01
3.37018788e-01 1.15666263e-01 -7.92724431e-01 1.18990970e+00
-1.85646210e-02 -1.35027218e+00 2.79913396e-01 -5.11536002e-02
8.15760493e-01 -1.38597682e-01 3.60140830e-01 -1.47369042e-01
1.77170895e-02 -8.76720488e-01 7.04536796e-01 5.63866913e-01
1.36229312e+00 -6.73822641e-01 2.30275303e-01 9.71715376e-02
-1.52899361e+00 1.25826240e-01 -6.41425908e-01 1.50197819e-01
-1.02096550e-01 6.06620967e-01 -1.41568020e-01 4.62982774e-01
8.96474898e-01 7.50464916e-01 -3.48476708e-01 5.91059566e-01
-3.05732876e-01 1.46583021e-01 -3.55546594e-01 4.74715441e-01
7.32734352e-02 -5.22220135e-01 2.66233236e-01 9.01179016e-01
5.33629835e-01 4.02981579e-01 1.91163033e-01 9.41517115e-01
-2.15468958e-01 2.84460038e-02 -3.14076036e-01 3.31262678e-01
3.93379122e-01 1.45160413e+00 -7.47237682e-01 -2.00281262e-01
-5.66954136e-01 1.47075546e+00 3.23504806e-01 5.03321767e-01
-9.41648781e-01 -4.82000679e-01 7.80109286e-01 1.60579294e-01
4.36933368e-01 -2.80984372e-01 -2.80512214e-01 -1.33114386e+00
2.92968363e-01 -6.67029083e-01 2.94183032e-03 -9.38247502e-01
-1.04211640e+00 6.00241780e-01 -2.42645413e-01 -1.35402143e+00
3.22753161e-01 -2.87080616e-01 -3.81076753e-01 1.09779930e+00
-1.97887039e+00 -1.48475862e+00 -6.71784699e-01 1.09854376e+00
3.49150687e-01 3.05297673e-01 5.35852253e-01 5.61797738e-01
-5.76379120e-01 5.82057059e-01 4.67651449e-02 1.09124027e-01
7.75229394e-01 -7.04996705e-01 4.08039719e-01 9.50497091e-01
-9.17308703e-02 5.34384251e-01 4.17411268e-01 -5.13041675e-01
-1.70669687e+00 -1.29965758e+00 7.64737189e-01 -5.19649088e-02
5.19944847e-01 -3.22612643e-01 -9.92791057e-01 5.13911366e-01
1.56760439e-01 4.15680110e-01 2.18410045e-01 -5.97018182e-01
-4.98836100e-01 -3.37168366e-01 -1.02795899e+00 8.49436939e-01
1.19601572e+00 -7.12018013e-01 -1.30734831e-01 3.14784646e-01
8.39001179e-01 -6.05104208e-01 -9.58637118e-01 3.20254505e-01
4.94202077e-01 -9.47435379e-01 1.30315948e+00 1.59814954e-01
6.34498954e-01 -6.90713227e-01 -7.04270601e-02 -1.09450352e+00
-7.85469115e-01 -8.36827397e-01 1.39973015e-01 9.78156149e-01
3.10936980e-02 -7.49303460e-01 4.07644600e-01 4.82701927e-01
7.26804659e-02 -6.02111459e-01 -9.90284383e-01 -4.30674791e-01
-2.32229710e-01 -1.42314173e-02 5.18891931e-01 8.86887550e-01
-4.68775630e-01 2.97992140e-01 -7.22168982e-01 4.29190338e-01
9.90584850e-01 5.33526123e-01 6.36714637e-01 -7.65001118e-01
-1.48968831e-01 -3.17897946e-01 -2.46494710e-01 -1.09177935e+00
-1.36174694e-01 -6.60530388e-01 7.35191479e-02 -1.26733744e+00
4.21615094e-01 -2.08432287e-01 -3.40786837e-02 5.92026114e-01
-2.22934559e-01 9.72703159e-01 1.66466206e-01 5.19553721e-01
-2.69157380e-01 7.47672439e-01 1.75731146e+00 2.04455983e-02
-1.43351153e-01 -5.06488204e-01 -7.05711246e-01 7.67686248e-01
6.20657086e-01 -3.25864524e-01 -4.27701622e-01 -8.79992604e-01
4.57834043e-02 9.71636102e-02 4.33962315e-01 -9.16182101e-01
4.52268630e-01 -1.93451494e-01 6.54905260e-01 -4.10920143e-01
3.86399269e-01 -8.50492537e-01 4.59864825e-01 4.46056694e-01
-3.15447807e-01 -1.51511291e-02 1.33192584e-01 3.58054459e-01
-2.50068069e-01 4.05704260e-01 1.14332891e+00 -2.87213810e-02
-5.87827444e-01 6.89298630e-01 -4.28541377e-02 -1.69561431e-01
8.27456474e-01 -4.65728864e-02 -3.76991421e-01 -3.24278086e-01
-4.59036201e-01 -1.45149112e-01 6.98778689e-01 2.89443314e-01
7.22123981e-01 -1.57513797e+00 -8.24220002e-01 4.96703118e-01
-6.45823404e-02 5.54505996e-02 5.10465145e-01 9.39591289e-01
-7.33986855e-01 2.33241886e-01 -4.38192099e-01 -4.76633936e-01
-1.21138263e+00 4.74111527e-01 2.40393445e-01 -2.84375817e-01
-1.10089505e+00 8.19179356e-01 6.44585907e-01 -1.90761924e-01
-7.24006444e-03 -2.34589398e-01 5.76001219e-02 -1.91850781e-01
6.97585106e-01 2.77163893e-01 -1.56195715e-01 -8.18516076e-01
-2.16580331e-01 1.27650690e+00 -1.30384669e-01 -4.14039679e-02
1.45205355e+00 -4.52724338e-01 -5.01373172e-01 -3.54228951e-02
1.41035724e+00 -2.67144162e-02 -1.47252917e+00 -6.20145082e-01
-4.99488533e-01 -8.60392570e-01 3.85254651e-01 -3.49953979e-01
-1.45869243e+00 7.53413856e-01 5.41857719e-01 -1.73761040e-01
1.77247107e+00 -3.09430987e-01 9.76863444e-01 2.47186601e-01
2.90864944e-01 -8.09615910e-01 1.30783230e-01 2.89516270e-01
1.14985263e+00 -9.58656073e-01 3.31113875e-01 -7.22679913e-01
-3.33536774e-01 1.04572296e+00 5.53503036e-01 -2.64333755e-01
5.30458331e-01 2.68483460e-01 9.45172310e-02 -2.93560382e-02
-5.72353363e-01 -1.12936525e-02 3.71712685e-01 4.50219482e-01
3.11589271e-01 -5.57034463e-02 -1.86593786e-01 1.97890952e-01
-1.58590630e-01 5.28726652e-02 4.23186749e-01 5.59042692e-01
-6.23861179e-02 -8.45137000e-01 -5.85476398e-01 1.41450465e-01
-3.37488174e-01 -4.26680714e-01 4.28341031e-02 4.19680715e-01
1.42423958e-01 5.79239428e-01 4.25169766e-02 -2.82440841e-01
4.13819432e-01 -3.20020705e-01 5.67841053e-01 -1.45049080e-01
-2.70464331e-01 3.92145246e-01 -4.24008548e-01 -9.54661012e-01
-4.24609989e-01 -3.42888206e-01 -9.26010370e-01 -6.24195814e-01
1.27763063e-01 -2.52430856e-01 4.56650227e-01 4.37799990e-01
5.90374172e-01 5.29407978e-01 7.34272718e-01 -1.38930392e+00
-2.78847724e-01 -7.17097580e-01 -4.79554147e-01 4.46605146e-01
4.04589087e-01 -3.90011281e-01 -2.93382823e-01 2.33830005e-01] | [10.985404014587402, -1.8236578702926636] |
abd46bed-1e94-4682-b263-b840cf176ca7 | generating-features-with-increased-crop | 2304.05096 | null | https://arxiv.org/abs/2304.05096v1 | https://arxiv.org/pdf/2304.05096v1.pdf | Generating Features with Increased Crop-related Diversity for Few-Shot Object Detection | Two-stage object detectors generate object proposals and classify them to detect objects in images. These proposals often do not contain the objects perfectly but overlap with them in many possible ways, exhibiting great variability in the difficulty levels of the proposals. Training a robust classifier against this crop-related variability requires abundant training data, which is not available in few-shot settings. To mitigate this issue, we propose a novel variational autoencoder (VAE) based data generation model, which is capable of generating data with increased crop-related diversity. The main idea is to transform the latent space such latent codes with different norms represent different crop-related variations. This allows us to generate features with increased crop-related diversity in difficulty levels by simply varying the latent norm. In particular, each latent code is rescaled such that its norm linearly correlates with the IoU score of the input crop w.r.t. the ground-truth box. Here the IoU score is a proxy that represents the difficulty level of the crop. We train this VAE model on base classes conditioned on the semantic code of each class and then use the trained model to generate features for novel classes. In our experiments our generated features consistently improve state-of-the-art few-shot object detection methods on the PASCAL VOC and MS COCO datasets. | ['Dimitris Samaras', 'Hieu Le', 'Jingyi Xu'] | 2023-04-11 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Xu_Generating_Features_With_Increased_Crop-Related_Diversity_for_Few-Shot_Object_Detection_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Xu_Generating_Features_With_Increased_Crop-Related_Diversity_for_Few-Shot_Object_Detection_CVPR_2023_paper.pdf | cvpr-2023-1 | ['few-shot-object-detection'] | ['computer-vision'] | [ 1.85828045e-01 -1.21905006e-01 -2.23685205e-01 -8.39480162e-02
-5.05195439e-01 -6.52056515e-01 5.25084317e-01 1.33019909e-01
-1.36798188e-01 2.34447479e-01 -5.95861375e-02 2.67836601e-01
1.48404628e-01 -1.14410782e+00 -8.56298566e-01 -9.16578889e-01
3.13661873e-01 2.92482853e-01 5.91805339e-01 -4.35603410e-01
1.94768801e-01 2.80221224e-01 -2.14790940e+00 4.58580852e-01
9.68343198e-01 8.85932624e-01 7.42891490e-01 6.91149175e-01
-1.91061154e-01 4.18899357e-01 -7.79760599e-01 -2.01802969e-01
4.75709558e-01 -5.10532320e-01 -1.31680593e-01 6.44748092e-01
5.23348868e-01 -1.57824874e-01 -1.13837399e-01 1.29908776e+00
9.31555107e-02 1.32915363e-01 1.23297417e+00 -1.24750721e+00
-8.55296552e-01 4.35334086e-01 -6.46659732e-01 -2.88248863e-02
-6.63657039e-02 2.07637519e-01 1.07764971e+00 -1.15410125e+00
8.38882029e-01 1.10169280e+00 5.28868258e-01 6.38504505e-01
-1.45823574e+00 -5.53142130e-01 -5.83266094e-02 1.81562394e-01
-1.34871006e+00 -7.66460598e-02 7.53813088e-01 -6.85835183e-01
5.15690327e-01 1.09448060e-02 7.73218811e-01 1.13735557e+00
2.41138533e-01 8.89097989e-01 7.96242356e-01 -5.71360409e-01
4.85936195e-01 3.93710703e-01 6.84324801e-02 6.46577656e-01
5.30465543e-01 1.35546327e-02 -4.41435963e-01 1.37826636e-01
5.42028427e-01 1.60970271e-01 -1.70205057e-01 -9.89331901e-01
-1.09830868e+00 1.16255510e+00 7.05731273e-01 3.66299063e-01
-4.45827752e-01 -4.80522253e-02 1.41380623e-01 -1.06262609e-01
3.47132593e-01 6.03830159e-01 -3.92183900e-01 5.61266065e-01
-1.00939214e+00 4.16740268e-01 5.50686598e-01 1.13177168e+00
9.38827217e-01 4.46145497e-02 -7.12159872e-01 8.33571792e-01
1.70706183e-01 4.33907151e-01 5.56814194e-01 -8.48454297e-01
2.14965373e-01 6.88740253e-01 1.66159749e-01 -9.18642342e-01
-2.46912707e-03 -4.93668467e-01 -6.73410535e-01 5.31539202e-01
5.58988631e-01 4.29637283e-02 -1.17937398e+00 1.60520351e+00
3.96285146e-01 -1.38560146e-01 1.62904546e-01 7.41640449e-01
5.95127165e-01 9.93466496e-01 6.83906749e-02 -1.79906458e-01
1.42384946e+00 -9.70532835e-01 -5.99762321e-01 -3.22202295e-01
6.17342949e-01 -7.95415163e-01 1.09226930e+00 2.33743042e-01
-6.52056634e-01 -8.29610705e-01 -1.24191797e+00 3.05185527e-01
-5.61620235e-01 5.61687708e-01 4.36047107e-01 6.76806748e-01
-5.53863168e-01 5.69929779e-01 -6.33894801e-01 -4.44065511e-01
4.92250353e-01 -2.54207611e-01 -7.69405589e-02 7.48651661e-03
-7.94786096e-01 8.61810029e-01 7.44563878e-01 -1.54444024e-01
-1.17997181e+00 -6.30500197e-01 -1.07658482e+00 2.76812851e-01
5.67250311e-01 -3.16437304e-01 1.01110172e+00 -1.10293949e+00
-1.28440797e+00 7.60239422e-01 3.44336964e-02 -3.43065023e-01
4.19372082e-01 1.65193304e-01 -7.80953243e-02 -1.49504811e-01
3.06497842e-01 1.12671697e+00 1.39083242e+00 -1.45998776e+00
-6.92043245e-01 -2.26001188e-01 -1.46889076e-01 1.76356379e-02
-4.53157187e-01 -4.04260337e-01 -1.43880233e-01 -7.67146468e-01
1.32605553e-01 -9.22705114e-01 -1.06008634e-01 5.83645701e-01
-1.83715522e-01 -2.70349443e-01 8.63951087e-01 -2.69112229e-01
8.81380677e-01 -2.23627448e+00 1.95131212e-01 -2.43378252e-01
-7.93641731e-02 3.83732319e-01 -5.14758766e-01 5.96235245e-02
9.82350335e-02 -2.26163521e-01 -4.25085992e-01 -7.88010135e-02
-6.81182221e-02 2.24424049e-01 -4.10711467e-01 3.38990182e-01
7.99499512e-01 7.95576453e-01 -1.01534176e+00 -4.91376221e-01
3.61515105e-01 2.56651998e-01 -6.63626552e-01 3.90742719e-01
-7.31312692e-01 -4.66181803e-03 -2.97118098e-01 7.03953803e-01
8.89599323e-01 -4.28099185e-02 -1.20685078e-01 -2.91993678e-01
-2.46198967e-01 -4.84200358e-01 -1.30178630e+00 1.38784647e+00
-2.74999887e-01 7.36502349e-01 -2.90016681e-01 -1.01291823e+00
1.30440843e+00 -5.60882837e-02 1.30046070e-01 -1.66730791e-01
1.87563911e-01 1.19150415e-01 -2.89508719e-02 -3.93049389e-01
6.12782538e-01 5.99777838e-03 -3.86899114e-02 -1.09614745e-01
5.06953776e-01 -4.72329766e-01 5.67856550e-01 -7.51696602e-02
7.49873698e-01 1.92819148e-01 5.38410962e-01 -2.68473685e-01
2.39305332e-01 1.27541348e-01 6.64901793e-01 8.97192478e-01
-4.09532160e-01 7.25250363e-01 5.98368168e-01 -2.18576372e-01
-1.33254814e+00 -1.25617814e+00 -3.43275309e-01 1.14934957e+00
1.99497670e-01 -1.16449647e-01 -8.12823594e-01 -7.17460275e-01
-6.03926294e-02 9.70555544e-01 -1.02850616e+00 -3.26198578e-01
-1.00789182e-01 -5.95563114e-01 7.93242082e-02 4.60393310e-01
4.19051915e-01 -1.22460973e+00 -1.01866257e+00 3.29478323e-01
-9.90235060e-02 -8.51886332e-01 -2.07531214e-01 4.78392988e-01
-6.70165002e-01 -9.86953020e-01 -1.02360213e+00 -8.57380748e-01
7.63602614e-01 5.52204370e-01 9.75917876e-01 -2.94794977e-01
-7.64478564e-01 1.98780615e-02 -7.12952912e-01 -8.37788105e-01
-6.40234351e-01 -1.28722504e-01 -1.74338356e-01 1.12416677e-01
4.32177991e-01 1.28285661e-01 -3.85364145e-01 2.71701396e-01
-1.16658092e+00 -6.74879849e-02 5.03279686e-01 1.20334351e+00
6.58304989e-01 1.55699655e-01 1.93116069e-01 -3.52710783e-01
2.20354185e-01 -5.65464795e-01 -8.93694997e-01 4.26676095e-01
-2.64243305e-01 2.65884429e-01 4.45791900e-01 -8.56992662e-01
-9.99951541e-01 5.21313608e-01 3.13860387e-01 -5.88296771e-01
-2.67231315e-01 4.40434031e-02 -1.25164956e-01 2.12751627e-01
1.10629141e+00 2.24897116e-01 -2.50583500e-01 -2.91667506e-02
6.45128548e-01 5.45440078e-01 3.98407966e-01 -2.06125394e-01
9.72731888e-01 3.79936248e-01 -9.14222524e-02 -9.56900358e-01
-1.09175670e+00 -5.01888275e-01 -7.86365032e-01 -3.43748271e-01
1.14528012e+00 -7.66278207e-01 2.90314928e-02 5.22112250e-01
-1.43233943e+00 -1.74461290e-01 -6.52929306e-01 3.14688653e-01
-6.24344170e-01 2.74243951e-03 -1.69468373e-01 -8.86770785e-01
-6.18427023e-02 -1.11120415e+00 1.14998591e+00 5.12353778e-01
7.11269602e-02 -4.43778634e-01 -4.48171049e-02 -6.56830594e-02
2.15914458e-01 3.19025666e-01 7.87096024e-01 -3.21240932e-01
-5.11458814e-01 -3.12043756e-01 -2.49302268e-01 5.63810647e-01
1.26622105e-02 3.25938433e-01 -9.73948717e-01 -1.85125366e-01
-9.40345451e-02 -4.66861755e-01 1.13150978e+00 5.86273074e-01
1.07263970e+00 2.79871300e-02 -3.52582693e-01 4.41097051e-01
1.60045934e+00 5.88224153e-04 4.25080776e-01 9.74759236e-02
4.90287185e-01 7.14513600e-01 9.29759860e-01 5.94840467e-01
-2.24866599e-01 7.49595165e-01 5.69081306e-01 1.56944931e-01
-1.29872650e-01 -1.52531773e-01 4.71263081e-01 3.40237737e-01
1.76086798e-01 -3.00733745e-01 -6.16297424e-01 6.27120733e-01
-1.82178605e+00 -9.62765098e-01 -8.65925848e-02 2.15038109e+00
6.50666177e-01 2.00762287e-01 -7.90197309e-03 6.50572404e-02
9.30113256e-01 1.89412832e-01 -5.82283318e-01 -2.15450272e-01
-1.57669410e-01 6.06340840e-02 4.90479201e-01 2.48830747e-02
-1.15822566e+00 1.06852353e+00 5.55364132e+00 9.03067350e-01
-9.00833011e-01 8.77086222e-02 3.01826537e-01 1.11331910e-01
8.68706033e-03 -5.35235181e-02 -1.04272616e+00 3.47417831e-01
4.25740153e-01 -2.17793524e-01 2.67890424e-01 1.28726852e+00
-2.20854327e-01 -1.86635017e-01 -1.04242849e+00 6.77924991e-01
3.23254198e-01 -1.24868083e+00 4.16371189e-02 -1.45506933e-01
1.01857531e+00 -2.69261956e-01 2.01061517e-01 5.96550167e-01
1.69597268e-01 -5.58534086e-01 9.73453939e-01 4.21988904e-01
6.32724464e-01 -5.22141278e-01 5.91727495e-01 5.32023013e-01
-1.33739460e+00 -3.64294767e-01 -1.01860428e+00 1.83651641e-01
-1.88855886e-01 6.03410363e-01 -7.76383281e-01 -2.93026082e-02
6.74264967e-01 4.72409427e-01 -7.31995463e-01 1.04477239e+00
-2.89610386e-01 3.60052168e-01 -5.59082069e-02 -3.23021889e-01
2.41431475e-01 -3.81425135e-02 7.31203675e-01 1.06655622e+00
7.22286344e-01 -5.04898019e-02 1.33249462e-01 1.52658975e+00
3.47983204e-02 -1.52862063e-02 -7.69498885e-01 -1.00176737e-01
5.43509841e-01 1.40826595e+00 -8.63695443e-01 -4.43177581e-01
-2.35201776e-01 1.15765154e+00 2.14402840e-01 1.05506040e-01
-8.05118203e-01 -5.42454004e-01 4.81569618e-01 -9.61361527e-02
8.34794343e-01 -4.20140475e-02 7.43345544e-02 -1.19274879e+00
1.93603616e-02 -6.98988795e-01 2.20182553e-01 -8.91337216e-01
-1.39801109e+00 4.87052083e-01 2.60437876e-02 -1.53582025e+00
-1.69268087e-01 -8.38791966e-01 -6.17623925e-01 5.59178829e-01
-1.17642677e+00 -1.14590514e+00 -7.60082901e-01 2.50741392e-01
1.09865880e+00 -2.64148563e-01 9.17443156e-01 -2.74019718e-01
-4.97166097e-01 5.40528715e-01 2.55824298e-01 -7.22326115e-02
5.88254631e-01 -1.22099864e+00 1.78912476e-01 9.62946057e-01
4.54924792e-01 1.06187813e-01 8.95871401e-01 -5.06704628e-01
-1.06488419e+00 -1.40868473e+00 5.97405970e-01 -5.76566875e-01
6.24374866e-01 -4.94416654e-01 -1.04796112e+00 2.38732710e-01
-1.04927607e-01 3.53012472e-01 2.39570409e-01 -2.59064049e-01
-6.05002761e-01 -5.93187027e-02 -1.05987048e+00 5.02212048e-01
7.13041484e-01 -2.63208270e-01 -7.10326314e-01 3.60291511e-01
8.31754327e-01 -1.83887422e-01 -4.33597535e-01 3.50671828e-01
2.91720688e-01 -9.77402687e-01 8.33572328e-01 -3.36650670e-01
6.71439052e-01 -3.66147965e-01 -3.57703328e-01 -1.52211785e+00
-7.54864872e-01 7.13472143e-02 -2.02566117e-01 1.11477804e+00
2.75282234e-01 5.49041014e-03 7.77272999e-01 2.04207171e-02
8.04344788e-02 -5.37598550e-01 -5.70367277e-01 -9.64012146e-01
3.49797755e-02 -2.18708351e-01 4.00762409e-01 7.45306373e-01
-2.29754522e-01 3.70715782e-02 -1.94388732e-01 1.58780172e-01
8.17770422e-01 1.64060071e-01 7.48534799e-01 -1.34015286e+00
-3.37756395e-01 -4.35235023e-01 -5.23235798e-01 -6.86910093e-01
1.57430395e-01 -8.83434296e-01 5.43368995e-01 -1.29256964e+00
3.90065640e-01 -1.73571751e-01 -1.72545418e-01 3.52472663e-01
-3.36788714e-01 2.20331863e-01 4.94755775e-01 1.81726262e-01
-3.53180915e-01 7.95685291e-01 1.10968435e+00 -4.67270285e-01
-4.54900056e-01 -5.19499891e-02 -2.61532426e-01 6.32541239e-01
7.58870542e-01 -7.03378737e-01 -1.24447726e-01 -1.80616841e-01
-7.80488774e-02 -1.79019913e-01 4.29768175e-01 -1.38189638e+00
-1.03719056e-01 -3.71188909e-01 6.97988331e-01 -7.10243225e-01
2.72575587e-01 -7.18203187e-01 -2.56900787e-01 6.93471014e-01
-3.75935763e-01 -5.65189302e-01 7.15768561e-02 8.42208683e-01
3.58409099e-02 -6.83627725e-01 1.13893998e+00 -5.90706244e-02
-9.19488251e-01 2.55336255e-01 -5.27576327e-01 -2.72646099e-01
1.34314191e+00 -3.03073496e-01 -1.07448041e-01 6.30524829e-02
-4.60612386e-01 -3.94742154e-02 4.72104371e-01 7.86101222e-01
5.09687901e-01 -1.38773203e+00 -7.93000042e-01 3.67532432e-01
8.44602764e-01 -5.96655272e-02 7.82024264e-02 2.06923887e-01
-2.97457576e-01 1.54166028e-01 -6.77563369e-01 -8.28062356e-01
-1.07072222e+00 8.77335012e-01 3.17437053e-01 -1.53303472e-02
-3.91118169e-01 9.39387619e-01 3.98537695e-01 -2.42330685e-01
2.26304352e-01 -3.61718714e-01 -4.64284331e-01 4.76181597e-01
4.65671539e-01 2.67530948e-01 -2.40628302e-01 -4.65628326e-01
-1.57740116e-02 5.46182036e-01 5.90190627e-02 7.71989599e-02
1.19865990e+00 2.89513350e-01 1.85475186e-01 5.67139566e-01
9.97740328e-01 -4.19298053e-01 -1.66349399e+00 -2.68639386e-01
-1.00581154e-01 -6.18056953e-01 1.00997336e-01 -3.82479787e-01
-9.76362586e-01 9.08610463e-01 1.06860280e+00 1.20095216e-01
8.95892978e-01 1.41107023e-01 3.38734984e-01 4.89293545e-01
4.23169911e-01 -1.35813105e+00 6.86963856e-01 4.58493829e-01
8.69112730e-01 -1.37818813e+00 -2.73346454e-01 -3.90716523e-01
-6.81924045e-01 1.24293363e+00 8.37275803e-01 -2.95816928e-01
4.38258648e-01 6.33992553e-02 -1.07133165e-01 -7.79787004e-02
-5.10747075e-01 -4.89259809e-01 3.79669011e-01 9.02332544e-01
2.08253443e-01 3.21214497e-01 -2.39975661e-01 4.71913248e-01
9.84525681e-02 -1.73375562e-01 6.95030332e-01 8.02108943e-01
-1.02834702e+00 -7.54282296e-01 -6.65072739e-01 4.59392399e-01
2.13225678e-01 1.44138783e-01 -1.60730168e-01 4.76369858e-01
6.51997983e-01 7.52226412e-01 2.49665484e-01 -2.35742897e-01
3.82729977e-01 3.91618870e-02 6.33986712e-01 -8.60721886e-01
-1.80791579e-02 -5.38820289e-02 -5.30892968e-01 -4.02044296e-01
-4.94410247e-02 -6.89179838e-01 -9.17853177e-01 2.95591503e-01
-8.47660005e-01 -2.85390705e-01 7.42631376e-01 6.08590424e-01
-6.07361235e-02 5.70180476e-01 6.76728666e-01 -1.22438228e+00
-9.74227726e-01 -9.95745242e-01 -7.25914657e-01 5.84917307e-01
1.65921271e-01 -9.75965500e-01 -3.89365613e-01 2.22182393e-01] | [9.698585510253906, 2.0349221229553223] |
88819c49-ca83-4611-b720-ad97e6ebe426 | poison-attack-and-defense-on-deep-source-code | 2210.17029 | null | https://arxiv.org/abs/2210.17029v1 | https://arxiv.org/pdf/2210.17029v1.pdf | Poison Attack and Defense on Deep Source Code Processing Models | In the software engineering community, deep learning (DL) has recently been applied to many source code processing tasks. Due to the poor interpretability of DL models, their security vulnerabilities require scrutiny. Recently, researchers have identified an emergent security threat, namely poison attack. The attackers aim to inject insidious backdoors into models by poisoning the training data with poison samples. Poisoned models work normally with clean inputs but produce targeted erroneous results with poisoned inputs embedded with triggers. By activating backdoors, attackers can manipulate the poisoned models in security-related scenarios. To verify the vulnerability of existing deep source code processing models to the poison attack, we present a poison attack framework for source code named CodePoisoner as a strong imaginary enemy. CodePoisoner can produce compilable even human-imperceptible poison samples and attack models by poisoning the training data with poison samples. To defend against the poison attack, we further propose an effective defense approach named CodeDetector to detect poison samples in the training data. CodeDetector can be applied to many model architectures and effectively defend against multiple poison attack approaches. We apply our CodePoisoner and CodeDetector to three tasks, including defect detection, clone detection, and code repair. The results show that (1) CodePoisoner achieves a high attack success rate (max: 100%) in misleading models to targeted erroneous behaviors. It validates that existing deep source code processing models have a strong vulnerability to the poison attack. (2) CodeDetector effectively defends against multiple poison attack approaches by detecting (max: 100%) poison samples in the training data. We hope this work can help practitioners notice the poison attack and inspire the design of more advanced defense techniques. | ['Xin Xia', 'Xing Hu', 'Zhi Jin', 'Ge Li', 'Huangzhao Zhang', 'Zhuo Li', 'Jia Li'] | 2022-10-31 | null | null | null | null | ['defect-detection'] | ['computer-vision'] | [-1.85080543e-01 -2.31541276e-01 -9.67923701e-02 1.68843091e-01
-4.22699571e-01 -1.15365875e+00 4.47878867e-01 3.82433742e-01
1.92537144e-01 -9.27805603e-02 -1.26121566e-01 -8.96374166e-01
3.57038975e-01 -9.88275290e-01 -8.63932908e-01 -3.80592376e-01
-3.97488147e-01 -2.04463825e-01 3.35197181e-01 -3.05303723e-01
4.78006035e-01 3.69633555e-01 -1.27410376e+00 6.52622759e-01
7.84688115e-01 2.77825981e-01 -2.54613042e-01 8.80857587e-01
-1.58137947e-01 1.20912921e+00 -1.37922943e+00 -5.73509097e-01
3.18375617e-01 -1.40016586e-01 -7.58872032e-01 -5.81711888e-01
3.25561687e-02 -5.25666535e-01 -1.10430248e-01 1.56636894e+00
2.40249842e-01 -8.00873578e-01 1.86168537e-01 -1.99243450e+00
-7.52139986e-01 1.00260270e+00 -7.74215877e-01 1.04457229e-01
4.10081983e-01 6.93986654e-01 6.56328559e-01 -7.24880397e-01
8.40666965e-02 1.47174275e+00 6.23867512e-01 1.09755790e+00
-1.09930146e+00 -9.25913692e-01 -8.16797558e-03 -4.68634188e-01
-1.30681825e+00 -8.13821703e-02 6.03507876e-01 -6.35261893e-01
1.23465014e+00 4.75495815e-01 9.52753723e-02 1.60924494e+00
8.92772615e-01 8.67941141e-01 5.72849870e-01 -2.19071284e-01
5.51847935e-01 2.54551321e-01 5.35436928e-01 9.51516032e-01
4.77984607e-01 6.20777249e-01 -2.81553239e-01 -1.12774134e+00
3.39272648e-01 6.07441813e-02 -2.43604958e-01 2.84421772e-01
-8.84009659e-01 1.11087918e+00 4.54314888e-01 1.96472928e-01
7.68737122e-02 5.21486640e-01 8.42799067e-01 5.07378161e-01
-1.35641843e-01 8.77339482e-01 -7.01007247e-01 -4.88428436e-02
-2.86346108e-01 1.17123283e-01 1.01993847e+00 1.00582254e+00
4.29113179e-01 6.67823970e-01 1.55393332e-01 2.69432157e-01
5.15337229e-01 7.68520594e-01 2.95782804e-01 -4.08705652e-01
4.27779764e-01 9.70682085e-01 -1.92635477e-01 -1.17440665e+00
-1.54273242e-01 -4.71090406e-01 -6.76259518e-01 6.68798089e-01
1.13405295e-01 -3.13727111e-01 -6.05006099e-01 1.55001473e+00
2.32848302e-01 2.86682770e-02 1.79793805e-01 3.64494592e-01
3.69785100e-01 5.60944200e-01 1.49627775e-01 3.61526817e-01
1.40485656e+00 -5.17021835e-01 -2.71770686e-01 -1.97501168e-01
1.10034251e+00 -3.81480038e-01 1.48115611e+00 7.63185799e-01
-4.63072091e-01 -3.11977357e-01 -1.43538010e+00 6.67138815e-01
-5.82311928e-01 -4.06389832e-01 6.83923483e-01 1.44956017e+00
-8.70284021e-01 3.20006102e-01 -1.03783691e+00 1.45593435e-01
4.64745402e-01 5.26418328e-01 -2.12979183e-01 4.53363406e-03
-1.03557372e+00 5.21556914e-01 1.48980200e-01 -3.03136647e-01
-1.84436071e+00 -9.40203369e-01 -1.00133061e+00 7.30576068e-02
1.14370346e-01 -4.30106044e-01 1.15558350e+00 -6.69189692e-01
-8.65649521e-01 5.49007356e-01 3.48357290e-01 -4.97460335e-01
2.86316335e-01 -2.34313458e-01 -5.52058399e-01 -5.39278723e-02
1.81834176e-02 2.42260359e-02 1.11631346e+00 -1.47542655e+00
-2.94033051e-01 -7.66332150e-02 4.14573848e-01 -7.39819348e-01
-8.60883057e-01 6.80657864e-01 4.56994027e-01 -6.74084067e-01
-4.19268131e-01 -5.16913712e-01 -4.64704707e-02 -1.92000195e-01
-7.13835716e-01 4.36482392e-02 1.10497737e+00 -1.67416960e-01
1.39506137e+00 -2.42413402e+00 -6.44088462e-02 5.98194376e-02
7.12837279e-01 6.79297328e-01 -3.86046082e-01 5.08670390e-01
-1.91459805e-01 8.11454594e-01 -5.15587330e-01 5.67897372e-02
1.80313125e-01 1.28469452e-01 -1.07316518e+00 5.67561507e-01
5.59444845e-01 8.98308575e-01 -1.01905572e+00 3.72588597e-02
-1.47721529e-01 1.51469469e-01 -9.34133887e-01 3.95462126e-01
-4.16482002e-01 -1.27635092e-01 -6.30057871e-01 1.19924438e+00
6.95319533e-01 8.91686976e-02 -3.36603135e-01 2.39310071e-01
1.23030439e-01 6.40871897e-02 -8.09344709e-01 8.26763153e-01
-4.37220991e-01 4.28002238e-01 3.09971869e-01 -2.92085290e-01
1.02203035e+00 2.59319425e-01 -2.16358393e-01 -2.03790247e-01
2.77861983e-01 2.90442944e-01 2.24599242e-01 -5.41417658e-01
1.47267804e-01 -2.20970251e-02 -6.01358294e-01 1.01257527e+00
-2.59693056e-01 -1.15131937e-01 -2.66907692e-01 5.65390170e-01
1.98988795e+00 -4.44540322e-01 1.00665614e-01 -3.62153500e-01
7.28434861e-01 -4.05642539e-02 4.90138292e-01 1.18953264e+00
-2.21197292e-01 4.79225330e-02 8.83051872e-01 -7.13431299e-01
-8.12060535e-01 -1.15966535e+00 1.06461853e-01 8.53841484e-01
-1.57421827e-01 -7.80864358e-01 -9.74178731e-01 -1.44277954e+00
2.93237679e-02 9.06004190e-01 -5.71292222e-01 -1.08302462e+00
-6.12812459e-01 -7.72262216e-01 1.75354445e+00 7.95794845e-01
5.93851149e-01 -1.11719525e+00 -7.48349607e-01 1.17908508e-01
1.16493016e-01 -5.21651387e-01 -3.88356030e-01 4.71194118e-01
-5.34671426e-01 -1.33711481e+00 1.61535308e-01 -4.54441339e-01
6.58190131e-01 6.25124499e-02 1.17336011e+00 1.08148801e+00
-5.48161209e-01 3.70455593e-01 -5.86770773e-01 -6.81218386e-01
-1.47726166e+00 -1.77592948e-01 2.67262280e-01 -4.07918632e-01
6.47550225e-01 -3.59871328e-01 1.07500158e-01 3.90268356e-01
-1.42187858e+00 -6.99896753e-01 2.06180304e-01 7.32067525e-01
-4.16203946e-01 3.61752063e-01 7.08254099e-01 -7.74734497e-01
9.53323364e-01 -7.76317775e-01 -8.84243786e-01 1.52178496e-01
-2.47554377e-01 1.09567367e-01 1.15070653e+00 -8.34048986e-01
-5.38085282e-01 -3.00110668e-01 -5.19307196e-01 -5.58534503e-01
-2.58981377e-01 4.08227712e-01 -5.80594361e-01 -3.39988917e-01
1.41917658e+00 1.52131319e-01 -1.82485387e-01 -3.56589347e-01
3.10596991e-02 5.96170068e-01 4.07259613e-01 -1.02213717e+00
1.43196750e+00 2.06066370e-01 -2.88838655e-01 -5.35980165e-01
-2.57546782e-01 1.49346843e-01 -1.15143880e-01 1.09832890e-01
5.13499975e-01 -5.10122716e-01 -8.86959910e-01 9.95840967e-01
-1.52078116e+00 -4.19832349e-01 1.68189555e-01 -3.80923033e-01
-9.18054879e-02 6.21515810e-01 -9.24217999e-01 -9.31383789e-01
-4.44710046e-01 -1.41617990e+00 9.09780860e-01 -2.82447398e-01
-3.88264716e-01 -8.55194867e-01 8.23732093e-02 -5.83411120e-02
4.71587539e-01 5.52799046e-01 1.68034840e+00 -9.67838526e-01
-4.64909375e-01 -5.02336204e-01 3.13602269e-01 5.08724272e-01
1.03369668e-01 3.89935911e-01 -1.07604468e+00 -4.57666218e-01
6.48848355e-01 -4.40599144e-01 2.86964536e-01 -5.60640037e-01
1.07804227e+00 -6.68153703e-01 -3.09818268e-01 7.33845294e-01
1.30629182e+00 3.94054323e-01 6.28934920e-01 3.25600654e-01
8.20330024e-01 3.13144505e-01 2.59146571e-01 5.17198443e-01
-3.53273988e-01 2.38613516e-01 1.14388359e+00 1.71051726e-01
2.75843501e-01 -3.37706387e-01 8.70901823e-01 3.49877179e-01
8.13355327e-01 -2.59548545e-01 -1.42098701e+00 4.11063552e-01
-1.26323366e+00 -7.31258452e-01 -5.39263666e-01 1.92417669e+00
9.77908313e-01 3.29840392e-01 -1.06997669e-01 5.60954273e-01
4.85954881e-01 -1.39115199e-01 -6.03591740e-01 -6.89610541e-01
1.56160995e-01 2.17171609e-01 2.54157096e-01 4.36637402e-01
-9.73126054e-01 1.02192760e+00 6.21002483e+00 6.89147234e-01
-9.93381083e-01 8.23500752e-02 2.19675273e-01 1.85165793e-01
-4.45202768e-01 5.71215898e-02 -7.66034782e-01 4.05044883e-01
1.08786678e+00 -2.60810733e-01 4.09665465e-01 1.36331654e+00
-3.32499504e-01 2.28464887e-01 -1.57694554e+00 3.68218094e-01
-3.00381873e-02 -1.14890885e+00 1.89358473e-01 1.03565946e-01
3.46556515e-01 -9.27385837e-02 1.99980333e-01 6.33041024e-01
9.39453781e-01 -1.28253853e+00 8.36606622e-01 -2.85768509e-01
4.03699279e-01 -9.41881835e-01 6.35265946e-01 5.96098304e-01
-9.89842296e-01 -6.32848799e-01 -3.35352361e-01 -1.90484688e-01
-3.13374192e-01 3.42145294e-01 -8.93749654e-01 1.80385321e-01
6.97624922e-01 3.36497843e-01 -9.35884953e-01 6.96094513e-01
-4.82217163e-01 8.36846948e-01 -1.11176834e-01 -8.95697717e-03
1.76538929e-01 6.27201200e-01 6.27291858e-01 1.19513667e+00
-6.22621104e-02 -2.96256214e-01 1.04912713e-01 1.55340493e+00
7.24466294e-02 -5.64733922e-01 -1.10662758e+00 -2.18248963e-01
6.80568755e-01 9.81736064e-01 -7.22567916e-01 -2.59635337e-02
-2.81703234e-01 6.69286847e-01 8.43869075e-02 2.95092225e-01
-1.04927611e+00 -6.44778371e-01 9.54462886e-01 -9.54865366e-02
-1.52877048e-01 -1.66893482e-01 -4.40725535e-01 -1.08629715e+00
-8.84338841e-02 -1.81650257e+00 3.73714954e-01 -4.83222991e-01
-1.47367990e+00 8.84232938e-01 -4.69910316e-02 -9.85957861e-01
1.34857938e-01 -6.93567812e-01 -1.13308883e+00 7.06261933e-01
-7.56749988e-01 -1.00551975e+00 8.48577693e-02 8.89092088e-01
3.27955067e-01 -6.35537267e-01 1.00274670e+00 1.11665623e-03
-7.28298128e-01 1.06738639e+00 -6.20505333e-01 6.49037778e-01
2.04206005e-01 -1.08622384e+00 1.14487851e+00 1.38127398e+00
-9.27272364e-02 1.42562473e+00 6.59918427e-01 -1.05991471e+00
-1.74282885e+00 -1.31734502e+00 2.36213893e-01 -1.06193829e+00
1.00101972e+00 -9.21125293e-01 -1.29479325e+00 7.04486430e-01
4.07716855e-02 -2.11178288e-02 7.10947990e-01 -3.39308023e-01
-1.21276689e+00 3.21206748e-01 -1.54175198e+00 7.95218229e-01
4.81721073e-01 -8.08024585e-01 -5.93130171e-01 2.76234508e-01
1.42496097e+00 -2.75550000e-02 -5.75280309e-01 1.48705825e-01
2.50432175e-02 -9.65472937e-01 8.44527543e-01 -8.74074221e-01
5.80369890e-01 -4.99474287e-01 -2.26254523e-01 -9.15802658e-01
-8.44726786e-02 -7.44341791e-01 -4.16985571e-01 1.32141912e+00
1.71466529e-01 -8.32917035e-01 3.42805356e-01 6.25527978e-01
-1.17163576e-01 -4.24916387e-01 -6.60791039e-01 -8.83188963e-01
5.86789966e-01 -6.16430342e-01 9.76826370e-01 1.17507386e+00
2.88440734e-01 -1.26877844e-01 -2.68175840e-01 7.75766134e-01
7.46407568e-01 -4.96859580e-01 8.56957734e-01 -9.38241422e-01
-6.65018380e-01 -3.45332623e-01 -2.61471927e-01 -3.87047827e-01
2.90979922e-01 -8.63207996e-01 5.07129449e-03 -5.96354067e-01
-4.64934781e-02 -2.36340940e-01 -8.70095193e-02 1.09692740e+00
-3.37806016e-01 -1.61934514e-02 1.91819504e-01 1.85256582e-02
-1.98924951e-02 2.57448405e-01 2.95264333e-01 -4.92138416e-01
-2.77189203e-02 7.58766662e-03 -9.27671909e-01 8.70158076e-01
7.55757213e-01 -1.10726929e+00 -4.66968805e-01 -4.80322778e-01
6.50609255e-01 -1.41748697e-01 7.03295469e-01 -9.94503617e-01
5.96932024e-02 -1.46113411e-01 -1.77282229e-01 -1.47655576e-01
-4.03547198e-01 -8.20052922e-01 -1.02032706e-01 1.12122798e+00
-1.77595049e-01 4.01264578e-01 6.88356280e-01 4.31385636e-01
1.76978365e-01 -6.13624811e-01 7.63517737e-01 -2.72163600e-01
-3.38804036e-01 8.22392851e-02 -9.53623891e-01 5.66805527e-02
1.10938478e+00 5.24485484e-02 -8.96359563e-01 1.17985129e-01
-5.10277748e-02 1.65163130e-01 7.34894991e-01 7.30270147e-01
9.68895495e-01 -1.06063247e+00 -5.45514703e-01 8.48834217e-01
2.51609683e-01 -4.45390671e-01 -2.61465371e-01 2.02065930e-01
-4.75340277e-01 -2.73423940e-01 -1.37527272e-01 -5.52291751e-01
-1.36477292e+00 1.44971859e+00 5.30016005e-01 3.28804292e-02
-4.91382927e-01 9.74811792e-01 4.74121690e-01 -6.58742011e-01
3.48176688e-01 -3.15697342e-01 3.08851272e-01 -6.23701453e-01
1.08573580e+00 1.92634374e-01 3.17234658e-02 -7.49069005e-02
-6.62695885e-01 1.00039072e-01 -3.70516807e-01 3.14053655e-01
1.10187888e+00 7.21043229e-01 -5.98951459e-01 7.79214352e-02
1.11750615e+00 2.33638600e-01 -8.26120675e-01 3.12781304e-01
1.88808784e-01 -5.51784813e-01 -2.48388126e-01 -1.08412659e+00
-1.02917862e+00 1.13823819e+00 2.25102827e-01 5.75950861e-01
9.70142126e-01 -2.43474901e-01 5.81469953e-01 3.62878025e-01
8.19830060e-01 -1.05749726e-01 6.21502459e-01 6.19253457e-01
8.20287466e-01 -7.39922404e-01 -4.91647184e-01 -2.97371626e-01
-3.20300043e-01 9.80587840e-01 1.09764969e+00 -2.36640185e-01
4.59177941e-01 1.32588649e+00 1.98327556e-01 -4.58458215e-01
-1.00068116e+00 8.10827732e-01 -3.80241364e-01 1.07106030e+00
1.19623311e-01 -2.37301826e-01 3.31648886e-01 9.78476822e-01
-6.21553569e-04 -5.20429015e-01 1.00476921e+00 1.14352548e+00
-3.93626988e-01 -1.36209786e+00 -9.22959447e-01 2.12301835e-02
-4.12635386e-01 -3.83931935e-01 -8.47146928e-01 7.46192932e-01
7.00942352e-02 1.41175914e+00 -4.60477620e-01 -1.00136364e+00
1.74169302e-01 -1.28383413e-02 -1.38146758e-01 -8.74988139e-01
-1.58203888e+00 -3.49155396e-01 -3.23514223e-01 -6.73676193e-01
5.40382624e-01 -7.92386606e-02 -1.27855134e+00 -7.14128792e-01
-3.27798218e-01 1.03364192e-01 3.67860913e-01 6.30904615e-01
3.22769284e-01 6.72063828e-01 7.46470332e-01 -3.32983226e-01
-1.16473544e+00 -5.69355965e-01 -3.65005076e-01 1.67951450e-01
6.39572203e-01 -4.72999960e-01 -9.31289136e-01 -1.70082629e-01] | [6.545122146606445, 7.851982593536377] |
bb77aeef-7a19-42aa-9ba9-5600641d0735 | boost-test-time-performance-with-closed-loop | 2203.10853 | null | https://arxiv.org/abs/2203.10853v2 | https://arxiv.org/pdf/2203.10853v2.pdf | Boost Test-Time Performance with Closed-Loop Inference | Conventional deep models predict a test sample with a single forward propagation, which, however, may not be sufficient for predicting hard-classified samples. On the contrary, we human beings may need to carefully check the sample many times before making a final decision. During the recheck process, one may refine/adjust the prediction by referring to related samples. Motivated by this, we propose to predict those hard-classified test samples in a looped manner to boost the model performance. However, this idea may pose a critical challenge: how to construct looped inference, so that the original erroneous predictions on these hard test samples can be corrected with little additional effort. To address this, we propose a general Closed-Loop Inference (CLI) method. Specifically, we first devise a filtering criterion to identify those hard-classified test samples that need additional inference loops. For each hard sample, we construct an additional auxiliary learning task based on its original top-$K$ predictions to calibrate the model, and then use the calibrated model to obtain the final prediction. Promising results on ImageNet (in-distribution test samples) and ImageNet-C (out-of-distribution test samples) demonstrate the effectiveness of CLI in improving the performance of any pre-trained model. | ['Mingkui Tan', 'YaoWei Wang', 'Peilin Zhao', 'Junzhou Huang', 'Haokun Li', 'Guanghui Xu', 'Yifan Zhang', 'Jiaxiang Wu', 'Shuaicheng Niu'] | 2022-03-21 | null | null | null | null | ['auxiliary-learning'] | ['methodology'] | [ 2.57529795e-01 1.08390607e-01 -1.92516536e-01 -8.07897925e-01
-7.25298047e-01 -1.85852185e-01 3.02504957e-01 1.22959644e-01
-3.48228186e-01 8.16049337e-01 -4.78220642e-01 -3.50725174e-01
-7.19504505e-02 -9.48907137e-01 -8.34953547e-01 -4.80799615e-01
3.51458669e-01 7.43597627e-01 5.04196048e-01 1.06895283e-01
2.06811845e-01 1.58628568e-01 -1.34461045e+00 3.78639281e-01
1.27178454e+00 1.22097898e+00 3.67800862e-01 4.67138946e-01
-2.20072605e-02 7.59942710e-01 -7.11097956e-01 -6.68226242e-01
4.95734066e-03 -5.26270151e-01 -7.87526369e-01 1.59955755e-01
2.43236929e-01 -4.98216003e-01 5.20421155e-02 1.35643148e+00
2.34392345e-01 -3.24754454e-02 5.72954178e-01 -1.13350987e+00
-4.88832712e-01 7.69633055e-01 -4.21182334e-01 -6.13462143e-02
7.41842091e-02 2.48603925e-01 9.26140070e-01 -1.01917720e+00
3.15580130e-01 1.09446406e+00 6.73061252e-01 4.47915286e-01
-1.11286151e+00 -8.12335491e-01 3.90598357e-01 5.91815650e-01
-1.34878945e+00 -2.20965862e-01 7.29477286e-01 -2.31280014e-01
4.26742762e-01 1.95164621e-01 6.57276511e-01 9.14083123e-01
2.43115529e-01 9.54290867e-01 9.57067311e-01 -3.08428824e-01
2.99839795e-01 4.01038885e-01 3.78409386e-01 8.70785892e-01
2.56455004e-01 -9.77684110e-02 -2.78944790e-01 1.67570170e-02
3.88857245e-01 -8.08418989e-02 -3.99358451e-01 2.28908230e-02
-8.25028419e-01 6.24924302e-01 5.22791266e-01 6.94440901e-02
-3.56418937e-01 -7.03166872e-02 8.90391096e-02 3.43879998e-01
5.76827407e-01 3.75964969e-01 -6.09769464e-01 1.11528486e-01
-1.25811338e+00 3.12798560e-01 7.54602432e-01 6.54538393e-01
1.03776026e+00 -3.17170143e-01 -3.33265215e-01 1.09437108e+00
3.97892386e-01 2.94005096e-01 3.47131729e-01 -4.60141420e-01
4.06214893e-01 6.38770461e-01 -5.11699542e-03 -8.79757881e-01
-4.21616584e-02 -7.42083073e-01 -9.99575794e-01 9.63663012e-02
4.01294589e-01 1.50088398e-02 -1.20189643e+00 1.52159238e+00
1.71168193e-01 3.01364124e-01 -2.72845298e-01 7.63904870e-01
4.03786004e-01 5.80115199e-01 -2.12944895e-02 -1.18857466e-01
1.05000412e+00 -1.03261685e+00 -3.72274935e-01 -4.00036275e-01
4.13387924e-01 -6.04553461e-01 1.13441908e+00 7.24666893e-01
-1.03550589e+00 -8.20683956e-01 -1.19875360e+00 2.47287452e-01
-1.51515573e-01 3.67499232e-01 2.32258797e-01 2.98486799e-01
-9.16010499e-01 7.09771812e-01 -7.71789610e-01 8.87015983e-02
4.61711407e-01 3.43705297e-01 -1.25023443e-02 -4.04689550e-01
-1.18913329e+00 8.18342209e-01 6.29195571e-01 6.01899624e-01
-9.69702661e-01 -4.62864459e-01 -7.02477336e-01 1.57254964e-01
4.64972138e-01 -5.17773986e-01 1.33166933e+00 -1.08646858e+00
-1.40244555e+00 5.32112479e-01 -4.79257047e-01 -4.51876491e-01
8.36672187e-01 -1.48892879e-01 -4.59014624e-01 -8.31648782e-02
1.33449480e-01 7.47822523e-01 1.10419464e+00 -1.29204893e+00
-7.92538404e-01 -1.27632171e-01 8.28398466e-02 -9.70119685e-02
-2.09128231e-01 -4.45453525e-01 -8.30662787e-01 -4.09169495e-01
3.79377514e-01 -8.71917903e-01 -2.94289410e-01 -3.34978923e-02
-7.38343894e-01 -3.92691076e-01 6.25446558e-01 -5.52811027e-01
1.29509211e+00 -2.16450548e+00 -2.32876495e-01 6.29418015e-01
3.32080245e-01 5.01268625e-01 -1.27563849e-01 -3.76462162e-01
6.73020706e-02 1.14781275e-01 -3.48469675e-01 -3.92114580e-01
1.11901276e-02 2.01882899e-01 -2.85834372e-01 1.97731629e-01
5.58490634e-01 8.01352620e-01 -8.07232559e-01 -4.25244153e-01
1.47529453e-01 1.67272687e-01 -9.36099768e-01 3.07122141e-01
-5.35684109e-01 2.49396592e-01 -3.95790815e-01 5.70682168e-01
8.24893415e-01 -6.04598522e-01 1.04690991e-01 -2.67058581e-01
2.72474498e-01 1.72033221e-01 -1.12799430e+00 9.37332034e-01
-2.94397950e-01 4.04463768e-01 -1.52585447e-01 -1.20079410e+00
1.01293087e+00 -7.98085928e-02 -8.45288783e-02 -5.94891727e-01
-4.74344939e-04 3.46578568e-01 1.90435112e-01 -4.66802001e-01
3.99943203e-01 -3.52319889e-02 2.33738869e-01 1.61521524e-01
-9.53214914e-02 -1.08021155e-01 1.02071315e-01 -1.11651063e-01
9.06804562e-01 1.34969831e-01 -1.65955760e-02 3.51717025e-02
7.52166152e-01 -1.35445029e-01 7.72040486e-01 9.64719355e-01
-1.97227210e-01 4.61269408e-01 5.82617640e-01 -4.72980529e-01
-8.03601980e-01 -1.03862095e+00 -3.59988749e-01 9.42638755e-01
1.58462867e-01 -2.15085208e-01 -8.36875558e-01 -9.81606543e-01
-9.91240516e-02 7.88916767e-01 -4.58249658e-01 -4.05714214e-01
-4.27196801e-01 -8.62759113e-01 3.48230720e-01 3.92269611e-01
9.73911166e-01 -1.13322532e+00 -1.34447664e-01 3.45637172e-01
-2.80238390e-01 -7.00792015e-01 -2.60944307e-01 4.00231004e-01
-6.25625610e-01 -1.17976320e+00 -5.32493830e-01 -8.10737848e-01
1.09961367e+00 -1.90209657e-01 1.06468260e+00 4.01131690e-01
2.70942658e-01 -4.17611122e-01 -3.20441335e-01 -2.13306665e-01
-5.39425731e-01 1.08187549e-01 -1.48944497e-01 1.46164283e-01
4.07403648e-01 -1.80237010e-01 -4.94472444e-01 4.66305882e-01
-7.54985631e-01 1.38170660e-01 7.41768241e-01 1.06026232e+00
8.00038218e-01 3.89986217e-01 8.73256087e-01 -9.40353632e-01
5.43505132e-01 -4.71613944e-01 -7.06454277e-01 4.17901218e-01
-8.42656791e-01 2.01798543e-01 9.81365144e-01 -5.58091700e-01
-8.61659527e-01 -9.91662145e-02 -6.05645895e-01 -5.55091262e-01
3.99744362e-02 8.16608250e-01 -1.79356530e-01 5.00839353e-02
4.20471549e-01 4.72323716e-01 -1.33891732e-01 -4.01389629e-01
-9.78208780e-02 6.55267298e-01 4.66922939e-01 -4.86808181e-01
7.07755566e-01 -7.55507499e-02 -2.80147612e-01 -4.51789618e-01
-1.14819980e+00 2.38348544e-02 -4.53268349e-01 -3.12657297e-01
6.73366427e-01 -6.01736188e-01 -6.31585896e-01 5.99026918e-01
-9.01137829e-01 -5.84250987e-01 5.00335880e-02 3.41487050e-01
-2.28626981e-01 2.62516022e-01 -6.62418127e-01 -5.36824107e-01
-1.69702560e-01 -1.45703769e+00 7.45558619e-01 1.40842482e-01
-2.66429424e-01 -8.45201969e-01 -3.44687402e-01 2.07378253e-01
2.76572406e-01 -2.28289694e-01 1.09297097e+00 -7.30727851e-01
-6.70812011e-01 -4.39255565e-01 -2.55687326e-01 8.39358091e-01
-1.26213402e-01 2.71517709e-02 -9.82530236e-01 -2.32985139e-01
-1.34864133e-02 -4.47910130e-01 9.90782440e-01 3.08387876e-01
1.84028864e+00 -1.69433713e-01 -2.68830299e-01 5.68019032e-01
1.07726026e+00 9.74923372e-02 6.64515853e-01 1.43660247e-01
4.22128677e-01 1.80687517e-01 7.71149874e-01 3.60929310e-01
3.24086845e-01 3.14863920e-01 3.64788055e-01 2.20203817e-01
-9.42651927e-03 -4.29944694e-01 2.64573097e-01 9.18342054e-01
2.61595964e-01 -3.38824600e-01 -8.01946163e-01 3.59444946e-01
-1.59791481e+00 -6.97306037e-01 3.31241600e-02 2.19526172e+00
1.06085050e+00 6.78774893e-01 -4.43589091e-01 3.88681591e-01
7.81141996e-01 -5.32349981e-02 -7.58744597e-01 -9.84329283e-02
2.09139779e-01 3.67660105e-01 1.63076773e-01 5.89810133e-01
-8.74662280e-01 8.50606322e-01 6.42878437e+00 1.02055740e+00
-1.32447791e+00 -8.91948119e-02 1.20072687e+00 1.54948860e-01
-4.87044156e-01 -2.15856601e-02 -1.13270283e+00 7.92495787e-01
8.62192512e-01 1.29142821e-01 3.84687543e-01 9.85118866e-01
1.39336944e-01 -2.97827274e-01 -1.32795656e+00 7.98107505e-01
1.77380815e-02 -1.19299066e+00 1.38083428e-01 -2.06265092e-01
5.46700358e-01 -5.71864136e-02 -3.06781270e-02 8.22326005e-01
2.25092605e-01 -8.02726686e-01 7.43031561e-01 6.20617032e-01
6.44519448e-01 -7.24113584e-01 8.74245584e-01 6.84562325e-01
-8.33295643e-01 3.99919339e-02 -5.95376968e-01 5.72500527e-02
-1.81068331e-01 1.12514806e+00 -1.12479663e+00 3.02554190e-01
5.66729665e-01 7.20442176e-01 -7.75960684e-01 1.11603272e+00
-5.32663584e-01 8.75206649e-01 -2.52660125e-01 -1.88889593e-01
1.53613359e-01 2.46033072e-03 8.87762904e-02 9.57133651e-01
4.44320202e-01 -1.04188830e-01 3.17023993e-01 1.09300661e+00
-2.88449079e-01 -3.21761280e-01 2.37122402e-02 2.03726262e-01
4.44584608e-01 9.80989873e-01 -5.65394640e-01 -6.95668817e-01
-1.57312125e-01 1.02982140e+00 5.44200420e-01 2.90243924e-01
-1.11172152e+00 -3.17091525e-01 2.96953946e-01 8.60331655e-02
2.62124509e-01 2.36587971e-01 -3.74614030e-01 -1.26482582e+00
2.66568244e-01 -8.70479345e-01 1.66758329e-01 -7.41956353e-01
-1.35171008e+00 6.70230329e-01 -2.92253017e-01 -1.24893928e+00
-3.94121945e-01 -4.71683234e-01 -4.95467186e-01 9.48236108e-01
-1.69601023e+00 -5.75594187e-01 -4.82538611e-01 5.37388325e-01
5.42955101e-01 1.57434732e-01 5.94660461e-01 3.79382193e-01
-7.41031766e-01 9.88590240e-01 -1.34340495e-01 1.66697338e-01
7.04812825e-01 -1.12000036e+00 1.73294753e-01 9.66966510e-01
-3.09503913e-01 5.17504096e-01 5.83585501e-01 -7.82123327e-01
-9.96583223e-01 -1.43351698e+00 9.14061308e-01 -1.69320539e-01
4.68820721e-01 -1.23305090e-01 -1.28104031e+00 7.02735543e-01
-2.43273869e-01 1.81841627e-01 3.75950277e-01 1.37017205e-01
-1.90120086e-01 -2.89405763e-01 -1.22634220e+00 6.09130740e-01
6.84397995e-01 -3.67480725e-01 -6.83535397e-01 2.71320462e-01
5.92314005e-01 -3.05329412e-01 -6.73844934e-01 6.38777852e-01
4.42103624e-01 -9.03514981e-01 5.82883477e-01 -9.02113616e-02
6.11719966e-01 -4.64559793e-01 7.70798475e-02 -1.49075866e+00
-3.87858957e-01 -1.17207408e-01 2.63742637e-02 1.03044081e+00
6.61968410e-01 -8.25537264e-01 9.86108899e-01 7.03900099e-01
-3.28543484e-01 -1.06086564e+00 -6.68845773e-01 -4.34664369e-01
-1.98048443e-01 -8.26756001e-01 6.24618709e-01 7.51591086e-01
-1.63147688e-01 3.35589424e-02 -2.78218389e-01 3.54475081e-01
5.82884431e-01 4.72033955e-02 7.30261803e-01 -1.11562538e+00
-5.92387319e-01 -3.78908634e-01 -2.13474274e-01 -1.38902473e+00
1.68121025e-01 -7.78146327e-01 5.49498916e-01 -1.35997486e+00
2.49754563e-01 -9.16563988e-01 -5.04038274e-01 6.48401260e-01
-6.59875095e-01 3.22698593e-01 1.61331985e-03 1.51794791e-01
-6.46036267e-01 5.32225251e-01 1.40175700e+00 -1.78567857e-01
-4.61149588e-02 3.16124588e-01 -5.91614008e-01 7.65143812e-01
6.67206585e-01 -3.55429530e-01 -4.00313348e-01 -3.94487321e-01
2.19223991e-01 4.90110070e-02 4.37941134e-01 -1.18726635e+00
2.77912021e-01 -1.78805903e-01 7.64778614e-01 -9.17539656e-01
1.13619134e-01 -6.80045366e-01 -1.03043571e-01 6.44420028e-01
-5.00674009e-01 -2.69756854e-01 -2.14171782e-02 5.09481430e-01
-4.11055267e-01 -5.20477951e-01 9.73443091e-01 2.76726410e-02
-6.73958123e-01 4.96491969e-01 -1.68722808e-01 -2.03988761e-01
8.21104407e-01 -1.76310614e-01 -2.69131452e-01 -1.60117894e-01
-9.03183818e-01 4.80012596e-01 2.42155045e-01 1.55321285e-01
7.52009630e-01 -1.07992125e+00 -5.73492646e-01 7.00059056e-01
8.26022867e-03 4.47623789e-01 1.09805673e-01 7.34999537e-01
-3.91996711e-01 1.00954637e-01 7.76004791e-02 -7.28434205e-01
-8.67071867e-01 3.66708755e-01 4.19591814e-01 -4.10810947e-01
-4.16176200e-01 9.46528494e-01 2.27280259e-02 -4.80068326e-01
3.28911841e-01 -6.25232577e-01 -8.70580971e-02 -1.36594623e-01
6.37912393e-01 -5.63305728e-02 2.16541409e-01 4.32212390e-02
-2.56013662e-01 2.15390995e-01 -4.20800328e-01 2.12831020e-01
1.24913847e+00 2.61979401e-02 -2.07267165e-01 3.42938215e-01
1.02466083e+00 -1.70216769e-01 -1.41192913e+00 -2.34206706e-01
-1.27534047e-01 -3.86148155e-01 -9.71710533e-02 -1.00833762e+00
-1.09961367e+00 7.80434668e-01 2.93500125e-01 2.05664784e-01
1.26975918e+00 -1.31160766e-01 7.36573100e-01 5.76571584e-01
2.83673882e-01 -9.96665597e-01 -2.37417854e-02 7.55696893e-01
6.01715088e-01 -1.17486107e+00 -3.17631662e-01 -3.72273266e-01
-2.65701473e-01 1.15075517e+00 8.53000641e-01 -1.84601128e-01
7.51077116e-01 1.52225271e-01 -2.80527174e-01 1.96921661e-01
-8.16066623e-01 1.62267804e-01 2.55394936e-01 2.03762412e-01
2.08084553e-01 1.41523570e-01 -1.22865222e-01 6.39604032e-01
-2.51901150e-01 3.89064610e-01 3.45441848e-01 6.21857285e-01
-6.44105077e-01 -9.42134321e-01 -2.88265467e-01 9.46171999e-01
-2.85907984e-01 -2.61128873e-01 -1.13112584e-01 4.64058578e-01
3.14226627e-01 9.07351494e-01 9.74258110e-02 -7.40968943e-01
2.14234352e-01 2.47046918e-01 2.48209208e-01 -5.48722684e-01
-2.95912474e-01 -1.19307779e-01 -7.69112855e-02 -3.96617919e-01
-2.40023397e-02 -2.95005620e-01 -1.10425901e+00 -4.21376526e-01
-4.33385551e-01 1.90011621e-01 3.44720483e-01 1.23026299e+00
1.83022916e-01 7.20607281e-01 6.23032749e-01 -7.27206230e-01
-9.16258693e-01 -1.06569505e+00 -3.40353549e-01 3.63616794e-01
3.09157848e-01 -5.18152952e-01 -5.56173384e-01 -6.17227890e-03] | [9.368674278259277, 3.7293245792388916] |
649865c2-e6ce-4a88-a922-9d6741f21cb4 | rapid-inr-storage-efficient-cpu-free-dnn | 2306.16699 | null | https://arxiv.org/abs/2306.16699v1 | https://arxiv.org/pdf/2306.16699v1.pdf | Rapid-INR: Storage Efficient CPU-free DNN Training Using Implicit Neural Representation | Implicit Neural Representation (INR) is an innovative approach for representing complex shapes or objects without explicitly defining their geometry or surface structure. Instead, INR represents objects as continuous functions. Previous research has demonstrated the effectiveness of using neural networks as INR for image compression, showcasing comparable performance to traditional methods such as JPEG. However, INR holds potential for various applications beyond image compression. This paper introduces Rapid-INR, a novel approach that utilizes INR for encoding and compressing images, thereby accelerating neural network training in computer vision tasks. Our methodology involves storing the whole dataset directly in INR format on a GPU, mitigating the significant data communication overhead between the CPU and GPU during training. Additionally, the decoding process from INR to RGB format is highly parallelized and executed on-the-fly. To further enhance compression, we propose iterative and dynamic pruning, as well as layer-wise quantization, building upon previous work. We evaluate our framework on the image classification task, utilizing the ResNet-18 backbone network and three commonly used datasets with varying image sizes. Rapid-INR reduces memory consumption to only 5% of the original dataset size and achieves a maximum 6$\times$ speedup over the PyTorch training pipeline, as well as a maximum 1.2x speedup over the DALI training pipeline, with only a marginal decrease in accuracy. Importantly, Rapid-INR can be readily applied to other computer vision tasks and backbone networks with reasonable engineering efforts. Our implementation code is publicly available at https://anonymous.4open.science/r/INR-4BF7. | ['Cong Hao', 'Stephen BR Fitzmeyer', 'Hang Yang', 'Hanqiu Chen'] | 2023-06-29 | null | null | null | null | ['image-compression', 'quantization'] | ['computer-vision', 'methodology'] | [ 6.03219569e-01 -7.60450512e-02 -5.46748526e-02 -3.21703404e-01
-3.49594653e-01 -2.52303123e-01 3.21482480e-01 1.98650986e-01
-9.23391759e-01 3.79917145e-01 -1.66362539e-01 -6.60750628e-01
8.49573389e-02 -1.13000572e+00 -1.04301643e+00 -4.62403506e-01
2.13632826e-02 2.20502540e-01 2.19966367e-01 1.85683176e-01
2.31260478e-01 8.62749457e-01 -1.86382091e+00 4.04846340e-01
5.03156185e-01 1.33241057e+00 3.85539651e-01 8.64153266e-01
-1.20643571e-01 1.01804256e+00 -5.58094263e-01 -4.71230298e-01
5.81697464e-01 2.11670786e-01 -6.55048370e-01 -1.03176765e-01
8.62556696e-01 -8.74491394e-01 -6.06191158e-01 9.41104949e-01
5.18510044e-01 2.70898491e-02 1.40255347e-01 -8.58156025e-01
-4.54522938e-01 3.84490579e-01 -5.27652621e-01 6.46072626e-02
-1.55159548e-01 7.51816034e-02 6.29701138e-01 -8.57212007e-01
3.06055337e-01 1.04458988e+00 9.01625276e-01 4.04511213e-01
-1.08874404e+00 -8.51607382e-01 -1.46031827e-01 2.37403125e-01
-1.50178230e+00 -5.90380132e-01 4.13195074e-01 -9.17029977e-02
1.40798771e+00 2.89166868e-01 8.83069158e-01 4.83468026e-01
-2.15153500e-01 5.52113652e-01 7.69737244e-01 -2.99540818e-01
3.11719298e-01 -3.63465369e-01 2.53595293e-01 8.82150054e-01
3.61591011e-01 -3.30451056e-02 -5.03003061e-01 -3.05930302e-02
1.22894812e+00 1.47530228e-01 -2.94134051e-01 8.93094093e-02
-9.01068389e-01 6.62760615e-01 6.92131042e-01 -2.33576775e-01
-3.56527865e-01 7.13868141e-01 6.11611724e-01 2.64242619e-01
4.20920104e-01 6.65984750e-02 -5.62964678e-01 -3.50064963e-01
-1.05363834e+00 2.93080866e-01 7.67775536e-01 9.36053813e-01
8.85773659e-01 1.91242069e-01 1.72304749e-01 1.01512051e+00
1.17894048e-02 2.80870974e-01 3.82512093e-01 -1.36777246e+00
4.80028123e-01 5.03126442e-01 -3.23080033e-01 -1.08164907e+00
-1.57524124e-01 -4.93356109e-01 -1.13345706e+00 3.51263553e-01
1.70796707e-01 1.02927364e-01 -1.02585280e+00 1.58633900e+00
2.86115170e-01 3.31758946e-01 3.68350670e-02 6.56897128e-01
9.52820003e-01 7.25026488e-01 -5.83812334e-02 2.60684073e-01
1.29707778e+00 -1.09252870e+00 -1.28828555e-01 -1.73215091e-01
6.14250779e-01 -7.30631232e-01 1.12382376e+00 5.86510599e-01
-1.41813648e+00 -5.19905090e-01 -1.27083337e+00 -6.33751392e-01
-2.60497123e-01 3.56225252e-01 7.56493509e-01 4.29464340e-01
-1.40829051e+00 9.74658489e-01 -1.07157242e+00 3.07866372e-03
8.31101060e-01 7.84493268e-01 -2.16152444e-01 -2.37740844e-01
-5.53166986e-01 3.83517087e-01 6.38110042e-01 5.13228513e-02
-4.78388160e-01 -9.49690938e-01 -8.80443513e-01 2.51529694e-01
1.22657225e-01 -6.59908354e-01 1.36625874e+00 -1.00273478e+00
-1.46522617e+00 5.74980915e-01 -1.49077937e-01 -9.29937005e-01
3.11493605e-01 -3.34031075e-01 1.11682504e-01 3.38467032e-01
-4.66940999e-01 1.20928097e+00 7.70727575e-01 -7.22003818e-01
-4.61216092e-01 -2.65080452e-01 2.82423615e-01 2.27821708e-01
-7.08562076e-01 -9.77791473e-02 -9.38891113e-01 -8.12837601e-01
1.90056115e-01 -8.52728128e-01 -2.95795411e-01 6.77140117e-01
-3.82561535e-02 -8.94858781e-03 8.97732317e-01 -6.97653413e-01
8.36989224e-01 -2.10141683e+00 -3.07899088e-01 1.72157004e-01
3.35483938e-01 6.27211034e-01 -9.14369822e-02 -1.28097624e-01
4.09808792e-02 7.33739361e-02 -4.76207703e-01 -4.99654770e-01
-2.62382925e-01 4.66700226e-01 -2.67091483e-01 3.48888636e-01
1.35132134e-01 7.82067716e-01 -3.85816067e-01 -2.57736355e-01
1.73988536e-01 9.64330733e-01 -7.98438311e-01 4.03086953e-02
-8.94534364e-02 -1.57633886e-01 -1.05229458e-02 5.43416619e-01
1.01664603e+00 -4.43717033e-01 9.04744714e-02 -5.16002357e-01
-1.65555418e-01 5.49511909e-01 -9.17607546e-01 1.65425718e+00
-5.42679071e-01 8.32625210e-01 3.36730957e-01 -1.18271279e+00
7.87305295e-01 2.59791948e-02 3.17835301e-01 -7.82501519e-01
1.89705715e-01 1.25565872e-01 -1.98447049e-01 -5.92193613e-03
7.44425774e-01 4.26973224e-01 3.91354620e-01 3.76773924e-01
-1.68403164e-01 1.58941671e-02 3.34988654e-01 1.30003646e-01
1.19336784e+00 7.93412030e-02 3.86098549e-02 4.85588498e-02
8.90628397e-02 8.54150876e-02 4.03531373e-01 7.06296325e-01
1.93383574e-01 4.92643625e-01 2.15273023e-01 -5.72688162e-01
-1.20355749e+00 -8.01923573e-01 -3.30843508e-01 9.43627417e-01
-7.27755055e-02 -6.67202234e-01 -9.20453489e-01 -1.92384481e-01
-4.94924635e-02 4.73405898e-01 -3.75783354e-01 1.74521700e-01
-8.52383614e-01 -5.90828478e-01 7.51079023e-01 7.69987762e-01
9.73222315e-01 -9.36686635e-01 -1.12100744e+00 1.13158599e-01
2.33270228e-01 -1.21601200e+00 -1.03911355e-01 3.06229025e-01
-1.33180058e+00 -9.14540589e-01 -6.30881488e-01 -7.79113173e-01
8.79079640e-01 5.54578066e-01 1.11611259e+00 4.95266169e-01
-5.80815375e-01 1.12970106e-01 -1.36752561e-01 -2.89013147e-01
-1.92534596e-01 -5.70967468e-03 -3.18331778e-01 -6.68107033e-01
2.31539458e-01 -8.11700165e-01 -8.74305308e-01 -6.45973999e-03
-1.11127281e+00 6.75267398e-01 5.94537377e-01 6.66836560e-01
8.87885749e-01 4.33063284e-02 3.88313755e-02 -7.87555456e-01
2.86407322e-01 -1.83595628e-01 -7.90633559e-01 -1.27688840e-01
-6.25854850e-01 -1.16785526e-01 8.12674642e-01 -2.56821513e-01
-7.46171057e-01 1.80910736e-01 -5.57083011e-01 -6.48773432e-01
-1.84164137e-01 5.53256571e-01 2.95416832e-01 -3.64339262e-01
4.90835249e-01 2.58586377e-01 1.84653580e-01 -5.72085917e-01
2.00722709e-01 5.78096092e-01 7.56177187e-01 -5.05524576e-01
4.49121654e-01 4.94307369e-01 5.35316840e-02 -8.76769423e-01
-5.66551685e-01 -2.58371145e-01 -3.05198997e-01 1.28164396e-01
5.17812729e-01 -1.07198036e+00 -6.85386896e-01 5.56705534e-01
-1.21155703e+00 -5.86294651e-01 -3.19120646e-01 3.28036338e-01
-3.85064125e-01 4.26592380e-01 -1.07949817e+00 -3.28121096e-01
-8.66863668e-01 -1.23151159e+00 8.72467220e-01 9.53470618e-02
1.48544908e-01 -5.68880320e-01 -4.42210197e-01 2.83942610e-01
5.66000402e-01 5.07382825e-02 7.53659546e-01 -1.74017847e-01
-8.95101547e-01 -3.55089828e-02 -7.47219145e-01 6.34476602e-01
-2.00425535e-01 -6.69900775e-02 -9.15874898e-01 -5.24684489e-01
-4.19579111e-02 -5.65578699e-01 1.03991938e+00 1.84704766e-01
2.06826591e+00 -5.39136052e-01 1.08682483e-01 1.18272638e+00
1.57224393e+00 1.49257809e-01 7.39112556e-01 4.28361297e-01
7.19599664e-01 2.80499190e-01 2.26219967e-01 6.65109336e-01
2.67529130e-01 4.74965721e-01 6.20051980e-01 -3.26657116e-01
-4.36504871e-01 -5.73547967e-02 2.85898566e-01 1.08470285e+00
-1.41717345e-01 -2.07997952e-02 -8.32508564e-01 2.32497811e-01
-1.41671002e+00 -7.00523198e-01 8.28330368e-02 2.18204951e+00
9.13309097e-01 8.49182755e-02 -3.15469831e-01 3.41720402e-01
4.03337628e-01 4.16521095e-02 -8.27285826e-01 -6.02213204e-01
3.34657691e-02 8.25414598e-01 9.75242913e-01 1.93297088e-01
-1.01159847e+00 8.41853261e-01 5.74927998e+00 1.04943991e+00
-1.52047050e+00 -2.72613727e-02 9.34230745e-01 -3.90157253e-01
6.83972463e-02 -1.36693820e-01 -7.71744370e-01 1.24239489e-01
1.10803998e+00 1.42311811e-01 7.83708036e-01 1.00015140e+00
3.59390751e-02 -1.13326460e-02 -8.65050375e-01 1.24728227e+00
-6.32410496e-02 -1.78846025e+00 1.88116357e-01 1.13017038e-01
4.04988825e-01 4.57105368e-01 1.21277556e-01 6.70220181e-02
5.10179885e-02 -1.04212427e+00 8.07218432e-01 1.25251606e-01
1.31695962e+00 -8.77687573e-01 5.08285165e-01 2.71685958e-01
-1.33903134e+00 4.73308042e-02 -7.48605311e-01 -1.77999988e-01
-1.84623256e-01 6.04614377e-01 -7.96977937e-01 2.48776630e-01
1.05275559e+00 8.61553133e-01 -6.53149843e-01 8.90267372e-01
7.59166777e-02 8.35310578e-01 -7.17851162e-01 2.74750918e-01
1.67916536e-01 -1.34492368e-01 7.94508830e-02 1.34182322e+00
5.42333364e-01 9.73363668e-02 -7.52966031e-02 5.74552178e-01
-4.27683145e-01 -2.08185008e-03 -3.41516644e-01 -7.50375306e-03
5.29081106e-01 1.11268616e+00 -8.87120962e-01 -5.88637590e-01
-4.21060473e-01 1.00430775e+00 5.59709311e-01 1.43987566e-01
-8.66053462e-01 -5.96148670e-01 5.56514025e-01 7.34077170e-02
6.39302135e-01 -4.25283909e-01 -4.56895471e-01 -9.04395163e-01
2.02533081e-01 -8.61883104e-01 -5.79668880e-02 -6.55948222e-01
-7.20537901e-01 6.54663265e-01 -1.22220084e-01 -1.04566956e+00
2.15656571e-02 -8.89917970e-01 -2.77421594e-01 5.72028220e-01
-1.61221862e+00 -8.03520679e-01 -6.43775642e-01 5.47216833e-01
5.20046771e-01 -2.00390583e-03 8.14546287e-01 6.20499790e-01
-6.85063362e-01 8.02903533e-01 1.33486167e-01 7.63919950e-02
2.24765524e-01 -8.37980270e-01 8.41564894e-01 6.15283847e-01
1.11423984e-01 8.05991769e-01 2.48171255e-01 -4.18088704e-01
-1.62940419e+00 -1.46931028e+00 3.25242668e-01 5.78179002e-01
2.79833347e-01 -3.31484795e-01 -1.00218511e+00 5.24214506e-01
7.35776946e-02 1.96667254e-01 5.51264763e-01 -2.89292365e-01
-5.11908412e-01 -2.23325849e-01 -1.19472361e+00 6.50133789e-01
1.15118802e+00 -4.41390991e-01 -7.63975382e-02 3.83734822e-01
8.85651290e-01 -8.23422670e-01 -9.33544457e-01 4.95817661e-01
5.90099573e-01 -1.06081629e+00 1.24547708e+00 1.17483288e-01
6.24113381e-01 -1.72714591e-01 -3.19390863e-01 -6.09119117e-01
-1.61072001e-01 -2.30623603e-01 -4.30650204e-01 8.48246753e-01
7.43492693e-02 -6.65464461e-01 1.22391021e+00 5.16886234e-01
-2.20371962e-01 -1.21886706e+00 -8.53491008e-01 -4.98451680e-01
-2.04679996e-01 -9.19060349e-01 6.23532355e-01 5.56451142e-01
-7.63057530e-01 -1.60900593e-01 2.59589087e-02 1.23201221e-01
5.48524201e-01 -9.39721465e-02 7.59086907e-01 -9.28956032e-01
-6.20374382e-01 -4.64219570e-01 -4.45442706e-01 -1.62322903e+00
-1.73258670e-02 -1.04781055e+00 -1.50381893e-01 -1.37513185e+00
7.31719797e-03 -8.46706450e-01 -1.49358511e-01 8.64322782e-01
2.89652020e-01 7.80972958e-01 4.25324738e-01 3.29719961e-01
-2.73039073e-01 5.25312960e-01 9.84062672e-01 -1.44719005e-01
9.04128700e-02 -3.34664196e-01 -5.43155789e-01 9.35530424e-01
1.09396613e+00 -3.60449106e-01 -5.42684734e-01 -1.02953398e+00
2.94075496e-02 -1.96378827e-01 6.25802338e-01 -1.40295815e+00
2.78588623e-01 2.60615975e-01 4.73019391e-01 -8.06301177e-01
7.15448439e-01 -7.32635677e-01 1.53937086e-01 5.86601496e-01
-1.96033776e-01 4.05138373e-01 6.52861118e-01 3.72278363e-01
-1.73745528e-01 -5.09266555e-01 7.11877704e-01 -2.03072622e-01
-5.87048292e-01 5.05558729e-01 -1.31299973e-01 -2.28413031e-01
5.99240184e-01 -4.82552141e-01 -4.18333620e-01 -8.87179822e-02
-3.27328861e-01 -3.09233993e-01 5.83538234e-01 -1.10809401e-01
9.65065062e-01 -9.11922395e-01 -4.24605846e-01 3.95967096e-01
-4.72199053e-01 6.00246191e-01 2.11649701e-01 4.11940575e-01
-1.38215709e+00 3.71275932e-01 -1.19965583e-01 -7.59334743e-01
-1.54607785e+00 1.38083249e-01 1.70195431e-01 -1.28051549e-01
-9.76258874e-01 1.06536603e+00 8.17813277e-02 -1.83661282e-01
4.88688529e-01 -4.43655312e-01 8.67189914e-02 -4.83061314e-01
6.93055809e-01 4.76259261e-01 3.55921745e-01 -4.71162617e-01
2.98117567e-02 5.78721285e-01 -2.71803468e-01 3.64174694e-02
1.49278092e+00 3.51945698e-01 -5.04189789e-01 -8.64553303e-02
1.47428453e+00 -3.13842326e-01 -1.44017982e+00 -1.21472143e-01
-3.22553843e-01 -4.36506599e-01 4.93239284e-01 -4.93627936e-01
-1.49255705e+00 9.28563178e-01 8.22589159e-01 -1.44136280e-01
1.54832673e+00 -4.26581115e-01 1.06762815e+00 7.77573168e-01
3.75014961e-01 -8.63912761e-01 -1.73977152e-01 6.65470600e-01
8.30897212e-01 -7.81720400e-01 4.16599125e-01 -5.03125608e-01
-8.73863026e-02 1.36594367e+00 6.34475648e-01 -3.39928538e-01
6.07681036e-01 7.31434405e-01 -2.16896132e-01 3.74458940e-03
-8.38113189e-01 2.98966765e-01 -7.28367642e-03 2.77222127e-01
5.75606525e-01 -7.38669410e-02 -1.47641346e-01 5.52405231e-02
-5.80854952e-01 3.15269798e-01 3.22371423e-01 1.32261419e+00
-2.49493808e-01 -1.09215283e+00 -2.58114964e-01 8.46331358e-01
-5.82172692e-01 -4.41533029e-01 1.20527864e-01 5.02159357e-01
2.93552577e-01 4.79680240e-01 6.01988435e-01 -3.80304873e-01
1.12208322e-01 -4.10611361e-01 4.64523792e-01 -4.22345966e-01
-6.90879643e-01 -1.46905169e-01 -7.11777955e-02 -7.79856324e-01
-4.20989603e-01 -3.69124383e-01 -1.40240359e+00 -5.69787562e-01
-3.13375928e-02 -2.91627139e-01 1.07400191e+00 4.29418623e-01
7.55064666e-01 5.73736072e-01 1.67971596e-01 -1.24728513e+00
-3.85758132e-01 -6.07202113e-01 6.38038814e-02 3.46397609e-02
1.83137119e-01 -2.90004820e-01 -1.86776623e-01 9.40197483e-02] | [8.602421760559082, 2.917926788330078] |
e3f81c0a-c032-46d5-9fe8-7365a01edde0 | icdar-2021-competition-on-scientific-table | 2105.14426 | null | https://arxiv.org/abs/2105.14426v2 | https://arxiv.org/pdf/2105.14426v2.pdf | ICDAR 2021 Competition on Scientific Table Image Recognition to LaTeX | Tables present important information concisely in many scientific documents. Visual features like mathematical symbols, equations, and spanning cells make structure and content extraction from tables embedded in research documents difficult. This paper discusses the dataset, tasks, participants' methods, and results of the ICDAR 2021 Competition on Scientific Table Image Recognition to LaTeX. Specifically, the task of the competition is to convert a tabular image to its corresponding LaTeX source code. We proposed two subtasks. In Subtask 1, we ask the participants to reconstruct the LaTeX structure code from an image. In Subtask 2, we ask the participants to reconstruct the LaTeX content code from an image. This report describes the datasets and ground truth specification, details the performance evaluation metrics used, presents the final results, and summarizes the participating methods. Submission by team VCGroup got the highest Exact Match accuracy score of 74% for Subtask 1 and 55% for Subtask 2, beating previous baselines by 5% and 12%, respectively. Although improvements can still be made to the recognition capabilities of models, this competition contributes to the development of fully automated table recognition systems by challenging practitioners to solve problems under specific constraints and sharing their approaches; the platform will remain available for post-challenge submissions at https://competitions.codalab.org/competitions/26979 . | ['Mayank Singh', 'Harsh Desai', 'Mrinal Anand', 'Pratik Kayal'] | 2021-05-30 | null | null | null | null | ['table-recognition'] | ['computer-vision'] | [ 6.88896626e-02 -3.04899421e-02 -9.41103324e-02 -5.72710752e-01
-1.32319558e+00 -1.08265829e+00 5.10148525e-01 5.10199368e-01
-4.27647047e-02 6.32879615e-01 2.97363065e-02 -1.94639295e-01
9.93814990e-02 -3.92647594e-01 -1.00958407e+00 -2.56604314e-01
1.97107747e-01 6.48095787e-01 -1.75264746e-01 3.15151364e-01
6.96130216e-01 5.83446026e-01 -1.42660344e+00 1.27359438e+00
7.16391623e-01 1.27609956e+00 4.83116023e-02 8.56845319e-01
-6.53890729e-01 1.01191771e+00 -8.79603982e-01 -8.78865480e-01
1.47018835e-01 -2.90022045e-01 -8.44804347e-01 4.70122173e-02
1.30789769e+00 5.79706207e-02 -3.75867605e-01 8.90633404e-01
3.69126767e-01 -1.60245538e-01 8.30421984e-01 -1.17819583e+00
-7.53523111e-01 7.29871988e-01 -4.63243276e-01 1.71384767e-01
6.81049585e-01 -1.85039774e-01 1.06581235e+00 -1.30618012e+00
9.07256842e-01 1.10634732e+00 7.61122763e-01 4.01127905e-01
-1.11350417e+00 -7.83557177e-01 1.10882148e-01 3.58882040e-01
-1.54798293e+00 -7.11098790e-01 2.68323213e-01 -7.84980893e-01
9.56683576e-01 7.32447445e-01 3.07094187e-01 1.07873011e+00
2.34110668e-01 9.76619244e-01 1.12451029e+00 -2.21111760e-01
7.12950751e-02 5.66057503e-01 3.10661584e-01 8.59320164e-01
5.74859202e-01 -6.91882789e-01 -7.71125734e-01 7.00039715e-02
4.67636555e-01 -5.13694704e-01 -2.98052579e-02 -3.96384239e-01
-1.42027998e+00 3.53309304e-01 1.02940738e-01 1.97371110e-01
7.55252764e-02 -3.81817162e-01 4.36142683e-01 1.16348699e-01
7.46754259e-02 8.31962645e-01 -2.79842108e-01 -1.71475574e-01
-1.04752827e+00 6.24967754e-01 9.62651789e-01 1.39990795e+00
9.55026969e-02 -2.67471492e-01 -6.38564229e-01 9.18902457e-01
-1.07366689e-01 3.93830061e-01 6.32098988e-02 -1.04864514e+00
9.72609222e-01 7.18043327e-01 2.11655553e-02 -1.26336253e+00
-2.93046206e-01 -5.59426904e-01 -9.49752033e-01 -1.96810201e-01
7.19495535e-01 4.24820900e-01 -7.89024889e-01 8.78268898e-01
-9.72889662e-02 -4.84417737e-01 -1.89395294e-01 5.66612482e-01
1.57609165e+00 5.30633390e-01 -2.37288415e-01 1.18111789e-01
1.79402018e+00 -9.30894732e-01 -9.47895527e-01 -8.56685489e-02
5.18121839e-01 -1.26062679e+00 9.59030926e-01 1.03481126e+00
-1.35210323e+00 -7.70199656e-01 -1.06981337e+00 -4.99620616e-01
-6.59347892e-01 7.36425340e-01 4.49830085e-01 7.02836990e-01
-7.90356338e-01 4.10972744e-01 -4.01330143e-01 -1.11008741e-01
7.10499823e-01 2.19015792e-01 -4.96855617e-01 -1.51334092e-01
-5.93563080e-01 6.33778095e-01 -5.91973290e-02 9.33591053e-02
-3.87835920e-01 -1.06423819e+00 -7.06920385e-01 -2.90565621e-02
2.36538440e-01 -4.31090564e-01 1.15600836e+00 -1.00785233e-01
-7.95994937e-01 1.55699277e+00 -1.70488283e-01 -4.66945618e-01
7.95828462e-01 -1.31296786e-02 -3.87622058e-01 1.17310204e-01
2.40977481e-01 6.78090870e-01 3.00643861e-01 -1.13868082e+00
-6.02139294e-01 -2.76925832e-01 -3.71084243e-01 -3.51366177e-02
-1.47564560e-01 3.60023677e-02 -1.15356767e+00 -9.56908286e-01
1.82345122e-01 -8.38074148e-01 3.38572055e-01 -4.05129641e-02
-7.50518858e-01 -4.25394863e-01 1.27067044e-01 -8.85943711e-01
1.08990610e+00 -2.15881300e+00 -6.80064186e-02 5.64491823e-02
4.60792452e-01 -1.61050811e-01 -4.70706597e-02 2.72148728e-01
-1.00305758e-01 3.08258921e-01 -4.69822213e-02 -5.46858609e-01
1.37468994e-01 -4.17091578e-01 -6.08432710e-01 3.48735154e-01
-4.93260846e-02 8.41924846e-01 -3.30407321e-01 -6.32949173e-01
-6.94070011e-02 2.95399010e-01 -3.17146182e-01 1.60187811e-01
3.27094400e-04 1.31515592e-01 -2.34049827e-01 9.62159455e-01
9.53277826e-01 -4.94907647e-01 2.01832116e-01 -4.83571321e-01
-1.76270291e-01 5.83726943e-01 -1.57060051e+00 1.42869413e+00
-9.16066766e-02 8.76862943e-01 2.33148932e-01 -9.71924305e-01
1.02273428e+00 -2.69230288e-02 3.53839636e-01 -8.27755988e-01
-1.70205757e-01 3.57149214e-01 -3.09229136e-01 -3.79177809e-01
4.40631211e-01 6.11501157e-01 -3.29976290e-01 3.96179929e-02
-2.89949417e-01 -4.12822098e-01 6.95523500e-01 4.59732026e-01
9.57091391e-01 1.28926784e-01 3.99039350e-02 -2.64654696e-01
8.27261508e-01 9.30271596e-02 1.27162516e-01 1.30855155e+00
2.64483467e-02 9.59869266e-01 6.35744095e-01 -7.48133481e-01
-1.01547921e+00 -1.06815410e+00 -4.58617717e-01 7.35426068e-01
-2.41995916e-01 -7.61077523e-01 -6.98257983e-01 -6.39527857e-01
3.94511610e-01 4.42657918e-01 -7.48302102e-01 3.66065353e-01
-6.50821030e-01 -4.86136943e-01 6.78749382e-01 4.87696230e-01
5.31610727e-01 -9.39956963e-01 -3.74088064e-02 -1.05684750e-01
-5.48744619e-01 -1.50646627e+00 -6.62023842e-01 1.86130121e-01
-6.87254667e-01 -1.09448230e+00 -6.40986264e-01 -8.83686900e-01
6.55629337e-01 6.27548099e-02 1.29374409e+00 -1.22197503e-02
-7.38400877e-01 2.76718229e-01 1.75461054e-01 -7.14281142e-01
-1.89588860e-01 2.10037231e-01 -2.74368227e-01 -2.31508717e-01
2.26175517e-01 1.49104789e-01 -2.74121851e-01 2.31633395e-01
-5.10026276e-01 1.91411600e-01 6.44861460e-01 4.82154906e-01
9.72715497e-01 -1.92381054e-01 -2.77558323e-02 -1.08530867e+00
3.99981022e-01 4.77141477e-02 -9.26372766e-01 5.61974883e-01
-6.74022675e-01 1.81093991e-01 5.81274211e-01 5.04999943e-02
-7.62520730e-01 3.03543895e-01 -4.28992957e-02 -1.94470719e-01
-9.10313651e-02 2.78362304e-01 -3.00641596e-01 5.18419733e-03
6.32687509e-01 3.89300674e-01 -1.76469088e-01 -9.43871975e-01
8.13509244e-03 7.88211286e-01 8.47304642e-01 -8.07776868e-01
6.42924190e-01 2.75889009e-01 -2.12749615e-02 -6.46757007e-01
-1.13720870e+00 -5.69951177e-01 -4.96764004e-01 8.58896300e-02
7.88171768e-01 -1.07749581e+00 -1.11878681e+00 4.30382371e-01
-1.17140818e+00 6.70143515e-02 -4.33537737e-02 2.29315668e-01
-3.73598814e-01 3.58497202e-01 -7.50638306e-01 -3.82419378e-01
-2.44476438e-01 -1.23000371e+00 1.22235203e+00 -1.30261123e-01
-4.41415787e-01 -5.47140002e-01 -2.41569400e-01 1.23512077e+00
8.75913352e-02 2.53182650e-01 8.06592107e-01 -7.65223861e-01
-9.36362088e-01 -3.71451497e-01 -4.21233803e-01 -2.87241489e-02
-3.48759033e-02 8.80183205e-02 -1.02651918e+00 -1.39015734e-01
-3.30273092e-01 -4.90525544e-01 1.03644776e+00 1.66550875e-01
2.08654928e+00 -3.77538174e-01 -3.67502511e-01 7.45773017e-01
8.87651682e-01 2.75878519e-01 6.41343772e-01 5.07831573e-01
7.83255398e-01 8.78176570e-01 3.08358997e-01 2.39099637e-01
4.73213524e-01 8.82625878e-01 -8.39783549e-02 1.41573533e-01
-4.55191582e-01 -8.12840536e-02 1.41078889e-01 9.03136849e-01
2.89895684e-01 -1.95187554e-01 -1.15474141e+00 2.77143121e-01
-1.39242280e+00 -8.29188228e-01 -6.45420611e-01 2.19942880e+00
1.07438087e+00 3.71387184e-01 1.38276994e-01 8.75930339e-02
4.60077286e-01 -1.45072237e-01 -3.22811037e-01 -5.09969831e-01
-3.74380708e-01 8.89056474e-02 3.97861600e-01 4.03260678e-01
-1.34493947e+00 6.16090536e-01 6.55276871e+00 1.05047786e+00
-6.82391107e-01 -3.02446753e-01 1.07176042e+00 -2.08272934e-01
-3.95888872e-02 -5.06181002e-01 -1.42329550e+00 3.91284078e-01
1.03049505e+00 -6.32961765e-02 3.80532384e-01 7.50558555e-01
-2.29628444e-01 1.08803809e-01 -1.36713982e+00 1.45527756e+00
3.63768011e-01 -1.91751397e+00 3.74336809e-01 4.11396250e-02
5.31493843e-01 -4.24098253e-01 3.11708361e-01 3.54075849e-01
-1.27131566e-01 -1.52522337e+00 8.92316341e-01 6.53190672e-01
1.01122224e+00 -5.91732025e-01 6.93051577e-01 -8.26986209e-02
-1.15707576e+00 3.42062324e-01 -1.90216482e-01 1.69713512e-01
-7.19180882e-01 6.85475469e-01 -7.25806594e-01 5.12523770e-01
1.09765244e+00 7.71750450e-01 -1.24568307e+00 1.03713238e+00
5.98992258e-02 3.12162936e-01 -3.51738557e-02 -2.67082691e-01
-2.35365078e-01 -3.93567830e-02 1.98458105e-01 1.76523101e+00
6.29479960e-02 -2.11380064e-01 2.66928989e-02 1.03955746e+00
-6.74457669e-01 3.22563648e-01 -4.67361122e-01 -3.86616558e-01
5.29514968e-01 1.29114223e+00 -9.29779589e-01 -5.80606222e-01
-2.86982238e-01 8.50965917e-01 2.67663807e-01 2.10870989e-02
-5.40308714e-01 -8.79623652e-01 3.56854767e-01 1.87210992e-01
4.48235929e-01 1.38135672e-01 -1.17621005e+00 -1.12831986e+00
4.88097131e-01 -1.44805372e+00 6.51176512e-01 -9.17756259e-01
-1.21177721e+00 6.03586972e-01 -1.45514205e-01 -1.03260398e+00
1.40903503e-01 -8.91163945e-01 1.29055962e-01 8.78381431e-01
-8.77425015e-01 -8.54425013e-01 -5.18291891e-01 2.25554392e-01
5.98838985e-01 -5.16604364e-01 8.45107496e-01 3.55153918e-01
-5.91861546e-01 1.17785895e+00 2.72342801e-01 4.78950441e-01
1.10344648e+00 -1.48236346e+00 5.57810664e-01 5.29297411e-01
4.88447636e-01 9.28225279e-01 8.04200530e-01 -7.54682422e-01
-1.69977736e+00 -8.73648286e-01 1.45971322e+00 -1.16003036e+00
4.99325663e-01 -1.02473128e+00 -9.94775295e-01 4.88813281e-01
1.49031639e-01 -3.54994349e-02 5.58194280e-01 1.82829514e-01
-7.79711485e-01 -4.17625666e-01 -9.87251103e-01 4.53910470e-01
9.08057630e-01 -5.18233716e-01 -3.98431152e-01 7.63050556e-01
1.47924259e-01 -1.06247354e+00 -1.00392246e+00 1.96059674e-01
8.79052699e-01 -6.23674989e-01 1.17724061e+00 -5.14475167e-01
8.06319952e-01 -2.83200055e-01 -6.56943163e-03 -5.66257834e-01
-3.79265279e-01 -4.74222422e-01 -1.12110659e-01 1.19503200e+00
5.74957192e-01 1.24807833e-02 1.12287652e+00 4.79653418e-01
-7.44661018e-02 -6.34397388e-01 -6.96540534e-01 -9.05161440e-01
3.65415812e-01 -3.70134652e-01 2.96333402e-01 9.08775866e-01
-9.63560715e-02 3.04902434e-01 -8.64509717e-02 -1.97587729e-01
7.32351422e-01 3.74646515e-01 8.09593260e-01 -1.12808156e+00
-1.97860971e-01 -8.69352579e-01 -1.62553251e-01 -8.23109150e-01
-1.27284423e-01 -1.31880307e+00 -2.05826998e-01 -1.74577570e+00
4.93974030e-01 -1.41976804e-01 -1.34358853e-01 5.92964768e-01
-8.83108079e-02 5.28388560e-01 4.53719407e-01 3.68394673e-01
-9.73921955e-01 -1.99455917e-01 1.02847695e+00 -6.42264783e-01
1.52792364e-01 -1.44304290e-01 -1.02392125e+00 2.17518851e-01
7.78600276e-01 -3.79275054e-01 -8.03679973e-02 -2.57830083e-01
4.20447052e-01 -1.53224856e-01 1.27608031e-01 -1.11459732e+00
3.64921838e-01 9.67990160e-02 1.01319146e+00 -9.85069215e-01
1.16715431e-01 -3.57484728e-01 4.80320342e-02 4.06195790e-01
-7.44319975e-01 2.85355031e-01 6.73371673e-01 1.16629034e-01
-8.64602178e-02 -1.32409766e-01 5.84964097e-01 -1.73782155e-01
-2.38145143e-01 -1.51620358e-01 -5.19358277e-01 3.91008496e-01
6.00012183e-01 -2.43758589e-01 -5.36095023e-01 -1.01151466e-01
-6.79494023e-01 2.39284098e-01 2.67427772e-01 5.92271030e-01
7.04728663e-01 -1.30503726e+00 -9.47956741e-01 2.15106934e-01
4.87308294e-01 -2.84657985e-01 2.14292005e-01 8.13069105e-01
-7.87693620e-01 1.02950871e+00 -3.15600723e-01 -4.78829622e-01
-1.62061918e+00 5.42092979e-01 1.49013922e-01 -4.28419828e-01
-5.34647286e-01 1.14658356e+00 2.32974797e-01 -5.03264725e-01
8.86956453e-01 -4.10774052e-01 -2.97573626e-01 1.67627200e-01
8.70917976e-01 2.93603092e-01 8.21590424e-01 -1.18739404e-01
-6.97468698e-01 4.50925767e-01 -6.25884473e-01 1.71646029e-01
1.09473920e+00 3.55097145e-01 -1.00433230e-01 5.56719959e-01
1.22440624e+00 5.77326715e-01 -6.33891225e-01 4.51329462e-02
1.12738572e-01 -3.25798482e-01 -2.26737782e-01 -1.37587368e+00
-9.23705161e-01 8.51177394e-01 4.49383825e-01 -1.27977403e-02
6.99508131e-01 -1.11334600e-01 4.26585168e-01 6.62214160e-01
-8.22426602e-02 -1.03291631e+00 1.05030939e-01 5.16697705e-01
1.28639615e+00 -1.25949633e+00 4.04824257e-01 -8.21016550e-01
-5.64113915e-01 1.26522934e+00 5.89812219e-01 2.42631361e-01
2.13709682e-01 6.70037389e-01 8.66689719e-03 -2.52893060e-01
-8.31207097e-01 4.35686529e-01 9.94888842e-01 4.15834218e-01
9.52539980e-01 -2.11256053e-02 -1.62546784e-01 8.38682473e-01
-5.74743211e-01 -1.19535223e-01 6.00009620e-01 8.98745298e-01
-8.25553834e-02 -1.13416326e+00 -7.68560469e-01 8.32943320e-01
-6.51502371e-01 -3.12800556e-01 -1.07990909e+00 6.58737957e-01
5.79785705e-02 7.43422687e-01 2.12688640e-01 -2.67738134e-01
7.39885211e-01 4.70648222e-02 6.91406608e-01 -6.11165345e-01
-9.40896690e-01 1.02840820e-02 2.92292446e-01 -6.09848440e-01
3.41354191e-01 -9.47102427e-01 -1.07332587e+00 -5.90022147e-01
5.56732833e-01 2.83514082e-01 7.38685668e-01 3.70354503e-01
5.91522098e-01 6.78526819e-01 4.42406274e-02 -1.28395230e-01
-4.66372013e-01 -6.77141488e-01 -2.06648827e-01 3.89306456e-01
1.59766808e-01 -4.91904765e-01 -1.85987651e-01 4.60599154e-01] | [11.687643051147461, 3.013688802719116] |
69a0893f-a406-444f-a2eb-115ddee5f88f | an-end-to-end-deep-learning-approach-for-1 | 2108.07453 | null | https://arxiv.org/abs/2108.07453v1 | https://arxiv.org/pdf/2108.07453v1.pdf | An End-to-End Deep Learning Approach for Epileptic Seizure Prediction | An accurate seizure prediction system enables early warnings before seizure onset of epileptic patients. It is extremely important for drug-refractory patients. Conventional seizure prediction works usually rely on features extracted from Electroencephalography (EEG) recordings and classification algorithms such as regression or support vector machine (SVM) to locate the short time before seizure onset. However, such methods cannot achieve high-accuracy prediction due to information loss of the hand-crafted features and the limited classification ability of regression and SVM algorithms. We propose an end-to-end deep learning solution using a convolutional neural network (CNN) in this paper. One and two dimensional kernels are adopted in the early- and late-stage convolution and max-pooling layers, respectively. The proposed CNN model is evaluated on Kaggle intracranial and CHB-MIT scalp EEG datasets. Overall sensitivity, false prediction rate, and area under receiver operating characteristic curve reaches 93.5%, 0.063/h, 0.981 and 98.8%, 0.074/h, 0.988 on two datasets respectively. Comparison with state-of-the-art works indicates that the proposed model achieves exceeding prediction performance. | ['Mohamad Sawan', 'Hemmings Wu', 'Shiqi Zhao', 'Jie Yang', 'Yankun Xu'] | 2021-08-17 | null | null | null | null | ['seizure-prediction'] | ['medical'] | [-1.31124482e-01 -3.35700423e-01 2.09836483e-01 -3.60787988e-01
-4.47562516e-01 -1.19062342e-01 2.55909353e-01 3.26860398e-01
-6.06436253e-01 1.01939380e+00 -1.52001709e-01 -2.70087242e-01
-3.70598495e-01 -4.28980321e-01 -1.45866022e-01 -6.71335876e-01
-7.09561408e-01 -1.05282098e-01 1.81669131e-01 1.89119969e-02
3.52103978e-01 5.45738935e-01 -1.16325521e+00 4.43493307e-01
1.01660049e+00 1.48182559e+00 1.88363329e-01 5.75909734e-01
2.97503114e-01 6.13740802e-01 -5.85100770e-01 1.54928386e-01
1.75738707e-01 -2.39364296e-01 -2.73362726e-01 -3.49781692e-01
-5.28887868e-01 -2.40388080e-01 -5.32434523e-01 9.57804263e-01
7.94139087e-01 -2.76726276e-01 7.66842484e-01 -1.29575872e+00
-3.58702868e-01 2.26765171e-01 -5.32455802e-01 6.96796715e-01
1.17409460e-01 3.40659805e-02 1.28816932e-01 -7.30228841e-01
-1.19771145e-01 1.04181252e-01 8.69898200e-01 5.69769442e-01
-8.12778831e-01 -1.24597740e+00 -4.20872927e-01 4.28580940e-01
-1.70745814e+00 -3.53105992e-01 5.73557019e-01 -7.72115171e-01
1.28701794e+00 1.70127794e-01 8.10288668e-01 8.02137077e-01
1.01978123e+00 4.39978689e-01 1.30466425e+00 -5.59893809e-02
1.93044394e-01 1.29340142e-01 2.05356330e-01 3.71958911e-01
1.34029090e-02 3.61601830e-01 -6.39489055e-01 -2.84438372e-01
7.31593430e-01 3.49664986e-01 -6.99551880e-01 3.26081991e-01
-1.09526825e+00 6.00852907e-01 4.96107340e-01 3.98623466e-01
-8.20383668e-01 -3.29315245e-01 5.74576795e-01 3.57855111e-01
4.57905650e-01 3.13938975e-01 -7.24622428e-01 -3.41569304e-01
-1.20587003e+00 -7.93485045e-02 6.55121207e-01 6.18849516e-01
2.23815262e-01 2.06118345e-01 -1.41766354e-01 5.96890450e-01
-7.96149969e-02 3.45616303e-02 1.15608549e+00 3.68155897e-01
1.94517195e-01 6.28146231e-01 9.28512812e-02 -6.89985216e-01
-9.73237693e-01 -8.25221002e-01 -1.16663921e+00 5.70060387e-02
-5.34878746e-02 -4.40705240e-01 -1.03168762e+00 1.18232429e+00
-4.58546340e-01 6.00858986e-01 3.55909705e-01 7.98421562e-01
8.47605824e-01 4.93545711e-01 2.30789483e-01 -4.59339172e-01
1.21962368e+00 -5.58095276e-01 -6.78580284e-01 -1.56329840e-01
6.71678722e-01 -6.05927765e-01 5.21048248e-01 8.50718021e-01
-6.67921841e-01 -2.57628202e-01 -1.18058741e+00 6.05706871e-01
-2.97331423e-01 6.31013632e-01 5.51554382e-01 4.22800303e-01
-8.99472117e-01 5.65890849e-01 -9.96721745e-01 -8.56561810e-02
6.75487876e-01 1.01446795e+00 -5.22199869e-01 3.79867941e-01
-1.14389741e+00 9.84492719e-01 4.75991815e-01 1.42253295e-01
-7.42408633e-01 -6.22550130e-01 -3.51289511e-01 1.40850753e-01
-4.91634220e-01 -2.21435633e-02 9.81725216e-01 -8.16267312e-01
-1.28527260e+00 3.53356063e-01 1.04408376e-01 -8.15806806e-01
3.80227238e-01 -1.63669780e-01 -7.62926817e-01 1.61254704e-02
-1.61236614e-01 3.69809687e-01 5.23626685e-01 -2.86725014e-01
-8.83831501e-01 -6.19676530e-01 -5.96242189e-01 5.91149628e-02
-3.69579673e-01 2.30895177e-01 3.87208998e-01 -5.85162103e-01
1.41716868e-01 -5.37924170e-01 -7.24888891e-02 -4.46036696e-01
-2.63274848e-01 -1.98374063e-01 8.63842964e-01 -1.03678393e+00
1.34181190e+00 -2.03296757e+00 -4.37550068e-01 1.03106886e-01
2.45471314e-01 4.78855729e-01 3.44808042e-01 2.00334296e-01
-5.24653077e-01 -4.38048840e-01 -8.17691609e-02 2.64928818e-01
-4.24744040e-01 -5.11108637e-01 -1.53722614e-01 6.67565167e-01
1.80481046e-01 8.03862512e-01 -5.41297197e-01 1.87475458e-01
3.59119475e-01 8.33366990e-01 1.67660471e-02 3.72196853e-01
5.49773335e-01 5.49343348e-01 -4.15598065e-01 6.09136164e-01
5.15109897e-01 -4.84564155e-02 -3.87712926e-01 -9.45811942e-02
-1.59869343e-01 1.19656138e-01 -7.67129064e-01 1.40904343e+00
-2.67386556e-01 9.54023182e-01 -6.11187935e-01 -1.30175745e+00
1.35816646e+00 7.70273268e-01 3.70503634e-01 -6.56340897e-01
3.61896127e-01 4.84317958e-01 4.11128521e-01 -7.07166553e-01
-2.16909811e-01 -1.58550441e-01 3.08930904e-01 2.07404885e-02
6.64250404e-02 4.24107879e-01 -4.60136205e-01 -3.05142313e-01
1.35249734e+00 -3.88781816e-01 6.92442417e-01 -4.94776696e-01
5.60114920e-01 -2.96134859e-01 4.71856534e-01 3.50264996e-01
-3.22382390e-01 2.99234629e-01 3.27993006e-01 -7.98718452e-01
-6.21663213e-01 -5.18393457e-01 -6.96732283e-01 4.88847315e-01
-1.26872241e-01 -1.92321524e-01 -8.23567510e-01 -3.50622386e-01
-2.56025493e-01 5.00989854e-01 -6.30190790e-01 -3.95945132e-01
-1.63745821e-01 -1.05922949e+00 5.37819505e-01 7.99908102e-01
6.59603953e-01 -1.15447462e+00 -1.13553250e+00 5.70353210e-01
4.60810542e-01 -8.30733836e-01 2.46560067e-01 6.52624071e-01
-8.69775534e-01 -9.73420501e-01 -8.19588244e-01 -9.22159910e-01
5.11226356e-01 -4.15363431e-01 4.83085543e-01 1.31556382e-02
-3.72502416e-01 -4.43921685e-01 -4.22801495e-01 -5.58930993e-01
3.63658875e-01 5.14777787e-02 4.48256999e-01 9.80388150e-02
1.03497636e+00 -9.16603327e-01 -8.73506069e-01 1.38224080e-01
-2.82217085e-01 1.39852837e-01 7.45114148e-01 1.01023018e+00
2.73225337e-01 7.80858025e-02 1.09863639e+00 -2.91157603e-01
8.83400738e-01 -6.17448986e-01 -6.38991594e-01 2.06282660e-01
-7.99864411e-01 -3.13018918e-01 8.86698008e-01 -5.96274912e-01
-3.35599214e-01 2.58640796e-01 -7.51702785e-02 -2.60973185e-01
-3.41676265e-01 3.62460971e-01 2.21038818e-01 -3.85198951e-01
6.42706215e-01 6.80524111e-01 -3.87467325e-01 -3.25327843e-01
-6.73331857e-01 1.33958149e+00 5.37361324e-01 2.42375329e-01
1.29817232e-01 -8.41121525e-02 -1.67923555e-01 -6.39712691e-01
-1.64726198e-01 -4.62991357e-01 -4.78149086e-01 3.48419435e-02
7.40985513e-01 -1.09423935e+00 -7.56097496e-01 6.80255055e-01
-1.09578681e+00 -6.47775084e-02 4.13244277e-01 1.02993822e+00
-4.06720072e-01 -1.41086206e-01 -5.24747550e-01 -9.95114326e-01
-9.74494696e-01 -1.38501108e+00 5.85967243e-01 3.23791116e-01
-2.26381227e-01 -5.50333738e-01 -3.46478641e-01 -3.43049169e-01
6.58225954e-01 4.33391869e-01 5.75434208e-01 -1.38363063e+00
-1.49257839e-01 -8.98206770e-01 -3.98254961e-01 2.24575862e-01
2.94491231e-01 -4.95230377e-01 -1.17185414e+00 -3.75970334e-01
1.06251672e-01 -6.96895793e-02 6.49770200e-01 6.59375370e-01
1.40440750e+00 -2.27916121e-01 -4.71101910e-01 8.11647952e-01
1.51715982e+00 1.07788491e+00 6.93824470e-01 3.50814372e-01
1.57363385e-01 -5.42940944e-03 1.28901437e-01 7.87783623e-01
-2.56514340e-03 2.54603922e-01 1.35209948e-01 1.11179307e-01
3.75769973e-01 2.68530160e-01 -1.08170686e-02 5.61570525e-01
-6.32992834e-02 1.02591373e-01 -1.23129153e+00 5.10941625e-01
-1.60786557e+00 -5.94749928e-01 -4.99916002e-02 2.25517678e+00
7.43908644e-01 3.09383124e-01 -1.14417359e-01 6.58752859e-01
5.75132310e-01 -5.47055244e-01 -5.74345171e-01 -2.08630085e-01
1.92429032e-02 6.35816395e-01 5.50041437e-01 1.63328469e-01
-1.12988567e+00 6.21644258e-01 5.11259890e+00 7.54619002e-01
-1.68127656e+00 1.92118883e-01 7.51452863e-01 -1.04731463e-01
6.06976748e-01 -3.81129563e-01 -5.11512935e-01 8.29647064e-01
1.15732849e+00 -3.14136147e-01 4.73543823e-01 7.85571873e-01
1.69666395e-01 -1.20925382e-01 -1.05484569e+00 1.54867351e+00
-2.65279468e-02 -1.32464957e+00 -4.62535262e-01 -6.84569925e-02
3.59706998e-01 1.89923182e-01 -1.31208837e-01 2.63457596e-01
-5.51139295e-01 -1.45510912e+00 2.14240640e-01 6.72308743e-01
1.02595770e+00 -1.11525738e+00 1.18649793e+00 4.10700083e-01
-1.08263135e+00 -3.83691192e-01 -1.92298383e-01 -5.48118174e-01
-2.94638664e-01 4.64024305e-01 -1.02697015e+00 1.13565661e-01
8.98842275e-01 6.96074009e-01 -4.39382702e-01 1.54751325e+00
-1.42852306e-01 6.95781112e-01 -2.42263347e-01 -3.76505762e-01
3.02138746e-01 2.61716574e-01 1.15669332e-01 1.23504329e+00
5.19972324e-01 6.36782110e-01 -1.76029697e-01 3.89888614e-01
1.69248700e-01 2.82640606e-01 -4.20825154e-01 9.84849185e-02
3.80298764e-01 1.10698843e+00 -7.47767806e-01 -1.26995146e-01
-3.80956292e-01 8.66874516e-01 1.17682293e-01 2.61939675e-01
-5.96136928e-01 -1.18355834e+00 2.94377923e-01 1.23777334e-02
-1.10969752e-01 1.35774970e-01 -7.19660044e-01 -1.07704043e+00
6.59960657e-02 -4.62666661e-01 5.54985460e-03 -7.44861126e-01
-9.62388575e-01 1.16845906e+00 -4.06911999e-01 -1.44353664e+00
-1.35307282e-01 -6.78371668e-01 -9.39013481e-01 1.13401556e+00
-1.34834445e+00 -6.66453779e-01 -3.94151479e-01 7.86772549e-01
4.98829275e-01 -6.07185841e-01 9.94065940e-01 1.88290700e-01
-5.95794320e-01 7.52549529e-01 1.52993158e-01 2.70209998e-01
3.55263621e-01 -8.52837086e-01 4.68620956e-02 7.01974213e-01
-4.41704094e-01 3.87277693e-01 4.36967015e-01 -4.20613885e-01
-9.39533710e-01 -8.82054150e-01 9.52620864e-01 1.84750229e-01
6.25138283e-01 -1.51231125e-01 -9.98017013e-01 5.15888989e-01
1.66206747e-01 2.50547767e-01 8.62166166e-01 -2.44351551e-01
2.50372261e-01 -2.46020749e-01 -1.34165537e+00 1.26016289e-01
2.60309011e-01 -3.61067444e-01 -6.17065668e-01 3.25753570e-01
9.75737814e-03 -3.78709584e-01 -1.06198609e+00 6.77339792e-01
7.35066175e-01 -8.41914535e-01 4.48720723e-01 -5.52193761e-01
1.56711847e-01 1.06022015e-01 4.15308326e-02 -1.40665174e+00
-3.04940313e-01 -6.48371220e-01 -4.77307215e-02 4.08665538e-01
5.61380446e-01 -8.70271742e-01 6.75148606e-01 6.27961695e-01
-2.93177843e-01 -1.44356954e+00 -1.06487000e+00 -5.77358544e-01
-4.52940874e-02 -4.10741955e-01 7.34841168e-01 9.11335349e-01
5.38712382e-01 4.28359248e-02 -3.50593895e-01 2.06526279e-01
2.67540872e-01 -2.40099475e-01 1.46316335e-01 -1.38273180e+00
1.44564584e-01 -2.64676601e-01 -1.05354762e+00 -2.80196816e-01
-1.23678192e-01 -7.40248203e-01 -1.96382567e-01 -1.30245602e+00
8.69852975e-02 -4.65181947e-01 -1.14529371e+00 8.29940617e-01
1.15976654e-01 2.34707400e-01 -3.80660444e-01 1.28560290e-01
1.84006803e-02 3.78326744e-01 7.06796825e-01 1.24002583e-01
-5.09417593e-01 2.36086264e-01 -3.32342297e-01 7.77491510e-01
1.46937037e+00 -5.06803036e-01 -4.41455036e-01 -2.79391885e-01
-2.78660625e-01 4.06349629e-01 2.41503924e-01 -1.49027419e+00
5.79689324e-01 1.23602092e-01 1.20136273e+00 -4.67345357e-01
2.94199288e-01 -5.20496964e-01 5.26710078e-02 7.46605635e-01
-1.25054508e-01 9.89344120e-02 4.16861892e-01 2.58280754e-01
-2.40575179e-01 4.74177226e-02 7.54082978e-01 1.41633958e-01
-7.36053288e-01 5.35626888e-01 -4.04614002e-01 -3.98647159e-01
1.37884521e+00 -6.51708484e-01 6.93267509e-02 -4.93036062e-02
-8.63334954e-01 -7.16812611e-02 -1.81473374e-01 3.48213077e-01
1.02599657e+00 -1.19397104e+00 -7.02297628e-01 7.48138428e-01
3.19339335e-01 -4.84467834e-01 1.76374763e-01 1.24814761e+00
-5.68658531e-01 5.99376738e-01 -7.59498894e-01 -4.26128238e-01
-1.17642641e+00 2.75725275e-02 5.80363452e-01 2.38747001e-01
-7.64705300e-01 1.07174826e+00 -6.84746134e-04 3.07353020e-01
4.06906962e-01 -4.65583652e-01 -3.70567977e-01 -2.13098243e-01
8.77298117e-01 1.59890756e-01 5.59985220e-01 -3.89143020e-01
-8.08549702e-01 2.89234549e-01 -1.74891159e-01 2.42249832e-01
1.53687835e+00 5.65981805e-01 -5.37561476e-02 2.60743424e-02
1.08165586e+00 -6.63684011e-01 -1.01591253e+00 1.81836829e-01
-3.45602781e-02 -2.59212077e-01 5.78542590e-01 -1.24970067e+00
-1.00923979e+00 1.19381607e+00 1.26430011e+00 1.34177189e-02
1.48633850e+00 -4.47010189e-01 8.00960243e-01 3.42287838e-01
4.70377803e-01 -8.18018973e-01 -5.62805712e-01 1.34071738e-01
8.30584705e-01 -9.84441638e-01 -1.73230812e-01 1.57195792e-01
-4.46024418e-01 1.46547854e+00 6.25125110e-01 -4.67080086e-01
1.09987998e+00 4.99436021e-01 -1.79001078e-01 1.15419496e-02
-7.72933602e-01 3.06606978e-01 3.14254582e-01 6.19932830e-01
5.68471968e-01 4.40985769e-01 -5.83723247e-01 1.37457514e+00
-3.91378552e-01 6.26309395e-01 3.50125194e-01 9.84724104e-01
-5.62461615e-01 -4.47018653e-01 3.32178809e-02 1.10185492e+00
-7.49803603e-01 -3.64477575e-01 -5.82590364e-02 6.26753271e-01
1.79205060e-01 9.34838176e-01 2.44694710e-01 -8.94216120e-01
1.18253663e-01 1.63132638e-01 3.21229875e-01 -3.15475345e-01
-6.80426061e-01 8.41641575e-02 -3.99222404e-01 -2.72806793e-01
9.16040689e-02 -2.67776221e-01 -1.33481622e+00 1.79010957e-01
-5.76886475e-01 4.05647427e-01 7.34095633e-01 9.75107014e-01
4.26649719e-01 5.04887521e-01 6.12595558e-01 -5.71553349e-01
-2.87349105e-01 -1.47482967e+00 -9.54048574e-01 -1.63419038e-01
6.09946907e-01 -7.32150495e-01 -2.81080008e-01 -5.96075803e-02] | [13.226664543151855, 3.5194737911224365] |
1d3bee02-6312-4e8a-a94a-d6b7d4b9a2c3 | scene-text-detection-with-scribble-lines | 2012.05030 | null | https://arxiv.org/abs/2012.05030v2 | https://arxiv.org/pdf/2012.05030v2.pdf | Scene Text Detection with Scribble Lines | Scene text detection, which is one of the most popular topics in both academia and industry, can achieve remarkable performance with sufficient training data. However, the annotation costs of scene text detection are huge with traditional labeling methods due to the various shapes of texts. Thus, it is practical and insightful to study simpler labeling methods without harming the detection performance. In this paper, we propose to annotate the texts by scribble lines instead of polygons for text detection. It is a general labeling method for texts with various shapes and requires low labeling costs. Furthermore, a weakly-supervised scene text detection framework is proposed to use the scribble lines for text detection. The experiments on several benchmarks show that the proposed method bridges the performance gap between the weakly labeling method and the original polygon-based labeling methods, with even better performance. We will release the weak annotations of the benchmarks in our experiments and hope it will benefit the field of scene text detection to achieve better performance with simpler annotations. | ['Xiang Bai', 'Xiaolin Wei', 'Rui Zhang', 'Minghui Liao', 'Yang Qiu', 'Wenqing Zhang'] | 2020-12-09 | null | null | null | null | ['scene-text-detection'] | ['computer-vision'] | [ 2.79902041e-01 -1.74791262e-01 -1.32393077e-01 -1.74432039e-01
-4.08454090e-01 -6.34645343e-01 6.12242758e-01 3.43271524e-01
-2.53260732e-01 2.11429477e-01 -2.64032912e-02 -3.73810709e-01
5.16957700e-01 -7.56283998e-01 -3.31698149e-01 -6.58791959e-01
5.00510514e-01 3.85782093e-01 8.89240921e-01 -1.15629323e-01
4.05064523e-01 2.20182136e-01 -1.40846050e+00 4.64782983e-01
7.00052679e-01 6.43922567e-01 3.81981820e-01 4.58481222e-01
-5.98286092e-01 8.90829444e-01 -5.17549634e-01 -4.17670161e-01
2.58636713e-01 -3.91848475e-01 -7.91514039e-01 5.56607723e-01
6.25091910e-01 -3.36447328e-01 -4.05527353e-01 1.24151540e+00
3.23109627e-01 -1.09423019e-01 6.08985722e-01 -1.05886269e+00
-2.37804607e-01 8.41251850e-01 -1.15032327e+00 -1.96922615e-01
4.31858271e-01 -2.27843374e-01 9.51097369e-01 -1.13102949e+00
3.46956491e-01 1.24781334e+00 7.92639673e-01 1.98560029e-01
-7.15415478e-01 -4.51618433e-01 4.08408731e-01 -3.24061769e-03
-1.60537934e+00 -1.86157495e-01 7.03476191e-01 -4.63715822e-01
5.44974029e-01 6.23546422e-01 3.56893837e-01 7.49236822e-01
-2.35510051e-01 1.35415757e+00 9.43251848e-01 -7.41052508e-01
-4.88104597e-02 4.08634752e-01 2.83057064e-01 1.17397487e+00
4.98380840e-01 -5.34652233e-01 -2.26095915e-01 -1.10518202e-01
7.15094686e-01 2.38450840e-01 -2.28763670e-01 -3.87176067e-01
-1.49105728e+00 7.33511925e-01 1.63367569e-01 4.66345012e-01
3.97293627e-01 8.73486623e-02 6.29526734e-01 -1.73493519e-01
6.48545742e-01 2.21055031e-01 -4.60494421e-02 -5.38146542e-03
-1.20796585e+00 -2.53690183e-02 8.21668565e-01 1.28129411e+00
6.56268835e-01 1.37253301e-02 -3.60449255e-01 1.04234445e+00
2.87633866e-01 5.73497474e-01 2.37596601e-01 -4.19267505e-01
6.15898013e-01 9.90040839e-01 1.15722179e-01 -1.23924780e+00
-5.55810034e-01 -1.28212467e-01 -8.63363743e-01 -9.45262611e-02
6.57659888e-01 -1.64681017e-01 -8.30744624e-01 8.52420986e-01
4.22904819e-01 -3.41301747e-02 -3.20377916e-01 8.29272389e-01
9.91789699e-01 7.05493569e-01 -2.62070060e-01 -2.67280415e-02
1.62570786e+00 -1.51315141e+00 -9.24190342e-01 -1.94221452e-01
1.29476464e+00 -1.19661367e+00 1.46548319e+00 2.99590886e-01
-7.35618234e-01 -4.84244645e-01 -9.59558010e-01 -2.12505862e-01
-4.55913603e-01 7.25445151e-01 6.48325682e-01 8.77661288e-01
-6.98678076e-01 1.99209854e-01 -8.08203101e-01 -9.48054790e-01
4.28134948e-01 9.25852731e-02 1.97138190e-01 -8.05494562e-03
-6.15589678e-01 4.35613126e-01 3.17491889e-01 -5.31523190e-02
-4.91776496e-01 -1.93337068e-01 -7.63924897e-01 -5.64296581e-02
6.25121593e-01 -3.52351695e-01 1.19641757e+00 -7.67771959e-01
-1.30959153e+00 1.03430355e+00 -9.25547257e-02 -1.06603429e-01
9.76836979e-01 -2.48569354e-01 3.45374979e-02 1.00177169e-01
2.25294530e-01 4.96084154e-01 8.38154137e-01 -1.21310842e+00
-9.25342321e-01 -1.08201183e-01 4.77731638e-02 3.55878204e-01
-7.65014350e-01 1.40867710e-01 -9.31123495e-01 -7.93090343e-01
3.17565531e-01 -9.52155352e-01 -2.83630520e-01 1.06410369e-01
-8.95311058e-01 -5.40793300e-01 1.47472227e+00 -1.25680044e-01
1.19064295e+00 -2.16678619e+00 -4.58746970e-01 -1.19987927e-01
3.07261825e-01 1.97405338e-01 9.76595506e-02 5.32485664e-01
2.55588472e-01 2.48105988e-01 -1.80941179e-01 -5.75448573e-01
2.41183955e-02 -3.77668515e-02 -4.76540118e-01 7.53645837e-01
-2.14628339e-01 7.59642005e-01 -8.58533561e-01 -9.80736375e-01
6.00350201e-01 6.50411472e-02 -1.63110584e-01 -2.88157389e-02
-3.23922783e-01 -4.05783467e-02 -7.18819916e-01 8.01580369e-01
7.36491919e-01 -3.95486683e-01 5.19976914e-02 -3.40084910e-01
-3.13708812e-01 -6.71149045e-02 -1.20886087e+00 1.40988195e+00
-2.94495057e-02 9.10250187e-01 2.99936850e-02 -1.02086866e+00
9.81622458e-01 3.44365388e-02 4.45679665e-01 -2.49273941e-01
3.07853371e-01 1.47660403e-02 -3.75506133e-01 -6.11028552e-01
7.25466371e-01 2.07842767e-01 -1.46787256e-01 5.06940544e-01
-6.85669482e-01 -5.24477839e-01 4.70831901e-01 3.26097965e-01
9.16578293e-01 1.10942230e-01 2.11323529e-01 -4.40480381e-01
5.53933740e-01 3.34285438e-01 2.67865926e-01 9.01742160e-01
-1.58769801e-01 8.04549992e-01 4.11446154e-01 -5.78387499e-01
-1.01717889e+00 -5.05508482e-01 -3.99367601e-01 1.26658750e+00
5.17726898e-01 -8.24044168e-01 -9.03277457e-01 -8.65192354e-01
-2.83215582e-01 3.03650916e-01 -3.77331197e-01 3.58785391e-01
-6.28907382e-01 -9.84496236e-01 7.84682989e-01 5.79880238e-01
8.05421889e-01 -6.99738681e-01 -3.97860229e-01 -2.32360452e-01
-3.06572258e-01 -1.59257197e+00 -6.10180140e-01 9.34377126e-03
-7.73261726e-01 -9.52798724e-01 -7.94022739e-01 -1.39921641e+00
9.97726977e-01 8.46133947e-01 7.61393309e-01 4.21702772e-01
-3.04544687e-01 3.38494599e-01 -8.31601918e-01 -4.73724514e-01
-2.14167655e-01 4.50610332e-02 -1.94375381e-01 7.85811432e-03
2.33284980e-01 1.83278069e-01 -3.74059469e-01 6.05048120e-01
-9.09488678e-01 4.83883202e-01 2.26340130e-01 6.22408688e-01
4.74934638e-01 4.60334748e-01 -1.51016250e-01 -1.30225432e+00
3.49843204e-01 1.17576443e-01 -6.39213383e-01 2.57722408e-01
-3.34452301e-01 -2.72562087e-01 6.16202474e-01 -3.72368902e-01
-1.07545817e+00 4.95038301e-01 1.07530892e-01 -8.29371810e-02
-2.95883983e-01 2.25257516e-01 -2.52870582e-02 -1.87460169e-01
7.32600689e-01 2.60064393e-01 -7.44528174e-01 -4.44094449e-01
2.65255958e-01 8.44847560e-01 2.13380724e-01 -5.61044037e-01
8.10072601e-01 9.59292948e-01 -9.90872607e-02 -1.23764348e+00
-1.07872462e+00 -9.23454762e-01 -7.33716428e-01 -1.18257821e-01
7.62779534e-01 -8.88353229e-01 -3.13994646e-01 5.40277719e-01
-1.20640147e+00 -4.82618719e-01 -9.62559283e-02 2.40270734e-01
-2.40282401e-01 1.25368297e+00 -7.98210323e-01 -8.96755278e-01
-2.86822230e-01 -1.06269634e+00 1.57467985e+00 -3.29520553e-02
1.06166899e-01 -9.28898633e-01 -1.70976534e-01 3.73078912e-01
-4.72939312e-02 6.30233288e-02 7.31611550e-01 -5.73892355e-01
-5.06606281e-01 -4.76732939e-01 -6.48913503e-01 -7.65194967e-02
1.47479132e-01 1.86716035e-01 -1.12963378e+00 -2.10337967e-01
3.12367920e-02 -1.60360336e-01 9.38506305e-01 2.07398385e-01
1.43186915e+00 -7.17295334e-02 -6.57796800e-01 4.62668180e-01
1.35809827e+00 -2.88057025e-03 5.34067035e-01 4.68084395e-01
1.29193580e+00 6.48933291e-01 7.95525491e-01 5.03619313e-01
2.90035427e-01 5.65625072e-01 2.77697295e-01 -5.52398860e-01
-2.02498779e-01 -2.78966010e-01 3.72906178e-01 8.64757121e-01
1.76805213e-01 -5.75422943e-01 -1.07006347e+00 2.99921006e-01
-2.18966508e+00 -6.06115997e-01 -1.02133441e+00 1.93388760e+00
5.96156657e-01 2.25415736e-01 1.53481856e-01 3.91650915e-01
1.09845114e+00 1.86545640e-01 -1.92963272e-01 -8.10432434e-02
-2.63477653e-01 -2.69638777e-01 5.96542954e-01 2.65524775e-01
-1.53892291e+00 1.35884023e+00 6.63141584e+00 1.13804793e+00
-9.34110999e-01 2.29388382e-02 5.71496308e-01 3.42102587e-01
2.16192693e-01 2.52873693e-02 -1.14018595e+00 2.97883332e-01
1.43019482e-01 1.06105983e-01 -1.08558491e-01 1.05426836e+00
3.95233840e-01 -1.91396475e-01 -1.07391727e+00 1.22095859e+00
3.24420184e-01 -1.16885209e+00 1.48679376e-01 -3.35272521e-01
9.41912949e-01 -8.05216655e-02 -2.80248404e-01 1.76178828e-01
1.73328042e-01 -7.21112430e-01 7.21133411e-01 -1.40539125e-01
5.46591222e-01 -3.33192706e-01 7.61728287e-01 5.49656987e-01
-1.54290891e+00 2.38225594e-01 -7.62109637e-01 -1.39302358e-01
-6.92131277e-03 8.22430015e-01 -9.86954749e-01 2.50489950e-01
6.21015012e-01 8.84740949e-01 -8.02833140e-01 1.17319810e+00
-2.44706273e-01 7.33168304e-01 -3.82968932e-01 -5.03771901e-01
2.01925516e-01 -2.80392319e-01 3.45145106e-01 1.69085479e+00
7.86699280e-02 -2.29349121e-01 8.98313284e-01 6.33466482e-01
-1.32538125e-01 6.77976966e-01 -7.76275694e-01 1.42744202e-02
1.87862754e-01 1.58209622e+00 -1.63449025e+00 -5.46180367e-01
-6.20974541e-01 1.06079066e+00 -8.24201033e-02 2.07762465e-01
-8.82261217e-01 -5.72711766e-01 -3.95648122e-01 2.78955072e-01
9.21785906e-02 -3.89985502e-01 -6.72379255e-01 -1.24937475e+00
2.16752831e-02 -5.89499235e-01 3.52762371e-01 -7.69064486e-01
-1.13861179e+00 4.15071487e-01 -1.54993981e-01 -1.34794486e+00
4.86254066e-01 -7.66566277e-01 -6.93089843e-01 1.74632177e-01
-1.37000144e+00 -1.39466524e+00 -6.39397919e-01 5.44583142e-01
1.07152522e+00 3.18954200e-01 4.05036271e-01 2.09114298e-01
-8.83985281e-01 6.01864934e-01 1.76186293e-01 5.88425517e-01
7.70925224e-01 -1.26142704e+00 3.66737455e-01 9.22580481e-01
4.03521657e-01 1.55129626e-01 6.33572698e-01 -6.81020379e-01
-1.33736658e+00 -1.27083731e+00 5.78445494e-01 -5.10666788e-01
7.69223630e-01 -7.30944395e-01 -7.86159515e-01 6.01759493e-01
3.09435815e-01 -1.27106145e-01 3.51240069e-01 -6.69119731e-02
-2.72371441e-01 1.29840940e-01 -8.17952514e-01 8.90009582e-01
9.23701465e-01 -3.22061896e-01 -4.31378037e-01 1.10272014e+00
6.71137512e-01 -4.46380854e-01 -3.33156735e-01 4.08127636e-01
2.40077242e-01 -8.37590039e-01 6.83729172e-01 1.58192307e-01
8.05436522e-02 -6.09916031e-01 -1.10061783e-02 -5.59115052e-01
-2.00880896e-02 -4.70202357e-01 2.04343811e-01 1.39697385e+00
3.58492643e-01 -4.20833826e-01 1.18833530e+00 1.94605082e-01
-2.14136541e-01 -3.92391831e-01 -3.78365904e-01 -7.69174635e-01
-6.49543628e-02 -4.32553023e-01 8.27645361e-02 1.25615907e+00
3.01313818e-01 6.86378956e-01 -3.14396590e-01 9.89414752e-02
5.12682974e-01 4.82883930e-01 1.01600194e+00 -1.12543154e+00
-8.11648928e-03 -5.25458634e-01 -2.10362539e-01 -1.64202845e+00
-1.40921563e-01 -8.38244736e-01 3.75812948e-01 -1.64412808e+00
5.76205850e-01 -4.91459072e-01 3.64274710e-01 5.76729715e-01
-2.16011733e-01 4.95168090e-01 1.58864826e-01 3.71939093e-01
-1.12720764e+00 3.76040459e-01 1.41709983e+00 -3.76061201e-01
-1.57895431e-01 -1.28766358e-01 -4.12108690e-01 1.30366433e+00
8.28121424e-01 -4.55189764e-01 -2.85987347e-01 -4.65678871e-01
2.99470454e-01 -4.47093934e-01 2.20483467e-02 -8.30962300e-01
5.45302570e-01 -6.76573962e-02 2.37828150e-01 -1.07860458e+00
5.54586351e-02 -7.68457592e-01 -5.99972665e-01 4.29390758e-01
-1.44640997e-01 -1.76999465e-01 1.88196395e-02 5.17510891e-01
2.47530695e-02 -4.81742680e-01 8.15892041e-01 -7.57891014e-02
-5.77260017e-01 1.53723389e-01 -5.50562620e-01 2.38651671e-02
1.07780981e+00 -3.97807896e-01 -3.64813179e-01 -2.74489284e-01
-2.94213384e-01 1.88894793e-01 7.20068991e-01 2.31411979e-01
4.91733104e-01 -9.59360421e-01 -6.54377222e-01 -9.35791731e-02
2.96228200e-01 4.13293958e-01 -2.28378549e-01 9.49939668e-01
-9.34763670e-01 4.48701829e-01 4.59466249e-01 -1.00543702e+00
-1.48621845e+00 6.34292126e-01 1.08705588e-01 -2.08292902e-01
-9.52032328e-01 6.30791664e-01 7.37302721e-01 -3.46318185e-01
5.52521825e-01 -5.16876578e-01 -2.03179836e-01 -1.44856483e-01
4.93557215e-01 5.01564264e-01 -1.64494421e-02 -4.82728362e-01
-1.44552454e-01 8.56157005e-01 -1.06046066e-01 3.81198823e-01
8.39669466e-01 -2.76945323e-01 -1.81779012e-01 4.40197200e-01
7.96062827e-01 4.25852776e-01 -8.54629457e-01 -2.10392848e-01
1.71894282e-01 -4.40406382e-01 3.01131271e-02 -3.86458814e-01
-9.28645432e-01 1.00168312e+00 4.78345007e-01 7.03271210e-01
8.99670720e-01 1.00640319e-01 7.38694847e-01 7.18772411e-01
2.96861321e-01 -1.14915466e+00 4.00636673e-01 4.72928673e-01
4.38195795e-01 -1.27614439e+00 3.21808428e-01 -1.12530148e+00
-4.51855212e-01 1.32808876e+00 7.49251485e-01 -1.04894176e-01
3.55584830e-01 6.01026595e-01 -1.01319507e-04 -3.41060191e-01
-1.87284231e-01 -4.96697038e-01 1.30354717e-01 3.75845909e-01
6.52315080e-01 4.69121970e-02 -4.72438574e-01 2.19406411e-01
9.81093049e-02 -5.66636503e-01 6.91705585e-01 9.58432019e-01
-8.24450672e-01 -1.19906020e+00 -6.53529346e-01 5.12062669e-01
-4.67283577e-01 -2.47225165e-01 -8.68739009e-01 8.89964521e-01
1.07658273e-02 1.16880083e+00 -9.53008160e-02 -2.39415735e-01
3.21369618e-01 -2.23249272e-01 3.50172371e-01 -9.29271579e-01
-3.22012365e-01 5.58475673e-01 1.11358017e-01 -1.63849965e-01
-5.91099560e-01 -5.20560741e-01 -1.56261671e+00 -2.81458080e-01
-8.35694134e-01 8.73911660e-03 4.28720146e-01 9.13908660e-01
-6.12457395e-02 1.93219364e-01 7.32409120e-01 -7.05659926e-01
-1.85228556e-01 -8.81195247e-01 -6.27852917e-01 3.51990730e-01
-1.29942924e-01 -4.32970911e-01 -3.87269437e-01 3.22969377e-01] | [12.029952049255371, 2.3013153076171875] |
5e676eae-7869-4071-a9b0-0048ebae8daf | deep-bag-of-sub-emotions-for-depression | 2103.01334 | null | https://arxiv.org/abs/2103.01334v1 | https://arxiv.org/pdf/2103.01334v1.pdf | Deep Bag-of-Sub-Emotions for Depression Detection in Social Media | This paper presents the Deep Bag-of-Sub-Emotions (DeepBoSE), a novel deep learning model for depression detection in social media. The model is formulated such that it internally computes a differentiable Bag-of-Features (BoF) representation that incorporates emotional information. This is achieved by a reinterpretation of classical weighting schemes like term frequency-inverse document frequency into probabilistic deep learning operations. An important advantage of the proposed method is that it can be trained under the transfer learning paradigm, which is useful to enhance conventional BoF models that cannot be directly integrated into deep learning architectures. Experiments were performed in the eRisk17 and eRisk18 datasets for the depression detection task; results show that DeepBoSE outperforms conventional BoF representations and it is competitive with the state of the art, achieving a F1-score over the positive class of 0.64 in eRisk17 and 0.65 in eRisk18. | ['Manuel Montes-y-Gomez', 'Fabio A. Gonzalez', 'Mario Ezra Aragon', 'Juan S. Lara'] | 2021-03-01 | null | null | null | null | ['probabilistic-deep-learning'] | ['computer-vision'] | [-3.54278415e-01 3.49053890e-01 -1.09509706e-01 -7.03898847e-01
-4.64149326e-01 1.21852033e-01 7.62122929e-01 7.43063033e-01
-5.58231652e-01 5.49235642e-01 3.37093621e-01 1.91681728e-01
-3.27095538e-01 -9.11085963e-01 -2.41816789e-01 -7.20677435e-01
-3.50568146e-01 4.36266214e-01 -3.55307251e-01 -5.38684845e-01
-6.54354021e-02 4.97642875e-01 -1.57422423e+00 4.27605152e-01
5.78790665e-01 1.25558758e+00 -3.77289027e-01 4.57756609e-01
-4.07982394e-02 8.00238490e-01 -7.00461805e-01 -8.11588824e-01
-4.82345968e-01 4.01215889e-02 -5.68822742e-01 -2.51264840e-01
2.31121387e-02 -5.92457235e-01 -2.82939225e-01 7.86414385e-01
7.27257013e-01 1.73397198e-01 1.06263101e+00 -1.29406691e+00
-7.99424648e-01 1.77789390e-01 -3.93053502e-01 9.00103003e-02
3.72645795e-01 -7.96755910e-01 1.11146533e+00 -1.23134923e+00
2.48558223e-01 1.59970474e+00 9.40868080e-01 5.32907963e-01
-1.03355074e+00 -4.76781756e-01 9.09293741e-02 1.56135634e-01
-1.17449367e+00 -9.16291475e-02 4.64220464e-01 -4.82638538e-01
1.62426996e+00 -1.11490853e-01 8.36400747e-01 1.27412784e+00
5.46506882e-01 1.03752661e+00 7.32543766e-01 -5.96559703e-01
2.62770474e-01 2.52120197e-01 4.76842552e-01 8.67107689e-01
2.68561840e-01 5.63256862e-03 -7.84779429e-01 -5.22609413e-01
2.37295896e-01 3.58965218e-01 1.94418982e-01 -1.74421012e-01
-5.61936498e-01 1.45052350e+00 5.06385148e-01 4.09357458e-01
-6.17140830e-01 2.73212075e-01 8.33194077e-01 1.98346272e-01
1.15446150e+00 2.67335232e-02 -3.79842967e-01 -1.59976050e-01
-7.18343616e-01 5.29375196e-01 5.33274174e-01 1.87564850e-01
4.33718145e-01 5.65707721e-02 -3.96118879e-01 1.28976405e+00
3.94251764e-01 4.20368582e-01 7.94121325e-01 -5.39498448e-01
-9.38394219e-02 7.97282636e-01 8.84756371e-02 -1.29963374e+00
-6.86443686e-01 -3.71423513e-01 -1.02319062e+00 2.62749612e-01
-3.37647140e-01 -1.67410627e-01 -7.75210440e-01 1.64297843e+00
9.73085389e-02 -1.65413499e-01 1.56011745e-01 4.27618533e-01
1.20206666e+00 7.14791417e-01 2.18014091e-01 9.16529223e-02
1.44575679e+00 -8.50462973e-01 -1.17394018e+00 -3.76024693e-01
8.00753951e-01 -3.59396189e-01 6.48050129e-01 7.76883483e-01
-9.66549456e-01 -4.07910079e-01 -1.11231792e+00 -2.77727187e-01
-9.15307760e-01 3.20568353e-01 1.00632191e+00 8.53501022e-01
-1.41506541e+00 6.48465395e-01 -7.43017256e-01 -4.46523070e-01
6.01704299e-01 4.38474238e-01 -5.54675102e-01 1.20393177e-02
-1.58669078e+00 1.04582345e+00 3.60067040e-01 -3.68996486e-02
-6.99963033e-01 -4.62526709e-01 -1.07769096e+00 4.47271317e-01
-4.42364573e-01 -5.83074927e-01 1.22134459e+00 -1.03035212e+00
-1.44907463e+00 1.26841223e+00 1.22935860e-03 -5.72667122e-01
2.25876093e-01 -6.59316957e-01 -5.35680532e-01 3.90136480e-01
-1.31874219e-01 7.81471431e-01 9.02120829e-01 -7.14558482e-01
-4.85667497e-01 -5.06800890e-01 -4.81893905e-02 3.27470563e-02
-8.63183916e-01 1.97738588e-01 -2.77803950e-02 -8.17905307e-01
-3.94840926e-01 -7.00918972e-01 -1.35202203e-02 5.53864613e-02
2.36306936e-02 -7.76608288e-01 5.75085044e-01 -6.29018247e-01
1.18173409e+00 -2.27402782e+00 -3.99438143e-02 1.98199209e-02
3.86316329e-01 6.01627171e-01 -1.18541420e-01 7.15945601e-01
-4.95107263e-01 -3.73990647e-02 1.74506500e-01 -7.69232988e-01
3.39374781e-01 2.05024213e-01 -1.81264743e-01 5.28804064e-01
6.49238408e-01 7.31690884e-01 -9.25185919e-01 -1.03878014e-01
2.33012646e-01 9.45356011e-01 -5.23091555e-01 3.18973631e-01
2.88968205e-01 -7.29521275e-01 -1.89369127e-01 3.76007468e-01
8.67781222e-01 1.82610720e-01 3.07404567e-02 1.08016931e-01
3.20472300e-01 2.23335609e-01 -6.80082798e-01 1.28264129e+00
-4.60325062e-01 5.04892111e-01 -1.19978793e-01 -1.41185045e+00
1.34564888e+00 4.16563690e-01 6.22318566e-01 -5.69658935e-01
2.50296414e-01 -9.71421674e-02 -4.52608526e-01 -3.19249511e-01
4.39080089e-01 -4.28222984e-01 -3.23449314e-01 1.93858251e-01
5.79795599e-01 1.84025720e-01 -5.01639470e-02 2.29049399e-01
9.99776244e-01 -2.13964045e-01 5.12361884e-01 -2.22134396e-01
3.86547118e-01 -7.66732633e-01 4.80402052e-01 5.98282993e-01
-2.93488860e-01 3.09272975e-01 8.11957896e-01 -4.99638140e-01
-2.94860721e-01 -1.13528919e+00 -3.00361931e-01 1.41889203e+00
-5.89535117e-01 -8.19107115e-01 -6.07750893e-01 -8.10200393e-01
2.10535064e-01 7.44179666e-01 -1.11271358e+00 -8.00412774e-01
2.77536750e-01 -1.18701446e+00 5.42925477e-01 7.89028943e-01
3.30788910e-01 -1.18879998e+00 -5.96408546e-01 2.04827502e-01
-7.65428022e-02 -6.76333070e-01 3.54425728e-01 6.21028602e-01
-7.47263193e-01 -6.39863968e-01 -9.00051057e-01 -7.03568459e-01
2.01116294e-01 -1.44591197e-01 1.32419741e+00 -1.12918891e-01
-2.89636165e-01 3.84540886e-01 -5.02362728e-01 -8.61371875e-01
-1.30664319e-01 -2.34961480e-01 4.86951731e-02 8.37360099e-02
1.10078168e+00 -5.07686675e-01 -6.83093488e-01 -3.57606918e-01
-7.99288630e-01 -4.14016724e-01 3.39268446e-01 1.20257974e+00
1.98692322e-01 1.45692617e-01 1.26341951e+00 -7.50406742e-01
9.81699705e-01 -6.72089159e-01 2.71641035e-02 -1.00199834e-01
-6.71970189e-01 -2.85589546e-01 1.20725170e-01 -1.49125785e-01
-9.73471940e-01 -3.31148744e-01 -5.70759654e-01 -1.67638615e-01
-1.25399053e-01 6.73356950e-01 1.48903485e-02 3.13953876e-01
4.88370985e-01 -1.49774641e-01 -1.08843170e-01 -4.45466965e-01
2.44014412e-01 1.03476238e+00 1.76013589e-01 -1.38238832e-01
-2.88059771e-01 3.75476450e-01 -2.74632126e-01 -7.61543274e-01
-8.69540274e-01 -6.66165411e-01 -3.15733314e-01 -1.81303367e-01
8.30563128e-01 -1.17908812e+00 -7.10434914e-01 8.16242397e-01
-1.02994537e+00 4.65447158e-02 -1.14393584e-01 3.31811696e-01
-4.26128119e-01 1.96109369e-01 -9.13768589e-01 -1.12617755e+00
-8.33784223e-01 -5.88820636e-01 1.39841211e+00 5.17671090e-03
-5.37092030e-01 -1.30885243e+00 2.38364667e-01 1.21906571e-01
3.44337821e-01 4.63655293e-01 1.05984414e+00 -9.28058982e-01
9.19364214e-01 -7.78778791e-01 -3.96921128e-01 8.82968843e-01
-5.27815148e-02 6.61499724e-02 -1.38199461e+00 -2.81016052e-01
6.37281984e-02 -8.79023314e-01 1.26387668e+00 4.89221752e-01
1.15548480e+00 7.32374862e-02 -1.55761197e-01 2.56725043e-01
1.33280218e+00 1.31874710e-01 8.36208522e-01 3.46790403e-01
1.18821062e-01 6.54675424e-01 4.72558588e-01 9.37075377e-01
4.87307549e-01 5.08930922e-01 5.16378462e-01 -5.09080172e-01
2.11941168e-01 4.35539782e-02 5.44384897e-01 4.98240411e-01
2.17014596e-01 -2.69252926e-01 -8.34436536e-01 6.85226142e-01
-1.97704625e+00 -7.60859787e-01 8.61309767e-02 1.83824718e+00
5.51519752e-01 7.23684579e-02 2.30600923e-01 4.11218137e-01
5.02729774e-01 1.93856329e-01 -1.55161947e-01 -1.27975547e+00
6.22195899e-02 7.59531736e-01 -1.67306513e-01 2.88192511e-01
-1.37703228e+00 7.56117642e-01 6.31152153e+00 7.77445078e-01
-9.30366218e-01 2.12788045e-01 7.45530486e-01 -8.15282613e-02
1.59023911e-01 -9.14444268e-01 -6.13568246e-01 3.08218718e-01
1.29541397e+00 2.15524621e-03 -9.85673815e-02 1.03744721e+00
1.18797801e-01 -9.17219296e-02 -9.85301077e-01 1.24302852e+00
3.26564103e-01 -7.35508800e-01 -1.50660500e-02 2.15220693e-02
3.30173165e-01 9.57617834e-02 2.46290103e-01 1.01904058e+00
1.93151578e-01 -1.19774282e+00 3.93744379e-01 5.73748350e-01
5.08287072e-01 -1.35554051e+00 1.25402129e+00 1.17712386e-01
-3.93896878e-01 -1.37400404e-01 -6.75380170e-01 -3.02720815e-01
-2.93897033e-01 1.11623931e+00 -8.75870049e-01 4.37744588e-01
1.05859375e+00 9.93501425e-01 -4.60668713e-01 6.61186934e-01
-9.73669738e-02 4.99932468e-01 -4.60515171e-02 -2.07013175e-01
4.62652028e-01 6.68397173e-02 -1.13761880e-01 1.77073944e+00
4.92406905e-01 -3.76755148e-01 -9.11222175e-02 4.07366544e-01
-1.34205773e-01 2.90524542e-01 -6.76832199e-01 -2.89809227e-01
-1.17506348e-01 1.50044763e+00 -4.12827909e-01 -5.23320258e-01
-2.46474534e-01 1.03335845e+00 6.93564475e-01 9.69424024e-02
-7.58078635e-01 -7.65736818e-01 7.75689244e-01 -2.27029011e-01
5.28571188e-01 3.45239431e-01 2.14690194e-01 -9.85293031e-01
-3.19382399e-01 -6.93321764e-01 5.45617223e-01 -8.74376297e-01
-1.67571700e+00 6.90837741e-01 -4.01105955e-02 -7.14697421e-01
-4.05201316e-01 -1.02883852e+00 -4.67410028e-01 7.60796726e-01
-1.54934657e+00 -1.09246099e+00 -1.21048979e-01 6.06271863e-01
7.71926194e-02 -2.38025397e-01 1.62927532e+00 4.23281044e-01
-5.60320377e-01 7.31558561e-01 4.97417986e-01 -2.37319502e-03
8.18439662e-01 -1.50427473e+00 1.53307423e-01 1.45820171e-01
-1.40438110e-01 4.61867690e-01 3.12877119e-01 -2.95470625e-01
-6.01685643e-01 -1.16676557e+00 1.43830967e+00 5.73645160e-03
4.91424918e-01 -6.44710183e-01 -8.37469459e-01 3.80120814e-01
4.10146117e-01 -3.14795375e-02 1.32309568e+00 5.54612041e-01
-4.65588897e-01 -2.37525061e-01 -1.47821689e+00 1.84617534e-01
5.32353044e-01 -5.06097198e-01 -8.32395136e-01 5.94250500e-01
3.07799459e-01 9.48456228e-02 -9.56971824e-01 2.96255887e-01
5.71674705e-01 -1.23045111e+00 1.12709796e+00 -7.85705566e-01
7.52464354e-01 5.92905283e-01 -1.93396360e-01 -1.67691278e+00
-5.47823071e-01 -2.13880599e-01 -4.31112587e-01 1.24557018e+00
-1.55388996e-01 -6.82239175e-01 3.92979115e-01 1.55953839e-01
3.19253504e-02 -1.12838972e+00 -9.62908208e-01 -5.18759131e-01
2.98855931e-01 -4.16482955e-01 2.96489805e-01 9.97790039e-01
3.81166011e-01 4.71826494e-01 -3.36532712e-01 -2.91308224e-01
9.76485685e-02 -3.64865839e-01 2.22552165e-01 -1.51974165e+00
-9.61840674e-02 -4.35593188e-01 -7.45134532e-01 -2.52614677e-01
5.45145333e-01 -9.32791531e-01 -2.19076097e-01 -1.79521036e+00
1.10440142e-01 1.43215284e-01 -9.02501345e-01 8.49282026e-01
-2.17453852e-01 3.58976960e-01 -2.17241913e-01 -5.53895831e-01
-5.24820805e-01 1.06514823e+00 3.95017117e-01 -1.02517277e-01
1.26321256e-01 -3.69700454e-02 -7.97414184e-01 9.53670919e-01
8.91372204e-01 -4.97955710e-01 -3.46351057e-01 1.04138823e-02
3.76072854e-01 -2.13633865e-01 3.10309917e-01 -7.49831915e-01
-4.73081261e-01 5.27022064e-01 6.86063766e-01 -6.58739686e-01
6.88120246e-01 -4.29780513e-01 -5.15526235e-01 4.39977467e-01
-3.60659778e-01 1.18049100e-01 3.69945735e-01 5.77072442e-01
-4.14301246e-01 -4.45918798e-01 6.00661993e-01 1.75955757e-01
-4.58111346e-01 -1.02062508e-01 -9.72602785e-01 -2.69464582e-01
7.82050788e-01 2.63017803e-01 -3.97798046e-02 -5.55460334e-01
-8.21700335e-01 -4.28145491e-02 -2.66626120e-01 3.98223400e-01
8.64159703e-01 -1.58123767e+00 -7.93639660e-01 2.59244055e-01
4.20186281e-01 -4.69287276e-01 2.91957766e-01 5.61079144e-01
-4.13157016e-01 3.61509949e-01 -2.44433776e-01 -2.74278820e-01
-1.23021197e+00 5.85996270e-01 3.73766690e-01 -6.76060140e-01
-5.02997279e-01 1.03190660e+00 2.24890783e-01 -5.65374196e-01
4.29859251e-01 -4.39527184e-01 -5.88111103e-01 5.62464774e-01
8.19633722e-01 2.89081365e-01 6.67681634e-01 -4.07899439e-01
-5.83582222e-01 3.84814963e-02 -3.70535433e-01 9.85483080e-02
1.75690305e+00 1.51326492e-01 -3.39747429e-01 3.67278725e-01
1.52731705e+00 -3.94045413e-01 -4.26224113e-01 -1.36372168e-02
-4.43366207e-02 -1.28990680e-01 7.62502730e-01 -1.08755565e+00
-9.24366593e-01 1.28717148e+00 1.06112266e+00 2.80527651e-01
1.14482665e+00 -2.72053450e-01 6.80722117e-01 6.96921527e-01
-1.41984433e-01 -1.25339055e+00 2.73122638e-01 6.24354064e-01
1.13723147e+00 -1.18761134e+00 -1.40354067e-01 1.15554154e-01
-6.20527506e-01 1.30592680e+00 3.25579345e-01 -4.99998152e-01
9.96760011e-01 1.54495776e-01 -3.80697101e-02 -5.18135965e-01
-8.80928338e-01 -8.37978125e-02 3.17231685e-01 7.37121999e-01
8.48395765e-01 8.74317959e-02 -5.24461031e-01 1.21802700e+00
-5.44979945e-02 2.12039351e-01 1.75861180e-01 8.13901424e-01
-2.93022782e-01 -9.52907860e-01 -1.30368263e-01 4.80969071e-01
-9.37315047e-01 -2.24815011e-02 -6.28591239e-01 6.37387812e-01
3.38060975e-01 1.05745292e+00 1.69307753e-01 -3.66463721e-01
2.35976204e-01 4.07874167e-01 2.60227919e-01 -6.64954543e-01
-8.39483142e-01 1.01482585e-01 3.74895930e-01 -4.65685785e-01
-4.81301069e-01 -3.97535294e-01 -1.15296924e+00 -1.06044531e-01
-2.12263957e-01 7.73379058e-02 6.76725984e-01 8.01374316e-01
4.60814059e-01 6.79430842e-01 5.27678430e-01 -8.98222446e-01
-3.69274139e-01 -1.08808291e+00 -9.72491801e-01 4.90836442e-01
4.72632885e-01 -1.07536924e+00 -3.88860673e-01 -3.73657554e-01] | [13.133757591247559, 5.770815372467041] |
91cff4aa-004d-409c-93df-22f1d35f2707 | namer-a-node-based-multitasking-framework-for | null | null | https://aclanthology.org/2021.naacl-demos.3 | https://aclanthology.org/2021.naacl-demos.3.pdf | NAMER: A Node-Based Multitasking Framework for Multi-Hop Knowledge Base Question Answering | We present NAMER, an open-domain Chinese knowledge base question answering system based on a novel node-based framework that better grasps the structural mapping between questions and KB queries by aligning the nodes in a query with their corresponding mentions in question. Equipped with techniques including data augmentation and multitasking, we show that the proposed framework outperforms the previous SoTA on CCKS CKBQA dataset. Moreover, we develop a novel data annotation strategy that facilitates the node-to-mention alignment, a dataset (https://github.com/ridiculouz/CKBQA) with such strategy is also published to promote further research. An online demo of NAMER (http://kbqademo.gstore.cn) is provided to visualize our framework and supply extra information for users, a video illustration (https://youtu.be/yetnVye{\_}hg4) of NAMER is also available. | ['Sen Hu', 'Yinnian Lin', 'Lei Zou', 'Ruoyu Zhang', 'Minhao Zhang'] | 2021-06-01 | null | null | null | naacl-2021-4 | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-4.03199643e-01 4.42440003e-01 -1.78127006e-01 -4.00917292e-01
-1.18784094e+00 -7.79880524e-01 1.35618061e-01 2.56108761e-01
-4.01881874e-01 9.63159919e-01 4.19879526e-01 -5.60844481e-01
-3.12999070e-01 -6.54473722e-01 -6.14004195e-01 -2.03840807e-01
3.56828451e-01 7.40874946e-01 6.32030904e-01 -7.76796579e-01
2.28359371e-01 -3.16645086e-01 -9.60981607e-01 3.66300285e-01
1.22535062e+00 8.10363054e-01 6.83043063e-01 5.85588813e-01
-3.47996831e-01 8.16266298e-01 -6.16896689e-01 -7.99947262e-01
-9.56147388e-02 -1.45943403e-01 -1.65847814e+00 -6.12228394e-01
4.08202350e-01 -6.48814887e-02 -5.96277475e-01 1.02793133e+00
5.81278086e-01 2.39114687e-01 2.17014313e-01 -1.22964215e+00
-1.23653531e+00 8.46824169e-01 -1.03768937e-01 4.48447794e-01
6.28053069e-01 -3.12552422e-01 1.64173281e+00 -7.06813216e-01
7.27053404e-01 7.48538733e-01 1.73766434e-01 4.94908601e-01
-3.74227434e-01 -6.84667408e-01 -2.03974359e-02 8.22667718e-01
-1.59976780e+00 -2.97556996e-01 6.58986270e-01 8.47391784e-02
9.15058851e-01 4.70301718e-01 2.99831212e-01 7.88343072e-01
-5.33878326e-01 9.09244061e-01 6.90144360e-01 -4.68995810e-01
-1.04503915e-01 -4.17864434e-02 4.91869837e-01 7.81775415e-01
2.05286462e-02 -7.60854363e-01 -5.38135171e-01 -2.65456975e-01
4.28562045e-01 -2.89664090e-01 -7.18610525e-01 6.89081401e-02
-1.11373019e+00 7.31449366e-01 5.81497252e-01 6.39822662e-01
-1.39328167e-01 1.91401601e-01 1.38018653e-01 4.65458453e-01
3.22134614e-01 6.94749057e-01 -9.70135212e-01 -1.27275333e-01
-2.63855070e-01 2.47946590e-01 1.05175698e+00 1.43723261e+00
1.04779959e+00 -6.81847334e-01 -1.79680094e-01 9.48036671e-01
2.09517896e-01 3.78992319e-01 3.72820050e-01 -1.09787738e+00
6.09092593e-01 9.49073553e-01 2.46314123e-01 -1.00129700e+00
-4.65522319e-01 -7.34314993e-02 -3.56272787e-01 -1.07979095e+00
5.31775594e-01 -1.85968861e-01 -4.28756177e-01 1.45725214e+00
7.24589407e-01 2.16713563e-01 3.80262882e-01 1.05266523e+00
1.84564435e+00 7.45054603e-01 -1.56406090e-02 3.19055974e-01
1.90563476e+00 -1.12212133e+00 -1.03914416e+00 1.32593870e-01
1.17006099e+00 -7.15328515e-01 1.38282084e+00 -1.10058330e-01
-7.65274405e-01 -2.97211707e-01 -5.91103375e-01 -8.06187391e-01
-7.75159299e-01 2.95263857e-01 2.78785795e-01 4.34317857e-01
-1.12882304e+00 -1.66905671e-01 -5.08289576e-01 -6.32270038e-01
1.36456490e-01 -2.55699940e-02 -3.94181073e-01 -1.27588302e-01
-1.87184381e+00 8.85359168e-01 5.82895458e-01 -4.79134843e-02
-6.96057498e-01 -8.58060002e-01 -7.96274424e-01 1.56104565e-02
1.04515505e+00 -6.01877749e-01 1.45802915e+00 -2.95141876e-01
-1.25790262e+00 8.10545266e-01 -4.01831269e-01 -2.31094852e-01
-1.05755955e-01 -4.87880856e-01 -4.91415858e-01 4.25238967e-01
1.39881179e-01 5.28936982e-01 2.27741808e-01 -9.10089254e-01
-6.63397193e-01 -4.19270486e-01 7.34936535e-01 2.98391998e-01
-4.67141926e-01 2.63523221e-01 -1.03531456e+00 -4.69627917e-01
-1.51839197e-01 -5.96546352e-01 2.81246903e-04 -4.21918988e-01
-6.14152908e-01 -9.12019432e-01 6.06921136e-01 -1.16873181e+00
1.64370799e+00 -1.69055712e+00 -9.21175554e-02 1.31820953e-02
4.68185663e-01 3.27808678e-01 -2.53743827e-01 1.05925202e+00
1.24286845e-01 3.36424559e-01 -1.20789379e-01 5.74794672e-02
-4.95401397e-02 2.15319172e-01 -1.46017760e-01 1.66523740e-01
-2.17712647e-03 1.44045806e+00 -1.03574944e+00 -6.53393686e-01
-3.89265299e-01 2.26363093e-01 -2.79232800e-01 2.69152999e-01
-6.81377351e-01 4.19868171e-01 -9.07947838e-01 9.34202194e-01
4.78566974e-01 -7.61331618e-01 3.64258736e-01 -3.46182287e-01
9.08178687e-02 7.15824127e-01 -8.42945993e-01 1.97470331e+00
-2.89140403e-01 3.68011057e-01 1.90912977e-01 -8.94065619e-01
1.05537307e+00 6.88556671e-01 2.23518103e-01 -6.83826149e-01
-6.10737763e-02 3.45631719e-01 -2.98787266e-01 -9.84611511e-01
8.60414267e-01 5.99566400e-01 -3.92378449e-01 3.29110026e-01
8.99455622e-02 -7.12291449e-02 2.14930415e-01 7.68756330e-01
1.26868451e+00 8.80253613e-02 3.84640574e-01 -3.77539515e-01
7.10647225e-01 5.37884474e-01 4.64437306e-01 6.32699728e-01
-7.30771273e-02 1.47155970e-01 4.05201733e-01 -8.78860950e-02
-7.70182192e-01 -6.62452400e-01 -1.34906948e-01 1.38427305e+00
2.23574698e-01 -7.40751445e-01 -7.17608631e-01 -7.04620957e-01
3.34694274e-02 7.87699521e-01 -6.50842309e-01 2.24572837e-01
-6.04191959e-01 -2.45236963e-01 1.02586544e+00 2.81140715e-01
5.78940034e-01 -1.00429273e+00 -2.40017906e-01 5.74400909e-02
-9.95789945e-01 -1.18100893e+00 -4.89929110e-01 -1.44752756e-01
-3.81713778e-01 -1.31707180e+00 -7.76465774e-01 -1.01879680e+00
4.52841580e-01 3.20126176e-01 1.39560330e+00 7.21905649e-01
-1.09683082e-01 7.51595020e-01 -1.22608721e+00 -2.31038660e-01
-1.54626489e-01 7.15167105e-01 -3.74266773e-01 -3.81158829e-01
6.51362240e-01 -2.42738426e-01 -7.02933371e-01 6.45055234e-01
-8.97660673e-01 2.38859311e-01 1.75922260e-01 5.47407746e-01
3.82735789e-01 -4.90200073e-01 1.13402522e+00 -9.66740847e-01
7.23015845e-01 -8.82716596e-01 -6.98339224e-01 9.10163403e-01
-4.10340428e-01 -1.72121301e-01 1.49295822e-01 9.68878567e-02
-1.02371883e+00 -3.75120252e-01 -6.55405343e-01 -1.15542766e-02
-2.19047159e-01 9.94283915e-01 -3.04170668e-01 2.22858697e-01
5.05740643e-01 1.45336881e-01 -2.96208411e-01 -8.91215682e-01
7.68402874e-01 1.16737401e+00 5.59623718e-01 -6.64619088e-01
6.09838307e-01 2.34195188e-01 -7.78224349e-01 -7.10415423e-01
-1.00641298e+00 -9.46581423e-01 -8.28696430e-01 -2.29858652e-01
8.12459052e-01 -1.02660739e+00 -1.03785872e+00 1.35627344e-01
-1.11678636e+00 -4.18357998e-01 9.39179212e-02 1.48579225e-01
-3.30390573e-01 4.31455106e-01 -7.01419592e-01 -4.49261308e-01
-4.87818301e-01 -6.27928138e-01 7.16447890e-01 4.02003556e-01
3.06906700e-02 -1.16115534e+00 1.10299319e-01 1.26863170e+00
2.92961329e-01 -2.65693158e-01 9.33078527e-01 -1.47643280e+00
-5.40224552e-01 -1.45328042e-04 -2.08835259e-01 -2.76384540e-02
1.94124743e-01 -2.40431637e-01 -6.03466690e-01 -2.03323022e-01
-4.13791060e-01 -5.01753390e-01 6.43578947e-01 -1.16613023e-01
1.04948711e+00 -4.35588270e-01 -3.81444812e-01 1.39056608e-01
1.35346818e+00 -8.31925869e-03 6.01647735e-01 5.40755630e-01
7.89868951e-01 7.92625904e-01 8.30539405e-01 3.06571841e-01
1.11434019e+00 7.10819423e-01 4.01821524e-01 2.02577770e-01
-1.19178466e-01 -3.22367400e-01 -6.86549246e-02 1.18525577e+00
-2.71305088e-02 -5.32446265e-01 -1.22064412e+00 9.44228232e-01
-1.87501001e+00 -5.58883309e-01 -4.45574492e-01 1.54313374e+00
1.00133479e+00 -5.52317262e-01 6.01532981e-02 -2.82634050e-01
6.23393536e-01 -2.57751681e-02 -3.71066600e-01 1.94472745e-02
-2.53736526e-01 1.15418479e-01 3.80632430e-01 6.89739525e-01
-7.48757839e-01 1.35532761e+00 4.78391838e+00 1.05129147e+00
-5.59131861e-01 5.69848776e-01 3.04335952e-01 2.18581811e-01
-6.22089803e-01 1.62008464e-01 -1.05865431e+00 2.15576708e-01
8.67719591e-01 -4.24183339e-01 1.68507725e-01 5.68457663e-01
4.89604026e-02 -8.12329061e-04 -4.66692895e-01 4.69583809e-01
9.62855145e-02 -1.70457900e+00 -3.41238752e-02 -2.85829335e-01
3.07940274e-01 2.99471200e-01 -2.69779384e-01 4.78442699e-01
5.22477508e-01 -7.08608270e-01 4.32517856e-01 3.13358516e-01
7.01947033e-01 -4.32704479e-01 8.20345521e-01 5.42440176e-01
-1.29995668e+00 -1.63453743e-02 -6.24924600e-01 1.51816368e-01
3.14676195e-01 3.00301731e-01 -9.09816384e-01 1.27827442e+00
1.15548050e+00 5.23626328e-01 -8.05720389e-01 9.27438378e-01
-6.58970952e-01 7.50332177e-01 -1.28893524e-01 -4.56712604e-01
9.94382128e-02 -2.39363134e-01 4.03678209e-01 1.13989460e+00
3.23763669e-01 6.03431463e-01 -1.10720195e-01 6.91012740e-01
-4.89764184e-01 5.42006314e-01 -1.43969581e-01 3.56968939e-02
9.55758154e-01 1.58767915e+00 -3.18069965e-01 -4.74033058e-01
-6.44444704e-01 8.04446876e-01 6.87732875e-01 3.44163299e-01
-7.74885714e-01 -5.58090985e-01 4.58804816e-01 -1.42005295e-01
4.29201245e-01 -5.27904853e-02 3.09990585e-01 -1.27783203e+00
-3.50230047e-03 -9.66102004e-01 1.06388628e+00 -1.06356931e+00
-1.21669650e+00 7.28991985e-01 -1.43647036e-02 -8.28207254e-01
6.18627556e-02 -3.74254405e-01 -2.25927711e-01 7.74686277e-01
-1.73042881e+00 -1.29274988e+00 -5.17791390e-01 1.07638729e+00
3.39650124e-01 1.47877634e-01 1.05299044e+00 6.08361304e-01
-6.13769531e-01 6.90969229e-01 1.08375072e-01 6.75838411e-01
7.94857860e-01 -1.11727762e+00 4.39110190e-01 6.03245139e-01
3.69894266e-01 5.91641366e-01 4.18930203e-01 -8.91420662e-01
-1.44955349e+00 -9.06733871e-01 1.17247677e+00 -7.77997732e-01
9.32186425e-01 -4.41021115e-01 -1.26929045e+00 9.75123763e-01
6.59483433e-01 -1.67888552e-01 9.61823106e-01 2.03670174e-01
-3.62715185e-01 -2.76425835e-02 -9.12543774e-01 4.07959223e-01
9.68207419e-01 -5.78964591e-01 -9.77920294e-01 4.88721251e-01
1.33433354e+00 -5.85776269e-01 -1.17420006e+00 4.12296832e-01
8.95292312e-02 -5.08205891e-01 6.44615948e-01 -8.09790015e-01
-6.76949993e-02 -5.61482131e-01 -2.61635512e-01 -8.65143001e-01
1.54260136e-02 -5.91117561e-01 -1.69437617e-01 1.46368647e+00
8.21311772e-01 -6.96899772e-01 7.63014019e-01 4.08740908e-01
-3.59369457e-01 -8.87321472e-01 -1.29534698e+00 -6.31953895e-01
1.27304703e-01 -4.61640775e-01 6.15445793e-01 1.08918691e+00
4.00207788e-01 3.92954320e-01 -1.33907199e-01 4.84898269e-01
1.82990767e-02 2.00276107e-01 6.04093850e-01 -8.33994925e-01
-4.07498553e-02 1.60483778e-01 1.81601778e-01 -1.52482367e+00
6.82538077e-02 -1.03014100e+00 -2.19045654e-01 -1.78523219e+00
7.71017149e-02 -7.29724407e-01 -1.24105088e-01 8.09277952e-01
-2.98356235e-01 1.61672398e-01 1.16782039e-01 4.21897352e-01
-1.07949042e+00 4.17074353e-01 1.26387095e+00 1.62455440e-01
9.18500964e-03 -1.27368510e-01 -9.37055826e-01 3.06460351e-01
9.87165511e-01 -5.60275733e-01 -1.89660832e-01 -7.50268161e-01
5.68419456e-01 2.66613513e-01 2.02570766e-01 -4.92175281e-01
5.90888500e-01 1.23656847e-01 -2.34229475e-01 -7.08608925e-01
4.20831382e-01 -6.63104653e-01 -3.42655927e-02 4.67868000e-01
-5.05775571e-01 3.96906972e-01 1.08987525e-01 3.40874553e-01
-3.58426601e-01 -5.00726223e-01 9.48424712e-02 -3.39526147e-01
-9.38918293e-01 3.03495765e-01 -2.84460902e-01 5.17181039e-01
8.18737030e-01 3.01634103e-01 -8.80788803e-01 -7.28960872e-01
-6.95687413e-01 9.27605033e-01 1.17600195e-01 5.74434876e-01
5.11353850e-01 -1.12543416e+00 -9.04617906e-01 -5.12914479e-01
5.17805934e-01 1.48552850e-01 5.64694166e-01 8.89511406e-01
-5.83684683e-01 9.00626540e-01 9.74202752e-02 -1.23722907e-02
-1.38123381e+00 4.04877394e-01 1.60236925e-01 -3.58115077e-01
-3.84311289e-01 1.16826093e+00 -2.88980305e-01 -8.61165941e-01
6.32743537e-02 -9.14771780e-02 -6.54762745e-01 1.20631076e-01
4.91717160e-01 3.37667376e-01 1.79733261e-01 -6.50944471e-01
-5.96248806e-01 1.64333567e-01 -4.94683012e-02 1.14921510e-01
1.23108387e+00 -7.10574329e-01 -5.26225090e-01 1.88454285e-01
1.03830791e+00 1.65465534e-01 -5.04996419e-01 -7.20896482e-01
4.06208962e-01 -2.36264452e-01 -4.38765943e-01 -8.91825795e-01
-9.03856039e-01 5.24551630e-01 8.49265829e-02 3.27327609e-01
1.13859224e+00 7.44913399e-01 9.54714537e-01 8.79117250e-01
2.76674747e-01 -9.55536187e-01 1.27195969e-01 7.40891695e-01
9.97479081e-01 -1.21442258e+00 -3.07151437e-01 -5.68215847e-01
-5.52824497e-01 8.87872636e-01 7.35441267e-01 2.55035520e-01
6.98659956e-01 -1.21172100e-01 4.86083001e-01 -4.91202474e-01
-9.51210558e-01 -6.40195310e-01 2.37142161e-01 6.07476711e-01
2.93884367e-01 -1.69444621e-01 -4.59637344e-01 9.05386269e-01
-4.92683411e-01 -1.36021540e-01 5.21966279e-01 7.99898088e-01
-4.42874372e-01 -1.17060423e+00 -5.88979237e-02 3.45739514e-01
-6.96138322e-01 -5.19439042e-01 -1.95995405e-01 8.56575727e-01
-1.07797444e-01 1.25400805e+00 9.99050140e-02 -3.45644027e-01
1.99535638e-01 1.54168651e-01 -4.97123040e-02 -7.78832853e-01
-6.57031536e-01 -2.98214316e-01 5.34391105e-01 -3.72910857e-01
-6.39133811e-01 -3.03216279e-01 -1.39083254e+00 -2.20299527e-01
-5.61269820e-01 7.46799171e-01 3.71603459e-01 7.50397444e-01
6.37782454e-01 2.56723046e-01 8.44247714e-02 1.68306172e-01
-8.40392858e-02 -1.22942173e+00 -4.14971858e-01 2.16909751e-01
-2.14221366e-02 -2.93263465e-01 -1.55405581e-01 -2.20820657e-03] | [10.753992080688477, 7.9877424240112305] |
c1238801-3646-48be-b108-7446c3305c4c | a-joint-framework-for-ancient-chinese-ws-and | null | null | https://aclanthology.org/2022.lt4hala-1.27 | https://aclanthology.org/2022.lt4hala-1.27.pdf | A Joint Framework for Ancient Chinese WS and POS Tagging Based on Adversarial Ensemble Learning | Ancient Chinese word segmentation and part-of-speech tagging tasks are crucial to facilitate the study of ancient Chinese and the dissemination of traditional Chinese culture. Current methods face problems such as lack of large-scale labeled data, individual task error propagation, and lack of robustness and generalization of models. Therefore, we propose a joint framework for ancient Chinese WS and POS tagging based on adversarial ensemble learning, called AENet. On the basis of pre-training and fine-tuning, AENet uses a joint tagging approach of WS and POS tagging and treats it as a joint sequence tagging task. Meanwhile, AENet incorporates adversarial training and ensemble learning, which effectively improves the model recognition efficiency while enhancing the robustness and generalization of the model. Our experiments demonstrate that AENet improves the F1 score of word segmentation by 4.48% and the score of part-of-speech tagging by 2.29% on test dataset compared with the baseline, which shows high performance and strong generalization. | ['Shuxun Yang'] | null | null | null | null | lt4hala-lrec-2022-6 | ['chinese-word-segmentation', 'culture'] | ['natural-language-processing', 'speech'] | [ 1.18132748e-01 -9.44505408e-02 1.10262394e-01 -3.64727259e-01
-8.34892631e-01 -7.42558062e-01 2.42383853e-01 -3.89107645e-01
-9.42187726e-01 7.34752655e-01 1.04164049e-01 -5.09899676e-01
5.20049751e-01 -6.35924935e-01 -4.60179448e-01 -7.13861465e-01
1.32414013e-01 1.51992321e-01 4.43664014e-01 -2.70283967e-01
-6.95815086e-02 7.03302398e-02 -5.58972299e-01 7.88100213e-02
1.36397159e+00 8.03625166e-01 3.48660648e-01 5.16654491e-01
-2.87257314e-01 3.67864996e-01 -9.89916205e-01 -6.34602308e-01
-6.99697137e-02 -3.71481299e-01 -7.46043742e-01 -2.06691369e-01
-2.69925624e-01 3.66929062e-02 -2.51081139e-01 1.29068673e+00
7.88836300e-01 1.59959063e-01 3.73556614e-01 -6.32603705e-01
-9.98488128e-01 9.36526716e-01 -3.09189707e-01 1.26457393e-01
-1.56472757e-01 -1.09737337e-01 1.00625718e+00 -6.56644881e-01
3.62579882e-01 8.02471340e-01 8.16540003e-01 9.20040369e-01
-8.99457812e-01 -9.78767991e-01 3.69669288e-01 -6.17268775e-03
-1.31410456e+00 -1.01954021e-01 5.68141699e-01 -4.44219671e-02
7.89714754e-01 7.42416382e-02 4.74961370e-01 1.15715921e+00
2.10984461e-02 1.20337379e+00 1.01247203e+00 -5.05665600e-01
2.38812715e-01 -1.31802186e-01 1.72200814e-01 5.86073399e-01
-5.40492944e-02 -8.30765441e-02 -1.62698448e-01 7.42516443e-02
5.63565612e-01 -2.08677500e-01 -2.70453617e-02 4.77791339e-01
-9.65539455e-01 7.49794841e-01 2.34630451e-01 5.88166833e-01
-1.10575914e-01 -8.41292441e-02 3.68744820e-01 -2.15189278e-01
6.91067636e-01 4.40351665e-01 -9.21302915e-01 -1.47261754e-01
-7.31252611e-01 -1.24120757e-01 5.98079443e-01 9.52135086e-01
2.76349366e-01 4.22391444e-01 -7.79110491e-02 1.16744435e+00
1.07305691e-01 1.00377119e+00 7.87593186e-01 -5.87347984e-01
3.19478422e-01 4.72828925e-01 -2.95475096e-01 -4.21663314e-01
-2.84666300e-01 -5.81223845e-01 -6.30093575e-01 -2.42644116e-01
2.65467204e-02 -8.36555779e-01 -1.49838209e+00 1.82400250e+00
6.93261400e-02 3.67467970e-01 2.02339217e-01 5.24299920e-01
5.30732155e-01 8.26090395e-01 8.08560312e-01 5.30046746e-02
1.34787703e+00 -8.80606949e-01 -7.82861173e-01 -6.40983701e-01
5.76959372e-01 -8.64717722e-01 1.02124691e+00 8.53343308e-02
-8.70197713e-01 -6.06925488e-01 -1.00194836e+00 1.37178764e-01
-2.61933982e-01 2.80632198e-01 4.39492941e-01 7.95687139e-01
-5.65821528e-01 3.77151608e-01 -1.13703775e+00 -4.59321067e-02
5.54870427e-01 5.31673431e-01 -2.53841311e-01 -8.61627832e-02
-1.35611641e+00 6.38318300e-01 8.25355113e-01 1.11616127e-01
-4.89141166e-01 -6.75289392e-01 -8.01718593e-01 2.73243319e-02
2.83328772e-01 -1.68824404e-01 1.32103658e+00 -9.53488648e-01
-1.62241292e+00 5.24733245e-01 1.30758792e-01 -1.96819127e-01
3.01635385e-01 -7.29927897e-01 -8.59426558e-01 -1.78887650e-01
2.11012974e-01 4.63736355e-01 2.88620681e-01 -9.43252921e-01
-8.31124961e-01 -3.11413050e-01 -3.63319248e-01 2.97477772e-03
-2.96838790e-01 1.67851672e-01 -6.96629584e-01 -1.01641941e+00
6.44744113e-02 -1.00982428e+00 -5.47442257e-01 -8.48580301e-01
-3.38453382e-01 -1.95312932e-01 7.67349780e-01 -1.24340785e+00
1.27740264e+00 -2.38396168e+00 -2.41494074e-01 7.45646581e-02
-2.46313706e-01 9.94391203e-01 -3.31941336e-01 4.14572768e-02
1.72319129e-01 5.02143145e-01 -5.20639122e-01 -2.42081255e-01
-1.87793553e-01 4.38081145e-01 -1.22251317e-01 1.69694975e-01
5.63408852e-01 9.25826967e-01 -1.03627491e+00 -6.04380786e-01
-1.14129536e-01 4.61208493e-01 -3.03797692e-01 3.48602414e-01
-2.64683306e-01 4.96203840e-01 -7.58421302e-01 6.69732571e-01
7.47313380e-01 7.83564299e-02 3.90340507e-01 1.76933855e-01
5.89359924e-02 3.87598872e-01 -9.18804646e-01 1.84696162e+00
-4.20955598e-01 1.60940215e-01 -3.46487552e-01 -8.62516522e-01
9.92619991e-01 2.81781852e-01 2.35187128e-01 -1.00193226e+00
1.60741404e-01 3.72695923e-01 2.19769225e-01 -4.08092201e-01
5.38142145e-01 -3.12793344e-01 -4.53877479e-01 3.27370077e-01
2.30799064e-01 3.46551508e-01 -8.80339369e-02 2.57784799e-02
1.15469217e+00 4.99004006e-01 1.78006962e-01 -1.91380262e-01
5.30931592e-01 -2.17058361e-02 1.09622598e+00 4.23925400e-01
-1.13308370e-01 5.88612139e-01 3.99166346e-02 -1.45510221e-02
-1.04747522e+00 -1.04811180e+00 -8.35507810e-02 1.28268468e+00
9.76219773e-02 -7.98446387e-02 -1.03817129e+00 -1.17848432e+00
-3.44768077e-01 8.24973106e-01 -4.88191277e-01 -2.31657878e-01
-1.08177245e+00 -1.02057254e+00 1.08128548e+00 9.93727267e-01
8.67483437e-01 -1.36217451e+00 3.49379927e-02 4.56912279e-01
-3.30094546e-01 -1.34676957e+00 -6.09220505e-01 2.04448447e-01
-6.52655959e-01 -7.77658284e-01 -8.85004938e-01 -1.19824874e+00
5.61450541e-01 -1.64882734e-01 7.90652335e-01 1.33695945e-01
9.64690447e-02 -2.53784418e-01 -8.59128356e-01 -5.17011940e-01
-4.98624921e-01 3.65735501e-01 -3.30248773e-02 -1.80371925e-01
4.39175963e-01 -2.72144735e-01 -4.20828462e-01 2.93174177e-01
-9.92525458e-01 -3.00279170e-01 8.62549961e-01 9.28625643e-01
4.26392525e-01 -5.43614551e-02 6.38376236e-01 -1.32010734e+00
3.07788700e-01 -3.84783626e-01 -3.70019674e-01 5.61451852e-01
-4.75282788e-01 3.14065441e-02 6.38182104e-01 -6.29134774e-01
-1.58893907e+00 5.46031035e-02 -6.79385722e-01 2.41829485e-01
-4.92441319e-02 3.67781401e-01 -6.06504619e-01 9.07701626e-02
3.72120202e-01 3.93131047e-01 -3.31645280e-01 -6.16937578e-01
3.78747106e-01 9.96600449e-01 6.74707413e-01 -5.53641021e-01
7.10903287e-01 3.09445504e-02 -5.76458871e-01 -6.47930861e-01
-1.02258337e+00 -4.24414963e-01 -8.16910863e-01 2.75957465e-01
1.14513350e+00 -1.05089128e+00 -3.37030172e-01 7.87812054e-01
-9.95759666e-01 -5.07220149e-01 1.65103704e-01 4.28433329e-01
3.07141617e-02 5.47317684e-01 -7.38193154e-01 -7.48325109e-01
-5.53230762e-01 -7.64787376e-01 7.44124115e-01 3.85044545e-01
4.07204106e-02 -9.90604222e-01 1.77653842e-02 2.45706588e-01
2.59836018e-01 2.30055116e-02 8.91523778e-01 -1.18098485e+00
-8.98922682e-02 -3.02492440e-01 4.27307524e-02 9.26200688e-01
2.82917172e-02 -2.26119742e-01 -9.55655098e-01 -8.59386548e-02
-1.89807877e-01 -8.57036486e-02 7.30276525e-01 2.74873525e-01
9.40702617e-01 2.52954252e-02 -3.42272520e-01 5.13614595e-01
1.32238984e+00 7.11077213e-01 6.86547935e-01 3.73618454e-01
7.10515678e-01 2.88126767e-01 6.73240066e-01 1.03614032e-01
2.41372049e-01 2.78637886e-01 -1.13527045e-01 -1.67014495e-01
-5.14337681e-02 -2.38048956e-01 4.03276056e-01 1.13133895e+00
-1.57136992e-01 -2.82171398e-01 -1.07376361e+00 6.35652244e-01
-1.62346315e+00 -6.33021474e-01 1.55179873e-01 1.91431653e+00
9.83274519e-01 2.00877130e-01 1.04512759e-02 -6.45399243e-02
9.82923329e-01 -3.78023647e-02 -2.71221936e-01 -1.75270721e-01
-1.84946135e-01 7.39857495e-01 6.33763194e-01 3.56132418e-01
-1.20288968e+00 1.62450707e+00 6.15710163e+00 1.09494615e+00
-1.05870414e+00 3.22133243e-01 5.21532178e-01 5.19877017e-01
-1.75621599e-01 -1.25584409e-01 -1.02174520e+00 6.80577457e-01
8.88051450e-01 2.76331395e-01 5.99674769e-02 7.37742305e-01
-3.43717635e-01 3.32611948e-01 -5.02753913e-01 2.02397257e-01
3.25109772e-02 -8.49870145e-01 -2.78833508e-01 -8.15524161e-03
9.32895243e-01 2.46834815e-01 -1.05979845e-01 7.02245355e-01
8.70563745e-01 -8.37905884e-01 7.10991859e-01 -1.87202217e-03
7.93355048e-01 -7.91675925e-01 1.10041571e+00 3.72061551e-01
-9.79782641e-01 1.62666515e-01 -2.79126465e-01 2.06994474e-01
4.02508199e-01 2.62259364e-01 -6.24039292e-01 6.04742348e-01
6.16785586e-01 3.68718505e-01 -3.86860073e-01 7.43383288e-01
-7.61382699e-01 1.20777881e+00 -1.57837838e-01 -2.99812555e-01
3.96160543e-01 -1.17710173e-01 1.59997568e-01 1.52257216e+00
2.54954159e-01 4.77625668e-01 3.20318729e-01 3.91517371e-01
-1.52883649e-01 7.23366067e-02 1.14105344e-01 -1.86701879e-01
7.41957307e-01 1.09403288e+00 -7.16183901e-01 -1.90020069e-01
-5.01177549e-01 1.07767558e+00 4.76019472e-01 4.24035549e-01
-1.02151597e+00 -6.27500296e-01 5.05950451e-01 -4.69200134e-01
7.31863439e-01 -3.08679163e-01 -4.95370686e-01 -1.12312877e+00
-1.11387469e-01 -7.79767871e-01 2.99239188e-01 -2.81697154e-01
-1.22771657e+00 8.22944760e-01 -4.41119552e-01 -8.62010658e-01
1.35702193e-01 -6.56131327e-01 -8.12440991e-01 8.72819602e-01
-1.46230352e+00 -1.49496031e+00 8.02287012e-02 2.24864811e-01
5.67869008e-01 -3.15829784e-01 8.66500795e-01 2.94904709e-01
-8.82004619e-01 7.81283319e-01 2.76337624e-01 8.80723834e-01
5.88408768e-01 -1.32548070e+00 8.88078868e-01 1.27780294e+00
1.28550023e-01 5.15358925e-01 2.14168757e-01 -9.96144652e-01
-6.61030889e-01 -1.33439517e+00 1.09590423e+00 -2.67509818e-01
6.49979949e-01 -4.68391299e-01 -1.05661595e+00 7.32874691e-01
-1.29506923e-02 -3.19040492e-02 1.04081011e+00 2.13879615e-01
-2.72135288e-01 1.30569711e-01 -8.99633646e-01 5.74016869e-01
9.34350312e-01 -4.05174673e-01 -7.82570958e-01 4.40175310e-02
9.39233482e-01 -2.36396283e-01 -9.77619469e-01 3.41735065e-01
4.84748870e-01 -3.97424430e-01 7.39559650e-01 -8.18184733e-01
2.08458409e-01 -6.25367835e-02 -2.48825744e-01 -1.33468020e+00
-5.61584949e-01 -4.22970682e-01 4.62363690e-01 1.64003015e+00
8.61956835e-01 -4.30528253e-01 7.36176491e-01 4.60530937e-01
-5.34860075e-01 -5.30005157e-01 -7.33675659e-01 -8.65723968e-01
4.30347115e-01 -5.20319104e-01 5.95634997e-01 1.01673722e+00
-1.74584582e-01 4.67988700e-01 -3.06496918e-01 3.86497140e-01
3.13608944e-01 -3.38693798e-01 2.63825923e-01 -9.24067676e-01
-2.60984808e-01 -7.29899034e-02 -1.07037924e-01 -8.52170467e-01
5.10065198e-01 -7.41082370e-01 5.64555228e-01 -1.29766381e+00
1.11438511e-02 -6.42182887e-01 -6.21954203e-01 7.10555673e-01
-8.01448286e-01 2.37990439e-01 2.47918412e-01 2.19552908e-02
-6.59152567e-01 5.08793235e-01 1.16338813e+00 1.24493249e-01
-1.81014866e-01 1.35158658e-01 -6.07612133e-01 6.37578964e-01
9.46908116e-01 -6.67115986e-01 1.20071538e-01 -7.13113546e-01
-2.75340796e-01 -2.28873745e-01 -1.95561171e-01 -8.61582518e-01
-1.95600037e-02 -8.12060386e-02 4.83988345e-01 -9.58331600e-02
7.02745840e-02 -5.73894083e-01 -2.01485023e-01 6.61861300e-01
-1.33052424e-01 -1.53541416e-01 2.65160233e-01 5.86597621e-01
-3.63892131e-02 -3.48877817e-01 6.63304806e-01 -2.95456976e-01
-1.07507420e+00 2.44796440e-01 -1.69072941e-01 3.84303868e-01
9.88249958e-01 2.65292218e-03 -3.13525051e-01 1.41166255e-01
-6.11701548e-01 2.56309628e-01 2.64506433e-02 4.32900608e-01
1.63541839e-01 -1.11635900e+00 -8.70796859e-01 2.68412650e-01
-1.03948064e-01 1.09455712e-01 3.22519690e-01 3.65591109e-01
-6.15016401e-01 1.01164274e-01 -2.37659901e-01 -1.90590292e-01
-1.08555114e+00 1.91102445e-01 -4.61948337e-03 -4.10872489e-01
-3.73085558e-01 1.09467053e+00 1.13772936e-01 -7.71989644e-01
1.20136011e-02 -1.04815550e-02 -2.53948838e-01 -4.85286266e-01
3.91869873e-01 2.80448496e-01 -8.96135420e-02 -6.16227150e-01
-5.33486247e-01 5.57349026e-01 -2.78098434e-01 -1.19514637e-01
1.34909129e+00 7.05919042e-02 1.26927778e-01 9.01643559e-02
9.07771468e-01 3.43830079e-01 -1.30036795e+00 -4.20299858e-01
2.00243920e-01 5.60243730e-04 -5.88072576e-02 -1.32336640e+00
-1.02988231e+00 8.92863691e-01 4.39772666e-01 -1.39392734e-01
1.09975481e+00 -6.17477186e-02 1.37268031e+00 1.63034588e-01
9.22667682e-02 -1.23971701e+00 -8.79023969e-02 7.22937286e-01
1.12089343e-01 -1.23371661e+00 -3.28105509e-01 -4.72269177e-01
-1.04332507e+00 7.22324252e-01 7.08746135e-01 -1.99393988e-01
5.89177787e-01 4.99977529e-01 6.02327347e-01 4.21979368e-01
-1.13067977e-01 -4.00826246e-01 5.02055526e-01 7.36554146e-01
4.84718353e-01 2.17524663e-01 -4.77267772e-01 1.20296216e+00
-1.90287396e-01 -2.78233916e-01 -4.94340211e-02 9.89792228e-01
-5.88586926e-01 -1.53061581e+00 1.07930370e-01 8.26866925e-03
-1.15923893e+00 -4.49812144e-01 -2.38338187e-01 9.11173761e-01
3.29883754e-01 9.32133377e-01 5.27425893e-02 -6.47389472e-01
3.75688672e-01 2.24774003e-01 1.74784973e-01 -6.97908759e-01
-7.68639028e-01 3.30826819e-01 1.27601698e-01 2.57366281e-02
-3.25438648e-01 -4.55230951e-01 -1.60924506e+00 -1.65633522e-02
-6.56875670e-01 4.47158277e-01 6.85418010e-01 1.36670697e+00
2.30790228e-01 6.53440833e-01 5.33420682e-01 -2.01202482e-01
-7.20483780e-01 -1.28765976e+00 -7.04508424e-01 4.53282416e-01
-3.20540890e-02 -1.55743107e-01 8.52672458e-02 2.48143733e-01] | [9.992293357849121, 10.08755111694336] |
08ad9184-d53e-4b9a-95cd-6f6b47e3c048 | transfer-learning-with-class-weighted-and | 2009.05977 | null | https://arxiv.org/abs/2009.05977v1 | https://arxiv.org/pdf/2009.05977v1.pdf | Transfer learning with class-weighted and focal loss function for automatic skin cancer classification | Skin cancer is by far in top-3 of the world's most common cancer. Among different skin cancer types, melanoma is particularly dangerous because of its ability to metastasize. Early detection is the key to success in skin cancer treatment. However, skin cancer diagnosis is still a challenge, even for experienced dermatologists, due to strong resemblances between benign and malignant lesions. To aid dermatologists in skin cancer diagnosis, we developed a deep learning system that can effectively and automatically classify skin lesions into one of the seven classes: (1) Actinic Keratoses, (2) Basal Cell Carcinoma, (3) Benign Keratosis, (4) Dermatofibroma, (5) Melanocytic nevi, (6) Melanoma, (7) Vascular Skin Lesion. The HAM10000 dataset was used to train the system. An end-to-end deep learning process, transfer learning technique, utilizing multiple pre-trained models, combining with class-weighted and focal loss were applied for the classification process. The result was that our ensemble of modified ResNet50 models can classify skin lesions into one of the seven classes with top-1, top-2 and top-3 accuracy 93%, 97% and 99%, respectively. This deep learning system can potentially be integrated into computer-aided diagnosis systems that support dermatologists in skin cancer diagnosis. | ['Lua T. Ngo', 'Duyen N. T. Le', 'Hieu X. Le', 'Hoan T. Ngo'] | 2020-09-13 | null | null | null | null | ['skin-cancer-classification'] | ['medical'] | [ 3.75779361e-01 -6.33538812e-02 -1.55091628e-01 1.72995985e-01
-7.53322244e-01 -3.69012237e-01 4.79000509e-01 4.70415175e-01
-3.60374242e-01 1.06480360e+00 9.33606625e-02 -4.31539983e-01
-1.67488873e-01 -9.06895399e-01 4.38081883e-02 -1.13441777e+00
1.83654100e-01 2.64140248e-01 2.84704119e-01 -1.51475653e-01
6.95938691e-02 7.37798333e-01 -9.75750685e-01 4.41964895e-01
1.09434593e+00 1.14837301e+00 -1.13434732e-01 9.66105700e-01
-1.71413511e-01 7.24629462e-01 -1.22360118e-01 -3.18514019e-01
6.10083230e-02 -3.11784536e-01 -8.44338477e-01 -9.69612822e-02
5.11506796e-01 -8.40035751e-02 -1.16276667e-01 9.89829719e-01
6.85137331e-01 -4.38409775e-01 1.10299683e+00 -7.33352304e-01
-3.53258222e-01 -5.23910671e-02 -9.04884577e-01 2.07770020e-02
3.70794952e-01 1.14172831e-01 4.77690220e-01 -6.41248584e-01
6.77821100e-01 6.66971862e-01 9.94502664e-01 9.12617445e-01
-7.18664646e-01 -3.12322527e-01 -3.27258855e-01 2.77163923e-01
-1.33857942e+00 3.85288037e-02 4.19923626e-02 -4.30547416e-01
9.06008840e-01 4.11741167e-01 9.23994303e-01 1.08122861e+00
7.08189368e-01 5.14723480e-01 1.43334103e+00 -5.14003158e-01
3.91028970e-01 4.15669888e-01 8.50779042e-02 7.59860098e-01
9.16815624e-02 -1.86657254e-02 -5.10958098e-02 -1.68639403e-02
6.75736725e-01 1.16942950e-01 -6.88341260e-02 3.04658592e-01
-4.34748054e-01 5.99992156e-01 7.12557435e-01 2.72725046e-01
-5.51771343e-01 -1.15568325e-01 4.91128266e-01 2.40262732e-01
3.98677468e-01 -1.24701835e-01 -3.05579811e-01 1.98479995e-01
-5.38247824e-01 -8.51320103e-02 6.30039573e-01 7.40085244e-02
2.28201255e-01 -2.52633184e-01 -7.91602358e-02 1.07004964e+00
2.39675805e-01 1.50673047e-01 8.32542896e-01 1.06592633e-01
-1.44072801e-01 9.67059255e-01 -3.03630203e-01 -3.29473227e-01
-5.71603239e-01 -3.53750527e-01 -1.37693799e+00 5.53198755e-01
3.28582823e-01 -1.54736340e-01 -1.62948811e+00 1.20159101e+00
4.19056803e-01 1.38983637e-01 1.66531742e-01 9.40234959e-01
8.62928450e-01 2.03683212e-01 4.31811064e-01 1.81472361e-01
1.52994931e+00 -8.73168051e-01 -2.06364214e-01 1.37509599e-01
5.40498912e-01 -8.68272126e-01 5.17090619e-01 4.76846159e-01
-8.67741346e-01 -2.33403385e-01 -8.71042609e-01 8.85558501e-02
-6.08347297e-01 -8.48047994e-03 8.35343719e-01 8.90750468e-01
-1.16423440e+00 4.28578377e-01 -7.83283591e-01 -1.25666904e+00
7.35701501e-01 3.37435097e-01 -7.79148579e-01 -2.09128290e-01
-1.06175089e+00 1.24812770e+00 2.06958115e-01 -2.33100146e-01
-8.13372970e-01 -7.62524545e-01 -4.67313141e-01 -3.83687615e-01
1.73451938e-02 -1.02608502e+00 8.67704630e-01 -1.20548904e+00
-1.41232145e+00 1.26674271e+00 -5.64573333e-02 -4.71689671e-01
6.30341887e-01 6.04538620e-02 -7.37298310e-01 2.30329797e-01
-1.86970830e-01 4.04870957e-01 4.83865410e-01 -7.81922698e-01
-1.12319398e+00 -5.17132461e-01 -1.43998459e-01 3.34681004e-01
-2.61834025e-01 -1.62974775e-01 -1.43980116e-01 -1.27975434e-01
-2.54464835e-01 -8.75293732e-01 -4.26339984e-01 5.43029368e-01
-8.54651093e-01 -3.22634876e-01 6.63003147e-01 -9.27478373e-01
1.03632522e+00 -1.82551169e+00 -5.41353635e-02 3.38765293e-01
2.23537266e-01 8.04247737e-01 4.10494804e-02 6.10157192e-01
-2.26147532e-01 4.06282634e-01 -3.62207480e-02 1.87741235e-01
-5.69963872e-01 -2.23270833e-01 5.17448366e-01 4.71632928e-01
3.62754554e-01 6.85046136e-01 -7.71471441e-01 -2.61386186e-01
5.31028211e-01 7.21433878e-01 3.49396139e-01 -4.28799503e-02
8.32755864e-02 1.04393344e-02 -2.28019312e-01 1.06026304e+00
8.26167464e-01 -2.29688764e-01 -7.28947446e-02 -3.03327888e-01
2.25770712e-01 -2.35832989e-01 -8.68413806e-01 1.23528326e+00
-5.53685904e-01 5.66445768e-01 8.57699588e-02 -2.83590347e-01
4.97570336e-01 4.47907865e-01 3.23711216e-01 -4.69044685e-01
8.79020020e-02 2.18767345e-01 1.78906471e-01 -7.85018265e-01
-3.06418151e-01 -4.50677127e-01 5.86522281e-01 -6.98233664e-04
-2.67849445e-01 2.79244244e-01 2.10206360e-01 -2.64347531e-02
1.30084884e+00 -3.48940015e-01 7.47910559e-01 -3.17792920e-03
6.63960278e-01 4.61183116e-02 3.35119575e-01 1.99806601e-01
-3.08897674e-01 6.51014507e-01 4.42537427e-01 -6.74447894e-01
-8.75783443e-01 -1.46358943e+00 -4.32536274e-01 4.41348583e-01
-1.51279867e-01 1.86403781e-01 -7.23942041e-01 -6.75612330e-01
8.55760649e-02 3.05080146e-01 -7.95811594e-01 -1.56407803e-01
1.10238798e-01 -9.17709887e-01 7.12376118e-01 4.51511621e-01
9.04963136e-01 -6.33069515e-01 -1.06107198e-01 8.10428038e-02
2.38757625e-01 -4.80958849e-01 -1.83841567e-02 1.85214803e-01
-4.21194226e-01 -1.70798385e+00 -1.14473844e+00 -1.07435787e+00
9.73475933e-01 8.89940038e-02 6.20843887e-01 1.28824741e-01
-1.28264058e+00 1.17985949e-01 -1.59147441e-01 -5.16301095e-01
-6.22011781e-01 -1.39582962e-01 5.54008707e-02 1.18957631e-01
6.67562485e-01 -2.12846503e-01 -8.14753890e-01 -1.33280143e-01
-8.52779150e-01 1.43988818e-01 9.92537022e-01 9.75863695e-01
5.47531068e-01 4.42211330e-01 5.29442489e-01 -1.13940358e+00
7.20698714e-01 -6.96470499e-01 7.17997551e-02 3.46743792e-01
-4.14270014e-01 -6.87099576e-01 8.04399133e-01 -2.41109118e-01
-1.01603127e+00 3.38972449e-01 -3.72453332e-01 1.50610565e-03
-5.10137498e-01 8.40320230e-01 1.77207008e-01 -3.52093786e-01
1.15781057e+00 2.51180470e-01 2.07670122e-01 -3.64135928e-03
-1.56869575e-01 1.08798635e+00 3.05265039e-01 2.52554268e-01
5.58830202e-01 4.80471104e-01 3.97084236e-01 -1.24376678e+00
-6.41021252e-01 -6.96404457e-01 -3.99584472e-01 -2.23041862e-01
9.79527950e-01 -1.03218842e+00 -1.96719617e-01 1.18914533e+00
-6.96021318e-01 -1.17023714e-01 1.84313133e-02 3.36511657e-02
4.54996787e-02 4.60782647e-01 -8.36396754e-01 -7.54535794e-01
-8.65987241e-01 -5.68541169e-01 8.51523519e-01 8.64316761e-01
-3.97885710e-01 -1.55787861e+00 1.49446279e-01 3.56168896e-01
5.83886027e-01 7.58096874e-01 1.10265982e+00 -5.66897511e-01
-1.19845264e-01 -8.79836798e-01 -3.50170255e-01 5.28363585e-01
7.05012798e-01 4.83151823e-01 -1.06030738e+00 -4.03810561e-01
-7.65143573e-01 -5.65193236e-01 1.01846015e+00 5.13578713e-01
9.37388480e-01 5.21506928e-02 -7.07881033e-01 6.49201691e-01
1.93608415e+00 2.83889592e-01 8.42524469e-01 7.51273185e-02
5.23289204e-01 3.48022193e-01 8.96278620e-02 6.11889511e-02
9.01024193e-02 7.16819428e-03 4.32872862e-01 -7.11969733e-01
-5.37353933e-01 -1.80836186e-01 -7.60525465e-02 1.31165102e-01
-2.77145982e-01 -2.58533061e-01 -9.81239676e-01 5.95274031e-01
-1.24135518e+00 -8.18015695e-01 -3.10769349e-01 2.21578479e+00
7.51735866e-01 1.61263824e-01 1.54942170e-01 3.85768473e-01
9.08989191e-01 -4.46427047e-01 -6.88754082e-01 -4.94627982e-01
-6.84625059e-02 5.85684896e-01 5.50476253e-01 2.14484930e-01
-1.29344034e+00 8.13699543e-01 5.56138039e+00 8.83098841e-01
-1.75426626e+00 -2.52648026e-01 7.64585078e-01 2.88207084e-01
7.15055913e-02 -4.29386824e-01 -6.50686026e-01 4.85838860e-01
5.35187840e-01 -2.45380729e-01 2.51584537e-02 5.22576392e-01
3.54977213e-02 -4.00756538e-01 -9.93220925e-01 7.22214758e-01
-3.37887332e-02 -1.40847778e+00 1.68932214e-01 -2.13187449e-02
6.91106379e-01 -5.37243783e-02 7.55096376e-02 8.16424750e-03
2.58962989e-01 -1.50639999e+00 -4.82548386e-01 7.76750207e-01
1.27323341e+00 -6.99428618e-01 1.05134070e+00 1.97654873e-01
-8.47295225e-01 6.89990669e-02 -3.14520746e-01 1.58605576e-01
-2.02278912e-01 6.41388834e-01 -1.33967268e+00 5.25085628e-01
5.41032732e-01 6.43332958e-01 -6.03446543e-01 1.43187082e+00
-1.21886380e-01 6.20012224e-01 -4.30947810e-01 -4.94963884e-01
1.25924498e-01 2.47914597e-01 1.71043530e-01 1.08854997e+00
2.52558589e-01 -1.29803285e-01 -1.49415314e-01 3.96436125e-01
2.20462039e-01 1.66849181e-01 -4.03161645e-01 -4.32664156e-02
2.92409450e-01 1.67446160e+00 -7.12403893e-01 -1.89274400e-02
-1.06277309e-01 1.00307989e+00 -3.98110300e-02 2.71115571e-01
-3.94362926e-01 -7.42313445e-01 5.53444862e-01 3.74532729e-01
-2.07150429e-01 6.67638779e-01 -2.53593326e-01 -7.53572762e-01
-2.95438260e-01 -6.03517771e-01 5.00057638e-01 -5.91704428e-01
-1.75965977e+00 5.73727429e-01 -7.80356169e-01 -1.07004118e+00
-8.16967562e-02 -1.05692029e+00 -1.16301787e+00 1.21684885e+00
-1.61115801e+00 -1.57038629e+00 -8.76622379e-01 6.01835608e-01
3.76168102e-01 -4.42917883e-01 1.28984344e+00 -1.87038612e-02
-7.21802115e-01 7.14313209e-01 1.76715761e-01 9.50197801e-02
7.59041786e-01 -1.63171029e+00 3.93834896e-04 2.72017747e-01
-7.00299680e-01 2.72611439e-01 1.38743758e-01 -5.19790769e-01
-1.01835585e+00 -1.44105184e+00 5.63675940e-01 -1.99585687e-02
6.04589641e-01 2.98447013e-01 -5.05197763e-01 -2.54657958e-02
2.41957381e-01 -2.05040933e-03 1.29353690e+00 -2.42792685e-02
-1.94950402e-01 -2.76202083e-01 -1.65712166e+00 6.18969560e-01
3.15449178e-01 -2.94786811e-01 6.68483078e-02 7.79568255e-01
-1.50694668e-01 -5.14738917e-01 -1.00188720e+00 1.79755405e-01
8.24962139e-01 -9.13119972e-01 8.39419901e-01 -7.13670969e-01
9.56132054e-01 1.11762099e-01 2.59960145e-01 -1.54720306e+00
-6.24134839e-01 -1.00608356e-01 3.63249213e-01 9.00263608e-01
5.37390947e-01 -6.30944073e-01 1.25525737e+00 5.47155559e-01
1.32286713e-01 -1.22103727e+00 -9.59479034e-01 -3.71198744e-01
4.12787855e-01 3.29872966e-01 7.01839924e-02 8.15891981e-01
-2.54930168e-01 3.10621351e-01 2.53497716e-02 1.05687425e-01
6.20489240e-01 -5.92145145e-01 4.38773066e-01 -1.19220161e+00
-2.08938345e-02 -6.58126593e-01 -8.76058221e-01 -5.54769970e-02
-4.64341849e-01 -8.74467790e-01 -6.08775616e-01 -2.18247437e+00
3.18925053e-01 -3.70922595e-01 -6.74964666e-01 6.49582982e-01
-1.61055312e-01 3.37953627e-01 -1.85069084e-01 -2.68510818e-01
-2.93944515e-02 -3.72895062e-01 1.24034214e+00 -2.97668666e-01
1.29334897e-01 4.07432437e-01 -8.94567311e-01 7.89092243e-01
7.89520800e-01 1.36395171e-01 -2.76996106e-01 -3.81768532e-02
2.04580292e-01 1.17895156e-01 3.00849974e-01 -1.24865603e+00
4.71117705e-01 -3.73143852e-01 9.34288859e-01 -4.14635211e-01
3.47253710e-01 -7.40902364e-01 9.24867392e-02 1.00288546e+00
-3.11344355e-01 -7.32198417e-01 2.08230793e-01 4.34999049e-01
-1.62908182e-01 -2.18449220e-01 1.36487925e+00 -2.91941464e-01
-8.93238544e-01 4.39344853e-01 -7.00511098e-01 -4.55820173e-01
1.77904201e+00 -6.11542165e-01 -4.96161312e-01 -1.77266553e-01
-8.36476028e-01 1.72030210e-01 3.75384599e-01 2.38903929e-02
8.26003790e-01 -1.08105934e+00 -1.04413188e+00 5.12702689e-02
2.55109876e-01 1.66288421e-01 7.25738525e-01 7.66000271e-01
-1.17504084e+00 8.81563127e-02 -4.74424005e-01 -4.45871502e-01
-1.60498655e+00 -1.19819283e-01 7.74880946e-01 -3.86475980e-01
-1.83050796e-01 1.31959915e+00 -3.69831115e-01 -1.89613089e-01
4.32592124e-01 7.13155344e-02 -2.95251071e-01 -1.87798411e-01
6.99643254e-01 4.87163186e-01 1.99204922e-01 -1.00694746e-01
-3.92473131e-01 6.22882545e-01 -9.39504802e-01 5.87081134e-01
1.01697958e+00 4.70920265e-01 -2.38497213e-01 7.93795884e-02
1.07407379e+00 -3.07989925e-01 -6.66635334e-01 7.46750161e-02
-3.80637944e-01 -1.04115970e-01 2.36951903e-01 -1.83036506e+00
-8.98728728e-01 8.95409584e-01 1.17584527e+00 2.00633451e-01
1.51851428e+00 -5.10404050e-01 9.00086462e-01 2.64574736e-01
2.11658746e-01 -1.14327502e+00 -1.69618845e-01 4.36897129e-02
5.16728997e-01 -1.40004778e+00 1.36028439e-01 -5.40055811e-01
-5.19823790e-01 1.40095830e+00 8.08724165e-01 -4.40510780e-01
6.88026130e-01 2.82764733e-01 4.88154083e-01 1.05839737e-01
-1.08006513e+00 -1.54302135e-01 4.50554527e-02 9.39602613e-01
5.29142916e-01 2.37606570e-01 -4.13928628e-01 2.77421027e-01
2.20995843e-01 3.07051659e-01 5.16569078e-01 6.77259684e-01
-4.16285008e-01 -9.70164478e-01 -8.63748938e-02 1.14113808e+00
-6.87954783e-01 -1.12030171e-01 -9.36535060e-01 1.01859975e+00
3.47799867e-01 6.66481376e-01 -5.06451242e-02 -4.32930976e-01
3.10706105e-02 -2.78114378e-01 4.79827553e-01 -5.07152081e-01
-7.64049947e-01 -1.25967309e-01 1.75206602e-01 3.75524051e-02
-4.60075885e-02 -1.35351732e-01 -9.72501934e-01 -3.26259106e-01
-2.63008386e-01 -2.97741175e-01 8.10167849e-01 7.14723766e-01
2.08063982e-02 3.76869529e-01 6.13724709e-01 -1.99299157e-02
-5.90451360e-01 -1.12833083e+00 -8.15962195e-01 3.76743644e-01
2.65149772e-01 -1.91422492e-01 -3.29229116e-01 3.38953622e-02] | [15.674290657043457, -2.9966049194335938] |
058088d4-8d18-440b-a50a-1419b8dd50a2 | from-perspective-x-ray-imaging-to-parallax | 2003.02959 | null | https://arxiv.org/abs/2003.02959v1 | https://arxiv.org/pdf/2003.02959v1.pdf | From Perspective X-ray Imaging to Parallax-Robust Orthographic Stitching | Stitching images acquired under perspective projective geometry is a relevant topic in computer vision with multiple applications ranging from smartphone panoramas to the construction of digital maps. Image stitching is an equally prominent challenge in medical imaging, where the limited field-of-view captured by single images prohibits holistic analysis of patient anatomy. The barrier that prevents straight-forward mosaicing of 2D images is depth mismatch due to parallax. In this work, we leverage the Fourier slice theorem to aggregate information from multiple transmission images in parallax-free domains using fundamental principles of X-ray image formation. The semantics of the stitched image are restored using a novel deep learning strategy that exploits similarity measures designed around frequency, as well as dense and sparse spatial image content. Our pipeline, not only stitches images, but also provides orthographic reconstruction that enables metric measurements of clinically relevant quantities directly on the 2D image plane. | ['Nassir Navab', 'Mehran Armand', 'Mathias Unberath', 'Xingtong Liu', 'Javad Fotouhi'] | 2020-03-05 | null | null | null | null | ['image-stitching'] | ['computer-vision'] | [ 8.65293086e-01 9.81106311e-02 -1.74878109e-02 -4.93669026e-02
-8.34990680e-01 -4.59286749e-01 4.97716010e-01 7.30201006e-02
-3.42479557e-01 3.83023620e-01 4.56939250e-01 -2.74547517e-01
-4.25762892e-01 -4.17304277e-01 -4.26442236e-01 -9.25995827e-01
-1.11743666e-01 1.21187262e-01 -1.20146617e-01 -5.67448959e-02
2.22343847e-01 5.38133562e-01 -8.53335738e-01 3.93640429e-01
4.68067706e-01 8.86062443e-01 2.94145346e-01 8.63856912e-01
3.14782679e-01 7.37933993e-01 -4.10246104e-01 -3.15205067e-01
5.34431517e-01 -3.07289273e-01 -4.29009318e-01 2.61490077e-01
7.17229128e-01 -9.14814353e-01 -8.02878380e-01 8.72804642e-01
5.31999052e-01 -3.90992850e-01 2.37212375e-01 -6.98458195e-01
-2.08741963e-01 2.62207597e-01 -8.21862161e-01 3.80742818e-01
4.30397391e-01 1.91373020e-01 3.39470237e-01 -4.35806364e-01
8.33321095e-01 6.57178342e-01 8.09323668e-01 8.17213953e-02
-1.15226972e+00 -3.92734647e-01 -8.60878706e-01 -2.80556474e-02
-9.16004539e-01 -1.85560361e-01 1.12568486e+00 -4.88055736e-01
4.93975431e-01 4.08673972e-01 7.96019197e-01 6.43451273e-01
1.04340601e+00 2.43987963e-01 1.31199539e+00 -7.52558827e-01
-2.16868192e-01 -1.76723495e-01 -3.66167814e-01 6.27002656e-01
2.43627846e-01 6.00083768e-01 -6.72351897e-01 -1.29703462e-01
1.35341668e+00 5.72887957e-01 -6.34644508e-01 -6.79469407e-01
-1.69564497e+00 4.97039258e-01 3.16052258e-01 2.73831725e-01
-2.74976104e-01 1.94897741e-01 1.75324738e-01 1.79670438e-01
1.11608408e-01 8.57166171e-01 2.28677973e-01 1.05886653e-01
-8.80560756e-01 -1.66621312e-01 1.90112159e-01 6.29957557e-01
3.85278076e-01 1.04731927e-02 1.94351390e-01 3.36364776e-01
6.92565693e-03 6.68662608e-01 6.69303834e-01 -1.17682910e+00
2.36206815e-01 2.32602373e-01 -1.84665322e-01 -1.12171781e+00
-6.05684519e-01 -5.67304313e-01 -1.00012290e+00 3.69762689e-01
4.51756120e-01 -1.94642798e-03 -6.03671968e-01 1.23269892e+00
5.05201697e-01 3.33590537e-01 4.29377370e-02 8.43543768e-01
7.02105820e-01 2.40706299e-02 -7.79313266e-01 -3.86465788e-01
1.39537001e+00 -4.54783678e-01 -8.27266812e-01 -1.95945337e-01
4.83107477e-01 -1.14514840e+00 5.20621896e-01 5.34376383e-01
-1.36370766e+00 -1.44952923e-01 -1.23667753e+00 -2.30018273e-01
3.21971834e-01 -1.55603573e-01 6.13742232e-01 6.29075885e-01
-8.93056154e-01 6.36738360e-01 -9.58383501e-01 1.21937841e-01
4.68615294e-01 4.56514686e-01 -7.18982756e-01 -5.23852706e-01
-6.30177379e-01 9.20352638e-01 -7.11540133e-02 -3.01808864e-01
-2.92362332e-01 -1.26779401e+00 -8.87559116e-01 -2.30052754e-01
1.47298172e-01 -5.77104449e-01 1.03589392e+00 -4.55294162e-01
-1.28925240e+00 1.07718623e+00 2.97504067e-01 -4.05667365e-01
7.86342561e-01 -2.14186057e-01 -2.21812725e-01 9.04644132e-01
2.09823400e-02 3.91239434e-01 1.05552757e+00 -1.08324075e+00
-1.26153365e-01 -6.77561522e-01 -2.56771892e-01 4.90342081e-01
-4.52113710e-02 -4.19311434e-01 -1.95232615e-01 -7.22807407e-01
5.71524024e-01 -8.08977008e-01 -2.49381423e-01 4.47823167e-01
-4.30315405e-01 1.06922030e+00 8.98766875e-01 -8.39577258e-01
9.88372028e-01 -2.03466225e+00 -1.00786127e-01 1.62672207e-01
6.28181636e-01 -1.29018262e-01 1.48194999e-01 4.40746665e-01
-4.10880327e-01 -6.83758676e-01 -3.63649160e-01 -2.10210741e-01
-7.93952286e-01 -7.23307952e-02 -3.26513052e-01 1.07943118e+00
-2.92584091e-01 1.00982571e+00 -1.13371646e+00 -4.79012251e-01
8.07615042e-01 5.14180541e-01 -5.32101393e-01 3.68855195e-03
3.77067477e-01 1.08237469e+00 -2.16024220e-01 5.31083226e-01
1.11233854e+00 -2.84987032e-01 3.71863782e-01 -5.70446551e-01
-2.45861456e-01 1.58789173e-01 -5.07500827e-01 2.33061790e+00
-6.69178784e-01 8.16753805e-01 -2.26015560e-02 -9.15349364e-01
2.58944333e-01 3.72168183e-01 1.32822704e+00 -1.00639451e+00
2.51243562e-01 3.14508468e-01 -1.28258960e-02 -5.05751491e-01
2.91287243e-01 -3.75167400e-01 5.59212342e-02 7.12528586e-01
-1.48086548e-01 -8.08895528e-01 -4.47266638e-01 -1.17607698e-01
1.22107625e+00 -2.00115398e-01 4.98386145e-01 -1.21845536e-01
1.07384488e-01 5.06290272e-02 -7.67423809e-02 6.24916553e-01
1.31080464e-01 1.34643912e+00 2.51307070e-01 -6.61615789e-01
-1.35484874e+00 -1.21114433e+00 -5.66184998e-01 -1.67614117e-01
3.97317678e-01 -1.34931758e-01 -5.58994532e-01 -3.96501124e-01
-2.49473467e-01 4.19255465e-01 -4.57611978e-01 -2.74544775e-01
-9.50639248e-01 -4.14536297e-01 3.40126157e-01 -8.31962079e-02
4.25127000e-01 -5.33858359e-01 -1.47058201e+00 1.42831147e-01
-3.32829952e-01 -1.34759879e+00 -5.45535147e-01 3.19763646e-02
-1.14564574e+00 -1.34373379e+00 -1.01789713e+00 -4.91897434e-01
6.55232430e-01 9.02626455e-01 1.11193633e+00 -9.71375629e-02
-9.37056422e-01 5.75322926e-01 -9.00720730e-02 -6.76299632e-02
-5.63626587e-01 -5.30104041e-01 -1.45909667e-01 -1.26531214e-01
-1.80514321e-01 -8.26223612e-01 -1.05770385e+00 2.43866891e-01
-1.13458967e+00 5.38568676e-01 7.53810823e-01 1.08671713e+00
5.44398248e-01 4.44185995e-02 -3.11162323e-01 -7.35686123e-01
3.10199689e-02 -1.67195976e-01 -5.43847978e-01 1.46837890e-01
-2.65944719e-01 -5.90447076e-02 2.74766415e-01 -1.84178367e-01
-8.34807396e-01 -1.05311647e-02 4.09705527e-02 -6.92585170e-01
1.36720240e-01 3.45613897e-01 4.29297030e-01 -4.66155052e-01
7.77020156e-01 4.60564971e-01 6.88705921e-01 2.83597186e-02
4.48683091e-02 1.79486170e-01 9.21955705e-01 -1.54369995e-01
7.13267505e-01 1.38463426e+00 5.00244558e-01 -1.31326187e+00
-6.70908988e-01 -5.44460535e-01 -8.32291305e-01 -3.44342977e-01
1.02297366e+00 -6.58652127e-01 -8.29567194e-01 4.41660583e-01
-1.04742110e+00 -3.87236178e-02 -5.12614667e-01 1.00521171e+00
-7.25054801e-01 8.14941704e-01 -4.71046448e-01 -1.21691592e-01
-2.39632919e-01 -1.34229016e+00 1.50454235e+00 -3.77346128e-02
-1.67614639e-01 -1.09608841e+00 2.99666345e-01 6.63996756e-01
2.20578879e-01 5.10237575e-01 8.07701349e-01 2.15536103e-01
-9.67199564e-01 -2.90986657e-01 1.19513243e-01 2.16088310e-01
5.24740875e-01 -5.87674499e-01 -1.01924145e+00 -3.43254000e-01
7.83193946e-01 -1.14204243e-01 3.98336142e-01 1.14413142e+00
1.02418256e+00 -1.14629686e-01 -2.95160443e-01 1.25545740e+00
1.29353607e+00 1.41987115e-01 8.68613958e-01 2.26683646e-01
7.61270463e-01 6.80765867e-01 4.61677641e-01 3.89711320e-01
-1.29652321e-01 9.43016589e-01 4.24568564e-01 -4.75772381e-01
-4.02861089e-01 -9.19637904e-02 -1.38212159e-01 6.30463064e-01
2.17521042e-01 1.55531093e-01 -8.85279179e-01 1.91569030e-01
-1.16646361e+00 -9.52650189e-01 1.38516560e-01 2.60695553e+00
7.23460853e-01 -2.45388165e-01 -6.04498863e-01 1.53442025e-01
3.58612150e-01 1.31177828e-01 -6.54501259e-01 2.54634738e-01
-2.47037932e-01 3.72711658e-01 9.35499549e-01 7.26146996e-01
-8.56019497e-01 1.24868855e-01 6.15957165e+00 5.68456650e-01
-1.52887225e+00 1.53501570e-01 5.68111479e-01 -2.21428558e-01
-4.20718372e-01 -1.11967221e-01 -4.89449836e-02 2.54892975e-01
3.29375237e-01 -5.09696752e-02 2.07507491e-01 1.33918807e-01
8.99919942e-02 -4.85003233e-01 -1.11522460e+00 1.51287484e+00
4.31637526e-01 -1.83168018e+00 -3.22788149e-01 5.70148766e-01
8.53402555e-01 9.14480910e-02 5.23152590e-01 -8.22177589e-01
-2.39365160e-01 -1.01926196e+00 3.50307494e-01 2.99542964e-01
1.27981377e+00 -5.18238008e-01 2.48356670e-01 3.87151353e-02
-6.67818308e-01 9.85413790e-02 -1.67836472e-02 4.52928156e-01
3.88390124e-01 8.97669733e-01 -1.14534366e+00 7.07881570e-01
3.79822969e-01 8.40586722e-01 -1.89828143e-01 9.43760037e-01
2.40198076e-01 -8.08796212e-02 -3.26917082e-01 9.81604457e-01
-3.38758305e-02 -3.89769137e-01 8.10080826e-01 6.50654614e-01
6.34789050e-01 1.77199662e-01 -2.13207662e-01 4.12963539e-01
2.43252248e-01 -4.66010094e-01 -1.01115823e+00 2.57425338e-01
1.95978388e-01 1.22165918e+00 -8.84601951e-01 -1.40975639e-01
-4.99650300e-01 8.67041826e-01 -5.16535282e-01 1.56162918e-01
-4.52491909e-01 1.88503906e-01 4.37694460e-01 5.79258502e-01
-1.61639806e-02 -3.71861517e-01 -5.31387627e-01 -1.34647226e+00
1.12150632e-01 -8.61925364e-01 2.33634144e-01 -1.03050876e+00
-7.91643560e-01 3.73934507e-01 1.22917399e-01 -1.85225797e+00
-3.79975647e-01 -4.82451111e-01 -5.32213569e-01 6.12597525e-01
-1.50363290e+00 -1.12930381e+00 -5.34090281e-01 6.68386817e-01
2.60798365e-01 -3.75315524e-03 7.12641597e-01 1.96839437e-01
3.38215649e-01 3.26513469e-01 4.38573182e-01 -1.88458953e-02
8.82315636e-01 -9.99835074e-01 1.94924966e-01 6.58208609e-01
1.35488631e-02 5.73030412e-01 6.21579051e-01 -6.63700998e-01
-1.73581851e+00 -5.66127479e-01 1.20955348e-01 -4.48118985e-01
2.69550443e-01 6.34659380e-02 -5.81643820e-01 4.03542519e-01
3.30368936e-01 2.41215050e-01 6.26125872e-01 -6.70726597e-01
-2.33280987e-01 -5.35863526e-02 -1.20361805e+00 5.10075927e-01
4.59280878e-01 -5.82864583e-01 -1.87881410e-01 4.66169298e-01
4.70131636e-01 -1.15962851e+00 -9.17714059e-01 1.93830267e-01
1.03947973e+00 -1.45006013e+00 1.50015569e+00 1.02143265e-01
8.89483988e-01 -3.21396589e-02 -9.66144577e-02 -1.22513461e+00
1.90270782e-01 -9.15506840e-01 2.31293887e-01 5.78878112e-02
-1.88287869e-01 -7.10477471e-01 9.87881124e-01 2.77157873e-01
-4.24069703e-01 -5.39030194e-01 -1.22412288e+00 -3.90440762e-01
-9.52610821e-02 -2.41576150e-01 1.28025904e-01 1.05062485e+00
2.12671265e-01 -1.70479685e-01 -3.76382113e-01 1.81473628e-01
1.07868671e+00 1.75395772e-01 5.24498701e-01 -7.50994742e-01
-6.08671188e-01 -2.88280755e-01 -5.74216008e-01 -9.13893878e-01
-4.70706016e-01 -8.93794119e-01 -4.13305104e-01 -9.66356754e-01
3.55355181e-02 -4.91640687e-01 2.15098590e-01 -1.68936059e-01
3.25638711e-01 5.27951837e-01 -2.16627568e-02 5.37663400e-01
2.54209816e-01 2.57424533e-01 1.77737105e+00 -2.92000022e-05
4.38155308e-02 -1.20932721e-02 -3.67096305e-01 6.67904079e-01
3.96641821e-01 -3.40802759e-01 -6.51493132e-01 -7.15732157e-01
2.64151692e-01 8.04595411e-01 5.50869167e-01 -1.09148264e+00
2.93477952e-01 5.54986745e-02 5.11244953e-01 -6.95662498e-01
7.12164760e-01 -1.05162644e+00 3.78373474e-01 7.53854990e-01
-3.44934165e-01 2.62668490e-01 -1.13721408e-01 4.02881175e-01
-3.08054686e-01 -1.41268387e-01 1.08526707e+00 -2.97719002e-01
2.66112536e-02 2.79603302e-01 1.31826848e-01 3.00452579e-02
9.85977948e-01 -5.71495891e-01 -2.59847850e-01 -4.19601202e-01
-2.17655182e-01 -4.51839179e-01 8.08097422e-01 -5.87316751e-02
1.13000453e+00 -1.14296412e+00 -5.68505347e-01 7.17469394e-01
3.14885490e-02 9.56938192e-02 8.08391213e-01 1.41830802e+00
-1.25938201e+00 3.83973539e-01 -2.80983418e-01 -1.13817930e+00
-1.39020753e+00 4.11018133e-01 5.13159037e-01 -4.36006993e-01
-1.31373870e+00 4.64450806e-01 6.39388382e-01 -2.09791318e-01
-2.12576568e-01 -3.12053829e-01 1.47871345e-01 -2.85967201e-01
7.46665537e-01 -4.20362465e-02 5.39852083e-01 -7.00372875e-01
-1.01784900e-01 1.13442302e+00 -2.00454429e-01 -5.22355556e-01
1.17354000e+00 -3.56367171e-01 1.17780946e-01 1.39102731e-02
1.47094822e+00 1.99510098e-01 -1.46734214e+00 -3.97895545e-01
-5.81212938e-01 -1.09223509e+00 3.18530619e-01 -4.82088059e-01
-1.04499650e+00 1.18122029e+00 7.41137266e-01 -5.43791382e-03
1.18232644e+00 -1.67909518e-01 9.41047192e-01 6.73069805e-02
8.78973678e-02 -6.22658730e-01 4.36218023e-01 -5.13299406e-02
7.58331776e-01 -1.15853536e+00 6.52744472e-01 -4.40773845e-01
-4.65363234e-01 1.34473968e+00 -8.29474330e-02 -7.87928849e-02
6.81368530e-01 5.47376037e-01 2.80832499e-01 -4.27360922e-01
-3.28910828e-01 4.48369890e-01 3.06370378e-01 7.73760021e-01
1.85958415e-01 -1.26699612e-01 1.63966954e-01 -5.27009189e-01
-3.30321908e-01 -1.32844254e-01 9.20594573e-01 1.10917759e+00
5.17738946e-02 -9.18438733e-01 -7.42456198e-01 3.27549875e-01
-5.73230326e-01 -1.69766754e-01 2.78969675e-01 7.20000923e-01
-1.83725819e-01 5.62093139e-01 1.21498205e-01 -1.72110423e-01
1.08391233e-02 -6.83139026e-01 1.05177438e+00 -5.25031924e-01
-3.43205184e-01 3.88807118e-01 -3.52889001e-01 -6.63895547e-01
-4.25441533e-01 -7.19517827e-01 -9.55766678e-01 -5.71571551e-02
8.06459710e-02 -3.59274864e-01 1.05047119e+00 9.22386169e-01
1.04486488e-01 4.26673651e-01 8.79796088e-01 -9.09506202e-01
-2.61210084e-01 -2.93284118e-01 -5.59877872e-01 3.55222523e-01
9.30020690e-01 -5.12209296e-01 -2.34221041e-01 9.46898535e-02] | [13.555532455444336, -2.764758586883545] |
baa4cfaa-41a0-4e2e-8e95-517bb5cadac4 | privacy-preserving-domain-adaptation-of | 2212.10520 | null | https://arxiv.org/abs/2212.10520v3 | https://arxiv.org/pdf/2212.10520v3.pdf | Privacy-Preserving Domain Adaptation of Semantic Parsers | Task-oriented dialogue systems often assist users with personal or confidential matters. For this reason, the developers of such a system are generally prohibited from observing actual usage. So how can they know where the system is failing and needs more training data or new functionality? In this work, we study ways in which realistic user utterances can be generated synthetically, to help increase the linguistic and functional coverage of the system, without compromising the privacy of actual users. To this end, we propose a two-stage Differentially Private (DP) generation method which first generates latent semantic parses, and then generates utterances based on the parses. Our proposed approach improves MAUVE by 2.5$\times$ and parse tree function type overlap by 1.3$\times$ relative to current approaches for private synthetic data generation, improving both on fluency and semantic coverage. We further validate our approach on a realistic domain adaptation task of adding new functionality from private user data to a semantic parser, and show overall gains of 8.5% points in accuracy with the new feature. | ['Jason Eisner', 'Tatsunori Hashimoto', 'Yu Su', 'Richard Shin', 'FatemehSadat Mireshghallah'] | 2022-12-20 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation', 'task-oriented-dialogue-systems'] | ['medical', 'miscellaneous', 'natural-language-processing'] | [ 4.70783204e-01 1.12137926e+00 2.51657128e-01 -7.18441546e-01
-1.15402508e+00 -9.25610542e-01 3.58232796e-01 9.84237641e-02
-1.69098884e-01 1.14546287e+00 1.42142102e-02 -3.18531364e-01
4.84192401e-01 -7.57023752e-01 -5.75377464e-01 -1.48130283e-01
4.18296874e-01 6.18945658e-01 -1.10116318e-01 -3.78793240e-01
-5.59623837e-02 -1.20617583e-01 -1.37555707e+00 5.38034678e-01
1.44054639e+00 5.89260519e-01 8.11619535e-02 6.72704101e-01
-3.21336955e-01 3.82265478e-01 -1.15131283e+00 -7.65192330e-01
3.00138742e-01 -4.86189663e-01 -9.56281722e-01 -6.70392215e-02
1.76594511e-01 -6.34762585e-01 1.15937471e-01 1.06945419e+00
5.99275529e-01 -5.20678386e-02 3.50300789e-01 -1.27213633e+00
-5.35100877e-01 9.21358526e-01 -2.02878281e-01 -4.77584541e-01
7.43409514e-01 1.72287926e-01 8.54351997e-01 -2.21671879e-01
5.16075134e-01 1.18385923e+00 2.95318961e-01 1.10944927e+00
-1.43138349e+00 -7.78300345e-01 -9.59705487e-02 -4.43867296e-01
-1.13380170e+00 -9.08140540e-01 4.96993154e-01 -4.17079449e-01
9.53397930e-01 3.89536142e-01 1.69613380e-02 1.52491176e+00
-3.84080917e-01 8.41045260e-01 1.01732731e+00 -6.27947986e-01
4.39296842e-01 8.75813425e-01 6.92560673e-02 7.62667656e-01
2.41455644e-01 -3.75328839e-01 -3.98709953e-01 -5.71698606e-01
6.71080574e-02 -5.03325760e-01 -1.93620831e-01 -1.96683973e-01
-6.79414928e-01 8.82178187e-01 -4.66513306e-01 1.34543315e-01
-5.58091849e-02 -1.16407253e-01 4.64540511e-01 4.15958971e-01
7.25031972e-01 8.16808760e-01 -8.34018946e-01 -4.59990740e-01
-7.29384065e-01 5.56174636e-01 1.27495253e+00 1.53494799e+00
5.55722177e-01 -2.20817491e-01 -3.16374242e-01 9.56005573e-01
2.86486913e-02 6.21947944e-01 7.26625085e-01 -1.08496702e+00
7.20560849e-01 6.67468011e-01 5.26142716e-01 -4.17129397e-01
-4.96961921e-02 -1.61633305e-02 -1.71670288e-01 -8.35593641e-02
6.58067107e-01 -7.74290442e-01 -4.72190470e-01 2.00614309e+00
1.81969509e-01 -4.32180971e-01 3.65285426e-01 2.48324528e-01
5.10175705e-01 4.60148692e-01 1.18857101e-01 -4.89787422e-02
1.46036220e+00 -6.25028253e-01 -7.49232233e-01 -2.47657627e-01
1.17456877e+00 -5.86117566e-01 1.42768323e+00 8.54023024e-02
-1.07887042e+00 -2.90624589e-01 -8.52892518e-01 8.68271440e-02
-1.73740134e-01 3.77852738e-01 4.81114358e-01 1.52650189e+00
-1.00899053e+00 4.71807599e-01 -6.83399618e-01 -2.90122330e-01
4.49268222e-01 3.04414093e-01 -3.10789913e-01 -3.34789902e-02
-1.39885175e+00 4.81899381e-01 2.14781389e-01 -5.97034335e-01
-4.30392683e-01 -8.07617605e-01 -1.06218624e+00 2.77775317e-01
3.98291975e-01 -6.27969623e-01 1.87312317e+00 -6.97396278e-01
-1.78861535e+00 6.36884332e-01 -2.43504211e-01 -4.23457861e-01
5.98052025e-01 -2.62616128e-01 -2.48249635e-01 -1.19809732e-01
3.87519807e-01 6.41291976e-01 6.11328244e-01 -1.04018736e+00
-7.21443236e-01 -3.95751953e-01 5.16700506e-01 2.88047075e-01
-3.99070203e-01 -1.37025803e-01 -3.73441987e-02 -3.41345817e-01
-4.51316535e-01 -1.01902425e+00 -8.02485049e-02 -2.85714120e-01
-5.69925249e-01 -1.32813513e-01 6.84174955e-01 -1.07180953e+00
9.22568798e-01 -2.27353764e+00 -3.05515409e-01 2.53139157e-02
1.29284188e-01 3.79495561e-01 -3.74149606e-02 3.26520622e-01
1.04825698e-01 4.45809275e-01 -3.58807117e-01 -5.71884394e-01
2.13424891e-01 -3.07235979e-02 -2.26780936e-01 -8.45118761e-02
8.63080993e-02 7.45816290e-01 -8.33511055e-01 -1.38903633e-01
-1.57950267e-01 6.52076676e-02 -6.94570303e-01 4.06304121e-01
-5.66581488e-01 3.58072579e-01 -6.69374824e-01 3.55265588e-01
5.85998058e-01 7.90265724e-02 3.27058643e-01 3.75969917e-01
2.52106756e-01 4.79374915e-01 -7.05882132e-01 1.92302179e+00
-8.24040949e-01 4.00577337e-01 1.46077260e-01 -6.79721296e-01
1.05630338e+00 6.11314416e-01 7.32294545e-02 -5.00480056e-01
3.00779361e-02 1.22773036e-01 -2.34220445e-01 -5.18290579e-01
4.97542262e-01 1.24323461e-02 -5.28343976e-01 8.19448471e-01
6.76149875e-02 -5.96935786e-02 -1.07520573e-01 2.93084681e-01
1.40484345e+00 1.16299897e-01 1.99141607e-01 -1.22079246e-01
3.92349988e-01 -7.87110701e-02 3.35285932e-01 6.44808590e-01
-4.47394736e-02 2.94957697e-01 8.62917721e-01 2.22273678e-01
-9.25759852e-01 -5.97083092e-01 2.54367858e-01 1.09958220e+00
-2.85507560e-01 -2.94646472e-01 -1.54700518e+00 -1.26167881e+00
-6.52475730e-02 1.54465616e+00 -3.05136055e-01 -3.29597414e-01
-3.50352913e-01 -2.20957786e-01 1.04100513e+00 2.11549595e-01
6.65095389e-01 -8.84031475e-01 -6.01179898e-01 2.46179968e-01
-5.91526508e-01 -1.30346048e+00 -4.53057826e-01 -2.55850524e-01
-5.59555650e-01 -7.31499076e-01 -6.72007918e-01 -4.38906729e-01
7.05371380e-01 -4.20222804e-02 6.54247344e-01 -1.80359006e-01
-1.82129607e-01 3.15156966e-01 -5.13902307e-01 -2.52586991e-01
-1.01890349e+00 4.00302112e-01 -8.10647830e-02 -1.17338389e-01
5.80141604e-01 -5.28524399e-01 -3.50839853e-01 1.42484650e-01
-6.78687990e-01 2.82253046e-02 4.05926883e-01 7.78681934e-01
-2.08769113e-01 -2.03163037e-03 9.02249634e-01 -1.77397966e+00
1.13377595e+00 -5.57923377e-01 -3.92003715e-01 3.17864537e-01
-7.38654315e-01 4.42504138e-01 9.13938880e-01 -1.61961354e-02
-1.60583591e+00 2.16130510e-01 -3.03344369e-01 1.31558582e-01
-5.51400781e-01 -2.93954853e-02 -5.57024419e-01 1.74921244e-01
9.86599803e-01 1.66915625e-01 1.76185250e-01 -4.90809649e-01
6.56866193e-01 1.33598518e+00 1.17117368e-01 -7.70893574e-01
5.28169811e-01 -7.55840838e-02 -7.45087802e-01 -6.97384536e-01
-3.87914628e-01 -2.03216627e-01 -2.96952367e-01 3.29018563e-01
5.39964736e-01 -8.55208218e-01 -6.69895351e-01 3.17612767e-01
-1.14184701e+00 -3.33015382e-01 -4.14362937e-01 -4.23565768e-02
-8.23162735e-01 4.27994132e-01 -2.31668204e-01 -9.73280430e-01
-4.04580086e-01 -1.24835813e+00 1.20644557e+00 5.29047884e-02
-7.24063396e-01 -6.31302297e-01 -1.39219701e-01 7.75973082e-01
4.49590594e-01 1.96463749e-01 1.02157938e+00 -9.96651292e-01
-1.77881360e-01 -4.71753776e-01 9.26187541e-03 4.78048950e-01
2.41865322e-01 -5.05836904e-01 -1.19347572e+00 -2.43921131e-01
1.74490914e-01 -4.12842095e-01 7.05532953e-02 -1.98022366e-01
1.10115206e+00 -7.40566671e-01 -3.59260857e-01 1.26864016e-01
8.24291527e-01 4.14768428e-01 4.18441415e-01 -8.60117227e-02
3.24065059e-01 1.13402736e+00 6.51938379e-01 5.93802273e-01
3.59621912e-01 7.61646450e-01 -1.06993988e-01 2.99990296e-01
1.31693870e-01 -6.39842391e-01 5.28524637e-01 1.51032373e-01
4.38215584e-01 -3.88583243e-01 -6.42707050e-01 5.54005802e-01
-1.57258558e+00 -5.40866196e-01 2.25066528e-01 2.24891686e+00
1.13994122e+00 -2.19663954e-03 5.97686954e-02 -2.45479867e-02
8.01851094e-01 -1.62154481e-01 -5.66271067e-01 -5.57964623e-01
4.31600541e-01 4.98976082e-01 5.27214944e-01 4.91173655e-01
-6.72230542e-01 9.41283882e-01 5.45745564e+00 7.05034792e-01
-6.82830095e-01 2.81464487e-01 7.80935109e-01 -6.01568855e-02
-5.33245981e-01 8.42638165e-02 -7.65956759e-01 7.02798247e-01
1.41709661e+00 -5.24572730e-01 4.31927562e-01 1.16658044e+00
4.72444408e-02 -1.77059439e-03 -1.31584501e+00 6.90905690e-01
-1.04343399e-01 -8.91321778e-01 -2.63313442e-01 1.74231514e-01
2.64474303e-01 -5.31549990e-01 4.35118079e-02 5.44981539e-01
6.24922693e-01 -8.84077430e-01 6.10495329e-01 5.14223240e-02
1.26013899e+00 -7.11202443e-01 7.54642963e-01 8.09194982e-01
-5.11299908e-01 8.68682116e-02 -2.70728357e-02 7.98803791e-02
2.21604779e-01 2.86354661e-01 -1.40357101e+00 3.14668804e-01
4.06595618e-01 -8.88734534e-02 -5.06637812e-01 2.02778831e-01
-3.35580409e-01 6.09874964e-01 -2.13067204e-01 -3.71353924e-01
-2.19314769e-01 9.50073153e-02 3.40632915e-01 1.15323699e+00
4.40810472e-01 2.65079290e-01 -2.87971526e-01 1.08508539e+00
-2.43001670e-01 1.40064657e-01 -8.80228460e-01 -3.25807810e-01
6.85198843e-01 1.04720724e+00 -1.16990209e-01 -2.84488916e-01
-2.13821590e-01 1.43217015e+00 1.90158471e-01 2.78077215e-01
-5.60781658e-01 -7.00071871e-01 6.84270978e-01 2.14250267e-01
-9.12215654e-03 7.02609867e-02 -1.88987941e-01 -1.06572080e+00
4.01302338e-01 -1.09415495e+00 1.25026420e-01 -4.91082191e-01
-9.43853498e-01 5.56817353e-01 2.00917292e-02 -8.04235876e-01
-8.03163826e-01 -2.47227162e-01 -1.75126374e-01 1.11083174e+00
-8.38200927e-01 -9.46972728e-01 -1.21783905e-01 2.56553829e-01
7.35140383e-01 -5.23339510e-01 1.44372559e+00 1.05358586e-01
-3.39202315e-01 1.38570166e+00 3.99524644e-02 -8.38478841e-03
6.01693153e-01 -1.32450056e+00 1.05908406e+00 5.76333463e-01
-2.26478830e-01 8.47978234e-01 8.06770980e-01 -6.66989744e-01
-1.00493419e+00 -1.12732458e+00 9.22218978e-01 -6.87441111e-01
1.99845210e-01 -9.71391976e-01 -6.80948615e-01 6.43482089e-01
1.45410284e-01 -4.76909071e-01 9.83171105e-01 5.36686406e-02
-3.25827956e-01 2.72430837e-01 -2.11940408e+00 6.49547994e-01
1.04207039e+00 -4.99978185e-01 -3.53661418e-01 4.60323721e-01
1.19133294e+00 -3.40177923e-01 -7.83993185e-01 -8.03601295e-02
3.71458620e-01 -8.04848850e-01 4.04105783e-01 -5.78296423e-01
2.52020329e-01 1.21112145e-01 -1.94846392e-01 -1.54776430e+00
1.91585526e-01 -1.15154541e+00 -8.99840612e-03 1.48779070e+00
9.08711672e-01 -9.06293511e-01 1.03651047e+00 1.48102522e+00
3.42583843e-02 -3.97946298e-01 -7.89987028e-01 -5.66226184e-01
2.88400557e-02 -3.06695729e-01 9.53243494e-01 8.97581041e-01
4.73875463e-01 4.34751391e-01 -2.48088479e-01 6.10572025e-02
4.54004973e-01 -4.05735254e-01 9.35953557e-01 -6.81975901e-01
-5.93134642e-01 2.79301088e-02 -1.12093478e-01 -1.13192868e+00
4.01202679e-01 -7.54646778e-01 7.15649724e-02 -1.00330520e+00
-1.00855567e-01 -7.10553586e-01 5.04930019e-01 5.70493877e-01
-1.96144462e-01 -3.67237270e-01 -1.32234320e-01 -2.32554242e-01
-1.23878434e-01 5.82016468e-01 5.69816768e-01 9.56940651e-03
-3.72376323e-01 3.24445784e-01 -1.27959204e+00 4.87767875e-01
1.09060681e+00 -4.65073705e-01 -8.21626961e-01 -2.48698428e-01
-2.67396215e-02 3.31974775e-01 -1.31535500e-01 -8.30749094e-01
-1.71566755e-01 9.72730815e-02 -1.93443879e-01 4.88544926e-02
2.35277995e-01 -6.37041330e-01 5.12797758e-02 2.47333065e-01
-6.63950562e-01 -4.36140984e-01 3.37590426e-01 5.05529046e-01
1.32296100e-01 -6.99733853e-01 4.03349876e-01 -3.07728142e-01
-2.33741418e-01 -7.61200264e-02 -4.11475569e-01 4.63064089e-02
1.18921196e+00 -1.06792770e-01 -5.55753291e-01 -7.13180542e-01
-6.58381939e-01 1.96484715e-01 5.31680226e-01 3.58717352e-01
2.52278656e-01 -8.50979388e-01 -5.79511642e-01 3.05120230e-01
3.29511434e-01 -1.05523117e-01 2.30751604e-01 1.18147634e-01
-4.52416122e-01 4.33280528e-01 -4.77127880e-02 -1.25474095e-01
-1.39804506e+00 2.06056699e-01 1.11015275e-01 1.85725291e-03
-5.31781137e-01 9.70845580e-01 2.22364515e-01 -7.86733270e-01
4.97693829e-02 -1.79912195e-01 -1.31006271e-01 -3.18009287e-01
4.63215560e-01 3.17436427e-01 1.53760433e-01 -3.34493369e-01
1.62319243e-02 -3.05673927e-01 -3.63623530e-01 -5.95433295e-01
9.70984995e-01 -1.14488043e-01 1.38585716e-01 1.57749712e-01
1.08989847e+00 3.12588811e-01 -1.04590976e+00 -1.10399529e-01
-1.46580890e-01 -5.93347073e-01 -4.73651171e-01 -1.03012609e+00
-6.56583130e-01 7.45339572e-01 4.99226332e-01 5.78065515e-01
8.86826515e-01 2.63594911e-02 1.10483897e+00 4.22248065e-01
8.40469837e-01 -9.14825857e-01 -3.29471409e-01 9.40290913e-02
6.08622730e-01 -9.26970840e-01 -5.20805478e-01 -7.30868340e-01
-9.15534019e-01 6.90277815e-01 7.08288014e-01 4.45180506e-01
1.21745549e-01 1.74256548e-01 -2.22465191e-02 6.45456016e-02
-6.71541274e-01 3.05678904e-01 -2.83623338e-01 9.94462490e-01
5.55387139e-01 3.69177639e-01 -2.93262839e-01 1.03655159e+00
-6.68560147e-01 -1.75492361e-01 7.94903219e-01 9.66590226e-01
-3.35445404e-01 -1.62903321e+00 -9.94757283e-03 6.66319370e-01
-5.57974637e-01 -1.41759634e-01 -7.14405119e-01 4.51335788e-01
1.13677513e-02 1.14124131e+00 -3.26541930e-01 -5.44170618e-01
5.12001157e-01 6.33463204e-01 1.94441363e-01 -1.15802431e+00
-5.05712807e-01 -4.83468980e-01 8.19549680e-01 -5.19316494e-01
2.21325129e-01 -9.79849279e-01 -1.06954253e+00 -2.29044482e-01
-3.89696389e-01 4.37567085e-01 7.78145969e-01 7.71296203e-01
7.64102280e-01 2.57507831e-01 7.03642130e-01 -4.21170779e-02
-8.95804226e-01 -1.11923277e+00 -4.48675275e-01 3.79673541e-01
-1.21587560e-01 -4.11028773e-01 -2.85656720e-01 1.62397012e-01] | [12.4743013381958, 7.83352518081665] |
9b7e4b2f-3f9c-4adc-8eca-5c8211959403 | dilated-neighborhood-attention-transformer | 2209.15001 | null | https://arxiv.org/abs/2209.15001v3 | https://arxiv.org/pdf/2209.15001v3.pdf | Dilated Neighborhood Attention Transformer | Transformers are quickly becoming one of the most heavily applied deep learning architectures across modalities, domains, and tasks. In vision, on top of ongoing efforts into plain transformers, hierarchical transformers have also gained significant attention, thanks to their performance and easy integration into existing frameworks. These models typically employ localized attention mechanisms, such as the sliding-window Neighborhood Attention (NA) or Swin Transformer's Shifted Window Self Attention. While effective at reducing self attention's quadratic complexity, local attention weakens two of the most desirable properties of self attention: long range inter-dependency modeling, and global receptive field. In this paper, we introduce Dilated Neighborhood Attention (DiNA), a natural, flexible and efficient extension to NA that can capture more global context and expand receptive fields exponentially at no additional cost. NA's local attention and DiNA's sparse global attention complement each other, and therefore we introduce Dilated Neighborhood Attention Transformer (DiNAT), a new hierarchical vision transformer built upon both. DiNAT variants enjoy significant improvements over strong baselines such as NAT, Swin, and ConvNeXt. Our large model is faster and ahead of its Swin counterpart by 1.6% box AP in COCO object detection, 1.4% mask AP in COCO instance segmentation, and 1.4% mIoU in ADE20K semantic segmentation. Paired with new frameworks, our large variant is the new state of the art panoptic segmentation model on COCO (58.5 PQ) and ADE20K (49.4 PQ), and instance segmentation model on Cityscapes (45.1 AP) and ADE20K (35.4 AP) (no extra data). It also matches the state of the art specialized semantic segmentation models on ADE20K (58.1 mIoU), and ranks second on Cityscapes (84.5 mIoU) (no extra data). | ['Humphrey Shi', 'Ali Hassani'] | 2022-09-29 | null | null | null | null | ['panoptic-segmentation'] | ['computer-vision'] | [ 4.58679870e-02 9.10457149e-02 -2.01627195e-01 -1.60103068e-01
-9.86327350e-01 -7.33807385e-01 7.44941175e-01 -8.16839710e-02
-8.07639360e-01 3.65895629e-01 1.33599833e-01 -3.95231664e-01
-7.97474477e-03 -8.28435421e-01 -8.52100492e-01 -5.69188535e-01
2.39864245e-01 6.00644350e-01 8.04642379e-01 -1.96667776e-01
-8.66418630e-02 2.68197268e-01 -1.31861913e+00 3.27684164e-01
1.02665234e+00 1.22301710e+00 3.52195293e-01 7.04092026e-01
-7.51606524e-02 5.91522634e-01 -3.15006793e-01 -4.82667595e-01
2.27142558e-01 -1.11146659e-01 -1.13522470e+00 -1.41593397e-01
9.40168858e-01 -4.31423001e-02 -8.68716910e-02 9.98242497e-01
5.08208036e-01 1.54159173e-01 5.33154666e-01 -1.19869757e+00
-9.70957637e-01 6.74217463e-01 -8.10756385e-01 4.11807269e-01
-2.24505663e-01 2.78171659e-01 1.34150410e+00 -9.80309188e-01
4.08799708e-01 1.41434097e+00 1.05218613e+00 4.68136489e-01
-1.24842608e+00 -6.14410818e-01 4.78086442e-01 1.25338122e-01
-1.15539587e+00 -2.74719775e-01 1.07847162e-01 -2.18041778e-01
1.34920442e+00 2.14077681e-01 5.89534640e-01 1.15858889e+00
-1.20776802e-01 1.21880615e+00 1.15091789e+00 -1.72613442e-01
2.75756791e-02 -1.04717627e-01 2.88893163e-01 7.04960465e-01
1.52144972e-02 -1.88081965e-01 -2.23733276e-01 2.10475236e-01
8.05413783e-01 -8.55674073e-02 -1.38621867e-01 8.29921588e-02
-1.14497554e+00 7.95831919e-01 8.99727404e-01 2.55924076e-01
-1.34530276e-01 3.91009629e-01 3.43950242e-01 -8.48024711e-02
5.34574449e-01 3.73879313e-01 -7.51230478e-01 -8.09160918e-02
-8.45878184e-01 1.74196497e-01 3.85114759e-01 9.53921318e-01
7.80230641e-01 8.04139823e-02 -3.15340906e-01 1.13226831e+00
1.66166261e-01 7.32762218e-01 5.11959374e-01 -1.05357254e+00
3.93865883e-01 5.87723553e-01 -2.54660249e-01 -5.66121340e-01
-5.15491366e-01 -6.51672661e-01 -8.29045236e-01 2.24632975e-02
4.73537117e-01 -6.54900074e-03 -1.63151944e+00 1.72763884e+00
8.00164193e-02 2.10815921e-01 -5.18648028e-02 6.85110569e-01
1.01862812e+00 7.65502274e-01 3.95712346e-01 4.98046964e-01
1.56201565e+00 -1.46791995e+00 -2.38106430e-01 -5.94694376e-01
3.55558664e-01 -7.54747450e-01 1.28106952e+00 2.36309230e-01
-1.15041828e+00 -7.80597508e-01 -6.29452169e-01 -5.23791313e-01
-6.51790738e-01 -5.70605956e-02 7.37016499e-01 4.84840304e-01
-1.55351961e+00 2.82422543e-01 -8.40056121e-01 -7.00058937e-01
7.76081502e-01 4.63330448e-01 -1.80182502e-01 -1.96125537e-01
-1.10296404e+00 7.26816058e-01 2.43786886e-01 9.25133973e-02
-9.17982459e-01 -8.23047936e-01 -9.76420164e-01 1.69562310e-01
5.84370553e-01 -7.68864036e-01 1.45827425e+00 -1.05368221e+00
-1.31120431e+00 9.70117807e-01 -9.49174240e-02 -7.94047534e-01
3.56293857e-01 -4.64221090e-01 -2.57095486e-01 2.00515822e-01
5.20073593e-01 1.46761644e+00 7.97584713e-01 -9.37919378e-01
-1.02863967e+00 -3.71252775e-01 2.93521345e-01 1.80371419e-01
-1.64261520e-01 -1.24708980e-01 -9.60177541e-01 -7.51762152e-01
-8.72144997e-02 -8.81226361e-01 -3.20418835e-01 -2.06005946e-01
-3.91827643e-01 -6.03699446e-01 9.12632763e-01 -4.97731149e-01
8.92573416e-01 -2.07340050e+00 4.51712757e-02 7.09518464e-03
2.77701855e-01 4.19282615e-01 -4.57253695e-01 -1.04812153e-01
8.02184045e-02 2.37178758e-01 -5.78650713e-01 -6.04720354e-01
3.41853336e-03 4.32514220e-01 -1.06604919e-01 7.57250413e-02
2.98590958e-01 1.34431982e+00 -8.96170080e-01 -3.82548213e-01
2.94841021e-01 4.99892473e-01 -6.35250926e-01 -2.50544637e-01
-2.42896006e-01 8.50312337e-02 -1.90299913e-01 9.07131135e-01
5.87396562e-01 -4.23751116e-01 -3.11893851e-01 -2.65712053e-01
-2.60947078e-01 1.39388219e-01 -7.56116986e-01 1.60147572e+00
-5.36534548e-01 6.06307566e-01 2.90765226e-01 -8.90929282e-01
6.24935150e-01 1.31634444e-01 1.96193993e-01 -1.09457302e+00
-2.50810944e-02 2.55140752e-01 -1.46910116e-01 -2.43829072e-01
5.12776852e-01 9.93207693e-02 -6.95503801e-02 -1.98136363e-02
2.95948029e-01 -8.37026983e-02 2.75139689e-01 2.88776904e-01
9.60105479e-01 3.99997793e-02 -4.20218408e-02 -4.93699938e-01
3.31502080e-01 2.73803286e-02 6.49606705e-01 1.04570115e+00
-4.16051328e-01 9.60029066e-01 4.23466831e-01 -3.62368435e-01
-8.55550945e-01 -1.14519835e+00 -1.95868850e-01 1.41243923e+00
1.35057658e-01 -3.64504009e-01 -8.61900747e-01 -9.55832422e-01
-6.43508732e-02 3.76031101e-01 -8.35393548e-01 1.50549278e-01
-6.69164896e-01 -8.56566966e-01 8.45385194e-01 1.08088470e+00
1.09172499e+00 -1.36050677e+00 -3.58171850e-01 1.41790897e-01
-2.93027639e-01 -1.24268186e+00 -6.75784886e-01 2.43280157e-01
-6.84897482e-01 -8.68782580e-01 -1.10630441e+00 -7.68631160e-01
2.53773719e-01 3.41399610e-01 1.31688964e+00 -3.35044265e-01
-2.62462735e-01 4.79186922e-01 -3.30924898e-01 -4.07888979e-01
9.63367224e-02 4.73823786e-01 -3.50374788e-01 4.22913954e-03
3.12154502e-01 -3.17090482e-01 -7.27106452e-01 4.01948631e-01
-8.71169090e-01 -7.01038819e-03 6.66298151e-01 9.31201518e-01
7.56037891e-01 -2.82203227e-01 4.59083289e-01 -1.01676631e+00
2.32132450e-01 -4.70748514e-01 -6.35556400e-01 1.27792150e-01
-5.42599142e-01 -1.48969471e-01 5.43353617e-01 -2.19655499e-01
-1.16895950e+00 -1.60527498e-01 -4.14676458e-01 -5.25410473e-01
-2.65862882e-01 3.62032086e-01 -7.34181181e-02 9.73731056e-02
6.06839895e-01 -7.82419555e-03 -3.97369862e-01 -6.50654435e-01
5.35833061e-01 4.54820246e-01 7.67531455e-01 -5.42461693e-01
4.39708263e-01 6.22859895e-01 -5.09671211e-01 -8.82717192e-01
-1.08180630e+00 -6.98655128e-01 -4.45218891e-01 3.49981993e-01
1.36949897e+00 -9.14802313e-01 -5.72630167e-01 8.12300563e-01
-8.33392560e-01 -7.78332293e-01 -4.11394715e-01 1.44889072e-01
-1.81447566e-01 2.00794697e-01 -7.95544982e-01 -4.47209030e-01
-4.52663511e-01 -1.28740406e+00 1.43410158e+00 4.82306808e-01
-1.18340701e-01 -9.60219324e-01 -2.39004001e-01 5.82674503e-01
5.55894017e-01 -6.96131885e-02 6.98186159e-01 -3.86496305e-01
-5.53467989e-01 2.01646939e-01 -8.69344652e-01 5.76078057e-01
-8.30588490e-02 -4.93817925e-02 -1.26256239e+00 -1.79544613e-01
-4.94575024e-01 -3.14269334e-01 1.42772317e+00 8.62367690e-01
1.25489092e+00 8.03449079e-02 -4.02951747e-01 9.17502165e-01
1.25383246e+00 2.36861706e-01 5.59641123e-01 5.18276274e-01
9.63150978e-01 3.36958706e-01 1.87780857e-01 -1.40833914e-01
6.72465205e-01 5.53258240e-01 6.10009015e-01 -5.63490868e-01
-4.55517322e-01 -1.00991257e-01 3.48149240e-01 4.75498199e-01
-4.10787255e-01 -2.01948419e-01 -1.10055912e+00 9.43389118e-01
-1.89365411e+00 -7.32037783e-01 -2.24624038e-01 1.79836309e+00
4.53828961e-01 3.52160037e-01 2.30331168e-01 -1.51805893e-01
5.44135034e-01 2.79153973e-01 -5.98859251e-01 -5.79330564e-01
-3.48040551e-01 7.06548333e-01 7.53457844e-01 5.39317727e-01
-1.44280875e+00 1.38738668e+00 5.82325220e+00 1.13482666e+00
-1.00152040e+00 2.58119553e-01 8.88749063e-01 1.25019804e-01
-1.05401665e-01 -2.25793958e-01 -1.02773654e+00 3.26040417e-01
7.52946496e-01 4.03852791e-01 2.92643368e-01 8.69331121e-01
-1.74976885e-01 -1.46616369e-01 -8.53617728e-01 9.31345105e-01
-1.10523984e-01 -1.39894283e+00 2.51838323e-02 8.56006294e-02
8.08242142e-01 7.13695705e-01 4.05465692e-01 6.53185725e-01
5.06176949e-01 -1.21457148e+00 7.45209813e-01 -7.67757073e-02
8.89680922e-01 -6.42273903e-01 8.87844741e-01 -1.20865006e-03
-1.43263090e+00 -2.07455546e-01 -2.67059594e-01 9.99371111e-02
4.64783609e-02 3.55620593e-01 -3.92486572e-01 3.46001685e-01
1.36308432e+00 9.11140144e-01 -7.09684253e-01 8.57321799e-01
-3.60679388e-01 9.39577341e-01 -5.84222674e-01 3.37916106e-01
1.11378825e+00 -1.62550770e-02 4.70943481e-01 1.55140686e+00
1.70176942e-02 2.29817983e-02 2.56442338e-01 7.88955390e-01
-3.54845971e-01 -1.41809672e-01 -2.20444232e-01 2.64264256e-01
2.06734955e-01 1.25604463e+00 -9.94564235e-01 -6.22213542e-01
-5.85409701e-01 1.07808626e+00 3.90984446e-01 5.53231180e-01
-1.00031579e+00 -2.54630417e-01 7.45387435e-01 -1.28845721e-01
6.72659814e-01 1.21650480e-01 -3.28416526e-01 -9.64886546e-01
-1.54214963e-01 -7.53925383e-01 7.84459591e-01 -7.07441270e-01
-1.41282070e+00 7.65715778e-01 -4.95026968e-02 -6.35714650e-01
3.63420039e-01 -7.69821644e-01 -5.79051375e-01 8.61935198e-01
-1.72955453e+00 -1.63303399e+00 -2.40412414e-01 8.20270479e-01
9.46991205e-01 1.62913918e-01 6.65246546e-01 4.05645341e-01
-8.13060582e-01 6.94436789e-01 -1.06803581e-01 2.94337839e-01
5.64097703e-01 -1.65662146e+00 8.33393991e-01 7.91904390e-01
2.73135185e-01 3.48517179e-01 1.25924841e-01 -2.60781765e-01
-6.94922864e-01 -1.36547983e+00 8.32946599e-01 -5.17722845e-01
7.77201414e-01 -3.50404084e-01 -9.51780498e-01 1.05326009e+00
3.47160816e-01 3.76301706e-01 1.01331770e-01 3.30545485e-01
-6.18338346e-01 -2.87753880e-01 -1.01905715e+00 5.58097124e-01
1.07726502e+00 -4.47356701e-01 -6.52325809e-01 2.01284513e-01
9.92728472e-01 -3.81492704e-01 -6.40869915e-01 4.82108831e-01
5.02676249e-01 -1.04561794e+00 1.13421261e+00 -1.72931075e-01
3.09523821e-01 -1.42088979e-01 -1.52812734e-01 -1.10646749e+00
-6.86873734e-01 -5.52491307e-01 2.42235899e-01 1.32885003e+00
6.54669464e-01 -8.26511800e-01 8.27761173e-01 3.59018862e-01
-5.71765125e-01 -8.36949289e-01 -8.41065705e-01 -7.57832348e-01
2.74030417e-01 -6.06176496e-01 4.10810441e-01 8.88388097e-01
-5.55024505e-01 5.57080448e-01 -3.72950807e-02 1.74130484e-01
5.06370723e-01 4.65755118e-03 4.00393903e-01 -1.01823282e+00
-3.24429899e-01 -8.15455198e-01 -1.53406262e-01 -1.36540043e+00
-1.13996640e-01 -8.09252381e-01 -8.41805562e-02 -1.75089097e+00
2.38291383e-01 -5.96239626e-01 -5.62568605e-01 8.51989746e-01
-3.29033434e-01 7.74951756e-01 2.94241846e-01 4.03387919e-02
-6.75524712e-01 4.05286580e-01 1.14997065e+00 -3.19961131e-01
-2.53867537e-01 -9.44277495e-02 -7.50704825e-01 1.04551435e+00
7.84557223e-01 -1.68176115e-01 -4.21994776e-01 -7.91354954e-01
-4.47961725e-02 -5.47473669e-01 5.41183054e-01 -9.54676509e-01
1.75792933e-01 1.68094158e-01 2.91123152e-01 -5.74037135e-01
3.02291274e-01 -5.44114113e-01 -4.13793266e-01 3.17778885e-01
-3.93958539e-02 1.72924548e-01 5.34357846e-01 4.50245649e-01
-3.88914138e-01 2.23326340e-01 9.46775317e-01 -4.11810249e-01
-1.24174166e+00 5.92019916e-01 -2.96425581e-01 4.28705275e-01
7.21405506e-01 -4.00050163e-01 -4.58808422e-01 5.14734676e-03
-9.69957650e-01 5.16615748e-01 4.91470173e-02 5.22803307e-01
4.10333365e-01 -9.05337155e-01 -5.97083986e-01 1.10232271e-01
1.04867287e-01 5.13246775e-01 3.78823787e-01 9.93928194e-01
-4.14681703e-01 5.13525248e-01 -4.72716987e-02 -9.30194438e-01
-1.05991590e+00 2.39619896e-01 3.83681387e-01 -4.63651240e-01
-7.55075693e-01 1.47110331e+00 8.85924220e-01 -5.25353253e-01
1.52770132e-01 -7.96599627e-01 -1.41620308e-01 8.85180235e-02
3.76824439e-01 4.93282288e-01 1.19009912e-01 -4.82919306e-01
-4.73054081e-01 7.75681973e-01 -2.73807853e-01 -3.10769165e-03
1.29380107e+00 -9.11026895e-02 -9.72339958e-02 1.18256874e-01
9.41825807e-01 -1.77171960e-01 -1.45943272e+00 -1.71742395e-01
-1.05009191e-02 9.07723443e-04 1.60246447e-01 -1.16182303e+00
-1.21712124e+00 1.05273581e+00 5.25012553e-01 2.06053138e-01
1.22847509e+00 3.38637680e-01 1.18353021e+00 1.08010871e-02
9.67038944e-02 -1.03324139e+00 -1.00298785e-01 7.66472578e-01
7.07172990e-01 -1.30761671e+00 -4.09804434e-01 -4.32419837e-01
-5.96333742e-01 6.06329501e-01 8.24719012e-01 -1.81947947e-01
4.60389644e-01 3.05073768e-01 1.08674511e-01 -2.51526624e-01
-5.39149046e-01 -6.76612616e-01 3.55871350e-01 6.94924116e-01
2.88018197e-01 1.92088723e-01 2.43293986e-01 5.12229502e-01
-4.41200025e-02 -2.42502823e-01 -9.09729744e-04 4.58162278e-01
-3.96090835e-01 -8.14784586e-01 -2.71999508e-01 3.58982116e-01
-5.76910734e-01 -5.43880224e-01 -2.57093579e-01 1.06658578e+00
5.35101414e-01 8.64562690e-01 2.56847024e-01 5.55945337e-02
4.87621099e-01 -1.48846090e-01 2.65352428e-01 -5.80864727e-01
-8.78584862e-01 4.13542926e-01 1.73473526e-02 -7.78937280e-01
-4.59605724e-01 -4.74985182e-01 -1.24127007e+00 -2.11479947e-01
-1.41179964e-01 -1.50430771e-02 2.80187935e-01 9.23000693e-01
3.90053779e-01 6.29396021e-01 -8.84905532e-02 -8.97076249e-01
-9.21029970e-02 -9.11165297e-01 -3.63155752e-01 3.16479027e-01
3.70265841e-01 -4.67507064e-01 -1.30129591e-01 -1.07922973e-02] | [9.510196685791016, 0.5610043406486511] |
8dd63f80-59e2-4ccf-8821-086353e14825 | road-damage-detection-acquisition-system | 1909.08991 | null | https://arxiv.org/abs/1909.08991v1 | https://arxiv.org/pdf/1909.08991v1.pdf | Road Damage Detection Acquisition System based on Deep Neural Networks for Physical Asset Management | Research on damage detection of road surfaces has been an active area of re-search, but most studies have focused so far on the detection of the presence of damages. However, in real-world scenarios, road managers need to clearly understand the type of damage and its extent in order to take effective action in advance or to allocate the necessary resources. Moreover, currently there are few uniform and openly available road damage datasets, leading to a lack of a common benchmark for road damage detection. Such dataset could be used in a great variety of applications; herein, it is intended to serve as the acquisition component of a physical asset management tool which can aid governments agencies for planning purposes, or by infrastructure mainte-nance companies. In this paper, we make two contributions to address these issues. First, we present a large-scale road damage dataset, which includes a more balanced and representative set of damages. This dataset is composed of 18,034 road damage images captured with a smartphone, with 45,435 in-stances road surface damages. Second, we trained different types of object detection methods, both traditional (an LBP-cascaded classifier) and deep learning-based, specifically, MobileNet and RetinaNet, which are amenable for embedded and mobile and implementations with an acceptable perfor-mance for many applications. We compare the accuracy and inference time of all these models with others in the state of the art. | ['G Ochoa-Ruiz', 'L. M. Aguilar-Lobo', 'J. A. Vega-Fernández', 'S. Natraj', 'A. A. Angulo'] | 2019-09-19 | null | null | null | null | ['road-damage-detection'] | ['computer-vision'] | [ 2.78100163e-01 -2.27439269e-01 -5.70084229e-02 1.40293479e-01
-6.48324251e-01 -1.54994261e-02 2.55054444e-01 3.03120632e-02
-2.70955205e-01 8.63970101e-01 5.94746061e-02 -5.08491457e-01
-2.39260688e-01 -1.64007854e+00 -5.37797570e-01 -7.75888264e-01
-3.05116642e-03 -4.37174849e-02 6.45345390e-01 -4.34910566e-01
4.55451846e-01 1.01324809e+00 -2.32085681e+00 5.89106558e-03
1.13009226e+00 1.15291822e+00 2.74326444e-01 6.33939028e-01
3.78613025e-01 5.75787485e-01 -6.59756184e-01 -3.04297358e-01
-4.29258347e-02 1.53062090e-01 -6.85300052e-01 -3.38372469e-01
4.40224200e-01 -7.52622187e-01 -6.08662665e-01 7.93986440e-01
8.84482324e-01 -2.56657111e-03 7.70748556e-01 -9.91431475e-01
-2.68602967e-01 -8.01683590e-02 -4.80855584e-01 6.63697004e-01
1.26316592e-01 2.14898244e-01 6.83709562e-01 -7.76506305e-01
5.41692711e-02 1.19806850e+00 6.72459304e-01 1.09659739e-01
-6.75654173e-01 -3.90089571e-01 -1.98542938e-01 7.55690277e-01
-1.34673667e+00 -3.93444419e-01 8.79572868e-01 -5.14349878e-01
9.96088207e-01 4.02661115e-01 4.76484239e-01 8.81084800e-01
6.97508976e-02 5.68685174e-01 1.13828659e+00 -2.87420094e-01
1.56250060e-01 -1.98253915e-01 6.83247074e-02 5.33957243e-01
5.98623991e-01 3.74953210e-01 -2.17568427e-01 -1.05109811e-01
5.73197365e-01 -8.31072107e-02 -2.79038399e-01 1.31062925e-01
-5.66071749e-01 6.23929143e-01 6.35995269e-01 2.70563960e-01
-7.18078256e-01 1.65308565e-01 2.81804740e-01 4.05615941e-02
6.36776507e-01 -2.53803402e-01 -5.83046712e-02 -3.19475383e-01
-5.50257266e-01 1.64245501e-01 4.89092261e-01 1.74976796e-01
9.59527731e-01 1.10091269e-01 4.55322955e-03 1.00530076e+00
2.28328690e-01 9.48360562e-01 -1.03800960e-01 -7.01788843e-01
5.23541629e-01 5.34721255e-01 2.44578216e-02 -1.35697734e+00
-3.81258756e-01 6.81271106e-02 -1.02689910e+00 5.30407548e-01
2.42799789e-01 -1.40249953e-01 -6.92801476e-01 1.09522629e+00
2.94973701e-01 3.49781245e-01 -1.76237360e-01 7.14977384e-01
6.33955300e-01 5.45439184e-01 -1.03883065e-01 3.74027729e-01
1.20654213e+00 -4.64841187e-01 -4.36038941e-01 -2.52064884e-01
5.90171993e-01 -5.86427033e-01 1.20031309e+00 4.99972790e-01
-8.27630877e-01 -4.98043060e-01 -9.49796438e-01 -2.12878846e-02
-7.11187065e-01 1.88379556e-01 3.27124417e-01 9.03392196e-01
-9.57510710e-01 5.23622394e-01 -5.48038483e-01 -6.85555100e-01
7.70498991e-01 -4.77943989e-03 -2.60658655e-02 -2.53698170e-01
-1.32130659e+00 1.38844979e+00 -9.70938355e-02 4.48811024e-01
-1.00253069e+00 -4.17146951e-01 -5.99268079e-01 -4.55755256e-02
2.11710125e-01 -3.12674612e-01 8.71642411e-01 -1.62387282e-01
-1.20927548e+00 8.38208258e-01 2.54681706e-02 -4.79471147e-01
3.34051847e-01 -3.21273297e-01 -6.20796680e-01 1.27288297e-01
-5.53599298e-02 2.86436766e-01 7.18240023e-01 -1.17192185e+00
-1.01592970e+00 -3.57903481e-01 4.15757447e-01 1.30064577e-01
-3.93950433e-01 1.62273925e-02 9.68960598e-02 -3.58296573e-01
-4.00431037e-01 -7.06672728e-01 5.24641760e-02 1.45094439e-01
-4.49769497e-01 2.81620417e-02 1.18510532e+00 -9.35625792e-01
1.57980990e+00 -1.93355036e+00 -4.12997901e-01 2.36805320e-01
-3.34099121e-03 7.71074891e-01 1.52342632e-01 7.23203659e-01
8.27012137e-02 1.43669859e-01 -6.02443874e-01 5.34681454e-02
-2.20378608e-01 3.94913822e-01 -4.85757649e-01 5.21414042e-01
2.24638566e-01 6.15871072e-01 -6.05955184e-01 -1.98821962e-01
6.07243419e-01 4.05610770e-01 2.19806321e-02 1.07064225e-01
4.96005714e-01 -9.37598944e-03 -2.86979318e-01 1.05899107e+00
7.04673767e-01 5.20143092e-01 -6.91586196e-01 -7.94484317e-02
-3.45149666e-01 3.37698489e-01 -1.10286868e+00 6.98416054e-01
-6.12770081e-01 8.66534531e-01 1.09244943e-01 -1.17681146e+00
1.00735438e+00 1.61232144e-01 4.26755428e-01 -6.39495850e-01
1.51559949e-01 3.34806055e-01 -2.27200791e-01 -9.55845237e-01
5.06455064e-01 2.08939821e-01 1.16076238e-01 2.62830734e-01
-6.92023754e-01 -9.56089795e-02 -7.95740075e-03 -3.05352449e-01
1.42626739e+00 -4.48501319e-01 -2.00213585e-02 -2.31013689e-02
5.99458337e-01 -5.74886166e-02 4.87062223e-02 6.13242209e-01
-3.66065204e-01 5.52346110e-01 3.56400520e-01 -5.01653671e-01
-7.53403425e-01 -8.64761114e-01 -3.78575742e-01 6.21101737e-01
2.45353997e-01 4.16697592e-01 -9.14826393e-01 -4.45637554e-01
2.62645274e-01 4.98266876e-01 -4.26809877e-01 -2.27723703e-01
-5.50833106e-01 -1.07161939e+00 1.05593431e+00 5.79647243e-01
9.88408208e-01 -1.27823949e+00 -9.83187377e-01 1.22962289e-01
-3.66758823e-01 -8.27922642e-01 1.62950307e-01 -1.55108616e-01
-8.76985550e-01 -1.44118679e+00 -4.82665271e-01 -4.86170262e-01
4.42676008e-01 8.87068033e-01 8.62104297e-01 5.32966971e-01
-4.46755767e-01 3.84398818e-01 -2.51563907e-01 -6.02988243e-01
-8.14167187e-02 1.34632766e-01 -1.58841923e-01 6.85344264e-02
2.84008652e-01 -5.32809138e-01 -7.32902408e-01 6.32740080e-01
-7.65361428e-01 -5.79190791e-01 7.48821855e-01 4.07826453e-01
3.13417077e-01 3.70921701e-01 8.48660648e-01 -3.43621552e-01
7.45559335e-01 -6.25926971e-01 -4.36763436e-01 -1.41877960e-02
-5.20887613e-01 -5.66338003e-01 8.13386217e-02 6.08957410e-02
-1.13085473e+00 -3.81545722e-01 -6.89632058e-01 1.77269414e-01
-6.25379503e-01 7.22165525e-01 -3.90212864e-01 -2.90368378e-01
7.94028938e-01 -4.70181108e-02 -7.86484405e-02 -5.85684121e-01
1.78507686e-01 1.19116592e+00 7.41817236e-01 -4.91326302e-01
7.28032649e-01 6.87471390e-01 6.17531165e-02 -1.43261921e+00
-5.17124832e-01 -6.80208027e-01 -6.31509066e-01 -6.02108359e-01
5.79559386e-01 -7.57524490e-01 -4.37661797e-01 1.36210537e+00
-1.16487443e+00 -5.06838918e-01 -1.21555634e-01 5.99323884e-02
-3.72229695e-01 5.29840052e-01 -3.94378960e-01 -9.40054655e-01
-3.69881302e-01 -7.47343481e-01 1.06130171e+00 5.03204875e-02
2.16484651e-01 -7.27269053e-01 1.08815454e-01 4.99183774e-01
6.81407213e-01 3.89879078e-01 8.45660865e-01 1.79847643e-01
-5.31447232e-01 -4.04157788e-01 -5.76118946e-01 6.12760544e-01
3.13397199e-01 2.75129467e-01 -1.25382888e+00 -2.24610977e-02
-2.55451530e-01 -7.70277902e-02 1.05694854e+00 4.40589309e-01
1.22269893e+00 -8.50752071e-02 -4.36636865e-01 6.77776039e-02
1.24714363e+00 8.87216702e-02 1.62674069e+00 4.52702314e-01
5.87320387e-01 9.27656531e-01 7.19102740e-01 2.74661928e-01
6.31790221e-01 6.87717080e-01 9.92518187e-01 -3.02448332e-01
-4.44278300e-01 -4.28488925e-02 4.14751440e-01 3.83500189e-01
-5.22869051e-01 -4.08245713e-01 -1.16607237e+00 9.03485477e-01
-1.67163086e+00 -1.32566047e+00 -7.51636505e-01 2.23494101e+00
4.09797430e-01 1.77383557e-01 2.21919253e-01 8.46452057e-01
9.97350156e-01 1.57656997e-01 -4.91542488e-01 -3.15644920e-01
-2.01336369e-01 9.63299572e-02 7.46846318e-01 2.05926627e-01
-1.18989623e+00 5.67629695e-01 6.48720360e+00 1.03739834e+00
-1.12329412e+00 6.41467944e-02 4.34188306e-01 4.36643124e-01
8.80824998e-02 -2.57825732e-01 -7.94201374e-01 6.09719694e-01
1.05165994e+00 2.71282822e-01 2.49834388e-01 7.08265424e-01
6.26952827e-01 -7.92517245e-01 -2.64080614e-01 7.55881250e-01
-1.41934648e-01 -1.08405280e+00 -4.00553256e-01 3.67782265e-01
5.30971527e-01 3.43871206e-01 -1.44427180e-01 3.14701088e-02
-1.18646257e-01 -8.01169336e-01 4.81890738e-01 7.71937966e-01
7.32534826e-01 -8.39019775e-01 9.55014944e-01 5.04623890e-01
-1.08594739e+00 -2.73431212e-01 -4.67615008e-01 -5.89077234e-01
4.06604528e-01 1.04098964e+00 -5.52476108e-01 4.45492774e-01
1.11499560e+00 5.63424647e-01 -4.77636039e-01 1.41613281e+00
-4.70538616e-01 8.93450916e-01 -4.25637603e-01 9.35705900e-02
-4.55758199e-02 7.21948370e-02 3.66339892e-01 9.59758043e-01
5.63838124e-01 3.95205840e-02 -1.15203522e-01 3.40466648e-01
1.56514600e-01 -3.37667793e-01 -9.69842792e-01 5.72425544e-01
6.59353316e-01 1.17925346e+00 -4.18164551e-01 -1.10960081e-01
-5.11797786e-01 4.61128145e-01 1.14636853e-01 1.30659595e-01
-9.25137699e-01 -7.10214734e-01 1.09219670e+00 6.89951181e-01
2.83766270e-01 -3.62248838e-01 -5.65140903e-01 -5.87784171e-01
4.34438661e-02 -5.28453052e-01 1.05911739e-01 -8.56194496e-01
-1.20464146e+00 1.90202594e-01 1.93730637e-01 -1.11746025e+00
2.84698009e-01 -6.57120287e-01 -9.39421773e-01 8.34171414e-01
-2.20870829e+00 -1.11505711e+00 -7.66908705e-01 3.58001590e-01
2.18444273e-01 1.83480114e-01 5.77074945e-01 8.74221861e-01
-9.12298322e-01 1.73472196e-01 4.22656424e-02 1.36236101e-01
5.03791392e-01 -8.50901723e-01 4.60565239e-01 1.06969535e+00
-2.99105793e-01 -3.22249010e-02 3.21166307e-01 -7.78391421e-01
-8.99731696e-01 -1.30695093e+00 7.85471261e-01 -5.10705352e-01
5.45736909e-01 1.28100082e-01 -1.18426263e+00 1.02195114e-01
-7.08480179e-02 -7.85500631e-02 1.62635475e-01 -2.30553657e-01
2.93741524e-01 -4.04172212e-01 -1.34289706e+00 2.65384525e-01
1.16264188e+00 -5.29025495e-01 -3.43771577e-01 5.63771464e-02
1.24495931e-01 -3.89587253e-01 -6.84455216e-01 6.01054847e-01
4.20029551e-01 -1.27214003e+00 1.03933883e+00 3.23629677e-02
3.70907575e-01 -3.78839254e-01 -1.01391003e-01 -1.07901609e+00
-1.98908791e-01 -3.06230169e-02 -3.40974212e-01 1.25411582e+00
5.14562055e-02 -9.54672992e-01 6.08897269e-01 3.57295841e-01
-6.76974714e-01 -1.03522646e+00 -1.17629194e+00 -8.15974534e-01
-6.89674169e-03 -8.07734549e-01 7.96437383e-01 3.16824019e-01
-5.85424423e-01 -1.26112565e-01 -2.19955161e-01 5.51482737e-01
3.76677364e-01 -2.60053933e-01 8.03718686e-01 -1.36985457e+00
6.18098676e-01 -5.68761647e-01 -6.52877092e-01 -9.23024893e-01
-1.90301642e-01 -1.68148771e-01 1.67810097e-01 -1.83493519e+00
-2.71482736e-01 -7.25681067e-01 -1.53043792e-01 7.44705617e-01
-4.65201512e-02 3.58492851e-01 -5.03398418e-01 3.05010587e-01
1.73125416e-01 5.83789170e-01 9.67262506e-01 -3.33094478e-01
-2.28908733e-02 2.66795725e-01 -6.05207801e-01 8.70691299e-01
8.94479334e-01 -3.20845246e-01 -2.12557614e-01 -4.31271106e-01
2.08075121e-01 -4.23336297e-01 1.08935785e+00 -1.44787133e+00
1.50643274e-01 -2.31942922e-01 -2.98926800e-01 -9.60394561e-01
1.55461043e-01 -7.31509030e-01 -2.83056386e-02 5.00924230e-01
4.22861934e-01 -1.62734538e-01 1.29125372e-01 5.25176287e-01
-3.81845176e-01 -2.59847283e-01 8.28356504e-01 3.02969754e-01
-1.10141802e+00 2.42654309e-01 -7.30220795e-01 -1.42110020e-01
1.22549093e+00 -6.79487228e-01 -7.68921137e-01 -2.86896378e-01
-6.44866601e-02 4.58083153e-02 3.69000375e-01 3.96693707e-01
7.88794518e-01 -1.20768809e+00 -7.22115874e-01 5.33761084e-02
1.41765133e-01 5.96010834e-02 5.95243037e-01 8.89301181e-01
-6.16737366e-01 2.28625789e-01 -3.27124357e-01 -4.28639978e-01
-1.05020404e+00 2.70978004e-01 3.05566281e-01 -2.41213106e-02
-4.73771036e-01 3.77078652e-01 -2.68721074e-01 -1.90706953e-01
5.35298660e-02 -4.62527543e-01 -5.59509575e-01 2.62957484e-01
5.86033642e-01 1.27108371e+00 6.27350509e-01 -9.56449211e-01
-3.02353263e-01 7.01225758e-01 6.32283807e-01 2.75034875e-01
1.23370683e+00 -3.80007476e-01 -1.71660990e-01 7.87120089e-02
7.34537482e-01 -2.04158977e-01 -1.13274014e+00 1.04481712e-01
-1.00450076e-01 -5.30102074e-01 4.14095998e-01 -7.10359037e-01
-1.16703498e+00 1.17451572e+00 9.64054763e-01 5.23885250e-01
1.29214489e+00 -1.55505329e-01 1.17583132e+00 5.21741092e-01
6.69205308e-01 -1.12923884e+00 6.45039091e-03 5.21847188e-01
1.06342065e+00 -1.11322844e+00 -1.24307536e-01 -5.44150949e-01
-3.49615753e-01 9.01093721e-01 5.76404750e-01 -3.80322188e-02
8.28094244e-01 1.44406587e-01 7.66691228e-04 -3.99608225e-01
-7.83563778e-02 -7.36890733e-01 1.21809289e-01 1.02498007e+00
-1.74898669e-01 2.79623847e-02 -2.23962158e-01 2.03414038e-01
1.64303616e-01 -6.69492334e-02 5.72476506e-01 9.75602329e-01
-1.10785508e+00 -9.96089160e-01 -7.28422821e-01 7.45351732e-01
-1.55446947e-01 1.58887535e-01 -2.13270307e-01 7.06576645e-01
4.20226365e-01 1.20614421e+00 -2.56579846e-01 -4.85878229e-01
9.50468481e-01 -4.65129942e-01 1.43007278e-01 -5.50748050e-01
-9.67699289e-02 -6.69144094e-01 4.88653988e-01 -4.63902801e-01
-5.03786445e-01 -8.19935679e-01 -8.60246480e-01 -6.75061166e-01
-5.77957094e-01 -4.94519264e-01 5.64118683e-01 9.82996047e-01
5.61576366e-01 3.21129709e-01 7.78713763e-01 -1.11375463e+00
-2.08837301e-01 -9.73143041e-01 -7.26151586e-01 1.67122204e-02
2.55692095e-01 -1.19840801e+00 -3.04546118e-01 -9.51757208e-02] | [7.419291973114014, 1.129409909248352] |
e3b9eff5-fc60-41bb-bfd5-15a84a135d6a | counterfactual-explanation-with-multi-agent | 2103.12983 | null | https://arxiv.org/abs/2103.12983v2 | https://arxiv.org/pdf/2103.12983v2.pdf | Counterfactual Explanation with Multi-Agent Reinforcement Learning for Drug Target Prediction | Motivation: Many high-performance DTA models have been proposed, but they are mostly black-box and thus lack human interpretability. Explainable AI (XAI) can make DTA models more trustworthy, and can also enable scientists to distill biological knowledge from the models. Counterfactual explanation is one popular approach to explaining the behaviour of a deep neural network, which works by systematically answering the question "How would the model output change if the inputs were changed in this way?". Most counterfactual explanation methods only operate on single input data. It remains an open problem how to extend counterfactual-based XAI methods to DTA models, which have two inputs, one for drug and one for target, that also happen to be discrete in nature. Methods: We propose a multi-agent reinforcement learning framework, Multi-Agent Counterfactual Drug target binding Affinity (MACDA), to generate counterfactual explanations for the drug-protein complex. Our proposed framework provides human-interpretable counterfactual instances while optimizing both the input drug and target for counterfactual generation at the same time. Results: We benchmark the proposed MACDA framework using the Davis dataset and find that our framework produces more parsimonious explanations with no loss in explanation validity, as measured by encoding similarity and QED. We then present a case study involving ABL1 and Nilotinib to demonstrate how MACDA can explain the behaviour of a DTA model in the underlying substructure interaction between inputs in its prediction, revealing mechanisms that align with prior domain knowledge. | ['Truyen Tran', 'Thin Nguyen', 'Thomas P Quinn', 'Tri Minh Nguyen'] | 2021-03-24 | null | null | null | null | ['counterfactual-explanation'] | ['miscellaneous'] | [ 5.27661681e-01 9.17580605e-01 -5.59044659e-01 -2.00895295e-01
-3.04668933e-01 -5.79620361e-01 7.47469902e-01 4.89301234e-02
-2.52064727e-02 1.60770977e+00 2.51964480e-01 -1.00144553e+00
-5.57349920e-01 -6.49433672e-01 -1.32049692e+00 -8.29164505e-01
-1.01232007e-01 9.24289286e-01 -4.36399341e-01 -1.89557657e-01
2.56332576e-01 3.96667153e-01 -1.21349967e+00 5.50663233e-01
1.01268566e+00 3.33568722e-01 2.41215862e-02 4.53673214e-01
3.65084074e-02 7.62007177e-01 -5.44246376e-01 -4.88257855e-01
2.68642992e-01 -7.84869850e-01 -1.19721329e+00 -4.33862716e-01
-6.24188334e-02 -2.17784002e-01 -1.15839794e-01 7.41903782e-01
3.94730330e-01 -2.68729091e-01 9.40885425e-01 -1.47369456e+00
-8.49765718e-01 9.20651138e-01 -4.78294492e-01 6.00624643e-02
1.17451906e-01 7.06720889e-01 1.28746283e+00 -7.58789107e-02
5.98949134e-01 1.63358724e+00 3.36813062e-01 1.15425205e+00
-1.54066205e+00 -7.09167182e-01 2.43676841e-01 2.67004609e-01
-6.81905985e-01 3.15686576e-02 6.31542206e-01 -1.37951866e-01
1.23651755e+00 5.90877533e-01 7.79282391e-01 1.39335752e+00
8.61660182e-01 7.68916845e-01 1.26296628e+00 -2.71915376e-01
6.70621872e-01 -2.05740288e-01 -1.39989600e-01 5.86392939e-01
6.35432541e-01 6.02537453e-01 -4.42850828e-01 -6.26092255e-01
6.41311884e-01 1.21799506e-01 -2.99762338e-01 -4.75318104e-01
-1.17620838e+00 1.20525181e+00 4.10326838e-01 2.84848474e-02
-8.24362218e-01 4.82452571e-01 2.56720304e-01 1.37350857e-01
1.51177803e-02 1.19533610e+00 -1.10205019e+00 3.61169934e-01
-3.98553163e-01 7.31848180e-01 5.48339963e-01 3.87905866e-01
4.30560082e-01 6.32323399e-02 -2.13481724e-01 1.27057403e-01
3.79759848e-01 4.26277608e-01 7.29971051e-01 -9.54345047e-01
1.00017838e-01 5.68445206e-01 2.99589902e-01 -4.35750484e-01
-4.82136399e-01 -4.14722204e-01 -6.78636074e-01 3.92596424e-01
4.57942247e-01 -2.82096088e-01 -1.06284308e+00 2.18213582e+00
3.10146570e-01 3.39854330e-01 4.67883945e-01 9.36680377e-01
4.69194293e-01 4.65799719e-01 5.01204550e-01 -4.43384826e-01
1.25794649e+00 -5.31555295e-01 -7.10696280e-01 -2.64300793e-01
8.84262145e-01 5.97835286e-03 8.38464975e-01 4.35654134e-01
-9.88783598e-01 -1.37055526e-02 -1.07235134e+00 2.70793796e-01
-3.65381449e-01 -5.38833737e-01 1.18692255e+00 5.45089185e-01
-5.63224554e-01 8.56886208e-01 -5.66798329e-01 1.71637349e-02
8.14014435e-01 8.17712724e-01 -1.92092076e-01 1.50120184e-01
-1.56717277e+00 1.02744961e+00 6.49996042e-01 -1.55430913e-01
-1.32661140e+00 -1.01145744e+00 -4.49616551e-01 2.71763593e-01
4.60909396e-01 -1.35523462e+00 1.07603180e+00 -1.27050316e+00
-1.26091003e+00 5.00010371e-01 -1.45052448e-01 -1.14501667e+00
4.11216140e-01 2.88253605e-01 -2.63052344e-01 -2.95748919e-01
1.58610269e-01 1.14006186e+00 5.03099978e-01 -1.11355948e+00
-3.02992761e-01 -6.27674639e-01 3.01353455e-01 2.09670991e-01
4.78279084e-01 -6.69538021e-01 6.04877591e-01 -5.18821955e-01
-2.92403549e-01 -9.40793395e-01 -4.75497693e-01 -1.64066210e-01
-7.68015325e-01 -2.06578776e-01 6.51751757e-01 -2.06515729e-01
8.63141954e-01 -1.50269926e+00 2.20608473e-01 -6.34627370e-03
3.97137374e-01 1.66872129e-01 -1.72880054e-01 1.80250034e-01
-9.31195855e-01 6.89469516e-01 -4.67881173e-01 4.51814622e-01
2.67541409e-02 5.30296147e-01 -6.46663666e-01 1.35371432e-01
3.53904456e-01 1.26086807e+00 -9.11156118e-01 -1.16103619e-01
-9.93695930e-02 6.12201877e-02 -8.23184431e-01 1.72249340e-02
-8.27086627e-01 4.66232359e-01 -5.15474856e-01 2.05023006e-01
7.27648497e-01 -2.23470807e-01 7.29869068e-01 1.82783417e-03
2.07996339e-01 3.53084236e-01 -7.83998728e-01 1.30764449e+00
-8.44756067e-02 2.84683347e-01 -8.10093164e-01 -1.16525316e+00
5.73864520e-01 4.95425045e-01 3.12196404e-01 -6.24647200e-01
1.58802882e-01 3.06558877e-01 9.12108123e-01 -3.38657498e-01
-2.99054384e-01 -8.69970500e-01 1.18734099e-01 7.32068181e-01
-3.03514630e-01 -3.83076072e-03 -2.72307545e-01 8.71552378e-02
1.19333267e+00 1.84597507e-01 8.96485448e-01 -3.25728893e-01
2.64659852e-01 3.05166841e-01 7.82379389e-01 9.63773131e-01
-1.88196033e-01 2.53951550e-01 9.41514075e-01 -9.80155468e-01
-1.20743418e+00 -7.81394601e-01 -5.20069338e-02 2.23716959e-01
-2.29172725e-02 3.30594271e-01 -8.89211774e-01 -1.03267097e+00
2.45275304e-01 1.42779183e+00 -1.13886571e+00 -7.29932606e-01
-4.01228517e-01 -1.24308538e+00 6.51575387e-01 2.40562901e-01
3.53382349e-01 -1.21069610e+00 -8.83276165e-01 2.78223723e-01
-6.80169761e-02 -2.80769438e-01 -9.51313749e-02 4.00626957e-01
-1.09169042e+00 -1.28460658e+00 -4.11599755e-01 8.28094408e-02
3.64640534e-01 -1.71495184e-01 8.73407125e-01 1.80318668e-01
-2.25458264e-01 -3.92051905e-01 1.61542147e-01 -1.03606927e+00
-5.97110569e-01 -3.03589493e-01 3.64505142e-01 -3.63887995e-01
5.57669580e-01 -4.41020340e-01 -8.27157319e-01 1.69305146e-01
-1.10154307e+00 4.70636219e-01 7.51884699e-01 1.27970755e+00
5.48488200e-01 -1.40007421e-01 1.06373596e+00 -1.29668307e+00
7.30839014e-01 -7.68105686e-01 -2.32167631e-01 4.32592630e-01
-1.05186713e+00 7.03763068e-01 8.84144664e-01 -4.92909670e-01
-1.11570263e+00 2.96287332e-02 -1.30253527e-02 -1.82221308e-01
-2.53810316e-01 3.34790200e-01 -4.03131962e-01 4.70829248e-01
9.13793206e-01 2.31067404e-01 1.08154565e-01 -2.16218874e-01
4.18930531e-01 2.32407197e-01 1.04798160e-01 -5.92347562e-01
2.97007769e-01 5.76339006e-01 3.51270735e-01 -8.78790617e-02
-6.87161922e-01 4.76763129e-01 -2.62816370e-01 3.85090113e-01
8.36226523e-01 -4.87763613e-01 -1.30147934e+00 -1.35463685e-01
-1.38969409e+00 -1.96079671e-01 -3.45325977e-01 4.56277072e-01
-1.02540004e+00 -4.07240093e-02 -4.86733625e-03 -7.05901504e-01
-2.15117753e-01 -1.32859325e+00 7.95462787e-01 3.48611474e-02
-5.43548107e-01 -1.06705701e+00 2.23042578e-01 2.81738222e-01
5.84222674e-02 5.10549724e-01 1.67967427e+00 -1.09414852e+00
-6.47660673e-01 2.84625083e-01 1.41397379e-02 -4.84140784e-01
2.21772835e-01 -3.08316350e-01 -1.01009619e+00 1.04650073e-01
-6.11848570e-02 -1.07798569e-01 6.41521871e-01 8.53484452e-01
1.18943179e+00 -9.30527329e-01 -6.06811881e-01 1.28617659e-01
1.20891273e+00 7.58261144e-01 7.64133573e-01 3.26593667e-01
3.43672186e-01 4.24699306e-01 6.30738139e-01 3.04101586e-01
4.47115190e-02 7.05844641e-01 9.53359246e-01 -2.38750398e-01
2.19492078e-01 -3.44828069e-01 1.03609130e-01 -4.20900553e-01
-3.52804810e-02 -3.45616877e-01 -9.04247582e-01 4.78356779e-01
-2.00638318e+00 -1.18497002e+00 -2.56096959e-01 1.94393551e+00
9.53948379e-01 1.50102779e-01 2.13256046e-01 3.35179195e-02
4.26605463e-01 -4.61934566e-01 -1.36789525e+00 -9.59282160e-01
-3.47999752e-01 2.00071946e-01 5.23898721e-01 6.80714965e-01
-3.83400291e-01 7.53558040e-01 6.47385168e+00 6.61843836e-01
-1.12073314e+00 -8.00490379e-02 1.16622341e+00 -1.41558379e-01
-9.85345781e-01 2.51160979e-01 -2.98511118e-01 3.14136922e-01
1.35607541e+00 -4.70640838e-01 2.89628625e-01 5.35716832e-01
8.06836188e-01 1.30199417e-01 -1.60422480e+00 4.68797177e-01
-4.94904131e-01 -2.04952502e+00 7.25629330e-01 3.76143992e-01
6.19118333e-01 -3.63211751e-01 1.31779253e-01 8.93458575e-02
6.78567231e-01 -1.38679564e+00 5.06696820e-01 4.84139532e-01
4.75039423e-01 -8.05349052e-01 7.46165097e-01 5.45149386e-01
-1.36259139e-01 -2.29533583e-01 -3.52201164e-01 -2.91502684e-01
-2.94898927e-01 2.41262242e-01 -1.50629437e+00 4.85394895e-01
1.72247306e-01 3.18934023e-01 -5.00664413e-02 6.73120439e-01
-2.25478262e-01 5.73132157e-01 9.64405611e-02 -2.20361993e-01
3.31009597e-01 7.61254579e-02 5.12791574e-01 6.23243451e-01
-2.98927799e-02 2.87397981e-01 -5.47325134e-01 1.41109645e+00
5.15453331e-02 -1.51300237e-01 -9.84330118e-01 -7.45982975e-02
3.11533153e-01 4.87583667e-01 -4.54137474e-01 -4.20951903e-01
2.06231609e-01 9.33082819e-01 1.14832520e-01 4.19421405e-01
-1.09614336e+00 1.74172103e-01 9.18429673e-01 -2.78247092e-02
-9.40622389e-02 6.79500937e-01 -5.38555920e-01 -9.35689807e-01
-4.59028572e-01 -1.40852809e+00 5.73239863e-01 -9.30821955e-01
-1.06566310e+00 2.59873509e-01 8.84694830e-02 -8.32912207e-01
-5.61863780e-01 -7.68055379e-01 -5.08023739e-01 9.94223058e-01
-1.31309664e+00 -9.49859321e-01 7.25268662e-01 1.76785544e-01
6.05574131e-01 -9.82288793e-02 1.03540564e+00 -3.30899984e-01
-5.74916601e-01 4.09846157e-01 1.12848513e-01 -5.56413054e-01
1.96755856e-01 -1.34436655e+00 4.24696863e-01 3.46595377e-01
-2.48227529e-02 9.11827803e-01 1.25967109e+00 -8.42159212e-01
-1.31089807e+00 -8.80525112e-01 7.04622746e-01 -6.49069250e-01
3.41831446e-01 -9.07356516e-02 -8.55457604e-01 6.41457856e-01
2.09999666e-01 -4.54196870e-01 7.93722093e-01 1.17360968e-02
-2.37072960e-01 2.23542511e-01 -1.58207452e+00 9.72498834e-01
9.06877458e-01 2.36916039e-02 -8.07989657e-01 4.73124117e-01
9.29851711e-01 -5.91164604e-02 -3.30695212e-01 1.37312680e-01
5.90527773e-01 -8.27951431e-01 9.59839463e-01 -1.79698765e+00
7.63588965e-01 -4.18221354e-01 4.29462316e-03 -1.55319238e+00
-2.12249637e-01 -6.71651185e-01 -1.01270407e-01 4.60958540e-01
8.77235770e-01 -8.00097227e-01 8.44771326e-01 9.63258564e-01
1.15557611e-01 -1.08604848e+00 -1.16289043e+00 -6.78053021e-01
4.51031029e-01 -3.34744632e-01 1.16440451e+00 9.76361990e-01
2.67459959e-01 5.31391740e-01 -5.41092157e-01 1.88782200e-01
4.92328912e-01 1.83962896e-01 6.03403330e-01 -1.14113557e+00
-6.85048819e-01 -3.19840819e-01 -1.85219780e-01 -5.10639012e-01
2.65364826e-01 -9.71643806e-01 -3.60298663e-01 -1.23881233e+00
5.58618307e-01 -1.17809922e-01 -8.91388729e-02 7.74566114e-01
-2.38454580e-01 -2.86399662e-01 -1.05530456e-01 1.86919615e-01
-4.63813636e-03 5.74428022e-01 1.25211942e+00 -3.31410110e-01
-1.38434367e-02 -2.09129363e-01 -1.24800730e+00 7.78979659e-01
7.90809155e-01 -8.73322785e-01 -6.72273815e-01 -1.53685987e-01
2.98357546e-01 3.75699252e-01 4.75343734e-01 -1.64333209e-01
-3.53766531e-01 -8.71327102e-01 4.68364954e-01 5.20682484e-02
8.49355087e-02 -8.31355989e-01 6.81943417e-01 1.20206082e+00
-7.15958655e-01 -2.28316151e-02 4.27156955e-01 8.30443978e-01
2.97611386e-01 -1.89375356e-01 7.29684114e-01 -3.23519498e-01
-2.45068565e-01 1.83266401e-01 -4.54617471e-01 -2.33117178e-01
1.16105771e+00 -2.46795848e-01 -5.67939401e-01 -4.32789803e-01
-7.64104724e-01 1.52145162e-01 1.73630163e-01 2.59546816e-01
6.87811613e-01 -1.10773170e+00 -6.40562832e-01 -2.10835189e-01
2.61002071e-02 -4.57993269e-01 2.33871892e-01 4.53585416e-01
-3.84201974e-01 9.21997666e-01 -4.64769036e-01 -2.03808606e-01
-9.52286780e-01 9.07787502e-01 8.64533722e-01 -4.40842301e-01
-1.86806157e-01 4.25202578e-01 8.37455750e-01 -5.11922002e-01
-4.65156853e-01 -2.33023316e-01 -8.88335630e-02 -4.13501620e-01
3.64982903e-01 -3.58454771e-02 -2.86474884e-01 -6.43331707e-02
-4.11560118e-01 -3.26096639e-02 -4.07835990e-01 -6.10235184e-02
1.45820320e+00 1.56144425e-01 2.15238601e-01 1.91515431e-01
6.66459978e-01 -5.42386234e-01 -9.71604764e-01 2.93542325e-01
5.02471849e-02 -2.81295061e-01 -2.02856913e-01 -1.50910950e+00
-5.27947485e-01 9.11420286e-01 7.01907277e-01 8.72304142e-02
8.71798813e-01 -5.30496836e-02 3.40705186e-01 5.82749248e-01
1.58809442e-02 -6.24729455e-01 -1.66194826e-01 -1.00132130e-01
1.10627913e+00 -1.17970729e+00 1.08677983e-01 -1.07594065e-01
-7.51844823e-01 8.71179819e-01 7.00539052e-01 9.66568366e-02
1.47321612e-01 -2.59199202e-01 -2.37664044e-01 -5.30795097e-01
-1.31708837e+00 2.42132619e-01 1.13341950e-01 6.15009189e-01
4.64510173e-01 3.02826464e-01 -6.94818318e-01 7.20508575e-01
-1.00731552e-01 1.16513826e-01 7.57571042e-01 3.79495323e-01
-1.50249884e-01 -1.27726650e+00 -2.84368575e-01 5.63293636e-01
-5.00127852e-01 -2.96539456e-01 -7.91500092e-01 1.01617396e+00
2.06923664e-01 5.69436550e-01 -2.35069662e-01 2.10738964e-02
1.43507361e-01 1.89322099e-01 4.23066109e-01 -4.76654381e-01
-3.92471135e-01 -8.55486989e-02 6.41242415e-02 -3.86179358e-01
-3.02417219e-01 -6.48343682e-01 -1.79386747e+00 -4.04192746e-01
-2.05517530e-01 3.72624248e-01 4.17335957e-01 1.36714184e+00
6.92471504e-01 6.26725078e-01 3.71339828e-01 -8.97902250e-02
-5.88887155e-01 -6.89236999e-01 -3.23254943e-01 3.85619193e-01
5.55496752e-01 -7.20883369e-01 -1.35430723e-01 -7.89720416e-02] | [8.570538520812988, 5.69355583190918] |
57b8e37b-f50c-448d-8167-8bfe9617541b | submanifold-sparse-convolutional-networks | 1706.01307 | null | http://arxiv.org/abs/1706.01307v1 | http://arxiv.org/pdf/1706.01307v1.pdf | Submanifold Sparse Convolutional Networks | Convolutional network are the de-facto standard for analysing spatio-temporal
data such as images, videos, 3D shapes, etc. Whilst some of this data is
naturally dense (for instance, photos), many other data sources are inherently
sparse. Examples include pen-strokes forming on a piece of paper, or (colored)
3D point clouds that were obtained using a LiDAR scanner or RGB-D camera.
Standard "dense" implementations of convolutional networks are very inefficient
when applied on such sparse data. We introduce a sparse convolutional operation
tailored to processing sparse data that differs from prior work on sparse
convolutional networks in that it operates strictly on submanifolds, rather
than "dilating" the observation with every layer in the network. Our empirical
analysis of the resulting submanifold sparse convolutional networks shows that
they perform on par with state-of-the-art methods whilst requiring
substantially less computation. | ['Laurens van der Maaten', 'Benjamin Graham'] | 2017-06-05 | null | null | null | null | ['3d-part-segmentation'] | ['computer-vision'] | [ 3.10735106e-01 -2.04750076e-02 1.30353704e-01 -1.99861407e-01
2.70679444e-02 -5.72403014e-01 9.59335268e-01 -3.73348862e-01
-3.32261980e-01 4.85030949e-01 2.86764175e-01 -4.32681859e-01
-2.58611858e-01 -9.89707947e-01 -8.99124503e-01 -5.46361566e-01
-4.16254848e-01 5.57009876e-01 1.92973673e-01 -1.10122792e-01
8.02411735e-02 1.24840438e+00 -1.63291478e+00 3.69912982e-01
-2.72196624e-02 1.00992560e+00 -5.35676777e-02 8.16250503e-01
-2.75426865e-01 7.93749630e-01 -2.00819075e-01 -8.34854394e-02
5.50007880e-01 -6.45182058e-02 -8.17170024e-01 2.67670214e-01
8.69777024e-01 -2.56840557e-01 -8.34922731e-01 9.96566117e-01
3.72234322e-02 3.29764709e-02 5.60494006e-01 -1.07613635e+00
-7.08655953e-01 2.36253798e-01 -4.36681062e-01 1.67151377e-01
3.54588091e-01 -8.45460966e-02 6.11666143e-01 -1.10066545e+00
1.00320256e+00 1.24428773e+00 1.14208543e+00 1.87564537e-01
-1.33184206e+00 -2.37112626e-01 -1.55992910e-01 -6.27007708e-02
-1.37646925e+00 -3.66821051e-01 7.81358659e-01 -5.52596927e-01
9.71704364e-01 3.70717883e-01 6.19252801e-01 9.67263281e-01
-5.58972508e-02 5.08001328e-01 9.74428594e-01 -2.34013155e-01
2.01876506e-01 -2.28494763e-01 -3.19841713e-01 7.07215428e-01
2.70219445e-01 2.79837161e-01 -4.43518758e-01 -1.88161388e-01
1.44665086e+00 4.95693654e-01 -2.76437635e-03 -7.05808282e-01
-1.61739421e+00 6.93460047e-01 6.85952425e-01 4.70333487e-01
-5.33161640e-01 3.54322404e-01 1.23023808e-01 4.81075794e-01
4.86007661e-01 -7.82415718e-02 -2.86354125e-01 -8.33397359e-02
-1.32870865e+00 1.88566551e-01 8.65307510e-01 1.04015756e+00
1.11732399e+00 2.35923216e-01 6.85550869e-01 3.55064273e-01
4.85049225e-02 3.68124068e-01 1.15149967e-01 -1.14147174e+00
3.33477467e-01 6.73146248e-01 -4.13278267e-02 -1.16074634e+00
-5.06638944e-01 -1.54544279e-01 -1.57779205e+00 7.36901164e-01
2.68885136e-01 -1.30273387e-01 -1.08417463e+00 1.08686757e+00
7.45764608e-03 4.68003362e-01 -1.96510419e-01 9.40125287e-01
8.20054114e-01 5.93825877e-01 -4.85981107e-01 1.03024177e-01
7.62944162e-01 -2.89406627e-01 -3.44311208e-01 -2.55075400e-03
4.55449939e-01 -5.95478415e-01 6.44342124e-01 4.04112101e-01
-1.12219858e+00 -4.36027765e-01 -8.56015861e-01 -1.86384201e-01
-6.51506305e-01 -8.62931609e-02 7.64898837e-01 1.60582989e-01
-1.21329105e+00 9.22216237e-01 -8.59412253e-01 -5.31647503e-01
8.03802848e-01 3.83466780e-01 -8.09509218e-01 -9.53189954e-02
-6.04962766e-01 7.82798588e-01 -1.10108014e-02 1.89832166e-01
-8.12415004e-01 -8.95622730e-01 -1.00667644e+00 -2.92886738e-02
1.11523576e-01 -5.52919745e-01 9.91285801e-01 -8.67370129e-01
-1.05224109e+00 1.04450846e+00 4.51791994e-02 -5.67212284e-01
4.55033153e-01 -1.46946236e-01 -2.15840936e-01 1.56779155e-01
-1.83298051e-01 7.10691988e-01 1.04660249e+00 -1.18172586e+00
-2.38050610e-01 -2.23867312e-01 3.38335991e-01 -9.47643444e-02
1.16418861e-01 7.29872957e-02 3.73128653e-02 -4.66902584e-01
3.09536725e-01 -8.31859231e-01 -5.41622460e-01 3.67743939e-01
-4.54656422e-01 1.76173244e-02 1.62218499e+00 -2.10869089e-01
5.01303375e-01 -2.03882742e+00 3.43409032e-01 4.53276128e-01
3.23360145e-01 2.57771641e-01 -9.65030864e-02 6.13780618e-01
-2.41001084e-01 1.02812462e-01 -5.94998062e-01 -3.57704371e-01
-1.81594431e-01 4.24694866e-01 -2.71705598e-01 9.33979750e-01
4.86554146e-01 8.97346258e-01 -8.70586693e-01 -2.41054639e-01
6.65364623e-01 8.27515006e-01 -3.38505030e-01 -5.50116822e-02
-1.27091274e-01 3.23944539e-01 -1.22412160e-01 8.68449271e-01
5.28661728e-01 -5.01764596e-01 6.16089534e-03 -7.33244792e-02
-4.08170551e-01 1.06985264e-01 -1.43466282e+00 2.12646461e+00
-2.33140841e-01 9.92935359e-01 4.69574302e-01 -1.11322856e+00
7.37514615e-01 2.80678958e-01 7.27282345e-01 -4.00405347e-01
1.40155256e-01 1.41578004e-01 -1.76852256e-01 -2.71920085e-01
4.70714182e-01 -2.08969280e-01 2.37239540e-01 4.76029038e-01
-2.12353393e-02 -3.80349308e-01 -1.43598631e-01 3.47446024e-01
1.26760817e+00 2.45901927e-01 8.05835053e-02 -3.80380899e-01
3.33681375e-01 2.01445654e-01 1.19383343e-01 4.27977771e-01
2.09794700e-01 1.17229986e+00 4.64607149e-01 -1.12678993e+00
-1.34446645e+00 -8.56461585e-01 -2.60147631e-01 4.61033374e-01
-1.07312329e-01 -3.51493508e-01 -5.08421779e-01 -3.18683416e-01
1.62373632e-01 -3.57317813e-02 -7.84456193e-01 5.44258893e-01
-6.93338454e-01 -3.42681706e-01 4.51275826e-01 6.46135092e-01
4.52825189e-01 -1.41603088e+00 -7.31510460e-01 8.51558149e-02
4.89133805e-01 -1.14605856e+00 1.42672151e-01 3.39936107e-01
-1.11076951e+00 -1.24128473e+00 -5.38172781e-01 -7.75674939e-01
8.51252615e-01 4.76042479e-01 1.24332070e+00 3.73271406e-01
-4.45835531e-01 4.80717838e-01 -1.78212658e-01 -4.02914584e-01
1.05650567e-01 6.54500425e-02 1.60032898e-01 6.43824339e-02
5.70903540e-01 -1.08407164e+00 -3.09765249e-01 -1.56907275e-01
-1.33467793e+00 -5.44623472e-02 6.02955282e-01 6.51132941e-01
7.12965846e-01 1.42063230e-01 1.84692070e-02 -9.97393787e-01
3.64952683e-01 -7.32980609e-01 -5.49580812e-01 -4.28144693e-01
2.04811513e-01 -2.91048944e-01 5.71157813e-01 -3.80429655e-01
-4.53041762e-01 4.69469547e-01 -1.68554768e-01 -9.47324157e-01
-7.11801827e-01 5.17502069e-01 3.21856409e-01 -4.00706202e-01
7.53805816e-01 1.89750511e-02 3.16977412e-01 -6.95980608e-01
4.12841618e-01 3.38010758e-01 7.94583142e-01 -2.83297002e-01
1.31387126e+00 1.09434557e+00 5.54955602e-01 -1.41009319e+00
-5.50378919e-01 -4.44251925e-01 -1.29028749e+00 7.43244663e-02
7.75975227e-01 -8.58033895e-01 -5.77011943e-01 4.09464568e-01
-1.16030872e+00 -5.67211449e-01 -6.72144473e-01 2.93786675e-01
-7.00802445e-01 1.81956455e-01 -4.92104173e-01 -4.75304663e-01
2.08374903e-01 -6.31684244e-01 1.37360084e+00 -7.84614608e-02
-1.97608188e-01 -1.05424583e+00 2.53217310e-01 -2.97109425e-01
5.96006691e-01 7.25387812e-01 7.21874118e-01 -1.29808381e-01
-8.57539415e-01 -4.41500783e-01 -4.10144180e-01 4.05551016e-01
1.87469479e-02 3.29788737e-02 -9.86208558e-01 -2.25166783e-01
-1.72400415e-01 -4.04448211e-01 7.91966558e-01 3.61993909e-01
1.49812531e+00 -3.82255852e-01 -1.42630190e-01 1.05399263e+00
1.57051075e+00 -3.84390116e-01 9.36072528e-01 2.50326276e-01
1.10189521e+00 4.99532342e-01 -1.89396262e-01 2.48037174e-01
1.28355607e-01 2.55950779e-01 8.02521825e-01 -4.29264277e-01
-1.44743085e-01 -3.21337208e-02 7.74046630e-02 8.22022855e-01
-5.52970171e-01 2.68354446e-01 -9.04370487e-01 7.55354226e-01
-1.78490710e+00 -1.23126996e+00 -3.27799320e-01 1.90323293e+00
2.63238221e-01 -7.73201659e-02 1.15558475e-01 4.52849835e-01
3.35414648e-01 4.92560506e-01 -4.16223496e-01 -5.07177472e-01
-3.64544868e-01 6.17145061e-01 6.72733128e-01 3.31837058e-01
-1.10991406e+00 7.26030111e-01 7.18256760e+00 8.84117782e-02
-1.08206952e+00 -1.18568838e-01 9.79448184e-02 -3.09274018e-01
-2.69338936e-01 -9.49881226e-02 -2.65692860e-01 2.45720789e-01
8.46376181e-01 1.77999452e-01 5.58170795e-01 7.75906384e-01
-2.51512416e-02 1.70978874e-01 -1.27030027e+00 1.11297655e+00
6.23245444e-03 -2.12839055e+00 6.35916740e-02 4.60400522e-01
1.02977288e+00 5.54432034e-01 -1.03048757e-01 -1.76570997e-01
4.01951939e-01 -1.67059863e+00 5.92117131e-01 7.33674347e-01
1.01033926e+00 -5.71827590e-01 5.41315615e-01 5.32602012e-01
-1.05387402e+00 2.20507964e-01 -7.32512772e-01 -4.67979610e-01
1.21759206e-01 7.10582495e-01 -4.27092165e-01 2.47648865e-01
9.96892750e-01 1.12916243e+00 -3.08056802e-01 9.23494637e-01
-1.70711689e-02 3.79075348e-01 -5.23961961e-01 2.60950834e-01
7.50611365e-01 -3.45047891e-01 6.51471734e-01 1.41375768e+00
3.32347095e-01 3.42611432e-01 7.74950087e-02 9.69846308e-01
-3.39529157e-01 -3.75350535e-01 -1.56830013e+00 -1.38616249e-01
1.44354761e-01 1.37052429e+00 -6.50565386e-01 -4.27398294e-01
-9.14419234e-01 7.51997709e-01 1.29461095e-01 5.53733766e-01
-1.21666722e-01 -4.25892264e-01 1.02095771e+00 3.90117526e-01
6.98537827e-01 -8.38366807e-01 -3.62671286e-01 -1.13654101e+00
-1.62365902e-02 -7.49705315e-01 -9.04619843e-02 -1.11870515e+00
-1.29897416e+00 5.31097770e-01 -1.22181267e-01 -1.34286785e+00
-2.19743788e-01 -9.71718252e-01 -7.98906446e-01 8.88938129e-01
-1.45858765e+00 -1.15853441e+00 -5.48581362e-01 1.06245887e+00
3.96509826e-01 -1.41907081e-01 7.85623491e-01 5.05467206e-02
-7.43385917e-03 -3.33546489e-01 3.65450867e-02 2.89221168e-01
-5.11690900e-02 -1.30880189e+00 6.74101889e-01 8.86555433e-01
6.11473501e-01 5.93494356e-01 2.77224183e-01 -6.25270963e-01
-1.83580196e+00 -1.04799747e+00 7.68975377e-01 -5.77416241e-01
8.18867683e-01 -4.53075856e-01 -1.10025716e+00 1.07721782e+00
2.27641970e-01 5.51028848e-01 2.80705601e-01 1.95991814e-01
-4.36895698e-01 3.17589700e-01 -1.09883356e+00 4.19593066e-01
1.45331144e+00 -9.59486425e-01 -7.35580921e-01 4.25403386e-01
5.65423846e-01 -5.57695329e-01 -8.88383389e-01 1.83054417e-01
4.80346441e-01 -1.21029484e+00 1.32446408e+00 -1.02854431e+00
5.86597800e-01 -4.80365217e-01 -3.32108974e-01 -1.30830109e+00
-6.16791666e-01 -8.34974706e-01 -5.15880466e-01 3.51274669e-01
-1.17502306e-02 -1.27693638e-01 1.17587912e+00 1.22822352e-01
-2.04973340e-01 -5.95477223e-01 -9.50397193e-01 -7.99273491e-01
1.62905365e-01 -6.02950215e-01 7.98823595e-01 1.21614575e+00
-4.12157416e-01 -1.38383046e-01 -4.16520089e-01 2.84839034e-01
8.47768247e-01 5.02585843e-02 7.84712255e-01 -1.72093558e+00
3.91428441e-01 -5.01490772e-01 -9.86719966e-01 -8.02165091e-01
1.56036213e-01 -8.89411092e-01 -1.45775050e-01 -1.55309081e+00
-3.03759187e-01 -5.33311307e-01 5.77774830e-03 5.65979242e-01
6.38470650e-01 7.25507617e-01 3.41275856e-02 3.77216548e-01
-2.51640499e-01 1.35479555e-01 1.01835203e+00 -1.43148407e-01
-3.17637436e-02 -2.90002882e-01 -4.45651710e-01 1.05762434e+00
5.72389603e-01 -2.74056077e-01 -2.14138135e-01 -7.14067757e-01
5.23496628e-01 -2.14200631e-01 9.32720482e-01 -1.15295708e+00
4.38377887e-01 -1.98849186e-01 7.39357591e-01 -9.10013735e-01
2.55229205e-01 -1.15525293e+00 3.93236756e-01 1.21293180e-01
-6.03179447e-02 1.84387892e-01 9.14893746e-02 4.61955637e-01
-2.11612701e-01 8.25151578e-02 6.53949499e-01 -7.04359472e-01
-6.99833870e-01 7.66613662e-01 -2.46460840e-01 -1.47725627e-01
6.08223021e-01 -5.28495133e-01 -2.83992589e-01 -2.90164113e-01
-5.98326445e-01 -2.53555179e-01 8.70947003e-01 4.06654179e-01
7.76254356e-01 -1.68840396e+00 -6.32827759e-01 6.49080992e-01
-2.75740057e-01 7.60525286e-01 -1.06990302e-03 8.14406455e-01
-9.33295786e-01 5.34548700e-01 -5.26674867e-01 -7.94732630e-01
-8.95046473e-01 5.28685153e-01 2.52491832e-01 3.62466544e-01
-1.49606764e+00 6.42173171e-01 3.74232195e-02 -5.40676177e-01
4.19009954e-01 -5.42109370e-01 1.53113693e-01 -5.65888435e-02
6.19165003e-01 3.81237447e-01 1.99063465e-01 -7.85259902e-01
-4.81500685e-01 6.65160954e-01 4.25425172e-01 1.27590045e-01
1.97387397e+00 2.41109401e-01 -3.95764500e-01 5.96042156e-01
1.36365950e+00 -1.43963605e-01 -1.40792835e+00 -3.80344778e-01
-8.80027413e-02 -6.28610551e-01 2.84575578e-02 -2.37978339e-01
-1.13125312e+00 1.15331984e+00 -1.36454618e-02 8.42146456e-01
7.84282625e-01 1.01544887e-01 7.09472001e-01 7.54093826e-01
5.28650582e-01 -8.66018593e-01 -9.90999117e-02 8.58591080e-01
1.07691205e+00 -1.16850996e+00 3.05027962e-01 -8.86052400e-02
-1.91894218e-01 1.24981809e+00 3.89858522e-02 -8.01011622e-01
9.77796137e-01 4.92061108e-01 6.21367944e-03 -8.35124612e-01
-4.92061615e-01 -1.67510152e-01 3.63572925e-01 9.36313510e-01
2.49555632e-01 5.06306216e-02 5.87089956e-01 -2.32398942e-01
-3.35293949e-01 1.30487114e-01 7.00067222e-01 1.18688548e+00
-3.35599840e-01 -7.28730440e-01 -3.78567427e-01 7.27244258e-01
-2.27688551e-01 -1.06086761e-01 -6.91620529e-01 8.87274504e-01
1.70042619e-01 2.84652114e-01 5.77462316e-01 -3.29846591e-01
2.20399842e-01 5.38416430e-02 3.69004846e-01 -7.56875873e-01
-4.23424721e-01 -3.05880457e-01 -4.77438532e-02 -1.02123344e+00
-1.09502625e+00 -8.65823269e-01 -8.98200214e-01 -7.49415457e-01
2.82282561e-01 -4.70915079e-01 8.34350169e-01 9.74456310e-01
3.20276976e-01 2.74055064e-01 5.71154475e-01 -1.62349343e+00
-5.03256097e-02 -7.86023140e-01 -8.68511558e-01 4.28058743e-01
7.83698678e-01 -7.00114727e-01 -5.43210804e-01 1.98728248e-01] | [8.060649871826172, -3.6980648040771484] |
9744c9ac-6450-458e-923a-2a4063e84056 | evolving-mario-levels-in-the-latent-space-of | 1805.00728 | null | http://arxiv.org/abs/1805.00728v1 | http://arxiv.org/pdf/1805.00728v1.pdf | Evolving Mario Levels in the Latent Space of a Deep Convolutional Generative Adversarial Network | Generative Adversarial Networks (GANs) are a machine learning approach
capable of generating novel example outputs across a space of provided training
examples. Procedural Content Generation (PCG) of levels for video games could
benefit from such models, especially for games where there is a pre-existing
corpus of levels to emulate. This paper trains a GAN to generate levels for
Super Mario Bros using a level from the Video Game Level Corpus. The approach
successfully generates a variety of levels similar to one in the original
corpus, but is further improved by application of the Covariance Matrix
Adaptation Evolution Strategy (CMA-ES). Specifically, various fitness functions
are used to discover levels within the latent space of the GAN that maximize
desired properties. Simple static properties are optimized, such as a given
distribution of tile types. Additionally, the champion A* agent from the 2009
Mario AI competition is used to assess whether a level is playable, and how
many jumping actions are required to beat it. These fitness functions allow for
the discovery of levels that exist within the space of examples designed by
experts, and also guide the search towards levels that fulfill one or more
specified objectives. | ['Adam Smith', 'Vanessa Volz', 'Simon M. Lucas', 'Jialin Liu', 'Sebastian Risi', 'Jacob Schrum'] | 2018-05-02 | null | null | null | null | ['snes-games'] | ['playing-games'] | [ 4.99571472e-01 1.92526177e-01 2.10258946e-01 1.54733062e-01
-8.76776695e-01 -7.66905606e-01 5.47105312e-01 -1.92059457e-01
-2.83151209e-01 8.77628744e-01 1.91579342e-01 -7.37593174e-02
-2.67094284e-01 -1.16525400e+00 -7.71254897e-01 -8.10727239e-01
1.04441643e-02 6.77116215e-01 2.07589939e-01 -6.56199515e-01
2.68328756e-01 3.07427913e-01 -1.89292395e+00 6.54740036e-01
1.01869512e+00 6.29905999e-01 1.81858554e-01 1.00823843e+00
1.02261811e-01 9.26284373e-01 -1.30419815e+00 -3.69200766e-01
4.83089894e-01 -1.07044697e+00 -3.83691967e-01 5.81827946e-02
1.35958374e-01 1.74635798e-01 1.13119856e-01 8.93850029e-01
5.08919835e-01 3.76997948e-01 8.40342462e-01 -1.39301968e+00
-3.10425907e-01 8.18613708e-01 9.74106975e-03 1.01435259e-01
2.42866933e-01 5.53900540e-01 8.77259076e-01 -2.32610002e-01
6.39559388e-01 9.57724690e-01 4.89524037e-01 8.88458312e-01
-1.32644153e+00 -6.40467584e-01 -2.90268779e-01 -1.13296404e-01
-1.19567704e+00 7.50561282e-02 6.77993536e-01 -4.82063919e-01
5.67623019e-01 5.06823897e-01 1.26830423e+00 1.10727561e+00
2.99163610e-01 7.02129424e-01 1.26365578e+00 -5.51457047e-01
8.67800832e-01 9.47571322e-02 -8.28727424e-01 5.38140237e-01
9.55254063e-02 3.46222132e-01 -7.03328311e-01 -3.93853448e-02
7.61293888e-01 -8.68107319e-01 8.09931103e-03 -4.61137116e-01
-8.19202960e-01 1.14280558e+00 3.55500370e-01 3.40834111e-01
-5.29303908e-01 2.58249938e-01 2.36861661e-01 1.19630143e-01
2.21586823e-01 1.54797232e+00 -1.93826139e-01 -5.98640919e-01
-1.11387205e+00 8.04894686e-01 7.18934000e-01 6.81132019e-01
6.37901425e-01 5.20370066e-01 -5.68457842e-01 5.57118714e-01
-1.98566630e-01 9.86398160e-02 6.71124697e-01 -1.18181455e+00
2.65561819e-01 7.98596919e-01 -9.50805750e-03 -6.39581919e-01
-1.02316834e-01 -5.48008859e-01 -4.29127425e-01 8.78400505e-01
3.25309366e-01 -6.81471288e-01 -1.11941195e+00 1.69681263e+00
2.29690716e-01 1.95137247e-01 2.09428310e-01 6.05579793e-01
6.19148552e-01 7.54169762e-01 -8.56151208e-02 9.69483256e-02
9.33492720e-01 -7.56610692e-01 -4.01960850e-01 -5.20476222e-01
4.29929018e-01 -4.20088887e-01 1.25880790e+00 4.60495204e-01
-1.38754904e+00 -6.26945734e-01 -1.19314289e+00 7.31797159e-01
-4.76861626e-01 -2.17120573e-01 7.37690449e-01 1.30581975e+00
-1.21482015e+00 5.58773100e-01 -5.82423508e-01 2.07449824e-01
3.89171988e-01 4.75882471e-01 2.87297666e-01 1.01176143e-01
-1.27419150e+00 8.51200998e-01 6.28072739e-01 -2.21398529e-02
-1.19559968e+00 -9.10966456e-01 -9.28224444e-01 6.24771975e-02
3.63613278e-01 -7.05108404e-01 1.07966316e+00 -1.36392152e+00
-1.81575298e+00 4.25786704e-01 4.73700941e-01 -5.59230447e-01
6.49995446e-01 3.33805412e-01 -5.38577698e-02 -1.45460859e-01
1.39933392e-01 9.84954238e-01 8.04854453e-01 -1.30833256e+00
-6.84246361e-01 1.84946463e-01 5.21344721e-01 5.12245953e-01
-1.29580110e-01 -2.82232672e-01 -3.08214575e-01 -8.41030180e-01
-4.89088535e-01 -1.11979425e+00 -5.76926351e-01 -6.98899627e-01
-3.67658019e-01 5.91565333e-02 1.57292292e-01 -5.09908795e-01
1.19190133e+00 -1.70870090e+00 5.81076860e-01 6.39989734e-01
-2.22282931e-01 2.32015643e-02 -2.12702304e-01 4.42923874e-01
1.35588914e-01 3.60311985e-01 -2.23139331e-01 1.30822331e-01
1.24642365e-01 1.25718147e-01 2.16404557e-01 -1.37671188e-01
2.34196916e-01 9.24878299e-01 -7.49920189e-01 -1.89489439e-01
6.34325519e-02 1.71763569e-01 -1.06192374e+00 2.63691276e-01
-7.25035429e-01 6.32515192e-01 -3.06734204e-01 3.36121023e-01
7.13841394e-02 1.18015498e-01 -1.13616161e-01 3.41832101e-01
2.75288951e-02 -3.82442563e-03 -1.13501084e+00 1.42774570e+00
-5.15414953e-01 4.47133273e-01 -3.12612325e-01 -7.32821107e-01
1.08557606e+00 -4.86736558e-02 4.46268588e-01 -4.02504504e-01
2.55245119e-01 2.51873117e-02 5.59949696e-01 -3.19258153e-01
6.78915143e-01 -2.97951907e-01 -4.56315041e-01 2.23915681e-01
-1.05355725e-01 -9.21895623e-01 7.47013927e-01 -1.50824621e-01
1.45884371e+00 3.30324560e-01 -8.99605155e-02 -1.58174992e-01
2.77707905e-01 3.86394262e-01 6.09988391e-01 1.07606018e+00
2.17245445e-01 8.59028995e-01 6.85272217e-01 -2.97334403e-01
-1.32359433e+00 -8.88587415e-01 2.95031875e-01 1.00840235e+00
-2.33258933e-01 -3.81685108e-01 -1.13280654e+00 -3.13709021e-01
-4.61319566e-01 1.06532228e+00 -1.00131857e+00 -3.13804239e-01
-4.88586694e-01 -7.87438691e-01 6.08678579e-01 3.53922009e-01
4.17239010e-01 -1.77891159e+00 -1.14049411e+00 3.44392270e-01
-3.73993069e-02 -6.82284236e-01 -3.41669142e-01 3.05662692e-01
-4.13016289e-01 -9.18808162e-01 -6.53735161e-01 -6.81864619e-01
5.84114790e-01 -8.15504730e-01 1.30979204e+00 1.13330133e-01
-3.98068279e-01 1.83911309e-01 -6.11396849e-01 -5.87813020e-01
-1.09881508e+00 2.79383332e-01 -2.22290203e-01 -2.83910841e-01
-1.20070308e-01 -3.70728701e-01 -3.69723499e-01 2.61248976e-01
-1.00480127e+00 4.19050366e-01 6.49640858e-01 9.36726093e-01
4.96676773e-01 6.69288456e-01 5.30205309e-01 -9.47411716e-01
1.14090407e+00 -2.73878872e-01 -6.95450008e-01 1.62546560e-01
-2.88026601e-01 1.96605340e-01 6.63310409e-01 -7.69964576e-01
-9.23205674e-01 -7.54984021e-02 -1.82958618e-01 -2.77221084e-01
-9.91633311e-02 5.42321086e-01 -2.73578554e-01 1.05030634e-01
1.09617662e+00 2.59182334e-01 -1.05249815e-01 2.40206420e-01
3.95293862e-01 3.87720168e-02 4.51202363e-01 -9.16548908e-01
9.04166579e-01 -1.04545593e-01 -1.59778178e-01 -4.07704979e-01
-4.51037794e-01 2.07133010e-01 -3.27072859e-01 -6.36961579e-01
1.01614213e+00 -5.38570881e-01 -4.80788887e-01 3.36565852e-01
-5.60233295e-01 -1.03744590e+00 -8.37187350e-01 1.72950774e-01
-1.03044331e+00 -5.33194721e-01 -2.89689392e-01 -8.07440281e-01
-2.40007229e-02 -1.35445940e+00 7.95755386e-01 6.67181730e-01
-4.54404891e-01 -7.68888056e-01 1.50190920e-01 2.99534917e-01
3.15468937e-01 8.01683068e-01 1.02195823e+00 -3.10754359e-01
-4.53627229e-01 -2.29471058e-01 7.48249292e-01 2.37761348e-01
1.26529634e-01 4.35126200e-02 -3.80354881e-01 -3.92341644e-01
-3.61286610e-01 -4.07568306e-01 1.39568344e-01 6.69881582e-01
1.09580219e+00 -5.20480216e-01 6.44891188e-02 6.26336753e-01
1.24007392e+00 8.31737876e-01 9.85109091e-01 8.74893069e-01
3.63592803e-01 5.10455728e-01 5.17963290e-01 3.44265550e-01
-9.60116088e-02 7.84063339e-01 5.09398997e-01 -1.12093510e-02
-7.30071515e-02 -4.34090823e-01 3.61275703e-01 2.39998609e-01
-1.43519700e-01 -3.34392309e-01 -7.85699308e-01 5.99853933e-01
-1.55638731e+00 -1.14617217e+00 3.44531775e-01 1.97054684e+00
9.71303701e-01 6.09156311e-01 5.24435878e-01 1.42379299e-01
7.30918884e-01 3.34371105e-02 -7.94566214e-01 -8.20168018e-01
-8.83285478e-02 9.42667127e-01 4.26935375e-01 1.57069474e-01
-7.21902728e-01 1.03547966e+00 6.52997351e+00 1.07462966e+00
-6.78020537e-01 -3.80718201e-01 9.28401470e-01 -3.35372776e-01
-6.59554839e-01 4.69971448e-03 -5.48591435e-01 6.66619718e-01
9.27547157e-01 -4.14922386e-01 8.38680029e-01 7.28020966e-01
2.45978817e-01 -1.15487151e-01 -8.00435841e-01 5.02190888e-01
8.46327394e-02 -1.74358284e+00 1.38845056e-01 3.23948473e-01
1.31551933e+00 -5.56784451e-01 3.61172289e-01 6.12540305e-01
8.50511909e-01 -1.48999488e+00 9.46972072e-01 3.55047017e-01
7.88455904e-01 -1.38170052e+00 7.06639469e-01 5.48460543e-01
-8.43529820e-01 -2.53720313e-01 -1.92251936e-01 1.16351508e-02
1.00140739e-02 -8.26866627e-02 -9.71834898e-01 2.53478408e-01
5.65811396e-01 -1.24992318e-01 -8.51596236e-01 1.06419766e+00
-4.80500907e-01 6.73852205e-01 -2.47090250e-01 -4.61641550e-01
5.53509414e-01 -2.58384854e-01 6.20338738e-01 6.33602440e-01
5.43925345e-01 8.73714015e-02 -4.05173786e-02 1.13354886e+00
5.53683192e-02 6.85896259e-03 -5.47787905e-01 -4.74001840e-02
4.72077489e-01 9.31742370e-01 -8.73262167e-01 -6.16849586e-02
2.38752738e-01 8.74326527e-01 3.90576161e-02 1.45631030e-01
-9.14224088e-01 -1.52602121e-01 5.78889132e-01 1.93196923e-01
4.30422604e-01 6.82137236e-02 -4.69207048e-01 -5.44992685e-01
-3.87943029e-01 -1.58391762e+00 2.99270242e-01 -9.50750053e-01
-1.00970650e+00 8.79977584e-01 1.88676342e-01 -1.37596512e+00
-7.61064172e-01 -2.99060255e-01 -9.21949685e-01 7.56011248e-01
-6.26621187e-01 -9.61986721e-01 -2.51167864e-01 2.48110130e-01
8.96324813e-01 -8.24516177e-01 6.69102848e-01 -3.00650448e-01
-3.12179506e-01 6.93441391e-01 -6.18167780e-02 -1.75047368e-02
-2.81133596e-02 -1.46946597e+00 5.16803145e-01 9.59742606e-01
3.94364893e-01 1.64671317e-01 1.14184654e+00 -7.99960017e-01
-9.57542002e-01 -9.70828593e-01 1.40912617e-02 -3.44089925e-01
4.39259768e-01 -4.38689500e-01 -4.28087354e-01 2.23506600e-01
1.97094232e-01 -6.41802132e-01 7.02590823e-01 -2.02104613e-01
3.87730718e-01 3.53757262e-01 -1.20238864e+00 1.02939963e+00
8.95091712e-01 1.17992401e-01 -2.54393816e-01 -2.35265531e-02
4.24867034e-01 -9.53743160e-01 -6.56580329e-01 3.81656498e-01
3.24603945e-01 -1.09257257e+00 7.94132471e-01 -7.78084517e-01
9.00334477e-01 -3.14446956e-01 6.32844269e-02 -1.96866453e+00
-2.95198411e-01 -7.73902655e-01 3.90253514e-01 1.10446250e+00
7.00914443e-01 -3.88532355e-02 1.49816382e+00 5.52871466e-01
-1.97994068e-01 -9.27045405e-01 -8.24142575e-01 -7.03578711e-01
3.28866988e-01 -3.42100352e-01 8.43360424e-01 5.52541494e-01
-2.93421656e-01 1.04058109e-01 -3.01427096e-01 -6.54614419e-02
4.05854106e-01 -1.27502054e-01 1.06457460e+00 -7.45481074e-01
-8.16199362e-01 -7.02306032e-01 -4.61490542e-01 -5.23772776e-01
1.26002789e-01 -8.03333104e-01 3.99523288e-01 -1.58325112e+00
-1.16921663e-01 -7.64418721e-01 1.35079786e-01 3.13787401e-01
-2.00642049e-01 3.08898032e-01 5.37180305e-01 -1.21181458e-01
-1.85902417e-02 5.42003810e-01 1.27439201e+00 -2.83838123e-01
-5.94260693e-01 1.85363755e-01 -8.97307038e-01 6.27974570e-01
9.59003925e-01 -5.28313875e-01 -5.60669243e-01 1.98699106e-02
5.34200728e-01 1.59824878e-01 4.88292091e-02 -1.47671461e+00
-4.02439423e-02 -4.02504414e-01 4.78919715e-01 -2.52293408e-01
3.69692862e-01 -4.53120828e-01 6.34127021e-01 5.45584023e-01
-5.76850772e-01 2.42529467e-01 4.88599569e-01 3.39659065e-01
2.52750553e-02 -7.34374762e-01 7.34203577e-01 -3.87113810e-01
-7.07248628e-01 -3.00425831e-02 -9.18453753e-01 3.42165112e-01
1.09069431e+00 -6.22268796e-01 9.90007371e-02 -5.68636835e-01
-6.82515085e-01 1.07054211e-01 5.99191248e-01 4.37527806e-01
4.12349731e-01 -1.43621075e+00 -9.29720759e-01 8.84977132e-02
1.28634665e-02 1.87715739e-01 1.49953097e-01 -1.66413754e-01
-8.61595571e-01 -2.92646110e-01 -7.28168011e-01 -2.91486979e-01
-9.65811849e-01 1.94804773e-01 5.99836707e-01 -6.61375642e-01
-5.13378203e-01 8.97295058e-01 3.90139148e-02 -3.87352198e-01
-2.36877084e-01 -1.24235479e-02 -2.91358829e-01 -1.41458690e-01
2.95180202e-01 1.90691799e-01 4.11631949e-02 -3.73658836e-01
-8.26795679e-03 5.40510416e-02 4.40205425e-01 -5.22702515e-01
1.49941421e+00 6.05269253e-01 3.31862718e-01 2.93062776e-01
6.05172634e-01 6.69747442e-02 -1.71924484e+00 4.73443568e-01
-1.05605885e-01 -4.10126060e-01 -3.31514239e-01 -1.04919982e+00
-9.44470108e-01 1.81396097e-01 5.71120143e-01 3.45755100e-01
1.17661834e+00 -1.74875915e-01 2.63051838e-01 -6.63483962e-02
3.52953255e-01 -1.16985834e+00 5.99114299e-01 3.48825663e-01
8.73468518e-01 -5.22801578e-01 -3.05333585e-01 8.42236634e-03
-9.09311414e-01 7.62409687e-01 9.73950326e-01 -2.26034433e-01
-9.46253315e-02 5.99866152e-01 -1.75143424e-02 -2.18013123e-01
-6.68536484e-01 -2.96282977e-01 2.91577309e-01 9.81125295e-01
1.65046891e-03 1.10598199e-01 -3.35893154e-01 4.10030782e-01
-1.11826086e+00 -2.76937515e-01 8.53463054e-01 8.37520480e-01
-2.25504488e-01 -1.19170678e+00 -4.83987272e-01 6.53411567e-01
-2.33767584e-01 -1.00561686e-01 -3.68473858e-01 9.10545230e-01
4.67475951e-01 9.38741207e-01 1.83109552e-01 -3.74124765e-01
4.35550272e-01 -6.63697794e-02 5.81321836e-01 -9.60160613e-01
-9.84734356e-01 -1.04551509e-01 1.33270502e-01 -1.14310026e-01
-1.49995014e-01 -7.59048045e-01 -9.93014038e-01 -7.98223093e-02
-2.39195570e-01 4.92896706e-01 5.13784587e-01 7.85530925e-01
-1.50549442e-01 1.01122272e+00 4.95882660e-01 -8.29453945e-01
-8.50870535e-02 -7.83667922e-01 -4.21185225e-01 4.39731240e-01
-4.49595988e-01 -6.57021940e-01 -1.14571273e-01 1.24403343e-01] | [3.6620688438415527, 1.554531455039978] |
6ec2d417-5a4a-4534-a7c7-e82ebf35baa9 | a-greedy-bit-flip-training-algorithm-for | null | null | https://aclanthology.org/2020.findings-emnlp.10 | https://aclanthology.org/2020.findings-emnlp.10.pdf | A Greedy Bit-flip Training Algorithm for Binarized Knowledge Graph Embeddings | This paper presents a simple and effective discrete optimization method for training binarized knowledge graph embedding model B-CP. Unlike the prior work using a SGD-based method and quantization of real-valued vectors, the proposed method directly optimizes binary embedding vectors by a series of bit flipping operations. On the standard knowledge graph completion tasks, the B-CP model trained with the proposed method achieved comparable performance with that trained with SGD as well as state-of-the-art real-valued models with similar embedding dimensions. | ['Masashi Shimbo', 'Koki Kishimoto', 'Katsuhiko Hayashi'] | 2020-11-01 | null | null | null | findings-of-the-association-for-computational | ['knowledge-graph-embeddings', 'knowledge-graph-embeddings'] | ['graphs', 'methodology'] | [ 1.16258010e-01 6.93039954e-01 -6.15708590e-01 -2.93356061e-01
-4.48158979e-01 -2.32749581e-01 4.55842227e-01 4.35964912e-01
-5.31750500e-01 6.84266210e-01 -9.32917744e-02 -4.04144019e-01
-3.62374753e-01 -1.15290904e+00 -7.64901280e-01 -5.43175161e-01
-4.25184071e-01 7.89683282e-01 2.96768527e-02 -1.24677755e-01
1.52776271e-01 3.76318753e-01 -9.59282815e-01 -3.30999255e-01
5.46679497e-01 1.05119455e+00 -2.17867658e-01 7.84040868e-01
2.48188861e-02 5.71214139e-01 -6.62299693e-01 -9.42130506e-01
4.46945012e-01 2.91131940e-02 -8.58517766e-01 -2.05618769e-01
3.90888005e-01 -3.57633144e-01 -1.26185751e+00 1.24651337e+00
4.27673846e-01 5.03944695e-01 9.40956593e-01 -1.40932560e+00
-1.57752371e+00 5.53549051e-01 -3.24377976e-03 4.90769446e-02
3.18249881e-01 -2.93577492e-01 1.34018314e+00 -8.74203980e-01
4.89106208e-01 1.37503290e+00 6.11609578e-01 1.94242507e-01
-1.38181841e+00 -2.71863401e-01 -2.01943681e-01 6.55072927e-01
-1.80667627e+00 -6.04094304e-02 7.54776895e-01 -6.57158196e-02
1.64185023e+00 -2.30481520e-01 8.37297916e-01 8.23248386e-01
3.42357457e-02 4.51507598e-01 7.52289951e-01 -7.89119661e-01
3.36329818e-01 1.94127470e-01 3.39005858e-01 1.42483246e+00
8.62064242e-01 1.33565769e-01 -6.65195882e-01 -3.65463316e-01
6.55995429e-01 9.13989991e-02 -3.06282550e-01 -6.62671685e-01
-1.27345443e+00 1.24121773e+00 4.69703168e-01 -7.25531727e-02
-4.16616440e-01 6.98725581e-01 2.63716370e-01 4.51098263e-01
2.62189984e-01 1.98621541e-01 -2.61440307e-01 -2.56275773e-01
-8.14762950e-01 -2.56414860e-01 1.33172500e+00 1.22618020e+00
8.40862930e-01 3.24582696e-01 -3.78333271e-01 5.09282351e-01
7.19722509e-01 4.72915709e-01 4.44337785e-01 -8.26775134e-01
3.95801485e-01 4.69919205e-01 -1.18631043e-01 -1.22020829e+00
7.14117438e-02 -5.77404946e-02 -6.35571420e-01 2.10924402e-01
-5.09735644e-02 1.30659804e-01 -1.26285946e+00 1.33495474e+00
1.13701642e-01 3.79036486e-01 4.91264552e-01 6.23836458e-01
9.27896738e-01 6.69869244e-01 -2.41372466e-01 2.13786513e-01
1.06031382e+00 -1.21451366e+00 -9.90836382e-01 1.04769163e-01
4.36322004e-01 -2.77649075e-01 8.61842573e-01 2.44420946e-01
-1.03811800e+00 -2.46945769e-01 -1.63635647e+00 -2.01834932e-01
-9.99924958e-01 -2.82061212e-02 1.15335321e+00 1.28673697e+00
-1.38265061e+00 6.40058458e-01 -9.35994208e-01 -2.59280175e-01
5.07664442e-01 4.88017321e-01 -6.01277947e-01 -3.51227313e-01
-1.41957378e+00 1.08072364e+00 8.95621359e-01 1.56871378e-01
-1.21351957e+00 -3.75942677e-01 -1.35146439e+00 4.48706299e-01
9.73282382e-02 -5.38567245e-01 5.69903433e-01 1.37609079e-01
-1.90036631e+00 7.73014545e-01 1.22602195e-01 -7.72380948e-01
-1.84091002e-01 4.97425124e-02 -5.18289626e-01 4.43856925e-01
-3.52598250e-01 5.23369253e-01 1.21330667e+00 -7.46500909e-01
1.00541607e-01 -2.14212060e-01 2.97131568e-01 1.75742134e-01
-9.11996365e-01 -4.49434012e-01 -5.56203663e-01 -6.87054574e-01
-7.61139626e-03 -6.44273400e-01 5.85792474e-02 1.25768557e-01
-2.29511365e-01 -1.86206818e-01 7.88772225e-01 -8.52701545e-01
1.27634990e+00 -1.93930948e+00 3.82883072e-01 5.51570475e-01
3.57785821e-01 4.47384983e-01 -6.40152320e-02 6.29134834e-01
9.10165980e-02 1.82392895e-01 -1.64243311e-01 -5.75197875e-01
7.92980909e-01 7.34788239e-01 -1.52746797e-01 5.87007821e-01
1.78351760e-01 1.12218034e+00 -1.19889355e+00 -6.81639433e-01
2.64011711e-01 8.88343573e-01 -4.60578531e-01 1.99089080e-01
-8.19783360e-02 -6.66736960e-01 -8.27114731e-02 8.36339712e-01
6.29587829e-01 -4.99192685e-01 3.98129791e-01 -1.81027830e-01
8.22586596e-01 1.50834858e-01 -1.29402685e+00 1.95896816e+00
-2.11915478e-01 7.75505483e-01 -2.55125642e-01 -1.37244439e+00
1.05862784e+00 4.21455234e-01 1.64049178e-01 -3.44072491e-01
-7.88504351e-03 -5.36841340e-03 -5.07550716e-01 2.00837534e-02
7.29321957e-01 4.74936962e-02 1.33891851e-01 3.38524222e-01
8.02461684e-01 -2.15336099e-01 4.41446379e-02 5.02593458e-01
1.40425193e+00 -6.59063160e-02 1.78478658e-01 1.10188305e-01
4.99573469e-01 -2.26731509e-01 3.14475656e-01 8.00338626e-01
-3.23563248e-01 2.87213475e-01 3.05149764e-01 -9.08776149e-02
-7.68422425e-01 -1.37381804e+00 -2.14273915e-01 6.99745238e-01
2.01561391e-01 -7.99548566e-01 -5.17844558e-01 -8.15169156e-01
4.28095996e-01 8.86649251e-01 -6.39642000e-01 -5.48750162e-01
4.78910282e-03 -6.84521198e-01 9.71543074e-01 6.59773946e-01
6.19829118e-01 -8.30910563e-01 2.12532923e-01 2.58258492e-01
2.29090154e-01 -9.68026996e-01 -4.58445370e-01 3.33862036e-01
-9.02539849e-01 -9.68370974e-01 -6.26811564e-01 -1.23630726e+00
8.20212245e-01 1.57775015e-01 8.76109421e-01 1.08264722e-02
-3.71677190e-01 8.20334256e-01 -4.22817856e-01 -1.28916487e-01
-1.48790672e-01 -1.39824241e-01 -5.18049300e-02 -9.53871831e-02
6.64748132e-01 -5.00460923e-01 -4.24981892e-01 -1.06700383e-01
-9.77895975e-01 -6.36204064e-01 5.59730291e-01 1.14427102e+00
8.23159873e-01 4.61940140e-01 2.84527421e-01 -5.64190209e-01
1.10217810e+00 -2.04571083e-01 -6.59301102e-01 5.76124668e-01
-1.29486251e+00 5.47649860e-01 3.38046819e-01 -4.06897992e-01
-6.43776476e-01 -2.61887759e-01 1.92824259e-01 -9.11520243e-01
4.66471612e-01 7.59521782e-01 1.09167166e-01 -8.84252191e-01
6.98287070e-01 3.34284246e-01 4.66106497e-02 -3.27625632e-01
9.44234073e-01 7.38304138e-01 4.58328962e-01 -5.09742320e-01
8.40520620e-01 3.40148032e-01 1.08387515e-01 -7.30503857e-01
-3.80907178e-01 -4.72902119e-01 -6.67159021e-01 2.70975679e-01
9.11609590e-01 -8.84594262e-01 -7.65963376e-01 2.34884858e-01
-1.13425410e+00 -2.51790792e-01 -4.31496352e-01 7.23068357e-01
-5.66623271e-01 6.72861814e-01 -6.16654217e-01 -4.62988824e-01
-6.13393366e-01 -6.75353646e-01 9.05657709e-01 -8.75338838e-02
2.05071822e-01 -1.62873280e+00 3.97360116e-01 3.80153716e-01
2.88552731e-01 9.24924091e-02 1.10102987e+00 -6.24121189e-01
-5.86337984e-01 -6.78315222e-01 -2.14762986e-01 7.43538380e-01
-9.23858359e-02 -3.47784877e-01 -5.27792454e-01 -5.32854736e-01
-3.54087889e-01 -6.58705294e-01 8.67908061e-01 -6.96167024e-03
9.57811594e-01 -4.42803442e-01 -2.77306288e-01 8.77492905e-01
1.80266571e+00 -1.74171910e-01 7.19092965e-01 2.42776111e-01
6.61899090e-01 -5.37334323e-01 4.07436311e-01 2.18899086e-01
6.08644724e-01 4.10480320e-01 4.87919360e-01 1.55342519e-01
-3.24600577e-01 -5.33284307e-01 3.30539286e-01 1.01730609e+00
-4.43729851e-03 -3.30232471e-01 -7.38033950e-01 8.31517398e-01
-1.98578894e+00 -8.63801837e-01 5.12442887e-01 2.24038363e+00
9.02799070e-01 2.64981538e-01 -4.22759861e-01 3.67698252e-01
6.79241836e-01 4.06072408e-01 -3.92430693e-01 -8.41880918e-01
-1.14779659e-01 7.98361719e-01 1.03985083e+00 7.89374650e-01
-8.59471560e-01 9.86306310e-01 7.99756098e+00 8.08165073e-01
-6.13558173e-01 3.34560335e-01 -2.98833817e-01 1.91203877e-01
-2.86745578e-01 1.61303252e-01 -7.57073164e-01 5.59952110e-02
1.01698387e+00 -2.80343771e-01 9.84169483e-01 9.34114397e-01
-7.08943605e-01 3.47459942e-01 -1.12206745e+00 1.25526571e+00
6.25002146e-01 -1.38217247e+00 1.98876023e-01 4.68046702e-02
8.79047275e-01 -3.24633271e-02 -9.62849781e-02 5.38905144e-01
5.77725887e-01 -1.37274528e+00 2.62356609e-01 5.13276279e-01
9.30055737e-01 -5.36580563e-01 6.96999490e-01 -3.32669288e-01
-1.15954459e+00 1.61182031e-01 -8.84490848e-01 8.02866369e-03
1.20719515e-01 5.65370500e-01 -1.10175049e+00 6.30248487e-01
3.52837741e-01 7.49479532e-01 -5.53045571e-01 8.20681512e-01
-8.22887897e-01 1.00149798e+00 -3.56529832e-01 -3.26988041e-01
1.48294359e-01 -3.49578351e-01 3.98106545e-01 1.33977652e+00
2.97884271e-02 -8.14415589e-02 -2.03508168e-01 8.32663953e-01
-3.16905290e-01 -2.14190215e-01 -6.74184382e-01 -8.45462382e-01
5.98672926e-01 9.67215359e-01 -3.12867612e-01 -5.14719844e-01
-6.37922823e-01 1.26155794e+00 7.67005622e-01 6.87307596e-01
-8.38785708e-01 -1.23738480e+00 4.51086223e-01 -3.53571326e-01
1.04802167e+00 -6.82305932e-01 -1.07201576e-01 -1.19662416e+00
4.49464377e-03 -3.51465613e-01 6.01239800e-01 -7.10030258e-01
-1.28983617e+00 2.45118991e-01 -7.39855245e-02 -6.13669515e-01
4.67926301e-02 -9.01199698e-01 -4.30949897e-01 9.55905497e-01
-1.77686846e+00 -9.86580312e-01 -1.76896304e-01 1.07112324e+00
-5.30823767e-01 -4.27552730e-01 1.47397053e+00 2.59341627e-01
-3.53203595e-01 1.02471411e+00 6.62657022e-01 4.01505619e-01
4.42001075e-01 -1.42437136e+00 4.76702422e-01 5.27454019e-01
5.60226858e-01 5.96811235e-01 2.65403807e-01 -5.86889505e-01
-1.87695217e+00 -8.79398406e-01 9.19181108e-01 -1.45599857e-01
7.89018989e-01 -3.43228787e-01 -7.19799757e-01 9.88014519e-01
1.90220580e-01 2.72383362e-01 7.45011210e-01 8.82240385e-02
-5.89356959e-01 -1.10251777e-01 -1.52339351e+00 2.97078282e-01
9.76486206e-01 -1.01348794e+00 -1.24202406e+00 7.53339469e-01
1.05578578e+00 -2.12492347e-01 -1.52257931e+00 5.73433675e-02
2.83193588e-01 -8.99635255e-02 1.42579424e+00 -8.12878489e-01
-1.40736803e-01 -1.48445129e-01 -4.33061212e-01 -1.33651328e+00
-5.73865891e-01 -6.93518817e-01 -1.15717208e+00 5.66722751e-01
2.38583535e-01 -1.06189811e+00 1.02748430e+00 5.06236732e-01
9.04994160e-02 -6.94163024e-01 -1.20159984e+00 -1.13424432e+00
-2.19749138e-01 -3.82299088e-02 3.86270195e-01 9.28632379e-01
5.01457095e-01 2.19721019e-01 -1.75917834e-01 3.33759099e-01
1.20034122e+00 -6.83392584e-02 5.12059629e-01 -8.99010539e-01
-4.48956728e-01 -1.44352093e-01 -1.31738722e+00 -9.72071588e-01
4.48113769e-01 -1.51637697e+00 -4.47064012e-01 -2.00555587e+00
-1.37413785e-01 -1.17312975e-01 -6.72551572e-01 9.17572081e-01
-4.67871688e-02 1.69191778e-01 -1.43611774e-01 -1.91397503e-01
-7.33724177e-01 1.01374912e+00 1.08115172e+00 -6.14171505e-01
4.08088043e-02 -5.53117096e-01 -3.76090467e-01 6.43399134e-02
5.83650529e-01 -7.00907528e-01 -7.66876936e-01 -3.20452243e-01
3.67422432e-01 -8.99478719e-02 4.05993581e-01 -7.84922123e-01
4.37336683e-01 1.51921839e-01 1.65313616e-01 -3.41280222e-01
8.87099206e-01 -8.02055478e-01 -1.19608916e-01 3.84185344e-01
1.10037483e-01 -1.35028347e-01 6.88825846e-02 1.05482900e+00
-3.89995009e-01 -5.04484057e-01 2.94432849e-01 6.02564067e-02
-7.70577729e-01 4.20100749e-01 -2.19887886e-02 1.36102200e-01
9.15937066e-01 -5.21473587e-01 -4.99116480e-01 -4.49400723e-01
-7.94644415e-01 1.04873426e-01 2.53598988e-01 1.79422304e-01
1.21539247e+00 -1.61208093e+00 -3.92506927e-01 3.19728225e-01
1.19939841e-01 -1.13426529e-01 -3.15533608e-01 4.30092931e-01
-7.77362406e-01 6.91476882e-01 -1.94604516e-01 -1.37822807e-01
-6.89427257e-01 7.82186747e-01 4.03281376e-02 -2.80646175e-01
-6.21184170e-01 8.35889757e-01 -8.70715976e-01 -5.62456250e-01
6.94166720e-01 -5.22190072e-02 1.65750146e-01 -2.54573762e-01
1.06338419e-01 6.94142282e-01 2.75413632e-01 -3.23836267e-01
-3.35761517e-01 2.97512770e-01 7.66168013e-02 -2.49918371e-01
1.39457428e+00 2.96766639e-01 -1.35830164e-01 2.15024754e-01
1.66074121e+00 -5.59985995e-01 -7.90898502e-01 -4.51896310e-01
-5.21645129e-01 -5.98777473e-01 5.40142179e-01 -5.64313889e-01
-9.77008939e-01 8.49529207e-01 4.73198503e-01 -1.86326331e-04
6.58458829e-01 -2.54785508e-01 7.15630114e-01 1.02062130e+00
5.46711624e-01 -1.29862690e+00 2.34226927e-01 4.82036948e-01
6.86752439e-01 -1.24389160e+00 3.27676594e-01 -2.33682632e-01
-3.04769099e-01 1.07304907e+00 4.78787422e-02 -1.58498481e-01
1.28458095e+00 -3.63987714e-01 -4.07490343e-01 -5.52901149e-01
-5.42076945e-01 -9.33083072e-02 4.14748728e-01 1.04159772e+00
-2.25116700e-01 1.69252083e-01 -4.98337984e-01 3.25705588e-01
-1.06902085e-01 3.61216962e-02 2.94678479e-01 1.20675075e+00
-2.97937900e-01 -1.09525263e+00 -2.25483328e-01 5.19803703e-01
-6.21014684e-02 -4.19993520e-01 -2.71071583e-01 5.85084617e-01
-6.31333649e-01 9.64921951e-01 -2.04374209e-01 -5.63879311e-01
1.54110879e-01 3.59780699e-01 8.08874249e-01 -6.24211609e-01
-1.45052418e-01 -1.01110220e+00 1.34643659e-01 -3.70282561e-01
-2.71839052e-01 -1.03017993e-01 -1.51105142e+00 -4.69397813e-01
-5.86427569e-01 1.16649643e-01 9.46777344e-01 3.65950912e-01
5.19630611e-01 3.25180054e-01 2.69750774e-01 -6.41139865e-01
-1.00675380e+00 -7.81181097e-01 -9.87357736e-01 5.82470059e-01
1.78010091e-02 -8.42531681e-01 -6.19632006e-01 -1.35421976e-01] | [8.729273796081543, 7.859460830688477] |
bd242a3c-6f49-4106-80a3-a0ff20ebae29 | paraphrase-identification-with-deep-learning | 2212.06933 | null | https://arxiv.org/abs/2212.06933v1 | https://arxiv.org/pdf/2212.06933v1.pdf | Paraphrase Identification with Deep Learning: A Review of Datasets and Methods | The rapid advancement of AI technology has made text generation tools like GPT-3 and ChatGPT increasingly accessible, scalable, and effective. This can pose serious threat to the credibility of various forms of media if these technologies are used for plagiarism, including scientific literature and news sources. Despite the development of automated methods for paraphrase identification, detecting this type of plagiarism remains a challenge due to the disparate nature of the datasets on which these methods are trained. In this study, we review traditional and current approaches to paraphrase identification and propose a refined typology of paraphrases. We also investigate how this typology is represented in popular datasets and how under-representation of certain types of paraphrases impacts detection capabilities. Finally, we outline new directions for future research and datasets in the pursuit of more effective paraphrase detection using AI. | ['Daniel E. Acuna', 'Cheng Qiu', 'Chao Zhou'] | 2022-12-13 | null | null | null | null | ['paraphrase-identification'] | ['natural-language-processing'] | [ 2.95021713e-01 -9.46501922e-03 -3.99073362e-01 2.94937585e-02
-9.21391606e-01 -9.85770464e-01 8.54061306e-01 8.01905215e-01
-1.33213818e-01 5.63235164e-01 7.87182927e-01 -4.80040461e-01
3.19744572e-02 -7.59186983e-01 -5.47008932e-01 -1.23980314e-01
5.46073377e-01 3.70661497e-01 1.51224378e-02 -2.59690851e-01
1.17797661e+00 2.63806731e-01 -1.35154974e+00 6.67967439e-01
1.08066845e+00 3.41949284e-01 -1.08546726e-01 4.29543048e-01
-5.12499630e-01 8.82033169e-01 -1.16653609e+00 -9.01684284e-01
6.81560189e-02 -4.55017567e-01 -1.16999006e+00 -2.36324206e-01
7.46327877e-01 -2.66080290e-01 -5.40460169e-01 1.25739729e+00
4.65322375e-01 -1.97655946e-01 7.07892179e-01 -1.31891572e+00
-8.07611942e-01 9.54749227e-01 -5.44314921e-01 6.03412032e-01
1.04509759e+00 3.43206301e-02 9.95086133e-01 -7.93805301e-01
9.05885577e-01 1.29187405e+00 7.73953736e-01 2.45025456e-01
-1.06179166e+00 -9.20297682e-01 -6.44976199e-01 3.30978751e-01
-1.06254530e+00 -5.26586950e-01 7.84989715e-01 -5.12211025e-01
7.85638094e-01 2.82897353e-01 6.58591986e-01 1.50819182e+00
2.29372710e-01 5.40624678e-01 1.10205960e+00 -4.76407230e-01
9.25499350e-02 2.35861003e-01 3.27395320e-01 4.69260186e-01
7.04734802e-01 -4.60273445e-01 -8.81073594e-01 -8.35989892e-01
4.08171058e-01 -2.26376683e-01 -1.36996418e-01 1.25136971e-01
-9.25869107e-01 1.32367992e+00 1.09042495e-01 3.93076420e-01
-2.58721501e-01 -2.43454993e-01 7.92398036e-01 5.04924834e-01
6.82781518e-01 1.20198107e+00 2.44784936e-01 -5.59042692e-01
-1.17670417e+00 6.87430739e-01 1.09179175e+00 1.03519559e+00
3.94002527e-01 -3.40274930e-01 -6.07844256e-03 9.58294034e-01
-6.48270175e-02 2.19475269e-01 8.28174412e-01 -1.01640344e+00
9.26277280e-01 6.99402034e-01 1.77069351e-01 -1.68268704e+00
7.12209493e-02 -1.51942134e-01 -2.79200286e-01 -3.50746274e-01
5.50760627e-01 3.09168287e-02 -3.04191321e-01 1.05396366e+00
7.01709911e-02 -2.33857363e-01 -1.09514751e-01 5.28434873e-01
7.86281586e-01 7.86621273e-01 -1.65086806e-01 -3.07841543e-02
1.46103263e+00 -5.98582745e-01 -6.33639336e-01 -5.28621733e-01
8.68288696e-01 -1.34081948e+00 1.19534814e+00 1.58794209e-01
-1.18465853e+00 5.08397296e-02 -9.23500538e-01 -2.06994221e-01
-3.10062468e-01 -3.75929445e-01 3.51312667e-01 8.52129400e-01
-5.68733096e-01 8.73795927e-01 -3.38381857e-01 -6.61989093e-01
6.84315264e-01 -3.95151049e-01 -2.21586347e-01 -2.23152459e-01
-1.27520287e+00 1.19732046e+00 2.09765330e-01 -4.60184395e-01
-3.07040304e-01 -8.53254855e-01 -5.36834180e-01 -3.07764988e-02
2.14603573e-01 -4.85522419e-01 1.46721959e+00 -9.31984007e-01
-9.84030664e-01 1.15743673e+00 -1.90744564e-01 -4.96935815e-01
4.46769357e-01 -4.33042616e-01 -2.25744262e-01 3.93305033e-01
5.14045954e-01 1.21740125e-01 9.23752189e-01 -6.69226944e-01
-3.24775428e-01 -2.63004512e-01 -1.34384394e-01 2.46533841e-01
-6.68902218e-01 7.17001200e-01 2.70579159e-01 -8.91562402e-01
-1.69508439e-02 -7.55929053e-01 1.46402061e-01 -1.13006592e-01
-4.73077148e-01 -3.08890909e-01 6.89449012e-01 -1.01043391e+00
1.46002746e+00 -1.88651788e+00 -1.15864769e-01 9.40430015e-02
4.05846596e-01 2.88832664e-01 3.31247523e-02 1.28552389e+00
2.38374487e-01 6.62232280e-01 -6.39398918e-02 2.94501744e-02
-3.34058613e-01 -7.39287362e-02 -6.51106656e-01 4.46402162e-01
-5.29118069e-02 8.36536646e-01 -1.11232948e+00 -5.26010633e-01
-1.31302804e-01 -2.49963477e-02 -2.17380196e-01 7.57033899e-02
-6.73066080e-02 -4.29458022e-02 -6.37779057e-01 6.37281537e-01
3.37967902e-01 -4.07705545e-01 -1.14839330e-01 1.73727989e-01
-7.12713376e-02 1.02277637e+00 -5.55395126e-01 1.29567599e+00
-1.88808486e-01 1.05664062e+00 -2.50166863e-01 -7.50943959e-01
8.83780956e-01 3.55555207e-01 1.29576670e-02 -4.30057287e-01
1.78313285e-01 3.93472046e-01 -1.18043847e-01 -5.86820006e-01
1.04543686e+00 -4.43390347e-02 -1.99008018e-01 1.05227864e+00
-2.72049814e-01 -3.01254928e-01 2.05636039e-01 7.70920753e-01
1.39475691e+00 -6.52641356e-01 3.72864127e-01 -1.27343491e-01
1.66922301e-01 6.57690704e-01 -4.11858298e-02 1.10577166e+00
-1.64281949e-01 4.90127206e-01 4.66189831e-01 -1.95698217e-01
-1.47161067e+00 -5.73084831e-01 -5.72435483e-02 7.20248580e-01
6.29854947e-02 -7.09604025e-01 -7.24728167e-01 -6.33551002e-01
2.40322754e-01 9.27322626e-01 -1.88483268e-01 -4.23548609e-01
-5.24935186e-01 -4.33035165e-01 1.20735788e+00 2.23214790e-01
2.54940569e-01 -1.21653652e+00 -5.19309461e-01 3.02351892e-01
-8.12028766e-01 -9.45575893e-01 -3.52527857e-01 -5.40715337e-01
-1.02439547e+00 -1.10532618e+00 -7.46427298e-01 -6.40101552e-01
4.15583760e-01 7.69662559e-01 1.20571065e+00 2.18040317e-01
-1.53742522e-01 2.17358321e-01 -6.69415414e-01 -3.39174896e-01
-1.32270241e+00 1.73027128e-01 -3.33643377e-01 -5.65403581e-01
5.98446667e-01 -5.73317349e-01 -3.67634147e-01 -3.73571515e-02
-9.01380122e-01 -3.30507569e-02 2.70507514e-01 5.71653366e-01
-3.96274298e-01 -6.69888929e-02 6.99178040e-01 -1.18632543e+00
1.48558068e+00 -1.00412786e+00 -1.11204602e-01 7.68868476e-02
-5.55210888e-01 -1.51328385e-01 5.50391614e-01 -5.46915531e-01
-9.08135474e-01 -5.66999912e-01 2.48516984e-02 -2.59172857e-01
-1.07171983e-01 8.75869811e-01 5.63833833e-01 -1.78585649e-01
1.20546520e+00 1.46714941e-01 1.80721387e-01 -5.55103898e-01
3.39419395e-01 1.06737983e+00 4.05535340e-01 -2.55580634e-01
8.20079744e-01 2.00801656e-01 -6.62284553e-01 -1.00629365e+00
-8.36319506e-01 -6.64284348e-01 1.03819855e-01 -2.63174772e-01
2.11310223e-01 -8.64612043e-01 -8.86909291e-02 3.75256032e-01
-1.47533977e+00 3.63337874e-01 -6.91794381e-02 5.89119233e-02
-1.92370534e-01 1.10546625e+00 -9.37901378e-01 -4.18503761e-01
-7.52757967e-01 -8.27359498e-01 5.27205646e-01 1.20516203e-01
-1.19729507e+00 -6.77710056e-01 2.79595971e-01 9.51341867e-01
4.08500105e-01 -2.06242222e-02 1.04252851e+00 -9.32890117e-01
-1.45164141e-02 -5.81328154e-01 -2.71411091e-01 -4.58808281e-02
9.30190645e-03 5.04933149e-02 -6.04523659e-01 -2.16529950e-01
2.10952848e-01 -5.83775520e-01 4.20260400e-01 -1.26898233e-02
7.11274445e-01 -9.70138133e-01 -2.65895188e-01 -2.97561903e-02
9.66491282e-01 -2.62707695e-02 5.36515474e-01 6.64581299e-01
4.58369702e-01 8.26088965e-01 4.76605505e-01 5.41936100e-01
4.64508161e-02 3.64111006e-01 6.93873540e-02 6.10133886e-01
4.65406217e-02 -6.33735955e-01 4.13162649e-01 9.38849092e-01
4.30382431e-01 -4.52249616e-01 -9.91158724e-01 7.05224097e-01
-1.70826054e+00 -1.47207487e+00 -3.18652689e-01 2.10735464e+00
1.09359205e+00 1.38328522e-01 2.38981858e-01 3.09332490e-01
1.04675245e+00 1.35314807e-01 -2.53551066e-01 -8.81072640e-01
-8.56687650e-02 1.62708223e-01 4.94097352e-01 4.62535657e-02
-5.55165768e-01 8.66427243e-01 6.87353468e+00 8.27100873e-01
-8.31429005e-01 -4.22516093e-02 2.49027923e-01 1.32436961e-01
-3.79564315e-01 2.15617165e-01 -5.66461682e-01 9.07031476e-01
1.01189315e+00 -7.79571116e-01 2.89870799e-01 9.21360850e-01
9.36408937e-02 -1.27439946e-01 -8.86205971e-01 8.41903269e-01
4.50977445e-01 -1.61657512e+00 2.58463919e-01 -1.64804250e-01
5.55128992e-01 -5.33342101e-02 -9.77833644e-02 2.33670380e-02
4.59912330e-01 -7.61288047e-01 5.62856615e-01 -2.16328070e-01
3.80220741e-01 -5.37595093e-01 4.50921953e-01 4.81111765e-01
-4.05542076e-01 5.31307571e-02 -5.77987015e-01 -3.36982220e-01
2.78340936e-01 6.94515526e-01 -1.00781786e+00 2.16554821e-01
4.47520077e-01 7.73867667e-01 -8.08888733e-01 1.27943182e+00
-2.93016523e-01 8.62242460e-01 -2.58014679e-01 -3.51318717e-01
7.41198361e-02 4.71620709e-02 9.16634619e-01 1.29209793e+00
4.37544644e-01 -2.02667043e-01 -1.60021082e-01 9.18666959e-01
-2.38232955e-01 1.31088242e-01 -9.41132307e-01 -8.40993404e-01
1.11320329e+00 1.15104687e+00 -6.55434549e-01 -4.85063940e-01
-3.71552259e-01 9.49863732e-01 3.43094885e-01 4.49313782e-02
-4.78126734e-01 -5.56258440e-01 3.42884004e-01 2.73934603e-01
-1.91162318e-01 -9.60623622e-02 -4.57498372e-01 -1.35040271e+00
-4.11623903e-02 -1.35991168e+00 5.70335150e-01 -8.88321519e-01
-1.79093599e+00 2.01201811e-01 1.79100603e-01 -9.62682366e-01
-4.43295032e-01 -2.79863812e-02 -8.99210095e-01 7.56225348e-01
-9.02026057e-01 -6.75503552e-01 -2.14904398e-01 2.07397372e-01
7.10133135e-01 -7.37074912e-02 5.62922478e-01 1.68205366e-01
-2.91201532e-01 6.44416928e-01 1.25644982e-01 2.12683469e-01
1.03170073e+00 -7.80084133e-01 9.81727540e-01 8.20500612e-01
2.69448936e-01 7.87945747e-01 1.01228642e+00 -1.19150293e+00
-1.31301892e+00 -6.88956499e-01 1.28961623e+00 -5.65479338e-01
1.19386399e+00 -1.29794821e-01 -1.27952158e+00 5.09403408e-01
2.05806166e-01 -9.24450815e-01 6.86766684e-01 1.34544700e-01
-8.05216908e-01 6.96282744e-01 -1.14010084e+00 7.04192162e-01
8.31640720e-01 -8.64883363e-01 -1.24946570e+00 6.79345369e-01
5.66191912e-01 -2.85347015e-01 -7.18788981e-01 -2.54081428e-01
5.44138491e-01 -7.22654402e-01 8.09536159e-01 -6.18074358e-01
1.18177164e+00 4.26268578e-01 4.89937723e-01 -1.30636168e+00
-2.80054629e-01 -8.15100372e-01 -5.06276339e-02 1.36443627e+00
1.81703344e-01 -6.06178045e-01 9.79897738e-01 7.00891852e-01
-3.12574506e-02 -1.08596370e-01 -8.31472337e-01 -6.88008666e-01
4.57486510e-01 -2.78680492e-02 1.89607009e-01 1.40670931e+00
7.30361760e-01 5.54451823e-01 -2.67662406e-01 -3.52862835e-01
5.37293077e-01 1.37199625e-01 8.05470765e-01 -1.14787650e+00
-1.05830856e-01 -6.50977075e-01 -3.22823673e-01 -6.27391517e-01
8.60630199e-02 -9.71551001e-01 -3.03288966e-01 -1.20410764e+00
4.96167660e-01 -1.61842749e-01 4.93053287e-01 1.60908237e-01
-2.64883906e-01 2.38860711e-01 2.25445762e-01 7.65460670e-01
-2.12144881e-01 2.13155091e-01 7.92012990e-01 -2.30770707e-02
-1.25074953e-01 -4.94097732e-02 -8.99693012e-01 6.88318849e-01
1.10038054e+00 -9.57973242e-01 -1.97608367e-01 -2.81736434e-01
7.05308259e-01 -2.54449043e-02 3.95979315e-01 -7.78363824e-01
2.41802320e-01 -1.44994274e-01 -5.04274741e-02 -4.24161494e-01
7.52075389e-03 -3.01444471e-01 4.80999313e-02 3.78137052e-01
-7.70103514e-01 6.17688656e-01 1.56501517e-01 5.91828346e-01
-5.11698425e-02 -8.33918333e-01 6.71454430e-01 -3.09545338e-01
-1.33492231e-01 -3.35059941e-01 -1.15703487e+00 6.64904237e-01
7.33193636e-01 -4.26042408e-01 -8.80302846e-01 -6.06276631e-01
5.97901531e-02 -2.35773157e-02 8.19803596e-01 5.94559968e-01
4.92445260e-01 -9.64696527e-01 -7.85041749e-01 -2.64113724e-01
2.99962282e-01 -5.30145407e-01 -3.20903063e-02 2.94485718e-01
-6.02448642e-01 1.31741285e-01 -4.59607571e-01 -7.15115666e-02
-1.50886321e+00 4.44472283e-01 -2.73518890e-01 -1.13458343e-01
-8.28363538e-01 5.08605421e-01 -2.28858754e-01 -3.96824814e-02
4.35620770e-02 2.77289450e-01 -2.66127586e-01 1.12719566e-01
6.70600593e-01 8.30599964e-01 -2.59701703e-02 -7.19445527e-01
-4.80801426e-02 -3.12114675e-02 -6.40310585e-01 -1.98761374e-01
9.24334347e-01 -1.67808443e-01 -2.25402400e-01 3.47087651e-01
1.21883094e+00 7.28081018e-02 -1.45657927e-01 -2.19263673e-01
3.46803725e-01 -7.12160528e-01 -1.46674573e-01 -5.82571626e-01
-3.23084563e-01 7.67487884e-01 -2.91112036e-01 5.25257587e-01
4.46656942e-01 2.37634685e-02 1.20586145e+00 4.14561510e-01
1.89287648e-01 -9.39197958e-01 2.99037129e-01 5.18001676e-01
1.00391424e+00 -9.59240139e-01 3.70128125e-01 -4.79498208e-01
-6.39125228e-01 1.13570809e+00 3.97719830e-01 -1.50311902e-01
1.63950846e-01 -5.06958133e-03 -8.21346864e-02 -4.52481300e-01
-6.11648142e-01 5.89882076e-01 9.76629108e-02 2.47693077e-01
5.77521086e-01 -3.26387107e-01 -8.94040108e-01 2.52777576e-01
-6.66724741e-01 -1.26648709e-01 1.17890990e+00 1.29147196e+00
-5.04769087e-01 -1.00529528e+00 -5.69788456e-01 8.87018859e-01
-8.88460159e-01 -2.53585249e-01 -1.09495676e+00 4.18737561e-01
-7.05503881e-01 1.14611888e+00 2.06725020e-02 -3.66750091e-01
-9.20629352e-02 1.66942298e-01 2.16574416e-01 -9.04565871e-01
-1.20232737e+00 -4.53060418e-01 4.79973704e-01 -3.52847785e-01
-1.51424214e-01 -7.07656264e-01 -7.57702470e-01 -1.03838825e+00
-5.79119861e-01 4.18942541e-01 5.08036613e-01 7.88253367e-01
7.07265794e-01 -5.20166278e-01 2.83476025e-01 -5.20547271e-01
-9.80361581e-01 -9.71940219e-01 -1.64096028e-01 8.30521166e-01
-1.86986551e-01 -3.81959051e-01 -6.13471687e-01 -1.30276218e-01] | [8.631721496582031, 10.0015287399292] |
97f0bec7-ef7d-49c7-9710-7b160bba50da | sequential-recommendation-with-diffusion | 2304.04541 | null | https://arxiv.org/abs/2304.04541v2 | https://arxiv.org/pdf/2304.04541v2.pdf | Sequential Recommendation with Diffusion Models | Generative models, such as Variational Auto-Encoder (VAE) and Generative Adversarial Network (GAN), have been successfully applied in sequential recommendation. These methods require sampling from probability distributions and adopt auxiliary loss functions to optimize the model, which can capture the uncertainty of user behaviors and alleviate exposure bias. However, existing generative models still suffer from the posterior collapse problem or the model collapse problem, thus limiting their applications in sequential recommendation. To tackle the challenges mentioned above, we leverage a new paradigm of the generative models, i.e., diffusion models, and present sequential recommendation with diffusion models (DiffRec), which can avoid the issues of VAE- and GAN-based models and show better performance. While diffusion models are originally proposed to process continuous image data, we design an additional transition in the forward process together with a transition in the reverse process to enable the processing of the discrete recommendation data. We also design a different noising strategy that only noises the target item instead of the whole sequence, which is more suitable for sequential recommendation. Based on the modified diffusion process, we derive the objective function of our framework using a simplification technique and design a denoise sequential recommender to fulfill the objective function. As the lengthened diffusion steps substantially increase the time complexity, we propose an efficient training strategy and an efficient inference strategy to reduce training and inference cost and improve recommendation diversity. Extensive experiment results on three public benchmark datasets verify the effectiveness of our approach and show that DiffRec outperforms the state-of-the-art sequential recommendation models. | ['Xiaofang Zhou', 'Pengpeng Zhao', 'Zhen Huang', 'Huanhuan Yuan', 'Hanwen Du'] | 2023-04-10 | null | null | null | null | ['sequential-recommendation'] | ['miscellaneous'] | [ 1.19403392e-01 -3.65515023e-01 -1.25503883e-01 -9.80488136e-02
-3.13744068e-01 -4.86849993e-01 6.75509751e-01 -5.98087311e-01
-2.14598626e-01 6.03772581e-01 3.06896210e-01 -2.77398080e-01
-1.33170441e-01 -9.67879713e-01 -7.73983777e-01 -1.05071592e+00
7.30240345e-01 2.82942235e-01 1.20877400e-01 -1.89924762e-01
9.62379351e-02 1.10633392e-03 -1.22130477e+00 1.11277901e-01
1.30171371e+00 8.88279915e-01 4.08136606e-01 2.07416326e-01
-2.42860317e-01 6.17511213e-01 -5.91998816e-01 -6.42135859e-01
2.32291490e-01 -8.91735494e-01 -1.77896351e-01 1.95079654e-01
-1.41722932e-01 -5.85547566e-01 -6.32880747e-01 1.09861660e+00
7.74542570e-01 4.56240445e-01 7.28424251e-01 -9.85261500e-01
-1.17458034e+00 8.11554134e-01 -4.60579872e-01 3.43347490e-02
1.05699845e-01 -4.83687334e-02 8.19252968e-01 -8.11960876e-01
2.44803786e-01 1.15629923e+00 7.14837730e-01 7.07311809e-01
-1.16234004e+00 -7.32544661e-01 4.80145305e-01 3.84724140e-02
-1.28179049e+00 -2.54460990e-01 8.00672591e-01 -1.70223758e-01
2.75720775e-01 2.37064078e-01 6.78458333e-01 1.55337262e+00
-1.65644893e-03 1.07400310e+00 1.02933562e+00 4.51394990e-02
2.50593841e-01 3.37296911e-02 -1.33666754e-01 4.46935177e-01
4.56623659e-02 1.31837815e-01 -1.40425414e-01 -4.67510372e-02
1.07481766e+00 8.34927440e-01 -4.75618869e-01 -1.32178947e-01
-9.70922053e-01 1.00635374e+00 3.27807099e-01 1.11591056e-01
-5.89278162e-01 1.39995813e-02 1.85119808e-01 3.73701066e-01
5.07848263e-01 3.17697995e-03 4.46193703e-02 4.81198207e-02
-1.07969892e+00 4.19497162e-01 6.08044326e-01 8.61608803e-01
3.68575066e-01 2.59741366e-01 -5.44719219e-01 9.83014524e-01
4.26589429e-01 6.38855755e-01 6.12495363e-01 -5.61673582e-01
3.87451768e-01 2.11872682e-01 1.54060498e-01 -1.00022221e+00
8.52607936e-02 -8.83797348e-01 -1.25964153e+00 -4.21247274e-01
2.32937738e-01 -2.99755305e-01 -1.02291059e+00 1.64460909e+00
2.90081590e-01 5.73515356e-01 -1.00174613e-01 9.93474185e-01
6.49990797e-01 9.04530764e-01 -1.28845781e-01 -4.14149761e-01
9.32670414e-01 -1.11458838e+00 -8.53752077e-01 2.00809091e-01
1.67068973e-01 -6.99013889e-01 1.02634251e+00 5.33529282e-01
-1.06189299e+00 -7.09513664e-01 -1.02052200e+00 2.41992146e-01
1.13692790e-01 1.16701327e-01 6.48204207e-01 6.69837356e-01
-6.95733726e-01 6.78366244e-01 -8.62065315e-01 6.51102811e-02
3.50693434e-01 2.09329352e-01 2.94841826e-01 -2.67195135e-01
-1.19179583e+00 3.06379586e-01 1.32688507e-03 3.96370739e-01
-9.97418165e-01 -6.00388885e-01 -4.77862507e-01 1.59751356e-01
6.34436607e-01 -9.36875999e-01 1.08007276e+00 -9.65569675e-01
-2.06979227e+00 -1.50008023e-01 -1.03298007e-02 -3.29746962e-01
6.38184547e-01 -3.13739479e-01 -5.78800619e-01 -2.69514620e-01
-2.88798481e-01 1.31886080e-01 1.28512144e+00 -1.26349914e+00
-6.13629460e-01 5.41953556e-02 1.96076736e-01 1.26722321e-01
-5.79022348e-01 -2.81140089e-01 -7.30354369e-01 -1.20384312e+00
-1.11021638e-01 -1.07756257e+00 -4.04383302e-01 -5.54486573e-01
-2.84384251e-01 -5.05681932e-02 6.62932038e-01 -7.42570281e-01
1.63168645e+00 -2.18731666e+00 3.92627090e-01 3.11504990e-01
1.94406435e-01 3.93763483e-01 -1.48672923e-01 3.32834542e-01
4.23068106e-01 1.00426413e-01 -1.28870219e-01 -6.94465101e-01
8.04736093e-02 5.13709784e-01 -5.28174400e-01 2.66590893e-01
-3.10300082e-01 9.66194034e-01 -9.13576186e-01 -1.48647144e-01
2.32730359e-01 6.73848391e-01 -9.95364368e-01 3.83301944e-01
-3.52722466e-01 6.39401257e-01 -6.44359350e-01 2.53685921e-01
7.71128833e-01 -3.90330553e-01 5.21710962e-02 -2.02444300e-01
1.42729878e-01 2.87139773e-01 -1.30859804e+00 1.62096274e+00
-6.41940355e-01 -2.80288696e-01 -3.76297347e-02 -8.88740420e-01
9.65565085e-01 2.92481989e-01 3.18578571e-01 -6.16901338e-01
7.64111653e-02 8.19731429e-02 -1.37333274e-01 -3.02800477e-01
5.31454623e-01 -2.05805138e-01 8.55282620e-02 5.18061996e-01
-3.75512280e-02 2.30140910e-01 -1.48229182e-01 1.33779451e-01
7.18685627e-01 3.47396046e-01 -1.57815441e-01 1.41002983e-01
4.96754616e-01 -6.24328792e-01 8.69130969e-01 9.01396394e-01
4.29087609e-01 6.75725043e-01 2.39975885e-01 -1.66512728e-01
-8.62279594e-01 -9.04650927e-01 2.20867559e-01 8.94577086e-01
3.28798980e-01 -4.71106023e-01 -6.15092933e-01 -9.56848919e-01
-1.82723179e-01 8.72847676e-01 -6.20989084e-01 -3.46442670e-01
-4.97376293e-01 -1.10510397e+00 1.19897120e-01 5.14866590e-01
4.75934923e-01 -8.03464830e-01 7.16962293e-02 4.03872043e-01
-2.63657987e-01 -7.24482715e-01 -9.38593686e-01 -2.50762403e-01
-7.86584973e-01 -3.92958373e-01 -1.07670307e+00 -4.06673640e-01
6.64312243e-01 4.16616976e-01 6.61442816e-01 2.67686509e-03
5.50334573e-01 3.02239686e-01 -7.78533816e-01 -6.03109859e-02
-4.71854508e-01 6.20872900e-02 5.18920971e-03 4.30347919e-01
-2.40054838e-02 -8.39653552e-01 -1.04221976e+00 5.32939553e-01
-1.12077641e+00 3.57299559e-02 6.38983727e-01 1.10656416e+00
7.71277487e-01 1.72868580e-01 7.54941702e-01 -9.63179588e-01
7.60675550e-01 -7.77030885e-01 -4.41539675e-01 1.71395406e-01
-9.65278089e-01 -8.19895193e-02 1.05990708e+00 -8.81802380e-01
-1.23407459e+00 -4.24039632e-01 -4.03697342e-01 -8.73361528e-01
1.75820515e-01 5.82987487e-01 -2.51294285e-01 1.63543031e-01
2.02852890e-01 6.10158503e-01 -1.07148616e-02 -7.70030499e-01
4.60466266e-01 5.74551582e-01 2.87983984e-01 -3.19344223e-01
7.68964946e-01 4.12894040e-01 -1.93576351e-01 -2.74868965e-01
-9.73009408e-01 -1.33734986e-01 4.84096818e-02 -4.08938192e-02
5.28686702e-01 -8.86584818e-01 -4.68665957e-01 4.32411760e-01
-8.19118738e-01 -2.32956007e-01 -3.15251052e-01 6.06120110e-01
-3.24209332e-01 5.49936116e-01 -7.18068719e-01 -6.72362089e-01
-5.79475999e-01 -1.14317548e+00 7.58898616e-01 3.64805311e-01
1.73645139e-01 -9.85436916e-01 5.41591039e-03 1.57195076e-01
6.23911381e-01 -8.44357312e-02 6.48112655e-01 -6.50915027e-01
-5.81135094e-01 -6.04125895e-02 2.49309149e-02 6.20950460e-01
1.37734935e-01 -2.90191203e-01 -4.88367617e-01 -4.35691059e-01
2.95990646e-01 6.39486164e-02 9.19820845e-01 3.39226037e-01
1.43045807e+00 -3.91763717e-01 -1.91867888e-01 7.99066246e-01
1.35088027e+00 3.36189181e-01 8.73052597e-01 1.66755304e-01
9.06283855e-01 3.48896272e-02 4.57285851e-01 6.04502082e-01
4.16418433e-01 6.48670673e-01 2.27544785e-01 6.87819198e-02
-9.73734632e-02 -6.43797338e-01 6.46970987e-01 1.26103365e+00
-4.12599802e-01 -7.49028504e-01 -1.52214944e-01 2.80807883e-01
-2.07358694e+00 -1.06128681e+00 1.58244699e-01 2.25397348e+00
6.57579720e-01 -3.19041200e-02 1.69532299e-01 -6.63615614e-02
6.29283249e-01 2.03502744e-01 -5.48629522e-01 -7.34730586e-02
6.07691519e-02 2.31273115e-01 3.45209122e-01 1.89082339e-01
-8.15451801e-01 7.65852511e-01 5.77712727e+00 1.31812871e+00
-1.16267562e+00 3.54508191e-01 4.48161900e-01 -2.36537203e-01
-7.85574436e-01 -9.98476222e-02 -8.80848885e-01 1.00326908e+00
7.71587968e-01 8.61397907e-02 8.42676461e-01 6.09663785e-01
2.74729967e-01 4.14504260e-01 -8.15069318e-01 1.01423442e+00
1.24289908e-01 -1.13915777e+00 2.54617304e-01 1.97606832e-01
1.03823137e+00 -2.09440812e-01 3.33957583e-01 5.97261786e-01
4.73443002e-01 -8.04335356e-01 6.31033719e-01 1.00069225e+00
4.85870272e-01 -9.26929653e-01 8.23388636e-01 4.68867034e-01
-9.37517405e-01 -6.87906146e-02 -4.88481551e-01 1.65576130e-01
4.73436922e-01 7.33160317e-01 -7.57904276e-02 9.31871235e-01
5.53847015e-01 9.25544918e-01 -1.26693457e-01 9.24213707e-01
-3.99976492e-01 9.88433719e-01 -2.48610258e-01 -6.19796440e-02
1.34374261e-01 -7.80583858e-01 6.16703570e-01 8.77336621e-01
7.15846479e-01 1.12913519e-01 2.57481694e-01 9.12404895e-01
-7.89005980e-02 1.06482930e-01 -6.44503832e-02 -4.86857705e-02
4.45222735e-01 1.05145144e+00 -5.05881667e-01 -2.00220883e-01
-6.40220821e-01 1.32560718e+00 1.13208771e-01 5.81411898e-01
-1.16620362e+00 -6.95900619e-02 2.72402048e-01 -2.80817337e-02
9.07866955e-01 -1.11560255e-01 -2.88989432e-02 -1.41165698e+00
-4.19302844e-02 -1.17975163e+00 1.24819353e-01 -2.55550295e-01
-1.52111042e+00 3.87480468e-01 -1.38756260e-01 -1.39641166e+00
-1.57407001e-01 -3.51024745e-03 -7.28613913e-01 7.79453218e-01
-1.39166451e+00 -1.13546884e+00 -1.47808790e-01 7.23157465e-01
5.41735947e-01 -1.01143315e-01 5.20504057e-01 6.41542971e-01
-7.92437375e-01 8.93005550e-01 5.95738947e-01 -3.60517770e-01
5.38960516e-01 -9.98572469e-01 2.24761993e-01 8.53791654e-01
1.26057997e-01 9.58178282e-01 5.40226400e-01 -6.28342569e-01
-1.42858374e+00 -1.26133788e+00 2.42695168e-01 -1.98382929e-01
4.35806274e-01 -3.45638096e-01 -8.59484255e-01 5.91014564e-01
2.26835497e-02 -2.83906460e-01 7.32900321e-01 -3.03148739e-02
9.12038237e-03 -9.01478976e-02 -8.59419525e-01 8.02401662e-01
1.22057498e+00 -1.93976149e-01 -5.10477304e-01 2.96698231e-02
8.54163289e-01 -2.98653603e-01 -9.93834078e-01 4.05381382e-01
6.00517273e-01 -8.78278732e-01 9.97376382e-01 -3.02578002e-01
5.34489810e-01 -3.52125406e-01 1.07364450e-02 -1.53747499e+00
-6.72729373e-01 -9.60938811e-01 -5.94994426e-01 1.55946517e+00
2.06969857e-01 -7.39074409e-01 5.05508542e-01 1.98784322e-01
-1.43399328e-01 -9.57929373e-01 -5.22625744e-01 -7.02290297e-01
-9.53563154e-02 -1.70907184e-01 8.48802090e-01 7.75674582e-01
-5.92376888e-01 3.39197606e-01 -9.69048083e-01 3.86954881e-02
5.52810133e-01 3.98191065e-01 8.39167833e-01 -8.97287488e-01
-7.50382900e-01 -3.30573559e-01 2.02642992e-01 -1.85931540e+00
-6.28643110e-02 -6.07092619e-01 -3.35880741e-02 -1.50576663e+00
1.24777324e-01 -5.79107106e-01 -5.95257044e-01 1.14758082e-01
-4.69261885e-01 9.85898972e-02 6.87229410e-02 4.54543620e-01
-4.65766877e-01 1.17140651e+00 1.58776009e+00 1.16096683e-01
-2.69365430e-01 4.33897644e-01 -8.50216150e-01 5.35460413e-01
5.78816533e-01 -5.21088302e-01 -7.60100126e-01 -3.57218623e-01
1.84055001e-01 -1.94299761e-02 2.47318879e-01 -6.14732325e-01
1.06554456e-01 -1.51352420e-01 2.71750778e-01 -5.51547229e-01
1.05241448e-01 -8.86908591e-01 4.68832701e-01 2.72260040e-01
-1.68018281e-01 -1.99374944e-01 -1.72998965e-01 1.05254710e+00
-7.96701089e-02 -1.77461833e-01 4.82634127e-01 3.20642032e-02
-2.55719513e-01 6.77372038e-01 -1.43206760e-01 -3.43983084e-01
7.56625295e-01 9.06230584e-02 -4.12959084e-02 -5.91050804e-01
-7.61776805e-01 2.23275736e-01 4.41113770e-01 5.31684756e-01
5.52919805e-01 -1.51167023e+00 -6.40785515e-01 1.09163947e-01
-3.82662117e-01 -6.02637976e-02 5.71209192e-01 1.01223254e+00
-6.06437065e-02 1.21012233e-01 2.27599204e-01 -1.99997097e-01
-6.98080838e-01 8.78064990e-01 8.82739350e-02 -5.31587780e-01
-7.15683699e-01 7.54008472e-01 4.56373185e-01 -2.55780220e-01
9.27223638e-02 -1.59771845e-01 -3.62210393e-01 -4.95188795e-02
4.49509710e-01 4.13718700e-01 -2.80768871e-01 -2.77640194e-01
1.00201502e-01 3.17529500e-01 -4.07741755e-01 3.11600026e-02
1.35770190e+00 -5.23462117e-01 1.00215390e-01 2.63019264e-01
8.37007105e-01 3.04950416e-01 -1.36732936e+00 -4.96788949e-01
-7.06227660e-01 -5.24522662e-01 4.37040359e-01 -6.03972375e-01
-1.33943129e+00 5.74103713e-01 4.75840360e-01 3.35419595e-01
1.21255779e+00 -3.47178966e-01 1.29340446e+00 -7.91224986e-02
2.47995719e-01 -8.78656983e-01 -6.83924509e-03 3.30042720e-01
6.98252857e-01 -8.85173440e-01 -4.86670174e-02 -2.95274913e-01
-6.58901751e-01 7.28315651e-01 4.06628758e-01 -3.22838426e-01
6.40966296e-01 2.24164464e-02 -3.20657015e-01 1.80389062e-01
-5.16401172e-01 -1.21785877e-02 4.07416642e-01 2.42301449e-01
9.93656442e-02 -2.63026003e-02 -6.27021074e-01 1.20236945e+00
1.04395941e-01 2.20157757e-01 1.36039734e-01 5.83516896e-01
-6.39721900e-02 -1.35178757e+00 -1.88906983e-01 5.56227326e-01
-4.89421070e-01 -2.55426794e-01 1.56965479e-01 3.34582001e-01
2.22396165e-01 9.87693012e-01 -9.89065543e-02 -6.96478605e-01
2.81134844e-01 -8.03766176e-02 3.83675635e-01 -3.61227542e-01
-5.39035678e-01 5.78240931e-01 -2.07049489e-01 -2.60341853e-01
-4.72842216e-01 -7.09021270e-01 -7.48537719e-01 -4.99236852e-01
-7.64229596e-01 2.66190767e-01 3.31296027e-01 1.02702296e+00
5.35505712e-01 8.81202281e-01 9.66425538e-01 -7.71277070e-01
-9.71439481e-01 -1.06223214e+00 -5.95129430e-01 4.06949580e-01
1.14645466e-01 -7.05605507e-01 -3.18998307e-01 -1.13123417e-01] | [10.24290657043457, 5.51597785949707] |
fae7e9dc-24c7-4ea7-8b36-e3510e818db6 | a-deeper-autoregressive-approach-to-non | 2305.12510 | null | https://arxiv.org/abs/2305.12510v1 | https://arxiv.org/pdf/2305.12510v1.pdf | A Deeper (Autoregressive) Approach to Non-Convergent Discourse Parsing | Online social platforms provide a bustling arena for information-sharing and for multi-party discussions. Various frameworks for dialogic discourse parsing were developed and used for the processing of discussions and for predicting the productivity of a dialogue. However, most of these frameworks are not suitable for the analysis of contentious discussions that are commonplace in many online platforms. A novel multi-label scheme for contentious dialog parsing was recently introduced by Zakharov et al. (2021). While the schema is well developed, the computational approach they provide is both naive and inefficient, as a different model (architecture) using a different representation of the input, is trained for each of the 31 tags in the annotation scheme. Moreover, all their models assume full knowledge of label collocations and context, which is unlikely in any realistic setting. In this work, we present a unified model for Non-Convergent Discourse Parsing that does not require any additional input other than the previous dialog utterances. We fine-tuned a RoBERTa backbone, combining embeddings of the utterance, the context and the labels through GRN layers and an asymmetric loss function. Overall, our model achieves results comparable with SOTA, without using label collocation and without training a unique architecture/model for each label. | ['Oren Tsur', 'Yoav Tulpan'] | 2023-05-21 | null | null | null | null | ['discourse-parsing'] | ['natural-language-processing'] | [-1.83428228e-01 6.17588043e-01 -2.03583650e-02 -7.00051248e-01
-5.32042921e-01 -8.65629077e-01 8.22298229e-01 5.10142267e-01
-5.55090785e-01 7.49031842e-01 5.96041083e-01 -1.87334821e-01
2.23150834e-01 -6.76006615e-01 -2.12155238e-01 -4.06609982e-01
4.01018232e-01 8.87683272e-01 1.61412314e-01 -5.07864356e-01
4.77359779e-02 4.74458672e-02 -1.27989912e+00 4.40618277e-01
5.86603403e-01 8.98328125e-01 4.15185988e-01 9.09756303e-01
-6.02916658e-01 1.05426002e+00 -7.64670074e-01 -7.20414400e-01
-1.81860611e-01 -4.22423959e-01 -1.35476744e+00 1.80388261e-02
1.65324673e-01 -2.97470033e-01 -1.53129905e-01 6.04338408e-01
4.15962309e-01 1.74607724e-01 7.27468550e-01 -8.69543493e-01
-4.19077933e-01 9.90321398e-01 7.92765766e-02 -5.36301695e-02
4.55824077e-01 -3.73935461e-01 1.39656937e+00 -7.01156199e-01
8.17065656e-01 1.31027782e+00 6.96498990e-01 7.68124580e-01
-9.39734459e-01 -3.47857088e-01 1.35213539e-01 -1.62521899e-02
-7.57390380e-01 -3.28625500e-01 7.25101352e-01 -5.80323279e-01
7.96690822e-01 3.16686273e-01 4.12176996e-01 1.41312993e+00
-1.46394759e-01 7.21924365e-01 9.59676266e-01 -6.66764379e-01
-1.91102102e-02 4.09388930e-01 5.87949753e-01 5.88111162e-01
-4.58436698e-01 -6.87986612e-01 -5.28154492e-01 -3.24274749e-01
2.22767398e-01 -1.55523568e-01 4.44899648e-02 -1.06595524e-01
-7.75156140e-01 1.17557120e+00 2.06638500e-01 7.96050429e-01
-3.23785804e-02 -1.82238042e-01 7.01714218e-01 4.29752678e-01
7.56253779e-01 4.95030880e-01 -3.99137348e-01 -3.10486823e-01
-5.49015164e-01 3.43282253e-01 1.34174132e+00 7.47823596e-01
6.82346761e-01 -6.57743692e-01 -3.99329275e-01 9.39810693e-01
3.97487611e-01 -9.52253267e-02 5.02388656e-01 -1.22132480e+00
6.02988601e-01 9.05850708e-01 2.45064482e-01 -8.08496177e-01
-9.51514840e-01 -1.84098743e-02 -5.44958115e-01 -2.42836162e-01
1.10178220e+00 -5.71510434e-01 -2.15498060e-01 1.92756391e+00
3.81351501e-01 -1.36400163e-01 2.34377563e-01 7.05691338e-01
1.15291703e+00 7.15631187e-01 2.39618674e-01 -1.27227396e-01
1.48485613e+00 -1.19412577e+00 -1.04457223e+00 -1.97033226e-01
9.24214184e-01 -1.03724122e+00 1.06211090e+00 1.38727322e-01
-1.09958839e+00 -3.26580167e-01 -7.27262139e-01 -7.06439912e-01
-5.05912483e-01 2.08390683e-01 6.01007760e-01 6.20943487e-01
-1.03776407e+00 4.74334091e-01 -4.52592850e-01 -4.52338636e-01
-7.18121082e-02 2.17873424e-01 -1.93046272e-01 3.37062508e-01
-1.39783978e+00 1.20525634e+00 1.61699027e-01 -1.39990775e-02
-3.89427930e-01 -3.36838424e-01 -8.14458430e-01 1.34079739e-01
2.57232189e-01 -2.87658304e-01 1.56558716e+00 -8.82530153e-01
-1.94145775e+00 1.17682910e+00 -3.52797098e-02 -4.75316107e-01
6.29454613e-01 -4.07577753e-01 9.76741239e-02 -1.93975925e-01
-1.84901893e-01 6.00325644e-01 3.32615107e-01 -1.04261386e+00
-4.76056159e-01 -1.77203387e-01 7.57218778e-01 4.78704721e-01
-4.03490722e-01 2.27243274e-01 -2.49066010e-01 -3.64564389e-01
-2.44275741e-02 -1.00841284e+00 -9.90015045e-02 -3.23422015e-01
-3.54277223e-01 -6.72637224e-01 7.21503973e-01 -9.15097713e-01
1.28375101e+00 -2.08521128e+00 3.85450989e-01 -3.42851549e-01
2.66117066e-01 2.81717926e-01 1.46920949e-01 8.69371772e-01
4.94666189e-01 9.78742242e-02 -1.39043584e-01 -9.65593219e-01
2.80425251e-01 1.20328240e-01 -2.68771380e-01 2.34190285e-01
8.76484439e-02 6.92164898e-01 -8.53653133e-01 -3.85606915e-01
1.68251067e-01 4.42950130e-01 -5.49364507e-01 5.98150730e-01
-5.57754159e-01 7.33020961e-01 -3.21510285e-01 1.01870649e-01
3.00982207e-01 -3.00692677e-01 8.04829180e-01 9.27845240e-02
-3.09794992e-01 7.13630438e-01 -8.68170261e-01 1.83046937e+00
-8.84873927e-01 7.62079477e-01 4.06572521e-01 -7.88724720e-01
1.00663745e+00 6.44856513e-01 3.65603000e-01 -4.03564870e-01
5.79396367e-01 6.84121251e-02 -6.31488189e-02 -5.70010185e-01
7.35015154e-01 -1.62561275e-02 -5.95904231e-01 8.91766965e-01
3.74417037e-01 4.86643985e-02 2.37761199e-01 2.28775501e-01
9.49722946e-01 -4.97583486e-02 1.21985339e-01 -2.38140732e-01
7.18609989e-01 -1.50630131e-01 4.53353077e-01 4.86435294e-01
-1.71295136e-01 5.14234841e-01 8.95782948e-01 -3.82440150e-01
-1.06937647e+00 -6.70799553e-01 -1.09819956e-01 1.67813349e+00
3.43812890e-02 -5.02233446e-01 -1.00314617e+00 -7.72745788e-01
-2.47940272e-01 6.02316737e-01 -4.92353410e-01 3.65094900e-01
-8.79758775e-01 -4.40290362e-01 6.05291247e-01 1.19908884e-01
4.07238185e-01 -1.14803123e+00 -5.09071410e-01 3.92218590e-01
-6.17022812e-01 -1.19132507e+00 -3.41059744e-01 1.54471546e-01
-4.64201927e-01 -1.19379675e+00 -4.70631868e-01 -1.03280091e+00
2.19788626e-01 -1.20407924e-01 1.17100179e+00 1.13031290e-01
1.09928772e-01 3.70879173e-01 -4.80171710e-01 -2.41738975e-01
-8.57099950e-01 4.21642572e-01 -2.70594805e-01 3.06972396e-02
3.61640662e-01 -2.76812792e-01 -3.32111925e-01 2.13019401e-01
-5.87459207e-01 8.79861712e-02 9.79111437e-03 9.77664292e-01
-2.25579172e-01 -6.16139472e-01 8.39538217e-01 -1.31664860e+00
1.00476134e+00 -6.81086957e-01 -2.51120299e-01 2.95641989e-01
-1.27983317e-01 -2.68769544e-02 6.96233630e-01 -2.63662517e-01
-1.30884147e+00 -2.01506525e-01 -5.36967278e-01 2.22846612e-01
-1.33072823e-01 2.69431263e-01 -5.21961786e-02 3.58526200e-01
5.64887643e-01 -5.34848034e-01 6.92666620e-02 -8.15431893e-01
5.50424874e-01 1.05349457e+00 1.23392254e-01 -6.66482210e-01
1.21195063e-01 9.84448865e-02 -4.96814072e-01 -8.11326802e-01
-1.08807099e+00 -6.26398146e-01 -6.70549333e-01 -2.49467805e-01
1.31897545e+00 -7.11213410e-01 -1.00707102e+00 4.88146693e-01
-1.61854243e+00 -4.80967194e-01 -1.21614337e-01 3.20359290e-01
-5.39216042e-01 5.04658818e-01 -9.73673880e-01 -9.17424083e-01
-2.52938360e-01 -1.07586777e+00 6.95859611e-01 3.28561604e-01
-6.20678842e-01 -1.33293247e+00 1.59301907e-01 6.87955439e-01
5.65244973e-01 1.26756012e-01 1.04780138e+00 -1.15266860e+00
-9.49297175e-02 3.56442877e-03 -2.74146348e-01 4.27867323e-01
1.14723347e-01 -2.45615065e-01 -1.30160129e+00 3.88722159e-02
2.24586636e-01 -8.84065866e-01 5.62693417e-01 7.16289654e-02
7.55066037e-01 -3.17409068e-01 -6.49370775e-02 -1.30648911e-01
8.58296096e-01 -7.04540759e-02 9.64709744e-02 1.92207962e-01
4.88056928e-01 1.10990894e+00 4.30927992e-01 4.96124297e-01
8.07340860e-01 7.05109000e-01 3.72983634e-01 1.33411542e-01
2.01801658e-02 -2.94765104e-02 3.20262015e-01 1.09749758e+00
1.05405487e-01 -5.98744571e-01 -9.92039621e-01 4.48409110e-01
-1.96635664e+00 -6.38846934e-01 -1.98548228e-01 1.88972723e+00
1.10826719e+00 -1.09030614e-02 1.52116463e-01 -2.00439870e-01
7.50863314e-01 3.68360698e-01 -1.84502512e-01 -7.73578286e-01
-1.44954875e-01 9.62342545e-02 1.15935050e-01 9.15305674e-01
-1.27104950e+00 9.09673512e-01 6.13392591e+00 3.66118580e-01
-8.25929046e-01 5.41563451e-01 6.28845215e-01 2.88962752e-01
-1.69763252e-01 1.26388529e-02 -8.97543609e-01 4.09386218e-01
1.26834965e+00 8.88165459e-02 2.28541464e-01 6.35490358e-01
3.28226052e-02 -1.10813305e-01 -1.18216777e+00 6.15033925e-01
1.63936131e-02 -1.31338096e+00 -4.50921685e-01 -2.78435767e-01
3.79259706e-01 -4.11456488e-02 -3.38197589e-01 4.75859761e-01
5.46324790e-01 -8.43845546e-01 6.41652524e-01 2.90516555e-01
2.52056658e-01 -2.79341787e-01 9.47907567e-01 5.86612046e-01
-6.63245559e-01 -2.52508730e-01 -1.57772496e-01 -3.51869285e-01
3.55505645e-01 1.70887694e-01 -9.21027660e-01 3.93373847e-01
2.93170005e-01 3.39052200e-01 -2.85861552e-01 3.59000385e-01
-2.69907385e-01 5.99774778e-01 -2.63060123e-01 -2.14029193e-01
4.87321466e-01 -2.04952940e-01 3.45084339e-01 1.46681452e+00
1.02300599e-01 -6.56620711e-02 3.96767050e-01 4.74416077e-01
-3.65274727e-01 4.33936477e-01 -3.98686022e-01 -7.98785314e-02
6.45396531e-01 1.29862010e+00 -4.36363518e-01 -1.90973476e-01
-5.83845258e-01 6.58229947e-01 7.24408209e-01 -5.06077567e-03
-6.13077581e-01 8.39589760e-02 4.64014024e-01 1.72850061e-02
-7.64962584e-02 -2.82528162e-01 -2.28723630e-01 -9.50695693e-01
-8.15316290e-02 -5.77582419e-01 5.03848016e-01 -1.41447634e-01
-1.41439748e+00 6.36545718e-01 -7.71492571e-02 -7.76745200e-01
-5.88832855e-01 -5.86862028e-01 -5.66332996e-01 9.06830192e-01
-1.42062044e+00 -1.24718797e+00 -2.29680791e-01 2.51660228e-01
7.29251742e-01 -6.69199899e-02 1.26173961e+00 4.04192507e-01
-5.25125325e-01 4.51470762e-01 3.55518609e-02 3.14312190e-01
1.03923702e+00 -1.41922319e+00 3.60399224e-02 1.44623429e-01
-1.22740984e-01 3.43176782e-01 7.87610173e-01 -2.77191848e-01
-9.90363955e-01 -6.90243483e-01 1.56055129e+00 -4.45084751e-01
9.77082074e-01 -7.38672376e-01 -9.74928975e-01 6.77525163e-01
8.43789160e-01 -5.73418319e-01 9.41594243e-01 5.43250084e-01
-1.44713029e-01 3.13267618e-01 -1.06996107e+00 3.63855869e-01
7.29930103e-01 -5.72449684e-01 -8.10786605e-01 6.20481372e-01
8.25198412e-01 -6.75268173e-01 -9.77361679e-01 -8.07943419e-02
3.98782104e-01 -1.03536892e+00 5.50710082e-01 -4.66686219e-01
5.23376822e-01 2.79377103e-01 -2.03148603e-01 -9.65102196e-01
6.38804510e-02 -5.41929543e-01 3.37071382e-02 1.63182485e+00
5.68987906e-01 -4.77353066e-01 7.15607047e-01 7.76271880e-01
-3.25685918e-01 -5.52790642e-01 -1.05836761e+00 -2.89626896e-01
4.02880639e-01 -1.30021244e-01 3.85910124e-01 1.06254876e+00
3.30808729e-01 8.30487728e-01 -4.49280977e-01 -2.59807080e-01
3.46670449e-02 2.65472513e-02 7.40493476e-01 -1.56619966e+00
-2.08019122e-01 -4.91029680e-01 1.03914835e-01 -1.05243421e+00
5.39126039e-01 -9.12681699e-01 1.12924827e-02 -1.61661804e+00
-1.42859980e-01 -8.28630149e-01 3.38198654e-02 2.35927001e-01
3.91734019e-02 8.71952157e-03 3.19261044e-01 2.76096076e-01
-5.58375359e-01 4.68186915e-01 9.15524125e-01 -6.82876855e-02
-3.30640316e-01 -2.07426976e-02 -4.63339716e-01 9.39813614e-01
8.38144481e-01 -3.77756596e-01 -2.49114171e-01 -4.50439155e-01
3.52234364e-01 2.99607605e-01 7.80055113e-03 -6.36955976e-01
3.09510529e-01 7.55994022e-02 -3.03800076e-01 -4.34605777e-01
6.94456339e-01 -6.52242482e-01 -3.16247120e-02 3.72936688e-02
-6.96406543e-01 -1.42712936e-01 -1.42007798e-01 4.45311308e-01
-3.34800184e-01 -8.37616801e-01 6.30532682e-01 -2.43835494e-01
-4.02979583e-01 -3.24438930e-01 -4.12641048e-01 2.67876208e-01
9.08517480e-01 3.38220507e-01 -5.55242777e-01 -4.62180078e-01
-1.02548969e+00 3.53638232e-01 3.98715615e-01 5.65411508e-01
-1.83548927e-01 -9.34634626e-01 -7.22062290e-01 -2.69978970e-01
-2.39817560e-01 1.73945114e-01 3.76879096e-01 5.32303274e-01
-4.66884077e-01 2.57317126e-01 -1.51318893e-01 -3.59140962e-01
-1.18863523e+00 1.00570321e-01 1.56202644e-01 -8.74422073e-01
-4.87087488e-01 7.32017636e-01 1.47379220e-01 -8.66220176e-01
4.82242048e-01 -1.07662410e-01 -7.40808904e-01 8.41077626e-01
4.46942568e-01 2.57009059e-01 -1.23826722e-02 -7.36404240e-01
-7.00565651e-02 1.56086311e-01 -1.20857053e-01 -2.30896294e-01
1.23093009e+00 -4.74646986e-01 -2.09753990e-01 1.05803394e+00
9.96036291e-01 6.49446175e-02 -1.10254073e+00 -4.65377241e-01
3.32320929e-01 1.19521126e-01 -2.40549460e-01 -6.40075028e-01
-5.56270003e-01 9.94550645e-01 1.61550328e-01 7.84714162e-01
6.19908929e-01 1.69576719e-01 7.83041775e-01 4.96827185e-01
7.34541640e-02 -1.36794162e+00 1.39204472e-01 1.09405732e+00
8.44198227e-01 -1.34195220e+00 -3.53427708e-01 -4.49018002e-01
-7.83541322e-01 1.30950022e+00 4.76014018e-01 3.13055627e-02
6.64141476e-01 1.18529692e-01 3.96751195e-01 -2.20683753e-01
-8.32755625e-01 -6.61401078e-02 -1.91919193e-01 2.90080488e-01
9.54880297e-01 5.29091433e-02 -7.86864638e-01 8.12192440e-01
-2.42718518e-01 -4.04167533e-01 6.96471989e-01 7.87504673e-01
-4.96497512e-01 -1.41541934e+00 -1.42095853e-02 2.39251718e-01
-6.18616104e-01 1.47627607e-01 -6.28300428e-01 6.51623666e-01
2.77058072e-02 1.14492869e+00 2.43280321e-01 -2.72057474e-01
2.15909109e-01 6.17986917e-01 5.91043197e-02 -9.25263405e-01
-1.26627183e+00 -1.98655009e-01 7.01473534e-01 -9.05494243e-02
-7.07314849e-01 -7.02658057e-01 -1.20013773e+00 -3.89072090e-01
-2.88901031e-01 3.90854150e-01 7.73735762e-01 1.11360204e+00
1.54245362e-01 3.74389768e-01 7.04116523e-01 -8.10723901e-01
-4.84910071e-01 -1.36015260e+00 -3.29761595e-01 4.24185961e-01
-1.94537137e-02 -7.24417508e-01 -3.19887877e-01 -6.70530424e-02] | [12.530037879943848, 7.946274757385254] |
3eecf93f-7697-493c-92e8-84f8b71fc053 | hi-transformer-hierarchical-interactive | 2106.01040 | null | https://arxiv.org/abs/2106.01040v3 | https://arxiv.org/pdf/2106.01040v3.pdf | Hi-Transformer: Hierarchical Interactive Transformer for Efficient and Effective Long Document Modeling | Transformer is important for text modeling. However, it has difficulty in handling long documents due to the quadratic complexity with input text length. In order to handle this problem, we propose a hierarchical interactive Transformer (Hi-Transformer) for efficient and effective long document modeling. Hi-Transformer models documents in a hierarchical way, i.e., first learns sentence representations and then learns document representations. It can effectively reduce the complexity and meanwhile capture global document context in the modeling of each sentence. More specifically, we first use a sentence Transformer to learn the representations of each sentence. Then we use a document Transformer to model the global document context from these sentence representations. Next, we use another sentence Transformer to enhance sentence modeling using the global document context. Finally, we use hierarchical pooling method to obtain document embedding. Extensive experiments on three benchmark datasets validate the efficiency and effectiveness of Hi-Transformer in long document modeling. | ['Yongfeng Huang', 'Tao Qi', 'Fangzhao Wu', 'Chuhan Wu'] | 2021-06-02 | null | https://aclanthology.org/2021.acl-short.107 | https://aclanthology.org/2021.acl-short.107.pdf | acl-2021-5 | ['document-embedding'] | ['methodology'] | [ 1.31176442e-01 -1.96816906e-01 -2.00405777e-01 -4.40437496e-01
-7.86682904e-01 -4.25163746e-01 6.37648821e-01 2.63022959e-01
-1.95197761e-01 2.24043071e-01 7.56350875e-01 -1.94133312e-01
1.41400155e-02 -9.37210262e-01 -4.71918166e-01 -6.86821103e-01
4.34383154e-01 1.52233317e-01 2.97903150e-01 5.31073613e-03
4.74419117e-01 5.08411266e-02 -9.44470286e-01 6.56309068e-01
7.55407929e-01 7.29549527e-01 7.55150557e-01 6.03361726e-01
-7.63872921e-01 9.15729165e-01 -6.43924415e-01 -2.74882823e-01
-3.74713361e-01 -2.09487870e-01 -7.90740371e-01 2.15150848e-01
-1.19961035e-02 -6.39224410e-01 -4.94268298e-01 7.47096777e-01
4.96868670e-01 1.36129797e-01 6.52864933e-01 -8.07371020e-01
-8.74330997e-01 9.78124976e-01 -6.50905013e-01 1.07042134e-01
1.99990079e-01 -3.41988266e-01 1.20329344e+00 -1.17359626e+00
2.10295260e-01 1.63707530e+00 3.90411258e-01 4.13945526e-01
-1.21407735e+00 -6.18683338e-01 7.83937395e-01 -7.50886202e-02
-1.14821243e+00 -1.09088108e-01 9.68483865e-01 -3.98775667e-01
1.02935255e+00 1.98212877e-01 4.81122017e-01 8.24975371e-01
3.61749977e-01 1.30379629e+00 6.03810072e-01 -3.51168692e-01
-3.62180173e-01 2.27341548e-01 7.20827639e-01 6.08691573e-01
-2.04282820e-01 -5.00549138e-01 -2.32946873e-01 -1.76796734e-01
5.89346945e-01 6.70500040e-01 -7.73857832e-02 1.37057886e-01
-1.04050887e+00 9.01134372e-01 4.93263364e-01 3.44415545e-01
-1.93587795e-01 1.94329992e-01 7.36146867e-01 1.52883247e-01
6.33412123e-01 -1.01472199e-01 -3.85084242e-01 1.28257304e-01
-8.11953187e-01 4.01170433e-01 4.83273804e-01 9.44904745e-01
5.49740255e-01 -1.67152807e-01 -7.37277389e-01 1.30138946e+00
5.29565156e-01 3.22288483e-01 7.25683212e-01 -3.64744753e-01
1.01701736e+00 7.40378380e-01 -1.40716791e-01 -1.20231545e+00
-2.22980499e-01 -4.69826847e-01 -1.02460766e+00 -5.01398802e-01
-3.24441522e-01 2.22200051e-01 -8.17711592e-01 1.49052680e+00
-5.07873148e-02 -7.45070055e-02 2.15259433e-01 4.54417437e-01
8.60382915e-01 1.43465972e+00 1.73842534e-01 -1.53260097e-01
1.49587858e+00 -1.40076673e+00 -6.57345235e-01 -2.52661347e-01
9.11658168e-01 -6.38047755e-01 1.45632803e+00 3.93342748e-02
-1.00351369e+00 -5.58171332e-01 -9.36104119e-01 -5.09642601e-01
-4.13901746e-01 4.68874186e-01 2.50761390e-01 2.03603104e-01
-5.57213306e-01 1.34900600e-01 -8.45617294e-01 8.72662142e-02
3.75358045e-01 3.40165734e-01 6.34039715e-02 -3.61358672e-01
-1.23163772e+00 4.10675973e-01 4.86179471e-01 1.60641775e-01
-8.07644010e-01 -6.48431778e-01 -1.12810266e+00 6.43490136e-01
4.85428870e-02 -6.50133431e-01 1.31910408e+00 -5.24288714e-01
-1.30175710e+00 3.14148277e-01 -6.94463491e-01 -2.69259512e-01
1.43395603e-01 -2.57072747e-01 -6.22432046e-02 3.25423852e-02
9.46791563e-03 5.07378936e-01 6.04482770e-01 -1.05911064e+00
-4.01082844e-01 -4.72643316e-01 4.05963778e-01 2.78914899e-01
-9.91302490e-01 1.90485373e-01 -9.01925385e-01 -8.50735486e-01
1.90023050e-01 -5.82643449e-01 -4.21167105e-01 -2.94856787e-01
-6.29155338e-01 -5.36689103e-01 1.00649452e+00 -7.29652464e-01
1.72186577e+00 -2.19004083e+00 1.50155529e-01 -6.63748682e-02
2.98444957e-01 3.91904116e-02 -3.11384499e-01 6.86884224e-01
1.87454924e-01 3.54788035e-01 -2.18090072e-01 -9.05637145e-01
1.14872232e-01 2.31640428e-01 -8.64865065e-01 -3.01626861e-01
9.10824612e-02 1.00604367e+00 -5.68584621e-01 -6.54412508e-01
1.50939703e-01 5.89928746e-01 -6.44179046e-01 2.69005299e-01
-3.81082952e-01 -1.99057102e-01 -9.31640446e-01 1.89382434e-01
8.62016201e-01 -3.46794456e-01 -9.26757529e-02 -9.33622271e-02
-3.48438346e-03 6.09114766e-01 -7.70158648e-01 1.47876275e+00
-8.38177860e-01 4.59153414e-01 -3.06694955e-01 -8.68556261e-01
9.91891861e-01 1.80804044e-01 1.26575157e-01 -4.84542936e-01
7.72894621e-02 -8.09852593e-03 -5.12953222e-01 -6.29262745e-01
6.71615303e-01 -2.14920551e-01 -1.08900346e-01 7.18796253e-01
-7.83628523e-02 4.44095135e-02 5.82914203e-02 6.38327122e-01
5.92383862e-01 -3.21239680e-01 -1.32361799e-01 -1.66080028e-01
8.43383431e-01 -5.14555216e-01 3.50409180e-01 4.27310765e-01
4.06216204e-01 6.74769223e-01 7.78286517e-01 -3.43347490e-01
-7.24880815e-01 -8.43409061e-01 1.11158185e-01 1.12549102e+00
1.45064116e-01 -1.04043150e+00 -7.38978326e-01 -7.94512749e-01
-4.36538979e-02 6.96764290e-01 -7.18239129e-01 -4.70742077e-01
-6.29436672e-01 -7.16545165e-01 2.35465810e-01 1.02128279e+00
5.28738320e-01 -1.04425526e+00 -7.54760951e-02 4.26196784e-01
-2.61722326e-01 -7.85819709e-01 -1.00911021e+00 -4.74803895e-02
-9.72450912e-01 -5.82290232e-01 -8.62929046e-01 -1.08369577e+00
8.21895897e-01 6.01819038e-01 7.77720392e-01 2.22533256e-01
7.08024502e-02 9.78836510e-03 -2.45947912e-01 -2.37771198e-01
-1.25152186e-01 3.28893751e-01 -3.44644189e-01 -1.72536328e-01
2.87976146e-01 -3.23116064e-01 -6.02657855e-01 1.34483576e-01
-1.12101960e+00 2.60644794e-01 5.14895380e-01 1.02985191e+00
5.48373163e-01 3.43615383e-01 5.59240282e-01 -9.45986867e-01
1.20991743e+00 -4.19807404e-01 -3.90928566e-01 5.53415358e-01
-4.39782917e-01 9.09506306e-02 9.34626460e-01 -4.26398188e-01
-1.18517101e+00 -3.14520389e-01 -3.35427314e-01 -2.58177966e-01
5.05024433e-01 1.06019247e+00 -3.23780447e-01 4.75564092e-01
-2.88819261e-02 8.73911798e-01 -3.52704227e-01 -7.20252097e-01
5.82880415e-02 1.05430412e+00 4.11440767e-02 -6.40896559e-01
5.19048035e-01 1.56650692e-01 -3.77711177e-01 -6.50225163e-01
-1.14849353e+00 -2.15386257e-01 -5.48812926e-01 3.24066907e-01
8.28995526e-01 -7.91000187e-01 -5.10446429e-01 5.06014645e-01
-1.44658518e+00 -1.73415411e-02 1.10600717e-01 2.74777859e-01
-8.28131884e-02 4.60620493e-01 -8.71070981e-01 -8.54325294e-01
-6.97424710e-01 -1.25996995e+00 1.27249885e+00 2.58287579e-01
2.09839925e-01 -1.30613029e+00 -1.29258588e-01 2.47341856e-01
2.89860457e-01 -3.08165997e-01 1.16741276e+00 -2.64882833e-01
-4.53161240e-01 -5.25123417e-01 -5.06400108e-01 5.67825019e-01
2.25427717e-01 -9.49398056e-03 -8.42568219e-01 -2.44152337e-01
-1.05433436e-02 -2.24633396e-01 1.24518788e+00 2.65875131e-01
1.71347976e+00 -4.76775408e-01 -3.95684332e-01 3.53799045e-01
1.21722996e+00 -5.09272097e-03 4.82046127e-01 1.97587788e-01
8.57217669e-01 6.33254230e-01 6.77877784e-01 4.63853747e-01
6.54859066e-01 4.82503563e-01 -1.12602085e-01 6.84947148e-02
2.15930194e-01 -7.89931476e-01 4.85671252e-01 1.26864588e+00
5.34757435e-01 -4.55950528e-01 -6.81124806e-01 3.89404535e-01
-1.83779299e+00 -6.65516198e-01 3.67780626e-02 1.72571945e+00
8.29749584e-01 4.73740131e-01 -1.93461761e-01 1.18203118e-01
4.90490854e-01 4.23731565e-01 -3.38338315e-01 -4.70982075e-01
6.59228191e-02 -3.69322240e-01 2.56185271e-02 5.09126604e-01
-7.80186772e-01 1.12048256e+00 6.10775280e+00 1.04804003e+00
-1.12085783e+00 1.00031486e-02 6.15102291e-01 -4.22765426e-02
-7.34562635e-01 9.20517021e-04 -1.14532137e+00 6.44881606e-01
6.42568171e-01 -6.24611795e-01 8.30050111e-02 7.71100461e-01
3.29017609e-01 4.08127546e-01 -8.65559876e-01 8.30416024e-01
2.90420920e-01 -1.41363168e+00 7.85647929e-01 -5.31389788e-02
5.06048739e-01 -3.22685927e-01 8.40519890e-02 6.22516632e-01
-7.58330822e-02 -8.29276741e-01 5.17826200e-01 4.73530143e-01
4.65832204e-01 -1.01880074e+00 6.18455827e-01 7.95760453e-01
-1.61945271e+00 -3.41538221e-01 -9.66511071e-01 1.42104372e-01
1.49889573e-01 4.93003458e-01 -3.65273356e-01 3.66380900e-01
3.63561630e-01 1.03059483e+00 -6.57631278e-01 7.59662092e-01
-1.53330773e-01 4.46213067e-01 -6.28520325e-02 -2.15825438e-01
4.74562913e-01 -1.54346898e-01 1.11070149e-01 1.47795451e+00
2.53620386e-01 1.03698500e-01 4.45897132e-01 7.02610135e-01
-2.65623778e-01 3.95127773e-01 -4.42645758e-01 -2.90066659e-01
4.75933015e-01 1.10602450e+00 -2.90650964e-01 -5.86752892e-01
-4.28660005e-01 1.09068191e+00 4.88010257e-01 4.62978184e-01
-6.46136642e-01 -7.05748916e-01 2.34564140e-01 -5.39782830e-02
1.76605538e-01 -3.67766291e-01 -1.46536544e-01 -1.35353267e+00
4.33312863e-01 -4.96918708e-01 4.25400257e-01 -9.58740413e-01
-1.13816607e+00 7.43212938e-01 -6.83699846e-02 -1.18177986e+00
5.52701950e-03 -3.56680274e-01 -9.77209508e-01 1.02724409e+00
-1.49611342e+00 -1.57838297e+00 -2.22634345e-01 5.10471940e-01
1.04087484e+00 8.21838379e-02 8.53924334e-01 1.19055629e-01
-8.11576903e-01 7.03898132e-01 3.30389142e-01 4.95933108e-02
3.52118134e-01 -1.23807931e+00 5.03509521e-01 4.13467824e-01
-4.51984107e-02 1.11078012e+00 2.01130375e-01 -4.19364542e-01
-1.18395936e+00 -1.26303661e+00 1.17865944e+00 -2.42335066e-01
5.08428812e-01 -7.88926601e-01 -1.10968554e+00 8.98566484e-01
1.79188982e-01 -3.72423887e-01 8.44792485e-01 7.23649412e-02
-4.35700297e-01 -3.68068546e-01 -7.70318627e-01 6.83701456e-01
5.07903278e-01 -8.26259196e-01 -8.31878006e-01 3.88409525e-01
1.39546156e+00 -6.86723813e-02 -7.23414063e-01 1.00927725e-01
5.51569998e-01 -3.81851792e-01 9.55886066e-01 -5.53880870e-01
7.14246213e-01 -6.96989372e-02 -1.72339171e-01 -1.30694950e+00
-3.53504509e-01 -3.75890471e-02 -1.73827022e-01 1.75536370e+00
3.65388304e-01 -6.68477833e-01 6.07194066e-01 4.81844515e-01
-1.81868345e-01 -1.27508378e+00 -3.12767416e-01 -6.69131398e-01
4.78197813e-01 -2.50596076e-01 9.23489451e-01 6.81587279e-01
9.72345695e-02 5.51932216e-01 -4.19453144e-01 -6.10250607e-02
3.63745004e-01 4.39048022e-01 6.66823804e-01 -9.39782321e-01
-2.45184734e-01 -3.95009935e-01 -1.07667278e-02 -1.86377859e+00
3.14927906e-01 -9.63401914e-01 2.41457615e-02 -1.98196971e+00
7.90700316e-01 -2.41705626e-01 -5.57729423e-01 2.02081546e-01
-5.53405404e-01 -2.77947694e-01 1.56287596e-01 2.97389328e-01
-5.83171010e-01 1.00494778e+00 1.46917689e+00 -4.37574774e-01
-1.14843719e-01 3.48862968e-02 -9.50893104e-01 5.34673452e-01
8.19379866e-01 -4.85729992e-01 -5.91166794e-01 -9.79244947e-01
3.24027706e-03 4.94082905e-02 1.32969320e-01 -5.39337099e-01
2.71538258e-01 -2.35423725e-02 4.75754291e-01 -1.03390527e+00
3.73884916e-01 -7.01569021e-01 -5.95407546e-01 4.90859807e-01
-7.91773319e-01 -4.01101187e-02 1.72435254e-01 5.50031960e-01
-5.17016351e-01 -5.06629884e-01 5.23184896e-01 -6.34155348e-02
-1.32882521e-01 5.97924113e-01 -4.35146570e-01 -2.96696037e-01
6.46773338e-01 6.06865659e-02 -2.96393156e-01 -1.82222635e-01
-5.23934364e-01 9.13795769e-01 1.45377994e-01 6.06751859e-01
8.86582255e-01 -1.41411102e+00 -6.28601015e-01 3.79826427e-01
2.03901678e-01 1.51258513e-01 4.52032030e-01 5.23639202e-01
-2.11963251e-01 5.92424691e-01 3.91011894e-01 -4.40917790e-01
-1.53999865e+00 4.72306997e-01 2.56962568e-01 -7.27472961e-01
-6.64928615e-01 9.22006488e-01 7.92493463e-01 -4.64534611e-01
3.17564845e-01 -4.57947910e-01 -4.68267530e-01 -5.00320606e-02
1.03922164e+00 -1.04504876e-01 -3.11068505e-01 -3.68882626e-01
-1.40968621e-01 8.47591937e-01 -7.39344478e-01 -9.81583968e-02
1.27119684e+00 -3.03612500e-01 -3.99656862e-01 6.16255939e-01
1.68523872e+00 -2.57432256e-02 -9.03983951e-01 -4.51244831e-01
2.00865194e-02 -5.67592382e-01 1.66263789e-01 -4.19610769e-01
-9.73589122e-01 1.55726802e+00 1.26987144e-01 2.72660851e-01
1.08152723e+00 -8.36195946e-02 1.19921994e+00 7.06726551e-01
-1.30514205e-01 -6.77741647e-01 2.97485411e-01 7.15264320e-01
1.18406105e+00 -7.54309535e-01 -1.90009445e-01 -4.59475607e-01
-5.92107713e-01 1.38455939e+00 6.93719685e-01 -5.84470145e-02
8.23104024e-01 1.56786665e-01 -2.39985555e-01 -3.65121961e-02
-1.02799082e+00 3.60495210e-01 1.44953996e-01 -1.16427382e-02
6.46501720e-01 -7.76995048e-02 -1.77562982e-01 1.23401010e+00
-7.26120472e-02 3.97952870e-02 2.79395431e-01 8.36142600e-01
-6.17204785e-01 -1.27495718e+00 -4.20596413e-02 6.15161955e-01
-2.91420311e-01 -3.74183118e-01 -4.56783831e-01 8.42510313e-02
-2.52734035e-01 8.24954867e-01 2.14792951e-03 -5.86051583e-01
3.97630990e-01 6.80984035e-02 2.12458238e-01 -8.70107472e-01
-6.24678075e-01 2.13245198e-01 -2.21145779e-01 -6.20753877e-02
1.68175966e-01 -4.84292597e-01 -1.41732955e+00 -2.19461888e-01
-4.78337198e-01 4.47621495e-01 4.56533700e-01 9.16150987e-01
2.89020777e-01 7.13026941e-01 9.12633836e-01 -5.32850683e-01
-6.40968859e-01 -1.25678027e+00 -6.33609056e-01 1.07525170e-01
3.69462997e-01 -3.31709325e-01 -3.48807484e-01 7.72012072e-03] | [11.148287773132324, 8.597188949584961] |
ca1ccdf0-29e9-48b9-92e1-825fe83dffcf | discrete-point-flow-networks-for-efficient | 2007.10170 | null | https://arxiv.org/abs/2007.10170v1 | https://arxiv.org/pdf/2007.10170v1.pdf | Discrete Point Flow Networks for Efficient Point Cloud Generation | Generative models have proven effective at modeling 3D shapes and their statistical variations. In this paper we investigate their application to point clouds, a 3D shape representation widely used in computer vision for which, however, only few generative models have yet been proposed. We introduce a latent variable model that builds on normalizing flows with affine coupling layers to generate 3D point clouds of an arbitrary size given a latent shape representation. To evaluate its benefits for shape modeling we apply this model for generation, autoencoding, and single-view shape reconstruction tasks. We improve over recent GAN-based models in terms of most metrics that assess generation and autoencoding. Compared to recent work based on continuous flows, our model offers a significant speedup in both training and inference times for similar or better performance. For single-view shape reconstruction we also obtain results on par with state-of-the-art voxel, point cloud, and mesh-based methods. | ['Roman Klokov', 'Edmond Boyer', 'Jakob Verbeek'] | 2020-07-20 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/4408_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123680681.pdf | eccv-2020-8 | ['3d-shape-representation', 'point-cloud-generation'] | ['computer-vision', 'computer-vision'] | [-2.86597684e-02 1.52915940e-01 3.31263423e-01 -1.75258219e-01
-9.45136309e-01 -8.01094472e-01 1.11183548e+00 -1.99658185e-01
2.45786875e-01 4.90056366e-01 6.11491576e-02 -2.40895063e-01
1.83289111e-01 -1.20780027e+00 -8.29339623e-01 -5.95076621e-01
3.05442274e-01 1.31875455e+00 1.30124658e-01 -2.89166253e-02
1.41238660e-01 1.01691329e+00 -1.39674723e+00 1.32736295e-01
5.34963369e-01 7.85063148e-01 2.39052437e-02 6.75603688e-01
-4.67912078e-01 1.23180822e-01 -2.57326990e-01 -3.81909609e-01
2.68538892e-01 -1.44123495e-01 -6.86912358e-01 1.80729240e-01
6.72223270e-01 -1.62141055e-01 5.91350570e-02 5.57792246e-01
6.53941274e-01 -1.31536633e-01 1.04993105e+00 -1.08638883e+00
-6.99804664e-01 1.20545991e-01 -4.15409058e-01 -2.25961372e-01
2.23909035e-01 1.81709379e-01 7.31989682e-01 -1.18002200e+00
9.64684129e-01 1.42880762e+00 7.88455188e-01 5.43021858e-01
-1.80411553e+00 -3.75446647e-01 -1.28000796e-01 -4.68913734e-01
-1.12547207e+00 -3.81959170e-01 9.51363206e-01 -6.40025973e-01
1.34396565e+00 8.41744766e-02 8.48503172e-01 9.62412357e-01
2.39489630e-01 6.72423422e-01 9.79246497e-01 -1.97247714e-01
2.79843867e-01 -8.94261003e-02 -6.72392130e-01 5.52941322e-01
-4.63088490e-02 1.86821088e-01 -1.93427831e-01 -5.16206980e-01
1.35603786e+00 -1.33187965e-01 2.91989595e-02 -9.58456159e-01
-1.21979928e+00 8.98121178e-01 3.71025920e-01 1.62275434e-01
-4.50720221e-01 6.87222362e-01 7.58801773e-02 1.27293080e-01
1.13838243e+00 1.11826397e-01 -2.67190218e-01 -3.69324714e-01
-1.18211329e+00 6.64910793e-01 7.52875865e-01 1.23730063e+00
6.65691435e-01 5.11393249e-01 -6.45214394e-02 8.17789793e-01
5.29452443e-01 7.29565024e-01 -1.83833793e-01 -1.08422720e+00
3.83300334e-01 5.90215564e-01 -7.31580853e-02 -7.04901755e-01
-9.03275609e-02 -4.57184821e-01 -8.21209073e-01 5.85613012e-01
-8.69546831e-02 2.70523697e-01 -1.29495883e+00 1.26884818e+00
4.25024778e-01 4.44507062e-01 -1.59854129e-01 5.38566828e-01
7.85704851e-01 6.58781588e-01 -1.49076834e-01 4.10337746e-02
9.42463219e-01 -6.23441160e-01 -1.91761345e-01 1.94772243e-01
2.42772579e-01 -1.04638195e+00 6.73167706e-01 3.60149056e-01
-1.75561833e+00 -5.46177864e-01 -6.68043792e-01 -2.22550586e-01
-1.62712246e-01 -2.49543980e-01 7.29982257e-01 8.19823802e-01
-1.49957204e+00 7.84523487e-01 -1.27078688e+00 -1.05997935e-01
7.84861088e-01 2.40112633e-01 -2.02533066e-01 -1.94032937e-01
-2.64869750e-01 6.71009362e-01 -2.37643093e-01 -4.78155524e-01
-9.92104173e-01 -1.15908074e+00 -8.56697679e-01 -4.59257551e-02
-2.00173527e-01 -1.58840024e+00 1.14889324e+00 -1.69236705e-01
-1.68101370e+00 9.51157451e-01 -3.46202731e-01 -3.25817227e-01
5.81230581e-01 2.09360346e-01 1.90857425e-01 -2.04923097e-02
-4.89488617e-02 9.57325101e-01 8.98442566e-01 -1.63510358e+00
5.23212999e-02 -3.99473876e-01 -8.27423856e-02 1.57885835e-01
3.65125746e-01 -3.36111903e-01 -4.64201719e-01 -6.98819697e-01
3.60206962e-01 -9.64836717e-01 -3.73001873e-01 2.80075550e-01
-7.06111193e-02 -1.60019755e-01 8.88515830e-01 -3.86613011e-01
4.22231823e-01 -1.81015265e+00 3.66724163e-01 2.85484016e-01
2.35180542e-01 -5.80742806e-02 -8.15910473e-02 4.75356311e-01
1.10512963e-02 4.37369317e-01 -6.20588720e-01 -1.12231100e+00
1.63604125e-01 5.13610780e-01 -4.54474986e-01 2.61320800e-01
3.34054440e-01 1.27015722e+00 -7.23171294e-01 -2.86426753e-01
5.97536802e-01 1.04010022e+00 -9.46999431e-01 1.40184358e-01
-4.78415042e-01 7.35031486e-01 -2.90037096e-01 5.32594323e-01
8.13783884e-01 -3.38422805e-01 -2.99899876e-01 4.91376370e-02
-7.70317316e-02 2.14859620e-01 -8.59484375e-01 2.12554383e+00
-7.68402100e-01 2.45653361e-01 1.13638034e-02 -6.85297310e-01
1.11441052e+00 4.90903556e-01 7.28895783e-01 -1.82090923e-01
-9.30650681e-02 2.52473235e-01 -3.35882664e-01 2.11636618e-01
5.00001132e-01 -4.99575943e-01 2.14204505e-01 5.08600473e-01
2.28426427e-01 -1.08822775e+00 -2.39944443e-01 2.33726382e-01
8.57895911e-01 7.00981379e-01 -6.68065697e-02 -2.14559779e-01
2.22576976e-01 -9.39647108e-02 -2.64570564e-02 4.77237284e-01
6.69045389e-01 1.26855874e+00 2.68634915e-01 -3.52782547e-01
-1.56270349e+00 -1.45090806e+00 -2.65799910e-01 2.96986639e-01
-3.10863405e-01 -2.86805928e-01 -5.47307611e-01 -2.87225813e-01
2.09259465e-01 1.00029123e+00 -5.37092984e-01 2.67424226e-01
-6.13483667e-01 -6.09917402e-01 3.17695409e-01 7.05060542e-01
8.60307738e-02 -1.02552676e+00 -4.15812552e-01 3.47483397e-01
2.48430595e-01 -1.05149150e+00 -1.67443022e-01 -2.50565797e-01
-1.54454911e+00 -6.97265804e-01 -9.22380447e-01 -3.88174891e-01
7.56159902e-01 -6.76167896e-04 1.70779288e+00 -1.10621564e-01
-2.29083180e-01 9.45524991e-01 -5.44711798e-02 -4.83025491e-01
-8.87200117e-01 9.90558118e-02 -3.49238575e-01 -3.67007136e-01
-1.03708617e-01 -1.15137911e+00 -4.94823337e-01 2.11609051e-01
-1.07494962e+00 2.18059376e-01 2.62242228e-01 7.22044408e-01
1.04694402e+00 -4.45527345e-01 1.63737699e-01 -9.32725132e-01
3.71473670e-01 -4.36049759e-01 -6.16592824e-01 -1.18836127e-01
-7.04242766e-01 1.51677117e-01 3.38116318e-01 -8.36454704e-02
-9.82522607e-01 3.96450534e-02 -4.99974489e-01 -1.00339091e+00
-3.88302714e-01 1.12761438e-01 1.82212815e-02 -2.08714470e-01
4.52709824e-01 4.20631945e-01 2.29144886e-01 -6.06083393e-01
6.80915654e-01 4.40538973e-02 2.66908526e-01 -7.99675405e-01
1.04533815e+00 8.89732957e-01 3.83882761e-01 -7.41993725e-01
-8.90944898e-02 -3.18044364e-01 -8.05251479e-01 1.06227966e-02
8.12572837e-01 -8.73398960e-01 -3.38149518e-01 3.62186462e-01
-1.39828658e+00 -3.05332214e-01 -6.36241794e-01 1.40692681e-01
-1.28655350e+00 3.09938371e-01 -4.11140740e-01 -7.68164217e-01
-4.61092889e-01 -1.14349580e+00 1.78502667e+00 -3.32601488e-01
-6.05529360e-02 -1.33548439e+00 2.86831498e-01 4.66527939e-02
8.23487699e-01 5.62312245e-01 1.14006531e+00 3.27960588e-02
-1.01185179e+00 6.01746179e-02 -5.18253446e-02 8.59651491e-02
4.21772078e-02 -2.02618148e-02 -1.02026057e+00 -4.07746673e-01
-5.11793830e-02 1.44191058e-02 6.70598447e-01 6.56773865e-01
1.16350484e+00 1.58842430e-02 -4.05699700e-01 9.68833864e-01
1.62776828e+00 2.05055438e-02 8.97648335e-01 -2.50216872e-01
7.18153894e-01 1.85811952e-01 -1.00834653e-01 4.03330475e-01
3.22762758e-01 8.35878253e-01 6.29269004e-01 -1.18143432e-01
-4.26389396e-01 -4.37400371e-01 -7.58832768e-02 1.07715154e+00
-5.40652633e-01 -2.14450136e-01 -9.16390181e-01 7.84558773e-01
-1.49182177e+00 -8.95096660e-01 -2.69369036e-01 2.13089514e+00
3.40908855e-01 8.59533250e-02 4.43307422e-02 -7.53302500e-02
2.67715961e-01 1.37480602e-01 -5.45718670e-01 -4.48128253e-01
2.16457080e-02 7.24634945e-01 2.38999158e-01 5.46044528e-01
-6.69384837e-01 8.28223765e-01 6.83073711e+00 7.13326752e-01
-9.84191179e-01 3.16027313e-01 6.37724876e-01 -7.49822259e-02
-1.06749082e+00 6.20514713e-02 -6.05787635e-01 1.05014361e-01
7.17982650e-01 -1.77627783e-02 4.74291891e-01 9.01598036e-01
-6.90493127e-03 3.90364051e-01 -1.14545858e+00 1.18683529e+00
1.64213330e-01 -1.74426138e+00 3.38223279e-01 3.94968331e-01
1.01646829e+00 3.52294594e-01 1.23475231e-01 1.71935916e-01
4.86602664e-01 -1.19933379e+00 7.28454947e-01 6.74770296e-01
1.04884994e+00 -7.04587460e-01 3.33284289e-01 3.45402777e-01
-1.18792975e+00 6.77831233e-01 -5.65320730e-01 3.26034993e-01
5.51220477e-01 5.88077903e-01 -9.20659542e-01 8.72396588e-01
3.29740942e-01 7.31608570e-01 -2.61708885e-01 9.78188813e-01
-5.05652800e-02 6.26536965e-01 -5.79118967e-01 4.04709518e-01
1.17883839e-01 -3.54830354e-01 8.25662732e-01 9.54908133e-01
7.58647621e-01 -5.53899221e-02 1.53705128e-03 1.56044936e+00
1.14111543e-01 2.21882872e-02 -1.03002882e+00 2.80644447e-01
3.64950448e-01 9.77416575e-01 -8.22552860e-01 -3.15615445e-01
-1.59174964e-01 8.27244759e-01 1.05281122e-01 1.86453834e-01
-6.86226368e-01 3.04587334e-01 5.42328238e-01 4.64745015e-01
6.28871262e-01 -6.56969488e-01 -6.46938026e-01 -1.05742836e+00
-3.71001549e-02 -3.05054545e-01 -1.54800728e-01 -1.02053821e+00
-1.41713214e+00 7.12473571e-01 3.74585092e-01 -1.35566485e+00
-7.34426856e-01 -4.81826544e-01 -5.34268439e-01 1.13220000e+00
-1.38008165e+00 -1.58196199e+00 -1.93978727e-01 3.72405827e-01
5.91844022e-01 -1.55740410e-01 1.13322139e+00 1.09129939e-02
3.69845986e-01 2.22703218e-01 -1.58007089e-02 -3.79278272e-01
1.51546434e-01 -1.48335409e+00 1.30895245e+00 5.02208889e-01
4.38271463e-01 4.96806353e-01 4.49936271e-01 -7.88867772e-01
-1.50502098e+00 -1.04424691e+00 5.24224460e-01 -8.97881150e-01
5.33532463e-02 -4.16610062e-01 -8.38350058e-01 7.07172096e-01
4.24404927e-02 1.76377654e-01 3.33402276e-01 -5.05349413e-02
-2.42262259e-01 4.15233433e-01 -1.29431748e+00 2.60719508e-01
1.42468262e+00 -2.88697630e-01 -4.76008415e-01 9.62668881e-02
6.18582129e-01 -7.59599984e-01 -1.16675341e+00 4.12222981e-01
4.49658275e-01 -9.92236316e-01 1.33535397e+00 -4.23007339e-01
6.66933417e-01 -1.65812671e-01 -2.47893840e-01 -1.63409162e+00
-4.97379094e-01 -5.08586645e-01 -1.99905351e-01 1.08955479e+00
2.03292519e-01 -5.35132468e-01 1.11709785e+00 4.69628781e-01
-4.73526657e-01 -8.70348036e-01 -1.08180606e+00 -8.41115057e-01
5.64167857e-01 -6.07074857e-01 1.00271404e+00 6.75497830e-01
-8.06501508e-01 1.25357285e-01 -1.59305558e-02 -1.19181767e-01
8.67435992e-01 5.17751336e-01 1.02292669e+00 -1.34499431e+00
-2.65346259e-01 -6.14485800e-01 -5.47426045e-01 -1.13641095e+00
1.46449432e-02 -1.30754292e+00 -3.49042267e-01 -2.08025146e+00
-1.24601774e-01 -7.90197909e-01 3.66206586e-01 2.50789911e-01
2.88428187e-01 4.18307245e-01 1.67652190e-01 1.05651334e-01
1.08847000e-01 8.44524324e-01 1.40222824e+00 -3.15693021e-02
-2.50337254e-02 1.45472575e-03 -3.08960378e-01 5.69098949e-01
5.97817838e-01 -3.84904683e-01 -5.45576513e-01 -7.15799332e-01
1.74126655e-01 2.93610126e-01 7.02397525e-01 -9.54457700e-01
-7.07371905e-02 1.99447591e-02 5.30814469e-01 -1.01293874e+00
7.88813174e-01 -8.35663497e-01 8.64520669e-01 2.51929671e-01
8.53416622e-02 3.36746871e-01 3.69934976e-01 5.24627090e-01
1.26173033e-03 1.57954525e-02 6.97180152e-01 -4.63983595e-01
-3.15998942e-02 7.07364976e-01 -4.00318131e-02 -1.83621328e-02
6.64133668e-01 -4.73411977e-01 2.35700086e-02 -5.45076907e-01
-8.23234975e-01 -3.17528456e-01 8.79261494e-01 3.68027627e-01
1.00472820e+00 -1.65983927e+00 -1.05680752e+00 4.41197306e-01
-4.86961268e-02 5.65191686e-01 3.21005583e-01 3.75721425e-01
-6.94963574e-01 3.67499888e-01 -1.37415707e-01 -1.08679116e+00
-9.83792841e-01 3.64143699e-01 2.98867792e-01 -3.25276256e-01
-7.94077873e-01 5.95843494e-01 2.65421927e-01 -8.67927790e-01
-4.01250541e-01 -4.95556384e-01 2.24686906e-01 -3.00361991e-01
-1.69357043e-02 3.28253239e-01 3.53088468e-01 -6.62955761e-01
-7.56028760e-03 9.78064239e-01 4.91769046e-01 -2.84114391e-01
1.65769815e+00 2.74960548e-01 -1.01068474e-01 4.90111649e-01
1.12064326e+00 5.75657412e-02 -1.30340290e+00 2.74842791e-02
-6.20618343e-01 -5.76956332e-01 4.58847769e-02 -5.68745852e-01
-1.16009569e+00 1.19899249e+00 5.15743434e-01 1.15305871e-01
6.29647017e-01 1.82344168e-01 6.89901769e-01 -2.52389401e-01
9.00007606e-01 -3.33201587e-01 -1.62121847e-01 5.55452049e-01
1.21899700e+00 -7.49928236e-01 1.57783136e-01 -7.05053091e-01
-2.48100564e-01 1.04336619e+00 1.26809433e-01 -5.40150821e-01
8.56568933e-01 4.61691082e-01 -2.71072894e-01 -4.88768905e-01
-8.69767547e-01 8.36430490e-02 4.85180676e-01 8.29928339e-01
3.45081955e-01 1.04487248e-01 1.92923918e-01 -1.21883169e-01
-4.94305432e-01 2.03492064e-02 2.84302950e-01 6.12726331e-01
5.47210053e-02 -1.40556931e+00 -3.12186539e-01 5.19553840e-01
-2.44723573e-01 -1.09497398e-01 -3.69432904e-02 6.31684840e-01
-1.51495367e-01 3.04216295e-01 4.19668555e-01 -1.71960816e-02
3.71891767e-01 2.34551132e-01 9.53212440e-01 -7.83391416e-01
-5.48015296e-01 3.30579817e-01 -1.09719411e-01 -5.79834223e-01
-5.44449806e-01 -8.67114246e-01 -9.73687768e-01 -2.51160502e-01
-1.24529071e-01 -6.30633980e-02 9.92969632e-01 4.98216569e-01
7.71577716e-01 4.10724849e-01 4.61465150e-01 -1.43322790e+00
-4.55348849e-01 -8.26920033e-01 -3.52060050e-01 4.42752928e-01
-9.80050862e-02 -8.29760373e-01 -1.02618076e-01 1.33052737e-01] | [8.893821716308594, -3.638495445251465] |
ce0cc6c1-9fb1-4400-899a-18da7361f6f6 | invpt-inverted-pyramid-multi-task-transformer | 2306.04842 | null | https://arxiv.org/abs/2306.04842v1 | https://arxiv.org/pdf/2306.04842v1.pdf | InvPT++: Inverted Pyramid Multi-Task Transformer for Visual Scene Understanding | Multi-task scene understanding aims to design models that can simultaneously predict several scene understanding tasks with one versatile model. Previous studies typically process multi-task features in a more local way, and thus cannot effectively learn spatially global and cross-task interactions, which hampers the models' ability to fully leverage the consistency of various tasks in multi-task learning. To tackle this problem, we propose an Inverted Pyramid multi-task Transformer, capable of modeling cross-task interaction among spatial features of different tasks in a global context. Specifically, we first utilize a transformer encoder to capture task-generic features for all tasks. And then, we design a transformer decoder to establish spatial and cross-task interaction globally, and a novel UP-Transformer block is devised to increase the resolutions of multi-task features gradually and establish cross-task interaction at different scales. Furthermore, two types of Cross-Scale Self-Attention modules, i.e., Fusion Attention and Selective Attention, are proposed to efficiently facilitate cross-task interaction across different feature scales. An Encoder Feature Aggregation strategy is further introduced to better model multi-scale information in the decoder. Comprehensive experiments on several 2D/3D multi-task benchmarks clearly demonstrate our proposal's effectiveness, establishing significant state-of-the-art performances. | ['Dan Xu', 'Hanrong Ye'] | 2023-06-08 | null | null | null | null | ['scene-understanding'] | ['computer-vision'] | [ 6.33632466e-02 -5.71005702e-01 4.27671224e-02 -4.82866764e-01
-8.66770267e-01 -3.73844802e-01 7.14250505e-01 -1.55178130e-01
-2.65390635e-01 3.15056622e-01 3.14203978e-01 1.59145996e-01
-3.86847407e-01 -4.23499554e-01 -6.98470294e-01 -6.31763816e-01
3.93297911e-01 1.15688637e-01 4.40249652e-01 -1.91410899e-01
2.48334199e-01 1.76719815e-01 -1.57112992e+00 7.24362552e-01
1.03405631e+00 9.83417273e-01 1.09863424e+00 3.69352341e-01
-1.41884908e-01 7.77424872e-01 -5.59196770e-01 -3.66450474e-02
-1.04464432e-02 -6.97744563e-02 -7.92097926e-01 2.71133125e-01
5.55924892e-01 -1.92993924e-01 -2.29552045e-01 8.88412952e-01
4.68475282e-01 8.83459598e-02 5.07798374e-01 -1.19253981e+00
-8.23630095e-01 1.13079205e-01 -7.90180922e-01 3.16749185e-01
1.68330476e-01 1.63239747e-01 1.08517337e+00 -1.08049667e+00
2.26602957e-01 1.53904641e+00 2.75261134e-01 1.60469845e-01
-1.19685400e+00 -8.25363994e-01 7.16002226e-01 3.74573290e-01
-1.37521625e+00 -3.24199706e-01 9.52483535e-01 -4.32192594e-01
9.91063654e-01 1.57910027e-02 4.69818592e-01 1.14014244e+00
6.23554766e-01 1.18042934e+00 1.16741896e+00 1.02697182e-02
-2.69884676e-01 -1.52403489e-01 -2.36801933e-02 5.06668150e-01
-4.85036559e-02 -1.93500981e-01 -9.47683036e-01 2.34578788e-01
1.09286153e+00 3.91907603e-01 -1.72710076e-01 -4.37148273e-01
-1.59005165e+00 6.13849938e-01 6.30007625e-01 6.10338748e-01
-4.14955467e-01 6.82315752e-02 5.77235997e-01 2.88618237e-01
6.35568857e-01 2.81211644e-01 -7.24916399e-01 1.71512201e-01
-6.38031542e-01 8.14142078e-02 1.03937685e-01 1.11056590e+00
9.89180267e-01 -1.89613640e-01 -4.97069240e-01 1.09747648e+00
2.84229934e-01 3.43393505e-01 4.98192340e-01 -4.81389672e-01
8.82359922e-01 8.38405013e-01 -2.90480375e-01 -9.69366729e-01
-5.69593847e-01 -6.54898286e-01 -1.13540995e+00 -2.39380345e-01
6.41304031e-02 1.67341053e-01 -8.30704033e-01 1.80091810e+00
2.77073443e-01 3.72003585e-01 3.56501900e-02 8.29362333e-01
6.87142730e-01 5.36688149e-01 3.24083060e-01 9.73154753e-02
1.66952157e+00 -1.41045916e+00 -7.07637191e-01 -6.93068326e-01
6.58525050e-01 -6.84730709e-01 1.13837075e+00 1.85538679e-01
-8.68770540e-01 -1.21065891e+00 -8.23841214e-01 -5.27391374e-01
-5.60628533e-01 3.08922499e-01 6.16098821e-01 -6.37795553e-02
-8.15092087e-01 1.19796488e-02 -5.85932851e-01 -6.55238777e-02
6.25014901e-01 3.02490532e-01 -5.01025081e-01 -3.64063203e-01
-1.13517332e+00 9.15826023e-01 4.16029245e-01 1.66804627e-01
-1.06053138e+00 -9.02806699e-01 -1.09446669e+00 7.18922168e-02
4.06397402e-01 -9.55648422e-01 9.80169296e-01 -7.26964474e-01
-1.11477268e+00 6.39729142e-01 -4.72899854e-01 1.12261079e-01
1.88240260e-01 -5.42161644e-01 -3.97412181e-01 2.28361040e-02
5.06598353e-01 7.91149378e-01 9.76306081e-01 -1.30942297e+00
-8.97374213e-01 -6.21925056e-01 2.51285583e-01 7.77570665e-01
-4.03738379e-01 -3.05981394e-02 -7.98907518e-01 -7.11600780e-01
1.76627249e-01 -6.09784186e-01 -2.55242020e-01 -1.90980271e-01
-4.61029828e-01 -2.37149313e-01 1.07053232e+00 -3.97539526e-01
1.03907454e+00 -2.29356933e+00 4.78712559e-01 -3.13358217e-01
5.62267184e-01 6.22289293e-02 -4.01696682e-01 1.69912159e-01
-1.84261039e-01 -6.79895282e-02 -6.38860911e-02 -8.04493070e-01
-1.89102620e-01 3.00496459e-01 -6.77526444e-02 2.19594926e-01
3.68689388e-01 1.25884473e+00 -8.41310620e-01 -4.70324427e-01
6.06869698e-01 4.54738498e-01 -3.83160442e-01 2.78376698e-01
4.48028259e-02 7.73309469e-01 -8.89699340e-01 5.37881374e-01
8.51373196e-01 -6.66215003e-01 -1.35705069e-01 -6.17747128e-01
-2.17425749e-01 1.90673470e-01 -1.05260158e+00 2.27772927e+00
-1.06070364e+00 3.67238104e-01 1.02018259e-01 -1.08153784e+00
8.67203772e-01 1.32152885e-01 4.61112350e-01 -1.12177837e+00
-1.51286066e-01 1.28292171e-02 -1.50526017e-01 -3.05647522e-01
4.62514520e-01 -5.29100411e-02 -2.67087221e-01 2.11383134e-01
2.44677216e-01 -2.94048898e-02 -1.56528205e-01 -8.86132941e-02
8.29231858e-01 8.70185941e-02 3.71005535e-01 -4.05544907e-01
7.69321144e-01 -3.59880149e-01 4.92398322e-01 6.10336065e-01
-1.57311007e-01 5.80098391e-01 2.46709347e-01 -4.83759940e-01
-7.49034464e-01 -8.98714662e-01 -7.93771520e-02 1.67717111e+00
6.18794441e-01 -3.99765879e-01 -1.93641186e-01 -8.28333676e-01
3.98600176e-02 3.40510488e-01 -8.21140707e-01 -3.11367661e-01
-4.82934207e-01 -7.16996312e-01 1.78971335e-01 7.02781916e-01
9.51891005e-01 -9.15811896e-01 -7.06370831e-01 3.84351820e-01
-4.03445482e-01 -1.36098242e+00 -6.57997489e-01 4.54452425e-01
-6.90893531e-01 -1.04405165e+00 -7.53246009e-01 -9.26690638e-01
4.72729594e-01 9.09149826e-01 1.02692819e+00 -2.19981834e-01
-5.80560043e-02 2.18201190e-01 -3.23778093e-01 -2.06680089e-01
2.21738115e-01 2.99893588e-01 -1.71304420e-01 3.05463642e-01
1.66085497e-01 -6.66600764e-01 -5.82867026e-01 6.86390281e-01
-9.43918884e-01 6.62269831e-01 9.96108830e-01 1.04761982e+00
6.16086721e-01 6.68114843e-03 6.89320922e-01 -6.82143271e-01
3.99456888e-01 -5.26414335e-01 -2.09304333e-01 5.95061362e-01
6.59478679e-02 9.01719853e-02 7.27853417e-01 -4.17380184e-01
-1.30580115e+00 3.50242443e-02 3.89606245e-02 -7.73357749e-01
-2.73846596e-01 2.91906297e-01 -6.29405141e-01 -1.21641427e-01
6.23994842e-02 6.63187683e-01 -3.94490123e-01 -6.77317619e-01
4.33978766e-01 3.90284598e-01 2.68273681e-01 -6.15583360e-01
6.18331015e-01 5.25315166e-01 -1.54038861e-01 -7.12765634e-01
-1.26128697e+00 -6.96770310e-01 -1.05865216e+00 -9.91659164e-02
1.14599395e+00 -1.39613116e+00 -5.91923475e-01 9.49149311e-01
-1.23276556e+00 -3.58319253e-01 4.42178063e-02 2.62141198e-01
-5.24872422e-01 7.65956640e-02 -3.11641276e-01 -3.02855045e-01
-3.16895097e-02 -1.44265759e+00 1.86127734e+00 2.93969154e-01
2.58602142e-01 -1.23892486e+00 -1.69398710e-01 4.78567839e-01
4.49901909e-01 -1.83920845e-01 8.20808351e-01 -5.11913002e-01
-7.73890316e-01 4.76468533e-01 -8.30949008e-01 2.56751448e-01
4.84854341e-01 -6.20343328e-01 -1.02824795e+00 -3.91640097e-01
1.00866266e-01 -3.83393407e-01 1.03238654e+00 3.82203102e-01
1.51414227e+00 1.81607500e-01 -6.15769088e-01 7.30727136e-01
1.30155420e+00 1.30013516e-03 4.84283805e-01 3.24296147e-01
1.21370435e+00 5.94787598e-01 7.47867882e-01 4.38151836e-01
9.46277022e-01 8.83922219e-01 3.92957419e-01 -4.14593160e-01
-1.99159309e-01 -2.00125292e-01 2.21081391e-01 8.44292462e-01
2.48534530e-01 -3.25647108e-02 -8.05085719e-01 5.60741544e-01
-1.94455433e+00 -6.42478943e-01 -1.12624867e-02 1.72933829e+00
5.67028522e-01 1.01450033e-01 -2.41577357e-01 -2.61804998e-01
5.88932812e-01 6.54692531e-01 -6.67244673e-01 7.04412833e-02
-1.86616257e-01 -5.82465231e-02 3.06416452e-01 2.23623037e-01
-1.45132685e+00 1.08670259e+00 5.51985884e+00 1.19392443e+00
-1.11565971e+00 3.60020876e-01 5.43723881e-01 -2.94428766e-02
-2.72769362e-01 -1.96267143e-01 -9.24792588e-01 3.99126500e-01
3.01213622e-01 -1.34738535e-01 -2.95152906e-02 6.63109958e-01
2.06870120e-02 -1.29838377e-01 -1.05924463e+00 1.15193892e+00
7.64841065e-02 -1.03312302e+00 2.05590695e-01 9.15987343e-02
7.14954972e-01 -5.70639223e-02 1.93873182e-01 5.38515151e-01
1.41052783e-01 -8.75983000e-01 7.01423705e-01 5.54528296e-01
7.55156755e-01 -6.06103122e-01 4.99776363e-01 4.96532381e-01
-2.01571894e+00 -3.16726923e-01 -3.02722782e-01 1.20234735e-01
2.06358820e-01 5.36766052e-01 -1.41382158e-01 1.07161331e+00
6.88784420e-01 1.26892805e+00 -9.30546165e-01 8.13791871e-01
6.68353140e-02 -1.63470387e-01 -7.21866637e-02 3.49600434e-01
5.11978805e-01 5.60888983e-02 2.84163386e-01 1.19443274e+00
1.79994002e-01 -1.82153791e-01 4.77311879e-01 7.15153873e-01
2.55585849e-01 5.16278557e-02 -6.83590353e-01 4.87983823e-01
3.59080821e-01 1.17332971e+00 -3.75483304e-01 -3.31547529e-01
-8.38313758e-01 1.27477777e+00 6.60342395e-01 4.47061539e-01
-9.23166871e-01 -6.17942065e-02 1.02007961e+00 -1.83959454e-01
4.04663026e-01 -5.55509925e-01 -4.23353374e-01 -1.33893776e+00
1.16840206e-01 -7.62178898e-01 4.03039515e-01 -8.22183788e-01
-1.32364845e+00 4.64927346e-01 1.36822075e-01 -1.31940758e+00
2.45697975e-01 -5.25252044e-01 -5.70926726e-01 9.89411950e-01
-1.92427933e+00 -1.77170706e+00 -4.77028042e-01 9.42144573e-01
1.15348661e+00 -6.12616725e-03 4.72683638e-01 3.78316075e-01
-7.37495244e-01 4.31959748e-01 -3.69239002e-02 -2.93015242e-01
8.00802648e-01 -1.10083365e+00 4.24438238e-01 7.12442815e-01
1.80195168e-01 5.02137780e-01 -4.44743261e-02 -5.02244294e-01
-1.34144008e+00 -1.31131327e+00 6.58783674e-01 -4.63905215e-01
5.27708769e-01 -6.03525400e-01 -1.19010925e+00 7.66497672e-01
2.56131701e-02 1.73572943e-01 4.51592863e-01 3.73775303e-01
-5.19379497e-01 -2.56535679e-01 -5.29887736e-01 5.40865421e-01
1.30720520e+00 -9.07078028e-01 -6.63411498e-01 3.12283128e-01
8.61068189e-01 -4.60032880e-01 -8.15566897e-01 5.27536094e-01
4.40928608e-01 -1.09312141e+00 1.23524666e+00 -3.97552431e-01
5.19368529e-01 -2.90091693e-01 -2.43729576e-01 -1.36771095e+00
-7.20207095e-01 -1.34265602e-01 -3.11766379e-02 1.17712462e+00
1.62715167e-01 -6.44990325e-01 3.30130905e-01 1.60810888e-01
-5.06976962e-01 -8.52337599e-01 -9.71762896e-01 -7.21261263e-01
-4.17222604e-02 -4.38034356e-01 5.09586096e-01 8.84716749e-01
-3.57395589e-01 6.68540239e-01 -5.88967204e-01 2.54261136e-01
5.58801115e-01 3.66527021e-01 7.66627491e-01 -1.05597532e+00
-9.13413092e-02 -5.36853552e-01 -1.87413186e-01 -1.76074934e+00
2.87616968e-01 -7.12911129e-01 -2.36087404e-02 -1.49940789e+00
4.77141052e-01 -4.97821867e-01 -6.02249980e-01 6.40836120e-01
-6.02320731e-01 -4.82854918e-02 3.00116688e-01 3.12114954e-01
-1.06845701e+00 1.00706255e+00 1.89014924e+00 7.85315931e-02
-2.52420679e-02 -1.91001222e-01 -9.38968301e-01 6.56130672e-01
4.55110401e-01 -3.20845217e-01 -5.99272668e-01 -8.99377942e-01
-3.51823360e-01 -4.10181917e-02 6.53086305e-01 -1.21678793e+00
4.01080042e-01 -2.80638486e-01 6.44298613e-01 -7.17558503e-01
5.70907414e-01 -8.83679509e-01 -1.18864231e-01 5.83657697e-02
-2.83019722e-01 3.20902988e-02 3.62559080e-01 6.84191167e-01
-5.94168425e-01 3.55638355e-01 7.34830618e-01 2.13935580e-02
-1.16431069e+00 5.76970279e-01 1.02820523e-01 3.77923548e-02
1.11302090e+00 -1.18195534e-01 -3.56814831e-01 -5.15988693e-02
-6.04259372e-01 7.16054678e-01 2.16808960e-01 9.13765669e-01
6.90883458e-01 -1.41343427e+00 -6.41052008e-01 4.82115120e-01
4.05487746e-01 3.44027251e-01 1.04969788e+00 8.46295595e-01
2.44659424e-01 5.55622697e-01 -3.90099108e-01 -9.54205692e-01
-1.17181635e+00 4.50934261e-01 3.02333206e-01 -7.26426840e-01
-5.79568148e-01 9.70807195e-01 1.32467914e+00 -3.85758996e-01
-1.38062283e-01 -2.51155436e-01 -3.08860570e-01 1.92955434e-01
5.60975611e-01 -1.32566035e-01 -1.01540424e-01 -8.41419935e-01
-4.51474249e-01 1.02434146e+00 -3.94836962e-01 2.66861349e-01
1.15984058e+00 -5.54727972e-01 1.01233862e-01 6.46808028e-01
1.39233446e+00 -4.68147993e-01 -1.66587889e+00 -7.28722215e-01
-2.24827901e-01 -6.10321522e-01 1.42123997e-01 -5.92295945e-01
-1.10251272e+00 1.15713632e+00 3.45192492e-01 7.06802448e-03
1.51233459e+00 2.43037105e-01 6.98029399e-01 1.80053875e-01
2.59862453e-01 -9.32205856e-01 5.36448538e-01 8.07610452e-01
1.11083448e+00 -1.42021286e+00 -1.59510851e-01 -5.60576081e-01
-9.49659705e-01 8.70954275e-01 1.13687444e+00 1.63885489e-01
6.39377773e-01 2.45921791e-01 -2.36590996e-01 -4.05527413e-01
-9.53255057e-01 -4.46431339e-01 6.23613656e-01 4.35361117e-01
2.76834279e-01 -1.94581486e-02 3.25885385e-01 7.91032732e-01
2.37340689e-01 -1.81164369e-01 -1.87228993e-01 8.19789767e-01
-3.68155748e-01 -9.19646919e-01 -7.74571300e-02 5.06610215e-01
-9.09693539e-02 -1.70379877e-01 1.59291804e-01 8.31760943e-01
3.45927387e-01 7.79658854e-01 1.10416368e-01 -5.81684470e-01
5.38362205e-01 -5.40077873e-02 3.20786625e-01 -6.77466035e-01
-6.39894307e-01 1.80010006e-01 -1.76551506e-01 -7.83123076e-01
-4.76800144e-01 -4.86854672e-01 -7.79825151e-01 1.02875479e-01
-1.28603712e-01 -3.18934053e-01 3.82030278e-01 1.21489227e+00
6.61766589e-01 1.07520187e+00 7.64431894e-01 -9.91411030e-01
-1.84099272e-01 -1.00324154e+00 -5.90072691e-01 3.82379711e-01
4.05572653e-01 -1.15949821e+00 -5.64279854e-02 -8.13497901e-02] | [9.725807189941406, 1.259037733078003] |
bb0d23dd-d0a2-44cd-b258-be25ae0ef79c | computational-narratology-extracting-tense | null | null | https://aclanthology.org/L14-1256 | https://aclanthology.org/L14-1256.pdf | Computational Narratology: Extracting Tense Clusters from Narrative Texts | Computational Narratology is an emerging field within the Digital Humanities. In this paper, we tackle the problem of extracting temporal information as a basis for event extraction and ordering, as well as further investigations of complex phenomena in narrative texts. While most existing systems focus on news texts and extract explicit temporal information exclusively, we show that this approach is not feasible for narratives. Based on tense information of verbs, we define temporal clusters as an annotation task and validate the annotation schema by showing that the task can be performed with high inter-annotator agreement. To alleviate and reduce the manual annotation effort, we propose a rule-based approach to robustly extract temporal clusters using a multi-layered and dynamic NLP pipeline that combines off-the-shelf components in a heuristic setting. Comparing our results against human judgements, our system is capable of predicting the tense of verbs and sentences with very high reliability: for the most prevalent tense in our corpus, more than 95{\%} of all verbs are annotated correctly. | ['Michael Gertz', 'Jannik Str{\\"o}tgen', 'Thomas B{\\"o}gel'] | 2014-05-01 | null | null | null | lrec-2014-5 | ['morphological-tagging'] | ['natural-language-processing'] | [ 4.85611223e-02 2.79476225e-01 4.24918439e-03 -2.64742553e-01
-7.08601952e-01 -1.11361182e+00 9.88201976e-01 5.67315578e-01
-6.77735031e-01 7.07872391e-01 6.62958562e-01 -3.35965067e-01
-1.84558079e-01 -7.10659862e-01 -2.98453808e-01 -3.21224213e-01
-1.55924484e-01 5.79665601e-01 5.74536443e-01 -3.36803734e-01
1.02711685e-01 3.30779970e-01 -1.50371683e+00 6.81094289e-01
5.48791468e-01 6.36099279e-01 1.39289666e-02 5.05967438e-01
-2.04554990e-01 1.36311173e+00 -4.98607069e-01 -6.55972838e-01
-1.46649539e-01 -5.68827093e-01 -1.24215615e+00 -9.52180475e-02
-3.71350527e-01 5.91081567e-02 -6.87168166e-02 4.29375589e-01
3.00282985e-01 8.08416009e-02 5.19449472e-01 -9.72808242e-01
1.96490198e-01 1.13950622e+00 -1.82831809e-01 3.47473085e-01
7.58530676e-01 -1.64972290e-01 1.15964532e+00 -4.21378165e-01
1.34954429e+00 7.52029479e-01 7.51545966e-01 1.01985060e-01
-1.15148902e+00 -2.29456961e-01 2.38824822e-02 4.41465884e-01
-1.25224650e+00 -4.85142648e-01 8.75716269e-01 -5.71234226e-01
1.30640221e+00 4.68732655e-01 9.37748730e-01 1.28624034e+00
-8.59034657e-02 7.01402068e-01 1.10521615e+00 -6.29353583e-01
3.29867572e-01 1.72347121e-03 3.73870730e-02 2.52332121e-01
-1.74563855e-01 -3.45258385e-01 -8.30812097e-01 -3.39877531e-02
3.08100522e-01 -6.36732817e-01 -5.16542047e-02 2.35954911e-01
-1.61818612e+00 5.44586837e-01 -2.66919851e-01 9.33019936e-01
-7.33980656e-01 -3.24015282e-02 8.45648825e-01 4.28338125e-02
6.89103425e-01 4.25243229e-01 -4.33097512e-01 -7.19238400e-01
-1.14287841e+00 4.13055867e-01 1.06725848e+00 6.31366909e-01
1.81318194e-01 -5.65230846e-01 -1.41750730e-03 7.17083216e-01
-3.77749130e-02 -3.33962366e-02 2.94605672e-01 -9.04142082e-01
4.54749256e-01 6.36708379e-01 3.20277840e-01 -1.07918906e+00
-8.39708924e-01 -2.56501734e-02 -2.68917829e-01 -2.34090641e-01
5.81396997e-01 -4.82464805e-02 -3.95487010e-01 1.75751662e+00
6.09943151e-01 -2.94440356e-03 8.90344307e-02 6.42098129e-01
7.06658959e-01 7.26431549e-01 3.96429569e-01 -1.01025832e+00
1.68572819e+00 -2.12647632e-01 -1.19426572e+00 -1.15569346e-01
8.16353440e-01 -7.71578372e-01 8.60235095e-01 6.16538882e-01
-1.06920969e+00 1.19371749e-01 -9.52653587e-01 -1.43513769e-01
-2.95314044e-01 -1.25449777e-01 7.65548706e-01 2.52586305e-01
-5.93154967e-01 7.44030118e-01 -1.10801327e+00 -6.24018013e-01
1.51196688e-01 3.84407155e-02 -4.17942762e-01 4.69597459e-01
-1.32430708e+00 1.20145130e+00 8.47727597e-01 7.64698535e-02
-2.20814228e-01 -4.82252479e-01 -6.70349121e-01 -1.73942879e-01
6.14517570e-01 -7.64117539e-02 1.38022137e+00 -6.77362382e-01
-1.25056136e+00 1.07431448e+00 -1.19071186e-01 -5.74941099e-01
6.24944389e-01 -3.08618918e-02 -5.69076598e-01 4.36311811e-01
2.59313136e-01 1.81830719e-01 5.36140129e-02 -1.00263226e+00
-6.89652741e-01 -8.27291608e-02 -8.72161463e-02 4.83887643e-02
-4.88793522e-01 6.98201895e-01 -3.66314650e-01 -5.64662695e-01
1.87768146e-01 -9.24344361e-01 3.40799103e-03 -5.07918656e-01
-2.88453490e-01 -5.42493820e-01 4.67318326e-01 -8.96545172e-01
1.85551739e+00 -1.96395957e+00 7.54272118e-02 1.36500046e-01
6.40420690e-02 -2.32745811e-01 5.62228739e-01 7.46173143e-01
2.72949021e-02 3.87349501e-02 -3.57442111e-01 -2.48536229e-01
2.74787486e-01 3.77530307e-01 -4.87099290e-01 4.78121281e-01
1.19950250e-01 7.89832532e-01 -1.17258739e+00 -1.09918976e+00
6.10582940e-02 2.71990508e-01 -2.83613950e-01 1.03834577e-01
-3.52157235e-01 2.16785103e-01 -3.65745753e-01 2.94320047e-01
-9.30139050e-02 -1.40804127e-01 9.30202007e-01 -2.15061437e-02
-5.84840417e-01 9.29698229e-01 -1.26628160e+00 1.65603757e+00
-3.92562300e-01 8.75060022e-01 -2.99070835e-01 -7.82549977e-01
6.17630005e-01 7.99353123e-01 7.96999216e-01 -6.39044702e-01
1.14640892e-01 3.08461279e-01 6.05267584e-02 -8.13597977e-01
7.72102416e-01 -3.34300667e-01 -6.05231464e-01 8.33259702e-01
7.48483688e-02 -1.01035684e-01 7.78935730e-01 2.57335186e-01
1.38397205e+00 5.47729254e-01 8.12680125e-01 -2.36616895e-01
1.41495332e-01 4.83598888e-01 6.33430421e-01 3.30247521e-01
-1.05680250e-01 3.65561694e-01 8.82599294e-01 -5.99664032e-01
-1.25388014e+00 -6.10894084e-01 -1.56434640e-01 9.77853477e-01
-2.65544385e-01 -9.19364631e-01 -7.13295043e-01 -6.07375503e-01
-6.87022090e-01 9.67491508e-01 -7.81495035e-01 3.96513790e-01
-9.06629622e-01 -8.80478799e-01 5.71734011e-01 3.20794791e-01
3.71793732e-02 -1.50289834e+00 -1.15558970e+00 6.50199950e-01
-8.08674932e-01 -1.46368468e+00 1.71005994e-01 3.50786597e-01
-2.07251281e-01 -1.05588400e+00 1.11947104e-01 -3.60379338e-01
1.45451464e-02 -4.41964239e-01 1.18714797e+00 -2.42145687e-01
-1.46425769e-01 1.44461477e-02 -7.37663031e-01 -4.75409359e-01
-5.88836372e-01 6.96131513e-02 -2.01527849e-01 -1.23854779e-01
4.71006036e-01 -7.86591887e-01 -2.46502399e-01 5.39192418e-03
-8.71267915e-01 2.91655689e-01 1.79060251e-02 2.85066217e-01
2.37247720e-01 1.97056949e-01 3.83395970e-01 -8.96375418e-01
5.79012632e-01 -6.89347625e-01 -3.75762433e-01 8.78431275e-02
-3.75806868e-01 -5.76727614e-02 5.07798076e-01 -5.43295205e-01
-1.08174193e+00 2.28154749e-01 -1.59492746e-01 2.66018957e-01
-1.23618238e-01 7.65576422e-01 3.23564559e-02 8.24430764e-01
7.44664729e-01 2.83555984e-02 -5.44756830e-01 -2.39942253e-01
3.96744877e-01 5.24108589e-01 6.52561605e-01 -5.43881774e-01
5.25027037e-01 7.24515736e-01 -2.10053120e-02 -6.07166648e-01
-9.92343187e-01 -4.59054083e-01 -1.03287733e+00 -6.94518328e-01
9.63088453e-01 -7.44762361e-01 -8.40315282e-01 5.21226553e-03
-1.38968730e+00 -4.83853757e-01 -3.81877631e-01 4.59589720e-01
-7.68226385e-01 2.59006202e-01 -5.97523153e-01 -1.08653998e+00
-1.27593488e-01 -6.75012410e-01 9.86322403e-01 -2.04334483e-01
-1.19989979e+00 -9.01326358e-01 2.39084542e-01 2.21851632e-01
-2.26311773e-01 9.70658839e-01 7.43905365e-01 -1.04199016e+00
-1.20125890e-01 -1.29041612e-01 2.32033014e-01 -4.39776748e-01
-1.54003784e-01 2.72382826e-01 -8.55125308e-01 4.62514490e-01
-3.29115242e-02 -3.03570688e-01 3.53091419e-01 1.63496810e-03
4.52306777e-01 -5.47293603e-01 -3.86656016e-01 -1.07927941e-01
1.21106052e+00 2.18798846e-01 8.30681384e-01 6.95428491e-01
3.19766879e-01 1.04508495e+00 7.40544736e-01 7.58001387e-01
6.76844716e-01 8.10431600e-01 1.07528888e-01 4.60129589e-01
1.02271199e-01 -1.11261763e-01 2.29023576e-01 8.88429046e-01
-3.29711288e-01 -4.16358948e-01 -1.42541504e+00 1.10459292e+00
-2.07709026e+00 -1.44474328e+00 -6.98238492e-01 1.70791984e+00
1.33074701e+00 5.43948472e-01 3.61618251e-01 7.08121002e-01
5.04010677e-01 2.37114951e-01 4.02208865e-01 -3.18886757e-01
-1.99507356e-01 2.03343034e-01 7.18010068e-02 3.98233682e-01
-1.11839771e+00 1.07529736e+00 6.33399487e+00 7.19422698e-01
-8.13205183e-01 2.83114821e-01 1.50483251e-01 -2.85982847e-01
-3.22548091e-01 3.33822876e-01 -3.86813253e-01 4.96495157e-01
1.33552265e+00 -1.19330496e-01 2.91423410e-01 5.12765169e-01
6.25236511e-01 -2.81651825e-01 -1.03798604e+00 6.01276040e-01
-6.55386820e-02 -1.51746321e+00 -5.80292344e-01 9.35753956e-02
3.83156747e-01 -2.29370147e-01 -6.94420218e-01 3.02483365e-02
3.54084373e-01 -7.30624437e-01 1.38579714e+00 5.11997461e-01
4.90465641e-01 -6.73555076e-01 6.39913738e-01 3.50420535e-01
-1.21814644e+00 1.94174405e-02 5.01163721e-01 -2.56792277e-01
7.17710793e-01 6.59658730e-01 -1.02098465e+00 5.84502101e-01
7.06016004e-01 7.38564849e-01 -3.41746867e-01 6.42818987e-01
-5.50287843e-01 9.03155744e-01 -7.19718695e-01 -1.98244646e-01
1.65504932e-01 1.35456130e-01 5.95014751e-01 1.65418637e+00
1.94798142e-01 5.53138733e-01 1.50945336e-01 6.30863726e-01
1.76237866e-01 3.23922306e-01 -4.67178404e-01 -2.38593444e-01
6.19775593e-01 1.34913611e+00 -1.44042075e+00 -3.23330998e-01
-7.47559443e-02 7.69505203e-01 4.71706867e-01 -1.51709035e-01
-9.36550558e-01 -1.62227392e-01 6.85876384e-02 2.64766634e-01
3.33531141e-01 -4.85502005e-01 -4.65885282e-01 -9.40811038e-01
3.76278341e-01 -6.58936381e-01 6.56421840e-01 -7.67702103e-01
-7.98670411e-01 7.48267889e-01 2.81037867e-01 -1.18956840e+00
-6.80076361e-01 -2.41973400e-01 -4.12954271e-01 1.52414560e-01
-1.01053846e+00 -1.18282485e+00 2.79213395e-03 3.05420488e-01
2.90062517e-01 5.22302151e-01 7.25172222e-01 3.51394951e-01
-3.88154328e-01 -1.18718199e-01 -4.19111282e-01 2.24596038e-01
5.82196116e-01 -1.23984468e+00 1.22068904e-01 1.08687353e+00
2.69297957e-01 4.02181149e-01 1.19151855e+00 -7.86894560e-01
-1.07073653e+00 -7.60840774e-01 1.81800056e+00 -6.25686228e-01
1.30185676e+00 -5.02606869e-01 -9.12739456e-01 7.07636178e-01
2.81233728e-01 -2.25232020e-01 7.93625891e-01 3.26733112e-01
-4.33544636e-01 4.20658529e-01 -6.75240993e-01 6.65145814e-01
1.25490117e+00 -5.91960371e-01 -1.01141930e+00 6.12394452e-01
5.42788804e-01 -4.79930997e-01 -1.24513590e+00 2.25885406e-01
4.87959027e-01 -9.24778163e-01 4.98555183e-01 -3.62963676e-01
7.20494270e-01 -4.47015554e-01 -6.68046921e-02 -6.87981606e-01
1.80122945e-02 -1.00575554e+00 -1.58987865e-01 1.73742533e+00
5.17626584e-01 -1.67439096e-02 3.07883650e-01 5.24916768e-01
-1.31914735e-01 -4.07829642e-01 -1.15612769e+00 -5.50259709e-01
-3.10680717e-01 -1.09201539e+00 1.88259497e-01 1.26661515e+00
8.95190954e-01 4.64505017e-01 -2.18508989e-01 -1.85259357e-02
1.55355319e-01 8.03316534e-02 4.17688191e-01 -1.33256960e+00
-4.86716107e-02 -4.84217584e-01 -2.34026238e-01 -1.07530929e-01
3.74909639e-01 -7.43878126e-01 2.27462545e-01 -1.74071312e+00
1.15986571e-01 -2.91221648e-01 2.49891058e-01 7.26185918e-01
1.31416634e-01 3.54813218e-01 4.50184830e-02 3.05889845e-01
-8.41392756e-01 5.63075505e-02 5.71905315e-01 2.45081186e-01
-4.20708179e-01 -5.41051567e-01 -3.55313271e-01 9.10925984e-01
5.27506769e-01 -7.07685471e-01 -1.09302372e-01 -7.27638882e-03
8.45294237e-01 7.75778145e-02 2.86883444e-01 -6.97315395e-01
3.39804649e-01 -3.72958541e-01 -2.48800013e-02 -6.27363920e-01
1.41484544e-01 -7.80262351e-01 6.08277857e-01 2.01607332e-01
-3.31137955e-01 -6.44175485e-02 2.65910983e-01 2.59614617e-01
-1.76398337e-01 -2.15557575e-01 4.02081162e-01 -1.94624588e-01
-6.02359414e-01 -3.31234336e-01 -7.59114802e-01 5.15950136e-02
1.21545804e+00 -1.05952613e-01 -1.92192942e-01 -1.97882742e-01
-8.70117843e-01 -9.41906348e-02 2.93777823e-01 2.76811808e-01
1.32699385e-01 -1.23736572e+00 -8.00847471e-01 -6.61417127e-01
1.34328410e-01 -1.80066451e-01 3.51584423e-03 9.40530121e-01
-5.00444829e-01 2.08626807e-01 7.09904507e-02 -3.17904234e-01
-1.19710243e+00 6.29749835e-01 4.57352819e-03 -6.82750463e-01
-7.66252339e-01 3.41635019e-01 -5.32704055e-01 1.50600672e-01
-2.60732144e-01 -1.92907795e-01 -6.06207967e-01 8.36480439e-01
3.93811852e-01 1.15662657e-01 1.52564242e-01 -8.91083598e-01
-6.00844979e-01 8.23762119e-02 2.04340190e-01 -8.16419303e-01
1.86272478e+00 -2.47048661e-01 -3.67454529e-01 8.66488695e-01
7.31867492e-01 2.15973258e-01 -9.61230755e-01 6.01438917e-02
7.36497700e-01 7.83344582e-02 -1.62758231e-01 -7.75248826e-01
-5.90334572e-02 2.39983454e-01 -2.42939368e-01 8.68152261e-01
1.10377645e+00 3.17690462e-01 6.86525047e-01 3.44627500e-02
2.92264521e-01 -1.46328104e+00 -1.95136875e-01 5.14748573e-01
9.05723631e-01 -8.75622272e-01 2.80789435e-01 -5.58242381e-01
-7.68808901e-01 1.01177514e+00 1.23491049e-01 1.08653612e-01
3.93700153e-01 5.21275520e-01 -1.23612583e-01 -4.05345440e-01
-1.10877883e+00 -3.97373825e-01 2.22124040e-01 1.65894106e-01
6.96313679e-01 -3.90992500e-02 -1.02663469e+00 7.66265154e-01
-6.28479421e-01 -1.04500048e-01 6.70419872e-01 1.18204308e+00
-2.04112947e-01 -9.54011738e-01 -2.30858758e-01 -4.32989635e-02
-7.03066051e-01 -1.29186541e-01 -5.26129544e-01 9.95199502e-01
3.38850051e-01 1.02323830e+00 1.26193032e-01 -2.09794849e-01
3.26783478e-01 3.69472474e-01 5.11580586e-01 -4.63789344e-01
-9.76579189e-01 2.76466489e-01 8.76422226e-01 -5.53372741e-01
-9.42236304e-01 -1.26152468e+00 -1.45606983e+00 -8.90654698e-02
-2.49245279e-02 2.60761410e-01 6.76151037e-01 1.40507555e+00
5.99828549e-02 5.19041419e-01 3.05672586e-01 -6.33847535e-01
2.61490166e-01 -9.69249070e-01 -1.84717476e-01 5.92830896e-01
-1.42010674e-01 -5.23574054e-01 -2.79083729e-01 6.88601315e-01] | [9.205469131469727, 9.32336711883545] |
7cf068d9-e4c2-4e40-83e5-7a4add4dc53c | a-fair-and-in-depth-evaluation-of-existing | 2305.14937 | null | https://arxiv.org/abs/2305.14937v1 | https://arxiv.org/pdf/2305.14937v1.pdf | A Fair and In-Depth Evaluation of Existing End-to-End Entity Linking Systems | Existing evaluations of entity linking systems often say little about how the system is going to perform for a particular application. There are four fundamental reasons for this: many benchmarks focus on named entities; it is hard to define which other entities to include; there are ambiguities in entity recognition and entity linking; many benchmarks have errors or artifacts that invite overfitting or lead to evaluation results of limited meaningfulness. We provide a more meaningful and fair in-depth evaluation of a variety of existing end-to-end entity linkers. We characterize the strengths and weaknesses of these linkers and how well the results from the respective publications can be reproduced. Our evaluation is based on several widely used benchmarks, which exhibit the problems mentioned above to various degrees, as well as on two new benchmarks, which address these problems. | ['Natalie Prange', 'Matthias Hertel', 'Hannah Bast'] | 2023-05-24 | null | null | null | null | ['entity-linking'] | ['natural-language-processing'] | [-1.79623112e-01 1.58407688e-01 -4.09659445e-01 -4.89420146e-01
-8.70870411e-01 -9.97594833e-01 5.89502990e-01 6.09907210e-01
-4.67069387e-01 1.06217027e+00 3.57020974e-01 -2.79486775e-01
-2.28128955e-01 -6.72085106e-01 -5.41735053e-01 -5.82023989e-03
-1.20126195e-01 7.90362954e-01 6.42032087e-01 -5.02029896e-01
3.55460495e-02 3.13587964e-01 -1.30363762e+00 7.52136335e-02
7.46189058e-01 4.50652510e-01 -4.45885956e-01 4.23727900e-01
-4.32477891e-01 7.32219040e-01 -9.90901053e-01 -9.48506355e-01
-9.18759778e-03 -4.88556251e-02 -1.26621509e+00 -6.08106911e-01
5.59240043e-01 8.87871757e-02 -6.30863309e-02 9.77147281e-01
6.33132458e-01 -6.28127828e-02 5.90494037e-01 -1.44897830e+00
-7.16637611e-01 9.81283009e-01 -1.71314538e-01 3.11101675e-01
7.30160236e-01 -3.18601787e-01 1.15680170e+00 -8.15079033e-01
1.11589921e+00 9.46892500e-01 1.10216856e+00 2.86300510e-01
-1.10618603e+00 -5.47680676e-01 -1.02760352e-01 -1.13252863e-01
-1.62686181e+00 -6.41388774e-01 1.86177213e-02 -4.14838344e-01
1.26713872e+00 5.33786595e-01 2.02330649e-01 6.68510199e-01
-1.18755624e-02 2.54653394e-01 6.52767897e-01 -3.20418835e-01
-5.15037253e-02 1.97343484e-01 3.51857036e-01 3.90740544e-01
9.27226782e-01 -1.33218139e-01 -2.10979626e-01 -5.16681492e-01
3.19724530e-01 -4.90075797e-01 -1.00982562e-01 -3.97599727e-01
-1.45441484e+00 5.04897058e-01 3.21948647e-01 5.87224960e-01
-4.72057275e-02 6.04753792e-02 6.83132529e-01 3.14076930e-01
-6.69348910e-02 8.60732019e-01 -6.47080123e-01 -3.30284417e-01
-1.03645515e+00 5.38663507e-01 1.29139030e+00 1.50593460e+00
8.43555391e-01 -3.89143646e-01 3.19771692e-02 6.15985215e-01
2.44002730e-01 6.85987473e-02 2.95367479e-01 -8.94913852e-01
6.96594477e-01 7.77541161e-01 4.15362865e-01 -9.49970663e-01
-6.79440439e-01 -2.22240612e-01 -1.94334164e-01 1.44291937e-01
7.32713580e-01 -2.33177632e-01 -5.53254664e-01 1.73710096e+00
9.45372283e-02 -1.54899850e-01 2.62828887e-01 5.55144846e-01
1.50758326e+00 2.34519631e-01 6.20860815e-01 1.39143229e-01
1.62574816e+00 -8.09385359e-01 -8.47155035e-01 -2.72403389e-01
1.06010675e+00 -1.19467616e+00 7.32411087e-01 -3.29997331e-01
-1.31505537e+00 -3.70741993e-01 -1.08676732e+00 -2.86297321e-01
-1.00801432e+00 -9.15114060e-02 6.89963758e-01 7.49245882e-01
-9.05362546e-01 6.23785973e-01 -6.76629782e-01 -7.41949141e-01
-1.93751886e-01 4.08571571e-01 -6.45680666e-01 4.36687410e-01
-1.45195639e+00 1.27615142e+00 8.53558779e-01 -2.30924964e-01
-7.32430490e-03 -7.96378791e-01 -7.65708447e-01 6.29053544e-03
4.75343764e-01 -7.23235309e-01 1.28322184e+00 -5.64252079e-01
-5.97127557e-01 1.05397642e+00 3.07270475e-02 -2.80926853e-01
4.27693009e-01 -1.49842247e-01 -9.01618004e-01 -2.67910749e-01
5.04035473e-01 5.58093309e-01 -3.62688363e-01 -1.09097099e+00
-8.18702281e-01 -1.84732378e-02 3.30045909e-01 7.08466768e-02
-8.40507224e-02 4.72467154e-01 -8.77080798e-01 -9.64663804e-01
-7.66062811e-02 -8.55598569e-01 -9.57383960e-02 -3.63550603e-01
-5.45284033e-01 -1.59146249e-01 5.88471413e-01 -4.59261984e-01
1.79435408e+00 -1.86030960e+00 -2.08060473e-01 -9.78517681e-02
2.90203065e-01 2.51506597e-01 3.06440257e-02 9.20773745e-01
-3.36398602e-01 7.02618659e-01 -7.99524710e-02 -2.97416467e-02
2.51901954e-01 1.34236082e-01 -2.08217427e-01 1.71769053e-01
3.16204548e-01 9.75408435e-01 -1.00153351e+00 -7.84641325e-01
-2.13565797e-01 4.58647132e-01 -2.93167591e-01 8.46987311e-03
-6.01468012e-02 -4.80642505e-02 -4.47722971e-01 7.62940288e-01
4.96627152e-01 -3.09253633e-01 2.57015169e-01 -4.05472726e-01
-1.89673364e-01 7.23820746e-01 -1.48783612e+00 1.53889441e+00
8.81769136e-02 5.20654142e-01 -2.01612368e-01 -3.27134430e-01
7.21561372e-01 5.15119672e-01 4.62134063e-01 -3.76601219e-01
-3.22736591e-01 6.17065549e-01 -1.11362226e-01 -5.44019103e-01
1.02568495e+00 1.05202563e-01 -3.21211457e-01 3.50591481e-01
1.10250019e-01 2.33057752e-01 8.41123044e-01 2.96009779e-01
1.30560732e+00 1.84854999e-01 4.55309808e-01 -2.43207902e-01
3.65058303e-01 3.62704843e-01 8.39679182e-01 5.99568725e-01
-1.25359222e-01 5.94618082e-01 5.40616572e-01 -3.39025944e-01
-1.17966771e+00 -1.02995276e+00 -2.59678453e-01 1.03867233e+00
1.94798529e-01 -9.25178289e-01 -6.02514088e-01 -8.83378506e-01
-1.62167195e-02 6.88959479e-01 -5.29555082e-01 1.62377790e-01
-5.88490188e-01 -7.39974797e-01 1.22375059e+00 8.43706369e-01
1.39796883e-01 -9.03176725e-01 -5.28019786e-01 3.89990509e-01
-3.47594708e-01 -1.25682247e+00 -2.04133719e-01 1.21120684e-01
-7.23389030e-01 -1.16865754e+00 -5.21225810e-01 -7.82937944e-01
5.42971015e-01 -2.51856316e-02 1.88842833e+00 3.44269037e-01
5.91611676e-02 1.67550579e-01 -2.90050566e-01 -3.43611360e-01
-4.90120679e-01 6.24172091e-01 -2.40538165e-01 -9.38032269e-01
7.54850328e-01 -1.93282023e-01 -3.87103558e-01 6.13430023e-01
-9.07526135e-01 -5.11001229e-01 4.82230753e-01 7.13215530e-01
3.14952433e-01 -1.59741059e-01 6.83690190e-01 -1.46803606e+00
5.82193494e-01 -6.75804496e-01 -4.58007246e-01 6.10362113e-01
-8.10429931e-01 8.01537558e-02 3.00469160e-01 -1.18853496e-02
-8.13911378e-01 -1.33681014e-01 -4.05995190e-01 2.60914564e-01
-1.69672042e-01 7.59352863e-01 -2.41165951e-01 -1.50692416e-02
7.83684671e-01 -5.37536681e-01 -5.04715919e-01 -5.13418674e-01
3.17042261e-01 3.92746717e-01 6.86128676e-01 -8.59650791e-01
8.06890011e-01 1.58825055e-01 -4.60733324e-01 -3.57743502e-01
-5.13986230e-01 -4.86296803e-01 -5.65430045e-01 3.06153148e-01
5.51779091e-01 -9.14333344e-01 -5.89588821e-01 5.41596767e-03
-1.10800564e+00 2.88208574e-03 -3.99790913e-01 2.76077271e-01
-2.72416234e-01 5.41840672e-01 -6.63832843e-01 -2.86999017e-01
-2.07146630e-01 -1.17654705e+00 8.31594169e-01 3.65029424e-01
-8.53899598e-01 -1.15745366e+00 2.27873117e-01 2.23463982e-01
5.10999143e-01 4.82429832e-01 9.99379992e-01 -1.07483113e+00
-2.45923564e-01 -2.91306108e-01 -3.00792724e-01 -2.05831379e-01
9.17846989e-03 5.86471081e-01 -5.99901080e-01 -2.66539484e-01
-7.18642890e-01 -1.14099354e-01 3.94439101e-01 -1.93547368e-01
3.40935469e-01 -4.03180756e-02 -7.40362704e-01 3.56093109e-01
1.55675113e+00 1.13344319e-01 7.76118398e-01 9.39439952e-01
4.60410029e-01 6.82456732e-01 8.17429245e-01 3.09330933e-02
8.23318005e-01 9.57846224e-01 2.34397471e-01 -2.06824511e-01
-6.67715967e-02 -1.81522131e-01 8.05600062e-02 6.63170099e-01
-1.83793396e-01 -4.54760581e-01 -1.18175030e+00 7.18902707e-01
-1.84462249e+00 -9.09332097e-01 -5.39943457e-01 2.28023267e+00
1.03310084e+00 2.35966176e-01 3.03875357e-01 -1.26603842e-01
7.51129270e-01 -1.75146610e-02 -6.64011836e-02 -2.04878539e-01
-4.24747527e-01 6.81208074e-02 7.26498604e-01 2.38552660e-01
-1.13754427e+00 9.86208856e-01 8.21366501e+00 4.03368235e-01
-8.54325533e-01 3.40290144e-02 1.03949554e-01 2.58685112e-01
-4.00530636e-01 5.11649013e-01 -1.14452314e+00 3.60455394e-01
1.10981905e+00 -4.40850198e-01 -1.00620434e-01 6.21833563e-01
-2.50535697e-01 2.34542653e-01 -1.25673282e+00 5.16596675e-01
-4.33386028e-01 -1.48145604e+00 -2.25577325e-01 -1.71268493e-01
5.21008313e-01 2.52367347e-01 -2.55613655e-01 3.46918464e-01
6.23017311e-01 -1.03523409e+00 8.40790570e-01 4.28420871e-01
8.04693222e-01 -5.40268242e-01 9.59349453e-01 -1.54705867e-01
-1.21733868e+00 2.82036871e-01 -2.99557865e-01 1.71828002e-01
3.30482811e-01 4.42218721e-01 -6.28858089e-01 7.27056801e-01
7.65449584e-01 3.82849097e-01 -9.35197115e-01 1.30076230e+00
-2.92484462e-01 3.25035095e-01 -3.89788479e-01 -4.98849200e-03
1.15697511e-01 1.75620541e-01 4.74640697e-01 1.69951892e+00
2.79859453e-01 -1.13862887e-01 -1.15504339e-01 6.61903083e-01
-3.39100331e-01 2.95510292e-01 -5.63535511e-01 -2.13658407e-01
9.73217547e-01 1.34354019e+00 -8.04113925e-01 -4.15194541e-01
-8.30736876e-01 4.52949673e-01 3.37585062e-01 3.18696409e-01
-1.05620503e+00 -9.04191256e-01 7.44937003e-01 2.42949411e-01
1.94951326e-01 -1.83735400e-01 -8.04094896e-02 -1.15485120e+00
8.31446350e-02 -9.74498272e-01 8.94955993e-01 -7.38710344e-01
-1.15085566e+00 7.06213355e-01 1.52713150e-01 -1.03564000e+00
-2.86699355e-01 -3.16770047e-01 -4.06717598e-01 7.32698381e-01
-1.45201361e+00 -9.45891500e-01 -3.07265997e-01 1.85989469e-01
-1.40475839e-01 1.28163114e-01 1.08810294e+00 1.07269907e+00
-5.92604935e-01 9.94665504e-01 -2.58649811e-02 6.33695245e-01
1.23753595e+00 -1.45545626e+00 9.61049080e-01 7.31511772e-01
1.51399210e-01 1.07091808e+00 8.61282051e-01 -6.29551828e-01
-1.08798778e+00 -9.81445312e-01 1.49391055e+00 -9.69716668e-01
8.01208675e-01 -5.39837591e-02 -1.01060760e+00 1.06107926e+00
3.19719046e-01 5.70207834e-02 8.57079268e-01 4.68462437e-01
-6.24557137e-01 1.89797953e-01 -1.14908600e+00 4.76000547e-01
1.10739505e+00 -2.87867874e-01 -8.68835568e-01 2.30488077e-01
5.13031304e-01 -7.00648546e-01 -1.58510566e+00 4.66299176e-01
5.43894768e-01 -9.71238017e-01 1.06302011e+00 -9.28248107e-01
1.62736461e-01 -5.78919649e-01 -3.46645340e-02 -8.92242968e-01
-3.07089835e-01 -5.83367169e-01 2.86105573e-02 1.97857404e+00
1.03868127e+00 -8.47617388e-01 6.87278628e-01 1.08619022e+00
-5.43829277e-02 -3.99067938e-01 -6.27126515e-01 -8.79106402e-01
1.32517904e-01 -9.28482264e-02 1.05645168e+00 1.39341915e+00
2.18395531e-01 5.88534594e-01 2.99173117e-01 1.56292751e-01
3.44696522e-01 -1.08094640e-01 8.37754428e-01 -1.17302871e+00
3.42062749e-02 -5.69151878e-01 -6.22430682e-01 -6.94318414e-01
-6.33957535e-02 -7.11441576e-01 -1.94625288e-01 -1.61990488e+00
1.50050834e-01 -1.04404068e+00 -3.18101794e-01 7.12891638e-01
-3.16298395e-01 2.83796698e-01 1.47247151e-01 4.83873338e-01
-8.76374602e-01 -1.86308712e-01 5.22955000e-01 -1.54195465e-02
-2.23927703e-02 -2.35522360e-01 -9.91402745e-01 6.50526941e-01
3.98451269e-01 -6.97711766e-01 -3.67517695e-02 -5.34769237e-01
8.25081646e-01 -8.36642683e-02 1.02708591e-02 -1.09814739e+00
4.06119704e-01 1.47877019e-02 1.80347487e-01 -6.06109858e-01
-2.97282729e-02 -6.49606407e-01 7.17962146e-01 2.02820107e-01
-2.52264649e-01 5.75410068e-01 2.73783863e-01 1.84997752e-01
-5.17509401e-01 -6.16049290e-01 4.82593834e-01 -6.02329932e-02
-1.00531244e+00 6.60584569e-02 -3.84047218e-02 5.45773923e-01
9.72726643e-01 -1.63614675e-01 -7.55694449e-01 -1.62299037e-01
-6.74176097e-01 4.01840359e-01 1.01303816e+00 5.86442232e-01
-1.65350690e-01 -1.43249416e+00 -7.56061018e-01 -3.52316409e-01
5.11357248e-01 -2.43724942e-01 -3.44547719e-01 7.77716339e-01
-7.87565291e-01 5.06378412e-01 -2.11555153e-01 -1.61113620e-01
-1.18308389e+00 4.54962283e-01 3.13643068e-01 -4.66493398e-01
-4.23094660e-01 5.03750503e-01 -2.43387982e-01 -6.79034352e-01
7.97970444e-02 -2.69409776e-01 -2.53462344e-01 1.96408927e-01
3.70507598e-01 5.16166151e-01 4.25036281e-01 -8.97992909e-01
-7.42072940e-01 4.48017716e-01 -1.26494929e-01 3.60293984e-02
1.22272372e+00 -3.13275284e-03 -1.76072925e-01 2.52935737e-01
9.52889919e-01 3.60801846e-01 -6.21092618e-01 -1.42847374e-01
5.79199970e-01 -2.21135661e-01 -5.28007865e-01 -9.41891730e-01
-9.28485990e-01 4.55399960e-01 1.98465645e-01 6.08478069e-01
8.12446058e-01 -2.30721049e-02 6.95935726e-01 3.06489497e-01
2.86249429e-01 -1.21283960e+00 -7.64801204e-01 6.02664530e-01
4.67860073e-01 -1.06726635e+00 3.53328079e-01 -6.90496743e-01
-5.16439497e-01 1.22294724e+00 5.71344137e-01 2.39793509e-01
3.66967887e-01 5.43650806e-01 2.46951893e-01 -3.64800334e-01
-5.75049281e-01 -2.88276941e-01 6.71838298e-02 5.40110767e-01
1.09147429e+00 -1.95253298e-01 -8.49708021e-01 6.11491323e-01
-2.10915312e-01 -2.75416207e-02 4.50379908e-01 1.14353693e+00
-9.26490203e-02 -1.57299411e+00 -3.26319844e-01 1.90070912e-01
-9.67990458e-01 -3.91788781e-03 -6.71005547e-01 1.26165080e+00
1.09856352e-01 8.43479395e-01 -1.45075023e-01 -2.91888505e-01
8.53583992e-01 1.12478305e-02 2.52130389e-01 -7.04368293e-01
-1.05467987e+00 -3.05891216e-01 8.12160254e-01 -4.45209652e-01
-5.35523653e-01 -7.95189619e-01 -1.65498102e+00 -5.52187324e-01
-6.37632966e-01 5.30210912e-01 4.98748511e-01 6.62110806e-01
6.61641479e-01 3.46995562e-01 -1.28915891e-01 -1.27232581e-01
-2.62777030e-01 -6.98739827e-01 -2.18486384e-01 6.13027573e-01
-1.91548631e-01 -6.73254490e-01 -2.09999215e-02 2.24800933e-05] | [9.397372245788574, 8.699065208435059] |
e8c12746-c7ac-4dfb-b4f3-ade57494259d | the-clickbait-challenge-2017-towards-a | 1812.10847 | null | http://arxiv.org/abs/1812.10847v1 | http://arxiv.org/pdf/1812.10847v1.pdf | The Clickbait Challenge 2017: Towards a Regression Model for Clickbait Strength | Clickbait has grown to become a nuisance to social media users and social
media operators alike. Malicious content publishers misuse social media to
manipulate as many users as possible to visit their websites using clickbait
messages. Machine learning technology may help to handle this problem, giving
rise to automatic clickbait detection. To accelerate progress in this
direction, we organized the Clickbait Challenge 2017, a shared task inviting
the submission of clickbait detectors for a comparative evaluation. A total of
13 detectors have been submitted, achieving significant improvements over the
previous state of the art in terms of detection performance. Also, many of the
submitted approaches have been published open source, rendering them
reproducible, and a good starting point for newcomers. While the 2017 challenge
has passed, we maintain the evaluation system and answer to new registrations
in support of the ongoing research on better clickbait detectors. | ['Benno Stein', 'Tim Gollub', 'Matthias Hagen', 'Martin Potthast'] | 2018-12-27 | null | null | null | null | ['clickbait-detection'] | ['natural-language-processing'] | [-2.30350286e-01 -1.72533616e-01 -4.40870404e-01 -3.21426272e-01
-1.07701290e+00 -9.17212665e-01 9.07844186e-01 6.29608512e-01
-6.97975636e-01 6.13108695e-01 -2.67786644e-02 -4.46950674e-01
2.29617104e-01 -4.45442379e-01 -4.78737801e-01 7.72358477e-02
3.37274224e-02 4.16181386e-01 1.16494131e+00 -6.76589832e-02
7.03612268e-01 -3.63613223e-03 -1.44867158e+00 4.64730889e-01
5.28269053e-01 8.84004176e-01 -3.33558396e-02 8.26996088e-01
-1.04778349e-01 5.02914786e-01 -4.86793607e-01 -5.97582340e-01
2.62283146e-01 -1.42801553e-01 -7.12273240e-01 -3.90299797e-01
9.48778927e-01 -2.46823907e-01 -3.86917740e-01 1.05924773e+00
1.38766766e-01 -1.07622758e-01 1.97471619e-01 -1.27874303e+00
-7.94722021e-01 2.50399232e-01 -3.73231083e-01 6.47888005e-01
4.56515372e-01 1.43813670e-01 1.64034045e+00 -6.19053781e-01
6.82433307e-01 8.05207312e-01 4.63592499e-01 3.67961854e-01
-1.37466979e+00 -5.93120217e-01 -1.08095482e-01 3.75320345e-01
-1.20493746e+00 -3.21704447e-01 1.91784516e-01 -6.01880193e-01
7.86660910e-01 1.03053164e+00 5.11849523e-01 1.32022643e+00
-3.00255567e-01 1.49635863e+00 1.19443178e+00 -2.20533729e-01
2.98186779e-01 3.98131490e-01 6.38153434e-01 3.20967197e-01
7.44884253e-01 -1.57929942e-01 -5.08991003e-01 -6.78724945e-01
1.44719705e-01 -5.97372577e-02 1.83073238e-01 -1.52704328e-01
-1.10105944e+00 1.10914254e+00 4.99344289e-01 5.36178946e-01
-1.19573534e-01 2.01035932e-01 4.69644070e-01 2.99739391e-01
5.59298575e-01 8.79171789e-01 -3.22301716e-01 -4.60273057e-01
-1.08930957e+00 8.75315070e-01 1.20289516e+00 6.62633300e-01
5.61302543e-01 -7.55185902e-01 -1.47653982e-01 9.02455270e-01
5.06740272e-01 4.25482184e-01 4.21915382e-01 -5.87527752e-01
5.01916885e-01 6.94635570e-01 4.54465330e-01 -1.12398994e+00
-4.91884425e-02 -3.57519329e-01 -7.76264817e-03 6.70052394e-02
7.16162622e-01 3.28545421e-01 -7.29595065e-01 8.94872248e-01
1.33332476e-01 -2.61801034e-01 -1.05631781e+00 7.27848589e-01
2.87203044e-01 3.79005224e-01 1.96737066e-01 2.23593593e-01
1.62186837e+00 -7.99960732e-01 -8.30175698e-01 -2.40473926e-01
7.87486792e-01 -1.32412720e+00 1.11849546e+00 3.09242129e-01
-9.69458342e-01 1.42759718e-02 -1.06666720e+00 -5.91686517e-02
-9.86656785e-01 -3.02389264e-01 6.31317735e-01 8.23903263e-01
-6.10979199e-01 3.90442073e-01 -6.83291435e-01 -6.64828598e-01
6.44631207e-01 -7.30593354e-02 1.79000065e-01 2.12165534e-01
-1.20975697e+00 9.52452362e-01 1.34376705e-01 -3.20076525e-01
-4.58776444e-01 -8.34632277e-01 8.86482000e-02 -2.33922169e-01
6.41160905e-01 -1.96544096e-01 1.69051552e+00 -3.70287716e-01
-7.77354538e-01 1.13685369e+00 1.82914287e-01 -5.62946677e-01
1.11881506e+00 -5.35950124e-01 -6.84506536e-01 -1.27649888e-01
5.19398630e-01 1.15889855e-01 7.06231594e-01 -5.47823787e-01
-1.05269718e+00 -2.71310538e-01 -1.37767896e-01 -4.68536288e-01
-4.06769305e-01 4.87768620e-01 -4.21853662e-01 -4.43423331e-01
1.52595609e-01 -1.06410992e+00 6.65224418e-02 -5.58234891e-03
-8.74001622e-01 -6.83112621e-01 9.67268527e-01 -7.25788176e-01
1.52897227e+00 -1.96442473e+00 -5.14785409e-01 1.28198475e-01
7.07838953e-01 7.36547112e-01 2.57063389e-01 8.67213249e-01
2.69276470e-01 8.63211811e-01 4.29302335e-01 -1.14225030e-01
5.29404581e-01 -4.22482222e-01 -5.21110177e-01 4.18197870e-01
-1.99670345e-01 9.25619006e-01 -8.74776065e-01 -2.03176096e-01
-3.49013507e-02 -4.33386005e-02 -4.26689267e-01 -3.58355530e-02
-5.08531094e-01 -8.16129148e-02 -1.09726989e+00 4.93439317e-01
6.89545691e-01 -7.56471753e-01 -1.71531096e-01 1.16919026e-01
-5.44143498e-01 7.90760756e-01 -7.85002053e-01 8.31042349e-01
1.03429854e-01 1.09070826e+00 3.51342373e-02 -1.49189755e-01
5.60596406e-01 -7.80822337e-02 2.97163188e-01 -7.05167472e-01
-2.03624200e-02 6.07652664e-01 -1.46148592e-01 -6.16041660e-01
6.61470115e-01 4.66746062e-01 6.49220198e-02 8.28810632e-01
-2.84833878e-01 3.04333538e-01 5.32088995e-01 7.18213618e-01
1.47982001e+00 -4.56330508e-01 1.47100568e-01 -8.55055749e-02
2.80617207e-01 8.47330913e-02 -3.36464435e-01 1.37823761e+00
-6.21514857e-01 5.06629467e-01 5.72192371e-01 -5.01437426e-01
-1.33509946e+00 -8.97520959e-01 -2.07132865e-02 1.59578192e+00
-2.68541515e-01 -7.69484818e-01 -6.31544113e-01 -9.15704310e-01
3.28940898e-01 6.33593857e-01 -5.56159437e-01 3.24637949e-01
-4.88810778e-01 -5.63518643e-01 7.43755460e-01 1.49980769e-01
7.48365939e-01 -6.89199090e-01 -4.55285572e-02 1.44840926e-01
-1.91641986e-01 -1.01444495e+00 -5.19638598e-01 -1.36709005e-01
-3.29843700e-01 -1.36935914e+00 -7.47352660e-01 -3.82206351e-01
8.23672563e-02 3.86570662e-01 6.75326109e-01 3.14650893e-01
-6.17385626e-01 4.07700688e-01 -2.83685863e-01 -4.81629342e-01
-2.86386281e-01 7.54615009e-01 -4.04320300e-01 2.69523030e-03
7.57651269e-01 -5.75205684e-02 -7.16252923e-01 7.14884400e-01
-1.06368506e+00 -5.16509950e-01 9.22116116e-02 3.09026629e-01
-2.80547559e-01 -6.12376392e-01 4.69924092e-01 -1.03609622e+00
1.10560858e+00 -5.66873193e-01 -9.58262205e-01 9.85953137e-02
-7.49950349e-01 -1.64758816e-01 -1.33288115e-01 -2.54945874e-01
-5.64094782e-01 -3.56880605e-01 -1.13086239e-01 3.39306921e-01
2.48431265e-02 3.64421338e-01 5.94617426e-01 -2.68170834e-01
9.83568907e-01 9.53253880e-02 -3.25176641e-02 -9.22235250e-01
2.56120622e-01 7.55402982e-01 -3.06209475e-01 1.12677850e-01
1.26139450e+00 3.82203817e-01 -5.27929544e-01 -8.77410531e-01
-1.13036251e+00 -9.78177011e-01 -1.54321089e-01 -1.66560978e-01
8.41420412e-01 -4.30116832e-01 -1.11197531e+00 4.31496501e-01
-9.63344693e-01 1.62310779e-01 2.68641412e-01 2.93207496e-01
6.70247972e-02 4.48420107e-01 -7.68589079e-01 -6.66256785e-01
2.17641458e-01 -7.09414482e-01 6.08800352e-01 -2.19835211e-02
-4.53875989e-01 -1.01249146e+00 4.96043772e-01 8.59558582e-01
1.04262590e+00 -9.82174724e-02 4.35168177e-01 -1.68294513e+00
-1.17681170e+00 -1.12573552e+00 -5.34651101e-01 2.57535875e-01
-2.29763821e-01 -4.34690565e-02 -1.03555417e+00 -1.27546579e-01
-3.15093398e-01 -4.23873067e-01 1.01798570e+00 -1.24650136e-01
8.53047967e-01 -3.78022045e-01 -7.02730060e-01 -3.75894070e-01
9.48007047e-01 -2.99546182e-01 5.99279761e-01 9.78160977e-01
2.87738055e-01 2.28010625e-01 4.29189414e-01 4.14358258e-01
2.05404177e-01 7.45609164e-01 5.07910192e-01 2.94752151e-01
-9.51979011e-02 -5.63687146e-01 -2.86601335e-01 6.33370340e-01
5.34417592e-02 -1.03655949e-01 -8.97706866e-01 6.51129410e-02
-1.80443895e+00 -9.73803818e-01 -6.38202548e-01 2.38651252e+00
4.09314036e-01 5.02633810e-01 4.71066624e-01 -1.21041089e-01
9.19852793e-01 5.45420110e-01 -3.29352736e-01 -2.53408141e-02
3.23760092e-01 -1.25635445e-01 9.04500544e-01 3.99831504e-01
-1.18317032e+00 7.32405126e-01 7.15558004e+00 8.39157522e-01
-1.03701556e+00 2.81007439e-01 2.54243284e-01 1.71255320e-01
-9.73062590e-03 4.68435064e-02 -1.24200654e+00 8.47258449e-01
8.05354536e-01 -4.16071206e-01 5.29679716e-01 1.13764656e+00
3.13646853e-01 -1.89339653e-01 -9.28851187e-01 6.94096982e-01
-3.64049859e-02 -1.60035169e+00 -4.31718439e-01 4.67365950e-01
2.37415567e-01 6.27335668e-01 2.12694898e-01 6.46066964e-01
2.69532681e-01 -5.26102364e-01 3.98487717e-01 2.37309068e-01
-7.07032830e-02 2.05078125e-01 3.08199406e-01 3.18484575e-01
-6.84606910e-01 -9.69213434e-03 -6.47427216e-02 -7.77857006e-02
1.13670640e-01 4.47606683e-01 -1.37508404e+00 -2.67128617e-01
5.41405201e-01 4.74812508e-01 -1.34677660e+00 1.82081366e+00
-1.16777398e-01 9.98747766e-01 -4.73625094e-01 -9.88843083e-01
4.93814498e-01 3.80249411e-01 9.15959477e-01 1.10075951e+00
-3.18861037e-01 -6.16752684e-01 1.74644470e-01 8.19574058e-01
-2.51535028e-01 1.03772059e-01 -2.52448916e-01 -7.09682107e-01
4.06335771e-01 1.25863838e+00 -7.25229383e-01 -4.57571179e-01
-3.56334060e-01 7.85472810e-01 3.46580148e-01 2.85006464e-01
-9.84822392e-01 -5.32973289e-01 3.92890424e-01 7.96635866e-01
2.09435076e-01 -3.87185663e-01 -4.53486927e-02 -1.34250748e+00
2.92391866e-01 -6.27929449e-01 4.76097703e-01 -6.46459401e-01
-1.48917735e+00 1.77497149e-01 -9.37274918e-02 -1.01859951e+00
1.09450772e-01 -5.17315686e-01 -2.88644284e-01 4.97591138e-01
-1.39312160e+00 -8.66875589e-01 -1.61888704e-01 -2.26095319e-02
5.56004941e-01 1.13183912e-03 4.05871034e-01 8.81087065e-01
-3.83524895e-01 4.06701595e-01 2.74996042e-01 1.44264057e-01
9.99607325e-01 -1.08662915e+00 7.16208160e-01 3.47740859e-01
2.13186249e-01 1.05479276e+00 9.49755669e-01 -7.81355619e-01
-1.32084167e+00 -8.13346803e-01 1.41764450e+00 -9.39092755e-01
1.68169057e+00 -6.87970281e-01 -9.01568890e-01 9.90594566e-01
2.80408829e-01 -8.94051939e-02 6.80628896e-01 5.98423660e-01
-7.63444901e-01 4.93121706e-02 -8.19774151e-01 6.17200613e-01
5.41664541e-01 -5.82043707e-01 -7.47637749e-01 1.04686940e+00
2.83626348e-01 1.05390340e-01 -3.48534167e-01 -4.23752666e-01
7.08941936e-01 -8.68564725e-01 7.32322514e-01 -8.34619880e-01
8.02032873e-02 -2.01307952e-01 8.98792595e-02 -8.20964098e-01
-3.91085565e-01 -8.42722893e-01 1.63198382e-01 1.19229531e+00
7.10587204e-01 -9.76363659e-01 1.11568296e+00 9.24276054e-01
1.32030189e-01 -1.32523775e-01 -9.58142817e-01 -1.00841141e+00
-2.80782461e-01 -4.57009614e-01 2.28050221e-02 8.52946520e-01
1.04048543e-01 4.91084546e-01 -4.34749156e-01 -4.24568176e-01
7.34954357e-01 -4.07404840e-01 8.96845400e-01 -1.22455657e+00
-1.07828349e-01 -6.00557685e-01 -5.25166631e-01 -1.43051958e+00
-5.74258506e-01 -1.09725893e+00 -3.48086983e-01 -1.12629080e+00
3.71819288e-01 -3.55779201e-01 -2.47832209e-01 2.78153419e-01
-4.44261990e-02 7.88423717e-01 3.67681354e-01 5.67584217e-01
-1.50178349e+00 -1.22423835e-01 9.74154055e-01 -2.75062740e-01
7.04918355e-02 4.66698647e-01 -7.23844945e-01 5.45007229e-01
6.60608292e-01 -6.91950679e-01 3.78011376e-01 7.20533729e-02
8.04985046e-01 -7.08048880e-01 5.18906534e-01 -8.13741148e-01
4.55418259e-01 1.35938346e-01 1.34660468e-01 -6.95383072e-01
2.54939854e-01 -5.60436726e-01 -3.19558591e-01 4.72953975e-01
-9.76565003e-01 -2.36044794e-01 -1.39989555e-01 8.25161099e-01
-6.19086921e-02 -4.48761731e-01 6.99595332e-01 -2.45379314e-01
-3.35137963e-01 5.01758866e-02 -7.94286549e-01 3.83130759e-01
9.90725160e-01 -6.05451465e-02 -8.19323719e-01 -4.60594058e-01
-8.80826175e-01 9.00229812e-02 1.69729274e-02 6.63542867e-01
6.23016153e-03 -1.01003397e+00 -5.01895845e-01 -5.95255801e-03
4.90158468e-01 -1.23838985e+00 -1.65377393e-01 8.61479282e-01
-6.47363603e-01 9.45628464e-01 7.66190439e-02 -4.83487070e-01
-1.30971456e+00 3.16790313e-01 -8.78782943e-02 -2.87987411e-01
-2.70672321e-01 6.49546146e-01 -4.18665379e-01 -1.61940739e-01
2.29611889e-01 -9.77305621e-02 2.53722966e-01 2.12585747e-01
9.77451444e-01 7.76459157e-01 3.12791109e-01 -1.65293869e-02
-2.46804252e-01 -2.46110350e-01 -7.64195681e-01 -5.01918532e-02
1.22525501e+00 -1.24797262e-01 -2.28506792e-02 4.96918350e-01
1.59440827e+00 2.36271203e-01 -4.68128502e-01 -1.32095084e-01
7.55776703e-01 -1.02003860e+00 -1.29804790e-01 -9.62466002e-01
-3.62611085e-01 4.18739825e-01 4.39639002e-01 1.48574829e+00
-5.14716171e-02 3.91658485e-01 9.26302493e-01 6.41829073e-01
7.27409199e-02 -1.19661951e+00 4.39789921e-01 4.79672015e-01
7.82264471e-01 -1.48365366e+00 2.22563341e-01 -5.48381746e-01
-1.28851771e-01 1.01539433e+00 2.94625729e-01 -3.15430015e-01
7.71471322e-01 -3.28909844e-01 -1.10869050e-01 -3.93877178e-01
-7.68360198e-01 4.73471321e-02 2.08366975e-01 1.63066745e-01
4.92705464e-01 2.63945032e-02 -9.33794558e-01 -3.85385822e-03
2.12160304e-01 1.68764740e-02 4.04001564e-01 9.74127650e-01
-8.55357170e-01 -1.27040088e+00 -2.59579837e-01 9.26918089e-01
-8.37769985e-01 -1.86926335e-01 -6.95849478e-01 9.89305794e-01
-6.58539176e-01 8.09104741e-01 -3.74136299e-01 -1.09692857e-01
3.80588651e-01 1.00989014e-01 3.73438024e-03 -5.68500161e-01
-8.59416604e-01 -3.68559025e-02 5.84861398e-01 -6.47005856e-01
2.51463622e-01 -7.29653299e-01 -3.44791710e-01 -3.41049016e-01
-6.79833710e-01 4.21610892e-01 1.19817483e+00 6.48812592e-01
4.55645293e-01 -9.68485326e-02 4.30435598e-01 -3.32306921e-01
-9.82735097e-01 -9.86815929e-01 -6.35497034e-01 4.87745315e-01
2.79275954e-01 -5.01464188e-01 -8.47306490e-01 -2.54151523e-01] | [7.746814727783203, 9.767853736877441] |
bb400e91-8a4e-4db6-968f-37d4d2780bf4 | himfr-a-hybrid-masked-face-recognition | 2209.08930 | null | https://arxiv.org/abs/2209.08930v1 | https://arxiv.org/pdf/2209.08930v1.pdf | HiMFR: A Hybrid Masked Face Recognition Through Face Inpainting | To recognize the masked face, one of the possible solutions could be to restore the occluded part of the face first and then apply the face recognition method. Inspired by the recent image inpainting methods, we propose an end-to-end hybrid masked face recognition system, namely HiMFR, consisting of three significant parts: masked face detector, face inpainting, and face recognition. The masked face detector module applies a pretrained Vision Transformer (ViT\_b32) to detect whether faces are covered with masked or not. The inpainting module uses a fine-tune image inpainting model based on a Generative Adversarial Network (GAN) to restore faces. Finally, the hybrid face recognition module based on ViT with an EfficientNetB3 backbone recognizes the faces. We have implemented and evaluated our proposed method on four different publicly available datasets: CelebA, SSDMNV2, MAFA, {Pubfig83} with our locally collected small dataset, namely Face5. Comprehensive experimental results show the efficacy of the proposed HiMFR method with competitive performance. Code is available at https://github.com/mdhosen/HiMFR | ['Md Baharul Islam', 'Md Imran Hosen'] | 2022-09-19 | null | null | null | null | ['facial-inpainting', 'image-inpainting'] | ['computer-vision', 'computer-vision'] | [ 3.21861178e-01 1.18906602e-01 2.36484006e-01 -4.76278216e-01
-8.04603696e-01 -4.06297177e-01 3.43665421e-01 -1.15233195e+00
-2.05198992e-02 5.78772604e-01 2.83994135e-02 2.13761944e-02
4.44241941e-01 -7.01372564e-01 -9.90605772e-01 -7.80817091e-01
3.12879205e-01 1.98943198e-01 -4.47020866e-02 -1.57644087e-03
-2.43830234e-02 8.79950047e-01 -1.76073325e+00 6.04054213e-01
5.65796375e-01 1.26929593e+00 7.00420290e-02 5.80132008e-01
3.97731699e-02 1.08956850e+00 -4.79829878e-01 -7.22973764e-01
6.67678833e-01 -5.42181432e-01 -6.33086145e-01 4.61940646e-01
6.69951081e-01 -7.83892632e-01 -6.09396696e-01 1.00139260e+00
6.35103106e-01 -2.00636446e-01 3.41234922e-01 -1.32835507e+00
-8.02640736e-01 4.22356188e-01 -8.25092137e-01 3.85225117e-02
3.01233023e-01 4.21347857e-01 1.28513545e-01 -1.42923570e+00
5.11870563e-01 1.73926413e+00 5.66460311e-01 1.07385218e+00
-9.59037304e-01 -9.50605631e-01 -2.17103601e-01 1.66237265e-01
-1.66257906e+00 -1.17032170e+00 7.34595597e-01 -1.99171454e-01
5.55263698e-01 2.51131386e-01 2.90838350e-02 1.06317282e+00
-1.48024902e-01 5.19246399e-01 1.21061468e+00 -2.80342817e-01
-6.90405443e-02 2.18484811e-02 -2.02716485e-01 9.89703894e-01
-2.20002189e-01 4.01540160e-01 -4.30853367e-01 -1.98716268e-01
7.85251677e-01 3.67889464e-01 -4.72458750e-01 2.08257213e-01
-4.78325397e-01 6.08081698e-01 4.36559170e-01 -1.54186279e-01
-4.76669312e-01 -2.97644436e-02 8.22851583e-02 4.28520173e-01
5.73763072e-01 -5.25214612e-01 -1.94662645e-01 6.25087976e-01
-1.16335273e+00 7.55863711e-02 5.41730285e-01 7.94776082e-01
9.95443463e-01 3.12918872e-01 -5.24503767e-01 1.12415433e+00
6.00436628e-01 5.27154088e-01 2.51343071e-01 -8.97899568e-01
1.39699057e-01 4.43340182e-01 -6.85033128e-02 -8.16064596e-01
1.80925414e-01 9.05368775e-02 -7.87507713e-01 5.09987533e-01
-1.95603296e-02 -1.30770892e-01 -1.29585683e+00 1.60939181e+00
5.30844450e-01 5.96517444e-01 1.43664569e-01 1.03005683e+00
1.21646380e+00 6.13390863e-01 -7.49706551e-02 -5.11183888e-02
1.63458252e+00 -1.22697890e+00 -5.67883432e-01 -3.40795487e-01
-1.83667824e-01 -1.15129995e+00 6.91458464e-01 1.05843529e-01
-1.32386208e+00 -8.74091089e-01 -8.84822547e-01 -2.43344814e-01
-8.70195627e-02 5.36452353e-01 8.07861239e-02 7.11303174e-01
-1.39839041e+00 3.84392500e-01 -6.59391642e-01 -2.30351701e-01
9.13780451e-01 5.81754088e-01 -6.61364675e-01 -4.16854888e-01
-8.01464498e-01 5.87737620e-01 8.32066908e-02 6.06566727e-01
-1.65141141e+00 -3.75966370e-01 -8.31637740e-01 1.32394999e-01
2.20040560e-01 -5.44566154e-01 1.14700019e+00 -1.44242656e+00
-1.54775274e+00 1.23168659e+00 -3.47974420e-01 -2.61637449e-01
6.45135760e-01 -7.45111611e-03 -5.39698124e-01 3.15000504e-01
-7.10923225e-02 8.02963674e-01 1.44474554e+00 -1.22798467e+00
-2.63340145e-01 -7.11353183e-01 -3.23935300e-01 -9.06510130e-02
1.99882947e-02 5.38181782e-01 -6.62111700e-01 -6.27447844e-01
-1.54999778e-01 -6.67999089e-01 2.69741654e-01 2.72359759e-01
-4.87936884e-01 1.08639821e-01 1.31549942e+00 -1.12796044e+00
7.06136405e-01 -2.33187842e+00 -9.09136534e-02 -1.19024599e-02
1.44935653e-01 5.96135557e-01 -4.69444245e-01 3.71325403e-01
-3.00296575e-01 -2.41402552e-01 -4.61093903e-01 -8.19308937e-01
-2.78485179e-01 -1.91689637e-02 -5.23078740e-01 6.68420255e-01
4.82431501e-01 8.67134154e-01 -2.20014974e-01 -2.28834972e-01
7.39845484e-02 1.01114213e+00 -4.06640261e-01 7.65537798e-01
-1.42513096e-01 3.88969809e-01 -1.51034966e-01 1.15062547e+00
1.41901410e+00 1.42536417e-01 4.12282944e-02 -1.21484838e-01
9.21785161e-02 -2.53119171e-01 -1.01225996e+00 1.35601759e+00
-1.92794725e-01 2.61744916e-01 6.97074890e-01 -4.48685080e-01
1.02964604e+00 3.93439084e-01 -3.69033553e-02 -4.42679316e-01
3.91728818e-01 7.21350536e-02 -3.57969314e-01 -6.44252658e-01
3.06948442e-02 1.06588155e-01 5.58257580e-01 2.86664724e-01
2.54695684e-01 4.62720424e-01 -1.80999950e-01 2.20661927e-02
1.16331041e+00 2.62110323e-01 -1.89894885e-01 -6.90659955e-02
8.70363653e-01 -4.26321924e-01 5.66357732e-01 1.66514531e-01
-2.00537294e-01 8.94923449e-01 2.56143451e-01 -3.91900539e-01
-7.42282152e-01 -9.62136686e-01 9.54090729e-02 1.01914656e+00
-1.97135597e-01 7.54471570e-02 -1.21971881e+00 -7.38088071e-01
8.44256729e-02 7.90159553e-02 -7.72210956e-01 -1.99278653e-01
-4.97341812e-01 -4.52165693e-01 6.78630531e-01 3.47535640e-01
1.03114533e+00 -1.56644559e+00 -1.03915989e-01 -1.78462982e-01
-7.61398450e-02 -1.08831656e+00 -8.04715037e-01 -3.66419435e-01
-3.80164772e-01 -1.18750882e+00 -7.51385212e-01 -1.27091074e+00
8.69197249e-01 4.26094085e-01 7.74964988e-01 5.02772331e-01
-4.30891007e-01 2.42665056e-02 -3.89313623e-02 -3.05012558e-02
-4.09956664e-01 -4.09311146e-01 -1.14128985e-01 6.20533705e-01
2.76882619e-01 -5.59824467e-01 -9.10616338e-01 3.42514396e-01
-8.99449110e-01 -2.63880044e-01 6.87309146e-01 8.24157178e-01
5.35481334e-01 -2.38440245e-01 5.59391022e-01 -9.87128198e-01
1.14360698e-01 -4.95105028e-01 -7.14626193e-01 2.33790979e-01
-1.34361207e-01 -3.99421930e-01 5.05360126e-01 -3.42623800e-01
-1.21981955e+00 4.65135098e-01 -5.77954948e-01 -1.01850867e+00
-1.95252955e-01 -2.25847319e-01 -8.10694277e-01 -3.70846778e-01
3.74480009e-01 5.21850526e-01 2.80188799e-01 -6.52860284e-01
1.78907096e-01 1.03620160e+00 9.00016844e-01 -2.72262961e-01
8.66590977e-01 6.85032070e-01 -3.74465644e-01 -6.70895159e-01
-4.35690790e-01 1.05993673e-01 -2.62352616e-01 -2.57929295e-01
8.63445759e-01 -1.25845420e+00 -9.59547281e-01 8.40050220e-01
-1.23898172e+00 -2.41606146e-01 1.11183196e-01 -1.11811787e-01
-3.42868537e-01 2.43916482e-01 -8.57632756e-01 -8.07056010e-01
-8.44747245e-01 -1.32975411e+00 1.34609377e+00 4.00360852e-01
4.88455176e-01 -3.01772624e-01 -1.96525112e-01 6.92332625e-01
5.42985797e-01 3.03988695e-01 3.35781634e-01 -2.54469424e-01
-8.02931726e-01 6.50414452e-02 -2.99246609e-01 6.74277067e-01
2.88203806e-01 5.97348996e-02 -1.51417029e+00 -5.27646601e-01
2.39451483e-01 -4.19578999e-01 1.14086580e+00 1.13977306e-01
1.30953336e+00 -5.47675371e-01 -3.44504863e-01 9.25853848e-01
1.42764521e+00 2.53954798e-01 1.16942620e+00 -3.21312070e-01
7.55102754e-01 4.74058896e-01 1.94373250e-01 2.06006557e-01
2.58015007e-01 5.11843979e-01 4.79486465e-01 -2.33099237e-01
-5.34309328e-01 -2.69542545e-01 8.21946800e-01 2.76102722e-01
2.56990880e-01 -2.27546796e-01 -5.38054049e-01 2.57980108e-01
-1.51946652e+00 -1.01921260e+00 2.54377514e-01 1.81655371e+00
7.36055553e-01 -3.91746014e-01 4.29788195e-02 -1.22786667e-02
1.07693458e+00 -8.14822130e-03 -5.49018323e-01 -1.51502460e-01
-1.49119556e-01 5.01721442e-01 2.48263732e-01 5.41429937e-01
-9.20837641e-01 1.21670270e+00 5.20935774e+00 7.99713016e-01
-1.31058919e+00 5.22096634e-01 1.00794327e+00 7.66209606e-03
2.83086866e-01 -1.41909420e-01 -8.40329707e-01 6.96595967e-01
7.46626377e-01 3.70710969e-01 9.71709967e-01 7.47500658e-01
1.95551813e-02 2.61730522e-01 -9.37200129e-01 1.03339911e+00
3.89409393e-01 -1.20909452e+00 2.79317331e-02 -1.45508066e-01
4.98647898e-01 -7.04397261e-02 1.91851020e-01 2.49376789e-01
5.96546978e-02 -1.29630220e+00 6.20588899e-01 6.62206352e-01
1.14781857e+00 -8.18992436e-01 4.13634002e-01 1.37733892e-01
-1.23976374e+00 -1.40950665e-01 -5.51938832e-01 1.94187894e-01
-2.74853796e-01 3.88018906e-01 -7.60827899e-01 3.02174270e-01
8.49316120e-01 3.31014842e-01 -4.79935735e-01 5.71442962e-01
-2.81039774e-01 5.31455040e-01 -1.06654316e-01 8.12288046e-01
-3.46019089e-01 1.68640241e-02 2.45700091e-01 7.42301524e-01
3.17562491e-01 2.61357963e-01 1.34102125e-02 1.03388178e+00
-7.06892252e-01 -2.09533364e-01 -5.13396740e-01 1.49902835e-01
7.03815937e-01 1.65748549e+00 -4.50035512e-01 -2.18550831e-01
-5.18097103e-01 1.28653193e+00 3.39410543e-01 2.82814920e-01
-9.85147059e-01 -1.57919109e-01 7.27066457e-01 2.02392370e-01
6.85082674e-01 4.51109320e-01 2.59818703e-01 -1.06443167e+00
3.14092785e-01 -1.23956096e+00 3.89188379e-01 -9.16870236e-01
-1.28801632e+00 1.25303507e+00 -5.25691807e-01 -8.76134038e-01
1.68139800e-01 -4.72573608e-01 -7.94652283e-01 1.13638771e+00
-1.50119877e+00 -1.52532363e+00 -7.46420324e-01 1.06038022e+00
6.68160200e-01 -5.18262684e-01 6.77476943e-01 6.66758537e-01
-1.06432915e+00 8.62882257e-01 -3.17794591e-01 4.01505530e-01
7.55366325e-01 -4.91095603e-01 5.39360821e-01 1.00994754e+00
-3.68000746e-01 3.87683094e-01 1.90766349e-01 -6.98615611e-01
-1.67850196e+00 -1.77134895e+00 4.34434801e-01 -2.24401414e-01
1.74593143e-02 -6.84278548e-01 -1.02460277e+00 9.94813979e-01
5.77795625e-01 4.61205840e-01 1.40583828e-01 -9.37677145e-01
-5.25185168e-01 -4.13102567e-01 -1.92197669e+00 3.58064413e-01
1.12519264e+00 -5.21638513e-01 -2.28581667e-01 4.49817598e-01
5.22877753e-01 -3.51120055e-01 -5.19783258e-01 4.85607833e-01
4.22596365e-01 -9.48455572e-01 9.98826504e-01 -1.38060972e-01
5.02560854e-01 -5.95678926e-01 -2.86877394e-01 -9.34777319e-01
-1.40952259e-01 -7.20073879e-01 -1.79248646e-01 1.74636376e+00
-3.64729352e-02 -7.75695741e-01 8.28786373e-01 2.65288711e-01
-2.35229190e-02 -6.85553730e-01 -8.68314981e-01 -5.95293164e-01
-2.67702192e-01 2.94879764e-01 1.04193962e+00 7.23639429e-01
-8.47798347e-01 6.68041855e-02 -5.87797284e-01 5.62461019e-01
7.70308793e-01 -9.08838864e-03 7.15375721e-01 -7.56476283e-01
-2.87494928e-01 -2.64320001e-02 -2.30990350e-01 -6.99765503e-01
4.53850180e-01 -8.98379028e-01 9.20656621e-02 -1.00772882e+00
3.72215033e-01 1.65564846e-02 -5.04843295e-02 9.81645346e-01
-9.42985415e-02 8.35642576e-01 1.61151234e-02 1.97889969e-01
-8.60372856e-02 6.53438270e-01 1.03886724e+00 -1.50880381e-01
1.96512669e-01 -1.99917972e-01 -8.38562489e-01 5.34851670e-01
6.51998878e-01 -5.39738595e-01 -1.22520901e-01 -6.13426507e-01
-7.43307590e-01 3.01454037e-01 6.63019240e-01 -1.06979692e+00
1.52905285e-01 1.84171066e-01 9.12619352e-01 -4.36646879e-01
5.94385862e-01 -8.44799519e-01 5.16067088e-01 4.10816133e-01
3.94844636e-02 2.94907391e-01 3.83561671e-01 9.93207693e-02
-1.79747164e-01 1.65556669e-01 1.30190587e+00 -4.50546183e-02
-6.05824828e-01 6.36215806e-01 -5.72673138e-03 -3.43575418e-01
1.23892450e+00 -5.88728115e-02 -5.80917299e-01 -1.36069683e-02
-6.09871387e-01 1.12143092e-01 4.56653476e-01 4.15435731e-01
1.02258730e+00 -1.33556283e+00 -1.08968866e+00 7.52392232e-01
-2.42244035e-01 -2.54682869e-01 6.04936779e-01 5.73981583e-01
-6.66072667e-01 -1.46950677e-01 -4.84132916e-01 -3.85268450e-01
-1.62534201e+00 7.40583360e-01 6.79612994e-01 2.99959928e-01
-3.03509653e-01 1.15199947e+00 2.64625102e-01 -3.80632609e-01
3.67325813e-01 3.23624879e-01 1.87492352e-02 -2.59989887e-01
9.45706248e-01 2.65865713e-01 2.46073425e-01 -9.49616492e-01
-6.34359777e-01 4.37707841e-01 -2.78635830e-01 9.63049158e-02
1.28090215e+00 -2.94445585e-02 -5.10717511e-01 -5.05006075e-01
1.17823541e+00 -4.08198796e-02 -1.36908865e+00 -1.59402296e-01
-5.54640651e-01 -5.91267526e-01 -9.50673148e-02 -8.47813725e-01
-1.91247404e+00 7.02141583e-01 1.02885699e+00 -2.73459703e-01
1.76779938e+00 -1.38973132e-01 8.46851230e-01 -1.02227360e-01
3.17321837e-01 -3.32214206e-01 -9.52468514e-02 1.75652280e-01
1.23341250e+00 -1.08194137e+00 -3.10433269e-01 -6.56883061e-01
-5.15152633e-01 8.20337951e-01 9.30589080e-01 -3.71567875e-01
7.38400042e-01 4.23180401e-01 3.82845789e-01 -2.14187615e-02
-7.74986327e-01 -8.02267790e-02 -2.39236783e-02 3.70698005e-01
1.31302744e-01 -2.20380753e-01 7.00079724e-02 6.33981168e-01
-6.40030503e-02 2.27250278e-01 2.07817435e-01 8.98637414e-01
-1.61663637e-01 -1.08561277e+00 -6.46942139e-01 1.82543889e-01
-7.44743586e-01 -8.16659704e-02 -6.77802384e-01 4.35240179e-01
3.94102335e-01 1.12854373e+00 -4.52828854e-02 -5.49500525e-01
1.95833266e-01 1.30386755e-01 6.03365421e-01 -5.77514410e-01
-8.79169285e-01 -4.63036261e-03 -3.11346680e-01 -7.40522325e-01
-1.62586406e-01 -3.86458635e-01 -1.08819950e+00 -6.18478656e-01
-6.97855353e-02 -1.72463566e-01 3.88750762e-01 7.41831124e-01
8.27960908e-01 2.78222263e-01 1.01147032e+00 -9.58338320e-01
-3.98296118e-01 -1.10789764e+00 -4.29414272e-01 2.14582250e-01
5.37905812e-01 -5.05850077e-01 -1.70477942e-01 2.34053403e-01] | [12.873003005981445, 0.059548269957304] |
179357c4-e6c3-4d59-8cb7-6d06fe8d531f | an-edge-enhanced-hierarchical-graph-to-tree | null | null | https://aclanthology.org/2021.findings-emnlp.127 | https://aclanthology.org/2021.findings-emnlp.127.pdf | An Edge-Enhanced Hierarchical Graph-to-Tree Network for Math Word Problem Solving | Math word problem solving has attracted considerable research interest in recent years. Previous works have shown the effectiveness of utilizing graph neural networks to capture the relationships in the problem. However, these works did not carefully take the edge label information and the long-range word relationship across sentences into consideration. In addition, during generation, they focus on the most relevant areas of the currently generated word, while neglecting the rest of the problem. In this paper, we propose a novel Edge-Enhanced Hierarchical Graph-to-Tree model (EEH-G2T), in which the math word problems are represented as edge-labeled graphs. Specifically, an edge-enhanced hierarchical graph encoder is used to incorporate edge label information. This encoder updates the graph nodes hierarchically in two steps: sentence-level aggregation and problem-level aggregation. Furthermore, a tree-structured decoder with a split attention mechanism is applied to guide the model to pay attention to different parts of the input problem. Experimental results on the MAWPS and Math23K dataset showed that our EEH-G2T can effectively improve performance compared with state-of-the-art methods. | ['Zhongyu Wei', 'Qi Zhang', 'Qinzhuo Wu'] | null | null | null | null | findings-emnlp-2021-11 | ['math-word-problem-solving', 'math-word-problem-solving', 'math-word-problem-solving'] | ['knowledge-base', 'reasoning', 'time-series'] | [ 3.17617595e-01 3.61964464e-01 7.63446167e-02 -3.73378247e-01
-5.55704236e-01 -3.14236641e-01 9.67142954e-02 3.86573434e-01
-8.45445246e-02 7.37450838e-01 3.47337961e-01 -4.12060350e-01
-1.07562393e-02 -1.23847055e+00 -7.72067606e-01 -2.08872989e-01
5.21360151e-02 3.59850943e-01 2.99028754e-01 -2.61288941e-01
3.63831133e-01 -2.05152839e-01 -9.08863127e-01 3.38708848e-01
1.39687097e+00 6.46571219e-01 6.85708106e-01 6.69673979e-01
-5.53261757e-01 1.11106408e+00 -7.27854192e-01 -6.36642933e-01
2.54616737e-02 -8.24489832e-01 -8.88986349e-01 3.30817252e-01
3.65504116e-01 -2.09526226e-01 -5.82968950e-01 1.13478577e+00
4.92978126e-01 2.75891900e-01 3.43064308e-01 -1.03107452e+00
-1.19270360e+00 1.25872755e+00 -6.65026248e-01 1.39701203e-01
4.21121150e-01 -2.92773265e-02 1.37499499e+00 -7.25384116e-01
4.91009325e-01 1.32489991e+00 2.06119746e-01 4.04048622e-01
-8.53920102e-01 -6.16237104e-01 6.37657821e-01 5.89587510e-01
-1.43645155e+00 9.23698768e-02 9.48723674e-01 -1.38620824e-01
1.32805943e+00 1.14905648e-01 8.03245068e-01 4.72363174e-01
4.47894275e-01 8.70549560e-01 6.86112761e-01 -4.24270570e-01
1.36666313e-01 -4.12553906e-01 5.22633076e-01 1.05087316e+00
3.13483894e-01 -5.41999102e-01 -4.31350350e-01 1.35801747e-01
6.87741637e-01 -3.87073845e-01 -2.98192710e-01 2.98734512e-02
-8.74595582e-01 7.86292017e-01 6.92779005e-01 1.64104491e-01
-4.12425458e-01 2.40995988e-01 1.99171394e-01 1.31480679e-01
6.23442411e-01 5.54711044e-01 -2.28213385e-01 1.37282312e-02
-7.34397411e-01 1.73184499e-01 7.13858545e-01 1.45410252e+00
6.68371856e-01 6.23606890e-02 -6.06240988e-01 6.94483876e-01
4.16632921e-01 -3.31444331e-02 3.58555079e-01 -5.02107799e-01
1.19518936e+00 1.01429296e+00 -5.05862176e-01 -1.16488695e+00
-1.67626023e-01 -6.73764288e-01 -8.52801740e-01 -6.05088592e-01
-5.90600818e-02 -2.42026299e-01 -1.07183003e+00 1.80933416e+00
1.96669385e-01 3.18845212e-01 1.29250273e-01 7.77088106e-01
1.06855345e+00 1.09405744e+00 4.97163413e-03 -1.10380068e-01
1.25954807e+00 -1.55538917e+00 -9.52878594e-01 -5.24228752e-01
7.95127809e-01 -4.92858082e-01 1.00408912e+00 9.82731357e-02
-1.50026560e+00 -6.11020088e-01 -9.18142319e-01 -3.63526940e-01
-2.41919607e-01 -9.57045481e-02 4.19296145e-01 2.95428187e-01
-1.22657704e+00 4.71850127e-01 -5.16405642e-01 -1.36570603e-01
2.28177756e-01 2.47823030e-01 1.26950905e-01 -3.69991899e-01
-1.58215678e+00 7.60834575e-01 6.65481091e-01 3.10474962e-01
-5.06340563e-01 -4.18784767e-01 -1.30341005e+00 4.54269201e-01
6.88810170e-01 -8.60421896e-01 1.25370538e+00 -3.91367316e-01
-1.34542215e+00 3.36261183e-01 -3.85383278e-01 -3.21393490e-01
-1.03828590e-03 1.37955025e-02 -2.28413031e-01 1.29756883e-01
8.85050148e-02 6.28410399e-01 4.78527635e-01 -1.03899515e+00
-4.37888622e-01 -2.19524711e-01 4.37717140e-01 6.49512410e-01
-4.21329677e-01 -2.05987737e-01 -1.06313765e+00 -8.22700441e-01
1.44530237e-01 -6.72610760e-01 -3.64326984e-01 -6.47343397e-01
-5.52190781e-01 -5.61639428e-01 3.74263763e-01 -8.93719852e-01
1.97555518e+00 -1.73775637e+00 6.36291981e-01 -3.26760742e-03
4.52637851e-01 1.83242932e-01 -4.74908799e-01 7.86701143e-01
-1.85510814e-02 3.40335459e-01 -3.00884157e-01 -2.90135652e-01
8.30405951e-02 2.68431723e-01 -1.71293505e-02 -1.38546586e-01
4.03709680e-01 1.32626808e+00 -1.11070752e+00 -6.87885702e-01
-1.22922234e-01 1.16884626e-01 -7.85658181e-01 3.10861945e-01
-4.76500869e-01 4.73812129e-03 -6.27744079e-01 3.21308494e-01
6.27046764e-01 -5.98137319e-01 2.55762458e-01 -8.82833749e-02
2.07324862e-01 5.58660984e-01 -9.29436743e-01 1.77110970e+00
-4.08819646e-01 2.72233576e-01 -1.03844255e-01 -8.26101303e-01
7.63241649e-01 5.01149930e-02 2.74612308e-02 -7.98355341e-01
1.13755733e-01 -1.74282268e-01 2.83932686e-01 -4.66679662e-01
7.92532027e-01 2.14879304e-01 -1.71773225e-01 3.56417209e-01
-4.85832505e-02 -3.11826020e-01 7.32244015e-01 7.64749944e-01
1.22414577e+00 -3.99073679e-03 2.20744386e-01 -2.14838441e-02
7.47606814e-01 -1.35729238e-01 5.02546072e-01 5.75501025e-01
2.09038213e-01 5.45431793e-01 7.00094342e-01 3.01733445e-02
-6.09917700e-01 -6.10330462e-01 4.95736152e-01 8.70772779e-01
3.66152436e-01 -1.06744266e+00 -9.80311930e-01 -7.44486392e-01
-2.64860064e-01 1.00449264e+00 -3.96856070e-01 -4.52386409e-01
-8.99111211e-01 -6.24623835e-01 2.04152301e-01 5.64099848e-01
7.03505933e-01 -1.17186093e+00 -3.40822637e-02 5.34677684e-01
-3.82278264e-01 -1.32882345e+00 -1.00506258e+00 -1.75839171e-01
-7.29518235e-01 -8.47547293e-01 -5.75743794e-01 -1.17173743e+00
1.06895781e+00 3.47794443e-01 1.25845182e+00 3.65658492e-01
-1.55196413e-01 1.67868510e-01 -6.82521939e-01 -1.80190638e-01
-2.17870429e-01 3.68865877e-01 -5.73144436e-01 -1.12518430e-01
9.44942608e-02 -3.13072205e-01 -3.27992469e-01 4.14454266e-02
-8.37150216e-01 4.57846344e-01 6.55109048e-01 6.93469644e-01
3.75981182e-01 4.68644053e-01 5.81997812e-01 -1.06259608e+00
1.16168094e+00 -4.40487444e-01 -5.67761242e-01 6.16761923e-01
-4.89081144e-01 1.28540158e-01 8.31864715e-01 -2.99591310e-02
-8.50424170e-01 -1.80729941e-01 -1.94328696e-01 -1.90532789e-01
2.96858817e-01 1.09306276e+00 -4.31608468e-01 1.54973730e-01
-7.86319077e-02 4.57514167e-01 -4.63267773e-01 -1.05929017e-01
4.23760861e-01 5.47291219e-01 2.14027405e-01 -7.00597525e-01
7.33779073e-01 -4.20871764e-01 -9.59021002e-02 -6.43086851e-01
-9.72543538e-01 -2.06878692e-01 -3.91075432e-01 -2.61639535e-01
1.12354219e+00 -7.33429492e-01 -4.51909155e-01 6.04771733e-01
-1.50005257e+00 -4.95801061e-01 -7.63985291e-02 3.60123545e-01
-1.66048959e-01 6.74867153e-01 -9.78473723e-01 -8.50321352e-01
-5.00319183e-01 -1.33271718e+00 1.05198658e+00 3.79071116e-01
-1.88375097e-02 -1.11517179e+00 -3.14203173e-01 5.78230441e-01
2.34190434e-01 -1.41067699e-01 1.43621206e+00 -4.50531632e-01
-8.64262879e-01 -6.34949608e-03 -5.13742268e-01 1.62894055e-01
2.33442560e-01 -3.66622120e-01 -2.40295261e-01 -6.77162930e-02
-1.95521012e-01 -3.12840343e-01 8.43433142e-01 1.89110428e-01
1.48876667e+00 -3.57350677e-01 -1.41368166e-01 3.45426798e-01
1.25324333e+00 2.27435932e-01 6.43787920e-01 -1.28819898e-01
1.09185910e+00 4.37945724e-01 3.47529024e-01 1.15453124e-01
9.97493029e-01 5.80523193e-01 5.07292092e-01 -7.32247159e-02
-3.10622841e-01 -5.41699827e-01 3.10014755e-01 1.53725827e+00
1.86768562e-01 -9.93257999e-01 -8.75663161e-01 5.22917449e-01
-1.90949786e+00 -5.10491848e-01 -5.29795945e-01 1.71936953e+00
9.44775283e-01 3.35659117e-01 -3.48874688e-01 -5.15157208e-02
1.02335143e+00 3.46320033e-01 -3.86643857e-01 -4.26295400e-01
1.35656282e-01 2.03123495e-01 3.29345435e-01 7.12439418e-01
-7.58652389e-01 1.32289004e+00 5.47432470e+00 9.40521836e-01
-6.11487329e-01 -2.00146481e-01 5.13870478e-01 2.21540898e-01
-3.99266511e-01 -6.64008409e-02 -7.78492153e-01 3.96790117e-01
6.35568619e-01 -5.75215220e-01 4.83685464e-01 4.43474025e-01
1.30720258e-01 -1.25117116e-02 -9.57084119e-01 6.95751667e-01
4.36843663e-01 -1.12194753e+00 4.14280295e-01 -6.08283952e-02
7.30272651e-01 -5.07154167e-01 -2.35772908e-01 5.61645269e-01
4.14643735e-01 -8.58532429e-01 5.22383571e-01 4.15022284e-01
5.00359178e-01 -7.81200886e-01 6.25919282e-01 4.21562463e-01
-1.68738484e+00 1.48532137e-01 -4.44155246e-01 -2.67820925e-01
5.06401122e-01 5.06070137e-01 -7.78581142e-01 1.07126427e+00
1.66700572e-01 8.70262980e-01 -6.41505539e-01 8.67291808e-01
-7.80169964e-01 6.94908082e-01 9.58146080e-02 -3.78309935e-01
6.28174663e-01 -3.73254210e-01 3.33229452e-01 1.07192492e+00
4.79999781e-01 5.78454614e-01 4.27194357e-01 9.50154185e-01
-3.89187753e-01 1.83894843e-01 -3.64693314e-01 -4.17836130e-01
2.94973999e-01 1.26563430e+00 -6.91310585e-01 -4.81122285e-01
-5.59245646e-01 1.12966692e+00 7.17553854e-01 2.67208725e-01
-9.90101397e-01 -6.94254398e-01 1.49502948e-01 8.89692530e-02
4.53806996e-01 -4.63826060e-01 -1.40171811e-01 -1.21332645e+00
7.93020427e-02 -6.42018974e-01 4.73154783e-01 -1.03392708e+00
-1.30451190e+00 6.02892518e-01 -4.19200808e-02 -7.22024262e-01
6.09160140e-02 -4.14425820e-01 -7.88344204e-01 1.06305027e+00
-1.55505836e+00 -9.42448735e-01 -2.79522270e-01 3.15470278e-01
9.23201501e-01 5.89538813e-02 4.97706592e-01 1.99235722e-01
-8.52010190e-01 5.37357807e-01 -5.36883235e-01 3.51517767e-01
3.44565958e-01 -1.36714399e+00 8.21352243e-01 1.08646715e+00
6.27630129e-02 7.13998854e-01 2.71334171e-01 -1.15467453e+00
-1.56108439e+00 -1.45671999e+00 1.31886995e+00 3.66530940e-02
7.37778962e-01 -5.24268985e-01 -1.05426693e+00 5.96375167e-01
7.08556175e-01 -3.23599786e-01 4.86781299e-01 -1.61610007e-01
9.04134512e-02 1.30821422e-01 -5.47034323e-01 8.09992373e-01
1.40548289e+00 -1.72309026e-01 -7.82184482e-01 5.51871538e-01
1.24613595e+00 -8.48083973e-01 -5.84263146e-01 2.85831273e-01
-1.59120142e-01 -3.20700467e-01 6.66969776e-01 -6.99053824e-01
9.68889058e-01 -2.07226321e-01 8.95781964e-02 -1.68405604e+00
-7.25843608e-01 -5.25179565e-01 -4.41593289e-01 1.27035308e+00
6.15668118e-01 -3.94678533e-01 7.84397006e-01 4.85129714e-01
-5.91154337e-01 -1.00923252e+00 -5.39074183e-01 -6.07340097e-01
-1.15326561e-01 -2.77087331e-01 6.31632626e-01 7.75356650e-01
2.47538969e-01 1.02674127e+00 -3.83512467e-01 3.21012884e-01
4.63844806e-01 2.55950093e-01 4.61426854e-01 -8.08157742e-01
-3.54255795e-01 -4.59738553e-01 -6.34778962e-02 -1.49947321e+00
4.31786448e-01 -1.32985175e+00 6.76249117e-02 -2.51511765e+00
8.20848867e-02 -9.74599570e-02 -1.27062514e-01 4.59249914e-01
-9.32563841e-01 -6.44958466e-02 2.99300224e-01 -3.40691239e-01
-8.58408868e-01 7.94663906e-01 1.79214454e+00 -3.76771748e-01
-1.16658919e-01 -2.79232234e-01 -9.27712500e-01 4.03969020e-01
8.59441400e-01 -2.64285505e-01 -8.32936883e-01 -1.03462410e+00
3.91059518e-01 2.69931465e-01 -7.49661401e-02 -8.08243573e-01
6.99361265e-01 -1.92912772e-01 -5.75153530e-03 -7.03675330e-01
2.13502795e-01 -6.94281220e-01 -2.70074844e-01 3.94437850e-01
-6.13105655e-01 4.61697221e-01 1.05844714e-01 6.24111354e-01
-2.70025611e-01 -4.01549816e-01 2.18843877e-01 -3.99937958e-01
-6.99008167e-01 4.31327939e-01 -3.45543712e-01 3.23810756e-01
1.11834848e+00 5.28096333e-02 -3.71906459e-01 -5.85976243e-01
-4.60541129e-01 8.34798038e-01 -1.27859816e-01 4.49981689e-01
8.62673998e-01 -1.33122265e+00 -9.09543812e-01 1.19820647e-01
9.23491791e-02 3.97430092e-01 3.41518670e-01 4.55338955e-01
-4.74337101e-01 4.60214525e-01 2.81327784e-01 -1.46433145e-01
-1.26689100e+00 6.63421571e-01 4.99280803e-02 -7.75346994e-01
-5.22944391e-01 1.02806437e+00 2.32058421e-01 -2.48322681e-01
2.76648760e-01 -5.29905319e-01 -4.17487383e-01 3.91152594e-03
3.06017876e-01 1.35083914e-01 2.89396513e-02 -3.62951994e-01
-1.36459932e-01 6.34813368e-01 -4.02263522e-01 2.25858718e-01
1.22133005e+00 -1.04356945e-01 -4.59351987e-01 1.18957363e-01
9.33950961e-01 -8.29143375e-02 -6.96926534e-01 -4.26919281e-01
-3.10021564e-02 -2.48953044e-01 1.38738513e-01 -6.73160911e-01
-1.25894248e+00 9.04774547e-01 -3.99304867e-01 3.82353336e-01
1.18531072e+00 -9.46009159e-02 1.14062500e+00 3.08630556e-01
2.21619949e-01 -1.05766571e+00 2.49908745e-01 9.35807049e-01
9.14064229e-01 -7.39212990e-01 -2.30439417e-02 -1.08925831e+00
-6.73869073e-01 1.05052757e+00 1.23073721e+00 -3.35334502e-02
2.38772362e-01 1.53149709e-01 -4.24997091e-01 -1.33178398e-01
-9.95087266e-01 -3.49501699e-01 3.71837437e-01 2.40138873e-01
5.38940191e-01 -1.45402461e-01 -7.85188735e-01 8.21378469e-01
-2.57686794e-01 -7.84107223e-02 5.72529614e-01 9.60789621e-01
-5.15176475e-01 -1.22926104e+00 -4.28565070e-02 5.14320433e-01
-1.32833213e-01 -6.68384612e-01 -6.30532026e-01 4.88954455e-01
7.86972325e-03 1.05998206e+00 1.19406683e-02 -4.55267131e-01
5.44685304e-01 8.93516988e-02 4.97290045e-01 -1.09711325e+00
-7.13821173e-01 -1.21393790e-02 2.55248904e-01 -3.62954378e-01
-9.21064243e-03 -2.47164652e-01 -1.48736286e+00 -1.38691872e-01
-4.21081036e-01 4.56310183e-01 1.91242784e-01 8.52022171e-01
4.38906252e-01 1.36941338e+00 5.02600908e-01 -4.87179935e-01
-3.22888315e-01 -9.83237028e-01 -3.93357575e-01 1.55996501e-01
-1.43875584e-01 -1.66505918e-01 1.18268095e-02 -1.50293902e-01] | [10.286176681518555, 8.246307373046875] |
15fbd765-da37-4189-9f12-b54524b21de6 | is-speech-pathology-a-biomarker-in-automatic | 2204.06450 | null | https://arxiv.org/abs/2204.06450v2 | https://arxiv.org/pdf/2204.06450v2.pdf | The effect of speech pathology on automatic speaker verification -- a large-scale study | With the advancements in deep learning (DL) and an increasing interest in data-driven speech processing methods, there is a major challenge in accessing pathological speech data. Public challenge data offers a potential remedy for this but may expose patient health information by re-identification attacks. Therefore, we investigate in this study whether or not pathological speech is more vulnerable to such re-identification than healthy speech. Our study is the first large-scale investigation on the effects of different speech pathology on automatic speaker verification (ASV) using a real-world pathological speech corpus of more than 2,000 test subjects with various speech and voice disorders from different ages. Utilizing a DL-based ASV method, we obtained a mean equal error rate (EER) of 0.89% with a standard deviation of 0.06%, which is a factor of three lower than comparable healthy speech databases. We further perform detailed analyses of external influencing factors on ASV such as age, pathology, recording environment, utterance length, and intelligibility, to explore their respective effect. Our experiments indicate that some types of speech pathology, in particular dysphonia, regardless of speech intelligibility, are more vulnerable to a breach of privacy compared to healthy speech. We also observe that the effect of pathology lies in the range of other factors, such as age, microphone, and recording environment. | ['Elmar Noeth', 'Seung Hee Yang', 'Andreas Maier', 'Maria Schuster', 'Tobias Weise', 'Soroosh Tayebi Arasteh'] | 2022-04-13 | null | null | null | null | ['text-independent-speaker-verification'] | ['speech'] | [-3.36679704e-02 1.39662221e-01 3.22329849e-01 -2.37646475e-01
-1.12937808e+00 -2.82672286e-01 3.54638338e-01 3.70996624e-01
-4.40329641e-01 4.21739697e-01 6.77180886e-01 -5.31698763e-01
-4.10490856e-02 -2.77185500e-01 -4.89765942e-01 -7.50514984e-01
3.85298356e-02 -1.80369895e-02 -8.52944031e-02 6.06329925e-02
-1.34308999e-02 3.25780749e-01 -1.47754574e+00 2.11027130e-01
7.68405855e-01 7.27590919e-01 1.42636701e-01 5.47395110e-01
-6.39907196e-02 2.27778226e-01 -1.22948921e+00 -5.00945508e-01
-4.87062857e-02 -1.19542047e-01 -7.40495205e-01 1.15467295e-01
1.56857133e-01 -3.95017117e-01 -3.88728827e-01 1.07137012e+00
1.20356619e+00 -2.36270893e-02 2.26877034e-01 -7.30959177e-01
-5.59563816e-01 7.95575082e-01 -9.39984694e-02 3.85323435e-01
4.76198643e-01 3.36344630e-01 5.33474565e-01 -5.16422749e-01
2.61545748e-01 1.19598293e+00 7.12262809e-01 8.28090012e-01
-1.12087703e+00 -6.66944623e-01 -3.36585753e-02 1.34599775e-01
-1.54592264e+00 -1.00290668e+00 6.88965678e-01 -4.42372292e-01
8.82318974e-01 4.13409382e-01 3.99389923e-01 1.45089328e+00
7.15210512e-02 1.47279173e-01 1.02001488e+00 -5.91517508e-01
2.44130120e-01 4.86375868e-01 1.56888768e-01 1.72427222e-01
-3.89381722e-02 1.94713101e-01 -5.51097095e-01 -5.10292351e-01
8.83157030e-02 -3.58001620e-01 -5.37675440e-01 3.86658520e-01
-1.04944456e+00 5.53861678e-01 -1.49477497e-01 6.60245121e-01
-2.64713228e-01 -3.93090874e-01 6.64696097e-01 4.23347712e-01
4.95132834e-01 2.76606768e-01 -6.04378641e-01 -4.30623949e-01
-6.25471830e-01 -1.70219749e-01 7.36698508e-01 2.78993458e-01
-8.73764530e-02 1.54296786e-01 3.14150453e-02 1.30983007e+00
3.36285770e-01 4.13475454e-01 1.00116432e+00 -5.18714428e-01
5.23407757e-01 1.38716057e-01 -2.45234370e-01 -9.08443928e-01
-3.62097710e-01 -4.92439181e-01 -6.02705121e-01 -1.40166031e-02
4.92091417e-01 -1.68396950e-01 -7.61841714e-01 1.94191885e+00
4.70540561e-02 -1.52606368e-01 1.63609341e-01 6.35941863e-01
8.97232354e-01 2.75347114e-01 5.16353510e-02 -5.19385278e-01
1.64920020e+00 -3.41549873e-01 -8.32354307e-01 -1.15036391e-01
5.50822735e-01 -8.35513711e-01 1.40728509e+00 5.22719443e-01
-8.45647216e-01 -3.63947958e-01 -8.90474916e-01 2.40817562e-01
-2.59758234e-01 -5.52881323e-02 4.00073230e-02 1.55304027e+00
-1.15660501e+00 2.85168082e-01 -6.87490523e-01 -5.58063030e-01
6.23541474e-02 3.55409294e-01 -5.83055794e-01 3.53008837e-01
-1.41269886e+00 7.83317149e-01 -8.71888101e-02 -2.54636239e-02
-5.83791852e-01 -5.89526594e-01 -6.53263748e-01 -4.04761061e-02
3.62532139e-02 -2.20224842e-01 1.13311350e+00 -5.18460810e-01
-1.49746537e+00 1.05139697e+00 -3.09209198e-01 -3.77473772e-01
4.67417508e-01 7.79277012e-02 -1.15491354e+00 -1.54868513e-01
-1.76302359e-01 1.38611458e-02 7.83347487e-01 -1.00011420e+00
-1.57419369e-01 -7.64041305e-01 -4.73359734e-01 -1.73225403e-02
-6.06389523e-01 4.41402704e-01 -8.26179087e-02 -6.01757765e-01
1.32674739e-01 -8.03650737e-01 1.64360791e-01 -1.36681139e-01
-6.87047064e-01 -1.94601655e-01 5.03323197e-01 -1.15048134e+00
1.35419512e+00 -2.54629374e+00 -3.47122312e-01 1.33772060e-01
2.48022974e-01 4.95867491e-01 6.17202148e-02 2.63175786e-01
-1.02953590e-01 4.58396047e-01 -1.46532387e-01 -5.59618950e-01
-1.32620605e-02 -4.11075614e-02 -8.20363909e-02 6.84848428e-01
-2.75074661e-01 1.99328601e-01 -3.23166221e-01 -1.76142216e-01
1.28677055e-01 7.17163563e-01 -3.85776639e-01 1.36550963e-01
3.09535086e-01 5.05616605e-01 5.37744835e-02 6.54593170e-01
6.13939345e-01 2.87272930e-01 -3.43957059e-02 -7.71717206e-02
-1.41997427e-01 7.95655847e-01 -7.89101660e-01 1.17008245e+00
-4.89901602e-01 7.79622674e-01 4.61694688e-01 -7.31810510e-01
1.02189660e+00 7.24241018e-01 1.54925868e-01 -6.37533128e-01
3.66940886e-01 4.17852372e-01 6.02335393e-01 -6.24249756e-01
1.96973696e-01 -3.04634243e-01 1.64303109e-01 2.36958653e-01
-2.70044953e-01 2.87383080e-01 -2.67529190e-01 -1.17178224e-01
1.14690840e+00 -9.92720604e-01 1.99792802e-01 -2.02935100e-01
5.89437664e-01 -7.40354419e-01 4.95484531e-01 5.24060726e-01
-7.42451310e-01 5.79609692e-01 4.20365959e-01 1.68320507e-01
-7.66327322e-01 -1.06250286e+00 -6.02244794e-01 6.57508135e-01
-4.67846602e-01 -3.21736395e-01 -8.39764833e-01 -3.37556511e-01
-2.92286389e-02 8.37666154e-01 -2.41660729e-01 -4.35180515e-01
-3.32964748e-01 -8.57137918e-01 9.60467577e-01 3.40200335e-01
2.82371104e-01 -1.00089490e+00 9.99241136e-04 9.72223133e-02
-2.25815207e-01 -1.20880616e+00 -6.23704791e-01 3.79837789e-02
-4.37337846e-01 -8.28928828e-01 -7.94297099e-01 -6.29236639e-01
4.01220143e-01 -1.92530572e-01 6.82676196e-01 4.68039066e-02
-2.35726357e-01 4.43102539e-01 -3.73704702e-01 -4.82817203e-01
-1.26185024e+00 7.16520846e-02 5.76804101e-01 5.43984398e-02
3.63633364e-01 -5.44401050e-01 -4.64822263e-01 3.21628004e-01
-6.06303394e-01 -6.70585334e-01 2.82054156e-01 5.67254424e-01
1.56848148e-01 3.23567867e-01 8.63280773e-01 -4.57976669e-01
1.06749976e+00 -5.07328987e-01 -2.10129425e-01 6.97796047e-02
-6.80033922e-01 -2.59119153e-01 3.60902935e-01 -5.02265334e-01
-9.67022657e-01 -1.46162659e-01 -8.20789814e-01 -3.52445468e-02
-5.54227114e-01 3.75594378e-01 -5.31471252e-01 2.00734526e-01
5.94597101e-01 3.36794734e-01 4.34078217e-01 -7.64684677e-01
-2.07694530e-01 1.51995444e+00 4.10488874e-01 -1.09023191e-01
4.20106441e-01 1.18604250e-01 -7.43627071e-01 -1.29474235e+00
-2.35467598e-01 -4.72863376e-01 -1.70776829e-01 -8.74571353e-02
8.05401921e-01 -6.88480258e-01 -8.20290983e-01 9.26610649e-01
-1.06846440e+00 -7.25693479e-02 2.14880794e-01 6.13294005e-01
-1.02020219e-01 6.15175009e-01 -5.73818862e-01 -1.04220963e+00
-6.15530491e-01 -1.43927419e+00 8.30445349e-01 -1.74863696e-01
-6.12400830e-01 -7.08277762e-01 -1.24350898e-01 8.86507273e-01
5.61260879e-01 -2.30515167e-01 1.15681100e+00 -1.09495211e+00
1.50387153e-01 -2.08774731e-01 1.85925052e-01 7.44448066e-01
6.75338864e-01 -1.58527032e-01 -1.27697611e+00 -3.66940677e-01
5.42507529e-01 -3.08740400e-02 2.63334572e-01 3.76556575e-01
9.96362686e-01 -3.80942583e-01 -9.37431455e-02 2.54782319e-01
7.60246098e-01 6.66726828e-01 6.11728311e-01 2.62133062e-01
4.61805910e-01 7.50917554e-01 -7.55186658e-03 2.97282577e-01
9.95430425e-02 8.97207320e-01 1.01367868e-01 9.19986591e-02
-3.35780591e-01 9.46006365e-03 6.12913489e-01 8.81360829e-01
3.86716187e-01 -3.62091154e-01 -1.09472704e+00 6.10995233e-01
-8.46070707e-01 -7.63784885e-01 -7.18960119e-03 2.45142746e+00
9.06883240e-01 -3.69035006e-02 4.77798074e-01 6.25099838e-01
9.76844847e-01 4.37859222e-02 -6.34026527e-01 -5.20272672e-01
-1.69331983e-01 -1.06521063e-01 3.37794036e-01 4.22303736e-01
-6.49743557e-01 3.54127258e-01 6.60524416e+00 4.35454190e-01
-1.57879055e+00 1.20734163e-01 7.94321001e-01 -2.09245592e-01
-2.16352418e-01 -7.06524074e-01 -6.46497011e-01 8.04817736e-01
1.39786077e+00 -1.54476166e-01 4.24796373e-01 6.46156847e-01
5.70528269e-01 2.12042764e-01 -7.68465042e-01 9.29496169e-01
2.21594766e-01 -7.54323781e-01 -3.34085137e-01 3.42297405e-01
-1.11656383e-01 1.82748631e-01 2.88138688e-01 1.66045912e-02
-2.74814099e-01 -8.45559359e-01 7.87861288e-01 -7.81214982e-02
8.40353370e-01 -5.81494153e-01 7.74677336e-01 2.08157331e-01
-5.75074673e-01 -2.39931241e-01 6.07778095e-02 2.19523981e-01
2.88895726e-01 7.05293238e-01 -1.11274374e+00 6.54038601e-03
8.06040108e-01 -3.20347510e-02 -2.29884386e-01 8.26217055e-01
3.97110358e-02 9.29258883e-01 -2.37635329e-01 8.62499103e-02
-3.55274200e-01 4.29265738e-01 8.50645542e-01 1.01039422e+00
5.20885050e-01 -8.35372508e-02 -4.80781913e-01 5.37613750e-01
-3.04327123e-02 3.28520387e-01 -3.95323068e-01 -2.97970682e-01
8.90140116e-01 6.10035181e-01 -3.22706014e-01 1.33380964e-01
-4.99024391e-01 8.13751161e-01 -5.98793328e-02 1.48146033e-01
-4.81094539e-01 -2.16661870e-01 1.13108361e+00 4.09052312e-01
-6.34925440e-02 -3.14765833e-02 -3.52754354e-01 -8.01553786e-01
3.59801054e-01 -1.38925576e+00 1.60728157e-01 -3.18572223e-01
-1.22201467e+00 8.32572401e-01 -4.01525676e-01 -9.74142671e-01
-1.84244007e-01 -4.38735574e-01 -5.29859066e-01 1.04060185e+00
-9.13589776e-01 -5.62670350e-01 1.29998460e-01 5.07592618e-01
5.28820693e-01 -5.45801282e-01 9.73591983e-01 6.08392239e-01
-8.84090781e-01 1.01736104e+00 1.16809309e-01 1.69849664e-01
6.43471539e-01 -7.77115107e-01 4.86149043e-01 6.94751084e-01
-4.18251380e-02 8.19757998e-01 6.72413647e-01 -5.18903792e-01
-1.07518399e+00 -7.76956856e-01 1.19419992e+00 -3.97320628e-01
6.14176869e-01 -6.17678106e-01 -1.13281810e+00 2.92509496e-01
-2.85065570e-03 -4.11425442e-01 1.00639403e+00 3.41845095e-01
-3.98706019e-01 -1.45099536e-01 -1.49912143e+00 5.27968705e-01
9.69267130e-01 -8.77023220e-01 -4.26417768e-01 1.49014726e-01
1.02177429e+00 -1.12378947e-01 -1.02371943e+00 1.94550738e-01
4.13965344e-01 -1.08639586e+00 8.73373747e-01 -1.49519667e-01
-1.52920693e-01 8.68684649e-02 -1.98214993e-01 -1.38032985e+00
-7.38650784e-02 -7.06774473e-01 3.52978826e-01 1.69617546e+00
6.56597793e-01 -1.04930186e+00 6.60524607e-01 9.71800387e-01
-2.66257405e-01 -5.07100821e-01 -1.39566791e+00 -9.10382688e-01
1.50127068e-01 -7.37612724e-01 7.36442447e-01 8.62713277e-01
2.63729513e-01 -7.07384571e-02 -2.67437756e-01 4.44275618e-01
1.86313808e-01 -7.33109176e-01 3.65198880e-01 -9.93130982e-01
-2.36062348e-01 -3.85416389e-01 -5.73150635e-01 -2.93690264e-01
2.66371101e-01 -7.19601274e-01 1.29996017e-02 -1.04661918e+00
-3.67002748e-02 -2.94448107e-01 -2.79894322e-01 2.51117080e-01
-6.40426800e-02 -3.56270045e-01 7.04186410e-03 8.41966271e-02
3.43011349e-01 4.38127518e-01 7.82273650e-01 -1.81669056e-01
-4.44902390e-01 3.90587270e-01 -8.83834779e-01 6.16004229e-01
1.03225613e+00 -4.19711024e-01 -5.30216455e-01 -3.92346382e-01
-5.77475727e-01 1.20012723e-01 3.74943689e-02 -8.97090375e-01
4.72389199e-02 3.39172810e-01 -1.82349831e-01 -1.95144996e-01
3.99823427e-01 -6.26681030e-01 -3.72413360e-03 6.05279982e-01
-4.53689307e-01 -1.53596932e-02 4.56858158e-01 3.31010640e-01
-1.44981354e-01 -9.93805751e-02 7.06460714e-01 1.28054887e-01
-9.67958048e-02 -1.36252865e-01 -9.10473406e-01 -2.02423602e-01
5.44615388e-01 -1.83884129e-01 -3.11585754e-01 -6.14732206e-01
-8.91911566e-01 -3.77729893e-01 2.02863514e-01 6.65579379e-01
3.74519885e-01 -8.56916130e-01 -6.31739020e-01 4.83027577e-01
2.77573578e-02 -5.17798007e-01 3.80999357e-01 9.50034142e-01
-1.06020205e-01 5.05806029e-01 2.86866844e-01 -3.38707954e-01
-1.85687363e+00 3.26678276e-01 5.17822325e-01 2.59979695e-01
-5.95942199e-01 9.38688278e-01 -6.24400042e-02 -4.94369239e-01
7.45152950e-01 -4.24506456e-01 1.05405465e-01 4.82797530e-03
6.36859477e-01 5.40037274e-01 6.05569482e-01 -8.55415523e-01
-6.09597683e-01 1.17468864e-01 -2.47200772e-01 -1.90813035e-01
8.07793140e-01 -3.96512657e-01 7.33619705e-02 3.24689478e-01
1.31527245e+00 3.46499503e-01 -5.01090288e-01 -4.32005860e-02
-4.62592579e-02 -3.69530410e-01 1.45553276e-01 -9.66991484e-01
-1.01949990e+00 6.62942350e-01 1.05367410e+00 5.89320421e-01
1.05611145e+00 3.91163647e-01 9.80782509e-01 -3.53115350e-02
2.72747278e-01 -8.40161026e-01 -2.75648892e-01 1.72202989e-01
1.01657879e+00 -1.11828029e+00 -5.95233381e-01 -4.22944695e-01
-5.31305790e-01 5.22665322e-01 3.62525851e-01 7.12748349e-01
8.74741793e-01 3.75129730e-01 6.39155984e-01 3.55615988e-02
-4.30677950e-01 1.05336249e-01 5.02604544e-02 8.44619572e-01
7.01656222e-01 2.25597069e-01 -5.06901920e-01 9.01341498e-01
-8.35111380e-01 -3.12928379e-01 3.25579256e-01 4.36435342e-01
-2.10752368e-01 -1.23643923e+00 -7.23475635e-01 5.38003445e-01
-7.66972482e-01 -1.84268326e-01 -5.25308669e-01 3.91038686e-01
3.69430636e-03 1.47351635e+00 2.36771077e-01 -6.28256857e-01
5.44628561e-01 3.25811595e-01 -2.97873039e-02 -5.82506657e-01
-6.36104584e-01 -2.89124250e-02 5.18007398e-01 -2.82483160e-01
-2.68720202e-02 -1.07652712e+00 -1.15118599e+00 -4.25097466e-01
-3.96290988e-01 6.27191886e-02 9.47854519e-01 9.92409170e-01
4.81979996e-01 5.20987570e-01 4.16006893e-01 -1.79138452e-01
-7.22613215e-01 -1.12598491e+00 -7.63502657e-01 3.94881189e-01
6.52344823e-01 -3.51643384e-01 -9.24893081e-01 -2.56670326e-01] | [14.280558586120605, 6.0951457023620605] |
9ac93ce0-2f32-49e9-9c8c-5c698e08a791 | clara-classifying-and-disambiguating-user | 2306.10376 | null | https://arxiv.org/abs/2306.10376v3 | https://arxiv.org/pdf/2306.10376v3.pdf | CLARA: Classifying and Disambiguating User Commands for Reliable Interactive Robotic Agents | In this paper, we focus on inferring whether the given user command is clear, ambiguous, or infeasible in the context of interactive robotic agents utilizing large language models (LLMs). To tackle this problem, we first present an uncertainty estimation method for LLMs to classify whether the command is certain (i.e., clear) or not (i.e., ambiguous or infeasible). Once the command is classified as uncertain, we further distinguish it between ambiguous or infeasible commands leveraging LLMs with situational aware context in a zero-shot manner. For ambiguous commands, we disambiguate the command by interacting with users via question generation with LLMs. We believe that proper recognition of the given commands could lead to a decrease in malfunction and undesired actions of the robot, enhancing the reliability of interactive robot agents. We present a dataset for robotic situational awareness, consisting pair of high-level commands, scene descriptions, and labels of command type (i.e., clear, ambiguous, or infeasible). We validate the proposed method on the collected dataset, pick-and-place tabletop simulation. Finally, we demonstrate the proposed approach in real-world human-robot interaction experiments, i.e., handover scenarios. | ['Minsuk Chang', 'Sungjoon Choi', 'Youngjae Yu', 'Sangbeom Park', 'Joonhyung Lee', 'Seungwon Lim', 'Jeongeun Park'] | 2023-06-17 | null | null | null | null | ['question-generation'] | ['natural-language-processing'] | [ 4.49425131e-01 5.08248329e-01 2.16338158e-01 -6.95434809e-01
-5.28070033e-01 -8.69703054e-01 5.53816199e-01 1.05194576e-01
-1.01128772e-01 8.77637744e-01 -3.26959267e-02 -5.41414797e-01
-3.59767407e-01 -5.05067110e-01 -8.43885303e-01 -4.89925921e-01
1.06527163e-02 6.42322183e-01 1.77031800e-01 -1.33515596e-01
4.91869599e-01 4.65788633e-01 -1.72510207e+00 1.36593536e-01
9.48715508e-01 8.52439761e-01 7.62251019e-01 5.93809068e-01
-2.14333888e-02 9.45542157e-01 -8.48431826e-01 2.87430227e-01
-5.58923045e-03 -1.07873037e-01 -1.15443146e+00 2.72796810e-01
-1.64610729e-01 -5.68818867e-01 1.08684346e-01 1.00184619e+00
-9.50876549e-02 2.65821636e-01 6.76848233e-01 -1.84817505e+00
6.95673600e-02 5.63504457e-01 1.18867241e-01 -4.44523603e-01
8.71034741e-01 2.54465669e-01 5.76717496e-01 -6.35434806e-01
5.98432422e-01 1.53388202e+00 -1.85798761e-02 2.74416953e-01
-9.42820013e-01 -3.57856393e-01 6.34320676e-01 2.39340961e-01
-1.49067926e+00 -4.51602995e-01 4.49275494e-01 -6.75997436e-01
6.59377515e-01 2.14388877e-01 3.86913083e-02 1.11128223e+00
2.76977152e-01 7.15328693e-01 9.36809838e-01 -2.47733384e-01
4.83334482e-01 2.65905917e-01 2.57803500e-01 5.73720455e-01
1.99272335e-01 -3.04178428e-02 -1.76541060e-01 -1.62590519e-01
6.87391758e-01 -1.19996250e-01 -2.48184144e-01 -6.05663896e-01
-1.53765786e+00 4.11349535e-01 -2.27504019e-02 -9.11488086e-02
-6.61655188e-01 -2.66498476e-01 2.01380953e-01 1.08144708e-01
-3.72104198e-01 3.25512797e-01 -3.65604311e-01 -3.25441033e-01
1.72091141e-01 1.77500010e-01 9.78513420e-01 1.65998507e+00
7.20593810e-01 -2.86956489e-01 -1.82912126e-01 6.72296166e-01
3.96939725e-01 3.89675498e-01 1.98832333e-01 -1.16723216e+00
5.17213881e-01 5.61675787e-01 9.21009421e-01 -8.75985503e-01
-6.27592683e-01 3.09166044e-01 -2.76110917e-01 3.59882832e-01
1.06890813e-01 -3.95042807e-01 -7.39006102e-01 1.52275980e+00
4.78813261e-01 1.49214044e-02 6.80471301e-01 1.04213059e+00
5.50853908e-01 7.56325543e-01 -4.57507670e-02 -2.23667920e-01
1.47597325e+00 -6.26714885e-01 -1.07722604e+00 -4.36211586e-01
4.99755859e-01 -5.11080921e-01 1.04157650e+00 7.29250014e-01
-3.47233891e-01 -4.80062246e-01 -1.03734839e+00 3.60409558e-01
-2.15661824e-01 3.65515858e-01 5.01462102e-01 6.11376017e-02
-6.14042521e-01 1.40891880e-01 -7.18867660e-01 -6.76406622e-01
-5.47372699e-01 3.37101877e-01 -3.75065625e-01 -2.35557988e-01
-9.90609407e-01 7.39579141e-01 8.12690318e-01 2.65392184e-01
-1.20244384e+00 2.56311208e-01 -1.19732201e+00 -2.68538326e-01
1.01816714e+00 -3.49515900e-02 1.54186583e+00 -3.18938255e-01
-1.68680656e+00 1.58104703e-01 -1.35794252e-01 -1.51596248e-01
2.54821420e-01 -2.05648884e-01 -3.75498563e-01 -4.02567014e-02
2.72440881e-01 5.61988592e-01 7.04546690e-01 -1.79226577e+00
-1.19675601e+00 -4.67267901e-01 6.19740307e-01 5.47359765e-01
3.70742381e-01 -1.26982287e-01 -2.83871025e-01 7.85018206e-02
4.93644655e-01 -1.32254624e+00 -2.46769923e-04 -3.48812073e-01
-7.46085882e-01 -3.50158393e-01 8.50541651e-01 -4.17082608e-01
8.30806732e-01 -2.24439740e+00 5.26855178e-02 1.94801152e-01
-2.07362011e-01 -2.18899265e-01 -9.00634304e-02 4.98983860e-01
2.53392816e-01 -1.52390793e-01 -3.05792242e-01 2.71932911e-02
8.39200988e-02 6.64368510e-01 -4.86291736e-01 1.79169923e-01
3.01765054e-01 5.34617119e-02 -1.06025684e+00 -4.62534070e-01
3.46398652e-01 -3.99416238e-02 -2.21735612e-01 7.58204699e-01
-2.72095174e-01 6.73966229e-01 -7.27772415e-01 7.61020541e-01
5.89149833e-01 3.41187149e-01 5.03018618e-01 -1.60962239e-01
-3.71056795e-01 2.60625124e-01 -1.57796097e+00 1.45374799e+00
-6.65978253e-01 4.02216643e-01 4.16235328e-01 -5.96926928e-01
8.73320401e-01 2.82107800e-01 -5.81630878e-02 -8.63051042e-02
1.22476466e-01 1.35391444e-01 5.74066956e-03 -8.62067163e-01
7.43983865e-01 4.06372637e-01 -5.32020509e-01 2.48287708e-01
-1.84126839e-01 -2.84337431e-01 1.55766970e-02 1.62724957e-01
9.61761951e-01 4.98696089e-01 4.04498488e-01 2.54947562e-02
4.27066743e-01 1.92796573e-01 4.49933052e-01 8.68721426e-01
-3.60837340e-01 2.95406461e-01 5.41164637e-01 -2.63334997e-02
-2.70197362e-01 -9.05150890e-01 2.39789695e-01 1.11575580e+00
7.74305582e-01 -1.69694945e-01 -6.72671795e-01 -6.00263357e-01
-1.35195032e-01 1.18151402e+00 -1.69817209e-01 -5.17251380e-02
-3.69331390e-01 -6.72947466e-02 3.19973499e-01 8.62380788e-02
3.57193470e-01 -1.16807294e+00 -1.09484172e+00 1.74159810e-01
-5.57391047e-01 -1.43054497e+00 -1.60852745e-01 3.58063400e-01
-3.60195935e-01 -1.18453503e+00 6.52112290e-02 -8.87921214e-01
9.80818808e-01 3.48722816e-01 6.09203160e-01 -1.48837686e-01
1.80140644e-01 5.43791354e-01 -6.19295895e-01 -3.45145971e-01
-6.00826144e-01 -3.61787766e-01 4.90812689e-01 6.05892092e-02
-2.37707142e-02 -7.62225091e-02 -2.08325505e-01 7.12664247e-01
-7.80685484e-01 7.18277916e-02 6.62408888e-01 5.65616131e-01
3.08115512e-01 2.86491185e-01 4.98771966e-01 -5.16904950e-01
1.09296846e+00 -5.56816936e-01 -5.66892147e-01 3.60965580e-01
-3.17743719e-01 2.86240935e-01 4.35633123e-01 -5.78094482e-01
-1.24207282e+00 4.68292356e-01 2.82912284e-01 -6.47003725e-02
-9.43060338e-01 6.52673125e-01 -3.65227014e-01 2.47903198e-01
3.88899535e-01 1.36518374e-01 -1.73488051e-01 -1.22918434e-01
2.81088293e-01 1.29867876e+00 7.72119462e-01 -8.92964184e-01
3.46533179e-01 8.22437778e-02 -2.72920787e-01 -6.21360421e-01
-1.16666846e-01 -3.58642399e-01 -4.98091221e-01 -4.41254079e-01
7.27668524e-01 -7.10625470e-01 -1.11248493e+00 3.31932575e-01
-1.36164260e+00 -2.59212434e-01 4.24442708e-01 5.04635215e-01
-7.35052705e-01 4.08301055e-01 -3.11950654e-01 -1.38955235e+00
2.86387682e-01 -1.57926130e+00 1.10423899e+00 1.66942626e-01
-4.12981212e-01 -2.05025420e-01 -5.55829883e-01 1.26388490e-01
-1.07011467e-01 1.60231203e-01 1.00448668e+00 -8.99167240e-01
-5.73553205e-01 -1.96106598e-01 -1.38764769e-01 -1.85548902e-01
4.95542854e-01 -1.44836962e-01 -5.56834042e-01 1.84415877e-02
-4.79250960e-02 -3.46729070e-01 -2.01099381e-01 -1.49657249e-01
6.59171879e-01 -5.73850632e-01 -5.49559295e-01 -1.89590752e-01
9.09814477e-01 9.31248844e-01 5.71554065e-01 1.30776614e-01
4.10078496e-01 8.82629216e-01 1.65660250e+00 6.50942862e-01
7.45064139e-01 6.27334595e-01 7.20891297e-01 5.67804337e-01
5.53505242e-01 -1.74337640e-01 5.51255107e-01 1.87682465e-01
3.10433626e-01 -4.82888460e-01 -8.85475576e-01 3.41910601e-01
-2.14267921e+00 -3.99324477e-01 5.72302938e-02 2.18113971e+00
6.84771776e-01 2.17152447e-01 -2.93963403e-01 -1.24398500e-01
9.13619339e-01 -2.48496160e-01 -7.14099944e-01 -5.01012206e-01
3.94580156e-01 -7.43146420e-01 1.19641863e-01 7.07909942e-01
-8.66221786e-01 1.05986857e+00 4.70999479e+00 3.04858238e-01
-8.37582111e-01 -4.29423064e-01 1.39125496e-01 3.05495799e-01
9.33070630e-02 3.45161073e-02 -6.41243935e-01 3.35712045e-01
5.45953929e-01 2.32329324e-01 8.11199486e-01 9.94377911e-01
2.86153644e-01 -7.20020473e-01 -1.65439200e+00 8.98403823e-01
5.35519011e-02 -5.20906866e-01 -9.83328000e-02 -3.53378415e-01
6.21785857e-02 -3.36077392e-01 -2.73760319e-01 7.08056629e-01
4.88781631e-01 -7.01831520e-01 1.04377770e+00 4.79337215e-01
5.58336020e-01 -5.09783268e-01 8.99228871e-01 9.68375683e-01
-9.53295290e-01 -3.02029788e-01 1.46076493e-02 -1.66944355e-01
2.38681406e-01 -1.19635947e-01 -1.34136069e+00 7.92298496e-01
4.72822070e-01 1.31574467e-01 -1.16305344e-01 6.32275701e-01
-3.37289095e-01 4.31476571e-02 -2.72049814e-01 -1.70723051e-01
-5.98538406e-02 -3.43657047e-01 6.88100398e-01 8.33851874e-01
4.25414592e-01 6.83661640e-01 5.85236311e-01 8.63759518e-01
6.24516368e-01 -3.48309606e-01 -7.52455235e-01 -1.32260304e-02
9.75709140e-01 9.51961517e-01 -4.60475117e-01 -3.35308671e-01
1.43658007e-02 1.14634573e+00 -1.04240160e-02 6.03710413e-01
-6.90500498e-01 -8.79239142e-01 6.60617888e-01 -4.73247230e-01
-4.24802527e-02 -4.18190032e-01 1.26777850e-02 -7.44234443e-01
2.02093437e-01 -7.58705974e-01 9.68296230e-02 -1.29114866e+00
-7.36549854e-01 8.12609673e-01 3.64030391e-01 -1.49919426e+00
-8.02721798e-01 -5.96028388e-01 -4.02635664e-01 6.84372783e-01
-1.22990489e+00 -6.39518499e-01 -6.54755116e-01 2.46356234e-01
7.66734660e-01 2.77125597e-01 7.31931746e-01 -2.44210318e-01
-4.98007625e-01 5.76570667e-02 -4.58231956e-01 -1.37986943e-01
8.16844106e-01 -8.11792195e-01 -1.10029243e-01 4.73732710e-01
-5.78207374e-01 7.69813418e-01 1.05144048e+00 -9.19915915e-01
-1.62909961e+00 -9.20639217e-01 4.64342952e-01 -4.14484769e-01
3.21967810e-01 -4.19284135e-01 -7.78058350e-01 9.30688083e-01
-1.08554415e-01 -3.41565788e-01 1.24784783e-01 -1.93909466e-01
1.31952122e-01 2.50315994e-01 -1.61412311e+00 8.76134217e-01
9.48001981e-01 -2.75041640e-01 -7.78288364e-01 1.71671554e-01
1.14097512e+00 -6.39720261e-01 -6.01573944e-01 6.38016045e-01
4.74903971e-01 -6.69200718e-01 6.05040014e-01 -3.20273101e-01
-1.85082797e-02 -7.78771579e-01 -4.08724755e-01 -1.41535699e+00
-1.34068457e-02 -4.41390067e-01 1.47777870e-01 1.17083609e+00
3.86440545e-01 -6.38232946e-01 2.52691656e-01 1.09403002e+00
-5.50818145e-01 -4.00713325e-01 -6.55207157e-01 -6.77231014e-01
-7.38939524e-01 -4.79313612e-01 8.24173927e-01 7.45556951e-01
6.28608644e-01 2.62098104e-01 -2.33523518e-01 9.19942081e-01
9.29588154e-02 1.41273722e-01 8.73935223e-01 -1.09251595e+00
3.22028726e-01 8.02041218e-02 -2.53120482e-01 -1.21717799e+00
3.33099306e-01 -2.86079854e-01 1.13836300e+00 -1.78690076e+00
-1.78676695e-01 -5.72481334e-01 1.01372667e-01 4.68003839e-01
5.83760217e-02 -6.37802064e-01 1.63822100e-01 1.93954840e-01
-7.25097775e-01 4.75432605e-01 7.73360550e-01 -1.34770691e-01
-5.32789826e-01 9.63457748e-02 -2.95216918e-01 7.20786452e-01
9.22459304e-01 -1.00717969e-01 -5.47613084e-01 -2.94405460e-01
-2.46276762e-02 6.38881326e-01 1.52140856e-01 -9.39139307e-01
2.73855299e-01 -7.86063135e-01 -2.56271541e-01 -5.80651343e-01
5.97002447e-01 -1.35383725e+00 -7.89644718e-02 3.18439901e-01
-4.72877949e-01 -5.25004715e-02 1.05019234e-01 6.43220663e-01
-7.56270140e-02 -4.53931749e-01 4.40701842e-02 -2.80467663e-02
-1.17219341e+00 -3.17309588e-01 -9.65774775e-01 -5.52625656e-01
1.19868827e+00 -1.11905530e-01 -3.11665416e-01 -6.17786109e-01
-6.65458560e-01 7.59386778e-01 2.92899221e-01 7.05580115e-01
8.66035819e-01 -9.57062721e-01 -3.63445669e-01 2.33731255e-01
6.10250473e-01 3.05993855e-01 -4.93819155e-02 4.78920937e-01
-1.95348769e-01 5.49076021e-01 -1.36147574e-01 -6.83421791e-01
-1.18080723e+00 5.12310088e-01 -8.49999115e-02 4.26695108e-01
-1.64751291e-01 2.36254096e-01 2.73762405e-01 -9.41346586e-01
7.72926807e-01 -8.82198930e-01 -6.83023632e-01 -3.00299704e-01
2.19009310e-01 1.61751226e-01 -2.07103342e-01 -6.54933214e-01
-5.73620081e-01 2.83518195e-01 2.58795589e-01 -2.68872947e-01
5.73697031e-01 -4.89225537e-01 -7.28380308e-02 7.52636254e-01
5.37838101e-01 -1.36062875e-01 -1.29556763e+00 -4.82191890e-02
2.16260970e-01 -2.86031902e-01 -4.98101443e-01 -1.11832213e+00
-1.65339708e-01 3.20407987e-01 5.73756814e-01 2.65963972e-01
7.18420684e-01 1.57793477e-01 2.76536494e-01 1.11998117e+00
1.20156908e+00 -1.05390263e+00 -1.08732298e-01 8.00952196e-01
1.17984140e+00 -1.47717679e+00 -2.13367686e-01 -6.61726594e-01
-1.11212611e+00 9.52350020e-01 1.03543568e+00 5.56608796e-01
2.72699185e-02 1.55834422e-01 1.02996767e-01 -3.11033297e-02
-6.23533905e-01 -5.16286716e-02 -2.56020010e-01 7.78787136e-01
-1.25184610e-01 4.49777871e-01 9.97315198e-02 7.77716398e-01
-1.31941125e-01 -6.27419502e-02 1.06796968e+00 1.41798878e+00
-7.60985196e-01 -4.31526572e-01 -6.84540153e-01 1.81639284e-01
3.70858848e-01 3.31429720e-01 -5.31642497e-01 5.89921355e-01
-8.38667154e-02 1.69346416e+00 1.14071272e-01 -7.82664359e-01
6.38332725e-01 1.26776248e-01 1.89221367e-01 -8.42618525e-01
4.40798774e-02 -2.01336965e-01 5.46289384e-01 -4.86884266e-01
3.29972133e-02 -4.94288087e-01 -1.91455448e+00 3.20037723e-01
-3.41887802e-01 1.85906991e-01 9.34604168e-01 1.01837659e+00
4.37418163e-01 3.85654420e-01 6.78979874e-01 -9.86209691e-01
-6.25438333e-01 -1.03172600e+00 -2.49095932e-01 2.29991496e-01
5.74936092e-01 -8.92412007e-01 -3.21957439e-01 1.23460786e-02] | [4.480828762054443, 0.8384336233139038] |
24dcac8a-4c6b-4594-b17c-4c80a10f9139 | algebraic-and-geometric-models-for-space | 2304.01150 | null | https://arxiv.org/abs/2304.01150v1 | https://arxiv.org/pdf/2304.01150v1.pdf | Algebraic and Geometric Models for Space Networking | In this paper we introduce some new algebraic and geometric perspectives on networked space communications. Our main contribution is a novel definition of a time-varying graph (TVG), defined in terms of a matrix with values in subsets of the real line P(R). We leverage semi-ring properties of P(R) to model multi-hop communication in a TVG using matrix multiplication and a truncated Kleene star. This leads to novel statistics on the communication capacity of TVGs called lifetime curves, which we generate for large samples of randomly chosen STARLINK satellites, whose connectivity is modeled over day-long simulations. Determining when a large subsample of STARLINK is temporally strongly connected is further analyzed using novel metrics introduced here that are inspired by topological data analysis (TDA). To better model networking scenarios between the Earth and Mars, we introduce various semi-rings capable of modeling propagation delay as well as protocols common to Delay Tolerant Networking (DTN), such as store-and-forward. Finally, we illustrate the applicability of zigzag persistence for featurizing different space networks and demonstrate the efficacy of K-Nearest Neighbors (KNN) classification for distinguishing Earth-Mars and Earth-Moon satellite systems using time-varying topology alone. | ['Robert Kassouf-Short', 'Tung Lam', 'Alan Hylton', 'Brian Heller', 'Robert Green', 'Justin Curry', 'Jacob Cleveland', 'Robert Cardona', 'William Bernardoni'] | 2023-04-03 | null | null | null | null | ['topological-data-analysis'] | ['graphs'] | [-1.47001401e-01 4.12246525e-01 -2.37939820e-01 6.10306039e-02
1.29464313e-01 -1.11214972e+00 1.09512186e+00 1.54777199e-01
-1.50004774e-02 9.29404438e-01 -8.17984119e-02 -7.20215440e-01
-1.06061828e+00 -1.15241671e+00 -5.83042681e-01 -8.44435871e-01
-1.56078911e+00 6.41972065e-01 5.01982868e-01 -8.82760882e-01
3.58861615e-03 1.08638430e+00 -1.23733187e+00 -5.52560091e-01
5.43571234e-01 8.11890721e-01 -1.92102507e-01 7.74437904e-01
2.90219933e-01 2.71765172e-01 -4.54786062e-01 1.74479097e-01
3.25933784e-01 -1.08496267e-02 -9.00502384e-01 -5.80730364e-02
-6.79265559e-02 3.54373127e-01 -8.63630116e-01 5.56626260e-01
1.53410152e-01 1.14752606e-01 5.40336072e-01 -1.65563285e+00
1.47550583e-01 9.81624067e-01 -2.31265295e-02 4.30509895e-01
3.23032320e-01 -1.48106545e-01 8.33998859e-01 -8.60653073e-02
9.96908605e-01 1.17100406e+00 8.82330656e-01 -2.07508162e-01
-1.20248985e+00 -2.34748840e-01 -2.46036530e-01 1.75712809e-01
-1.42837226e+00 -5.45862973e-01 4.91755366e-01 -1.43770427e-01
5.37978232e-01 9.42214489e-01 1.09915113e+00 7.02000082e-01
4.63389307e-01 -1.52466118e-01 8.62923741e-01 -1.50684416e-01
3.72081816e-01 -4.59654778e-01 1.44031018e-01 3.94487828e-01
6.38614535e-01 2.56833583e-01 -2.00265884e-01 -4.66363162e-01
6.90451145e-01 -4.55645919e-01 -5.26147664e-01 -5.09088039e-01
-1.67002022e+00 6.29751921e-01 7.99982250e-01 5.86893260e-01
-3.46038729e-01 3.57378900e-01 1.23967342e-01 8.80587399e-01
1.86826259e-01 2.91978776e-01 -1.21314712e-01 2.26015210e-01
-7.54134059e-01 -1.18398637e-01 1.21757293e+00 1.15762520e+00
5.63731015e-01 -1.90431729e-01 3.13286602e-01 1.71655819e-01
3.66265208e-01 6.07678473e-01 -1.96065232e-01 -9.90520895e-01
2.50390023e-02 2.41820797e-01 1.57680929e-01 -1.17775118e+00
-1.11164880e+00 -1.04782581e+00 -9.75582659e-01 -1.71565726e-01
2.14959309e-01 1.43895298e-01 -2.12149620e-01 1.48904812e+00
5.83078206e-01 1.96595773e-01 3.59422624e-01 7.31632173e-01
4.45415497e-01 5.84305286e-01 -6.16514683e-01 -4.89822716e-01
9.54522610e-01 -4.20347422e-01 -1.34224638e-01 1.54708728e-01
1.11270642e+00 -1.08122781e-01 4.46898229e-02 1.68307766e-01
-1.03616357e+00 2.95141846e-01 -1.19912410e+00 4.09482509e-01
-2.91162401e-01 -6.75244689e-01 5.42858779e-01 5.47500908e-01
-1.59675646e+00 8.55969608e-01 -8.19614589e-01 -1.08449709e+00
-1.86298758e-01 2.84119815e-01 -2.25285262e-01 9.56492499e-02
-1.42553890e+00 6.53034329e-01 2.55436182e-01 -7.26618841e-02
-9.65913177e-01 -3.73276502e-01 -5.09016573e-01 -3.67444366e-01
1.18642129e-01 -9.21537638e-01 6.48699164e-01 -3.27158064e-01
-1.04045033e+00 5.37528872e-01 3.09434950e-01 -1.14384425e+00
5.06834209e-01 7.79464841e-01 -1.03505278e+00 4.81575459e-01
-1.56124994e-01 6.98010847e-02 3.95095408e-01 -1.34102225e+00
-2.43303373e-01 -3.07563126e-01 3.09993833e-01 -1.23216145e-01
-4.56450023e-02 -1.78814396e-01 -1.91863254e-02 -3.24022949e-01
7.28350759e-01 -1.26367962e+00 -3.23014170e-01 1.97938859e-01
-5.84317863e-01 -3.80458347e-02 9.00108695e-01 -8.84127542e-02
1.11246395e+00 -2.03337717e+00 2.90968388e-01 9.47509825e-01
5.11419654e-01 -2.71317989e-01 -1.69174880e-01 1.22373521e+00
1.16324257e-02 1.16276786e-01 -2.90250838e-01 4.20941077e-02
-3.02752793e-01 6.10215008e-01 -2.08904877e-01 1.31582594e+00
-8.15744698e-01 2.57137448e-01 -1.15650702e+00 -2.17909381e-01
-1.10809840e-01 2.11646184e-01 -3.05297207e-02 -6.83475375e-01
-2.91987836e-01 5.70035636e-01 -5.54039776e-01 6.06296599e-01
7.68621981e-01 -3.64976078e-01 2.63498813e-01 5.69234714e-02
-5.71846664e-01 3.62446815e-01 -9.21999991e-01 1.54963851e+00
-2.85401106e-01 6.88787043e-01 5.26916087e-01 -1.00576830e+00
9.95216966e-01 1.02705806e-01 6.28662348e-01 -3.50391299e-01
2.04540074e-01 2.51136690e-01 -9.42303147e-03 -9.40955952e-02
6.25564575e-01 2.31704682e-01 -7.93436263e-03 4.36826706e-01
-2.94417500e-01 -8.47634375e-02 2.64732242e-01 8.22419882e-01
1.69732583e+00 -8.95675063e-01 -1.53087363e-01 -9.60872233e-01
3.28449011e-01 4.01538499e-02 2.19742671e-01 6.82626963e-01
-4.27696807e-03 -4.24018353e-02 5.44567525e-01 -1.05083808e-01
-9.83305514e-01 -1.13306069e+00 -3.97328585e-01 3.75532299e-01
5.64070761e-01 -7.50538349e-01 -1.01107620e-02 -1.66702300e-01
3.58641744e-01 6.65831685e-01 -5.09417713e-01 -2.21871212e-01
-5.88395856e-02 -5.85028827e-01 6.62838519e-01 -5.32082736e-01
2.00481474e-01 -5.82519174e-02 -6.09987937e-02 2.47281760e-01
-1.03565924e-01 -1.24361396e+00 -1.37080802e-02 -1.56083748e-01
-1.20330441e+00 -1.29503667e+00 -1.77575946e-01 -4.42465276e-01
4.96749252e-01 7.09296703e-01 1.07738924e+00 2.64047056e-01
7.03471974e-02 9.71163929e-01 -5.75629771e-01 3.33547384e-01
-4.69603807e-01 -5.47915027e-02 7.26857305e-01 1.61935482e-02
-6.85873449e-01 -9.90447164e-01 -8.83091331e-01 8.82861078e-01
-7.82532275e-01 -3.70768338e-01 1.61843345e-01 -5.86236455e-02
3.23326170e-01 3.71652037e-01 2.50174761e-01 -1.81164518e-01
9.51217189e-02 -1.31616604e+00 -7.19851434e-01 1.39731005e-01
-5.66431224e-01 -1.70498509e-02 2.85711050e-01 -1.64399967e-01
2.87014432e-03 -4.26028967e-01 6.14617050e-01 -9.02456976e-03
7.72105217e-01 7.47327387e-01 3.11156631e-01 -1.04220605e+00
6.68536127e-01 1.60092041e-01 1.80672541e-01 -9.52688530e-02
6.38380706e-01 5.83380580e-01 3.49224329e-01 -4.80223447e-01
1.08069134e+00 1.28840923e+00 9.08071458e-01 -1.56736314e+00
6.00751229e-02 -2.25585163e-01 -3.84881109e-01 -5.57415187e-01
-4.49051196e-03 -8.94334078e-01 -9.60358739e-01 1.61066279e-01
-9.68443394e-01 -2.83588171e-01 -4.81142253e-01 7.57267296e-01
-5.89155555e-01 5.24066389e-01 -2.34676048e-01 -9.60721672e-01
-6.61028028e-02 -2.63537019e-01 4.16000217e-01 -2.28757247e-01
-2.06351429e-02 -1.34038532e+00 3.13059002e-01 -1.68057039e-01
6.64880931e-01 6.50719345e-01 4.56485420e-01 -5.86231828e-01
-1.01087630e+00 -7.39008561e-02 8.94393474e-02 -3.31506222e-01
-2.39077628e-01 2.44956866e-01 -5.14868386e-02 -8.81313503e-01
-2.02592090e-01 3.87623042e-01 5.88377297e-01 3.81803006e-01
5.60410738e-01 -6.40096784e-01 -9.53750968e-01 9.66564596e-01
1.36768711e+00 -3.09996873e-01 7.49026954e-01 5.90046883e-01
1.49830163e-01 6.53761506e-01 4.87472057e-01 6.89380288e-01
5.91118395e-01 4.52521741e-01 9.76950765e-01 3.82724434e-01
8.41808133e-03 2.80586760e-02 4.40412760e-01 1.03886974e+00
-1.26299784e-01 -5.85785151e-01 -9.50466812e-01 8.43294501e-01
-1.42042398e+00 -7.04424083e-01 -7.99132049e-01 2.37293053e+00
3.28551769e-01 1.97100148e-01 3.84478480e-01 6.56422302e-02
9.30958927e-01 3.59797567e-01 -3.56486291e-01 2.38783032e-01
-7.03735828e-01 -3.02672923e-01 1.31983614e+00 5.98375380e-01
-5.57947040e-01 2.20764711e-01 6.54114771e+00 5.30904889e-01
-9.82650936e-01 -1.85286909e-01 -1.00215629e-01 1.09414579e-02
-8.43406320e-01 7.19341218e-01 -4.62670296e-01 4.74567294e-01
1.41547310e+00 -8.06305528e-01 4.65672523e-01 3.11584294e-01
5.45507014e-01 -1.60490349e-01 -7.24028587e-01 8.15655053e-01
-2.86455780e-01 -1.39456618e+00 -2.97868624e-02 4.95481342e-01
6.19311690e-01 6.85257435e-01 -1.38042331e-01 -5.98449290e-01
6.49432063e-01 -4.70962107e-01 5.42080224e-01 8.22929144e-01
8.71316612e-01 -3.58768970e-01 7.09647536e-02 1.45714313e-01
-1.44439280e+00 -1.32368699e-01 -2.26011887e-01 3.27877514e-02
3.99824202e-01 1.05034900e+00 -8.20615232e-01 1.19878447e+00
5.13169765e-01 1.00376248e+00 -3.57557207e-01 1.64428401e+00
3.96165729e-01 5.40697992e-01 -1.16108537e+00 2.14973390e-02
2.90098459e-01 -5.58414102e-01 1.53953886e+00 6.01058364e-01
9.65156257e-01 2.45179072e-01 -2.81965107e-01 4.94973719e-01
9.44639032e-04 -1.99996129e-01 -9.17464614e-01 -9.20792371e-02
1.15129018e+00 1.29032505e+00 -9.57165420e-01 1.24680296e-01
-5.55659086e-02 3.59622568e-01 -4.93942142e-01 3.43755156e-01
-4.39422011e-01 -5.68171918e-01 8.05055141e-01 7.83674598e-01
1.40652940e-01 -1.04169834e+00 1.36417925e-01 -7.74204791e-01
-3.52787912e-01 -2.43125156e-01 2.59479851e-01 -7.62798309e-01
-9.03691351e-01 5.80114305e-01 -2.84392387e-02 -1.85330725e+00
2.10375860e-01 4.77607548e-02 -6.59307182e-01 2.96855152e-01
-1.25329459e+00 -8.95218074e-01 -3.12985361e-01 7.53735721e-01
-6.42836094e-01 2.48802663e-03 4.19208020e-01 2.09389180e-02
-9.20041129e-02 6.44555092e-02 9.26524580e-01 -5.15159011e-01
2.64249474e-01 -8.33453774e-01 6.47132874e-01 6.37403846e-01
-1.62278965e-01 4.28858906e-01 1.11037743e+00 -8.49395692e-01
-1.96773028e+00 -1.24550009e+00 6.62635207e-01 -2.95370013e-01
1.34775949e+00 -2.23006949e-01 -3.85275185e-01 7.74392724e-01
-1.03093587e-01 2.92972952e-01 1.84337601e-01 -1.35904640e-01
-8.68791621e-03 -2.30453163e-01 -1.31810558e+00 3.17394614e-01
1.46987307e+00 -2.70557523e-01 -1.13945100e-02 8.29111159e-01
7.85367250e-01 -5.10915369e-02 -1.29256952e+00 4.22326535e-01
1.97958022e-01 -7.97331810e-01 1.10321438e+00 -1.58909798e-01
-6.04711115e-01 -6.50940180e-01 -1.62249312e-01 -1.40778673e+00
-3.95686358e-01 -1.35830450e+00 7.21175317e-03 8.66627991e-01
1.70382768e-01 -1.11645687e+00 5.87031364e-01 -2.98159987e-01
-2.64932722e-01 -1.95819542e-01 -1.36988974e+00 -1.49545825e+00
-7.99463540e-02 -1.98407546e-01 6.20043337e-01 1.26892126e+00
2.86511928e-01 1.08485095e-01 1.01861786e-02 6.25102043e-01
1.09160984e+00 -2.15843365e-01 6.20957017e-01 -1.92545497e+00
8.26636106e-02 -8.97681154e-03 -8.21974516e-01 -1.04829383e+00
-1.24811716e-01 -1.25849497e+00 -5.82133949e-01 -1.47950220e+00
-6.45991862e-01 -1.28516185e+00 1.90786272e-01 -2.48404518e-01
1.33665121e+00 1.59993395e-02 -2.44581863e-01 7.13955343e-01
-8.56557786e-01 5.94427943e-01 1.08082604e+00 4.57160957e-02
-2.57671867e-02 2.02238530e-01 -9.91099551e-02 3.60366791e-01
7.20740676e-01 -3.99567395e-01 -3.97056222e-01 -1.72112435e-02
8.02165329e-01 7.33950675e-01 6.65518165e-01 -1.39325202e+00
6.31373942e-01 9.02218595e-02 -4.75586236e-01 -8.24804127e-01
4.14222687e-01 -9.60445881e-01 8.55566800e-01 6.95764720e-01
1.97796822e-01 -4.29928377e-02 -2.04459742e-01 1.19161749e+00
1.73749179e-01 2.62944013e-01 6.79139316e-01 1.54034957e-01
-1.92942813e-01 4.81795639e-01 -2.29059741e-01 -1.97308883e-01
1.43350363e+00 -1.16439864e-01 -8.79553735e-01 -7.78966784e-01
-1.03771746e+00 6.57153189e-01 9.54702497e-01 -5.85969500e-02
1.59450740e-01 -1.21063793e+00 -7.33062744e-01 -1.63278282e-01
3.03566456e-01 -4.90082830e-01 1.86065227e-01 1.32993746e+00
-7.03975737e-01 5.46054542e-01 -4.32283841e-02 -7.36869633e-01
-1.07333732e+00 3.08265090e-01 5.36569178e-01 2.48240411e-01
-5.13288975e-01 6.12213016e-01 -5.69533885e-01 -2.10726395e-01
7.22923279e-02 -3.40815574e-01 1.95013825e-02 1.51355192e-01
-4.33012620e-02 7.44124889e-01 1.40198886e-01 -8.52188647e-01
-5.54048419e-01 4.54155445e-01 4.42377657e-01 -2.14536533e-01
1.40895367e+00 -7.99155474e-01 -8.07007432e-01 5.30690551e-01
1.02870035e+00 2.12330762e-02 -6.38454854e-01 -3.30537319e-01
1.82323717e-02 -2.02261716e-01 -2.54269913e-02 -2.07563758e-01
-8.50481212e-01 -3.51229347e-02 3.17039847e-01 1.26579261e+00
5.51930606e-01 4.54356253e-01 6.20610058e-01 6.72827721e-01
1.07896364e+00 -6.17149234e-01 -5.10750175e-01 6.51167214e-01
9.53788936e-01 -3.48481566e-01 5.42288134e-03 -5.40930390e-01
2.87018806e-01 1.25953627e+00 -3.27810109e-01 -1.62258908e-01
1.14422166e+00 -8.89874399e-02 -6.10193074e-01 -5.43718815e-01
-8.13063800e-01 -2.72455126e-01 -3.80618572e-01 5.40486634e-01
-2.46555910e-01 2.21049875e-01 -6.47611558e-01 -1.42071679e-01
-7.64589190e-01 -2.83789039e-01 1.20338154e+00 9.47508633e-01
-7.75648355e-01 -1.07337773e+00 -5.75529695e-01 5.85909188e-01
1.89903393e-01 6.54383898e-02 -2.93454289e-01 8.21058691e-01
-5.97958624e-01 1.16668868e+00 3.19611073e-01 -7.32289433e-01
-5.22575639e-02 -6.37987316e-01 3.60955209e-01 -7.87387043e-02
9.88623500e-02 -7.57119656e-01 7.24889696e-01 -4.01773244e-01
-5.59560955e-01 -7.80622780e-01 -1.05769181e+00 -9.64399874e-01
-2.97490746e-01 7.72112489e-01 1.08059394e+00 5.56437492e-01
5.42862654e-01 2.15909094e-01 1.17492628e+00 -5.44074416e-01
-3.02033544e-01 -6.09504342e-01 -1.32640541e+00 -2.07889587e-01
7.93365598e-01 -6.94836617e-01 -1.19318414e+00 -5.60910165e-01] | [6.934648513793945, 4.985941410064697] |
72d68bb3-d3d9-434f-b9a4-9d8960fd9445 | an-empirical-investigation-of-global-and | 1904.06834 | null | http://arxiv.org/abs/1904.06834v1 | http://arxiv.org/pdf/1904.06834v1.pdf | An Empirical Investigation of Global and Local Normalization for Recurrent Neural Sequence Models Using a Continuous Relaxation to Beam Search | Globally normalized neural sequence models are considered superior to their
locally normalized equivalents because they may ameliorate the effects of label
bias. However, when considering high-capacity neural parametrizations that
condition on the whole input sequence, both model classes are theoretically
equivalent in terms of the distributions they are capable of representing.
Thus, the practical advantage of global normalization in the context of modern
neural methods remains unclear. In this paper, we attempt to shed light on this
problem through an empirical study. We extend an approach for search-aware
training via a continuous relaxation of beam search (Goyal et al., 2017b) in
order to enable training of globally normalized recurrent sequence models
through simple backpropagation. We then use this technique to conduct an
empirical study of the interaction between global normalization, high-capacity
encoders, and search-aware optimization. We observe that in the context of
inexact search, globally normalized neural models are still more effective than
their locally normalized counterparts. Further, since our training approach is
sensitive to warm-starting with pre-trained models, we also propose a novel
initialization strategy based on self-normalization for pre-training globally
normalized models. We perform analysis of our approach on two tasks: CCG
supertagging and Machine Translation, and demonstrate the importance of global
normalization under different conditions while using search-aware training. | ['Taylor Berg-Kirkpatrick', 'Chris Dyer', 'Kartik Goyal'] | 2019-04-15 | an-empirical-investigation-of-global-and-1 | https://aclanthology.org/N19-1171 | https://aclanthology.org/N19-1171.pdf | naacl-2019-6 | ['ccg-supertagging'] | ['natural-language-processing'] | [ 6.89272523e-01 2.04356194e-01 -3.68413329e-01 -3.32402617e-01
-9.28342342e-01 -5.81286192e-01 7.35323429e-01 7.12560117e-02
-9.55281317e-01 7.05101073e-01 4.97955889e-01 -5.10761917e-01
-3.49725746e-02 -6.86593473e-01 -9.69092429e-01 -6.63976192e-01
3.52109224e-01 4.80495155e-01 -1.48672340e-02 -4.33678389e-01
1.63651973e-01 4.45262492e-01 -1.33317208e+00 1.27933905e-01
6.08003020e-01 6.21073067e-01 3.65547508e-01 6.16415322e-01
1.89244688e-01 5.99345744e-01 -5.26045859e-01 -5.02334416e-01
2.42847234e-01 -6.09245837e-01 -1.00676417e+00 -3.14003527e-01
4.92092580e-01 -1.45052835e-01 -7.79558495e-02 1.15025961e+00
7.54357398e-01 5.01940310e-01 3.98605645e-01 -6.58405066e-01
-5.95349491e-01 1.07339907e+00 -6.79898411e-02 4.44727868e-01
-2.06438512e-01 1.72080219e-01 1.24969554e+00 -7.18636811e-01
8.44903708e-01 1.07551789e+00 7.91015804e-01 6.83927000e-01
-1.43867695e+00 -3.68558824e-01 1.75577879e-01 8.99738893e-02
-1.18163574e+00 -8.26750576e-01 3.87013942e-01 -2.35826857e-02
1.29487085e+00 4.07889277e-01 2.08077982e-01 1.22673059e+00
-1.26014739e-01 5.50600290e-01 8.62352550e-01 -8.09328914e-01
4.82516624e-02 -3.17917801e-02 1.33244112e-01 4.01474208e-01
-2.98737753e-02 2.40447745e-01 -5.61309457e-01 -1.48301035e-01
6.31692827e-01 -3.49266618e-01 -3.97549629e-01 -1.97486386e-01
-1.27883589e+00 9.68566895e-01 5.09355724e-01 6.98775411e-01
-3.97056431e-01 2.86178976e-01 4.48667139e-01 3.08374971e-01
5.46715438e-01 8.58937562e-01 -6.78324699e-01 -3.82916898e-01
-1.09394956e+00 7.27390200e-02 6.66388452e-01 5.89163363e-01
7.55989909e-01 2.62929928e-02 -3.66217226e-01 1.02136660e+00
8.60590935e-02 2.79670745e-01 9.67436016e-01 -1.00320899e+00
4.43063974e-01 2.36998960e-01 -1.87324479e-01 -7.31602311e-01
-5.21921456e-01 -1.06767762e+00 -7.09019184e-01 -2.89857566e-01
4.03291345e-01 -1.24075100e-01 -9.59886193e-01 2.38547850e+00
-1.00529902e-02 8.79656151e-03 5.19698486e-02 8.44490767e-01
3.03533196e-01 6.07234418e-01 1.67369857e-01 -3.10828626e-01
9.95014787e-01 -1.18524015e+00 -7.26830602e-01 -3.90070289e-01
1.11698079e+00 -5.87132514e-01 1.40918040e+00 2.69976445e-03
-1.19287562e+00 -2.42605120e-01 -9.19967115e-01 -1.97643489e-01
-3.31007868e-01 -8.20651837e-03 2.99557418e-01 7.08630860e-01
-1.33546889e+00 8.52477551e-01 -8.61557066e-01 -6.98636353e-01
-2.68557481e-02 4.07377541e-01 -1.88841335e-02 1.69109613e-01
-1.36417043e+00 1.39040947e+00 5.99915266e-01 1.53977767e-01
-5.26236594e-01 -4.58143383e-01 -5.84620774e-01 2.86792874e-01
3.37478310e-01 -6.22885108e-01 1.48333883e+00 -1.08147025e+00
-1.61188984e+00 7.12424040e-01 -2.78571069e-01 -6.40447199e-01
2.22183943e-01 -1.22788116e-01 1.90159846e-02 -1.95520252e-01
-2.03317076e-01 5.16613185e-01 4.10286933e-01 -9.01884973e-01
-2.78420120e-01 -2.56800324e-01 7.07298443e-02 3.59384209e-01
-5.67685127e-01 2.16703922e-01 -4.20101255e-01 -8.87157857e-01
-3.29037160e-02 -1.02319741e+00 -2.99838811e-01 -6.92475140e-01
-2.43498757e-01 -1.81223918e-02 1.68807119e-01 -6.98428988e-01
1.42795646e+00 -1.97192383e+00 4.25316572e-01 3.02637935e-01
-2.01483741e-01 3.68854582e-01 -5.46529710e-01 4.43557024e-01
-1.45349532e-01 4.08453375e-01 -5.41133702e-01 -5.35455048e-01
1.12298235e-01 4.50985312e-01 -1.81220785e-01 4.26970869e-01
-2.76316758e-02 1.22531784e+00 -7.86566257e-01 -2.60747731e-01
-1.96546376e-01 4.75133300e-01 -6.93767071e-01 -1.39376611e-01
-2.60387182e-01 2.35541865e-01 8.06674734e-02 3.07488978e-01
1.68419659e-01 -3.35975856e-01 4.38742906e-01 -7.81517178e-02
-5.86856529e-02 6.28014147e-01 -7.01318860e-01 1.79388320e+00
-7.07715094e-01 6.46444917e-01 -3.93298827e-02 -1.02211678e+00
5.14271379e-01 2.53418803e-01 2.08834205e-02 -1.08751035e+00
2.73189545e-01 3.46922785e-01 1.56100959e-01 -7.92623237e-02
6.87144756e-01 -3.40867966e-01 2.11469308e-01 7.39539266e-01
1.85109898e-01 3.36175561e-01 3.05764765e-01 7.16344118e-02
9.78168964e-01 1.34136721e-01 4.44696173e-02 -3.38694662e-01
3.01596701e-01 -2.66298831e-01 3.35390806e-01 1.13257062e+00
9.57501009e-02 6.71487391e-01 1.69321761e-01 -1.93191357e-02
-1.12733483e+00 -5.41965127e-01 -6.99772015e-02 1.76527894e+00
-1.69315502e-01 -5.46971381e-01 -1.02713037e+00 -5.72088122e-01
-4.56938297e-01 8.37796926e-01 -6.39503598e-01 -4.14828718e-01
-1.06049383e+00 -1.13310552e+00 9.04331625e-01 4.83544737e-01
1.20108187e-01 -1.10932624e+00 -7.36117363e-01 2.76489198e-01
-4.78818774e-01 -9.18638170e-01 -5.10958672e-01 6.18299961e-01
-1.14373982e+00 -5.50811410e-01 -7.92863727e-01 -6.88342512e-01
4.16934758e-01 -8.22876543e-02 1.19294405e+00 3.58117938e-01
3.22762519e-01 1.90556675e-01 -3.08172554e-01 9.65175703e-02
-6.54130816e-01 9.59136367e-01 -8.35710615e-02 -2.24078044e-01
2.77513005e-02 -6.85474873e-01 -3.31560522e-01 1.76756963e-01
-1.07641923e+00 -4.19418812e-02 7.88473248e-01 1.00797451e+00
3.39369535e-01 -4.91167158e-01 4.28492457e-01 -9.72718060e-01
9.08200979e-01 -4.12987798e-01 -4.40269470e-01 4.20238167e-01
-8.28753114e-01 5.28258562e-01 4.52697128e-01 -4.84567136e-01
-9.86429274e-01 -2.53618509e-01 -4.53479141e-01 -2.20537648e-01
1.67357147e-01 9.26728427e-01 6.70669824e-02 -1.63821414e-01
9.24342453e-01 3.88275594e-01 -1.11659437e-01 -7.24856973e-01
6.51029646e-01 4.23464090e-01 5.57682812e-01 -6.61784351e-01
5.46882451e-01 4.36826944e-02 -2.36420795e-01 -5.43545425e-01
-7.08814859e-01 -1.86074659e-01 -5.13161600e-01 8.63968804e-02
5.68883181e-01 -4.85867739e-01 -4.84208196e-01 3.39504451e-01
-1.10833204e+00 -6.97469532e-01 -3.61278296e-01 4.28009897e-01
-6.01248622e-01 3.99612814e-01 -7.86095738e-01 -5.72566569e-01
-2.65408278e-01 -1.09295261e+00 8.25980186e-01 -1.78527862e-01
-3.33037972e-01 -1.13959610e+00 2.75243133e-01 8.05665180e-02
8.85003150e-01 -2.32594416e-01 1.07352364e+00 -1.03211761e+00
-4.28880036e-01 6.15200214e-02 5.53461863e-03 4.23538476e-01
-9.18761715e-02 -3.43261689e-01 -1.06105828e+00 -3.60131204e-01
1.59802675e-01 -2.66101837e-01 1.01171482e+00 2.03593820e-01
8.09231520e-01 -6.21493340e-01 -1.50311396e-01 7.68168688e-01
1.37110567e+00 -1.98261619e-01 4.59590077e-01 7.52612650e-01
5.59311867e-01 4.91336554e-01 2.79721946e-01 8.54475722e-02
2.04944611e-01 9.69911516e-01 4.23086613e-01 3.80473281e-03
-1.56824037e-01 -2.60130733e-01 3.12295020e-01 9.46832895e-01
-1.40418366e-01 -4.97112364e-01 -8.48680675e-01 4.49189782e-01
-1.88674581e+00 -8.04492295e-01 3.23555499e-01 2.27596235e+00
1.20766103e+00 1.51863784e-01 1.12165719e-01 -1.27187401e-01
6.92803502e-01 2.04133824e-01 -3.31959248e-01 -6.40253782e-01
-4.03105289e-01 4.33622628e-01 8.33679199e-01 7.57787585e-01
-9.59115565e-01 1.05409682e+00 6.76883888e+00 9.84192610e-01
-1.37517095e+00 4.55270469e-01 6.06737018e-01 -3.21010053e-01
-4.55838174e-01 -6.24072812e-02 -7.20620275e-01 3.70661795e-01
1.52651668e+00 2.12592399e-03 7.64519453e-01 4.77508664e-01
-3.21692936e-02 1.80907458e-01 -1.12749743e+00 6.58519268e-01
3.86690013e-02 -1.34234262e+00 5.88283725e-02 7.18403384e-02
6.69055939e-01 5.79745531e-01 1.44613922e-01 3.42341930e-01
3.53699774e-01 -1.14099741e+00 7.91908622e-01 2.71222860e-01
4.90614414e-01 -5.35420179e-01 6.66604936e-01 6.21514976e-01
-4.75835770e-01 -3.37637812e-02 -2.90346473e-01 -1.39040425e-01
1.38400555e-01 3.78451586e-01 -7.08652973e-01 3.96258801e-01
3.43890935e-01 2.95573533e-01 -6.16153121e-01 7.89104879e-01
-1.38776883e-01 8.41745675e-01 -6.84137285e-01 1.00158574e-03
4.81328726e-01 -8.39942619e-02 4.07915890e-01 1.51883185e+00
1.97551623e-01 -1.91444010e-01 -2.87809670e-01 6.98473215e-01
-3.07550400e-01 3.07971746e-01 -1.86685622e-01 -5.38598821e-02
4.28782910e-01 9.42683101e-01 -6.82308555e-01 -3.16338062e-01
-2.16267392e-01 8.05536926e-01 6.83850050e-01 6.20512784e-01
-7.02415109e-01 -7.57203251e-02 4.29294407e-01 -1.71332523e-01
3.01680833e-01 -2.26754978e-01 -4.58426207e-01 -1.18450534e+00
9.16737877e-03 -1.07095706e+00 2.67672300e-01 -4.63354379e-01
-7.43227124e-01 7.45214999e-01 -1.30718902e-01 -5.48978090e-01
-6.60011530e-01 -2.68457472e-01 -2.47142270e-01 1.04928505e+00
-1.70880675e+00 -1.01009881e+00 2.76008964e-01 2.91264564e-01
3.99965882e-01 2.99770683e-01 8.38273168e-01 4.47234392e-01
-5.52219570e-01 8.80729795e-01 4.86287326e-01 -1.10518076e-01
6.41413391e-01 -9.68437970e-01 5.47890663e-01 1.07162750e+00
4.03628111e-01 1.07952976e+00 7.59971619e-01 -4.91768777e-01
-1.16347396e+00 -8.77433777e-01 1.26720822e+00 -3.96926850e-01
5.17549455e-01 -4.38182861e-01 -1.11250603e+00 8.36759210e-01
3.11824948e-01 -3.13829124e-01 4.81885254e-01 4.24968094e-01
-2.23353982e-01 2.81093329e-01 -6.53315365e-01 6.47675157e-01
1.25036800e+00 -6.76170588e-01 -5.48199236e-01 3.44479322e-01
8.34738493e-01 -3.75647515e-01 -6.97935164e-01 5.37961364e-01
4.04220968e-01 -8.43490183e-01 8.16275358e-01 -6.32809401e-01
1.72931343e-01 1.64728865e-01 -4.01357859e-01 -1.26258886e+00
-4.24371332e-01 -6.51085138e-01 1.10533997e-01 1.16504431e+00
7.01930702e-01 -8.06840897e-01 7.80090094e-01 3.45629901e-01
-1.78202599e-01 -7.82526255e-01 -9.43191826e-01 -8.83759558e-01
3.66268843e-01 -3.13041955e-01 5.01616061e-01 9.84620750e-01
-5.99803291e-02 3.59157681e-01 -3.99801999e-01 -1.72139183e-01
1.18243784e-01 -1.63529038e-01 3.78310561e-01 -7.50862360e-01
-5.57217002e-01 -7.31870234e-01 7.94416014e-03 -1.21169150e+00
3.25584024e-01 -1.14670360e+00 3.00201535e-01 -1.21366882e+00
1.68172032e-01 -3.87544334e-01 -5.80005169e-01 6.25671744e-01
-1.13018379e-01 4.89848226e-01 3.28663409e-01 4.38322127e-01
-3.91722858e-01 4.13417250e-01 8.42564046e-01 1.87441483e-01
-1.94504950e-02 -1.11981720e-01 -6.53688788e-01 4.07559544e-01
8.55183959e-01 -5.20375311e-01 -3.02778900e-01 -7.30059564e-01
5.81775784e-01 -2.78428793e-01 2.30051503e-01 -5.51654398e-01
4.04720008e-01 -2.34896946e-03 -1.86179906e-01 -1.42477110e-01
3.23071063e-01 -5.67155838e-01 -3.77606526e-02 4.32243675e-01
-9.14760649e-01 1.92863554e-01 2.21775815e-01 1.94891930e-01
-1.61322311e-01 -5.65181315e-01 6.42987609e-01 -1.88951403e-01
-3.58428717e-01 -1.65266573e-01 -4.92301732e-01 2.01470226e-01
2.56593347e-01 -1.55441657e-01 -1.89765379e-01 -4.55773979e-01
-7.98721611e-01 -8.88116956e-02 6.19548321e-01 2.84053624e-01
1.14822321e-01 -1.05020881e+00 -5.35685122e-01 1.50741726e-01
-1.77376717e-01 -3.45661521e-01 -1.81589112e-01 9.57430422e-01
-2.48584703e-01 7.59619117e-01 5.63636608e-02 -3.77724022e-01
-1.00198221e+00 5.43562889e-01 4.70730603e-01 -4.62346077e-01
-1.66554049e-01 9.08103406e-01 2.62024045e-01 -7.87075758e-01
4.37212944e-01 -4.57975388e-01 3.40071805e-02 -4.39497316e-03
4.07736637e-02 2.50555217e-01 4.89037961e-01 -7.10985065e-01
-3.39265585e-01 3.85457814e-01 -6.99161589e-02 -3.76696587e-01
1.26640475e+00 -2.18181655e-01 -1.52465582e-01 3.55502695e-01
1.39130783e+00 -4.95188264e-03 -8.68200481e-01 -5.12355268e-01
3.56344491e-01 -1.18593298e-01 4.23812717e-02 -1.02917969e+00
-9.80192959e-01 8.48746300e-01 4.82742369e-01 4.72792722e-02
1.13859212e+00 -5.36674820e-02 6.30829096e-01 5.72001100e-01
-1.06327899e-01 -9.90361452e-01 -3.11486065e-01 8.78199458e-01
5.58486998e-01 -8.48091364e-01 -3.85107845e-01 -2.54957657e-02
-1.71370924e-01 1.00610363e+00 2.12238595e-01 1.99387938e-01
1.28345624e-01 1.38622969e-01 1.73261434e-01 3.30906361e-02
-7.41518974e-01 -1.86274275e-01 1.11639775e-01 2.50059664e-01
5.20867705e-01 -2.58814037e-01 -4.85846639e-01 3.50595832e-01
-3.51719528e-01 -7.71656260e-02 2.52469271e-01 9.18803394e-01
-3.58030498e-01 -1.33697939e+00 -1.83305085e-01 2.27770716e-01
-6.91672087e-01 -7.59175956e-01 -2.89559931e-01 6.50000274e-01
-2.13296339e-01 6.54787123e-01 -1.09032422e-01 -2.98626184e-01
8.69353190e-02 4.21900660e-01 5.22024333e-01 -6.39087498e-01
-9.50771391e-01 1.23901412e-01 2.63879925e-01 -4.78409022e-01
-4.52871382e-01 -5.65858305e-01 -8.88646543e-01 -1.61027148e-01
-6.13942623e-01 2.56196797e-01 8.83003414e-01 1.13553572e+00
4.87329543e-01 4.88176107e-01 1.49970978e-01 -7.69974470e-01
-1.00499845e+00 -1.05427480e+00 -5.22769280e-02 3.35898012e-01
4.13426608e-01 -3.08414429e-01 -3.43859881e-01 -1.10501878e-01] | [11.505279541015625, 9.579493522644043] |
56a1b681-9750-4baf-95e2-753f97f4edcd | automated-essay-scoring-using-efficient | 2102.13136 | null | https://arxiv.org/abs/2102.13136v1 | https://arxiv.org/pdf/2102.13136v1.pdf | Automated essay scoring using efficient transformer-based language models | Automated Essay Scoring (AES) is a cross-disciplinary effort involving Education, Linguistics, and Natural Language Processing (NLP). The efficacy of an NLP model in AES tests it ability to evaluate long-term dependencies and extrapolate meaning even when text is poorly written. Large pretrained transformer-based language models have dominated the current state-of-the-art in many NLP tasks, however, the computational requirements of these models make them expensive to deploy in practice. The goal of this paper is to challenge the paradigm in NLP that bigger is better when it comes to AES. To do this, we evaluate the performance of several fine-tuned pretrained NLP models with a modest number of parameters on an AES dataset. By ensembling our models, we achieve excellent results with fewer parameters than most pretrained transformer-based models. | ['Amir Jafari', 'Akanksha Malhotra', 'Christopher M Ormerod'] | 2021-02-25 | null | null | null | null | ['automated-essay-scoring'] | ['natural-language-processing'] | [-1.22482002e-01 -9.89668742e-02 -1.73573911e-01 -4.18916553e-01
-1.04390788e+00 -8.75637591e-01 4.97034848e-01 4.43433642e-01
-5.80757141e-01 9.57308710e-01 5.52436352e-01 -7.56976008e-01
-2.46564075e-01 -6.38331413e-01 -7.32136548e-01 1.40361965e-01
3.65234882e-01 7.09695876e-01 9.60050300e-02 -4.72079098e-01
3.54015768e-01 2.09059700e-01 -1.10125339e+00 4.30297315e-01
1.14296532e+00 5.80560327e-01 -2.40082115e-01 7.19856679e-01
-4.28866267e-01 1.10734880e+00 -6.56429350e-01 -1.13387299e+00
-1.23683959e-01 -1.98328882e-01 -9.95220482e-01 -8.48528743e-01
8.36759865e-01 -4.06629667e-02 -2.78910577e-01 7.21557856e-01
2.92749673e-01 4.55934517e-02 6.47702813e-01 -9.05797184e-01
-9.44473088e-01 9.66843605e-01 -1.88715875e-01 4.94542480e-01
6.66411221e-01 5.00033498e-02 1.50560498e+00 -9.58539307e-01
3.62215549e-01 9.29313362e-01 1.10081160e+00 2.82972038e-01
-1.22309780e+00 -7.69347787e-01 -6.22999705e-02 3.98946673e-01
-1.13192785e+00 -3.51363629e-01 6.34591997e-01 -3.05167019e-01
1.22017407e+00 1.10470213e-01 7.25955784e-01 1.24627829e+00
5.25701821e-01 9.43254530e-01 1.27857137e+00 -7.28362799e-01
-2.72434443e-01 1.69989243e-01 6.44545615e-01 1.00244725e+00
1.40046760e-01 -8.67419317e-02 -1.14268923e+00 -3.78500186e-02
2.07913056e-01 -6.23902321e-01 5.68412989e-02 3.02868903e-01
-1.05271959e+00 8.03666949e-01 -4.58759395e-03 3.64610940e-01
-4.20518853e-02 6.60284758e-02 2.11767167e-01 7.16520011e-01
6.35216713e-01 8.49522948e-01 -8.87325048e-01 -6.29248321e-01
-1.40078211e+00 3.74914050e-01 1.19093108e+00 6.35616422e-01
2.78153002e-01 -1.04345299e-01 -3.94409359e-01 9.74107623e-01
1.63562492e-01 3.29804063e-01 6.94067657e-01 -7.77053833e-01
7.61942983e-01 7.95283139e-01 -2.06054285e-01 -8.56943667e-01
-3.34489107e-01 -5.09568691e-01 -2.97578186e-01 -3.05119127e-01
8.43925714e-01 -2.44433403e-01 -7.07374811e-01 1.70159650e+00
-2.75894821e-01 2.25310966e-01 7.37887295e-03 3.36724311e-01
1.04909182e+00 7.52222240e-01 4.05708045e-01 1.89795569e-02
1.38443482e+00 -9.39605355e-01 -6.40729547e-01 -5.59830546e-01
9.61219370e-01 -1.00602305e+00 1.54145598e+00 6.74582183e-01
-1.38397324e+00 -3.78730476e-01 -9.75167930e-01 -6.11752748e-01
-6.22136354e-01 2.97606140e-01 6.59904242e-01 7.90474772e-01
-1.08924973e+00 6.48759365e-01 -5.50993741e-01 -1.29509479e-01
3.32180560e-01 4.11887288e-01 -2.20566675e-01 -1.24207744e-02
-1.45421898e+00 1.47229171e+00 2.39997953e-01 -1.60944983e-01
-4.29438591e-01 -1.44438803e+00 -7.75416970e-01 5.05570412e-01
-1.86875891e-02 -7.02815175e-01 1.53575158e+00 -4.15762275e-01
-1.74539042e+00 9.69487548e-01 -2.64696866e-01 -4.64845836e-01
3.64558429e-01 -3.10221285e-01 -2.38751829e-01 -3.60379905e-01
-6.60736412e-02 3.45484227e-01 2.35930651e-01 -3.62343371e-01
-4.68029648e-01 -2.12176263e-01 -1.98654503e-01 2.13273570e-01
-9.28883970e-01 3.24664086e-01 -2.36881319e-02 -4.74091798e-01
-3.45351070e-01 -7.01736510e-01 2.08519295e-01 -4.77050245e-01
2.20236331e-02 -9.09433722e-01 2.90610939e-01 -9.08299029e-01
1.58820319e+00 -1.50341058e+00 -6.33114427e-02 -1.32992407e-02
-1.75558627e-02 3.08985889e-01 -2.60122657e-01 3.52102429e-01
4.47071455e-02 2.45184094e-01 4.94274721e-02 -3.08894634e-01
5.06819248e-01 6.79037422e-02 -6.51741922e-01 -2.38656942e-02
6.63631111e-02 1.27865160e+00 -8.76619399e-01 -6.46269321e-01
-1.68548316e-01 2.34221548e-01 -5.63434362e-01 5.50714284e-02
-2.05718398e-01 -8.06461647e-02 -3.41397822e-01 5.14098704e-01
-2.18330175e-02 -7.85508379e-02 5.03446050e-02 2.24867761e-01
-2.90307134e-01 1.17901182e+00 -6.15395844e-01 1.78360987e+00
-6.25566363e-01 1.07024527e+00 -4.90146786e-01 -8.49895120e-01
8.42362583e-01 3.64976466e-01 1.74273074e-01 -7.77401626e-01
8.11041668e-02 3.95460010e-01 2.65894830e-01 -4.26421314e-01
6.46321774e-01 -2.61785567e-01 -3.32884759e-01 8.68597984e-01
3.52441609e-01 -6.45717978e-01 5.09425938e-01 2.84730077e-01
1.33426607e+00 2.09073201e-01 1.45033047e-01 -5.42294383e-01
7.12166786e-01 2.88004696e-01 2.89333731e-01 5.48039436e-01
-4.03778553e-02 1.61626562e-01 4.49963301e-01 -5.45723557e-01
-9.75833058e-01 -1.00253582e+00 -3.20862472e-01 1.53461933e+00
-6.87795699e-01 -7.20651865e-01 -6.87179565e-01 -8.13251138e-01
-4.88376431e-03 1.10405874e+00 -3.99831504e-01 -1.25812113e-01
-6.79591954e-01 -9.65885520e-01 1.04815722e+00 4.80644524e-01
2.53517777e-01 -1.15864050e+00 -3.03979039e-01 3.21641356e-01
-3.03343266e-01 -1.19505107e+00 -4.24173862e-01 3.57536972e-02
-6.77000940e-01 -8.74056756e-01 -3.27278584e-01 -8.83245468e-01
1.90324157e-01 -5.12383282e-01 1.73788905e+00 -7.58048438e-04
-2.37156413e-02 4.73475724e-01 4.17186245e-02 -1.03787124e+00
-2.79828727e-01 5.52930832e-01 5.71164191e-02 -6.72870100e-01
1.15236795e+00 -6.02549314e-01 8.34322497e-02 -2.40046546e-01
-5.16209662e-01 -6.74196631e-02 5.56009471e-01 8.05162728e-01
1.82973325e-01 -1.13275006e-01 6.49539113e-01 -9.81840074e-01
1.48972106e+00 -1.86915979e-01 -3.69500667e-01 6.89792514e-01
-8.49515021e-01 1.74238428e-01 8.47290456e-01 -4.81842399e-01
-1.07900214e+00 -4.16136414e-01 -1.40355319e-01 1.75430909e-01
1.65908128e-01 9.98240471e-01 3.61441135e-01 -3.27546239e-01
6.67635322e-01 1.76975518e-01 -4.87952054e-01 -3.97019148e-01
1.94360152e-01 5.56767046e-01 5.11347353e-01 -1.25105274e+00
7.50006437e-01 -4.17636186e-01 -5.64950034e-02 -5.63803852e-01
-1.62591660e+00 -4.22173142e-02 -6.94496095e-01 -7.39697814e-02
4.47012186e-01 -7.78976083e-01 -8.40296388e-01 1.10401593e-01
-1.19900560e+00 -5.84143043e-01 -2.84405798e-01 5.10865986e-01
-2.32607290e-01 1.45927891e-01 -9.02513206e-01 -3.14006776e-01
-5.58866382e-01 -8.61765683e-01 4.82396811e-01 4.04987484e-01
-6.47653282e-01 -1.52298272e+00 6.00878358e-01 7.55579948e-01
4.19111043e-01 -2.82741398e-01 1.23943329e+00 -9.77956772e-01
-6.28224611e-02 -1.62127107e-01 -5.69711514e-02 3.95013005e-01
-3.83627832e-01 1.31667778e-01 -8.75559390e-01 1.87411383e-01
-6.70438260e-02 -8.13821614e-01 7.26721883e-01 1.42521903e-01
1.23921740e+00 -2.82146692e-01 1.30339608e-01 3.93371105e-01
1.00050974e+00 -4.01614934e-01 2.45028585e-01 4.48762000e-01
5.39649725e-01 6.08916640e-01 2.42878184e-01 -1.26213312e-01
9.07133937e-01 2.33578980e-01 -5.61563015e-01 4.31167811e-01
-1.80508077e-01 -7.39698529e-01 6.98143184e-01 1.49058652e+00
4.08737697e-02 -3.10504943e-01 -1.48904121e+00 7.64502287e-01
-1.52868545e+00 -6.78291738e-01 -2.87028402e-01 1.89274585e+00
1.43806267e+00 2.90814072e-01 -2.99753875e-01 1.43192327e-02
-1.04284566e-03 2.95978803e-02 -1.16323516e-01 -1.09048450e+00
1.34589793e-02 1.10735118e+00 1.19970992e-01 5.53046823e-01
-7.64308274e-01 1.18319929e+00 7.06693840e+00 8.24349999e-01
-8.94771516e-01 5.92868179e-02 4.52317417e-01 -7.48414695e-02
-4.85940784e-01 -9.83930901e-02 -1.04397595e+00 3.36566895e-01
1.63348210e+00 -6.44844115e-01 3.16252172e-01 3.68776798e-01
-1.86313856e-02 1.11221246e-01 -1.39390409e+00 6.56235516e-01
3.37858081e-01 -1.13248563e+00 1.55844823e-01 -3.14123034e-01
8.10790300e-01 7.61805847e-02 2.08830729e-01 8.39622617e-01
8.46963406e-01 -1.52941668e+00 5.08740664e-01 4.35486913e-01
5.85376084e-01 -6.14548981e-01 6.76569402e-01 4.89309937e-01
-7.91554272e-01 -1.18814185e-01 -4.05627221e-01 -6.47001922e-01
-1.80280417e-01 4.03375119e-01 -8.32612574e-01 1.81074813e-01
5.83221674e-01 6.31961823e-01 -9.35849309e-01 8.20862353e-01
-9.50903893e-01 1.13880825e+00 -4.25437093e-01 -4.27810550e-01
3.05850834e-01 -1.77778780e-01 2.14666381e-01 1.38726330e+00
3.04385453e-01 3.83123726e-01 2.02463940e-02 8.15964162e-01
-5.95367432e-01 3.94601315e-01 -2.59915352e-01 -2.32582271e-01
4.13197786e-01 1.30044997e+00 -1.21162958e-01 -2.69904017e-01
-5.24778306e-01 5.38661122e-01 8.36626530e-01 1.74236998e-01
-6.26169026e-01 -2.26604775e-01 2.97529876e-01 -7.45592043e-02
-1.45802110e-01 -2.95436263e-01 -8.39970171e-01 -1.33887482e+00
-7.28077888e-02 -1.09251940e+00 7.12246895e-01 -9.07039225e-01
-1.57017791e+00 1.82897761e-01 -1.96698323e-01 -5.49967289e-01
-2.22210124e-01 -9.22524929e-01 -8.14066172e-01 1.04333794e+00
-1.70499420e+00 -1.20262861e+00 -1.11808283e-02 4.01017308e-01
5.06757617e-01 -2.78807610e-01 1.10452080e+00 2.33128756e-01
-5.99223316e-01 1.03255570e+00 -1.44911095e-01 2.99887717e-01
1.13611126e+00 -1.47529447e+00 3.00232410e-01 8.52754116e-01
6.24307334e-01 6.39661491e-01 5.71535647e-01 -4.62496102e-01
-1.23333764e+00 -6.42066896e-01 1.92492568e+00 -1.31097865e+00
1.28998423e+00 -2.93395877e-01 -9.33391571e-01 1.05966544e+00
4.00024533e-01 -2.21540645e-01 1.00381839e+00 6.02744997e-01
-5.51636159e-01 -2.63593569e-02 -7.80506492e-01 6.06593013e-01
7.29656219e-01 -8.15146208e-01 -1.37971520e+00 3.25435132e-01
4.05970395e-01 -6.26001775e-01 -1.21153259e+00 2.31753439e-01
7.62605548e-01 -4.10853654e-01 9.19707358e-01 -1.06235671e+00
1.01458633e+00 2.14785665e-01 3.86445165e-01 -1.55304277e+00
-3.97256136e-01 -5.40215194e-01 -1.67524099e-01 1.27359641e+00
8.67137253e-01 -4.66530502e-01 9.22097325e-01 1.06385207e+00
-2.27625698e-01 -9.29061830e-01 -6.91316545e-01 -5.64378262e-01
9.12904501e-01 -6.94157481e-01 5.58797479e-01 1.20113790e+00
5.85637808e-01 8.48985553e-01 2.27816328e-01 -1.76732436e-01
4.88873839e-01 4.18957137e-03 5.08344948e-01 -1.59108281e+00
1.07225971e-02 -7.79134572e-01 -4.04390395e-02 -7.76826620e-01
6.75156951e-01 -1.25413752e+00 -1.76822200e-01 -1.45518506e+00
2.98857629e-01 -4.27777231e-01 -3.27366859e-01 9.22910988e-01
-3.14773202e-01 2.66500175e-01 1.52147308e-01 -2.13184640e-01
-4.42553848e-01 3.82616311e-01 1.07502246e+00 -2.77433395e-01
-1.37577593e-01 -2.05133691e-01 -8.54821980e-01 7.55561054e-01
8.58422279e-01 -6.10273480e-01 -3.57456744e-01 -7.65533507e-01
7.61356413e-01 -1.65927872e-01 -3.89495529e-02 -7.06211150e-01
7.71201491e-01 -7.59340748e-02 6.12061620e-01 -2.91691482e-01
1.12995863e-01 -3.76415282e-01 -5.43104231e-01 2.32961494e-02
-6.44234300e-01 4.30925667e-01 6.00707650e-01 -1.93074659e-01
-3.51983607e-01 -4.15442437e-01 3.21947515e-01 -1.22851335e-01
-3.38959455e-01 1.04574300e-01 -2.01313883e-01 6.33726776e-01
5.75816453e-01 1.75777391e-01 -5.15014410e-01 -2.38872722e-01
-2.63960242e-01 4.20284748e-01 -3.83219346e-02 4.90926087e-01
4.16758060e-01 -1.05383444e+00 -1.21523380e+00 1.56108841e-01
-1.86766341e-01 -1.11127831e-01 -9.84549969e-02 7.98073113e-01
-6.13632560e-01 1.07683754e+00 -9.62591022e-02 -1.43990293e-01
-1.34200633e+00 -9.27595701e-03 2.32779667e-01 -1.02768517e+00
-2.10852161e-01 1.11991644e+00 -4.26601946e-01 -8.55950415e-01
1.04767449e-01 -2.88042128e-01 -5.57594001e-01 9.03451256e-03
5.47116876e-01 3.50672692e-01 4.19131249e-01 -1.65299922e-01
-1.68923642e-02 4.10660177e-01 -2.80652344e-01 -2.35549465e-01
1.53047335e+00 4.77733552e-01 -2.80597299e-01 4.80579704e-01
7.40207195e-01 3.07473511e-01 -5.62557757e-01 -2.72050649e-01
5.37333667e-01 -7.26590380e-02 4.33570474e-01 -1.33067977e+00
-4.74144787e-01 9.31821525e-01 -1.00451522e-01 -1.18259341e-01
6.79469109e-01 -4.60140526e-01 9.76203501e-01 7.19027102e-01
1.31609634e-01 -1.21745396e+00 -1.01528630e-01 1.08356202e+00
6.84264183e-01 -9.66050148e-01 2.81395435e-01 -9.23266187e-02
-4.36286628e-01 1.34593248e+00 7.68141031e-01 -3.74122635e-02
5.59820414e-01 4.84426916e-01 -2.73556020e-02 -2.64853895e-01
-1.15416718e+00 2.72645891e-01 7.79913485e-01 4.81532440e-02
1.04668450e+00 2.18446739e-02 -6.01241887e-01 1.07735658e+00
-1.10518861e+00 2.95535743e-01 7.52576172e-01 4.63778555e-01
-1.64706871e-01 -1.27875245e+00 -1.89048901e-01 5.81939340e-01
-9.44347739e-01 -7.16963768e-01 -6.84641123e-01 6.69786334e-01
-2.04942584e-01 8.71353745e-01 -5.96485026e-02 -1.69887125e-01
3.62310410e-01 8.13863695e-01 7.86898673e-01 -6.93243980e-01
-1.18892217e+00 -8.15842986e-01 3.15769047e-01 -1.92111596e-01
-7.32734874e-02 -6.28913105e-01 -8.98317277e-01 -5.78121305e-01
-8.60737413e-02 3.75184417e-01 6.52042806e-01 1.33304560e+00
4.21699360e-02 5.84743798e-01 -6.32696897e-02 -8.22619200e-02
-8.91466260e-01 -1.11104798e+00 -1.12652749e-01 -1.45391375e-02
-5.74075878e-02 -1.76107600e-01 -2.79086173e-01 -2.30335090e-02] | [11.234313011169434, 9.363542556762695] |
c279e444-d8e2-44fb-b769-ddd916823419 | entropy-enhanced-multimodal-attention-model | 1908.08191 | null | https://arxiv.org/abs/1908.08191v1 | https://arxiv.org/pdf/1908.08191v1.pdf | Entropy-Enhanced Multimodal Attention Model for Scene-Aware Dialogue Generation | With increasing information from social media, there are more and more videos available. Therefore, the ability to reason on a video is important and deserves to be discussed. TheDialog System Technology Challenge (DSTC7) (Yoshino et al. 2018) proposed an Audio Visual Scene-aware Dialog (AVSD) task, which contains five modalities including video, dialogue history, summary, and caption, as a scene-aware environment. In this paper, we propose the entropy-enhanced dynamic memory network (DMN) to effectively model video modality. The attention-based GRU in the proposed model can improve the model's ability to comprehend and memorize sequential information. The entropy mechanism can control the attention distribution higher, so each to-be-answered question can focus more specifically on a small set of video segments. After the entropy-enhanced DMN secures the video context, we apply an attention model that in-corporates summary and caption to generate an accurate answer given the question about the video. In the official evaluation, our system can achieve improved performance against the released baseline model for both subjective and objective evaluation metrics. | ['Lun-Wei Ku', 'Yun-Nung Chen', 'Chao-Chun Hsu', 'Kuan-Yen Lin'] | 2019-08-22 | null | null | null | null | ['scene-aware-dialogue'] | ['computer-vision'] | [ 5.81541061e-02 1.08974732e-01 -5.50201908e-02 -2.63453394e-01
-6.41378820e-01 -3.94208074e-01 6.23615444e-01 -1.85326964e-01
-4.41114962e-01 6.25901759e-01 8.80270362e-01 -1.54236667e-02
3.39159846e-01 -2.74129152e-01 -5.29745877e-01 -3.15577090e-01
3.32947046e-01 -5.28881215e-02 4.47659701e-01 -8.83768275e-02
3.60445768e-01 -5.01266867e-02 -1.47392511e+00 8.91077518e-01
8.63335609e-01 1.19781911e+00 9.62314487e-01 1.11507857e+00
-1.44860491e-01 1.59728312e+00 -6.35240257e-01 -3.74327838e-01
-3.86638731e-01 -7.44873703e-01 -1.19590116e+00 8.35195258e-02
2.80871958e-01 -9.14619148e-01 -1.01241755e+00 9.03189540e-01
5.67372561e-01 5.24512231e-01 4.58374739e-01 -1.36564946e+00
-9.61108446e-01 6.42073691e-01 -3.78534943e-01 6.54831886e-01
7.74778306e-01 5.49135506e-01 6.87312543e-01 -8.84269357e-01
6.76618695e-01 1.51212740e+00 -1.44697890e-01 7.66632199e-01
-4.32774425e-01 -3.46268415e-01 2.62790442e-01 1.14634979e+00
-9.68796432e-01 -6.19478524e-01 5.76451898e-01 -1.68192029e-01
1.08217430e+00 5.29004991e-01 6.25331223e-01 1.27912807e+00
1.84308127e-01 1.30237472e+00 5.55129230e-01 -7.54978284e-02
9.98584367e-03 1.54633552e-01 -1.24262176e-01 2.52936006e-01
-5.95246553e-01 -2.40660802e-01 -8.69568348e-01 3.74077529e-01
4.99715030e-01 3.63004301e-03 -5.95269442e-01 2.03619316e-01
-1.21267605e+00 6.74470484e-01 3.88143986e-01 9.68618840e-02
-5.74630558e-01 -2.72323694e-02 5.43929219e-01 1.11991182e-01
3.62308025e-01 2.47068539e-01 -2.27433890e-01 -5.91575027e-01
-7.38517225e-01 -4.65399101e-02 6.49257004e-01 1.01457870e+00
2.60363162e-01 -7.28173628e-02 -9.65132356e-01 7.17870891e-01
5.59228063e-01 6.26356661e-01 4.56252486e-01 -1.55231214e+00
6.77832067e-01 4.60514337e-01 -6.92361146e-02 -1.18031383e+00
3.27338874e-02 1.97607547e-01 -8.55214119e-01 -4.45257306e-01
-2.43527070e-03 -3.31470281e-01 -5.67981362e-01 1.75106537e+00
2.15701073e-01 2.44462952e-01 1.41092494e-01 1.20678282e+00
1.35087669e+00 1.35421193e+00 3.49113494e-01 -5.17206132e-01
1.53887105e+00 -1.32573247e+00 -1.37099195e+00 -3.55365574e-01
1.25261649e-01 -8.21062088e-01 1.06454980e+00 1.27126321e-01
-1.32180405e+00 -7.06061125e-01 -9.03565228e-01 -3.93168241e-01
-5.14134094e-02 -2.25676537e-01 5.95980212e-02 -6.76607117e-02
-1.25749993e+00 4.92741987e-02 -5.07124305e-01 -4.93644923e-01
2.75719464e-01 -8.60524401e-02 -1.27235234e-01 -2.14238927e-01
-1.77653575e+00 9.03910577e-01 5.48630238e-01 1.37439564e-01
-8.96024227e-01 -3.46811384e-01 -9.44983363e-01 2.32782736e-01
6.73956096e-01 -7.38898933e-01 1.53034210e+00 -9.95988369e-01
-1.60561955e+00 5.78610897e-01 -2.51833111e-01 -6.52698517e-01
3.33451599e-01 -3.39577466e-01 -4.71935809e-01 1.10059512e+00
-2.07095623e-01 1.31521487e+00 8.37942660e-01 -9.42692697e-01
-5.74451745e-01 3.52246165e-02 4.66955483e-01 7.08447456e-01
-4.70607907e-01 3.42734039e-01 -8.13405633e-01 -5.03487110e-01
-4.21947181e-01 -6.84984982e-01 3.43858778e-01 -1.88777149e-02
-3.51861298e-01 -1.83342412e-01 1.25749362e+00 -1.26672506e+00
1.64187849e+00 -2.27636933e+00 4.00864631e-01 -6.19446278e-01
2.92609036e-01 3.28059196e-01 -2.07427353e-01 3.80519241e-01
2.01696754e-01 7.43914023e-02 1.50148258e-01 -3.07775140e-01
-1.95166990e-01 4.85861499e-04 -3.60107124e-01 -4.91903676e-03
2.02349812e-01 9.90494370e-01 -8.07186902e-01 -8.54309618e-01
4.39720422e-01 5.34859896e-01 -6.77691519e-01 7.73772120e-01
-5.30037761e-01 4.94971842e-01 -2.90674418e-01 2.51901507e-01
5.12522876e-01 -5.78947544e-01 -1.78462148e-01 -3.23875546e-01
5.13597280e-02 8.07580501e-02 -7.78793633e-01 1.83761823e+00
-4.59801435e-01 1.14114881e+00 1.46691918e-01 -3.27921033e-01
3.35670620e-01 7.18762577e-01 2.59215504e-01 -8.58866572e-01
2.67455310e-01 -4.34998333e-01 -2.21687347e-01 -1.13448608e+00
7.89770842e-01 5.53358197e-01 7.16804946e-03 3.62552702e-01
2.24141166e-01 3.51446331e-01 1.06422327e-01 9.02373612e-01
7.61164308e-01 -1.21666174e-02 1.14502236e-01 2.40178302e-01
7.80538023e-01 -3.01661402e-01 1.67048931e-01 5.49914956e-01
-4.98676807e-01 7.24515855e-01 4.07221019e-01 7.15780854e-02
-8.77595067e-01 -6.75096512e-01 4.10000145e-01 1.40442801e+00
6.40716791e-01 -4.95027661e-01 -1.07808065e+00 -5.47150075e-01
-5.23615897e-01 1.00492370e+00 -5.38958251e-01 -5.77461958e-01
-3.63860101e-01 -6.23795576e-02 1.94908604e-01 3.38066041e-01
1.24002063e+00 -1.64946604e+00 -7.12291479e-01 3.42085101e-02
-1.17633629e+00 -1.37318003e+00 -1.08594310e+00 -7.08653033e-01
-3.31530750e-01 -8.74968290e-01 -8.98216188e-01 -6.02943361e-01
2.44606555e-01 5.15064001e-01 8.92476439e-01 6.68010563e-02
1.90950215e-01 8.07289481e-01 -6.01743340e-01 -6.11197352e-02
-4.39524829e-01 -1.17184713e-01 -3.32207412e-01 3.49220127e-01
3.49674970e-01 -1.28204614e-01 -8.29745829e-01 3.55926841e-01
-1.04607725e+00 5.80677569e-01 3.60955566e-01 5.08846879e-01
2.86075175e-01 -4.32921559e-01 6.35199130e-01 -2.72849381e-01
6.85054004e-01 -8.57753754e-01 2.73469351e-02 4.49669302e-01
-6.33042008e-02 -3.43917273e-02 3.54266852e-01 -4.52231348e-01
-1.47491717e+00 -2.88959980e-01 -1.40386283e-01 -6.33455098e-01
-1.60164610e-02 5.17753005e-01 -6.10693812e-01 3.43785375e-01
1.11033559e-01 4.53306586e-01 2.13793479e-02 -1.18735805e-01
3.35819513e-01 9.42108274e-01 6.93228900e-01 -5.28158098e-02
2.17869073e-01 1.22234032e-01 -5.42556942e-01 -7.63243198e-01
-8.74381840e-01 -3.61202508e-01 -1.91059768e-01 -1.02956438e+00
1.42982471e+00 -9.31251884e-01 -1.11814249e+00 4.81230229e-01
-1.74183249e+00 -2.24512085e-01 9.99014676e-02 5.09861290e-01
-7.20975518e-01 6.09622836e-01 -7.89268672e-01 -9.59793389e-01
-3.80756885e-01 -1.09135008e+00 7.76028872e-01 6.24463499e-01
-1.62032157e-01 -7.93224454e-01 -2.13229790e-01 7.64727533e-01
5.13899267e-01 -7.57841095e-02 3.26650858e-01 -6.78873420e-01
-9.64663684e-01 8.31913650e-02 -4.85030085e-01 3.55644315e-01
-2.81161159e-01 -1.09297879e-01 -1.12941206e+00 -1.32760629e-01
3.66101027e-01 -4.95157957e-01 8.60578597e-01 3.97447944e-01
1.55762804e+00 -7.38541961e-01 1.12828873e-01 2.77293712e-01
8.46238971e-01 3.99442673e-01 1.00966620e+00 1.50827825e-01
6.00744724e-01 6.33868098e-01 8.68071258e-01 5.11449873e-01
8.55129480e-01 5.78336239e-01 6.77773535e-01 2.39214540e-01
-3.97450358e-01 -4.52457637e-01 6.19567335e-01 1.05790770e+00
7.89553672e-02 -9.64123011e-01 -4.95842606e-01 3.68341982e-01
-2.01328015e+00 -1.44625473e+00 1.34602711e-02 1.68009222e+00
7.20614851e-01 -1.27668649e-01 -2.38656420e-02 -2.89028227e-01
1.04486525e+00 4.52798069e-01 -7.76011467e-01 -2.57383674e-01
-2.50406325e-01 -7.61483371e-01 -9.92264301e-02 6.59034550e-01
-9.27064240e-01 7.45388508e-01 5.60154629e+00 9.05459046e-01
-1.00732875e+00 1.96608156e-01 8.33025098e-01 -2.16581687e-01
-2.02927679e-01 -2.77624816e-01 -5.22008061e-01 9.08122003e-01
1.30478704e+00 -5.51746130e-01 6.18635833e-01 5.55578113e-01
4.47797924e-01 -3.92375827e-01 -7.63000429e-01 1.19445121e+00
5.72936296e-01 -1.23454773e+00 2.31126919e-01 -1.42533526e-01
3.49149227e-01 -3.63698840e-01 3.39155227e-01 5.18730462e-01
-2.39656866e-01 -9.36777651e-01 6.84351563e-01 8.52000475e-01
8.62314582e-01 -5.82287908e-01 6.95017874e-01 4.58573967e-01
-1.17860889e+00 -7.80930370e-02 -2.25487947e-01 1.69307664e-01
6.92563295e-01 4.71348204e-02 -6.05744600e-01 4.05993372e-01
7.29331911e-01 7.20579445e-01 -5.27316093e-01 9.29062605e-01
-1.61289915e-01 5.96746087e-01 4.05679382e-02 -2.45481536e-01
8.02146420e-02 3.14852804e-01 7.30012715e-01 1.23469305e+00
2.68954843e-01 5.74386299e-01 -9.55986902e-02 5.18115580e-01
-2.91479737e-01 2.05514790e-03 -2.93619365e-01 -8.82023349e-02
5.61127543e-01 1.13377130e+00 -3.67303938e-01 -4.83560890e-01
-4.64739442e-01 1.33399272e+00 1.20528102e-01 4.41803515e-01
-1.15751135e+00 -2.99959123e-01 2.00479329e-01 -1.44447252e-01
2.11090297e-01 1.55280940e-02 1.95272908e-01 -1.35930216e+00
-1.20253928e-01 -8.78106296e-01 6.61376417e-01 -1.56012797e+00
-9.64326143e-01 8.53415251e-01 1.18983425e-01 -1.06672823e+00
-3.87316197e-01 -1.05857901e-01 -4.47269410e-01 6.70343995e-01
-1.43834555e+00 -7.57219851e-01 -6.73950732e-01 7.35474110e-01
1.13744283e+00 -5.26426993e-02 4.17552888e-01 4.48520869e-01
-4.99505430e-01 4.13377106e-01 -3.17390233e-01 -1.78488661e-02
1.01062334e+00 -8.42520654e-01 -6.10297062e-02 8.37205172e-01
-2.55046487e-01 6.06331378e-02 8.49351585e-01 -6.97406709e-01
-1.34269822e+00 -9.93965864e-01 8.75548005e-01 -5.62239349e-01
4.29448128e-01 -8.85607526e-02 -1.09111357e+00 4.89611864e-01
1.10690975e+00 -4.27387744e-01 6.10138595e-01 -6.99540496e-01
3.82921100e-02 1.61613286e-01 -1.10568237e+00 6.28353119e-01
8.84926915e-01 -7.59148717e-01 -7.71232784e-01 2.40543649e-01
1.35661089e+00 -4.52175379e-01 -5.41529834e-01 6.59897923e-02
2.22769871e-01 -9.90197241e-01 7.52817452e-01 -5.09781361e-01
9.78843451e-01 -6.82310313e-02 -2.14784667e-01 -1.09643924e+00
-9.47731733e-02 -7.93837130e-01 -7.21958876e-01 1.46725214e+00
1.84117869e-01 1.62781309e-02 2.75347054e-01 9.32329237e-01
-2.05498531e-01 -3.45498353e-01 -8.24328899e-01 -2.80684888e-01
-4.96838152e-01 -2.13399321e-01 5.10452330e-01 6.31346524e-01
2.93669850e-01 6.75551057e-01 -9.07730818e-01 -1.54023208e-02
1.50035903e-01 -1.48857996e-01 4.69474763e-01 -6.49527907e-01
-1.37652040e-01 -2.20360652e-01 -1.44557834e-01 -1.63422680e+00
1.71613917e-01 -5.64460397e-01 -9.99199133e-03 -1.74933755e+00
5.70402741e-01 7.19260871e-01 -1.62248105e-01 -1.03842560e-02
-4.17240113e-01 -6.42843992e-02 6.98583663e-01 1.12444766e-01
-1.53100979e+00 9.36961055e-01 1.63664567e+00 -1.54171199e-01
-1.10757984e-02 -3.14475209e-01 -6.55870557e-01 4.75437105e-01
5.18877804e-01 -5.70042655e-02 -5.05185068e-01 -5.10965049e-01
4.25911024e-02 7.65874505e-01 3.21595103e-01 -8.48553836e-01
6.64786339e-01 -2.75720298e-01 2.57767320e-01 -8.98413420e-01
7.12513983e-01 -8.42520714e-01 -7.05994070e-02 3.46181601e-01
-8.90605628e-01 1.45369157e-01 2.97902673e-01 7.08219707e-01
-4.81842428e-01 -1.32635916e-02 5.81535220e-01 -6.60583610e-03
-9.40228760e-01 5.15354097e-01 -6.77478790e-01 1.62733227e-01
1.09448266e+00 -4.68912572e-02 -6.64085746e-01 -1.34583795e+00
-8.12663674e-01 8.98251176e-01 1.49161994e-01 6.86286628e-01
1.00297201e+00 -1.49089420e+00 -7.87887692e-01 -2.79956758e-01
9.76686738e-03 -2.50574201e-01 1.16813898e+00 6.42025769e-01
-3.87238264e-01 5.92534721e-01 -1.90676287e-01 -4.81131196e-01
-1.30756450e+00 8.79960895e-01 1.24650612e-01 -3.47075500e-02
-3.12677264e-01 7.88003564e-01 4.57999557e-01 1.67380840e-01
4.34728205e-01 1.14580132e-01 -7.24048853e-01 2.78102249e-01
1.01254082e+00 4.15080220e-01 -5.35387516e-01 -7.55522788e-01
-1.15530461e-01 -2.13400424e-02 -1.25094101e-01 -3.06968957e-01
9.80685592e-01 -9.68273759e-01 1.09927855e-01 4.00904775e-01
1.24622655e+00 -3.63147169e-01 -1.62452793e+00 -2.02431068e-01
-4.36151624e-01 -4.86559629e-01 5.52175641e-02 -1.02515543e+00
-1.02079427e+00 1.08842897e+00 4.44438428e-01 4.86623555e-01
1.34933090e+00 1.38742626e-01 1.18502581e+00 2.36870438e-01
-1.82583064e-01 -1.21239674e+00 6.62599146e-01 7.20413148e-01
1.33989251e+00 -1.29365027e+00 -3.70260477e-01 -2.98015058e-01
-1.26662672e+00 1.05927920e+00 1.03399968e+00 3.23751748e-01
4.33214635e-01 -1.99962482e-01 -9.65383649e-02 6.22323565e-02
-1.30606735e+00 9.06606391e-02 3.64852637e-01 4.89276797e-01
2.23984290e-02 -4.40942526e-01 8.23233798e-02 7.98663676e-01
9.04438496e-02 7.21814795e-05 6.53636932e-01 6.17863655e-01
-6.48988724e-01 -3.41695964e-01 -3.09422642e-01 2.31186062e-01
-2.98984379e-01 -1.70148328e-01 -4.08853948e-01 2.83540338e-01
-4.64667976e-01 1.38844573e+00 2.60838449e-01 -5.79065681e-01
3.38695291e-03 9.83006880e-02 1.12664446e-01 -3.88008475e-01
-3.80246013e-01 -1.62026640e-02 1.27054125e-01 -6.75755262e-01
-6.20987892e-01 -5.14401436e-01 -1.11931026e+00 -5.87460995e-01
-1.57777652e-01 1.22493021e-01 4.03068781e-01 1.02227521e+00
7.24460304e-01 7.80239046e-01 7.31830955e-01 -6.31533325e-01
-1.50370821e-01 -1.10254717e+00 -1.31998494e-01 1.95130199e-01
3.72483283e-01 -3.83406520e-01 -5.04683197e-01 2.96738774e-01] | [10.610542297363281, 0.920678973197937] |
e58ebc3d-5218-4611-924b-995a16e1ba95 | toward-achieving-robust-low-level-and-high | null | null | https://ieeexplore.ieee.org/document/8517116 | https://ieeexplore.ieee.org/document/8517116 | Toward Achieving Robust Low-Level and High-Level Scene Parsing | In this paper, we address the challenging task of scene segmentation. We first discuss and compare two widely used approaches to retain detailed spatial information from pre-trained convolutional context network (CNN)-“dilation” and “skip”. Then, we demonstrate that the parsing performance of “skip” network can be noticeably improved by modifying the parameterization of skip layers. Furthermore, we introduce a “dense skip” architecture to retain a rich set of low-level information from the pre-trained CNN, which is essential to improve the low-level parsing performance. Meanwhile, we propose a CCN and place it on top of pre-trained CNNs, which is used to aggregate contexts for high-level feature maps so that robust high-level parsing can be achieved. We name our segmentation network enhanced fully convolutional network (EFCN) based on its significantly enhanced structure over FCN. Extensive experimental studies justify each contribution separately. Without bells and whistles, EFCN achieves state-of-the-arts on segmentation datasets of ADE20K, Pascal Context, SUN-RGBD, and Pascal VOC 2012. | ['Gang Wang', 'Henghui Ding', 'Xudong Jiang', 'Ting Liu', 'Bing Shuai'] | 2019-03-01 | null | null | null | journal-2019-3 | ['scene-parsing'] | ['computer-vision'] | [ 4.82745230e-01 1.76361710e-01 -2.44973488e-02 -6.86881840e-01
-5.55855870e-01 -7.41357327e-01 3.03640872e-01 -7.59691074e-02
-6.42068326e-01 4.23818588e-01 6.12347685e-02 -4.49799925e-01
2.44142458e-01 -8.00996423e-01 -1.01155007e+00 -6.43292487e-01
1.65604427e-01 -3.02322835e-01 4.65841830e-01 -7.27331862e-02
7.39987940e-02 4.30092961e-01 -1.33270502e+00 3.41444254e-01
9.05580401e-01 1.17590511e+00 5.11403739e-01 6.84391320e-01
-4.29868549e-01 7.29906082e-01 -5.09033799e-01 -2.80698210e-01
2.26814240e-01 -3.31177592e-01 -1.07237971e+00 -1.65677086e-01
7.22486794e-01 -4.38291252e-01 -2.73426920e-01 1.03625309e+00
2.08051533e-01 1.39208632e-02 1.36042997e-01 -6.33819163e-01
-4.59936708e-01 6.50394678e-01 -5.31614900e-01 1.68471772e-03
-8.60261843e-02 4.66572158e-02 9.93786037e-01 -6.54189527e-01
5.91366172e-01 1.07292628e+00 6.85123205e-01 6.35968447e-01
-9.71305966e-01 -5.19981027e-01 7.37365842e-01 -1.95789337e-02
-9.21363890e-01 3.06174513e-02 8.78283918e-01 -1.57906979e-01
8.53764415e-01 2.43486911e-01 5.94670713e-01 8.77269685e-01
-1.46461993e-01 1.14347315e+00 8.79826963e-01 -2.86312461e-01
1.78837702e-01 -5.63427508e-01 4.37165231e-01 7.33913660e-01
1.73253998e-01 -4.93108593e-02 -1.96805418e-01 4.15649354e-01
6.96393549e-01 4.29343097e-02 -1.17758714e-01 -1.47655383e-01
-1.07691920e+00 6.52045846e-01 1.06336856e+00 3.46726298e-01
-2.08991215e-01 2.46970698e-01 3.38450670e-01 -2.42479861e-01
2.71314889e-01 3.13425243e-01 -7.04067767e-01 1.15804844e-01
-1.07751191e+00 2.61672940e-02 5.33763647e-01 1.24198747e+00
8.80022883e-01 -1.68720111e-01 -3.21745306e-01 7.93917835e-01
3.19926557e-03 3.36312532e-01 1.33402094e-01 -1.09759474e+00
6.02263212e-01 7.69186616e-01 -2.61438280e-01 -6.66576028e-01
-5.58467805e-01 -5.42292416e-01 -8.32455635e-01 -1.20183319e-01
4.70110089e-01 -3.42869312e-01 -1.40190983e+00 1.63539255e+00
2.95994520e-01 3.94156516e-01 1.09192528e-01 8.38766038e-01
1.09491205e+00 6.05089486e-01 3.19931000e-01 3.31591070e-01
1.31651342e+00 -1.55184603e+00 -5.30140758e-01 -5.93110383e-01
5.78400552e-01 -5.49263060e-01 1.08409929e+00 5.87121472e-02
-9.23961520e-01 -8.11797023e-01 -1.03897703e+00 -4.90957916e-01
-5.68382859e-01 2.28471845e-01 8.94261122e-01 3.93250048e-01
-1.02731001e+00 8.22678983e-01 -9.74430561e-01 -2.44218677e-01
8.40678930e-01 3.22638035e-01 -2.20507666e-01 -3.68359625e-01
-8.76543164e-01 2.77093679e-01 6.00158513e-01 5.56593001e-01
-6.91006541e-01 -7.00747848e-01 -1.10654747e+00 -2.09642276e-02
5.59756756e-01 -4.07064855e-01 1.21811163e+00 -7.94789135e-01
-1.32620609e+00 7.43487477e-01 -1.68158814e-01 -3.39122295e-01
3.43900532e-01 -5.87032199e-01 8.73534754e-02 4.35619175e-01
1.47666365e-01 1.10163891e+00 3.52044195e-01 -1.29860198e+00
-9.67845976e-01 -3.03726852e-01 4.21505511e-01 3.35561186e-02
-7.47781396e-02 -1.59623727e-01 -1.04807460e+00 -7.03149259e-01
4.27026778e-01 -9.65357125e-01 -5.17109692e-01 -1.70135945e-01
-7.95168698e-01 -4.21670265e-02 1.14607882e+00 -7.42717266e-01
1.07801211e+00 -2.26953197e+00 1.11784153e-01 -2.05085292e-01
4.83717695e-02 5.61588943e-01 -3.48154873e-01 -1.10660912e-02
7.10305125e-02 3.40753883e-01 -7.34586358e-01 -6.99162662e-01
-1.80439115e-01 6.94630206e-01 -2.27670819e-01 1.18446767e-01
6.70643926e-01 1.29538321e+00 -8.07759643e-01 -5.48003674e-01
4.47446525e-01 5.13767719e-01 -6.07331336e-01 1.51691094e-01
-4.37111437e-01 8.40274334e-01 -5.09339094e-01 8.49948645e-01
9.46288407e-01 -1.84516340e-01 1.35210499e-01 -2.63029367e-01
-3.24322701e-01 2.96206445e-01 -8.12579334e-01 2.12145591e+00
-3.32701504e-01 4.67156202e-01 2.77153105e-01 -9.54424024e-01
6.84502304e-01 -3.33769828e-01 1.65075943e-01 -1.00556302e+00
1.17939077e-01 5.84250316e-02 -2.78923661e-01 -1.98986501e-01
6.93719506e-01 4.19226706e-01 -2.73069888e-01 -1.51260182e-01
1.87049862e-02 -1.36080254e-02 2.65587002e-01 1.48196727e-01
8.71053755e-01 4.65209633e-01 -9.20110717e-02 -3.61611545e-01
5.70318103e-01 1.54121136e-02 8.03062558e-01 8.14559519e-01
-3.09880286e-01 7.76626766e-01 7.97384202e-01 -3.68430585e-01
-7.23392427e-01 -8.52731168e-01 -2.29195163e-01 1.21187007e+00
4.05896991e-01 -2.35793591e-01 -1.22635198e+00 -9.32606876e-01
-2.47961715e-01 4.60224122e-01 -8.76171529e-01 1.83551311e-01
-1.10414803e+00 -8.16454113e-01 6.51204407e-01 1.17706740e+00
1.15457261e+00 -1.18284726e+00 -6.82758987e-01 2.61184633e-01
-1.35259181e-01 -1.58082438e+00 -2.33445019e-01 5.03614545e-01
-1.01939774e+00 -1.11759233e+00 -7.67379403e-01 -1.06417191e+00
5.61263263e-01 2.04991713e-01 1.19279242e+00 2.79478103e-01
-2.39316911e-01 -2.29553711e-02 -6.67668343e-01 -2.68460572e-01
1.44389123e-01 4.35522020e-01 -8.14863205e-01 -2.97200024e-01
2.63990968e-01 -4.13636625e-01 -8.74779165e-01 1.96991578e-01
-9.19690669e-01 2.46932179e-01 7.28506804e-01 6.95323706e-01
1.06109667e+00 -3.15250188e-01 2.50267595e-01 -1.07445264e+00
-1.16614588e-01 -8.29478800e-02 -6.63053155e-01 1.54929847e-01
-1.76425785e-01 -6.27583712e-02 6.68110013e-01 2.45807860e-02
-1.13512814e+00 4.71964717e-01 -5.00014186e-01 -4.19705398e-02
-4.66930926e-01 2.74990618e-01 -3.84149551e-01 -1.10725209e-01
-2.16312874e-02 4.96997125e-02 -6.04218841e-01 -7.75434494e-01
7.26317286e-01 3.35191071e-01 1.11912298e+00 -7.98243761e-01
4.53543723e-01 6.20157182e-01 7.36152083e-02 -6.93331659e-01
-1.33154190e+00 -4.05604661e-01 -1.13082802e+00 2.63020486e-01
1.42255914e+00 -9.87637699e-01 -4.65714037e-01 8.59494150e-01
-1.22387612e+00 -9.18910742e-01 -2.27136374e-01 8.28671735e-03
-2.76843429e-01 2.75062710e-01 -8.73315394e-01 -3.49133402e-01
-3.17523271e-01 -1.23852909e+00 1.28237367e+00 7.22424865e-01
4.43712920e-01 -9.01392519e-01 -3.11001241e-01 2.60729194e-01
2.86349654e-01 6.63924932e-01 8.20006371e-01 -4.38377023e-01
-7.76235342e-01 1.06368341e-01 -7.17976153e-01 5.83264470e-01
-1.65459402e-02 5.37242107e-02 -1.20150912e+00 -1.06816478e-02
-2.75800258e-01 -9.85867381e-02 1.41117728e+00 5.28387129e-01
1.71834159e+00 1.25216931e-01 -4.06239569e-01 1.25959349e+00
1.52681446e+00 2.94251051e-02 7.33734131e-01 3.27132314e-01
1.21485019e+00 6.34937644e-01 5.54289401e-01 7.11702369e-03
6.60654128e-01 3.53806764e-01 7.99231410e-01 -6.43979788e-01
-3.95408571e-01 -2.63483196e-01 -2.03667253e-01 7.69249797e-01
-6.00100611e-04 -8.19913372e-02 -8.39421749e-01 7.45540023e-01
-1.75002265e+00 -3.77168000e-01 -3.46210986e-01 1.58615005e+00
7.11419404e-01 1.84506372e-01 -9.77987945e-02 -1.57562122e-01
7.75642633e-01 4.03548062e-01 -7.42314339e-01 -4.20500517e-01
-3.61100435e-01 5.60993254e-01 7.44497657e-01 4.57690984e-01
-1.61195290e+00 1.45221817e+00 5.81871319e+00 6.14771962e-01
-1.13778794e+00 -1.05698600e-01 7.99600542e-01 1.99202254e-01
-2.02958733e-02 2.27474850e-02 -9.12919581e-01 4.43772048e-01
6.34361327e-01 5.76801300e-01 1.18259996e-01 9.34325516e-01
-7.86715075e-02 -3.35222006e-01 -1.04126060e+00 6.18548512e-01
-3.60016495e-01 -1.31493998e+00 -1.51767686e-01 -2.20948830e-02
9.08621728e-01 2.95569271e-01 -7.25898668e-02 4.36413616e-01
3.37638557e-01 -1.01651263e+00 8.15290332e-01 2.29312018e-01
7.70936668e-01 -7.94849157e-01 8.64351511e-01 9.74476412e-02
-1.60643530e+00 -2.05857933e-01 -4.99494642e-01 6.63817972e-02
1.58565536e-01 7.19915867e-01 -2.81789750e-02 8.38675201e-01
9.41461205e-01 8.12176168e-01 -8.41757953e-01 8.56509387e-01
-7.00490415e-01 8.32309663e-01 -3.36011857e-01 1.89901441e-01
7.99822032e-01 -1.05482176e-01 -5.09354584e-02 1.64251685e+00
-4.22149189e-02 3.97699624e-01 2.22033754e-01 8.84031475e-01
-3.61681163e-01 -1.75475955e-01 -1.16876177e-01 1.84322372e-01
3.20504457e-01 1.41514242e+00 -1.13183892e+00 -4.80470210e-01
-4.65258211e-01 1.12719202e+00 3.89347851e-01 3.97688001e-01
-9.08043087e-01 -6.01318359e-01 6.52348459e-01 -3.77874672e-01
7.69140005e-01 -3.59602988e-01 -6.93894804e-01 -1.04130721e+00
1.40111327e-01 -3.90540987e-01 2.77191073e-01 -5.08229554e-01
-8.94253731e-01 7.06849456e-01 -5.14908992e-02 -5.96835494e-01
2.60697573e-01 -8.82709980e-01 -7.91203320e-01 8.79370689e-01
-2.14401770e+00 -1.39342141e+00 -5.81285119e-01 5.16466439e-01
5.57575643e-01 5.69403648e-01 4.68768984e-01 3.20206165e-01
-9.91989374e-01 5.41611731e-01 -4.71135601e-02 8.41795087e-01
4.41392213e-01 -1.56583166e+00 8.34019303e-01 1.16892004e+00
-1.36613965e-01 5.46689391e-01 1.50408214e-02 -5.92291355e-01
-1.07280684e+00 -1.67194510e+00 6.31531835e-01 -2.24932507e-01
2.73671865e-01 -5.78854382e-01 -8.29031825e-01 7.45735288e-01
1.14287771e-01 4.57037866e-01 3.09624106e-01 -9.85418353e-03
-4.64782685e-01 5.26291830e-03 -1.01762295e+00 5.21722734e-01
1.31434548e+00 -3.93331409e-01 -5.44063270e-01 1.40802264e-01
1.34563780e+00 -7.61178076e-01 -6.41077697e-01 6.61805391e-01
3.45324963e-01 -1.16145694e+00 1.02353752e+00 -3.67848635e-01
6.86305046e-01 -2.87175268e-01 -3.71845961e-01 -8.77076924e-01
-8.65068734e-02 -3.36182237e-01 6.33548722e-02 1.35467994e+00
3.46568555e-01 -4.34705704e-01 9.04762626e-01 4.46407616e-01
-5.65236747e-01 -8.98896754e-01 -6.68000579e-01 -6.98492587e-01
5.37640572e-01 -5.66533625e-01 7.64886856e-01 7.01916754e-01
-6.40488088e-01 1.09648377e-01 3.69197503e-02 3.08476269e-01
5.85044563e-01 3.12936038e-01 6.13589406e-01 -9.70482051e-01
4.46619913e-02 -3.33650678e-01 -8.67723003e-02 -1.55952930e+00
1.12315826e-01 -7.40181088e-01 3.14815491e-01 -1.67445958e+00
7.94416144e-02 -4.39629823e-01 -3.50099713e-01 7.52057076e-01
-4.92014021e-01 4.55231249e-01 5.10996997e-01 -1.03479341e-01
-8.99698496e-01 2.89706230e-01 1.49583972e+00 4.70569059e-02
-2.37822458e-01 -2.39348263e-01 -8.86181533e-01 8.71870697e-01
7.31108010e-01 -3.06868345e-01 -1.86377108e-01 -8.91697526e-01
-2.17652246e-01 -2.17547640e-01 6.17365718e-01 -1.09840572e+00
1.04268692e-01 -4.61902060e-02 4.84802544e-01 -9.82794940e-01
-1.38156032e-02 -6.39890552e-01 -5.71885049e-01 2.84982145e-01
-2.94932663e-01 -1.80193529e-01 5.70418239e-01 4.85567451e-01
-3.42036545e-01 -1.27717733e-01 8.61622095e-01 -2.27201506e-01
-1.18449652e+00 3.83269221e-01 1.62690908e-01 1.96140245e-01
8.42109501e-01 -9.65122953e-02 -3.82037491e-01 2.00442687e-01
-7.72356570e-01 4.86963123e-01 4.24503714e-01 2.86770403e-01
2.88978338e-01 -1.02706420e+00 -3.42414051e-01 8.73476043e-02
-4.98109609e-02 1.00952542e+00 4.29183722e-01 5.17127991e-01
-8.14409792e-01 4.93781477e-01 8.96339957e-03 -8.16524327e-01
-9.23738420e-01 4.29939717e-01 1.62035882e-01 -3.27689886e-01
-7.51960516e-01 1.17976892e+00 4.88280654e-01 -7.08015382e-01
3.20530355e-01 -9.92469192e-01 -5.73440865e-02 -2.05648229e-01
4.88282144e-01 -8.60137306e-03 -2.54497351e-03 -5.31720936e-01
-5.09439766e-01 8.61120641e-01 5.06783370e-03 3.21038246e-01
1.46441209e+00 -2.05360144e-01 -2.00305492e-01 -7.23251179e-02
1.20359147e+00 -8.56683031e-02 -1.97787070e+00 -2.18903005e-01
5.89330904e-02 -1.43589675e-01 -4.22025472e-02 -1.03728068e+00
-1.40098786e+00 1.23705697e+00 3.85246903e-01 -8.19509849e-02
1.44970274e+00 3.73357646e-02 1.09108794e+00 4.03342664e-01
1.17824338e-01 -1.09842849e+00 -6.20255098e-02 7.91209221e-01
5.31811357e-01 -1.21207392e+00 -1.25852004e-01 -7.63956785e-01
-3.98273677e-01 1.10717392e+00 8.10839057e-01 -3.16442162e-01
5.46408176e-01 4.86604720e-01 1.33523822e-01 -6.83556357e-03
-1.23183273e-01 -5.96998811e-01 2.38287285e-01 6.31664693e-01
2.91789085e-01 9.71242413e-02 -6.35288581e-02 9.16313946e-01
-1.56926200e-01 -1.87193230e-01 2.66533822e-01 1.03839064e+00
-4.67533469e-01 -1.01050556e+00 -9.42531601e-02 2.37892400e-02
-5.73785365e-01 -3.20971519e-01 -4.36206341e-01 1.04922247e+00
4.87531185e-01 9.47051048e-01 1.99178398e-01 -2.97705293e-01
4.94752556e-01 -7.35592097e-02 2.54385561e-01 -6.31117463e-01
-6.51976347e-01 8.80653858e-02 -1.39261976e-01 -9.73352551e-01
-5.54201841e-01 -4.70339775e-01 -1.46823359e+00 -2.19456907e-02
-2.20951200e-01 -2.61951357e-01 8.53061199e-01 1.07972276e+00
3.40226531e-01 9.52262819e-01 2.58560449e-01 -9.51247573e-01
-3.59817874e-03 -7.14764833e-01 -2.75900871e-01 3.24161321e-01
3.72526407e-01 -2.22535163e-01 -2.46400516e-02 9.36166346e-02] | [9.545391082763672, 0.2804185152053833] |
d879393f-5f7b-4540-a3f4-b7852877682b | encoding-program-as-image-evaluating-visual | 2111.01097 | null | https://arxiv.org/abs/2111.01097v3 | https://arxiv.org/pdf/2111.01097v3.pdf | Code2Snapshot: Using Code Snapshots for Learning Representations of Source Code | There are several approaches for encoding source code in the input vectors of neural models. These approaches attempt to include various syntactic and semantic features of input programs in their encoding. In this paper, we investigate Code2Snapshot, a novel representation of the source code that is based on the snapshots of input programs. We evaluate several variations of this representation and compare its performance with state-of-the-art representations that utilize the rich syntactic and semantic features of input programs. Our preliminary study on the utility of Code2Snapshot in the code summarization and code classification tasks suggests that simple snapshots of input programs have comparable performance to state-of-the-art representations. Interestingly, obscuring input programs have insignificant impacts on the Code2Snapshot performance, suggesting that, for some tasks, neural models may provide high performance by relying merely on the structure of input programs. | ['Mohammad Amin Alipour', 'Md Rafiqul Islam Rabin'] | 2021-11-01 | null | null | null | null | ['code-classification', 'method-name-prediction'] | ['computer-code', 'natural-language-processing'] | [ 1.75355449e-01 1.29968703e-01 -3.83888930e-01 -5.53237438e-01
-4.45905715e-01 -5.44461787e-01 5.48716068e-01 5.75037837e-01
-2.28753075e-01 1.25633568e-01 5.48076928e-01 -5.46186030e-01
1.23592913e-01 -7.74486601e-01 -8.92944694e-01 -1.15509160e-01
2.73767877e-02 -2.95831561e-01 2.91348577e-01 -3.86593342e-01
8.22207749e-01 1.25571951e-01 -1.90194321e+00 1.07592154e+00
7.48989403e-01 5.68384767e-01 3.24274093e-01 8.49526048e-01
-8.62393498e-01 1.21635580e+00 -1.03372025e+00 -4.02555108e-01
-9.69532654e-02 -5.02623975e-01 -7.37102211e-01 -4.87631768e-01
6.40963376e-01 -1.08119726e-01 -4.92372394e-01 1.15425110e+00
5.84893823e-02 -4.83643450e-02 4.84694451e-01 -9.63755250e-01
-1.14693463e+00 8.87421191e-01 -2.23002076e-01 4.38013643e-01
6.12283707e-01 -1.71984185e-03 9.99838173e-01 -9.55517948e-01
6.10048354e-01 9.61651325e-01 1.00368452e+00 3.64755362e-01
-1.42425966e+00 -1.91914633e-01 9.77733508e-02 3.08812596e-02
-9.57835674e-01 -4.33087111e-01 5.93702137e-01 -8.37105572e-01
1.72947061e+00 3.56065035e-01 4.24178660e-01 7.86763906e-01
4.82567519e-01 8.12798381e-01 4.41109657e-01 -4.97271150e-01
8.94299448e-02 2.87991643e-01 9.57894981e-01 1.12628222e+00
7.04730213e-01 -2.11785495e-01 -3.25836301e-01 -5.77215374e-01
2.82759726e-01 2.23443836e-01 -2.51991749e-01 -3.38547021e-01
-1.00526524e+00 9.39071894e-01 6.84690297e-01 5.46116114e-01
-1.67252526e-01 5.52606285e-01 9.04490113e-01 3.03911477e-01
3.85858595e-01 8.74208272e-01 -2.63180494e-01 -3.32190245e-01
-1.04955304e+00 1.96892411e-01 6.72747135e-01 1.16446912e+00
9.56547260e-01 3.97673965e-01 -3.71548682e-01 8.46843421e-01
-4.86822501e-02 1.20928183e-01 8.26681077e-01 -7.99936533e-01
7.90212572e-01 1.19470286e+00 -3.40102017e-01 -1.12677157e+00
-1.74941033e-01 -1.67522773e-01 -3.56046557e-01 1.40707687e-01
4.96948184e-03 9.79389995e-02 -1.06274223e+00 1.29788041e+00
-6.33565962e-01 -6.40658200e-01 1.29613638e-01 1.86210886e-01
1.21516621e+00 7.31117845e-01 -1.15277849e-01 3.89958799e-01
9.52941358e-01 -1.19496274e+00 -4.03861046e-01 -3.63738835e-01
9.59209561e-01 -4.80057627e-01 1.03031027e+00 5.91233931e-02
-9.12867367e-01 -8.25660884e-01 -1.14980733e+00 -1.07992023e-01
-6.08390093e-01 6.43431842e-01 8.69265437e-01 7.32171834e-01
-1.18726671e+00 8.38893175e-01 -7.85706282e-01 -5.68130314e-01
4.10797715e-01 2.43621141e-01 -3.95164996e-01 -1.09029703e-01
-4.26756769e-01 8.97511661e-01 5.56528211e-01 -5.74586093e-01
-7.62080014e-01 -5.18344820e-01 -1.25123572e+00 4.43245441e-01
3.86977661e-03 -3.31912726e-01 1.35941315e+00 -1.27598917e+00
-8.70976150e-01 6.67193532e-01 -4.39235479e-01 -3.88686955e-01
-2.37029105e-01 -1.64174065e-01 -5.68186715e-02 -2.05172561e-02
-1.51430115e-01 2.32338786e-01 4.84599948e-01 -1.10795963e+00
-3.43365669e-01 -7.68950880e-02 4.39180493e-01 -1.86540946e-01
-4.77265984e-01 2.95498401e-01 -3.12849045e-01 -6.92946076e-01
-9.85353962e-02 -7.25415528e-01 -1.37545884e-01 -1.22871689e-01
-3.25184196e-01 -6.46179318e-02 4.99662519e-01 -7.89415002e-01
1.68382573e+00 -2.43070698e+00 2.56081492e-01 -5.49408868e-02
1.43458188e-01 4.13796246e-01 -4.22212422e-01 7.35392988e-01
-2.71832496e-01 6.04176998e-01 -3.56941134e-01 -1.61011606e-01
-1.38402656e-01 2.67417341e-01 -4.15622890e-01 1.93407699e-01
2.40328327e-01 9.08468723e-01 -7.23512471e-01 -1.38121009e-01
-1.86861917e-01 1.69787139e-01 -9.28628147e-01 2.85505414e-01
-2.09306359e-01 -4.70190585e-01 -3.99623692e-01 5.68474948e-01
2.74798632e-01 -1.04844041e-01 1.26284612e-02 1.77401483e-01
-1.41828522e-01 6.38416648e-01 -4.08375770e-01 1.80582988e+00
-4.32194144e-01 1.15061069e+00 -3.21808130e-01 -1.02205431e+00
9.07796800e-01 2.25332052e-01 8.14043358e-02 -4.57841337e-01
-1.40160233e-01 1.28884703e-01 1.39499262e-01 -7.52368212e-01
9.08827364e-01 4.66000408e-01 -3.99927557e-01 6.03752732e-01
4.70770836e-01 -6.62589744e-02 4.32264179e-01 4.06358480e-01
1.52536726e+00 2.78104365e-01 6.09825313e-01 -3.18874329e-01
2.97691703e-01 3.77523094e-01 1.88457191e-01 1.00107372e+00
2.39644259e-01 5.98235726e-01 8.47040474e-01 -6.61923766e-01
-1.13300955e+00 -6.94602549e-01 2.63595462e-01 1.30418479e+00
-3.41706663e-01 -1.08994508e+00 -9.42349613e-01 -8.01243186e-01
1.22930631e-01 9.40633893e-01 -9.27564621e-01 -4.09363747e-01
-6.13691807e-01 -7.61339486e-01 9.50064600e-01 8.22143197e-01
4.72822562e-02 -8.63754988e-01 -1.10418344e+00 2.01049805e-01
1.21375605e-01 -4.11648780e-01 -8.71750116e-02 5.93578041e-01
-1.10422611e+00 -1.19496381e+00 -5.32566309e-01 -6.54423058e-01
1.01414108e+00 3.94149482e-01 1.26114953e+00 6.58918083e-01
-1.12932503e-01 3.71946573e-01 -6.41414881e-01 -3.58201563e-01
-9.45440888e-01 2.27162689e-01 -5.52102506e-01 -6.51420414e-01
4.30978835e-01 -2.92063892e-01 3.11104134e-02 -2.94117153e-01
-9.30742443e-01 8.88958350e-02 6.42288268e-01 6.48094177e-01
-5.06051518e-02 -1.31469801e-01 1.71523571e-01 -1.31101298e+00
7.80065894e-01 -8.01282227e-01 -2.04033881e-01 2.68689036e-01
-4.64857429e-01 4.49067533e-01 9.18276846e-01 -3.53377998e-01
-9.20285523e-01 4.03728411e-02 -1.79752246e-01 -2.58233905e-01
-1.47696570e-01 8.74565721e-01 3.15146118e-01 -1.52901977e-01
1.06239951e+00 4.22392547e-01 -7.19680786e-02 -7.23485887e-01
2.04145610e-01 6.31762981e-01 3.71027440e-01 -6.07762992e-01
4.24709529e-01 1.63819210e-03 -5.65392971e-01 -6.31564319e-01
-3.23566049e-01 -3.72386366e-01 -6.10987544e-01 1.93337351e-01
5.24320185e-01 -7.26186872e-01 2.24312216e-01 3.29554707e-01
-1.50152326e+00 -1.07104100e-01 -2.08577573e-01 1.14536196e-01
-4.73352283e-01 2.13418573e-01 -6.04017794e-01 -3.84859502e-01
-1.09397613e-01 -1.40957844e+00 7.30762243e-01 5.46169654e-02
-4.96264845e-01 -6.78909123e-01 6.22382127e-02 -2.04656005e-01
7.28519201e-01 2.10869879e-01 1.63245344e+00 -1.06941545e+00
-5.41016579e-01 -2.32718751e-01 -6.25216365e-02 3.65741074e-01
1.89075172e-01 3.48695129e-01 -1.06120265e+00 -1.61229312e-01
-1.64597780e-01 -9.00233313e-02 1.24217236e+00 -2.66450997e-02
1.31714094e+00 -6.51862562e-01 -4.20353711e-01 6.32994235e-01
1.78689778e+00 4.71413076e-01 6.87025189e-01 3.03161651e-01
6.78900242e-01 5.14176250e-01 6.85002729e-02 4.33308542e-01
2.27047697e-01 3.48356783e-01 5.67632258e-01 1.83547840e-01
-1.69482633e-01 -3.02766979e-01 7.26153433e-01 1.08074093e+00
9.79376808e-02 -1.40750930e-01 -1.11707616e+00 9.21637714e-01
-1.66490364e+00 -1.10343134e+00 -3.16770107e-01 1.91935956e+00
6.16771400e-01 3.78353558e-02 -2.05881894e-01 -2.44748350e-02
5.95197320e-01 4.21723127e-01 -2.90822446e-01 -1.17681968e+00
2.01818019e-01 1.16846748e-01 4.16563660e-01 -8.33290443e-02
-7.26462483e-01 5.25545418e-01 7.53265238e+00 4.30232197e-01
-1.12660074e+00 1.14549018e-01 1.62340105e-01 -4.13475297e-02
-6.47583842e-01 1.26424238e-01 -5.16622603e-01 5.77516377e-01
1.26793885e+00 -5.03406048e-01 4.97841984e-01 1.36524451e+00
-3.92269015e-01 -3.96854013e-01 -1.55151951e+00 5.92950106e-01
5.26830256e-01 -1.64551842e+00 1.38381407e-01 -1.86644778e-01
8.25163245e-01 3.36379975e-01 -1.74718603e-01 7.24474728e-01
3.78789574e-01 -8.31677377e-01 8.65944505e-01 4.91848856e-01
7.35014558e-01 -5.98052800e-01 7.99554050e-01 3.13366950e-01
-1.14603138e+00 -4.54687566e-01 -5.72659612e-01 -1.27303794e-01
-4.38944429e-01 2.39812657e-01 -7.62368858e-01 4.12071586e-01
5.77305675e-01 7.05778241e-01 -1.49526143e+00 1.05904031e+00
-1.94785390e-02 5.00329077e-01 1.51562870e-01 -2.59334832e-01
3.19839180e-01 4.06408250e-01 3.02444875e-01 1.82694340e+00
6.02805734e-01 -2.34348431e-01 -7.05332160e-02 1.23437750e+00
-1.18194163e-01 1.46185281e-02 -1.32727802e+00 -5.24850547e-01
3.08551520e-01 6.70666456e-01 -5.64185381e-01 -8.22788477e-01
-8.08883429e-01 7.27980375e-01 5.00145078e-01 3.02522242e-01
-6.44251704e-01 -1.04081690e+00 8.28753531e-01 -5.99097237e-02
4.92853016e-01 -6.49612099e-02 -3.98444861e-01 -1.23793185e+00
6.64713383e-02 -1.04969049e+00 2.08816335e-01 -9.79053855e-01
-5.58281898e-01 8.29400718e-01 2.21494049e-01 -1.03336608e+00
-3.85960668e-01 -7.28108585e-01 -1.03191066e+00 6.79759026e-01
-1.04297650e+00 -6.38279438e-01 -3.40209603e-01 7.81994164e-02
8.24386239e-01 -5.06951630e-01 9.62115705e-01 -3.39018628e-02
-3.31324279e-01 6.02493823e-01 4.33926791e-01 3.92441899e-01
1.98399141e-01 -1.26065648e+00 8.77048612e-01 1.11120045e+00
2.71894485e-01 1.31480300e+00 6.08356595e-01 -5.75221658e-01
-1.49585927e+00 -1.11499894e+00 7.64564276e-01 -6.13749564e-01
2.58827388e-01 -2.28661180e-01 -1.14904833e+00 9.74049091e-01
5.31641185e-01 -1.07824251e-01 7.65641451e-01 2.33545192e-02
-7.79793620e-01 3.56339872e-01 -9.71759439e-01 2.86948591e-01
8.78306210e-01 -8.99807215e-01 -1.17117393e+00 -8.27162936e-02
7.86276460e-01 -2.66821712e-01 -5.80993176e-01 1.19472608e-01
3.98914129e-01 -1.19782782e+00 7.27292120e-01 -7.42646277e-01
9.96426702e-01 -1.37061045e-01 -3.85647714e-01 -1.62888050e+00
-4.50591773e-01 3.96494754e-02 -2.86177471e-02 1.19010329e+00
4.95246977e-01 -4.03522968e-01 4.48049903e-01 4.35947806e-01
-5.41584849e-01 -6.15962148e-01 -3.68465453e-01 -5.37628174e-01
1.54897854e-01 -2.85839736e-01 6.24152303e-01 8.40821326e-01
4.23323900e-01 3.63752320e-02 1.94274217e-01 -2.71568269e-01
4.47084457e-02 2.80486792e-01 7.77284622e-01 -1.05800903e+00
-4.24284607e-01 -5.46793163e-01 -6.51175857e-01 -6.53807759e-01
3.00961792e-01 -1.34064150e+00 3.30610238e-02 -1.78887391e+00
4.90527213e-01 -9.55838859e-02 -2.58794010e-01 8.19007814e-01
-2.96191573e-02 -1.65872619e-01 3.83663386e-01 2.98438698e-01
-4.07466859e-01 2.22293153e-01 3.71738464e-01 -4.54768270e-01
9.21012163e-02 -3.00508171e-01 -8.78502488e-01 8.20630550e-01
7.74322689e-01 -8.73112977e-01 -4.39977407e-01 -9.94149625e-01
2.54433036e-01 -8.50474313e-02 1.36231884e-01 -1.21666968e+00
1.08802013e-01 5.55411577e-02 2.66370803e-01 -4.12955463e-01
8.52364153e-02 -5.50563514e-01 1.91871986e-01 7.33558655e-01
-6.00753248e-01 6.50102437e-01 5.75888157e-01 3.79357666e-01
-4.09772217e-01 -1.02797377e+00 5.61500788e-01 -7.75227249e-01
-9.58047748e-01 -4.58207637e-01 -9.74755108e-01 -3.34989130e-02
7.22742736e-01 -4.01657581e-01 -7.71259427e-01 8.51506963e-02
-7.50994757e-02 -4.61956412e-01 9.01849270e-01 7.04248667e-01
6.71697021e-01 -1.17879212e+00 -2.95104265e-01 4.62569088e-01
6.02720618e-01 -4.99682218e-01 -3.05463314e-01 2.91409582e-01
-8.39971483e-01 7.39732027e-01 -5.58091819e-01 -3.58442068e-01
-1.40134132e+00 7.84691870e-01 2.04083309e-01 -4.50027660e-02
-4.60899174e-01 7.15634108e-01 1.80523276e-01 -5.28606355e-01
1.20409846e-01 -7.96881855e-01 -3.34591717e-01 5.68130426e-03
6.35740101e-01 1.68412477e-01 2.89121896e-01 -4.97482687e-01
-3.42140734e-01 3.21132928e-01 -1.69089243e-01 3.05668682e-01
1.51746857e+00 5.02775192e-01 -3.17245156e-01 6.89601421e-01
1.43848050e+00 6.57046884e-02 -7.64611125e-01 -1.70783028e-01
3.44987303e-01 -6.84314787e-01 -1.83246225e-01 -6.26314223e-01
-9.12882745e-01 1.08316600e+00 2.47865662e-01 1.12040199e-01
8.56744766e-01 -1.57330155e-01 1.94201052e-01 8.65588486e-01
7.13545978e-01 -7.33881056e-01 8.63236934e-02 8.18360925e-01
8.20891619e-01 -8.67273927e-01 1.86655328e-01 -4.90202643e-02
-5.04783690e-01 1.45447373e+00 7.54531145e-01 -3.29673111e-01
1.93508342e-01 4.08958614e-01 -2.12708548e-01 -3.26153368e-01
-1.06463301e+00 9.49401110e-02 2.06760496e-01 5.33442378e-01
8.59014452e-01 -2.34599054e-01 -1.44398376e-01 5.58688879e-01
5.08495141e-03 -1.44029528e-01 1.09798086e+00 1.42728078e+00
-6.42284870e-01 -1.00705850e+00 -3.20652723e-01 8.57587934e-01
-6.27833545e-01 -4.40925598e-01 -6.04363441e-01 8.01260889e-01
2.87593398e-02 5.93018174e-01 1.18738048e-01 -6.74626648e-01
4.40814674e-01 4.14233118e-01 2.72419870e-01 -1.32749009e+00
-1.30086565e+00 -5.72304308e-01 1.64591357e-01 -6.44282222e-01
-1.47697926e-01 -3.80420327e-01 -1.19339025e+00 -2.34381557e-01
-2.82792479e-01 1.92830488e-01 9.17160273e-01 4.66054171e-01
7.24884212e-01 8.29085290e-01 1.52761593e-01 -1.02435613e+00
-5.28512299e-01 -8.80135000e-01 -5.38709722e-02 3.22651178e-01
6.48925245e-01 -4.49933946e-01 -1.61789998e-01 3.00388008e-01] | [7.607367515563965, 7.897158145904541] |
7dd9cd7e-468b-4f67-9a83-f8293a4de0f4 | preliminary-study-on-using-vector | 2106.13479 | null | https://arxiv.org/abs/2106.13479v1 | https://arxiv.org/pdf/2106.13479v1.pdf | Preliminary study on using vector quantization latent spaces for TTS/VC systems with consistent performance | Generally speaking, the main objective when training a neural speech synthesis system is to synthesize natural and expressive speech from the output layer of the neural network without much attention given to the hidden layers. However, by learning useful latent representation, the system can be used for many more practical scenarios. In this paper, we investigate the use of quantized vectors to model the latent linguistic embedding and compare it with the continuous counterpart. By enforcing different policies over the latent spaces in the training, we are able to obtain a latent linguistic embedding that takes on different properties while having a similar performance in terms of quality and speaker similarity. Our experiments show that the voice cloning system built with vector quantization has only a small degradation in terms of perceptive evaluations, but has a discrete latent space that is useful for reducing the representation bit-rate, which is desirable for data transferring, or limiting the information leaking, which is important for speaker anonymization and other tasks of that nature. | ['Junichi Yamagishi', 'Hieu-Thi Luong'] | 2021-06-25 | null | null | null | null | ['voice-cloning'] | ['speech'] | [ 2.37132147e-01 5.47401369e-01 -2.08040431e-01 -4.55941945e-01
-4.19911623e-01 -4.38845724e-01 6.19896114e-01 -1.72757953e-02
-3.47915053e-01 7.34754086e-01 4.49569851e-01 -2.52346903e-01
1.37707889e-01 -8.11245739e-01 -6.50839269e-01 -1.00428998e+00
1.27459764e-01 1.37050822e-01 2.07021143e-02 -5.96538782e-02
-1.95526779e-01 7.96430588e-01 -1.67303479e+00 2.22596988e-01
5.82341015e-01 9.85289872e-01 -8.97443593e-02 4.36034828e-01
-2.34060004e-01 6.70071065e-01 -1.02142501e+00 -3.50664020e-01
2.29818299e-01 -6.43983960e-01 -7.27064312e-01 2.95987502e-02
2.48397186e-01 -1.71072289e-01 -2.03597799e-01 1.31567085e+00
6.42110527e-01 1.18132055e-01 6.47450864e-01 -1.25451016e+00
-5.98598480e-01 9.15497482e-01 3.50078672e-01 -3.27061087e-01
2.24182606e-01 -3.02663743e-02 9.80837762e-01 -2.86283255e-01
3.89489740e-01 1.51317036e+00 3.02283913e-01 9.85025823e-01
-1.53683889e+00 -6.48925722e-01 2.69486066e-02 -1.63993567e-01
-1.36580408e+00 -1.07186949e+00 7.81141281e-01 -1.84786066e-01
5.19243479e-01 6.34906828e-01 2.10983291e-01 1.23759401e+00
-1.70402393e-01 6.34512246e-01 8.25901806e-01 -5.74678481e-01
5.33018887e-01 8.36984336e-01 -1.80295363e-01 5.57568312e-01
-6.37977496e-02 -1.24069722e-02 -4.63580519e-01 -2.67253369e-01
4.85287994e-01 -2.39675835e-01 -6.54331744e-01 -4.79511529e-01
-9.60516453e-01 8.75954151e-01 2.40955591e-01 6.59330785e-01
-2.34643623e-01 2.07296863e-01 4.91110712e-01 7.47951031e-01
3.21822852e-01 5.07349312e-01 -2.53814948e-03 -9.47655216e-02
-1.04601538e+00 -1.44082293e-01 8.84282112e-01 8.75910461e-01
5.07620037e-01 5.59628785e-01 -2.73483843e-01 8.23799551e-01
2.22801194e-01 3.50780010e-01 1.00239718e+00 -9.36112404e-01
3.22026640e-01 3.08977574e-01 -4.76880185e-02 -8.12964141e-01
7.91151002e-02 -4.60431337e-01 -1.04245377e+00 3.58831406e-01
1.74416974e-01 -1.18038394e-01 -6.05929017e-01 1.97075236e+00
-5.28209023e-02 -1.50227025e-01 5.14409781e-01 7.48521805e-01
4.91048336e-01 9.53684151e-01 -2.22440735e-01 -3.90525013e-01
1.11203647e+00 -8.49670410e-01 -1.06190813e+00 6.90906569e-02
5.87388694e-01 -6.64781272e-01 1.41134989e+00 2.13032857e-01
-1.05771899e+00 -5.66312790e-01 -1.23185503e+00 -1.01591326e-01
-3.70110035e-01 1.86308026e-01 1.91547558e-01 1.12988079e+00
-1.32137430e+00 5.93269229e-01 -5.95212877e-01 -2.66586095e-01
-3.18943374e-02 5.13160944e-01 -4.38957959e-01 3.94825578e-01
-1.40466070e+00 5.70312142e-01 4.72123951e-01 -2.72496194e-02
-5.30077517e-01 -2.38534376e-01 -7.87458599e-01 5.39183140e-01
1.71854645e-02 -4.35165644e-01 8.88323784e-01 -1.13744247e+00
-2.21729922e+00 4.95979548e-01 -1.18179537e-01 -6.33833826e-01
6.87020302e-01 1.43848673e-01 -4.50875849e-01 1.34730963e-02
-3.18194956e-01 7.76376784e-01 1.33285248e+00 -1.05163622e+00
-3.35653752e-01 -7.31620193e-02 -3.17392237e-02 2.78365556e-02
-9.42539871e-01 -1.28831908e-01 -1.56382397e-01 -8.14064503e-01
-6.99715689e-02 -9.48568344e-01 6.55107796e-02 1.51361331e-01
-4.21118826e-01 -1.33887544e-01 8.64100158e-01 -5.40955007e-01
9.69919980e-01 -2.66584754e+00 4.46861982e-01 1.84495643e-01
-1.63363054e-01 3.96843076e-01 -1.09968754e-02 3.52423191e-01
-1.58022031e-01 3.93015265e-01 -3.08489770e-01 -8.39310467e-01
1.89470753e-01 3.55233669e-01 -8.51286113e-01 5.40988505e-01
-6.68713078e-02 6.01408005e-01 -4.95939255e-01 -2.23285019e-01
1.02032669e-01 6.14776194e-01 -7.06202030e-01 3.78924757e-01
-1.48653105e-01 3.29379946e-01 -1.90698341e-01 3.56391668e-02
4.05918598e-01 3.09587151e-01 1.36813685e-01 9.67946053e-02
1.46330772e-02 4.82940048e-01 -1.06037295e+00 1.57758439e+00
-7.27879524e-01 8.76219451e-01 2.97977716e-01 -9.37746108e-01
1.15308714e+00 6.86579883e-01 1.62181482e-01 -3.97493720e-01
2.04771280e-01 1.60295293e-01 -2.76555389e-01 -8.39728415e-02
5.32229960e-01 -2.34721541e-01 2.68752035e-02 5.78043401e-01
8.00334588e-02 -7.31862932e-02 -4.07663107e-01 -2.23869681e-01
5.29804051e-01 -5.83276689e-01 5.71444146e-02 -2.59247690e-01
6.94237888e-01 -7.94245422e-01 3.45052004e-01 4.13696289e-01
8.72640014e-02 2.40474105e-01 6.51021183e-01 2.72671971e-02
-1.05952752e+00 -8.92781079e-01 -1.27810165e-01 8.18035424e-01
-1.54327437e-01 -1.93595529e-01 -9.88265693e-01 -3.56797606e-01
-1.32513657e-01 1.15210354e+00 -2.26674542e-01 -7.08919525e-01
-3.60046864e-01 -1.14212684e-01 9.54063416e-01 7.70667568e-02
1.77075312e-01 -9.86013234e-01 -4.13703829e-01 1.16026022e-01
-6.12258948e-02 -1.04441822e+00 -5.20083368e-01 4.29483265e-01
-7.31348813e-01 -9.07711461e-02 -9.62869644e-01 -5.47694802e-01
5.39524674e-01 -5.61562777e-02 5.84553421e-01 -2.12241128e-01
2.36813232e-01 2.66257972e-02 -3.30295652e-01 -2.55007774e-01
-1.13822746e+00 4.36515898e-01 4.18177128e-01 4.54801410e-01
3.51977460e-02 -7.15632856e-01 -5.40476553e-02 2.19701052e-01
-1.22349489e+00 -1.54889107e-01 2.11861297e-01 8.26967299e-01
1.98638916e-01 3.11834246e-01 5.28721809e-01 -6.27407014e-01
8.58093560e-01 -2.34649628e-02 -7.62448132e-01 2.20262364e-01
-5.53451598e-01 7.28817642e-01 1.03331411e+00 -4.79096293e-01
-7.93296993e-01 -5.29786311e-02 -2.50510424e-01 -5.56441247e-01
-2.07196668e-01 -1.95915624e-01 -6.70933783e-01 7.37343803e-02
5.48888028e-01 3.76823038e-01 2.33456969e-01 -5.98190725e-01
7.91709840e-01 1.11414790e+00 2.76876211e-01 -1.66160151e-01
6.86409950e-01 2.58309335e-01 -2.20694900e-01 -1.25653195e+00
-2.53548950e-01 -2.08755612e-01 -5.45417130e-01 1.83693513e-01
8.99948597e-01 -6.62550569e-01 -6.87673569e-01 2.70623058e-01
-1.14952528e+00 -9.65823755e-02 -9.11624432e-01 5.33478498e-01
-7.11424768e-01 2.69911706e-01 -4.49398249e-01 -8.30019057e-01
-2.86337674e-01 -1.54508078e+00 9.15584803e-01 -3.02138656e-01
-1.89560786e-01 -1.03511643e+00 -1.55667424e-01 1.25524746e-02
4.82213527e-01 -1.73968911e-01 1.23842561e+00 -6.26552105e-01
-4.16695803e-01 -7.92364404e-02 3.08486670e-01 9.07190859e-01
4.17318970e-01 -2.37177446e-01 -1.52865708e+00 -5.21967828e-01
3.70549321e-01 -2.30287954e-01 8.68058681e-01 1.15311958e-01
1.17380083e+00 -8.93692672e-01 1.02418445e-01 8.16086590e-01
1.02606058e+00 1.02625348e-01 5.94743550e-01 -1.87964126e-01
3.25780243e-01 8.79102707e-01 3.75070423e-02 1.12224288e-01
-3.48651797e-01 1.00513029e+00 3.06399226e-01 -1.60845548e-01
-9.45703313e-02 -3.01079214e-01 8.26411963e-01 1.00470114e+00
4.28827405e-01 -3.25268686e-01 -2.84199476e-01 3.39123011e-01
-1.61673689e+00 -1.01715088e+00 6.73193395e-01 2.42294288e+00
1.02505243e+00 1.75209239e-01 -1.61376819e-02 4.70291853e-01
8.27496231e-01 3.37403417e-01 -3.97171348e-01 -1.03349328e+00
-5.97310299e-03 -1.14374496e-01 5.38805783e-01 6.96099937e-01
-7.41346538e-01 8.37729871e-01 6.61081696e+00 1.01472104e+00
-1.72402763e+00 8.14921483e-02 5.20339906e-01 -1.30178258e-01
-5.96841037e-01 -2.71120876e-01 -7.07678497e-01 4.86260891e-01
1.34991324e+00 -2.05847383e-01 5.93748808e-01 9.00968611e-01
1.86680093e-01 5.00248313e-01 -1.35354793e+00 9.39109027e-01
-7.96701387e-02 -1.26594555e+00 4.21459675e-01 2.98099786e-01
4.07416970e-01 -3.83308142e-01 3.91183138e-01 -5.05719222e-02
-1.26982912e-01 -1.17780292e+00 1.00000775e+00 4.09148455e-01
1.12427282e+00 -9.06994402e-01 6.89743102e-01 5.74106991e-01
-6.62048042e-01 2.77332664e-02 -5.56834340e-01 2.03498885e-01
-2.96256994e-03 5.29730380e-01 -9.43894565e-01 2.27142096e-01
1.14663832e-01 1.68992981e-01 -2.46881768e-01 5.02600610e-01
-2.68750906e-01 6.34606540e-01 -2.53765076e-01 -1.50071844e-01
2.04022273e-01 -2.47073308e-01 7.19285667e-01 9.57365513e-01
3.96944076e-01 -5.03406823e-01 -1.88868716e-01 9.92996514e-01
-3.94573212e-01 2.11535051e-01 -9.62826490e-01 -2.69373238e-01
5.07959902e-01 6.60125434e-01 -3.99921507e-01 -2.01222539e-01
4.95774299e-02 1.28943765e+00 1.79799154e-01 3.40901613e-01
-7.08028972e-01 -5.79486787e-01 8.78972948e-01 4.72266674e-02
1.33358672e-01 -1.04208410e-01 -7.27482652e-03 -1.16724682e+00
9.81029272e-02 -7.27150261e-01 -2.37999558e-01 -4.68131989e-01
-8.69157612e-01 8.99146080e-01 -2.00526431e-01 -1.14079559e+00
-7.57924020e-01 -3.11824292e-01 -2.24639818e-01 1.11137927e+00
-1.50659084e+00 -8.45259190e-01 2.42639259e-01 6.41781509e-01
3.27436119e-01 -5.77231169e-01 1.22817945e+00 2.84356385e-01
-4.15749550e-01 1.21222556e+00 4.85793263e-01 2.89980397e-02
6.20204687e-01 -1.16484010e+00 3.58643144e-01 7.13105738e-01
4.14801121e-01 6.36569083e-01 8.87937725e-01 3.20688933e-02
-1.00108957e+00 -1.24338627e+00 9.99809921e-01 1.00532554e-01
2.84303904e-01 -7.52805412e-01 -9.68689144e-01 4.77664798e-01
2.53463864e-01 -2.64448196e-01 7.25437224e-01 -9.26763415e-02
-2.83668280e-01 -2.26977423e-01 -1.25288284e+00 6.74149334e-01
5.89466989e-01 -1.08708966e+00 -4.44487095e-01 1.59550786e-01
1.36345732e+00 1.12217195e-01 -7.77075291e-01 -2.38217428e-01
3.98201466e-01 -8.12897861e-01 9.42044556e-01 -5.20849526e-01
-5.31962998e-02 -5.42705543e-02 -2.51416355e-01 -1.47780454e+00
-1.86183408e-01 -7.38206208e-01 1.53179631e-01 1.60008299e+00
4.71294403e-01 -8.70483816e-01 7.38236845e-01 4.06336457e-01
1.53968349e-01 -2.45728374e-01 -1.31398773e+00 -1.03580225e+00
1.56795040e-01 -2.08563924e-01 8.96892786e-01 9.12376642e-01
-1.76115915e-01 3.39932531e-01 -5.89730203e-01 2.77657479e-01
3.88521761e-01 3.80535796e-02 6.98552549e-01 -9.31254387e-01
-2.45882273e-01 -4.07037646e-01 -6.06452882e-01 -1.20128953e+00
6.35233402e-01 -1.08408558e+00 -1.17488205e-02 -8.63249779e-01
-5.93517721e-01 -4.12196249e-01 -2.60125250e-01 4.27286863e-01
4.46591765e-01 4.72374121e-03 3.40616256e-01 2.19059512e-01
6.00770973e-02 9.72292483e-01 8.95711362e-01 -2.45337367e-01
-3.84434849e-01 2.84940273e-01 -4.46251392e-01 5.43386400e-01
7.69154370e-01 -5.61618328e-01 -6.51894748e-01 -3.67344350e-01
-1.33347645e-01 1.24645688e-01 1.50215387e-01 -1.08377504e+00
1.57306239e-01 2.22962797e-02 -2.73270369e-01 -1.84570923e-01
7.34336913e-01 -1.26530027e+00 9.21790227e-02 5.88156283e-01
-9.35777426e-01 -4.76892740e-01 -3.67672704e-02 3.19630533e-01
-5.75962007e-01 -5.74454367e-01 9.82411563e-01 1.53643996e-01
-1.40754342e-01 6.73316419e-02 -5.52169025e-01 -3.36644977e-01
7.26003170e-01 -1.57892644e-01 3.31300013e-02 -7.09658623e-01
-7.56422758e-01 -2.15156958e-01 6.46310270e-01 5.77588737e-01
4.27820295e-01 -1.52567911e+00 -4.50562626e-01 6.21565282e-01
1.84662804e-01 -2.59394616e-01 -1.94304720e-01 2.62733608e-01
-3.59693885e-01 7.51409054e-01 -3.33178669e-01 -4.77019161e-01
-1.23607361e+00 6.17992759e-01 3.58248621e-01 7.99326971e-02
-5.74985981e-01 6.61822677e-01 1.31474018e-01 -5.05879104e-01
7.31672049e-01 -7.29837120e-01 -2.21154407e-01 2.62824446e-01
4.95836586e-01 1.31470874e-01 1.45569488e-01 -7.98178196e-01
-4.24061716e-02 1.12223290e-01 3.03990692e-01 -3.08847070e-01
1.06855953e+00 -2.30580747e-01 -1.08448990e-01 7.59455681e-01
1.65317452e+00 3.74380857e-01 -9.09195423e-01 -3.11681747e-01
-1.72515213e-01 -3.43383819e-01 3.13336462e-01 -1.69374287e-01
-1.01793158e+00 1.34526742e+00 7.40871131e-01 5.88864148e-01
1.00921571e+00 -2.21260145e-01 7.42442191e-01 4.65214580e-01
3.10457438e-01 -8.77486467e-01 -2.90461034e-01 2.97774047e-01
1.00281608e+00 -7.33686924e-01 -4.34109330e-01 -2.06696972e-01
-5.21045506e-01 1.09518349e+00 -2.19470993e-01 2.72626787e-01
5.75225055e-01 2.49281719e-01 4.65868451e-02 3.52792382e-01
-6.22277915e-01 2.19007507e-01 2.51292884e-01 4.62732941e-01
2.60240167e-01 2.34942541e-01 -5.59745841e-02 3.02441716e-01
-6.55549586e-01 -3.59936923e-01 5.13167858e-01 4.29104567e-01
-4.80457634e-01 -1.36093485e+00 -4.44239467e-01 -5.32348193e-02
-3.42494875e-01 -1.41541986e-02 -4.44618702e-01 2.37251520e-01
-1.17573226e-02 8.50437641e-01 1.53097168e-01 -3.34684134e-01
2.61708468e-01 3.92534107e-01 -9.59318411e-03 -6.47303760e-01
-4.36167598e-01 2.54733916e-02 8.92232284e-02 -5.01743317e-01
-3.23824197e-01 -3.99979800e-01 -1.10028839e+00 -3.39862853e-01
-4.14118797e-01 5.35745442e-01 9.18566227e-01 6.82093322e-01
2.74739534e-01 6.79889381e-01 8.84326756e-01 -3.76891285e-01
-9.50500727e-01 -7.43550479e-01 -9.41518307e-01 2.77262211e-01
7.67287672e-01 -2.61201888e-01 -4.90552396e-01 1.61431596e-01] | [14.872294425964355, 6.4548845291137695] |
c74e61e9-4c21-4e5c-aef0-eaa2ab4d0bc6 | mixup-mil-novel-data-augmentation-for | 2211.05862 | null | https://arxiv.org/abs/2211.05862v3 | https://arxiv.org/pdf/2211.05862v3.pdf | MixUp-MIL: Novel Data Augmentation for Multiple Instance Learning and a Study on Thyroid Cancer Diagnosis | Multiple instance learning exhibits a powerful approach for whole slide image-based diagnosis in the absence of pixel- or patch-level annotations. In spite of the huge size of hole slide images, the number of individual slides is often rather small, leading to a small number of labeled samples. To improve training, we propose and investigate different data augmentation strategies for multiple instance learning based on the idea of linear interpolations of feature vectors (known as MixUp). Based on state-of-the-art multiple instance learning architectures and two thyroid cancer data sets, an exhaustive study is conducted considering a range of common data augmentation strategies. Whereas a strategy based on to the original MixUp approach showed decreases in accuracy, the use of a novel intra-slide interpolation method led to consistent increases in accuracy. | ['Anton Hittmair', 'Gertie Janneke Oostingh', 'Sebastien Couillard-Despres', 'Christina Kreutzer', 'Lea Maria Stangassinger', 'Maximilian Tschuchnig', 'Lukas Koller', 'Michael Gadermayr'] | 2022-11-10 | null | null | null | null | ['multiple-instance-learning'] | ['methodology'] | [ 5.92062116e-01 3.27997118e-01 -3.20675075e-01 -3.15286756e-01
-1.36580479e+00 -1.25783160e-01 5.67858458e-01 5.47999442e-01
-5.37407041e-01 8.39602113e-01 -1.77767098e-01 -1.92076728e-01
-1.90275803e-01 -6.71715975e-01 -6.31498814e-01 -1.20514131e+00
2.39003062e-01 6.07902765e-01 3.88057351e-01 -2.46382207e-02
6.74508661e-02 5.24254203e-01 -1.67250919e+00 6.09830856e-01
5.73935091e-01 8.41781318e-01 -8.31616595e-02 7.94084013e-01
-4.61288512e-01 6.19611084e-01 -6.00751460e-01 -3.62963527e-01
3.40084434e-01 -5.12050688e-01 -7.53885508e-01 4.00636315e-01
7.13833928e-01 -1.97318196e-01 2.79392060e-02 5.97595513e-01
6.31056666e-01 -3.08023572e-01 5.01337230e-01 -1.05538988e+00
-4.06916410e-01 4.58012551e-01 -6.88039422e-01 1.33858904e-01
-1.44363120e-01 9.99268368e-02 7.89680600e-01 -6.95684195e-01
6.89189255e-01 5.52450120e-01 9.98075306e-01 6.02549374e-01
-1.51908255e+00 -3.27339977e-01 -1.45942882e-01 3.40625755e-02
-1.42582881e+00 -3.43831807e-01 5.13734281e-01 -1.98830858e-01
7.78953373e-01 3.88490021e-01 8.04619908e-01 6.34868324e-01
1.69640630e-02 7.83823252e-01 1.35487199e+00 -8.58308136e-01
1.09521084e-01 5.30731320e-01 7.34309852e-02 6.44197285e-01
9.22605395e-02 -1.72914326e-01 -1.70128495e-01 -2.16475487e-01
8.35529804e-01 2.19170839e-01 -1.36761880e-02 -3.79694253e-01
-1.21286237e+00 6.88952684e-01 4.48908389e-01 7.90941894e-01
-1.56324223e-01 2.98090558e-02 5.18839419e-01 3.16655964e-01
6.06310785e-01 4.61240470e-01 -4.26882476e-01 1.98364094e-01
-1.21341157e+00 1.65394112e-01 5.78122377e-01 2.93565124e-01
8.82262409e-01 -2.16972008e-01 -4.66589451e-01 7.01582670e-01
-7.34586269e-02 -6.95648566e-02 8.09780598e-01 -5.08056641e-01
4.55840528e-02 1.17760754e+00 -6.53770342e-02 -5.48148513e-01
-5.65671444e-01 -6.67977989e-01 -7.19619393e-01 2.36519381e-01
8.98039639e-01 1.53852239e-01 -1.23494327e+00 1.36369085e+00
4.03865695e-01 4.91991431e-01 -8.44403878e-02 4.60371494e-01
6.38960958e-01 2.01548070e-01 -2.22460572e-02 -3.02010268e-01
1.35170817e+00 -9.20075834e-01 -6.20758176e-01 4.58364300e-02
1.18938565e+00 -4.15623367e-01 1.13172424e+00 3.52558166e-01
-9.07295823e-01 -3.02563518e-01 -9.70236838e-01 -1.33487619e-02
-7.05617785e-01 2.99177408e-01 5.55590332e-01 9.17989314e-01
-1.21288049e+00 3.63266706e-01 -8.45383167e-01 -2.99177647e-01
8.09832275e-01 5.81366062e-01 -6.65795088e-01 -2.29148000e-01
-4.58258957e-01 8.73234510e-01 1.29540592e-01 -3.18102874e-02
-6.12589657e-01 -1.15106916e+00 -6.03697121e-01 -1.21117547e-01
2.65004516e-01 -4.42648649e-01 1.05451560e+00 -9.48287964e-01
-1.35403740e+00 1.14816678e+00 -1.98539585e-01 -4.68578547e-01
6.66017890e-01 1.76710710e-01 -3.14122066e-02 1.74996674e-01
-2.25599945e-01 8.86413395e-01 6.67603970e-01 -1.05672550e+00
-5.95390439e-01 -3.10668945e-01 -1.99854046e-01 -1.74628701e-02
-6.63051724e-01 -3.06892753e-01 -3.43229830e-01 -3.81088197e-01
8.88430863e-04 -9.19618785e-01 -5.46239316e-01 7.53794461e-02
-2.80150771e-01 -3.33765447e-02 8.54252636e-01 -3.84032696e-01
1.02916610e+00 -2.02339172e+00 -6.10345453e-02 1.74408630e-01
1.84782743e-01 3.33235145e-01 -2.13057101e-01 7.19868764e-02
-1.54140979e-01 3.21496516e-01 -3.63350213e-01 -4.59093392e-01
-2.98502177e-01 2.76917070e-01 3.09373140e-01 4.85228777e-01
4.51775223e-01 1.03984559e+00 -8.24297547e-01 -5.87363183e-01
3.35735530e-01 6.02759480e-01 -4.23028201e-01 -1.13854177e-01
-1.86974719e-01 4.33003783e-01 -1.78615645e-01 8.39862943e-01
4.34829235e-01 -5.73189199e-01 3.61516960e-02 -1.23040780e-01
-5.00103086e-02 -4.92424257e-02 -1.00902236e+00 1.57149303e+00
-2.95334905e-01 4.15684640e-01 -2.41989225e-01 -9.90840316e-01
7.30261564e-01 6.39672935e-01 6.39379323e-01 -6.04428113e-01
9.35265049e-02 4.50828493e-01 1.71244740e-01 -4.65303153e-01
2.11478412e-01 -3.77248526e-01 2.57860392e-01 3.79789352e-01
1.81164946e-02 -5.18925935e-02 4.72851723e-01 -1.97316147e-02
1.33688939e+00 -1.61874682e-01 5.36401927e-01 -1.65426552e-01
5.69457471e-01 1.17557071e-01 3.22657675e-01 6.89213276e-01
-1.56668842e-01 9.18838620e-01 6.35015070e-01 -6.15549266e-01
-1.09476364e+00 -4.84728098e-01 -4.22549754e-01 7.40441561e-01
-4.23479855e-01 -8.91480148e-02 -6.42465770e-01 -9.11522150e-01
6.87317178e-02 1.38858944e-01 -1.22033572e+00 3.50035459e-01
-4.04664636e-01 -1.36125171e+00 6.42627656e-01 5.01687467e-01
2.62062222e-01 -8.86575758e-01 -5.06760478e-01 2.66903281e-01
9.41557884e-02 -9.11701500e-01 2.14483384e-02 3.44415843e-01
-1.09641445e+00 -1.14942122e+00 -9.58038330e-01 -6.27810776e-01
1.05710971e+00 3.44132222e-02 1.03723431e+00 5.68088770e-01
-7.64241517e-01 2.26963937e-01 -3.61579657e-01 -6.45026028e-01
-6.40795648e-01 2.24789932e-01 -4.17631149e-01 1.55291110e-01
3.69559348e-01 -3.36138487e-01 -5.51318586e-01 3.74573991e-02
-1.18802083e+00 -2.44335290e-02 1.16531301e+00 1.14087737e+00
1.02559865e+00 -2.55120128e-01 6.59988523e-01 -1.34069991e+00
2.74152398e-01 -9.62594301e-02 -3.72515559e-01 4.18027759e-01
-5.92432201e-01 -1.31840423e-01 3.02269667e-01 -5.81173837e-01
-6.48027718e-01 1.76711440e-01 -1.17647126e-01 -1.73907652e-01
-2.73990303e-01 5.86895049e-01 2.78189719e-01 -6.01880133e-01
7.08665788e-01 1.98438510e-01 4.51499701e-01 -1.12499811e-01
-1.34395817e-02 2.39978626e-01 1.15386017e-01 1.09868571e-01
4.83843535e-01 5.34516335e-01 2.00089619e-01 -8.49037528e-01
-7.22303033e-01 -5.81832349e-01 -9.02321219e-01 -1.87488243e-01
5.90820134e-01 -5.94901919e-01 -3.60897273e-01 6.06854737e-01
-4.01595652e-01 -6.17590964e-01 -6.43103778e-01 3.11081022e-01
-3.54046166e-01 5.38409837e-02 -7.49799430e-01 -4.38411146e-01
-1.97983071e-01 -1.32394123e+00 1.16605699e+00 2.58644372e-01
-2.21830495e-02 -1.15340805e+00 2.46951848e-01 3.70330393e-01
6.17869198e-01 4.37152117e-01 9.33452487e-01 -1.07352293e+00
-6.04984522e-01 -8.81094873e-01 -1.13772161e-01 1.38903990e-01
3.75767440e-01 9.77210253e-02 -1.13824749e+00 -2.65001178e-01
-3.00424337e-01 -3.88888270e-01 7.93731213e-01 5.19722581e-01
1.20149505e+00 3.60742286e-02 -5.91290593e-01 3.26269895e-01
1.68223274e+00 -1.61755249e-01 8.24271321e-01 4.59704429e-01
4.24321294e-01 3.99727553e-01 3.56338501e-01 2.95310348e-01
9.61492881e-02 5.65579712e-01 4.25768524e-01 -3.77986103e-01
-3.14253688e-01 3.03613573e-01 -2.62360960e-01 2.91476846e-01
-9.83582437e-03 -1.48365080e-01 -9.48996305e-01 8.41648817e-01
-1.32457769e+00 -6.87828362e-01 -1.03753105e-01 2.21362758e+00
8.25682461e-01 1.37518451e-01 1.05415069e-01 6.09360039e-01
3.88780594e-01 -1.85270220e-01 -3.47598195e-01 -5.93830384e-02
-8.46280903e-02 2.10054025e-01 4.59320247e-01 2.13512152e-01
-9.38667536e-01 2.36284941e-01 7.06975222e+00 8.53379905e-01
-1.34840417e+00 1.34165108e-01 1.05233705e+00 -2.68823147e-01
-6.56648427e-02 -2.84772545e-01 -9.76553082e-01 1.50897160e-01
9.70981896e-01 1.98484078e-01 -1.32499933e-01 4.57903147e-01
-5.09350263e-02 -2.27208406e-01 -1.02482331e+00 4.73715484e-01
1.78968489e-01 -1.73478878e+00 9.55331884e-03 2.96521813e-01
8.26851726e-01 1.28427586e-02 -3.58986631e-02 1.96345657e-01
-2.13208705e-01 -8.81303430e-01 4.55250069e-02 3.73453051e-01
7.48545468e-01 -5.76292872e-01 1.16687918e+00 1.63276821e-01
-7.16269076e-01 -2.42270213e-02 5.82843088e-03 1.52129263e-01
-2.35318303e-01 7.27931201e-01 -1.29649174e+00 5.29275656e-01
4.59690213e-01 3.36605996e-01 -1.06375730e+00 1.20361602e+00
2.15911478e-01 7.02596784e-01 -4.15444583e-01 8.53081644e-02
2.95959234e-01 1.99167445e-01 1.05575733e-01 1.27576137e+00
2.75728762e-01 -1.03717491e-01 -2.07177773e-02 4.14918184e-01
-2.89462646e-03 2.98923433e-01 -2.56837845e-01 3.89002822e-02
2.70232588e-01 1.75635457e+00 -1.04782808e+00 -3.53938341e-01
-7.15096831e-01 5.79352796e-01 2.94016033e-01 4.68592974e-04
-4.11525756e-01 -1.44442573e-01 -1.09271191e-01 4.69260007e-01
3.40747982e-01 3.33289713e-01 -3.60157579e-01 -4.97576147e-01
-1.18597724e-01 -8.43796432e-01 4.20523882e-01 -3.02118450e-01
-1.16590774e+00 4.45940405e-01 -1.80428252e-01 -1.30398297e+00
-1.82150289e-01 -5.75782657e-01 -7.96949267e-01 5.83436787e-01
-1.69213998e+00 -1.54973853e+00 -3.62232417e-01 2.69351065e-01
2.86507905e-01 -1.54026419e-01 1.12602997e+00 3.37810069e-01
-5.40547073e-01 1.00681984e+00 2.48748004e-01 5.16219698e-02
7.61067927e-01 -1.33388674e+00 -2.21526280e-01 2.96019167e-01
2.02913433e-02 3.22203040e-01 6.07950449e-01 -2.42364362e-01
-1.17134106e+00 -9.65545058e-01 7.39013076e-01 -4.68503207e-01
5.48897147e-01 -3.90644595e-02 -1.32370591e+00 7.40468144e-01
1.35794684e-01 6.15799546e-01 1.18273520e+00 2.78558303e-02
3.75797637e-02 -2.18616515e-01 -1.52323616e+00 4.56830591e-01
1.70075670e-01 -1.23107560e-01 -5.16343154e-02 3.32385451e-01
3.31566006e-01 -4.58882540e-01 -1.12889850e+00 5.34149408e-01
4.92657542e-01 -9.08037663e-01 7.50153720e-01 -2.26857603e-01
2.78102070e-01 -2.34129846e-01 9.74498168e-02 -1.06576955e+00
-1.58811420e-01 -4.47635837e-02 1.05521260e-02 1.24009252e+00
7.23027647e-01 -6.06092036e-01 1.39900482e+00 4.86464202e-01
-2.79370815e-01 -1.16652441e+00 -1.08653772e+00 -3.52515668e-01
5.50018586e-02 1.46457285e-01 2.82887936e-01 9.34393466e-01
4.36181650e-02 -1.66226581e-01 -5.66646159e-02 -7.80447572e-02
4.69431937e-01 -1.83993839e-02 7.03379571e-01 -1.24285614e+00
-3.47186774e-01 -5.12615860e-01 -6.47685349e-01 -1.78243384e-01
-1.79031059e-01 -9.17510509e-01 -1.23068482e-01 -1.46026230e+00
4.87779200e-01 -5.63997209e-01 -5.90898514e-01 7.94119418e-01
-4.24310505e-01 8.88085306e-01 -1.61662042e-01 -3.06394324e-02
-5.26702762e-01 7.38531426e-02 1.14444184e+00 -1.97741136e-01
-2.70067722e-01 1.35370687e-01 -6.79670453e-01 4.32475924e-01
6.74701512e-01 -4.55340683e-01 -1.69961348e-01 -1.99503273e-01
7.69316033e-02 2.00508162e-02 3.07230115e-01 -1.09327090e+00
2.18616173e-01 1.59390017e-01 5.37055016e-01 -6.98533058e-01
3.36529255e-01 -6.67225480e-01 1.70694873e-01 6.27452910e-01
-4.39117670e-01 -2.19318792e-01 5.14737546e-01 5.34516513e-01
-3.97477359e-01 -3.08144987e-01 8.02161515e-01 -3.94735634e-01
-3.94836485e-01 1.75782099e-01 -3.40181470e-01 -5.23477912e-01
1.15525055e+00 -6.98182166e-01 -2.72913843e-01 1.53891683e-01
-8.25032473e-01 1.25560969e-01 4.67625171e-01 -1.90012977e-02
3.01102251e-01 -1.08656991e+00 -8.51719320e-01 1.84295610e-01
4.29715127e-01 1.25854775e-01 4.49679017e-01 1.36889970e+00
-4.18824732e-01 3.55855852e-01 -1.43599033e-01 -8.50399017e-01
-1.60049093e+00 3.80294353e-01 4.78331834e-01 -9.49400783e-01
-4.41667199e-01 8.28691900e-01 -4.15676506e-03 -5.24851918e-01
8.55893344e-02 -1.95704713e-01 -2.10443079e-01 1.90010652e-01
7.71596134e-01 2.56983787e-01 6.37821317e-01 -1.31802827e-01
-1.04502060e-01 2.35669166e-01 -5.18002391e-01 2.09026605e-01
1.49975562e+00 1.46090731e-01 -6.97928146e-02 5.29056013e-01
9.81729209e-01 -1.61609203e-01 -1.21469092e+00 -2.34061569e-01
1.31443486e-01 -3.92179191e-01 5.30391671e-02 -9.22200143e-01
-1.11023355e+00 7.50083208e-01 9.85053599e-01 1.02299489e-01
1.14911902e+00 -1.66933149e-01 2.34133840e-01 2.91520268e-01
2.13317662e-01 -6.50375545e-01 2.43499994e-01 2.77911574e-02
4.68614727e-01 -1.46440244e+00 2.95902282e-01 -3.45799088e-01
-3.37122649e-01 9.89000976e-01 5.58784127e-01 4.01232205e-02
4.77788299e-01 6.80097401e-01 2.77666211e-01 -9.65866819e-02
-8.17208827e-01 -1.14698470e-01 2.31916998e-02 4.74751323e-01
7.55697072e-01 -2.03263789e-01 -2.21604213e-01 3.10033351e-01
3.69632810e-01 3.21138561e-01 6.22416914e-01 1.07081163e+00
-3.17531645e-01 -1.37738705e+00 -4.09946084e-01 9.29828584e-01
-7.70655215e-01 1.25875264e-01 -3.07805270e-01 1.35802722e+00
1.76214129e-01 3.80688131e-01 2.13529602e-01 -3.51559818e-02
1.23787381e-01 3.05574238e-01 7.78342247e-01 -5.67420959e-01
-9.17504847e-01 1.46867812e-01 -2.11871073e-01 -6.60398006e-02
-7.92335749e-01 -8.63895416e-01 -1.19179738e+00 1.18172653e-01
-6.07980609e-01 -1.06407486e-01 5.56901634e-01 1.19090104e+00
1.54141307e-01 8.15606356e-01 2.80963093e-01 -8.76701236e-01
-2.61617005e-01 -1.24622416e+00 -6.28635526e-01 4.67241883e-01
4.74779099e-01 -4.22002643e-01 -2.55584091e-01 1.64193183e-01] | [15.066178321838379, -2.9501848220825195] |
8cec16e5-3006-421b-ae82-c33ea0b1976c | transflow-transformer-as-flow-learner | 2304.11523 | null | https://arxiv.org/abs/2304.11523v1 | https://arxiv.org/pdf/2304.11523v1.pdf | TransFlow: Transformer as Flow Learner | Optical flow is an indispensable building block for various important computer vision tasks, including motion estimation, object tracking, and disparity measurement. In this work, we propose TransFlow, a pure transformer architecture for optical flow estimation. Compared to dominant CNN-based methods, TransFlow demonstrates three advantages. First, it provides more accurate correlation and trustworthy matching in flow estimation by utilizing spatial self-attention and cross-attention mechanisms between adjacent frames to effectively capture global dependencies; Second, it recovers more compromised information (e.g., occlusion and motion blur) in flow estimation through long-range temporal association in dynamic scenes; Third, it enables a concise self-learning paradigm and effectively eliminate the complex and laborious multi-stage pre-training procedures. We achieve the state-of-the-art results on the Sintel, KITTI-15, as well as several downstream tasks, including video object detection, interpolation and stabilization. For its efficacy, we hope TransFlow could serve as a flexible baseline for optical flow estimation. | ['Dongfang Liu', 'Huaijin Chen', 'Yingjie Victor Chen', 'Tong Geng', 'Siqi Ma', 'Qifan Wang', 'Yawen Lu'] | 2023-04-23 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Lu_TransFlow_Transformer_As_Flow_Learner_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Lu_TransFlow_Transformer_As_Flow_Learner_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-object-detection', 'motion-estimation', 'self-learning'] | ['computer-vision', 'computer-vision', 'natural-language-processing'] | [-3.57309401e-01 -8.38914931e-01 -3.50244373e-01 -1.30966336e-01
-2.92933434e-01 -3.17957401e-01 3.14204782e-01 -3.33510011e-01
-3.75652432e-01 7.83122540e-01 4.17530596e-01 -7.68962502e-02
2.29294032e-01 -5.77823579e-01 -5.85036874e-01 -6.39293492e-01
4.72396277e-02 -1.79496542e-01 4.78677601e-01 -8.81257877e-02
3.52087349e-01 5.27371883e-01 -1.41382480e+00 6.57677650e-02
9.83545065e-01 1.28158462e+00 2.16714978e-01 6.26820743e-01
1.55540556e-01 1.34113169e+00 -2.95565009e-01 -3.62004846e-01
1.38932198e-01 -2.26452589e-01 -6.91380024e-01 -2.56148763e-02
1.16627538e+00 -1.01772583e+00 -9.96614099e-01 9.79312181e-01
4.60465699e-01 3.40780884e-01 3.40427220e-01 -1.10995555e+00
-7.85533428e-01 -2.96266526e-02 -6.83082819e-01 6.73313916e-01
3.02730560e-01 7.91296303e-01 8.05514574e-01 -1.07949209e+00
6.21523499e-01 1.38354349e+00 5.29429615e-01 6.12521708e-01
-1.10349441e+00 -8.60602498e-01 7.50334114e-02 7.53828168e-01
-1.06158578e+00 -8.49151909e-01 6.33523643e-01 -7.48678029e-01
7.47762501e-01 6.58347504e-03 8.81342411e-01 9.28461194e-01
2.44317532e-01 1.05142879e+00 5.06316185e-01 6.32638484e-02
-1.55867383e-01 -4.30485308e-01 -2.46962849e-02 9.22838867e-01
1.61888584e-01 3.41986388e-01 -5.73897302e-01 3.29003155e-01
1.32734478e+00 -1.12630673e-01 -8.15223873e-01 -1.23588786e-01
-1.46029794e+00 5.14606118e-01 7.57839024e-01 3.47958431e-02
-1.63264185e-01 5.09319663e-01 4.66862470e-01 1.74277066e-03
3.94296765e-01 2.13778302e-01 -2.34030530e-01 -2.11402327e-01
-8.72439206e-01 1.40754849e-01 2.36866310e-01 9.28084075e-01
8.99265885e-01 4.84759003e-01 -5.50028265e-01 5.85539877e-01
3.98393720e-01 5.34278154e-01 4.67870384e-01 -1.42402864e+00
6.27965093e-01 2.40084842e-01 4.02382463e-01 -1.19103932e+00
-1.58047423e-01 -3.31294358e-01 -1.09833920e+00 2.70970494e-01
6.00224912e-01 -3.62929367e-02 -6.06783569e-01 1.66526473e+00
2.61436135e-01 8.28329623e-01 -3.08622748e-01 1.30851161e+00
1.04945326e+00 6.29905224e-01 -6.87213838e-02 -3.21706325e-01
1.04756391e+00 -1.44119060e+00 -8.58021319e-01 -3.70433718e-01
1.37203172e-01 -9.00283456e-01 8.48968208e-01 1.72406629e-01
-1.26095688e+00 -9.54971254e-01 -8.72405410e-01 -6.55167818e-01
2.08806023e-01 1.57035381e-01 1.07596183e+00 1.50942221e-01
-1.06574047e+00 5.98006368e-01 -1.01064050e+00 4.20142012e-03
8.14010739e-01 2.48907059e-01 -3.71830374e-01 -1.64178342e-01
-1.05020857e+00 6.20325208e-01 4.20351103e-02 4.83156383e-01
-9.09059525e-01 -9.38349843e-01 -1.17641270e+00 2.02607047e-02
1.32321268e-01 -1.12970650e+00 9.89449441e-01 -7.70049453e-01
-1.66125059e+00 6.24250412e-01 -6.40389621e-01 -2.70125747e-01
7.85423040e-01 -4.66496974e-01 -2.27100581e-01 2.38115326e-01
2.08736986e-01 8.07784200e-01 1.03940856e+00 -5.86033523e-01
-6.08744621e-01 -1.27337962e-01 -1.28806382e-01 1.90550387e-02
-1.98123410e-01 -2.12788787e-02 -7.82328129e-01 -8.01701844e-01
-2.93599181e-02 -6.35670960e-01 -1.34217307e-01 6.08723938e-01
-1.63971856e-01 -8.50463137e-02 9.23304558e-01 -5.83182812e-01
1.40862179e+00 -1.96470571e+00 5.51854335e-02 -4.79366213e-01
5.19475937e-01 7.79712200e-01 -1.78249910e-01 -2.26694360e-01
6.38506413e-02 -3.11720818e-01 -9.77813974e-02 -4.02271062e-01
-4.73008424e-01 -9.80375484e-02 -2.97781050e-01 6.81269825e-01
4.26264375e-01 1.28081524e+00 -1.16055441e+00 -6.81051254e-01
8.94202471e-01 7.28308976e-01 -7.68279791e-01 2.50987560e-01
-8.77612904e-02 9.23744977e-01 -3.95511419e-01 5.35797656e-01
8.42545688e-01 -3.79359514e-01 -4.95897204e-01 -6.20947361e-01
-2.19929501e-01 2.69128472e-01 -1.05631614e+00 2.02010584e+00
-3.73321682e-01 1.31155169e+00 -7.01601803e-02 -6.83622181e-01
6.41215324e-01 8.41472521e-02 6.95201993e-01 -6.95946813e-01
2.25254968e-01 9.23503861e-02 -1.28219843e-01 -7.41101980e-01
4.32530642e-01 3.60253334e-01 5.81045091e-01 1.37446404e-01
2.11571902e-01 1.65299311e-01 3.33538413e-01 9.59457606e-02
8.06518495e-01 3.39469403e-01 1.59764364e-01 -3.19392651e-01
1.05714488e+00 -4.54268217e-01 8.95786285e-01 3.31036150e-01
-8.22528481e-01 8.20438385e-01 1.00672036e-01 -8.28092277e-01
-7.34756112e-01 -1.06366658e+00 -1.22823648e-01 4.85419393e-01
6.07455373e-01 -2.66161054e-01 -3.84811908e-01 -3.55273455e-01
1.11230075e-01 3.95128652e-02 -5.26477516e-01 -2.05727652e-01
-9.10057187e-01 -4.35455024e-01 2.72290856e-01 6.58504963e-01
9.17883337e-01 -1.00892055e+00 -4.80267078e-01 4.02989715e-01
-6.53062701e-01 -1.46000493e+00 -1.21207154e+00 -6.27773702e-01
-9.05150771e-01 -1.18847406e+00 -7.58245051e-01 -7.67927885e-01
3.96703005e-01 6.84207618e-01 1.23598158e+00 1.83997035e-01
-3.97026807e-01 -1.31393105e-01 1.54887319e-01 2.69108891e-01
1.44348472e-01 -2.56862700e-01 -2.10042283e-01 2.71503717e-01
2.10502863e-01 -4.89781618e-01 -1.23535216e+00 4.51072991e-01
-4.88230437e-01 1.32337753e-02 1.18683986e-01 1.00330257e+00
2.41595879e-01 -4.78862017e-01 2.27466643e-01 -3.57310921e-01
1.37354657e-01 -1.14077553e-02 -8.47500145e-01 7.25070685e-02
-2.29939267e-01 -1.09952614e-01 5.94049335e-01 -4.14767712e-01
-1.20404124e+00 -2.75317729e-02 -2.50706933e-02 -7.87642896e-01
2.02617332e-01 -1.76499560e-01 3.45455622e-03 -5.49826562e-01
4.80511516e-01 2.27023661e-01 1.23390384e-01 -1.93776757e-01
1.69618651e-01 3.31032634e-01 8.81413460e-01 -3.41406584e-01
6.30476654e-01 7.44202375e-01 2.59416908e-01 -6.95023119e-01
-9.28175271e-01 -6.02530122e-01 -6.15498722e-01 -3.75628769e-01
9.83576655e-01 -1.18271780e+00 -1.39990366e+00 9.08163011e-01
-1.49244785e+00 -2.57096797e-01 1.54304113e-02 7.59220421e-01
-4.97530758e-01 5.88557303e-01 -1.12063551e+00 -4.91201848e-01
-4.69865352e-01 -1.39506626e+00 1.00618160e+00 5.17053187e-01
9.81775522e-02 -1.22196329e+00 -8.92780051e-02 4.32741553e-01
5.52892804e-01 7.59232044e-02 1.68526769e-01 5.40269971e-01
-1.22241855e+00 3.15218657e-01 -6.76593184e-01 3.71356308e-01
3.06887716e-01 3.38971496e-01 -1.00545454e+00 -4.16415721e-01
-1.21907495e-01 -3.10677677e-01 1.15921152e+00 7.66138911e-01
1.10954630e+00 -1.50207773e-01 -7.63143674e-02 1.33577693e+00
1.28240359e+00 1.24434479e-01 9.13729489e-01 1.82674974e-01
1.18229508e+00 4.45145875e-01 5.60184121e-01 2.95628726e-01
4.88011479e-01 8.19445252e-01 3.97364974e-01 -3.13229114e-01
-6.34080827e-01 -1.68747723e-01 3.42168689e-01 6.61698937e-01
-1.44672960e-01 -2.99516153e-02 -5.30726433e-01 3.87193024e-01
-2.04515791e+00 -1.24409282e+00 -3.87012929e-01 2.01050520e+00
7.34377563e-01 -1.53823093e-01 -5.68973348e-02 -2.23244518e-01
7.47647703e-01 4.84112114e-01 -7.42017984e-01 2.71057189e-01
-2.65994519e-01 3.89783271e-03 3.22780430e-01 8.61986041e-01
-1.34116387e+00 1.22123230e+00 6.16484308e+00 5.94014287e-01
-1.36920702e+00 8.22931062e-03 8.30434620e-01 -1.26262084e-01
3.32966596e-02 -2.37021372e-01 -6.92477047e-01 7.81842649e-01
8.04396719e-02 -3.28326933e-02 4.39079344e-01 3.42861444e-01
5.00659645e-01 -1.62005901e-01 -1.04829919e+00 1.37724829e+00
7.87868258e-03 -1.83665311e+00 -6.27327710e-02 -1.69100702e-01
9.16700006e-01 1.23736531e-01 -1.26032019e-02 -1.77044436e-01
7.84696713e-02 -8.93754900e-01 6.50146544e-01 4.97226208e-01
8.74440491e-01 -4.29639220e-01 6.18701160e-01 -9.29812491e-02
-1.38598275e+00 -2.00465191e-02 -3.35798919e-01 -1.91819251e-01
3.96943599e-01 6.91814840e-01 1.35244587e-02 4.03932244e-01
6.62839532e-01 1.57241189e+00 -4.52619284e-01 1.35130298e+00
-1.95211187e-01 2.28441641e-01 -4.02474254e-02 3.46756339e-01
1.49590448e-01 -4.02039379e-01 4.72404599e-01 1.03348601e+00
1.26943484e-01 -8.09708387e-02 2.42947005e-02 8.73889506e-01
8.98386911e-03 -2.23353505e-01 -2.34720215e-01 4.76064414e-01
3.49467486e-01 1.07381153e+00 -2.59710014e-01 -4.19714302e-01
-5.67611396e-01 9.35984969e-01 3.42166036e-01 6.31187677e-01
-8.95903647e-01 -2.93748558e-01 1.24473059e+00 -5.18172840e-03
2.68357217e-01 -3.47155571e-01 -4.45365086e-02 -1.85837936e+00
2.01345772e-01 -3.94445151e-01 2.06641153e-01 -9.71835136e-01
-1.11790752e+00 6.04599714e-01 -5.48523664e-01 -1.58801246e+00
-2.17458636e-01 -6.47751451e-01 -7.00322568e-01 8.95615995e-01
-1.94278526e+00 -8.66148055e-01 -7.92668462e-01 9.05048907e-01
6.11069322e-01 5.12981205e-04 2.34269425e-01 7.84010470e-01
-8.53655457e-01 4.56602454e-01 -6.40348047e-02 5.14754057e-01
1.07449675e+00 -9.27542269e-01 5.20812690e-01 1.34510839e+00
-1.40704801e-02 4.13559169e-01 3.89071107e-01 -3.23831528e-01
-1.35464668e+00 -1.16309166e+00 6.53728962e-01 -4.09986705e-01
7.91205287e-01 5.32190409e-03 -8.41503322e-01 6.02990270e-01
7.15344213e-03 8.16431880e-01 -2.53095757e-02 -4.96058077e-01
-3.54440331e-01 -5.58334410e-01 -7.06489027e-01 4.99870241e-01
1.21580458e+00 -5.27742565e-01 -2.32714508e-02 1.15043767e-01
6.56845033e-01 -6.70362115e-01 -6.16740525e-01 3.27600867e-01
6.87766671e-01 -1.50712812e+00 1.33514214e+00 -3.24377596e-01
7.05139756e-01 -5.04529417e-01 2.69677848e-01 -9.58145976e-01
-5.78847766e-01 -1.09347618e+00 -5.33515036e-01 1.17774475e+00
-2.13368699e-01 -6.54389739e-01 9.66153324e-01 5.51518261e-01
-1.22534536e-01 -5.33922493e-01 -8.31310034e-01 -4.33551282e-01
-2.54699379e-01 -2.93610901e-01 3.56494993e-01 9.85384166e-01
-3.98291111e-01 3.47744793e-01 -7.02891827e-01 -1.13955714e-01
9.09598649e-01 1.58080831e-01 8.70144784e-01 -9.43672299e-01
-1.91195756e-01 -6.43032491e-01 -7.14258730e-01 -1.84068024e+00
2.69971699e-01 -3.67533892e-01 -8.99132527e-03 -1.30537331e+00
-1.61330953e-01 -3.09365004e-01 -2.56713271e-01 -1.11852465e-02
-4.93280560e-01 4.12884027e-01 2.90975899e-01 4.46565628e-01
-6.86987996e-01 6.92410827e-01 2.07986403e+00 -1.60793170e-01
-8.05972815e-02 9.07551944e-02 -3.98482382e-01 5.83607554e-01
3.88695031e-01 1.33154113e-02 -3.70540738e-01 -8.30681384e-01
-3.92693132e-01 3.65860015e-01 6.23326302e-01 -9.52100217e-01
4.70872819e-01 -3.30112338e-01 6.23774230e-01 -4.53562200e-01
3.24237317e-01 -5.98189056e-01 -1.56420827e-01 5.26383758e-01
-4.03871387e-02 1.33500010e-01 1.90892100e-01 4.34570253e-01
-5.99948406e-01 1.65279880e-01 9.92456555e-01 -3.39715518e-02
-1.06765807e+00 9.61924911e-01 9.01000649e-02 3.08719605e-01
8.47797692e-01 -2.22567588e-01 -6.16047561e-01 -4.68852162e-01
-3.15044016e-01 4.52128321e-01 4.09317434e-01 4.74449188e-01
7.48704255e-01 -1.38453650e+00 -7.53053069e-01 5.93320727e-01
-9.74107981e-02 2.41682172e-01 4.54247504e-01 9.75771606e-01
-9.00194407e-01 4.30631667e-01 -4.74952966e-01 -9.16547418e-01
-9.09904003e-01 4.88924623e-01 4.96131271e-01 7.17361569e-02
-8.05035889e-01 1.05012679e+00 5.42550147e-01 3.27915221e-01
3.26614588e-01 -3.50092709e-01 -1.26811758e-01 -1.26118615e-01
9.42902923e-01 6.01768374e-01 -1.40327379e-01 -7.48842537e-01
-4.30715322e-01 9.65873957e-01 1.43604726e-01 2.26266578e-01
9.40510333e-01 -4.46058303e-01 -1.56853437e-01 1.17849670e-01
1.40125215e+00 -2.42634937e-01 -2.06059241e+00 -2.19273627e-01
-4.37707692e-01 -1.23933017e+00 3.17996621e-01 -3.11290503e-01
-1.72822809e+00 1.28528655e+00 4.47234154e-01 -2.13518307e-01
1.20882761e+00 -3.60408336e-01 1.04089916e+00 4.98319715e-02
1.27781913e-01 -5.91817319e-01 2.04191417e-01 5.27343392e-01
7.76392996e-01 -1.63228273e+00 5.88715971e-02 -6.14013433e-01
-3.55165154e-01 1.24530208e+00 9.98095691e-01 -1.23632967e-01
5.58185160e-01 5.14380038e-02 2.17888817e-01 2.76573777e-01
-8.64061594e-01 -2.87622690e-01 5.08135378e-01 7.55732238e-01
5.57407081e-01 -4.29281920e-01 1.71699911e-01 -2.81296790e-01
1.43992111e-01 3.26303542e-01 4.54571366e-01 2.42656112e-01
-9.10758823e-02 -7.16869414e-01 -5.38344793e-02 1.53530717e-01
-3.97852778e-01 -2.71723896e-01 2.53027678e-01 6.27537012e-01
-4.79699709e-02 8.00137341e-01 2.87662297e-01 -2.06443876e-01
1.71820268e-01 -7.19563961e-01 6.67722821e-01 -8.12357813e-02
-4.30765480e-01 1.12202428e-01 -1.22884803e-01 -1.06115866e+00
-6.35425687e-01 -4.81929988e-01 -8.65913212e-01 -6.22222483e-01
-1.92707375e-01 -2.60159403e-01 4.78290170e-02 7.75676370e-01
4.18563038e-01 4.80960876e-01 6.33456171e-01 -1.19480634e+00
-9.11621228e-02 -7.64271021e-01 -8.83891732e-02 6.45694733e-01
9.49095666e-01 -8.15282285e-01 -3.97760570e-01 2.65867770e-01] | [8.94531536102295, -1.8159418106079102] |
18433aa5-7073-41ba-b68b-45b866f8f809 | gazeonce-real-time-multi-person-gaze | 2204.09480 | null | https://arxiv.org/abs/2204.09480v1 | https://arxiv.org/pdf/2204.09480v1.pdf | GazeOnce: Real-Time Multi-Person Gaze Estimation | Appearance-based gaze estimation aims to predict the 3D eye gaze direction from a single image. While recent deep learning-based approaches have demonstrated excellent performance, they usually assume one calibrated face in each input image and cannot output multi-person gaze in real time. However, simultaneous gaze estimation for multiple people in the wild is necessary for real-world applications. In this paper, we propose the first one-stage end-to-end gaze estimation method, GazeOnce, which is capable of simultaneously predicting gaze directions for multiple faces (>10) in an image. In addition, we design a sophisticated data generation pipeline and propose a new dataset, MPSGaze, which contains full images of multiple people with 3D gaze ground truth. Experimental results demonstrate that our unified framework not only offers a faster speed, but also provides a lower gaze estimation error compared with state-of-the-art methods. This technique can be useful in real-time applications with multiple users. | ['Feng Lu', 'Yunfei Liu', 'Mingfang Zhang'] | 2022-04-20 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Zhang_GazeOnce_Real-Time_Multi-Person_Gaze_Estimation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Zhang_GazeOnce_Real-Time_Multi-Person_Gaze_Estimation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['gaze-estimation'] | ['computer-vision'] | [ 8.09241310e-02 -5.04707843e-02 9.14944112e-02 -6.66272342e-01
-2.63780922e-01 -2.13542759e-01 2.14158148e-01 -5.91195107e-01
-2.42801756e-01 3.34245473e-01 -2.27486715e-01 -1.41439721e-01
3.06105614e-01 -1.39298514e-01 -5.73717892e-01 -4.97473150e-01
3.17035317e-01 2.34502062e-01 1.31264761e-01 -4.96627651e-02
4.00640666e-01 1.46380454e-01 -2.20775628e+00 -4.80787270e-02
8.06654871e-01 1.17954290e+00 1.56626791e-01 7.87046790e-01
2.08362371e-01 5.66913188e-01 -4.40556169e-01 -5.40634096e-01
1.73777878e-01 -4.95208830e-01 -4.90709335e-01 2.37555936e-01
1.46226251e+00 -8.48025858e-01 2.80128747e-01 9.58019733e-01
7.17541695e-01 -1.11823030e-01 3.19026142e-01 -1.74314964e+00
-5.41062176e-01 -2.08831251e-01 -1.32273316e+00 1.28473103e-01
7.62366354e-01 4.63077992e-01 6.76742852e-01 -9.62173402e-01
2.90856272e-01 1.33850837e+00 4.58938599e-01 9.70167160e-01
-1.03318048e+00 -1.14835620e+00 3.41518074e-01 2.55950749e-01
-1.27997482e+00 -9.91616964e-01 6.74713135e-01 -3.97706866e-01
4.74483162e-01 1.55379966e-01 6.10891104e-01 1.09374714e+00
1.21456813e-02 9.12965417e-01 1.30250883e+00 -5.94085932e-01
-2.24913180e-01 -2.30208840e-02 -1.43704683e-01 9.08513486e-01
5.79323098e-02 2.67345142e-02 -1.11542583e+00 3.71972084e-01
7.15744317e-01 8.96392018e-02 -4.74484354e-01 -3.44958544e-01
-1.15693033e+00 4.00603712e-01 6.67017996e-01 -2.55840272e-01
-2.98406780e-01 9.08819288e-02 -1.71334714e-01 4.43848148e-02
6.77775443e-01 -1.38255209e-01 -1.87155828e-01 -2.04893127e-01
-1.13624430e+00 4.76944327e-01 4.67493206e-01 1.05127180e+00
8.03574681e-01 -3.58285815e-01 -1.34850919e-01 4.70257670e-01
6.96304858e-01 8.68915975e-01 6.84800595e-02 -9.14075136e-01
4.28133398e-01 6.13665521e-01 3.76653880e-01 -8.88187647e-01
-4.41312641e-01 2.06885770e-01 -5.23515761e-01 6.92507565e-01
6.75817668e-01 -3.14572901e-01 -8.55508327e-01 1.90045106e+00
7.77831256e-01 2.64529884e-01 -5.64300358e-01 1.38028765e+00
8.20992708e-01 2.38859251e-01 -1.00347258e-01 -3.47609133e-01
1.47570527e+00 -1.07588196e+00 -8.51026714e-01 -4.60781723e-01
6.85774535e-02 -7.04497278e-01 1.25293291e+00 5.32504916e-01
-1.25422788e+00 -6.75019860e-01 -7.76260018e-01 -3.63610119e-01
5.05239069e-02 2.72480845e-01 5.48375964e-01 1.05964839e+00
-1.47024524e+00 2.99955737e-02 -7.17719078e-01 -4.93131995e-01
5.62948585e-01 6.71077132e-01 -3.70368570e-01 3.31069045e-02
-5.42675972e-01 6.99581087e-01 -3.26464415e-01 3.14517826e-01
-5.66197217e-01 -6.75448895e-01 -9.66156304e-01 9.00940895e-02
4.87966329e-01 -7.53201663e-01 1.57466578e+00 -1.33862114e+00
-1.75148380e+00 1.19591463e+00 -1.09078395e+00 2.52519429e-01
5.00795364e-01 -4.55891520e-01 -3.07588875e-01 -1.87743697e-02
-1.12323664e-01 1.13576126e+00 1.26866484e+00 -1.22301960e+00
-8.76809239e-01 -7.63301432e-01 2.55422592e-01 3.55494529e-01
-2.44992211e-01 5.26341498e-01 -6.28995836e-01 1.54518172e-01
-2.43977621e-01 -9.71512973e-01 3.27380568e-01 5.48917830e-01
-4.19545412e-01 -4.76125777e-01 8.34135711e-01 -3.94448847e-01
1.15523362e+00 -1.90516996e+00 1.16466001e-01 -2.07053833e-02
7.74299741e-01 2.38768518e-01 1.49724158e-02 -1.55889466e-01
-4.49462458e-02 -1.80608138e-01 1.91665843e-01 -1.08675587e+00
8.97805300e-03 -3.92856508e-01 5.23866266e-02 5.09248435e-01
2.22060561e-01 9.76147175e-01 -7.96506286e-01 -6.62788808e-01
1.72001258e-01 6.13865197e-01 -4.98397321e-01 4.22443539e-01
-7.10895956e-02 5.64742327e-01 -1.27137974e-02 7.07822204e-01
1.00041437e+00 -5.81669748e-01 2.53861323e-02 -9.78076365e-03
-1.83353886e-01 -8.67120177e-02 -8.93264830e-01 1.63488245e+00
-3.84534419e-01 1.01778615e+00 1.26728371e-01 5.45653813e-02
6.57455087e-01 -2.42838431e-02 1.05426081e-01 -7.24376798e-01
5.41086555e-01 -1.65600225e-01 8.49301089e-03 -5.78033447e-01
6.07577562e-01 2.79543906e-01 3.06276739e-01 9.43909168e-01
1.68930460e-02 4.61084276e-01 1.90424263e-01 -1.54997289e-01
4.36210901e-01 3.63725305e-01 1.31529316e-01 2.04043952e-03
6.84231043e-01 -6.47997916e-01 3.25033039e-01 2.47868404e-01
-6.23725176e-01 8.80311608e-01 4.55741316e-01 -3.74729067e-01
-7.59531498e-01 -7.97317326e-01 1.68428987e-01 1.55177629e+00
4.82080758e-01 -3.35964739e-01 -1.25651014e+00 -7.20817268e-01
-2.20003784e-01 3.28228682e-01 -8.88235927e-01 4.22244400e-01
-4.43114936e-01 -2.42180407e-01 2.19568536e-02 2.72456080e-01
3.75210226e-01 -1.04800117e+00 -1.02453315e+00 -5.25893807e-01
-2.84057915e-01 -9.46483493e-01 -1.02059531e+00 -8.35334897e-01
-3.09376240e-01 -1.29619896e+00 -1.00217581e+00 -5.83366036e-01
9.22035277e-01 6.57978833e-01 1.21025741e+00 3.54881465e-01
5.18962778e-02 2.27808177e-01 -7.91362077e-02 -6.94280088e-01
1.03010789e-01 1.62154824e-01 3.31083119e-01 5.27909040e-01
9.04250264e-01 -1.59189209e-01 -9.85658884e-01 4.29468870e-01
-2.93829679e-01 2.00593591e-01 3.94353718e-01 6.14432216e-01
2.82460004e-01 -6.52529418e-01 3.19111317e-01 -7.00165093e-01
6.49091005e-01 -1.23249695e-01 -9.12414849e-01 4.66426790e-01
-7.63861239e-01 -1.80999026e-01 -5.70137687e-02 -4.32551235e-01
-1.24015069e+00 1.18810810e-01 6.34122267e-02 -7.46232092e-01
-3.29899520e-01 -4.44726981e-02 -1.02875821e-01 -2.52554744e-01
8.18738997e-01 -1.50635660e-01 4.83473003e-01 -3.93427402e-01
1.91384524e-01 1.01545715e+00 4.06889141e-01 -7.01984689e-02
5.14333487e-01 4.51127112e-01 -4.31170724e-02 -6.45858765e-01
-1.04517174e+00 -4.53612089e-01 -8.52346361e-01 -7.13305473e-01
7.26112664e-01 -1.09212279e+00 -1.58825684e+00 9.52782929e-01
-1.14227104e+00 -3.01938742e-01 5.50846934e-01 1.05322637e-01
-4.58370656e-01 1.11916535e-01 -5.57616986e-02 -1.06078494e+00
-7.25160301e-01 -1.23552740e+00 1.58039880e+00 8.28673363e-01
-1.64532691e-01 -6.35809720e-01 -1.34195155e-02 3.39604259e-01
3.45813453e-01 -9.43536032e-03 -8.68945792e-02 9.33527648e-02
-6.86809480e-01 -1.75606944e-02 -5.07252932e-01 -9.03564617e-02
7.22249746e-02 3.18517089e-01 -1.37404478e+00 -3.59959394e-01
-4.63411778e-01 -4.35929388e-01 4.57720727e-01 5.41580856e-01
1.02323902e+00 -7.10672960e-02 -4.90957469e-01 6.11543596e-01
1.07396674e+00 -3.79671872e-01 5.48064709e-01 -6.13125637e-02
9.67766762e-01 8.05767834e-01 7.75842369e-01 3.18956256e-01
8.54275107e-01 8.57535362e-01 5.41522622e-01 -2.07331821e-01
-1.62733123e-01 -1.51567921e-01 2.93526173e-01 1.11338653e-01
-2.23435670e-01 -4.12202954e-01 -9.60061252e-01 3.77146333e-01
-1.74862134e+00 -1.01042366e+00 -3.22124511e-01 2.30341649e+00
7.59633005e-01 -1.81950673e-01 6.23210609e-01 -9.49214622e-02
7.95356691e-01 -6.67930916e-02 -7.75605023e-01 -1.15946196e-01
3.23522568e-01 1.35459946e-02 2.02301562e-01 3.22993279e-01
-9.65581119e-01 8.49049747e-01 6.78340769e+00 1.30884796e-01
-1.53814328e+00 2.21885055e-01 5.70136130e-01 -5.92660725e-01
4.62320596e-02 -4.08382088e-01 -1.10255492e+00 7.47196794e-01
9.07749712e-01 1.58197880e-01 5.83376586e-01 6.34928763e-01
2.50990778e-01 -4.91530567e-01 -1.23412538e+00 1.59364450e+00
6.49136484e-01 -8.70351076e-01 -5.74438155e-01 3.59489262e-01
4.61567461e-01 8.53849724e-02 3.76427352e-01 -1.12089425e-01
-1.87499747e-01 -1.03676522e+00 9.43346798e-01 8.04227889e-01
1.44813716e+00 -6.64317787e-01 4.12842572e-01 3.89299035e-01
-9.01248038e-01 -7.61239529e-02 -2.01289658e-03 -1.60687655e-01
3.58667910e-01 -4.01421858e-04 -8.31301093e-01 1.33465184e-02
9.30848897e-01 7.37072945e-01 -1.03038406e+00 1.04732358e+00
-3.08746368e-01 1.60175279e-01 -4.17412072e-01 -2.00429305e-01
-3.44166517e-01 4.42083217e-02 2.45724961e-01 6.11099064e-01
3.04851234e-01 -1.22506000e-01 -3.39220583e-01 9.32797253e-01
-1.65202394e-01 -2.94931948e-01 -3.33328903e-01 4.22934920e-01
6.11496031e-01 1.35229504e+00 -3.33944499e-01 3.31119113e-02
-4.83128726e-01 1.08071673e+00 5.28852403e-01 3.79980147e-01
-7.99749136e-01 -2.88097173e-01 1.10828090e+00 4.50473100e-01
2.14699522e-01 1.68858487e-02 -1.17736436e-01 -1.17771173e+00
1.66637287e-01 -8.41049135e-01 -4.20758873e-02 -1.29316378e+00
-1.10912621e+00 5.80584943e-01 -1.62786528e-01 -1.22914255e+00
-6.04583204e-01 -7.86996126e-01 -5.52203119e-01 1.26818669e+00
-1.57599556e+00 -1.62038064e+00 -1.02329993e+00 7.67167270e-01
4.26194876e-01 -9.92226675e-02 7.31163740e-01 1.20891511e-01
-9.32101429e-01 1.06771064e+00 -3.45479459e-01 -2.13778354e-02
1.20977402e+00 -1.16012716e+00 6.61069989e-01 9.19730544e-01
5.51298857e-02 7.38841236e-01 6.18602335e-01 -2.09574938e-01
-1.29590607e+00 -5.55014431e-01 1.06743050e+00 -9.71242845e-01
1.96742583e-02 -6.29898906e-01 -6.52637720e-01 7.28008449e-01
5.46348095e-01 9.26792771e-02 6.44935012e-01 5.44553161e-01
-2.16516018e-01 -3.61467004e-01 -1.01623404e+00 7.48615384e-01
1.08555233e+00 -4.68661070e-01 -1.56759590e-01 -1.06577836e-01
2.67620772e-01 -8.75907779e-01 -3.90501082e-01 4.91515808e-02
9.51080024e-01 -1.37220228e+00 8.42474639e-01 -2.39227474e-01
2.41000816e-01 -4.35676605e-01 5.80867469e-01 -1.02630985e+00
-9.14687067e-02 -8.08469951e-01 -5.91910362e-01 1.13075340e+00
1.02410451e-01 -4.99380946e-01 8.04933369e-01 9.72598910e-01
5.42995811e-01 -6.22514606e-01 -5.21699488e-01 -1.25310168e-01
-4.52287674e-01 -1.65741295e-01 8.73063147e-01 4.47345942e-01
1.40189063e-02 6.11248255e-01 -7.02190757e-01 1.52478337e-01
9.24252868e-01 8.64034370e-02 1.54577899e+00 -1.52903938e+00
2.23083034e-01 -4.00140047e-01 -4.44437027e-01 -1.33022296e+00
2.51867861e-01 7.85302222e-02 1.93896964e-01 -1.07438016e+00
2.57106483e-01 -2.18448430e-01 -8.22821409e-02 4.00432110e-01
-6.12829745e-01 5.91858447e-01 1.91833213e-01 2.54916191e-01
-8.62295926e-01 2.65763849e-01 1.31446517e+00 2.92287707e-01
-2.33574644e-01 1.32107377e-01 -8.79243314e-01 5.32423854e-01
3.91414970e-01 -1.76229492e-01 -4.23922002e-01 -6.48404658e-01
3.65073085e-01 -2.98226565e-01 5.08510292e-01 -8.68882000e-01
5.44347405e-01 -1.59516364e-01 6.34779036e-01 -8.15506816e-01
4.52026069e-01 -6.51542723e-01 -1.24805585e-01 -2.57253975e-01
3.35602276e-02 2.48685107e-03 1.09185681e-01 5.27732491e-01
2.32653208e-02 3.66883166e-02 7.93024659e-01 2.64196187e-01
-6.60998642e-01 6.21599317e-01 2.71383554e-01 -1.12429574e-01
1.14383495e+00 -3.86339068e-01 -4.23282713e-01 -4.86999542e-01
-3.10809761e-01 3.45552564e-01 9.80141521e-01 6.80259228e-01
5.51338434e-01 -1.09379351e+00 -5.81149697e-01 5.79740822e-01
3.37770969e-01 1.40268892e-01 3.21776479e-01 1.02764797e+00
-4.30748403e-01 2.46007964e-01 -4.23129171e-01 -1.08429611e+00
-1.74951851e+00 5.47322989e-01 4.14583862e-01 5.26815295e-01
-1.46837547e-01 1.10282505e+00 3.72721165e-01 -2.39208620e-02
1.73241168e-01 8.90240353e-03 -4.15409774e-01 6.58655390e-02
1.18444288e+00 1.30761042e-01 -7.61273056e-02 -1.23100853e+00
-3.64378303e-01 7.49343038e-01 -1.92303762e-01 -5.48958518e-02
1.00617170e+00 -8.29539061e-01 -5.52913034e-03 3.43647093e-01
8.34164560e-01 6.41115010e-02 -1.74025977e+00 -2.62041122e-01
-5.09485722e-01 -1.04251373e+00 7.64432475e-02 -7.01515138e-01
-1.15496111e+00 1.08735299e+00 8.49042714e-01 1.79374963e-02
1.40445864e+00 -6.28756434e-02 6.20983005e-01 5.12061380e-02
3.54194820e-01 -7.42028713e-01 -1.95033066e-02 1.64151683e-01
7.38230288e-01 -1.83082235e+00 -8.90664831e-02 -3.47943276e-01
-8.04900825e-01 9.07953739e-01 1.12401819e+00 3.53925496e-01
6.52187228e-01 -1.74406216e-01 4.75308865e-01 -3.33671182e-01
-7.26191103e-01 -5.27316034e-01 6.88896358e-01 8.02618444e-01
4.47894037e-01 -2.50107437e-01 3.75904649e-01 2.51010656e-01
-2.65768200e-01 2.25007921e-01 3.62666100e-01 5.61944783e-01
-1.75252184e-01 -9.73136485e-01 -5.44230640e-01 2.64558583e-01
-4.29907739e-01 4.21598852e-02 -3.71654809e-01 5.57639003e-01
7.86337405e-02 1.10890663e+00 4.82117474e-01 -2.98423648e-01
1.70234919e-01 2.76977066e-02 8.17208946e-01 -4.55728292e-01
-2.85410762e-01 -5.44839650e-02 -3.41821015e-01 -8.34965587e-01
-7.05388546e-01 -7.75438011e-01 -5.73850513e-01 -7.45430887e-01
-4.66170073e-01 -5.17701447e-01 6.33114398e-01 7.77855337e-01
7.63352931e-01 1.30936295e-01 6.12443924e-01 -1.54191971e+00
-4.06429432e-02 -1.31506538e+00 -5.00739694e-01 3.56824160e-01
8.29008222e-01 -1.09261012e+00 -1.24253318e-01 4.77240771e-01] | [14.115507125854492, 0.07170946896076202] |
15b8e49e-4016-4705-88bc-69daf5579589 | zero-shot-federated-learning-with-new-classes | 2106.10019 | null | https://arxiv.org/abs/2106.10019v1 | https://arxiv.org/pdf/2106.10019v1.pdf | Zero-Shot Federated Learning with New Classes for Audio Classification | Federated learning is an effective way of extracting insights from different user devices while preserving the privacy of users. However, new classes with completely unseen data distributions can stream across any device in a federated learning setting, whose data cannot be accessed by the global server or other users. To this end, we propose a unified zero-shot framework to handle these aforementioned challenges during federated learning. We simulate two scenarios here -- 1) when the new class labels are not reported by the user, the traditional FL setting is used; 2) when new class labels are reported by the user, we synthesize Anonymized Data Impressions by calculating class similarity matrices corresponding to each device's new classes followed by unsupervised clustering to distinguish between new classes across different users. Moreover, our proposed framework can also handle statistical heterogeneities in both labels and models across the participating users. We empirically evaluate our framework on-device across different communication rounds (FL iterations) with new classes in both local and global updates, along with heterogeneous labels and models, on two widely used audio classification applications -- keyword spotting and urban sound classification, and observe an average deterministic accuracy increase of ~4.041% and ~4.258% respectively. | ['Satheesh K. Perepu', 'Gautham Krishna Gudur'] | 2021-06-18 | null | null | null | null | ['sound-classification'] | ['audio'] | [ 1.21244445e-01 -1.76648811e-01 -1.58645973e-01 -4.37420934e-01
-1.27071953e+00 -1.14922404e+00 1.70008272e-01 2.59396523e-01
-1.41724721e-01 8.19753349e-01 1.07809663e-01 -1.10354714e-01
-1.80741683e-01 -6.24280930e-01 -6.95816934e-01 -7.82765746e-01
-2.30733603e-01 5.06110251e-01 2.95180883e-02 3.24913293e-01
-3.50836605e-01 4.35353339e-01 -1.41586792e+00 4.20269489e-01
6.82244658e-01 1.30606425e+00 -4.03383851e-01 6.28818691e-01
-3.01465020e-02 4.99854773e-01 -8.75260353e-01 -6.81877017e-01
5.63923955e-01 -1.23679467e-01 -6.05342090e-01 5.39264530e-02
2.51978278e-01 -4.42190677e-01 -2.57024676e-01 1.18354487e+00
7.70315468e-01 3.61218373e-03 3.79957348e-01 -1.41933322e+00
-2.81533152e-01 9.29115474e-01 -4.61241275e-01 -1.32575959e-01
3.27074170e-01 -5.72178960e-02 6.77666724e-01 -2.82966465e-01
4.85345840e-01 9.86387074e-01 6.02861643e-01 4.03151691e-01
-1.20167649e+00 -1.29831302e+00 2.85009623e-01 7.59061798e-02
-1.60969996e+00 -6.70669496e-01 8.04377019e-01 -2.60349721e-01
-1.10424221e-01 8.19198608e-01 -2.10364535e-02 1.17119372e+00
-3.45870823e-01 6.62293494e-01 9.88126338e-01 1.19156606e-01
6.68958664e-01 6.95132852e-01 -2.04627011e-02 -2.29884647e-02
1.19504988e-01 -2.50140339e-01 -7.85743952e-01 -9.42456603e-01
-1.67326033e-01 3.68129343e-01 -2.29438320e-01 -3.80688339e-01
-1.11859787e+00 4.47750151e-01 -9.78477895e-02 -4.56297249e-02
-2.70858109e-01 -1.69588223e-01 6.69585168e-01 5.97974896e-01
6.12068772e-01 -1.60568118e-01 -9.26624537e-01 -2.82329100e-04
-7.56777287e-01 6.75334036e-02 8.59722435e-01 1.27772951e+00
7.83351600e-01 -2.02927321e-01 -3.66949320e-01 7.08424866e-01
-7.05605522e-02 7.18090832e-01 6.23616576e-01 -8.71882260e-01
6.70493126e-01 2.26049095e-01 5.70938408e-01 -1.19221556e+00
-5.93425110e-02 -6.89739466e-01 -1.07078743e+00 -3.61234039e-01
3.01887661e-01 -6.97404027e-01 -1.69786096e-01 2.02716875e+00
8.08872223e-01 5.47559679e-01 9.52384546e-02 5.30176818e-01
6.10710919e-01 3.04008633e-01 -2.69777298e-01 -5.72260916e-01
1.24150157e+00 -5.16800046e-01 -1.01544523e+00 4.41789299e-01
5.87131023e-01 -6.77726448e-01 7.31407523e-01 5.91747344e-01
-8.18979383e-01 -2.58824438e-01 -7.35059917e-01 5.11812747e-01
-4.05951262e-01 1.10091105e-01 4.35679585e-01 1.12208819e+00
-7.85097182e-01 4.36781436e-01 -6.55432701e-01 -1.80945024e-01
8.49066079e-01 7.06004143e-01 -3.29227328e-01 -1.29690632e-01
-1.02940428e+00 -4.86176640e-01 -3.20091546e-02 -3.64308834e-01
-8.56232166e-01 -7.99189448e-01 -2.46047959e-01 3.82519253e-02
4.50947613e-01 -5.36859334e-01 1.17936563e+00 -6.71930909e-01
-1.26432347e+00 5.70973635e-01 -3.22553307e-01 -4.31837380e-01
7.12649226e-01 1.21409275e-01 -9.99404252e-01 -1.75519764e-01
8.37005302e-02 -1.49682954e-01 1.11395085e+00 -1.43611968e+00
-1.11407959e+00 -7.60590613e-01 -8.88034329e-02 -6.61722943e-02
-7.93512344e-01 -1.09457947e-01 -3.47381651e-01 -7.89536774e-01
7.36687407e-02 -9.86851692e-01 8.62922817e-02 -8.32891150e-04
-8.56293976e-01 1.25555499e-02 1.24633420e+00 -6.22555614e-01
1.34056926e+00 -2.39993072e+00 -4.11281824e-01 4.41911668e-01
4.32120562e-01 8.32974166e-02 2.50274777e-01 2.60984063e-01
2.28050545e-01 1.79574788e-01 -5.03407270e-02 -6.99407637e-01
1.15868174e-01 5.43341115e-02 -4.69227523e-01 4.75339562e-01
-6.49139047e-01 3.73402089e-01 -8.28031778e-01 -1.02877706e-01
-3.42021883e-02 1.95793390e-01 -4.53381866e-01 3.04673284e-01
1.14352060e-02 7.68255353e-01 -5.83566666e-01 6.85966849e-01
1.16369355e+00 -2.05002978e-01 5.38650155e-01 -2.11204618e-01
4.40025061e-01 1.44890085e-01 -1.70904684e+00 1.69367313e+00
-6.44637764e-01 5.33112288e-02 6.19418204e-01 -7.38108158e-01
6.80097342e-01 6.38729572e-01 6.24806285e-01 -2.23024935e-01
2.60456353e-01 1.51493818e-01 -5.27610898e-01 -2.68935055e-01
1.44395575e-01 1.29574239e-01 -3.90391558e-01 1.17799056e+00
-2.74879802e-02 6.99567258e-01 -5.09643078e-01 1.05277732e-01
1.35550523e+00 -8.05988133e-01 5.49337305e-02 -3.22473794e-02
3.28589469e-01 -7.11732209e-01 7.28425026e-01 1.06366491e+00
-2.69774288e-01 4.48053718e-01 1.18261889e-01 -3.29532087e-01
-4.50720996e-01 -1.16364384e+00 -1.12068087e-01 1.24575496e+00
2.89844394e-01 -3.92697096e-01 -5.81687510e-01 -1.07556748e+00
2.38875896e-01 4.17021334e-01 -4.73043472e-01 -1.21238306e-01
8.39778408e-02 -6.17340803e-01 7.63734162e-01 -7.27310851e-02
5.79216063e-01 -5.32209098e-01 -2.57092714e-01 1.89933822e-01
-3.82088095e-01 -1.08499551e+00 -6.56274140e-01 5.20217754e-02
-5.00834942e-01 -8.63437533e-01 -2.04512775e-01 -4.78010476e-01
4.68174577e-01 3.64743084e-01 6.12968028e-01 -4.20203596e-01
-2.13732128e-03 5.34267902e-01 -2.80030608e-01 -3.07587892e-01
-3.51634949e-01 3.05358142e-01 5.28436303e-01 1.08368361e+00
2.86562830e-01 -1.02512622e+00 -5.99603593e-01 6.53155327e-01
-9.49288785e-01 -5.52305400e-01 -1.49294034e-01 4.86793160e-01
5.21097541e-01 5.70930362e-01 8.16274762e-01 -1.26342320e+00
6.37734830e-01 -9.48519886e-01 -1.54856741e-01 3.26395094e-01
-2.95077682e-01 -2.28314430e-01 9.04374242e-01 -7.48620927e-01
-8.08298707e-01 1.95959732e-01 1.60056949e-02 -4.14019883e-01
-5.60870469e-01 -2.39450440e-01 -6.75487161e-01 -2.17958670e-02
4.80682641e-01 3.26731205e-02 -3.49048913e-01 -8.39286625e-01
4.73544836e-01 1.50426555e+00 6.56455457e-01 -5.84831834e-01
1.09007573e+00 7.91445136e-01 -5.56277335e-01 -4.80192453e-01
-3.65209848e-01 -5.42331457e-01 -3.00549775e-01 1.93317443e-01
3.52960348e-01 -1.31855929e+00 -8.81926477e-01 6.90320075e-01
-9.43720043e-01 1.91962019e-01 -5.35245359e-01 3.90633881e-01
-1.21606782e-01 3.32156152e-01 -3.22876126e-01 -8.99612725e-01
-5.07199824e-01 -1.01481104e+00 1.06395841e+00 8.32387507e-02
-1.89016536e-01 -8.34400356e-01 -2.74179190e-01 3.79549086e-01
4.20749009e-01 3.13547164e-01 8.06107104e-01 -1.15216839e+00
-3.63517970e-01 -4.95763540e-01 1.13191731e-01 1.11296326e-01
7.86775649e-01 -6.30943775e-01 -1.53760755e+00 -7.09526002e-01
1.06603719e-01 -2.51724809e-01 -1.07455269e-01 -1.02209441e-01
1.58937371e+00 -9.54572201e-01 -4.93201137e-01 7.00065672e-01
9.43217635e-01 2.00867906e-01 -2.11930946e-02 -3.52303922e-01
6.63352191e-01 4.39927191e-01 4.41071928e-01 1.04842556e+00
4.13614929e-01 8.29702735e-01 3.66160572e-01 1.15858965e-01
2.19010949e-01 -4.94745165e-01 1.20715611e-01 9.32540476e-01
8.82423043e-01 -4.33301926e-01 -3.08442920e-01 5.42137980e-01
-1.75223589e+00 -5.90250134e-01 2.55043715e-01 2.74707627e+00
7.92831421e-01 -2.22552240e-01 2.55909979e-01 3.46834749e-01
1.00636554e+00 -7.36516640e-02 -1.08653998e+00 -3.96733768e-02
-1.17301397e-01 1.04462266e-01 6.98598325e-01 1.99969739e-01
-1.00071847e+00 4.23626870e-01 4.66254520e+00 1.14501667e+00
-1.10722029e+00 6.82939529e-01 9.11518157e-01 -4.96594697e-01
-3.26198161e-01 -2.15819269e-01 -5.52162707e-01 8.94097328e-01
8.61124396e-01 -2.32406259e-01 6.83561623e-01 7.44699776e-01
7.63998479e-02 3.55450541e-01 -1.10976243e+00 1.49290335e+00
-1.55888960e-01 -1.18351078e+00 -9.76231545e-02 1.63960725e-01
8.09894383e-01 6.33478761e-02 3.56723249e-01 1.13602065e-01
5.86798310e-01 -4.94983613e-01 6.38051212e-01 3.47221166e-01
1.17940116e+00 -9.62550759e-01 4.91802752e-01 7.02575445e-01
-1.26441300e+00 -4.70446765e-01 -4.96384613e-02 2.11804599e-01
-9.91928726e-02 7.53288567e-01 -5.77355266e-01 7.12198794e-01
1.08443356e+00 3.92526925e-01 -5.59278846e-01 8.02066684e-01
4.54081595e-01 8.00694525e-01 -7.04844236e-01 2.97637641e-01
-3.83862138e-01 8.59488025e-02 6.49671555e-01 6.51103914e-01
4.71748918e-01 -9.32795405e-02 2.30537608e-01 4.29717958e-01
-6.52659535e-01 3.51395875e-01 -6.11112237e-01 4.09610271e-01
1.20978916e+00 1.20371687e+00 -3.65034133e-01 -2.57040322e-01
-9.71733406e-02 1.20783687e+00 8.87337956e-04 5.06331444e-01
-6.79692388e-01 -6.38599992e-01 1.12703478e+00 2.22525701e-01
2.70088255e-01 2.67760396e-01 -1.95923641e-01 -1.20951009e+00
2.92081237e-01 -9.61273909e-01 5.95380962e-01 -3.31041485e-01
-1.59431314e+00 5.06912410e-01 -3.16791266e-01 -1.36382377e+00
-1.41480476e-01 2.66239852e-01 -4.41561669e-01 5.36465466e-01
-9.39281523e-01 -1.06873870e+00 -1.51821017e-01 1.22282994e+00
6.82338029e-02 -3.51701558e-01 1.04917157e+00 8.35308671e-01
-5.23314655e-01 1.49294996e+00 7.57307053e-01 4.40543517e-02
8.43775928e-01 -7.99677253e-01 1.27160370e-01 6.23259187e-01
4.04960841e-01 4.83977765e-01 2.94056654e-01 -3.99902910e-01
-1.34651923e+00 -1.68202090e+00 5.56601703e-01 -4.07527804e-01
2.99882799e-01 -9.88838494e-01 -8.01694453e-01 5.54163516e-01
-2.70705014e-01 5.29121935e-01 1.20739627e+00 3.15906480e-02
-4.80078191e-01 -7.20290065e-01 -1.82895720e+00 3.08904976e-01
1.02269971e+00 -8.56399715e-01 3.46105278e-01 6.00136697e-01
1.04207575e+00 -2.17865795e-01 -9.71827149e-01 1.62130311e-01
4.71171528e-01 -8.42718303e-01 7.17025220e-01 -6.35019600e-01
-7.23906338e-01 -4.72750962e-01 -5.57357371e-01 -1.11913741e+00
8.31781179e-02 -1.41819251e+00 -4.62944776e-01 2.06191206e+00
3.32913697e-01 -9.04890835e-01 8.20941269e-01 7.28482902e-01
5.64813972e-01 -3.35617185e-01 -1.33023489e+00 -4.83722985e-01
-2.94007659e-01 -7.42409706e-01 1.49790037e+00 1.11057532e+00
-1.50901571e-01 7.69294873e-02 -7.06542909e-01 5.31242847e-01
8.15572739e-01 5.15366346e-02 1.06770277e+00 -1.29757953e+00
-4.80999798e-01 2.70849735e-01 -2.72733599e-01 -8.63843322e-01
1.52790919e-01 -8.48978877e-01 -4.18445677e-01 -5.46915233e-01
-9.46235135e-02 -1.14679301e+00 -5.66619277e-01 5.78286231e-01
3.06821585e-01 2.77529895e-01 1.70616377e-02 2.88972765e-01
-9.19703066e-01 3.93631101e-01 3.76418442e-01 -2.69392222e-01
-4.27418649e-01 7.52993822e-01 -9.81328309e-01 3.69246215e-01
9.53447223e-01 -8.13315570e-01 -6.37712002e-01 -3.71941000e-01
-1.40702575e-02 1.31381899e-01 1.13063119e-01 -1.17905223e+00
2.18314931e-01 1.99003711e-01 9.85321328e-02 -4.90093768e-01
-2.69141365e-02 -1.43866062e+00 5.53021789e-01 -2.13604383e-02
-4.32491958e-01 -3.19656998e-01 6.48428351e-02 1.00453031e+00
1.16842948e-01 3.20815057e-01 4.61426646e-01 2.05086932e-01
4.05077077e-02 7.03099787e-01 -2.25805119e-01 1.46597341e-01
1.17936945e+00 -3.36195119e-02 -3.92832141e-03 -7.01713502e-01
-8.88310730e-01 2.99407721e-01 4.92357105e-01 5.09391010e-01
1.71926737e-01 -1.48190641e+00 -3.23794603e-01 3.93588841e-01
1.95931822e-01 1.12019576e-01 6.46838605e-01 5.51655650e-01
4.07910109e-01 9.20679793e-02 2.95803279e-01 -5.65668881e-01
-1.41482508e+00 6.56793177e-01 1.20720051e-01 -4.16633748e-02
-3.41346055e-01 6.59500659e-01 1.13975234e-01 -8.62075865e-01
7.08541214e-01 -1.49547383e-01 2.50297487e-01 2.74261773e-01
5.68203330e-01 6.42365754e-01 4.82879758e-01 -6.81590557e-01
-2.63899118e-01 2.56567389e-01 1.40185766e-02 1.45827532e-01
8.92007351e-01 -7.78180838e-01 1.73265487e-01 6.17927253e-01
1.68474531e+00 4.68154579e-01 -1.05603945e+00 -8.95448864e-01
-4.65443552e-01 -6.20518386e-01 -2.42402151e-01 -6.93425536e-01
-1.40835106e+00 5.47553837e-01 1.15828562e+00 2.95398921e-01
1.22426820e+00 -3.95920873e-02 1.20963800e+00 2.71224007e-02
8.97971213e-01 -8.58112335e-01 -5.04251242e-01 -2.23959357e-01
2.59863622e-02 -1.06612933e+00 -3.23009729e-01 -2.39556298e-01
-4.90719169e-01 4.52980429e-01 1.21960089e-01 6.04962945e-01
1.08872163e+00 2.44372934e-01 1.41133562e-01 2.44406119e-01
-7.32908785e-01 3.99232924e-01 -8.45096707e-02 8.20488811e-01
-2.99095005e-01 4.08619851e-01 3.19145143e-01 1.25588357e+00
-3.51877302e-01 -1.22548208e-01 3.04891855e-01 8.92258763e-01
1.46874055e-01 -1.17888951e+00 -4.77939069e-01 8.35480750e-01
-7.22052693e-01 3.15878689e-01 -1.54755488e-01 -1.39750063e-01
4.59203303e-01 1.32549775e+00 -5.23816794e-03 -8.47111404e-01
1.08409807e-01 1.38132006e-01 -2.97013879e-01 -4.64405954e-01
-7.52727568e-01 -1.08914919e-01 -2.88365871e-01 -5.19671261e-01
-6.19237348e-02 -6.92658722e-01 -9.67926025e-01 -3.52023989e-01
-2.58978695e-01 5.25193214e-01 6.48235381e-01 7.18423724e-01
8.83863807e-01 2.13525787e-01 1.56562173e+00 -3.67923021e-01
-7.37538517e-01 -5.14646292e-01 -1.11340058e+00 4.75159615e-01
4.82720494e-01 -4.21116531e-01 -6.71179950e-01 -3.04548759e-02] | [5.87308406829834, 6.297885894775391] |
0a7a0135-3e88-4d6e-b5df-730a9833579b | long-term-stock-prediction-based-on-financial | null | null | http://cs230.stanford.edu/projects_winter_2021/reports/70728801.pdf | http://cs230.stanford.edu/projects_winter_2021/reports/70728801.pdf | Long Term Stock Prediction based on Financial Statements | This paper proposes a model with LSTM and fully connected layers to predict long term stock trendings based on financial statements. Two data augmentation techniques are applied on structured data: 1) adding random noise to data fields; 2) erasing partial information from training examples. The performance of the proposed approach is demonstrated on real-world data of about 7,000 stocks listed on Nasdaq. | ['Shujia Liu'] | 2021-11-01 | null | null | null | journal-2021-11 | ['stock-prediction'] | ['time-series'] | [-4.10778850e-01 1.98490947e-01 -2.57490277e-01 -7.26423264e-01
-2.79341429e-01 -3.64058256e-01 3.46577942e-01 2.18595594e-01
-6.48126006e-01 8.91627967e-01 5.13647735e-01 -8.15129399e-01
2.18493417e-01 -1.26874566e+00 -7.08503962e-01 -5.25311470e-01
-9.03237343e-01 9.62884575e-02 -2.80757882e-02 -2.28598982e-01
5.70456266e-01 4.99825090e-01 -1.25069642e+00 4.71797943e-01
3.46617073e-01 1.60051179e+00 -2.85149276e-01 2.57889330e-01
-7.55986571e-01 1.75907362e+00 -7.54187465e-01 -4.51232284e-01
8.83705795e-01 3.24935049e-01 -5.84990084e-01 7.05591142e-02
4.82731879e-01 -7.60287642e-01 -3.17943782e-01 1.29851627e+00
9.04052034e-02 1.38940841e-01 -7.11403266e-02 -7.96842277e-01
-9.77260768e-01 1.21213925e+00 -6.53137326e-01 9.78097439e-01
-4.41017866e-01 -5.57836667e-02 1.23343349e+00 -8.31570625e-01
2.81736761e-01 7.63387322e-01 9.02237415e-01 1.67082205e-01
-7.42012084e-01 -8.29568088e-01 5.33450603e-01 1.14642069e-01
-8.38456035e-01 -2.64759600e-01 9.77064312e-01 -4.62382615e-01
1.41775966e+00 -4.06211652e-02 6.35755181e-01 4.08469856e-01
4.77458566e-01 7.00077355e-01 1.04822624e+00 -6.03208058e-02
4.43399340e-01 -4.77740578e-02 9.41452980e-01 2.07918927e-01
7.14687645e-01 2.95999795e-01 -7.77973533e-01 -3.48146528e-01
7.28047132e-01 2.92629093e-01 1.66312214e-02 2.91103125e-01
-9.85793114e-01 8.82055640e-01 5.25630116e-01 5.70442975e-01
-1.13813424e+00 1.47759706e-01 6.34426236e-01 8.69616628e-01
9.42146659e-01 5.35343349e-01 -1.06052136e+00 7.09404796e-02
-1.21541750e+00 4.03463453e-01 6.32875741e-01 7.67798543e-01
4.17605311e-01 1.00608635e+00 3.87913048e-01 2.44352847e-01
7.59299621e-02 2.66841471e-01 1.11028123e+00 -1.35338396e-01
8.08296859e-01 6.81062341e-01 2.77998477e-01 -8.47480953e-01
-4.98499393e-01 -7.99954414e-01 -8.40133488e-01 4.47433621e-01
1.00724865e-02 -6.13702655e-01 -1.18069112e+00 9.89948273e-01
-4.86128509e-01 2.26723716e-01 6.37927711e-01 4.41904873e-01
1.02310860e+00 9.16267037e-01 9.35161263e-02 -1.10457458e-01
9.05434787e-01 -6.98767841e-01 -1.30231988e+00 -4.48776066e-01
4.61140543e-01 -1.12071559e-01 4.22019303e-01 6.22787476e-01
-1.28318822e+00 -4.71270084e-01 -1.14844549e+00 3.24171960e-01
-1.02122426e+00 9.06439964e-03 7.32914925e-01 3.45895767e-01
-1.10024643e+00 7.30723023e-01 -8.63300323e-01 5.60114324e-01
4.69520003e-01 5.96143842e-01 -1.07492015e-01 6.21271014e-01
-1.51351547e+00 7.39421844e-01 1.16280007e+00 5.63239276e-01
-2.95308322e-01 -6.39444232e-01 -9.96695220e-01 4.01104122e-01
4.64193821e-02 2.00005583e-02 1.37933898e+00 -9.36890721e-01
-1.13889575e+00 4.32240576e-01 2.40526676e-01 -1.41900659e+00
1.28221586e-01 -6.29746556e-01 -8.49350631e-01 -3.68662894e-01
-3.56492698e-01 1.48899481e-01 2.87260443e-01 -5.55374324e-01
-7.80470014e-01 -5.65864384e-01 -2.77386904e-01 -1.66183114e-01
-9.45419192e-01 1.40084237e-01 4.85836744e-01 -1.42904568e+00
4.72166836e-01 -1.02517225e-01 -5.20394623e-01 -7.01660991e-01
-1.48494795e-01 -2.89268434e-01 9.68569577e-01 -1.39867818e+00
1.31757641e+00 -1.66120219e+00 -8.35878670e-01 3.60834301e-01
6.60477430e-02 4.21747297e-01 2.49568839e-02 7.50075430e-02
-6.71785712e-01 1.66783258e-01 -5.00645116e-02 1.08647883e-01
-1.89746127e-01 -2.83668749e-02 -9.83815551e-01 7.60777891e-02
2.09805205e-01 1.25059676e+00 -5.25036693e-01 2.46616200e-01
-2.07524318e-02 -1.05124950e-01 2.63338536e-02 9.01809111e-02
-3.43765110e-01 -2.07458064e-01 -1.04955055e-01 6.72306538e-01
9.84143615e-01 -2.63044626e-01 -1.52695522e-01 2.95870841e-01
-3.88320863e-01 7.14981437e-01 -1.27962995e+00 1.16057074e+00
1.83106884e-01 6.07001007e-01 -4.06063080e-01 -8.65373969e-01
1.48582685e+00 5.23519397e-01 4.73695457e-01 -6.93025768e-01
6.98720440e-02 4.34700102e-01 -2.00008869e-01 -3.49258125e-01
8.03570926e-01 -5.79373956e-01 1.86641701e-02 5.19155324e-01
-4.41171601e-02 4.48414624e-01 2.35120490e-01 -1.48132592e-01
9.70276237e-01 -1.62872702e-01 1.71730638e-01 -3.57950598e-01
3.64797503e-01 9.31394249e-02 8.72996211e-01 4.19935942e-01
5.66994324e-02 1.43324852e-01 4.77907658e-01 -1.43278587e+00
-9.87981617e-01 -5.80603421e-01 4.71919365e-02 7.06503391e-01
-8.63639474e-01 2.04022169e-01 -3.42263848e-01 -7.14964271e-01
2.44241327e-01 1.02261031e+00 -9.21223819e-01 2.61389405e-01
-6.59737945e-01 -1.14829755e+00 3.45731825e-01 1.08485830e+00
1.02727067e+00 -1.46119463e+00 -7.65820324e-01 4.95872796e-01
4.95938599e-01 -8.10139835e-01 3.21877003e-02 6.94414914e-01
-1.45845377e+00 -9.33101058e-01 -7.74533391e-01 -8.34253311e-01
4.63741362e-01 -2.31260821e-01 1.29353237e+00 -1.12279460e-01
7.27793038e-01 -5.67981839e-01 -7.45581836e-02 -8.69957149e-01
-1.00420021e-01 2.49837279e-01 -8.56601447e-02 -4.24055979e-02
9.86507297e-01 -3.81710827e-01 -1.42288610e-01 -4.91273880e-01
-9.57494557e-01 -2.24739477e-01 6.91664875e-01 7.13963807e-01
2.23125100e-01 4.88409847e-01 1.06434989e+00 -1.02378249e+00
9.19886470e-01 -5.49912333e-01 -1.24007308e+00 1.63460150e-01
-9.57069218e-01 9.59484503e-02 4.22591120e-01 -1.28917456e-01
-1.28475821e+00 -5.29622845e-02 -2.13556081e-01 -3.54602575e-01
1.02752449e-04 1.40077710e+00 2.57356405e-01 2.73664743e-01
1.61177754e-01 4.77552027e-01 -1.38135254e-01 -8.74473453e-01
-1.87276274e-01 6.69018984e-01 8.72485936e-01 2.26818770e-01
5.74922383e-01 3.37407529e-01 -3.33187252e-01 -4.74485874e-01
-9.18416023e-01 -2.80324936e-01 -9.48629737e-01 1.39819458e-01
4.58287627e-01 -1.07648826e+00 -3.75274956e-01 1.22927928e+00
-8.52558255e-01 -1.11323379e-01 -6.88077569e-01 5.75144112e-01
6.10281453e-02 -1.76033564e-02 -1.12623894e+00 -1.07751250e+00
-9.01515961e-01 -1.56069234e-01 2.19571024e-01 1.48561701e-01
1.63167238e-01 -1.33613074e+00 1.10732660e-01 -2.63346821e-01
6.88913703e-01 3.27375412e-01 5.90664923e-01 -1.61512339e+00
-3.70518535e-01 -9.02124584e-01 4.94320802e-02 7.96848059e-01
2.63946891e-01 -3.69219303e-01 -9.73300159e-01 -3.72561693e-01
4.72793341e-01 -2.86525607e-01 1.18118393e+00 5.33882022e-01
1.04472017e+00 -7.21699715e-01 1.96921974e-01 4.57707226e-01
1.34724867e+00 6.30336523e-01 4.77899134e-01 8.53946507e-01
5.09079754e-01 6.07781291e-01 4.28358972e-01 3.91762316e-01
2.26590157e-01 -4.66808140e-01 3.78367275e-01 -6.27705529e-02
6.84880853e-01 -2.20955729e-01 3.08739096e-01 9.65092838e-01
2.07724702e-02 -1.78403426e-02 -1.33334994e+00 6.88504577e-01
-1.64494228e+00 -1.00068080e+00 -1.64896045e-02 2.00995302e+00
5.13261497e-01 8.67431462e-01 -5.04394881e-02 5.10453224e-01
5.63966155e-01 6.17279410e-01 -8.32613349e-01 -9.27197468e-03
-7.30252504e-01 2.20839888e-01 1.05354023e+00 4.82441522e-02
-1.46999240e+00 6.19860768e-01 7.52659321e+00 -4.63955197e-03
-1.43802953e+00 -2.57885158e-01 1.16133523e+00 -2.03820422e-01
-4.27461386e-01 -4.01655495e-01 -8.68298173e-01 4.62440193e-01
1.43989301e+00 -3.92344832e-01 -3.41205120e-01 1.04719281e+00
8.95273462e-02 3.29981178e-01 -6.06283069e-01 4.86549795e-01
-1.49548233e-01 -1.99025214e+00 2.60696262e-01 1.91252112e-01
8.55678916e-01 2.78062075e-01 1.66086167e-01 2.41561368e-01
6.92949891e-01 -7.52365947e-01 7.63586640e-01 7.96823144e-01
1.59316063e-01 -9.94613767e-01 1.56193948e+00 3.21335405e-01
-9.87049878e-01 -6.53339505e-01 -2.45506793e-01 -4.70593601e-01
1.88470446e-02 7.12162018e-01 -6.31604671e-01 4.09103572e-01
9.52813148e-01 1.15370929e+00 -6.57325566e-01 9.65051472e-01
-8.16631988e-02 9.41564918e-01 -7.14395866e-02 3.11222933e-02
6.11583591e-01 -1.25554234e-01 1.83437169e-01 8.27548981e-01
2.36840054e-01 2.25035578e-01 -2.04134703e-01 6.13667250e-01
-8.76220614e-02 7.79275075e-02 -6.18535399e-01 -2.49462456e-01
7.34578073e-02 6.16246581e-01 -6.91753924e-01 -7.87194848e-01
-7.67412543e-01 1.97266221e-01 5.87071106e-02 5.92973411e-01
-1.62500367e-01 -4.32684243e-01 2.54673660e-01 2.58868217e-01
5.68129420e-01 -1.21297836e-01 -8.58600974e-01 -1.14763999e+00
3.06132615e-01 -5.06649792e-01 9.18331921e-01 -7.43002236e-01
-1.47808707e+00 8.85617077e-01 -1.24582626e-01 -1.22730958e+00
-5.42172313e-01 -8.92984867e-01 -8.80280137e-01 1.10012734e+00
-1.80036700e+00 -8.29287529e-01 1.95426345e-01 2.32142836e-01
4.96469736e-01 -9.84492183e-01 5.67780137e-01 3.11319739e-01
-5.70575595e-01 1.77540392e-01 3.79385978e-01 8.19189072e-01
-1.18320510e-01 -1.50778925e+00 1.60374951e+00 1.25017011e+00
1.10070549e-01 7.31535316e-01 4.32777405e-01 -1.01850927e+00
-5.83438993e-01 -1.24068284e+00 1.22460151e+00 1.86649129e-01
1.11282897e+00 -6.29083812e-02 -1.59376335e+00 1.44411683e+00
5.54506898e-01 1.28934547e-01 5.59108794e-01 -1.78419262e-01
-1.51928499e-01 -2.36416250e-01 -1.28778410e+00 -6.02480248e-02
1.55335262e-01 -4.65517566e-02 -1.08378792e+00 -3.01012248e-02
6.90921187e-01 -2.02136114e-01 -8.05558681e-01 4.96911347e-01
2.02829316e-01 -6.50013387e-01 5.21788955e-01 -1.12141132e+00
1.84955806e-01 1.66656785e-02 1.81537390e-01 -1.20441127e+00
-1.81095451e-01 -5.32309413e-01 -6.49856567e-01 9.85837340e-01
4.96110439e-01 -1.05532551e+00 1.28553259e+00 1.00190496e+00
-7.72412196e-02 -7.14398146e-01 -9.24720407e-01 -7.86584258e-01
4.58999336e-01 -2.67056316e-01 1.29041159e+00 1.27122271e+00
-1.41865149e-01 -1.87285226e-02 -3.61281812e-01 9.37966853e-02
3.45258683e-01 3.00959940e-03 1.80574626e-01 -1.37551999e+00
1.92948848e-01 -4.14412498e-01 -4.41091359e-01 -7.86952138e-01
1.50963262e-01 -2.92307556e-01 -3.64883572e-01 -1.27245104e+00
-2.37263113e-01 -2.01722234e-01 -1.15237260e+00 7.56837487e-01
5.08312061e-02 -1.89724162e-01 -1.75979990e-03 3.28050882e-01
-9.08940509e-02 3.64163399e-01 4.78523195e-01 -1.68200091e-01
-2.14464307e-01 4.59821075e-01 -5.90341866e-01 9.55806136e-01
1.22173810e+00 -2.70977437e-01 -1.24274097e-01 -3.30929458e-01
4.43122387e-01 -5.54716401e-02 1.23516299e-01 -1.06705332e+00
2.87233263e-01 -2.32941657e-02 6.58847928e-01 -1.58096147e+00
-1.22692607e-01 -8.69002521e-01 -1.35087714e-01 8.83351922e-01
-6.80216193e-01 7.92578518e-01 5.64459026e-01 2.01039463e-01
-8.56602788e-01 -4.06487644e-01 3.90936375e-01 -4.52527374e-01
-1.05149341e+00 1.95066541e-01 -3.68990034e-01 -4.91133928e-01
8.13704491e-01 -6.33777007e-02 -2.86161482e-01 -3.16260040e-01
-8.22208047e-01 3.55024755e-01 -1.72077358e-01 4.90929365e-01
7.35936761e-01 -1.42356443e+00 -7.61729598e-01 7.26446152e-01
-3.23426008e-01 -1.42989336e-02 -1.14945732e-01 4.53218166e-03
-6.73392653e-01 9.26001668e-01 -4.55666482e-01 2.36994073e-01
-8.47704887e-01 7.76769161e-01 4.97196108e-01 -6.93221569e-01
-5.79640746e-01 9.66822445e-01 -4.67669964e-01 -2.14792341e-01
5.12206256e-01 -9.51365113e-01 -7.48182595e-01 4.31437105e-01
8.55619311e-01 2.57803828e-01 5.45837402e-01 -3.84862512e-01
-3.66411284e-02 -1.48638457e-01 -4.51286703e-01 -3.49926889e-01
2.15370703e+00 2.98288018e-01 -2.73370981e-01 9.59497273e-01
7.60474086e-01 -4.53379810e-01 -1.14764762e+00 -6.84162736e-01
9.01185989e-01 -3.29048514e-01 3.69014561e-01 -7.17247605e-01
-1.61156726e+00 7.31473625e-01 6.78778172e-01 6.86177611e-01
1.24639821e+00 -7.17927158e-01 1.16497159e+00 9.10505652e-01
-1.93740875e-02 -1.27487493e+00 -2.94697285e-01 6.44119263e-01
8.59234154e-01 -1.11060607e+00 3.51401679e-02 3.44440311e-01
-4.77972090e-01 1.32478976e+00 5.17418385e-01 -4.38119590e-01
1.21673787e+00 5.88055551e-01 5.85872531e-01 -5.76292276e-01
-1.00280094e+00 2.71512985e-01 1.43254787e-01 9.93053764e-02
4.72374588e-01 -4.48965222e-01 -8.15995038e-03 7.44069874e-01
-4.28909391e-01 4.24883068e-01 5.95158219e-01 1.41761088e+00
-5.27893901e-01 -5.17172515e-01 -5.32079995e-01 9.84722495e-01
-1.19376993e+00 -5.23545146e-01 -2.41299182e-01 6.57102883e-01
-2.54534483e-01 4.42196757e-01 7.49674261e-01 -3.06896716e-01
4.15916502e-01 4.98710722e-01 -7.32876420e-01 -4.62704659e-01
-1.14131653e+00 3.03244554e-02 -5.94615340e-02 2.44058259e-02
-4.86481965e-01 -9.88217950e-01 -1.28473437e+00 -9.95801836e-02
-4.44860011e-02 4.32654321e-01 5.64931750e-01 5.46486080e-01
2.92850882e-01 7.22759545e-01 6.39214635e-01 -5.30026615e-01
-1.03440225e+00 -1.44292581e+00 -1.07864082e+00 2.49855191e-01
8.56420517e-01 -4.93782796e-02 -3.23046237e-01 3.64523262e-01] | [4.46391487121582, 4.212879180908203] |
6117df5d-06c7-439e-bd19-595e5c9d745d | high-resolution-image-synthesis-with-latent | 2112.10752 | null | https://arxiv.org/abs/2112.10752v2 | https://arxiv.org/pdf/2112.10752v2.pdf | High-Resolution Image Synthesis with Latent Diffusion Models | By decomposing the image formation process into a sequential application of denoising autoencoders, diffusion models (DMs) achieve state-of-the-art synthesis results on image data and beyond. Additionally, their formulation allows for a guiding mechanism to control the image generation process without retraining. However, since these models typically operate directly in pixel space, optimization of powerful DMs often consumes hundreds of GPU days and inference is expensive due to sequential evaluations. To enable DM training on limited computational resources while retaining their quality and flexibility, we apply them in the latent space of powerful pretrained autoencoders. In contrast to previous work, training diffusion models on such a representation allows for the first time to reach a near-optimal point between complexity reduction and detail preservation, greatly boosting visual fidelity. By introducing cross-attention layers into the model architecture, we turn diffusion models into powerful and flexible generators for general conditioning inputs such as text or bounding boxes and high-resolution synthesis becomes possible in a convolutional manner. Our latent diffusion models (LDMs) achieve a new state of the art for image inpainting and highly competitive performance on various tasks, including unconditional image generation, semantic scene synthesis, and super-resolution, while significantly reducing computational requirements compared to pixel-based DMs. Code is available at https://github.com/CompVis/latent-diffusion . | ['Björn Ommer', 'Patrick Esser', 'Dominik Lorenz', 'Andreas Blattmann', 'Robin Rombach'] | 2021-12-20 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Rombach_High-Resolution_Image_Synthesis_With_Latent_Diffusion_Models_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Rombach_High-Resolution_Image_Synthesis_With_Latent_Diffusion_Models_CVPR_2022_paper.pdf | cvpr-2022-1 | ['layout-to-image-generation'] | ['computer-vision'] | [ 2.66567320e-01 8.59274864e-02 3.46433744e-02 -5.90983815e-02
-8.15553606e-01 -3.26106608e-01 8.38158548e-01 -3.31275284e-01
-3.17619294e-01 5.44293463e-01 2.82906622e-01 -7.45346844e-02
2.72388607e-01 -1.14833546e+00 -1.01309454e+00 -6.97193444e-01
3.93118858e-01 2.71673769e-01 1.86434656e-01 -2.36103803e-01
4.64998791e-03 5.98500967e-01 -1.54188538e+00 5.36318839e-01
9.45029020e-01 8.84662628e-01 5.79258561e-01 8.39178920e-01
-2.83041559e-02 1.02259660e+00 -4.73575801e-01 -3.47898990e-01
3.86763304e-01 -6.14458025e-01 -6.39508009e-01 2.90310711e-01
6.22148037e-01 -8.59697223e-01 -6.63067281e-01 9.05248821e-01
4.98218745e-01 1.86223030e-01 5.58525026e-01 -7.66244590e-01
-1.27262259e+00 4.22898561e-01 -5.75618327e-01 7.11329058e-02
5.11521846e-02 4.82154518e-01 9.92163420e-01 -1.03916681e+00
7.52852261e-01 1.20568466e+00 5.32482088e-01 7.55880237e-01
-1.70359838e+00 -3.45583439e-01 9.24873650e-02 -4.73754592e-02
-1.20671260e+00 -6.71029270e-01 6.73327684e-01 -2.63865113e-01
1.08241868e+00 1.54845148e-01 6.11453295e-01 1.21140122e+00
1.18623963e-02 8.74826193e-01 9.87342775e-01 -3.93252730e-01
2.08510250e-01 -1.52384698e-01 -4.70947772e-01 7.53377259e-01
-1.07598990e-01 1.76473752e-01 -6.17489457e-01 1.33725241e-01
1.40100586e+00 -3.73326428e-02 -3.96253556e-01 -8.29671696e-02
-1.12910151e+00 1.01235294e+00 6.56769037e-01 1.39300436e-01
-5.80661178e-01 4.39441472e-01 1.23011887e-01 2.30054811e-01
6.22054935e-01 5.08335292e-01 -1.53849587e-01 9.31321830e-02
-1.32162225e+00 3.94474834e-01 3.51072490e-01 9.03158426e-01
6.98975146e-01 4.23347682e-01 -2.42766276e-01 8.70413244e-01
-4.75103222e-02 3.51899564e-01 3.79671574e-01 -1.39392149e+00
1.84447706e-01 2.46844128e-01 1.15012191e-01 -7.48582065e-01
1.11998674e-02 -5.01101553e-01 -1.05087078e+00 5.49990773e-01
2.31779411e-01 -3.42134590e-04 -1.18808079e+00 1.64787233e+00
1.83366746e-01 8.45083743e-02 -5.67578077e-02 9.50287879e-01
6.55229449e-01 1.03659320e+00 -8.83485228e-02 -7.17335939e-02
1.24591196e+00 -1.32916689e+00 -5.88876486e-01 -3.77026528e-01
3.78151774e-01 -8.04659247e-01 1.28922760e+00 5.35310090e-01
-1.60319555e+00 -6.86704993e-01 -1.02358472e+00 -7.32341707e-01
-2.04601157e-02 1.13502489e-02 6.58239186e-01 1.69721290e-01
-1.40480673e+00 8.69983733e-01 -1.08376718e+00 1.59578607e-03
6.68439925e-01 1.51588455e-01 -1.50509551e-01 -2.57491261e-01
-8.90171230e-01 7.20002651e-01 1.56369340e-02 -6.53140619e-03
-1.12908852e+00 -9.23168480e-01 -8.89961064e-01 8.93673208e-03
1.03974797e-01 -1.16505706e+00 1.20383310e+00 -1.08868015e+00
-1.76459241e+00 7.39994049e-01 -8.26247185e-02 -6.90234721e-01
7.24920154e-01 -3.19597423e-01 -1.11159980e-01 3.52481157e-01
1.05625160e-01 1.14774764e+00 1.25360191e+00 -1.02869260e+00
-3.13537776e-01 1.58709455e-02 1.31495774e-01 1.56319842e-01
-3.29675317e-01 -8.74739438e-02 -6.37916863e-01 -9.93184805e-01
-1.16092309e-01 -7.10116088e-01 -4.49807256e-01 3.92527193e-01
-1.11518554e-01 1.94238245e-01 8.27407420e-01 -8.39131057e-01
1.01703906e+00 -2.18871784e+00 5.58064401e-01 -1.68354169e-01
4.24042076e-01 1.74441934e-01 -3.24565560e-01 2.29619950e-01
-7.25524640e-03 1.17775677e-02 -5.35473943e-01 -7.10472584e-01
-6.18369952e-02 3.51929367e-01 -5.14559150e-01 3.01451951e-01
4.51135218e-01 1.10212290e+00 -5.59646070e-01 -2.66653508e-01
3.94177407e-01 8.45448732e-01 -9.91158009e-01 3.32074225e-01
-4.03425306e-01 6.34965539e-01 -1.99445069e-01 3.55881900e-01
6.03129745e-01 -5.17597497e-01 -1.00689568e-01 -3.05397600e-01
-1.97481930e-01 2.79526204e-01 -9.70342696e-01 2.01593900e+00
-9.41577792e-01 7.38414764e-01 3.31283003e-01 -8.49181473e-01
5.69374561e-01 1.09862588e-01 2.64122903e-01 -8.78591299e-01
4.45826650e-02 1.06557094e-01 -2.96673149e-01 -4.75648753e-02
6.21906281e-01 -1.33334219e-01 2.48001456e-01 4.87443864e-01
1.03800192e-01 -5.32439172e-01 3.70837480e-01 2.54032224e-01
9.76311982e-01 3.28263998e-01 -1.82194918e-01 -1.06655747e-01
1.42226577e-01 -7.89785534e-02 2.11679205e-01 6.83681190e-01
3.51688474e-01 9.80366409e-01 3.58671486e-01 -5.30370116e-01
-1.64935172e+00 -1.13223040e+00 -1.99403331e-01 8.99559081e-01
-1.86415091e-01 -3.74805629e-01 -1.00628960e+00 -1.25497386e-01
-1.83319598e-01 8.02089512e-01 -6.71836138e-01 -4.47781906e-02
-6.50015533e-01 -6.35820925e-01 4.25727993e-01 6.43073261e-01
6.92693889e-01 -1.00985479e+00 -5.00251770e-01 3.55090797e-01
-9.64218006e-02 -1.15667653e+00 -6.01156831e-01 1.00957394e-01
-9.78108883e-01 -5.03525019e-01 -1.08574843e+00 -6.52971506e-01
7.96244740e-01 2.00656146e-01 1.28929055e+00 2.54402429e-01
-5.15057743e-01 9.02576223e-02 -7.55215390e-03 1.28235132e-01
-5.49865723e-01 1.52545739e-02 -2.03473464e-01 -1.04546279e-01
-3.81593585e-01 -8.74546409e-01 -1.07825732e+00 7.14627132e-02
-1.28812134e+00 4.26744670e-01 6.60737336e-01 1.07234359e+00
7.24084318e-01 -1.78202037e-02 8.13046768e-02 -8.43739212e-01
5.51524639e-01 -1.98797747e-01 -7.70143986e-01 -1.72109708e-01
-5.49010456e-01 3.59520525e-01 7.25291371e-01 -4.56508249e-01
-1.23983276e+00 -1.02041408e-01 -4.28754717e-01 -7.08669066e-01
7.69376233e-02 1.85812712e-01 1.88402936e-01 4.10514837e-03
8.45570445e-01 4.05569524e-01 6.56604841e-02 -4.12764400e-01
8.08327258e-01 1.95126116e-01 5.24077773e-01 -5.89394808e-01
6.40049398e-01 6.92207873e-01 -1.15728617e-01 -8.08893025e-01
-7.73128211e-01 1.73645332e-01 -4.29778785e-01 1.33363828e-01
9.76808310e-01 -1.12695742e+00 -3.45784724e-01 4.64263827e-01
-1.11058772e+00 -8.82075965e-01 -5.45723915e-01 5.11010587e-02
-6.97825253e-01 2.36449808e-01 -1.05646002e+00 -2.88267493e-01
-3.36249799e-01 -1.29837704e+00 1.18439996e+00 -1.49511576e-01
-2.72181123e-01 -9.48763430e-01 -2.57480532e-01 3.20477605e-01
6.38703585e-01 1.96721733e-01 8.81987154e-01 4.36437845e-01
-1.02152777e+00 1.23615183e-01 -2.84189999e-01 6.15794122e-01
-8.65645707e-02 2.44255029e-02 -8.93498898e-01 -1.89681619e-01
5.68214431e-02 -3.24317366e-01 1.21259344e+00 6.47142291e-01
1.32043278e+00 -3.94474179e-01 -4.47021499e-02 1.12974751e+00
1.40667844e+00 -2.48072654e-01 7.77353406e-01 2.16013134e-01
6.97709560e-01 3.25209677e-01 4.73428927e-02 4.07064289e-01
2.08345547e-01 6.13262892e-01 2.46125713e-01 -4.69670266e-01
-7.21417308e-01 -3.15928161e-01 2.35023305e-01 7.08959043e-01
-8.15695599e-02 -2.42025033e-01 -7.06412792e-01 4.70094174e-01
-1.64594734e+00 -9.52976525e-01 -8.32468271e-02 1.95875823e+00
1.02531469e+00 1.55667603e-01 -6.90758377e-02 -3.98875438e-02
4.13798481e-01 4.18728322e-01 -6.37806058e-01 -4.35497254e-01
-2.13236317e-01 5.27988851e-01 5.18077314e-01 8.43336999e-01
-8.39231610e-01 1.13546097e+00 6.38177156e+00 8.68038177e-01
-1.19929135e+00 3.03110421e-01 1.05959284e+00 -4.83684182e-01
-6.06046140e-01 -1.00925989e-01 -5.61804116e-01 3.18477213e-01
7.91924536e-01 1.54594570e-01 8.18807781e-01 8.12653899e-01
2.62923717e-01 -2.21235037e-04 -9.55987394e-01 1.02694440e+00
-1.91231385e-01 -1.82628953e+00 4.02745456e-01 -1.75963603e-02
1.15199864e+00 1.27901316e-01 4.91708696e-01 4.11177706e-03
5.33842444e-01 -1.12176943e+00 8.71850133e-01 4.34813112e-01
1.06157827e+00 -7.83087432e-01 3.06271054e-02 3.30230594e-01
-8.06150198e-01 -9.12142769e-02 -5.53766310e-01 -8.80283341e-02
3.18985492e-01 7.66306520e-01 -1.78287119e-01 1.18880652e-01
6.78206980e-01 6.11648977e-01 -1.45404950e-01 3.97876441e-01
-4.14451987e-01 4.31276590e-01 -2.45582581e-01 4.62856561e-01
4.01430875e-01 -2.29417264e-01 3.32963079e-01 1.04640520e+00
3.98522645e-01 2.65614808e-01 -1.44077197e-01 1.35590589e+00
-2.90929794e-01 -2.54830450e-01 -5.24785995e-01 -3.34693752e-02
1.72294378e-01 1.06582844e+00 -6.72784567e-01 -4.41368818e-01
-4.06896651e-01 1.56779897e+00 4.76731807e-01 5.28695524e-01
-9.32736695e-01 -1.82584047e-01 6.56368852e-01 3.25215042e-01
5.59369802e-01 -4.97541398e-01 -5.79273880e-01 -1.29047835e+00
-3.27090807e-02 -9.40939903e-01 -4.98152636e-02 -8.94164622e-01
-1.02157903e+00 8.44918728e-01 -3.35283250e-01 -9.08245802e-01
-2.47832894e-01 -4.69208390e-01 -4.28319335e-01 9.82078493e-01
-1.51805723e+00 -1.16888297e+00 -3.07346553e-01 6.24587536e-01
8.11993122e-01 1.38057366e-01 7.69850552e-01 3.51853758e-01
-4.04222310e-01 4.43484604e-01 1.25351965e-01 -1.11858740e-01
4.66789693e-01 -1.08784890e+00 8.36892128e-01 1.00853956e+00
1.75835580e-01 6.15568340e-01 6.82029426e-01 -5.15135348e-01
-1.40413225e+00 -1.04978251e+00 4.83334839e-01 -4.16218877e-01
5.47637284e-01 -4.22819167e-01 -8.78616631e-01 6.00120246e-01
4.45412308e-01 1.39855549e-01 2.74471819e-01 -1.47976071e-01
-3.33866656e-01 1.15634821e-01 -8.34855556e-01 8.71451080e-01
1.24874341e+00 -6.06391490e-01 -1.03912085e-01 3.94169211e-01
7.31762528e-01 -5.89001715e-01 -7.70794511e-01 5.97976334e-02
2.72636443e-01 -1.13618302e+00 1.17391682e+00 -1.99992210e-01
1.15743732e+00 -2.02752754e-01 1.11664943e-02 -1.19940734e+00
-5.36067963e-01 -8.83082509e-01 -3.28884542e-01 9.37121987e-01
3.72926086e-01 -3.75508726e-01 7.99740434e-01 6.31200492e-01
-2.91239053e-01 -9.08370435e-01 -5.27955472e-01 -4.42481369e-01
1.54433623e-01 -3.60540748e-01 4.73994017e-01 7.53829062e-01
-6.46348357e-01 2.26823777e-01 -4.72307295e-01 1.28295287e-01
7.14282513e-01 1.70941010e-01 6.71971262e-01 -5.77998936e-01
-7.61391878e-01 -6.97394907e-01 7.86630288e-02 -1.67294443e+00
2.61758324e-02 -7.63075113e-01 -9.07102302e-02 -1.71121538e+00
-1.10460199e-01 -2.26171449e-01 1.00103110e-01 3.65654826e-01
-1.11352690e-01 6.31273925e-01 1.38644367e-01 3.03232193e-01
-1.00355648e-01 7.68531144e-01 1.68116426e+00 -8.55101570e-02
-1.52816102e-01 -3.50299805e-01 -7.37257004e-01 7.25656450e-01
5.66438854e-01 -2.24325195e-01 -5.01484871e-01 -1.10200405e+00
2.16618642e-01 2.72052616e-01 4.79231238e-01 -8.97080898e-01
8.42107683e-02 -4.96188551e-02 6.60576701e-01 -2.76386470e-01
6.23478174e-01 -4.12261128e-01 1.00243598e-01 3.49320322e-01
-3.33670795e-01 1.20211095e-02 2.39688560e-01 3.51210535e-01
-3.62450331e-01 -1.50831428e-03 1.12190485e+00 -3.93096298e-01
-6.28839970e-01 4.51209307e-01 -3.40554476e-01 -1.06459819e-01
6.40508175e-01 -3.14301491e-01 -1.38281643e-01 -5.33817232e-01
-7.50165999e-01 -1.66648477e-01 7.43299603e-01 1.81328401e-01
6.23872399e-01 -1.30489850e+00 -7.71507740e-01 3.43309551e-01
-3.74589056e-01 4.00834948e-01 5.67823827e-01 4.95585769e-01
-9.34681356e-01 9.45532545e-02 -2.96599716e-01 -4.48143572e-01
-7.21913457e-01 5.70857942e-01 2.56463021e-01 -2.55618513e-01
-9.60180581e-01 1.26570070e+00 6.55381680e-01 -6.13049651e-03
-4.35042083e-02 -2.71246582e-01 4.45204973e-01 -2.42789060e-01
5.65206826e-01 1.84292138e-01 -4.63387296e-02 -2.86520064e-01
1.19138956e-01 4.11847949e-01 -2.03713238e-01 -3.42753053e-01
1.49833870e+00 -7.08644092e-02 -1.33352652e-01 -1.29880399e-01
1.13652134e+00 2.85654496e-02 -2.01725745e+00 -1.68061465e-01
-5.85738719e-01 -4.31321561e-01 4.84347701e-01 -5.45860708e-01
-1.23095834e+00 1.03391969e+00 2.92186648e-01 -6.84034154e-02
1.37332761e+00 -7.25405961e-02 1.14031267e+00 2.85293162e-02
4.87499051e-02 -1.01059711e+00 3.99052352e-01 3.56284320e-01
1.18210769e+00 -9.88233268e-01 9.43404585e-02 -3.04789573e-01
-5.60180008e-01 9.90006030e-01 4.94670242e-01 -5.41601241e-01
3.64414006e-01 6.32001758e-01 -3.94859090e-02 1.34141184e-02
-9.44762945e-01 6.95024207e-02 2.00315133e-01 4.37535346e-01
3.76104653e-01 -1.61026135e-01 -7.23160356e-02 1.72462493e-01
-3.07333618e-01 -1.24600224e-01 4.64500666e-01 7.12252438e-01
-3.00019711e-01 -1.13962281e+00 -1.58898771e-01 2.85522401e-01
-4.25921589e-01 -4.72019225e-01 7.59086534e-02 5.63214183e-01
-6.80280775e-02 6.53633177e-01 3.51285547e-01 4.08991426e-03
1.20633692e-01 -2.41058215e-01 6.67997479e-01 -5.71107507e-01
-2.91219741e-01 2.56179869e-01 -2.62199789e-02 -7.65238941e-01
-1.46032602e-01 -4.46507305e-01 -1.05779803e+00 -5.54666162e-01
-8.81620944e-02 -2.55686253e-01 4.79703039e-01 6.40500784e-01
5.81672192e-01 7.97561049e-01 4.33232903e-01 -1.06892180e+00
-4.84937131e-01 -7.60439813e-01 -3.04471016e-01 3.61273557e-01
4.31296468e-01 -3.49288613e-01 -9.46495980e-02 3.35627198e-01] | [11.374889373779297, -0.4910448491573334] |
9515f00c-2daa-4ff6-a410-ca648f234453 | spherical-transformer-for-lidar-based-3d | 2303.12766 | null | https://arxiv.org/abs/2303.12766v1 | https://arxiv.org/pdf/2303.12766v1.pdf | Spherical Transformer for LiDAR-based 3D Recognition | LiDAR-based 3D point cloud recognition has benefited various applications. Without specially considering the LiDAR point distribution, most current methods suffer from information disconnection and limited receptive field, especially for the sparse distant points. In this work, we study the varying-sparsity distribution of LiDAR points and present SphereFormer to directly aggregate information from dense close points to the sparse distant ones. We design radial window self-attention that partitions the space into multiple non-overlapping narrow and long windows. It overcomes the disconnection issue and enlarges the receptive field smoothly and dramatically, which significantly boosts the performance of sparse distant points. Moreover, to fit the narrow and long windows, we propose exponential splitting to yield fine-grained position encoding and dynamic feature selection to increase model representation ability. Notably, our method ranks 1st on both nuScenes and SemanticKITTI semantic segmentation benchmarks with 81.9% and 74.8% mIoU, respectively. Also, we achieve the 3rd place on nuScenes object detection benchmark with 72.8% NDS and 68.5% mAP. Code is available at https://github.com/dvlab-research/SphereFormer.git. | ['Jiaya Jia', 'Jianhui Liu', 'Fanbin Lu', 'Yukang Chen', 'Xin Lai'] | 2023-03-22 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Lai_Spherical_Transformer_for_LiDAR-Based_3D_Recognition_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Lai_Spherical_Transformer_for_LiDAR-Based_3D_Recognition_CVPR_2023_paper.pdf | cvpr-2023-1 | ['lidar-semantic-segmentation'] | ['computer-vision'] | [-1.16240092e-01 -3.74696285e-01 -2.87149549e-01 -3.93829733e-01
-7.45709598e-01 -6.70835018e-01 3.18138719e-01 1.45925835e-01
-3.17299128e-01 2.11571544e-01 1.52275283e-02 -1.08554617e-01
-1.84259370e-01 -7.45869637e-01 -7.45930791e-01 -7.15724289e-01
3.16282988e-01 5.34835875e-01 6.35291338e-01 -5.47770746e-02
3.62246007e-01 7.38857210e-01 -1.77275002e+00 -1.71521213e-02
1.19473028e+00 1.09489799e+00 5.99393725e-01 7.79490545e-03
-2.89119661e-01 -1.31221414e-01 -3.23615134e-01 -3.75494994e-02
4.47503030e-01 4.56556171e-01 -2.80341864e-01 -1.22129016e-01
7.66578615e-01 -2.40427792e-01 -3.48820269e-01 1.10666502e+00
5.45715094e-01 3.03662121e-01 6.40838861e-01 -1.10194492e+00
-7.05068529e-01 2.75618106e-01 -1.20095813e+00 3.74232501e-01
-6.48033544e-02 5.73424473e-02 9.22420084e-01 -1.34664679e+00
2.78346866e-01 1.26093078e+00 4.92537647e-01 1.50387570e-01
-1.07379425e+00 -1.06162202e+00 5.79856813e-01 2.79963285e-01
-1.80109417e+00 -3.77115488e-01 7.61090279e-01 -2.33230501e-01
9.15352166e-01 4.06247437e-01 4.68477964e-01 6.83574200e-01
-3.47429723e-01 9.31823850e-01 7.92656958e-01 5.83850406e-02
4.35112556e-03 -1.77700177e-01 2.13129103e-01 2.64889032e-01
4.78929281e-01 -4.05211784e-02 -3.85491937e-01 4.34708335e-02
9.76417899e-01 3.67603838e-01 -2.70252615e-01 -3.99048775e-01
-1.12685585e+00 7.56091118e-01 8.70598257e-01 2.76992857e-01
-3.18752885e-01 -2.06712618e-01 3.47167440e-02 -3.48720580e-01
4.49026525e-01 1.29548013e-01 -3.43364984e-01 1.89866573e-01
-9.49126244e-01 1.21715963e-01 2.16031417e-01 1.20126712e+00
8.43851805e-01 -3.58708091e-02 -1.39286771e-01 1.06280267e+00
3.09515297e-01 1.02317595e+00 1.64048612e-01 -8.96605313e-01
7.18748748e-01 6.59979880e-01 -1.04027763e-02 -1.16816962e+00
-3.77894074e-01 -7.53172457e-01 -7.39260077e-01 -1.39279366e-02
2.91280210e-01 2.67828792e-01 -1.11512363e+00 1.40317953e+00
5.30683577e-01 4.22685176e-01 -3.38297725e-01 1.18760860e+00
1.09394634e+00 8.19646239e-01 -1.01165183e-01 7.83118606e-02
1.22017622e+00 -8.65098774e-01 -2.46360436e-01 -4.08367425e-01
1.44462556e-01 -8.14622879e-01 1.04538047e+00 1.88437581e-01
-1.03017664e+00 -6.83609664e-01 -8.62423003e-01 -3.92937392e-01
-9.65733901e-02 1.95151359e-01 5.36477864e-01 2.86923289e-01
-6.90409064e-01 3.21939826e-01 -8.99479449e-01 -1.36493012e-01
8.67213905e-01 4.38606918e-01 2.81938966e-02 -4.02779937e-01
-6.56179249e-01 3.73281986e-01 1.91334903e-01 1.82898100e-02
-3.78606617e-01 -9.23790157e-01 -7.18495667e-01 1.21915780e-01
4.54120070e-01 -3.86655271e-01 8.42861891e-01 -1.72062993e-01
-9.94792104e-01 7.63267457e-01 -4.63505715e-01 -1.98020592e-01
3.02741855e-01 -3.76754194e-01 -1.20856769e-01 -1.11942030e-02
4.19472635e-01 1.02416182e+00 6.95800304e-01 -1.29000592e+00
-7.64755785e-01 -9.39972818e-01 -2.68500686e-01 5.55108845e-01
-2.37270921e-01 -3.77337068e-01 -9.07048702e-01 -6.64002776e-01
8.50743651e-01 -8.69947374e-01 -1.40198350e-01 -5.92041314e-02
-3.55015993e-01 -4.12897766e-01 8.16372037e-01 -2.11008966e-01
1.12156057e+00 -2.42159224e+00 8.68019555e-03 2.52567589e-01
3.93892735e-01 2.58100957e-01 -1.25146866e-01 -6.63463585e-03
1.11374788e-01 -1.87663082e-03 -1.65562019e-01 -2.85519660e-01
-2.87302304e-02 2.06618711e-01 -3.42209131e-01 4.13964123e-01
1.90702736e-01 9.02388990e-01 -6.81057692e-01 -3.82293642e-01
4.47354913e-01 6.40623450e-01 -6.24282420e-01 -1.40096024e-01
7.86280558e-02 3.78783643e-01 -8.57883990e-01 9.70300794e-01
1.29136193e+00 -3.08839351e-01 -5.30066669e-01 -4.29874182e-01
-3.07532609e-01 1.38625860e-01 -1.25509012e+00 1.96685398e+00
-2.58114129e-01 3.88277292e-01 6.03904873e-02 -6.44817889e-01
1.20045996e+00 -3.87147218e-01 4.85928953e-01 -8.53478253e-01
1.06681049e-01 3.74813288e-01 -1.51697233e-01 -1.73920065e-01
5.51402211e-01 8.33706334e-02 1.69203550e-01 -1.58486471e-01
-1.43399432e-01 -1.54118434e-01 -9.27694738e-02 -9.17626172e-03
7.74752021e-01 5.12108617e-02 -1.64291225e-02 -1.97450832e-01
3.85305822e-01 -5.67335039e-02 6.24445558e-01 5.56121826e-01
-1.93773106e-01 1.01537991e+00 -5.27353585e-03 -1.66741520e-01
-6.28745079e-01 -1.35783434e+00 -6.02062166e-01 1.05739069e+00
7.84068704e-01 -2.59513259e-01 -4.35548782e-01 -4.19274390e-01
4.27855283e-01 7.24014223e-01 -2.55394757e-01 -9.36040953e-02
-5.30559182e-01 -6.26593590e-01 2.46501416e-01 6.70089185e-01
4.04457390e-01 -6.76727176e-01 -3.97034585e-01 -4.17049341e-02
-1.77741230e-01 -1.07505584e+00 -5.71569920e-01 1.30788401e-01
-9.62130547e-01 -7.92160571e-01 -7.45894551e-01 -8.36451709e-01
5.53664148e-01 9.80766356e-01 9.02021945e-01 -2.72645295e-01
-9.52998847e-02 -5.53298555e-02 -3.65032673e-01 -3.99007261e-01
5.77851593e-01 1.87324807e-01 1.43659674e-02 -1.36482343e-01
7.21974492e-01 -6.32872224e-01 -8.20116043e-01 6.08923316e-01
-6.09305739e-01 -3.97002436e-02 6.73268139e-01 5.64452708e-01
1.07973611e+00 -2.45528281e-01 3.61003578e-01 -3.60702366e-01
1.32522225e-01 -6.25886083e-01 -6.11546695e-01 -1.26512513e-01
-3.69876564e-01 -1.73309222e-01 2.12869078e-01 -3.40913177e-01
-7.38620162e-01 1.06005110e-01 -2.19911411e-01 -8.35736215e-01
-3.65764081e-01 2.56744120e-02 -3.39813173e-01 -2.52882630e-01
5.19237816e-01 2.55103648e-01 -3.23677093e-01 -7.36102760e-01
2.89600998e-01 7.56108999e-01 4.75906134e-01 -6.62375450e-01
7.97697365e-01 6.90174460e-01 -2.36881614e-01 -8.86390805e-01
-8.06058943e-01 -8.62095118e-01 -4.59250242e-01 7.82948881e-02
6.63043559e-01 -1.13644052e+00 -6.44930243e-01 3.16499978e-01
-1.06779635e+00 4.75208694e-03 -3.92498583e-01 4.86672044e-01
-1.40677646e-01 2.88338035e-01 -3.13468456e-01 -6.48826480e-01
-3.19163471e-01 -1.27505183e+00 1.44389749e+00 4.85778958e-01
1.21288538e-01 -2.27295384e-01 -2.64485151e-01 4.89051163e-01
1.75834998e-01 8.95151272e-02 4.49429274e-01 -5.21791697e-01
-8.62135053e-01 -9.19215679e-02 -7.82530844e-01 1.20595023e-01
-1.41930627e-02 -2.13827342e-01 -1.03100121e+00 -1.54662073e-01
2.20703594e-02 1.97694823e-02 1.12605762e+00 6.32682323e-01
1.40428269e+00 2.08512709e-01 -5.65940142e-01 9.49017107e-01
1.16158545e+00 2.74805218e-01 3.41458440e-01 3.93778831e-02
9.96593058e-01 5.22763968e-01 8.91266763e-01 3.83039683e-01
5.89946151e-01 6.62626803e-01 6.58513725e-01 4.95409872e-03
-1.01361565e-01 -3.68376702e-01 -1.51579872e-01 6.88028574e-01
-1.61386475e-01 -2.06781596e-01 -1.05364871e+00 5.81203222e-01
-1.76892591e+00 -6.88450456e-01 -3.05200070e-01 2.28074169e+00
5.06490588e-01 2.47214898e-01 -6.21506386e-02 -5.26597016e-02
8.55618477e-01 2.27110311e-01 -8.20229173e-01 3.31772417e-01
-2.26527438e-01 2.35218480e-01 6.79307163e-01 4.72851992e-01
-1.11602938e+00 1.03242707e+00 4.20669460e+00 1.36715519e+00
-1.01660228e+00 1.45735934e-01 5.45749724e-01 -4.38488364e-01
-2.64056832e-01 -2.34071046e-01 -1.26177263e+00 6.12780809e-01
2.81819940e-01 1.24868087e-01 1.70364872e-01 7.77132273e-01
1.58337608e-01 -9.38105434e-02 -6.89787090e-01 1.31305194e+00
-6.53812736e-02 -1.23578429e+00 1.26044124e-01 1.54433087e-01
7.45992959e-01 6.84895396e-01 2.39539251e-01 8.47336724e-02
-9.99867469e-02 -9.89337504e-01 8.26996267e-01 2.21104741e-01
7.16279089e-01 -8.77850294e-01 4.31068897e-01 4.91130859e-01
-1.45507169e+00 -9.05259699e-02 -6.64832532e-01 1.90601766e-01
3.14225592e-02 7.40849137e-01 -4.81098831e-01 4.66086477e-01
9.83208954e-01 8.35488081e-01 -4.66385603e-01 1.33086622e+00
1.88864842e-02 3.33787948e-01 -9.94064569e-01 2.27626208e-02
3.03960115e-01 -3.81525695e-01 8.34036171e-01 1.06943786e+00
5.18245518e-01 2.43801951e-01 3.96732539e-01 9.11401451e-01
4.31109406e-02 1.35234728e-01 -4.94987309e-01 3.58417749e-01
9.64556873e-01 1.14016390e+00 -7.64418960e-01 -6.88840672e-02
-3.39303404e-01 6.54769003e-01 2.29270160e-01 4.05155271e-01
-8.98531437e-01 -3.73609960e-01 8.16701770e-01 3.56698632e-01
5.77797651e-01 -4.81797844e-01 -6.45519376e-01 -1.27825594e+00
2.21224651e-01 -4.64813292e-01 2.42948294e-01 -6.92456543e-01
-1.19171989e+00 5.97724259e-01 -2.14903802e-03 -1.31501460e+00
4.30297375e-01 -4.30538476e-01 -4.56022978e-01 9.20159519e-01
-1.64819932e+00 -1.23593426e+00 -6.13502979e-01 4.89503503e-01
7.14407384e-01 1.49969757e-01 2.74253964e-01 5.87452650e-01
-5.10362387e-01 5.46337426e-01 1.46688908e-01 -1.09234229e-01
6.09993219e-01 -9.09078479e-01 4.42054659e-01 6.46515012e-01
2.16994867e-01 6.45511448e-01 3.66662234e-01 -6.92007661e-01
-1.19810855e+00 -1.10566163e+00 6.30842507e-01 -4.43421632e-01
4.52332467e-01 -4.32041287e-01 -1.23415339e+00 3.49495232e-01
-3.64528358e-01 1.20546646e-01 3.46472532e-01 1.11468367e-01
-4.42025870e-01 -3.12341481e-01 -1.02167284e+00 4.84878540e-01
1.36121798e+00 -1.96095243e-01 -4.62131560e-01 2.85051733e-01
7.46565819e-01 -7.29235709e-01 -6.58675373e-01 8.32658291e-01
3.26943040e-01 -9.42842305e-01 1.37640250e+00 -3.51812840e-02
1.51559383e-01 -6.72420979e-01 -4.93794411e-01 -9.07078028e-01
-5.75996876e-01 -1.72442302e-01 -1.32906422e-01 1.20199955e+00
3.23432833e-01 -6.95747018e-01 9.59348857e-01 2.06115350e-01
-4.27117676e-01 -9.72908080e-01 -1.10096383e+00 -6.95656538e-01
4.54710461e-02 -6.41692460e-01 7.81516314e-01 6.99315548e-01
-4.98213738e-01 3.94422323e-01 1.33045435e-01 5.05433321e-01
7.55970240e-01 6.10703170e-01 5.30201077e-01 -1.30609632e+00
7.75802135e-02 -6.17248178e-01 -4.00808781e-01 -1.67893445e+00
-1.33437619e-01 -9.36812580e-01 -1.17863975e-01 -1.34694362e+00
8.58429000e-02 -8.80199850e-01 -3.50255340e-01 4.06146616e-01
-2.42519081e-01 4.38163966e-01 2.43835911e-01 4.84460711e-01
-5.99681020e-01 8.04624498e-01 1.29172671e+00 -3.68814990e-02
-3.36681694e-01 1.18481435e-01 -7.26064026e-01 7.73974657e-01
1.03383672e+00 -4.56761837e-01 -4.17526871e-01 -8.14197004e-01
-1.18407048e-01 -2.89483786e-01 4.75472331e-01 -1.13585806e+00
3.01837146e-01 -1.18041545e-01 6.02551758e-01 -1.19107461e+00
5.60711443e-01 -7.29458451e-01 -1.04531579e-01 1.57745957e-01
1.01492591e-01 -2.61901706e-01 2.71888524e-01 6.86811566e-01
-1.45365283e-01 -5.77383786e-02 8.69865298e-01 1.83668613e-01
-7.82081544e-01 7.21630931e-01 3.33970755e-01 1.52738199e-01
1.05028772e+00 -5.26134253e-01 -2.07765698e-01 5.65262549e-02
-6.22901142e-01 6.43360615e-01 6.19661272e-01 5.69933414e-01
7.25582123e-01 -1.32223570e+00 -5.95941424e-01 4.93005812e-01
1.87297225e-01 8.15713167e-01 5.43591142e-01 8.48698974e-01
-2.34406024e-01 3.55324805e-01 -8.98583680e-02 -1.11262488e+00
-1.05783904e+00 2.75834739e-01 4.73012514e-02 3.83120060e-01
-6.44386232e-01 1.16026080e+00 5.76030314e-01 -3.63161325e-01
2.16467932e-01 -5.26934981e-01 -2.41683677e-01 1.51886478e-01
3.43978077e-01 4.81633663e-01 5.50228581e-02 -7.97974229e-01
-5.58364213e-01 1.10814190e+00 -1.63841963e-01 3.73613060e-01
1.33056509e+00 -1.86152458e-01 1.42368615e-01 1.87784597e-01
1.12966168e+00 1.29729524e-01 -1.53543735e+00 -3.83674294e-01
-2.55737960e-01 -8.57739568e-01 2.96116978e-01 -4.25859928e-01
-1.18810213e+00 9.96259689e-01 6.73352420e-01 1.45304576e-01
9.63122547e-01 4.28830981e-01 9.40007746e-01 1.79761037e-01
4.33156520e-01 -7.74401248e-01 -2.25695029e-01 6.66734576e-01
9.13232684e-01 -1.36287642e+00 7.61520565e-02 -7.96977639e-01
-5.91691375e-01 7.70321488e-01 7.01298296e-01 -2.92489439e-01
5.80595315e-01 2.22820550e-01 -2.06638023e-01 -2.77045906e-01
-2.37997562e-01 -3.84691447e-01 5.50451338e-01 6.92689717e-01
1.32834405e-01 1.69558376e-01 -9.14478227e-02 6.48450196e-01
-1.76685750e-01 -3.01589847e-01 -1.32007658e-01 5.52339017e-01
-7.69009352e-01 -6.37133300e-01 -4.68187928e-01 6.06591523e-01
-4.97578308e-02 -1.72587320e-01 4.68535349e-03 6.72120810e-01
3.57379198e-01 7.64324963e-01 3.64953339e-01 -3.65347415e-01
5.11291802e-01 -3.93462121e-01 3.19815993e-01 -6.13802910e-01
-1.91014037e-01 4.63183075e-01 -3.59863967e-01 -6.40552759e-01
-1.20774835e-01 -9.36358273e-01 -1.53167129e+00 -2.92206615e-01
-5.21175325e-01 -1.22503599e-03 6.19471014e-01 5.35022676e-01
7.38920569e-01 3.10880899e-01 5.71084976e-01 -1.17169631e+00
-5.45316041e-01 -7.24765599e-01 -4.77520734e-01 1.81349069e-02
1.61870807e-01 -9.80087161e-01 -2.47473866e-01 -5.16382575e-01] | [7.951814651489258, -3.2977027893066406] |
c777bf5d-58f5-481b-ad5d-34f8cd8311b2 | real-time-seismic-intensity-prediction-using | 2306.14336 | null | https://arxiv.org/abs/2306.14336v1 | https://arxiv.org/pdf/2306.14336v1.pdf | Real-time Seismic Intensity Prediction using Self-supervised Contrastive GNN for Earthquake Early Warning | Seismic intensity prediction in a geographical area from early or initial seismic waves received by a few seismic stations is a critical component of an effective Earthquake Early Warning (EEW) system. State-of-the-art deep learning-based techniques for this task suffer from limited accuracy in the prediction and, more importantly, require input waveforms of a large time window from a handful number of seismic stations, which is not practical for EEW systems. To overcome the above limitations, in this paper, we propose a novel deep learning approach, Seismic Contrastive Graph Neural Network (SC-GNN) for highly accurate seismic intensity prediction using a small portion of initial seismic waveforms received by a few seismic stations. The SC-GNN comprises two key components: (i) a graph neural network (GNN) to propagate spatiotemporal information through the nodes of a graph-like structure of seismic station distribution and wave propagation, and (ii) a self-supervised contrastive learning component to train the model with larger time windows and make predictions using shorter initial waveforms. The efficacy of our proposed model is thoroughly evaluated through experiments on three real-world seismic datasets, showing superior performance over existing state-of-the-art techniques. In particular, the SC-GNN model demonstrates a substantial reduction in mean squared error (MSE) and the lowest standard deviation of the error, indicating its robustness, reliability, and a strong positive relationship between predicted and actual values. More importantly, the model maintains superior performance even with 5s input waveforms, making it particularly efficient for EEW systems. | ['Mohammed Eunus Ali', 'A. F. M. Saiful Amin', 'Md. Forkan Uddin', 'Md. Anu Zakaria', 'Kazi Noshin', 'Rafid Umayer Murshed'] | 2023-06-25 | null | null | null | null | ['contrastive-learning', 'contrastive-learning'] | ['computer-vision', 'methodology'] | [-3.75411399e-02 1.03302975e-03 3.21169347e-01 -2.92459782e-02
-8.40976477e-01 -1.01134941e-01 3.36189181e-01 3.89984965e-01
-4.29906845e-01 3.90036494e-01 1.61673769e-01 -5.84527552e-01
-4.39064652e-01 -1.10088301e+00 -5.87676525e-01 -9.40163195e-01
-1.10026002e+00 1.80771440e-01 7.32407093e-01 -6.80331826e-01
2.76712686e-01 6.97238684e-01 -8.54916334e-01 -4.71491590e-02
5.84823728e-01 1.11800969e+00 1.74201168e-02 6.70986533e-01
2.50867009e-01 1.20185256e+00 -6.48641467e-01 -1.04226083e-01
5.48494421e-02 -2.57758886e-01 -3.93654346e-01 -5.30003488e-01
-3.77949834e-01 -4.35983539e-01 -8.78869534e-01 5.45667768e-01
8.85358572e-01 3.71884763e-01 6.64049029e-01 -1.07683372e+00
-2.40634263e-01 8.15581322e-01 -7.72929609e-01 6.32130444e-01
4.22704816e-02 9.63312462e-02 7.49243677e-01 -9.52930808e-01
-3.19357179e-02 5.89291334e-01 1.45655584e+00 -7.82160610e-02
-8.47244680e-01 -6.84076965e-01 -5.00372052e-02 3.69463891e-01
-1.57812071e+00 -1.48773089e-01 1.33849704e+00 -5.72043419e-01
1.06172311e+00 -5.58610857e-02 6.84891582e-01 6.29574060e-01
4.23568428e-01 4.85606343e-01 3.23228806e-01 9.51721054e-03
3.94850820e-01 -7.97017694e-01 -1.90217763e-01 4.99748677e-01
1.67162612e-01 2.16652676e-01 -6.53803468e-01 -1.61378369e-01
8.65240574e-01 7.48263523e-02 -4.81546789e-01 4.84654278e-01
-1.00477362e+00 9.55996394e-01 8.24658632e-01 4.29391354e-01
-7.90899992e-01 3.56353521e-01 5.34638047e-01 2.59648889e-01
8.06730747e-01 1.64650202e-01 -5.89992255e-02 -1.40011013e-01
-1.21151543e+00 4.64305142e-03 6.83067977e-01 3.40178579e-01
4.81377959e-01 8.66113842e-01 1.37915641e-01 6.76652849e-01
4.06324178e-01 7.65790224e-01 1.51598051e-01 -4.07429338e-01
7.38604546e-01 2.96451241e-01 -1.51700690e-01 -1.75797272e+00
-1.00351191e+00 -7.26480007e-01 -1.38723016e+00 1.68893933e-01
2.77399093e-01 -6.29819393e-01 -4.87739801e-01 1.58262277e+00
1.63378827e-02 7.50808537e-01 2.15026230e-01 7.54737437e-01
7.92506158e-01 1.31019354e+00 2.56121121e-02 -5.07971533e-02
9.20424998e-01 -5.20115674e-01 -4.65726525e-01 -5.54943085e-01
6.21490002e-01 -2.93753892e-01 4.28943813e-01 1.47571504e-01
-9.27257359e-01 -4.37891006e-01 -1.08669269e+00 5.70908368e-01
-1.60569966e-01 -1.97831243e-01 4.84503567e-01 1.43808261e-01
-1.09629071e+00 7.44761169e-01 -9.78200078e-01 2.94717193e-01
2.75600135e-01 3.47587287e-01 -1.39247939e-01 1.55556470e-01
-1.74870932e+00 5.59873164e-01 2.92446077e-01 7.52874136e-01
-8.55307817e-01 -8.41797054e-01 -9.52164352e-01 4.38686103e-01
-6.73317611e-02 -7.39242360e-02 9.00821149e-01 -3.42816532e-01
-1.14528012e+00 1.45270899e-02 3.31170976e-01 -5.86990058e-01
3.41770589e-01 4.29883339e-02 -5.44885635e-01 4.48755980e-01
-1.83549687e-01 -2.55895048e-01 5.64502954e-01 -1.17023969e+00
-5.87771595e-01 -4.60572988e-02 -3.45911950e-01 -1.82937413e-01
-4.25944716e-01 -3.20963174e-01 -1.07549086e-01 -9.83591437e-01
2.50086337e-01 -4.35031861e-01 -4.12320167e-01 -3.12135965e-01
-2.28866681e-01 -2.08221823e-01 5.77718735e-01 -1.07447135e+00
1.62068391e+00 -2.12650180e+00 -2.24319845e-01 5.21720946e-01
2.98143506e-01 3.37127298e-01 -2.16714531e-01 1.02673376e+00
-1.16960697e-01 -1.47063196e-01 -6.04303241e-01 -1.17033914e-01
-3.52751762e-01 -5.22263721e-02 -3.74633193e-01 5.76567173e-01
3.23385686e-01 6.01626098e-01 -1.07346499e+00 -2.12796956e-01
2.48890147e-02 7.00507581e-01 -2.08575889e-01 4.15688396e-01
3.82940888e-01 2.58938730e-01 -5.90064108e-01 3.97393942e-01
6.93853021e-01 -3.60731855e-02 -2.29097798e-01 -7.42041543e-02
-2.32415363e-01 -3.24352458e-02 -1.19959295e+00 1.03671825e+00
-3.64834547e-01 9.04374421e-01 -1.73681363e-01 -1.48964179e+00
1.51353216e+00 5.59745431e-01 7.98341036e-01 -8.79286468e-01
-2.05045529e-02 4.18826252e-01 -1.08126342e-01 -7.64485538e-01
2.09227443e-01 -1.20781846e-01 -4.88809124e-02 5.20570636e-01
-1.72469854e-01 1.24986976e-01 1.31707624e-01 2.02418432e-01
1.40161848e+00 -4.82057124e-01 -2.20878329e-02 -1.07886977e-01
3.28403056e-01 -1.61748096e-01 6.36698544e-01 8.81336451e-01
9.51098278e-02 7.01125681e-01 5.42674303e-01 -6.22937143e-01
-9.36162651e-01 -8.20220947e-01 1.89381808e-01 8.80865812e-01
5.37416860e-02 1.57019272e-01 -3.48104537e-01 -1.77855134e-01
-1.52044937e-01 5.77504814e-01 -4.76396501e-01 -3.90609980e-01
-1.07074940e+00 -1.10049045e+00 1.01261234e+00 9.34934020e-01
5.28539181e-01 -1.03874385e+00 -6.17328823e-01 7.29441762e-01
-2.64667779e-01 -8.83687675e-01 -1.38158783e-01 1.83854431e-01
-8.17026734e-01 -7.69422233e-01 -1.01479459e+00 -9.21479642e-01
4.40155625e-01 1.04895160e-01 9.90908861e-01 3.31490636e-01
2.94235915e-01 1.04011865e-02 -5.74651539e-01 -3.33060205e-01
-3.63146424e-01 -3.94924134e-02 -3.43124986e-01 2.50486612e-01
-7.20138922e-02 -8.44012320e-01 -9.00058985e-01 1.34497330e-01
-8.94507527e-01 -1.06544934e-01 5.62460065e-01 8.61458361e-01
1.74806535e-01 3.96677017e-01 1.10867965e+00 -4.46940720e-01
5.23160577e-01 -1.11542988e+00 -6.05155289e-01 -2.20180620e-02
-2.65770108e-01 -3.00573736e-01 9.41720963e-01 -4.50215042e-01
-9.44865048e-01 -1.20235652e-01 -7.00127363e-01 1.52583688e-01
3.03146571e-01 1.29821014e+00 4.67144340e-01 -2.38068193e-01
6.39206409e-01 4.84323204e-01 -2.39731565e-01 -3.13592911e-01
-3.04928273e-01 5.67839980e-01 8.86716008e-01 -1.09202780e-01
8.93346608e-01 3.92750353e-01 2.89237589e-01 -1.04406393e+00
-4.19191301e-01 -4.54131633e-01 -5.55927515e-01 -6.75531209e-01
4.72870648e-01 -8.30988228e-01 -7.48315811e-01 9.94077265e-01
-1.33802223e+00 -6.99167609e-01 8.19946975e-02 7.08884537e-01
-1.07949287e-01 3.92228007e-01 -1.00833106e+00 -1.18844485e+00
-6.61683679e-01 -6.07847035e-01 8.59970748e-01 2.42866516e-01
1.29246339e-01 -1.46451080e+00 2.61372447e-01 -2.97622532e-01
7.81385303e-01 8.65232110e-01 7.67593384e-01 -7.57519960e-01
-2.07562104e-01 -7.47397244e-01 -2.33693644e-01 -9.92018823e-03
-4.68495972e-02 2.85127736e-03 -7.12052345e-01 -3.44512254e-01
1.46046598e-02 1.11194186e-01 8.85047257e-01 5.65244853e-01
6.24368310e-01 -2.89312333e-01 -1.26935333e-01 7.03998625e-01
1.56311715e+00 5.16470611e-01 6.62680149e-01 4.18344885e-01
6.66116178e-01 4.14587468e-01 2.89483041e-01 9.03689325e-01
3.85837734e-01 1.51440695e-01 5.63257575e-01 -5.53981245e-01
2.35961881e-02 -1.61958598e-02 1.32664680e-01 1.26502585e+00
-4.92922306e-01 -5.58514118e-01 -1.52821541e+00 8.59442234e-01
-1.79297805e+00 -1.17603016e+00 -4.20769811e-01 2.01438332e+00
6.06761873e-01 3.56698275e-01 1.04459794e-02 6.65020764e-01
5.04942119e-01 6.27244413e-01 -3.20113838e-01 -3.98917012e-02
-1.67383939e-01 1.16018564e-01 5.53541660e-01 3.53339821e-01
-1.04341149e+00 4.22639847e-01 5.73544121e+00 7.97753215e-01
-1.40758598e+00 -3.33149284e-02 5.86183786e-01 4.33107764e-01
-2.80776978e-01 -4.01919872e-01 -3.10412765e-01 5.42423010e-01
1.04471529e+00 -1.44650221e-01 7.13750273e-02 5.36303580e-01
5.55562615e-01 8.37257132e-02 -6.02107346e-01 7.18038738e-01
-1.45912215e-01 -1.59098113e+00 -2.64490247e-01 -3.09660405e-01
5.15761852e-01 3.70179504e-01 -2.77079552e-01 1.71197727e-01
9.84695852e-02 -7.54320085e-01 7.25295842e-01 7.10083127e-01
5.18727720e-01 -9.73756313e-01 1.22226262e+00 3.88685435e-01
-1.76544869e+00 -2.78586090e-01 -2.60177314e-01 -2.17856675e-01
6.65633976e-01 8.48440945e-01 -5.29081762e-01 6.88193619e-01
6.38799667e-01 7.64924049e-01 -1.78007767e-01 1.40286756e+00
-2.32393488e-01 1.32798803e+00 -4.70554322e-01 -1.42545447e-01
5.81936896e-01 1.25065848e-01 4.26199466e-01 1.62779391e+00
6.27151728e-01 4.64522719e-01 3.78120095e-02 3.49321932e-01
1.74500421e-01 -1.68787494e-01 -6.33832574e-01 2.84589529e-01
7.31255114e-01 8.06005180e-01 -7.88042724e-01 -1.99152350e-01
-5.64461410e-01 2.86972731e-01 1.06635936e-01 7.23858833e-01
-8.73048067e-01 -9.16394472e-01 3.20923366e-02 1.55268162e-01
2.62016594e-01 -5.58084846e-01 -4.97875154e-01 -4.19096291e-01
-1.89081877e-01 -2.87438750e-01 5.64105451e-01 -5.77064157e-01
-1.19245994e+00 1.04931879e+00 -1.00742146e-01 -1.50000429e+00
-2.93491334e-01 -2.45803833e-01 -1.20156586e+00 7.12333024e-01
-1.71593654e+00 -9.22447085e-01 -4.93971705e-01 6.04502022e-01
3.67998302e-01 -1.71368077e-01 4.16067630e-01 6.34914041e-01
-5.85604131e-01 4.46852565e-01 2.47855455e-01 7.31392801e-01
5.48301488e-02 -9.09604013e-01 8.15467000e-01 1.17648125e+00
-1.70658380e-01 -5.14884554e-02 7.04626799e-01 -6.33966386e-01
-1.30045950e+00 -1.09335101e+00 8.95069301e-01 3.73830110e-01
9.71409678e-01 6.95544109e-02 -1.33238149e+00 3.12634021e-01
8.39747209e-03 9.88304019e-02 4.67761636e-01 -4.38215166e-01
5.59123158e-02 -2.76246101e-01 -6.63539708e-01 4.48940098e-01
5.02932787e-01 -3.05418342e-01 -4.92885143e-01 2.33069837e-01
5.64683914e-01 -2.25261465e-01 -1.05246568e+00 6.55916572e-01
3.35070968e-01 -8.19892645e-01 1.07334352e+00 -2.53669284e-02
5.46018064e-01 -1.56129047e-01 1.28271848e-01 -1.44420516e+00
-5.31618595e-01 -5.91807961e-01 -2.07238168e-01 9.46661174e-01
5.25254846e-01 -8.82279098e-01 6.34170711e-01 3.54934394e-01
-6.27754927e-01 -1.03313696e+00 -1.08723164e+00 -8.14676762e-01
4.22581211e-02 -7.47537553e-01 5.48049271e-01 8.39357734e-01
1.75166363e-03 -4.88796458e-02 -5.62782168e-01 7.57205844e-01
7.74984181e-01 -1.35812402e-01 2.50927001e-01 -1.14000857e+00
-2.54348040e-01 -5.72510183e-01 -4.59465265e-01 -9.91279066e-01
-2.84145832e-01 -4.95884687e-01 3.34744483e-01 -1.67383587e+00
-2.46962503e-01 -4.75132555e-01 -4.99290228e-01 5.39553165e-01
-2.14934334e-01 3.10479909e-01 -1.64452881e-01 2.24747047e-01
-9.54829007e-02 6.32353663e-01 7.71610379e-01 -2.29387775e-01
-2.30179787e-01 3.06936324e-01 -1.05147995e-02 8.18708479e-01
8.28250170e-01 -8.29591036e-01 -2.66897887e-01 -7.80157924e-01
4.99502093e-01 7.07515955e-01 2.97317207e-01 -1.30031943e+00
6.59003854e-01 -8.14421922e-02 8.53713006e-02 -8.16804945e-01
2.86346488e-02 -5.10089934e-01 9.81110036e-02 8.30501854e-01
-7.96286687e-02 1.75879851e-01 2.72260606e-01 5.32693744e-01
-5.78751206e-01 -2.23297939e-01 5.90872407e-01 3.33449930e-01
-9.38205957e-01 4.25281554e-01 -7.77512491e-01 1.39056280e-01
6.65318847e-01 -1.38605013e-01 -2.71054894e-01 -5.89798570e-01
-2.95266062e-01 2.25484744e-01 -4.03502584e-01 -3.69679295e-02
1.07158613e+00 -1.23810148e+00 -1.15541637e+00 2.21476078e-01
-8.74658078e-02 9.92994085e-02 6.10258877e-01 1.02883458e+00
-1.06575215e+00 -1.43454269e-01 7.42431805e-02 -5.10612547e-01
-5.88543653e-01 1.05564058e-01 3.12386960e-01 -3.06774497e-01
-9.38785553e-01 1.01952994e+00 -1.13124497e-01 7.42337927e-02
1.98123738e-01 -1.42255530e-01 -4.10317302e-01 7.50937983e-02
7.60906577e-01 6.76673174e-01 1.72712117e-01 -7.10804403e-01
-3.92877996e-01 7.22360134e-01 6.28352582e-01 4.39131819e-02
1.85217988e+00 5.55494763e-02 1.97360337e-01 3.41510236e-01
1.22721791e+00 -2.31002450e-01 -1.55810547e+00 -3.55312228e-01
1.89351946e-01 -1.44082502e-01 3.16602886e-01 -2.38707215e-01
-1.60251510e+00 9.87143278e-01 3.78319293e-01 6.30105078e-01
1.41224372e+00 -1.12465985e-01 1.28025413e+00 4.52920794e-01
1.93638161e-01 -8.58885169e-01 2.56443590e-01 6.68009400e-01
1.00644934e+00 -1.00081944e+00 -1.31583869e-01 1.04818113e-01
-4.77397472e-01 1.48125148e+00 -2.97418032e-02 -4.01837111e-01
9.87571359e-01 7.29545653e-01 1.06124103e-01 -4.19049680e-01
-4.11691844e-01 1.27825633e-01 1.97685614e-01 6.42737329e-01
2.47542143e-01 -1.95741206e-01 -4.97911461e-02 7.82530010e-01
7.92295635e-02 -2.51797765e-01 4.54356700e-01 9.41395938e-01
-7.06940293e-01 -1.80165038e-01 -3.52865070e-01 2.94378430e-01
-3.88328344e-01 -2.22277164e-01 3.22063625e-01 8.02746892e-01
-4.23118353e-01 1.08524299e+00 1.84175279e-02 -6.15405381e-01
4.67042476e-01 -5.24996877e-01 -4.33007807e-01 -3.03462781e-02
-5.62945306e-01 4.13984619e-02 -6.19210117e-02 -3.82501990e-01
-2.08983332e-01 -4.74853069e-01 -1.55135238e+00 -4.84688431e-01
-2.91333824e-01 2.29206115e-01 3.96612465e-01 1.02421701e+00
1.23561054e-01 7.92290926e-01 1.10548139e+00 -1.36719143e+00
-2.76785284e-01 -9.92116570e-01 -8.87019157e-01 2.57163346e-01
5.39231896e-01 -3.89821023e-01 -6.49658620e-01 -2.99089737e-02] | [6.907083034515381, 2.64308762550354] |
682b9f20-4136-49c8-8076-b12e31f28144 | hdnet-human-depth-estimation-for-multi-person | 2007.08943 | null | https://arxiv.org/abs/2007.08943v1 | https://arxiv.org/pdf/2007.08943v1.pdf | HDNet: Human Depth Estimation for Multi-Person Camera-Space Localization | Current works on multi-person 3D pose estimation mainly focus on the estimation of the 3D joint locations relative to the root joint and ignore the absolute locations of each pose. In this paper, we propose the Human Depth Estimation Network (HDNet), an end-to-end framework for absolute root joint localization in the camera coordinate space. Our HDNet first estimates the 2D human pose with heatmaps of the joints. These estimated heatmaps serve as attention masks for pooling features from image regions corresponding to the target person. A skeleton-based Graph Neural Network (GNN) is utilized to propagate features among joints. We formulate the target depth regression as a bin index estimation problem, which can be transformed with a soft-argmax operation from the classification output of our HDNet. We evaluate our HDNet on the root joint localization and root-relative 3D pose estimation tasks with two benchmark datasets, i.e., Human3.6M and MuPoTS-3D. The experimental results show that we outperform the previous state-of-the-art consistently under multiple evaluation metrics. Our source code is available at: https://github.com/jiahaoLjh/HumanDepth. | ['Gim Hee Lee', 'Jiahao Lin'] | 2020-07-17 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/3074_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123630613.pdf | eccv-2020-8 | ['3d-multi-person-pose-estimation-absolute', '3d-multi-person-pose-estimation-root-relative'] | ['computer-vision', 'computer-vision'] | [-3.63738745e-01 1.76375553e-01 -7.36395642e-02 -3.67815048e-01
-6.24597549e-01 2.48435959e-02 3.08117270e-01 -4.04695928e-01
-6.49546862e-01 3.85543168e-01 3.50508720e-01 5.53036273e-01
3.76917988e-01 -4.71735299e-01 -5.97578824e-01 -3.20867747e-01
-6.40537962e-02 7.11909592e-01 1.58148468e-01 8.38527549e-03
-6.11114874e-03 2.58427739e-01 -8.66219461e-01 -2.98162937e-01
3.56332242e-01 1.02271521e+00 -9.51552391e-02 7.61224151e-01
6.73854172e-01 6.24627471e-01 -7.61095703e-01 -5.10661066e-01
3.82697374e-01 -1.75823078e-01 -4.90426481e-01 2.60072976e-01
7.88098335e-01 -7.12972403e-01 -8.95329475e-01 8.86313617e-01
1.00462162e+00 2.27958888e-01 6.91171646e-01 -1.44819844e+00
-5.09462714e-01 1.21929711e-02 -1.05985785e+00 1.16736051e-02
7.45359898e-01 3.72411698e-01 9.11925316e-01 -1.10148466e+00
6.79576874e-01 1.65473831e+00 8.23494434e-01 5.75558305e-01
-6.41909182e-01 -5.35704255e-01 1.60668492e-01 2.44008869e-01
-1.82059979e+00 -9.29703116e-02 8.16234529e-01 -3.95889461e-01
1.02163541e+00 -1.62373185e-01 9.72267568e-01 1.12385201e+00
4.64538813e-01 9.78643537e-01 6.03974998e-01 -2.64711976e-01
-2.59883821e-01 -5.70059776e-01 -7.18482733e-02 1.30926597e+00
1.96494922e-01 8.52187164e-03 -8.17789733e-01 1.17056578e-01
1.34650755e+00 1.16762239e-02 -1.24134094e-01 -3.57811719e-01
-1.26953185e+00 7.09395707e-01 9.29188013e-01 -3.96186948e-01
-5.35007834e-01 7.70169079e-01 2.14720339e-01 -2.61994720e-01
5.76837838e-01 -1.82066813e-01 -2.85926133e-01 1.25780767e-02
-5.16099513e-01 7.14541495e-01 6.69945776e-01 1.03656638e+00
4.19324219e-01 -4.24199849e-01 -3.83376211e-01 7.19739676e-01
7.27750063e-01 5.29383302e-01 -7.14773759e-02 -1.09600675e+00
7.91993201e-01 7.68665075e-01 9.96653810e-02 -1.15495956e+00
-7.49395013e-01 -4.84909236e-01 -8.05495381e-01 1.93519462e-02
7.39044905e-01 -4.22173351e-01 -1.02296090e+00 1.59218585e+00
7.90841937e-01 2.64814906e-02 -5.75918257e-01 1.50243461e+00
8.70261371e-01 5.16627014e-01 2.69895736e-02 6.68308973e-01
1.53344190e+00 -1.46601212e+00 -4.21965331e-01 -7.03338563e-01
3.41192961e-01 -4.32696939e-01 8.15895617e-01 2.14028254e-01
-1.07124674e+00 -7.91987956e-01 -9.76036012e-01 -5.42904317e-01
-1.01172924e-01 6.48184836e-01 4.39760804e-01 3.02367836e-01
-8.98687065e-01 4.31565136e-01 -1.04126537e+00 -3.97909820e-01
2.51900673e-01 3.32196116e-01 -5.26815474e-01 -5.50146066e-02
-1.25749350e+00 8.81636322e-01 1.33681431e-01 7.37589180e-01
-7.21746802e-01 -2.35309914e-01 -1.13056982e+00 -2.89368808e-01
5.42431653e-01 -1.36411464e+00 1.24384594e+00 -1.25486910e-01
-1.35026956e+00 1.09577525e+00 5.72237112e-02 -1.90497547e-01
8.90702307e-01 -8.46443057e-01 1.43927023e-01 2.70536304e-01
4.69655991e-01 1.02046824e+00 7.12202966e-01 -8.34474862e-01
-4.72797811e-01 -7.65191197e-01 -1.16156980e-01 5.72367430e-01
6.94612190e-02 -1.00965716e-01 -1.30313027e+00 -6.84667349e-01
2.16634870e-01 -1.13971627e+00 -3.14151049e-01 5.73006928e-01
-1.21270001e+00 -5.17239392e-01 2.92616338e-01 -1.04558671e+00
1.03448880e+00 -1.62197232e+00 6.20176911e-01 2.83444911e-01
5.76411784e-01 -3.68712991e-01 7.73024783e-02 1.89583853e-01
3.70080471e-01 -3.69246334e-01 6.82743415e-02 -9.30730164e-01
2.74609268e-01 -2.33880699e-01 6.41794026e-01 1.01654220e+00
1.81657523e-02 1.16175878e+00 -5.67123234e-01 -7.69328713e-01
2.73830801e-01 8.69992554e-01 -3.78359407e-01 4.10241067e-01
7.56799132e-02 4.30903137e-01 -4.54260767e-01 8.02700520e-01
4.19525892e-01 -4.37736779e-01 -6.28119707e-02 -3.54370266e-01
3.07402909e-01 -6.81236899e-03 -1.25105953e+00 2.21376586e+00
1.09897284e-02 2.58853257e-01 -2.88446009e-01 -2.46918783e-01
7.84095824e-01 3.33444662e-02 4.89836991e-01 -3.06904405e-01
4.47231114e-01 -1.30907178e-01 -3.20260197e-01 -2.64509887e-01
3.52561831e-01 5.01244426e-01 -4.68218833e-01 2.04143018e-01
1.40438840e-01 2.30487674e-01 8.55530500e-02 6.28217533e-02
8.17090154e-01 5.32710671e-01 3.22649479e-01 1.45314127e-01
3.59293908e-01 -3.67119521e-01 6.07341290e-01 5.47114372e-01
-4.85237569e-01 8.57163012e-01 5.84456503e-01 -5.57400167e-01
-9.50989842e-01 -1.31677401e+00 4.86160457e-01 9.90471900e-01
3.41012895e-01 -5.69640577e-01 -1.01568127e+00 -8.06990266e-01
2.08327875e-01 -1.99907161e-02 -7.56259024e-01 -1.15884438e-01
-8.06462169e-01 -3.47276449e-01 6.13865077e-01 8.17229688e-01
9.47100401e-01 -8.07401955e-01 -7.01388240e-01 -5.21035530e-02
-4.98325169e-01 -1.34819436e+00 -1.07403088e+00 -4.45913523e-01
-4.71285284e-01 -1.21024883e+00 -1.30983591e+00 -8.18268120e-01
7.10612714e-01 -8.01064819e-02 1.06363499e+00 -3.29477377e-02
-6.17773056e-01 4.90551770e-01 1.03074007e-01 -1.20459214e-01
5.58264852e-01 3.60443175e-01 1.69510320e-01 -1.92046821e-01
4.98906791e-01 -2.72969842e-01 -1.13235676e+00 4.25479382e-01
6.90776333e-02 1.44410223e-01 5.84622204e-01 4.31895852e-01
6.71968162e-01 -1.82696775e-01 9.90258008e-02 -3.21395755e-01
4.63083208e-01 -7.42237493e-02 -3.55883420e-01 1.12383582e-01
5.12779057e-02 -1.64318159e-01 -1.27545387e-01 -2.36802518e-01
-7.46590197e-01 4.14828241e-01 -1.53414622e-01 -3.76389235e-01
-9.52681527e-02 1.91934466e-01 -3.55738878e-01 -4.13148887e-02
5.41855395e-01 -1.45791382e-01 1.32203018e-02 -4.34113264e-01
2.81684250e-01 3.77331257e-01 8.34395051e-01 -5.40463746e-01
8.65545809e-01 3.54355276e-01 1.82206795e-01 -6.55050933e-01
-1.04240918e+00 -5.10052443e-01 -9.96133506e-01 -7.23745048e-01
1.34041595e+00 -1.25443530e+00 -1.04978263e+00 9.77262735e-01
-1.39037216e+00 -4.00185943e-01 2.54224658e-01 4.92902100e-01
-5.52334905e-01 4.83551651e-01 -9.02733922e-01 -6.91756248e-01
-6.74437881e-01 -9.89123106e-01 1.65289426e+00 3.16864043e-01
-4.37163651e-01 -7.53003180e-01 1.23271653e-02 5.88861227e-01
-2.45534480e-01 7.11035252e-01 2.71356165e-01 -3.17442464e-03
-6.22503161e-01 -7.05805361e-01 -2.96939760e-01 1.91329848e-02
-2.82286882e-01 -4.38065320e-01 -6.06588066e-01 -4.00168061e-01
-5.96101582e-01 -3.82376194e-01 8.47954094e-01 7.11621046e-01
8.98432612e-01 -1.63701206e-01 -3.08971643e-01 7.25464642e-01
1.00021529e+00 -5.36601543e-01 4.80755657e-01 2.90241569e-01
1.16221035e+00 6.45296514e-01 7.02860713e-01 6.73013568e-01
1.01400638e+00 9.88923132e-01 3.66379172e-01 -1.07539751e-01
-3.51005048e-01 -5.29912353e-01 4.11735713e-01 3.85090321e-01
-4.74243730e-01 -2.08925888e-01 -8.36541772e-01 1.15347631e-01
-1.97036612e+00 -4.12219554e-01 4.25674394e-02 1.93141830e+00
5.03174245e-01 4.11974698e-01 6.36873841e-01 -3.40647787e-01
9.41942215e-01 2.89342195e-01 -7.45780587e-01 4.85928059e-01
3.25571567e-01 -1.25942662e-01 4.27417517e-01 5.04037082e-01
-1.44687748e+00 1.06141281e+00 4.75966549e+00 3.59455377e-01
-4.77246016e-01 4.91483770e-02 4.08463925e-01 -4.03327286e-01
6.11658633e-01 -4.16552871e-01 -1.17780936e+00 3.18613052e-01
2.47251898e-01 4.03062522e-01 2.29850337e-01 7.71501005e-01
1.50334686e-01 -1.97164983e-01 -1.07501137e+00 1.31231034e+00
1.80119500e-01 -6.34799719e-01 -3.26930225e-01 1.43673077e-01
4.32341486e-01 -3.39751951e-02 1.51312556e-02 1.86605126e-01
3.10221929e-02 -9.38560843e-01 7.97572136e-01 8.35255027e-01
7.01308787e-01 -9.91967201e-01 6.73332810e-01 2.45630920e-01
-1.61106503e+00 2.11236298e-01 -3.90790343e-01 -7.57650062e-02
4.97339875e-01 5.51751494e-01 -6.87854052e-01 2.06482366e-01
9.19578791e-01 7.45189071e-01 -7.21184492e-01 1.02910042e+00
-9.46764946e-01 -1.44656569e-01 -3.96698117e-01 -1.25058413e-01
-5.76980151e-02 6.00072630e-02 4.54601943e-01 1.06221974e+00
1.92756444e-01 -1.78737760e-01 4.34679121e-01 8.08505654e-01
-1.41353503e-01 -1.38220683e-01 -1.53325215e-01 2.95499802e-01
3.55897784e-01 1.28356183e+00 -5.85963190e-01 -8.48761499e-02
-1.58716068e-01 1.51145124e+00 5.26386142e-01 4.22408253e-01
-1.14391875e+00 -5.08842647e-01 7.50408947e-01 1.76324159e-01
7.50205889e-02 -5.96490085e-01 -2.05295496e-02 -1.21844983e+00
3.02215397e-01 -4.79187608e-01 5.13390660e-01 -8.27287853e-01
-1.36219203e+00 3.09102327e-01 8.07433426e-02 -9.38863337e-01
-4.13684189e-01 -6.50149226e-01 -3.96801025e-01 8.95770311e-01
-8.51613939e-01 -1.49826324e+00 -7.01933205e-01 6.61159337e-01
2.66266167e-01 1.49644822e-01 4.05174792e-01 2.09704503e-01
-7.89745390e-01 8.74173462e-01 -8.67649913e-01 8.12776268e-01
8.31866562e-01 -1.32999766e+00 1.06860507e+00 7.75924504e-01
-2.57508963e-01 4.33880508e-01 4.21121448e-01 -1.00153697e+00
-1.36713731e+00 -1.09889817e+00 1.07524049e+00 -6.27634048e-01
2.29974121e-01 -7.54836619e-01 -1.88365787e-01 8.29950929e-01
-1.52517125e-01 3.12024206e-02 1.85622409e-01 1.50265023e-02
-1.58316940e-01 1.60634831e-01 -9.66340184e-01 7.55979240e-01
1.50766814e+00 -2.84075856e-01 -3.10064822e-01 4.76677150e-01
6.37363970e-01 -8.79985690e-01 -1.01026487e+00 1.14994437e-01
7.68973529e-01 -7.06400394e-01 1.47808266e+00 -3.05173725e-01
4.45160568e-01 -3.20113569e-01 5.86654022e-02 -1.01380420e+00
-4.12763983e-01 -2.57384211e-01 -5.31059742e-01 6.62662506e-01
2.53685191e-02 -2.78641462e-01 1.26702154e+00 5.61390817e-01
1.43667340e-01 -9.73297834e-01 -1.01630545e+00 -4.35692966e-01
-3.44488740e-01 -2.39691004e-01 1.63673133e-01 1.97813272e-01
-4.18145508e-01 4.71232206e-01 -7.49832094e-01 3.96207333e-01
1.21636820e+00 -3.21696132e-01 1.20591319e+00 -9.94529366e-01
-4.70841557e-01 -2.37405092e-01 -7.46423483e-01 -1.59040093e+00
4.77714241e-02 -6.02093697e-01 2.11337224e-01 -1.80476725e+00
2.51338571e-01 2.72437453e-01 4.01815362e-02 5.04570603e-01
-3.68351370e-01 4.31205004e-01 2.64615327e-01 2.67665107e-02
-7.71727920e-01 6.02999151e-01 1.56746542e+00 -1.40511498e-01
-1.49662867e-01 1.56034037e-01 -3.07669044e-01 9.22629476e-01
7.44164526e-01 -3.29118699e-01 -1.07563250e-01 -4.97602522e-01
4.29851040e-02 5.17826788e-02 1.13408959e+00 -1.38070881e+00
4.28619057e-01 1.76412269e-01 1.13532710e+00 -1.01020479e+00
9.06295896e-01 -3.57476979e-01 -1.46891981e-01 6.59168363e-01
-2.69278586e-01 1.61729887e-01 -3.65721017e-01 5.68742752e-01
3.02465856e-01 2.12700918e-01 5.75056314e-01 -4.13538337e-01
-8.29755843e-01 8.36833358e-01 1.18168816e-01 1.32229581e-01
8.46436262e-01 -2.34236300e-01 -1.17096804e-01 -5.80117702e-01
-8.69582593e-01 5.27859032e-01 1.80074558e-01 5.28998792e-01
9.94150877e-01 -1.58768570e+00 -8.52759242e-01 -9.36858505e-02
5.98958284e-02 3.12925547e-01 3.77854228e-01 9.24591780e-01
-7.14891315e-01 3.65668356e-01 -1.34676948e-01 -7.11176097e-01
-1.24621749e+00 3.82980034e-02 6.21732950e-01 -3.60747486e-01
-7.30729401e-01 1.27732503e+00 5.91012612e-02 -6.48907781e-01
7.22339272e-01 -1.10273346e-01 -2.30321172e-03 -2.29337528e-01
3.81265461e-01 6.81674421e-01 -4.18429047e-01 -7.59542108e-01
-6.97381973e-01 9.54241872e-01 7.98426345e-02 -2.09381402e-01
1.11755276e+00 -1.76237926e-01 6.24407409e-03 6.43684808e-03
1.41402841e+00 -4.76222008e-01 -1.65906692e+00 -2.41407052e-01
-4.51369017e-01 -5.11735380e-01 -2.27080584e-01 -6.51288092e-01
-1.18853104e+00 7.68125951e-01 6.64596796e-01 -6.00173295e-01
8.69089127e-01 3.04654866e-01 1.10628438e+00 3.75137329e-01
4.30371642e-01 -1.28648508e+00 4.88725841e-01 3.92716080e-01
9.97767329e-01 -1.11750412e+00 3.12891006e-01 -5.31920075e-01
-5.46273530e-01 1.01305127e+00 1.17578006e+00 -4.54955012e-01
5.32250404e-01 -1.11030325e-01 4.85539548e-02 -3.90248746e-01
-2.39794999e-01 -2.42071047e-01 6.99346781e-01 6.00095391e-01
3.96747291e-01 8.45902860e-02 -3.34050395e-02 4.85680878e-01
-3.82819384e-01 -5.50728627e-02 -5.77442646e-02 8.47028911e-01
-5.26435196e-01 -6.87536120e-01 -5.55822134e-01 1.17763229e-01
-2.09886044e-01 2.59452432e-01 -7.79610455e-01 1.01433098e+00
1.79673448e-01 7.13285029e-01 -3.38155739e-02 -5.95480442e-01
5.86934030e-01 -1.72142357e-01 8.10337245e-01 -4.71290976e-01
-2.17903078e-01 1.79609507e-01 1.82690024e-01 -8.80713105e-01
-1.79188341e-01 -5.34373939e-01 -1.32765377e+00 -3.91366601e-01
-5.55769391e-02 -4.18564528e-01 4.63920265e-01 7.35698819e-01
1.31159350e-01 4.86599088e-01 2.29336411e-01 -1.39751220e+00
-5.37296116e-01 -1.07872868e+00 -5.00349224e-01 2.42469117e-01
5.54058105e-02 -1.00837672e+00 4.77423333e-02 -2.45937705e-01] | [7.0937957763671875, -0.8029274344444275] |
558497ad-f429-444f-8852-b6b9c58f8203 | a-character-level-length-control-algorithm | 2205.14522 | null | https://arxiv.org/abs/2205.14522v2 | https://arxiv.org/pdf/2205.14522v2.pdf | A Character-Level Length-Control Algorithm for Non-Autoregressive Sentence Summarization | Sentence summarization aims at compressing a long sentence into a short one that keeps the main gist, and has extensive real-world applications such as headline generation. In previous work, researchers have developed various approaches to improve the ROUGE score, which is the main evaluation metric for summarization, whereas controlling the summary length has not drawn much attention. In our work, we address a new problem of explicit character-level length control for summarization, and propose a dynamic programming algorithm based on the Connectionist Temporal Classification (CTC) model. Results show that our approach not only achieves higher ROUGE scores but also yields more complete sentences. | ['Lili Mou', 'Xiang Zhang', 'Puyuan Liu'] | 2022-05-28 | null | null | null | null | ['headline-generation', 'abstractive-sentence-summarization'] | ['natural-language-processing', 'natural-language-processing'] | [ 2.59098232e-01 -5.51369712e-02 -5.12921870e-01 -2.20532104e-01
-8.93487096e-01 -3.35959285e-01 4.55363393e-01 6.04840338e-01
-3.38187814e-01 1.03036714e+00 8.34016681e-01 -1.67168289e-01
1.25267029e-01 -7.34336436e-01 -2.66355723e-01 -5.15016019e-01
1.86110839e-01 7.99518749e-02 4.55627382e-01 -3.36832970e-01
8.11266005e-01 7.26925507e-02 -9.96481419e-01 2.78063059e-01
1.40521955e+00 4.69815403e-01 2.46646762e-01 8.02752197e-01
-3.12684208e-01 5.50882578e-01 -1.08577120e+00 -4.51029688e-01
-2.00196430e-01 -8.25003147e-01 -9.02325690e-01 -1.09141342e-01
2.24820942e-01 -3.20820953e-03 -3.98273647e-01 1.00904250e+00
5.98439157e-01 2.92146534e-01 5.57021618e-01 -7.79991448e-01
-5.49212158e-01 1.03280044e+00 -8.79768550e-01 4.80201155e-01
5.19364119e-01 -1.05724566e-01 1.25813735e+00 -2.71917969e-01
4.59891707e-01 1.31716049e+00 4.66190368e-01 3.83503795e-01
-1.09049809e+00 -2.99047679e-01 3.74087185e-01 2.81143427e-01
-1.02188587e+00 -1.73424140e-01 8.04865181e-01 -6.54296950e-02
1.02538919e+00 5.75834513e-01 8.13212097e-01 6.94835544e-01
7.49548078e-01 1.07382309e+00 6.25894070e-01 -6.69594288e-01
1.11898713e-01 -2.36713603e-01 6.15408063e-01 3.86227727e-01
2.72583067e-01 -5.79786301e-01 -3.92891407e-01 -2.30907544e-01
2.66917944e-01 -1.55599505e-01 -3.21047783e-01 4.25413579e-01
-1.39450240e+00 8.77088249e-01 5.40590063e-02 3.32519412e-01
-3.68803054e-01 -3.45326662e-02 8.33305120e-01 1.12618223e-01
6.83477700e-01 5.75008512e-01 -1.77033320e-01 -3.96059245e-01
-1.09680617e+00 3.45449358e-01 8.42018604e-01 9.42327976e-01
3.06738853e-01 -4.82882559e-02 -7.36145973e-01 9.69176531e-01
-1.92120567e-01 2.99891323e-01 7.76533484e-01 -7.87518382e-01
8.75995278e-01 5.45177400e-01 -1.01576954e-01 -1.12372410e+00
-2.44628057e-01 -4.18545604e-01 -1.04807293e+00 -7.29549766e-01
-2.67602473e-01 -3.65351260e-01 -3.89148831e-01 1.40740085e+00
-9.15570185e-02 4.69870083e-02 1.50673211e-01 4.89271134e-01
6.96512103e-01 1.34422612e+00 -1.13211926e-02 -9.42948461e-01
1.10472679e+00 -1.44493270e+00 -8.08271110e-01 -1.39527544e-01
5.64311564e-01 -8.28523576e-01 9.75951374e-01 5.00932395e-01
-1.14970303e+00 -4.17777628e-01 -1.21686780e+00 -7.34118894e-02
1.78193450e-01 1.28093749e-01 5.41710377e-01 4.79771554e-01
-1.00499177e+00 7.58100092e-01 -6.51206851e-01 -5.46138644e-01
-5.85541017e-02 1.10884830e-01 5.75357117e-03 1.45137295e-01
-1.08836889e+00 7.98425972e-01 9.11362171e-01 -9.42080021e-02
-1.03731364e-01 -2.70652145e-01 -5.43421507e-01 2.57408440e-01
3.41686010e-01 -7.15528190e-01 1.41473222e+00 -5.63221633e-01
-1.67744887e+00 3.11140358e-01 -4.83353704e-01 -5.28118134e-01
3.12780917e-01 -2.55964249e-01 -3.68197739e-01 2.11555868e-01
5.42232282e-02 5.30155480e-01 5.69686294e-01 -8.01702082e-01
-8.35654557e-01 -1.65820807e-01 -7.53886178e-02 4.39732999e-01
-7.33819962e-01 7.81320110e-02 -4.31701839e-01 -9.78633225e-01
-2.32278206e-03 -7.61055648e-01 -4.62747574e-01 -9.28280175e-01
-6.85140729e-01 -5.71889997e-01 4.96174842e-01 -8.71428251e-01
2.11616564e+00 -1.79869771e+00 3.73805583e-01 -2.00570136e-01
-2.40382049e-02 4.87899184e-01 -8.63757357e-02 8.23589325e-01
1.86647937e-01 3.31461310e-01 -3.85361940e-01 -2.47773156e-01
-2.89095432e-01 3.10016237e-02 -5.58503985e-01 -3.21406089e-02
-5.39196059e-02 9.00049806e-01 -1.04419005e+00 -9.56521034e-01
-1.86845839e-01 -2.50119358e-01 -5.90920210e-01 1.88989967e-01
-3.78250122e-01 2.17555389e-01 -5.77709675e-01 4.74581607e-02
5.18005192e-01 7.35733584e-02 1.44446746e-01 2.81900112e-02
-3.49309295e-01 3.23368937e-01 -6.69987559e-01 1.64397883e+00
-3.31536740e-01 5.74012578e-01 -5.58976948e-01 -9.42885995e-01
1.07309377e+00 9.96164158e-02 2.95504898e-01 -5.01741111e-01
9.64460149e-02 -5.12217879e-02 -1.67675436e-01 -5.75324118e-01
1.34238684e+00 -2.34615281e-02 -3.86986017e-01 5.04559040e-01
-4.18128073e-01 -3.36764634e-01 6.83173001e-01 5.04073620e-01
1.13567698e+00 -2.02284262e-01 4.75939393e-01 -1.94532946e-01
6.62393034e-01 2.17557281e-01 7.32665777e-01 5.81595838e-01
-7.45460996e-03 7.49684632e-01 9.00166452e-01 1.72291428e-01
-1.17401123e+00 -6.89461231e-01 1.66747078e-01 8.23663831e-01
1.94305122e-01 -7.22097695e-01 -1.08270013e+00 -5.70924938e-01
-2.86192864e-01 1.11892319e+00 -1.70019239e-01 -3.37995827e-01
-9.49009836e-01 -8.21670413e-01 6.18046165e-01 3.42174262e-01
5.30024230e-01 -1.04451835e+00 -2.46630952e-01 4.75689262e-01
-7.10248411e-01 -7.75488496e-01 -1.14698350e+00 -2.30483457e-01
-1.25047815e+00 -5.00686765e-01 -9.55115318e-01 -8.95719111e-01
5.10746181e-01 5.63630104e-01 7.48169541e-01 -8.66954774e-02
1.85143769e-01 -5.26991412e-02 -7.04690993e-01 -4.45349306e-01
-5.30296683e-01 7.60625541e-01 -1.50505543e-01 -1.55794263e-01
1.03942402e-01 -5.04022658e-01 -4.28879738e-01 -7.18245804e-02
-9.67863441e-01 2.41053119e-01 6.64990604e-01 7.61457384e-01
4.56220895e-01 -3.94445099e-02 1.09086955e+00 -9.20708537e-01
1.44329381e+00 -3.33971769e-01 -2.39513636e-01 4.45091069e-01
-7.82524586e-01 2.36227363e-01 9.43724632e-01 -2.60905981e-01
-1.14775085e+00 -5.07497251e-01 -3.18444848e-01 1.11657724e-01
2.06876025e-01 8.72256398e-01 1.81800816e-02 5.21569848e-01
3.79736453e-01 6.66202366e-01 -1.19864643e-01 -3.52371633e-01
2.17956662e-01 9.45608079e-01 5.37506104e-01 -3.39394391e-01
3.99940073e-01 3.10196187e-02 -2.16896698e-01 -1.09019673e+00
-9.86774504e-01 -6.49754584e-01 -6.01508260e-01 -6.85766339e-03
5.98124862e-01 -5.39966106e-01 -3.42342436e-01 3.04942101e-01
-1.62678266e+00 1.02838017e-01 -2.19462663e-01 4.15414602e-01
-4.97567385e-01 1.10252154e+00 -6.62360549e-01 -6.39841974e-01
-9.32172954e-01 -7.63872802e-01 7.78904855e-01 4.62069184e-01
-4.54733700e-01 -9.12873864e-01 3.70742023e-01 1.64019063e-01
1.93218246e-01 1.72550380e-01 1.05281532e+00 -6.02298617e-01
-2.76226819e-01 -2.41876170e-01 -1.22448184e-01 4.30458367e-01
-6.46100268e-02 1.13063082e-01 -4.03942227e-01 -2.48149291e-01
1.99417006e-02 2.02290267e-01 1.14976335e+00 5.12420654e-01
9.99033511e-01 -6.24866426e-01 -1.90218687e-01 3.03825974e-01
1.22507977e+00 4.65707690e-01 8.97585988e-01 3.27436596e-01
5.35963893e-01 5.21859467e-01 1.05373418e+00 5.58072269e-01
3.86060506e-01 4.24459338e-01 -6.96043894e-02 2.85703391e-01
5.49561568e-02 -5.40143311e-01 5.31297743e-01 1.67370915e+00
1.76395774e-01 -6.76267982e-01 -4.92536753e-01 5.96670508e-01
-2.25183249e+00 -1.18348897e+00 -2.87905425e-01 1.97194374e+00
1.04822695e+00 3.22241873e-01 1.15404874e-01 1.28463089e-01
9.65470195e-01 5.74756384e-01 -3.42188507e-01 -8.69014025e-01
-1.38279825e-01 -1.27947852e-01 3.40360433e-01 4.06084478e-01
-7.91390121e-01 1.06755638e+00 6.81849480e+00 1.11909032e+00
-1.04840469e+00 -1.24208659e-01 5.80917478e-01 -4.87828925e-02
-4.29729819e-01 1.14976533e-01 -8.59498858e-01 6.59747183e-01
9.60208476e-01 -9.80709732e-01 -1.30480751e-02 6.22193456e-01
7.12803662e-01 -3.47499907e-01 -7.85135627e-01 8.84195387e-01
3.02976102e-01 -1.41341090e+00 3.91130596e-01 -8.67165700e-02
9.44383264e-01 -4.60110515e-01 -3.76081675e-01 2.53696203e-01
-5.57659827e-02 -5.51667511e-01 6.78322613e-01 6.32710755e-01
6.37125671e-01 -1.00755048e+00 8.24577272e-01 6.46556556e-01
-1.09339416e+00 8.52078572e-02 -5.64866006e-01 -4.82145026e-02
4.65603083e-01 7.02534616e-01 -7.15706229e-01 7.42684603e-01
3.15682255e-02 8.82875204e-01 -5.38408875e-01 1.42827201e+00
-3.08752865e-01 9.88706410e-01 7.41870701e-02 -6.15007401e-01
1.92396566e-01 -5.01049757e-01 8.60253274e-01 1.64008963e+00
3.31202447e-01 3.24420661e-01 2.82097042e-01 2.06519663e-01
-5.06900959e-02 5.72488785e-01 -4.32170808e-01 -8.60383213e-02
3.71401548e-01 9.19234157e-01 -6.94854081e-01 -4.27023947e-01
-3.49571630e-02 1.08330619e+00 1.38523072e-01 1.50643736e-01
-8.28269541e-01 -8.97560894e-01 2.15897672e-02 -6.82196394e-02
1.38455153e-01 -3.16916555e-01 -5.39022446e-01 -1.28658164e+00
2.10912883e-01 -5.11368394e-01 3.11335564e-01 -4.02177960e-01
-8.86469424e-01 4.83050138e-01 1.14194550e-01 -1.25483286e+00
-1.58023953e-01 2.04433858e-01 -1.26584029e+00 6.75066411e-01
-1.28660750e+00 -8.19394648e-01 -6.16084673e-02 -7.87253827e-02
1.05546916e+00 -1.55486256e-01 4.88885581e-01 1.47062883e-01
-1.07606673e+00 6.35225415e-01 4.22367513e-01 -2.38885760e-01
7.67488122e-01 -1.08353698e+00 4.87981349e-01 1.06004894e+00
-2.44938955e-01 5.60220480e-01 1.03865314e+00 -7.47729242e-01
-1.13805473e+00 -1.12321329e+00 1.21603835e+00 1.54194266e-01
4.74711329e-01 2.37733230e-01 -8.00810754e-01 3.91911060e-01
5.18586099e-01 -1.11293077e+00 4.77264106e-01 1.85218174e-02
1.82718217e-01 -2.60694087e-01 -7.58643270e-01 7.92900264e-01
8.49126279e-01 -4.13588211e-02 -8.70350480e-01 3.85451883e-01
1.18584120e+00 -1.43154860e-01 -6.00099504e-01 6.33826628e-02
3.10449094e-01 -7.52649069e-01 4.24097776e-01 -4.06314939e-01
8.59062135e-01 -4.93803956e-02 2.70374417e-01 -1.62215042e+00
-4.26859587e-01 -8.17430854e-01 -1.34025574e-01 1.50571311e+00
4.32440519e-01 -4.63470697e-01 7.40263820e-01 2.35271305e-01
-4.10157949e-01 -7.95519292e-01 -7.18237996e-01 -8.31272483e-01
8.18260461e-02 -1.16898097e-01 4.31120455e-01 4.58185077e-01
5.75761914e-01 9.23952281e-01 -4.92384166e-01 -3.43846530e-01
2.69864410e-01 2.14864254e-01 7.21543074e-01 -1.04159403e+00
-4.63191085e-02 -7.40301967e-01 -1.25055745e-01 -1.51851702e+00
1.88522100e-01 -8.58749509e-01 1.85260549e-01 -1.92763388e+00
6.45820141e-01 -5.74595928e-02 3.20529789e-02 -4.28744219e-02
-4.99873012e-01 -3.59839380e-01 1.98639199e-01 3.20947737e-01
-9.25837934e-01 9.49520707e-01 1.22441101e+00 -3.86574388e-01
-5.54413080e-01 2.08813369e-01 -9.97456610e-01 4.79247630e-01
9.73277807e-01 -4.49105620e-01 -3.62539798e-01 -4.92965341e-01
3.37410532e-02 4.40074891e-01 -4.43264127e-01 -8.74844372e-01
3.79256159e-01 -4.84062910e-01 -3.14199090e-01 -1.09162676e+00
-7.21440166e-02 -9.80470181e-02 -2.35202342e-01 5.84291458e-01
-6.42999709e-01 4.00787622e-01 7.57883638e-02 8.71413767e-01
-4.41267908e-01 -7.85642564e-01 6.74389303e-01 -1.86459310e-02
-4.75620270e-01 2.42216289e-01 -3.87944967e-01 4.71525900e-02
1.14227688e+00 -2.07839787e-01 -2.72159576e-01 -4.19802159e-01
-2.33814985e-01 6.04423106e-01 2.84510165e-01 4.25470978e-01
6.50315285e-01 -1.06577516e+00 -1.15684795e+00 -4.46621239e-01
-5.22383004e-02 -3.70297134e-02 1.69542983e-01 6.37788415e-01
-8.95174444e-01 8.96684587e-01 1.60060674e-01 -4.40823883e-01
-1.39401662e+00 5.23279250e-01 -2.61374831e-01 -7.27951765e-01
-7.95678496e-01 4.38137740e-01 -8.57423916e-02 2.66611092e-02
2.73376644e-01 -2.00932205e-01 -6.43890500e-01 3.07843864e-01
6.43084943e-01 7.07759619e-01 -6.31591305e-02 -3.88983369e-01
1.71311691e-01 3.69189560e-01 -4.79701102e-01 -1.86928913e-01
1.23343229e+00 -4.04609263e-01 -4.50060934e-01 5.49775124e-01
1.17369390e+00 1.88092157e-01 -6.41425848e-01 -2.21218452e-01
2.73991913e-01 -3.47165197e-01 -8.76555592e-02 -2.45703250e-01
-6.10157490e-01 7.11740553e-01 -1.26868293e-01 5.25303304e-01
1.04168940e+00 -2.74292946e-01 1.51888335e+00 5.04794419e-01
3.27158779e-01 -1.31626856e+00 1.88529998e-01 9.33263898e-01
9.79622424e-01 -7.87579358e-01 2.13866085e-01 -3.01555902e-01
-8.81202281e-01 1.16448021e+00 4.85615313e-01 -8.13961625e-02
1.05749443e-01 -1.04952946e-01 -4.11211222e-01 2.51475871e-01
-8.87746036e-01 1.65689409e-01 -1.21753387e-01 2.20073611e-01
4.28540438e-01 1.56181663e-01 -1.22225833e+00 5.94860137e-01
-4.24092531e-01 -1.71484873e-01 9.48820114e-01 7.71612644e-01
-8.87812555e-01 -1.14734268e+00 -3.18382561e-01 7.49710202e-01
-5.18017411e-01 -2.15542331e-01 -3.42111111e-01 3.92511077e-02
-3.71500850e-01 1.11885917e+00 4.76015508e-02 -4.18178439e-01
5.59004426e-01 -1.29547358e-01 2.27612749e-01 -8.55109692e-01
-6.16956830e-01 1.18800297e-01 2.14452326e-01 1.87891796e-01
-1.88763872e-01 -7.95039117e-01 -1.25047302e+00 -5.95283985e-01
-6.41871631e-01 7.58405030e-01 4.21811074e-01 8.66245091e-01
3.23459208e-01 6.45187736e-01 1.02128267e+00 -6.77406251e-01
-7.16422379e-01 -1.17703509e+00 -5.28760731e-01 7.00441375e-02
1.57474160e-01 1.44788861e-01 -1.15369193e-01 1.07725868e-02] | [12.594905853271484, 9.466304779052734] |
56cc1e37-c3ef-40bf-ae05-f45fd726d733 | mitosis-detection-in-intestinal-crypt-images | 1608.07616 | null | http://arxiv.org/abs/1608.07616v1 | http://arxiv.org/pdf/1608.07616v1.pdf | Mitosis Detection in Intestinal Crypt Images with Hough Forest and Conditional Random Fields | Intestinal enteroendocrine cells secrete hormones that are vital for the
regulation of glucose metabolism but their differentiation from intestinal stem
cells is not fully understood. Asymmetric stem cell divisions have been linked
to intestinal stem cell homeostasis and secretory fate commitment. We monitored
cell divisions using 4D live cell imaging of cultured intestinal crypts to
characterize division modes by means of measurable features such as orientation
or shape. A statistical analysis of these measurements requires annotation of
mitosis events, which is currently a tedious and time-consuming task that has
to be performed manually. To assist data processing, we developed a learning
based method to automatically detect mitosis events. The method contains a
dual-phase framework for joint detection of dividing cells (mothers) and their
progeny (daughters). In the first phase we detect mother and daughters
independently using Hough Forest whilst in the second phase we associate mother
and daughters by modelling their joint probability as Conditional Random Field
(CRF). The method has been evaluated on 32 movies and has achieved an AUC of
72%, which can be used in conjunction with manual correction and dramatically
speed up the processing pipeline. | ['Anika Böttcher', 'Michael Sterr', 'Heiko Lickert', 'Lichao Wang', 'Gerda Bortsova', 'Fausto Milletari', 'Tingying Peng', 'Nassir Navab', 'Fabian Theis'] | 2016-08-26 | null | null | null | null | ['mitosis-detection'] | ['medical'] | [ 3.89490962e-01 -2.18880828e-03 -5.33657335e-02 -3.64161253e-01
-5.85110486e-01 -8.56914997e-01 5.97908139e-01 1.14286745e+00
-7.07454503e-01 5.94989836e-01 2.37578556e-01 -1.51219100e-01
4.24468666e-01 -8.24221790e-01 -7.14750767e-01 -9.66501713e-01
-2.68354893e-01 9.56975698e-01 4.90534663e-01 4.37598735e-01
2.73502797e-01 2.03328490e-01 -1.23070395e+00 2.86454320e-01
6.27200127e-01 4.06662881e-01 1.59973040e-01 1.16858649e+00
-6.66814446e-02 3.85944217e-01 -2.81852186e-01 -1.10663623e-02
-7.69575089e-02 -5.72209120e-01 -5.52048862e-01 -5.17292693e-02
6.86161667e-02 -2.53880084e-01 7.10972071e-01 7.69485533e-01
4.85146284e-01 -2.96136856e-01 9.71012414e-01 -8.51449788e-01
2.17059523e-01 3.47985506e-01 -4.54971045e-01 4.33383882e-01
4.59280610e-01 -1.10603333e-01 4.42330122e-01 -5.32844067e-01
1.00189197e+00 8.08319628e-01 5.52165508e-01 2.68389583e-01
-1.60001695e+00 -3.12212080e-01 -4.07747060e-01 -2.96361268e-01
-1.11322534e+00 -3.93861711e-01 -3.12617719e-02 -9.24228549e-01
8.30099702e-01 6.01829514e-02 9.55675066e-01 4.11547750e-01
3.28775078e-01 6.43026292e-01 1.20151114e+00 -7.12501705e-01
5.69015324e-01 1.62156671e-01 -5.98748066e-02 5.89201391e-01
4.64244485e-01 -3.75486501e-02 -4.43423897e-01 -8.35108943e-03
8.32353830e-01 -2.91104436e-01 -2.11663954e-02 -3.05603713e-01
-1.46495271e+00 8.56636405e-01 -5.97471446e-02 4.51187551e-01
-3.21204156e-01 6.74666390e-02 3.46411675e-01 -5.86076193e-02
3.61602008e-01 -2.56477147e-02 -3.61807525e-01 -2.24623337e-01
-1.08366048e+00 1.12678081e-01 9.43747640e-01 4.42156196e-01
3.69623989e-01 -6.56945050e-01 6.82476684e-02 6.14748597e-01
5.52684307e-01 1.45919830e-01 5.23545504e-01 -8.81743014e-01
-4.60507303e-01 9.43121135e-01 -4.81733233e-02 -6.50136828e-01
-5.08210242e-01 -5.64694181e-02 -6.40650332e-01 4.83400762e-01
1.19542885e+00 -3.39124426e-02 -6.77625418e-01 1.34748387e+00
8.10377836e-01 -1.22563273e-01 -1.84645485e-02 8.18127155e-01
6.55579031e-01 5.31124592e-01 3.11856449e-01 -2.56234348e-01
1.70958257e+00 -4.36659813e-01 -4.71895993e-01 1.60696939e-01
6.53758764e-01 -9.97511506e-01 3.19122583e-01 5.09136140e-01
-9.15258050e-01 -2.73364205e-02 -1.14579928e+00 -1.01888843e-01
-5.61390400e-01 1.00528531e-01 7.32958317e-01 6.96109176e-01
-8.26964974e-01 5.48793554e-01 -1.47126293e+00 -7.95996130e-01
4.49392349e-01 5.45292437e-01 -6.18133128e-01 2.21047580e-01
-3.95357132e-01 5.50108194e-01 1.88321784e-01 -3.01397979e-01
-5.11324286e-01 -4.14094955e-01 -8.31293344e-01 -3.13269377e-01
-2.58592039e-01 -9.45177674e-01 1.18664801e+00 -5.41478157e-01
-1.58710349e+00 1.39778376e+00 -3.44535172e-01 -4.10812229e-01
5.91246665e-01 1.97107106e-01 1.69884875e-01 4.21821982e-01
1.94157273e-01 8.92631352e-01 2.32048512e-01 -1.02180028e+00
-1.12281966e+00 -5.73363662e-01 -4.86317515e-01 2.74408728e-01
3.70225579e-01 1.73961505e-01 -5.20189643e-01 -3.54136974e-01
4.43152606e-01 -8.91817391e-01 -1.05384499e-01 2.43233025e-01
-3.10449619e-02 1.21101856e-01 3.84759754e-01 -7.39087999e-01
6.23700619e-01 -1.90540385e+00 -1.65355708e-02 1.93055887e-02
1.12850636e-01 2.67296424e-03 6.45027876e-01 1.78103864e-01
4.05355126e-01 1.13934480e-01 -3.32165025e-02 -2.27201521e-01
-1.59701467e-01 4.18100730e-02 2.91516542e-01 9.46366012e-01
-4.64735404e-02 4.99345720e-01 -1.13979495e+00 -7.77922928e-01
4.00576681e-01 7.76769519e-01 -3.17443103e-01 5.37361018e-02
-2.58350462e-01 8.61544251e-01 1.04077728e-02 8.34992230e-01
4.27881241e-01 -1.25487998e-01 4.95470613e-01 -2.87782475e-02
-4.04958636e-01 2.57162541e-01 -1.11171377e+00 1.28889179e+00
2.20199469e-02 5.53942502e-01 3.51866931e-01 -5.15152216e-01
7.87242949e-01 3.68878603e-01 4.48299825e-01 -3.87995303e-01
2.79071003e-01 3.99701744e-01 -1.80533230e-01 -4.02376086e-01
2.59545922e-01 -5.44940054e-01 1.71428785e-01 4.08680707e-01
2.20406085e-01 -1.32814437e-01 6.10228002e-01 -1.07108727e-01
7.82018185e-01 6.20535910e-01 6.19836986e-01 -4.29954350e-01
3.45141143e-01 3.54683101e-01 5.31026006e-01 2.56245643e-01
-3.82789642e-01 7.33536839e-01 6.65227413e-01 -4.48543131e-01
-1.00725985e+00 -9.60440099e-01 -4.36664522e-01 8.43607903e-01
1.23836018e-01 1.44033246e-02 -7.09712446e-01 -3.36745530e-01
-8.51629972e-02 4.68059599e-01 -6.05928242e-01 4.11313564e-01
-4.49184328e-01 -1.27402973e+00 3.37867916e-01 3.20701152e-01
6.92753121e-02 -6.38417482e-01 -9.14865077e-01 4.76114929e-01
-9.03495681e-03 -7.98207521e-01 -1.51968837e-01 6.91021383e-01
-9.70361412e-01 -1.39828503e+00 -9.02324319e-01 -9.77205694e-01
8.66389930e-01 -1.04915507e-01 8.53270531e-01 -7.11851418e-02
-5.13020873e-01 9.21061204e-04 -3.25524688e-01 -5.09893894e-01
-7.22484231e-01 -1.79135025e-01 -3.17981653e-02 -3.27462375e-01
4.59958613e-01 -5.79539776e-01 -8.80288422e-01 4.40852106e-01
-8.06717575e-01 1.15609907e-01 5.29220402e-01 5.78282058e-01
1.13970196e+00 -2.54126519e-01 2.07969815e-01 -9.80916262e-01
-5.28645776e-02 -4.00889128e-01 -1.01051557e+00 -1.93636328e-01
-3.22608769e-01 -2.57747471e-01 4.82552886e-01 -1.74128070e-01
-8.09138358e-01 5.43548763e-01 -7.92989209e-02 5.90463877e-01
-5.45250416e-01 4.35714096e-01 2.79024065e-01 1.10292681e-01
4.63652909e-01 2.61706173e-01 3.10915351e-01 -3.19949895e-01
-1.40484527e-01 5.35100937e-01 5.91556847e-01 -1.84034884e-01
1.91141382e-01 7.30796933e-01 2.74556160e-01 -1.18297470e+00
-3.83517027e-01 -8.90877366e-01 -7.05168307e-01 -1.81496799e-01
9.90331829e-01 -9.43485081e-01 -7.77729332e-01 6.90852046e-01
-7.12547600e-01 -5.77594697e-01 -6.34972565e-03 9.06030416e-01
-5.02415180e-01 3.55415046e-01 -8.53519857e-01 -5.74537873e-01
-2.84733534e-01 -8.89032722e-01 1.08534491e+00 6.85400605e-01
-4.13847208e-01 -1.15832949e+00 6.55476272e-01 3.37237537e-01
3.64979565e-01 7.36284494e-01 8.43138814e-01 -7.57817447e-01
-3.90267342e-01 -3.73580694e-01 9.04879496e-02 -3.04015100e-01
-1.03718780e-01 4.26228344e-01 -7.28747904e-01 2.02291226e-03
-3.64462107e-01 2.54525226e-02 8.39444995e-01 8.79910946e-01
1.04741707e-01 7.98314512e-02 -5.53373218e-01 6.40933454e-01
1.43766952e+00 1.57193005e-01 4.16156918e-01 5.12273908e-01
-1.11124597e-01 7.88463354e-01 5.65088987e-01 3.41413230e-01
4.55687940e-01 4.76096898e-01 2.46296927e-01 -1.34761602e-01
-1.56332687e-01 -1.36036500e-01 3.14869016e-01 4.24439490e-01
-1.40831143e-01 -1.35649443e-01 -8.20914805e-01 7.33003199e-01
-1.26442170e+00 -6.77460134e-01 -4.70417410e-01 2.37881398e+00
9.48749423e-01 1.54013649e-01 4.61669385e-01 2.03253821e-01
8.34187746e-01 -5.17074049e-01 -1.82284266e-01 -2.26545319e-01
-4.29450162e-02 9.30778235e-02 5.30261457e-01 6.38945997e-01
-1.21570635e+00 4.81239140e-01 6.19979906e+00 1.58647135e-01
-1.19598687e+00 -3.09516191e-01 7.14975536e-01 2.24665478e-01
1.35034055e-01 2.86218315e-01 -9.66894329e-01 4.52264220e-01
5.46099186e-01 -1.11928545e-02 1.00034745e-02 3.59475583e-01
1.37220919e-01 -1.06005633e+00 -1.00195849e+00 4.18950945e-01
-4.29193705e-01 -1.29878569e+00 -6.00828588e-01 2.67218679e-01
4.22149807e-01 4.81759831e-02 -6.49730325e-01 2.60802880e-02
2.12858379e-01 -8.17156315e-01 4.81613636e-01 4.75128174e-01
8.89447153e-01 -3.28903139e-01 1.05974579e+00 4.07135308e-01
-1.01355958e+00 4.84378397e-01 -6.13101125e-02 -1.06227640e-02
3.79366964e-01 6.27024710e-01 -1.58684766e+00 4.55095759e-03
4.67570871e-01 3.07595372e-01 -4.79548156e-01 1.53429592e+00
7.38716349e-02 7.32720733e-01 -7.52358079e-01 -9.77633148e-02
-3.02380741e-01 -3.73471737e-01 5.63685417e-01 1.31689179e+00
3.43402475e-01 -1.81870505e-01 -9.29544866e-02 6.30958319e-01
2.43560433e-01 3.44928682e-01 -1.23165749e-01 -1.77701905e-01
2.72906125e-01 1.64740121e+00 -1.78440845e+00 -1.81180269e-01
-1.96622059e-01 7.38364577e-01 -1.41998723e-01 -2.29493350e-01
-4.17085528e-01 -1.50038645e-01 2.20048130e-01 3.93124163e-01
3.08374584e-01 -4.29236852e-02 -9.29276869e-02 -7.95471191e-01
-3.06181461e-01 -2.53449261e-01 5.66993773e-01 -2.45074436e-01
-7.22230911e-01 8.85895863e-02 -2.30455607e-01 -9.61272597e-01
-2.44021565e-01 -5.09447455e-01 -6.49421573e-01 5.04507959e-01
-1.27000177e+00 -1.24442852e+00 -3.43991637e-01 -1.86630383e-01
1.41314387e-01 5.54114819e-01 1.16876805e+00 2.14454174e-01
-3.86308998e-01 2.45719738e-02 1.44205585e-01 8.14799294e-02
7.02136755e-01 -1.85109675e+00 -1.54206604e-01 4.42625642e-01
2.03443854e-03 4.12889898e-01 1.03324318e+00 -7.34829605e-01
-1.04070222e+00 -9.36970592e-01 1.22669601e+00 -2.18997523e-01
1.34002730e-01 -2.14931369e-01 -4.88346249e-01 5.44731200e-01
-4.46170568e-02 7.67601654e-02 1.27072155e+00 -2.86818683e-01
2.51103491e-01 2.08865568e-01 -1.23089159e+00 1.93560138e-01
1.00343689e-01 2.27656007e-01 -2.50766188e-01 2.69255817e-01
-6.31753728e-02 -7.98586488e-01 -1.05371809e+00 2.48787299e-01
8.80442381e-01 -1.19129837e+00 4.79810357e-01 1.44042701e-01
2.38955200e-01 -8.05204451e-01 5.61388768e-02 -7.02281475e-01
1.67590395e-01 -3.27615499e-01 2.70875096e-01 1.32124388e+00
5.10035574e-01 -3.19692254e-01 8.21623206e-01 1.42761379e-01
-5.23322634e-02 -5.97082436e-01 -1.05129099e+00 -2.48775631e-01
-5.41286208e-02 2.05852062e-01 8.46005883e-03 7.56575167e-01
3.33012253e-01 2.56408691e-01 4.22369897e-01 -6.41708151e-02
5.65536022e-01 1.79809928e-01 7.63948202e-01 -1.14275587e+00
-1.91548616e-01 -4.53925908e-01 -5.33204436e-01 -6.13517940e-01
-6.14565015e-01 -8.15861464e-01 5.78820594e-02 -1.57143629e+00
4.02996451e-01 -2.64671355e-01 2.99420774e-01 2.11819082e-01
-4.20282707e-02 4.42471206e-01 -4.66413289e-01 1.13330089e-01
-6.11958206e-01 -2.74573416e-01 8.37192357e-01 3.19768190e-01
-3.20478559e-01 3.53752598e-02 -3.29397410e-01 8.67954910e-01
6.54532313e-01 -6.17894590e-01 3.22067440e-01 2.10206553e-01
4.53957200e-01 3.00261557e-01 3.20353925e-01 -9.76027131e-01
3.18568140e-01 3.67571235e-01 6.23936772e-01 -7.03831613e-01
-3.40823568e-02 -5.02497494e-01 4.72872913e-01 8.13489377e-01
-1.99175864e-01 -4.42785233e-01 -1.39778443e-02 5.34649551e-01
-3.34433585e-01 -3.28983843e-01 9.78911042e-01 -2.57680178e-01
-2.03872845e-01 -2.23755121e-01 -1.02808607e+00 -1.20097689e-01
1.33953357e+00 -4.49799299e-01 -4.92875017e-02 -2.01435521e-01
-8.72024357e-01 2.59392411e-01 1.16471910e+00 -3.20070952e-01
1.63191274e-01 -7.20954478e-01 -7.47036517e-01 1.37068793e-01
1.87715385e-02 4.79424208e-01 3.80286351e-02 1.36679542e+00
-1.36420596e+00 2.62556762e-01 -1.72306448e-01 -1.06119823e+00
-1.51498616e+00 1.49508253e-01 2.70218730e-01 -5.20888925e-01
-3.05060148e-01 8.67167234e-01 -2.75458694e-01 -1.92916974e-01
-4.57574949e-02 -3.61371100e-01 -7.32045293e-01 4.80052084e-01
3.55311692e-01 5.05784631e-01 1.95490550e-02 -6.89798236e-01
-3.50233346e-01 4.48143601e-01 -2.30813604e-02 3.23804170e-02
1.40021598e+00 -2.96585113e-01 -2.93825984e-01 5.05972326e-01
8.95058513e-01 1.54002413e-01 -1.25188828e+00 2.07358003e-01
2.51130819e-01 -2.17357546e-01 6.18988574e-02 -8.00592601e-01
-7.25729465e-01 5.81656814e-01 5.37542284e-01 2.12012276e-01
8.71288896e-01 3.48575190e-02 7.95699716e-01 -4.48236108e-01
2.86179692e-01 -9.39550042e-01 -4.22335446e-01 8.25370923e-02
1.68448433e-01 -1.34734869e+00 1.88890904e-01 -5.78802824e-01
-2.40606695e-01 1.34549785e+00 3.40218097e-02 -1.47269890e-01
4.12822574e-01 5.95162570e-01 2.15350956e-01 -1.22142196e-01
-7.59271801e-01 1.94590650e-02 -5.53198270e-02 6.30752981e-01
9.37953591e-01 7.52993375e-02 -8.41050506e-01 5.46853542e-01
1.97971030e-03 4.17243749e-01 4.67855036e-01 1.14933169e+00
-5.36137998e-01 -1.16700161e+00 -3.70082945e-01 3.43667746e-01
-1.11918342e+00 3.07122827e-01 -2.10813239e-01 7.41088867e-01
2.54946619e-01 6.38616860e-01 1.72145486e-01 1.69030398e-01
-1.56089783e-01 5.44075035e-02 6.28170729e-01 -5.56093514e-01
-5.77641845e-01 6.79961860e-01 1.59227312e-01 -1.69821501e-01
-7.97528028e-01 -1.17172503e+00 -1.47520113e+00 1.02632977e-01
-3.20253015e-01 -4.48310934e-02 9.93244946e-01 7.50286639e-01
6.77182227e-02 2.20715195e-01 1.57433823e-01 -8.55162323e-01
-5.37239434e-03 -8.29568207e-01 -7.72159159e-01 1.88229978e-01
1.92991421e-01 -4.55610752e-01 -6.23629093e-01 8.35063100e-01] | [14.612926483154297, -3.1864142417907715] |
fb50a242-437d-4794-b801-8f526bbbb466 | shrec-22-track-sketch-based-3d-shape | 2207.04945 | null | https://arxiv.org/abs/2207.04945v1 | https://arxiv.org/pdf/2207.04945v1.pdf | SHREC'22 Track: Sketch-Based 3D Shape Retrieval in the Wild | Sketch-based 3D shape retrieval (SBSR) is an important yet challenging task, which has drawn more and more attention in recent years. Existing approaches address the problem in a restricted setting, without appropriately simulating real application scenarios. To mimic the realistic setting, in this track, we adopt large-scale sketches drawn by amateurs of different levels of drawing skills, as well as a variety of 3D shapes including not only CAD models but also models scanned from real objects. We define two SBSR tasks and construct two benchmarks consisting of more than 46,000 CAD models, 1,700 realistic models, and 145,000 sketches in total. Four teams participated in this track and submitted 15 runs for the two tasks, evaluated by 7 commonly-adopted metrics. We hope that, the benchmarks, the comparative results, and the open-sourced evaluation code will foster future research in this direction among the 3D object retrieval community. | ['Hongyuan Wang', 'Ji Zhang', 'Qunying Zhou', 'Yan Wang', 'Haiqin Chen', 'Ying Tang', 'Feng Wang', 'Yang Wang', 'Zihao Xin', 'Zheng Zhang', 'Jianning Wang', 'Haoyang Luo', 'Minh-Triet Tran', 'Hai-Dang Nguyen', 'Tuan-Luc Huynh', 'Nhat-Khang Ngo', 'Thien-Tri Cao', 'Khoi-Nguyen Nguyen-Ngoc', 'Chi-Bien Chu', 'Nhat Hoang-Xuan', 'Yi Fang', 'Boulbaba Ben Amor', 'Jiaxin Chen', 'Shuaihang Yuan', 'Jie Qin'] | 2022-07-11 | null | null | null | null | ['3d-object-retrieval'] | ['computer-vision'] | [-9.74397287e-02 -6.23067379e-01 -2.81158164e-02 -2.50282496e-01
-9.04118538e-01 -1.09714580e+00 1.03371882e+00 -2.48882353e-01
-3.83971073e-02 1.58345774e-01 1.64551094e-01 -2.63287853e-02
-1.51064834e-02 -8.77073467e-01 -4.10075814e-01 -2.66979560e-02
-4.10008766e-02 8.29580903e-01 4.63477612e-01 -4.10690755e-02
7.44747460e-01 1.23520505e+00 -1.41866696e+00 2.67281175e-01
4.49807256e-01 9.76530313e-01 1.05529472e-01 5.34385324e-01
-4.01748389e-01 6.85502812e-02 -7.91208804e-01 -8.47686827e-01
5.30776620e-01 1.30903780e-01 -4.44098830e-01 7.14397654e-02
1.03859758e+00 -3.82063955e-01 -4.61681038e-01 8.80601466e-01
8.83814931e-01 1.01211816e-02 8.83139431e-01 -1.25473499e+00
-9.16949928e-01 7.91867748e-02 -6.19606555e-01 -7.09024072e-02
5.20057023e-01 2.43087724e-01 9.16513324e-01 -1.40346599e+00
1.01040792e+00 1.68726039e+00 5.42707920e-01 4.19558853e-01
-1.26053023e+00 -9.26651716e-01 5.41223474e-02 -1.84767157e-01
-1.60315442e+00 -2.86013424e-01 1.01783383e+00 -2.79679596e-01
6.07418120e-01 2.14746282e-01 6.71056271e-01 1.13835239e+00
-2.98795879e-01 1.12872469e+00 9.44570839e-01 -2.41110578e-01
2.11756155e-01 -1.49084860e-02 1.18840663e-02 5.09542346e-01
4.18107510e-01 -7.58105591e-02 -3.82869899e-01 -5.71546316e-01
1.28089738e+00 6.61368072e-02 -5.25853112e-02 -8.88836443e-01
-1.35238516e+00 5.84560633e-01 1.90594420e-01 3.32020938e-01
-7.82204792e-02 2.29208022e-01 3.84621590e-01 5.52987874e-01
5.33153355e-01 5.97413063e-01 3.56133189e-03 -1.84931368e-01
-8.69447410e-01 8.54933560e-01 9.10618484e-01 1.56613338e+00
3.11759025e-01 -2.21523210e-01 -2.08993956e-01 1.25483012e+00
1.11488990e-01 8.87251616e-01 -7.45483041e-02 -9.61942494e-01
5.79153478e-01 6.49796009e-01 3.22720647e-01 -1.25827956e+00
1.89180762e-01 -1.16910264e-01 -6.88752949e-01 2.68166274e-01
2.95530230e-01 3.85160387e-01 -8.85101616e-01 1.39547539e+00
1.65890276e-01 4.96914461e-02 -5.09570241e-01 1.03998017e+00
1.03994608e+00 5.25504053e-01 -2.16265738e-01 4.81146961e-01
1.25021267e+00 -1.03033507e+00 -3.19243282e-01 9.01224613e-02
1.31066874e-01 -1.31839979e+00 1.34964645e+00 4.07493979e-01
-1.43810761e+00 -6.38899386e-01 -1.07881951e+00 -1.65021449e-01
-5.56812525e-01 1.58606008e-01 6.07565939e-01 4.50165421e-01
-9.77049172e-01 6.72211826e-01 -3.37555677e-01 -5.47869205e-01
6.00700200e-01 -9.29204896e-02 -3.55406195e-01 -5.51032722e-01
-8.50429058e-01 8.47281277e-01 -2.73189157e-01 -7.36775026e-02
-7.91786253e-01 -8.74285340e-01 -3.54075491e-01 -2.04224512e-01
3.16976964e-01 -7.78209090e-01 1.29393470e+00 -1.27952591e-01
-1.22323096e+00 1.40198493e+00 1.80772468e-01 2.48790398e-01
1.03797424e+00 -2.50153214e-01 -3.19082886e-01 2.93995515e-02
-5.07271104e-02 6.69832647e-01 9.00863707e-01 -1.47721505e+00
-4.96890619e-02 -4.52007383e-01 1.83685496e-01 2.76986845e-02
-5.93903884e-02 1.36188284e-01 -1.06536543e+00 -1.04498672e+00
1.35296240e-01 -9.52667534e-01 -6.61682039e-02 6.78386569e-01
-2.45898426e-01 -5.04980385e-01 7.84175456e-01 -1.35525808e-01
9.91580367e-01 -2.13366199e+00 1.01648092e-01 2.08217248e-01
1.73173502e-01 4.10990894e-01 -6.70850039e-01 7.99645841e-01
1.80557355e-01 4.15507942e-01 -9.82551053e-02 -5.56468964e-01
4.39564407e-01 -8.47539827e-02 -6.78335369e-01 1.07078413e-02
2.53028959e-01 1.02142060e+00 -8.94576550e-01 -4.75841969e-01
1.29553005e-01 2.48152316e-01 -2.17579588e-01 2.77902961e-01
-2.62105823e-01 -1.22764401e-01 -8.90561223e-01 1.03097200e+00
1.05240822e+00 -2.75843143e-01 -1.35406643e-01 -1.35714009e-01
1.59736887e-01 -2.23433763e-01 -1.30704725e+00 2.10412121e+00
-2.39069775e-01 5.51741183e-01 -8.69770534e-03 -3.76442879e-01
1.36048484e+00 3.30121629e-03 1.97526723e-01 -5.90497017e-01
-2.82157511e-01 3.14081222e-01 -4.78913724e-01 -1.25977457e-01
6.30708516e-01 -1.04576740e-02 -2.15832755e-01 9.19019878e-01
-2.49191850e-01 -1.15897930e+00 1.44661486e-01 2.87833005e-01
1.23286593e+00 1.65550932e-01 -1.58357292e-01 -1.09702803e-01
2.28074923e-01 5.12887463e-02 -1.82566062e-01 1.00307989e+00
-7.79026523e-02 1.10932982e+00 4.14358497e-01 -8.04068029e-01
-1.26937056e+00 -1.34828448e+00 -5.15265241e-02 7.09735513e-01
3.99746269e-01 -4.91694033e-01 -2.16212288e-01 -5.72784126e-01
6.09117985e-01 3.80373240e-01 -6.05833113e-01 9.37972665e-02
-7.27166235e-01 -7.53901899e-02 6.10348463e-01 5.53081572e-01
3.48546147e-01 -1.22924578e+00 -3.34954798e-01 3.37205790e-02
3.17913562e-01 -1.03055573e+00 -8.23143303e-01 -6.83981359e-01
-9.16279256e-01 -1.04450548e+00 -1.40613627e+00 -6.83254242e-01
6.07738316e-01 8.59130681e-01 1.86799276e+00 4.08885330e-01
-5.50458968e-01 7.95886636e-01 -2.56633312e-01 -6.22217000e-01
-1.99996427e-01 2.10493878e-01 -2.97606975e-01 -4.19363052e-01
1.70139208e-01 -6.18070424e-01 -6.06142640e-01 6.88504040e-01
-1.02954054e+00 4.48624492e-02 6.85670614e-01 5.12811065e-01
6.51519120e-01 -5.14647007e-01 4.79785740e-01 -6.55271947e-01
9.02658165e-01 -1.56402245e-01 -7.95597076e-01 5.17207861e-01
-2.11791351e-01 -1.76328585e-01 3.21085334e-01 -6.78441405e-01
-6.71868145e-01 -3.85003448e-01 1.05457209e-01 -7.94122934e-01
-1.15128718e-01 1.70777842e-01 -1.68269292e-01 -1.53535515e-01
4.73118603e-01 1.44025773e-01 -8.09871331e-02 -8.83520901e-01
4.81972575e-01 4.28763181e-01 2.49218732e-01 -1.20667100e+00
1.15717590e+00 4.02683198e-01 -5.24997897e-02 -7.41278231e-01
-5.59756696e-01 -2.72901058e-01 -4.94346380e-01 -1.05240807e-01
1.50701672e-01 -7.32724905e-01 -4.70933735e-01 5.66539526e-01
-1.52510941e+00 -2.98744053e-01 -1.75829023e-01 4.89937589e-02
-4.97375011e-01 3.83430690e-01 -3.87211740e-01 -7.55479574e-01
-5.10617852e-01 -1.08528936e+00 1.67987764e+00 -5.81350960e-02
-2.35058203e-01 -5.62802911e-01 2.43632406e-01 6.58674613e-02
6.78006649e-01 2.71105856e-01 1.31871855e+00 -3.91041368e-01
-1.04336452e+00 -7.33996570e-01 -7.94369698e-01 1.45857148e-02
-3.29775400e-02 2.04580501e-01 -8.47347021e-01 -2.97016710e-01
-5.12128294e-01 -6.45164788e-01 6.64717913e-01 -8.24970827e-02
1.43700659e+00 4.61513966e-01 -3.99373233e-01 1.57539129e-01
1.31969833e+00 9.69454572e-02 6.05661452e-01 -2.09137797e-01
4.81521279e-01 4.52862561e-01 5.99300146e-01 4.33649063e-01
2.51510352e-01 8.92370999e-01 2.34202474e-01 1.23807369e-02
-3.88744056e-01 -4.41354096e-01 -2.52232522e-01 8.91552508e-01
-2.81311780e-01 -3.22583318e-01 -9.64135289e-01 4.17106748e-01
-1.52714598e+00 -9.31907773e-01 1.76348954e-01 2.27798533e+00
6.66784763e-01 1.63489670e-01 2.11003765e-01 -1.02700122e-01
7.36379087e-01 5.03906429e-01 -6.82346940e-01 -5.80111817e-02
-2.62627602e-02 4.50614721e-01 -2.03985527e-01 -4.88681905e-02
-9.10128951e-01 9.28782821e-01 6.89801693e+00 1.15622127e+00
-1.03504682e+00 -4.78730321e-01 4.79754537e-01 -1.32032186e-01
-6.24349713e-01 -2.28646189e-01 -5.38974166e-01 3.32234353e-01
6.83674365e-02 -2.85744965e-01 5.60124159e-01 9.77070749e-01
-1.93829939e-01 2.27067754e-01 -1.29581463e+00 1.35268795e+00
1.65191308e-01 -1.36984301e+00 4.86715943e-01 -5.60021140e-02
7.31850803e-01 -1.27541840e-01 1.35141119e-01 3.47405106e-01
3.06736618e-01 -1.14634442e+00 8.01974952e-01 8.93306255e-01
1.10356104e+00 -4.43243563e-01 3.47776324e-01 1.54013947e-01
-1.34074473e+00 4.13423181e-01 -5.80684900e-01 3.05581570e-01
2.95932088e-02 4.29369062e-01 -3.93257111e-01 6.13784373e-01
4.83267993e-01 7.97077775e-01 -6.84560657e-01 1.42591965e+00
1.03068709e-01 1.01906821e-01 -3.08954686e-01 -4.64864910e-01
1.41609609e-01 -2.41709292e-01 4.49265480e-01 1.13953626e+00
3.94481122e-01 1.90527514e-01 2.07060017e-02 1.08599794e+00
-3.54805678e-01 9.55877155e-02 -9.29161906e-01 -3.23699862e-01
7.51943886e-01 1.23044276e+00 -8.72826397e-01 -3.98325890e-01
-2.85118610e-01 7.89680839e-01 3.24453503e-01 3.44640672e-01
-6.48298681e-01 -5.27712107e-01 6.71052635e-01 1.15465730e-01
2.53997326e-01 -5.57065785e-01 -3.13422024e-01 -1.04314435e+00
2.52587557e-01 -9.09792662e-01 7.48103708e-02 -1.15436721e+00
-1.97153044e+00 5.80622077e-01 1.19268149e-01 -1.31363201e+00
1.47303790e-01 -6.14410698e-01 -7.20849574e-01 9.15663302e-01
-1.47944844e+00 -1.22781074e+00 -8.01369846e-01 4.44398746e-02
7.70338774e-01 -1.24775909e-01 9.15545702e-01 4.50076520e-01
-1.04499497e-01 4.76426125e-01 -5.08481376e-02 1.48295432e-01
1.05103242e+00 -1.06395686e+00 1.23299837e+00 -8.33609849e-02
3.51191014e-01 7.05803514e-01 2.59121746e-01 -5.50950348e-01
-1.85117483e+00 -8.01873505e-01 6.87876999e-01 -9.48595226e-01
5.64275026e-01 -6.50113344e-01 -9.68550622e-01 1.52932718e-01
-1.06668353e-01 2.15977252e-01 2.27678567e-01 -1.36141777e-01
-6.30415976e-01 1.66184083e-01 -1.13240457e+00 7.74566174e-01
1.85859334e+00 -6.04958594e-01 -6.27162278e-01 2.41037652e-01
4.66936350e-01 -6.21482551e-01 -9.66073096e-01 3.21618199e-01
1.09419429e+00 -7.85397410e-01 1.45580316e+00 -8.07400286e-01
6.38248265e-01 -8.73439945e-03 -2.21986458e-01 -9.99567986e-01
-1.24828257e-01 -2.97207892e-01 8.82489793e-03 1.03900850e+00
1.03773229e-01 -2.47047409e-01 9.27059293e-01 7.06666648e-01
5.15401363e-02 -1.01536775e+00 -4.39584732e-01 -1.00833213e+00
3.81127328e-01 -4.27007616e-01 1.06489313e+00 8.06612015e-01
-7.77876198e-01 2.32894614e-01 -1.05218515e-01 -3.95821452e-01
6.35167658e-01 7.71414459e-01 1.37594104e+00 -1.59242499e+00
-1.78964976e-02 -9.87353086e-01 -3.84405881e-01 -1.44750452e+00
1.51588544e-01 -8.77633035e-01 -4.14302558e-01 -1.45820916e+00
2.61420816e-01 -9.18300927e-01 1.26819000e-01 2.17953593e-01
9.20112580e-02 4.62307125e-01 5.12866378e-01 3.73045385e-01
-6.50482416e-01 5.35540104e-01 1.81562650e+00 -4.18087661e-01
1.84551537e-01 1.20197803e-01 -4.42016363e-01 4.84275311e-01
2.46682569e-01 -2.60987520e-01 -4.88365144e-01 -7.59502172e-01
4.09542536e-03 6.55880347e-02 6.33049190e-01 -6.11262083e-01
1.01259388e-01 -1.97655126e-01 5.41299701e-01 -1.03930962e+00
6.92054570e-01 -8.81142974e-01 3.46432924e-01 5.19835278e-02
-3.28273326e-01 2.49062747e-01 2.06023589e-01 4.48825985e-01
-4.46279496e-02 -3.01748574e-01 4.89772350e-01 -4.19857711e-01
-4.45733279e-01 6.11031950e-01 2.34012470e-01 2.46902660e-01
7.65507400e-01 -4.40687478e-01 -3.12491626e-01 -4.03839707e-01
-4.74833637e-01 6.99427798e-02 8.50558221e-01 7.54852295e-01
8.84532809e-01 -1.83513701e+00 -7.70802259e-01 1.00843661e-01
3.95834863e-01 7.13586807e-03 4.20685299e-02 8.83990079e-02
-5.28623998e-01 3.65742356e-01 -2.50917763e-01 -4.75305289e-01
-1.03605533e+00 4.57239985e-01 1.07802609e-02 -1.54111251e-01
-6.54346526e-01 5.27747273e-01 3.78667489e-02 -5.66865861e-01
4.07049209e-01 -2.05967993e-01 2.73241878e-01 -3.51446122e-02
4.87739593e-01 5.25543332e-01 1.28226638e-01 -9.75494385e-02
-1.93744093e-01 1.07209611e+00 -6.76296428e-02 -1.05726942e-02
1.44352615e+00 4.87780184e-01 1.72309771e-01 3.75641942e-01
1.20725989e+00 6.81768581e-02 -9.98554647e-01 -4.38229471e-01
1.94777846e-01 -9.90234375e-01 -4.61752534e-01 -8.91580641e-01
-9.53457355e-01 9.52902913e-01 3.85755658e-01 3.80573839e-01
5.80264211e-01 1.20553538e-01 7.60230482e-01 5.39802313e-01
9.78406012e-01 -6.76713705e-01 4.70589280e-01 4.00052041e-01
1.67151248e+00 -1.05150509e+00 1.37797534e-01 -4.65184122e-01
-3.99751753e-01 1.20524824e+00 4.88513201e-01 -5.74782014e-01
5.46452463e-01 1.40522480e-01 2.19464954e-02 -4.28508431e-01
-7.50338674e-01 8.48889574e-02 3.97829592e-01 4.36270028e-01
3.42735738e-01 -5.43405525e-02 3.53596322e-02 4.63570714e-01
6.41416013e-02 2.36752648e-02 -1.74965970e-02 1.04316413e+00
2.55724806e-02 -1.42136145e+00 -3.09255362e-01 5.46055973e-01
2.05887944e-01 1.76956400e-01 -1.00448501e+00 9.86986995e-01
-5.27478039e-01 3.74055386e-01 5.71366660e-02 -1.30999714e-01
9.98141527e-01 -1.19753830e-01 8.23662221e-01 -4.46528763e-01
-4.08399969e-01 -1.98780581e-01 1.45459753e-02 -5.12183666e-01
-1.42075673e-01 -6.26806676e-01 -6.38779044e-01 -5.03999949e-01
-2.61156082e-01 -1.19253881e-01 6.80673897e-01 2.77656645e-01
6.77117944e-01 -1.16383694e-01 6.79417789e-01 -1.30598676e+00
-8.45058799e-01 -8.86654437e-01 -6.16074860e-01 5.88656902e-01
-1.75972059e-01 -1.01372814e+00 -1.52057856e-01 -3.19249779e-01] | [8.556492805480957, -3.5636508464813232] |
45a21190-341e-4963-8f58-bfbda7120061 | consensus-neural-network-for-medical-imaging | 1906.03639 | null | https://arxiv.org/abs/1906.03639v1 | https://arxiv.org/pdf/1906.03639v1.pdf | Consensus Neural Network for Medical Imaging Denoising with Only Noisy Training Samples | Deep neural networks have been proved efficient for medical image denoising. Current training methods require both noisy and clean images. However, clean images cannot be acquired for many practical medical applications due to naturally noisy signal, such as dynamic imaging, spectral computed tomography, arterial spin labeling magnetic resonance imaging, etc. In this paper we proposed a training method which learned denoising neural networks from noisy training samples only. Training data in the acquisition domain was split to two subsets and the network was trained to map one noisy set to the other. A consensus loss function was further proposed to efficiently combine the outputs from both subsets. A mathematical proof was provided that the proposed training scheme was equivalent to training with noisy and clean samples when the noise in the two subsets was uncorrelated and zero-mean. The method was validated on Low-dose CT Challenge dataset and NYU MRI dataset and achieved improved performance compared to existing unsupervised methods. | ['Quanzheng Li', 'Kyungsang Kim', 'Kuang Gong', 'Dufan Wu'] | 2019-06-09 | null | null | null | null | ['medical-image-denoising'] | ['computer-vision'] | [ 6.03585780e-01 1.39157489e-01 2.20315173e-01 -6.05266988e-01
-7.81220019e-01 3.22188176e-02 1.35714829e-01 -1.04722090e-01
-6.41484618e-01 9.83639836e-01 1.15215495e-01 9.31167901e-02
-3.87906671e-01 -6.30866051e-01 -5.33205390e-01 -1.14146340e+00
-3.15166414e-01 4.22006458e-01 -4.28230762e-02 4.92800921e-02
-2.43948147e-01 8.32189620e-02 -1.06501925e+00 2.24127933e-01
9.63758767e-01 9.58441854e-01 4.78385359e-01 5.13565123e-01
5.35324886e-02 1.09221768e+00 -5.20451725e-01 2.35898018e-01
4.75435704e-01 -8.16524565e-01 -5.66800237e-01 9.25002098e-02
1.83117166e-01 -3.60999256e-01 -5.35680771e-01 1.49730217e+00
9.01221514e-01 4.48573589e-01 3.76959056e-01 -6.40338123e-01
-3.64285350e-01 6.22130454e-01 -3.34545493e-01 3.74394536e-01
-3.79059047e-01 -9.12483707e-02 2.13070750e-01 -6.10548496e-01
7.36550927e-01 7.21086919e-01 8.01936626e-01 7.81912863e-01
-1.17465746e+00 -5.58691919e-01 -2.09628150e-01 2.63169855e-01
-8.47394526e-01 -2.89228618e-01 7.24541068e-01 -2.54630357e-01
4.98622864e-01 1.18914858e-01 5.16514540e-01 9.64843214e-01
3.38857085e-01 6.14711106e-01 1.61370492e+00 -2.01770589e-01
1.94075018e-01 -1.35301009e-01 2.23808840e-01 5.27744591e-01
2.44173437e-01 2.52353281e-01 -1.23910576e-01 6.27156496e-02
7.42662668e-01 5.28638400e-02 -6.16558135e-01 -1.50263935e-01
-1.04225647e+00 6.55031025e-01 6.33833289e-01 7.71303833e-01
-7.35851467e-01 4.25155880e-03 5.97836614e-01 4.11302119e-01
4.16436642e-01 3.36079955e-01 -3.72018181e-02 9.26472321e-02
-1.10253048e+00 -3.16206545e-01 6.11065805e-01 3.57321262e-01
3.39823872e-01 4.09258872e-01 -2.83129103e-02 9.16847110e-01
5.33587337e-02 4.03212309e-01 8.31845999e-01 -1.16838133e+00
1.61132663e-01 -1.34827912e-01 -1.02726795e-01 -9.51232076e-01
-5.61470211e-01 -9.08536196e-01 -1.65365374e+00 1.59783646e-01
2.72055089e-01 -2.69808412e-01 -1.25471902e+00 1.69547987e+00
2.04884320e-01 5.31261504e-01 5.56304641e-02 1.29391015e+00
8.98164570e-01 4.97189641e-01 -6.48487434e-02 -6.54890776e-01
7.43452370e-01 -7.40948021e-01 -1.24646235e+00 -3.10118292e-02
1.99804083e-01 -5.45321584e-01 4.46131289e-01 8.67944479e-01
-1.26702952e+00 -5.27456045e-01 -1.29438198e+00 2.31786802e-01
1.88904881e-01 -7.93376490e-02 4.46215659e-01 5.96814871e-01
-9.10798907e-01 1.14974093e+00 -1.18425417e+00 1.56739995e-01
6.62011087e-01 4.15438533e-01 -4.64884490e-01 -3.86854261e-01
-1.32322752e+00 8.64617467e-01 2.98270077e-01 5.05492985e-01
-1.09619856e+00 -4.05274361e-01 -6.70498371e-01 -1.47490576e-01
9.35453027e-02 -6.22030854e-01 1.03295529e+00 -1.31241250e+00
-1.34131539e+00 5.36566257e-01 1.04759419e-02 -7.14268684e-01
6.32245779e-01 -1.33364409e-01 -4.28378493e-01 4.77779031e-01
1.75650910e-01 1.13056056e-01 1.05141759e+00 -1.30820572e+00
-7.91078061e-02 -4.46519315e-01 -5.16417921e-01 4.92337830e-02
7.65065625e-02 -2.75083780e-01 2.89064343e-03 -7.44037628e-01
7.05548406e-01 -5.63239813e-01 -4.94586021e-01 -4.22595233e-01
-2.59794176e-01 3.35205257e-01 6.14979506e-01 -1.01639283e+00
8.17555368e-01 -2.15647411e+00 4.66479175e-02 5.05217493e-01
4.41504210e-01 2.64631987e-01 -8.97708014e-02 -1.48987919e-01
-5.22271335e-01 -1.29373387e-01 -6.36639833e-01 -2.67464340e-01
-4.27488923e-01 5.30382812e-01 1.09786697e-01 7.98156083e-01
-7.20091909e-02 4.34246302e-01 -1.19214153e+00 -3.99447143e-01
2.46561423e-01 6.92072034e-01 -4.94775437e-02 2.32371494e-01
3.71978253e-01 1.13349545e+00 -2.60749727e-01 1.84097588e-01
8.69228065e-01 -7.95375258e-02 3.62945288e-01 -4.27412331e-01
2.86769003e-01 -1.55003127e-02 -1.15024734e+00 1.79200852e+00
-3.58486056e-01 5.41246951e-01 6.12593174e-01 -1.62447965e+00
7.91141868e-01 6.15752161e-01 8.82144153e-01 -9.48928833e-01
3.98086548e-01 3.45801502e-01 2.96107173e-01 -8.63572776e-01
-8.28662887e-02 -7.29458451e-01 4.07537788e-01 2.86455512e-01
2.36510769e-01 -4.41343412e-02 8.66161883e-02 -3.02799661e-02
1.23827648e+00 -1.10773377e-01 -1.20816298e-01 -2.13705242e-01
4.19284791e-01 -1.20627210e-01 7.86155820e-01 1.22713590e+00
-5.61879516e-01 9.55680907e-01 1.52884856e-01 -4.85579878e-01
-9.15807486e-01 -9.44898367e-01 -3.08870614e-01 3.84039670e-01
2.04526633e-01 1.10536732e-01 -8.79206717e-01 -6.44485056e-01
-5.91525257e-01 4.17372197e-01 -6.21998310e-01 -2.67442495e-01
-8.45486462e-01 -1.01408052e+00 4.22209680e-01 2.74540722e-01
6.92254663e-01 -9.55108643e-01 -4.68585998e-01 5.49273372e-01
-3.86774331e-01 -9.96700168e-01 -3.82214516e-01 7.02985406e-01
-1.13918102e+00 -1.13226318e+00 -7.72878826e-01 -9.44460750e-01
8.39015841e-01 9.37159210e-02 9.60797131e-01 2.40296602e-01
-3.02277416e-01 8.65977705e-02 -1.72811225e-01 -1.65105611e-01
-4.71135050e-01 -3.03653955e-01 2.41025090e-01 -2.75342702e-03
-3.00721806e-02 -7.95648932e-01 -7.71654010e-01 6.12311028e-02
-1.23241174e+00 -1.91585138e-01 5.71533978e-01 1.45997882e+00
8.00217807e-01 4.58099872e-01 7.32223928e-01 -9.83259261e-01
6.84856772e-01 -4.52235550e-01 -1.32640004e-01 3.99273587e-03
-4.23776418e-01 1.07905723e-01 7.54283905e-01 -3.59578699e-01
-1.11913300e+00 1.40161723e-01 -4.66861665e-01 -3.07053775e-01
-2.67901123e-01 6.48643374e-01 7.36914352e-02 1.50150422e-03
6.99044704e-01 3.92974675e-01 3.84417713e-01 -4.01016057e-01
-5.32156378e-02 5.17229497e-01 8.62796307e-01 -4.56923842e-01
3.49797606e-01 5.31543851e-01 8.62400681e-02 -9.18241382e-01
-7.07247198e-01 -2.40577266e-01 -5.82134724e-01 -1.55716076e-01
8.75926614e-01 -6.21272445e-01 -4.31071192e-01 6.00362659e-01
-9.01183248e-01 -3.40723664e-01 -5.10191619e-01 9.41316485e-01
-4.07029063e-01 7.78604746e-01 -8.63635600e-01 -5.61170340e-01
-6.22697413e-01 -1.30823517e+00 3.51267785e-01 1.21633820e-01
1.65604174e-01 -9.85480189e-01 1.04096919e-01 1.57777548e-01
6.99945569e-01 4.80341017e-01 7.35665262e-01 -6.19641423e-01
-5.99799380e-02 -3.39831293e-01 -7.07089435e-03 1.09428895e+00
3.44000667e-01 -6.46041334e-01 -8.16260815e-01 -3.31796348e-01
9.23758030e-01 -3.41263592e-01 9.99967039e-01 1.00859058e+00
1.22589207e+00 2.16567479e-02 3.12784463e-02 7.55811632e-01
1.48066044e+00 3.41647863e-01 7.68499434e-01 1.81314200e-01
5.18447638e-01 3.65278006e-01 6.51687235e-02 -5.90599626e-02
-2.63880938e-01 1.39365464e-01 3.37538451e-01 -2.28848636e-01
-3.02336723e-01 3.06425244e-01 -1.49302641e-02 1.13048995e+00
-1.18192472e-01 9.40499753e-02 -6.94705963e-01 5.53280115e-01
-1.72883856e+00 -1.08579159e+00 -3.49033356e-01 2.09432173e+00
8.51707399e-01 1.63386047e-01 -3.43721479e-01 2.90890694e-01
7.15765655e-01 -1.49133995e-01 -5.52614152e-01 8.30423608e-02
-2.56974339e-01 7.05115139e-01 5.07849097e-01 4.88247335e-01
-1.16588092e+00 2.41118111e-02 6.39632320e+00 7.51383841e-01
-1.39691293e+00 6.66932523e-01 7.98925281e-01 -2.14530341e-02
4.40281965e-02 -4.13381308e-01 1.44850567e-01 4.69403088e-01
8.29752505e-01 2.70530373e-01 3.30896348e-01 5.35411894e-01
4.44709957e-01 -2.87446976e-01 -6.39522016e-01 1.00097311e+00
-1.31724775e-01 -1.04813337e+00 -4.64946866e-01 -4.31169987e-01
8.46103251e-01 2.92412132e-01 -6.24886192e-02 1.76642880e-01
1.28342718e-01 -1.29677749e+00 2.04428524e-01 7.12156951e-01
5.18632054e-01 -5.37236929e-01 1.16811740e+00 6.26869321e-01
-5.10214269e-01 1.41462535e-01 -3.23123276e-01 8.12138692e-02
2.45640159e-01 9.79413152e-01 -5.23784518e-01 8.14919412e-01
6.63954318e-01 6.76640034e-01 -8.65799189e-02 1.06698024e+00
-1.14079662e-01 8.71011078e-01 -3.51346105e-01 5.56010664e-01
2.82825351e-01 -5.59647381e-01 6.26256883e-01 9.63001430e-01
1.98348358e-01 2.64297277e-01 2.65954584e-01 4.69187856e-01
3.34139238e-03 -1.71608955e-01 -4.26365346e-01 3.93671691e-01
-4.80853654e-02 1.25951576e+00 -9.23452199e-01 -5.68851233e-01
-2.14330450e-01 9.06239688e-01 -9.22807232e-02 4.73114520e-01
-5.51432014e-01 -3.59936327e-01 -1.19885013e-01 1.17349088e-01
5.31777218e-02 -1.72606423e-01 -5.51679075e-01 -1.17387700e+00
-1.27569407e-01 -7.26199508e-01 2.74375886e-01 -5.52443802e-01
-1.50093222e+00 9.95137632e-01 -6.84155971e-02 -1.04211235e+00
-2.70778090e-02 -4.39472228e-01 -5.55462718e-01 1.07101762e+00
-1.28937399e+00 -5.29277265e-01 -4.78612721e-01 5.83449543e-01
2.93191224e-01 -6.95332587e-02 6.66543007e-01 8.38016331e-01
-5.02907097e-01 1.65347531e-01 5.49583912e-01 2.44271815e-01
6.40143812e-01 -1.44932485e+00 -3.23604643e-01 9.88294542e-01
-3.06997001e-01 6.96989894e-01 7.18482256e-01 -7.30739057e-01
-8.80782247e-01 -1.04765427e+00 2.74945676e-01 3.20422411e-01
2.34288782e-01 2.25006446e-01 -1.19575667e+00 2.92101979e-01
4.63205129e-01 5.53427935e-01 3.81811976e-01 -5.16149998e-01
2.76787221e-01 -1.98907167e-01 -1.50719035e+00 1.06992926e-02
6.02272034e-01 -2.06224874e-01 -5.69486976e-01 4.76858974e-01
2.10715517e-01 -7.30678439e-01 -1.05528867e+00 4.73052680e-01
3.05274040e-01 -8.21928144e-01 8.33582163e-01 -5.96832633e-01
3.56500030e-01 -2.32919455e-01 -6.55483231e-02 -1.58775496e+00
-2.60494083e-01 -3.89289737e-01 -7.41641074e-02 5.33261120e-01
-5.73111176e-02 -5.24972439e-01 6.74866378e-01 1.82405159e-01
-4.88804579e-01 -5.89257300e-01 -1.11663258e+00 -6.22745037e-01
6.14901334e-02 -3.72065067e-01 4.90873605e-02 9.61738288e-01
-4.64906365e-01 2.32186556e-01 -5.91300309e-01 1.13738589e-01
1.14387047e+00 -2.85769254e-01 1.92394890e-02 -9.04301286e-01
-3.62161338e-01 -1.51012555e-01 -2.77242541e-01 -7.86263824e-01
-5.48678311e-03 -1.01772034e+00 4.39830601e-01 -1.73540938e+00
1.33186251e-01 -4.71896291e-01 -6.26248777e-01 2.86783814e-01
-2.65033573e-01 4.39937651e-01 -1.60783604e-01 2.63173640e-01
-2.72687525e-01 3.34635466e-01 1.47113526e+00 -3.97090673e-01
-3.05837281e-02 1.46748647e-01 -3.44236284e-01 5.80888569e-01
8.33931327e-01 -7.60351777e-01 -5.23119509e-01 -3.78498524e-01
-1.69301257e-01 2.58521467e-01 3.07759017e-01 -1.17172146e+00
2.27162957e-01 2.51296669e-01 6.01069808e-01 -4.10322815e-01
6.65704086e-02 -1.03383994e+00 4.42409247e-01 7.61115372e-01
-3.93579572e-01 -4.60246593e-01 -1.29546180e-01 4.50540304e-01
-5.15202880e-01 -5.58422863e-01 1.19601548e+00 -5.69738686e-01
-3.71054083e-01 1.78169161e-01 -2.33799636e-01 2.17517056e-02
6.67080164e-01 -2.66035236e-02 2.94546504e-02 -5.47879457e-01
-1.43379045e+00 -6.09364943e-04 -1.34687781e-01 -5.41966334e-02
8.76835585e-01 -1.26606643e+00 -7.12624907e-01 2.08335638e-01
-6.90339684e-01 1.77648827e-01 6.75952256e-01 1.41349387e+00
-5.52239597e-01 -4.23942953e-02 -3.82071584e-01 -7.21049070e-01
-9.57678914e-01 4.09681767e-01 7.81986535e-01 -4.94382501e-01
-8.52435291e-01 5.46621025e-01 -1.19195290e-01 -4.26239431e-01
2.64828593e-01 -2.79922605e-01 -2.75056511e-02 -2.37509161e-01
5.91906011e-01 2.59415358e-01 5.11325479e-01 -5.85487664e-01
-3.29895616e-02 2.78373003e-01 -1.91794019e-02 -3.09847929e-02
1.71551633e+00 -1.62502304e-01 -3.09194386e-01 2.07926601e-01
1.39079416e+00 -5.62260449e-01 -1.04869998e+00 -4.27620232e-01
-1.04609102e-01 -1.58851922e-01 5.33941031e-01 -8.26150417e-01
-1.45682192e+00 8.72560918e-01 1.22525775e+00 2.62505382e-01
1.44782484e+00 -5.31341612e-01 9.74234283e-01 2.91351199e-01
2.27282181e-01 -1.10116017e+00 -6.18087389e-02 3.24413627e-01
6.98324084e-01 -1.36908662e+00 6.15235558e-03 -7.05674216e-02
-5.34824967e-01 1.22811413e+00 2.96701491e-01 -4.59034622e-01
7.88156688e-01 3.86094958e-01 1.84233591e-01 -4.24015701e-01
-1.63711339e-01 2.41486691e-02 1.70331672e-01 6.90198779e-01
3.19503069e-01 -1.28248394e-01 -5.43559730e-01 7.60796309e-01
2.83432335e-01 4.23281997e-01 4.67549533e-01 1.04990959e+00
-3.63606364e-01 -8.53403270e-01 -5.36657691e-01 7.90320694e-01
-7.96867073e-01 9.38646775e-03 4.01748985e-01 5.43003201e-01
3.46420556e-01 8.28327000e-01 -2.19143748e-01 -1.20327957e-01
2.95515060e-01 -2.60071736e-02 6.61913812e-01 -5.19966245e-01
-6.54455006e-01 3.38836938e-01 -1.47361889e-01 -4.48843688e-01
-8.27364624e-01 -3.43460590e-01 -1.25476670e+00 1.43702462e-01
-1.83385462e-01 3.24798167e-01 7.36531377e-01 9.65676486e-01
-1.25926986e-01 9.35797215e-01 6.44887388e-01 -8.67615581e-01
-7.53351569e-01 -1.21474981e+00 -8.09906662e-01 6.67853951e-01
6.10836625e-01 -4.16195393e-01 -3.27384621e-01 2.67343849e-01] | [13.311063766479492, -2.486466884613037] |
5e58ccd7-27fb-4509-9e22-bb12f2aa259c | exploring-data-redundancy-in-real-world-image | 2306.14113 | null | https://arxiv.org/abs/2306.14113v1 | https://arxiv.org/pdf/2306.14113v1.pdf | Exploring Data Redundancy in Real-world Image Classification through Data Selection | Deep learning models often require large amounts of data for training, leading to increased costs. It is particularly challenging in medical imaging, i.e., gathering distributed data for centralized training, and meanwhile, obtaining quality labels remains a tedious job. Many methods have been proposed to address this issue in various training paradigms, e.g., continual learning, active learning, and federated learning, which indeed demonstrate certain forms of the data valuation process. However, existing methods are either overly intuitive or limited to common clean/toy datasets in the experiments. In this work, we present two data valuation metrics based on Synaptic Intelligence and gradient norms, respectively, to study the redundancy in real-world image data. Novel online and offline data selection algorithms are then proposed via clustering and grouping based on the examined data values. Our online approach effectively evaluates data utilizing layerwise model parameter updates and gradients in each epoch and can accelerate model training with fewer epochs and a subset (e.g., 19%-59%) of data while maintaining equivalent levels of accuracy in a variety of datasets. It also extends to the offline coreset construction, producing subsets of only 18%-30% of the original. The codes for the proposed adaptive data selection and coreset computation are available (https://github.com/ZhenyuTANG2023/data_selection). | ['Xiaosong Wang', 'Shaoting Zhang', 'Zhenyu Tang'] | 2023-06-25 | null | null | null | null | ['active-learning', 'active-learning'] | ['methodology', 'natural-language-processing'] | [ 1.31349072e-01 -1.27650037e-01 -2.73584545e-01 -7.24916041e-01
-8.97130430e-01 -3.23891997e-01 1.83319375e-02 5.09011269e-01
-7.16251552e-01 8.87638390e-01 -3.21819365e-01 -1.05528042e-01
-7.37812221e-01 -6.69605494e-01 -4.38427567e-01 -9.64777291e-01
-1.02799274e-01 5.49772799e-01 -5.48523851e-02 1.61618561e-01
2.14707479e-01 3.95681500e-01 -1.46890628e+00 2.81485487e-02
1.19203472e+00 1.43594170e+00 4.84275043e-01 2.99477547e-01
-2.95857359e-02 5.82548916e-01 -5.28153658e-01 -4.05663401e-01
2.52300650e-01 -9.95043889e-02 -5.94197333e-01 -5.33047654e-02
7.20791742e-02 -3.45065922e-01 1.20637808e-02 1.11616683e+00
1.00012004e+00 6.05161786e-02 2.00122193e-01 -1.32022548e+00
-2.59576768e-01 6.45875096e-01 -4.27687496e-01 2.89862573e-01
-5.02250433e-01 1.47234485e-01 6.91720486e-01 -9.25315559e-01
3.01194191e-01 6.24042392e-01 4.20910478e-01 6.98659599e-01
-1.10662341e+00 -1.02060640e+00 -2.74338741e-02 4.40628827e-01
-1.26037157e+00 -6.82600796e-01 7.99540341e-01 -2.60719091e-01
5.49384892e-01 3.90755892e-01 7.59385228e-01 1.00096679e+00
1.56611335e-02 8.93969655e-01 1.15076125e+00 -6.22660965e-02
6.70419693e-01 1.58844382e-01 4.15829808e-01 5.12425780e-01
5.69947600e-01 -3.80374007e-02 -7.62227833e-01 -2.06884146e-01
5.71219087e-01 2.50677556e-01 -1.71436444e-01 -1.61087826e-01
-1.12320626e+00 6.11798406e-01 3.84219676e-01 9.05355588e-02
-3.93774629e-01 -6.42309040e-02 6.87500894e-01 2.59197682e-01
5.08729219e-01 4.75567192e-01 -6.44943655e-01 -2.17130974e-01
-1.01523483e+00 1.19321279e-01 3.64439219e-01 9.25502241e-01
8.93155754e-01 -1.13671437e-01 -1.41196866e-02 9.38088536e-01
-1.08966671e-01 1.78359061e-01 6.36864364e-01 -8.87365937e-01
3.54534447e-01 7.08697855e-01 -9.05825384e-03 -9.98986065e-01
-6.53997600e-01 -7.58909523e-01 -1.24476683e+00 6.53801039e-02
1.27836972e-01 -3.12942982e-01 -6.50846004e-01 1.79323304e+00
6.57737195e-01 7.67875090e-02 -3.22150797e-01 8.38571072e-01
7.56587684e-01 8.84259269e-02 8.41707513e-02 -5.34056664e-01
9.73456800e-01 -8.92012715e-01 -7.23661125e-01 5.08064069e-02
5.83062649e-01 -4.04429585e-01 1.21233177e+00 6.37380540e-01
-1.16690040e+00 -4.30728257e-01 -9.43003178e-01 9.94957387e-02
-1.33380070e-01 2.41311803e-01 9.67947602e-01 4.97308403e-01
-1.06092739e+00 7.55531073e-01 -1.10090113e+00 1.46320239e-01
1.00993443e+00 6.24623299e-01 5.39773889e-03 -1.45856617e-02
-1.02965891e+00 4.21104640e-01 3.92156720e-01 2.19728529e-01
-1.09359217e+00 -8.76914442e-01 -2.82105505e-01 6.56615421e-02
2.15019405e-01 -6.85669899e-01 1.06776667e+00 -1.06463516e+00
-1.09400618e+00 6.32430136e-01 1.57107159e-01 -3.74126226e-01
4.91507381e-01 -7.44165629e-02 -2.72016317e-01 1.47003934e-01
-4.44158949e-02 5.78694999e-01 6.61446512e-01 -9.12916660e-01
-5.37223339e-01 -6.62731946e-01 -5.11272289e-02 3.49217713e-01
-1.00410950e+00 -8.97142515e-02 -3.90194327e-01 -6.68050468e-01
1.32651955e-01 -7.85477757e-01 -4.53672141e-01 3.40971082e-01
-2.99899995e-01 -1.19637601e-01 3.37713391e-01 -4.01414543e-01
1.36273086e+00 -2.10533977e+00 -1.04992129e-01 2.70701557e-01
7.46760368e-01 1.41332030e-01 3.52548100e-02 -7.19968230e-02
1.25187002e-02 1.87398329e-01 -2.44859427e-01 -4.62865382e-01
-1.96884364e-01 1.15344405e-01 7.33153522e-02 4.52751160e-01
2.42878497e-02 6.96692288e-01 -8.58148277e-01 -7.84003437e-01
-1.18425399e-01 2.46787429e-01 -4.88056630e-01 1.14887454e-01
1.37041926e-01 5.64567924e-01 -5.66343665e-01 8.68699431e-01
6.60327613e-01 -5.93588591e-01 1.57199919e-01 -2.53023744e-01
1.52721122e-01 8.83394778e-02 -1.04139233e+00 1.89495003e+00
-3.41580540e-01 2.29286492e-01 1.02734819e-01 -1.17222738e+00
7.77965963e-01 2.79908448e-01 9.17235255e-01 -8.62811089e-01
2.90284663e-01 5.23837566e-01 1.64328307e-01 -4.50570077e-01
2.62601525e-01 1.69001222e-01 9.10307616e-02 5.70641816e-01
1.07369803e-01 5.62015533e-01 3.17452669e-01 1.47280604e-01
1.24124992e+00 -2.98637122e-01 1.13334797e-01 -4.50005829e-01
2.48013511e-02 -4.78382176e-03 8.95680726e-01 6.08261108e-01
-5.33810556e-01 3.68398786e-01 1.74084529e-01 -3.68994862e-01
-1.02428150e+00 -8.33415866e-01 -4.05880183e-01 9.45934594e-01
2.93562889e-01 -2.88314015e-01 -8.24484169e-01 -4.28406179e-01
-1.46356747e-01 2.24251971e-01 -5.80645919e-01 -4.25033540e-01
-3.74895811e-01 -1.21427810e+00 4.09565121e-01 3.74351859e-01
5.50793350e-01 -1.05302978e+00 -8.06049883e-01 1.11489080e-01
-6.77298307e-02 -6.91515744e-01 -2.35231802e-01 4.10519749e-01
-1.38812065e+00 -9.61700201e-01 -5.54131627e-01 -6.64918542e-01
1.03036010e+00 2.04754889e-01 1.04530251e+00 3.32504928e-01
-5.17976522e-01 -7.77805001e-02 -4.58737552e-01 -4.63641465e-01
1.84845343e-01 1.71740025e-01 2.56114513e-01 -1.02263002e-03
1.70910314e-01 -5.88527441e-01 -1.05136454e+00 3.20415944e-01
-9.48234856e-01 3.65869462e-01 7.20197499e-01 1.11994171e+00
1.06516075e+00 1.57423779e-01 9.44057167e-01 -1.13281238e+00
5.65919816e-01 -6.46212399e-01 -3.75216842e-01 2.99672484e-01
-1.06008589e+00 -1.15508556e-01 8.40479016e-01 -6.01100624e-01
-8.67238700e-01 3.24179120e-02 -1.20934388e-02 -4.96111959e-01
7.39133731e-02 4.57900554e-01 -1.95832089e-01 -1.74271852e-01
7.82724440e-01 1.03615955e-01 5.13153598e-02 -3.68885815e-01
-7.14312633e-03 6.54824853e-01 1.15715325e-01 -5.57318389e-01
4.08174276e-01 4.35838521e-01 -3.00218880e-01 -2.99641103e-01
-7.51923800e-01 -3.04059833e-01 -3.48747283e-01 -3.60638827e-01
2.68111050e-01 -7.90783107e-01 -6.28158748e-01 6.58401549e-01
-5.82302392e-01 -3.56349081e-01 -3.86893868e-01 6.43524230e-01
-4.41385090e-01 9.97187495e-02 -5.73609233e-01 -6.64164722e-01
-9.83878016e-01 -1.16941762e+00 6.86506867e-01 3.42128247e-01
6.14530444e-02 -7.48981535e-01 -1.22936741e-01 3.55497718e-01
5.71105003e-01 1.74677610e-01 9.81768787e-01 -6.35509968e-01
-4.48295832e-01 -6.41198680e-02 -6.50596768e-02 1.82163104e-01
2.11514354e-01 -1.85666323e-01 -1.00293076e+00 -5.66261947e-01
1.88576072e-01 -7.39091396e-01 6.17354512e-01 4.73087460e-01
1.87516081e+00 -3.63061190e-01 -2.96664089e-01 7.29755282e-01
1.38092327e+00 3.70903403e-01 4.84107971e-01 2.53126919e-01
3.40231836e-01 4.51500922e-01 6.95190012e-01 8.26563060e-01
6.78975806e-02 2.82071143e-01 4.58199531e-01 -3.36456567e-01
9.68484208e-02 1.68151796e-01 -1.35364607e-01 1.25112498e+00
1.45146579e-01 5.04013710e-02 -9.26411748e-01 4.04926777e-01
-1.80274367e+00 -6.25325859e-01 2.21155137e-01 2.32385945e+00
1.37880647e+00 1.39784470e-01 -3.98584008e-02 2.64725864e-01
7.45772302e-01 -3.57272297e-01 -1.25773335e+00 1.43100157e-01
-5.74310906e-02 3.65269780e-01 4.62091237e-01 -1.57279924e-01
-8.93872082e-01 4.09210354e-01 5.23606157e+00 1.10675502e+00
-1.35874367e+00 4.86726820e-01 1.30686677e+00 -7.44680703e-01
-3.39404196e-01 -1.10982485e-01 -5.29119730e-01 7.68079519e-01
7.77584255e-01 -3.89091104e-01 4.90084410e-01 9.87534225e-01
3.98957759e-01 -9.35398638e-02 -9.00989473e-01 1.38753819e+00
-2.44881138e-01 -1.32955801e+00 -1.43397525e-01 -9.44999456e-02
7.53018558e-01 1.85888886e-01 3.40037942e-02 8.73061717e-02
-2.89210328e-03 -9.01090622e-01 6.30653858e-01 5.57498097e-01
9.75363791e-01 -8.16546857e-01 4.52865452e-01 5.81570506e-01
-8.43390107e-01 -3.34053010e-01 -4.79802817e-01 3.21975023e-01
-1.68478563e-01 1.08571947e+00 -4.09031510e-01 3.42994153e-01
1.01622069e+00 5.81801593e-01 -7.62722313e-01 1.42021656e+00
2.90218681e-01 6.92472339e-01 -3.67984712e-01 -2.05541998e-01
-2.02774867e-01 -5.34358323e-02 1.90776303e-01 5.41134298e-01
2.95550257e-01 -5.64763881e-02 -4.65355143e-02 6.09690189e-01
-3.41961771e-01 3.43518227e-01 6.24951161e-02 9.37605575e-02
8.29318166e-01 1.63321924e+00 -7.76848912e-01 -4.32704240e-01
-2.37642806e-02 5.57086527e-01 6.57358587e-01 1.03320390e-01
-8.94440889e-01 -4.44193274e-01 4.64539766e-01 -4.62873131e-02
-3.17644656e-01 -6.93937019e-02 -6.96173847e-01 -1.15989184e+00
2.36767694e-01 -8.86784017e-01 4.32732284e-01 -3.96974564e-01
-1.36606789e+00 6.72606707e-01 -1.81523308e-01 -1.41051674e+00
2.10337579e-01 -4.23993617e-01 -4.20608073e-01 5.31274736e-01
-1.33783126e+00 -4.93227601e-01 -7.88797021e-01 7.74208307e-01
3.32910597e-01 -1.80919722e-01 7.66978264e-01 9.20643806e-01
-9.15246367e-01 9.34985578e-01 3.95368397e-01 4.68578073e-04
5.97353280e-01 -9.27540779e-01 2.89583132e-02 4.26166385e-01
-2.03621052e-02 8.01269829e-01 2.28706524e-01 -3.91572833e-01
-1.47990394e+00 -1.16063368e+00 3.30678672e-01 1.77444384e-01
4.70036983e-01 -3.76147628e-01 -7.68932343e-01 9.11696032e-02
-1.87634125e-01 2.41668269e-01 9.11198676e-01 4.04905714e-02
1.67959243e-01 -6.63130403e-01 -1.25672472e+00 5.09063482e-01
1.21085763e+00 -1.53987601e-01 2.78470874e-01 5.80995739e-01
5.41231036e-01 -4.05712932e-01 -1.02594340e+00 4.92359877e-01
3.65597755e-01 -7.76405334e-01 5.42796791e-01 -5.80516994e-01
2.29039982e-01 -1.02200627e-01 1.22763917e-01 -1.00684810e+00
-3.17022264e-01 -5.42971909e-01 -1.71160266e-01 1.15510809e+00
5.92181504e-01 -6.91554427e-01 1.00397038e+00 7.77164102e-01
-2.90425986e-01 -1.40335536e+00 -1.08204663e+00 -5.36598742e-01
-1.35753468e-01 -4.97345746e-01 5.02857983e-01 1.08094156e+00
-6.14600666e-02 8.02396014e-02 -2.28997856e-01 -2.74612039e-01
9.13236916e-01 2.44258031e-01 3.48908842e-01 -1.07670772e+00
-3.73016208e-01 -3.67422253e-01 -2.70814747e-01 -6.80513442e-01
-1.96178600e-01 -1.04205847e+00 -1.48578867e-01 -1.33818769e+00
4.49091345e-01 -1.22393370e+00 -6.96569741e-01 7.38295794e-01
-2.47729033e-01 3.44779156e-02 -9.44109634e-02 6.32485271e-01
-8.38853180e-01 5.34750521e-01 1.26850939e+00 -1.69282451e-01
-1.33872358e-02 -3.97855379e-02 -8.57606590e-01 3.21537495e-01
1.18292117e+00 -5.88323414e-01 -7.09729016e-01 -6.26808405e-01
4.24587578e-01 -9.53912269e-03 2.60377288e-01 -1.12260127e+00
4.59034294e-01 -1.66774988e-01 5.53209364e-01 -3.51473629e-01
7.58304223e-02 -7.43080556e-01 3.22217554e-01 4.50407445e-01
-5.47629893e-01 4.84060347e-02 -3.40798087e-02 4.15990531e-01
-1.23514190e-01 -2.18899265e-01 8.88604820e-01 -1.73345000e-01
-5.86740434e-01 9.07495201e-01 1.81995019e-01 1.77634522e-01
1.01990092e+00 -1.40712142e-01 -3.43194515e-01 9.24238637e-02
-8.22171211e-01 4.76001680e-01 3.00750256e-01 1.18640870e-01
7.79764473e-01 -1.38086390e+00 -4.53979433e-01 1.62658971e-02
5.93783706e-02 4.07579213e-01 7.40272045e-01 1.29911935e+00
-3.94005060e-01 4.47759628e-02 -3.75397325e-01 -6.70850456e-01
-1.09941590e+00 4.13352638e-01 3.44273120e-01 -1.80856735e-01
-3.88741732e-01 8.50825727e-01 9.43176262e-03 -2.90184021e-01
4.66947287e-01 -1.08983450e-01 -6.11695983e-02 1.83291495e-01
4.67965156e-01 4.61120337e-01 5.26259899e-01 -2.80199964e-02
-3.28397095e-01 5.11896573e-02 -2.11031973e-01 3.47229064e-01
1.43903482e+00 -3.64440419e-02 -4.22989912e-02 3.82883012e-01
1.31909192e+00 -4.59826171e-01 -1.29707158e+00 -2.44430512e-01
-7.23648891e-02 -4.07994300e-01 1.93657026e-01 -7.34164834e-01
-1.57011545e+00 8.24964941e-01 1.08475411e+00 -4.82184291e-02
1.64614081e+00 -1.78078026e-01 8.66067350e-01 4.11831796e-01
7.30221093e-01 -1.44214451e+00 3.79805975e-02 -7.39590824e-02
5.34966707e-01 -1.27354801e+00 1.01854488e-01 2.25220006e-02
-5.82357109e-01 1.00981939e+00 6.89801216e-01 8.69511664e-02
7.72449195e-01 4.83193636e-01 -5.89499474e-02 -2.52942711e-01
-9.98356819e-01 2.26711169e-01 -2.00188830e-01 1.80698588e-01
2.68618762e-01 2.11561859e-01 -4.86332804e-01 9.63640511e-01
-9.13194865e-02 2.22580552e-01 1.38958797e-01 1.04617405e+00
-2.49704286e-01 -8.76911938e-01 9.54389349e-02 1.12801886e+00
-4.56294388e-01 -2.86899149e-01 3.99412364e-02 2.20084339e-01
2.99283087e-01 8.07940364e-01 6.71993494e-02 -5.54512203e-01
3.01446021e-02 -3.67306083e-01 2.77745754e-01 -4.36168343e-01
-6.80486679e-01 2.94445716e-02 -2.04726696e-01 -4.80335891e-01
-4.95619863e-01 -5.49704552e-01 -1.52403307e+00 -2.68694669e-01
-6.60722673e-01 1.00911908e-01 7.25042284e-01 5.76009154e-01
6.33783400e-01 4.51769084e-01 9.72337127e-01 -5.53821027e-01
-8.09230387e-01 -7.91148543e-01 -7.44980097e-01 5.47094166e-01
-1.42700344e-01 -7.53815293e-01 -3.53674084e-01 -1.16339587e-01] | [6.066474914550781, 6.427331924438477] |
dfa61274-8400-4c67-8bc0-ed5c2686130d | learning-without-human-scores-for-blind-image | null | null | http://openaccess.thecvf.com/content_cvpr_2013/html/Xue_Learning_without_Human_2013_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2013/papers/Xue_Learning_without_Human_2013_CVPR_paper.pdf | Learning without Human Scores for Blind Image Quality Assessment | General purpose blind image quality assessment (BIQA) has been recently attracting significant attention in the fields of image processing, vision and machine learning. Stateof-the-art BIQA methods usually learn to evaluate the image quality by regression from human subjective scores of the training samples. However, these methods need a large number of human scored images for training, and lack an explicit explanation of how the image quality is affected by image local features. An interesting question is then: can we learn for effective BIQA without using human scored images? This paper makes a good effort to answer this question. We partition the distorted images into overlapped patches, and use a percentile pooling strategy to estimate the local quality of each patch. Then a quality-aware clustering (QAC) method is proposed to learn a set of centroids on each quality level. These centroids are then used as a codebook to infer the quality of each patch in a given image, and subsequently a perceptual quality score of the whole image can be obtained. The proposed QAC based BIQA method is simple yet effective. It not only has comparable accuracy to those methods using human scored images in learning, but also has merits such as high linearity to human perception of image quality, real-time implementation and availability of image local quality map. | ['Wufeng Xue', 'Xuanqin Mou', 'Lei Zhang'] | 2013-06-01 | null | null | null | cvpr-2013-6 | ['blind-image-quality-assessment'] | ['computer-vision'] | [ 1.40769362e-01 -5.26393652e-01 1.84574112e-01 -5.63974261e-01
-1.27957428e+00 -3.96620601e-01 8.10580030e-02 1.63011178e-01
-4.14434224e-01 5.62483370e-01 1.35181472e-01 3.01937182e-02
-5.39622188e-01 -6.45883918e-01 -3.94111961e-01 -9.45698798e-01
-3.57718728e-02 1.17973842e-01 4.84224647e-01 1.30345421e-02
5.77472806e-01 4.10499454e-01 -1.61499596e+00 3.46306115e-01
1.21528172e+00 1.26283181e+00 2.50690132e-01 7.82687843e-01
1.42069116e-01 7.35708952e-01 -1.10098505e+00 -2.61384338e-01
3.07989120e-01 -6.56563759e-01 -7.32793629e-01 4.35036689e-01
4.51379299e-01 -2.36782506e-01 6.22765981e-02 1.38521159e+00
7.42221177e-01 -6.95070438e-03 8.86916220e-01 -1.01760173e+00
-6.34607255e-01 -6.26239926e-02 -4.97567654e-01 2.99950689e-01
3.60487640e-01 2.69248337e-01 9.18601573e-01 -7.80273139e-01
3.30725722e-02 1.01288724e+00 3.62297773e-01 2.58485526e-01
-1.00517464e+00 -5.26006401e-01 -3.49132270e-01 8.78592730e-01
-1.46570623e+00 -2.92183846e-01 6.93188727e-01 -4.55008715e-01
5.26423275e-01 2.33730331e-01 7.10924327e-01 1.12351380e-01
2.64381707e-01 6.87175870e-01 1.60586548e+00 -6.16895556e-01
5.66115141e-01 2.42941529e-01 -2.15595275e-01 8.31787467e-01
-7.98764601e-02 5.11958003e-02 -4.17062998e-01 4.40217890e-02
6.85606301e-01 -1.09900177e-01 -5.17075479e-01 -4.18345213e-01
-1.05027616e+00 7.90313959e-01 9.10070240e-01 3.55379373e-01
-5.21424115e-01 -1.49626300e-01 4.21848893e-02 3.85044277e-01
1.90247998e-01 4.44741517e-01 -2.21023649e-01 1.49564594e-01
-1.20077038e+00 5.61800264e-02 3.79252791e-01 3.01991284e-01
9.86710787e-01 -2.90764123e-01 -2.53412545e-01 1.01751149e+00
3.96284610e-01 6.66580856e-01 5.54785848e-01 -1.24286366e+00
3.30236673e-01 6.30029500e-01 2.17691392e-01 -1.20022035e+00
-1.25577688e-01 -7.87047371e-02 -7.53075778e-01 1.03948975e+00
4.04829532e-01 1.94773495e-01 -9.86413002e-01 1.08244371e+00
6.06805496e-02 -6.83422908e-02 -1.02740325e-01 1.18080068e+00
6.04924858e-01 8.21982443e-01 -3.02918762e-01 -4.72683847e-01
1.19325066e+00 -8.02812755e-01 -6.80711925e-01 3.89497168e-02
-1.28217429e-01 -9.24253702e-01 1.09778738e+00 8.74736845e-01
-1.20826399e+00 -8.94309044e-01 -1.22184300e+00 2.06740692e-01
-3.09907734e-01 3.04787278e-01 3.22964132e-01 9.59368587e-01
-1.40364957e+00 5.28118789e-01 -4.80399758e-01 -1.57013997e-01
4.83150750e-01 5.31538486e-01 -2.44483978e-01 -3.56226623e-01
-8.59123647e-01 9.97474015e-01 1.61279112e-01 2.11051963e-02
-9.56645131e-01 -2.32563391e-01 -6.79123998e-01 -9.57175344e-02
1.46950006e-01 -4.77465719e-01 1.05059302e+00 -1.36488724e+00
-1.48916221e+00 7.46482670e-01 -4.27148819e-01 -2.19729364e-01
1.03845090e-01 7.18682334e-02 -4.99992996e-01 6.66267335e-01
1.11170679e-01 6.58225775e-01 1.18460095e+00 -1.49033213e+00
-8.50110590e-01 -5.34980059e-01 -1.35309950e-01 3.30161899e-01
-3.65291089e-02 2.76166350e-01 -6.59277797e-01 -3.80505919e-01
2.76329666e-01 -4.06305403e-01 -3.24600525e-02 2.97019988e-01
2.00354889e-01 -3.14774215e-01 3.37849945e-01 -9.37787354e-01
1.22624040e+00 -2.02895617e+00 5.17831370e-02 4.32218671e-01
1.20666973e-01 5.38893044e-01 -1.75850943e-01 6.00458868e-03
2.70526379e-01 -1.22907765e-01 -4.03675914e-01 -1.91717744e-02
-2.47526869e-01 1.89196831e-03 1.98723912e-01 4.96043205e-01
3.42418015e-01 6.35729134e-01 -7.00281262e-01 -8.51319194e-01
4.81206089e-01 4.07911420e-01 -4.69498158e-01 4.73860353e-01
2.73063213e-01 2.35655978e-01 -1.43288866e-01 7.05701709e-01
8.36099982e-01 -9.31546614e-02 -2.16214061e-01 -4.71174151e-01
6.67412505e-02 -1.98328689e-01 -1.41880012e+00 1.29586041e+00
-2.92255223e-01 6.87260270e-01 1.61590297e-02 -1.01094532e+00
1.03107178e+00 3.85683924e-01 3.59960854e-01 -1.01353800e+00
8.09894577e-02 2.12156966e-01 4.39040177e-02 -7.68629491e-01
1.02511607e-01 -3.13869864e-01 2.71746427e-01 3.08785766e-01
2.83994794e-01 -5.45136094e-01 1.05412014e-01 -1.00825615e-01
8.04447472e-01 -3.12042028e-01 5.13188124e-01 8.22984725e-02
8.92032921e-01 -2.15268716e-01 5.15291870e-01 4.27059263e-01
-6.93727851e-01 8.75696480e-01 1.40963942e-01 -2.36489996e-01
-1.09587729e+00 -1.32169259e+00 -6.50821626e-02 6.48973763e-01
5.91886461e-01 -6.46812469e-03 -1.00485098e+00 -5.81874788e-01
-3.57533395e-01 1.24605848e-02 -4.13025290e-01 -7.00833928e-03
-2.68997550e-01 -7.67055929e-01 -7.18036201e-03 1.79217950e-01
7.36910760e-01 -1.37622869e+00 -4.16473448e-01 5.82189597e-02
-3.21824074e-01 -6.54990435e-01 -3.40185493e-01 -1.65585622e-01
-8.41140449e-01 -1.16180873e+00 -8.90363991e-01 -1.01001394e+00
7.92558789e-01 4.61447358e-01 1.02776313e+00 2.65938789e-01
-3.90924335e-01 2.49316067e-01 -4.15991455e-01 -2.25992411e-01
-2.00311929e-01 -6.51778996e-01 -7.25468993e-02 3.90950620e-01
3.13385427e-01 -2.13765070e-01 -9.88808393e-01 6.51111901e-01
-8.24413657e-01 -5.03977537e-01 9.12432015e-01 7.88014710e-01
8.13430309e-01 8.60461175e-01 5.22435844e-01 -2.96814322e-01
7.35594392e-01 9.58631784e-02 -5.98243177e-01 4.43197608e-01
-6.42126381e-01 -1.60919145e-01 4.89380300e-01 -1.73609003e-01
-1.09112251e+00 -7.62478486e-02 -1.88386682e-02 -4.87101823e-02
-2.36926541e-01 2.78949231e-01 -4.34495360e-01 -4.02865291e-01
7.59930491e-01 1.87347963e-01 8.30711648e-02 -2.74664044e-01
2.51944005e-01 9.73012805e-01 7.05882430e-01 -1.97736546e-01
6.86473250e-01 3.97681057e-01 -2.20004156e-01 -6.60158396e-01
-5.87251604e-01 -8.54392767e-01 -5.57352483e-01 -5.07306933e-01
1.07035148e+00 -8.04988801e-01 -7.47959077e-01 6.92118287e-01
-8.87561202e-01 -8.44077170e-02 -8.55131261e-03 5.86381972e-01
-6.49915040e-01 6.43856585e-01 -4.16185945e-01 -9.74299014e-01
-3.90017062e-01 -1.43014336e+00 8.55653882e-01 5.19834399e-01
2.63557225e-01 -8.60206246e-01 -1.92093313e-01 7.80555010e-01
3.42203945e-01 -1.99110419e-01 8.87205541e-01 1.81739107e-01
-6.30244434e-01 -3.41182530e-01 -6.28634810e-01 9.89138246e-01
3.89580160e-01 -2.24058524e-01 -9.35941041e-01 -3.77902359e-01
2.13212863e-01 -3.34831029e-01 5.80099165e-01 7.57977784e-01
1.10236406e+00 -2.44461179e-01 1.60977483e-01 3.81700069e-01
1.78911769e+00 3.86198908e-01 8.71930778e-01 2.35497966e-01
2.38123462e-01 5.49825907e-01 8.16874266e-01 5.24787754e-02
1.57341123e-01 8.14156830e-01 4.65007782e-01 -3.89548659e-01
-2.83301651e-01 2.02709973e-01 3.54230672e-01 9.03040349e-01
-8.86652172e-02 -4.47451929e-03 -7.09245503e-01 6.39089704e-01
-1.52352428e+00 -9.51945424e-01 -2.16535613e-04 2.37442470e+00
8.70123267e-01 1.16481811e-01 2.35740930e-01 6.84333444e-01
7.10935712e-01 -2.79303759e-01 -4.51686651e-01 -1.83653653e-01
-8.14618319e-02 3.18160087e-01 2.94811040e-01 6.13657176e-01
-1.02595103e+00 5.56026101e-01 6.67314863e+00 9.61665452e-01
-1.04463041e+00 2.22089551e-02 7.36274540e-01 1.59965575e-01
1.51282221e-01 -1.31764650e-01 -2.15127289e-01 5.93274653e-01
6.67608798e-01 1.14683472e-01 6.58424437e-01 3.33775938e-01
4.88781422e-01 -6.15476727e-01 -7.04340816e-01 1.31209254e+00
5.03350079e-01 -7.51256168e-01 1.63110182e-01 5.35001121e-02
8.73296201e-01 -3.70718658e-01 2.55919993e-01 -1.90386698e-01
6.02099411e-02 -1.38256049e+00 4.43037242e-01 9.05225396e-01
6.07617140e-01 -9.11285400e-01 1.18139362e+00 2.27922469e-01
-9.37932968e-01 -3.60879600e-01 -7.00025022e-01 2.27472577e-02
-5.13998307e-02 6.04138255e-01 -5.99143386e-01 4.14823532e-01
1.01140451e+00 3.07978272e-01 -1.11158729e+00 1.77483571e+00
-2.59750724e-01 6.56689942e-01 9.93534997e-02 7.08931088e-02
4.21633832e-02 -1.97844133e-01 1.65578365e-01 7.53640294e-01
3.86596859e-01 1.98414713e-01 -6.66675195e-02 6.02133155e-01
3.41758102e-01 2.92343140e-01 -1.38018141e-02 4.72531617e-01
2.77937740e-01 1.20121908e+00 -6.70306444e-01 -4.21470255e-01
-4.75878775e-01 1.11836874e+00 1.97736900e-02 3.45545292e-01
-2.86814779e-01 -5.75384557e-01 2.99595445e-01 -4.32197452e-02
3.55775833e-01 -2.73402203e-02 -3.71880025e-01 -7.50151873e-01
1.20373689e-01 -1.23383081e+00 3.28023314e-01 -1.21849287e+00
-1.37444317e+00 6.05876088e-01 -3.69252324e-01 -1.63402951e+00
-1.03964813e-01 -8.59411299e-01 -6.40975893e-01 1.16771424e+00
-1.52491343e+00 -7.76055634e-01 -5.59880674e-01 1.00541890e+00
4.74843442e-01 -4.74194318e-01 6.63566530e-01 1.85099751e-01
-3.24348062e-02 5.36112547e-01 1.06632702e-01 1.22480325e-01
8.79280448e-01 -1.48389637e+00 -3.50730509e-01 8.83938134e-01
1.80291310e-01 3.18295330e-01 6.40754700e-01 -1.21094726e-01
-8.09501290e-01 -7.44605005e-01 6.39911294e-01 -4.67853963e-01
1.98939651e-01 2.54660189e-01 -9.33052540e-01 -3.93501371e-01
2.59354889e-01 1.19578436e-01 5.82251072e-01 -2.82761633e-01
-1.69450223e-01 -6.36525095e-01 -1.48567462e+00 1.72728419e-01
3.27596813e-01 -6.61114395e-01 -7.52321243e-01 5.37075065e-02
3.18026632e-01 1.76746786e-01 -7.84581006e-01 2.87628472e-01
3.27329963e-01 -1.39121842e+00 9.61876571e-01 1.78436130e-01
2.20802292e-01 -9.02195930e-01 -1.41985834e-01 -1.38236690e+00
-3.60830903e-01 7.19532277e-03 3.81432235e-01 1.05538094e+00
3.12179506e-01 -1.77463964e-01 6.17296517e-01 1.01363465e-01
1.73558563e-01 -6.57712698e-01 -8.71564448e-01 -5.77892363e-01
-1.28749654e-01 -2.19080150e-01 4.59598780e-01 4.78262007e-01
-5.47936484e-02 1.95211068e-01 -2.73707598e-01 4.73682374e-01
9.14615691e-01 9.23328102e-03 4.46419984e-01 -1.09036899e+00
-3.84750068e-01 -5.06143034e-01 -9.53059793e-01 -6.59543157e-01
-4.70475703e-01 -5.00990808e-01 3.43094319e-01 -1.99000859e+00
3.31564397e-01 -2.38270313e-01 -4.92891580e-01 1.45696729e-01
-5.20856142e-01 8.17547262e-01 8.95852223e-02 3.47318381e-01
-9.45356846e-01 2.79594809e-01 1.43927085e+00 -3.46171141e-01
-7.44637027e-02 2.96757221e-01 -4.32721525e-01 4.49878782e-01
7.57134140e-01 -1.91434413e-01 -3.21853280e-01 -2.37494886e-01
1.40131831e-01 1.60886586e-01 4.57302898e-01 -1.49662971e+00
3.15212965e-01 -3.52604641e-03 6.80363297e-01 -4.07312691e-01
2.89706945e-01 -9.42006588e-01 -3.59302700e-01 2.90083200e-01
-7.43869320e-02 -1.39595866e-01 -2.20977902e-01 3.59190613e-01
-8.26532960e-01 -4.61990595e-01 1.22399259e+00 -2.64423043e-01
-8.30936849e-01 1.78894579e-01 -2.49599382e-01 -3.06761563e-01
8.58420253e-01 -5.09387612e-01 -1.40080322e-02 -6.47952020e-01
-5.30088663e-01 8.62493366e-03 4.07640368e-01 9.32227448e-02
1.01205516e+00 -1.36009896e+00 -7.16491699e-01 2.84348160e-01
2.19366297e-01 -2.24135235e-01 4.03991312e-01 5.56864142e-01
-6.49635732e-01 1.23877510e-01 -4.65593368e-01 -7.28838861e-01
-1.48394287e+00 5.92153311e-01 4.80271459e-01 1.59638114e-02
-1.05570026e-01 7.95431495e-01 -3.00678220e-02 -1.41232088e-01
3.23330551e-01 -6.71968907e-02 -4.99177217e-01 -7.50354975e-02
8.58299851e-01 3.95977199e-01 2.92114019e-01 -8.29629779e-01
-1.89068764e-01 1.08110654e+00 2.97343552e-01 -3.22138429e-01
1.03263211e+00 -3.70077252e-01 -3.15540284e-01 2.87915170e-01
1.25311768e+00 -9.12403241e-02 -1.17865574e+00 -3.14430386e-01
-6.66446015e-02 -8.05837631e-01 2.87457436e-01 -1.31795311e+00
-1.12209272e+00 1.18224585e+00 1.43211949e+00 2.96571106e-01
1.89774644e+00 1.07063301e-01 4.80538160e-01 -5.55713400e-02
4.27479684e-01 -1.19784498e+00 5.98832250e-01 -3.96263264e-02
9.16121364e-01 -1.44787872e+00 7.97710195e-02 -1.40182704e-01
-7.90796757e-01 9.24072862e-01 3.15596879e-01 -2.09983870e-01
6.04562342e-01 -3.44128698e-01 6.27135277e-01 -1.52495697e-01
-2.56621659e-01 -3.29098463e-01 7.88087904e-01 1.05627584e+00
1.06858365e-01 2.05964550e-01 -1.93169922e-01 2.44635820e-01
-2.01430589e-01 -1.00508323e-02 3.16143304e-01 4.18447971e-01
-9.65855062e-01 -9.32267368e-01 -8.07124436e-01 3.89840275e-01
-4.95932847e-01 3.24172825e-02 -8.14699233e-02 1.10112853e-01
6.28358483e-01 1.46141708e+00 -7.92944357e-02 -3.26612651e-01
2.92014420e-01 -2.22935930e-01 6.16516352e-01 -4.58556294e-01
-3.01571012e-01 1.04267739e-01 -5.81293941e-01 -5.24858296e-01
-7.14720964e-01 -5.10859668e-01 -9.32248294e-01 4.41726111e-03
-2.71913856e-01 3.40390474e-01 7.96096087e-01 8.46834660e-01
-2.28047282e-01 2.59465873e-01 9.46075678e-01 -7.82072425e-01
-2.95643181e-01 -9.90811825e-01 -8.05421829e-01 5.25874436e-01
6.30227625e-01 -4.79979873e-01 -5.05641818e-01 2.23780841e-01] | [11.79484748840332, -1.9105793237686157] |
df972d75-25af-4eb8-a74f-df1bbc3d0a9b | an-application-of-cascaded-3d-fully | 1803.05431 | null | http://arxiv.org/abs/1803.05431v2 | http://arxiv.org/pdf/1803.05431v2.pdf | An application of cascaded 3D fully convolutional networks for medical image segmentation | Recent advances in 3D fully convolutional networks (FCN) have made it
feasible to produce dense voxel-wise predictions of volumetric images. In this
work, we show that a multi-class 3D FCN trained on manually labeled CT scans of
several anatomical structures (ranging from the large organs to thin vessels)
can achieve competitive segmentation results, while avoiding the need for
handcrafting features or training class-specific models.
To this end, we propose a two-stage, coarse-to-fine approach that will first
use a 3D FCN to roughly define a candidate region, which will then be used as
input to a second 3D FCN. This reduces the number of voxels the second FCN has
to classify to ~10% and allows it to focus on more detailed segmentation of the
organs and vessels.
We utilize training and validation sets consisting of 331 clinical CT images
and test our models on a completely unseen data collection acquired at a
different hospital that includes 150 CT scans, targeting three anatomical
organs (liver, spleen, and pancreas). In challenging organs such as the
pancreas, our cascaded approach improves the mean Dice score from 68.5 to
82.2%, achieving the highest reported average score on this dataset. We compare
with a 2D FCN method on a separate dataset of 240 CT scans with 18 classes and
achieve a significantly higher performance in small organs and vessels.
Furthermore, we explore fine-tuning our models to different datasets.
Our experiments illustrate the promise and robustness of current 3D FCN based
semantic segmentation of medical images, achieving state-of-the-art results.
Our code and trained models are available for download:
https://github.com/holgerroth/3Dunet_abdomen_cascade. | ['Kensaku MORI', 'Kazunari Misawa', 'Yuichiro Hayashi', 'Hirohisa ODA', 'Michitaka Fujiwara', 'Holger R. Roth', 'Ying Yang', 'Xiangrong Zhou', 'Masahiro Oda', 'Natsuki Shimizu'] | 2018-03-14 | null | null | null | null | ['3d-medical-imaging-segmentation'] | ['medical'] | [-1.46118356e-02 2.55908459e-01 -2.04649180e-01 -4.93112415e-01
-9.23268259e-01 -6.56094968e-01 2.87041128e-01 4.93078589e-01
-3.83136332e-01 4.45823610e-01 1.89685836e-01 -5.20245671e-01
1.97856888e-01 -6.98651731e-01 -5.98915219e-01 -5.69493473e-01
-3.76169860e-01 9.19833362e-01 4.47493494e-01 2.64351964e-01
-2.25419700e-01 7.96371222e-01 -5.47529876e-01 4.04848367e-01
6.50745511e-01 1.07860494e+00 1.48115054e-01 7.35707581e-01
-1.40627638e-01 5.02507567e-01 -2.62694657e-01 -1.66916475e-01
4.61781621e-01 -4.65969414e-01 -9.42479908e-01 7.76323602e-02
4.26833510e-01 -4.27277178e-01 -7.96462595e-02 6.92440093e-01
5.84055245e-01 -2.04686120e-01 7.20627069e-01 -7.47038066e-01
-3.28866661e-01 7.07710862e-01 -5.22766948e-01 4.72852826e-01
-1.25655338e-01 1.95182607e-01 5.87550581e-01 -6.38128161e-01
5.65241158e-01 9.59377944e-01 9.77251291e-01 6.73620045e-01
-1.24953210e+00 -5.92113853e-01 -2.15688884e-01 -4.51250762e-01
-1.33063436e+00 7.25679398e-02 2.77942717e-01 -5.67656517e-01
7.41124034e-01 1.07715726e-01 8.84007454e-01 6.59068763e-01
3.56438458e-01 6.81303918e-01 1.05719984e+00 -1.16843231e-01
1.65925458e-01 -1.87743127e-01 1.05362320e-02 9.97614384e-01
3.60970169e-01 -7.15072453e-02 2.60537684e-01 -2.67423540e-01
1.18874753e+00 9.69932973e-02 -3.19009155e-01 -6.51062965e-01
-1.30605912e+00 1.05447209e+00 9.80887473e-01 4.12519276e-01
-5.40798843e-01 1.25868246e-01 4.33550298e-01 -2.30886266e-01
4.43347126e-01 2.98939824e-01 -5.27959943e-01 3.73296529e-01
-9.54107761e-01 -5.63641824e-02 8.42242122e-01 6.00991189e-01
2.95720845e-01 -3.87770385e-01 -2.99873441e-01 6.49949610e-01
2.86091954e-01 2.02510893e-01 5.48091173e-01 -8.63785923e-01
2.07831427e-01 5.87704837e-01 -2.66407669e-01 -3.46258849e-01
-1.01544261e+00 -7.31572330e-01 -7.83696055e-01 2.47405797e-01
7.54410744e-01 -1.97751150e-01 -1.52423406e+00 1.37697744e+00
5.73938906e-01 5.68346642e-02 -3.31247002e-01 1.20113027e+00
1.09689999e+00 1.50784522e-01 3.23940814e-01 1.22634865e-01
1.35751879e+00 -1.01679695e+00 -3.35341990e-02 1.59055114e-01
8.24059963e-01 -5.54989934e-01 8.56513977e-01 1.54967397e-01
-1.20638108e+00 -2.26600707e-01 -7.59575605e-01 1.11445077e-01
-5.78394495e-02 -6.70040920e-02 7.98927605e-01 6.98159516e-01
-1.19140232e+00 6.97856963e-01 -1.41249561e+00 -4.50004339e-01
1.01509988e+00 5.67369342e-01 -2.42047787e-01 -1.78975314e-01
-8.26982856e-01 1.11052001e+00 4.07897264e-01 -1.41303033e-01
-1.29576945e+00 -1.20526755e+00 -6.97775781e-01 4.43389975e-02
2.60527104e-01 -9.69136178e-01 1.26451159e+00 -7.42852926e-01
-1.23182654e+00 1.14690089e+00 2.61315316e-01 -6.64632201e-01
8.90789926e-01 2.18577072e-01 6.81435615e-02 6.33545518e-01
9.62403938e-02 8.80377531e-01 2.42361546e-01 -1.11014485e+00
-3.23507756e-01 -2.73159176e-01 -1.12665139e-01 1.12704329e-01
2.75862545e-01 -1.01203665e-01 -5.25379062e-01 -5.45772970e-01
2.30573714e-01 -1.03366840e+00 -5.88113248e-01 4.15641248e-01
-3.31147850e-01 1.41119957e-01 3.57702583e-01 -6.33451223e-01
5.70426047e-01 -1.71874440e+00 -3.49957012e-02 3.99980217e-01
5.26538670e-01 5.18299825e-02 8.09982494e-02 -3.18560392e-01
-1.00677364e-01 2.47328818e-01 -4.98827279e-01 -2.17887670e-01
-3.73231560e-01 1.76427677e-01 4.57287192e-01 6.36197627e-01
1.21239431e-01 1.10952830e+00 -8.00846457e-01 -7.60383964e-01
4.98849541e-01 5.90235770e-01 -6.48801148e-01 -9.33046173e-03
5.78220859e-02 8.98825049e-01 -5.14361501e-01 7.93642998e-01
6.64351285e-01 -7.21908212e-01 2.12462574e-01 -2.85972863e-01
1.06308676e-01 1.95014372e-01 -5.99057317e-01 1.96357810e+00
-4.97485787e-01 9.92225632e-02 1.85767680e-01 -9.45852458e-01
6.87809348e-01 3.11959475e-01 1.07625127e+00 -4.27466571e-01
4.23568249e-01 3.72884691e-01 2.81085551e-01 -1.71759635e-01
-3.99989873e-01 -5.99515200e-01 -7.45929182e-02 3.20915788e-01
3.85275275e-01 -5.34998238e-01 1.23736434e-01 8.49468634e-02
1.10804093e+00 -4.70544882e-02 2.14956135e-01 -7.28995919e-01
4.20251906e-01 3.86294991e-01 3.09618205e-01 6.59465969e-01
-5.13762653e-01 9.31996405e-01 4.90779221e-01 -6.23358965e-01
-1.02451861e+00 -1.32888198e+00 -5.81300855e-01 6.53604507e-01
2.55089849e-02 5.76035865e-02 -8.13734055e-01 -1.02723694e+00
1.84993610e-01 4.46720988e-01 -8.87733459e-01 1.49067283e-01
-7.30380416e-01 -1.01552963e+00 6.37909472e-01 6.12086713e-01
2.78022707e-01 -7.52254367e-01 -7.71369636e-01 3.40911329e-01
-1.42049849e-01 -1.10874343e+00 -3.55092138e-01 4.59196121e-01
-1.23276842e+00 -1.31396246e+00 -1.02060914e+00 -7.98236191e-01
8.59352887e-01 -1.36652321e-01 1.31521976e+00 2.48693526e-01
-5.47519803e-01 2.17372924e-01 -1.74391195e-01 -3.21887434e-01
-5.63042104e-01 1.46372959e-01 -3.60251248e-01 -5.98920643e-01
2.49441545e-02 -3.18388522e-01 -9.22821999e-01 3.37952733e-01
-7.01521873e-01 1.19227163e-01 6.58218086e-01 8.52876782e-01
9.07877386e-01 -3.68883252e-01 2.86223739e-01 -9.13073778e-01
1.58944786e-01 -5.91203809e-01 -4.71056342e-01 1.87781334e-01
-3.37548822e-01 -4.57251407e-02 5.29515326e-01 -3.61618757e-01
-7.81823397e-01 3.52190733e-01 -3.01614434e-01 -4.60282236e-01
-2.77369499e-01 3.81679595e-01 5.08188665e-01 -3.79249513e-01
6.74793482e-01 -4.48708721e-02 2.15287536e-01 -4.09110010e-01
1.56677231e-01 5.88788427e-02 3.61672670e-01 -4.64119285e-01
5.19846141e-01 6.42267764e-01 3.21518868e-01 -2.39277452e-01
-8.22816133e-01 -3.23278069e-01 -9.59309757e-01 -1.18438825e-01
1.10375226e+00 -8.51904750e-01 -4.27583158e-01 3.02607846e-02
-6.81496203e-01 -7.56085098e-01 -4.19357002e-01 7.71454990e-01
-4.77322221e-01 1.33931875e-01 -1.07377398e+00 1.38831893e-02
-6.84572339e-01 -1.41937971e+00 9.90575075e-01 1.11849859e-01
-1.99414268e-01 -1.23846877e+00 -6.30922467e-02 1.95535481e-01
5.98665059e-01 6.80145025e-01 9.73637760e-01 -8.95374596e-01
-4.22354788e-01 -1.82265565e-02 -3.15366030e-01 1.83971822e-01
2.06766278e-01 -2.38706201e-01 -5.81166089e-01 -2.19202071e-01
-1.17623366e-01 -3.83181691e-01 9.89688396e-01 9.96604204e-01
1.33091581e+00 2.96650022e-01 -4.90783423e-01 8.40047419e-01
1.40550351e+00 6.13188893e-02 9.39101428e-02 7.06963167e-02
6.74072921e-01 2.26782203e-01 2.90185452e-01 3.06133002e-01
3.19159716e-01 2.85320640e-01 6.43326640e-01 -5.72297871e-01
-4.93348867e-01 -6.31810203e-02 -3.55708569e-01 4.95530456e-01
-8.49899277e-02 -5.10678403e-02 -1.17923307e+00 5.77081561e-01
-1.27650213e+00 -2.95792013e-01 -1.18894458e-01 1.98306978e+00
8.00386846e-01 2.01697007e-01 2.09817335e-01 -4.15302366e-01
7.43847013e-01 -2.01085955e-01 -6.07394040e-01 -1.70419186e-01
2.56193727e-01 6.69821382e-01 7.32385159e-01 5.45782328e-01
-1.43792129e+00 6.90609455e-01 6.28214216e+00 5.32101035e-01
-1.40772235e+00 3.04709017e-01 1.11280906e+00 -2.36098245e-01
-1.37195319e-01 -2.41690382e-01 -5.23296297e-01 2.49745697e-01
8.25541139e-01 8.17938745e-02 1.21902063e-01 6.26499414e-01
1.58452600e-01 -2.07305908e-01 -1.07503057e+00 5.97863793e-01
-1.51929989e-01 -1.42818594e+00 -4.09495309e-02 2.92515028e-02
6.74493670e-01 6.24738753e-01 -3.22233707e-01 1.85370147e-01
3.70576918e-01 -1.35916483e+00 5.21803856e-01 7.93645754e-02
9.00028169e-01 -3.74220103e-01 8.17726076e-01 1.45893559e-01
-1.07930315e+00 2.88493454e-01 -2.76149750e-01 4.42637831e-01
2.36543745e-01 6.22942209e-01 -1.27884078e+00 4.03780282e-01
7.53702581e-01 4.18882310e-01 -4.41674948e-01 1.38652849e+00
-4.16745432e-02 5.83459795e-01 -6.36822701e-01 1.31310612e-01
4.39029187e-01 9.86771509e-02 1.67009607e-01 1.29485738e+00
2.95418531e-01 4.10729617e-01 2.88793594e-01 9.40553963e-01
-2.95195162e-01 2.43193984e-01 -8.16855058e-02 4.44822252e-01
-6.88660052e-03 1.54240978e+00 -1.36527395e+00 -5.31444073e-01
-3.65312248e-01 7.34633446e-01 2.07551837e-01 -1.64779443e-02
-1.10220885e+00 1.97024435e-01 2.61530280e-01 1.94887504e-01
3.85300845e-01 -4.34100581e-03 -4.47296441e-01 -1.17580307e+00
-3.81861269e-01 -3.18122238e-01 7.70607412e-01 -4.23966020e-01
-1.25155663e+00 6.43565238e-01 -4.79830056e-02 -1.00944078e+00
1.32260486e-01 -6.11386597e-01 -5.08326173e-01 9.02276397e-01
-1.50590193e+00 -1.05719364e+00 -5.84521651e-01 5.29444516e-01
3.38176131e-01 4.52608615e-01 8.13950241e-01 4.12777364e-01
-1.54416561e-01 3.41902882e-01 -5.64281493e-02 4.38210756e-01
5.24795651e-01 -1.39239132e+00 1.56144917e-01 4.29547280e-01
-2.40319550e-01 2.67372876e-01 1.69748157e-01 -6.33890390e-01
-1.00452209e+00 -1.20903873e+00 3.22165400e-01 -5.15404820e-01
4.83192772e-01 -1.94789171e-01 -6.69323266e-01 7.89959848e-01
-1.07092090e-01 7.38159239e-01 7.90985107e-01 -3.18791956e-01
-4.26961705e-02 1.89716652e-01 -1.71474588e+00 1.16785318e-01
9.25810993e-01 1.30328432e-01 -5.20906150e-01 4.40959275e-01
5.83425164e-01 -9.85203028e-01 -1.46976161e+00 6.74396634e-01
5.58545530e-01 -8.52709591e-01 1.13230777e+00 -3.94257993e-01
4.80078369e-01 -1.17426902e-01 1.06244013e-01 -1.29060841e+00
-2.90516555e-01 4.43174206e-02 1.40005529e-01 4.09053802e-01
5.05971432e-01 -5.34732997e-01 1.00380182e+00 6.80109560e-01
-5.14277875e-01 -1.06651783e+00 -9.23474491e-01 -4.32086140e-01
7.47285903e-01 -2.31967792e-01 3.77013952e-01 9.45170462e-01
-2.89384484e-01 -1.25377357e-01 3.75706673e-01 1.35970709e-03
7.34419525e-01 3.58264029e-01 2.22280011e-01 -1.16481614e+00
-1.67202592e-01 -6.38329864e-01 -2.27133512e-01 -8.43581796e-01
-2.46463105e-01 -1.39433670e+00 -2.19764516e-01 -1.76195061e+00
5.23758769e-01 -8.15927982e-01 -4.43991750e-01 7.06579208e-01
-3.91986668e-02 5.87303817e-01 1.99813500e-01 1.57639846e-01
-3.76478612e-01 1.81660488e-01 1.84578133e+00 -6.95504323e-02
-1.24838442e-01 7.79497325e-02 -5.45392871e-01 7.09604561e-01
8.98571670e-01 -4.77010995e-01 -1.12209976e-01 -2.75882840e-01
-6.83051169e-01 3.77288550e-01 5.21218061e-01 -8.83777082e-01
-1.50661841e-01 1.46982089e-01 1.12690175e+00 -5.53671420e-01
7.46770054e-02 -8.11612785e-01 4.11667563e-02 1.11339974e+00
-2.71751553e-01 -2.68741488e-01 4.16188538e-01 7.02117532e-02
1.65308418e-04 -2.83250138e-02 1.20377600e+00 -7.13416338e-01
-2.76537269e-01 6.53958499e-01 -2.40982860e-01 1.74741089e-01
1.15835404e+00 -1.02464352e-02 -6.36803806e-02 -1.49314433e-01
-1.13515925e+00 3.30437183e-01 4.11722034e-01 3.27043282e-03
3.89540046e-01 -1.00831425e+00 -7.90927231e-01 9.81582925e-02
-1.52774110e-01 5.15337646e-01 2.33591229e-01 1.20382440e+00
-1.13429368e+00 5.76324821e-01 -3.25266600e-01 -1.08120012e+00
-9.24641311e-01 2.74263978e-01 9.28995550e-01 -6.07636034e-01
-8.52296174e-01 8.21474612e-01 4.35050815e-01 -7.14273512e-01
-1.05126746e-01 -9.25473213e-01 7.53209144e-02 -3.40166062e-01
1.23944171e-01 1.05605787e-02 2.55280375e-01 -5.78068674e-01
-6.91662133e-01 5.08284390e-01 6.30981326e-02 3.05484712e-01
1.40674341e+00 9.86078605e-02 7.30635822e-02 -1.18966743e-01
1.38000560e+00 -3.13715309e-01 -1.39063203e+00 -9.32874829e-02
-1.34404585e-01 -2.29373455e-01 2.73303181e-01 -1.17824757e+00
-1.49998403e+00 7.92030632e-01 7.73857355e-01 -2.99816504e-02
1.02996361e+00 4.54332620e-01 8.64458621e-01 -2.96278596e-01
3.75659615e-01 -3.66359115e-01 -1.29068241e-01 2.38160923e-01
5.40749967e-01 -1.39961147e+00 7.81376362e-02 -5.28827906e-01
-6.02904975e-01 1.32271755e+00 6.07792497e-01 -3.30262125e-01
7.67316401e-01 4.36387390e-01 2.39044547e-01 -3.57891798e-01
-3.58094186e-01 -4.69924808e-02 3.91769439e-01 3.17378968e-01
5.67748666e-01 3.69500637e-01 -2.42847502e-01 4.45759296e-01
-1.20505663e-02 9.84830335e-02 4.19675857e-01 8.76799107e-01
-3.34164321e-01 -9.07077789e-01 -2.40463346e-01 6.97075486e-01
-8.79099667e-01 -1.62892893e-01 3.64359631e-03 1.09862733e+00
4.47240993e-02 2.95209378e-01 1.30266801e-01 1.41885161e-01
1.55246064e-01 -8.51849318e-02 7.61325479e-01 -7.05378473e-01
-9.24855590e-01 3.43701661e-01 -1.26953587e-01 -6.85180902e-01
-4.13157046e-01 -7.64781058e-01 -1.74614131e+00 -1.19712338e-01
-1.27585471e-01 -4.27372232e-02 7.10339010e-01 7.80391932e-01
7.11829737e-02 6.04749918e-01 3.64027023e-01 -9.94157970e-01
-3.98209780e-01 -8.52937639e-01 -3.69571894e-01 4.88566697e-01
3.07097048e-01 -6.30437315e-01 -2.37129658e-01 -2.06265524e-01] | [14.671149253845215, -2.4317781925201416] |
1aa4e169-5132-4334-bdf5-11ce8072e1cd | neural-machine-translation-for-code | 2305.13504 | null | https://arxiv.org/abs/2305.13504v1 | https://arxiv.org/pdf/2305.13504v1.pdf | Neural Machine Translation for Code Generation | Neural machine translation (NMT) methods developed for natural language processing have been shown to be highly successful in automating translation from one natural language to another. Recently, these NMT methods have been adapted to the generation of program code. In NMT for code generation, the task is to generate output source code that satisfies constraints expressed in the input. In the literature, a variety of different input scenarios have been explored, including generating code based on natural language description, lower-level representations such as binary or assembly (neural decompilation), partial representations of source code (code completion and repair), and source code in another language (code translation). In this paper we survey the NMT for code generation literature, cataloging the variety of methods that have been explored according to input and output representations, model architectures, optimization techniques used, data sets, and evaluation methods. We discuss the limitations of existing methods and future research directions | ['Clayton T. Morrison', 'Dharma KC'] | 2023-05-22 | null | null | null | null | ['nmt', 'code-generation', 'code-translation'] | ['computer-code', 'computer-code', 'computer-code'] | [ 6.88289106e-01 3.62142742e-01 -3.17082107e-01 -4.46089566e-01
-7.17129171e-01 -7.37112045e-01 5.57689071e-01 3.12418014e-01
2.25703850e-01 5.47390997e-01 1.68575183e-01 -8.08107555e-01
3.12536508e-01 -8.86819839e-01 -7.99900293e-01 1.85980070e-02
7.49625266e-02 3.77494335e-01 -4.57845926e-01 -3.09515446e-01
4.99693662e-01 1.48195848e-01 -1.69833422e+00 6.10893548e-01
1.13502419e+00 2.74293125e-01 4.30116117e-01 9.75176930e-01
-7.15359807e-01 8.80730033e-01 -9.25397098e-01 -4.47829992e-01
1.63092315e-01 -9.15531397e-01 -9.30203855e-01 -2.82169074e-01
-6.89590815e-03 1.28686070e-01 9.71310213e-02 1.23039007e+00
1.58546269e-01 -2.17371717e-01 7.61381626e-01 -1.39557695e+00
-1.35382068e+00 1.08506978e+00 -1.42737463e-01 -2.81348050e-01
7.77596772e-01 -5.93282329e-03 7.22693801e-01 -1.18845296e+00
7.89700329e-01 1.20262027e+00 5.02714753e-01 1.06025529e+00
-1.56096411e+00 -2.54938364e-01 -3.58780384e-01 -2.24554613e-01
-1.28193355e+00 -2.53186494e-01 3.04487526e-01 -8.44330251e-01
1.75189567e+00 2.37143397e-01 3.77852291e-01 9.34764802e-01
7.16898680e-01 7.33626127e-01 6.88023865e-01 -8.14145446e-01
2.11456463e-01 3.05217773e-01 1.07856505e-01 9.31784391e-01
3.41921479e-01 2.15120181e-01 -2.02022374e-01 -3.48700911e-01
5.22185385e-01 -5.94691634e-01 4.75088786e-03 -3.55291933e-01
-1.42431533e+00 1.07889032e+00 3.39173645e-01 3.83550733e-01
-1.63999200e-01 2.66013950e-01 7.41347730e-01 6.43100381e-01
3.74394357e-02 1.00332475e+00 -5.14117420e-01 -3.84361535e-01
-1.00959289e+00 3.80348474e-01 1.02575850e+00 1.67132533e+00
6.82868779e-01 6.74512506e-01 -5.14001012e-01 6.96532488e-01
2.49493048e-01 3.74366492e-01 1.00213289e+00 -6.66351438e-01
1.00624299e+00 8.42422247e-01 -1.14440642e-01 -8.70628774e-01
-1.36985481e-01 3.27184587e-03 -6.90517664e-01 3.14886212e-01
-1.85758501e-01 -3.48365575e-01 -1.04159355e+00 1.67431581e+00
-3.42037052e-01 -7.35167027e-01 7.54998028e-01 4.96244490e-01
7.85465300e-01 1.09536302e+00 -2.20956266e-01 -7.34064803e-02
9.64838445e-01 -1.18233800e+00 -4.38290864e-01 -4.74616408e-01
8.58197391e-01 -9.49545562e-01 1.08300912e+00 1.11323662e-01
-1.26179922e+00 -8.14327836e-01 -1.05653310e+00 -2.48344630e-01
-5.58932245e-01 3.81314218e-01 5.92397869e-01 8.03174675e-01
-1.41091788e+00 5.34705579e-01 -5.40637434e-01 -3.45641404e-01
-2.69500948e-02 3.81508648e-01 -1.57072261e-01 1.79086067e-02
-9.63249922e-01 1.03832126e+00 7.25384533e-01 -1.73796982e-01
-1.10000980e+00 -4.83407170e-01 -1.36376631e+00 1.69598937e-01
-2.71324813e-01 -9.32983875e-01 1.72556555e+00 -1.36748600e+00
-1.53486347e+00 8.35358322e-01 -2.23835662e-01 -6.94642007e-01
1.82022423e-01 2.15621993e-01 -3.44245344e-01 -4.36657161e-01
2.17552066e-01 1.01907027e+00 6.82479620e-01 -9.14681673e-01
-3.91738415e-01 2.46318713e-01 8.66613612e-02 -1.03479758e-01
-1.28545344e-01 4.54760998e-01 2.06817393e-04 -9.23115134e-01
-2.79649436e-01 -1.00799751e+00 -3.40875924e-01 -2.69932538e-01
-3.90228599e-01 -5.32141328e-02 3.57994497e-01 -7.60554314e-01
1.38786006e+00 -2.09866381e+00 7.02268064e-01 7.86414742e-02
-2.78104544e-01 8.39508474e-02 -4.91424650e-01 7.57553518e-01
-3.38270813e-01 5.01164436e-01 -7.44597256e-01 -4.65088636e-02
2.58677155e-01 2.29203597e-01 -6.44838631e-01 -2.75327265e-01
6.49814248e-01 1.25182343e+00 -8.20663393e-01 -3.99053097e-01
-2.35682249e-01 2.16577202e-01 -7.85080194e-01 4.78642344e-01
-5.33019841e-01 1.10213064e-01 -2.94396609e-01 6.33490920e-01
1.41352847e-01 -3.48779224e-02 3.22297662e-02 4.71454889e-01
-4.31006610e-01 4.71709222e-01 -6.36886179e-01 1.92886078e+00
-8.66932034e-01 1.04977465e+00 -3.42146307e-01 -5.79406977e-01
1.31332588e+00 5.04337072e-01 -1.88420609e-01 -3.44684243e-01
-4.47274968e-02 4.30672854e-01 2.55935878e-01 -7.96571434e-01
8.41468632e-01 2.27074713e-01 -5.12281775e-01 7.46472895e-01
1.78846896e-01 -6.87933624e-01 6.79188311e-01 -4.39716019e-02
1.00524318e+00 5.06832719e-01 6.50574684e-01 -3.10725421e-01
4.30402130e-01 6.46767020e-01 1.81740269e-01 7.65944541e-01
4.19076949e-01 6.89181805e-01 5.21460176e-01 -3.46290559e-01
-1.54623902e+00 -6.91144824e-01 3.93163294e-01 8.86298120e-01
-4.11243975e-01 -6.37924373e-01 -1.28940988e+00 -3.25449228e-01
-2.55964220e-01 1.15194762e+00 -5.58321118e-01 -4.95961756e-01
-9.14391577e-01 -6.69100344e-01 9.87157762e-01 6.38585091e-01
9.60768238e-02 -1.42197740e+00 -9.22669768e-01 3.94737214e-01
-2.35142708e-01 -6.30693853e-01 -4.30730909e-01 2.70529568e-01
-1.16588533e+00 -6.68732762e-01 -6.34776711e-01 -1.18559325e+00
1.21303415e+00 -6.30744025e-02 1.38521159e+00 3.25589180e-01
-3.05787623e-01 -2.41817236e-02 -5.01959801e-01 -3.20156872e-01
-1.54369152e+00 2.07411870e-01 -1.20964617e-01 -6.03256404e-01
9.87339094e-02 -2.90589839e-01 2.75688648e-01 -1.45302638e-01
-1.21321595e+00 5.89332402e-01 9.99764800e-01 8.88107061e-01
1.42748803e-01 -2.31053919e-01 3.25954348e-01 -8.57603133e-01
1.30235910e+00 -6.13357604e-01 -6.50534809e-01 4.09661680e-01
-8.16981494e-01 5.97702384e-01 1.04386055e+00 -4.27737027e-01
-1.03699052e+00 2.24112704e-01 -1.83571026e-01 -5.37830926e-02
-8.17878246e-02 1.11534262e+00 1.89626306e-01 -5.09603298e-04
1.17737329e+00 6.69596493e-01 -9.29082930e-02 -1.29998118e-01
4.72161502e-01 6.55279636e-01 4.31973666e-01 -1.01848257e+00
9.28650558e-01 -4.65642810e-01 -3.71346653e-01 -1.36384189e-01
-4.97259423e-02 3.65619719e-01 -7.62005329e-01 1.83620632e-01
7.33327091e-01 -5.16152263e-01 2.50480533e-01 7.69509375e-02
-1.69929612e+00 -2.85020500e-01 -2.95585960e-01 1.52149037e-01
-8.47671270e-01 7.57492185e-02 -6.17917895e-01 -5.93028247e-01
-6.85844123e-01 -1.73596585e+00 9.62470829e-01 2.87149251e-01
-8.04643393e-01 -7.12191761e-01 9.05249044e-02 -2.61024266e-01
6.71992540e-01 2.94142127e-01 1.79505348e+00 -4.12577540e-01
-4.72289801e-01 -1.78054228e-01 -4.90511395e-02 2.52037972e-01
1.91783890e-01 3.87374550e-01 -4.33443815e-01 -1.45314276e-01
-1.98375598e-01 -2.06870794e-01 1.64113194e-01 5.99215925e-02
7.51185477e-01 -6.69486880e-01 -2.08359495e-01 5.32156587e-01
1.44524813e+00 6.20185137e-01 5.21928310e-01 2.33950242e-01
2.92722076e-01 6.22777045e-01 3.88412148e-01 4.00663167e-02
1.13155022e-01 5.45935154e-01 2.76170850e-01 1.48073882e-01
-1.17593758e-01 -4.49305773e-01 9.25792873e-01 1.12467897e+00
2.22583473e-01 -1.30085126e-01 -1.24029458e+00 5.80277085e-01
-1.73951197e+00 -6.35895848e-01 -1.77828953e-01 1.78170693e+00
1.01709151e+00 6.46640453e-03 -1.42316401e-01 -1.72308356e-01
8.25064361e-01 -4.23344076e-01 -4.82344687e-01 -1.20679021e+00
1.37462109e-01 1.84176132e-01 1.09531470e-01 2.90591925e-01
-4.32719141e-01 8.64841521e-01 7.59750271e+00 4.12760377e-01
-8.71867120e-01 -2.29665581e-02 3.19019347e-01 3.48976791e-01
-5.23499906e-01 1.83582708e-01 -5.02899110e-01 2.02949807e-01
1.43810320e+00 -8.85702312e-01 8.19754779e-01 1.20156217e+00
9.63652506e-02 2.21733049e-01 -1.68091881e+00 7.42507637e-01
2.94057667e-01 -1.41334999e+00 5.08985102e-01 -3.35833013e-01
1.08476150e+00 -1.27536222e-01 -1.45903647e-01 7.77831376e-01
4.77317899e-01 -1.04277468e+00 1.13091719e+00 3.38910758e-01
1.00636363e+00 -6.34509206e-01 5.90437293e-01 2.80314773e-01
-1.11921024e+00 -4.03596371e-01 -2.48331338e-01 -2.16802269e-01
-3.77559662e-02 2.87586600e-01 -1.10010874e+00 6.83093011e-01
3.48271936e-01 4.70039099e-01 -8.14656973e-01 8.04462969e-01
-4.47607219e-01 2.70815134e-01 3.36976558e-01 -4.11998779e-01
1.43429592e-01 -6.76922174e-03 5.00245690e-01 1.50826919e+00
8.01529169e-01 -4.35713470e-01 2.20216867e-02 1.79141426e+00
1.56410076e-02 1.30978838e-01 -9.69506979e-01 -6.13066196e-01
2.39942312e-01 8.85398746e-01 -5.92434704e-01 -5.72339654e-01
-3.91665757e-01 9.80294764e-01 1.22558303e-01 2.71856874e-01
-8.89259040e-01 -9.13958013e-01 5.18394053e-01 -3.88258904e-01
-1.77419454e-01 -4.88862932e-01 -3.90063494e-01 -1.22485793e+00
1.86383441e-01 -1.29499173e+00 8.78959242e-03 -1.23312390e+00
-6.60759449e-01 1.23758471e+00 3.51278156e-01 -1.36075497e+00
-1.00071430e+00 -6.31141484e-01 -5.22711456e-01 1.03176701e+00
-9.49655056e-01 -6.99686766e-01 3.38339359e-02 -6.85752854e-02
1.13946974e+00 -6.95491970e-01 1.12592840e+00 2.13350117e-01
-3.93592507e-01 5.04144430e-01 -5.81342615e-02 2.94639170e-01
-4.17716503e-02 -1.05337238e+00 1.23074126e+00 1.18818235e+00
1.16835177e-01 1.19082129e+00 7.42305756e-01 -6.42863274e-01
-1.64063108e+00 -1.42671359e+00 1.10214126e+00 -3.36564124e-01
4.51174140e-01 -4.98895884e-01 -6.97302103e-01 6.52093768e-01
4.00702089e-01 -6.95235312e-01 4.40083951e-01 -6.13355219e-01
-4.48098242e-01 5.28479695e-01 -9.94717062e-01 8.41012359e-01
7.78991282e-01 -6.70959651e-01 -6.13507390e-01 1.51236996e-01
9.95807946e-01 -7.11904466e-01 -6.63225770e-01 2.09608003e-01
2.50573128e-01 -5.07814586e-01 5.64397514e-01 -7.66310036e-01
1.12533200e+00 -5.67020774e-01 -1.80806756e-01 -1.47795141e+00
-2.39042178e-01 -6.69432461e-01 -1.07084073e-01 1.18461800e+00
1.00758672e+00 -2.70073533e-01 2.88232923e-01 7.23001063e-01
-6.55485511e-01 -5.97097158e-01 -5.22110045e-01 -7.68462956e-01
1.98960438e-01 -2.46403754e-01 9.08182383e-01 7.55304992e-01
3.85330230e-01 4.51172888e-01 -9.69100296e-02 -1.76568285e-01
-1.26231790e-01 5.61872184e-01 7.25640893e-01 -9.24435914e-01
-4.80101645e-01 -7.57465661e-01 -1.31136313e-01 -7.85584450e-01
4.34545636e-01 -1.57507825e+00 4.71304476e-01 -1.66385579e+00
1.98545098e-01 3.55589241e-02 3.40804130e-01 7.33478785e-01
1.76610231e-01 -2.06038356e-01 2.06451207e-01 2.95276850e-01
2.80612975e-01 3.24413657e-01 7.88821578e-01 -5.89353800e-01
-2.47019738e-01 -1.98638737e-01 -8.12778950e-01 2.36103103e-01
9.43110645e-01 -9.13985491e-01 -4.23582405e-01 -1.01805198e+00
8.02552342e-01 3.52844656e-01 -9.84826032e-03 -9.14909542e-01
-6.09612130e-02 -3.37377489e-01 -3.54835279e-02 5.71058923e-03
-1.80183232e-01 -5.49739361e-01 3.81816834e-01 7.39071906e-01
-8.41525912e-01 1.02135170e+00 4.70661879e-01 2.39591883e-03
-3.89263660e-01 -9.92511094e-01 6.36230648e-01 -4.03728426e-01
-6.76052392e-01 -2.44613856e-01 -1.03147149e+00 -1.78260982e-01
8.67828071e-01 -2.66921431e-01 -2.08411157e-01 -1.51610345e-01
-3.19208622e-01 -3.92375477e-02 6.07468128e-01 1.06963062e+00
8.80777836e-01 -1.41210830e+00 -8.44640970e-01 4.18810964e-01
4.23456341e-01 -3.66246551e-01 -5.13476074e-01 3.87753606e-01
-8.41171324e-01 6.93888664e-01 -5.75435758e-01 -2.19052792e-01
-1.02655172e+00 8.07110012e-01 3.80011499e-01 -1.17414504e-01
-3.00429404e-01 4.28934574e-01 -2.11712897e-01 -8.87460053e-01
-2.85380900e-01 -7.43981719e-01 -7.58982301e-02 -5.69704950e-01
3.94587159e-01 -5.73095074e-03 2.24261373e-01 -5.36990881e-01
-5.91611825e-02 3.51317674e-01 1.84950322e-01 -1.93343952e-01
9.85947549e-01 4.18498933e-01 -6.63593054e-01 3.71354073e-01
1.15556085e+00 -3.71332526e-01 -3.62438858e-01 1.20660529e-01
4.49954152e-01 -1.42547071e-01 -5.96861005e-01 -6.76456213e-01
-7.48189032e-01 9.03815627e-01 3.04572791e-01 2.04508141e-01
1.00105131e+00 -4.45424527e-01 5.20556808e-01 8.46839368e-01
5.78564286e-01 -8.89513314e-01 -4.41691168e-02 9.40387309e-01
1.19630516e+00 -7.23965585e-01 -5.01263201e-01 -1.30832717e-01
-4.78086114e-01 1.57055604e+00 8.06366801e-01 -1.16815045e-02
1.99209917e-02 5.00015199e-01 -2.22631693e-02 1.71599582e-01
-9.44550872e-01 2.29277357e-01 1.25088096e-01 7.85718858e-01
9.23681498e-01 -1.20290138e-01 -3.21919382e-01 3.14219922e-01
-6.56208336e-01 -4.38859314e-02 1.04513204e+00 1.20997572e+00
-1.18989773e-01 -1.50694215e+00 -5.25305212e-01 5.55766761e-01
-1.72612384e-01 -5.10570288e-01 -7.21142471e-01 4.78540808e-01
-2.40571913e-04 1.00107253e+00 -1.99911475e-01 -5.11573672e-01
3.95136505e-01 3.07479739e-01 3.86754692e-01 -1.36984205e+00
-9.73422408e-01 -6.20804727e-01 7.31836781e-02 -1.34976849e-01
-8.27636942e-02 -5.39335370e-01 -1.51916909e+00 7.33386874e-02
-1.26338914e-01 4.47031051e-01 9.26793993e-01 6.45968139e-01
6.72832370e-01 6.72415018e-01 2.67061144e-01 -8.74642968e-01
-2.96458721e-01 -8.83717179e-01 3.34828407e-01 7.05447719e-02
3.26384962e-01 -1.00376248e-01 6.62187859e-02 7.05959916e-01] | [7.767967224121094, 7.800108909606934] |
4cbadca4-c5dd-4eb7-915c-1342bf4e01f5 | model-based-demosaicking-for-acquisitions-by | 2306.01357 | null | https://arxiv.org/abs/2306.01357v1 | https://arxiv.org/pdf/2306.01357v1.pdf | Model-based demosaicking for acquisitions by a RGBW color filter array | Microsatellites and drones are often equipped with digital cameras whose sensing system is based on color filter arrays (CFAs), which define a pattern of color filter overlaid over the focal plane. Recent commercial cameras have started implementing RGBW patterns, which include some filters with a wideband spectral response together with the more classical RGB ones. This allows for additional light energy to be captured by the relevant pixels and increases the overall SNR of the acquisition. Demosaicking defines reconstructing a multi-spectral image from the raw image and recovering the full color components for all pixels. However, this operation is often tailored for the most widespread patterns, such as the Bayer pattern. Consequently, less common patterns that are still employed in commercial cameras are often neglected. In this work, we present a generalized framework to represent the image formation model of such cameras. This model is then exploited by our proposed demosaicking algorithm to reconstruct the datacube of interest with a Bayesian approach, using a total variation regularizer as prior. Some preliminary experimental results are also presented, which apply to the reconstruction of acquisitions of various RGBW cameras. | ['Magnus O Ulfarsson', 'Mauro Dalla Mura', 'Daniele Picone', 'Matthieu Muller'] | 2023-06-02 | null | null | null | null | ['demosaicking'] | ['computer-vision'] | [ 4.77976739e-01 -3.39979023e-01 4.44838583e-01 -1.97884247e-01
-1.05829790e-01 -4.78495270e-01 4.83081490e-01 -3.03800106e-01
-9.17299092e-01 7.38265872e-01 -2.24020749e-01 5.73355854e-02
-1.55266166e-01 -8.34656835e-01 -6.60513520e-01 -9.60824013e-01
3.74759853e-01 1.34329617e-01 2.72078782e-01 2.63653956e-02
3.62583809e-03 6.11989915e-01 -1.89302289e+00 -1.02535468e-02
8.40869784e-01 8.49889696e-01 4.93341595e-01 5.43621600e-01
7.51497075e-02 4.62095588e-01 -3.70232582e-01 -3.58655393e-01
6.13024890e-01 -5.52657425e-01 2.14440376e-01 6.20101452e-01
3.40455055e-01 -5.74460685e-01 -1.91056356e-01 1.33751464e+00
-2.26287935e-02 2.08985120e-01 5.65557182e-01 -6.31161809e-01
8.02330524e-02 1.84914753e-01 -6.06101036e-01 -2.23480463e-01
3.89112532e-01 2.29476377e-01 6.78907692e-01 -8.95409882e-01
4.09477264e-01 8.04019392e-01 4.04024184e-01 1.67639181e-01
-1.19937813e+00 -3.68632376e-01 -3.02990347e-01 3.31100255e-01
-1.59852505e+00 -3.91937047e-01 9.15763021e-01 -2.67489642e-01
2.30543360e-01 3.50265026e-01 1.18533409e+00 7.09531784e-01
6.09244704e-02 2.16535196e-01 1.58116746e+00 -5.56523263e-01
1.81488857e-01 1.50901154e-01 -5.92610650e-02 4.49740201e-01
7.26379275e-01 7.11222962e-02 -5.40579259e-01 2.90453117e-02
5.76026917e-01 2.72428691e-01 -7.07501948e-01 -4.24380600e-01
-8.65465701e-01 4.95000094e-01 2.87882626e-01 4.05957073e-01
-5.16771913e-01 4.56390856e-03 -4.21013832e-01 -1.18430309e-01
1.85732260e-01 1.09638453e-01 1.67508289e-01 1.67502970e-01
-1.10432947e+00 -3.29945683e-02 6.50038600e-01 8.53766978e-01
1.14890385e+00 1.82519257e-02 3.88214409e-01 5.46249926e-01
7.60819554e-01 9.33998227e-01 7.49815255e-02 -9.48169589e-01
5.54544143e-02 6.37998462e-01 2.87062734e-01 -7.28874683e-01
-2.68641561e-01 -3.09835762e-01 -7.07736433e-01 5.91232181e-01
6.07608855e-01 -2.45386049e-01 -6.06088638e-01 1.20751870e+00
3.81723434e-01 2.42364854e-01 1.64715856e-01 1.18921697e+00
2.61914015e-01 7.14069247e-01 -5.79967558e-01 -3.85128498e-01
1.20074189e+00 -7.84469694e-02 -7.77348101e-01 -2.24731475e-01
-2.77343065e-01 -9.09982979e-01 7.75943875e-01 9.81546402e-01
-9.30297554e-01 -3.86274546e-01 -1.12474823e+00 1.65541649e-01
-9.33638960e-02 5.73026121e-01 1.24872804e-01 1.07941902e+00
-9.86084998e-01 2.58530587e-01 -7.29471505e-01 -2.86317348e-01
-1.50068775e-01 -1.95802152e-02 -1.64212674e-01 -2.73231953e-01
-6.48384094e-01 8.66177380e-01 3.02875668e-01 6.00315511e-01
-7.70879626e-01 -3.91721696e-01 -3.59356344e-01 -1.05979927e-01
3.05627316e-01 -3.16469043e-01 6.32037461e-01 -1.23814082e+00
-1.84702587e+00 7.87171841e-01 2.22723380e-01 -3.28535318e-01
6.20983541e-01 -1.94214195e-01 -3.32738966e-01 4.06072676e-01
-3.99822533e-01 1.73260905e-02 1.11987138e+00 -1.44192672e+00
-4.87446874e-01 -4.29722935e-01 4.74020001e-03 2.38442317e-01
-1.54056042e-01 -1.94054529e-01 -5.79047740e-01 -2.12266684e-01
4.86610949e-01 -9.10316527e-01 -7.60308653e-02 1.42065018e-01
-2.03169242e-01 4.82387304e-01 5.61191142e-01 -5.89602888e-01
8.42506111e-01 -2.45663667e+00 4.33066368e-01 4.62205023e-01
-4.12615426e-02 5.30314222e-02 2.33962670e-01 6.11456752e-01
1.65954620e-01 -7.79536128e-01 -4.76167411e-01 -3.43293041e-01
-3.69329989e-01 2.67416894e-01 1.38787646e-02 1.00489056e+00
-3.46784174e-01 -1.49783134e-01 -4.69858557e-01 -3.05247515e-01
5.65137565e-01 4.88595068e-01 -4.22245443e-01 3.95689279e-01
-2.25338623e-01 4.49733883e-01 -1.84468538e-01 4.19127315e-01
1.14552081e+00 3.84036928e-01 3.51922095e-01 -4.60843503e-01
-7.43047893e-01 -5.36590159e-01 -1.57800186e+00 1.71335304e+00
-2.61805207e-01 4.47748721e-01 6.92979515e-01 -7.28886485e-01
1.11121058e+00 7.51694664e-02 5.53598523e-01 -2.23063916e-01
8.79266933e-02 3.65073115e-01 -1.02741517e-01 -6.13264441e-01
5.49848199e-01 -4.03976649e-01 3.85779798e-01 2.74635316e-03
-4.48989086e-02 -4.10679132e-01 2.39592001e-01 -5.98918609e-02
6.82144821e-01 3.31563264e-01 2.21212178e-01 -3.38195175e-01
6.85642242e-01 -1.44344434e-01 5.17951310e-01 3.58689576e-01
3.81760925e-01 7.80728042e-01 1.57327965e-01 -2.13510036e-01
-9.68458712e-01 -8.71428609e-01 -1.96297392e-01 1.38620660e-01
3.58902127e-01 -1.35788530e-01 -7.43351042e-01 -7.24526413e-04
-2.43460819e-01 6.34640396e-01 -2.08439067e-01 2.92821992e-02
-2.54687548e-01 -9.08804595e-01 1.62262633e-01 -3.68190289e-01
7.98203051e-01 -4.61362690e-01 -1.21696723e+00 9.22206566e-02
-5.29399253e-02 -1.04979622e+00 1.65983841e-01 3.09235509e-02
-9.42562640e-01 -1.50311744e+00 -7.58205295e-01 -8.47944394e-02
6.23460650e-01 5.40370464e-01 6.85263336e-01 -1.62503719e-01
-3.15975994e-01 8.63379598e-01 -7.67406940e-01 -7.49801248e-02
-1.59528971e-01 -6.42306626e-01 -8.95857885e-02 8.77886653e-01
2.73419291e-01 -5.20395696e-01 -6.35534823e-01 1.81893229e-01
-1.22528243e+00 1.32038757e-01 7.32825577e-01 5.50640345e-01
3.78950685e-01 1.42173097e-02 -3.95841211e-01 -6.77994430e-01
3.04491948e-02 -1.10982522e-01 -1.19868600e+00 1.47986814e-01
-2.41527006e-01 -3.26742113e-01 6.70938194e-01 -9.37740412e-03
-1.40015221e+00 3.82124752e-01 -7.08731264e-02 -4.06300008e-01
-3.71277809e-01 3.00266325e-01 -3.74448508e-01 -4.03236061e-01
5.62713444e-01 3.99610579e-01 1.67303905e-01 -6.28480971e-01
1.05734006e-01 5.24467528e-01 3.96625072e-01 -3.04492474e-01
9.35452580e-01 7.81009972e-01 3.12184960e-01 -1.60484314e+00
-3.00108641e-01 -5.14471650e-01 -3.26282591e-01 -8.06427062e-01
9.19523418e-01 -9.51207459e-01 -5.48823118e-01 7.55711854e-01
-9.66675460e-01 -2.25902069e-02 -2.77599812e-01 9.23342288e-01
-3.11522901e-01 6.81083083e-01 -2.86553890e-01 -1.10795033e+00
2.47512385e-01 -1.26244295e+00 7.95441508e-01 3.92187536e-01
4.91084754e-01 -6.07389569e-01 9.48991179e-02 1.65853128e-01
2.18850717e-01 6.34718314e-02 4.90619391e-01 3.70232433e-01
-9.18429077e-01 -1.62748501e-01 1.75649766e-02 6.07823372e-01
-1.00304596e-01 2.49038488e-01 -1.07449138e+00 -2.00897291e-01
3.79705817e-01 1.32869571e-01 7.61717260e-01 4.81161743e-01
8.80886972e-01 3.70421946e-01 5.67680188e-02 9.24978435e-01
2.08135772e+00 2.44320527e-01 8.18774581e-01 2.18898967e-01
3.53836030e-01 6.74068809e-01 7.51534343e-01 7.73045719e-01
-1.08746588e-02 8.45654368e-01 9.11751568e-01 -4.61838581e-02
-9.04749148e-03 2.95203716e-01 4.56259251e-01 4.20347214e-01
-6.04794443e-01 -2.72574097e-01 -5.84922731e-01 1.92008361e-01
-1.46651649e+00 -7.39063382e-01 -6.71653807e-01 2.51511049e+00
4.33595806e-01 -3.55026454e-01 -5.74226156e-02 3.06944311e-01
7.28747666e-01 1.21943712e-01 -5.74766397e-02 2.63100892e-01
-4.37499255e-01 3.38449270e-01 8.29431534e-01 5.81207752e-01
-7.73987472e-01 3.54645878e-01 5.38602877e+00 4.88347560e-01
-9.65613544e-01 -4.12061214e-02 -7.39180520e-02 -5.65692410e-02
-3.69874448e-01 1.66898772e-01 -6.57263398e-01 6.38588786e-01
5.43804348e-01 2.85536706e-01 6.07134461e-01 3.95948738e-01
3.89562547e-01 -8.44465256e-01 -5.33626914e-01 1.31359303e+00
2.31247082e-01 -6.26567483e-01 -1.54616252e-01 2.37135187e-01
5.08282542e-01 -2.15796992e-01 2.71745324e-02 -6.44604266e-01
-1.07127495e-01 -1.95428923e-01 8.08858216e-01 9.94097352e-01
6.69339240e-01 -6.81048334e-01 5.32129824e-01 2.56491870e-01
-8.61996949e-01 -2.12448344e-01 -6.90551758e-01 -3.53339803e-03
1.65600881e-01 1.01778352e+00 -3.93176556e-01 8.61429870e-01
6.71486080e-01 7.36669421e-01 -4.87423450e-01 1.37447655e+00
-2.83174336e-01 4.60144848e-01 -6.88785434e-01 5.39027946e-03
1.44376650e-01 -1.34727848e+00 7.68482029e-01 8.74025345e-01
1.00725281e+00 4.10390437e-01 -1.43987626e-01 9.45475936e-01
2.76677281e-01 -4.24053930e-02 -6.48833692e-01 2.47042313e-01
7.99392834e-02 1.42652345e+00 -7.94749379e-01 -1.24751039e-01
-7.40024209e-01 8.55402112e-01 -4.43018258e-01 5.51406741e-01
-6.94041669e-01 -2.32113972e-01 5.92464149e-01 2.41693780e-01
2.14145973e-01 -4.88784164e-01 2.14316294e-01 -1.40859258e+00
-2.66910829e-02 -7.60291517e-01 1.75378826e-02 -9.87652242e-01
-8.86785150e-01 2.80123532e-01 1.44178972e-01 -1.67143500e+00
1.82888553e-01 -1.01817560e+00 -1.93161353e-01 8.39292109e-01
-1.39560866e+00 -1.04882443e+00 -5.78952491e-01 9.39048886e-01
2.04721555e-01 -8.30854010e-03 5.07671893e-01 4.24182236e-01
-4.22367007e-01 -3.38109463e-01 4.71305221e-01 -3.33758414e-01
6.44007385e-01 -1.11090350e+00 -7.04635620e-01 1.51173460e+00
1.34252116e-01 4.13239777e-01 1.01472867e+00 -4.87603605e-01
-1.83242786e+00 -5.21890938e-01 4.85545583e-02 2.63408244e-01
3.12246650e-01 -1.48293957e-01 -3.42501253e-01 4.86750603e-01
3.84728163e-01 -2.19683275e-01 5.24936616e-01 -4.52316135e-01
1.74514189e-01 -6.82491422e-01 -1.00124633e+00 2.02303872e-01
5.03707707e-01 -2.82548606e-01 -3.94416094e-01 6.02526478e-02
-2.21953899e-01 -2.05783412e-01 -5.56933939e-01 7.81446472e-02
6.62847638e-01 -1.72765756e+00 8.40579510e-01 4.48619783e-01
6.20202683e-02 -8.09513390e-01 -2.98720092e-01 -1.36095560e+00
-7.86902104e-03 -3.10703903e-01 4.17780071e-01 1.00916982e+00
-4.74253818e-02 -7.87377536e-01 6.08639121e-01 3.33171904e-01
-1.36219725e-01 3.89458984e-01 -9.81687903e-01 -3.63250911e-01
-6.61700726e-01 -2.87181824e-01 1.98661104e-01 5.93956113e-01
-3.47828388e-01 -2.04450727e-01 -5.46178579e-01 5.46308398e-01
1.30589128e+00 2.37345085e-01 7.02843785e-01 -1.35139632e+00
-6.50219381e-01 -6.42047971e-02 -5.17959535e-01 -8.34506452e-01
-1.50858849e-01 -3.61877173e-01 2.01580688e-01 -1.02610457e+00
-1.24651827e-01 -2.20504880e-01 2.46469498e-01 -4.17850971e-01
8.22083503e-02 3.98850769e-01 2.65573114e-01 9.94590372e-02
-6.69543967e-02 3.32724512e-01 8.86815429e-01 1.16828375e-01
2.83449199e-02 1.74953803e-01 5.27821295e-02 8.12719226e-01
3.51168156e-01 -1.72971830e-01 -2.90232033e-01 -5.15167892e-01
5.75520515e-01 4.20722254e-02 4.91061449e-01 -1.51542878e+00
4.71643448e-01 -9.85858124e-03 4.24910694e-01 -5.78622818e-01
7.29020953e-01 -1.57879770e+00 1.00417602e+00 4.49224055e-01
3.68057787e-01 -3.10382754e-01 -2.34348565e-01 6.57459855e-01
-3.86995822e-01 -7.64728069e-01 1.13077056e+00 -3.39743644e-01
-7.48667598e-01 -1.08593427e-01 -7.28735268e-01 -6.14114761e-01
1.04530251e+00 -3.25411737e-01 4.79095951e-02 -3.22082996e-01
-5.76636672e-01 -4.05453682e-01 8.85830462e-01 -5.03833294e-01
8.31434250e-01 -9.01251137e-01 -4.19177324e-01 5.46768367e-01
3.24237975e-03 -1.65439025e-01 4.67580795e-01 9.19865787e-01
-1.16998613e+00 -7.03595430e-02 -4.16961461e-01 -5.70642412e-01
-1.17487586e+00 2.89928496e-01 5.05359292e-01 3.99905056e-01
-6.56901300e-01 5.79241276e-01 -6.68294355e-02 1.73417285e-01
-1.20168529e-01 -4.41541880e-01 -3.33740771e-01 2.87709922e-01
4.51904714e-01 5.99511266e-01 9.99326035e-02 -6.92160666e-01
-2.10419789e-01 1.05876577e+00 8.24518502e-01 -3.60742420e-01
1.40083861e+00 -2.86167204e-01 -4.69164699e-01 4.69716221e-01
6.58464968e-01 4.26902294e-01 -1.42238295e+00 3.98995355e-03
-3.37685734e-01 -8.86954188e-01 3.12047005e-01 -3.46496344e-01
-1.21714103e+00 9.15655792e-01 7.17032611e-01 3.26943070e-01
1.56747663e+00 -5.12162924e-01 3.12505141e-02 2.80500025e-01
7.32850194e-01 -1.04820585e+00 -3.58680993e-01 -3.88781987e-02
5.48167109e-01 -6.70168936e-01 3.01310003e-01 -3.74529183e-01
-3.81626606e-01 1.54293120e+00 1.90476418e-01 -3.01046818e-01
5.10158598e-01 1.94118261e-01 -2.10112050e-01 -1.47162536e-02
-1.36999741e-01 -7.00824738e-01 -1.44430071e-01 3.14952701e-01
1.93381399e-01 -9.05025192e-03 -5.20270646e-01 7.53080100e-02
3.68102014e-01 2.16850005e-02 1.01152694e+00 5.36948681e-01
-5.47623694e-01 -1.11639547e+00 -1.01458657e+00 1.67918578e-01
-2.33768389e-01 5.99203631e-02 -2.22318381e-01 7.28679001e-01
3.28419626e-01 9.42278981e-01 -8.97652581e-02 -9.68480706e-02
5.12658060e-01 1.98775940e-02 7.72391796e-01 -3.72834980e-01
-3.43197078e-01 5.94670296e-01 3.55053283e-02 -5.37476659e-01
-9.65643346e-01 -1.03825581e+00 -5.38218439e-01 -5.17475419e-02
-2.40232259e-01 1.60167113e-01 9.00593579e-01 5.46856403e-01
-3.90464514e-01 1.62399366e-01 6.01635396e-01 -1.11034942e+00
-2.65574753e-01 -9.42057610e-01 -1.46331298e+00 1.90421507e-01
3.13604176e-01 -6.67070687e-01 -6.09419703e-01 2.39962846e-01] | [10.214545249938965, -2.543914556503296] |
097e7580-edf9-4d53-985c-2eeb515fe21b | robust-and-controllable-object-centric | 2210.05519 | null | https://arxiv.org/abs/2210.05519v1 | https://arxiv.org/pdf/2210.05519v1.pdf | Robust and Controllable Object-Centric Learning through Energy-based Models | Humans are remarkably good at understanding and reasoning about complex visual scenes. The capability to decompose low-level observations into discrete objects allows us to build a grounded abstract representation and identify the compositional structure of the world. Accordingly, it is a crucial step for machine learning models to be capable of inferring objects and their properties from visual scenes without explicit supervision. However, existing works on object-centric representation learning either rely on tailor-made neural network modules or strong probabilistic assumptions in the underlying generative and inference processes. In this work, we present \ours, a conceptually simple and general approach to learning object-centric representations through an energy-based model. By forming a permutation-invariant energy function using vanilla attention blocks readily available in Transformers, we can infer object-centric latent variables via gradient-based MCMC methods where permutation equivariance is automatically guaranteed. We show that \ours can be easily integrated into existing architectures and can effectively extract high-quality object-centric representations, leading to better segmentation accuracy and competitive downstream task performance. Further, empirical evaluations show that \ours's learned representations are robust against distribution shift. Finally, we demonstrate the effectiveness of \ours in systematic compositional generalization, by re-composing learned energy functions for novel scene generation and manipulation. | ['Liam Paull', 'Yoshua Bengio', 'Marco Pavone', 'Renhao Wang', 'Boris Ivanovic', 'Tong Che', 'Ruixiang Zhang'] | 2022-10-11 | null | null | null | null | ['scene-generation'] | ['computer-vision'] | [ 4.19548631e-01 1.22013882e-01 -1.77478328e-01 -6.05804861e-01
-8.05724740e-01 -7.45563209e-01 9.50695157e-01 -1.33961961e-01
-1.10896945e-01 4.67698723e-01 2.06437781e-01 -7.42988139e-02
-6.51889741e-02 -9.10070539e-01 -1.17143118e+00 -7.00383902e-01
2.46310625e-02 8.52295995e-01 6.08633608e-02 3.38169950e-04
2.38361210e-01 6.66565537e-01 -1.70888150e+00 5.08921802e-01
8.59830379e-01 7.32071221e-01 4.80347008e-01 7.99771488e-01
-7.48509467e-02 7.50062048e-01 -2.55264521e-01 -4.25384223e-01
1.93353131e-01 -6.53397322e-01 -9.78228927e-01 3.30729216e-01
6.60457015e-01 -2.63477445e-01 -1.77847892e-01 1.04042816e+00
-4.17948840e-03 4.00098979e-01 1.06288564e+00 -1.11479175e+00
-1.09624445e+00 5.97981095e-01 -4.74271059e-01 9.85050201e-02
-4.65045422e-02 3.66687268e-01 1.36796272e+00 -8.58653963e-01
7.89939821e-01 1.49148619e+00 3.52068901e-01 5.88501215e-01
-1.85615766e+00 -2.25410804e-01 6.06775820e-01 1.99637488e-01
-1.16880381e+00 -4.07554060e-01 9.05488253e-01 -5.53609312e-01
1.05727386e+00 2.21828610e-01 6.49486780e-01 1.17582166e+00
-3.06659881e-02 1.04595053e+00 9.58211780e-01 -2.97410339e-01
3.59786332e-01 1.03827193e-02 -8.66295025e-02 8.40341151e-01
2.92445838e-01 -1.80831492e-01 -4.46293473e-01 1.44013047e-01
9.73430336e-01 1.29946709e-01 -5.40706664e-02 -8.39612544e-01
-1.38826036e+00 8.29096317e-01 8.51713717e-01 1.10297181e-01
-3.41515541e-01 7.51950383e-01 1.80149794e-01 -1.97077438e-01
4.09721464e-01 5.86522877e-01 -3.30727875e-01 2.23129794e-01
-8.90495896e-01 3.62515181e-01 4.95325118e-01 1.19760442e+00
9.42301869e-01 2.04861641e-01 -2.78659016e-01 5.86260557e-01
3.93966317e-01 3.88094425e-01 2.60069579e-01 -1.27022839e+00
1.61363706e-01 5.11577070e-01 -5.63149489e-02 -7.22216666e-01
-1.05807178e-01 -4.52505022e-01 -7.54035592e-01 2.27004051e-01
2.02289879e-01 1.68907613e-01 -1.26935196e+00 2.11180449e+00
9.64329392e-02 2.14309663e-01 -3.86985503e-02 8.03621769e-01
4.70677853e-01 5.87163270e-01 2.64312804e-01 1.57042459e-01
1.36449885e+00 -8.32404196e-01 -1.97113812e-01 -3.25108916e-01
1.96020707e-01 -2.63103068e-01 1.18986475e+00 2.01518282e-01
-1.24046147e+00 -6.90773129e-01 -8.96687210e-01 -4.62770909e-01
-3.47102016e-01 -9.65900347e-03 1.05338371e+00 4.33709919e-01
-1.07955718e+00 6.18685007e-01 -1.18473494e+00 -1.83276325e-01
9.56067979e-01 1.79561928e-01 -2.71172583e-01 -1.19673796e-01
-4.95188177e-01 6.91007495e-01 5.07421136e-01 3.73879597e-02
-1.32096159e+00 -8.22771192e-01 -1.14410019e+00 1.30364478e-01
3.11517358e-01 -1.28504694e+00 1.17345345e+00 -9.70826924e-01
-1.37233067e+00 9.36584771e-01 -5.23049057e-01 -4.50225353e-01
3.07382017e-01 -1.63626611e-01 2.33821332e-01 3.21836829e-01
2.01185092e-01 1.09773314e+00 1.08520758e+00 -1.63277161e+00
-4.30373162e-01 -2.60949492e-01 2.04489559e-01 1.32002920e-01
7.96741918e-02 -3.57293785e-01 -4.13692981e-01 -7.87048817e-01
2.71615267e-01 -8.52527320e-01 -2.75726110e-01 2.04819351e-01
-4.01789665e-01 -2.67241359e-01 3.99083346e-01 -4.79809731e-01
4.14575905e-01 -2.05351257e+00 5.99880219e-01 -3.46030109e-02
2.85403937e-01 -1.17590897e-01 -1.45491138e-01 2.06707209e-01
1.19497076e-01 2.78192848e-01 -6.37973785e-01 -5.00649273e-01
3.97191465e-01 4.63700354e-01 -6.22668326e-01 3.57628584e-01
8.09030116e-01 1.49029207e+00 -9.91697848e-01 -3.35025758e-01
3.61506850e-01 4.63165104e-01 -1.05321729e+00 1.68260455e-01
-7.93287098e-01 5.47431171e-01 -2.45376185e-01 5.20653486e-01
4.84631449e-01 -4.69941020e-01 2.22412959e-01 -2.33561441e-01
3.06827456e-01 5.11258245e-01 -8.51599991e-01 2.12108541e+00
-6.01700187e-01 6.71865106e-01 -1.20375879e-01 -1.32069480e+00
5.58764577e-01 -1.05213001e-01 5.77887557e-02 -3.84986520e-01
-1.41489282e-02 -6.50156885e-02 -8.34382549e-02 -3.31593126e-01
3.58093500e-01 -4.61052567e-01 -1.62376791e-01 4.98275995e-01
4.13662344e-01 -5.32695711e-01 1.39912561e-01 3.29505146e-01
6.60459578e-01 6.90439224e-01 1.41239226e-01 -3.68518621e-01
1.62281856e-01 3.66863795e-02 4.22638535e-01 9.13998008e-01
1.78368405e-01 7.16816187e-01 3.46543312e-01 -2.83602625e-01
-1.12162793e+00 -1.71807539e+00 -7.07268491e-02 1.23489702e+00
2.77069807e-02 -1.19603701e-01 -6.20198011e-01 -5.51972270e-01
5.20602567e-03 9.89309132e-01 -7.82600820e-01 -2.83704698e-01
-6.33192062e-01 -7.48747706e-01 3.06365192e-01 8.99184108e-01
4.73922312e-01 -1.11696184e+00 -5.17400384e-01 1.38612553e-01
-1.74672976e-01 -1.03464055e+00 -1.14086695e-01 1.88881308e-01
-9.52792585e-01 -7.68843532e-01 -5.42371869e-01 -7.64400601e-01
9.45087314e-01 2.84076154e-01 1.26569235e+00 -1.38520032e-01
-6.24940515e-01 4.77890879e-01 -6.64660856e-02 -2.64816463e-01
-3.16408902e-01 5.37361167e-02 -2.56924957e-01 2.04813723e-02
1.93434760e-01 -8.16826165e-01 -5.90080261e-01 -1.35726035e-01
-9.51371670e-01 2.69788384e-01 6.32304609e-01 6.87501669e-01
7.62979567e-01 -1.01205833e-01 3.25494647e-01 -9.81730759e-01
1.72017828e-01 -4.37426686e-01 -5.90394258e-01 1.42755613e-01
-1.90933496e-01 5.37967443e-01 5.65146506e-01 -3.19810390e-01
-1.30879784e+00 1.20426096e-01 -3.34672369e-02 -4.34849292e-01
-4.15940702e-01 2.69509703e-01 -3.97826850e-01 3.91318649e-01
5.52420318e-01 4.41507876e-01 -2.32418507e-01 -4.38923627e-01
1.12449849e+00 -3.02163139e-02 7.83487439e-01 -9.99224424e-01
8.54688048e-01 9.35563445e-01 2.84716976e-03 -6.56935990e-01
-1.05209064e+00 -2.37067670e-01 -8.61148894e-01 2.42487743e-01
1.23941278e+00 -1.04006124e+00 -6.78134978e-01 2.81084210e-01
-1.04972374e+00 -5.00772536e-01 -5.53247035e-01 3.00472766e-01
-9.40297604e-01 1.91834882e-01 -5.41533113e-01 -6.57135606e-01
1.11450531e-01 -1.03964365e+00 1.35132504e+00 2.94074584e-02
-3.27471644e-01 -1.16302824e+00 -4.80175465e-02 4.51309800e-01
7.23899826e-02 3.05470258e-01 1.25402510e+00 -2.46553168e-01
-1.17650688e+00 2.19635636e-01 -3.57412010e-01 3.73898417e-01
1.19887598e-01 6.19898364e-02 -1.20042169e+00 -2.35138148e-01
-9.47762281e-02 -3.20252389e-01 1.37772691e+00 4.81383651e-01
1.56058931e+00 -4.21845973e-01 -3.03711921e-01 9.62400913e-01
1.45289791e+00 -3.18080425e-01 6.13487899e-01 -5.69145791e-02
9.36968684e-01 6.71831608e-01 -1.40441190e-02 1.43557176e-01
5.76686561e-01 4.17386919e-01 4.32494491e-01 1.52010992e-02
-3.35055470e-01 -4.94127452e-01 2.49356002e-01 5.54432333e-01
-1.66510299e-01 -2.06226408e-01 -6.74796641e-01 7.13023007e-01
-1.82232821e+00 -1.19982386e+00 8.12562257e-02 1.81810248e+00
7.96847463e-01 9.05286297e-02 -1.02723658e-01 -2.76044101e-01
4.49223489e-01 8.98823813e-02 -8.25199842e-01 -2.15203777e-01
-6.99725896e-02 4.26116049e-01 2.31390759e-01 5.35427809e-01
-1.07447278e+00 1.17727256e+00 6.86272573e+00 6.02720737e-01
-7.12426424e-01 8.50928128e-02 4.59516734e-01 -2.43389219e-01
-8.69374454e-01 6.13427646e-02 -6.06515229e-01 1.53885245e-01
6.01901293e-01 -3.63584273e-02 5.39928973e-01 9.06225264e-01
-1.10488914e-01 -1.78048722e-02 -1.49238968e+00 9.01541770e-01
1.05834030e-01 -1.67510247e+00 5.83793342e-01 1.46818399e-01
9.14856195e-01 8.54784772e-02 2.02862799e-01 2.88864791e-01
8.91546667e-01 -1.19065249e+00 9.54148769e-01 6.16782725e-01
5.97405672e-01 -5.74962735e-01 1.77705571e-01 2.89506614e-01
-1.15409100e+00 8.26029330e-02 -5.43838322e-01 -1.08698852e-01
1.16013683e-01 4.55060214e-01 -6.66866899e-01 4.93436605e-01
5.83537221e-01 8.84351730e-01 -4.63406324e-01 6.11809313e-01
-5.43986261e-01 6.04540706e-01 -1.89710811e-01 8.64619166e-02
1.96688518e-01 -1.81838751e-01 4.45466518e-01 1.25776994e+00
4.77626063e-02 2.31402609e-02 5.24194092e-02 1.71216989e+00
-3.70216191e-01 -3.96106750e-01 -5.88589549e-01 -1.00381203e-01
1.24528073e-01 1.27834046e+00 -9.82403517e-01 -5.28603494e-01
-2.09764972e-01 1.17243826e+00 6.23679280e-01 6.18522108e-01
-8.36737096e-01 1.26100879e-03 9.41180825e-01 -1.20122783e-01
7.36779809e-01 -5.78503788e-01 -4.43440914e-01 -1.44775665e+00
-2.84179132e-02 -6.20059252e-01 1.55892745e-01 -8.76989007e-01
-1.51139212e+00 2.08025679e-01 2.26399004e-01 -6.72660530e-01
-4.07099396e-01 -8.72544527e-01 -5.02610207e-01 8.79840374e-01
-1.39981902e+00 -1.47060692e+00 -1.87789798e-01 5.64845085e-01
7.73615956e-01 1.67757630e-01 6.90123200e-01 -1.48557335e-01
-3.60312402e-01 2.81417042e-01 -3.25476564e-02 1.32711127e-01
2.11936980e-01 -1.52859890e+00 6.90491080e-01 1.02525926e+00
7.03173995e-01 9.25889432e-01 6.28778756e-01 -5.19770443e-01
-1.49008811e+00 -1.20889521e+00 3.10315996e-01 -9.05024707e-01
5.91014326e-01 -9.03759837e-01 -9.59387600e-01 1.11203504e+00
1.37622422e-02 1.07810311e-01 5.92453241e-01 2.31004387e-01
-7.70702481e-01 1.04118079e-01 -9.47875142e-01 7.31357574e-01
1.44169104e+00 -8.77808571e-01 -8.93123567e-01 2.66842604e-01
8.37440372e-01 -1.06680110e-01 -5.66445768e-01 2.87472874e-01
3.73638541e-01 -8.39355707e-01 1.20061123e+00 -9.63489234e-01
5.17436981e-01 -3.21329564e-01 -3.12198818e-01 -1.10702062e+00
-6.78775549e-01 -3.59705955e-01 -1.51338264e-01 1.15592873e+00
3.61325979e-01 -5.03184974e-01 6.98902905e-01 5.26201665e-01
-2.33643577e-01 -4.98033226e-01 -6.00842714e-01 -7.26871848e-01
3.86358470e-01 -6.84987664e-01 5.13625801e-01 6.88543856e-01
-4.70918864e-01 3.75125706e-01 -4.78150882e-02 3.47513765e-01
7.91709602e-01 5.82532883e-01 7.57596970e-01 -1.15041566e+00
-6.28256857e-01 -6.21074617e-01 -4.42401081e-01 -1.39123404e+00
4.37356770e-01 -1.28075635e+00 3.33158553e-01 -1.86720109e+00
5.10132909e-01 -3.12632173e-01 -2.56836653e-01 3.74467134e-01
-1.81679681e-01 3.29613090e-01 1.82423487e-01 2.48536989e-01
-5.82085729e-01 7.99588203e-01 1.26337636e+00 -2.83557743e-01
8.60652477e-02 -2.44067997e-01 -9.08716500e-01 9.27674830e-01
6.48625493e-01 -3.29421639e-01 -6.44907892e-01 -7.10427225e-01
3.32646430e-01 -6.07257724e-01 9.57838297e-01 -7.22222447e-01
-1.52735487e-01 -3.69241178e-01 6.14380717e-01 -4.24381435e-01
5.06150901e-01 -4.45370644e-01 -1.51477745e-02 1.91219568e-01
-4.07750517e-01 -4.07286584e-01 1.09772511e-01 7.29191303e-01
1.46112056e-03 -1.02478683e-01 7.12400794e-01 -5.24208426e-01
-9.06351984e-01 3.51677060e-01 -2.76796877e-01 2.77649283e-01
8.96348476e-01 -1.08303368e-01 -2.31832221e-01 -1.48526505e-01
-7.75886476e-01 -4.62614559e-02 6.41757429e-01 3.88166815e-01
6.63436174e-01 -1.19805574e+00 -6.73143923e-01 2.37147659e-01
1.11603327e-01 2.80202746e-01 1.64316252e-01 4.75155145e-01
-4.32437360e-01 2.51944482e-01 -2.17837140e-01 -9.19806063e-01
-6.64600015e-01 6.89060628e-01 2.64879972e-01 9.93588939e-02
-7.84599841e-01 1.13871241e+00 8.20316195e-01 -5.01356125e-01
-1.56180367e-01 -7.32450664e-01 3.94674093e-01 -1.12178758e-01
3.12009811e-01 1.99678186e-02 -3.37005287e-01 -4.85781997e-01
-3.13306719e-01 5.24835169e-01 3.69556136e-02 -3.69699985e-01
1.40407038e+00 -1.68979019e-01 -1.80757597e-01 5.78307152e-01
1.19435060e+00 -1.50319651e-01 -1.67993474e+00 -1.07238978e-01
-1.15942933e-01 -3.37019086e-01 -4.83626202e-02 -5.39354444e-01
-8.33668411e-01 1.26233184e+00 4.98696752e-02 -4.46355827e-02
9.11955774e-01 6.09699547e-01 4.62008297e-01 5.14074743e-01
4.00211722e-01 -7.49222279e-01 3.38854700e-01 2.53821015e-01
8.96560192e-01 -1.23221231e+00 7.25807156e-04 -4.00465935e-01
-5.50612211e-01 8.00267279e-01 3.95320624e-01 -4.18950140e-01
4.78610903e-01 -1.97262410e-02 -3.70441139e-01 -3.09196353e-01
-7.88453579e-01 -4.71713096e-01 6.15735352e-01 8.19705307e-01
2.21954405e-01 1.09281704e-01 4.93426055e-01 2.58458763e-01
-2.89328575e-01 -4.36744004e-01 2.45250493e-01 7.51147270e-01
-5.39549768e-01 -8.33522618e-01 -5.36233885e-04 3.91233236e-01
-1.08921364e-01 -3.32466036e-01 -2.28722930e-01 7.33784318e-01
2.06521764e-01 5.23276269e-01 3.44633609e-01 1.62068620e-01
-6.58858120e-02 1.84398472e-01 1.00973797e+00 -9.96644497e-01
1.14493589e-04 -3.69099453e-02 -2.21846759e-01 -6.64444685e-01
-5.65719962e-01 -8.04712772e-01 -1.41723287e+00 -8.40375274e-02
4.91657928e-02 -3.00553709e-01 5.47534943e-01 9.98150945e-01
4.88689482e-01 8.03108871e-01 1.90095499e-01 -1.12781119e+00
-2.79170036e-01 -5.95713019e-01 -2.97939479e-01 5.99486470e-01
3.41347009e-01 -7.20720291e-01 -2.24814326e-01 5.90149522e-01] | [10.005575180053711, 0.6707603931427002] |
8aa39a42-007b-4e0b-8eed-36d654d68cf4 | end-to-end-multi-view-lipreading | 1709.00443 | null | http://arxiv.org/abs/1709.00443v1 | http://arxiv.org/pdf/1709.00443v1.pdf | End-to-End Multi-View Lipreading | Non-frontal lip views contain useful information which can be used to enhance
the performance of frontal view lipreading. However, the vast majority of
recent lipreading works, including the deep learning approaches which
significantly outperform traditional approaches, have focused on frontal mouth
images. As a consequence, research on joint learning of visual features and
speech classification from multiple views is limited. In this work, we present
an end-to-end multi-view lipreading system based on Bidirectional Long-Short
Memory (BLSTM) networks. To the best of our knowledge, this is the first model
which simultaneously learns to extract features directly from the pixels and
performs visual speech classification from multiple views and also achieves
state-of-the-art performance. The model consists of multiple identical streams,
one for each view, which extract features directly from different poses of
mouth images. The temporal dynamics in each stream/view are modelled by a BLSTM
and the fusion of multiple streams/views takes place via another BLSTM. An
absolute average improvement of 3% and 3.8% over the frontal view performance
is reported on the OuluVS2 database when the best two (frontal and profile) and
three views (frontal, profile, 45) are combined, respectively. The best
three-view model results in a 10.5% absolute improvement over the current
multi-view state-of-the-art performance on OuluVS2, without using external
databases for training, achieving a maximum classification accuracy of 96.9%. | ['Yujiang Wang', 'Zuwei Li', 'Maja Pantic', 'Stavros Petridis'] | 2017-09-01 | null | null | null | null | ['lipreading'] | ['computer-vision'] | [-1.46467611e-02 -5.79628795e-02 -5.32465518e-01 -1.39449537e-01
-1.15163314e+00 -1.12780161e-01 8.16223323e-01 -2.84188449e-01
-3.77313823e-01 3.62176001e-01 4.04615521e-01 -8.29135403e-02
5.82933545e-01 -4.76448052e-02 -5.73612213e-01 -7.98515379e-01
3.92302066e-01 -4.33280831e-03 3.52385223e-01 1.26245737e-01
1.71670273e-01 2.91693628e-01 -2.10492849e+00 7.57424533e-01
3.66176724e-01 1.29303825e+00 4.07183826e-01 9.59850371e-01
-5.63247763e-02 2.86434859e-01 -3.65907639e-01 -5.52199483e-01
-1.47559464e-01 -2.51061141e-01 -3.69264930e-01 2.53304780e-01
7.93375611e-01 -5.75389326e-01 -3.21536779e-01 6.42669201e-01
1.18946850e+00 -1.50060013e-01 5.71351290e-01 -1.11515760e+00
-5.79108953e-01 -1.29816905e-01 -6.49040222e-01 1.67511985e-01
4.57721025e-01 3.74053210e-01 8.17021668e-01 -1.52322149e+00
4.48370576e-01 1.26175225e+00 3.92034262e-01 9.63752568e-01
-9.56037700e-01 -5.42007685e-01 1.79278165e-01 4.35391903e-01
-9.59790170e-01 -1.51557958e+00 7.16163456e-01 -2.21880153e-01
1.29485941e+00 -2.06491530e-01 6.21438920e-01 1.39040589e+00
1.61193132e-01 1.26674175e+00 1.27186441e+00 -3.74291122e-01
-8.16875249e-02 2.75207847e-01 -2.87808836e-01 5.40804923e-01
-2.28288442e-01 2.83425987e-01 -1.31806397e+00 1.63053468e-01
2.81782329e-01 -1.26495004e-01 -4.44731444e-01 -4.59241450e-01
-9.73814785e-01 7.65411198e-01 -2.18849182e-02 -8.25345609e-03
-4.50702995e-01 -7.65133649e-02 6.44870818e-01 7.09151849e-02
7.91597307e-01 -5.73764026e-01 -5.58941543e-01 -3.02729845e-01
-1.22758293e+00 -2.62547553e-01 5.86346507e-01 6.93454981e-01
1.63797036e-01 7.60424510e-02 2.68228110e-02 1.37083662e+00
9.09778118e-01 8.82270873e-01 8.06849778e-01 -6.24795973e-01
5.77307701e-01 1.53876632e-01 -2.58098423e-01 -3.61683071e-01
-1.70001477e-01 7.50264600e-02 -5.37890494e-01 5.09801090e-01
2.75559187e-01 -3.82627062e-02 -1.10571659e+00 1.70806718e+00
1.64234996e-01 4.80398834e-02 3.74112666e-01 6.98396921e-01
1.24723768e+00 5.92171073e-01 -1.25037760e-01 -4.10755783e-01
1.43937266e+00 -1.22960305e+00 -8.89493227e-01 -4.53118831e-01
-1.65753886e-02 -1.08243358e+00 8.77725244e-01 5.32379508e-01
-1.55685186e+00 -6.68069422e-01 -1.15037179e+00 -1.56709790e-01
-4.62485403e-01 4.24009115e-01 9.75601599e-02 7.35488057e-01
-1.48947442e+00 1.11732166e-02 -6.57468200e-01 -4.51586276e-01
4.16413844e-01 4.74800617e-01 -6.15380764e-01 -1.31710961e-01
-9.95814562e-01 9.42603111e-01 -4.46928829e-01 -1.03734314e-01
-8.78562868e-01 -4.88936126e-01 -1.05364192e+00 -2.96252161e-01
2.09681883e-01 -6.00826323e-01 1.41268933e+00 -7.50209093e-01
-1.94357240e+00 1.18009257e+00 -9.74212050e-01 -3.48713070e-01
5.46437979e-01 -1.72550231e-01 -4.37768549e-01 4.07026023e-01
-9.06113014e-02 9.93569493e-01 1.25875926e+00 -1.37649083e+00
-5.98961473e-01 -6.51466072e-01 -3.91467571e-01 3.71838182e-01
-1.97493017e-01 1.49824008e-01 -7.72577584e-01 -3.24624270e-01
-5.36475405e-02 -7.77465403e-01 7.29999781e-01 2.89078504e-01
-2.97107059e-03 -3.17413509e-01 8.84337664e-01 -1.00612509e+00
7.95312822e-01 -2.11968350e+00 6.25964254e-02 -7.02562392e-01
2.72036064e-02 5.26050091e-01 2.07488444e-02 4.10018831e-01
2.26179156e-02 7.15790316e-02 9.17520076e-02 -1.24817026e+00
-1.33094952e-01 -3.07055153e-02 -1.77325737e-02 6.79142654e-01
9.22001153e-03 8.87095273e-01 -4.58600789e-01 -4.92243677e-01
4.93347734e-01 9.02726531e-01 -3.26319307e-01 3.38525563e-01
5.90575263e-02 -7.15397298e-02 3.62711698e-01 8.36776316e-01
7.95088828e-01 -6.42964765e-02 -4.08122577e-02 -3.19386423e-01
2.94437353e-02 3.28239202e-01 -7.16257930e-01 1.81808472e+00
-8.08015823e-01 1.12300479e+00 5.10930061e-01 -5.73243320e-01
6.49129272e-01 9.13902044e-01 3.34568769e-01 -8.27061296e-01
-5.51945381e-02 2.68572867e-01 -2.74291515e-01 -8.17472935e-01
1.77979872e-01 -4.32473809e-01 5.44911146e-01 3.69274110e-01
4.03849006e-01 1.71668865e-02 -1.37334391e-01 -1.84884846e-01
1.94315344e-01 6.10837489e-02 1.71016768e-01 3.02921623e-01
9.09891307e-01 -9.38671708e-01 1.11834973e-01 1.22216560e-01
-5.90398014e-01 8.26212883e-01 3.10474098e-01 1.13371603e-01
-9.50085044e-01 -1.29041696e+00 -4.64485809e-02 1.08895838e+00
5.18997274e-02 -1.69087395e-01 -6.67886376e-01 -6.13528550e-01
-2.89642643e-02 4.37013537e-01 -5.43717861e-01 -5.50661720e-02
-2.25897789e-01 -2.77503461e-01 6.02499485e-01 5.86933970e-01
6.61525011e-01 -1.24990308e+00 -5.35145462e-01 -1.61399215e-01
-4.88806516e-01 -1.34978950e+00 -8.03711414e-01 -2.24746451e-01
-7.39912927e-01 -9.14012969e-01 -1.33875358e+00 -8.00781071e-01
1.23465806e-01 4.44809675e-01 7.76594222e-01 -3.00096124e-01
-4.33062986e-02 5.16183734e-01 3.10055185e-02 -4.64633465e-01
-5.23288667e-01 -1.71241716e-01 3.11508924e-01 3.51455271e-01
3.04382682e-01 -3.75048816e-01 -7.61894047e-01 3.68275106e-01
-5.37534595e-01 7.11870417e-02 7.81203091e-01 1.12646091e+00
3.65023196e-01 -8.72663677e-01 7.85682917e-01 1.81399345e-01
5.05402625e-01 -2.37134874e-01 -2.42503747e-01 1.76719919e-01
-7.67438054e-01 -1.63763717e-01 1.34118453e-01 -5.31893492e-01
-1.11869884e+00 -2.41028190e-01 -3.96160454e-01 -7.35263228e-01
-3.49781007e-01 -1.07919062e-02 -2.87978768e-01 4.16382737e-02
2.49921624e-02 6.83593214e-01 8.17234039e-01 -5.93294859e-01
3.18081677e-01 1.29385352e+00 2.56774664e-01 1.78683534e-01
-1.95928201e-01 6.00194752e-01 -3.52442116e-01 -1.16332781e+00
-2.59694427e-01 -7.39722133e-01 -4.91210788e-01 -5.15772402e-01
9.34809387e-01 -1.00311542e+00 -1.14135981e+00 1.19241214e+00
-1.09490728e+00 -3.26777250e-01 1.86526388e-01 5.36238849e-01
-8.68778229e-01 4.66613322e-01 -4.89154726e-01 -1.04290187e+00
-4.37332302e-01 -1.57501948e+00 1.47048593e+00 1.56039149e-01
2.48907506e-02 -9.58214760e-01 -1.46170557e-01 9.07055974e-01
3.60672683e-01 -5.55836141e-01 4.19613689e-01 -4.36524838e-01
-1.53432980e-01 7.44120823e-03 -1.49947360e-01 7.72474706e-01
2.16698065e-01 -7.63930231e-02 -1.70827222e+00 -5.30600369e-01
3.45320478e-02 -4.81438935e-01 1.19335711e+00 7.05223858e-01
4.93112326e-01 -1.60454154e-01 -4.60369796e-01 3.58031452e-01
1.07247555e+00 2.45952383e-01 4.57942545e-01 -1.65579408e-01
1.80156142e-01 6.62277043e-01 3.33143860e-01 3.09718400e-01
6.17285550e-01 1.00213158e+00 5.74310482e-01 -3.00842285e-01
-8.75646949e-01 -3.39541614e-01 9.80820894e-01 7.23883450e-01
1.92425400e-01 -4.49062258e-01 -6.86712921e-01 8.05865347e-01
-1.43072629e+00 -1.01877260e+00 4.44221705e-01 2.25937533e+00
5.72124660e-01 -4.56189662e-02 3.45430046e-01 3.05620760e-01
7.13299155e-01 5.49593806e-01 -7.05556691e-01 -4.80256349e-01
-1.60280645e-01 -2.50527859e-02 2.35143572e-01 7.65044689e-01
-1.04601133e+00 8.48709404e-01 6.30374002e+00 6.72737598e-01
-1.62267721e+00 2.22300112e-01 5.92455328e-01 -5.46398222e-01
2.31532037e-01 -5.83960235e-01 -1.02101040e+00 4.00539935e-01
1.05840063e+00 3.34960967e-01 2.24750146e-01 5.03842175e-01
4.48617637e-01 -3.95691842e-01 -9.19716656e-01 1.39889789e+00
9.16221976e-01 -1.08974469e+00 -1.03633534e-02 4.23011035e-01
3.41020614e-01 4.21029866e-01 5.68116724e-01 -1.49230406e-01
-4.66643542e-01 -9.95239139e-01 7.20671833e-01 5.68118453e-01
1.15919960e+00 -6.33766413e-01 6.17881119e-01 3.22504878e-01
-1.16356719e+00 -1.15421057e-01 1.79146469e-01 5.53220451e-01
4.55064774e-01 -7.33172102e-03 -8.71382177e-01 2.39084661e-01
6.83245957e-01 8.05829883e-01 -3.18650901e-01 7.85250723e-01
-2.11585071e-02 4.21632856e-01 -1.41469046e-01 -5.61771020e-02
6.54688999e-02 3.51438463e-01 6.34812891e-01 1.09681690e+00
1.55370802e-01 -3.31263006e-01 -3.40021402e-01 4.07561302e-01
-5.29074110e-02 2.70198360e-02 -8.14402103e-01 2.28228971e-01
1.01197384e-01 9.63499546e-01 -1.43189228e-03 -3.62138599e-01
-8.01731110e-01 1.03453803e+00 -8.92177969e-02 3.72730583e-01
-6.46658003e-01 -4.16664593e-02 1.05477357e+00 2.26068482e-01
6.20639443e-01 -1.37586847e-01 -1.13567002e-01 -1.09522164e+00
2.23114923e-01 -9.08150971e-01 1.41903996e-01 -1.07248628e+00
-1.06615663e+00 5.60394824e-01 -1.99431673e-01 -1.03432155e+00
-5.43436527e-01 -7.87907302e-01 -3.00884336e-01 1.03654194e+00
-1.89146030e+00 -1.34253228e+00 -1.72399089e-01 7.46697962e-01
1.15396667e+00 -4.49012011e-01 8.28768492e-01 1.43342435e-01
-1.91806376e-01 9.43983495e-01 4.51390184e-02 -2.36822367e-02
1.03501701e+00 -8.44639361e-01 3.01673442e-01 5.67494512e-01
6.98973238e-02 3.15907538e-01 3.36432546e-01 -3.31048369e-01
-1.43790960e+00 -4.66124773e-01 1.12600994e+00 -1.63405344e-01
2.51779467e-01 -4.39932108e-01 -6.04986548e-01 2.67072141e-01
7.10930705e-01 1.47307888e-01 7.96989918e-01 -2.82523036e-01
-3.93016338e-01 -2.70456433e-01 -1.20005631e+00 5.23263454e-01
7.95658171e-01 -9.29492593e-01 -4.87395018e-01 -4.81505953e-02
4.02060896e-01 -2.42379472e-01 -7.77239382e-01 4.26907986e-01
1.00578415e+00 -1.29265916e+00 1.09211373e+00 -2.52521276e-01
1.78582698e-01 1.40832856e-01 -3.61970812e-01 -1.28351843e+00
4.97320116e-01 -5.12132525e-01 -3.56432170e-01 1.13721776e+00
4.42411810e-01 -9.03851449e-01 6.96132302e-01 -1.24040112e-01
2.08276194e-02 -1.10406768e+00 -1.33635724e+00 -5.88136256e-01
4.04706709e-02 -4.62282300e-01 1.40580058e-01 1.11407302e-01
2.05431268e-01 3.45329106e-01 -4.88848150e-01 -1.05191350e-01
6.09474957e-01 2.62421109e-02 6.74566865e-01 -9.46907818e-01
-1.02496207e-01 -5.75143397e-01 -2.76878119e-01 -1.46575522e+00
3.77085567e-01 -7.33493447e-01 1.04181267e-01 -1.63241363e+00
2.14659482e-01 2.77474761e-01 1.20059080e-01 3.31474364e-01
-2.58727614e-02 2.22317606e-01 4.30333525e-01 1.23710401e-01
-1.18200339e-01 6.90934062e-01 1.18886971e+00 -2.61922330e-01
-6.57019615e-02 2.78657407e-01 -3.36326689e-01 6.61353409e-01
6.90947056e-01 1.13349155e-01 -4.41226572e-01 -2.69648582e-01
-4.70546186e-01 4.33956265e-01 4.47365820e-01 -7.38107741e-01
4.87840205e-01 3.50368381e-01 3.90252382e-01 -8.54723811e-01
1.34434271e+00 -3.25346887e-01 -4.99349266e-01 2.32527733e-01
-1.46413416e-01 6.43619299e-02 4.99103457e-01 5.36236465e-01
-4.72321123e-01 1.71223596e-01 1.13719833e+00 -4.03486602e-02
-5.95249593e-01 3.20639849e-01 -6.30861402e-01 3.52140679e-03
8.19361329e-01 -4.41981733e-01 -2.21002370e-01 -8.02672207e-01
-1.01949465e+00 -5.67626432e-02 3.29976857e-01 7.15501130e-01
1.08540845e+00 -1.12014568e+00 -6.47335708e-01 6.15291297e-01
6.12013228e-02 -4.86278057e-01 3.37284327e-01 1.09104359e+00
2.11770445e-01 7.23153234e-01 -1.38810158e-01 -9.61002350e-01
-1.93895614e+00 5.40147007e-01 6.00118399e-01 4.17105556e-02
-5.80676317e-01 7.91775167e-01 1.31808087e-01 -1.04605362e-01
4.58161503e-01 -4.60692572e-05 -4.01195109e-01 6.81395113e-01
4.99637276e-01 4.18518811e-01 2.31670752e-01 -1.13667977e+00
-4.58387196e-01 9.15905178e-01 4.21154266e-03 -7.38950133e-01
1.10501707e+00 -4.53483820e-01 3.62659216e-01 6.29545033e-01
1.53001881e+00 1.56414196e-01 -1.30861008e+00 -8.35592151e-02
-5.86556971e-01 -2.88224190e-01 1.45170629e-01 -1.01561129e+00
-1.22837651e+00 1.66421461e+00 9.50428665e-01 -1.18451804e-01
1.08065486e+00 2.05638900e-01 8.04178834e-01 -2.24565268e-01
-3.47777754e-02 -9.45496798e-01 3.90156776e-01 3.01965714e-01
1.18332720e+00 -1.47120202e+00 -4.46572006e-01 -1.27959803e-01
-8.44262123e-01 1.07135534e+00 4.42990601e-01 5.22594213e-01
7.64961362e-01 3.26709718e-01 3.91374350e-01 1.33600861e-01
-1.04900968e+00 -3.19378436e-01 3.16822976e-01 7.49359787e-01
4.12609756e-01 -1.31116495e-01 1.73029825e-01 2.66087621e-01
1.74461782e-01 9.41644758e-02 9.99379307e-02 4.71511245e-01
-2.15128571e-01 -8.82646561e-01 -2.96724856e-01 1.44115031e-01
-6.09016180e-01 -3.21071446e-01 -2.55335420e-01 7.41166592e-01
4.38709781e-02 1.31500006e+00 2.48364676e-02 -2.09995493e-01
3.00127149e-01 6.19764864e-01 4.68566895e-01 -1.47692829e-01
-2.66130149e-01 3.88273269e-01 8.48121494e-02 -5.94649136e-01
-4.36490178e-01 -9.57636952e-01 -8.80335093e-01 -2.33752415e-01
-3.45878124e-01 -4.07382458e-01 1.02146256e+00 6.92009330e-01
5.23454189e-01 1.90498888e-01 5.72347045e-01 -1.26522458e+00
-5.88437259e-01 -1.24527216e+00 -3.57231647e-01 6.23309277e-02
9.81794119e-01 -9.51692104e-01 -6.85922205e-01 4.29902598e-02] | [14.326679229736328, 5.009440898895264] |
18945d10-4a4d-4267-a9ce-6d8a0c33979a | feature-fusion-vision-transformer-fine | 2107.02341 | null | https://arxiv.org/abs/2107.02341v3 | https://arxiv.org/pdf/2107.02341v3.pdf | Feature Fusion Vision Transformer for Fine-Grained Visual Categorization | The core for tackling the fine-grained visual categorization (FGVC) is to learn subtle yet discriminative features. Most previous works achieve this by explicitly selecting the discriminative parts or integrating the attention mechanism via CNN-based approaches.However, these methods enhance the computational complexity and make the modeldominated by the regions containing the most of the objects. Recently, vision trans-former (ViT) has achieved SOTA performance on general image recognition tasks. Theself-attention mechanism aggregates and weights the information from all patches to the classification token, making it perfectly suitable for FGVC. Nonetheless, the classifi-cation token in the deep layer pays more attention to the global information, lacking the local and low-level features that are essential for FGVC. In this work, we proposea novel pure transformer-based framework Feature Fusion Vision Transformer (FFVT)where we aggregate the important tokens from each transformer layer to compensate thelocal, low-level and middle-level information. We design a novel token selection mod-ule called mutual attention weight selection (MAWS) to guide the network effectively and efficiently towards selecting discriminative tokens without introducing extra param-eters. We verify the effectiveness of FFVT on three benchmarks where FFVT achieves the state-of-the-art performance. | ['Yongsheng Gao', 'Xiaohan Yu', 'Jun Wang'] | 2021-07-06 | null | null | null | null | ['fine-grained-visual-categorization'] | ['computer-vision'] | [-1.33095145e-01 -3.93396914e-01 -2.37242997e-01 -3.52341652e-01
-6.63107753e-01 -1.47915736e-01 6.77284002e-01 4.48384993e-02
-5.84283113e-01 3.82553220e-01 1.97185263e-01 1.59985662e-01
-1.35417163e-01 -8.70764911e-01 -5.27419746e-01 -1.04611409e+00
3.82101566e-01 -2.86802109e-02 4.87128913e-01 -2.36897599e-02
3.21756423e-01 5.28935075e-01 -2.02048564e+00 6.43984497e-01
1.16002476e+00 1.58664942e+00 5.89228213e-01 3.19587201e-01
-4.88912463e-01 8.10265243e-01 -4.52464879e-01 -2.50921279e-01
1.84051991e-01 -2.11393818e-01 -6.11030996e-01 1.35710791e-01
7.14135647e-01 -2.32285872e-01 -2.89758950e-01 1.10540211e+00
4.87000108e-01 4.17782664e-02 6.33297324e-01 -1.22276533e+00
-6.62518799e-01 4.45150554e-01 -6.67545915e-01 2.16905326e-01
-2.24939510e-01 9.50644836e-02 1.13185883e+00 -1.19482303e+00
3.27368021e-01 1.18954265e+00 5.35004854e-01 2.72431821e-01
-1.00291276e+00 -6.16324782e-01 6.97643995e-01 6.81808233e-01
-1.44283915e+00 -3.76281798e-01 8.59502017e-01 -4.33849782e-01
9.04665649e-01 3.12407076e-01 5.67259312e-01 6.63029373e-01
2.96645731e-01 9.88770366e-01 1.01315272e+00 -2.37943754e-01
-6.64315075e-02 1.19739184e-02 4.28938538e-01 6.96858764e-01
1.27566949e-01 -1.47600204e-01 -4.19576645e-01 7.30009899e-02
6.10577106e-01 6.52934849e-01 -2.09827542e-01 -1.45134494e-01
-1.02068925e+00 5.87856352e-01 8.53311718e-01 6.20774627e-01
-6.63583755e-01 3.46840352e-01 3.48898172e-01 2.51047909e-01
2.12331295e-01 7.85584822e-02 -3.94239604e-01 2.93773085e-01
-8.14047873e-01 -4.56679845e-03 4.14024070e-02 7.90018678e-01
1.12180555e+00 2.29177959e-02 -8.00592363e-01 9.64742422e-01
4.16403502e-01 3.23160917e-01 6.78690910e-01 -7.01415896e-01
3.61949146e-01 9.04109657e-01 -1.40010998e-01 -1.04763377e+00
-1.53704202e-02 -7.46145129e-01 -1.15270174e+00 1.69419542e-01
7.13373795e-02 1.72049299e-01 -1.31125808e+00 1.49145615e+00
3.69838588e-02 -8.96087140e-02 -3.86432081e-01 1.01806486e+00
1.02165854e+00 5.36068738e-01 2.61541009e-01 7.11778030e-02
1.43480885e+00 -1.15811408e+00 -4.96683657e-01 -1.98800921e-01
3.71458828e-01 -8.36324096e-01 1.03206682e+00 2.31736705e-01
-8.69608521e-01 -9.86856759e-01 -9.83969331e-01 -2.25313172e-01
-3.99613440e-01 5.19800186e-01 4.19998467e-01 6.23819046e-02
-1.06786382e+00 4.91927624e-01 -5.70930481e-01 -1.55465752e-01
6.62084997e-01 3.83028835e-01 -4.19980317e-01 -2.60187119e-01
-8.65555882e-01 6.95065558e-01 2.43122876e-01 2.84815133e-01
-9.22841549e-01 -5.20308435e-01 -6.15963578e-01 3.13831925e-01
2.30643809e-01 -5.21261811e-01 8.55141521e-01 -1.18259180e+00
-1.03014100e+00 6.16689682e-01 -4.34494317e-01 -1.66101485e-01
3.30580682e-01 -6.89987689e-02 7.24871978e-02 -3.40577736e-02
1.62377134e-01 7.55149662e-01 1.00586927e+00 -1.22121024e+00
-9.20388341e-01 -4.31747258e-01 2.24746644e-01 8.74068215e-02
-5.56882381e-01 -1.34135157e-01 -6.14011884e-01 -9.10335243e-01
1.87630877e-01 -4.36072916e-01 -6.45741671e-02 8.65939483e-02
-4.33625728e-01 -7.40053654e-01 1.09023094e+00 -4.29783553e-01
1.19273865e+00 -2.29024124e+00 1.84590220e-01 1.12941093e-03
5.18018067e-01 4.94348228e-01 -2.89256603e-01 4.14267927e-01
1.03597596e-01 -8.66268128e-02 8.59829634e-02 -5.21097302e-01
8.07324871e-02 1.87975436e-01 -8.52000490e-02 2.64136702e-01
3.64285290e-01 1.01983833e+00 -6.20248318e-01 -7.09359050e-01
4.70442533e-01 4.33362931e-01 -5.80268800e-01 2.79335648e-01
2.13728901e-02 -1.16439506e-01 -6.52764976e-01 7.78442979e-01
7.95641363e-01 -1.20949760e-01 -1.93343297e-01 -8.17285955e-01
-3.31029564e-01 -3.89861390e-02 -9.26719666e-01 1.44537425e+00
-4.95204508e-01 3.49256486e-01 2.25089833e-01 -9.80310202e-01
9.28285956e-01 -6.66852444e-02 3.91551673e-01 -6.28268600e-01
3.54576975e-01 1.14142358e-01 -5.17697707e-02 -4.62999523e-01
3.65730226e-01 3.86676416e-02 8.07983950e-02 1.56930730e-01
3.59700531e-01 2.29297951e-01 -8.56477320e-02 9.89655182e-02
7.66287446e-01 -9.70022567e-03 3.02707911e-01 -4.18051183e-01
7.87132740e-01 -1.21228829e-01 7.64666796e-01 7.85140932e-01
-4.18599874e-01 6.23933554e-01 -1.30216209e-02 -3.71184111e-01
-6.29251122e-01 -8.02789271e-01 6.49011433e-02 1.34424436e+00
3.15408409e-01 -4.27873641e-01 -6.63582683e-01 -8.77062023e-01
1.25680149e-01 2.30375424e-01 -8.21200252e-01 -4.14828956e-01
-3.24433565e-01 -4.55265135e-01 2.48133004e-01 7.13011444e-01
1.01973569e+00 -1.05212092e+00 -5.30219615e-01 5.47450520e-02
-3.05658609e-01 -8.03206623e-01 -7.20530391e-01 3.09778482e-01
-4.28363740e-01 -8.76574874e-01 -7.01874375e-01 -9.91983891e-01
6.72445893e-01 7.16265321e-01 7.53938317e-01 2.49992549e-01
-2.90505022e-01 4.00754586e-02 -3.94590020e-01 -2.94340461e-01
3.73174965e-01 9.75085795e-02 -2.16298670e-01 4.40109164e-01
5.64849675e-01 -2.01409504e-01 -7.93293774e-01 3.57112706e-01
-5.77094555e-01 1.03338853e-01 8.34546387e-01 1.14249849e+00
8.27897072e-01 3.33711840e-02 4.64004427e-01 -3.76557767e-01
2.89804876e-01 -6.61582425e-02 -2.81795710e-01 4.13769037e-01
-3.65362287e-01 -3.86070018e-03 8.80783498e-01 -3.55087847e-01
-9.82175410e-01 -3.59670222e-02 -2.19639391e-01 -7.25883901e-01
-2.20547244e-01 1.76405758e-01 -5.26199162e-01 -4.97597367e-01
1.39700726e-01 4.27356482e-01 -3.81455660e-01 -6.36889458e-01
2.00970039e-01 7.58437335e-01 3.91196102e-01 -4.94005799e-01
6.26029730e-01 3.94411951e-01 -2.90162504e-01 -5.87058306e-01
-9.07676101e-01 -5.54306805e-01 -6.25312924e-01 -1.34014487e-01
9.54829752e-01 -8.30026865e-01 -7.97624528e-01 7.30427325e-01
-1.05033743e+00 3.35329920e-02 -3.90966356e-01 2.88607746e-01
-2.14384422e-01 2.66417027e-01 -4.32952106e-01 -6.25520825e-01
-4.51924503e-01 -1.37038767e+00 1.26089025e+00 2.77378082e-01
2.61767775e-01 -6.81061566e-01 -3.56870979e-01 1.08539470e-01
6.25428677e-01 -1.35837257e-01 9.25104260e-01 -3.35126191e-01
-7.51358688e-01 -2.92590875e-02 -8.78267109e-01 4.96416062e-01
2.12323159e-01 -9.69924554e-02 -1.28027129e+00 -1.55925468e-01
-1.16620332e-01 -2.97124565e-01 1.61127937e+00 4.53770816e-01
1.52028024e+00 -4.48337384e-02 -3.40318769e-01 7.00033665e-01
1.59921181e+00 1.55402958e-01 5.36405742e-01 1.74518734e-01
1.11636651e+00 5.49975753e-01 6.94694459e-01 3.36137682e-01
7.04482853e-01 7.07547009e-01 5.15398622e-01 -3.41846019e-01
-4.65237439e-01 -7.00728148e-02 3.36177558e-01 7.86905527e-01
-2.78902203e-01 -1.86431244e-01 -5.31896651e-01 7.01629877e-01
-2.06351399e+00 -9.39793289e-01 5.65690883e-02 1.98921633e+00
6.31335735e-01 -4.63509783e-02 -1.90828681e-01 2.17776984e-01
7.15792954e-01 2.33693510e-01 -4.52123165e-01 -2.77901679e-01
-1.85323238e-01 4.28134613e-02 3.61754090e-01 2.48622373e-01
-1.29484320e+00 1.05009222e+00 5.41350126e+00 1.41721320e+00
-1.31005657e+00 1.74947590e-01 5.78969181e-01 1.33024186e-01
-2.74736434e-01 -2.67138869e-01 -8.69717300e-01 3.32040429e-01
1.77999109e-01 1.46990582e-01 2.02160463e-01 7.10147262e-01
8.13364983e-02 -7.53000155e-02 -7.27948964e-01 1.15024006e+00
1.14912994e-01 -1.13947845e+00 3.14275980e-01 -6.84801042e-02
5.66360235e-01 5.99099621e-02 4.67294604e-02 3.42594624e-01
1.79157659e-01 -8.52529883e-01 8.38437378e-01 6.63300157e-01
8.30920219e-01 -7.53694475e-01 1.12852991e+00 1.89112395e-01
-1.62072277e+00 -3.65024090e-01 -6.27350092e-01 2.51115650e-01
-3.42116028e-01 7.12294996e-01 -3.79145205e-01 7.54251063e-01
9.03671503e-01 9.71057773e-01 -6.39017820e-01 1.04896533e+00
-4.14168313e-02 3.72337133e-01 -7.79566914e-02 1.64664268e-01
5.04556894e-01 2.49883849e-02 1.66535810e-01 1.23960567e+00
3.40848356e-01 -6.09164052e-02 5.26227474e-01 6.10362113e-01
-6.02077879e-02 1.42832138e-02 -2.50399441e-01 9.98236984e-02
2.72336006e-01 1.36689222e+00 -6.28480256e-01 -4.69426811e-01
-3.63615811e-01 1.02975130e+00 4.52357888e-01 3.40446085e-01
-5.95308661e-01 -7.91593492e-01 8.54594588e-01 -5.95807843e-02
7.73947954e-01 -1.24330735e-02 -3.30098927e-01 -1.27379560e+00
-3.31825204e-02 -6.55262887e-01 3.79968345e-01 -4.44208384e-01
-1.34336352e+00 8.46844137e-01 -2.01519504e-01 -1.16196942e+00
1.60300016e-01 -5.30123413e-01 -6.40713513e-01 9.41822410e-01
-1.73917830e+00 -1.55951822e+00 -6.86850727e-01 8.13734710e-01
7.72098839e-01 -7.94997737e-02 4.98865724e-01 4.38482821e-01
-6.68109536e-01 8.05038214e-01 -8.49082097e-02 9.33678299e-02
6.85881913e-01 -1.18624842e+00 2.40554027e-02 8.45517457e-01
-1.76935151e-01 6.24588609e-01 3.77617031e-01 -4.55043197e-01
-1.27846634e+00 -1.35922110e+00 9.84133422e-01 1.90415680e-02
3.24186236e-01 -3.52027267e-01 -8.47819149e-01 2.61164725e-01
1.84076235e-01 2.89984852e-01 1.85752839e-01 -4.98230234e-02
-5.49561381e-01 -4.86824751e-01 -1.16255116e+00 2.68076569e-01
1.06633461e+00 -6.32755637e-01 -5.79678237e-01 -4.95928712e-02
6.15734756e-01 2.76643693e-01 -6.49408281e-01 4.67278421e-01
6.31890416e-01 -1.04902649e+00 9.14206743e-01 -1.72330841e-01
3.34726900e-01 -6.62252903e-01 -4.04003084e-01 -1.24352860e+00
-8.07385623e-01 3.47798243e-02 -2.31294949e-02 1.42708826e+00
-1.31052434e-01 -7.03900456e-01 5.60441375e-01 1.52399642e-02
-3.63390476e-01 -9.10175741e-01 -7.44032681e-01 -5.23873687e-01
-1.73500821e-01 -3.83060187e-01 7.82910109e-01 7.28905559e-01
-3.99801254e-01 1.37374431e-01 -3.41848910e-01 -2.02569559e-01
6.78917170e-01 4.98476177e-01 3.79266262e-01 -1.21999383e+00
-1.04898825e-01 -6.26924872e-01 -6.78601921e-01 -1.13353169e+00
-7.93621242e-02 -8.64058971e-01 1.93654761e-01 -1.80453396e+00
4.40007329e-01 -4.99162346e-01 -8.31227243e-01 8.78115892e-01
-3.02765667e-01 4.92245793e-01 5.11125267e-01 2.34231845e-01
-7.46144891e-01 8.41134667e-01 1.48305798e+00 -5.07933557e-01
1.15128301e-01 -2.96782196e-01 -8.35614622e-01 6.45089388e-01
5.18774450e-01 -1.95543870e-01 -3.59100312e-01 -5.13193965e-01
-5.99479735e-01 -5.95298290e-01 5.62265635e-01 -9.41299498e-01
4.50604498e-01 -2.54397660e-01 7.29750574e-01 -9.13768411e-01
3.13419461e-01 -9.00843501e-01 -1.25914901e-01 4.37673688e-01
-1.30795240e-01 -1.42141491e-01 -2.63044331e-02 3.24121296e-01
-5.56064069e-01 1.35481551e-01 8.42971385e-01 -5.02572814e-03
-1.10043216e+00 6.64601445e-01 -3.55597824e-01 -2.67415524e-01
1.00835669e+00 -2.21580476e-01 -3.87118846e-01 -6.77516907e-02
-5.25590122e-01 2.86246210e-01 4.16503012e-01 3.66423815e-01
8.43111634e-01 -1.45268929e+00 -5.78735054e-01 4.47883427e-01
3.41039687e-01 -1.94055170e-01 5.25509775e-01 9.59627271e-01
-4.88771424e-02 5.68503261e-01 -3.78111452e-01 -6.47047579e-01
-1.31258416e+00 6.00399673e-01 5.02632380e-01 -1.26820073e-01
-4.46762294e-01 1.02425051e+00 8.46790671e-01 -1.57498285e-01
1.90284699e-01 -5.00309825e-01 -5.38318515e-01 2.40670338e-01
6.22331738e-01 2.75136888e-01 8.98438394e-02 -8.13716590e-01
-5.80052912e-01 9.47639585e-01 -3.61281961e-01 4.82277364e-01
1.15932155e+00 -2.24883124e-01 -1.82289153e-01 1.72163412e-01
1.27400482e+00 -1.45924836e-01 -1.29166138e+00 -4.20216143e-01
-2.83646047e-01 -5.96902966e-01 4.23111677e-01 -6.75600827e-01
-1.46830654e+00 1.09644854e+00 7.02648997e-01 -3.71941179e-02
1.41515601e+00 -1.30737469e-01 7.57583082e-01 8.01973268e-02
5.26698172e-01 -9.43177819e-01 -1.74841415e-02 4.26574856e-01
9.34800446e-01 -1.15303886e+00 -2.00617939e-01 -3.27960819e-01
-5.86189985e-01 9.57455993e-01 9.03445899e-01 -3.50760430e-01
8.63827765e-01 1.40590072e-01 1.30868377e-02 -1.41306758e-01
-8.53616238e-01 -6.11503601e-01 6.82806253e-01 4.91163045e-01
2.07359493e-01 9.31893587e-02 -2.08138436e-01 7.05123127e-01
3.17593366e-01 -2.64280085e-02 -1.61584139e-01 7.44156420e-01
-7.39597023e-01 -9.95446146e-01 -2.19789162e-01 7.68581092e-01
-3.86850178e-01 -1.39205232e-01 -4.17498618e-01 4.49845821e-01
8.08737278e-01 9.96935964e-01 6.98586330e-02 -8.05591226e-01
2.47718930e-01 -3.08875650e-01 4.22974616e-01 -3.61417741e-01
-9.05272007e-01 2.59364694e-01 -2.12426364e-01 -6.64984941e-01
-4.30876821e-01 -3.88169676e-01 -1.08223724e+00 -3.26382548e-01
-4.95521396e-01 1.86344072e-01 2.73575872e-01 1.07855916e+00
4.72356319e-01 8.12668383e-01 6.76002443e-01 -1.10586905e+00
-2.62927473e-01 -9.28750753e-01 -4.90160257e-01 2.10148633e-01
5.35522819e-01 -8.52836609e-01 -1.54110938e-01 -2.28360280e-01] | [9.646004676818848, 1.91808021068573] |
41753141-703e-446f-bb95-7153ff70ed5f | direction-of-arrival-estimation-for-multiple | 1710.10059 | null | http://arxiv.org/abs/1710.10059v2 | http://arxiv.org/pdf/1710.10059v2.pdf | Direction of arrival estimation for multiple sound sources using convolutional recurrent neural network | This paper proposes a deep neural network for estimating the directions of
arrival (DOA) of multiple sound sources. The proposed stacked convolutional and
recurrent neural network (DOAnet) generates a spatial pseudo-spectrum (SPS)
along with the DOA estimates in both azimuth and elevation. We avoid any
explicit feature extraction step by using the magnitudes and phases of the
spectrograms of all the channels as input to the network. The proposed DOAnet
is evaluated by estimating the DOAs of multiple concurrently present sources in
anechoic, matched and unmatched reverberant conditions. The results show that
the proposed DOAnet is capable of estimating the number of sources and their
respective DOAs with good precision and generate SPS with high signal-to-noise
ratio. | ['Tuomas Virtanen', 'Archontis Politis', 'Sharath Adavanne'] | 2017-10-27 | null | null | null | null | ['direction-of-arrival-estimation'] | ['audio'] | [-2.50281483e-01 -7.65590012e-01 1.10110438e+00 2.74946477e-04
-9.40111756e-01 -6.46220088e-01 3.80411893e-01 -1.45883545e-01
4.90416177e-02 5.23626387e-01 5.17392814e-01 -2.41074458e-01
-4.70274031e-01 -6.23367488e-01 -4.21101004e-01 -9.44225729e-01
-5.22089362e-01 -4.45827752e-01 -2.51710892e-01 -1.18669599e-01
-1.93433568e-01 6.43163383e-01 -1.69959772e+00 -5.85173769e-03
3.31933826e-01 1.36764276e+00 -7.25801662e-02 1.15674484e+00
3.39940518e-01 9.43305314e-01 -1.30855775e+00 -2.22512912e-02
4.79937881e-01 -4.73634273e-01 2.45407000e-01 -5.39088130e-01
6.43673956e-01 -3.54222596e-01 -4.82173741e-01 7.12734103e-01
1.20789230e+00 2.93204635e-01 4.75159883e-01 -8.38684499e-01
5.75038195e-02 4.85925466e-01 -3.08717757e-01 5.04119873e-01
1.87360495e-01 -1.41145721e-01 5.91014504e-01 -1.18426728e+00
-1.80621728e-01 7.86973119e-01 1.00343347e+00 -3.68549675e-01
-5.92975020e-01 -7.62523353e-01 -6.54027939e-01 -1.16125435e-01
-1.48277557e+00 -6.84229255e-01 9.51512754e-01 -1.19138584e-01
9.02844310e-01 2.45317653e-01 5.90942681e-01 8.72478187e-01
2.03599453e-01 9.94662270e-02 9.11431611e-01 -5.79573333e-01
2.56504774e-01 -3.46720964e-01 7.45440796e-02 4.59644854e-01
1.22350782e-01 3.80702347e-01 -6.46829426e-01 -3.26475590e-01
7.45219290e-01 -2.24299356e-01 -3.95887583e-01 1.04296483e-01
-1.41231465e+00 3.34140003e-01 4.73369926e-01 6.79486334e-01
-8.62271249e-01 5.97405851e-01 9.18224156e-02 2.43705824e-01
3.96978348e-01 6.37856185e-01 -2.97255486e-01 -6.18818961e-02
-9.33090031e-01 2.11237743e-01 8.34523618e-01 5.86459219e-01
4.24324185e-01 1.08973479e+00 1.41961187e-01 9.68200386e-01
4.35559690e-01 1.21967149e+00 2.26443619e-01 -7.40971386e-01
4.02283460e-01 -4.92744893e-01 5.69151461e-01 -1.56805098e+00
-7.16280162e-01 -1.34635139e+00 -9.46360469e-01 -1.49113193e-01
3.16121608e-01 -1.13901126e+00 -8.23902309e-01 1.60012603e+00
2.53815979e-01 5.72616100e-01 3.57780308e-01 9.10374045e-01
9.31398094e-01 1.09289765e+00 -2.81777263e-01 -2.98357129e-01
9.82937098e-01 -6.21921837e-01 -1.11901855e+00 -4.01918143e-01
7.41421953e-02 -1.20404446e+00 -1.13783970e-01 4.23771679e-01
-8.26848865e-01 -8.45512927e-01 -1.11298573e+00 5.83946705e-01
-9.92477462e-02 5.45263946e-01 4.55604017e-01 9.09187138e-01
-1.06370592e+00 7.19140843e-02 -5.84504843e-01 5.08733690e-01
-2.88641572e-01 -4.57401387e-02 -9.50171053e-02 2.98864722e-01
-1.27003050e+00 3.74613881e-01 -4.44333941e-01 8.46780539e-01
-8.89516175e-01 -8.53158176e-01 -7.96969771e-01 3.22107494e-01
-3.16452026e-01 -4.12923813e-01 1.10882115e+00 -7.25323796e-01
-1.25840712e+00 -1.84214756e-01 -2.16353595e-01 -4.82938826e-01
-1.21476697e-02 -4.83425468e-01 -1.26818919e+00 5.46489917e-02
1.48237988e-01 -2.12067306e-01 7.88178444e-01 -1.14488924e+00
-6.39515400e-01 3.57036330e-02 -4.06519651e-01 2.32823759e-01
1.41361907e-01 2.00969838e-02 2.42854595e-01 -8.03745031e-01
7.65829206e-01 -5.16170502e-01 -3.21954429e-01 -3.90934169e-01
-4.59160089e-01 3.87209088e-01 2.41999358e-01 -6.25839293e-01
9.60626960e-01 -2.34203291e+00 -5.11829257e-01 2.59416133e-01
1.73711270e-01 3.81818712e-01 -1.64283946e-01 6.98152006e-01
-3.06076437e-01 -4.55668032e-01 2.94997424e-01 -1.17130481e-01
-2.14157879e-01 -6.07812464e-01 -6.38685763e-01 4.60637212e-01
-3.31298672e-02 1.38230026e-01 -9.48525250e-01 3.08265448e-01
2.75881857e-01 9.19019103e-01 -2.19843403e-01 5.68321168e-01
4.51479703e-01 3.49393964e-01 -2.57376641e-01 2.57795334e-01
1.02522683e+00 1.26014680e-01 -1.55248831e-03 -3.15698922e-01
-3.26295346e-01 4.72786188e-01 -1.54151058e+00 1.00682330e+00
-1.18676114e+00 1.19645333e+00 2.83957303e-01 -5.74448705e-01
1.44157064e+00 7.23912835e-01 3.74056309e-01 -6.52916729e-01
1.98037952e-01 7.40957141e-01 2.59847015e-01 -4.74805951e-01
3.05870563e-01 -1.82798922e-01 6.90218806e-02 5.51703572e-01
3.19724500e-01 -6.97200140e-03 -3.11292708e-01 -3.59139711e-01
1.02082896e+00 -3.49549949e-01 4.13580298e-01 -9.48242769e-02
4.97730434e-01 -6.80764616e-01 6.14802063e-01 8.32748711e-01
8.11560601e-02 6.08816206e-01 2.76884511e-02 -6.95207357e-01
-8.27929914e-01 -1.26866758e+00 2.77354792e-02 7.50362098e-01
-2.28213444e-01 9.41067487e-02 -3.83221805e-01 5.01390360e-02
-2.26314738e-01 6.53762102e-01 -2.64174789e-01 3.50455552e-01
-7.79852152e-01 -6.39342308e-01 8.20192039e-01 2.32581317e-01
8.19213092e-01 -5.84711015e-01 -5.13229191e-01 5.00715435e-01
-3.02542388e-01 -1.18228245e+00 -5.82639091e-02 3.54901135e-01
-2.12529153e-01 -6.96336567e-01 -9.21586692e-01 -6.29349232e-01
2.32132807e-01 5.64548790e-01 6.89039648e-01 -6.35939777e-01
-5.57136275e-02 3.74064863e-01 -1.85130805e-01 -8.62250447e-01
-1.12408444e-01 -4.56220180e-01 3.39134008e-01 5.41863203e-01
-1.24734670e-01 -1.23176730e+00 -8.74211311e-01 4.02230769e-01
-2.53334165e-01 -2.61560619e-01 3.69787514e-01 4.11493301e-01
6.50393814e-02 2.14957491e-01 9.47810352e-01 -4.70344499e-02
6.61717772e-01 -6.98890448e-01 -8.06050420e-01 -2.51034319e-01
1.29854545e-01 -4.91932303e-01 7.93013573e-01 -9.27382931e-02
-1.31090784e+00 -2.96943426e-01 -4.80980515e-01 -3.01979870e-01
-3.04972082e-01 5.18583119e-01 1.53491899e-01 5.11052972e-03
8.82303894e-01 3.25639427e-01 -6.77890301e-01 -5.22821903e-01
-1.17434062e-01 7.69532323e-01 5.17076254e-01 9.39758420e-02
9.29168284e-01 2.54167497e-01 2.75741279e-01 -1.28007901e+00
-7.40195334e-01 -5.80746770e-01 -1.64363444e-01 -3.84467304e-01
3.92134577e-01 -1.21335185e+00 -4.19024497e-01 9.01198030e-01
-1.42113376e+00 1.35044560e-01 2.05610231e-01 1.28414023e+00
-5.83549589e-02 -5.56064434e-02 -3.90054464e-01 -1.31297553e+00
-6.19540513e-01 -8.11439335e-01 5.22296309e-01 2.10756615e-01
-3.05368602e-02 -7.63269186e-01 3.24252874e-01 7.14181224e-03
9.62516248e-01 4.10269946e-01 4.65452611e-01 -5.61179876e-01
-5.61558425e-01 -5.86107016e-01 1.22186569e-02 3.28474611e-01
4.70285654e-01 -3.76856148e-01 -1.36195648e+00 1.25583708e-01
3.81897956e-01 1.32147074e-01 2.78434545e-01 8.93164635e-01
3.72296512e-01 -4.22518253e-01 1.94946438e-01 9.30032253e-01
1.72024393e+00 5.61912775e-01 6.19967699e-01 -7.05721602e-02
4.69053775e-01 2.13178173e-01 1.90214738e-01 7.07038820e-01
-1.33806258e-01 3.70961547e-01 3.82772952e-01 -8.00954252e-02
-2.87678897e-01 -2.29817554e-02 -6.69458807e-02 1.18557084e+00
-1.13179326e-01 -7.70053446e-01 -8.93208206e-01 8.20228517e-01
-1.09505641e+00 -8.80266368e-01 -4.74029481e-01 1.95908415e+00
2.89709538e-01 -3.74277890e-01 -2.22307548e-01 4.45263922e-01
5.50896049e-01 6.36679292e-01 -1.31821753e-02 -3.94935369e-01
-8.95692557e-02 5.53918898e-01 4.55110103e-01 4.80877191e-01
-1.06892359e+00 1.95270985e-01 6.39037752e+00 3.24755639e-01
-1.45727921e+00 -7.07095340e-02 3.62445205e-01 3.77259329e-02
-2.16499329e-01 -6.11734331e-01 -3.50685626e-01 1.80326730e-01
1.31092203e+00 2.26580761e-02 2.92726159e-01 5.97729087e-01
3.34065348e-01 -1.41134873e-01 -4.35702741e-01 8.79693031e-01
-1.65904880e-01 -1.26353025e+00 -3.52179170e-01 -3.82356972e-01
7.35327005e-01 3.28200996e-01 3.28000277e-01 -4.15506475e-02
2.14229912e-01 -7.94768274e-01 7.01016486e-01 8.08907330e-01
7.15371370e-01 -8.92547429e-01 1.01018119e+00 2.61367828e-01
-1.21657968e+00 -2.71972150e-01 -8.45558345e-02 -1.28263876e-01
1.77513853e-01 1.10789955e+00 -1.26826906e+00 6.79633915e-01
5.80375254e-01 2.17927441e-01 1.02670230e-01 1.44778895e+00
-1.80645943e-01 1.07367766e+00 -6.01268828e-01 -2.90733546e-01
3.53407860e-01 2.00381041e-01 7.59939432e-01 1.24781346e+00
1.05555081e+00 8.68364722e-02 -4.59508955e-01 4.75990981e-01
-4.62316396e-03 8.64109490e-03 -5.72976708e-01 1.24196798e-01
8.17088783e-01 9.55481410e-01 -3.45274180e-01 9.61987004e-02
4.26496193e-02 1.82868302e-01 -6.01861119e-01 9.12159443e-01
-6.17468297e-01 -9.01902735e-01 6.15999579e-01 -7.79986382e-02
5.48812449e-01 -5.30598283e-01 -7.13434666e-02 -5.51383018e-01
-2.29029227e-02 -8.02243233e-01 -2.14734972e-01 -1.02905250e+00
-9.48609591e-01 9.64464784e-01 -4.11394864e-01 -1.73747003e+00
-5.75309932e-01 -2.36401796e-01 -7.86671340e-01 1.34539700e+00
-1.37663913e+00 -7.71297038e-01 -2.70938426e-01 2.63194799e-01
8.92348588e-02 -4.61373717e-01 9.88863289e-01 5.07077694e-01
-1.82691261e-01 2.72600055e-01 3.16485971e-01 4.31849658e-01
3.73933256e-01 -9.89445508e-01 5.05074918e-01 9.80298638e-01
7.63482377e-02 7.99389958e-01 1.10330737e+00 -1.56478390e-01
-1.16640615e+00 -1.01081502e+00 1.00324249e+00 1.93786353e-01
4.02689874e-01 -2.46745542e-01 -3.46823543e-01 3.39328855e-01
2.57095248e-01 5.87382078e-01 7.63306379e-01 -1.40884653e-01
-3.05028111e-01 -5.70984066e-01 -7.47924924e-01 1.40464991e-01
3.35570991e-01 -3.82352650e-01 -2.06742585e-01 1.77287176e-01
4.87800747e-01 -5.62646627e-01 -5.66466868e-01 2.52688140e-01
6.36284649e-01 -1.44905126e+00 1.17727745e+00 4.70898151e-02
2.90519208e-01 -4.04930025e-01 -5.37240744e-01 -1.62348175e+00
-2.44624361e-01 -7.96089113e-01 3.41769904e-02 1.09534764e+00
3.88299435e-01 -9.63723183e-01 9.75481346e-02 -4.61782604e-01
-2.17826918e-01 -5.85055590e-01 -1.10318172e+00 -5.52979112e-01
-5.03029168e-01 -8.02577078e-01 8.32166195e-01 7.71299303e-01
-5.77564538e-01 2.71028519e-01 -7.82656550e-01 1.10728002e+00
7.57697642e-01 3.55083682e-02 7.04190969e-01 -9.18782949e-01
-2.86261827e-01 1.07637241e-01 -3.28387588e-01 -9.95436847e-01
-4.28707540e-01 -3.14868689e-01 3.14018577e-01 -1.43623114e+00
-1.04377615e+00 -5.38345933e-01 -5.52511692e-01 -1.08088315e-01
1.73827469e-01 2.91234523e-01 -1.27530709e-01 -1.65215999e-01
-1.63204055e-02 3.63232821e-01 7.81639934e-01 1.91944316e-01
-2.00250342e-01 7.36187339e-01 -1.57276332e-01 8.56200397e-01
6.34070396e-01 -5.64926684e-01 -3.06900561e-01 -7.77304530e-01
5.28361380e-01 7.19548404e-01 4.04426336e-01 -1.67733181e+00
3.06925416e-01 3.73793304e-01 6.74729705e-01 -7.61531651e-01
5.06900191e-01 -6.70137048e-01 3.63052756e-01 2.06770390e-01
-3.19955915e-01 -1.36687741e-01 2.57837147e-01 6.34585023e-01
-5.06347716e-01 1.71628788e-01 6.42866254e-01 4.21279818e-02
-2.33668730e-01 -1.12901200e-02 -7.14357674e-01 -2.93938547e-01
4.33817446e-01 1.79742321e-01 -2.69221485e-01 -1.06896734e+00
-3.39326382e-01 -3.44238698e-01 -8.58275354e-01 -6.39558583e-02
7.92257071e-01 -1.33064604e+00 -8.10383797e-01 4.21049923e-01
-3.66368055e-01 -3.34423959e-01 8.46883953e-01 5.76872110e-01
-7.19126284e-01 7.07691431e-01 -1.90272465e-01 -3.87956679e-01
-8.72051537e-01 -1.71762079e-01 1.19407630e+00 7.97423199e-02
-1.63750321e-01 1.15799487e+00 5.09231351e-02 -4.99202549e-01
2.28798240e-01 -1.28197670e-01 -5.51682591e-01 6.01788377e-03
7.13568330e-01 6.93662167e-01 1.43565029e-01 -5.61692774e-01
-3.01277667e-01 5.96274197e-01 5.16545713e-01 -4.41313148e-01
1.47431839e+00 -1.43678755e-01 -8.88905525e-02 4.38432485e-01
1.52905571e+00 9.23173368e-01 -8.49397957e-01 -2.90274411e-01
-5.84111154e-01 -3.47091138e-01 4.08186793e-01 -1.04128993e+00
-1.10408473e+00 1.09229672e+00 9.94652033e-01 3.18037242e-01
1.32710624e+00 -5.25433600e-01 8.57346654e-01 6.11938417e-01
3.76482904e-01 -5.39362609e-01 -6.52552843e-02 7.04608917e-01
7.79217243e-01 -5.39049804e-01 -1.97382912e-01 3.54164951e-02
-5.26265278e-02 1.38434434e+00 7.34330118e-02 -1.90602899e-01
8.98801506e-01 4.99629945e-01 9.27559674e-01 -1.98023602e-01
-4.93020356e-01 1.76969022e-02 2.03575328e-01 7.36268163e-01
5.89790702e-01 -1.12475427e-02 2.92218715e-01 4.32118326e-01
-5.67987859e-01 -4.48788166e-01 6.06861115e-01 5.07422626e-01
-4.74973202e-01 -2.99600244e-01 -8.26670110e-01 1.58978581e-01
-7.06274271e-01 -2.07761809e-01 1.96355402e-01 3.48653376e-01
1.71155870e-01 1.24813259e+00 1.70701370e-01 -5.08220971e-01
4.61399049e-01 -1.23361573e-02 -1.41303852e-01 -1.52295187e-01
-6.58108115e-01 6.04533792e-01 5.18325865e-01 -2.36083493e-01
-2.65019655e-01 -3.80526751e-01 -9.18662071e-01 1.16675355e-01
-2.55798757e-01 3.16331238e-01 8.24853301e-01 6.82324886e-01
3.48199457e-01 1.04918635e+00 1.49221075e+00 -7.55028844e-01
-2.48296246e-01 -1.07538426e+00 -8.13929677e-01 -4.04312938e-01
1.23751736e+00 -1.40162930e-01 -6.82186127e-01 -3.18247616e-01] | [15.258953094482422, 5.604907989501953] |
5be65edf-0895-4360-9be4-6fb0bc24dead | espnet-se-speech-enhancement-for-robust | 2207.09514 | null | https://arxiv.org/abs/2207.09514v1 | https://arxiv.org/pdf/2207.09514v1.pdf | ESPnet-SE++: Speech Enhancement for Robust Speech Recognition, Translation, and Understanding | This paper presents recent progress on integrating speech separation and enhancement (SSE) into the ESPnet toolkit. Compared with the previous ESPnet-SE work, numerous features have been added, including recent state-of-the-art speech enhancement models with their respective training and evaluation recipes. Importantly, a new interface has been designed to flexibly combine speech enhancement front-ends with other tasks, including automatic speech recognition (ASR), speech translation (ST), and spoken language understanding (SLU). To showcase such integration, we performed experiments on carefully designed synthetic datasets for noisy-reverberant multi-channel ST and SLU tasks, which can be used as benchmark corpora for future research. In addition to these new tasks, we also use CHiME-4 and WSJ0-2Mix to benchmark multi- and single-channel SE approaches. Results show that the integration of SE front-ends with back-end tasks is a promising research direction even for tasks besides ASR, especially in the multi-channel scenario. The code is available online at https://github.com/ESPnet/ESPnet. The multi-channel ST and SLU datasets, which are another contribution of this work, are released on HuggingFace. | ['Shinji Watanabe', 'Yanmin Qian', 'Yu Tsao', 'Zhong-Qiu Wang', 'Robin Scheibler', 'Brian Yan', 'Yoshiki Masuyama', 'Zhaoheng Ni', 'Samuele Cornell', 'Wangyou Zhang', 'Chenda Li', 'Xuankai Chang', 'Yen-Ju Lu'] | 2022-07-19 | null | null | null | null | ['spoken-language-understanding', 'robust-speech-recognition', 'speech-separation', 'spoken-language-understanding'] | ['natural-language-processing', 'speech', 'speech', 'speech'] | [ 2.30629534e-01 1.13833219e-01 3.15202683e-01 -4.06181246e-01
-1.28160310e+00 -4.34273928e-01 6.87018692e-01 -3.03870112e-01
-5.98639190e-01 5.32279193e-01 4.09741610e-01 -5.17229021e-01
2.21545711e-01 -7.21034184e-02 -5.01112163e-01 -6.32856190e-01
6.08608797e-02 1.65579364e-01 1.96928456e-01 -5.71806550e-01
-3.48113775e-01 2.62766242e-01 -1.48540294e+00 4.87987608e-01
7.09353089e-01 7.99845219e-01 5.31651914e-01 1.05309260e+00
1.90948024e-01 4.30049390e-01 -5.90273142e-01 -3.33845079e-01
4.07154337e-02 -5.01655161e-01 -4.89209414e-01 -2.45008450e-02
2.50755340e-01 -1.38304159e-01 -4.14248437e-01 9.33788061e-01
1.24605060e+00 3.63126785e-01 1.66812167e-01 -9.30343032e-01
-3.80969733e-01 8.92064393e-01 -1.87710643e-01 1.91752747e-01
3.81625235e-01 2.05041766e-01 9.41712499e-01 -1.16189802e+00
3.67406577e-01 1.26751804e+00 7.14311123e-01 6.06325984e-01
-9.65607703e-01 -7.58146048e-01 -2.08844487e-02 3.26223642e-01
-1.10205317e+00 -1.35154557e+00 6.50917053e-01 1.14251137e-01
1.34279716e+00 5.67979753e-01 3.17688972e-01 1.60881543e+00
-4.30879653e-01 1.14307094e+00 1.20722163e+00 -6.28257394e-01
1.80295810e-01 2.34194249e-01 2.35831842e-01 1.31945282e-01
-4.18276608e-01 4.67917651e-01 -7.76836395e-01 2.10282549e-01
2.40015000e-01 -7.08771467e-01 -6.92671239e-01 4.51489836e-01
-1.23519123e+00 4.75521594e-01 1.75609477e-02 6.18045926e-01
-3.65179479e-01 -3.97584550e-02 5.27921200e-01 6.06757939e-01
8.44540298e-01 1.81184098e-01 -7.52631128e-01 -5.50459504e-01
-1.16237593e+00 -1.45221734e-02 7.86975026e-01 9.98278320e-01
3.52565914e-01 4.24276322e-01 -3.86498481e-01 1.38331115e+00
3.36436749e-01 6.21457815e-01 4.40420330e-01 -7.03824043e-01
6.42765939e-01 -4.48047936e-01 -1.68603390e-01 -1.51857302e-01
-3.21332425e-01 -8.30857694e-01 -8.33846867e-01 -4.37489804e-03
-1.14788366e-02 -5.31998158e-01 -9.72187996e-01 1.67774308e+00
1.44891113e-01 5.57662249e-01 3.68792623e-01 9.73829210e-01
1.02755404e+00 1.03378272e+00 -1.09740637e-01 -1.92791358e-01
1.46962333e+00 -1.27863336e+00 -1.16864598e+00 -2.89707690e-01
4.26213473e-01 -1.14738667e+00 9.74713564e-01 4.80607361e-01
-1.40676677e+00 -7.50601411e-01 -8.74279559e-01 3.18839215e-02
-3.13351661e-01 4.04326946e-01 2.12122262e-01 8.78747523e-01
-1.54569411e+00 3.88174236e-01 -8.87489557e-01 -3.98053199e-01
-1.09448232e-01 5.13103679e-02 -3.18513095e-01 1.14178762e-01
-1.58878434e+00 7.02910185e-01 3.21643233e-01 2.21436739e-01
-7.86050439e-01 -6.01996899e-01 -8.67104471e-01 1.54838368e-01
3.78250927e-01 -4.04626906e-01 1.75805426e+00 -8.05528581e-01
-1.97873902e+00 6.20224774e-01 -2.91160494e-01 -4.68153894e-01
3.89796793e-01 -3.48849922e-01 -1.04911828e+00 -2.19135787e-02
-3.60037386e-01 4.45549697e-01 8.60573292e-01 -1.00659823e+00
-5.02218664e-01 -1.94845706e-01 -5.13994455e-01 3.86347175e-01
-2.32459396e-01 6.72129571e-01 -5.78434885e-01 -1.04577982e+00
-6.98538348e-02 -7.04610348e-01 9.64169763e-03 -6.09862924e-01
-5.59177339e-01 9.66416746e-02 8.40556264e-01 -1.18469238e+00
1.21128631e+00 -2.52104163e+00 1.86345577e-01 -5.42903040e-03
-2.76822597e-01 7.74598718e-01 -4.36686099e-01 5.95890641e-01
-2.90054023e-01 -1.27778858e-01 -4.08203423e-01 -1.01330411e+00
2.07564101e-01 -8.81280378e-02 -1.39885589e-01 1.78365752e-01
1.65172741e-01 7.27318406e-01 -5.11934876e-01 4.37244549e-02
4.48168039e-01 8.85177195e-01 -2.51136154e-01 2.16040716e-01
5.31081185e-02 6.27347171e-01 6.68371245e-02 3.00939023e-01
8.66694510e-01 3.45466793e-01 -1.19130373e-01 -5.41738570e-02
-3.03285450e-01 7.82684565e-01 -1.43130291e+00 1.78984571e+00
-8.37021768e-01 8.26221883e-01 7.21729159e-01 -6.76614046e-01
8.35657179e-01 9.72061336e-01 3.70174721e-02 -7.24807203e-01
6.52475879e-02 4.85665947e-01 7.57822394e-02 -2.87585080e-01
6.70854867e-01 -1.65821180e-01 3.45170707e-01 3.53063762e-01
5.63961565e-01 -1.76310271e-01 1.04674742e-01 2.11272374e-01
1.00161111e+00 -1.90892652e-01 2.17909932e-01 -2.16359715e-03
7.37851858e-01 -5.72799146e-01 1.08441800e-01 6.08852208e-01
-2.61068881e-01 8.78235221e-01 -3.78980092e-03 4.54910487e-01
-9.28738177e-01 -1.21413612e+00 -1.53539822e-01 1.29794300e+00
-4.26517338e-01 -5.45436919e-01 -9.16858912e-01 -2.72929668e-01
-5.68124592e-01 1.00290406e+00 -3.06018572e-02 2.01552525e-01
-4.17520672e-01 -6.90097392e-01 8.81658435e-01 4.70774740e-01
3.41395706e-01 -1.24831116e+00 2.19961420e-01 4.49542135e-01
-3.87486517e-01 -1.55134380e+00 -6.56399786e-01 3.94493014e-01
-3.75522912e-01 -5.02757907e-01 -1.10387230e+00 -7.57822573e-01
-5.97595833e-02 2.73906797e-01 8.44179273e-01 -2.41934717e-01
2.12827206e-01 4.70326692e-01 -7.41647780e-01 -3.37865412e-01
-9.09244776e-01 1.66520074e-01 5.22410758e-02 1.63299531e-01
4.75709252e-02 -5.06379068e-01 -3.23913872e-01 3.94349277e-01
-7.83937812e-01 9.77262110e-02 5.07255435e-01 8.17242742e-01
3.95915449e-01 -1.30264506e-01 8.34458768e-01 -5.34242511e-01
8.63875628e-01 -2.35853210e-01 -4.96896744e-01 1.26558021e-01
-1.55019641e-01 -3.07984650e-01 5.30038595e-01 -4.13624644e-01
-1.58947372e+00 -2.12523669e-01 -1.28725398e+00 -1.40515685e-01
-5.34631431e-01 4.18959051e-01 -4.86420870e-01 3.23684394e-01
5.20850599e-01 4.94166136e-01 -1.55298576e-01 -8.93495023e-01
5.22582352e-01 1.47332990e+00 6.76058471e-01 -2.69881487e-01
4.89384145e-01 3.34279845e-03 -7.41047263e-01 -1.37747884e+00
-4.14924800e-01 -7.90716469e-01 -1.99895233e-01 -4.06664982e-02
6.08610153e-01 -1.19259179e+00 -2.29718044e-01 8.94919813e-01
-1.35256970e+00 -6.41317248e-01 -1.60723194e-01 8.18038285e-01
-5.32722533e-01 3.38220000e-01 -8.89452636e-01 -9.79348183e-01
-5.27851045e-01 -1.38783956e+00 1.23506379e+00 -8.49562362e-02
-1.19559862e-01 -9.22835648e-01 1.00430809e-01 4.69016314e-01
8.16401005e-01 -6.91142797e-01 1.85632974e-01 -8.66424084e-01
-1.44465134e-01 6.15788549e-02 3.53672728e-02 9.62917089e-01
-1.23045154e-01 -1.60826653e-01 -1.70825386e+00 -4.40043092e-01
8.79056677e-02 -1.74533293e-01 9.61644053e-01 5.47925115e-01
7.04740167e-01 -4.30681668e-02 -1.41149024e-02 7.11364329e-01
7.78795540e-01 3.43030691e-01 8.76774609e-01 -2.60599647e-02
2.95773834e-01 4.21814740e-01 4.34607178e-01 2.75745630e-01
2.03092411e-01 1.04657495e+00 -1.07120425e-01 -4.19602662e-01
-7.71239936e-01 3.87333371e-02 8.38855267e-01 1.50724769e+00
9.45955515e-02 -6.21146679e-01 -8.32617581e-01 3.12531412e-01
-1.46970963e+00 -9.00108993e-01 -3.18950385e-01 1.99364102e+00
8.23181927e-01 -1.21039480e-01 1.62210003e-01 4.43222046e-01
7.82409191e-01 3.38528544e-01 -4.98787500e-02 -4.89674568e-01
-4.44119185e-01 5.77085912e-01 1.32197618e-01 8.80636096e-01
-1.05625963e+00 1.05877280e+00 5.47007656e+00 1.30373192e+00
-1.18847263e+00 7.25192785e-01 5.14514446e-01 2.61363443e-02
-9.87047106e-02 -4.66436952e-01 -8.07896078e-01 3.44402313e-01
1.48906183e+00 2.23240092e-01 6.83289111e-01 4.01688397e-01
5.63337505e-01 1.49976701e-01 -8.12690496e-01 9.20182943e-01
-9.21935737e-02 -8.67842972e-01 -5.52615106e-01 -3.00046116e-01
3.95097971e-01 7.38403261e-01 1.23250656e-01 4.38080341e-01
5.52854650e-02 -6.41345382e-01 8.87783706e-01 -3.23737264e-02
1.03520310e+00 -5.20468056e-01 7.19248235e-01 2.10289910e-01
-1.27149069e+00 2.45998561e-01 3.32636386e-02 2.10472435e-01
5.32670617e-01 6.07630432e-01 -8.53101313e-01 8.37156534e-01
6.30300522e-01 5.45431495e-01 -1.48936972e-01 9.99362767e-01
-6.90366387e-01 1.06923687e+00 -4.28670138e-01 2.54707366e-01
1.75671369e-01 6.21053204e-02 1.06941223e+00 1.82143497e+00
5.24033010e-01 -1.56806454e-01 -3.84492934e-01 6.26266301e-01
-7.55195618e-02 2.15463653e-01 -3.50270271e-01 -5.00012785e-02
4.44813788e-01 1.26988280e+00 -3.69834542e-01 -2.95473516e-01
-5.46604574e-01 1.21610951e+00 -1.01798140e-01 6.81705773e-01
-8.42459321e-01 -5.34950376e-01 8.55550051e-01 -1.82775572e-01
3.08085501e-01 -3.40690464e-01 4.93755452e-02 -1.07077837e+00
-2.96177454e-02 -1.13107765e+00 -4.22084145e-02 -8.82740140e-01
-1.06054091e+00 1.09489596e+00 -2.49497876e-01 -9.80581343e-01
-2.70247996e-01 -6.34738386e-01 -5.32296300e-01 1.13059556e+00
-1.78094423e+00 -1.05765522e+00 1.15827993e-01 6.01984501e-01
8.99489462e-01 -2.22575694e-01 8.04822266e-01 8.40896428e-01
-6.28092289e-01 7.25381613e-01 3.58817428e-01 -3.57460752e-02
8.50272417e-01 -1.07271898e+00 9.90441501e-01 1.09002280e+00
4.23990846e-01 2.54366249e-01 7.14019060e-01 -3.58482361e-01
-9.75335062e-01 -1.18449748e+00 9.37542379e-01 2.21693655e-03
7.07285523e-01 -8.15004706e-01 -1.02324712e+00 6.74866855e-01
6.48087323e-01 -1.62058458e-01 5.34795761e-01 1.05846025e-01
3.14314291e-03 2.38125827e-02 -7.41124392e-01 5.15482843e-01
1.07462287e+00 -7.22794771e-01 -3.21057141e-01 1.40190542e-01
1.07752991e+00 -4.89659429e-01 -6.51314378e-01 2.53222346e-01
2.12186530e-01 -1.03779137e+00 9.69956696e-01 -2.58313328e-01
5.70892915e-03 -1.74152181e-01 -3.39376956e-01 -2.02816725e+00
-1.65551566e-02 -1.12555194e+00 6.12781905e-02 1.63360131e+00
8.30741405e-01 -8.50173295e-01 1.02639720e-01 -8.01707655e-02
-7.71112144e-01 -2.67261386e-01 -9.39092696e-01 -1.05147147e+00
-2.24611327e-01 -1.12117541e+00 4.97852683e-01 6.14503205e-01
-1.63561925e-01 5.42921126e-01 -5.34919739e-01 3.71724427e-01
3.03995192e-01 -5.90405464e-01 6.04347706e-01 -6.60815001e-01
-7.16158390e-01 -4.05379444e-01 8.37128013e-02 -1.04295552e+00
2.31626913e-01 -9.47866619e-01 3.03993583e-01 -1.38227713e+00
-4.51521873e-01 -1.62421972e-01 -3.51526916e-01 3.78726751e-01
-2.17098668e-01 1.15108520e-01 4.11942482e-01 -3.77209187e-01
-4.46235746e-01 8.73535514e-01 1.16831982e+00 3.88829671e-02
-3.30931067e-01 3.30208033e-01 -4.54488456e-01 3.81135315e-01
9.85859990e-01 -3.10200512e-01 -2.55832374e-01 -3.92061055e-01
-4.02173936e-01 2.94611454e-01 1.56275928e-01 -1.14387369e+00
2.91888267e-01 5.25005817e-01 -2.88547486e-01 -4.85555023e-01
8.50583136e-01 -5.89529574e-01 1.52055770e-01 7.32336864e-02
-2.74259716e-01 -2.10411802e-01 5.04046261e-01 1.17367566e-01
-6.23048782e-01 -1.84158962e-02 8.73227954e-01 2.02849835e-01
-6.79607093e-01 -9.71518864e-04 -6.29394174e-01 6.34976849e-02
5.35479486e-01 7.53747672e-02 -3.37780923e-01 -7.08021164e-01
-9.36561525e-01 -8.86257738e-02 -3.67180295e-02 4.58259940e-01
6.84699833e-01 -9.09690619e-01 -1.21215320e+00 4.44237471e-01
-7.75664579e-03 -3.80628139e-01 5.81983507e-01 1.13705945e+00
6.07584715e-02 5.25581360e-01 8.70718732e-02 -3.18380654e-01
-1.42616093e+00 1.75478607e-01 4.57556814e-01 -1.09867565e-01
-5.08012235e-01 1.03827059e+00 1.11213408e-01 -6.36543870e-01
5.14010727e-01 -2.51008511e-01 4.98102792e-03 -9.05081481e-02
6.33070409e-01 5.49740255e-01 5.69020450e-01 -6.60696566e-01
-3.06058079e-01 2.63025109e-02 1.15384221e-01 -7.49253929e-01
1.48720968e+00 -4.21408981e-01 1.37069821e-01 3.37289631e-01
1.14753687e+00 3.01327795e-01 -8.81765485e-01 -3.42698276e-01
-1.42925680e-01 6.22925302e-03 4.40693885e-01 -1.29440701e+00
-1.03107774e+00 1.20390558e+00 6.22378647e-01 1.04412712e-01
1.43738604e+00 -1.33434795e-02 8.42860401e-01 4.75163348e-02
9.17551368e-02 -1.08115649e+00 -2.24003032e-01 7.98672915e-01
1.21179855e+00 -1.18977404e+00 -6.88218713e-01 -5.57025313e-01
-7.04610467e-01 8.00694168e-01 1.80526689e-01 4.96472448e-01
6.35705054e-01 8.38644862e-01 3.29226524e-01 1.97298422e-01
-7.25274980e-01 -6.06201947e-01 3.01965892e-01 6.72238827e-01
7.72503972e-01 1.88930169e-01 -9.35043544e-02 7.03385472e-01
-3.39625180e-01 -1.61399692e-01 3.08561444e-01 5.39745212e-01
-2.65043974e-01 -1.42238176e+00 -6.81275845e-01 -3.81487305e-03
-5.38100481e-01 -6.74324274e-01 -2.09996417e-01 5.96539199e-01
-2.96027184e-01 1.32160950e+00 -2.68006116e-01 -4.01990771e-01
5.28227448e-01 2.35924616e-01 2.63916641e-01 -5.99837124e-01
-8.68634045e-01 5.59225976e-01 5.43111265e-01 -4.38205242e-01
-1.35903567e-01 -6.60438359e-01 -1.07433116e+00 -1.02971710e-01
-4.46578115e-01 8.93514007e-02 1.00695586e+00 7.21386254e-01
3.29971015e-01 1.00251913e+00 3.95632267e-01 -9.21351850e-01
-4.77615058e-01 -1.52178550e+00 -7.38918185e-01 1.91241875e-01
5.97791851e-01 -2.77348250e-01 -4.49810475e-01 -7.28143230e-02] | [14.780576705932617, 6.066534042358398] |
c8e62af7-7408-49c1-ae1c-890d7a41839f | multivariate-confidence-calibration-for | 2004.13546 | null | https://arxiv.org/abs/2004.13546v1 | https://arxiv.org/pdf/2004.13546v1.pdf | Multivariate Confidence Calibration for Object Detection | Unbiased confidence estimates of neural networks are crucial especially for safety-critical applications. Many methods have been developed to calibrate biased confidence estimates. Though there is a variety of methods for classification, the field of object detection has not been addressed yet. Therefore, we present a novel framework to measure and calibrate biased (or miscalibrated) confidence estimates of object detection methods. The main difference to related work in the field of classifier calibration is that we also use additional information of the regression output of an object detector for calibration. Our approach allows, for the first time, to obtain calibrated confidence estimates with respect to image location and box scale. In addition, we propose a new measure to evaluate miscalibration of object detectors. Finally, we show that our developed methods outperform state-of-the-art calibration models for the task of object detection and provides reliable confidence estimates across different locations and scales. | ['Fabian Küppers', 'Jan Kronenberger', 'Amirhossein Shantia', 'Anselm Haselhoff'] | 2020-04-28 | null | null | null | null | ['classifier-calibration', 'classifier-calibration'] | ['computer-vision', 'miscellaneous'] | [ 6.27357438e-02 -2.61605650e-01 -5.79887331e-02 -7.05304027e-01
-6.01750135e-01 -5.84593892e-01 5.22836506e-01 3.94266516e-01
-8.28917444e-01 7.69821584e-01 -5.76817811e-01 -2.06341043e-01
-3.67718190e-02 -5.67459106e-01 -8.97947371e-01 -6.48210704e-01
1.99277118e-01 2.82902837e-01 6.97529733e-01 4.15574968e-01
4.28755999e-01 7.97224700e-01 -1.57052791e+00 -1.73252337e-02
5.17953515e-01 1.10803103e+00 5.69612384e-02 7.77939737e-01
1.80647463e-01 3.13469261e-01 -8.74099553e-01 -5.08108079e-01
-4.96336930e-02 -2.32453108e-01 -2.77447522e-01 -2.23256260e-01
8.95909131e-01 -2.87741035e-01 1.38231024e-01 1.45087254e+00
4.54890639e-01 -2.94653833e-01 1.00455260e+00 -1.45886433e+00
-4.22887206e-01 7.55365849e-01 -7.65464664e-01 4.09733415e-01
-1.45100310e-01 -2.99410880e-01 4.30886447e-01 -8.11248422e-01
2.72970259e-01 1.17974567e+00 8.42438221e-01 3.96296561e-01
-1.26241720e+00 -1.10254979e+00 2.36512169e-01 1.34437844e-01
-1.50221825e+00 -3.08385223e-01 5.75069726e-01 -6.14501476e-01
4.49030906e-01 5.16883424e-03 3.80283833e-01 1.23013437e+00
1.60613969e-01 3.62224132e-01 1.40565717e+00 -7.71293283e-01
4.98396665e-01 6.81411266e-01 5.42771280e-01 5.00244379e-01
9.41996396e-01 3.98219705e-01 -3.28430980e-01 9.17953551e-02
8.97077024e-01 -3.91888678e-01 -6.65441751e-02 -7.48749256e-01
-1.07788348e+00 8.29898000e-01 5.39806843e-01 4.21152592e-01
-1.91291347e-02 3.17336291e-01 2.04985738e-01 -1.44570485e-01
4.75716859e-01 4.52279508e-01 -2.97067642e-01 1.70717016e-01
-9.76784468e-01 1.79731548e-01 8.79199266e-01 1.11991060e+00
4.07219321e-01 4.35138829e-02 -2.52474636e-01 6.93527997e-01
1.79226547e-01 6.31211519e-01 2.16429099e-01 -7.61211812e-01
1.13711968e-01 2.31737107e-01 2.34215945e-01 -1.01804423e+00
-3.89952540e-01 -6.17648304e-01 -6.29987895e-01 6.91600084e-01
6.33987367e-01 -7.06001557e-03 -6.58910990e-01 1.58867872e+00
2.50522256e-01 2.62368321e-01 -2.47897670e-01 8.31248403e-01
5.57543218e-01 -4.89710318e-03 1.55473381e-01 -1.33941948e-01
1.22353387e+00 -6.13441586e-01 -4.76796955e-01 -2.11788580e-01
1.15908287e-01 -7.02864707e-01 5.30250072e-01 6.67026877e-01
-7.40214169e-01 -7.41369188e-01 -1.40914166e+00 4.09266710e-01
-5.79896450e-01 3.35642964e-01 5.88233292e-01 1.26390040e+00
-6.21515989e-01 5.28093398e-01 -6.74046099e-01 -2.37764165e-01
4.75339949e-01 2.36611560e-01 -3.11502606e-01 2.06380457e-01
-6.78694546e-01 1.40832055e+00 7.14250803e-01 4.61268388e-02
-7.14134037e-01 -5.24147570e-01 -6.36966646e-01 -2.24302486e-01
2.20760196e-01 -2.75432646e-01 1.49438345e+00 -1.01205277e+00
-1.08268416e+00 7.99803257e-01 2.41179377e-01 -6.81531012e-01
8.31206381e-01 -2.51834005e-01 -3.90388370e-01 -2.04401270e-01
-1.06287703e-01 8.50906432e-01 1.00826979e+00 -1.34122264e+00
-8.64287019e-01 -3.27186823e-01 -1.69470608e-01 -2.60226816e-01
-3.56432885e-01 1.93540767e-01 -2.69022167e-01 -5.43135464e-01
1.30939186e-01 -8.08169484e-01 -2.84011243e-03 3.77849430e-01
-1.87491447e-01 8.98173153e-02 7.15924263e-01 -3.10680628e-01
1.03888142e+00 -2.08827043e+00 -1.96619257e-01 2.32421577e-01
8.90643820e-02 2.68253356e-01 1.33930609e-01 -2.63145685e-01
-2.51086831e-01 3.63574699e-02 -3.27281237e-01 -3.69902909e-01
-5.52573800e-02 4.30785306e-02 -3.73080403e-01 7.65748143e-01
2.33675674e-01 4.29981589e-01 -8.05180788e-01 -7.35269010e-01
5.31248927e-01 6.01642311e-01 -1.48051500e-01 1.21227033e-01
1.75656214e-01 2.17983961e-01 -1.27829731e-01 5.50281048e-01
1.05367720e+00 2.31498927e-01 -7.56580085e-02 -4.47043836e-01
-3.26113313e-01 -1.81446478e-01 -1.51160157e+00 1.10464394e+00
-3.25116843e-01 8.16285014e-01 -1.62759349e-01 -9.20916200e-01
1.03262460e+00 8.54916498e-02 7.11546317e-02 -2.72217095e-01
5.95565617e-01 3.52396578e-01 6.54541254e-02 3.57208028e-02
5.54074705e-01 -2.09929962e-02 3.04787308e-02 1.36298373e-01
3.17522019e-01 -4.48592216e-01 1.19427793e-01 -1.82392523e-01
6.01390302e-01 1.52929783e-01 6.76697612e-01 -2.94555157e-01
3.34858567e-01 -7.92805627e-02 2.28007555e-01 1.07554829e+00
-3.83367479e-01 6.97597921e-01 4.07876700e-01 -2.63458967e-01
-1.18569744e+00 -1.10581982e+00 -7.80534983e-01 7.81035721e-01
9.60557833e-02 2.26140589e-01 -1.17588043e+00 -5.81427038e-01
3.09165418e-01 7.88568735e-01 -8.08456838e-01 -1.58931956e-01
-3.67921472e-01 -7.06273139e-01 6.26598954e-01 9.75205481e-01
4.21871483e-01 -7.56017447e-01 -9.60509002e-01 1.00492917e-01
2.97527999e-01 -1.19912028e+00 -1.84534267e-01 3.43040764e-01
-1.08728385e+00 -1.33493090e+00 -7.61375129e-01 -5.37602007e-01
8.20313990e-01 2.14721352e-01 1.17977893e+00 -8.52815732e-02
-4.51638520e-01 3.99325848e-01 -2.47136727e-01 -9.37106252e-01
-5.69057584e-01 -1.16068564e-01 2.72242725e-01 -2.25318551e-01
4.21960831e-01 -1.42224655e-01 -3.30640018e-01 7.06358671e-01
-6.29264474e-01 -2.85618126e-01 6.96131289e-01 5.94802439e-01
5.88939846e-01 -1.09551251e-01 4.75477368e-01 -9.69558954e-01
5.93258142e-01 -1.23684742e-01 -1.39359701e+00 3.86988908e-01
-8.70647669e-01 2.30863497e-01 1.73540294e-01 -8.22360694e-01
-9.41963673e-01 2.90899128e-01 5.80475181e-02 -4.73122656e-01
-4.27447200e-01 -7.66191557e-02 2.87976712e-01 -5.36486983e-01
1.08564556e+00 -4.61402893e-01 -3.79008859e-01 -3.87049764e-01
2.97679096e-01 6.09110236e-01 8.40161085e-01 -5.23941696e-01
6.96040452e-01 3.62819016e-01 1.59646720e-01 -6.21601403e-01
-7.60536790e-01 -4.94387358e-01 -1.01478004e+00 -3.18822592e-01
6.57258987e-01 -7.14799285e-01 -7.95741200e-01 5.50565064e-01
-1.30128181e+00 8.99450183e-02 1.29941590e-02 7.61185944e-01
-4.60231423e-01 3.24104488e-01 -1.09710038e-01 -1.26585162e+00
-1.25234619e-01 -9.95430887e-01 8.83967161e-01 2.58806288e-01
-9.10416096e-02 -6.91547394e-01 -2.43216399e-02 -1.97379351e-01
3.89007360e-01 2.90598333e-01 3.20722878e-01 -5.22145033e-01
-3.12611729e-01 -5.92598855e-01 -5.07582009e-01 5.29030561e-01
-1.69051796e-01 3.60286444e-01 -1.30863500e+00 -1.85167953e-01
-6.68643229e-03 -1.28350571e-01 9.26145136e-01 6.91647291e-01
1.26155448e+00 1.15315087e-01 -5.18353343e-01 2.83610433e-01
1.15244651e+00 1.07955441e-01 4.27183807e-01 3.59285533e-01
1.51493683e-01 4.62821156e-01 9.23345029e-01 2.51193434e-01
-2.42346883e-01 7.18345046e-01 5.57350934e-01 9.96556282e-02
-1.66153073e-01 -1.18814157e-02 -1.17753565e-01 9.53426883e-02
-3.89790982e-02 -3.75844315e-02 -7.41091311e-01 4.31835145e-01
-1.75278378e+00 -8.28231573e-01 -3.49409550e-01 2.61709070e+00
7.05685496e-01 5.78452885e-01 1.35435432e-01 3.91910851e-01
1.07274771e+00 -3.82057101e-01 -5.59093297e-01 -2.08712742e-01
-2.39645690e-02 2.13461280e-01 9.36773777e-01 4.16989923e-01
-1.34907508e+00 5.81479609e-01 7.48635483e+00 5.58519006e-01
-9.57957268e-01 1.14460848e-01 4.83687282e-01 2.03828201e-01
3.60328376e-01 -1.80643469e-01 -1.14666688e+00 3.09958488e-01
8.12994897e-01 2.38637626e-01 -4.13579680e-02 1.42632186e+00
-3.34875584e-01 -5.47114968e-01 -1.33852720e+00 9.31251228e-01
1.96523234e-01 -9.03644502e-01 -4.29535985e-01 -3.19610924e-01
5.92541039e-01 -2.23780558e-01 2.26007015e-01 1.78406686e-01
2.88718283e-01 -7.68064797e-01 1.21918964e+00 5.29626429e-01
7.76651025e-01 -7.79985964e-01 9.74459648e-01 2.26005435e-01
-9.35545862e-01 -5.36164604e-02 -7.09679782e-01 1.68049276e-01
-2.74377555e-01 8.14656675e-01 -1.11628211e+00 -1.55445840e-02
8.56545687e-01 2.38831192e-01 -1.00781035e+00 1.54797900e+00
-4.08330441e-01 5.76606214e-01 -3.42797548e-01 -3.88637036e-01
-1.92119196e-01 1.96733907e-01 3.09742749e-01 1.43628049e+00
4.16394591e-01 -4.18343425e-01 -4.23548728e-01 1.15638614e+00
1.66961730e-01 -1.50800899e-01 -5.27592599e-01 2.86590368e-01
6.32403195e-01 1.16724133e+00 -9.10902202e-01 -3.13796550e-01
-1.88465491e-01 6.14193261e-01 3.20743531e-01 6.92854896e-02
-1.08034658e+00 -3.22339445e-01 1.83433890e-01 -1.74557760e-01
3.67764175e-01 -2.60993421e-01 -4.29758370e-01 -7.69262075e-01
-6.14099354e-02 -5.49252927e-01 4.46515441e-01 -8.87332082e-01
-1.15814304e+00 6.25180006e-01 5.46281099e-01 -1.12840450e+00
-1.52982071e-01 -1.19895446e+00 -3.03266674e-01 7.59013712e-01
-1.31307185e+00 -1.07383895e+00 -5.23509562e-01 1.72251284e-01
3.59553158e-01 -2.11566925e-01 7.80445278e-01 1.41035929e-01
-4.60950792e-01 7.61554301e-01 8.55841935e-02 1.97104037e-01
1.04836762e+00 -1.17896509e+00 3.52638274e-01 1.03894603e+00
2.41302118e-01 5.54181755e-01 1.03894866e+00 -5.05692303e-01
-7.63228476e-01 -9.13217485e-01 3.48789483e-01 -8.02512109e-01
5.55032253e-01 -2.52238840e-01 -7.95508623e-01 4.17080700e-01
-2.09358796e-01 2.25961074e-01 3.22617233e-01 2.34672800e-01
-8.37901473e-01 -4.24196810e-01 -1.27170479e+00 2.84900576e-01
6.87345922e-01 -2.02546254e-01 -6.29218161e-01 9.26892981e-02
3.99911255e-01 -5.14520526e-01 -5.65194905e-01 5.18833995e-01
9.06152785e-01 -1.14669299e+00 1.04595351e+00 -2.81928033e-01
-8.03300589e-02 -4.15864795e-01 -9.50908735e-02 -1.13894987e+00
-1.51100263e-01 2.67398655e-01 4.01337482e-02 1.20409226e+00
4.60837781e-01 -6.71047270e-01 5.92586696e-01 5.36069274e-01
-7.90939257e-02 -9.91851166e-02 -1.05660224e+00 -9.80392396e-01
3.98853160e-02 -7.23818839e-01 2.39151374e-01 5.21521628e-01
-3.52389574e-01 -9.83719304e-02 -2.01251417e-01 4.36129808e-01
9.50002015e-01 -4.13945504e-02 5.80395103e-01 -1.47401583e+00
-2.55351514e-01 -7.33092666e-01 -6.89136207e-01 -3.64715070e-01
2.42342670e-02 -2.68448621e-01 3.44467402e-01 -1.00961339e+00
2.86662489e-01 -4.40081775e-01 -3.27861577e-01 1.99276492e-01
-8.44677091e-02 5.32533467e-01 -9.22850221e-02 -2.79265027e-02
-3.74298960e-01 -1.21044390e-01 5.66411853e-01 -6.70393333e-02
2.37981498e-01 2.56830156e-01 -4.92008984e-01 8.78247142e-01
8.84561658e-01 -8.93204093e-01 -4.37656865e-02 -3.32278073e-01
1.48232967e-01 -4.77655470e-01 8.54209125e-01 -1.54314768e+00
3.33585680e-01 -4.22724225e-02 7.38446176e-01 -8.12830865e-01
2.22196192e-01 -1.09215224e+00 1.58804163e-01 5.51603794e-01
-2.50494868e-01 3.12336814e-03 4.38793272e-01 6.99683845e-01
-9.40389410e-02 -8.69915903e-01 1.30538845e+00 6.07841797e-02
-6.72291934e-01 -1.31010801e-01 -2.33222753e-01 -2.83643246e-01
1.11160433e+00 -2.79730260e-01 -4.89013284e-01 -3.53323817e-02
-2.80579120e-01 -1.43088520e-01 4.40838993e-01 3.67793560e-01
5.39772928e-01 -1.16646481e+00 -4.75028276e-01 5.98006845e-02
3.80705506e-01 -1.77733541e-01 -2.38667682e-01 6.20108962e-01
-4.11685795e-01 3.72282863e-01 -4.31396812e-01 -8.34840536e-01
-1.47104847e+00 8.10694635e-01 4.60302591e-01 3.39753211e-01
-2.20858753e-02 9.19477940e-01 -9.18551069e-03 -1.15887664e-01
6.30008817e-01 -6.55487061e-01 -1.86770335e-01 5.86045049e-02
6.46438658e-01 4.17089999e-01 2.42819384e-01 -5.38388908e-01
-3.86162460e-01 7.50539303e-01 6.51606023e-02 -1.78996041e-01
8.91109467e-01 2.53418922e-01 2.11219743e-01 6.22648954e-01
7.38134861e-01 -3.02340060e-01 -1.23787248e+00 -4.99940664e-02
1.46456603e-02 -5.88052690e-01 1.26217112e-01 -7.94120669e-01
-7.55305827e-01 1.08814514e+00 1.08701396e+00 -1.30927470e-02
8.81859541e-01 -1.30548984e-01 -3.23426634e-01 6.02176547e-01
5.69016397e-01 -1.20490456e+00 3.76798324e-02 2.43997037e-01
9.91096020e-01 -1.43382621e+00 3.30914348e-01 -4.06860650e-01
-4.94133741e-01 1.15667844e+00 7.42143333e-01 -1.57991499e-01
5.92875957e-01 5.81042290e-01 -3.62739339e-02 2.81013310e-01
-1.51448607e-01 -6.89466745e-02 4.50316846e-01 9.53518569e-01
5.40886879e-01 4.49839309e-02 -2.33447149e-01 4.52754617e-01
-2.67814305e-02 2.66179368e-02 4.27997679e-01 7.84335136e-01
-7.27551401e-01 -8.75630796e-01 -8.23727965e-01 3.01355153e-01
-2.99418062e-01 4.29644138e-02 -3.31946701e-01 1.02785218e+00
1.60414368e-01 9.12585378e-01 -5.51999593e-03 -1.77450880e-01
4.82387304e-01 3.29578519e-02 7.25876749e-01 -4.36604053e-01
-3.41673374e-01 -1.83115616e-01 -9.74240825e-02 -1.56293735e-01
-6.04467452e-01 -6.52374923e-01 -7.32980549e-01 4.70447317e-02
-1.03012908e+00 9.57128778e-03 1.30723560e+00 5.94365835e-01
-2.64949858e-01 5.52542567e-01 1.61182836e-01 -9.78767574e-01
-7.39720583e-01 -1.20490098e+00 -5.69509387e-01 1.22091651e-01
1.79888383e-01 -1.24324572e+00 -4.66932774e-01 -5.14229164e-02] | [8.606608390808105, 2.0122761726379395] |
497fbab9-5ee2-4fcb-b2fd-70f2c6e28073 | learning-to-have-an-ear-for-face-super | 1909.12780 | null | https://arxiv.org/abs/1909.12780v3 | https://arxiv.org/pdf/1909.12780v3.pdf | Learning to Have an Ear for Face Super-Resolution | We propose a novel method to use both audio and a low-resolution image to perform extreme face super-resolution (a 16x increase of the input size). When the resolution of the input image is very low (e.g., 8x8 pixels), the loss of information is so dire that important details of the original identity have been lost and audio can aid the recovery of a plausible high-resolution image. In fact, audio carries information about facial attributes, such as gender and age. To combine the aural and visual modalities, we propose a method to first build the latent representations of a face from the lone audio track and then from the lone low-resolution image. We then train a network to fuse these two representations. We show experimentally that audio can assist in recovering attributes such as the gender, the age and the identity, and thus improve the correctness of the high-resolution image reconstruction process. Our procedure does not make use of human annotation and thus can be easily trained with existing video datasets. Moreover, we show that our model builds a factorized representation of images and audio as it allows one to mix low-resolution images and audio from different videos and to generate realistic faces with semantically meaningful combinations. | ['Simon Jenni', 'Paolo Favaro', 'Givi Meishvili'] | 2019-09-27 | learning-to-have-an-ear-for-face-super-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Meishvili_Learning_to_Have_an_Ear_for_Face_Super-Resolution_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Meishvili_Learning_to_Have_an_Ear_for_Face_Super-Resolution_CVPR_2020_paper.pdf | cvpr-2020-6 | ['audio-super-resolution', 'audio-super-resolution'] | ['audio', 'music'] | [ 4.15669978e-01 2.64712840e-01 1.19014084e-01 -3.75205129e-01
-9.38097894e-01 -2.93817759e-01 4.53145087e-01 -2.67834485e-01
-3.11082214e-01 6.90657496e-01 4.70452487e-01 3.80464017e-01
1.40708283e-01 -8.88433099e-01 -8.75871897e-01 -6.47047162e-01
1.13963716e-01 2.87302673e-01 1.25030577e-01 -3.03498376e-02
-1.09447800e-01 5.49898386e-01 -2.40182424e+00 7.01478779e-01
4.18184578e-01 1.03403652e+00 1.14724688e-01 7.68675327e-01
1.15155824e-01 5.79530656e-01 -5.24681032e-01 -3.87653708e-01
3.41499627e-01 -3.55681688e-01 -6.23570800e-01 9.95012820e-02
8.89226675e-01 -7.75444090e-01 -1.52800843e-01 9.52552855e-01
5.16741872e-01 -1.13665983e-02 4.29887831e-01 -9.74048078e-01
-4.54139262e-01 5.19960225e-01 -7.48053968e-01 -8.14146101e-02
8.00033867e-01 -2.51954794e-01 6.15797222e-01 -1.08551431e+00
8.80554795e-01 1.53642797e+00 6.14214182e-01 6.90533817e-01
-1.31558275e+00 -9.45924461e-01 -8.53773057e-02 2.08019972e-01
-1.70723808e+00 -9.47414935e-01 6.75765514e-01 -1.75791800e-01
2.39270806e-01 2.69222975e-01 6.51289821e-01 1.26740944e+00
-2.87589967e-01 2.09785908e-01 9.41813111e-01 -4.33647543e-01
-3.80244926e-02 2.69739270e-01 -5.50292134e-01 6.14858031e-01
-4.96850582e-03 6.78014830e-02 -7.67636895e-01 -2.61477679e-01
1.16308653e+00 -8.99692848e-02 -3.14776242e-01 -5.28828874e-02
-1.09483409e+00 4.84045208e-01 3.33259493e-01 2.85765439e-01
-3.79272521e-01 9.68834981e-02 4.74304594e-02 8.57345015e-02
4.44886982e-01 1.65046215e-01 3.52702960e-02 1.23635806e-01
-1.13606155e+00 1.54519036e-01 2.48368204e-01 6.19922340e-01
8.88433576e-01 5.17059639e-02 1.31883085e-01 8.86196613e-01
3.96141820e-02 4.53352511e-01 2.84150302e-01 -1.42792010e+00
1.45205170e-01 1.00869231e-01 4.22277272e-01 -1.10371184e+00
-1.95242435e-01 1.34065121e-01 -7.90556252e-01 3.13259125e-01
5.99218071e-01 8.23910087e-02 -8.37865114e-01 1.88687325e+00
3.65458995e-01 4.77781922e-01 -1.48343205e-01 9.28431273e-01
8.78447652e-01 5.66343248e-01 -1.03100725e-01 -4.37299013e-01
1.47550583e+00 -4.34234023e-01 -8.62718463e-01 -8.18050504e-02
-9.49601978e-02 -7.83494890e-01 9.75999415e-01 2.87004262e-01
-1.39344823e+00 -9.06190455e-01 -9.67023075e-01 -2.61624217e-01
1.28568774e-02 2.14688897e-01 3.06350589e-01 3.65477800e-01
-1.21267402e+00 6.08369470e-01 -6.25939250e-01 -2.37283781e-01
3.49588901e-01 3.84842604e-01 -7.45054305e-01 -2.38829449e-01
-1.05121982e+00 5.70010006e-01 3.53602469e-01 -6.74615800e-02
-7.75561512e-01 -6.23747945e-01 -8.08106840e-01 6.95220083e-02
2.14622915e-01 -6.49540782e-01 8.23765934e-01 -1.37612164e+00
-1.34494889e+00 8.42801511e-01 -3.74420881e-01 -1.23108260e-01
4.35734034e-01 -1.02293745e-01 -3.52267951e-01 7.98515260e-01
-2.38921512e-02 1.00900793e+00 1.35345280e+00 -1.36084521e+00
-6.03675187e-01 -5.68812251e-01 4.63943742e-02 1.06032133e-01
-4.09923524e-01 1.99530274e-01 -5.39897680e-01 -5.25473535e-01
1.40358895e-01 -7.65887082e-01 1.51108995e-01 3.53097051e-01
1.77226979e-02 3.15711588e-01 7.50969708e-01 -9.80445802e-01
7.72408605e-01 -2.38494325e+00 8.41462240e-02 3.41436006e-02
1.83650285e-01 -1.29955709e-02 -2.34529138e-01 -1.24563754e-01
-2.25173831e-01 1.11049853e-01 4.06021997e-03 -4.95510608e-01
-5.00083208e-01 1.37595460e-01 -3.38940948e-01 4.26893562e-01
1.82856813e-01 4.52086300e-01 -7.59858966e-01 -6.97906613e-01
6.65907785e-02 1.30821240e+00 -6.71604097e-01 3.70124340e-01
7.37402514e-02 6.70207024e-01 -1.77323088e-01 5.15242040e-01
8.37643802e-01 6.33489434e-03 1.74059972e-01 -4.50866729e-01
5.86978719e-02 1.00420743e-01 -1.24731266e+00 1.63389945e+00
-5.29581249e-01 5.49306154e-01 4.75463629e-01 -4.16203946e-01
8.96040678e-01 5.72486579e-01 5.67264020e-01 -6.07566953e-01
4.35227305e-02 2.21544486e-02 -5.93164444e-01 -2.73022115e-01
5.36500454e-01 -4.34084535e-01 1.95769995e-01 4.59726870e-01
1.51896209e-01 3.10432240e-02 -1.07583024e-01 5.79938740e-02
6.29722536e-01 1.79698184e-01 7.71688968e-02 2.82632798e-01
4.24918264e-01 -6.68065131e-01 4.47493970e-01 2.47956693e-01
1.70030177e-01 1.03331590e+00 4.12931263e-01 -3.82666200e-01
-1.38367116e+00 -1.10682940e+00 -2.19983816e-01 1.13717639e+00
-6.55980483e-02 -6.03161812e-01 -9.64247823e-01 -3.01877022e-01
-3.21753055e-01 3.07591051e-01 -7.61855364e-01 -1.07010104e-01
-5.31217635e-01 -3.52817923e-01 4.63299245e-01 4.39163387e-01
3.98224145e-01 -9.00823832e-01 -4.14299101e-01 -1.85536951e-01
-6.17982030e-01 -1.39838266e+00 -2.57089049e-01 -4.73764390e-01
-7.27658570e-01 -9.00227964e-01 -5.99428296e-01 -7.45056152e-01
8.17794561e-01 2.32524678e-01 9.66179907e-01 1.26227260e-01
-2.85794497e-01 3.39604706e-01 -1.89023986e-01 7.77293295e-02
-4.24576312e-01 -4.50214237e-01 2.76139766e-01 4.90640610e-01
4.23163688e-03 -8.95863414e-01 -4.68149096e-01 1.65443614e-01
-8.50558400e-01 2.14805335e-01 3.07265759e-01 7.41868913e-01
7.40527868e-01 2.03090698e-01 5.27691185e-01 -6.90691292e-01
8.83190036e-02 -1.98665380e-01 -4.36385989e-01 -4.28411774e-02
-1.17780305e-01 -5.58273755e-02 6.94516122e-01 -5.56729078e-01
-1.35636032e+00 3.91243219e-01 -2.69055128e-01 -7.34751821e-01
-3.11191529e-01 -9.77880582e-02 -3.13594371e-01 -2.74619117e-04
6.24204099e-01 -1.75607298e-02 7.47312456e-02 -6.79467320e-01
4.88190204e-01 7.81530261e-01 9.91556644e-01 -4.45480227e-01
7.30343282e-01 8.11329484e-01 -1.20094500e-01 -9.21764433e-01
-6.63754940e-01 -2.81121191e-02 -7.56765127e-01 -1.83258057e-01
7.57809043e-01 -1.34932971e+00 -7.64687717e-01 2.20993891e-01
-1.04816985e+00 9.21682715e-02 -3.69686365e-01 5.02621055e-01
-7.09819496e-01 2.68119335e-01 -6.68449879e-01 -9.19339359e-01
-2.66162436e-02 -9.73557651e-01 1.38947451e+00 1.95308685e-01
-9.60428938e-02 -4.58385617e-01 -3.51570368e-01 4.48364705e-01
1.58861935e-01 4.36621785e-01 5.96659839e-01 1.00054018e-01
-6.20681465e-01 -8.85688812e-02 -3.39425594e-01 1.39016882e-01
6.85088485e-02 1.41277924e-01 -1.43750823e+00 -3.85122061e-01
8.44392404e-02 -4.36395496e-01 7.73657084e-01 2.08948135e-01
1.20167804e+00 -5.42780995e-01 -8.38470906e-02 7.48243272e-01
1.23830640e+00 -1.95415393e-01 8.65718126e-01 -8.32446814e-02
7.68360555e-01 8.77835035e-01 6.32194817e-01 5.85950315e-01
3.16094667e-01 1.00057781e+00 3.99426699e-01 -1.31772116e-01
-4.82394785e-01 -5.53754091e-01 5.64153194e-01 4.71315295e-01
-5.18620014e-01 4.29289967e-01 -3.41564626e-01 4.92712170e-01
-1.44244385e+00 -1.23661149e+00 2.62472600e-01 2.44272542e+00
1.06067526e+00 -2.67540842e-01 2.67005861e-01 2.69650877e-01
8.65364969e-01 3.95811442e-03 -2.51295805e-01 -1.51845247e-01
-2.36594751e-01 2.59416908e-01 5.57854809e-02 6.13938928e-01
-7.85987556e-01 7.97973871e-01 6.83934164e+00 7.22184420e-01
-1.14425266e+00 8.88735503e-02 6.55672014e-01 -4.88302380e-01
-4.11106706e-01 -2.16112584e-01 -6.63621485e-01 3.58166635e-01
1.11130154e+00 -2.22512446e-02 7.22320795e-01 6.79465950e-01
1.73553869e-01 -4.33863662e-02 -1.07069480e+00 1.35169601e+00
2.87861615e-01 -1.07704902e+00 2.61493176e-01 4.24546450e-02
5.11223555e-01 -5.68705916e-01 2.54090488e-01 -2.04425469e-01
-1.27299398e-01 -1.26904809e+00 8.14496934e-01 6.32554829e-01
1.52770555e+00 -1.06597710e+00 4.57841069e-01 1.49218455e-01
-1.21651602e+00 -2.29405999e-01 -6.08332634e-01 1.51117742e-01
4.04556580e-02 3.93884629e-01 -7.51015186e-01 3.66784424e-01
9.81635153e-01 5.52641511e-01 -5.54840147e-01 4.96691346e-01
-2.34226570e-01 1.61555305e-01 -4.32112843e-01 8.60048354e-01
-5.19709826e-01 1.35169120e-03 3.97287667e-01 8.76310885e-01
6.75387800e-01 2.96717703e-01 -6.52661622e-02 8.99708509e-01
-2.48314425e-01 6.96232617e-02 -7.20447183e-01 6.92984909e-02
5.38755715e-01 1.34689784e+00 -5.29959559e-01 -2.34821051e-01
-3.60875607e-01 9.19588745e-01 1.80202931e-01 2.68816829e-01
-6.99858785e-01 -1.53083250e-01 7.08302021e-01 5.30679524e-01
4.37944382e-01 3.20514217e-02 -1.13993166e-02 -1.21317494e+00
5.48313037e-02 -9.38618183e-01 2.45753393e-01 -1.15919673e+00
-7.30895340e-01 8.08842659e-01 -7.78230280e-02 -1.14005828e+00
-6.44091666e-01 -2.33307049e-01 -1.38451323e-01 8.13008785e-01
-1.39370310e+00 -1.32452691e+00 -5.63368261e-01 8.50148916e-01
2.81748176e-01 5.92380837e-02 1.02973676e+00 5.28403461e-01
-1.62084252e-01 7.12510467e-01 -3.68867427e-01 9.97497141e-02
1.00661671e+00 -7.38834977e-01 -4.17304300e-02 5.71007431e-01
2.03265294e-01 5.34406841e-01 8.35566401e-01 -4.81061012e-01
-1.20430553e+00 -9.90151346e-01 7.15531588e-01 -4.44780082e-01
1.71317697e-01 -5.37589371e-01 -9.44819212e-01 6.77706838e-01
-2.14471534e-01 2.32227594e-01 6.18022203e-01 -1.00914761e-03
-6.06980264e-01 -3.28197062e-01 -1.39480031e+00 4.31958854e-01
9.38821673e-01 -9.47865486e-01 -3.52575123e-01 -2.91391723e-02
6.86028540e-01 -2.85522491e-01 -1.07209599e+00 2.97096997e-01
8.69423330e-01 -1.32019532e+00 1.19058573e+00 -2.11467162e-01
4.64014411e-01 -3.96239370e-01 -2.32391939e-01 -9.71102417e-01
-1.77358165e-01 -4.11210150e-01 -1.28382862e-01 1.42211723e+00
5.17624952e-02 -1.44656122e-01 6.63274288e-01 4.53732789e-01
4.10139322e-01 -3.42506886e-01 -9.12782371e-01 -3.59433830e-01
-3.01360101e-01 -1.87426746e-01 8.55010390e-01 8.25744152e-01
-1.35199890e-01 3.01271617e-01 -7.91615725e-01 3.67252201e-01
7.03896523e-01 1.65015072e-01 7.47309089e-01 -1.39939427e+00
-1.93501398e-01 1.29515901e-01 -4.60673928e-01 -6.15906358e-01
2.06140563e-01 -4.72277999e-01 -1.38132811e-01 -1.06350470e+00
3.66933227e-01 -1.72705576e-01 -8.44784230e-02 5.76038480e-01
2.92080492e-02 9.94796455e-01 1.60476908e-01 1.77209809e-01
-3.05157840e-01 3.95734489e-01 1.16421700e+00 1.87252611e-01
-9.66193974e-02 -2.07575887e-01 -7.86835134e-01 8.85200560e-01
4.43796277e-01 -3.00375730e-01 -2.20981881e-01 -1.87579095e-01
2.78745174e-01 4.03158695e-01 4.78157103e-01 -9.65563953e-01
-7.62763098e-02 7.07187206e-02 9.45652366e-01 -2.97456414e-01
9.09332812e-01 -8.56157959e-01 5.04690707e-01 -4.93956693e-02
-3.15282136e-01 -2.76287854e-01 1.34837225e-01 4.57683951e-01
-3.35978329e-01 -6.82167942e-03 1.05312157e+00 -2.30460376e-01
-2.17141569e-01 2.13796362e-01 1.88416336e-02 -3.47442985e-01
7.02107191e-01 -3.77362311e-01 -1.25169814e-01 -6.85378969e-01
-9.37799454e-01 -3.10688674e-01 9.29525435e-01 4.69725370e-01
8.11373234e-01 -1.62517285e+00 -7.75019526e-01 5.18043101e-01
-1.09496310e-01 -1.58912450e-01 4.25670952e-01 6.02749884e-01
-2.52161294e-01 -6.89544976e-02 -6.81811512e-01 -4.76067692e-01
-1.67616510e+00 7.50248790e-01 1.30599707e-01 2.42322981e-01
-7.01868951e-01 6.09847546e-01 4.89008993e-01 1.21449545e-01
2.84780264e-01 7.23806173e-02 -5.31858802e-01 4.03288156e-01
1.20550215e+00 2.03150094e-01 -2.04117253e-01 -1.11916363e+00
-1.94352776e-01 9.41582382e-01 1.55400902e-01 -4.19469029e-01
1.53127217e+00 -4.07868177e-01 -2.12496132e-01 4.43335414e-01
1.18067920e+00 4.42807347e-01 -1.48694396e+00 -9.25680026e-02
-6.01984322e-01 -9.15001690e-01 -5.92715747e-04 -4.08447385e-01
-1.23693550e+00 9.98374999e-01 5.84204495e-01 2.37048417e-02
1.54089797e+00 -3.02292854e-02 5.83785474e-01 -1.21287815e-01
5.11527658e-01 -1.00199902e+00 2.32260436e-01 1.93122067e-02
9.65597391e-01 -9.93032098e-01 3.81586328e-02 -5.41822016e-01
-4.76339698e-01 1.17833781e+00 4.52600598e-01 1.62093490e-01
2.94555336e-01 2.68422782e-01 -3.16436253e-02 5.70423342e-02
-7.10007548e-01 -1.95081577e-01 3.06751579e-01 7.64212251e-01
3.33698750e-01 -1.21969335e-01 2.01144069e-01 5.73038638e-01
-5.25592744e-01 5.47349863e-02 7.36234367e-01 3.56917650e-01
-4.75393862e-01 -8.51143420e-01 -7.32616723e-01 1.11685127e-01
-7.79292703e-01 1.38459578e-01 -3.87807488e-01 3.06159735e-01
4.06183213e-01 7.98712313e-01 2.99199611e-01 -4.11055505e-01
2.85623055e-02 1.01851702e-01 7.78825223e-01 -4.50606942e-01
-1.12009585e-01 3.33372027e-01 -2.49457397e-02 -8.54710639e-01
-5.07321954e-01 -4.03671026e-01 -1.10132182e+00 -3.98394525e-01
-1.41281918e-01 5.08001186e-02 6.15408421e-01 4.99039739e-01
3.63050044e-01 3.00371021e-01 6.16036355e-01 -1.22381032e+00
1.25141919e-01 -7.56357312e-01 -8.25421512e-01 5.57551980e-01
6.12266243e-01 -7.21997082e-01 -3.95740569e-01 4.30302829e-01] | [12.848860740661621, -0.130252406001091] |
a6f0c2c6-5cfe-405c-a329-0c699f55888a | a-vessel-segmentation-based-cyclegan-for | 2306.02901 | null | https://arxiv.org/abs/2306.02901v1 | https://arxiv.org/pdf/2306.02901v1.pdf | A Vessel-Segmentation-Based CycleGAN for Unpaired Multi-modal Retinal Image Synthesis | Unpaired image-to-image translation of retinal images can efficiently increase the training dataset for deep-learning-based multi-modal retinal registration methods. Our method integrates a vessel segmentation network into the image-to-image translation task by extending the CycleGAN framework. The segmentation network is inserted prior to a UNet vision transformer generator network and serves as a shared representation between both domains. We reformulate the original identity loss to learn the direct mapping between the vessel segmentation and the real image. Additionally, we add a segmentation loss term to ensure shared vessel locations between fake and real images. In the experiments, our method shows a visually realistic look and preserves the vessel structures, which is a prerequisite for generating multi-modal training data for image registration. | ['Vincent Christlein', 'Andreas Maier', 'Aline Sindel'] | 2023-06-05 | null | null | null | null | ['image-registration', 'image-to-image-translation', 'image-to-image-translation'] | ['computer-vision', 'computer-vision', 'miscellaneous'] | [ 5.14785051e-01 5.61491191e-01 -5.40558659e-02 -3.58761936e-01
-8.13980222e-01 -6.70596600e-01 5.29713035e-01 -5.47940612e-01
-4.28991616e-01 6.36402249e-01 1.40124317e-02 -3.67874563e-01
6.60625458e-01 -8.55103433e-01 -9.29992378e-01 -7.55026817e-01
5.75718641e-01 -7.07094520e-02 2.02038601e-01 1.01247020e-02
-7.41094798e-02 4.28983480e-01 -9.30716395e-01 1.86519578e-01
8.86062920e-01 9.50140536e-01 -7.23486990e-02 6.38836443e-01
6.74557611e-02 7.70699382e-01 -3.69347841e-01 -7.57132649e-01
9.21475470e-01 -1.08634722e+00 -8.11243296e-01 1.67420477e-01
1.09327793e+00 -6.33518219e-01 -4.88445669e-01 1.19586146e+00
5.89549482e-01 -3.23575616e-01 4.17084932e-01 -1.31313062e+00
-6.79539919e-01 2.48753175e-01 -7.33909428e-01 8.91078189e-02
-1.07421018e-01 5.37916958e-01 5.39643466e-01 -3.40589613e-01
7.78246224e-01 8.70714605e-01 4.67178047e-01 6.94610298e-01
-1.44453418e+00 -6.31950319e-01 -5.32540500e-01 -2.01834679e-01
-1.10782874e+00 -6.52130961e-01 6.75083458e-01 -6.92145467e-01
9.98541191e-02 1.09311365e-01 8.66483271e-01 7.77366102e-01
2.00600907e-01 2.90153593e-01 1.35920143e+00 -3.21115851e-01
-3.20196807e-01 -5.42611368e-02 -5.32855988e-01 9.31250751e-01
1.42818525e-01 5.38547575e-01 -8.44275355e-02 7.13948905e-02
1.57005668e+00 -3.16319734e-01 -4.53496367e-01 -3.21316689e-01
-1.14446390e+00 7.11940110e-01 6.86717749e-01 4.29532751e-02
-1.41557500e-01 4.16235059e-01 1.08958468e-01 2.24398479e-01
1.06046967e-01 4.55565304e-01 1.16650790e-01 4.24235731e-01
-8.24487269e-01 -8.82493705e-02 2.18360037e-01 7.43128419e-01
6.82911098e-01 2.48966917e-01 -2.89726466e-01 4.94733542e-01
2.34253258e-01 4.16091710e-01 5.93956470e-01 -1.29515231e+00
1.46947742e-01 5.56854725e-01 -8.36027637e-02 -7.22694933e-01
2.14465130e-02 -4.68598008e-01 -8.69374990e-01 5.21514714e-01
8.29192877e-01 -1.92153305e-01 -1.03514826e+00 1.80392838e+00
5.32562554e-01 5.31625390e-01 3.35810445e-02 1.06554615e+00
8.42073858e-01 1.20877114e-03 -3.38373967e-02 -7.36116469e-02
1.27604640e+00 -1.14085352e+00 -5.28154671e-01 -1.27778620e-01
6.41573846e-01 -8.28976274e-01 1.06597149e+00 -2.57718921e-01
-1.45822072e+00 -4.07619148e-01 -9.73922551e-01 -4.26706672e-01
3.34676541e-02 2.27811605e-01 5.31158626e-01 5.89319944e-01
-1.28014600e+00 1.42119229e-01 -6.82186663e-01 9.23223048e-02
8.24385226e-01 2.61242241e-01 -7.25615442e-01 -1.28928840e-01
-9.81446147e-01 9.55513835e-01 1.35289878e-01 8.92813653e-02
-8.77340555e-01 -8.20423722e-01 -9.74747419e-01 -4.49618727e-01
-2.51550794e-01 -1.19840932e+00 8.37291896e-01 -1.20822644e+00
-1.54751086e+00 1.63850629e+00 -1.06018342e-01 -5.50218403e-01
9.81247425e-01 4.97920573e-01 -1.52302951e-01 5.62994063e-01
1.12719767e-01 1.02336895e+00 9.75057840e-01 -1.16579580e+00
-3.96120846e-01 -8.95665810e-02 -1.33375106e-02 -7.76684880e-02
3.16775739e-01 5.13049215e-02 -3.35199982e-01 -7.38395751e-01
-5.35487682e-02 -8.87049615e-01 -1.80814430e-01 5.62432706e-01
-6.21801019e-01 5.62207699e-01 3.50980997e-01 -8.72193396e-01
6.07378125e-01 -1.99599326e+00 -5.76585568e-02 2.89735138e-01
5.64473987e-01 1.96068272e-01 -6.58501029e-01 -3.48295361e-01
-4.29021865e-01 4.40835161e-03 -1.56658486e-01 -3.80292177e-01
-6.32586539e-01 3.41154300e-02 -3.08665365e-01 8.19915950e-01
1.58714831e-01 1.38623595e+00 -8.16397846e-01 -7.69969523e-01
1.41807556e-01 6.12888694e-01 -5.44227660e-01 3.18954408e-01
1.10921167e-01 1.14534533e+00 -8.85808393e-02 3.98304313e-01
7.56376445e-01 -2.87470043e-01 -2.20564034e-04 -4.41863805e-01
6.02411442e-02 4.47412999e-03 -4.95761842e-01 1.69514430e+00
-4.37150717e-01 6.59147978e-01 -1.63579136e-01 -6.68615282e-01
6.86960816e-01 2.46012658e-01 6.16224170e-01 -8.72670352e-01
4.70326096e-01 2.45529935e-01 2.44349599e-01 -2.49784052e-01
-6.66842563e-03 -4.86899465e-01 1.96553037e-01 3.64951342e-01
5.86981289e-02 -3.79983217e-01 -2.92584468e-02 -8.25020969e-02
5.70399284e-01 1.45811036e-01 -1.75560951e-01 2.24615410e-01
4.97771293e-01 -4.18405198e-02 5.90031087e-01 2.89481401e-01
-2.26451084e-01 1.14856195e+00 7.74064839e-01 -3.21392298e-01
-1.40381825e+00 -1.15374327e+00 -2.49835134e-01 2.08332866e-01
3.09684724e-01 2.03143746e-01 -9.18686748e-01 -7.34976411e-01
-3.94811481e-01 3.50951731e-01 -5.80833912e-01 -3.50124240e-01
-4.45708662e-01 -4.62180287e-01 1.02011871e+00 1.32318199e-01
7.53769934e-01 -5.00678062e-01 -4.27762508e-01 -2.53902338e-02
-6.00396335e-01 -1.36251891e+00 -1.02208507e+00 -5.97432017e-01
-6.29356086e-01 -1.44912040e+00 -9.12419617e-01 -1.17427254e+00
1.12535143e+00 5.19062281e-02 8.67537022e-01 1.21074401e-01
-4.34855282e-01 -1.62170038e-01 -1.77403763e-02 -9.74540114e-02
-7.60017633e-01 -4.72410858e-01 -2.70405501e-01 3.92530113e-01
-1.32758096e-01 -6.80956841e-01 -1.04282320e+00 3.78304154e-01
-9.98257995e-01 3.12041044e-01 5.52693248e-01 1.09303379e+00
9.32883382e-01 -1.85237378e-01 1.63320675e-01 -8.08757424e-01
1.40389115e-01 1.89950436e-01 -9.44061160e-01 3.52010608e-01
-3.18599075e-01 -1.91797748e-01 2.61358410e-01 -5.59727073e-01
-7.28578985e-01 2.86825687e-01 -5.29771438e-03 -5.20183206e-01
1.04691394e-01 8.85458067e-02 -2.24266008e-01 -8.62110376e-01
5.57515621e-01 3.94694656e-01 6.13111556e-01 -4.32057530e-02
7.47212708e-01 4.82774436e-01 1.05750215e+00 -2.51167506e-01
9.66231823e-01 8.14332187e-01 7.31732547e-02 -3.04904133e-01
-8.32752168e-01 2.43200455e-03 -7.26580918e-01 -8.40032622e-02
1.06611049e+00 -1.12260211e+00 -6.26899183e-01 7.99564362e-01
-1.33128333e+00 -6.55526757e-01 -5.01702905e-01 5.00267923e-01
-7.88250923e-01 3.77393156e-01 -6.33061588e-01 1.88297387e-02
-2.50280887e-01 -1.37417316e+00 1.02365804e+00 4.33771640e-01
2.49863178e-01 -1.04439008e+00 6.83138967e-02 5.66306412e-01
3.81628543e-01 6.55658603e-01 9.61773813e-01 6.29022568e-02
-8.29171896e-01 -1.39571503e-01 -6.34337604e-01 6.91323996e-01
9.52473283e-02 -1.73860013e-01 -8.74925435e-01 -2.22207561e-01
1.53583819e-02 -3.41303974e-01 8.77038717e-01 5.72793007e-01
8.02904069e-01 -3.96825254e-01 -4.13095951e-02 1.23840773e+00
1.43868649e+00 -6.89109936e-02 1.35480404e+00 5.34693748e-02
1.01286936e+00 6.54497147e-01 2.31913209e-01 -1.61149129e-01
4.95139837e-01 7.36257315e-01 4.11535233e-01 -1.05989301e+00
-9.28112388e-01 -4.60638702e-01 1.50043011e-01 3.12758952e-01
4.10981774e-02 6.23213537e-02 -5.87717116e-01 5.15937567e-01
-1.23432565e+00 -8.40683818e-01 -1.63639933e-01 2.27185702e+00
1.30441606e+00 -3.66739094e-01 -7.30885565e-03 -5.53575754e-01
8.25420737e-01 -1.59104556e-01 -6.05035007e-01 2.10829049e-01
-3.91349196e-01 2.82749861e-01 7.85144031e-01 6.91396654e-01
-1.10931516e+00 1.27565420e+00 6.73360920e+00 5.03896415e-01
-1.58670938e+00 3.18404347e-01 8.49806428e-01 -7.62113854e-02
-3.22198570e-01 1.37524381e-01 -2.90610969e-01 3.82564157e-01
5.84969699e-01 -1.45506352e-01 1.80452436e-01 2.48135403e-01
3.42088819e-01 9.46490318e-02 -8.81177485e-01 1.04347932e+00
6.05562367e-02 -1.56410551e+00 2.48561129e-01 4.59206879e-01
8.62032175e-01 2.23403051e-02 2.28143975e-01 -6.55103981e-01
3.70375752e-01 -1.26484585e+00 5.05573273e-01 7.09980667e-01
1.55741191e+00 -4.81196225e-01 7.06289947e-01 -3.16763490e-01
-6.15439117e-01 4.71243411e-01 -2.66428173e-01 5.13495207e-01
4.12946701e-01 4.40664172e-01 -7.09462941e-01 3.61632556e-01
1.14557862e-01 6.85292423e-01 -7.30948925e-01 1.35378075e+00
-4.43064570e-01 1.47430435e-01 -1.43329158e-01 9.51884270e-01
-1.15248382e-01 -6.47474766e-01 5.87880313e-01 3.43079627e-01
1.78093120e-01 -3.69805187e-01 -1.31405160e-01 1.39718771e+00
-1.59807220e-01 -9.88165215e-02 -7.26400554e-01 -6.53886497e-02
2.40763634e-01 1.19024253e+00 -3.83683711e-01 -6.34391010e-02
-4.22666222e-01 1.06596708e+00 -1.38936996e-01 4.66263264e-01
-9.08456564e-01 -3.51280242e-01 7.09471107e-01 3.95363599e-01
7.74615854e-02 3.51597257e-02 -2.34475896e-01 -1.27815473e+00
-2.49885842e-02 -6.60840392e-01 -2.87682302e-02 -1.03963912e+00
-1.20003235e+00 7.39344060e-01 -5.80625653e-01 -1.58593810e+00
-1.55154869e-01 -2.29081154e-01 -8.05765390e-01 1.22162247e+00
-1.96406567e+00 -1.95565426e+00 -3.84880990e-01 7.37935781e-01
-4.44916695e-01 -2.17750654e-01 7.16512322e-01 3.27960849e-01
-4.69081700e-01 9.95085239e-01 -2.90949225e-01 6.85967207e-01
9.74985600e-01 -9.24578309e-01 3.18712622e-01 1.24559605e+00
-1.81835860e-01 4.97468144e-01 2.46666893e-01 -4.92159307e-01
-7.88748205e-01 -1.47609806e+00 5.55727363e-01 -4.96894538e-01
5.96442640e-01 5.47052138e-02 -5.72495461e-01 8.68368566e-01
1.53914675e-01 4.54680681e-01 6.70459747e-01 -7.34881759e-01
-4.15366501e-01 -3.20413075e-02 -1.28378069e+00 7.64364064e-01
6.22569621e-01 -7.30650723e-01 -2.08414719e-01 3.90229613e-01
7.44127214e-01 -8.14008772e-01 -1.18081439e+00 3.06476187e-02
6.22714221e-01 -7.50481606e-01 1.15142703e+00 -4.41885233e-01
7.06854165e-01 -5.34626365e-01 3.69313121e-01 -1.19052815e+00
2.75181532e-01 -9.98522222e-01 4.79276985e-01 1.15687680e+00
4.35776561e-01 -9.45133865e-01 8.47538888e-01 6.87713623e-01
-4.53062821e-03 -2.38439500e-01 -1.09382582e+00 -4.82937098e-01
3.97190332e-01 1.35646775e-01 4.83580321e-01 1.03997552e+00
-4.96928751e-01 1.16052896e-01 -4.23868090e-01 4.20438200e-02
9.10660505e-01 1.35703251e-01 9.13851917e-01 -7.74112582e-01
-2.94501513e-01 -5.78929663e-01 -6.01738989e-01 -1.02929986e+00
2.20888659e-01 -1.17335320e+00 -3.10410947e-01 -1.04696548e+00
1.13909461e-01 -4.94745731e-01 -2.41356809e-02 5.72027802e-01
1.20172232e-01 1.03755212e+00 -1.68612357e-02 2.76293308e-01
-1.55160194e-02 5.05637586e-01 2.27217531e+00 -1.97559178e-01
-2.01752916e-01 -7.29349405e-02 -9.41699922e-01 5.08772433e-01
6.75778627e-01 -4.48443472e-01 -4.42693859e-01 -5.01973808e-01
-1.39389634e-01 2.33206332e-01 9.34515774e-01 -6.30744576e-01
1.64682478e-01 -2.29723938e-02 2.41734967e-01 -1.05375059e-01
1.63717419e-01 -6.61603808e-01 1.63189426e-01 4.26351577e-01
-3.55037034e-01 -4.76085037e-01 -6.17689155e-02 3.11132461e-01
-4.99936849e-01 2.00592071e-01 1.35294783e+00 -1.80386633e-01
-1.11918136e-01 6.62743747e-01 3.64323169e-01 2.17024967e-01
9.43083107e-01 -4.42122608e-01 -6.91785395e-01 -2.85904855e-01
-7.37401962e-01 1.09295271e-01 9.98531699e-01 1.79061949e-01
8.65575314e-01 -1.35125232e+00 -8.23091209e-01 5.02677500e-01
2.05940261e-01 1.87899917e-01 2.52730370e-01 1.39087462e+00
-9.71590638e-01 2.64310651e-03 -5.94578803e-01 -5.45790792e-01
-1.15554643e+00 1.90862596e-01 1.13285768e+00 6.86925054e-02
-7.36605287e-01 6.96904898e-01 4.81192380e-01 -2.75629193e-01
-1.18495472e-01 -1.39318421e-01 8.46648663e-02 -4.10611182e-01
5.66705346e-01 -1.78946942e-01 -2.12417319e-01 -9.07996774e-01
6.75206855e-02 8.23345661e-01 9.35882889e-03 -9.27733332e-02
9.93055105e-01 -3.62348974e-01 -4.80166405e-01 -1.22707240e-01
1.35157013e+00 1.54362842e-02 -1.50947952e+00 -3.06708634e-01
-6.83868468e-01 -7.24574327e-01 3.55963349e-01 -7.77955115e-01
-1.56695724e+00 8.46581995e-01 8.85967731e-01 -3.35529894e-01
1.09986973e+00 -1.25630558e-01 1.14764524e+00 -4.43552345e-01
2.34722093e-01 -4.28395659e-01 3.23125608e-02 6.51192963e-02
7.85375655e-01 -1.33251321e+00 -2.56816387e-01 -6.67706788e-01
-6.92918479e-01 8.48224759e-01 6.21377826e-01 -9.65243727e-02
4.23705488e-01 1.19317286e-01 6.09841049e-01 1.42007163e-02
-8.49167109e-02 -3.60315025e-01 4.62368995e-01 1.13891017e+00
2.04762205e-01 -1.74308121e-01 -1.24000154e-01 1.54952884e-01
-3.18808377e-01 1.05400823e-01 8.91960502e-01 1.88945517e-01
2.74385959e-01 -1.40188611e+00 1.20601378e-01 2.49784291e-02
-5.50552189e-01 -3.45777035e-01 -2.46520966e-01 6.77985907e-01
1.67465746e-01 6.42195642e-01 3.10774505e-01 -1.56836376e-01
1.48887590e-01 -3.61153096e-01 8.00453663e-01 -4.42231178e-01
-4.71791208e-01 2.68778771e-01 -1.86722234e-01 -7.66245246e-01
-4.98735398e-01 -2.64756322e-01 -1.06200731e+00 -2.40827501e-01
-1.09194942e-01 -2.89994657e-01 5.41947246e-01 6.63474381e-01
3.63571882e-01 4.36293632e-01 7.37477481e-01 -5.19067705e-01
1.82159711e-02 -5.34047723e-01 -6.04640126e-01 5.60451984e-01
5.23017704e-01 -4.09927845e-01 -2.16556370e-01 5.68088114e-01] | [15.565449714660645, -3.734544038772583] |
c5f20e49-51e1-4d45-b0f9-654e81779797 | itnlp-aikf-at-semeval-2016-task-3-a-quesiton | null | null | https://aclanthology.org/S16-1139 | https://aclanthology.org/S16-1139.pdf | ITNLP-AiKF at SemEval-2016 Task 3 a quesiton answering system using community QA repository | null | ["Chang{'}e Jia"] | 2016-06-01 | null | null | null | semeval-2016-6 | ['question-similarity'] | ['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] | [-7.351032257080078, 3.6671559810638428] |
2baf19f8-457c-4e79-ac3f-73efa739c375 | a-supervised-model-for-extraction-of | null | null | https://aclanthology.info/papers/W14-0802/w14-0802 | https://www.aclweb.org/anthology/W14-0802 | A Supervised Model for Extraction of Multiword Expressions, Based on Statistical Context Features | null | ['Ronaldo Martins', 'Meghdad Farahmand'] | 2014-04-01 | null | https://aclanthology.org/W14-0802 | https://aclanthology.org/W14-0802.pdf | ws-2014-4 | ['cross-lingual-information-retrieval'] | ['natural-language-processing'] | [-2.44508207e-01 3.89024585e-01 -2.65282035e-01 -2.15905145e-01
-8.60921741e-02 -7.76765764e-01 4.48510379e-01 -7.23253429e-01
-5.48377395e-01 1.31954515e+00 3.66348401e-02 -9.49533224e-01
-2.40340635e-01 -1.05564880e+00 -8.44053447e-01 -8.75781775e-01
-7.42435038e-01 6.86515033e-01 1.44298598e-01 -6.52004302e-01
8.47113907e-01 6.29996777e-01 -1.62287033e+00 5.77558100e-01
6.81926727e-01 5.49597681e-01 5.66466339e-02 1.04480565e+00
6.13576770e-02 1.59847176e+00 -6.05450153e-01 -4.56729174e-01
1.43710867e-01 -1.88022718e-01 -5.35770595e-01 -3.36825520e-01
7.91536197e-02 -5.88986814e-01 -3.10633808e-01 6.07356608e-01
1.14270830e+00 -5.53557873e-02 1.07025254e+00 -1.51067197e+00
-7.97060013e-01 5.44234335e-01 2.30399013e-01 1.20632313e-01
6.10446513e-01 -1.79472014e-01 3.51212233e-01 -1.36621380e+00
6.32680655e-01 8.64870071e-01 9.47046995e-01 5.99273622e-01
-1.20647264e+00 -5.60552299e-01 -7.17137158e-01 -2.57835120e-01
-1.59567785e+00 -6.30042374e-01 6.23617955e-02 -3.34735900e-01
1.69142139e+00 6.72587395e-01 1.35647905e+00 1.47341585e+00
1.36649036e+00 4.72517729e-01 1.13904178e+00 -3.16340059e-01
3.65351558e-01 6.43754780e-01 -2.29148138e-02 5.51889002e-01
1.19686353e+00 6.76312149e-01 -5.60917914e-01 -7.92779744e-01
1.11365557e+00 -3.12949389e-01 2.60771513e-01 -7.06989706e-01
-1.06566191e+00 7.88466871e-01 -1.33808568e-01 2.20882341e-01
-3.46203089e-01 -9.92989019e-02 2.24548295e-01 4.12961632e-01
-2.94082850e-01 4.77862865e-01 -8.89795244e-01 -1.87686309e-01
-1.01472771e+00 1.91863477e-01 1.30615628e+00 1.45811689e+00
4.72722203e-03 3.70100170e-01 1.66493103e-01 2.89523482e-01
5.29429734e-01 1.05927479e+00 4.36951250e-01 -1.33365822e+00
-9.55119133e-02 4.20878716e-02 6.12505674e-01 -1.13473701e+00
-7.12930143e-01 -2.41300568e-01 -9.54724610e-01 5.18771529e-01
-1.59270123e-01 -4.59180593e-01 -7.59686172e-01 4.79171664e-01
-1.81984559e-01 -3.19489211e-01 5.45619786e-01 2.69257158e-01
7.28577793e-01 1.45816103e-01 9.25893057e-03 -4.73371416e-01
8.38296890e-01 -1.26899445e+00 -1.11426139e+00 2.33304992e-01
6.82289064e-01 -1.11704338e+00 4.06794518e-01 4.29345578e-01
-1.87200487e+00 -9.69799384e-02 -1.08317137e+00 1.19022481e-01
-7.97166348e-01 -2.57194549e-01 6.66624963e-01 1.49922979e+00
-1.63972652e+00 6.47176981e-01 -6.05021298e-01 5.40335439e-02
-4.73218322e-01 9.07614470e-01 -3.47142309e-01 2.90655226e-01
-1.28751957e+00 9.52449858e-01 -1.04219764e-01 3.03568155e-01
-3.30324143e-01 -1.84501112e-01 -1.03385210e+00 -6.24127567e-01
-2.59417385e-01 -1.07312346e+00 1.36302292e+00 1.09815992e-01
-1.30379498e+00 1.00962746e+00 -4.14315462e-01 -1.14109404e-01
6.61359608e-01 -1.52096570e-01 -8.18174660e-01 1.58321753e-01
2.87593510e-02 3.47163439e-01 5.30214727e-01 -9.69232559e-01
-6.31655395e-01 -1.00153096e-01 -2.54188716e-01 2.48816490e-01
1.43904146e-03 2.05205917e-01 6.38460100e-01 -9.53604504e-02
1.91062525e-01 -8.61961663e-01 -3.21419090e-01 -7.61404037e-01
-1.77353188e-01 -3.84736449e-01 5.59762537e-01 -2.92163283e-01
1.33848476e+00 -1.84365213e+00 -4.54683900e-01 2.94456817e-03
2.53527164e-01 3.89016122e-02 2.33101144e-01 1.23831403e+00
-2.43849114e-01 1.06667292e+00 3.98064762e-01 9.52541176e-03
3.13382268e-01 5.07273376e-01 2.86674555e-02 2.51206756e-01
9.43330750e-02 1.15280068e+00 -1.09214199e+00 -5.20940363e-01
5.57267785e-01 4.60968643e-01 -3.84588838e-01 6.24593735e-01
9.30012882e-01 2.99172718e-02 -8.94021094e-02 1.33241999e+00
1.05377018e+00 -2.90812284e-01 -2.46628653e-02 3.44544262e-01
-6.44541264e-01 6.00222833e-02 -7.41804183e-01 9.49429095e-01
1.31349964e-02 1.99757457e-01 1.74998537e-01 -6.80728197e-01
3.94654661e-01 1.10545933e+00 4.45659161e-01 -9.24297392e-01
-1.69569388e-01 9.00310338e-01 1.03920102e-01 -9.62983370e-01
6.40755475e-01 -1.72503099e-01 7.69825354e-02 1.40852645e-01
-3.61913115e-01 -3.84120375e-01 5.52458428e-02 4.19574350e-01
1.00224769e+00 -1.33707494e-01 5.48685491e-01 -8.72328460e-01
6.88946426e-01 -2.01589778e-01 3.00597936e-01 1.23978651e+00
-2.90862411e-01 9.21478033e-01 3.19022417e-01 -5.37574828e-01
-5.72780192e-01 -1.20098352e+00 -4.80556756e-01 9.86326516e-01
-1.58388391e-01 -3.87671381e-01 -7.10155964e-01 -2.63155133e-01
-1.85071185e-01 4.93483543e-01 -5.93591571e-01 5.23159266e-01
-5.14705896e-01 -6.85349941e-01 7.65277922e-01 3.17972392e-01
6.49624616e-02 -1.49228525e+00 -5.70813596e-01 3.34754556e-01
-1.07132711e-01 -5.39811492e-01 -1.57590687e-01 6.12205923e-01
-1.19366693e+00 -1.01526415e+00 -8.75530541e-01 -1.03603637e+00
7.18889177e-01 2.44572356e-01 1.18463767e+00 2.84042627e-01
-2.87709564e-01 6.49904788e-01 -1.31030381e-02 -2.67939389e-01
-3.16364132e-02 -5.78750819e-02 5.01899779e-01 -1.07851994e+00
6.46393359e-01 -3.92214775e-01 -1.00622165e+00 5.35368681e-01
-3.85821253e-01 -2.04611585e-01 7.85597622e-01 1.04281425e+00
1.42742008e-01 -1.33182362e-01 2.41971418e-01 -1.16735983e+00
1.11498392e+00 -4.85333532e-01 1.30341025e-02 2.62960047e-01
-8.53362501e-01 -4.53584343e-01 2.27881029e-01 -1.33037493e-02
-6.33760452e-01 -2.51423597e-01 -4.14545417e-01 3.22794706e-01
-4.06866640e-01 3.34393084e-02 2.24873364e-01 -5.26000679e-01
5.21005154e-01 1.50258854e-01 -4.07556817e-02 1.21735156e-01
6.30975887e-02 3.75316888e-01 3.67786407e-01 -7.01224983e-01
5.19809723e-01 1.93325520e-01 -3.00995037e-02 -1.28791058e+00
2.66786069e-01 -1.64763972e-01 -6.60750508e-01 -2.69126564e-01
7.75121748e-01 -9.50198710e-01 -6.35705650e-01 5.21708906e-01
-9.54387486e-01 -1.36340424e-01 -5.25335252e-01 6.30965471e-01
-7.02474535e-01 2.94757709e-02 -4.17852998e-01 -1.17976546e+00
-4.78847146e-01 -1.38869667e+00 9.59160328e-01 3.85633826e-01
-2.05811903e-01 -1.19790471e+00 3.38975668e-01 8.99848714e-02
4.29175526e-01 7.39965662e-02 5.03531873e-01 -2.80004859e-01
-2.98304528e-01 -2.85470933e-01 1.59573574e-02 -1.65876746e-01
-3.09365243e-02 4.80770379e-01 -5.48391879e-01 -4.37962860e-01
2.44738594e-01 -4.86500829e-01 -7.51581416e-02 7.11018145e-01
4.09077644e-01 -6.02920949e-01 -7.12544680e-01 5.56552589e-01
1.27709007e+00 4.72497940e-01 5.88340998e-01 6.77552521e-01
9.28869769e-02 5.17871320e-01 4.48484391e-01 2.08113909e-01
1.07506551e-01 1.08483166e-01 1.14217520e-01 -2.17137560e-01
9.45900455e-02 -1.75644696e-01 4.44304377e-01 1.09980536e+00
-8.23837519e-01 -2.15223029e-01 -4.88283783e-01 6.68056428e-01
-1.69156313e+00 -1.31570101e+00 -6.70152605e-01 9.76623476e-01
3.78074080e-01 3.38884830e-01 1.58041969e-01 1.21232189e-01
5.47049284e-01 -4.76693302e-01 -1.27305150e-01 -9.16352808e-01
-3.82233024e-01 4.21136379e-01 7.14838982e-01 6.51479006e-01
-6.10602558e-01 5.08361042e-01 1.25505590e+01 5.49717486e-01
-5.10860123e-02 1.27314463e-01 3.14815581e-01 1.99914098e-01
-2.24020526e-01 1.62632689e-01 -1.04165065e+00 9.26906094e-02
1.64754760e+00 -4.36735392e-01 3.93212438e-01 6.44444764e-01
3.94769877e-01 -1.96429595e-01 -9.78620768e-01 5.21999180e-01
1.12709433e-01 -1.50397491e+00 -3.76861334e-01 9.64888573e-01
5.80093563e-01 -1.81895152e-01 6.81941509e-01 4.33209062e-01
7.39425182e-01 -1.14519978e+00 7.88427055e-01 3.18626672e-01
8.87266219e-01 -8.18826377e-01 1.07315242e+00 2.76328743e-01
-1.04355454e+00 1.15487054e-02 -5.90443075e-01 -9.22581375e-01
1.62244976e-01 2.02188566e-01 -5.61583102e-01 4.06869203e-01
6.19966745e-01 2.15250537e-01 -8.53338003e-01 1.45581341e+00
-4.29231405e-01 4.22687940e-02 -3.00976187e-01 -4.39922899e-01
4.83290881e-01 5.28384298e-02 3.26606274e-01 1.02295005e+00
2.89785862e-01 5.45581102e-01 -2.62208551e-01 2.92286038e-01
6.63503349e-01 5.84383309e-03 -1.38496268e+00 -8.68683681e-02
3.54620427e-01 9.69217539e-01 -3.99494410e-01 -3.56748998e-02
-6.11456513e-01 6.62115753e-01 -1.66942239e-01 5.75146377e-01
-2.92761147e-01 -8.89406264e-01 4.33969021e-01 -2.30117049e-02
-2.88469017e-01 -1.35338321e-01 -3.23372036e-01 -1.02459288e+00
-6.06416404e-01 -3.90122980e-01 7.32179061e-02 -8.63867998e-01
-1.79313409e+00 6.85844719e-01 -3.50327820e-01 -1.03178740e+00
-9.29034173e-01 -9.72984850e-01 -2.90242791e-01 6.67755604e-01
-6.59234166e-01 -1.30476952e+00 1.77773401e-01 4.65253919e-01
1.82401046e-01 -5.88519633e-01 1.09945750e+00 -1.42670095e-01
-3.18711132e-01 1.08882773e+00 3.51253241e-01 -3.05974871e-01
8.32965672e-01 -1.16271269e+00 4.23785180e-01 1.09189451e-02
-6.59259140e-01 1.28460062e+00 6.81860268e-01 -6.81785107e-01
-1.16639841e+00 -1.52893141e-01 1.38110602e+00 -9.15031374e-01
5.27874053e-01 -1.02810413e-01 1.46300867e-01 6.57190859e-01
9.07351732e-01 -5.09461224e-01 1.08935249e+00 -1.13616258e-01
4.93123859e-01 8.23448122e-01 -1.38805819e+00 5.99272728e-01
1.21222687e+00 -4.81854081e-01 -8.27020168e-01 3.28937650e-01
9.94227231e-01 -5.60786903e-01 -1.34066856e+00 6.99696779e-01
1.14132881e+00 -1.04883564e+00 1.54237401e+00 -1.16912270e+00
4.08816934e-01 1.70770600e-01 -2.18455836e-01 -6.52935803e-01
-5.24122715e-01 -8.07741940e-01 -3.97186369e-01 3.40213925e-01
8.64144087e-01 -1.16800272e+00 5.53630650e-01 1.39166510e+00
-3.58625412e-01 -4.51715857e-01 -8.63314152e-01 -1.06149745e+00
-4.66870656e-03 1.81816339e-01 3.85523647e-01 8.18638563e-01
7.09820271e-01 2.02331603e-01 -2.05201268e-01 -4.37674336e-02
6.59316301e-01 -4.87512946e-01 5.82342803e-01 -8.74011815e-01
2.49149442e-01 -3.85976404e-01 -2.94489831e-01 -2.90531307e-01
-3.26795012e-01 -4.87107635e-01 -8.59293997e-01 -1.53084815e+00
8.49664882e-02 1.97367743e-01 -2.33256146e-01 3.44385266e-01
6.77823246e-01 2.45789096e-01 -2.27382466e-01 1.17054820e-01
-1.86927155e-01 -9.06788632e-02 9.18564737e-01 1.10557920e-03
-2.17724726e-01 3.92412305e-01 -1.01716614e+00 9.60534811e-01
7.85404518e-02 -6.09580100e-01 -5.51487744e-01 1.64939865e-01
5.71885824e-01 3.57775450e-01 4.49405432e-01 -9.70509052e-01
9.87445056e-01 -4.32846308e-01 6.12257063e-01 -1.43385470e+00
-2.65485227e-01 -9.00021493e-01 3.61826897e-01 6.29400074e-01
4.08042461e-01 3.53796333e-01 2.26811334e-01 -9.13986787e-02
-2.01378226e-01 -6.44658387e-01 4.43001747e-01 -7.45609999e-01
-4.55086857e-01 -3.90635371e-01 -1.02313221e+00 -9.67486948e-03
7.88939476e-01 -4.61360693e-01 -6.47767961e-01 -1.95199717e-02
-1.44608676e+00 7.30553791e-02 8.33678424e-01 1.34732248e-02
6.98022604e-01 -1.44444084e+00 -1.11748077e-01 6.18802965e-01
-5.42432606e-01 -4.98194307e-01 2.20674574e-01 1.06156552e+00
-1.31528914e+00 1.26915324e+00 -5.21896303e-01 -2.05424115e-01
-8.96585643e-01 7.19095051e-01 3.15837234e-01 1.75098896e-01
-3.93049836e-01 7.00151443e-01 1.12239346e-02 -6.58594191e-01
1.24011397e-01 2.97559589e-01 -4.87763733e-01 -2.52128989e-01
9.22277749e-01 1.17918742e+00 1.32526740e-01 -3.91715407e-01
-5.58191121e-01 4.39446121e-01 2.08604142e-01 -6.12734079e-01
1.38662541e+00 -2.20989168e-01 -7.44191706e-01 5.00318348e-01
7.50003338e-01 -7.33495504e-02 -3.28966863e-02 1.27011871e+00
4.80256975e-03 -5.47239244e-01 -3.11336160e-01 -7.82959402e-01
-3.42039734e-01 5.93214929e-01 6.17733955e-01 4.59080935e-01
9.69997942e-01 -6.47576571e-01 7.70057857e-01 7.64313757e-01
7.83133745e-01 -1.22997761e+00 -6.36951804e-01 3.69483382e-01
9.39836800e-01 -9.30032432e-01 1.30400375e-01 -3.14221263e-01
-5.55534840e-01 1.08689797e+00 5.38173854e-01 -1.62142932e-01
1.17091513e+00 6.96656525e-01 -1.75611794e-01 -4.53250527e-01
-1.04549897e+00 9.83817875e-02 2.10727617e-01 1.32608259e+00
8.26061428e-01 1.41275764e-01 -1.40426433e+00 8.46834600e-01
-6.34404242e-01 4.76996809e-01 9.47268784e-01 1.55018985e+00
-5.12773991e-01 -1.41692734e+00 -4.41162825e-01 2.31142953e-01
-2.82854587e-01 -3.49930972e-01 -8.62966061e-01 1.09929097e+00
1.97083559e-02 1.37330377e+00 -4.48758155e-01 -4.63731140e-01
8.27405751e-01 3.19600284e-01 4.41484600e-01 -1.25142381e-01
-8.94688129e-01 6.48750722e-01 3.84649158e-01 -1.05672014e+00
-7.94483483e-01 -1.27049911e+00 -1.08188879e+00 -1.16740942e+00
-3.07217389e-01 4.65777248e-01 2.31674016e-01 3.19533348e-01
-2.09303293e-02 1.56313553e-03 6.67447627e-01 -1.00111675e+00
-1.97397545e-01 -6.07395887e-01 -1.20775104e+00 -5.56946039e-01
3.94425273e-01 -5.85869551e-01 -1.15223050e+00 -6.24734610e-02] | [-1.5391768217086792, 15.86919116973877] |
e090fb15-839b-4a31-9c16-d5c15bdc4a9a | retroxpert-decompose-retrosynthesis | 2011.02893 | null | https://arxiv.org/abs/2011.02893v1 | https://arxiv.org/pdf/2011.02893v1.pdf | RetroXpert: Decompose Retrosynthesis Prediction like a Chemist | Retrosynthesis is the process of recursively decomposing target molecules into available building blocks. It plays an important role in solving problems in organic synthesis planning. To automate or assist in the retrosynthesis analysis, various retrosynthesis prediction algorithms have been proposed. However, most of them are cumbersome and lack interpretability about their predictions. In this paper, we devise a novel template-free algorithm for automatic retrosynthetic expansion inspired by how chemists approach retrosynthesis prediction. Our method disassembles retrosynthesis into two steps: i) identify the potential reaction center of the target molecule through a novel graph neural network and generate intermediate synthons, and ii) generate the reactants associated with synthons via a robust reactant generation model. While outperforming the state-of-the-art baselines by a significant margin, our model also provides chemically reasonable interpretation. | ['Junzhou Huang', 'Yang Yu', 'Jinyu Yang', 'Shuangjia Zheng', 'Peilin Zhao', 'Qianggang Ding', 'Chaochao Yan'] | 2020-11-04 | null | http://proceedings.neurips.cc/paper/2020/hash/819f46e52c25763a55cc642422644317-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/819f46e52c25763a55cc642422644317-Paper.pdf | neurips-2020-12 | ['retrosynthesis'] | ['medical'] | [ 7.21802831e-01 4.18269902e-01 -5.27049184e-01 9.87346545e-02
-3.45793724e-01 -1.15162742e+00 8.35780263e-01 6.23549163e-01
3.59703489e-02 9.64506567e-01 3.42981219e-01 -7.33136952e-01
2.76242226e-01 -8.89730930e-01 -6.00990415e-01 -7.56199241e-01
1.92051485e-01 5.08356690e-01 2.92476982e-01 -4.18591410e-01
5.35041213e-01 8.70387197e-01 -9.86174464e-01 1.70409888e-01
1.00210834e+00 7.51752436e-01 2.73882002e-01 4.34238553e-01
-1.28131062e-01 8.89577806e-01 -3.86193484e-01 -4.68050361e-01
2.86388278e-01 -6.44010603e-01 -9.18499231e-01 -2.41281822e-01
-1.37854457e-01 5.98566607e-02 -2.31054097e-01 9.41320300e-01
3.38590801e-01 3.55711937e-01 9.47771847e-01 -8.94650280e-01
-3.73759776e-01 1.06399608e+00 -5.48858531e-02 -1.81799904e-01
5.13451099e-01 -8.45120009e-03 1.14855838e+00 -1.16117537e+00
7.00285017e-01 9.87623632e-01 2.91251183e-01 4.73383874e-01
-1.14192462e+00 -8.30280602e-01 1.94763571e-01 -8.86914581e-02
-1.21475041e+00 -5.28861582e-01 7.36904323e-01 -3.78334582e-01
1.17957568e+00 2.77369142e-01 7.57388592e-01 8.75770092e-01
4.62423533e-01 3.10045987e-01 5.35294533e-01 -2.04428598e-01
3.69158685e-01 -2.35554442e-01 -5.29824734e-01 8.89351964e-01
8.92745331e-02 1.03543952e-01 -3.08720350e-01 -4.01750170e-02
5.42413175e-01 -5.01264818e-04 -1.00460447e-01 -1.32077456e-01
-1.40961647e+00 8.28496635e-01 7.90596306e-01 1.74518406e-01
-2.82372832e-01 1.80327475e-01 3.01490247e-01 -4.17074233e-01
2.82236993e-01 1.17874372e+00 -3.62455398e-01 4.04791951e-01
-9.99141514e-01 3.50953313e-03 7.64121115e-01 1.17810249e+00
6.13953352e-01 9.36639383e-02 1.61871910e-01 2.72844166e-01
2.46312216e-01 -2.30154157e-01 2.11642757e-01 -4.50491697e-01
4.65799034e-01 6.44744813e-01 1.17843866e-01 -7.43121505e-01
-7.35640168e-01 3.36328633e-02 -7.21925318e-01 -2.37052038e-01
3.15143704e-01 -1.60094544e-01 -1.05948329e+00 1.24640989e+00
5.47063828e-01 -2.00506747e-01 1.78389885e-02 5.75574100e-01
9.91892517e-01 1.38255596e+00 3.50218922e-01 -5.93941152e-01
1.00371933e+00 -1.42875969e+00 -4.30575997e-01 1.62734047e-01
6.64788544e-01 -9.54064786e-01 3.29240501e-01 6.72472596e-01
-1.21208715e+00 -3.91767651e-01 -1.30754340e+00 -3.39642555e-01
-7.81812549e-01 -7.04673352e-03 9.63583529e-01 6.65268481e-01
-6.86195135e-01 1.04160762e+00 -4.32204098e-01 2.37713903e-02
1.01290360e-01 5.95693111e-01 -2.76516140e-01 2.77813166e-01
-9.98483598e-01 1.01212907e+00 1.22637141e+00 1.50925651e-01
-1.45893312e+00 -8.10503781e-01 -8.38913739e-01 1.12796567e-01
9.14900243e-01 -7.71430373e-01 1.25514781e+00 -8.13241780e-01
-1.81480837e+00 1.88846454e-01 -1.46369515e-02 -2.19481200e-01
2.80725479e-01 4.46882010e-01 -3.23620200e-01 4.82634418e-02
-1.49731403e-02 1.06448102e+00 6.92032397e-01 -1.13203704e+00
-3.68857175e-01 1.49695158e-01 8.09475556e-02 2.37853482e-01
3.02808404e-01 -5.83034642e-02 -2.18721882e-01 -7.81400144e-01
1.80943072e-01 -1.03078902e+00 -8.17375839e-01 -3.04994524e-01
-9.10746813e-01 -2.71579862e-01 1.19363248e-01 -3.17791671e-01
1.24573731e+00 -1.51648736e+00 5.26344955e-01 3.13115358e-01
4.16262656e-01 2.62321383e-01 -1.35462899e-02 1.07540548e+00
-6.42772377e-01 4.41622406e-01 -1.35646850e-01 2.94922918e-01
-1.24644451e-01 -3.80218387e-01 -5.34244716e-01 2.08042562e-01
1.47529379e-01 7.86984682e-01 -1.42930448e+00 -3.02919984e-01
3.37158561e-01 8.33308324e-02 -5.20832360e-01 -2.71816794e-02
-9.79696929e-01 6.09085977e-01 -5.78846395e-01 8.78794789e-01
3.14655602e-01 -3.05935919e-01 8.64499152e-01 -4.63299006e-01
-4.68493074e-01 8.20114911e-01 -8.22853506e-01 1.76864636e+00
-1.66194975e-01 -8.56510848e-02 -6.16845608e-01 -7.72234738e-01
8.72635245e-01 4.30155218e-01 4.90919143e-01 -2.25063384e-01
1.95461854e-01 2.33997852e-01 1.24972612e-01 -1.63881027e-03
9.31569636e-01 -6.12019420e-01 -5.05975727e-03 3.45612824e-01
-2.09166203e-02 -4.36915368e-01 6.99200094e-01 3.22134227e-01
6.73868477e-01 4.60970283e-01 1.04131424e+00 -1.74519479e-01
7.37743020e-01 2.17504680e-01 4.00693834e-01 3.79505575e-01
1.57101378e-01 3.35741043e-01 4.42506582e-01 -6.84299648e-01
-1.26907814e+00 -8.95510077e-01 5.51160455e-01 9.94787872e-01
1.49815127e-01 -1.02058530e+00 -6.36736691e-01 -9.05177712e-01
-4.61473376e-01 7.28254795e-01 -3.50027382e-01 -9.65752229e-02
-3.48469645e-01 -6.01408064e-01 4.32760715e-01 5.20150423e-01
1.47372574e-01 -8.85936260e-01 3.58658195e-01 5.38947761e-01
-7.62439296e-02 -7.07081497e-01 -8.25473487e-01 3.55112493e-01
-6.85674787e-01 -1.19121265e+00 -3.91300142e-01 -7.11314201e-01
8.34095001e-01 5.96572399e-01 7.05994844e-01 -6.24507107e-03
-2.57986844e-01 -3.87613624e-01 -2.39336804e-01 -6.85426891e-01
-8.65956187e-01 2.22583160e-01 -9.61220860e-02 -5.11581540e-01
-3.78100991e-01 -4.39818114e-01 -7.06910968e-01 2.10322484e-01
-8.55245292e-01 4.00801569e-01 5.78332424e-01 3.26957732e-01
8.25965226e-01 3.55792761e-01 6.18159592e-01 -8.67219746e-01
3.78645182e-01 -5.53864598e-01 -9.75851893e-01 4.11382675e-01
-8.98202240e-01 4.46011007e-01 1.28353798e+00 -2.44105294e-01
-1.05368161e+00 6.41717970e-01 -1.74244896e-01 1.78572416e-01
5.86451925e-02 6.42653942e-01 -4.52606469e-01 -1.18832372e-01
5.76086879e-01 1.37529358e-01 -3.29224974e-01 -1.31281912e-01
7.67891526e-01 2.26283502e-02 8.31005722e-02 -6.74307048e-01
8.30426455e-01 1.87131956e-01 6.87811434e-01 -7.26254463e-01
-7.06387460e-01 -3.43669206e-01 -5.89630663e-01 2.40503065e-02
8.84754121e-01 -6.70925498e-01 -1.06279874e+00 -1.55332327e-01
-1.13311696e+00 -2.24147230e-01 1.18932433e-01 1.88702479e-01
-5.37539899e-01 6.38988018e-01 -4.04053420e-01 -3.34333271e-01
-5.58931172e-01 -1.56950366e+00 9.27642286e-01 8.08850825e-02
-3.42501372e-01 -8.38444412e-01 5.74318767e-02 4.02381927e-01
-2.69021511e-01 2.59302288e-01 1.16272247e+00 -7.77152121e-01
-9.95625019e-01 8.66734460e-02 -7.47498348e-02 -3.04534674e-01
5.19326806e-01 3.17757726e-01 -6.46891177e-01 -1.25903646e-02
-7.20606387e-01 -1.69030756e-01 9.16730106e-01 1.77606538e-01
1.36195850e+00 -4.88067806e-01 -6.03615701e-01 4.35341746e-01
1.14096963e+00 7.36739337e-01 5.44247687e-01 2.23630980e-01
8.55693161e-01 7.42738008e-01 6.12596095e-01 3.66091430e-01
5.02833612e-02 3.10185045e-01 7.52360582e-01 1.37536421e-01
1.13265544e-01 -8.94293189e-01 3.91273111e-01 4.08590138e-01
-4.96238500e-01 -7.90835559e-01 -8.01688433e-01 -3.32979709e-02
-1.71556675e+00 -8.53184462e-01 -8.23200196e-02 2.03859282e+00
9.77439404e-01 1.77925900e-01 4.98166382e-01 1.57933757e-01
5.93688905e-01 3.04519743e-01 -4.14278656e-01 -5.07063866e-01
3.96634430e-01 5.49829841e-01 7.50842631e-01 5.47077298e-01
-9.16113973e-01 1.61229146e+00 6.74665737e+00 8.38810861e-01
-1.06905854e+00 -5.53628147e-01 4.49518651e-01 1.55161619e-01
-3.69430810e-01 5.09355545e-01 -8.32246900e-01 2.36118332e-01
7.79928863e-01 -3.72181267e-01 6.30932868e-01 7.32025921e-01
3.82197708e-01 1.62754819e-01 -1.42719138e+00 5.18383026e-01
-2.22717628e-01 -2.16074514e+00 5.57395697e-01 -1.29141241e-01
7.15649486e-01 -5.54080009e-01 -1.91476271e-01 -5.04366197e-02
3.59476089e-01 -1.10586774e+00 9.49508607e-01 6.39387667e-02
6.31985307e-01 -9.67034221e-01 -2.81452667e-02 9.61522013e-02
-1.49622965e+00 -1.09512324e-03 -1.69598594e-01 8.63210931e-02
9.40949544e-02 2.77390361e-01 -1.63058424e+00 1.02711165e+00
-2.05191523e-01 8.05276692e-01 -3.50988805e-01 8.00085008e-01
-8.03722680e-01 2.47582838e-01 7.88920969e-02 -3.66506100e-01
5.26892424e-01 -6.42165899e-01 2.79118121e-01 1.06800365e+00
3.72753471e-01 2.27090538e-01 3.72094780e-01 9.03921604e-01
-5.02897501e-01 2.61555076e-01 -6.16449952e-01 -8.12215984e-01
4.48778808e-01 1.38054895e+00 -1.31065381e+00 -4.66418207e-01
-6.18753470e-02 7.57192135e-01 -4.88625169e-02 3.31944019e-01
-9.76309717e-01 -3.47724289e-01 1.13588341e-01 -3.82278003e-02
2.28060037e-01 -3.01018089e-01 1.16378739e-01 -8.22777271e-01
-5.21068573e-01 -9.85010564e-01 3.09810936e-01 -7.68280268e-01
-6.04526699e-01 3.45675796e-01 -2.21925676e-01 -9.48906004e-01
9.71621368e-03 -7.32100487e-01 -6.85551524e-01 4.68605250e-01
-1.25851846e+00 -1.06955528e+00 1.29972294e-01 1.47943676e-01
8.09181929e-01 6.27637133e-02 7.60377526e-01 -5.96559234e-02
-7.42879272e-01 6.66978136e-02 -1.42250165e-01 -4.85080659e-01
8.31653297e-01 -1.28426850e+00 5.67771137e-01 8.33530426e-01
-3.66640091e-02 8.85951698e-01 7.86776483e-01 -9.53333199e-01
-1.59473073e+00 -1.39003265e+00 7.92838335e-01 -1.49087727e-01
9.48657930e-01 -4.53870356e-01 -1.49552777e-01 6.90981209e-01
8.00991580e-02 -5.28216064e-01 8.57982516e-01 -1.74335197e-01
-3.97539854e-01 7.37053752e-02 -7.86133528e-01 1.10300910e+00
1.07479417e+00 -1.33585945e-01 -3.78671646e-01 8.50327671e-01
8.75622213e-01 -4.72764879e-01 -8.66876543e-01 5.74895740e-02
4.08020616e-01 -5.76477706e-01 1.27077472e+00 -6.85879171e-01
7.81286955e-01 -4.33401495e-01 2.12676704e-01 -1.13800585e+00
-4.82225001e-01 -1.29685569e+00 7.00960457e-02 7.20691621e-01
9.40784037e-01 -3.93153071e-01 8.04501593e-01 3.87558460e-01
-5.94953120e-01 -4.74768162e-01 -2.64835566e-01 -8.05026591e-01
-5.86762466e-02 7.05541521e-02 8.01920533e-01 8.67509842e-01
5.19634068e-01 9.24668849e-01 -3.68467927e-01 7.09467456e-02
1.94869816e-01 1.81095585e-01 5.73540986e-01 -1.06067050e+00
-6.73759356e-02 -6.33768797e-01 1.99914947e-01 -1.01502371e+00
2.73846835e-01 -1.39604020e+00 7.23733455e-02 -1.57037652e+00
1.04473181e-01 -3.51870477e-01 -7.97324181e-02 5.04184306e-01
-6.84729889e-02 -1.49968460e-01 4.75868806e-02 -3.92836705e-03
-5.68966866e-01 4.63320971e-01 1.33991337e+00 -3.63380969e-01
-4.40915048e-01 -1.08628690e-01 -8.66960645e-01 6.58776164e-01
9.12393987e-01 -5.52267134e-01 -7.05525994e-01 2.62894452e-01
8.77537847e-01 3.30512375e-01 -2.22843677e-01 -4.78468806e-01
2.58591086e-01 -6.20391607e-01 2.56066233e-01 -8.31546962e-01
2.59386808e-01 -6.66417897e-01 4.89056319e-01 7.46630788e-01
-3.26892972e-01 -2.92116869e-02 1.43031269e-01 6.63996577e-01
2.03592151e-01 -4.91364062e-01 5.08394301e-01 -5.05590141e-01
-3.74824524e-01 7.69663155e-01 -8.80665958e-01 -5.87329030e-01
1.04620636e+00 -1.66454867e-01 -4.51892704e-01 -4.03745770e-02
-6.62949562e-01 8.93269554e-02 4.87380445e-01 1.90698192e-01
5.12393653e-01 -9.87770557e-01 8.51150379e-02 -3.94719094e-01
9.93163511e-02 1.35431111e-01 -3.46784413e-01 6.10239565e-01
-9.32897925e-01 8.42312515e-01 1.42291695e-01 8.33185390e-02
-1.18404686e+00 1.30241740e+00 1.40086249e-01 -3.91336322e-01
3.90679725e-02 7.95235932e-01 3.87617201e-01 -3.08860481e-01
9.29961056e-02 -5.37693083e-01 -9.97695476e-02 1.35428816e-01
3.76072347e-01 4.46726084e-01 1.52669370e-01 -2.62621224e-01
-3.55823874e-01 2.16361806e-01 -3.43782276e-01 2.85548240e-01
1.24782622e+00 4.28192586e-01 -2.07317337e-01 -2.37689406e-01
6.98422968e-01 8.84936526e-02 -8.72573256e-01 2.71475434e-01
-2.90319286e-02 7.94982761e-02 -6.81589693e-02 -8.01409543e-01
-5.55458188e-01 6.08257473e-01 -4.19917375e-01 1.49559736e-01
8.64419639e-01 -1.56414315e-01 8.11326623e-01 8.42020988e-01
9.49608162e-02 -9.31561828e-01 3.43427837e-01 3.38048041e-01
8.93224299e-01 -1.09621978e+00 5.17881215e-01 -9.24924850e-01
-3.45435798e-01 1.57070327e+00 3.11842471e-01 2.68345237e-01
5.62453717e-02 -2.91412950e-01 -7.21845686e-01 -1.64161667e-01
-5.73796809e-01 -4.91821729e-02 3.10618758e-01 3.11645210e-01
6.68965757e-01 -1.59987405e-01 -5.49515963e-01 2.63980985e-01
-2.21658677e-01 -2.53343344e-01 5.22052526e-01 9.57068682e-01
-6.14357769e-01 -1.52490008e+00 -1.45912796e-01 -2.51798779e-01
-3.10123324e-01 -7.22060502e-01 -1.23637271e+00 7.36531377e-01
1.86350018e-01 1.03786433e+00 -7.30544269e-01 -3.97652984e-01
6.27214238e-02 1.72718745e-02 9.01799917e-01 -9.09027994e-01
-6.82229638e-01 2.86182642e-01 6.08881176e-01 -4.46804672e-01
-3.90669018e-01 -2.25095123e-01 -1.57357872e+00 -2.87293971e-01
-5.25033295e-01 5.20095050e-01 6.69211447e-01 9.49512780e-01
2.06794262e-01 5.49977243e-01 8.90001893e-01 -1.00371718e+00
-1.60268396e-01 -4.15012985e-01 -1.45501480e-01 -3.50773513e-01
-1.24499891e-02 -4.25663620e-01 -2.76035685e-02 1.61454141e-01] | [4.502028942108154, 6.103488922119141] |
74a34e56-2dab-4940-9379-4072b7ca02ab | towards-personalized-cold-start | 2306.17256 | null | https://arxiv.org/abs/2306.17256v2 | https://arxiv.org/pdf/2306.17256v2.pdf | Towards Personalized Cold-Start Recommendation with Prompts | Recommender systems play a crucial role in helping users discover information that aligns with their interests based on their past behaviors. However, developing personalized recommendation systems becomes challenging when historical records of user-item interactions are unavailable, leading to what is known as the system cold-start recommendation problem. This issue is particularly prominent in start-up businesses or platforms with insufficient user engagement history. Previous studies focus on user or item cold-start scenarios, where systems could make recommendations for new users or items but are still trained with historical user-item interactions in the same domain, which cannot solve our problem. To bridge the gap, our research introduces an innovative and effective approach, capitalizing on the capabilities of pre-trained language models. We transform the recommendation process into sentiment analysis of natural languages containing information of user profiles and item attributes, where the sentiment polarity is predicted with prompt learning. By harnessing the extensive knowledge housed within language models, the prediction can be made without historical user-item interaction records. A benchmark is also introduced to evaluate the proposed method under the cold-start setting, and the results demonstrate the effectiveness of our method. To the best of our knowledge, this is the first study to tackle the system cold-start recommendation problem. The benchmark and implementation of the method are available at https://github.com/JacksonWuxs/PromptRec. | ['Ninghao Liu', 'Xiao Huang', 'Wenlin Yao', 'Huachi Zhou', 'Xuansheng Wu'] | 2023-06-29 | null | null | null | null | ['sentiment-analysis'] | ['natural-language-processing'] | [-1.47142053e-01 -3.50996256e-01 -6.70063615e-01 -3.60392660e-01
-2.71277457e-01 -7.56990075e-01 4.75166917e-01 2.17700839e-01
-2.89019823e-01 2.96166718e-01 4.01117593e-01 -3.86735737e-01
-8.76461416e-02 -6.43087983e-01 -2.51503468e-01 -3.11276138e-01
3.33346665e-01 1.70024097e-01 -1.40910089e-01 -5.94910502e-01
4.51701522e-01 -6.80508511e-03 -1.52262449e+00 4.23533708e-01
9.80205536e-01 8.28724027e-01 2.87279099e-01 4.16083664e-01
-3.84514987e-01 8.36673796e-01 -1.29322201e-01 -5.09169221e-01
3.53940368e-01 -6.06026053e-02 -6.83120549e-01 -2.72034913e-01
-7.53965080e-02 -3.04870695e-01 -9.01159495e-02 7.17474282e-01
3.84325087e-01 5.24447381e-01 2.59525061e-01 -1.07651794e+00
-8.94400656e-01 7.15832829e-01 -1.98900253e-01 1.77706495e-01
6.84796691e-01 -1.59253314e-01 1.42615855e+00 -9.93296146e-01
5.04620194e-01 5.12894094e-01 4.83013809e-01 5.43721914e-01
-9.30779517e-01 -6.90383971e-01 7.61154354e-01 1.23839580e-01
-1.00577641e+00 -3.27999502e-01 6.51632786e-01 -6.49576962e-01
1.05564475e+00 4.49583232e-01 6.52205169e-01 1.25872469e+00
-8.36650655e-02 9.00100172e-01 8.07758808e-01 -2.36251608e-01
2.41261318e-01 6.90588355e-01 7.50088096e-01 -5.59190102e-02
7.45893717e-02 -1.70764819e-01 -3.42030585e-01 -1.59013197e-01
1.56257689e-01 9.01451945e-01 -1.66333050e-01 -1.47673870e-02
-8.11090231e-01 6.41279936e-01 2.06610352e-01 7.49973834e-01
-4.86386716e-01 -6.68746889e-01 1.50863439e-01 5.53639114e-01
7.33810127e-01 9.01309073e-01 -7.63591111e-01 -3.06898594e-01
-7.54505277e-01 2.27688149e-01 1.29486001e+00 1.07070506e+00
6.55060053e-01 -2.60487884e-01 -3.16157266e-02 7.64427483e-01
3.30303341e-01 5.09204865e-01 7.36831963e-01 -3.82752061e-01
2.44121432e-01 7.01308966e-01 4.64691490e-01 -9.74704683e-01
-3.56345266e-01 -6.51693702e-01 -4.47284520e-01 -6.07104719e-01
3.23161006e-01 -4.21908528e-01 -4.72839922e-01 1.47712803e+00
2.88741827e-01 3.66026014e-01 2.01266333e-01 8.29034209e-01
8.39420617e-01 6.87225699e-01 8.22546799e-03 -6.90413535e-01
1.19829011e+00 -1.01072121e+00 -8.63292098e-01 -3.13103467e-01
8.00730467e-01 -9.61701632e-01 1.32820892e+00 5.62828839e-01
-5.78959107e-01 -4.71863806e-01 -6.91564083e-01 2.75830477e-01
-5.34940124e-01 -4.93196677e-03 6.88382030e-01 4.57206100e-01
-5.21727383e-01 4.56581354e-01 -5.22476614e-01 -6.80014670e-01
3.40646282e-02 2.82223761e-01 7.56665841e-02 -5.82973212e-02
-1.30742252e+00 7.17887342e-01 -8.28366354e-02 1.04628056e-01
-3.01509261e-01 -8.05719078e-01 -3.10663104e-01 -5.54310568e-02
7.85471976e-01 -3.53412658e-01 1.62906873e+00 -1.34160316e+00
-1.55013108e+00 1.09704204e-01 -1.90680549e-01 -1.52517244e-01
3.49020325e-02 -6.85151637e-01 -8.52851570e-01 -6.96354866e-01
-1.48530796e-01 -5.17218351e-01 5.10502934e-01 -1.02589560e+00
-8.05679023e-01 -3.05562407e-01 3.93311679e-01 2.11385593e-01
-8.25594485e-01 1.18455619e-01 -6.36099398e-01 -4.21100080e-01
-2.88235575e-01 -1.04606974e+00 -3.69244337e-01 -1.09661794e+00
-1.51795819e-01 -2.73513019e-01 4.99747485e-01 -4.51028287e-01
1.90764070e+00 -2.03980851e+00 -1.65010288e-01 4.43337500e-01
1.76904067e-01 4.72912252e-01 -5.43530881e-02 8.33431184e-01
1.53288141e-01 2.04445824e-01 5.42700768e-01 -1.33143157e-01
-5.80143277e-03 -1.79652423e-02 -6.51525557e-01 -1.88541063e-03
-4.51982141e-01 7.81274021e-01 -1.06373489e+00 8.99264216e-02
-3.44287045e-02 2.37441048e-01 -7.58240461e-01 5.80477655e-01
-3.22211683e-01 5.02707303e-01 -8.04812729e-01 5.30273616e-01
2.84170657e-01 -5.63136518e-01 5.73248088e-01 -2.13058576e-01
-2.38914207e-01 3.69151741e-01 -1.06685781e+00 1.43880141e+00
-7.84822345e-01 -3.89991738e-02 -4.21827137e-01 -7.95700014e-01
8.49670589e-01 3.54037791e-01 6.71284497e-01 -8.67790222e-01
7.55282491e-03 1.99329779e-01 2.19500735e-02 -8.08283806e-01
7.58380294e-01 -1.60827115e-02 -7.40909055e-02 6.84074819e-01
-1.20896637e-01 4.56714451e-01 1.76268801e-01 3.61330628e-01
1.06488729e+00 9.75917503e-02 4.05854613e-01 1.25280261e-01
4.42645222e-01 -1.93689063e-01 6.14409208e-01 7.95123219e-01
9.61665213e-02 1.80106610e-01 8.23327377e-02 -3.67581248e-01
-7.03991532e-01 -4.71123040e-01 1.45959184e-01 1.58455384e+00
1.33607998e-01 -1.05228317e+00 -2.89283991e-01 -7.77938247e-01
-8.33420604e-02 8.16633165e-01 -5.16284943e-01 -6.32881001e-02
-3.11538160e-01 -6.32476628e-01 -2.40899831e-01 3.39451879e-01
-1.14639990e-01 -9.10159826e-01 -6.53148144e-02 2.83364654e-01
-1.72090039e-01 -9.28358018e-01 -5.35571992e-01 -5.57309166e-02
-7.94826686e-01 -9.73283470e-01 -4.52517539e-01 -4.51883674e-01
5.95137358e-01 5.91715217e-01 1.11640930e+00 2.25435361e-01
3.45597565e-01 4.37751293e-01 -9.73170817e-01 -2.74490565e-01
5.05927438e-03 4.09083277e-01 4.41879988e-01 1.95094243e-01
8.50699246e-01 -6.60226047e-01 -7.67685175e-01 5.09171724e-01
-7.16174901e-01 2.96532065e-02 4.21322465e-01 6.99096680e-01
4.65390384e-01 -5.90709075e-02 8.56332004e-01 -1.57680929e+00
7.82591045e-01 -1.17377651e+00 -3.35895389e-01 3.78095955e-01
-1.25050473e+00 -2.77750522e-01 1.08188105e+00 -6.01137996e-01
-1.11548710e+00 -1.71551287e-01 -2.12678999e-01 -4.46555875e-02
-6.08326457e-02 9.68352854e-01 2.85153333e-02 3.61169934e-01
5.98973811e-01 -1.13752693e-01 -6.32137895e-01 -8.39965820e-01
4.81797218e-01 9.82900143e-01 8.72984901e-02 -3.11650127e-01
6.88942969e-01 2.33185008e-01 -8.49709153e-01 -5.02104044e-01
-1.38116741e+00 -1.09096599e+00 -3.97041291e-01 -2.13726401e-01
3.23229641e-01 -8.43870640e-01 -6.64368987e-01 2.69093122e-02
-5.80439091e-01 -1.23997763e-01 -2.73588985e-01 6.67006195e-01
-1.75453633e-01 1.10818237e-01 -4.00673926e-01 -9.61100936e-01
-6.14153326e-01 -7.64934659e-01 3.24929327e-01 4.96533245e-01
-4.45456773e-01 -1.01284516e+00 4.03431833e-01 4.44058508e-01
5.24825573e-01 -3.53637189e-01 5.08038461e-01 -1.40977490e+00
-1.95682213e-01 -5.74553788e-01 2.85480738e-01 2.67192334e-01
4.10874426e-01 -1.52677104e-01 -8.55549872e-01 -4.36635643e-01
1.35470271e-01 2.58465447e-02 3.53245974e-01 7.55278487e-03
8.80293548e-01 -5.72935343e-01 -3.69404644e-01 2.87838042e-01
1.36719501e+00 3.14033061e-01 1.91794991e-01 3.33936542e-01
8.34383667e-01 4.49990332e-01 1.04108274e+00 8.04907262e-01
6.68168426e-01 7.63254762e-01 6.29004687e-02 8.87799717e-04
4.70088005e-01 -3.59074354e-01 4.71067935e-01 1.13336122e+00
-1.85782641e-01 -5.42497635e-01 -9.44319069e-01 4.13598090e-01
-2.19213700e+00 -8.88446450e-01 -2.22499520e-02 2.24214053e+00
7.06635535e-01 4.57010418e-02 3.21995199e-01 -1.56340331e-01
2.93113142e-01 -1.08847350e-01 -6.01150453e-01 -3.58628273e-01
3.27257901e-01 6.49217814e-02 2.03818172e-01 3.06343943e-01
-8.78370166e-01 9.42243874e-01 5.00356722e+00 3.72489899e-01
-1.15028811e+00 2.22605690e-01 2.65329540e-01 -4.58653957e-01
-5.41046321e-01 3.09228480e-01 -9.77678537e-01 6.39370084e-01
1.15336430e+00 -3.19682717e-01 6.14268839e-01 8.05473983e-01
4.36978519e-01 6.84133768e-02 -1.23672032e+00 8.60274851e-01
2.15729430e-01 -9.33642805e-01 -3.81119370e-01 4.92678024e-03
9.38843906e-01 2.07307070e-01 2.09805876e-01 6.19311988e-01
4.35556948e-01 -6.14059806e-01 2.69502401e-01 9.33511853e-01
1.96197405e-01 -6.41569853e-01 8.32571566e-01 6.22396052e-01
-9.79842663e-01 -4.76778239e-01 -1.38075918e-01 -3.29061449e-01
8.83383304e-02 7.34269261e-01 -9.24286306e-01 6.41328216e-01
7.32912123e-01 1.21468484e+00 -3.92192602e-01 8.79236937e-01
-1.09766804e-01 9.38922882e-01 -9.35122296e-02 -1.81417599e-01
-1.05595984e-01 -3.99218678e-01 3.24090987e-01 1.04903102e+00
6.04704201e-01 3.46073776e-01 4.06900376e-01 4.19716746e-01
-3.12801301e-02 8.80288184e-01 -6.96680486e-01 -2.00825170e-01
4.38174337e-01 1.54287505e+00 -4.13228661e-01 -2.81140894e-01
-8.55067670e-01 7.11928248e-01 2.08501592e-01 5.08205771e-01
-5.22229910e-01 -1.07767969e-01 6.82227314e-01 3.73411536e-01
2.42105663e-01 2.42526233e-02 -2.25101262e-01 -1.57302129e+00
4.08543646e-02 -1.06629348e+00 4.23556685e-01 -3.10012162e-01
-1.61080587e+00 5.70895016e-01 -3.33310276e-01 -1.43368530e+00
-4.16272581e-01 -6.96242675e-02 -6.17608428e-01 5.91754913e-01
-1.37883472e+00 -1.02325094e+00 -1.97439626e-01 5.45922399e-01
4.87107158e-01 -3.18034381e-01 8.58396590e-01 6.41428351e-01
-8.03193152e-01 6.92326427e-01 5.59205234e-01 -2.32802451e-01
9.75142181e-01 -1.05279779e+00 1.92290366e-01 7.77616382e-01
2.74659693e-01 1.16591275e+00 7.47907698e-01 -7.92824149e-01
-1.73164833e+00 -9.10453677e-01 1.18099761e+00 -8.40399623e-01
8.10544848e-01 -5.40279746e-01 -9.19767976e-01 6.51889980e-01
1.08789444e-01 -2.36229032e-01 1.44429016e+00 9.43564653e-01
-3.13740581e-01 -2.92350739e-01 -6.43563688e-01 6.40426934e-01
8.67333770e-01 -5.31277716e-01 -5.45131862e-01 4.46935296e-01
6.93924248e-01 -1.70681372e-01 -1.00341237e+00 1.10576011e-01
8.13717484e-01 -5.69186330e-01 6.28350258e-01 -8.80736649e-01
2.22452015e-01 -2.41826147e-01 -2.15486243e-01 -1.32580495e+00
-4.39613372e-01 -8.19934487e-01 -4.35629487e-01 1.50052369e+00
7.17617333e-01 -4.60251868e-01 4.61711884e-01 1.06632435e+00
2.61801071e-02 -8.12063992e-01 -9.03887078e-02 -4.80472088e-01
-2.19351396e-01 -4.98687565e-01 8.01913381e-01 1.12566888e+00
4.43135500e-01 4.56255674e-01 -8.17290604e-01 -2.36898996e-02
-8.67678672e-02 6.32170856e-01 1.02853358e+00 -1.29551685e+00
-5.14164269e-01 -8.78588259e-02 3.24223578e-01 -1.06029403e+00
1.06478959e-01 -9.94886696e-01 -2.37845704e-02 -1.36990786e+00
2.65984207e-01 -5.92040241e-01 -7.89645910e-01 5.95638752e-01
-2.02507675e-01 3.44340317e-02 1.82364121e-01 5.14318824e-01
-9.99170601e-01 1.50167927e-01 9.44022894e-01 1.85545906e-01
-7.92679667e-01 5.62946916e-01 -1.31050956e+00 6.65509164e-01
9.66439009e-01 -3.95784795e-01 -7.58580506e-01 -1.35171056e-01
9.69641805e-01 -2.35965356e-01 -2.95988798e-01 -4.93520945e-01
3.57263803e-01 -4.95401710e-01 -1.43117279e-01 -2.21817926e-01
-7.39150867e-02 -1.00486553e+00 2.59115994e-01 -1.32594183e-01
-5.44618070e-01 4.98812757e-02 -8.02300572e-02 5.52790642e-01
-5.08690067e-03 -1.56622529e-01 2.15068802e-01 1.36721572e-02
-7.02812731e-01 4.38928545e-01 -3.76731902e-01 -3.06950975e-02
8.98171782e-01 -2.99183764e-02 -2.43575051e-01 -5.20072937e-01
-7.47381806e-01 3.34560663e-01 4.46971834e-01 9.02642667e-01
3.89060766e-01 -1.15662134e+00 -4.23481435e-01 1.32379040e-01
4.54820752e-01 -5.08840024e-01 3.89481664e-01 1.00664401e+00
3.73155117e-01 4.31151569e-01 1.88772798e-01 -1.42144132e-02
-1.03651619e+00 7.42345452e-01 -9.40639991e-03 -4.89161104e-01
-3.90577704e-01 6.65185988e-01 1.27454251e-01 -3.71564597e-01
8.56040418e-02 -1.46621227e-01 -7.69178271e-01 3.71942014e-01
6.19141638e-01 8.57956707e-02 8.55180919e-02 -5.43726265e-01
-1.98336437e-01 2.34850347e-01 -7.06916809e-01 2.58360803e-01
1.58820641e+00 -4.28099126e-01 4.80582826e-02 7.53520966e-01
8.89124155e-01 3.62173170e-01 -6.71494722e-01 -7.67495871e-01
1.91961721e-01 -6.35784328e-01 7.89422467e-02 -1.03331161e+00
-1.05940032e+00 4.11273599e-01 5.95916569e-01 4.35314775e-01
1.07856154e+00 -8.69131535e-02 1.05311430e+00 5.88621855e-01
2.78639704e-01 -1.33064914e+00 1.00834183e-02 7.06134140e-01
4.68261272e-01 -1.34398770e+00 -8.33619386e-02 3.66468690e-02
-8.61006320e-01 7.01096535e-01 8.77454638e-01 1.30108058e-01
1.11892295e+00 7.73243979e-03 1.43350706e-01 -5.75064048e-02
-9.99534786e-01 -2.93954521e-01 4.74328399e-01 6.11923076e-02
8.76777112e-01 1.90522343e-01 -5.33881962e-01 1.49866748e+00
-1.81484416e-01 2.55555928e-01 4.23054129e-01 9.92689967e-01
-1.44049689e-01 -1.54912817e+00 2.16038018e-01 8.29277456e-01
-6.66332126e-01 -3.26863319e-01 -2.85782725e-01 1.07436471e-01
6.19847737e-02 1.26908886e+00 -3.40676934e-01 -8.45153570e-01
6.16279185e-01 1.88035667e-01 -2.15087280e-01 -1.00095749e+00
-1.03618598e+00 2.09975094e-01 7.20469952e-02 -6.51469171e-01
-3.24283779e-01 -6.85833991e-01 -1.00953102e+00 -2.42556542e-01
-7.02686489e-01 5.00436842e-01 4.73950207e-01 9.48167503e-01
5.59495628e-01 2.69323051e-01 1.13046455e+00 -4.14630026e-01
-4.67814565e-01 -8.53359818e-01 -5.81172884e-01 6.04657531e-01
2.53480613e-01 -4.13361400e-01 -2.18877807e-01 -7.44243860e-02] | [10.149100303649902, 5.699491500854492] |
72f8fa8f-ab11-4283-96ec-571c0c031b6e | compressive-quantization-for-fast-object | null | null | http://openaccess.thecvf.com/content_iccv_2017/html/Yu_Compressive_Quantization_for_ICCV_2017_paper.html | http://openaccess.thecvf.com/content_ICCV_2017/papers/Yu_Compressive_Quantization_for_ICCV_2017_paper.pdf | Compressive Quantization for Fast Object Instance Search in Videos | Most of current visual search systems focus on image-to-image (point-to-point) search such as image and object retrieval. Nevertheless, fast image-to-video (point-to-set) search is much less exploited. This paper tackles object instance search in videos, where efficient point-to-set matching is essential. Through jointly optimizing vector quantization and hashing, we propose compressive quantization method to compress M object proposals extracted from each video into only k binary codes, where k<< M. Then the similarity between the query object and the whole video can be determined by the Hamming distance between the query's binary code and the video's best-matched binary code. Our compressive quantization not only enables fast search but also significantly reduces the memory cost of storing the video features. Despite the high compression ratio, our proposed compressive quantization still can effectively retrieve small objects in large video datasets. Systematic experiments on three benchmark datasets verify the effectiveness and efficiency of our compressive quantization.
| ['Tan Yu', 'Zhenzhen Wang', 'Junsong Yuan'] | 2017-10-01 | null | null | null | iccv-2017-10 | ['set-matching', 'instance-search'] | ['computer-vision', 'computer-vision'] | [ 2.60504067e-01 -7.50999153e-01 -6.78605020e-01 -1.75690934e-01
-9.24516916e-01 -4.74868089e-01 2.99112469e-01 4.95294273e-01
-2.85782725e-01 2.18011022e-01 2.72603091e-02 1.10386029e-01
-3.54311019e-01 -5.82049370e-01 -7.55067110e-01 -7.12217093e-01
-1.70901477e-01 1.79867610e-01 5.23828268e-01 3.68752390e-01
6.90344751e-01 4.30342525e-01 -1.81443143e+00 8.36454481e-02
2.75867552e-01 1.55974889e+00 6.37141764e-01 6.12617552e-01
8.91229808e-02 6.17844403e-01 -4.68677670e-01 -1.95379376e-01
4.56813723e-01 -2.47885093e-01 -4.87837195e-01 1.83099061e-01
7.12879360e-01 -7.49739408e-01 -9.15928960e-01 1.25164962e+00
4.59719151e-01 1.14151865e-01 3.96935046e-01 -1.42625153e+00
-9.10997450e-01 1.42835259e-01 -6.23658776e-01 3.93576711e-01
5.74716806e-01 3.46440412e-02 8.88051271e-01 -1.05300081e+00
6.60151124e-01 1.15613401e+00 2.55022883e-01 5.70139773e-02
-6.54375613e-01 -6.36660337e-01 -4.21951056e-01 8.09758782e-01
-2.15814495e+00 -5.68018854e-01 5.67294717e-01 -1.87045768e-01
9.01432037e-01 3.61581802e-01 7.97955453e-01 1.24521600e-02
4.31932323e-02 7.08647072e-01 4.85816777e-01 -2.63678402e-01
1.47916704e-01 -4.71863374e-02 -3.97848904e-01 8.23629618e-01
2.76953995e-01 -3.69464494e-02 -7.13668346e-01 -3.40139002e-01
8.34912121e-01 4.69380409e-01 -6.15141809e-01 -3.31534266e-01
-1.33752823e+00 7.41354764e-01 5.01277447e-01 2.25179195e-01
-4.85477895e-01 6.20243967e-01 3.65214229e-01 2.87387103e-01
-4.39880759e-01 -2.15191558e-01 9.69307348e-02 -1.38326302e-01
-1.37150776e+00 2.36940607e-01 4.14516538e-01 1.27061069e+00
9.14320469e-01 -3.06529641e-01 -2.66182929e-01 6.14079356e-01
4.15204138e-01 8.93506289e-01 4.58572924e-01 -1.09285522e+00
5.16604483e-01 3.65059823e-01 1.70556411e-01 -1.58256841e+00
4.37440485e-01 -1.02657735e-01 -7.37320244e-01 -2.61188149e-01
-1.43106818e-01 5.09981811e-01 -6.10302269e-01 1.19965863e+00
4.64466393e-01 3.21770161e-01 -9.84095559e-02 1.25059891e+00
9.00163352e-01 8.90301049e-01 -2.57975072e-01 -4.46510762e-01
1.44757259e+00 -9.38054144e-01 -6.22676075e-01 -2.35847216e-02
4.16869521e-01 -9.73946214e-01 4.15527791e-01 -1.79405920e-02
-1.38384521e+00 -6.54292464e-01 -1.05385900e+00 -8.22994038e-02
8.40559304e-02 6.01924434e-02 2.64532454e-02 4.57488030e-01
-9.27440345e-01 1.83907598e-01 -5.21797240e-01 -1.59590378e-01
4.06621248e-01 6.28770888e-01 -5.59252501e-01 -5.82917571e-01
-9.94023681e-01 3.24120790e-01 6.29761934e-01 -2.88778245e-01
-8.69511127e-01 -3.40740353e-01 -7.21101403e-01 4.48525846e-01
3.74599457e-01 -4.74602193e-01 8.83078456e-01 -6.64410949e-01
-6.68782413e-01 7.81345725e-01 -4.41893756e-01 -4.53000873e-01
-7.75339827e-03 1.04716033e-01 -2.94480383e-01 9.54682708e-01
2.60026664e-01 9.25878882e-01 1.28080571e+00 -9.75180805e-01
-7.93601096e-01 -3.45272303e-01 -2.80574322e-01 2.60883212e-01
-5.66821814e-01 1.32332578e-01 -1.05120540e+00 -5.37169397e-01
4.74014193e-01 -7.27566957e-01 9.84024778e-02 6.01170957e-01
-4.47860844e-02 -3.58006895e-01 1.24265301e+00 -4.90173489e-01
1.42685819e+00 -2.40357852e+00 -6.60749152e-02 2.37242848e-01
1.94374144e-01 2.98065454e-01 -7.72168636e-02 6.38332605e-01
2.66366571e-01 -1.33071676e-01 2.81646308e-02 -1.78805918e-01
-1.88409954e-01 1.04915380e-01 -2.17708513e-01 7.64348447e-01
-1.93806633e-01 1.11308599e+00 -7.84703195e-01 -1.22641742e+00
2.08675236e-01 4.41557735e-01 -5.93387663e-01 2.12840080e-01
2.37380758e-01 -1.10698454e-01 -4.89760369e-01 1.01173484e+00
8.22098315e-01 -5.05418837e-01 -1.28976077e-01 -4.65978324e-01
7.35590905e-02 -2.89654911e-01 -1.07922101e+00 1.64208400e+00
1.23083897e-01 5.30999780e-01 5.56610711e-02 -9.30329561e-01
8.44994962e-01 2.46512413e-01 5.99296331e-01 -8.79320860e-01
-1.32952958e-01 4.17017967e-01 -3.42631459e-01 -5.86745977e-01
7.48389482e-01 1.48720860e-01 8.49768892e-03 2.57159382e-01
-1.40952408e-01 -8.79265666e-02 2.06515491e-01 3.04829746e-01
9.29786086e-01 -3.86245817e-01 4.65515584e-01 3.46010700e-02
7.38753378e-01 -1.49024464e-02 3.68289948e-01 7.56290734e-01
-3.19094062e-01 5.69342732e-01 -1.65033266e-01 -2.66203374e-01
-1.08518875e+00 -9.09332454e-01 -1.28790095e-01 6.65003896e-01
9.79127586e-01 -6.89256907e-01 -4.73743707e-01 -2.73836732e-01
2.87494421e-01 -8.50746185e-02 -2.13855058e-01 -1.42651975e-01
-5.22580624e-01 1.15084626e-01 3.66762698e-01 2.36476272e-01
5.30134022e-01 -8.97698224e-01 -9.63796079e-01 1.77524835e-01
-2.32090041e-01 -1.21115148e+00 -1.07991195e+00 -3.18668246e-01
-9.43572104e-01 -9.44716513e-01 -9.53469217e-01 -1.30450201e+00
4.71579552e-01 9.98546124e-01 6.91339850e-01 6.28026903e-01
-5.29551148e-01 4.33214039e-01 -4.89323795e-01 1.46637216e-01
-1.33965373e-01 -3.80952835e-01 9.64149460e-02 -2.70962305e-02
4.40811068e-01 -3.40383381e-01 -1.22834170e+00 6.08213782e-01
-1.17698598e+00 -4.65912104e-01 5.03042996e-01 5.92705607e-01
8.81162405e-01 3.56276482e-01 1.13096714e-01 3.07244241e-01
2.05638185e-01 -4.16049272e-01 -5.13618112e-01 3.80284578e-01
-5.32017767e-01 -1.17080711e-01 2.95142710e-01 -4.42877054e-01
-9.95039195e-02 9.85745639e-02 2.49858171e-01 -1.10903239e+00
1.79845870e-01 4.06945497e-01 9.56992507e-02 -4.74704832e-01
8.36163834e-02 9.07720625e-01 1.94734428e-02 -2.42594153e-01
2.97963560e-01 8.62054765e-01 8.18148851e-01 -1.82762101e-01
8.48441362e-01 5.96442401e-01 1.14685275e-01 -7.49541044e-01
9.26156044e-02 -9.07667160e-01 -4.24250573e-01 -2.03218848e-01
6.54409766e-01 -1.09852195e+00 -9.48189914e-01 1.18491285e-01
-1.01720190e+00 4.30317461e-01 -6.19560666e-02 5.56019008e-01
-5.83489597e-01 9.60148752e-01 -4.33060408e-01 -6.21122539e-01
-5.43265879e-01 -1.36309981e+00 1.47165775e+00 1.31207213e-01
1.72966063e-01 -4.56740469e-01 -2.92550743e-01 2.65122503e-01
2.84271210e-01 -2.48654068e-01 8.29755843e-01 -9.08889771e-02
-1.20492208e+00 -5.57654202e-01 -5.72806120e-01 -6.34370148e-02
9.10076499e-02 -3.12119663e-01 -2.92697340e-01 -7.12679148e-01
-3.73193622e-02 -4.19347167e-01 8.06233346e-01 2.54680127e-01
1.28874934e+00 -5.53886473e-01 -5.15820444e-01 6.97146297e-01
1.85460734e+00 2.90814430e-01 8.12320530e-01 2.08360523e-01
5.28871119e-01 2.74788924e-02 1.17121482e+00 8.45927358e-01
2.30324402e-01 9.34204876e-01 6.81821823e-01 4.82506067e-01
-1.47790298e-01 -3.32008868e-01 2.51783729e-01 9.57650244e-01
3.52556318e-01 -3.17995787e-01 -4.17679071e-01 7.94445634e-01
-1.65886569e+00 -1.16623318e+00 1.41463250e-01 2.42410207e+00
6.69852018e-01 -1.36778146e-01 -3.07799056e-02 2.79931217e-01
9.19289112e-01 2.83974022e-01 -4.47880775e-01 -1.62759256e-02
-1.26659453e-01 -5.60050383e-02 5.86888373e-01 1.47588640e-01
-9.07721579e-01 6.42204225e-01 5.54942894e+00 1.45035434e+00
-1.08176100e+00 1.65289178e-01 3.09805036e-01 -2.66019106e-01
-1.03752822e-01 1.47234902e-01 -8.46742868e-01 6.54371798e-01
5.46507418e-01 -3.58448833e-01 3.97590607e-01 9.67060685e-01
-1.08867608e-01 -3.05894375e-01 -1.06253469e+00 1.82986295e+00
3.74390990e-01 -1.61017609e+00 4.02351230e-01 2.75409311e-01
7.32591808e-01 -2.38987580e-01 2.01454982e-01 -1.22637466e-01
-7.37926662e-01 -9.47622478e-01 1.01224506e+00 2.93940395e-01
1.21796536e+00 -7.04034209e-01 5.38121819e-01 2.97536850e-01
-1.75655353e+00 -3.43085140e-01 -7.24371016e-01 3.98282021e-01
2.76556671e-01 3.38145912e-01 -6.53349698e-01 2.31170252e-01
1.03448629e+00 7.11537600e-01 -3.57409656e-01 1.44549465e+00
5.73785424e-01 1.47318184e-01 -3.73080760e-01 -5.97134605e-02
3.73572916e-01 -7.71518424e-02 5.81362247e-01 9.87393200e-01
8.54746699e-01 2.67866284e-01 3.80234689e-01 3.68314445e-01
-1.80823281e-01 4.45841789e-01 -7.43436635e-01 -1.72025338e-01
9.38672185e-01 8.80201042e-01 -7.25044429e-01 -5.95287502e-01
-3.94545674e-01 1.26573336e+00 -6.64897636e-02 -3.00497469e-02
-6.88968480e-01 -6.37393415e-01 5.15178204e-01 2.05445752e-01
1.06010664e+00 -2.69626170e-01 4.33059961e-01 -7.09806681e-01
2.71153331e-01 -9.13592398e-01 2.84562886e-01 -8.76342595e-01
-7.23842621e-01 2.02919438e-01 -3.14476378e-02 -1.63610780e+00
-1.48884237e-01 -1.10594958e-01 -1.09385513e-01 6.01229250e-01
-1.27267039e+00 -8.78987908e-01 -6.26374960e-01 8.81587744e-01
5.37648141e-01 -1.85460746e-01 4.76081580e-01 5.46538889e-01
1.33001670e-01 7.43481755e-01 4.24978167e-01 3.77217345e-02
6.21308088e-01 -4.51109171e-01 1.05451286e-01 3.98822397e-01
1.47217974e-01 6.79082811e-01 3.78580749e-01 -7.19236672e-01
-2.02820134e+00 -9.30331409e-01 7.54319847e-01 6.99140579e-02
3.90045434e-01 1.45149171e-01 -9.08670783e-01 5.90487421e-02
9.02220905e-02 3.79642040e-01 4.78746712e-01 -1.07106030e+00
-3.79189074e-01 -2.95504510e-01 -1.21862423e+00 3.54957700e-01
9.21729267e-01 -7.71927714e-01 -3.56602132e-01 4.73900765e-01
7.74897635e-01 -3.03858161e-01 -1.12897241e+00 4.09914464e-01
7.32249439e-01 -9.53831732e-01 1.48059618e+00 1.25690280e-02
3.14150482e-01 -5.77375054e-01 -9.16420698e-01 -4.13304895e-01
-3.54694903e-01 -4.08397377e-01 -5.14156103e-01 8.04738462e-01
-4.24297422e-01 -1.80107579e-01 8.66761029e-01 -4.80160601e-02
4.29899186e-01 -9.02688622e-01 -1.18887615e+00 -8.16600978e-01
-6.37032509e-01 -5.25554270e-02 8.20953906e-01 6.44288003e-01
-1.85236394e-01 -1.82034642e-01 -6.03021920e-01 3.43891531e-01
9.22114432e-01 5.41966915e-01 6.05504215e-01 -8.53892386e-01
-3.54863286e-01 -3.03725004e-01 -1.08231342e+00 -1.72097957e+00
-1.97449073e-01 -8.70286107e-01 5.12487777e-02 -1.31337893e+00
5.97541273e-01 -4.76765811e-01 -1.06690027e-01 1.11168906e-01
-1.17468521e-01 7.44288504e-01 3.44708860e-01 8.06925356e-01
-9.84896302e-01 6.67810202e-01 1.23604739e+00 -3.40723544e-01
2.13405341e-01 -4.08218294e-01 -2.00065091e-01 6.17734678e-02
4.04917747e-01 -7.11437523e-01 -3.29975903e-01 -4.07309622e-01
-1.14185356e-01 6.69680417e-01 5.15715539e-01 -1.09305942e+00
6.15785122e-01 -1.66791692e-01 2.58150250e-01 -9.66132104e-01
7.01454759e-01 -9.52501655e-01 2.73015559e-01 7.93613911e-01
-9.19276029e-02 3.76280367e-01 -1.27957731e-01 9.38605130e-01
-5.58846056e-01 -5.99961102e-01 6.24179900e-01 -1.45707414e-01
-1.02279043e+00 6.85589433e-01 4.10300344e-02 -3.23628396e-01
1.20538533e+00 -7.34011531e-01 -1.05321810e-01 -6.35898232e-01
-1.96474567e-01 2.71076858e-01 7.05548882e-01 3.46091390e-01
1.24355006e+00 -1.61371565e+00 -5.54096401e-01 2.89144605e-01
3.68807405e-01 -2.55692095e-01 3.82220358e-01 6.92747474e-01
-7.62004256e-01 7.80884564e-01 -7.01668561e-02 -1.00658083e+00
-1.77459931e+00 9.33642566e-01 -9.21229199e-02 4.21071559e-01
-5.11183143e-01 8.50419819e-01 3.61546166e-02 4.32726890e-01
4.03823197e-01 1.27989694e-01 1.94448009e-01 -2.22003207e-01
8.32234800e-01 3.70622724e-01 -2.17852294e-01 -1.09963000e+00
-5.14023542e-01 1.11811161e+00 -5.75012192e-02 4.71484428e-03
9.06541586e-01 -3.94771695e-01 -2.86574543e-01 -1.09541662e-01
1.86204743e+00 -1.02034070e-01 -9.20111418e-01 -5.16351104e-01
-2.51805335e-01 -1.28816807e+00 1.20636694e-01 1.31141350e-01
-1.18392003e+00 7.61328578e-01 8.77278924e-01 -4.28998321e-02
1.26821733e+00 1.70725673e-01 1.35845089e+00 5.68648279e-01
7.22034454e-01 -8.63871336e-01 3.61170202e-01 -8.11718926e-02
7.17227995e-01 -1.30890620e+00 4.25348133e-01 -3.26081723e-01
-3.40230405e-01 1.06542587e+00 2.26217002e-01 -1.04967363e-01
6.69265807e-01 -1.69865727e-01 -3.05876166e-01 -2.02966422e-01
-6.04455531e-01 -5.64408908e-03 3.10886949e-01 4.50360060e-01
-2.84136057e-01 -2.07161367e-01 -1.51894912e-01 -1.66554362e-01
8.92033130e-02 1.34261400e-01 9.52491537e-02 1.25756097e+00
-1.00328135e+00 -9.21170056e-01 -7.36056507e-01 4.81953502e-01
-3.47420424e-01 -3.31675380e-01 1.69243887e-01 3.93912286e-01
-1.72779206e-02 9.55877542e-01 1.86345011e-01 -3.66246343e-01
-1.77297547e-01 -3.90481502e-01 5.63184321e-01 -2.25910798e-01
-1.78492114e-01 2.95368850e-01 -6.92629039e-01 -5.99660635e-01
-4.10942316e-01 -5.10499001e-01 -1.25058568e+00 -3.37909132e-01
-3.94537807e-01 3.07415962e-01 7.43620098e-01 4.88145053e-01
3.89873117e-01 -3.10992450e-01 7.81640768e-01 -8.77432466e-01
-7.85188615e-01 -4.33637470e-01 -5.95375896e-01 5.27695119e-01
5.50915956e-01 -5.33593953e-01 -1.95394576e-01 1.48317367e-01] | [10.121065139770508, 0.5733950734138489] |
02c1c323-a820-47a5-a7ee-c10a36a1a48e | road-redesign-technique-achieving-enhanced | 2302.07440 | null | https://arxiv.org/abs/2302.07440v1 | https://arxiv.org/pdf/2302.07440v1.pdf | Road Redesign Technique Achieving Enhanced Road Safety by Inpainting with a Diffusion Model | Road infrastructure can affect the occurrence of road accidents. Therefore, identifying roadway features with high accident probability is crucial. Here, we introduce image inpainting that can assist authorities in achieving safe roadway design with minimal intervention in the current roadway structure. Image inpainting is based on inpainting safe roadway elements in a roadway image, replacing accident-prone (AP) features by using a diffusion model. After object-level segmentation, the AP features identified by the properties of accident hotspots are masked by a human operator and safe roadway elements are inpainted. With only an average time of 2 min for image inpainting, the likelihood of an image being classified as an accident hotspot drops by an average of 11.85%. In addition, safe urban spaces can be designed considering human factors of commuters such as gaze saliency. Considering this, we introduce saliency enhancement that suggests chrominance alteration for a safe road view. | ['Dongsoo Har', 'TaeYoung Kim', 'Medhavi Mishra', 'Sumit Mishra'] | 2023-02-15 | null | null | null | null | ['image-inpainting'] | ['computer-vision'] | [ 6.02188051e-01 6.30255401e-01 -3.14463854e-01 -1.11640714e-01
-6.03215337e-01 -1.41709819e-01 3.77831697e-01 1.86824083e-01
-7.51191258e-01 6.65247977e-01 3.30743015e-01 -6.07262313e-01
-1.34921327e-01 -1.03022265e+00 -7.91948378e-01 -6.90335691e-01
3.05539012e-01 -3.06568980e-01 4.03549105e-01 -8.26827362e-02
5.85668266e-01 7.55533576e-01 -1.81881416e+00 1.21422306e-01
1.12138259e+00 5.99032283e-01 4.87253577e-01 5.32499909e-01
1.08547151e-01 5.56280971e-01 -5.53777933e-01 -2.56324977e-01
3.10257703e-01 -2.99007297e-01 -7.19379723e-01 3.15100372e-01
3.72018725e-01 -5.83545506e-01 -7.45519042e-01 9.67644453e-01
3.11300665e-01 4.52057868e-01 7.75762320e-01 -1.13184178e+00
-2.28394791e-01 1.27313778e-01 -6.09115064e-01 5.23870647e-01
1.66455343e-01 3.64864856e-01 4.82956171e-01 -7.52041996e-01
3.54019403e-01 1.13411760e+00 8.83421972e-02 1.86565280e-01
-1.02095425e+00 -4.48181003e-01 1.62362307e-01 5.74222803e-01
-1.47161365e+00 -2.77119786e-01 9.36314225e-01 -1.93850175e-01
7.10107088e-01 7.03735828e-01 8.94241810e-01 3.40269148e-01
4.30308163e-01 6.51321650e-01 8.48069966e-01 -3.38880092e-01
4.59750295e-02 1.25542343e-01 -1.66156098e-01 5.13089836e-01
1.85378581e-01 1.00447215e-01 -2.00802654e-01 2.04036400e-01
6.96332157e-01 3.19987237e-01 -4.55643177e-01 6.65384680e-02
-1.03811753e+00 5.64140141e-01 8.01264048e-01 2.99335122e-02
-8.34226131e-01 -5.70839830e-03 1.32135168e-01 -1.11167789e-01
6.19102061e-01 3.45862329e-01 1.91665828e-01 -1.68186054e-01
-9.83339608e-01 9.52562168e-02 -1.72799513e-01 7.88332939e-01
1.16543877e+00 -2.02289149e-02 -3.41424108e-01 6.90186024e-01
-9.38837305e-02 7.40260959e-01 -1.41615301e-01 -1.35505939e+00
3.92883927e-01 6.00138903e-01 1.53992146e-01 -1.06021845e+00
-3.22616726e-01 -1.48337334e-01 -5.95502257e-01 5.15106976e-01
1.44950897e-01 -1.17331594e-01 -9.83249485e-01 1.18179584e+00
3.09001833e-01 1.90957800e-01 -4.51854169e-01 8.33007812e-01
2.28408769e-01 6.41015351e-01 3.88702005e-01 -1.16780996e-02
1.64920926e+00 -9.21816230e-01 -9.91751313e-01 -4.54630971e-01
5.27128696e-01 -7.44373143e-01 1.24528480e+00 7.95068964e-02
-1.15062952e+00 -4.05591190e-01 -7.54226267e-01 -1.28980815e-01
-5.90717494e-01 -5.86871356e-02 8.77142251e-02 7.01723576e-01
-9.20154393e-01 2.08451122e-01 -2.63145477e-01 -3.99450094e-01
5.27223527e-01 1.42489761e-01 -1.41934738e-01 -2.51243055e-01
-1.23880851e+00 1.35843396e+00 8.79783034e-02 -7.36955414e-03
-6.86378062e-01 -8.48733008e-01 -8.15180361e-01 2.28842744e-03
4.59393203e-01 -6.21751547e-01 9.38278437e-01 -9.13891196e-01
-9.08192456e-01 8.19063663e-01 -4.75753188e-01 -5.03949285e-01
7.25158453e-01 -4.74395037e-01 -5.74889421e-01 6.60055101e-01
2.59275109e-01 8.79424751e-01 1.28143525e+00 -1.51823556e+00
-1.23217332e+00 -6.75897375e-02 2.16061890e-01 5.84128082e-01
1.22118499e-02 6.04840890e-02 -4.55009848e-01 -7.10824609e-01
-1.95781335e-01 -8.44098568e-01 -3.30443501e-01 3.41958284e-01
-4.93863612e-01 1.73839033e-02 1.17333555e+00 -9.45331275e-01
1.52623856e+00 -2.40158343e+00 -4.53967601e-01 4.53685403e-01
4.20183510e-01 3.37110847e-01 4.96500134e-02 1.40660526e-02
4.27864753e-02 1.66447833e-01 -4.51640368e-01 2.48195659e-02
-4.30566490e-01 1.76557481e-01 -2.49248996e-01 4.60109800e-01
3.78264189e-01 7.27765739e-01 -9.26545858e-01 -5.91675699e-01
7.61756837e-01 5.72721362e-01 -1.69370756e-01 6.89015985e-02
4.38763440e-01 3.71220291e-01 -2.73036033e-01 4.38954026e-01
1.02172124e+00 4.22933608e-01 -6.24161959e-01 -1.31483972e-01
-7.35556066e-01 3.09832036e-01 -6.81887388e-01 1.01403522e+00
-5.96712768e-01 9.56007242e-01 2.86500854e-03 -4.81128186e-01
6.85710847e-01 2.50262264e-02 2.44072676e-01 -1.10671711e+00
-1.69680089e-01 -8.44081864e-02 -2.06969395e-01 -6.99839950e-01
9.16269660e-01 5.40457428e-01 1.62791237e-01 4.59656417e-01
-8.98613036e-01 -1.52963758e-01 7.86332600e-03 2.44602755e-01
9.08177972e-01 -4.36010331e-01 -2.64717728e-01 -3.57569978e-02
2.94158489e-01 8.64770114e-02 1.80752829e-01 7.38412738e-01
-2.62646109e-01 9.24510837e-01 2.53160626e-01 -3.75547737e-01
-1.21854484e+00 -1.08866453e+00 -8.81538019e-02 8.63255680e-01
3.95282596e-01 1.93775147e-02 -1.21294689e+00 -5.27586639e-01
-3.56736720e-01 1.29161251e+00 -7.58888245e-01 -5.95548391e-01
-6.98133111e-01 -3.78860861e-01 3.35358799e-01 1.87546328e-01
8.80967915e-01 -1.07258499e+00 -1.23328149e+00 2.45970145e-01
-5.23591518e-01 -8.41999471e-01 -7.63589203e-01 -1.96038023e-01
-5.04673779e-01 -1.07598937e+00 -1.28494513e+00 -7.30079234e-01
1.17597520e+00 1.07759535e+00 7.58978903e-01 3.23060334e-01
-3.46899062e-01 3.56527150e-01 -2.25098640e-01 -4.26258832e-01
-3.63119781e-01 5.17706834e-02 -4.33935642e-01 2.09618047e-01
3.43837142e-01 -2.12408364e-01 -1.24593151e+00 3.93541485e-01
-7.54659772e-01 3.79877150e-01 6.10270500e-01 2.03556851e-01
5.81279218e-01 5.57781994e-01 3.00334275e-01 -4.62681890e-01
4.52580690e-01 -6.10349298e-01 -2.59728372e-01 -4.41884845e-02
-4.88654554e-01 -4.51951802e-01 1.79232255e-01 -2.30427593e-01
-1.22203982e+00 -1.81710273e-01 9.85446423e-02 -7.85690024e-02
-6.21782005e-01 -9.93402302e-02 -1.95032492e-01 1.14906825e-01
4.17845815e-01 1.48220479e-01 -1.89631939e-01 -2.45389156e-02
3.26630712e-01 6.99454725e-01 5.57844341e-01 8.18564594e-02
6.95289373e-01 8.21085691e-01 3.35555188e-02 -1.14500868e+00
-5.26134789e-01 -5.08828163e-01 -5.48280537e-01 -8.74552011e-01
1.01809084e+00 -6.76073372e-01 -4.26959753e-01 3.90323013e-01
-1.02990234e+00 -2.62737811e-01 -4.03845727e-01 5.89913905e-01
-3.25273663e-01 2.74295598e-01 9.49075669e-02 -8.21183503e-01
4.55520749e-02 -9.63347435e-01 8.28884184e-01 4.16391194e-01
-4.44889963e-01 -5.62686861e-01 -4.85434115e-01 4.97893572e-01
2.59255648e-01 7.34108984e-02 1.07866848e+00 1.92744717e-01
-7.87038743e-01 -1.18450359e-01 -6.38191521e-01 2.05904543e-01
4.22712117e-01 6.84071034e-02 -9.41445291e-01 2.46135339e-01
-3.14795971e-01 5.00908613e-01 7.84668505e-01 8.40971649e-01
1.21353579e+00 -2.92537570e-01 -4.77282703e-01 1.52683631e-01
7.54095972e-01 4.79435474e-01 1.26310182e+00 5.93718588e-01
8.15593302e-01 1.31307900e+00 1.17797446e+00 3.46056908e-01
5.19767582e-01 4.28744316e-01 6.49638057e-01 -9.31179464e-01
-3.47737730e-01 -2.75198877e-01 1.11729071e-01 -3.28081474e-02
-2.50502646e-01 -1.19881466e-01 -8.52552056e-01 1.15593064e+00
-1.53464186e+00 -1.01374757e+00 -7.19523489e-01 2.43740344e+00
4.52916056e-01 1.91794425e-01 1.47918805e-01 2.95028567e-01
1.10011554e+00 3.70859616e-02 -4.94617730e-01 -4.27694410e-01
-1.41082197e-01 -2.28688896e-01 1.04221749e+00 7.20641375e-01
-8.50391328e-01 8.04594755e-01 5.93474579e+00 9.57763016e-01
-9.32383299e-01 9.48720947e-02 1.07185948e+00 -1.74225867e-02
-6.43294096e-01 6.39571100e-02 -3.81040275e-01 6.10785902e-01
7.40608513e-01 3.22823673e-02 2.76032090e-01 4.73633528e-01
1.11637104e+00 -1.01826370e+00 -2.03423694e-01 7.56187737e-01
-4.79975045e-02 -1.11597574e+00 -5.76194655e-03 1.64733648e-01
5.58822572e-01 -3.57435495e-01 4.23873901e-01 -2.73104250e-01
-3.15633237e-01 -8.07424188e-01 8.12030733e-01 6.82995498e-01
6.68416739e-01 -1.37316489e+00 5.36605597e-01 1.15461782e-01
-1.15189838e+00 -7.09949341e-03 -1.26114458e-01 8.44266117e-02
6.32960916e-01 6.52735114e-01 -7.88415015e-01 2.34456286e-01
8.98355722e-01 3.42788547e-01 -7.47693062e-01 1.35905123e+00
-5.49424052e-01 5.21730065e-01 -2.26266205e-01 3.64115447e-01
4.18610632e-01 -2.01780051e-01 6.20672822e-01 1.03070056e+00
4.22557235e-01 8.21153745e-02 -3.86192530e-01 8.28436852e-01
3.54848891e-01 2.60942541e-02 -8.44441235e-01 5.64358950e-01
4.20251369e-01 8.74757409e-01 -8.68085265e-01 -3.44768405e-01
-5.59751153e-01 1.14950430e+00 -3.24558675e-01 6.75428391e-01
-1.00289762e+00 -7.30182290e-01 6.40085459e-01 7.64215469e-01
2.27822706e-01 8.86159614e-02 -6.72863424e-01 -2.64619440e-01
-1.39631582e-02 -3.07170898e-01 -1.30924702e-01 -1.12554181e+00
-3.89279544e-01 4.01171565e-01 2.88911641e-01 -1.38074338e+00
3.11186671e-01 6.74924329e-02 -8.54855657e-01 9.25785363e-01
-1.93409300e+00 -9.22750711e-01 -4.08531189e-01 6.85128748e-01
8.23746979e-01 3.32405329e-01 2.68639863e-01 4.50417578e-01
-5.70642233e-01 2.18050450e-01 -1.72793955e-01 -3.44526142e-01
5.46984315e-01 -1.03410351e+00 5.53851664e-01 9.65366185e-01
-6.33123875e-01 2.70387322e-01 8.41036618e-01 -8.86456907e-01
-5.66310823e-01 -1.44111454e+00 1.21631193e+00 -1.07022189e-01
1.30232453e-01 9.12274048e-02 -1.03143537e+00 2.29170293e-01
3.03181350e-01 -4.60677266e-01 3.36069196e-01 -6.53370202e-01
2.41667047e-01 1.37824520e-01 -1.38583624e+00 1.29600191e+00
9.68410432e-01 -5.71957052e-01 -4.25940841e-01 1.39176607e-01
5.30066550e-01 7.17251450e-02 -3.95706028e-01 -3.18242842e-03
2.29061931e-01 -8.44781518e-01 9.77945626e-01 1.59030691e-01
2.00935632e-01 -5.13763368e-01 5.08532941e-01 -1.18342638e+00
-4.06574368e-01 -7.73507178e-01 3.15672755e-01 8.72751832e-01
3.25964183e-01 -5.31449258e-01 7.31944740e-01 1.06318557e+00
-3.96046132e-01 -3.84194851e-01 -1.00642705e+00 -3.52912158e-01
-2.72341818e-01 -5.03611445e-01 5.25917351e-01 5.18166125e-01
-3.67694348e-01 -3.03177774e-01 -4.48093206e-01 4.24130261e-01
5.69315076e-01 -3.89945358e-01 6.42943263e-01 -7.45923996e-01
6.27275288e-01 -4.87061739e-01 -2.15822697e-01 -1.05385435e+00
-6.02414943e-02 -2.41775960e-01 2.09280048e-02 -1.58543968e+00
-5.02567776e-02 -1.58398286e-01 -1.81347087e-01 4.40309942e-01
-3.93238604e-01 3.19885761e-01 -1.44078150e-01 -2.47525191e-03
-1.52345300e-01 4.22289431e-01 1.49042976e+00 -3.58877480e-01
-2.51343995e-01 1.11523874e-01 -4.62198466e-01 8.03705454e-01
1.16328728e+00 -6.12064958e-01 -5.49562454e-01 -1.96898744e-01
5.46911582e-02 -2.06365258e-01 6.39385164e-01 -8.83933604e-01
8.91545191e-02 -3.24350744e-01 -1.47598600e-02 -9.83371615e-01
2.97961473e-01 -1.07502103e+00 1.45683261e-02 6.22811437e-01
-1.50038794e-01 1.28847912e-01 3.44953686e-01 5.77830970e-01
-7.86265880e-02 -3.07676762e-01 8.54728937e-01 1.71161547e-01
-9.26082492e-01 -1.24007627e-01 -1.48045039e+00 -2.30757073e-01
1.28976333e+00 -8.95597339e-01 -1.29918113e-01 -5.51540911e-01
-4.96360987e-01 2.69817770e-01 5.72046280e-01 3.83240193e-01
9.02101159e-01 -8.77948105e-01 -5.92584372e-01 3.19339246e-01
1.24124624e-01 -1.28797978e-01 7.57003844e-01 1.00940990e+00
-6.25879884e-01 3.28185707e-01 -3.77130419e-01 -2.13124305e-01
-1.39222109e+00 5.21890223e-01 -2.22431906e-02 2.40172595e-01
-7.81855643e-01 4.72689837e-01 5.68898141e-01 4.27590162e-01
5.22374325e-02 -3.49982500e-01 -3.95293564e-01 -1.76973175e-02
6.59703493e-01 9.68661726e-01 -1.57462254e-01 -1.01919961e+00
-2.47761920e-01 7.08426118e-01 -1.23953752e-01 -1.00571193e-01
7.40920365e-01 -8.24145913e-01 2.22703949e-01 8.18818510e-02
9.05655205e-01 4.63572033e-02 -1.70353293e+00 2.23481685e-01
-2.56448925e-01 -8.51081789e-01 3.72591555e-01 -5.18878698e-01
-1.07184196e+00 8.71935904e-01 9.25272346e-01 2.34597117e-01
1.28873289e+00 -2.86898073e-02 1.18241012e+00 -5.93962558e-02
-3.94185707e-02 -1.31075156e+00 -2.48030156e-01 1.10806488e-01
8.82693887e-01 -1.14405060e+00 -8.06125030e-02 -6.36187196e-01
-8.40739548e-01 8.16514373e-01 4.07285482e-01 1.67083798e-03
5.02239227e-01 -2.12859318e-01 1.12883924e-02 -3.26498747e-01
-4.79972251e-02 -4.81848478e-01 3.57058853e-01 8.47359121e-01
-1.34783506e-01 2.59493031e-02 -4.00390089e-01 -1.21679902e-01
1.55994883e-02 -3.68536800e-01 5.63226998e-01 9.21267807e-01
-7.69213498e-01 -6.24117374e-01 -6.29839480e-01 4.22385782e-01
-2.63879478e-01 -1.51933104e-01 -9.72494658e-04 6.69999182e-01
3.92608613e-01 1.20708966e+00 4.08288926e-01 -3.47777866e-02
5.65301418e-01 -2.38178417e-01 -2.41867863e-02 -4.68934298e-01
-4.43003953e-01 -6.03827462e-02 4.11903292e-01 -4.89113241e-01
-3.91857684e-01 -5.24202883e-01 -1.29986787e+00 -4.16518271e-01
-1.73364073e-01 -9.37745795e-02 5.33944964e-01 9.18769300e-01
4.25592810e-01 7.32176006e-01 6.88569188e-01 -8.99368584e-01
6.70425653e-01 -4.29869413e-01 -6.69069648e-01 2.62289077e-01
5.98597527e-01 -6.89309478e-01 -4.57776427e-01 2.39142343e-01] | [8.644720077514648, -1.2019010782241821] |
35b9cef4-b645-41ed-be5c-784046046eb2 | measuring-board-game-distance | 2301.03913 | null | https://arxiv.org/abs/2301.03913v1 | https://arxiv.org/pdf/2301.03913v1.pdf | Measuring Board Game Distance | This paper presents a general approach for measuring distances between board games within the Ludii general game system. These distances are calculated using a previously published set of general board game concepts, each of which represents a common game idea or shared property. Our results compare and contrast two different measures of distance, highlighting the subjective nature of such metrics and discussing the different ways that they can be interpreted. | ['Cameron Browne', 'Éric Piette', 'Dennis J. N. J. Soemers', 'Matthew Stephenson'] | 2023-01-10 | null | null | null | null | ['board-games'] | ['playing-games'] | [-4.57964778e-01 2.98986193e-02 1.83113322e-01 1.36171177e-01
-3.64048928e-01 -1.03362238e+00 5.83700836e-01 3.00009459e-01
-5.52017987e-01 6.88790321e-01 1.15268536e-01 -3.58481586e-01
-7.69769251e-01 -1.13410449e+00 3.39040250e-01 -3.56277406e-01
-2.54031718e-01 3.40342999e-01 7.81983912e-01 -1.09251106e+00
6.31753564e-01 2.05414891e-01 -1.70384252e+00 8.24286370e-04
3.88347089e-01 8.58906090e-01 1.19359858e-01 1.03453171e+00
-3.42442095e-02 1.00552678e+00 -1.19926667e+00 -6.70650721e-01
5.43848753e-01 -6.56269491e-01 -8.18476558e-01 -4.38927382e-01
-1.93293303e-01 -1.46125287e-01 -2.52641737e-01 1.17927706e+00
4.65993643e-01 4.05313432e-01 6.86279953e-01 -1.37184966e+00
-2.22628281e-01 5.82313836e-01 -1.21478669e-01 6.10490263e-01
1.30803514e+00 -1.43460467e-01 1.33836782e+00 -1.90149412e-01
7.05082834e-01 5.07217407e-01 1.01055539e+00 4.13762629e-02
-7.91681945e-01 -4.19164717e-01 -2.30063945e-01 5.92400245e-02
-1.80030429e+00 9.62622464e-02 4.55927432e-01 -7.57435501e-01
1.04251349e+00 7.68274486e-01 1.20628667e+00 1.66410923e-01
3.73588026e-01 2.42094812e-03 1.21427858e+00 -3.98081392e-01
4.63568360e-01 -5.61800413e-02 8.99413675e-02 3.25110078e-01
6.71443224e-01 1.11569658e-01 -3.55498135e-01 -3.46042693e-01
8.93995166e-01 -4.36891407e-01 -1.08169384e-01 -3.66664439e-01
-5.51907957e-01 9.15839195e-01 -2.44615853e-01 6.73911154e-01
-8.89962614e-02 4.23642993e-02 7.52602696e-01 7.70447314e-01
3.71330738e-01 9.44722712e-01 -2.12890431e-01 -1.08698606e+00
-5.44788659e-01 6.81884110e-01 1.11723685e+00 8.55274618e-01
5.16579092e-01 -2.32131541e-01 3.53444934e-01 5.41730881e-01
1.11957625e-01 -2.15116963e-01 6.34395123e-01 -7.19694495e-01
7.38702044e-02 3.77470374e-01 9.35814232e-02 -1.62385368e+00
-4.90088403e-01 -1.90499313e-02 3.26950662e-02 8.64465177e-01
5.48547089e-01 -8.74032155e-02 1.87648401e-01 1.39137840e+00
-1.88708976e-01 -1.10645659e-01 -1.51354983e-01 4.33041781e-01
1.25952280e+00 1.37122542e-01 -1.25373110e-01 2.47179270e-01
1.13470268e+00 -4.58135515e-01 -4.20945972e-01 1.49322599e-01
9.70983148e-01 -7.04805791e-01 9.31509078e-01 6.33500099e-01
-1.40799928e+00 -4.71418142e-01 -1.28275692e+00 2.48731509e-01
-7.58965194e-01 -6.87135220e-01 6.95807636e-01 1.36375403e+00
-1.61863196e+00 8.90062749e-01 -3.18877488e-01 -7.72011638e-01
-3.38519156e-01 3.94738823e-01 -2.69933790e-01 5.11751592e-01
-1.06744492e+00 9.46381211e-01 8.40725124e-01 -5.68196118e-01
-9.78980865e-03 -1.88304007e-01 -7.21115530e-01 -3.58319022e-02
4.42290336e-01 -1.82746395e-01 1.13989711e+00 -2.91639864e-01
-1.81440461e+00 1.22333741e+00 1.05390239e+00 6.54233014e-03
4.14238989e-01 1.24241203e-01 -7.26820350e-01 -1.66451007e-01
2.61028767e-01 3.74097400e-03 -3.76679972e-02 -7.76014566e-01
-9.98733521e-01 2.08216250e-01 9.06462133e-01 5.99689662e-01
-2.87039336e-02 3.67372006e-01 -2.26182222e-01 -5.80216527e-01
-1.31029576e-01 -5.07097006e-01 -2.00382233e-01 -3.27809215e-01
-5.64390942e-02 -6.89023212e-02 -8.25452805e-02 -2.60939986e-01
1.76363647e+00 -2.07598114e+00 1.10416599e-01 4.23098385e-01
6.08576775e-01 -5.84144108e-02 4.65274230e-02 1.05104613e+00
-1.35412857e-01 2.59064734e-01 4.90057826e-01 4.76777434e-01
2.48300552e-01 9.90455225e-02 2.99926072e-01 5.22844076e-01
-7.92267561e-01 6.26866043e-01 -1.18104362e+00 -2.15112656e-01
3.67358327e-01 -2.42508247e-01 -4.47197258e-01 -6.30552173e-02
4.18160081e-01 -1.04633041e-01 -2.09161222e-01 4.01978225e-01
4.32078630e-01 3.62222075e-01 2.75372565e-01 4.85798866e-01
-3.50535363e-01 5.56441844e-01 -1.46080804e+00 1.65043640e+00
-1.13810316e-01 8.98343801e-01 -4.49163049e-01 -7.36457944e-01
1.13376284e+00 2.25803182e-01 6.31874502e-01 -5.89080155e-01
5.85506916e-01 1.78382039e-01 4.54552084e-01 -2.50579238e-01
9.78095055e-01 -1.69876143e-01 -5.93406856e-01 6.78419650e-01
-5.52369431e-02 -6.28831446e-01 4.95032281e-01 2.87753828e-02
1.53947401e+00 -1.73819304e-01 1.36294091e+00 -8.50910842e-01
4.65431690e-01 1.61968153e-02 2.97727399e-02 8.91260564e-01
-4.56064850e-01 4.77813393e-01 8.16729009e-01 -4.84405756e-01
-7.03187823e-01 -1.19623470e+00 1.66561157e-01 1.27111137e+00
6.23459816e-01 -1.40982163e+00 -5.80647111e-01 -2.80827314e-01
-1.51447207e-01 4.09061044e-01 -8.28796983e-01 -1.29487440e-01
-2.85250042e-02 -9.12814364e-02 7.11024046e-01 4.28041220e-01
-4.52323584e-03 -8.83848250e-01 -1.01045227e+00 1.72864199e-01
9.85975340e-02 -6.29070520e-01 -7.35700130e-02 1.36960477e-01
-5.25949597e-01 -1.17772913e+00 -9.85607952e-02 -5.92981219e-01
-1.04105912e-01 6.85798824e-01 1.65752661e+00 4.29208070e-01
-1.80565730e-01 6.09662652e-01 -8.57539058e-01 -3.24990571e-01
-2.69629508e-01 2.33004298e-02 -4.11437033e-03 -9.08067822e-01
5.89873195e-01 -6.92664027e-01 -2.74879068e-01 3.94879222e-01
-7.08247304e-01 -2.72802144e-01 -6.24010563e-02 3.41214657e-01
7.03621656e-02 5.19911349e-01 3.30028273e-02 -6.16896391e-01
1.54004574e+00 -5.79661906e-01 -2.75138915e-01 -1.90015703e-01
-1.89413309e-01 -4.02617246e-01 2.70545155e-01 -2.29288325e-01
-2.72602558e-01 -9.55669820e-01 -2.88276464e-01 6.01938426e-01
-1.35733113e-01 7.21459091e-01 -3.21707249e-01 -5.91090977e-01
7.32932389e-01 -1.64486036e-01 8.55580065e-03 -1.41105190e-01
5.54181039e-01 4.88882452e-01 2.62264729e-01 -5.70190847e-01
6.31435931e-01 2.61868328e-01 -2.06254274e-01 -8.27115297e-01
9.80925709e-02 -7.50808299e-01 -5.74462831e-01 -7.38363028e-01
5.34314454e-01 -4.61858511e-01 -9.26504016e-01 4.44153041e-01
-6.75351381e-01 -2.61464387e-01 -9.13228035e-01 4.49091792e-01
-8.94603252e-01 4.61323231e-01 -4.77578640e-01 -5.87124288e-01
2.58725107e-01 -8.67090762e-01 5.05593300e-01 2.66279131e-01
-9.26924109e-01 -1.48848689e+00 8.47748399e-01 -1.62648231e-01
2.59591639e-01 3.55206072e-01 4.31916744e-01 -8.04530025e-01
3.50621283e-01 -6.77096844e-01 1.03591017e-01 -5.14125563e-02
5.00351727e-01 -1.93059724e-02 -1.46439180e-01 -2.74502695e-01
1.10253125e-01 1.56711891e-01 -3.16060521e-02 1.92262545e-01
5.31384587e-01 1.22746199e-01 -2.71673590e-01 3.32180083e-01
1.85627830e+00 6.90838397e-01 9.85850871e-01 1.10916221e+00
2.35639572e-01 2.89507747e-01 7.44963944e-01 7.37003446e-01
2.45542765e-01 9.02471483e-01 2.21001163e-01 8.27740803e-02
3.25220048e-01 8.60656574e-02 1.79409370e-01 1.20156920e+00
-7.07436919e-01 -4.58737701e-01 -8.95698965e-01 3.89103472e-01
-1.69240212e+00 -1.16281366e+00 -3.19175631e-01 2.18618608e+00
2.27478519e-01 5.30049622e-01 9.02507484e-01 3.84636611e-01
7.36317337e-01 2.49433011e-01 3.00130814e-01 -8.83186162e-01
-5.02091348e-02 7.01617777e-01 6.13552272e-01 5.05227685e-01
-6.64025605e-01 9.14154053e-01 8.63527298e+00 1.22661889e+00
-6.45509601e-01 1.71064496e-01 -5.04840873e-02 -9.67701524e-02
-6.79180175e-02 1.63987845e-01 8.11863318e-02 3.37037832e-01
6.35174096e-01 -9.14541900e-01 2.90335715e-02 7.25383937e-01
-1.77865446e-01 -5.35670519e-01 -7.25737512e-01 1.09606636e+00
-3.84227745e-02 -1.12589705e+00 -2.74106652e-01 3.70385230e-01
4.02613103e-01 -2.66288817e-01 -9.23448429e-02 7.10834563e-02
8.51645648e-01 -1.02041066e+00 8.07417035e-01 5.27032316e-01
7.04584539e-01 -8.61540556e-01 7.50454247e-01 -2.50869215e-01
-1.66593289e+00 3.40886086e-01 -4.34168875e-01 -1.24081695e+00
-6.85438886e-02 -1.16138004e-01 -1.39618427e-01 8.58186781e-01
7.43588567e-01 4.89127725e-01 -5.18111944e-01 1.46283007e+00
-8.59842151e-02 9.59426090e-02 -1.85886055e-01 -4.00534511e-01
4.69933331e-01 -6.15657985e-01 5.80186784e-01 7.52490282e-01
6.43210411e-01 2.26400718e-01 -3.07003986e-02 4.13655698e-01
8.05556476e-01 5.37186623e-01 -1.03209174e+00 2.83772331e-02
5.06611168e-01 1.05193174e+00 -1.34612453e+00 -4.14837971e-02
-7.59302318e-01 8.08219254e-01 5.76722510e-02 -3.23689669e-01
-9.87630010e-01 -9.34491277e-01 1.35629392e+00 2.33791858e-01
2.78863199e-02 -3.35925609e-01 -5.33710361e-01 -7.08561063e-01
-3.18564892e-01 -6.88589036e-01 4.87861156e-01 -5.97245991e-01
-1.10622203e+00 8.00884366e-01 3.00508171e-01 -1.67856622e+00
-2.50691473e-01 -8.86606634e-01 -9.92854118e-01 6.62472725e-01
-2.91381687e-01 -2.27011755e-01 -3.64295661e-01 4.42235500e-01
-3.11291553e-02 -5.66333950e-01 8.18315685e-01 -9.63211339e-03
-1.18061982e-01 4.75504667e-01 3.59723806e-01 -1.38420269e-01
4.14453894e-01 -1.36430848e+00 7.42716551e-01 3.84112418e-01
1.24034733e-01 3.57801497e-01 8.81957352e-01 -3.66295546e-01
-6.29965842e-01 -1.25859067e-01 2.29226917e-01 -5.72602391e-01
1.07398689e+00 -3.10002327e-01 -3.49847198e-01 3.99063438e-01
3.21132332e-01 -7.52239645e-01 1.32562232e+00 4.67599034e-01
-1.02613620e-01 3.23863387e-01 -1.12253535e+00 8.67469192e-01
1.29487634e+00 -4.81617659e-01 -7.70116448e-01 -3.58619094e-02
2.11554915e-01 -2.09066764e-01 -1.19459379e+00 -2.46987119e-01
6.52287602e-01 -1.70956063e+00 8.67377281e-01 -3.93765241e-01
5.86323477e-02 -2.24339679e-01 -2.43823618e-01 -1.56553042e+00
-7.63882637e-01 -8.38100433e-01 6.83369279e-01 8.39390099e-01
8.42098966e-02 -8.39922071e-01 9.49710965e-01 3.77208740e-01
8.74406472e-02 -4.67638075e-01 -9.92841601e-01 -1.43143559e+00
3.71703863e-01 -8.25983763e-01 9.82301831e-01 1.01717007e+00
1.38082743e+00 3.02958619e-02 -5.08615337e-02 -5.61436474e-01
-1.16398763e-02 -5.49988270e-01 8.77974272e-01 -1.69942880e+00
-4.39992070e-01 -1.28599739e+00 -1.83710539e+00 -7.47517824e-01
-2.64566183e-01 -6.50609314e-01 -1.98495358e-01 -1.44433844e+00
-1.30764693e-01 -2.71335900e-01 -2.58880675e-01 -1.40036732e-01
2.46997438e-02 4.90162104e-01 2.62825638e-01 1.48082927e-01
-7.93767810e-01 -2.46920101e-02 7.00005651e-01 2.61263877e-01
-2.13985741e-01 -2.40830272e-01 -1.12951171e+00 9.68344748e-01
9.36571181e-01 -3.94525111e-01 -6.34941697e-01 1.52025670e-01
9.15343285e-01 -2.75409948e-02 -1.72414616e-01 -1.61082673e+00
3.32836986e-01 -2.84025043e-01 -5.33371329e-01 3.83315161e-02
1.89541847e-01 -5.73361754e-01 5.94658792e-01 7.19745278e-01
2.13127062e-01 5.58703780e-01 2.26691216e-01 3.34261805e-01
-3.76596749e-01 -6.64455235e-01 3.42768252e-01 -2.61599332e-01
-1.11428523e+00 -2.23471418e-01 -9.95851457e-01 3.00259084e-01
1.43782198e+00 -1.17829263e+00 -7.67025501e-02 -8.64297330e-01
-7.78117955e-01 -2.74667948e-01 8.40270221e-01 2.41976961e-01
3.96926403e-01 -1.47091293e+00 -5.89248061e-01 -3.69873047e-01
5.24328589e-01 -9.25946176e-01 4.34459373e-02 4.66875702e-01
-1.38887954e+00 1.56909660e-01 -1.03187919e+00 1.59010515e-02
-1.26827145e+00 4.32423949e-01 4.11863953e-01 -3.73529494e-01
-6.79452181e-01 9.59261060e-01 1.54183060e-01 -2.63624191e-01
-3.97327453e-01 -2.44938925e-01 -5.64743519e-01 8.09917971e-02
4.86401618e-01 7.99273968e-01 2.64463555e-02 -8.06836426e-01
-4.11202222e-01 6.36710763e-01 8.99350420e-02 -3.56775224e-01
9.16687787e-01 -1.81209147e-01 -3.19344878e-01 8.74726295e-01
6.29787207e-01 2.62385517e-01 -4.13647979e-01 2.55457401e-01
7.22212419e-02 -7.27561772e-01 -3.42028677e-01 -3.34468156e-01
-7.84539163e-01 -9.90491454e-03 3.42581272e-01 1.20432484e+00
1.10426140e+00 -1.32963276e-02 3.41872305e-01 8.88723210e-02
9.72868681e-01 -1.25937283e+00 -1.02116488e-01 6.63450778e-01
5.52763343e-01 -4.84808028e-01 1.28685147e-01 -7.48989522e-01
-6.13769054e-01 1.21189690e+00 4.07343537e-01 -6.03109956e-01
8.97357643e-01 4.61317509e-01 -2.98383720e-02 -6.39207542e-01
-3.16809297e-01 -6.25748098e-01 -6.97959289e-02 1.18728507e+00
6.32678270e-01 3.08551729e-01 -1.28991103e+00 8.58322740e-01
-1.12184989e+00 -2.92375416e-01 1.19511366e+00 1.27941966e+00
-5.65228939e-01 -1.25260842e+00 -2.06498593e-01 4.65740204e-01
-2.27516025e-01 -9.83450487e-02 -7.75330961e-01 1.41123223e+00
4.35823128e-02 1.30899072e+00 1.23770222e-01 -1.19048369e+00
5.25457025e-01 -4.99104232e-01 7.05775142e-01 -8.28192294e-01
-8.30502212e-01 -5.25565803e-01 2.97044814e-01 -5.32231271e-01
-2.25339919e-01 -6.93530560e-01 -7.28398561e-01 -9.69566822e-01
-6.32928312e-01 5.40748894e-01 1.05477676e-01 8.42006505e-01
-1.93541393e-01 5.05877197e-01 2.06790552e-01 -5.70079505e-01
1.05791874e-01 -7.33904541e-01 -1.53227985e+00 4.05058235e-01
-5.50405741e-01 -9.82766569e-01 -2.58478105e-01 -6.38223767e-01] | [3.4816489219665527, 1.4325557947158813] |
abbcaa78-1b31-48d3-b78b-8f5e234c178e | weakly-supervised-unconstrained-action-unit | 1903.10143 | null | https://arxiv.org/abs/1903.10143v4 | https://arxiv.org/pdf/1903.10143v4.pdf | Unconstrained Facial Action Unit Detection via Latent Feature Domain | Facial action unit (AU) detection in the wild is a challenging problem, due to the unconstrained variability in facial appearances and the lack of accurate annotations. Most existing methods depend on either impractical labor-intensive labeling or inaccurate pseudo labels. In this paper, we propose an end-to-end unconstrained facial AU detection framework based on domain adaptation, which transfers accurate AU labels from a constrained source domain to an unconstrained target domain by exploiting labels of AU-related facial landmarks. Specifically, we map a source image with label and a target image without label into a latent feature domain by combining source landmark-related feature with target landmark-free feature. Due to the combination of source AU-related information and target AU-free information, the latent feature domain with transferred source label can be learned by maximizing the target-domain AU detection performance. Moreover, we introduce a novel landmark adversarial loss to disentangle the landmark-free feature from the landmark-related feature by treating the adversarial learning as a multi-player minimax game. Our framework can also be naturally extended for use with target-domain pseudo AU labels. Extensive experiments show that our method soundly outperforms lower-bounds and upper-bounds of the basic model, as well as state-of-the-art approaches on the challenging in-the-wild benchmarks. The code is available at https://github.com/ZhiwenShao/ADLD. | ['Xuequan Lu', 'Tat-Jen Cham', 'Zhiwen Shao', 'Jianfei Cai', 'Lizhuang Ma'] | 2019-03-25 | null | null | null | null | ['action-unit-detection', 'facial-action-unit-detection'] | ['computer-vision', 'computer-vision'] | [ 3.37192118e-01 2.60878056e-01 -2.69606918e-01 -4.91952509e-01
-1.37271261e+00 -5.64777374e-01 3.10730666e-01 -4.59078342e-01
-3.32311749e-01 5.77714801e-01 -3.68391983e-02 3.39744568e-01
3.77373546e-01 -5.36528945e-01 -7.95391858e-01 -9.78018939e-01
1.80175382e-04 2.89643466e-01 -7.82882273e-02 -1.60530359e-02
-4.04225349e-01 2.74569482e-01 -1.27701223e+00 1.06659174e-01
4.81537610e-01 1.57600594e+00 -3.23928803e-01 1.84217066e-01
3.43901604e-01 5.99715233e-01 -3.42967659e-01 -4.66332793e-01
8.68556976e-01 -6.56826019e-01 -5.52784145e-01 3.48730683e-01
7.60703146e-01 -6.50302708e-01 -2.48872161e-01 1.33615446e+00
5.41986525e-01 1.37749594e-02 5.35396576e-01 -1.65622330e+00
-6.10507309e-01 -1.00492954e-01 -9.89359021e-01 -3.25362742e-01
3.44385564e-01 1.73749283e-01 1.12199569e+00 -1.02323186e+00
6.70895934e-01 1.30082524e+00 4.26908731e-01 1.02419400e+00
-1.20963192e+00 -1.19922018e+00 1.40524492e-01 -7.89221600e-02
-1.72415924e+00 -6.83904946e-01 9.64144349e-01 -4.14389253e-01
2.03330919e-01 -1.99355446e-02 2.83721179e-01 1.16239345e+00
-3.58789146e-01 9.08834338e-01 1.01620078e+00 -2.51884937e-01
9.04399231e-02 1.54183619e-02 -4.59981203e-01 1.31808341e+00
-3.25324178e-01 2.94250220e-01 -4.51159358e-01 -2.86160797e-01
8.51074576e-01 1.23397909e-01 -1.77450925e-01 -5.66224635e-01
-6.80218101e-01 1.00185072e+00 5.64737797e-01 -8.84295106e-02
-1.02012292e-01 2.32369468e-01 3.71166617e-01 1.71644464e-01
6.69404984e-01 -7.22044110e-02 -4.03378189e-01 1.75083220e-01
-6.96262896e-01 1.54925168e-01 3.62412512e-01 1.09676111e+00
1.03549683e+00 -1.40213922e-01 -2.85956025e-01 8.00795555e-01
4.44405764e-01 6.65214777e-01 2.42387891e-01 -1.24802828e+00
2.65621901e-01 6.15846753e-01 2.16296524e-01 -7.57263899e-01
-2.42973492e-01 -6.34736866e-02 -5.32758117e-01 5.93028128e-01
7.76321054e-01 -4.67243314e-01 -8.09782147e-01 2.36510444e+00
7.07985938e-01 2.88884819e-01 -1.30313054e-01 1.07832158e+00
6.39337182e-01 4.31138158e-01 4.82741743e-02 -2.57730484e-01
1.36610782e+00 -1.11690152e+00 -5.20516038e-01 -2.61955053e-01
5.90737820e-01 -5.33179045e-01 1.04612911e+00 1.42817870e-01
-8.50359976e-01 -2.92637199e-01 -7.57477105e-01 -1.02393232e-01
-3.77281159e-02 4.13093984e-01 4.34399575e-01 6.30626678e-01
-8.18700671e-01 1.38478741e-01 -6.87381804e-01 -2.25658238e-01
8.70265484e-01 4.49603707e-01 -7.80274272e-01 -1.87317684e-01
-1.02992654e+00 3.66360486e-01 9.30926651e-02 1.38436913e-01
-1.15074193e+00 -6.98246658e-01 -1.17059267e+00 -2.36974329e-01
8.08533251e-01 -3.61756802e-01 1.34205520e+00 -1.44262469e+00
-1.66116917e+00 1.21101928e+00 -1.00642316e-01 6.91994056e-02
7.32552588e-01 -7.27563202e-02 -3.20924185e-02 3.31666946e-01
3.44074368e-01 7.16633260e-01 1.12370670e+00 -1.17373788e+00
-7.01240063e-01 -6.31780326e-01 1.09748587e-01 2.28562772e-01
-3.41336608e-01 8.64493698e-02 -5.79152942e-01 -4.82221603e-01
-2.14267507e-01 -1.00477350e+00 1.19328097e-01 9.13537085e-01
-2.32507482e-01 -1.97609350e-01 7.75806725e-01 -5.63797116e-01
7.59137392e-01 -2.37845325e+00 1.26481861e-01 9.35593843e-02
3.31194490e-01 1.99813321e-01 -4.09470528e-01 -6.35081306e-02
4.24853601e-02 -2.37990230e-01 -2.66553104e-01 -7.73871720e-01
7.05968365e-02 1.05108336e-01 -5.52642047e-02 8.56540442e-01
3.26288342e-01 8.30383539e-01 -1.26677346e+00 -6.20026469e-01
2.95887757e-02 3.48497391e-01 -4.84854639e-01 3.66342664e-01
-1.74095482e-01 6.59762263e-01 -6.13513172e-01 1.18259597e+00
7.31130064e-01 8.01396370e-02 -1.62324235e-01 -1.81880191e-01
3.51345688e-01 -3.85507554e-01 -1.01032543e+00 1.77271914e+00
-4.64757353e-01 3.58953327e-01 4.64978933e-01 -6.89173996e-01
8.90428185e-01 3.14017087e-01 5.73487699e-01 -6.09579623e-01
3.64533484e-01 2.46292114e-01 -3.74940664e-01 -2.46139467e-01
-1.46524251e-01 -4.08717602e-01 -1.49784029e-01 4.20250565e-01
2.17064604e-01 1.92511007e-01 -1.49292201e-01 -1.15333371e-01
1.03038299e+00 3.42650354e-01 2.99815357e-01 7.48050138e-02
6.31353736e-01 -4.68201518e-01 9.23720360e-01 2.07888797e-01
-7.84982502e-01 7.14398742e-01 6.83155954e-01 -3.01831931e-01
-6.27493203e-01 -1.17165995e+00 -2.13989764e-01 1.33798575e+00
8.92817974e-02 -1.10591859e-01 -8.55697513e-01 -1.33946931e+00
1.06949411e-01 9.92711186e-02 -1.13862765e+00 -2.95074403e-01
-2.29558825e-01 -3.88380736e-01 7.44070530e-01 4.25036043e-01
4.69359666e-01 -9.95326936e-01 -3.11487257e-01 -1.62325948e-01
-8.93079415e-02 -1.16490448e+00 -1.05679774e+00 -1.64298788e-01
-2.71609873e-01 -1.20409930e+00 -1.05044007e+00 -7.21084058e-01
1.01595354e+00 7.29916571e-03 7.15491951e-01 -1.21672429e-01
-3.54048312e-01 3.79806221e-01 -3.01641643e-01 -2.88848102e-01
-2.07407013e-01 -4.02982295e-01 1.29580781e-01 7.80509889e-01
3.90951842e-01 -2.70802885e-01 -7.52852857e-01 5.37349164e-01
-8.20056677e-01 -2.33251527e-01 3.94055873e-01 1.00430191e+00
7.37113178e-01 -4.31734055e-01 4.68900919e-01 -7.67735660e-01
1.55368775e-01 -3.96428883e-01 -7.36247182e-01 9.28441510e-02
-1.52107626e-01 -1.86132655e-01 3.82978976e-01 -5.89347064e-01
-8.35398138e-01 6.40606403e-01 -2.86345743e-02 -9.85681236e-01
-4.47506048e-02 -9.26176086e-03 -6.83146775e-01 -4.39528048e-01
5.88662744e-01 -3.48800160e-02 4.01333869e-01 -1.98311612e-01
3.95054996e-01 5.66077054e-01 4.40623224e-01 -7.03066707e-01
9.25119936e-01 7.56381750e-01 7.25455284e-02 -4.66382474e-01
-1.13626373e+00 -4.50360954e-01 -5.95140338e-01 -2.86072284e-01
8.20220292e-01 -1.09768796e+00 -5.43513417e-01 6.03272319e-01
-1.04038739e+00 -5.16176164e-01 -4.28473800e-01 1.22909129e-01
-7.49017775e-01 3.88742000e-01 -4.20490563e-01 -6.95425808e-01
-3.31903785e-01 -1.26359892e+00 1.65469575e+00 1.83272794e-01
-2.99762847e-04 -7.81854451e-01 8.05212930e-02 3.27880234e-01
-8.12353864e-02 8.14832211e-01 1.87765613e-01 -3.49948972e-01
-3.31356227e-01 -5.79419196e-01 -5.13870656e-01 6.70308590e-01
3.51397187e-01 -2.58391559e-01 -1.12177765e+00 -3.68178934e-01
-1.40836045e-01 -8.87468457e-01 6.64526463e-01 1.23929359e-01
9.14685905e-01 -4.09617156e-01 -1.52645633e-01 8.51835728e-01
1.28037143e+00 -1.82040259e-01 3.63734871e-01 -3.23749892e-02
8.27758789e-01 6.19411290e-01 9.84401286e-01 6.46627188e-01
2.72055805e-01 9.71517444e-01 6.46938562e-01 -1.93907484e-01
-7.58306906e-02 -4.65363324e-01 5.82922876e-01 -2.89686143e-01
9.95215550e-02 6.50335252e-02 -5.41574299e-01 4.36679661e-01
-1.85085475e+00 -7.82504678e-01 4.37171727e-01 2.18618131e+00
1.03181195e+00 -2.99246848e-01 4.85358983e-01 -3.01027149e-01
6.59124017e-01 1.96165234e-01 -7.32935607e-01 -9.23248455e-02
1.25413612e-01 4.26477827e-02 4.75734591e-01 6.28409445e-01
-1.28908229e+00 1.15380883e+00 4.54547310e+00 1.04064453e+00
-9.98204172e-01 5.65919518e-01 7.27312624e-01 -2.55095690e-01
1.25131458e-01 -3.43095720e-01 -8.78067255e-01 5.10938942e-01
4.18603092e-01 -1.78373486e-01 4.11148101e-01 1.19201767e+00
1.31266098e-02 1.33100256e-01 -1.36313391e+00 1.11935198e+00
2.79326439e-01 -8.21626306e-01 -1.77080929e-01 3.17574441e-01
7.75001109e-01 -2.23364681e-01 2.44193196e-01 3.08657944e-01
1.78427130e-01 -8.49246800e-01 7.70491540e-01 1.80382594e-01
1.47144318e+00 -7.07664311e-01 5.59471309e-01 7.47008473e-02
-1.26009667e+00 3.10625006e-02 -3.59060377e-01 5.07368892e-02
-6.29605502e-02 8.51004794e-02 -5.49447358e-01 1.41320378e-01
5.25839984e-01 7.42424905e-01 -2.65756100e-01 6.51239514e-01
-5.35342634e-01 2.85293192e-01 -3.55236560e-01 2.94080555e-01
2.69129932e-01 -2.52725154e-01 3.76561582e-01 9.65085804e-01
1.80807263e-01 1.03768319e-01 3.85495692e-01 8.78287077e-01
-4.58487839e-01 2.39493147e-01 -6.57211006e-01 2.29788601e-01
2.28615388e-01 1.46947825e+00 -3.34404111e-01 8.22361261e-02
-4.92496789e-01 1.40650487e+00 4.85845268e-01 2.74303049e-01
-1.02828753e+00 -1.27342239e-01 1.02508593e+00 1.17852598e-01
7.08927512e-02 1.62981808e-01 1.56631306e-01 -1.07743359e+00
2.00589702e-01 -7.79807210e-01 3.86787415e-01 -4.62842494e-01
-1.41395223e+00 6.27289057e-01 -1.08299911e-01 -1.56112504e+00
-2.54426181e-01 -6.77741170e-01 -6.13385141e-01 8.24632823e-01
-1.51496530e+00 -1.60201097e+00 -4.83633846e-01 1.00012231e+00
2.78748930e-01 -1.47291213e-01 9.72304165e-01 3.64491999e-01
-6.60706639e-01 1.35781479e+00 6.29678695e-03 5.89422107e-01
9.97461259e-01 -1.07052922e+00 -1.86556473e-01 7.03032613e-01
-2.15793885e-02 9.26677957e-02 3.09498399e-01 -2.70117402e-01
-1.05898654e+00 -1.37139201e+00 4.34017658e-01 -5.06558776e-01
7.85617828e-01 -6.69457078e-01 -5.78806937e-01 8.43117058e-01
-3.16382259e-01 8.35797429e-01 7.15043902e-01 -2.35840052e-01
-7.33517051e-01 -3.69918883e-01 -1.49801385e+00 5.60529530e-01
1.07211828e+00 -6.12861335e-01 2.46951655e-02 4.99723703e-01
4.85372305e-01 -5.50148308e-01 -6.64762199e-01 2.77202040e-01
7.40086436e-01 -7.82819986e-01 7.19156504e-01 -4.01045799e-01
3.16822231e-01 -2.81272292e-01 -2.58296996e-01 -9.84430552e-01
-7.16204150e-03 -8.03308845e-01 1.69058479e-02 1.32061160e+00
2.57478714e-01 -5.99613965e-01 1.02313292e+00 7.28892207e-01
1.90083578e-01 -8.60833943e-01 -1.36640728e+00 -7.62710690e-01
1.27888113e-01 -2.90798038e-01 3.39501888e-01 6.90588832e-01
-1.01154275e-01 1.12934716e-01 -6.13212883e-01 3.13942224e-01
9.00156796e-01 6.05642237e-02 8.10996234e-01 -9.40058947e-01
-3.49295825e-01 -3.61011505e-01 -6.91408634e-01 -8.31434727e-01
6.38716400e-01 -7.33265758e-01 3.31934035e-01 -9.03551102e-01
3.09707373e-01 -4.43074554e-01 -2.89895266e-01 9.87643123e-01
-2.11023003e-01 8.42949152e-01 2.33882740e-01 2.84653623e-03
-7.06190109e-01 7.50106812e-01 1.29288566e+00 -2.16143802e-01
-3.46588567e-02 5.12579158e-02 -5.35242319e-01 7.98176110e-01
5.85396707e-01 -5.56968153e-01 -2.93485552e-01 -1.60973623e-01
-1.84656098e-01 3.09954546e-02 4.14250195e-01 -6.73085392e-01
-9.31287855e-02 -1.18340619e-01 1.84553906e-01 1.76301658e-01
5.81024826e-01 -9.22415078e-01 -4.69133466e-01 2.31800660e-01
-2.21844271e-01 -5.68929851e-01 1.54271185e-01 6.63974404e-01
-2.28467062e-01 -6.16232045e-02 1.19192171e+00 5.68349659e-02
-6.86593950e-01 8.30944002e-01 3.11353683e-01 2.83468246e-01
1.43730319e+00 -1.05222955e-01 -9.30666700e-02 -5.60126185e-01
-8.07855844e-01 3.11267972e-01 5.74906945e-01 5.13381302e-01
6.10955834e-01 -1.59203386e+00 -8.47418547e-01 4.43848670e-01
4.27570254e-01 4.59426530e-02 2.25064039e-01 8.63412380e-01
-2.26336226e-01 -1.31196648e-01 -3.42432708e-01 -5.13612866e-01
-1.52726877e+00 5.73839962e-01 5.97551942e-01 -1.10896118e-01
-1.62127987e-01 1.18038499e+00 8.81032288e-01 -4.88098919e-01
4.37688857e-01 1.99151143e-01 2.81246781e-01 -1.56562822e-03
8.01329911e-01 2.25108027e-01 -2.80576617e-01 -1.16818142e+00
-3.25443715e-01 7.78482258e-01 6.86300360e-03 -4.28263061e-02
9.62279856e-01 -1.22987576e-01 9.05870721e-02 7.80377090e-02
1.72775114e+00 3.51905301e-02 -1.77317023e+00 -5.48653483e-01
-4.04209554e-01 -7.18272567e-01 -6.15379214e-02 -6.42183602e-01
-1.32657623e+00 7.75095761e-01 8.26544940e-01 -5.14212072e-01
1.39176464e+00 2.91854978e-01 7.01592326e-01 -5.64943440e-02
6.31691039e-01 -7.26362586e-01 3.92460495e-01 1.40575334e-01
9.95636523e-01 -1.64100838e+00 -3.02181751e-01 -5.44322968e-01
-6.60063326e-01 7.97238290e-01 8.28219891e-01 -2.41478756e-01
6.49753094e-01 8.38666856e-02 3.61256868e-01 -6.28399663e-03
-2.64393926e-01 -4.62039173e-01 3.63595486e-01 6.91052616e-01
1.17298968e-01 7.06327558e-02 -2.35683713e-02 7.00667799e-01
2.29691237e-01 1.13361768e-01 1.11740649e-01 7.39009559e-01
-9.21272635e-02 -1.23483491e+00 -3.15568238e-01 1.20469309e-01
-5.19353271e-01 9.18143466e-02 -5.81609428e-01 8.46005559e-01
4.62407917e-01 7.98002601e-01 2.13945867e-03 -3.21152508e-01
2.02272803e-01 -3.94410491e-02 6.19130194e-01 -7.63413131e-01
-1.29733086e-02 1.92946076e-01 -2.40920737e-01 -9.82508242e-01
-2.19505668e-01 -8.03046405e-01 -1.26141548e+00 -2.16927752e-02
-2.43255928e-01 -1.12317629e-01 2.79020578e-01 7.15425253e-01
2.95036942e-01 -7.60369524e-02 8.84038806e-01 -1.06791842e+00
-6.54455364e-01 -9.29113448e-01 -8.13341677e-01 6.10781670e-01
5.82308114e-01 -8.92723441e-01 -4.29821849e-01 -5.29756434e-02] | [13.62904167175293, 1.539353370666504] |
8f529f89-5abe-435c-990e-da044d745059 | the-self-learning-ai-controller-for-adaptive | 2204.05227 | null | https://arxiv.org/abs/2204.05227v1 | https://arxiv.org/pdf/2204.05227v1.pdf | The self-learning AI controller for adaptive power beaming with fiber-array laser transmitter system | In this study we consider adaptive power beaming with fiber-array laser transmitter system in presence of atmospheric turbulence. For optimization of power transition through the atmosphere fiber-array is traditionally controlled by stochastic parallel gradient descent (SPGD) algorithm where control feedback is provided via radio frequency link by an optical-to-electrical power conversion sensor, attached to a cooperative target. The SPGD algorithm continuously and randomly perturbs voltages applied to fiber-array phase shifters and fiber tip positioners in order to maximize sensor signal, i.e. uses, so-called, "blind" optimization principle. In opposite to this approach a perspective artificially intelligent (AI) control systems for synthesis of optimal control can utilize various pupil- or target-plane data available for the analysis including wavefront sensor data, photo-voltaic array (PVA) data, other optical or atmospheric parameters, and potentially can eliminate well-known drawbacks of SPGD-based controllers. In this study an optimal control is synthesized by a deep neural network (DNN) using target-plane PVA sensor data as its input. A DNN training is occurred online in sync with control system operation and is performed by applying of small perturbations to DNN's outputs. This approach does not require initial DNN's pre-training as well as guarantees optimization of system performance in time. All theoretical results are verified by numerical experiments. | ['G. A. Filimonov', 'A. M. Vorontsov'] | 2022-04-08 | null | null | null | null | ['self-learning'] | ['natural-language-processing'] | [ 2.84685671e-01 1.13210671e-01 3.43482822e-01 -3.17643583e-03
2.80516744e-01 -8.20856690e-01 1.20153137e-01 -4.83724684e-01
-4.89266604e-01 1.26858759e+00 -3.27407867e-01 -7.63179362e-02
-6.72433197e-01 -3.82402927e-01 -6.10971093e-01 -1.12090111e+00
-5.69041027e-03 2.98326492e-01 -3.01449865e-01 -3.31576198e-01
3.97846758e-01 9.40376043e-01 -1.66770649e+00 -5.38020015e-01
6.91500962e-01 9.03009832e-01 4.45495546e-01 1.21059537e+00
4.44835871e-01 2.23080069e-01 -7.34980762e-01 3.60584348e-01
6.79896653e-01 -3.46916765e-01 -6.94794953e-02 1.92289501e-01
5.37776232e-01 -4.28635895e-01 -9.05463547e-02 1.36737478e+00
8.42874408e-01 3.32625881e-02 6.50213361e-01 -9.24408734e-01
-2.26243228e-01 1.66400686e-01 -1.73314139e-01 2.93178678e-01
7.21297488e-02 9.16400194e-01 5.97281337e-01 -4.74541068e-01
5.90342104e-01 7.78759062e-01 2.88696051e-01 5.83940148e-01
-1.39195216e+00 -6.00009024e-01 -4.85584408e-01 4.89697978e-02
-1.24642730e+00 -4.69260722e-01 9.21172202e-01 -6.54438019e-01
8.31215501e-01 1.54106751e-01 9.01431620e-01 3.14582407e-01
5.19473493e-01 -5.71194030e-02 1.13324070e+00 -6.93673670e-01
8.46870095e-02 4.98833120e-01 -1.21024556e-01 8.41696739e-01
2.44021595e-01 7.45875478e-01 -3.23883176e-01 2.09777057e-02
7.45073736e-01 -7.34834313e-01 -8.22020948e-01 8.65789801e-02
-9.01192009e-01 3.99029553e-01 4.36550170e-01 5.07852912e-01
-6.20592535e-01 9.07767117e-02 -4.81358439e-01 8.52450907e-01
1.40915196e-02 1.14697826e+00 -4.20993298e-01 4.34959270e-02
-9.41843510e-01 3.56463820e-01 7.67996013e-01 5.36199152e-01
7.27998376e-01 7.99513400e-01 -2.17519253e-01 2.93441087e-01
4.86206442e-01 1.01406050e+00 5.92334151e-01 -1.07268000e+00
1.57996386e-01 4.72140610e-01 5.86277962e-01 -9.07687604e-01
-7.32064843e-01 -9.98157442e-01 -6.24730647e-01 8.72780383e-01
2.93191671e-01 -9.24886048e-01 -7.76675284e-01 1.71277750e+00
3.15189064e-01 1.54238850e-01 3.44235778e-01 1.26733875e+00
2.17464209e-01 7.21913815e-01 -8.17463517e-01 -9.94885743e-01
8.55295122e-01 -4.61924285e-01 -1.02502096e+00 1.52942672e-01
5.40791126e-03 -8.02152574e-01 5.74465275e-01 7.78544664e-01
-1.12506950e+00 -5.28121114e-01 -1.19311631e+00 5.64972818e-01
-2.23565146e-01 4.19014424e-01 -2.97387630e-01 6.04349852e-01
-1.28218091e+00 9.08451915e-01 -4.57046777e-01 -9.62846726e-02
1.33059612e-02 8.03209722e-01 1.06193423e-01 6.85247123e-01
-9.28175628e-01 8.04150283e-01 3.54882628e-01 4.02172118e-01
-7.82450438e-01 -8.61834228e-01 6.90587312e-02 -5.85594364e-02
9.95562673e-02 -9.48870778e-01 9.68563437e-01 -1.11498022e+00
-2.49725866e+00 3.59937936e-01 -1.12160623e-01 -6.47823393e-01
2.15220079e-01 -1.20651722e-01 -2.12922007e-01 4.07218486e-01
-5.57540059e-01 3.65456879e-01 1.61941397e+00 -1.26508081e+00
-5.81372857e-01 -2.84554362e-01 -3.08848619e-01 1.95463195e-01
-4.42692518e-01 -2.97723174e-01 6.58773482e-01 -1.77383438e-01
1.27882689e-01 -1.03794885e+00 -1.67071193e-01 -1.37422875e-01
-7.73313820e-01 -1.19918771e-01 1.07335997e+00 -2.02198043e-01
1.16050506e+00 -1.64770043e+00 3.34236294e-01 2.28262693e-01
6.28467202e-02 6.11840189e-01 -1.52524803e-02 5.89320421e-01
-1.71356261e-01 -1.89406753e-01 -9.02183056e-02 7.06393272e-02
-3.38021517e-01 -1.30408213e-01 -1.99873149e-01 6.54222250e-01
2.05122411e-01 3.89683783e-01 -6.02693081e-01 -9.95966569e-02
3.56864691e-01 -7.65525736e-03 -5.75578451e-01 5.46746969e-01
-5.91661870e-01 1.00927866e+00 -4.35030967e-01 2.24208385e-01
6.57830536e-01 1.49625897e-01 -2.00934201e-01 -5.64105749e-01
-8.82869363e-01 -3.65963727e-01 -1.22722673e+00 1.15239668e+00
-3.85708898e-01 9.03507948e-01 6.92721248e-01 -1.11917424e+00
1.05158770e+00 5.00288665e-01 5.08407295e-01 -1.93272784e-01
5.16671419e-01 3.26473296e-01 4.60766017e-01 -9.23317552e-01
1.94829136e-01 -2.03076556e-01 7.67033994e-01 2.45172247e-01
3.52818191e-01 -5.82735121e-01 -1.32377362e-02 -3.57089907e-01
8.17329109e-01 -1.32644996e-01 8.19943547e-02 -5.00659645e-01
1.04863763e+00 1.95395947e-01 4.91228044e-01 5.23991585e-01
-3.80791165e-02 9.41348523e-02 1.63845256e-01 -1.03528917e-01
-1.07736814e+00 -6.63230479e-01 -1.83114737e-01 2.43356645e-01
-2.40444094e-02 4.45544243e-01 -6.63097084e-01 1.57125041e-01
-9.79177095e-03 6.40819073e-01 1.24951251e-01 2.99263373e-02
-6.92086279e-01 -3.07242006e-01 7.77854100e-02 -5.06484807e-01
5.41620851e-01 -6.49451733e-01 -7.65464187e-01 4.12076980e-01
6.38421059e-01 -1.10267997e+00 2.57365972e-01 -1.15624815e-01
-9.38074708e-01 -9.94658232e-01 -4.00223672e-01 -3.56688380e-01
6.31684303e-01 -3.70297313e-01 8.06081235e-01 -2.09840387e-01
-3.90211552e-01 6.46692276e-01 8.77894536e-02 -7.57066488e-01
-3.42678249e-01 -1.76834837e-01 7.84220397e-01 2.79264271e-01
-2.97577549e-02 -1.10547984e+00 -7.18664765e-01 9.74549726e-03
-2.17344642e-01 -2.08891883e-01 6.90129042e-01 6.67321622e-01
1.21501610e-01 3.95346045e-01 3.83919835e-01 -2.08141118e-01
8.36218119e-01 -1.24094613e-01 -1.63896215e+00 -2.33773202e-01
-8.02049756e-01 -2.42218319e-02 1.07945979e+00 -3.30014586e-01
-6.40972674e-01 4.67458926e-02 1.38616055e-01 -9.11267459e-01
-7.83100352e-02 1.97506428e-01 1.53568044e-01 -7.07355201e-01
9.35876250e-01 3.28644663e-01 4.14710820e-01 -1.67897448e-01
3.29903997e-02 7.42298722e-01 5.14880419e-01 -3.23142767e-01
1.59202170e+00 1.39425844e-01 6.93359673e-01 -1.30518806e+00
-4.00102228e-01 4.45684157e-02 -3.15177828e-01 -7.20188558e-01
6.91961884e-01 -5.78284681e-01 -1.35152650e+00 5.71990073e-01
-1.34249461e+00 -1.03440478e-01 -2.37311780e-01 9.47807014e-01
-3.46599281e-01 -1.03828296e-01 -2.26323679e-01 -1.35733545e+00
-5.27228594e-01 -9.12808537e-01 5.07326245e-01 8.79081190e-01
4.09811586e-01 -7.28782356e-01 -4.92643472e-03 6.85023218e-02
5.97698689e-01 3.39343518e-01 5.14917970e-01 3.37914899e-02
-1.01998842e+00 -1.88814268e-01 3.77617091e-01 9.81420875e-01
-1.18451551e-01 1.61541432e-01 -1.07917452e+00 -3.95203948e-01
5.08989990e-01 3.96876857e-02 2.80138671e-01 9.14480865e-01
6.35644138e-01 -5.04240751e-01 -4.07310039e-01 5.62668681e-01
1.97982061e+00 3.91558766e-01 1.41621873e-01 2.49502491e-02
4.79624689e-01 5.53168595e-01 4.15490329e-01 4.19411927e-01
-3.09698582e-01 5.91494679e-01 7.70909727e-01 1.29548118e-01
6.01137914e-02 5.74168563e-01 4.08652306e-01 5.05183995e-01
-3.02968740e-01 -6.49032891e-01 -4.21081722e-01 2.04644710e-01
-1.28559065e+00 -9.03760612e-01 -2.89692849e-01 2.22835827e+00
7.17928231e-01 7.53367413e-03 -2.85117030e-01 1.68178737e-01
8.96645784e-01 -2.01540351e-01 -6.94500744e-01 -4.55262601e-01
2.06791013e-02 3.44504446e-01 1.03889120e+00 9.19766247e-01
-5.95771194e-01 5.76484561e-01 5.03885651e+00 4.39152837e-01
-1.77434397e+00 -7.98290148e-02 -1.66838557e-01 -5.05131423e-01
-7.19670802e-02 -2.61821270e-01 -9.58028257e-01 6.32230759e-01
7.29047716e-01 -2.24863052e-01 9.51116920e-01 3.41118932e-01
9.99979377e-01 -1.05867662e-01 -7.49718666e-01 9.97034371e-01
-3.59736830e-01 -1.24754059e+00 -4.20809209e-01 1.07596450e-01
7.28067756e-01 -6.77604824e-02 2.08978191e-01 -2.73301482e-01
-1.73404813e-01 -6.03411138e-01 3.61901253e-01 1.11686969e+00
6.10277593e-01 -3.78248155e-01 5.31225681e-01 6.05095148e-01
-6.19743884e-01 -6.17427170e-01 -2.91704774e-01 -3.27178955e-01
2.98624963e-01 9.63207066e-01 -9.91059780e-01 4.85811830e-01
1.81831643e-01 5.44715226e-01 9.90026370e-02 8.50452900e-01
-1.37574494e-01 8.71618450e-01 -6.92292035e-01 -5.31763196e-01
2.30168402e-01 -7.90631473e-01 1.62359703e+00 4.38581437e-01
6.01387024e-01 2.46880472e-01 -1.69286087e-01 1.23442662e+00
1.87879711e-01 -2.78139383e-01 -6.95903063e-01 -1.37206629e-01
7.17275858e-01 1.44074142e+00 6.17372841e-02 -6.26620948e-02
1.48856685e-01 2.46600032e-01 -3.08862209e-01 6.69197917e-01
-4.43072230e-01 -6.27703488e-01 1.04025948e+00 3.69448423e-01
1.05039567e-01 -5.73395133e-01 1.81913525e-01 -5.78700423e-01
-1.33173885e-02 -3.49111021e-01 -3.76180798e-01 -1.07092500e+00
-9.34721172e-01 4.61110890e-01 -2.64432818e-01 -1.59700572e+00
-2.70234346e-01 -8.75793457e-01 -5.41571617e-01 1.25609434e+00
-1.87687206e+00 -4.83862281e-01 -2.73744583e-01 6.49794400e-01
1.93660095e-01 -6.68503106e-01 4.22973126e-01 2.02854890e-02
-6.26673818e-01 2.37126853e-02 3.75465333e-01 -2.29165182e-01
4.92172182e-01 -1.39853311e+00 -7.82521188e-01 1.05442321e+00
-4.42131490e-01 3.67134511e-01 1.28716481e+00 -3.56179029e-01
-1.66136670e+00 -8.89251173e-01 3.79264653e-01 1.35067940e-01
6.23631477e-01 6.07894398e-02 -3.63628983e-01 1.40691102e-01
9.18147743e-01 -5.51410876e-02 5.55241248e-03 -7.44876027e-01
8.41786444e-01 -7.54571319e-01 -1.27254736e+00 4.56079245e-01
6.96137905e-01 7.15573952e-02 -4.18336213e-01 8.68033290e-01
5.81432998e-01 -5.54144263e-01 -8.08370352e-01 4.36400682e-01
2.29181886e-01 -8.39938343e-01 6.18465602e-01 -3.98190379e-01
1.00198783e-01 -7.34284401e-01 1.54967502e-01 -1.69933116e+00
-2.39824042e-01 -1.55617094e+00 -7.68682584e-02 9.93275046e-01
3.31583321e-01 -1.11156452e+00 8.65257859e-01 -2.52733398e-02
-5.24688601e-01 -6.05953991e-01 -9.44814324e-01 -7.03827262e-01
-2.89703488e-01 -4.15128767e-02 9.99531075e-02 7.51925111e-01
-3.03893715e-01 4.65131432e-01 -1.67233542e-01 1.12054265e+00
8.67078125e-01 -1.95780337e-01 6.51809156e-01 -1.27681208e+00
-2.07865417e-01 -5.67588806e-01 -5.67741215e-01 -7.70535409e-01
1.66513488e-01 -6.05792165e-01 1.73322633e-01 -1.18807256e+00
-9.92218971e-01 -2.66411781e-01 -2.55785063e-02 -1.84446946e-01
1.70454934e-01 5.51766790e-02 1.14650212e-01 1.77273631e-01
4.93933201e-01 4.04328704e-01 1.45948243e+00 1.28017411e-01
-5.09527743e-01 4.42115158e-01 7.49932835e-03 3.79101694e-01
8.84288013e-01 -3.11597615e-01 -6.37534142e-01 -3.48899633e-01
3.03215295e-01 3.47640723e-01 3.99233162e-01 -1.60675204e+00
6.55779004e-01 -8.65888372e-02 1.33855343e-01 -2.56566226e-01
3.13748986e-01 -1.07959616e+00 1.12424707e-02 8.45429420e-01
1.40373306e-02 -1.09371684e-01 -8.58011842e-02 5.10070384e-01
-1.48543164e-01 -5.31124234e-01 1.08189178e+00 -1.92137420e-01
-1.40808731e-01 -2.36227214e-02 -6.86590314e-01 -2.96918154e-01
1.01379657e+00 -3.08577299e-01 -3.79381329e-01 -3.28294158e-01
-6.30396307e-01 2.57563144e-01 1.45011153e-02 -1.86583117e-01
3.03079337e-01 -8.58604252e-01 -6.80534542e-01 3.74023169e-01
-6.50398135e-01 -1.54642344e-01 -9.78531092e-02 1.04543734e+00
-7.74141908e-01 6.04009390e-01 -1.84528887e-01 -9.10545826e-01
-1.23150897e+00 2.69863456e-01 1.05064762e+00 2.04718858e-01
-1.01963148e-01 9.71738994e-01 -5.29587269e-01 -6.71934709e-02
-1.20452449e-01 -3.63098294e-01 -5.56504846e-01 -1.18494898e-01
1.11353487e-01 2.07919776e-01 -6.75032809e-02 -1.14282645e-01
-3.27664241e-02 1.21330869e+00 5.67625761e-01 -2.21533597e-01
1.16746521e+00 -2.10230783e-01 -4.07681346e-01 1.92173749e-01
9.75057065e-01 2.56021172e-01 -1.08311427e+00 -2.02822201e-02
-4.81644720e-01 -2.87808120e-01 6.02388203e-01 -8.35707605e-01
-1.21168530e+00 6.29560232e-01 1.11548305e+00 5.26355028e-01
1.52469742e+00 -6.92087770e-01 4.25311804e-01 7.91936457e-01
-1.19080674e-02 -1.32951105e+00 -4.56964746e-02 1.88278198e-01
7.41475046e-01 -8.35700989e-01 5.89354523e-02 -5.85350432e-02
-6.37372509e-02 1.55100918e+00 6.16271973e-01 -4.45034325e-01
8.04782510e-01 3.59704494e-01 9.05326158e-02 -2.01876223e-01
-7.29794741e-01 -1.36548251e-01 1.91198409e-01 3.59640241e-01
5.13716079e-02 -2.62989730e-01 -4.90263700e-01 -3.86272073e-02
-3.54396820e-01 2.93135822e-01 8.45164001e-01 6.79542720e-01
-9.22631443e-01 -8.50684106e-01 -6.59105957e-01 5.27573526e-01
-3.17394674e-01 1.40683353e-01 2.63103664e-01 5.57640672e-01
4.44419831e-01 9.45524931e-01 9.13127735e-02 -2.93043405e-01
5.08884728e-01 -1.03615671e-01 5.35641789e-01 -3.82233381e-01
-5.16551256e-01 -2.84138024e-02 -1.03393704e-01 -4.09927577e-01
-4.77356106e-01 -4.71424967e-01 -9.15057659e-01 1.08484872e-01
-5.99309325e-01 3.66154611e-01 1.01114047e+00 1.01381326e+00
4.95265543e-01 5.49188793e-01 1.09816909e+00 -9.47202861e-01
-6.40316486e-01 -1.12471700e+00 -6.42227948e-01 -4.43034828e-01
7.89748788e-01 -7.67474115e-01 -8.62373412e-01 6.94977026e-03] | [5.465681552886963, 2.5140044689178467] |
ab6af541-a6fc-4c22-9b53-e1bf8e58307a | quantifying-morphological-computation-based | 1503.05113 | null | http://arxiv.org/abs/1503.05113v1 | http://arxiv.org/pdf/1503.05113v1.pdf | Quantifying Morphological Computation based on an Information Decomposition of the Sensorimotor Loop | The question how an agent is affected by its embodiment has attracted growing
attention in recent years. A new field of artificial intelligence has emerged,
which is based on the idea that intelligence cannot be understood without
taking into account embodiment. We believe that a formal approach to
quantifying the embodiment's effect on the agent's behaviour is beneficial to
the fields of artificial life and artificial intelligence. The contribution of
an agent's body and environment to its behaviour is also known as morphological
computation. Therefore, in this work, we propose a quantification of
morphological computation, which is based on an information decomposition of
the sensorimotor loop into shared, unique and synergistic information. In
numerical simulation based on a formal representation of the sensorimotor loop,
we show that the unique information of the body and environment is a good
measure for morphological computation. The results are compared to our
previously derived quantification of morphological computation. | ['Johannes Rauh', 'Keyan Ghazi-Zahedi'] | 2015-03-17 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 2.57396609e-01 1.14989594e-01 1.90539107e-01 6.19450137e-02
6.80901229e-01 -3.99356484e-01 9.77811038e-01 3.44385177e-01
-6.25090718e-01 4.55476820e-01 4.33947533e-01 -3.41555439e-02
-5.25998116e-01 -1.09229004e+00 -2.69640386e-01 -6.95527434e-01
-1.16370618e-01 7.19308853e-02 -1.35202929e-01 -6.17527664e-01
3.11240107e-01 6.84395373e-01 -1.61331892e+00 -2.90554315e-01
7.64165998e-01 6.44999981e-01 2.11779952e-01 5.61781943e-01
1.72406211e-02 9.03159738e-01 -6.55201793e-01 -4.36104201e-02
-1.66801419e-02 -9.64149058e-01 -6.17078781e-01 1.84346102e-02
-3.49842429e-01 5.66782132e-02 1.82823822e-01 1.15695190e+00
1.67400435e-01 5.45188300e-02 7.37357318e-01 -1.06494522e+00
-5.72528541e-01 5.06947875e-01 -8.71620700e-02 2.31655128e-03
6.71519995e-01 2.17027649e-01 7.59485304e-01 -2.89432436e-01
7.01864362e-01 1.24674904e+00 4.67645794e-01 3.57129753e-01
-1.26902354e+00 -1.59335628e-01 -4.21354920e-02 1.24161042e-01
-1.42189097e+00 -9.82194394e-02 1.02061582e+00 -6.77742720e-01
6.68026209e-01 4.51185167e-01 1.28745282e+00 4.19354379e-01
3.81985515e-01 4.43937927e-01 1.30973828e+00 -7.63424039e-01
5.40304422e-01 2.30033129e-01 1.43857047e-01 5.39210916e-01
5.11027992e-01 4.45104837e-01 -2.60970861e-01 2.80641437e-01
8.77493441e-01 -1.76832750e-01 -1.50320366e-01 -4.85375345e-01
-1.24587262e+00 6.65065825e-01 3.86024356e-01 1.02018499e+00
-4.75793719e-01 4.83937204e-01 -2.17883158e-02 3.51586938e-01
8.76334831e-02 7.62524188e-01 -1.26829505e-01 -3.73284638e-01
-4.03338820e-01 4.00246382e-01 8.02063346e-01 1.80117890e-01
6.66140139e-01 -1.08710110e-01 4.93163258e-01 1.15655489e-01
4.28405404e-01 5.62790096e-01 5.06249070e-01 -9.02239263e-01
-1.37451053e-01 1.35634935e+00 2.10258700e-02 -1.21460629e+00
-6.71933353e-01 -4.77366298e-01 -5.72968066e-01 9.35344815e-01
6.56634092e-01 -1.47223100e-01 -2.90008128e-01 2.13693309e+00
3.79475355e-01 -2.17891604e-01 2.47954190e-01 9.02755916e-01
6.92924619e-01 1.45367533e-01 3.15398909e-02 -1.90786839e-01
1.17353809e+00 -3.02394688e-01 -9.53843176e-01 1.69135988e-01
7.68126070e-01 -4.58639681e-01 6.26514733e-01 3.36579591e-01
-1.37942815e+00 -4.08176422e-01 -1.26991928e+00 1.11320101e-01
-7.12612629e-01 -3.63980532e-02 9.97338295e-01 7.39525795e-01
-9.19185162e-01 6.97705328e-01 -7.60449111e-01 -7.48663545e-01
-8.26913640e-02 3.91289532e-01 -3.49847674e-01 6.97723389e-01
-8.97200346e-01 1.55818772e+00 4.13269252e-01 -1.91378780e-02
-3.12185526e-01 -1.88209489e-01 -8.36482525e-01 -1.05338976e-01
2.44531453e-01 -9.68353868e-01 8.36099148e-01 -1.14814842e+00
-1.64698827e+00 7.01568544e-01 2.38827825e-01 -4.39007580e-01
4.75656331e-01 1.04564235e-01 -2.21648633e-01 1.46672070e-01
-3.89775664e-01 2.95592576e-01 4.93692547e-01 -1.36595833e+00
-1.40525907e-01 -7.46189177e-01 4.17286128e-01 2.95749068e-01
-1.13186121e-01 3.45186107e-02 2.19104946e-01 -7.30973125e-01
3.33898306e-01 -8.31568539e-01 -3.77962142e-01 2.33311392e-02
2.66175196e-02 -8.97794664e-02 6.67443812e-01 -2.56260157e-01
1.13243294e+00 -2.09413004e+00 9.06983197e-01 1.11594275e-01
4.81178284e-01 7.75274560e-02 6.88728765e-02 7.06980526e-01
6.34637997e-02 6.68943897e-02 -2.75443763e-01 -4.53477018e-02
2.20604807e-01 2.46928096e-01 2.92712543e-02 6.21054351e-01
-8.66679326e-02 8.14602613e-01 -8.37518096e-01 -4.63745713e-01
3.94796729e-01 6.16180062e-01 -4.98037875e-01 -9.38096866e-02
2.66221687e-02 5.76506078e-01 -7.55326629e-01 2.89261639e-01
2.83072859e-01 1.44253954e-01 2.17329502e-01 -1.74800493e-02
-4.34148282e-01 1.63805466e-02 -1.06049573e+00 1.56161034e+00
-2.99321502e-01 3.97741675e-01 1.15692846e-01 -9.87723947e-01
8.80573750e-01 4.73338872e-01 4.87134367e-01 -7.98196256e-01
6.61802292e-01 1.18014410e-01 6.84904218e-01 -4.74174291e-01
3.72021109e-01 -6.59447730e-01 -1.52604923e-01 9.36797559e-01
-2.20256075e-01 -7.39916444e-01 3.87713820e-01 1.04380138e-01
1.07697272e+00 3.65313768e-01 1.02340841e+00 -4.65398818e-01
7.83538461e-01 -1.96248949e-01 1.88821986e-01 1.01939552e-01
-2.43609399e-01 -1.97368726e-01 3.91571999e-01 -3.43904346e-01
-7.57428944e-01 -1.09962761e+00 -8.08707103e-02 5.41701972e-01
4.24669832e-01 -2.26469874e-01 -1.13710046e+00 1.73832938e-01
4.27319370e-02 7.50439525e-01 -9.97584999e-01 -4.37546253e-01
-4.88685012e-01 -6.47931755e-01 2.16902062e-01 2.64151573e-01
4.48405147e-01 -1.42921090e+00 -1.65540218e+00 2.42514536e-01
2.17373818e-01 -7.65143752e-01 4.35462087e-01 2.80729253e-02
-1.20672441e+00 -8.91902268e-01 -2.88918346e-01 -1.86823726e-01
7.08570778e-01 7.71015063e-02 8.63668561e-01 4.80904013e-01
-3.00654858e-01 5.61638951e-01 -5.43984294e-01 -5.96012831e-01
-7.92289793e-01 -3.85702163e-01 1.91959247e-01 -1.73498511e-01
1.43471137e-01 -9.10624146e-01 -3.52476478e-01 9.12020579e-02
-1.07605767e+00 3.51946115e-01 5.13685346e-01 2.10555360e-01
9.46669206e-02 3.97635698e-01 3.98064375e-01 -2.17693090e-01
7.19346642e-01 -2.42266238e-01 -2.42787838e-01 -4.53167930e-02
-2.44173378e-01 2.40385681e-01 2.90454209e-01 -3.67247850e-01
-1.18826413e+00 -6.04006611e-02 2.01899797e-01 4.12984788e-01
-3.19036692e-01 6.33186996e-01 -3.92716527e-01 -2.47159138e-01
2.53311753e-01 1.17881820e-01 2.97148973e-01 -5.54981709e-01
5.54508686e-01 3.52340132e-01 4.25897330e-01 -5.98387897e-01
6.58931613e-01 6.69548035e-01 4.79377329e-01 -9.76683021e-01
-3.45133357e-02 -6.23929165e-02 -8.42239559e-01 -4.91390526e-01
9.92568612e-01 -1.89443439e-01 -1.04541075e+00 3.92401695e-01
-1.14168406e+00 -2.11906031e-01 -5.33465326e-01 8.44021976e-01
-9.65124726e-01 3.08212966e-01 -2.78622478e-01 -1.18562555e+00
-8.52456689e-02 -1.13128114e+00 7.31302559e-01 2.13318467e-01
-7.70961463e-01 -1.20574462e+00 3.47708821e-01 1.51045829e-01
3.51651698e-01 4.96381044e-01 8.83658171e-01 -2.96036243e-01
-4.88351017e-01 -1.46096736e-01 1.90815568e-01 5.77364117e-02
1.01299904e-01 -9.04839560e-02 -6.23462737e-01 1.22584715e-01
4.74316627e-01 1.91713616e-01 5.44601679e-01 1.32395625e-01
6.51670545e-02 -1.56322811e-02 -2.92170614e-01 2.38970160e-01
1.66556263e+00 5.21445632e-01 7.71545351e-01 4.26028937e-01
2.67567426e-01 1.01189744e+00 4.10643965e-01 5.06554425e-01
1.08084925e-01 8.92831326e-01 6.93556666e-01 4.09851596e-02
-2.31057718e-01 1.23917952e-01 3.78400862e-01 7.69973993e-01
-7.98563063e-01 -1.67770032e-02 -6.00049317e-01 3.33676338e-01
-1.63549292e+00 -1.20015705e+00 -1.22451887e-01 2.05733109e+00
5.84411085e-01 -4.85852659e-02 1.32966593e-01 6.04217589e-01
4.23333943e-01 -2.32813716e-01 -2.31336579e-01 -5.71312189e-01
-1.01838551e-01 8.59238729e-02 4.17828187e-02 5.59074402e-01
-5.58225155e-01 5.69503784e-01 7.10791206e+00 3.01904798e-01
-9.64941382e-01 3.98155935e-02 -6.32708892e-02 7.56231770e-02
-3.90973389e-01 -1.37941658e-01 -2.38790922e-02 4.41229492e-01
5.25627375e-01 -6.36811554e-01 5.72119296e-01 4.04260278e-01
5.14638722e-01 -6.77989423e-01 -1.08486795e+00 5.67852736e-01
7.65851373e-03 -7.33052731e-01 -1.24015912e-01 4.45059866e-01
3.16766858e-01 -6.10698462e-01 -1.18803175e-03 -2.59346515e-01
2.48298511e-01 -1.07312775e+00 1.01103246e+00 8.35103393e-01
1.55341402e-01 -5.51672101e-01 6.88618600e-01 5.21536827e-01
-1.21253443e+00 -1.07151993e-01 1.99445873e-01 -8.94974887e-01
3.75226647e-01 2.72334665e-01 -4.56460327e-01 3.12759519e-01
8.42534527e-02 4.72278774e-01 -4.77685302e-01 8.96482646e-01
-2.71046102e-01 1.11894332e-01 -3.92454475e-01 -3.64349842e-01
-4.04945724e-02 -7.32314527e-01 8.82099807e-01 6.94051862e-01
3.96213412e-01 4.00839955e-01 -3.42444092e-01 1.48615170e+00
5.91196239e-01 2.50103563e-01 -7.70753860e-01 -3.93115520e-01
-4.52668481e-02 1.05096149e+00 -1.03377414e+00 -3.52351457e-01
-1.11028820e-01 5.75597703e-01 -3.37415636e-02 9.96496454e-02
-6.93771124e-01 -3.91937606e-02 7.07166791e-01 -1.49114519e-01
-1.18090793e-01 -4.50299412e-01 -5.57417810e-01 -7.22271323e-01
-2.32707828e-01 -5.20832181e-01 -2.42418692e-01 -7.28457689e-01
-6.33168936e-01 2.39550680e-01 1.80924937e-01 -9.22133565e-01
-3.31167877e-01 -6.43997371e-01 -5.93132973e-01 6.28209293e-01
-5.22071600e-01 -1.13339150e+00 -2.18574390e-01 3.79751325e-01
-1.94860116e-01 4.36221026e-02 1.01481795e+00 -1.62953839e-01
-3.64809662e-01 7.87448362e-02 -3.62661123e-01 -1.40045866e-01
-2.20778093e-01 -1.29820347e+00 -3.60364802e-02 8.12080622e-01
6.90154848e-05 8.20510030e-01 1.33063889e+00 -4.92039293e-01
-1.58189166e+00 -2.37026900e-01 9.76433754e-01 -5.86668611e-01
7.80812144e-01 1.04345217e-01 -4.23883945e-01 6.44803405e-01
2.06067145e-01 -4.80244756e-01 6.48822069e-01 -1.09931238e-01
2.79851183e-02 5.13573885e-02 -1.34735429e+00 8.63344073e-01
1.17494988e+00 -3.40000778e-01 -1.22413588e+00 -2.40622461e-01
5.93989193e-01 2.03895956e-01 -9.89483774e-01 2.64995962e-01
8.26652586e-01 -1.39076376e+00 1.00233817e+00 -1.13551356e-01
1.86863720e-01 -4.66856807e-01 -1.31056130e-01 -1.16437912e+00
-2.00390160e-01 -2.96187103e-01 8.10435116e-02 7.50086725e-01
-9.40893292e-02 -9.34473097e-01 4.50947702e-01 3.34584534e-01
1.26206249e-01 -3.78903031e-01 -8.14321816e-01 -7.61933386e-01
1.86618030e-01 -4.93987739e-01 7.00136244e-01 8.50394249e-01
8.38702917e-01 2.15911463e-01 1.25973180e-01 -3.45426828e-01
6.46368980e-01 2.09481567e-01 7.16424167e-01 -1.64643073e+00
-2.87519872e-01 -9.60053861e-01 -1.03231120e+00 -3.88868749e-01
-1.40637264e-01 -7.07183123e-01 -2.75644124e-01 -1.67164385e+00
2.47711629e-01 -1.64877266e-01 -1.85886249e-01 1.32604048e-01
3.56400639e-01 3.56069922e-01 6.21060252e-01 -2.17511207e-02
-2.63118565e-01 4.59239095e-01 1.63133287e+00 2.76988596e-01
-2.39439711e-01 -4.05706763e-01 -5.01769781e-01 1.16207194e+00
8.96676540e-01 -8.53114054e-02 -3.10497075e-01 4.39819545e-02
6.51795447e-01 5.78745417e-02 4.99033779e-01 -1.26966417e+00
-3.52525711e-03 -1.92117676e-01 -6.53439015e-02 -2.92437047e-01
6.12908065e-01 -1.18018544e+00 6.54060364e-01 1.12580061e+00
-1.57873482e-01 -4.05175760e-02 3.34364325e-02 1.56038493e-01
-1.33192807e-01 -3.72996598e-01 4.95007455e-01 -3.04800659e-01
-6.84699893e-01 -5.24870336e-01 -6.62749648e-01 -4.44797635e-01
1.20723546e+00 -4.51963842e-01 -1.80656731e-01 -3.08522910e-01
-7.75017738e-01 -3.82671058e-01 9.68966782e-01 5.01339212e-02
3.88970375e-01 -1.19732141e+00 -3.67566109e-01 -4.14156467e-02
-1.11552484e-01 -7.14213073e-01 2.64729597e-02 1.05904138e+00
-6.78569436e-01 5.27010798e-01 -6.19467914e-01 -3.97338383e-02
-1.35824370e+00 7.91046143e-01 2.94051737e-01 -7.88749233e-02
-4.19567406e-01 3.74936312e-01 4.00185913e-01 -1.58784121e-01
-1.79486185e-01 -6.52322292e-01 -5.43425381e-01 1.34919463e-02
5.75051486e-01 6.60978496e-01 -5.77192962e-01 -1.18395066e+00
-4.18258876e-01 1.11964834e+00 7.48641431e-01 -4.37233567e-01
1.01982856e+00 -2.14301199e-01 -6.85335517e-01 5.58393598e-01
8.28986526e-01 2.76847869e-01 -4.97955501e-01 2.74951637e-01
-1.24437533e-01 -1.73773140e-01 2.55891662e-02 -9.50717688e-01
-6.93395734e-01 8.08600783e-01 4.17904586e-01 6.76344275e-01
1.23277593e+00 -5.54972440e-02 3.51426154e-01 2.40697175e-01
7.66274929e-01 -9.96699631e-01 3.20993848e-02 2.04292059e-01
1.15014112e+00 -6.77587688e-01 2.61727095e-01 -5.00529885e-01
-2.96114981e-01 1.07298386e+00 3.32705826e-02 -1.13464862e-01
6.40144110e-01 5.14546394e-01 -4.73637842e-02 -2.82508194e-01
-4.69938725e-01 -6.64102376e-01 1.73206717e-01 5.58239222e-01
5.14275968e-01 4.35286105e-01 -9.81846571e-01 4.86696273e-01
-6.28437877e-01 3.85280728e-01 4.62822825e-01 9.51247215e-01
-7.60245860e-01 -1.08367932e+00 -6.23233080e-01 -2.56322533e-01
-1.56916454e-01 1.52287394e-01 -7.71281362e-01 1.04603028e+00
8.12097251e-01 1.00034988e+00 -6.34247530e-03 -3.51983070e-01
1.41782416e-02 -2.34990060e-01 1.09027004e+00 -4.49249357e-01
-6.62738562e-01 -2.15975240e-01 8.63921568e-02 -6.61444187e-01
-1.02396989e+00 -7.91808367e-01 -1.75436711e+00 -3.96438867e-01
-8.87666196e-02 2.43702710e-01 1.03054702e+00 1.08353305e+00
-1.49132892e-01 6.72273338e-01 -1.17068730e-01 -9.79456782e-01
-2.02242002e-01 -9.24168766e-01 -8.18221867e-01 4.34171945e-01
1.40776083e-01 -1.02830505e+00 -6.11429513e-01 1.05552711e-02] | [5.626086711883545, 4.174091339111328] |
993bf131-dfee-4563-bcf9-199ed32706d8 | a-transition-based-dependency-parser-using-a | null | null | https://aclanthology.org/P13-1014 | https://aclanthology.org/P13-1014.pdf | A Transition-Based Dependency Parser Using a Dynamic Parsing Strategy | null | ['Giorgio Satta', 'Francesco Sartorio', 'Joakim Nivre'] | 2013-08-01 | null | null | null | acl-2013-8 | ['transition-based-dependency-parsing'] | ['natural-language-processing'] | [-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01] | [-7.295925140380859, 3.758568286895752] |
c1ff6717-9818-4f01-9fac-084d7990da02 | distilled-dual-encoder-model-for-vision | 2112.08723 | null | https://arxiv.org/abs/2112.08723v2 | https://arxiv.org/pdf/2112.08723v2.pdf | Distilled Dual-Encoder Model for Vision-Language Understanding | We propose a cross-modal attention distillation framework to train a dual-encoder model for vision-language understanding tasks, such as visual reasoning and visual question answering. Dual-encoder models have a faster inference speed than fusion-encoder models and enable the pre-computation of images and text during inference. However, the shallow interaction module used in dual-encoder models is insufficient to handle complex vision-language understanding tasks. In order to learn deep interactions of images and text, we introduce cross-modal attention distillation, which uses the image-to-text and text-to-image attention distributions of a fusion-encoder model to guide the training of our dual-encoder model. In addition, we show that applying the cross-modal attention distillation for both pre-training and fine-tuning stages achieves further improvements. Experimental results demonstrate that the distilled dual-encoder model achieves competitive performance for visual reasoning, visual entailment and visual question answering tasks while enjoying a much faster inference speed than fusion-encoder models. Our code and models will be publicly available at https://github.com/kugwzk/Distilled-DualEncoder. | ['Furu Wei', 'Bing Qin', 'Ming Liu', 'Haichao Zhu', 'Wenhui Wang', 'Zekun Wang'] | 2021-12-16 | null | null | null | null | ['visual-entailment'] | ['reasoning'] | [-1.12974823e-01 2.82843024e-01 2.95200776e-02 -5.76817334e-01
-8.13343167e-01 -4.33151573e-01 8.62785518e-01 -1.66740432e-01
-3.83666217e-01 2.16762543e-01 3.11095923e-01 -8.01775157e-01
3.22228372e-01 -7.75067389e-01 -1.22897565e+00 -3.82790446e-01
5.45801997e-01 5.41357100e-01 8.53156857e-03 -1.43794175e-02
-5.44683561e-02 5.47060333e-02 -1.55087769e+00 9.36176896e-01
7.63524413e-01 9.56190228e-01 5.71307242e-01 1.23994625e+00
-2.94693470e-01 1.58209062e+00 -2.47273296e-01 -7.07985997e-01
5.20089231e-02 -4.09524798e-01 -1.07172847e+00 7.93527365e-02
8.62323999e-01 -1.04780138e+00 -9.07296300e-01 9.65375364e-01
3.02363843e-01 4.42101322e-02 5.89717090e-01 -1.22062826e+00
-1.51062393e+00 3.60973001e-01 -6.32437587e-01 3.54547352e-01
3.40272546e-01 5.02181232e-01 1.26644909e+00 -1.20385695e+00
2.81218827e-01 1.68388534e+00 1.78561434e-01 6.66046202e-01
-1.14259994e+00 -4.89943266e-01 4.34020638e-01 6.24477148e-01
-1.11115909e+00 -3.74996185e-01 5.00000834e-01 -4.63922381e-01
1.09045923e+00 -4.24302518e-02 4.98339504e-01 1.04682481e+00
1.53231040e-01 1.38458180e+00 8.57394516e-01 -3.36306602e-01
-1.72588319e-01 1.30339742e-01 -5.02113104e-02 1.19460118e+00
-1.15762532e-01 -1.16277307e-01 -3.06855708e-01 2.57145107e-01
9.25947547e-01 3.82845074e-01 -3.26922566e-01 2.27356832e-02
-1.24104786e+00 9.72822189e-01 9.36495602e-01 -8.20725923e-04
-3.97554874e-01 5.75879633e-01 3.54666263e-01 1.57603562e-01
3.87413561e-01 3.38002145e-02 -3.60427976e-01 1.27455622e-01
-5.78939736e-01 1.32100374e-01 4.49126482e-01 9.82927144e-01
7.73956180e-01 -3.13192271e-02 -6.40482008e-01 8.02996218e-01
5.55236101e-01 6.95428729e-01 2.62374222e-01 -1.20499599e+00
5.63939095e-01 5.44192672e-01 -2.39082843e-01 -6.29797697e-01
-7.18885195e-03 -7.47578219e-02 -1.03519571e+00 3.23508501e-01
2.44129032e-01 3.70144588e-03 -1.40225327e+00 1.51735306e+00
2.32438385e-01 2.24747639e-02 1.98136359e-01 9.92763221e-01
1.26101172e+00 8.33685935e-01 1.08225010e-01 4.58556145e-01
1.81884050e+00 -1.58615494e+00 -6.99480891e-01 -4.47150290e-01
4.43917215e-01 -6.91942334e-01 1.33954334e+00 -7.24942535e-02
-1.59016049e+00 -8.18057120e-01 -7.32041359e-01 -1.12507725e+00
-4.10281330e-01 2.97684640e-01 5.37726223e-01 -2.75440603e-01
-1.06987989e+00 -4.45712395e-02 -8.12049270e-01 -9.03735757e-02
7.59533048e-01 7.12337941e-02 -2.68419564e-01 -4.92877960e-01
-1.03683269e+00 9.33545291e-01 2.31248811e-01 6.77479580e-02
-1.28812754e+00 -7.71673143e-01 -1.16060925e+00 2.74054259e-01
2.61130095e-01 -1.43214822e+00 1.54238236e+00 -8.40715706e-01
-1.24052846e+00 1.10027087e+00 -3.83931339e-01 -4.56425428e-01
4.18988556e-01 -3.64675045e-01 1.78465173e-01 5.44419408e-01
1.39253914e-01 1.21019399e+00 9.99569595e-01 -1.22288418e+00
-5.90259194e-01 -3.43659669e-01 7.42434442e-01 4.25303608e-01
4.10845093e-02 -4.53843623e-01 -1.01028287e+00 -3.58294636e-01
-2.39568025e-01 -5.81265152e-01 3.35891806e-02 4.87483978e-01
-3.59029323e-01 -4.28508162e-01 8.98823738e-01 -8.43055487e-01
6.65703773e-01 -2.12038898e+00 3.06436598e-01 -5.17347634e-01
6.23747110e-01 3.04174989e-01 -4.53584433e-01 1.87836498e-01
-8.05867240e-02 -1.23182014e-01 -4.34093140e-02 -7.77918041e-01
4.84960042e-02 6.48127973e-01 -4.92554665e-01 1.88548192e-01
5.37824094e-01 1.57044923e+00 -8.23774576e-01 -5.40503085e-01
5.69343865e-01 7.24780738e-01 -8.19721282e-01 6.01622641e-01
-6.17745817e-01 1.59283713e-01 -2.54899144e-01 6.92061722e-01
6.03649437e-01 -8.28599691e-01 -2.30565950e-01 -6.27212882e-01
2.55512446e-01 3.01843077e-01 -3.24664146e-01 1.91582847e+00
-8.08346748e-01 9.32369351e-01 1.34383619e-01 -1.09223211e+00
3.11752707e-01 1.34306401e-01 -7.09331110e-02 -1.11102903e+00
1.65539891e-01 -3.15620154e-01 -1.64794639e-01 -7.72191882e-01
3.90696228e-01 -6.06379919e-02 2.15008840e-01 5.40234745e-01
4.90310729e-01 -2.66941905e-01 2.29643941e-01 7.09987104e-01
6.38950408e-01 1.63832471e-01 9.44481790e-03 1.77291960e-01
6.26756310e-01 -2.69983381e-01 -5.51281460e-02 7.71922827e-01
-3.29732485e-02 6.33394659e-01 5.31246305e-01 -4.09506202e-01
-1.09834754e+00 -1.33521867e+00 1.94257826e-01 1.39566278e+00
1.16640665e-01 -4.19581115e-01 -5.44369459e-01 -6.52115107e-01
2.46326119e-01 7.34964728e-01 -9.37384367e-01 -2.06770867e-01
-1.06491134e-01 -1.14221223e-01 5.84222257e-01 8.57379496e-01
7.17878401e-01 -9.00917351e-01 -4.18784320e-01 -2.87579536e-01
-4.69882220e-01 -1.49959958e+00 -5.86125314e-01 9.17335972e-02
-5.57889581e-01 -1.14594686e+00 -7.73695946e-01 -9.38002527e-01
8.30837309e-01 4.77759749e-01 1.25089037e+00 2.92904437e-01
-4.09711182e-01 7.72206128e-01 -1.70104921e-01 -3.87528896e-01
-2.19370961e-01 -2.55125731e-01 -7.07394540e-01 -1.84877381e-01
3.53930205e-01 -3.42978150e-01 -8.84231389e-01 -2.61891373e-02
-1.06734931e+00 6.17801249e-01 7.34027267e-01 1.10061383e+00
5.51547289e-01 -5.65685511e-01 7.96341076e-02 -5.12243629e-01
4.49144959e-01 -3.87914926e-01 -4.43070590e-01 5.35659492e-01
-1.48294657e-01 3.94744098e-01 3.81746560e-01 -2.77280658e-01
-1.13785684e+00 -1.68636188e-01 -6.16438150e-01 -9.27334309e-01
-7.04196095e-02 2.96780825e-01 -7.44734099e-03 2.51828134e-01
2.38462508e-01 3.99773657e-01 2.16136351e-01 -2.06761852e-01
1.02037716e+00 5.32419443e-01 9.11765814e-01 -4.75918651e-01
6.50342464e-01 6.70586586e-01 -3.72507483e-01 -5.19181073e-01
-1.41378796e+00 -3.94829512e-01 -4.13833112e-01 -5.34686968e-02
1.45655727e+00 -1.45016205e+00 -1.20243609e+00 3.08360815e-01
-1.57437778e+00 -6.29631877e-01 -3.26013058e-01 2.78140008e-01
-6.95345819e-01 3.16180140e-01 -7.89105237e-01 -5.80761254e-01
-5.04227877e-01 -1.31071484e+00 1.47083902e+00 2.28127077e-01
3.81729901e-01 -1.27676344e+00 -3.00493330e-01 8.91112268e-01
3.09474200e-01 -1.30456284e-01 8.56485307e-01 -1.59874469e-01
-8.64274144e-01 2.58345515e-01 -8.47530067e-01 4.48185444e-01
-1.03191875e-01 -1.52907655e-01 -1.23777449e+00 -1.76020712e-01
-1.59147441e-01 -9.06695068e-01 1.36023915e+00 3.81332844e-01
1.64881265e+00 -2.69655257e-01 -6.05880842e-02 8.44517291e-01
1.34325528e+00 -8.43868554e-02 7.57375896e-01 -4.39364500e-02
1.03385937e+00 2.95336246e-01 1.70582935e-01 2.36240447e-01
1.19438338e+00 3.24931592e-01 7.36757278e-01 -5.35583436e-01
-4.62408185e-01 -3.82545084e-01 2.70521373e-01 5.44023037e-01
5.46540059e-02 -1.02132238e-01 -8.71611655e-01 6.95373297e-01
-2.00522661e+00 -1.21003103e+00 -1.82216652e-02 1.52505863e+00
7.96305656e-01 -9.71249789e-02 -1.71194345e-01 -3.70534688e-01
2.72909135e-01 1.98802143e-01 -6.81263864e-01 -5.30131519e-01
1.11089736e-01 1.05151057e-01 1.79569766e-01 9.02452528e-01
-8.98922205e-01 9.49557781e-01 5.82490587e+00 5.00204325e-01
-9.56084490e-01 3.63249570e-01 7.54020214e-01 -1.03117064e-01
-5.89561939e-01 -1.20895505e-01 -5.55772841e-01 1.91734314e-01
6.98951542e-01 -1.69779155e-02 4.17865157e-01 6.55715168e-01
-1.60965696e-01 -1.42556518e-01 -1.23776782e+00 1.31782901e+00
4.49093223e-01 -1.58604324e+00 3.92978221e-01 -1.03782684e-01
6.77538097e-01 3.65350276e-01 2.72184551e-01 6.47456825e-01
5.53745985e-01 -1.15099919e+00 7.00870156e-01 5.78753769e-01
8.09132278e-01 -5.81105232e-01 5.20467222e-01 2.26755157e-01
-1.23906684e+00 -9.61819366e-02 -3.58713925e-01 -1.12630762e-01
1.81117877e-01 4.45646197e-01 -5.58095872e-01 4.84428018e-01
7.57108688e-01 8.10520589e-01 -5.25197625e-01 2.99619108e-01
-4.35734272e-01 1.85914472e-01 6.09458191e-03 3.18152666e-01
4.29753631e-01 4.24459279e-02 -9.96704400e-03 1.23901498e+00
-1.06116883e-01 7.65463412e-02 3.72487009e-02 1.31293559e+00
-3.04272175e-01 -5.00300288e-01 -6.15271747e-01 -2.54886568e-01
8.04280583e-03 1.15563297e+00 -1.09716497e-01 -8.41676891e-01
-8.92462730e-01 1.40321362e+00 6.19720697e-01 6.91849291e-01
-1.12737751e+00 -2.49487936e-01 8.90507102e-01 -2.19162956e-01
6.73890948e-01 -1.91711560e-01 -2.70744681e-01 -1.50468254e+00
-1.17359206e-01 -7.41064072e-01 4.20131981e-01 -1.34114265e+00
-1.29760683e+00 5.76461017e-01 -6.84126690e-02 -6.61433995e-01
-3.40064406e-01 -8.77704024e-01 -5.94780803e-01 9.99019146e-01
-1.93292642e+00 -1.63848698e+00 -7.19123483e-01 9.49125350e-01
8.34474444e-01 1.75259456e-01 6.29639149e-01 2.14532018e-01
-3.59805346e-01 4.51765180e-01 -1.79133683e-01 3.35756838e-01
5.60758710e-01 -1.44310808e+00 5.31818449e-01 6.41227424e-01
1.85474589e-01 5.56842327e-01 2.72206396e-01 -1.36328623e-01
-1.60008764e+00 -1.15818679e+00 6.26688302e-01 -5.21641433e-01
6.66762531e-01 -4.82207745e-01 -8.84222746e-01 1.17331064e+00
9.21021044e-01 2.70835876e-01 5.05821228e-01 5.41147143e-02
-6.86012566e-01 8.97440463e-02 -6.29327774e-01 6.38284266e-01
7.26710975e-01 -1.28795588e+00 -8.06300581e-01 4.66584355e-01
8.98154020e-01 -5.86063564e-01 -7.11721718e-01 1.40837312e-01
5.29219151e-01 -9.74721730e-01 1.28252912e+00 -4.80843455e-01
9.88800228e-01 -1.96150973e-01 -9.36375111e-02 -1.02635932e+00
-3.62391949e-01 2.72011999e-02 -5.26184916e-01 8.61237943e-01
2.59921730e-01 -3.80023658e-01 3.92385542e-01 3.27155024e-01
-5.67757338e-02 -9.96110141e-01 -4.34835732e-01 -1.75043255e-01
1.45989463e-01 -3.78325373e-01 2.10724130e-01 6.50331855e-01
-2.44139314e-01 8.71544957e-01 -3.31130147e-01 1.51345268e-01
6.59894407e-01 3.70553344e-01 8.75823259e-01 -6.75290525e-01
-6.10245109e-01 -4.12853837e-01 -7.55849630e-02 -1.73678327e+00
3.30821007e-01 -9.77053940e-01 3.16322297e-02 -2.24817109e+00
5.27331829e-01 3.84199500e-01 1.00293197e-02 7.35600293e-01
-4.69623089e-01 3.49220932e-01 4.22543138e-01 -6.52955100e-02
-7.78650224e-01 7.52542377e-01 1.73986530e+00 -5.29591739e-01
3.96632671e-01 -6.01931393e-01 -7.34830379e-01 6.64927006e-01
4.64063615e-01 1.05939351e-01 -6.66750968e-01 -1.14957166e+00
9.52871963e-02 8.16571191e-02 9.54172611e-01 -5.37687540e-01
3.44559163e-01 7.40018860e-02 5.95411420e-01 -8.51884604e-01
6.14908934e-01 -5.85128009e-01 -5.51277995e-01 4.61365759e-01
-4.50084686e-01 6.96266815e-02 4.36301798e-01 5.02210677e-01
-4.15656775e-01 1.10332943e-01 6.35300994e-01 -2.61196285e-01
-7.94915378e-01 4.54281449e-01 -2.97263384e-01 1.72762558e-01
8.81084919e-01 1.18966565e-01 -6.26169026e-01 -6.48567200e-01
-8.58849943e-01 7.06089139e-01 2.60260791e-01 4.18742955e-01
1.01268220e+00 -1.27504027e+00 -7.13920772e-01 1.18647344e-01
2.50960171e-01 4.10739422e-01 5.87862015e-01 9.00889695e-01
-5.34841478e-01 5.83476245e-01 -2.62244701e-01 -7.97632694e-01
-1.24965703e+00 7.66191244e-01 4.93296355e-01 -1.60349682e-01
-7.17444181e-01 1.11655068e+00 1.00043249e+00 -2.32206225e-01
3.69751483e-01 -6.52877092e-01 2.35454589e-01 -2.56702572e-01
8.82418275e-01 -1.80919692e-01 -3.60895663e-01 -3.07716548e-01
-3.21231037e-01 4.41166520e-01 -3.70006412e-01 -5.43249249e-02
1.05622685e+00 -3.32325011e-01 -3.61391753e-02 3.20827067e-01
1.45684397e+00 -5.87916553e-01 -1.59379542e+00 -3.06501865e-01
-6.87891483e-01 -3.21789116e-01 2.61212647e-01 -6.76221967e-01
-1.11417031e+00 1.50336540e+00 2.78259605e-01 -1.27548575e-01
1.21067595e+00 4.19432878e-01 6.71634614e-01 4.97296512e-01
-2.95514673e-01 -6.69112086e-01 6.67554021e-01 6.66143835e-01
1.15669203e+00 -1.63709164e+00 -2.11098209e-01 -5.03123365e-02
-8.56015325e-01 1.02636981e+00 9.96270418e-01 -4.73243035e-02
5.97652793e-01 2.59314299e-01 4.46917601e-02 -3.93805772e-01
-1.39439118e+00 -6.04769528e-01 5.35289764e-01 4.57230479e-01
4.61765617e-01 -1.54301077e-01 3.39311868e-01 -2.68336739e-02
2.34765440e-01 -1.59401249e-03 2.15749651e-01 7.23578811e-01
-2.41225302e-01 -6.75848722e-01 -2.32143179e-01 3.68677348e-01
-2.05182031e-01 -3.90285909e-01 -2.64025956e-01 5.95985651e-01
1.33300051e-01 7.68708825e-01 4.91822064e-01 -1.53702602e-01
5.26964590e-02 4.90103662e-02 8.78025651e-01 -4.78348613e-01
-2.59778857e-01 -1.91675559e-01 -1.06154509e-01 -6.52101278e-01
-4.29634422e-01 -4.83689532e-02 -1.35903955e+00 -4.25438821e-01
-1.59019887e-01 -1.49415836e-01 6.11106038e-01 1.09487247e+00
4.75812435e-01 9.49585497e-01 2.33178228e-01 -7.28515387e-01
-2.96163291e-01 -9.70794976e-01 -7.25614727e-02 4.78979230e-01
8.26169133e-01 -4.44473654e-01 -1.44373089e-01 4.01579529e-01] | [10.821964263916016, 1.6595840454101562] |
d87fd5b7-4ff0-4078-93c3-b6f8aa3b843a | contextualizing-argument-quality-assessment | 2305.12280 | null | https://arxiv.org/abs/2305.12280v1 | https://arxiv.org/pdf/2305.12280v1.pdf | Contextualizing Argument Quality Assessment with Relevant Knowledge | Automatic assessment of the quality of arguments has been recognized as a challenging task with significant implications for misinformation and targeted speech. While real world arguments are tightly anchored in context, existing efforts to judge argument quality analyze arguments in isolation, ultimately failing to accurately assess arguments. We propose SPARK: a novel method for scoring argument quality based on contextualization via relevant knowledge. We devise four augmentations that leverage large language models to provide feedback, infer hidden assumptions, supply a similar-quality argument, or a counterargument. We use a dual-encoder Transformer architecture to enable the original argument and its augmentation to be considered jointly. Our experiments in both in-domain and zero-shot setups show that SPARK consistently outperforms baselines across multiple metrics. We make our code available to encourage further work on argument assessment. | ['Fred Morstatter', 'Filip Ilievski', 'Zhivar Sourati', 'Darshan Deshpande'] | 2023-05-20 | null | null | null | null | ['misinformation'] | ['miscellaneous'] | [ 3.28513592e-01 4.29413974e-01 -7.31178164e-01 -7.45783687e-01
-1.87773478e+00 -9.73963559e-01 9.83719826e-01 6.44069910e-01
-5.87427318e-01 8.15459430e-01 1.26038086e+00 -1.07771909e+00
2.63183396e-02 -7.70265520e-01 -8.53466868e-01 5.23531996e-02
6.05379701e-01 6.60434723e-01 6.35033548e-02 -3.31044585e-01
6.49286389e-01 -1.88275740e-01 -1.00956702e+00 9.73908961e-01
7.44631827e-01 6.69047236e-01 -3.91213447e-01 8.03885043e-01
-5.11247754e-01 1.38985002e+00 -9.70361829e-01 -1.31107330e+00
-2.53751967e-02 -2.20241711e-01 -1.37799108e+00 -4.86213297e-01
5.35206437e-01 -5.07079124e-01 3.54345068e-02 9.89517987e-01
6.54219806e-01 -2.46010154e-01 3.86138529e-01 -7.76540935e-01
-9.01717842e-01 1.62937689e+00 -8.70197713e-02 6.17106616e-01
3.97509754e-01 -1.00720666e-01 1.74384487e+00 -8.65173161e-01
8.06132495e-01 1.70448148e+00 6.59240544e-01 2.50472605e-01
-1.28600204e+00 -4.54992414e-01 4.56142128e-01 1.68053687e-01
-2.21604519e-02 -7.02911913e-01 8.27132523e-01 -3.99026334e-01
7.09706664e-01 2.80714840e-01 3.59803170e-01 1.66076076e+00
-3.93417269e-01 9.79814589e-01 1.23896694e+00 -2.51916796e-01
1.36011422e-01 -1.48349836e-01 6.61693275e-01 2.83259541e-01
2.76422322e-01 -4.00027782e-02 -5.03018320e-01 -8.08880270e-01
-3.25412937e-02 -4.55144346e-01 2.50618756e-01 1.37725145e-01
-1.15895295e+00 1.38620019e+00 5.33947289e-01 5.98496348e-02
-4.26739216e-01 1.46304533e-01 6.73413336e-01 3.87925416e-01
6.16222143e-01 7.80820489e-01 -4.70367819e-01 -6.85183525e-01
-5.27470589e-01 8.21943760e-01 9.80741560e-01 2.49080867e-01
2.91377544e-01 -3.65206927e-01 -5.88587105e-01 9.93059754e-01
1.19776763e-01 4.49674308e-01 7.14868307e-02 -1.29884481e+00
1.10249829e+00 5.84267855e-01 4.98364180e-01 -8.69444609e-01
-8.41404647e-02 -4.54922676e-01 1.91126719e-01 1.78936139e-01
5.88668168e-01 -2.42782637e-01 -4.39835817e-01 1.85662735e+00
5.03181517e-01 -3.02978694e-01 4.63970244e-01 9.00879681e-01
9.56734300e-01 4.23543781e-01 2.31011838e-01 2.44453996e-01
1.51571441e+00 -9.70467210e-01 -7.78560221e-01 -4.64786649e-01
6.16905808e-01 -1.11918771e+00 1.50380683e+00 2.87916623e-02
-1.27017176e+00 2.31749602e-02 -9.87192929e-01 -4.78071213e-01
4.04556133e-02 -1.70939133e-01 6.03592098e-01 4.10296708e-01
-3.29505503e-01 6.42886877e-01 -4.71824318e-01 3.92987758e-01
6.77165210e-01 -3.74334514e-01 -1.07000269e-01 2.10028246e-01
-1.58524251e+00 1.03905594e+00 -6.08493462e-02 -1.46429718e-01
-8.50864828e-01 -5.82975447e-01 -9.15356576e-01 3.05800862e-03
5.91413319e-01 -5.76641381e-01 1.92755353e+00 -7.32663095e-01
-1.31348503e+00 8.92478526e-01 -1.42831698e-01 -8.28706801e-01
7.22560644e-01 -7.81383753e-01 -2.80545861e-01 -1.29901677e-01
7.53358543e-01 1.32167891e-01 4.58799392e-01 -9.34741139e-01
-4.92346644e-01 -1.78452611e-01 7.64012635e-01 2.77723432e-01
-8.77274126e-02 5.79491973e-01 3.28211010e-01 -6.61198556e-01
-1.09126568e-01 -5.90548575e-01 -3.11028510e-02 -2.32125878e-01
-7.57203400e-01 -5.25414705e-01 6.05919421e-01 -6.42663598e-01
1.29759407e+00 -1.75305331e+00 -2.07447916e-01 -1.39727101e-01
7.08817169e-02 4.35977131e-01 -1.69857427e-01 4.05600935e-01
7.42579848e-02 5.07586241e-01 -1.70786157e-01 2.12075543e-02
3.43560636e-01 1.46897301e-01 -7.53012002e-01 1.51445404e-01
3.44604880e-01 1.21768117e+00 -1.13194561e+00 -6.00795686e-01
-1.62883043e-01 3.72398853e-01 -7.16697156e-01 1.70329317e-01
-3.24533015e-01 -6.32956699e-02 -7.40351737e-01 5.47155023e-01
3.54221731e-01 -7.02600658e-01 1.44958407e-01 -3.33933592e-01
-1.11682331e-02 1.48461235e+00 -8.06550741e-01 1.49001789e+00
-5.15209556e-01 4.17730689e-01 3.50670457e-01 -9.27887738e-01
6.12998545e-01 2.40662068e-01 -2.46503428e-01 -7.87032306e-01
2.57384986e-01 4.29387420e-01 1.61643028e-01 -3.31072658e-01
3.22295904e-01 -1.34318948e-01 -4.26283568e-01 1.07486129e+00
-5.02396047e-01 1.55751603e-02 2.18322799e-01 7.02968776e-01
1.21985817e+00 1.92358661e-02 1.57936886e-01 -9.98917818e-02
2.97375679e-01 3.09778720e-01 5.16676903e-01 1.08323169e+00
-1.40802950e-01 3.76509488e-01 8.53397310e-01 -6.32003963e-01
-1.21487236e+00 -1.09313834e+00 -6.91307560e-02 1.28118348e+00
-7.52334818e-02 -4.10538763e-01 -5.80037415e-01 -1.19182038e+00
1.68758973e-01 9.97923672e-01 -6.85839057e-01 3.45237106e-01
-8.05477679e-01 -4.92536873e-01 6.45171165e-01 6.45104885e-01
2.37048239e-01 -1.09247780e+00 -7.96977580e-01 5.56547821e-01
-7.34982908e-01 -1.08489883e+00 -2.69723892e-01 1.92606598e-01
-3.67810816e-01 -1.63163173e+00 -2.65111923e-01 -3.44620228e-01
2.02419445e-01 6.67117238e-02 1.42967677e+00 3.85983706e-01
4.43604449e-03 -3.63941975e-02 -3.36557865e-01 -5.63002348e-01
-8.22900891e-01 -1.55849248e-01 -5.34717143e-01 -4.57953274e-01
3.07578474e-01 -2.44408011e-01 -5.17055154e-01 8.25350638e-03
-6.05013490e-01 -1.73121691e-02 3.87354881e-01 1.25545084e+00
1.46490887e-01 -8.29215944e-01 7.43216693e-01 -1.47719574e+00
1.45359290e+00 -4.61283386e-01 -3.74408036e-01 1.63310975e-01
-6.69702351e-01 3.79878700e-01 3.67083818e-01 -5.55306189e-02
-1.38047850e+00 -8.22386086e-01 -6.71794772e-01 3.84367019e-01
2.16829538e-01 6.29221737e-01 1.73510045e-01 5.03863454e-01
1.05810344e+00 -7.69804716e-01 3.19051705e-02 -4.17256087e-01
7.01970458e-01 6.66679263e-01 5.61353147e-01 -1.15360546e+00
6.33390069e-01 5.12238324e-01 -6.70523882e-01 -9.26916953e-03
-2.01387978e+00 -2.59671509e-01 5.25179543e-02 1.73735321e-01
5.70076287e-01 -9.34869230e-01 -6.04256690e-01 -2.39643335e-01
-1.46176362e+00 -2.80316830e-01 -3.32221061e-01 3.50860208e-01
-2.46066883e-01 3.56506616e-01 -1.04174793e+00 -7.50688136e-01
-6.15832627e-01 -1.10258198e+00 9.20119286e-01 -1.42403454e-01
-6.70215607e-01 -9.14676487e-01 1.89274684e-01 9.36345041e-01
5.17938137e-01 1.22299455e-01 1.05262959e+00 -9.92337644e-01
-2.19969656e-02 -1.45973504e-01 -4.60734606e-01 2.56616354e-01
-1.09935887e-01 -2.50315219e-01 -9.91896152e-01 2.29879782e-01
-5.07546291e-02 -9.74495351e-01 1.09779072e+00 3.15785825e-01
7.80498207e-01 -7.04132438e-01 -7.40554780e-02 -1.02754580e-02
7.24965215e-01 -2.69108742e-01 3.09546739e-01 9.15632367e-01
1.53031677e-01 6.80830598e-01 8.28133821e-01 3.49665940e-01
4.14774090e-01 5.56612611e-01 4.15457428e-01 -1.63673356e-01
-2.61503100e-01 -5.43862700e-01 2.55050242e-01 2.92760968e-01
4.02227342e-01 -4.40093905e-01 -7.61174858e-01 7.43357897e-01
-1.91568911e+00 -1.46914816e+00 -1.24690555e-01 1.73717141e+00
9.94890273e-01 6.70002520e-01 5.85714653e-02 2.22796440e-01
8.29264522e-01 5.02070010e-01 -4.15941387e-01 -8.35529923e-01
-4.50878263e-01 2.14855969e-01 1.34974152e-01 1.02410066e+00
-1.14006710e+00 9.47021782e-01 6.58107424e+00 3.81977558e-01
-7.11669147e-01 3.04334253e-01 9.14831460e-01 -1.36650428e-01
-1.19542360e+00 6.65680468e-01 -7.10796535e-01 4.15731877e-01
8.02552402e-01 -3.34371686e-01 3.96700986e-02 8.14421296e-01
2.02210173e-02 2.40334138e-01 -7.91751862e-01 2.51697749e-01
-7.81273320e-02 -1.71371520e+00 -3.71776172e-03 -9.47802663e-02
4.92313683e-01 3.05500507e-01 2.34071925e-01 3.71315807e-01
1.23467660e+00 -8.79323125e-01 1.03039706e+00 -2.34190091e-01
3.69325310e-01 -3.57818067e-01 7.53050148e-01 2.00673625e-01
-5.74333906e-01 -2.93732077e-01 -1.16480105e-01 -4.13855702e-01
7.20036447e-01 6.80878520e-01 -1.04019690e+00 9.24242511e-02
2.87127525e-01 4.83701080e-01 -2.74390101e-01 4.11449403e-01
-8.83586466e-01 1.07870591e+00 -1.56769991e-01 -3.07259977e-01
7.25633681e-01 4.57196623e-01 6.68303967e-01 1.05598998e+00
-2.83103362e-02 2.23525345e-01 2.02589303e-01 1.05089128e+00
-3.50269049e-01 1.84823856e-01 -3.35597038e-01 -3.26674283e-01
7.47964621e-01 1.21791875e+00 -2.80449420e-01 -7.37021327e-01
-3.85739505e-01 4.05170500e-01 7.69686162e-01 1.05944946e-01
-8.88606369e-01 -1.81970559e-02 7.68564820e-01 -5.53599000e-02
1.44221738e-01 2.79236674e-01 -4.29023832e-01 -1.19796944e+00
2.30236411e-01 -1.05529153e+00 8.35266650e-01 -5.97212553e-01
-1.48258674e+00 5.92339337e-01 -2.22517028e-01 -7.12683856e-01
-5.45551717e-01 -4.93754297e-01 -8.06453764e-01 7.23012984e-01
-1.87436831e+00 -1.01819527e+00 2.13812023e-01 2.25927737e-02
7.20567822e-01 9.02419314e-02 8.49448740e-01 5.61341345e-02
-1.28543481e-01 3.69571537e-01 -4.08860087e-01 4.37335610e-01
8.87025237e-01 -1.19627464e+00 9.58730519e-01 7.32339382e-01
4.43294913e-01 7.18145549e-01 9.28318977e-01 -7.73356736e-01
-9.08778906e-01 -5.42889297e-01 1.41284716e+00 -8.74930084e-01
9.50790405e-01 -1.54912755e-01 -9.09561157e-01 4.85290617e-01
4.21954900e-01 -1.94068372e-01 8.23587537e-01 9.03371990e-01
-1.09710765e+00 3.75073850e-01 -9.55092192e-01 4.57775652e-01
1.00189030e+00 -8.48746479e-01 -1.29642701e+00 3.50733757e-01
8.09262633e-01 -4.44176286e-01 -4.53559726e-01 3.27931494e-01
6.99758172e-01 -1.00226021e+00 1.07243335e+00 -9.34397399e-01
9.03498709e-01 5.96507899e-02 -1.71828657e-01 -1.11915743e+00
-2.11983755e-01 -4.88154501e-01 -2.48756651e-02 1.10701621e+00
1.01697767e+00 -4.72608000e-01 8.03476691e-01 7.52729893e-01
-1.63766086e-01 -6.47655785e-01 -7.71292388e-01 -3.41237307e-01
3.10822934e-01 -5.12619615e-01 7.93630123e-01 8.78880620e-01
3.23702306e-01 9.35574234e-01 -2.06694737e-01 -1.63189560e-01
5.66731572e-01 6.05730653e-01 7.49713123e-01 -1.34105158e+00
-2.88460582e-01 -6.38073027e-01 3.35508466e-01 -9.70527530e-01
5.64437926e-01 -1.03128600e+00 -1.72840536e-01 -1.87906170e+00
3.96940500e-01 -5.40503502e-01 -3.16144466e-01 3.69206041e-01
-5.72297454e-01 -5.91819398e-02 -3.29722427e-02 3.71656120e-02
-7.29319930e-01 2.20450088e-01 9.61215317e-01 -2.71071464e-01
2.93441743e-01 -6.99236915e-02 -1.37004268e+00 7.57909298e-01
6.33058429e-01 -6.15374446e-01 -1.55868486e-01 -9.38114107e-01
7.80279577e-01 4.14900668e-02 5.02818286e-01 -3.67205083e-01
1.29213840e-01 -1.41806930e-01 6.94732070e-02 -5.82365453e-01
2.89796621e-01 -7.52084032e-02 -5.42760253e-01 2.31438711e-01
-1.31201923e+00 3.79783839e-01 -1.08083956e-01 6.83754921e-01
-3.89970243e-01 -3.35104197e-01 4.22172219e-01 -1.87660232e-01
-2.37657785e-01 -1.07954599e-01 2.76389997e-02 8.81677330e-01
3.57419163e-01 2.98258156e-01 -9.11298156e-01 -6.86446011e-01
-4.53899920e-01 1.85996234e-01 3.03347379e-01 4.23311472e-01
5.22052884e-01 -1.27983010e+00 -1.32551360e+00 -3.30743194e-01
1.14549145e-01 -3.98012877e-01 -7.30040297e-02 5.30004799e-01
-1.13247618e-01 3.59014720e-01 3.03463429e-01 -1.19120099e-01
-1.30012202e+00 3.38802725e-01 -1.82019293e-01 -5.23742199e-01
-6.77202284e-01 1.03798890e+00 -1.95344031e-01 -4.55175430e-01
2.64283478e-01 -2.44204685e-01 -3.01315874e-01 1.17735766e-01
7.62144923e-01 1.53920889e-01 -8.51154551e-02 -3.99054229e-01
-2.54214644e-01 3.63827683e-02 -3.03893626e-01 -5.33408225e-01
1.38894236e+00 8.13147575e-02 8.12130794e-02 2.18348399e-01
1.05274069e+00 5.67957044e-01 -1.04750645e+00 -4.61312592e-01
3.28444511e-01 -5.75204968e-01 -4.56062742e-02 -1.23562479e+00
-2.54500628e-01 9.03322160e-01 -9.00014862e-02 3.41156602e-01
2.53244936e-01 2.73705512e-01 9.43562508e-01 4.87310082e-01
-5.05411364e-02 -1.22302401e+00 3.35655570e-01 6.92249894e-01
9.18716252e-01 -1.44989145e+00 -9.63505507e-02 -2.64319897e-01
-8.64213526e-01 9.61091757e-01 2.61087120e-01 -8.22141618e-02
2.33398810e-01 6.20065928e-01 5.69306374e-01 -5.30644298e-01
-1.24819112e+00 -9.19915549e-03 8.49802420e-02 -2.12854333e-03
1.10913050e+00 5.18516637e-02 -6.58681870e-01 6.85050011e-01
-3.68215621e-01 -4.78832304e-01 3.98225427e-01 8.37632835e-01
-4.73821521e-01 -1.38990629e+00 -2.33577237e-01 3.98985267e-01
-1.07649326e+00 -5.16856134e-01 -8.12140167e-01 3.29901844e-01
-4.72207189e-01 1.32382524e+00 -2.51019418e-01 1.75064057e-02
3.96834016e-01 5.76766720e-03 8.29566792e-02 -5.78927636e-01
-1.00474691e+00 -9.91493389e-02 1.18585110e+00 -6.22766316e-01
-4.82001990e-01 -7.39441633e-01 -1.22757065e+00 -2.24861309e-01
-4.46535647e-01 8.88644814e-01 7.17122853e-01 9.76389229e-01
4.69819337e-01 3.51417929e-01 1.71731099e-01 -6.14931174e-02
-1.25692260e+00 -8.23055506e-01 3.06957930e-01 8.18777740e-01
4.48055536e-01 -7.62091219e-01 -5.32746792e-01 -3.87872994e-01] | [9.749860763549805, 9.475616455078125] |
cf77e654-cae5-4c65-ace9-800f257c6c08 | 190600772 | 1906.00772 | null | https://arxiv.org/abs/1906.00772v1 | https://arxiv.org/pdf/1906.00772v1.pdf | Dynamic Service Composition Orchestrated by Cognitive Agents in Mobile & Pervasive Computing | Automatic service composition in mobile and pervasive computing faces many challenges due to the complex nature of the environment. Common approaches address service composition from optimization perspectives which are not feasible in practice due to the intractability of the problem, limited computational resources of smart devices, service host's mobility, and time constraints. Our main contribution is the development of a cognitively-inspired agent-based service composition model focused on bounded rationality rather than optimality, which allows the system to compensate for limited resources by selectively filtering out continuous streams of data. The evaluation of our approach shows promising results when compared against state-of-the-art service composition models. | ['Oscar J. Romero'] | 2019-05-31 | null | null | null | null | ['service-composition'] | ['miscellaneous'] | [-7.73152485e-02 2.14674354e-01 -1.81244582e-01 -4.59787436e-02
2.84311250e-02 -4.83845264e-01 6.07234538e-01 -3.03964287e-01
-2.78876096e-01 6.99823201e-01 2.03121737e-01 -2.49093071e-01
-6.49795771e-01 -7.49834716e-01 7.07086995e-02 -6.89122200e-01
-2.20911548e-01 9.41760898e-01 5.67152977e-01 -6.71600819e-01
-9.50817168e-02 3.99440229e-01 -2.05737305e+00 1.61028937e-01
9.16141033e-01 1.14454842e+00 4.80837196e-01 4.95042890e-01
-1.47120684e-01 5.62091172e-01 -2.98355460e-01 -9.24846753e-02
2.75530636e-01 -4.00438786e-01 -6.31550252e-01 3.93548369e-01
-9.62499022e-01 -1.51098087e-01 2.69740611e-01 1.12786722e+00
3.93017679e-01 3.27546090e-01 3.22006941e-01 -1.99120855e+00
-4.77076054e-01 5.37621319e-01 3.63330454e-01 8.47385749e-02
7.92566001e-01 -1.25321031e-01 1.03747702e+00 -3.91636282e-01
9.72624958e-01 7.83904314e-01 3.18776548e-01 9.27620232e-01
-1.01662850e+00 -8.13284293e-02 9.13510442e-01 6.02640450e-01
-1.40323937e+00 -7.06064522e-01 5.06829083e-01 7.87510425e-02
1.31994569e+00 9.69168544e-01 8.13746572e-01 8.27548802e-01
-1.42019883e-01 6.46559000e-01 8.95068049e-01 -4.43803668e-01
9.24604893e-01 1.07218698e-01 -3.36999327e-01 9.67647731e-02
8.58209804e-02 -1.17864519e-01 -3.38508785e-01 -4.83156890e-01
1.31823435e-01 2.31520921e-01 -8.76162648e-02 -6.25621974e-01
-8.76475692e-01 4.05959606e-01 -1.71300247e-01 5.21639705e-01
-1.03013456e+00 -2.88354397e-01 1.92730248e-01 4.66983408e-01
3.80896270e-01 1.61416054e-01 -7.78012395e-01 -6.61530674e-01
-6.31952286e-01 4.95928109e-01 1.33775961e+00 1.26685190e+00
1.35957584e-01 6.60351440e-02 2.89866120e-01 5.15152335e-01
5.76797545e-01 2.06442609e-01 4.90551949e-01 -1.20200670e+00
-6.48818463e-02 6.84857249e-01 6.22757733e-01 -5.17632782e-01
-4.99856412e-01 -2.00927019e-01 -2.93068498e-01 1.02077380e-01
2.02582136e-01 -1.65789962e-01 -3.27176124e-01 1.51413882e+00
5.45103312e-01 2.54961491e-01 2.47457191e-01 1.22800398e+00
3.45205665e-01 2.75109410e-01 -8.23849514e-02 -8.93863499e-01
1.17753994e+00 -9.44473624e-01 -8.19003940e-01 2.25624427e-01
1.41631529e-01 -2.78718352e-01 1.19912553e+00 3.54045361e-01
-1.26584220e+00 6.13551557e-01 -7.28845894e-01 6.67019844e-01
-2.39598751e-01 -7.18630314e-01 7.12934554e-01 1.00009692e+00
-1.12126362e+00 -4.25865427e-02 -6.87070847e-01 -5.52728593e-01
9.58284810e-02 5.79404831e-01 2.16282815e-01 2.32050475e-02
-8.19754183e-01 6.01891935e-01 -1.48220599e-01 -1.53530553e-01
-8.81226778e-01 -2.57426560e-01 -3.80954593e-01 3.84311408e-01
8.80689561e-01 -7.81973064e-01 1.39423752e+00 -1.27546453e+00
-2.07668710e+00 3.07753980e-01 -6.55796565e-03 -2.35184431e-01
8.79549563e-01 5.70797682e-01 -6.36072457e-01 -6.32548258e-02
6.65246844e-02 -3.39220315e-01 3.49974215e-01 -1.38569844e+00
-1.28749025e+00 -2.77970642e-01 4.12287444e-01 6.91154838e-01
-2.23670959e-01 3.13902736e-01 -3.00342768e-01 -5.07683046e-02
3.32698077e-01 -1.04390812e+00 -9.09366369e-01 -7.59365439e-01
2.60768145e-01 -1.66789085e-01 5.99605918e-01 -9.77002084e-02
1.10385728e+00 -1.94126308e+00 2.42197320e-01 4.02221411e-01
-1.99168205e-01 -1.75605893e-01 -2.54533976e-01 6.27027988e-01
6.06642365e-01 9.69844759e-02 2.61862040e-01 -2.20292717e-01
4.83704001e-01 5.97102284e-01 -2.88404375e-02 2.77350008e-01
-6.10316336e-01 3.87052774e-01 -9.35424387e-01 6.76363036e-02
-2.46392921e-01 1.99898034e-01 -7.98969209e-01 2.14985281e-01
-7.96295762e-01 4.29243386e-01 -6.10000372e-01 1.17145503e+00
3.83187741e-01 1.51325846e-02 7.28238225e-01 7.15372145e-01
-1.92370370e-01 5.52647889e-01 -1.50813615e+00 1.47310674e+00
-5.84098279e-01 -2.09437892e-01 7.45923281e-01 -8.68635297e-01
1.88030422e-01 9.32672560e-01 1.02947521e+00 -8.38174164e-01
7.83617496e-02 5.03329039e-01 3.83757770e-01 -6.01905525e-01
3.51728231e-01 8.61753151e-02 5.44614578e-03 9.14291978e-01
-5.05753398e-01 7.05443472e-02 1.70050457e-01 -6.75559938e-02
1.39880681e+00 1.49951071e-01 1.61323339e-01 -3.43308806e-01
7.79916108e-01 -1.16424426e-01 9.85821724e-01 7.28517830e-01
-5.71108401e-01 8.96759331e-02 3.45753163e-01 -4.88527238e-01
-7.29685247e-01 -6.02501094e-01 4.82845813e-01 1.57160044e+00
5.16833782e-01 -4.14673299e-01 -8.79198194e-01 -7.49868453e-01
-1.25578001e-01 7.25327373e-01 -2.08605349e-01 3.00420493e-01
-4.44415778e-01 -6.96002841e-01 2.26086704e-03 -1.67133678e-02
-3.75094190e-02 -1.16766584e+00 -9.40708339e-01 5.21422625e-01
-3.42860639e-01 -1.47760510e+00 -2.35729754e-01 -2.37872481e-01
-5.44438303e-01 -8.56399655e-01 -9.16099250e-02 -2.63872772e-01
5.92116356e-01 6.65622592e-01 9.98347461e-01 4.28342730e-01
3.90505977e-02 6.78422928e-01 -5.66251993e-01 -4.24688727e-01
-1.41625315e-01 1.61965370e-01 5.74477315e-01 3.16074729e-01
2.91452497e-01 -9.27663445e-01 -7.51878858e-01 6.43644929e-01
-8.64686251e-01 -2.39103064e-01 -7.68080074e-03 4.84102070e-01
5.07893026e-01 5.82537591e-01 9.79244351e-01 -6.91357613e-01
1.03046322e+00 -8.05696368e-01 -5.54411769e-01 1.82040542e-01
-1.09287965e+00 -5.46634674e-01 5.89193046e-01 -4.71574008e-01
-1.10235846e+00 -2.23709345e-01 -1.23607308e-01 3.52866650e-01
-1.34282887e-01 3.06222200e-01 -6.37262046e-01 -6.11713678e-02
1.19966239e-01 2.83346653e-01 9.11668167e-02 -3.63075465e-01
1.57288760e-01 8.87859344e-01 -4.64046597e-02 -7.24375188e-01
3.73830944e-01 8.38345170e-01 -1.22600392e-01 -6.32591784e-01
6.38125390e-02 -5.87942541e-01 -2.78290734e-02 -3.93641710e-01
2.66334742e-01 -3.96497846e-01 -1.32348490e+00 1.67611577e-02
-7.17182219e-01 -3.34585547e-01 -3.24483991e-01 1.35478884e-01
-8.84983480e-01 1.91585660e-01 -4.37178612e-01 -1.36289763e+00
-3.34761351e-01 -1.36581087e+00 5.24833143e-01 1.47779778e-01
-2.33529031e-01 -5.70936084e-01 -2.86277711e-01 4.03922677e-01
9.47355807e-01 -3.31118822e-01 6.04286551e-01 -9.09731746e-01
-8.81745279e-01 -4.01796460e-01 2.15277091e-01 -3.26954395e-01
1.17525809e-01 -1.73114195e-01 -4.48969036e-01 -2.13609755e-01
2.59331107e-01 4.02718008e-01 -3.36200744e-01 1.50628731e-01
7.32988954e-01 -6.83982074e-01 -3.48302543e-01 3.42934549e-01
1.41166866e+00 6.79214060e-01 4.48277831e-01 8.74463975e-01
4.78639677e-02 8.09310138e-01 8.78452837e-01 9.76510763e-01
9.34021890e-01 1.10921013e+00 1.04426110e+00 6.57689512e-01
1.96107075e-01 2.06541136e-01 2.32397050e-01 6.74967289e-01
-7.63874114e-01 -2.73322493e-01 -7.55844355e-01 5.45836926e-01
-2.66312766e+00 -1.10344613e+00 1.75200868e-02 2.13800669e+00
2.02194348e-01 1.49797024e-02 5.37076831e-01 4.01206464e-01
2.28073269e-01 -4.21358109e-01 -3.67348582e-01 -6.18632913e-01
-7.04265982e-02 -5.38886607e-01 3.11899006e-01 5.40137231e-01
-5.10878921e-01 7.49035597e-01 6.57304525e+00 3.09680402e-01
-7.43781745e-01 8.65269363e-01 -6.90544862e-03 -4.98723775e-01
-6.73030198e-01 -3.60976309e-02 -2.63416439e-01 7.47028530e-01
1.18445373e+00 -4.21399176e-01 1.38802588e+00 9.28070068e-01
5.63120067e-01 -7.87815005e-02 -9.42272902e-01 7.06676364e-01
-6.94992095e-02 -1.34192562e+00 -5.48840284e-01 3.51549417e-01
7.87058055e-01 1.85708851e-01 -2.06552610e-01 -8.71394388e-03
3.96905363e-01 -4.97553796e-01 1.29149318e+00 4.23383921e-01
9.89087448e-02 -6.85561776e-01 6.80646241e-01 7.40009367e-01
-9.50517774e-01 -7.50872850e-01 1.08338907e-01 -7.09306002e-01
5.35549045e-01 2.66302764e-01 -4.65242475e-01 3.86487007e-01
9.07817543e-01 -1.73319504e-01 1.64716795e-01 8.94648135e-01
3.90480198e-02 1.00647904e-01 -3.19884956e-01 -4.33574677e-01
9.96208005e-03 -5.54241836e-01 7.74157107e-01 7.89436281e-01
3.49208534e-01 4.09382045e-01 4.39204127e-01 4.33235794e-01
2.79936284e-01 1.55114278e-01 -4.48069930e-01 1.90824002e-01
8.32112312e-01 1.07117248e+00 -9.27063346e-01 -3.15532595e-01
-6.19437754e-01 8.47678721e-01 -1.72979549e-01 6.47655189e-01
-5.59507310e-01 2.18277067e-01 1.14803457e+00 2.67470509e-01
4.42914143e-02 -2.04364464e-01 -5.54194033e-01 -1.12826478e+00
1.55780911e-01 -1.18074000e+00 4.55772817e-01 -3.85920942e-01
-8.61380219e-01 8.32293570e-01 -5.12442708e-01 -1.03150880e+00
-4.79879230e-01 -5.26939444e-02 -3.15097868e-01 4.64387238e-01
-1.37161815e+00 -9.01013017e-01 -9.46873799e-02 7.51107156e-01
6.10377133e-01 -5.64864397e-01 1.30293787e+00 6.07880950e-01
-3.69576514e-01 1.95237160e-01 1.41273364e-01 -8.96054745e-01
1.20221488e-02 -8.78610015e-01 -2.16165017e-02 9.65044618e-01
-3.32302123e-01 2.10214928e-01 1.10400712e+00 -3.08532476e-01
-1.90117884e+00 -2.80172765e-01 1.26280200e+00 -3.18531543e-01
6.53984606e-01 -3.50483149e-01 -3.85054469e-01 4.70721722e-01
-8.71199369e-02 6.21237792e-02 8.22742522e-01 3.07942554e-03
1.95908211e-02 -3.71664017e-01 -1.64052796e+00 1.01225388e+00
1.48224652e+00 -2.33321339e-01 7.07439408e-02 2.95618355e-01
6.97368324e-01 -7.17367139e-03 -3.32878053e-01 2.21999228e-01
7.44170785e-01 -9.23287868e-01 6.61117673e-01 -6.90194666e-01
-3.11268866e-01 -4.74979460e-01 -7.63547659e-01 -1.02918029e+00
-2.88971901e-01 -1.01821280e+00 -3.37702930e-01 1.03831422e+00
4.46538925e-01 -9.92750525e-01 9.27638888e-01 1.39057136e+00
-1.74087703e-01 -3.78924370e-01 -1.38228750e+00 -9.13392127e-01
-8.48425686e-01 -6.55572474e-01 1.40395570e+00 7.95392811e-01
4.72123444e-01 -2.13575378e-01 1.92671120e-02 2.22613931e-01
4.57393289e-01 2.47072861e-01 3.47215801e-01 -1.28088069e+00
-4.12219912e-01 -5.76973498e-01 -2.04314739e-01 -4.06944335e-01
2.60691255e-01 -4.87328172e-01 3.09674174e-01 -1.67549026e+00
-1.33276999e-01 -1.04461277e+00 -4.52424258e-01 2.32297823e-01
6.44322574e-01 1.48834229e-01 3.14121455e-01 4.03911680e-01
-1.13058662e+00 4.23333347e-01 7.72297204e-01 1.08191036e-01
-6.16536200e-01 6.33536756e-01 -7.98198283e-01 7.72368908e-01
7.53470242e-01 -3.46568495e-01 -6.15650713e-01 -3.04348201e-01
5.84361672e-01 3.87974471e-01 -3.55667353e-01 -5.03567815e-01
4.12083805e-01 -9.81798708e-01 -8.68548751e-01 2.30522275e-01
4.12867278e-01 -1.47230530e+00 7.15058446e-01 5.57699382e-01
-7.47575909e-02 2.37880379e-01 -6.27708018e-01 5.39246917e-01
2.87223229e-04 -3.03138465e-01 4.22243737e-02 -2.16844991e-01
-7.00241625e-01 2.56673604e-01 -9.46037412e-01 -5.43541193e-01
1.32099092e+00 -2.00660408e-01 -1.43074259e-01 -6.24757409e-01
-8.65792096e-01 2.29016021e-02 7.74114549e-01 6.73983753e-01
3.78351718e-01 -1.27697039e+00 -3.52878600e-01 -8.20306912e-02
-5.63139953e-02 -5.18428028e-01 3.86070386e-02 9.66125429e-01
-4.48172599e-01 3.38390291e-01 -3.24529320e-01 -1.80380777e-01
-1.30041397e+00 9.49082255e-01 2.17060149e-01 7.11115971e-02
-5.63851178e-01 1.77860796e-01 -4.12597716e-01 -4.16809082e-01
4.83906388e-01 3.86029370e-02 -7.08455071e-02 -5.98471016e-02
5.99003255e-01 7.04017758e-01 2.28154793e-01 -7.31863260e-01
-8.29446375e-01 -2.19881877e-01 5.06630838e-01 -7.02649355e-01
1.29783738e+00 -8.71679723e-01 -4.96838123e-01 -3.13344114e-02
1.63349062e-01 -3.98085103e-04 -9.25128996e-01 -2.37983614e-01
3.10161650e-01 -6.28150403e-01 -9.19454619e-02 -9.71442044e-01
-8.20173383e-01 -1.85918495e-01 4.84378070e-01 9.79095817e-01
1.25742781e+00 -2.45325699e-01 4.94874537e-01 9.56550017e-02
9.68844056e-01 -1.53861070e+00 -3.70603502e-01 4.29821998e-01
5.68481207e-01 -7.99813569e-01 -3.79142225e-01 -5.88473439e-01
-6.67558074e-01 5.36109626e-01 2.70644963e-01 1.48066059e-01
7.99280167e-01 4.83777612e-01 9.65955555e-02 -4.84739803e-02
-1.37550139e+00 -6.19506538e-01 -3.37237567e-01 1.12777686e+00
1.09596765e-02 4.97415990e-01 -1.06482148e+00 1.21155405e+00
1.24789894e-01 6.79168059e-03 8.59953284e-01 1.12217569e+00
-2.33092502e-01 -1.45328426e+00 -2.71755427e-01 4.10801359e-02
-6.31089926e-01 1.97378904e-01 -9.96964648e-02 1.29444227e-01
2.83246458e-01 1.63209260e+00 -1.99939132e-01 -1.27288699e-01
5.23832262e-01 -4.15247306e-02 1.03661560e-01 -5.64673483e-01
-6.17822111e-01 3.63566607e-01 4.80168968e-01 -7.70328522e-01
-5.53344309e-01 -1.17366278e+00 -1.45252228e+00 -2.45624423e-01
6.60360791e-03 2.88533360e-01 1.10876966e+00 1.12173533e+00
6.16855323e-01 6.84032679e-01 7.59730637e-01 -8.41736019e-01
-4.92613703e-01 -4.15085554e-01 -5.34458697e-01 4.03655529e-01
2.00554356e-02 -4.89954740e-01 -1.99995384e-01 -1.93389982e-01] | [8.616639137268066, 6.927917957305908] |
9c471d22-f96f-4b8d-b2ed-c7f949986be3 | modeling-label-correlations-for-ultra-fine | 2212.01581 | null | https://arxiv.org/abs/2212.01581v1 | https://arxiv.org/pdf/2212.01581v1.pdf | Modeling Label Correlations for Ultra-Fine Entity Typing with Neural Pairwise Conditional Random Field | Ultra-fine entity typing (UFET) aims to predict a wide range of type phrases that correctly describe the categories of a given entity mention in a sentence. Most recent works infer each entity type independently, ignoring the correlations between types, e.g., when an entity is inferred as a president, it should also be a politician and a leader. To this end, we use an undirected graphical model called pairwise conditional random field (PCRF) to formulate the UFET problem, in which the type variables are not only unarily influenced by the input but also pairwisely relate to all the other type variables. We use various modern backbones for entity typing to compute unary potentials, and derive pairwise potentials from type phrase representations that both capture prior semantic information and facilitate accelerated inference. We use mean-field variational inference for efficient type inference on very large type sets and unfold it as a neural network module to enable end-to-end training. Experiments on UFET show that the Neural-PCRF consistently outperforms its backbones with little cost and results in a competitive performance against cross-encoder based SOTA while being thousands of times faster. We also find Neural- PCRF effective on a widely used fine-grained entity typing dataset with a smaller type set. We pack Neural-PCRF as a network module that can be plugged onto multi-label type classifiers with ease and release it in https://github.com/modelscope/adaseq/tree/master/examples/NPCRF. | ['Kewei Tu', 'Pengjun Xie', 'Weiqi Wu', 'Yong Jiang', 'Chengyue Jiang'] | 2022-12-03 | null | null | null | null | ['entity-typing', 'type'] | ['natural-language-processing', 'speech'] | [-4.13483791e-02 2.47066051e-01 -5.75301230e-01 -6.94676638e-01
-6.91667199e-01 -7.77978301e-01 4.02190119e-01 1.36585802e-01
-5.37107825e-01 1.23593140e+00 1.70319110e-01 -4.87508774e-01
1.56129375e-01 -1.15162921e+00 -1.35403204e+00 -4.56371963e-01
-1.33972168e-02 9.72264528e-01 -2.04525977e-01 1.05486922e-01
-1.40877068e-01 -2.49037609e-01 -1.35409331e+00 5.05372107e-01
8.93503785e-01 8.45948756e-01 2.08024099e-01 5.68919182e-01
-2.92357177e-01 8.30982029e-01 -4.86159921e-01 -1.08420050e+00
-2.35866100e-01 3.59549582e-01 -1.11553538e+00 -6.19910121e-01
6.58043802e-01 -6.62585124e-02 -1.17945194e-01 9.48582232e-01
4.57615077e-01 -5.39546609e-02 9.28382754e-01 -1.32812858e+00
-8.89680028e-01 8.92180979e-01 -5.57020724e-01 -3.26790437e-02
3.89123440e-01 -6.08983561e-02 1.64182508e+00 -9.13686752e-01
8.51548493e-01 1.42249155e+00 1.02953124e+00 7.95294762e-01
-1.43132150e+00 -7.34426796e-01 1.64070547e-01 3.37063044e-01
-1.28651094e+00 -2.66360044e-01 4.06005114e-01 -4.09938663e-01
1.30312061e+00 3.35891783e-01 1.95206627e-01 1.56172788e+00
3.10296744e-01 9.95017350e-01 9.38960314e-01 -1.87185779e-01
1.92310616e-01 1.26054734e-01 6.20653629e-01 8.47508609e-01
2.95676082e-01 -1.83393419e-01 -4.04008776e-01 -6.39205158e-01
2.42591187e-01 -3.34291875e-01 2.31859293e-02 6.48319647e-02
-1.01186204e+00 7.79445231e-01 4.00869370e-01 -7.05755949e-02
-2.24944890e-01 4.00579035e-01 3.18091661e-01 6.60321768e-03
4.95556980e-01 4.02942955e-01 -1.05807519e+00 -1.74745247e-01
-5.77840090e-01 6.69778407e-01 1.17127037e+00 1.17042625e+00
9.79506731e-01 -4.28696305e-01 -4.24674928e-01 1.15555561e+00
3.29076856e-01 6.16710544e-01 -1.24164507e-01 -8.53830278e-01
5.11103392e-01 4.94591981e-01 4.96848002e-02 -4.05596733e-01
-2.85842478e-01 -3.34690213e-01 -7.58525908e-01 -2.65272796e-01
4.31027263e-01 -5.30788898e-01 -8.17950189e-01 2.02421737e+00
5.66777945e-01 5.23911774e-01 -1.40959188e-01 5.35648108e-01
7.98671126e-01 6.06885791e-01 4.28910792e-01 2.36729652e-01
1.81979740e+00 -5.69754481e-01 -5.05573630e-01 -3.71463507e-01
7.61240661e-01 -3.64591330e-01 7.11487830e-01 1.08681381e-01
-8.96935046e-01 -1.49302870e-01 -6.75564945e-01 -4.29052740e-01
-6.04132950e-01 -8.38842988e-02 8.13151598e-01 3.94404054e-01
-7.95588315e-01 6.03820324e-01 -7.14963853e-01 1.67874902e-01
5.85046172e-01 2.94489563e-01 -3.77101332e-01 -1.46338135e-01
-1.69037735e+00 7.08555222e-01 4.06681329e-01 2.51653910e-01
-4.21321452e-01 -1.35131466e+00 -9.13388371e-01 1.11994945e-01
4.03841436e-01 -1.34594703e+00 1.30868411e+00 -6.52165592e-01
-8.86188507e-01 7.35670149e-01 -8.34962010e-01 -3.95609617e-01
9.07344371e-02 -9.53125432e-02 -2.83050835e-01 -3.60070407e-01
3.23086530e-01 6.24238431e-01 3.52459222e-01 -1.02808666e+00
-8.60498130e-01 -2.33051404e-01 2.90737450e-01 4.26233932e-02
2.97604967e-02 -9.61266086e-02 -2.52362192e-01 -5.34271896e-01
-5.89463413e-01 -1.00333214e+00 4.83540930e-02 -3.73002291e-01
-6.99981272e-01 -8.97653282e-01 4.14176196e-01 -7.27062166e-01
1.18765199e+00 -1.72694528e+00 2.76896119e-01 6.47427812e-02
2.63135314e-01 1.56142572e-02 9.72950459e-02 1.67652696e-01
-1.44369183e-02 4.28127110e-01 -4.29588377e-01 -4.62854892e-01
4.87462401e-01 5.91553450e-01 -1.20015778e-01 1.52771458e-01
3.32491487e-01 1.09448111e+00 -1.02691317e+00 -4.03832048e-01
-3.57307047e-01 5.53360403e-01 -1.03938973e+00 1.92605197e-01
-7.42463350e-01 1.72579419e-02 -4.01589781e-01 5.50722599e-01
8.29905629e-01 -3.87626141e-01 4.29567069e-01 -2.45384708e-01
3.00448202e-02 7.33803213e-01 -1.02637684e+00 1.41690266e+00
-6.82567179e-01 1.83083788e-01 6.64776340e-02 -7.47467458e-01
4.05025959e-01 1.88039958e-01 1.16666563e-01 -3.06215405e-01
-2.68987238e-01 1.43824294e-01 -1.73087254e-01 -4.67425644e-01
6.26638412e-01 -1.13178059e-01 -4.48684782e-01 2.43835226e-01
3.01445901e-01 6.16931319e-01 2.76885837e-01 3.13000828e-01
1.00496638e+00 3.33628178e-01 -4.62344624e-02 -2.22708866e-01
1.36285394e-01 -1.11858539e-01 1.23364425e+00 1.01491213e+00
3.76694769e-01 2.44171247e-01 7.93735325e-01 -2.29084656e-01
-1.03097141e+00 -1.20214558e+00 -5.20478845e-01 1.22261775e+00
-2.88532991e-02 -6.52910173e-01 -5.19288123e-01 -9.35592175e-01
3.49828988e-01 8.94518435e-01 -7.37160861e-01 2.71893501e-01
-4.85214323e-01 -9.79646862e-01 8.04047108e-01 6.84155643e-01
2.30461568e-01 -1.02207768e+00 2.52685755e-01 2.08278537e-01
-6.59618556e-01 -9.41403270e-01 -2.93179572e-01 2.18709856e-01
-1.30073026e-01 -9.80759919e-01 -4.18800443e-01 -6.57885134e-01
5.53717613e-01 -5.75378537e-01 1.43357897e+00 6.49273992e-02
3.14330421e-02 -1.35755077e-01 -2.16200694e-01 -3.00496399e-01
-2.09190264e-01 4.80884045e-01 8.83166939e-02 -7.66094178e-02
5.46843886e-01 -4.98749465e-01 -3.57713461e-01 -5.55177443e-02
-3.57356995e-01 1.03892595e-01 4.65030402e-01 1.19960451e+00
5.34006774e-01 -5.53150922e-02 3.39219123e-01 -1.73202658e+00
1.17514335e-01 -1.02458417e+00 -5.01060724e-01 5.12364566e-01
-5.24741828e-01 4.26139534e-01 6.27414346e-01 -1.83298051e-01
-1.37282228e+00 -3.37956399e-01 -6.13314033e-01 -2.36991853e-01
-2.30912313e-01 9.08181190e-01 -4.08985287e-01 6.74497962e-01
3.06268334e-01 8.62749591e-02 -4.27328587e-01 -5.61031878e-01
2.86214620e-01 8.89566123e-01 5.42605579e-01 -1.20095229e+00
6.19436145e-01 1.74174290e-02 -1.57234922e-01 -5.79613149e-01
-1.09524703e+00 -4.14449483e-01 -4.90200698e-01 2.27825209e-01
8.67848754e-01 -1.11289191e+00 -1.04581487e+00 6.48758054e-01
-1.27886915e+00 -6.26832068e-01 3.51367146e-01 1.08893953e-01
-2.55193621e-01 2.40226492e-01 -9.15933549e-01 -6.10730886e-01
-3.86809707e-01 -9.20368791e-01 1.32870650e+00 1.72532067e-01
-2.54935712e-01 -1.37606609e+00 -1.72148850e-02 3.35072368e-01
8.05401579e-02 4.45042271e-03 1.27821219e+00 -5.56304336e-01
-6.37925446e-01 1.80780873e-01 -1.95426330e-01 1.13851637e-01
-2.27828190e-01 1.10216126e-01 -9.31065738e-01 -8.73859879e-03
-7.78538167e-01 -1.08657368e-01 8.79391551e-01 2.98478961e-01
1.13227117e+00 -8.00942898e-01 -5.46396852e-01 8.87902260e-01
1.44778264e+00 -2.99858838e-01 6.66426659e-01 3.93095091e-02
1.20982039e+00 3.77562433e-01 4.59804595e-01 2.11142063e-01
9.98938978e-01 7.68478274e-01 5.44201322e-02 1.81933090e-01
2.07641214e-01 -5.95392823e-01 4.61226493e-01 4.88421828e-01
2.80176532e-02 -3.64005268e-01 -7.78540671e-01 4.63654190e-01
-1.86139941e+00 -1.30979466e+00 -3.82003635e-01 1.83917296e+00
1.32236445e+00 7.67070875e-02 1.16948910e-01 -4.44402486e-01
7.29649842e-01 -5.87851107e-02 -4.57936198e-01 -5.29869735e-01
-6.27638176e-02 4.00835037e-01 7.47334540e-01 7.60313451e-01
-1.07022297e+00 9.73589480e-01 5.14310694e+00 1.04417789e+00
-7.60482371e-01 2.68409878e-01 6.62593246e-01 1.54010832e-01
-8.36002290e-01 2.49549180e-01 -1.34829116e+00 8.89859378e-01
9.89975691e-01 -1.20429821e-01 4.78717089e-01 8.23634803e-01
-2.02941015e-01 2.55596060e-02 -1.19705868e+00 4.64329541e-01
-4.46046293e-01 -1.55091691e+00 -1.59221187e-01 2.33593062e-01
3.84262234e-01 2.26014286e-01 -1.66025221e-01 7.77183175e-01
7.36847043e-01 -9.62494016e-01 8.30909669e-01 5.24355888e-01
1.11827612e+00 -7.66269982e-01 7.20485270e-01 3.53203058e-01
-1.14952564e+00 6.36198893e-02 -4.59772021e-01 -8.66139680e-02
1.01493180e-01 8.87026966e-01 -9.31755781e-01 8.00171435e-01
5.64633489e-01 8.37587833e-01 -3.13612670e-01 6.71832383e-01
-5.16776919e-01 8.10213387e-01 -3.06813002e-01 -2.68972933e-01
-1.12985060e-01 2.15953421e-02 5.32200456e-01 1.44568443e+00
-8.88369791e-03 4.43798862e-02 8.23815819e-04 1.27219892e+00
-3.96691203e-01 -4.36183959e-01 -3.64782572e-01 1.48991153e-01
9.81513441e-01 1.44285727e+00 8.83786753e-02 -5.86241901e-01
-5.46214819e-01 7.66055644e-01 8.67703319e-01 4.45007533e-01
-1.11455381e+00 -3.27322990e-01 1.02631700e+00 -1.39947524e-02
5.63615441e-01 6.92127869e-02 -1.56227782e-01 -1.52946067e+00
8.90950114e-02 -6.28288925e-01 6.73255980e-01 -7.20996678e-01
-1.85012007e+00 2.19288737e-01 2.47923974e-02 -5.21167338e-01
-4.13818270e-01 -8.26625109e-01 -4.03708875e-01 1.28875554e+00
-1.54721117e+00 -1.25555372e+00 1.96849570e-01 3.31623852e-01
1.77877888e-01 1.98044464e-01 1.01915073e+00 4.67367351e-01
-9.89070773e-01 1.20646799e+00 -1.50826260e-01 6.21005893e-01
5.80845237e-01 -1.86517537e+00 6.64577305e-01 6.90012634e-01
-1.37004361e-01 1.06061327e+00 5.74053824e-01 -8.43600571e-01
-1.63773537e+00 -1.42626250e+00 1.45839965e+00 -8.35732043e-01
7.16834724e-01 -9.19570744e-01 -9.52661335e-01 1.04143405e+00
9.41104293e-02 -5.56708872e-02 7.54622936e-01 1.07105935e+00
-8.09876740e-01 9.23058838e-02 -9.82442379e-01 3.07505459e-01
1.14923263e+00 -6.07450604e-01 -7.61063039e-01 2.74604052e-01
6.45320892e-01 -5.71387887e-01 -1.28179634e+00 2.47906297e-01
7.26964712e-01 -6.23480439e-01 8.24053347e-01 -5.44204056e-01
7.27397501e-01 -1.38594836e-01 -8.33604336e-02 -1.50711298e+00
-6.06215715e-01 -3.35706711e-01 -2.62710750e-01 1.69480765e+00
9.42332387e-01 -9.46662188e-01 5.74591696e-01 8.78144979e-01
-1.20217450e-01 -7.82857895e-01 -8.63388956e-01 -4.83485460e-01
2.91934818e-01 -4.00165647e-01 6.47940874e-01 1.11779845e+00
1.77107900e-01 6.51661217e-01 -2.61184186e-01 3.91406506e-01
8.34257841e-01 2.08168849e-01 4.95292127e-01 -1.11237562e+00
-7.51815915e-01 -2.77133822e-01 -6.81824684e-02 -8.32217813e-01
8.44472766e-01 -1.40373635e+00 6.53253123e-03 -1.40128183e+00
4.75657612e-01 -1.07711887e+00 -1.88494578e-01 9.44182932e-01
-5.97908258e-01 -2.92571373e-02 -1.71908617e-01 -1.57860339e-01
-3.16429168e-01 1.62484124e-01 7.67057300e-01 -3.56777519e-01
4.12995458e-01 -9.13713947e-02 -9.03520167e-01 3.72014403e-01
3.96837711e-01 -5.81186771e-01 3.59826535e-02 -6.03584647e-01
7.34361291e-01 1.53866634e-01 4.31442052e-01 -4.29424554e-01
1.01874247e-01 -9.51146707e-03 4.07531381e-01 -4.45349514e-01
2.37889349e-01 -2.98236817e-01 3.71243387e-01 6.00108644e-03
-4.26985204e-01 3.61540169e-02 -1.05871998e-01 4.27353650e-01
8.36874098e-02 -2.65635818e-01 3.57340574e-01 -9.40374583e-02
-8.98305893e-01 5.57654381e-01 -8.68785009e-03 3.50734562e-01
4.71039712e-01 2.47339457e-01 -7.52741277e-01 1.48190558e-01
-5.04533648e-01 3.89697850e-01 5.20906091e-01 3.38664800e-01
1.20963961e-01 -1.18707871e+00 -8.22098255e-01 2.25090027e-01
7.84348994e-02 3.47525209e-01 4.15570498e-01 7.77407765e-01
-1.11278363e-01 1.96268111e-01 1.28031313e-01 -5.51303923e-01
-1.06998563e+00 4.87382263e-01 3.11505586e-01 -4.35213596e-01
-4.30994451e-01 1.14545774e+00 3.05583477e-01 -8.66323948e-01
3.58820632e-02 -1.38986692e-01 -6.78905845e-02 1.20942257e-01
4.95416552e-01 7.69221708e-02 5.28246723e-02 -5.02302825e-01
-5.50737977e-01 8.23367164e-02 -3.29595029e-01 1.79446414e-01
1.40106833e+00 2.67676681e-01 -4.41436559e-01 5.40382802e-01
1.32427132e+00 1.72748223e-01 -1.13059402e+00 -1.17889024e-01
-1.04761675e-01 -1.34088635e-01 -1.96014538e-01 -1.01657033e+00
-8.42199802e-01 4.98506516e-01 -8.97557884e-02 3.70153189e-02
5.48446953e-01 8.50666389e-02 8.54167104e-01 1.75315857e-01
5.42162418e-01 -8.24999809e-01 -7.99688220e-01 8.44864070e-01
2.57237822e-01 -8.40689778e-01 -2.26345479e-01 -6.01078570e-01
-5.64236641e-01 7.04233527e-01 5.05569637e-01 -2.51620654e-02
5.26484787e-01 6.94119155e-01 -2.11990669e-01 3.13172452e-02
-1.38539195e+00 -6.26548231e-02 2.93338597e-01 5.61623573e-01
7.86847353e-01 5.03614485e-01 -1.37379721e-01 8.73029232e-01
-3.68422598e-01 1.30740091e-01 3.86995375e-01 6.00378573e-01
1.88453406e-01 -1.49138129e+00 1.97724868e-02 9.04661536e-01
-7.08482981e-01 -5.90499580e-01 -8.23746696e-02 4.76434320e-01
3.48445773e-01 6.59681559e-01 2.47075900e-01 -4.60625380e-01
6.10248558e-02 3.09894681e-01 7.11197019e-01 -7.17388332e-01
-6.78042114e-01 -3.50735664e-01 7.49677062e-01 -2.99472779e-01
-1.93745330e-01 -9.29330468e-01 -1.24368787e+00 -5.95528305e-01
-9.18687582e-02 2.11180314e-01 4.23722684e-01 1.08094680e+00
6.48479402e-01 4.32941258e-01 3.88981432e-01 -4.53753710e-01
-4.80358303e-01 -1.05346763e+00 -3.95141393e-01 5.16401649e-01
2.57613957e-01 -1.00775170e+00 -2.41159573e-01 -7.97536150e-02] | [9.680726051330566, 8.752251625061035] |
597a11eb-4c9b-4717-989d-810b58d14931 | krylov-methods-are-nearly-optimal-for-low | 2304.03191 | null | https://arxiv.org/abs/2304.03191v1 | https://arxiv.org/pdf/2304.03191v1.pdf | Krylov Methods are (nearly) Optimal for Low-Rank Approximation | We consider the problem of rank-$1$ low-rank approximation (LRA) in the matrix-vector product model under various Schatten norms: $$ \min_{\|u\|_2=1} \|A (I - u u^\top)\|_{\mathcal{S}_p} , $$ where $\|M\|_{\mathcal{S}_p}$ denotes the $\ell_p$ norm of the singular values of $M$. Given $\varepsilon>0$, our goal is to output a unit vector $v$ such that $$ \|A(I - vv^\top)\|_{\mathcal{S}_p} \leq (1+\varepsilon) \min_{\|u\|_2=1}\|A(I - u u^\top)\|_{\mathcal{S}_p}. $$ Our main result shows that Krylov methods (nearly) achieve the information-theoretically optimal number of matrix-vector products for Spectral ($p=\infty$), Frobenius ($p=2$) and Nuclear ($p=1$) LRA. In particular, for Spectral LRA, we show that any algorithm requires $\Omega\left(\log(n)/\varepsilon^{1/2}\right)$ matrix-vector products, exactly matching the upper bound obtained by Krylov methods [MM15, BCW22]. Our lower bound addresses Open Question 1 in [Woo14], providing evidence for the lack of progress on algorithms for Spectral LRA and resolves Open Question 1.2 in [BCW22]. Next, we show that for any fixed constant $p$, i.e. $1\leq p =O(1)$, there is an upper bound of $O\left(\log(1/\varepsilon)/\varepsilon^{1/3}\right)$ matrix-vector products, implying that the complexity does not grow as a function of input size. This improves the $O\left(\log(n/\varepsilon)/\varepsilon^{1/3}\right)$ bound recently obtained in [BCW22], and matches their $\Omega\left(1/\varepsilon^{1/3}\right)$ lower bound, to a $\log(1/\varepsilon)$ factor. | ['Shyam Narayanan', 'Ainesh Bakshi'] | 2023-04-06 | null | null | null | null | ['open-question'] | ['natural-language-processing'] | [ 3.75476718e-01 2.16023773e-01 -2.38639899e-02 3.34970653e-01
-1.13083494e+00 -6.92893386e-01 -3.96410003e-02 5.47043458e-02
-7.17456758e-01 8.06115270e-01 -3.33341300e-01 -6.99065924e-01
-6.99534416e-01 -9.41476822e-01 -7.53409147e-01 -1.01517940e+00
-9.23307180e-01 2.59650469e-01 -1.47146285e-01 -5.07309973e-01
3.00174177e-01 3.16992283e-01 -1.62531686e+00 -3.41485441e-01
8.88280213e-01 1.63276482e+00 -2.10261509e-01 9.55796301e-01
-7.80560076e-02 4.10319090e-01 -1.47123381e-01 -2.52626032e-01
9.29782629e-01 -8.02901983e-01 -9.58914399e-01 -1.54458150e-01
5.36640525e-01 4.22274888e-01 -3.19556832e-01 1.87778819e+00
2.65004337e-01 3.57747257e-01 8.35929155e-01 -8.32801461e-01
-3.25342357e-01 5.30300021e-01 -1.07391310e+00 1.93111926e-01
2.16560140e-01 7.01075792e-02 1.35872722e+00 -9.61334944e-01
4.48449194e-01 7.90789604e-01 7.59964466e-01 1.00374594e-03
-1.46413279e+00 -1.16122842e+00 2.26744246e-02 -3.20804685e-01
-1.87010884e+00 -2.22006097e-01 5.08916855e-01 -4.96324003e-01
5.65902948e-01 7.84362555e-01 4.67283070e-01 -1.81638867e-01
-6.84262291e-02 1.99744761e-01 1.18649960e+00 -7.00703859e-01
1.29759824e-02 -6.57817647e-02 2.19258398e-01 1.23296428e+00
4.60260987e-01 -1.53819084e-01 -2.15049982e-01 -3.00819695e-01
1.03347588e+00 -3.24988723e-01 -6.03243649e-01 1.00324638e-01
-1.00103831e+00 1.02440596e+00 1.06876597e-01 5.98573446e-01
-4.04877663e-02 5.50186276e-01 4.32903171e-02 5.94567537e-01
2.21376449e-01 5.36389053e-01 -3.56771469e-01 3.95513792e-03
-1.08152127e+00 3.02377462e-01 6.83595538e-01 1.11362839e+00
1.05677176e+00 2.28324205e-01 1.75463289e-01 8.04929018e-01
8.71281698e-02 1.25166440e+00 -1.89966127e-01 -1.49900556e+00
5.78404546e-01 2.99516827e-01 1.55912131e-01 -9.21494722e-01
-4.12971795e-01 -3.30669940e-01 -1.28520751e+00 2.96475440e-01
8.75025988e-01 -1.20624818e-01 -3.23425621e-01 1.99515343e+00
-2.54128635e-01 -3.66317064e-01 -1.92638442e-01 6.81917548e-01
1.90261126e-01 9.25203204e-01 -5.28943062e-01 -1.17318010e+00
1.40984142e+00 -2.55773753e-01 -2.17250779e-01 -3.78002524e-02
7.48417974e-01 -1.07051158e+00 1.10174263e+00 7.55452514e-01
-1.72419834e+00 -2.96604753e-01 -9.90343928e-01 4.10173297e-01
5.75353354e-02 -1.73405111e-01 6.37013674e-01 8.07681918e-01
-1.14702344e+00 6.33726239e-01 -2.70488769e-01 3.29137564e-01
8.40479136e-02 5.93105197e-01 -5.18226266e-01 -1.99777365e-01
-8.84374499e-01 1.44940272e-01 -1.07960962e-01 1.36328235e-01
-3.43226850e-01 -6.24038219e-01 -6.70621276e-01 -1.25700459e-01
5.87590218e-01 -1.41609818e-01 5.83857536e-01 -3.96207958e-01
-6.96224928e-01 7.03657866e-01 -7.69648254e-02 -2.12282121e-01
2.26110294e-01 2.90910661e-01 -1.55794621e-01 4.03360665e-01
2.59382248e-01 -6.34675622e-02 6.04571283e-01 -1.12250161e+00
-4.85022604e-01 -7.45510757e-01 -3.50723043e-02 -7.02473894e-02
-4.27433848e-02 7.69955963e-02 -1.48721829e-01 -4.60936546e-01
8.40287209e-01 -1.05536699e+00 -5.29833496e-01 -4.74286705e-01
-1.01591095e-01 5.41247278e-02 8.56108963e-02 -4.17882383e-01
1.55343378e+00 -2.19731617e+00 1.99883848e-01 1.03673840e+00
5.57928383e-01 1.49508908e-01 2.42986307e-01 5.88499069e-01
-4.41252291e-01 4.86193717e-01 -5.16439199e-01 8.99673998e-02
-1.21038541e-01 -1.15198493e-01 -2.43010998e-01 7.31017470e-01
-7.29304910e-01 -5.36557287e-02 -7.33294487e-01 -1.85052305e-01
-3.19757789e-01 2.66862690e-01 -5.75202465e-01 -4.65730011e-01
-3.99900898e-02 1.62464812e-01 -2.67451018e-01 3.17645550e-01
1.12670970e+00 -3.75572801e-01 3.09072137e-01 -1.85941815e-01
-4.14960623e-01 -1.98228031e-01 -2.05595779e+00 1.38362038e+00
-2.84109056e-01 2.65851676e-01 1.02078891e+00 -1.30985117e+00
5.21286249e-01 2.19598889e-01 8.91085267e-01 -7.70279527e-01
3.43599141e-01 7.25533485e-01 -2.00058401e-01 1.79255530e-02
2.48477340e-01 -7.63447762e-01 -3.82028162e-01 5.79467833e-01
-1.30343020e-01 -2.33783409e-01 5.87432146e-01 4.71215129e-01
1.54531169e+00 -5.13723850e-01 -1.15176551e-02 -8.23094666e-01
7.16643572e-01 -2.02319115e-01 6.94002509e-01 1.02444053e+00
-1.87193230e-01 3.05988282e-01 1.09168649e+00 -1.17754877e-01
-9.96272206e-01 -9.42780077e-01 -4.77682233e-01 1.32461464e+00
2.11874217e-01 -6.56993747e-01 -8.14548254e-01 -1.02281630e-01
-4.07899231e-01 4.01954532e-01 -5.86953640e-01 3.27160835e-01
-7.17077971e-01 -1.13102460e+00 6.71707869e-01 4.36280549e-01
5.71489990e-01 -7.29550302e-01 -3.05713177e-01 2.20444910e-02
-2.34175976e-02 -6.56159282e-01 -7.14630127e-01 5.39644003e-01
-8.32874477e-01 -1.06969857e+00 -7.19550848e-01 -4.64853346e-01
9.51870322e-01 1.91284612e-01 6.05050266e-01 -2.98787653e-02
-4.59325552e-01 3.77462715e-01 -1.43869326e-01 -4.59953956e-02
7.05512147e-03 -5.78980327e-01 4.85837609e-01 -2.73726463e-01
-2.09586829e-01 -8.81225705e-01 -8.90604198e-01 2.48456717e-01
-9.84811783e-01 -3.30339134e-01 3.16208452e-01 1.02745748e+00
1.00389683e+00 3.88454914e-01 8.98623168e-02 -7.81510592e-01
4.53761399e-01 8.69900286e-02 -1.23964465e+00 -2.81965852e-01
-9.22806859e-01 2.95805633e-01 9.93916094e-01 -1.18260071e-01
-1.77895576e-01 -1.73948884e-01 -1.58244371e-01 -4.22135800e-01
5.75680435e-01 6.04572356e-01 1.88989639e-01 -3.19164276e-01
9.71257746e-01 4.12459105e-01 -1.50141105e-01 -4.81448948e-01
4.99873281e-01 1.70751408e-01 6.30410373e-01 -7.98135400e-01
8.35240602e-01 5.54268956e-01 6.78869605e-01 -8.96583438e-01
-6.66712165e-01 -3.43107998e-01 -1.21496685e-01 1.72825679e-01
5.15997767e-01 -8.18365514e-01 -1.45411921e+00 -3.09210300e-01
-4.94420737e-01 -1.35709256e-01 -4.95046437e-01 6.10074997e-01
-9.02184904e-01 5.15452981e-01 -5.86659670e-01 -1.40965438e+00
-4.26177353e-01 -1.13419223e+00 4.33203459e-01 -2.03066617e-01
-6.72692657e-02 -5.20102024e-01 -1.85326099e-01 5.37107527e-01
3.20597529e-01 1.48737863e-01 1.24729967e+00 -2.04657629e-01
-5.44604182e-01 -4.05879527e-01 -5.95877349e-01 6.92268372e-01
-5.33632636e-01 -5.87680101e-01 -3.46595019e-01 -3.62767130e-01
2.97949076e-01 2.12253541e-01 8.40212166e-01 2.62391865e-01
9.80694294e-01 -1.04143929e+00 -1.29953802e-01 3.97283375e-01
1.69588733e+00 5.60177788e-02 6.54447436e-01 -3.49879026e-01
3.87221456e-01 3.67015749e-02 2.38569528e-01 6.56929672e-01
-2.43362874e-01 3.58859628e-01 2.46383190e-01 2.61709273e-01
1.98968947e-01 2.82995045e-01 3.53822172e-01 1.01048064e+00
-3.96543026e-01 3.73199791e-01 -6.76100373e-01 7.57636786e-01
-1.47699714e+00 -8.93495083e-01 -4.06172514e-01 2.83867979e+00
8.81942570e-01 3.17657620e-01 2.81097323e-01 4.79665637e-01
4.14432436e-01 1.89183116e-01 9.40681249e-02 -6.27375185e-01
-2.01961681e-01 1.36249113e+00 1.14243913e+00 7.21794546e-01
-6.88149631e-01 5.59244037e-01 4.10207176e+00 1.44600773e+00
-6.59158051e-01 1.55510485e-01 3.63039255e-01 -5.11045814e-01
-4.78554070e-01 2.97542810e-01 -4.84226525e-01 5.75191081e-01
8.95941675e-01 -6.31766617e-02 6.62084341e-01 8.08344960e-01
-3.88002880e-02 -4.58121330e-01 -8.83668900e-01 1.23629165e+00
4.44058180e-02 -1.26250052e+00 -5.04284561e-01 6.33228898e-01
7.43389964e-01 -2.72526920e-01 2.39840165e-01 6.24678060e-02
4.35170859e-01 -1.21358848e+00 3.64914000e-01 -1.20182641e-01
1.35314870e+00 -1.05758047e+00 4.41631615e-01 3.34785342e-01
-1.75687671e+00 -2.83962905e-01 -2.57409036e-01 -1.69138923e-01
5.59291393e-02 8.07093740e-01 -2.47754604e-01 5.73605478e-01
8.16344619e-01 -2.44564995e-01 4.29388762e-01 5.85557938e-01
2.60848284e-01 3.95029783e-01 -7.07321227e-01 2.04361200e-01
3.70030314e-01 -8.23534369e-01 8.06347907e-01 9.81211066e-01
5.45711637e-01 9.82205868e-01 3.69529486e-01 5.29498935e-01
-2.92113781e-01 5.74661016e-01 -5.75352371e-01 -5.34611456e-02
3.14527184e-01 7.32872486e-01 -8.09446037e-01 -1.28931910e-01
-1.61064923e-01 5.35662055e-01 -1.10513896e-01 2.62279153e-01
-4.46442366e-01 -1.01872027e+00 8.50700498e-01 7.73759842e-01
5.14388800e-01 -5.85677147e-01 -5.76400995e-01 -8.86060715e-01
2.04490066e-01 -7.37074137e-01 5.55920243e-01 -1.61777645e-01
-8.60562325e-01 3.49752516e-01 1.13587871e-01 -1.01407647e+00
6.33108839e-02 -6.04271293e-01 -4.15203236e-02 1.04547906e+00
-4.88449395e-01 -2.96144664e-01 4.40367758e-01 5.29531896e-01
-2.57010102e-01 -1.13398619e-02 8.59779596e-01 3.03549528e-01
-2.59024382e-01 7.57201195e-01 4.70620066e-01 1.47984818e-01
1.67103991e-01 -7.63116300e-01 -5.00109196e-01 9.61610973e-01
-1.53549211e-02 7.53633380e-01 9.55837965e-01 -4.08262491e-01
-1.76378965e+00 -4.49271202e-01 7.81939089e-01 -4.51880768e-02
6.83490157e-01 -1.94700703e-01 -6.35258496e-01 6.00720167e-01
-9.54747498e-02 1.37559250e-01 8.45064640e-01 -6.37349486e-03
-5.62172294e-01 -2.24256799e-01 -1.34708905e+00 6.68557227e-01
1.18772435e+00 -5.47049165e-01 9.13939327e-02 3.01577449e-01
2.31370121e-01 -1.88865542e-01 -1.32828057e+00 6.60346687e-01
3.56541514e-01 -1.13394284e+00 9.43747520e-01 -9.98986959e-02
-1.12687498e-01 -4.22971308e-01 -7.42777944e-01 -4.77344096e-01
-2.95864403e-01 -1.00773108e+00 1.07187882e-01 4.01844412e-01
6.63710177e-01 -6.39834285e-01 6.03345931e-01 5.15461385e-01
-1.27665401e-01 -8.57262969e-01 -1.48848963e+00 -7.45128989e-01
6.06188536e-01 -6.60941005e-01 -2.04048499e-01 7.78348744e-01
4.86096293e-01 1.13971112e-02 -2.73564517e-01 -1.09993480e-02
6.65747404e-01 -2.91253813e-02 4.00311619e-01 -1.17913365e+00
-7.60875106e-01 -7.72889256e-01 -2.08052427e-01 -9.59045947e-01
-4.30324674e-01 -1.00711012e+00 -1.60361841e-01 -9.85995054e-01
3.38054061e-01 -9.42538917e-01 -2.61558115e-01 4.14810300e-01
3.91384363e-01 4.88344252e-01 2.85995275e-01 9.10554230e-02
-4.41394746e-01 9.68012512e-02 1.15123868e+00 -3.97910997e-02
-1.92117929e-01 -3.39041263e-01 -8.97145987e-01 7.78598189e-01
2.40818232e-01 -2.74810731e-01 -1.93156213e-01 -5.14036790e-02
9.81136620e-01 8.46279502e-01 -5.47863916e-02 -1.11011815e+00
9.25190449e-02 -1.71243295e-01 -4.32166010e-02 -3.08286428e-01
5.98688841e-01 -2.77185559e-01 2.36632943e-01 2.79264927e-01
-8.20013881e-03 1.04106769e-01 5.38188219e-02 5.58970094e-01
1.51608780e-01 -3.26295793e-01 1.06575513e+00 -3.37950587e-01
2.00394932e-02 2.64888644e-01 -2.67925322e-01 2.06622064e-01
7.35716105e-01 -1.91902652e-01 -2.09007695e-01 -5.22585988e-01
-7.65277743e-01 9.98886675e-02 3.47172827e-01 -6.55896187e-01
1.34716004e-01 -1.24282861e+00 -6.69218183e-01 3.20317268e-01
-1.00161858e-01 3.47121745e-01 6.21649086e-01 1.28439665e+00
-6.33749485e-01 2.99467325e-01 2.65458792e-01 -5.13209701e-01
-8.69327903e-01 4.54066992e-01 7.26362616e-02 -4.37170744e-01
-2.40079284e-01 1.46541595e+00 2.01347973e-02 1.72530174e-01
-1.70504637e-02 -8.00718367e-02 5.99902868e-01 -6.06934130e-02
5.19763768e-01 7.15800762e-01 -3.01751494e-03 -7.55818009e-01
-2.55343467e-01 6.89477324e-01 -1.58109263e-01 -6.56991065e-01
9.80997145e-01 4.67572995e-02 -6.79532170e-01 1.64624572e-01
1.55402470e+00 4.89737958e-01 -7.74845481e-01 -1.45158619e-01
-4.98415947e-01 -3.49615335e-01 -3.57585549e-01 -2.18773186e-01
-9.63305473e-01 8.55708718e-01 7.26263940e-01 4.51301068e-01
1.13832235e+00 1.52353793e-01 6.51475608e-01 2.70747215e-01
8.66297483e-01 -1.63649225e+00 -6.76028617e-03 6.86114430e-01
6.18730187e-01 -6.74894214e-01 2.39045233e-01 -5.40524483e-01
-1.81854606e-01 8.07054043e-01 7.71742687e-02 -3.01310658e-01
1.01558971e+00 -3.47991474e-02 -4.16518509e-01 -2.81585485e-01
-1.45379573e-01 -2.42275760e-01 9.50350761e-02 -2.58093864e-01
7.53491879e-01 1.54734582e-01 -9.29847419e-01 8.10115695e-01
-5.45803964e-01 -3.42402816e-01 5.69858849e-01 9.20298576e-01
-8.09107661e-01 -1.42325687e+00 -5.98127127e-01 7.21183300e-01
-6.00795269e-01 -3.17254305e-01 2.81924069e-01 5.49849212e-01
3.07413399e-01 8.54555011e-01 -2.43260145e-01 -2.39439249e-01
1.43234938e-01 4.26982701e-01 8.62294614e-01 -3.86033952e-01
-3.01972389e-01 5.34453392e-01 8.64855479e-03 -5.36892295e-01
1.78175241e-01 -6.29985452e-01 -1.79456806e+00 -6.23152316e-01
-1.41522795e-01 6.27448678e-01 5.46782136e-01 7.32616782e-01
1.43748239e-01 -1.04398526e-01 6.76288366e-01 -4.18981969e-01
-5.15634656e-01 -8.77908766e-01 -1.24835134e+00 2.81914651e-01
-7.46124163e-02 -6.00386202e-01 -7.43167043e-01 -1.55422941e-01] | [6.542399883270264, 4.704171180725098] |
5ddf9838-2398-4cd9-bc22-31653d4870fd | infrared-safety-of-a-neural-net-top-tagging | 1806.01263 | null | http://arxiv.org/abs/1806.01263v2 | http://arxiv.org/pdf/1806.01263v2.pdf | Infrared Safety of a Neural-Net Top Tagging Algorithm | Neural network-based algorithms provide a promising approach to jet
classification problems, such as boosted top jet tagging. To date, NN-based top
taggers demonstrated excellent performance in Monte Carlo studies. In this
paper, we construct a top-jet tagger based on a Convolutional Neural Network
(CNN), and apply it to parton-level boosted top samples, with and without an
additional gluon in the final state. We show that the jet observable defined by
the CNN obeys the canonical definition of infrared safety: it is unaffected by
the presence of the extra gluon, as long as it is soft or collinear with one of
the quarks. Our results indicate that the CNN tagger is robust with respect to
possible mis-modeling of soft and collinear final-state radiation by Monte
Carlo generators. | ['Suyong Choi', 'Maxim Perelstein', 'Seung J. Lee'] | 2018-06-04 | null | null | null | null | ['jet-tagging'] | ['graphs'] | [-3.62579584e-01 -7.46604130e-02 -3.36035609e-01 -3.39010715e-01
-5.95142484e-01 -7.94684947e-01 1.01232278e+00 2.16236711e-02
-3.28181773e-01 6.21864021e-01 2.77191490e-01 -4.99875724e-01
5.64624043e-03 -1.06704080e+00 -8.50890458e-01 -9.24230695e-01
-1.86479747e-01 1.02840030e+00 5.48877358e-01 -3.99880469e-01
-1.88663974e-01 5.49606860e-01 -9.42100406e-01 4.45374936e-01
2.08127841e-01 1.26944327e+00 -6.13312721e-01 6.95075989e-01
1.77052125e-01 7.39530325e-01 -3.47890675e-01 -7.96193004e-01
7.00354636e-01 -6.79551661e-01 -6.00589395e-01 -1.07327139e+00
3.95534873e-01 3.32631469e-02 -8.84174168e-01 1.35545969e+00
5.14604151e-01 7.94757083e-02 5.09643257e-01 -1.10036993e+00
-5.92272580e-01 9.10893321e-01 4.45890516e-01 1.60087049e-01
-4.84556705e-01 -1.99391261e-01 1.23518312e+00 -6.66688263e-01
8.15877020e-01 1.27310491e+00 8.77031386e-01 2.16445968e-01
-1.27423966e+00 -6.61006093e-01 -7.30014801e-01 -2.40968093e-01
-8.93659651e-01 -1.61555782e-01 6.26217842e-01 -6.21157110e-01
9.49849844e-01 1.26822725e-01 7.04044163e-01 1.07823443e+00
1.31905687e+00 5.93666658e-02 8.65391612e-01 -7.55887210e-01
4.54644501e-01 -5.07444851e-02 7.64981732e-02 7.11066067e-01
3.66213351e-01 8.45971406e-01 -6.21590018e-01 -5.93276441e-01
8.01097691e-01 -5.92217445e-01 4.90248978e-01 -7.85630643e-01
-1.30993462e+00 1.18401861e+00 6.70934916e-01 2.29947507e-01
2.70443708e-01 6.13018334e-01 8.41030419e-01 2.11097628e-01
9.24942315e-01 7.35704303e-01 -9.26496863e-01 1.36749342e-01
-7.21950054e-01 4.31857288e-01 7.21638918e-01 7.57497072e-01
3.13877255e-01 1.77891374e-01 -4.19500142e-01 4.20704514e-01
1.76951334e-01 4.05888081e-01 1.56193867e-01 -7.95340419e-01
2.12486818e-01 1.12064190e-01 2.38225400e-01 -2.40176246e-01
-7.86217511e-01 -8.04163516e-01 -7.78962135e-01 5.57861626e-01
3.92054439e-01 -3.41329575e-01 -1.10782206e+00 1.52543354e+00
3.72660607e-01 -4.93014514e-01 -1.99717730e-02 9.23094153e-01
7.86606371e-01 4.37861323e-01 1.41587943e-01 1.30962446e-01
1.27246869e+00 -9.33691025e-01 -2.62212127e-01 6.25124900e-03
6.64837599e-01 -8.58824551e-01 -2.47573480e-01 3.27087760e-01
-9.71307933e-01 -6.07004225e-01 -9.80615437e-01 -6.92575648e-02
-3.77743334e-01 1.01847827e-01 1.47187042e+00 7.34886467e-01
-8.14345121e-01 1.00112009e+00 -8.26888740e-01 -1.88166723e-01
-3.80725712e-01 3.95087332e-01 -2.79958218e-01 4.90721375e-01
-1.50893807e+00 8.17904770e-01 8.15295577e-01 1.36819988e-01
-7.57958770e-01 -5.54565012e-01 -4.84013677e-01 8.80383849e-02
-6.49814382e-02 -7.06454515e-01 1.89639425e+00 -9.40619111e-01
-1.29016805e+00 5.68960607e-01 4.04800355e-01 -4.79173660e-01
4.64809120e-01 2.64811479e-02 -8.32544506e-01 -3.76527682e-02
5.66611029e-02 2.81390786e-01 4.03977633e-01 -6.68027103e-01
-3.44053686e-01 2.03863513e-02 7.21352547e-02 -3.56161833e-01
3.74679565e-01 8.32402110e-01 -7.48060048e-02 -7.22759843e-01
5.51710904e-01 -1.24159265e+00 -2.40997180e-01 -3.20394933e-01
-5.63799858e-01 9.83449817e-03 7.88502619e-02 -3.82466227e-01
3.31236184e-01 -1.72765481e+00 -1.08477890e-01 3.50804150e-01
2.55785018e-01 2.74529099e-01 8.96413624e-02 6.28318250e-01
-3.95956546e-01 2.20667124e-01 3.99417222e-01 1.79574639e-01
4.58924860e-01 -6.24442026e-02 -1.52567802e-02 3.59407246e-01
-5.20697981e-02 1.01474094e+00 -7.08899081e-01 1.13165695e-02
1.94565840e-02 2.37196669e-01 -5.82761705e-01 -8.56650770e-02
-4.16954130e-01 4.34393585e-01 -4.11065578e-01 4.17352974e-01
8.45313072e-01 3.21736664e-01 1.64662108e-01 -1.99455678e-01
-5.83213866e-01 7.90807724e-01 -7.83643305e-01 1.10530090e+00
-3.91939431e-02 4.22175735e-01 1.55377895e-01 -6.48132026e-01
8.87084663e-01 4.56755102e-01 4.26185012e-01 -1.18991220e+00
1.63637400e-01 3.82169068e-01 5.75850964e-01 1.35037884e-01
7.55337536e-01 -8.12637985e-01 -5.21189392e-01 3.07947665e-01
6.11719608e-01 3.33587021e-01 -6.60363818e-03 4.33248311e-01
1.19422495e+00 1.86011344e-02 4.52414788e-02 -4.29090202e-01
-2.14224502e-01 1.80118397e-01 8.53464842e-01 1.17787397e+00
1.72771271e-02 3.54267895e-01 8.27990890e-01 -7.44616807e-01
-1.49723327e+00 -1.14512742e+00 -4.66874748e-01 1.39050555e+00
-1.76364765e-01 -5.71543932e-01 -2.43909940e-01 -6.10861182e-01
1.28925338e-01 4.29882944e-01 -4.79534179e-01 -4.27964270e-01
-5.56412876e-01 -1.12535203e+00 6.36760950e-01 7.43815482e-01
2.15622544e-01 -9.26162839e-01 2.89858103e-01 4.78877157e-01
4.10535872e-01 -8.47292542e-01 6.67764768e-02 1.12410593e+00
-6.00864172e-01 -8.05973887e-01 -1.23743922e-01 -4.65234548e-01
2.49175519e-01 -2.96007097e-01 1.18196023e+00 -1.91000268e-01
-1.01976857e-01 -3.94458115e-01 -4.48080063e-01 -2.61716485e-01
-1.07644212e+00 -2.22695529e-01 3.64642084e-01 -4.47356611e-01
9.52893719e-02 3.70600894e-02 -1.30240142e-01 1.05762795e-01
-3.47763330e-01 2.05266818e-01 4.89577234e-01 7.97340691e-01
1.68778986e-01 1.50258452e-01 6.59048557e-02 -5.09628415e-01
-1.07679009e-01 -2.89135844e-01 -1.18701005e+00 -3.47406603e-02
-1.06970571e-01 4.45585042e-01 7.47259259e-01 7.57312123e-03
-1.05129933e+00 -9.34148729e-02 -2.63021976e-01 3.14185359e-02
5.64449988e-02 1.93958908e-01 -2.31408834e-01 -8.35795939e-01
4.48097646e-01 -4.37989950e-01 -8.64230931e-01 -9.34180081e-01
3.10533017e-01 8.84112343e-02 5.78032792e-01 -7.91954994e-01
9.39239860e-01 1.45193622e-01 6.35651648e-01 -7.67889842e-02
-1.05954790e+00 -3.28085005e-01 -1.05404735e+00 -8.56129453e-02
1.06626666e+00 -7.55148709e-01 -5.66136122e-01 4.95058745e-01
-1.33007789e+00 -5.56742102e-02 -5.14759645e-02 1.01384616e+00
-3.89100790e-01 1.06366025e-02 -1.02642894e+00 -6.27828717e-01
-3.50029826e-01 -1.04401433e+00 8.63021731e-01 -1.04195096e-01
4.52567577e-01 -1.03559887e+00 3.22381258e-01 -1.18981682e-01
3.70138794e-01 2.15774029e-01 1.44048917e+00 -1.16389370e+00
-6.25837922e-01 -4.23950106e-01 -3.32295835e-01 3.69422346e-01
-5.63498914e-01 -1.41498461e-01 -7.37934589e-01 -2.96487182e-01
-2.20725730e-01 -1.67930365e-01 1.32549906e+00 3.13070804e-01
4.98798996e-01 -1.02225907e-01 -2.97434360e-01 6.21339619e-01
1.32721353e+00 1.35666996e-01 -4.47734632e-02 1.97638497e-01
5.50047219e-01 -2.79448126e-02 5.14926136e-01 1.26876920e-01
-3.38506490e-01 1.05337119e+00 4.55670446e-01 -4.66372855e-02
-1.44726709e-01 -2.05133200e-01 5.04985750e-01 9.64668989e-01
-1.96371883e-01 -3.98649126e-01 -8.03744793e-01 -8.76310840e-03
-1.71386421e+00 -7.23371744e-01 -8.68853927e-01 2.19902277e+00
5.53715348e-01 7.35922158e-01 -1.92720056e-01 -3.40497375e-01
6.27339840e-01 -1.18273281e-01 7.90605471e-02 -9.14832473e-01
9.32066590e-02 9.33086693e-01 9.99351084e-01 4.86260980e-01
-1.41426194e+00 8.12784731e-01 7.26035070e+00 7.05561161e-01
-7.24920928e-01 6.35113001e-01 1.37808457e-01 -8.57105926e-02
-2.64267415e-01 4.35515076e-01 -9.75783110e-01 3.34710062e-01
1.01582193e+00 2.40972027e-01 2.05652773e-01 6.98262036e-01
-1.33373216e-01 -7.20348628e-03 -1.15574014e+00 1.28036499e-01
-2.41958067e-01 -1.52166176e+00 -3.46499115e-01 2.51679197e-02
8.42268884e-01 8.39036405e-01 -5.60849011e-01 6.58120275e-01
7.94715643e-01 -5.53067505e-01 1.31576872e+00 4.13421065e-01
5.34874499e-01 -9.98079360e-01 1.12674987e+00 1.53768063e-01
-1.09522653e+00 2.47568399e-01 -7.11261988e-01 -2.06007943e-01
6.64350539e-02 9.44405198e-01 -7.12591350e-01 7.61891544e-01
4.41388071e-01 -1.76860079e-01 -6.05698168e-01 1.26910901e+00
-1.88668966e-01 1.02263427e+00 -2.79487550e-01 4.71227169e-02
4.12532926e-01 -2.44802207e-01 6.42243743e-01 1.18732429e+00
4.91081983e-01 -3.27235490e-01 5.42634010e-01 8.34330797e-01
-2.10943833e-01 -2.14989841e-01 -5.20415068e-01 -3.12369466e-01
-9.10777897e-02 1.42876530e+00 -1.05637562e+00 -2.70957589e-01
-3.61137062e-01 5.37951827e-01 1.12685367e-01 3.00697178e-01
-5.46508789e-01 -3.18131238e-01 6.89572275e-01 -3.08786422e-01
3.72226924e-01 -4.32261556e-01 -1.75253432e-02 -1.09177005e+00
-2.56893098e-01 -1.07233949e-01 2.44788662e-01 -8.58853221e-01
-1.15219390e+00 1.89944476e-01 -1.91739768e-01 -8.17317963e-01
-1.63122073e-01 -1.32456958e+00 -6.90675259e-01 8.38321269e-01
-9.86927450e-01 -1.33845210e+00 4.65596795e-01 -1.40349910e-01
-2.83796519e-01 -3.95177901e-01 8.45785737e-01 2.25209996e-01
-2.47292995e-01 2.68996775e-01 6.62521183e-01 5.41561723e-01
7.67537892e-01 -1.40619528e+00 9.41244423e-01 7.91336358e-01
2.40416273e-01 3.53794873e-01 8.75650465e-01 -1.16963434e+00
-1.48440576e+00 -1.28114879e+00 1.30550885e+00 -3.44048321e-01
1.05784714e+00 -1.12180531e+00 -3.81052166e-01 8.74489784e-01
1.01390824e-01 6.44020259e-01 2.95572996e-01 4.36672091e-01
-4.45187151e-01 1.13602608e-01 -8.53682995e-01 -1.24303229e-01
8.52021992e-01 -5.55529296e-01 -3.84406269e-01 8.62712026e-01
8.06758285e-01 -4.21993315e-01 -8.07385147e-01 5.10994196e-01
3.42247277e-01 -9.21578169e-01 9.05861795e-01 -1.25366330e+00
-6.40955241e-03 -2.17559487e-02 -5.99859767e-02 -1.41978192e+00
-9.86310005e-01 -5.18118620e-01 2.65807241e-01 9.86372530e-01
1.26503631e-01 -4.38440412e-01 5.47971845e-01 4.68008891e-02
-6.33978426e-01 -2.07994394e-02 -1.47330761e+00 -1.08979619e+00
7.07091451e-01 -6.40147269e-01 2.66812921e-01 5.93517065e-01
-1.39644176e-01 -4.09678482e-02 -5.26145279e-01 3.96699011e-01
3.34257513e-01 4.63059321e-02 1.02514058e-01 -1.80991173e+00
-6.02615356e-01 -2.01920226e-01 -6.31321371e-01 -1.66226760e-01
1.06706373e-01 -1.65354598e+00 1.98725387e-01 -1.14481390e+00
3.49315673e-01 -4.90538061e-01 -5.85877657e-01 4.19803977e-01
7.73604035e-01 6.92976356e-01 6.64633838e-03 -3.34715545e-01
-6.74768925e-01 3.11569452e-01 8.16899538e-01 6.14069737e-02
6.86896682e-01 1.42572373e-01 2.84450185e-02 1.10031664e+00
7.58383155e-01 -9.85245943e-01 6.43827140e-01 -5.35142422e-03
1.04705036e+00 3.27232063e-01 8.56905639e-01 -1.12198651e+00
-7.58299232e-03 -6.84347656e-03 8.05077553e-01 -4.29329574e-01
4.20370936e-01 -2.39383116e-01 7.54194781e-02 3.49248379e-01
-3.41735214e-01 -9.31192711e-02 1.22838989e-01 1.14952408e-01
-4.80970554e-02 -8.37931693e-01 9.19832468e-01 -3.84180427e-01
-2.65878975e-01 1.27318546e-01 -5.76690197e-01 5.77710830e-02
3.90318722e-01 7.86658347e-01 -5.51327050e-01 1.65193051e-01
-6.25372529e-01 -3.31282914e-01 3.73383611e-01 3.47252995e-01
-3.16190213e-01 -1.84791875e+00 -3.64417255e-01 3.23618054e-01
1.44461825e-01 -9.57978666e-01 -2.43033305e-01 5.79021037e-01
-5.32107830e-01 9.98554468e-01 -2.13634402e-01 -9.59187970e-02
-9.55059528e-01 4.98339206e-01 6.36445343e-01 -1.81006178e-01
-6.35730088e-01 9.41339850e-01 -6.06393791e-04 -8.12145174e-01
-1.81116283e-01 -5.41966736e-01 6.46502018e-01 -1.79592110e-02
1.24846391e-01 6.17362410e-02 7.95525312e-01 -6.67165756e-01
-1.71081156e-01 -3.75073627e-02 2.27904484e-01 1.14306035e-02
8.32120299e-01 8.07711780e-01 -3.94678831e-01 4.49893296e-01
1.11858189e+00 3.41616660e-01 -6.32652164e-01 -1.13744885e-01
3.13124321e-02 -7.59070665e-02 4.69406605e-01 -1.21002913e+00
-7.50080526e-01 7.50108361e-01 4.50686842e-01 4.55646873e-01
6.15862757e-02 4.48356301e-01 7.00265050e-01 6.57880545e-01
4.33463275e-01 -9.80941892e-01 -2.96178430e-01 9.77368891e-01
5.83567858e-01 -1.00208473e+00 -2.85452306e-01 3.15449797e-02
3.72698009e-02 1.35527563e+00 4.22202200e-01 -1.09794319e-01
3.85503501e-01 3.92586023e-01 -1.92539096e-01 -3.52444768e-01
-9.27821696e-01 -1.63911134e-01 7.40401387e-01 8.39683861e-02
4.38820362e-01 4.34574842e-01 -2.04023883e-01 6.01888835e-01
-5.07182837e-01 -3.12425464e-01 4.74113226e-01 6.74381673e-01
-7.37435043e-01 -1.40083086e+00 -4.93342876e-01 5.77632308e-01
-5.66642165e-01 -3.22247088e-01 -4.89767373e-01 8.74768972e-01
6.47312760e-01 5.02546668e-01 1.13336883e-01 -2.11909547e-01
-1.89704597e-01 6.23173296e-01 7.50484049e-01 -8.22037458e-01
-7.91884363e-01 1.46379083e-01 4.91437227e-01 -4.03473049e-01
-2.13156790e-02 -5.85911214e-01 -1.13622034e+00 -1.77641302e-01
-5.15471399e-01 6.26029849e-01 9.59684074e-01 1.14421666e+00
-1.74991474e-01 5.82613885e-01 3.05206478e-01 -1.06493771e+00
-5.76843083e-01 -8.78569007e-01 -8.12141716e-01 -1.76244929e-01
1.35910779e-01 -8.48406434e-01 -3.71824741e-01 -8.79662693e-01] | [15.698616981506348, 2.920226573944092] |
83afc892-1c6b-407f-9522-ebf15da4476e | event-centric-query-expansion-in-web-search | 2305.19019 | null | https://arxiv.org/abs/2305.19019v1 | https://arxiv.org/pdf/2305.19019v1.pdf | Event-Centric Query Expansion in Web Search | In search engines, query expansion (QE) is a crucial technique to improve search experience. Previous studies often rely on long-term search log mining, which leads to slow updates and is sub-optimal for time-sensitive news searches. In this work, we present Event-Centric Query Expansion (EQE), a novel QE system that addresses these issues by mining the best expansion from a significant amount of potential events rapidly and accurately. This system consists of four stages, i.e., event collection, event reformulation, semantic retrieval and online ranking. Specifically, we first collect and filter news headlines from websites. Then we propose a generation model that incorporates contrastive learning and prompt-tuning techniques to reformulate these headlines to concise candidates. Additionally, we fine-tune a dual-tower semantic model to function as an encoder for event retrieval and explore a two-stage contrastive training approach to enhance the accuracy of event retrieval. Finally, we rank the retrieved events and select the optimal one as QE, which is then used to improve the retrieval of event-related documents. Through offline analysis and online A/B testing, we observe that the EQE system significantly improves many metrics compared to the baseline. The system has been deployed in Tencent QQ Browser Search and served hundreds of millions of users. The dataset and baseline codes are available at https://open-event-hub.github.io/eqe . | ['Tianhua Zhou', 'Xiang Chen', 'Jin Ma', 'Zhe Zhang', 'Xiaoling Bai', 'Yangfan Zhang', 'Weijie Cui', 'Yanan Zhang'] | 2023-05-30 | null | null | null | null | ['semantic-retrieval'] | ['natural-language-processing'] | [-2.15285629e-01 -5.39234579e-01 -5.68642914e-01 -2.04892457e-01
-1.36129642e+00 -6.49463236e-01 6.89192653e-01 3.71301591e-01
-6.56313598e-01 4.96383041e-01 6.21775329e-01 -2.42788792e-01
-3.89803797e-01 -9.09509659e-01 -7.22030222e-01 -7.43629993e-04
-1.87489226e-01 5.55330217e-01 6.68417513e-01 -5.33661008e-01
4.17981535e-01 8.77103433e-02 -1.48406172e+00 4.78840917e-01
9.24845338e-01 1.30464959e+00 3.70569676e-01 6.03464365e-01
-1.62073568e-01 6.28670573e-01 -5.30038178e-01 -2.08609134e-01
1.24348060e-03 -2.15408996e-01 -9.47615981e-01 -5.26490510e-01
-1.13188706e-01 -4.35245425e-01 -6.61269188e-01 5.65135181e-01
8.52982759e-01 3.46319258e-01 2.48160720e-01 -9.37014520e-01
-5.05960226e-01 6.83510303e-01 -4.29326475e-01 8.71227622e-01
7.28913605e-01 -2.93762445e-01 1.38059509e+00 -9.88333941e-01
6.35878503e-01 9.69430745e-01 2.39903241e-01 -1.18093668e-02
-7.83488512e-01 -8.50358129e-01 1.22619346e-01 4.94290829e-01
-1.54358745e+00 -3.64889473e-01 6.06810153e-01 -6.31991178e-02
1.23275340e+00 4.34665143e-01 7.16926575e-01 1.09313083e+00
-4.74942401e-02 1.12673879e+00 4.09048796e-01 -3.15741628e-01
1.39612556e-01 1.46609649e-01 2.00405702e-01 2.83978671e-01
-2.00862810e-01 2.14304224e-01 -8.98194611e-01 -4.26439345e-01
5.35074890e-01 3.31424505e-01 -2.48803988e-01 3.83229613e-01
-9.26174521e-01 9.12469208e-01 5.47706485e-01 1.84530258e-01
-7.77588069e-01 -1.40337437e-01 6.93211496e-01 3.45841527e-01
4.74926680e-01 6.02653325e-01 -5.65130472e-01 -2.44308397e-01
-1.03598976e+00 3.52193326e-01 9.63666558e-01 8.68901253e-01
4.96499658e-01 -6.29475296e-01 -7.36784637e-01 1.00338745e+00
1.56104833e-01 3.97668272e-01 6.18937135e-01 -6.42796934e-01
5.12081802e-01 5.56821108e-01 1.92067772e-01 -9.50173974e-01
-1.20890975e-01 -7.06379652e-01 -4.38074261e-01 -8.08772504e-01
-2.98533082e-01 8.99065062e-02 -6.71382129e-01 1.57117927e+00
2.26230085e-01 2.66527474e-01 -3.73540044e-01 8.68463993e-01
7.40438104e-01 1.04616427e+00 4.00981843e-01 -3.14120233e-01
1.57687497e+00 -1.12310040e+00 -6.44539297e-01 -1.15765475e-01
3.51891994e-01 -9.49689627e-01 1.46348989e+00 3.41352761e-01
-1.10911584e+00 -4.46789205e-01 -7.27556884e-01 -1.66529194e-01
-4.02705908e-01 1.30002886e-01 5.84389985e-01 2.98577137e-02
-7.65090346e-01 2.26002768e-01 -7.69662201e-01 -3.73853624e-01
1.76684141e-01 5.21655148e-03 2.85225600e-01 1.25873089e-01
-1.89992487e+00 4.12762105e-01 5.36890745e-01 -5.77700973e-01
-7.26562679e-01 -8.65724206e-01 -3.53777528e-01 5.03631711e-01
5.95709264e-01 -7.49103129e-01 1.65658462e+00 -2.61338294e-01
-1.11331463e+00 4.87951279e-01 -4.13438231e-01 -5.18302917e-01
1.58857018e-01 -6.79081500e-01 -7.27704883e-01 3.95512819e-01
3.12135905e-01 3.50702107e-01 5.33245742e-01 -6.23125911e-01
-1.00734830e+00 -8.87045562e-02 9.05259550e-02 4.05602723e-01
-6.03810728e-01 4.08366442e-01 -1.22175777e+00 -8.13606322e-01
-2.20356703e-01 -5.34179091e-01 -6.57670870e-02 -6.10786080e-01
-2.43800342e-01 -7.56331325e-01 4.11956936e-01 -8.95716250e-01
2.28764296e+00 -2.11167145e+00 -3.61851126e-01 4.83926922e-01
5.80179654e-02 3.95846367e-02 -3.49825583e-02 7.86441088e-01
1.28858313e-01 1.53925046e-01 5.90121210e-01 -5.31336628e-02
1.91647142e-01 -2.56281912e-01 -8.05538058e-01 -2.93128997e-01
-1.53883457e-01 1.14555430e+00 -9.87923384e-01 -7.02918470e-01
-1.28529727e-01 2.70558447e-01 -5.69325566e-01 2.35947043e-01
-3.78730893e-01 1.69759601e-01 -8.24424982e-01 7.25245476e-01
9.39334184e-02 -9.37553346e-01 -2.77860969e-01 -2.26751149e-01
-9.19201672e-02 9.40929949e-01 -9.02850151e-01 1.72489452e+00
-7.18999565e-01 1.71280771e-01 -5.06308019e-01 -6.94582701e-01
5.42853117e-01 3.43713582e-01 7.40214705e-01 -1.49301243e+00
-1.13889702e-01 4.13310885e-01 -8.61782014e-01 -4.37830061e-01
8.87803137e-01 5.02727985e-01 -1.92494407e-01 6.16605043e-01
-1.30226448e-01 5.58293581e-01 5.69073915e-01 6.35396242e-01
1.23059261e+00 -1.78982154e-01 1.71358466e-01 3.06637496e-01
3.61950666e-01 1.48367345e-01 2.55402595e-01 9.23625350e-01
2.78612107e-01 2.58589983e-01 1.53843304e-02 -1.02414250e-01
-8.26414466e-01 -9.37447906e-01 3.79855298e-02 1.77025080e+00
4.11111683e-01 -8.57640266e-01 -2.62356490e-01 -5.05897403e-01
-1.82467952e-01 8.61175716e-01 -1.08802535e-01 -4.65469450e-01
-6.63415968e-01 -6.98728919e-01 3.77755284e-01 4.49137092e-01
5.29933631e-01 -1.11151350e+00 -5.26629269e-01 4.32893187e-01
-7.88108766e-01 -6.72946751e-01 -1.03778005e+00 4.07873988e-02
-6.14707232e-01 -8.49811435e-01 -6.50046885e-01 -7.15147674e-01
2.51875706e-02 4.61681962e-01 1.54720116e+00 -2.15712027e-03
-2.02633366e-01 3.49223047e-01 -8.68360460e-01 -3.01505387e-01
2.96259582e-01 5.55717409e-01 -3.60024631e-01 -3.10384661e-01
7.08100975e-01 -4.42830324e-01 -1.41021693e+00 4.79085088e-01
-1.01555300e+00 -3.60479474e-01 7.68111408e-01 7.09348619e-01
8.69882166e-01 1.38350397e-01 8.10905576e-01 -7.24015057e-01
1.01555252e+00 -1.03886247e+00 -4.97389495e-01 5.66296518e-01
-1.05006421e+00 -3.48917097e-02 3.94257933e-01 -5.31462431e-01
-1.01750100e+00 -4.00208235e-01 -2.56640732e-01 -1.12548389e-01
3.63019258e-01 9.86094832e-01 4.10024792e-01 4.22427922e-01
7.93365121e-01 4.94166195e-01 -7.30237544e-01 -6.56428874e-01
2.17537373e-01 7.88337588e-01 3.53679568e-01 -4.31738287e-01
4.44607884e-01 2.78296411e-01 -7.55081654e-01 -3.93583894e-01
-1.01789141e+00 -1.30562055e+00 3.62454653e-01 -9.07761604e-02
3.78618389e-01 -1.15913343e+00 -5.51882148e-01 -9.06824097e-02
-9.19100702e-01 -6.76596314e-02 -3.87870401e-01 6.13267779e-01
-2.34465569e-01 1.87458396e-01 -7.87779629e-01 -5.49526274e-01
-8.93372178e-01 -6.18997633e-01 1.36243093e+00 4.81200665e-01
-1.76868692e-01 -5.62403977e-01 3.88084024e-01 2.18244970e-01
5.79296589e-01 -4.69390661e-01 5.55437028e-01 -9.26437199e-01
-6.71128392e-01 -5.13454616e-01 -3.25115532e-01 -2.32938826e-01
-3.08011442e-01 -4.26092505e-01 -5.68203032e-01 -3.18664551e-01
-2.31232837e-01 -4.25962418e-01 1.06187820e+00 3.71087670e-01
1.38622868e+00 -4.18055564e-01 -6.71800137e-01 5.83811164e-01
1.24104071e+00 3.74700189e-01 6.22370720e-01 6.63213849e-01
3.22920643e-03 1.18116356e-01 1.03241217e+00 9.01968598e-01
6.28861129e-01 7.79970169e-01 -2.71575786e-02 2.07608216e-03
4.28968742e-02 -7.90158451e-01 2.98659146e-01 8.77803683e-01
9.86274183e-02 -6.33041322e-01 -6.09099030e-01 5.87785661e-01
-1.70427680e+00 -1.17768443e+00 4.87224162e-01 2.12384582e+00
1.16008818e+00 1.97253019e-01 3.05095226e-01 -9.95238125e-02
5.54663479e-01 1.42516002e-01 -6.29900336e-01 1.79395050e-01
6.50633425e-02 5.39810658e-01 4.08052802e-01 1.77579775e-01
-1.04333055e+00 7.85630822e-01 5.58804655e+00 1.30480826e+00
-1.03520977e+00 2.57295698e-01 5.33732712e-01 -2.52298564e-01
-4.38210875e-01 -7.15114316e-03 -1.07328713e+00 7.24305093e-01
1.10121214e+00 -6.24919772e-01 5.32715261e-01 8.36551070e-01
2.98721343e-01 -2.53134798e-02 -7.73200452e-01 9.48432565e-01
-6.31099567e-02 -1.39890802e+00 1.58815801e-01 -1.28525645e-01
5.33085883e-01 1.94091514e-01 1.66318249e-02 9.23677385e-01
1.86963215e-01 -6.16392195e-01 5.54679811e-01 6.62803769e-01
7.34801233e-01 -6.11961126e-01 5.04949570e-01 4.64274526e-01
-1.57795274e+00 -3.39356869e-01 -4.71460521e-02 5.62566698e-01
5.64423144e-01 6.51004612e-01 -6.00084007e-01 6.07680857e-01
1.03357387e+00 5.46704590e-01 -5.04295588e-01 1.40837669e+00
1.90294150e-03 8.28860700e-01 -7.42885113e-01 -2.13790238e-01
1.04380444e-01 9.84168425e-02 5.25917888e-01 1.23094678e+00
6.45686567e-01 2.00404823e-01 3.44695628e-01 5.76821208e-01
-2.70037353e-01 4.07513976e-01 -4.72455099e-03 -1.31495282e-01
9.83633637e-01 1.10781074e+00 -5.21081686e-01 -4.85389262e-01
-5.26510596e-01 1.04432881e+00 1.96529329e-01 4.86534029e-01
-1.14570427e+00 -7.11651087e-01 -8.51665139e-02 2.67050892e-01
4.09669459e-01 1.75003007e-01 3.83349627e-01 -1.21593821e+00
2.87364751e-01 -8.32895219e-01 1.11790621e+00 -7.08090305e-01
-1.19584751e+00 6.94857657e-01 6.80567175e-02 -1.18470812e+00
-5.33669710e-01 2.01317847e-01 -5.76538324e-01 7.51238644e-01
-1.72353196e+00 -8.63982737e-01 -3.88338506e-01 6.82435036e-01
7.97235906e-01 -6.38472289e-02 5.53324342e-01 1.10256052e+00
-2.24467248e-01 7.50817358e-01 5.79762310e-02 -1.57937974e-01
8.04246664e-01 -9.87047553e-01 2.77537555e-01 5.07324934e-01
2.66405702e-01 7.03496516e-01 3.11500430e-01 -8.16556096e-01
-1.12711799e+00 -1.04351747e+00 1.29815304e+00 -1.69828877e-01
7.26144612e-01 -3.93338166e-02 -7.81081796e-01 4.84320581e-01
6.76812679e-02 -3.31022739e-01 6.09738111e-01 2.53498375e-01
-2.97280699e-01 -3.22812259e-01 -4.16150957e-01 5.01235366e-01
1.03695381e+00 -7.58443892e-01 -6.19909585e-01 6.72506273e-01
1.10008323e+00 -2.17646807e-01 -8.05835843e-01 3.40045542e-01
5.42936385e-01 -5.40158749e-01 1.42085826e+00 -4.68626589e-01
2.05451965e-01 -6.68226928e-02 4.77818027e-02 -9.89731312e-01
-2.40307361e-01 -7.99191952e-01 -6.48002982e-01 1.04198599e+00
6.10455275e-01 -4.57197994e-01 5.70090890e-01 1.39632761e-01
-5.55889532e-02 -8.01329970e-01 -4.48388427e-01 -6.01473570e-01
-4.87155169e-01 -2.85949200e-01 6.09581947e-01 5.28964758e-01
5.84889352e-02 7.18128741e-01 -2.61459440e-01 -7.87914619e-02
2.19133079e-01 4.82473940e-01 3.51642847e-01 -1.06463170e+00
-4.19340372e-01 -5.12050152e-01 5.02910495e-01 -1.74609375e+00
-4.02680635e-01 -9.11100328e-01 -8.48850384e-02 -1.42623544e+00
4.55857933e-01 -5.40207326e-01 -8.00018251e-01 1.23497412e-01
-3.86040419e-01 -9.08768848e-02 -3.91625613e-01 5.98221481e-01
-1.35897624e+00 5.94623387e-01 8.82833958e-01 4.09684815e-02
-6.43455207e-01 3.90483707e-01 -7.85828292e-01 8.98351818e-02
7.16669202e-01 -4.82539713e-01 -6.09819174e-01 -3.38316113e-01
9.02242482e-01 3.50198805e-01 2.32999161e-01 -6.57563090e-01
6.92810774e-01 1.58365555e-02 2.39864811e-01 -9.67656553e-01
4.53177206e-02 -5.63545942e-01 1.15636609e-01 2.10858971e-01
-8.28869939e-01 3.25339079e-01 -1.22919351e-01 8.39652598e-01
-6.66981697e-01 -1.82214528e-01 8.97789448e-02 -1.45927042e-01
-8.69904935e-01 6.61332250e-01 -1.92899648e-02 3.98736686e-01
6.45071626e-01 3.69518906e-01 -2.43820399e-01 -7.10913122e-01
-5.41945517e-01 6.81317210e-01 -2.71466047e-01 4.68878329e-01
5.35334885e-01 -1.41264367e+00 -7.36471951e-01 -5.75917810e-02
4.02621865e-01 -1.88910812e-01 2.81184107e-01 8.36550891e-01
-2.68243015e-01 7.54511178e-01 4.35185134e-01 -3.08085531e-01
-9.89760280e-01 5.09826064e-01 -2.02359036e-01 -8.51754785e-01
-5.02376735e-01 9.00952518e-01 -1.30454257e-01 3.48696448e-02
6.51749551e-01 -2.88643949e-02 -5.28687239e-01 2.09921703e-01
7.84820318e-01 4.80572581e-01 1.81452245e-01 -4.54793088e-02
-1.82350859e-01 1.95879444e-01 -3.46547067e-01 -2.92432725e-01
1.25913918e+00 -5.36964595e-01 1.38265818e-01 4.87975962e-02
1.24025440e+00 5.58450446e-02 -6.14016533e-01 -6.17244303e-01
3.83285016e-01 -3.01167220e-01 2.07880750e-01 -9.49041903e-01
-8.13053906e-01 3.45874310e-01 4.86651629e-01 4.93910015e-01
1.62949562e+00 3.63424152e-01 1.62224472e+00 5.11086226e-01
1.61286235e-01 -1.41675007e+00 3.51643473e-01 5.65580904e-01
8.10980678e-01 -1.21079934e+00 -1.36711821e-01 -1.48696706e-01
-3.95052820e-01 7.22639561e-01 3.79838616e-01 8.90465975e-02
8.53368580e-01 -1.43335491e-01 -3.42285097e-01 -4.54787016e-01
-1.13422024e+00 -3.42316180e-01 6.48935258e-01 -2.61356533e-01
3.42080921e-01 -2.15495512e-01 -6.59959435e-01 8.63395214e-01
-1.58570215e-01 2.74272710e-01 -4.57854718e-01 1.00453842e+00
-5.30957520e-01 -8.91572177e-01 1.60434674e-02 7.72111833e-01
-6.48703933e-01 -4.18865204e-01 4.02319431e-02 4.17189837e-01
-2.94952959e-01 9.67765749e-01 1.01961002e-01 -6.08135879e-01
5.67510247e-01 2.13238016e-01 -8.16389471e-02 -3.45364749e-01
-7.09939718e-01 4.64716941e-01 7.63924494e-02 -8.86691451e-01
-6.23336248e-02 -5.13870895e-01 -1.08560073e+00 -1.79556161e-01
-5.85620046e-01 7.09700644e-01 4.69893932e-01 4.87997204e-01
8.34882796e-01 5.76116383e-01 9.58159387e-01 -1.44411996e-01
-7.78074980e-01 -1.01808596e+00 -2.48535618e-01 5.19164324e-01
-2.95251776e-02 -5.36720693e-01 -3.34539562e-01 -9.34219360e-02] | [11.481141090393066, 7.589568614959717] |
15f5ae8d-c189-4f8d-8c1b-6939502745e5 | counterfactual-explanation-for-fairness-in | 2307.04386 | null | https://arxiv.org/abs/2307.04386v1 | https://arxiv.org/pdf/2307.04386v1.pdf | Counterfactual Explanation for Fairness in Recommendation | Fairness-aware recommendation eliminates discrimination issues to build trustworthy recommendation systems.Explaining the causes of unfair recommendations is critical, as it promotes fairness diagnostics, and thus secures users' trust in recommendation models. Existing fairness explanation methods suffer high computation burdens due to the large-scale search space and the greedy nature of the explanation search process. Besides, they perform score-based optimizations with continuous values, which are not applicable to discrete attributes such as gender and race. In this work, we adopt the novel paradigm of counterfactual explanation from causal inference to explore how minimal alterations in explanations change model fairness, to abandon the greedy search for explanations. We use real-world attributes from Heterogeneous Information Networks (HINs) to empower counterfactual reasoning on discrete attributes. We propose a novel Counterfactual Explanation for Fairness (CFairER) that generates attribute-level counterfactual explanations from HINs for recommendation fairness. Our CFairER conducts off-policy reinforcement learning to seek high-quality counterfactual explanations, with an attentive action pruning reducing the search space of candidate counterfactuals. The counterfactual explanations help to provide rational and proximate explanations for model fairness, while the attentive action pruning narrows the search space of attributes. Extensive experiments demonstrate our proposed model can generate faithful explanations while maintaining favorable recommendation performance. | ['Guandong Xu', 'Qing Li', 'Dianer Yu', 'Qian Li', 'Xiangmeng Wang'] | 2023-07-10 | null | null | null | null | ['fairness', 'causal-inference', 'counterfactual-explanation', 'fairness', 'causal-inference'] | ['computer-vision', 'knowledge-base', 'miscellaneous', 'miscellaneous', 'miscellaneous'] | [-1.33922407e-02 4.67129618e-01 -1.04067254e+00 -7.28594542e-01
1.42564671e-02 -1.04292259e-01 4.09574717e-01 8.32343940e-03
-1.68809459e-01 1.11050284e+00 6.69648588e-01 -8.33322227e-01
-5.70044518e-01 -9.72145379e-01 -2.97909766e-01 -1.52250916e-01
3.20437849e-02 2.29234576e-01 -4.93635386e-01 -1.85652927e-01
6.20105088e-01 9.35158357e-02 -1.42229092e+00 2.62884498e-01
1.71675766e+00 7.25303471e-01 -4.22888905e-01 3.26283485e-01
-1.33616582e-01 6.76490605e-01 -2.47827545e-01 -1.09381056e+00
4.44401652e-01 -5.09906769e-01 -7.51912951e-01 -5.10115743e-01
3.45741004e-01 -6.81845367e-01 -1.27381012e-01 1.07438970e+00
3.13165247e-01 3.59346241e-01 5.40878773e-01 -1.82993686e+00
-1.33387816e+00 1.15744746e+00 -4.19592500e-01 2.70656273e-02
2.71677405e-01 2.38537654e-01 1.55868936e+00 -3.20110589e-01
1.38977215e-01 1.66655123e+00 4.55195040e-01 1.05620682e+00
-1.24335051e+00 -1.16691995e+00 7.17803478e-01 3.91053140e-01
-7.64048696e-01 -3.72974128e-01 6.55586541e-01 9.43962578e-03
4.75658059e-01 8.70429575e-01 8.51995707e-01 1.03259456e+00
1.47974446e-01 4.30526137e-01 1.39652240e+00 -2.48134777e-01
5.43663204e-01 1.76848918e-01 3.33357483e-01 6.34350121e-01
9.91771996e-01 7.29559779e-01 -4.47081625e-01 -6.68634057e-01
6.61530614e-01 4.15402293e-01 -1.87759981e-01 -1.33898824e-01
-8.54517579e-01 1.34291315e+00 5.33219993e-01 -4.94995683e-01
-6.42250836e-01 1.42159671e-01 6.52511567e-02 3.34323108e-01
3.20743412e-01 8.93219948e-01 -5.83067715e-01 1.95311755e-01
-4.98894691e-01 4.96878207e-01 7.76326716e-01 5.10374844e-01
6.15085423e-01 1.30994067e-01 -6.20008290e-01 5.64731598e-01
6.38973892e-01 5.55581450e-01 3.62864733e-01 -1.50406551e+00
2.42354333e-01 5.89614868e-01 2.66154915e-01 -1.10500264e+00
-1.51000440e-01 -4.16729659e-01 -8.21620345e-01 2.89103359e-01
4.14310545e-01 -1.74229071e-01 -5.37493229e-01 1.77757275e+00
4.39034015e-01 8.27715918e-02 -1.26363173e-01 1.37807095e+00
4.95216131e-01 1.30306050e-01 5.86496055e-01 -5.13367474e-01
1.33679700e+00 -6.74453139e-01 -7.94180751e-01 -1.08403370e-01
2.15097293e-01 -4.02897358e-01 1.25847781e+00 2.09097803e-01
-7.65830815e-01 -1.85824841e-01 -6.52348757e-01 1.72936961e-01
-1.25644598e-02 -3.80416274e-01 1.39727855e+00 9.37754631e-01
-3.95330429e-01 7.67270684e-01 -6.87441677e-02 9.02763531e-02
8.36114287e-01 4.87313002e-01 3.10265645e-02 -5.61642274e-02
-1.66445160e+00 6.43700242e-01 -2.94170342e-02 -1.31389439e-01
-5.83393633e-01 -8.93735647e-01 -5.88920116e-01 5.97115159e-01
8.74365449e-01 -1.09903741e+00 9.60819483e-01 -1.10522878e+00
-1.55954623e+00 1.48057818e-01 -8.14353675e-02 -5.35900712e-01
5.54304361e-01 -9.66918245e-02 -6.01300597e-01 -2.78848380e-01
2.02957720e-01 5.08046389e-01 8.83962333e-01 -1.34109294e+00
-5.75255752e-01 -3.41360420e-01 4.13833499e-01 1.82889521e-01
-2.16204345e-01 -1.93148866e-01 5.13409078e-01 -6.67238474e-01
-1.19025141e-01 -8.23478878e-01 -7.13908553e-01 -1.24147646e-02
-5.28636456e-01 -1.02363303e-01 1.28751129e-01 -3.90085220e-01
1.27540576e+00 -1.72964859e+00 -6.32385492e-01 5.44693887e-01
5.73108971e-01 8.14739466e-02 -1.61416709e-01 -2.15120018e-01
-9.79275182e-02 7.13419199e-01 2.19018504e-01 2.19522715e-01
3.95274043e-01 2.11546808e-01 -5.83638310e-01 2.30709583e-01
-3.76006693e-01 8.75560999e-01 -9.04275715e-01 -3.88205081e-01
-9.87494737e-02 -1.12780213e-01 -1.22518456e+00 2.25191027e-01
-3.22345048e-01 2.72522178e-02 -7.83748269e-01 3.85528505e-01
7.81430185e-01 -5.71335740e-02 3.08747023e-01 -7.85975680e-02
-1.20189106e-02 6.38886571e-01 -1.10184574e+00 8.93333435e-01
-2.73014665e-01 -1.62143990e-01 -3.76188010e-01 -6.26406431e-01
8.95417273e-01 1.11935943e-01 1.18748024e-01 -8.01447809e-01
1.17161684e-01 1.71899080e-01 4.83877689e-01 -3.00870121e-01
3.80576402e-01 -7.15525210e-01 -5.92503734e-02 9.13200557e-01
-5.70034444e-01 2.45681897e-01 -3.60497445e-01 3.72428209e-01
3.98237050e-01 -2.15777025e-01 7.89474845e-01 -3.55633646e-01
4.32126224e-01 -2.93819815e-01 1.10175669e+00 1.08029938e+00
-4.61500764e-01 9.53020975e-02 6.65383995e-01 -8.37034822e-01
-6.83188975e-01 -9.33034956e-01 1.54586285e-01 1.20351946e+00
3.12579811e-01 -2.65069306e-01 -3.78420681e-01 -1.25603342e+00
5.84915102e-01 1.35424519e+00 -8.90692830e-01 -7.23595679e-01
-1.09071843e-02 -8.67548764e-01 2.57459790e-01 3.57361406e-01
4.76196855e-01 -8.19050670e-01 -4.34366465e-01 -1.64056256e-01
-2.73243815e-01 -2.70765066e-01 -7.33069181e-01 -3.98542553e-01
-8.93223643e-01 -1.16302729e+00 -2.79760566e-02 3.69153738e-01
5.43832481e-01 3.67923081e-01 1.06679082e+00 6.53381586e-01
1.28280699e-01 -6.92760721e-02 -9.76754799e-02 -6.08318388e-01
-2.71757007e-01 -4.09227788e-01 4.72808421e-01 -7.34121129e-02
6.03014827e-01 -4.66230214e-01 -1.01913142e+00 3.54505688e-01
-3.39540660e-01 2.42235586e-02 4.89093810e-01 1.05241287e+00
3.16110969e-01 -2.17109263e-01 1.07475603e+00 -1.52866852e+00
1.11616302e+00 -8.02318633e-01 -4.09253210e-01 3.62691134e-01
-1.72673357e+00 2.52328485e-01 1.10079122e+00 -3.49162519e-01
-1.55256760e+00 -5.96845686e-01 2.86811501e-01 -2.76173488e-03
-4.67211753e-02 1.21107452e-01 -3.39783996e-01 2.36953929e-01
7.93185949e-01 -3.69163007e-01 8.49183649e-02 -2.78152108e-01
7.96675682e-01 5.33183038e-01 2.15937912e-01 -8.96232188e-01
6.69574678e-01 3.77902806e-01 -6.10400438e-02 1.76567256e-01
-1.11256325e+00 7.34996051e-02 4.45410330e-03 -4.59578447e-02
5.08340716e-01 -3.99023801e-01 -1.41359901e+00 -5.78996420e-01
-8.44194949e-01 5.60399219e-02 -3.64308983e-01 6.91515923e-01
-5.00023842e-01 1.70046821e-01 -3.04692745e-01 -1.10123444e+00
-4.13011968e-01 -7.03226745e-01 3.09575260e-01 4.47770596e-01
-6.99869573e-01 -7.91026354e-01 -1.28141209e-01 6.48348212e-01
3.50504160e-01 -1.23848356e-01 1.20162499e+00 -8.21431577e-01
-5.33379197e-01 1.31616980e-01 -4.04274255e-01 -3.33512098e-01
-4.83795448e-05 4.35940735e-02 -8.46163154e-01 3.56838703e-02
-3.65172088e-01 -7.40709482e-03 6.74860775e-01 4.92660761e-01
1.50340486e+00 -1.24704313e+00 -1.23665608e-01 5.30083537e-01
1.03950942e+00 1.61871314e-01 4.95522499e-01 3.29542816e-01
4.56160665e-01 7.41050780e-01 8.73952568e-01 8.30019593e-01
6.31960869e-01 4.64049727e-01 7.36624420e-01 -1.48849770e-01
1.85788780e-01 -6.76106751e-01 -6.32632943e-03 3.12402658e-02
-5.86228073e-01 2.01081485e-01 -2.74844199e-01 1.32475346e-01
-2.08613396e+00 -1.46159732e+00 -1.72253385e-01 2.42154408e+00
6.47454321e-01 -5.14742881e-02 2.90711731e-01 -1.24012753e-01
6.80673182e-01 -7.59903118e-02 -8.21677506e-01 -8.65680635e-01
2.98254520e-01 -1.87369645e-01 4.51216787e-01 8.31771970e-01
-4.50480878e-01 7.48380303e-01 5.98312664e+00 4.79066759e-01
-7.60615230e-01 -7.39261732e-02 9.55639958e-01 -4.21924233e-01
-1.32204413e+00 3.74346703e-01 -3.50891352e-01 5.31373978e-01
7.55131721e-01 -6.34866416e-01 7.99601495e-01 9.20544088e-01
7.44090855e-01 3.06326538e-01 -9.31956053e-01 5.14251709e-01
-4.18168426e-01 -1.37133992e+00 5.28039575e-01 1.87455967e-01
8.39736700e-01 -6.97879553e-01 9.71108451e-02 3.26369643e-01
9.54247713e-01 -1.25033414e+00 7.56296992e-01 6.86525702e-01
6.80909693e-01 -9.22619820e-01 7.60011673e-01 2.34902695e-01
-5.19484699e-01 -6.34128034e-01 -6.97437882e-01 -5.72842002e-01
-1.12299576e-01 5.57765126e-01 -6.59469724e-01 3.73729229e-01
4.35754567e-01 4.22266871e-01 -2.39674151e-01 6.38879895e-01
-5.20460188e-01 6.67592466e-01 3.43075067e-01 -2.11982340e-01
-1.45067945e-01 -3.19252968e-01 2.76340574e-01 7.03975737e-01
1.05482578e-01 7.19713688e-01 -1.00803517e-01 1.27613950e+00
-2.29645729e-01 2.39638805e-01 -3.88053417e-01 3.08965147e-01
9.80245948e-01 1.08217800e+00 -3.03308100e-01 -3.93942863e-01
-9.08325016e-02 5.27308464e-01 2.94263154e-01 4.78918165e-01
-8.79004896e-01 3.46193314e-02 1.08278644e+00 1.44656047e-01
-3.69727552e-01 7.08699167e-01 -8.97353351e-01 -1.29049206e+00
-5.24110019e-01 -1.23233759e+00 8.94051194e-01 -4.53464985e-01
-1.62636995e+00 2.52458453e-01 -1.43410072e-01 -8.05412650e-01
-8.95582065e-02 -3.54099087e-02 -1.08241093e+00 9.55514073e-01
-1.66668165e+00 -8.38955224e-01 -1.20002806e-01 4.60859299e-01
2.21995577e-01 -2.11789086e-01 7.90884316e-01 -1.16603812e-02
-4.98826444e-01 9.50066447e-01 -2.50609189e-01 -4.11133468e-01
7.48260856e-01 -1.24954593e+00 1.32753223e-01 6.47082448e-01
-9.18109119e-02 1.38633418e+00 8.51407349e-01 -6.92038059e-01
-9.14950788e-01 -8.65598798e-01 9.67217803e-01 -2.03448281e-01
2.35671446e-01 4.14462835e-02 -8.28498542e-01 4.49983507e-01
-1.99003279e-01 -6.75015971e-02 1.18594038e+00 7.50107288e-01
-7.67205298e-01 -2.02553391e-01 -1.48874795e+00 1.07394493e+00
1.25098324e+00 -1.98172867e-01 -5.50998628e-01 -3.42585370e-02
8.67188334e-01 2.86609739e-01 -4.81537193e-01 9.20729041e-02
1.03145564e+00 -1.15806293e+00 1.06749225e+00 -1.34690011e+00
7.84228265e-01 -1.74694255e-01 2.57164668e-02 -1.53988850e+00
-9.10338521e-01 -7.67340183e-01 -2.03769818e-01 1.03832603e+00
4.15086418e-01 -9.12283242e-01 7.81872869e-01 1.43485165e+00
2.47771531e-01 -8.40707064e-01 -5.88997424e-01 -4.50920552e-01
-4.25860174e-02 -2.92476922e-01 1.60296357e+00 1.31118429e+00
3.69624376e-01 2.85723120e-01 -6.20677948e-01 1.38257205e-01
1.02590966e+00 7.69748449e-01 6.37789369e-01 -1.43008745e+00
-4.61085111e-01 -5.65732241e-01 3.36693376e-01 -6.48720384e-01
3.61936331e-01 -6.79092884e-01 -2.33590752e-01 -1.19674742e+00
5.31029463e-01 -7.72851646e-01 -3.53692204e-01 6.30615771e-01
-8.64143848e-01 -1.86245814e-01 3.48037302e-01 2.07223877e-01
-6.41066611e-01 7.80380189e-01 1.37118673e+00 5.00538424e-02
9.43490788e-02 2.55694002e-01 -1.68419933e+00 7.97534943e-01
8.78770590e-01 -6.57238722e-01 -6.91749871e-01 1.70275480e-01
2.92133063e-01 2.36629009e-01 4.51274723e-01 7.41455853e-02
-1.97341517e-01 -1.17615509e+00 3.39433998e-01 3.11423510e-01
-2.96639860e-01 -7.33899474e-01 1.67599365e-01 7.70577610e-01
-9.91451681e-01 -1.39530554e-01 -3.76165628e-01 8.53894532e-01
3.28969479e-01 -1.21001475e-01 5.86669922e-01 -2.48795494e-01
-3.83244038e-01 4.65663940e-01 -8.42919350e-02 -4.23223563e-02
7.17148542e-01 -1.88550174e-01 -6.85741127e-01 -7.43170977e-01
-4.92914200e-01 2.90703773e-01 3.88182074e-01 3.26950967e-01
5.55504978e-01 -1.51137924e+00 -7.43266046e-01 2.62177102e-02
5.32061346e-02 -7.31650591e-01 3.64398181e-01 3.73216510e-01
1.86569721e-01 3.93519402e-01 -4.93531078e-01 4.10784870e-01
-1.13629520e+00 7.52968669e-01 3.13754916e-01 -1.90310135e-01
-1.47806346e-01 6.62354767e-01 3.53209645e-01 -6.51819587e-01
-1.30227998e-01 1.30241811e-01 -4.30479288e-01 -3.74293745e-01
5.27692795e-01 8.25507522e-01 -5.94194531e-01 2.91973613e-02
-2.21240118e-01 -7.73623437e-02 -2.96862312e-02 1.74094737e-01
1.07599330e+00 -4.39441115e-01 -1.36307791e-01 -1.45852298e-01
4.78432834e-01 3.89373273e-01 -1.24419713e+00 1.54375806e-01
-1.67734414e-01 -1.35614800e+00 -8.65169838e-02 -1.30423355e+00
-9.70453680e-01 5.84099948e-01 1.18711412e-01 3.55847210e-01
9.77477789e-01 -4.69454169e-01 5.58945477e-01 1.21880598e-01
2.33670384e-01 -9.04774427e-01 -3.12270582e-01 -1.20588459e-01
6.83670640e-01 -1.43084228e+00 3.66724938e-01 -6.53548121e-01
-8.42553794e-01 9.13194656e-01 9.93553996e-01 1.97343305e-01
3.16215634e-01 -4.93788421e-01 1.87930185e-02 -1.49849623e-01
-1.04504406e+00 1.75385028e-01 4.65888023e-01 5.91508865e-01
6.14566565e-01 6.61935151e-01 -8.51732969e-01 1.43537927e+00
-4.28119302e-01 -2.62548868e-02 6.13579988e-01 -1.06947258e-01
-3.52209657e-01 -1.01821721e+00 -4.66985136e-01 1.03507805e+00
-5.80159962e-01 -3.41539502e-01 -4.81663018e-01 4.43854094e-01
2.81741351e-01 1.07687676e+00 -2.08641082e-01 -2.79888123e-01
3.19607466e-01 -2.44443178e-01 3.48964483e-02 -3.56720179e-01
-5.69955766e-01 -3.59888077e-01 3.55302811e-01 -8.17100763e-01
-7.81474113e-02 -5.69149196e-01 -1.37440026e+00 -1.10371554e+00
-4.76191103e-01 7.28400648e-01 5.57077825e-02 1.03785586e+00
5.81200123e-01 2.95310080e-01 8.07437241e-01 -3.36560681e-02
-1.12651026e+00 -4.25387114e-01 -5.03642857e-01 6.88070893e-01
2.32964158e-01 -7.87227035e-01 -3.42180043e-01 -4.81485158e-01] | [9.448477745056152, 5.6040215492248535] |
e0476a50-a8fe-4bb9-a045-3284dda1b4e2 | 190503646 | 1905.03646 | null | https://arxiv.org/abs/1905.03646v3 | https://arxiv.org/pdf/1905.03646v3.pdf | TE141K: Artistic Text Benchmark for Text Effect Transfer | Text effects are combinations of visual elements such as outlines, colors and textures of text, which can dramatically improve its artistry. Although text effects are extensively utilized in the design industry, they are usually created by human experts due to their extreme complexity; this is laborious and not practical for normal users. In recent years, some efforts have been made toward automatic text effect transfer; however, the lack of data limits the capabilities of transfer models. To address this problem, we introduce a new text effects dataset, TE141K, with 141,081 text effect/glyph pairs in total. Our dataset consists of 152 professionally designed text effects rendered on glyphs, including English letters, Chinese characters, and Arabic numerals. To the best of our knowledge, this is the largest dataset for text effect transfer to date. Based on this dataset, we propose a baseline approach called text effect transfer GAN (TET-GAN), which supports the transfer of all 152 styles in one model and can efficiently extend to new styles. Finally, we conduct a comprehensive comparison in which 14 style transfer models are benchmarked. Experimental results demonstrate the superiority of TET-GAN both qualitatively and quantitatively and indicate that our dataset is effective and challenging. | ['Wenjing Wang', 'Shuai Yang', 'Jiaying Liu'] | 2019-05-08 | null | null | null | null | ['text-effects-transfer'] | ['natural-language-processing'] | [ 5.30575454e-01 -4.05760258e-01 -7.28418678e-02 -1.90777764e-01
-2.65652061e-01 -5.07430971e-01 6.18893027e-01 -6.38321698e-01
9.20560062e-02 8.66021216e-01 3.70087832e-01 -1.84195060e-02
2.14749262e-01 -7.07104802e-01 -6.84155941e-01 -5.53544104e-01
5.56039453e-01 1.10044040e-01 1.79549053e-01 -3.60162348e-01
5.20952284e-01 4.22041476e-01 -1.17891097e+00 3.79559010e-01
1.12746131e+00 6.51780963e-01 4.24117208e-01 5.23745477e-01
-6.86235055e-02 1.03161919e+00 -9.10830081e-01 -6.40312254e-01
-1.05449175e-02 -8.67656887e-01 -3.44767630e-01 2.91347504e-04
4.60219413e-01 -7.58240283e-01 -5.04265070e-01 7.58005202e-01
9.30848718e-01 6.19986951e-02 9.42230105e-01 -1.18825614e+00
-1.51882458e+00 5.37725747e-01 -8.35272014e-01 -4.33331966e-01
3.24368387e-01 2.76913762e-01 8.34145904e-01 -1.05074728e+00
6.10969305e-01 1.44539988e+00 5.49307406e-01 7.85897553e-01
-1.05646038e+00 -8.74821424e-01 1.95627790e-02 -1.15258247e-01
-8.68595421e-01 -1.75714940e-01 1.17879581e+00 -2.89543062e-01
5.24261594e-01 3.86772543e-01 8.46640229e-01 1.28460813e+00
2.44003862e-01 1.22369754e+00 1.20961285e+00 -4.95861351e-01
2.32483968e-02 -2.55268496e-02 -5.10131478e-01 3.72862756e-01
-2.59205461e-01 2.08566096e-02 -5.85459411e-01 2.77596980e-01
1.25141883e+00 -5.05252294e-02 -3.93495023e-01 -4.64509614e-02
-1.26073039e+00 6.61406755e-01 4.67722654e-01 2.04388410e-01
-1.82336625e-02 2.97203660e-01 3.67630750e-01 3.84325385e-01
8.21997941e-01 4.12525952e-01 -3.04704726e-01 -3.61101151e-01
-4.96073544e-01 3.13258529e-01 5.63280940e-01 1.33295441e+00
4.09611106e-01 2.71918952e-01 -4.94295299e-01 1.56198621e+00
-3.69002372e-02 8.94475102e-01 1.35803178e-01 -4.85863209e-01
4.27218527e-01 5.18060207e-01 7.23206475e-02 -1.04707432e+00
-8.56166482e-02 8.64837319e-02 -1.03574276e+00 2.92836577e-01
2.16076776e-01 -4.75200117e-01 -8.69880259e-01 1.57163262e+00
7.97243342e-02 -3.47396970e-01 -5.08962691e-01 9.41702724e-01
1.08277130e+00 9.66083527e-01 -6.80779144e-02 5.58816679e-02
9.24782157e-01 -1.28794301e+00 -9.45206881e-01 -3.09054051e-02
4.83347774e-01 -1.19239879e+00 1.62481546e+00 4.70177740e-01
-1.19292974e+00 -6.57554507e-01 -9.89260316e-01 -1.58781871e-01
-1.82362184e-01 3.29477578e-01 6.38839304e-01 7.44326949e-01
-9.08314228e-01 2.58457035e-01 -1.64172083e-01 -2.28308648e-01
5.86718082e-01 1.59142926e-01 4.85203378e-02 -1.33775517e-01
-1.15205526e+00 6.39464200e-01 -1.06312841e-01 -2.47802008e-02
-4.60599780e-01 -1.10344112e+00 -4.53720123e-01 -1.03187792e-01
2.71269470e-01 -6.48068845e-01 1.30297124e+00 -1.26600301e+00
-2.00762486e+00 5.28200567e-01 1.56232581e-01 2.55340397e-01
6.63611054e-01 -3.57911587e-01 -5.84238410e-01 -2.38027036e-01
-1.47749439e-01 7.88670301e-01 9.88762200e-01 -1.65835023e+00
-5.63795686e-01 -2.18328219e-02 1.53510407e-01 3.60253990e-01
-6.52876019e-01 1.72780022e-01 -7.57721007e-01 -1.49652839e+00
-5.43805957e-01 -1.14296961e+00 1.05015770e-01 2.59968102e-01
-5.16098380e-01 -5.66765554e-02 1.22617722e+00 -7.88822412e-01
1.28687835e+00 -2.16056347e+00 2.30433866e-01 -6.48104623e-02
1.95760846e-01 1.92479387e-01 -1.13776654e-01 6.80941820e-01
8.25252905e-02 1.71376958e-01 -2.15544105e-01 -2.92148530e-01
2.65229315e-01 -7.76576847e-02 -3.49362314e-01 2.91368943e-02
-9.27980244e-02 1.19734359e+00 -6.78856432e-01 -4.43769157e-01
5.34381509e-01 5.45475602e-01 -7.31866896e-01 1.34178191e-01
-3.17085028e-01 7.30461478e-01 -6.42724276e-01 7.28131294e-01
8.09224188e-01 -6.20354526e-02 -1.02357961e-01 -4.54369895e-02
-9.86169577e-02 -9.64852944e-02 -7.66746938e-01 1.67753482e+00
-7.58886456e-01 9.41423297e-01 -4.21407878e-01 -2.99679130e-01
1.03080165e+00 8.99357647e-02 3.83882284e-01 -9.60711360e-01
1.93778500e-01 1.55976072e-01 -1.27544552e-01 -3.40514719e-01
7.88293183e-01 -2.52804220e-01 -1.36112049e-01 4.15584892e-01
-4.14947987e-01 -6.95294857e-01 8.42744634e-02 3.83467600e-02
6.77749276e-01 1.83739170e-01 -1.12507738e-01 -2.48301491e-01
1.69013083e-01 -2.47874528e-01 3.70749176e-01 6.06504619e-01
2.76443273e-01 9.86262679e-01 4.38538224e-01 -3.37331563e-01
-1.19820440e+00 -8.55599225e-01 3.45523395e-02 1.13753676e+00
1.98127434e-01 -3.13827813e-01 -9.64771807e-01 -8.18636119e-01
-1.44239873e-01 7.86818922e-01 -7.69741952e-01 -6.49150759e-02
-7.54716396e-01 -4.95902002e-01 4.99670088e-01 7.90049970e-01
9.72388446e-01 -1.58670998e+00 -2.51133353e-01 2.36809179e-02
-1.59279048e-01 -7.61036575e-01 -1.13330340e+00 -4.91329044e-01
-6.81653738e-01 -6.53039992e-01 -1.36892855e+00 -9.21417117e-01
6.20246053e-01 9.29451659e-02 1.14907575e+00 1.91847622e-01
-1.26900211e-01 1.95716694e-01 -6.22046590e-01 -6.43898726e-01
-5.22589386e-01 -1.06793940e-01 -3.44609886e-01 7.33781548e-04
-1.21405378e-01 -3.82502437e-01 -7.76502073e-01 5.82230926e-01
-9.96612906e-01 4.76489395e-01 9.66490328e-01 8.80998611e-01
2.66108304e-01 -1.29786693e-02 6.20781124e-01 -1.25159931e+00
6.66027665e-01 7.50665814e-02 -2.96071053e-01 2.95384467e-01
-2.61871010e-01 -3.68874460e-01 9.92946744e-01 -5.92532635e-01
-1.70489979e+00 -4.19383973e-01 -1.87569454e-01 -2.09819108e-01
-2.94258120e-03 3.04571509e-01 -5.32175958e-01 -2.87030071e-01
4.99361098e-01 1.33709401e-01 -3.38106096e-01 -4.72630829e-01
4.97736156e-01 7.74867415e-01 2.66102850e-01 -7.02449620e-01
7.56432056e-01 2.97319353e-01 -8.45999457e-03 -1.01335812e+00
-5.33953428e-01 2.76220560e-01 -3.78851503e-01 -5.13195634e-01
6.86123371e-01 -4.39544439e-01 -3.97739321e-01 1.18983996e+00
-1.09350502e+00 -8.24395657e-01 -1.61661386e-01 3.22003737e-02
-5.07697463e-01 2.82269806e-01 -8.18112731e-01 -2.42716163e-01
-2.46273622e-01 -9.68679905e-01 1.02451611e+00 1.76865265e-01
-1.56028062e-01 -1.10865259e+00 -1.47750244e-01 2.64021516e-01
4.77428824e-01 3.51582766e-01 1.00784969e+00 4.50682491e-01
-3.40124726e-01 2.78016627e-02 -3.85188609e-01 2.53849298e-01
6.69920266e-01 1.91314578e-01 -7.80704021e-01 -2.19464526e-01
-3.22625518e-01 -3.31161916e-01 7.09572196e-01 5.88013291e-01
1.58474970e+00 -9.16273370e-02 -2.89422512e-01 4.43774998e-01
1.14245784e+00 7.91940570e-01 1.10122359e+00 1.46557584e-01
1.28185987e+00 4.76102293e-01 6.50495827e-01 6.57771230e-01
1.07673340e-01 7.79033363e-01 4.56529250e-03 -5.98054469e-01
-6.95418417e-01 -5.90770245e-01 6.39823824e-02 9.92566884e-01
-4.26214218e-01 -6.70810103e-01 -5.26993811e-01 3.97201121e-01
-1.50458479e+00 -6.54171169e-01 -2.41752610e-01 1.75281107e+00
9.72053766e-01 7.04461616e-03 -1.17402980e-02 1.79125592e-02
8.02752435e-01 1.22328229e-01 -6.87698185e-01 -3.96743387e-01
-2.78674752e-01 2.76231349e-01 3.10172051e-01 7.12692216e-02
-8.70566368e-01 1.06220186e+00 6.53007174e+00 1.24417233e+00
-1.15668428e+00 -1.62499428e-01 8.17090929e-01 2.33064517e-02
-4.94464129e-01 -2.23608673e-01 -4.33332115e-01 8.09371233e-01
8.95891041e-02 2.32237820e-02 6.02694333e-01 5.04667699e-01
1.93618089e-01 2.13908434e-01 -9.28826869e-01 1.21012616e+00
2.95041621e-01 -1.19999385e+00 4.50512767e-01 -1.12741768e-01
1.43037891e+00 -8.65325809e-01 5.62959373e-01 3.52307349e-01
2.77258873e-01 -1.01134419e+00 8.02889228e-01 4.26764458e-01
1.39631176e+00 -1.05117321e+00 7.95783922e-02 -1.50094181e-01
-1.13347554e+00 3.93698253e-02 -2.34966978e-01 -5.05165346e-02
1.56505167e-01 4.70905274e-01 -2.36103266e-01 4.93745327e-01
7.63490140e-01 1.24064767e+00 -4.95234519e-01 7.54785359e-01
-5.11394799e-01 7.26372182e-01 1.49090067e-01 -3.16109657e-01
1.60226543e-02 -2.41180882e-01 3.83577436e-01 1.08295667e+00
4.33424503e-01 6.75246939e-02 -1.15743376e-01 9.21767890e-01
-4.60965514e-01 2.82288998e-01 -6.16013408e-01 2.03313679e-02
3.32122028e-01 1.05097759e+00 -8.26828718e-01 -3.61157924e-01
-6.18652582e-01 1.29917598e+00 2.74217520e-02 5.66372514e-01
-1.13829780e+00 -6.32924259e-01 3.83212894e-01 -6.73593879e-02
2.48162597e-01 1.06780909e-01 -8.26461732e-01 -1.00561273e+00
4.33144020e-03 -1.15983534e+00 -1.45077631e-02 -1.21684027e+00
-1.41042709e+00 2.34738857e-01 -1.60689056e-01 -1.50398993e+00
1.65865183e-01 -6.46397889e-01 -5.82076848e-01 8.27235937e-01
-1.00200891e+00 -1.49578559e+00 -4.22920167e-01 4.22255874e-01
9.68732893e-01 -1.70528248e-01 4.06803787e-01 3.71723980e-01
-3.13343525e-01 8.45545173e-01 2.71177739e-01 -4.09185588e-02
1.17418921e+00 -1.20380557e+00 6.01841092e-01 5.41347623e-01
-3.23203653e-01 4.34219182e-01 5.30623496e-01 -7.56531000e-01
-1.44687319e+00 -1.01715219e+00 4.00834531e-01 -3.28906149e-01
2.93338269e-01 -4.64235216e-01 -6.30867064e-01 6.48929119e-01
5.67786634e-01 -5.92857778e-01 5.04207075e-01 -4.95624496e-03
1.72150880e-02 2.58707628e-03 -1.02652478e+00 1.23090732e+00
1.26695085e+00 -1.54711887e-01 -1.86603531e-01 7.46775568e-02
7.26893604e-01 -6.06443048e-01 -8.43923390e-01 4.36423242e-01
6.88630939e-01 -9.53203559e-01 7.76783645e-01 5.15487716e-02
1.20001805e+00 -1.89518452e-01 1.47567138e-01 -1.80922294e+00
-4.47217762e-01 -4.74991053e-01 2.18047902e-01 1.64123440e+00
2.10885331e-01 -5.34781694e-01 6.07740879e-01 5.48017144e-01
-3.97443086e-01 -5.85414290e-01 -2.60887742e-01 -5.19913375e-01
3.29351753e-01 -1.96086213e-01 9.68331933e-01 1.23600864e+00
-3.96006525e-01 4.15103734e-01 -1.05178058e+00 -4.32310075e-01
4.12644833e-01 2.56668508e-01 1.06926727e+00 -9.33053017e-01
-4.24617797e-01 -7.38858283e-01 3.79574858e-02 -1.53680980e+00
2.87916027e-02 -5.03232718e-01 -2.62041353e-02 -1.69845176e+00
3.04639220e-01 -6.26726687e-01 1.00534245e-01 4.60468411e-01
-4.16140318e-01 5.34783840e-01 2.60650992e-01 1.27758071e-01
-1.64285749e-01 1.09567225e+00 2.22933149e+00 -3.40253532e-01
-2.44670972e-01 -3.01077098e-01 -7.76337445e-01 6.54817045e-01
7.48281956e-01 -8.13061967e-02 -6.20794833e-01 -6.70601130e-01
1.33547932e-01 3.44360508e-02 1.14432827e-01 -5.91076136e-01
-1.40789360e-01 -2.79474646e-01 7.49870539e-01 -6.30705595e-01
1.79125190e-01 -5.84817052e-01 1.49367720e-01 1.78909600e-01
-3.56644541e-01 -2.03942999e-01 3.75379801e-01 2.85822004e-01
-1.03024453e-01 7.23704100e-02 7.20302880e-01 9.74077433e-02
-8.03694487e-01 3.42131287e-01 -4.22052026e-01 -3.50431912e-02
8.38043451e-01 -2.32632712e-01 -3.18764001e-01 -6.10655010e-01
-6.32025152e-02 -2.95990240e-03 7.50868261e-01 7.14851260e-01
7.54262805e-01 -1.64998460e+00 -7.35120296e-01 8.98687169e-02
1.66115269e-01 -1.29183561e-01 6.84196830e-01 4.20019776e-01
-6.50689304e-01 8.54558051e-02 -4.43828017e-01 -2.20965013e-01
-1.32747853e+00 5.45077324e-01 -4.78719585e-02 -6.91073760e-03
-7.75750339e-01 5.61502993e-01 1.06952000e+00 -5.25649488e-01
-2.00208217e-01 -3.06588173e-01 -2.27730900e-01 -3.93277138e-01
5.52745163e-01 5.44190407e-01 -1.79430708e-01 -2.78219104e-01
2.53217727e-01 8.25541198e-01 -1.37612075e-01 -1.14176199e-02
1.13660383e+00 -6.84344247e-02 -2.28734240e-01 3.71040612e-01
8.35503399e-01 2.64410943e-01 -1.58822346e+00 6.44595996e-02
-6.17930710e-01 -9.22879159e-01 -1.96927145e-01 -1.07363510e+00
-1.38065588e+00 7.66524732e-01 2.88686126e-01 -6.77561536e-02
1.59057868e+00 -2.81111509e-01 1.02404606e+00 1.08621113e-01
2.25169450e-01 -1.20215702e+00 6.12831652e-01 4.63255137e-01
1.30265009e+00 -8.18803906e-01 -3.62467080e-01 -6.77434683e-01
-7.77301610e-01 8.61499488e-01 8.65696371e-01 -1.13112614e-01
2.87641287e-01 3.78403783e-01 7.53089041e-02 1.02677494e-01
-4.04762864e-01 3.75373989e-01 4.19435471e-01 5.80149114e-01
7.00736821e-01 1.41890913e-01 -4.12181258e-01 2.86745131e-01
-3.10379595e-01 8.20982307e-02 4.16242003e-01 9.76568162e-01
-3.46174836e-02 -1.21671474e+00 -4.48306262e-01 5.13225913e-01
-3.91558021e-01 -2.83462554e-01 -7.01547503e-01 8.22618186e-01
-8.54190812e-02 9.23517704e-01 -2.63364851e-01 -2.97840238e-01
5.55715144e-01 -4.28567350e-01 8.34973693e-01 -1.56598628e-01
-4.18366045e-01 1.69376701e-01 4.41800170e-02 -2.22414173e-02
-2.90423959e-01 -4.14782107e-01 -9.36750829e-01 -6.86940491e-01
-9.42249075e-02 -5.05272269e-01 4.13282156e-01 7.06457913e-01
1.00783795e-01 1.00431657e+00 7.77452409e-01 -9.51394379e-01
4.87276018e-02 -1.28126621e+00 -7.53788888e-01 6.46713793e-01
-1.34677738e-01 -8.27168226e-01 -3.92972827e-02 4.25896883e-01] | [11.612492561340332, -0.4409361779689789] |
3d4ef8a1-1e9d-4e25-a640-bf4611186d18 | equivariant-multi-view-networks | 1904.00993 | null | https://arxiv.org/abs/1904.00993v2 | https://arxiv.org/pdf/1904.00993v2.pdf | Equivariant Multi-View Networks | Several popular approaches to 3D vision tasks process multiple views of the input independently with deep neural networks pre-trained on natural images, achieving view permutation invariance through a single round of pooling over all views. We argue that this operation discards important information and leads to subpar global descriptors. In this paper, we propose a group convolutional approach to multiple view aggregation where convolutions are performed over a discrete subgroup of the rotation group, enabling, thus, joint reasoning over all views in an equivariant (instead of invariant) fashion, up to the very last layer. We further develop this idea to operate on smaller discrete homogeneous spaces of the rotation group, where a polar view representation is used to maintain equivariance with only a fraction of the number of input views. We set the new state of the art in several large scale 3D shape retrieval tasks, and show additional applications to panoramic scene classification. | ['Christine Allen-Blanchette', 'Kostas Daniilidis', 'Yinshuang Xu', 'Carlos Esteves'] | 2019-04-01 | equivariant-multi-view-networks-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Esteves_Equivariant_Multi-View_Networks_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Esteves_Equivariant_Multi-View_Networks_ICCV_2019_paper.pdf | iccv-2019-10 | ['3d-shape-retrieval'] | ['computer-vision'] | [ 3.13890755e-01 1.21636055e-02 1.46650240e-01 -3.92226517e-01
-4.96097267e-01 -1.12056541e+00 1.13952529e+00 -2.62477348e-04
-5.05124390e-01 -4.50623222e-02 3.46605659e-01 -1.56757161e-01
-1.97261441e-02 -8.45417857e-01 -8.62763464e-01 -6.97357595e-01
-4.40750532e-02 4.34903383e-01 1.53356403e-01 -1.61720797e-01
2.40823612e-01 1.18841636e+00 -1.55703938e+00 4.83898669e-01
4.21422161e-02 7.94088960e-01 -2.81336874e-01 9.55160856e-01
2.57292330e-01 5.97500682e-01 -4.41590577e-01 -3.09251189e-01
7.82120526e-01 3.57795879e-02 -8.93318534e-01 3.73504132e-01
1.19737220e+00 -6.06042206e-01 -5.32492280e-01 9.98211205e-01
5.12966394e-01 1.89175367e-01 6.67467475e-01 -9.68353331e-01
-8.79832447e-01 2.17322931e-01 -5.35243213e-01 3.83901373e-02
2.67818302e-01 -1.24325082e-01 1.29762638e+00 -1.08140564e+00
9.50853467e-01 1.54494417e+00 3.01693767e-01 2.76267797e-01
-1.31168735e+00 -7.94948786e-02 2.43947729e-01 2.09882241e-02
-1.04208255e+00 -5.13576627e-01 8.13297629e-01 -1.84364647e-01
1.34803367e+00 3.27653438e-01 7.63931870e-01 6.59591556e-01
2.95191020e-01 5.15756547e-01 1.08993053e+00 -2.86270738e-01
8.42717364e-02 -3.54198009e-01 1.14250287e-01 9.68492210e-01
3.78967375e-01 -3.81365418e-01 -5.03789365e-01 -9.77909341e-02
1.04133415e+00 2.70801693e-01 -1.12833075e-01 -1.16489244e+00
-1.46285331e+00 7.34919369e-01 6.24047756e-01 4.43092883e-02
-2.79170811e-01 2.19807290e-02 2.65036196e-01 6.27686918e-01
5.53488553e-01 3.61532360e-01 -3.26898426e-01 5.12018323e-01
-6.44728541e-01 1.74013153e-01 8.36782515e-01 7.59264648e-01
8.05927634e-01 -2.91818261e-01 -8.37184489e-03 6.25963330e-01
1.62371360e-02 6.58912122e-01 -6.39163703e-02 -1.11346555e+00
3.07495981e-01 5.95181227e-01 -1.67865664e-01 -8.44888628e-01
-3.83697420e-01 -2.26532102e-01 -9.67358351e-01 6.97679162e-01
4.43954498e-01 2.34103695e-01 -1.21124196e+00 1.92181611e+00
1.49488807e-01 -4.34925765e-01 1.01231821e-01 7.23812580e-01
5.14728665e-01 2.89991945e-01 -3.68409961e-01 1.82410523e-01
1.55258226e+00 -7.22222328e-01 6.36035725e-02 2.56160200e-02
1.09893583e-01 -8.30073357e-01 7.23782480e-01 3.92052114e-01
-1.45254230e+00 -5.07693827e-01 -1.24186826e+00 -4.49857295e-01
-5.84346592e-01 9.87095311e-02 5.97667813e-01 3.85493577e-01
-1.54314351e+00 4.53411281e-01 -7.54664242e-01 -6.84491277e-01
5.97421229e-01 5.49438000e-01 -1.01009977e+00 -2.92137891e-01
-7.15467751e-01 9.32529569e-01 2.93941703e-02 -1.30825460e-01
-7.72226572e-01 -5.50539672e-01 -1.00839782e+00 2.52269536e-01
1.53435543e-01 -1.07911885e+00 9.36363995e-01 -6.37675405e-01
-1.28739727e+00 1.37122703e+00 -1.20697729e-01 -3.10261041e-01
4.65063334e-01 -1.98001638e-01 1.51873857e-01 4.82123941e-01
-5.91313094e-02 7.75301754e-01 1.17533100e+00 -1.14082217e+00
-2.02696249e-01 -8.10738742e-01 6.65685713e-01 5.26472330e-01
-2.64717638e-02 -1.32991478e-01 -5.44315457e-01 -3.55375290e-01
6.32626414e-01 -1.07406890e+00 -1.38107315e-01 1.87857270e-01
-1.47996813e-01 -2.54075706e-01 7.10166574e-01 -4.26894695e-01
4.13522720e-01 -2.02144361e+00 5.93571484e-01 4.11159337e-01
5.54856539e-01 -6.36733100e-02 -3.20025682e-01 3.26543629e-01
-3.21441680e-01 -3.72290872e-02 -1.31558597e-01 -4.03248698e-01
3.50247733e-02 1.55427635e-01 -4.74738121e-01 8.66791964e-01
3.76320213e-01 1.00715017e+00 -6.12562776e-01 2.42784433e-02
5.34916043e-01 4.80758995e-01 -7.59230971e-01 -3.28106177e-03
4.85598929e-02 1.59612060e-01 -1.92532644e-01 4.95152175e-01
9.84835744e-01 -2.69133240e-01 3.38523611e-02 -4.29804325e-01
-7.59311914e-02 3.18305254e-01 -1.26401854e+00 2.00183964e+00
-5.55715919e-01 5.38429201e-01 1.19967967e-01 -1.07964492e+00
6.77808523e-01 3.88632715e-02 7.02658892e-01 -4.50081378e-01
-2.34446768e-02 9.18653011e-02 -2.08075762e-01 3.28669809e-02
4.68412757e-01 -2.44842544e-02 -2.69903034e-01 7.53134966e-01
6.15094066e-01 -4.91858095e-01 -4.08461951e-02 2.25174591e-01
9.13569450e-01 2.49916352e-02 4.86953080e-01 -3.09146017e-01
6.74886763e-01 -4.82293844e-01 -6.52912110e-02 9.95935142e-01
8.13143924e-02 8.89335155e-01 6.58017397e-01 -7.73167610e-01
-1.43070447e+00 -1.40776289e+00 2.11935434e-02 8.70848358e-01
7.90309310e-02 -2.83861905e-01 -6.29269302e-01 -9.07763898e-01
-7.58684066e-04 1.71735585e-01 -8.15518498e-01 -9.99845490e-02
-7.25345314e-01 -3.65139395e-01 3.56265098e-01 3.62456530e-01
6.87965214e-01 -7.76292145e-01 -7.31322646e-01 -1.26846895e-01
3.73522520e-01 -1.13794208e+00 -3.69923502e-01 4.10386980e-01
-7.45718896e-01 -1.15367913e+00 -8.37439060e-01 -6.16642416e-01
8.46740484e-01 8.26042116e-01 1.06381595e+00 -2.95117348e-01
-6.62932396e-02 6.83900416e-01 1.85713973e-02 -3.76900375e-01
-1.33521259e-01 7.06326663e-02 5.51130883e-02 4.54535559e-02
3.84736955e-01 -7.33623564e-01 -5.06657839e-01 9.58648622e-02
-1.09168017e+00 9.39818546e-02 5.21137953e-01 7.47550070e-01
5.32671034e-01 -3.72546911e-01 -1.92881703e-01 -7.40611613e-01
1.81052998e-01 1.10265419e-01 -7.92830944e-01 4.58582848e-01
1.34477884e-01 4.74675417e-01 5.76511383e-01 -1.70683414e-01
-8.50956738e-01 1.10649757e-01 1.13939293e-01 -7.59811878e-01
-2.83371001e-01 -4.54511531e-02 -3.31776947e-01 -2.90274650e-01
4.95227337e-01 1.39921635e-01 1.54573694e-01 -3.60210150e-01
7.72676229e-01 1.67424068e-01 4.74776208e-01 -2.87418038e-01
9.52892840e-01 1.13479078e+00 5.91437399e-01 -7.65840054e-01
-8.54286730e-01 -2.33065084e-01 -1.04269755e+00 6.04945831e-02
1.11026955e+00 -1.03875887e+00 -7.19449818e-01 6.53456032e-01
-1.20476758e+00 5.53820394e-02 -4.51909214e-01 5.24152994e-01
-8.94859791e-01 3.78670752e-01 -3.60111803e-01 -2.85619110e-01
-2.73317546e-01 -1.09532952e+00 1.48594820e+00 2.01814342e-02
-4.83618192e-02 -6.48865163e-01 -2.01853197e-02 8.92705098e-03
2.72724181e-01 3.57806236e-02 1.18488061e+00 -7.74278224e-01
-8.44037652e-01 -8.97575021e-02 -4.19858575e-01 4.13468987e-01
9.37356055e-02 -2.26570129e-01 -1.19208801e+00 -4.71927255e-01
2.78499871e-01 -1.93244681e-01 1.33015251e+00 3.00539374e-01
1.19216585e+00 3.19655053e-02 9.68966484e-02 8.07009101e-01
1.36726308e+00 -1.48861960e-01 5.31409085e-01 6.01969697e-02
8.81998181e-01 4.87847179e-01 2.95663979e-02 1.88710615e-01
1.31015360e-01 3.56787771e-01 4.80091542e-01 -2.54285008e-01
-2.27716461e-01 -7.06789419e-02 1.86363965e-01 5.98641634e-01
-4.16465253e-01 -2.06458420e-01 -5.54554522e-01 3.57525587e-01
-1.52369845e+00 -1.07369781e+00 4.34188515e-01 2.53954220e+00
9.96177346e-02 1.43796980e-01 -1.57306001e-01 -1.91506177e-01
4.65658963e-01 7.98826456e-01 -4.33369964e-01 -5.56216002e-01
-4.76288885e-01 6.68673873e-01 8.09079528e-01 5.92348754e-01
-1.47783220e+00 8.68634582e-01 6.46268368e+00 3.93862188e-01
-1.31221163e+00 -1.24725856e-01 3.42052102e-01 -5.27457260e-02
-4.83550370e-01 7.28501827e-02 -4.27557021e-01 -3.61925513e-01
1.70079395e-01 2.01507822e-01 4.88751680e-01 5.63223064e-01
-4.87776905e-01 4.23679538e-02 -1.36452317e+00 9.57894981e-01
5.29819906e-01 -1.35118949e+00 5.82351029e-01 2.80407667e-01
9.23468769e-01 3.54840040e-01 6.50201812e-02 -2.35498309e-01
3.85836452e-01 -8.02976906e-01 6.23389602e-01 5.96290231e-01
7.68067300e-01 -7.38067865e-01 3.71299833e-01 -6.17476553e-02
-1.17145598e+00 -1.30131990e-02 -5.99641383e-01 1.30927101e-01
-3.10177542e-02 1.33100212e-01 -4.99453276e-01 7.95203209e-01
6.58925235e-01 9.03344214e-01 -6.68825567e-01 6.12035096e-01
-6.03070594e-02 -3.28195542e-01 -5.65427780e-01 2.96861708e-01
2.62652218e-01 -1.86982512e-01 9.40421164e-01 7.56515563e-01
4.16671753e-01 -2.78067719e-02 -1.19485013e-01 6.10814214e-01
-3.47822249e-01 -2.33137012e-01 -1.17941070e+00 3.68599236e-01
-8.37271065e-02 1.20537519e+00 -7.32450902e-01 -4.58486676e-01
-7.42834210e-01 1.17530048e+00 3.30818981e-01 4.08202708e-01
-2.40549445e-01 -2.83090800e-01 8.77497613e-01 -1.95259348e-01
7.43466258e-01 -6.22106016e-01 -1.02443255e-01 -1.57134831e+00
5.03944792e-02 -8.46710145e-01 3.89943480e-01 -7.92336047e-01
-1.16124606e+00 2.65477747e-01 1.70499220e-01 -1.31606674e+00
-2.19570622e-01 -9.27549124e-01 -6.41297698e-01 9.66491699e-01
-1.42899060e+00 -1.34202540e+00 -1.49912030e-01 7.77154326e-01
3.50540966e-01 -2.88582563e-01 9.35244381e-01 -1.11803085e-01
2.41753250e-01 2.46980429e-01 -1.49552763e-01 5.47194146e-02
8.33183885e-01 -1.34207702e+00 7.71126807e-01 8.00787270e-01
4.38819915e-01 8.09733570e-01 3.05164695e-01 -9.99560505e-02
-1.68580365e+00 -7.68692255e-01 9.42605793e-01 -5.86615741e-01
3.81195992e-01 -6.27900064e-01 -6.94112301e-01 7.10258126e-01
4.29134607e-01 9.15864632e-02 3.65780532e-01 -2.82958876e-02
-8.35179389e-01 -1.62994385e-01 -9.21357691e-01 6.95294499e-01
1.16679180e+00 -9.54361379e-01 -8.01620126e-01 5.35802804e-02
5.08146107e-01 -4.14312571e-01 -8.11902344e-01 4.77421403e-01
7.77005196e-01 -1.11384559e+00 1.38656950e+00 -7.44425178e-01
2.62503177e-01 -4.35429215e-01 -5.20969808e-01 -1.16658068e+00
-3.16651881e-01 -3.43757659e-01 7.93200508e-02 6.70480072e-01
1.09764703e-01 -5.99119246e-01 5.22010446e-01 1.44914404e-01
1.60859060e-02 -5.11207104e-01 -9.24237967e-01 -2.89535850e-01
7.32415766e-02 1.45015977e-02 5.31823099e-01 7.20192909e-01
-5.69188833e-01 2.98870802e-01 -1.55899972e-01 3.63399267e-01
7.01905131e-01 4.32148993e-01 9.40167129e-01 -1.38662863e+00
-3.00610542e-01 -4.94476706e-01 -6.64066017e-01 -9.77657974e-01
1.79481804e-01 -1.08448064e+00 -3.94807905e-01 -1.33362567e+00
3.61703306e-01 2.45978594e-01 -3.61619949e-01 4.55766350e-01
2.00550184e-01 5.78724444e-01 3.72207344e-01 9.90119278e-02
-5.71160257e-01 1.90405771e-01 1.44207144e+00 -2.28543684e-01
-1.54381944e-02 -1.12032689e-01 -5.93132555e-01 7.86267638e-01
5.15191734e-01 -2.18809061e-02 -3.22180927e-01 -7.34490573e-01
2.46874988e-01 -2.12255552e-01 8.40949893e-01 -8.12662482e-01
1.40200034e-01 9.07334909e-02 8.78008723e-01 -6.37329459e-01
5.25022209e-01 -7.73430824e-01 -2.21176725e-02 2.27248058e-01
-3.14149439e-01 2.48123378e-01 4.11613956e-02 3.88228357e-01
-1.78913847e-01 1.85713485e-01 8.85645628e-01 -4.76253062e-01
-6.02089107e-01 5.40150881e-01 -1.02235407e-01 -3.68563116e-01
6.69591427e-01 -1.01039685e-01 -3.44261348e-01 -4.69538212e-01
-8.40255320e-01 -1.47840530e-01 8.38669717e-01 5.07137716e-01
4.37792093e-01 -1.38122296e+00 -4.61343646e-01 7.71461546e-01
1.38289377e-01 8.62771180e-03 3.11658233e-01 5.70894420e-01
-6.22789919e-01 6.00882709e-01 -7.09579051e-01 -7.56412387e-01
-1.51186645e+00 4.90235925e-01 3.81442040e-01 -2.63595760e-01
-5.95543742e-01 5.74263453e-01 5.04406273e-01 -5.63205361e-01
-4.76569645e-02 -3.64451021e-01 -2.19422221e-01 3.22212875e-01
2.86060661e-01 3.44366431e-02 3.32652956e-01 -8.73424411e-01
-4.31408793e-01 1.08546066e+00 -1.18405990e-01 -2.48501614e-01
1.55962455e+00 -9.22595337e-02 -5.08228183e-01 2.03855067e-01
1.62383103e+00 2.29578987e-01 -1.26477456e+00 -3.52059275e-01
-4.79073435e-01 -4.23847795e-01 -1.54259786e-01 -2.25537449e-01
-8.90754879e-01 1.09372318e+00 4.43177611e-01 3.14532012e-01
1.11834347e+00 3.64385307e-01 1.35327056e-01 9.00110483e-01
2.26303950e-01 -7.00388014e-01 -4.39269319e-02 8.25947225e-01
1.00888431e+00 -1.15197694e+00 1.78532958e-01 -1.27068833e-01
-2.68878549e-01 1.32643008e+00 2.34763563e-01 -7.27221727e-01
7.46041298e-01 -1.60282224e-01 -9.01636407e-02 -3.64501774e-01
-5.60603559e-01 -3.45277697e-01 5.44289470e-01 4.27451968e-01
2.12514967e-01 -5.22574969e-02 8.95682573e-02 -1.05576769e-01
-1.40065968e-01 -6.05788112e-01 2.30830744e-01 8.20372105e-01
-3.70389938e-01 -1.25651121e+00 -2.54343420e-01 3.47240508e-01
-5.18438399e-01 -7.41854683e-02 -7.77793705e-01 8.61889601e-01
-2.28982512e-02 3.57809007e-01 3.74967515e-01 -1.15506612e-02
4.95551020e-01 2.42935807e-01 9.59888637e-01 -5.10300159e-01
-2.28018299e-01 1.51130453e-01 -2.06745625e-01 -7.06973970e-01
-7.36979187e-01 -8.05201113e-01 -5.68281114e-01 -1.64348200e-01
1.62549660e-01 -3.86783749e-01 5.88950932e-01 7.84888446e-01
3.76172423e-01 2.78264225e-01 9.20918286e-01 -1.19392550e+00
-6.07526541e-01 -5.89231968e-01 -5.13712049e-01 6.18565679e-01
7.34479547e-01 -4.84592617e-01 -3.00366312e-01 -1.13089137e-01] | [8.861673355102539, 2.336280584335327] |
6492b47b-eec9-4694-a408-390c38dfe6b2 | a-self-attention-joint-model-for-spoken | 1905.11393 | null | https://arxiv.org/abs/1905.11393v1 | https://arxiv.org/pdf/1905.11393v1.pdf | A Self-Attention Joint Model for Spoken Language Understanding in Situational Dialog Applications | Spoken language understanding (SLU) acts as a critical component in goal-oriented dialog systems. It typically involves identifying the speakers intent and extracting semantic slots from user utterances, which are known as intent detection (ID) and slot filling (SF). SLU problem has been intensively investigated in recent years. However, these methods just constrain SF results grammatically, solve ID and SF independently, or do not fully utilize the mutual impact of the two tasks. This paper proposes a multi-head self-attention joint model with a conditional random field (CRF) layer and a prior mask. The experiments show the effectiveness of our model, as compared with state-of-the-art models. Meanwhile, online education in China has made great progress in the last few years. But there are few intelligent educational dialog applications for students to learn foreign languages. Hence, we design an intelligent dialog robot equipped with different scenario settings to help students learn communication skills. | ['Mengyang Chen', 'Jin Zeng', 'Jie Lou'] | 2019-05-27 | null | null | null | null | ['goal-oriented-dialog'] | ['natural-language-processing'] | [-1.38607189e-01 3.63317132e-01 -1.23863697e-01 -6.56592190e-01
-3.22156847e-01 -1.87243953e-01 5.22984266e-01 -1.50709718e-01
-3.59541804e-01 7.01200664e-01 4.38422561e-01 -4.27233070e-01
2.02251121e-01 -6.58868492e-01 -6.52369857e-02 -5.89869499e-01
5.98194063e-01 5.09660721e-01 3.19834441e-01 -5.12634933e-01
1.93065301e-01 -4.65477966e-02 -1.38846493e+00 -3.29998672e-01
1.47841883e+00 5.07937253e-01 9.86404657e-01 1.94071606e-01
-9.34004009e-01 7.15713084e-01 -6.31558657e-01 -5.86592257e-02
-5.69336474e-01 -5.58435798e-01 -9.77744401e-01 4.41871375e-01
-3.45849872e-01 -5.54533482e-01 -2.34198600e-01 1.21829998e+00
4.18826520e-01 3.74496549e-01 3.25059414e-01 -1.18991637e+00
-6.55169964e-01 8.93519640e-01 -2.84231901e-01 -2.29345948e-01
5.23805797e-01 -1.35211065e-01 7.49084234e-01 -3.31601650e-01
2.75654167e-01 1.59099638e+00 2.86045857e-02 9.48870301e-01
-6.12294853e-01 -7.61132896e-01 5.42173266e-01 1.80163532e-01
-1.12825286e+00 -2.30527222e-01 8.22762847e-01 -3.64142358e-01
6.40261233e-01 -1.23887330e-01 2.66665161e-01 7.09842622e-01
-3.61999422e-02 1.11987138e+00 9.26022172e-01 -5.00393629e-01
8.47108066e-02 5.32909036e-01 5.85009813e-01 7.57773221e-01
-3.68399650e-01 -6.38131261e-01 -3.24250698e-01 2.15806037e-01
6.28331542e-01 2.69052267e-01 -2.13168979e-01 -6.62323646e-03
-9.44493413e-01 1.02330756e+00 2.54297405e-02 4.57382590e-01
-6.71620890e-02 -6.28712893e-01 8.39148313e-02 1.08673133e-01
4.15992856e-01 -6.80685788e-03 -5.41738391e-01 -2.67389059e-01
-1.81180269e-01 -5.74196130e-02 9.54357982e-01 1.31698954e+00
6.94818735e-01 -4.78982776e-02 -1.67610213e-01 1.24629819e+00
9.02396023e-01 2.60403663e-01 8.69069278e-01 -5.95907092e-01
3.58896583e-01 8.54528129e-01 2.46227793e-02 -4.04290110e-01
-3.39785725e-01 1.82354107e-01 -5.57750106e-01 -3.15609723e-01
2.74371833e-01 -6.25418842e-01 -7.19267368e-01 1.92905331e+00
6.37246192e-01 1.91013917e-01 3.56228650e-01 9.66349125e-01
1.43222153e+00 7.73241162e-01 6.29755080e-01 -3.36203605e-01
1.71126521e+00 -1.12796080e+00 -1.33524907e+00 -4.45918620e-01
6.00256324e-01 -9.96608794e-01 1.19756603e+00 2.74752766e-01
-8.14293981e-01 -6.00780845e-01 -8.52420509e-01 -2.30087206e-01
-3.19533855e-01 2.46598855e-01 8.22755992e-01 7.32931733e-01
-7.07677722e-01 -6.32656738e-02 -7.31947005e-01 -4.75319952e-01
-8.37930739e-02 4.21507299e-01 -1.21766977e-01 -2.57850513e-02
-1.53645003e+00 7.92025805e-01 2.51041561e-01 -6.13056645e-02
-6.25685751e-01 -4.98197265e-02 -1.13204765e+00 2.23739728e-01
6.09291732e-01 -4.66960333e-02 1.64853013e+00 -3.53442460e-01
-2.21305156e+00 6.19991899e-01 -4.39404726e-01 2.14312542e-02
1.20568968e-01 -3.49310338e-01 -3.64522636e-01 -3.04931313e-01
2.22316280e-01 7.31760859e-01 1.53321907e-01 -7.68380702e-01
-6.80943370e-01 -4.09252912e-01 3.08334064e-02 7.16334641e-01
-3.39578241e-01 2.50109076e-01 -6.36545300e-01 -1.82761088e-01
5.41950226e-01 -6.80242479e-01 -3.63763094e-01 -5.97531736e-01
-2.91557103e-01 -1.01043642e+00 9.40899253e-01 -7.87763476e-01
1.33448577e+00 -1.94899106e+00 -3.68646532e-02 -3.85554284e-01
6.54375507e-03 2.61226654e-01 1.63663447e-01 3.68630946e-01
3.59029204e-01 -9.14284065e-02 -1.00505777e-01 -4.66584206e-01
2.50666469e-01 4.71946925e-01 -2.81730503e-01 -5.49901687e-02
-2.17185840e-01 3.62231821e-01 -8.13532174e-01 -7.84294546e-01
4.44766849e-01 1.02081977e-01 -3.14288408e-01 8.11785221e-01
-4.44712460e-01 7.39984632e-01 -1.01742721e+00 4.83726531e-01
6.96731746e-01 -5.10398299e-02 2.96558052e-01 5.51177919e-01
-5.08679032e-01 7.03403354e-01 -1.13934827e+00 1.90707076e+00
-5.90567470e-01 5.84946461e-02 6.12650454e-01 -9.11354005e-01
1.21267056e+00 6.57613873e-01 1.51487216e-01 -6.35305464e-01
2.90187538e-01 -1.37697542e-02 -3.93749811e-02 -6.40618384e-01
4.63360369e-01 5.14585152e-03 -2.41000026e-01 4.77872610e-01
4.35277104e-01 -6.41277656e-02 -1.29288342e-02 2.58833230e-01
3.89495999e-01 1.18730858e-01 3.12866539e-01 -2.58635104e-01
9.56145227e-01 -2.35370338e-01 7.00751841e-01 4.53098446e-01
-4.77131307e-01 2.64026612e-01 3.44259858e-01 3.64994630e-02
-2.65831053e-01 -7.10932016e-01 -5.47982827e-02 1.29409969e+00
6.78938806e-01 -2.53894597e-01 -9.91143167e-01 -7.79581785e-01
-4.29986119e-01 8.63376141e-01 9.08436924e-02 -4.71657701e-02
-5.24903297e-01 -3.04750204e-01 9.00763497e-02 1.42630473e-01
1.09108031e+00 -1.39153349e+00 -1.17865734e-01 5.08641243e-01
-2.86541164e-01 -1.27326584e+00 -5.44486880e-01 2.09103152e-03
-4.42195714e-01 -7.70267963e-01 -6.09320819e-01 -1.29157770e+00
4.63798404e-01 4.72253203e-01 8.02440047e-01 1.30045563e-01
1.20055251e-01 1.88762665e-01 -3.92596126e-01 -4.81273234e-01
-3.30624849e-01 1.76993385e-01 1.67845041e-01 -3.40157241e-01
8.05747449e-01 -4.56327140e-01 -3.11532319e-01 6.03486121e-01
-6.74396217e-01 3.57850075e-01 5.48019946e-01 6.80244744e-01
1.44824677e-03 5.21980633e-04 8.55974376e-01 -8.52967024e-01
7.60144234e-01 -5.22436202e-01 -6.47588432e-01 4.71542031e-01
-5.44026554e-01 -2.87394449e-02 4.74427789e-01 -2.14990661e-01
-1.76677299e+00 8.51592720e-02 -6.10683680e-01 1.44299259e-02
-9.39691246e-01 3.57423246e-01 -9.41005349e-01 1.90732807e-01
-2.68605649e-01 3.89723539e-01 1.24941409e-01 -6.77669883e-01
1.62809212e-02 1.31508958e+00 2.88659513e-01 -6.91195369e-01
4.21663858e-02 -1.90975890e-01 -4.90690172e-01 -1.04222178e+00
-9.49242532e-01 -7.12614775e-01 -6.07345939e-01 -2.00365096e-01
1.21546626e+00 -1.07432032e+00 -1.12566066e+00 6.67055547e-01
-1.20062959e+00 -3.22829187e-01 4.99111623e-01 7.75270939e-01
-1.45070001e-01 4.53368574e-01 -8.50950658e-01 -1.08342969e+00
-1.50249809e-01 -1.54364657e+00 8.58771920e-01 1.15687239e+00
8.95690694e-02 -9.97663140e-01 -1.25922814e-01 7.52832711e-01
2.22610429e-01 -7.18597591e-01 7.31999576e-01 -9.54085112e-01
-5.11776328e-01 3.56396049e-01 -1.20700374e-01 8.14701468e-02
3.12992752e-01 -5.68574905e-01 -1.06268394e+00 2.84409046e-01
2.91261792e-01 -3.92797112e-01 4.21819359e-01 2.70550907e-01
8.48368645e-01 -3.96566063e-01 -3.44601601e-01 -7.22372457e-02
7.81670868e-01 7.93455660e-01 2.97099978e-01 1.56150624e-01
6.79343402e-01 9.45790887e-01 9.62129831e-01 3.46639097e-01
8.45392704e-01 6.51250362e-01 -5.96457161e-02 2.52457798e-01
2.23149851e-01 -4.16843355e-01 2.80068457e-01 1.14743078e+00
6.33108377e-01 -1.91245630e-01 -9.95473146e-01 4.87214744e-01
-1.94512784e+00 -4.47867513e-01 -8.08948725e-02 1.81496859e+00
1.17645860e+00 -2.97957212e-02 -2.70187944e-01 -3.80409151e-01
8.90201092e-01 1.74090639e-01 -3.20545286e-01 -2.54212976e-01
3.13274294e-01 -9.08555239e-02 5.67846894e-02 7.48848855e-01
-1.20104611e+00 1.39667344e+00 4.68958235e+00 8.55597377e-01
-8.69386673e-01 9.00078341e-02 5.72881460e-01 8.28261614e-01
-2.24915743e-01 1.00660861e-01 -1.31238890e+00 5.52153826e-01
6.07899666e-01 9.92838889e-02 1.95820108e-01 1.09001637e+00
1.26440212e-01 -4.01140153e-01 -5.68009794e-01 7.77178049e-01
-3.26806679e-02 -6.23070121e-01 -3.03066701e-01 -4.14398238e-02
3.94222289e-01 -4.51493680e-01 -3.23544651e-01 8.38626266e-01
7.68781722e-01 -8.45576763e-01 7.10472390e-02 2.89860517e-01
2.23630011e-01 -6.08870804e-01 7.66584814e-01 8.03925931e-01
-1.24882698e+00 2.17821777e-01 -4.38384295e-01 -3.78254116e-01
3.41636449e-01 3.45105290e-01 -1.01628315e+00 3.77143353e-01
5.21450639e-01 4.24190968e-01 -7.17489496e-02 7.12447166e-01
-5.86230576e-01 9.11334634e-01 -2.69376367e-01 -4.83744383e-01
3.06091279e-01 -7.79502749e-01 3.30240548e-01 7.10179806e-01
1.61356166e-01 6.19129658e-01 8.11485589e-01 7.73030996e-01
1.07151851e-01 3.16714317e-01 -2.71097422e-01 1.19605213e-02
6.35050416e-01 1.17005086e+00 -7.50253022e-01 -2.03913987e-01
-7.01819539e-01 8.73555660e-01 1.36515394e-01 2.25700691e-01
-4.20448899e-01 -3.77036005e-01 6.83945596e-01 -4.99262303e-01
-5.65062985e-02 -3.71266782e-01 -3.42325807e-01 -1.14097905e+00
-4.18656647e-01 -5.14959931e-01 2.79678553e-01 -6.18385613e-01
-9.63152647e-01 1.91023186e-01 -7.68353930e-03 -9.17785347e-01
-2.86630481e-01 -4.01701212e-01 -1.01645494e+00 1.07671046e+00
-1.43950260e+00 -9.23793137e-01 -3.52185965e-01 4.09675866e-01
1.13909650e+00 -1.03565753e-01 9.76889074e-01 1.10465124e-01
-1.06445765e+00 2.31192142e-01 -1.90065667e-01 2.42284030e-01
7.15953827e-01 -1.23821032e+00 2.88039558e-02 5.90386152e-01
-2.12832749e-01 6.55965924e-01 6.13104701e-01 -6.77152455e-01
-9.12675798e-01 -4.62939531e-01 1.22099388e+00 1.69002235e-01
4.40670431e-01 -3.29604208e-01 -1.01080871e+00 7.06203938e-01
6.60154819e-01 -6.78137004e-01 7.15094030e-01 2.84506768e-01
4.00567591e-01 2.95931280e-01 -9.19994652e-01 6.08288884e-01
7.51164794e-01 -2.75060743e-01 -8.75624001e-01 2.99941421e-01
1.12496567e+00 -6.24307990e-01 -5.67349553e-01 2.26337940e-01
-9.54966899e-03 -7.60308146e-01 5.74202955e-01 -2.86331028e-01
1.41505063e-01 -9.20308381e-02 9.67404246e-02 -1.13916159e+00
1.06011078e-01 -4.99842405e-01 3.15679610e-01 1.89995348e+00
4.65615019e-02 -6.58500373e-01 8.78669858e-01 1.11348498e+00
-5.25761485e-01 -4.49987203e-01 -5.62435687e-01 -1.30451649e-01
-1.44024298e-01 -1.15436792e-01 7.44113564e-01 1.01300061e+00
4.70649928e-01 8.39245200e-01 -2.98389882e-01 1.62849054e-01
2.27415517e-01 2.78282732e-01 7.15992689e-01 -1.36338890e+00
1.34379417e-01 -3.92562300e-01 2.87914723e-01 -1.69479454e+00
5.87049544e-01 -1.42176837e-01 3.66175950e-01 -1.63810182e+00
-4.20693159e-02 -6.38386726e-01 9.60727185e-02 6.14366591e-01
-6.16004288e-01 -8.64640653e-01 -1.16085283e-01 9.62681621e-02
-7.82820284e-01 1.06398320e+00 1.34472382e+00 1.08852061e-02
-5.73849797e-01 3.82595479e-01 -5.86361468e-01 8.15548003e-01
9.10988033e-01 7.36163696e-03 -5.72435915e-01 -1.47278130e-01
-6.96779430e-01 6.61647439e-01 -3.21943760e-01 -6.98669970e-01
5.75290084e-01 -4.71932471e-01 -2.47621536e-02 -6.81052625e-01
1.78834990e-01 -7.68292010e-01 -6.61096156e-01 2.40927845e-01
-4.01258498e-01 -2.59918541e-01 4.13220637e-02 1.52804449e-01
-4.62395549e-01 -4.51202303e-01 5.40589035e-01 -2.74074614e-01
-1.13035798e+00 1.34335920e-01 -5.61133564e-01 -1.92295060e-01
1.02744055e+00 2.54534602e-01 -2.22704843e-01 -6.34740710e-01
-5.99792957e-01 1.03238630e+00 -1.00978971e-01 6.43460572e-01
5.20128846e-01 -8.73820066e-01 -3.54355186e-01 3.51678073e-01
-1.64788634e-01 5.95901906e-01 8.40904176e-01 5.41161954e-01
-2.67761588e-01 8.15251589e-01 -4.53598760e-02 -4.76924837e-01
-1.33753741e+00 1.77286521e-01 -9.27927345e-03 -4.45293963e-01
-2.77910829e-01 9.25874412e-01 4.92404670e-01 -1.09427249e+00
6.91051781e-01 4.75621894e-02 -1.04796708e+00 -8.09407532e-02
4.62406427e-01 7.72450166e-03 -2.84657061e-01 -5.74691474e-01
-1.25889182e-01 8.45132619e-02 -2.04891503e-01 -1.75975978e-01
8.19466472e-01 -6.27037168e-01 5.07532395e-02 2.98737139e-01
6.97163522e-01 -2.38858126e-02 -1.22399402e+00 -4.51478302e-01
8.96174274e-03 6.86851740e-02 -3.92595530e-02 -6.99317336e-01
-7.06010044e-01 9.87416923e-01 3.31838280e-01 1.45918146e-01
7.34943509e-01 1.03826314e-01 9.75296557e-01 4.26586300e-01
3.92697364e-01 -1.10057366e+00 -1.59985498e-01 9.45067167e-01
3.88713032e-01 -1.55810750e+00 -3.72142524e-01 -7.84850597e-01
-8.88921142e-01 8.51854682e-01 1.45766914e+00 4.99619722e-01
7.94995427e-01 7.09286854e-02 3.66197586e-01 1.74621627e-01
-5.17023206e-01 -4.20140803e-01 7.70963058e-02 3.41137826e-01
6.62237227e-01 1.19953312e-01 -6.87843204e-01 1.05207622e+00
-3.52687985e-01 -2.84589082e-01 4.18037295e-01 1.14392543e+00
-8.98614049e-01 -1.47470713e+00 -4.65656728e-01 8.40848684e-02
-3.37872565e-01 -1.01741180e-01 -2.29688659e-01 6.24225378e-01
-1.62874818e-01 1.41695797e+00 -1.07865520e-01 -3.87362301e-01
8.59118924e-02 4.06551093e-01 -1.89266682e-01 -8.41447592e-01
-2.95785636e-01 2.43240312e-01 2.22169571e-02 -1.54832751e-01
-3.31670523e-01 -5.06291747e-01 -1.77566469e+00 1.28694341e-01
-5.22153497e-01 7.82075405e-01 7.16547251e-01 1.34528637e+00
-5.53168468e-02 4.58733559e-01 7.12974489e-01 -4.71859962e-01
-4.14925873e-01 -1.41146946e+00 -6.79778874e-01 -8.61993581e-02
-1.24717012e-01 -7.39781380e-01 -1.62497550e-01 -2.62654185e-01] | [12.735067367553711, 7.631462097167969] |
c6a41dfd-2494-40bf-a57c-22c8d0739723 | scaling-scaling-laws-with-board-games | 2104.03113 | null | https://arxiv.org/abs/2104.03113v2 | https://arxiv.org/pdf/2104.03113v2.pdf | Scaling Scaling Laws with Board Games | The largest experiments in machine learning now require resources far beyond the budget of all but a few institutions. Fortunately, it has recently been shown that the results of these huge experiments can often be extrapolated from the results of a sequence of far smaller, cheaper experiments. In this work, we show that not only can the extrapolation be done based on the size of the model, but on the size of the problem as well. By conducting a sequence of experiments using AlphaZero and Hex, we show that the performance achievable with a fixed amount of compute degrades predictably as the game gets larger and harder. Along with our main result, we further show that the test-time and train-time compute available to an agent can be traded off while maintaining performance. | ['Andy L. Jones'] | 2021-04-07 | null | null | null | null | ['board-games'] | ['playing-games'] | [-3.33644271e-01 1.72378331e-01 1.82721972e-01 -2.14937791e-01
-7.35981762e-01 -7.46662855e-01 4.27852839e-01 3.16501319e-01
-8.41710329e-01 8.63773763e-01 -2.44137168e-01 -5.20155370e-01
2.06773635e-02 -5.89829922e-01 -7.31873512e-01 -5.42427182e-01
-2.50525385e-01 7.81502187e-01 3.62363607e-01 -8.30683485e-02
2.19739959e-01 4.36969459e-01 -1.41389668e+00 2.44470015e-02
3.96874279e-01 8.52393687e-01 -3.26704420e-02 1.01433253e+00
4.55472708e-01 7.31892169e-01 -6.54136121e-01 -4.15652961e-01
7.39161730e-01 -4.42732453e-01 -9.79075372e-01 1.62296444e-01
1.12731621e-01 -4.54267710e-01 -3.63462836e-01 7.99831152e-01
3.03712577e-01 2.73234814e-01 8.94703642e-02 -1.44646990e+00
1.70010239e-01 8.39005411e-01 -6.03380203e-01 1.49516210e-01
2.65270602e-02 4.19485986e-01 8.91033769e-01 -2.23884135e-01
4.87076402e-01 8.49778652e-01 5.53014278e-01 2.07443759e-01
-1.09443843e+00 -5.21370649e-01 1.56390473e-01 -7.39909150e-03
-1.24655712e+00 -2.76654005e-01 3.83753568e-01 -2.58322030e-01
9.70323682e-01 3.49829286e-01 6.81327045e-01 4.11128104e-01
2.21289806e-02 3.33261371e-01 1.06022739e+00 -5.45284450e-01
4.72692490e-01 1.42208695e-01 -1.24207295e-01 9.09354687e-01
3.65094990e-01 1.91081151e-01 -3.80030945e-02 -1.48078293e-01
7.85397708e-01 -1.64804101e-01 -1.34267211e-01 -5.39774895e-01
-1.09730303e+00 9.98912930e-01 3.84024858e-01 2.34251067e-01
-1.50042623e-01 3.35878164e-01 3.22810829e-01 6.37095332e-01
2.81601161e-01 9.24962997e-01 -6.94928706e-01 -3.68524313e-01
-8.52626383e-01 3.36471170e-01 1.34171593e+00 8.08445632e-01
6.84929907e-01 -1.05405331e-01 4.24319208e-01 2.19231650e-01
-1.05313875e-01 -7.66744614e-02 3.28133285e-01 -1.39814878e+00
4.29441243e-01 3.31698924e-01 4.84323561e-01 -7.21744955e-01
-7.27847457e-01 -4.61644918e-01 -3.11018705e-01 3.27043563e-01
1.03591287e+00 -6.36701703e-01 -4.20400023e-01 1.69365537e+00
4.12463784e-01 5.76672181e-02 -2.93522358e-01 7.76443660e-01
-1.30979627e-01 4.46281284e-01 -1.28561422e-01 -1.62226349e-01
1.00877845e+00 -1.00336301e+00 -1.00982659e-01 -3.33252817e-01
1.34792852e+00 -7.18321025e-01 9.67059910e-01 5.75346768e-01
-1.25292742e+00 -1.05594561e-01 -1.09755266e+00 1.14582211e-01
-1.57545671e-01 -3.82455260e-01 1.20134449e+00 8.10889602e-01
-1.13143575e+00 8.09563637e-01 -1.08528256e+00 -3.33036810e-01
1.55410185e-01 5.24696648e-01 -2.10909724e-01 -3.43041718e-02
-7.26651847e-01 1.07019579e+00 4.34497118e-01 2.82693692e-02
-6.31296277e-01 -5.58456302e-01 -3.40642691e-01 3.92519027e-01
6.12147152e-01 -5.98614037e-01 1.35701764e+00 -9.91711259e-01
-1.35354781e+00 6.55012012e-01 3.53573292e-01 -4.12234396e-01
8.28550279e-01 1.59273982e-01 2.91563869e-01 -2.06024423e-02
-3.18011224e-01 3.52673620e-01 1.89464763e-01 -8.61056387e-01
-8.18938911e-01 -5.58391452e-01 8.03589404e-01 2.76185006e-01
-3.00823480e-01 1.41370177e-01 -2.91258425e-01 -1.89960808e-01
-2.99386561e-01 -1.34473586e+00 -6.47673845e-01 -1.53652623e-01
1.22101985e-01 -1.94995150e-01 2.42025375e-01 -4.71160889e-01
9.20655072e-01 -2.07995057e+00 1.82917401e-01 2.01954499e-01
3.45848143e-01 -2.10474461e-01 -2.63339847e-01 3.78927648e-01
3.61278816e-03 3.20379198e-01 -4.64240797e-02 -3.28500569e-02
2.27492988e-01 2.25853816e-01 1.61991164e-01 7.21678197e-01
-5.63507080e-02 5.00575244e-01 -7.43855238e-01 -4.81062174e-01
-5.32873161e-02 -2.09151670e-01 -1.00397813e+00 -1.76611952e-02
-1.22819901e-01 7.83547387e-02 -5.38364232e-01 1.27903014e-01
3.69037032e-01 -4.74258721e-01 4.78338540e-01 3.30995798e-01
-1.79503132e-02 2.37515301e-01 -1.26820064e+00 1.39633405e+00
-4.92356449e-01 4.37364787e-01 3.79292279e-01 -1.09476030e+00
1.66422188e-01 2.15977773e-01 3.85704547e-01 -4.94036317e-01
3.12705994e-01 1.71188504e-01 5.51001370e-01 -1.86927900e-01
2.66744912e-01 -1.31729901e-01 -9.69931334e-02 7.58346856e-01
-2.00691059e-01 -2.30419457e-01 6.06896520e-01 1.61042482e-01
1.30640507e+00 -8.68283138e-02 -3.83387730e-02 -3.33745152e-01
-1.86445133e-03 4.22721475e-01 3.14113617e-01 6.85203552e-01
-1.33914500e-01 -7.77772069e-02 8.92006099e-01 -5.48601925e-01
-1.59235585e+00 -5.97096622e-01 -1.07460447e-01 1.37274468e+00
-1.15159072e-01 -3.30582052e-01 -1.07247388e+00 -6.02244914e-01
2.76624840e-02 4.61044014e-01 -4.77609843e-01 3.54062617e-02
-7.55528152e-01 -1.03538263e+00 3.69439065e-01 5.71902454e-01
1.15467645e-01 -7.31925666e-01 -6.46403253e-01 1.29195943e-01
4.18253660e-01 -1.20718849e+00 -1.95998371e-01 3.26132417e-01
-9.34188724e-01 -1.08444822e+00 -5.02683520e-01 -5.57503939e-01
7.03901887e-01 2.02312976e-01 1.14750266e+00 5.40356398e-01
-1.24732845e-01 1.15452141e-01 -3.28487009e-01 -4.50939476e-01
-3.41839880e-01 3.74218136e-01 5.88266039e-03 -6.96409702e-01
-1.94871146e-02 -6.78456306e-01 -3.90633017e-01 2.43398368e-01
-7.63933957e-01 2.17782021e-01 5.60529292e-01 8.24864566e-01
-8.76635015e-02 1.91483602e-01 3.58630180e-01 -9.01888371e-01
4.04882878e-01 -4.64431971e-01 -1.14423072e+00 1.11551210e-01
-5.35802782e-01 2.44377494e-01 7.82036841e-01 -4.59286600e-01
-5.51789939e-01 -4.39481996e-02 3.10288578e-01 -2.34777153e-01
3.38925235e-02 4.96101588e-01 1.16356701e-01 -3.35493118e-01
5.32938361e-01 -2.80119658e-01 -3.57766263e-02 -2.27481648e-01
2.43977696e-01 3.47317457e-01 3.40088718e-02 -7.98454285e-01
6.80761933e-01 1.18043169e-01 2.16101050e-01 -5.88824928e-01
-7.95704782e-01 -7.71904439e-02 -3.47546220e-01 -6.44941404e-02
3.03982347e-01 -7.29676068e-01 -1.17173946e+00 5.43327071e-02
-7.96214104e-01 -9.35928226e-01 -1.87940285e-01 6.09744728e-01
-6.84950233e-01 1.16097838e-01 -7.95544565e-01 -5.44613302e-01
2.33729661e-01 -1.07222378e+00 4.47140545e-01 1.71267629e-01
-7.01957121e-02 -8.45190465e-01 -2.32706983e-02 5.63103855e-01
3.54696721e-01 9.65269133e-02 9.35989082e-01 -8.46419990e-01
-7.73265004e-01 -4.66886818e-01 -1.59568250e-01 1.43824652e-01
-4.25819904e-01 -6.23934567e-02 -5.53205609e-01 -6.00351632e-01
8.23762566e-02 -5.52889049e-01 4.26435679e-01 1.86649516e-01
1.13491380e+00 -4.75105137e-01 -1.29501149e-01 6.36637390e-01
1.49079430e+00 2.90499367e-02 3.82501692e-01 6.40363395e-01
6.08533978e-01 5.44059634e-01 6.24017179e-01 4.66601670e-01
2.34843284e-01 5.15995920e-01 2.54819423e-01 -4.69294973e-02
4.67476398e-01 9.27114710e-02 -1.79846063e-02 7.53642023e-01
-4.02250051e-01 -7.62274861e-02 -1.12268567e+00 4.29684252e-01
-1.96267462e+00 -7.15056717e-01 2.67098218e-01 2.27286506e+00
9.07261372e-01 4.26527262e-01 5.66522121e-01 1.11382410e-01
4.30171967e-01 -2.64527977e-01 -5.86476862e-01 -5.88759005e-01
4.03706968e-01 2.62424558e-01 8.86525691e-01 6.75019145e-01
-6.51855111e-01 8.44721317e-01 7.32120371e+00 6.68876767e-01
-9.70875978e-01 9.07888040e-02 8.12311471e-01 -7.87502527e-01
2.46911481e-01 1.22385405e-01 -3.39322418e-01 2.99683571e-01
1.43150485e+00 -5.70579886e-01 9.66577053e-01 1.20883930e+00
-7.75266364e-02 -3.43199551e-01 -1.49667299e+00 5.89558423e-01
-3.12933177e-01 -1.04426217e+00 -5.48144281e-01 4.44858253e-01
7.62223542e-01 2.17648000e-01 -1.76970288e-01 5.50348759e-01
8.63167226e-01 -1.10556948e+00 5.93348324e-01 -8.21349695e-02
9.84829590e-02 -1.20204663e+00 8.93814862e-01 7.78697014e-01
-7.05091000e-01 -2.54125029e-01 -4.42830443e-01 -6.69886112e-01
-4.47235823e-01 2.69709527e-01 -7.75780499e-01 2.46132806e-01
4.68933195e-01 -1.69618681e-01 -6.62174702e-01 1.23742902e+00
1.77380592e-02 4.77294177e-01 -6.12003982e-01 -3.31571162e-01
5.10312438e-01 -2.06581891e-01 -4.16753478e-02 1.04019487e+00
1.83087751e-01 4.14716274e-01 3.26467782e-01 5.09880662e-01
-2.05470964e-01 -3.87687981e-02 -2.92625219e-01 -3.33472639e-01
4.60808992e-01 1.36986387e+00 -8.89278591e-01 -2.44994462e-01
-3.30089986e-01 5.29777050e-01 6.64726079e-01 1.25895455e-01
-8.54937375e-01 -2.54558474e-01 4.29402858e-01 2.34158084e-01
3.39561820e-01 -5.20593822e-01 -2.55019248e-01 -1.10063493e+00
1.88461378e-01 -9.50493276e-01 3.22622389e-01 -5.24035871e-01
-9.27275717e-01 3.36357713e-01 -8.77987593e-02 -6.56526685e-01
-3.29840839e-01 -6.39337778e-01 -2.15380162e-01 6.69411123e-01
-1.07653522e+00 -5.13814867e-01 1.14054777e-01 8.63267481e-02
2.58087158e-01 2.04729095e-01 7.77340829e-01 1.84894010e-01
-6.96443141e-01 4.97136056e-01 2.46535704e-01 1.79503769e-01
3.45364809e-01 -1.22862101e+00 3.11215490e-01 6.69165373e-01
1.44804820e-01 3.79640579e-01 1.00793171e+00 -1.41433277e-03
-1.49916816e+00 -5.93710840e-01 4.77916807e-01 -5.60628533e-01
1.00156105e+00 -3.39859396e-01 -4.45286363e-01 1.00891268e+00
1.96510196e-01 7.61600733e-02 4.02873576e-01 6.19300842e-01
-1.88006699e-01 -8.81864652e-02 -9.83952045e-01 5.93852043e-01
8.55394840e-01 -2.83750772e-01 -2.20695570e-01 4.79252011e-01
6.31373167e-01 -6.34835243e-01 -9.80635405e-01 -7.10251331e-02
5.23341238e-01 -9.25651371e-01 6.37552142e-01 -1.03783119e+00
3.02064538e-01 7.48663396e-02 -3.65981124e-02 -1.21922421e+00
-2.88054824e-01 -5.05452991e-01 -2.38633086e-03 7.02693403e-01
4.85221714e-01 -4.37913984e-01 1.12195730e+00 1.12500906e+00
1.51344121e-01 -8.15254807e-01 -8.11398268e-01 -8.94940555e-01
3.57723176e-01 -1.75742969e-01 4.60734218e-01 1.00851095e+00
3.28879267e-01 2.92385221e-01 4.61735018e-02 8.93174261e-02
5.80985546e-01 1.26068816e-01 7.81322658e-01 -1.27513146e+00
-7.42472827e-01 -5.09720802e-01 -3.72286558e-01 -6.62531197e-01
-9.66774523e-02 -5.19062698e-01 3.35192159e-02 -9.94616449e-01
4.45104003e-01 -7.24412978e-01 -2.45101631e-01 4.44748670e-01
-1.81424804e-02 1.91845857e-02 6.35911763e-01 1.40793324e-01
-8.29077899e-01 1.09760575e-02 1.31732428e+00 3.48019540e-01
-6.73871338e-02 -6.49405047e-02 -6.28428638e-01 1.02483177e+00
8.71373415e-01 -6.17708802e-01 -2.30630219e-01 -4.97597128e-01
6.95004106e-01 3.37534040e-01 1.71631336e-01 -1.06595516e+00
2.58986562e-01 -3.42825651e-01 2.08294302e-01 9.50341746e-02
2.49181300e-01 -7.67818749e-01 -1.25730097e-01 5.83783865e-01
-4.24611300e-01 2.55969286e-01 2.15633094e-01 3.36172551e-01
2.84716547e-01 -3.88485670e-01 7.61221468e-01 -3.61559302e-01
-1.33545756e-01 1.03692643e-01 -2.46966615e-01 1.54279009e-01
1.32005334e+00 7.00843707e-02 -2.86222309e-01 -3.11133236e-01
-6.48993552e-01 4.46544558e-01 7.00515151e-01 -7.14402795e-02
-2.05746770e-01 -8.52289617e-01 -5.94582736e-01 -1.95820987e-01
-4.05116618e-01 -1.48651138e-01 1.03154644e-01 9.48701859e-01
-8.47509503e-01 3.83667082e-01 -3.30583185e-01 -1.46360755e-01
-1.33728123e+00 8.52198243e-01 2.81953096e-01 -5.76448321e-01
-4.53943461e-01 7.29069293e-01 -1.15308780e-02 -2.00609162e-01
3.97628903e-01 -3.26569527e-01 4.61199552e-01 -2.63517618e-01
5.08247375e-01 3.31018537e-01 5.73842712e-02 6.62910566e-02
-2.17518955e-01 8.16298500e-02 -3.53291690e-01 -2.52561271e-01
1.80904233e+00 1.61181971e-01 1.74374264e-02 1.60990685e-01
9.23967659e-01 -1.47986159e-01 -1.28370249e+00 6.42071888e-02
-6.69880733e-02 -3.63607287e-01 1.72687009e-01 -5.32836139e-01
-1.05551028e+00 6.79283857e-01 2.57922918e-01 6.71036243e-01
1.03853714e+00 -7.46581629e-02 5.02064645e-01 7.23352075e-01
8.06319058e-01 -1.24092829e+00 -1.58178255e-01 3.37248743e-01
6.07521951e-01 -9.72064495e-01 2.00328544e-01 -2.08077669e-01
-3.52667212e-01 8.92966628e-01 4.43363577e-01 -3.78928393e-01
3.81823897e-01 6.52798891e-01 -3.84181470e-01 -7.87664726e-02
-1.11939764e+00 7.61898607e-02 -5.03969073e-01 2.62934625e-01
2.41437912e-01 8.60689953e-02 -2.65448958e-01 5.40762484e-01
-6.31784439e-01 7.60648102e-02 7.10726917e-01 1.10217810e+00
-7.55556226e-01 -1.06975472e+00 -3.42139482e-01 3.28757107e-01
-5.26189506e-01 3.72828916e-02 -3.21106076e-01 1.07031095e+00
1.29131833e-02 7.94302702e-01 2.54647493e-01 -1.20394930e-01
1.66583717e-01 -1.74222484e-01 9.58542049e-01 -4.92926031e-01
-6.03006840e-01 -2.15317532e-01 4.10699278e-01 -7.79283822e-01
-1.66186586e-01 -3.30951929e-01 -1.08636701e+00 -1.24911535e+00
-4.64151174e-01 3.52449566e-01 5.86240888e-01 1.03299272e+00
9.04466733e-02 3.78745943e-01 6.12200499e-01 -8.57923687e-01
-1.01878130e+00 -8.48651707e-01 -7.72985220e-01 1.26977295e-01
-5.51344454e-02 -2.83635587e-01 -5.68640471e-01 -6.46544471e-02] | [5.080963134765625, 2.957099199295044] |
1b8f6195-3787-49d4-8005-12deab0521bc | optimizing-feature-set-for-click-through-rate | 2301.10909 | null | https://arxiv.org/abs/2301.10909v1 | https://arxiv.org/pdf/2301.10909v1.pdf | Optimizing Feature Set for Click-Through Rate Prediction | Click-through prediction (CTR) models transform features into latent vectors and enumerate possible feature interactions to improve performance based on the input feature set. Therefore, when selecting an optimal feature set, we should consider the influence of both feature and its interaction. However, most previous works focus on either feature field selection or only select feature interaction based on the fixed feature set to produce the feature set. The former restricts search space to the feature field, which is too coarse to determine subtle features. They also do not filter useless feature interactions, leading to higher computation costs and degraded model performance. The latter identifies useful feature interaction from all available features, resulting in many redundant features in the feature set. In this paper, we propose a novel method named OptFS to address these problems. To unify the selection of feature and its interaction, we decompose the selection of each feature interaction into the selection of two correlated features. Such a decomposition makes the model end-to-end trainable given various feature interaction operations. By adopting feature-level search space, we set a learnable gate to determine whether each feature should be within the feature set. Because of the large-scale search space, we develop a learning-by-continuation training scheme to learn such gates. Hence, OptFS generates the feature set only containing features which improve the final prediction results. Experimentally, we evaluate OptFS on three public datasets, demonstrating OptFS can optimize feature sets which enhance the model performance and further reduce both the storage and computational cost. | ['Xue Liu', 'Xiuqiang He', 'Liang Chen', 'Dugang Liu', 'Xing Tang', 'Fuyuan Lyu'] | 2023-01-26 | null | null | null | null | ['click-through-rate-prediction'] | ['miscellaneous'] | [ 1.47862837e-01 -6.28911018e-01 -4.33553159e-01 -5.68937898e-01
-5.41286886e-01 -5.11313736e-01 2.43875176e-01 -1.82344560e-02
-4.33017731e-01 4.81489390e-01 1.19790219e-01 -1.77586630e-01
-4.24855292e-01 -1.06552756e+00 -3.44200760e-01 -8.67297709e-01
1.09344907e-01 2.47005261e-02 4.47296590e-01 9.27915238e-03
3.34317267e-01 3.63948226e-01 -1.75874901e+00 4.47527885e-01
1.09971273e+00 1.13571322e+00 4.61524069e-01 -9.15932376e-03
-2.15544954e-01 1.40471458e-01 -3.91185671e-01 1.20756933e-02
3.55486721e-01 -2.74453044e-01 -4.58279192e-01 -1.14386797e-01
-1.60661802e-01 -2.82616377e-01 -2.17314973e-01 8.96245480e-01
4.11392421e-01 2.33091682e-01 3.69840115e-01 -1.26804721e+00
-3.42197835e-01 5.76633155e-01 -2.00865164e-01 1.66170433e-01
2.67383218e-01 2.69012928e-01 1.45039070e+00 -1.07345366e+00
4.20503706e-01 9.06647861e-01 2.29173660e-01 3.38338315e-01
-1.00946748e+00 -9.23537970e-01 4.44477707e-01 4.28340405e-01
-1.61410332e+00 -1.30473703e-01 7.67799318e-01 -2.33940050e-01
9.10561025e-01 6.91182852e-01 7.87414134e-01 4.91536260e-01
2.84444571e-01 8.98562074e-01 7.63151348e-01 -3.02149087e-01
1.89442411e-01 6.87909871e-02 4.05727684e-01 7.66180992e-01
9.17389765e-02 1.53026775e-01 -6.21096373e-01 -3.06930423e-01
5.05427182e-01 6.75869763e-01 -3.65646511e-01 -2.25364611e-01
-1.08869696e+00 9.06048536e-01 5.78834534e-01 2.07525656e-01
-2.59610951e-01 -1.61605477e-01 8.08248967e-02 4.84359175e-01
-3.77773829e-02 4.41367269e-01 -7.57553399e-01 -6.78468123e-02
-6.09384239e-01 1.58238873e-01 5.09472549e-01 6.27666295e-01
1.17926490e+00 -3.27844948e-01 -5.72777748e-01 7.07176208e-01
3.96792948e-01 2.47854307e-01 5.91982007e-01 -2.03678414e-01
5.74203014e-01 1.24879587e+00 5.58346417e-03 -1.08042800e+00
-4.39850271e-01 -5.13418257e-01 -5.80374658e-01 -1.94346815e-01
-1.87376142e-02 -1.08101092e-01 -9.91968215e-01 1.63129807e+00
3.88624370e-01 -7.01925680e-02 -2.97267020e-01 8.92005622e-01
6.94669127e-01 5.46457589e-01 -1.86231375e-01 -3.70994955e-01
1.24917209e+00 -8.65791023e-01 -4.77792412e-01 -6.34797812e-02
1.03020191e+00 -6.90744936e-01 1.24944389e+00 4.38449413e-01
-4.18653816e-01 -4.67923701e-01 -1.11035240e+00 1.90224558e-01
-4.08338279e-01 4.49706823e-01 1.06122828e+00 4.17525142e-01
-4.50957745e-01 7.63979137e-01 -8.12425911e-01 2.74538964e-01
4.14383560e-01 8.37515771e-01 -1.85067937e-01 -8.70130435e-02
-1.45922327e+00 3.66029143e-01 4.59613472e-01 2.10322008e-01
-4.86516774e-01 -4.37854618e-01 -5.08330166e-01 4.47238117e-01
4.74540591e-01 -3.12673122e-01 8.46836627e-01 -6.69701993e-01
-1.33658636e+00 -1.77038163e-01 -4.71833974e-01 -7.98625574e-02
2.98326880e-01 -1.19318947e-01 -6.17260039e-01 -4.13577139e-01
-3.20389569e-02 3.25105011e-01 7.92934179e-01 -4.87986952e-01
-1.01587760e+00 -1.18087403e-01 4.03891057e-02 2.49872714e-01
-7.54499674e-01 -1.28762826e-01 -7.02329934e-01 -2.54128009e-01
3.36512536e-01 -9.08082187e-01 -3.80232364e-01 -2.87212670e-01
-5.96225441e-01 -3.82636786e-01 7.09812582e-01 -7.30969012e-02
1.81927586e+00 -2.30212855e+00 -1.04488350e-01 5.91694117e-01
2.84765065e-01 9.93545055e-02 -9.59052294e-02 4.62294787e-01
1.49493292e-02 1.82908118e-01 1.68075740e-01 9.07772183e-02
-1.29164234e-01 2.09608927e-01 -2.75069058e-01 1.28904223e-01
2.32798159e-01 8.84558022e-01 -6.43647373e-01 -4.49002951e-01
5.81148341e-02 1.92103148e-01 -9.29764569e-01 1.59193531e-01
-1.01102330e-01 1.97170645e-01 -1.22778285e+00 5.83113849e-01
7.37501025e-01 -3.74891102e-01 -1.07748128e-01 -2.64033347e-01
-2.49568015e-01 2.79218167e-01 -1.37921226e+00 1.20633817e+00
-4.87444282e-01 1.33508593e-01 -6.63435400e-01 -5.85571289e-01
1.12754822e+00 -5.84890544e-02 5.96934676e-01 -6.47677183e-01
9.78761464e-02 2.89349496e-01 3.95278335e-01 -3.79672974e-01
1.40761375e-01 1.70495197e-01 -4.02998589e-02 3.87616932e-01
-2.88007129e-02 4.62923646e-01 2.82997221e-01 -5.84545024e-02
1.28047562e+00 -1.25977635e-01 1.58120289e-01 -9.67967615e-04
7.55097687e-01 -2.01165661e-01 8.17487776e-01 7.18722105e-01
1.94852769e-01 4.37664330e-01 4.41796094e-01 -5.82698226e-01
-3.25818717e-01 -6.33802474e-01 -1.96231470e-01 1.18231046e+00
2.78440505e-01 -9.28586423e-01 -3.42867911e-01 -1.16470301e+00
-3.34092863e-02 5.49606144e-01 -5.84427357e-01 -6.08258545e-01
-4.51234519e-01 -8.43073964e-01 -5.28837889e-02 3.10775161e-01
4.32762474e-01 -1.05692184e+00 -5.53224862e-01 1.82489067e-01
7.91361183e-02 -4.14688349e-01 -7.81569898e-01 5.40582716e-01
-7.40241349e-01 -1.04808187e+00 -6.72604516e-02 -5.49123108e-01
8.64130914e-01 3.31597865e-01 5.73604763e-01 3.63381356e-01
-2.24264070e-01 -3.66832167e-01 -5.68300962e-01 -8.02605599e-02
1.79643914e-01 2.57680655e-01 -7.96532705e-02 2.90698677e-01
5.90644598e-01 -4.14621800e-01 -7.75694370e-01 5.04376113e-01
-7.94029951e-01 1.11500658e-01 9.23483372e-01 1.23998177e+00
7.30336666e-01 4.94087040e-01 5.06145954e-01 -7.38726258e-01
5.27352929e-01 -3.71140599e-01 -5.60778856e-01 3.88516217e-01
-7.88439631e-01 4.56287742e-01 1.05643630e+00 -4.64578658e-01
-9.10346031e-01 4.83305663e-01 2.38804449e-03 -2.68800259e-01
1.76640883e-01 8.81447971e-01 -4.27994370e-01 -7.18293786e-02
4.37631905e-01 6.31487727e-01 -3.22707295e-01 -5.54713964e-01
3.83835547e-02 6.50217474e-01 -2.84342110e-01 -2.00021893e-01
8.14919829e-01 -1.26663903e-02 9.48694199e-02 -3.02109480e-01
-7.04301357e-01 -5.06814957e-01 -5.66709638e-01 2.09324732e-01
3.20642114e-01 -6.86059535e-01 -7.75020838e-01 1.92497652e-02
-7.52411366e-01 3.52930665e-01 -8.39675665e-02 6.31816924e-01
-3.09325997e-02 2.04217285e-01 -1.39338031e-01 -5.84810972e-01
-4.53814566e-01 -1.53548181e+00 8.55814815e-01 3.63768429e-01
-8.05780143e-02 -4.81241554e-01 -2.08776012e-01 -2.73133785e-01
2.49675557e-01 -1.38245732e-01 1.04377806e+00 -8.97557855e-01
-8.56544316e-01 -5.58587074e-01 -9.11657512e-02 8.52670595e-02
4.62117404e-01 -4.56989892e-02 -6.97458684e-01 -1.72478050e-01
-1.53348163e-01 -2.83049364e-02 9.88141954e-01 1.09113485e-01
1.45248163e+00 -3.17149848e-01 -7.47739375e-01 6.74862802e-01
1.24205673e+00 4.20963377e-01 3.44898224e-01 2.90289402e-01
6.22485876e-01 8.32021460e-02 1.07134748e+00 7.68466771e-01
-4.10592183e-03 8.07823777e-01 2.00551048e-01 1.73487246e-01
2.79573798e-01 -4.09037828e-01 2.33838126e-01 6.49060190e-01
1.56819865e-01 -1.43860132e-01 -5.88687897e-01 1.91380560e-01
-1.83027112e+00 -5.60138285e-01 1.11683078e-01 2.36907458e+00
9.65944409e-01 2.46480450e-01 -1.44375741e-01 2.71012992e-01
5.54645121e-01 -1.16207741e-01 -6.58730388e-01 -8.04959461e-02
1.75176978e-01 1.81786582e-01 2.93554783e-01 1.78857476e-01
-9.59293067e-01 7.95319915e-01 4.91414738e+00 1.11491156e+00
-1.44357920e+00 -2.55878925e-01 4.36602890e-01 -2.12295741e-01
-5.31278431e-01 1.98770255e-01 -1.29699647e+00 6.23757243e-01
4.48244601e-01 -2.56568074e-01 4.19453949e-01 1.03846467e+00
1.77197814e-01 1.16399646e-01 -1.07314694e+00 8.50277722e-01
-3.79609853e-01 -1.12972653e+00 4.30482060e-01 5.62857501e-02
4.52495009e-01 -1.86290249e-01 2.15752739e-02 6.03853941e-01
-1.22366482e-02 -9.46764231e-01 3.06408077e-01 4.20196116e-01
6.04544878e-01 -9.03808177e-01 6.93871319e-01 3.88925284e-01
-1.38658595e+00 -4.87996876e-01 -4.08097625e-01 2.41881399e-03
-2.03447491e-01 6.84359491e-01 -8.50481093e-01 5.70694625e-01
5.89932561e-01 6.87620163e-01 -8.83011520e-01 1.32582557e+00
-6.25315756e-02 6.23964906e-01 -5.93725264e-01 -4.73058403e-01
1.40900359e-01 5.20528741e-02 3.00779283e-01 8.28860879e-01
5.21924496e-01 5.68395257e-02 5.39431751e-01 7.23773837e-01
2.58583724e-01 4.31979448e-01 -5.94325922e-02 -1.17542610e-01
7.91387856e-01 1.28443980e+00 -7.21866131e-01 9.97705385e-02
-4.71969366e-01 6.84832752e-01 3.28570962e-01 2.33329698e-01
-7.95414984e-01 -8.16820204e-01 4.92906064e-01 3.64207476e-02
5.23095667e-01 1.35925442e-01 -2.94754475e-01 -1.20567441e+00
2.70239294e-01 -5.58801711e-01 3.85307193e-01 -3.61844636e-02
-9.60803449e-01 7.72192776e-01 -1.34406179e-01 -1.77372861e+00
-1.13402113e-01 -3.27474147e-01 -6.65595591e-01 9.18766260e-01
-1.29246652e+00 -9.23743129e-01 -4.80656892e-01 6.26725852e-01
4.16794330e-01 -1.52350411e-01 7.76432276e-01 1.92967355e-01
-6.81133389e-01 1.03638887e+00 1.70418203e-01 -1.72464445e-01
5.90309739e-01 -8.23871613e-01 1.55889526e-01 5.92548847e-01
4.08846661e-02 1.07206845e+00 3.38990211e-01 -6.08636320e-01
-1.60531187e+00 -1.06677890e+00 8.89582276e-01 -4.43349145e-02
3.06237906e-01 -4.25740004e-01 -8.09093237e-01 4.04486567e-01
-6.37481987e-01 3.53984416e-01 8.31453204e-01 4.23653573e-01
-1.96257740e-01 -3.39727879e-01 -8.88886392e-01 8.04298043e-01
1.03353608e+00 -1.69494197e-01 -3.06922823e-01 1.61596790e-01
7.86220431e-01 -1.32627115e-01 -6.85604692e-01 5.44268727e-01
7.75714040e-01 -7.65353203e-01 6.42730176e-01 -4.30545717e-01
2.92064190e-01 -5.84491789e-01 -6.33520260e-02 -1.22644269e+00
-6.48542225e-01 -5.56303740e-01 -1.32428616e-01 1.25984192e+00
6.66829050e-01 -8.63271058e-01 7.61200428e-01 6.31040752e-01
-1.01130977e-01 -1.40977407e+00 -8.54204953e-01 -5.47855794e-01
-5.59068143e-01 -2.06055209e-01 1.11204684e+00 6.28802299e-01
1.45162232e-02 3.65712434e-01 -2.85467505e-01 -5.88299595e-02
5.34985848e-02 7.75778949e-01 5.69700122e-01 -1.38944411e+00
-6.09946549e-01 -3.15557361e-01 -3.13124776e-01 -1.20487976e+00
-2.04348400e-01 -9.56638038e-01 8.88072420e-03 -1.21640539e+00
3.52809578e-01 -9.11955655e-01 -7.99631476e-01 9.38734591e-01
-5.00627995e-01 -2.50264019e-01 1.10174842e-01 3.01331729e-01
-6.05407238e-01 6.60176516e-01 1.35306406e+00 4.36461903e-02
-6.67852998e-01 3.11122268e-01 -8.03034723e-01 3.00465941e-01
6.50423408e-01 -5.80860317e-01 -4.46960449e-01 -1.26267269e-01
1.96510032e-01 -1.05446868e-01 -6.45382851e-02 -9.41742837e-01
2.13715404e-01 -5.86075664e-01 7.88134515e-01 -6.37056351e-01
1.70833036e-01 -9.04359341e-01 1.11444265e-01 6.42186522e-01
-3.65952402e-01 -2.87512213e-01 2.96855271e-02 5.20656168e-01
-2.73024291e-01 -2.59022027e-01 2.93146700e-01 1.62568450e-01
-8.48654687e-01 5.04626334e-01 2.97837332e-02 -5.18154263e-01
9.30160880e-01 -2.60436803e-01 -8.22097883e-02 -3.65422517e-02
-6.11408472e-01 4.40789640e-01 1.28185004e-01 5.15296221e-01
6.51988328e-01 -1.45305407e+00 -3.93662632e-01 7.68876255e-01
2.89495081e-01 -7.88236558e-02 3.51107001e-01 6.27699852e-01
-2.38431469e-02 6.83634996e-01 -3.94118018e-02 -5.04688919e-01
-1.40561271e+00 5.09368598e-01 -2.33753324e-02 -4.62206155e-01
-4.19519544e-01 9.77446795e-01 3.21011901e-01 -2.99511552e-01
7.47215142e-03 -3.32461953e-01 -5.40118814e-01 4.20352370e-02
6.48433149e-01 1.14855273e-02 1.36417612e-01 -1.54228613e-01
-5.19368410e-01 6.00774407e-01 -5.27635455e-01 4.16093528e-01
1.21632385e+00 -1.74766332e-02 6.87587336e-02 1.22969724e-01
1.41226029e+00 1.29693180e-01 -1.05866230e+00 -3.47579956e-01
-1.50368661e-01 -7.67373085e-01 2.56320953e-01 -7.82673478e-01
-1.16770267e+00 5.15592158e-01 7.35110939e-01 -8.18518363e-03
1.46373439e+00 -3.69355679e-01 7.67066061e-01 5.99730968e-01
3.73164713e-01 -8.24860752e-01 -1.40336171e-01 5.89549601e-01
6.29097402e-01 -1.03640521e+00 1.83137238e-01 -7.16376424e-01
-5.51715016e-01 1.20227396e+00 8.10743690e-01 -9.03861299e-02
8.52438748e-01 1.86676681e-02 -3.79482359e-01 -5.23210242e-02
-9.60846841e-01 -1.80410251e-01 5.11666059e-01 -9.98453889e-03
2.35779971e-01 2.53176577e-02 -7.12557673e-01 1.26465559e+00
-3.45954984e-01 1.96717143e-01 -1.60070676e-02 9.44018066e-01
-4.96454537e-01 -1.40789092e+00 1.76881596e-01 1.08095181e+00
-2.81056315e-01 -3.22697997e-01 -1.29393101e-01 4.27440017e-01
4.09522235e-01 8.45669270e-01 -1.25788629e-01 -1.06863391e+00
3.58367413e-01 9.08305943e-02 9.60680544e-02 -6.28913522e-01
-6.74581110e-01 1.31826967e-01 -7.75853097e-02 -5.45410872e-01
2.84779072e-01 -5.56023598e-01 -1.42899072e+00 1.33872509e-01
-1.06108522e+00 4.37211752e-01 3.29765797e-01 1.08388221e+00
6.79645300e-01 4.84848440e-01 1.18917477e+00 -3.95879626e-01
-8.52329969e-01 -8.04473996e-01 -5.49410582e-01 1.28949985e-01
1.84275478e-01 -9.33005571e-01 -2.64949799e-01 -4.50150102e-01] | [10.119778633117676, 5.379974842071533] |
b172319b-048c-4ba7-a074-f7874bbe2a1e | cldice-a-topology-preserving-loss-function | 2003.07311 | null | https://arxiv.org/abs/2003.07311v7 | https://arxiv.org/pdf/2003.07311v7.pdf | clDice -- A Novel Topology-Preserving Loss Function for Tubular Structure Segmentation | Accurate segmentation of tubular, network-like structures, such as vessels, neurons, or roads, is relevant to many fields of research. For such structures, the topology is their most important characteristic; particularly preserving connectedness: in the case of vascular networks, missing a connected vessel entirely alters the blood-flow dynamics. We introduce a novel similarity measure termed centerlineDice (short clDice), which is calculated on the intersection of the segmentation masks and their (morphological) skeleta. We theoretically prove that clDice guarantees topology preservation up to homotopy equivalence for binary 2D and 3D segmentation. Extending this, we propose a computationally efficient, differentiable loss function (soft-clDice) for training arbitrary neural segmentation networks. We benchmark the soft-clDice loss on five public datasets, including vessels, roads and neurons (2D and 3D). Training on soft-clDice leads to segmentation with more accurate connectivity information, higher graph similarity, and better volumetric scores. | ['Ulrich Bauer', 'Josien P. W. Pluim', 'Johannes C. Paetzold', 'Alexander Unger', 'Andrey Zhylka', 'Ivan Ezhov', 'Bjoern H. Menze', 'Anjany Sekuboyina', 'Suprosanna Shit'] | 2020-03-16 | null | null | null | null | ['graph-similarity'] | ['graphs'] | [-2.43475810e-02 4.17650938e-01 -1.06088318e-01 -3.49847645e-01
1.57407179e-01 -7.84170270e-01 3.96951348e-01 6.44738972e-01
-3.25515300e-01 5.28199673e-01 -1.25708461e-01 -4.92049336e-01
-1.56339526e-01 -1.08746350e+00 -7.20328212e-01 -5.26272297e-01
-3.75013828e-01 4.08456534e-01 7.49957383e-01 -5.15879616e-02
7.49778524e-02 9.58803058e-01 -9.47113514e-01 -4.86338258e-01
1.15634239e+00 1.01512277e+00 -1.45492479e-01 4.15315062e-01
-4.48933840e-01 3.28675061e-01 -5.53778559e-03 -8.56476724e-01
4.92662400e-01 -2.28740320e-01 -9.67901468e-01 -1.58450574e-01
6.17010772e-01 6.94433674e-02 -2.91384071e-01 1.27942228e+00
3.18706900e-01 3.64275300e-03 9.70543087e-01 -1.35520530e+00
-4.01047736e-01 6.29248321e-01 -6.48968995e-01 2.45073244e-01
-3.05866480e-01 1.15083657e-01 8.32604408e-01 -5.69150269e-01
8.62437963e-01 1.25750399e+00 1.09439337e+00 2.48940334e-01
-1.62300158e+00 -2.91857123e-01 6.47449270e-02 -2.44030505e-01
-1.20227098e+00 1.52548747e-02 6.94542110e-01 -7.98639476e-01
1.54951364e-01 1.48967519e-01 8.14073145e-01 4.99854416e-01
2.55745769e-01 3.58717740e-01 9.23061430e-01 2.34037340e-02
2.04754248e-01 -7.94350132e-02 5.00166476e-01 1.01580417e+00
5.68154573e-01 -2.51603991e-01 1.32327452e-01 3.01276714e-01
1.27911675e+00 -2.82927334e-01 -3.07480574e-01 -7.46487737e-01
-9.47995186e-01 6.34125233e-01 6.76656544e-01 2.04349071e-01
-1.88979313e-01 -1.81137383e-01 4.48942572e-01 3.25241238e-01
4.32729900e-01 2.14342192e-01 -1.87634677e-02 2.46479973e-01
-6.60008013e-01 -1.03614002e-01 8.89305055e-01 6.89740419e-01
6.60610557e-01 -2.37999976e-01 -1.52569890e-01 6.75614297e-01
4.76054102e-02 1.40108600e-01 1.91377804e-01 -1.16883898e+00
1.75492331e-01 8.99550498e-01 -6.43140733e-01 -1.15993273e+00
-9.10731137e-01 -6.43614769e-01 -1.47743344e+00 2.25429580e-01
8.60126257e-01 -2.25023758e-02 -7.42301345e-01 1.81929088e+00
7.14086950e-01 1.25648454e-01 -1.13332227e-01 8.50148320e-01
1.05842876e+00 1.31266713e-01 -4.86942083e-02 3.13292705e-02
9.56304669e-01 -7.45921612e-01 -2.34606162e-01 2.71089464e-01
7.25048065e-01 -2.67503321e-01 8.91521215e-01 -1.13684215e-01
-1.16662002e+00 -1.93213299e-01 -8.26894581e-01 -3.34318548e-01
-4.09230858e-01 -2.12339282e-01 6.10369265e-01 6.64591491e-01
-1.21887541e+00 9.91046607e-01 -9.06507730e-01 -3.51434141e-01
9.46572542e-01 2.74576664e-01 -5.40162921e-01 2.32061118e-01
-7.61562049e-01 8.16873908e-01 1.03460126e-01 -2.66001016e-01
-3.65150601e-01 -1.23518097e+00 -8.81325841e-01 8.43765587e-02
9.12430063e-02 -7.28360295e-01 6.29376292e-01 -3.91185611e-01
-1.18926430e+00 1.25085819e+00 1.54502288e-01 -7.38112807e-01
1.07727480e+00 4.36293513e-01 5.35164624e-02 4.62264866e-01
1.89487040e-01 7.96340942e-01 4.30492073e-01 -1.01739693e+00
-2.02461660e-01 -4.25596267e-01 -4.27635983e-02 -8.65083039e-02
-3.56489480e-01 -4.89435345e-01 -3.01233828e-01 -5.52455366e-01
4.35915172e-01 -4.75402832e-01 -2.94627279e-01 8.84058356e-01
-8.05940211e-01 -2.48580500e-01 7.16258407e-01 -7.59115160e-01
8.75607073e-01 -1.98760748e+00 2.50715137e-01 6.36471570e-01
7.35290051e-01 1.49914786e-01 -1.02261111e-01 -2.56453902e-01
4.60130572e-02 3.39756399e-01 -7.41987646e-01 -1.63064212e-01
-2.19791412e-01 1.27315789e-01 1.94299653e-01 6.00300908e-01
1.62566245e-01 9.61127043e-01 -7.23806858e-01 -7.83044994e-01
1.19947016e-01 4.70022887e-01 -3.49106371e-01 -4.17479724e-01
9.66558829e-02 4.98387247e-01 -4.13600057e-01 4.20397252e-01
8.51513267e-01 -4.37021822e-01 -6.15438670e-02 -2.82113165e-01
-9.86714661e-02 -2.64481008e-01 -9.16515887e-01 1.34233248e+00
-2.70422608e-01 7.07885265e-01 3.16809177e-01 -1.05691922e+00
1.11419261e+00 -2.03575864e-01 8.58104527e-01 -6.71456873e-01
3.68738413e-01 1.58163786e-01 -1.30863726e-01 -2.57651240e-01
-5.03592975e-02 1.65559091e-02 2.82390684e-01 1.41436219e-01
-8.26460049e-02 -8.33313242e-02 5.20359099e-01 4.85839546e-01
9.82200623e-01 -2.89122999e-01 3.85583341e-02 -6.10023558e-01
5.41763544e-01 -1.88212693e-01 5.05337238e-01 3.68584961e-01
-4.86439198e-01 5.97105801e-01 1.17695498e+00 -2.02519923e-01
-1.05057037e+00 -1.28948402e+00 -6.52874231e-01 3.90467614e-01
5.04590034e-01 1.80940360e-01 -1.00348449e+00 -7.68083453e-01
3.84917587e-01 2.30881169e-01 -6.01466119e-01 -2.24180296e-01
-5.21725059e-01 -5.07198393e-01 6.98774219e-01 4.31447595e-01
6.87587738e-01 -5.47427297e-01 -4.66167271e-01 3.62938121e-02
-1.82236005e-02 -1.27192354e+00 -7.96186507e-01 -1.02567777e-01
-1.16110158e+00 -1.35284376e+00 -9.84974802e-01 -1.02550435e+00
8.72449398e-01 6.68527838e-03 1.06719542e+00 2.35932857e-01
-4.27371293e-01 6.46920800e-02 4.06796671e-03 1.70889199e-01
-4.41231310e-01 1.17870271e-01 -4.57808614e-01 8.46588388e-02
-2.71782756e-01 -8.72176230e-01 -5.56085646e-01 5.78639269e-01
-6.98351562e-01 1.23072445e-01 3.48006010e-01 4.11414117e-01
1.02622509e+00 7.52375200e-02 5.84884167e-01 -9.14022982e-01
4.08818781e-01 -3.29031825e-01 -7.76236594e-01 3.09014499e-01
-4.82286423e-01 9.35304239e-02 6.29619598e-01 -3.07315528e-01
-4.80401963e-01 4.07216372e-03 -6.50148615e-02 -4.10713196e-01
3.16896662e-02 3.49221766e-01 -6.91738427e-02 -6.86575055e-01
4.62622464e-01 8.73361453e-02 4.71573293e-01 -4.71503735e-01
2.25823343e-01 5.49412239e-03 8.66879642e-01 -4.47938949e-01
7.47424185e-01 7.98063338e-01 9.34838235e-01 -9.94976640e-01
-4.50252444e-01 -2.53829032e-01 -9.66577888e-01 -2.60761946e-01
8.42220664e-01 -2.12944970e-01 -7.38500953e-01 4.84122038e-01
-8.20323169e-01 -5.80110967e-01 -5.48375010e-01 2.76699990e-01
-4.47132260e-01 7.06657112e-01 -5.27374864e-01 -4.24169660e-01
-4.77229506e-01 -9.84357774e-01 6.33470058e-01 2.87472337e-01
1.60221383e-01 -1.43853164e+00 2.40742080e-02 -6.82879910e-02
4.19984847e-01 9.79352415e-01 1.44745111e+00 -4.71630007e-01
-2.91556299e-01 -1.60253607e-02 -5.85763454e-01 3.96554172e-01
-1.36743216e-02 2.94658333e-01 -3.39955389e-01 6.48780093e-02
-4.43776071e-01 1.03035592e-01 1.18467546e+00 7.77166009e-01
1.10921955e+00 -2.88852125e-01 -3.32113266e-01 8.77262175e-01
1.35461426e+00 -2.01700926e-01 4.95866448e-01 -9.54706520e-02
7.97979534e-01 8.45931053e-01 -5.74645698e-02 6.99122846e-02
4.87886518e-01 4.07317132e-01 5.42113483e-01 -5.04175127e-01
-5.82677901e-01 -3.18158716e-02 -4.11897190e-02 6.47821188e-01
-8.73333365e-02 8.08639824e-02 -6.41073108e-01 6.20805502e-01
-1.51093423e+00 -6.31489694e-01 -7.01228440e-01 2.21343088e+00
8.12180817e-01 2.30081141e-01 4.53301072e-01 -4.29161303e-02
1.15298414e+00 -2.18574107e-01 -9.65837061e-01 -2.42021218e-01
-6.04263306e-01 2.20487475e-01 6.39576018e-01 5.12808919e-01
-1.25407934e+00 7.25264430e-01 5.57286167e+00 6.34730458e-01
-1.11571956e+00 8.75833556e-02 9.66866016e-01 2.18669772e-01
-3.32278877e-01 -2.81782478e-01 -5.59275150e-01 4.88392621e-01
3.50132257e-01 -2.27003559e-01 1.19555732e-02 5.89932561e-01
2.07573235e-01 7.05221295e-02 -9.64165926e-01 5.35587788e-01
-3.91009092e-01 -1.51332057e+00 1.55843139e-01 1.54516935e-01
8.59112024e-01 -2.44452083e-03 -1.24248397e-02 -2.84284353e-01
2.56886810e-01 -9.94784474e-01 6.68769658e-01 5.00432312e-01
9.35991108e-01 -6.84761524e-01 5.57825804e-01 -2.14652903e-02
-1.23823369e+00 4.24639165e-01 -4.23176080e-01 5.45707643e-01
3.46868038e-01 1.00993621e+00 -5.30643940e-01 5.16777277e-01
4.24133897e-01 9.72968519e-01 -5.91798067e-01 1.73395598e+00
1.62929490e-01 4.24515069e-01 -5.07588923e-01 5.43220453e-02
1.28669962e-01 -7.34121382e-01 6.54838920e-01 1.08779836e+00
1.28882810e-01 -9.36412364e-02 1.03467047e-01 1.10880744e+00
-4.54993784e-01 3.20802063e-01 -4.78496313e-01 2.23976150e-01
4.24049139e-01 1.41343939e+00 -1.42228055e+00 -9.53761339e-02
-1.85212269e-01 5.74053764e-01 2.13626787e-01 1.50127202e-01
-6.02828324e-01 -4.48260784e-01 6.19218349e-01 3.91700357e-01
1.95017099e-01 -1.55132979e-01 -8.05975676e-01 -7.74773598e-01
-8.33017472e-03 -5.32258712e-02 3.06384206e-01 -2.18234569e-01
-1.46720064e+00 4.28534031e-01 -3.18966657e-01 -8.10814083e-01
6.81767762e-01 -6.39779687e-01 -8.30654979e-01 4.96143579e-01
-1.56327486e+00 -8.99498284e-01 -5.22240460e-01 4.90467161e-01
1.38294930e-02 1.40626565e-01 2.61276960e-01 4.31956202e-01
-7.32228875e-01 5.98897398e-01 -7.43543878e-02 4.64461267e-01
1.88531354e-01 -1.31011093e+00 4.13440824e-01 7.68188536e-01
-1.79843321e-01 2.77029485e-01 2.44095370e-01 -7.40665674e-01
-8.59406769e-01 -1.51631355e+00 6.00462079e-01 5.04883658e-03
6.18289590e-01 -5.45030572e-02 -9.67141092e-01 5.04252791e-01
-3.61455530e-01 2.02037543e-01 2.58251131e-01 -3.94322455e-01
-2.98377156e-01 -1.55943424e-01 -1.46914542e+00 6.20721281e-01
1.28287458e+00 1.40817724e-02 -5.54757640e-02 4.85879183e-01
8.23095977e-01 -3.14350992e-01 -1.33975339e+00 3.56078804e-01
5.39393961e-01 -9.80999351e-01 1.14229465e+00 -4.44044083e-01
4.06747639e-01 -1.63348570e-01 3.11863869e-01 -1.12204790e+00
-2.17389837e-01 -6.57790840e-01 2.56484896e-02 9.42421734e-01
4.89019960e-01 -8.09248805e-01 8.44772518e-01 3.04512471e-01
-4.69668746e-01 -9.54711735e-01 -1.04819798e+00 -1.03989685e+00
5.49122632e-01 1.01866990e-01 5.27556598e-01 1.01288855e+00
-4.01960999e-01 9.98414755e-02 3.75152737e-01 -8.90396908e-02
9.73800063e-01 -3.69590670e-02 5.04813552e-01 -1.92875552e+00
2.02258393e-01 -1.21485531e+00 -7.86150932e-01 -9.86175239e-01
1.51557654e-01 -1.55196548e+00 -4.27961379e-01 -1.75231946e+00
1.99541841e-02 -7.23504305e-01 7.36396760e-02 5.37779331e-01
2.83249944e-01 3.10638964e-01 -1.63943499e-01 5.22578061e-02
-2.70385683e-01 5.31457961e-01 1.99287379e+00 -2.20786080e-01
-3.24004948e-01 2.08281130e-01 -5.65394819e-01 7.78977573e-01
7.91784704e-01 -2.28657484e-01 -1.66778684e-01 -9.67192054e-02
-1.23838924e-01 -6.54815361e-02 7.79058993e-01 -1.00163114e+00
3.44939590e-01 1.30327910e-01 1.05697289e-01 -5.18338740e-01
-4.80677783e-02 -6.35114491e-01 -1.69652365e-02 6.98545277e-01
-3.89439911e-01 -2.54634470e-01 -5.44981472e-02 3.60342652e-01
6.81117848e-02 -1.76467299e-01 1.21621585e+00 -4.41285782e-02
-3.55204761e-01 5.85801482e-01 1.87146198e-02 4.68932360e-01
1.20288479e+00 -5.48046172e-01 -4.75734293e-01 -9.62803885e-02
-7.99968064e-01 3.50680470e-01 5.65380991e-01 -9.54670459e-02
6.29268169e-01 -9.44161057e-01 -7.42628336e-01 1.37935653e-01
-1.07947156e-01 3.32468301e-01 2.05487058e-01 1.33336806e+00
-9.06742454e-01 1.95661396e-01 -4.82683927e-01 -7.75220037e-01
-1.17049003e+00 1.28996670e-01 7.29470253e-01 -5.75774759e-02
-1.01803803e+00 8.63479435e-01 2.16242447e-01 -3.95430535e-01
3.92775565e-01 -7.26299047e-01 -3.44301105e-01 1.83474377e-03
-3.69765908e-02 7.47355163e-01 1.21326953e-01 -7.55456746e-01
-1.75775841e-01 9.06473815e-01 1.74649149e-01 2.65929580e-01
1.24246407e+00 -9.86167863e-02 -3.73430431e-01 -2.01306981e-03
1.32788730e+00 -3.77155900e-01 -1.45132017e+00 -2.69904166e-01
-9.19975787e-02 -1.21860281e-01 9.51994136e-02 -4.82392877e-01
-1.73907566e+00 8.48334074e-01 4.34464842e-01 4.40844655e-01
8.45096767e-01 1.65356785e-01 9.77619112e-01 1.01103723e-01
2.25402981e-01 -7.34739780e-01 1.11421589e-02 4.62730259e-01
6.13089740e-01 -7.27273166e-01 -1.75573215e-01 -9.95618045e-01
-2.65583336e-01 1.08091891e+00 5.62249959e-01 -5.95603660e-02
1.05093622e+00 1.34135662e-02 -2.48793676e-01 -2.65839845e-01
-9.15971249e-02 -2.46932566e-01 2.65441477e-01 7.22849309e-01
3.78080201e-03 1.41465425e-01 -3.03641081e-01 3.71146321e-01
-3.57908934e-01 -3.27786356e-01 6.10478282e-01 3.73478651e-01
-5.80075264e-01 -6.63316011e-01 1.49134666e-01 7.89475024e-01
-2.21421033e-01 1.74375549e-01 -6.16872847e-01 6.81273043e-01
1.85514186e-02 5.79576015e-01 3.76697451e-01 -5.63765317e-02
3.36404383e-01 -4.63310540e-01 3.96208435e-01 -2.13924333e-01
-3.74813646e-01 -2.29588822e-01 -1.66571781e-01 -4.26226288e-01
-1.96331248e-01 -5.31717777e-01 -1.71967208e+00 -5.20126164e-01
-1.18620954e-01 -1.81020394e-01 6.33426428e-01 6.77128792e-01
4.28306043e-01 4.41140831e-01 6.93888843e-01 -3.33532304e-01
-2.76100039e-01 -3.75696778e-01 -7.78531492e-01 5.87646961e-01
2.32944995e-01 -7.40955412e-01 -3.15111428e-01 -3.18932831e-02] | [14.293086051940918, -2.6550652980804443] |
6587bfe7-4311-4c62-a740-73fe55c96d66 | yes-we-can-annotating-english-modal-verbs | null | null | https://aclanthology.org/L12-1458 | https://aclanthology.org/L12-1458.pdf | Yes we can!? Annotating English modal verbs | This paper presents an annotation scheme for English modal verbs together with sense-annotated data from the news domain. We describe our annotation scheme and discuss problematic cases for modality annotation based on the inter-annotator agreement during the annotation. Furthermore, we present experiments on automatic sense tagging, showing that our annotations do provide a valuable training resource for NLP systems. | ['Ines Rehbein', 'Josef Ruppenhofer'] | 2012-05-01 | null | null | null | lrec-2012-5 | ['subjectivity-analysis'] | ['natural-language-processing'] | [ 1.74796849e-01 7.64456093e-01 -5.37968874e-01 -4.83145386e-01
-1.05885875e+00 -1.15697479e+00 5.57457745e-01 4.61024493e-01
-7.77814031e-01 1.38667846e+00 9.47151601e-01 -1.71738997e-01
5.63221574e-02 -3.91615212e-01 -3.25674444e-01 -4.06304985e-01
2.04858467e-01 7.22553909e-01 5.54369211e-01 -6.29837394e-01
1.21350855e-01 4.18144651e-02 -1.13170958e+00 8.53020191e-01
4.30228055e-01 6.78617537e-01 3.04101557e-01 1.12506837e-01
-7.49188960e-01 1.13788903e+00 -6.17776692e-01 -4.44039106e-01
-2.19831899e-01 -2.89933056e-01 -1.74590063e+00 7.98821636e-03
-1.30253345e-01 6.49143100e-01 -2.92646550e-02 1.16623747e+00
4.34083581e-01 8.24184045e-02 3.81399006e-01 -1.21800745e+00
-3.29886079e-01 1.45080602e+00 -2.04480723e-01 1.70664728e-01
1.37944055e+00 -5.53517342e-01 1.33607066e+00 -4.26009655e-01
1.58327699e+00 8.66770327e-01 7.76301444e-01 8.20830107e-01
-9.62844431e-01 -5.11432476e-02 -2.25876510e-01 3.16905767e-01
-1.40157783e+00 -4.37188506e-01 8.21838140e-01 -3.54180455e-01
1.27092612e+00 2.85190940e-01 7.56962121e-01 1.29450512e+00
-2.42379084e-01 1.10935211e+00 1.01402199e+00 -1.02316141e+00
-6.47132285e-03 2.98974395e-01 4.35969204e-01 4.64171886e-01
-4.50873114e-02 -3.25463593e-01 -6.26695871e-01 -4.16983068e-01
6.53752983e-01 -1.11029315e+00 -4.01979655e-01 -2.13554829e-01
-1.52847540e+00 6.66447043e-01 -1.17440663e-01 9.77391124e-01
-3.51082146e-01 -1.30498946e-01 1.09479570e+00 3.62592548e-01
4.34124887e-01 5.67131996e-01 -9.56927180e-01 -3.22924942e-01
-5.03917277e-01 2.90073752e-02 1.17460465e+00 1.13802195e+00
2.30688840e-01 -2.97227591e-01 8.44250619e-02 1.09285760e+00
2.21465051e-01 2.75556713e-01 5.05167603e-01 -1.06209588e+00
4.53745008e-01 6.22087479e-01 4.52489048e-01 -6.21030331e-01
-9.77226555e-01 3.72661442e-01 -1.71413586e-01 -5.63697219e-01
5.78256488e-01 -1.54026046e-01 -5.94321787e-01 1.60293305e+00
2.59459823e-01 -5.98119080e-01 8.02240729e-01 5.92159986e-01
1.27197981e+00 4.58358169e-01 6.54659152e-01 -1.04303312e+00
1.83691657e+00 -4.71215993e-01 -1.45788217e+00 -9.64537337e-02
1.22406352e+00 -1.07546043e+00 9.21466112e-01 2.31196478e-01
-1.01512814e+00 1.36162207e-01 -3.58708292e-01 -9.47341248e-02
-4.31030065e-01 1.61447778e-01 9.62168813e-01 3.70574504e-01
-7.94898570e-01 -6.51422814e-02 -5.02992272e-01 -1.12904680e+00
-1.96574122e-01 1.91957414e-01 -8.45694304e-01 3.69810730e-01
-1.62062240e+00 1.20126259e+00 1.40079165e+00 -1.77910984e-01
-4.05011512e-02 1.00698374e-01 -1.05232918e+00 -5.05071521e-01
5.31780660e-01 -3.01358163e-01 1.44670701e+00 -9.28786993e-01
-1.04371750e+00 1.63867545e+00 -1.48837984e-01 -3.87535632e-01
1.16913930e-01 1.78576887e-01 -9.30149913e-01 2.17755720e-01
3.47192913e-01 5.41584432e-01 -1.09291993e-01 -1.30928767e+00
-7.88399160e-01 -1.15277886e-01 2.74422288e-01 2.56521940e-01
-2.34349985e-02 7.07229376e-01 -1.51878774e-01 -8.53983879e-01
4.39776629e-01 -8.85801554e-01 5.94761632e-02 -6.90242767e-01
-5.00965774e-01 -8.87109041e-01 6.30891740e-01 -5.13397396e-01
1.36032617e+00 -2.12964559e+00 3.41290757e-02 1.48266345e-01
-2.79776692e-01 -3.42985481e-01 -2.65709963e-02 8.31995308e-01
-3.96071851e-01 1.49428263e-01 8.20773691e-02 1.62872836e-01
1.17999420e-01 8.65311384e-01 -5.43069124e-01 1.52450159e-01
-3.52785170e-01 6.31341100e-01 -1.11141551e+00 -1.06806993e+00
-7.26085976e-02 1.95582554e-01 -1.07526667e-01 -9.58248153e-02
-3.36323529e-01 5.50313234e-01 -6.06218219e-01 7.62646377e-01
-1.08122557e-01 -3.00773457e-02 8.72211993e-01 -5.79017758e-01
-1.73407853e-01 5.79964936e-01 -1.05310142e+00 1.99710846e+00
-2.96993285e-01 6.09885991e-01 2.31781453e-02 -9.03411329e-01
5.03503025e-01 1.09036219e+00 6.02487743e-01 -2.96112299e-01
3.02387297e-01 3.92439961e-01 -3.07622164e-01 -9.59227026e-01
8.76993418e-01 -4.88273561e-01 -8.31098378e-01 3.44543576e-01
4.72526550e-01 -3.13174367e-01 6.44991755e-01 1.34456679e-01
8.31744432e-01 2.58019090e-01 1.08520544e+00 -3.43331575e-01
3.95355821e-01 6.04001164e-01 7.82547653e-01 2.91064411e-01
-2.84389973e-01 7.04988651e-03 6.84122920e-01 -5.49398184e-01
-7.88498819e-01 -5.67356706e-01 -5.00092208e-01 1.43838489e+00
1.69450313e-01 -7.71997988e-01 -6.04437709e-01 -9.81308520e-01
-8.36233914e-01 8.67228627e-01 -7.02979803e-01 5.39445996e-01
-3.51030648e-01 -6.79916620e-01 9.59749222e-01 7.22000897e-01
2.30220005e-01 -1.39665151e+00 -3.13165724e-01 2.75141805e-01
-9.93406832e-01 -1.45959425e+00 9.17490870e-02 6.03554308e-01
-3.41038764e-01 -1.37600875e+00 -1.02499053e-01 -1.30625510e+00
2.66570061e-01 -3.33809614e-01 1.23210931e+00 -8.74683037e-02
3.93020511e-01 5.39384544e-01 -8.33391547e-01 -3.81286591e-01
-6.73252165e-01 1.37888610e-01 1.74676239e-01 -8.24417174e-01
7.49477804e-01 -2.16664955e-01 2.70491332e-01 3.09985042e-01
-9.65151489e-01 -1.81852356e-01 -1.75322831e-01 8.23934674e-01
6.14993572e-01 -1.46868199e-01 3.14726681e-01 -1.13527751e+00
6.61134064e-01 -3.00511986e-01 -3.21283966e-01 5.49093485e-01
3.95206064e-02 2.10864276e-01 -1.56052917e-01 -3.02353173e-01
-1.46947432e+00 4.48885918e-01 -4.74669874e-01 5.89664519e-01
-3.97380292e-01 7.03055084e-01 -3.45280588e-01 2.07313392e-02
1.02634597e+00 -1.55721322e-01 -6.66725993e-01 -3.89055580e-01
4.45416123e-01 6.28200948e-01 5.04385293e-01 -6.82419002e-01
1.23438872e-01 3.31652910e-01 -2.78602540e-01 -7.07399786e-01
-1.29718530e+00 -7.20310390e-01 -7.74524093e-01 -1.59488365e-01
1.09700811e+00 -8.13890815e-01 -4.93029803e-01 -1.58504903e-01
-1.51028085e+00 -5.84225208e-02 -6.60846591e-01 7.41872907e-01
-9.11247790e-01 4.93318349e-01 -8.30444813e-01 -8.05477560e-01
2.36342289e-02 -6.33018494e-01 9.22613800e-01 -6.01908751e-02
-9.03164029e-01 -1.56869745e+00 3.30373615e-01 4.11757469e-01
-1.69135302e-01 1.38557106e-01 6.67090535e-01 -1.33050752e+00
3.97599339e-01 -2.42928304e-02 -2.85599157e-02 -2.97053456e-01
-1.50723532e-01 -3.47639352e-01 -7.24226892e-01 2.73391873e-01
-4.22496915e-01 -8.99859726e-01 5.41839361e-01 1.28423989e-01
3.42923969e-01 -3.54905963e-01 -7.13269472e-01 -1.87404558e-01
1.38892317e+00 1.85232222e-01 5.63385308e-01 6.23015583e-01
4.31232035e-01 6.88818634e-01 9.02336955e-01 2.92909980e-01
3.90848845e-01 6.69123352e-01 -2.40528844e-02 2.47696742e-01
2.28134841e-01 -1.99484363e-01 6.76094294e-02 8.81112397e-01
-3.60061437e-01 -5.36207557e-01 -1.10848892e+00 7.34260857e-01
-1.98462152e+00 -1.20518434e+00 -4.15368646e-01 1.26897764e+00
1.44198012e+00 1.37084750e-02 2.38779992e-01 1.74757585e-01
7.11261451e-01 1.29525423e-01 4.43336755e-01 -3.01594257e-01
-6.69374049e-01 -6.54871911e-02 5.01573563e-01 5.64916611e-01
-1.29561758e+00 1.42537785e+00 8.31232929e+00 8.49747002e-01
-4.03700173e-01 5.55536211e-01 -5.34326851e-01 6.12738848e-01
-3.90163481e-01 2.05992997e-01 -5.92908323e-01 5.73805757e-02
5.44407666e-01 3.88485119e-02 2.61980742e-02 7.61283159e-01
-9.56603438e-02 -4.69988614e-01 -7.22450018e-01 8.16634834e-01
1.92143977e-01 -1.26338518e+00 -2.74351448e-01 -2.56930232e-01
5.59066892e-01 2.08168760e-01 -9.15311575e-01 1.25686273e-01
5.91091514e-01 -1.75997525e-01 7.71106839e-01 2.55604118e-01
6.98979974e-01 -4.60025579e-01 1.23684931e+00 1.68030456e-01
-1.14735067e+00 2.91677862e-01 -1.07518278e-01 4.46687452e-02
6.82679594e-01 3.14462483e-01 -7.25016057e-01 5.44522941e-01
4.12196785e-01 3.59319985e-01 -2.26970300e-01 7.59683371e-01
-6.73367620e-01 5.59655964e-01 -5.64342797e-01 -1.46396816e-01
2.01764377e-03 3.38695377e-01 8.89013290e-01 1.52437484e+00
8.52051079e-02 5.13906419e-01 5.76057315e-01 3.61992605e-02
-6.57163188e-02 5.69370806e-01 -6.47857785e-01 -2.18349919e-01
4.61396754e-01 9.90798950e-01 -1.25362289e+00 -4.81880546e-01
-2.73054868e-01 9.26708460e-01 7.35459700e-02 9.96660516e-02
-5.05255938e-01 -5.09117603e-01 5.20824119e-02 -2.61784315e-01
-7.30656274e-03 4.22014743e-02 -7.45930895e-02 -1.38490558e+00
-1.45260975e-01 -3.82515639e-01 1.07073426e+00 -1.17423069e+00
-1.32401335e+00 9.56567764e-01 3.90517354e-01 -1.27847302e+00
-4.43553984e-01 -5.53417087e-01 1.01847380e-01 1.73752606e-01
-9.67799306e-01 -1.39643264e+00 2.10202545e-01 5.24441004e-01
2.14723542e-01 -1.22572385e-01 1.28250265e+00 1.00166604e-01
1.40812963e-01 1.95690215e-01 -4.98496205e-01 4.63320047e-01
8.06231022e-01 -1.29471958e+00 -4.08478886e-01 3.11221838e-01
3.72569114e-01 5.13145685e-01 1.39347732e+00 -6.91946745e-01
-6.42866611e-01 -3.40865076e-01 1.68030000e+00 -6.76821530e-01
1.06244612e+00 2.53365993e-01 -8.41477036e-01 1.18985116e+00
7.75378883e-01 -1.05428472e-01 1.25228930e+00 6.77496314e-01
-5.39843321e-01 6.45740569e-01 -1.32036829e+00 8.06369409e-02
1.27359724e+00 -6.82104111e-01 -1.60074997e+00 8.60161483e-01
6.21294320e-01 -7.31024623e-01 -9.92910385e-01 3.21291566e-01
3.28398347e-01 -2.56068110e-01 4.94566977e-01 -7.59498715e-01
-1.30007163e-01 -5.49035907e-01 -3.84415507e-01 -1.27887166e+00
7.84959644e-02 -3.57988983e-01 5.14955521e-01 1.41568100e+00
7.35620737e-01 -1.09284788e-01 3.11712950e-01 5.26512563e-01
-5.37067801e-02 2.46741831e-01 -1.13467872e+00 -5.73833227e-01
-2.97154844e-01 -1.07301819e+00 1.07324809e-01 1.80162513e+00
1.39537466e+00 7.98463106e-01 -2.00189456e-01 -6.87722713e-02
1.63383111e-01 -1.71362907e-01 -2.55880002e-02 -1.27630055e+00
-3.80964875e-02 -5.71189448e-02 -6.73642576e-01 -4.53131616e-01
8.06188285e-01 -8.79526317e-01 1.58643767e-01 -1.82802153e+00
2.38401011e-01 -1.29425734e-01 3.53587000e-03 1.24805677e+00
2.86989152e-01 5.87853730e-01 -2.44622268e-02 1.45211875e-01
-1.10167563e+00 -7.64080510e-02 6.92681074e-01 -1.57450605e-02
-2.93430656e-01 -3.99988025e-01 -5.67053914e-01 1.15954149e+00
6.23266876e-01 -6.38050079e-01 -7.42397876e-03 -3.58199328e-01
1.06313217e+00 1.63027316e-01 -4.47231624e-03 -4.54413891e-01
1.58310816e-01 -4.99538839e-01 1.80812553e-02 -4.83746797e-01
1.42815411e-01 -1.14102805e+00 7.17734173e-02 8.72936249e-02
-5.27396917e-01 -1.57590620e-02 2.95883983e-01 2.15154916e-01
-4.93068159e-01 -6.39242589e-01 3.31107140e-01 -2.89638400e-01
-1.33106363e+00 -6.12067103e-01 -7.70083308e-01 4.38734412e-01
1.01887727e+00 2.54670102e-02 -1.32658318e-01 -2.33059064e-01
-1.55839968e+00 1.62987411e-01 5.27919948e-01 2.34791443e-01
2.53267959e-02 -1.49306417e+00 -2.40832701e-01 -4.37502623e-01
8.12285304e-01 -6.90731585e-01 -4.62376624e-02 6.37039840e-01
-3.54036123e-01 3.78402799e-01 -2.12990180e-01 -2.76444316e-01
-1.45291483e+00 6.28230393e-01 -4.98709530e-02 -2.18237326e-01
-2.37123698e-01 8.47233772e-01 -4.75085407e-01 -5.26534915e-01
2.29789495e-01 -3.04334521e-01 -8.22325587e-01 5.83871663e-01
5.52062869e-01 -2.62087613e-01 -7.23788738e-02 -1.20710671e+00
-7.61863589e-01 3.23605746e-01 3.80830228e-01 -6.14776492e-01
1.09840512e+00 -5.24290383e-01 -4.13451463e-01 1.01641548e+00
7.61346400e-01 3.61953706e-01 5.88923693e-02 -3.61739755e-01
5.71907699e-01 4.77419347e-02 -1.76390305e-01 -9.77461934e-01
-4.96698856e-01 -7.31462240e-02 2.19305813e-01 7.28678524e-01
9.73267436e-01 8.80537689e-01 5.82418859e-01 7.81906009e-01
7.26075649e-01 -1.84299004e+00 -6.41772568e-01 9.02952909e-01
8.14176500e-01 -1.17856431e+00 -4.14676853e-02 -8.49242032e-01
-1.13467467e+00 1.06814337e+00 2.10402772e-01 4.94120717e-01
5.35547614e-01 4.78791982e-01 5.52694201e-01 -5.39280176e-01
-5.83935499e-01 -6.33008957e-01 5.17855138e-02 7.39224017e-01
1.02133131e+00 2.89678574e-05 -1.30432653e+00 6.48803055e-01
-2.26469651e-01 2.30028093e-01 5.18515825e-01 1.26634562e+00
-4.97451454e-01 -1.41434836e+00 -4.17251617e-01 -6.69464767e-02
-8.69966090e-01 -1.10913314e-01 -7.61444449e-01 1.01098013e+00
2.23460540e-01 9.03868437e-01 -2.24860217e-02 -1.44534051e-01
5.11524200e-01 5.30181646e-01 8.26247454e-01 -7.86128104e-01
-3.94897103e-01 2.86167562e-01 1.16190612e+00 -4.14894193e-01
-1.57390344e+00 -7.02459991e-01 -1.59180021e+00 8.77912939e-02
-5.98206639e-01 9.04193878e-01 2.23850697e-01 1.47062218e+00
-4.07526612e-01 4.05733697e-02 -8.46954733e-02 -4.54326898e-01
2.56130368e-01 -1.18460798e+00 -7.20664620e-01 5.31663060e-01
-1.79810196e-01 -6.70991182e-01 -2.76169777e-01 6.35717452e-01] | [10.102240562438965, 9.406415939331055] |
302020f8-e37c-44c0-8061-dc34e365b8e1 | investigation-into-the-effectiveness-of-long | 1603.07893 | null | http://arxiv.org/abs/1603.07893v3 | http://arxiv.org/pdf/1603.07893v3.pdf | Investigation Into The Effectiveness Of Long Short Term Memory Networks For Stock Price Prediction | The effectiveness of long short term memory networks trained by
backpropagation through time for stock price prediction is explored in this
paper. A range of different architecture LSTM networks are constructed trained
and tested. | ['Hengjian Jia'] | 2016-03-25 | null | null | null | null | ['stock-price-prediction'] | ['time-series'] | [-7.68881917e-01 -3.24234903e-01 -3.15828711e-01 -4.40764755e-01
2.62688220e-01 -3.28070283e-01 6.27493799e-01 -5.51609159e-01
-5.43849349e-01 8.57444942e-01 1.44888788e-01 -8.49358916e-01
-2.94043869e-02 -9.80974495e-01 -3.88262331e-01 -2.53062516e-01
-8.37608159e-01 2.29490235e-01 2.06832796e-01 -5.00050306e-01
7.05005050e-01 6.53924823e-01 -7.67655253e-01 4.15180147e-01
5.90143092e-02 1.47729278e+00 -2.78736770e-01 6.53812408e-01
-4.51864809e-01 1.88897419e+00 -9.30226445e-01 -3.36560696e-01
5.23443520e-01 -4.83298972e-02 -4.67410326e-01 -3.81150007e-01
-1.53954208e-01 -5.75143039e-01 -4.69705731e-01 7.98999727e-01
3.16783786e-01 2.31101424e-01 9.91789922e-02 -7.82207608e-01
-1.11420047e+00 9.67123985e-01 1.78485438e-02 1.03068185e+00
-1.73147753e-01 5.42701967e-02 9.21778321e-01 -1.16711354e+00
3.49710047e-01 8.44977081e-01 1.11778629e+00 8.19260553e-02
-9.67435718e-01 -1.08866394e+00 1.18675053e-01 1.17857501e-01
-6.13032997e-01 -5.31172678e-02 8.94564450e-01 -2.57046819e-01
2.04336214e+00 3.73410508e-02 1.01181018e+00 7.91742921e-01
1.03595972e+00 5.75434029e-01 1.31221426e+00 -3.58157694e-01
-5.82201481e-02 2.08816439e-01 5.36352217e-01 6.00711286e-01
2.85031468e-01 1.11905265e+00 -6.21757150e-01 -1.74870834e-01
9.64163125e-01 2.38581419e-01 3.49589825e-01 6.60771728e-01
-8.39535832e-01 9.91271853e-01 7.91886687e-01 9.67981517e-01
-6.90495610e-01 2.66466200e-01 4.79514927e-01 1.05026948e+00
9.60314155e-01 9.97160554e-01 -9.98673797e-01 -2.19889522e-01
-1.29251277e+00 3.43207151e-01 1.04082346e+00 2.78278112e-01
5.94594963e-02 1.38625693e+00 1.58127636e-01 5.84284663e-01
2.54253566e-01 4.38092768e-01 1.32265949e+00 -3.24539125e-01
3.40126157e-01 2.32702225e-01 1.51257485e-01 -1.14173293e+00
-6.89581811e-01 -7.61496902e-01 -5.63706517e-01 7.03305006e-01
6.97879717e-02 -4.75784838e-01 -1.14124513e+00 7.53180146e-01
-7.23967671e-01 5.24680138e-01 6.88028276e-01 3.32751602e-01
7.62512505e-01 1.22945523e+00 1.36375636e-01 -1.57727554e-01
6.94308579e-01 -1.14722705e+00 -9.71778989e-01 -5.78151822e-01
3.87168586e-01 -5.37284493e-01 1.19085215e-01 2.16733307e-01
-1.41333556e+00 -6.58323646e-01 -9.81993318e-01 3.21108192e-01
-1.15066338e+00 -4.08263683e-01 6.67801142e-01 4.11661714e-01
-1.15094936e+00 1.38534379e+00 -7.22896934e-01 6.28541291e-01
1.84271067e-01 5.27204037e-01 3.21101278e-01 1.01706612e+00
-1.94991946e+00 1.42667866e+00 6.73182964e-01 7.67353237e-01
-2.79580116e-01 -3.86919260e-01 -4.31318909e-01 -1.77153736e-01
-8.83749783e-01 -3.07886183e-01 1.83409810e+00 -1.16041291e+00
-1.84186399e+00 4.36931074e-01 3.39165628e-01 -1.25269926e+00
2.73622662e-01 -4.27510729e-03 -1.08760428e+00 -3.15347314e-01
-5.61001539e-01 2.46479273e-01 9.58570659e-01 -2.10368782e-01
-5.95036209e-01 -1.00721821e-01 -5.45930743e-01 -2.96256781e-01
-5.30514605e-02 6.04124665e-01 8.36913466e-01 -1.35218573e+00
4.12766561e-02 -5.00155985e-01 -2.56815732e-01 -7.94192433e-01
2.79325485e-01 -2.89186507e-01 8.17251801e-01 -1.09802079e+00
1.13993967e+00 -1.67008972e+00 -7.70235777e-01 5.15960813e-01
-3.57675493e-01 4.23823029e-01 1.27439782e-01 3.64877939e-01
-6.21640146e-01 2.37251490e-01 1.59236506e-01 -8.20441172e-02
2.29257733e-01 4.16934878e-01 -1.20947421e+00 -5.95319755e-02
2.12860852e-01 1.64532089e+00 -5.08718967e-01 1.59022838e-01
3.95479575e-02 4.18459237e-01 4.11044687e-01 1.67178035e-01
-2.45303437e-01 -6.88491613e-02 -2.27700561e-01 7.22395301e-01
4.12539542e-01 -3.70679572e-02 -4.93825436e-01 3.74121964e-01
-7.38470256e-01 6.38855159e-01 -4.48458761e-01 7.71630943e-01
-1.99334458e-01 1.02818847e+00 -9.56635594e-01 -7.24895537e-01
1.50013876e+00 7.93565392e-01 -4.94243503e-02 -1.20772576e+00
1.58505350e-01 8.60694885e-01 1.70823410e-01 -2.37535834e-01
4.34009254e-01 -8.21097076e-01 3.27597588e-01 4.37026083e-01
-4.60780784e-02 4.45642650e-01 -1.59764245e-01 -8.38515580e-01
6.72933042e-01 8.06399621e-03 -2.09993988e-01 -1.16720058e-01
2.96837121e-01 5.45889139e-03 3.59107435e-01 4.37806517e-01
-2.54126072e-01 -3.94916013e-02 2.70883679e-01 -1.39620471e+00
-9.78446484e-01 -7.12334752e-01 -3.99535805e-01 1.10059905e+00
-8.28746676e-01 4.22048032e-01 5.77722676e-02 1.79838035e-02
2.71597266e-01 1.15102541e+00 -7.00967431e-01 1.34470075e-01
-8.11644733e-01 -7.42206216e-01 8.07973266e-01 9.50738430e-01
7.59165645e-01 -2.06154203e+00 -1.25407374e+00 8.59202981e-01
8.54233265e-01 -7.50463843e-01 2.68730223e-01 6.37791634e-01
-1.73534703e+00 -4.68412399e-01 -1.22586632e+00 -9.02743936e-01
-7.29990453e-02 -6.14089131e-01 1.23763072e+00 -1.74536556e-01
3.32045734e-01 -1.82212681e-01 -6.73917979e-02 -9.68011379e-01
3.56792919e-02 -1.06602289e-01 -1.09044194e-01 -3.41933101e-01
6.57027125e-01 -5.91382682e-01 -4.72060859e-01 -1.06472872e-01
-3.66168857e-01 -4.76202965e-01 6.35491550e-01 7.03299165e-01
1.85903758e-01 3.51625949e-01 7.68813789e-01 -7.88007021e-01
1.01455033e+00 -3.83712471e-01 -1.16239750e+00 2.51424722e-02
-1.21498168e+00 5.71910255e-02 5.36926925e-01 -2.36133486e-01
-1.02190030e+00 -7.66610622e-01 -4.24324989e-01 -4.01499182e-01
4.18996274e-01 1.14886355e+00 1.21198392e+00 -5.65674186e-01
1.00180313e-01 6.75505280e-01 3.34108481e-03 -5.35700679e-01
-2.95273036e-01 7.40011334e-02 2.69056588e-01 3.73220623e-01
5.40324092e-01 -7.74642378e-02 -2.46737957e-01 -7.17863083e-01
-5.84249735e-01 2.18635470e-01 -5.73679328e-01 -4.20209795e-01
4.27453816e-01 -5.82848430e-01 -1.23879537e-01 1.12968898e+00
-1.01592612e+00 -5.50204813e-01 -4.42537695e-01 5.89669824e-01
-1.47396356e-01 -7.30945468e-01 -1.48407400e+00 -1.07989049e+00
-7.56553710e-01 -3.64910513e-01 1.14203477e-02 9.71568972e-02
-2.41006777e-01 -1.85467780e+00 6.56328559e-01 -6.45326793e-01
1.09877384e+00 3.54458630e-01 2.67632008e-01 -1.00230229e+00
-3.04826468e-01 -9.46293652e-01 -3.84779880e-03 4.52176660e-01
-1.27930433e-01 -2.71776225e-02 -1.00234294e+00 -1.60890400e-01
4.65587854e-01 -7.08829686e-02 1.24439943e+00 7.93587804e-01
5.96536458e-01 -5.81227660e-01 1.10294804e-01 6.97001696e-01
1.39986038e+00 7.50060380e-01 8.20432544e-01 1.12382019e+00
2.15211824e-01 3.40170264e-01 3.89166437e-02 1.55663773e-01
-1.05490401e-01 -3.62405241e-01 -1.44405991e-01 4.79748920e-02
6.41210914e-01 -2.66861916e-01 6.90472126e-01 1.18123293e+00
-3.29838425e-01 1.07587673e-01 -1.06523895e+00 1.09437205e-01
-1.37318361e+00 -1.46092319e+00 3.07020575e-01 1.20512605e+00
5.41021705e-01 9.93654251e-01 2.41081908e-01 3.22547376e-01
2.33006194e-01 7.34536052e-01 -6.06631577e-01 -1.07840347e+00
-2.94563740e-01 8.54410172e-01 9.62781727e-01 3.74412358e-01
-1.11041343e+00 9.05827880e-01 8.63257122e+00 1.31481022e-01
-1.75073087e+00 -7.07909539e-02 7.41747916e-01 1.06785655e-01
-1.28832549e-01 -1.57764047e-01 -7.34436393e-01 5.13490677e-01
1.96445847e+00 -1.89219847e-01 -2.35830143e-01 8.17926764e-01
1.06942549e-01 2.98780024e-01 -5.05679250e-01 7.77891815e-01
-2.74330288e-01 -1.56082845e+00 1.13924742e-02 -1.17669441e-01
8.53714287e-01 5.04928470e-01 7.12460399e-01 5.84345698e-01
1.18794516e-01 -1.04510367e+00 8.96075428e-01 1.25185752e+00
8.67363289e-02 -9.98178780e-01 1.02459621e+00 1.20648921e-01
-1.15908599e+00 -4.46317613e-01 -6.44050598e-01 -8.95467877e-01
9.97017995e-02 2.81254709e-01 -3.80342990e-01 -1.67226732e-01
5.85137546e-01 1.00511205e+00 -5.12708724e-01 1.08402336e+00
-2.06849054e-01 8.54168594e-01 -2.27514356e-01 -4.42888439e-01
1.05241692e+00 1.23943982e-03 3.27548772e-01 1.25170112e+00
4.35373962e-01 3.22215050e-01 -3.27880234e-02 8.65752816e-01
3.04013908e-01 -6.16398938e-02 -7.86008298e-01 -6.18727326e-01
-5.18233068e-02 4.13913161e-01 -6.26571715e-01 -4.09812599e-01
-4.62603480e-01 6.05001092e-01 -1.34702086e-01 6.89209461e-01
-5.99594712e-01 -6.72784686e-01 3.12308609e-01 4.90481332e-02
5.56393266e-01 -2.58512080e-01 -6.28552377e-01 -9.25515831e-01
-2.48531222e-01 -1.77953728e-02 6.99687660e-01 -9.39344764e-01
-1.62196636e+00 9.57802117e-01 -2.62399733e-01 -1.17400801e+00
-7.74297595e-01 -1.18783331e+00 -1.28841746e+00 1.20580220e+00
-2.14408159e+00 -6.38618588e-01 6.97129011e-01 3.76774848e-01
5.21656811e-01 -1.01135194e+00 9.66171145e-01 -9.99096259e-02
-5.79539895e-01 2.60063410e-01 2.96087801e-01 4.60826784e-01
-1.61129475e-01 -1.21560824e+00 1.23813796e+00 5.65187752e-01
1.31678388e-01 5.23482978e-01 6.45348549e-01 -8.55408072e-01
-6.50175452e-01 -1.22757304e+00 1.29758835e+00 7.51143089e-03
1.27080119e+00 4.29775864e-01 -1.18040788e+00 1.35520196e+00
7.49661267e-01 -8.22528750e-02 3.70626718e-01 -2.78704882e-01
-2.95788467e-01 -1.53905153e-01 -9.35983300e-01 7.49580115e-02
8.53800029e-02 -6.01867318e-01 -1.37436938e+00 -1.38843268e-01
5.37868619e-01 -2.78992593e-01 -9.16010737e-01 2.49390841e-01
6.83821797e-01 -1.01379633e+00 8.86263251e-01 -8.30895603e-01
1.66984588e-01 4.96372312e-01 2.23314732e-01 -1.26775396e+00
-4.91245449e-01 -5.66010594e-01 -7.25289404e-01 3.96557778e-01
1.05542374e+00 -1.67192197e+00 8.48787189e-01 9.76278603e-01
-2.29855865e-01 -8.63850415e-01 -7.82409191e-01 -1.15892935e+00
5.72515607e-01 -5.60328364e-01 6.22160971e-01 1.04216540e+00
2.64535338e-01 1.67829961e-01 -2.33157232e-01 -2.87583023e-01
2.78169811e-01 5.49427688e-01 -4.20004278e-01 -1.23676670e+00
1.00222267e-01 -9.08207715e-01 -4.08248991e-01 -6.47406161e-01
3.71897757e-01 -5.94456553e-01 -5.31280220e-01 -1.21434569e+00
-5.49385369e-01 -6.32555559e-02 -1.23734891e+00 4.45423692e-01
4.58753437e-01 5.23305424e-02 1.94104984e-01 5.42385221e-01
3.88706140e-02 2.31256440e-01 7.91670382e-01 -1.81456476e-01
-2.32709259e-01 5.10186911e-01 2.38810033e-01 9.94450867e-01
1.54103827e+00 -2.86820143e-01 -7.29636475e-02 -2.99880981e-01
5.01273632e-01 1.80832565e-01 3.00324976e-01 -9.90247369e-01
2.91213423e-01 9.17475075e-02 1.05538201e+00 -1.06049633e+00
3.40300173e-01 -6.24154687e-01 -5.51379509e-02 1.14486241e+00
-5.51482260e-01 1.01805520e+00 5.87669969e-01 1.46338716e-01
-7.80456245e-01 -3.85720193e-01 5.32387614e-01 -7.98623741e-01
-1.20443594e+00 2.41530567e-01 -5.42275429e-01 -3.10896128e-01
7.52243400e-01 -6.74183547e-01 -4.03384641e-02 -2.88512886e-01
-9.17843342e-01 1.79397553e-01 -5.25701165e-01 2.92665064e-01
1.15465760e+00 -1.49396038e+00 -5.13368487e-01 5.44566691e-01
-5.79722285e-01 -1.28189981e+00 -3.40117753e-01 4.66694653e-01
-8.77117813e-01 1.34132695e+00 -7.09520757e-01 2.04427719e-01
-5.51746905e-01 4.22084332e-01 1.03409886e+00 -3.33151102e-01
-8.00439835e-01 1.23298156e+00 -7.52730966e-01 -3.40279266e-02
2.54629582e-01 -7.77639091e-01 -8.22631955e-01 2.21490055e-01
7.29226470e-01 3.65219295e-01 -2.32295692e-03 -5.92613518e-01
1.90999061e-01 4.02641535e-01 9.87727866e-02 -3.43531787e-01
1.86072123e+00 4.07229096e-01 -4.26634014e-01 1.54175889e+00
1.11741126e+00 -8.43551040e-01 -7.00684905e-01 -7.57030621e-02
1.13020229e+00 -2.53528267e-01 3.86327684e-01 -8.96889091e-01
-1.16399407e+00 7.61404395e-01 8.72765839e-01 7.02931106e-01
9.06245828e-01 -8.50679040e-01 1.54529846e+00 6.91733718e-01
8.29938278e-02 -1.36173213e+00 -1.26262709e-01 1.05060220e+00
9.89585876e-01 -8.97490501e-01 -3.26625526e-01 7.80562639e-01
-2.48232752e-01 1.97203898e+00 1.03343232e-02 -1.15163386e+00
1.45968187e+00 2.95103014e-01 3.85869861e-01 -4.54501331e-01
-1.23356664e+00 2.62178872e-02 3.31813574e-01 -1.29390031e-01
7.05107510e-01 -1.39978573e-01 1.59787625e-01 4.24823910e-01
-6.23670757e-01 3.71040553e-01 3.49813759e-01 1.07944703e+00
-6.80845559e-01 -7.62639821e-01 -3.84368151e-01 5.98930120e-01
-1.03241360e+00 -3.31750244e-01 -4.18812454e-01 6.86018109e-01
-4.24808562e-01 3.44043493e-01 7.00906932e-01 -3.89817834e-01
4.74018574e-01 7.07776546e-01 2.84559140e-03 -2.68767357e-01
-1.13230932e+00 -8.09121318e-03 1.80256233e-01 -1.32890716e-01
-4.30925071e-01 -6.02107167e-01 -1.19751728e+00 -4.75482047e-01
-2.72286862e-01 1.62309855e-01 4.22779471e-01 5.69649100e-01
-1.93349048e-01 6.09366596e-01 7.73053944e-01 -7.60732234e-01
-8.29201698e-01 -1.05357385e+00 -1.09509599e+00 -2.18453377e-01
9.56296444e-01 -2.69260228e-01 -4.96845245e-01 -2.69957721e-01] | [4.461664199829102, 4.22921085357666] |
7085a109-23b6-4743-9314-2e9169ae2ba6 | deep-latent-variable-models-for-semi | 2301.02275 | null | https://arxiv.org/abs/2301.02275v1 | https://arxiv.org/pdf/2301.02275v1.pdf | Deep Latent Variable Models for Semi-supervised Paraphrase Generation | This paper explores deep latent variable models for semi-supervised paraphrase generation, where the missing target pair is modelled as a latent paraphrase sequence. We present a novel unsupervised model named variational sequence auto-encoding reconstruction (VSAR), which performs latent sequence inference given an observed text. To leverage information from text pairs, we introduce a supervised model named dual directional learning (DDL). Combining VSAR with DDL (DDL+VSAR) enables us to conduct semi-supervised learning; however, the combined model suffers from a cold-start problem. To combat this issue, we propose to deal with better weight initialisation, leading to a two-stage training scheme named knowledge reinforced training. Our empirical evaluations suggest that the combined model yields competitive performance against the state-of-the-art supervised baselines on complete data. Furthermore, in scenarios where only a fraction of the labelled pairs are available, our combined model consistently outperforms the strong supervised model baseline (DDL and Transformer) by a significant margin. | ['Noura Al Moubayed', 'Lei Shi', 'Olanrewaju Tahir Aduragba', 'Zhongtian Sun', 'Anoushka Harit', 'Alexandra I. Cristea', 'Jialin Yu'] | 2023-01-05 | null | null | null | null | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 5.78628480e-01 3.29746246e-01 -5.40048838e-01 -2.23022312e-01
-1.24020028e+00 -7.15132475e-01 1.09993958e+00 -4.73340834e-03
-2.44197190e-01 7.95311928e-01 6.14345789e-01 -3.16906720e-01
3.19653630e-01 -4.82484043e-01 -1.03286266e+00 -5.62243521e-01
5.82415879e-01 7.18018115e-01 3.11564300e-02 -3.58273461e-02
1.78038612e-01 -1.22678392e-02 -1.19067681e+00 6.21084869e-01
8.85606885e-01 4.49836046e-01 3.76566291e-01 4.65987235e-01
-3.59672338e-01 1.03294265e+00 -2.55900055e-01 -7.98841596e-01
2.45096684e-01 -7.87292898e-01 -1.07824683e+00 1.38231948e-01
1.90156683e-01 -4.19042528e-01 -2.19150722e-01 7.49440432e-01
3.61708105e-01 -6.25311537e-03 9.11521316e-01 -1.35555255e+00
-4.80177939e-01 9.37782168e-01 -6.13222480e-01 -8.00017864e-02
5.89312434e-01 1.12112770e-02 1.35684597e+00 -1.13336968e+00
7.81104326e-01 1.22876811e+00 6.89114928e-01 3.76039952e-01
-1.59942710e+00 -4.11745667e-01 1.05984092e-01 2.78420210e-01
-9.42458987e-01 -5.70834875e-01 9.87648129e-01 -4.14769709e-01
1.08030534e+00 -9.01646242e-02 4.23224151e-01 1.80495596e+00
-1.23346239e-01 1.37946951e+00 1.13456726e+00 -7.76205063e-01
2.32523367e-01 2.37999454e-01 1.62457317e-01 6.91562116e-01
-7.08851889e-02 3.15275416e-02 -7.81136930e-01 -4.31123614e-01
4.26845700e-01 1.54620809e-02 -5.98801747e-02 -8.47639441e-01
-1.28376198e+00 9.80917573e-01 8.27186555e-02 5.58158979e-02
-4.85386163e-01 -2.45950520e-02 6.33159876e-01 5.12900710e-01
6.59039140e-01 3.54235202e-01 -2.74235696e-01 -3.42705786e-01
-1.56430984e+00 2.98592210e-01 8.12701225e-01 1.03142166e+00
7.41476476e-01 -8.40845555e-02 -4.15322095e-01 1.10028636e+00
2.48432532e-01 2.41931289e-01 6.61281705e-01 -8.80652666e-01
8.60893667e-01 3.26890439e-01 1.73403829e-01 -3.34248185e-01
1.16967961e-01 -5.08174002e-01 -6.30736291e-01 -1.78606585e-01
2.75528848e-01 -3.58159468e-02 -9.91219759e-01 1.80574846e+00
3.87050845e-02 4.14479613e-01 2.27814510e-01 6.32433355e-01
4.97241974e-01 7.39690840e-01 -1.59632683e-01 -4.34614241e-01
8.96834791e-01 -1.44145441e+00 -6.04900420e-01 -3.98367733e-01
7.31906474e-01 -6.96826935e-01 1.07796121e+00 3.25837195e-01
-1.29404175e+00 -4.52485979e-01 -1.05222368e+00 -2.03146726e-01
8.66623446e-02 1.11864768e-01 3.36099774e-01 2.73196012e-01
-9.73975539e-01 7.24327981e-01 -1.06401849e+00 -3.72429818e-01
2.30772480e-01 -7.90539309e-02 -2.91606367e-01 -1.94563940e-01
-1.23137963e+00 8.71187568e-01 4.69227970e-01 -2.14293331e-01
-1.10333192e+00 -7.16190755e-01 -1.09840441e+00 2.27997340e-02
3.89311671e-01 -9.23883736e-01 1.34401619e+00 -9.96020675e-01
-1.65378022e+00 8.72358024e-01 -7.92860568e-01 -7.59721100e-01
8.64621103e-01 -4.51865852e-01 2.44799986e-01 1.26895100e-01
2.51892060e-01 5.97366273e-01 1.03035915e+00 -1.40028214e+00
-1.91013470e-01 5.21975197e-02 -8.07654560e-02 2.55869567e-01
-2.65733749e-01 -4.76143621e-02 -4.82588917e-01 -8.99992406e-01
-6.87089860e-02 -1.08860898e+00 -2.55187243e-01 -2.88197368e-01
-4.33865696e-01 -2.06776470e-01 6.13474429e-01 -8.86860490e-01
1.11403584e+00 -1.83133137e+00 7.43336618e-01 -1.11020943e-02
7.35865012e-02 3.97774875e-01 -3.85335863e-01 9.15803254e-01
-2.28367507e-01 -1.52173966e-01 -5.38806260e-01 -1.15197706e+00
6.34793118e-02 2.85946995e-01 -6.56115294e-01 4.99932356e-02
1.49312407e-01 1.28953254e+00 -1.19950223e+00 -5.40273607e-01
1.15839802e-01 1.37474805e-01 -5.92214167e-01 3.94580245e-01
-4.72985685e-01 2.33796269e-01 -1.06137395e-01 3.58049095e-01
4.24858838e-01 -5.95504045e-01 2.94131726e-01 8.72105733e-02
3.45687360e-01 5.91228008e-01 -7.68229246e-01 2.15744281e+00
-6.35413408e-01 5.64777672e-01 -2.46133357e-01 -1.05350137e+00
7.18840480e-01 5.01167357e-01 2.06628472e-01 -3.29158604e-01
-3.33941668e-01 1.67128742e-01 -4.85743225e-01 -2.98195601e-01
5.27924776e-01 -2.19513327e-01 5.19060642e-02 7.62324035e-01
2.96944171e-01 1.40324160e-01 9.19240639e-02 8.37462485e-01
1.12250328e+00 6.42069459e-01 2.87954628e-01 1.01638846e-01
3.85181725e-01 -1.47434682e-01 5.06326377e-01 1.10371530e+00
4.79134917e-03 6.19430840e-01 6.94192290e-01 4.51663360e-02
-1.41064811e+00 -1.22969949e+00 3.17602754e-01 8.69265258e-01
-3.93962190e-02 -6.74631238e-01 -6.96524918e-01 -8.93526375e-01
-1.33275852e-01 8.96202683e-01 -5.04881740e-01 -1.27943784e-01
-4.95980114e-01 -5.27259409e-01 5.53488135e-01 6.31926000e-01
3.71916115e-01 -9.88348365e-01 -9.27592888e-02 2.35690340e-01
-6.61349535e-01 -1.14423490e+00 -2.88056046e-01 2.20646366e-01
-9.27272499e-01 -6.56983137e-01 -1.02542067e+00 -6.97961748e-01
4.66103196e-01 3.14478576e-01 1.22046483e+00 -3.21684688e-01
2.65979350e-01 6.40729740e-02 -6.84772372e-01 -1.17883477e-02
-8.44703078e-01 3.88797253e-01 1.63718332e-02 -5.44040836e-02
2.83884466e-01 -7.76588321e-01 -4.36604291e-01 -3.03651560e-02
-7.70971954e-01 5.92041254e-01 8.24020505e-01 1.33156788e+00
3.96667331e-01 -6.42038643e-01 6.08347774e-01 -1.02379799e+00
5.82069099e-01 -6.38077974e-01 -3.13615203e-01 4.58794385e-01
-6.52406693e-01 5.28312802e-01 7.39663482e-01 -3.77082080e-01
-1.10660648e+00 9.61276740e-02 -1.03677861e-01 -9.00755823e-01
-2.30485685e-02 7.44128883e-01 -4.58188690e-02 5.18797278e-01
5.15862226e-01 6.16168261e-01 1.68818325e-01 -6.87653303e-01
5.94373822e-01 8.48574281e-01 5.37477672e-01 -3.64618570e-01
8.92271757e-01 3.09025407e-01 -2.07289472e-01 -5.22711456e-01
-9.85754132e-01 -5.84367454e-01 -8.20891500e-01 2.00141832e-01
4.98204589e-01 -1.15954792e+00 6.45674393e-02 5.78593373e-01
-1.15751469e+00 -5.19827008e-01 -1.64219916e-01 2.99853861e-01
-7.72297263e-01 9.00544822e-01 -8.22316825e-01 -7.09063113e-01
-4.73929584e-01 -9.88642931e-01 1.29170907e+00 -4.39909488e-01
-5.09073377e-01 -1.10871625e+00 2.97355711e-01 6.75835669e-01
8.52341130e-02 -3.88188995e-02 8.63992572e-01 -9.59932446e-01
-5.33801794e-01 -2.37041675e-02 -5.07834107e-02 4.67943311e-01
7.25169703e-02 -3.12503129e-01 -8.64504457e-01 -5.13670683e-01
-7.43523836e-02 -7.83086479e-01 1.25225306e+00 9.70164500e-03
8.21921110e-01 -3.32713127e-01 -3.60346973e-01 5.98584533e-01
1.07469594e+00 -4.20518219e-01 4.44188118e-01 2.98598707e-01
7.90385246e-01 5.14276862e-01 5.12626529e-01 3.86248261e-01
4.46450442e-01 8.80652308e-01 8.84948149e-02 -7.36188218e-02
-3.93756442e-02 -7.73801208e-01 6.78594470e-01 1.00647342e+00
3.76954347e-01 -3.46663177e-01 -7.70814300e-01 7.66287267e-01
-2.18316031e+00 -1.19218516e+00 4.22432348e-02 2.13122082e+00
1.12563586e+00 1.85749784e-01 2.77991563e-01 2.13562638e-01
4.65353608e-01 4.21215951e-01 -5.13058126e-01 -3.09368789e-01
-1.24673806e-01 1.90662459e-01 3.80009443e-01 6.95362687e-01
-8.33465695e-01 1.11195707e+00 6.36171913e+00 1.14523184e+00
-7.24912286e-01 2.55893528e-01 2.10531101e-01 -2.25703567e-01
-7.36224115e-01 3.79505366e-01 -7.66455770e-01 7.17394769e-01
1.05239701e+00 -6.82815239e-02 3.99307847e-01 6.73708379e-01
2.65533268e-01 1.24313317e-01 -1.20056772e+00 8.86309922e-01
3.30996722e-01 -1.41307449e+00 3.29670608e-01 -1.58910230e-02
9.07054961e-01 1.79681629e-01 -1.87351912e-01 5.28900206e-01
6.36702299e-01 -6.40034974e-01 4.94327426e-01 3.34532887e-01
6.89548790e-01 -5.77750444e-01 6.67953074e-01 7.48355448e-01
-8.72191370e-01 5.59543595e-02 -2.46239603e-01 5.62622994e-02
5.44176698e-01 5.60172439e-01 -9.13223922e-01 8.93289447e-01
1.28922731e-01 1.09174764e+00 -5.51707268e-01 6.86553836e-01
-5.97115576e-01 9.87709701e-01 -8.15034285e-02 2.21116647e-01
2.82010734e-01 -2.26879925e-01 7.45412111e-01 1.17048180e+00
1.02448668e-02 -3.42977732e-01 1.05066553e-01 9.78704870e-01
-1.88405886e-01 -1.19593963e-01 -5.77276289e-01 -1.27643317e-01
5.57769835e-01 9.00219679e-01 -3.70386153e-01 -6.14311814e-01
-4.24275070e-01 1.75070512e+00 6.43865883e-01 3.98437142e-01
-7.15683520e-01 -1.62896663e-01 2.89316744e-01 -1.77378267e-01
3.99748325e-01 -6.82536736e-02 -1.98764503e-01 -1.64009106e+00
1.35209292e-01 -9.21172023e-01 3.05439800e-01 -9.07019556e-01
-1.34902906e+00 3.69639277e-01 2.66651690e-01 -1.13508725e+00
-9.23473895e-01 -1.16763078e-01 -6.03814065e-01 9.79324043e-01
-1.62818551e+00 -1.43728781e+00 3.82241383e-02 3.54021728e-01
9.62977290e-01 -2.45128930e-01 7.02274859e-01 -9.00424421e-02
-3.23994815e-01 8.70474756e-01 5.04951715e-01 5.81619181e-02
7.75015891e-01 -1.38015568e+00 9.63666379e-01 1.01951587e+00
3.90488148e-01 7.24531233e-01 8.45482767e-01 -7.72262275e-01
-1.07234704e+00 -1.06315994e+00 1.27722859e+00 -6.82589889e-01
6.14839137e-01 -6.78743362e-01 -1.03270316e+00 9.39231813e-01
2.45424449e-01 -4.86606807e-01 7.07761824e-01 1.53120846e-01
-5.57500422e-01 4.38034505e-01 -7.14752138e-01 6.78877771e-01
1.05486429e+00 -9.95973885e-01 -1.07204807e+00 4.42327142e-01
1.01812446e+00 -1.99278593e-01 -4.68186587e-01 2.60379136e-01
5.31276584e-01 -9.33342814e-01 8.90943825e-01 -6.53752565e-01
1.02886021e+00 4.05690223e-02 1.11535918e-02 -1.38462806e+00
-2.28058562e-01 -8.05289567e-01 -6.36925638e-01 1.33532846e+00
5.03686190e-01 -3.48161906e-01 9.04205620e-01 2.24014595e-01
-1.23483976e-02 -6.40904725e-01 -7.39647985e-01 -1.05587280e+00
1.66950449e-01 -1.13437466e-01 2.01049298e-01 1.02146471e+00
2.82035649e-01 7.17128992e-01 -8.34715128e-01 -2.05268458e-01
8.15161347e-01 2.72756279e-01 9.38455820e-01 -7.91692734e-01
-8.67074251e-01 -6.60665482e-02 7.09141605e-03 -1.51829123e+00
5.71326137e-01 -1.17271292e+00 -1.49104416e-01 -1.72527099e+00
6.51161194e-01 3.19503695e-02 -2.56945819e-01 4.06236142e-01
-5.66346109e-01 1.42105982e-01 1.99612945e-01 5.27308047e-01
-5.89510679e-01 8.81527781e-01 6.06915355e-01 2.36457586e-02
-2.71678090e-01 -4.73330952e-02 -4.71817464e-01 4.08880830e-01
5.67707121e-01 -4.54756081e-01 -6.03657424e-01 -3.75576168e-01
1.31777257e-01 3.54538500e-01 4.32193488e-01 -4.65224326e-01
2.32488811e-01 7.94593617e-02 4.51276936e-02 -8.15262258e-01
4.71641570e-01 -4.03952897e-01 1.79955482e-01 2.45639741e-01
-9.07984793e-01 -1.02126084e-01 -1.78793371e-01 6.90126479e-01
-3.04459870e-01 -4.23997998e-01 5.68527758e-01 -5.38438559e-02
-3.42932492e-01 -7.88636059e-02 -3.50179464e-01 1.49697170e-01
7.07691550e-01 -2.73541421e-01 -1.46093234e-01 -6.23959899e-01
-5.04783690e-01 3.44432861e-01 6.88379943e-01 5.68357646e-01
4.25289184e-01 -1.29451585e+00 -8.04813623e-01 2.32943684e-01
3.00542027e-01 -2.04202503e-01 9.75795910e-02 6.55804396e-01
-3.50182154e-03 5.61333597e-01 1.74964666e-01 -5.29922366e-01
-1.23578465e+00 7.20829070e-01 -6.50205910e-02 -7.96175897e-01
-6.81429803e-01 9.15928125e-01 5.93909994e-02 -4.86083627e-01
1.24858178e-01 1.20220631e-01 3.97147611e-02 9.32526663e-02
2.05838129e-01 2.88916647e-01 -4.90537882e-02 -5.30689836e-01
-5.61290644e-02 -9.49748419e-03 -6.27730072e-01 -5.22107005e-01
1.20063972e+00 -3.03970933e-01 1.30184278e-01 6.42876387e-01
1.28518236e+00 -9.23851281e-02 -1.33184373e+00 -7.56728828e-01
1.19643398e-01 -3.53305131e-01 -6.89872429e-02 -6.71493948e-01
-4.35465991e-01 9.19615805e-01 -1.58656016e-02 -2.97458589e-01
7.24752188e-01 -9.40335616e-02 1.02616048e+00 3.71729046e-01
3.01101983e-01 -8.47078919e-01 3.28094333e-01 6.29938722e-01
6.41574740e-01 -1.29387712e+00 -2.01071613e-02 -2.59998530e-01
-9.91157591e-01 6.61059916e-01 2.70049363e-01 -7.79445916e-02
1.75464645e-01 -3.78285423e-02 -3.26316170e-02 1.74772531e-01
-1.17371500e+00 -6.43538982e-02 1.26838624e-01 3.19418013e-01
3.14469993e-01 -3.66421849e-01 -2.62943685e-01 4.45109308e-01
-2.46198550e-01 1.62674099e-01 4.66534108e-01 1.00825250e+00
-1.55630499e-01 -1.62419975e+00 -1.34082735e-02 4.70151305e-01
-4.46854055e-01 -5.54136574e-01 -5.30326784e-01 3.25191945e-01
-5.53145945e-01 9.21716094e-01 -1.29611596e-01 -2.73514599e-01
-3.41189764e-02 5.55556118e-01 4.03536648e-01 -6.63296938e-01
-4.19176698e-01 1.00774713e-01 7.01104999e-02 -4.83604789e-01
-4.19514507e-01 -7.24542320e-01 -6.89758241e-01 -7.37256706e-02
-3.71185660e-01 3.27670753e-01 3.51706803e-01 1.25064266e+00
4.14038062e-01 2.75733411e-01 7.52500892e-01 -8.98837805e-01
-9.92379427e-01 -1.07906032e+00 -2.91423470e-01 5.86723149e-01
4.65080202e-01 -4.95189309e-01 -4.58439142e-01 3.07982862e-01] | [11.647695541381836, 9.209799766540527] |
8f698e6e-9ebc-46e9-869c-cd95f64c926e | multi-team-a-multi-attention-multi-decoder | null | null | https://aclanthology.org/W19-4206 | https://aclanthology.org/W19-4206.pdf | Multi-Team: A Multi-attention, Multi-decoder Approach to Morphological Analysis. | This paper describes our submission to SIGMORPHON 2019 Task 2: Morphological analysis and lemmatization in context. Our model is a multi-task sequence to sequence neural network, which jointly learns morphological tagging and lemmatization. On the encoding side, we exploit character-level as well as contextual information. We introduce a multi-attention decoder to selectively focus on different parts of character and word sequences. To further improve the model, we train on multiple datasets simultaneously and use external embeddings for initialization. Our final model reaches an average morphological tagging F1 score of 94.54 and a lemma accuracy of 93.91 on the test data, ranking respectively 3rd and 6th out of 13 teams in the SIGMORPHON 2019 shared task. | ['Ahmet {\\"U}st{\\"u}n', 'Rob van der Goot', 'Gosse Bouma', 'Gertjan van Noord'] | 2019-08-01 | null | null | null | ws-2019-8 | ['morphological-tagging'] | ['natural-language-processing'] | [ 1.50377288e-01 -6.69615641e-02 -2.45692059e-02 -2.55032599e-01
-1.18681288e+00 -9.96938407e-01 2.19186768e-01 6.68065727e-01
-1.24398911e+00 6.49127245e-01 4.09381896e-01 -4.19290990e-01
4.22910959e-01 -4.93818969e-01 -8.89474988e-01 -4.57411021e-01
2.43787825e-01 5.03600240e-01 5.84608950e-02 1.90601200e-01
1.12592839e-01 3.52427930e-01 -5.54860771e-01 7.32131183e-01
7.47927547e-01 5.74947298e-01 2.81515807e-01 8.19792807e-01
-1.35120139e-01 3.00980181e-01 -6.62623107e-01 -1.20019877e+00
1.93346277e-01 -1.94609314e-01 -8.24540377e-01 -3.29823911e-01
7.15814114e-01 8.29057544e-02 -2.33792782e-01 1.27547908e+00
6.75777853e-01 1.08442090e-01 4.90286142e-01 -4.85601127e-01
-1.19052374e+00 1.34765029e+00 -3.84731382e-01 5.66438615e-01
8.47135931e-02 1.94522172e-01 1.46189165e+00 -1.15752256e+00
7.87930608e-01 9.98148441e-01 8.36974919e-01 4.67879951e-01
-1.33272469e+00 -4.45087880e-01 4.61093992e-01 4.33045663e-02
-1.15558350e+00 -3.03463906e-01 2.55710393e-01 -3.48000020e-01
1.65035427e+00 -2.58790791e-01 4.76340592e-01 1.33159459e+00
3.71993929e-01 8.93332481e-01 9.63226855e-01 -4.08797354e-01
-1.59751490e-01 -3.67750525e-01 2.37436339e-01 8.38778615e-01
3.48411113e-01 -3.02334666e-01 -6.13583446e-01 9.90704969e-02
5.93286812e-01 -4.98461783e-01 2.06071153e-01 3.21582347e-01
-1.41107070e+00 7.51688302e-01 2.17181563e-01 3.16172659e-01
-4.67825770e-01 1.94075003e-01 6.36822879e-01 2.46840045e-01
3.81847352e-01 7.86910713e-01 -7.93682456e-01 -3.05655986e-01
-7.70565748e-01 -2.53503919e-02 6.88150942e-01 8.78887177e-01
4.57608014e-01 1.02911450e-01 -2.63553739e-01 1.00259078e+00
-2.91604139e-02 4.75230902e-01 5.75357437e-01 -4.72775966e-01
7.37819791e-01 2.20456660e-01 -3.69549066e-01 -2.73917586e-01
-4.02114630e-01 -3.30000609e-01 -4.38921630e-01 -2.16903150e-01
1.76061586e-01 -5.27880788e-01 -1.10293937e+00 1.80552804e+00
-4.71085943e-02 -1.14735961e-02 3.92570719e-02 4.34962541e-01
7.39433348e-01 5.71605206e-01 5.46241939e-01 8.30288157e-02
1.49785805e+00 -1.18255758e+00 -6.79195940e-01 -4.78083104e-01
8.56155694e-01 -1.10437119e+00 1.11306071e+00 3.16962481e-01
-1.34580636e+00 -5.49987197e-01 -9.67562258e-01 -4.56318170e-01
-4.95485336e-01 5.08757949e-01 3.92487139e-01 2.58392453e-01
-1.09628117e+00 7.16196060e-01 -1.00225759e+00 -3.97700638e-01
2.84250796e-01 4.37165022e-01 -5.34037113e-01 5.19237630e-02
-1.02054238e+00 1.09093845e+00 9.23565567e-01 -1.50495440e-01
-6.46402836e-01 -8.27406168e-01 -8.80246580e-01 -2.58409623e-02
-1.38645172e-01 -4.72545534e-01 1.24971473e+00 -6.09618247e-01
-1.43936336e+00 1.28805912e+00 1.33260101e-01 -7.29287446e-01
1.47387177e-01 -7.30790794e-01 -4.57324594e-01 9.22705382e-02
1.29926294e-01 9.55998182e-01 2.27644265e-01 -8.65990758e-01
-8.19006741e-01 1.04611188e-01 -1.63914055e-01 1.52621761e-01
-2.97773004e-01 5.42836249e-01 -4.98685092e-01 -9.58245993e-01
-2.26528212e-01 -8.67488146e-01 -2.27720037e-01 -6.01298094e-01
-4.52798665e-01 -2.80881137e-01 1.64631866e-02 -1.22950006e+00
1.16973042e+00 -2.19144058e+00 3.23030174e-01 -2.23555192e-01
-1.37260169e-01 5.93349397e-01 -6.36021733e-01 7.12631881e-01
-2.51232177e-01 3.84802967e-01 -4.22235072e-01 -8.24436903e-01
1.02067724e-01 1.92142367e-01 -9.12898183e-02 3.95368814e-01
7.65979648e-01 1.11957908e+00 -8.93040895e-01 -4.71324742e-01
-7.90401176e-02 2.98849463e-01 -6.65913463e-01 2.33291447e-01
-2.32142448e-01 9.17146653e-02 1.29053608e-01 5.71931899e-01
5.44158161e-01 2.59007782e-01 4.04159904e-01 -1.26096979e-01
-2.60550678e-01 8.76032174e-01 -6.81977451e-01 2.15542197e+00
-8.59531283e-01 5.82849383e-01 -4.91469493e-03 -6.70246363e-01
6.98823810e-01 4.60673898e-01 2.69338250e-01 -4.80658412e-01
2.69201756e-01 3.22895646e-01 3.30313414e-01 -4.22402740e-01
5.64293385e-01 -3.20826292e-01 -5.47025323e-01 2.87826389e-01
6.23216152e-01 1.62473068e-01 3.95529628e-01 1.86984703e-01
1.39749336e+00 2.57667214e-01 4.13732439e-01 -3.24766427e-01
2.18293667e-01 -1.72479704e-01 7.61775315e-01 5.16024530e-01
-5.76720536e-02 6.56502426e-01 5.70001304e-01 -2.09435448e-01
-1.19172490e+00 -1.34564841e+00 -8.50593075e-02 1.42450202e+00
-3.84714156e-01 -5.47897637e-01 -7.14554846e-01 -1.09610784e+00
5.17446734e-02 8.89908969e-01 -7.12964714e-01 -1.81185052e-01
-1.17692804e+00 -7.50760496e-01 1.09879708e+00 9.19236958e-01
2.27639675e-01 -1.57258797e+00 -1.05960064e-01 4.64976221e-01
-1.11127816e-01 -1.40348911e+00 -6.11027539e-01 8.04947615e-01
-5.63261449e-01 -7.46513903e-01 -3.25944275e-01 -1.40353954e+00
3.38506222e-01 -4.58149135e-01 1.13823295e+00 -1.04624882e-01
-1.90017503e-02 -9.22354534e-02 -4.77178067e-01 -7.24245384e-02
-3.63702714e-01 6.56615257e-01 -3.22265774e-02 -7.81283155e-02
2.78838664e-01 -5.39670408e-01 -2.72090677e-02 -4.66089338e-01
-6.75081491e-01 -1.20263301e-01 1.00003922e+00 8.55242312e-01
7.46110737e-01 -6.16928399e-01 5.15597165e-01 -1.04353952e+00
5.14115989e-01 -3.60720485e-01 -5.23397684e-01 2.80867040e-01
-3.84707563e-02 8.67461972e-03 9.26737428e-01 -5.24217963e-01
-8.83784175e-01 2.33874992e-01 -7.34851003e-01 -7.90081546e-02
-1.71207070e-01 6.64703250e-01 -4.86332536e-01 2.88437307e-01
3.25474054e-01 4.08007316e-02 -6.54909909e-01 -1.05020237e+00
5.95784485e-01 4.43375558e-01 9.54008162e-01 -6.66998148e-01
6.13004327e-01 -5.58131672e-02 -3.81624579e-01 -6.59098268e-01
-1.09191310e+00 -4.11062837e-01 -1.14926124e+00 5.63718319e-01
1.31157076e+00 -9.80004013e-01 -3.63328665e-01 4.43225890e-01
-1.49633563e+00 -7.85544693e-01 -3.47991049e-01 4.91558164e-01
-2.68646896e-01 3.50632906e-01 -1.07801199e+00 -2.02827483e-01
-6.22229695e-01 -9.23928499e-01 9.69003379e-01 -8.28132853e-02
-2.50148177e-01 -1.16651678e+00 2.82240450e-01 -3.84674445e-02
-3.91856655e-02 3.59998234e-02 1.27095652e+00 -1.20654440e+00
-1.52315959e-01 -1.43623024e-01 -2.73840427e-01 5.94313979e-01
6.51805010e-03 3.48930731e-02 -6.18466616e-01 -2.82210439e-01
-4.75570172e-01 -4.67054337e-01 1.28008115e+00 3.33522595e-02
9.87077832e-01 -5.34275435e-02 -8.26117173e-02 8.47881436e-01
1.55178010e+00 -6.17026957e-03 4.07464564e-01 4.48195398e-01
8.53521645e-01 2.31437549e-01 3.69407356e-01 1.49917960e-01
2.55293518e-01 4.51799214e-01 2.39534929e-01 2.21289337e-01
-7.13933170e-01 -5.69149435e-01 7.40307271e-01 1.34844339e+00
4.73606437e-01 -6.02956772e-01 -1.10197341e+00 9.66228008e-01
-1.42048407e+00 -6.18711948e-01 -1.63461775e-01 1.91808224e+00
1.12566304e+00 2.32720912e-01 -1.71572268e-02 -2.58034855e-01
8.87996793e-01 9.45769027e-02 -2.55356103e-01 -9.09328163e-01
-3.36673856e-01 9.04624403e-01 6.18979037e-01 6.99750364e-01
-1.48834598e+00 1.76344204e+00 6.30462265e+00 8.78280759e-01
-8.82225275e-01 3.46392900e-01 4.67464104e-02 -2.07003713e-01
-3.39572638e-01 1.54786219e-03 -1.17229211e+00 5.36280930e-01
1.26208425e+00 -5.51566295e-02 2.56673515e-01 4.86625224e-01
-4.47560519e-01 1.75240129e-01 -1.01915860e+00 6.00978136e-01
1.51884496e-01 -1.18572485e+00 3.15480649e-01 5.12752449e-03
8.55040193e-01 5.63236117e-01 5.95965050e-02 4.42356706e-01
7.81985879e-01 -1.06216979e+00 8.60603809e-01 2.79626846e-01
9.68837857e-01 -1.00564504e+00 8.73227537e-01 2.57925894e-02
-1.28087807e+00 1.10727400e-01 -3.43510628e-01 1.11643285e-01
5.19697607e-01 4.74311262e-01 -8.84100974e-01 4.66908962e-01
1.47253379e-01 7.00224400e-01 -7.41021812e-01 1.17486942e+00
-9.40921426e-01 8.66445422e-01 -3.28851908e-01 1.03998678e-02
5.57398915e-01 6.04765601e-02 3.64752352e-01 2.11415672e+00
1.16584010e-01 -8.95603895e-02 3.18889380e-01 2.73586035e-01
-6.69030726e-01 3.53859395e-01 -2.36333147e-01 -2.88442045e-01
5.53247452e-01 1.33935583e+00 -7.95209765e-01 -5.60214341e-01
-3.13532323e-01 1.26061356e+00 1.12131763e+00 5.90264946e-02
-9.92661834e-01 -6.99255526e-01 8.58236611e-01 -4.10438746e-01
7.27263808e-01 -6.43095791e-01 -5.46479046e-01 -1.20936704e+00
-2.32147217e-01 -8.06782246e-01 4.74003851e-01 -2.93024570e-01
-1.38863349e+00 7.02498913e-01 -3.68339628e-01 -4.08442259e-01
3.55569839e-01 -9.65226531e-01 -6.73986852e-01 8.31986308e-01
-1.31680644e+00 -1.55639720e+00 4.16308165e-01 8.13302398e-02
5.87518990e-01 -2.58987635e-01 9.50102329e-01 4.88889456e-01
-8.63374531e-01 9.99244928e-01 1.65720791e-01 6.52551830e-01
9.58773196e-01 -1.73967648e+00 1.22653985e+00 1.28196788e+00
6.87753439e-01 7.45097399e-01 7.33115152e-02 -8.38947833e-01
-1.03437650e+00 -1.39067602e+00 1.67734766e+00 -5.08332253e-01
1.05454290e+00 -6.79963410e-01 -8.47681940e-01 1.04060030e+00
6.58133090e-01 -1.40067622e-01 9.21521783e-01 2.91484237e-01
-5.28142095e-01 4.04808342e-01 -7.82636940e-01 7.17495322e-01
1.33253062e+00 -7.85224676e-01 -9.44788694e-01 3.75343323e-01
8.60636652e-01 -3.96483511e-01 -9.53164220e-01 1.71534449e-01
3.47568184e-01 -2.40365252e-01 6.86670125e-01 -8.99743021e-01
5.83246946e-01 8.34468231e-02 -3.85901868e-01 -1.43099499e+00
-6.27077222e-01 -6.15731239e-01 1.20653078e-01 1.53853488e+00
8.54798615e-01 -2.43688419e-01 5.75223386e-01 -2.88237184e-01
-6.98381662e-01 -7.44464159e-01 -8.91440153e-01 -8.90812278e-01
6.96663797e-01 -4.14373368e-01 2.24797338e-01 8.20853293e-01
-1.17126092e-01 5.72814643e-01 -6.79421127e-02 2.64375418e-01
2.46172532e-01 4.89249341e-02 1.39426947e-01 -8.55294824e-01
-3.55027258e-01 -4.69759285e-01 -2.10121751e-01 -1.09341609e+00
6.07027471e-01 -1.54863703e+00 2.17497244e-01 -1.66994989e+00
1.13372877e-01 -1.09141216e-01 -6.64913535e-01 8.16714108e-01
-3.88064653e-01 5.90028703e-01 3.21931511e-01 -3.19720924e-01
-6.74353302e-01 3.25937986e-01 6.66632295e-01 1.14462987e-01
2.59402901e-01 -4.74212378e-01 -5.18113434e-01 7.02920496e-01
1.09346616e+00 -7.49410868e-01 3.55033219e-01 -1.18985152e+00
3.88481647e-01 -3.86276722e-01 -6.73798844e-03 -8.68433595e-01
7.19779059e-02 5.29348962e-02 3.35252851e-01 -6.81036592e-01
3.03398848e-01 -1.64990246e-01 -4.10895169e-01 4.55073923e-01
-4.48958337e-01 5.94476044e-01 4.25344020e-01 2.38630310e-01
2.76671983e-02 -4.37588692e-01 9.61189210e-01 -1.32448718e-01
-6.66512966e-01 3.34147334e-01 -6.57207549e-01 5.02367139e-01
5.93873918e-01 3.38387609e-01 -1.31147966e-01 4.18981403e-01
-9.35403705e-01 9.88355577e-02 4.58115399e-01 3.61091048e-01
3.90353382e-01 -1.25844860e+00 -9.27405119e-01 1.18071988e-01
-1.25501350e-01 -4.17052507e-01 -1.88528374e-01 5.94077528e-01
-6.18864954e-01 2.25828588e-01 -3.53700161e-01 1.12688087e-01
-1.16038036e+00 4.67141539e-01 3.19046415e-02 -6.32561684e-01
-5.49382031e-01 1.54276097e+00 -2.10856855e-01 -4.28009480e-01
3.11831236e-02 -3.39389116e-01 1.80440154e-02 3.04163426e-01
2.03551576e-01 8.26765597e-02 2.38273337e-01 -5.98147511e-01
-6.25675976e-01 4.95694667e-01 -4.32418913e-01 -4.64241326e-01
1.45266378e+00 2.86323130e-01 -1.95395738e-01 3.56999815e-01
1.35629463e+00 4.60341692e-01 -1.03566647e+00 -1.70506850e-01
5.99317312e-01 6.31492957e-02 -2.05088183e-01 -1.19949377e+00
-8.32624257e-01 1.16270041e+00 -1.10359743e-01 -2.89474130e-01
6.93461239e-01 -1.08489603e-01 1.22114408e+00 4.97423321e-01
-1.99577704e-01 -1.30394506e+00 8.90674815e-02 1.22256374e+00
7.18512535e-01 -8.76236260e-01 -2.13176265e-01 -2.04344660e-01
-7.27330208e-01 1.01842940e+00 5.96055388e-01 -6.85285330e-01
2.93017060e-01 6.65847063e-01 1.10262029e-01 1.37621045e-01
-8.39721262e-01 -5.09626865e-01 2.79436558e-02 6.05800569e-01
8.62933576e-01 3.21513899e-02 -5.87999105e-01 9.28848982e-01
-5.19610286e-01 -5.86914599e-01 2.84959376e-01 7.25186944e-01
-3.89394760e-01 -1.47839761e+00 3.29385489e-01 1.74069405e-01
-9.76526022e-01 -8.03198993e-01 -5.80153048e-01 6.60436451e-01
5.01570642e-01 6.00759864e-01 1.85515955e-01 -2.78635681e-01
3.26470762e-01 5.92728615e-01 6.25148892e-01 -1.16957283e+00
-1.37303066e+00 -5.84301166e-02 6.42807782e-01 -2.05336541e-01
2.75592536e-01 -9.37347412e-01 -1.42600310e+00 -1.11017562e-01
2.63898168e-03 1.13736108e-01 6.55196548e-01 8.66166711e-01
3.24626893e-01 6.83019459e-01 1.93530902e-01 -6.38614297e-01
-5.67032099e-01 -1.21682131e+00 -3.07041973e-01 2.86782920e-01
-1.01332981e-02 -1.58655569e-01 -3.36252432e-03 8.70944783e-02] | [10.424283981323242, 10.005356788635254] |
e5c97a79-f288-4fe6-aa74-9bf65708d2cb | an-initial-investigation-for-detecting | 2104.02518 | null | https://arxiv.org/abs/2104.02518v2 | https://arxiv.org/pdf/2104.02518v2.pdf | An Initial Investigation for Detecting Partially Spoofed Audio | All existing databases of spoofed speech contain attack data that is spoofed in its entirety. In practice, it is entirely plausible that successful attacks can be mounted with utterances that are only partially spoofed. By definition, partially-spoofed utterances contain a mix of both spoofed and bona fide segments, which will likely degrade the performance of countermeasures trained with entirely spoofed utterances. This hypothesis raises the obvious question: 'Can we detect partially-spoofed audio?' This paper introduces a new database of partially-spoofed data, named PartialSpoof, to help address this question. This new database enables us to investigate and compare the performance of countermeasures on both utterance- and segmental- level labels. Experimental results using the utterance-level labels reveal that the reliability of countermeasures trained to detect fully-spoofed data is found to degrade substantially when tested with partially-spoofed data, whereas training on partially-spoofed data performs reliably in the case of both fully- and partially-spoofed utterances. Additional experiments using segmental-level labels show that spotting injected spoofed segments included in an utterance is a much more challenging task even if the latest countermeasure models are used. | ['Nicholas Evans', 'Jose Patino', 'Junichi Yamagishi', 'Erica Cooper', 'Xin Wang', 'Lin Zhang'] | 2021-04-06 | null | null | null | null | ['voice-anti-spoofing'] | ['audio'] | [ 4.33703780e-01 1.99142039e-01 -1.56737670e-01 -1.75108224e-01
-7.73885190e-01 -7.91237533e-01 4.00234371e-01 1.43706858e-01
-1.94508180e-01 3.72326553e-01 3.23631555e-01 -6.98938549e-01
2.47061774e-01 -4.97539908e-01 -5.76511860e-01 -6.15667105e-01
-2.65855163e-01 2.78936416e-01 4.84516591e-01 -4.77241009e-01
-1.28757907e-02 3.96855146e-01 -1.44320655e+00 5.45421660e-01
4.45815086e-01 4.37754720e-01 8.21773782e-02 9.62409616e-01
1.65658951e-01 6.14573002e-01 -1.38140726e+00 -3.97104025e-01
1.65046051e-01 -2.76607841e-01 -9.03947055e-01 9.33417454e-02
4.47084546e-01 -3.82206827e-01 -3.97914886e-01 1.37430739e+00
3.88367355e-01 -4.54451680e-01 1.02823965e-01 -1.29310489e+00
1.28133476e-01 8.74973476e-01 -8.09263960e-02 4.89935070e-01
6.84848845e-01 1.42768830e-01 7.82718420e-01 -2.84311086e-01
4.14814681e-01 1.61033499e+00 4.26639497e-01 3.57158214e-01
-1.01948142e+00 -9.86110032e-01 -4.07194793e-01 -1.03168488e-01
-9.88169968e-01 -7.55086005e-01 7.94109404e-01 -3.47455233e-01
7.82105327e-01 4.73750263e-01 1.92632660e-01 1.48146760e+00
1.62573710e-01 5.36428392e-01 1.00223601e+00 -4.16231275e-01
-1.45816341e-01 3.42239559e-01 1.68817222e-01 5.26254356e-01
4.08023059e-01 7.87101328e-01 -4.14031863e-01 -5.56418598e-01
3.98911312e-02 -3.44164073e-01 -5.31282306e-01 2.41207462e-02
-1.14829469e+00 9.97013330e-01 4.92389426e-02 7.36168206e-01
-2.33605728e-01 -2.69805610e-01 7.07770944e-01 7.84660280e-01
5.14476120e-01 6.50714219e-01 -4.22847569e-01 -3.24990839e-01
-1.09758496e+00 1.77931011e-01 9.79075134e-01 4.14330512e-01
6.09510481e-01 4.94261324e-01 4.50451493e-01 8.34858537e-01
1.55442715e-01 7.98287451e-01 4.18008476e-01 -6.22899175e-01
5.29462278e-01 -2.14733094e-01 1.60028130e-01 -1.29678476e+00
-3.21950197e-01 -5.58005273e-01 -2.64515728e-01 -9.97462571e-02
3.04806292e-01 -3.11625004e-01 -7.86228657e-01 1.62180960e+00
2.84479529e-01 -2.87563168e-02 1.08802684e-01 6.65697038e-01
3.16975206e-01 6.38794363e-01 -3.62535238e-01 -4.59255278e-01
1.23548460e+00 -7.29858041e-01 -1.08383930e+00 -2.65788764e-01
7.01095045e-01 -1.16787672e+00 7.39426076e-01 4.32207912e-01
-5.62139094e-01 -6.04326665e-01 -1.12599051e+00 9.64248896e-01
-3.89864504e-01 -5.90531528e-01 2.28161663e-01 1.53253722e+00
-7.75389433e-01 3.06519270e-01 -5.41176379e-01 -6.56530187e-02
-1.16769083e-01 1.48987144e-01 -6.09398246e-01 -3.83088142e-02
-1.69448936e+00 7.86420405e-01 5.19927263e-01 -4.54213768e-01
-1.33236384e+00 -3.10117841e-01 -8.31003845e-01 -5.66180423e-02
2.61432230e-01 1.68530956e-01 1.27787805e+00 -5.77770174e-01
-1.09447384e+00 7.28874326e-01 -1.62783489e-02 -7.64176071e-01
2.49503851e-01 1.57732032e-02 -1.28193796e+00 5.75156152e-01
7.92977214e-02 1.74852997e-01 1.45984602e+00 -1.23128736e+00
-6.46183550e-01 -1.04309529e-01 -8.44751373e-02 -3.80578130e-01
-3.83569211e-01 5.06939888e-01 2.01810047e-01 -8.21016788e-01
2.57165223e-01 -1.10767448e+00 3.76001835e-01 -8.83381069e-01
-6.67454004e-01 1.10849820e-01 1.37457943e+00 -9.26675498e-01
1.29479921e+00 -2.47205830e+00 -5.81481397e-01 1.27148643e-01
-1.58012450e-01 9.74519193e-01 -9.78715159e-03 6.41809464e-01
-4.66788203e-01 5.29340029e-01 -4.77035865e-02 -1.90626889e-01
-1.60854504e-01 1.93684235e-01 -7.18849361e-01 7.25824594e-01
1.83869549e-03 1.22643106e-01 -9.85332429e-01 -3.76998603e-01
1.70711264e-01 4.06757683e-01 -6.96765721e-01 1.90601408e-01
1.77702487e-01 4.12774205e-01 -9.09337923e-02 6.78301334e-01
5.50284684e-01 4.39346790e-01 -9.21694934e-02 8.07099417e-02
1.74480349e-01 8.65986466e-01 -9.69146252e-01 9.40386117e-01
-3.12177151e-01 8.90532017e-01 5.49648404e-01 -8.62679303e-01
1.01921725e+00 9.95011210e-01 3.52372825e-01 -1.91218674e-01
8.92706662e-02 5.63245356e-01 2.29887694e-01 -6.09563470e-01
5.03497243e-01 -5.45396447e-01 -3.21663916e-01 5.61622024e-01
2.15323240e-01 -4.91275728e-01 -9.70487818e-02 1.31810248e-01
1.41634500e+00 -7.58546591e-01 -1.37342848e-02 1.15691954e-02
4.56817597e-01 -2.70382404e-01 3.87302011e-01 1.05685389e+00
-8.86283815e-01 3.03523481e-01 1.75370336e-01 6.01773784e-02
-8.96488309e-01 -8.86620998e-01 -3.60221684e-01 1.08969426e+00
-7.74127916e-02 -3.75218362e-01 -9.15692091e-01 -9.60300326e-01
-5.73654026e-02 7.26418197e-01 -1.55835405e-01 -3.09291333e-01
-4.18318540e-01 -4.63328212e-01 1.18303347e+00 -2.95276284e-01
3.13507259e-01 -1.05127704e+00 -3.73105764e-01 4.71587986e-01
-5.60711026e-01 -1.32947493e+00 -1.86521739e-01 4.10166085e-01
-3.92091185e-01 -9.96150255e-01 -3.19097877e-01 -5.95774710e-01
7.71251321e-02 6.45517588e-01 6.80301666e-01 3.10325176e-01
8.39268342e-02 9.50027257e-03 -7.92231381e-01 -4.71800089e-01
-1.39130187e+00 -1.87246248e-01 2.94004083e-01 5.23036718e-02
1.99402660e-01 -2.95901716e-01 2.10460365e-01 8.06874275e-01
-1.02607179e+00 -6.90017283e-01 6.79149851e-02 8.93830776e-01
-3.66287947e-01 7.82838225e-01 8.16894472e-01 -6.52354538e-01
7.22373843e-01 -4.53114003e-01 -6.94068149e-02 -1.06018133e-01
-1.95281774e-01 -8.35276395e-02 4.00295138e-01 -4.66334790e-01
-6.66210830e-01 -3.94438267e-01 -7.10053027e-01 -4.25182313e-01
-6.57705724e-01 1.75429836e-01 -1.80627331e-01 -6.69055283e-02
7.75800347e-01 1.30907536e-01 6.79750368e-02 -5.84036827e-01
-5.37238568e-02 1.27101886e+00 6.18730545e-01 -1.84785903e-01
9.09744680e-01 4.40568477e-01 -5.75644195e-01 -1.32298374e+00
-5.15517294e-01 -7.59702981e-01 -4.72507253e-02 -1.79286003e-01
4.42956835e-01 -5.97923338e-01 -2.74611652e-01 7.24974453e-01
-1.35840845e+00 1.64869145e-01 1.84661802e-02 6.62273765e-01
-2.73655206e-01 8.02691221e-01 -7.57484853e-01 -9.81770992e-01
3.86294350e-02 -1.38511598e+00 9.00103152e-01 -4.85485077e-01
-3.54530722e-01 -8.55283320e-01 1.43343255e-01 5.70573151e-01
4.10781085e-01 9.55562964e-02 5.87219954e-01 -1.43573284e+00
3.39599669e-01 -8.00362051e-01 3.03404331e-01 8.97656143e-01
5.24508119e-01 3.54554355e-02 -1.27490556e+00 -6.45043075e-01
5.31379223e-01 -3.77401739e-01 7.88091302e-01 -3.71433422e-03
3.82226884e-01 -4.07481313e-01 -2.39592895e-01 1.62861064e-01
6.58311129e-01 4.75664139e-01 3.81103039e-01 -9.85501613e-03
2.84695804e-01 1.01986158e+00 7.48704374e-01 2.80869514e-01
-2.62662560e-01 7.17844367e-01 7.62211919e-01 3.13600332e-01
-3.47207971e-02 -4.34944779e-01 7.32876599e-01 1.14316070e+00
6.50622189e-01 -6.10267997e-01 -9.22798574e-01 5.30827641e-01
-8.56694281e-01 -1.33271718e+00 -1.14656184e-02 2.04853225e+00
6.15217268e-01 5.74935138e-01 3.36902499e-01 1.20932257e+00
1.34869027e+00 7.37709343e-01 8.96958709e-02 -8.18495750e-01
-2.19135568e-01 1.60214081e-02 3.91165733e-01 8.99685562e-01
-1.49088347e+00 8.10985386e-01 6.90197420e+00 8.93850148e-01
-1.33406246e+00 3.36722910e-01 1.71895131e-01 2.44308874e-01
-2.79906392e-01 9.48961638e-03 -6.89740360e-01 8.56201589e-01
1.45750320e+00 3.48240525e-01 2.31876224e-01 6.89880192e-01
2.94675857e-01 2.36370876e-01 -4.30871993e-01 6.51352704e-01
1.61680296e-01 -1.02248180e+00 -3.21479350e-01 4.13683355e-01
3.61881584e-01 -1.84844434e-02 9.33409333e-02 2.31919855e-01
3.03287655e-01 -7.38961756e-01 8.43602657e-01 -4.64358926e-01
5.41644633e-01 -7.80570686e-01 7.97053933e-01 7.97102988e-01
-8.43445659e-01 -2.28728682e-01 -2.02261373e-01 4.07883152e-02
4.75304097e-01 7.38274217e-01 -1.48057866e+00 4.23769355e-01
4.80788976e-01 2.11199626e-01 -2.34101623e-01 7.74318576e-01
-3.76677811e-01 1.25156522e+00 -4.23738897e-01 2.67000347e-01
2.40605146e-01 6.64871573e-01 9.53776062e-01 1.49553728e+00
2.64343917e-01 -5.77813625e-01 8.02574456e-02 1.89663336e-01
3.86613399e-01 -3.09585243e-01 -9.19097781e-01 -4.25704062e-01
8.13977301e-01 5.44793606e-01 -3.89832705e-01 -2.93277204e-01
5.00442237e-02 6.83876336e-01 -3.22935402e-01 1.51968256e-01
-7.64884114e-01 -4.88938749e-01 7.36170292e-01 1.80506781e-01
2.57503569e-01 -1.05971284e-01 3.38380992e-01 -9.46277022e-01
-4.30382580e-01 -1.45169103e+00 3.11279744e-01 -3.98570210e-01
-8.24736953e-01 7.96746194e-01 -1.65678874e-01 -9.96463776e-01
-6.82219148e-01 -3.08501601e-01 -3.04059714e-01 5.65214574e-01
-1.12753427e+00 -5.84399164e-01 2.54159331e-01 6.27765119e-01
4.70375687e-01 -2.58904546e-01 8.87289226e-01 3.99772376e-01
-3.41014832e-01 6.08551264e-01 -2.75632977e-01 3.23080242e-01
7.47207463e-01 -6.11901462e-01 4.64019775e-01 1.01912701e+00
5.25468409e-01 4.63323921e-01 1.40727818e+00 -7.86715627e-01
-9.89193320e-01 -7.73983717e-01 1.11607158e+00 -2.05064356e-01
7.13200629e-01 -4.17017937e-02 -9.79044020e-01 5.15185416e-01
-8.87804292e-03 -1.95055245e-03 5.60803294e-01 -2.94666171e-01
-7.96502590e-01 2.05894545e-01 -1.48346972e+00 -1.29818782e-01
6.23094857e-01 -1.26685596e+00 -8.84546995e-01 4.65120435e-01
1.13127899e+00 -3.59236896e-02 -5.65946460e-01 5.83645761e-01
3.98297191e-01 -1.26350164e+00 9.99768317e-01 -4.75182980e-01
-1.85676828e-01 -5.33577502e-02 -4.64603782e-01 -1.49090171e+00
2.92014331e-01 -9.32551980e-01 -2.60949284e-02 1.21148610e+00
1.69731289e-01 -7.98334539e-01 6.30141973e-01 -4.13203508e-01
-1.68081343e-01 -8.79265293e-02 -1.15585768e+00 -1.23974991e+00
-6.16873242e-02 -9.65160429e-01 6.48697972e-01 1.11990035e+00
3.78172696e-01 -1.90364465e-01 -7.06365824e-01 6.18779123e-01
5.01573265e-01 -4.60940838e-01 6.94916666e-01 -7.47972548e-01
-5.23890316e-01 -1.36171758e-01 -7.76639640e-01 -9.31538165e-01
2.80587852e-01 -5.15395105e-01 2.69930542e-01 -5.52702188e-01
-5.76573610e-01 -3.03593487e-01 -1.01878054e-01 1.90833762e-01
-5.97906671e-02 2.57671088e-01 2.42578834e-01 2.06065655e-01
1.51071593e-01 -1.55166656e-01 8.33112955e-01 -3.17748964e-01
1.10467926e-01 4.45155293e-01 -8.62869322e-02 6.89753234e-01
8.08587551e-01 -9.78793442e-01 -2.62390137e-01 -6.37752488e-02
-3.57511103e-01 4.78680819e-01 2.83265263e-01 -1.14942884e+00
-2.27139488e-01 -4.35314551e-02 -5.94807684e-01 -6.36116087e-01
5.05777776e-01 -7.06656575e-01 4.19108979e-02 1.03617442e+00
-3.26309621e-01 -1.50034577e-01 -4.67492118e-02 5.24191201e-01
-6.01329148e-01 -4.90823478e-01 9.79286969e-01 2.94434819e-02
-3.20391655e-01 -2.13405907e-01 -6.71209335e-01 -8.19474459e-02
6.73340917e-01 -1.30604491e-01 -4.84860241e-01 -7.90004194e-01
-6.82700932e-01 -6.04847133e-01 4.22892302e-01 4.37244892e-01
6.28095448e-01 -8.73668611e-01 -6.78289175e-01 6.90430522e-01
-7.77104124e-02 -7.75673568e-01 1.24459669e-01 4.91944253e-01
-4.75590646e-01 7.64449775e-01 5.79658076e-02 -6.40625238e-01
-1.64307475e+00 9.68192995e-01 7.90324509e-02 -2.15266243e-01
-8.02487582e-02 1.00588763e+00 -1.80331856e-01 -5.04163682e-01
4.05884564e-01 -8.28810036e-02 6.35153651e-02 1.77974954e-01
6.64410889e-01 3.67716461e-01 2.74688482e-01 -1.19748998e+00
-4.19121832e-01 -6.79528639e-02 1.77514600e-03 -3.58785421e-01
8.01651776e-01 -2.60669664e-02 6.36471584e-02 3.24040681e-01
1.67037499e+00 5.00541985e-01 -7.83105969e-01 -1.91467941e-01
-1.49978418e-02 -7.40349412e-01 4.96294722e-02 -5.22564054e-01
-7.94856071e-01 1.20405710e+00 4.33973730e-01 9.91319001e-01
6.97266757e-01 -5.28951213e-02 9.03771400e-01 -5.41656464e-03
8.19210410e-01 -6.26151741e-01 3.13910961e-01 4.72832888e-01
6.72971129e-01 -8.88452470e-01 -3.87910843e-01 -5.51814377e-01
-3.68148386e-01 7.77376533e-01 -3.29523906e-02 7.04045370e-02
8.78822088e-01 3.69572669e-01 2.28096232e-01 -2.97686812e-02
-5.05674720e-01 1.09182715e-01 -1.48521855e-01 7.06041396e-01
3.04337814e-02 3.62891018e-01 1.18829288e-01 3.35280821e-02
-7.31107593e-01 -8.32177758e-01 6.81147695e-01 1.12201643e+00
-1.00318599e+00 -1.07571793e+00 -1.19343984e+00 2.67074525e-01
-8.15122545e-01 1.12084739e-01 -4.92024392e-01 4.63465065e-01
2.47864008e-01 1.82165599e+00 -1.87961161e-01 -1.07028317e+00
-6.13095518e-03 2.15716034e-01 -8.16220045e-02 -7.55150557e-01
-6.77261889e-01 2.68248141e-01 5.14421701e-01 -3.87984753e-01
-4.33730781e-01 -4.95650142e-01 -7.31931984e-01 -5.74246764e-01
-8.80629241e-01 5.12272477e-01 1.08466065e+00 9.58161712e-01
-8.79780054e-02 3.57814193e-01 1.07826817e+00 -7.41574049e-01
-9.80704010e-01 -1.12649524e+00 -7.29005814e-01 6.11238360e-01
1.12217784e+00 -6.35209560e-01 -1.07367623e+00 -1.82128787e-01] | [14.125818252563477, 5.888576030731201] |
5403c12f-4281-44de-ba0e-1c5968320e6f | single-image-layer-separation-using-relative | null | null | http://openaccess.thecvf.com/content_cvpr_2014/html/Li_Single_Image_Layer_2014_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2014/papers/Li_Single_Image_Layer_2014_CVPR_paper.pdf | Single Image Layer Separation using Relative Smoothness | This paper addresses extracting two layers from an image where one layer is smoother than the other. This problem arises most notably in intrinsic image decomposition and reflection interference removal. Layer decomposition from a single-image is inherently ill-posed and solutions require additional constraints to be enforced. We introduce a novel strategy that regularizes the gradients of the two layers such that one has a long tail distribution and the other a short tail distribution. While imposing the long tail distribution is a common practice, our introduction of the short tail distribution on the second layer is unique. We formulate our problem in a probabilistic framework and describe an optimization scheme to solve this regularization with only a few iterations. We apply our approach to the intrinsic image and reflection removal problems and demonstrate high quality layer separation on par with other techniques but being significantly faster than prevailing methods. | ['Yu Li', 'Michael S. Brown'] | 2014-06-01 | null | null | null | cvpr-2014-6 | ['intrinsic-image-decomposition', 'reflection-removal'] | ['computer-vision', 'computer-vision'] | [ 9.08088863e-01 1.71125963e-01 2.70225763e-01 -1.16737135e-01
-6.89416170e-01 -3.38066459e-01 4.01409000e-01 -3.42367023e-01
-5.48636198e-01 6.72704279e-01 2.06347376e-01 -4.48906645e-02
-3.09268683e-01 -3.76856416e-01 -4.83925790e-01 -1.31446481e+00
1.62599072e-01 1.62677258e-01 3.64844441e-01 1.75684858e-02
1.44258931e-01 5.75148821e-01 -1.49882269e+00 3.36847514e-01
8.15589547e-01 8.56796086e-01 2.81178534e-01 8.07743430e-01
-1.41185537e-01 5.93463361e-01 -2.52957404e-01 -1.25036418e-01
3.76332313e-01 -3.56018305e-01 -6.94986999e-01 7.30941534e-01
5.24108589e-01 -1.01305030e-01 9.15367305e-02 1.13829446e+00
3.99991006e-01 -1.87733152e-03 8.87578666e-01 -8.91213775e-01
-5.93218692e-02 5.09781726e-02 -1.06014395e+00 -1.44217119e-01
2.01507479e-01 -2.87279963e-01 6.67217195e-01 -7.66303778e-01
4.47846770e-01 8.82986784e-01 8.93710554e-01 4.03930932e-01
-1.71017718e+00 -2.67276675e-01 2.49890774e-01 -4.73013818e-01
-1.27959239e+00 -5.88338912e-01 8.77035022e-01 -6.80590630e-01
6.68412447e-01 2.95090139e-01 1.41931191e-01 7.91842401e-01
3.51235494e-02 6.64582133e-01 1.39270902e+00 -6.36864781e-01
-1.24906644e-01 4.65133488e-01 2.96102226e-01 6.11193240e-01
5.15422285e-01 -1.28603414e-01 -3.12159479e-01 -2.22453699e-01
6.27285838e-01 -1.60796896e-01 -5.14980614e-01 -5.57190657e-01
-9.03667450e-01 5.52873731e-01 -1.02555737e-01 1.69642717e-01
-3.36749226e-01 9.30777472e-03 -4.12540920e-02 1.22638062e-01
6.13681138e-01 2.00155884e-01 -2.15833157e-01 4.27504003e-01
-1.32206726e+00 3.11558574e-01 8.57099414e-01 7.24768400e-01
9.80441749e-01 -1.38119638e-01 2.07884386e-01 1.08907413e+00
4.91150171e-01 6.05886579e-01 -1.34000614e-01 -1.00196302e+00
2.09857479e-01 1.76719740e-01 4.32959616e-01 -8.18244576e-01
-2.82625705e-01 -3.75535905e-01 -9.37438130e-01 7.64129817e-01
6.56961024e-01 -2.28669792e-01 -8.61806154e-01 1.58014655e+00
1.40713423e-01 -7.06618465e-03 5.92940906e-03 7.43138373e-01
3.53568614e-01 6.50027931e-01 -4.30820465e-01 -5.97008467e-01
1.24077141e+00 -6.47998095e-01 -7.37307966e-01 -1.66433662e-01
1.46927312e-01 -1.14799690e+00 6.73301399e-01 8.02552402e-01
-1.34497964e+00 -1.74225509e-01 -1.00573337e+00 -2.64716260e-02
1.29920870e-01 3.05753440e-01 3.66559297e-01 7.86923945e-01
-1.05475485e+00 6.60491943e-01 -6.18909955e-01 -1.40986517e-01
4.76510599e-02 4.67620373e-01 -4.38846231e-01 -2.05360912e-02
-4.84546304e-01 5.47939837e-01 -8.56642574e-02 2.09715843e-01
-1.09508105e-01 -6.66179657e-01 -8.02029192e-01 -2.30733156e-01
2.63523132e-01 -5.57520032e-01 8.20306182e-01 -1.10563219e+00
-1.60728621e+00 1.16788542e+00 -4.03281510e-01 -1.81058124e-01
6.15165889e-01 -3.16300362e-01 -1.01322733e-01 3.10031533e-01
-1.69278353e-01 -1.13503017e-01 1.51310956e+00 -1.86077416e+00
-3.02666068e-01 -1.32183984e-01 -1.95228532e-01 1.94191381e-01
-1.01687260e-01 2.22238097e-02 -5.31126440e-01 -5.80135345e-01
5.60266197e-01 -7.77178049e-01 -2.50490665e-01 -1.02691632e-02
-4.38034862e-01 5.15117884e-01 6.32281721e-01 -6.20237648e-01
1.05544770e+00 -2.19264865e+00 4.06695753e-01 3.52665097e-01
4.90528107e-01 8.67102444e-02 -7.56339356e-02 4.76119101e-01
-3.48599792e-01 -1.65319741e-01 -7.14012504e-01 -6.87699080e-01
-1.22129716e-01 1.06928997e-01 -1.88540384e-01 9.03644204e-01
1.60519212e-01 2.95455679e-02 -5.83974242e-01 -3.05514276e-01
1.45965606e-01 7.53071308e-01 -7.86575854e-01 1.78751856e-01
2.11624324e-01 3.52746904e-01 -1.77215397e-01 2.43253633e-01
1.21763337e+00 -1.57489955e-01 2.64401644e-01 -5.66878140e-01
-5.54998934e-01 -5.48853353e-02 -1.66922736e+00 1.12157786e+00
-4.58959252e-01 6.26114130e-01 1.01351595e+00 -9.83624876e-01
6.85152113e-01 3.82617205e-01 7.97032773e-01 -2.69315124e-01
-2.31438652e-02 1.00950196e-01 -1.87988281e-01 -5.82186282e-01
4.23446387e-01 -7.86831915e-01 3.70193541e-01 4.41760004e-01
-1.49924830e-01 -3.23128432e-01 4.50584143e-02 2.03839257e-01
1.01111805e+00 1.46186307e-01 2.18111407e-02 -5.62956750e-01
7.44987786e-01 -4.98663306e-01 5.21811485e-01 8.01080823e-01
6.75439462e-02 1.15157294e+00 5.97398937e-01 -4.24679443e-02
-8.48954201e-01 -1.08568180e+00 -3.28686684e-01 6.22197866e-01
1.44714698e-01 -3.90846610e-01 -8.57271075e-01 -4.47882026e-01
-3.57375070e-02 2.56657124e-01 -5.16681433e-01 1.53251529e-01
-4.94125783e-01 -1.10384429e+00 2.18975306e-01 1.89434197e-02
4.74297225e-01 -5.82390487e-01 -3.11739117e-01 1.85083621e-03
-2.10929364e-01 -1.34559441e+00 -4.37273264e-01 3.08283031e-01
-7.51282275e-01 -1.11300337e+00 -1.06801701e+00 -6.03548169e-01
9.84474361e-01 3.84377599e-01 1.27248943e+00 9.98277441e-02
-4.02938098e-01 7.35130548e-01 3.50433923e-02 -1.86497033e-01
-1.21236920e-01 -4.18054998e-01 1.52828693e-02 6.14219129e-01
-6.35033399e-02 -6.50091231e-01 -4.80193824e-01 1.56615451e-01
-9.65694129e-01 1.77087441e-01 5.01063466e-01 7.28132010e-01
4.00891006e-01 4.24106002e-01 -7.32733011e-02 -1.18490565e+00
3.52837682e-01 -8.42851102e-02 -6.86896265e-01 -1.07112288e-01
-4.09737647e-01 1.07890271e-01 3.74133080e-01 -2.07116291e-01
-1.50591898e+00 3.28202009e-01 -1.25580132e-01 -1.55315846e-01
-2.20858753e-01 1.61169320e-01 -2.75972128e-01 -4.46317345e-01
4.02994573e-01 3.20452005e-02 1.23506479e-01 -8.54382098e-01
7.15268776e-02 3.39882910e-01 4.66574043e-01 -5.90876698e-01
1.12257576e+00 9.54853237e-01 2.70891517e-01 -1.74615669e+00
-9.17489946e-01 -8.53203475e-01 -7.14400053e-01 -2.80825585e-01
9.28299963e-01 -6.74973309e-01 -6.66527748e-01 5.32082081e-01
-1.02697241e+00 -3.77445489e-01 -3.18821549e-01 3.39432359e-01
-5.05208850e-01 6.88463092e-01 -6.41942024e-01 -1.09913278e+00
4.17750441e-02 -1.05313206e+00 1.06237662e+00 3.40057686e-02
1.45570934e-01 -1.20105517e+00 2.62378514e-01 4.13053155e-01
2.17016086e-01 2.19680533e-01 5.68404257e-01 2.44788796e-01
-5.22284925e-01 -7.77077153e-02 -5.25122583e-01 6.48060679e-01
1.49967596e-01 6.21726960e-02 -1.23161077e+00 -3.29486936e-01
5.74785829e-01 1.34940729e-01 1.25568116e+00 6.63996279e-01
7.46819973e-01 -7.07290992e-02 -1.84070721e-01 9.52486157e-01
1.74658036e+00 -3.17238659e-01 8.53364229e-01 2.35070527e-01
8.13406944e-01 8.59487295e-01 1.94899693e-01 3.15372407e-01
-7.00627640e-02 6.20443344e-01 2.87686020e-01 -4.49201077e-01
-2.40526751e-01 4.66869622e-01 4.17207062e-01 7.74502456e-01
-3.56489658e-01 -1.16578832e-01 -6.28196657e-01 4.36361820e-01
-1.66671455e+00 -9.10178781e-01 -5.95297754e-01 2.39419222e+00
6.65559769e-01 1.59831375e-01 2.96711782e-03 4.82989043e-01
3.27643186e-01 2.28419378e-01 1.29520431e-01 -1.83505222e-01
-3.46329212e-01 1.63572967e-01 7.38552809e-01 1.24107468e+00
-1.19226074e+00 4.19506252e-01 7.25465631e+00 5.49470663e-01
-9.85670030e-01 -9.32992697e-02 6.81531131e-02 2.07060147e-02
-4.07236844e-01 4.90263849e-02 -8.65499735e-01 3.55190367e-01
4.19865310e-01 3.96993130e-01 2.37275466e-01 7.93181136e-02
2.26922289e-01 -4.94598478e-01 -7.70047724e-01 9.42159891e-01
1.42433539e-01 -8.74936342e-01 -1.39202520e-01 2.86971629e-01
6.17226958e-01 -1.76264018e-01 -2.99024023e-02 -4.80917633e-01
5.87374531e-02 -9.26994324e-01 4.55523610e-01 7.56830931e-01
6.33914471e-01 -7.05652654e-01 3.70731294e-01 3.68949562e-01
-1.10671008e+00 2.81453013e-01 -1.92429125e-01 -1.60879076e-01
1.45800829e-01 1.11473739e+00 -5.33830002e-02 6.59997046e-01
5.45376062e-01 6.52835786e-01 -1.30590439e-01 9.99219835e-01
-1.41413972e-01 3.90721202e-01 -5.06442070e-01 7.16486037e-01
1.44899208e-02 -9.20170188e-01 1.05087209e+00 1.56006062e+00
1.49085864e-01 1.50307089e-01 2.85540551e-01 5.68895817e-01
3.56754303e-01 -1.62223071e-01 -5.94548762e-01 2.31235340e-01
-4.37705517e-01 1.44452000e+00 -8.47290695e-01 -1.46610871e-01
-7.92070687e-01 1.17449462e+00 6.40505999e-02 8.18383396e-01
-4.04292703e-01 -4.00850952e-01 8.39351654e-01 4.55678105e-01
3.56413782e-01 -3.52571636e-01 -3.17867339e-01 -1.39461172e+00
1.56157449e-01 -6.76153123e-01 1.78552732e-01 -5.54833472e-01
-1.25636137e+00 4.17339683e-01 4.13456410e-02 -9.51214492e-01
1.00136556e-01 -1.07798135e+00 -5.07081270e-01 1.19288993e+00
-1.90747976e+00 -9.63496149e-01 -1.51026547e-01 6.30817294e-01
1.69171244e-01 1.97155118e-01 5.56866348e-01 5.45221210e-01
-5.35547554e-01 1.44406348e-01 2.60481924e-01 -1.40701950e-01
5.82830071e-01 -1.37484384e+00 -2.66218275e-01 1.06986129e+00
-2.21742809e-01 7.16769457e-01 1.08943200e+00 -4.91269410e-01
-1.26062155e+00 -5.26217699e-01 8.05076420e-01 -1.87227473e-01
4.04998541e-01 -4.36967343e-01 -9.72729385e-01 6.36227846e-01
2.76345074e-01 -9.77581888e-02 6.22933924e-01 1.80407707e-03
-3.30727428e-01 7.55558610e-02 -1.02099681e+00 3.95265102e-01
5.42139232e-01 -5.26632905e-01 -2.36551002e-01 3.81811202e-01
-1.28440991e-01 -9.45830569e-02 -5.32759786e-01 3.68593395e-01
6.34455323e-01 -1.39504266e+00 1.26508617e+00 2.71688006e-03
9.93395075e-02 -5.45178294e-01 -2.53762543e-01 -9.61289585e-01
-3.88873041e-01 -9.56074417e-01 1.03821665e-01 1.11181116e+00
3.43369603e-01 -7.64658093e-01 9.12646592e-01 4.74740833e-01
3.76632661e-02 -3.95258099e-01 -4.51689273e-01 -8.37530732e-01
-1.09976396e-01 -3.34209442e-01 -3.71553808e-01 6.82898104e-01
-3.67514968e-01 1.29905537e-01 -6.58517063e-01 5.06693482e-01
1.38176239e+00 1.56735182e-01 6.22556508e-01 -1.26785743e+00
-4.43951666e-01 -3.57798636e-01 -4.86801527e-02 -1.09131384e+00
-4.38396744e-02 -5.87256193e-01 3.74358863e-01 -1.44335032e+00
3.04796040e-01 -2.97120243e-01 -1.52261496e-01 -8.39650929e-02
-9.14600492e-03 6.06041670e-01 2.77195331e-02 1.21945895e-01
-1.55128762e-01 2.23089799e-01 9.02835488e-01 6.67774081e-02
-2.37609446e-01 2.39837110e-01 -5.57798922e-01 1.40800107e+00
4.47389662e-01 -4.12435204e-01 -4.22241271e-01 -3.98981988e-01
4.55343336e-01 -2.51551002e-01 3.66325736e-01 -9.98947799e-01
1.93926260e-01 9.11913142e-02 3.56827945e-01 -4.96550351e-01
6.13851249e-01 -9.44721699e-01 1.41318664e-01 4.30187173e-02
5.16876318e-02 -5.43214858e-01 6.16024993e-02 5.40978491e-01
-3.82800102e-02 -5.74926019e-01 1.23191738e+00 -2.82305688e-01
-1.80510670e-01 8.00628290e-02 -6.25871539e-01 1.90108120e-02
5.37621498e-01 -3.28365356e-01 2.59378850e-01 -4.09317493e-01
-9.12640333e-01 -1.36518165e-01 7.34928250e-01 -2.30807751e-01
6.28266811e-01 -9.43216920e-01 -7.43379354e-01 4.23346460e-01
-2.94917524e-01 -2.18038276e-01 2.61798888e-01 1.22576189e+00
-5.72463632e-01 -1.89268976e-01 2.11016685e-01 -6.32042050e-01
-1.58605254e+00 2.38236383e-01 4.82233316e-01 -4.28902984e-01
-1.11326599e+00 9.58509505e-01 7.57550776e-01 -1.59838527e-01
1.75001249e-01 -1.66385202e-03 -2.78506756e-01 2.32903078e-01
6.45923078e-01 3.89138520e-01 -5.66537566e-02 -8.70025992e-01
-3.26993048e-01 9.93334889e-01 -1.62303373e-01 -3.04186374e-01
1.41567934e+00 -4.35384482e-01 -4.54996049e-01 5.34035206e-01
1.22858620e+00 5.65370083e-01 -1.52067733e+00 -1.63333431e-01
-1.27055468e-02 -5.58206260e-01 3.69045615e-01 -3.28992039e-01
-1.09364951e+00 7.72393107e-01 2.17501611e-01 5.51313996e-01
1.30855691e+00 -2.79184163e-01 4.92439836e-01 2.32617781e-02
-1.38311684e-01 -1.05178976e+00 -2.10945169e-03 5.37216246e-01
8.89715910e-01 -1.05797982e+00 5.78621566e-01 -9.20926929e-01
-2.11649805e-01 1.05326807e+00 -2.00091247e-02 -5.29054403e-01
9.54472601e-01 8.08445513e-01 1.88790023e-01 -2.64068425e-01
-4.04849738e-01 -5.08453310e-01 5.06943941e-01 6.32794559e-01
6.92535162e-01 -2.73933679e-01 -1.98064551e-01 9.56499204e-02
3.03804666e-01 -2.47077614e-01 5.38627982e-01 9.84863281e-01
-2.32587367e-01 -1.22311783e+00 -6.30044222e-01 3.16837490e-01
-7.98886657e-01 -1.73644453e-01 -1.62259519e-01 6.14615023e-01
9.75364968e-02 8.53039980e-01 -1.56355515e-01 2.37710908e-01
3.12742174e-01 4.37855758e-02 7.98126221e-01 -6.19568825e-01
-3.19064111e-01 7.92641163e-01 1.19235143e-01 -4.94433463e-01
-7.90319324e-01 -8.71398091e-01 -8.88357043e-01 4.65256050e-02
-4.15217429e-01 -3.06015778e-02 6.06851161e-01 8.60949993e-01
-1.27800286e-01 4.98902917e-01 4.59838033e-01 -1.38030863e+00
-6.81856051e-02 -4.67644006e-01 -1.15303123e+00 3.18695128e-01
1.03885758e+00 -5.90017438e-01 -8.05263519e-01 3.23053032e-01] | [10.756985664367676, -2.748857259750366] |
28c32a96-1a0e-4ba4-a21f-cb30e351a906 | learning-system-parameters-from-turing | 2108.08542 | null | https://arxiv.org/abs/2108.08542v1 | https://arxiv.org/pdf/2108.08542v1.pdf | Learning System Parameters from Turing Patterns | The Turing mechanism describes the emergence of spatial patterns due to spontaneous symmetry breaking in reaction-diffusion processes and underlies many developmental processes. Identifying Turing mechanisms in biological systems defines a challenging problem. This paper introduces an approach to the prediction of Turing parameter values from observed Turing patterns. The parameter values correspond to a parametrized system of reaction-diffusion equations that generate Turing patterns as steady state. The Gierer-Meinhardt model with four parameters is chosen as a case study. A novel invariant pattern representation based on resistance distance histograms is employed, along with Wasserstein kernels, in order to cope with the highly variable arrangement of local pattern structure that depends on the initial conditions which are assumed to be unknown. This enables to compute physically plausible distances between patterns, to compute clusters of patterns and, above all, model parameter prediction: for small training sets, classical state-of-the-art methods including operator-valued kernels outperform neural networks that are applied to raw pattern data, whereas for large training sets the latter are more accurate. Excellent predictions are obtained for single parameter values and reasonably accurate results for jointly predicting all parameter values. | ['Christoph Schnörr', 'David Schnörr'] | 2021-08-19 | null | null | null | null | ['parameter-prediction'] | ['miscellaneous'] | [ 1.60962135e-01 -1.57400936e-01 1.93851054e-01 7.19816014e-02
2.49288557e-03 -6.29567325e-01 1.08643270e+00 4.53318477e-01
-4.56686199e-01 6.29268467e-01 -2.36041158e-01 -1.44540340e-01
-4.75977719e-01 -6.18141711e-01 -4.76374924e-01 -1.31292462e+00
-4.34370309e-01 7.46904850e-01 5.28606176e-01 -2.16691121e-01
5.47702253e-01 8.67856860e-01 -1.57270861e+00 1.20798334e-01
5.53971529e-01 5.46318352e-01 1.92680731e-01 1.13438070e+00
-1.61763474e-01 5.32879114e-01 -4.54720914e-01 -1.08697675e-02
1.34738088e-01 -6.14863157e-01 -6.90303624e-01 -6.02968261e-02
-1.34569675e-01 4.25198972e-01 -1.48864761e-01 6.33226275e-01
2.78291821e-01 1.90560207e-01 1.53637111e+00 -7.01044321e-01
-6.02774203e-01 1.65581048e-01 -1.03336096e-01 3.01284492e-01
6.27855361e-02 3.24841827e-01 6.17973268e-01 -7.69756258e-01
6.09405935e-01 9.20749664e-01 6.40740275e-01 5.33499658e-01
-1.68820310e+00 2.75655329e-01 -5.58574080e-01 1.11786857e-01
-1.58619773e+00 -1.48413897e-01 5.26351511e-01 -7.98477590e-01
1.09778154e+00 2.50974298e-01 7.03570545e-01 8.95428479e-01
6.10809863e-01 1.80071041e-01 1.20857012e+00 -6.34865046e-01
7.05694973e-01 -2.06352234e-01 -1.93622023e-01 7.69533217e-01
3.11156809e-01 1.53963014e-01 -2.22874254e-01 -2.94210285e-01
1.09505785e+00 -2.57322192e-01 -2.48508543e-01 -5.33088028e-01
-1.20247149e+00 3.28066200e-01 1.67725846e-01 8.20590496e-01
-4.36228991e-01 -1.49358930e-02 3.31280977e-02 3.31172884e-01
3.43036354e-01 7.28206635e-01 -5.02782822e-01 -1.58956677e-01
-6.32679164e-01 5.17731249e-01 1.02741694e+00 4.51854795e-01
8.13845873e-01 -1.39274687e-01 3.82728651e-02 8.97164702e-01
-7.78991207e-02 4.86145526e-01 5.77281535e-01 -5.35481811e-01
-5.74107915e-02 7.97111154e-01 1.92814574e-01 -7.10879207e-01
-6.73592985e-01 -2.03348011e-01 -1.19558012e+00 1.85169786e-01
9.53964889e-01 1.01366945e-01 -8.49062920e-01 1.38941824e+00
4.00713414e-01 2.29052424e-01 2.36335799e-01 7.52030492e-01
4.38080505e-02 8.77674580e-01 -3.22214007e-01 -2.79391319e-01
1.02026820e+00 -4.10868436e-01 -2.68966019e-01 3.11520278e-01
6.83134377e-01 -5.34049153e-01 8.97400916e-01 5.38504243e-01
-8.17559600e-01 -2.33234286e-01 -9.24977362e-01 3.24983358e-01
-6.77993298e-01 -1.00173458e-01 2.32296064e-01 5.06021976e-01
-9.32780802e-01 1.25024736e+00 -9.03299809e-01 -7.30235279e-01
6.15772512e-03 3.39775920e-01 -4.38612401e-01 3.13047767e-01
-7.72414565e-01 1.04657793e+00 3.24610829e-01 1.50680378e-01
-7.13714182e-01 -5.11048436e-01 -4.33402747e-01 5.06593334e-03
-1.82158902e-01 -4.26664770e-01 8.46602619e-01 -5.73967755e-01
-1.92930436e+00 8.50212693e-01 4.41016182e-02 -3.70150656e-01
6.01395428e-01 5.20068705e-01 8.85256082e-02 1.42514169e-01
-4.78449970e-01 3.78359973e-01 6.98489189e-01 -1.11983991e+00
5.29749133e-02 -4.54621240e-02 -7.99926877e-01 -6.34252131e-02
-8.87334719e-03 -2.31299609e-01 1.10086702e-01 -2.90889144e-01
2.83663899e-01 -9.71947908e-01 -3.24101925e-01 -1.46811247e-01
-1.18878096e-01 -3.89441043e-01 4.47122812e-01 -3.23779643e-01
8.21783543e-01 -1.85927749e+00 6.62963986e-01 4.00800854e-01
4.23424095e-02 9.42809284e-02 -6.14938624e-02 8.70646417e-01
9.48584545e-03 5.62296659e-02 -5.57246625e-01 -1.25684971e-02
1.54308556e-02 3.30208421e-01 -1.44181296e-01 7.92576611e-01
2.63453066e-01 5.54325163e-01 -5.81410170e-01 -2.79188544e-01
2.75616288e-01 4.97522384e-01 -1.11865945e-01 2.88933367e-01
-2.79429048e-01 5.15345097e-01 -2.66317695e-01 2.89498299e-01
3.95868748e-01 -2.37794191e-01 1.99846566e-01 4.63732898e-01
-4.46145654e-01 -3.58888537e-01 -7.70992458e-01 1.19294775e+00
-3.06031942e-01 5.55882692e-01 -2.21715391e-01 -1.10755217e+00
1.31159210e+00 3.19313198e-01 4.64692086e-01 -3.07709903e-01
2.03711495e-01 3.89185995e-01 3.90012950e-01 -5.50101638e-01
-1.04712322e-01 -4.10598218e-01 1.66963041e-01 3.64528120e-01
2.19700448e-02 -4.97331589e-01 1.92754343e-01 -3.57182145e-01
1.35769618e+00 -3.82824875e-02 2.98760474e-01 -8.26491177e-01
6.57203257e-01 -1.22109778e-01 2.69750625e-01 6.26317561e-01
2.79898625e-02 7.36261487e-01 8.66567135e-01 -9.83854234e-01
-1.61873043e+00 -1.02183402e+00 -3.12448770e-01 4.79393333e-01
1.07330762e-01 -1.92235096e-03 -1.03993857e+00 4.73627821e-02
-8.38996172e-02 2.81010985e-01 -1.07980561e+00 -1.50839403e-01
-5.30763090e-01 -1.03396201e+00 6.55650914e-01 2.26866193e-02
2.73245126e-01 -1.17991233e+00 -8.32230747e-01 2.46584207e-01
3.03521395e-01 -5.92142761e-01 1.98126242e-01 4.06758547e-01
-8.67377281e-01 -1.04907382e+00 -9.19451773e-01 -4.14671928e-01
7.05815375e-01 -3.82187188e-01 6.47912979e-01 2.21316665e-01
-8.03495526e-01 2.97112707e-02 -1.98254839e-01 -1.68348283e-01
-8.02785695e-01 1.43881604e-01 3.04127187e-01 -3.19883451e-02
9.47374403e-02 -6.64987564e-01 -5.48925877e-01 5.62943518e-01
-1.09830868e+00 -1.02799632e-01 6.63475573e-01 8.73343349e-01
5.00809848e-01 6.78943694e-02 9.34071094e-02 -3.22276890e-01
6.78725898e-01 -3.06448698e-01 -7.47029424e-01 4.59916949e-01
-3.73371303e-01 5.12042344e-01 9.53820825e-01 -7.34564424e-01
-8.64367962e-01 1.16418064e-01 2.44597107e-01 -2.06117243e-01
-6.52156293e-01 9.54868123e-02 2.34200343e-01 -2.69675076e-01
9.57131863e-01 6.55254722e-01 1.83740512e-01 -4.54439193e-01
-1.86160859e-02 3.67132038e-01 3.52032632e-01 -8.63062084e-01
5.08968890e-01 4.37899590e-01 6.56627476e-01 -1.43556619e+00
9.24694911e-03 -2.80700117e-01 -1.02843177e+00 -3.15649420e-01
8.08232427e-01 3.67072262e-02 -7.20335305e-01 8.81966472e-01
-1.17878795e+00 -8.44335616e-01 -3.83987427e-01 3.20834756e-01
-1.07613969e+00 3.22848111e-01 -8.62945676e-01 -1.07274854e+00
5.67511171e-02 -8.03219259e-01 8.57635558e-01 1.01467684e-01
-1.30700365e-01 -1.29442143e+00 7.79519439e-01 -3.06733996e-01
5.72674453e-01 3.51238936e-01 1.23353839e+00 -5.49286723e-01
-3.74376029e-01 -3.00225824e-01 2.07877263e-01 3.07911802e-02
5.32674566e-02 4.99877661e-01 -7.38667786e-01 -9.29193944e-02
-4.18495908e-02 2.96068043e-02 7.70593107e-01 4.10862237e-01
7.68743634e-01 -1.16271630e-01 -3.74191076e-01 3.50510210e-01
1.26895726e+00 1.75071359e-01 5.57511330e-01 1.19064495e-01
2.68715382e-01 1.02110493e+00 2.26994026e-02 4.74430799e-01
-1.78941712e-01 5.87079048e-01 1.55306652e-01 2.01085910e-01
2.59714156e-01 -3.66646163e-02 7.23472163e-02 7.96174228e-01
-4.01281774e-01 -1.71239465e-01 -1.42030215e+00 3.73543829e-01
-1.93013740e+00 -9.27016616e-01 -3.52521867e-01 2.51186371e+00
5.40267110e-01 8.95631760e-02 2.37889096e-01 1.89793572e-01
8.45512569e-01 -3.27853948e-01 -3.73364180e-01 -4.59099203e-01
-4.22494471e-01 7.52604157e-02 3.85194033e-01 6.35060251e-01
-8.39882672e-01 7.18833804e-01 7.26347733e+00 6.68002248e-01
-1.19666719e+00 -1.32266551e-01 5.95904112e-01 1.78364411e-01
1.20145097e-01 -4.95812893e-02 -2.84863412e-01 4.66231614e-01
1.04996848e+00 4.05624602e-03 5.09078503e-01 2.77361274e-01
2.13979274e-01 -2.20508009e-01 -1.00274539e+00 8.83034647e-01
-1.81526914e-01 -1.27810669e+00 -1.51844278e-01 3.90650004e-01
6.90720737e-01 -7.00422674e-02 -4.35425229e-02 -3.40706527e-01
8.97824019e-02 -1.06572342e+00 5.50380528e-01 1.01602018e+00
3.65949303e-01 -4.32731450e-01 7.22084284e-01 6.41701519e-01
-9.50226009e-01 -6.08074702e-02 -8.33117366e-01 -3.70754153e-01
-5.07213846e-02 7.16871142e-01 -8.40053856e-01 1.50693059e-01
2.39575878e-01 3.93198788e-01 -4.82392550e-01 1.26808918e+00
7.23861307e-02 5.84549904e-01 -8.08884978e-01 -5.82684755e-01
1.67886660e-01 -5.20521402e-01 4.02484804e-01 1.31318879e+00
4.44939792e-01 2.75144223e-02 -5.03857732e-01 1.09117234e+00
4.61235374e-01 3.41196895e-01 -9.42213833e-01 4.41976078e-02
1.92007825e-01 1.02115571e+00 -1.31076908e+00 -1.50744870e-01
7.99026862e-02 9.24489737e-01 5.10785162e-01 4.69250679e-01
-3.34958643e-01 -2.90330470e-01 5.74739575e-01 2.58474737e-01
3.27783078e-01 -5.90370953e-01 -7.74027109e-02 -1.10285485e+00
-1.94519237e-01 -7.32894987e-02 -3.94758508e-02 -4.36974138e-01
-1.33700383e+00 4.27987576e-01 1.86275452e-01 -9.15597200e-01
-1.85410529e-01 -1.09421599e+00 -9.14926529e-01 7.71323740e-01
-8.71601522e-01 -7.25481868e-01 1.62888095e-01 1.74968228e-01
2.16295943e-01 -8.89973789e-02 8.36916924e-01 -4.43887979e-01
-5.25418997e-01 -5.48141589e-03 7.10509121e-01 -2.02111006e-01
1.56313181e-01 -1.49576783e+00 3.10779452e-01 5.92091322e-01
9.78431031e-02 5.60643494e-01 1.04775131e+00 -5.35586774e-01
-1.31599605e+00 -6.18202865e-01 9.40743387e-01 -4.78735745e-01
7.14709401e-01 -5.70161402e-01 -1.20879745e+00 -8.84603709e-02
5.76853082e-02 2.49942183e-01 5.75556159e-01 -2.73289621e-01
-2.90069997e-01 1.63092926e-01 -1.09557080e+00 6.79570019e-01
8.10131907e-01 -3.05247068e-01 -4.52096909e-01 3.38245064e-01
-1.43283382e-01 8.07324573e-02 -9.23416793e-01 6.81642517e-02
6.28924131e-01 -1.12092626e+00 5.89770734e-01 -6.84638739e-01
4.11188751e-01 -4.69797641e-01 1.22699678e-01 -1.31112802e+00
-5.13756394e-01 -7.25077152e-01 6.74403906e-02 7.50047088e-01
4.83120888e-01 -7.69660890e-01 6.53561056e-01 4.66895521e-01
3.56928855e-01 -8.30617607e-01 -1.36461127e+00 -9.86857653e-01
4.75524217e-01 1.03656091e-01 4.54203159e-01 7.96452224e-01
4.41437453e-01 2.70688552e-02 -4.35492061e-02 8.19562301e-02
7.40365386e-01 -7.18191117e-02 5.17408848e-01 -1.31856740e+00
-3.78490180e-01 -9.73507643e-01 -9.55269635e-01 -5.95781922e-01
1.50499493e-01 -5.12935579e-01 3.17675382e-01 -1.06813300e+00
-1.06194854e-01 -4.15310264e-01 -2.81790018e-01 7.46962503e-02
1.22904763e-01 6.96635917e-02 -2.18361676e-01 3.55202109e-01
-3.18970144e-01 5.03158092e-01 8.71901333e-01 1.44605130e-01
-1.86668098e-01 -1.24323085e-01 3.86563122e-01 5.32785058e-01
9.61590648e-01 -2.25909188e-01 -2.57712752e-01 1.29748195e-01
2.63787031e-01 1.70001835e-01 3.27008456e-01 -1.19962621e+00
2.17025653e-01 -3.46742153e-01 4.09278214e-01 1.10028945e-02
1.00829601e-01 -6.27450228e-01 3.64583582e-01 6.08244896e-01
-4.20551628e-01 1.53902933e-01 -3.17938589e-02 6.22494996e-01
7.30478484e-03 -4.14642304e-01 9.87612903e-01 -1.03308752e-01
-6.15160652e-02 7.06106871e-02 -1.11465240e+00 -3.43919337e-01
1.10800040e+00 -2.53886253e-01 -3.16929758e-01 -1.57166064e-01
-7.38567114e-01 -2.91975468e-01 9.04982924e-01 -1.11385983e-04
4.87172067e-01 -1.07807708e+00 -5.74034750e-01 3.21695417e-01
1.18118703e-01 -2.77390182e-02 2.91994542e-01 8.74500692e-01
-1.17658746e+00 3.98064077e-01 -3.47262025e-01 -7.20829189e-01
-1.05770969e+00 7.60658681e-01 7.24298656e-01 -2.08956882e-01
-3.74246657e-01 5.71098506e-01 1.28050700e-01 -5.19278109e-01
-3.93315941e-01 -6.92504764e-01 2.76689157e-02 -3.63742739e-01
3.30368876e-01 4.42109436e-01 -1.06560187e-02 -5.76181412e-01
-2.10174605e-01 9.12648678e-01 3.42073709e-01 -2.10826010e-01
1.29060125e+00 1.76999271e-01 -3.53158176e-01 8.98883641e-01
8.72612178e-01 -4.50162768e-01 -1.52471566e+00 -3.86178419e-02
2.48837858e-01 -2.09366322e-01 -4.50808436e-01 -5.28781652e-01
-3.30297202e-01 1.12004113e+00 4.15679514e-01 6.10041142e-01
6.57370448e-01 2.28026599e-01 -4.16535139e-02 6.51576102e-01
3.03833932e-01 -9.70661402e-01 2.09001862e-02 5.44976056e-01
8.60956609e-01 -7.17742264e-01 -2.28622660e-01 -2.22823322e-01
-2.27116913e-01 1.52223980e+00 3.57685268e-01 -2.97909379e-01
5.55639327e-01 4.44683164e-01 -8.87394622e-02 4.65873489e-03
-1.03927767e+00 -7.92287476e-03 7.37044588e-02 6.27682149e-01
4.29114580e-01 1.19958341e-01 -3.61239493e-01 -8.49730000e-02
-1.45696728e-02 -2.89721876e-01 5.05355179e-01 7.93116033e-01
-5.54188192e-01 -1.03137231e+00 -4.63306129e-01 3.22015762e-01
-1.55110613e-01 1.71191573e-01 -6.53272569e-01 5.70310652e-01
5.73167391e-02 4.13247168e-01 2.58823693e-01 -2.45594904e-01
1.58552453e-01 4.28893238e-01 8.85206282e-01 -1.55756027e-01
-3.71746540e-01 -1.42750785e-01 -3.87928724e-01 -1.16255276e-01
-1.97954372e-01 -8.00420940e-01 -1.14216423e+00 -3.13604027e-01
-1.66597322e-01 3.75983804e-01 6.58790708e-01 1.05178511e+00
1.92645192e-01 1.57682840e-02 4.27449942e-01 -1.10089684e+00
-3.68603796e-01 -1.07829857e+00 -9.35345769e-01 4.45771456e-01
2.84640670e-01 -6.90141320e-01 -4.84837115e-01 1.76809788e-01] | [6.070559501647949, 4.199647426605225] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.