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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
d6efb8f1-a4e5-4b5d-bf19-e5859e7ef15e | deep-residual-learning-for-image-recognition | 1512.03385 | null | http://arxiv.org/abs/1512.03385v1 | http://arxiv.org/pdf/1512.03385v1.pdf | Deep Residual Learning for Image Recognition | Deeper neural networks are more difficult to train. We present a residual
learning framework to ease the training of networks that are substantially
deeper than those used previously. We explicitly reformulate the layers as
learning residual functions with reference to the layer inputs, instead of
learning unreferenced functions. We provide comprehensive empirical evidence
showing that these residual networks are easier to optimize, and can gain
accuracy from considerably increased depth. On the ImageNet dataset we evaluate
residual nets with a depth of up to 152 layers---8x deeper than VGG nets but
still having lower complexity. An ensemble of these residual nets achieves
3.57% error on the ImageNet test set. This result won the 1st place on the
ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100
and 1000 layers.
The depth of representations is of central importance for many visual
recognition tasks. Solely due to our extremely deep representations, we obtain
a 28% relative improvement on the COCO object detection dataset. Deep residual
nets are foundations of our submissions to ILSVRC & COCO 2015 competitions,
where we also won the 1st places on the tasks of ImageNet detection, ImageNet
localization, COCO detection, and COCO segmentation. | ['Xiangyu Zhang', 'Shaoqing Ren', 'Jian Sun', 'Kaiming He'] | 2015-12-10 | deep-residual-learning-for-image-recognition-1 | http://openaccess.thecvf.com/content_cvpr_2016/html/He_Deep_Residual_Learning_CVPR_2016_paper.html | http://openaccess.thecvf.com/content_cvpr_2016/papers/He_Deep_Residual_Learning_CVPR_2016_paper.pdf | cvpr-2016-6 | ['retinal-oct-disease-classification'] | ['computer-vision'] | [ 1.59325898e-01 2.49316648e-01 -4.39586341e-02 -3.71373057e-01
-5.43741703e-01 -5.44872046e-01 3.88906121e-01 -3.14377964e-01
-8.73889327e-01 4.52866942e-01 -1.59757003e-01 -4.13832039e-01
1.91881016e-01 -5.49961567e-01 -1.00516462e+00 -4.65056151e-01
-2.17815340e-01 2.01522186e-01 1.95303887e-01 -2.21798971e-01
1.12360872e-01 6.65548503e-01 -1.34425533e+00 6.65173352e-01
3.20756972e-01 1.28476369e+00 3.73127684e-02 1.04347491e+00
3.43105584e-01 1.05088234e+00 -8.20891619e-01 -3.40613246e-01
4.07370716e-01 -7.23669454e-02 -1.03724301e+00 -8.49172175e-02
9.32679355e-01 -1.69900417e-01 -5.18288732e-01 1.01344764e+00
4.52592105e-01 -2.54949201e-02 6.55436039e-01 -1.11560535e+00
-8.32256973e-01 6.46784306e-01 -3.76127571e-01 5.66481471e-01
-7.44557679e-02 2.41877571e-01 1.17768180e+00 -9.93085206e-01
6.29500985e-01 1.17307305e+00 9.03886735e-01 8.12619627e-01
-1.08706665e+00 -5.46337664e-01 4.02362257e-01 8.42494220e-02
-1.26411545e+00 -3.65925550e-01 1.46265015e-01 -4.12643671e-01
1.36073339e+00 3.11392516e-01 4.31191295e-01 1.11963618e+00
-2.66309846e-02 1.20269179e+00 7.76100934e-01 -2.87271172e-01
-2.17755988e-01 1.90795958e-01 2.24566773e-01 8.78570616e-01
2.63691247e-01 1.48574769e-01 1.05600595e-01 5.21000147e-01
8.41136217e-01 -6.14484772e-02 -2.59658515e-01 -1.66205168e-01
-1.16952968e+00 9.30310190e-01 1.20388341e+00 3.66020113e-01
-4.77354042e-02 6.87199175e-01 4.38578308e-01 3.66633475e-01
3.28235775e-01 6.91787958e-01 -6.02954507e-01 4.19020325e-01
-7.27249265e-01 1.35872036e-01 6.56289577e-01 1.01073134e+00
5.30763507e-01 5.32861471e-01 -4.09065112e-02 9.00525272e-01
1.24959722e-01 2.91762441e-01 5.09766936e-01 -9.83577669e-01
6.54489875e-01 5.26230931e-01 -2.67642587e-01 -8.40501845e-01
-5.59555769e-01 -8.82964671e-01 -8.71254206e-01 6.19197130e-01
5.40453792e-01 -1.35856375e-01 -1.42459440e+00 1.45998418e+00
-3.63508254e-01 -1.87639624e-01 1.04294829e-01 7.89571524e-01
1.23853505e+00 5.32043576e-01 1.00570276e-01 5.01267612e-01
1.28450811e+00 -1.19085526e+00 -2.11818025e-01 -5.12685597e-01
8.90334725e-01 -5.79805315e-01 7.89374709e-01 4.42905813e-01
-1.10300434e+00 -6.68372691e-01 -1.31599140e+00 -1.70568421e-01
-8.28406870e-01 4.06333983e-01 8.69675457e-01 5.74844480e-01
-1.47375345e+00 7.94588149e-01 -4.76867557e-01 -2.31665045e-01
9.76341546e-01 5.90121210e-01 -5.26405632e-01 5.01455218e-02
-9.70433116e-01 1.14517915e+00 4.56230313e-01 2.99130738e-01
-1.01874018e+00 -6.83531165e-01 -9.74103510e-01 3.92397642e-02
5.44224940e-02 -3.37113827e-01 1.07473338e+00 -1.08354127e+00
-8.95633638e-01 1.47744024e+00 2.49313518e-01 -1.07063091e+00
7.46695399e-01 -1.37518495e-01 -3.41379315e-01 1.14002451e-01
1.73830628e-01 1.31314874e+00 6.97046936e-01 -1.03367198e+00
-8.40311527e-01 -1.25050634e-01 3.51961017e-01 8.93214066e-03
1.64356783e-01 -1.75827742e-01 -4.41482037e-01 -5.99409044e-01
2.27696523e-02 -7.74675071e-01 -4.52451348e-01 -6.67095035e-02
-4.71338511e-01 -3.73304963e-01 6.62182033e-01 -3.89163822e-01
7.03276575e-01 -2.10292387e+00 -1.39527887e-01 -7.63140200e-03
5.23631454e-01 4.75077659e-01 -5.20886242e-01 -4.25407887e-01
-5.32167912e-01 4.78882521e-01 -8.80777761e-02 -2.77193248e-01
3.81066911e-02 4.51172739e-02 -3.63923520e-01 6.19447410e-01
3.24432701e-01 1.26231229e+00 -6.24659300e-01 -1.16216585e-01
3.60454172e-01 5.11759043e-01 -4.88800168e-01 -1.78413272e-01
8.36554356e-03 -5.80715500e-02 -9.35695320e-02 6.89235985e-01
5.93664587e-01 -4.29922491e-01 -2.31772795e-01 -4.04363781e-01
-1.09665327e-01 7.51739517e-02 -8.28001499e-01 1.30560374e+00
-3.50745171e-01 1.27537405e+00 -4.97820368e-03 -1.30746257e+00
6.98708594e-01 3.01445778e-02 2.49802798e-01 -1.03159451e+00
2.70203114e-01 2.73487002e-01 1.24695092e-01 -1.07642531e-01
4.44769830e-01 2.31424257e-01 -1.49523318e-02 -3.17828581e-02
4.23303097e-01 -3.82944718e-02 3.15637648e-01 2.04502925e-01
1.05517793e+00 -1.15740806e-01 1.57248348e-01 -3.50872070e-01
3.63223195e-01 -3.32795829e-02 1.50370404e-01 1.06223845e+00
-4.11234498e-01 8.72260392e-01 7.01223135e-01 -9.07185197e-01
-1.00549841e+00 -9.06903207e-01 -5.06916881e-01 1.16496265e+00
-1.08822986e-01 -2.21414015e-01 -6.92280650e-01 -9.54148531e-01
3.32020037e-02 3.47617894e-01 -1.10394132e+00 4.60405089e-02
-6.97749555e-01 -7.91042626e-01 1.14382374e+00 8.74423146e-01
6.38541996e-01 -1.42272317e+00 -5.26351810e-01 1.01357698e-03
2.96311110e-01 -1.37659323e+00 -2.81787664e-01 7.86828995e-01
-7.18929648e-01 -1.34749901e+00 -9.02171612e-01 -1.23253024e+00
7.11300790e-01 -2.21496262e-02 1.29598272e+00 2.75759280e-01
-8.68214846e-01 3.88354748e-01 -9.06929746e-02 -3.93794894e-01
-9.83459279e-02 2.81168342e-01 -3.44900161e-01 -4.45535868e-01
2.46003583e-01 -9.51146800e-03 -7.15082765e-01 3.39520037e-01
-7.53211498e-01 -2.11108550e-01 5.45399666e-01 8.99974704e-01
4.27310377e-01 -2.23238423e-01 4.10137206e-01 -1.05721462e+00
2.37022549e-01 -2.76626796e-01 -8.12067270e-01 7.30871558e-02
-3.44157279e-01 -2.89528398e-03 4.44213182e-01 -2.28910804e-01
-5.70181012e-01 2.73840129e-01 -5.75717330e-01 -3.41370344e-01
-2.85364330e-01 -2.23485939e-02 2.25152493e-01 -4.25667733e-01
8.78161013e-01 -1.96246386e-01 -4.18751538e-01 -2.54668027e-01
3.97179395e-01 1.27763137e-01 6.77379489e-01 -1.43723130e-01
4.87044215e-01 4.08305049e-01 -1.60897121e-01 -8.17053974e-01
-1.17385697e+00 -3.63372922e-01 -4.79166985e-01 1.06596857e-01
1.31528461e+00 -1.01115906e+00 -9.80316818e-01 5.32449126e-01
-1.20118308e+00 -6.90151632e-01 -3.54255170e-01 2.77336240e-01
-2.69477338e-01 -8.38174596e-02 -8.21790159e-01 -2.65823781e-01
-1.60212070e-01 -1.38697851e+00 7.58376300e-01 1.01232983e-01
2.26779794e-03 -1.24599147e+00 -2.62115240e-01 1.56807914e-01
4.73612040e-01 4.31885123e-01 6.60423815e-01 -6.56159043e-01
-5.18188953e-01 -2.46627718e-01 -6.74497902e-01 7.48027742e-01
-3.59929115e-01 -1.33617729e-01 -1.38253176e+00 -2.53033549e-01
-2.95539916e-01 -6.39790177e-01 1.79331565e+00 5.94957352e-01
1.51021552e+00 -6.07525408e-02 -2.60182500e-01 1.09267962e+00
1.52765572e+00 8.74280706e-02 9.18783426e-01 5.62136471e-01
9.38359380e-01 4.92357641e-01 4.36810516e-02 -7.68820420e-02
1.93830743e-01 4.36617911e-01 8.80447209e-01 -4.78503466e-01
-4.96136874e-01 9.27111357e-02 1.22867309e-01 1.97271124e-01
-1.39379084e-01 -2.60577142e-01 -8.16884875e-01 7.02370584e-01
-1.60314286e+00 -9.22901511e-01 -2.81914622e-01 1.82314157e+00
4.47785288e-01 4.48631883e-01 -1.03107408e-01 1.47940544e-02
6.93181515e-01 3.43660831e-01 -5.03708839e-01 -5.64515352e-01
-3.07200164e-01 5.83068252e-01 1.07421327e+00 6.04011595e-01
-1.58890629e+00 1.29420495e+00 7.33696938e+00 6.87312365e-01
-1.16669500e+00 1.49946928e-03 9.54706371e-01 -6.87388331e-02
1.09151386e-01 -4.27738339e-01 -1.00660276e+00 1.23897783e-01
6.94139242e-01 5.26835084e-01 3.23358208e-01 1.25111771e+00
-5.80197036e-01 1.85886934e-01 -1.32913351e+00 1.20941889e+00
-9.53144487e-03 -1.76279414e+00 -3.69518367e-03 9.62830707e-02
9.59244251e-01 7.38267839e-01 4.52820241e-01 7.87339449e-01
4.15878683e-01 -1.88292623e+00 7.88166881e-01 6.06094971e-02
9.01210666e-01 -6.65738404e-01 8.80790412e-01 -1.00718312e-01
-1.25311017e+00 5.35748294e-03 -8.34389865e-01 1.82544626e-02
-3.77745420e-01 2.47190952e-01 -6.88608527e-01 -2.41622142e-02
9.01825249e-01 8.62493038e-01 -9.06668842e-01 1.10115182e+00
-3.94027680e-01 3.23562711e-01 -2.26857632e-01 -1.27650294e-02
6.99748278e-01 4.71350551e-01 2.77469903e-01 1.59723341e+00
-1.20251976e-01 -2.39198029e-01 -3.88336182e-03 9.73811924e-01
-7.57065594e-01 -3.58642995e-01 -7.43561268e-01 4.38367203e-02
1.44346714e-01 1.45375800e+00 -8.77825797e-01 -3.56556475e-01
-1.99862018e-01 1.08463442e+00 4.31282371e-01 5.56748748e-01
-9.09125149e-01 -7.11164832e-01 6.33111119e-01 -2.08635956e-01
5.88533700e-01 1.73755363e-01 -3.28871757e-01 -1.12111795e+00
-2.23850071e-01 -7.86934197e-01 3.43340486e-01 -5.28390110e-01
-1.13002050e+00 9.55622613e-01 -3.71063620e-01 -7.43173778e-01
-1.02752216e-01 -1.43907130e+00 -7.36645699e-01 7.32746601e-01
-1.51522994e+00 -9.57243085e-01 -2.11983576e-01 6.79136038e-01
5.23461163e-01 -2.78283864e-01 7.18054831e-01 4.39662039e-01
-4.98711616e-01 9.11830783e-01 -3.12729180e-01 7.44295835e-01
2.92738229e-01 -1.56296277e+00 7.94973433e-01 6.57801986e-01
3.90446275e-01 5.25640488e-01 2.72523314e-01 -8.84048939e-02
-9.23179030e-01 -1.23932517e+00 5.98467290e-01 -6.84504032e-01
5.22161961e-01 -4.19310719e-01 -6.83123946e-01 1.11532128e+00
2.58069009e-01 6.35903955e-01 9.13474634e-02 7.90810883e-02
-5.29526472e-01 -2.37241611e-02 -1.13676369e+00 3.82447124e-01
1.06094480e+00 -5.81315279e-01 -5.79037130e-01 3.45923901e-01
7.35033154e-01 -5.86803496e-01 -5.03882766e-01 6.10287011e-01
5.47414899e-01 -8.69533718e-01 1.21825755e+00 -7.36061513e-01
2.77706176e-01 -1.85483426e-01 -2.38700271e-01 -1.14352536e+00
-3.42731237e-01 -1.31532818e-01 2.42333427e-01 7.71000564e-01
6.46742582e-01 -6.42690063e-01 9.91554260e-01 2.06997488e-02
-2.54300654e-01 -8.35858345e-01 -6.87505782e-01 -6.88590348e-01
2.56101817e-01 -7.09597290e-01 4.71471027e-02 8.26049685e-01
-6.58979952e-01 4.27440584e-01 -1.52450576e-01 9.94021073e-03
6.59123659e-01 -4.19028193e-01 5.45565069e-01 -1.09476268e+00
-1.72448888e-01 -9.36978161e-01 -8.41022730e-01 -1.18866503e+00
1.46382570e-01 -1.22235382e+00 8.39208886e-02 -1.62003291e+00
-4.03345078e-02 -2.82946110e-01 -5.30024827e-01 7.37326443e-01
1.71699032e-01 1.12020755e+00 3.66907477e-01 1.12258818e-03
-8.21038842e-01 1.50977075e-02 1.11231899e+00 -4.35552567e-01
2.48392314e-01 -8.44713748e-02 -8.33301246e-01 1.06277955e+00
8.64355445e-01 -3.95896792e-01 -2.26269528e-01 -7.50810027e-01
2.34245971e-01 -5.73783696e-01 7.14039087e-01 -1.11539280e+00
7.56092519e-02 4.89216417e-01 1.08601511e+00 -7.66592264e-01
1.54576838e-01 -5.32752216e-01 -5.35999060e-01 7.91717172e-01
-6.04576766e-01 -6.46607280e-02 5.26606083e-01 1.82064518e-01
-2.15698138e-01 -3.44406724e-01 1.20278203e+00 -5.61517715e-01
-1.17562294e+00 4.06459838e-01 -3.08466136e-01 3.56718659e-01
9.69376326e-01 -2.51500070e-01 -4.98345733e-01 -1.12597927e-01
-1.13750315e+00 2.93814093e-01 -5.94072565e-02 4.57221538e-01
7.44012594e-01 -1.23259830e+00 -5.98302484e-01 1.70419440e-01
-1.99701870e-03 -3.08393110e-02 -4.43724655e-02 6.08599961e-01
-9.60424423e-01 8.05577457e-01 -3.55539650e-01 -8.34285021e-01
-1.07664871e+00 3.55789423e-01 8.31019044e-01 -3.68974775e-01
-5.77920020e-01 1.43566060e+00 6.65052593e-01 -6.12002194e-01
7.57808447e-01 -5.26320696e-01 -3.20485473e-01 1.20599598e-01
5.90802431e-01 2.58724481e-01 2.18904629e-01 -3.93956482e-01
-6.70441687e-01 4.97719646e-01 -3.04686159e-01 2.79591858e-01
1.38001347e+00 2.96061218e-01 1.58330634e-01 -3.78442071e-02
1.59184599e+00 -2.40594447e-01 -1.37617528e+00 9.25020576e-02
1.31162211e-01 -7.45742097e-02 6.69993907e-02 -9.73361552e-01
-1.48778081e+00 1.14199317e+00 9.02084351e-01 1.29312500e-01
6.84147358e-01 2.09913641e-01 4.11027372e-01 5.72952569e-01
4.90611196e-02 -9.76893783e-01 2.96762943e-01 8.50280523e-01
9.53853369e-01 -1.39153600e+00 -1.82187885e-01 1.52099118e-01
-4.56231922e-01 1.14778936e+00 9.97512102e-01 -7.17121601e-01
3.31799686e-01 2.92348951e-01 9.68067814e-03 -3.91725898e-01
-5.97512841e-01 -5.48338473e-01 5.45366585e-01 6.84414387e-01
5.65598369e-01 3.97610925e-02 2.56031841e-01 2.37387046e-01
-6.49250746e-02 -5.08054256e-01 3.31025034e-01 4.58052188e-01
-5.83006322e-01 -6.79049551e-01 -1.56773105e-02 3.91715020e-01
-7.72421181e-01 -4.77443218e-01 -4.82369155e-01 1.29823041e+00
2.63629138e-01 6.05757058e-01 1.53818354e-01 -2.72169501e-01
4.50733602e-01 -2.74338514e-01 6.81196511e-01 -6.18856072e-01
-7.27603495e-01 -2.55648375e-01 -5.58740571e-02 -8.69540989e-01
-1.84620783e-01 -9.13376957e-02 -1.24817872e+00 -2.63146847e-01
-1.74156696e-01 -8.18348154e-02 7.42348015e-01 7.86116779e-01
-9.18103009e-02 9.36026394e-01 3.00592542e-01 -1.05459356e+00
-3.80754650e-01 -9.30889785e-01 -3.78244162e-01 1.46509558e-01
5.63225925e-01 -3.38576168e-01 -5.06775320e-01 -2.44352549e-01] | [9.369797706604004, 1.7578859329223633] |
58fd8527-b58b-4662-9dec-631b1135a042 | the-effects-of-skin-lesion-segmentation-on | 2008.12602 | null | https://arxiv.org/abs/2008.12602v1 | https://arxiv.org/pdf/2008.12602v1.pdf | The Effects of Skin Lesion Segmentation on the Performance of Dermatoscopic Image Classification | Malignant melanoma (MM) is one of the deadliest types of skin cancer. Analysing dermatoscopic images plays an important role in the early detection of MM and other pigmented skin lesions. Among different computer-based methods, deep learning-based approaches and in particular convolutional neural networks have shown excellent classification and segmentation performances for dermatoscopic skin lesion images. These models can be trained end-to-end without requiring any hand-crafted features. However, the effect of using lesion segmentation information on classification performance has remained an open question. In this study, we explicitly investigated the impact of using skin lesion segmentation masks on the performance of dermatoscopic image classification. To do this, first, we developed a baseline classifier as the reference model without using any segmentation masks. Then, we used either manually or automatically created segmentation masks in both training and test phases in different scenarios and investigated the classification performances. Evaluated on the ISIC 2017 challenge dataset which contained two binary classification tasks (i.e. MM vs. all and seborrheic keratosis (SK) vs. all) and based on the derived area under the receiver operating characteristic curve scores, we observed four main outcomes. Our results show that 1) using segmentation masks did not significantly improve the MM classification performance in any scenario, 2) in one of the scenarios (using segmentation masks for dilated cropping), SK classification performance was significantly improved, 3) removing all background information by the segmentation masks significantly degraded the overall classification performance, and 4) in case of using the appropriate scenario (using segmentation for dilated cropping), there is no significant difference of using manually or automatically created segmentation masks. | ['Isabella Ellinger', 'Georg Langs', 'Rupert Ecker', 'Philipp Tschandl', 'Amirreza Mahbod'] | 2020-08-28 | null | null | null | null | ['skin-lesion-segmentation'] | ['medical'] | [ 6.29349709e-01 1.94471523e-01 8.85242224e-02 5.92462858e-03
-4.68045354e-01 -5.95896661e-01 7.15526223e-01 4.26555932e-01
-9.15185153e-01 5.65398991e-01 -3.49523008e-01 -7.23484993e-01
-2.69506037e-01 -6.73312962e-01 -4.98880208e-01 -9.84322608e-01
2.14537382e-01 5.84950894e-02 4.20306742e-01 1.64639935e-01
4.18337345e-01 5.72633147e-01 -1.63423777e+00 3.35340440e-01
1.15675044e+00 8.89235973e-01 1.79214746e-01 9.47612762e-01
-3.62341791e-01 4.35097277e-01 -6.47916973e-01 -4.14800823e-01
4.69104201e-01 -4.66188461e-01 -8.52707505e-01 2.54989833e-01
5.76553285e-01 1.04022235e-01 3.17929149e-01 1.00664663e+00
6.81891143e-01 -4.60660636e-01 8.19823682e-01 -6.22649789e-01
5.50119728e-02 1.57084838e-01 -6.33898795e-01 5.19235693e-02
1.49429068e-01 5.14836609e-01 3.93809646e-01 -2.23612487e-01
8.43340337e-01 6.37262583e-01 7.93525219e-01 4.77509081e-01
-1.44812703e+00 -2.10834309e-01 -1.61580667e-01 1.58538654e-01
-1.25633681e+00 -1.27659544e-01 1.58952475e-01 -8.43815506e-01
6.17933989e-01 6.72739685e-01 8.11331272e-01 8.65512788e-01
1.11787684e-01 4.16146129e-01 1.90824819e+00 -8.41046035e-01
1.60567462e-01 5.84029913e-01 2.11125940e-01 5.11223733e-01
2.95205384e-01 1.12633966e-01 1.97320521e-01 1.13984626e-02
5.18156946e-01 -3.60236883e-01 -3.64975959e-01 3.71671617e-02
-8.66265178e-01 6.79743230e-01 2.96813816e-01 4.63679880e-01
-3.52128804e-01 -1.68613285e-01 4.20095593e-01 2.26071060e-01
3.93640012e-01 4.88739371e-01 -1.30852148e-01 4.66087461e-02
-1.04956770e+00 -1.75759554e-01 6.48276031e-01 -4.40466143e-02
4.37029272e-01 -4.04990196e-01 -3.77800196e-01 9.33392107e-01
-2.01444328e-01 2.06827849e-01 4.56255764e-01 -1.42343774e-01
1.89484924e-01 8.80670249e-01 -6.50827214e-02 -4.07052338e-01
-7.70078719e-01 -4.32948560e-01 -6.38372302e-01 5.61975241e-01
9.06217396e-01 -4.57412779e-01 -1.63752282e+00 1.26116097e+00
3.33920449e-01 -1.84439152e-01 -4.33033779e-02 9.34993684e-01
6.30747080e-01 -1.12919265e-03 3.38618129e-01 -3.32148783e-02
1.39399529e+00 -6.34359658e-01 -4.41040576e-01 1.29092216e-01
8.67490828e-01 -1.03198171e+00 1.07272792e+00 5.91723561e-01
-7.76478469e-01 -3.29981655e-01 -1.19111609e+00 3.01253229e-01
-5.97777486e-01 5.45709610e-01 4.17214155e-01 1.20178306e+00
-1.12414312e+00 5.28010666e-01 -7.65867651e-01 -8.13157558e-01
3.25855851e-01 4.99814481e-01 -4.74068582e-01 4.98952568e-02
-8.54028165e-01 1.10469806e+00 2.44182736e-01 -3.23697389e-03
-3.32914650e-01 -5.59531689e-01 -4.40957457e-01 -2.25455076e-01
1.46211401e-01 -5.22077322e-01 8.03868175e-01 -1.47181773e+00
-1.40869141e+00 1.17859626e+00 -1.21395864e-01 -5.89242041e-01
1.07656229e+00 4.10550535e-02 -3.90930712e-01 1.97020724e-01
-3.30458730e-01 4.64285225e-01 7.60116696e-01 -1.12300801e+00
-7.77006507e-01 -3.80197436e-01 1.98208809e-01 2.90302467e-02
-1.27480119e-01 -1.07844837e-01 -2.65101790e-01 -4.00978893e-01
-2.91237503e-01 -1.12244427e+00 -2.77700573e-01 -3.80525030e-02
-5.23756266e-01 6.01248406e-02 5.06189108e-01 -9.43775594e-01
1.17609727e+00 -2.08069277e+00 -2.14708433e-01 5.35580814e-01
1.43355042e-01 8.01042914e-01 -1.45577379e-02 3.19123328e-01
-1.87223792e-01 4.83314872e-01 -2.79856235e-01 -2.36187112e-02
-3.83949995e-01 -2.72047818e-01 5.39166033e-01 4.35310423e-01
2.79347479e-01 5.15146673e-01 -4.51910973e-01 -3.70539218e-01
7.85129666e-01 5.58632612e-01 -3.31678353e-02 -1.52663380e-01
-1.81153744e-01 4.64514375e-01 1.19917564e-01 3.70297641e-01
7.33003378e-01 -7.79111236e-02 2.46506035e-01 -2.65889198e-01
-1.93245068e-01 -1.70142233e-01 -1.22944629e+00 1.23667729e+00
-8.26925933e-01 8.97727251e-01 -7.48557299e-02 -5.72345197e-01
6.80636108e-01 3.11738282e-01 2.22604305e-01 -8.25203121e-01
7.36557543e-02 3.17704320e-01 5.06533563e-01 -7.43323684e-01
-6.68787509e-02 -1.18516043e-01 5.26895165e-01 1.23942792e-01
-3.10895771e-01 8.18545148e-02 3.55551720e-01 -1.75081059e-01
1.08491766e+00 -1.64629087e-01 3.13897580e-01 -2.26805776e-01
7.96195447e-01 3.07714641e-01 -9.78126656e-03 5.96808910e-01
-4.57251966e-02 8.05196583e-01 7.14536846e-01 -2.82515973e-01
-6.16237223e-01 -9.49681342e-01 -5.47743440e-01 4.05545771e-01
8.20199698e-02 -4.91067860e-03 -1.10924888e+00 -7.53780246e-01
-1.06447235e-01 6.07598007e-01 -8.50366890e-01 1.45037115e-01
-9.66043994e-02 -1.21334457e+00 4.35093880e-01 5.52949123e-02
5.83672166e-01 -7.60457695e-01 -7.38962054e-01 -1.57687604e-01
3.23321670e-01 -8.47085297e-01 2.29319874e-02 2.08711505e-01
-6.23311639e-01 -1.66173327e+00 -1.00296760e+00 -4.78828400e-01
8.95456672e-01 8.93334020e-03 6.39612854e-01 2.20751971e-01
-8.83102179e-01 2.48449877e-01 -3.94419432e-01 -5.87792754e-01
-5.40176630e-01 1.13516413e-01 -5.16715765e-01 3.59896928e-01
4.08157349e-01 -3.32515016e-02 -8.21655095e-01 1.38291374e-01
-1.00856102e+00 1.45128772e-01 1.00885248e+00 5.63451350e-01
3.86646539e-01 1.84120368e-02 2.41365105e-01 -1.49377692e+00
7.03003824e-01 -9.78519991e-02 -5.05279362e-01 3.56778532e-01
-6.01346076e-01 -1.31499678e-01 5.26392817e-01 -4.15729225e-01
-7.16920137e-01 2.56611466e-01 -2.63414174e-01 -6.59789592e-02
-6.41152918e-01 4.37996864e-01 2.24239320e-01 -4.42148030e-01
9.57921743e-01 1.49049647e-02 2.97191352e-01 -3.88615251e-01
-2.06920311e-01 9.42680299e-01 -6.51821941e-02 -3.11605196e-04
3.98694754e-01 4.42075521e-01 2.19469637e-01 -1.08921695e+00
-3.98123354e-01 -6.21825337e-01 -6.66087389e-01 -3.15033406e-01
1.11629903e+00 -4.05553907e-01 -3.50234747e-01 7.67143190e-01
-7.65017509e-01 -4.75355774e-01 -5.39812073e-02 3.42161566e-01
-8.03842768e-02 4.90523010e-01 -1.88957736e-01 -9.29460227e-01
-4.08236653e-01 -1.18709838e+00 7.06392765e-01 5.79543173e-01
-2.79288024e-01 -1.23330390e+00 -9.43258479e-02 4.04364049e-01
5.27263939e-01 7.54028380e-01 1.07560742e+00 -7.16982067e-01
-1.82172373e-01 -5.07514179e-01 -3.00509572e-01 5.40956438e-01
2.48042434e-01 3.73129368e-01 -1.07367241e+00 -2.06210002e-01
-4.74549353e-01 2.19357654e-01 9.12343919e-01 3.42779994e-01
1.09170306e+00 -1.89070553e-02 -2.93027312e-01 4.57976013e-01
1.77945673e+00 2.55247593e-01 8.61312747e-01 3.65238309e-01
3.80855262e-01 9.60041463e-01 3.96498501e-01 -3.89683992e-02
-1.89288378e-01 4.31903511e-01 3.64296407e-01 -6.53217137e-01
-4.77504969e-01 1.47328094e-01 6.61098137e-02 -2.91495740e-01
-3.98761690e-01 -1.28588796e-01 -9.41633821e-01 6.73745930e-01
-1.31028342e+00 -4.54898506e-01 -6.10876381e-01 2.60819602e+00
5.21235108e-01 3.29058021e-01 2.82352030e-01 3.19312155e-01
8.71170282e-01 -1.85239121e-01 -4.47365433e-01 -7.39076018e-01
-5.28579354e-02 7.61918604e-01 8.82454515e-01 3.55983704e-01
-1.11329913e+00 5.32274604e-01 5.24690580e+00 9.64682102e-01
-1.79122865e+00 -4.03198972e-02 6.81016684e-01 -7.30361640e-02
1.60506353e-01 -6.13819957e-02 -4.68324721e-01 6.78785205e-01
7.35528409e-01 3.56076866e-01 4.58268076e-02 2.10032746e-01
3.30589592e-01 -8.66389215e-01 -7.67319679e-01 7.04091430e-01
-1.74037188e-01 -1.33414924e+00 -9.51424241e-02 1.77218080e-01
5.70420384e-01 -1.40709341e-01 6.20819628e-02 -1.36942133e-01
-1.12013616e-01 -1.19790232e+00 1.98138475e-01 6.35857999e-01
1.08294547e+00 -4.59291577e-01 1.23457205e+00 2.02439547e-01
-5.20654678e-01 2.13350356e-01 1.12902485e-01 2.13983849e-01
-2.14810595e-01 6.77208781e-01 -1.13003898e+00 5.08000016e-01
3.91970485e-01 1.01913512e-01 -1.07204199e+00 1.54844320e+00
-1.91397414e-01 8.59039485e-01 -2.75593698e-01 -1.62279695e-01
3.14657331e-01 -1.19305588e-01 5.23288906e-01 1.39820659e+00
-4.40524407e-02 -5.45108676e-01 -4.60168600e-01 6.87688768e-01
5.12504697e-01 4.33125734e-01 -1.84885338e-01 -3.48491222e-02
2.06016041e-02 1.48050642e+00 -9.59583521e-01 3.27484589e-03
-3.79470527e-01 8.52446616e-01 -2.11888358e-01 4.20647502e-01
-5.77733696e-01 -6.45136654e-01 2.29708061e-01 8.69814217e-01
1.37445569e-01 2.88372546e-01 -5.21005929e-01 -5.67439437e-01
-4.54877615e-02 -5.80004513e-01 5.15773773e-01 -3.81333292e-01
-1.01911974e+00 4.78819162e-01 -2.55145639e-01 -8.96944404e-01
3.88796218e-02 -9.08311427e-01 -9.21221018e-01 1.23679519e+00
-1.43821120e+00 -1.16917157e+00 -7.07759202e-01 2.73638397e-01
2.03102320e-01 1.59555465e-01 7.23581195e-01 7.79008046e-02
-5.62639415e-01 6.98769689e-01 8.14932659e-02 1.32623225e-01
8.27797115e-01 -1.52779830e+00 -1.88213587e-01 7.03880131e-01
-9.15796235e-02 3.70435715e-01 5.82590759e-01 -4.56838280e-01
-6.09385192e-01 -8.94283414e-01 3.86912346e-01 -1.07081898e-01
3.83686662e-01 -1.39038503e-01 -5.69411993e-01 6.24204986e-02
1.01289287e-01 -2.91833788e-01 8.32662940e-01 -2.72851810e-03
-8.27194080e-02 5.82514703e-02 -1.46149862e+00 6.30157113e-01
4.20656323e-01 -7.50542060e-02 -3.21737118e-02 4.07157063e-01
-1.27691692e-02 -4.06677783e-01 -9.07861173e-01 6.45467460e-01
5.90830564e-01 -1.27732503e+00 6.98036790e-01 -3.19381654e-01
4.80954617e-01 -1.37659147e-01 3.09866309e-01 -1.20480680e+00
1.29320353e-01 -2.61023253e-01 4.97784317e-01 1.01105452e+00
7.38210082e-01 -9.64526117e-01 9.98514473e-01 5.89899302e-01
1.43261597e-01 -1.15643573e+00 -7.38125563e-01 -5.06786704e-01
2.01783597e-01 5.03537878e-02 7.46665075e-02 6.37279093e-01
-3.74847025e-01 -8.19436014e-02 2.72643805e-01 -4.76263911e-02
2.70373046e-01 -9.88557860e-02 8.22944641e-01 -1.02413857e+00
-2.37961501e-01 -6.39319003e-01 -7.18198001e-01 -9.90520269e-02
-3.85738194e-01 -8.86155844e-01 -5.00217199e-01 -1.86342645e+00
8.04631785e-02 -5.62649667e-01 -1.60489276e-01 4.03638095e-01
-4.48503554e-01 2.87185222e-01 2.93645322e-01 -6.57558739e-02
7.38281310e-02 -5.94773054e-01 1.20982611e+00 2.34496742e-02
-2.68517703e-01 4.39597458e-01 -5.06802738e-01 5.70194840e-01
9.27218318e-01 -1.88914500e-02 -2.08106279e-01 -8.88895467e-02
-5.67085445e-02 -1.79021403e-01 6.56881690e-01 -1.16242111e+00
1.87317163e-01 -4.28845808e-02 5.59783399e-01 -2.54250228e-01
1.12158649e-01 -6.70152664e-01 2.30358735e-01 8.55803967e-01
-1.54218420e-01 -5.44401228e-01 4.70508367e-01 3.85471880e-01
1.03059150e-02 -5.71966290e-01 1.15568912e+00 -9.58201960e-02
-7.72587299e-01 -2.30600491e-01 -5.34442544e-01 -3.08350891e-01
1.47664905e+00 -7.50171900e-01 -4.02111292e-01 -1.12697370e-01
-9.10018027e-01 -5.36866076e-02 7.36186147e-01 1.22661166e-01
1.18917078e-01 -3.82457644e-01 -6.56648815e-01 6.90039322e-02
1.90873623e-01 -8.56003836e-02 5.05263388e-01 1.34007442e+00
-9.94200349e-01 4.29345310e-01 -2.65013963e-01 -7.31739342e-01
-1.77852082e+00 1.91699907e-01 7.66147435e-01 -4.68713701e-01
-1.69412509e-01 8.14625621e-01 -1.97679818e-01 -2.77865350e-01
1.78420097e-01 -2.89431125e-01 -3.67560267e-01 1.05451196e-01
2.16051653e-01 3.63684982e-01 4.93547857e-01 -2.71815360e-01
-1.51121542e-01 7.21939027e-01 -2.87856817e-01 3.68257537e-02
9.24522221e-01 2.75072217e-01 -5.31233661e-03 2.57381529e-01
9.85099554e-01 1.85231909e-01 -7.97004223e-01 3.25265229e-01
8.29646885e-02 -4.79372650e-01 1.38882324e-01 -1.50317490e+00
-1.04711449e+00 9.95794237e-01 1.16481471e+00 2.70141542e-01
1.25004399e+00 -3.59866619e-01 4.40933645e-01 -2.81095356e-01
1.82864562e-01 -1.00712240e+00 -5.12513280e-01 -1.52003616e-01
4.32367325e-01 -1.19286156e+00 5.08372746e-02 -8.08085144e-01
-6.75908566e-01 1.11663675e+00 5.99998474e-01 -3.05401850e-02
4.70453203e-01 5.92602044e-02 3.45007896e-01 -1.00333206e-01
-3.26811612e-01 -6.33168936e-01 4.81582373e-01 7.37674475e-01
4.61518079e-01 1.66322172e-01 -9.23683167e-01 2.96368927e-01
-3.27789187e-02 9.60703716e-02 5.66943109e-01 7.51156807e-01
-9.96477306e-02 -1.22717357e+00 -3.31229180e-01 1.12021637e+00
-6.98241174e-01 2.68290229e-02 -8.32347870e-01 1.15854490e+00
5.31237841e-01 8.67730856e-01 2.51129828e-02 -3.45706701e-01
2.85343051e-01 4.74954434e-02 5.20285189e-01 -6.81982577e-01
-1.11399281e+00 -2.77297229e-01 2.78159529e-01 -2.09835634e-01
-3.37194830e-01 -5.36244154e-01 -9.42831874e-01 -1.35780677e-01
-5.28676271e-01 -2.68240459e-02 1.20166874e+00 8.41303110e-01
2.13511512e-01 4.90276039e-01 2.57733583e-01 -2.82488525e-01
-4.10977244e-01 -1.09855318e+00 -4.91219610e-01 5.06746471e-01
1.62466764e-01 -4.31116968e-01 -5.59494078e-01 3.50543554e-03] | [15.551468849182129, -3.0700836181640625] |
830ea51c-f5fc-4fc1-a487-ed312d77c216 | gpr1200-a-benchmark-for-general-purpose | 2111.13122 | null | https://arxiv.org/abs/2111.13122v1 | https://arxiv.org/pdf/2111.13122v1.pdf | GPR1200: A Benchmark for General-Purpose Content-Based Image Retrieval | Even though it has extensively been shown that retrieval specific training of deep neural networks is beneficial for nearest neighbor image search quality, most of these models are trained and tested in the domain of landmarks images. However, some applications use images from various other domains and therefore need a network with good generalization properties - a general-purpose CBIR model. To the best of our knowledge, no testing protocol has so far been introduced to benchmark models with respect to general image retrieval quality. After analyzing popular image retrieval test sets we decided to manually curate GPR1200, an easy to use and accessible but challenging benchmark dataset with a broad range of image categories. This benchmark is subsequently used to evaluate various pretrained models of different architectures on their generalization qualities. We show that large-scale pretraining significantly improves retrieval performance and present experiments on how to further increase these properties by appropriate fine-tuning. With these promising results, we hope to increase interest in the research topic of general-purpose CBIR. | ['Klaus Jung', 'Nico Hezel', 'Kai Uwe Barthel', 'Konstantin Schall'] | 2021-11-25 | null | null | null | null | ['content-based-image-retrieval'] | ['computer-vision'] | [ 3.26034911e-02 -4.74955410e-01 -3.65034431e-01 -5.54484665e-01
-1.06770205e+00 -3.45875055e-01 6.97020471e-01 6.78808987e-02
-7.09321380e-01 4.58190650e-01 -1.63759105e-02 -2.51857311e-01
-5.31221867e-01 -8.38100553e-01 -5.04990339e-01 -5.49143374e-01
-9.16347280e-02 5.78070283e-01 3.19732845e-01 -4.06066269e-01
3.92981797e-01 8.32111835e-01 -1.70480466e+00 2.56644547e-01
4.39780563e-01 1.24161935e+00 1.85608357e-01 3.23775768e-01
3.47209126e-01 4.35764879e-01 -5.84700286e-01 -2.82449424e-01
4.46239740e-01 -2.67613292e-01 -1.09196067e+00 -5.46863496e-01
6.96108043e-01 -5.40974319e-01 -4.63705212e-01 8.33476961e-01
1.01123941e+00 3.87503654e-01 8.38840663e-01 -8.89955163e-01
-1.27519703e+00 2.57464528e-01 -2.00545952e-01 4.80922192e-01
1.11230619e-01 -1.24860838e-01 1.19053721e+00 -7.48266041e-01
7.79270530e-01 9.08873737e-01 5.78324437e-01 5.28910816e-01
-9.49863911e-01 -5.86200535e-01 -2.60405093e-01 5.14669299e-01
-1.73622549e+00 -3.12871784e-01 6.20599985e-01 6.15324453e-02
9.35051262e-01 3.78611326e-01 2.46104747e-01 1.07028198e+00
3.54300551e-02 8.14787209e-01 9.55105841e-01 -5.02304673e-01
-9.68370680e-03 2.01654494e-01 3.00750971e-01 4.82902139e-01
2.02327728e-01 2.02380210e-01 -1.71812445e-01 -1.72819365e-02
5.99001467e-01 7.52060413e-02 -5.19419551e-01 -6.40097022e-01
-1.15105724e+00 8.64736676e-01 1.27726746e+00 8.75736535e-01
-1.81945786e-01 2.96567082e-01 4.43825245e-01 6.70282602e-01
2.16319263e-01 7.63350606e-01 -4.54897821e-01 3.93189371e-01
-1.00539112e+00 1.65879175e-01 5.33926427e-01 8.49456191e-01
7.04847872e-01 -3.37367386e-01 -3.32229525e-01 1.36726224e+00
1.16794772e-01 5.49597859e-01 6.25364304e-01 -7.79458404e-01
-9.37968567e-02 3.76919538e-01 -8.61818045e-02 -1.21839106e+00
-3.80103230e-01 -5.96743166e-01 -9.40818429e-01 -5.25647476e-02
4.36296493e-01 5.00871003e-01 -1.03353214e+00 1.53841114e+00
-2.20006138e-01 -1.82949990e-01 -1.91214960e-02 1.24212205e+00
1.19554389e+00 4.98146743e-01 9.84621886e-03 2.99763292e-01
1.35258901e+00 -8.69912028e-01 -2.56039947e-01 1.47637621e-01
7.16739833e-01 -8.96728218e-01 1.18922949e+00 2.71406740e-01
-8.52896631e-01 -7.10036993e-01 -1.18619108e+00 -3.33080798e-01
-8.59057963e-01 1.76873356e-01 7.08713233e-01 5.63013077e-01
-1.44990134e+00 6.45405710e-01 -4.30386096e-01 -6.49563134e-01
4.61486578e-01 4.64490324e-01 -5.56748748e-01 -4.90986228e-01
-1.39867473e+00 1.11631596e+00 3.41874063e-01 1.63824454e-01
-9.43166077e-01 -5.01862049e-01 -5.62954664e-01 2.38978222e-01
9.51140225e-02 -7.99578488e-01 1.07897961e+00 -8.42768192e-01
-1.12619483e+00 1.22258115e+00 1.32250458e-01 -6.14103675e-01
2.76569098e-01 -5.94338737e-02 -4.10196722e-01 1.33910164e-01
1.14586279e-01 1.27892232e+00 4.79847699e-01 -1.29274416e+00
-3.69765937e-01 -3.70740473e-01 1.73820764e-01 1.27085328e-01
-6.47626460e-01 -1.25313073e-01 -9.30710793e-01 -5.67128062e-01
-8.14222246e-02 -1.06652403e+00 -1.34651765e-01 2.44616177e-02
-1.15032889e-01 -3.83838087e-01 6.44586325e-01 -3.38128477e-01
8.61650467e-01 -2.02757335e+00 -1.97401956e-01 4.65691090e-01
-1.19470291e-01 4.22501236e-01 -7.20797598e-01 3.34840566e-01
-1.19669847e-01 2.92194009e-01 2.32566789e-01 -3.13328430e-02
-2.53028572e-02 7.58829713e-02 -3.73772144e-01 4.76743072e-01
-1.36761799e-01 1.18536973e+00 -6.59084380e-01 -3.87352824e-01
3.12520713e-01 7.73987770e-01 -4.70389128e-01 -5.97551130e-02
-1.59924768e-03 7.22282901e-02 -5.80573440e-01 6.03111207e-01
6.59341514e-01 -5.74760914e-01 -3.63956094e-01 -2.96414047e-01
4.79518622e-01 1.11408439e-02 -5.98179817e-01 2.00611329e+00
-5.27887225e-01 9.13252711e-01 -5.02338052e-01 -1.06659043e+00
7.38792360e-01 1.33527026e-01 4.35901463e-01 -1.30848587e+00
2.29803532e-01 5.69765806e-01 1.36004761e-01 -2.14299545e-01
7.99242139e-01 3.01043838e-01 1.42327398e-01 2.74107218e-01
1.83377847e-01 -1.08331703e-01 2.84509659e-01 1.26934990e-01
9.71526325e-01 -3.60033453e-01 -2.46412819e-03 -3.82047862e-01
6.61858916e-01 1.48769855e-01 -3.68305445e-01 1.09344590e+00
-1.39460415e-01 9.20521975e-01 -2.05412313e-01 -7.48631179e-01
-9.76883650e-01 -7.12601960e-01 -6.35521054e-01 1.37583089e+00
4.10439610e-01 -2.13223904e-01 -5.60570598e-01 -4.58976775e-01
-1.79776117e-01 3.71658504e-01 -7.19574332e-01 -2.41092563e-01
-3.76782477e-01 -8.08271110e-01 5.14140368e-01 3.87002975e-01
7.46474862e-01 -1.46064663e+00 -2.97699124e-01 -1.35103881e-01
2.41805762e-02 -9.49871898e-01 -3.15026313e-01 6.55785352e-02
-6.41637385e-01 -1.08260131e+00 -1.36765695e+00 -1.14215147e+00
4.67114925e-01 6.06990159e-01 1.46024680e+00 6.78382516e-01
-5.15904069e-01 6.21793747e-01 -5.11444807e-01 -1.37665346e-01
-1.10882208e-01 5.42700231e-01 -3.34160656e-01 -4.45910841e-01
5.21847606e-01 -1.69528157e-01 -1.10784853e+00 7.30627656e-01
-1.17359412e+00 -5.90974689e-01 7.27829456e-01 9.80618536e-01
6.90028846e-01 9.58378688e-02 4.08625752e-01 -5.01517773e-01
8.93036425e-01 -1.57904595e-01 -6.33543670e-01 5.19987762e-01
-8.01556587e-01 1.21338852e-01 2.39666373e-01 -3.34930718e-01
-4.73503530e-01 -3.53129238e-01 -3.59483749e-01 -3.43446225e-01
-2.56781131e-01 6.03854299e-01 3.34903419e-01 -4.78934556e-01
1.08245099e+00 2.84731239e-01 -2.30143577e-01 -2.89838940e-01
4.21350956e-01 6.58585072e-01 3.51629734e-01 -5.98618209e-01
5.12486160e-01 4.03526604e-01 -5.16020646e-03 -7.01039970e-01
-6.79020941e-01 -7.88813114e-01 -3.16410780e-01 2.49949157e-01
7.73472190e-01 -8.64160240e-01 -6.06093466e-01 2.82883197e-01
-8.82357538e-01 -3.57885331e-01 6.17126487e-02 3.04180801e-01
-4.10368443e-01 2.16484725e-01 -6.87627077e-01 -1.71471044e-01
-6.51760459e-01 -1.42577088e+00 1.29224968e+00 1.01729417e-02
7.61947036e-02 -9.12843287e-01 1.25239164e-01 2.91043013e-01
9.19896066e-01 -3.55767906e-01 7.57239163e-01 -8.46465588e-01
-7.88718462e-01 -4.26168352e-01 -6.22174144e-01 4.00077313e-01
-4.81154397e-02 -2.78831214e-01 -1.03786886e+00 -5.68599224e-01
-2.91387677e-01 -6.05879724e-01 1.27071559e+00 4.24603730e-01
1.57146966e+00 2.03432620e-01 -4.65465844e-01 5.55368721e-01
1.67360079e+00 2.65762031e-01 9.86552238e-01 7.62576759e-01
3.06847811e-01 2.98334926e-01 4.40841645e-01 -2.50606328e-01
7.83556700e-02 1.00076151e+00 3.93655032e-01 -3.14689368e-01
-2.24472821e-01 1.13425747e-01 -2.36553267e-01 2.42161140e-01
-1.66470572e-01 -5.38278222e-01 -9.54768538e-01 7.07642972e-01
-1.49117303e+00 -9.42109406e-01 3.77448499e-01 2.12804914e+00
6.45169914e-01 -1.74429461e-01 -2.64327645e-01 -1.00257106e-01
4.19398308e-01 5.04140258e-02 -3.23836118e-01 -5.67478612e-02
-9.75041538e-02 3.27431232e-01 4.71325576e-01 2.54566610e-01
-1.33848834e+00 1.02246678e+00 6.50810575e+00 1.28750360e+00
-1.37170625e+00 1.09006779e-03 8.55121732e-01 1.29655093e-01
-1.42859772e-01 -3.82546484e-01 -8.24552596e-01 4.14513424e-02
8.02442908e-01 1.53183281e-01 3.96578789e-01 1.10081637e+00
-4.03800994e-01 1.62450239e-01 -1.30831289e+00 1.19691277e+00
2.29357749e-01 -1.41796756e+00 3.74009490e-01 1.28507480e-01
7.11210251e-01 4.95239943e-01 4.60586160e-01 5.79880357e-01
1.58692747e-01 -1.33361232e+00 1.50033325e-01 3.10781896e-01
6.75866246e-01 -5.78576803e-01 1.04328597e+00 -5.26652969e-02
-8.20882320e-01 1.40535399e-01 -8.93107533e-01 5.20789087e-01
-2.78476208e-01 2.44354308e-01 -6.97471619e-01 4.69988704e-01
1.10606790e+00 6.03759706e-01 -9.95705962e-01 1.36353016e+00
5.59679344e-02 1.21427201e-01 -3.95097524e-01 -6.19571134e-02
4.59552109e-01 1.09445922e-01 2.95316856e-02 1.05293000e+00
4.72553521e-01 -1.27272040e-01 -8.16091299e-02 5.63261390e-01
-3.73078465e-01 4.11300540e-01 -8.29564035e-01 1.38413876e-01
2.35987455e-01 1.29412770e+00 -7.42463112e-01 -2.59997487e-01
-3.12748104e-01 9.05244589e-01 2.55137444e-01 5.54065228e-01
-5.95918953e-01 -3.05170327e-01 4.45032448e-01 5.71155325e-02
3.72849673e-01 1.51863754e-01 1.41882181e-01 -1.03531885e+00
-1.35796309e-01 -9.89070415e-01 5.94816506e-01 -1.01888549e+00
-1.58041191e+00 1.09812593e+00 1.02539835e-02 -1.10204864e+00
-3.50175500e-01 -8.83645177e-01 -1.28774345e-01 8.60391736e-01
-1.82374763e+00 -1.06874502e+00 -4.30909216e-01 9.16878343e-01
3.53220552e-01 -4.55345511e-01 1.04446149e+00 3.79361093e-01
-1.49797797e-01 8.77310157e-01 4.10077810e-01 3.74104440e-01
9.27323759e-01 -9.02689397e-01 9.00300369e-02 3.42915148e-01
6.10114217e-01 1.13628185e+00 4.12996143e-01 -8.13658815e-03
-1.15199387e+00 -8.21702540e-01 5.23127973e-01 -4.87767994e-01
5.21399379e-01 -2.87062116e-02 -8.61920595e-01 5.16720951e-01
5.49487650e-01 2.52371848e-01 5.11178374e-01 2.43317574e-01
-5.61782956e-01 -3.83784831e-01 -1.07123160e+00 6.32782459e-01
9.00793672e-01 -6.66345537e-01 -4.01182920e-01 5.08339882e-01
4.86768544e-01 -3.06713521e-01 -9.83162761e-01 7.92201817e-01
5.93868554e-01 -1.07738101e+00 1.48296905e+00 -3.97721916e-01
2.38252357e-01 -6.43627495e-02 -3.70407999e-01 -1.14543486e+00
-1.61327839e-01 1.52692080e-01 6.17002606e-01 8.02879333e-01
3.94291669e-01 -6.32642627e-01 8.69334698e-01 4.05807018e-01
6.54289871e-02 -7.31821239e-01 -7.28448033e-01 -8.10908616e-01
3.78816158e-01 -2.98486471e-01 5.47232211e-01 9.06872153e-01
-4.19891447e-01 4.22275424e-01 -1.75660938e-01 3.73334595e-04
2.53049970e-01 4.11067724e-01 8.10604036e-01 -1.26544774e+00
4.26877402e-02 -8.51887882e-01 -5.74820757e-01 -1.12366915e+00
1.99334785e-01 -1.07211161e+00 -2.71766763e-02 -1.44501054e+00
4.77118403e-01 -6.88701570e-01 -7.23808348e-01 5.03882766e-01
3.37099358e-02 8.99373353e-01 2.45093871e-02 4.24989104e-01
-9.11432803e-01 3.74210328e-01 1.30378556e+00 -5.70621014e-01
4.23304252e-02 -1.18182994e-01 -6.93212509e-01 1.71397835e-01
7.57480025e-01 -3.77224237e-01 -7.39035070e-01 -5.82836092e-01
1.85254186e-01 -2.48050421e-01 4.25970525e-01 -1.03868854e+00
1.69644773e-01 3.32536250e-01 5.94048798e-01 -5.94037056e-01
3.63965422e-01 -1.07378268e+00 5.36333621e-02 2.32638150e-01
-7.44239867e-01 -6.96270773e-03 2.85377651e-01 3.49229902e-01
-6.09519899e-01 -3.59267622e-01 6.40453756e-01 -2.26675093e-01
-1.05639696e+00 5.13507009e-01 2.07344210e-03 -2.01413244e-01
6.63789392e-01 -6.66534081e-02 -3.86292428e-01 -6.50376141e-01
-6.02415383e-01 9.16762948e-02 5.83431125e-01 6.01727724e-01
6.04065776e-01 -1.39555061e+00 -5.58613539e-01 8.14441144e-02
5.40941715e-01 -3.47806334e-01 1.11492641e-01 5.27113378e-01
-8.06968093e-01 9.87401247e-01 -3.18777651e-01 -9.14069116e-01
-1.19371212e+00 9.94591475e-01 5.09027600e-01 -4.58939612e-01
-3.80592853e-01 7.74034262e-01 2.61678189e-01 -4.81290191e-01
3.30252290e-01 -2.35458031e-01 -3.84816051e-01 -8.70158300e-02
6.23912156e-01 -1.08916322e-02 5.15760005e-01 -5.89418530e-01
-3.43245834e-01 7.55165935e-01 -3.44129920e-01 1.05923988e-01
1.32408345e+00 -7.62384851e-03 -2.09655300e-01 -1.72229111e-01
1.75718904e+00 -3.92884105e-01 -4.68502939e-01 -2.18252853e-01
6.64214417e-02 -4.72246230e-01 1.78690746e-01 -8.44212234e-01
-1.43870473e+00 8.51774454e-01 1.13125193e+00 1.69236213e-01
1.33132720e+00 -2.39975061e-02 6.11862242e-01 1.10389245e+00
6.92376673e-01 -7.27884412e-01 2.38262787e-01 4.35601443e-01
1.15988851e+00 -1.65613961e+00 -6.95396513e-02 1.03867657e-01
-2.59030014e-01 9.25403953e-01 3.67989421e-01 -3.71478260e-01
8.02667797e-01 -2.37172827e-01 3.08016747e-01 -3.81233364e-01
-3.56800884e-01 -3.23416829e-01 7.67653048e-01 5.34681022e-01
5.26209414e-01 -3.97061497e-01 -3.71408284e-01 -8.98338184e-02
-1.40818566e-01 2.15341851e-01 -1.60199404e-01 4.60315883e-01
-2.46292263e-01 -1.31921601e+00 -1.87828705e-01 3.75926703e-01
-6.70483589e-01 -4.57069874e-01 -3.43081534e-01 1.20906949e+00
-2.83589184e-01 7.87634790e-01 -1.01931535e-01 -1.73033789e-01
1.88431457e-01 -1.59920052e-01 5.04850149e-01 -2.50811100e-01
-5.31973362e-01 -2.13680938e-01 -1.48386404e-01 -6.33399725e-01
-6.55499220e-01 -9.86067131e-02 -6.80569828e-01 -3.53156239e-01
-4.26104605e-01 3.15396667e-01 6.75171494e-01 6.49958730e-01
3.99263978e-01 1.94504589e-01 9.97488648e-02 -9.20002043e-01
-4.20066983e-01 -1.06963098e+00 -4.40869778e-01 7.01641679e-01
2.05739230e-01 -6.08822107e-01 -1.78679034e-01 -3.70303363e-01] | [10.655444145202637, 0.6212829947471619] |
4cd0c174-7310-44aa-b185-4ca16eceefc5 | syntax-aware-multi-task-graph-convolutional | null | null | https://aclanthology.org/D19-6204 | https://aclanthology.org/D19-6204.pdf | Syntax-aware Multi-task Graph Convolutional Networks for Biomedical Relation Extraction | In this paper we tackle two unique challenges in biomedical relation extraction. The first challenge is that the contextual information between two entity mentions often involves sophisticated syntactic structures. We propose a novel graph convolutional networks model that incorporates dependency parsing and contextualized embedding to effectively capture comprehensive contextual information. The second challenge is that most of the benchmark data sets for this task are quite imbalanced because more than 80{\%} mention pairs are negative instances (i.e., no relations). We propose a multi-task learning framework to jointly model relation identification and classification tasks to propagate supervision signals from each other and apply a focal loss to focus training on ambiguous mention pairs. By applying these two strategies, experiments show that our model achieves state-of-the-art F-score on the 2013 drug-drug interaction extraction task. | ['Heng Ji', 'Diya Li'] | 2019-11-01 | null | null | null | ws-2019-11 | ['drug-drug-interaction-extraction'] | ['natural-language-processing'] | [ 4.17940110e-01 5.08443415e-01 -6.72266185e-01 -5.36846280e-01
-9.65327740e-01 -3.02685529e-01 3.55403483e-01 6.81897283e-01
-4.95867074e-01 1.04643869e+00 6.36108145e-02 -5.70231259e-01
-1.32900938e-01 -6.98643506e-01 -8.05117965e-01 -5.28836548e-01
-1.69195458e-01 5.13463676e-01 2.50225812e-02 -1.94855750e-01
-1.66159198e-01 2.97092259e-01 -6.05357587e-01 4.63667989e-01
1.04604530e+00 6.95784271e-01 -2.43739650e-01 4.55537409e-01
-6.45341724e-02 9.40561056e-01 -7.05627143e-01 -9.43465889e-01
-1.24578565e-01 -3.47928047e-01 -1.17173052e+00 -2.97639400e-01
9.72512236e-04 -6.28315285e-02 -2.40803435e-01 1.10733819e+00
6.25443041e-01 -2.67164707e-01 6.43031359e-01 -1.07782233e+00
-6.04071736e-01 8.58093798e-01 -6.90161109e-01 3.54647309e-01
3.74055207e-01 5.47684692e-02 1.46235049e+00 -7.48000264e-01
9.21572864e-01 9.49373722e-01 7.79122770e-01 3.99095058e-01
-1.22754407e+00 -6.92055225e-01 3.76300454e-01 3.11123490e-01
-1.33294082e+00 -2.64940441e-01 6.54928803e-01 -2.65114427e-01
1.36192501e+00 1.71864018e-01 2.13066190e-01 1.29798234e+00
3.48018795e-01 6.27309859e-01 6.75488770e-01 -2.62483388e-01
-7.28031397e-02 -6.01486526e-02 4.77536708e-01 7.31599689e-01
4.93262291e-01 -3.06233466e-01 -2.72513390e-01 -2.74892032e-01
2.35006854e-01 -5.70549536e-03 -2.25551441e-01 1.38881251e-01
-8.61592472e-01 8.33747208e-01 5.37348270e-01 3.68292302e-01
-4.66774642e-01 -9.94144380e-02 4.49180871e-01 1.41689271e-01
6.87758505e-01 4.83624935e-01 -9.39095736e-01 1.95360482e-01
-4.04153019e-01 7.05206767e-02 8.89469624e-01 9.09128845e-01
4.65048999e-01 -6.34780407e-01 -2.93972582e-01 7.97601283e-01
2.64418900e-01 3.11802328e-02 1.15925983e-01 -3.07110965e-01
9.47800219e-01 8.21098626e-01 -3.47037911e-01 -8.80804420e-01
-7.84478068e-01 -5.18894076e-01 -9.80581641e-01 -2.95076728e-01
3.04910690e-01 -4.29961801e-01 -1.05931020e+00 1.85223162e+00
5.16257405e-01 3.39084715e-01 2.12962925e-01 5.27480066e-01
1.57507396e+00 2.24974856e-01 6.35796607e-01 -1.33788228e-01
1.73459089e+00 -9.38630164e-01 -1.23673153e+00 -4.92611706e-01
9.63499129e-01 -5.41794896e-01 5.43003976e-01 -5.41687720e-02
-8.76543880e-01 -4.48566936e-02 -1.17682695e+00 -3.91787440e-01
-5.01889706e-01 1.10627323e-01 9.01111126e-01 2.15940222e-01
-5.21926403e-01 7.31319785e-01 -9.65081275e-01 2.24257205e-02
7.94424236e-01 5.91691792e-01 -6.53654397e-01 -1.18441872e-01
-1.67938673e+00 1.13961148e+00 4.21103418e-01 2.14532524e-01
-2.78710365e-01 -9.85353768e-01 -1.11371350e+00 1.87152058e-01
5.43975413e-01 -8.29855502e-01 9.38233912e-01 -3.72164428e-01
-1.13687706e+00 1.04722381e+00 -2.53823429e-01 -4.88612354e-01
2.72332668e-01 -2.75374264e-01 -4.68213797e-01 1.19966939e-02
2.28081435e-01 8.82415175e-02 3.65522355e-02 -8.82883489e-01
-3.88306022e-01 -4.82707113e-01 6.53618500e-02 4.47661504e-02
-1.35840133e-01 2.22620636e-01 -5.10806322e-01 -4.98498261e-01
-7.72864446e-02 -6.52333438e-01 -5.78331828e-01 -3.59972924e-01
-1.04802442e+00 -5.21916926e-01 4.67146367e-01 -6.91485226e-01
1.34888029e+00 -1.76098192e+00 2.36554071e-01 7.05284104e-02
6.10043883e-01 4.34164196e-01 -2.11423755e-01 3.58359456e-01
-4.77467835e-01 3.99100512e-01 -3.11528385e-01 -4.70458895e-01
-1.53234094e-01 3.00399512e-01 7.47006014e-02 2.39805102e-01
9.18315947e-01 1.21385503e+00 -9.85172689e-01 -5.55806637e-01
-2.65280634e-01 7.81061590e-01 -3.86002958e-01 3.55910242e-01
-1.75929293e-01 5.54110408e-01 -5.62541246e-01 6.47396922e-01
7.30168462e-01 -6.95710242e-01 7.16964424e-01 -4.60649639e-01
5.35802245e-01 8.89149308e-01 -8.51440966e-01 1.67056048e+00
-2.38289297e-01 6.92912266e-02 -6.53949678e-02 -1.11473155e+00
6.27712488e-01 4.61496115e-01 5.68857789e-01 -4.02664930e-01
2.19901159e-01 9.59298834e-02 9.85157266e-02 -6.17692888e-01
3.90918329e-02 -9.06745493e-02 -1.68140382e-01 1.15689725e-01
4.07097220e-01 2.70787090e-01 1.24877915e-01 2.75254458e-01
1.52766693e+00 5.32146357e-02 6.55309498e-01 -3.38475853e-02
5.32822371e-01 -3.55293363e-01 1.31607437e+00 3.21772337e-01
-1.54914021e-01 4.43731368e-01 1.12385762e+00 -4.29254472e-01
-4.52639580e-01 -7.47736573e-01 -2.00035781e-01 8.14390361e-01
-8.55194107e-02 -7.84018993e-01 -4.67265725e-01 -1.38176191e+00
-8.60417485e-02 3.29689771e-01 -8.73762548e-01 -2.94549048e-01
-6.31111026e-01 -1.45238876e+00 5.77897847e-01 6.51900411e-01
9.47309509e-02 -8.48996460e-01 8.42664018e-02 3.68160605e-01
-2.29646131e-01 -1.61065304e+00 -2.64411479e-01 6.72560930e-01
-5.71345866e-01 -1.60796309e+00 -2.75132507e-01 -7.58133471e-01
7.36057341e-01 -3.78963619e-01 1.59787083e+00 2.06453502e-01
-2.16092482e-01 -4.53041345e-01 -2.76738971e-01 -5.74880779e-01
-2.72533149e-01 4.44764137e-01 -3.80127102e-01 -3.24724436e-01
7.96009123e-01 -4.65746045e-01 -6.50929034e-01 -2.21324693e-02
-7.72303581e-01 -8.56241658e-02 6.14862263e-01 1.06122911e+00
8.03483903e-01 -2.22628057e-01 9.83537078e-01 -1.68498576e+00
5.30025959e-01 -7.17106640e-01 -2.24123329e-01 4.94027644e-01
-5.61319768e-01 1.03249550e-01 5.93499243e-01 -1.75200731e-01
-8.10044050e-01 1.24971263e-01 -5.80575764e-01 1.20429575e-01
-1.34995684e-01 8.91010344e-01 -5.42378962e-01 2.76821166e-01
3.90232503e-01 -4.32594150e-01 -3.56567651e-01 -3.99337858e-01
3.51987273e-01 5.58605909e-01 4.05361414e-01 -3.82119417e-01
4.62540120e-01 3.93878073e-02 1.68932348e-01 -4.17621046e-01
-1.29514134e+00 -3.96585315e-01 -7.34166801e-01 6.95620418e-01
1.11139166e+00 -1.06666338e+00 -8.09382439e-01 2.93587089e-01
-1.38178444e+00 -2.80270457e-01 2.51519363e-02 5.25531948e-01
5.51002566e-03 3.35413516e-01 -1.14689732e+00 -3.86084408e-01
-8.23375642e-01 -1.34105062e+00 1.04875803e+00 9.53911543e-02
-1.83995426e-01 -1.06908190e+00 2.11771443e-01 4.33077902e-01
-5.22925481e-02 5.17428219e-01 1.12644386e+00 -1.18236661e+00
-3.75710189e-01 -1.53382629e-01 -3.93512666e-01 -7.08931834e-02
4.40786898e-01 -6.89487606e-02 -9.46310759e-01 -7.01546669e-02
-3.23928982e-01 -3.13531399e-01 1.03575516e+00 1.66234583e-01
1.03285289e+00 -3.25251609e-01 -6.89234495e-01 7.14367390e-01
1.20784938e+00 1.36397406e-01 5.70294738e-01 1.65139124e-01
1.07122946e+00 4.99358118e-01 5.20106018e-01 8.24257731e-02
7.13811278e-01 7.03072250e-01 3.42370033e-01 -3.94160628e-01
6.35457342e-04 -8.73139054e-02 -7.08748028e-02 7.34987438e-01
-7.53350696e-03 -3.06213886e-01 -9.43998754e-01 4.95699316e-01
-1.90066051e+00 -5.08617878e-01 -2.93774992e-01 1.77403724e+00
1.55192828e+00 2.44220018e-01 -2.08508343e-01 -1.35955423e-01
6.68964446e-01 7.58923441e-02 -4.59760010e-01 -2.76920259e-01
-1.18637934e-01 6.86045468e-01 3.98500830e-01 4.60074514e-01
-1.49348068e+00 1.10500550e+00 5.81018496e+00 6.21961772e-01
-7.83734977e-01 7.86607116e-02 8.73406827e-01 1.88532948e-01
-1.77580133e-01 -2.43348435e-01 -1.02313411e+00 4.26504493e-01
1.07817352e+00 1.53991938e-01 -2.16501027e-01 3.86892110e-01
-1.20213129e-01 3.03381532e-01 -1.27303255e+00 6.79472566e-01
-8.65507796e-02 -1.33034778e+00 -1.20940864e-01 -4.12025452e-02
6.96166277e-01 1.00826308e-01 -1.15191318e-01 3.55198145e-01
6.00628376e-01 -1.34733450e+00 -1.31260380e-01 1.43733516e-01
9.02036488e-01 -7.61036396e-01 1.19495654e+00 -8.47811326e-02
-1.16020095e+00 3.31433982e-01 -1.21918067e-01 2.36792341e-01
2.99280465e-01 1.12690270e+00 -8.94865215e-01 1.16593778e+00
4.55186695e-01 9.35988188e-01 -4.63894039e-01 8.10534120e-01
-8.01127255e-01 4.63457227e-01 -2.17020154e-01 1.39008299e-01
-4.68298234e-02 1.32870590e-02 2.46495172e-01 1.42529869e+00
-2.56947219e-01 2.66768932e-01 1.66440293e-01 7.57093489e-01
-7.09108949e-01 2.02518642e-01 -6.37193203e-01 3.64410058e-02
3.95562738e-01 1.38237667e+00 -5.97254574e-01 -2.87254184e-01
-6.73209548e-01 6.81656718e-01 7.83826709e-01 2.35576376e-01
-9.25001740e-01 -5.97215235e-01 8.25857341e-01 -4.00377065e-01
2.72959203e-01 2.62293428e-01 -3.99304599e-01 -1.29322839e+00
6.17278628e-02 -7.75096655e-01 7.67809033e-01 -1.16187155e-01
-1.52375972e+00 7.43934095e-01 -3.81689310e-01 -7.32120335e-01
-1.34452760e-01 -6.86586201e-01 -5.40575206e-01 9.03304517e-01
-1.94309449e+00 -1.28144908e+00 1.10255860e-01 2.72607654e-01
-4.19503078e-03 9.63479728e-02 1.16688454e+00 7.79022038e-01
-1.04009926e+00 1.03132927e+00 -3.79757464e-01 6.01058841e-01
8.21714222e-01 -1.45493543e+00 6.32489324e-01 6.67784870e-01
8.54371190e-02 7.42461801e-01 3.65484774e-01 -7.91924357e-01
-9.38271582e-01 -1.30130517e+00 1.35758555e+00 -6.08864486e-01
6.06014669e-01 -4.86241341e-01 -1.17812300e+00 9.69062507e-01
1.22607887e-01 3.86795968e-01 1.23324931e+00 5.37012637e-01
-5.58615685e-01 2.17333972e-01 -1.04248440e+00 3.57868373e-01
1.04999447e+00 -5.32257795e-01 -7.10188925e-01 6.54333830e-01
9.78315175e-01 -6.46008492e-01 -1.21804094e+00 8.46186638e-01
1.03047200e-01 -4.44762409e-01 1.09041691e+00 -1.21065819e+00
7.22547233e-01 -3.16834785e-02 2.22161397e-01 -1.27541137e+00
-1.70644477e-01 -5.73932648e-01 -4.25144374e-01 1.22917151e+00
1.00220072e+00 -6.35460198e-01 7.12021887e-01 6.29978776e-01
-5.94863445e-02 -1.32246304e+00 -9.19172883e-01 -3.34423959e-01
2.65083969e-01 1.56668890e-02 6.41049623e-01 1.36989248e+00
2.17681006e-01 1.03350854e+00 -2.12238818e-01 1.99710965e-01
3.34767073e-01 1.48368210e-01 3.68267208e-01 -1.19273508e+00
-3.73094916e-01 -1.30231932e-01 -3.82515728e-01 -6.26032293e-01
4.15237159e-01 -1.01426721e+00 -1.96276784e-01 -1.70764863e+00
5.32798111e-01 -3.81532013e-01 -7.11153030e-01 8.89293909e-01
-9.60109293e-01 1.03131674e-01 -2.33472690e-01 -3.45024079e-01
-6.72388494e-01 4.00902152e-01 1.11124647e+00 -2.71432132e-01
-2.19163552e-01 3.90585465e-03 -1.11440194e+00 6.18406713e-01
6.69803917e-01 -7.45765567e-01 -1.84239492e-01 -4.03148383e-01
4.13171798e-01 2.73138043e-02 3.45341712e-02 -2.80698270e-01
1.62110060e-01 3.26975621e-02 2.38219887e-01 -4.36217755e-01
1.12952717e-01 -6.71608090e-01 -1.52928129e-01 3.15572441e-01
-4.27194804e-01 -3.69162224e-02 2.37579018e-01 6.89173341e-01
-2.19075486e-01 -2.68350970e-02 4.32558358e-01 1.06999360e-01
-1.01232208e-01 5.31124592e-01 2.25593016e-01 4.45278317e-01
8.88713896e-01 6.22342169e-01 -5.95208585e-01 7.85739161e-03
-9.29407060e-01 4.88215894e-01 -1.60737559e-01 2.71649510e-01
3.74769241e-01 -1.18935239e+00 -9.73334372e-01 -7.04329973e-03
1.73341364e-01 4.14068758e-01 8.38663876e-02 9.47382152e-01
-1.80226311e-01 4.01110440e-01 1.95732340e-01 -2.36211866e-01
-1.53558242e+00 5.41494191e-01 2.89511532e-01 -1.17183506e+00
-6.06505334e-01 1.12412894e+00 2.45520815e-01 -6.59364879e-01
1.47158369e-01 -3.51320803e-01 -5.36529183e-01 1.03981934e-01
4.75534499e-01 -1.95283890e-01 4.57532585e-01 -4.57154155e-01
-8.38818491e-01 2.84719616e-01 -5.69984317e-01 5.00872552e-01
1.74321699e+00 3.68648797e-01 -3.39111269e-01 5.37828878e-02
1.51289237e+00 -1.79191321e-01 -8.48913252e-01 -4.78757054e-01
3.76572251e-01 -2.56280117e-02 -2.37087935e-01 -1.08880067e+00
-1.23126400e+00 7.89060414e-01 1.22846164e-01 -1.27972618e-01
9.43325162e-01 2.51744092e-01 7.91729450e-01 3.53709489e-01
-8.40736404e-02 -8.13963473e-01 -2.18025625e-01 5.74415684e-01
5.67292571e-01 -1.43500435e+00 3.36494803e-01 -1.00071943e+00
-5.23580074e-01 8.18442285e-01 6.15011513e-01 1.63444474e-01
5.83664238e-01 4.88149345e-01 1.64940767e-02 -5.45592964e-01
-8.35927188e-01 -3.02035332e-01 4.93805498e-01 4.21706915e-01
8.71735334e-01 3.84032447e-03 -6.73913181e-01 1.29703510e+00
2.16378644e-01 -8.80094990e-02 1.70625657e-01 8.23300660e-01
2.88246810e-01 -1.48658800e+00 3.06435972e-01 5.20689249e-01
-1.39360476e+00 -5.08802116e-01 -4.95963037e-01 5.93546093e-01
-9.89750102e-02 9.73371923e-01 -3.26264441e-01 -3.85907173e-01
5.47748446e-01 9.19176489e-02 4.01222080e-01 -8.69861901e-01
-8.50419760e-01 -3.35199907e-02 4.42859411e-01 -5.66287637e-01
-4.55800980e-01 -2.64500260e-01 -1.44363308e+00 -4.64446004e-03
-5.64630091e-01 2.65235215e-01 2.87709951e-01 1.02975047e+00
6.31563962e-01 1.22398436e+00 3.42968583e-01 -1.42617851e-01
-3.91715169e-01 -9.75549340e-01 -3.01317602e-01 5.41944087e-01
3.97005022e-01 -5.29935956e-01 -1.75698310e-01 -2.42863074e-01] | [8.803067207336426, 8.794048309326172] |
08770932-74bf-4631-a9c7-0e06145bc7ff | reconstruction-of-the-external-stimuli-from | 1711.06550 | null | http://arxiv.org/abs/1711.06550v1 | http://arxiv.org/pdf/1711.06550v1.pdf | Reconstruction of the External Stimuli from Brain Signals | Despite the rapid advances in Brain-computer Interfacing (BCI) and continuous
effort to improve the accuracy of brain decoding systems, the urge for the
systems to reconstruct the experiences of the users has been widely
acknowledged. This urge has been investigated by some researchers during the
past years in terms of reconstruction of the naturalistic images, abstract
images, video and audio. In this study, we try to tackle this issue by
regressing the stimuli spectrogram using the spectrogram analysis of the brain
signals. The results of our regression-based method suggest the feasibility of
such reconstructions using the neuroimaging techniques that are appropriate for
out-of-lab scenarios. | ['Pouya Ghaemmaghami'] | 2017-11-14 | null | null | null | null | ['brain-decoding', 'brain-decoding'] | ['medical', 'miscellaneous'] | [ 1.47937521e-01 -4.39139120e-02 6.96066022e-01 -3.22353512e-01
-1.88640639e-01 -1.18457407e-01 4.30223167e-01 -3.31056476e-01
-6.36130393e-01 9.28059101e-01 2.50237137e-01 -3.02652419e-01
-1.51049823e-01 -1.54370815e-01 -4.25217599e-01 -4.96002495e-01
-2.78916359e-01 -2.73598820e-01 -3.52027901e-02 -1.51299551e-01
5.37621379e-01 6.18895352e-01 -1.91635191e+00 3.03586602e-01
6.60493433e-01 9.51041162e-01 4.32589442e-01 5.42442322e-01
1.02100678e-01 5.68444371e-01 -8.11324835e-01 -2.59819776e-01
-1.06678270e-01 -4.25780028e-01 -4.68306214e-01 -2.22425222e-01
-1.02772266e-01 -5.34163594e-01 -5.03965676e-01 1.02812874e+00
7.97835827e-01 -1.75861090e-01 5.18593788e-01 -1.08666873e+00
-3.34789246e-01 2.62714148e-01 -2.61800766e-01 5.83725929e-01
9.07597899e-01 1.18684433e-02 6.20820150e-02 -8.25540841e-01
4.77459639e-01 1.01048899e+00 3.81050855e-01 3.33795041e-01
-1.17453420e+00 -1.12735569e+00 -3.76905739e-01 6.47547305e-01
-1.69372368e+00 -6.78038120e-01 6.97419703e-01 -4.71167803e-01
9.12170291e-01 1.43866628e-01 1.02767229e+00 1.33846390e+00
7.41184592e-01 2.14551479e-01 1.53640115e+00 -4.81557578e-01
1.80587769e-01 6.77467227e-01 1.08272403e-01 2.91297361e-02
1.80900261e-01 9.09514725e-02 -6.46967471e-01 7.15903714e-02
9.60605145e-01 -3.20646852e-01 -6.62910640e-01 2.24262267e-01
-8.96750152e-01 4.90548581e-01 -1.08996987e-01 8.16734314e-01
-5.58885574e-01 -1.48031369e-01 3.51262242e-01 2.79574484e-01
4.23031390e-01 3.59721303e-01 3.07192630e-03 -5.68966627e-01
-1.04800236e+00 2.31979545e-02 7.90839791e-01 5.78576326e-01
-5.54787032e-02 2.31803402e-01 1.37193203e-01 7.09639132e-01
4.00268078e-01 1.72539800e-01 8.05157185e-01 -5.21078050e-01
1.24343976e-01 3.12815644e-02 -2.88425153e-03 -1.07587218e+00
-5.96781433e-01 -9.79789943e-02 -4.78277147e-01 2.40354747e-01
2.84350008e-01 -1.23050980e-01 -5.95095336e-01 1.48110127e+00
-6.11051656e-02 2.52996266e-01 -2.37414151e-01 9.64574456e-01
5.14540851e-01 5.36213756e-01 2.15066835e-01 -5.20788610e-01
1.13216424e+00 -7.26552606e-02 -1.29634368e+00 6.84002787e-02
9.70958397e-02 -7.56408393e-01 1.05219793e+00 8.21325958e-01
-9.85423326e-01 -5.84234416e-01 -1.15426826e+00 4.22695398e-01
-1.14053689e-01 -7.16122165e-02 4.45735306e-01 9.48153615e-01
-1.18745005e+00 4.29073274e-01 -7.90812254e-01 -5.01479805e-01
2.39058048e-01 3.27280164e-01 -5.93039751e-01 3.52571368e-01
-9.55878377e-01 1.29017794e+00 1.50736600e-01 4.37986612e-01
-6.85007513e-01 -4.22074974e-01 -3.75904769e-01 -1.39288232e-01
-2.72576511e-01 1.17637910e-01 7.39402533e-01 -9.63791490e-01
-1.67884552e+00 6.72146916e-01 -3.69316749e-02 -1.72798067e-01
3.55352759e-01 -8.23193491e-02 -1.02213943e+00 3.11502725e-01
-2.80962169e-01 3.78131568e-01 7.53164113e-01 -8.57000113e-01
-1.24086261e-01 -4.96935874e-01 -3.17065805e-01 -1.21253133e-01
-2.12036580e-01 3.01275373e-01 1.63381398e-01 -5.15689015e-01
1.49934217e-01 -8.32799435e-01 2.52357185e-01 -2.12075442e-01
-1.24423824e-01 2.14712888e-01 6.12681806e-01 -1.02773881e+00
1.05777061e+00 -2.37442207e+00 -1.83472171e-01 1.54598698e-01
-6.99604005e-02 2.15720460e-01 9.22373310e-02 4.47214305e-01
-4.18502659e-01 -1.13256633e-01 1.73623115e-01 4.58265021e-02
-3.91392797e-01 -6.21363409e-02 -2.35024691e-01 8.62174213e-01
8.12421218e-02 2.68710852e-01 -3.88898373e-01 -4.06494707e-01
2.67220467e-01 7.59816706e-01 -4.54624861e-01 4.88640815e-01
6.91595614e-01 1.02782631e+00 -7.96575844e-02 2.62845755e-01
6.55419350e-01 4.21340138e-01 3.09636630e-02 -3.75148743e-01
-4.16484714e-01 3.09854537e-01 -9.25894618e-01 1.62437296e+00
-2.38927096e-01 1.13542390e+00 2.20062718e-01 -1.23518169e+00
6.49134696e-01 8.94466579e-01 6.20873570e-01 -1.14347458e+00
4.09339100e-01 1.92093909e-01 5.97000241e-01 -1.07265186e+00
8.51250887e-02 -3.71299297e-01 5.92806399e-01 3.80032629e-01
2.63638049e-01 -2.06761047e-01 -2.37788513e-01 6.59652893e-03
6.22002602e-01 2.79734522e-01 6.30085245e-02 -5.15766919e-01
4.57376152e-01 -3.06869477e-01 -1.12078346e-01 2.88639575e-01
-3.07294965e-01 4.61497843e-01 2.15327218e-01 -9.13664997e-02
-7.31463671e-01 -8.21521163e-01 -6.44919395e-01 3.67560297e-01
-8.04226324e-02 -1.77038237e-01 -1.06059611e+00 2.54105121e-01
-3.45145613e-01 1.07677341e+00 -2.69679129e-01 -2.87335545e-01
-8.21494311e-02 -6.14010274e-01 4.98584390e-01 4.19533178e-02
2.26848349e-01 -1.23881876e+00 -1.12555563e+00 5.64280987e-01
-3.10544431e-01 -1.12425554e+00 3.14466536e-01 2.39411592e-01
-8.76341343e-01 -7.44645238e-01 -5.99508703e-01 -3.74164492e-01
3.65330905e-01 1.46259740e-01 4.63726044e-01 -4.89009842e-02
-6.66277349e-01 6.53134406e-01 -2.34159261e-01 -6.78509593e-01
-2.25043952e-01 -5.87151885e-01 3.12235504e-01 3.08232084e-02
4.07915860e-01 -1.14350975e+00 -6.30378485e-01 2.59610862e-01
-8.85976851e-01 5.10333739e-02 3.90588343e-01 2.66419649e-01
2.05729604e-01 6.83113858e-02 8.35191369e-01 -2.64245123e-01
1.12881029e+00 -6.49569154e-01 -2.50191897e-01 -8.40070322e-02
-4.74898398e-01 -2.92146415e-01 3.59279186e-01 -7.27763236e-01
-1.02445674e+00 -3.81399363e-01 -2.78970212e-01 -3.56330648e-02
-4.64165986e-01 4.88568038e-01 5.56875691e-02 -2.95597553e-01
6.73349917e-01 2.10675567e-01 2.00622037e-01 -3.02341700e-01
-3.55008096e-01 1.25536549e+00 6.31567180e-01 -2.61310607e-01
1.30879030e-01 1.91298470e-01 -3.84462833e-01 -1.27862382e+00
-1.32504418e-01 1.15753531e-01 -4.67121631e-01 -9.75136042e-01
7.99463034e-01 -8.84087622e-01 -7.85835624e-01 4.46451724e-01
-1.18683720e+00 -5.41875213e-02 3.57546896e-01 1.14704752e+00
-4.82843518e-01 3.10407609e-01 -3.80313069e-01 -1.32912910e+00
-2.23284647e-01 -1.49999809e+00 4.44390178e-01 2.00571239e-01
-6.85143650e-01 -4.77773011e-01 -5.27969711e-02 2.37229377e-01
7.03468442e-01 1.54865518e-01 8.04678082e-01 -3.47332567e-01
-8.19849446e-02 -3.35656762e-01 -8.06767587e-03 2.37577289e-01
1.18833616e-01 -1.75361603e-01 -1.40316677e+00 -1.10144295e-01
7.28594780e-01 -2.05632493e-01 -1.07689455e-01 6.09703004e-01
1.26123857e+00 4.07314673e-03 -4.31419052e-02 2.88906991e-01
1.31265664e+00 6.58062816e-01 1.17224360e+00 6.58521280e-02
4.76916246e-02 8.58402848e-01 1.71000287e-01 4.43719804e-01
4.44324426e-02 6.50694847e-01 2.52394259e-01 2.03007475e-01
-4.69371956e-03 -3.44637223e-02 3.78334075e-01 8.50067914e-01
-3.50117683e-01 1.70661673e-01 -6.32746160e-01 3.15297842e-01
-1.33346546e+00 -8.22890997e-01 -3.35884929e-01 2.33055520e+00
5.30408263e-01 4.53991033e-02 -1.92705598e-02 5.17694592e-01
5.06217241e-01 -3.85953665e-01 -2.36262262e-01 -7.87306428e-01
1.92906201e-01 2.87991881e-01 1.84589297e-01 1.71326831e-01
-2.92028964e-01 2.53129095e-01 7.76395321e+00 5.64549446e-01
-1.56814015e+00 1.87771127e-01 4.79256779e-01 -1.71421632e-01
-1.79057289e-02 -2.33485103e-01 -7.01354370e-02 8.82148743e-01
1.52585864e+00 -1.17375758e-02 1.01459396e+00 3.88305873e-01
7.53913164e-01 -8.68142068e-01 -9.50691223e-01 1.19373715e+00
2.21404254e-01 -6.45162046e-01 -4.39547628e-01 1.92467541e-01
-7.89112896e-02 -1.31363690e-01 2.03598700e-02 -1.59458313e-02
-7.32394218e-01 -1.32363331e+00 1.08908153e+00 7.92489350e-01
9.19910789e-01 -4.82822537e-01 5.71226060e-01 6.19421184e-01
-4.72059399e-01 1.32395308e-02 -2.11699679e-01 -4.81103420e-01
2.31856063e-01 3.34660679e-01 -6.67694092e-01 3.89858149e-02
8.01428437e-01 9.67036188e-03 -3.15186709e-01 1.07296586e+00
1.67364642e-01 9.75185275e-01 -3.49001527e-01 -1.15007386e-01
-1.77820265e-01 -6.33145988e-01 3.23110819e-01 1.05846095e+00
5.68704963e-01 4.94786710e-01 -9.02290285e-01 9.81569350e-01
8.58076036e-01 2.73192883e-01 -7.97312737e-01 -2.69220442e-01
1.32715568e-01 1.02325559e+00 -7.05472291e-01 1.48985624e-01
-6.11306727e-01 8.74781311e-01 -4.32601571e-02 3.67608100e-01
-8.37396562e-01 -3.43002707e-01 4.05266821e-01 3.72388482e-01
-2.18773603e-01 -4.11130548e-01 -5.60012102e-01 -9.77360129e-01
5.47386259e-02 -8.54263544e-01 -2.82073140e-01 -1.29419506e+00
-6.78573251e-01 8.21946323e-01 4.31336433e-01 -1.12092650e+00
-1.97091281e-01 -6.79945171e-01 -5.26834309e-01 1.04512227e+00
-9.20030773e-01 -2.91268617e-01 -2.38162175e-01 9.07477558e-01
3.45510304e-01 -1.90179050e-01 1.03492963e+00 6.17481709e-01
-4.39387798e-01 1.30332500e-01 -3.04581255e-01 -5.47822833e-01
4.99439299e-01 -4.20555234e-01 -3.42091769e-01 5.57722867e-01
-3.10869720e-02 7.52082705e-01 1.17797458e+00 -3.93880218e-01
-1.37585735e+00 -5.40389642e-02 6.46172285e-01 1.16075374e-01
5.54571688e-01 -3.78986150e-01 -8.20878446e-01 3.75780344e-01
5.10638952e-01 -3.99311304e-01 8.15210164e-01 -4.13230389e-01
1.20649211e-01 -1.79260120e-01 -1.36756134e+00 6.85762823e-01
6.50316656e-01 -6.89961553e-01 -5.86997271e-01 3.22299674e-02
-1.28424466e-01 8.00986886e-02 -7.79146254e-01 1.45837903e-01
9.68234539e-01 -1.17793369e+00 7.73341835e-01 -1.87228695e-01
2.13674754e-01 1.52377963e-01 -1.53653845e-01 -1.50508261e+00
5.44050597e-02 -3.88978064e-01 4.72842604e-01 9.20685709e-01
2.00498044e-01 -9.18980956e-01 2.02958882e-01 1.08188069e+00
1.23438664e-01 -2.94044405e-01 -1.22564232e+00 -4.38017935e-01
-2.19942495e-01 -8.00079405e-01 2.22553149e-01 4.09354329e-01
8.22839558e-01 1.08619086e-01 -4.22441691e-01 -5.05207293e-02
3.70165855e-01 -6.17706954e-01 4.43675488e-01 -1.32197762e+00
-2.84831319e-02 -3.38321358e-01 -7.33051658e-01 -4.05166447e-01
5.32204844e-02 -4.30457503e-01 -4.29507461e-04 -1.24069202e+00
7.78983235e-02 -9.25447568e-02 -1.89835176e-01 -4.96982224e-02
2.10911781e-01 1.13884561e-01 5.21268249e-02 2.02568155e-02
3.04060221e-01 3.42005342e-01 1.04470849e+00 2.24209324e-01
-5.61597906e-02 -1.61654785e-01 -6.38302803e-01 4.60804433e-01
6.65653229e-01 -5.26613176e-01 -6.12697840e-01 -1.69555947e-01
-5.84637262e-02 4.30463850e-01 4.13753510e-01 -1.40905201e+00
1.73143297e-01 2.41375282e-01 4.30948853e-01 -5.01231194e-01
7.31354117e-01 -1.27532816e+00 9.09148335e-01 4.05589700e-01
-1.51533902e-01 2.41995230e-02 3.89328748e-01 2.67367244e-01
-2.61339813e-01 -3.79887432e-01 7.70500779e-01 2.08680760e-02
-2.33028173e-01 -3.90244395e-01 -9.94207680e-01 -4.03795838e-01
9.19552088e-01 -7.29303777e-01 8.14896822e-02 -8.33948016e-01
-6.35791421e-01 -5.34830809e-01 -3.39655280e-02 1.65042594e-01
5.99770069e-01 -1.04080868e+00 -4.46236342e-01 5.27612031e-01
-2.11558655e-01 -7.40960956e-01 3.91821831e-01 1.46171474e+00
-5.15166461e-01 6.17371082e-01 -9.48267400e-01 -3.62916112e-01
-1.07196629e+00 2.52806395e-01 2.93316633e-01 4.45119858e-01
-8.36845696e-01 3.71854007e-01 1.06811924e-02 5.47720373e-01
3.15685689e-01 -1.02227509e-01 -5.42017102e-01 8.48798007e-02
5.89910984e-01 3.84048671e-01 3.41868311e-01 -7.02548742e-01
-3.97852749e-01 6.07361645e-02 2.07681715e-01 -6.65443718e-01
1.49293792e+00 -1.52532697e-01 -1.20446496e-01 7.24223018e-01
1.12227404e+00 -2.78588128e-03 -9.30282176e-01 3.82990509e-01
-1.86846495e-01 -5.87387800e-01 3.43532205e-01 -8.10833573e-01
-9.77811158e-01 9.52214241e-01 9.61738467e-01 2.45075285e-01
1.23638153e+00 -4.79873925e-01 3.17597300e-01 4.57956418e-02
8.58240008e-01 -1.19693089e+00 -2.71584600e-01 -1.35280127e-02
1.15792358e+00 -6.71978235e-01 -5.22909388e-02 -1.55835435e-01
-6.26640320e-01 1.14133251e+00 1.79333284e-01 -2.46239454e-01
1.10996795e+00 5.20894110e-01 6.22976199e-02 -1.56533122e-01
-4.69712228e-01 2.33724371e-01 -1.94297594e-04 6.87810004e-01
7.89558291e-01 1.78171530e-01 -9.32812333e-01 9.65897024e-01
-2.87990987e-01 4.69926268e-01 7.68607080e-01 7.64787138e-01
-2.39888698e-01 -5.32067358e-01 -7.90546298e-01 4.74996865e-01
-7.36066997e-01 -5.19088469e-03 -1.33220449e-01 7.67323673e-01
-2.07640469e-01 1.30784607e+00 -7.57070631e-02 -4.59201783e-01
4.38139498e-01 4.38886762e-01 7.24169731e-01 -1.14917450e-01
-4.21684414e-01 1.61928460e-01 7.45405331e-02 -6.30882144e-01
-4.23993945e-01 -7.21497297e-01 -9.96763170e-01 -1.08066015e-03
-4.41374034e-01 -2.70808749e-02 1.41271377e+00 1.06281424e+00
2.92034954e-01 5.12734830e-01 3.43423426e-01 -1.08388090e+00
-1.98650792e-01 -1.37052619e+00 -1.02172029e+00 3.01945478e-01
-7.43151903e-02 -8.24376106e-01 -3.71990114e-01 -9.46798176e-02] | [13.186948776245117, 3.3296055793762207] |
40915926-ad40-40c9-9073-01d03067b4a6 | neural-symbolic-regression-using-control | 2306.04718 | null | https://arxiv.org/abs/2306.04718v1 | https://arxiv.org/pdf/2306.04718v1.pdf | Neural Symbolic Regression using Control Variables | Symbolic regression (SR) is a powerful technique for discovering the analytical mathematical expression from data, finding various applications in natural sciences due to its good interpretability of results. However, existing methods face scalability issues when dealing with complex equations involving multiple variables. To address this challenge, we propose SRCV, a novel neural symbolic regression method that leverages control variables to enhance both accuracy and scalability. The core idea is to decompose multi-variable symbolic regression into a set of single-variable SR problems, which are then combined in a bottom-up manner. The proposed method involves a four-step process. First, we learn a data generator from observed data using deep neural networks (DNNs). Second, the data generator is used to generate samples for a certain variable by controlling the input variables. Thirdly, single-variable symbolic regression is applied to estimate the corresponding mathematical expression. Lastly, we repeat steps 2 and 3 by gradually adding variables one by one until completion. We evaluate the performance of our method on multiple benchmark datasets. Experimental results demonstrate that the proposed SRCV significantly outperforms state-of-the-art baselines in discovering mathematical expressions with multiple variables. Moreover, it can substantially reduce the search space for symbolic regression. The source code will be made publicly available upon publication. | ['Huajie Shao', 'Minghan Chen', 'Hairong Qi', 'Enze Xu', 'Hongjue Zhao', 'Xieting Chu'] | 2023-06-07 | null | null | null | null | ['symbolic-regression'] | ['knowledge-base'] | [ 4.43980843e-01 -2.79724717e-01 -4.38336790e-01 -4.67108488e-01
-8.40422928e-01 -3.70555103e-01 2.73102760e-01 -1.26336366e-01
5.79728149e-02 9.36876833e-01 -4.66704726e-01 -4.66768265e-01
-6.40260875e-02 -8.65004957e-01 -8.87247264e-01 -7.31354773e-01
2.55901247e-01 3.15021425e-01 -1.31391212e-01 -2.54679888e-01
1.47107810e-01 6.46577895e-01 -1.35356569e+00 -5.74221415e-03
9.87247288e-01 1.27952385e+00 -1.43839180e-01 2.15113223e-01
-2.66302913e-01 1.23180032e+00 -5.36927938e-01 -1.85379922e-01
4.08479273e-02 -4.13685381e-01 -6.40218794e-01 -3.88169378e-01
-4.87840697e-02 -1.76733330e-01 7.12604448e-02 1.12674236e+00
1.81197822e-01 3.27394098e-01 5.25337875e-01 -1.48382521e+00
-7.02074826e-01 7.41635501e-01 -8.08606923e-01 -1.38437226e-01
3.18460196e-01 9.00455415e-02 1.18626428e+00 -8.29265893e-01
3.18410605e-01 1.65939939e+00 4.15381461e-01 3.28113347e-01
-1.54109848e+00 -1.26724732e+00 1.23337507e-01 3.24592702e-02
-1.63980174e+00 -4.03591692e-01 1.01405883e+00 -3.27658743e-01
9.18173134e-01 2.40957707e-01 2.47233689e-01 9.84699667e-01
-7.84636065e-02 7.35567570e-01 1.00751185e+00 -1.52826384e-01
1.69384211e-01 4.81602177e-02 2.86490232e-01 8.43890965e-01
7.87398499e-03 -3.44920903e-02 -4.15235460e-01 -2.72523403e-01
8.03129554e-01 -1.09772719e-02 -2.02994309e-02 -2.33940836e-02
-8.49279344e-01 1.00059152e+00 3.31353545e-01 5.04814684e-02
-4.06644762e-01 4.38797414e-01 4.59917456e-01 3.85893464e-01
3.12850803e-01 6.69002891e-01 -7.05061376e-01 -8.29880461e-02
-6.83472514e-01 5.62976539e-01 5.66659570e-01 1.10424972e+00
8.07971001e-01 2.89719254e-01 -2.00047806e-01 9.75203514e-01
1.60779402e-01 3.90737474e-01 3.65366280e-01 -7.93043733e-01
5.32076120e-01 9.78507817e-01 -7.54813552e-02 -1.24715996e+00
-4.04264748e-01 -3.14218313e-01 -9.80125427e-01 9.89712030e-02
2.05429494e-01 -2.00830862e-01 -6.70124769e-01 1.71314812e+00
4.26721931e-01 4.13148850e-01 1.53404295e-01 9.14944649e-01
7.78998435e-01 7.82769680e-01 1.84041467e-02 -2.22963706e-01
1.15765941e+00 -8.37586701e-01 -6.57647669e-01 2.43992079e-02
6.17823005e-01 -1.86128676e-01 9.74505424e-01 6.49585783e-01
-9.05403554e-01 -5.88612258e-01 -9.85830486e-01 -3.32208008e-01
-2.26175249e-01 4.10562783e-01 9.57754672e-01 1.58608720e-01
-5.80514729e-01 6.33094847e-01 -9.55448091e-01 5.50412655e-01
6.58631265e-01 6.42273307e-01 -1.56923309e-01 1.55028149e-01
-1.34741831e+00 4.89406914e-01 2.89813012e-01 3.05350214e-01
-6.82121217e-01 -8.37002635e-01 -1.24915946e+00 2.36647934e-01
7.64227927e-01 -3.35525990e-01 1.26301527e+00 -9.66321945e-01
-1.65010118e+00 2.98120409e-01 -7.69109070e-01 -3.24738979e-01
2.93852091e-01 -1.46522135e-01 -3.42075616e-01 -2.17791155e-01
2.44503275e-01 1.25216737e-01 7.22447574e-01 -1.04549348e+00
-4.53941435e-01 -4.26501811e-01 -2.83682290e-02 -4.92951393e-01
-1.86954494e-04 2.53953427e-01 -4.59157974e-01 -5.67795038e-01
1.17651567e-01 -7.53333688e-01 -3.24586332e-01 -1.55962318e-01
-5.82909584e-01 -6.02579713e-01 6.47913516e-01 -5.79648495e-01
1.52606964e+00 -2.07798290e+00 5.99775374e-01 5.02924383e-01
4.25016075e-01 -2.77517596e-04 5.45480326e-02 -9.79407057e-02
-3.83030295e-01 2.82240629e-01 -4.04980391e-01 -3.36909145e-01
-9.88578945e-02 -6.00336008e-02 -5.91792047e-01 2.44332165e-01
8.68524790e-01 1.04943657e+00 -8.95634294e-01 -5.05723059e-01
-1.05325647e-01 2.99960524e-01 -2.40813643e-01 1.66122347e-01
-6.05909169e-01 4.29067940e-01 -8.76548707e-01 8.89133751e-01
6.84550703e-01 -5.40194511e-01 1.66136816e-01 -4.73649660e-03
-4.61288495e-03 1.90585688e-01 -1.31659830e+00 1.28416216e+00
-5.85890830e-01 4.93422091e-01 -2.63268799e-01 -1.25241125e+00
1.35366559e+00 7.60009885e-02 3.44757438e-01 -6.92987144e-01
3.80853146e-01 3.31375360e-01 -1.07131958e-01 -5.47944665e-01
2.65413284e-01 -2.14286953e-01 -3.06949466e-01 2.67458469e-01
-3.11359763e-01 -7.87268505e-02 3.46055120e-01 -6.95246384e-02
9.40132380e-01 4.05683398e-01 5.46905041e-01 3.36527437e-01
1.07531762e+00 -4.05639112e-02 9.35396552e-01 4.01086777e-01
2.89918542e-01 8.30681175e-02 1.18371272e+00 -3.53572905e-01
-6.09919310e-01 -6.69263303e-01 2.98487488e-02 9.28603888e-01
3.20014916e-02 -3.82394046e-01 -2.94793844e-01 -3.57177019e-01
3.36232930e-01 9.36127484e-01 -7.57029653e-01 -8.99783298e-02
-6.97155416e-01 -5.06022096e-01 6.45631433e-01 7.75479734e-01
4.97753799e-01 -1.12797177e+00 -3.16898316e-01 2.08211899e-01
-6.57293424e-02 -1.08872855e+00 1.15562506e-01 1.81684703e-01
-7.39507198e-01 -9.35352206e-01 -3.61081332e-01 -6.10823691e-01
6.89015448e-01 1.12134337e-01 7.81512737e-01 3.55101883e-01
-1.71373129e-01 -4.96143907e-01 -1.33493394e-01 -3.32806975e-01
-4.68564212e-01 1.67601556e-01 -8.36334676e-02 7.45189190e-02
5.59436262e-01 -6.00388825e-01 5.62640317e-02 1.99372679e-01
-7.32352912e-01 4.32169512e-02 6.36598587e-01 8.88069332e-01
9.90101397e-01 2.12154433e-01 6.09798908e-01 -9.14289653e-01
7.84737349e-01 -7.27847695e-01 -1.14582801e+00 3.21992099e-01
-4.72624660e-01 5.07104814e-01 1.16171873e+00 -6.11346185e-01
-6.98561072e-01 2.98449755e-01 2.80771405e-01 -8.16586912e-01
1.51570931e-01 8.62748265e-01 -2.69251466e-01 1.26252100e-01
4.08539116e-01 3.46410394e-01 4.91356440e-02 -2.82704532e-01
1.19341835e-01 5.54901302e-01 5.42085767e-01 -8.62914205e-01
9.44977641e-01 1.10991681e-02 6.66429341e-01 -4.47543889e-01
-6.35682285e-01 -9.92764607e-02 -4.31879848e-01 1.65312201e-01
6.58565164e-01 -7.93597162e-01 -1.17366934e+00 4.97171193e-01
-1.26619923e+00 -3.07283789e-01 4.51852560e-01 1.59554929e-01
-3.20149273e-01 1.10981904e-01 -2.61008561e-01 -9.57580566e-01
-3.83354068e-01 -1.41379154e+00 1.18659866e+00 2.34326318e-01
-4.52103734e-01 -7.22250879e-01 -2.81554580e-01 2.56675571e-01
6.87804893e-02 6.53980374e-01 1.11527956e+00 -6.55113220e-01
-6.26201689e-01 -1.60874173e-01 -4.04795319e-01 1.99357182e-01
2.77441770e-01 2.08507329e-01 -5.97022533e-01 1.79847479e-01
-2.25430772e-01 -5.54884076e-01 6.33330464e-01 1.34887174e-01
1.75790226e+00 -2.81548679e-01 -3.91667038e-01 7.77769089e-01
1.18886971e+00 2.33657271e-01 3.01304668e-01 2.72100061e-01
8.23575377e-01 3.90200675e-01 7.48449683e-01 4.38746780e-01
3.57785016e-01 7.33239770e-01 2.66674012e-01 -9.37311202e-02
4.83512193e-01 -2.63632357e-01 2.81785995e-01 3.74235332e-01
-4.38226350e-02 8.40475485e-02 -9.46174026e-01 1.97821617e-01
-1.93702626e+00 -7.77607441e-01 -3.08955461e-01 1.87322307e+00
1.12792206e+00 1.00942127e-01 7.85371065e-02 2.92120248e-01
4.47042048e-01 -9.51912850e-02 -9.72925067e-01 -3.82197708e-01
-5.44944555e-02 5.78233063e-01 3.71570766e-01 4.08739865e-01
-9.02367234e-01 1.13691628e+00 4.83933449e+00 7.82890320e-01
-1.51425660e+00 -3.29582572e-01 6.25737548e-01 -2.28707314e-01
-1.98814765e-01 -1.74905598e-01 -8.76490712e-01 4.43195462e-01
7.57012129e-01 -2.66326934e-01 8.14057708e-01 1.01522768e+00
4.72733408e-01 2.33978480e-01 -1.21063805e+00 9.62461412e-01
-1.29037470e-01 -1.21013319e+00 2.68861428e-02 -3.04553032e-01
4.55599427e-01 -4.38843966e-01 -9.98528004e-02 6.30674362e-01
2.47135743e-01 -1.43910861e+00 6.09829187e-01 4.53833491e-01
9.39521611e-01 -9.62722003e-01 5.03607213e-01 2.24042252e-01
-1.29349387e+00 -1.94572315e-01 -1.42595023e-01 -1.89777911e-01
-1.17677383e-01 3.48867148e-01 -7.49981284e-01 5.34411371e-01
3.87446761e-01 9.28297698e-01 -4.59510654e-01 2.97078162e-01
-7.04979539e-01 5.06337285e-01 -1.66917399e-01 -4.36859459e-01
5.11335693e-02 -4.15347934e-01 2.12864026e-01 8.96393299e-01
1.66957244e-01 2.69674987e-01 -3.62512134e-02 1.59948170e+00
-1.89115703e-01 -7.76198111e-04 -3.98224354e-01 -1.84549347e-01
6.89289808e-01 1.21409154e+00 -3.76449525e-01 -3.04631054e-01
-3.99652869e-01 7.03751445e-01 5.77500939e-01 5.10991752e-01
-1.09073567e+00 -7.33552098e-01 5.37878811e-01 -2.31842279e-01
3.99131000e-01 -2.21566185e-01 -5.94045281e-01 -1.07524002e+00
3.29732627e-01 -1.29672396e+00 1.31173953e-01 -8.17386389e-01
-8.92507017e-01 3.92960995e-01 1.10219397e-01 -9.77292418e-01
-5.45699894e-01 -6.70648456e-01 -4.36150849e-01 1.15300107e+00
-1.56392729e+00 -1.09718871e+00 -4.20510918e-01 4.33557302e-01
4.28561807e-01 -2.05457166e-01 7.53571630e-01 8.12908411e-02
-1.05747259e+00 9.37017977e-01 -1.99115630e-02 3.20632756e-01
2.45642528e-01 -1.02298975e+00 3.98026466e-01 7.08407223e-01
-2.63515234e-01 1.12093127e+00 4.60884184e-01 -6.53631806e-01
-1.77026439e+00 -1.19281864e+00 7.76984811e-01 -1.55156985e-01
9.06568110e-01 -5.47652125e-01 -1.18116450e+00 6.98637962e-01
-6.39548957e-01 2.90127490e-02 5.18432498e-01 1.77571416e-01
-4.27260041e-01 -1.78155333e-01 -9.97978806e-01 5.93046844e-01
6.87973857e-01 -2.66531229e-01 -2.30334848e-01 -5.94752049e-03
1.02298987e+00 -7.13298023e-01 -9.81350660e-01 2.97978371e-01
6.18716598e-01 -3.34969729e-01 7.86849797e-01 -8.45846117e-01
1.14240479e+00 -4.69148040e-01 -5.21531962e-02 -1.00018311e+00
-6.98710978e-02 -8.60285521e-01 -3.73371333e-01 1.16456521e+00
6.49385691e-01 -8.48107815e-01 4.59401876e-01 8.88102949e-01
1.19377032e-01 -1.20548820e+00 -7.51344800e-01 -7.06348836e-01
7.89564196e-03 -5.00797510e-01 1.14929295e+00 1.00056481e+00
-2.46300906e-01 5.89841962e-01 -3.10027331e-01 2.63897538e-01
3.24117750e-01 5.32932460e-01 1.19041491e+00 -1.14359426e+00
-5.61181009e-01 -5.42079270e-01 -2.90728480e-01 -1.07238078e+00
6.82581782e-01 -9.85805154e-01 -5.80497086e-02 -9.48058665e-01
9.13290400e-03 -4.09437239e-01 -2.83127904e-01 7.81285584e-01
-4.50744778e-01 -3.21350753e-01 -5.03105745e-02 4.40264307e-03
-3.24966162e-01 6.16638780e-01 1.09599590e+00 -1.92287430e-01
-5.21263719e-01 1.03557453e-01 -1.01007009e+00 4.15002763e-01
7.86325812e-01 -5.84763229e-01 -3.13901693e-01 -1.79006055e-01
2.69954115e-01 3.25316757e-01 4.89483893e-01 -5.76403856e-01
1.96649544e-02 -5.67907453e-01 3.71458292e-01 -5.14051259e-01
2.37503469e-01 -6.59108281e-01 3.76168601e-02 2.80100942e-01
-4.70786422e-01 1.42356843e-01 4.01177198e-01 2.65434891e-01
-1.96224332e-01 -1.81389481e-01 5.63430369e-01 2.16022208e-01
-4.04060781e-01 1.92194417e-01 3.94650735e-02 -1.30521357e-01
9.31646705e-01 1.89424187e-01 -2.91162152e-02 -1.61328956e-01
-2.30213076e-01 6.43273234e-01 -8.98237154e-03 4.41263556e-01
6.55436635e-01 -1.26401651e+00 -6.10360920e-01 2.65409499e-01
-1.61682982e-02 6.33505344e-01 -3.36635798e-01 7.72720277e-01
-2.76184976e-01 4.16117996e-01 1.80676609e-01 -5.32969892e-01
-1.42685556e+00 6.42209113e-01 2.62377083e-01 -3.70363951e-01
-3.32254291e-01 7.10167289e-01 -9.23995301e-02 -4.29664046e-01
1.39223918e-01 -7.81704783e-01 -1.59282193e-01 -2.55670816e-01
5.83510995e-01 4.47957605e-01 -3.42432171e-01 -3.92692029e-01
-3.38546842e-01 5.67310810e-01 -1.24663517e-01 2.10813627e-01
1.65743101e+00 4.20032799e-01 -4.69525695e-01 6.03236318e-01
1.39612699e+00 -1.73212931e-01 -9.26313281e-01 -3.62544447e-01
-8.02290291e-02 -3.00160676e-01 1.61627755e-02 -6.35890484e-01
-1.16389203e+00 7.69411445e-01 -9.75836217e-02 -1.06625289e-01
1.34986210e+00 -1.05957754e-01 7.29071736e-01 6.70920908e-01
-1.83890834e-02 -6.92178845e-01 -1.23373650e-01 4.65770692e-01
1.04715443e+00 -1.26244819e+00 2.31900334e-01 -4.37232763e-01
-5.96341133e-01 1.40049338e+00 7.47902691e-01 -2.84067970e-02
1.66728154e-01 2.69200176e-01 -2.46701613e-01 -2.61410952e-01
-7.23753095e-01 3.04499537e-01 1.62181988e-01 1.75310433e-01
4.11786765e-01 -2.42950357e-02 -7.80853555e-02 1.16342330e+00
-3.65264773e-01 5.12414217e-01 1.74801648e-01 7.08812892e-01
4.81711663e-02 -1.13827693e+00 -3.30730945e-01 5.09589016e-01
-3.67987782e-01 -7.91995749e-02 -6.28798842e-01 8.86792719e-01
-1.35585889e-01 8.26130271e-01 -2.00447127e-01 -4.71161753e-01
3.68556440e-01 -1.20259032e-01 2.23382950e-01 -4.09803212e-01
-4.62704360e-01 -1.11928344e-01 -5.84556423e-02 -4.98826325e-01
-9.26157609e-02 -8.08140278e-01 -1.81822944e+00 -1.76304519e-01
-2.93948412e-01 -9.77958813e-02 5.50056696e-01 1.00021482e+00
3.60228568e-01 7.85306573e-01 7.91039407e-01 -4.69477654e-01
-8.46612275e-01 -7.94160724e-01 -1.62166849e-01 2.79856950e-01
5.82308888e-01 -9.58443940e-01 -3.36258769e-01 -7.98110366e-02] | [8.519248962402344, 6.860581874847412] |
46797de5-ac87-4712-b4a3-2401f1d0fc9d | ernie-music-text-to-waveform-music-generation | 2302.04456 | null | https://arxiv.org/abs/2302.04456v1 | https://arxiv.org/pdf/2302.04456v1.pdf | ERNIE-Music: Text-to-Waveform Music Generation with Diffusion Models | In recent years, there has been an increased popularity in image and speech generation using diffusion models. However, directly generating music waveforms from free-form text prompts is still under-explored. In this paper, we propose the first text-to-waveform music generation model that can receive arbitrary texts using diffusion models. We incorporate the free-form textual prompt as the condition to guide the waveform generation process of diffusion models. To solve the problem of lacking such text-music parallel data, we collect a dataset of text-music pairs from the Internet with weak supervision. Besides, we compare the effect of two prompt formats of conditioning texts (music tags and free-form texts) and prove the superior performance of our method in terms of text-music relevance. We further demonstrate that our generated music in the waveform domain outperforms previous works by a large margin in terms of diversity, quality, and text-music relevance. | ['Hua Wu', 'Hao Tian', 'Yu Sun', 'Yekun Chai', 'Shuohuan Wang', 'Chao Pang', 'Pengfei Zhu'] | 2023-02-09 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 3.36147636e-01 -2.92744428e-01 -1.32645756e-01 -9.71463174e-02
-1.02515185e+00 -7.12912381e-01 8.56185913e-01 -4.00700361e-01
1.07802665e-02 4.03112233e-01 4.59456414e-01 -2.43845403e-01
-3.40442091e-01 -6.00515425e-01 -5.65654993e-01 -6.64242387e-01
2.48058155e-01 4.34673935e-01 9.26253498e-02 -2.32487693e-01
1.09711334e-01 -1.28397509e-01 -1.43055499e+00 2.07231984e-01
1.03710353e+00 7.04269052e-01 5.02406120e-01 6.78851426e-01
-1.74546257e-01 7.22425222e-01 -9.03861523e-01 -5.03436446e-01
3.46876830e-01 -9.36026692e-01 -1.55345038e-01 -1.14394791e-01
3.67616653e-01 -2.78981835e-01 -6.31564617e-01 1.03747022e+00
1.03340757e+00 7.45253861e-02 8.05560172e-01 -1.15105236e+00
-9.19722021e-01 1.33261502e+00 -6.87665999e-01 -9.12594348e-02
3.39391679e-01 -7.18072802e-02 1.28862834e+00 -7.29085445e-01
5.61638057e-01 1.09047282e+00 3.70231509e-01 6.12522840e-01
-9.92133021e-01 -1.06745601e+00 -1.73213810e-01 3.27858388e-01
-1.28016651e+00 -5.83571792e-01 1.11865628e+00 -3.72546762e-01
4.90663022e-01 2.12689564e-01 4.82089549e-01 1.40996635e+00
-5.59760444e-02 9.59430039e-01 8.80479157e-01 -5.32062173e-01
5.77266067e-02 -1.80905774e-01 -2.85268337e-01 3.60527575e-01
-1.37701511e-01 2.07767546e-01 -1.13460624e+00 -1.56951204e-01
7.43508935e-01 -4.84426826e-01 -2.42804259e-01 2.29295820e-01
-1.24559617e+00 7.77623355e-01 -2.02659652e-01 3.20494235e-01
-2.50719607e-01 1.39884561e-01 1.80824295e-01 4.38044101e-01
3.45747411e-01 3.31229657e-01 -1.00890018e-01 -4.00350183e-01
-1.24368799e+00 3.51470202e-01 8.52667093e-01 1.26689827e+00
7.70754740e-02 5.73379874e-01 -6.14560902e-01 9.33522344e-01
2.21905589e-01 1.06600475e+00 6.65596187e-01 -8.16917777e-01
6.36116385e-01 -1.84069440e-01 3.54719535e-02 -9.96708572e-01
-3.38354311e-03 -6.19723022e-01 -9.33144212e-01 -3.98298085e-01
2.08600506e-01 -3.96478623e-01 -4.80236590e-01 1.77862048e+00
8.50234702e-02 5.29675007e-01 6.83890805e-02 9.68463302e-01
7.63872325e-01 8.63881886e-01 -4.43258286e-01 -4.13843930e-01
1.06270993e+00 -1.14197552e+00 -1.00927103e+00 2.60458440e-01
2.54927933e-01 -1.30833900e+00 1.13142800e+00 6.77057803e-01
-1.15230787e+00 -7.23864853e-01 -1.01447141e+00 1.83515757e-01
3.85651141e-01 4.04587626e-01 3.16890270e-01 6.00490093e-01
-7.85079300e-01 6.82469189e-01 -3.73777151e-01 1.26271639e-02
1.21352091e-01 -6.10478446e-02 3.15946549e-01 7.55985230e-02
-1.33717632e+00 2.90685683e-01 -6.31411523e-02 -3.13415319e-01
-9.28991795e-01 -6.38583541e-01 -3.28933179e-01 5.87878451e-02
3.33638668e-01 -7.09055543e-01 1.36591244e+00 -7.42342412e-01
-1.86009860e+00 2.40912229e-01 3.54673415e-02 -3.62307936e-01
5.73702455e-01 -1.79289654e-01 -6.78472340e-01 2.68381447e-01
1.45367667e-01 5.92682064e-01 1.43620694e+00 -9.32692468e-01
-5.27173340e-01 4.37191017e-02 -1.58649161e-01 1.41756073e-01
-6.85955286e-01 -9.20126513e-02 -6.64427280e-01 -1.38442433e+00
-1.33516222e-01 -1.14939344e+00 1.06350929e-01 -1.42982289e-01
-5.25696158e-01 -2.64803052e-01 6.48433447e-01 -5.26537418e-01
1.45531631e+00 -2.38695550e+00 -2.13491009e-03 1.28221259e-01
-9.64209959e-02 -1.69343781e-02 -6.49088919e-01 6.10862792e-01
1.50341019e-01 4.23774607e-02 3.77076231e-02 -4.10952181e-01
4.07778829e-01 3.65866795e-02 -8.77210677e-01 -4.49468046e-02
2.15127338e-02 8.10738742e-01 -8.87460649e-01 -6.06439173e-01
-2.68148273e-01 4.01084393e-01 -6.62980855e-01 2.96576530e-01
-4.65133548e-01 6.72298908e-01 -5.85521460e-01 3.08229268e-01
3.43942493e-01 -2.36942127e-01 1.85204465e-02 -4.40738723e-02
2.67340802e-02 5.06617665e-01 -1.06048119e+00 2.15117979e+00
-3.05974156e-01 7.40955710e-01 -2.03165948e-01 -4.82681453e-01
1.06261361e+00 5.87528229e-01 5.36174953e-01 -7.73473263e-01
-6.70680180e-02 2.75303602e-01 2.31153712e-01 -4.96503830e-01
7.52089679e-01 -2.17092663e-01 1.02708107e-02 8.19203258e-01
-3.88570875e-02 -3.98845255e-01 3.62941414e-01 2.79875875e-01
1.06365085e+00 1.24430411e-01 -3.41168970e-01 1.67344898e-01
2.71388013e-02 -2.70097077e-01 1.70830250e-01 8.78205180e-01
3.09212506e-01 9.49596584e-01 2.43971333e-01 4.78254855e-01
-1.00362849e+00 -9.13329601e-01 -1.60329267e-02 1.13636827e+00
7.91934431e-02 -6.77390516e-01 -6.74646854e-01 -4.06030208e-01
-2.48721242e-01 8.12327504e-01 -7.93389305e-02 7.06938952e-02
-3.70010793e-01 -5.59039176e-01 1.06317675e+00 3.45981300e-01
1.76409528e-01 -1.11970973e+00 2.40404401e-02 3.06723177e-01
-5.60068011e-01 -1.09276927e+00 -1.07850337e+00 -1.60570130e-01
-7.52668381e-01 -4.71293747e-01 -1.19297540e+00 -6.20741785e-01
1.51757807e-01 4.36034620e-01 7.82269657e-01 -2.17557326e-01
3.48935951e-03 2.59703904e-01 -7.21390009e-01 -4.71197039e-01
-4.63761091e-01 3.18719685e-01 7.09636807e-02 2.02577919e-01
-1.70734897e-01 -9.14903879e-01 -4.80212450e-01 3.46218288e-01
-1.11929107e+00 2.56016433e-01 6.54744446e-01 6.51524484e-01
2.89132953e-01 1.07313342e-01 9.27374125e-01 -5.69398403e-01
1.03085458e+00 -4.05304939e-01 -4.11781758e-01 -1.06608227e-01
-6.73709214e-01 1.18484132e-01 6.25433743e-01 -9.59001899e-01
-9.61144149e-01 -2.38771111e-01 -2.01179594e-01 -5.81861794e-01
2.31988370e-01 7.33960032e-01 1.15116693e-01 4.46252197e-01
6.90759957e-01 3.74902308e-01 -3.71049136e-01 -5.54683030e-01
5.84076762e-01 9.22306240e-01 4.54641670e-01 -8.93089116e-01
1.06314552e+00 2.89723098e-01 -2.43296981e-01 -8.18460047e-01
-7.89112151e-01 -2.38362417e-01 -5.65294437e-02 -2.40584418e-01
5.75507164e-01 -8.88783813e-01 -3.54501754e-01 3.84275854e-01
-1.31336987e+00 -2.52285451e-01 -3.22283238e-01 6.55520976e-01
-6.41815007e-01 6.14729583e-01 -8.31245780e-01 -8.59376132e-01
-5.53873658e-01 -8.68580997e-01 1.12096596e+00 4.86733479e-04
-2.82698423e-01 -5.47948480e-01 1.32207930e-01 2.79607892e-01
4.45039928e-01 -3.54502887e-01 9.30281162e-01 -4.85142499e-01
-7.67240584e-01 7.51815364e-02 -1.39256520e-02 9.53147709e-02
5.58074303e-02 -1.87953457e-01 -1.02654207e+00 -6.69979304e-02
1.17092445e-01 -2.22742185e-01 7.24167049e-01 3.57752383e-01
7.72087932e-01 -3.04879963e-01 7.64239058e-02 5.16172588e-01
8.95422757e-01 2.11126789e-01 5.40640116e-01 8.25843867e-03
5.69240749e-01 4.62938488e-01 5.64985454e-01 7.37828314e-01
1.70748442e-01 8.66776109e-01 -1.45264000e-01 1.33357616e-02
-4.98567998e-01 -6.07079864e-01 6.86621368e-01 1.65235567e+00
1.07982717e-01 -5.83264232e-01 -5.06087363e-01 4.91118968e-01
-1.75250506e+00 -1.19256628e+00 -1.57361805e-01 1.93702102e+00
1.26485455e+00 7.19886869e-02 9.02549103e-02 2.75039285e-01
7.73189902e-01 2.14134306e-01 -4.58930463e-01 2.47857302e-01
-3.35845828e-01 1.88794717e-01 4.04585227e-02 1.57593265e-01
-6.61132157e-01 8.14953983e-01 6.30190849e+00 1.52075863e+00
-1.26123297e+00 1.68156445e-01 1.32909894e-01 -3.57145727e-01
-6.60823345e-01 -8.35917071e-02 -7.77930439e-01 5.96705914e-01
8.29626024e-01 -4.06957060e-01 8.07503104e-01 4.48229671e-01
4.03053284e-01 4.11886513e-01 -1.02265918e+00 1.18123889e+00
2.38312602e-01 -1.19008601e+00 2.75078624e-01 -5.38941137e-02
1.02001393e+00 -2.08590403e-02 3.09709966e-01 2.60871828e-01
2.07221568e-01 -6.97577953e-01 1.08696651e+00 5.55854261e-01
1.06997716e+00 -3.92399341e-01 1.28371015e-01 4.80947673e-01
-1.10066128e+00 1.34776339e-01 -1.40384689e-01 1.54425815e-01
4.49112356e-01 8.59893978e-01 -6.31262958e-01 8.36779416e-01
2.48336092e-01 7.15836406e-01 -1.49094224e-01 1.02011478e+00
-3.44048440e-01 1.11626124e+00 -2.56532460e-01 5.22399321e-03
4.64678295e-02 -3.01412791e-01 6.75162077e-01 1.06562126e+00
9.44740295e-01 7.69646280e-03 2.15077356e-01 9.76933956e-01
-1.03792667e-01 2.71390349e-01 -4.86412615e-01 -5.67766130e-01
5.38458645e-01 1.07293034e+00 -4.78320003e-01 -2.68330246e-01
-2.99687624e-01 9.72140193e-01 -2.91669637e-01 5.42504430e-01
-1.00068867e+00 -4.19663966e-01 1.10354237e-01 6.95654899e-02
4.76150602e-01 -3.22327614e-01 -2.12472156e-01 -1.11984348e+00
1.48293242e-01 -1.17861712e+00 -9.16035660e-03 -1.13019383e+00
-1.67380524e+00 6.36612713e-01 -2.64328301e-01 -1.50229037e+00
-4.58855510e-01 -4.53728586e-02 -4.99710172e-01 8.17516923e-01
-1.46129704e+00 -1.04385161e+00 -1.53532857e-02 5.93626082e-01
6.91240191e-01 -4.54622924e-01 6.26453817e-01 6.72001541e-01
-8.30302015e-02 6.37555063e-01 1.66904151e-01 4.27696146e-02
1.14112854e+00 -7.51268148e-01 3.85874242e-01 4.87724721e-01
7.49270737e-01 5.76690972e-01 6.41024053e-01 -7.85612881e-01
-1.44108498e+00 -9.95101154e-01 9.08593893e-01 -3.03383440e-01
8.77209723e-01 -2.70028621e-01 -6.20509744e-01 2.01117843e-01
5.06043911e-01 -8.00740421e-01 7.29888260e-01 8.48390087e-02
-2.94056386e-01 -1.54033795e-01 -4.52029407e-01 9.68927622e-01
1.14993358e+00 -6.94698513e-01 -4.24340487e-01 1.69467673e-01
1.02926648e+00 -3.13674390e-01 -6.12726331e-01 1.69185400e-01
5.35544991e-01 -5.73810637e-01 7.09803700e-01 -2.46618986e-02
6.71434402e-01 -4.41781610e-01 -1.73511446e-01 -1.16218412e+00
-1.09294154e-01 -1.43197536e+00 -8.72775465e-02 1.54123139e+00
5.41290104e-01 -6.76794276e-02 5.31923890e-01 -1.82126999e-01
-2.92702973e-01 -2.95558244e-01 -6.75783515e-01 -9.36566174e-01
-1.39625967e-01 -8.55719864e-01 7.36606658e-01 1.02679646e+00
1.32302135e-01 8.32357585e-01 -9.27238703e-01 -1.18822202e-01
4.38795090e-01 4.54613566e-01 8.66865396e-01 -1.07940936e+00
-9.23976719e-01 -4.84537542e-01 1.95031688e-01 -1.70001650e+00
4.36608568e-02 -1.12990820e+00 1.45411223e-01 -1.30906165e+00
1.58834606e-01 -5.93451738e-01 -1.10604629e-01 2.27629676e-01
-1.11785784e-01 1.25999555e-01 5.16202033e-01 4.60215390e-01
-5.39179504e-01 8.76879573e-01 1.55788159e+00 -2.74261445e-01
-2.54097641e-01 1.56396076e-01 -7.17145741e-01 4.18811917e-01
6.41533315e-01 -7.66325593e-01 -7.25238621e-01 -6.07279837e-01
4.51378316e-01 4.53918368e-01 -5.59371449e-02 -9.21654880e-01
5.09055316e-01 -8.34115148e-02 -9.17063355e-02 -6.44879162e-01
4.38430876e-01 -4.30968404e-01 1.12807460e-01 -1.13291636e-01
-6.23360574e-01 3.10292700e-03 8.51336792e-02 6.91563427e-01
-1.60651371e-01 -4.57858831e-01 2.48023704e-01 3.34599316e-01
1.13127373e-01 3.12685400e-01 -3.92001182e-01 3.21183860e-01
3.43132585e-01 1.46461576e-01 -1.59250036e-01 -9.41054046e-01
-3.99712950e-01 -1.83977619e-01 1.18288726e-01 6.85575426e-01
4.84620750e-01 -1.56024683e+00 -9.23245788e-01 1.53338432e-01
-5.31902015e-02 -4.78045672e-01 2.21486047e-01 8.51897061e-01
1.56535849e-01 3.74836951e-01 3.29338133e-01 -5.31567097e-01
-1.10291970e+00 4.08457547e-01 -3.26110423e-01 -1.35116637e-01
-6.22659624e-01 5.20686448e-01 2.77141333e-01 -1.35701746e-01
4.31809306e-01 -4.58719209e-02 -5.79384454e-02 -2.26088148e-02
4.89120930e-01 2.38861710e-01 -1.63648009e-01 -2.64010221e-01
3.67566556e-01 5.95227182e-01 2.13641748e-01 -9.12667453e-01
1.08993363e+00 6.02948945e-04 2.15924114e-01 5.84379017e-01
6.34551585e-01 5.85350215e-01 -1.02617550e+00 -2.98399270e-01
-6.50262758e-02 -4.09163266e-01 -4.66375574e-02 -8.25969696e-01
-1.05733621e+00 9.58963394e-01 3.66133720e-01 2.21895546e-01
1.07032764e+00 -4.48317677e-02 1.14528275e+00 3.75007182e-01
2.57331520e-01 -1.13943088e+00 5.65051079e-01 6.61315978e-01
9.27973926e-01 -6.65170491e-01 -3.88092011e-01 -1.99950323e-01
-7.00645089e-01 7.90987492e-01 1.73316911e-01 2.67089188e-01
4.38232094e-01 5.72313905e-01 2.89253265e-01 6.91381544e-02
-8.82931590e-01 -1.88852742e-01 4.76038814e-01 6.49639428e-01
6.08552158e-01 -4.43506092e-02 -4.47522163e-01 9.42271113e-01
-6.61676466e-01 4.24423926e-02 4.00008321e-01 4.74739820e-01
-3.22874278e-01 -1.38376021e+00 -4.16004270e-01 3.35759372e-01
-4.89893943e-01 -6.40306234e-01 -3.77082944e-01 6.46690205e-02
-7.52477124e-02 1.28855121e+00 -2.59232968e-01 -4.53809500e-01
1.25643104e-01 4.82995361e-02 4.69629318e-01 -5.40209830e-01
-5.24979651e-01 8.25569630e-01 4.22977433e-02 2.42789984e-02
-3.37962121e-01 -5.58927536e-01 -1.28963518e+00 -3.42463791e-01
-7.41832078e-01 3.08882713e-01 6.95775747e-01 7.40296423e-01
5.04781067e-01 7.33930826e-01 7.99899280e-01 -7.25595415e-01
-8.94896865e-01 -1.12693751e+00 -8.51143897e-01 4.33272690e-01
9.08421911e-03 -3.30955863e-01 -3.30486149e-01 4.72340614e-01] | [15.539417266845703, 5.74280309677124] |
a259e58f-9f2a-4f57-b343-da36cf17d177 | extracting-dynamical-models-from-data | 2110.06917 | null | https://arxiv.org/abs/2110.06917v5 | https://arxiv.org/pdf/2110.06917v5.pdf | Extracting Dynamical Models from Data | The problem of determining the underlying dynamics of a system when only given data of its state over time has challenged scientists for decades. In this paper, the approach of using machine learning to model the {\em updates} of the phase space variables is introduced; this is done as a function of the phase space variables. (More generally, the modeling is done over the jet space of the variables.) This approach is shown to accurately replicate the dynamics for the examples of the harmonic oscillator, the pendulum, and the Duffing oscillator; the underlying differential equation is also accurately recovered in each example. In addition, the results in no way depend on how the data is sampled over time (i.e., regularly or irregularly). It is demonstrated that this approach (named "FJet") is similar to the model resulting from a Taylor series expansion of the Runge-Kutta (RK) numerical integration scheme. This analogy confers the advantage of explicitly revealing the appropriate functions to use in the modeling, as well as revealing the error estimate of the updates. Thus, this new approach can be thought of as a way to determine the coefficients of an RK scheme by machine learning. Finally, it is shown in the undamped harmonic oscillator example that the stability of the updates is stable for $10^9$ times longer than with $4$th-order RK. | ['Michael F. Zimmer'] | 2021-10-13 | null | null | null | null | ['numerical-integration'] | ['miscellaneous'] | [-2.36498371e-01 6.47431910e-02 6.61622435e-02 3.29107434e-01
-4.55909878e-01 -5.92674434e-01 6.54097736e-01 1.96817163e-02
-4.81035709e-01 1.11354256e+00 -6.74700916e-01 -2.49957159e-01
-3.10554177e-01 -6.23423755e-01 -5.88498712e-01 -1.25899959e+00
-2.69428670e-01 3.28857630e-01 -1.38085365e-01 -5.19124031e-01
3.20513070e-01 4.90825444e-01 -1.46984231e+00 -6.40745819e-01
7.33281791e-01 9.04259682e-01 -2.45531142e-01 1.02858520e+00
9.79019403e-02 5.75335741e-01 -6.87941134e-01 1.26666695e-01
1.58370152e-01 -6.84910476e-01 -6.73283160e-01 -1.30966946e-01
-1.67765781e-01 8.78916159e-02 -2.99797654e-01 9.49604809e-01
3.22597891e-01 1.94564849e-01 8.39867890e-01 -8.87394786e-01
2.17786944e-03 1.27676323e-01 -3.13644558e-01 3.12292039e-01
3.61572593e-01 5.95144965e-02 5.87178886e-01 -5.77786446e-01
4.26435739e-01 6.02714419e-01 1.04812622e+00 1.64604709e-01
-1.53768337e+00 -3.64492893e-01 -4.79413390e-01 6.38028011e-02
-1.57440507e+00 -1.44560888e-01 1.02190208e+00 -7.26523161e-01
5.29749870e-01 3.09544265e-01 9.96373296e-01 2.66754419e-01
7.79511154e-01 1.84414700e-01 1.07610953e+00 -8.53111565e-01
4.61672306e-01 2.52230525e-01 2.26840705e-01 5.84794819e-01
1.55482069e-01 4.51392323e-01 -1.34217635e-01 -4.24398839e-01
7.83499479e-01 -4.43887562e-01 -4.02983785e-01 -5.04173815e-01
-6.53204978e-01 1.03977466e+00 3.13148685e-02 5.36228180e-01
-2.66659975e-01 4.04408813e-01 3.46009821e-01 4.70524490e-01
6.05374694e-01 8.12986493e-01 -5.00492513e-01 -3.16881508e-01
-9.19805467e-01 5.99315763e-01 1.06029069e+00 5.02275825e-01
7.98661172e-01 4.41485971e-01 5.91774642e-01 3.14662158e-01
1.35931671e-01 5.87401688e-01 6.43831313e-01 -1.15388584e+00
-1.82689011e-01 4.67429943e-02 4.69618887e-01 -7.98706174e-01
-6.61690712e-01 -3.96471739e-01 -7.54353762e-01 5.28997242e-01
8.68556201e-01 -4.67976063e-01 -3.86152178e-01 1.61032891e+00
4.75766867e-01 1.71598807e-01 6.20171279e-02 4.17611301e-01
9.64436159e-02 9.21714365e-01 -4.83352542e-01 -9.55018163e-01
9.67709363e-01 -1.90496519e-01 -1.19487834e+00 3.80171925e-01
7.75024652e-01 -6.97646856e-01 5.54605186e-01 4.24335599e-01
-1.34051299e+00 -6.92470253e-01 -1.20990431e+00 5.19553423e-01
-2.94822484e-01 -5.43123521e-02 2.43336678e-01 3.10714871e-01
-1.12897789e+00 1.19045281e+00 -9.55511272e-01 -4.70841751e-02
-3.52062792e-01 3.29839230e-01 1.05391845e-01 6.88706517e-01
-1.38063657e+00 1.19197130e+00 2.29240820e-01 3.39031667e-01
-1.25256389e-01 -9.54507351e-01 -6.81868792e-01 -1.85142130e-01
-2.23152608e-01 -3.52197707e-01 1.48885715e+00 -1.01263785e+00
-1.76953316e+00 4.53778207e-01 -4.02535468e-01 -3.77769738e-01
6.04889214e-01 2.51304090e-01 -3.16458166e-01 2.39687726e-01
1.08552612e-02 -1.91545248e-01 1.05672383e+00 -1.01308882e+00
-3.01990122e-01 -1.35214135e-01 -1.36773050e-01 -9.18639451e-02
1.69817984e-01 -3.76600862e-01 2.49065399e-01 -4.38991427e-01
2.58643389e-01 -1.03107500e+00 -2.91582793e-01 -1.77256852e-01
1.82252619e-02 -2.23820120e-01 6.01169884e-01 -6.35975659e-01
1.39656568e+00 -2.12636948e+00 2.95822561e-01 3.37801784e-01
-8.01428109e-02 7.99054578e-02 7.34303772e-01 9.25622106e-01
-4.45852429e-01 -3.59158181e-02 -4.76298183e-01 9.17915702e-02
-2.11055145e-01 1.20459557e-01 -4.94749874e-01 8.79916191e-01
2.69817561e-01 4.20322031e-01 -7.89984405e-01 -2.30670080e-01
8.40257630e-02 4.32094514e-01 -3.10082793e-01 -1.15401186e-01
6.01750091e-02 7.51320720e-01 -4.49132413e-01 -1.89897642e-01
3.45791787e-01 -2.93381840e-01 1.94257170e-01 -1.20471507e-01
-6.76806509e-01 1.34531055e-02 -1.44651747e+00 1.11675072e+00
-5.00900924e-01 7.97912717e-01 1.46113187e-01 -1.44739115e+00
9.21965599e-01 7.52417862e-01 8.02036881e-01 -5.86902857e-01
1.52423978e-01 5.07220924e-01 8.17749351e-02 -5.72005570e-01
1.07187867e-01 -5.29736996e-01 -6.08389974e-02 4.59967703e-01
-1.44218206e-01 -5.83125293e-01 3.23842347e-01 -1.38367757e-01
6.85982168e-01 8.94021839e-02 6.89159095e-01 -6.66070223e-01
7.07457125e-01 3.91929775e-01 3.13385695e-01 4.01221782e-01
-4.09507938e-02 1.35910973e-01 6.38776600e-01 -4.95384097e-01
-1.19881773e+00 -5.82368255e-01 -8.91119659e-01 1.63410589e-01
2.20101122e-02 -2.30234504e-01 -7.68863380e-01 3.25361580e-01
7.06100985e-02 6.47969246e-01 -7.11380899e-01 -2.44816318e-01
-9.03568387e-01 -7.98960745e-01 2.50685543e-01 1.53646111e-01
3.23280066e-01 -7.73719728e-01 -6.35987401e-01 4.40397531e-01
2.19643805e-02 -6.71349466e-01 -6.72878325e-02 3.01703930e-01
-1.03655875e+00 -1.19390392e+00 -5.11521339e-01 -6.37269795e-01
4.06090170e-01 -3.98827881e-01 7.14174092e-01 -1.40229464e-01
-3.35274190e-01 6.40352845e-01 6.48652390e-03 -3.26118499e-01
-6.05947316e-01 -2.20284745e-01 3.59309286e-01 -2.37121955e-02
-1.33870319e-01 -7.61157334e-01 -4.00163502e-01 7.11933821e-02
-7.40264893e-01 -5.63829482e-01 -6.89129904e-02 9.21684027e-01
5.31323135e-01 4.14019465e-01 6.09180093e-01 -4.84496295e-01
5.63825667e-01 -3.34366143e-01 -1.07597148e+00 -2.57303327e-01
-7.23514736e-01 2.30730936e-01 1.06937468e+00 -6.03046656e-01
-6.12251163e-01 1.07154533e-01 -5.35443723e-02 -1.70551494e-01
2.96796679e-01 4.80217844e-01 4.33428586e-01 -4.91448373e-01
6.13713205e-01 3.94216537e-01 4.53226238e-01 -4.73171115e-01
-1.05669666e-02 4.54224110e-01 3.36649269e-01 -5.07922113e-01
7.62127578e-01 4.83923823e-01 3.96482617e-01 -1.23984814e+00
-3.37835163e-01 -2.90410012e-01 -6.30817950e-01 -2.07889169e-01
6.29730761e-01 -4.79134858e-01 -1.17909884e+00 6.37075067e-01
-1.13172829e+00 -3.00089896e-01 -7.57782578e-01 6.29700005e-01
-1.01214874e+00 2.67687231e-01 -7.17889428e-01 -1.37755668e+00
-7.69489482e-02 -9.80085671e-01 5.61100781e-01 2.70660102e-01
-3.71054947e-01 -1.30804741e+00 5.84453285e-01 -2.81079262e-01
2.93768197e-01 4.66576397e-01 9.43720162e-01 -1.61006033e-01
-1.62610352e-01 -5.36431313e-01 5.37968099e-01 1.66770175e-01
-1.32052839e-01 4.07697320e-01 -8.67753625e-01 -4.23750699e-01
8.56795967e-01 2.23537877e-01 2.47918263e-01 8.46480370e-01
5.17476201e-01 -3.39376122e-01 -3.79069149e-01 5.26175499e-01
1.68756080e+00 6.60215974e-01 2.30983078e-01 2.17343062e-01
1.63779855e-01 4.76958454e-01 3.63538295e-01 3.47038120e-01
-6.35170043e-02 8.23028147e-01 3.67729217e-02 6.16788305e-02
4.94639695e-01 4.45209891e-02 2.29710057e-01 9.36874151e-01
-9.25156474e-02 3.56319964e-01 -7.64221668e-01 3.06014419e-01
-1.57734895e+00 -1.00106621e+00 -2.75463909e-01 2.30603814e+00
8.97188008e-01 8.60340372e-02 2.59875059e-01 7.58669913e-01
6.10411286e-01 -1.17541216e-02 -5.56412756e-01 -7.81078815e-01
1.36194751e-01 4.81568247e-01 4.78898227e-01 9.47352827e-01
-9.57872093e-01 1.20026805e-01 7.32170248e+00 6.32393956e-01
-1.36751008e+00 -1.70812070e-01 2.77072281e-01 3.16627771e-01
1.13957629e-01 9.94910002e-02 -7.97442734e-01 8.24884713e-01
1.16209280e+00 -6.55610323e-01 4.49723423e-01 6.25272810e-01
6.11897886e-01 -2.82907784e-01 -6.94373786e-01 6.86883211e-01
-2.50816405e-01 -1.27810454e+00 -6.22755349e-01 5.40048555e-02
5.76388359e-01 -5.44058621e-01 -1.04368173e-01 1.74375057e-01
-1.39233425e-01 -7.77834415e-01 6.83977067e-01 8.13033640e-01
7.44028032e-01 -7.02942193e-01 6.92441940e-01 7.92893112e-01
-1.23580384e+00 -1.58600152e-01 -5.84224574e-02 -4.58731741e-01
2.94045091e-01 5.51242173e-01 -6.66651428e-01 5.55463195e-01
3.82634908e-01 5.39762795e-01 1.36441942e-02 9.96284723e-01
2.40076467e-01 7.83437729e-01 -6.57967746e-01 -3.20103109e-01
7.81056136e-02 -7.57662416e-01 6.09589517e-01 7.71880567e-01
4.67369080e-01 3.19300413e-01 -1.10870600e-01 7.53848493e-01
6.49811983e-01 -3.90853621e-02 -5.34738421e-01 1.00005507e-01
2.47127101e-01 8.68228912e-01 -3.91642690e-01 -1.83001664e-02
-1.12715326e-01 3.40806127e-01 -1.36204481e-01 7.24094033e-01
-5.80039740e-01 -6.15988791e-01 4.84285295e-01 4.08045918e-01
5.44458508e-01 -4.23234940e-01 -7.19822720e-02 -7.56031752e-01
6.26845881e-02 -3.93109143e-01 1.64482802e-01 -6.28875613e-01
-9.30690229e-01 5.13216615e-01 1.97647825e-01 -1.40269649e+00
-9.91818368e-01 -4.50107604e-01 -7.41076827e-01 1.17266083e+00
-1.36772788e+00 -2.02498794e-01 3.73496801e-01 3.99866730e-01
2.57078707e-02 1.86929226e-01 8.65482092e-01 1.20190181e-01
-5.51921606e-01 8.80207866e-02 8.41453552e-01 -1.27958916e-02
1.43033549e-01 -1.33260047e+00 -1.67598039e-01 5.00673652e-01
-2.95608371e-01 5.26867449e-01 1.47212505e+00 -2.24624738e-01
-1.47009099e+00 -5.38866878e-01 8.16253126e-01 -2.19022781e-01
9.99522209e-01 3.24140936e-02 -1.01778173e+00 3.89452457e-01
9.74946003e-03 2.27264702e-01 2.98415959e-01 -2.96678483e-01
4.43734288e-01 -4.57355022e-01 -1.04804039e+00 2.23565236e-01
1.66488051e-01 -5.13275623e-01 -5.01074672e-01 2.21435145e-01
4.94452894e-01 -4.48187649e-01 -1.03570700e+00 2.86018401e-01
5.43513298e-01 -7.60514379e-01 6.82421684e-01 -5.36988914e-01
6.33409768e-02 -5.03464520e-01 1.86657831e-01 -1.34079480e+00
-1.70637175e-01 -1.15111041e+00 -4.23219323e-01 7.15954065e-01
2.78133512e-01 -1.10609031e+00 5.46469629e-01 2.26639807e-01
1.50110364e-01 -1.01985073e+00 -1.32100844e+00 -7.39584208e-01
5.09815633e-01 -1.61300734e-01 -6.49427399e-02 9.48256254e-01
4.44775283e-01 7.92592298e-03 1.54676782e-02 3.83099914e-01
5.84852993e-01 1.79068163e-01 4.01509315e-01 -1.17501795e+00
-5.60378790e-01 -4.02386397e-01 -4.70562458e-01 -9.53894258e-01
1.15852728e-02 -4.82581615e-01 -4.01280373e-02 -9.10714090e-01
-5.80407441e-01 -5.86959183e-01 -1.24259079e-02 -3.48470539e-01
-5.80373220e-02 -6.26202375e-02 -1.15942001e-01 4.27565157e-01
2.83014357e-01 3.73354942e-01 1.16616642e+00 2.93001026e-01
-3.86141896e-01 5.62708318e-01 -9.52067822e-02 7.89261162e-01
7.60179281e-01 -3.38402420e-01 -2.89094776e-01 2.60102302e-01
2.51817554e-01 5.77547252e-01 3.50678355e-01 -1.02568531e+00
3.48995477e-01 1.69319913e-01 1.20809607e-01 -5.00826061e-01
4.68151867e-01 -6.36367381e-01 5.12013555e-01 7.83107102e-01
-1.64304867e-01 3.38125855e-01 2.94153601e-01 5.49665034e-01
-3.59529763e-01 -5.04476130e-01 1.03007829e+00 5.25074042e-02
-1.83740035e-01 -8.85259435e-02 -7.13695228e-01 7.19654933e-02
9.34172809e-01 -2.72224903e-01 5.63591942e-02 -6.11881614e-01
-8.69954288e-01 -2.20601670e-02 3.10428083e-01 -6.30879521e-01
2.17655152e-02 -1.17172718e+00 -3.74352634e-01 5.19003868e-01
-3.79915714e-01 -3.73357683e-02 9.14764702e-02 9.77096319e-01
-6.69415593e-01 5.09764552e-01 1.65904745e-01 -7.66718209e-01
-7.55168140e-01 3.52403134e-01 9.29011881e-01 -3.56645703e-01
-4.22989726e-01 3.60606939e-01 -3.68244320e-01 -1.57805979e-01
-3.00586224e-02 -4.62506860e-01 -1.68399289e-01 1.36856198e-01
5.00110626e-01 3.74853790e-01 -3.53444889e-02 -4.76756662e-01
8.78949165e-02 1.15318346e+00 4.48217779e-01 -2.85756290e-01
1.14425468e+00 -1.25000710e-02 -2.36729756e-01 9.71299946e-01
1.33647907e+00 6.72082081e-02 -1.26524675e+00 -8.36570561e-02
-1.41127929e-01 2.10604593e-01 -1.72756001e-01 -2.00157031e-01
-6.95616782e-01 7.47221172e-01 4.15281504e-01 1.01523244e+00
1.02774167e+00 -3.89814705e-01 4.35968876e-01 2.22312495e-01
3.24814886e-01 -1.02252829e+00 -4.15589839e-01 5.18174291e-01
7.11081684e-01 -8.01846862e-01 6.39331788e-02 -2.41769522e-01
-7.37723801e-03 1.53557968e+00 -1.27988057e-02 -5.87793052e-01
1.09596777e+00 4.72450048e-01 -5.18703386e-02 7.94495270e-02
-5.26102006e-01 2.42185920e-01 2.02569421e-02 2.48927280e-01
4.43167388e-01 -1.53180540e-01 -6.83604479e-01 1.70824841e-01
-2.85485357e-01 5.75642437e-02 7.06742108e-01 8.71720433e-01
-4.44012612e-01 -9.41660643e-01 -5.92360139e-01 4.33249362e-02
-2.49902621e-01 2.70762861e-01 1.60317183e-01 1.31640685e+00
-6.77319765e-02 6.36848569e-01 1.60370305e-01 4.93212081e-02
4.43998605e-01 4.06405300e-01 4.66103524e-01 -1.31316349e-01
-3.42230946e-01 1.13260567e-01 -1.39912486e-01 -3.61144960e-01
-3.78747433e-01 -8.00904214e-01 -1.40370250e+00 -2.96783179e-01
-3.34701359e-01 8.34172785e-01 6.58245802e-01 1.06879997e+00
-1.37132645e-01 4.36990976e-01 9.84773934e-01 -8.82699668e-01
-9.26427543e-01 -7.58927763e-01 -9.55996931e-01 2.20848694e-02
8.08449745e-01 -6.64138019e-01 -1.09793293e+00 -4.12642322e-02] | [6.469962120056152, 3.469964027404785] |
c6ecae4e-a85d-426c-a984-90838b6eff3f | an-lstm-model-for-twitter-sentiment-analysis | 2212.01791 | null | https://arxiv.org/abs/2212.01791v1 | https://arxiv.org/pdf/2212.01791v1.pdf | An LSTM model for Twitter Sentiment Analysis | Sentiment analysis on social media such as Twitter provides organizations and individuals an effective way to monitor public emotions towards them and their competitors. As a result, sentiment analysis has become an important and challenging task. In this work, we have collected seven publicly available and manually annotated twitter sentiment datasets. We create a new training and testing dataset from the collected datasets. We develop an LSTM model to classify sentiment of a tweet and evaluate the model with the new dataset. | ['Md Parvez Mollah'] | 2022-12-04 | null | null | null | null | ['twitter-sentiment-analysis'] | ['natural-language-processing'] | [-1.36222467e-01 -3.50695133e-01 -5.37416518e-01 -9.26132381e-01
-3.67239416e-01 -4.01123464e-01 6.75631523e-01 5.33579826e-01
-6.67836905e-01 7.70511746e-01 3.34227771e-01 -9.28688571e-02
7.11497307e-01 -8.64041209e-01 -2.45793253e-01 -2.48153076e-01
3.08053583e-01 8.87494013e-02 -1.71066433e-01 -6.10885382e-01
4.62373644e-01 7.34587461e-02 -1.43700075e+00 5.05023122e-01
4.37534839e-01 1.49276245e+00 -3.58658671e-01 3.73003602e-01
-4.85934854e-01 1.05051327e+00 -7.98465371e-01 -5.52525222e-01
-6.84527084e-02 -1.90297574e-01 -7.33026326e-01 7.68980430e-03
-1.07648134e-01 1.64251819e-01 1.17839649e-01 9.83352304e-01
5.23266554e-01 2.56002486e-01 2.37830207e-01 -1.27124643e+00
-5.98291516e-01 6.43083870e-01 -7.06690967e-01 2.09662393e-01
2.72364378e-01 -2.92642772e-01 8.45220923e-01 -7.30643511e-01
3.83710384e-01 8.21785033e-01 6.10677600e-01 3.44925016e-01
-7.01642871e-01 -8.75854969e-01 3.82450610e-01 -1.27936423e-01
-8.48585486e-01 -3.16974729e-01 9.61451352e-01 -3.05713236e-01
7.75118291e-01 5.33186495e-02 7.20243692e-01 1.14371002e+00
2.82287508e-01 8.47723603e-01 1.44603407e+00 -1.29135087e-01
-4.50411774e-02 7.92880237e-01 4.03050989e-01 2.88736850e-01
1.25190625e-02 -6.00906968e-01 -5.87691605e-01 1.39195427e-01
-2.15347707e-01 1.86289668e-01 2.32791066e-01 4.57613707e-01
-8.43697488e-01 1.04126596e+00 7.59736776e-01 3.48660618e-01
-4.10359293e-01 1.61250278e-01 7.99062073e-01 4.25600737e-01
1.15984118e+00 4.13103044e-01 -6.16611838e-01 -3.07455897e-01
-6.19091928e-01 -1.06080391e-01 7.29511142e-01 3.70651156e-01
6.54931605e-01 9.97600630e-02 4.26878154e-01 9.30618048e-01
3.00116092e-01 7.38451004e-01 9.48518813e-01 5.15800156e-03
2.93957889e-01 1.04554129e+00 5.79636171e-02 -1.60118353e+00
-6.78646922e-01 -4.20867264e-01 -7.39289343e-01 -3.21006149e-01
-2.07429081e-01 -6.01646185e-01 -5.52713871e-01 1.34087121e+00
2.37911597e-01 4.63991202e-02 9.94878486e-02 4.78733748e-01
1.23486805e+00 7.84795642e-01 1.63633302e-01 -1.91245288e-01
9.71868873e-01 -9.68650877e-01 -8.21670473e-01 -4.78747219e-01
7.04969764e-01 -6.20377719e-01 1.14151955e+00 5.34900486e-01
-7.22671747e-01 -6.03008866e-01 -1.03730953e+00 1.62766367e-01
-1.09164560e+00 3.08700483e-02 9.30285871e-01 6.28307879e-01
-7.03688443e-01 2.96810657e-01 -6.02007508e-01 -6.04882874e-02
6.48091078e-01 4.61533338e-01 -2.56805837e-01 4.15425122e-01
-1.47443509e+00 6.57721877e-01 1.18547738e-01 4.08160716e-01
-1.29979104e-01 2.76834071e-02 -9.54640925e-01 -3.28224570e-01
-1.10996822e-02 1.00010090e-01 1.47194254e+00 -1.58516598e+00
-1.42077756e+00 1.02627861e+00 -2.20158324e-01 -3.62827808e-01
9.90273431e-02 -7.33403936e-02 -8.08179438e-01 -4.57825541e-01
3.27915728e-01 2.85135955e-01 6.62888288e-01 -9.99389410e-01
-5.78484416e-01 -4.06125069e-01 -5.99113889e-02 -1.70545578e-02
-9.68693376e-01 5.10042429e-01 3.16971168e-02 -4.29174364e-01
-2.03847393e-01 -9.73301232e-01 -2.60086238e-01 -1.08421433e+00
-6.44423664e-01 -3.82975429e-01 9.00892675e-01 -2.45580450e-01
1.20896173e+00 -1.96418488e+00 -3.61855298e-01 2.17494592e-01
1.81736857e-01 2.70799994e-01 6.42404631e-02 3.18763167e-01
-1.43253252e-01 5.12249351e-01 3.60716343e-01 -6.63002014e-01
-1.34958640e-01 -2.29959935e-01 -3.68111402e-01 3.15951228e-01
2.38616511e-01 7.92418599e-01 -7.94562697e-01 -4.07060444e-01
-6.66254237e-02 4.87672061e-01 -3.00767660e-01 2.20478311e-01
-1.21151790e-01 6.46518052e-01 -8.99794877e-01 6.26612127e-01
4.01580036e-01 -4.58674103e-01 -1.03383712e-01 5.06314300e-02
-3.43962193e-01 2.20918223e-01 -5.77727079e-01 9.74325538e-01
-8.25822592e-01 1.00285852e+00 -4.51363832e-01 -8.31539869e-01
1.36404967e+00 1.57630831e-01 4.60228145e-01 -9.67417419e-01
8.52067173e-01 2.96824813e-01 -3.83158684e-01 -5.71137130e-01
7.45279789e-01 -4.66533780e-01 -7.01753199e-01 7.46633947e-01
-1.85921967e-01 -1.85198799e-01 2.31830865e-01 -8.51823203e-03
4.49994773e-01 -3.87801558e-01 1.10792018e-01 7.32368156e-02
9.18926120e-01 -1.52713358e-02 2.06461549e-01 1.39977440e-01
-3.03262949e-01 5.89882396e-02 5.23544848e-01 -8.75787199e-01
-4.21741456e-01 -1.31921666e-02 -6.50843009e-02 1.36353862e+00
-6.86982125e-02 -3.87210071e-01 -4.24557239e-01 -8.67711246e-01
-2.29763523e-01 4.05448824e-01 -8.96120846e-01 1.50731914e-02
-2.95908660e-01 -9.03975964e-01 3.08259819e-02 5.69346607e-01
8.11937809e-01 -1.39218402e+00 -3.32872331e-01 1.40474122e-02
-3.39968890e-01 -1.16396165e+00 -1.16019249e-01 2.11210459e-01
-2.09018871e-01 -8.82702589e-01 -5.09903878e-02 -1.04978240e+00
6.43638551e-01 2.88612157e-01 1.05592620e+00 6.92836270e-02
2.97254205e-01 -2.50005901e-01 -5.20567954e-01 -1.26125979e+00
-6.24452047e-02 5.02476335e-01 1.18592776e-01 4.07545328e-01
8.05394471e-01 -9.50123593e-02 -3.75786811e-01 3.22484791e-01
-8.08585286e-01 -1.45259991e-01 -6.26152530e-02 5.02858996e-01
4.33796465e-01 1.68874368e-01 1.18873084e+00 -1.31769764e+00
1.12421870e+00 -6.70111775e-01 -4.59914595e-01 -3.17207634e-01
-3.76933813e-01 -5.01719832e-01 9.49027181e-01 -1.11291021e-01
-9.38581228e-01 -1.13110371e-01 -3.17391038e-01 2.47723013e-01
1.16410755e-01 1.23344803e+00 2.06098706e-01 -8.62102062e-02
4.56489444e-01 -1.81517810e-01 -3.25787127e-01 -1.40842497e-01
1.13694496e-01 1.24974465e+00 6.36197813e-03 -2.21054014e-02
3.58091265e-01 5.76796830e-01 -4.26408350e-01 -6.84967637e-01
-1.71065426e+00 -5.35595119e-01 -4.08374608e-01 -5.17075360e-01
8.55059385e-01 -8.64291608e-01 -9.31886673e-01 7.41156220e-01
-8.34108412e-01 -1.11895293e-01 -3.95531813e-03 2.96362936e-01
5.83288632e-02 -3.05868298e-01 -5.34147501e-01 -7.15471685e-01
-5.53412378e-01 -1.14500904e+00 8.07711363e-01 4.43907827e-01
-4.79259282e-01 -1.16416705e+00 4.32057619e-01 4.86338824e-01
8.42472911e-01 6.03710651e-01 -2.80301794e-02 -1.09968698e+00
3.51022094e-01 -9.63224173e-01 -2.03363568e-01 6.97971642e-01
3.48491639e-01 2.47984484e-01 -1.03827131e+00 -1.37699723e-01
3.38156372e-01 -1.02538884e+00 1.03258228e+00 1.21726848e-01
1.32007051e+00 -2.90401876e-01 -1.95361838e-01 3.42049509e-01
1.18107557e+00 5.24156541e-02 1.63282514e-01 8.70939851e-01
6.53989792e-01 8.04202914e-01 7.87022114e-01 4.94163424e-01
6.10492766e-01 -4.33131233e-02 2.96198666e-01 -3.70070964e-01
8.53633404e-01 -8.36768895e-02 4.08194929e-01 1.16859818e+00
1.13554567e-01 -3.98401529e-01 -1.18434393e+00 3.16722512e-01
-1.64839244e+00 -8.46021116e-01 -2.32307255e-01 1.61331379e+00
8.00631642e-01 5.19327164e-01 4.58548777e-02 2.57861078e-01
7.04865634e-01 5.51024318e-01 -4.73043919e-01 -8.56402814e-01
-2.34104529e-01 4.30997938e-01 3.66485834e-01 3.45736355e-01
-1.50174069e+00 1.16462529e+00 6.15013266e+00 1.41925722e-01
-1.93063390e+00 -1.76372882e-02 1.20748603e+00 -2.11209118e-01
-1.49404258e-01 -5.44338584e-01 -7.86826670e-01 5.97409487e-01
1.22502649e+00 -2.84759462e-01 -1.62320673e-01 1.07435513e+00
3.71833712e-01 1.59202598e-03 -5.51523328e-01 8.44207942e-01
3.49763691e-01 -1.20980179e+00 -2.37950370e-01 -1.85171634e-01
1.09776878e+00 4.43901509e-01 4.06345665e-01 5.63812971e-01
2.32088670e-01 -1.28343713e+00 2.84928262e-01 2.68817455e-01
5.96567273e-01 -9.52965021e-01 1.33257282e+00 3.76010723e-02
-7.95414746e-01 -3.78363505e-02 5.37976287e-02 -4.94213343e-01
2.51110971e-01 5.47373116e-01 -6.89945102e-01 -1.03024133e-01
6.58623636e-01 1.32444477e+00 -7.92001069e-01 3.23991537e-01
-4.11271334e-01 6.78807735e-01 2.76132394e-02 -6.12236321e-01
4.13397193e-01 -1.28637448e-01 -1.61413565e-01 1.21115935e+00
-1.64549917e-01 -3.22659433e-01 3.86161387e-01 1.87297650e-02
-4.82969791e-01 4.71459329e-01 -5.08297205e-01 -4.74770755e-01
2.80436296e-02 1.70836222e+00 -9.74316478e-01 -4.72088516e-01
-5.49138546e-01 5.13920426e-01 2.78727323e-01 1.18617266e-01
-9.43375766e-01 -6.11039579e-01 3.72242957e-01 -1.39005885e-01
-2.27978200e-01 -3.60678602e-03 -2.73421615e-01 -1.26107252e+00
-1.13380097e-01 -1.00723934e+00 2.43945375e-01 -5.74741125e-01
-1.33771873e+00 1.04350781e+00 -6.37534916e-01 -9.65206087e-01
-8.36071149e-02 -6.67494357e-01 -7.90503085e-01 5.16249955e-01
-1.85405231e+00 -1.19469905e+00 -5.84433436e-01 5.41412115e-01
4.49552119e-01 -2.93169975e-01 6.74859822e-01 4.48673278e-01
-9.31438804e-01 3.63322765e-01 -1.61496073e-01 4.27795947e-01
8.55236232e-01 -1.03339946e+00 4.68469918e-01 3.32774758e-01
-2.22594708e-01 5.67596138e-01 6.14417017e-01 -4.40647572e-01
-1.10880125e+00 -1.14091802e+00 1.03736317e+00 -5.31073511e-01
1.16695964e+00 -3.15828800e-01 -3.57113510e-01 1.02026772e+00
4.28425461e-01 -1.11571774e-01 1.31065750e+00 2.12716058e-01
9.04351845e-02 -1.47056624e-01 -1.07065463e+00 2.96999991e-01
2.20078707e-01 -5.37374079e-01 -2.07832947e-01 5.53675234e-01
7.68703163e-01 -1.10692836e-01 -6.82589710e-01 1.84130922e-01
5.44147789e-01 -8.13072681e-01 4.02095377e-01 -8.55786741e-01
8.06211412e-01 4.04123180e-02 -6.23337328e-02 -1.63880169e+00
3.84056777e-01 -3.36053163e-01 3.59022647e-01 1.22590601e+00
6.72682345e-01 -1.00718868e+00 1.03655469e+00 6.30872726e-01
-8.93713068e-03 -8.67133200e-01 -2.10334122e-01 1.17859975e-01
-9.80694443e-02 -7.08836854e-01 5.59947252e-01 1.42159247e+00
3.77613157e-01 7.52864361e-01 -2.50499696e-01 -4.29011643e-01
1.69112738e-02 3.16665232e-01 9.57348049e-01 -1.33553362e+00
3.00471663e-01 -3.90055835e-01 -3.26447695e-01 -5.38167238e-01
5.15181184e-01 -7.70014524e-01 -2.63964504e-01 -1.43676007e+00
-1.35696474e-02 -4.30204988e-01 -6.03492916e-01 5.15393496e-01
5.38965762e-02 8.09159636e-01 -1.42003372e-01 -2.24417001e-01
-9.36816454e-01 4.50420469e-01 1.03769314e+00 -2.62450486e-01
-2.95186698e-01 1.36329055e-01 -1.27092934e+00 1.08938205e+00
1.61581850e+00 -4.12242055e-01 7.81352073e-02 -2.75022149e-01
1.25549388e+00 -3.65877450e-01 -2.57834107e-01 -5.20039618e-01
9.36145037e-02 -2.03399882e-01 4.45019931e-01 -6.92685664e-01
3.13653052e-01 -4.41605598e-01 -6.10542953e-01 2.38983154e-01
-5.70947111e-01 3.80786777e-01 2.50691444e-01 2.22761169e-01
-7.53540874e-01 1.55566439e-01 4.67790127e-01 5.76896966e-03
-5.45777798e-01 2.59666115e-01 -3.50200355e-01 7.83265531e-02
9.28163588e-01 2.05153480e-01 -5.16206563e-01 -6.58501506e-01
-6.57325506e-01 6.53819978e-01 2.93506414e-01 6.91381633e-01
4.99882728e-01 -1.18664253e+00 -6.61622167e-01 1.83202401e-01
1.37171775e-01 -1.76521465e-01 -4.23445962e-02 7.87146628e-01
-2.53669620e-01 2.82402962e-01 -1.83421418e-01 -2.23978579e-01
-1.19027221e+00 3.10158789e-01 1.28282860e-01 -2.06378683e-01
2.99990714e-01 1.07386243e+00 -2.76067853e-01 -7.34839022e-01
-1.42458174e-02 -2.60763139e-01 -1.08027279e+00 7.43254483e-01
6.15092218e-01 4.96881753e-02 1.45713136e-01 -1.12110364e+00
-2.37053454e-01 3.31069648e-01 -1.76488042e-01 1.93621770e-01
1.62739682e+00 -5.23240364e-04 -4.19943124e-01 8.92350137e-01
1.48726070e+00 3.43283474e-01 -3.02733600e-01 7.10873380e-02
-9.05359280e-04 -4.70235348e-01 2.94001609e-01 -4.98639435e-01
-1.28951752e+00 7.41344988e-01 2.93902427e-01 8.17922831e-01
9.39287245e-01 -2.13962436e-01 1.14723718e+00 5.26810944e-01
3.90602648e-02 -1.54116309e+00 4.13323522e-01 8.65523040e-01
6.30131245e-01 -1.74158287e+00 -9.07123610e-02 -2.40638871e-02
-1.03476799e+00 8.98021996e-01 7.05995977e-01 -3.01070005e-01
9.85588729e-01 3.67133319e-02 5.97966850e-01 -4.50631738e-01
-6.85082436e-01 -1.32544547e-01 2.47250553e-02 5.57487346e-02
9.96553659e-01 1.29447710e-02 -3.70210260e-01 9.00221109e-01
-7.00700760e-01 2.67412178e-02 7.23205924e-01 9.50665891e-01
-4.21649724e-01 -1.04487479e+00 -2.34613363e-02 7.16166258e-01
-1.01727962e+00 -1.07136238e-02 -7.50666976e-01 4.27646190e-01
-8.88807625e-02 1.31482494e+00 1.64777506e-02 -8.09662998e-01
3.18840891e-01 -1.46727145e-01 -4.68065321e-01 -8.02990139e-01
-1.10731852e+00 -2.30751693e-01 3.99697214e-01 -3.86946648e-01
-8.59013617e-01 -4.67490643e-01 -1.26796687e+00 -2.49094114e-01
-4.93156254e-01 4.18906093e-01 1.25071335e+00 7.54545331e-01
3.22844297e-01 4.83404726e-01 1.29519856e+00 -8.65585983e-01
-1.78904936e-01 -9.25534070e-01 -2.83807278e-01 5.42724133e-01
4.98995304e-01 -4.80481714e-01 -4.81398642e-01 -1.69001102e-01] | [11.179594993591309, 6.943390846252441] |
0961f8a5-09cc-4715-a8df-106216bb7f38 | efficient-multi-order-gated-aggregation | 2211.03295 | null | https://arxiv.org/abs/2211.03295v2 | https://arxiv.org/pdf/2211.03295v2.pdf | Efficient Multi-order Gated Aggregation Network | Since the recent success of Vision Transformers (ViTs), explorations toward ViT-style architectures have triggered the resurgence of ConvNets. In this work, we explore the representation ability of modern ConvNets from a novel view of multi-order game-theoretic interaction, which reflects inter-variable interaction effects w.r.t.~contexts of different scales based on game theory. Within the modern ConvNet framework, we tailor the two feature mixers with conceptually simple yet effective depthwise convolutions to facilitate middle-order information across spatial and channel spaces respectively. In this light, a new family of pure ConvNet architecture, dubbed MogaNet, is proposed, which shows excellent scalability and attains competitive results among state-of-the-art models with more efficient use of parameters on ImageNet and multifarious typical vision benchmarks, including COCO object detection, ADE20K semantic segmentation, 2D\&3D human pose estimation, and video prediction. Typically, MogaNet hits 80.0\% and 87.8\% top-1 accuracy with 5.2M and 181M parameters on ImageNet, outperforming ParC-Net-S and ConvNeXt-L while saving 59\% FLOPs and 17M parameters. The source code is available at \url{https://github.com/Westlake-AI/MogaNet}. | ['Stan Z. Li', 'Jiangbin Zheng', 'ZhiYuan Chen', 'Di wu', 'Haitao Lin', 'Cheng Tan', 'Zicheng Liu', 'Zedong Wang', 'Siyuan Li'] | 2022-11-07 | null | null | null | null | ['3d-human-pose-estimation', 'video-prediction'] | ['computer-vision', 'computer-vision'] | [-1.47645220e-01 -3.01566198e-02 8.24395642e-02 -2.57202178e-01
-2.48224348e-01 -6.50987625e-01 5.51803172e-01 -3.75440717e-01
-8.75989199e-01 3.87281388e-01 -3.45397145e-01 -3.33508462e-01
-1.28487155e-01 -4.85232323e-01 -7.60231674e-01 -5.17366052e-01
-2.70318627e-01 4.32052195e-01 5.31430602e-01 -4.93844718e-01
1.40187442e-01 1.24540828e-01 -1.33587885e+00 -4.62718979e-02
6.55683935e-01 1.29570460e+00 5.41637301e-01 9.10837650e-01
1.18921638e-01 7.06147373e-01 -4.43145961e-01 -8.38454902e-01
6.48881137e-01 -8.88811871e-02 -6.78187132e-01 -1.18877828e-01
7.24177957e-01 -3.42503071e-01 -7.25343347e-01 1.18059635e+00
6.06840909e-01 -7.95555674e-03 3.98024708e-01 -1.44478738e+00
-5.40273488e-01 6.62258387e-01 -9.29911673e-01 5.73454261e-01
-2.01736957e-01 5.55782676e-01 1.10517871e+00 -7.73911178e-01
5.36580861e-01 1.25966859e+00 7.26421833e-01 7.22499192e-01
-1.06413019e+00 -8.72780502e-01 3.63603145e-01 3.31998497e-01
-1.34090340e+00 -2.03207418e-01 5.00281274e-01 -1.96114674e-01
1.14393187e+00 5.43652289e-02 1.00347400e+00 1.06797123e+00
4.44503665e-01 1.03136158e+00 1.04799736e+00 8.05211961e-02
1.50051981e-01 -4.43726152e-01 8.51496384e-02 7.63123691e-01
2.84719855e-01 2.08000898e-01 -5.80220103e-01 3.26488405e-01
1.01540279e+00 -1.98473155e-01 -1.45873008e-02 -3.35009545e-01
-8.96782815e-01 8.40617597e-01 6.84275270e-01 4.98089045e-02
-1.63948193e-01 9.18759584e-01 4.85526085e-01 1.77700624e-01
3.26804280e-01 4.01225477e-01 -5.98747551e-01 -4.01786685e-01
-8.22445869e-01 5.77499986e-01 5.26755333e-01 1.24329114e+00
4.42106158e-01 1.84565783e-01 1.28622055e-02 6.84373081e-01
4.05743152e-01 5.99194407e-01 1.94504634e-01 -1.07180822e+00
4.35197324e-01 3.81772280e-01 -2.16523036e-01 -7.94872463e-01
-7.16351092e-01 -6.96286857e-01 -8.88284385e-01 3.16153646e-01
5.03216803e-01 -2.43688196e-01 -1.41275024e+00 1.73446083e+00
1.59002155e-01 2.80412644e-01 -6.52031749e-02 1.13829744e+00
9.60410833e-01 3.91470611e-01 2.45198667e-01 3.11438292e-01
1.51626158e+00 -1.04635346e+00 -3.38394910e-01 -5.83518803e-01
3.65903169e-01 -6.25340939e-01 6.47421360e-01 5.25269687e-01
-1.29760861e+00 -6.61819279e-01 -1.03354526e+00 -6.76661655e-02
-3.65531296e-01 -6.57937974e-02 1.01017284e+00 6.93037152e-01
-1.62947214e+00 5.11686504e-01 -8.84295642e-01 -2.65715778e-01
8.96478891e-01 6.30846620e-01 -8.86274129e-02 -7.77419806e-02
-1.00626636e+00 7.32363224e-01 3.43826771e-01 2.20063671e-01
-1.06715631e+00 -7.38578320e-01 -6.91907465e-01 -1.23325966e-01
6.61470532e-01 -7.90553093e-01 1.39441085e+00 -7.68285871e-01
-1.35653210e+00 7.66853869e-01 3.58750582e-01 -8.82213593e-01
7.46196330e-01 -4.41836119e-01 2.95320875e-03 1.47925720e-01
2.19386891e-01 1.44023776e+00 7.41022050e-01 -8.27852309e-01
-8.02707434e-01 -4.79779840e-01 3.62305373e-01 4.68279511e-01
4.37000394e-02 1.56523404e-03 -1.05227435e+00 -6.01358771e-01
-3.47146653e-02 -1.17016494e+00 -7.46986032e-01 1.19572237e-01
-3.83104265e-01 -2.36800969e-01 7.20851302e-01 -2.41918951e-01
7.44006753e-01 -1.93229771e+00 1.96629047e-01 -5.89929372e-02
8.55037391e-01 5.47131896e-01 -3.13480675e-01 -2.25013886e-02
1.57875344e-01 5.63897379e-02 5.61864767e-03 -5.67662656e-01
8.55236202e-02 2.51009315e-01 1.61350563e-01 4.52810138e-01
1.32744268e-01 1.49443698e+00 -8.16188335e-01 -4.53322202e-01
5.02453327e-01 3.38724941e-01 -7.94767678e-01 -2.79816002e-01
-1.95570111e-01 2.49222934e-01 -5.28798938e-01 5.60806513e-01
8.21029425e-01 -3.37142825e-01 6.67895451e-02 -8.80118087e-02
-2.29478881e-01 -2.12102085e-01 -1.08917093e+00 1.83406031e+00
6.80382084e-03 8.73138070e-01 6.14366122e-02 -9.82916772e-01
6.85263455e-01 -5.79217039e-02 4.88911659e-01 -1.05938613e+00
6.61783874e-01 9.19387713e-02 2.99114048e-01 -2.41778955e-01
7.32725978e-01 3.47236067e-01 -5.83215915e-02 -1.73520178e-01
4.66882437e-01 -1.66693479e-01 3.41465235e-01 3.18226308e-01
1.08347809e+00 1.78967286e-02 4.31157090e-02 -6.50563717e-01
3.18197086e-02 -4.72424813e-02 3.74843210e-01 1.02132869e+00
-5.37280262e-01 4.25416738e-01 5.58827043e-01 -5.09919167e-01
-9.43511128e-01 -1.01361513e+00 1.34066604e-02 1.18912315e+00
5.73363602e-01 -5.33324778e-01 -9.11036193e-01 -3.99630159e-01
-6.19923696e-02 2.39506975e-01 -7.64765739e-01 -3.72273428e-03
-6.25782192e-01 -7.82077253e-01 9.36399043e-01 8.75324070e-01
1.08228314e+00 -9.67119515e-01 -1.04621458e+00 2.22865790e-01
9.31754038e-02 -1.36618876e+00 -3.52356881e-01 4.18050557e-01
-6.92317128e-01 -1.19249034e+00 -7.28028893e-01 -7.81213224e-01
9.32000056e-02 2.59701431e-01 1.22618973e+00 -1.33483544e-01
-5.13703883e-01 2.66863853e-01 -3.94081503e-01 -5.34370363e-01
2.47062609e-01 2.14918435e-01 -2.49796733e-02 -4.00523335e-01
2.39156783e-01 -5.23138046e-01 -1.04391217e+00 3.48304421e-01
-5.93748868e-01 2.76250631e-01 6.40345454e-01 7.68332839e-01
4.92272198e-01 -2.68366635e-01 1.74263075e-01 -4.14145350e-01
2.50452727e-01 -1.97896659e-01 -7.73566425e-01 -7.77381957e-02
-5.11805654e-01 -1.14439301e-01 3.99878770e-01 -4.23781455e-01
-7.59258986e-01 -8.56800377e-02 -4.14179973e-02 -7.50663042e-01
-6.98109940e-02 2.66724914e-01 3.65826897e-02 -3.33724707e-01
6.34300828e-01 2.74850249e-01 -1.48546934e-01 -1.66538328e-01
5.93144417e-01 1.34005859e-01 6.47452474e-01 -4.66651231e-01
6.24040246e-01 4.99334663e-01 5.50348386e-02 -6.82398617e-01
-5.39914370e-01 -3.98716956e-01 -3.01283777e-01 -4.73481685e-01
1.25916815e+00 -9.51768994e-01 -1.01429415e+00 1.06751263e+00
-1.11689270e+00 -7.91172624e-01 -1.85514644e-01 3.26117158e-01
-5.91428757e-01 3.82455513e-02 -7.21279323e-01 -5.31821191e-01
-3.83658856e-01 -1.57438612e+00 1.00933206e+00 4.74058479e-01
1.43011257e-01 -8.34664524e-01 -3.32419932e-01 4.16576415e-01
3.75790030e-01 5.74023686e-02 2.69941568e-01 -4.20146793e-01
-8.72582257e-01 1.80878237e-01 -5.75560153e-01 1.80591375e-01
-4.71827745e-01 -3.22858751e-01 -8.98892820e-01 -4.41839814e-01
-2.79741615e-01 -3.83059084e-01 9.93320465e-01 8.78915906e-01
1.22567129e+00 1.62278116e-01 -1.95427388e-01 1.17075622e+00
1.47126842e+00 4.22471464e-01 6.28399134e-01 4.19135749e-01
9.67464566e-01 1.55112907e-01 3.53871197e-01 5.30315936e-01
5.98432481e-01 6.56469762e-01 9.40877974e-01 -3.83580290e-02
-1.96017787e-01 4.34152856e-02 -7.69606903e-02 3.78277689e-01
-4.60190147e-01 -4.61500406e-01 -1.13726938e+00 4.50241357e-01
-1.91545701e+00 -6.86116338e-01 -1.26208529e-01 1.54448450e+00
3.34412158e-01 4.79034841e-01 1.40733495e-01 -4.30084616e-01
5.07179439e-01 4.11295831e-01 -9.44276392e-01 -1.49361059e-01
-1.25359952e-01 4.09736991e-01 1.20632851e+00 2.57392645e-01
-1.27100539e+00 1.44547355e+00 5.59478474e+00 1.12163806e+00
-1.04979610e+00 1.28921986e-01 9.27371502e-01 -2.29601517e-01
3.10467780e-01 -2.85814106e-01 -9.22666371e-01 2.33953759e-01
5.42812824e-01 6.13484867e-02 5.26224613e-01 8.46431851e-01
-1.85165666e-02 -2.14623034e-01 -8.83978784e-01 1.40991294e+00
-1.20056786e-01 -1.51708007e+00 -8.13172311e-02 2.30513155e-01
6.53994977e-01 8.01306069e-01 4.83958542e-01 4.28932309e-01
6.47848606e-01 -1.16046906e+00 1.04257786e+00 1.11254908e-01
8.69805038e-01 -7.66011655e-01 6.06902778e-01 9.13265198e-02
-1.49835348e+00 -1.82034373e-01 -3.72733533e-01 -1.64183393e-01
-1.96215808e-02 -1.45332798e-01 -3.59057069e-01 3.73047620e-01
1.30787146e+00 8.34756672e-01 -7.83732355e-01 1.03921962e+00
7.46198595e-02 3.02344054e-01 -4.40072060e-01 -1.74373999e-01
8.77072930e-01 -7.84250572e-02 5.51785171e-01 1.13147557e+00
-2.16976013e-02 3.81585121e-01 2.10320249e-01 6.24764383e-01
-1.36675999e-01 -2.92020231e-01 -3.95767361e-01 3.91208410e-01
3.55809927e-01 1.24734449e+00 -1.22363317e+00 -2.75428832e-01
-3.22062045e-01 9.34642911e-01 2.28264436e-01 3.88740689e-01
-1.25458610e+00 -3.58741507e-02 9.97195542e-01 -2.04451442e-01
7.23687828e-01 -2.78664052e-01 -3.96326810e-01 -9.34649050e-01
-7.29449168e-02 -8.51225257e-01 1.08686678e-01 -7.05996573e-01
-9.30958092e-01 7.83790112e-01 1.26053616e-01 -9.33724880e-01
-2.16587894e-02 -1.04807532e+00 -5.30156851e-01 4.31182981e-01
-1.24702322e+00 -1.18060982e+00 -2.95190811e-01 8.09549272e-01
7.67312765e-01 -2.35363871e-01 3.18518072e-01 2.97291040e-01
-4.82937187e-01 9.34865177e-01 -8.66480637e-03 4.16093647e-01
2.10894808e-01 -1.38502145e+00 1.03822446e+00 8.18578422e-01
1.60793558e-01 2.35096589e-01 6.73656285e-01 -4.15071577e-01
-1.38136995e+00 -1.08678436e+00 1.83325306e-01 -4.04533297e-01
7.70000875e-01 -4.73343223e-01 -3.44661623e-01 5.16631842e-01
3.43069345e-01 2.30926409e-01 5.95458783e-02 -4.79652882e-02
-3.49104226e-01 -1.52005060e-02 -8.60635698e-01 1.03860986e+00
1.56546199e+00 -2.04033285e-01 -2.08257176e-02 5.41153178e-02
7.88407028e-01 -7.54647493e-01 -5.13455510e-01 3.25916886e-01
5.91692090e-01 -9.78821337e-01 1.08391070e+00 -3.57432812e-01
3.50378424e-01 -3.33172046e-02 -3.52863610e-01 -1.05174482e+00
-3.54895324e-01 -6.91155791e-01 -1.32984286e-02 5.36545753e-01
2.40080267e-01 -5.47907710e-01 1.06868386e+00 4.64936048e-01
-3.68058681e-01 -1.15744603e+00 -1.15619028e+00 -8.07230234e-01
1.64610207e-01 -7.91512847e-01 3.69873673e-01 6.03378952e-01
-4.56535339e-01 2.75607079e-01 -3.76702785e-01 5.00512719e-02
7.97499001e-01 -3.56888235e-01 7.35532641e-01 -9.78026271e-01
-4.19277012e-01 -8.43143404e-01 -8.90734851e-01 -1.46701348e+00
-2.31121510e-01 -6.46652758e-01 -5.43260425e-02 -1.30999398e+00
1.93194762e-01 -2.89998472e-01 -2.60343879e-01 4.61364388e-01
6.88484982e-02 7.15148032e-01 7.13524342e-01 -8.62232819e-02
-1.06448638e+00 4.04640138e-01 1.46160150e+00 -1.20878473e-01
-4.42507230e-02 -2.43660629e-01 -7.17002630e-01 9.32576060e-01
1.04152775e+00 -1.73803642e-01 -6.32592440e-01 -7.73849249e-01
1.26420155e-01 2.98720784e-02 9.03689146e-01 -1.27104390e+00
3.33993167e-01 3.76134105e-02 3.71002793e-01 -6.06290698e-01
6.09761834e-01 -4.93003577e-01 2.30184812e-02 5.47752261e-01
-1.94056064e-01 2.97411352e-01 4.52366173e-01 5.90124488e-01
1.24584645e-01 1.88613489e-01 8.47701192e-01 -2.95865715e-01
-1.41391158e+00 8.10952425e-01 -2.03769743e-01 3.79414678e-01
1.04476798e+00 -4.92168546e-01 -5.21152139e-01 -1.46078214e-01
-6.98079944e-01 5.13788223e-01 1.30584016e-01 6.01113498e-01
6.87081039e-01 -9.12908912e-01 -6.23805702e-01 -1.75142512e-02
6.17881212e-03 2.05162525e-01 5.33593416e-01 7.34997332e-01
-7.34814405e-01 3.30480576e-01 -3.95236969e-01 -8.87275636e-01
-1.20279944e+00 2.06449896e-01 5.93167543e-01 -2.12959617e-01
-6.25206172e-01 1.32468629e+00 6.69647634e-01 -5.45002222e-02
1.71879336e-01 -3.81334543e-01 -1.27752304e-01 -8.45264047e-02
3.14807117e-01 2.68349856e-01 -1.85371131e-01 -5.40754855e-01
-5.40357053e-01 5.54676056e-01 -3.67474318e-01 -2.19436228e-01
1.27671683e+00 -1.12705313e-01 2.84306139e-01 2.30745859e-02
1.00605679e+00 -7.78903544e-01 -1.91195023e+00 -1.30520716e-01
-2.08150104e-01 -2.51049370e-01 2.31283128e-01 -7.68640161e-01
-1.55567932e+00 7.28391886e-01 8.53393018e-01 -8.80646259e-02
1.08407152e+00 3.43540400e-01 6.30651116e-01 3.75907332e-01
6.87291861e-01 -1.19328022e+00 2.93199569e-01 7.93882847e-01
5.75414240e-01 -1.26214099e+00 -2.12422192e-01 -2.14860350e-01
-7.23357260e-01 6.72341764e-01 1.14532506e+00 -4.59711313e-01
7.30797052e-01 2.51532733e-01 6.50067851e-02 -5.78869104e-01
-7.25773752e-01 -4.42566663e-01 1.03211857e-01 6.29456699e-01
1.29598334e-01 2.59330750e-01 -1.30929649e-01 4.16310787e-01
-3.79781008e-01 -2.21650094e-01 1.95928514e-01 6.71262443e-01
-2.02668369e-01 -6.67115629e-01 -2.28593331e-02 5.88953137e-01
-4.82144415e-01 -3.49222898e-01 -7.47492164e-02 1.14040625e+00
4.13148820e-01 7.70229697e-01 1.61743909e-01 -5.98650813e-01
2.22008377e-01 -4.85383809e-01 5.53565800e-01 -2.77200997e-01
-6.01693153e-01 1.77833587e-01 -1.11927137e-01 -9.17159200e-01
-3.83452743e-01 -6.79210067e-01 -1.03152692e+00 -6.19820654e-01
-3.23742092e-01 -2.74937809e-01 5.18855155e-01 9.11802292e-01
2.35618174e-01 7.08903193e-01 8.17742869e-02 -1.15705526e+00
-4.64295596e-01 -8.19615304e-01 -6.15078986e-01 -3.82719114e-02
2.48130616e-02 -7.63449430e-01 1.77787989e-01 -2.37626910e-01] | [9.400206565856934, 1.1540828943252563] |
334085a0-d276-4caa-8f5f-ea6f7497223d | coordinating-flexible-ramping-products-with | 2208.00036 | null | https://arxiv.org/abs/2208.00036v1 | https://arxiv.org/pdf/2208.00036v1.pdf | Coordinating Flexible Ramping Products with Dynamics of the Natural Gas Network | In electricity networks with high penetration levels of renewable resources, Flexible Ramping Products (FRPs) are among the utilized measures for dealing with the potential fluctuations in the net demand. This paper investigates the impacts of FRPs on the operation of interdependent electricity and natural gas networks. To accurately model and reflect the effects of variations in the natural gas fuel demand on the natural gas network, a dynamic Optimal Gas Flow (OGF) formulation is utilized. The non-convex dynamic model of the natural gas system is represented in a convex form via a tight relaxation scheme. An improved distributed optimization method is proposed to solve the coordinated operation problem in a privacy-preserving manner, where the two infrastructures only share limited information. We introduce the Inexact Varying Alternating Direction Method of Multipliers (IV-ADMM) and show that compared with the classic ADMM, it converges considerably faster and in fewer iterations. Through a comparison of day-ahead and real-time operation planning results, it is concluded that without accounting for natural gas network dynamics, the FRP model is not a trustworthy tool in day-ahead planning against uncertainties. | ['Saeed D. Manshadi', 'Reza Bayani'] | 2022-07-29 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-3.25183243e-01 1.04835993e-02 7.46426592e-03 -1.29007012e-01
-1.14213906e-01 -8.69411945e-01 3.95253092e-01 -5.72512411e-02
-5.47422469e-03 1.18470633e+00 -9.45588872e-02 -3.75598490e-01
-6.65686905e-01 -1.21862924e+00 -4.30189192e-01 -1.01043403e+00
-4.08330888e-01 7.97784925e-01 -8.08423102e-01 -2.60192901e-01
1.66924819e-02 7.15864599e-01 -3.95956516e-01 -3.60958129e-01
9.90202248e-01 9.29043889e-01 9.79042128e-02 8.61248448e-02
1.41974121e-01 3.95150721e-01 -4.21864808e-01 -2.43228272e-01
1.23094249e+00 -1.88572444e-02 -6.18185401e-01 1.84676245e-01
-6.11821711e-01 -7.07733989e-01 -4.92713034e-01 1.44828773e+00
3.93620223e-01 6.09931886e-01 2.51971632e-01 -1.82263291e+00
-2.76844829e-01 9.17050242e-01 -9.10117686e-01 1.03726618e-01
2.05105334e-03 1.04247056e-01 9.62019086e-01 -6.77602470e-01
2.75005519e-01 9.63236153e-01 3.74767572e-01 -1.19273901e-01
-1.26032794e+00 -4.23824519e-01 3.87275696e-01 -3.08304317e-02
-1.60827863e+00 1.12136640e-01 4.76962298e-01 -2.32231066e-01
1.23386192e+00 7.35549629e-01 9.10701156e-01 2.23273356e-02
3.66721004e-01 3.11018288e-01 8.87962878e-01 -7.62828365e-02
4.14481968e-01 4.50083949e-02 -6.13687634e-01 8.15316886e-02
4.43458080e-01 2.46155217e-01 -1.21468477e-01 -3.99845541e-01
3.79170150e-01 6.14749119e-02 -5.65327466e-01 -1.73058406e-01
-8.55317533e-01 8.44796598e-01 4.47180182e-01 3.53295267e-01
-8.47789884e-01 -2.31810417e-02 3.11343223e-01 3.29572886e-01
6.92046463e-01 3.61361772e-01 -5.78958511e-01 7.60461614e-02
-1.30660975e+00 2.95875013e-01 1.05371380e+00 1.19306636e+00
3.80520344e-01 7.43303359e-01 -3.54665145e-02 1.50177097e-02
2.41076320e-01 6.62400365e-01 -1.36534020e-01 -8.81211758e-01
8.18526924e-01 1.25095779e-02 5.52389026e-01 -1.22277570e+00
-3.57638210e-01 -6.48553491e-01 -1.28962767e+00 2.83462346e-01
2.69216329e-01 -7.13838577e-01 -2.22443432e-01 1.33027565e+00
3.01420540e-01 -3.83421220e-02 -2.37885788e-01 7.10878074e-01
-4.08585578e-01 1.07590985e+00 -5.23676634e-01 -9.24676597e-01
7.54783809e-01 -9.02460456e-01 -1.14146495e+00 2.01418325e-01
3.28332007e-01 -3.97163659e-01 -2.69392669e-01 4.63415056e-01
-1.48813176e+00 9.86958668e-02 -7.08637059e-01 5.07398546e-01
-6.97867513e-01 -2.34752655e-01 3.46940428e-01 5.61569631e-01
-1.16820776e+00 8.52142632e-01 -6.45703256e-01 1.75085738e-01
3.82415354e-01 4.34675157e-01 -2.18413264e-01 -1.42366812e-01
-1.26711655e+00 1.38780880e+00 5.94787717e-01 1.04394019e+00
-7.39131629e-01 -1.03480160e+00 -5.58725178e-01 4.65825737e-01
6.19834423e-01 -5.79979002e-01 6.74801290e-01 -3.38638157e-01
-1.13984978e+00 -1.67697832e-01 2.20275089e-01 -7.58519769e-01
1.14930403e+00 1.94805339e-01 -4.04714108e-01 5.85110905e-03
-2.59558946e-01 -1.99179873e-01 6.96421325e-01 -1.06481934e+00
-3.94190460e-01 -3.20436478e-01 -9.53881368e-02 3.84046704e-01
8.56241137e-02 -2.51433372e-01 3.39575768e-01 -7.27357030e-01
-2.50131458e-01 -8.45175326e-01 -6.73157513e-01 -1.05623327e-01
-5.33604980e-01 3.20973456e-01 9.60937083e-01 -1.21300125e+00
1.22041214e+00 -1.77112901e+00 2.15225533e-01 1.06374109e+00
-1.50951758e-01 2.07937926e-01 2.25819513e-01 1.00112081e+00
-3.09652597e-01 2.63796866e-01 -5.19959629e-01 -3.68902147e-01
4.77207720e-01 7.38147259e-01 -3.78313869e-01 6.98738754e-01
-3.93644661e-01 7.83987880e-01 -1.05077410e+00 4.22734290e-01
5.19480288e-01 2.07204342e-01 8.91376361e-02 -2.93266565e-01
-1.31107092e-01 4.43378896e-01 -5.16961873e-01 5.68434417e-01
1.39730430e+00 2.37455234e-01 4.38272983e-01 -1.56112164e-01
-3.03457081e-01 -5.40460765e-01 -1.37701929e+00 1.56649876e+00
-6.06539607e-01 3.00143391e-01 7.27956653e-01 -1.27208638e+00
7.98822820e-01 5.42221189e-01 8.68480802e-01 -5.66668451e-01
-1.84704214e-01 2.64658481e-01 -1.56452626e-01 -9.68036652e-02
5.20331860e-01 -1.67195797e-01 3.01284790e-01 6.10723078e-01
-4.68809903e-01 -7.18460917e-01 1.88972950e-01 9.98059139e-02
7.61545837e-01 -4.38706130e-01 3.04493487e-01 -6.28791273e-01
5.38163781e-01 -2.44359359e-01 7.50843406e-01 3.04348022e-01
2.78891660e-02 1.08918861e-01 4.54046398e-01 -4.63267803e-01
-8.62808883e-01 -6.90016627e-01 -1.72851115e-01 1.00814596e-01
-2.69621938e-01 5.20707965e-02 -4.07095760e-01 -5.72062790e-01
5.08645177e-01 1.36754382e+00 -3.21240157e-01 2.63687193e-01
-2.10655764e-01 -1.22035134e+00 -2.25311115e-01 1.45239249e-01
5.63920379e-01 -2.40008727e-01 -6.26014024e-02 4.77218717e-01
1.04213513e-01 -9.19585884e-01 -6.27372921e-01 1.10567976e-02
-6.58121407e-01 -8.70842040e-01 -6.87988818e-01 1.32627701e-02
1.21964490e+00 1.30184859e-01 1.08161843e+00 -1.20176122e-01
-2.11931467e-01 4.56508636e-01 5.55175878e-02 -3.73578250e-01
-6.73452616e-02 -2.63722569e-01 2.18447134e-01 7.75949582e-02
-3.15880746e-01 -7.56263435e-01 -5.29183149e-01 2.33162433e-01
-8.66809011e-01 -1.97683468e-01 3.77006754e-02 7.72458196e-01
6.17622912e-01 1.38133299e+00 7.70693481e-01 -9.56213653e-01
6.48670316e-01 -8.50451469e-01 -1.33699203e+00 4.25574034e-01
-1.23881471e+00 -2.41740033e-01 7.63124764e-01 2.96665251e-01
-1.19375074e+00 9.10149440e-02 5.13792992e-01 -3.81796330e-01
5.67609012e-01 9.22559619e-01 -1.61611930e-01 -4.97422516e-01
-3.88305634e-01 2.93107182e-01 3.99687365e-02 -2.79772937e-01
3.68954748e-01 2.18496203e-01 1.90503046e-01 -4.04492289e-01
1.57430863e+00 6.75405324e-01 6.16550624e-01 -4.57871705e-01
-2.17953295e-01 -1.42017290e-01 -4.28741604e-01 -2.47525454e-01
2.00121537e-01 -1.04447949e+00 -7.59234428e-01 5.82133710e-01
-9.02589321e-01 -6.66438192e-02 -7.37326503e-01 4.96792704e-01
-3.46078873e-01 3.67361218e-01 -2.29543358e-01 -1.18909717e+00
-5.26459396e-01 -7.91444004e-01 1.23491153e-01 -2.76627839e-02
3.88158143e-01 -1.38305628e+00 1.46062061e-01 2.54727900e-01
1.05522919e+00 8.44425857e-01 5.94717026e-01 -2.82492101e-01
-8.07675064e-01 -4.47070450e-01 6.31676521e-03 5.93639374e-01
4.80847448e-01 1.65592074e-01 -4.80516762e-01 -1.09240484e+00
3.53956670e-01 4.42121089e-01 -2.60900818e-02 3.12271923e-01
1.16919839e+00 -9.64689851e-01 -6.85925260e-02 6.64516866e-01
2.07369304e+00 4.50988829e-01 6.40969694e-01 1.71111092e-01
4.61564690e-01 5.97768366e-01 8.17261159e-01 1.27048659e+00
3.90237778e-01 3.85013103e-01 1.13135529e+00 9.26034339e-03
1.13507617e+00 8.57110396e-02 1.52901977e-01 5.73648155e-01
-2.44270608e-01 -5.65154374e-01 -6.33544445e-01 6.83350086e-01
-1.92133391e+00 -1.07442939e+00 -1.27159670e-01 2.36194062e+00
1.63209319e-01 -2.52936959e-01 -7.96412006e-02 -6.74687549e-02
7.13387191e-01 3.01803678e-01 -7.48739660e-01 -6.83664858e-01
-3.29593450e-01 7.59189799e-02 1.32603669e+00 3.11837733e-01
-5.35623431e-01 -2.25547686e-01 6.04077387e+00 6.99190676e-01
-6.88557029e-01 -2.88620740e-02 7.20061183e-01 -2.02304676e-01
-6.42150044e-01 1.60550028e-01 -4.88314033e-02 7.72019804e-01
1.17242050e+00 -1.06325412e+00 1.14060867e+00 4.37022358e-01
1.09336782e+00 -5.35856858e-02 -7.12188244e-01 4.97723103e-01
-5.55325329e-01 -1.36817145e+00 -5.15549839e-01 5.75576663e-01
1.34353578e+00 3.04450125e-01 -8.01613703e-02 -2.77145773e-01
5.00569880e-01 -9.92927015e-01 4.84563351e-01 6.76017463e-01
2.92187452e-01 -1.28382301e+00 1.06346834e+00 2.23945901e-01
-1.56108606e+00 -3.28318655e-01 -1.99261829e-01 -5.68696558e-02
8.01316798e-01 9.59678650e-01 -4.68170106e-01 1.47715056e+00
5.16641021e-01 6.47806942e-01 4.16908920e-01 9.99726176e-01
-1.21277198e-01 1.68624789e-01 -6.89156830e-01 6.01581872e-01
3.27967912e-01 -9.90745246e-01 5.20620406e-01 4.86081898e-01
5.86433530e-01 1.77858710e-01 2.36648396e-01 1.01710856e+00
1.93288755e-02 -2.13071749e-01 -6.39754772e-01 -2.41798371e-01
7.46943593e-01 1.24537814e+00 -2.98437923e-01 2.36512348e-01
-6.24292195e-01 6.04568064e-01 -4.38224256e-01 5.58250844e-01
-6.85611308e-01 -1.62962839e-01 7.12318718e-01 -2.29025245e-01
2.63217926e-01 -4.22589749e-01 -2.40068302e-01 -9.60727632e-01
2.16234609e-01 -4.71610010e-01 4.31452751e-01 -6.33579731e-01
-1.46982205e+00 1.61151052e-01 1.26788719e-02 -1.27347338e+00
-3.90714705e-01 -9.69488081e-03 -1.04978073e+00 1.57714772e+00
-2.05301118e+00 -8.16676497e-01 2.48323813e-01 7.08481908e-01
1.21643558e-01 -1.61513656e-01 8.87984931e-01 2.91510075e-01
-9.76477206e-01 -7.56916450e-03 1.02696168e+00 -2.48617381e-01
-3.02275807e-01 -1.50596035e+00 3.73305172e-01 1.39600503e+00
-3.49306136e-01 1.11096330e-01 6.51121795e-01 -5.20566583e-01
-1.70241225e+00 -1.20838273e+00 5.51447332e-01 4.98653203e-01
8.95950437e-01 2.00974811e-02 -7.59424210e-01 7.75655210e-01
6.86499178e-01 3.22745055e-01 3.95471364e-01 -7.83975303e-01
4.16765839e-01 -3.45493317e-01 -2.03318787e+00 -1.78913116e-01
3.59716892e-01 -3.90776694e-01 2.72218529e-02 8.90458167e-01
2.81275511e-01 -4.69336718e-01 -1.22646332e+00 -9.91361141e-02
-4.92871292e-02 -2.80968964e-01 8.36550772e-01 -2.68764585e-01
-5.31790376e-01 -5.86083949e-01 -1.80917457e-01 -1.89415574e+00
-3.31872314e-01 -1.46249521e+00 -3.47489566e-01 1.36708653e+00
3.79055083e-01 -1.31034446e+00 6.95899010e-01 1.49273276e+00
2.05465317e-01 -4.12208080e-01 -1.42984939e+00 -1.01290131e+00
1.50677577e-01 1.12565294e-01 1.12828803e+00 1.40470433e+00
1.37453958e-01 -7.70393372e-01 -6.23579323e-01 8.04854274e-01
8.94838214e-01 1.61793202e-01 1.33592203e-01 -6.02881372e-01
-1.64762646e-01 6.72015697e-02 -1.51916891e-02 -2.06100151e-01
9.14270058e-02 -7.94000506e-01 -3.41196269e-01 -1.47589386e+00
-4.62195843e-01 -5.26643991e-01 -5.28072536e-01 4.66873318e-01
5.22057950e-01 -4.64289844e-01 4.71228659e-01 -1.03094779e-01
3.68477017e-01 8.91376078e-01 9.69704926e-01 -4.15575445e-01
5.17051145e-02 4.48430449e-01 -3.02611858e-01 3.62961859e-01
9.39315319e-01 -4.13538128e-01 -7.13886082e-01 -6.61700130e-01
3.16892684e-01 5.86892247e-01 -1.08970389e-01 -4.07228649e-01
3.14308167e-01 -7.20117509e-01 -1.69367477e-01 -8.81215394e-01
2.17105746e-01 -1.86902618e+00 1.15649879e+00 7.51784801e-01
1.88268974e-01 4.98876780e-01 1.51619464e-01 6.96890950e-01
-2.71017969e-01 -3.95842075e-01 4.04746979e-01 -4.22252595e-01
-3.02955002e-01 7.91068614e-01 -3.30196321e-01 -2.34869719e-01
1.29056668e+00 1.06867686e-01 -5.00785828e-01 -5.94920814e-01
-7.87325978e-01 1.24255288e+00 2.35055372e-01 -6.65694326e-02
3.69698107e-01 -1.07956040e+00 -8.59623909e-01 -1.90800875e-01
-9.75444198e-01 2.91798025e-01 6.21412694e-01 7.23935723e-01
-6.46301627e-01 4.48537767e-01 -6.52645156e-03 8.19229037e-02
-3.86032313e-01 5.71911395e-01 5.75011373e-01 -6.01169467e-01
-2.81513304e-01 6.22335553e-01 -4.44694370e-01 -2.42020771e-01
-1.17910422e-01 -2.07627848e-01 3.08278948e-01 4.70629036e-01
4.03234273e-01 1.15974712e+00 2.86869854e-01 -3.35367531e-01
-2.48677880e-01 5.96690699e-02 1.54380932e-01 1.88001886e-01
1.94118440e+00 -6.95748031e-01 -6.06587112e-01 -8.68921280e-02
9.58159745e-01 -1.84401885e-01 -1.14778209e+00 -6.23151287e-02
-2.25536555e-01 -9.56006706e-01 5.60028017e-01 -8.53434503e-01
-2.14844489e+00 9.25375596e-02 1.25243083e-01 7.31843531e-01
1.37404168e+00 -1.17221951e+00 7.16980815e-01 1.57886043e-01
7.19640195e-01 -1.30149817e+00 -1.13843572e+00 5.36327958e-02
1.10358274e+00 -7.44250774e-01 5.87915003e-01 -3.73443276e-01
-4.98251975e-01 1.16293931e+00 1.34508125e-03 2.09728070e-02
1.00138354e+00 5.30943334e-01 -6.64148152e-01 1.67196974e-01
-5.43701351e-01 4.83753532e-01 -3.75202417e-01 4.25305039e-01
-5.93901515e-01 5.05179524e-01 -2.46551752e-01 3.38329673e-02
9.61738359e-03 3.14320140e-02 1.02787137e+00 1.18656147e+00
5.21820664e-01 -9.54640031e-01 -2.76911676e-01 4.68125761e-01
-3.81680101e-01 -9.72722396e-02 2.82720447e-01 5.64654589e-01
-2.59062313e-02 8.37973177e-01 2.97737624e-02 3.08860034e-01
3.93040180e-01 -2.51324236e-01 -1.07589111e-01 -2.10139185e-01
-6.17743611e-01 -3.13842833e-01 -7.28680640e-02 -5.47346711e-01
-4.53224391e-01 -1.07048213e+00 -9.81917322e-01 -9.27781105e-01
-6.24312401e-01 4.42457408e-01 9.24010575e-01 9.46226656e-01
1.74254790e-01 4.15824980e-01 1.56729925e+00 -8.51223350e-01
-1.13234746e+00 -2.19408855e-01 -1.44419897e+00 -3.81665915e-01
1.78925857e-01 -1.50954559e-01 -8.70014548e-01 -6.53617203e-01] | [5.660068035125732, 2.554194211959839] |
1b06585a-ce8f-4d8c-a407-a9f416e95960 | hybrid-decentralized-optimization-first-and | 2210.07703 | null | https://arxiv.org/abs/2210.07703v1 | https://arxiv.org/pdf/2210.07703v1.pdf | Hybrid Decentralized Optimization: First- and Zeroth-Order Optimizers Can Be Jointly Leveraged For Faster Convergence | Distributed optimization has become one of the standard ways of speeding up machine learning training, and most of the research in the area focuses on distributed first-order, gradient-based methods. Yet, there are settings where some computationally-bounded nodes may not be able to implement first-order, gradient-based optimization, while they could still contribute to joint optimization tasks. In this paper, we initiate the study of hybrid decentralized optimization, studying settings where nodes with zeroth-order and first-order optimization capabilities co-exist in a distributed system, and attempt to jointly solve an optimization task over some data distribution. We essentially show that, under reasonable parameter settings, such a system can not only withstand noisier zeroth-order agents but can even benefit from integrating such agents into the optimization process, rather than ignoring their information. At the core of our approach is a new analysis of distributed optimization with noisy and possibly-biased gradient estimators, which may be of independent interest. Experimental results on standard optimization tasks confirm our analysis, showing that hybrid first-zeroth order optimization can be practical. | ['Dan Alistarh', 'Giorgi Nadiradze', 'Shayan Talaei'] | 2022-10-14 | null | null | null | null | ['distributed-optimization'] | ['methodology'] | [-3.84761244e-01 1.80692181e-01 -9.73784849e-02 -2.12919444e-01
-8.82534802e-01 -6.18184149e-01 1.97729185e-01 3.04177821e-01
-7.62915432e-01 1.15779865e+00 -2.90553957e-01 -3.52744192e-01
-4.07596201e-01 -5.13052940e-01 -9.27944660e-01 -1.13198984e+00
-4.40542579e-01 9.00387108e-01 -5.12416884e-02 -2.13932414e-02
8.79447237e-02 3.99631411e-01 -1.04275143e+00 -2.41789803e-01
9.83026624e-01 7.78721035e-01 1.76754802e-01 1.08184993e+00
3.32833052e-01 6.94008946e-01 -9.40373778e-01 -2.10051507e-01
5.01642704e-01 -4.00299758e-01 -5.78401446e-01 -4.46569398e-02
2.57558405e-01 -6.10486209e-01 1.79972202e-01 1.33403182e+00
7.60586143e-01 3.27366441e-01 3.25611442e-01 -1.54205477e+00
-2.31623963e-01 8.74335110e-01 -4.18044150e-01 9.69269499e-02
-3.78079489e-02 -6.85299635e-02 1.08461452e+00 -5.73296666e-01
5.00720263e-01 1.12193239e+00 7.05336750e-01 4.05244201e-01
-1.09073102e+00 -3.32216680e-01 3.39474648e-01 -1.14248440e-01
-1.22974575e+00 -5.82819879e-01 7.08184838e-01 3.74372303e-03
7.67380774e-01 5.25870800e-01 5.22020280e-01 4.73292232e-01
1.38048157e-01 1.00519300e+00 9.31443870e-01 -3.77615541e-01
6.20188832e-01 4.22176421e-01 -1.26828015e-01 6.91165149e-01
6.98981643e-01 -1.65077209e-01 -6.12444699e-01 -7.82414198e-01
3.91940385e-01 -8.57444480e-03 -5.06993771e-01 -4.96909261e-01
-1.08911824e+00 1.11673605e+00 5.33463359e-01 2.95951188e-01
-7.68454373e-01 2.40558848e-01 1.83225751e-01 6.07840896e-01
1.08054912e+00 4.04137790e-01 -6.42823219e-01 -1.34966910e-01
-8.80633175e-01 4.22919571e-01 1.53171420e+00 6.54347599e-01
9.91871834e-01 3.38687152e-02 -3.22630927e-02 5.13538361e-01
4.59956735e-01 5.89755535e-01 3.93168777e-01 -1.10241902e+00
6.40046000e-01 2.28009492e-01 4.67438310e-01 -9.44089711e-01
-4.51714307e-01 -6.96276784e-01 -8.63865376e-01 2.72200197e-01
5.15866637e-01 -9.26341593e-01 -8.67936760e-02 1.62865961e+00
8.69428098e-01 -1.01140499e-01 3.44343483e-02 1.23738778e+00
-1.54149503e-01 5.53959072e-01 -6.09657407e-01 -5.28074026e-01
8.62124920e-01 -1.10266447e+00 -4.58982050e-01 -1.63543835e-01
9.43136930e-01 -4.89664197e-01 4.28610504e-01 3.19247574e-01
-1.20234668e+00 3.20390731e-01 -9.24849987e-01 3.44063938e-01
-2.07089875e-02 -5.13273571e-03 6.33406699e-01 8.18782270e-01
-1.44556010e+00 7.34292746e-01 -9.45334613e-01 -2.41977453e-01
3.90534937e-01 5.22675335e-01 -1.87616229e-01 5.04218191e-02
-7.87533581e-01 7.50620723e-01 -8.42851177e-02 4.65944171e-01
-1.09296453e+00 -7.38479078e-01 -3.41315836e-01 1.12031497e-01
7.05573320e-01 -7.73457229e-01 1.32706559e+00 -1.14053130e+00
-1.61258030e+00 3.11772972e-01 -3.37128043e-01 -5.12547553e-01
8.89632285e-01 -1.01551712e-01 4.95615840e-01 3.81857120e-02
1.94894612e-01 -8.34177658e-02 8.84012222e-01 -1.19498897e+00
-7.56804049e-01 -7.49372900e-01 2.09324792e-01 7.36880541e-01
-7.45786369e-01 -1.23169273e-01 1.92721024e-01 -8.32080096e-02
-1.84797227e-01 -1.11791289e+00 -6.80996954e-01 2.08468705e-01
-5.61040163e-01 -2.72011310e-01 4.98418838e-01 -4.23081279e-01
8.06174994e-01 -1.80647576e+00 2.33354256e-01 4.51157093e-01
6.73910677e-01 5.54164015e-02 9.10927504e-02 5.08784831e-01
5.64965189e-01 3.77050906e-01 -1.38856247e-01 -9.67871010e-01
2.60873258e-01 2.82013237e-01 1.13548204e-01 1.09142661e+00
-3.58902663e-01 5.50648332e-01 -1.01391578e+00 -2.91065156e-01
-2.24369958e-01 1.74841210e-01 -6.38185918e-01 9.37297717e-02
-1.46298498e-01 3.77001703e-01 -1.02049327e+00 4.68587548e-01
5.32426417e-01 -5.25650501e-01 4.62950081e-01 3.04444104e-01
8.47771093e-02 -1.30440757e-01 -1.52458930e+00 1.24513817e+00
-5.19644856e-01 6.33645713e-01 1.23120832e+00 -1.58817828e+00
1.77427992e-01 3.16154778e-01 7.92030871e-01 -1.56241924e-01
2.00005975e-02 4.41278309e-01 7.46200457e-02 -1.87725440e-01
4.04874831e-01 2.44295478e-01 2.96687901e-01 9.26050544e-01
-2.25097790e-01 1.32800832e-01 1.98246807e-01 2.16031015e-01
1.25699806e+00 -6.47113383e-01 8.44385773e-02 -4.84296054e-01
2.55385369e-01 -1.02645688e-01 4.25183803e-01 1.30729318e+00
-3.01317036e-01 9.35433507e-02 4.76887643e-01 -9.59227756e-02
-9.28244531e-01 -5.00188172e-01 1.12479113e-01 1.49773037e+00
5.23429960e-02 -2.66397148e-01 -7.62054443e-01 -6.48945391e-01
3.43858272e-01 1.16982065e-01 -4.13162887e-01 2.67033637e-01
-3.34222615e-01 -1.01628315e+00 3.92413169e-01 -1.34092927e-01
4.81658757e-01 -4.77335095e-01 -7.41679490e-01 3.84306908e-01
-9.93122458e-02 -6.80524468e-01 -5.45559704e-01 3.52524400e-01
-9.10979092e-01 -9.60755527e-01 -1.13070989e+00 -2.82416433e-01
1.02206194e+00 3.73817295e-01 1.15412903e+00 4.30021822e-01
-1.10424146e-01 7.73824811e-01 -1.32346198e-01 -5.49182653e-01
-3.14932495e-01 4.38710481e-01 2.12436065e-01 2.19597206e-01
-3.13461661e-01 -4.86865759e-01 -7.05374897e-01 2.11506724e-01
-7.56819308e-01 -5.35596013e-01 3.67316067e-01 8.58224452e-01
2.15071753e-01 7.91875124e-02 6.19011104e-01 -8.03081512e-01
1.08179569e+00 -7.38503158e-01 -8.74394178e-01 4.07813758e-01
-7.80977905e-01 1.98243573e-01 8.91988516e-01 -5.21462977e-01
-7.55959809e-01 -7.08201900e-02 3.52282345e-01 -3.53766173e-01
3.55533928e-01 8.29682589e-01 5.99708669e-02 -5.60472250e-01
7.81651497e-01 1.00432828e-01 4.01851803e-01 -3.52216631e-01
3.75340879e-01 6.91183209e-01 -2.85172790e-01 -8.38455498e-01
6.09122992e-01 7.08357871e-01 8.33586231e-02 -9.04913187e-01
-5.08720756e-01 -3.96963120e-01 -5.50082289e-02 -2.53670007e-01
9.66982841e-02 -8.05849493e-01 -9.92612004e-01 4.89068687e-01
-1.20118213e+00 -6.11585319e-01 -3.08535904e-01 3.54767323e-01
-2.48110458e-01 2.26297751e-01 -3.50949913e-01 -1.38031232e+00
-3.34055901e-01 -1.08572733e+00 8.45248342e-01 3.56832325e-01
2.52514631e-01 -1.34446681e+00 1.38363138e-01 1.18804693e-01
8.50836575e-01 -1.45988986e-01 1.10544540e-01 -1.14059603e+00
-6.89811587e-01 -3.48238647e-01 2.56312191e-01 2.26722851e-01
-1.42882556e-01 -1.42915308e-01 -6.09632790e-01 -6.81469560e-01
3.36167037e-01 -4.77833599e-01 5.31391203e-01 5.75618804e-01
7.85785317e-01 -8.40261459e-01 -4.80824530e-01 5.55191398e-01
1.32702386e+00 -3.83283317e-01 -2.00312957e-01 1.06038190e-01
5.23864985e-01 4.66531575e-01 3.68385762e-01 1.00589943e+00
6.25943124e-01 4.74557877e-01 5.49911320e-01 -1.16407657e-02
4.44852203e-01 1.10997371e-01 2.81553388e-01 5.84596097e-01
1.60573458e-03 -4.27051604e-01 -5.91240644e-01 6.07534170e-01
-2.30154276e+00 -6.39034033e-01 5.82085922e-03 2.34356308e+00
1.07355762e+00 -5.50796688e-01 1.58182636e-01 -2.84506828e-01
8.30132067e-01 7.16976747e-02 -9.56404269e-01 -1.35839015e-01
1.47746308e-02 -2.17531398e-01 9.84598041e-01 5.64661920e-01
-8.33717227e-01 3.50699723e-01 6.35986710e+00 7.95595050e-01
-1.02922678e+00 4.70626622e-01 7.62156308e-01 -3.99877071e-01
-1.68045983e-01 8.99247825e-02 -7.69694030e-01 5.30959487e-01
9.33072805e-01 -5.03178895e-01 1.00771439e+00 1.16014314e+00
2.01584220e-01 -4.75103796e-01 -1.21820450e+00 6.99485540e-01
1.52303511e-02 -1.09966373e+00 -7.01874197e-01 5.53330243e-01
1.04702377e+00 5.50146818e-01 -1.83006108e-01 1.07272705e-02
7.53556073e-01 -7.67239809e-01 4.72131103e-01 2.78466761e-01
-6.94417804e-02 -5.52204549e-01 7.45751679e-01 1.07219791e+00
-8.14019680e-01 -2.41297007e-01 -5.59362292e-01 -2.33093947e-01
1.21947706e-01 1.15667331e+00 -5.36575496e-01 5.00617504e-01
6.74607813e-01 2.62721390e-01 -2.40596130e-01 1.18522596e+00
2.86658462e-02 5.10358632e-01 -1.26394069e+00 -7.35112727e-01
7.67606571e-02 -2.96545327e-01 7.67534912e-01 5.96710145e-01
3.93085867e-01 -1.04821444e-01 4.75536287e-01 5.48414707e-01
-3.90276462e-01 1.61363646e-01 -5.23628354e-01 4.90367152e-02
6.31615043e-01 1.38670218e+00 -5.68559766e-01 -5.71042597e-01
-1.20942801e-01 9.75863218e-01 7.12741017e-01 5.72329879e-01
-7.07173228e-01 -3.25342476e-01 7.03587532e-01 -3.91555279e-01
1.87224597e-01 -2.73313791e-01 -2.12073416e-01 -1.52149832e+00
4.75828916e-01 -8.83399606e-01 2.68850088e-01 -1.08390905e-01
-1.47240567e+00 1.25491127e-01 -1.48925081e-01 -4.93133962e-01
-4.45352018e-01 -3.01512152e-01 -7.11524129e-01 8.43747020e-01
-1.45094991e+00 -7.33079493e-01 1.04389377e-01 6.70692980e-01
2.68273532e-01 -7.66855478e-02 5.29637635e-01 2.97720939e-01
-6.04378283e-01 5.26087761e-01 9.03168976e-01 -2.09390178e-01
5.36505520e-01 -1.39900351e+00 -5.43023407e-01 8.68118465e-01
2.74157107e-01 6.28031611e-01 7.77404487e-01 -4.89808589e-01
-1.91038966e+00 -7.34129846e-01 8.15011084e-01 -1.48981914e-01
8.74215543e-01 -3.75023156e-01 -5.55413604e-01 4.58323896e-01
8.57636258e-02 4.10469800e-01 4.03822333e-01 3.09849948e-01
3.50988090e-01 -3.67486268e-01 -1.27173960e+00 3.95783216e-01
7.46122479e-01 -3.01182300e-01 2.29603142e-01 1.10808468e+00
1.74783632e-01 -3.96997452e-01 -6.59933805e-01 -8.67245868e-02
2.76875377e-01 -7.56850421e-01 8.68548036e-01 -4.52389657e-01
-9.75143090e-02 -2.15156265e-02 -2.83210892e-02 -1.67419875e+00
3.79522830e-01 -1.27299035e+00 -3.82607520e-01 9.45448041e-01
5.10851979e-01 -1.29329956e+00 1.02909517e+00 9.75731611e-01
3.19240868e-01 -6.28934681e-01 -1.34640598e+00 -8.85460973e-01
2.71631807e-01 -1.14195593e-01 2.29596257e-01 6.67365730e-01
-3.36555615e-02 1.09016620e-01 -5.20450830e-01 3.97439599e-01
9.94529486e-01 -1.26596764e-02 1.07306743e+00 -9.37448382e-01
-6.75807595e-01 -4.82596844e-01 7.77815729e-02 -1.29533923e+00
2.67691463e-01 -6.98849618e-01 4.83887136e-01 -1.35938215e+00
-3.15450467e-02 -7.09656954e-01 -4.83585745e-02 4.16513264e-01
-1.68325394e-01 -6.26658797e-02 2.70724863e-01 4.35190052e-01
-9.03359652e-01 4.63934988e-01 9.75727975e-01 3.23121510e-02
-1.69348717e-01 3.47337395e-01 -6.74222350e-01 4.81999665e-01
6.63619280e-01 -8.20934296e-01 -5.31086028e-01 -5.74138284e-01
4.53012049e-01 2.32647896e-01 3.86857361e-01 -5.40137291e-01
9.51258719e-01 -1.62245110e-01 -2.68325638e-02 6.26690835e-02
1.68529332e-01 -9.27274823e-01 -2.00221583e-01 5.24203479e-01
-3.82816434e-01 -5.16197691e-03 -4.03081685e-01 6.65933967e-01
2.36314666e-02 -5.72967172e-01 4.07862931e-01 -2.03643680e-01
1.60786748e-01 3.24032336e-01 -5.62138200e-01 1.14651367e-01
1.06427312e+00 3.04070562e-01 -3.89430702e-01 -7.65804529e-01
-6.67131186e-01 6.95179939e-01 3.50740671e-01 -4.64734524e-01
2.74647087e-01 -8.01917315e-01 -9.09446895e-01 -1.45931587e-01
-6.47931814e-01 2.19439134e-01 1.65035110e-02 1.18971765e+00
-3.75823975e-01 2.58310810e-02 5.17213702e-01 -6.60906494e-01
-1.06723535e+00 1.74032301e-01 5.37934959e-01 -5.31593442e-01
-1.28744900e-01 8.37065041e-01 -1.36506364e-01 -6.39104605e-01
2.79722095e-01 -5.33803478e-02 4.11017060e-01 3.63042057e-02
2.61343569e-01 5.22736728e-01 7.92108402e-02 2.79620420e-02
-4.21317756e-01 1.17265485e-01 9.01413560e-02 -4.73573357e-01
1.40095150e+00 -4.48698848e-01 -4.76832598e-01 3.45833540e-01
1.25472760e+00 1.03217261e-02 -1.14662671e+00 -3.64312798e-01
-9.91770551e-02 -4.56009984e-01 1.97502613e-01 -5.23039758e-01
-1.24194419e+00 7.01509714e-01 2.08835185e-01 7.82463133e-01
8.46900105e-01 -1.20464742e-01 4.12585706e-01 1.04325020e+00
5.74353278e-01 -1.20824420e+00 -2.46648520e-01 3.41448307e-01
6.83480382e-01 -1.53404713e+00 2.42401361e-01 1.48752639e-02
-2.38809571e-01 1.04352987e+00 2.73264647e-01 -4.29866403e-01
8.06481838e-01 2.25523919e-01 -1.96393847e-01 -2.91444249e-02
-1.01544094e+00 4.24736105e-02 -1.13635205e-01 4.27589327e-01
6.85034394e-02 -4.36290652e-02 -4.01351660e-01 3.07188004e-01
9.87555012e-02 -2.18289956e-01 5.29125631e-01 1.10024250e+00
-5.60967028e-01 -9.87757325e-01 -4.21196997e-01 6.10432088e-01
-6.91552877e-01 1.40549317e-02 -3.38623255e-01 5.52878916e-01
-4.50506300e-01 1.18226087e+00 -6.33240193e-02 3.30560356e-01
-1.78330258e-01 -2.78107017e-01 2.80794561e-01 -3.32639217e-01
-6.93242967e-01 1.12626851e-01 3.27433944e-02 -6.01372778e-01
-4.00509804e-01 -6.25668883e-01 -8.48577857e-01 -6.08606935e-01
-5.49975157e-01 5.91881812e-01 1.15080023e+00 1.12281179e+00
5.73607862e-01 1.68976739e-01 1.05908048e+00 -1.12852633e+00
-1.47999799e+00 -8.83311987e-01 -7.12902784e-01 -2.16535509e-01
7.04325378e-01 -1.36098877e-01 -9.03382957e-01 -3.03964615e-01] | [6.24462890625, 4.925557613372803] |
1fe73d32-1bb2-4759-9157-a706e2bf5a2b | grasping-the-inconspicuous | 2211.08182 | null | https://arxiv.org/abs/2211.08182v1 | https://arxiv.org/pdf/2211.08182v1.pdf | Grasping the Inconspicuous | Transparent objects are common in day-to-day life and hence find many applications that require robot grasping. Many solutions toward object grasping exist for non-transparent objects. However, due to the unique visual properties of transparent objects, standard 3D sensors produce noisy or distorted measurements. Modern approaches tackle this problem by either refining the noisy depth measurements or using some intermediate representation of the depth. Towards this, we study deep learning 6D pose estimation from RGB images only for transparent object grasping. To train and test the suitability of RGB-based object pose estimation, we construct a dataset of RGB-only images with 6D pose annotations. The experiments demonstrate the effectiveness of RGB image space for grasping transparent objects. | ['Markus Vincze', 'Markus Leitner', 'Stefan Thalhammer', 'Hrishikesh Gupta'] | 2022-11-15 | null | null | null | null | ['transparent-objects', '6d-pose-estimation-1'] | ['computer-vision', 'computer-vision'] | [ 1.55846864e-01 2.67160833e-01 2.25498855e-01 -5.44938207e-01
-4.41241384e-01 -6.16626382e-01 6.22051470e-02 -2.91410744e-01
-2.63260275e-01 3.30056846e-01 -2.18766123e-01 3.82572934e-02
-2.16483802e-01 -7.00195849e-01 -8.98094118e-01 -8.97365093e-01
3.55663784e-02 5.86725116e-01 4.49901819e-01 -9.42296609e-02
2.25881234e-01 6.22970283e-01 -1.64873147e+00 5.37366748e-01
3.51173013e-01 1.42146122e+00 5.60457826e-01 6.22132421e-01
-1.46826237e-01 1.18143417e-01 -5.36288202e-01 -4.81610775e-01
8.29985261e-01 3.72617990e-01 -5.31846762e-01 2.72211969e-01
2.43234217e-01 -9.94773388e-01 -2.63915867e-01 8.50926816e-01
4.76476580e-01 -3.06366026e-01 6.45839989e-01 -1.20214403e+00
-8.63234580e-01 4.76621032e-01 -3.87003630e-01 -5.16945839e-01
6.99514925e-01 1.73298836e-01 3.48120391e-01 -6.93245292e-01
4.46677595e-01 1.40428257e+00 4.87580746e-01 8.72244775e-01
-7.48176873e-01 -2.67915390e-02 2.41772398e-01 1.17308043e-01
-8.90532792e-01 -1.03543274e-01 9.50154424e-01 -2.81683564e-01
6.12177432e-01 2.09665313e-01 6.14351094e-01 1.41492712e+00
6.83132857e-02 7.33164012e-01 1.25908756e+00 -3.48630548e-01
1.77587926e-01 7.57564530e-02 -1.97993040e-01 6.15438759e-01
5.04575610e-01 -1.56561360e-01 -3.22076291e-01 -1.45972436e-02
9.25534606e-01 2.65919834e-01 -1.72342062e-01 -9.66048658e-01
-1.24338937e+00 2.35902876e-01 7.39331841e-01 -2.02951640e-01
-2.92941421e-01 4.88974601e-01 1.86363041e-01 2.50738412e-01
2.20423058e-01 2.20281824e-01 -6.82127237e-01 -9.37302485e-02
1.29892305e-01 2.83027887e-02 8.13580453e-01 1.29520822e+00
5.57604432e-01 -3.10634971e-01 8.43379796e-02 3.88383418e-01
6.42146111e-01 6.44273102e-01 2.00817019e-01 -1.10516942e+00
5.13634801e-01 7.89536476e-01 5.49567044e-01 -6.39021635e-01
-4.67297167e-01 2.95316279e-01 -2.64182031e-01 5.17169058e-01
7.82947063e-01 3.07594329e-01 -1.16758096e+00 1.10196221e+00
6.53841138e-01 -7.23457515e-01 2.26918399e-01 1.19110584e+00
9.93158221e-01 1.85581222e-01 -3.70076895e-01 2.04545498e-01
9.55652773e-01 -7.16092944e-01 -4.59597111e-01 -9.81378253e-04
2.76373565e-01 -6.77411377e-01 1.23941326e+00 8.04998279e-01
-9.26640272e-01 -3.21677655e-01 -7.25044429e-01 -3.33807290e-01
-4.38103080e-01 4.29723442e-01 8.85976017e-01 7.08920062e-01
-6.46635771e-01 5.07201254e-01 -1.03164494e+00 -1.24743022e-01
3.06304812e-01 5.79463363e-01 -5.47568381e-01 -2.88468987e-01
-4.53387946e-01 1.00469863e+00 4.32517827e-01 7.42558777e-01
-8.36980164e-01 -2.53146350e-01 -4.96571094e-01 -4.37927097e-01
3.90476733e-01 -2.19161242e-01 1.20078588e+00 -4.28433985e-01
-1.91541910e+00 1.07338309e+00 2.81609118e-01 1.54964119e-01
8.87858272e-01 -8.48239660e-01 6.31373882e-01 5.65705657e-01
-3.29190463e-01 3.47989738e-01 1.05521762e+00 -2.02011776e+00
1.66365370e-01 -8.02548230e-01 4.88962203e-01 6.10179342e-02
-3.95888954e-01 -3.52300048e-01 -3.55355650e-01 -1.79893017e-01
1.05975056e+00 -1.00955117e+00 1.43725704e-02 8.21340978e-01
-5.15837967e-01 -8.39547664e-02 1.24271500e+00 -5.43761551e-01
-2.21391782e-01 -2.00998688e+00 3.17864180e-01 1.47950754e-01
1.32615179e-01 7.71825835e-02 -7.20127448e-02 2.99067348e-01
3.21964681e-01 -2.84220040e-01 -3.49133238e-02 -4.46729302e-01
1.69807494e-01 4.14397180e-01 -3.92938614e-01 6.02037251e-01
-5.75435162e-02 8.61487150e-01 -8.08245838e-01 -4.07150507e-01
3.53034526e-01 5.80713391e-01 -2.46031523e-01 4.65837508e-01
-4.90276843e-01 6.90431535e-01 -8.44338953e-01 1.26978016e+00
1.01766992e+00 2.41791368e-01 2.25650165e-02 -5.66793263e-01
7.26657659e-02 -1.72254920e-01 -9.26462471e-01 1.94635308e+00
-4.33970958e-01 1.56018957e-01 3.44843328e-01 -7.51286387e-01
1.17648125e+00 4.36454386e-01 7.99406528e-01 -3.69193017e-01
3.80673736e-01 4.82517868e-01 -1.93989098e-01 -1.07283378e+00
4.82263453e-02 1.50834069e-01 3.37104589e-01 2.12110192e-01
-3.01363230e-01 -1.04022586e+00 -3.59705269e-01 -2.72892058e-01
9.87702847e-01 1.00496340e+00 -3.54058623e-01 2.13476166e-01
-3.81024331e-02 -1.85876891e-01 4.14746739e-02 4.53195930e-01
-9.31987241e-02 9.51878548e-01 2.35051736e-01 -6.84159815e-01
-1.07931912e+00 -1.48082614e+00 -2.93181866e-01 6.48745239e-01
5.66920638e-01 2.36692175e-01 -6.48228943e-01 -5.45616448e-01
5.77883124e-01 -3.91490348e-02 -6.28738463e-01 -1.34732664e-01
-6.13317013e-01 -5.69649339e-01 -6.41758218e-02 6.33055747e-01
6.84447467e-01 -1.03713274e+00 -1.32200575e+00 -4.78539383e-03
1.97281018e-02 -1.46430564e+00 5.12979150e-01 3.64385068e-01
-1.22510529e+00 -1.11838841e+00 -9.71040785e-01 -4.19766515e-01
8.91259253e-01 6.41590714e-01 7.47598469e-01 9.95141044e-02
-5.69938183e-01 8.46646488e-01 -9.81326699e-01 -6.45695448e-01
-1.81829259e-01 -2.67644376e-01 3.24267238e-01 -3.74850482e-01
-8.34632386e-03 -4.35597658e-01 -8.26474130e-01 3.26473385e-01
-8.31746995e-01 -2.82317489e-01 8.55228424e-01 2.19612926e-01
2.59938210e-01 -4.20498163e-01 1.55981481e-01 -1.86142519e-01
7.55121037e-02 -3.86311747e-02 -6.26264930e-01 3.76457363e-01
2.66933411e-01 -2.45101660e-01 6.63446039e-02 -6.78995252e-01
-9.92018342e-01 5.45577109e-01 -3.81093696e-02 -5.98464072e-01
-1.22636296e-01 -1.14637174e-01 -3.21666479e-01 -3.69332761e-01
5.47000885e-01 -2.40257800e-01 -2.31442247e-02 -5.78028023e-01
2.57595688e-01 8.97876322e-01 3.66984785e-01 -7.58846879e-01
9.25123811e-01 8.98909628e-01 9.05337334e-02 -9.24766481e-01
-8.03458869e-01 -3.46610039e-01 -1.04342461e+00 -3.94415021e-01
8.93175483e-01 -6.18830860e-01 -1.44028127e+00 5.65832913e-01
-1.46497011e+00 -5.27130783e-01 -9.50123817e-02 6.32624149e-01
-8.59015942e-01 4.31471765e-01 -5.24371326e-01 -1.24168515e+00
-2.14242250e-01 -1.29339445e+00 1.72170413e+00 -1.41802337e-02
1.49887785e-01 -2.92738944e-01 -4.36310381e-01 4.95313436e-01
4.10456359e-02 5.52603424e-01 6.36703551e-01 3.81000578e-01
-9.48249459e-01 -4.71388847e-01 -3.49545538e-01 3.72387379e-01
3.38048339e-01 -2.62764264e-02 -1.30994916e+00 -2.48487920e-01
1.84199452e-01 -7.11763859e-01 6.73519492e-01 2.25010067e-01
1.34034872e+00 -4.70982082e-02 -2.80934960e-01 5.22153735e-01
1.39742160e+00 -1.95022374e-02 4.85130578e-01 3.75167996e-01
9.85848367e-01 9.17996347e-01 8.21694791e-01 4.90046680e-01
8.85082707e-02 6.68813169e-01 1.36028039e+00 1.61512688e-01
6.58390224e-02 2.36296877e-01 5.74568138e-02 4.49084938e-01
-6.56753063e-01 -2.01526180e-01 -8.19106638e-01 2.30408549e-01
-1.48045361e+00 -3.05631727e-01 -3.17931414e-01 2.28167987e+00
7.19590545e-01 2.23431528e-01 -2.22957835e-01 4.31425154e-01
2.22223718e-02 -5.68651021e-01 -6.13563836e-01 -2.56871909e-01
-1.46487253e-02 -5.50740249e-02 6.11145139e-01 -1.41032794e-02
-6.75531387e-01 7.95591116e-01 5.92393017e+00 -1.46098673e-01
-9.65005219e-01 -6.67206123e-02 -9.61273611e-02 4.53669988e-02
-2.26932496e-01 -2.71440119e-01 -3.83019775e-01 1.73073471e-01
-1.52775139e-01 8.63499284e-01 2.25079149e-01 1.11700797e+00
-1.30392894e-01 -5.97639143e-01 -1.28613985e+00 1.00967968e+00
-1.25418440e-01 -4.09642965e-01 -6.20863289e-02 -1.80066943e-01
6.14864349e-01 -1.99509580e-02 2.02991381e-01 -2.17824653e-01
6.81910962e-02 -6.87984943e-01 1.17919314e+00 6.71408474e-01
5.14296830e-01 -5.51594719e-02 6.89711988e-01 3.83738399e-01
-8.06402028e-01 -3.90919864e-01 -8.51577818e-01 -7.64002129e-02
-5.91440685e-02 6.34304106e-01 -8.93438756e-01 4.17042226e-01
1.31210744e+00 1.63136661e-01 -3.64450127e-01 1.16004682e+00
-3.93733531e-01 -1.32986501e-01 -4.88477439e-01 -2.18284294e-01
1.93248820e-02 -1.35732338e-01 4.08900797e-01 4.86796021e-01
2.52115130e-01 9.81124714e-02 -3.29218395e-02 8.83027315e-01
8.37389529e-02 -5.00202894e-01 -6.57150745e-01 1.43772393e-01
7.81682208e-02 1.03665423e+00 -9.55303133e-01 8.39248300e-02
-1.53550103e-01 1.21283364e+00 2.25750342e-01 4.49561924e-01
-5.61159909e-01 -1.81612730e-01 4.31116581e-01 -2.32476637e-01
1.89059064e-01 -9.21576321e-01 -3.29377383e-01 -9.70616460e-01
7.06448555e-01 -4.09340829e-01 -4.55649585e-01 -1.18492985e+00
-1.13937414e+00 3.69617939e-01 2.18362242e-01 -1.28905654e+00
2.88425952e-01 -1.42247057e+00 -8.32296237e-02 5.25174499e-01
-1.36562049e+00 -1.41298413e+00 -8.92675817e-01 2.42065206e-01
5.30315518e-01 3.60108644e-01 8.50987434e-01 -3.37553024e-01
8.96378234e-02 2.68549547e-02 2.10123077e-01 -1.07636876e-01
5.84471047e-01 -1.06662703e+00 6.78260848e-02 8.73676389e-02
-2.19581276e-01 5.46248198e-01 7.91872919e-01 -5.11654615e-01
-2.27011085e+00 -4.32696700e-01 -1.84863791e-01 -1.08545482e+00
1.67569771e-01 -7.11096823e-01 -7.49917984e-01 5.47215223e-01
-3.74941677e-01 2.81751931e-01 -9.56845284e-02 -4.80313957e-01
-4.12409067e-01 -8.99209529e-02 -1.60772038e+00 3.99760455e-01
1.24588966e+00 -3.23748261e-01 -6.05962992e-01 5.94465554e-01
8.21340024e-01 -8.79477561e-01 -9.96475518e-01 7.74017274e-01
1.11462355e+00 -1.08272517e+00 1.38293982e+00 -2.27034122e-01
4.78587687e-01 -1.09182872e-01 -4.51336652e-01 -9.29124773e-01
3.74524176e-01 -3.35467458e-01 -2.60843158e-01 8.04011166e-01
-5.44786043e-02 -5.27296960e-01 1.23526013e+00 8.71432066e-01
-2.89705187e-01 -6.60869062e-01 -1.04617429e+00 -8.61847401e-01
-7.49224797e-02 -2.90840417e-01 4.01064336e-01 3.56126606e-01
-1.99989691e-01 -5.52425623e-01 1.14619963e-01 4.62446719e-01
9.34864461e-01 5.49977541e-01 8.26134562e-01 -1.29753053e+00
-6.01651631e-02 -7.83274621e-02 -5.46477735e-01 -9.26157713e-01
-9.51182842e-02 -2.98327863e-01 6.02446198e-01 -1.89142191e+00
1.54041365e-01 -7.36773312e-01 3.91964614e-01 4.91256446e-01
1.66915264e-02 4.72669333e-01 1.59836799e-01 1.82626337e-01
-3.64545375e-01 7.01756001e-01 1.65713918e+00 -8.96638632e-02
-5.60301393e-02 1.63795352e-01 2.23036185e-01 7.04858780e-01
8.55755270e-01 -1.45550132e-01 -2.01574802e-01 -1.05927253e+00
3.10336053e-01 -1.98633209e-01 6.57574356e-01 -9.50127125e-01
-3.94425362e-01 -1.68935165e-01 6.02590084e-01 -6.14481926e-01
9.65032041e-01 -1.40244746e+00 -3.85032594e-01 6.16701901e-01
2.25550998e-02 -3.34520817e-01 5.49355112e-02 5.34575641e-01
2.98253804e-01 -4.98897851e-01 4.22287762e-01 -6.43890917e-01
-4.45116132e-01 2.56407112e-01 -3.14393453e-02 -5.55001438e-01
1.06799591e+00 -6.12764716e-01 -2.74656534e-01 -8.72160792e-02
-5.83946586e-01 -1.45913228e-01 6.54472947e-01 4.40725356e-01
1.01175618e+00 -9.85312164e-01 -2.81312078e-01 -1.17172092e-01
2.08030686e-01 5.84633648e-01 -6.13507070e-02 5.94647527e-01
-8.75714898e-01 1.14950612e-01 -4.40923661e-01 -9.04670835e-01
-1.13556552e+00 6.45319223e-01 6.49209097e-02 7.72281945e-01
-7.45250642e-01 9.97166216e-01 -3.64430398e-02 -7.12925851e-01
5.03457904e-01 -7.90796459e-01 2.76667148e-01 -5.85685253e-01
-3.69983651e-02 4.76449728e-01 2.65831292e-01 -2.84935147e-01
-1.82452619e-01 8.83847237e-01 3.63518119e-01 -1.44618064e-01
1.49602711e+00 -1.47322848e-01 -1.97926730e-01 4.83883679e-01
1.22193897e+00 -2.58977532e-01 -1.64283478e+00 1.52163267e-01
-4.10771556e-02 -7.69556344e-01 -3.54155272e-01 -4.68531281e-01
-8.98768425e-01 1.06390119e+00 7.18275487e-01 1.18085235e-01
7.20389903e-01 2.66734660e-01 5.29952645e-01 1.16121936e+00
9.57989335e-01 -9.85373557e-01 6.29236877e-01 3.22492599e-01
1.35906029e+00 -1.60279536e+00 1.37590036e-01 -9.55412328e-01
-2.20150217e-01 1.42062700e+00 8.09798956e-01 -6.38630912e-02
4.22190875e-01 4.35191423e-01 3.92036498e-01 -1.74748734e-01
-4.68073748e-02 -1.47453442e-01 4.57700491e-02 1.07846832e+00
1.78531677e-01 -1.44378632e-01 2.36089662e-01 -4.65181433e-02
-5.59800342e-02 -2.20079109e-01 3.02621335e-01 1.46675658e+00
-5.13759077e-01 -8.41037750e-01 -7.29578733e-01 1.09684758e-01
-8.18437487e-02 5.67434192e-01 -5.91042340e-01 6.78272843e-01
4.91549261e-02 8.43255758e-01 -2.29662284e-01 -2.36930653e-01
4.29627776e-01 -1.79064319e-01 1.39676082e+00 -5.32695591e-01
-2.53863513e-01 -1.69233292e-01 -3.51045042e-01 -7.77119696e-01
-6.61350906e-01 -3.39207709e-01 -1.33843851e+00 3.59023839e-01
-5.03109872e-01 -4.34843034e-01 1.54208946e+00 7.68780053e-01
-2.61737943e-01 2.20792189e-01 6.14922822e-01 -1.58969200e+00
-1.00683856e+00 -1.05000091e+00 -3.99681389e-01 3.61067712e-01
4.26080137e-01 -9.79669869e-01 -2.14969963e-01 -2.17000558e-03] | [6.023335933685303, -1.1052803993225098] |
6ec9db3c-0e60-4e7c-a021-49ccfeb562ba | instruction-tuning-with-gpt-4 | 2304.03277 | null | https://arxiv.org/abs/2304.03277v1 | https://arxiv.org/pdf/2304.03277v1.pdf | Instruction Tuning with GPT-4 | Prior work has shown that finetuning large language models (LLMs) using machine-generated instruction-following data enables such models to achieve remarkable zero-shot capabilities on new tasks, and no human-written instructions are needed. In this paper, we present the first attempt to use GPT-4 to generate instruction-following data for LLM finetuning. Our early experiments on instruction-tuned LLaMA models show that the 52K English and Chinese instruction-following data generated by GPT-4 leads to superior zero-shot performance on new tasks to the instruction-following data generated by previous state-of-the-art models. We also collect feedback and comparison data from GPT-4 to enable a comprehensive evaluation and reward model training. We make our data generated using GPT-4 as well as our codebase publicly available. | ['Jianfeng Gao', 'Michel Galley', 'Pengcheng He', 'Chunyuan Li', 'Baolin Peng'] | 2023-04-06 | null | null | null | null | ['instruction-following'] | ['natural-language-processing'] | [-3.72685790e-02 -1.31786019e-01 -8.01541388e-01 -5.09065449e-01
-1.12286079e+00 -3.24251175e-01 6.26792490e-01 3.73246148e-02
-5.78406334e-01 7.15431154e-01 2.41301611e-01 -1.03750575e+00
4.62690175e-01 -4.84185070e-01 -1.00259352e+00 -3.47748632e-03
-7.14389607e-02 6.94891334e-01 4.10219401e-01 -7.52754688e-01
5.17026603e-01 1.37215868e-01 -1.56868100e+00 5.81159592e-01
1.10849643e+00 5.23398578e-01 5.17320037e-01 1.01905835e+00
-2.74809271e-01 6.82158470e-01 -5.76023757e-01 -6.52776286e-02
1.83358535e-01 -9.28686708e-02 -8.14552248e-01 -4.84205037e-01
5.99462628e-01 -4.31544036e-01 -3.64823282e-01 5.19191206e-01
3.47051531e-01 5.00861764e-01 4.11724210e-01 -9.36070800e-01
-9.71086204e-01 1.13879979e+00 -4.89349276e-01 3.85943979e-01
8.33672658e-02 6.88519955e-01 6.60674572e-01 -8.11366796e-01
3.32588583e-01 1.50965524e+00 3.93187821e-01 9.87297297e-01
-1.33238816e+00 -8.64974558e-01 -1.60699748e-02 -1.10242389e-01
-1.13036871e+00 -7.13158190e-01 2.43789583e-01 -7.37133771e-02
1.95781028e+00 2.28653491e-01 4.97782975e-02 1.31551397e+00
9.28010643e-01 1.01415253e+00 9.75588322e-01 -7.64878631e-01
4.30013612e-02 -1.29633129e-01 4.21637177e-01 8.83552313e-01
8.31035245e-03 6.33193552e-01 -5.49687564e-01 -4.43013050e-02
4.50500935e-01 -3.24725211e-01 8.90075937e-02 -2.67164469e-01
-1.28171873e+00 9.81584311e-01 3.29084575e-01 2.84644812e-01
1.65997982e-01 6.63430810e-01 6.02712274e-01 4.41973031e-01
2.15991914e-01 1.03842854e+00 -9.37486589e-01 -7.55294561e-01
-9.56326604e-01 1.56569824e-01 6.74919426e-01 1.32625771e+00
7.35105932e-01 4.38613981e-01 -5.13088703e-01 6.04017556e-01
3.96449380e-02 5.57286799e-01 1.23907030e+00 -9.94453490e-01
8.54264617e-01 3.20666939e-01 -1.23941720e-01 -1.99046075e-01
-4.32922989e-01 -2.28863969e-01 -1.54256150e-01 3.49707007e-02
-1.58670044e-03 -3.30328077e-01 -1.01694131e+00 1.76360798e+00
-3.57937574e-01 9.69178081e-02 1.66418791e-01 2.37360448e-01
6.98624194e-01 6.73672140e-01 1.73851028e-01 -1.63217690e-02
1.04830742e+00 -1.46253765e+00 -4.05851543e-01 -6.92802489e-01
1.45411801e+00 -7.74165869e-01 1.81077409e+00 2.77598649e-01
-1.01028049e+00 -9.13301051e-01 -1.14201272e+00 -1.14292793e-01
-3.55681598e-01 -1.90792847e-02 8.98837030e-01 7.99413681e-01
-1.20882607e+00 5.52771568e-01 -1.13188756e+00 -3.17649096e-01
2.94611305e-01 3.67134541e-01 -1.30863592e-01 -1.41220316e-02
-1.01488662e+00 1.04796004e+00 6.14170671e-01 -7.15269506e-01
-1.28775346e+00 -7.99697638e-01 -1.10368407e+00 1.34220481e-01
5.06262541e-01 -3.24189991e-01 2.03760409e+00 -4.60716158e-01
-1.67062783e+00 5.90866983e-01 -1.58134460e-01 -6.21297538e-01
-4.78133559e-02 -3.04776639e-01 -4.70551610e-01 -4.98089135e-01
-1.16675340e-01 1.00405324e+00 8.40918064e-01 -1.09483182e+00
-7.08728492e-01 1.59232780e-01 4.15623887e-03 -7.13475496e-02
-5.36107481e-01 2.53101271e-02 -4.31342870e-01 -6.57372415e-01
-7.73222387e-01 -1.08505094e+00 -8.89647603e-02 -9.75793183e-01
-1.70502409e-01 -2.72629529e-01 8.19976270e-01 -2.42919683e-01
1.62793922e+00 -1.97760272e+00 -2.35312805e-01 -3.33189309e-01
6.74501657e-02 6.83500648e-01 -9.70189810e-01 2.09394619e-01
-6.39170408e-02 3.48756611e-01 1.94917753e-01 -4.42227036e-01
2.40483060e-01 4.94640291e-01 -4.47757572e-01 -2.09395394e-01
1.36147007e-01 1.32942152e+00 -1.15487075e+00 -4.17819589e-01
1.24478146e-01 -1.13429363e-04 -1.12703764e+00 1.45214275e-01
-7.23551452e-01 4.02950570e-02 -4.36367720e-01 5.62400877e-01
1.60989091e-01 -5.16124368e-02 -1.59732401e-01 2.12212756e-01
-1.58967316e-01 6.38954997e-01 -2.95685977e-01 1.95571780e+00
-1.08781004e+00 6.08808458e-01 -3.73586565e-01 -4.18148339e-01
8.69241953e-01 -4.42229240e-04 -2.18327209e-01 -1.02833128e+00
1.50487140e-01 3.14938545e-01 2.32490763e-01 -2.87151009e-01
1.11505246e+00 1.26732692e-01 -6.81762516e-01 6.46785736e-01
3.18201959e-01 -6.20800853e-01 4.97282624e-01 1.83644712e-01
1.18815160e+00 2.54547089e-01 2.26097241e-01 -5.54132164e-01
3.03414375e-01 2.46545494e-01 1.79322749e-01 1.18510497e+00
-2.16081351e-01 2.62871068e-02 -1.27646634e-02 -4.10618216e-01
-8.50646257e-01 -6.84811354e-01 -4.38229516e-02 2.03159952e+00
-6.38474584e-01 -7.28429675e-01 -9.70348537e-01 -8.31783473e-01
1.25038087e-01 1.66845298e+00 -5.60851157e-01 -6.46044314e-01
-8.00240040e-01 -5.25598347e-01 6.46737993e-01 6.38126493e-01
3.80712867e-01 -1.16993594e+00 -7.38315284e-01 2.49588326e-01
-4.70811762e-02 -7.04793692e-01 -1.31852090e+00 8.39193344e-01
-1.10117602e+00 -5.91249406e-01 -1.10785998e-01 -8.56203914e-01
2.75413543e-01 2.53505677e-01 1.80352962e+00 2.66996473e-01
-1.49327308e-01 2.04630256e-01 -2.96162188e-01 -4.10621077e-01
-1.11693156e+00 6.02018118e-01 3.49162757e-01 -9.60965037e-01
7.68768251e-01 -2.39128992e-01 -4.54546697e-02 3.27921212e-01
-7.78238177e-01 8.06807578e-02 7.46518016e-01 1.02886486e+00
3.64940405e-01 -1.96066469e-01 5.68973362e-01 -9.81997609e-01
1.02634323e+00 -1.99762464e-01 -7.31030524e-01 2.85916120e-01
-8.93498600e-01 7.64312148e-01 8.18657696e-01 -6.86949372e-01
-1.20334280e+00 -2.64264017e-01 -1.85134947e-01 -4.10721660e-01
-1.83294147e-01 5.33089280e-01 1.96961507e-01 -3.04655910e-01
1.16399193e+00 1.95129842e-01 -2.63629109e-01 -5.02056599e-01
8.40082765e-01 7.03667283e-01 6.41908228e-01 -1.22600734e+00
4.68702465e-01 -4.47716683e-01 -3.34720403e-01 -4.07185853e-01
-7.41340339e-01 -4.77218479e-02 -5.32958031e-01 4.87043470e-01
5.50341070e-01 -7.57674456e-01 -7.39300370e-01 1.92327574e-01
-6.92558765e-01 -1.34720814e+00 -3.67228776e-01 1.80186301e-01
-9.03188169e-01 7.92655125e-02 -1.02217638e+00 -2.54203081e-01
-4.79093701e-01 -1.91560245e+00 1.03306293e+00 2.48162910e-01
-4.88060296e-01 -1.08174181e+00 2.43571281e-01 2.78493047e-01
7.95557976e-01 -5.37673831e-01 1.27283931e+00 -6.85473144e-01
-3.79415214e-01 -1.79594219e-01 5.97246215e-02 2.37739712e-01
1.09498546e-01 -7.72147626e-02 -8.47474754e-01 -5.75918674e-01
-2.19992891e-01 -9.74428415e-01 9.43886101e-01 1.86713681e-01
1.33279204e+00 -2.17966318e-01 -3.87684196e-01 8.35517645e-01
1.09006453e+00 2.18096346e-01 1.87470675e-01 3.35788399e-01
7.28189290e-01 -1.18176341e-01 8.81132782e-01 2.54679143e-01
4.10080582e-01 6.56403124e-01 1.74641892e-01 3.72021437e-01
-4.32043761e-01 -5.90143323e-01 7.44536459e-01 1.16481364e+00
3.91213357e-01 -2.64037997e-01 -1.20384967e+00 4.55493629e-01
-1.53325236e+00 -3.34498644e-01 1.50837272e-01 2.20360827e+00
1.42366934e+00 5.37146509e-01 -3.81223470e-01 -6.23978496e-01
-1.51410997e-01 -3.99828590e-02 -5.71157932e-01 -1.17895615e+00
3.05994809e-01 7.59977996e-01 8.10814142e-01 8.13918829e-01
-9.49137092e-01 1.59351456e+00 7.54470301e+00 1.45974922e+00
-9.94370639e-01 1.19533293e-01 6.37919962e-01 -3.49013478e-01
-4.34818506e-01 3.77413742e-02 -1.28975153e+00 2.88379550e-01
1.76726484e+00 -4.68066573e-01 7.17973590e-01 9.15062606e-01
8.61598700e-02 -1.88333869e-01 -1.48830724e+00 7.18570530e-01
5.15363179e-02 -1.30831873e+00 3.25512737e-01 -4.64222096e-02
1.17401314e+00 5.27246535e-01 3.74317229e-01 1.48525965e+00
1.08371162e+00 -1.07381177e+00 4.79717791e-01 3.14698786e-01
1.17513275e+00 -9.40553010e-01 4.90843803e-01 6.93651736e-01
-9.33209538e-01 -1.78450301e-01 -4.13743913e-01 -2.82545596e-01
-1.99060366e-01 -8.22416767e-02 -1.22824550e+00 7.39047527e-02
3.86780858e-01 3.93311679e-01 -1.16247404e+00 3.96843940e-01
-1.80191487e-01 8.14447284e-01 4.74233665e-02 -1.36257663e-01
3.70447576e-01 4.30700004e-01 1.44859463e-01 1.36297250e+00
3.04555804e-01 -1.26462445e-01 3.49714130e-01 7.75427401e-01
-3.72944027e-01 -1.03858970e-01 -5.36997914e-01 -2.92037517e-01
4.58744466e-01 1.05006361e+00 -2.13182613e-01 -6.29572093e-01
-5.94853401e-01 8.21700335e-01 6.80810809e-01 2.59820521e-01
-7.02683926e-01 -3.48617613e-01 8.10688198e-01 -4.11828272e-02
4.47325612e-04 -3.63928676e-01 -1.92725345e-01 -1.19634318e+00
-8.10945988e-01 -1.43564105e+00 1.40625939e-01 -9.45380151e-01
-1.12709641e+00 6.71787262e-01 3.56690347e-01 -6.23363137e-01
-7.90232241e-01 -9.40420628e-01 -6.46862626e-01 9.17606711e-01
-1.18950117e+00 -7.75816083e-01 6.09386265e-02 3.12629372e-01
9.76443410e-01 -5.70701480e-01 1.16402996e+00 4.59780842e-02
-6.59705997e-01 1.30806279e+00 3.48561466e-01 -2.37666622e-01
1.16593313e+00 -1.25598133e+00 1.22074306e+00 7.92094648e-01
-5.89985363e-02 8.92557144e-01 8.07558596e-01 -8.24669659e-01
-1.61443245e+00 -1.05602801e+00 7.33151853e-01 -1.18755770e+00
9.75764096e-01 -4.02757734e-01 -8.39458108e-01 1.16223252e+00
4.78265971e-01 -1.65250435e-01 5.87025464e-01 2.07053587e-01
-2.52281487e-01 1.75935805e-01 -7.79781461e-01 9.09942091e-01
1.00414610e+00 -4.21369553e-01 -8.59244287e-01 9.73694175e-02
1.27419960e+00 -7.75190651e-01 -9.46748078e-01 2.82221019e-01
1.48562238e-01 -6.23752952e-01 8.61310482e-01 -8.79025698e-01
4.11878288e-01 2.80856282e-01 -3.01734924e-01 -1.88240802e+00
-4.19557154e-01 -5.74005127e-01 -4.15283740e-01 8.84676695e-01
6.91191137e-01 -5.19297719e-01 6.12714291e-01 5.40018559e-01
-8.00352514e-01 -6.07802987e-01 -5.76256812e-01 -9.37732875e-01
7.63639510e-01 -5.57772994e-01 8.11837435e-01 6.00773096e-01
2.92282760e-01 5.59500694e-01 -2.30171204e-01 -5.81869841e-01
4.02145177e-01 1.24789611e-01 1.03946888e+00 -6.77700758e-01
-6.43803298e-01 -6.25353754e-01 3.53822976e-01 -1.39075363e+00
4.97411042e-01 -1.08445573e+00 4.89313990e-01 -8.61823380e-01
4.03569758e-01 -5.34470677e-01 -3.25289786e-01 9.43997800e-01
-5.26268601e-01 -6.57952949e-02 2.03293025e-01 6.57876804e-02
-8.70097518e-01 6.88367009e-01 1.39925790e+00 -1.47139162e-01
-3.49276215e-01 -3.46467137e-01 -8.29922199e-01 4.79212672e-01
9.37610090e-01 -3.28841686e-01 -6.10698462e-01 -7.82429576e-01
-1.70046523e-01 3.00190113e-02 -5.43231130e-01 -9.83441114e-01
-4.63051200e-02 -4.93534386e-01 1.03273384e-01 -3.91028970e-01
5.19663543e-02 -2.25635350e-01 -5.96435726e-01 5.84859550e-01
-5.83878994e-01 2.59965450e-01 1.06864035e+00 2.65522093e-01
7.78167173e-02 -5.00926912e-01 6.19175494e-01 -2.34279692e-01
-1.22026086e+00 1.88249841e-01 -5.57133913e-01 2.68850833e-01
7.50848651e-01 2.49859408e-01 -7.78196454e-01 -1.51736274e-01
-3.05813074e-01 5.17887354e-01 7.73426116e-01 9.50018048e-01
5.08215606e-01 -1.32732749e+00 -6.63307786e-01 4.97347146e-01
6.32550657e-01 -5.22633255e-01 5.13996705e-02 3.93163264e-01
-4.22499537e-01 1.09569013e+00 -2.74910599e-01 -3.69357467e-01
-1.05932629e+00 1.12902284e+00 1.16786122e-01 -6.24052405e-01
-2.94051349e-01 9.31951463e-01 3.14115852e-01 -8.55935156e-01
2.20247172e-02 -9.51693654e-01 1.93535849e-01 -6.71703517e-01
6.56256080e-01 -5.35381734e-02 1.51847512e-01 -2.75092542e-01
-1.21121623e-01 -2.65343394e-02 -4.37000483e-01 7.60923233e-03
9.23252881e-01 2.02390864e-01 1.93955168e-01 6.26763821e-01
1.13334060e+00 -5.96543252e-02 -1.28808343e+00 -2.71102279e-01
8.38474836e-03 -3.34412426e-01 3.20344597e-01 -9.90867853e-01
-5.48849881e-01 9.68120396e-01 4.90134627e-01 -4.02220964e-01
8.44363511e-01 -2.45688736e-01 9.72925901e-01 7.64080346e-01
8.11785936e-01 -1.12312400e+00 4.10031348e-01 1.21167040e+00
6.31293654e-01 -1.49794865e+00 -4.51115251e-01 2.17861861e-01
-3.91559184e-01 8.76202703e-01 1.20040810e+00 9.41308439e-02
3.60417426e-01 5.61962664e-01 -9.38982293e-02 3.07378083e-01
-1.41353190e+00 9.17215273e-02 2.69661665e-01 7.04673827e-01
8.45624983e-01 4.50195491e-01 -5.87230325e-02 8.13844919e-01
-4.35334474e-01 2.32734114e-01 6.49435759e-01 1.22275925e+00
-7.30700552e-01 -1.49305880e+00 -4.42719817e-01 8.65259349e-01
-2.34343991e-01 -7.70567536e-01 4.82894704e-02 7.42130876e-01
-2.63739914e-01 9.67657566e-01 7.90400431e-02 -6.60064340e-01
2.96049267e-01 3.28171074e-01 7.40940332e-01 -1.13333917e+00
-7.03115284e-01 -2.05692559e-01 2.08201319e-01 -7.15079665e-01
6.45209849e-01 -1.52473092e-01 -1.44419229e+00 -7.67988443e-01
-1.07298724e-01 1.63790181e-01 4.26019490e-01 7.15604067e-01
4.82757479e-01 1.05867696e+00 1.65538132e-01 -1.07836902e+00
-1.02207446e+00 -1.21283138e+00 -1.28785610e-01 1.49550989e-01
2.87578881e-01 -5.25738418e-01 -1.36369497e-01 -4.37338836e-02] | [10.57967758178711, 8.44201946258545] |
742f4d5d-3a63-41d4-b0de-09e44c2ba817 | making-table-understanding-work-in-practice | 2109.05173 | null | https://arxiv.org/abs/2109.05173v1 | https://arxiv.org/pdf/2109.05173v1.pdf | Making Table Understanding Work in Practice | Understanding the semantics of tables at scale is crucial for tasks like data integration, preparation, and search. Table understanding methods aim at detecting a table's topic, semantic column types, column relations, or entities. With the rise of deep learning, powerful models have been developed for these tasks with excellent accuracy on benchmarks. However, we observe that there exists a gap between the performance of these models on these benchmarks and their applicability in practice. In this paper, we address the question: what do we need for these models to work in practice? We discuss three challenges of deploying table understanding models and propose a framework to address them. These challenges include 1) difficulty in customizing models to specific domains, 2) lack of training data for typical database tables often found in enterprises, and 3) lack of confidence in the inferences made by models. We present SigmaTyper which implements this framework for the semantic column type detection task. SigmaTyper encapsulates a hybrid model trained on GitTables and integrates a lightweight human-in-the-loop approach to customize the model. Lastly, we highlight avenues for future research that further close the gap towards making table understanding effective in practice. | ['Çağatay Demiralp', 'Paul Groth', 'Isil Dillig', 'James Gale', 'Sneha Gathani', 'Madelon Hulsebos'] | 2021-09-11 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [-7.62736350e-02 5.50552964e-01 -3.56651813e-01 -6.75299704e-01
-6.87221229e-01 -8.07637930e-01 5.43869317e-01 8.50288391e-01
-7.98594281e-02 5.31212509e-01 3.19122642e-01 -7.17836380e-01
-9.90191177e-02 -1.15783858e+00 -9.89746749e-01 3.48033577e-01
1.33124098e-01 8.50593209e-01 2.48090193e-01 -4.00097787e-01
2.70710737e-01 2.38576263e-01 -1.60473657e+00 8.79984021e-01
6.80926919e-01 1.29440510e+00 -1.94147214e-01 4.98484105e-01
-5.75372815e-01 1.15104342e+00 -6.01009250e-01 -7.88665235e-01
3.15992862e-01 5.31485006e-02 -1.03346872e+00 -1.76483765e-01
5.27736008e-01 -3.52136612e-01 -5.96411824e-02 9.04273093e-01
8.60955343e-02 -1.79727390e-01 4.10066426e-01 -1.41279435e+00
-5.51520348e-01 1.19955063e+00 -2.20540583e-01 1.67060435e-01
3.70428234e-01 -1.79325730e-01 1.30878794e+00 -7.84872949e-01
7.67553389e-01 1.39691091e+00 8.09928775e-01 1.85016662e-01
-1.02600169e+00 -4.28158134e-01 3.98859411e-01 1.28146544e-01
-1.06617153e+00 -6.27014697e-01 1.24446221e-01 -4.35971230e-01
1.39057946e+00 4.37102467e-01 2.08288282e-01 8.14882576e-01
2.79318132e-02 8.66275132e-01 5.69196641e-01 -4.08056110e-01
2.49391332e-01 7.16614902e-01 3.34830970e-01 6.09182000e-01
7.37080634e-01 -6.51857495e-01 -6.77833140e-01 -2.50935983e-02
6.45087540e-01 -2.13660687e-01 1.83759660e-01 -6.59569800e-01
-1.25204611e+00 7.04559207e-01 3.78815383e-01 1.04049958e-01
-1.94269940e-01 9.62555595e-03 6.60321653e-01 2.22104371e-01
2.37860844e-01 7.62089968e-01 -1.00266075e+00 -1.83057532e-01
-5.90892911e-01 4.89775538e-01 1.46965408e+00 1.50540841e+00
5.89368224e-01 -5.37708998e-01 2.73349937e-02 5.49943328e-01
2.74846911e-01 1.63642496e-01 -8.18914920e-02 -8.47815931e-01
8.82801473e-01 1.13903880e+00 1.64052501e-01 -1.03481317e+00
-4.50488538e-01 -3.01296920e-01 -2.96432436e-01 -1.67640239e-01
6.45710766e-01 9.04870853e-02 -7.04459429e-01 1.43777978e+00
2.04649106e-01 -2.87151664e-01 2.55225867e-01 4.53068107e-01
9.76360083e-01 3.55509877e-01 2.26381391e-01 3.33008200e-01
1.67244983e+00 -7.61840045e-01 -8.77891421e-01 -6.92194641e-01
1.00286555e+00 -6.06406927e-01 1.05795693e+00 1.76284358e-01
-1.12578070e+00 -4.32409197e-01 -1.23769927e+00 -5.40036857e-01
-1.25155425e+00 1.45566076e-01 1.12060142e+00 6.67157888e-01
-7.89230347e-01 5.67232370e-01 -8.21739197e-01 -7.89244354e-01
3.47224444e-01 3.69552165e-01 -3.29311579e-01 1.05287582e-01
-1.27789915e+00 9.92854714e-01 5.81005991e-01 5.74754700e-02
-2.40249217e-01 -1.00898433e+00 -1.03798139e+00 4.56187546e-01
1.00098753e+00 -8.22330475e-01 1.37884247e+00 -2.79065162e-01
-6.25585437e-01 9.98361528e-01 -1.80857524e-01 -7.44913101e-01
4.74237770e-01 -4.40116853e-01 -5.53754926e-01 -3.89291435e-01
3.75124425e-01 4.50913727e-01 -1.43439636e-01 -1.25700271e+00
-9.74009573e-01 -4.19535607e-01 4.29551274e-01 7.45320618e-02
-7.34688565e-02 3.56200486e-02 -8.60413671e-01 -3.54829788e-01
1.62080511e-01 -4.73056883e-01 -4.75518107e-02 -1.50468349e-01
-5.83355606e-01 -2.06051856e-01 4.27145034e-01 -7.10824251e-01
1.60387242e+00 -1.89112139e+00 -4.77207989e-01 4.32416975e-01
2.99969137e-01 -5.94658852e-02 3.52862686e-01 7.74332404e-01
1.76542494e-02 5.19385099e-01 2.06164438e-02 4.28699255e-02
4.50720012e-01 8.87533352e-02 -5.15642464e-01 -2.74387628e-01
3.34991962e-01 9.21532869e-01 -6.59571946e-01 -3.77345622e-01
6.50452599e-02 1.31648898e-01 -7.13766038e-01 6.75453097e-02
-6.06459558e-01 -3.72398883e-01 -3.41610461e-01 9.48110402e-01
5.81467271e-01 -5.37706017e-01 6.44046605e-01 -3.84433746e-01
7.81722069e-02 1.07313418e+00 -1.66145790e+00 1.32066143e+00
-3.39110225e-01 3.48129600e-01 6.77676201e-02 -8.36498022e-01
9.10424173e-01 -2.43168399e-02 2.03275397e-01 -6.34918571e-01
-1.73818275e-01 4.13907379e-01 3.66568789e-02 -4.41392601e-01
6.97044730e-01 3.49709511e-01 -2.33173415e-01 4.50529218e-01
-1.21319853e-01 5.61180040e-02 6.21761084e-01 3.14271927e-01
1.20293427e+00 2.98939999e-02 4.34735358e-01 -5.06385982e-01
5.13065755e-01 2.80569971e-01 4.60804880e-01 9.43881869e-01
1.89519659e-01 2.51145959e-01 9.63110089e-01 -8.03108037e-01
-8.89175177e-01 -8.51992309e-01 -6.37883320e-02 1.24028540e+00
8.59406292e-02 -9.92378473e-01 -7.01112926e-01 -8.95199955e-01
4.76224840e-01 9.16841269e-01 -8.08509767e-01 5.75174531e-03
-5.47948241e-01 -6.51649475e-01 3.05221558e-01 1.13388491e+00
4.95577782e-01 -9.40475702e-01 -6.17974222e-01 3.58834952e-01
-3.23359936e-01 -1.46427584e+00 9.30759162e-02 4.39120680e-01
-7.27233768e-01 -1.25740099e+00 4.00471121e-01 -5.94730020e-01
4.00950581e-01 -5.44362999e-02 1.91452539e+00 1.54012635e-01
-9.51681472e-03 1.95470527e-01 -7.18106329e-02 -1.02103400e+00
-5.21110356e-01 5.39917052e-01 -4.14149016e-01 -5.74543893e-01
1.01558459e+00 -9.84068066e-02 -5.00453353e-01 3.97949398e-01
-8.38694036e-01 1.15394115e-01 4.69951451e-01 3.33071738e-01
3.21052372e-01 6.36563376e-02 4.33374584e-01 -1.60383332e+00
6.65193915e-01 -5.36752760e-01 -6.06310427e-01 7.79731095e-01
-8.96037161e-01 2.74643004e-01 4.13509995e-01 2.99077719e-01
-8.24312210e-01 -2.50048250e-01 -2.82572269e-01 2.74994850e-01
-1.09173298e-01 7.94499695e-01 -5.20116389e-01 5.13693571e-01
5.99521279e-01 -3.83866578e-01 -3.82970013e-02 -6.24666870e-01
3.30485940e-01 5.32217801e-01 5.79036057e-01 -8.34070206e-01
4.81550276e-01 5.35475850e-01 -2.97815531e-01 -1.85661286e-01
-1.01412690e+00 -3.50143313e-01 -6.13028109e-01 3.29050004e-01
6.39675379e-01 -1.12736893e+00 -9.45917189e-01 3.33090089e-02
-8.31231654e-01 -4.75508064e-01 -2.28405952e-01 -5.48326410e-02
-4.00772899e-01 -4.56343405e-02 -6.44736052e-01 -4.48635280e-01
-3.53892207e-01 -1.06818843e+00 1.04222262e+00 1.07126467e-01
-6.05699301e-01 -1.16625345e+00 -4.81188625e-01 4.36361581e-01
5.89918077e-01 2.44110189e-02 1.37609351e+00 -1.17591429e+00
-8.96237612e-01 -2.73733884e-01 -4.58621860e-01 -1.35178491e-01
9.41520855e-02 5.74943535e-02 -8.43517721e-01 1.44989178e-01
-3.46551836e-01 -2.72204012e-01 5.73352695e-01 5.59218787e-02
1.26463902e+00 -2.68837899e-01 -5.96458316e-01 5.11984110e-01
1.32010686e+00 3.42239201e-01 4.89777714e-01 9.16615367e-01
6.38965607e-01 7.22794235e-01 8.27616155e-01 4.03856963e-01
1.01898742e+00 4.77072924e-01 3.20866376e-01 -9.57521573e-02
1.89968914e-01 -4.86398697e-01 -2.87755966e-01 3.69079262e-01
3.36480200e-01 -3.76957119e-01 -1.16964972e+00 4.84719843e-01
-2.04511881e+00 -5.08279622e-01 -1.93644166e-02 2.02058816e+00
7.56534696e-01 6.51595056e-01 -3.38376127e-02 6.82813823e-02
3.38723153e-01 -1.57523945e-01 -5.92616975e-01 -6.25525653e-01
3.45148542e-03 5.90757728e-02 5.94784737e-01 4.04383093e-01
-1.29579604e+00 9.77711439e-01 6.43479252e+00 2.33070061e-01
-8.08977723e-01 -3.91546696e-01 8.92815411e-01 3.01993698e-01
-5.39643645e-01 2.91112572e-01 -1.33380914e+00 1.23804472e-01
9.65485632e-01 -2.96410739e-01 3.59640270e-01 1.01620150e+00
-2.50135243e-01 -5.54432608e-02 -1.68794489e+00 6.69908106e-01
-1.29423290e-01 -1.66948831e+00 4.43164110e-02 -7.24690408e-02
1.32013291e-01 -2.30244830e-01 -1.27970666e-01 6.61282003e-01
4.95624810e-01 -1.10529411e+00 6.91847742e-01 2.33197466e-01
4.41382229e-01 -5.26254594e-01 8.14120650e-01 -8.17599148e-02
-1.21617842e+00 -7.73881301e-02 -2.64122218e-01 -4.99743856e-02
-3.24155778e-01 4.11232859e-01 -1.09943593e+00 4.85863239e-01
9.81649458e-01 7.05902755e-01 -1.06447601e+00 5.63104093e-01
5.82343340e-02 1.37561560e-01 -3.81562293e-01 2.04606690e-02
-3.91358919e-02 3.48001793e-02 -9.05657113e-02 1.22275496e+00
1.18748851e-01 -3.82016063e-01 5.90175251e-03 1.10206628e+00
-2.77494371e-01 -2.34892461e-02 -5.24271011e-01 -2.74517745e-01
7.59099722e-01 1.21623814e+00 -1.03980660e+00 -6.62217021e-01
-6.58709943e-01 2.36686796e-01 2.71139354e-01 1.64002135e-01
-5.73590398e-01 -3.79530907e-01 7.39067435e-01 3.22493613e-01
4.35109019e-01 1.47670180e-01 -1.00682998e+00 -1.06830573e+00
3.25827807e-01 -1.36921883e+00 9.76022840e-01 -7.16000676e-01
-1.03303146e+00 3.95200849e-01 1.82922676e-01 -5.52356720e-01
-4.32619035e-01 -8.56212854e-01 -2.31340975e-01 7.46529400e-01
-1.17089581e+00 -9.32242572e-01 -3.96345019e-01 2.10823670e-01
4.18903083e-01 3.07819713e-02 9.32262719e-01 3.46368730e-01
-4.67499793e-01 6.67259991e-01 -2.20487848e-01 4.81641471e-01
8.09643030e-01 -1.66646087e+00 1.14724040e+00 8.02657962e-01
2.31121778e-02 1.17101109e+00 8.24868917e-01 -6.59235537e-01
-1.53757834e+00 -8.57451379e-01 1.20982003e+00 -9.14248884e-01
7.88396597e-01 -8.86988759e-01 -1.11731946e+00 1.27269471e+00
2.40129307e-01 -1.90403149e-01 6.73858881e-01 7.89101064e-01
-6.24002337e-01 -4.05548543e-01 -9.91920710e-01 5.84117234e-01
9.93612409e-01 -4.60985154e-01 -4.81265455e-01 3.28751355e-01
6.05622530e-01 -7.71735609e-01 -1.23009574e+00 6.52083576e-01
7.57095993e-01 -1.07712376e+00 9.12737310e-01 -6.90933526e-01
5.17066538e-01 -1.69749290e-01 -1.89321861e-01 -9.13051248e-01
-8.36328790e-02 -3.07873160e-01 -4.32956517e-01 1.39348090e+00
8.72083247e-01 -6.06423080e-01 9.25366104e-01 1.25571084e+00
3.91394868e-02 -8.39997172e-01 -8.40531364e-02 -4.02339190e-01
-2.16902292e-04 -3.37759525e-01 1.12312019e+00 8.81906867e-01
2.08860427e-01 3.91535074e-01 3.84781182e-01 1.15653008e-01
2.22794592e-01 3.59031349e-01 1.15452957e+00 -1.30218172e+00
-1.70170069e-01 -3.47103626e-01 -1.58319890e-01 -9.67359364e-01
-1.87781766e-01 -6.79714978e-01 -2.76867896e-01 -1.96501505e+00
1.51451737e-01 -5.87337077e-01 -2.98957556e-01 5.74960589e-01
-2.30853096e-01 -3.41607362e-01 2.03988105e-01 -4.18856973e-03
-7.06084669e-01 -2.50527591e-01 6.61872625e-01 -6.53647631e-02
1.75299924e-02 -1.99824069e-02 -1.48640716e+00 6.35496378e-01
7.19703674e-01 -2.27267519e-01 -4.86024231e-01 -4.76122081e-01
8.27996016e-01 -1.39560893e-01 -5.64881787e-02 -9.88279223e-01
2.99347609e-01 -2.61746813e-02 4.96360779e-01 -8.45415115e-01
6.76523596e-02 -7.12168396e-01 -7.55441636e-02 2.24649265e-01
-6.03053153e-01 5.89693189e-01 5.33592522e-01 2.40547687e-01
-2.80685753e-01 2.94852722e-02 1.59858465e-01 -5.09339690e-01
-9.52695549e-01 -5.89181036e-02 -2.21856266e-01 3.66273284e-01
5.68190694e-01 -3.32953297e-02 -6.81310475e-01 -2.79776841e-01
-4.76277202e-01 5.23840487e-01 4.83009070e-01 8.10205996e-01
9.06534642e-02 -8.48507941e-01 -3.03090483e-01 2.92175949e-01
5.31045794e-01 2.46897861e-01 -2.97630429e-01 5.70898712e-01
-8.01550508e-01 7.71630645e-01 -1.39110878e-01 -3.89895260e-01
-9.27248061e-01 7.82716870e-01 3.22924197e-01 -7.21408725e-01
-4.18024123e-01 7.11863101e-01 2.45334715e-01 -7.67124832e-01
6.01503134e-01 -8.13504159e-01 -2.05452725e-01 6.89739287e-02
4.57969934e-01 6.76463395e-02 7.60803998e-01 9.33099091e-02
-6.09638572e-01 -1.27232760e-01 -5.39691865e-01 2.14541718e-01
1.25286341e+00 -1.90182015e-01 -1.82493210e-01 5.95691681e-01
7.21686602e-01 7.31755123e-02 -8.33524108e-01 -2.13458866e-01
7.07933247e-01 -2.43928432e-01 -4.32147920e-01 -1.28618443e+00
-6.89386725e-01 7.04145193e-01 1.92968145e-01 5.99178195e-01
8.13672900e-01 -3.68617289e-02 6.76403821e-01 7.15965211e-01
1.73320383e-01 -1.24752343e+00 -4.12382573e-01 4.74575520e-01
6.57739282e-01 -1.50741684e+00 1.56807840e-01 -8.51529777e-01
-6.85462117e-01 1.27731931e+00 9.38580930e-01 2.67034203e-01
7.25270092e-01 6.89419985e-01 2.78582066e-01 -4.90313917e-01
-1.17661107e+00 -2.24531099e-01 1.71528816e-01 4.19819891e-01
9.43492651e-01 -4.85475138e-02 3.26287039e-02 6.52747095e-01
-5.05414546e-01 7.33542666e-02 4.45251167e-01 1.27659059e+00
-1.60255477e-01 -1.12907219e+00 -1.88957468e-01 7.95876324e-01
-6.71312451e-01 -2.57110417e-01 -6.54757679e-01 1.19213820e+00
-7.85279274e-02 1.02245867e+00 3.46728951e-01 -2.97061056e-01
6.35769725e-01 1.59978256e-01 3.16000849e-01 -7.77418375e-01
-7.94473648e-01 -1.43555582e-01 6.42318189e-01 -7.23351657e-01
1.17244691e-01 -5.48276126e-01 -1.08559477e+00 -4.33670312e-01
4.73733619e-02 1.83030069e-01 6.31105661e-01 8.66426468e-01
7.96205401e-01 5.78945458e-01 -2.11738259e-01 1.50765419e-01
-4.92469966e-01 -8.63955796e-01 -2.54161358e-01 4.71363991e-01
-5.23866750e-02 -6.15161657e-01 1.01734348e-01 3.79844150e-03] | [9.5758056640625, 7.878114700317383] |
490fa070-a89c-4177-ad59-c772b795c548 | revisiting-class-incremental-learning-with | 2303.07338 | null | https://arxiv.org/abs/2303.07338v1 | https://arxiv.org/pdf/2303.07338v1.pdf | Revisiting Class-Incremental Learning with Pre-Trained Models: Generalizability and Adaptivity are All You Need | Class-incremental learning (CIL) aims to adapt to emerging new classes without forgetting old ones. Traditional CIL models are trained from scratch to continually acquire knowledge as data evolves. Recently, pre-training has achieved substantial progress, making vast pre-trained models (PTMs) accessible for CIL. Contrary to traditional methods, PTMs possess generalizable embeddings, which can be easily transferred. In this work, we revisit CIL with PTMs and argue that the core factors in CIL are adaptivity for model updating and generalizability for knowledge transferring. 1) We first reveal that frozen PTM can already provide generalizable embeddings for CIL. Surprisingly, a simple baseline (SimpleCIL) which continually sets the classifiers of PTM to prototype features can beat state-of-the-art even without training on the downstream task. 2) Due to the distribution gap between pre-trained and downstream datasets, PTM can be further cultivated with adaptivity via model adapting. We propose ADapt And Merge (ADAM), which aggregates the embeddings of PTM and adapted models for classifier construction. ADAM is a general framework that can be orthogonally combined with any parameter-efficient tuning method, which holds the advantages of PTM's generalizability and adapted model's adaptivity. 3) Additionally, we find previous benchmarks are unsuitable in the era of PTM due to data overlapping and propose four new benchmarks for assessment, namely ImageNet-A, ObjectNet, OmniBenchmark, and VTAB. Extensive experiments validate the effectiveness of ADAM with a unified and concise framework. | ['Ziwei Liu', 'De-Chuan Zhan', 'Han-Jia Ye', 'Da-Wei Zhou'] | 2023-03-13 | null | null | null | null | ['class-incremental-learning'] | ['computer-vision'] | [-1.44498482e-01 -7.20100775e-02 -4.50414628e-01 -3.81813109e-01
-3.36881697e-01 -5.03783524e-01 7.07936168e-01 1.30928019e-02
-4.06337053e-01 6.47820771e-01 -6.50704429e-02 -4.07141335e-02
-1.36210948e-01 -6.82672262e-01 -8.21244657e-01 -4.06407714e-01
-2.77637672e-02 5.22572696e-01 5.44982493e-01 -3.43778104e-01
-9.02316570e-02 3.88846278e-01 -1.66028798e+00 3.48742425e-01
9.39749599e-01 1.13803625e+00 1.28353760e-01 4.95795667e-01
-1.15660608e-01 8.18097651e-01 -2.70539463e-01 -6.94449246e-01
2.57601321e-01 -2.09707871e-01 -1.08317161e+00 -2.02539414e-01
4.68404830e-01 -2.18263894e-01 -5.78712285e-01 5.40309370e-01
3.79862964e-01 1.10718943e-01 6.86070859e-01 -1.45478010e+00
-1.25326264e+00 6.96924806e-01 2.79455539e-02 1.16315663e-01
-7.53319487e-02 4.27455753e-01 9.41373587e-01 -1.43832171e+00
6.55516505e-01 9.94834006e-01 1.16146028e+00 8.34536672e-01
-1.10852826e+00 -6.14270151e-01 5.82503259e-01 6.00474358e-01
-1.47418177e+00 -5.09381890e-01 4.85225827e-01 -3.70511562e-01
1.07356644e+00 1.93389073e-01 8.71677816e-01 1.19449127e+00
7.16620907e-02 1.10016990e+00 7.18882620e-01 -3.53998125e-01
2.35192701e-01 4.81323004e-01 2.44245619e-01 5.66446006e-01
2.35055566e-01 1.38613611e-01 -7.70721793e-01 1.20562591e-01
3.57633084e-01 3.67493294e-02 -2.69963652e-01 -5.61251938e-01
-1.26446557e+00 5.31129837e-01 6.82915688e-01 2.92330235e-01
-1.38867766e-01 1.37477010e-01 6.02133334e-01 6.56429291e-01
4.76328790e-01 6.13654613e-01 -9.22480643e-01 -2.06551224e-01
-6.00893319e-01 -1.02633938e-01 6.16217196e-01 1.15806425e+00
9.37765598e-01 1.09441645e-01 -1.63740456e-01 8.37065279e-01
2.64559705e-02 1.97105840e-01 1.01627994e+00 -5.79443336e-01
2.90794432e-01 9.54562902e-01 -3.01518977e-01 -5.54772854e-01
-1.07600167e-01 -6.63969934e-01 -9.23982739e-01 -2.06241131e-01
-1.45762473e-01 1.73878863e-01 -9.36714470e-01 1.62564898e+00
3.62185359e-01 5.15276194e-01 2.84895658e-01 1.88315600e-01
8.76105487e-01 5.16262412e-01 8.06404278e-02 4.96013202e-02
6.91623211e-01 -1.47544730e+00 -1.49501100e-01 -2.96039790e-01
7.26270854e-01 -3.42721224e-01 1.39644194e+00 3.64357859e-01
-7.61502802e-01 -9.27292466e-01 -1.11215401e+00 7.37113804e-02
-6.45054042e-01 -1.49587005e-01 6.51814878e-01 4.10209179e-01
-1.24685740e+00 7.00680971e-01 -8.06594074e-01 -5.26804566e-01
6.72361612e-01 3.23636442e-01 -4.99189794e-01 -1.81919336e-01
-1.11853576e+00 9.13859606e-01 6.92361295e-01 1.08849950e-01
-1.12334895e+00 -1.09361064e+00 -6.13790274e-01 -7.45395897e-03
-8.04815907e-03 -9.25301194e-01 1.41281545e+00 -1.06221175e+00
-1.68151414e+00 6.78093433e-01 -4.01716791e-02 -5.15866339e-01
5.65180063e-01 -3.07955980e-01 -4.75463957e-01 -1.37262851e-01
-2.25525741e-02 9.38109219e-01 9.40030038e-01 -1.16531301e+00
-7.91371524e-01 1.09768204e-01 1.68321542e-02 1.67586029e-01
-1.14944589e+00 -5.20293832e-01 -6.27059758e-01 -3.86802852e-01
-1.63937539e-01 -7.92414963e-01 -5.40289171e-02 -8.37534107e-03
-5.31362602e-03 -4.02254850e-01 7.89242864e-01 -1.59986034e-01
1.60579133e+00 -2.20191622e+00 2.21193358e-01 -1.43073529e-01
2.67168075e-01 5.61228395e-01 -5.16893923e-01 4.23076093e-01
-8.60020444e-02 1.18692830e-01 -2.93823928e-01 -5.35935581e-01
3.74739468e-02 4.80497569e-01 -4.20011491e-01 7.96545856e-03
2.47964576e-01 1.34590328e+00 -1.00922263e+00 -3.02634716e-01
-7.54908770e-02 2.80200720e-01 -6.24838352e-01 3.69580351e-02
-2.39286810e-01 2.10954979e-01 -1.70492828e-01 6.49920881e-01
5.42933702e-01 -4.53621536e-01 -1.63473979e-01 -2.48359144e-01
1.14220278e-02 6.56407773e-02 -9.19639707e-01 1.60719025e+00
-5.52132905e-01 5.66842258e-01 -6.22883379e-01 -1.11093676e+00
9.08402562e-01 1.53899923e-01 3.10343742e-01 -6.05852127e-01
-2.27102160e-01 2.62195408e-01 -1.79667890e-01 -3.42234969e-01
5.47053754e-01 1.59374803e-01 6.08232543e-02 3.03229511e-01
4.17092949e-01 -4.20093760e-02 5.26166447e-02 1.36152297e-01
1.17373943e+00 9.49414074e-02 2.10754156e-01 -7.53205046e-02
4.96021122e-01 -1.12683780e-01 7.74404824e-01 7.87821054e-01
-3.73878360e-01 3.75130534e-01 -8.70721638e-02 -8.65837455e-01
-7.85830319e-01 -1.35990512e+00 -3.10273439e-01 1.37026787e+00
1.01639606e-01 -5.49590945e-01 -4.02098805e-01 -1.12691641e+00
1.47919685e-01 6.83044314e-01 -8.19273531e-01 -6.52512193e-01
-4.83258396e-01 -8.45555544e-01 4.78667200e-01 7.07487166e-01
8.02400887e-01 -1.06652796e+00 -2.89566368e-01 3.24737996e-01
-1.51695823e-02 -8.77908230e-01 -5.02372146e-01 1.15014091e-01
-1.09502530e+00 -1.03215647e+00 -3.37408632e-01 -8.99529159e-01
7.09739804e-01 1.72350869e-01 1.28753912e+00 8.54285061e-02
-9.70223621e-02 9.11986351e-01 -4.95139003e-01 -4.79716688e-01
-5.40128052e-01 6.51342869e-01 2.76947469e-01 1.90878436e-01
2.73500770e-01 -6.59101963e-01 -7.01443851e-01 4.05593306e-01
-9.31253076e-01 1.98808715e-01 7.65289187e-01 9.27373290e-01
7.08866835e-01 -1.51094511e-01 9.26704526e-01 -1.01642895e+00
4.81602490e-01 -5.97291350e-01 -1.45869434e-01 5.80670357e-01
-1.05853391e+00 -1.26368418e-01 7.59217203e-01 -9.23882186e-01
-1.14142871e+00 -9.27473530e-02 1.24381008e-02 -3.99683714e-01
1.48577020e-01 6.06286645e-01 3.16619035e-03 -2.18433768e-01
9.07471836e-01 4.26762938e-01 -1.73675597e-01 -4.85703826e-01
6.22985899e-01 6.72904849e-01 5.78990817e-01 -8.05639267e-01
8.38848650e-01 4.42572773e-01 -5.07346332e-01 -5.25380492e-01
-9.61241722e-01 -2.39919633e-01 -7.81565130e-01 -3.95973288e-02
1.66010484e-01 -1.04326856e+00 -3.68069023e-01 7.08021164e-01
-8.64887118e-01 -7.52148688e-01 -8.83004665e-01 3.32597315e-01
-4.58851248e-01 1.01571508e-01 -5.92868745e-01 -6.89615384e-02
-4.47864592e-01 -7.96385944e-01 6.04690611e-01 2.09188193e-01
-1.99555203e-01 -1.27275515e+00 2.97681123e-01 9.16257352e-02
9.28890765e-01 -7.60077462e-02 8.30403984e-01 -5.81327438e-01
-5.26815593e-01 -2.86441982e-01 -1.07988650e-02 8.27162743e-01
2.64140368e-01 -1.98345389e-02 -1.15290070e+00 -5.59372067e-01
-2.15854138e-01 -5.06044805e-01 1.09186554e+00 -2.85491459e-02
1.20151353e+00 -3.94922823e-01 -5.49090683e-01 8.44762266e-01
1.25407636e+00 3.37562896e-02 5.84310353e-01 6.18783414e-01
6.46662354e-01 1.18435159e-01 3.14405322e-01 2.66225457e-01
8.20117295e-01 4.36596632e-01 2.45323181e-01 1.18714526e-01
-5.74775219e-01 -4.26306844e-01 6.54133856e-01 1.43723476e+00
4.53785397e-02 -6.08773641e-02 -9.54809964e-01 8.92043173e-01
-1.92398107e+00 -6.87693238e-01 3.48828584e-01 2.15293550e+00
1.21404815e+00 1.79413632e-01 -3.30896854e-01 -3.48812155e-02
2.84270465e-01 -2.77806193e-01 -1.08762133e+00 -3.32196653e-01
-1.35289639e-01 3.79106879e-01 2.05762863e-01 3.34888607e-01
-1.00101984e+00 1.16798770e+00 6.46749163e+00 8.63819480e-01
-1.21530032e+00 4.31545436e-01 5.75310290e-01 -9.88718718e-02
-2.75337666e-01 1.30895570e-01 -1.21824086e+00 4.17312384e-01
1.10726559e+00 -5.33195674e-01 2.64471233e-01 1.04820287e+00
-5.10504007e-01 3.62584025e-01 -1.58915114e+00 8.70518386e-01
6.85442239e-02 -1.58262980e+00 5.67103684e-01 -4.43507761e-01
1.12517393e+00 4.64252919e-01 4.63076621e-01 1.00219274e+00
3.78366530e-01 -7.73696899e-01 6.61485374e-01 7.99793482e-01
9.64432359e-01 -4.27282304e-01 6.02775395e-01 2.84448594e-01
-1.36518586e+00 -3.39438915e-01 -4.45763081e-01 -3.19468118e-02
-1.52183056e-01 4.27727222e-01 -8.69238138e-01 6.31114542e-01
8.27994406e-01 1.26394093e+00 -1.04203701e+00 9.91458178e-01
-3.05860609e-01 8.34968746e-01 -2.28072688e-01 1.97334945e-01
1.45643160e-01 2.20233023e-01 1.80274189e-01 1.18284452e+00
2.74356753e-01 -3.88890952e-01 6.50679544e-02 5.74716985e-01
-3.57790023e-01 6.09897040e-02 -4.40838486e-01 1.94341958e-01
7.36867309e-01 1.04959571e+00 -2.44202480e-01 -5.78180075e-01
-4.21881407e-01 1.05011523e+00 7.17112601e-01 4.53230739e-01
-6.76499784e-01 -2.60354340e-01 6.76961124e-01 5.69014773e-02
4.32757020e-01 1.59888733e-02 -8.82674977e-02 -1.33628285e+00
1.70985892e-01 -7.92966545e-01 4.57051098e-01 -4.84178424e-01
-1.65056884e+00 8.91568124e-01 7.57354274e-02 -1.36177278e+00
-6.04496803e-03 -4.39905584e-01 -5.63178182e-01 2.81634420e-01
-1.84694278e+00 -1.46004426e+00 -4.48887825e-01 6.88377142e-01
7.31956959e-01 -4.78066832e-01 8.79945755e-01 3.12590331e-01
-8.20612729e-01 1.13082612e+00 3.65654051e-01 -2.87052952e-02
8.38007689e-01 -9.76648450e-01 5.50576329e-01 6.17964804e-01
3.25397909e-01 5.55493534e-01 1.00142285e-01 -2.94614732e-01
-1.11281359e+00 -1.65560567e+00 8.91562521e-01 -9.34523582e-01
7.47684479e-01 -3.45520437e-01 -1.09129596e+00 1.00779319e+00
3.26286480e-02 3.17875862e-01 6.42299294e-01 2.08489284e-01
-6.20020449e-01 -6.92555487e-01 -8.27768862e-01 5.32622159e-01
1.32736862e+00 -3.99003953e-01 -6.03671014e-01 3.59208703e-01
9.95639741e-01 -2.46933341e-01 -1.06519270e+00 5.68118691e-01
5.03162742e-01 -8.75898063e-01 9.38895822e-01 -7.11607993e-01
1.33515641e-01 -1.86976373e-01 -5.78577556e-02 -1.55264652e+00
-5.44587255e-01 -5.82365513e-01 -4.07502323e-01 1.43492270e+00
4.52643096e-01 -1.07856762e+00 5.62634051e-01 4.58227664e-01
-4.80292320e-01 -9.46618795e-01 -1.08812642e+00 -1.19796693e+00
4.13695723e-01 -5.71160913e-01 9.75966632e-01 1.04915106e+00
-3.82106937e-03 4.08365458e-01 3.28806899e-02 -6.79686964e-02
2.62830794e-01 -1.24404401e-01 9.52758014e-01 -1.41037416e+00
-3.48319232e-01 -5.19836128e-01 -4.74574029e-01 -1.20954859e+00
2.19662443e-01 -1.32039773e+00 -2.17983052e-01 -1.29321063e+00
4.17473793e-01 -9.77374911e-01 -8.74320865e-01 9.08877969e-01
-2.31485248e-01 8.22322369e-02 1.42478526e-01 5.86399972e-01
-8.90155077e-01 9.99307454e-01 1.15696275e+00 -3.52640480e-01
-4.76477027e-01 -1.75944299e-01 -7.29991555e-01 6.39393985e-01
8.25095177e-01 -3.89331043e-01 -7.74289310e-01 -7.04843760e-01
4.09892410e-01 -9.12691891e-01 3.01272392e-01 -1.25388694e+00
3.54045421e-01 2.76113637e-02 2.86719680e-01 -1.13936186e-01
1.34117469e-01 -6.49114728e-01 1.49301678e-01 3.70271623e-01
-1.78011611e-01 4.89797033e-02 5.43475091e-01 6.02211177e-01
-3.73558104e-01 -1.62234470e-01 7.22104728e-01 3.64153385e-02
-1.16869283e+00 7.91526198e-01 2.40117349e-02 1.92894399e-01
1.12288880e+00 -4.07253802e-01 -4.33834791e-01 -5.50169274e-02
-6.41490817e-01 3.69838566e-01 3.49737495e-01 7.14144349e-01
8.49029660e-01 -1.59859610e+00 -6.89393878e-01 3.33459884e-01
5.65415323e-01 1.66648045e-01 3.11006606e-01 8.13302457e-01
-3.30746084e-01 1.57976598e-01 -3.10861431e-02 -7.17435181e-01
-7.81876326e-01 6.44707620e-01 4.76604134e-01 -3.37897360e-01
-6.22616649e-01 1.02942896e+00 3.27749819e-01 -8.60366046e-01
2.27898568e-01 -4.98092651e-01 8.70798677e-02 -3.58146951e-02
5.53644419e-01 1.42907754e-01 1.32437304e-01 -2.28160664e-01
-3.12351972e-01 5.06645799e-01 -5.07673621e-01 4.32197034e-01
1.57334089e+00 -1.84495419e-01 -1.23135364e-02 6.72356308e-01
1.15073454e+00 -4.88966256e-01 -1.47537136e+00 -6.06937528e-01
-6.77596312e-03 -8.43069628e-02 -2.89873868e-01 -7.86420166e-01
-9.73423898e-01 8.38742316e-01 6.27807260e-01 -1.76468030e-01
1.13477159e+00 8.29614773e-02 8.28284562e-01 6.54581010e-01
4.04437602e-01 -1.21350825e+00 4.88966942e-01 6.93121016e-01
9.73184526e-01 -1.17933297e+00 -2.02274084e-01 1.35720715e-01
-4.97778982e-01 1.02215266e+00 9.36785638e-01 2.26214692e-01
9.18328583e-01 -3.54645811e-02 -7.17056468e-02 1.00857690e-01
-1.26172507e+00 1.42316520e-02 2.98010945e-01 7.37307608e-01
5.87069467e-02 -5.07794879e-02 1.82010874e-01 6.20687425e-01
-1.63284451e-01 2.04964504e-01 2.37966955e-01 7.53258646e-01
-3.48598391e-01 -1.11824727e+00 1.83900058e-01 3.83134753e-01
3.47525477e-01 -1.43061668e-01 -1.87407434e-01 9.96368229e-01
4.82822657e-01 6.40249014e-01 -2.79734489e-02 -6.13167107e-01
5.98979235e-01 2.28700086e-01 4.46906656e-01 -6.52877927e-01
-4.23851103e-01 -8.45903635e-01 -3.08004200e-01 -4.71358478e-01
-2.18412444e-01 -4.57655072e-01 -8.91055822e-01 -2.82825679e-01
-3.20791334e-01 1.47000223e-01 2.60468960e-01 7.50791728e-01
8.41271222e-01 4.66880500e-01 5.98043084e-01 -6.00905061e-01
-6.29216552e-01 -1.02563250e+00 -1.41212374e-01 4.65755224e-01
2.19797552e-01 -5.88157654e-01 -3.78427714e-01 3.88605446e-01] | [9.821043014526367, 3.360435962677002] |
f7cc63c2-31e8-46b0-9d21-569da5b10efc | machine-learning-for-the-prediction-of-safe | 2302.10952 | null | https://arxiv.org/abs/2302.10952v1 | https://arxiv.org/pdf/2302.10952v1.pdf | Machine learning for the prediction of safe and biologically active organophosphorus molecules | Drug discovery is a complex process with a large molecular space to be considered. By constraining the search space, the fragment-based drug design is an approach that can effectively sample the chemical space of interest. Here we propose a framework of Recurrent Neural Networks (RNN) with an attention model to sample the chemical space of organophosphorus molecules using the fragment-based approach. The framework is trained with a ZINC dataset that is screened for high druglikeness scores. The goal is to predict molecules with similar biological action modes as organophosphorus pesticides or chemical warfare agents yet less toxic to humans. The generated molecules contain a starting fragment of PO2F but have a bulky hydrocarbon side chain limiting its binding effectiveness to the targeted protein. | ['Anguang Hu', 'Mohammad Sajjad Ghaemi', 'Hsu Kiang Ooi', 'Hang Hu'] | 2023-02-21 | null | null | null | null | ['drug-discovery'] | ['medical'] | [ 7.61263549e-01 -8.17135721e-02 -6.26364350e-01 1.83001179e-02
-5.82142413e-01 -5.29849350e-01 4.25668061e-01 -2.37609278e-02
-3.95139217e-01 1.52206278e+00 3.97992991e-02 -8.00624073e-01
-3.06235343e-01 -6.97994828e-01 -9.24416244e-01 -9.80680227e-01
-9.29606855e-02 3.96636039e-01 -5.74489944e-02 -2.82370020e-02
5.29324591e-01 1.14118946e+00 -1.24926317e+00 4.31826562e-01
8.88494432e-01 3.51590186e-01 4.10367131e-01 3.02583963e-01
-4.63535003e-02 6.96275115e-01 -5.40309906e-01 4.28156070e-02
2.28336290e-01 -3.29598337e-01 -8.34104061e-01 -7.59124279e-01
1.62822604e-01 1.40312016e-01 -2.94267982e-01 1.03674233e+00
6.87203526e-01 3.09651136e-01 1.04485500e+00 -5.88156819e-01
-6.79182529e-01 4.44346905e-01 1.05434276e-01 2.20105704e-02
6.08072937e-01 3.35096180e-01 7.92546391e-01 -1.25277102e+00
7.67179251e-01 1.30997825e+00 4.08459395e-01 8.81075501e-01
-1.30421948e+00 -9.04793561e-01 -1.89215615e-01 2.71684885e-01
-1.44954395e+00 -2.50346094e-01 5.76315045e-01 -4.80539083e-01
1.56454873e+00 4.04855222e-01 3.84858549e-01 1.48380792e+00
7.42150009e-01 3.62244666e-01 4.88222957e-01 -1.38176769e-01
7.35814214e-01 -1.19637243e-01 -9.73764211e-02 5.14591709e-02
3.90015453e-01 4.22446668e-01 -5.22985876e-01 -6.01410329e-01
4.47000265e-01 4.30057049e-01 -4.41378117e-01 -2.23852873e-01
-6.80811882e-01 1.02725315e+00 4.65451360e-01 2.02351630e-01
-8.98205400e-01 1.76930606e-01 1.92928046e-01 4.02332138e-04
1.45833055e-02 1.07673311e+00 -7.54218876e-01 4.59613621e-01
-5.17415822e-01 4.59132522e-01 7.68363476e-01 4.60614651e-01
4.23965514e-01 8.29258487e-02 -1.18698046e-01 3.33490700e-01
6.96341336e-01 1.97106779e-01 4.39104706e-01 -2.06149429e-01
2.82181859e-01 3.81090432e-01 2.74063945e-01 -4.26800460e-01
-4.68229979e-01 -2.30759531e-01 -5.81904113e-01 4.48880434e-01
3.21310073e-01 -4.86662090e-01 -1.02724171e+00 1.63704920e+00
4.32998091e-01 1.48417316e-02 4.36973423e-01 5.78541458e-01
9.15932357e-01 1.01129758e+00 7.84169436e-01 -5.07952034e-01
1.26270890e+00 -4.96859968e-01 -5.56701839e-01 2.54775494e-01
5.83794475e-01 -3.35699171e-01 5.12117684e-01 7.20623672e-01
-8.33178103e-01 -7.45133385e-02 -1.10323024e+00 3.85430217e-01
-8.31767857e-01 -3.71751189e-01 9.90949631e-01 7.67268002e-01
-5.27804077e-01 1.29002786e+00 -6.17792904e-01 2.12268502e-01
5.09008169e-01 9.88463938e-01 -5.08671880e-01 3.31748351e-02
-1.61195719e+00 1.01943243e+00 8.10627401e-01 3.76584306e-02
-1.16132438e+00 -8.62133682e-01 -7.19279051e-01 9.31200571e-03
5.08656614e-02 -6.13357127e-01 7.26047635e-01 -6.51413858e-01
-1.77584660e+00 -1.00149065e-01 7.80437812e-02 -7.00310946e-01
1.10692576e-01 -4.78517488e-02 -2.10026741e-01 1.26472518e-01
2.92914268e-02 6.91728592e-01 3.50265771e-01 -4.53853488e-01
-2.52968073e-01 -3.25035453e-01 -3.48702073e-01 2.45783895e-01
8.98054168e-02 3.90818119e-01 4.50557977e-01 -5.16065657e-01
-2.54863650e-01 -8.88459980e-01 -8.71782720e-01 -5.22862554e-01
-8.60341430e-01 -3.19246054e-01 3.21843833e-01 -1.81184083e-01
8.65502775e-01 -1.41739106e+00 4.56511229e-01 4.66775656e-01
1.20708803e-02 5.53504765e-01 -3.30792218e-01 1.06409240e+00
-9.24279094e-01 4.94586766e-01 1.55868247e-01 8.84209335e-01
-4.10997391e-01 -5.94788432e-01 -2.45971680e-01 6.68440044e-01
3.56512517e-01 6.95279360e-01 -8.53457391e-01 3.07583153e-01
-2.17122361e-01 7.32695043e-01 -5.52406967e-01 -3.05655487e-02
-8.12906504e-01 6.00387692e-01 -1.25879145e+00 7.48676479e-01
6.71924591e-01 -1.33499905e-01 2.18559414e-01 9.14923027e-02
-2.29203761e-01 4.92745191e-01 -9.07636046e-01 1.27364171e+00
1.19323820e-01 1.09688193e-02 -8.14232051e-01 -6.19990408e-01
9.57755983e-01 6.62731826e-01 5.29766560e-01 -5.28951406e-01
-1.14609338e-02 4.90781069e-01 3.04091185e-01 -5.13283968e-01
-1.78774014e-01 -4.82824832e-01 3.92939419e-01 5.63198738e-02
8.28201417e-03 3.65256369e-01 2.07407355e-01 -3.29021871e-01
1.45036685e+00 3.46321762e-01 3.54381382e-01 -3.23564827e-01
6.39977157e-01 -1.80229247e-01 7.18536556e-01 4.43094254e-01
3.54308695e-01 5.77061586e-02 4.84864861e-01 -9.04163897e-01
-1.16858459e+00 -4.68666583e-01 -4.82725531e-01 7.70115614e-01
-4.83514041e-01 -1.83977365e-01 -7.91677237e-01 -4.22253728e-01
-1.41975224e-01 6.13696218e-01 -6.10202134e-01 -2.29908034e-01
-2.89399236e-01 -8.08371007e-01 6.32825136e-01 2.07448825e-02
-7.97221735e-02 -1.50990915e+00 -2.03941822e-01 7.20515668e-01
7.06018090e-01 -3.52186292e-01 -5.74980713e-02 1.05075371e+00
-6.82199359e-01 -1.28987634e+00 -9.38183129e-01 -6.38823032e-01
3.92405927e-01 -3.93187463e-01 5.63522339e-01 -3.18446130e-01
-4.06374902e-01 -5.42050302e-01 -2.33684972e-01 -1.05019426e+00
-4.42926675e-01 -6.45172689e-03 2.44552150e-01 -3.53008032e-01
6.65493906e-01 -6.50870740e-01 -6.03593826e-01 -3.47930565e-02
-9.18150663e-01 -3.92698884e-01 6.60067856e-01 8.16541076e-01
7.41501689e-01 -1.72597155e-01 1.11681175e+00 -1.04190063e+00
8.20371211e-01 -8.06720734e-01 -5.32410145e-01 1.82345301e-01
-1.44917622e-01 3.96472275e-01 8.53750050e-01 -8.37495804e-01
-6.10607147e-01 7.20164180e-01 -4.02569801e-01 -2.04887599e-01
-2.81647235e-01 7.26844609e-01 -6.84238017e-01 -2.62540996e-01
1.12546742e+00 1.85500398e-01 -2.61062831e-01 -4.78067458e-01
-9.50417891e-02 3.11259091e-01 -6.89299330e-02 -5.20334303e-01
4.74539548e-01 -8.06022435e-02 3.60422999e-01 -9.52133894e-01
-1.61412925e-01 -9.23285335e-02 -2.62559235e-01 2.00799972e-01
7.43889034e-01 -6.98984623e-01 -1.24552524e+00 -1.88168082e-02
-1.59033203e+00 -4.36178632e-02 6.61193300e-03 8.94493043e-01
-5.07983804e-01 2.17600897e-01 -4.66220230e-01 -7.35811949e-01
-8.26924384e-01 -1.16809976e+00 4.59907621e-01 1.22060575e-01
-3.52285206e-01 -5.61585605e-01 3.83372843e-01 -9.45616737e-02
-1.03899509e-01 6.38794899e-01 1.22287834e+00 -1.42557454e+00
-8.26792359e-01 -2.28602529e-01 2.86201686e-01 -7.14811264e-03
1.31516233e-01 3.57005820e-02 -7.32333124e-01 -1.51007632e-02
-4.64013256e-02 -2.67248571e-01 6.95635080e-01 7.05003440e-01
1.03686035e+00 -4.47613448e-01 -5.85291862e-01 5.86847901e-01
1.29038846e+00 1.25399470e+00 1.16973126e+00 4.09677476e-01
6.27016008e-01 4.52273339e-01 5.69764614e-01 3.89793247e-01
-7.42049217e-01 4.58104968e-01 4.88200068e-01 5.29705845e-02
3.21027994e-01 -1.66826159e-01 3.01678091e-01 -4.18254495e-01
-6.01353310e-02 -4.34264034e-01 -9.08426225e-01 1.82156965e-01
-1.47053981e+00 -1.17231858e+00 -1.97654933e-01 2.33392525e+00
1.37021327e+00 3.40359621e-02 2.10107714e-01 -1.19284168e-01
8.32804263e-01 -4.97504950e-01 -9.40518498e-01 -6.71106279e-01
7.84345120e-02 7.89100826e-01 6.70300364e-01 4.92156833e-01
-8.83364618e-01 9.93070304e-01 7.33282518e+00 1.10777903e+00
-1.08389378e+00 -8.85858595e-01 6.98964238e-01 4.43342365e-02
-4.24756825e-01 5.97096421e-02 -1.28664196e+00 5.02741754e-01
1.44750488e+00 -2.84004271e-01 1.88474506e-01 6.45276904e-01
5.65511107e-01 2.98573107e-01 -1.49340320e+00 6.56958163e-01
-5.56650519e-01 -1.84008324e+00 4.92303401e-01 2.92092621e-01
4.79093790e-01 -6.82397708e-02 6.76607192e-02 -2.88110692e-02
-9.16740950e-03 -1.68590224e+00 1.32930994e-01 8.62395763e-01
8.66145849e-01 -1.16758859e+00 3.34334135e-01 5.00480175e-01
-5.12776792e-01 -1.80279285e-01 -6.65718555e-01 -1.17332386e-02
-3.22156817e-01 3.91485810e-01 -1.21236527e+00 3.48367691e-02
1.66024730e-01 5.03254354e-01 -6.51086122e-02 1.31336963e+00
-1.93292484e-01 1.56063825e-01 -1.68288097e-01 -8.03475261e-01
4.80105877e-01 -3.74148697e-01 4.89250243e-01 8.75911474e-01
1.56951115e-01 2.89745748e-01 5.87147325e-02 1.17129922e+00
-2.90005922e-01 4.11246777e-01 -1.23168886e+00 -2.58396894e-01
4.81823951e-01 7.74796009e-01 -4.29286063e-01 -1.25329122e-01
-4.89027277e-02 4.47399706e-01 -1.23831101e-01 5.64287186e-01
-5.66597819e-01 -8.48467290e-01 5.83821177e-01 -8.83793756e-02
2.27871984e-01 5.15368104e-01 2.12391987e-01 -4.23461884e-01
-5.28775573e-01 -1.17955792e+00 7.37280399e-02 -4.66724932e-01
-1.00056350e+00 5.42857707e-01 -1.67997777e-01 -9.91967738e-01
-1.60433307e-01 -7.33759701e-01 -6.98141813e-01 1.46401000e+00
-1.05720365e+00 -7.38495588e-01 7.50906944e-01 3.75648052e-01
6.72385216e-01 -5.15558839e-01 1.20009875e+00 2.15350732e-01
-5.40561080e-01 1.84082940e-01 4.54603583e-01 -5.46685219e-01
4.87958074e-01 -9.39733863e-01 3.39267999e-01 2.25247115e-01
-3.55044961e-01 1.15248585e+00 1.03839469e+00 -1.11507976e+00
-1.37834942e+00 -1.18873167e+00 8.76587391e-01 -4.76477355e-01
3.78070831e-01 -2.55042791e-01 -8.30259442e-01 2.56060034e-01
-1.23278461e-01 -3.11617970e-01 1.26264262e+00 -3.02122355e-01
-1.98968261e-01 6.07054830e-01 -1.30676782e+00 7.55919933e-01
5.82503140e-01 -2.63770789e-01 -4.33777511e-01 8.09044242e-01
8.71075571e-01 -2.67752141e-01 -8.28206420e-01 3.33477914e-01
4.63958383e-01 -3.52704495e-01 1.29820251e+00 -1.44740927e+00
3.22761983e-01 -5.15445352e-01 1.34021848e-01 -9.53914821e-01
-4.79840010e-01 -8.11568916e-01 4.27048746e-03 3.73913825e-01
9.23995495e-01 -4.34747249e-01 1.29555249e+00 7.83399343e-01
1.07810810e-01 -1.10792196e+00 -9.14309859e-01 -7.24267900e-01
3.83277833e-01 1.44342959e-01 5.50329328e-01 5.90001822e-01
3.48468900e-01 9.48220670e-01 -4.44564819e-01 -2.52180193e-02
1.75389871e-01 -4.91791844e-01 1.47918329e-01 -1.50064754e+00
-1.02668822e-01 -1.88591540e-01 -3.64589304e-01 -5.74758947e-01
2.70736396e-01 -8.28096688e-01 -4.87139374e-01 -1.05215251e+00
1.22453600e-01 8.49587619e-02 -2.69326508e-01 4.48459148e-01
4.41413939e-01 -3.03161442e-01 -3.89595956e-01 -8.53662416e-02
-3.20720941e-01 3.71039242e-01 8.49301159e-01 -4.80896145e-01
-7.72874653e-01 4.45922285e-01 -9.17163789e-01 6.30500734e-01
7.42290497e-01 -8.58138740e-01 -5.22868991e-01 7.01454699e-01
7.10980237e-01 3.87557894e-01 -2.01807067e-01 -6.77416801e-01
2.31749684e-01 -6.07338786e-01 7.89856851e-01 -7.76459157e-01
1.78765714e-01 -8.26449990e-01 6.48786306e-01 9.61319685e-01
-5.69196343e-01 -4.45466697e-01 2.06498221e-01 7.39015937e-01
2.13405583e-02 -7.46755362e-01 7.54371524e-01 -4.25467938e-01
-4.71130908e-02 6.86485469e-01 -9.25620377e-01 -7.30226994e-01
8.42227876e-01 -6.13846183e-01 2.15814829e-01 1.26246726e-02
-6.85228229e-01 -1.49562016e-01 8.51527601e-02 1.44989088e-01
9.40292478e-01 -1.20939815e+00 -5.65514088e-01 3.15950252e-03
5.47589883e-02 -2.41565585e-01 7.81869665e-02 1.82303727e-01
-7.34718859e-01 9.62171555e-01 -2.39693269e-01 -2.25765221e-02
-1.25804102e+00 9.83093560e-01 8.09638381e-01 -1.58389866e-01
-3.19866747e-01 7.51153946e-01 4.38360095e-01 -3.24657738e-01
3.28505903e-01 -1.67325959e-01 -7.89680302e-01 -6.68196380e-02
1.04727507e+00 2.54722834e-01 2.36889109e-01 -2.38570914e-01
-4.96022552e-01 1.94368199e-01 -3.69987667e-01 5.52514315e-01
1.80677319e+00 8.68947387e-01 -9.76152197e-02 -1.36144623e-01
1.29584098e+00 -3.22056681e-01 -9.05016184e-01 1.05274595e-01
3.58171165e-01 -1.14789546e-01 -5.53919226e-02 -9.84211683e-01
-3.40148926e-01 3.99913818e-01 7.13250458e-01 -2.77283788e-01
4.59190726e-01 -3.69384527e-01 2.91955739e-01 1.17659664e+00
1.39414757e-01 -1.09632242e+00 5.27655892e-02 6.12693071e-01
9.81929362e-01 -6.37205839e-01 1.59411907e-01 -1.96556702e-01
-1.85626298e-01 1.39128816e+00 2.35398650e-01 -1.54933780e-01
3.89470786e-01 -1.91900805e-01 -6.01231635e-01 -7.37964571e-01
-9.91289556e-01 3.03514153e-01 3.26423079e-01 7.62480557e-01
7.16731668e-01 -3.97686571e-01 -7.39642322e-01 5.84995806e-01
4.08053607e-01 -2.39009578e-02 4.91201729e-01 7.07714021e-01
-6.77851796e-01 -1.50876069e+00 -5.88760018e-01 1.84454650e-01
-9.15141523e-01 -6.42586529e-01 -8.64492655e-01 4.90085036e-01
-7.44764805e-02 6.77083552e-01 -6.55871689e-01 -2.51533911e-02
3.08101416e-01 2.83656836e-01 1.87373385e-01 -8.06014419e-01
-7.47618139e-01 3.24202538e-01 9.87897217e-02 -2.82735705e-01
-6.86224625e-02 -2.79158741e-01 -1.41048002e+00 2.53074884e-01
-6.75897062e-01 6.06510937e-01 8.10432196e-01 6.96062088e-01
4.55254972e-01 3.76020461e-01 9.15502310e-01 -9.26129699e-01
-3.04283500e-01 -8.46513689e-01 -9.12334085e-01 -6.60476089e-02
4.00678515e-01 -4.95776743e-01 -1.88445106e-01 -2.04323843e-01] | [4.922391414642334, 5.712667942047119] |
5fd2f22e-e481-41ec-ab3a-4b8a6dd4efb4 | pyabsa-open-framework-for-aspect-based | 2208.01368 | null | https://arxiv.org/abs/2208.01368v2 | https://arxiv.org/pdf/2208.01368v2.pdf | PyABSA: A Modularized Framework for Reproducible Aspect-based Sentiment Analysis | The advancement of aspect-based sentiment analysis (ABSA) has urged the lack of a user-friendly framework that can largely lower the difficulty of reproducing state-of-the-art ABSA performance, especially for beginners. To meet the demand, we present \our, a modularized framework built on PyTorch for reproducible ABSA. To facilitate ABSA research, PyABSA supports several ABSA subtasks, including aspect term extraction, aspect sentiment classification, and end-to-end aspect-based sentiment analysis. Concretely, PyABSA integrates 29 models and 26 datasets. With just a few lines of code, the result of a model on a specific dataset can be reproduced. With a modularized design, PyABSA can also be flexiblely extended to considered models, datasets, and other related tasks. Besides, PyABSA highlights its data augmentation and annotation features, which significantly address data scarity. All are welcome to have a try at \url{https://github.com/yangheng95/PyABSA}. | ['Ke Li', 'Heng Yang'] | 2022-08-02 | null | null | null | null | ['classification', 'term-extraction', 'aspect-based-sentiment-analysis'] | ['methodology', 'natural-language-processing', 'natural-language-processing'] | [-2.47121722e-01 3.27895321e-02 -1.73789844e-01 -6.68293834e-01
-8.31329048e-01 -7.64706135e-01 5.92372477e-01 2.40980431e-01
-2.16629710e-02 9.56695378e-02 1.76545426e-01 -5.01833856e-01
1.47225305e-01 -9.06017482e-01 -2.87138700e-01 -3.11241239e-01
2.73675472e-01 2.27638692e-01 -1.43277645e-01 -6.17636979e-01
2.88099110e-01 -5.50498515e-02 -1.74448073e+00 6.77044749e-01
4.54686552e-01 1.06109083e+00 -1.83679014e-01 3.40544492e-01
-3.98892075e-01 3.39703977e-01 -5.09121120e-01 -9.29283798e-01
1.63902283e-01 -1.57673076e-01 -8.72291386e-01 -2.87020564e-01
1.15117297e-01 1.29613802e-01 3.88229311e-01 6.58966839e-01
4.61433381e-01 -1.95210725e-01 3.27591628e-01 -1.56901801e+00
-5.97450137e-01 7.92200863e-01 -4.68619585e-01 -1.80607408e-01
5.49532175e-01 3.24969620e-01 1.53606105e+00 -1.05646110e+00
5.61882913e-01 8.39861572e-01 8.56060922e-01 4.31510895e-01
-8.31938863e-01 -5.20356596e-01 3.79305452e-01 -5.54889888e-02
-1.04691184e+00 -5.67681611e-01 6.83440447e-01 -2.76542664e-01
1.53587067e+00 6.91913962e-01 1.04756761e+00 1.12740695e+00
1.05484193e-02 1.13322330e+00 1.13543975e+00 -3.28221947e-01
2.01381773e-01 1.65256947e-01 6.37045026e-01 5.71209550e-01
3.48398894e-01 -3.91835600e-01 -8.74219239e-01 -4.04576749e-01
-2.13719279e-01 -1.29353270e-01 1.28282048e-02 -2.42955342e-01
-1.17125416e+00 7.16776013e-01 -1.40387490e-01 3.26949805e-01
-1.31883532e-01 -3.85878384e-01 6.93571627e-01 4.37344193e-01
6.45096421e-01 7.26042509e-01 -1.09477985e+00 -6.25672996e-01
-7.90408909e-01 4.89444911e-01 1.19163883e+00 1.34126663e+00
8.21090758e-01 1.57869495e-02 -1.63084999e-01 8.64293694e-01
4.11967337e-01 5.84739447e-01 3.96201909e-01 -5.94110668e-01
1.70724198e-01 1.20235610e+00 -3.20903540e-01 -6.46183491e-01
-4.56809610e-01 -3.93434972e-01 -4.22311813e-01 2.70066023e-01
1.38794452e-01 1.34209348e-02 -5.67971230e-01 1.27476144e+00
5.02213478e-01 -5.86727500e-01 -7.86364600e-02 5.94603181e-01
1.19142139e+00 3.33082020e-01 -9.64486822e-02 1.49125010e-01
1.87588775e+00 -1.25854087e+00 -5.11876285e-01 -3.22219878e-01
7.73497581e-01 -1.26732421e+00 1.67279136e+00 5.76312780e-01
-1.14941835e+00 -8.87990445e-02 -1.13739538e+00 -2.54328787e-01
-9.65688288e-01 6.37494400e-02 1.10396385e+00 7.14766860e-01
-1.07092130e+00 2.53265560e-01 -1.03024960e+00 -6.29369795e-01
3.73048604e-01 1.09556831e-01 -2.09003150e-01 1.17129259e-01
-8.43141377e-01 4.83773768e-01 -2.57862490e-02 -1.83291808e-02
-4.34879631e-01 -1.07105458e+00 -1.02612889e+00 -2.60758162e-01
4.72559452e-01 -1.12298691e+00 1.38341248e+00 -7.76854873e-01
-1.39985013e+00 1.14310598e+00 -3.61690104e-01 2.22389959e-02
1.57779127e-01 -4.84247118e-01 -4.57087100e-01 -4.23762500e-01
2.35980242e-01 2.86788315e-01 6.35859966e-01 -8.97424757e-01
-4.08589691e-01 -4.77131426e-01 3.96904230e-01 1.19512238e-01
-4.97283250e-01 4.67977524e-01 -7.12778211e-01 -7.02964246e-01
-3.73541057e-01 -8.81634057e-01 -2.24710003e-01 -2.10659936e-01
-6.75511003e-01 -2.94978023e-01 6.78303599e-01 -3.47239017e-01
1.43831682e+00 -2.24322557e+00 -2.82415926e-01 1.88362747e-02
9.65216309e-02 1.02559716e-01 -4.37123716e-01 7.49005914e-01
-1.35209247e-01 3.15942556e-01 -3.77652198e-01 -7.66483784e-01
3.63858193e-01 -2.25867823e-01 -3.47447515e-01 1.05016537e-01
1.79680869e-01 1.01482058e+00 -6.65353298e-01 -9.06843990e-02
5.54446317e-02 5.78374088e-01 -6.36491477e-01 2.06337258e-01
-4.77311134e-01 5.54318912e-02 -5.26133180e-01 1.22131765e+00
7.99741149e-01 -3.36751878e-01 1.79468729e-02 -1.95011035e-01
-3.11673820e-01 7.05900788e-01 -1.01475906e+00 2.01063800e+00
-6.05580091e-01 4.55412269e-01 1.86887696e-01 -4.10694033e-01
8.88530672e-01 7.75953382e-02 5.18292367e-01 -5.33687115e-01
2.07505688e-01 4.66547236e-02 -3.68609399e-01 -2.82254905e-01
9.05524611e-01 1.58407465e-02 -3.18490565e-01 9.65497315e-01
-9.17876791e-03 -4.16541755e-01 5.26401222e-01 2.07468495e-01
9.63272810e-01 3.24908584e-01 5.06008148e-01 -3.54589611e-01
4.48796213e-01 2.48140544e-01 2.80113041e-01 4.09718364e-01
-3.72829698e-02 7.36679554e-01 6.04816616e-01 -5.81588268e-01
-9.33067918e-01 -7.82345116e-01 -2.98010767e-01 1.47517490e+00
-3.98871303e-01 -1.36937761e+00 -5.58323920e-01 -8.41896594e-01
-8.30178484e-02 8.58845294e-01 -7.73923576e-01 2.31059223e-01
-1.79320082e-01 -1.07457876e+00 3.90425563e-01 2.66460687e-01
3.76528025e-01 -1.13015926e+00 -3.55077952e-01 -1.74722493e-01
-6.80597946e-02 -8.70207489e-01 -3.62370074e-01 5.50344549e-02
-5.83817422e-01 -1.02822673e+00 -1.87860534e-01 -5.78446269e-01
2.96453357e-01 4.07815635e-01 1.75793791e+00 1.58735320e-01
-2.13401586e-01 6.07061744e-01 -7.24223137e-01 -8.53260219e-01
-2.06461594e-01 6.09707177e-01 -3.60351622e-01 -3.55188370e-01
8.57800364e-01 -6.85654998e-01 -7.74933875e-01 2.45280653e-01
-1.02848327e+00 4.43129092e-02 3.14250201e-01 2.02209130e-01
6.05899334e-01 -4.32078719e-01 3.17934811e-01 -1.22592270e+00
7.70100713e-01 -5.72144270e-01 -4.18222427e-01 1.06499739e-01
-1.10480428e+00 -4.30843949e-01 4.08431739e-01 2.53995538e-01
-8.03502798e-01 -3.54868561e-01 -5.55363774e-01 3.64745498e-01
-2.23889008e-01 8.53988230e-01 -2.37089455e-01 3.83063704e-01
5.57277262e-01 1.11099802e-01 1.02617741e-01 -6.16226614e-01
6.07943594e-01 7.71477282e-01 -1.83781147e-01 -4.37241346e-01
4.77119297e-01 5.85329235e-01 -3.69308114e-01 -7.70144284e-01
-1.09752703e+00 -5.41441858e-01 -3.73576641e-01 3.63175832e-02
4.20097589e-01 -1.16879284e+00 -5.60950518e-01 7.49122143e-01
-7.65487611e-01 -2.78671324e-01 -4.07955527e-01 -3.63729782e-02
-3.95555109e-01 1.91191450e-01 -4.89704669e-01 -4.87532258e-01
-1.01109302e+00 -1.12879848e+00 1.21442974e+00 1.59377143e-01
-8.08041394e-01 -9.94365871e-01 3.46701294e-01 8.13875794e-01
7.18342662e-01 1.02779232e-01 6.59388423e-01 -8.14052343e-01
-3.50728631e-01 -4.26712722e-01 1.90353855e-01 3.50117147e-01
2.34040588e-01 6.35927737e-01 -1.35520554e+00 -2.93011904e-01
-5.38405105e-02 -3.83865625e-01 4.95869964e-01 1.65141508e-01
9.54661429e-01 -1.29254565e-01 -1.38621584e-01 6.79903269e-01
1.12280631e+00 -1.95197627e-01 4.43910718e-01 8.99341822e-01
5.34232557e-01 4.15026844e-01 8.87745023e-01 7.10062027e-01
8.92038822e-01 4.10613865e-01 4.69985604e-01 -5.21527454e-02
-1.52813628e-01 1.44745395e-01 6.62233651e-01 1.33357668e+00
-1.61846876e-02 -2.39644852e-02 -8.73386443e-01 6.46663487e-01
-1.53287339e+00 -5.27057290e-01 -5.54766953e-01 1.97066700e+00
8.67934406e-01 2.64455657e-02 3.44441891e-01 -6.78242519e-02
-1.88203871e-01 5.88897586e-01 -3.57908219e-01 -8.32290590e-01
-3.59496772e-01 3.00679594e-01 -2.33583689e-01 2.86261469e-01
-1.07859778e+00 8.81892204e-01 6.42249107e+00 8.13672304e-01
-9.73757923e-01 2.06578702e-01 5.81569016e-01 -3.89996350e-01
-8.36019754e-01 1.55598745e-01 -7.64169157e-01 1.56732544e-01
1.00344467e+00 -4.37215179e-01 3.48406911e-01 1.26307619e+00
5.92198670e-02 -1.34070665e-01 -1.02951300e+00 6.01373136e-01
1.57454893e-01 -1.23148394e+00 1.48240075e-01 -1.44464239e-01
5.94808161e-01 4.14059639e-01 2.26020500e-01 7.45807648e-01
1.62511468e-01 -7.88793921e-01 5.50145328e-01 1.93176046e-01
6.31027043e-01 -6.19888902e-01 8.21249545e-01 -1.45739600e-01
-1.30102921e+00 3.63515198e-01 -1.35535687e-01 -2.70120740e-01
6.14227802e-02 7.95529604e-01 -6.51723325e-01 9.34056878e-01
1.00228286e+00 1.03900826e+00 -9.41640913e-01 8.28572154e-01
-2.39061341e-01 6.56083047e-01 -1.32411167e-01 -2.51121879e-01
4.06265706e-02 -4.31855768e-01 6.00375235e-01 1.42043233e+00
2.06940979e-01 -3.47014576e-01 -1.20906599e-01 5.78202069e-01
2.35208794e-01 6.91887677e-01 -7.42638588e-01 -2.66387850e-01
1.75640419e-01 1.76449215e+00 -6.07430875e-01 -2.46978313e-01
-9.15125489e-01 6.75891817e-01 2.78888077e-01 2.08345383e-01
-6.69358432e-01 -3.62953752e-01 1.08809054e+00 8.63285437e-02
3.12684000e-01 -6.58383891e-02 -7.95479417e-01 -1.34999967e+00
3.22179884e-01 -1.41837978e+00 5.82013845e-01 -8.83883178e-01
-1.31670952e+00 9.86560225e-01 -2.33216926e-01 -1.24121094e+00
-8.11761245e-02 -6.54583812e-01 -6.69874787e-01 8.37670922e-01
-1.39820015e+00 -1.57942557e+00 -3.49443674e-01 4.71955568e-01
5.29595971e-01 -1.80044159e-01 1.18805647e+00 2.93402463e-01
-6.08436525e-01 7.29427993e-01 -3.14182401e-01 1.26980646e-02
9.62505281e-01 -1.28211689e+00 8.82723212e-01 6.33032084e-01
3.02002784e-02 9.89460647e-01 8.05556536e-01 -3.70779246e-01
-1.63844347e+00 -9.58287954e-01 1.08428121e+00 -1.08933997e+00
1.14099324e+00 -5.81178427e-01 -6.13905966e-01 8.86085868e-01
6.10979319e-01 -3.79811287e-01 1.32591927e+00 8.30544055e-01
-6.41513348e-01 -3.47944260e-01 -7.21506953e-01 7.79206276e-01
8.35538745e-01 -5.66346824e-01 -4.85550314e-01 2.71851867e-01
7.69199729e-01 -2.33854264e-01 -1.06415641e+00 5.72971165e-01
7.02784777e-01 -1.27591479e+00 5.97516418e-01 -5.90147018e-01
6.44048810e-01 -3.45405072e-01 -1.72187075e-01 -1.11805701e+00
-7.22229183e-02 -6.53349042e-01 7.68076926e-02 1.59517801e+00
9.45174873e-01 -7.83834755e-01 6.03800058e-01 7.63568997e-01
-2.40908593e-01 -1.20952666e+00 -5.53236783e-01 -4.87411767e-01
2.03674391e-01 -1.12365639e+00 9.68479753e-01 9.38494503e-01
2.59028405e-01 5.54035783e-01 7.89726898e-02 -1.11494213e-01
2.15709493e-01 7.13593841e-01 1.18865621e+00 -7.88001716e-01
-4.33301479e-01 -7.18243420e-01 1.69833156e-03 -7.08832681e-01
-3.04201096e-01 -1.07203341e+00 -3.88590485e-01 -1.49430418e+00
4.91614699e-01 -4.44618493e-01 -2.68556148e-01 8.62043500e-01
-3.54039431e-01 5.14354944e-01 2.25200117e-01 4.07777615e-02
-7.50829637e-01 6.44857168e-01 1.20191681e+00 -1.39019817e-01
-1.47723526e-01 2.65350282e-01 -1.39282751e+00 6.73503220e-01
9.32921231e-01 -4.52298284e-01 -3.96302581e-01 -3.59414607e-01
9.11616683e-01 -6.88028455e-01 -4.14800085e-02 -6.01603806e-01
-1.81415573e-01 1.10527091e-01 -2.08429739e-01 -6.72532141e-01
3.20729941e-01 -4.82555866e-01 1.56869628e-02 -5.00402078e-02
9.51998606e-02 4.55473065e-01 5.32272816e-01 3.25867459e-02
-3.91627014e-01 -1.32986486e-01 1.41525596e-01 -1.78083137e-01
-3.56449872e-01 3.28806281e-01 -3.78592640e-01 1.45193249e-01
8.03555310e-01 8.74361247e-02 -6.44725561e-01 -2.20941663e-01
-4.85997200e-01 1.12437479e-01 8.00860107e-01 6.25801444e-01
1.77274853e-01 -9.56940830e-01 -5.87333381e-01 3.84570509e-01
6.49651289e-01 1.22017592e-01 2.93239355e-01 9.81570303e-01
-2.47448713e-01 3.85609746e-01 1.21776909e-01 -3.56344521e-01
-1.53595746e+00 5.87835670e-01 -1.50334954e-01 -3.91316652e-01
-3.78044575e-01 7.84647107e-01 -9.27399769e-02 -9.40581143e-01
-2.28786096e-01 -3.72577608e-01 -1.85906917e-01 2.81668663e-01
6.95326746e-01 8.58461335e-02 5.73754847e-01 -3.27586502e-01
-5.16067326e-01 4.34681475e-01 -3.24219346e-01 -4.32350338e-02
1.70833480e+00 -6.24655336e-02 -5.90813339e-01 8.06883335e-01
9.15906250e-01 6.20618701e-01 -7.12105930e-01 5.27333170e-02
-2.75807440e-01 -2.30128616e-01 -2.43276060e-01 -1.26645434e+00
-1.10774899e+00 5.17955005e-01 8.57374370e-02 6.47863328e-01
1.34449565e+00 1.27536535e-01 6.15263760e-01 1.62564605e-01
3.14644784e-01 -7.96582878e-01 -2.10322112e-01 7.13021934e-01
1.01864791e+00 -1.20326006e+00 3.89032990e-01 -4.33909327e-01
-8.78434539e-01 8.26905012e-01 4.88717467e-01 1.75120980e-01
9.04111147e-01 5.44856846e-01 6.83898032e-01 -5.25924027e-01
-1.08744764e+00 -1.32446736e-01 2.65080184e-01 5.74277401e-01
8.64573181e-01 1.36566699e-01 -4.78423268e-01 1.15786803e+00
-8.43327522e-01 -2.50459641e-01 4.78928864e-01 1.18080771e+00
1.73752531e-01 -1.52231562e+00 -8.68063942e-02 4.62922513e-01
-5.16604841e-01 -6.62416756e-01 -4.48970705e-01 9.24473524e-01
-2.35199988e-01 1.07293594e+00 -2.88057089e-01 -2.84634799e-01
5.08363068e-01 2.64235288e-01 -4.25567813e-02 -7.20811546e-01
-1.11440504e+00 5.19457869e-02 4.93332505e-01 -8.29407096e-01
-4.69286799e-01 -9.17964697e-01 -8.81640077e-01 -5.21545112e-01
-7.78239146e-02 1.36475399e-01 9.84607875e-01 7.52142191e-01
9.04204190e-01 3.56918126e-01 4.68036026e-01 -2.13275582e-01
3.97221409e-02 -1.13986111e+00 -3.92735690e-01 2.44295835e-01
-4.71612141e-02 -1.49708644e-01 -3.68371934e-01 -4.21105549e-02] | [11.470094680786133, 6.755216121673584] |
671a1336-410a-43a7-b844-03deb61232a6 | corn-yield-prediction-based-on-remotely | 2211.13286 | null | https://arxiv.org/abs/2211.13286v1 | https://arxiv.org/pdf/2211.13286v1.pdf | Corn Yield Prediction based on Remotely Sensed Variables Using Variational Autoencoder and Multiple Instance Regression | In the U.S., corn is the most produced crop and has been an essential part of the American diet. To meet the demand for supply chain management and regional food security, accurate and timely large-scale corn yield prediction is attracting more attention in precision agriculture. Recently, remote sensing technology and machine learning methods have been widely explored for crop yield prediction. Currently, most county-level yield prediction models use county-level mean variables for prediction, ignoring much detailed information. Moreover, inconsistent spatial resolution between crop area and satellite sensors results in mixed pixels, which may decrease the prediction accuracy. Only a few works have addressed the mixed pixels problem in large-scale crop yield prediction. To address the information loss and mixed pixels problem, we developed a variational autoencoder (VAE) based multiple instance regression (MIR) model for large-scaled corn yield prediction. We use all unlabeled data to train a VAE and the well-trained VAE for anomaly detection. As a preprocess method, anomaly detection can help MIR find a better representation of every bag than traditional MIR methods, thus better performing in large-scale corn yield prediction. Our experiments showed that variational autoencoder based multiple instance regression (VAEMIR) outperformed all baseline methods in large-scale corn yield prediction. Though a suitable meta parameter is required, VAEMIR shows excellent potential in feature learning and extraction for large-scale corn yield prediction. | ['Zhou Zhang', 'Yuchi Ma', 'Zeyu Cao'] | 2022-11-23 | null | null | null | null | ['crop-yield-prediction', 'crop-yield-prediction'] | ['computer-vision', 'miscellaneous'] | [-2.57832348e-01 -3.49265695e-01 -5.05408704e-01 -2.46118858e-01
-3.71519089e-01 -3.28763068e-01 1.03647470e-01 5.84196746e-01
3.26058902e-02 7.37508953e-01 -2.95697749e-01 -4.71536756e-01
1.47065818e-02 -1.69424284e+00 -7.90177584e-01 -8.70005786e-01
-7.14579821e-02 2.17657819e-01 -2.07876619e-02 -4.37230468e-01
-1.36925101e-01 4.71897274e-01 -1.78503084e+00 4.30101305e-02
1.31084049e+00 1.22706628e+00 5.25780678e-01 3.38537395e-01
-3.67642432e-01 1.35717928e-01 -2.40314007e-01 6.81394637e-02
1.90499008e-01 -2.96610773e-01 -2.08406866e-01 7.14223757e-02
2.02702042e-02 -6.25481188e-01 4.56490600e-03 1.27687788e+00
2.32709095e-01 1.27536476e-01 1.02440858e+00 -1.29369664e+00
-8.69334340e-01 9.53258395e-01 -1.20383120e+00 -1.52044252e-01
-4.64543611e-01 -3.56446356e-02 9.24798727e-01 -8.25061262e-01
3.64613086e-02 1.08420885e+00 5.67943513e-01 1.07124411e-01
-1.34354627e+00 -8.18128824e-01 2.15898424e-01 7.96167031e-02
-1.24385428e+00 8.81255418e-03 5.31287432e-01 -5.19332170e-01
9.01805341e-01 1.05949134e-01 9.70614195e-01 5.81135094e-01
4.99497414e-01 9.29341435e-01 6.57006204e-01 -9.76032764e-02
1.68912023e-01 -1.62436783e-01 1.63926035e-01 4.25568551e-01
7.14551568e-01 5.84332705e-01 1.08740635e-01 -9.48260725e-02
1.10296226e+00 4.55796868e-01 -3.91043611e-02 -4.27743107e-01
-1.08593535e+00 1.21363533e+00 7.13498592e-01 2.13535622e-01
-9.23737407e-01 -2.20219675e-03 3.46903443e-01 3.47233176e-01
7.17389405e-01 3.19107682e-01 -9.09719765e-01 3.06226969e-01
-1.14222324e+00 3.47228438e-01 4.97525007e-01 7.65044451e-01
7.94688463e-01 8.00175428e-01 3.90735082e-02 9.22879338e-01
4.32760030e-01 1.06332314e+00 4.93993551e-01 -6.17034853e-01
1.16101176e-01 8.01511884e-01 -1.00741349e-02 -1.36032641e+00
-2.74379611e-01 -1.66857690e-01 -1.41144288e+00 3.17133635e-01
2.26635858e-01 -2.19353944e-01 -9.38999176e-01 1.40518212e+00
1.90779418e-01 1.43850669e-01 2.79897332e-01 8.98002625e-01
8.38703632e-01 1.25763893e+00 5.47845304e-01 -3.12953979e-01
1.04715109e+00 -3.96766543e-01 -9.09020305e-01 -6.53871000e-02
7.48198628e-01 -5.04915595e-01 4.42130566e-01 1.90445080e-01
-5.53258181e-01 -6.01678014e-01 -1.05309236e+00 3.92010570e-01
-8.26485395e-01 5.67413867e-01 9.95844960e-01 2.90274560e-01
-4.99559075e-01 8.83052945e-01 -1.06286085e+00 -3.87975782e-01
3.78845066e-01 5.92972897e-02 -4.56992149e-01 3.37377526e-02
-1.07127404e+00 9.53460276e-01 9.13712084e-01 5.34047127e-01
-6.80375755e-01 -6.97926462e-01 -1.29464066e+00 3.32583368e-01
2.52284795e-01 -1.10492982e-01 7.85782576e-01 -7.93501437e-01
-1.13152122e+00 4.54828411e-01 -1.68142974e-01 -4.50429589e-01
-2.02850759e-01 -3.08760643e-01 -6.56352282e-01 -4.72018242e-01
5.77368811e-02 6.52171552e-01 9.13856089e-01 -1.18617630e+00
-7.03840256e-01 -6.72938108e-01 -8.28468561e-01 -1.94413960e-01
-1.38675854e-01 -3.41924310e-01 4.95463490e-01 -7.11677969e-01
7.16196775e-01 -5.72763681e-01 -4.17941779e-01 -2.21131861e-01
-2.51546465e-02 -7.04517886e-02 9.84336972e-01 -1.02408910e+00
1.18424582e+00 -2.06057644e+00 -1.49957277e-02 3.06041121e-01
-8.48776251e-02 3.71588349e-01 -4.17229056e-01 -1.04779499e-02
-2.28698179e-01 1.74176872e-01 -5.28215587e-01 6.08110011e-01
-1.17234468e-01 4.50503469e-01 -3.03644031e-01 2.43405044e-01
7.39891469e-01 9.40899968e-01 -8.76748204e-01 -2.27649480e-01
6.70985937e-01 8.49812403e-02 -4.10139352e-01 2.70598888e-01
-4.25564766e-01 -1.73937611e-03 -6.35513723e-01 1.24512315e+00
1.23806620e+00 8.27422589e-02 -8.61555412e-02 -3.52189481e-01
-3.35630834e-01 -8.40518236e-01 -1.10979664e+00 1.26783860e+00
-2.58613110e-01 3.36780220e-01 -1.13512076e-04 -1.42991209e+00
1.45931435e+00 3.00097346e-01 6.90014958e-01 -4.85031962e-01
7.65935332e-02 4.00340438e-01 1.49320439e-01 -2.13834673e-01
7.41885900e-01 3.06214452e-01 6.96701109e-02 -6.55215681e-02
1.60706684e-01 -1.73597947e-01 4.70184796e-02 -2.79043674e-01
3.92640412e-01 5.18830121e-01 6.51714385e-01 -4.55733031e-01
2.34155446e-01 5.23040056e-01 9.13565338e-01 3.50405991e-01
-3.90014052e-01 4.45883542e-01 3.74875277e-01 -6.23048842e-01
-1.05797648e+00 -8.99171829e-01 -4.37847257e-01 1.05670416e+00
-1.54874753e-02 1.18983231e-01 -2.62499571e-01 -3.30406457e-01
5.43025970e-01 8.34625423e-01 -4.31466371e-01 -2.08746076e-01
-1.02893107e-01 -1.46005273e+00 1.30488962e-01 8.45129550e-01
6.92092776e-01 -1.34664917e+00 -1.97982520e-01 6.72925115e-01
-2.08416879e-01 -6.09928787e-01 2.33177662e-01 4.66184556e-01
-1.20320404e+00 -9.71872091e-01 -8.52263689e-01 -4.60752606e-01
3.74382854e-01 2.55834252e-01 1.24066377e+00 -2.42304325e-01
-2.69322753e-01 -2.49297693e-01 -4.80144620e-01 -9.70961452e-01
-4.78707463e-01 1.10445298e-01 8.34738687e-02 -4.68815804e-01
1.02044082e+00 -3.72751176e-01 -3.01630855e-01 -2.95757223e-03
-8.17225397e-01 -3.98044199e-01 6.89549923e-01 1.26070321e+00
9.30633247e-01 2.50475317e-01 7.33693361e-01 -4.33937192e-01
2.92726219e-01 -9.52300191e-01 -1.17968845e+00 3.82990062e-01
-7.38190532e-01 3.58099528e-02 2.90964723e-01 -3.52225840e-01
-9.42349136e-01 8.48793909e-02 -4.62888032e-02 -2.60972321e-01
-3.63635927e-01 1.06050026e+00 -1.93853348e-01 4.13877428e-01
3.36550206e-01 2.13523448e-01 9.31832939e-02 -2.76011109e-01
7.54829794e-02 4.93407249e-01 2.91623175e-01 -3.49650681e-01
7.09706724e-01 -3.91507834e-01 1.09883353e-01 -1.07044852e+00
-4.42869097e-01 -3.30507278e-01 -8.15462112e-01 9.65148509e-02
9.45900381e-01 -1.14832473e+00 -5.59570312e-01 6.05407596e-01
-1.06782138e+00 -3.07354927e-01 -1.49474382e-01 8.75931203e-01
-2.24152774e-01 -9.85424519e-02 -2.85000652e-01 -8.28263521e-01
-4.71695274e-01 -1.01560807e+00 8.77220869e-01 3.55709672e-01
7.79850706e-02 -7.36248493e-01 -1.10969238e-01 -2.49676809e-01
5.27399778e-01 4.50840265e-01 1.15109897e+00 -5.05015671e-01
-4.05424744e-01 -3.31261992e-01 -6.30393028e-01 3.98146033e-01
3.82431567e-01 4.71641123e-01 -8.57075691e-01 -1.98896732e-02
-4.34822649e-01 -1.67073384e-01 1.11502266e+00 1.30856478e+00
1.37877250e+00 2.11438015e-01 -2.98744798e-01 7.40794837e-01
1.65297151e+00 4.78676736e-01 5.15744448e-01 2.94219762e-01
6.92812622e-01 3.72605592e-01 9.87428904e-01 5.76858759e-01
1.27655938e-02 -1.29210949e-01 7.52894104e-01 -2.13539183e-01
6.75022423e-01 -2.18742296e-01 3.19282115e-01 6.88850462e-01
-3.24938267e-01 -2.21143305e-01 -6.62511826e-01 7.12129354e-01
-2.13912725e+00 -1.34628224e+00 -1.33127347e-01 2.02947330e+00
3.89296055e-01 -4.55237240e-01 -2.23738357e-01 2.79183328e-01
5.44859171e-01 2.90481627e-01 -7.85012782e-01 -6.55018687e-01
-3.79589379e-01 1.97919965e-01 1.05605018e+00 -1.22684287e-02
-1.44037771e+00 1.24944651e+00 5.58902836e+00 7.87385702e-01
-1.02400291e+00 -3.74463439e-01 7.00629354e-01 4.19683337e-01
-1.19496346e-01 -2.62549102e-01 -8.39010656e-01 1.43678486e-01
6.20302618e-01 2.82002687e-02 2.09644258e-01 1.22231925e+00
4.46950793e-01 -3.10246676e-01 -6.88829720e-01 6.39769673e-01
-5.28862059e-01 -9.83759165e-01 7.20241070e-02 -3.11515406e-02
8.70323896e-01 -1.20113529e-02 -1.70631304e-01 5.95146596e-01
7.46895671e-01 -1.14980006e+00 -3.05888392e-02 4.14192170e-01
6.58200026e-01 -1.06363165e+00 1.15475345e+00 2.64917046e-01
-1.71017778e+00 -1.14727259e-01 -1.09635305e+00 1.12436585e-01
-2.56099522e-01 9.42929566e-01 -7.41459280e-02 6.84896171e-01
8.22402418e-01 1.07566690e+00 -2.07885355e-01 7.54968882e-01
1.64568454e-01 7.67874718e-01 -4.37429160e-01 8.57025832e-02
3.13838422e-01 -7.67063141e-01 2.82993078e-01 8.15773249e-01
9.18412447e-01 3.12106669e-01 4.99942362e-01 1.09623694e+00
2.21670598e-01 2.87575036e-01 -7.99268723e-01 -3.71881694e-01
4.70111221e-01 1.00578022e+00 -4.36704427e-01 5.90080805e-02
-4.45464194e-01 7.88645685e-01 -1.37415275e-01 4.03374046e-01
-4.89627838e-01 -3.67253810e-01 1.03735375e+00 -3.17119956e-01
3.89681160e-01 -3.28167081e-02 -2.95080602e-01 -1.30791187e+00
-2.48423785e-01 -7.84014463e-01 2.42236778e-01 -5.71468890e-01
-1.19839644e+00 2.18672767e-01 -9.06478241e-02 -1.21842551e+00
-4.03065711e-01 -8.44108582e-01 -6.07825279e-01 1.17427611e+00
-1.76070213e+00 -1.16527641e+00 -3.32034469e-01 1.82291195e-01
5.51338911e-01 -6.52027726e-01 1.50347865e+00 -1.47708714e-01
-9.58869517e-01 2.70623535e-01 5.56508720e-01 2.84877151e-01
3.64959359e-01 -1.21031618e+00 1.66789457e-01 9.79922593e-01
-1.53762728e-01 2.07981467e-02 4.71833646e-01 -8.83017182e-01
-1.29912376e+00 -1.54506266e+00 6.27587676e-01 2.87288100e-01
6.57675922e-01 3.41645867e-01 -1.26846945e+00 8.17689359e-01
-2.16171443e-01 2.92597950e-01 4.48824465e-01 2.57354736e-01
3.30473743e-02 -4.17121887e-01 -1.26139128e+00 1.84755549e-01
3.32092822e-01 7.37073794e-02 -1.40855148e-01 -1.57010064e-01
5.69195330e-01 -1.52073279e-01 -1.21132529e+00 7.95543253e-01
6.39078081e-01 -4.36133534e-01 8.93472135e-01 -5.04741073e-01
9.14408267e-01 -6.26221150e-02 -5.16840219e-01 -1.88840342e+00
-6.71219707e-01 2.84529746e-01 -3.00410897e-01 1.24185002e+00
4.18479264e-01 -4.71340805e-01 7.06110239e-01 4.91229594e-01
1.51826441e-01 -4.23925191e-01 -1.99219808e-01 -6.26948893e-01
5.06469607e-01 -2.84975380e-01 1.11943591e+00 9.76709247e-01
-3.44954371e-01 -3.09278190e-01 -3.04674506e-01 5.78105450e-01
7.10351408e-01 4.43634391e-01 5.17433703e-01 -1.73426187e+00
3.17255914e-01 -5.67666054e-01 -5.17413199e-01 -5.35547912e-01
1.78234518e-01 -7.57618546e-01 -5.71095794e-02 -1.48643053e+00
4.57911082e-02 -4.73494172e-01 -5.08624256e-01 6.56816244e-01
-5.40100396e-01 -3.23667526e-01 -2.30160877e-01 -2.77715504e-01
5.64432800e-01 8.29918027e-01 1.15254486e+00 -4.59354937e-01
-6.87982857e-01 2.95131505e-01 -4.44167703e-01 5.99354804e-01
1.06770194e+00 -3.67257059e-01 -2.05456629e-01 -5.30720234e-01
-5.42037301e-02 3.25594097e-01 2.13732958e-01 -7.31476665e-01
-3.82242739e-01 -8.39339972e-01 1.06033576e+00 -1.33462739e+00
-2.23589525e-01 -1.14481139e+00 -5.86624928e-02 3.82855684e-01
-3.11844777e-02 1.10850297e-01 6.39225543e-01 4.17650521e-01
-4.33466107e-01 -1.87894136e-01 5.23484349e-01 -2.23797500e-01
-1.33462787e+00 8.38754773e-01 -4.54920143e-01 -6.56374037e-01
1.01160502e+00 -6.19051009e-02 9.14354846e-02 9.54064429e-02
-7.15331078e-01 6.70348644e-01 -5.90299815e-02 5.91752291e-01
8.19678545e-01 -1.48147774e+00 -1.27103686e+00 6.68955505e-01
3.06981862e-01 3.16645026e-01 2.34615386e-01 2.62311757e-01
-7.78316557e-01 2.62795091e-01 -4.80495453e-01 -6.91640437e-01
-8.96378219e-01 5.28767288e-01 2.75407106e-01 -2.44874686e-01
-2.81665921e-01 5.85460246e-01 -1.27865136e-01 -5.91911554e-01
-1.28494263e-01 -6.06808782e-01 -5.79651594e-01 3.29408944e-01
4.73028809e-01 2.43906915e-01 -3.81629281e-02 -4.71253604e-01
5.63744754e-02 2.99256355e-01 2.79353499e-01 2.99999833e-01
1.61160147e+00 5.62247895e-02 -1.87029243e-01 6.04660571e-01
7.77161300e-01 -6.95962667e-01 -9.42762375e-01 -2.20041439e-01
1.15662441e-01 -3.35152954e-01 6.76768064e-01 -6.16247773e-01
-1.65145338e+00 9.95147943e-01 9.18707311e-01 7.83472583e-02
1.13456821e+00 -6.27869606e-01 7.57199764e-01 6.53855622e-01
1.50756672e-01 -1.04967010e+00 -6.16513491e-01 6.39385819e-01
8.05472732e-01 -2.25165009e+00 3.55051383e-02 -9.27931964e-02
-6.59000933e-01 1.28073764e+00 8.43200028e-01 -1.91657528e-01
1.18670166e+00 3.70374709e-01 1.87222898e-01 2.04270244e-01
-3.95083249e-01 -3.65880758e-01 1.20519109e-01 1.00912321e+00
6.65784061e-01 7.86409199e-01 9.38014463e-02 6.95547938e-01
-2.88633779e-02 5.11063673e-02 3.12015712e-02 6.49463475e-01
-6.82406306e-01 -9.78905678e-01 -6.00911379e-01 1.05253172e+00
-1.76302433e-01 -2.24577963e-01 2.31272861e-01 7.02986360e-01
-6.03246428e-02 9.70934331e-01 3.68538409e-01 -2.00417787e-01
1.72253683e-01 1.70559138e-01 1.58190787e-01 -3.68739963e-01
-1.91093668e-01 1.69497088e-01 -5.41193247e-01 -5.03092766e-01
-3.38073790e-01 -6.33750677e-01 -9.62743998e-01 -7.42645264e-01
-7.85582542e-01 -5.72778396e-02 7.83185542e-01 8.25907230e-01
1.15545304e-03 5.95962286e-01 7.23849237e-01 -8.60586822e-01
-6.90064728e-01 -1.19466257e+00 -1.31212330e+00 -1.90057456e-02
3.30228895e-01 -6.38984919e-01 2.28470080e-02 -5.93510717e-02] | [9.363112449645996, -1.6017647981643677] |
6d101a77-04e6-45a9-b7de-faa79fdf8dbb | joint-entity-and-relation-extraction-for | null | null | https://aclanthology.org/2020.coling-main.137 | https://aclanthology.org/2020.coling-main.137.pdf | Joint Entity and Relation Extraction for Legal Documents with Legal Feature Enhancement | In recent years, the plentiful information contained in Chinese legal documents has attracted a great deal of attention because of the large-scale release of the judgment documents on China Judgments Online. It is in great need of enabling machines to understand the semantic information stored in the documents which are transcribed in the form of natural language. The technique of information extraction provides a way of mining the valuable information implied in the unstructured judgment documents. We propose a Legal Triplet Extraction System for drug-related criminal judgment documents. The system extracts the entities and the semantic relations jointly and benefits from the proposed legal lexicon feature and multi-task learning framework. Furthermore, we manually annotate a dataset for Named Entity Recognition and Relation Extraction in Chinese legal domain, which contributes to training supervised triplet extraction models and evaluating the model performance. Our experimental results show that the legal feature introduction and multi-task learning framework are feasible and effective for the Legal Triplet Extraction System. The F1 score of triplet extraction finally reaches 0.836 on the legal dataset. | ['Hongfei Lin', 'Zhihao Yang', 'Yuanyuan Sun', 'Yanguang Chen'] | 2020-12-01 | null | null | null | coling-2020-8 | ['joint-entity-and-relation-extraction'] | ['natural-language-processing'] | [ 1.09192871e-01 -9.85384583e-02 -5.64280450e-01 -5.02889395e-01
-1.12870300e+00 -5.51644325e-01 4.58698303e-01 1.48119867e-01
-5.38435757e-01 1.04577613e+00 3.84350002e-01 -3.84256572e-01
-4.35333848e-01 -7.20820904e-01 -6.49795979e-02 -5.46459794e-01
3.64697576e-01 5.88813663e-01 2.38215122e-02 -2.61819273e-01
4.34715301e-01 4.23830032e-01 -9.36508238e-01 6.40247047e-01
1.63660681e+00 9.80102122e-01 1.36659648e-02 -3.74831371e-02
-5.39913177e-01 1.19468224e+00 -6.70668960e-01 -1.20439184e+00
1.38781175e-01 -2.08424449e-01 -6.63261473e-01 -9.92330238e-02
-1.13853641e-01 -3.03311259e-01 -7.68994763e-02 8.78948390e-01
5.18455744e-01 -3.78793329e-01 8.29836786e-01 -1.12475789e+00
-7.73448348e-01 6.66741729e-01 -5.94598353e-01 7.37057179e-02
3.43644232e-01 -4.09900784e-01 1.12267268e+00 -7.31993496e-01
9.78808880e-01 1.00836408e+00 3.49750131e-01 4.33730073e-02
-3.57707709e-01 -8.41408491e-01 -2.17423335e-01 5.04513800e-01
-1.29749131e+00 -4.62636590e-01 6.10859573e-01 -6.19725049e-01
1.22513831e+00 -1.99977368e-01 1.73495889e-01 9.08469141e-01
5.05614400e-01 8.84791374e-01 1.12030995e+00 -3.90667737e-01
-2.92232126e-01 1.89875975e-01 5.14965415e-01 6.67993963e-01
6.92467213e-01 -1.49026036e-01 -4.29921657e-01 -3.78862828e-01
2.96414822e-01 9.24855098e-02 2.89004952e-01 2.52566993e-01
-7.55058825e-01 9.06191349e-01 -1.05300181e-01 5.34365594e-01
-3.68214756e-01 -3.57469410e-01 7.71453857e-01 1.52932018e-01
6.30852520e-01 2.71414399e-01 -6.23500705e-01 -2.61538863e-01
-6.18652582e-01 1.16399623e-01 5.63295901e-01 1.30172789e+00
4.67558712e-01 -4.59677190e-01 -2.17552319e-01 1.01203966e+00
2.57430196e-01 7.91919827e-01 3.02940518e-01 -6.20947778e-01
1.08943796e+00 1.00284958e+00 -4.41456661e-02 -1.07317030e+00
-3.87647390e-01 -4.92613167e-02 -5.40636003e-01 -6.23227656e-01
2.04779372e-01 -4.89149511e-01 -6.75021648e-01 1.05447435e+00
3.32915157e-01 -4.30286467e-01 3.50180954e-01 3.29691410e-01
9.31656182e-01 3.64739239e-01 3.54729176e-01 -3.79180312e-01
1.63946950e+00 -7.04411924e-01 -1.41242170e+00 7.95407146e-02
8.62928092e-01 -1.02411175e+00 4.56692576e-01 2.48957530e-01
-3.85847986e-01 -2.11272791e-01 -8.33621740e-01 -4.28569853e-01
-5.48847914e-01 7.46705294e-01 1.15157306e+00 7.66734004e-01
3.20609920e-02 1.83833569e-01 -5.36853373e-01 -1.66336030e-01
8.53997231e-01 2.45717540e-01 -3.25857699e-01 -4.07372147e-01
-1.38520312e+00 1.14076722e+00 6.38078272e-01 2.07596973e-01
-2.41607562e-01 -1.58986479e-01 -9.25170064e-01 -6.61169291e-02
7.49605656e-01 -1.78592369e-01 8.91875327e-01 -2.58480281e-01
-7.17258811e-01 8.20066452e-01 -2.99688429e-01 -7.71866962e-02
3.46820444e-01 -1.67840049e-01 -9.89778817e-01 3.31671506e-01
8.61832619e-01 -2.69323736e-02 5.57183862e-01 -5.53714514e-01
-8.05873394e-01 -6.81216359e-01 -2.24033743e-01 -1.33981463e-02
-6.18052244e-01 7.62922645e-01 -4.48448956e-01 -5.13469219e-01
-2.21406296e-01 -6.64497018e-01 -8.39806125e-02 -5.94351709e-01
-5.83253384e-01 -7.52391517e-01 7.50957191e-01 -9.50081885e-01
1.43540156e+00 -1.85214818e+00 -4.63725507e-01 3.34504843e-01
1.80849463e-01 3.65907729e-01 2.09427513e-02 7.71789193e-01
1.75565362e-01 1.86784863e-01 -2.83993274e-01 6.91608116e-02
-8.47070757e-03 2.26355463e-01 -2.82620013e-01 4.92912717e-02
2.78114438e-01 1.06118608e+00 -8.23845446e-01 -1.09002149e+00
-2.41214395e-01 1.79877311e-01 -6.89525083e-02 -1.20853037e-01
3.10902335e-02 2.11754099e-01 -1.40098059e+00 9.64366794e-01
7.53481448e-01 -2.35405445e-01 3.10609847e-01 -2.39648402e-01
-2.91881356e-02 4.67761517e-01 -6.66620553e-01 1.65365720e+00
-1.43029079e-01 7.09088966e-02 -6.11509085e-01 -7.64468968e-01
9.63994563e-01 6.09986603e-01 6.92146838e-01 -1.04546559e+00
2.83238709e-01 4.16691631e-01 -9.00121927e-02 -1.24900603e+00
4.16873276e-01 -1.67581365e-01 -4.91908312e-01 6.36692762e-01
-2.06589058e-01 1.75735697e-01 8.81169140e-01 4.23476279e-01
9.70067680e-01 2.83605546e-01 4.90973800e-01 1.24779744e-02
6.20292604e-01 3.59122068e-01 9.39029276e-01 9.33248401e-02
6.09518066e-02 -1.66257575e-01 6.00778222e-01 -4.15770441e-01
-7.99131513e-01 -5.31892300e-01 -5.86670458e-01 6.15743697e-01
-1.73134208e-01 -5.55286348e-01 -4.31283385e-01 -1.06794167e+00
1.32990599e-01 5.10213315e-01 -4.34266537e-01 1.14924274e-01
-6.78697407e-01 -9.91677880e-01 8.97534311e-01 5.20119488e-01
8.52026284e-01 -1.16072774e+00 -1.85976297e-01 3.68086129e-01
-5.90073764e-01 -1.68338120e+00 -3.29182208e-01 -1.32379055e-01
-5.75672328e-01 -1.63176811e+00 -3.86455327e-01 -7.20047057e-01
5.29783428e-01 -1.02540120e-01 4.56270456e-01 -2.39807013e-02
-1.97929129e-01 -4.14818287e-01 -5.43744385e-01 -4.25713360e-01
-1.97782382e-01 1.31828696e-01 -2.96910614e-01 -6.86195642e-02
1.06090963e+00 -1.93155304e-01 8.59927759e-02 -9.41473525e-03
-8.88035119e-01 -4.08931747e-02 7.05108941e-01 4.50466782e-01
1.88084528e-01 1.36691406e-01 1.06256282e+00 -1.43803203e+00
1.18204916e+00 -3.73553604e-01 -4.85813230e-01 9.63510215e-01
-8.26605916e-01 2.14514986e-01 6.96722806e-01 1.19053036e-01
-1.56444597e+00 -1.71260521e-01 -2.24228576e-01 4.25082833e-01
3.66917849e-02 9.34083521e-01 -3.29619855e-01 4.03181136e-01
4.84170169e-01 6.34065480e-04 -5.14425933e-01 -4.45247918e-01
2.98462182e-01 1.41486084e+00 2.73131251e-01 -7.46395886e-01
5.76508880e-01 2.17225030e-01 -1.39170885e-01 -3.68697643e-01
-1.29376149e+00 -6.41964436e-01 -1.02303672e+00 1.00247353e-01
1.06422496e+00 -9.02205288e-01 -8.39634061e-01 4.34749991e-01
-1.30174923e+00 4.68625993e-01 2.18864545e-01 5.74728787e-01
-3.30773294e-02 6.44183278e-01 -8.11668038e-01 -8.31248462e-01
-5.08520961e-01 -1.07632208e+00 1.02787316e+00 2.19484624e-02
9.94754955e-03 -7.97051311e-01 8.10084864e-02 1.13268399e+00
4.18356173e-02 2.58022338e-01 1.32477403e+00 -1.02486777e+00
-4.94404048e-01 -3.43470365e-01 -7.19654381e-01 9.94611010e-02
5.09355783e-01 -1.52072206e-01 -8.19253266e-01 2.60547161e-01
3.81431058e-02 -5.14224827e-01 8.47757638e-01 9.02332179e-03
5.81814647e-01 -2.49915227e-01 -4.24165159e-01 1.56459048e-01
1.29969716e+00 8.49279940e-01 6.37498558e-01 3.06956530e-01
8.51059854e-01 5.84269166e-01 8.35074067e-01 3.75685155e-01
8.08541894e-01 4.79033858e-01 -4.40367103e-01 1.68534741e-01
2.67759949e-01 -2.26745069e-01 3.32130864e-02 9.69356418e-01
-2.77784407e-01 3.46797556e-02 -1.06026340e+00 4.36587870e-01
-1.93463969e+00 -1.09800315e+00 -2.36358538e-01 1.55192602e+00
1.08184564e+00 -6.48951605e-02 -1.28035292e-01 -6.51519299e-02
6.12669468e-01 -2.44439393e-01 -5.37732661e-01 -3.38137358e-01
-2.27273971e-01 1.71043769e-01 3.76398623e-01 1.26444176e-01
-1.20683217e+00 1.30682492e+00 5.24664116e+00 1.19595623e+00
-6.40094459e-01 1.61712199e-01 4.06947106e-01 4.19517040e-01
-2.31332287e-01 7.54272491e-02 -1.25475514e+00 4.87840146e-01
8.38941455e-01 -5.05715072e-01 -1.56996623e-01 5.96559823e-01
3.06428522e-01 -1.87105358e-01 -6.74874187e-01 8.64915192e-01
2.72744715e-01 -1.27927661e+00 2.07274109e-01 5.01053810e-01
7.61784256e-01 -1.30329803e-01 -2.63595313e-01 3.35427940e-01
2.97357947e-01 -8.58004451e-01 2.37133831e-01 5.23339868e-01
8.60610604e-01 -7.41238475e-01 9.70889211e-01 5.77914000e-01
-1.17140412e+00 -3.00762653e-01 -6.02274776e-01 2.16216743e-01
4.11288589e-01 9.59768891e-01 -7.98428297e-01 1.10643029e+00
2.42888659e-01 1.22205412e+00 -5.42732954e-01 9.32416677e-01
-6.98960185e-01 1.10696651e-01 -9.33639705e-03 -2.07409725e-01
1.91416338e-01 -6.48574173e-01 -2.53840089e-01 1.16673052e+00
1.99627399e-01 3.77806664e-01 2.06415966e-01 5.33454955e-01
-2.86714941e-01 7.45541155e-01 -7.71197855e-01 -5.21596432e-01
4.72150654e-01 1.20836484e+00 -6.24266803e-01 -6.07375145e-01
-5.86622536e-01 5.90021551e-01 3.32378328e-01 2.31370896e-01
-7.16398895e-01 -6.79515064e-01 2.28350777e-02 -4.67735708e-01
2.17774853e-01 -6.45706281e-02 -4.43041176e-01 -1.54912555e+00
4.84768987e-01 -9.04086053e-01 5.57374895e-01 -5.67527890e-01
-1.38560557e+00 8.23094666e-01 -1.56729168e-03 -1.29927623e+00
-2.92698115e-01 -4.45553511e-01 -1.95100293e-01 6.56416595e-01
-1.67214203e+00 -1.45836031e+00 2.43595704e-01 5.75665534e-01
2.52379239e-01 -4.22899932e-01 8.34458828e-01 7.63165653e-01
-8.22095573e-01 4.51978505e-01 1.50845036e-01 6.75826490e-01
8.78375113e-01 -8.17485511e-01 1.09195672e-01 7.24315524e-01
2.04781462e-02 1.06646502e+00 -1.37037069e-01 -1.21599185e+00
-9.84058499e-01 -8.91129076e-01 1.88341165e+00 -4.16261762e-01
8.07651222e-01 -2.17227355e-01 -5.12693226e-01 5.91788173e-01
2.01331958e-01 -5.79293311e-01 1.26404524e+00 2.23000944e-01
-5.26128888e-01 -9.89697874e-02 -1.25375164e+00 2.50701189e-01
9.75253761e-01 -8.89373183e-01 -1.08762741e+00 9.15063918e-01
4.13178980e-01 -4.99787508e-03 -1.04107988e+00 3.36310178e-01
5.24731636e-01 -3.91423136e-01 7.39743948e-01 -7.77142346e-01
6.02170944e-01 -1.45805374e-01 9.12604332e-02 -7.57877707e-01
1.76684231e-01 -2.03656390e-01 3.47502291e-01 1.29458141e+00
1.01142502e+00 -6.06078386e-01 4.12042707e-01 9.78382707e-01
-2.01014116e-01 -7.07495272e-01 -8.72402668e-01 -7.31095016e-01
-1.15901351e-01 -3.67045045e-01 6.86733484e-01 1.21123135e+00
5.19312322e-01 7.66532242e-01 -5.32156885e-01 -2.97702640e-01
5.86606085e-01 7.35177100e-01 3.74800086e-01 -1.25575602e+00
7.89404586e-02 2.22054273e-01 -4.40992340e-02 -8.02444696e-01
3.61052871e-01 -1.08932912e+00 -5.67735136e-01 -1.92218435e+00
7.25396514e-01 -6.04898334e-01 -2.61444226e-02 8.26233625e-01
-2.29543358e-01 -1.83944479e-01 8.40711370e-02 5.48006237e-01
-8.71165633e-01 6.44692421e-01 1.54597914e+00 -1.96272016e-01
9.28273425e-03 -8.09266195e-02 -1.07263112e+00 5.41891277e-01
6.28013194e-01 -8.14020097e-01 -4.41571385e-01 -5.20374775e-01
4.25291419e-01 1.85188428e-01 -3.61362576e-01 -3.58146220e-01
4.83739972e-01 -1.87839568e-01 2.24749967e-01 -8.86560977e-01
1.69334158e-01 -9.33715403e-01 -8.73272792e-02 2.12080911e-01
-2.06745982e-01 -2.70352270e-02 -1.96159586e-01 2.73876071e-01
-3.17800730e-01 -3.69098753e-01 1.35729954e-01 -3.29738855e-02
-4.28280860e-01 3.37099701e-01 -1.72594503e-01 2.78994262e-01
9.72779036e-01 -8.47018734e-02 -8.44401598e-01 1.39631167e-01
-6.13602996e-01 4.20738637e-01 -1.21114679e-01 4.34273839e-01
5.76119781e-01 -1.31570542e+00 -7.90344357e-01 1.43134892e-01
3.48339409e-01 -4.38531935e-01 9.88916010e-02 7.48723149e-01
-3.62969071e-01 8.75312328e-01 -1.59693241e-01 3.22798416e-02
-1.24250972e+00 4.10315812e-01 -2.50538826e-01 -7.56659806e-01
-4.36040342e-01 1.56254724e-01 -3.70038748e-01 -4.30891156e-01
-1.91491365e-01 -2.04903722e-01 -7.75777519e-01 4.09837306e-01
5.23641825e-01 1.17076881e-01 1.70840070e-01 -8.71146202e-01
-4.25968111e-01 7.61308789e-01 -5.74658215e-01 8.05238336e-02
1.70975649e+00 3.15244615e-01 -5.20172656e-01 -1.03011735e-01
1.27664614e+00 2.22742215e-01 -4.31635797e-01 -2.84652114e-01
6.83680296e-01 -5.22318304e-01 -4.79318619e-01 -1.03050864e+00
-9.04845774e-01 7.58590817e-01 -3.09174471e-02 -7.49957636e-02
9.16933894e-01 -1.69567093e-02 1.13005507e+00 8.91001284e-01
7.34767914e-01 -1.45110404e+00 -3.37740928e-01 5.58001876e-01
6.45818055e-01 -1.13768053e+00 3.22315305e-01 -8.66919219e-01
-9.10693645e-01 1.09263706e+00 2.14599907e-01 2.58413434e-01
7.17737198e-01 3.47996563e-01 3.48708391e-01 -6.04591787e-01
-4.58494037e-01 -2.72057772e-01 2.96210736e-01 5.08492768e-01
4.27969247e-01 -1.38299838e-01 -1.00576472e+00 9.23261225e-01
-5.88078387e-02 2.77072430e-01 3.15204859e-01 8.82634163e-01
-4.28537130e-01 -1.75154269e+00 1.95117984e-02 7.78554916e-01
-8.77457500e-01 -5.41028619e-01 -5.73709130e-01 5.71343243e-01
3.20817888e-01 1.34671664e+00 -7.06898391e-01 -1.27110645e-01
3.68414611e-01 2.53030837e-01 2.86964774e-01 -6.47751868e-01
-6.85374320e-01 2.82954216e-01 6.30524039e-01 -2.57791132e-01
-7.44414270e-01 -6.47735536e-01 -1.55890429e+00 1.68352142e-01
-4.99135673e-01 6.34676576e-01 4.99173880e-01 1.56725276e+00
2.94935465e-01 1.84673309e-01 5.65607607e-01 2.80129492e-01
-3.40771496e-01 -1.03172827e+00 -6.30341411e-01 4.58643138e-01
-2.84752458e-01 -5.10646105e-01 1.84928790e-01 5.31786941e-02] | [9.322656631469727, 8.709813117980957] |
bff2b062-c5eb-4995-9f8a-2de1f1ec85c8 | building-a-culture-of-reproducibility-in | 2212.13534 | null | https://arxiv.org/abs/2212.13534v1 | https://arxiv.org/pdf/2212.13534v1.pdf | Building a Culture of Reproducibility in Academic Research | Reproducibility is an ideal that no researcher would dispute "in the abstract", but when aspirations meet the cold hard reality of the academic grind, reproducibility often "loses out". In this essay, I share some personal experiences grappling with how to operationalize reproducibility while balancing its demands against other priorities. My research group has had some success building a "culture of reproducibility" over the past few years, which I attempt to distill into lessons learned and actionable advice, organized around answering three questions: why, what, and how. I believe that reproducibility efforts should yield easy-to-use, well-packaged, and self-contained software artifacts that allow others to reproduce and generalize research findings. At the core, my approach centers on self interest: I argue that the primary beneficiaries of reproducibility efforts are, in fact, those making the investments. I believe that (unashamedly) appealing to self interest, augmented with expectations of reciprocity, increases the chances of success. Building from repeatability, social processes and standardized tools comprise the two important additional ingredients that help achieve aspirational ideals. The dogfood principle nicely ties these ideas together. | ['Jimmy Lin'] | 2022-12-27 | null | null | null | null | ['culture'] | ['speech'] | [-2.95503587e-01 5.05717006e-03 -2.87957311e-01 -4.12265837e-01
-2.84404039e-01 -5.33385098e-01 3.94130766e-01 3.00582558e-01
-2.27483884e-01 8.25055838e-01 7.43398368e-01 -3.96164447e-01
-4.94704038e-01 -3.97517443e-01 -5.57712495e-01 -3.09406668e-01
5.90367079e-01 -1.55923516e-01 -4.85053629e-01 -3.70644718e-01
9.07841325e-01 2.99101651e-01 -1.34417307e+00 -2.87683636e-01
1.02759528e+00 1.07442832e-03 -1.18397690e-01 2.74116248e-01
-1.37963682e-01 1.26637506e+00 -6.11836255e-01 -7.46941090e-01
-1.33988783e-01 -6.52665973e-01 -8.35008562e-01 -3.34171325e-01
-8.07728991e-02 -1.58824474e-01 5.22760093e-01 8.86993945e-01
3.29332769e-01 -3.08506131e-01 5.98065369e-02 -1.23239398e+00
-1.16984046e+00 6.64971352e-01 -2.83554137e-01 1.41421640e-02
2.82575697e-01 5.58888376e-01 9.07932997e-01 -5.83788455e-01
7.19243288e-01 7.90176809e-01 8.90066385e-01 3.12721968e-01
-1.45756197e+00 -7.61360168e-01 -2.39109963e-01 -3.67436856e-01
-1.22921836e+00 -8.42393577e-01 2.18967661e-01 -8.51315856e-01
5.05181313e-01 5.00547886e-01 9.92877483e-01 1.06408119e+00
4.73611087e-01 -7.45397210e-02 1.23711896e+00 -3.21791232e-01
1.46866724e-01 5.87903142e-01 2.94185430e-01 1.72315225e-01
1.07733202e+00 -2.09143326e-01 -4.82579827e-01 -4.33680974e-02
8.21899951e-01 1.46801680e-01 -5.28524220e-01 1.98044509e-01
-1.45223868e+00 4.48741585e-01 -9.80644375e-02 9.49251413e-01
-3.55612874e-01 1.09564848e-01 2.65604347e-01 3.95450443e-01
3.85744780e-01 7.90252805e-01 -3.06258611e-02 -9.70130503e-01
-9.44662809e-01 -7.19079971e-02 1.02959049e+00 4.21383530e-01
3.24182808e-01 -2.98791736e-01 9.32611823e-02 6.08858645e-01
4.50025082e-01 4.41264778e-01 2.68223822e-01 -1.44704020e+00
-3.45767438e-01 6.72865152e-01 5.85747659e-01 -1.20697331e+00
-2.40201682e-01 -8.70960891e-01 -3.42616916e-01 3.71920913e-01
4.46674764e-01 -1.49341285e-01 1.85649842e-02 1.55929410e+00
7.45666772e-02 -4.45778936e-01 -1.21543951e-01 9.40699756e-01
4.86372799e-01 1.42958820e-01 3.79660279e-01 -1.62336200e-01
9.66447592e-01 -2.81163394e-01 -9.80297685e-01 1.64560482e-01
7.16010273e-01 -1.01962602e+00 1.24382424e+00 2.15759292e-01
-1.37884343e+00 -1.96156964e-01 -1.10229588e+00 -1.22894160e-01
-4.97349864e-03 -9.97059867e-02 5.14535308e-01 9.93831456e-01
-9.78705585e-01 8.69213283e-01 -5.12426138e-01 -6.49522305e-01
2.77935237e-01 -2.59148180e-01 -3.74523997e-01 3.03708643e-01
-6.60513937e-01 8.44914734e-01 -4.56275553e-01 3.22844297e-01
-2.02085689e-01 -8.57300758e-01 -1.31196663e-01 -1.97249860e-01
3.40267837e-01 -8.98298681e-01 8.82533848e-01 -9.22267318e-01
-1.14405465e+00 1.08412004e+00 -8.27878043e-02 1.83825612e-01
5.73873103e-01 -1.85976192e-01 -3.57605457e-01 -2.92696416e-01
5.63667178e-01 4.70428057e-02 2.21002307e-02 -1.10707271e+00
-2.46233895e-01 -4.41730231e-01 -3.58953118e-01 -3.21095347e-01
-4.13915277e-01 3.73353958e-01 4.37272102e-01 -2.18244433e-01
4.89626080e-03 -5.00782430e-01 6.07218072e-02 -3.16123366e-02
4.54866700e-03 -1.73069984e-01 3.75614285e-01 -4.40489113e-01
1.07261288e+00 -2.23542404e+00 -4.17664886e-01 2.66421158e-02
6.78297997e-01 -2.24966690e-01 3.26048881e-01 9.91076946e-01
2.68113464e-01 7.37287760e-01 5.87286472e-01 1.48577422e-01
3.36054862e-01 -1.97636068e-01 -2.27840960e-01 4.78888422e-01
-1.32166117e-01 8.74746740e-01 -1.19451904e+00 -1.81900725e-01
-4.52748805e-01 5.30705273e-01 -9.50244814e-02 -1.02170147e-02
4.02861744e-01 4.90481287e-01 -1.44436583e-01 3.79024535e-01
5.05254388e-01 -8.89727294e-01 5.44657826e-01 1.75958052e-01
-9.14855838e-01 2.74571449e-01 -9.11913335e-01 1.26651335e+00
-9.95639861e-02 5.91588318e-01 2.36483395e-01 -5.13496101e-01
1.29896033e+00 3.78813118e-01 2.22338855e-01 -8.54125023e-01
9.90568474e-02 4.42584723e-01 3.54895949e-01 -7.40180731e-01
5.60620964e-01 -1.34309843e-01 2.64503181e-01 1.09920835e+00
-4.46806312e-01 -2.13137612e-01 -1.53308421e-01 1.99274257e-01
7.67566979e-01 5.39780557e-01 1.64396316e-01 -5.58667541e-01
-3.19771585e-03 7.55357230e-03 6.66869998e-01 5.66179335e-01
-4.25073862e-01 1.23180620e-01 5.50744653e-01 -2.79920042e-01
-1.32409477e+00 -8.21969271e-01 2.49961410e-02 7.90488601e-01
-1.75052866e-01 -3.96216720e-01 -5.23395002e-01 1.89575419e-01
-1.07956819e-01 8.85557652e-01 -5.96041381e-01 2.33394757e-01
1.01563260e-02 -1.44443125e-01 3.10480177e-01 1.55531257e-01
4.43656474e-01 -6.42353714e-01 -8.64171386e-01 2.88678378e-01
-1.55631462e-02 -6.45591080e-01 -1.78006336e-01 -4.14044321e-01
-7.97729492e-01 -9.03016031e-01 -4.01903093e-01 -2.98772007e-01
3.49222809e-01 7.49739230e-01 8.67252171e-01 5.88343203e-01
-3.91188920e-01 4.51419413e-01 -1.66658670e-01 -6.23970926e-01
-5.64472139e-01 -5.14079630e-01 -2.20020398e-01 -3.72212499e-01
5.77330172e-01 -8.94176483e-01 -3.93834919e-01 2.47669026e-01
-6.61470592e-01 1.51813090e-01 5.59528530e-01 5.49009144e-01
1.22070327e-01 -5.35148799e-01 1.10057127e+00 -7.52623677e-01
7.54596114e-01 -4.77593273e-01 7.33575821e-02 4.38030452e-01
-1.18620157e+00 -4.35013950e-01 1.78398371e-01 4.78892811e-02
-7.63276756e-01 -1.05992913e+00 1.94196358e-01 2.59573460e-01
1.93525329e-01 3.49456370e-01 2.26663008e-01 -7.93148950e-02
7.99149930e-01 -1.66842312e-01 8.00134540e-01 -1.48022279e-01
1.86684996e-01 6.07732594e-01 1.26315698e-01 -7.45615304e-01
3.19687992e-01 3.59016538e-01 -3.35286736e-01 -8.55639040e-01
-5.61430216e-01 -1.27845883e-01 -4.19910327e-02 -4.23879266e-01
5.38970530e-01 -9.20195758e-01 -1.16436183e+00 1.25361914e-02
-7.30483890e-01 -3.40554893e-01 -5.17789602e-01 8.04016769e-01
-4.76374298e-01 2.66792655e-01 -3.02560568e-01 -8.19251001e-01
-3.80811393e-01 -6.97089136e-01 3.42918128e-01 6.05740964e-01
-8.64521742e-01 -9.23225403e-01 1.75435200e-01 7.93073297e-01
1.11211383e+00 6.16111159e-01 5.74807405e-01 -4.04302984e-01
-5.61669767e-01 -1.63134694e-01 -2.75032789e-01 2.30476275e-01
4.78921890e-01 5.44531345e-01 -6.30809307e-01 1.48990139e-01
5.02671361e-01 -2.67087191e-01 -1.89360276e-01 -7.17795193e-02
5.59428871e-01 -5.54202557e-01 -7.56110027e-02 1.22019857e-01
1.49814403e+00 -2.01528557e-02 5.60248494e-01 5.00864565e-01
3.46232593e-01 7.67651081e-01 4.14431483e-01 4.96620148e-01
6.20992243e-01 2.35753451e-02 -1.95397437e-01 1.71050385e-01
1.05449952e-01 -2.19838515e-01 -1.28973052e-01 8.74175131e-01
-2.43620187e-01 3.04728180e-01 -8.99385512e-01 5.89423299e-01
-1.69002318e+00 -1.15820730e+00 -5.57124197e-01 2.31284046e+00
9.87153351e-01 2.24148408e-01 2.35159189e-01 -2.25421488e-01
4.61967856e-01 -1.24174044e-01 -2.20656246e-01 -6.29934251e-01
1.11700661e-01 -8.77921060e-02 -6.70477971e-02 2.95988053e-01
-1.42861195e-02 4.53546524e-01 6.82263994e+00 1.51246944e-02
-1.10765231e+00 2.38696158e-01 6.23974562e-01 -9.24576968e-02
-1.02771461e+00 1.89654887e-01 -4.48588580e-01 4.42774355e-01
1.15774751e+00 -8.23882878e-01 4.91406992e-02 4.88716245e-01
8.87768090e-01 -2.05251873e-01 -9.61564422e-01 3.42086643e-01
-1.33989036e-01 -1.53821564e+00 -6.97542489e-01 1.00183681e-01
5.85883558e-01 -1.81609288e-01 1.07465215e-01 -3.62831622e-01
6.78554714e-01 -1.34042680e+00 7.49644637e-01 8.57480526e-01
3.83995712e-01 -2.78818607e-01 5.44600010e-01 7.15773180e-02
-9.43902880e-02 1.53516963e-01 -2.34168753e-01 -6.86483324e-01
-6.00483790e-02 9.81123865e-01 -6.99373782e-01 3.31347674e-01
6.34144545e-01 5.33721447e-01 -2.28262141e-01 1.06867087e+00
-2.35804334e-01 5.68710208e-01 4.72683124e-02 -4.36108619e-01
-1.52009204e-01 -3.91410559e-01 4.25967664e-01 1.02490902e+00
2.19495222e-01 4.72460836e-02 -7.10418582e-01 1.25562727e+00
2.52647221e-01 3.76004964e-01 -7.32590795e-01 -8.79814625e-01
7.99254715e-01 1.33947515e+00 -6.98851526e-01 1.08897857e-01
-4.13028210e-01 1.46249339e-01 1.35816798e-01 2.01785639e-01
-3.71030331e-01 -2.16146782e-01 6.56602740e-01 3.05286616e-01
-7.73936585e-02 -4.79258001e-01 -9.86196220e-01 -8.24048042e-01
-4.53509763e-02 -1.05138433e+00 -3.41031015e-01 -8.12356353e-01
-1.12085712e+00 -1.15423173e-01 -6.20045662e-01 -7.73576558e-01
2.25979373e-01 2.48194076e-02 -6.60969198e-01 1.11151230e+00
-8.44132900e-01 -1.12872362e+00 -4.65421319e-01 -2.39770308e-01
-2.65633225e-01 6.69169664e-01 7.49402642e-01 2.67220065e-02
-3.11800718e-01 6.17601871e-01 3.38368684e-01 -2.88450867e-01
1.03776515e+00 -6.58596814e-01 8.93243179e-02 4.53544408e-01
-2.56894261e-01 1.39511549e+00 7.68792868e-01 -6.10478938e-01
-1.59825122e+00 -4.08935636e-01 1.44523454e+00 -5.48166037e-01
1.10172951e+00 -9.98282284e-02 -8.17942977e-01 8.02832484e-01
3.53284031e-01 -7.39807963e-01 1.24850404e+00 4.65678185e-01
-3.61762166e-01 -8.22777823e-02 -1.06410336e+00 8.22946429e-01
1.02388525e+00 -5.64113021e-01 -4.84080762e-01 5.17190039e-01
8.11543345e-01 1.31039277e-01 -1.33978510e+00 -3.29741359e-01
8.93375933e-01 -1.26943791e+00 5.08939385e-01 -3.78878951e-01
5.36601365e-01 -1.39096409e-01 -5.28212916e-03 -7.64357805e-01
-2.96360642e-01 -9.77235556e-01 1.13347077e+00 1.62959242e+00
3.98022443e-01 -1.03173411e+00 6.16802275e-01 1.22944474e+00
-1.41958103e-01 -4.90207016e-01 -3.60019624e-01 -6.23821974e-01
1.58413365e-01 -1.40158087e-01 8.72008801e-01 1.46963191e+00
6.31772459e-01 1.82324752e-01 -1.36335671e-01 -3.04381192e-01
6.13102674e-01 6.34216294e-02 1.16416717e+00 -1.34770501e+00
-1.69186264e-01 -4.70892578e-01 -2.44881108e-01 -1.50203377e-01
-6.92085683e-01 -5.07790804e-01 -3.62342566e-01 -1.57326806e+00
5.49932837e-01 -6.41964197e-01 7.84142613e-02 2.77987003e-01
-5.28296977e-02 -1.16881674e-04 1.49155825e-01 6.24479949e-01
-4.11375791e-01 -1.22128166e-01 1.39668405e+00 3.27265322e-01
-4.04207140e-01 -5.20843327e-01 -1.94527471e+00 1.47352546e-01
8.51630449e-01 -2.60916948e-01 -3.09201658e-01 -1.73118770e-01
7.93768525e-01 -4.80812639e-02 5.77603400e-01 -7.99416602e-01
4.26008850e-01 -6.87849283e-01 9.72618312e-02 7.56129846e-02
-6.15790077e-02 -2.93582499e-01 1.12487555e+00 4.76552099e-01
-4.45850164e-01 -5.02231270e-02 3.30033414e-02 -1.54709026e-01
2.25626737e-01 -3.10906142e-01 3.80214781e-01 -1.82511196e-01
5.24575524e-02 -4.24185216e-01 7.79982135e-02 -2.16523483e-02
8.89082313e-01 -4.77096051e-01 -1.01651359e+00 -6.14840209e-01
-5.41782975e-01 5.47945425e-02 1.22582614e+00 1.39021546e-01
1.70466974e-01 -7.78350711e-01 -8.07294250e-01 -8.09571669e-02
-2.22144097e-01 -3.56592983e-01 2.50817239e-01 1.42934573e+00
-4.59414035e-01 3.02330285e-01 -3.48913163e-01 -4.02351588e-01
-7.98584580e-01 2.45409653e-01 1.35220945e-01 3.93525004e-01
-7.16929138e-01 8.39572430e-01 -2.82161504e-01 1.17386997e-01
-9.47550461e-02 1.08832709e-01 1.45674516e-02 6.50343671e-02
7.47555852e-01 6.17078125e-01 -1.24507956e-01 -7.89732635e-02
-1.72907904e-01 4.07069117e-01 -1.89586729e-02 -3.07151407e-01
1.60972989e+00 -2.84348160e-01 -8.60335588e-01 7.80354023e-01
7.67063618e-01 6.13293231e-01 -8.31410110e-01 3.45536262e-01
-1.73058230e-02 -9.38910902e-01 -3.34709227e-01 -1.10435891e+00
-2.75984228e-01 6.34941399e-01 8.14996138e-02 4.47062731e-01
5.81691980e-01 -2.68270165e-01 1.67190880e-01 -6.44894168e-02
3.28561872e-01 -1.14653051e+00 6.35495409e-02 -3.23569208e-01
9.32446241e-01 -7.75604784e-01 3.18150789e-01 -1.78734630e-01
-5.30929744e-01 1.10242784e+00 3.05059314e-01 2.62586623e-01
3.79408568e-01 1.14952594e-01 7.20537081e-02 -2.99227476e-01
-8.08962226e-01 3.36161524e-01 -6.90551102e-01 5.90070426e-01
1.04959071e+00 2.41193876e-01 -1.48196363e+00 4.82976288e-01
-2.74939597e-01 7.46610343e-01 9.04498279e-01 7.95029581e-01
-8.79118204e-01 -1.23370647e+00 -6.32885218e-01 1.73155770e-01
-6.49930477e-01 2.07872540e-01 -5.62676251e-01 1.00226724e+00
1.15852971e-02 1.24938679e+00 -1.63294554e-01 -1.57655686e-01
2.68430188e-02 -9.16074309e-03 2.62218356e-01 -7.45096058e-02
-7.33596921e-01 -4.91441451e-02 3.13792288e-01 -4.82473254e-01
-4.56112683e-01 -7.53413558e-01 -8.91811430e-01 -1.19986057e+00
1.69798192e-02 5.70761621e-01 1.02684045e+00 7.38296390e-01
8.92965198e-01 1.79459617e-01 5.85301638e-01 1.12907752e-01
-6.88073874e-01 -7.46334016e-01 -4.88232434e-01 1.58887655e-01
-1.77946985e-02 -1.38122946e-01 -3.37530881e-01 -2.44610712e-01] | [8.971681594848633, 6.321878433227539] |
b7b2d667-fbcc-4b56-bcc1-e42d2def7a87 | multi-scale-user-behavior-network-for-entire | 2208.01889 | null | https://arxiv.org/abs/2208.01889v2 | https://arxiv.org/pdf/2208.01889v2.pdf | Multi-Scale User Behavior Network for Entire Space Multi-Task Learning | Modelling the user's multiple behaviors is an essential part of modern e-commerce, whose widely adopted application is to jointly optimize click-through rate (CTR) and conversion rate (CVR) predictions. Most of existing methods overlook the effect of two key characteristics of the user's behaviors: for each item list, (i) contextual dependence refers to that the user's behaviors on any item are not purely determinated by the item itself but also are influenced by the user's previous behaviors (e.g., clicks, purchases) on other items in the same sequence; (ii) multiple time scales means that users are likely to click frequently but purchase periodically. To this end, we develop a new multi-scale user behavior network named Hierarchical rEcurrent Ranking On the Entire Space (HEROES) which incorporates the contextual information to estimate the user multiple behaviors in a multi-scale fashion. Concretely, we introduce a hierarchical framework, where the lower layer models the user's engagement behaviors while the upper layer estimates the user's satisfaction behaviors. The proposed architecture can automatically learn a suitable time scale for each layer to capture the dynamic user's behavioral patterns. Besides the architecture, we also introduce the Hawkes process to form a novel recurrent unit which can not only encode the items' features in the context but also formulate the excitation or discouragement from the user's previous behaviors. We further show that HEROES can be extended to build unbiased ranking systems through combinations with the survival analysis technique. Extensive experiments over three large-scale industrial datasets demonstrate the superiority of our model compared with the state-of-the-art methods. | ['Yong Yu', 'Zhewen Su', 'Zekun Zhu', 'Zaifan Jiang', 'Yuanbo Chen', 'Weinan Zhang', 'Xianyu Chen', 'Jiarui Jin'] | 2022-08-03 | null | null | null | null | ['survival-analysis'] | ['miscellaneous'] | [-9.91610065e-02 -7.56880462e-01 -5.28454661e-01 -5.52003741e-01
-3.12928587e-01 -2.96667427e-01 3.16512913e-01 7.44272843e-02
-2.96146929e-01 3.03796768e-01 2.49487296e-01 -2.06540972e-01
-2.59779036e-01 -7.86731243e-01 -5.04864872e-01 -5.52786350e-01
-2.75172383e-01 2.64649272e-01 2.65889674e-01 -4.81694132e-01
1.69354320e-01 -4.89046937e-03 -1.47727728e+00 1.78509474e-01
6.30108833e-01 1.16947317e+00 3.32391918e-01 5.54666519e-01
3.04186027e-02 7.85757601e-01 -4.87800151e-01 -1.69240102e-01
-4.28299867e-02 -4.43028271e-01 -1.91195533e-01 -1.06799930e-01
-3.36111069e-01 -2.77783841e-01 -4.56057101e-01 8.34380448e-01
1.31144121e-01 2.91720510e-01 6.09607577e-01 -1.00862801e+00
-8.91592741e-01 7.32117295e-01 -6.55707300e-01 2.92852730e-01
1.47421181e-01 -1.56275541e-01 1.39951408e+00 -5.43216825e-01
1.09611750e-01 1.05884635e+00 5.22993088e-01 3.61310631e-01
-1.06948185e+00 -6.17381036e-01 8.38980436e-01 2.39544332e-01
-1.20822477e+00 8.07588100e-02 9.31535959e-01 -3.03415149e-01
4.97511774e-01 4.16037619e-01 5.73989987e-01 1.55139458e+00
4.25736338e-01 1.19486761e+00 8.60454559e-01 1.47898132e-02
1.21154897e-01 3.87686342e-01 5.70809901e-01 3.07306081e-01
-4.02066596e-02 9.52293500e-02 -2.37606660e-01 -9.97461006e-02
7.59263873e-01 5.96715689e-01 1.80134341e-01 -4.76252995e-02
-9.90066111e-01 9.22377765e-01 5.06629944e-01 3.36396366e-01
-5.22046626e-01 9.85564515e-02 2.89830893e-01 3.13292533e-01
3.48759979e-01 1.38985157e-01 -7.30260789e-01 -2.16536328e-01
-6.27490044e-01 8.65548104e-02 8.72648954e-01 9.14853215e-01
6.05045795e-01 -1.76001713e-02 -5.05551040e-01 9.20907080e-01
4.83821362e-01 1.96935207e-01 8.21301818e-01 -2.25830540e-01
3.69589776e-01 6.14504457e-01 3.43427271e-01 -1.04467642e+00
-4.04412031e-01 -9.85089839e-01 -9.74618733e-01 -6.47064567e-01
2.06174970e-01 -8.68094712e-02 -7.40977108e-01 2.03801084e+00
2.80722547e-02 2.80300736e-01 -2.91800469e-01 7.15877891e-01
3.39115262e-01 8.48894179e-01 2.32336074e-01 -5.39498091e-01
1.26424050e+00 -7.94481158e-01 -7.82492697e-01 -1.23995833e-01
3.35722744e-01 -5.57593703e-01 1.21284080e+00 4.19384807e-01
-7.88659990e-01 -6.88628972e-01 -8.40555847e-01 2.63063908e-01
-2.57082194e-01 4.64779586e-01 9.79166508e-01 5.12975633e-01
-4.27120090e-01 6.51896417e-01 -8.28208745e-01 -8.84260684e-02
1.99887324e-02 3.67532283e-01 5.65151274e-01 4.25310194e-01
-1.65083158e+00 5.83800793e-01 -1.35057047e-01 4.04993773e-01
-8.69841158e-01 -4.34525788e-01 -3.68476927e-01 2.66523927e-01
5.22622049e-01 -3.29435796e-01 1.36816835e+00 -8.72579217e-01
-1.57761991e+00 -3.65126841e-02 -3.10319811e-01 -2.76186198e-01
4.45487410e-01 -3.03023338e-01 -8.07908177e-01 -6.52114391e-01
-8.03079605e-02 -2.47537881e-01 7.38836706e-01 -9.11710918e-01
-9.46160138e-01 -4.96356577e-01 -9.35018957e-02 2.71725893e-01
-6.85931444e-01 -1.55594125e-01 -8.12976182e-01 -6.98140621e-01
-7.50395432e-02 -1.00135016e+00 -3.23363900e-01 -7.67384827e-01
-5.23818493e-01 -5.52188516e-01 6.63969755e-01 -4.41569686e-01
1.98769808e+00 -2.10136890e+00 1.41578600e-01 3.98670971e-01
-8.60452652e-03 6.43095076e-02 -1.15902677e-01 5.78901231e-01
2.01944098e-01 1.36829019e-01 4.88754421e-01 -1.63627058e-01
1.31614268e-01 1.40308693e-01 -4.22337890e-01 2.55397707e-01
-1.82942048e-01 9.97452676e-01 -7.07373023e-01 5.69727086e-02
9.50412452e-02 4.91068423e-01 -5.59519529e-01 3.07573140e-01
-9.95065719e-02 3.32207650e-01 -1.00421548e+00 4.67997015e-01
1.69315785e-01 -7.46408463e-01 2.33732566e-01 -1.75710380e-01
-1.97794139e-01 4.07886803e-01 -1.20313966e+00 1.17822206e+00
-6.78441703e-01 -9.46515352e-02 -2.40752652e-01 -7.58070469e-01
8.03198457e-01 1.90022260e-01 5.31886637e-01 -8.36540997e-01
8.32378417e-02 8.49043578e-02 1.41983986e-01 -2.40332931e-01
4.59530354e-01 1.50872484e-01 -3.45299602e-01 5.66442609e-01
-1.12964891e-01 6.90809548e-01 1.90191224e-01 1.04566604e-01
9.95187581e-01 2.40853820e-02 3.69257450e-01 -7.47534111e-02
5.35369456e-01 -6.41488910e-01 8.19615841e-01 9.64796662e-01
-1.25533924e-01 7.43034296e-03 4.92776126e-01 -3.17043543e-01
-8.46728742e-01 -1.03257811e+00 -1.77795783e-01 1.52626657e+00
4.20909196e-01 -3.94230574e-01 -2.46033579e-01 -6.55290186e-01
4.92091589e-02 8.01780283e-01 -8.62953246e-01 -4.01344538e-01
-3.55654895e-01 -9.85593677e-01 1.30208004e-02 5.67758858e-01
4.30070311e-01 -1.01163733e+00 -1.50722235e-01 4.96286243e-01
-2.15885237e-01 -7.24063754e-01 -9.41991329e-01 2.32513949e-01
-7.47992694e-01 -6.60301626e-01 -4.95023429e-01 -4.81494427e-01
3.77105623e-01 1.97093651e-01 8.87818575e-01 -1.42552197e-01
-1.34784132e-01 2.03807905e-01 -2.75244266e-01 -2.22675741e-01
-2.22533867e-02 4.50070232e-01 3.15098315e-01 5.06797433e-01
5.69562137e-01 -6.15497530e-01 -9.86020088e-01 8.95332932e-01
-8.26445460e-01 -1.28807545e-01 7.64058352e-01 8.05863738e-01
5.55943489e-01 4.39538091e-01 7.92967916e-01 -1.13542330e+00
9.38858449e-01 -8.76405120e-01 -3.52884203e-01 4.55648690e-01
-9.03833270e-01 -4.82909866e-02 8.83038759e-01 -1.02329838e+00
-1.11749578e+00 -1.02824979e-01 -6.61557168e-02 -2.42095768e-01
1.05021849e-01 6.25821948e-01 2.48416383e-02 5.01342177e-01
1.31767660e-01 3.93982291e-01 -4.21885908e-01 -6.17805600e-01
2.00414166e-01 7.58188009e-01 8.13556612e-02 -3.85816544e-01
6.52075231e-01 1.55044375e-02 -3.79490614e-01 -5.17227471e-01
-1.00795496e+00 -7.43572354e-01 -3.31044644e-01 -2.21337304e-01
5.98391056e-01 -9.83224869e-01 -1.35992062e+00 2.96370536e-01
-7.51775086e-01 1.98347196e-02 4.86451127e-02 6.05664968e-01
-2.41651058e-01 -6.50013760e-02 -1.04807377e+00 -1.16733158e+00
-1.84496865e-01 -9.88484323e-01 9.73455191e-01 4.06664699e-01
-2.39500031e-01 -1.20790410e+00 -1.40514765e-02 1.57088086e-01
4.97578293e-01 -3.78812283e-01 9.19765353e-01 -9.83268678e-01
-5.77230632e-01 -4.77243334e-01 -5.48506677e-02 3.18189025e-01
4.21206832e-01 -1.72244042e-01 -5.22054970e-01 -2.05597982e-01
2.72304535e-01 -4.34973538e-02 5.90188682e-01 5.13976872e-01
1.17077899e+00 -4.78810549e-01 -2.76650101e-01 3.27441007e-01
1.33460164e+00 5.40337443e-01 4.55540925e-01 -4.70549660e-03
6.89294875e-01 4.16365296e-01 6.99175477e-01 6.60934448e-01
4.40931588e-01 8.73428762e-01 2.82681733e-01 -1.59522928e-02
6.25242829e-01 -7.95266271e-01 6.70768440e-01 1.23639190e+00
-1.64566040e-01 -3.83661896e-01 -2.11866543e-01 6.56800568e-02
-2.16486025e+00 -1.06375241e+00 -1.65031925e-01 2.41886353e+00
6.22815669e-01 1.82773486e-01 2.34416947e-01 -2.13915870e-01
7.43276119e-01 2.01139137e-01 -9.58894193e-01 -2.00139225e-01
1.97378859e-01 -2.16803014e-01 5.15779078e-01 3.01315635e-01
-1.00778651e+00 7.67876804e-01 6.14286900e+00 1.00261676e+00
-8.64758492e-01 1.00952469e-01 7.57574499e-01 -1.72977552e-01
-4.64705229e-01 -3.09407681e-01 -1.28700018e+00 8.41278374e-01
1.01051104e+00 -1.41757950e-01 7.41217554e-01 8.01793158e-01
4.71893460e-01 2.31699258e-01 -1.10648441e+00 7.02535808e-01
-2.37155929e-02 -6.85464203e-01 2.33335029e-02 2.91660339e-01
5.38362265e-01 -1.30088896e-01 4.12572920e-01 8.54391634e-01
5.23179293e-01 -6.97678685e-01 5.21477938e-01 7.07044244e-01
4.06282812e-01 -6.59631133e-01 4.62519467e-01 6.66553140e-01
-1.49956870e+00 -4.46906209e-01 -3.15739870e-01 2.82816906e-02
3.33892107e-01 6.61588192e-01 -4.18777406e-01 4.51903969e-01
4.89851058e-01 1.03817570e+00 -5.13160706e-01 6.19822741e-01
-1.67847455e-01 8.20464253e-01 -1.41537294e-01 -5.58788657e-01
2.78804839e-01 -5.60204327e-01 4.14649755e-01 9.56108749e-01
5.29099822e-01 2.99712680e-02 1.79510415e-01 8.53953958e-01
7.13882446e-02 2.89194047e-01 -2.75671244e-01 -3.23076367e-01
2.02154025e-01 1.33388841e+00 -4.42527533e-01 -1.76429972e-01
-6.72241569e-01 1.02409399e+00 1.72350258e-01 7.29343057e-01
-1.09891176e+00 -3.07937473e-01 6.78543806e-01 5.66986464e-02
3.50947440e-01 -9.35660377e-02 -5.69074824e-02 -1.39339292e+00
7.17399046e-02 -5.57498336e-01 4.18107450e-01 -4.78623390e-01
-1.54300249e+00 4.71377522e-01 -4.84499991e-01 -1.37973166e+00
-1.55274197e-01 -4.72361565e-01 -5.84053636e-01 9.30773139e-01
-1.34270513e+00 -9.19024706e-01 7.84289613e-02 7.07024634e-01
7.17733920e-01 -1.40240297e-01 6.11203074e-01 4.82080489e-01
-9.00584280e-01 5.07802665e-01 2.32667685e-01 -8.85180309e-02
4.94398892e-01 -1.08354354e+00 1.74202889e-01 4.32685822e-01
2.44010285e-01 1.20757163e+00 6.07577324e-01 -7.98112035e-01
-1.36491883e+00 -1.09602714e+00 9.23105299e-01 -5.01518428e-01
9.28024590e-01 -5.52323461e-01 -8.21295083e-01 8.93852234e-01
-3.55289549e-01 -3.83610845e-01 8.51154387e-01 7.79169500e-01
-2.42253020e-01 -3.62491280e-01 -5.59134245e-01 7.30453789e-01
1.00435627e+00 -4.75319356e-01 -3.79958421e-01 2.58819461e-01
6.63526654e-01 -3.86663117e-02 -7.83836961e-01 4.23069030e-01
8.28813612e-01 -8.27825665e-01 8.96168172e-01 -8.08535933e-01
3.98162752e-01 -6.17831796e-02 -1.95296094e-01 -1.26206088e+00
-7.94074714e-01 -6.04593158e-01 -4.42132890e-01 1.24203944e+00
5.49553216e-01 -6.76275611e-01 6.24474943e-01 7.11871088e-01
1.38522193e-01 -9.79129255e-01 -6.42550290e-01 -5.01791537e-01
-4.44047898e-01 -4.11921680e-01 6.47817373e-01 5.53438365e-01
6.02166802e-02 7.43294358e-01 -1.08381259e+00 9.77922007e-02
2.89346606e-01 3.03347349e-01 4.12328959e-01 -1.30264974e+00
-5.74199140e-01 -4.20285970e-01 2.12205537e-02 -1.61002016e+00
-1.65445000e-01 -6.19235754e-01 -3.82658467e-02 -1.08397114e+00
4.93914217e-01 -4.80112880e-01 -1.11099362e+00 1.66226596e-01
-2.59461641e-01 -2.36085132e-01 -9.36034918e-02 5.10419607e-01
-9.57481384e-01 6.83272839e-01 1.32597792e+00 2.29764685e-01
-5.39700449e-01 8.31900954e-01 -1.01044059e+00 4.81651336e-01
4.74519789e-01 -2.39481971e-01 -6.65763080e-01 -1.01818461e-02
6.37288034e-01 3.17718327e-01 -5.57411350e-02 -6.43264890e-01
2.30644479e-01 -2.68370390e-01 3.92585725e-01 -4.92090553e-01
3.29111904e-01 -9.07993972e-01 1.52438864e-01 3.48822206e-01
-8.16551387e-01 1.45069391e-01 -3.27744722e-01 1.19029355e+00
-1.31537080e-01 1.57205433e-01 3.40461165e-01 7.64655173e-02
-3.97706002e-01 7.78869748e-01 -3.95228833e-01 -3.03569436e-01
9.49200213e-01 1.98787510e-01 -3.75143401e-02 -5.96190631e-01
-7.94289052e-01 4.70706433e-01 -6.89730346e-02 9.38193917e-01
3.61240923e-01 -1.49498022e+00 -2.48727515e-01 2.42551282e-01
7.44004175e-02 -6.82814062e-01 5.68205118e-01 6.86198652e-01
4.09372509e-01 6.26727700e-01 1.94392577e-01 -4.22233343e-01
-1.02977061e+00 5.35666466e-01 5.22964075e-02 -7.51319408e-01
-1.94894552e-01 5.79464495e-01 4.33840632e-01 -2.52642989e-01
3.07818472e-01 -3.10927302e-01 -7.21311927e-01 2.64534712e-01
4.50037330e-01 3.74582767e-01 -2.77827471e-01 -4.70970541e-01
-1.22208983e-01 4.53084618e-01 -6.00426733e-01 6.43682480e-02
1.27329421e+00 -4.05972213e-01 3.08383018e-01 1.29630053e+00
1.23716259e+00 -1.12049490e-01 -1.19116282e+00 -5.05389690e-01
-1.47818699e-01 -3.76322150e-01 -1.69702861e-02 -9.04901206e-01
-1.04051507e+00 6.41938746e-01 4.54849482e-01 5.44091940e-01
1.04468143e+00 -1.56659171e-01 9.99073446e-01 5.43432385e-02
5.55653811e-01 -1.38475204e+00 1.65606603e-01 3.98321390e-01
5.85324168e-01 -1.17224634e+00 -2.59418488e-01 -2.06640825e-01
-1.01410127e+00 8.43596756e-01 6.22985959e-01 -8.22786689e-02
9.59277093e-01 -1.01243190e-01 -3.88969146e-02 -2.22151782e-02
-1.16682959e+00 -3.96190723e-03 4.45987135e-01 -1.44312710e-01
7.27784097e-01 3.05210173e-01 -4.67972577e-01 1.23785198e+00
3.38476896e-02 -1.60833299e-02 1.26740620e-01 4.04216498e-01
-1.75395563e-01 -1.15383530e+00 2.89541662e-01 8.14641356e-01
-5.45885980e-01 -9.35467985e-03 7.65220746e-02 4.75489885e-01
-1.18751958e-01 8.09579551e-01 1.22035444e-02 -7.98990905e-01
5.08508623e-01 -5.67441955e-02 -1.20078377e-01 -6.67350769e-01
-7.35261738e-01 4.19983596e-01 -2.41448075e-01 -7.11190522e-01
-5.59921376e-02 -8.39451790e-01 -9.26478088e-01 -1.33118734e-01
-3.49410564e-01 3.17856342e-01 6.17903173e-01 8.31172168e-01
3.06554675e-01 8.98206711e-01 1.16173816e+00 -4.93521988e-01
-9.99147713e-01 -1.05976367e+00 -1.00079918e+00 6.94082499e-01
2.71972895e-01 -6.76850557e-01 -5.63698173e-01 -3.08803231e-01] | [10.102543830871582, 5.511308670043945] |
6184b7d5-421a-42e8-a06c-969ca5042382 | learning-reward-machines-a-study-in-partially | 2112.09477 | null | https://arxiv.org/abs/2112.09477v1 | https://arxiv.org/pdf/2112.09477v1.pdf | Learning Reward Machines: A Study in Partially Observable Reinforcement Learning | Reinforcement learning (RL) is a central problem in artificial intelligence. This problem consists of defining artificial agents that can learn optimal behaviour by interacting with an environment -- where the optimal behaviour is defined with respect to a reward signal that the agent seeks to maximize. Reward machines (RMs) provide a structured, automata-based representation of a reward function that enables an RL agent to decompose an RL problem into structured subproblems that can be efficiently learned via off-policy learning. Here we show that RMs can be learned from experience, instead of being specified by the user, and that the resulting problem decomposition can be used to effectively solve partially observable RL problems. We pose the task of learning RMs as a discrete optimization problem where the objective is to find an RM that decomposes the problem into a set of subproblems such that the combination of their optimal memoryless policies is an optimal policy for the original problem. We show the effectiveness of this approach on three partially observable domains, where it significantly outperforms A3C, PPO, and ACER, and discuss its advantages, limitations, and broader potential. | ['Sheila A. McIlraith', 'Margarita P. Castro', 'Richard Valenzano', 'Toryn Q. Klassen', 'Ethan Waldie', 'Rodrigo Toro Icarte'] | 2021-12-17 | null | null | null | null | ['problem-decomposition'] | ['miscellaneous'] | [ 1.70005828e-01 5.49330950e-01 -4.43707138e-01 1.04302960e-02
-7.72664845e-01 -6.36654496e-01 6.14623308e-01 -2.92276423e-02
-7.60264218e-01 1.12560165e+00 1.01719268e-01 -3.15221757e-01
-3.55802357e-01 -5.34392118e-01 -6.77354276e-01 -8.75442386e-01
-4.58446860e-01 7.58667350e-01 -7.75342062e-02 -2.31485337e-01
3.29438448e-01 3.59477073e-01 -1.61709011e+00 -7.02276081e-02
8.73948038e-01 9.96169746e-01 5.94401836e-01 9.41316783e-01
1.64885744e-01 1.32232559e+00 -6.94824636e-01 4.87894028e-01
3.35826069e-01 -5.35047829e-01 -1.24462616e+00 4.80874002e-01
-4.53482270e-01 -3.38974565e-01 -1.61982626e-01 1.06498277e+00
2.47708023e-01 4.47622269e-01 6.51113331e-01 -1.28215635e+00
-3.81989390e-01 4.17613149e-01 2.39740927e-02 -1.41030937e-01
5.20563602e-01 4.67700303e-01 1.15032375e+00 -3.99825629e-03
5.68783343e-01 1.39998865e+00 -1.96237803e-01 8.47546995e-01
-1.53915608e+00 1.50620446e-01 3.50374848e-01 1.12799227e-01
-8.19645524e-01 -3.97288263e-01 4.29262072e-01 -9.20213237e-02
1.36377394e+00 2.11757883e-01 7.06124723e-01 8.94642889e-01
2.41903618e-01 1.14599502e+00 1.43998146e+00 -5.00754118e-01
1.02875328e+00 1.10096529e-01 -3.14443409e-01 7.12399423e-01
-6.83511347e-02 5.78902543e-01 -2.30070725e-01 -3.72176290e-01
8.17084968e-01 -1.10832319e-01 -1.75262094e-01 -7.00745344e-01
-1.12083089e+00 9.00443912e-01 8.88006389e-02 1.51956588e-01
-7.03539073e-01 3.76123995e-01 9.61862877e-03 8.77061188e-01
-1.15833774e-01 1.17244101e+00 -7.37624288e-01 -2.55647749e-01
-1.83756679e-01 5.19526184e-01 1.10768163e+00 6.02706075e-01
8.59272063e-01 3.93557668e-01 -6.14724942e-02 5.69220603e-01
2.97315031e-01 8.06764066e-01 5.12010932e-01 -1.76544917e+00
2.56051153e-01 3.93405348e-01 8.53291988e-01 -3.19095373e-01
-2.79618710e-01 -2.34666653e-02 -1.54899985e-01 6.78113222e-01
1.40041739e-01 -4.94277477e-01 -6.09578490e-01 1.80100489e+00
9.61933061e-02 1.49660960e-01 7.38802016e-01 8.80725324e-01
6.59107510e-03 8.78686547e-01 -6.59731627e-02 -6.03813589e-01
8.23164880e-01 -9.64554787e-01 -6.46321774e-01 -5.64257264e-01
6.79230154e-01 -4.09031985e-03 9.08538401e-01 5.50286114e-01
-1.19539237e+00 -5.80330156e-02 -1.12546945e+00 5.67614794e-01
1.53212657e-03 -1.81922570e-01 5.25511861e-01 1.54663816e-01
-1.30696797e+00 8.95390749e-01 -8.08388472e-01 -1.02047443e-01
1.83852371e-02 7.32084453e-01 5.11988290e-02 1.99695870e-01
-1.09875333e+00 1.14473224e+00 7.14451730e-01 -3.06738317e-01
-1.56787884e+00 1.32646799e-01 -8.64225924e-01 -5.06955162e-02
1.01145089e+00 -4.70967114e-01 2.06810904e+00 -1.34648061e+00
-2.15771818e+00 5.19124925e-01 -5.78670762e-02 -7.09036648e-01
7.40480572e-02 -5.79787120e-02 -2.71109968e-01 3.02136570e-01
9.34155583e-02 5.69133162e-01 1.02213871e+00 -1.31369519e+00
-1.12591898e+00 -8.85618553e-02 5.50187290e-01 6.87201679e-01
-2.44361535e-02 -1.85913607e-01 -1.54188901e-01 -1.85869247e-01
-3.45903784e-01 -1.17991149e+00 -8.30080092e-01 -6.76253736e-01
2.29259115e-02 -4.31380868e-01 6.82338834e-01 -1.08007587e-01
8.72660697e-01 -1.85249782e+00 6.03114367e-01 1.95119366e-01
4.08437252e-02 1.24262966e-01 -5.59898734e-01 4.13632721e-01
3.31903845e-01 -1.15549713e-02 -6.96001276e-02 2.31845360e-02
1.91768423e-01 7.32342303e-01 -4.00109380e-01 3.16285282e-01
1.80927411e-01 9.04048085e-01 -1.19772625e+00 -2.57538855e-01
1.16764814e-01 -2.92336583e-01 -3.95874709e-01 8.35748732e-01
-1.12776339e+00 5.70931077e-01 -9.53705132e-01 3.33053499e-01
-2.22959489e-01 -3.01951289e-01 6.16196513e-01 9.81911898e-01
-8.51374716e-02 2.69182503e-01 -1.30467832e+00 1.39241099e+00
-3.88111681e-01 2.31093630e-01 2.88050234e-01 -1.24997449e+00
9.33339775e-01 5.87197244e-01 6.98070049e-01 -8.32547128e-01
6.50389493e-02 1.84029430e-01 -9.49140117e-02 -6.20006144e-01
3.18478286e-01 -4.61732373e-02 -3.68899643e-01 8.98744106e-01
-5.28214090e-02 -3.30591410e-01 1.51134849e-01 -1.10755853e-01
1.39562249e+00 9.37497243e-02 5.32138288e-01 -4.94445451e-02
4.88788188e-01 1.63411364e-01 6.24926090e-01 1.35548925e+00
-1.45603374e-01 -3.19648117e-01 7.12691784e-01 -3.86316806e-01
-8.47042263e-01 -8.47852826e-01 6.40681624e-01 1.16392183e+00
1.71516970e-01 3.42062190e-02 -6.11619890e-01 -7.95775652e-01
-6.76453784e-02 8.99217904e-01 -4.78088289e-01 -1.67902097e-01
-6.22044086e-01 -2.45348677e-01 -6.73039109e-02 2.57514715e-01
4.91293252e-01 -1.88024199e+00 -1.23277783e+00 5.38780272e-01
-8.02371092e-03 -7.83012807e-01 -5.02674162e-01 5.54389000e-01
-9.83708620e-01 -1.06725657e+00 -4.61643070e-01 -8.09126496e-01
6.61898136e-01 1.83997557e-01 1.15843880e+00 3.82593013e-02
1.58049598e-01 9.20854747e-01 -2.20626935e-01 -3.17484379e-01
-6.52369022e-01 -9.84416083e-02 2.97405928e-01 -9.07014832e-02
-1.16146944e-01 -3.27996433e-01 -2.62186646e-01 1.73800081e-01
-9.09408391e-01 -1.34370878e-01 7.91291237e-01 8.42033863e-01
6.12023950e-01 2.80574918e-01 6.69110596e-01 -7.66725600e-01
1.03894699e+00 -3.21247995e-01 -9.86171901e-01 4.54003036e-01
-7.27231264e-01 9.13022220e-01 9.54814374e-01 -5.43912113e-01
-9.48469639e-01 3.11185658e-01 4.79985684e-01 -2.74726778e-01
-3.51150721e-01 3.17720920e-01 -3.31039071e-01 9.26097184e-02
4.75588232e-01 6.39244795e-01 1.84979454e-01 -2.49301389e-01
4.36408311e-01 6.90683842e-01 4.13262397e-01 -1.06313539e+00
3.08574110e-01 -3.07546239e-02 2.54309084e-02 -6.44098341e-01
-8.24471354e-01 -3.44198406e-01 -1.78534284e-01 -2.00042546e-01
6.03697479e-01 -5.48308969e-01 -1.19788945e+00 9.10524055e-02
-9.01504397e-01 -8.46421897e-01 -7.30827272e-01 2.99684763e-01
-1.51785564e+00 2.75569260e-02 -3.15351129e-01 -1.20605636e+00
-4.74277548e-02 -1.30680263e+00 5.87866187e-01 3.53255749e-01
-1.63182557e-01 -8.83175910e-01 3.61045837e-01 7.57754818e-02
1.80583701e-01 8.59688818e-02 1.04265165e+00 -6.82969987e-01
-6.49359226e-01 2.47853532e-01 3.98618877e-01 3.03880692e-01
2.06019869e-03 -5.39043367e-01 -5.14271200e-01 -6.60627484e-01
8.36194679e-02 -8.71177971e-01 3.86671722e-01 3.08845371e-01
9.36437368e-01 -9.08827305e-01 -1.05994612e-01 1.17781430e-01
1.45297718e+00 8.28177989e-01 2.71231711e-01 6.74318910e-01
-1.24774545e-01 3.37176502e-01 9.39394057e-01 5.49607217e-01
2.62219548e-01 5.75171173e-01 6.82350159e-01 4.00363982e-01
6.67278886e-01 -2.18570754e-01 7.67780483e-01 2.54635721e-01
-1.81656554e-01 -2.47225724e-02 -6.38485968e-01 3.47588569e-01
-2.52361679e+00 -1.13303864e+00 7.87569523e-01 2.29265332e+00
8.63959134e-01 1.54812410e-01 4.22930837e-01 -1.57513738e-01
3.64469141e-01 1.33699134e-01 -1.18228233e+00 -8.70531678e-01
1.30704865e-01 1.68757185e-01 3.50244254e-01 6.72240913e-01
-9.89545941e-01 1.03192151e+00 7.27230549e+00 4.27693963e-01
-9.33248341e-01 -1.83868021e-01 4.30739105e-01 1.74476765e-02
-1.36922807e-01 1.06071927e-01 -4.74055737e-01 1.12484328e-01
1.19593012e+00 -2.81356901e-01 1.30329227e+00 1.03538167e+00
5.12446761e-01 -3.69826794e-01 -1.47613752e+00 6.39064968e-01
-4.40567076e-01 -1.13468325e+00 -2.75891393e-01 2.26207867e-01
9.07435179e-01 -2.90218485e-03 2.00695246e-01 7.75260627e-01
1.19900441e+00 -1.18617213e+00 7.24727035e-01 4.26380217e-01
3.91909242e-01 -9.67023671e-01 3.60858321e-01 9.69476104e-01
-8.34473789e-01 -7.11685240e-01 -3.21506411e-01 -4.59011763e-01
-3.56573611e-01 -2.99023807e-01 -1.01375425e+00 2.15357378e-01
2.24325284e-01 4.85303372e-01 8.83398056e-02 7.25412011e-01
-5.75692594e-01 3.31905186e-01 -2.22188588e-02 -6.60661459e-01
6.44245148e-01 -3.78778070e-01 5.44309080e-01 7.41993308e-01
5.17873047e-03 4.49966788e-01 8.47679257e-01 7.95359254e-01
1.86722115e-01 -2.44229361e-01 -8.45521152e-01 -3.55882376e-01
4.13273036e-01 9.18271244e-01 -4.16657090e-01 -2.94107080e-01
-2.09555015e-01 8.93888593e-01 9.18160737e-01 5.49805999e-01
-4.18372601e-01 -1.33001193e-01 7.74995565e-01 -4.46440816e-01
3.56077373e-01 -1.59078270e-01 2.47253940e-01 -1.02635396e+00
-4.23523933e-01 -1.43101418e+00 4.78136361e-01 -6.33139968e-01
-8.01049531e-01 4.37605560e-01 -2.16012016e-01 -1.11687160e+00
-9.30448234e-01 -3.81289452e-01 -3.93173069e-01 5.49966514e-01
-1.29840446e+00 -3.19736719e-01 4.73356247e-01 6.16220772e-01
5.45783401e-01 -4.89816666e-01 1.15028608e+00 -7.22559690e-01
-4.16868418e-01 -7.61708096e-02 3.71314496e-01 -1.73038825e-01
-9.96003076e-02 -1.78088856e+00 -1.66901797e-01 4.80625719e-01
5.18165901e-02 8.07157829e-02 8.48078728e-01 -3.87506574e-01
-1.80495560e+00 -1.06236577e+00 4.58566695e-01 -1.69055417e-01
7.47708678e-01 1.36812508e-01 -3.88040841e-01 7.27607727e-01
2.22569212e-01 -1.31944433e-01 1.18645392e-01 -5.16873263e-02
8.81553888e-02 2.82667447e-02 -1.10119045e+00 8.46405029e-01
6.69936717e-01 -3.39826673e-01 -8.31135392e-01 1.28724128e-01
8.10996532e-01 -3.50951880e-01 -5.92185915e-01 -7.41539374e-02
1.06653437e-01 -5.47199786e-01 7.14139819e-01 -1.23486173e+00
3.41712713e-01 -2.80891806e-01 -8.62204134e-02 -1.77121222e+00
-3.09520513e-01 -1.14475942e+00 -5.94723582e-01 3.62537384e-01
3.12812179e-01 -7.72672176e-01 7.38925397e-01 6.86478496e-01
2.28496298e-01 -7.86610901e-01 -9.42625999e-01 -1.01748395e+00
-1.49042383e-01 -1.02648109e-01 5.17961681e-01 4.64074224e-01
3.39927673e-01 4.18247908e-01 -4.59254533e-01 1.99808359e-01
6.20112658e-01 4.55956966e-01 5.82682371e-01 -8.45947206e-01
-8.50209296e-01 -2.41978034e-01 1.97056234e-01 -1.42926276e+00
5.97768724e-01 -6.75234139e-01 6.10262513e-01 -1.40078819e+00
3.55470665e-02 -4.29210246e-01 -4.33109969e-01 7.29709625e-01
2.44131118e-01 -8.56548727e-01 4.32396650e-01 2.44448677e-01
-1.40957117e+00 5.70806384e-01 1.46283984e+00 -2.96687037e-01
-6.69084191e-01 3.34621906e-01 -7.00753927e-01 6.90941691e-01
1.01014614e+00 -5.41941881e-01 -5.80614150e-01 -1.95571184e-01
5.18923663e-02 1.10402882e+00 6.16715439e-02 -6.59150422e-01
1.16588175e-01 -1.03977752e+00 -8.87063369e-02 -5.30078299e-02
2.32200235e-01 -8.73573422e-01 -1.41240269e-01 8.59165788e-01
-8.43863845e-01 1.65975586e-01 -2.47041807e-01 6.85809374e-01
4.69267666e-02 -5.93391299e-01 6.37732506e-01 -5.12568116e-01
-7.93334782e-01 1.78861231e-01 -8.93721402e-01 2.92197526e-01
1.30274081e+00 3.36273044e-01 -7.42069706e-02 -7.95016468e-01
-8.92634153e-01 7.34487891e-01 3.13042104e-01 2.06036925e-01
7.45735466e-01 -9.95291412e-01 -3.98817033e-01 -5.62583916e-02
-1.01056203e-01 -1.42604351e-01 -4.40869004e-01 2.69453347e-01
9.50190648e-02 8.05500925e-01 -1.28894940e-01 -1.90230280e-01
-1.05081713e+00 6.35654390e-01 6.46750689e-01 -8.05311322e-01
-4.84909624e-01 2.46062994e-01 -2.02592388e-01 -5.13846993e-01
4.40275282e-01 -2.67628223e-01 -4.23253983e-01 -6.03900552e-01
4.69846308e-01 1.49123490e-01 -4.09014970e-01 -4.41637859e-02
6.01703301e-02 -8.39328617e-02 -1.46678612e-02 -6.30843520e-01
1.58446598e+00 -3.09653897e-02 -4.09922525e-02 3.53648871e-01
8.08942318e-01 -6.02094769e-01 -1.64888752e+00 -5.59597731e-01
4.82289344e-01 -1.98878616e-01 -7.60169234e-03 -9.08051908e-01
-7.12101579e-01 2.18908489e-01 3.16738278e-01 4.89961416e-01
1.00304592e+00 5.47038428e-02 4.31484699e-01 1.06901002e+00
8.22979927e-01 -1.55168140e+00 5.78639627e-01 8.66813064e-01
8.07648540e-01 -9.90935981e-01 -2.70413220e-01 5.01915455e-01
-1.03133965e+00 1.01033282e+00 6.35303140e-01 -3.53392780e-01
2.53474891e-01 3.16901892e-01 -2.11501434e-01 -1.84440939e-03
-1.38404620e+00 -6.10758841e-01 -1.09630702e-02 8.11778724e-01
-3.15026641e-01 3.23314309e-01 -2.05604769e-02 3.08734626e-01
1.95470564e-02 1.43298954e-01 5.32861412e-01 1.37306690e+00
-1.01790214e+00 -1.38955808e+00 -4.49223906e-01 4.82159913e-01
-1.39023498e-01 5.65304756e-01 -4.49804306e-01 3.93900365e-01
-1.89588577e-01 1.18800879e+00 -1.31989449e-01 -2.20026225e-01
2.31969267e-01 -1.40071854e-01 6.61424339e-01 -8.26547086e-01
-2.42319867e-01 -4.59207930e-02 2.03690395e-01 -9.14581597e-01
-3.59838933e-01 -7.17292845e-01 -1.54933226e+00 2.61865500e-02
2.32504621e-01 5.48048198e-01 2.70918489e-01 1.29877102e+00
1.44539297e-01 3.51183653e-01 1.16390419e+00 -6.09333634e-01
-1.35039008e+00 -3.85663748e-01 -6.84289575e-01 2.00375959e-01
6.36653185e-01 -4.92960542e-01 -3.03688347e-01 -2.25287974e-01] | [4.116504669189453, 1.9092142581939697] |
bcf89ec1-8c8a-4eb6-9131-52fb778b0ec9 | autoregressive-stylized-motion-synthesis-with | null | null | http://openaccess.thecvf.com//content/CVPR2021/html/Wen_Autoregressive_Stylized_Motion_Synthesis_With_Generative_Flow_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Wen_Autoregressive_Stylized_Motion_Synthesis_With_Generative_Flow_CVPR_2021_paper.pdf | Autoregressive Stylized Motion Synthesis With Generative Flow | Motion style transfer is an important problem in many computer graphics and computer vision applications, including human animation, games, and robotics. Most existing deep learning methods for this problem are supervised and trained by registered motion pairs. In addition, these methods are often limited to yielding a deterministic output, given a pair of style and content motions. In this paper, we propose an unsupervised approach for motion style transfer by synthesizing stylized motions autoregressively using a generative flow model M. M is trained to maximize the exact likelihood of a collection of unlabeled motions, based on an autoregressive context of poses in previous frames and a control signal representing the movement of a root joint. Thanks to invertible flow transformations, latent codes that encode deep properties of motion styles are efficiently inferred by M. By combining the latent codes (from an input style motion S) with the autoregressive context and control signal (from an input content motion C), M outputs a stylized motion which transfers style from S to C. Moreover, our model is probabilistic and is able to generate various plausible motions with a specific style. We evaluate the proposed model on motion capture datasets containing different human motion styles. Experiment results show that our model outperforms the state-of-the-art methods, despite not requiring manually labeled training data. | ['Yong-Jin Liu', 'Yanan sun', 'Lin Gao', 'Hongbo Fu', 'Zhipeng Yang', 'Yu-Hui Wen'] | 2021-06-19 | null | null | null | cvpr-2021-1 | ['motion-style-transfer'] | ['computer-code'] | [ 2.18311936e-01 -9.93917417e-03 -2.39266992e-01 -1.47701308e-01
-3.94180119e-01 -5.62313914e-01 9.21816826e-01 -5.74391961e-01
-2.60657340e-01 5.70422888e-01 4.10299331e-01 1.30139723e-01
4.35357094e-01 -9.02572513e-01 -9.54018652e-01 -7.01494515e-01
4.02749866e-01 5.20029604e-01 2.64740467e-01 -1.02261409e-01
1.41421616e-01 4.07223284e-01 -1.26950753e+00 1.90624073e-01
6.56806946e-01 4.44437802e-01 4.88107800e-01 8.91724348e-01
-1.51643634e-01 8.54314446e-01 -3.51036042e-01 -2.22326472e-01
1.66498378e-01 -9.52637672e-01 -7.83951223e-01 4.14216101e-01
2.84622312e-01 -5.17773330e-01 -4.83983427e-01 1.02051830e+00
1.15661845e-01 4.83281821e-01 1.00552118e+00 -1.11914122e+00
-6.54864311e-01 3.12199682e-01 -4.59667265e-01 -3.16398978e-01
4.36961442e-01 4.52544749e-01 8.91443253e-01 -7.49310791e-01
1.18606675e+00 1.57776189e+00 2.51486689e-01 9.69968557e-01
-1.50739217e+00 -2.66880304e-01 7.26481602e-02 -8.39369521e-02
-9.53334272e-01 -7.73742348e-02 1.00506043e+00 -6.77375436e-01
4.33431715e-01 -1.00780159e-01 8.35929751e-01 1.37074327e+00
1.07395932e-01 1.07329583e+00 7.27093279e-01 -3.21552038e-01
5.68532884e-01 -2.15876266e-01 -5.03308594e-01 7.42282510e-01
-1.74462348e-01 1.59816638e-01 -4.55225617e-01 -1.22313663e-01
1.42114830e+00 -1.49310917e-01 -2.94160962e-01 -6.98635101e-01
-1.48055792e+00 8.73736918e-01 3.73895317e-01 2.65726089e-01
-4.95527387e-01 4.86339331e-01 1.87747270e-01 -2.03354105e-01
1.44593984e-01 2.12909058e-01 -6.66545108e-02 -9.82593298e-02
-8.86652708e-01 5.13529956e-01 7.64707685e-01 1.00091147e+00
8.80639076e-01 4.72289741e-01 -4.08965915e-01 4.08826470e-01
3.70983779e-01 7.02172816e-01 4.92486000e-01 -1.42540526e+00
3.48724842e-01 1.96226254e-01 3.14108104e-01 -1.13716602e+00
4.58739512e-02 -1.50828052e-03 -8.81586254e-01 3.23510557e-01
5.46025038e-01 -2.65282214e-01 -7.37416029e-01 1.88475239e+00
2.58454382e-01 6.58340752e-01 1.32128047e-02 1.04480147e+00
3.99632066e-01 9.25934553e-01 2.56773651e-01 -1.43857161e-02
1.04444122e+00 -9.40081179e-01 -6.92882359e-01 -2.78890342e-01
4.32958126e-01 -8.12878907e-01 1.12954128e+00 1.55616015e-01
-1.07268524e+00 -8.78840804e-01 -7.02199459e-01 -1.05646469e-01
4.02463615e-01 2.47391313e-01 1.40947729e-01 2.81571597e-01
-8.57449293e-01 7.84278214e-01 -1.03846931e+00 -4.07045275e-01
9.73850787e-02 -1.05033629e-02 -3.03547859e-01 1.06389336e-01
-8.72250795e-01 5.44253230e-01 2.17206463e-01 5.41766407e-03
-1.08050704e+00 -2.93362677e-01 -1.06528175e+00 -6.97038695e-02
-2.94902883e-02 -1.27122974e+00 1.11925507e+00 -1.11158204e+00
-2.09703231e+00 5.43525040e-01 -3.65006059e-01 -2.92617857e-01
7.13914573e-01 -5.35080254e-01 1.29309237e-01 2.43924603e-01
2.60069549e-01 1.15606570e+00 1.10450459e+00 -1.36755776e+00
-4.86951649e-01 6.33467734e-02 -1.64001793e-01 3.39417934e-01
-7.97083676e-02 -2.68882900e-01 -6.57108426e-01 -9.55953658e-01
-4.94479164e-02 -1.22160208e+00 -4.20837462e-01 1.77743897e-01
-4.17327642e-01 7.52579495e-02 9.43588674e-01 -5.80446362e-01
7.18801677e-01 -1.98766398e+00 9.81262922e-01 9.50283371e-03
-3.08973908e-01 1.20846324e-01 -2.63355672e-01 9.88431051e-02
2.46788293e-01 -1.52316749e-01 -5.01659930e-01 -5.47378838e-01
-9.55657810e-02 4.79865134e-01 -5.55001438e-01 3.18737268e-01
2.79839069e-01 9.51855361e-01 -1.29363632e+00 -3.94452512e-01
5.61476409e-01 5.39241970e-01 -8.23950052e-01 6.89019799e-01
-6.42236114e-01 1.25057137e+00 -4.77928907e-01 5.02383448e-02
1.40522465e-01 -9.97817591e-02 2.52959788e-01 -1.27213404e-01
7.02007040e-02 -7.63976052e-02 -1.22869444e+00 2.23504758e+00
-4.36223358e-01 5.14457524e-01 -3.32419425e-01 -7.42810190e-01
1.08715332e+00 3.71269137e-01 3.84179801e-01 1.82493106e-01
2.58901805e-01 8.48134086e-02 -2.67334461e-01 -5.88964045e-01
4.80754226e-01 -5.79947606e-02 -1.61883563e-01 4.77429509e-01
2.34183937e-01 -6.39526367e-01 -4.50937785e-02 2.66717747e-02
7.19113290e-01 9.11335051e-01 8.69961306e-02 1.18512446e-02
7.73926079e-01 -1.55069843e-01 6.32129014e-01 4.59712893e-01
7.86394812e-03 9.28837359e-01 4.06127304e-01 -4.66260225e-01
-1.28079331e+00 -1.01833618e+00 5.27720511e-01 7.47002304e-01
1.33688748e-01 -1.02640964e-01 -9.19377089e-01 -5.02356410e-01
-3.85473311e-01 7.18443990e-01 -4.92994189e-01 -1.50835484e-01
-1.05166185e+00 -1.38180166e-01 2.79708117e-01 5.83300829e-01
5.79982162e-01 -1.40500677e+00 -6.78331196e-01 3.80731344e-01
-4.23954517e-01 -1.36254394e+00 -8.61413240e-01 -5.53141594e-01
-1.10369885e+00 -6.65578127e-01 -1.19927824e+00 -9.36832428e-01
5.88952661e-01 2.43886393e-02 9.24594164e-01 -2.10852295e-01
-1.27668474e-02 2.57046282e-01 -5.30286022e-02 2.46239394e-01
-8.65415096e-01 -7.17453212e-02 2.43503451e-02 5.21323800e-01
-2.25574523e-01 -6.51918054e-01 -6.47238076e-01 2.34900653e-01
-1.00981438e+00 3.31590593e-01 5.53744197e-01 9.45555449e-01
5.74899971e-01 -6.45644724e-01 8.77548009e-02 -8.15606833e-01
2.01002076e-01 -4.77754265e-01 -4.99795169e-01 -3.35443467e-02
1.60871208e-01 5.74873865e-01 6.88069880e-01 -6.78602815e-01
-1.37388015e+00 6.66286051e-01 -1.30189126e-02 -8.29350293e-01
-5.58978498e-01 5.73311895e-02 -2.94636160e-01 4.50947076e-01
7.25235999e-01 2.51384288e-01 -5.97253926e-02 -3.73103648e-01
1.01480305e+00 6.41554147e-02 1.01551878e+00 -8.45232308e-01
8.21917117e-01 7.02461898e-01 2.50272632e-01 -9.36138391e-01
-3.19014639e-01 -7.67333955e-02 -8.68483365e-01 -3.87651265e-01
1.38093877e+00 -7.31977880e-01 -3.60623062e-01 6.23916328e-01
-1.54481065e+00 -4.96898025e-01 -3.05178016e-01 6.83432341e-01
-1.12136841e+00 6.21645927e-01 -7.16061354e-01 -6.48939431e-01
-8.14291369e-03 -1.38005435e+00 1.30946720e+00 1.79929480e-01
-6.96575582e-01 -1.12568212e+00 3.29477698e-01 -2.19578408e-02
5.57524227e-02 6.29344165e-01 8.52138042e-01 1.70165926e-01
-7.28306055e-01 7.37906843e-02 2.41442949e-01 1.51711613e-01
4.50802654e-01 2.07209557e-01 -7.46990383e-01 -5.39923869e-02
-1.02409288e-01 -1.31665751e-01 6.92383170e-01 5.48736930e-01
8.17646086e-01 -3.60380352e-01 -2.47810125e-01 6.59434795e-01
1.06336260e+00 1.81073070e-01 5.85487902e-01 5.79415932e-02
1.12464881e+00 8.29325795e-01 3.68844897e-01 2.88153261e-01
1.14692241e-01 8.45247447e-01 2.60275215e-01 1.47311419e-01
-1.96742132e-01 -6.83999300e-01 5.88118970e-01 6.77900136e-01
-3.72338831e-01 -1.43774211e-01 -5.52361965e-01 5.88153183e-01
-2.06601262e+00 -1.02681339e+00 -1.45535722e-01 2.10690379e+00
6.68543935e-01 8.04305151e-02 1.46342829e-01 -1.13095775e-01
8.20844233e-01 2.51549035e-01 -5.79159260e-01 -1.42296821e-01
1.39114046e-02 2.10495591e-01 2.84359306e-02 6.60182476e-01
-9.45200682e-01 1.19732404e+00 5.34925747e+00 5.90315819e-01
-1.16284561e+00 -8.36244076e-02 4.69348431e-01 2.40112305e-01
-6.03705645e-01 1.17917970e-01 -4.70537513e-01 4.74849790e-01
4.54075754e-01 2.96070389e-02 3.47501725e-01 8.58612537e-01
3.94141793e-01 7.02781528e-02 -1.34942627e+00 8.07724655e-01
-1.29646197e-01 -1.56796014e+00 5.22270143e-01 -1.52753368e-01
1.05966556e+00 -5.19923925e-01 1.43214568e-01 9.80557967e-03
4.96037185e-01 -6.67599797e-01 9.81792510e-01 7.99243271e-01
6.51841223e-01 -5.47747433e-01 3.37075919e-01 4.20103908e-01
-1.16257703e+00 2.55164742e-01 -1.32368356e-01 -3.42509821e-02
5.10876060e-01 6.78108111e-02 -4.19812322e-01 4.68343496e-01
3.68662596e-01 9.26634550e-01 2.85012405e-02 5.18293679e-01
-6.98291719e-01 5.11623204e-01 4.87239845e-02 6.67530224e-02
2.45157853e-01 -5.15594363e-01 7.82531738e-01 1.04602718e+00
4.02514100e-01 -3.95571105e-02 1.42213225e-01 1.24411047e+00
3.78108211e-02 -1.02083929e-01 -9.46000695e-01 3.12280178e-01
2.98231870e-01 1.13288856e+00 -7.24942982e-01 -4.97071922e-01
-1.85470089e-01 1.39261520e+00 1.30244851e-01 5.97538710e-01
-6.77635789e-01 1.11336201e-01 6.80979192e-01 -1.71588764e-01
2.17343703e-01 -5.78074932e-01 -2.93823779e-01 -1.25397241e+00
-2.68042624e-01 -4.94640201e-01 -3.77255157e-02 -8.52905989e-01
-1.03613734e+00 6.05814099e-01 1.04811803e-01 -1.56985450e+00
-9.34143543e-01 -3.32405835e-01 -8.96753848e-01 8.59571040e-01
-9.19225037e-01 -1.05070877e+00 -2.85397470e-01 4.98343498e-01
9.17031169e-01 -2.55002290e-01 7.08989501e-01 -8.07059035e-02
-1.09122045e-01 1.45317465e-01 -2.05182523e-01 3.52650374e-01
6.14271641e-01 -1.15699375e+00 8.41210365e-01 9.64143097e-01
3.53490502e-01 3.84789437e-01 8.21044922e-01 -7.48899996e-01
-1.13194323e+00 -1.24798632e+00 5.34367561e-01 -4.37810212e-01
4.61051434e-01 -5.89748099e-02 -9.08197522e-01 6.48913205e-01
3.12348068e-01 -1.32903736e-02 2.28932738e-01 -7.99135089e-01
8.72993618e-02 3.90603811e-01 -8.11163068e-01 9.04139400e-01
1.09775841e+00 -4.15429264e-01 -6.34320796e-01 -4.03015986e-02
6.34448171e-01 -5.34542441e-01 -5.27371049e-01 1.33727595e-01
4.99020219e-01 -7.98166454e-01 9.10305321e-01 -7.25124776e-01
8.46898198e-01 -4.93807107e-01 4.14964221e-02 -1.35510790e+00
-3.57960761e-01 -6.87710643e-01 -1.12179130e-01 1.12757099e+00
-2.51184832e-02 4.55836877e-02 9.52894270e-01 5.41083515e-01
-1.64448190e-02 -3.30247492e-01 -5.45114219e-01 -6.29903018e-01
1.22241013e-01 -2.86858886e-01 4.47383881e-01 7.99284518e-01
-6.02005780e-01 3.98396164e-01 -7.52941549e-01 1.09351344e-01
7.26459324e-01 1.73028439e-01 1.22262323e+00 -8.28375816e-01
-6.84817612e-01 -4.39903647e-01 -3.40835959e-01 -1.41083694e+00
5.34707606e-01 -6.19283974e-01 2.69278765e-01 -1.39932954e+00
-1.17833361e-01 -1.13339782e-01 4.46550578e-01 9.38808098e-02
-3.46234977e-01 6.45912588e-02 5.02663434e-01 4.85115290e-01
-1.50112715e-02 9.15321410e-01 1.55481720e+00 -1.33380517e-01
-4.28650737e-01 4.77745943e-03 8.08602571e-02 1.15304852e+00
5.54517746e-01 -3.97441983e-01 -4.12528455e-01 -6.29899323e-01
-2.23291337e-01 5.38577914e-01 4.86496240e-01 -9.30356145e-01
7.35068843e-02 -4.33810085e-01 3.53904128e-01 -2.65839934e-01
4.10985112e-01 -5.09849548e-01 4.39662039e-01 7.10475922e-01
-3.91170114e-01 -2.99092736e-02 -1.90213323e-01 7.71515548e-01
-1.03609726e-01 -1.58892721e-01 6.98269546e-01 -2.77007163e-01
-8.14485192e-01 3.13935906e-01 -5.89722753e-01 -3.09971045e-04
7.64077008e-01 2.40175314e-02 1.68458298e-01 -6.69409037e-01
-9.38100934e-01 -1.90222785e-01 5.98181546e-01 6.78184867e-01
6.51348770e-01 -1.74810672e+00 -6.34040773e-01 2.03312591e-01
-1.75741196e-01 3.18286180e-01 1.48676619e-01 1.97818056e-01
-7.76076078e-01 3.12876180e-02 -6.12411678e-01 -9.12682295e-01
-7.41162956e-01 4.83570665e-01 1.98748574e-01 -1.10623628e-01
-6.03555202e-01 5.04109561e-01 5.98600268e-01 -2.69061089e-01
-9.09297094e-02 -5.09227693e-01 -2.23567128e-01 -4.20947164e-01
2.48669118e-01 3.83276910e-01 -5.79329014e-01 -9.75672901e-01
1.42853707e-01 9.16524827e-01 5.17183006e-01 -7.52983689e-01
9.51319396e-01 6.61451444e-02 1.33846045e-01 5.78047216e-01
1.13558137e+00 -1.35143831e-01 -1.81513608e+00 -1.85541272e-01
1.34408334e-03 -5.02966881e-01 -5.06100297e-01 1.61452293e-01
-1.15156198e+00 9.33204055e-01 2.73914605e-01 -3.96052331e-01
7.63285697e-01 -1.21954069e-01 8.64638209e-01 1.32414013e-01
5.31947911e-01 -6.99512124e-01 5.56238890e-01 4.73073691e-01
9.19547379e-01 -8.86601925e-01 -4.03169364e-01 -2.83208877e-01
-8.98507118e-01 1.19926131e+00 6.26166582e-01 -5.85639000e-01
4.22704756e-01 -8.36613402e-02 4.31327820e-02 2.82468081e-01
-3.91119748e-01 -7.00173751e-02 4.48750108e-01 6.62371755e-01
3.15976918e-01 5.95938824e-02 -1.77864105e-01 2.33968765e-01
-1.71017215e-01 1.45477951e-01 4.77274209e-01 7.24463403e-01
-2.34809488e-01 -1.28091013e+00 -5.28053343e-01 -2.30489016e-01
-3.00220978e-02 3.33833933e-01 -2.07977742e-01 5.34439206e-01
1.45246694e-02 6.35545909e-01 1.93386093e-01 -2.82937437e-01
2.14211851e-01 -5.06511889e-02 5.24779081e-01 -6.82329416e-01
-6.80691823e-02 4.44070339e-01 -3.12931836e-01 -6.17718935e-01
-6.82531893e-01 -8.48606110e-01 -1.36228490e+00 -9.14961994e-02
2.41677880e-01 4.68056323e-03 2.39703566e-01 9.53808606e-01
1.27838507e-01 4.35358495e-01 6.01226330e-01 -1.45477962e+00
-2.64213771e-01 -8.08477998e-01 -2.19771683e-01 8.78792584e-01
1.65580094e-01 -6.30165219e-01 -1.94668807e-02 9.35112834e-01] | [7.45280122756958, -0.18319538235664368] |
9a2a84c6-53d0-4be1-9886-f574a8bdf8e4 | multiview-detection-with-shadow-transformer | 2108.05888 | null | https://arxiv.org/abs/2108.05888v1 | https://arxiv.org/pdf/2108.05888v1.pdf | Multiview Detection with Shadow Transformer (and View-Coherent Data Augmentation) | Multiview detection incorporates multiple camera views to deal with occlusions, and its central problem is multiview aggregation. Given feature map projections from multiple views onto a common ground plane, the state-of-the-art method addresses this problem via convolution, which applies the same calculation regardless of object locations. However, such translation-invariant behaviors might not be the best choice, as object features undergo various projection distortions according to their positions and cameras. In this paper, we propose a novel multiview detector, MVDeTr, that adopts a newly introduced shadow transformer to aggregate multiview information. Unlike convolutions, shadow transformer attends differently at different positions and cameras to deal with various shadow-like distortions. We propose an effective training scheme that includes a new view-coherent data augmentation method, which applies random augmentations while maintaining multiview consistency. On two multiview detection benchmarks, we report new state-of-the-art accuracy with the proposed system. Code is available at https://github.com/hou-yz/MVDeTr. | ['Liang Zheng', 'Yunzhong Hou'] | 2021-08-12 | null | null | null | null | ['multiview-detection'] | ['computer-vision'] | [-1.30146205e-01 -3.42119902e-01 -5.95472902e-02 -5.02902925e-01
-7.61433005e-01 -7.09865868e-01 6.77284598e-01 -1.97446421e-01
-5.63875400e-02 1.29532173e-01 7.79207349e-02 -2.63540745e-02
4.66919333e-01 -5.48675239e-01 -9.02261496e-01 -4.35069263e-01
4.15846556e-01 2.65853018e-01 5.97845912e-01 -1.35742590e-01
1.46393478e-01 3.17979813e-01 -1.59049368e+00 4.62403238e-01
6.94985092e-01 1.01221907e+00 4.65150088e-01 5.53122401e-01
1.94847777e-01 6.55565023e-01 -3.11334342e-01 -6.00050449e-01
5.87828696e-01 4.02419753e-02 -1.85200766e-01 4.58870888e-01
1.12875187e+00 -7.17959344e-01 -5.12602985e-01 1.11189735e+00
2.98769951e-01 -3.02334893e-02 5.48455119e-01 -1.30489182e+00
-8.35959256e-01 2.06217155e-01 -9.59170163e-01 1.40911102e-01
4.37303603e-01 2.44794816e-01 1.02330399e+00 -1.61316252e+00
6.99204206e-01 1.28763509e+00 5.00418186e-01 3.13050866e-01
-1.08390629e+00 -6.38405979e-01 6.77844465e-01 2.70899713e-01
-1.36217809e+00 -4.36099648e-01 8.64144206e-01 -3.66279602e-01
7.21602678e-01 2.39668712e-01 6.47905409e-01 9.89131808e-01
2.07920954e-01 8.39005351e-01 9.73380506e-01 -2.46724159e-01
-1.75513308e-02 2.82941967e-01 -6.15722090e-02 7.23923743e-01
3.75394464e-01 -1.66768894e-01 -6.24269843e-01 -4.85988297e-02
6.71725392e-01 7.34195232e-01 -4.09899890e-01 -9.40022945e-01
-1.44196701e+00 7.19247580e-01 2.65055418e-01 -1.93934411e-01
-1.17286190e-01 -2.47850753e-02 2.45502919e-01 7.79856974e-03
4.99054164e-01 -1.85727820e-01 -3.12016815e-01 6.35760009e-01
-6.40721798e-01 2.57690102e-01 4.24001843e-01 1.30744803e+00
8.63418460e-01 6.40651584e-02 1.40931085e-02 6.59222543e-01
4.21215266e-01 7.44177759e-01 2.17621818e-01 -1.05166984e+00
8.18222106e-01 7.00775862e-01 -5.06873578e-02 -1.14050531e+00
-1.65802792e-01 -5.06418884e-01 -9.18530405e-01 3.97703171e-01
3.03621322e-01 5.53051420e-02 -6.98771954e-01 1.51774764e+00
6.58474445e-01 9.50320587e-02 -1.54524356e-01 1.02859294e+00
9.35614407e-01 5.02925217e-01 -5.37190914e-01 -5.72071150e-02
1.36922884e+00 -1.05578661e+00 -6.86342895e-01 -4.43553299e-01
2.20782816e-01 -1.06498551e+00 8.94985557e-01 3.34771931e-01
-1.24467242e+00 -5.11732459e-01 -1.27593291e+00 -1.72693416e-01
-2.43404657e-01 3.27856332e-01 3.66340131e-01 3.65037501e-01
-8.16612184e-01 1.08827733e-01 -6.39245510e-01 -2.91527957e-01
2.20281139e-01 1.55427441e-01 -6.10110819e-01 -2.95180470e-01
-6.15323365e-01 7.37919152e-01 -1.42180383e-01 5.87534979e-02
-9.42111909e-01 -6.37452722e-01 -1.06705391e+00 -1.15276009e-01
5.80418825e-01 -9.11986887e-01 1.14501584e+00 -6.47999823e-01
-1.18083632e+00 7.86348104e-01 -3.65244120e-01 -1.18722161e-02
6.67937696e-01 -3.64910483e-01 -2.79658526e-01 -8.94732494e-03
2.21068740e-01 3.37294251e-01 1.06660903e+00 -1.63241291e+00
-5.82555234e-01 -6.72658861e-01 1.50886253e-01 5.55264592e-01
-2.10679173e-01 -3.88942380e-03 -8.82506609e-01 -7.40769148e-01
6.49696767e-01 -8.12761784e-01 -7.43690953e-02 2.97767997e-01
-4.97325897e-01 7.53669813e-02 1.16002059e+00 -3.94603312e-01
8.03092241e-01 -2.06991577e+00 8.23375359e-02 -2.29610056e-01
6.19778097e-01 6.15182817e-02 1.46656364e-01 3.76178890e-01
-4.90202159e-02 -2.63119429e-01 -1.18434757e-01 -7.93794572e-01
-2.05904797e-01 -3.80661152e-02 -3.55755836e-01 9.42411900e-01
-1.20269820e-01 6.22606993e-01 -6.95125580e-01 -4.78585482e-01
6.83041155e-01 4.64698225e-01 -5.73987424e-01 2.22605333e-01
6.33740947e-02 2.48045638e-01 -2.06525236e-01 9.58583713e-01
1.21160376e+00 -5.07414103e-01 1.67640179e-01 -7.66164720e-01
-2.16608301e-01 -3.47566217e-01 -1.59454656e+00 1.77713609e+00
-3.71738791e-01 7.07855761e-01 1.30429521e-01 -7.03240693e-01
7.60462224e-01 1.21970572e-01 3.50882858e-01 -4.13811237e-01
4.40768227e-02 1.08425125e-01 -3.46926540e-01 -1.42250746e-01
6.99223399e-01 2.18732014e-01 1.76402435e-01 2.80327678e-01
1.65890127e-01 -5.04433587e-02 -2.20188811e-01 3.82911623e-01
8.40589285e-01 4.79043983e-02 6.50323987e-01 4.41823229e-02
5.32879770e-01 -4.59288985e-01 7.56900489e-01 6.69538856e-01
-8.86251479e-02 1.19264984e+00 -1.22705391e-02 -5.60892880e-01
-9.35113966e-01 -1.38201988e+00 -2.25147698e-02 8.81594360e-01
6.31157637e-01 -4.10776377e-01 -3.03333133e-01 -8.37015867e-01
9.66450199e-03 2.68029064e-01 -4.53412294e-01 2.70528108e-01
-5.49235642e-01 -4.17054564e-01 -1.43717825e-01 4.95006621e-01
4.84902173e-01 -3.43288064e-01 -6.45539343e-01 -1.22970000e-01
-3.39486152e-01 -1.45255387e+00 -8.32368433e-01 -2.75071152e-02
-6.32432044e-01 -1.14212108e+00 -8.59795749e-01 -5.35332263e-01
8.01550329e-01 1.05990601e+00 9.80353177e-01 -9.20117795e-02
-1.48487628e-01 5.69174945e-01 -2.92403072e-01 -3.01697552e-01
-2.22858526e-02 -2.50428826e-01 3.92159641e-01 3.24077010e-01
-9.68805254e-02 -6.74542546e-01 -8.88639212e-01 4.98414785e-01
-8.70593190e-01 2.57286698e-01 4.70549196e-01 8.13223660e-01
6.69585824e-01 -5.57080090e-01 -1.92804430e-02 -8.52232397e-01
-3.53264250e-02 -1.28180757e-01 -8.16570401e-01 4.04948354e-01
-2.99399704e-01 -2.72101343e-01 7.30373085e-01 -3.99103880e-01
-1.02065945e+00 3.84571075e-01 2.10453510e-01 -1.07225060e+00
-5.13765290e-02 -1.63241565e-01 -5.03231645e-01 -3.02192509e-01
4.38598812e-01 3.76853973e-01 -1.30705252e-01 -3.74285161e-01
6.58406079e-01 3.54291677e-01 6.53308213e-01 -2.85862803e-01
1.06091678e+00 9.48033929e-01 -6.32828102e-02 -8.21649790e-01
-7.88875461e-01 -7.71769404e-01 -7.18634546e-01 -3.44393492e-01
7.96624064e-01 -1.25502086e+00 -6.77389920e-01 5.30247331e-01
-1.49909008e+00 1.43472090e-01 -9.60338302e-03 5.80993712e-01
-5.78850448e-01 7.18282998e-01 -3.34199756e-01 -5.90874910e-01
-2.86445260e-01 -1.42844641e+00 1.27218199e+00 -9.22480673e-02
2.10330352e-01 -6.80335581e-01 -2.06946477e-01 2.48871773e-01
2.12182924e-01 1.92144036e-01 3.54250699e-01 -3.80702406e-01
-1.13464308e+00 -1.48206100e-01 -3.41722727e-01 4.22580928e-01
1.49149895e-01 -7.88670219e-03 -1.18950188e+00 -4.02522624e-01
2.26682022e-01 -6.69485405e-02 8.80218744e-01 3.92754942e-01
1.02118778e+00 -3.40590507e-01 -3.94271493e-01 1.05645239e+00
1.62736785e+00 -9.45327878e-02 3.16082507e-01 2.54480779e-01
1.14226067e+00 3.81030411e-01 7.36348033e-01 5.83786786e-01
7.27159262e-01 1.10646236e+00 8.31645310e-01 -2.56962746e-01
-1.75283298e-01 -8.20829999e-03 3.23745281e-01 7.89871991e-01
-6.78465590e-02 -2.49250785e-01 -7.41184890e-01 2.78021276e-01
-2.03123212e+00 -1.01845622e+00 -2.23419577e-01 2.41461205e+00
2.50514299e-01 -2.33352762e-02 -4.11829911e-02 -2.89918274e-01
6.92561090e-01 6.60098910e-01 -5.49824953e-01 3.35236937e-01
-2.66174316e-01 -4.96492743e-01 7.36807942e-01 4.77096111e-01
-1.15561187e+00 5.78856885e-01 5.44526625e+00 5.26966810e-01
-1.09030628e+00 2.97104090e-01 4.56655711e-01 -3.92238908e-02
-2.30593517e-01 1.05772391e-01 -1.08508217e+00 3.14261883e-01
-3.93243879e-02 4.64259088e-02 2.85087734e-01 9.13309097e-01
-1.11994974e-01 5.57367243e-02 -1.17927790e+00 1.25581205e+00
5.97316623e-01 -1.54040921e+00 2.41776451e-01 1.78233292e-02
8.20473731e-01 1.91248968e-01 3.05230200e-01 -9.82572734e-02
1.32671013e-01 -2.80101329e-01 1.01734865e+00 4.20042455e-01
5.72740793e-01 -4.59092647e-01 4.50988352e-01 1.89930916e-01
-1.70393431e+00 -9.49229151e-02 -3.76395911e-01 2.49867260e-01
3.49537611e-01 5.88387251e-01 -5.11717141e-01 7.76072025e-01
8.10250044e-01 9.94308054e-01 -7.35356808e-01 8.01727057e-01
2.99411658e-02 -1.56220749e-01 -1.49980009e-01 5.76195359e-01
-1.97560176e-01 -1.56167075e-02 8.34961772e-01 9.66267347e-01
4.22298044e-01 -4.51978408e-02 3.94969374e-01 7.00941205e-01
-4.15141508e-03 -4.88945320e-02 -1.00223732e+00 8.47231805e-01
5.62641799e-01 1.44429255e+00 -6.05944335e-01 -4.89566386e-01
-1.07985353e+00 1.17664480e+00 3.42010647e-01 4.75016415e-01
-9.77190673e-01 1.25049397e-01 7.08988965e-01 3.66429090e-01
4.17334884e-01 -3.00762445e-01 -1.82858825e-01 -1.79738235e+00
4.24674213e-01 -1.02636063e+00 3.25687915e-01 -7.48129547e-01
-1.23415852e+00 3.91907841e-01 4.91161607e-02 -1.96203971e+00
-6.35056198e-02 -5.66055298e-01 -6.88732862e-01 6.60305440e-01
-1.43216085e+00 -1.45134568e+00 -6.43329620e-01 7.37846017e-01
9.54787016e-01 -1.41926855e-01 3.62855494e-01 4.58971053e-01
-4.46999580e-01 6.26813829e-01 2.64583994e-02 -1.54982775e-01
1.05507255e+00 -1.14750350e+00 4.61652309e-01 1.05638719e+00
5.11001050e-02 6.11118317e-01 5.97408652e-01 -3.96404296e-01
-1.85442054e+00 -1.22662055e+00 4.19383407e-01 -6.34648621e-01
4.25129294e-01 -5.13012409e-01 -8.16748798e-01 1.04681909e+00
3.86752903e-01 5.58187962e-01 3.00032645e-01 -5.92633672e-02
-5.84255099e-01 -4.18888509e-01 -8.66350293e-01 5.72120428e-01
1.07446980e+00 -5.42045116e-01 -2.36519396e-01 3.46974701e-01
7.41654456e-01 -8.10059130e-01 -5.79969287e-01 6.27013505e-01
7.59454250e-01 -1.24151897e+00 1.45494628e+00 2.90387385e-02
2.64137954e-01 -6.58025503e-01 -7.17400730e-01 -1.21512711e+00
-1.92892045e-01 -1.56100631e-01 -3.03061366e-01 1.16637146e+00
1.59609795e-01 -9.06172454e-01 5.57085276e-01 3.43400389e-01
-2.44035900e-01 -7.25873649e-01 -7.91893303e-01 -7.91964293e-01
-5.36832273e-01 -2.47483566e-01 4.04142529e-01 9.35558200e-01
-3.70753348e-01 3.29425752e-01 -6.38977587e-01 6.51509047e-01
9.87063468e-01 3.69534731e-01 1.30943513e+00 -1.04392517e+00
-4.68159497e-01 -2.38684401e-01 -5.18213987e-01 -1.16411829e+00
-1.53390318e-01 -7.11141586e-01 -1.65127188e-01 -1.45658016e+00
3.82324725e-01 4.74410094e-02 1.56969324e-01 4.10794169e-02
-2.64142632e-01 2.75433004e-01 6.81980550e-01 2.55965918e-01
-8.89785528e-01 6.07955992e-01 1.22534013e+00 -1.19545884e-01
-3.15127037e-02 4.88646291e-02 -4.07788098e-01 1.06499684e+00
6.06979609e-01 -2.00264767e-01 -3.85516047e-01 -6.38694227e-01
-4.37199436e-02 1.96242511e-01 6.05905056e-01 -1.06546640e+00
4.61599618e-01 -2.91746389e-02 7.00885057e-01 -1.13665247e+00
8.71854782e-01 -8.74735177e-01 3.26281160e-01 1.86170831e-01
1.18085973e-01 7.10682988e-01 -1.25553319e-02 7.95069516e-01
-2.26275444e-01 2.74838388e-01 8.14833879e-01 -3.91465992e-01
-6.11818790e-01 5.99963844e-01 -5.16293682e-02 -1.18575186e-01
1.18317425e+00 -1.79494768e-01 -6.75889075e-01 -4.76998866e-01
-5.49963653e-01 1.66037649e-01 8.04924428e-01 4.27330077e-01
1.00991619e+00 -1.38965619e+00 -6.31022036e-01 2.88019270e-01
4.09277380e-01 8.14161301e-02 5.56839764e-01 9.01008368e-01
-3.01755846e-01 1.21270053e-01 5.18959276e-02 -9.97600853e-01
-1.54242837e+00 7.91904926e-01 3.33092809e-01 9.28587187e-03
-8.08281600e-01 5.88284194e-01 1.01311660e+00 -5.45656502e-01
2.37863258e-01 -3.62974435e-01 1.50049748e-02 8.54976773e-02
6.64593041e-01 3.78130168e-01 7.31751621e-02 -8.17806661e-01
-4.53005046e-01 9.38886404e-01 -2.01505765e-01 -4.57456373e-02
1.00433159e+00 -4.78970468e-01 1.38308644e-01 5.28852522e-01
1.17003322e+00 1.46088064e-01 -1.32982254e+00 -4.52403516e-01
-7.62537897e-01 -9.19653654e-01 -1.58920363e-01 -2.33535111e-01
-1.19312489e+00 9.37532067e-01 7.37388313e-01 -4.42413315e-02
1.05441761e+00 2.30818591e-03 4.41323698e-01 2.65047908e-01
4.38180685e-01 -6.26141965e-01 2.53851354e-01 4.02901888e-01
1.00112283e+00 -1.67395270e+00 3.37704033e-01 -7.45714664e-01
-6.36638165e-01 1.01535749e+00 9.09281969e-01 -3.72297727e-02
7.61320949e-01 1.07026294e-01 2.70979125e-02 -2.81593204e-01
-6.27353966e-01 -1.93046317e-01 2.72941619e-01 4.94667619e-01
2.22202376e-01 6.87260106e-02 -4.37473617e-02 1.59728467e-01
3.31960678e-01 -5.02446055e-01 5.41323543e-01 7.41552889e-01
-2.68197507e-01 -9.03653204e-01 -5.79691350e-01 3.28992277e-01
-3.35340887e-01 -1.34817332e-01 -8.54913592e-02 8.28739226e-01
-2.87033827e-03 7.91535556e-01 -1.01810507e-01 -5.55174112e-01
5.00502348e-01 -3.38502884e-01 4.98971969e-01 -6.57698452e-01
-2.09821880e-01 1.84358954e-01 -2.99729615e-01 -7.89899170e-01
-4.13250506e-01 -9.63525057e-01 -8.45622361e-01 -2.64267087e-01
-3.24580997e-01 -5.58956802e-01 4.77516234e-01 6.00339055e-01
2.74640441e-01 4.42554653e-01 8.96968603e-01 -1.25370395e+00
-5.72081327e-01 -6.67086005e-01 -5.55126429e-01 3.28357130e-01
7.12895989e-01 -7.99591064e-01 -3.82184267e-01 -9.68897808e-03] | [7.949124336242676, -2.1132357120513916] |
292d172d-86b6-495f-aeb0-a558ce31386f | adversarial-robustness-of-neural-statistical | 2203.07983 | null | https://arxiv.org/abs/2203.07983v1 | https://arxiv.org/pdf/2203.07983v1.pdf | Adversarial Robustness of Neural-Statistical Features in Detection of Generative Transformers | The detection of computer-generated text is an area of rapidly increasing significance as nascent generative models allow for efficient creation of compelling human-like text, which may be abused for the purposes of spam, disinformation, phishing, or online influence campaigns. Past work has studied detection of current state-of-the-art models, but despite a developing threat landscape, there has been minimal analysis of the robustness of detection methods to adversarial attacks. To this end, we evaluate neural and non-neural approaches on their ability to detect computer-generated text, their robustness against text adversarial attacks, and the impact that successful adversarial attacks have on human judgement of text quality. We find that while statistical features underperform neural features, statistical features provide additional adversarial robustness that can be leveraged in ensemble detection models. In the process, we find that previously effective complex phrasal features for detection of computer-generated text hold little predictive power against contemporary generative models, and identify promising statistical features to use instead. Finally, we pioneer the usage of $\Delta$MAUVE as a proxy measure for human judgement of adversarial text quality. | ['Paula Branco', 'Herna Viktor', 'Nathalie Japkowicz', 'Evan Crothers'] | 2022-03-02 | null | null | null | null | ['adversarial-text'] | ['adversarial'] | [ 2.40854710e-01 -2.40773950e-02 1.61180511e-01 -1.56310678e-01
-9.34013605e-01 -9.63285923e-01 1.25419903e+00 4.83156085e-01
-4.66556102e-01 5.00067413e-01 5.72041392e-01 -7.07411647e-01
-3.41158807e-02 -9.71292019e-01 -4.70163673e-01 -3.17074001e-01
1.95126534e-01 2.86007792e-01 -1.15600929e-01 -5.72862983e-01
5.75659275e-01 4.20378894e-01 -1.01733196e+00 3.11974436e-01
8.21544588e-01 4.71805513e-01 -5.02984285e-01 9.29137528e-01
2.05498517e-01 8.05438936e-01 -1.17287362e+00 -9.71150577e-01
-6.26412556e-02 -4.24965143e-01 -5.31759858e-01 -4.16311145e-01
5.56404233e-01 -4.23418164e-01 -7.40950882e-01 9.98138011e-01
8.21818709e-01 -1.98649019e-01 9.09281194e-01 -1.04390645e+00
-9.35731709e-01 8.31351399e-01 -1.60699621e-01 7.46176541e-01
6.10301137e-01 5.32033622e-01 1.02163470e+00 -5.83463192e-01
6.81034625e-01 1.36693323e+00 8.08797121e-01 3.74530524e-01
-1.30425501e+00 -8.34895968e-01 -3.36303502e-01 -2.91790068e-01
-6.53689146e-01 -7.08320379e-01 6.40977979e-01 -5.90947092e-01
7.43562520e-01 3.47380370e-01 4.59454298e-01 1.85618675e+00
6.68045700e-01 6.58537924e-01 1.10477865e+00 -2.83247679e-01
3.15548154e-03 2.41752490e-01 2.92445004e-01 6.69907689e-01
5.59062541e-01 4.71314400e-01 -5.95205069e-01 -7.13773072e-01
1.12423584e-01 -2.81529248e-01 -2.05863625e-01 4.33701307e-01
-7.26221561e-01 1.37838507e+00 3.37515116e-01 5.54166615e-01
3.48116457e-02 -5.19455131e-03 5.74730992e-01 3.62827808e-01
7.78504133e-01 1.03714252e+00 -1.12826444e-01 -2.80428588e-01
-1.01418924e+00 5.44214368e-01 1.01944757e+00 2.38838255e-01
3.46155524e-01 2.68934935e-01 -2.27470741e-01 6.12215936e-01
8.18431452e-02 7.76162922e-01 6.02177143e-01 -2.01086700e-01
4.88036573e-01 5.27413726e-01 -1.89063251e-01 -1.39436328e+00
-3.48832160e-01 -6.25834048e-01 -5.79640269e-01 9.58702564e-02
4.62689698e-01 -3.47723365e-01 -6.64186001e-01 1.29903114e+00
-1.18515097e-01 -4.13469017e-01 -2.18045384e-01 3.39988738e-01
5.62211573e-01 4.28958297e-01 1.06058314e-01 1.78203851e-01
1.06896353e+00 -2.34462574e-01 -3.50155801e-01 -6.29404247e-01
5.70085168e-01 -8.89526486e-01 9.62486029e-01 2.74391770e-01
-8.38536859e-01 -1.70920759e-01 -1.02751005e+00 2.58936763e-01
-7.24413097e-01 -4.10125554e-01 5.52752256e-01 1.51574588e+00
-6.25610471e-01 8.44722748e-01 -3.07620466e-01 -1.46371156e-01
6.41130269e-01 3.62544437e-03 -2.57106334e-01 1.91726889e-02
-1.45972359e+00 1.29336023e+00 2.07910061e-01 -2.08335057e-01
-7.80650318e-01 -5.45572698e-01 -6.24542773e-01 -7.82224834e-02
1.16161995e-01 -4.08088893e-01 9.35423076e-01 -6.96669340e-01
-8.35318446e-01 7.99749970e-01 1.06356777e-01 -6.07461214e-01
5.87032795e-01 -3.09004188e-01 -6.80570424e-01 -3.46119441e-02
1.07855611e-01 -5.06414510e-02 1.40059841e+00 -1.02786732e+00
-1.83164954e-01 -5.09353995e-01 -2.31871307e-01 -2.10811153e-01
-6.64504111e-01 3.14636260e-01 3.52594823e-01 -9.95768964e-01
-3.70272726e-01 -9.56426024e-01 4.50212024e-02 -6.12573564e-01
-6.24949217e-01 -1.23118982e-01 9.22426462e-01 -9.09467340e-01
1.35170436e+00 -1.79155469e+00 -3.57433736e-01 4.01429474e-01
5.24066985e-01 5.05270422e-01 6.28231168e-02 6.62422061e-01
2.02071935e-01 9.89420116e-01 -5.83499707e-02 5.00589656e-03
2.44868115e-01 -3.74479234e-01 -7.31099904e-01 4.41121876e-01
3.43469977e-01 1.22584832e+00 -7.27413714e-01 -2.43132651e-01
1.09684160e-02 4.79122698e-01 -4.46976125e-01 -9.27341580e-02
-1.04482695e-01 -9.38044582e-03 -5.24207950e-01 5.88167727e-01
1.29385233e-01 -9.21983123e-02 -2.15702951e-01 2.66233891e-01
2.93035328e-01 5.32616735e-01 -4.12641913e-01 6.68252289e-01
-1.47350669e-01 1.06372905e+00 -1.00046411e-01 -4.00418311e-01
8.38718474e-01 -6.94460869e-02 -2.12113485e-01 -5.12632430e-01
5.29015601e-01 9.73712951e-02 3.35341871e-01 -2.31031716e-01
7.13227630e-01 -1.25874966e-01 -2.82388151e-01 9.62014556e-01
-4.59944792e-02 -2.68738955e-01 5.22750579e-02 6.03276491e-01
1.48609817e+00 -3.12776119e-01 8.96367282e-02 -1.85337719e-02
7.71132037e-02 5.58680715e-03 -1.22026280e-01 1.25449884e+00
-2.13391244e-01 3.16369802e-01 5.40407419e-01 -1.49153084e-01
-1.23419607e+00 -9.63698149e-01 -2.58415580e-01 1.37039220e+00
-4.31797147e-01 -5.08821309e-01 -7.69228339e-01 -8.94085050e-01
2.68094271e-01 1.04210615e+00 -7.24134386e-01 -5.65904558e-01
-5.07293761e-01 -1.07229483e+00 1.15390253e+00 2.92750418e-01
6.36213496e-02 -1.25025761e+00 -2.54138112e-01 2.51987994e-01
-9.00256187e-02 -8.03916097e-01 -3.75841588e-01 5.14646992e-02
-5.99253953e-01 -1.10064828e+00 -3.09593648e-01 -9.63411853e-02
2.37530053e-01 1.88814439e-02 1.19529724e+00 1.96391523e-01
-3.96715373e-01 4.52530235e-01 -4.77581292e-01 -8.20923388e-01
-1.23251545e+00 2.21516177e-01 3.78261358e-02 -2.67870486e-01
3.61754000e-01 -3.80061984e-01 -2.57982641e-01 1.19705677e-01
-1.06425261e+00 -7.04959095e-01 5.95518589e-01 8.65648508e-01
-5.08110166e-01 1.35614619e-01 7.41070986e-01 -1.22764826e+00
1.55004144e+00 -6.18894517e-01 3.09744906e-02 -7.56592602e-02
-7.48465121e-01 3.02802045e-02 7.37930059e-01 -5.16626835e-01
-8.43730450e-01 -6.79513454e-01 -2.35607609e-01 1.03818607e-02
-1.48170829e-01 6.48952067e-01 1.17445089e-01 -2.11793512e-01
1.40987360e+00 2.62333721e-01 8.34766701e-02 -2.21782386e-01
3.47021788e-01 5.89692950e-01 2.86575824e-01 -2.94845402e-01
1.26492405e+00 1.51451603e-01 -3.58029932e-01 -9.46906149e-01
-7.10308909e-01 -1.41568676e-01 -2.59975463e-01 -1.49424657e-01
4.10250574e-01 -6.10657275e-01 -5.38261771e-01 5.66448033e-01
-9.60263073e-01 -2.53189318e-02 2.59630799e-01 -6.46692812e-02
-1.08508363e-01 5.30088365e-01 -6.94805026e-01 -9.02960300e-01
-6.24378443e-01 -8.10350537e-01 7.28239059e-01 -1.09716669e-01
-6.88897252e-01 -1.09900367e+00 8.65345895e-02 7.37181187e-01
6.64981067e-01 5.22230744e-01 9.86840367e-01 -1.30246019e+00
-1.84478909e-01 -1.00112116e+00 2.74314918e-03 4.32394207e-01
-2.84916341e-01 1.64556175e-01 -1.16066527e+00 -3.60273778e-01
7.58520737e-02 -4.81878161e-01 1.01855707e+00 3.58005501e-02
6.59744263e-01 -6.86762333e-01 -2.68377185e-01 1.53365657e-01
9.36151624e-01 -1.12873554e-01 6.79251850e-01 3.45747888e-01
5.47598600e-01 5.17644286e-01 2.37768684e-02 3.38287413e-01
-1.36950299e-01 9.85355824e-02 1.94144443e-01 2.65319198e-01
3.36647600e-01 -5.93117118e-01 6.68581128e-01 4.61837143e-01
2.07056805e-01 -7.03942657e-01 -1.09360969e+00 2.18966007e-01
-1.26238132e+00 -1.41377759e+00 -7.71774948e-02 1.84943640e+00
6.14907503e-01 9.03892100e-01 3.24132204e-01 3.38096231e-01
7.07058012e-01 5.12861669e-01 -4.48330045e-01 -6.01889551e-01
-2.57305533e-01 3.26496452e-01 4.93230939e-01 3.44256848e-01
-1.17677331e+00 7.91639209e-01 7.11883497e+00 8.45719039e-01
-8.59484076e-01 -1.12674229e-01 8.23060751e-01 -4.97892313e-02
-3.62926811e-01 -2.80290812e-01 -9.10256982e-01 7.27654994e-01
1.18005025e+00 -4.13036436e-01 3.90462130e-01 8.41188014e-01
8.21915343e-02 -5.89369908e-02 -9.27109301e-01 4.39242721e-01
5.09348989e-01 -1.17910528e+00 9.51583311e-02 3.18042129e-01
5.72159469e-01 2.11315528e-02 4.27457362e-01 5.05453348e-01
7.59633183e-01 -1.29185402e+00 5.54596126e-01 3.09980005e-01
5.32044828e-01 -7.41137385e-01 7.48274028e-01 5.05464971e-01
-2.51714796e-01 -2.31048137e-01 -1.64612576e-01 4.72575650e-02
-7.68686179e-03 5.89621246e-01 -1.21744430e+00 4.63311234e-03
1.33101791e-01 2.77468979e-01 -1.15424526e+00 3.73787522e-01
-2.73350745e-01 1.08979702e+00 -1.68383569e-01 -4.11083877e-01
2.58677661e-01 3.71259391e-01 8.87681603e-01 1.36164129e+00
-3.18931192e-02 -3.02243650e-01 -9.28846672e-02 7.68754065e-01
-2.34067395e-01 6.96300045e-02 -1.05951810e+00 -8.42295051e-01
4.08388793e-01 1.10975933e+00 -5.83349049e-01 -4.49060537e-02
-1.27689159e-02 7.29988456e-01 2.07930744e-01 1.25685677e-01
-5.50554514e-01 -5.09713888e-01 3.32013488e-01 3.48225385e-01
-1.38326198e-01 -1.23184912e-01 -6.44214690e-01 -1.11221004e+00
-9.47503373e-02 -1.39833713e+00 2.75686651e-01 -5.67543030e-01
-1.66986895e+00 4.98711169e-01 -4.57109034e-01 -5.97812116e-01
-5.56838870e-01 -6.42734528e-01 -9.05940890e-01 1.03051388e+00
-6.67853832e-01 -1.14133787e+00 7.35546798e-02 3.27909917e-01
3.18034351e-01 -6.49831295e-01 6.93967462e-01 -3.78734201e-01
-4.19422925e-01 7.39325881e-01 2.39420190e-01 6.30401611e-01
7.48840451e-01 -1.06863058e+00 1.00413942e+00 9.26805913e-01
2.92397141e-01 7.48205900e-01 8.86769712e-01 -1.16577280e+00
-1.21382058e+00 -8.75374377e-01 7.28104711e-01 -1.34340370e+00
1.06742013e+00 -4.79764730e-01 -8.23390961e-01 5.31144798e-01
-2.22919527e-02 -6.21151447e-01 8.37663233e-01 4.36110139e-01
-7.36126423e-01 4.11007494e-01 -1.14493906e+00 6.58076048e-01
6.89147055e-01 -9.96540964e-01 -6.81214571e-01 2.37103149e-01
4.81911242e-01 1.18292950e-01 -6.64068580e-01 8.84506628e-02
7.36855328e-01 -9.65873480e-01 9.56866562e-01 -8.71183217e-01
7.36825347e-01 4.31108654e-01 1.05633505e-01 -1.44473851e+00
-4.92183566e-01 -7.51139820e-01 -9.92664546e-02 1.07141769e+00
6.91588461e-01 -7.68141627e-01 7.67979205e-01 7.13989317e-01
1.24940507e-01 -4.92300630e-01 -4.93154317e-01 -4.54301327e-01
4.70843911e-01 -5.73525786e-01 1.77968815e-01 1.10945058e+00
1.25221804e-01 5.61402500e-01 -3.32263976e-01 -1.95624992e-01
6.80326283e-01 -4.95685667e-01 8.59637499e-01 -1.26313496e+00
-3.59082729e-01 -8.50248694e-01 -5.42988896e-01 -2.29417220e-01
1.29409924e-01 -1.01063001e+00 -2.47859463e-01 -1.09539580e+00
1.98779717e-01 -1.04490347e-01 1.30990297e-01 1.47577241e-01
-5.94163656e-01 4.44791526e-01 3.76173019e-01 2.51487583e-01
-1.82873592e-01 1.39539093e-01 9.64551091e-01 -2.69262880e-01
4.25523743e-02 1.11333944e-01 -1.07468820e+00 6.66945755e-01
9.66720462e-01 -5.55829763e-01 -1.47367641e-01 5.36794998e-02
5.78421533e-01 -2.60254085e-01 5.34171641e-01 -7.13222504e-01
-1.14761636e-01 1.02985255e-01 8.46520483e-01 -3.27835470e-01
1.11846745e-01 -9.78850499e-02 -4.45546657e-01 4.19617474e-01
-4.74640101e-01 2.50160605e-01 2.06085995e-01 8.50589991e-01
2.17084780e-01 -4.37540591e-01 7.03915954e-01 -1.53742358e-01
6.65179715e-02 7.61341229e-02 -7.34783411e-01 4.41563785e-01
6.19408488e-01 -1.18469819e-01 -7.92949319e-01 -6.39589250e-01
-5.98183095e-01 -2.37233430e-01 4.48227525e-01 6.35130346e-01
3.70999217e-01 -7.93252587e-01 -1.03806949e+00 1.68074787e-01
-1.09709360e-01 -9.44739759e-01 -3.32748406e-02 3.00627530e-01
-3.07219595e-01 3.95451039e-01 -3.80262434e-02 -1.44077629e-01
-1.26901519e+00 5.19548893e-01 5.73089831e-02 -3.95010322e-01
-1.98845521e-01 6.96753383e-01 -1.77899197e-01 -1.73013493e-01
8.36666953e-03 5.61063766e-01 -6.99217916e-02 2.61204422e-01
5.39702177e-01 6.46602750e-01 7.56368339e-02 -6.10340893e-01
-5.95695004e-02 -3.81069303e-01 -4.71242815e-01 -2.66241908e-01
1.14893055e+00 3.62491786e-01 3.01726889e-02 3.60434771e-01
1.18524480e+00 3.26744199e-01 -4.79198128e-01 -3.48231383e-02
5.50055020e-02 -4.68638539e-01 6.63982984e-03 -1.16482580e+00
-3.00782949e-01 8.48159850e-01 2.51932800e-01 8.65486622e-01
4.60103929e-01 -8.69705454e-02 7.50092089e-01 3.93585384e-01
-3.84790786e-02 -9.58100855e-01 3.04525971e-01 6.47020876e-01
8.96282077e-01 -1.21019292e+00 2.97866940e-01 -4.61451858e-02
-5.73232412e-01 9.78579700e-01 3.67624998e-01 -8.74757618e-02
5.72824538e-01 2.32261330e-01 -1.18040055e-01 -3.07701319e-01
-6.05374217e-01 1.50273353e-01 4.41789001e-01 5.53992867e-01
5.25985837e-01 1.22617096e-01 -2.38975465e-01 3.03822100e-01
-8.39507043e-01 -6.52482212e-01 6.07856333e-01 7.53713548e-01
-6.14141405e-01 -8.42604697e-01 -5.72047949e-01 1.22162652e+00
-9.56618488e-01 -4.46827292e-01 -1.07774472e+00 6.16028070e-01
-2.98117608e-01 1.11361277e+00 -2.91802585e-01 -6.44179225e-01
-1.13655888e-02 4.63043153e-01 2.76235640e-01 -4.80597854e-01
-1.17149222e+00 -2.53921211e-01 4.91405755e-01 3.02405115e-02
1.96528733e-01 -8.13798070e-01 -2.27906078e-01 -8.35421085e-01
-6.00067556e-01 -1.11563735e-01 7.45694637e-01 9.52910364e-01
2.36700103e-01 1.10009052e-01 5.53902268e-01 -6.24002397e-01
-9.35320377e-01 -1.32783604e+00 -2.26855323e-01 6.57800674e-01
2.16931164e-01 -2.55713493e-01 -7.87591577e-01 -1.49362683e-01] | [8.213116645812988, 10.043889045715332] |
4bc33b1b-2574-42c6-a39a-b40259691003 | vision-transformers-for-single-image-dehazing | 2204.03883 | null | https://arxiv.org/abs/2204.03883v1 | https://arxiv.org/pdf/2204.03883v1.pdf | Vision Transformers for Single Image Dehazing | Image dehazing is a representative low-level vision task that estimates latent haze-free images from hazy images. In recent years, convolutional neural network-based methods have dominated image dehazing. However, vision Transformers, which has recently made a breakthrough in high-level vision tasks, has not brought new dimensions to image dehazing. We start with the popular Swin Transformer and find that several of its key designs are unsuitable for image dehazing. To this end, we propose DehazeFormer, which consists of various improvements, such as the modified normalization layer, activation function, and spatial information aggregation scheme. We train multiple variants of DehazeFormer on various datasets to demonstrate its effectiveness. Specifically, on the most frequently used SOTS indoor set, our small model outperforms FFA-Net with only 25% #Param and 5% computational cost. To the best of our knowledge, our large model is the first method with the PSNR over 40 dB on the SOTS indoor set, dramatically outperforming the previous state-of-the-art methods. We also collect a large-scale realistic remote sensing dehazing dataset for evaluating the method's capability to remove highly non-homogeneous haze. | ['Xin Du', 'Hui Qian', 'Zhuqing He', 'Yuda Song'] | 2022-04-08 | null | null | null | null | ['image-dehazing'] | ['computer-vision'] | [ 3.12249839e-01 -3.64903718e-01 2.98148721e-01 -2.26205468e-01
-6.22324228e-01 -4.38824221e-02 4.12498444e-01 -1.70324668e-01
-3.35842371e-01 4.81126219e-01 1.19056456e-01 -2.59915650e-01
5.42209633e-02 -9.00226414e-01 -7.49097764e-01 -1.06366038e+00
-3.19560766e-02 -2.96247959e-01 4.83139366e-01 -4.47373867e-01
1.15882464e-01 1.37251273e-01 -1.78971517e+00 2.58201420e-01
1.26586616e+00 1.24356699e+00 2.00241774e-01 6.67165101e-01
3.33824545e-01 9.96458828e-01 -7.33208477e-01 -2.99283773e-01
4.91238475e-01 -2.06144020e-01 -2.96585232e-01 1.42056840e-02
1.22144842e+00 -7.61680424e-01 -5.91219068e-01 1.14392173e+00
6.50902867e-01 1.25596344e-01 5.19671321e-01 -1.09519637e+00
-1.25740921e+00 1.84170455e-01 -5.34893990e-01 3.93972784e-01
-2.94314414e-01 3.14037144e-01 5.78325152e-01 -1.19438720e+00
2.86434181e-02 9.86649930e-01 7.48857379e-01 3.01458210e-01
-7.43155897e-01 -9.66060936e-01 1.56999290e-01 4.74158496e-01
-1.60780644e+00 -4.55443799e-01 6.30183339e-01 -2.79446244e-01
8.21996570e-01 1.00531213e-01 4.81239825e-01 6.68414354e-01
3.41409355e-01 7.47663200e-01 1.23975372e+00 -2.41733328e-01
2.09368482e-01 -1.50283813e-01 -5.52085340e-02 7.64528394e-01
3.73817414e-01 7.48284906e-02 -4.06274974e-01 3.37062955e-01
5.08547246e-01 2.22746521e-01 -5.06613731e-01 1.30873397e-01
-1.06384361e+00 6.81588650e-01 7.95392990e-01 1.19183786e-01
-2.22497627e-01 4.58695590e-01 -2.73893625e-01 3.16627741e-01
8.60091209e-01 4.46976781e-01 -1.15522303e-01 3.90395045e-01
-1.09543586e+00 2.86468834e-01 3.71651828e-01 8.95586133e-01
8.39981258e-01 5.17190635e-01 -2.00410932e-01 8.22192729e-01
6.42762706e-02 6.93608940e-01 1.23658039e-01 -9.65495169e-01
2.80720621e-01 1.12627372e-01 2.17623830e-01 -1.08107340e+00
-1.11003574e-02 -5.71643472e-01 -1.19271517e+00 5.25349498e-01
1.37187392e-01 -8.43158551e-03 -1.67827833e+00 1.23849154e+00
-3.44062299e-02 5.08681476e-01 -9.06976163e-02 1.03191721e+00
9.74676371e-01 1.10798025e+00 -3.14670086e-01 7.20397308e-02
1.07658947e+00 -1.26318419e+00 -8.87867212e-01 -4.40190762e-01
-5.04736193e-02 -8.01165581e-01 9.29848969e-01 7.48347819e-01
-1.01448309e+00 -5.72686076e-01 -1.49541223e+00 -2.46764138e-01
-6.42619848e-01 -1.28149390e-01 5.43596268e-01 6.87741697e-01
-1.32833648e+00 4.64751601e-01 -6.09431803e-01 -2.35665485e-01
6.51196837e-01 -1.25888353e-02 -9.62965116e-02 -4.94648904e-01
-1.34968698e+00 9.59897518e-01 2.89975345e-01 4.54718262e-01
-1.40419936e+00 -9.36416805e-01 -8.24943006e-01 1.01686731e-01
4.41480368e-01 -6.85434043e-01 9.97534275e-01 -8.34417582e-01
-1.41196454e+00 4.42888349e-01 2.40986850e-02 -6.17819607e-01
2.08260313e-01 -4.55460757e-01 -5.79016328e-01 1.90724492e-01
-6.60466626e-02 7.42172658e-01 1.48355794e+00 -1.55382085e+00
-6.75445557e-01 -4.66662943e-02 1.81030363e-01 1.06643893e-01
-5.46457291e-01 -2.48740107e-01 -5.65876365e-01 -9.60159481e-01
-2.52087060e-02 -6.39725447e-01 -1.68048665e-01 2.77195066e-01
-2.43011221e-01 1.68876797e-01 7.89800704e-01 -8.10073614e-01
1.16138399e+00 -2.23817134e+00 -1.89559296e-01 -1.93931371e-01
6.36764824e-01 6.21145546e-01 -1.76453918e-01 2.01297939e-01
4.41351868e-02 3.28425199e-01 -4.76483852e-01 -4.63411450e-01
-1.09487966e-01 5.95444702e-02 -3.14602852e-01 6.58333480e-01
1.89672530e-01 7.21627951e-01 -8.73118281e-01 -2.86927789e-01
5.21831036e-01 7.16420174e-01 -4.15140688e-01 2.56714821e-01
-1.77827597e-01 -2.85292715e-02 2.39809528e-02 9.28254008e-01
9.97474849e-01 -4.56923917e-02 -6.04488909e-01 -2.33980030e-01
-4.05315280e-01 -4.03522551e-02 -6.82160497e-01 1.44942963e+00
-4.29728746e-01 9.86407876e-01 1.99315930e-03 -5.88761985e-01
6.19052231e-01 1.30948037e-01 1.89191416e-01 -8.65756691e-01
-1.05803031e-02 1.03301562e-01 -2.92063296e-01 -5.09892166e-01
7.29911506e-01 -7.45473132e-02 3.04375947e-01 -1.40644148e-01
-1.18738875e-01 -5.18411934e-01 -1.61542952e-01 1.12154439e-01
9.74819720e-01 -2.37974510e-01 2.94565577e-02 -2.42916912e-01
1.98500067e-01 -9.09345224e-02 3.81867945e-01 9.40814614e-01
-3.48511279e-01 1.13842440e+00 -1.22190237e-01 -6.00608945e-01
-9.84243453e-01 -1.28875756e+00 -8.52622464e-02 8.94769311e-01
4.23735529e-01 -3.12313527e-01 -8.05719316e-01 -1.96366966e-01
-1.44170299e-01 6.58641398e-01 -6.75824583e-01 -2.85486490e-01
-3.36940676e-01 -9.01137471e-01 5.72779417e-01 4.85761203e-02
1.30712426e+00 -6.28426313e-01 -2.78160006e-01 -6.37037680e-02
-2.44767025e-01 -1.25708425e+00 -5.56754768e-01 -8.12829286e-02
-4.95962620e-01 -7.52664387e-01 -8.76064539e-01 -8.70579004e-01
4.54061359e-01 1.06601715e+00 1.02847469e+00 3.23260635e-01
-8.73261914e-02 7.47575564e-03 -4.88295645e-01 -8.24702263e-01
-7.42979720e-02 -1.37531562e-02 -2.71942765e-01 1.01633012e-01
1.82942390e-01 -8.00174713e-01 -1.16255391e+00 3.06560308e-01
-1.33921230e+00 1.22571290e-01 7.35309958e-01 7.34448791e-01
4.41073358e-01 7.75427222e-01 2.52106078e-02 -5.00756741e-01
3.73851657e-01 -6.48416638e-01 -5.13595045e-01 -2.31995899e-02
-7.95844913e-01 -2.75152922e-01 5.00091672e-01 -3.46693605e-01
-1.19729567e+00 -3.49900275e-01 -8.83022696e-02 -7.00560331e-01
-1.30491093e-01 5.18738985e-01 -6.04550801e-02 -4.58821088e-01
8.12527120e-01 4.89099175e-01 -1.84934318e-01 -3.99821728e-01
3.57243210e-01 7.27806151e-01 8.59563470e-01 -8.13342780e-02
1.42414045e+00 7.79323220e-01 -9.85029638e-02 -1.18963611e+00
-1.02217567e+00 -4.20769304e-01 -1.65067807e-01 -1.33784354e-01
1.09674680e+00 -1.40002418e+00 -4.19714332e-01 1.07620966e+00
-9.81464565e-01 -4.91167337e-01 1.06664404e-01 2.57076323e-01
4.32430096e-02 3.81566137e-01 -6.13412917e-01 -7.29639888e-01
-5.12497842e-01 -9.26102102e-01 1.08011055e+00 7.60658756e-02
6.68378592e-01 -7.13964462e-01 -2.63191015e-01 4.57246900e-01
9.78866398e-01 4.31751460e-01 5.87191999e-01 2.47088552e-01
-1.05159175e+00 3.86675037e-02 -5.02326548e-01 9.07370031e-01
2.87625670e-01 -5.17800972e-02 -1.27020431e+00 -3.86178881e-01
3.56219374e-02 -6.28520697e-02 1.51385570e+00 4.55877662e-01
1.30982125e+00 -3.05668473e-01 1.32893026e-01 1.18926954e+00
1.51463234e+00 1.42717093e-01 1.16597939e+00 6.71016216e-01
1.06314373e+00 2.54021615e-01 4.47173715e-01 1.04788572e-01
4.81794596e-01 4.90984231e-01 1.09404910e+00 -5.90842903e-01
-4.34565336e-01 1.19451378e-02 3.94429058e-01 7.18765855e-01
-2.44940028e-01 -5.73914170e-01 -8.07404041e-01 6.73363090e-01
-1.59410739e+00 -8.20434093e-01 6.30594492e-02 1.91703856e+00
6.82756186e-01 9.41966698e-02 -4.38100934e-01 -4.13692072e-02
4.59817201e-01 7.05879390e-01 -3.54935408e-01 6.18823767e-02
-2.41992086e-01 2.53062725e-01 8.64959955e-01 5.72663188e-01
-1.46587527e+00 1.13716424e+00 6.25873852e+00 8.42025399e-01
-1.09765959e+00 2.47168973e-01 5.35053074e-01 -5.18608466e-02
-2.10568979e-01 -2.91726649e-01 -5.00114322e-01 7.70290673e-01
6.07079387e-01 1.77862123e-01 6.58455908e-01 6.90490961e-01
2.25030884e-01 -8.20023865e-02 -5.87488770e-01 1.13368511e+00
4.55792308e-01 -1.34826982e+00 2.48056263e-01 -1.97049472e-02
1.05305970e+00 2.69586831e-01 5.04426599e-01 1.91683084e-01
2.69290715e-01 -1.22011900e+00 9.83349741e-01 4.82014000e-01
7.83475578e-01 -5.70354104e-01 9.93521333e-01 3.18052480e-03
-1.09525931e+00 4.63656709e-02 -5.68724453e-01 -1.00966126e-01
-7.47637525e-02 8.12983036e-01 -5.27669787e-01 5.35605967e-01
1.23791146e+00 8.39493215e-01 -6.95516765e-01 1.19014549e+00
-3.74728084e-01 9.11791205e-01 -1.82046235e-01 4.75052744e-01
4.29453373e-01 2.04479182e-03 4.21645224e-01 1.20453286e+00
5.53296864e-01 2.33628526e-01 -4.49597798e-02 6.74459934e-01
-6.92119449e-02 -5.10133028e-01 -5.05829751e-01 2.16642573e-01
3.87379885e-01 9.59593296e-01 -2.42177352e-01 -2.27474496e-01
-4.33209002e-01 1.04315472e+00 -1.24802321e-01 6.80002332e-01
-1.05681396e+00 -5.02236366e-01 1.21242476e+00 2.87841428e-02
5.38222373e-01 -4.52311367e-01 -3.84172648e-02 -1.15898848e+00
-4.86583263e-02 -8.78828228e-01 -7.49938539e-04 -1.20808017e+00
-1.26626599e+00 5.91896653e-01 5.16504683e-02 -1.23218596e+00
3.92736137e-01 -7.01240420e-01 -5.49236774e-01 8.35838675e-01
-2.27975011e+00 -1.20674920e+00 -1.17523634e+00 7.89615393e-01
8.07233930e-01 5.34372143e-02 3.75717968e-01 5.61626196e-01
-5.34696698e-01 4.86921251e-01 1.96121201e-01 5.09337001e-02
8.49916279e-01 -1.21984136e+00 6.40920162e-01 1.48922276e+00
-3.45530123e-01 4.79589581e-01 9.81169224e-01 -4.32742804e-01
-1.30967116e+00 -1.57213426e+00 4.37549412e-01 -4.46548700e-01
5.31347394e-01 -3.80494326e-01 -1.10477793e+00 5.80926180e-01
4.04330343e-01 2.27677941e-01 1.41792461e-01 -5.76707304e-01
-5.50665617e-01 -5.35544574e-01 -8.83205533e-01 7.65236914e-01
1.05129099e+00 -4.15632337e-01 -3.91743571e-01 2.09960803e-01
1.14175928e+00 -4.73993123e-01 -7.31962144e-01 4.99954879e-01
4.19477254e-01 -1.16398597e+00 1.25585651e+00 1.76557422e-01
5.58940232e-01 -7.09303796e-01 -3.11517835e-01 -1.54401171e+00
-7.12365091e-01 -7.27293551e-01 -1.89064279e-01 9.79655862e-01
9.53454524e-02 -6.68719649e-01 7.08963454e-01 3.73863466e-02
-5.62388778e-01 -4.89749759e-01 -6.24490798e-01 -7.70473659e-01
1.10165529e-01 -3.67028803e-01 7.51776338e-01 8.85482967e-01
-9.79965568e-01 6.11839956e-03 -9.53420401e-01 6.11168981e-01
1.07689977e+00 -2.05223098e-01 7.09407210e-01 -9.09625709e-01
5.97504675e-02 -2.58397758e-01 -3.46682727e-01 -9.39542830e-01
-1.74698174e-01 -3.36919457e-01 4.17060941e-01 -1.69496620e+00
-1.16827108e-01 -6.98089600e-02 -4.23772633e-01 6.07144654e-01
-4.50015008e-01 6.07614696e-01 1.45071745e-01 2.90480822e-01
-4.10753518e-01 7.41965294e-01 1.33747184e+00 -7.24761486e-01
1.04985744e-01 -2.54573107e-01 -8.60193551e-01 5.95315397e-01
1.01225471e+00 -4.98063654e-01 -6.16520345e-01 -9.60586965e-01
8.62669498e-02 -5.90212762e-01 4.84282583e-01 -1.25404131e+00
2.90650070e-01 -2.91979849e-01 4.55477953e-01 -4.14731622e-01
5.39757967e-01 -7.43006468e-01 1.70716420e-01 2.53600746e-01
5.54247312e-02 -1.52543649e-01 2.17346162e-01 5.67942619e-01
-4.08749938e-01 2.91993767e-01 8.69862974e-01 -2.25626558e-01
-1.08562982e+00 7.00957656e-01 -4.98182029e-01 -2.24470973e-01
6.68447316e-01 -5.19979775e-01 -9.32072103e-01 -6.01857066e-01
9.30986926e-03 1.15948528e-01 5.59758902e-01 2.92665154e-01
1.07285583e+00 -1.05653512e+00 -8.61209691e-01 1.31661713e-01
2.59032845e-01 3.02704602e-01 3.88226181e-01 7.09174514e-01
-8.97805929e-01 1.37231708e-01 -1.61350757e-01 -3.63183856e-01
-1.10654473e+00 6.99447095e-01 5.32814801e-01 2.31743708e-01
-7.41750896e-01 8.38311911e-01 5.65048695e-01 6.26122802e-02
1.45376548e-01 -5.16253114e-01 -5.95104098e-02 -2.76752949e-01
8.02617252e-01 4.77737844e-01 2.97410250e-01 -3.94000411e-01
-1.68249071e-01 3.81419331e-01 -9.35451761e-02 1.51295841e-01
1.38807130e+00 -3.49178165e-01 -2.45231375e-01 1.66606709e-01
9.93734956e-01 -1.25471145e-01 -1.62477577e+00 -3.11092794e-01
-5.06527364e-01 -9.59242761e-01 7.05121636e-01 -6.94314659e-01
-1.44267917e+00 8.93825948e-01 8.36282432e-01 2.24047527e-01
1.55601561e+00 -3.98348808e-01 1.02528811e+00 4.94179398e-01
4.83932525e-01 -7.26831138e-01 3.59458029e-01 5.30466497e-01
7.97271669e-01 -1.33831418e+00 1.70351297e-01 -4.35770482e-01
-3.90960395e-01 7.07667232e-01 8.21311653e-01 -1.20635062e-01
6.79070592e-01 1.42055824e-01 3.17506909e-01 -3.26113790e-01
-5.18790185e-01 -4.15802836e-01 3.36032480e-01 7.67115712e-01
-4.97012734e-02 -1.32757872e-01 3.42564315e-01 -1.19511150e-01
-1.34014696e-01 -3.62567231e-02 6.18618190e-01 8.50471795e-01
-8.67214620e-01 -3.27159733e-01 -5.53321421e-01 2.24321008e-01
-5.46104550e-01 -6.73621058e-01 -7.01362044e-02 8.26839209e-01
4.74178493e-01 1.22781670e+00 -8.48390535e-02 -5.90071499e-01
2.21019909e-01 -5.76800585e-01 4.09952223e-01 -4.12408412e-01
-3.35446000e-01 -1.53626710e-01 -1.82220533e-01 -5.43845117e-01
-5.57049632e-01 -1.16165571e-01 -6.57417715e-01 -6.32211030e-01
-2.54509300e-01 -4.63219918e-02 5.78254223e-01 7.79314160e-01
2.46134505e-01 6.86458170e-01 5.57486653e-01 -1.32927990e+00
-9.63007361e-02 -1.00374436e+00 -6.97571456e-01 2.14351296e-01
9.55458999e-01 -6.92192197e-01 -6.57540083e-01 1.41498312e-01] | [10.94719409942627, -3.0973522663116455] |
6369fddb-2c59-4a14-9808-fb4f00270876 | freestyle-layout-to-image-synthesis | 2303.14412 | null | https://arxiv.org/abs/2303.14412v1 | https://arxiv.org/pdf/2303.14412v1.pdf | Freestyle Layout-to-Image Synthesis | Typical layout-to-image synthesis (LIS) models generate images for a closed set of semantic classes, e.g., 182 common objects in COCO-Stuff. In this work, we explore the freestyle capability of the model, i.e., how far can it generate unseen semantics (e.g., classes, attributes, and styles) onto a given layout, and call the task Freestyle LIS (FLIS). Thanks to the development of large-scale pre-trained language-image models, a number of discriminative models (e.g., image classification and object detection) trained on limited base classes are empowered with the ability of unseen class prediction. Inspired by this, we opt to leverage large-scale pre-trained text-to-image diffusion models to achieve the generation of unseen semantics. The key challenge of FLIS is how to enable the diffusion model to synthesize images from a specific layout which very likely violates its pre-learned knowledge, e.g., the model never sees "a unicorn sitting on a bench" during its pre-training. To this end, we introduce a new module called Rectified Cross-Attention (RCA) that can be conveniently plugged in the diffusion model to integrate semantic masks. This "plug-in" is applied in each cross-attention layer of the model to rectify the attention maps between image and text tokens. The key idea of RCA is to enforce each text token to act on the pixels in a specified region, allowing us to freely put a wide variety of semantics from pre-trained knowledge (which is general) onto the given layout (which is specific). Extensive experiments show that the proposed diffusion network produces realistic and freestyle layout-to-image generation results with diverse text inputs, which has a high potential to spawn a bunch of interesting applications. Code is available at https://github.com/essunny310/FreestyleNet. | ['Wenjun Zhang', 'Li Song', 'Qianru Sun', 'Zhiwu Huang', 'Han Xue'] | 2023-03-25 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Xue_Freestyle_Layout-to-Image_Synthesis_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Xue_Freestyle_Layout-to-Image_Synthesis_CVPR_2023_paper.pdf | cvpr-2023-1 | ['layout-to-image-generation'] | ['computer-vision'] | [ 3.72167170e-01 3.52513999e-01 -1.16273696e-02 -4.44297671e-01
-3.36768776e-01 -7.33807027e-01 6.91612959e-01 -4.18592006e-01
-8.23293254e-02 4.67606574e-01 -3.70302424e-02 -2.38610938e-01
3.14713657e-01 -1.03703785e+00 -1.06214464e+00 -6.26609206e-01
5.21179020e-01 3.38369370e-01 2.23808497e-01 -2.90925354e-01
1.17864959e-01 4.42033291e-01 -1.61609721e+00 5.62406242e-01
9.21323001e-01 8.06400001e-01 7.95383990e-01 4.92697209e-01
-3.56465489e-01 6.51270330e-01 -7.78260648e-01 -5.72543561e-01
7.45326281e-02 -5.27473032e-01 -7.39157259e-01 3.50672871e-01
5.79773188e-01 -3.22175264e-01 -1.71508044e-01 1.10983253e+00
2.26303220e-01 -2.18424536e-02 7.27931917e-01 -1.28905404e+00
-1.17379439e+00 6.59424961e-01 -3.84551018e-01 -2.14216337e-01
1.04569368e-01 4.43061352e-01 9.94966090e-01 -9.31791782e-01
8.94475281e-01 1.22898042e+00 2.82991022e-01 8.25876176e-01
-1.27380145e+00 -7.12022126e-01 3.45849514e-01 -5.18669374e-02
-1.47763634e+00 -2.98233449e-01 8.99116039e-01 -4.11276966e-01
6.06006026e-01 3.66050899e-01 5.67094445e-01 1.63552988e+00
-1.47329159e-02 1.08008587e+00 9.89698052e-01 -4.50069278e-01
3.07493478e-01 5.12319922e-01 -3.27650607e-01 6.74435258e-01
2.00565994e-01 -1.99577540e-01 -4.01372612e-01 4.04743671e-01
8.61630738e-01 -8.62288326e-02 -2.43910059e-01 -4.83056188e-01
-1.08878946e+00 7.66191602e-01 7.41420388e-01 2.36837342e-01
9.91428941e-02 1.16143487e-01 1.00432955e-01 5.45830280e-02
3.40216219e-01 5.42695105e-01 -3.99991870e-01 3.64705503e-01
-8.55032861e-01 2.14582026e-01 6.31379008e-01 1.30413985e+00
9.67408597e-01 1.49470001e-01 -2.74677426e-01 8.99836779e-01
7.10825771e-02 6.67100906e-01 4.73634720e-01 -7.61631012e-01
3.97842020e-01 6.73051059e-01 3.89002785e-02 -6.59025013e-01
-9.47474539e-02 -4.90522385e-01 -8.62568736e-01 2.63829410e-01
3.60678524e-01 -1.08494632e-01 -1.28267336e+00 1.82766461e+00
1.82901725e-01 -6.34791255e-02 -4.70002033e-02 9.41735148e-01
6.41386688e-01 6.89113200e-01 2.08658725e-02 3.84733438e-01
1.26596415e+00 -1.18709874e+00 -3.88659120e-01 -6.03910267e-01
5.61267972e-01 -6.63469553e-01 1.61659896e+00 9.60856006e-02
-8.01657677e-01 -7.92886138e-01 -1.01596117e+00 -1.82899833e-01
-7.17998803e-01 3.15390527e-01 5.55334330e-01 6.05214596e-01
-1.03139150e+00 4.51175630e-01 -4.66793865e-01 -4.95393604e-01
5.61765432e-01 1.48904160e-01 -2.17875093e-01 -2.74999380e-01
-1.03885531e+00 6.11025631e-01 4.79797095e-01 7.87621140e-02
-1.12307322e+00 -5.09623587e-01 -9.30994332e-01 1.08360082e-01
6.35893047e-01 -8.64557326e-01 9.56055462e-01 -1.57846320e+00
-1.36124599e+00 1.04867756e+00 -5.02717607e-02 -1.34534284e-01
5.90817153e-01 -1.43244624e-01 -2.42155373e-01 -6.71181306e-02
4.45959181e-01 1.21694374e+00 1.10516083e+00 -1.63650119e+00
-4.12522823e-01 -1.03785232e-01 1.52688757e-01 1.60345674e-01
-4.83666241e-01 -1.86947629e-01 -8.09492648e-01 -9.24198389e-01
-2.90229082e-01 -9.89631355e-01 -7.30067566e-02 6.64486513e-02
-8.62391829e-01 -4.69019935e-02 8.14904988e-01 -4.17956501e-01
9.93193746e-01 -2.25444865e+00 1.20631233e-01 1.67154938e-01
-1.03000708e-01 2.13157862e-01 -5.11378706e-01 2.40240291e-01
3.27680409e-02 4.25250143e-01 -2.83867806e-01 -3.76010209e-01
2.59770919e-03 3.61136436e-01 -4.90669191e-01 -1.70246493e-02
5.37683487e-01 1.13581634e+00 -8.88004661e-01 -3.16088468e-01
2.43629545e-01 4.32574064e-01 -7.65881360e-01 3.73260409e-01
-7.55568326e-01 6.35012329e-01 -5.24880767e-01 6.05109334e-01
6.47622526e-01 -5.25848210e-01 1.79089755e-01 -1.48820400e-01
3.20179425e-02 -9.19614211e-02 -9.24886823e-01 1.98025286e+00
-5.70566177e-01 4.70285207e-01 -1.60835400e-01 -6.67605698e-01
8.98718357e-01 -1.67252332e-01 -1.09357707e-01 -5.41125298e-01
9.89107266e-02 1.77587166e-01 -1.33637309e-01 -4.09256250e-01
4.88492012e-01 8.09191316e-02 -2.51225203e-01 2.56216884e-01
1.56595886e-01 -2.39468783e-01 1.50331065e-01 1.76758319e-01
8.26994300e-01 4.94829923e-01 -5.16621545e-02 -3.23679239e-01
3.86768878e-01 3.23976316e-02 4.05726433e-01 9.71940041e-01
1.72776982e-01 9.67238307e-01 4.19984251e-01 -3.34337741e-01
-1.26169968e+00 -1.21213055e+00 8.74561220e-02 1.08159399e+00
3.16130817e-01 -2.09025681e-01 -1.02312565e+00 -9.40998912e-01
-1.61322564e-01 7.62146175e-01 -7.47617960e-01 -1.42053723e-01
-3.67380917e-01 -4.30177718e-01 3.66101712e-01 4.25769299e-01
5.40316045e-01 -1.45501065e+00 -4.15890396e-01 -7.51898363e-02
-8.23465660e-02 -1.11867678e+00 -5.71470737e-01 1.19923040e-01
-4.86136585e-01 -6.84528351e-01 -9.23702836e-01 -8.86766970e-01
1.01157963e+00 1.67598829e-01 1.08809304e+00 1.06700517e-01
-2.60677278e-01 9.01342481e-02 -3.90842736e-01 -2.11244136e-01
-4.07959819e-01 3.91725898e-01 -3.24167252e-01 3.09897959e-01
1.79042086e-01 -4.42088902e-01 -7.03786969e-01 2.94758946e-01
-1.21933711e+00 5.48989832e-01 8.02726984e-01 8.22012544e-01
5.00031650e-01 2.78603733e-02 4.91494030e-01 -1.09495604e+00
3.68862092e-01 -4.81663495e-01 -5.80009341e-01 4.50747311e-01
-3.27104568e-01 2.11226746e-01 9.22808290e-01 -6.53681755e-01
-1.11417508e+00 1.51902899e-01 -1.70503482e-01 -6.75066233e-01
-3.83321345e-01 1.84961706e-01 -6.81826293e-01 3.32990140e-01
6.12398863e-01 3.63647610e-01 -2.35671058e-01 -4.98096377e-01
8.26034427e-01 6.25049651e-01 5.82286894e-01 -7.43370771e-01
8.33905280e-01 5.24631143e-01 -3.60942364e-01 -7.54300058e-01
-1.09006441e+00 -8.05465579e-02 -5.88613153e-01 5.85216982e-03
1.06276786e+00 -8.50896776e-01 -2.50719845e-01 5.73420048e-01
-1.17919230e+00 -8.70821953e-01 -3.44704330e-01 -1.22024290e-01
-5.16133249e-01 7.25692064e-02 -3.36936265e-01 -6.01001740e-01
-6.32602489e-03 -1.13388813e+00 1.17553294e+00 3.82897705e-01
-2.37686291e-01 -9.33870554e-01 -4.52765793e-01 3.73768002e-01
3.70236129e-01 2.74164770e-02 1.01731777e+00 -5.53108037e-01
-9.57160354e-01 8.63348991e-02 -4.24211055e-01 6.29674494e-01
6.98366910e-02 4.89974320e-02 -1.12641287e+00 -1.44895568e-01
-1.81410134e-01 -3.81133586e-01 9.00570095e-01 1.54058129e-01
1.49827933e+00 -3.81258667e-01 -3.91446710e-01 7.85768032e-01
1.42157924e+00 9.90007371e-02 8.47310007e-01 2.26990402e-01
9.69978750e-01 6.81808591e-01 3.62925649e-01 1.49917558e-01
3.12101841e-01 5.92494309e-01 3.93309474e-01 -3.20328921e-01
-5.11417449e-01 -7.46304691e-01 3.10891479e-01 4.95918095e-01
3.75923842e-01 -5.93516827e-01 -6.65469766e-01 6.33079767e-01
-1.70599890e+00 -7.53851175e-01 5.33057079e-02 2.02512860e+00
8.51858556e-01 2.43599772e-01 -1.99671596e-01 -3.90810043e-01
8.53078902e-01 1.48424178e-01 -7.10098982e-01 -3.78827542e-01
-3.15207958e-01 1.88524038e-01 2.95330882e-01 3.72171670e-01
-1.01127243e+00 1.41127789e+00 5.05061150e+00 1.17068493e+00
-1.28882015e+00 4.24251668e-02 9.32974756e-01 3.88514847e-02
-5.59535384e-01 7.25266859e-02 -9.12557960e-01 6.62599683e-01
4.38103855e-01 1.36222631e-01 6.10791445e-01 9.26531851e-01
1.15087651e-01 5.01079150e-02 -9.77691710e-01 6.36012018e-01
2.61395782e-01 -1.35674560e+00 5.15389562e-01 -2.10112073e-02
1.01272416e+00 -1.63646713e-01 2.16572642e-01 3.82752508e-01
4.16536897e-01 -1.03454375e+00 1.04655993e+00 4.40025270e-01
1.24453402e+00 -5.08567035e-01 3.36717784e-01 4.09437567e-01
-7.39752650e-01 -1.31207272e-01 -4.70416218e-01 1.39466316e-01
3.80375981e-02 5.35379946e-01 -6.28515184e-01 4.17734683e-01
6.58609033e-01 5.36659539e-01 -7.12225795e-01 6.60384715e-01
-6.82332814e-01 3.53196681e-01 -1.44873604e-01 2.82968618e-02
3.04401398e-01 -1.40366733e-01 2.12524563e-01 1.11780429e+00
5.80241203e-01 -2.98001468e-01 3.22476402e-02 1.52299428e+00
-3.54795575e-01 3.02029564e-03 -7.87256241e-01 3.82032096e-02
3.30158472e-01 1.37892711e+00 -8.15188110e-01 -4.25861537e-01
-4.67185110e-01 1.52197671e+00 4.10309672e-01 6.71639681e-01
-8.74557853e-01 -4.06199932e-01 4.65828270e-01 3.96028370e-01
5.81005692e-01 1.50496006e-01 -4.21485394e-01 -1.33148611e+00
1.51202949e-02 -7.89688289e-01 1.22350909e-01 -1.30442846e+00
-1.41094565e+00 7.32688785e-01 -1.27243698e-01 -8.53986442e-01
6.30796254e-02 -6.75911546e-01 -7.89098442e-01 9.56101596e-01
-1.41673005e+00 -1.68225002e+00 -4.35139388e-01 4.63259012e-01
7.14161336e-01 -7.57276788e-02 6.11271560e-01 8.76855254e-02
-6.17537379e-01 6.03280008e-01 -1.48301855e-01 2.62870401e-01
6.71254098e-01 -1.23884165e+00 5.58154941e-01 8.69403720e-01
3.21752280e-01 6.67046487e-01 5.59656143e-01 -7.12931395e-01
-9.81670380e-01 -1.52062213e+00 8.08624864e-01 -5.89479208e-01
6.24558330e-01 -8.49859834e-01 -8.18884373e-01 7.41803169e-01
1.77700505e-01 -7.13524595e-02 2.67787904e-01 -1.51553452e-01
-4.49476391e-01 -7.19633996e-02 -7.76066482e-01 9.93296146e-01
1.14405835e+00 -5.33315778e-01 -3.01567286e-01 3.22246820e-01
7.53427207e-01 -2.37900123e-01 -3.29137504e-01 3.42273414e-01
3.19994271e-01 -9.45846677e-01 9.42049742e-01 -4.53379750e-01
7.35393882e-01 -4.20721233e-01 7.19474163e-03 -1.22540867e+00
-1.98457241e-01 -6.07065678e-01 2.63242841e-01 1.46991694e+00
6.19806349e-01 -3.70434374e-01 7.72776425e-01 5.33113182e-01
-2.50308484e-01 -6.25168502e-01 -5.41434765e-01 -7.60417521e-01
2.50651330e-01 -3.54000330e-01 6.74111485e-01 9.74228621e-01
-4.19116110e-01 6.23549759e-01 -4.74209130e-01 1.10269427e-01
3.62922281e-01 1.92294687e-01 8.71523857e-01 -8.20989609e-01
-4.12386805e-01 -4.65547442e-01 -3.32979560e-02 -1.43534994e+00
2.92930186e-01 -1.14228559e+00 -6.45874860e-03 -1.48631883e+00
2.41294920e-01 -6.68337524e-01 -6.44432828e-02 6.11145854e-01
-2.78280646e-01 3.11190158e-01 2.95320749e-01 2.16760784e-01
-5.48749506e-01 5.77482879e-01 1.61129355e+00 -3.19172561e-01
-1.21225871e-01 -1.38468742e-01 -8.74617577e-01 5.63226700e-01
8.85249972e-01 -4.25898999e-01 -6.06218874e-01 -6.40524805e-01
2.84940988e-01 -2.89143652e-01 5.39423883e-01 -8.52662206e-01
2.08744705e-02 -2.42823109e-01 4.18375999e-01 -2.36616328e-01
2.10882366e-01 -6.19331419e-01 9.52904075e-02 1.36564314e-01
-3.37319493e-01 -2.69579053e-01 5.78814708e-02 3.96197557e-01
-8.37849826e-02 -3.39024395e-01 6.26197100e-01 -3.42410743e-01
-9.25673783e-01 1.39970854e-01 -1.23185121e-01 1.36390477e-01
9.57162738e-01 -2.65217602e-01 -5.66259265e-01 -2.71914005e-01
-6.97450280e-01 1.51676834e-01 9.16088879e-01 7.61021376e-01
4.86607552e-01 -1.29997933e+00 -4.79289770e-01 5.22130311e-01
4.80682522e-01 2.48805568e-01 3.80085498e-01 3.65067214e-01
-3.80830497e-01 2.61016458e-01 -2.02000692e-01 -5.59891880e-01
-7.56809831e-01 8.32587004e-01 1.99444160e-01 -1.13006197e-01
-5.57438195e-01 8.49360287e-01 8.98879170e-01 -4.70051289e-01
4.29861955e-02 -3.47027987e-01 2.43559688e-01 -2.50862330e-01
2.79538393e-01 -3.03672791e-01 -2.46760890e-01 -4.82058138e-01
-7.90428519e-02 4.67526078e-01 -8.49201381e-02 4.97041531e-02
9.91099298e-01 -1.20201968e-01 4.34486195e-02 3.15638542e-01
1.12182379e+00 7.60323927e-02 -1.65085089e+00 -3.76913957e-02
-1.47453710e-01 -3.82746398e-01 -2.64211923e-01 -1.06128109e+00
-1.11225200e+00 1.07986236e+00 3.51061821e-01 -3.47552635e-02
1.00301874e+00 2.27678970e-01 6.83109224e-01 1.81607977e-01
2.55783617e-01 -1.09238291e+00 3.77444476e-01 3.66767108e-01
9.61216390e-01 -1.13329470e+00 -5.51055014e-01 -5.21524012e-01
-9.33260620e-01 8.07478130e-01 8.19844842e-01 -9.16108489e-02
2.13034153e-01 2.31674761e-01 2.22574860e-01 -9.47763845e-02
-6.17810309e-01 -3.72118920e-01 2.54313469e-01 7.61476755e-01
2.31168494e-01 1.19756930e-01 2.54945010e-01 6.29111469e-01
-2.66441077e-01 -4.36049215e-02 5.09598553e-01 5.71163177e-01
-3.47159714e-01 -1.11859345e+00 -1.70437202e-01 3.09186786e-01
-1.94203213e-01 -2.66564429e-01 -4.53539997e-01 7.43681729e-01
4.99348521e-01 7.24366665e-01 1.27993941e-01 -2.67487556e-01
2.16805562e-01 4.82780002e-02 2.90748030e-01 -9.18419421e-01
-4.35464650e-01 4.68082726e-02 -9.13386419e-02 -3.93363595e-01
-2.78462321e-02 -3.30374867e-01 -1.00939596e+00 -8.85344893e-02
-2.06528053e-01 -2.06791759e-01 5.81035316e-01 6.87699497e-01
4.07131732e-01 5.88457406e-01 4.55237955e-01 -7.31356144e-01
-2.90558398e-01 -7.14420140e-01 -5.26623845e-01 8.04854035e-01
7.79727921e-02 -4.48268503e-01 -2.70672500e-01 2.41814792e-01] | [11.299942016601562, -0.14474837481975555] |
06c1d154-15fc-4383-9c6e-a5cfcaebdbe2 | combinatorial-3d-shape-generation-via | 2004.07414 | null | https://arxiv.org/abs/2004.07414v2 | https://arxiv.org/pdf/2004.07414v2.pdf | Combinatorial 3D Shape Generation via Sequential Assembly | Sequential assembly with geometric primitives has drawn attention in robotics and 3D vision since it yields a practical blueprint to construct a target shape. However, due to its combinatorial property, a greedy method falls short of generating a sequence of volumetric primitives. To alleviate this consequence induced by a huge number of feasible combinations, we propose a combinatorial 3D shape generation framework. The proposed framework reflects an important aspect of human generation processes in real life -- we often create a 3D shape by sequentially assembling unit primitives with geometric constraints. To find the desired combination regarding combination evaluations, we adopt Bayesian optimization, which is able to exploit and explore efficiently the feasible regions constrained by the current primitive placements. An evaluation function conveys global structure guidance for an assembly process and stability in terms of gravity and external forces simultaneously. Experimental results demonstrate that our method successfully generates combinatorial 3D shapes and simulates more realistic generation processes. We also introduce a new dataset for combinatorial 3D shape generation. All the codes are available at \url{https://github.com/POSTECH-CVLab/Combinatorial-3D-Shape-Generation}. | ['Jaesik Park', 'Minsu Cho', 'Jinhwi Lee', 'Hyunsoo Chung', 'Jungtaek Kim'] | 2020-04-16 | null | null | null | null | ['3d-shape-generation'] | ['computer-vision'] | [ 2.28763986e-02 1.95868254e-01 2.70709753e-01 1.25049740e-01
-3.24786842e-01 -8.10962856e-01 8.49839389e-01 3.03231716e-01
6.33731112e-02 5.63694835e-01 -1.36534691e-01 -1.26315162e-01
-1.58567578e-01 -9.99927223e-01 -6.87442720e-01 -6.61796927e-01
-9.70693212e-03 1.02228093e+00 1.02600664e-01 -2.44843662e-01
4.23134476e-01 9.19440567e-01 -1.63297975e+00 -2.23407805e-01
9.73724186e-01 8.24464917e-01 7.34982669e-01 4.30634201e-01
-8.74202996e-02 -2.32419237e-01 -3.37105751e-01 -2.82029808e-01
6.29580438e-01 -1.86360225e-01 -4.19895411e-01 3.83145481e-01
3.01638432e-02 -4.97134328e-02 1.68705836e-01 8.84826601e-01
5.49853504e-01 1.91958934e-01 1.10998452e+00 -1.16800821e+00
-4.00101721e-01 3.04598987e-01 -4.80659962e-01 -5.38137734e-01
5.45066893e-01 2.88757920e-01 9.44749057e-01 -1.15809703e+00
7.60276556e-01 1.29437494e+00 3.58239740e-01 4.01064217e-01
-1.28843606e+00 -3.11786562e-01 3.66572551e-02 -4.92898345e-01
-1.31073236e+00 -2.08593592e-01 9.85749602e-01 -5.66326022e-01
5.18819273e-01 3.60314846e-01 1.08169687e+00 7.75814593e-01
3.81504714e-01 8.69261920e-01 7.11734295e-01 -4.14352924e-01
3.97197872e-01 -2.83966988e-01 -4.64232504e-01 8.48305464e-01
3.15723330e-01 1.57951728e-01 -2.88123727e-01 -3.28106225e-01
1.25786185e+00 -1.14364073e-01 -9.12416652e-02 -7.60644197e-01
-1.45740163e+00 6.08066380e-01 4.20736909e-01 1.28332183e-01
-4.65582967e-01 2.74562150e-01 -5.48423789e-02 -1.75982311e-01
2.04256266e-01 7.43301213e-01 -3.18397880e-01 -4.03194353e-02
-4.82796222e-01 7.44515240e-01 8.62853289e-01 1.23898947e+00
8.09819520e-01 1.38108749e-02 1.31478813e-02 8.88283491e-01
2.46077448e-01 6.31676197e-01 -9.33025107e-02 -9.86625433e-01
3.09176326e-01 6.73072696e-01 2.25033849e-01 -1.08522797e+00
-4.31196183e-01 -2.52869129e-01 -9.73070502e-01 3.37185204e-01
3.74884427e-01 8.52383580e-03 -7.91454673e-01 1.24871898e+00
7.72334397e-01 -1.95318326e-01 -4.42012131e-01 9.09597218e-01
5.30721128e-01 6.07306361e-01 -3.30726922e-01 -1.75932586e-01
1.03277779e+00 -7.09289670e-01 -1.96959049e-01 6.08583242e-02
1.57260939e-01 -9.60818887e-01 8.40304017e-01 3.97326499e-01
-1.26234484e+00 -4.14820671e-01 -7.41112053e-01 1.47378549e-01
7.73556083e-02 2.11579055e-01 6.28402114e-01 4.07634050e-01
-8.20870101e-01 6.42289579e-01 -6.15878701e-01 -1.30157292e-01
4.56838220e-01 3.24166387e-01 -1.39044598e-01 1.30583286e-01
-4.61815566e-01 6.88809335e-01 4.91421223e-01 1.88958317e-01
-8.79395485e-01 -5.95921397e-01 -6.92822099e-01 -2.45025963e-01
5.78145146e-01 -1.11348665e+00 1.16618359e+00 -1.28656313e-01
-1.62266791e+00 6.21204138e-01 -3.22596021e-02 2.80256737e-02
7.22544551e-01 -1.05663642e-01 5.71229875e-01 -4.48288284e-02
6.13970980e-02 9.94360149e-01 1.02371883e+00 -1.56417286e+00
-2.39564151e-01 -2.65680969e-01 2.35853661e-02 3.89176935e-01
2.42853969e-01 -5.46307981e-01 -4.31216866e-01 -8.54260981e-01
5.06722867e-01 -1.10127616e+00 -8.12987924e-01 1.03534006e-01
-7.75473535e-01 -3.21380287e-01 3.34402174e-01 -2.49352068e-01
7.44934797e-01 -1.69687271e+00 5.18031180e-01 4.58461523e-01
9.61414650e-02 -1.40726015e-01 -1.00263488e-02 8.25942278e-01
2.99509406e-01 1.86564654e-01 -5.32243609e-01 -4.62888956e-01
6.79379031e-02 1.11488760e-01 -2.84523904e-01 3.14956158e-01
2.01150000e-01 9.07181025e-01 -9.68816519e-01 -4.06086475e-01
4.98142540e-01 3.53042334e-01 -7.66862988e-01 2.52165377e-01
-6.73934937e-01 8.14527512e-01 -8.94027114e-01 9.77482319e-01
7.44009316e-01 6.95211142e-02 9.03663412e-02 -4.96468693e-02
-2.31709316e-01 -1.15180939e-01 -1.22998762e+00 1.85726655e+00
-3.07174027e-01 -1.37545064e-01 1.20229319e-01 -6.87328637e-01
1.35279202e+00 4.21429239e-02 6.00776076e-01 -6.10828400e-02
4.07502234e-01 4.13789332e-01 -2.00854793e-01 -1.18777469e-01
5.74603736e-01 -1.15208037e-01 -2.34125435e-01 4.24901426e-01
-2.75362283e-01 -1.31959391e+00 1.30882993e-01 -9.94070619e-02
7.62818038e-01 5.89216232e-01 4.53859955e-01 -4.03890431e-01
3.14555615e-01 -1.71109922e-02 3.80893528e-01 5.68778396e-01
3.26369077e-01 7.76660562e-01 4.07412469e-01 -4.75112110e-01
-1.33982730e+00 -1.17706311e+00 -3.63624021e-02 3.93198729e-01
4.97979045e-01 -2.88432360e-01 -5.43847919e-01 -2.15073571e-01
1.70466259e-01 6.95463359e-01 -3.65575552e-01 2.02747017e-01
-6.95020437e-01 -4.31216389e-01 1.30993472e-02 7.93397203e-02
3.00461091e-02 -1.24024642e+00 -9.40222979e-01 3.94243211e-01
-7.95690715e-02 -8.54093075e-01 -3.90832156e-01 -8.78081545e-02
-1.12997854e+00 -1.00849485e+00 -8.06515872e-01 -6.52924120e-01
9.96013284e-01 3.04817647e-01 1.04915297e+00 3.10095519e-01
-6.13019764e-01 3.65514159e-01 -3.96065116e-01 -4.75435019e-01
-5.87819874e-01 -5.42174205e-02 8.40059817e-02 -1.89872369e-01
-5.29797256e-01 -9.17164922e-01 -6.55613899e-01 3.66581142e-01
-7.07372367e-01 4.99168724e-01 5.86050749e-01 6.06396019e-01
9.61386085e-01 7.37770274e-02 3.32945526e-01 -1.16496168e-01
5.92447281e-01 -2.22080216e-01 -8.01209986e-01 2.82687563e-02
1.55573219e-01 1.24590650e-01 5.15533507e-01 -3.15041274e-01
-9.08213973e-01 3.18079859e-01 -7.57044926e-02 -5.90270102e-01
-3.51641715e-01 1.54201061e-01 -2.88302213e-01 1.36159554e-01
3.28836441e-01 2.60471731e-01 2.49231197e-02 -5.37276328e-01
6.19054139e-01 3.69496271e-02 3.04442346e-01 -1.19860971e+00
9.86360133e-01 5.59757769e-01 4.39680636e-01 -1.04444730e+00
-2.22549871e-01 -2.59708583e-01 -9.38420117e-01 -4.82266814e-01
6.85978651e-01 -3.87684464e-01 -8.64222705e-01 5.29876471e-01
-1.46076155e+00 -2.77058154e-01 -3.54020804e-01 1.78739727e-01
-1.02959526e+00 4.32429880e-01 -9.86277163e-02 -1.12128830e+00
-3.38370740e-01 -1.42982292e+00 1.38537848e+00 -1.30108558e-02
-2.57405549e-01 -5.89011133e-01 -1.96379311e-02 -3.78799438e-03
-2.62454655e-02 7.91550159e-01 8.85842681e-01 -4.48564999e-02
-9.39699411e-01 -1.55604780e-01 1.13816932e-01 -2.23106816e-01
3.44445407e-01 1.66787609e-01 -4.49852318e-01 -1.21868081e-01
-2.42437631e-01 -8.47310722e-02 4.28016037e-01 4.76805300e-01
1.17097688e+00 -1.67082578e-01 -4.05824363e-01 4.64088261e-01
1.34752011e+00 1.82141140e-01 4.11897808e-01 4.09056470e-02
6.66225493e-01 7.85318196e-01 6.27864242e-01 8.71201277e-01
2.39155546e-01 8.62305880e-01 8.68946612e-01 3.99635613e-01
-2.79045943e-03 -5.31133354e-01 -6.44924715e-02 7.04929948e-01
-4.78578717e-01 -2.71022260e-01 -9.50179696e-01 4.42805916e-01
-1.65898657e+00 -6.53505266e-01 -2.83062279e-01 2.36593533e+00
7.07526803e-01 3.07453070e-02 1.47906914e-01 5.57848327e-02
7.37928510e-01 -1.10300742e-01 -3.72638166e-01 -2.56412417e-01
2.70059574e-02 1.12849131e-01 2.27692276e-01 5.92288733e-01
-8.32450807e-01 9.60710287e-01 5.39824390e+00 1.06829309e+00
-7.82473743e-01 -3.91850173e-01 4.85399872e-01 -2.98159495e-02
-6.20504856e-01 2.40050759e-02 -8.00969958e-01 3.02067131e-01
-8.46998543e-02 -1.98767096e-01 3.77555221e-01 6.68273926e-01
2.91981518e-01 -3.16389233e-01 -9.90966797e-01 1.06553948e+00
-2.24991366e-01 -1.38663030e+00 3.94243479e-01 3.06003392e-01
8.56717706e-01 -3.87868345e-01 -5.67610338e-02 -2.75738925e-01
2.72334725e-01 -8.49970877e-01 1.33667374e+00 5.87434888e-01
6.77044690e-01 -6.30939186e-01 1.36127278e-01 7.02908456e-01
-1.39211833e+00 1.35050490e-01 -4.17944282e-01 -4.35071960e-02
6.32093489e-01 9.39384162e-01 -1.18865848e+00 8.35826099e-01
2.52866238e-01 3.53125215e-01 -2.17284590e-01 1.31349266e+00
-2.83316195e-01 -6.48438409e-02 -6.45566702e-01 -3.01421106e-01
9.29093510e-02 -7.96155512e-01 1.15600669e+00 7.50811100e-01
7.92969227e-01 6.90757334e-02 4.07617182e-01 1.15499759e+00
2.72291929e-01 1.53763205e-01 -9.47373629e-01 3.80503051e-02
6.30137801e-01 1.13353479e+00 -1.28427565e+00 1.56014003e-02
3.14081609e-01 7.47859657e-01 2.05020711e-01 5.28861135e-02
-7.51506269e-01 8.08216408e-02 2.82542735e-01 1.80416584e-01
3.01566124e-01 -8.62572908e-01 -7.31135011e-01 -8.98956418e-01
1.07450470e-01 -4.89268601e-01 -1.81725353e-01 -7.39279389e-01
-9.70667303e-01 4.08230811e-01 3.03904027e-01 -1.27620053e+00
-1.29964352e-01 -4.30641472e-01 -5.36953449e-01 5.64076126e-01
-8.96935999e-01 -1.23157990e+00 -2.10193291e-01 1.34021729e-01
9.21184838e-01 2.31782719e-02 6.16833806e-01 -2.40040138e-01
-1.32132530e-01 6.82321936e-02 -4.11330909e-01 -4.72391874e-01
1.34505123e-01 -1.18766546e+00 5.72565317e-01 6.46854997e-01
8.05813968e-02 4.26668406e-01 8.70305598e-01 -1.01906610e+00
-1.67449749e+00 -7.75179386e-01 4.26505208e-01 -4.67049569e-01
4.15274560e-01 -4.02847618e-01 -5.09724736e-01 2.21726134e-01
-1.13958001e-01 -4.93223637e-01 8.26406851e-02 -3.54890585e-01
2.14487463e-01 3.18557680e-01 -1.08030999e+00 9.66363788e-01
1.60570753e+00 2.62786269e-01 -5.12076080e-01 2.28387296e-01
6.41386092e-01 -5.58401227e-01 -7.45766878e-01 7.26592243e-01
6.73221231e-01 -9.22227144e-01 1.05356371e+00 4.53329384e-02
4.00602251e-01 -5.06639898e-01 -1.62913308e-01 -1.40846753e+00
-1.32020295e-01 -9.83594298e-01 -8.64038765e-02 8.38597298e-01
2.59524703e-01 -5.47590852e-01 7.23769069e-01 3.92076612e-01
-4.77707297e-01 -1.07656884e+00 -9.51000094e-01 -9.04053390e-01
-6.70404732e-02 -5.60536504e-01 8.26196551e-01 4.80742514e-01
-3.95722568e-01 -5.51345311e-02 -6.77136630e-02 1.29349947e-01
9.25043523e-01 5.46184719e-01 1.12600911e+00 -1.21706748e+00
-2.74962425e-01 -7.90302932e-01 -1.78928971e-01 -1.37615132e+00
-1.18728913e-01 -8.79128098e-01 2.28852481e-01 -1.70972991e+00
-2.42106870e-01 -1.13305426e+00 5.31472445e-01 2.56498665e-01
1.21184357e-01 8.43130052e-02 4.81699049e-01 2.35576302e-01
-8.59258473e-02 9.72879231e-01 1.90318727e+00 8.63034427e-02
-5.08333743e-01 1.56160817e-01 -3.97119939e-01 7.79818535e-01
8.43448818e-01 -1.74902797e-01 -3.02747905e-01 -4.02396202e-01
3.60328496e-01 2.56445169e-01 4.22852695e-01 -7.50388026e-01
5.13841547e-02 -4.43647057e-01 1.73023909e-01 -1.16496098e+00
6.64238811e-01 -7.45939314e-01 5.89682341e-01 5.45257747e-01
1.99603319e-01 1.12012476e-02 8.84770229e-02 5.43362498e-01
3.32669318e-01 -3.77657562e-01 6.71033263e-01 -4.16903734e-01
-3.45333189e-01 5.26917517e-01 -2.14288965e-01 -2.67990559e-01
1.13371146e+00 -5.27361572e-01 1.37525797e-01 -1.38676599e-01
-7.77670979e-01 2.37304255e-01 8.44059289e-01 4.14570034e-01
9.71259654e-01 -1.39242172e+00 -7.67571390e-01 2.60972649e-01
-1.46236673e-01 7.14779377e-01 3.40013206e-02 6.26395881e-01
-9.09825623e-01 1.25522733e-01 -1.63830876e-01 -9.02586222e-01
-8.57055128e-01 3.50755990e-01 1.03721268e-01 1.17519230e-03
-8.11087012e-01 9.25259650e-01 5.03427207e-01 -6.47615373e-01
-1.26726657e-01 -4.95522350e-01 1.20532013e-01 -5.85700339e-03
4.73020002e-02 5.16983271e-01 -7.56171793e-02 -4.92612004e-01
-1.69629499e-01 9.73267198e-01 5.38653314e-01 -1.87188774e-01
1.41528857e+00 3.74440923e-02 -2.86909878e-01 1.61999926e-01
6.64001405e-01 1.33756310e-01 -1.34125078e+00 2.88040072e-01
-1.76608130e-01 -7.25143015e-01 -5.48121929e-01 -2.69873887e-01
-8.51864874e-01 5.67918599e-01 -2.47610927e-01 2.67636955e-01
7.04887092e-01 2.56736457e-01 6.55084014e-01 2.58570939e-01
1.01713908e+00 -7.31665492e-01 2.90001690e-01 5.96674621e-01
1.59520078e+00 -7.01650739e-01 7.69789517e-02 -8.85113955e-01
-3.54771346e-01 1.23535788e+00 4.20996875e-01 -4.27453578e-01
6.48536980e-01 1.78842768e-01 -4.19270337e-01 -2.98227191e-01
-4.86071378e-01 -7.95137659e-02 2.86496997e-01 5.76099157e-01
-6.47624880e-02 3.22299331e-01 -3.89487475e-01 1.40736535e-01
-6.34682357e-01 -3.80015016e-01 5.51054955e-01 8.42413783e-01
-5.09099901e-01 -1.26518619e+00 -5.63463807e-01 2.97823966e-01
8.93700942e-02 2.41495520e-01 -2.79845178e-01 5.55328667e-01
1.54689699e-01 6.86615109e-01 -4.32957821e-02 -7.43175298e-02
2.78466225e-01 -2.64185756e-01 8.52104604e-01 -8.11635554e-01
-2.31042981e-01 3.94372165e-01 6.33335114e-02 -4.60977465e-01
-2.23690912e-01 -7.51432598e-01 -1.17021191e+00 -3.91086750e-02
-3.25677127e-01 -3.81534360e-02 7.01584041e-01 6.18686974e-01
2.90474117e-01 1.96537465e-01 6.10960424e-01 -1.81246185e+00
-3.27889323e-01 -6.29565001e-01 -4.83650565e-01 6.99394792e-02
-1.62357375e-01 -1.08439565e+00 -2.00782910e-01 4.03801277e-02] | [8.79296588897705, -3.616032838821411] |
86ec4ecf-c3df-4a3a-8f88-f58a6d9460fb | towards-grand-unification-of-object-tracking | 2207.07078 | null | https://arxiv.org/abs/2207.07078v4 | https://arxiv.org/pdf/2207.07078v4.pdf | Towards Grand Unification of Object Tracking | We present a unified method, termed Unicorn, that can simultaneously solve four tracking problems (SOT, MOT, VOS, MOTS) with a single network using the same model parameters. Due to the fragmented definitions of the object tracking problem itself, most existing trackers are developed to address a single or part of tasks and overspecialize on the characteristics of specific tasks. By contrast, Unicorn provides a unified solution, adopting the same input, backbone, embedding, and head across all tracking tasks. For the first time, we accomplish the great unification of the tracking network architecture and learning paradigm. Unicorn performs on-par or better than its task-specific counterparts in 8 tracking datasets, including LaSOT, TrackingNet, MOT17, BDD100K, DAVIS16-17, MOTS20, and BDD100K MOTS. We believe that Unicorn will serve as a solid step towards the general vision model. Code is available at https://github.com/MasterBin-IIAU/Unicorn. | ['Huchuan Lu', 'Ping Luo', 'Zehuan Yuan', 'Dong Wang', 'Peize Sun', 'Yi Jiang', 'Bin Yan'] | 2022-07-14 | null | null | null | null | ['visual-object-tracking', 'multi-object-tracking-and-segmentation'] | ['computer-vision', 'computer-vision'] | [-4.68090773e-01 -3.09032708e-01 -3.47013265e-01 5.80516197e-02
-1.02271877e-01 -9.20267105e-01 5.91871977e-01 -4.55679357e-01
-4.74279851e-01 6.37243807e-01 -1.20723672e-01 -6.75609112e-02
-2.02883363e-01 -2.23728120e-01 -5.57810009e-01 -4.07089859e-01
-1.82638928e-01 4.98225451e-01 7.25336552e-01 -5.32926992e-02
-3.05050135e-01 3.25223655e-01 -1.46884310e+00 -2.74659097e-01
2.80622691e-01 8.43455732e-01 2.96257883e-01 9.38232541e-01
1.27427325e-01 6.51556969e-01 -3.61244887e-01 -5.59313178e-01
4.83102262e-01 7.43586645e-02 -7.71593809e-01 -3.11674923e-01
1.11767375e+00 -1.88076958e-01 -8.67572784e-01 1.17734873e+00
6.43377125e-01 -1.14431113e-01 3.84948611e-01 -1.93725872e+00
-6.82383597e-01 5.69385946e-01 -4.70896602e-01 5.74647188e-01
2.65875109e-03 3.79215360e-01 1.03955138e+00 -5.26247084e-01
8.11616600e-01 1.21932888e+00 1.39815831e+00 8.68640542e-01
-1.27022350e+00 -8.79973233e-01 2.83489048e-01 4.59915958e-02
-1.22461748e+00 -5.48365653e-01 1.26332849e-01 -7.79034972e-01
6.11029506e-01 2.15443209e-01 8.27833593e-01 1.46233821e+00
3.27325016e-01 9.91766512e-01 6.58836305e-01 1.10210419e-01
-2.10358799e-01 -1.05794400e-01 5.53959668e-01 7.16027141e-01
8.59206259e-01 4.35073137e-01 -5.53082943e-01 -1.75566778e-01
6.50811255e-01 5.71501032e-02 -3.62877578e-01 -6.87765300e-01
-1.47301817e+00 6.11808956e-01 4.48892772e-01 2.44996727e-01
-1.04266763e-01 7.91112363e-01 5.02651334e-01 4.33304697e-01
2.45859906e-01 2.20747381e-01 -5.95159709e-01 -2.38315046e-01
-9.12495911e-01 4.14590865e-01 9.09142077e-01 1.25953186e+00
5.85885704e-01 2.92773247e-01 -2.88376808e-01 1.90545261e-01
6.03131235e-01 8.89761150e-01 8.91176164e-02 -1.22384799e+00
-6.84543550e-02 1.42644837e-01 3.46461833e-01 -8.36039960e-01
-6.92906201e-01 -6.57731295e-01 -3.87781978e-01 1.77161813e-01
4.87618238e-01 -4.80744272e-01 -7.47799158e-01 2.08502364e+00
5.39116859e-01 6.33215606e-01 -2.64219314e-01 9.03049409e-01
1.13960814e+00 2.16328248e-01 1.61358878e-01 3.61472398e-01
1.54282451e+00 -1.23957837e+00 -6.56672537e-01 -2.48604745e-01
5.75334013e-01 -7.43115306e-01 5.05585134e-01 3.14736599e-03
-7.66663671e-01 -7.19754577e-01 -8.35840285e-01 1.32230192e-01
-5.62228739e-01 1.62094653e-01 8.69515777e-01 6.33158267e-01
-1.65611899e+00 4.53297913e-01 -1.02118897e+00 -9.87904549e-01
2.46142656e-01 4.72684830e-01 -1.62059486e-01 2.88713545e-01
-9.57676470e-01 9.42407787e-01 3.93996328e-01 5.85013181e-02
-1.03750336e+00 -7.81625450e-01 -6.22080803e-01 -2.96415418e-01
4.70448196e-01 -1.05982912e+00 1.47991800e+00 -2.20150113e-01
-1.00255883e+00 8.36858869e-01 6.32510185e-02 -6.10421002e-01
5.54053962e-01 -4.07592773e-01 -4.26764905e-01 -3.29854280e-01
3.47131938e-01 1.14764917e+00 7.54473448e-01 -1.03016615e+00
-6.74739122e-01 -5.08355089e-02 1.19887449e-01 -1.65161807e-02
-2.45646492e-01 1.14832118e-01 -7.69316614e-01 -4.33547348e-01
-4.55757886e-01 -1.14664316e+00 9.34524238e-02 5.28702378e-01
-3.78529668e-01 -4.07849967e-01 1.31766284e+00 -1.83868393e-01
1.03748381e+00 -2.11304641e+00 3.46986026e-01 -4.08877194e-01
7.39842474e-01 6.47611797e-01 -3.83196384e-01 4.31158662e-01
2.55647749e-01 -2.37026364e-01 2.15105996e-01 -6.20898008e-01
3.62714797e-01 3.38160455e-01 -3.84236290e-03 7.13113546e-01
-2.44268477e-01 1.23820925e+00 -1.10616517e+00 -6.03973091e-01
2.19552830e-01 4.97876763e-01 -2.92616189e-01 -1.38156444e-01
-2.91098267e-01 4.30171818e-01 -4.97116834e-01 7.76033878e-01
5.89928687e-01 -7.09425747e-01 -2.66027125e-03 -3.40141624e-01
-2.42911085e-01 -4.84719239e-02 -1.30245352e+00 1.61323047e+00
2.12019697e-01 1.09651315e+00 4.39345181e-01 -5.32191575e-01
5.22511303e-01 2.29766309e-01 9.44040298e-01 -3.60538721e-01
2.05337673e-01 3.56869213e-02 3.12052225e-03 -1.74616575e-01
7.00626373e-01 5.34665287e-01 1.75751969e-01 4.53440189e-01
4.12086099e-01 5.83046198e-01 3.52989256e-01 4.85886246e-01
1.38011789e+00 3.75449806e-01 -3.79374978e-04 -3.46347272e-01
2.49471828e-01 1.94193602e-01 6.44376040e-01 1.19178128e+00
-8.34213138e-01 3.22839826e-01 1.03075445e-01 -5.52754402e-01
-8.17758262e-01 -1.20135152e+00 -2.14541659e-01 1.21725607e+00
3.04355770e-01 -7.90743470e-01 -4.30855751e-01 -6.83086574e-01
5.31682789e-01 4.42319810e-02 -6.37448847e-01 9.07557607e-02
-4.71846253e-01 -4.58592623e-01 1.09388256e+00 5.13861656e-01
4.30140227e-01 -1.22201014e+00 -6.44513249e-01 8.93237963e-02
-1.17576301e-01 -1.46658623e+00 -7.32993782e-01 1.06959939e-01
-5.99161386e-01 -1.33311224e+00 -7.13077366e-01 -5.44165313e-01
1.47347689e-01 6.92149758e-01 1.22276294e+00 2.93850124e-01
-2.39919186e-01 8.96892667e-01 -1.27777472e-01 -4.03773397e-01
1.63366590e-02 4.03744072e-01 4.60858047e-01 -3.23366821e-01
4.27500457e-01 -3.19901139e-01 -3.67383599e-01 5.29565811e-01
-3.85963887e-01 -2.34402567e-01 2.88252950e-01 4.14084494e-01
2.84468681e-01 -3.74256581e-01 2.27871493e-01 -5.16768992e-01
2.79035896e-01 -6.56117022e-01 -9.68899846e-01 1.49022102e-01
-4.20854121e-01 -3.95915836e-01 2.36797363e-01 -6.83976114e-01
-5.15285730e-01 1.02234241e-02 1.83115795e-01 -8.79027128e-01
8.87323096e-02 1.03773065e-02 2.68597335e-01 -5.88263571e-01
5.54164290e-01 2.05670342e-01 6.00102358e-04 -5.39559662e-01
3.62062424e-01 2.29007289e-01 7.51418173e-01 -4.63891894e-01
1.32804859e+00 6.88975751e-01 -1.10519357e-01 -7.37365425e-01
-7.94162214e-01 -7.19232738e-01 -7.08994448e-01 -5.66688716e-01
1.06164229e+00 -1.04606867e+00 -1.00136220e+00 6.16228282e-01
-1.05904114e+00 -5.37873387e-01 -3.09775412e-01 5.50528288e-01
-2.98436016e-01 4.53799367e-01 -5.85653365e-01 -5.38538158e-01
-2.87648469e-01 -9.00002420e-01 1.08772409e+00 4.27469939e-01
-3.19741331e-02 -1.17656636e+00 3.47578406e-01 3.45782712e-02
7.39718318e-01 2.36877918e-01 -1.53569967e-01 -5.03314912e-01
-9.58339453e-01 1.00146279e-01 -4.76282477e-01 -3.56037468e-02
8.20978731e-02 1.46773651e-01 -7.25476980e-01 -8.42069566e-01
-4.68787104e-01 -3.60721260e-01 1.01081979e+00 6.98249400e-01
4.75101054e-01 1.86408430e-01 -8.14076960e-01 1.02630520e+00
1.30633640e+00 3.75898206e-03 6.14539646e-02 4.71652269e-01
8.31682861e-01 4.81394166e-03 4.28137630e-01 1.43588826e-01
6.36609852e-01 9.27205145e-01 6.85665071e-01 5.61718158e-02
-6.12777770e-01 -9.50741954e-03 7.01110721e-01 6.53411567e-01
-2.55719339e-03 -3.48747075e-01 -8.43671143e-01 6.69070780e-01
-2.16279602e+00 -1.00215805e+00 -5.74389338e-01 1.66137409e+00
2.67509878e-01 3.46890800e-02 5.36766768e-01 -7.38731086e-01
7.98733950e-01 5.14597118e-01 -6.70374572e-01 3.55668545e-01
1.35167353e-02 -2.75218636e-01 8.34595025e-01 2.64901131e-01
-1.18065906e+00 1.16112530e+00 6.87368727e+00 5.46969771e-01
-9.33488369e-01 3.77912939e-01 -5.94702482e-01 -1.92440689e-01
3.15111101e-01 6.45762589e-03 -1.47719049e+00 6.91751182e-01
8.32244992e-01 5.84386929e-04 4.41198349e-01 6.85754299e-01
-1.83994219e-01 3.01095545e-01 -1.06124592e+00 9.29518819e-01
-2.18660578e-01 -1.45489085e+00 -4.51163441e-01 2.77407885e-01
4.11174625e-01 1.00886333e+00 4.22944315e-02 4.27301586e-01
8.55852842e-01 -5.14440775e-01 6.89398944e-01 3.92252803e-01
6.08979702e-01 1.23695515e-01 5.36286056e-01 1.81413949e-01
-1.80972958e+00 1.77178755e-01 -5.64652383e-01 1.73092887e-01
2.52815127e-01 -6.21137209e-02 -3.97117764e-01 6.54051483e-01
1.25819719e+00 1.26035011e+00 -8.04893076e-01 1.49424231e+00
8.11847076e-02 4.62903351e-01 -5.71188450e-01 9.53375548e-02
3.97465438e-01 1.18879221e-01 1.03653848e+00 1.26563704e+00
6.46531507e-02 -4.43559945e-01 3.80964994e-01 7.19904959e-01
-6.18542694e-02 -5.81746995e-01 -7.68150926e-01 -1.41635705e-02
8.01062703e-01 1.51145530e+00 -7.10350215e-01 -1.59354582e-01
-5.90212882e-01 4.78887290e-01 3.81526411e-01 3.36737961e-01
-1.37573922e+00 7.52909034e-02 1.14395344e+00 -1.01232797e-01
7.91395962e-01 -2.97355145e-01 3.80530864e-01 -1.31593728e+00
-2.04366565e-01 -7.41031945e-01 4.27832276e-01 -7.28730381e-01
-1.23763502e+00 4.95621085e-01 9.12605319e-04 -1.28625774e+00
2.30136901e-01 -6.86219513e-01 -6.77537560e-01 4.16026682e-01
-1.52117920e+00 -1.17406976e+00 -5.92692673e-01 8.08657467e-01
3.16227257e-01 -3.14091355e-01 3.69591832e-01 8.06414843e-01
-7.69840956e-01 4.98166174e-01 1.40322700e-01 4.24308956e-01
8.10985029e-01 -1.19758224e+00 7.54248321e-01 8.76895964e-01
2.68905878e-01 4.97154564e-01 6.65977597e-01 -7.70763040e-01
-1.48213029e+00 -1.44672656e+00 6.63029432e-01 -9.13792074e-01
1.05310607e+00 -4.01854485e-01 -5.56692779e-01 1.41696072e+00
5.54066002e-01 2.25712121e-01 2.15277627e-01 2.95549780e-01
-3.31141889e-01 6.81923255e-02 -8.47059965e-01 5.37980914e-01
1.65267646e+00 -2.31490523e-01 -4.87305939e-01 5.00683010e-01
7.66368270e-01 -6.08498633e-01 -1.00334144e+00 1.42138496e-01
6.05003595e-01 -7.70442069e-01 1.22703803e+00 -6.96261644e-01
-4.82569605e-01 -7.98883319e-01 -1.01518102e-01 -1.03591394e+00
-8.11067224e-01 -9.62823033e-01 -6.01436377e-01 1.27520072e+00
5.36377206e-02 -8.22817802e-01 9.08135116e-01 -1.09530598e-01
-2.71692127e-01 -3.20348203e-01 -9.70960438e-01 -1.32495248e+00
-3.23599786e-01 -2.49164149e-01 5.17212510e-01 9.08160329e-01
-8.08375955e-01 2.55764216e-01 -7.18914926e-01 2.48203367e-01
1.28913319e+00 -1.28224552e-01 1.22031188e+00 -1.71497035e+00
-2.61098415e-01 -3.21826160e-01 -5.06598353e-01 -1.30593204e+00
4.15801965e-02 -9.40300584e-01 -2.12950990e-01 -1.47683811e+00
1.88432664e-01 -6.83488369e-01 -3.13132793e-01 7.30426431e-01
6.62956759e-02 4.72892046e-01 7.25989878e-01 5.61902404e-01
-1.32322419e+00 3.18985701e-01 1.27927971e+00 -2.31815875e-02
1.43157512e-01 2.04519972e-01 -7.03294754e-01 6.41221464e-01
7.16191649e-01 -8.80824864e-01 -1.55959249e-01 -8.26883435e-01
-1.30476728e-02 -3.19056392e-01 7.93758631e-01 -1.38260877e+00
9.12490249e-01 2.51302570e-01 2.12702408e-01 -9.26156640e-01
5.23504019e-01 -8.48115802e-01 6.39622808e-01 7.53668070e-01
1.14635572e-01 3.43358278e-01 3.39082152e-01 6.77046359e-01
1.49208292e-01 4.96396795e-03 6.42052233e-01 4.65863943e-02
-1.17009211e+00 5.95743299e-01 -1.01603858e-01 4.18810815e-01
9.83933270e-01 -3.92227888e-01 -9.29183960e-01 -1.89199939e-01
-7.22080410e-01 8.78612578e-01 4.93827850e-01 8.22607577e-01
1.05339833e-01 -1.47446179e+00 -6.36851132e-01 -2.04758734e-01
-2.29286347e-02 -3.68746251e-01 -4.53360714e-02 1.21443951e+00
-2.42842212e-01 6.66573465e-01 -4.66788203e-01 -1.04033446e+00
-1.30484605e+00 4.10959512e-01 5.93897641e-01 -3.42233986e-01
-8.86898875e-01 7.82271981e-01 5.57058379e-02 -6.81385458e-01
5.17938793e-01 1.30739138e-02 4.02049236e-02 -1.00789638e-02
2.87988275e-01 4.68686700e-01 -4.21118170e-01 -8.06253195e-01
-6.52156949e-01 7.67927885e-01 -6.06691353e-02 4.33916241e-01
1.09950244e+00 -1.52530357e-01 2.38924667e-01 1.01755843e-01
7.13584423e-01 -3.47039521e-01 -1.56666601e+00 -4.92794931e-01
1.45220846e-01 -2.12368771e-01 -8.52440521e-02 -4.22964066e-01
-1.34195364e+00 3.74479920e-01 6.67460620e-01 3.94019842e-01
6.00648522e-01 3.25554073e-01 7.90522873e-01 5.03782332e-01
5.08455217e-01 -6.83284581e-01 -1.37449205e-01 7.79325426e-01
3.54237586e-01 -1.13304484e+00 2.50620265e-02 -3.76774631e-02
-2.96854794e-01 7.10448921e-01 1.05699325e+00 -1.45839766e-01
7.11750269e-01 4.60321605e-01 5.33579569e-03 -6.25485182e-01
-8.34279060e-01 -5.62355995e-01 2.34608814e-01 9.29549515e-01
2.31713071e-01 -2.19096094e-01 1.02679282e-01 -1.85772434e-01
-1.07757375e-01 9.58479419e-02 1.77409723e-01 9.73340392e-01
-4.92106885e-01 -1.06257439e+00 -4.15903568e-01 2.97444940e-01
-3.36836487e-01 1.77644208e-01 -4.86768782e-01 1.46808839e+00
1.32067770e-01 6.38718843e-01 -1.52946904e-01 -4.91737872e-01
3.34234118e-01 -1.26446262e-01 4.79400933e-01 -3.56306493e-01
-8.01177323e-01 -1.51687548e-01 1.50157344e-02 -8.04891169e-01
-7.33773828e-01 -8.68472815e-01 -9.09932435e-01 -8.51010323e-01
-4.36408311e-01 9.44061801e-02 2.79464185e-01 5.85874736e-01
4.78938788e-01 6.26348436e-01 -1.53954968e-01 -1.00963998e+00
-5.11947274e-01 -7.37764418e-01 -4.54013973e-01 4.69181724e-02
6.27383232e-01 -1.18226230e+00 -1.43594697e-01 -2.55333930e-01] | [6.337717056274414, -2.0606565475463867] |
dee20e28-4dde-4798-9428-be8c24639fe4 | a-multilingual-study-of-compressive-cross | 1810.10639 | null | http://arxiv.org/abs/1810.10639v1 | http://arxiv.org/pdf/1810.10639v1.pdf | A Multilingual Study of Compressive Cross-Language Text Summarization | Cross-Language Text Summarization (CLTS) generates summaries in a language
different from the language of the source documents. Recent methods use
information from both languages to generate summaries with the most informative
sentences. However, these methods have performance that can vary according to
languages, which can reduce the quality of summaries. In this paper, we propose
a compressive framework to generate cross-language summaries. In order to
analyze performance and especially stability, we tested our system and
extractive baselines on a dataset available in four languages (English, French,
Portuguese, and Spanish) to generate English and French summaries. An automatic
evaluation showed that our method outperformed extractive state-of-art CLTS
methods with better and more stable ROUGE scores for all languages. | ['Juan-Manuel Torres-Moreno', 'Stéphane Huet', 'Elvys Linhares Pontes'] | 2018-10-24 | null | null | null | null | ['cross-language-text-summarization'] | ['natural-language-processing'] | [-9.87160206e-03 -5.29008731e-02 -1.82640374e-01 -8.98300186e-02
-1.37012625e+00 -9.60646570e-01 1.00466645e+00 6.35736644e-01
-4.30039167e-01 1.27498329e+00 1.18119168e+00 -1.10955141e-01
2.11378694e-01 -6.08774424e-01 -3.21560472e-01 -7.40978941e-02
3.60446185e-01 4.26445901e-01 2.94105679e-01 -3.50363821e-01
7.76990056e-01 1.95869327e-01 -1.14937270e+00 8.42071950e-01
1.61044896e+00 2.14665197e-02 2.26568401e-01 1.00548589e+00
-6.17817700e-01 6.54190183e-01 -1.46040010e+00 -6.42473221e-01
-2.79263705e-01 -8.52917910e-01 -7.10530460e-01 -1.53925121e-01
4.34821039e-01 -2.73204725e-02 -7.53188357e-02 7.94833422e-01
6.92247450e-01 -3.00515443e-02 8.65880787e-01 -8.22483420e-01
-7.56991982e-01 1.21353352e+00 -4.32110935e-01 4.92112935e-02
1.00915956e+00 -8.19996819e-02 9.50784504e-01 -7.60239482e-01
8.89229417e-01 1.52030659e+00 4.81267720e-01 6.02555990e-01
-1.13438392e+00 -3.33610624e-01 5.41509688e-02 -2.74797648e-01
-1.05130994e+00 -7.28241205e-01 5.30153751e-01 -1.29272908e-01
1.20649242e+00 6.95026457e-01 5.73591709e-01 1.27840686e+00
5.27897656e-01 9.46504891e-01 9.21547949e-01 -7.22817242e-01
1.26594871e-01 3.06974381e-01 1.92688987e-01 2.44244382e-01
7.23153055e-01 -5.66968322e-01 -6.61799848e-01 -4.53635067e-01
-1.27370268e-01 -6.89220130e-01 -3.63854647e-01 3.44744593e-01
-1.56538582e+00 9.20378447e-01 -1.52696282e-01 6.80638909e-01
-4.39017862e-01 -2.98901260e-01 9.30837154e-01 9.91873220e-02
7.28223264e-01 8.61453235e-01 -1.78663567e-01 -2.18718722e-01
-1.32854116e+00 4.55618918e-01 1.30162835e+00 1.00449085e+00
2.54711866e-01 1.55336082e-01 -8.89870286e-01 8.52771580e-01
-3.28586698e-02 8.57435644e-01 8.39950860e-01 -7.62610197e-01
9.41513062e-01 5.89428067e-01 2.36462042e-01 -1.05354166e+00
-1.36236072e-01 -3.65021378e-01 -8.30388427e-01 -4.72958177e-01
-9.09931436e-02 -3.87885988e-01 -5.28940320e-01 1.31029963e+00
-2.83618867e-01 -6.51805520e-01 7.19636619e-01 2.09979534e-01
1.33156300e+00 1.20576799e+00 -1.85690448e-01 -6.20161176e-01
1.08810830e+00 -1.15541613e+00 -1.05874777e+00 -3.61421973e-01
7.29688168e-01 -1.19277263e+00 1.16782582e+00 1.68985099e-01
-1.29003346e+00 -4.91407186e-01 -1.08645999e+00 -1.53744578e-01
-3.22086900e-01 7.34324753e-01 6.27207384e-02 6.30101025e-01
-1.16522574e+00 5.26160479e-01 -5.18396199e-01 -6.23511970e-01
-1.40168712e-01 -2.37807274e-01 -3.88221860e-01 1.74394011e-01
-1.10210097e+00 9.20896709e-01 8.75255883e-01 -4.88692641e-01
-3.55979204e-01 -3.90932739e-01 -8.51779938e-01 6.51584268e-02
1.10438526e-01 -9.28666770e-01 1.35843885e+00 -4.84837025e-01
-1.43540347e+00 4.45563465e-01 -5.85164011e-01 -6.00096583e-01
6.62292063e-01 -3.97290796e-01 -3.91263396e-01 2.41436601e-01
6.33559167e-01 5.56499004e-01 3.47444147e-01 -1.17550230e+00
-7.54573762e-01 4.77081947e-02 -2.47218683e-01 2.87116200e-01
-2.83863127e-01 1.45975068e-01 -4.42584395e-01 -7.69095361e-01
-3.98920715e-01 -6.98755562e-01 -6.34318441e-02 -1.08204234e+00
-8.65068734e-01 -4.24153894e-01 5.80156267e-01 -1.00560725e+00
1.82291007e+00 -1.85162473e+00 2.80918200e-02 -3.33863795e-01
-1.82074517e-01 2.72675574e-01 -2.94866145e-01 1.13753867e+00
4.29853350e-01 5.65154552e-01 -2.05605507e-01 -5.03478169e-01
-8.08724575e-03 -2.36678347e-02 -5.65206409e-01 -1.99994162e-01
4.81397174e-02 8.54416490e-01 -1.02943575e+00 -8.95762205e-01
-1.18824385e-01 1.35256872e-01 -3.67689013e-01 9.29978266e-02
-2.61670977e-01 2.34149233e-01 -3.90564084e-01 1.75111115e-01
2.87861884e-01 2.81111479e-01 5.68616129e-02 -6.80092126e-02
-4.88251030e-01 7.08685398e-01 -7.35470593e-01 1.59855652e+00
-4.80472058e-01 7.75684357e-01 -5.39127171e-01 -2.33631089e-01
1.07717216e+00 3.54509830e-01 -3.83361545e-03 -3.98309022e-01
-1.82484582e-01 5.77640533e-01 -2.21467942e-01 -4.21202719e-01
1.34441018e+00 1.94387153e-01 -6.25985563e-01 6.22101307e-01
-8.44971314e-02 -6.94180429e-01 1.25967109e+00 7.50352025e-01
8.72935116e-01 -1.31031945e-01 7.02522755e-01 -2.24097207e-01
7.28283942e-01 2.97659606e-01 3.86186898e-01 9.71981227e-01
2.38185778e-01 6.72854662e-01 7.18425572e-01 1.76715359e-01
-1.00353765e+00 -9.25006270e-01 2.92499393e-01 6.98205113e-01
-3.14140916e-01 -1.13050711e+00 -9.42707837e-01 -8.09366107e-01
-2.59943306e-01 1.53554571e+00 -2.63252378e-01 -1.60895497e-01
-6.63875103e-01 -5.01051843e-01 9.22736943e-01 1.34429410e-01
5.05176842e-01 -1.27176285e+00 -4.97414351e-01 2.33647972e-01
-8.44187260e-01 -1.18618035e+00 -8.58160913e-01 -4.62099582e-01
-7.92850614e-01 -7.16672301e-01 -8.71888340e-01 -5.86415350e-01
3.29401821e-01 9.33151245e-02 1.36090374e+00 -3.87622327e-01
4.04907525e-01 3.34444284e-01 -5.32049298e-01 -6.81815445e-01
-1.22268462e+00 7.90283799e-01 -1.87128577e-02 -4.76924688e-01
5.28653152e-02 -8.84325653e-02 3.35130957e-03 -3.48399103e-01
-1.11901391e+00 1.32458434e-01 5.46235085e-01 4.85517293e-01
2.29292139e-01 -1.74759105e-01 9.60265696e-01 -9.80477870e-01
1.79174042e+00 -2.02671736e-01 -1.28235623e-01 6.31907403e-01
-3.90813738e-01 3.63840491e-01 8.86545062e-01 -9.28068906e-02
-1.30958855e+00 -4.68989432e-01 -8.50872993e-02 4.43069905e-01
2.18221247e-02 7.76255071e-01 -4.41306047e-02 7.43086755e-01
7.21923113e-01 5.31546831e-01 -4.38313723e-01 -5.00498295e-01
3.10866654e-01 9.76991534e-01 5.68481624e-01 -3.78430456e-01
4.15368557e-01 -3.40831988e-02 -5.38635910e-01 -1.06161606e+00
-8.14194322e-01 -2.16974914e-01 -6.34983778e-01 -9.10496712e-02
6.46109402e-01 -9.12096202e-01 1.82044953e-01 3.05654079e-01
-1.59148431e+00 7.04051629e-02 -4.78266954e-01 4.57189262e-01
-2.98423678e-01 7.13109076e-01 -4.79121208e-01 -5.62202334e-01
-1.00075471e+00 -9.75380361e-01 1.22909188e+00 3.37846518e-01
-8.39603662e-01 -1.11000121e+00 3.48647475e-01 9.09504220e-02
3.90633672e-01 3.04293543e-01 8.29136074e-01 -7.97297299e-01
1.72681451e-01 -2.81125009e-01 -7.13165477e-03 2.30329871e-01
4.30396020e-01 5.23033023e-01 -3.81721914e-01 -3.06181788e-01
-1.73306569e-01 -1.25779107e-01 9.57562387e-01 3.51211846e-01
3.67215574e-01 -7.71299541e-01 -3.44918340e-01 2.81717405e-02
1.07192516e+00 2.82108150e-02 4.59365845e-01 3.87160510e-01
3.88564169e-01 6.08668566e-01 6.88433468e-01 3.83844614e-01
5.37921548e-01 2.61702836e-01 -4.42589611e-01 2.38971561e-01
-3.62463146e-01 -6.17694974e-01 8.83986652e-01 1.51135969e+00
4.13619995e-01 -8.33387733e-01 -6.46422565e-01 7.80663550e-01
-1.72359908e+00 -9.90312874e-01 -1.92643002e-01 1.99305701e+00
1.04760516e+00 2.23491773e-01 1.87446713e-01 -6.01695962e-02
6.50994778e-01 4.93489206e-01 -1.61149129e-01 -8.70314360e-01
-5.80330491e-01 -1.44632548e-01 1.04366623e-01 5.69280028e-01
-6.06045544e-01 1.02802742e+00 6.64340067e+00 8.11792135e-01
-1.07989800e+00 -2.00063348e-01 2.41167769e-01 -1.62105516e-01
-7.98365176e-01 -3.68914120e-02 -8.87737691e-01 5.31766295e-01
1.25768471e+00 -1.14691508e+00 -2.21762002e-01 4.97615308e-01
4.15243238e-01 -3.48194361e-01 -8.91365349e-01 6.35580778e-01
5.35211325e-01 -1.51213109e+00 7.05953181e-01 -3.00878912e-01
1.03275526e+00 3.73354703e-02 -4.67521369e-01 4.18377191e-01
3.97527069e-01 -6.22797310e-01 8.71112704e-01 5.60354650e-01
7.61029720e-01 -8.87666702e-01 8.91726434e-01 6.17648602e-01
-9.33005989e-01 4.16226149e-01 -4.38643366e-01 1.96357399e-01
3.92436117e-01 6.20649517e-01 -9.37020481e-01 1.00560570e+00
3.95504951e-01 8.85691822e-01 -1.05593526e+00 8.36431265e-01
-4.53429133e-01 5.38175762e-01 -4.20811698e-02 -5.16110063e-01
3.06011975e-01 -1.99211165e-01 1.07336164e+00 1.87752533e+00
6.67999923e-01 -3.12025517e-01 3.16025585e-01 4.89102572e-01
-1.43548355e-01 5.97301960e-01 -8.90444994e-01 -2.90689617e-01
5.34722805e-01 9.99509513e-01 -6.78049564e-01 -8.53392780e-01
1.54052302e-02 9.69852626e-01 7.88726807e-02 2.93845236e-01
-4.94504839e-01 -7.66306043e-01 -9.33108479e-02 -9.94214714e-02
-4.26823907e-02 -3.01398546e-01 -3.07692587e-01 -1.42766857e+00
1.77434176e-01 -1.06227529e+00 3.74716073e-01 -7.32709110e-01
-1.04058826e+00 1.07154942e+00 3.64409149e-01 -1.18545163e+00
-6.78081214e-01 2.03739867e-01 -7.01514184e-01 7.98308253e-01
-1.17433476e+00 -8.68169188e-01 3.92258465e-02 -3.30441482e-02
9.80771542e-01 -4.28146929e-01 7.84251034e-01 -1.36501761e-02
-5.16778171e-01 4.57214355e-01 3.61410886e-01 -5.57110012e-02
1.10264802e+00 -1.26046252e+00 7.82235086e-01 1.22344589e+00
9.28019732e-02 5.86441517e-01 1.00430357e+00 -9.92329419e-01
-9.22852337e-01 -1.18108249e+00 1.68739200e+00 -3.45779508e-01
4.74495530e-01 -3.32869776e-02 -7.60637283e-01 5.05442798e-01
1.09716594e+00 -9.91378367e-01 6.03507876e-01 -1.65718853e-01
-1.93126947e-02 8.98967683e-02 -9.26955402e-01 9.83844638e-01
6.40862644e-01 -2.01754376e-01 -9.91250873e-01 2.83643156e-01
9.43731785e-01 -2.98305571e-01 -5.36962092e-01 1.73202157e-01
3.37984443e-01 -1.00042593e+00 4.52253550e-01 -3.10480982e-01
6.85681522e-01 -3.10891747e-01 3.69780138e-02 -1.97888792e+00
2.55419384e-03 -6.83538914e-01 5.23896553e-02 1.81307757e+00
6.97437346e-01 -7.99926221e-01 7.53926560e-02 1.41073406e-01
-2.71529615e-01 -2.91652292e-01 -5.75661838e-01 -6.86643124e-01
3.40194255e-01 -3.27400789e-02 6.71281993e-01 3.79713714e-01
1.77494824e-01 1.05901432e+00 -1.79840997e-01 -4.32861269e-01
3.19141507e-01 3.01058173e-01 9.09820139e-01 -1.00846756e+00
3.21593434e-01 -6.82094753e-01 2.94700354e-01 -8.12546015e-01
4.04528052e-01 -9.54207778e-01 -9.55995754e-04 -2.24327326e+00
3.89454335e-01 1.29104167e-01 4.25600052e-01 2.24947304e-01
-4.97581273e-01 -5.74792251e-02 3.94544303e-01 1.67295352e-01
-7.26368785e-01 7.70058095e-01 1.01817322e+00 -3.03594679e-01
-5.77302814e-01 -4.11533043e-02 -9.48851407e-01 4.99005526e-01
1.12267768e+00 -6.04488850e-01 -3.11488569e-01 -5.48201323e-01
1.31871492e-01 8.90801549e-02 -4.32072997e-01 -1.00883079e+00
1.92171380e-01 -1.79893877e-02 9.21110734e-02 -1.14809525e+00
-1.46517232e-01 9.24618728e-03 -1.66754972e-03 5.34394145e-01
-6.49455547e-01 5.68730354e-01 4.92502004e-01 1.13314025e-01
-4.86211121e-01 -5.66716015e-01 4.21056747e-01 -1.65689752e-01
-1.75222699e-02 -4.13039744e-01 -7.04330564e-01 4.30016816e-01
6.45614803e-01 -6.19251467e-02 -5.96316040e-01 -6.11204863e-01
6.73666596e-02 2.24915057e-01 7.11431623e-01 5.94035387e-01
6.10944092e-01 -1.18159318e+00 -1.50321507e+00 -2.44775549e-01
1.00507826e-01 -3.52325499e-01 -4.99447510e-02 5.84482968e-01
-7.36153662e-01 9.15382683e-01 -1.21969149e-01 -3.85734409e-01
-1.32280636e+00 2.35986263e-01 -1.27587348e-01 -8.95037174e-01
-4.69985008e-01 1.19362578e-01 -2.28761196e-01 -6.16547346e-01
-1.42070249e-01 -3.71267796e-01 -5.93140662e-01 4.90807384e-01
7.11505115e-01 6.95777416e-01 2.23429859e-01 -7.19515026e-01
-1.88420877e-01 3.05562109e-01 -1.81257173e-01 -4.96596634e-01
9.78604376e-01 -2.56468713e-01 -4.03010309e-01 7.01804101e-01
9.07529116e-01 9.20855761e-01 -3.20033878e-01 8.76162387e-03
4.43343729e-01 -1.90336779e-01 -4.02222931e-01 -8.06422174e-01
-3.95286411e-01 4.36230421e-01 -3.57108980e-01 4.34541911e-01
9.80358303e-01 -2.24707320e-01 9.58503604e-01 4.06331807e-01
1.16383731e-01 -1.15188336e+00 -9.83769372e-02 8.76095235e-01
1.31438851e+00 -1.00990677e+00 3.30952793e-01 -9.30392966e-02
-1.01749098e+00 1.23700345e+00 3.41387391e-01 1.20495521e-01
6.58513512e-03 3.27062272e-02 1.58218965e-01 2.05919817e-01
-9.86251235e-01 1.84421390e-01 5.83254457e-01 3.40213865e-01
9.63106453e-01 6.76358715e-02 -9.72200394e-01 4.66473132e-01
-7.70779669e-01 -2.51343876e-01 1.07680488e+00 7.52817929e-01
-3.86423618e-01 -1.18136215e+00 -5.16054153e-01 5.56808174e-01
-6.54882193e-01 -1.29122078e-01 -1.03768718e+00 8.28065455e-01
-5.13051093e-01 1.52634776e+00 -2.10209951e-01 2.17525773e-02
6.10537887e-01 9.08447336e-03 2.29258582e-01 -8.67519259e-01
-9.00348186e-01 2.43786961e-01 5.82266927e-01 1.06285401e-02
-4.21120197e-01 -8.89008582e-01 -1.17817926e+00 -3.71548176e-01
-1.88208312e-01 8.61980259e-01 3.82670850e-01 6.79806530e-01
4.40660566e-01 6.79181635e-01 5.22286773e-01 -7.09370792e-01
-3.89635861e-01 -1.45406818e+00 -1.83241814e-01 1.13902017e-01
2.63985753e-01 2.17280462e-01 -3.48221689e-01 3.12386621e-02] | [12.39795207977295, 9.518388748168945] |
540a5859-90f3-4606-a3e3-bb455dcbdea0 | fine-grained-vr-sketching-dataset-and | 2209.10008 | null | https://arxiv.org/abs/2209.10008v1 | https://arxiv.org/pdf/2209.10008v1.pdf | Fine-Grained VR Sketching: Dataset and Insights | We present the first fine-grained dataset of 1,497 3D VR sketch and 3D shape pairs of a chair category with large shapes diversity. Our dataset supports the recent trend in the sketch community on fine-grained data analysis, and extends it to an actively developing 3D domain. We argue for the most convenient sketching scenario where the sketch consists of sparse lines and does not require any sketching skills, prior training or time-consuming accurate drawing. We then, for the first time, study the scenario of fine-grained 3D VR sketch to 3D shape retrieval, as a novel VR sketching application and a proving ground to drive out generic insights to inform future research. By experimenting with carefully selected combinations of design factors on this new problem, we draw important conclusions to help follow-on work. We hope our dataset will enable other novel applications, especially those that require a fine-grained angle such as fine-grained 3D shape reconstruction. The dataset is available at tinyurl.com/VRSketch3DV21. | ['Yi-Zhe Song', 'Tao Xiang', 'Yongxin Yang', 'Yulia Gryaditskaya', 'Ling Luo'] | 2022-09-20 | null | null | null | null | ['3d-shape-reconstruction'] | ['computer-vision'] | [ 5.86910024e-02 -2.05666393e-01 -1.27074137e-01 -2.09576711e-01
-5.52355707e-01 -1.19043946e+00 8.61902535e-01 -9.41680223e-02
4.04166788e-01 3.42439383e-01 3.05516154e-01 -6.26633525e-01
-5.30229986e-01 -1.00823069e+00 -4.39249337e-01 -1.25127643e-01
-4.57861163e-02 8.78679335e-01 1.18734002e-01 -4.54682440e-01
7.11059272e-01 1.33472073e+00 -1.60777187e+00 2.10028589e-01
3.99939954e-01 9.09892440e-01 -4.33619246e-02 7.11603403e-01
-4.14155334e-01 1.33020774e-01 -5.30058324e-01 -4.33387876e-01
6.10496044e-01 2.20729843e-01 -4.01939362e-01 -4.12283614e-02
1.06318760e+00 -4.40051019e-01 -2.27044106e-01 4.55894709e-01
6.77292705e-01 6.89742938e-02 9.58650410e-01 -1.15131664e+00
-7.48680472e-01 6.15533814e-02 -5.16238809e-01 -1.66448340e-01
7.49827147e-01 -9.95483994e-03 1.14515305e+00 -1.23396444e+00
1.15801156e+00 1.35883152e+00 8.37934077e-01 4.53607231e-01
-1.15491807e+00 -6.80897176e-01 7.16257468e-02 -3.78106564e-01
-1.51980352e+00 -3.45135480e-01 1.17158818e+00 -4.37360466e-01
7.80941188e-01 4.40405577e-01 7.77879953e-01 1.13989735e+00
-2.60588318e-01 6.51909530e-01 1.01860666e+00 -3.35437864e-01
3.12842935e-01 -3.83953899e-02 -7.61857927e-02 7.25267708e-01
4.10996795e-01 -3.25550176e-02 -3.16849589e-01 -4.76162344e-01
1.67019951e+00 2.44005233e-01 1.30843937e-01 -9.87113416e-01
-1.12104559e+00 6.16022825e-01 2.09906727e-01 4.29418743e-01
-1.82565674e-01 1.70398980e-01 2.19903991e-01 6.18623078e-01
3.07997555e-01 6.77339792e-01 -3.98177445e-01 -3.51129621e-01
-1.15013766e+00 7.73568332e-01 1.00615251e+00 1.45920038e+00
5.66584766e-01 -7.12669939e-02 9.22465250e-02 9.67511475e-01
4.59208637e-02 6.27868652e-01 -3.31701458e-01 -1.18712544e+00
4.53800321e-01 6.09596550e-01 1.58334181e-01 -1.09281445e+00
-2.35300720e-01 -6.27657324e-02 -7.69667208e-01 4.23890769e-01
4.06837076e-01 2.91891634e-01 -6.49440825e-01 9.62715745e-01
1.64220244e-01 -2.73074985e-01 -6.46552503e-01 9.40345287e-01
1.16775155e+00 1.45245105e-01 -3.00713569e-01 4.20167804e-01
1.27183330e+00 -4.65600967e-01 -1.53661802e-01 4.09653515e-01
2.99655795e-01 -1.07093656e+00 1.52900481e+00 5.00652015e-01
-1.30069649e+00 -4.06586856e-01 -1.15236819e+00 -3.83607060e-01
-4.61921036e-01 -5.59765995e-02 1.13851249e+00 7.58843124e-01
-1.03492153e+00 8.85036528e-01 -1.80275247e-01 -7.37884581e-01
6.76711380e-01 9.02837291e-02 -4.81767952e-01 -2.70173639e-01
-5.60846269e-01 8.19875777e-01 -4.90291715e-01 -2.52166092e-01
-3.70522439e-01 -1.03139925e+00 -3.82487744e-01 -3.51476103e-01
4.14011568e-01 -1.02258873e+00 1.05601323e+00 2.38991827e-02
-1.40171754e+00 1.10284853e+00 5.08239269e-02 3.19969207e-01
7.27889955e-01 -6.75040409e-02 -9.80246887e-02 8.08307379e-02
-7.66987056e-02 4.62041646e-01 8.66448402e-01 -1.58910704e+00
1.35267228e-01 -5.17001927e-01 2.71560073e-01 -1.15652308e-01
2.59409308e-01 -3.69500786e-01 -3.38865310e-01 -9.41084564e-01
3.08334738e-01 -8.34288001e-01 -4.29676324e-02 5.77068090e-01
-2.02442080e-01 -3.08130771e-01 8.41521084e-01 -1.14610061e-01
9.13688600e-01 -1.87249398e+00 1.52270555e-01 5.77902138e-01
4.02986646e-01 5.09322360e-02 -3.03974181e-01 1.05276537e+00
1.38031662e-01 5.56437671e-01 3.53378169e-02 -2.95149148e-01
4.04399514e-01 -4.24702652e-02 -5.74090898e-01 2.84146637e-01
2.24065527e-01 1.08486807e+00 -7.06188738e-01 -4.17409658e-01
4.61873800e-01 4.52203065e-01 -5.22793710e-01 1.21374145e-01
-1.72918975e-01 1.68259621e-01 -8.37923944e-01 1.03674877e+00
8.93548965e-01 -6.02679066e-02 -3.94071378e-02 -4.42584723e-01
-2.00165376e-01 -7.72084296e-02 -1.35795641e+00 2.06995893e+00
-5.40163875e-01 3.94350022e-01 1.85321823e-01 -4.97391939e-01
1.52417266e+00 -2.53455807e-02 4.75136191e-01 -8.22035372e-01
-4.15810980e-02 4.30005491e-01 -6.37044907e-01 -1.63797006e-01
8.44238818e-01 -3.57231587e-01 -3.40258867e-01 7.18911350e-01
-2.01176122e-01 -1.25351334e+00 -2.08951250e-01 3.78754675e-01
9.84994113e-01 4.08372402e-01 2.99305528e-01 -4.50737566e-01
4.31648642e-02 -6.88167065e-02 -3.10534328e-01 8.60774577e-01
1.65275350e-01 8.90205979e-01 2.60827959e-01 -6.67813540e-01
-1.65862513e+00 -1.20325828e+00 -3.54496658e-01 7.07429111e-01
8.90032277e-02 -6.74186587e-01 -9.21035483e-02 -2.80617863e-01
6.24920905e-01 3.15040112e-01 -5.70340693e-01 4.76916999e-01
-5.56726873e-01 2.30439246e-01 5.64338326e-01 5.71108758e-01
-1.35452184e-03 -8.50296140e-01 -2.19549552e-01 -1.46710575e-01
5.80504775e-01 -7.06285179e-01 -3.83403033e-01 -9.69180614e-02
-1.13004684e+00 -8.84411335e-01 -1.07866287e+00 -5.47820926e-01
5.25207043e-01 5.93135774e-01 1.46940064e+00 5.37837207e-01
-3.71010482e-01 8.19928885e-01 -4.08326685e-01 -2.39563406e-01
-4.26730849e-02 2.04387262e-01 1.28161442e-02 -6.71421945e-01
8.07682872e-02 -1.10797727e+00 -5.67841113e-01 5.12503088e-01
-7.33812153e-01 6.30160794e-02 4.00979102e-01 4.87106591e-01
6.14030421e-01 -2.43978098e-01 5.38944602e-01 -8.88946950e-01
9.77753639e-01 -2.46503174e-01 -4.13655818e-01 1.84094638e-01
-2.51951158e-01 1.55813247e-01 4.40841675e-01 -3.99224430e-01
-6.49854422e-01 -2.26092711e-01 -2.79980659e-01 -7.12831259e-01
-2.51616567e-01 -3.44227254e-02 -3.58140729e-02 -3.53156388e-01
6.88424289e-01 -6.33469746e-02 -2.90710717e-01 -8.26022148e-01
8.01062405e-01 5.09362161e-01 2.70861953e-01 -1.16619480e+00
1.13774943e+00 5.70409775e-01 3.30545515e-01 -1.11375082e+00
-3.87829654e-02 -1.75151080e-01 -9.07195628e-01 -3.35974284e-02
3.68385971e-01 -5.52334368e-01 -8.60092223e-01 1.65654942e-02
-1.07055008e+00 -4.67348188e-01 -5.71367800e-01 -1.41272485e-01
-8.87077153e-01 3.47240150e-01 -2.77687639e-01 -9.22805667e-01
-2.90589660e-01 -7.46342063e-01 1.64848053e+00 -6.06505759e-02
-5.97576320e-01 -8.27732742e-01 9.43717286e-02 1.70966476e-01
5.69857657e-01 2.83118427e-01 1.04110622e+00 -7.07380027e-02
-9.65849578e-01 -2.57601023e-01 -4.73589778e-01 -4.07277942e-01
2.72936542e-02 3.25840980e-01 -8.78397286e-01 -1.99292466e-01
-6.36404872e-01 -4.32913244e-01 5.54125845e-01 1.34639004e-02
1.23831439e+00 9.76758823e-02 -3.01544130e-01 5.67285597e-01
1.38212967e+00 -5.61004132e-02 5.68460941e-01 -1.34491185e-02
8.84584486e-01 5.33286929e-01 5.05682051e-01 6.09907031e-01
5.46069086e-01 8.43909979e-01 2.08860308e-01 -7.96385109e-02
-2.81809777e-01 -7.27195561e-01 -5.12831569e-01 7.27334499e-01
-6.37897789e-01 -9.04653687e-03 -7.26472676e-01 3.12999696e-01
-1.43929923e+00 -1.05972648e+00 1.22908242e-01 2.22792697e+00
2.58235157e-01 -2.24938672e-02 4.32708412e-01 2.99367100e-01
2.74658471e-01 4.84265596e-01 -4.50739413e-01 -6.67654514e-01
-6.04001880e-02 7.33005583e-01 3.08236122e-01 3.41619819e-01
-6.82825327e-01 9.27845418e-01 6.29327154e+00 8.23217928e-01
-9.21349764e-01 -4.50197518e-01 1.10203989e-01 -1.64079428e-01
-1.18918502e+00 9.35679078e-02 -4.54345703e-01 -2.44669281e-02
1.50866941e-01 3.51311043e-02 7.68403053e-01 6.89646065e-01
2.53857020e-02 1.52248651e-01 -1.21765864e+00 1.39076746e+00
-2.31860191e-01 -1.81528282e+00 2.92904198e-01 1.60194486e-01
5.03470302e-01 -2.68011063e-01 -5.46591207e-02 1.41442373e-01
2.55037725e-01 -1.29287243e+00 9.15586710e-01 8.14370215e-01
1.36326587e+00 -5.77813804e-01 -4.40093502e-03 6.60552904e-02
-1.57708335e+00 3.68139178e-01 -3.28656346e-01 -9.29556563e-02
3.82933505e-02 4.04634506e-01 -6.51723087e-01 6.14085913e-01
4.86684293e-01 8.69972706e-01 -4.19809759e-01 9.13010836e-01
3.45943958e-01 -2.38850228e-02 -4.73269224e-01 -4.16995406e-01
-9.99419019e-02 -2.39861682e-01 5.42645156e-01 1.08119905e+00
4.03714925e-01 4.20805067e-01 8.26441537e-05 9.95798767e-01
-5.66369630e-02 4.65348028e-02 -1.22562730e+00 -3.40116054e-01
1.03700364e+00 1.19129658e+00 -8.92878830e-01 -4.63877842e-02
-2.41809517e-01 9.87958193e-01 2.94491619e-01 2.93683618e-01
-2.09491491e-01 -4.44038361e-01 5.77104151e-01 5.43278635e-01
5.84101021e-01 -8.78093958e-01 -7.92766094e-01 -1.17145371e+00
1.65302485e-01 -6.81209087e-01 -3.28229293e-02 -1.07261014e+00
-1.69498360e+00 3.35872561e-01 -4.04149890e-02 -1.28003252e+00
-1.20079719e-01 -6.44949973e-01 -6.52474761e-01 8.75324428e-01
-1.35888612e+00 -1.33883309e+00 -3.81541133e-01 5.12325644e-01
4.78290856e-01 -8.08434784e-02 1.07964766e+00 3.05162966e-01
1.64367303e-01 6.43568814e-01 -1.68524757e-01 -2.72719473e-01
5.94607532e-01 -9.75097358e-01 9.43412125e-01 -1.13595583e-01
2.89758265e-01 8.88566196e-01 5.51092267e-01 -8.14886689e-01
-2.26635647e+00 -4.87537771e-01 5.91682971e-01 -9.14594829e-01
5.62512696e-01 -7.94525445e-01 -5.72715998e-01 3.70241076e-01
-3.67666364e-01 9.57873240e-02 3.67244750e-01 4.39627022e-01
-6.42075956e-01 2.48916969e-01 -1.47991943e+00 6.50860608e-01
1.88651526e+00 -9.15776193e-01 -6.41035497e-01 -8.61015171e-02
4.73286182e-01 -5.11913538e-01 -1.51971376e+00 2.19891906e-01
1.35120392e+00 -8.82232726e-01 1.46044660e+00 -5.64168096e-01
3.93862814e-01 -1.71330422e-01 -6.45659685e-01 -1.03827047e+00
-4.88630831e-01 -6.09423935e-01 -2.06693355e-02 1.02538800e+00
1.41022176e-01 -2.12692544e-01 1.07671928e+00 5.90894163e-01
-3.28604355e-02 -1.01636302e+00 -5.79081237e-01 -7.42089689e-01
3.02259535e-01 -6.67349339e-01 1.14599586e+00 1.00286078e+00
-3.37931275e-01 8.53551086e-03 -1.21881887e-01 -2.66737580e-01
6.15974128e-01 5.95549822e-01 1.33290327e+00 -1.79911244e+00
8.31314996e-02 -7.83643067e-01 -4.28035170e-01 -1.19735503e+00
-1.91786885e-01 -8.08341682e-01 -6.73507333e-01 -1.65803742e+00
-2.02213809e-01 -1.19475555e+00 3.88792962e-01 2.00608313e-01
4.23703998e-01 4.85830843e-01 5.00395834e-01 1.34662837e-01
-4.09547925e-01 4.44189161e-01 1.75860739e+00 -5.93357496e-02
-1.76583156e-01 1.50148617e-03 -9.05721188e-01 3.41600597e-01
4.80214983e-01 -2.90809870e-02 -6.18212879e-01 -3.00463915e-01
3.51096302e-01 2.89409608e-01 4.95585531e-01 -5.01985550e-01
-2.28946265e-02 -3.14019769e-01 4.70484346e-01 -9.55821693e-01
6.83126450e-01 -9.93786097e-01 3.89634222e-01 -1.93490997e-01
-7.91278854e-02 1.31223530e-01 2.62031823e-01 1.77630588e-01
3.91595095e-01 1.67301521e-02 2.92301774e-01 -5.30086517e-01
-6.51128888e-01 6.27722144e-01 9.29555148e-02 1.00739412e-01
7.82860041e-01 -8.24936748e-01 -3.14713925e-01 -4.31448221e-01
-5.89093506e-01 -1.71356142e-01 1.13526976e+00 6.47421479e-01
7.73723423e-01 -1.83625436e+00 -5.98687410e-01 5.08242607e-01
2.28226617e-01 -5.70618175e-02 2.86248237e-01 1.37636408e-01
-4.92762983e-01 5.24664938e-01 -4.63953793e-01 -5.52040279e-01
-1.01606441e+00 4.80762839e-01 -3.38612795e-02 1.92405894e-01
-9.71122026e-01 6.90154314e-01 -2.42135212e-01 -7.32406557e-01
6.57107830e-02 -5.54594040e-01 2.44739339e-01 1.52846768e-01
3.00254375e-01 5.46305537e-01 2.62486964e-01 -2.79663473e-01
-4.70710397e-01 1.31432712e+00 1.90618888e-01 -6.29628003e-02
1.74164295e+00 9.94483307e-02 2.45589375e-01 5.22738814e-01
1.12404931e+00 5.61476350e-01 -1.21680164e+00 5.83090745e-02
-2.28822097e-01 -9.09735143e-01 -1.97955340e-01 -6.79850340e-01
-7.78773487e-01 8.21166754e-01 1.13418266e-01 3.66397381e-01
6.16951585e-01 4.09696460e-01 5.68291664e-01 4.00267512e-01
9.87953365e-01 -6.23741925e-01 5.81464320e-02 2.85279602e-01
1.44388199e+00 -9.09410894e-01 4.97440517e-01 -3.60619545e-01
-4.79115456e-01 1.28215241e+00 1.62260711e-01 -6.69273019e-01
6.61268771e-01 5.51532507e-01 -2.88613528e-01 -6.78639889e-01
-5.22231698e-01 1.18672952e-01 5.14678359e-01 8.12073052e-01
5.58194697e-01 2.95491785e-01 4.92442325e-02 4.36124265e-01
-5.20006359e-01 2.16950506e-01 3.03870708e-01 9.59369421e-01
-1.94152847e-01 -1.56789923e+00 -2.63536632e-01 7.01263428e-01
3.56769383e-01 5.44244163e-02 -7.40321755e-01 9.50382233e-01
-2.16598213e-01 3.69856954e-01 1.07372664e-01 -3.32832515e-01
9.49040830e-01 -1.33332685e-01 1.25655782e+00 -4.41164166e-01
-3.28067392e-01 -4.48542774e-01 3.12206626e-01 -6.57306492e-01
-1.34113088e-01 -7.31870353e-01 -8.04721296e-01 -9.62631166e-01
-6.21676538e-03 -2.63525516e-01 8.03657055e-01 3.51370096e-01
6.25036120e-01 -2.68891603e-02 4.81710851e-01 -1.61760807e+00
-1.51206791e-01 -5.19308269e-01 -8.31274450e-01 3.74570251e-01
2.89375305e-01 -1.10709858e+00 -7.19964728e-02 -6.00335777e-01] | [8.666247367858887, -3.724386692047119] |
be65c067-71cb-4388-beda-0aff905632fe | computer-assisted-analysis-of-biomedical | 2106.04381 | null | https://arxiv.org/abs/2106.04381v1 | https://arxiv.org/pdf/2106.04381v1.pdf | Computer-Assisted Analysis of Biomedical Images | Nowadays, the amount of heterogeneous biomedical data is increasing more and more thanks to novel sensing techniques and high-throughput technologies. In reference to biomedical image analysis, the advances in image acquisition modalities and high-throughput imaging experiments are creating new challenges. This huge information ensemble could overwhelm the analytic capabilities needed by physicians in their daily decision-making tasks as well as by biologists investigating complex biochemical systems. In particular, quantitative imaging methods convey scientifically and clinically relevant information in prediction, prognosis or treatment response assessment, by also considering radiomics approaches. Therefore, the computational analysis of medical and biological images plays a key role in radiology and laboratory applications. In this regard, frameworks based on advanced Machine Learning and Computational Intelligence can significantly improve traditional Image Processing and Pattern Recognition approaches. However, conventional Artificial Intelligence techniques must be tailored to address the unique challenges concerning biomedical imaging data. This thesis aims at proposing novel and advanced computer-assisted methods for biomedical image analysis, also as an instrument in the development of Clinical Decision Support Systems, by always keeping in mind the clinical feasibility of the developed solutions. In conclusion, the ultimate goal of these research studies is to gain clinically and biologically useful insights that can guide differential diagnosis and therapies, leading towards biomedical data integration for personalized medicine. As a matter of fact, the proposed computer-assisted bioimage analysis methods can be beneficial for the definition of imaging biomarkers, as well as for quantitative medicine and biology. | ['Leonardo Rundo'] | 2021-06-04 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [ 6.22390449e-01 -2.44928673e-01 -2.54311919e-01 -6.60955161e-02
-2.99086332e-01 -2.15320945e-01 1.87973365e-01 8.28117192e-01
-4.53749269e-01 7.85441101e-01 -1.57987624e-01 -1.64299697e-01
-4.55492735e-01 -7.13890076e-01 6.10868409e-02 -1.18945122e+00
-9.85686630e-02 7.47677445e-01 -5.33431657e-02 -7.73911625e-02
2.19038770e-01 9.83237386e-01 -1.54028499e+00 1.21753655e-01
1.00923741e+00 1.17430103e+00 3.32238078e-01 5.06528914e-01
-1.79726183e-01 5.55003464e-01 -2.36845419e-01 -3.08438577e-02
-2.57175207e-01 -6.35074794e-01 -2.29627639e-01 -5.85716590e-02
-4.66760993e-01 1.17222726e-01 2.87739068e-01 9.11478043e-01
7.26682723e-01 2.75049973e-02 6.93701386e-01 -4.65118945e-01
-3.59508038e-01 2.67744154e-01 -3.53751868e-01 2.10071355e-01
3.06627601e-01 2.16313586e-01 1.76304668e-01 -5.35698593e-01
5.38560748e-01 7.28695095e-01 2.49842912e-01 5.31715214e-01
-1.04346585e+00 -3.12836468e-01 -2.82631427e-01 5.04546046e-01
-9.62433159e-01 -4.01314020e-01 7.56442249e-01 -8.05572748e-01
3.52193952e-01 6.18465841e-01 6.75436854e-01 7.70971656e-01
5.48700690e-01 2.46008858e-01 1.35035288e+00 -4.31109548e-01
4.09650624e-01 3.07072669e-01 1.25510186e-01 5.61143219e-01
4.88707542e-01 -3.16018984e-02 -2.06734613e-01 5.86432917e-03
3.92399043e-01 3.19087148e-01 -2.79470831e-01 6.01303168e-02
-1.25647080e+00 4.60487187e-01 1.78278491e-01 1.13919306e+00
-6.99041724e-01 -2.53859788e-01 3.14857036e-01 -2.43280888e-01
3.36474389e-01 4.53112304e-01 -2.28496060e-01 7.59329945e-02
-8.37239563e-01 -1.19337611e-01 5.03872335e-01 -2.06032712e-02
5.70191205e-01 -9.73375456e-04 -1.45003423e-01 6.00489199e-01
2.92800218e-01 7.84948170e-01 8.89380813e-01 -7.78290212e-01
-2.15333313e-01 8.92159641e-01 -1.36030987e-01 -1.41126978e+00
-7.59535849e-01 -6.64106667e-01 -1.17721713e+00 8.39312598e-02
3.94480109e-01 2.11799905e-01 -4.51014519e-01 1.11809146e+00
5.84492803e-01 -9.15417820e-02 -5.32905571e-02 1.03860295e+00
7.11692929e-01 3.40104848e-01 4.79331791e-01 -5.60240984e-01
1.74019241e+00 -8.79061371e-02 -7.61827469e-01 1.72515675e-01
6.33606493e-01 -5.72967112e-01 6.61065578e-01 4.95891690e-01
-6.22488141e-01 -3.81369710e-01 -8.67408633e-01 2.32772887e-01
-5.03862262e-01 3.18197668e-01 6.08071148e-01 6.79588854e-01
-4.59035248e-01 5.39978683e-01 -9.48081434e-01 -4.69120383e-01
4.12331313e-01 3.53318632e-01 -3.13585430e-01 -2.22159505e-01
-1.07158220e+00 1.01728821e+00 2.23043844e-01 4.31861997e-01
-3.51678461e-01 -7.30108619e-01 -5.43947041e-01 -5.88716269e-02
5.15382644e-03 -8.33745599e-01 4.87489820e-01 -5.64435899e-01
-1.49100161e+00 9.60673213e-01 -1.76589966e-01 -5.38502216e-01
4.23556358e-01 2.38469660e-01 -4.25709128e-01 5.38294554e-01
-2.16495842e-02 2.06126437e-01 2.70018131e-01 -9.29620683e-01
-1.42077953e-01 -9.95325983e-01 -5.43370843e-01 -8.62013623e-02
-4.45617348e-01 -1.08201303e-01 8.86662751e-02 -3.35641325e-01
3.07301044e-01 -6.72480941e-01 -5.93882322e-01 -4.32431214e-02
-1.26774281e-01 2.87200809e-01 6.74086750e-01 -6.51901066e-01
9.36869085e-01 -1.90206873e+00 2.63824075e-01 3.73042643e-01
2.50497133e-01 4.43019331e-01 3.54547650e-01 2.56998569e-01
1.84963867e-01 8.11343715e-02 -2.54263412e-02 1.47235826e-01
-4.18783754e-01 -1.07851081e-01 2.27858156e-01 6.29519939e-01
1.01729207e-01 7.13996351e-01 -7.80616939e-01 -8.03485930e-01
7.82540202e-01 6.77588701e-01 -6.66883886e-02 -8.47872049e-02
-2.56979614e-01 1.28210425e+00 -8.81726325e-01 7.58626699e-01
2.31604010e-01 -4.89232600e-01 3.52155328e-01 -4.14907753e-01
-5.03082946e-02 -4.14770305e-01 -5.72357237e-01 1.52745497e+00
-3.90738815e-01 1.82734758e-01 1.94174740e-02 -1.51955807e+00
1.14406872e+00 4.75251943e-01 1.25205851e+00 -1.14261520e+00
5.15111744e-01 3.06765139e-01 3.92264165e-02 -9.57881808e-01
-1.66681617e-01 -5.82554400e-01 3.39262694e-01 1.16222650e-01
-4.91781950e-01 1.27678752e-01 2.62019813e-01 -2.67296344e-01
7.62571216e-01 -8.84344056e-02 2.71941185e-01 -1.18471980e-01
1.26780427e+00 1.30660161e-01 3.72534692e-01 1.99940816e-01
-2.64607251e-01 1.24301359e-01 1.18553586e-01 -4.44960952e-01
-7.17074931e-01 -6.08118057e-01 -3.65551203e-01 4.45942402e-01
1.60051987e-01 2.48588666e-01 -7.16846704e-01 1.58260122e-01
-7.62894601e-02 6.26545250e-02 -4.21376824e-01 -1.17290534e-01
-1.58579350e-01 -1.09092057e+00 3.07848036e-01 -2.45744409e-03
3.47746104e-01 -8.81919026e-01 -8.39117706e-01 5.90986848e-01
8.46378654e-02 -1.14764225e+00 5.06653011e-01 -9.67981145e-02
-1.17701626e+00 -1.25500977e+00 -9.65874374e-01 -2.65359312e-01
5.12365937e-01 7.73215666e-02 6.73480809e-01 2.87865669e-01
-8.96983266e-01 1.11144215e-01 -4.43285704e-01 -7.46368110e-01
-6.99460506e-01 1.65410456e-03 -1.55185521e-01 2.12682486e-01
2.31860325e-01 -5.76383471e-01 -1.11968875e+00 2.92715818e-01
-8.87412727e-01 1.76035743e-02 8.53127718e-01 6.40089393e-01
7.81678319e-01 -4.45969515e-02 8.52184176e-01 -8.82193446e-01
4.32601988e-01 -4.71535563e-01 -5.17246127e-01 2.94436544e-01
-7.04518020e-01 -2.19746344e-02 6.11211181e-01 -5.16334884e-02
-1.03902090e+00 -8.85072649e-02 -2.41870463e-01 -3.62702571e-02
-4.62213993e-01 8.70299935e-01 -5.79543747e-02 -6.14054613e-02
8.44102383e-01 1.37831315e-01 5.10713935e-01 -2.91258663e-01
-1.38227329e-01 6.98503435e-01 3.06757689e-01 -4.80346352e-01
6.22246452e-02 7.10676968e-01 6.69335365e-01 -1.17939067e+00
-5.39857745e-01 -6.39915049e-01 -4.25156772e-01 -6.44406915e-01
1.04563212e+00 -4.15158033e-01 -1.06116748e+00 2.94148177e-01
-7.86963761e-01 9.86545905e-02 -1.10873789e-01 7.70670295e-01
-2.84257174e-01 6.02324367e-01 -4.94850129e-01 -9.62575376e-01
-5.33122301e-01 -1.42840064e+00 8.64101708e-01 4.25133556e-01
-2.32317820e-01 -1.12176788e+00 6.11489266e-02 8.83130610e-01
5.01013577e-01 7.49103069e-01 1.11650550e+00 -3.94064695e-01
-5.17753422e-01 -4.58450794e-01 5.13825007e-03 7.02039376e-02
2.54032880e-01 -1.07609108e-01 -8.82296383e-01 1.70779228e-01
2.00323135e-01 2.32237931e-02 5.94664931e-01 6.33766830e-01
1.08099520e+00 9.75977555e-02 -4.69357789e-01 1.80697680e-01
1.43437219e+00 6.49736822e-01 6.33410335e-01 1.28358617e-01
3.52375388e-01 9.18646991e-01 7.07371116e-01 5.96220374e-01
2.01769806e-02 5.60065031e-01 4.05087382e-01 -1.90834180e-01
1.49809886e-02 3.29551876e-01 -1.56728238e-01 6.55688345e-01
-4.27209407e-01 -7.60928020e-02 -9.95563686e-01 1.56705827e-01
-1.56566525e+00 -1.19775164e+00 -4.77362126e-01 2.28674531e+00
6.92541182e-01 -1.89733773e-01 -1.75173193e-01 5.10739088e-01
6.58568799e-01 -3.09594661e-01 -5.40804267e-01 -2.45362833e-01
-2.69881099e-01 3.06837916e-01 2.36833751e-01 1.68037072e-01
-7.36442268e-01 2.16657102e-01 5.61954260e+00 4.27295595e-01
-1.76166129e+00 6.20689802e-02 9.01631653e-01 1.57050416e-01
-2.64265835e-01 -4.22407001e-01 -3.43508720e-01 5.04106462e-01
9.58417296e-01 1.36292269e-02 1.86651558e-01 5.12800336e-01
7.91642249e-01 -6.29666030e-01 -7.34362900e-01 9.60965216e-01
-1.14177719e-01 -1.47321320e+00 -9.84185487e-02 3.28295618e-01
3.76342773e-01 -3.20640057e-01 1.83693096e-01 -4.39870209e-01
-6.12655163e-01 -8.95607829e-01 8.00865963e-02 1.03068435e+00
5.87787271e-01 -3.87785941e-01 9.68570530e-01 3.43410283e-01
-6.32046163e-01 -5.80765828e-02 -1.29960790e-01 -1.72547325e-01
5.58611974e-02 1.42310941e+00 -7.78621435e-01 6.69971466e-01
3.75153184e-01 5.60725570e-01 -3.64245057e-01 9.22485113e-01
1.67499021e-01 3.25042248e-01 -2.55677272e-02 -1.97672576e-01
-2.22770616e-01 -4.54730123e-01 3.58043224e-01 9.11651850e-01
3.35109383e-01 1.92679480e-01 -3.19793709e-02 7.82068491e-01
4.50266212e-01 4.89674658e-01 -5.13205051e-01 -4.41724330e-01
1.49601534e-01 1.39041257e+00 -1.29711330e+00 -4.04283047e-01
-2.23017305e-01 4.53427613e-01 -3.51146698e-01 1.40878707e-01
-5.66124320e-01 7.81549513e-02 4.30035442e-01 4.35795397e-01
-3.45597863e-01 -2.60331154e-01 -5.47933340e-01 -1.07889485e+00
-3.81840914e-01 -6.94518745e-01 2.53642201e-01 -4.33724821e-01
-9.45666850e-01 2.96135604e-01 -8.77786651e-02 -9.64892983e-01
-9.02819857e-02 -8.02934229e-01 -2.78842211e-01 7.96817362e-01
-1.38008952e+00 -9.29572582e-01 -5.20798385e-01 2.39580169e-01
-3.53160244e-03 -2.65065255e-03 9.93656635e-01 5.47426045e-01
-6.41441941e-01 -1.23554952e-01 4.69518572e-01 -2.07253098e-01
5.24580777e-01 -7.97079325e-01 -7.54735470e-01 4.96535569e-01
-2.59308338e-01 3.87477219e-01 7.29831934e-01 -4.38517511e-01
-1.51408565e+00 -6.27564669e-01 5.66661060e-01 1.26190454e-01
4.33875114e-01 1.84943616e-01 -7.52789080e-01 -1.06843114e-01
-3.09199333e-01 1.00709625e-01 1.16481829e+00 -1.64854050e-01
3.74573737e-01 -3.89371008e-01 -1.33276939e+00 3.81850481e-01
2.30214864e-01 -2.42424920e-01 -1.48135856e-01 4.07666832e-01
-3.12401950e-02 -2.04338536e-01 -1.40410888e+00 5.38113654e-01
8.34822476e-01 -1.04440176e+00 9.67469096e-01 -5.16367197e-01
3.00392926e-01 -2.82201171e-01 1.89042464e-01 -7.03640580e-01
-4.33743075e-02 1.04288682e-01 2.04863623e-01 7.64785707e-01
3.68870646e-02 -5.98972976e-01 1.00370777e+00 6.53369427e-01
9.79807824e-02 -1.03585613e+00 -9.90406573e-01 -4.33526218e-01
-2.10829481e-01 -3.16999942e-01 2.68963844e-01 7.21025646e-01
2.41609618e-01 -1.13681391e-01 1.52144566e-01 -2.06929818e-01
6.79656267e-01 1.11764073e-01 3.49644959e-01 -1.51553738e+00
-1.78209051e-01 -6.21995866e-01 -9.36715543e-01 3.09209619e-02
-1.30124599e-01 -6.29845917e-01 -4.44146425e-01 -1.54537416e+00
1.65929928e-01 -5.96855402e-01 -4.88990754e-01 -6.92912564e-02
-1.51166143e-02 6.44125402e-01 3.03712976e-03 3.26236933e-01
-1.53360635e-01 1.78030103e-01 1.45224106e+00 -3.49682063e-01
-8.97681713e-02 -4.74867485e-02 -5.04885077e-01 5.03825724e-01
9.10212278e-01 -2.14471281e-01 -2.08497345e-01 2.09565833e-01
3.66968602e-01 3.52316707e-01 4.33533102e-01 -1.30046642e+00
2.13387772e-01 -1.98885351e-01 5.96361578e-01 -2.75147885e-01
1.06539249e-01 -9.97792125e-01 5.35990179e-01 1.11674035e+00
-1.22112691e-01 -4.65558231e-01 -8.60661119e-02 3.70823175e-01
-3.29153508e-01 -1.82398856e-01 8.81621242e-01 -3.30828547e-01
-6.35161340e-01 2.44887527e-02 -6.72866881e-01 -4.09334451e-01
1.48561668e+00 -5.66993356e-01 -2.11407676e-01 -7.26838335e-02
-1.14035988e+00 -5.95422722e-02 3.07385266e-01 -1.57829747e-01
5.51624179e-01 -9.07017946e-01 -5.09521067e-01 3.96238752e-02
1.25949904e-01 -6.43680319e-02 9.93737578e-01 1.45776451e+00
-8.70236993e-01 5.75521767e-01 -4.52231854e-01 -8.44743073e-01
-1.17448580e+00 6.83517098e-01 3.57666016e-01 -4.10451233e-01
-9.49441046e-02 1.00727431e-01 -1.66662913e-02 2.08603442e-01
-8.84032398e-02 -2.31961727e-01 -5.78812420e-01 2.82458812e-01
6.18250787e-01 5.70126712e-01 3.37759405e-01 -6.10660076e-01
-4.28007782e-01 7.06227958e-01 3.13205183e-01 8.03228021e-02
1.25276268e+00 -2.05560237e-01 -3.01567644e-01 5.31799734e-01
8.11749816e-01 -1.45384163e-01 -3.49195242e-01 1.57375455e-01
2.21552357e-01 -1.21237881e-01 7.97348395e-02 -1.12335801e+00
-9.71893013e-01 1.11391640e+00 7.76482463e-01 2.28726894e-01
1.35897529e+00 -1.74737480e-02 3.82273525e-01 1.75702095e-01
3.98305297e-01 -9.21364427e-01 -4.34063971e-02 -2.91605353e-01
7.27784634e-01 -1.26407993e+00 3.11789155e-01 -5.39899647e-01
-3.51200849e-01 1.35696971e+00 8.85189250e-02 4.76945728e-01
5.78480721e-01 3.55337948e-01 2.86450773e-01 -2.48257682e-01
-2.73357600e-01 -2.01635420e-01 2.64218688e-01 5.92523038e-01
8.81369710e-01 -4.71464451e-03 -8.84401560e-01 4.23864692e-01
3.84439528e-01 5.46262920e-01 5.47160618e-02 6.86500371e-01
-5.76835811e-01 -1.24417233e+00 -7.23053217e-01 4.27444458e-01
-7.51693070e-01 2.93197244e-01 -1.71267942e-01 5.18092871e-01
2.18786702e-01 9.44314539e-01 -2.99003512e-01 -1.95614457e-01
2.14859527e-02 -2.98649967e-02 5.18683612e-01 -3.04699153e-01
-3.83303881e-01 1.40601769e-01 -2.62307137e-01 -3.92336071e-01
-9.11625028e-01 -5.01587391e-01 -1.41051316e+00 4.29619066e-02
-2.94331282e-01 2.37529948e-01 1.30130231e+00 1.17622519e+00
4.49735314e-01 7.59395361e-01 5.15964329e-01 -7.41206110e-01
-2.14399964e-01 -8.78933132e-01 -5.22528529e-01 2.91911870e-01
-1.48091346e-01 -5.40507674e-01 9.16080549e-02 1.94745466e-01] | [14.873992919921875, -2.7301135063171387] |
fc93237d-7989-4434-83e3-1a8d3dc5c64f | system-design-for-an-integrated-lifelong | 2212.04603 | null | https://arxiv.org/abs/2212.04603v1 | https://arxiv.org/pdf/2212.04603v1.pdf | System Design for an Integrated Lifelong Reinforcement Learning Agent for Real-Time Strategy Games | As Artificial and Robotic Systems are increasingly deployed and relied upon for real-world applications, it is important that they exhibit the ability to continually learn and adapt in dynamically-changing environments, becoming Lifelong Learning Machines. Continual/lifelong learning (LL) involves minimizing catastrophic forgetting of old tasks while maximizing a model's capability to learn new tasks. This paper addresses the challenging lifelong reinforcement learning (L2RL) setting. Pushing the state-of-the-art forward in L2RL and making L2RL useful for practical applications requires more than developing individual L2RL algorithms; it requires making progress at the systems-level, especially research into the non-trivial problem of how to integrate multiple L2RL algorithms into a common framework. In this paper, we introduce the Lifelong Reinforcement Learning Components Framework (L2RLCF), which standardizes L2RL systems and assimilates different continual learning components (each addressing different aspects of the lifelong learning problem) into a unified system. As an instantiation of L2RLCF, we develop a standard API allowing easy integration of novel lifelong learning components. We describe a case study that demonstrates how multiple independently-developed LL components can be integrated into a single realized system. We also introduce an evaluation environment in order to measure the effect of combining various system components. Our evaluation environment employs different LL scenarios (sequences of tasks) consisting of Starcraft-2 minigames and allows for the fair, comprehensive, and quantitative comparison of different combinations of components within a challenging common evaluation environment. | ['Aswin Raghavan', 'Jesse Hostetler', 'Michael Piacentino', 'Ajay Divakaran', 'Roberto Corizzo', 'Christopher Kanan', 'Zsolt Kira', 'Michael Baron', 'Nathalie Japkowicz', 'Sahana Joshi', 'James Smith', 'Mustafa Burak Gurbuz', 'Cameron E. Taylor', 'Tyler L. Hayes', 'Gianmarco J. Gallardo', 'Kamil Faber', 'Abrar Rahman', 'Zachary Daniels', 'Indranil Sur'] | 2022-12-08 | null | null | null | null | ['real-time-strategy-games', 'starcraft'] | ['playing-games', 'playing-games'] | [-3.84132057e-01 -7.36940478e-04 -1.25767350e-01 2.31230352e-02
-5.23575366e-01 -6.92322254e-01 7.89536834e-01 4.08280529e-02
-6.61579072e-01 9.53823686e-01 -3.21466178e-01 -2.44335592e-01
-4.60944682e-01 -5.77046394e-01 -8.59802842e-01 -4.94021088e-01
-4.81986493e-01 5.53423584e-01 5.84189177e-01 -5.24073899e-01
-2.73681190e-02 5.91803849e-01 -1.86202061e+00 -5.72953112e-02
7.16099024e-01 5.67899644e-01 5.51734269e-01 1.04284525e+00
2.51577526e-01 8.84760559e-01 -5.87216973e-01 3.42458859e-02
4.66217041e-01 -2.67124236e-01 -9.13296580e-01 -4.73701432e-02
5.50425500e-02 -2.36867487e-01 -6.85838610e-02 5.03134787e-01
6.68345034e-01 4.94044900e-01 3.43716770e-01 -1.62042642e+00
-1.96242213e-01 3.14443767e-01 -4.56336625e-02 1.72416985e-01
2.27828205e-01 7.86820054e-01 6.46884084e-01 -5.52433074e-01
4.29951310e-01 9.51341391e-01 8.32798839e-01 5.87660611e-01
-1.14479411e+00 -3.99668813e-01 1.45058721e-01 2.42248535e-01
-1.01985228e+00 -4.90241200e-01 3.23815763e-01 -3.71574342e-01
1.32158279e+00 -2.01163888e-01 8.17899287e-01 1.01648223e+00
5.59614658e-01 8.96194100e-01 1.17791867e+00 -4.99816835e-01
5.61786771e-01 1.25629276e-01 -3.22451331e-02 5.71551621e-01
2.67406255e-01 7.03847468e-01 -5.51004589e-01 2.32418612e-01
7.00614393e-01 -1.43186033e-01 2.75879145e-01 -6.68844283e-01
-9.75471795e-01 4.39473331e-01 1.30047306e-01 4.98337805e-01
-3.94270778e-01 4.82697070e-01 3.15530628e-01 8.39677811e-01
8.50727484e-02 8.30018044e-01 -7.01183796e-01 -6.57543004e-01
-5.43147206e-01 5.40498316e-01 9.10972595e-01 9.07587528e-01
1.11134708e+00 2.62555867e-01 -1.49350360e-01 6.74396455e-01
-9.87099260e-02 -9.09140483e-02 8.76752973e-01 -1.29251003e+00
-1.46131963e-01 4.33312505e-01 3.49994391e-01 -2.08165407e-01
-7.28578448e-01 -8.77603292e-01 -9.65137631e-02 6.23154104e-01
2.08079547e-01 -5.30014396e-01 -6.07787788e-01 1.96139133e+00
2.78021932e-01 2.83511609e-01 4.10296053e-01 4.89903629e-01
2.43566409e-01 6.67329311e-01 1.45870596e-01 -2.72712380e-01
7.81187773e-01 -1.23901212e+00 -2.16952026e-01 -6.48491740e-01
7.64285266e-01 -3.15999717e-01 1.58065498e+00 6.43931568e-01
-1.06532192e+00 -6.85974956e-01 -1.13994849e+00 3.24126869e-01
-6.00462615e-01 -3.16836119e-01 6.66350961e-01 5.13442874e-01
-1.24338341e+00 8.26586843e-01 -9.04945016e-01 -5.25990963e-01
-5.58968075e-02 4.04398590e-01 -2.14086190e-01 1.20218843e-01
-1.24762356e+00 1.27437794e+00 6.06041014e-01 -5.10323346e-01
-1.53723168e+00 -6.86995566e-01 -7.53135026e-01 -8.10787678e-02
7.47199416e-01 -6.25708282e-01 1.91842008e+00 -9.27454650e-01
-1.76321149e+00 5.97596884e-01 6.63029194e-01 -5.79370201e-01
4.41108346e-01 -5.67458808e-01 -2.92000502e-01 -2.51747757e-01
1.62924267e-02 6.42511010e-01 8.00463498e-01 -1.14134741e+00
-8.80082011e-01 -1.26704589e-01 4.79312003e-01 4.44839418e-01
-3.77891898e-01 -4.55089599e-01 -3.00722003e-01 -6.25071943e-01
-6.81593001e-01 -1.17595351e+00 -1.97971120e-01 -3.55178207e-01
5.56103945e-01 -4.08993334e-01 8.13664615e-01 -1.02627531e-01
1.08516574e+00 -2.25741887e+00 2.04560101e-01 -3.97545934e-01
-1.96570083e-01 4.47114795e-01 -4.19757575e-01 6.57592773e-01
5.56109734e-02 -2.54044831e-01 -1.25063762e-01 -3.17787766e-01
2.07201079e-01 6.25368416e-01 3.72517332e-02 1.06847674e-01
1.81875601e-02 9.41236794e-01 -1.17913997e+00 -2.36864552e-01
2.78708220e-01 1.74717773e-02 -5.93389332e-01 3.91124040e-01
-6.46808326e-01 2.04796836e-01 -1.92720786e-01 4.91573364e-01
-3.45679633e-02 -3.22430930e-03 9.55854133e-02 4.07653898e-01
-1.54136822e-01 -1.18094862e-01 -1.06159198e+00 1.98522639e+00
-8.31535995e-01 4.11060214e-01 -1.21340603e-01 -8.83289933e-01
7.67288744e-01 2.17112780e-01 4.90312546e-01 -7.87087142e-01
-1.77928023e-02 2.41726533e-01 -1.20334672e-02 -5.57417333e-01
5.28325379e-01 -2.63675630e-01 -4.75644410e-01 5.76170564e-01
6.16431236e-01 -4.24114704e-01 5.75920463e-01 -4.19832580e-02
1.27324450e+00 5.26931345e-01 3.67025942e-01 -3.17846477e-01
2.49324441e-01 9.28963274e-02 5.95338106e-01 1.11662269e+00
-5.57427764e-01 1.20510142e-02 2.42425054e-01 -6.61513269e-01
-8.68615746e-01 -1.11140811e+00 2.82656610e-01 1.62990487e+00
1.46900704e-02 -3.05197030e-01 -6.80892110e-01 -7.05477059e-01
2.14485005e-01 1.16764927e+00 -5.51168978e-01 -5.06532311e-01
-4.06540453e-01 -7.33929038e-01 5.76126218e-01 2.92925209e-01
6.44644797e-01 -1.54512358e+00 -1.26011109e+00 3.24900687e-01
3.39133650e-01 -8.73691499e-01 -1.63426325e-01 5.87409914e-01
-8.46673012e-01 -1.06450200e+00 -3.93071681e-01 -7.15484321e-01
-2.52851062e-02 1.71737820e-01 1.31089211e+00 -1.32646039e-01
-2.61494488e-01 1.01261079e+00 -5.11500835e-01 -3.89555931e-01
-5.80178320e-01 4.45325583e-01 5.05188882e-01 -5.40830076e-01
-3.31541039e-02 -8.49176884e-01 -3.87040883e-01 1.50172800e-01
-9.20196652e-01 -6.03304319e-02 7.67949104e-01 8.99128377e-01
3.80318493e-01 2.00765997e-01 1.13358486e+00 -7.36756027e-01
9.77003872e-01 -5.84006250e-01 -7.08853781e-01 4.32001889e-01
-1.20922828e+00 3.26528966e-01 7.24081397e-01 -5.32013297e-01
-1.03857648e+00 -1.20440975e-01 -5.37695140e-02 -3.06580186e-01
-8.36083218e-02 6.05513215e-01 5.90347908e-02 -1.84314903e-02
8.72882962e-01 3.75196725e-01 1.81468442e-01 -3.07772070e-01
5.71681619e-01 3.45694661e-01 5.18835366e-01 -9.39545929e-01
5.23222327e-01 -2.90513456e-01 -2.52703041e-01 -5.32967985e-01
-7.02052593e-01 6.77731214e-03 -3.25360388e-01 -5.33151209e-01
3.40047926e-01 -6.94774687e-01 -7.82499790e-01 7.67758846e-01
-5.40534019e-01 -1.38977385e+00 -1.03848445e+00 2.43686974e-01
-1.13050640e+00 -1.30476505e-01 -5.22819757e-01 -7.06422091e-01
-1.36366755e-01 -1.15589237e+00 6.16871297e-01 5.18015504e-01
-1.37139827e-01 -9.50018823e-01 6.52680039e-01 -1.36993453e-01
5.53583443e-01 3.07793878e-02 8.06086540e-01 -5.96254587e-01
-1.32113844e-01 -8.97948891e-02 4.10401613e-01 4.99602795e-01
6.40556067e-02 -3.33536834e-01 -8.48259985e-01 -7.65704751e-01
-1.54564708e-01 -9.75567698e-01 5.79914749e-01 7.91969225e-02
8.03538561e-01 -1.65177211e-01 1.65157288e-01 4.65873122e-01
1.32488060e+00 3.47017199e-01 4.55255121e-01 8.43857467e-01
1.94046244e-01 4.62120801e-01 9.23014998e-01 4.40259576e-01
4.39442605e-01 5.69939613e-01 5.65622866e-01 2.28501871e-01
-1.70669794e-01 -3.41944754e-01 7.76555479e-01 6.71529114e-01
1.08620405e-01 9.28726271e-02 -9.29095447e-01 6.62846327e-01
-2.06834888e+00 -7.37773418e-01 6.85251653e-01 2.22458100e+00
9.87940729e-01 3.55340123e-01 3.13432962e-01 -8.52374211e-02
2.12911159e-01 -5.33272102e-02 -1.09969151e+00 -6.43568575e-01
2.65495747e-01 1.86756775e-01 2.09941313e-01 5.99279523e-01
-9.12944496e-01 1.20166433e+00 7.01021528e+00 9.53652740e-01
-9.83642101e-01 2.16393262e-01 3.99630785e-01 -1.75376266e-01
1.39227271e-01 6.08740076e-02 -7.38711476e-01 1.78639174e-01
1.09354675e+00 -5.21460116e-01 7.45766878e-01 1.12881434e+00
6.73806816e-02 -3.58367205e-01 -1.05813849e+00 7.65483737e-01
-2.82684654e-01 -1.10695100e+00 -2.59410739e-01 -2.83231109e-01
7.37342298e-01 3.30896705e-01 1.24332540e-01 1.22916245e+00
8.42213094e-01 -7.66258597e-01 7.04480171e-01 6.06089473e-01
7.91943192e-01 -9.13477719e-01 4.34729248e-01 7.63805091e-01
-8.39227676e-01 -5.22043407e-01 7.74421245e-02 -3.96622300e-01
-2.28262916e-01 2.68242240e-01 -8.20208549e-01 3.82231295e-01
8.99379015e-01 7.43303418e-01 -8.58109713e-01 9.13605928e-01
-2.66449332e-01 4.59612608e-01 -6.81605414e-02 -7.20546097e-02
2.97144622e-01 1.58202127e-01 5.52962840e-01 1.05771399e+00
3.04602325e-01 -2.50439584e-01 4.23983723e-01 4.77057666e-01
1.64511427e-01 -2.57136822e-01 -6.70531511e-01 -4.73557189e-02
4.01428074e-01 1.26528454e+00 -4.34198052e-01 -1.77464142e-01
-1.49966210e-01 7.37630844e-01 5.24447203e-01 1.61393866e-01
-6.34111762e-01 -2.85838008e-01 7.19237804e-01 -1.57274585e-02
-3.14778946e-02 -4.51879114e-01 6.83621094e-02 -8.98792684e-01
-3.00183922e-01 -1.20815444e+00 4.24732983e-01 -8.61124575e-01
-1.14583135e+00 5.72537065e-01 1.67488173e-01 -1.04223549e+00
-6.84425294e-01 -2.92313188e-01 -4.32482779e-01 3.64417672e-01
-1.50590932e+00 -8.43864024e-01 -4.15761203e-01 6.40713930e-01
8.98367465e-01 -5.31681538e-01 9.95169222e-01 4.05856594e-02
-6.81632221e-01 4.52956468e-01 1.09431781e-01 -7.28476465e-01
7.80593753e-01 -1.37924922e+00 3.46776128e-01 6.86809361e-01
-1.42240137e-01 3.53541404e-01 7.56122708e-01 -5.92133701e-01
-1.42228150e+00 -1.15833008e+00 2.57803619e-01 -2.61904269e-01
6.08170390e-01 -2.28235230e-01 -7.99053788e-01 6.45036697e-01
2.32033670e-01 -8.36404115e-02 3.75959218e-01 1.60371393e-01
2.09771544e-02 -5.50643206e-01 -9.87626731e-01 7.80343950e-01
9.72070873e-01 -3.23609829e-01 -6.28357887e-01 1.49076268e-01
9.61515903e-01 -1.87201053e-01 -7.53139675e-01 5.27767420e-01
5.50945044e-01 -1.09112787e+00 7.51712918e-01 -5.00432789e-01
8.14760625e-02 -2.44735330e-01 7.07967877e-02 -1.72174478e+00
-3.64988625e-01 -9.05247271e-01 -4.24149573e-01 9.25993145e-01
2.17208967e-01 -6.71674430e-01 6.64309621e-01 2.66030788e-01
-2.90977478e-01 -7.75867224e-01 -8.24179053e-01 -1.33319342e+00
2.70352095e-01 -5.92079222e-01 4.05621499e-01 7.16994882e-01
7.22117200e-02 3.59279275e-01 -4.41576868e-01 -1.56816155e-01
3.30128849e-01 -2.36714542e-01 1.07500517e+00 -9.66674387e-01
-7.46720076e-01 -3.46359611e-01 -5.65461852e-02 -7.96695769e-01
3.69906053e-02 -7.07024932e-01 3.17351460e-01 -1.24050462e+00
-3.71445268e-02 -6.84756517e-01 -5.66630542e-01 8.04725409e-01
-1.82305444e-02 -1.94910496e-01 4.15402949e-01 2.43558243e-01
-1.07092309e+00 8.58322859e-01 1.04504597e+00 1.77370682e-01
-5.37517548e-01 9.49334800e-02 -6.21274650e-01 7.41784692e-01
1.03597224e+00 -3.60822588e-01 -8.62933159e-01 -3.69853199e-01
4.01189238e-01 9.65219662e-02 2.02261969e-01 -1.71417975e+00
4.02592838e-01 -2.49992982e-01 -3.89788002e-02 -2.36102822e-03
2.02071026e-01 -7.66987324e-01 4.32854444e-02 7.71573305e-01
-2.67271787e-01 3.76942843e-01 5.80677032e-01 4.42911506e-01
2.35388596e-02 -3.88960421e-01 6.90511644e-01 -1.57917082e-01
-1.19956398e+00 1.14733987e-01 -7.22566009e-01 1.68864980e-01
1.58084416e+00 -1.37893364e-01 -2.59512246e-01 -4.02561724e-01
-9.68627512e-01 5.79938233e-01 5.69713533e-01 6.93621993e-01
4.02929485e-01 -9.65020239e-01 -4.50774044e-01 1.78324685e-01
2.06039086e-01 -2.38998905e-01 6.15922548e-02 4.14039820e-01
-4.13351536e-01 1.53927565e-01 -5.96585214e-01 -3.56952846e-01
-7.96835124e-01 5.80213010e-01 7.05980957e-01 -7.43133128e-01
-3.52705300e-01 6.26951039e-01 -2.03167964e-02 -8.95924330e-01
3.13906103e-01 5.05356453e-02 -3.57818194e-02 -1.89213321e-01
2.82634079e-01 5.09924829e-01 9.98904854e-02 1.31758764e-01
-2.11445987e-01 1.19767591e-01 -1.14243187e-01 -3.16893697e-01
1.63667762e+00 -2.96804588e-02 3.05258989e-01 8.82748723e-01
4.70530719e-01 -5.07427871e-01 -1.89752507e+00 -1.73240170e-01
1.55972853e-01 8.65611061e-02 -7.24159852e-02 -9.80211616e-01
-6.63335323e-01 5.86630464e-01 8.08979034e-01 7.91913271e-02
1.17653728e+00 -2.94175059e-01 5.43915212e-01 6.92794442e-01
1.03058302e+00 -1.39411056e+00 6.85221553e-01 8.20014954e-01
9.48765039e-01 -1.01769459e+00 -3.72266248e-02 3.72955322e-01
-6.33682966e-01 9.82808709e-01 1.02085268e+00 -2.73403138e-01
4.90354270e-01 3.00453782e-01 -1.72757596e-01 -3.64794425e-04
-1.16103649e+00 -3.84333134e-01 -2.27214828e-01 7.70738125e-01
8.44504982e-02 -7.28786737e-02 -2.92016745e-01 6.47549868e-01
-1.63660049e-01 4.78704363e-01 5.41661739e-01 1.32508349e+00
-5.81721306e-01 -1.48367798e+00 -5.64031079e-02 2.34578550e-01
-7.70445168e-02 3.06717753e-01 8.83698871e-04 8.99587452e-01
4.66182172e-01 7.39530444e-01 -1.39367610e-01 -5.01307786e-01
4.38754946e-01 6.66929483e-02 6.94201767e-01 -8.30961287e-01
-7.47892261e-01 -3.38791013e-01 9.56361741e-02 -5.70797384e-01
-2.41568968e-01 -6.80197299e-01 -1.04984617e+00 -1.76287457e-01
-1.49531877e-02 1.78819031e-01 4.80143189e-01 9.34107780e-01
5.00030994e-01 7.51726151e-01 4.45785075e-01 -9.28454518e-01
-7.91844904e-01 -7.86365628e-01 -5.93635499e-01 1.59095332e-01
1.73953831e-01 -8.82142782e-01 -5.17207868e-02 -2.47579068e-02] | [4.130638122558594, 1.4674049615859985] |
754d616d-d0fd-4fb5-b92f-11cb0b4804e3 | single-channel-speech-enhancement-using | 1605.01329 | null | http://arxiv.org/abs/1605.01329v1 | http://arxiv.org/pdf/1605.01329v1.pdf | Single Channel Speech Enhancement Using Outlier Detection | Distortion of the underlying speech is a common problem for single-channel
speech enhancement algorithms, and hinders such methods from being used more
extensively. A dictionary based speech enhancement method that emphasizes
preserving the underlying speech is proposed. Spectral patches of clean speech
are sampled and clustered to train a dictionary. Given a noisy speech spectral
patch, the best matching dictionary entry is selected and used to estimate the
noise power at each time-frequency bin. The noise estimation step is formulated
as an outlier detection problem, where the noise at each bin is assumed present
only if it is an outlier to the corresponding bin of the best matching
dictionary entry. This framework assigns higher priority in removing spectral
elements that strongly deviate from a typical spoken unit stored in the trained
dictionary. Even without the aid of a separate noise model, this method can
achieve significant noise reduction for various non-stationary noises, while
effectively preserving the underlying speech in more challenging noisy
environments. | ['Eunjoon Cho', 'Bowon Lee', 'Ronald Schafer', 'Bernard Widrow'] | 2016-05-04 | null | null | null | null | ['noise-estimation'] | ['medical'] | [ 5.18350005e-01 -2.80035436e-01 1.65856734e-01 5.27821295e-02
-8.65682602e-01 -2.07437843e-01 3.22125033e-02 3.63228083e-01
-4.32430685e-01 4.67756659e-01 5.23411572e-01 -2.64498349e-02
-3.77667472e-02 -5.03773510e-01 -3.72827947e-01 -1.36122549e+00
9.44933444e-02 -1.12705976e-01 1.44952446e-01 -2.74183512e-01
8.68697241e-02 4.15112078e-01 -1.77534091e+00 4.61626165e-02
9.24749792e-01 8.50208938e-01 5.95206976e-01 6.05603993e-01
8.21695030e-02 2.17988878e-01 -1.05910742e+00 2.99139749e-02
2.55670846e-01 -7.44141459e-01 -1.69931296e-02 3.76388878e-01
3.21069956e-01 -1.07431039e-01 -4.46910679e-01 1.58902538e+00
9.25641596e-01 5.51808178e-01 2.71068752e-01 -6.22796834e-01
-1.09318793e-01 4.11179870e-01 -2.54611343e-01 5.16967714e-01
9.24689993e-02 -2.37450913e-01 7.63772190e-01 -1.05304217e+00
3.09553534e-01 7.50732422e-01 6.68791890e-01 3.44330847e-01
-1.27195895e+00 -5.08214593e-01 -1.46832958e-01 1.76833153e-01
-1.50541914e+00 -9.78397012e-01 1.10787678e+00 2.46537086e-02
7.86718309e-01 5.25475383e-01 6.13227367e-01 6.76623642e-01
-9.68261510e-02 4.98331249e-01 8.83948147e-01 -5.38549721e-01
4.55527097e-01 -2.82129496e-02 -7.16003925e-02 8.76922831e-02
2.05360606e-01 1.26159981e-01 -7.78474808e-01 -2.24495322e-01
3.40129077e-01 -2.94822663e-01 -7.02495217e-01 -1.40656903e-01
-8.16941142e-01 5.20918071e-01 5.51005118e-02 7.34010637e-01
-6.67114079e-01 -3.31594497e-01 4.14438576e-01 5.65551937e-01
6.49810195e-01 2.56905109e-01 -1.42364964e-01 -2.15827569e-01
-1.46355510e+00 2.56852150e-01 5.77768683e-01 6.48445189e-01
6.22130036e-01 6.23069823e-01 7.34578818e-02 1.15164578e+00
1.20468542e-01 6.57301664e-01 2.64176667e-01 -8.59347820e-01
1.96941063e-01 1.04342982e-01 2.83057746e-02 -1.04295051e+00
-5.96400090e-02 -8.71733725e-01 -9.66615736e-01 1.19920067e-01
1.87565118e-01 2.71339267e-02 -8.36000025e-01 1.44145894e+00
4.76465434e-01 4.20903504e-01 1.25548720e-01 8.45227361e-01
4.58168745e-01 8.94589067e-01 -3.51831734e-01 -1.06009483e+00
9.20599043e-01 -5.56427598e-01 -1.19355178e+00 -1.95323735e-01
-1.99815370e-02 -1.28778660e+00 6.95305407e-01 7.32197464e-01
-1.22395396e+00 -6.47333622e-01 -1.18539166e+00 2.42487416e-01
6.15179911e-02 -1.51060976e-03 -1.86311617e-01 7.90357828e-01
-1.03928196e+00 3.83970588e-01 -5.34102738e-01 -9.35688149e-03
-7.86211938e-02 1.13980584e-01 -1.78998515e-01 -1.26487747e-01
-1.09649026e+00 7.18888879e-01 1.37227222e-01 1.23005971e-01
-8.23992014e-01 -5.96298993e-01 -8.56111586e-01 9.07164663e-02
3.24916601e-01 -1.33220837e-01 1.02229691e+00 -9.51520920e-01
-1.34512842e+00 4.49539810e-01 -6.90480411e-01 -4.25570786e-01
1.08034469e-01 -1.26310224e-02 -1.08702660e+00 1.98498890e-01
-9.22025833e-03 -3.08142900e-01 1.88237166e+00 -1.34095967e+00
-4.92131054e-01 -3.70860338e-01 -7.47344196e-01 4.05315101e-01
-3.32647592e-01 1.31655708e-01 -5.45823157e-01 -1.24266636e+00
7.84561336e-01 -5.44723868e-01 -1.19224794e-01 -4.02769744e-01
-3.23394388e-01 2.54235923e-01 1.03160369e+00 -1.14056408e+00
1.66981244e+00 -2.78502965e+00 4.48016897e-02 4.98163790e-01
1.05431698e-01 5.84322274e-01 -6.82103410e-02 3.12414348e-01
-8.19185525e-02 -3.62632334e-01 -2.97549188e-01 -6.05429769e-01
-1.74237370e-01 2.44402885e-01 -2.38637730e-01 6.73522949e-01
-2.03080967e-01 -2.36489251e-02 -8.06864381e-01 -1.42822683e-01
2.83433795e-01 7.71848083e-01 -5.18936098e-01 2.34307200e-01
2.52002686e-01 4.49802309e-01 1.29887834e-01 4.46048737e-01
1.00658119e+00 5.40841699e-01 3.58514749e-02 -4.34421986e-01
-3.69902045e-01 5.23801088e-01 -1.82330549e+00 1.33476198e+00
-3.06323707e-01 7.30898082e-01 8.80225003e-01 -1.07276165e+00
1.21338987e+00 6.95968449e-01 5.09668946e-01 -7.53210127e-01
-1.73707962e-01 4.85982478e-01 2.62729883e-01 -4.56964493e-01
6.81168437e-01 -2.66293198e-01 4.14114505e-01 2.49436237e-02
9.81803164e-02 -1.76819056e-01 1.35048211e-01 -4.81434241e-02
8.91886353e-01 -6.77867949e-01 2.16765866e-01 -3.62715691e-01
6.11731052e-01 -5.77853382e-01 7.83501029e-01 7.13097394e-01
-4.28859621e-01 6.42539918e-01 -1.30800635e-01 1.00585625e-01
-1.34343600e+00 -1.07996118e+00 -3.85944605e-01 1.08340693e+00
3.30115706e-02 -3.81312996e-01 -8.29172373e-01 -1.06593138e-02
-1.10637330e-01 5.18390536e-01 -1.10646114e-01 -1.83601126e-01
-5.31340241e-01 -4.85612452e-01 4.20878291e-01 -3.18102986e-02
3.32197756e-01 -7.56046295e-01 -2.05701664e-01 6.12462580e-01
-4.29404229e-01 -8.74824703e-01 -7.95905709e-01 5.71018398e-01
-6.44593120e-01 -6.00950062e-01 -7.68284678e-01 -9.51169074e-01
7.24840403e-01 4.71352130e-01 5.45236945e-01 1.08233765e-01
7.83160999e-02 1.19861186e-01 -4.50007141e-01 -3.78356308e-01
-8.13917160e-01 -3.41789067e-01 5.29339194e-01 3.61828804e-01
2.35036060e-01 -5.11968911e-01 -5.09335518e-01 2.77810663e-01
-7.68188834e-01 -5.94221711e-01 1.90198809e-01 9.43089008e-01
7.21700728e-01 8.76827240e-01 7.40741253e-01 -2.58904159e-01
7.65229702e-01 -1.94428608e-01 -4.11542803e-01 -4.01537031e-01
-3.62200975e-01 -4.37190890e-01 7.69957840e-01 -4.84844208e-01
-1.04917622e+00 -1.87774539e-01 -4.96229678e-01 -2.98722386e-01
-3.28805327e-01 5.80295384e-01 -5.39888799e-01 -6.39448762e-02
5.92379451e-01 7.01143026e-01 6.36731833e-02 -8.53632331e-01
-1.10359564e-01 9.60457027e-01 8.75258386e-01 -6.45112395e-02
8.73290658e-01 3.13225925e-01 -4.28999364e-01 -1.64491737e+00
-2.99684644e-01 -1.04864311e+00 -3.68711889e-01 -2.07208380e-01
3.04861516e-01 -8.33437860e-01 -4.48033726e-03 6.49349153e-01
-9.68461752e-01 5.93557470e-02 -4.63357300e-01 6.12397730e-01
-1.48629501e-01 7.13334799e-01 -2.67971486e-01 -1.22895551e+00
-2.64627576e-01 -1.22121656e+00 7.26333439e-01 1.18370540e-01
-1.65610373e-01 -7.78884709e-01 -1.81239527e-02 1.42854705e-01
6.30674243e-01 -4.81766939e-01 6.92468286e-01 -6.00918651e-01
-8.07810426e-02 -3.29151005e-01 4.84883398e-01 8.55805099e-01
5.70845008e-01 -2.08646730e-01 -1.21951854e+00 -7.14578569e-01
8.18279326e-01 3.47607851e-01 7.06307292e-01 6.50876343e-01
7.05595911e-01 -1.94144413e-01 -3.13898213e-02 5.11654675e-01
1.14153111e+00 6.43622696e-01 7.38857210e-01 1.27590463e-01
2.72127897e-01 3.31468403e-01 5.42229772e-01 4.69504148e-01
-3.52028817e-01 6.94485128e-01 2.16189444e-01 -3.59576881e-01
-3.65615666e-01 8.22494365e-03 4.79005188e-01 1.35785258e+00
2.72277504e-01 -2.52975017e-01 -5.38183391e-01 9.33686614e-01
-1.25465810e+00 -1.18313468e+00 -1.10896761e-02 2.50121260e+00
1.02485013e+00 2.16874585e-01 -8.31823796e-02 8.99579525e-01
8.75948012e-01 3.88912439e-01 -4.64021802e-01 -2.82817423e-01
-5.03720164e-01 4.66129869e-01 1.59963518e-01 7.38554001e-01
-1.07391584e+00 6.02344692e-01 6.37066650e+00 1.05879867e+00
-1.16874862e+00 3.34878592e-03 1.61930785e-01 -5.92377819e-02
-1.29115343e-01 -2.25177094e-01 -5.49253106e-01 5.75899184e-01
8.15944314e-01 -2.06016630e-01 6.52267754e-01 5.21817267e-01
9.06673610e-01 -2.78063536e-01 -5.81105947e-01 1.12332845e+00
1.44264355e-01 -9.03028965e-01 -2.53643811e-01 2.19577812e-02
7.43206859e-01 -1.73151061e-01 1.78705975e-01 -9.25467014e-02
-3.97086561e-01 -7.08485305e-01 6.19259357e-01 2.87326068e-01
5.35095811e-01 -9.60022390e-01 6.09597504e-01 4.32584256e-01
-1.26269948e+00 8.77911747e-02 -5.01252890e-01 6.12479560e-02
1.09464779e-01 1.05362284e+00 -7.08000004e-01 4.38868135e-01
7.83928275e-01 4.18622553e-01 1.76210925e-01 1.35283279e+00
-1.06561661e-01 9.53725755e-01 -4.30731803e-01 5.09537876e-01
-7.95978159e-02 -3.03611457e-01 1.23623884e+00 1.25178456e+00
6.11547112e-01 3.77558991e-02 1.59776434e-01 4.09089863e-01
1.17848506e-02 2.40198329e-01 -5.03552318e-01 1.47325918e-01
8.31647456e-01 8.23986828e-01 -4.16760325e-01 -2.22880989e-01
-5.02978325e-01 9.53357339e-01 -4.12207901e-01 5.77562869e-01
-3.49581599e-01 -5.43760359e-01 1.10222208e+00 3.81333008e-02
5.62128603e-01 -1.97425961e-01 -2.58215815e-01 -7.65448093e-01
1.29040331e-01 -1.39125383e+00 4.01841564e-04 -4.90833372e-01
-1.07999980e+00 5.85214853e-01 -6.35269940e-01 -1.52291501e+00
-7.09479600e-02 -2.01211721e-01 -5.99188924e-01 1.21394956e+00
-1.29512799e+00 -4.22692955e-01 -2.78945584e-02 8.53872776e-01
6.48113668e-01 -3.26237470e-01 9.13007677e-01 7.64761806e-01
-3.87231857e-01 6.01139665e-01 5.31322420e-01 -1.35314047e-01
6.67773783e-01 -1.08881283e+00 1.61008000e-01 1.35726810e+00
1.24815643e-01 6.32680535e-01 1.19156241e+00 -8.52860570e-01
-1.02112162e+00 -9.95468318e-01 1.11905897e+00 3.20929587e-01
4.77040082e-01 -2.22196952e-01 -1.31687129e+00 -2.73153838e-02
2.44436309e-01 -3.97039279e-02 7.51172602e-01 -2.39297956e-01
1.57459140e-01 -2.85849184e-01 -1.08954966e+00 4.50560600e-01
6.82226002e-01 -9.36711371e-01 -5.52056670e-01 -3.57729234e-02
4.68662441e-01 -4.24942225e-01 -7.86932826e-01 1.48963332e-01
2.88554300e-02 -8.58587325e-01 8.42734814e-01 6.56794291e-03
-3.46523345e-01 -7.69711018e-01 -3.92249942e-01 -1.79338670e+00
-2.20184416e-01 -1.11892200e+00 -2.32213274e-01 1.28890908e+00
2.45204762e-01 -2.64109373e-01 6.19496226e-01 1.39161907e-02
-4.80891258e-01 -1.23906054e-01 -1.15852082e+00 -8.87071133e-01
-4.55357760e-01 -5.42836487e-01 3.72218758e-01 9.44998205e-01
5.42562641e-02 2.10296968e-03 -5.93477070e-01 7.08756328e-01
9.13973451e-01 -2.84085751e-01 5.51140070e-01 -9.80569839e-01
-2.52763003e-01 -2.01648355e-01 -3.66338730e-01 -1.14314246e+00
-1.63999811e-01 -5.20457268e-01 5.94440103e-01 -1.25258815e+00
-4.13046211e-01 -1.59221083e-01 -2.89987177e-01 -2.70020012e-02
-3.88557345e-01 2.26977199e-01 2.36694366e-02 1.03740849e-01
4.26630750e-02 6.31424606e-01 1.10466206e+00 -3.21202844e-01
-3.81727785e-01 4.21476781e-01 -3.78144830e-01 6.89315975e-01
6.53794646e-01 -5.69286108e-01 -4.43972290e-01 -2.39090428e-01
-2.62641937e-01 1.06984116e-01 1.13601230e-01 -1.28088081e+00
5.47184050e-01 2.19756171e-01 2.33404282e-02 -7.49392867e-01
4.00142699e-01 -1.11884308e+00 2.20567122e-01 3.15964311e-01
-7.14959875e-02 -1.70867726e-01 3.16674888e-01 5.46195567e-01
-7.39243865e-01 -3.95877630e-01 1.22085357e+00 2.03095645e-01
-6.80604637e-01 4.16040011e-02 -7.29700148e-01 -1.89363003e-01
4.97523695e-01 -4.78811502e-01 2.83676118e-01 -5.73111534e-01
-1.12827182e+00 -4.81491327e-01 3.42473418e-01 1.12319924e-01
7.95276046e-01 -1.04644418e+00 -7.19851434e-01 8.15014899e-01
-1.64633021e-01 -2.07894206e-01 4.48380917e-01 7.52155721e-01
-1.00542203e-01 -3.73311006e-02 2.94068903e-01 -6.10671222e-01
-1.36876988e+00 3.48124892e-01 5.72519243e-01 1.88120008e-01
-7.20615566e-01 8.24540317e-01 -1.70261532e-01 1.68771252e-01
6.19096994e-01 -3.28521729e-01 -5.39245456e-02 -4.40393314e-02
8.72372329e-01 5.18134654e-01 7.09462762e-01 -9.79930878e-01
-1.25756353e-01 3.44853818e-01 1.50195882e-01 -1.35502473e-01
1.19917095e+00 -5.14316201e-01 -1.92167491e-01 3.78184199e-01
1.09511137e+00 7.21461177e-01 -1.21953404e+00 -5.66273093e-01
-5.80806844e-02 -7.27730811e-01 6.49329126e-01 -6.15471244e-01
-9.81923878e-01 5.22738993e-01 8.67742896e-01 4.74820912e-01
1.66553116e+00 -4.26575631e-01 7.93385863e-01 2.45008394e-02
1.29553601e-01 -1.37499392e+00 7.19186366e-02 6.35472953e-01
7.73914993e-01 -9.47766781e-01 -3.15958172e-01 -2.97255069e-01
-2.44192511e-01 9.46328938e-01 2.22171411e-01 7.13492259e-02
7.37264633e-01 6.04695797e-01 3.14674675e-01 7.72670656e-02
-8.21190476e-02 -3.44346821e-01 2.70090491e-01 7.82151341e-01
2.32091755e-01 -1.24585129e-01 -1.60477981e-01 3.36537033e-01
-2.35957846e-01 -4.92090404e-01 3.24687660e-01 6.99769378e-01
-9.46886420e-01 -1.04085660e+00 -8.68890822e-01 2.77926356e-01
-4.65289682e-01 -3.91658753e-01 5.65402023e-02 1.78381324e-01
2.93042749e-01 1.28774679e+00 1.77598417e-01 -3.06485325e-01
7.31597483e-01 1.52485922e-01 1.36542246e-01 -5.07418990e-01
-6.08532906e-01 1.03586817e+00 -7.86049962e-02 -2.33097285e-01
-2.68781036e-01 -8.35589230e-01 -1.14297235e+00 -3.37022871e-01
-4.12531465e-01 3.47152352e-01 6.35068536e-01 6.98187768e-01
-1.08740118e-03 5.12344122e-01 8.47954154e-01 -6.29533470e-01
-4.63984311e-01 -9.10573542e-01 -1.09690690e+00 5.42822778e-01
9.53419209e-01 -4.03569192e-01 -6.86054170e-01 3.02705944e-01] | [15.00318431854248, 5.833864212036133] |
4dbc5814-02f8-497d-8029-80a098f22a51 | functions-that-emerge-through-end-to-end | 1703.02239 | null | http://arxiv.org/abs/1703.02239v2 | http://arxiv.org/pdf/1703.02239v2.pdf | Functions that Emerge through End-to-End Reinforcement Learning - The Direction for Artificial General Intelligence - | Recently, triggered by the impressive results in TV-games or game of Go by
Google DeepMind, end-to-end reinforcement learning (RL) is collecting
attentions. Although little is known, the author's group has propounded this
framework for around 20 years and already has shown various functions that
emerge in a neural network (NN) through RL. In this paper, they are introduced
again at this timing.
"Function Modularization" approach is deeply penetrated subconsciously. The
inputs and outputs for a learning system can be raw sensor signals and motor
commands. "State space" or "action space" generally used in RL show the
existence of functional modules. That has limited reinforcement learning to
learning only for the action-planning module. In order to extend reinforcement
learning to learning of the entire function on a huge degree of freedom of a
massively parallel learning system and to explain or develop human-like
intelligence, the author has believed that end-to-end RL from sensors to motors
using a recurrent NN (RNN) becomes an essential key. Especially in the higher
functions, this approach is very effective by being free from the need to
decide their inputs and outputs.
The functions that emerge, we have confirmed, through RL using a NN cover a
broad range from real robot learning with raw camera pixel inputs to
acquisition of dynamic functions in a RNN. Those are (1)image recognition,
(2)color constancy (optical illusion), (3)sensor motion (active recognition),
(4)hand-eye coordination and hand reaching movement, (5)explanation of brain
activities, (6)communication, (7)knowledge transfer, (8)memory, (9)selective
attention, (10)prediction, (11)exploration. The end-to-end RL enables the
emergence of very flexible comprehensive functions that consider many things in
parallel although it is difficult to give the boundary of each function
clearly. | ['Katsunari Shibata'] | 2017-03-07 | null | null | null | null | ['color-constancy', 'game-of-go'] | ['computer-vision', 'playing-games'] | [ 7.48312753e-03 3.88511479e-01 -6.84676692e-02 -1.28997892e-01
-2.02177651e-02 -6.03339612e-01 7.47205257e-01 -5.10291278e-01
-4.06876802e-01 8.94941270e-01 1.90437317e-01 -2.40787551e-01
-4.16599661e-01 -5.68446994e-01 -9.03026819e-01 -8.40420544e-01
-1.41518325e-01 8.64199474e-02 1.32337451e-01 -6.99346900e-01
4.92025793e-01 6.19983017e-01 -1.91066539e+00 4.15078178e-02
4.03181672e-01 8.67365897e-01 7.26166129e-01 8.08150649e-01
-1.58630207e-01 1.25653017e+00 -3.24470609e-01 3.92808527e-01
1.15600355e-01 -6.85147345e-01 -1.07455659e+00 9.68394205e-02
-4.82611090e-01 -1.09232709e-01 -1.92585811e-01 8.10859621e-01
3.24437559e-01 1.90354362e-01 4.04626727e-01 -1.18620145e+00
-8.49424899e-01 4.68524218e-01 -4.34576049e-02 -2.19434813e-01
5.10609329e-01 5.04014075e-01 7.89182425e-01 -5.71034729e-01
6.70186579e-01 1.20572293e+00 2.34392151e-01 9.79960561e-01
-1.05799365e+00 -4.18525398e-01 -5.49532212e-02 3.44673872e-01
-8.60593855e-01 -2.51913834e-02 5.11545956e-01 -3.54440540e-01
1.09887636e+00 1.59987763e-01 1.07761395e+00 1.25758815e+00
3.89227599e-01 8.56009901e-01 1.30223930e+00 -4.69203472e-01
3.72337520e-01 1.45756736e-01 -2.60381073e-01 5.48168242e-01
-3.48967552e-01 4.19019699e-01 -6.27361655e-01 6.33710325e-01
1.34090865e+00 1.74667567e-01 -3.22590441e-01 -2.92552084e-01
-1.24321198e+00 7.52846301e-01 6.49154007e-01 4.66385305e-01
-4.60447997e-01 5.06205320e-01 2.39531800e-01 7.40907907e-01
-3.39668840e-01 7.88205147e-01 -8.06755424e-01 -5.19164681e-01
-3.65520149e-01 -1.01965934e-01 6.55771077e-01 6.56900346e-01
8.46856773e-01 3.32725555e-01 3.13756317e-01 6.09187663e-01
3.25543493e-01 3.77026051e-01 1.04027033e+00 -1.24183023e+00
-9.60496143e-02 8.14905405e-01 3.23539339e-02 -5.45203507e-01
-8.08323979e-01 -3.89036477e-01 -7.50729978e-01 8.92774999e-01
3.35303903e-01 -3.05132836e-01 -7.22510695e-01 1.92788029e+00
2.20315456e-01 -2.99322188e-01 3.00875992e-01 1.00785351e+00
6.90908551e-01 8.11899483e-01 -1.46681190e-01 -1.71860129e-01
1.11291122e+00 -9.13504541e-01 -5.04130185e-01 -4.70717810e-02
5.13878107e-01 -2.31774792e-01 1.30452752e+00 8.03909242e-01
-1.05145538e+00 -8.31311166e-01 -1.11625755e+00 9.38276723e-02
-6.93416536e-01 -2.34712839e-01 8.97926033e-01 1.36645317e-01
-1.33957648e+00 8.35980773e-01 -6.05161130e-01 -6.33005142e-01
1.59879580e-01 7.99296498e-01 -5.77307224e-01 2.79708713e-01
-1.20025980e+00 1.05795562e+00 4.47353870e-01 4.77994233e-02
-1.07514322e+00 -1.50307342e-01 -5.90289295e-01 -1.28069296e-01
4.32790220e-01 -8.47633719e-01 9.49373066e-01 -1.56239605e+00
-2.16996121e+00 6.75155580e-01 2.09595114e-01 -3.99122983e-01
1.65198371e-01 -3.15368474e-01 -2.21841127e-01 -8.11167434e-03
-1.99212804e-01 1.05528140e+00 7.93472230e-01 -1.20843232e+00
-5.99219084e-01 -4.48024482e-01 2.91321456e-01 4.00370061e-01
1.04914226e-01 -2.66039640e-01 7.08820522e-02 -3.37408066e-01
2.27625504e-01 -9.11393642e-01 -7.67971501e-02 -4.45360392e-02
-9.22670811e-02 -3.86162013e-01 8.00551772e-01 -2.77527124e-01
7.60309994e-01 -2.04545999e+00 5.27618349e-01 -2.08099216e-01
1.03492215e-01 8.36989209e-02 -7.91868046e-02 6.40114903e-01
-4.04790998e-01 -2.02980161e-01 7.93166310e-02 3.18269879e-01
1.90169958e-03 4.70093787e-01 -1.58159822e-01 2.86029786e-01
9.01790187e-02 1.08531809e+00 -8.98601115e-01 -4.08997647e-02
3.03440988e-01 4.09764886e-01 -3.94302875e-01 3.36896300e-01
-2.50494629e-01 8.53962600e-01 -4.33149725e-01 5.31422138e-01
-1.11509763e-01 -2.30366126e-01 -1.90589890e-01 1.39794961e-01
-3.20356280e-01 -7.27175251e-02 -1.06782460e+00 1.79506159e+00
-4.64143455e-01 8.09441924e-01 8.35884735e-02 -1.15793455e+00
8.97440612e-01 4.46255386e-01 4.65613693e-01 -1.01254857e+00
2.60797322e-01 7.98453540e-02 3.94794285e-01 -8.91396224e-01
9.08171609e-02 -1.60015121e-01 2.42270604e-02 4.70182061e-01
2.88040549e-01 -1.77381396e-01 -1.79148853e-01 -2.05027759e-01
1.17438567e+00 8.73931348e-01 4.83237743e-01 -3.83331701e-02
5.17322958e-01 9.17609856e-02 1.59323364e-01 4.32956010e-01
-9.79131609e-02 3.32018435e-01 3.05128574e-01 -4.35792804e-01
-8.00445318e-01 -1.11122417e+00 2.23286048e-01 1.35441124e+00
-1.37137277e-02 2.43767262e-01 -7.40572512e-01 -1.65137723e-01
-3.74118537e-01 6.42009974e-01 -7.73310244e-01 -2.89516330e-01
-5.08939862e-01 -9.42191631e-02 2.84580559e-01 4.32027668e-01
6.84683025e-01 -2.12841702e+00 -1.18062484e+00 2.35678062e-01
1.91839322e-01 -7.56422758e-01 1.04110174e-01 8.95782411e-01
-9.83299494e-01 -9.18478668e-01 -6.13970816e-01 -8.28230679e-01
5.48557818e-01 4.17529717e-02 7.28510082e-01 -5.85560650e-02
-3.43583494e-01 7.54137516e-01 -4.08842593e-01 -4.49126005e-01
-2.26225957e-01 -2.83483416e-01 2.83929139e-01 -2.97515631e-01
-7.03930035e-02 -9.81043935e-01 -5.22461832e-01 2.62444049e-01
-9.88516212e-01 1.16473660e-01 9.63596165e-01 7.46207535e-01
3.77301872e-01 -2.10425153e-01 8.42313111e-01 -5.30845702e-01
7.01943099e-01 -3.52277398e-01 -3.29256326e-01 5.82712926e-02
-4.84401345e-01 2.73114383e-01 7.60079622e-01 -6.05539262e-01
-8.34558725e-01 1.82935804e-01 -6.67409077e-02 -1.73278794e-01
-5.37650943e-01 2.42373586e-01 -1.95491672e-01 5.39622195e-02
8.10024619e-01 7.91542947e-01 3.59613001e-01 -1.19772395e-02
5.66882968e-01 4.30575758e-01 5.89767098e-01 -2.98073143e-01
5.97092688e-01 2.17002735e-01 1.37243450e-01 -9.32615161e-01
-2.86726505e-01 -7.33972788e-02 -6.60793006e-01 -4.89990622e-01
1.04737675e+00 -6.21745527e-01 -1.50122809e+00 3.86012495e-01
-1.02985811e+00 -7.56615341e-01 -6.90284908e-01 6.14257455e-01
-1.04149044e+00 -6.83244541e-02 -5.83142519e-01 -8.83774221e-01
-9.08673853e-02 -1.06830108e+00 6.40455067e-01 6.52836978e-01
-3.82525891e-01 -1.02249682e+00 1.38853714e-02 3.95540856e-02
5.39053202e-01 2.12652311e-01 9.74079370e-01 -3.62650424e-01
-6.61344349e-01 2.05731429e-02 4.16913517e-02 4.67356622e-01
2.08132833e-01 -1.78177834e-01 -8.35688412e-01 2.34049279e-02
1.83653980e-01 -6.45771801e-01 4.44084257e-01 3.17481905e-01
1.09144986e+00 -1.76276565e-01 -9.64429602e-02 4.36301172e-01
1.39215815e+00 7.19590843e-01 1.17119312e+00 4.12373751e-01
4.54261184e-01 8.50360930e-01 3.83939564e-01 2.42405146e-01
2.04894960e-01 3.26666415e-01 1.00696290e+00 -1.24402493e-01
4.41989899e-02 -2.10727334e-01 9.45040703e-01 8.06862533e-01
-5.34056008e-01 1.23958908e-01 -3.93651426e-01 1.37334600e-01
-1.88079739e+00 -1.13550532e+00 1.06329858e-01 2.07008481e+00
6.54281557e-01 2.61348605e-01 1.29561514e-01 9.43436623e-02
4.28201407e-01 -2.47465491e-01 -9.91891444e-01 -8.42005312e-01
-2.23343428e-02 1.88418537e-01 1.84389442e-01 2.66961694e-01
-5.01635611e-01 1.21417093e+00 6.50043774e+00 5.99127710e-01
-1.37360513e+00 -2.56180078e-01 2.77554899e-01 1.77250281e-01
5.89017640e-04 -4.49773893e-02 -5.08811593e-01 2.62126714e-01
7.61702240e-01 4.42880660e-01 9.96143043e-01 9.02480781e-01
2.76502132e-01 -3.68158966e-01 -1.09989297e+00 1.27647436e+00
-1.98137969e-01 -1.10339153e+00 -1.12084605e-01 1.96780220e-01
3.63217980e-01 6.19979613e-02 -2.72783600e-02 6.23787642e-01
4.02786553e-01 -1.46813130e+00 6.33029282e-01 9.52076256e-01
6.13941729e-01 -5.73105514e-01 2.61215717e-01 7.70582497e-01
-8.66178811e-01 -4.75102723e-01 -3.29994231e-01 -4.65179533e-01
-7.65906349e-02 -5.98101392e-02 -5.76355815e-01 1.38405293e-01
4.30497289e-01 5.47362268e-01 -2.15674549e-01 5.93746424e-01
-5.48801720e-01 4.05680448e-01 -2.91914195e-02 -5.63832939e-01
4.50104803e-01 -2.19788134e-01 4.08790499e-01 8.47456872e-01
1.97126284e-01 2.87910908e-01 -3.95817131e-01 1.01550829e+00
2.92462826e-01 -1.80027097e-01 -8.22458923e-01 -1.24204777e-01
-1.29890129e-01 1.38685977e+00 -8.39498878e-01 2.88245138e-02
-1.72373563e-01 1.15023768e+00 2.33947664e-01 4.01816845e-01
-8.97251844e-01 -3.54773581e-01 6.19726181e-01 -4.66255471e-02
3.44631463e-01 -3.24623704e-01 4.51594591e-02 -6.93277180e-01
-3.23752731e-01 -8.18508148e-01 -2.49005973e-01 -1.17897475e+00
-8.19742799e-01 4.72631812e-01 -2.76748776e-01 -1.22496498e+00
-6.19835138e-01 -9.21205878e-01 -5.75510621e-01 4.10919368e-01
-1.04271710e+00 -8.91559064e-01 -2.20705017e-01 1.10812187e+00
7.55101144e-01 -3.40974122e-01 9.94658887e-01 -7.16845617e-02
-2.25280613e-01 2.96481196e-02 -2.07512230e-01 4.10428550e-03
2.28354841e-01 -1.25947928e+00 -2.57009953e-01 2.42304921e-01
2.68313587e-01 4.98800188e-01 5.03443003e-01 -1.78729311e-01
-1.80612683e+00 -4.74240512e-01 3.59551460e-01 -3.61145824e-01
4.22356874e-01 -2.32567698e-01 -6.28547907e-01 6.07178807e-01
3.79312873e-01 -2.41169199e-01 2.65328079e-01 -8.94889385e-02
8.98697674e-02 -2.69439995e-01 -9.45794702e-01 7.67730355e-01
9.46596920e-01 -3.83305341e-01 -6.87126994e-01 1.55643567e-01
7.20406473e-01 -1.69957370e-01 -3.49480420e-01 1.79973077e-02
8.70451152e-01 -1.23713398e+00 6.97810173e-01 -7.45875120e-01
5.28438509e-01 -3.75514448e-01 -2.10648149e-01 -1.32092535e+00
-4.46949840e-01 -7.39724457e-01 1.18115932e-01 8.11078727e-01
4.74592149e-01 -9.04102743e-01 5.66914260e-01 2.72941530e-01
-3.46835971e-01 -9.22865450e-01 -6.58292592e-01 -6.13599658e-01
-1.17761366e-01 -6.43092573e-01 1.99253708e-01 8.05588782e-01
4.59262282e-01 6.68359220e-01 -2.90949911e-01 -2.27305427e-01
4.88353148e-02 -7.12302327e-02 8.60450983e-01 -1.18793607e+00
-7.09958911e-01 -5.99352658e-01 -3.85475636e-01 -1.34036803e+00
-4.49293219e-02 -8.05160880e-01 7.89885819e-02 -1.74498904e+00
-8.83858278e-02 6.07900321e-02 -2.87543535e-01 5.82602382e-01
5.12404025e-01 -1.77045569e-01 2.49275222e-01 2.27510288e-01
-6.17519081e-01 4.01791573e-01 1.74848855e+00 3.52949977e-01
-2.57198334e-01 -6.76746741e-02 -7.43459821e-01 9.31601048e-01
8.83329690e-01 -2.39457756e-01 -5.88538349e-01 -1.57904699e-01
6.39889061e-01 4.60836381e-01 4.48852539e-01 -1.27886713e+00
4.17795062e-01 -1.92793712e-01 7.20480442e-01 -1.38542220e-01
4.93621916e-01 -1.05376339e+00 3.34161147e-02 9.93828475e-01
-2.59712964e-01 2.00223312e-01 1.44191002e-02 4.42803264e-01
-1.41731068e-01 -1.50716171e-01 6.77688301e-01 -6.21542215e-01
-1.15735888e+00 -1.86059415e-01 -6.85621202e-01 -1.91515818e-01
1.04411244e+00 -7.26377606e-01 -1.28562227e-01 -7.28220344e-01
-1.07256591e+00 1.61874831e-01 -7.26585314e-02 5.20110846e-01
6.99947476e-01 -1.09877610e+00 -1.82746768e-01 2.91880339e-01
-3.45972478e-01 -1.79931164e-01 5.51529117e-02 8.72530341e-01
-4.21905428e-01 5.18444479e-01 -7.22963572e-01 -5.77945709e-01
-8.09612274e-01 5.16425252e-01 3.64993840e-01 -1.34476230e-01
-4.50355947e-01 8.11764479e-01 3.82597327e-01 -4.56180781e-01
2.06833079e-01 -4.35560197e-01 -5.10178447e-01 -8.93018320e-02
3.91062587e-01 3.29596013e-01 -5.94132304e-01 -3.25069696e-01
-7.84043595e-02 7.75229990e-01 4.01229560e-01 -2.05924734e-01
1.51377892e+00 -1.11247472e-01 -1.48539245e-01 9.19385076e-01
9.33184147e-01 -4.81172979e-01 -1.43456137e+00 2.90009230e-01
-1.76043704e-01 3.18107635e-01 -2.75576234e-01 -1.08523560e+00
-9.28842843e-01 1.04888821e+00 7.15149999e-01 3.91413212e-01
1.20860910e+00 -4.66012321e-02 6.07877493e-01 5.76729596e-01
7.46675968e-01 -1.44614911e+00 6.10897958e-01 7.82488108e-01
1.14488518e+00 -1.21921158e+00 -2.84945130e-01 3.21185201e-01
-8.36046875e-01 1.49950743e+00 6.52659595e-01 -3.92645121e-01
4.84890252e-01 2.32432917e-01 -9.70684066e-02 -3.07604969e-01
-9.04153466e-01 -4.26213562e-01 -2.69839056e-02 7.00303078e-01
3.80779147e-01 -3.81740481e-02 -7.99549520e-02 5.28554738e-01
-4.05450016e-01 2.39251360e-01 3.68990660e-01 7.36730754e-01
-9.63029385e-01 -8.88366938e-01 -2.88263738e-01 -7.50945732e-02
-2.11798176e-01 2.74771303e-01 -2.81929344e-01 1.36934555e+00
4.76911008e-01 7.82958448e-01 -2.76429445e-01 -4.67522115e-01
2.07042575e-01 9.30226967e-02 7.85144329e-01 -4.74439949e-01
-6.73998117e-01 -1.16562545e-01 -3.86336923e-01 -8.30417931e-01
-4.36076343e-01 -3.80007744e-01 -1.94941282e+00 -4.58852053e-02
1.09803379e-01 -7.84118101e-02 9.25126612e-01 1.24158382e+00
8.20634663e-02 7.28994191e-01 5.17178297e-01 -1.01055479e+00
-3.71429861e-01 -9.25956488e-01 -7.59553254e-01 8.55062380e-02
2.12100908e-01 -5.44515312e-01 -3.82756948e-01 -7.12255994e-03] | [4.107966899871826, 1.3666601181030273] |
174791c6-a929-4a75-979d-af965be0b140 | mtr-multi-agent-motion-prediction-with | 2306.17770 | null | https://arxiv.org/abs/2306.17770v1 | https://arxiv.org/pdf/2306.17770v1.pdf | MTR++: Multi-Agent Motion Prediction with Symmetric Scene Modeling and Guided Intention Querying | Motion prediction is crucial for autonomous driving systems to understand complex driving scenarios and make informed decisions. However, this task is challenging due to the diverse behaviors of traffic participants and complex environmental contexts. In this paper, we propose Motion TRansformer (MTR) frameworks to address these challenges. The initial MTR framework utilizes a transformer encoder-decoder structure with learnable intention queries, enabling efficient and accurate prediction of future trajectories. By customizing intention queries for distinct motion modalities, MTR improves multimodal motion prediction while reducing reliance on dense goal candidates. The framework comprises two essential processes: global intention localization, identifying the agent's intent to enhance overall efficiency, and local movement refinement, adaptively refining predicted trajectories for improved accuracy. Moreover, we introduce an advanced MTR++ framework, extending the capability of MTR to simultaneously predict multimodal motion for multiple agents. MTR++ incorporates symmetric context modeling and mutually-guided intention querying modules to facilitate future behavior interaction among multiple agents, resulting in scene-compliant future trajectories. Extensive experimental results demonstrate that the MTR framework achieves state-of-the-art performance on the highly-competitive motion prediction benchmarks, while the MTR++ framework surpasses its precursor, exhibiting enhanced performance and efficiency in predicting accurate multimodal future trajectories for multiple agents. | ['Bernt Schiele', 'Dengxin Dai', 'Li Jiang', 'Shaoshuai Shi'] | 2023-06-30 | null | null | null | null | ['motion-prediction'] | ['computer-vision'] | [ 1.13523759e-01 -2.10851222e-01 -6.27062976e-01 -1.70046493e-01
-9.04252827e-01 -2.78775930e-01 8.63764822e-01 -1.35884270e-01
-5.11337757e-01 3.21462333e-01 5.75317264e-01 -3.21819156e-01
1.80424735e-01 -7.17309475e-01 -6.06436670e-01 -4.88715678e-01
-1.59392267e-01 3.91253233e-01 7.86719799e-01 -3.64977866e-01
2.05020308e-01 2.54115582e-01 -1.77680099e+00 3.13117713e-01
6.69724286e-01 7.28519082e-01 6.59264147e-01 1.00169909e+00
2.79134214e-01 1.40113258e+00 1.79628655e-01 -3.61908376e-01
6.31954595e-02 -1.15716182e-01 -6.53437734e-01 -1.07304618e-01
1.03959940e-01 -5.72943926e-01 -8.43864024e-01 4.86921489e-01
2.27773890e-01 5.01725614e-01 2.86780268e-01 -1.74041295e+00
-3.31363827e-01 3.11111122e-01 -2.70310640e-01 1.33232921e-01
5.77464402e-01 7.94904649e-01 9.96865571e-01 -9.48444307e-01
6.85128272e-01 1.11284852e+00 4.94402647e-01 6.98529422e-01
-7.14173317e-01 -4.07977819e-01 4.78002876e-01 9.07160401e-01
-1.39019191e+00 -6.88175321e-01 6.32141769e-01 -3.30092728e-01
1.32161522e+00 2.56446302e-01 6.64485574e-01 1.12643206e+00
3.34562302e-01 1.29020190e+00 2.72584707e-01 3.91500384e-01
1.78821638e-01 -3.47081333e-01 2.82115769e-02 8.41001511e-01
-3.81360829e-01 2.61715382e-01 -8.49218488e-01 9.45489630e-02
3.52562994e-01 1.99708402e-01 -2.60318797e-02 -3.83417606e-01
-1.77805316e+00 6.47581279e-01 3.95198613e-01 -2.03396752e-01
-6.61237419e-01 5.69190025e-01 3.79285187e-01 -8.77555460e-02
-6.69582263e-02 -7.56128505e-02 -6.45868480e-02 -7.53018439e-01
-7.05742180e-01 5.19918501e-01 3.84628296e-01 1.39315557e+00
9.41155672e-01 2.58427486e-02 -4.43584323e-01 5.31683326e-01
5.67301095e-01 8.44087064e-01 4.09284890e-01 -1.35855639e+00
8.27657640e-01 6.72887564e-01 2.42883474e-01 -1.09930706e+00
-5.37451506e-01 -4.74323519e-02 -5.60033023e-01 -1.16085373e-01
4.00638431e-02 -1.82521006e-03 -7.99797773e-01 1.68265462e+00
4.14312154e-01 7.48058736e-01 3.77158433e-01 9.82046545e-01
7.70633221e-01 9.64205503e-01 4.79715765e-01 9.98843312e-02
1.18122864e+00 -1.49042737e+00 -5.41730702e-01 -6.36723876e-01
9.71426785e-01 -3.55931371e-01 8.92914355e-01 -3.88264507e-02
-1.03091133e+00 -6.92772210e-01 -8.13809037e-01 -2.39783049e-01
-1.08413234e-01 9.71936360e-02 4.74005252e-01 2.17675447e-01
-1.11348236e+00 3.50292847e-02 -1.22447896e+00 -3.61853421e-01
2.17336193e-01 3.91852885e-01 -3.80582243e-01 -3.45954865e-01
-1.00719595e+00 8.01466048e-01 3.20071608e-01 1.31087705e-01
-1.05153155e+00 -5.84624052e-01 -1.33683169e+00 -1.86004788e-01
3.82163614e-01 -8.86757135e-01 1.49839413e+00 -2.31190130e-01
-1.54674518e+00 1.20411046e-01 -7.42161989e-01 -7.26891518e-01
5.55730462e-01 -2.04235971e-01 -7.11373508e-01 1.27420351e-01
4.35983390e-01 1.29912305e+00 6.88145399e-01 -1.05888748e+00
-1.36107492e+00 -1.86917499e-01 -9.91374329e-02 5.59744895e-01
-2.27986602e-05 -4.39474940e-01 -9.99614656e-01 -2.44383380e-01
-1.34137779e-01 -1.16720331e+00 -5.86301565e-01 -2.93657482e-01
-3.36554796e-02 -2.48514518e-01 1.29747009e+00 -4.71451551e-01
1.35114110e+00 -2.05973005e+00 2.43513376e-01 -3.34496275e-02
3.20595473e-01 -2.35366449e-02 -4.69792336e-01 6.09941781e-01
5.96829891e-01 -5.07252872e-01 -2.36272112e-01 -6.43389702e-01
3.27640921e-01 3.83302659e-01 -3.68827611e-01 3.15681428e-01
1.74997509e-01 1.46381044e+00 -9.47847724e-01 -5.91557741e-01
6.65839076e-01 3.52142811e-01 -6.67799354e-01 2.19376773e-01
-3.93712491e-01 5.56913614e-01 -5.62068284e-01 6.49483383e-01
3.01913381e-01 -2.35660553e-01 -2.60599609e-02 -1.34774327e-01
-3.84464949e-01 1.97368816e-01 -9.78591621e-01 1.78858447e+00
-4.99923766e-01 7.58863151e-01 -1.44334674e-01 -4.21175539e-01
6.29326582e-01 1.67461678e-01 7.37592399e-01 -9.90577459e-01
-2.21341774e-01 2.11149524e-03 -3.26946169e-01 -7.84194291e-01
1.18122590e+00 3.27317327e-01 -4.02632684e-01 2.15812296e-01
-4.33999926e-01 2.33720228e-01 2.74700224e-01 2.23286226e-01
1.17869270e+00 3.69390100e-01 7.59736001e-02 3.72566283e-01
6.40467942e-01 5.80152452e-01 6.87967837e-01 6.07607603e-01
-6.29427135e-01 2.14788347e-01 -1.36369467e-01 -5.05291820e-01
-7.82713950e-01 -9.45283055e-01 4.37646538e-01 1.32398248e+00
8.93122435e-01 -5.85809290e-01 -3.04856867e-01 -5.77252150e-01
-2.28702620e-01 1.01372433e+00 -3.72675687e-01 -2.09530026e-01
-9.70306635e-01 -5.02722204e-01 4.20097679e-01 8.00967395e-01
8.10546577e-01 -9.82360244e-01 -9.42697108e-01 3.17238986e-01
-9.00159836e-01 -1.51300240e+00 -7.30659127e-01 -6.30097449e-01
-3.79848957e-01 -7.88563907e-01 -3.52162391e-01 -7.38710046e-01
1.57397673e-01 7.92484760e-01 8.33484173e-01 -7.04327822e-02
2.07893163e-01 7.17772245e-01 -4.13742334e-01 1.04959369e-01
-4.93169039e-01 1.45838842e-01 -1.23343311e-01 1.88291118e-01
4.43722039e-01 -2.51696944e-01 -9.55776453e-01 5.79907000e-01
-6.76573753e-01 5.21028876e-01 6.38770163e-01 4.30850923e-01
4.56016570e-01 -1.05749607e-01 4.27900791e-01 -1.18667357e-01
1.20575882e-01 -8.38864028e-01 -3.25128406e-01 6.07985780e-02
-4.15149629e-01 -1.33789062e-01 4.70190138e-01 -2.93220788e-01
-1.10298634e+00 2.82250673e-01 -3.97843421e-01 -2.98556954e-01
-2.85637230e-01 3.43553901e-01 -8.11604187e-02 9.74214673e-02
1.11564659e-01 5.78293622e-01 -1.67316552e-02 1.56398490e-01
6.35550618e-01 4.23705518e-01 8.31195891e-01 -1.32402092e-01
6.59970880e-01 6.39504373e-01 -1.65997185e-02 -8.70240748e-01
-1.44253150e-01 -8.19562495e-01 -4.98208970e-01 -6.77558661e-01
1.21098971e+00 -1.18596578e+00 -1.04783714e+00 3.34944069e-01
-1.03314269e+00 -5.93235672e-01 -4.13502455e-02 7.10462630e-01
-9.05755043e-01 4.54519719e-01 -5.98265529e-01 -6.57301724e-01
3.61486375e-02 -1.66590893e+00 1.44874191e+00 9.61860791e-02
-2.40253299e-01 -9.83632565e-01 1.40667811e-01 7.87653506e-01
2.93167472e-01 -1.79343462e-01 5.93550503e-01 -2.75879830e-01
-1.25261903e+00 -1.07856967e-01 -9.45243463e-02 -5.44189930e-01
-1.70014381e-01 -3.58433604e-01 -5.68907976e-01 -1.57123655e-01
-6.13002300e-01 -5.81758246e-02 8.50684524e-01 3.03990066e-01
5.71485102e-01 -2.24085137e-01 -6.80005193e-01 4.29524243e-01
1.05303442e+00 3.12505424e-01 7.82173216e-01 4.14373875e-01
9.58758533e-01 5.95584750e-01 1.06248176e+00 4.08276260e-01
1.53358829e+00 1.01604891e+00 7.64835238e-01 2.12680131e-01
-1.16812319e-01 -5.02944648e-01 9.01239038e-01 7.53300369e-01
3.33631635e-02 -4.26246881e-01 -9.98995662e-01 8.05975080e-01
-2.51452994e+00 -1.35735798e+00 -3.75978798e-01 1.91375041e+00
2.52384041e-02 -1.14922903e-01 5.15144765e-01 -3.22004706e-01
3.62617046e-01 4.54712391e-01 -7.43223906e-01 3.27538401e-02
2.17504531e-01 -6.98825002e-01 5.24810910e-01 8.05028975e-01
-1.15635145e+00 1.35640275e+00 5.85321140e+00 7.87967563e-01
-8.09947670e-01 1.04132332e-01 2.23345950e-01 -1.06719425e-02
-3.79694134e-01 -1.09176561e-01 -1.01705611e+00 3.59563798e-01
9.67434645e-01 5.60166538e-02 3.47736210e-01 7.31390476e-01
6.89810872e-01 -1.15423314e-01 -9.21171069e-01 9.64408457e-01
-5.64221106e-02 -1.57743847e+00 1.34091496e-01 4.23665345e-02
5.07059872e-01 4.14406508e-01 1.46937966e-01 5.97504139e-01
4.47455198e-01 -8.35196793e-01 8.86767864e-01 7.33150184e-01
2.58705884e-01 -7.64047682e-01 5.02405763e-01 5.94607592e-01
-1.89541435e+00 -4.24031764e-01 3.68032418e-02 -4.97617908e-02
7.23287404e-01 -1.51512429e-01 -9.22100246e-01 6.04034483e-01
5.16151845e-01 1.24221969e+00 -5.65769970e-01 7.36481965e-01
2.44681031e-01 3.85414034e-01 -1.08572468e-01 -5.90734892e-02
6.23095512e-01 -1.68304339e-01 8.43686819e-01 1.13028371e+00
4.52062339e-01 -3.26279067e-02 5.85391879e-01 4.91380751e-01
3.75961334e-01 -1.26622245e-01 -7.93712616e-01 2.88928479e-01
5.97956777e-01 1.09184933e+00 -4.56562847e-01 -2.62165457e-01
-6.65849745e-01 1.02355170e+00 2.34926954e-01 4.46310848e-01
-1.30394733e+00 2.04149485e-01 1.07597756e+00 1.74568687e-03
3.88983876e-01 -6.93832934e-01 1.09248742e-01 -9.35261250e-01
-9.12644044e-02 -4.99794811e-01 3.39901805e-01 -6.67253315e-01
-6.43737316e-01 4.48053449e-01 -1.14307425e-03 -1.53943729e+00
-5.93077183e-01 -2.85703003e-01 -5.74513137e-01 2.32852429e-01
-1.62671459e+00 -1.76737511e+00 -3.61132801e-01 7.32586563e-01
1.06261671e+00 -1.60262764e-01 2.99907595e-01 3.42248440e-01
-7.56634772e-01 2.76380181e-01 -3.95730734e-01 -1.95122570e-01
2.84478396e-01 -9.46441829e-01 7.09057391e-01 1.05869281e+00
-1.26265049e-01 1.63706109e-01 5.35358965e-01 -7.65647888e-01
-2.01437020e+00 -1.70676792e+00 8.49823952e-01 -6.67132974e-01
5.87054431e-01 4.55153063e-02 -5.37203908e-01 9.77559745e-01
1.70431599e-01 -2.57247329e-01 5.88675857e-01 -4.86946970e-01
4.06509303e-02 1.23973712e-01 -5.39413750e-01 1.34003258e+00
1.29681075e+00 -3.69119078e-01 -2.56784230e-01 9.67773721e-02
8.46217036e-01 -4.26712334e-01 -8.00359130e-01 4.89785641e-01
5.92546344e-01 -1.00480151e+00 1.14618230e+00 -2.77586758e-01
2.87150949e-01 -7.63714612e-01 -4.44356054e-01 -7.51941323e-01
-4.65455532e-01 -7.08431780e-01 -4.13253814e-01 8.30982685e-01
6.05478466e-01 -3.55696708e-01 9.19122815e-01 8.72517526e-01
-5.21852255e-01 -6.32007301e-01 -1.00181842e+00 -3.88743013e-01
-4.99414235e-01 -1.23531139e+00 7.54790783e-01 4.64364767e-01
9.39076617e-02 2.92989045e-01 -7.48514116e-01 6.45246863e-01
3.83579701e-01 2.70380050e-01 1.26519489e+00 -3.77311856e-01
-9.40266997e-02 -5.01133800e-01 -6.26602292e-01 -1.90672374e+00
2.23561659e-01 -8.40831280e-01 3.68595272e-01 -1.74544454e+00
7.85375386e-02 -2.29291633e-01 1.84675664e-01 4.35050249e-01
-2.36553714e-01 2.55641729e-01 3.84216607e-01 3.29744697e-01
-1.39941478e+00 9.84497666e-01 1.24195218e+00 -2.72386730e-01
-4.55833346e-01 1.31027892e-01 -1.31627426e-01 6.86761141e-01
6.46669626e-01 -5.33457287e-02 -7.82187223e-01 -5.38219512e-01
3.10529266e-02 5.00478148e-01 8.32064986e-01 -1.25676548e+00
7.39696324e-01 -4.41549242e-01 -1.32170811e-01 -1.17162049e+00
8.02469254e-01 -8.78442645e-01 2.87566632e-01 4.54451382e-01
-2.71216780e-01 5.20703256e-01 1.91307709e-01 9.80977952e-01
-1.95156589e-01 3.42776090e-01 2.27554858e-01 2.33263835e-01
-1.77863932e+00 7.02816546e-01 -9.83299196e-01 -1.48607567e-01
1.39363682e+00 -3.98275673e-01 -3.44482183e-01 -7.69639432e-01
-5.79260588e-01 8.78082514e-01 4.70289081e-01 8.37927341e-01
1.03550541e+00 -1.53845572e+00 -5.89073718e-01 8.95781294e-02
4.91006434e-01 -2.00896174e-01 5.93283892e-01 1.21232200e+00
-4.42631453e-01 5.64470530e-01 1.25118583e-01 -9.98045325e-01
-1.21796703e+00 4.86685663e-01 1.66037768e-01 -1.25013739e-01
-8.62001181e-01 4.22477514e-01 3.25457543e-01 -4.79773670e-01
4.88559008e-02 -3.46401155e-01 -3.61281872e-01 -5.11080623e-01
5.84455013e-01 7.86231756e-01 -3.42487186e-01 -1.22917616e+00
-3.65293741e-01 5.15991688e-01 6.49264455e-02 -3.02995563e-01
8.24716747e-01 -8.14165533e-01 3.67451519e-01 1.34834751e-01
1.00615311e+00 -1.63549915e-01 -1.72240543e+00 -6.64529726e-02
-5.80039807e-02 -2.46695772e-01 -5.11528477e-02 -4.24046844e-01
-7.77107060e-01 6.40684724e-01 3.27230096e-01 -1.26959547e-01
9.83768702e-01 -3.40492162e-03 1.45923531e+00 4.58526611e-01
5.71737826e-01 -8.91730666e-01 1.10602796e-01 6.53899908e-01
5.55567205e-01 -1.35841382e+00 -5.62227368e-01 -3.87398064e-01
-1.25951350e+00 7.56633461e-01 8.04940164e-01 3.98726881e-01
3.46686035e-01 1.24032693e-02 1.14430273e-02 -6.29355684e-02
-1.17734063e+00 -5.17499983e-01 4.16178912e-01 7.29289174e-01
-2.13662624e-01 2.14599185e-02 2.75336891e-01 4.16446209e-01
-7.02755451e-02 -4.43435498e-02 1.90524116e-01 8.86009812e-01
-5.07516444e-01 -8.92916501e-01 -2.00753212e-01 1.35044321e-01
3.18505645e-01 1.03977650e-01 8.78019333e-02 7.26329029e-01
2.48630643e-02 1.30619168e+00 7.71096349e-02 -9.38226104e-01
2.73612410e-01 -1.37131318e-01 2.43410487e-02 -6.48025647e-02
-1.65600687e-01 -9.50444192e-02 3.04831237e-01 -1.20686924e+00
-4.82701451e-01 -9.72339094e-01 -1.50263822e+00 -4.66891110e-01
2.91412055e-01 -1.41191781e-01 2.84273267e-01 1.11031139e+00
8.51873815e-01 3.59091252e-01 4.61310476e-01 -1.07723463e+00
-2.31781900e-02 -4.18149740e-01 1.56695306e-01 2.50762284e-01
5.40651381e-01 -5.88518679e-01 3.15987617e-01 2.61109054e-01] | [5.904587745666504, 0.7840268015861511] |
54790de6-d473-4702-b13b-f9ba4d656477 | extraction-of-sleep-information-from-clinical | 2204.09601 | null | https://arxiv.org/abs/2204.09601v1 | https://arxiv.org/pdf/2204.09601v1.pdf | Extraction of Sleep Information from Clinical Notes of Alzheimer's Disease Patients Using Natural Language Processing | Alzheimer's Disease (AD) is the most common form of dementia in the United States. Sleep is one of the lifestyle-related factors that has been shown critical for optimal cognitive function in old age. . However, there is a lack of research studying the association between sleep and AD incidence. A major bottleneck for conducting such research is that the traditional way to acquire sleep information is time-consuming, inefficient, non-scalable, and limited to patients' subjective experience. In this study, we developed a rule-based NLP algorithm and machine learning models to automate the extraction of sleep-related concepts, including snoring, napping, sleep problem, bad sleep quality, daytime sleepiness, night wakings, and sleep duration, from the clinical notes of patients diagnosed with AD. We trained and validated the proposed models on the clinical notes retrieved from the University of Pittsburgh of Medical Center (UPMC). The results show that the rule-based NLP algorithm consistently achieved the best performance for all sleep concepts. | ['Yanshan Wang', 'Shyam Visweswaran', 'David Oniani', 'Samual Viggiano', 'Sonish Sivarajkumar', 'Haneef Ahamed Mohammad'] | 2022-03-08 | null | null | null | null | ['sleep-quality-prediction'] | ['medical'] | [-3.06737959e-01 -3.79156560e-01 -2.26398528e-01 -5.75916827e-01
-5.15700459e-01 -2.33489931e-01 -1.64444432e-01 5.57478964e-01
-6.35611296e-01 1.24390185e+00 5.94479740e-01 -5.60233533e-01
-4.27872807e-01 -5.64188004e-01 2.31467664e-01 -3.45118761e-01
-2.25270629e-01 6.65907145e-01 1.49506480e-01 1.72395572e-01
2.08556205e-01 3.40933055e-01 -1.33064342e+00 5.62874461e-03
1.29347920e+00 1.01467741e+00 6.15297079e-01 5.04259408e-01
-1.56264275e-01 3.50192517e-01 -5.66786349e-01 -9.96041484e-03
-2.00136844e-02 -4.55250949e-01 -6.58246338e-01 -6.26391917e-02
-1.71809867e-01 -3.50735128e-01 1.74176712e-02 8.20717990e-01
6.77468836e-01 2.80984342e-02 3.24733555e-01 -1.13077152e+00
-5.98459482e-01 -2.43206412e-01 3.31038505e-01 1.02479768e+00
4.29827064e-01 -5.49316742e-02 8.19261849e-01 -5.35499692e-01
2.17650190e-01 8.20262551e-01 5.62137604e-01 7.68037796e-01
-9.83805478e-01 -6.74119294e-01 -2.45286107e-01 8.62087429e-01
-1.18103158e+00 -6.60543680e-01 2.04271853e-01 -2.86421180e-01
1.14963007e+00 2.90992409e-01 1.66814649e+00 6.91176116e-01
7.68733501e-01 2.15909943e-01 1.11817265e+00 -4.30026144e-01
8.00778329e-01 7.97895715e-02 5.43778777e-01 5.89931369e-01
5.47461271e-01 -3.05699497e-01 -2.41174832e-01 -5.70342124e-01
7.82902837e-02 8.02608311e-01 -1.83437914e-01 1.31117404e-01
-6.70804083e-01 7.35228002e-01 1.67278677e-01 3.18842113e-01
-1.69193447e-01 -6.25840366e-01 1.30187586e-01 1.81216180e-01
4.94727314e-01 3.26036692e-01 -7.97100842e-01 -4.22436327e-01
-1.31298018e+00 5.59424646e-02 8.34317267e-01 5.51064074e-01
2.04696581e-01 -5.42241037e-01 3.14004272e-02 9.43139851e-01
8.02500486e-01 8.37093174e-01 1.00020897e+00 -8.92848194e-01
1.07406005e-01 1.12208700e+00 2.83375889e-01 -5.61316550e-01
-9.28046942e-01 -3.16585958e-01 -5.47218442e-01 -1.90424263e-01
9.98726711e-02 -9.26834419e-02 -6.96252942e-01 1.31247818e+00
-2.43877023e-01 -1.76249072e-01 1.50572196e-01 6.17379725e-01
8.33873391e-01 1.69791579e-01 3.97421807e-01 -8.53181720e-01
1.86644685e+00 -7.87910044e-01 -9.79701519e-01 -8.46985579e-01
3.43749404e-01 -6.33342147e-01 1.15417218e+00 2.01258898e-01
-9.40593362e-01 -1.27768785e-01 -9.82468903e-01 -2.37710059e-01
-4.20827299e-01 2.67553687e-01 4.83453870e-01 7.29849041e-01
-1.15034735e+00 3.36039126e-01 -1.19430649e+00 -1.13278282e+00
6.37140512e-01 4.60148931e-01 -2.20661774e-01 -2.26275906e-01
-8.78776312e-01 1.19181645e+00 -7.50595145e-03 -1.51401207e-01
-3.08807850e-01 -7.12387323e-01 -4.04235572e-01 8.04339126e-02
2.84082443e-02 -1.05609024e+00 1.27993083e+00 -3.59362394e-01
-8.97041976e-01 7.89175451e-01 -9.43896592e-01 -4.89225298e-01
-2.12199301e-01 -4.60219681e-01 -9.54675674e-01 1.78061068e-01
5.33769548e-01 1.69301391e-01 1.91615596e-01 -3.90833437e-01
-6.94722474e-01 -9.62222397e-01 -4.48449105e-01 1.57326963e-02
-3.72221678e-01 2.10070163e-01 2.00614959e-01 -1.48096412e-01
1.32016003e-01 -9.33637261e-01 -2.40833074e-01 2.64559388e-01
-1.42095432e-01 -4.15251732e-01 4.60583001e-01 -7.88359344e-01
1.69548559e+00 -2.31929779e+00 -5.23378849e-01 -2.24111304e-01
4.75571096e-01 2.32014969e-01 6.32068932e-01 1.63607702e-01
1.86309144e-01 1.25355765e-01 -2.54950553e-01 -3.85216177e-01
-2.57394850e-01 7.18793213e-01 2.25271955e-01 2.28503570e-01
-3.53003353e-01 6.64507210e-01 -1.13971615e+00 -7.28481650e-01
3.03599983e-01 3.16839516e-01 -5.06316960e-01 2.18818769e-01
2.91562285e-02 1.47933096e-01 -4.73395973e-01 6.66161656e-01
1.61655322e-01 -6.87262952e-01 2.06099391e-01 1.21977471e-01
-3.29857647e-01 8.61060262e-01 -3.79297823e-01 1.31512308e+00
-1.22381069e-01 4.54414785e-01 -3.11591893e-01 -3.26948166e-01
5.94783306e-01 2.28779614e-01 3.88870835e-01 -7.19331741e-01
6.53638765e-02 2.06470326e-01 2.00200945e-01 -1.03867137e+00
-1.31631032e-01 -4.74994540e-01 3.35690826e-01 4.36154068e-01
-3.73498708e-01 4.40264791e-01 4.56158131e-01 -6.91557210e-03
1.56432009e+00 -6.40401006e-01 1.26444614e+00 -3.56571078e-01
4.35530186e-01 6.84067756e-02 9.22897220e-01 2.81009704e-01
-6.62020028e-01 3.83551627e-01 1.60104424e-01 -4.43866372e-01
-6.66303933e-01 -1.24198449e+00 -1.49266303e-01 6.23926520e-01
-4.03312266e-01 -9.99065280e-01 -3.14520180e-01 -3.99047792e-01
-2.72109926e-01 9.97232616e-01 -2.77571410e-01 -3.07040900e-01
-1.79332525e-01 -9.97831464e-01 5.29962853e-02 4.90935296e-01
8.07344198e-01 -8.88065517e-01 -7.98768699e-01 3.16285580e-01
-5.68582654e-01 -7.41880834e-01 -5.30369401e-01 3.22499812e-01
-1.47628725e+00 -1.37376618e+00 -4.84327376e-01 -9.24682558e-01
7.11267710e-01 3.14296812e-01 9.52720106e-01 1.22690670e-01
-4.18555081e-01 3.19023073e-01 -1.85504228e-01 -6.12249970e-01
-1.66392550e-01 -9.64917541e-02 1.95851490e-01 -6.58558846e-01
1.13766432e+00 -1.00271678e+00 -1.26601052e+00 2.34588593e-01
-4.17966098e-01 -4.24520671e-01 9.18142736e-01 1.63518086e-01
5.76465964e-01 2.99173415e-01 6.10525846e-01 -3.35298806e-01
7.55074978e-01 -7.25165188e-01 5.00006089e-03 1.48596302e-01
-1.34462321e+00 -1.18172705e-01 2.91218013e-01 -7.62768984e-02
-6.95509911e-01 -1.19374236e-02 -2.28903428e-01 4.63214695e-01
-6.08469248e-01 2.47747213e-01 -3.45512956e-01 4.97665912e-01
4.46179599e-01 4.02588695e-01 2.98635960e-01 -7.28914857e-01
-3.25020254e-01 1.05389631e+00 1.09552152e-01 1.62688926e-01
2.03279674e-01 4.97122735e-01 -6.14082143e-02 -1.10093904e+00
-1.38382435e+00 -1.05672276e+00 -4.05149072e-01 1.09871872e-01
1.25606871e+00 -8.19140732e-01 -4.92819488e-01 5.98106571e-02
-6.16928518e-01 -1.65547431e-01 -1.16466612e-01 7.12491512e-01
-2.40439564e-01 3.74823779e-01 -1.53846443e-01 -5.67746997e-01
-8.91954184e-01 -8.33522856e-01 4.43241864e-01 4.24041808e-01
-8.10454369e-01 -1.10233188e+00 4.53978837e-01 6.92172050e-01
4.25934762e-01 -3.35153103e-01 1.50894165e+00 -7.70741940e-01
-1.51358515e-01 -1.15804821e-02 -1.35074005e-01 3.89685750e-01
7.89518774e-01 -6.24750197e-01 -2.53811240e-01 7.92461820e-03
6.32442653e-01 2.59136617e-01 2.11181879e-01 6.14783406e-01
5.49417436e-01 -5.19060194e-01 -5.79311013e-01 9.67324376e-02
1.30773365e+00 6.90750360e-01 8.21050584e-01 9.38286901e-01
-9.74422097e-02 -1.85389936e-01 2.64175296e-01 3.65797520e-01
8.36306930e-01 2.93054968e-01 9.08779204e-02 4.57623541e-01
-2.85598170e-03 3.64424109e-01 4.36844468e-01 5.46029568e-01
-1.41772434e-01 8.85699764e-02 -9.29782152e-01 4.74530101e-01
-1.54617476e+00 -7.84728289e-01 -1.62624806e-01 1.90849221e+00
7.87296832e-01 2.59805292e-01 3.64860266e-01 2.17847556e-01
2.17864618e-01 -4.42828000e-01 -5.23231745e-01 -1.80676416e-01
3.03993732e-01 4.91434723e-01 1.70033351e-01 -6.55189380e-02
-6.52813435e-01 2.82960176e-01 7.05771494e+00 -2.39104450e-01
-6.54561937e-01 3.86616439e-01 1.25683192e-02 -5.26725531e-01
2.38726884e-01 -1.67942882e-01 -8.52871656e-01 8.75770986e-01
1.44036508e+00 -2.79535979e-01 2.84552425e-01 1.03164089e+00
1.05049789e+00 -5.14203250e-01 -1.07300043e+00 7.94889450e-01
1.93056926e-01 -8.73483121e-01 -3.65985364e-01 1.91420838e-01
7.32116550e-02 -2.59599816e-02 -4.54880804e-01 1.80505589e-01
-1.90970927e-01 -6.93528771e-01 2.74756819e-01 8.09530735e-01
4.12504196e-01 -1.51895463e-01 9.40813720e-01 7.76353180e-02
-9.94431555e-01 -4.82356429e-01 -1.35134220e-01 -3.62485498e-01
3.64790529e-01 7.72762120e-01 -1.14414740e+00 -1.39859989e-01
1.13128877e+00 8.82317960e-01 -9.14231479e-01 1.59519196e+00
-3.67883861e-01 8.76365304e-01 -2.61981606e-01 -2.29969114e-01
-2.29846105e-01 -1.61246374e-01 3.52364093e-01 5.30664146e-01
6.27924085e-01 3.20764810e-01 -9.56557468e-02 6.05435967e-01
4.13705438e-01 -4.50610258e-02 -1.67294458e-01 -1.49981260e-01
4.43028331e-01 1.02186549e+00 -8.59042943e-01 -1.12998448e-01
-8.33956659e-01 6.37638032e-01 -2.02639863e-01 -3.08111440e-02
-2.97594130e-01 1.59800097e-01 7.25740612e-01 7.57504523e-01
2.20114663e-02 -4.13800955e-01 -3.62906724e-01 -8.17057073e-01
5.74099086e-02 -6.04898870e-01 7.94071436e-01 -8.95047605e-01
-1.46344483e+00 2.04437748e-01 -1.38890535e-01 -1.02948022e+00
1.32403404e-01 -4.01348561e-01 -7.02010870e-01 5.60692906e-01
-1.23460519e+00 -5.76709092e-01 -4.92911339e-01 6.84410512e-01
6.56357110e-01 -3.08711618e-01 1.20804715e+00 5.16762674e-01
-6.08197868e-01 -5.44304401e-02 -1.57230720e-02 -2.17942461e-01
8.14199090e-01 -1.41590416e+00 -2.78010160e-01 4.17969704e-01
-6.39023960e-01 1.07041633e+00 6.96096539e-01 -9.45867717e-01
-7.72450328e-01 -1.02291536e+00 1.63101184e+00 -3.28380644e-01
6.38393879e-01 1.15773484e-01 -5.04570901e-01 6.18545592e-01
-5.27741425e-02 -6.77957416e-01 1.78366828e+00 1.72119081e-01
3.66460323e-01 -4.17256594e-01 -1.24472785e+00 4.66042489e-01
9.11952794e-01 -3.25805962e-01 -1.42724288e+00 4.52008367e-01
4.04036433e-01 3.39435220e-01 -8.76921594e-01 -1.51214615e-01
5.68206429e-01 -9.05856669e-01 1.09848964e+00 -4.83994305e-01
1.57948583e-01 -1.89303055e-01 -1.06832631e-01 -1.07726979e+00
-4.38920110e-01 -1.97044998e-01 -1.41231120e-01 8.30748498e-01
3.60911578e-01 -6.84901297e-01 7.09812939e-01 1.09720385e+00
-4.36804354e-01 -6.79784298e-01 -9.28481519e-01 -8.40784311e-01
-3.68169785e-01 -6.74767271e-02 3.76982421e-01 2.43361950e-01
2.22875178e-01 6.86534107e-01 2.31333449e-01 9.06474143e-02
4.96613115e-01 -1.00790657e-01 2.14272011e-02 -1.80478156e+00
2.28049397e-01 -3.69739137e-03 -5.22506177e-01 -3.45817119e-01
-1.40216216e-01 -7.75081813e-01 -2.58200347e-01 -2.55507207e+00
6.61070585e-01 -1.62951559e-01 -3.61049205e-01 5.70535481e-01
-5.12840971e-02 -2.28113294e-01 -3.13971698e-01 4.62256312e-01
-5.89077115e-01 3.15814465e-01 9.28686738e-01 1.82659794e-02
-6.59701705e-01 5.72981954e-01 -1.12758005e+00 7.93951333e-01
1.25403655e+00 -6.35248005e-01 -7.49897838e-01 4.01924411e-03
3.52295071e-01 -3.06293994e-01 3.36028934e-01 -1.17940545e+00
2.85143763e-01 -1.60249293e-01 5.44419646e-01 -7.05498874e-01
4.61514980e-01 -9.45229769e-01 -9.27853398e-03 7.66078472e-01
1.43833756e-01 1.78820997e-01 -1.07621988e-02 5.41080475e-01
2.81320184e-01 -2.11569265e-01 6.25468373e-01 -4.60930288e-01
-4.43359375e-01 4.96049337e-02 -1.00872982e+00 1.65315256e-01
8.19382966e-01 -1.93838879e-01 -4.18865323e-01 -1.69604331e-01
-1.27230203e+00 2.02848107e-01 2.41433844e-01 2.62272537e-01
7.27454722e-01 -9.66907740e-01 8.13874230e-03 1.03661910e-01
1.24864750e-01 -2.44303957e-01 8.23549405e-02 1.17020047e+00
-3.99265230e-01 7.46783853e-01 -4.34324563e-01 -1.33341163e-01
-1.55515766e+00 5.38372517e-01 1.81220308e-01 -2.47802719e-01
-1.01839352e+00 1.19454868e-01 -5.26519716e-01 3.34553719e-01
3.37249905e-01 -7.75306940e-01 -4.65127289e-01 -7.67070577e-02
9.04866695e-01 8.46411467e-01 1.50404781e-01 -8.83424357e-02
-7.96738982e-01 2.47679338e-01 -4.91199754e-02 1.18373871e-01
1.63653445e+00 -4.87532318e-01 -5.39892137e-01 5.74473441e-01
8.34079444e-01 -2.45756730e-01 -3.46681237e-01 2.50907212e-01
2.13756800e-01 -4.23541181e-02 1.64509803e-01 -9.22017097e-01
-5.17390311e-01 4.51450884e-01 1.22261000e+00 3.60090584e-01
1.18756866e+00 2.93261260e-01 1.32385862e+00 6.75249398e-01
1.94291532e-01 -9.70855713e-01 -8.58143046e-02 2.14579906e-02
5.90081036e-01 -1.08029163e+00 3.38567406e-01 -8.09959546e-02
-1.00548372e-01 1.05531323e+00 2.78009087e-01 1.66376531e-01
1.30106449e+00 8.56487174e-03 -7.45972572e-03 -3.08247209e-01
-5.83101869e-01 -2.68913478e-01 1.93450153e-01 5.81286550e-01
2.87001461e-01 2.53541261e-01 -7.59758055e-01 1.21407509e+00
-5.08853018e-01 6.28473341e-01 3.70922148e-01 1.30223966e+00
-9.23166633e-01 -1.30036497e+00 -2.11101517e-01 1.31561518e+00
-5.96217811e-01 -2.67053008e-01 -4.45946932e-01 5.38167119e-01
4.25747782e-01 1.33437037e+00 5.42089380e-02 1.48426116e-01
3.96490484e-01 7.17058182e-01 2.14160204e-01 -1.05189681e+00
-9.70628560e-02 -1.94392070e-01 2.06449062e-01 -4.31095630e-01
-7.46773243e-01 -9.90017533e-01 -1.56990647e+00 -3.15263309e-02
2.11954325e-01 3.92343491e-01 3.50181401e-01 1.44140267e+00
6.19797945e-01 4.14677888e-01 -1.48561954e-01 1.26968876e-01
5.14608510e-02 -9.54733729e-01 -7.96648383e-01 -8.28477144e-02
4.81819779e-01 -8.35070550e-01 -5.91370344e-01 3.19567174e-01] | [13.600671768188477, 3.400681734085083] |
3a7d5b26-7f41-4f6b-a55b-abc48d9027db | evaluating-and-modeling-attribution-for-cross | 2305.14332 | null | https://arxiv.org/abs/2305.14332v1 | https://arxiv.org/pdf/2305.14332v1.pdf | Evaluating and Modeling Attribution for Cross-Lingual Question Answering | Trustworthy answer content is abundant in many high-resource languages and is instantly accessible through question answering systems, yet this content can be hard to access for those that do not speak these languages. The leap forward in cross-lingual modeling quality offered by generative language models offers much promise, yet their raw generations often fall short in factuality. To improve trustworthiness in these systems, a promising direction is to attribute the answer to a retrieved source, possibly in a content-rich language different from the query. Our work is the first to study attribution for cross-lingual question answering. First, we collect data in 5 languages to assess the attribution level of a state-of-the-art cross-lingual QA system. To our surprise, we find that a substantial portion of the answers is not attributable to any retrieved passages (up to 50% of answers exactly matching a gold reference) despite the system being able to attend directly to the retrieved text. Second, to address this poor attribution level, we experiment with a wide range of attribution detection techniques. We find that Natural Language Inference models and PaLM 2 fine-tuned on a very small amount of attribution data can accurately detect attribution. Based on these models, we improve the attribution level of a cross-lingual question-answering system. Overall, we show that current academic generative cross-lingual QA systems have substantial shortcomings in attribution and we build tooling to mitigate these issues. | ['Xinyi Wang', 'Jonathan Herzig', 'Roee Aharoni', 'Livio Baldini Soares', 'Sebastian Ruder', 'Tom Kwiatkowski', 'Jonathan H. Clark', 'John Wieting', 'Benjamin Muller'] | 2023-05-23 | null | null | null | null | ['cross-lingual-question-answering'] | ['natural-language-processing'] | [-3.43263894e-01 2.86028773e-01 -1.62060916e-01 -3.50116402e-01
-1.96621108e+00 -9.07083690e-01 7.55018651e-01 2.14030683e-01
-2.19250619e-01 8.34020019e-01 2.67551363e-01 -4.91555065e-01
1.22878011e-02 -7.15339601e-01 -7.24661231e-01 -1.38062403e-01
7.54386246e-01 1.08508694e+00 1.36791036e-01 -7.56183445e-01
1.50646880e-01 -8.20844769e-02 -1.10786104e+00 3.95914972e-01
1.23114622e+00 6.18066967e-01 -3.47270310e-01 6.14744961e-01
-4.71094698e-01 1.20910037e+00 -9.43273544e-01 -1.31306350e+00
-2.07734764e-01 -4.72357750e-01 -1.28470945e+00 -4.49985951e-01
1.04954505e+00 2.70063430e-02 6.53291270e-02 8.80480468e-01
5.18247783e-01 -3.46695065e-01 8.63723695e-01 -1.37615371e+00
-1.39385235e+00 7.21241474e-01 -1.40850125e-02 1.45070031e-01
7.16485023e-01 1.36325225e-01 1.38409054e+00 -8.77987385e-01
6.79702759e-01 1.41104150e+00 8.22088659e-01 5.12557209e-01
-1.21255541e+00 -6.93005145e-01 -4.99768764e-01 1.78552076e-01
-1.43491232e+00 -6.83629036e-01 5.03904581e-01 -4.67266291e-01
6.27640843e-01 3.23640376e-01 -1.50428578e-01 1.31735575e+00
2.55387872e-01 5.45287549e-01 1.55499172e+00 -6.30281985e-01
-1.42912462e-01 7.69188583e-01 1.73205793e-01 8.38430345e-01
-3.30312476e-02 -3.30484807e-01 -8.32940817e-01 -4.04015750e-01
-9.84296575e-02 -6.31234348e-01 -3.59731793e-01 1.49628729e-01
-9.30581868e-01 1.16228807e+00 3.86856794e-01 3.08798581e-01
-1.27487645e-01 1.01389885e-01 2.06936151e-01 6.37327254e-01
6.28317475e-01 8.76197040e-01 -5.61273336e-01 -1.29787102e-01
-1.14097393e+00 4.43099409e-01 1.28263783e+00 9.25149977e-01
7.60546625e-01 -2.56152719e-01 -3.84006113e-01 8.11641574e-01
5.17204523e-01 9.26716447e-01 4.26932424e-01 -1.25249052e+00
5.28740346e-01 5.81685543e-01 1.89713776e-01 -1.19601154e+00
2.27321744e-01 -6.12140834e-01 -3.75277132e-01 -3.75301093e-02
8.67237508e-01 2.38196671e-01 -2.98259705e-01 1.61787558e+00
1.80361435e-01 -7.67601013e-01 8.72186795e-02 7.55081534e-01
1.01623178e+00 3.25550228e-01 2.49527708e-01 1.30710855e-01
1.77265429e+00 -8.08718503e-01 -8.04329336e-01 -2.80651152e-01
7.66199827e-01 -1.34889925e+00 1.63398659e+00 1.35630190e-01
-1.01335335e+00 -5.54540217e-01 -8.13347161e-01 -4.27276403e-01
-5.98898709e-01 -6.14750497e-02 3.20615798e-01 1.00292265e+00
-1.04669452e+00 2.85587102e-01 -2.37209648e-01 -2.20419869e-01
2.68070221e-01 -2.49459952e-01 -2.57977337e-01 -2.77523667e-01
-1.67913866e+00 1.37912273e+00 -2.21613050e-01 -3.98218811e-01
-6.36204123e-01 -7.98232257e-01 -7.46940553e-01 -1.53890699e-01
2.70344704e-01 -7.45315552e-01 1.39271557e+00 -8.45103621e-01
-1.03025091e+00 1.18229103e+00 -6.49350822e-01 -2.68049061e-01
5.73462605e-01 -1.24751940e-01 -5.59235215e-01 6.58432543e-02
8.12439978e-01 3.61349821e-01 8.06525171e-01 -1.43553853e+00
-2.42256090e-01 -3.37125242e-01 2.54515689e-02 -6.04538135e-02
-3.38987499e-01 3.32100004e-01 -1.80639118e-01 -3.55919421e-01
-1.56723514e-01 -6.90482140e-01 4.73664463e-01 -2.37101868e-01
-3.27370912e-01 -5.51998019e-01 6.11509681e-01 -1.01264000e+00
1.18120658e+00 -1.48537338e+00 -3.30244929e-01 -6.56190142e-02
3.20428520e-01 5.34790009e-02 1.67402811e-02 6.05221510e-01
3.59389752e-01 6.94988549e-01 -1.67357065e-02 -6.54013276e-01
5.89437842e-01 2.54947931e-01 -8.07274938e-01 3.66783679e-01
2.88991481e-02 1.25954604e+00 -1.02780712e+00 -9.42772567e-01
-4.39041048e-01 3.78814220e-01 -3.04347932e-01 3.75437975e-01
-2.88811326e-01 8.80372897e-02 -2.52041906e-01 8.49245131e-01
1.47671208e-01 -4.16445464e-01 -3.75251472e-01 -1.42376320e-02
2.11150035e-01 8.94381702e-01 -5.66182792e-01 1.43243921e+00
-6.91608369e-01 7.45806277e-01 -1.31850779e-01 -2.54271746e-01
9.21976447e-01 6.15082026e-01 -5.10758236e-02 -9.57043529e-01
-1.20369621e-01 7.16352701e-01 -1.95784092e-01 -4.86881286e-01
8.20769191e-01 -4.16125655e-01 -3.12564105e-01 8.96224082e-01
2.53236800e-01 -5.60437620e-01 1.38015881e-01 6.46539450e-01
8.08466792e-01 6.37679687e-03 -9.67648253e-02 -2.56907076e-01
5.71882725e-01 4.75133598e-01 -1.16468668e-02 9.94011879e-01
-2.23212048e-01 4.44574237e-01 3.02210122e-01 1.80823915e-02
-1.01438975e+00 -9.66393769e-01 -2.44817108e-01 1.19823658e+00
-1.51401058e-01 -5.52486300e-01 -8.05393279e-01 -7.25092232e-01
-5.89274913e-02 1.28700340e+00 -5.41191757e-01 -1.48743123e-01
-3.14383119e-01 -3.59563768e-01 1.19363809e+00 2.05615461e-01
2.87970394e-01 -9.05029714e-01 -1.07847139e-01 2.25264117e-01
-7.98067272e-01 -1.08589911e+00 -4.66556787e-01 -2.75111109e-01
-3.41567665e-01 -8.88899028e-01 -4.03628796e-01 -3.39962065e-01
1.22384444e-01 -4.13833372e-02 2.10482121e+00 2.87907034e-01
2.47518331e-01 5.65893829e-01 -2.26587862e-01 -4.45655167e-01
-1.20920241e+00 3.04836571e-01 -2.37036854e-01 -4.46488529e-01
7.64169514e-01 -1.06602743e-01 -3.05467308e-01 4.31894779e-01
-5.58438540e-01 -5.26416302e-01 2.47629687e-01 7.58239269e-01
3.48092973e-01 -3.42796028e-01 8.10170472e-01 -1.05079401e+00
1.16815543e+00 -6.51311159e-01 -2.46312886e-01 5.55315971e-01
-9.70519483e-01 1.70388490e-01 6.56179607e-01 -1.13718584e-01
-8.92978191e-01 -8.07107687e-01 -3.15486729e-01 -1.30812168e-01
-3.79506424e-02 5.85795641e-01 7.05648260e-03 -8.02521929e-02
9.13529873e-01 -2.80629210e-02 1.28838466e-02 -4.20339704e-01
7.24619627e-01 8.34458828e-01 5.52052498e-01 -8.90608251e-01
9.73167360e-01 1.44438714e-01 -5.49953938e-01 -3.58177155e-01
-1.39362431e+00 -4.53884721e-01 -3.39170158e-01 -3.38979721e-01
7.21634865e-01 -8.15001011e-01 -8.16990435e-01 -9.01770964e-02
-1.30105078e+00 -1.39211237e-01 -1.45927012e-01 -7.06351772e-02
-3.21061432e-01 2.63622344e-01 -6.57784760e-01 -7.90747821e-01
-5.47341943e-01 -1.00322342e+00 1.17227721e+00 -1.27866223e-01
-8.62183034e-01 -1.24967730e+00 2.98200935e-01 1.31869268e+00
7.44962096e-01 -1.42364740e-01 1.04365242e+00 -8.46791565e-01
-5.88874459e-01 -4.76244777e-01 -1.43832460e-01 3.59674245e-01
-7.87605867e-02 2.74484586e-02 -1.02291107e+00 -2.96793673e-02
1.61601231e-01 -9.39110041e-01 4.18503910e-01 -2.60163367e-01
5.85047483e-01 -5.22643447e-01 1.03946768e-01 -7.79133588e-02
1.21538770e+00 -6.37323678e-01 4.31839496e-01 3.32324922e-01
6.22432828e-01 7.71830499e-01 3.03411931e-01 -3.36395264e-01
1.09928107e+00 8.28248620e-01 1.07117027e-01 -4.85249721e-02
-5.66799343e-01 -6.01713657e-01 3.84179443e-01 1.19530952e+00
1.73421741e-01 -3.22124690e-01 -9.89236236e-01 7.24226773e-01
-1.48576009e+00 -1.18856108e+00 -6.11855149e-01 2.13793302e+00
1.29970264e+00 9.86197293e-02 -1.83920845e-01 -2.29845271e-01
2.13314861e-01 -1.58537492e-01 2.02920549e-02 -5.90980172e-01
-4.35579866e-01 4.31381226e-01 2.96026379e-01 9.23279405e-01
-6.07414365e-01 1.01069653e+00 6.87070370e+00 1.00637102e+00
-6.39054656e-01 6.19541705e-01 4.31751668e-01 4.79270756e-01
-8.91401470e-01 2.23657817e-01 -8.86771321e-01 4.87325609e-01
1.39056516e+00 -2.21970275e-01 2.43858993e-01 6.56115830e-01
-9.53082666e-02 4.17010002e-02 -1.03045690e+00 4.95094538e-01
5.83015800e-01 -1.19936168e+00 1.12465583e-01 -1.42875880e-01
6.55001044e-01 -5.65518178e-02 1.23343393e-01 6.65542781e-01
6.95872962e-01 -1.47247529e+00 1.00445843e+00 6.05091214e-01
5.54853082e-01 -5.81055343e-01 9.76180375e-01 5.40551424e-01
-5.66016734e-01 5.37329674e-01 -2.81858742e-01 7.87714422e-02
3.07277381e-01 4.64240074e-01 -9.35545623e-01 3.79778385e-01
6.48244202e-01 8.58225748e-02 -1.01036561e+00 4.68552768e-01
-6.85370922e-01 9.55918252e-01 -8.55370611e-02 -2.41126984e-01
1.63921967e-01 1.06367096e-01 3.46918374e-01 1.18983209e+00
1.20523170e-01 -1.68310374e-01 -9.65975076e-02 1.31038511e+00
-6.12493634e-01 1.17211901e-01 -5.28891087e-01 -8.22401885e-03
6.72339320e-01 1.11943686e+00 9.05315429e-02 -5.43345392e-01
-3.63598496e-01 8.79784048e-01 6.38592422e-01 2.18491375e-01
-7.12298810e-01 -7.76366740e-02 2.74138927e-01 5.55433072e-02
-3.18752974e-01 -5.91961071e-02 -4.50857729e-01 -1.22291815e+00
1.45723596e-01 -1.48645937e+00 4.83364373e-01 -1.22151947e+00
-1.86562848e+00 6.20878279e-01 -3.98505986e-01 -6.27557039e-01
-6.67092502e-01 -3.25606108e-01 -1.33070439e-01 1.49258387e+00
-1.70492435e+00 -1.39710999e+00 -1.87536791e-01 6.99639976e-01
3.01007032e-01 -1.07848234e-01 1.16054833e+00 4.80475515e-01
6.67848065e-02 9.18906152e-01 -1.66956410e-01 2.26546749e-01
1.41804290e+00 -1.41400754e+00 1.74463019e-01 8.07692468e-01
6.89782143e-01 1.12283516e+00 9.75851119e-01 -5.66651404e-01
-1.27221334e+00 -8.11516225e-01 1.90892386e+00 -1.66459620e+00
1.04895234e+00 -2.57585645e-01 -1.32740176e+00 6.59704626e-01
6.88222229e-01 -4.45052952e-01 8.94842267e-01 4.28863615e-01
-9.57437575e-01 1.42133191e-01 -1.07118154e+00 5.38300693e-01
3.10046315e-01 -1.22484255e+00 -1.12793863e+00 7.04734802e-01
7.85734296e-01 -3.54072541e-01 -1.06886053e+00 -2.51240446e-03
3.14844996e-01 -7.60674536e-01 1.00189829e+00 -7.46888697e-01
8.22483182e-01 -3.84702921e-01 -2.64010549e-01 -1.25543976e+00
-9.73467082e-02 -4.89147812e-01 5.56375608e-02 1.72345114e+00
8.55002105e-01 -7.01975763e-01 4.49742287e-01 1.05904686e+00
-9.44198593e-02 -3.45645934e-01 -9.63539541e-01 -6.70915782e-01
6.89468443e-01 -5.90559185e-01 5.36666036e-01 1.13835859e+00
-1.66947901e-01 7.70068109e-01 -1.45855218e-01 8.69386122e-02
8.85466993e-01 3.81727785e-01 8.39282572e-01 -9.02610600e-01
-2.25161940e-01 -3.28725725e-01 1.95075244e-01 -6.76208794e-01
4.14788216e-01 -1.02167833e+00 2.56210081e-02 -1.53561795e+00
2.08828956e-01 -7.05972612e-01 2.73379594e-01 4.89141673e-01
-4.89002436e-01 6.03637516e-01 -5.47869690e-02 5.94004452e-01
-5.53563535e-01 4.30047423e-01 1.14552867e+00 -2.27977663e-01
5.06556690e-01 -2.33013138e-01 -1.05320489e+00 5.24153650e-01
5.42752206e-01 -8.23001742e-01 -2.33238503e-01 -6.78299725e-01
5.78536808e-01 -1.28037520e-02 7.16627657e-01 -5.27209044e-01
4.63966966e-01 1.63175330e-01 1.61007163e-03 -3.61744374e-01
2.65141308e-01 -5.55968285e-01 -6.41503707e-02 7.09461793e-02
-6.04829431e-01 2.02047959e-01 -1.10989176e-01 2.14499757e-01
-4.24701512e-01 -3.65513474e-01 4.57358807e-01 -2.44456157e-01
-3.01407635e-01 -1.03224225e-01 -2.66126305e-01 8.87521684e-01
1.75337493e-01 1.72905773e-01 -9.50862110e-01 -6.96381629e-01
-1.13844663e-01 1.36172950e-01 5.91575861e-01 4.21066701e-01
2.36839890e-01 -1.33420098e+00 -1.14010894e+00 -3.90772909e-01
4.74828988e-01 -7.57516921e-01 -8.75317827e-02 7.53768981e-01
-2.25801274e-01 7.38120675e-01 2.14320943e-01 -3.85503799e-01
-8.72367024e-01 2.92406410e-01 3.39877397e-01 -2.79768288e-01
-4.85459436e-03 9.83225107e-01 -4.99566913e-01 -8.42266381e-01
-1.86993137e-01 2.32107267e-01 -1.04896218e-01 1.52243137e-01
4.26983476e-01 1.27819628e-01 3.94461393e-01 -9.35159445e-01
-3.42805922e-01 2.37121835e-01 1.76126897e-01 -5.24795532e-01
7.22602308e-01 -3.10103953e-01 -5.08455336e-01 5.43925703e-01
1.18818879e+00 5.72193980e-01 -4.30934876e-01 -3.21852595e-01
-4.06559184e-02 -5.56759834e-01 3.80937494e-02 -1.18794763e+00
-6.13537133e-01 9.36368227e-01 1.28048256e-01 4.49104905e-01
4.72585946e-01 4.17120785e-01 9.01200950e-01 2.09125981e-01
4.94167805e-01 -8.96125257e-01 9.77920964e-02 5.42068362e-01
9.76575196e-01 -1.57606995e+00 -1.87806740e-01 -4.57484350e-02
-6.78024530e-01 9.22023296e-01 4.65438992e-01 3.24553251e-01
1.25089750e-01 -1.15503304e-01 6.17194295e-01 -6.50919437e-01
-7.71248817e-01 -5.69644384e-02 6.67787850e-01 4.18890536e-01
1.00526333e+00 1.98651955e-01 -1.54465735e-01 5.17761052e-01
-9.57689106e-01 -4.69506681e-01 4.94750559e-01 3.05975139e-01
-2.29668647e-01 -1.16832292e+00 -5.30698478e-01 9.60599259e-02
-7.79445529e-01 -5.75092018e-01 -7.11512983e-01 7.76710510e-01
-1.22348554e-01 1.54713917e+00 -4.35092449e-01 -4.31274809e-02
3.54550153e-01 4.56286132e-01 2.53681988e-01 -5.23283839e-01
-9.41390812e-01 -5.29968619e-01 4.07800138e-01 -4.72137809e-01
-3.99972111e-01 -5.79956770e-01 -6.98329091e-01 -6.92123771e-01
-3.25103760e-01 5.78370273e-01 6.44027889e-01 1.21361184e+00
1.66442707e-01 2.33518243e-01 1.16600372e-01 2.04244196e-01
-7.71485388e-01 -9.80941176e-01 -7.23821521e-02 5.60648322e-01
2.37734616e-01 -3.15194935e-01 -5.82476377e-01 2.52778709e-01] | [11.292728424072266, 8.192617416381836] |
f9dbb25a-8bc7-4062-9ca0-8728052d1827 | safeamc-adversarial-training-for-robust | 2105.13746 | null | https://arxiv.org/abs/2105.13746v1 | https://arxiv.org/pdf/2105.13746v1.pdf | SafeAMC: Adversarial training for robust modulation recognition models | In communication systems, there are many tasks, like modulation recognition, which rely on Deep Neural Networks (DNNs) models. However, these models have been shown to be susceptible to adversarial perturbations, namely imperceptible additive noise crafted to induce misclassification. This raises questions about the security but also the general trust in model predictions. We propose to use adversarial training, which consists of fine-tuning the model with adversarial perturbations, to increase the robustness of automatic modulation recognition (AMC) models. We show that current state-of-the-art models benefit from adversarial training, which mitigates the robustness issues for some families of modulations. We use adversarial perturbations to visualize the features learned, and we found that in robust models the signal symbols are shifted towards the nearest classes in constellation space, like maximum likelihood methods. This confirms that robust models not only are more secure, but also more interpretable, building their decisions on signal statistics that are relevant to modulation recognition. | ['Pascal Frossard', 'Gérôme Bovet', 'Javier Maroto'] | 2021-05-28 | null | null | null | null | ['automatic-modulation-recognition'] | ['time-series'] | [ 8.85576010e-01 4.28162873e-01 2.33877480e-01 -3.20276231e-01
-7.85750270e-01 -9.33996081e-01 8.96415114e-01 -5.29229164e-01
-6.66900277e-02 8.49449098e-01 1.10491402e-01 -8.03146064e-01
-2.71385312e-01 -7.43236899e-01 -9.37690496e-01 -8.72345328e-01
-5.54178894e-01 -7.56832436e-02 -2.06260070e-01 -3.80405396e-01
1.56266674e-01 7.70185888e-01 -1.12053919e+00 5.52931309e-01
4.24953848e-01 9.54597712e-01 -6.50802910e-01 1.23225868e+00
5.91069639e-01 7.68949926e-01 -1.27320588e+00 -6.60075665e-01
4.14062530e-01 -4.40963268e-01 -5.16404688e-01 -2.34038964e-01
3.14913481e-01 -3.79273146e-01 -7.57871866e-01 1.25613523e+00
4.96827543e-01 -3.93419325e-01 8.02844524e-01 -1.36780703e+00
-8.16972017e-01 5.68871975e-01 7.25382045e-02 -6.46183714e-02
1.43235937e-01 4.32551920e-01 6.37189329e-01 -3.94621491e-01
3.73355895e-01 1.64965034e+00 7.79550374e-01 7.01265097e-01
-1.31142974e+00 -9.31457341e-01 -3.42060119e-01 3.11831266e-01
-1.44090879e+00 -9.02784586e-01 5.84806740e-01 -2.64774829e-01
5.25374174e-01 7.61331379e-01 -1.54307927e-03 1.62691927e+00
4.39836144e-01 2.93280959e-01 1.28263915e+00 -6.57080293e-01
2.36724555e-01 3.55438501e-01 -6.24830663e-01 3.14608514e-01
4.19990979e-02 8.55768323e-01 -2.68872619e-01 -4.44423854e-01
5.07384658e-01 -5.50400734e-01 -5.68218112e-01 -2.03359246e-01
-9.12046134e-01 9.01971102e-01 3.32185715e-01 2.36150339e-01
-7.57115409e-02 5.82496464e-01 1.00904085e-01 7.87775338e-01
1.30895391e-01 8.78484726e-01 -3.09329897e-01 2.45555975e-02
-6.11642122e-01 -6.14384981e-03 8.60655427e-01 6.47599399e-01
3.39777291e-01 5.39036989e-01 -3.03473175e-01 5.40283024e-01
4.24743801e-01 8.33787084e-01 2.41126686e-01 -1.07376063e+00
3.12121898e-01 -1.08743221e-01 -2.12411031e-01 -1.18734109e+00
-3.62741590e-01 -7.94820487e-01 -1.08795071e+00 6.59434319e-01
3.81269664e-01 -3.28545660e-01 -1.01875424e+00 1.88148880e+00
-3.08038533e-01 1.04523592e-01 4.08688754e-01 5.86101711e-01
1.86410189e-01 5.09243965e-01 -1.23007745e-01 1.36997730e-01
8.54377508e-01 -2.49073103e-01 -5.55776775e-01 -7.00435862e-02
6.21544361e-01 -8.26225638e-01 6.31561875e-01 4.18705016e-01
-7.42154360e-01 -4.81095403e-01 -1.73270571e+00 4.75699157e-01
-4.51291919e-01 -3.08327585e-01 5.04064143e-01 1.71771789e+00
-9.83085573e-01 9.45617735e-01 -3.69177639e-01 5.76916598e-02
7.48833060e-01 5.74498296e-01 -4.33585942e-01 7.34994709e-02
-1.65525854e+00 1.05896270e+00 3.02875489e-01 1.07692644e-01
-1.15125704e+00 -3.41385275e-01 -7.12979555e-01 8.44573602e-02
-1.32972106e-01 -2.89963812e-01 1.10462904e+00 -1.21086347e+00
-1.47254813e+00 5.38481474e-01 2.61535615e-01 -1.08294773e+00
4.68770862e-01 2.96809852e-01 -1.20498729e+00 1.70207649e-01
-6.78196311e-01 5.46228290e-01 1.53855264e+00 -1.45530474e+00
-2.88911998e-01 1.00355841e-01 2.45223477e-01 -3.94669563e-01
-2.73316473e-01 2.85490099e-02 2.01127514e-01 -9.62923110e-01
-2.15075552e-01 -9.86179709e-01 -2.01752186e-01 4.49263491e-02
-5.95131695e-01 6.07090652e-01 1.01738274e+00 -6.85503900e-01
1.10704982e+00 -2.26507807e+00 -3.64688009e-01 8.28495383e-01
-3.60013284e-02 6.51152968e-01 -2.40869284e-01 3.63621414e-01
-4.69484240e-01 5.98333001e-01 -2.67433733e-01 1.83584663e-04
4.00398016e-01 2.58887231e-01 -7.93665648e-01 7.30961442e-01
4.33743745e-01 9.33504760e-01 -4.61527467e-01 2.42583856e-01
1.49187431e-01 4.56021070e-01 -4.28748190e-01 -4.61610891e-02
-1.05272047e-01 5.30064583e-01 6.49017692e-02 4.99440223e-01
8.74213099e-01 8.40032026e-02 2.68089205e-01 4.64191586e-02
5.46886265e-01 2.56501108e-01 -9.27362859e-01 8.92610729e-01
-4.77628946e-01 1.22460806e+00 -4.07415554e-02 -1.09650564e+00
7.29566038e-01 4.53496993e-01 -2.09856153e-01 -6.61218762e-01
2.24199921e-01 1.73225030e-01 5.49439073e-01 -4.00426835e-02
2.67648488e-01 -1.90534964e-02 -1.28639594e-01 4.46286052e-01
3.48798595e-02 -2.31286451e-01 -4.69631791e-01 8.32872391e-02
1.07542872e+00 -3.96562964e-01 2.04836264e-01 2.50015315e-02
4.60526347e-01 -4.34304655e-01 2.35988826e-01 1.23703277e+00
-2.62673907e-02 6.27655387e-01 5.03232777e-01 -6.78228885e-02
-1.12230480e+00 -1.02024460e+00 -3.67127597e-01 5.63528955e-01
-1.66827887e-01 -2.08584309e-01 -7.95182168e-01 -6.16000950e-01
-2.62724282e-03 1.07316542e+00 -5.79471231e-01 -8.74393284e-01
-3.22435886e-01 -8.29929769e-01 1.44278300e+00 2.12209791e-01
2.84424812e-01 -7.23496854e-01 4.16087396e-02 1.24894962e-01
4.19821113e-01 -1.08533800e+00 2.48817564e-03 3.85311544e-01
-3.43625426e-01 -9.27935600e-01 -5.33204317e-01 -2.65628517e-01
6.54158652e-01 -3.32448304e-01 8.65202427e-01 1.98922500e-01
8.39893054e-03 3.41630995e-01 -3.18182379e-01 -5.30043364e-01
-1.45109820e+00 -2.47709110e-01 4.09535587e-01 1.47570282e-01
9.95324180e-03 -8.00192118e-01 -1.98651537e-01 4.78558391e-01
-1.06391072e+00 -3.81290913e-01 6.76762640e-01 8.41406047e-01
1.08998055e-02 3.61230671e-01 5.67403197e-01 -8.78017783e-01
6.98892772e-01 -3.26928616e-01 -3.56730521e-01 1.86188355e-01
-6.20232701e-01 1.37850568e-01 9.36806858e-01 -5.76089561e-01
-3.64620894e-01 -3.74147713e-01 -4.75757629e-01 -2.84954250e-01
-2.94196635e-01 1.26071591e-02 -3.61384869e-01 -9.22381699e-01
1.14118469e+00 1.57976225e-01 1.32295853e-02 -1.09188281e-01
4.50842977e-01 1.05704463e+00 6.14534914e-01 -2.88362950e-01
1.45085859e+00 3.12259763e-01 2.81326056e-01 -8.13020170e-01
-4.17764127e-01 5.62562346e-01 -2.43348703e-01 -7.60000497e-02
3.42438608e-01 -7.28547871e-01 -6.59168482e-01 4.69663858e-01
-1.02643919e+00 -2.33234286e-01 -2.71536022e-01 2.80957729e-01
-6.13456368e-01 4.29281175e-01 -4.05333042e-01 -8.85633588e-01
1.62589420e-02 -9.39154625e-01 6.93434060e-01 -7.74817690e-02
-3.45676601e-01 -9.12434995e-01 -4.42790270e-01 7.44884908e-02
7.72403955e-01 3.69650245e-01 1.35878873e+00 -9.79481220e-01
-6.04098618e-01 -4.25094306e-01 8.03498626e-02 8.49152923e-01
9.03680325e-02 -6.50420859e-02 -1.60298896e+00 -5.13149559e-01
6.89828992e-02 -9.84973311e-02 5.84704995e-01 5.80940135e-02
1.48767805e+00 -7.91154444e-01 -1.45421490e-01 9.49747920e-01
8.50186467e-01 4.27459061e-01 1.24483335e+00 2.43070051e-01
2.96184301e-01 3.37648690e-01 8.49253312e-02 -2.59005688e-02
-3.00682157e-01 7.88256824e-01 6.78372681e-01 -2.44067803e-01
-4.83263992e-02 -3.09978664e-01 6.20224833e-01 3.10938567e-01
3.55278760e-01 -5.18242657e-01 -5.94577670e-01 -1.03960391e-02
-1.33501017e+00 -1.06257868e+00 1.72849327e-01 2.17926836e+00
8.00294399e-01 5.66219151e-01 -2.86296010e-01 6.03348434e-01
5.31416059e-01 2.92439371e-01 -4.54822451e-01 -7.52143204e-01
-6.23638391e-01 5.53840399e-01 8.58713269e-01 5.66252649e-01
-1.33516669e+00 5.34097254e-01 7.49274731e+00 1.19749129e+00
-1.12949216e+00 -7.43603781e-02 8.54203463e-01 1.04479060e-01
-3.25218350e-01 -2.67941318e-02 -2.24572808e-01 6.29341662e-01
1.40172100e+00 2.64111552e-02 7.81280518e-01 4.28124070e-01
-3.53933834e-02 4.22118843e-01 -1.19363356e+00 7.07965195e-01
1.54308509e-02 -1.47608781e+00 2.18682691e-01 2.68430889e-01
6.34873390e-01 -4.60257202e-01 6.26223743e-01 3.86640429e-01
3.85252386e-01 -1.73280728e+00 5.20230651e-01 6.71536565e-01
8.46349001e-01 -9.39343214e-01 7.70441055e-01 2.19791070e-01
-3.55043769e-01 -7.77605325e-02 -4.33025539e-01 1.07083812e-01
-1.33848876e-01 4.27005738e-01 -1.11359704e+00 3.26768488e-01
2.69148260e-01 4.21340838e-02 -8.69874418e-01 6.97721899e-01
-3.68802071e-01 9.08898354e-01 -2.87918270e-01 1.05818078e-01
6.32331148e-02 2.28600800e-01 6.84810996e-01 1.10074878e+00
4.83821064e-01 -3.64640176e-01 -4.53715384e-01 8.35906029e-01
-1.76362708e-01 -5.79428315e-01 -8.28208745e-01 -2.54294634e-01
6.72130764e-01 7.63308585e-01 -5.68132102e-02 4.44515375e-03
1.54708475e-02 9.32744443e-01 -3.53179783e-01 4.22952980e-01
-6.52345836e-01 -5.83654523e-01 8.10576916e-01 -2.30350852e-01
3.09699029e-01 -1.74208239e-01 -2.18419597e-01 -6.18056774e-01
-6.46563396e-02 -1.59782636e+00 -2.91971546e-02 -6.58170640e-01
-1.25500011e+00 3.99166912e-01 -4.21327531e-01 -1.55974436e+00
-6.14793420e-01 -8.22541416e-01 -5.01100302e-01 9.34489727e-01
-1.25001431e+00 -8.98231208e-01 3.86532992e-01 5.52043557e-01
-2.00076476e-01 -5.29237270e-01 1.23348677e+00 1.24363735e-01
-1.20390937e-01 1.22095644e+00 5.29723465e-01 2.14174896e-01
5.96942306e-01 -1.07387364e+00 8.36813450e-01 9.84387100e-01
4.22352284e-01 7.13985324e-01 8.81379128e-01 -7.26435333e-02
-1.15983510e+00 -1.18753004e+00 3.95186126e-01 -5.77141285e-01
7.12806344e-01 -6.37799025e-01 -7.56861448e-01 4.77545559e-01
1.67802244e-01 -5.06382398e-02 9.41846848e-01 -2.17589006e-01
-5.34451008e-01 1.43090963e-01 -1.48073781e+00 7.17934012e-01
8.59464765e-01 -1.05121863e+00 -3.48637938e-01 4.02261674e-01
7.24497795e-01 -3.97929043e-01 -7.87586629e-01 3.53557110e-01
5.37219405e-01 -9.90140259e-01 1.19070268e+00 -8.51314187e-01
4.89075817e-02 -4.40578043e-01 -6.45292819e-01 -1.53431475e+00
-1.23364195e-01 -1.05671477e+00 -4.60130334e-01 1.05105460e+00
8.27395439e-01 -8.24340045e-01 5.19177377e-01 2.82226533e-01
2.14184180e-01 -3.23129565e-01 -1.19008851e+00 -1.01114833e+00
2.10114926e-01 -7.60889232e-01 8.15751016e-01 9.47095990e-01
-2.34775990e-01 -8.85550827e-02 -7.80184448e-01 8.06971312e-01
4.76865113e-01 -3.87798607e-01 8.66813004e-01 -9.32996094e-01
-5.91021597e-01 -2.68621415e-01 -1.01991439e+00 -9.26668048e-01
1.53510422e-01 -8.01353097e-01 -1.65560499e-01 -6.51898026e-01
-6.05627060e-01 -4.47770327e-01 -3.20468575e-01 3.13407630e-01
1.56583577e-01 5.78170896e-01 1.93442434e-01 -1.15130983e-01
-7.32841566e-02 3.64885628e-01 7.76742935e-01 -4.09650892e-01
3.47917020e-01 3.21410239e-01 -1.12162662e+00 7.24728286e-01
9.61658716e-01 -5.91132045e-01 -1.65931329e-01 -1.96872726e-02
1.40959233e-01 -3.02383810e-01 6.52380526e-01 -1.36359215e+00
-2.72784177e-02 1.77306890e-01 8.48518729e-01 6.97374567e-02
5.67286253e-01 -9.89683032e-01 2.91244626e-01 5.68778694e-01
-4.77428913e-01 -4.79367346e-01 2.98936218e-01 5.70252836e-01
4.01029289e-02 -3.22018206e-01 9.43201184e-01 4.04855371e-01
-3.98260742e-01 -3.06524709e-02 -6.53079391e-01 -1.37257949e-01
7.28385925e-01 -1.68819621e-01 -5.40058732e-01 -1.06171906e+00
-8.22304904e-01 -2.55202681e-01 4.19828802e-01 4.64691609e-01
3.71345073e-01 -1.43020129e+00 -6.12133622e-01 5.60550272e-01
1.98204461e-02 -6.73888087e-01 2.84962766e-02 4.34726745e-01
-3.32213670e-01 4.84873593e-01 9.74203367e-03 -4.99360442e-01
-1.15144050e+00 5.19256055e-01 7.25531876e-01 1.32525608e-01
-1.24258190e-01 7.17507839e-01 -2.93878675e-01 -4.00650710e-01
3.83738607e-01 -7.11245984e-02 1.21264502e-01 -4.30261433e-01
4.57978398e-01 2.05134209e-02 1.96752846e-01 -5.55373788e-01
-3.35097730e-01 3.02655876e-01 7.19357505e-02 -1.26858130e-01
8.31058621e-01 3.65090854e-02 1.84191331e-01 -1.31090581e-01
1.36869264e+00 1.89057529e-01 -1.08374131e+00 -1.96854696e-01
-7.09386989e-02 -6.39708281e-01 -6.87168539e-02 -1.02162349e+00
-7.35134602e-01 9.89729762e-01 1.03747118e+00 6.52775705e-01
1.10532796e+00 -4.81855869e-01 3.94422144e-01 6.76107407e-01
4.83253390e-01 -8.15059245e-01 -3.93942088e-01 5.37222266e-01
9.39114690e-01 -8.44394505e-01 -6.81297481e-02 -1.16011739e-01
-3.74252766e-01 1.09354973e+00 -9.70262364e-02 1.18332587e-01
6.00618899e-01 2.96537906e-01 2.14993075e-01 1.95956916e-01
-6.16094768e-01 1.61674708e-01 2.78387636e-01 1.21713912e+00
-1.07624449e-01 1.90094829e-01 2.76946485e-01 4.41440165e-01
-7.42571056e-01 -5.35784781e-01 6.27219558e-01 6.87353432e-01
-2.00470626e-01 -1.31740308e+00 -7.85610378e-01 4.40553695e-01
-5.95296800e-01 -2.28367150e-01 -7.35610723e-01 7.35397935e-01
1.51768446e-01 1.30392957e+00 -1.61163211e-01 -8.12233150e-01
1.40201658e-01 5.16239274e-03 4.54879761e-01 -2.25915819e-01
-3.35810184e-01 -6.40120625e-01 3.97006422e-01 -3.35607082e-01
-1.18228100e-01 -2.54154921e-01 -6.67280674e-01 -5.36765099e-01
-2.27855444e-01 -4.67649400e-02 6.55843973e-01 1.03155541e+00
4.80436116e-01 6.47358954e-01 1.15808642e+00 -7.84023941e-01
-1.04766488e+00 -9.39124703e-01 -5.93670487e-01 2.16499910e-01
7.70320117e-01 -3.59800309e-01 -7.50220180e-01 -1.51814133e-01] | [5.63763952255249, 7.7013936042785645] |
e07d7fdd-df14-4235-b6e9-13a020f32113 | mapa-project-ready-to-go-open-source-datasets | null | null | https://aclanthology.org/2022.legal-1.12 | https://aclanthology.org/2022.legal-1.12.pdf | MAPA Project: Ready-to-Go Open-Source Datasets and Deep Learning Technology to Remove Identifying Information from Text Documents | This paper presents the outcomes of the MAPA project, a set of annotated corpora for 24 languages of the European Union and an open-source customisable toolkit able to detect and substitute sensitive information in text documents from any domain, using state-of-the art, deep learning-based named entity recognition techniques. In the context of the project, the toolkit has been developed and tested on administrative, legal and medical documents, obtaining state-of-the-art results. As a result of the project, 24 dataset packages have been released and the de-identification toolkit is available as open source. | ['Pierre Zweigenbaum', 'Patrick Paroubek', 'Manuel Herranz', 'Cyril Grouin', 'Lucie Gianola', 'Aitor García Pablos', 'Montse Cuadros', 'Khalid Choukri', 'Victoria Arranz'] | null | null | null | null | legal-lrec-2022-6 | ['de-identification'] | ['natural-language-processing'] | [-2.62933791e-01 3.91797423e-01 -7.84061328e-02 -6.74564362e-01
-1.08257723e+00 -6.24965370e-01 9.63945210e-01 5.52284718e-01
-8.25066626e-01 7.92863190e-01 4.01979685e-01 -8.72131735e-02
-1.99125394e-01 -6.03521824e-01 -2.15787172e-01 -1.94331035e-01
1.83085520e-02 9.50309932e-01 3.96457553e-01 -2.16716975e-01
2.97295719e-01 7.72091210e-01 -1.33665621e+00 8.45495522e-01
3.28252375e-01 6.64318085e-01 -3.77385169e-01 3.09023887e-01
-4.55255151e-01 6.10856593e-01 -7.65478194e-01 -7.63508379e-01
2.55823791e-01 1.13822468e-01 -1.48927867e+00 -4.98953462e-01
3.12503070e-01 -6.32203817e-02 -1.37305543e-01 7.57491052e-01
1.00630236e+00 -7.92428926e-02 2.06006244e-01 -5.89468777e-01
-5.49316108e-01 7.34471321e-01 1.29034340e-01 1.92292124e-01
4.28619385e-01 -1.84322476e-01 7.84372568e-01 -8.46828759e-01
1.03116024e+00 1.00490773e+00 1.00903237e+00 5.72030485e-01
-9.55252349e-01 -7.10304618e-01 -6.95286810e-01 -8.93189982e-02
-1.55069923e+00 -9.13445294e-01 1.65765554e-01 -4.97973830e-01
1.60116625e+00 2.18298584e-01 -5.89590194e-03 1.18770230e+00
5.17078303e-02 7.25164890e-01 9.72315848e-01 -5.65976262e-01
3.97362620e-01 5.47817111e-01 2.72503644e-01 4.45681244e-01
4.97036040e-01 -5.90119287e-02 -4.41307157e-01 -5.61986268e-01
3.51074457e-01 -5.57403564e-01 2.07244620e-01 -2.62421250e-01
-1.00245988e+00 6.68072522e-01 -2.35534608e-01 1.11838531e+00
-5.67936480e-01 -4.39074963e-01 9.79898632e-01 3.97166051e-02
6.51629150e-01 6.23870850e-01 -8.14728260e-01 -4.48927581e-01
-1.04906559e+00 3.39662194e-01 1.13641453e+00 7.59237707e-01
5.07001221e-01 -2.23942459e-01 -2.26771906e-01 1.00326777e+00
5.22436500e-01 9.95092019e-02 8.11494231e-01 -5.18416822e-01
4.14005220e-01 1.06213653e+00 3.01926970e-01 -5.34991205e-01
-8.80363524e-01 -8.84327367e-02 -3.57902467e-01 5.94313331e-02
3.82189512e-01 -3.34155381e-01 -8.47245455e-01 1.43391025e+00
4.59654748e-01 -1.49140760e-01 4.81655091e-01 4.77721542e-01
1.28599751e+00 2.85329819e-01 3.64966452e-01 8.86197761e-03
1.33552432e+00 -2.99459934e-01 -1.00044727e+00 4.73640785e-02
8.42412233e-01 -7.00931609e-01 2.53677189e-01 2.85151571e-01
-1.10286999e+00 -4.17077303e-01 -8.45748901e-01 -2.67246097e-01
-1.20595503e+00 1.92790061e-01 6.44842029e-01 1.16993654e+00
-1.09202027e+00 2.14117676e-01 -7.89940953e-01 -8.53252590e-01
3.81338567e-01 3.16343158e-01 -8.46196592e-01 4.79150563e-01
-1.63142872e+00 1.09049213e+00 1.17692304e+00 -1.52666509e-01
-4.08081830e-01 -8.04045439e-01 -8.31067145e-01 -1.97746322e-01
4.34473790e-02 -1.82236120e-01 1.38545942e+00 -4.29394245e-01
-1.24452579e+00 1.87149298e+00 3.54961395e-01 -6.57050312e-01
8.20074618e-01 -2.55548686e-01 -1.18559039e+00 -1.47817001e-01
6.19806111e-01 1.36420622e-01 -4.14059348e-02 -6.57741725e-01
-9.46460724e-01 -6.26741946e-01 -3.78482699e-01 -3.09185624e-01
-1.74412623e-01 1.04317081e+00 -3.67531747e-01 -5.64724147e-01
-6.49600148e-01 -6.44116282e-01 -7.76474401e-02 -5.57373881e-01
-3.09391201e-01 -4.31824148e-01 5.10043025e-01 -1.18256795e+00
1.30776727e+00 -2.18642378e+00 -4.35155392e-01 4.98341471e-02
-4.46844958e-02 9.64060843e-01 2.57487804e-01 6.83572948e-01
-3.61665010e-01 1.47675961e-01 -3.23637545e-01 -1.20094910e-01
3.62891287e-01 1.27984926e-01 -2.42107548e-02 5.74046910e-01
1.04972824e-01 3.49420965e-01 -9.15909052e-01 -4.29273546e-01
3.19651842e-01 4.31872815e-01 3.07645649e-01 6.24409765e-02
2.48047382e-01 3.63259390e-02 -5.82391083e-01 2.70961672e-01
9.03672516e-01 2.88362414e-01 3.81906092e-01 1.89920172e-01
-5.26392758e-01 6.20216608e-01 -1.52733839e+00 1.90713048e+00
-2.84254849e-01 4.95932341e-01 2.12749541e-01 -7.81727135e-01
1.18175220e+00 7.63838947e-01 3.86522710e-01 -4.54953581e-01
8.43440816e-02 5.60099125e-01 -5.62470436e-01 -6.00040436e-01
6.71081126e-01 8.28938261e-02 -2.77951628e-01 3.40800554e-01
5.26959181e-01 4.19094712e-01 3.89405102e-01 -2.09949180e-01
1.29250813e+00 6.76057041e-02 9.37201262e-01 -2.57946044e-01
1.06351388e+00 -1.62283797e-02 6.51275158e-01 4.14825052e-01
-4.46885794e-01 3.08082700e-01 4.51781303e-02 -7.56151378e-01
-1.20713651e+00 -7.17460573e-01 -7.75151014e-01 1.14697838e+00
-9.01614547e-01 -2.56471604e-01 -9.61199224e-01 -8.54011655e-01
-1.01073653e-01 9.94862854e-01 -6.39007568e-01 3.49772543e-01
-4.74612206e-01 -7.06884086e-01 1.41964102e+00 4.60733652e-01
6.89639509e-01 -1.56786120e+00 -7.39256561e-01 5.62548757e-01
2.62521207e-01 -1.10894704e+00 1.37102380e-01 1.82941347e-01
-1.64053395e-01 -1.20176995e+00 -6.62047923e-01 -7.11828291e-01
3.85591527e-03 -7.69427776e-01 1.11009014e+00 -4.88899946e-01
-5.13908982e-01 3.80192876e-01 -1.25867650e-01 -6.40203238e-01
-8.52172494e-01 5.45124650e-01 -1.54585317e-01 -1.12117920e-02
9.74305868e-01 -1.25960246e-01 7.23378966e-03 -1.74991339e-02
-9.61499453e-01 -7.36241519e-01 4.48481619e-01 5.29624224e-01
2.79642075e-01 -3.33132029e-01 7.92962909e-01 -1.21524978e+00
8.25620055e-01 -4.72193778e-01 -7.70787776e-01 5.19260347e-01
-6.88983202e-01 3.38739082e-02 2.59414136e-01 3.41103524e-01
-1.52904868e+00 3.18330646e-01 -7.28273690e-01 4.46275622e-01
-1.09226739e+00 5.25841892e-01 -5.66645563e-01 2.76793718e-01
9.26131666e-01 1.52750641e-01 -4.93677080e-01 -8.58814359e-01
4.04438555e-01 1.47638202e+00 8.37445796e-01 -3.08217496e-01
1.89535335e-01 2.34384447e-01 -4.32752550e-01 -8.97287309e-01
-5.49160838e-01 -9.69803870e-01 -9.48213935e-01 1.49815634e-01
1.02790594e+00 -9.03159738e-01 -1.10013850e-01 8.85600686e-01
-1.32123899e+00 4.40274104e-02 -2.79226154e-01 1.94206983e-01
-1.46016762e-01 2.13962525e-01 -4.87550706e-01 -8.71240795e-01
-7.78733671e-01 -5.89776933e-01 1.03218889e+00 2.25110754e-01
-5.22500992e-01 -1.05221903e+00 7.58458734e-01 3.25780988e-01
3.25274527e-01 4.32819635e-01 3.46279651e-01 -1.94499230e+00
3.29527557e-01 -5.67112803e-01 -3.04998495e-02 3.58863443e-01
-5.15088886e-02 -1.21044219e-01 -1.24365890e+00 -1.38137370e-01
-1.57886729e-01 -2.04997048e-01 5.27791381e-01 -2.39687875e-01
4.35638666e-01 -2.21014068e-01 -5.16721070e-01 2.82914311e-01
1.29016685e+00 3.50373000e-01 8.25673223e-01 1.09709144e+00
1.68447241e-01 5.96838415e-01 4.36932534e-01 6.92518413e-01
2.51560003e-01 6.32501066e-01 -2.93150365e-01 1.10690571e-01
1.59707069e-01 1.91118732e-01 2.20866478e-03 2.19926834e-01
2.32239012e-02 4.54944856e-02 -1.65666699e+00 8.94926786e-01
-1.69283760e+00 -1.09282076e+00 -4.47145462e-01 2.18417954e+00
1.12480652e+00 -1.05778733e-02 3.42865959e-02 -3.58837768e-02
8.79154503e-01 8.76287222e-02 -1.47175848e-01 -8.98304045e-01
-1.64611712e-01 5.77192843e-01 7.01591074e-01 8.42868090e-02
-1.59908688e+00 7.56093681e-01 7.26875496e+00 7.47890294e-01
-8.68830681e-01 4.86098379e-01 4.46166694e-01 2.20635042e-01
2.29051486e-01 -4.77422714e-01 -1.02909303e+00 3.70566785e-01
1.76663947e+00 -4.10553247e-01 4.65593636e-02 9.93703902e-01
-1.42075785e-03 3.26108515e-01 -8.40702891e-01 5.53424001e-01
-4.20995010e-03 -1.59213352e+00 -3.68244290e-01 4.95929085e-02
2.97414422e-01 6.53041661e-01 -3.15391481e-01 5.51869690e-01
3.76853466e-01 -6.63474917e-01 6.96949244e-01 5.11371493e-01
6.89789295e-01 -9.01045382e-01 1.21342182e+00 3.30684632e-02
-6.19068265e-01 -1.27081156e-01 -1.64454028e-01 4.36072379e-01
5.24695255e-02 6.33780122e-01 -8.42943311e-01 9.20047998e-01
1.03345466e+00 2.10460231e-01 -8.22917521e-01 1.20197284e+00
-1.48516700e-01 5.11099815e-01 -1.93787903e-01 2.08388224e-01
2.11313605e-01 3.59399766e-01 4.64661837e-01 2.03966117e+00
-1.24084912e-01 1.36521703e-03 -2.78758630e-02 6.44165933e-01
-1.64327845e-01 5.00050902e-01 -4.70765203e-01 -1.55457795e-01
4.70541596e-01 1.31558633e+00 -5.42292595e-01 -5.17549694e-01
-4.64473337e-01 8.94837677e-01 2.99699724e-01 -5.51549308e-02
-5.01848280e-01 -7.94455886e-01 5.17424047e-01 -1.32930472e-01
5.27626991e-01 2.81386524e-01 -3.00546139e-01 -6.77461445e-01
-8.81761312e-02 -9.45381641e-01 8.47511649e-01 -3.84255975e-01
-1.22149038e+00 6.82413876e-01 6.16949312e-02 -6.16605163e-01
-5.18039107e-01 -8.47911358e-01 -2.52013415e-01 9.55895424e-01
-8.68367434e-01 -1.31745005e+00 2.38326803e-01 4.65173781e-01
-2.97793988e-02 -7.72738457e-01 1.44431567e+00 7.62342870e-01
-5.93804657e-01 6.70206606e-01 5.84714592e-01 6.89259470e-01
1.08517706e+00 -1.33336473e+00 8.98717403e-01 5.86714089e-01
-1.26411626e-02 7.76519179e-01 3.79237831e-01 -5.98306298e-01
-7.93194771e-01 -1.16343606e+00 1.72999644e+00 -6.06962860e-01
9.73538697e-01 -4.99632597e-01 -9.53194261e-01 7.73331583e-01
5.76950371e-01 -1.20858096e-01 1.15141475e+00 4.01379801e-02
-3.98378879e-01 1.15395330e-01 -1.80419469e+00 -3.32515270e-01
5.87265134e-01 -7.93281615e-01 -1.14025033e+00 3.88114750e-01
2.57809192e-01 -4.96808499e-01 -1.24537885e+00 1.52202055e-01
2.72984862e-01 -1.07031918e+00 7.48755813e-01 -6.49839044e-01
-1.76342353e-01 -1.17428266e-01 1.15206644e-01 -7.79103756e-01
-1.06524691e-01 -7.49833524e-01 1.23154156e-01 2.12314439e+00
6.86934829e-01 -9.61206257e-01 3.08238596e-01 9.49440122e-01
-2.71512210e-01 8.83475319e-02 -1.64277220e+00 -6.73512220e-01
1.25823885e-01 -4.97791529e-01 8.11101735e-01 1.33949554e+00
1.27642408e-01 3.32336873e-01 9.99091044e-02 8.03360641e-02
1.70518439e-02 -6.41435981e-01 5.84818661e-01 -1.33108568e+00
4.09513246e-03 -4.33182269e-01 -8.06533217e-01 1.86710089e-01
4.61579353e-01 -9.24628913e-01 -1.34496689e-01 -1.56007564e+00
7.66175985e-02 -7.35765025e-02 -5.20480871e-01 1.15149724e+00
3.19216192e-01 -9.49177369e-02 -9.21473131e-02 8.88473168e-02
-6.36017323e-01 -3.16752493e-01 -3.21384780e-02 -2.38755420e-01
-3.80252860e-02 -7.12745711e-02 -5.97279131e-01 4.87244219e-01
7.31054485e-01 -5.92305779e-01 3.20852757e-01 -9.66158882e-02
1.62214547e-01 -5.19404650e-01 -9.29198228e-03 -1.09565055e+00
-1.60294041e-01 2.33785227e-01 4.51118022e-01 -7.63922215e-01
-2.18546674e-01 -6.39196813e-01 3.95890981e-01 4.17658120e-01
-4.28445637e-01 -5.64822555e-02 6.12515569e-01 6.68834448e-02
-1.43040642e-01 -6.03484213e-01 6.54007435e-01 -2.47741073e-01
-1.05841660e+00 -1.12086430e-01 -6.21141613e-01 2.25779280e-01
1.15053797e+00 4.28780504e-02 -2.63591856e-01 4.12123501e-01
-6.48820639e-01 -1.54419124e-01 3.82915705e-01 5.97298622e-01
-2.84366254e-02 -1.19221139e+00 -9.89970982e-01 1.34136304e-01
4.54100490e-01 -5.26760161e-01 1.08834647e-01 2.84843683e-01
-6.59069777e-01 9.55752909e-01 -4.25547749e-01 -4.47596908e-02
-1.15369618e+00 7.10696757e-01 3.84488970e-01 -6.58578455e-01
-4.51956451e-01 4.65893805e-01 -6.86630785e-01 -1.14351749e+00
1.12354316e-01 -1.34393722e-01 -6.63785398e-01 3.28775585e-01
9.58438396e-01 6.38603210e-01 6.26687229e-01 -9.41196978e-01
-8.93940806e-01 -1.45647585e-01 -7.84038827e-02 -1.79283962e-01
1.62646139e+00 2.28911757e-01 -2.74390042e-01 3.52240711e-01
1.22260118e+00 1.61080901e-02 -3.22070003e-01 9.63470025e-04
7.71039784e-01 -6.37542754e-02 1.71145692e-01 -1.24822450e+00
-7.07247615e-01 4.22367215e-01 1.15883267e+00 2.44610548e-01
7.48576283e-01 -1.17596470e-01 4.68739063e-01 5.21427631e-01
2.32478976e-01 -1.42614245e+00 -1.09784794e+00 6.67000651e-01
8.04914117e-01 -1.03229260e+00 -1.74589213e-02 3.32622454e-02
-4.22795057e-01 1.16707599e+00 5.46328723e-02 3.26491505e-01
5.76475203e-01 3.19316000e-01 2.91451991e-01 -3.06406170e-01
-2.17999488e-01 -1.55884638e-01 1.15335256e-01 9.83515620e-01
9.59158599e-01 -8.00500438e-03 -9.81875002e-01 9.01225805e-01
5.94788678e-02 2.90482968e-01 2.74596334e-01 1.10760844e+00
-1.05717763e-01 -1.41584182e+00 -5.40710926e-01 1.29028067e-01
-1.30749619e+00 -1.65540099e-01 -9.74176228e-01 1.01946068e+00
3.74503791e-01 8.39336276e-01 -1.49904236e-01 9.69070271e-02
7.17550635e-01 4.96064544e-01 -2.26792078e-02 -6.64861202e-01
-1.40229344e+00 -4.60067153e-01 9.91241157e-01 -5.36100268e-01
-4.97566462e-01 -9.77525830e-01 -1.46512389e+00 -1.05021529e-01
-2.04060543e-02 2.66409725e-01 9.18149114e-01 7.88973391e-01
6.67084992e-01 1.08153291e-01 9.32719037e-02 -4.72788006e-01
-4.57490981e-01 -1.08093286e+00 -6.03408575e-01 5.27759969e-01
-2.46302724e-01 -3.13507825e-01 8.39419514e-02 -1.86558351e-01] | [9.675333023071289, 9.620084762573242] |
a62df25d-ebaa-44bd-bfc2-2edb40d22d12 | esports-pro-players-behavior-during-the-game | 1908.06402 | null | https://arxiv.org/abs/1908.06402v1 | https://arxiv.org/pdf/1908.06402v1.pdf | eSports Pro-Players Behavior During the Game Events: Statistical Analysis of Data Obtained Using the Smart Chair | Today's competition between the professional eSports teams is so strong that in-depth analysis of players' performance literally crucial for creating a powerful team. There are two main approaches to such an estimation: obtaining features and metrics directly from the in-game data or collecting detailed information about the player including data on his/her physical training. While the correlation between the player's skill and in-game data has already been covered in many papers, there are very few works related to analysis of eSports athlete's skill through his/her physical behavior. We propose the smart chair platform which is to collect data on the person's behavior on the chair using an integrated accelerometer, a gyroscope and a magnetometer. We extract the important game events to define the players' physical reactions to them. The obtained data are used for training machine learning models in order to distinguish between the low-skilled and high-skilled players. We extract and figure out the key features during the game and discuss the results. | ['Andrey Somov', 'Evgeny Burnaev', 'Anton Smerdov'] | 2019-08-18 | null | null | null | null | ['sensor-modeling', 'skills-evaluation', 'skills-assessment', 'fps-games'] | ['computer-vision', 'computer-vision', 'computer-vision', 'playing-games'] | [-2.95391798e-01 -2.68012792e-01 -3.83041948e-01 -3.55800271e-01
-3.30402315e-01 -1.79207906e-01 3.45464535e-02 4.45569605e-01
-7.79932857e-01 3.89521599e-01 3.05510104e-01 1.67159364e-01
-4.85100061e-01 -1.03928316e+00 -1.09568775e-01 -4.84464735e-01
1.25753060e-01 4.94119108e-01 4.23821539e-01 -6.78193092e-01
4.88736272e-01 3.59718442e-01 -2.01125336e+00 1.82427749e-01
4.09254342e-01 8.82786751e-01 1.91810608e-01 9.28722024e-01
1.89123213e-01 9.33693469e-01 -9.43513870e-01 -3.68997484e-01
1.90364853e-01 -5.31761885e-01 -5.98329365e-01 2.42896061e-02
-2.48945549e-01 3.19756955e-01 -1.96046177e-02 7.52867818e-01
5.16949713e-01 2.73842961e-01 2.23565340e-01 -1.02270639e+00
9.85665321e-02 5.51379979e-01 -4.47542518e-01 5.00043035e-01
1.03199959e+00 -6.27721148e-03 4.64965522e-01 -2.47324049e-01
2.47899756e-01 6.62581146e-01 1.07984042e+00 2.93526977e-01
-8.14060569e-01 -6.64743781e-01 -3.74906301e-01 5.70135295e-01
-1.32684350e+00 -6.76924735e-02 1.05385566e+00 -7.99803853e-01
4.07527596e-01 4.86370414e-01 1.50268292e+00 1.03627777e+00
3.03762048e-01 2.89744526e-01 1.39587426e+00 -3.83219421e-01
3.89989257e-01 3.34449947e-01 3.86917323e-01 4.27436411e-01
3.10065418e-01 -2.67292350e-03 -1.08496380e+00 6.87513426e-02
6.76794291e-01 1.44823059e-01 1.57509312e-01 -7.69006610e-02
-6.81729794e-01 4.64879990e-01 -9.19127092e-02 3.97557229e-01
-6.92682624e-01 -1.12371035e-02 5.10103166e-01 2.53384739e-01
2.54384249e-01 5.19131839e-01 -3.74137163e-01 -1.17436886e+00
-9.27010238e-01 5.93394935e-01 7.68113315e-01 1.54231355e-01
5.44364810e-01 -2.68692642e-01 1.41649067e-01 5.82514405e-01
9.77639928e-02 2.47328445e-01 6.70024812e-01 -7.58567274e-01
4.25213426e-01 1.14364266e+00 1.19194672e-01 -1.37936425e+00
-8.30491543e-01 -3.79779428e-01 -1.69942275e-01 3.78815651e-01
6.10763073e-01 -3.58276248e-01 -8.52609798e-02 1.40123689e+00
4.06584948e-01 1.15459405e-01 -4.05392081e-01 1.16976941e+00
8.73805046e-01 -5.33524081e-02 -2.13037580e-02 -1.23374082e-01
1.71674740e+00 -3.76673639e-01 -8.58381152e-01 -2.19075009e-01
4.46688294e-01 -4.80465412e-01 1.15911651e+00 6.63336158e-01
-1.22560871e+00 -9.84317899e-01 -9.22899187e-01 3.08214486e-01
-3.47241879e-01 9.77221727e-02 8.31570983e-01 1.14404106e+00
-4.21925902e-01 1.00942194e+00 -1.22287965e+00 -3.01914096e-01
-2.48408809e-01 7.45711625e-01 -5.34341753e-01 6.73737466e-01
-1.16947663e+00 1.06160402e+00 1.41186386e-01 8.86152908e-02
-4.47666377e-01 -4.81082052e-01 -7.82388985e-01 -2.35044301e-01
4.59021807e-01 -3.37549210e-01 1.01710773e+00 -6.38582110e-01
-1.99083102e+00 9.92168009e-01 3.37971985e-01 -9.43187699e-02
5.91237903e-01 -3.94218773e-01 -5.44593573e-01 -1.90095365e-01
3.02432150e-01 -5.61750114e-01 2.41806000e-01 -4.29901391e-01
-8.05249155e-01 -9.04666066e-01 6.53827637e-02 3.93257141e-01
-3.13355893e-01 1.15138374e-01 -4.91594851e-01 -2.27353722e-01
3.76784027e-01 -8.94557476e-01 -1.25980809e-01 -9.09283578e-01
-4.19976294e-01 -2.41506368e-01 2.38696530e-01 -6.28386021e-01
1.62589025e+00 -1.94895053e+00 1.79809600e-01 3.08749676e-01
2.43049309e-01 2.04241812e-01 7.83520877e-01 6.72202647e-01
-6.20281883e-03 -4.51670885e-01 4.61462110e-01 -2.00379565e-01
-1.55276181e-02 2.78737277e-01 3.84411275e-01 4.88288015e-01
-5.20038188e-01 5.01571536e-01 -7.70914793e-01 -4.85596508e-01
4.42592651e-01 1.57878727e-01 -3.64425987e-01 2.58648127e-01
6.45281851e-01 5.76416492e-01 -6.75226450e-01 3.91336650e-01
1.01740547e-01 5.78815579e-01 2.18897030e-01 9.20099244e-02
-3.40816528e-01 5.52555382e-01 -1.46714222e+00 1.61557746e+00
-3.25562268e-01 3.35087031e-01 2.68091988e-02 -1.18734300e+00
1.12809229e+00 3.90283138e-01 8.48565221e-01 -7.62290716e-01
4.91048932e-01 1.73118338e-01 -6.66600093e-03 -1.10420728e+00
7.30177283e-01 -4.29386824e-01 -7.70207763e-01 3.20468068e-01
-1.13254845e-01 8.80717039e-02 4.91141349e-01 -2.08136141e-01
1.14055395e+00 3.82072896e-01 3.63399684e-01 -3.15042257e-01
5.12144446e-01 6.70859665e-02 6.93691969e-01 5.55828571e-01
-3.18020403e-01 1.76648587e-01 4.64201242e-01 -5.66963613e-01
-3.48496348e-01 -8.99720848e-01 3.24895471e-01 1.15446925e+00
3.01503450e-01 -8.07054818e-01 -9.05012786e-01 1.83415599e-02
-9.92105380e-02 -2.17796564e-02 -7.98996568e-01 -5.92505395e-01
-5.47312379e-01 -5.96927643e-01 5.15404165e-01 6.84549868e-01
4.92185742e-01 -9.82565522e-01 -1.04040480e+00 3.19285393e-01
-3.19045573e-01 -8.41170371e-01 1.67019695e-01 1.30992293e-01
-8.82789791e-01 -1.26288986e+00 8.49732012e-02 -4.06097293e-01
2.37281807e-02 -1.22231379e-01 1.13728011e+00 -1.36029288e-01
-3.35381627e-01 2.81210840e-01 -3.83498520e-01 -7.11309373e-01
-3.77454865e-03 2.41943389e-01 6.43452704e-01 -5.21607064e-02
8.19919348e-01 -8.62917066e-01 -6.34861708e-01 4.31529194e-01
-1.78464614e-02 7.52782798e-04 3.15941304e-01 -1.42365377e-02
3.49426597e-01 2.62584656e-01 -7.04595307e-03 -6.98606193e-01
9.43445265e-01 -3.74418676e-01 2.92449836e-02 -3.62671405e-01
-3.69577110e-01 -2.77084619e-01 3.35342556e-01 -4.40382957e-01
-6.19182289e-01 9.40200910e-02 -3.29005271e-01 2.51309812e-01
-2.82022506e-01 4.39074695e-01 -1.69939965e-01 3.01222444e-01
8.87024283e-01 -2.01168098e-02 -6.39466420e-02 -8.24071646e-01
-2.72430927e-01 7.70138800e-01 9.46177959e-01 -7.55702913e-01
5.62910378e-01 4.12540406e-01 5.34664802e-02 -6.89437628e-01
-1.00722587e+00 -9.89280939e-01 -8.28925014e-01 -9.16427851e-01
9.38798547e-01 -6.62957907e-01 -1.77096009e+00 4.23229903e-01
-3.20819467e-01 -1.28414333e-01 -5.42617440e-01 1.08349776e+00
-7.31316984e-01 1.46017283e-01 -6.08239055e-01 -1.26295018e+00
-9.70249996e-02 -8.93823683e-01 8.65000069e-01 6.62040830e-01
-7.14480042e-01 -8.50864410e-01 7.16718078e-01 9.79957104e-01
4.58577238e-02 5.07417977e-01 -5.97148351e-02 -2.17454076e-01
3.24926674e-01 -6.72623336e-01 6.81028962e-01 6.25739470e-02
1.71906263e-01 -2.51538277e-01 -7.05316901e-01 4.12482396e-02
5.00937283e-01 -1.47247300e-01 1.31828412e-01 4.72491950e-01
8.05720270e-01 3.64774346e-01 -2.12048933e-01 5.48935652e-01
1.25845933e+00 -2.02509910e-02 6.91305757e-01 6.25842631e-01
7.05820203e-01 6.65211737e-01 9.10274148e-01 5.54030538e-01
2.67083168e-01 1.12739396e+00 2.52815187e-01 1.57455415e-01
2.35660627e-01 -2.92263359e-01 6.48032427e-01 1.10415268e+00
-1.25996447e+00 3.50437403e-01 -6.93435133e-01 2.66509563e-01
-1.90639079e+00 -1.24604225e+00 -8.05591106e-01 2.27438331e+00
6.63426399e-01 4.75908160e-01 8.39080751e-01 9.91632521e-01
7.35309184e-01 -1.62819952e-01 -1.03140354e-01 -8.07162642e-01
3.81293178e-01 5.95480561e-01 6.34948194e-01 2.22037598e-01
-8.22390974e-01 6.08177602e-01 6.38437271e+00 6.51742339e-01
-1.02449834e+00 7.16777071e-02 -4.27066907e-02 -4.80189323e-01
5.31351566e-01 -4.06793356e-02 -7.46899843e-01 6.18445098e-01
1.16041458e+00 -6.66584820e-02 9.10059139e-02 9.43871200e-01
6.75282598e-01 -5.81863165e-01 -7.69526184e-01 1.08886111e+00
-4.06277850e-02 -7.78712869e-01 -8.17351401e-01 3.12029570e-01
1.63201496e-01 -3.64684582e-01 -2.15566874e-01 5.21367431e-01
1.19844355e-01 -7.22192764e-01 7.33913004e-01 8.47910166e-01
3.70233387e-01 -8.51184547e-01 8.30084980e-01 6.77483320e-01
-1.24685931e+00 -1.34621724e-01 -2.27792412e-02 -1.32769752e+00
1.01608641e-01 6.78139508e-01 -4.94065672e-01 5.41207790e-01
7.62950301e-01 5.03339648e-01 -5.92901289e-01 6.60228014e-01
-1.95041820e-01 8.76606584e-01 -4.02007282e-01 -1.66547731e-01
-1.03785634e-01 -5.21529019e-01 4.02481973e-01 7.38978684e-01
2.91879773e-01 -5.78857400e-03 2.84087747e-01 4.22797441e-01
6.17418289e-01 3.90416652e-01 -3.42924416e-01 2.12506518e-01
-7.42440112e-03 1.39770710e+00 -7.21015871e-01 -1.92740485e-01
-9.91521850e-02 7.87757814e-01 1.16599247e-01 -4.41787601e-01
-7.63493955e-01 -3.60134393e-01 1.02578020e+00 7.69583106e-01
-4.30042982e-01 -3.52948755e-01 -5.02543092e-01 -9.38355923e-01
1.98501959e-01 -6.16759121e-01 4.20871258e-01 -5.32804668e-01
-8.63250017e-01 1.80761456e-01 -2.43163139e-01 -1.30241251e+00
-3.45812529e-01 -8.15531373e-01 -8.58980954e-01 8.59664977e-01
-4.84167695e-01 -5.73130369e-01 -7.10932314e-01 7.17258036e-01
3.34827632e-01 -1.70706958e-02 8.02698493e-01 4.61478442e-01
-8.13218296e-01 2.59368181e-01 -3.94292831e-01 3.85140866e-01
5.22979200e-01 -1.24018610e+00 -4.60024625e-02 5.11991799e-01
1.22985825e-01 3.92861426e-01 1.04797494e+00 -6.07642114e-01
-1.45447695e+00 -1.46697372e-01 8.15197289e-01 -8.98776770e-01
5.77926517e-01 -3.90111595e-01 -2.78576225e-01 3.46655577e-01
-3.48249674e-01 -3.01909536e-01 1.24075413e+00 5.36875427e-01
3.55163246e-01 -3.46932471e-01 -7.36568272e-01 4.12187755e-01
1.15590084e+00 -5.20788133e-01 -9.02911246e-01 2.59014145e-02
-3.57512176e-01 -5.36665320e-01 -1.18519723e+00 1.47082284e-01
9.02432024e-01 -1.48942101e+00 9.47879970e-01 -6.49159729e-01
4.38773185e-01 -1.71768934e-01 2.73827404e-01 -1.18376553e+00
-2.80709207e-01 -4.01783198e-01 2.54108846e-01 9.26312506e-01
-1.66256860e-01 -1.49731219e-01 1.30658627e+00 6.25236571e-01
3.51207843e-03 -6.87893450e-01 -5.74618995e-01 -5.16448438e-01
-4.25069630e-01 -9.72961485e-01 3.54964793e-01 7.93208599e-01
8.16666007e-01 3.82505804e-01 -5.53724647e-01 -1.07892714e-01
3.52037221e-01 1.57371014e-02 1.26292872e+00 -1.75092018e+00
-6.85898960e-01 -2.69154400e-01 -1.23486781e+00 -3.15132558e-01
-3.65955532e-01 -4.66727734e-01 -2.92442650e-01 -1.45237052e+00
1.93849668e-01 -2.00211048e-01 -3.08130026e-01 1.68703329e-02
-1.65603459e-01 5.18605053e-01 -1.77241668e-01 -2.77966410e-02
-6.80286825e-01 -1.78850830e-01 1.10968268e+00 5.79611778e-01
-4.96282578e-01 5.75299203e-01 -7.95521677e-01 9.68277931e-01
7.08843112e-01 -5.12769401e-01 -3.35770756e-01 1.17712542e-01
7.19175637e-01 3.43250155e-01 1.48240775e-01 -1.61517215e+00
8.68453085e-02 -3.75195682e-01 2.96668887e-01 -2.61063367e-01
6.82143211e-01 -5.29162049e-01 3.21473211e-01 7.64443219e-01
-5.39209656e-02 -4.05987632e-03 -1.37820274e-01 2.30669796e-01
-4.63552326e-01 -1.64341196e-01 4.06277388e-01 -2.71327406e-01
-5.63815951e-01 2.35105958e-02 -5.53077757e-01 -8.54004472e-02
1.13246715e+00 -8.81558299e-01 1.84637174e-01 -4.02148038e-01
-1.26202941e+00 -7.05879107e-02 3.57917935e-01 3.90512675e-01
2.28770226e-02 -1.16446865e+00 -4.79604214e-01 1.49666607e-01
8.24135635e-03 -5.11367083e-01 6.45893753e-01 1.17693603e+00
-7.17876732e-01 -9.93279815e-02 -7.57535756e-01 -4.76689488e-01
-1.57585156e+00 3.91085327e-01 2.81353474e-01 -4.42068189e-01
-5.10795593e-01 6.19007289e-01 -5.20061076e-01 -2.31239572e-01
-2.19221801e-01 -1.47651762e-01 -8.38013589e-01 1.96977124e-01
5.64745784e-01 8.23911071e-01 2.20907792e-01 -7.68042147e-01
-5.21863997e-01 7.80440032e-01 4.74191427e-01 -2.34759107e-01
1.27917457e+00 -1.00438513e-01 1.00270100e-01 1.06452549e+00
6.41568363e-01 5.72398961e-01 -8.73704314e-01 1.34945735e-01
1.12422153e-01 -5.47850728e-01 -1.40725598e-01 -2.23124415e-01
-9.33342457e-01 8.27801764e-01 7.31946111e-01 5.85025370e-01
9.96294558e-01 -2.18096286e-01 6.64685011e-01 3.22870873e-02
6.81631625e-01 -1.88713229e+00 1.34647995e-01 2.29842544e-01
2.45254993e-01 -8.69833291e-01 7.52667859e-02 -4.77092087e-01
-7.18889356e-01 8.32329273e-01 5.02502620e-01 -5.22298217e-01
7.21674025e-01 4.63436633e-01 2.66327918e-01 -4.85478282e-01
-3.20902318e-01 -4.94600832e-01 2.11306706e-01 5.96545577e-01
5.35081327e-01 3.87344867e-01 -8.26342463e-01 1.24147642e+00
-1.16740942e+00 2.47505516e-01 6.73549950e-01 8.81466091e-01
-6.13654554e-01 -1.26737571e+00 -6.36774957e-01 3.93308312e-01
-8.96283627e-01 7.30617702e-01 -4.04082805e-01 8.32335711e-01
7.72578895e-01 1.19386554e+00 8.53268132e-02 -9.22718167e-01
9.35943902e-01 1.92891598e-01 5.91476142e-01 -9.14661229e-01
-1.14355254e+00 -3.69557589e-01 2.89907366e-01 -8.64690542e-01
-6.15005851e-01 -8.03056419e-01 -1.11301970e+00 -6.92679644e-01
-1.33607179e-01 5.30174553e-01 1.01194596e+00 1.04390514e+00
-1.43382967e-01 7.32073367e-01 6.09961212e-01 -1.02884102e+00
3.14008482e-02 -1.18317032e+00 -1.15816665e+00 8.13458085e-01
-4.71997738e-01 -1.08765733e+00 -8.32044184e-02 -1.34036884e-01] | [6.886430740356445, 0.37750786542892456] |
c0283a3a-9a75-4caf-b72a-5152b35156f5 | a-penalized-poisson-likelihood-approach-to | 2306.06756 | null | https://arxiv.org/abs/2306.06756v1 | https://arxiv.org/pdf/2306.06756v1.pdf | A Penalized Poisson Likelihood Approach to High-Dimensional Semi-Parametric Inference for Doubly-Stochastic Point Processes | Doubly-stochastic point processes model the occurrence of events over a spatial domain as an inhomogeneous Poisson process conditioned on the realization of a random intensity function. They are flexible tools for capturing spatial heterogeneity and dependence. However, implementations of doubly-stochastic spatial models are computationally demanding, often have limited theoretical guarantee, and/or rely on restrictive assumptions. We propose a penalized regression method for estimating covariate effects in doubly-stochastic point processes that is computationally efficient and does not require a parametric form or stationarity of the underlying intensity. We establish the consistency and asymptotic normality of the proposed estimator, and develop a covariance estimator that leads to a conservative statistical inference procedure. A simulation study shows the validity of our approach under less restrictive assumptions on the data generating mechanism, and an application to Seattle crime data demonstrates better prediction accuracy compared with existing alternatives. | ['Ali Shojaie', 'Jon Wakefield', 'Si Cheng'] | 2023-06-11 | null | null | null | null | ['point-processes'] | ['methodology'] | [ 2.56336898e-01 -5.12603283e-01 -3.01261783e-01 -2.79814005e-01
-6.47303820e-01 -3.53107095e-01 5.63656926e-01 4.04520601e-01
-4.69216824e-01 1.10523486e+00 1.91441521e-01 -4.96537268e-01
-5.73518753e-01 -1.01179206e+00 -6.43335760e-01 -7.79842675e-01
-1.16672151e-01 4.76695508e-01 1.54685050e-01 3.12301397e-01
2.71135360e-01 4.70285177e-01 -1.33889437e+00 -5.31124055e-01
9.48506892e-01 4.40302759e-01 3.43025565e-01 5.29801130e-01
-7.27803111e-02 6.53734148e-01 -4.06785071e-01 -2.76892930e-01
-1.21506840e-01 -1.63550690e-01 -3.12998563e-01 -1.81645807e-02
-1.71368256e-01 -1.96668193e-01 4.91051041e-02 7.43586063e-01
4.23057467e-01 8.20203424e-02 1.33919311e+00 -1.25160158e+00
-5.54568112e-01 2.24234700e-01 -9.75835979e-01 5.49444437e-01
2.64940709e-01 -2.03250974e-01 5.27688324e-01 -9.70395923e-01
1.13848686e-01 1.00398564e+00 1.24831843e+00 -5.28707616e-02
-1.55540335e+00 -5.09798348e-01 5.50241359e-02 -3.73243809e-01
-1.84515727e+00 -3.33093405e-01 3.09615642e-01 -5.83326697e-01
6.01682067e-01 1.88739657e-01 3.79604906e-01 1.10524178e+00
4.03885484e-01 3.57093096e-01 1.13026202e+00 -2.02425629e-01
5.48515260e-01 -1.93079978e-01 1.73878089e-01 1.27263501e-01
7.09290385e-01 -6.57118335e-02 -1.67091832e-01 -9.06947672e-01
1.30875170e+00 3.55923712e-01 1.82876885e-01 -2.55954355e-01
-6.05626464e-01 1.01858079e+00 -3.22700948e-01 1.89293623e-01
-1.00656235e+00 2.69830823e-01 -1.17844185e-02 -1.67320758e-01
9.00867760e-01 -6.18772328e-01 -6.81048185e-02 -2.62288809e-01
-1.15122640e+00 5.21819890e-01 6.77025199e-01 9.32955265e-01
4.74491209e-01 -1.79278944e-02 -4.51010704e-01 7.54928052e-01
2.39927977e-01 1.06886876e+00 -3.09392214e-01 -8.52164924e-01
5.63865185e-01 5.75501937e-03 7.52834797e-01 -9.44912553e-01
-4.06611145e-01 -2.67138481e-01 -1.03149962e+00 -9.83067378e-02
6.85342014e-01 -4.54323828e-01 -5.99432886e-01 1.88446176e+00
3.28865260e-01 5.90425253e-01 -4.93780673e-01 3.27660650e-01
-7.15307742e-02 6.76283360e-01 7.03500330e-01 -4.65129495e-01
9.67148960e-01 8.28050822e-02 -8.65147531e-01 1.94061309e-01
4.96782660e-01 -3.90529215e-01 9.23297763e-01 -9.56058726e-02
-1.42454135e+00 -1.00668244e-01 -9.82571691e-02 3.35703522e-01
1.08125754e-01 -2.35768333e-01 5.10482311e-01 8.71718824e-01
-9.67317343e-01 3.54695469e-01 -9.53376114e-01 -3.18586558e-01
3.53225261e-01 1.51769683e-01 -3.19185480e-02 6.68363348e-02
-9.15069282e-01 4.90372270e-01 -3.45470339e-01 -6.05824180e-02
-6.39843106e-01 -1.21072340e+00 -7.25985587e-01 3.64595860e-01
-6.27246127e-02 -8.66855562e-01 8.40309799e-01 -5.92418969e-01
-9.12928045e-01 5.08669734e-01 -6.99649394e-01 -3.53601605e-01
7.24190295e-01 1.16311640e-01 -9.55602676e-02 -5.66904843e-02
7.68453121e-01 -3.83271515e-01 6.63604081e-01 -1.14425480e+00
-6.33355498e-01 -3.67526352e-01 -3.38250309e-01 2.29714423e-01
-8.34682658e-02 4.40093160e-01 -2.21341327e-01 -7.95887291e-01
-7.84428939e-02 -7.01383293e-01 -5.30012131e-01 -1.88268706e-01
-1.34717256e-01 5.13375811e-02 1.74151093e-01 -5.98037422e-01
1.47413135e+00 -2.30244350e+00 -2.52416760e-01 4.43293840e-01
-2.18779996e-01 -4.53583181e-01 1.85698882e-01 6.25745952e-01
5.53358831e-02 2.33618513e-01 -7.18127668e-01 -5.44045031e-01
-2.19586477e-01 1.41299918e-01 -1.65825397e-01 1.05500126e+00
9.30884853e-02 3.69748414e-01 -9.08595920e-01 -5.23731470e-01
3.10659736e-01 5.06159306e-01 -4.48691726e-01 -1.51216045e-01
4.77627724e-01 3.95959795e-01 -4.26808357e-01 3.17169040e-01
8.75727117e-01 -3.10162336e-01 -6.55558854e-02 7.23905265e-01
-5.22512496e-01 -9.06982571e-02 -1.55734563e+00 8.02993417e-01
-6.77266061e-01 7.36338347e-02 2.56362557e-01 -8.46242607e-01
6.84019148e-01 4.34620053e-01 7.64812231e-01 -2.55588770e-01
-1.79783881e-01 1.14800416e-01 -4.31644619e-01 -1.44577533e-01
3.13553303e-01 -6.75936222e-01 -2.19541475e-01 5.02800107e-01
-2.77167618e-01 -1.86769105e-02 -8.70393291e-02 -2.55611390e-01
1.30287099e+00 -1.64640397e-01 3.76877815e-01 -6.25526130e-01
1.00857668e-01 -2.14162752e-01 6.43601358e-01 1.34537590e+00
-1.17299296e-01 5.79404533e-01 5.01559079e-01 -4.78196144e-02
-1.25178826e+00 -1.57595897e+00 -6.69701099e-01 8.65776241e-01
1.86730288e-02 3.42623979e-01 -4.12708342e-01 -1.93491448e-02
3.31514746e-01 9.34377372e-01 -7.99084425e-01 1.68750748e-01
-3.89927506e-01 -1.39876497e+00 6.27297759e-01 6.75746441e-01
1.13213874e-01 -5.93162775e-01 -2.87232786e-01 3.65902454e-01
-6.63965195e-02 -8.22955966e-01 -9.55365971e-02 -9.44084581e-03
-1.05704880e+00 -7.99732447e-01 -1.06967878e+00 -3.50791305e-01
6.23499095e-01 4.20649052e-01 9.17491674e-01 -2.62805074e-01
6.89281002e-02 5.95351040e-01 1.44830883e-01 -4.69737560e-01
-1.39736049e-02 -2.72797287e-01 1.40253350e-01 1.45530537e-01
3.99869770e-01 -6.05978549e-01 -6.85785294e-01 2.17598379e-01
-8.50120425e-01 -5.02919376e-01 1.42845854e-01 6.76320016e-01
6.08391702e-01 2.95288235e-01 7.33473480e-01 -8.28244865e-01
6.36665046e-01 -1.03684974e+00 -6.61956966e-01 -2.90042460e-02
-1.76923826e-01 -3.66812110e-01 2.00578526e-01 -3.55420709e-01
-1.66557896e+00 -1.63086146e-01 1.74073890e-01 -2.58853912e-01
-4.21781152e-01 6.50005639e-01 9.20569524e-03 2.97698379e-01
4.57817674e-01 6.16173409e-02 -2.20822573e-01 -3.58094811e-01
-1.43581268e-03 3.10909241e-01 2.91630119e-01 -6.81016147e-01
6.18396282e-01 1.07284176e+00 3.02258402e-01 -1.15517497e+00
-3.69851112e-01 -6.58702016e-01 -5.55041909e-01 1.35454470e-02
7.15878010e-01 -1.09067082e+00 -5.87538362e-01 5.09250224e-01
-1.01494014e+00 -5.62629580e-01 -3.10731977e-01 8.28794897e-01
-8.94319475e-01 2.20034719e-01 -7.99488187e-01 -1.52054656e+00
3.92754704e-01 -5.41055799e-01 1.14287555e+00 -1.58865541e-01
-3.78733128e-01 -1.61056173e+00 3.74066085e-01 -3.06313718e-03
4.05666500e-01 3.04565012e-01 8.87456954e-01 -1.25640854e-01
-2.24187940e-01 -3.18038166e-01 -3.03779453e-01 -2.92435914e-01
2.99653798e-01 1.92359075e-01 -5.69645643e-01 6.25088625e-03
2.12457180e-01 4.79090303e-01 4.19915617e-01 1.30432165e+00
1.31136525e+00 -1.16537258e-01 -4.17461306e-01 3.84077638e-01
1.65786469e+00 -3.60835972e-03 6.20924175e-01 1.83615491e-01
5.00359952e-01 8.47716808e-01 4.47212368e-01 1.08990121e+00
5.56364000e-01 3.97001326e-01 5.58992624e-02 -1.41428977e-01
4.72160608e-01 -2.87782729e-01 9.27864481e-03 1.89666972e-01
-3.57563853e-01 -2.14192361e-01 -1.22781193e+00 1.11288118e+00
-1.81779778e+00 -1.54218411e+00 -9.51926291e-01 2.60012126e+00
7.24757552e-01 -3.30492586e-01 4.42970276e-01 -6.42084926e-02
1.14967382e+00 -7.76679218e-02 -1.91005632e-01 -2.68940926e-01
-3.29843402e-01 2.53360838e-01 1.21070158e+00 6.74437225e-01
-9.60915625e-01 5.54743886e-01 7.80117989e+00 8.08976531e-01
-3.99940938e-01 4.17993993e-01 5.68855047e-01 -2.07748823e-02
-4.72811788e-01 -1.22865930e-01 -6.57790005e-01 7.05432415e-01
1.08648407e+00 -3.34810615e-01 -1.97051167e-01 4.09402400e-01
9.83350992e-01 -3.51758301e-01 -6.25251710e-01 6.48527026e-01
-5.72172165e-01 -9.50669169e-01 -2.88679838e-01 4.35490906e-01
8.08764935e-01 -1.90003410e-01 3.88479382e-01 -6.95324317e-02
8.93512487e-01 -1.02142358e+00 5.95310509e-01 7.57590771e-01
6.48131013e-01 -1.00765121e+00 6.83550835e-01 5.67240357e-01
-9.90939975e-01 -8.54082257e-02 -4.35390472e-01 -4.83683497e-01
7.20734894e-01 8.15635979e-01 -1.85699299e-01 1.48817509e-01
6.47805989e-01 3.03426385e-01 -8.95595457e-03 1.27429068e+00
1.43245850e-02 1.04182005e+00 -6.06243014e-01 1.25211388e-01
1.83963716e-01 -3.62861991e-01 4.58368421e-01 1.32187188e+00
7.68943608e-01 2.20203444e-01 -2.63843298e-01 8.54558170e-01
4.59706962e-01 2.50016272e-01 -9.74492431e-01 5.13839722e-01
7.46477425e-01 5.98396122e-01 -7.32143879e-01 -5.39706238e-02
-8.46720695e-01 6.34467781e-01 1.16099097e-01 6.88933015e-01
-9.94428456e-01 1.58657029e-01 5.55024505e-01 5.41049480e-01
3.31220865e-01 -3.51382852e-01 -8.86655688e-01 -9.00792122e-01
-8.56627617e-03 -1.57865644e-01 4.02359664e-01 -4.48225379e-01
-1.57594609e+00 -2.63180882e-01 5.26883364e-01 -8.84983778e-01
-2.16027573e-01 -1.24898545e-01 -8.73894930e-01 1.38997984e+00
-1.23780823e+00 -8.56609464e-01 7.05438182e-02 8.55223477e-01
1.70215786e-01 3.88119072e-01 6.16760373e-01 2.56613851e-01
-4.87498283e-01 3.99677940e-02 5.24733186e-01 -3.85870934e-01
2.37098560e-01 -1.24646413e+00 2.57855982e-01 6.58678234e-01
-4.02761847e-01 6.85151637e-01 1.00139654e+00 -1.13596082e+00
-8.37001026e-01 -1.09246564e+00 9.89603341e-01 -5.35336316e-01
1.00726581e+00 -2.50379503e-01 -1.03878045e+00 6.63947582e-01
-2.26216331e-01 -2.41674080e-01 1.13058817e+00 4.09393221e-01
2.34193966e-01 3.66550893e-01 -1.37158430e+00 6.60522401e-01
9.40789282e-01 -4.55138713e-01 -3.37087661e-01 3.76965493e-01
1.09410226e-01 3.22857946e-01 -8.89677942e-01 2.91896909e-01
3.91006738e-01 -8.06709766e-01 9.06776547e-01 -3.70017946e-01
7.49461204e-02 -1.11036278e-01 -2.06374705e-01 -9.31449175e-01
-7.24895358e-01 -2.49511346e-01 3.63094032e-01 1.29526687e+00
4.56268072e-01 -6.39271259e-01 7.73399472e-01 9.14745629e-01
4.50149536e-01 -2.28644997e-01 -1.30088246e+00 -9.19549346e-01
4.79837120e-01 -8.01347852e-01 6.28441930e-01 9.71628129e-01
-3.26185554e-01 -1.56106472e-01 -5.72782099e-01 6.08536482e-01
1.18508613e+00 -2.49112859e-01 6.03445768e-01 -1.49129939e+00
-1.21936902e-01 -1.06909394e-01 -1.88741684e-01 -5.62395573e-01
4.74921256e-01 -1.67663500e-01 -2.30047223e-03 -1.47082520e+00
5.20919085e-01 -8.21993351e-01 -3.48612405e-02 3.30080986e-02
-2.98418283e-01 6.58922344e-02 -5.70563793e-01 3.83514404e-01
-1.61145739e-02 6.23291373e-01 5.80178142e-01 2.60054082e-01
-3.93178552e-01 6.40121102e-01 -4.45107251e-01 8.74213457e-01
7.39638567e-01 -6.50049090e-01 -5.20822167e-01 -3.52603763e-01
3.25700164e-01 5.98267198e-01 7.44130015e-01 -4.80050594e-01
1.55141279e-01 -6.15311861e-01 3.19893450e-01 -6.56881750e-01
2.92369425e-01 -9.56735432e-01 6.80749178e-01 2.43327796e-01
-5.81899546e-02 2.29540855e-01 1.04933314e-01 1.01604247e+00
6.21417630e-03 -3.09389085e-01 5.43204069e-01 6.84539378e-02
3.49171609e-02 4.31629539e-01 -8.62206697e-01 -9.86168534e-02
1.12292850e+00 -2.76569784e-01 1.51945874e-01 -6.52260661e-01
-7.98613608e-01 4.12897905e-04 6.24382734e-01 -2.11804122e-01
4.29665446e-01 -1.29525042e+00 -8.19811165e-01 -9.25160497e-02
-2.43982479e-01 -3.41685683e-01 4.10362244e-01 1.07496941e+00
-3.38180453e-01 2.59370148e-01 1.50832355e-01 -6.12618983e-01
-9.93731976e-01 6.43977404e-01 9.80776921e-02 -7.20329955e-02
-4.54925805e-01 3.57434481e-01 6.59120500e-01 -1.99148938e-01
-2.94030994e-01 -5.12322746e-02 -3.53463478e-02 1.98774692e-02
4.79399681e-01 9.20296907e-01 -3.41489553e-01 -8.55198622e-01
-4.41821009e-01 3.75891536e-01 6.05057240e-01 -6.38826132e-01
1.31188834e+00 -6.39559865e-01 -1.49945304e-01 9.68640804e-01
6.46518588e-01 1.53235435e-01 -1.38282537e+00 -8.57659280e-02
8.40932205e-02 -5.33179700e-01 -6.91836923e-02 1.02664810e-02
-4.26003247e-01 6.31680608e-01 2.56801933e-01 4.76170182e-01
8.06474268e-01 -9.14447829e-02 7.34090358e-02 -3.17743629e-01
6.03024125e-01 -9.06128943e-01 -6.47642851e-01 3.45248818e-01
5.25571465e-01 -8.30678284e-01 -8.87056813e-02 -7.51530409e-01
-3.72162759e-01 4.49701786e-01 2.18313141e-03 -4.30058181e-01
1.06049883e+00 5.51511168e-01 -4.55413073e-01 -4.56601493e-02
-5.62618196e-01 -2.09271878e-01 -1.87039301e-01 8.23355615e-01
4.24494773e-01 1.88012108e-01 -7.42573977e-01 4.41540003e-01
2.10799590e-01 1.30470410e-01 5.96497893e-01 9.31021452e-01
-2.69804448e-01 -5.83997130e-01 -7.06359208e-01 4.81658280e-01
-1.10447800e+00 -3.04751277e-01 2.26371214e-01 8.43752205e-01
-2.68007219e-01 1.02334499e+00 4.73784208e-01 6.82417512e-01
3.86302441e-01 -3.75502668e-02 2.52604872e-01 -6.45948708e-01
-2.65887856e-01 1.86041713e-01 -1.52131662e-01 -1.79551154e-01
-4.86155301e-01 -1.45382488e+00 -8.22314501e-01 -6.67369902e-01
-3.19089413e-01 1.99750930e-01 4.51972306e-01 9.52555180e-01
1.93520099e-01 9.76591557e-02 7.50514388e-01 -6.59337938e-01
-4.36222941e-01 -7.69776762e-01 -1.06689417e+00 2.74044484e-01
1.97728693e-01 -9.54779744e-01 -5.16341507e-01 -9.54571888e-02] | [7.066011428833008, 4.216282844543457] |
4b57baac-f573-40ba-98d7-76b07ae06902 | connectivity-estimation-of-high-dimensional | 2005.07083 | null | https://arxiv.org/abs/2005.07083v1 | https://arxiv.org/pdf/2005.07083v1.pdf | Connectivity estimation of high dimensional data recorded from neuronal cells | The main result of this thesis is the development of a novel connectivity estimation method, called Total Spiking Probability Edges (TSPE). Based on cross-correlation and edge filtering at different time scales this method is proposed and the theoretical framework is outlined in this work. TSPE enables the classification between inhibitory and excitatory connections by using recorded action potentials. To compare this method learning about state of the art algorithms to estimate connectivity is necessary. After a research, promising algorithms are implemented and evaluated for further research topics, among others in the biomems lab of UAS Aschaffenburg. To evaluate these algorithms in silico networks are used, because of their known connectivity. This makes it possible to validate the correctness of our algorithm results. Therefore, a biophysically representative neuronal network simulation is needed first. Datasets were simulated in different ways and analysed in order to develop an evaluation framework. After a successful evaluation with in silico networks, in vitro experiments and their analyses complete this project. | ['Stefano De Blasi'] | 2020-05-01 | null | null | null | null | ['connectivity-estimation'] | ['graphs'] | [ 4.37619947e-02 2.86615919e-02 4.76830900e-01 1.40220374e-01
5.10513246e-01 -4.40549403e-01 4.29009825e-01 2.44367495e-01
-7.40355730e-01 1.28998172e+00 -4.94672626e-01 4.20594960e-02
-5.43701351e-01 -8.28130960e-01 -6.22673035e-01 -7.75303304e-01
-7.40009427e-01 4.49131668e-01 6.15930915e-01 -1.29655764e-01
4.93700325e-01 9.31913257e-01 -1.78275800e+00 -7.64224678e-03
4.80634212e-01 5.45834303e-01 5.65603137e-01 5.13772368e-01
8.15705433e-02 6.84433132e-02 -6.22083545e-01 1.37683675e-01
8.33170190e-02 -7.23530114e-01 -6.56715930e-01 -4.24590677e-01
-2.77842224e-01 3.97858769e-01 6.94028288e-02 8.87041211e-01
5.74088097e-01 -9.04050618e-02 8.42510879e-01 -1.29444063e+00
3.44703138e-01 1.07039511e+00 1.12668246e-01 4.24096525e-01
2.35389933e-01 -1.69981569e-01 4.04161274e-01 -3.96505028e-01
1.14673686e+00 6.88801110e-01 8.94221961e-01 3.53244334e-01
-1.59674668e+00 -7.93189585e-01 -1.68756932e-01 3.14753950e-01
-1.42373526e+00 2.94678435e-02 5.14643431e-01 -3.96243542e-01
1.07523561e+00 1.34182051e-01 1.35192561e+00 1.08447504e+00
5.48177063e-01 1.54517904e-01 1.90024400e+00 -2.01573834e-01
6.28072262e-01 2.04879329e-01 2.38497719e-01 1.07742175e-01
7.54892588e-01 2.06328154e-01 -2.27679029e-01 1.33176982e-01
1.00918615e+00 -4.70486075e-01 -3.46347600e-01 -1.49145409e-01
-1.17654085e+00 1.10588096e-01 2.63808817e-01 1.20953560e+00
-3.68764549e-01 1.65226609e-01 3.55824172e-01 4.54650879e-01
6.59223786e-03 3.90786171e-01 -3.88290733e-01 -2.07334131e-01
-1.05515850e+00 3.99069101e-01 1.16555882e+00 3.80765140e-01
6.15597725e-01 2.41598543e-02 2.27103084e-01 3.73449236e-01
2.69116968e-01 3.92769098e-01 3.45122784e-01 -7.81643271e-01
-1.42057031e-01 6.32754445e-01 -2.42133737e-01 -1.01801491e+00
-8.39642882e-01 -6.99563622e-01 -9.02483106e-01 8.44886363e-01
4.72617328e-01 -3.57713223e-01 -5.80415726e-01 1.52310872e+00
-6.53216317e-02 4.50524241e-01 1.44145906e-01 7.19325125e-01
4.77380037e-01 3.54106009e-01 -1.04043834e-01 -5.16893029e-01
8.31479490e-01 -7.17569739e-02 -7.97424495e-01 6.18104458e-01
5.14508605e-01 -5.22899449e-01 1.37402758e-01 6.13095224e-01
-1.03840339e+00 -2.26488471e-01 -1.07024741e+00 8.21820140e-01
-8.82025659e-01 -4.10968903e-03 6.40081584e-01 6.74556792e-01
-1.39373934e+00 1.18604791e+00 -1.03499591e+00 -9.21806574e-01
1.46427408e-01 6.57353163e-01 -6.34233356e-01 5.77359200e-01
-1.03352308e+00 1.04003000e+00 4.99013186e-01 2.58287881e-03
-6.65736973e-01 -2.90479720e-01 -3.17791700e-01 7.81223848e-02
-4.86394823e-01 -6.57480717e-01 3.27743560e-01 -7.91134715e-01
-1.46357834e+00 8.69745493e-01 8.43077824e-02 -7.92583048e-01
6.55657411e-01 5.52751601e-01 -5.99143021e-02 2.25143880e-01
-7.33120963e-02 7.27312088e-01 7.18047619e-02 -1.32634616e+00
-2.47110546e-01 -3.75986546e-01 -2.16467202e-01 -2.49026105e-01
-9.28834528e-02 -1.62729383e-01 9.80279446e-02 -5.44631243e-01
3.10018182e-01 -8.00409853e-01 -1.94524869e-01 -4.29193079e-01
-1.37441829e-01 2.12098479e-01 7.47322261e-01 -2.56632686e-01
8.48138869e-01 -1.62014139e+00 3.10528517e-01 6.64957285e-01
1.29222244e-01 1.30198210e-01 1.75875034e-02 7.19076097e-01
-2.20797151e-01 1.81454774e-02 -3.57098699e-01 5.83920181e-02
-4.63963628e-01 1.34547248e-01 1.23170339e-01 4.59468961e-01
6.20098300e-02 3.94135475e-01 -6.00301683e-01 -4.09150809e-01
4.49850112e-01 6.97860062e-01 -1.05524689e-01 -2.13807583e-01
2.64302015e-01 7.11372137e-01 -2.53158957e-01 3.10822606e-01
8.06802154e-01 2.40879253e-01 -8.01182445e-03 -2.25342423e-01
-6.50400519e-01 -2.34840572e-01 -1.39674914e+00 1.37857997e+00
-1.30253091e-01 8.82837057e-01 1.76267549e-01 -1.50331354e+00
1.35439682e+00 4.63376611e-01 7.67318726e-01 -4.99818265e-01
5.97361743e-01 4.83919710e-01 3.82660031e-01 -1.97237730e-01
-1.74560741e-01 2.06544980e-01 6.90080523e-01 1.57237306e-01
3.76123875e-01 2.81952824e-02 8.14280927e-01 1.48455769e-01
1.37257814e+00 3.83048981e-01 2.27511302e-02 -9.81357694e-01
6.81881905e-01 -1.18980482e-02 1.92409635e-01 4.91364628e-01
-2.92229921e-01 2.26533994e-01 6.65935934e-01 -1.49284109e-01
-8.74985993e-01 -8.59457433e-01 -6.44430816e-01 -6.04515746e-02
1.89219579e-01 -1.91764504e-01 -1.20868564e+00 1.62799209e-01
-2.01272547e-01 3.22581790e-02 -7.53833711e-01 2.99860626e-01
-1.26720339e-01 -7.96056271e-01 6.17745638e-01 1.02480486e-01
6.26780748e-01 -1.39547718e+00 -9.42640781e-01 4.37772959e-01
3.15604448e-01 -9.38363791e-01 1.06461298e+00 5.99648833e-01
-1.28885937e+00 -1.14808643e+00 -9.52609837e-01 -7.50509262e-01
5.54619253e-01 -2.48984262e-01 8.00045192e-01 1.12900980e-01
-6.58504665e-01 2.98395783e-01 -3.78225058e-01 -4.47567582e-01
-2.10939944e-01 6.83121905e-02 -2.42679147e-03 -4.92891461e-01
2.96664268e-01 -1.46689332e+00 -5.15267491e-01 3.94294202e-01
-6.97096527e-01 -2.82937646e-01 7.12447047e-01 3.72465461e-01
7.12509453e-01 1.79635659e-02 6.04576111e-01 -8.11522007e-01
6.92783773e-01 -1.01685368e-01 -9.47528303e-01 1.23969354e-02
-5.72351456e-01 -3.83135043e-02 4.95840549e-01 -3.85696381e-01
-4.32098538e-01 8.17427561e-02 -2.44052649e-01 9.53679532e-02
-6.18666410e-01 3.70799929e-01 1.65594608e-01 -5.93197644e-01
6.23303771e-01 1.25255480e-01 1.20187245e-01 -1.07888401e-01
-3.49945307e-01 2.77323604e-01 1.46183267e-01 -2.78122962e-01
3.51568311e-01 5.85141599e-01 6.14385486e-01 -1.14157403e+00
3.81910503e-01 -2.90928125e-01 -5.96053898e-01 -7.53002048e-01
5.57124674e-01 -4.49133635e-01 -9.79264081e-01 7.90824175e-01
-1.20325851e+00 -5.50333381e-01 -1.11420117e-01 8.92249048e-01
-7.75548875e-01 4.11550924e-02 -5.90779483e-01 -1.14116693e+00
-1.76042795e-01 -9.71049130e-01 3.67102474e-01 4.24379796e-01
-2.49883711e-01 -1.13076341e+00 4.52550411e-01 -3.47620398e-01
5.69150150e-01 4.57359910e-01 2.33104840e-01 -3.48766059e-01
-5.32038629e-01 -8.14407319e-02 -1.04615986e-02 8.77041742e-02
-2.82063544e-01 4.88449216e-01 -8.38133216e-01 -2.18423426e-01
-1.50611058e-01 2.16103420e-01 8.97428155e-01 7.52003014e-01
7.16444731e-01 5.90262353e-01 -9.75165427e-01 3.59861195e-01
1.94112170e+00 1.51222363e-01 1.04847777e+00 4.77514863e-01
-1.75744534e-01 7.37423897e-01 3.73408198e-01 3.35062951e-01
-2.69103706e-01 5.05140781e-01 6.80072606e-01 1.76490545e-01
-1.51655883e-01 4.94403094e-01 2.65544802e-01 7.20858216e-01
-7.53981769e-01 -1.53409794e-01 -8.16068351e-01 5.53009212e-01
-1.63296449e+00 -1.08976531e+00 -7.19988465e-01 2.28151083e+00
2.23372221e-01 4.88876581e-01 1.70796841e-01 5.46053350e-01
1.04001236e+00 -6.01747632e-01 -1.46682363e-03 -3.40519786e-01
-4.28370923e-01 5.39284945e-01 5.79460621e-01 5.29881120e-01
-8.15583587e-01 5.93223751e-01 6.59753180e+00 5.91764390e-01
-1.36162066e+00 -1.00055419e-01 2.38321185e-01 3.88486415e-01
1.25905603e-01 1.56902879e-01 -8.11208010e-01 6.32532358e-01
1.08242786e+00 -7.24815354e-02 5.57067871e-01 1.70276567e-01
4.34190959e-01 -6.97154403e-01 -5.30161262e-01 9.77905452e-01
-1.47993356e-01 -1.39649832e+00 -3.46786499e-01 1.19275101e-01
4.85028386e-01 3.83849263e-01 -4.35181111e-01 -8.10702331e-03
-1.24583095e-01 -8.36344481e-01 2.94048756e-01 1.06905091e+00
7.27463961e-02 -5.31716645e-01 1.05882919e+00 2.08550379e-01
-1.26456630e+00 7.03619421e-02 -4.85130340e-01 -4.65176702e-01
1.68303028e-01 9.55345094e-01 -2.96423584e-01 6.33434653e-01
8.24627221e-01 8.01737070e-01 -4.07446086e-01 1.78472567e+00
-1.01467997e-01 5.79206347e-01 -8.62744689e-01 -5.48977852e-01
-1.01580039e-01 -5.00788033e-01 8.14996958e-01 1.33320880e+00
4.87775505e-01 -2.88783073e-01 -5.96751869e-01 1.49183786e+00
3.85563254e-01 1.25501797e-01 -9.66669619e-01 -1.63268715e-01
4.46134716e-01 1.69861245e+00 -1.63249063e+00 5.60167357e-02
7.82134235e-02 6.35212719e-01 4.21875976e-02 3.98804188e-01
-6.24075949e-01 -7.41026878e-01 2.98326820e-01 2.15484381e-01
-5.42957857e-02 -3.49390239e-01 -3.76429886e-01 -6.84818089e-01
-2.12102786e-01 4.57071746e-03 -3.59392583e-01 -7.35512316e-01
-1.01434100e+00 7.04847991e-01 1.70291677e-01 -1.02034020e+00
1.34931549e-01 -9.44942534e-01 -7.84584463e-01 9.20587838e-01
-9.72425580e-01 -6.34652019e-01 -3.42953980e-01 3.98757488e-01
-6.54370785e-02 -1.88919734e-02 8.53897154e-01 2.99672812e-01
-4.59826797e-01 -1.36111721e-01 1.45587534e-01 -1.09750986e-01
2.36767992e-01 -1.28366911e+00 -1.99543517e-02 6.23478115e-01
1.97025202e-02 4.67432201e-01 8.67790520e-01 -7.05269933e-01
-6.92026198e-01 -4.54232991e-01 8.30110669e-01 1.05546311e-01
7.24757075e-01 -2.05291823e-01 -6.80382013e-01 1.96975097e-01
5.07131636e-01 -1.61098346e-01 4.59538490e-01 -3.07107657e-01
4.21865731e-01 -1.32741481e-02 -1.06832254e+00 5.28453290e-01
9.38096762e-01 -2.83608381e-02 -3.68953168e-01 3.00250351e-02
-2.25550935e-01 1.05379775e-01 -1.10311997e+00 6.59469783e-01
6.60283804e-01 -1.46194744e+00 5.12061238e-01 5.26929135e-03
-1.58805862e-01 -3.09508741e-01 4.55409884e-01 -1.13599706e+00
1.01390645e-01 -4.95710194e-01 4.18601662e-01 1.19792414e+00
5.00763655e-01 -1.13472056e+00 7.41494715e-01 -1.84427984e-02
-4.98853251e-03 -7.64363468e-01 -1.06366205e+00 -9.98570442e-01
-2.95675341e-02 -1.86813354e-01 -2.75989085e-01 8.27692091e-01
5.71831703e-01 1.42653003e-01 5.62882900e-01 -1.40510634e-01
6.02169335e-01 -2.61018425e-01 4.36877102e-01 -1.76222932e+00
4.54258779e-03 -8.15225661e-01 -1.40387368e+00 -2.31773570e-01
2.18085065e-01 -6.38587236e-01 -1.96509391e-01 -1.75201726e+00
-1.25861242e-01 -3.27120841e-01 -4.25915033e-01 1.38036117e-01
4.69140083e-01 4.26078260e-01 -1.99217066e-01 -5.16105592e-02
-1.76800802e-01 -3.95330824e-02 9.59543586e-01 2.52916008e-01
-7.28819519e-02 9.67598930e-02 3.85623932e-01 4.22777325e-01
1.22077024e+00 -5.58740497e-01 -3.24302047e-01 4.02299166e-01
4.31638509e-01 -2.19784498e-01 5.04719436e-01 -2.02251530e+00
2.97912866e-01 4.75092709e-01 2.84607530e-01 -6.54267132e-01
3.60632420e-01 -1.08828950e+00 4.90189195e-01 9.84821737e-01
-1.47463694e-01 5.78250289e-02 3.64020288e-01 4.21687096e-01
-3.67717475e-01 -4.45355564e-01 7.68915534e-01 -1.76277488e-01
-4.83433038e-01 -1.29784271e-01 -9.28844869e-01 -3.85111779e-01
1.43816411e+00 -8.10357869e-01 -2.15108454e-01 -1.73956826e-01
-1.27094185e+00 -9.07600820e-02 6.12249553e-01 -4.50389981e-01
4.23797518e-01 -7.59094238e-01 -4.90258843e-01 5.70751205e-02
-2.74858087e-01 -6.44592583e-01 2.90618002e-01 1.48946297e+00
-1.08732033e+00 3.70359123e-01 -1.23749423e+00 -6.45561695e-01
-1.35264599e+00 3.09979200e-01 5.06211638e-01 -8.61010104e-02
-1.72624812e-01 6.43402636e-01 -5.13412535e-01 -1.19105488e-01
9.01222304e-02 -3.81865859e-01 -7.55179346e-01 -6.43146876e-03
4.64368276e-02 4.56956476e-01 1.61500603e-01 -4.20224935e-01
-3.62398803e-01 9.06099200e-01 6.12848997e-01 -4.12918329e-01
1.35033524e+00 -6.07153699e-02 -5.94617367e-01 6.32655203e-01
9.02936459e-01 -1.91519633e-01 -1.01514184e+00 8.63602459e-01
1.00934498e-01 8.33057165e-02 -1.20975256e-01 -6.31802499e-01
-9.72595930e-01 9.92017925e-01 9.58769679e-01 6.78160846e-01
1.06235588e+00 -4.05039430e-01 1.14872113e-01 4.93113101e-01
7.11015522e-01 -1.11988378e+00 -3.18838298e-01 4.76822764e-01
6.68578446e-01 -5.57269990e-01 -1.82565436e-01 -7.63009310e-01
-4.93391901e-02 1.61665189e+00 3.40148300e-01 -5.86949468e-01
8.81497800e-01 9.93977368e-01 -1.96459070e-01 -1.73562527e-01
-5.35213470e-01 -6.75050080e-01 -4.56809044e-01 1.05695128e+00
7.08905220e-01 -3.74968648e-02 -1.28039563e+00 2.99680114e-01
1.03232011e-01 3.78244668e-01 7.23569393e-01 7.87174523e-01
-5.93904614e-01 -1.44042122e+00 -1.95949554e-01 3.66825372e-01
-5.07011950e-01 8.79945457e-02 -7.63886511e-01 1.23005569e+00
1.74283266e-01 6.93617821e-01 -4.15622592e-02 -4.52443033e-01
2.26628438e-01 -7.38833398e-02 8.72130394e-01 -1.53672487e-01
-9.10697281e-01 -1.04208060e-01 -1.17674991e-01 -4.33168381e-01
-1.08608603e+00 -4.53298241e-01 -1.54840779e+00 -9.68288034e-02
-6.28779054e-01 4.61446792e-01 1.23391318e+00 8.17105472e-01
2.40395620e-01 7.34246254e-01 1.30339026e-01 -1.07161438e+00
4.49346572e-01 -9.85966802e-01 -1.09591627e+00 1.51382804e-01
-5.05219162e-01 -9.31807280e-01 -6.04528427e-01 -3.72045487e-02] | [8.078728675842285, 2.8183159828186035] |
74b04219-3a78-4e63-83bf-367b0bc6e67d | d-2-im-net-learning-detail-disentangled | 2012.06650 | null | https://arxiv.org/abs/2012.06650v2 | https://arxiv.org/pdf/2012.06650v2.pdf | D$^2$IM-Net: Learning Detail Disentangled Implicit Fields from Single Images | We present the first single-view 3D reconstruction network aimed at recovering geometric details from an input image which encompass both topological shape structures and surface features. Our key idea is to train the network to learn a detail disentangled reconstruction consisting of two functions, one implicit field representing the coarse 3D shape and the other capturing the details. Given an input image, our network, coined D$^2$IM-Net, encodes it into global and local features which are respectively fed into two decoders. The base decoder uses the global features to reconstruct a coarse implicit field, while the detail decoder reconstructs, from the local features, two displacement maps, defined over the front and back sides of the captured object. The final 3D reconstruction is a fusion between the base shape and the displacement maps, with three losses enforcing the recovery of coarse shape, overall structure, and surface details via a novel Laplacian term. | ['Hao Zhang', 'Manyi Li'] | 2020-12-11 | null | null | null | null | ['single-view-3d-reconstruction'] | ['computer-vision'] | [ 3.38621497e-01 4.45852101e-01 2.15766251e-01 -4.02772933e-01
-9.47453499e-01 -5.67733586e-01 5.91507554e-01 -8.95824432e-02
-6.13171756e-02 4.81524259e-01 3.47160518e-01 2.42223069e-01
-2.91905291e-02 -1.06891739e+00 -1.06024003e+00 -9.75694239e-01
1.47429287e-01 7.20773757e-01 -7.58807063e-02 7.61420727e-02
3.52546722e-01 6.24246359e-01 -1.12646353e+00 2.60871768e-01
4.04763550e-01 1.29172492e+00 1.45207852e-01 6.01130247e-01
9.10505280e-02 6.84638262e-01 8.61733332e-02 -4.15892571e-01
4.20654416e-01 -1.40269920e-01 -6.33692384e-01 1.83205023e-01
5.90847611e-01 -5.62430620e-01 -7.15686440e-01 9.16672111e-01
2.21623167e-01 -2.19727650e-01 7.96634197e-01 -5.56786239e-01
-7.98012197e-01 2.02271957e-02 -4.27212864e-01 -2.25209519e-01
4.13674057e-01 1.76129743e-01 1.20783198e+00 -1.04733503e+00
1.01042140e+00 1.16687524e+00 6.32781625e-01 3.50842625e-01
-1.72096992e+00 -2.95015365e-01 -7.50493258e-02 -3.12638730e-01
-1.11782503e+00 -4.68936801e-01 1.20577765e+00 -7.43478417e-01
7.12133944e-01 2.30038026e-03 6.48237824e-01 9.51331258e-01
3.80216032e-01 4.16957974e-01 1.17255569e+00 -3.24911028e-02
1.29638448e-01 -8.39350745e-02 -9.35087651e-02 1.08028376e+00
1.34351984e-01 3.86646688e-01 -5.73528469e-01 -1.71925679e-01
1.34700310e+00 4.08924460e-01 -6.28973305e-01 -7.91401923e-01
-1.13718081e+00 7.35170484e-01 6.28525913e-01 9.54694673e-02
-4.61273611e-01 1.59119070e-01 -1.34144410e-01 1.69515863e-01
8.03803146e-01 1.89422727e-01 -2.68462926e-01 2.75822818e-01
-6.74065351e-01 3.08748007e-01 6.40910804e-01 6.42038405e-01
1.37348974e+00 -2.87841678e-01 1.08489037e-01 2.97390282e-01
5.41808188e-01 7.62851000e-01 -2.35483021e-01 -1.10480082e+00
5.52395225e-01 9.30215657e-01 9.05730203e-02 -1.18388844e+00
-7.89020769e-03 -2.77189106e-01 -1.08682942e+00 4.19289529e-01
1.52289405e-01 1.94741890e-01 -9.75567043e-01 1.96400797e+00
6.40914589e-02 -7.19076321e-02 -2.95379996e-01 8.34604919e-01
7.24093676e-01 4.70241457e-01 -5.45399189e-01 2.48035043e-01
1.07772994e+00 -4.54712033e-01 -3.22698534e-01 -2.62025386e-01
-1.39826983e-01 -4.87145692e-01 3.53624195e-01 -1.51538383e-02
-1.65232861e+00 -3.66677165e-01 -1.12715530e+00 -5.83860993e-01
-8.64124522e-02 1.14768773e-01 2.82739010e-02 -2.02703655e-01
-1.18420446e+00 8.15780997e-01 -8.98694277e-01 1.92772284e-01
5.91636717e-01 3.75125259e-01 -8.43249023e-01 -2.44070113e-01
-7.14037001e-01 6.72524631e-01 -2.35655203e-01 1.10126309e-01
-1.13073850e+00 -7.72939563e-01 -1.18571591e+00 1.25815347e-01
6.77856356e-02 -1.30476201e+00 7.94557571e-01 -5.03309369e-01
-1.42785156e+00 1.33261907e+00 -2.29808643e-01 1.44630373e-01
4.90811229e-01 3.09620827e-01 2.50305086e-01 3.29410762e-01
2.98128098e-01 4.03072059e-01 1.11258316e+00 -1.61514020e+00
-1.52271226e-01 -9.56162572e-01 -1.85622852e-02 1.51838019e-01
5.26275456e-01 -9.36727524e-01 -4.19862658e-01 -5.37736356e-01
5.53249061e-01 -5.42876184e-01 -9.70637519e-03 5.13960540e-01
-6.09790325e-01 4.92816240e-01 5.36159754e-01 -1.08249521e+00
6.49541080e-01 -2.30493307e+00 1.03926635e+00 3.39040935e-01
8.44566107e-01 -4.32333499e-01 -3.08487505e-01 3.24491709e-01
-1.33894831e-01 5.66787943e-02 -6.87724233e-01 -7.64906347e-01
-5.65079115e-02 6.64247721e-02 -2.47256726e-01 7.24948347e-01
4.73728150e-01 1.22172153e+00 -8.03732693e-01 2.57153243e-01
2.80723810e-01 8.31973672e-01 -7.42269516e-01 5.77741146e-01
-2.38411173e-01 8.41237962e-01 -7.05511987e-01 2.06611544e-01
8.87719750e-01 -3.90022457e-01 5.25746495e-02 -3.98865730e-01
-2.13023037e-01 5.02901316e-01 -8.59755158e-01 2.11892772e+00
-5.18144608e-01 3.35453868e-01 7.18719065e-01 -1.05989277e+00
9.12842512e-01 3.42757016e-01 5.75898051e-01 -6.55059099e-01
1.37910962e-01 3.16421330e-01 -6.66498542e-01 -3.39614511e-01
7.61412084e-02 -5.67344666e-01 -5.98399267e-02 4.67206925e-01
2.77963459e-01 -5.01046479e-01 -6.75124764e-01 2.88068205e-01
1.17773843e+00 2.57061034e-01 -8.67411569e-02 -2.76654251e-02
5.81478238e-01 -5.37466168e-01 1.64201140e-01 1.00389607e-01
3.88810068e-01 1.05299628e+00 7.61622369e-01 -7.10851192e-01
-1.34065866e+00 -1.53540528e+00 3.61342654e-02 2.05375418e-01
2.41721243e-01 -6.61280453e-02 -5.33732414e-01 -6.61194026e-01
2.95474142e-01 2.47136891e-01 -9.46797192e-01 -3.84324670e-01
-6.23122215e-01 1.03953153e-01 8.82826149e-02 4.03183028e-02
5.09131253e-01 -8.82798910e-01 -4.75238174e-01 -2.51298230e-02
-3.25454354e-01 -9.46725488e-01 -8.92126977e-01 1.24886051e-01
-1.01693368e+00 -1.13165951e+00 -6.68535471e-01 -6.86294198e-01
9.87246513e-01 2.48895824e-01 1.20804918e+00 1.62635982e-01
-2.75377512e-01 2.30105579e-01 2.50397831e-01 4.39233869e-01
-2.43127838e-01 -2.21477881e-01 -3.65439624e-01 5.94948292e-01
-3.95243496e-01 -1.19235134e+00 -6.23404801e-01 -1.75591856e-02
-9.23143625e-01 4.75783795e-01 6.20820999e-01 9.01343703e-01
1.02672434e+00 -3.78540993e-01 -2.09935717e-02 -8.06847513e-01
-5.46646416e-02 -3.49031508e-01 -3.91233474e-01 -3.50113325e-02
-1.57392696e-01 5.31403065e-01 4.92252231e-01 1.62967190e-01
-8.52559566e-01 1.32393315e-01 -2.96664953e-01 -5.78142643e-01
-8.88516828e-02 1.49338424e-01 -5.37468672e-01 -1.57053992e-01
3.09801191e-01 5.02608001e-01 3.14153880e-01 -9.08235431e-01
3.96261483e-01 1.56207606e-01 4.92350221e-01 -4.83820200e-01
8.39338839e-01 8.58285248e-01 7.02055991e-02 -5.96958458e-01
-9.15475011e-01 -1.78167164e-01 -8.54649186e-01 1.11396663e-01
1.09720457e+00 -9.45219159e-01 -7.62690246e-01 4.94484276e-01
-1.28215110e+00 -2.47780234e-01 -6.59198880e-01 1.11672208e-01
-9.49125171e-01 8.25517699e-02 -6.49425149e-01 -1.97128043e-01
-3.27046722e-01 -1.16316497e+00 1.75494313e+00 -3.55564982e-01
2.74583519e-01 -9.85572159e-01 1.82625636e-01 3.06287020e-01
3.77048790e-01 6.68701530e-01 1.35407567e+00 1.42872274e-01
-1.32164037e+00 -1.14291318e-01 -3.97485882e-01 3.45628440e-01
1.30195200e-01 -4.54365432e-01 -9.69293892e-01 -2.95507848e-01
4.69076782e-01 -2.05199748e-01 9.73574221e-01 4.78453606e-01
8.50248694e-01 -3.66819412e-01 -1.47872820e-01 1.15846968e+00
1.71694243e+00 -2.88238883e-01 5.02246857e-01 -3.70353818e-01
9.22489107e-01 3.21079046e-01 -4.23256338e-01 2.22639889e-01
5.70517600e-01 5.07320166e-01 7.40121305e-01 -1.30778812e-02
-3.82171214e-01 -7.72979677e-01 3.38683516e-01 9.61437762e-01
-2.66308278e-01 7.33943880e-02 -5.81211925e-01 3.03480834e-01
-1.50081539e+00 -8.50336194e-01 1.94033533e-01 2.33963108e+00
5.78355849e-01 2.34284401e-02 -3.73389304e-01 -1.79820150e-01
3.58319014e-01 5.61419427e-01 -8.73854756e-01 6.98463246e-02
-5.47197387e-02 3.26605350e-01 1.28288120e-01 1.01634443e+00
-6.09566629e-01 4.56830293e-01 5.69449997e+00 3.16948563e-01
-1.29255605e+00 -9.35564637e-02 4.89963293e-01 3.27200443e-02
-1.09112883e+00 1.25589877e-01 -2.58558422e-01 2.17333317e-01
3.85614038e-01 1.66834712e-01 6.95710540e-01 1.70229018e-01
-5.52605391e-02 2.01269269e-01 -1.34059608e+00 9.25046384e-01
1.33415818e-01 -1.74246418e+00 4.22942817e-01 3.04536909e-01
5.66811621e-01 1.44069806e-01 1.05882473e-01 -2.58486629e-01
1.10929981e-01 -1.03001130e+00 1.02211964e+00 9.87806082e-01
1.47851205e+00 -4.42018270e-01 3.21402490e-01 6.78136289e-01
-1.07681441e+00 3.62961978e-01 1.10671893e-02 -2.29550246e-02
2.78524280e-01 5.95389903e-01 -2.32601658e-01 7.67822027e-01
3.44163150e-01 1.14469731e+00 2.03560274e-02 2.79179454e-01
-2.24356309e-01 7.20070228e-02 -2.51381341e-02 7.05130816e-01
-8.79091471e-02 -6.76381707e-01 9.54202950e-01 5.76934874e-01
1.97274998e-01 3.17686141e-01 -3.87782939e-02 1.57096946e+00
-4.08780992e-01 -5.36201239e-01 -8.74924779e-01 1.46198988e-01
3.72776762e-02 1.16097426e+00 -3.89362693e-01 -2.20811144e-01
-2.85490721e-01 1.13840199e+00 7.67135203e-01 5.32856107e-01
-3.42032462e-01 2.28063650e-02 8.48125160e-01 4.55605835e-01
4.69767958e-01 -4.47607458e-01 -6.28322661e-01 -1.63500321e+00
2.35414997e-01 -5.53291261e-01 -2.10384041e-01 -8.27491522e-01
-1.35652626e+00 5.02188206e-01 -3.22993636e-01 -9.35199678e-01
1.14886537e-01 -6.28169179e-01 -4.58611101e-01 1.37122095e+00
-1.43289542e+00 -1.20238984e+00 -2.39863321e-01 4.50564831e-01
2.12670237e-01 2.37301156e-01 7.78926909e-01 2.05375195e-01
-2.09781483e-01 7.66630620e-02 -1.75706193e-01 1.47043392e-01
1.78364247e-01 -1.28364182e+00 5.63980937e-01 3.69439065e-01
-2.35254839e-02 5.37484884e-01 1.49127632e-01 -5.86497009e-01
-1.81011033e+00 -8.65155101e-01 1.05942690e+00 -6.02320850e-01
1.33273333e-01 -6.60704136e-01 -7.38215685e-01 9.07078743e-01
-9.92595106e-02 3.14424038e-01 6.37713373e-02 -5.65190077e-01
-6.45981133e-01 1.12860978e-01 -1.27677226e+00 1.44465923e-01
1.17113614e+00 -1.00046289e+00 -5.11592329e-01 -1.60605103e-01
7.86490977e-01 -6.00892723e-01 -9.91038442e-01 1.55903056e-01
5.83461404e-01 -1.14441836e+00 1.14615011e+00 -2.99787343e-01
8.81609201e-01 -1.70045078e-01 -5.20046175e-01 -1.37512553e+00
-5.73726177e-01 -4.84237880e-01 -1.53655216e-01 6.31641030e-01
3.90399635e-01 -5.49468935e-01 8.09222877e-01 4.03133094e-01
-1.52579218e-01 -1.03130734e+00 -1.00813365e+00 -8.34612921e-02
1.30130127e-01 -1.42769054e-01 4.80859488e-01 6.36692524e-01
-5.90130448e-01 7.59399891e-01 -2.65989721e-01 9.46434811e-02
9.24808145e-01 5.16309202e-01 3.48477930e-01 -1.21438062e+00
-3.35025281e-01 -3.04849386e-01 -3.45850378e-01 -1.54530430e+00
5.68689220e-02 -1.34590232e+00 1.51176136e-02 -1.38165367e+00
4.51944530e-01 -2.61798680e-01 2.68069804e-01 2.67205596e-01
1.81099370e-01 2.90015429e-01 1.89041838e-01 1.58367142e-01
-5.79482876e-02 9.73788679e-01 1.81260335e+00 -1.05286032e-01
2.74571944e-02 -1.00580491e-01 -7.69701242e-01 7.40000546e-01
6.10168427e-02 -3.70757073e-01 2.70900708e-02 -8.95211279e-01
4.04886067e-01 6.87217653e-01 8.83223832e-01 -4.78418112e-01
8.73347446e-02 1.38202727e-01 5.78415573e-01 -3.51741910e-01
5.24695337e-01 -9.48993325e-01 3.78346831e-01 2.32172653e-01
-3.79684150e-01 -8.55483711e-02 -2.43013099e-01 7.52337158e-01
-1.80596769e-01 3.25358719e-01 8.39495420e-01 -2.72239178e-01
1.84905306e-02 1.03381646e+00 3.90653044e-01 1.31516755e-01
5.52141070e-01 -9.03223380e-02 3.02523840e-02 -4.02116746e-01
-1.00346339e+00 -1.76032960e-01 1.03532779e+00 1.31532490e-01
1.06882191e+00 -1.48618901e+00 -8.14055085e-01 9.92493927e-01
-2.26685390e-01 6.33742452e-01 3.99298131e-01 6.50869966e-01
-6.00178480e-01 2.52109557e-01 -1.97837174e-01 -5.74573457e-01
-6.91090643e-01 3.29481691e-01 6.09273076e-01 -4.02050227e-01
-1.00700128e+00 7.96825230e-01 7.42453933e-01 -7.82767773e-01
-5.84375747e-02 -4.66420382e-01 1.83935225e-01 -2.92798430e-01
2.77712107e-01 1.81634679e-01 -7.80513743e-03 -9.04444396e-01
-1.76605493e-01 1.21068990e+00 3.74694169e-01 -3.51837277e-01
1.80011940e+00 -1.11268051e-01 -5.30145526e-01 4.45818007e-01
1.86726499e+00 1.42734379e-01 -1.79117560e+00 -4.54576463e-01
-3.98105741e-01 -5.07453978e-01 9.19434726e-02 -6.86438918e-01
-1.29921472e+00 1.12782037e+00 9.11929086e-02 1.44647826e-02
1.01549494e+00 3.19558024e-01 9.07055259e-01 -2.20537335e-01
6.08802080e-01 -1.31135494e-01 1.48632944e-01 6.44189239e-01
1.21479714e+00 -8.81791949e-01 -1.23678580e-01 -3.60725492e-01
-1.17653690e-01 1.04032552e+00 1.07752085e-01 -7.90673852e-01
9.02170300e-01 1.56095222e-01 -4.59687710e-01 -7.82804430e-01
-7.81728506e-01 1.49455845e-01 7.00890303e-01 3.08338970e-01
1.25442043e-01 9.01877508e-02 2.30500251e-01 5.17185032e-01
-2.88562663e-02 -8.91366154e-02 9.58030373e-02 5.59646606e-01
-2.48207539e-01 -9.42714393e-01 -9.13168304e-03 4.36775655e-01
-1.68032661e-01 1.97662890e-01 -5.06086767e-01 4.20233369e-01
4.56923872e-01 1.37728646e-01 3.54104817e-01 -4.82777923e-01
6.55711114e-01 -1.40963972e-01 7.22207487e-01 -8.16394448e-01
-2.27707028e-01 -2.05814406e-01 -2.93679029e-01 -8.60251546e-01
-1.61251098e-01 -4.89666939e-01 -1.17526865e+00 -2.88265318e-01
1.34772152e-01 -1.09764241e-01 5.98287284e-01 8.87864053e-01
5.09521484e-01 3.14741641e-01 9.47749555e-01 -1.45724559e+00
-5.30746102e-01 -6.25797749e-01 -7.70846784e-01 6.57120585e-01
1.11327100e+00 -5.72857559e-01 -6.11866951e-01 8.95294696e-02] | [8.852066993713379, -3.486140251159668] |
eb2199c8-5794-49bc-88ab-d45e56dfa6b7 | multi-frequency-image-reconstruction-for | 1703.03608 | null | http://arxiv.org/abs/1703.03608v1 | http://arxiv.org/pdf/1703.03608v1.pdf | Multi-frequency image reconstruction for radio-interferometry with self-tuned regularization parameters | As the world's largest radio telescope, the Square Kilometer Array (SKA) will
provide radio interferometric data with unprecedented detail. Image
reconstruction algorithms for radio interferometry are challenged to scale well
with TeraByte image sizes never seen before. In this work, we investigate one
such 3D image reconstruction algorithm known as MUFFIN (MUlti-Frequency image
reconstruction For radio INterferometry). In particular, we focus on the
challenging task of automatically finding the optimal regularization parameter
values. In practice, finding the regularization parameters using classical grid
search is computationally intensive and nontrivial due to the lack of ground-
truth. We adopt a greedy strategy where, at each iteration, the optimal
parameters are found by minimizing the predicted Stein unbiased risk estimate
(PSURE). The proposed self-tuned version of MUFFIN involves parallel and
computationally efficient steps, and scales well with large- scale data.
Finally, numerical results on a 3D image are presented to showcase the
performance of the proposed approach. | ['Rémi Flamary', 'André Ferrari', 'Chiara Ferrari', 'Rita Ammanouil', 'David Mary'] | 2017-03-10 | null | null | null | null | ['radio-interferometry'] | ['miscellaneous'] | [ 1.54089063e-01 -1.38275757e-01 4.02827054e-01 1.79311067e-01
-1.15466964e+00 -3.07131797e-01 2.62933046e-01 -6.75870717e-01
-6.22749925e-01 9.30728853e-01 1.38098402e-02 -4.51603800e-01
-4.44975138e-01 -4.74230975e-01 -4.66392905e-01 -8.74795794e-01
-1.65865585e-01 5.97721756e-01 -1.11092143e-01 1.72655806e-01
3.89019668e-01 5.32873571e-01 -1.00668657e+00 -7.43224740e-01
9.64198470e-01 1.03122735e+00 4.17007744e-01 5.43461502e-01
4.92240459e-01 2.30722100e-01 -1.29056036e-01 4.28550914e-02
6.79944575e-01 -6.90401435e-01 -5.63198030e-01 3.66602957e-01
3.28061551e-01 -3.47084582e-01 -2.58502904e-02 1.25585783e+00
5.88680923e-01 1.52029516e-02 7.02565491e-01 -5.56314766e-01
1.44456662e-02 7.36659914e-02 -1.11282182e+00 3.50365520e-01
1.01537928e-01 6.45605847e-02 7.27955878e-01 -1.21189547e+00
4.99165297e-01 7.64132380e-01 7.43003607e-01 -6.78556487e-02
-1.22641587e+00 -5.41955292e-01 -7.31854439e-01 -7.61461183e-02
-1.90472591e+00 -3.04392576e-01 6.15347028e-01 -5.29048324e-01
5.72946012e-01 5.90055762e-03 4.24324572e-01 3.31159383e-01
3.88173848e-01 5.22183068e-02 1.51813412e+00 -5.22014201e-01
1.20111145e-02 -8.82776454e-02 -2.43715420e-01 8.47443223e-01
5.89045584e-01 3.35304409e-01 -3.37627143e-01 -4.05677408e-01
1.16283619e+00 -2.65408397e-01 -5.12168705e-01 -2.32875377e-01
-1.17249441e+00 8.27275097e-01 3.32776010e-01 3.73171359e-01
-5.27311921e-01 1.83829978e-01 -9.84106064e-02 2.49180183e-01
7.21781433e-01 6.60570323e-01 -1.43133685e-01 1.88797623e-01
-1.15184426e+00 1.54890642e-01 4.05251920e-01 6.05963469e-01
1.16400158e+00 1.90844208e-01 3.12686533e-01 5.62091768e-01
2.58539975e-01 8.47427070e-01 3.29858840e-01 -9.13746297e-01
2.41437390e-01 -3.44143733e-02 5.28609335e-01 -1.03117466e+00
-6.61510050e-01 -8.06011856e-01 -1.25959742e+00 2.34702721e-01
4.71452653e-01 -4.40763921e-01 -5.64347506e-01 1.24375188e+00
4.01029527e-01 5.78270018e-01 1.84153378e-01 1.12222970e+00
6.05560616e-02 4.65411276e-01 -4.87161398e-01 -5.08287668e-01
1.42870116e+00 -3.70495260e-01 -4.23953593e-01 -4.40055251e-01
5.74719131e-01 -1.05543876e+00 6.37346566e-01 2.48950616e-01
-7.80999124e-01 -3.02437276e-01 -9.93732393e-01 3.41953129e-01
5.70115507e-01 2.81334966e-01 5.03177047e-01 4.89725977e-01
-9.59253013e-01 5.34836471e-01 -5.29086649e-01 -1.99462846e-01
2.82671154e-01 9.14641246e-02 -2.84872025e-01 -4.41205129e-02
-8.67709935e-01 8.50204289e-01 3.23960245e-01 1.91512227e-01
-5.11968076e-01 -6.59231901e-01 -5.79273880e-01 -2.06567481e-01
4.57809269e-01 -8.43750238e-01 9.25322592e-01 -4.81264174e-01
-1.58924961e+00 8.78345430e-01 -1.01916514e-01 -7.82181680e-01
4.41091299e-01 -3.46544713e-01 1.17758930e-01 2.79996634e-01
2.56572902e-01 1.41761854e-01 9.89623845e-01 -8.94389987e-01
-3.59030366e-01 -4.65451598e-01 -4.91915137e-01 2.60312438e-01
3.28916788e-01 -1.43404409e-01 -2.76370138e-01 -4.51492071e-01
7.50670612e-01 -1.04888523e+00 -6.54792249e-01 -2.88243055e-01
-2.72459626e-01 3.88705879e-01 3.74518335e-01 -5.76725721e-01
9.68952119e-01 -2.21895146e+00 3.80829945e-02 2.94680387e-01
8.39492083e-02 -9.23753306e-02 1.65297776e-01 1.33179188e-01
1.24944806e-01 -1.31492451e-01 -5.56083381e-01 -1.23270117e-01
-6.02190077e-01 -2.36071110e-01 -1.47971272e-01 1.02859616e+00
-8.44831392e-02 6.34050429e-01 -5.99779308e-01 -3.97641659e-01
2.71573126e-01 1.86293840e-01 -6.71064258e-01 8.30696225e-02
2.31866539e-01 1.00595593e+00 -7.45998323e-01 3.57506633e-01
9.88107860e-01 -6.79361045e-01 2.02086791e-01 -3.84943932e-01
-4.76649642e-01 -1.71311647e-01 -1.20924652e+00 1.55424273e+00
-8.11670661e-01 5.16008914e-01 2.95367390e-01 -9.48805034e-01
8.90402377e-01 2.20988899e-01 4.52554017e-01 -5.28865755e-01
1.96621552e-01 5.33011079e-01 -1.31347954e-01 -2.95794159e-01
4.55011010e-01 -7.73673058e-01 -2.69871056e-01 4.76457983e-01
-3.38266969e-01 -6.53056085e-01 -3.78807366e-01 -3.24245878e-02
1.04875338e+00 -2.08050907e-01 1.11927629e+00 -5.16350985e-01
6.07649267e-01 1.71494707e-01 4.56343442e-01 9.20735955e-01
-1.12756370e-02 6.97438180e-01 8.23784992e-02 -2.51123220e-01
-1.25766337e+00 -9.24656272e-01 -3.38507086e-01 -1.38485178e-01
2.58389413e-01 1.08597785e-01 -4.52478707e-01 -8.89360383e-02
-1.48524433e-01 2.60748714e-01 -1.82383239e-01 5.03688633e-01
-6.32268310e-01 -1.18446577e+00 2.38486260e-01 -2.61704326e-01
9.26506281e-01 -6.02079928e-01 -6.73174739e-01 2.18337446e-01
-3.09151262e-01 -1.34726274e+00 -3.91289085e-01 -2.46255100e-01
-1.06655300e+00 -1.06678438e+00 -7.88635433e-01 -4.25658941e-01
6.29645944e-01 7.46550858e-01 1.03144729e+00 -9.83398482e-02
-5.54940879e-01 1.32696167e-01 -6.04542457e-02 2.09277585e-01
-1.08387031e-01 -2.71738499e-01 1.99839145e-01 1.29449368e-01
-2.22857714e-01 -9.40550864e-01 -7.38273323e-01 4.76210505e-01
-6.72061443e-01 1.28429070e-01 1.04880452e+00 9.41669703e-01
8.69599342e-01 1.14115387e-01 4.60759997e-01 -9.52704847e-01
3.38336289e-01 -4.32922095e-01 -1.29020977e+00 -1.34117678e-01
-4.59718376e-01 1.39204726e-01 4.84846085e-01 8.63022078e-03
-1.02961445e+00 1.84887663e-01 1.14239477e-01 -2.25567192e-01
2.09524080e-01 6.53743863e-01 3.08553517e-01 -6.66501701e-01
7.17134237e-01 3.02368909e-01 7.20534846e-02 -3.76911044e-01
3.19682896e-01 5.33440530e-01 6.22289896e-01 -1.42979041e-01
1.31596053e+00 8.09765041e-01 4.86189425e-01 -1.21770573e+00
-9.78378117e-01 -7.36818075e-01 -3.01713258e-01 -1.82501838e-01
7.87073553e-01 -1.19578445e+00 -6.74113035e-01 3.69337946e-01
-9.17162836e-01 -7.48459026e-02 3.30873840e-02 9.33197796e-01
-7.53093481e-01 6.99813366e-01 -1.97581485e-01 -9.05209184e-01
-4.25347060e-01 -7.88768649e-01 9.17363107e-01 1.84697777e-01
1.89987943e-01 -6.38779521e-01 1.87526226e-01 3.49362075e-01
4.06421274e-01 1.67915121e-01 4.68578011e-01 1.80210739e-01
-1.14242876e+00 -1.34113163e-01 -5.74600875e-01 1.58989504e-01
-9.53343324e-03 -8.57209861e-01 -5.74652016e-01 -3.32789123e-01
6.87608600e-01 -2.28493720e-01 6.85670316e-01 9.04714882e-01
8.36626887e-01 -6.82993233e-02 -1.36993304e-01 9.30764616e-01
1.79385829e+00 -2.44176075e-01 6.83620632e-01 1.79063290e-01
3.06228399e-01 1.74357444e-01 9.30311322e-01 7.77738988e-01
-1.09661289e-01 8.49619091e-01 3.05409193e-01 1.45689949e-01
-1.07832076e-02 1.26205653e-01 1.04927747e-02 8.46542001e-01
-2.20692858e-01 -5.04960120e-03 -7.33394027e-01 3.50267380e-01
-1.57828391e+00 -9.06016886e-01 -1.47049814e-01 2.46382284e+00
3.22006553e-01 -1.09117188e-01 -3.66521090e-01 -2.01922089e-01
4.74302918e-01 2.13736922e-01 -3.89731526e-01 3.49052429e-01
-1.95773706e-01 2.31818438e-01 1.06809831e+00 8.29772770e-01
-1.02685452e+00 8.30310643e-01 6.07182503e+00 9.89924788e-01
-8.34494710e-01 2.59264439e-01 3.57610643e-01 7.20870346e-02
-1.05693929e-01 2.21557260e-01 -7.58352220e-01 1.53114974e-01
6.62506104e-01 -3.75411749e-01 6.50794625e-01 4.60959166e-01
5.25750697e-01 -7.51015067e-01 -1.92573741e-01 1.32449722e+00
-2.30364241e-02 -1.48590815e+00 -3.69540989e-01 3.13139826e-01
8.36698949e-01 3.22583139e-01 -9.24065486e-02 -2.94468433e-01
7.91080669e-02 -6.61401570e-01 1.93098843e-01 4.89385158e-01
8.03074777e-01 -8.46667528e-01 6.18872344e-01 4.93447065e-01
-1.10224962e+00 1.35621101e-01 -3.71068418e-01 -9.51741487e-02
5.62120259e-01 1.22989607e+00 -9.71481323e-01 8.70891213e-01
4.30554956e-01 4.47921902e-01 -1.65774077e-01 1.22878122e+00
-1.55976713e-01 4.00284976e-01 -5.96072614e-01 1.30067095e-01
2.62619913e-01 -9.53448832e-01 8.43736708e-01 6.05345547e-01
1.14643705e+00 4.64618802e-01 1.76166803e-01 4.38002020e-01
-2.68505737e-02 -7.16824979e-02 -9.09765363e-01 3.02256167e-01
4.70772833e-01 1.21066248e+00 -7.51377344e-01 -1.04779061e-02
-4.31962192e-01 9.55178261e-01 1.17688507e-01 3.24395508e-01
-5.08828938e-01 -1.85611397e-02 5.94847500e-01 1.99554861e-01
2.63703406e-01 -6.38226867e-01 -2.39813581e-01 -1.09758592e+00
-1.31008074e-01 -6.01132929e-01 2.28725389e-01 -7.27575004e-01
-1.15161586e+00 6.06860459e-01 -1.14958242e-01 -1.59301579e+00
-2.67857999e-01 -3.69428188e-01 -3.35127831e-01 1.04097617e+00
-1.57497752e+00 -6.77119315e-01 -2.48290375e-01 2.37489000e-01
3.25168043e-01 -3.13987248e-02 5.33723831e-01 1.67740554e-01
-2.06492469e-01 2.84096599e-02 4.70752537e-01 -1.89528853e-01
4.28946704e-01 -7.18515515e-01 3.16546857e-01 1.02154052e+00
-5.81149124e-02 2.21416876e-01 1.03821862e+00 -6.84556723e-01
-1.50135279e+00 -8.52211058e-01 6.66585326e-01 2.27203041e-01
8.64044189e-01 4.07777056e-02 -6.74569607e-01 5.08632898e-01
-6.42395392e-02 2.35929430e-01 2.68799990e-01 -1.68994337e-01
9.67932045e-02 4.11243103e-02 -1.23693728e+00 2.61788726e-01
9.04809952e-01 -4.79991883e-01 -1.72103718e-01 5.36609054e-01
2.30760723e-01 -3.51120800e-01 -5.58053434e-01 4.99898702e-01
3.49626303e-01 -1.14262569e+00 1.14577901e+00 1.97488517e-01
8.16124678e-02 -6.26986682e-01 -7.11295754e-02 -1.29463780e+00
-1.51161626e-01 -9.88838434e-01 4.55002815e-01 4.44846153e-01
2.34058529e-01 -8.57504904e-01 8.67676675e-01 -2.61950225e-01
-8.08641966e-03 -3.94481897e-01 -1.33251286e+00 -1.01391900e+00
-3.69572520e-01 -2.93676138e-01 -1.45734064e-02 5.71947753e-01
-3.97420555e-01 4.39395934e-01 -8.81662726e-01 7.84236789e-01
1.40701926e+00 5.37440956e-01 8.37681770e-01 -1.09175241e+00
-6.53421521e-01 5.34957089e-02 -5.04605889e-01 -1.27230084e+00
-1.08577296e-01 -7.16144681e-01 2.58653939e-01 -9.62528944e-01
2.32387781e-01 -5.44513047e-01 3.20393175e-01 -4.82997566e-01
-2.39331499e-01 6.09784663e-01 -1.26909120e-02 3.93315077e-01
-1.70663565e-01 4.36001301e-01 1.21575880e+00 2.92339802e-01
-9.27395746e-02 2.84585476e-01 -3.17953795e-01 8.64436746e-01
8.39305758e-01 -6.43036187e-01 -1.91258684e-01 -4.26133513e-01
4.03378695e-01 6.51618600e-01 5.19634843e-01 -1.17166960e+00
1.51895866e-01 -1.24849007e-03 4.98639084e-02 -7.05543160e-01
4.16611880e-01 -6.30849123e-01 5.31407058e-01 5.49888253e-01
4.24134851e-01 -1.34067819e-01 -7.19599426e-03 7.05704629e-01
-2.47542888e-01 -4.02000338e-01 1.32717681e+00 -1.64329335e-01
-4.83144939e-01 2.84903318e-01 -3.52307379e-01 6.56699836e-02
7.68085718e-01 1.02003299e-01 4.56026085e-02 -5.21859288e-01
-5.17117500e-01 -5.33312447e-02 3.94115537e-01 -4.66451019e-01
6.63923740e-01 -1.13670695e+00 -9.42201972e-01 1.43993542e-01
-7.67641738e-02 -1.50958583e-01 5.52674174e-01 9.70619678e-01
-7.97736287e-01 3.44528794e-01 1.78406805e-01 -5.21434903e-01
-1.07724690e+00 1.40573844e-01 3.86734247e-01 -5.94739914e-01
-8.60678971e-01 8.13577652e-01 2.66445756e-01 -2.17756748e-01
-4.80190188e-01 -1.38754835e-02 1.06093504e-01 -1.27215236e-01
5.38126111e-01 2.69212514e-01 -1.61839247e-01 -4.83516127e-01
-1.79832831e-01 1.02986193e+00 2.02127442e-01 -4.59771723e-01
1.27934146e+00 -5.08210182e-01 -2.60051310e-01 -1.47972908e-02
1.15191615e+00 1.73682049e-01 -1.18668270e+00 -4.33671564e-01
3.26160267e-02 -7.32196391e-01 3.70538890e-01 -1.44687310e-01
-7.20839262e-01 4.91408557e-01 6.99898660e-01 -3.01092546e-02
1.26860034e+00 3.05859409e-02 7.00882494e-01 5.03950417e-01
7.34969258e-01 -5.78108490e-01 -1.31726131e-01 4.77907777e-01
7.65737116e-01 -1.26597273e+00 5.35684168e-01 -5.84053695e-01
-2.51389205e-01 8.06344569e-01 4.57601659e-02 -4.19959366e-01
6.91397667e-01 7.81691000e-02 4.57186997e-02 -2.52141625e-01
-1.42805159e-01 -3.27306658e-01 5.50798550e-02 1.34890825e-01
1.47361040e-01 1.62362918e-01 -6.59792125e-01 -1.15554184e-01
-1.66851461e-01 6.89801797e-02 6.74744606e-01 6.37617946e-01
-7.39954591e-01 -9.62528646e-01 -6.99449420e-01 5.18965364e-01
-3.12596023e-01 -2.04140231e-01 3.37171495e-01 6.75119579e-01
-4.82619792e-01 7.62502611e-01 -1.64153297e-02 -3.25006023e-02
8.12228918e-02 -2.66160995e-01 6.66178465e-01 -3.48020375e-01
-1.04961075e-01 3.06363493e-01 2.96939034e-02 -4.69192654e-01
-4.94979888e-01 -6.77060723e-01 -1.05602372e+00 -1.69193879e-01
-4.23597157e-01 3.84928107e-01 6.83295131e-01 1.03505862e+00
2.17109755e-01 -7.72057921e-02 1.13424647e+00 -1.02612162e+00
-6.53167486e-01 -8.88098180e-01 -7.98176408e-01 -1.95156708e-01
3.47782463e-01 -5.09877145e-01 -7.29908764e-01 -4.00964558e-01] | [10.744527816772461, -2.3272340297698975] |
962da5fc-0aed-47bf-8860-4a7f2dddf5f8 | cross-attention-of-disentangled-modalities | 2207.13820 | null | https://arxiv.org/abs/2207.13820v1 | https://arxiv.org/pdf/2207.13820v1.pdf | Cross-Attention of Disentangled Modalities for 3D Human Mesh Recovery with Transformers | Transformer encoder architectures have recently achieved state-of-the-art results on monocular 3D human mesh reconstruction, but they require a substantial number of parameters and expensive computations. Due to the large memory overhead and slow inference speed, it is difficult to deploy such models for practical use. In this paper, we propose a novel transformer encoder-decoder architecture for 3D human mesh reconstruction from a single image, called FastMETRO. We identify the performance bottleneck in the encoder-based transformers is caused by the token design which introduces high complexity interactions among input tokens. We disentangle the interactions via an encoder-decoder architecture, which allows our model to demand much fewer parameters and shorter inference time. In addition, we impose the prior knowledge of human body's morphological relationship via attention masking and mesh upsampling operations, which leads to faster convergence with higher accuracy. Our FastMETRO improves the Pareto-front of accuracy and efficiency, and clearly outperforms image-based methods on Human3.6M and 3DPW. Furthermore, we validate its generalizability on FreiHAND. | ['Tae-Hyun Oh', 'Kim Youwang', 'Junhyeong Cho'] | 2022-07-27 | null | null | null | null | ['3d-hand-pose-estimation', '3d-hand-pose-estimation'] | ['computer-vision', 'graphs'] | [ 7.79590458e-02 8.10449645e-02 -1.43302411e-01 -1.10887833e-01
-7.24804640e-01 -1.30033419e-01 3.12764645e-01 -2.36344591e-01
-3.40596199e-01 3.87507558e-01 2.66032487e-01 -1.52506977e-01
2.54533857e-01 -8.64593506e-01 -1.02837634e+00 -2.76543111e-01
1.74783871e-01 6.86961293e-01 3.11456591e-01 -3.54735889e-02
-1.84047237e-01 2.68506825e-01 -1.80023789e+00 3.72046858e-01
7.48255789e-01 1.09079075e+00 4.00272787e-01 7.16447175e-01
5.20148240e-02 8.66964161e-01 -3.05654228e-01 -5.72652340e-01
2.29239330e-01 -3.52086186e-01 -7.61104226e-01 1.19033165e-01
7.30386853e-01 -8.03163886e-01 -5.24708450e-01 7.98933804e-01
7.75926352e-01 -3.68166864e-01 5.14624894e-01 -9.35545027e-01
-4.40259963e-01 4.21406984e-01 -6.18721783e-01 -1.81157485e-01
2.82002419e-01 1.20409057e-01 9.78566647e-01 -1.11276090e+00
5.74682415e-01 1.43509078e+00 8.57019424e-01 5.10282159e-01
-1.18677819e+00 -6.94655597e-01 -1.36768622e-02 3.65741491e-01
-1.62123001e+00 -6.20986283e-01 5.14099002e-01 -3.78591299e-01
1.00074494e+00 6.20191395e-02 1.08015275e+00 9.12139773e-01
2.31936783e-01 9.63514388e-01 7.80078292e-01 -2.46819660e-01
-5.00574633e-02 -2.61359602e-01 -5.74734092e-01 1.13495290e+00
6.61063567e-02 2.05608547e-01 -7.76744604e-01 6.07310757e-02
1.35595775e+00 -1.26990706e-01 -1.84282571e-01 -4.12831008e-01
-1.39015937e+00 5.55899322e-01 5.16992450e-01 -3.32616776e-01
-2.46750280e-01 6.46196187e-01 5.21011353e-01 2.20563650e-01
4.51413482e-01 5.65757044e-02 -4.21908647e-01 -2.91640967e-01
-1.08055604e+00 3.74956757e-01 5.66997230e-01 1.23454523e+00
6.73950195e-01 -4.70830761e-02 1.40438050e-01 7.56180644e-01
3.18916887e-01 8.18448961e-01 8.32469165e-02 -1.06518495e+00
5.08626342e-01 4.07278359e-01 -1.58920258e-01 -9.13582087e-01
-3.21477771e-01 -4.22078609e-01 -9.42976654e-01 1.55044273e-01
3.67073089e-01 2.68857211e-01 -9.50994670e-01 1.51992500e+00
4.01797920e-01 1.18592151e-01 -2.35935211e-01 1.23321998e+00
8.73565972e-01 3.93106550e-01 -1.26982078e-01 1.62781879e-01
1.61481786e+00 -1.02058101e+00 -5.41990638e-01 -1.41899019e-01
5.13394654e-01 -8.28129768e-01 1.25525689e+00 3.70267302e-01
-1.41488147e+00 -5.36205530e-01 -9.14372325e-01 -6.58684492e-01
2.05912843e-01 3.02590966e-01 7.35298216e-01 3.48420024e-01
-8.79718721e-01 5.46940863e-01 -1.19728100e+00 -1.92201555e-01
4.62609529e-01 5.51723540e-01 -2.67678171e-01 -1.04719982e-01
-9.93317783e-01 8.53942454e-01 -5.12351729e-02 1.63406029e-01
-8.57994854e-01 -8.51333737e-01 -1.12781692e+00 7.21335039e-02
3.48355502e-01 -1.10960925e+00 1.39589441e+00 -6.42535269e-01
-1.78195465e+00 9.81540203e-01 -3.07024717e-01 -3.50032419e-01
7.09502041e-01 -4.29686457e-01 9.77488011e-02 3.16163868e-01
-6.26349226e-02 9.25234020e-01 8.35520625e-01 -8.85041177e-01
-5.26342988e-01 -3.64806712e-01 8.81728679e-02 3.40280056e-01
-9.51784104e-02 -2.03145087e-01 -7.81071246e-01 -6.83676541e-01
2.22187668e-01 -9.53066647e-01 -3.71493995e-02 5.81423163e-01
-2.88841367e-01 4.61768396e-02 3.28296304e-01 -6.24712110e-01
1.07781923e+00 -2.14181232e+00 4.52550054e-01 1.03141805e-02
5.96590459e-01 -2.72811353e-02 4.78808843e-02 1.32584304e-01
3.08565557e-01 -2.22706616e-01 -9.90569964e-02 -5.96766472e-01
4.47504409e-02 3.72196943e-01 -2.78101955e-02 5.96235156e-01
4.06529289e-03 1.04998994e+00 -7.62258112e-01 -9.31811810e-01
3.26636493e-01 6.93402648e-01 -1.10385931e+00 7.10852370e-02
-2.82629997e-01 2.81508952e-01 -2.69730419e-01 7.75370121e-01
6.17553473e-01 -6.04644656e-01 1.40818208e-01 -5.94755232e-01
-5.29950596e-02 5.20005047e-01 -1.02408671e+00 2.05213380e+00
-5.78652859e-01 5.84833324e-01 1.81942910e-01 -6.92982674e-01
5.16244292e-01 3.73315871e-01 3.72065455e-01 -7.03975976e-01
2.38223255e-01 2.69841701e-01 -1.08756602e-01 -4.12387192e-01
5.27242124e-01 -1.47453710e-01 6.56359717e-02 2.71353662e-01
-7.99530298e-02 -1.36000067e-01 -1.20077796e-01 2.95091532e-02
8.79117370e-01 4.03690100e-01 2.38434464e-01 -2.96737134e-01
1.81542307e-01 -1.48200601e-01 5.75309038e-01 3.43504518e-01
1.21843427e-01 8.61757040e-01 3.48889619e-01 -4.45759833e-01
-1.20585203e+00 -1.32360494e+00 -1.64890185e-01 7.46087313e-01
3.44342321e-01 -8.10473263e-01 -6.33970320e-01 -2.57883102e-01
8.67616832e-02 2.18227282e-02 -5.37688434e-01 1.07048720e-01
-8.05777669e-01 -3.19436133e-01 7.46509314e-01 7.40769386e-01
6.06568754e-01 -5.76668441e-01 -1.05886781e+00 1.23199850e-01
-4.96043563e-01 -1.32240450e+00 -6.12688839e-01 -1.80157945e-01
-9.16012406e-01 -1.00974560e+00 -8.62705171e-01 -7.29124725e-01
7.53007174e-01 5.96201746e-03 1.27850878e+00 1.77870095e-01
-4.09659177e-01 8.85528401e-02 -1.02579348e-01 -1.48567885e-01
-1.79769635e-01 2.72589207e-01 -2.54849903e-02 -3.39274913e-01
-1.92842156e-01 -6.93669319e-01 -9.06693637e-01 4.33304995e-01
-5.54550827e-01 6.97043538e-01 7.06428468e-01 9.40809071e-01
7.98180521e-01 -1.22779138e-01 2.30854936e-02 -6.23088002e-01
-2.22730022e-02 -9.73786712e-02 -6.87261164e-01 6.70105591e-02
-3.80180359e-01 2.55948156e-01 5.61952055e-01 -4.49786603e-01
-8.91699851e-01 2.34920397e-01 -4.34221953e-01 -6.88771188e-01
3.56581360e-01 2.83975422e-01 -1.87751606e-01 -1.68339629e-02
4.14094567e-01 2.11208805e-01 1.25262812e-01 -6.67453170e-01
3.06790441e-01 4.77556854e-01 6.86525524e-01 -7.40144491e-01
6.36527777e-01 5.91946602e-01 5.56311905e-02 -8.99582326e-01
-6.46505356e-01 -1.95546099e-03 -4.68800366e-01 -1.73508361e-01
9.12258446e-01 -1.25378120e+00 -9.53151464e-01 4.94602650e-01
-1.30832684e+00 -4.37182635e-01 -2.90123820e-01 5.01680672e-01
-7.30486333e-01 2.93348640e-01 -1.07982516e+00 -4.51681107e-01
-4.28814381e-01 -1.31646037e+00 1.47273338e+00 -3.30217898e-01
-3.69748652e-01 -5.98145425e-01 -2.83093870e-01 3.60517085e-01
2.23299667e-01 5.48650697e-02 8.94480288e-01 3.06570709e-01
-9.99160230e-01 2.15847701e-01 -4.07027841e-01 2.35327370e-02
-3.50277275e-02 -2.44960845e-01 -8.60177815e-01 -2.32113719e-01
-3.03549886e-01 -3.23195070e-01 6.92392588e-01 4.53667969e-01
1.34599900e+00 -2.79074550e-01 -2.78226227e-01 1.02712941e+00
1.18230391e+00 -4.11163568e-01 8.62457812e-01 -6.96591660e-02
9.68120337e-01 2.54854739e-01 3.35558116e-01 5.26468277e-01
8.52088869e-01 9.37403619e-01 2.92280287e-01 -2.60802418e-01
-5.11460066e-01 -6.41693771e-01 2.38540962e-01 1.20286608e+00
-2.60814846e-01 -1.17224306e-01 -8.38584065e-01 4.43288743e-01
-1.81139827e+00 -7.16838121e-01 -2.79251412e-02 2.10578704e+00
9.30465937e-01 9.41055194e-02 8.83599669e-02 2.16023192e-01
2.62687951e-01 -4.37934371e-03 -3.90315562e-01 -1.01607315e-01
1.03525132e-01 5.16208231e-01 6.38769031e-01 5.50078988e-01
-8.29218566e-01 1.00019240e+00 6.94208384e+00 9.21453476e-01
-1.03560865e+00 1.50792897e-01 3.42931211e-01 -4.28639889e-01
-3.17632705e-01 -1.76800847e-01 -6.98816121e-01 4.20073092e-01
5.54339290e-01 3.63929346e-02 5.12222826e-01 6.65346742e-01
-4.78867814e-02 -5.26898280e-02 -1.30086935e+00 1.38879561e+00
1.07402608e-01 -1.38042843e+00 1.63128138e-01 2.85971552e-01
3.53644103e-01 4.16668169e-02 -8.97892285e-03 -9.61690247e-02
1.42564788e-01 -1.12118924e+00 1.04923105e+00 2.88440585e-01
1.27295816e+00 -5.97585201e-01 3.82289886e-01 2.84104198e-01
-1.52943659e+00 2.48493269e-01 -4.03847784e-01 -2.59148747e-01
3.55958849e-01 6.73098505e-01 -5.79856992e-01 4.04079795e-01
6.90217733e-01 7.17948258e-01 -2.00882435e-01 6.42493427e-01
-2.57490337e-01 2.96663642e-01 -5.93201399e-01 1.98346093e-01
-2.73489475e-01 2.50130683e-01 2.38003045e-01 9.55529451e-01
3.56292874e-01 1.89754352e-01 2.14347824e-01 8.54457736e-01
-1.22317918e-01 -9.08008814e-02 -3.60201925e-01 2.48106614e-01
5.16761422e-01 7.84807265e-01 -5.76758862e-01 -3.51510227e-01
-4.49214488e-01 1.20275629e+00 4.25316006e-01 6.85085133e-02
-1.11356294e+00 -1.13565445e-01 8.13551009e-01 5.90518773e-01
4.63922769e-01 -4.27489102e-01 -4.42156881e-01 -1.29878104e+00
1.77232668e-01 -8.57359231e-01 1.00419912e-02 -6.67498946e-01
-9.09222126e-01 5.38915157e-01 -5.61094508e-02 -1.30274391e+00
-1.79056957e-01 -4.60141659e-01 8.71552229e-02 5.29774427e-01
-1.42955148e+00 -1.22640240e+00 -3.25251400e-01 5.94173491e-01
4.92313325e-01 2.61393577e-01 8.03607702e-01 8.49391758e-01
-2.68074989e-01 9.37460780e-01 -4.43250477e-01 1.64970964e-01
5.31924903e-01 -9.42216516e-01 7.14983523e-01 5.10446966e-01
3.01908106e-02 4.11510020e-01 5.07383704e-01 -6.54892623e-01
-1.80286038e+00 -8.98613334e-01 1.00902510e+00 -3.83649349e-01
2.27121025e-01 -5.93564391e-01 -5.50011396e-01 7.09788442e-01
-1.75457075e-01 1.35208786e-01 2.95866877e-01 8.77822265e-02
-4.39009428e-01 -7.50246644e-02 -8.82667482e-01 7.89769709e-01
1.61544573e+00 -6.57441258e-01 -3.28228176e-01 7.19422847e-02
7.27681160e-01 -9.25780177e-01 -1.08656204e+00 4.90050018e-01
9.80023384e-01 -9.21724081e-01 1.19111085e+00 9.95981134e-03
8.97306383e-01 -2.63214886e-01 -2.71077693e-01 -9.18044806e-01
-3.63497615e-01 -5.28853416e-01 -4.35197383e-01 6.82615519e-01
1.64788738e-01 -4.25470322e-01 9.18414474e-01 3.77501160e-01
-1.08785599e-01 -1.04380870e+00 -1.03785050e+00 -6.38906777e-01
-1.13762245e-02 -3.64535809e-01 7.57838964e-01 5.36854684e-01
-6.09306479e-03 3.23278606e-01 -6.67831540e-01 2.52734683e-02
7.94470072e-01 1.99075267e-01 9.70160782e-01 -1.01380682e+00
-4.70843256e-01 -2.26651430e-01 -5.64988434e-01 -1.91157103e+00
2.66999584e-02 -8.97864342e-01 7.98466206e-02 -1.41264939e+00
2.43949860e-01 -5.33848882e-01 2.88213491e-01 4.94608670e-01
9.63004380e-02 4.82966512e-01 1.45487145e-01 1.91683561e-01
-5.77992320e-01 8.27339113e-01 1.55236232e+00 -4.94264886e-02
1.47966683e-01 -4.18828666e-01 -2.90490806e-01 8.65039170e-01
4.29721862e-01 -3.08433533e-01 -4.18637276e-01 -1.11680698e+00
2.94703841e-01 2.76858956e-01 5.99667966e-01 -9.92025793e-01
2.01205686e-01 1.55325875e-01 4.68514770e-01 -6.67081773e-01
7.04122961e-01 -7.05708981e-01 3.59987170e-01 5.76661706e-01
-1.32951751e-01 3.04078132e-01 -1.43029814e-04 2.96340674e-01
-6.19851649e-02 3.46792281e-01 8.04118752e-01 -2.16201872e-01
-4.23456758e-01 4.84292626e-01 -2.71546990e-01 2.57329792e-01
5.82779586e-01 -3.68164927e-01 1.33853760e-02 -2.87772030e-01
-5.25458574e-01 9.41182598e-02 8.25007558e-01 2.22920671e-01
8.87434006e-01 -1.52975214e+00 -7.92822182e-01 4.41422552e-01
-2.72214133e-02 2.86983192e-01 3.59603554e-01 8.45014393e-01
-9.29867923e-01 2.02522799e-01 -1.56528935e-01 -8.55799258e-01
-1.41293442e+00 3.03098381e-01 3.95178795e-01 -1.66145518e-01
-1.26076686e+00 8.01290810e-01 4.39978272e-01 -2.66063213e-01
3.03717017e-01 -5.93533933e-01 5.12557924e-01 -3.81150901e-01
3.76895934e-01 4.75210071e-01 -5.29823825e-02 -5.75541854e-01
-4.18715507e-01 8.93634617e-01 1.81454867e-01 -1.46009013e-01
9.41655576e-01 -2.41260100e-02 5.43224253e-02 4.92855199e-02
1.24945557e+00 -5.85313141e-02 -1.32259548e+00 -2.46437147e-01
-5.61781585e-01 -7.10699797e-01 7.10899159e-02 -4.62623388e-01
-1.31527233e+00 9.72204804e-01 4.45974171e-01 -4.64820504e-01
1.01589978e+00 1.62796915e-01 1.37267876e+00 2.23329321e-01
7.41800070e-01 -9.17959392e-01 -7.60513172e-02 4.75920379e-01
6.73346996e-01 -9.50028598e-01 3.26335877e-01 -7.90516138e-01
-1.56346440e-01 7.62429178e-01 5.73028028e-01 -6.09230213e-02
6.22224450e-01 6.54137313e-01 -1.52171820e-01 -3.22619379e-01
-8.18452477e-01 -6.98247999e-02 3.29589456e-01 3.77200633e-01
4.91093636e-01 2.40438282e-01 -1.31840870e-01 3.40768486e-01
-5.57739019e-01 2.05006212e-01 4.53090332e-02 6.90873504e-01
-7.68606439e-02 -1.04723406e+00 -2.07404524e-01 3.26033145e-01
-3.74858528e-01 -2.57538110e-01 9.18293372e-02 7.61110544e-01
2.21029267e-01 5.57249486e-01 2.32223973e-01 -4.49918211e-01
3.89522433e-01 -2.93263286e-01 1.12525427e+00 -4.09290910e-01
-3.92807186e-01 1.65838465e-01 1.70019895e-01 -8.42182696e-01
-2.72311240e-01 -4.18710440e-01 -1.28784204e+00 -8.15273166e-01
-2.14855596e-01 -2.10114300e-01 4.31143671e-01 7.60639250e-01
6.72123671e-01 6.19187355e-01 6.68027326e-02 -9.57152426e-01
-5.30663013e-01 -5.86006463e-01 -3.86578023e-01 2.10538670e-01
2.71070451e-01 -8.97960126e-01 3.11511327e-02 1.93941787e-01] | [7.126330852508545, -1.0513135194778442] |
803b88ac-1ec2-4e5f-a220-6530383b60ec | gloss-attention-for-gloss-free-sign-language | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Yin_Gloss_Attention_for_Gloss-Free_Sign_Language_Translation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Yin_Gloss_Attention_for_Gloss-Free_Sign_Language_Translation_CVPR_2023_paper.pdf | Gloss Attention for Gloss-Free Sign Language Translation | Most sign language translation (SLT) methods to date require the use of gloss annotations to provide additional supervision information, however, the acquisition of gloss is not easy. To solve this problem, we first perform an analysis of existing models to confirm how gloss annotations make SLT easier. We find that it can provide two aspects of information for the model, 1) it can help the model implicitly learn the location of semantic boundaries in continuous sign language videos, 2) it can help the model understand the sign language video globally. We then propose gloss attention, which enables the model to keep its attention within video segments that have the same semantics locally, just as gloss helps existing models do. Furthermore, we transfer the knowledge of sentence-to-sentence similarity from the natural language model to our gloss attention SLT network (GASLT) to help it understand sign language videos at the sentence level. Experimental results on multiple large-scale sign language datasets show that our proposed GASLT model significantly outperforms existing methods. Our code is provided in https://github.com/YinAoXiong/GASLT. | ['Zhou Zhao', 'Tao Jin', 'Weike Jin', 'Li Tang', 'Tianyun Zhong', 'Aoxiong Yin'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['sign-language-translation'] | ['computer-vision'] | [ 1.81309357e-01 -1.65617868e-01 -5.74539185e-01 -5.16099274e-01
-6.49020493e-01 -5.31988263e-01 5.65734766e-02 -6.43012166e-01
-4.50747102e-01 4.44235533e-01 7.99072623e-01 -1.47201970e-01
2.69039005e-01 -1.87328175e-01 -6.95273459e-01 -4.43835020e-01
3.21609229e-01 2.73556918e-01 5.10762036e-01 -9.36487243e-02
1.70507804e-01 -1.07116304e-01 -1.27192390e+00 4.88443702e-01
1.05016911e+00 6.48268342e-01 4.03616458e-01 4.43237633e-01
-1.40799791e-01 1.18669403e+00 -3.44183505e-01 -2.41997033e-01
2.15899572e-01 -6.40268207e-01 -8.72421384e-01 -8.71184468e-02
9.78411853e-01 -8.11170995e-01 -5.80534756e-01 1.10033417e+00
4.09282029e-01 -1.11240543e-01 4.03715760e-01 -1.37605202e+00
-1.04235327e+00 6.63898051e-01 -1.92872599e-01 3.75644863e-03
2.36245975e-01 5.56343138e-01 1.22238648e+00 -7.38925397e-01
1.04158604e+00 1.30483031e+00 4.40154225e-01 8.06338489e-01
-3.29955876e-01 -6.88627064e-01 5.93569875e-01 4.43741024e-01
-9.96932685e-01 -3.02192748e-01 6.82474434e-01 -2.58284420e-01
9.38660920e-01 1.52101573e-02 1.05199599e+00 1.06117654e+00
-1.24025263e-01 1.56997168e+00 1.08946788e+00 -2.43064851e-01
-2.19880030e-01 -2.88137823e-01 1.77977413e-01 9.71185267e-01
4.85640503e-02 -1.37922153e-01 -7.16494262e-01 4.72138822e-01
7.97033310e-01 -5.66948438e-03 -5.45273960e-01 -2.78548151e-01
-1.25070667e+00 4.67849255e-01 6.24128997e-01 3.32107127e-01
-1.04239993e-01 4.76794511e-01 4.36560124e-01 4.43802029e-01
6.42763004e-02 -1.05877398e-02 -3.59289080e-01 -2.72048056e-01
-4.94644374e-01 -6.46457598e-02 4.58532810e-01 1.24227595e+00
3.11269075e-01 -1.64771155e-01 -2.89733142e-01 5.02430499e-01
6.33722961e-01 8.15474689e-01 4.95894283e-01 -9.67100143e-01
8.65368664e-01 6.25631154e-01 -1.49394542e-01 -4.33824331e-01
5.25611229e-02 -4.72125709e-02 -2.93613493e-01 2.15659346e-02
6.19180381e-01 9.12950654e-03 -1.21894836e+00 1.75805068e+00
-1.11840799e-01 1.96394652e-01 -6.53022006e-02 1.30789018e+00
9.52332616e-01 2.99546450e-01 1.65070742e-01 2.41844252e-01
1.26852059e+00 -1.45718658e+00 -9.14474487e-01 -3.35323304e-01
8.07832539e-01 -6.75807178e-01 1.55930007e+00 2.74108015e-02
-7.52921224e-01 -3.30934852e-01 -6.41109049e-01 -5.84669411e-01
-2.29190215e-01 3.96118432e-01 5.29907644e-01 7.87379518e-02
-9.58014846e-01 8.53049755e-02 -9.88103449e-01 -7.58367538e-01
6.35159135e-01 -6.92836335e-03 -1.80452898e-01 -5.49572885e-01
-1.07017338e+00 9.14401591e-01 4.85485867e-02 3.59556824e-01
-5.81398308e-01 -2.43388861e-01 -9.91936266e-01 -1.73431486e-01
5.05687237e-01 -8.16325605e-01 1.47090888e+00 -1.35896552e+00
-1.51322615e+00 8.37395251e-01 -6.83289528e-01 -1.82386532e-01
6.80711746e-01 -5.84058344e-01 5.97279854e-02 5.63214242e-01
1.24696650e-01 9.70874906e-01 7.84561276e-01 -1.07659638e+00
-7.01580822e-01 -6.63675144e-02 1.68748304e-01 2.89140284e-01
-8.14524367e-02 2.30188206e-01 -8.83779407e-01 -7.72591174e-01
6.88156486e-02 -9.10629690e-01 2.50186771e-01 5.74437082e-01
-2.11247280e-01 -4.71447259e-01 9.32202458e-01 -1.17829621e+00
1.13727260e+00 -1.92868125e+00 3.27573150e-01 -3.99234481e-02
1.32398993e-01 3.61112118e-01 -4.25384939e-01 4.81024534e-01
2.80979931e-01 3.75533961e-02 -4.15751308e-01 -2.85750568e-01
-1.82677470e-02 5.19628525e-01 -3.72568548e-01 1.94463637e-02
2.86588520e-01 1.43883705e+00 -1.10807621e+00 -6.58711612e-01
1.42745927e-01 4.59034234e-01 -6.19516551e-01 1.01727419e-01
-4.15204525e-01 6.54899597e-01 -5.86060524e-01 8.35346103e-01
2.30179787e-01 -4.35120106e-01 1.90133095e-01 -3.92948925e-01
-2.80704945e-02 4.51121509e-01 -5.02209067e-01 1.95133483e+00
-2.43762478e-01 1.01620042e+00 -1.67563781e-01 -7.72413790e-01
3.18553925e-01 2.47107610e-01 3.32997531e-01 -6.08459771e-01
2.15436190e-01 3.38932067e-01 3.73805501e-02 -1.02470803e+00
-6.54277429e-02 7.63885304e-02 2.45435163e-01 5.36015332e-01
-8.54457766e-02 1.25288695e-01 3.96773338e-01 2.54449278e-01
7.82448471e-01 4.96946871e-01 -1.13925375e-01 1.98606059e-01
3.23896587e-01 6.79209381e-02 5.87769508e-01 5.92025757e-01
-5.40928662e-01 4.88938153e-01 2.52173424e-01 -2.10005060e-01
-6.92377090e-01 -9.02651787e-01 4.06809628e-01 9.50016916e-01
4.22531992e-01 -5.09496570e-01 -7.04700053e-01 -9.87171352e-01
-1.48301259e-01 6.67573392e-01 -3.79453182e-01 3.65771130e-02
-8.64409864e-01 8.73677209e-02 4.51083720e-01 9.76912141e-01
8.89717698e-01 -1.24825764e+00 -7.25345671e-01 -2.76807934e-01
-6.52034760e-01 -1.38098228e+00 -1.18147147e+00 -5.86340785e-01
-7.86823690e-01 -1.22407568e+00 -8.79937470e-01 -1.16559768e+00
9.72288251e-01 3.36468279e-01 6.77689731e-01 4.60843861e-01
-1.14012972e-01 5.89063108e-01 -7.34682202e-01 -2.63755232e-01
-2.13026389e-01 -5.90622835e-02 -2.30832860e-01 -1.37896165e-01
7.24197209e-01 -3.60965103e-01 -5.05768955e-01 4.44232404e-01
-8.67800415e-01 5.13009548e-01 7.16362178e-01 7.69007385e-01
3.64791632e-01 -4.27527636e-01 3.18693191e-01 -4.08380926e-01
4.76545721e-01 1.06550731e-01 -3.64556849e-01 6.52013898e-01
-3.01614851e-01 2.06741348e-01 1.78278953e-01 -5.57814896e-01
-8.41145933e-01 -6.01808764e-02 -1.96878418e-01 -4.88741100e-01
-2.38937847e-02 5.71792603e-01 -2.10264802e-01 -4.12922166e-02
-8.92113969e-02 4.47528362e-01 2.27789983e-01 -5.91520727e-01
4.25477922e-01 6.90505803e-01 3.37580651e-01 -4.09680635e-01
7.39694476e-01 5.35375774e-01 -4.73347992e-01 -5.92282653e-01
-1.08333790e+00 -4.27976876e-01 -8.39742303e-01 -4.07613039e-01
1.02865744e+00 -8.13153625e-01 -6.13561690e-01 6.10174537e-01
-1.02668440e+00 -7.97265708e-01 -1.71415284e-01 6.24095976e-01
-6.40899539e-01 7.38896787e-01 -6.13608539e-01 -5.29960215e-01
-1.56510353e-01 -1.22335732e+00 1.20664299e+00 2.33033687e-01
-2.13146552e-01 -8.85722697e-01 -3.18013191e-01 7.45500863e-01
4.16655749e-01 -2.16779709e-01 8.85103524e-01 -3.42506170e-02
-1.06710601e+00 -3.32181528e-02 -4.63756353e-01 5.88553131e-01
3.68918955e-01 -1.41635478e-01 -6.38831675e-01 -2.46625900e-01
-3.70774180e-01 -4.73552227e-01 1.13701892e+00 4.63782459e-01
9.57634807e-01 -4.19668466e-01 -1.61566094e-01 6.95310295e-01
1.05384362e+00 -7.95518979e-02 5.88680267e-01 1.06861845e-01
8.83906960e-01 4.37839597e-01 7.63789952e-01 -2.64606237e-01
8.80105317e-01 6.36607289e-01 2.61275828e-01 -1.65051550e-01
-8.13333511e-01 -7.54181325e-01 8.03301632e-01 1.22667062e+00
-2.92923659e-01 -1.90935299e-01 -8.01101267e-01 6.42957568e-01
-2.20789742e+00 -8.84361327e-01 2.27658585e-01 1.78059757e+00
8.39930177e-01 -7.75906816e-02 6.31151721e-02 -2.97350258e-01
3.75005692e-01 6.74315542e-02 -7.40466654e-01 -7.46608200e-03
-2.40083441e-01 -2.53663957e-01 4.77518737e-01 7.19494522e-01
-8.34691167e-01 1.57689202e+00 5.60247421e+00 3.97487968e-01
-1.35188735e+00 7.85181671e-02 -1.63112134e-01 -2.03093603e-01
-1.74088374e-01 1.25642776e-01 -6.46073997e-01 5.04665434e-01
1.59784347e-01 -1.03532158e-01 3.95833075e-01 6.11267507e-01
4.70127314e-01 -6.14455417e-02 -1.19279289e+00 8.98855627e-01
4.17897850e-01 -8.99932027e-01 5.85024714e-01 -4.13864851e-02
5.19244134e-01 3.85360450e-01 -1.27092749e-01 2.01097280e-01
2.33728692e-01 -7.97578633e-01 7.40781426e-01 8.14123154e-01
7.96085656e-01 -1.16328545e-01 8.27239454e-01 2.94493407e-01
-1.33384359e+00 -3.79309729e-02 -2.19988655e-02 -1.23723894e-02
6.21930659e-01 -2.70316184e-01 -6.49762750e-01 3.01276684e-01
6.22378230e-01 1.32421947e+00 -6.21891499e-01 1.23494279e+00
-1.05964100e+00 8.09445739e-01 -4.40216869e-01 -1.19016685e-01
2.95161158e-01 -1.72485009e-01 6.20867550e-01 9.90803897e-01
2.86116660e-01 -6.35867240e-03 2.55941778e-01 6.30398095e-01
-1.16238967e-02 8.37526098e-02 -5.14740348e-01 -2.76482552e-01
1.22745559e-01 3.03738058e-01 -3.82282734e-01 -5.25488257e-01
-6.78946376e-01 1.28902805e+00 1.51704162e-01 7.63520122e-01
-6.75552368e-01 -3.51480514e-01 8.41693521e-01 6.51802868e-02
3.73115808e-01 -4.50095117e-01 8.42946619e-02 -1.59914172e+00
5.31805038e-01 -8.42777133e-01 3.76016378e-01 -1.28645873e+00
-1.26159418e+00 1.72443151e-01 -1.32167742e-01 -1.43682182e+00
-1.47611484e-01 -9.20597732e-01 -4.50380355e-01 6.71155930e-01
-2.00327730e+00 -1.66771400e+00 -4.41265136e-01 6.45844162e-01
9.20485914e-01 1.66581511e-01 3.83431405e-01 2.69884974e-01
-3.06229055e-01 7.11200237e-01 -2.31515691e-01 7.83917725e-01
9.68876004e-01 -8.98430288e-01 2.36833587e-01 1.04510558e+00
3.07735652e-01 7.61757314e-01 2.10629255e-01 -9.28572237e-01
-1.22671497e+00 -9.31957304e-01 1.17083001e+00 -7.57571399e-01
7.75779247e-01 7.09181353e-02 -6.97417080e-01 1.02106893e+00
2.30569735e-01 -3.19061071e-01 3.71231526e-01 -1.81047127e-01
-5.12691379e-01 8.78995508e-02 -5.26980281e-01 7.59104908e-01
1.65623212e+00 -7.65568614e-01 -8.82667601e-01 4.63654101e-01
7.97215819e-01 -4.22578484e-01 -5.16743541e-01 2.63471007e-01
9.19323862e-01 -5.18612981e-01 6.05923772e-01 -9.00827825e-01
6.03038609e-01 -5.48249066e-01 -4.13899980e-02 -1.11190712e+00
1.64741352e-01 -3.48409921e-01 -7.69131109e-02 8.62276733e-01
3.70110124e-01 -6.49770677e-01 3.91477525e-01 6.45760357e-01
-2.20339850e-01 -5.80293000e-01 -8.18474710e-01 -9.16830063e-01
-2.27497332e-02 -4.69195098e-01 3.90426785e-01 7.88311362e-01
2.83559728e-02 2.04578951e-01 -4.84618038e-01 1.71199560e-01
6.44647479e-01 2.49466494e-01 8.21093023e-01 -8.54548335e-01
-4.36596200e-02 -5.81706107e-01 -3.55793804e-01 -1.76993048e+00
1.08270220e-01 -1.07819760e+00 8.05158392e-02 -2.10829544e+00
1.51246533e-01 5.24867559e-03 -1.28050923e-01 1.03195155e+00
-1.99045554e-01 3.00963353e-02 5.31364143e-01 4.29035276e-01
-7.90410519e-01 6.68785214e-01 1.84266555e+00 -2.41002932e-01
4.06185724e-02 -1.13736339e-01 -5.28941810e-01 8.42900395e-01
6.61619782e-01 -2.11816788e-01 -3.65539283e-01 -1.20704460e+00
-1.46572720e-02 -3.87129933e-01 5.52313328e-01 -7.03205526e-01
3.71377021e-01 -1.56976387e-01 -2.85907954e-01 -4.53634351e-01
1.15348712e-01 -8.97280991e-01 -5.41852593e-01 4.20097262e-01
-4.58048552e-01 -2.56578743e-01 -3.99121363e-03 3.91850740e-01
-4.97991025e-01 -4.86421399e-02 4.26402032e-01 -3.39385793e-02
-1.04234624e+00 3.31015110e-01 -2.91369647e-01 1.52871609e-01
5.19095540e-01 -1.76744536e-01 -4.37584251e-01 -7.94113994e-01
-6.29975259e-01 8.90267849e-01 5.08939683e-01 8.72196496e-01
6.72813535e-01 -1.28890145e+00 -5.28115869e-01 1.41894937e-01
4.46551949e-01 3.05167437e-02 6.73340335e-02 1.05212605e+00
-4.77660179e-01 4.99652803e-01 -1.32369980e-01 -7.17884898e-01
-1.64135110e+00 1.46347448e-01 3.43448848e-01 2.02227101e-01
-9.96918738e-01 8.53373826e-01 1.86582729e-01 -3.34209234e-01
5.01879692e-01 -1.08832037e+00 1.85938198e-02 -2.94468790e-01
4.14284766e-01 -2.88066804e-01 -6.97943509e-01 -7.85670400e-01
-2.00657085e-01 1.22329140e+00 -1.40795568e-02 -1.37133047e-01
1.18592858e+00 -2.70846486e-01 -6.78329989e-02 3.89638096e-01
9.96888041e-01 -1.93629429e-01 -1.41856480e+00 -5.61923504e-01
-4.38012294e-02 -5.78791916e-01 -1.96500018e-01 -1.05372453e+00
-1.10604787e+00 1.20418727e+00 4.13588315e-01 -6.54581904e-01
1.22121418e+00 1.82871953e-01 1.15744269e+00 6.67924881e-01
3.92489105e-01 -1.16671526e+00 1.38836220e-01 6.95184052e-01
1.12988508e+00 -1.45179343e+00 -4.18826103e-01 -5.15081048e-01
-8.89766812e-01 9.16843057e-01 5.70752501e-01 8.75694081e-02
5.19467950e-01 -5.56471944e-02 6.32672191e-01 -1.21076941e-01
-6.05437398e-01 -6.18520081e-01 4.97561187e-01 5.51879406e-01
2.73111731e-01 -1.95858121e-01 -3.21783990e-01 3.75988901e-01
-1.96417436e-01 5.42356193e-01 1.28311470e-01 1.00187457e+00
-4.66368884e-01 -1.15606523e+00 1.78412676e-01 1.71642974e-01
1.74202278e-01 -2.43227988e-01 -7.44669855e-01 9.24296439e-01
7.37575516e-02 6.08682334e-01 -1.90487713e-01 -7.26062655e-02
3.67008209e-01 4.07132298e-01 6.39062464e-01 -4.37820852e-01
-1.62387475e-01 9.20858532e-02 1.60346627e-01 -6.72268093e-01
-7.86076248e-01 -6.86092317e-01 -1.42640114e+00 2.63132751e-01
-7.65621662e-05 -2.49365449e-01 4.39043045e-01 1.32023787e+00
1.71319351e-01 3.78339767e-01 -3.84848081e-02 -3.06420654e-01
-2.80999959e-01 -1.06560314e+00 -9.66472253e-02 5.83223522e-01
3.89097154e-01 -5.94537199e-01 -2.87223279e-01 5.01455665e-01] | [9.226713180541992, -6.529380798339844] |
5a2314a9-5a1b-4889-8743-75f5f879e10b | neural-graph-reasoning-complex-logical-query | 2303.14617 | null | https://arxiv.org/abs/2303.14617v1 | https://arxiv.org/pdf/2303.14617v1.pdf | Neural Graph Reasoning: Complex Logical Query Answering Meets Graph Databases | Complex logical query answering (CLQA) is a recently emerged task of graph machine learning that goes beyond simple one-hop link prediction and solves a far more complex task of multi-hop logical reasoning over massive, potentially incomplete graphs in a latent space. The task received a significant traction in the community; numerous works expanded the field along theoretical and practical axes to tackle different types of complex queries and graph modalities with efficient systems. In this paper, we provide a holistic survey of CLQA with a detailed taxonomy studying the field from multiple angles, including graph types (modality, reasoning domain, background semantics), modeling aspects (encoder, processor, decoder), supported queries (operators, patterns, projected variables), datasets, evaluation metrics, and applications. Refining the CLQA task, we introduce the concept of Neural Graph Databases (NGDBs). Extending the idea of graph databases (graph DBs), NGDB consists of a Neural Graph Storage and a Neural Graph Engine. Inside Neural Graph Storage, we design a graph store, a feature store, and further embed information in a latent embedding store using an encoder. Given a query, Neural Query Engine learns how to perform query planning and execution in order to efficiently retrieve the correct results by interacting with the Neural Graph Storage. Compared with traditional graph DBs, NGDBs allow for a flexible and unified modeling of features in diverse modalities using the embedding store. Moreover, when the graph is incomplete, they can provide robust retrieval of answers which a normal graph DB cannot recover. Finally, we point out promising directions, unsolved problems and applications of NGDB for future research. | ['Jure Leskovec', 'Zhaocheng Zhu', 'Michael Cochez', 'Mikhail Galkin', 'Hongyu Ren'] | 2023-03-26 | null | null | null | null | ['logical-reasoning'] | ['reasoning'] | [ 2.70610917e-02 5.20834267e-01 -5.60489595e-01 -2.81266481e-01
-5.92931271e-01 -5.40420592e-01 4.71545935e-01 5.95330238e-01
1.01710871e-01 3.36611658e-01 1.23381652e-01 -3.70055646e-01
-4.45290089e-01 -1.57136691e+00 -8.44813108e-01 -2.33547345e-01
-2.62494981e-01 9.15496051e-01 5.67469895e-01 -4.06061232e-01
-1.10999234e-02 5.21109521e-01 -1.77648318e+00 5.48842728e-01
4.78120208e-01 1.24697137e+00 -5.31957187e-02 5.52779734e-01
-4.59980518e-01 1.40751398e+00 -2.78129399e-01 -8.90426099e-01
-2.85356282e-03 -1.15497470e-01 -1.05555415e+00 -4.09613729e-01
4.71951723e-01 -1.69202209e-01 -8.34787607e-01 1.14873171e+00
2.83034205e-01 8.25460441e-03 3.40177596e-01 -1.77446306e+00
-8.15742791e-01 7.43772089e-01 3.34905237e-01 -7.99774602e-02
8.99093628e-01 -3.09086516e-02 1.50990200e+00 -6.94214880e-01
9.73345041e-01 1.49606168e+00 2.76260644e-01 2.36079723e-01
-1.00340438e+00 -1.09909646e-01 -5.03545208e-03 8.56605649e-01
-1.50118887e+00 -2.80993700e-01 8.34454000e-01 -2.62153119e-01
1.42737043e+00 4.70382631e-01 6.70047343e-01 6.41613781e-01
2.56552696e-01 9.28708792e-01 3.25840801e-01 -4.00914431e-01
3.10492963e-01 1.07286714e-01 4.09203231e-01 1.45171463e+00
1.40706137e-01 -9.18441117e-02 -9.73333120e-01 -5.34265220e-01
3.48943323e-01 3.13966349e-02 -2.02221438e-01 -9.64371562e-01
-1.06239557e+00 9.55571115e-01 7.36163199e-01 2.90183499e-02
-3.71958137e-01 3.29840541e-01 5.75259566e-01 9.18887317e-01
-3.00612357e-02 2.73111165e-01 -4.21607018e-01 2.95102060e-01
-4.67289478e-01 2.95862466e-01 1.33960998e+00 1.48018742e+00
1.05618894e+00 -2.46530563e-01 -4.43511575e-01 2.93348491e-01
6.34182572e-01 4.98901457e-01 -7.22922161e-02 -7.37024546e-01
6.35648012e-01 1.22923124e+00 -5.69451630e-01 -1.33836293e+00
-3.00605834e-01 -2.32068315e-01 -7.06656694e-01 -2.21322522e-01
-1.59675404e-02 6.44020140e-01 -4.40177172e-01 1.56540060e+00
3.68606180e-01 -2.38537773e-01 1.79942161e-01 7.31401265e-01
1.02007043e+00 4.78591591e-01 -8.35121498e-02 -1.95890889e-01
1.53559482e+00 -1.03142548e+00 -7.65820444e-01 -1.44871116e-01
9.71435905e-01 -1.70645908e-01 1.13788068e+00 5.08751810e-01
-1.14599419e+00 -2.54719406e-01 -1.06293130e+00 -6.24030054e-01
-9.01911259e-01 -3.21666300e-01 1.02466249e+00 3.42289537e-01
-1.47534406e+00 2.21817181e-01 -6.48952067e-01 -4.12813723e-01
1.30305633e-01 4.17143464e-01 -5.35746515e-01 -7.11895049e-01
-1.49144089e+00 6.98150516e-01 6.89389169e-01 -1.21212468e-01
-7.42971539e-01 -4.40039366e-01 -1.15791070e+00 2.86278129e-01
1.04280257e+00 -1.21965611e+00 8.54364872e-01 -8.37806091e-02
-1.12472820e+00 7.42286026e-01 -1.70749485e-01 -5.19740403e-01
5.80715314e-02 2.27890044e-01 -6.55475020e-01 3.31260681e-01
-9.08504613e-03 2.39054576e-01 6.25355184e-01 -6.66069090e-01
-2.91387320e-01 -7.75160313e-01 4.20319825e-01 4.70381491e-02
-2.36728355e-01 -1.60447657e-01 -1.01603508e+00 -2.36903936e-01
3.31397384e-01 -4.65999365e-01 1.09052189e-01 1.20022289e-01
-5.93441069e-01 -6.54094458e-01 7.58207619e-01 -4.14092064e-01
1.59684920e+00 -2.08335018e+00 5.17612517e-01 4.80111957e-01
8.96156967e-01 -2.07770035e-01 -6.19733408e-02 1.09377074e+00
2.01848343e-01 -8.69649500e-02 9.92677212e-02 -7.86233041e-03
5.30669928e-01 4.99837518e-01 -5.59649229e-01 1.61180094e-01
4.17769738e-02 1.51528990e+00 -8.34068418e-01 -8.57618332e-01
-1.10981643e-01 1.66523635e-01 -7.66748488e-01 2.89366394e-01
-8.55895221e-01 -3.86228055e-01 -7.07950652e-01 1.22645330e+00
3.89144689e-01 -7.84649432e-01 3.86395723e-01 -5.24117053e-01
5.41916907e-01 6.25846982e-02 -1.19237173e+00 1.84056091e+00
-6.20383732e-02 2.14479402e-01 4.67473939e-02 -1.07054746e+00
9.31785285e-01 1.51340798e-01 3.69525194e-01 -8.59804332e-01
-2.78761059e-01 1.48079276e-01 -5.46447396e-01 -6.45967960e-01
6.48825943e-01 2.06359342e-01 -2.49767959e-01 3.11665237e-01
4.53318357e-01 -1.35456413e-01 3.56316954e-01 6.70694411e-01
1.67582834e+00 -3.44760597e-01 2.18597487e-01 1.95203364e-01
7.00272381e-01 1.74747720e-01 -2.66548358e-02 8.25826585e-01
1.31258339e-01 -3.60525213e-02 9.13565814e-01 -5.38744032e-01
-4.63948488e-01 -1.25213516e+00 3.29489022e-01 1.29796588e+00
1.98743865e-01 -1.18995929e+00 -3.09200078e-01 -7.72397697e-01
3.78463686e-01 4.11609083e-01 -1.95376486e-01 -4.66931581e-01
-3.36576849e-01 -1.35987461e-01 6.70322537e-01 3.74385118e-01
3.63825679e-01 -1.07202983e+00 -1.23109557e-01 1.58317300e-04
-5.32211326e-02 -1.16333580e+00 -6.18363358e-02 -7.99544752e-02
-8.39691699e-01 -1.51793075e+00 1.78358644e-01 -7.40973473e-01
4.09141779e-01 8.83424133e-02 1.65313947e+00 4.12072450e-01
-1.38355225e-01 8.90277624e-01 -2.56232381e-01 -6.36618733e-02
-3.44682813e-01 5.21480404e-02 -2.38431618e-01 -1.01680920e-01
4.10296321e-01 -4.28617984e-01 -1.84996590e-01 1.06453814e-01
-1.19667339e+00 -9.08601061e-02 6.49409354e-01 7.25761712e-01
9.39243257e-01 2.15101898e-01 5.42173386e-02 -1.17105579e+00
9.41690326e-01 -7.18431234e-01 -9.81370628e-01 9.88011956e-01
-9.75943089e-01 5.84745705e-01 5.00914037e-01 4.12230156e-02
-4.18707818e-01 -2.68718958e-01 1.73198655e-01 -6.64662659e-01
2.49974668e-01 1.18243957e+00 -4.70582843e-01 -9.25321728e-02
6.55741632e-01 2.16341704e-01 1.12652920e-01 -3.74784350e-01
8.16088915e-01 8.54642093e-02 4.47513521e-01 -6.54110491e-01
7.68727243e-01 2.64871866e-01 5.53760648e-01 -3.56942862e-01
-4.59286362e-01 -3.30447227e-01 -4.56015885e-01 -1.37619227e-01
6.54864788e-01 -7.06247568e-01 -1.20294702e+00 -6.44255951e-02
-1.18210888e+00 -1.64075032e-01 -4.53444302e-01 2.33898729e-01
-6.98801577e-01 3.52937549e-01 -6.34368181e-01 -5.97094297e-01
-3.57399762e-01 -1.11286020e+00 1.29211104e+00 -3.30646783e-01
1.32427737e-01 -1.14621532e+00 -5.39037213e-02 4.54830021e-01
4.41871166e-01 2.03389198e-01 1.74886465e+00 -7.89889455e-01
-1.51333237e+00 -4.01140392e-01 -3.50082219e-01 -6.65808618e-02
-3.85769248e-01 -4.29403692e-01 -5.97128570e-01 -2.79709309e-01
-1.66583523e-01 -5.72059810e-01 6.61291063e-01 -1.09381452e-01
9.77847695e-01 -5.24408758e-01 -5.02102971e-01 6.50376856e-01
1.62641811e+00 -1.63722202e-01 4.11849707e-01 -2.09432349e-01
7.79340029e-01 5.16069412e-01 3.91568959e-01 1.55147418e-01
7.66317487e-01 6.12182796e-01 7.95170069e-01 3.91792268e-01
-2.10630387e-01 -6.55957282e-01 3.41787517e-01 9.25918579e-01
2.16753766e-01 -5.25585115e-01 -1.07826447e+00 5.98531775e-02
-1.89329696e+00 -8.07411671e-01 -9.22591463e-02 1.94313109e+00
4.70112354e-01 -1.09622360e-03 -9.39555839e-02 1.28128007e-01
3.86121809e-01 2.82546818e-01 -7.09095418e-01 -1.54296547e-01
-2.69395769e-01 2.08033621e-01 4.27802145e-01 6.71611071e-01
-5.34625590e-01 1.09799767e+00 6.22083521e+00 8.06028485e-01
-8.23849499e-01 6.46392480e-02 -8.11577514e-02 1.89287975e-01
-8.19947720e-01 2.75528342e-01 -6.73994720e-01 -1.08495325e-01
9.73344803e-01 -3.65649670e-01 1.17312646e+00 9.17074144e-01
-7.06625044e-01 1.88219100e-01 -1.58955276e+00 1.17847061e+00
2.70542473e-01 -1.67944598e+00 4.14527506e-01 -1.32043539e-02
-4.52347100e-02 1.12814434e-01 -1.96411937e-01 7.83138871e-01
2.17706993e-01 -1.01489365e+00 5.21971762e-01 1.10271800e+00
8.77277732e-01 -3.47169727e-01 5.48423707e-01 3.80231917e-01
-1.38938999e+00 -2.50945956e-01 -3.59292805e-01 8.06878284e-02
-2.03737691e-02 4.53034848e-01 -7.90260434e-01 1.14469516e+00
5.13677478e-01 6.64093018e-01 -9.81721580e-01 5.79710066e-01
-2.07849711e-01 1.18149981e-01 -1.80326864e-01 -1.10667430e-01
-2.22135961e-01 -8.35749432e-02 4.04893160e-01 6.07064068e-01
1.66083023e-01 6.35434091e-02 3.19407612e-01 1.03712106e+00
-2.57266372e-01 2.31748104e-01 -1.17077565e+00 -5.27746797e-01
5.80465794e-01 8.72540116e-01 -3.19733620e-01 -3.80794346e-01
-5.98047316e-01 7.50506818e-01 6.78328812e-01 5.47663629e-01
-4.92403388e-01 -2.83189774e-01 2.48747513e-01 5.46773747e-02
-6.10644966e-02 -1.86086774e-01 1.62532613e-01 -1.37635088e+00
4.11215931e-01 -1.00789344e+00 1.19737089e+00 -1.01812255e+00
-1.26275504e+00 5.58206499e-01 2.83834010e-01 -8.16846848e-01
-5.25744498e-01 -7.58194923e-01 2.29472667e-02 6.14910483e-01
-1.43243968e+00 -1.33559465e+00 -4.36354548e-01 1.22161770e+00
-1.06795691e-01 -4.21531826e-01 1.06814146e+00 3.75464916e-01
-1.65696844e-01 6.11792088e-01 -4.91482288e-01 4.22788113e-02
3.12457025e-01 -1.19692934e+00 8.65343139e-02 4.47096735e-01
4.75029320e-01 6.02579594e-01 3.09041113e-01 -6.98706627e-01
-2.67922020e+00 -9.30899918e-01 1.04183125e+00 -4.65285122e-01
9.28768694e-01 -5.18642426e-01 -9.10339713e-01 9.82578039e-01
-8.55322629e-02 3.35952282e-01 3.96345824e-01 2.09433287e-01
-6.57077432e-01 -3.40339482e-01 -1.03125036e+00 4.85516906e-01
1.43112433e+00 -1.15334833e+00 -2.86358297e-01 7.96641529e-01
1.36266518e+00 -5.01408398e-01 -1.26177716e+00 3.90444368e-01
1.84907824e-01 -9.22105789e-01 1.02248764e+00 -9.51118946e-01
-1.22467980e-01 -1.86193272e-01 -6.60343528e-01 -5.43754220e-01
-1.30251497e-01 -6.06529355e-01 -1.00803483e+00 7.21731126e-01
3.30539435e-01 -7.26792216e-01 9.76493180e-01 6.89096987e-01
-2.15645730e-01 -1.03219211e+00 -9.06554222e-01 -5.09218574e-01
-6.92204237e-01 -8.74423742e-01 9.86373544e-01 7.34909236e-01
2.16816515e-01 5.74639142e-01 6.68600351e-02 4.34922487e-01
5.56152642e-01 3.80204946e-01 8.50716949e-01 -1.32552111e+00
-4.66595531e-01 -2.95034021e-01 -9.10516977e-01 -1.05269253e+00
3.10301363e-01 -1.51566637e+00 -6.99093759e-01 -1.87906146e+00
-1.48552418e-01 -1.86585128e-01 -1.97125912e-01 7.19814658e-01
4.85133469e-01 -4.31429267e-01 1.42428949e-01 1.85901836e-01
-1.06468225e+00 4.84816849e-01 1.17755818e+00 -5.87319553e-01
1.13402590e-01 -9.75256115e-02 -4.58768010e-01 1.18379839e-01
8.02906156e-02 -2.10484415e-01 -9.29960728e-01 -5.45809627e-01
1.06098700e+00 7.18953192e-01 5.93081534e-01 -7.79075861e-01
9.73234296e-01 -4.24490385e-02 -3.54149193e-01 -7.27362335e-01
4.24588561e-01 -1.03927362e+00 3.16562295e-01 3.61368120e-01
-3.12477350e-01 4.92024928e-01 -2.14144900e-01 7.94853568e-01
-6.81745648e-01 6.86097443e-02 -1.08120270e-01 -1.36438444e-01
-9.57610130e-01 6.91287696e-01 2.12006271e-01 1.62312254e-01
8.74145865e-01 1.33212805e-01 -6.48533285e-01 -5.07914901e-01
-9.68432009e-01 4.84286845e-01 2.66252816e-01 3.42801750e-01
9.86423850e-01 -1.47280955e+00 -2.54840761e-01 5.22862911e-01
5.88794589e-01 -5.45846485e-03 1.68098614e-01 8.85728061e-01
-5.01840293e-01 6.93826079e-01 1.82567850e-01 -4.36065018e-01
-8.98300350e-01 1.21294248e+00 2.37307608e-01 -6.49788260e-01
-4.91946191e-01 5.88723838e-01 -1.50913417e-01 -5.34830511e-01
5.92980266e-01 -5.41411042e-01 -4.38228995e-02 2.85650883e-02
1.92484841e-01 3.02278638e-01 4.29172367e-01 -2.80246943e-01
-3.42729002e-01 1.49890095e-01 2.67835289e-01 2.55214989e-01
1.11359286e+00 2.21555799e-01 -7.40950525e-01 3.85174364e-01
1.23006558e+00 -1.95555851e-01 -2.53431618e-01 -6.85950637e-01
3.28216434e-01 -2.13849619e-01 -1.49076417e-01 -6.37787819e-01
-9.80307102e-01 6.74695253e-01 1.84057444e-01 8.22034299e-01
1.18905246e+00 6.27176106e-01 5.77454269e-01 9.86970305e-01
7.00771391e-01 -6.99070871e-01 1.66960180e-01 6.47455096e-01
1.08152246e+00 -9.09779668e-01 -1.25605553e-01 -6.02667511e-01
-3.53744447e-01 1.18028057e+00 4.46757317e-01 3.14074278e-01
8.93659592e-01 -2.12602261e-02 -3.55824381e-01 -1.13289082e+00
-1.17359781e+00 -2.63693810e-01 5.87432206e-01 5.59906960e-01
-1.20029919e-01 -6.72311410e-02 8.71766806e-02 5.90599597e-01
-3.07115823e-01 4.98351268e-02 -7.17143640e-02 9.12595034e-01
-2.38547474e-01 -1.31138766e+00 -3.25084627e-02 6.08066797e-01
1.84412584e-01 -9.21396539e-02 -3.88848126e-01 8.55339706e-01
-1.05570167e-01 8.00105929e-01 -1.38507396e-01 -6.50883436e-01
6.51628971e-01 3.44435692e-01 4.28128362e-01 -6.97521567e-01
-1.81805015e-01 -5.55660367e-01 2.49763355e-01 -1.14631128e+00
2.39014383e-02 -2.06859902e-01 -1.29713440e+00 -2.67719030e-01
-2.46398881e-01 3.62194389e-01 5.55640996e-01 7.13902116e-01
6.52370691e-01 3.85907203e-01 1.15031138e-01 1.18896134e-01
-6.44904613e-01 -4.82430071e-01 -7.74900854e-01 2.45124862e-01
1.98442101e-01 -5.20044684e-01 -8.50757435e-02 -2.12474048e-01] | [9.124780654907227, 7.731012344360352] |
00121970-7864-4a82-921b-4b531a392e38 | explain-adapt-and-retrain-how-to-improve-the | 2303.14939 | null | https://arxiv.org/abs/2303.14939v1 | https://arxiv.org/pdf/2303.14939v1.pdf | Explain, Adapt and Retrain: How to improve the accuracy of a PPM classifier through different explanation styles | Recent papers have introduced a novel approach to explain why a Predictive Process Monitoring (PPM) model for outcome-oriented predictions provides wrong predictions. Moreover, they have shown how to exploit the explanations, obtained using state-of-the art post-hoc explainers, to identify the most common features that induce a predictor to make mistakes in a semi-automated way, and, in turn, to reduce the impact of those features and increase the accuracy of the predictive model. This work starts from the assumption that frequent control flow patterns in event logs may represent important features that characterize, and therefore explain, a certain prediction. Therefore, in this paper, we (i) employ a novel encoding able to leverage DECLARE constraints in Predictive Process Monitoring and compare the effectiveness of this encoding with Predictive Process Monitoring state-of-the art encodings, in particular for the task of outcome-oriented predictions; (ii) introduce a completely automated pipeline for the identification of the most common features inducing a predictor to make mistakes; and (iii) show the effectiveness of the proposed pipeline in increasing the accuracy of the predictive model by validating it on different real-life datasets. | ['Fabrizio Maria Maggi', 'Chiara Ghidini', 'Chiara Di Francescomarino', 'Williams Rizzi'] | 2023-03-27 | null | null | null | null | ['predictive-process-monitoring'] | ['time-series'] | [ 5.67419052e-01 7.53593326e-01 1.01700783e-01 -3.06296021e-01
-1.95492059e-01 -1.60736114e-01 8.39586318e-01 9.75349784e-01
1.99121758e-01 3.91712010e-01 1.32340327e-01 -5.32999277e-01
-6.92808211e-01 -8.55237544e-01 -6.93333328e-01 -1.04578950e-01
-2.12360770e-01 7.58717835e-01 4.13569987e-01 3.66734684e-01
5.83383739e-01 5.92679560e-01 -1.89644110e+00 6.38359904e-01
7.45563269e-01 8.66422892e-01 -1.29667088e-01 5.25427818e-01
-1.98138341e-01 1.35480821e+00 -3.78496796e-01 -4.41817977e-02
-8.20373595e-02 -5.79355717e-01 -8.73923659e-01 5.43532148e-02
-1.46830052e-01 8.45738128e-02 1.58220723e-01 4.73604679e-01
-2.70136833e-01 -2.23614067e-01 5.09198368e-01 -1.44141901e+00
6.43273145e-02 9.92097795e-01 1.21386223e-01 7.25666434e-02
4.96454865e-01 3.89910966e-01 8.81196499e-01 -3.54704320e-01
7.54015744e-01 1.03587961e+00 6.35964751e-01 4.34891284e-01
-1.37385476e+00 -3.21120024e-01 2.66883492e-01 3.78365546e-01
-1.03687966e+00 -2.92214632e-01 5.15882790e-01 -7.13724136e-01
1.24777651e+00 3.80114406e-01 6.85173094e-01 1.11655939e+00
4.23393965e-01 6.59064770e-01 9.97441471e-01 -4.47282553e-01
6.20967031e-01 3.32116246e-01 3.69269609e-01 6.14675343e-01
4.67782915e-01 2.16135442e-01 -6.64168119e-01 -2.30217606e-01
2.04386473e-01 3.27721059e-01 -2.89289296e-01 -1.87041849e-01
-1.07184649e+00 4.46248770e-01 -6.27310723e-02 6.58965111e-01
-7.30444074e-01 1.06699087e-01 3.59039307e-01 9.63674039e-02
1.91706002e-01 8.41034055e-01 -5.60889781e-01 -4.66528773e-01
-9.92839575e-01 2.22287416e-01 1.29319692e+00 8.01677525e-01
8.74293804e-01 -2.13762164e-01 -5.62611699e-01 -1.80011272e-01
2.35580996e-01 -1.32835940e-01 6.12345576e-01 -5.09891987e-01
1.34260818e-01 1.23972726e+00 9.87845734e-02 -7.60999978e-01
-4.70809877e-01 -4.15481150e-01 -5.40848315e-01 5.68046309e-02
2.16643170e-01 1.19306184e-01 -5.50117314e-01 1.25295830e+00
-1.54067665e-01 3.46186817e-01 8.26809481e-02 5.20599425e-01
-3.45111787e-02 4.69711095e-01 1.82586685e-01 -4.28717434e-01
1.12891257e+00 -8.25372338e-01 -4.72698271e-01 -9.72593799e-02
6.73061192e-01 -4.46990967e-01 8.71790886e-01 5.87336123e-01
-7.34765232e-01 -7.74647653e-01 -9.39764738e-01 6.55022919e-01
-1.42190352e-01 1.19956404e-01 6.27694547e-01 3.71505201e-01
-5.69210708e-01 1.44570971e+00 -1.20631218e+00 -6.14815414e-01
1.67100772e-01 3.50712985e-01 -2.23289594e-01 1.29153013e-01
-7.41206825e-01 8.53616953e-01 5.40541887e-01 -2.19802454e-01
-1.04171956e+00 -9.44544911e-01 -5.57328582e-01 7.54666805e-01
5.25486827e-01 -5.32753348e-01 1.04640412e+00 -9.44057882e-01
-1.18769836e+00 9.33070779e-02 -4.83615935e-01 -8.23384225e-01
7.50364602e-01 -4.97591704e-01 -3.30120116e-01 4.55236696e-02
-6.65150359e-02 1.04939178e-01 6.65249228e-01 -1.26321769e+00
-7.65458465e-01 -1.96143553e-01 -2.93828160e-01 -5.27004182e-01
-1.31775469e-01 -2.73424268e-01 -2.16430682e-03 1.76840685e-02
9.78493243e-02 -8.36081862e-01 -3.95000100e-01 -6.06652319e-01
-8.89653146e-01 -1.62481934e-01 7.32560515e-01 -4.31496024e-01
1.12067604e+00 -2.06333089e+00 6.03798404e-02 5.16218185e-01
2.13660061e-01 7.80998915e-02 4.19144630e-01 8.70134592e-01
-1.35983214e-01 4.34562504e-01 -2.18644530e-01 -6.66403294e-01
1.22824594e-01 3.75817746e-01 -6.92472458e-01 2.34814174e-02
6.57175362e-01 3.96757782e-01 -7.17559576e-01 -2.32611746e-01
6.33698881e-01 2.20814288e-01 -2.90873379e-01 6.41358376e-01
-4.40019965e-01 6.74191594e-01 -3.69926631e-01 3.49489689e-01
1.31222099e-01 -9.99501497e-02 4.20838326e-01 1.83186993e-01
-4.04982924e-01 4.15481299e-01 -1.08872044e+00 6.56922340e-01
-4.66020077e-01 3.15750092e-01 -8.00302744e-01 -6.43019557e-01
1.17664957e+00 4.56129849e-01 5.61100066e-01 -2.14373514e-01
-1.00633122e-01 1.37739837e-01 -7.63596296e-02 -4.00771618e-01
2.23978385e-01 -4.32643741e-02 2.24484280e-01 5.36047220e-01
-1.05800830e-01 2.42873281e-01 2.95962453e-01 -1.41185030e-01
1.72472596e+00 1.94803774e-01 6.67719424e-01 -5.80137037e-02
9.93383408e-01 5.48812933e-02 6.93554044e-01 9.18293417e-01
1.29432574e-01 5.91245174e-01 1.07974756e+00 -8.32024217e-01
-8.86469185e-01 -6.04511499e-01 3.02275777e-01 5.44443905e-01
-1.79865733e-01 -8.16664219e-01 -5.27339876e-01 -8.36935043e-01
9.42126364e-02 1.38345683e+00 -7.85598338e-01 -3.06874156e-01
-3.95760953e-01 -2.69257367e-01 3.73591423e-01 4.93264914e-01
7.76969418e-02 -1.16126227e+00 -9.78482962e-01 5.49169302e-01
1.61940515e-01 -1.12638247e+00 4.91230845e-01 7.48766303e-01
-1.02796733e+00 -1.56265342e+00 3.61392558e-01 1.92787752e-01
6.13863051e-01 -4.73039508e-01 1.08451760e+00 3.00001115e-01
-6.84209913e-02 2.00993747e-01 -5.77069759e-01 -5.66222548e-01
-1.02390075e+00 1.89397484e-01 -1.21304616e-01 1.34366825e-01
4.51500207e-01 -5.96573353e-01 -5.45733012e-02 3.34329754e-01
-8.85372519e-01 6.22484349e-02 7.59035349e-01 2.44135752e-01
6.63958549e-01 2.56437182e-01 2.15035453e-01 -1.23794651e+00
6.16527081e-01 -3.08252722e-01 -6.68372512e-01 3.62071574e-01
-1.34067845e+00 6.16635919e-01 1.12174273e+00 -1.36091337e-01
-7.95015693e-01 3.06448489e-01 -3.71454917e-02 -5.11386096e-01
-6.45871639e-01 4.66502517e-01 -9.30775777e-02 4.37912971e-01
4.57911581e-01 4.07281935e-01 -2.39098683e-01 -5.02999604e-01
-1.48733661e-01 1.39730185e-01 1.01427935e-01 -4.21067476e-01
7.49483407e-01 4.29319412e-01 4.74651098e-01 -1.88609183e-01
-6.29797757e-01 -5.11464536e-01 -6.98733449e-01 -8.38760734e-02
5.92451334e-01 -3.30829054e-01 -9.41778243e-01 -1.12046473e-01
-1.15144098e+00 -2.12552428e-01 -6.94464326e-01 3.19727153e-01
-8.48498821e-01 1.99114203e-01 -5.00692368e-01 -9.99520361e-01
-3.39434654e-01 -1.18679106e+00 9.20450091e-01 2.52047367e-02
-8.49907637e-01 -8.66835892e-01 3.51592191e-02 -2.36256178e-02
3.32085460e-01 2.27364123e-01 1.13148391e+00 -1.41440713e+00
-6.62142694e-01 -3.68502438e-01 5.31283543e-02 1.25546172e-01
7.41171166e-02 3.42162997e-01 -9.56876814e-01 1.74089834e-01
-4.03248444e-02 5.01385450e-01 5.55526257e-01 -4.05183211e-02
1.15885830e+00 -2.54937470e-01 -6.24964416e-01 2.38761604e-01
1.34484792e+00 -2.15274356e-02 6.89512968e-01 6.16930068e-01
3.11754912e-01 6.64501131e-01 6.59992635e-01 7.44515717e-01
8.75917673e-02 6.92219615e-01 6.66486979e-01 3.91016334e-01
2.05278412e-01 -5.97598374e-01 4.88769174e-01 2.86940336e-01
-2.15040982e-01 6.36193752e-02 -1.04092872e+00 4.46670473e-01
-2.15783453e+00 -1.03076279e+00 -5.14702797e-01 2.15217423e+00
2.20064983e-01 4.23955441e-01 -6.96425214e-02 6.32666349e-01
3.97273719e-01 -2.45146513e-01 -9.16131213e-02 -7.39579618e-01
2.89953202e-01 1.92617103e-01 4.86329019e-01 3.38540107e-01
-9.58693683e-01 5.86304188e-01 5.20626497e+00 8.04527253e-02
-1.00395799e+00 -1.53193936e-01 2.11135328e-01 3.32930118e-01
-2.11667627e-01 3.62602830e-01 -8.75427067e-01 3.75119656e-01
1.41375923e+00 -3.03480506e-01 1.06699802e-01 1.30517578e+00
4.73650068e-01 9.42566395e-02 -1.89498544e+00 3.03426713e-01
-1.45816967e-01 -1.31221354e+00 1.54566392e-01 1.74196824e-01
4.54011649e-01 -4.76937741e-01 -3.78506303e-01 2.40710348e-01
1.46959439e-01 -1.22077525e+00 7.53682315e-01 9.83331323e-01
-1.75005734e-01 -7.07809865e-01 1.24977362e+00 5.41384816e-01
-1.00089467e+00 -4.36726838e-01 -1.89820379e-01 -3.69552195e-01
1.03467070e-01 9.56304848e-01 -1.58183038e+00 8.22307229e-01
3.73241186e-01 6.71766996e-01 -7.75659561e-01 1.03093636e+00
-7.33897984e-01 8.90066922e-01 -6.67625591e-02 1.43558597e-02
1.37658104e-01 3.72038811e-01 6.21905923e-01 1.29054260e+00
5.43487310e-01 -5.54288507e-01 1.15012459e-01 1.13393772e+00
4.76603597e-01 -5.52050546e-02 -4.79576945e-01 -1.38049170e-01
2.75428474e-01 9.05714333e-01 -6.06564522e-01 -3.58306378e-01
9.31239426e-02 8.06583166e-01 1.71721280e-01 -9.39152911e-02
-6.02278054e-01 1.97107285e-01 5.87793946e-01 5.42533219e-01
4.83330548e-01 2.48894282e-02 -5.26352167e-01 -7.70205498e-01
2.07163841e-02 -6.63084984e-01 3.31152767e-01 -6.58852041e-01
-1.08732259e+00 7.53382146e-01 -2.12506905e-01 -1.20516050e+00
-7.22044826e-01 -2.52703577e-01 -7.40360320e-01 8.60092282e-01
-1.39278853e+00 -9.44143891e-01 -5.32363296e-01 2.38967985e-01
4.07820255e-01 8.45142975e-02 1.04088020e+00 -2.81476051e-01
-4.57357556e-01 -1.28073469e-01 -4.24409539e-01 -2.63374031e-01
3.78882974e-01 -1.40801024e+00 1.97414950e-01 8.85144651e-01
-3.40321171e-03 6.91247642e-01 1.18947756e+00 -7.49862611e-01
-1.37962019e+00 -1.37008107e+00 1.22756875e+00 -4.54138607e-01
6.29667521e-01 1.68139711e-01 -1.08442640e+00 9.85291600e-01
-1.28836483e-01 -3.53852183e-01 6.56434834e-01 2.31718943e-01
-1.54844984e-01 -1.01330481e-01 -1.07621932e+00 1.86184362e-01
6.40966892e-01 -1.75508291e-01 -8.44067574e-01 1.93020761e-01
5.37482560e-01 4.83883247e-02 -8.91632617e-01 3.83259594e-01
2.43853033e-01 -1.40927160e+00 4.02618736e-01 -6.72036946e-01
9.73548353e-01 -5.17445505e-01 1.93359241e-01 -1.20781207e+00
-1.59331053e-01 -5.54037333e-01 -4.54603523e-01 1.22493339e+00
4.30167288e-01 -4.98575628e-01 6.24001682e-01 6.37193322e-01
-1.79469973e-01 -7.61148751e-01 -8.11974287e-01 -6.18795931e-01
-5.89451075e-01 -7.14304268e-01 7.88015723e-01 6.87325001e-01
2.37147883e-01 3.25565115e-02 -3.10195267e-01 5.27015865e-01
3.40915054e-01 3.37141424e-01 8.44774723e-01 -1.61565924e+00
-5.11327267e-01 -2.78524548e-01 -8.36758316e-01 -3.01343501e-01
1.28568515e-01 -5.42360246e-01 -1.75566897e-02 -1.44372368e+00
9.80730206e-02 -3.23254734e-01 -3.58494341e-01 6.92213297e-01
-2.00838551e-01 -5.18998384e-01 4.40270394e-01 5.06742597e-01
-4.92005110e-01 2.45325655e-01 5.07211447e-01 5.35922013e-02
-3.92901182e-01 2.65818149e-01 -5.42558551e-01 7.77694821e-01
6.25110805e-01 -7.95799077e-01 -1.92415699e-01 2.07210809e-01
2.83484012e-01 4.90715392e-02 6.10353529e-01 -1.54400826e+00
2.84807920e-01 -1.08538561e-01 3.26269001e-01 -3.61991346e-01
-6.31101951e-02 -1.25901818e+00 6.31406963e-01 1.10880554e+00
-5.52544475e-01 -5.78923523e-02 1.83587402e-01 6.71466529e-01
-4.46572512e-01 -5.56958199e-01 3.17442089e-01 -2.45165173e-02
-7.69105136e-01 -1.72086999e-01 -5.50471783e-01 -6.61665916e-01
1.29500234e+00 -1.60807699e-01 -1.72895521e-01 -2.92731315e-01
-8.10608327e-01 -7.46609494e-02 4.59804237e-01 2.34948575e-01
3.98061961e-01 -5.23662031e-01 -3.64178360e-01 2.04213277e-01
3.63328665e-01 -2.93915361e-01 -1.08961109e-02 9.73567426e-01
-3.59250665e-01 5.95624030e-01 -2.20292136e-01 -5.74768186e-01
-1.24222994e+00 7.13015378e-01 3.91865551e-01 -8.82865548e-01
-8.06969345e-01 1.97042406e-01 -3.73091429e-01 -1.65421724e-01
-1.85135782e-01 -6.10804021e-01 -2.39972129e-01 -2.36657977e-01
5.48130572e-01 2.44510025e-01 3.99388522e-01 3.33665386e-02
-5.12293458e-01 6.31380007e-02 1.17370360e-01 2.09471524e-01
1.57355452e+00 2.40598395e-01 -2.82922536e-02 6.86077535e-01
4.17254448e-01 -4.12331457e-04 -1.13896525e+00 1.90021753e-01
5.72650254e-01 -4.66185927e-01 -3.22561085e-01 -9.23361957e-01
-6.10675454e-01 9.64961469e-01 3.26712400e-01 6.14690244e-01
9.79000449e-01 -8.44602659e-02 2.59459645e-01 3.36052865e-01
3.41425985e-01 -6.46758974e-01 -3.58204365e-01 4.21388298e-01
5.71396649e-01 -7.39868164e-01 -2.89665884e-03 -7.92608559e-01
-7.72154331e-01 1.66970587e+00 4.45367187e-01 1.32942870e-01
2.52507985e-01 3.83122534e-01 -3.20838809e-01 -2.42911845e-01
-1.17104757e+00 -1.26329452e-01 -5.56218885e-02 5.42570174e-01
4.32097733e-01 2.02675179e-01 -1.71860784e-01 9.44806457e-01
-1.50178671e-01 5.62490582e-01 7.89369822e-01 9.45424259e-01
-4.57629085e-01 -1.19304991e+00 -2.89807141e-01 5.14092624e-01
-1.99981734e-01 1.53877705e-01 -6.68463051e-01 8.31983626e-01
2.19154984e-01 1.04566872e+00 -4.15485818e-03 -6.26871109e-01
7.54538774e-01 6.41938329e-01 9.05090272e-02 -6.96353078e-01
-9.93833721e-01 -5.29841602e-01 2.78949827e-01 -7.72088766e-01
-1.42707422e-01 -8.08341801e-01 -1.13764524e+00 -2.97121733e-01
1.32145047e-01 1.39273435e-01 6.13020062e-01 1.39584923e+00
6.86013877e-01 1.14670634e+00 4.38667417e-01 -4.61301535e-01
-6.04422390e-01 -9.73634541e-01 -1.79239452e-01 7.46918678e-01
-2.59723544e-01 -5.90002298e-01 -4.31303263e-01 7.69995973e-02] | [8.663803100585938, 5.962955474853516] |
b0f317bd-3c05-43c1-913d-c6d4ce972920 | learned-sorted-table-search-and-static | 2107.09480 | null | https://arxiv.org/abs/2107.09480v6 | https://arxiv.org/pdf/2107.09480v6.pdf | Learned Sorted Table Search and Static Indexes in Small Model Space | Machine Learning Techniques, properly combined with Data Structures, have resulted in Learned Static Indexes, innovative and powerful tools that speed-up Binary Search, with the use of additional space with respect to the table being searched into. Such space is devoted to the Machine Learning Model. Although in their infancy, they are methodologically and practically important, due to the pervasiveness of Sorted Table Search procedures. In modern applications, model space is a key factor and, in fact, a major open question concerning this area is to assess to what extent one can enjoy the speed-up of Binary Search achieved by Learned Indexes while using constant or nearly constant space models. In this paper, we investigate the mentioned question by (a) introducing two new models, i.e., the Learned k-ary Search Model and the Synoptic Recursive Model Index, respectively; (b) systematically exploring the time-space trade-offs of a hierarchy of existing models, i.e., the ones in the reference software platform Searching on Sorted Data, together with the new ones proposed here. By adhering and extending the current benchmarking methodology, we experimentally show that the Learned k-ary Search Model can speed up Binary Search in constant additional space. Our second model, together with the bi-criteria Piece-wise Geometric Model index, can achieve a speed-up of Binary Search with a model space of 0:05% more than the one taken by the table, being competitive in terms of time-space trade-off with existing proposals. The Synoptic Recursive Model Index and the bi-criteria Piece-wise Geometric Model complement each other quite well across the various levels of the internal memory hierarchy. Finally, our findings stimulate research in this area, since they highlight the need for further studies regarding the time-space relation in Learned Indexes. | ['Raffaele Giancarlo', 'Giosuè Lo Bosco', 'Domenico Amato'] | 2021-07-19 | null | null | null | null | ['table-search'] | ['natural-language-processing'] | [ 4.86086085e-02 -1.56360477e-01 -4.65705901e-01 5.14782779e-02
-5.57001054e-01 -7.53768027e-01 7.17914343e-01 6.94266796e-01
-6.97688639e-01 6.42682910e-01 -2.02842250e-01 -6.34555042e-01
-8.21397305e-01 -1.07109141e+00 -6.43948138e-01 -6.03464901e-01
-1.05909787e-01 7.70491123e-01 6.98219419e-01 -4.52932149e-01
6.51296318e-01 7.26024330e-01 -2.13696861e+00 1.26963556e-01
8.15933049e-01 1.33815193e+00 1.08115777e-01 3.99203151e-01
-2.31881186e-01 3.83017629e-01 -4.62669075e-01 -3.94150406e-01
4.14916337e-01 -1.61618516e-01 -7.51279056e-01 -3.90513271e-01
3.20928007e-01 1.96865380e-01 -7.56893679e-02 7.10811675e-01
2.24533796e-01 -6.27778769e-02 2.98772037e-01 -1.12956846e+00
-4.04326320e-02 5.52275360e-01 -2.67810494e-01 4.54656839e-01
2.08493337e-01 3.58334109e-02 1.10326111e+00 -4.72652197e-01
6.48733735e-01 9.64512885e-01 5.17980695e-01 9.71015543e-02
-1.33686483e+00 -4.74762291e-01 5.63719384e-02 4.75178748e-01
-1.56289291e+00 -2.49041989e-01 4.20288175e-01 -4.19580668e-01
1.14562249e+00 8.31672370e-01 5.93248248e-01 4.72231686e-01
3.46517682e-01 3.55126381e-01 1.38599563e+00 -7.83478439e-01
4.06822324e-01 5.04910946e-01 2.22174257e-01 5.79754949e-01
5.83355844e-01 3.81800681e-01 -6.80168748e-01 -1.79956719e-01
6.46419406e-01 -3.17979127e-01 -8.70115757e-02 -8.96017432e-01
-1.00883508e+00 7.48987675e-01 3.76008004e-01 8.18850219e-01
-8.88582766e-02 5.74724423e-03 4.75782812e-01 5.68645060e-01
3.46161753e-01 8.86191845e-01 -7.83434987e-01 -3.71033788e-01
-1.31375277e+00 1.85099304e-01 1.03865361e+00 7.43882895e-01
8.57833803e-01 -1.71985164e-01 8.19814429e-02 5.44187784e-01
-2.30193976e-03 2.99544632e-01 5.85834503e-01 -4.98719841e-01
3.33938211e-01 1.04760194e+00 -1.57061681e-01 -1.02066541e+00
-6.45805061e-01 -7.61514306e-01 -7.28103399e-01 2.83148229e-01
6.44919991e-01 5.14142990e-01 -4.10986215e-01 1.57288098e+00
3.41975182e-01 -2.76127130e-01 -1.11871541e-01 4.56290543e-01
1.96185172e-01 3.03432733e-01 -3.33353728e-01 -5.85966289e-01
1.54324996e+00 -7.35600054e-01 -4.18997020e-01 -7.65252039e-02
7.47649550e-01 -6.78366840e-01 1.30680060e+00 6.53804362e-01
-1.20566034e+00 -6.20841920e-01 -1.20390749e+00 2.03287259e-01
-7.62586474e-01 -2.28442475e-01 7.08450735e-01 9.87098753e-01
-1.41017294e+00 8.01751137e-01 -8.31047654e-01 -3.93936247e-01
-8.33029076e-02 5.52540839e-01 -1.94478601e-01 4.30578262e-01
-1.22021079e+00 1.06943631e+00 4.53866750e-01 -1.86814487e-01
-2.40888551e-01 -6.52529538e-01 -3.87410522e-01 2.97843218e-01
8.56106341e-01 -4.58115458e-01 9.63169277e-01 -3.86617482e-01
-1.17650187e+00 9.57351446e-01 -1.12138782e-02 -8.33994269e-01
7.31518269e-01 -1.93491485e-02 -2.52794892e-01 -2.65128128e-02
-8.15234259e-02 1.48745179e-01 4.06977594e-01 -1.10489821e+00
-5.72709203e-01 -5.52934885e-01 1.47894219e-01 6.58114115e-03
-5.87641835e-01 -2.37774923e-01 -5.65605998e-01 -5.13040900e-01
1.37869313e-01 -9.89057720e-01 -2.36509860e-01 -2.85022765e-01
1.22701906e-01 -2.42390111e-01 3.38442773e-01 -2.74515152e-01
2.11581850e+00 -2.10121655e+00 2.80541331e-01 4.92850840e-01
6.83277249e-02 1.78845808e-01 1.74808711e-01 5.47603905e-01
3.90647240e-02 2.31094867e-01 -2.01153368e-01 4.72392403e-02
1.38835028e-01 1.54233575e-01 -2.89243758e-01 3.92386168e-01
-1.81104109e-01 7.26768613e-01 -6.36223793e-01 -5.10000944e-01
1.06815919e-01 9.84149799e-02 -6.98374033e-01 -1.57361448e-01
-3.60229254e-01 -1.21938817e-01 -2.23057956e-01 4.34371829e-01
3.77681136e-01 -2.81360805e-01 2.28104323e-01 -1.29415527e-01
-6.21270418e-01 3.91602784e-01 -1.42360365e+00 1.37300944e+00
-4.59825188e-01 3.14607352e-01 -1.34580076e-01 -1.00092995e+00
1.07733941e+00 -3.58011052e-02 2.89775252e-01 -1.11561644e+00
-6.18262403e-02 8.09600353e-01 5.64886630e-02 -1.24407627e-01
7.74368048e-01 -1.23240957e-02 -1.62191838e-01 3.62499475e-01
-2.00521007e-01 -3.96091528e-02 5.18511534e-01 -1.55470610e-01
9.79213953e-01 2.83258110e-02 4.46203858e-01 -6.69347286e-01
8.17102134e-01 7.31024891e-02 6.12405650e-02 8.88352871e-01
8.81545767e-02 1.81522533e-01 6.67227089e-01 -6.40449047e-01
-8.52174580e-01 -6.79900169e-01 -5.16026318e-01 1.17585087e+00
1.88508451e-01 -6.08664691e-01 -7.01613486e-01 -5.04388213e-01
8.44277442e-02 5.42620063e-01 -6.44595027e-01 -2.24360853e-01
-8.60133946e-01 -5.93428195e-01 4.70410109e-01 2.39950418e-01
3.97850901e-01 -9.59536850e-01 -1.40250850e+00 2.32707918e-01
1.27915800e-01 -7.47213721e-01 1.44735947e-02 6.93072855e-01
-1.30114746e+00 -1.04265857e+00 -3.00991386e-01 -2.12450892e-01
2.52098404e-02 -4.64117378e-02 1.39623237e+00 3.91943246e-01
-2.41589516e-01 3.93202275e-01 -4.11735028e-01 -1.90617427e-01
-4.22786444e-01 6.71011448e-01 -1.74154565e-02 -2.05094755e-01
3.56685370e-01 -6.12266779e-01 -4.85689789e-01 5.18438041e-01
-1.16082835e+00 -1.78181589e-01 7.20396519e-01 7.58587956e-01
5.55053949e-01 2.48600945e-01 1.41970396e-01 -6.80089712e-01
4.93581057e-01 -3.18885416e-01 -1.06842673e+00 3.65201533e-01
-1.44937015e+00 6.84706032e-01 4.75909740e-01 -2.16793060e-01
-4.67194140e-01 -1.80963844e-01 -2.82684471e-02 -5.43797798e-02
8.33731741e-02 5.66171348e-01 -1.24186948e-01 -1.37501925e-01
7.52608418e-01 4.32572275e-01 -5.50792133e-03 -6.42189622e-01
2.80132651e-01 3.10472280e-01 3.44545364e-01 -5.94352305e-01
8.15056145e-01 3.87507051e-01 2.89926261e-01 -5.93223393e-01
-2.81694531e-01 -5.14498413e-01 -7.79346049e-01 2.52909139e-02
5.48294008e-01 -3.24944258e-01 -7.09887922e-01 1.30554065e-01
-6.97454274e-01 -2.30265588e-01 -3.53436083e-01 1.58157364e-01
-7.08609819e-01 2.98152328e-01 -3.71216655e-01 -8.03880036e-01
-2.03148261e-01 -1.26586032e+00 9.35348094e-01 -8.39624107e-02
-3.62438351e-01 -9.24848795e-01 6.31735250e-02 1.58876047e-01
6.65554166e-01 4.55509238e-02 1.34625173e+00 -9.62697208e-01
-8.38287413e-01 -1.43214926e-01 -8.01408961e-02 -4.77814041e-02
-1.95396379e-01 -2.89991319e-01 -6.09288514e-01 -3.17854136e-01
3.30562830e-01 9.37252864e-02 8.77786934e-01 4.21828963e-02
9.04797852e-01 -3.12672526e-01 -3.81223530e-01 5.35619199e-01
1.77335477e+00 5.59870541e-01 5.81072509e-01 1.14772820e+00
3.36182006e-02 6.64153278e-01 7.10811913e-01 4.95391458e-01
1.69085294e-01 1.19451976e+00 5.39683282e-01 2.19876785e-03
2.17289999e-02 -2.39345729e-02 -1.54020533e-01 7.78082907e-01
-8.26155841e-02 6.69128902e-04 -1.14424455e+00 3.15843940e-01
-1.68393660e+00 -7.21765995e-01 -2.17099208e-02 2.60760522e+00
8.30721080e-01 6.99062526e-01 2.87336409e-01 8.89737308e-01
2.98042119e-01 1.79595649e-01 -3.35643291e-01 -6.58789873e-01
-1.61146045e-01 3.31929177e-01 5.50872743e-01 5.77056527e-01
-8.90334725e-01 6.63624108e-01 5.81328773e+00 1.08096921e+00
-1.20493269e+00 -1.32627293e-01 5.68473876e-01 -1.06612377e-01
-2.20834613e-01 5.96913472e-02 -1.05007982e+00 4.71884549e-01
1.39003372e+00 -3.08533341e-01 5.64493954e-01 8.73184562e-01
-2.05882132e-01 -2.93930888e-01 -1.24940193e+00 9.36175823e-01
-4.15068157e-02 -1.35806239e+00 8.25481340e-02 3.92133296e-01
2.52847880e-01 -2.57050157e-01 -2.56872689e-03 3.64632457e-01
-3.26350540e-01 -9.52489614e-01 8.84407222e-01 3.21871370e-01
6.89666927e-01 -6.37213767e-01 7.22500086e-01 4.11941111e-01
-1.31998217e+00 -1.77931949e-01 -3.90486419e-02 -1.42343119e-01
-3.22717816e-01 4.42431509e-01 -3.11812878e-01 7.15711236e-01
7.54050314e-01 7.02216327e-02 -9.15628374e-01 1.14750767e+00
3.05752873e-01 4.18663234e-01 -4.48217720e-01 -2.78234869e-01
3.03946108e-01 -1.19735986e-01 4.79509711e-01 1.18625474e+00
3.16529572e-01 -3.07452470e-01 -3.34329009e-01 6.31762981e-01
3.94633383e-01 3.01655412e-01 -3.57558966e-01 2.47303486e-01
5.30428290e-01 8.81087661e-01 -1.10350394e+00 -2.80239582e-01
-1.23775855e-01 5.14106095e-01 1.93967059e-01 -1.56340729e-02
-6.71738148e-01 -2.71606624e-01 2.18838930e-01 5.72687984e-01
4.49706852e-01 -1.95606291e-01 -6.46142423e-01 -7.98484206e-01
3.61358732e-01 -1.14330792e+00 5.81360936e-01 -1.02062583e-01
-7.69942999e-01 1.01181245e+00 3.51882100e-01 -1.06348026e+00
-3.19291294e-01 -6.48553848e-01 1.54728562e-01 6.86851740e-01
-1.37479520e+00 -4.22418803e-01 -1.51403919e-01 4.36385274e-01
4.32453245e-01 -7.52978995e-02 8.20047200e-01 3.99772942e-01
-1.80025548e-01 7.14005947e-01 3.13554108e-01 -5.55519283e-01
3.54503393e-01 -1.21664524e+00 2.24217698e-01 3.87943089e-01
3.88938278e-01 7.43538916e-01 7.70084918e-01 -2.18597949e-01
-1.49871993e+00 -3.86175603e-01 1.13509309e+00 -5.94950080e-01
8.21811020e-01 -4.84786063e-01 -1.09566808e+00 5.53638116e-02
-1.04953304e-01 -1.83916211e-01 4.65281874e-01 2.89907396e-01
-5.00276089e-01 -5.16001165e-01 -9.53920305e-01 4.05577123e-01
8.90709519e-01 -4.06832337e-01 -5.74628294e-01 8.63403082e-03
7.06962049e-01 -3.19707781e-01 -8.49479854e-01 6.61204875e-01
6.39763832e-01 -1.45829046e+00 1.09115195e+00 -2.17110306e-01
2.96767801e-01 -1.37375668e-01 -2.42318600e-01 -7.74000049e-01
-9.13283005e-02 -5.48672855e-01 -2.62306243e-01 9.82932508e-01
5.37545681e-01 -8.73642206e-01 8.98110032e-01 3.27622503e-01
1.37780443e-01 -1.41111219e+00 -1.30686390e+00 -1.13713932e+00
2.02127129e-01 -3.16933542e-01 7.12065995e-01 6.49677157e-01
-4.38376367e-02 3.00107151e-01 -2.34686956e-02 -3.97583187e-01
2.48150557e-01 3.17119747e-01 6.34715199e-01 -1.51613963e+00
-6.09240115e-01 -1.09861791e+00 -4.50529635e-01 -9.87126768e-01
-4.21313107e-01 -8.18256736e-01 -3.65521401e-01 -1.11912405e+00
2.92885886e-03 -7.95463681e-01 -5.97028196e-01 1.86736599e-01
2.07473245e-02 2.11718623e-02 4.30550903e-01 5.41056752e-01
-4.76191789e-01 1.37222067e-01 7.35110700e-01 1.09893896e-01
-4.28358912e-01 1.03434086e-01 -6.35784507e-01 5.17386198e-01
6.53128147e-01 -3.96250576e-01 -3.55613738e-01 -1.44575551e-01
6.53385162e-01 1.93047315e-01 8.08423758e-02 -1.25297391e+00
5.10495305e-01 7.30047375e-03 -9.26450118e-02 -6.02415264e-01
2.35213399e-01 -9.85401273e-01 2.32180938e-01 9.66432691e-01
-3.05775404e-01 5.35029352e-01 5.50799608e-01 3.82190615e-01
-3.09495151e-01 -3.56110334e-01 5.60833693e-01 -7.37357885e-02
-5.70479274e-01 -1.73752129e-01 -1.83750793e-01 -7.34182820e-02
8.91700804e-01 -6.13949776e-01 -2.32175693e-01 6.30323663e-02
-5.86188853e-01 -5.68607636e-02 4.90165323e-01 4.32147682e-01
2.66581625e-02 -9.61764038e-01 -2.66827911e-01 3.55050206e-01
5.78478724e-02 -3.94227237e-01 -7.39488825e-02 9.46083665e-01
-5.59875131e-01 1.07249141e+00 -4.70644124e-02 -6.99672163e-01
-1.39238572e+00 1.04889619e+00 1.99007154e-01 -9.38862860e-01
-2.59290069e-01 6.10666752e-01 -8.91419724e-02 -3.12839419e-01
4.62929994e-01 -4.63437438e-01 -5.56911668e-03 3.64803284e-01
2.83646852e-01 6.31774426e-01 5.44027925e-01 -1.94617599e-01
-5.84417939e-01 8.15927982e-01 1.37445018e-01 -2.32026745e-02
1.16288936e+00 5.84035292e-02 -4.32770491e-01 6.63095951e-01
1.02141452e+00 1.67308122e-01 -6.81439579e-01 -2.26566166e-01
6.41148090e-01 -3.38432908e-01 -8.99069309e-02 -1.06548190e+00
-6.86716855e-01 7.25876451e-01 8.33235085e-01 7.09138274e-01
1.46006203e+00 1.12535758e-02 2.67044663e-01 3.00482780e-01
7.64889836e-01 -8.49440753e-01 -9.99995768e-02 4.05060977e-01
8.29035282e-01 -8.52494359e-01 1.18868716e-01 -4.36559826e-01
-3.81849073e-02 1.13597262e+00 4.10846114e-01 6.86814412e-02
5.34737945e-01 4.85303789e-01 -1.28391773e-01 -1.72106847e-01
-9.27643299e-01 -2.61671066e-01 3.84165615e-01 1.51288375e-01
2.77126074e-01 -1.30601838e-01 -9.06549275e-01 4.60894763e-01
-5.97285390e-01 -8.33437219e-02 3.55605856e-02 8.85567248e-01
-5.35314202e-01 -1.50365222e+00 -6.11424208e-01 2.72597879e-01
-3.04650068e-01 -2.59202808e-01 -2.88588196e-01 1.26603770e+00
1.99821740e-01 8.24678242e-01 3.49899940e-02 -5.47867537e-01
5.86303532e-01 1.07939087e-01 4.74219322e-01 -2.83299714e-01
-1.10588109e+00 2.11748835e-02 -3.40667032e-02 -6.17642522e-01
-2.03329884e-02 -4.37628210e-01 -7.40406454e-01 -2.74844348e-01
-4.36394960e-01 5.31279266e-01 7.61590183e-01 7.63606191e-01
3.46803427e-01 3.52287889e-01 3.94450784e-01 -6.04946792e-01
-9.43651855e-01 -6.67058468e-01 -3.40738863e-01 5.77467084e-02
1.16869316e-01 -7.52711475e-01 -5.53292096e-01 -5.05535364e-01] | [8.361262321472168, 4.243413925170898] |
2581c8b3-c4ef-4be5-b4fa-4c90b56c754c | ga-drl-graph-neural-network-augmented-deep | 2307.00777 | null | https://arxiv.org/abs/2307.00777v1 | https://arxiv.org/pdf/2307.00777v1.pdf | GA-DRL: Graph Neural Network-Augmented Deep Reinforcement Learning for DAG Task Scheduling over Dynamic Vehicular Clouds | Vehicular clouds (VCs) are modern platforms for processing of computation-intensive tasks over vehicles. Such tasks are often represented as directed acyclic graphs (DAGs) consisting of interdependent vertices/subtasks and directed edges. In this paper, we propose a graph neural network-augmented deep reinforcement learning scheme (GA-DRL) for scheduling DAG tasks over dynamic VCs. In doing so, we first model the VC-assisted DAG task scheduling as a Markov decision process. We then adopt a multi-head graph attention network (GAT) to extract the features of DAG subtasks. Our developed GAT enables a two-way aggregation of the topological information in a DAG task by simultaneously considering predecessors and successors of each subtask. We further introduce non-uniform DAG neighborhood sampling through codifying the scheduling priority of different subtasks, which makes our developed GAT generalizable to completely unseen DAG task topologies. Finally, we augment GAT into a double deep Q-network learning module to conduct subtask-to-vehicle assignment according to the extracted features of subtasks, while considering the dynamics and heterogeneity of the vehicles in VCs. Through simulating various DAG tasks under real-world movement traces of vehicles, we demonstrate that GA-DRL outperforms existing benchmarks in terms of DAG task completion time. | ['Huaiyu Dai', 'Seyyedali Hosseinalipour', 'Manman Luo', 'Zhibin Gao', 'Lianfen Huang', 'Zhang Liu'] | 2023-07-03 | null | null | null | null | ['graph-attention'] | ['graphs'] | [-2.67853022e-01 -8.82402137e-02 -2.95396358e-01 -2.31842488e-01
-2.56321549e-01 -4.33876425e-01 4.33632344e-01 -5.16783856e-02
-2.07149580e-01 7.40694046e-01 6.57360628e-02 -7.66389489e-01
-2.03061193e-01 -7.61392772e-01 -9.33764100e-01 -6.22820377e-01
-4.79777783e-01 7.28972614e-01 6.49351895e-01 -3.17944854e-01
2.90011205e-02 8.06093991e-01 -1.56951952e+00 1.85723364e-01
9.95776296e-01 8.86064410e-01 7.35536039e-01 7.82005310e-01
-1.89078927e-01 9.98339415e-01 -5.71178973e-01 -7.48717636e-02
1.81241095e-01 2.44878247e-01 -6.86666906e-01 -2.84647383e-02
-1.36829391e-01 -1.02286153e-01 -6.50229335e-01 6.44705713e-01
1.58580795e-01 4.25857723e-01 8.22836816e-01 -1.97497463e+00
-4.96618271e-01 2.18282387e-01 -6.64978027e-01 6.23828053e-01
1.63821690e-02 4.45116997e-01 7.79683053e-01 -5.90818107e-01
5.41801095e-01 1.29414749e+00 1.27778009e-01 3.48471016e-01
-7.90243447e-01 -7.85078108e-01 7.97675371e-01 8.25556517e-01
-1.06726992e+00 -2.46991336e-01 8.87759089e-01 -7.28802145e-01
1.07233441e+00 1.66755125e-01 6.39427543e-01 8.27855766e-01
8.60357404e-01 1.03073955e+00 7.46191263e-01 6.10114187e-02
4.29822206e-01 -3.60704482e-01 2.54162252e-01 5.45704424e-01
2.74989367e-01 1.66732535e-01 -2.02517450e-01 -1.08706124e-01
4.14519548e-01 1.08989932e-01 4.50189412e-01 -5.61122596e-01
-9.14708138e-01 7.95616388e-01 6.05669737e-01 -2.81422734e-01
-7.56920755e-01 4.85256106e-01 7.64102638e-01 2.58753061e-01
3.60630363e-01 2.50945147e-02 -4.65830952e-01 2.32059553e-01
-4.60857511e-01 4.47177708e-01 5.16434252e-01 1.49826038e+00
1.00249493e+00 4.01791453e-01 -6.89945459e-01 1.63685307e-01
3.71173501e-01 4.32867497e-01 8.94699842e-02 -7.51346409e-01
6.26496434e-01 6.32509828e-01 2.28476971e-01 -1.08575904e+00
-6.61222577e-01 -2.91477978e-01 -9.01312113e-01 1.50630727e-01
-1.22206107e-01 -4.48325276e-01 -1.16260278e+00 1.67667294e+00
5.79518437e-01 4.30281818e-01 -4.46366929e-02 1.04048705e+00
7.47267902e-01 7.37856805e-01 5.69551110e-01 -5.25971465e-02
1.36132467e+00 -1.20626247e+00 -7.08946764e-01 -2.15800643e-01
6.34491742e-01 -2.31866091e-01 7.51403153e-01 -1.29700974e-01
-7.38883078e-01 -6.65391266e-01 -8.30532312e-01 3.24736089e-01
-4.97606367e-01 -9.76558551e-02 7.77602375e-01 2.06927866e-01
-1.24292910e+00 5.45867607e-02 -7.80920386e-01 2.02708989e-01
5.14968395e-01 7.13613331e-01 -7.41680190e-02 2.60590445e-02
-1.35954010e+00 4.75410998e-01 2.88899809e-01 3.70483994e-01
-1.59993887e+00 -5.20405591e-01 -9.38702822e-01 5.88990562e-02
9.37003613e-01 -7.52204418e-01 1.09792674e+00 -9.62231278e-01
-1.26928449e+00 2.18400419e-01 -3.82184684e-01 -3.10959995e-01
2.23935813e-01 3.23508739e-01 -6.11713171e-01 -1.67421743e-01
3.64167482e-01 4.64633763e-01 1.04932511e+00 -1.21865094e+00
-1.05534840e+00 -4.40563053e-01 4.06336099e-01 4.79097664e-01
1.48729965e-01 4.00456116e-02 -5.39779544e-01 -3.00268292e-01
-5.68035245e-01 -1.19739103e+00 -5.80777466e-01 -7.38532662e-01
-5.68567753e-01 -7.95363665e-01 8.74643683e-01 -4.59389001e-01
1.14605618e+00 -2.13567591e+00 2.77160436e-01 1.07225209e-01
7.07281709e-01 1.92106411e-01 -3.13361734e-01 4.69173342e-01
3.62964839e-01 -2.04128638e-01 3.76946986e-01 -2.25290343e-01
8.34845677e-02 4.74455833e-01 2.36321017e-02 2.69545227e-01
3.16123873e-01 1.18317842e+00 -1.22074664e+00 -5.53688467e-01
1.89811870e-01 -8.66044760e-02 -3.97675455e-01 3.48849595e-01
-5.41339040e-01 5.58645248e-01 -9.38494205e-01 7.14036286e-01
9.18275654e-01 -2.50334561e-01 3.54072094e-01 1.45278499e-01
2.74384096e-02 -3.21618877e-02 -1.01155710e+00 1.09633923e+00
-5.86520612e-01 7.17781067e-01 2.13484183e-01 -1.07510042e+00
7.28731811e-01 -8.46236199e-02 4.20287073e-01 -9.23635364e-01
-1.35501519e-01 -1.78973198e-01 1.20300934e-01 -6.35667205e-01
7.99111068e-01 5.31375051e-01 -2.69409299e-01 1.12672582e-01
-1.98969766e-01 3.95377547e-01 3.42350572e-01 4.66786861e-01
1.15909600e+00 -8.09537321e-02 -2.79776633e-01 -4.20850366e-01
4.37918186e-01 1.81107789e-01 8.04055393e-01 6.74137414e-01
-4.10882175e-01 -1.83810622e-01 7.69771338e-01 -9.13438559e-01
-8.25538516e-01 -8.30187798e-01 5.40467680e-01 1.62803984e+00
6.39273822e-01 -6.70363009e-02 -3.84627134e-01 -8.42474282e-01
4.50206131e-01 4.88677740e-01 -8.06837082e-01 -1.49870396e-01
-5.87047338e-01 -5.27266681e-01 1.07694544e-01 7.23722935e-01
6.43579513e-02 -1.59994400e+00 -6.19836628e-01 4.31051075e-01
2.48692796e-01 -1.42615545e+00 -7.31217563e-01 2.51910806e-01
-3.95656824e-01 -1.15630722e+00 -3.28863621e-01 -1.01862812e+00
6.61946952e-01 5.75622559e-01 1.08492029e+00 4.05815318e-02
7.76363984e-02 1.65202133e-02 -2.16329843e-01 -3.74212325e-01
-2.70043194e-01 4.22418714e-01 1.55377686e-01 3.49274725e-01
2.81911135e-01 -4.07145023e-01 -8.68830740e-01 4.77981269e-01
-5.61722517e-01 2.15494320e-01 6.24430716e-01 6.94973767e-01
6.76286459e-01 1.08529061e-01 8.18971574e-01 -7.67717242e-01
6.67606533e-01 -1.01775682e+00 -9.96876240e-01 1.44065619e-01
-3.79717469e-01 7.57816881e-02 9.84432280e-01 -3.76779288e-01
-8.88432503e-01 8.98410454e-02 3.51696640e-01 -7.96739638e-01
-1.55614212e-01 5.53976834e-01 -1.79188341e-01 9.80817899e-02
4.14570943e-02 2.26511240e-01 -1.76686600e-01 -2.69622933e-02
1.74478829e-01 4.86234426e-01 1.58278570e-01 -6.10035896e-01
5.43666065e-01 2.15721041e-01 3.72376144e-01 -4.16623443e-01
-1.91038236e-01 -4.43181753e-01 -2.87356555e-01 -5.47392726e-01
9.87227023e-01 -1.05617428e+00 -1.39522374e+00 4.62824136e-01
-1.16948664e+00 -8.96161795e-01 8.34700614e-02 2.66837656e-01
-4.78709847e-01 1.24027006e-01 -6.03561103e-01 -7.29424357e-01
-7.20558316e-02 -1.74705422e+00 1.31971657e+00 1.98157236e-01
4.65677619e-01 -9.38946009e-01 -2.46422458e-02 2.98846006e-01
3.32344979e-01 2.15797603e-01 9.94032264e-01 -5.04528284e-01
-8.74141037e-01 1.78197473e-01 -5.12598753e-01 -1.76863596e-01
-6.73497990e-02 7.63652548e-02 -4.72174019e-01 -4.94727939e-01
-8.68201673e-01 2.30871588e-02 6.25159919e-01 6.44754708e-01
1.33676207e+00 -1.83167934e-01 -7.06387699e-01 4.78148967e-01
1.14469707e+00 4.30960208e-01 3.48756611e-01 4.09860522e-01
1.21866047e+00 6.76077187e-01 8.71695340e-01 3.16745073e-01
1.02127790e+00 8.15796971e-01 1.14102125e+00 -1.30259320e-01
-5.89003824e-02 -1.18293270e-01 3.17033589e-01 7.72929430e-01
-1.37263268e-01 -4.60521668e-01 -1.05259001e+00 9.13893878e-01
-2.04921770e+00 -6.16757214e-01 -4.25270528e-01 1.79593360e+00
1.35573909e-01 2.39741564e-01 3.31826806e-01 -4.51336175e-01
9.08336461e-01 2.41213158e-01 -7.13134766e-01 -7.64534056e-01
3.36431921e-01 -2.83304483e-01 7.46241450e-01 3.49645197e-01
-9.78796303e-01 1.21675825e+00 5.09239721e+00 9.60074663e-01
-9.36406255e-01 1.11840673e-01 6.17441356e-01 1.15978591e-01
-3.98846030e-01 -1.38795763e-01 -7.91571558e-01 8.87183189e-01
9.96068358e-01 -2.31124610e-01 6.28055692e-01 8.12854648e-01
4.46913719e-01 1.20436721e-01 -9.01251554e-01 5.49637616e-01
-5.19907534e-01 -1.49253786e+00 1.29966438e-01 1.56753272e-01
8.50045264e-01 3.06314856e-01 1.17059082e-01 7.23003745e-01
8.55215430e-01 -9.83849645e-01 7.40995407e-01 3.18491310e-01
7.55746901e-01 -9.35733855e-01 6.88412130e-01 3.25084060e-01
-1.69172573e+00 -4.82664168e-01 -3.02331567e-01 -7.59971002e-03
9.85758081e-02 3.00964057e-01 -9.15786207e-01 1.03273022e+00
5.57291746e-01 3.46822172e-01 -4.49893355e-01 7.64711797e-01
5.92933930e-02 4.15758163e-01 8.06301609e-02 -3.02903861e-01
5.91079593e-01 -1.85725465e-01 5.86287200e-01 1.08482122e+00
-1.24498039e-01 -9.45936367e-02 7.46975362e-01 4.95818615e-01
-4.64472920e-02 -3.23030800e-02 -6.72986209e-01 2.49771878e-01
8.22745383e-01 1.30998659e+00 -7.33019471e-01 -4.70155269e-01
-3.47794384e-01 5.56768775e-01 5.03685236e-01 7.56105661e-01
-1.00738645e+00 -3.10568452e-01 9.79856551e-01 4.63806167e-02
3.48500222e-01 -2.58761734e-01 6.37760982e-02 -6.10760331e-01
5.04683554e-02 -5.41576862e-01 2.54671007e-01 -7.98673213e-01
-9.24729407e-01 9.37988400e-01 -2.36835815e-02 -1.06493545e+00
-3.41133446e-01 -4.08140481e-01 -8.45870078e-01 8.79767299e-01
-1.88678873e+00 -1.32081604e+00 -4.60393786e-01 8.78127515e-01
8.45096946e-01 -3.73519570e-01 1.10111713e-01 2.90025920e-01
-8.66545618e-01 3.93264055e-01 -5.69886118e-02 -1.12882882e-01
1.99191496e-01 -1.16617239e+00 8.79299641e-01 6.67810678e-01
-3.85666460e-01 -1.17634021e-01 4.71823812e-01 -9.68260169e-01
-1.93632460e+00 -1.74458873e+00 5.12527108e-01 -1.99436277e-01
7.64210820e-01 -5.92271149e-01 -6.94844842e-01 8.56815219e-01
1.13771722e-01 4.07741368e-01 -7.94559196e-02 2.89391279e-02
1.17459238e-01 -3.23986650e-01 -6.35223269e-01 6.16958320e-01
1.22506273e+00 -2.06507534e-01 -1.49233431e-01 6.60675168e-01
1.08221436e+00 -5.76225281e-01 -3.51946652e-01 4.56405759e-01
1.81875244e-01 -4.29280549e-01 6.68045461e-01 -1.09992254e+00
1.67402208e-01 -4.39897180e-01 3.45273018e-02 -1.38982224e+00
-7.89018750e-01 -6.40736222e-01 -2.87055194e-01 6.75155520e-01
2.83144355e-01 -6.67554677e-01 8.64962399e-01 1.28487572e-01
-8.30157518e-01 -9.14765060e-01 -1.08678424e+00 -7.96139061e-01
-2.37315595e-01 -2.88962603e-01 1.16283047e+00 6.91930056e-01
-5.01829624e-01 2.59314656e-01 -3.11179549e-01 5.13328254e-01
6.62582040e-01 2.54647553e-01 8.80934000e-01 -1.19615448e+00
-6.13985471e-02 -3.17002207e-01 -5.33155799e-01 -9.08322215e-01
7.23741472e-01 -9.68955457e-01 8.07238147e-02 -1.56682110e+00
5.61399460e-02 -8.84218037e-01 -4.73598003e-01 3.47587883e-01
-3.18769693e-01 -1.21129021e-01 6.85732886e-02 2.41912380e-02
-1.15867949e+00 6.54738605e-01 1.50694287e+00 -2.94978589e-01
-4.14025187e-01 2.23331988e-01 -3.55960250e-01 3.68023545e-01
8.26588154e-01 -3.50298077e-01 -5.68882406e-01 -6.74067318e-01
2.67099321e-01 5.11441290e-01 3.44781995e-01 -6.28171444e-01
5.33698201e-01 -7.27055669e-01 -4.18236442e-02 -7.90592015e-01
-3.68343038e-03 -7.48984993e-01 3.55652094e-01 4.35433745e-01
-8.87455791e-02 5.16142070e-01 3.88749629e-01 1.00126648e+00
6.13263585e-02 5.13909459e-01 1.55673668e-01 2.75623828e-01
-1.27693069e+00 8.97457898e-01 -8.53211164e-01 2.96701249e-02
1.58754253e+00 -1.08879447e-01 -1.82571411e-01 -7.89400116e-02
-7.85196781e-01 1.08972216e+00 5.03506139e-02 6.74772859e-01
4.95147228e-01 -1.40137899e+00 -7.17556059e-01 2.47782886e-01
1.15916982e-01 1.87263727e-01 7.31685042e-01 7.13233650e-01
-4.12972629e-01 4.58157808e-01 -4.24694061e-01 -7.38052428e-01
-1.14439738e+00 1.05481243e+00 7.80998468e-02 -6.55095577e-01
-4.52879041e-01 5.11616647e-01 7.44145572e-01 -3.03891122e-01
-1.21071404e-02 -1.72225416e-01 -4.51667935e-01 2.67220140e-02
3.66928428e-02 4.86092299e-01 2.30852380e-01 -6.69453442e-01
-5.82386613e-01 1.29573017e-01 -2.43096039e-01 5.39175808e-01
1.07332492e+00 -1.68380737e-01 1.48331255e-01 3.45240757e-02
8.18380296e-01 -3.02325368e-01 -1.60183001e+00 -3.95121127e-02
-1.22799195e-01 -2.42386997e-01 2.42386665e-02 -4.25080091e-01
-1.56403637e+00 5.90670347e-01 2.01124787e-01 4.23018694e-01
8.87537956e-01 5.93052916e-02 8.07263434e-01 1.03641283e-02
4.96257275e-01 -1.03036118e+00 9.85798016e-02 6.85371697e-01
6.06073499e-01 -1.12582850e+00 -4.22137618e-01 -3.29208136e-01
-8.85304749e-01 7.15689480e-01 9.26110387e-01 -3.03619385e-01
5.62891841e-01 2.46200442e-01 -1.38207063e-01 -4.75763708e-01
-1.14242196e+00 -2.43627295e-01 2.59455085e-01 8.30382884e-01
-3.99533510e-01 6.56641364e-01 -7.14852335e-03 4.36397105e-01
1.71807945e-01 2.44871471e-02 4.51150715e-01 7.11347580e-01
-2.10115820e-01 -6.59469962e-01 -1.05619237e-01 7.27472126e-01
-8.86986703e-02 -8.55578983e-04 1.53994501e-01 7.27105021e-01
7.13360012e-02 8.94662261e-01 3.13785881e-01 -6.26830101e-01
2.60614276e-01 -4.12262082e-01 3.86474021e-02 -5.52625358e-01
-5.46745837e-01 -3.77470553e-01 7.09355101e-02 -7.21643209e-01
9.00100097e-02 -5.75271070e-01 -1.17013800e+00 -5.55908859e-01
-1.23793766e-01 2.12324411e-01 5.59150875e-01 9.05869186e-01
1.02850437e+00 1.17933083e+00 1.08463764e+00 -1.23251033e+00
-2.35479698e-01 -5.21999121e-01 -5.34800172e-01 1.90822467e-01
6.15969777e-01 -1.09390128e+00 -4.30623628e-02 -3.80839139e-01] | [5.932770252227783, 1.0224945545196533] |
3f28cbd1-12de-4799-99bc-38cf3d6df855 | unsupervised-3d-shape-reconstruction-by-part | 2303.01999 | null | https://arxiv.org/abs/2303.01999v1 | https://arxiv.org/pdf/2303.01999v1.pdf | Unsupervised 3D Shape Reconstruction by Part Retrieval and Assembly | Representing a 3D shape with a set of primitives can aid perception of structure, improve robotic object manipulation, and enable editing, stylization, and compression of 3D shapes. Existing methods either use simple parametric primitives or learn a generative shape space of parts. Both have limitations: parametric primitives lead to coarse approximations, while learned parts offer too little control over the decomposition. We instead propose to decompose shapes using a library of 3D parts provided by the user, giving full control over the choice of parts. The library can contain parts with high-quality geometry that are suitable for a given category, resulting in meaningful decompositions with clean geometry. The type of decomposition can also be controlled through the choice of parts in the library. Our method works via a self-supervised approach that iteratively retrieves parts from the library and refines their placements. We show that this approach gives higher reconstruction accuracy and more desirable decompositions than existing approaches. Additionally, we show how the decomposition can be controlled through the part library by using different part libraries to reconstruct the same shapes. | ['Daniel Ritchie', 'Siddhartha Chaudhuri', 'Matthew Fisher', 'Paul Guerrero', 'Xianghao Xu'] | 2023-03-03 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Xu_Unsupervised_3D_Shape_Reconstruction_by_Part_Retrieval_and_Assembly_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Xu_Unsupervised_3D_Shape_Reconstruction_by_Part_Retrieval_and_Assembly_CVPR_2023_paper.pdf | cvpr-2023-1 | ['3d-shape-reconstruction'] | ['computer-vision'] | [-1.66032419e-01 1.87220648e-01 -2.12016866e-01 -1.08575456e-01
-3.67950082e-01 -1.11016512e+00 7.48434663e-01 1.86878473e-01
1.50716200e-01 1.50911212e-01 2.07625523e-01 5.82781807e-02
-2.06325024e-01 -9.55240607e-01 -7.21835077e-01 -4.72544104e-01
3.12871486e-01 1.25797582e+00 4.15207863e-01 -1.99460208e-01
1.93066016e-01 1.24176300e+00 -1.60005152e+00 2.51570255e-01
7.75366068e-01 9.07627165e-01 3.82298201e-01 2.31806472e-01
-4.76737082e-01 -5.82448542e-02 -4.96021360e-01 -9.35160667e-02
7.95908868e-01 -1.37913093e-01 -6.44957423e-01 6.07780337e-01
2.02767774e-02 -1.65995836e-01 1.77827582e-01 7.24698365e-01
1.14555538e-01 -5.87278828e-02 1.00678623e+00 -9.35906768e-01
-2.62718052e-01 3.64054650e-01 -2.95806646e-01 -8.44608843e-01
3.75968963e-01 5.83156683e-02 8.48536551e-01 -7.40611196e-01
9.75053370e-01 1.41489375e+00 5.64936996e-01 4.95515287e-01
-1.80223405e+00 -4.25718904e-01 1.27525717e-01 -4.51801419e-01
-1.45047200e+00 -3.88979316e-01 1.06079769e+00 -6.87738955e-01
4.06787038e-01 3.89319539e-01 8.90281439e-01 4.39583600e-01
-1.71162829e-01 7.25136220e-01 8.18063498e-01 -4.70365793e-01
4.77104545e-01 2.06766427e-01 -1.34388208e-01 7.34112799e-01
2.71944314e-01 -3.37724030e-01 -1.07338414e-01 -5.14785647e-01
1.46418500e+00 1.53070673e-01 -2.36812755e-01 -1.37866759e+00
-1.15548885e+00 5.53103566e-01 3.99354935e-01 9.04499665e-02
-4.45342243e-01 1.02280073e-01 1.59282386e-01 2.56890833e-01
2.60863602e-01 8.52071404e-01 -5.51756382e-01 1.08436927e-01
-7.60013640e-01 5.98932326e-01 9.95734155e-01 1.26514339e+00
1.06017232e+00 -1.54983506e-01 3.24079096e-02 9.36574221e-01
1.30625606e-01 3.45274329e-01 -2.87564732e-02 -1.27427042e+00
9.66992080e-02 8.55447531e-01 3.32412422e-01 -7.36842334e-01
-1.33403078e-01 -2.56404169e-02 -4.77906018e-01 7.55752802e-01
3.21601927e-01 2.38418773e-01 -1.13191330e+00 1.48791814e+00
4.65406239e-01 -5.55276871e-01 -3.92774403e-01 6.10254288e-01
4.01541412e-01 3.84541839e-01 -3.09009820e-01 2.46366978e-01
1.21018016e+00 -6.32477820e-01 -2.20345438e-01 6.20326921e-02
2.38676220e-01 -7.76806772e-01 1.09663093e+00 6.04501486e-01
-1.12469041e+00 -3.20748836e-01 -8.76578867e-01 -1.37473479e-01
-1.28498986e-01 3.76812458e-01 6.51520431e-01 5.40858805e-01
-7.87025154e-01 8.36177051e-01 -8.91452849e-01 -7.51994848e-02
5.50248384e-01 4.67264056e-01 -3.26633751e-01 -7.75552467e-02
-1.48165882e-01 7.93014765e-01 2.75205940e-01 -4.40468431e-01
-7.77000487e-01 -8.02047729e-01 -7.70986378e-01 -1.27562046e-01
4.75639939e-01 -1.10705292e+00 1.05998194e+00 -8.36949229e-01
-1.60996330e+00 8.10409606e-01 1.89076550e-02 6.92914128e-02
4.07101184e-01 -5.97488433e-02 5.70836246e-01 4.78793755e-02
-1.06958494e-01 9.33838665e-01 1.23807311e+00 -1.74018192e+00
-3.02467272e-02 -2.53136784e-01 4.44296688e-01 1.22017555e-01
9.01229307e-02 -4.97036994e-01 -6.75549090e-01 -8.62228632e-01
5.02965093e-01 -9.80543971e-01 -4.01449144e-01 6.20478690e-01
-3.46456856e-01 -1.20631516e-01 8.29457581e-01 -4.79157388e-01
6.32000506e-01 -2.12100077e+00 6.19627416e-01 4.24523801e-01
1.91820160e-01 -7.86842033e-02 -1.55530766e-01 5.74327290e-01
1.63290873e-01 2.25882277e-01 -2.49477923e-01 -5.99361300e-01
1.29879594e-01 5.54528415e-01 -1.86839104e-01 2.48094335e-01
2.42575824e-01 5.65989614e-01 -6.88063741e-01 -2.69308984e-01
3.09505314e-01 4.25024480e-01 -1.02402520e+00 2.04514876e-01
-7.25262165e-01 4.80956823e-01 -7.86867797e-01 6.46370828e-01
6.60492003e-01 1.23922050e-01 -8.56756698e-03 -4.97874796e-01
-9.00979433e-03 3.51925850e-01 -1.50819814e+00 1.79183090e+00
-5.71431100e-01 -4.03252617e-02 5.02480328e-01 -7.23083615e-01
1.28156388e+00 7.82285184e-02 6.26561821e-01 1.64859071e-01
4.11144458e-04 3.09861660e-01 -2.51572549e-01 -4.55505177e-02
3.53188396e-01 -2.82173395e-01 3.00680399e-02 6.99073195e-01
-1.23456702e-01 -1.13078034e+00 9.45761576e-02 2.05155555e-02
8.65371943e-01 5.99723637e-01 2.17391908e-01 -2.72960275e-01
5.19572794e-02 1.16215773e-01 3.75809431e-01 3.71798009e-01
6.69083714e-01 8.84512603e-01 5.05258322e-01 -3.75498354e-01
-1.43526673e+00 -1.03512454e+00 -2.24147424e-01 5.55194080e-01
1.37895793e-01 -5.41077137e-01 -5.53601384e-01 -3.41336399e-01
2.54259616e-01 5.93334556e-01 -4.99491543e-01 2.24341471e-02
-6.45388365e-01 -1.59118950e-01 7.67349545e-03 5.77020645e-01
-7.15373605e-02 -9.51227188e-01 -7.80364931e-01 -3.99861932e-02
8.61982852e-02 -8.45996559e-01 -5.94061196e-01 3.18317264e-01
-1.25346744e+00 -1.00339675e+00 -5.90892076e-01 -6.68216586e-01
1.36021829e+00 2.32911333e-01 1.06818855e+00 6.17365576e-02
-1.54203489e-01 6.35849774e-01 -4.17146444e-01 -3.05949122e-01
-7.74452507e-01 7.63485581e-02 -4.91758548e-02 -1.33366466e-01
-3.87164861e-01 -1.05565250e+00 -1.23434611e-01 2.88268983e-01
-1.00583148e+00 4.12224203e-01 5.07380009e-01 5.68113029e-01
9.55914438e-01 9.33400914e-02 -4.76995036e-02 -6.35584593e-01
5.24549663e-01 4.25467268e-03 -5.98192513e-01 1.29025996e-01
-1.25868514e-01 7.43553400e-01 7.14668393e-01 -7.21722126e-01
-7.53906488e-01 5.67860126e-01 2.23006487e-01 -9.32802856e-01
-2.41707921e-01 1.79191872e-01 -4.97011781e-01 -1.55109242e-01
5.42521596e-01 5.16120829e-02 3.74135256e-01 -1.05035591e+00
7.36689031e-01 2.00298503e-01 3.78042728e-01 -1.17072105e+00
8.32149982e-01 4.72318918e-01 7.81985000e-02 -7.54351377e-01
-1.68109834e-01 -1.75200880e-01 -1.13350737e+00 1.21533342e-01
3.82061154e-01 -5.68640053e-01 -3.15225124e-01 7.41229802e-02
-1.24115753e+00 -3.93769324e-01 -6.67180121e-01 -2.71088686e-02
-7.60386527e-01 3.26372921e-01 -2.75659204e-01 -6.64435863e-01
-1.26994580e-01 -1.32842958e+00 1.44194901e+00 -3.28513682e-01
-5.58003485e-01 -5.90115070e-01 -1.23857863e-01 -1.28509283e-01
8.23544040e-02 4.11329627e-01 1.34943485e+00 -2.04487979e-01
-9.47787344e-01 -3.43397319e-01 2.14686006e-01 7.52695724e-02
5.03519416e-01 1.60390392e-01 -6.07537627e-01 -1.86468452e-01
-1.41957417e-01 -6.70012161e-02 6.40582144e-01 2.28113681e-01
1.37977421e+00 -4.72957671e-01 -5.87325156e-01 6.49913132e-01
1.16746414e+00 -1.28619547e-03 5.03895760e-01 1.68688670e-01
7.55353689e-01 7.03361690e-01 1.95795521e-01 4.50027883e-01
2.37277448e-01 8.80656898e-01 4.01777893e-01 1.48147330e-01
-2.89861143e-01 -4.25352573e-01 5.29121086e-02 4.21392590e-01
-2.86668837e-01 3.59318405e-01 -7.64304996e-01 3.44835371e-01
-1.63592398e+00 -5.33467054e-01 2.48502210e-01 2.44263029e+00
9.80820179e-01 4.21083570e-02 4.39227432e-01 2.16917902e-01
3.16198111e-01 -1.84621140e-01 -4.44405913e-01 -2.97391295e-01
3.49387407e-01 1.52512521e-01 4.45394665e-01 5.46197891e-01
-7.40704775e-01 7.69244552e-01 6.54147196e+00 8.96129906e-01
-1.05366552e+00 -3.29025030e-01 1.91494793e-01 -5.87735437e-02
-9.45792317e-01 2.64276385e-01 -6.51542068e-01 5.58523424e-02
-4.79960293e-02 5.03850542e-02 6.43441498e-01 8.64877045e-01
7.49323294e-02 3.85230891e-02 -1.29924977e+00 8.64922166e-01
-1.51844233e-01 -1.41595054e+00 4.67799932e-01 1.36763409e-01
5.72852314e-01 -5.04631698e-01 -3.42109025e-01 -1.24848306e-01
3.38307083e-01 -8.32502782e-01 1.15533710e+00 6.90635145e-01
6.43829942e-01 -7.49047279e-01 3.26241367e-02 8.17373335e-01
-1.15819943e+00 1.35553837e-01 -4.11985695e-01 1.62674546e-01
2.90320404e-02 6.57588303e-01 -8.19124937e-01 4.87846583e-01
3.58647168e-01 3.19253564e-01 -5.02420545e-01 1.03901601e+00
-1.44340694e-01 1.77776352e-01 -8.49829316e-01 -3.15071754e-02
-2.64491260e-01 -5.36744893e-01 8.69988084e-01 8.85450542e-01
1.73560485e-01 2.22029597e-01 6.34525418e-01 1.33671772e+00
1.18103221e-01 2.09029820e-02 -5.95645964e-01 1.32730184e-02
6.84036911e-01 1.09977031e+00 -9.65689600e-01 -2.06309721e-01
1.11915551e-01 7.37420857e-01 3.22616071e-01 1.59883842e-01
-2.73372680e-01 -2.33263940e-01 5.37681699e-01 6.45646513e-01
6.26275599e-01 -7.53180265e-01 -7.67001033e-01 -9.88672912e-01
1.01108491e-01 -9.38951075e-01 -1.43947363e-01 -7.82853365e-01
-9.76678967e-01 4.53196138e-01 5.56865573e-01 -1.30901861e+00
-2.20912591e-01 -5.95536232e-01 -4.39593971e-01 7.11710036e-01
-7.72670567e-01 -1.27092445e+00 -7.00917393e-02 1.87430829e-01
5.40374637e-01 3.47862430e-02 8.53886127e-01 -1.46162137e-01
9.13738273e-03 7.48117268e-02 -2.41927445e-01 -1.87984586e-01
3.47256303e-01 -1.21719706e+00 3.17824990e-01 3.75682205e-01
3.11386615e-01 7.91775644e-01 7.38549769e-01 -6.96733832e-01
-1.48810899e+00 -7.38328934e-01 3.81203294e-01 -5.71725547e-01
5.44755235e-02 -4.95085597e-01 -9.40266848e-01 4.40164208e-01
-3.38115275e-01 -2.19032645e-01 1.97440058e-01 2.03974500e-01
-4.64474022e-01 -9.68564395e-03 -1.28386235e+00 7.25679517e-01
9.74883199e-01 -3.91048729e-01 -6.23477578e-01 5.41570969e-02
6.05750263e-01 -5.08175790e-01 -7.63781071e-01 3.81014585e-01
6.49431109e-01 -7.06236005e-01 1.22147453e+00 -3.26040506e-01
1.56488270e-01 -5.93771219e-01 9.54776406e-02 -1.38030696e+00
-4.10790384e-01 -5.34393013e-01 -1.62461549e-01 9.43401933e-01
2.25816652e-01 -2.52958149e-01 7.56350875e-01 8.10644031e-01
-2.02572316e-01 -9.12623286e-01 -5.12379706e-01 -8.31903875e-01
1.42853744e-02 -3.60306084e-01 9.74090815e-01 6.13769948e-01
-2.86907822e-01 3.80038098e-02 1.31151840e-01 -2.87346058e-02
5.35929263e-01 6.02366686e-01 1.19643092e+00 -1.46526766e+00
-4.59479809e-01 -7.37617671e-01 -2.38622531e-01 -1.50134361e+00
-4.44687083e-02 -8.34386170e-01 1.09296128e-01 -1.50914693e+00
-1.50144966e-02 -1.07576764e+00 5.77429354e-01 8.64484608e-01
2.11181283e-01 -7.54395276e-02 2.43468955e-01 4.13305432e-01
8.49333331e-02 6.47639215e-01 1.48618972e+00 -7.47579038e-02
-6.03344977e-01 1.65568590e-01 -5.63522756e-01 9.24691916e-01
5.77338636e-01 -2.74566323e-01 -3.59191895e-01 -5.50949693e-01
-1.20898180e-01 -5.01400977e-02 3.28201532e-01 -7.32265115e-01
-1.14556864e-01 -2.97369897e-01 3.66087615e-01 -7.15793788e-01
6.13463223e-01 -8.81761432e-01 5.98778069e-01 2.21122637e-01
-2.60946453e-01 -1.13911301e-01 2.22148508e-01 3.33941400e-01
2.66368628e-01 -5.66800237e-01 7.88955748e-01 -4.33197230e-01
-5.70800863e-02 3.13825071e-01 -1.44087955e-01 -3.79046649e-01
7.20384836e-01 -5.82141697e-01 5.41133165e-01 -3.86166036e-01
-7.25482047e-01 -3.26629751e-03 1.06451690e+00 4.27171797e-01
4.82027829e-01 -1.47479892e+00 -3.13786745e-01 3.70907158e-01
-7.28879794e-02 6.14101648e-01 -5.95887423e-01 3.19758087e-01
-6.12887084e-01 5.18771149e-02 -2.09475428e-01 -6.54543459e-01
-1.03706729e+00 5.52562952e-01 2.81870961e-01 1.44186437e-01
-8.48064125e-01 4.52688217e-01 2.42145658e-01 -6.84386909e-01
1.10208385e-01 -6.97328031e-01 1.37710661e-01 -3.50696594e-02
8.02860688e-03 2.53660530e-01 -3.61450128e-02 -4.41993415e-01
-1.44645587e-01 9.41105783e-01 2.47735545e-01 -1.82680055e-01
1.54613423e+00 1.88405529e-01 -2.58811504e-01 3.51736575e-01
8.57101917e-01 4.10194397e-01 -1.42053318e+00 -1.09133311e-01
-2.13499889e-01 -5.90576589e-01 -1.25163883e-01 -5.40036142e-01
-9.19645488e-01 6.12288952e-01 -2.19828755e-01 1.13228522e-01
9.14051175e-01 3.64352345e-01 5.47457278e-01 3.71548086e-01
8.56619835e-01 -6.82268143e-01 1.00273751e-01 3.65653515e-01
1.41712046e+00 -4.43234175e-01 2.96146065e-01 -8.84016812e-01
-4.36069995e-01 1.33636558e+00 2.44849980e-01 -5.02642453e-01
7.26185918e-01 4.57085580e-01 -3.33006680e-01 -2.82300711e-01
-3.30553979e-01 -1.31030411e-01 7.22939491e-01 7.63769031e-01
7.30840042e-02 1.51228622e-01 -5.72562590e-02 5.31655371e-01
-4.59358573e-01 -2.51522183e-01 2.56839871e-01 9.14899528e-01
-4.67225581e-01 -1.72941935e+00 -6.21058524e-01 4.79039431e-01
2.07890287e-01 4.24539059e-01 -5.72026372e-01 5.46840727e-01
2.26409942e-01 2.54932553e-01 -2.74957325e-02 -1.53215379e-01
4.75365072e-01 3.51615250e-02 1.02304840e+00 -1.01484549e+00
-1.90687478e-01 3.98640811e-01 9.22478437e-02 -4.57509995e-01
-2.32397556e-01 -8.75377119e-01 -1.08254123e+00 4.97548543e-02
-3.32887888e-01 -1.63244978e-02 4.64812875e-01 8.25175583e-01
4.81116116e-01 8.33523273e-02 5.70325255e-01 -1.40871263e+00
-5.89268148e-01 -5.19928575e-01 -4.27262723e-01 3.66252571e-01
1.58499241e-01 -1.00455892e+00 9.13440157e-03 2.65163422e-01] | [8.768264770507812, -3.6005115509033203] |
1540d753-af6d-49b7-9de9-6b6cf62119b5 | t360rrd-a-dataset-for-360-degree-rotated | 2303.01894 | null | https://arxiv.org/abs/2303.01894v3 | https://arxiv.org/pdf/2303.01894v3.pdf | TRR360D: A dataset for 360 degree rotated rectangular box table detection | To address the problem of scarcity and high annotation costs of rotated image table detection datasets, this paper proposes a method for building a rotated image table detection dataset. Based on the ICDAR2019MTD modern table detection dataset, we refer to the annotation format of the DOTA dataset to create the TRR360D rotated table detection dataset. The training set contains 600 rotated images and 977 annotated instances, and the test set contains 240 rotated images and 499 annotated instances. The AP50(T<90) evaluation metric is defined, and this dataset is available for future researchers to study rotated table detection algorithms and promote the development of table detection technology. The TRR360D rotated table detection dataset was created by constraining the starting point and annotation direction, and is publicly available at https://github.com/vansin/TRR360D. | ['Minglei Tong', 'Wenxing Hu'] | 2023-03-03 | null | null | null | null | ['2d-object-detection', 'table-detection'] | ['computer-vision', 'miscellaneous'] | [ 2.04743624e-01 -2.63302810e-02 -4.43137020e-01 -1.54089242e-01
-9.43251729e-01 -8.46216142e-01 3.43034565e-01 -6.64023608e-02
-1.20722599e-01 2.60283649e-01 1.87192619e-01 -1.20024957e-01
-7.75380954e-02 -8.66257250e-01 -5.22960901e-01 -3.81779134e-01
1.64903075e-01 6.47020578e-01 9.45054814e-02 -5.79473823e-02
2.33071581e-01 3.40523303e-01 -1.40214336e+00 5.15603483e-01
5.26983202e-01 1.27289832e+00 2.80206650e-01 5.78711271e-01
2.51579821e-01 8.77068758e-01 -9.77796853e-01 -6.50209606e-01
7.46245980e-01 -5.19733846e-01 -5.72485685e-01 2.45261878e-01
3.79040748e-01 -4.38397378e-01 -7.03026533e-01 9.08996582e-01
6.21009588e-01 -7.39798099e-02 5.40496826e-01 -1.40384436e+00
-9.47948694e-01 8.59222472e-01 -8.87118101e-01 4.26491767e-01
6.10196233e-01 -3.13210845e-01 8.41561139e-01 -1.16920280e+00
1.05859351e+00 9.65027511e-01 3.92759353e-01 2.70240009e-01
-7.26441860e-01 -1.05895829e+00 -4.04256098e-02 3.45475227e-01
-2.08540177e+00 -3.05463880e-01 4.82944191e-01 -4.27838892e-01
6.16407096e-01 4.66883242e-01 6.22015357e-01 1.06898928e+00
-7.92607814e-02 9.27773058e-01 4.49175805e-01 -3.60551208e-01
-9.30654109e-02 9.32097659e-02 -1.30304173e-01 5.79004765e-01
8.05094004e-01 -5.00490427e-01 -7.28378475e-01 2.00014576e-01
8.47382605e-01 -2.20474347e-01 -1.74997419e-01 -3.04859489e-01
-1.59808064e+00 6.19664192e-01 3.16636860e-01 -1.84557922e-02
-3.64674665e-02 -5.83112776e-01 6.21259630e-01 7.11729452e-02
3.50760996e-01 3.62211227e-01 -4.32590902e-01 -4.01422232e-02
-6.10639334e-01 2.50454724e-01 2.05076456e-01 1.64545429e+00
4.67186660e-01 -1.93459406e-01 -3.71393621e-01 9.43046510e-01
3.92351002e-02 7.84027636e-01 2.62886375e-01 -8.91277492e-01
9.18146372e-01 9.77004766e-01 2.00201757e-02 -1.23947990e+00
-3.73790413e-01 7.90113434e-02 -1.04907966e+00 -6.68291807e-01
3.47835422e-01 3.70597661e-01 -9.51863289e-01 9.07753348e-01
2.53532082e-01 -5.23881078e-01 -1.19460374e-02 8.12341154e-01
1.28153336e+00 5.59708595e-01 -5.29562056e-01 -5.91010191e-02
1.86465120e+00 -6.06791019e-01 -1.16457427e+00 9.46331471e-02
8.19054484e-01 -1.05402577e+00 1.23289299e+00 6.67389214e-01
-6.69117033e-01 -5.24552941e-01 -1.14255250e+00 -3.17525089e-01
-7.11264789e-01 8.62082362e-01 3.66709977e-01 4.94758934e-01
-6.21833980e-01 -4.18089122e-01 -3.90606403e-01 -4.11532938e-01
6.19136214e-01 2.08496433e-02 -5.15263796e-01 -4.03183162e-01
-1.10646582e+00 7.55825698e-01 5.83748877e-01 4.21691865e-01
-5.87553144e-01 -4.06836629e-01 -8.96742940e-01 -4.04885709e-01
8.36427152e-01 -2.44135216e-01 8.64547312e-01 -1.85821038e-02
-8.19689691e-01 1.41809547e+00 3.21048856e-01 -1.88537344e-01
7.26330459e-01 -1.58981532e-01 -5.14813960e-01 2.18152791e-01
4.44632560e-01 7.60383964e-01 3.67768526e-01 -1.03254044e+00
-5.57865977e-01 -4.95137542e-01 -3.29624206e-01 4.21809763e-01
-1.92956120e-01 1.61035284e-01 -1.27892172e+00 -1.05335128e+00
2.86252737e-01 -9.73927140e-01 2.53468752e-01 -2.81180561e-01
-8.48042488e-01 -6.03757538e-02 7.79689729e-01 -8.13990951e-01
1.44437253e+00 -2.09683490e+00 -2.86017299e-01 2.08279803e-01
1.04315609e-01 3.58440503e-02 -2.46850494e-02 2.84305900e-01
-2.65132576e-01 3.02913338e-01 -2.22696170e-01 8.74975175e-02
-2.60661274e-01 -6.49850518e-02 -2.96843767e-01 5.00717402e-01
-3.72951501e-03 6.25027061e-01 -5.67617178e-01 -8.92920852e-01
5.05756021e-01 4.45743144e-01 -3.02090019e-01 -2.60012690e-04
3.00134689e-01 2.11682811e-01 -4.42234337e-01 1.15203702e+00
9.36844230e-01 -6.29134625e-02 3.38869601e-01 -6.56278133e-01
1.96135361e-02 2.64251679e-02 -1.47968507e+00 1.67758131e+00
1.58518478e-01 6.62569165e-01 -5.22959709e-01 -5.38535118e-01
1.32600212e+00 2.01142371e-01 8.44709933e-01 -1.05196559e+00
2.71400958e-01 -8.30317587e-02 -2.24049971e-01 -4.89148885e-01
1.01411331e+00 8.26298714e-01 -4.48274404e-01 -2.68835537e-02
-2.69953072e-01 -3.00375491e-01 7.82596946e-01 4.32229877e-01
9.56723690e-01 2.68834651e-01 4.69694167e-01 -7.53372237e-02
5.75051725e-01 4.45807606e-01 6.84271514e-01 7.37591803e-01
-3.73115540e-01 1.04986894e+00 4.92544144e-01 -6.55389309e-01
-1.25534654e+00 -1.04261577e+00 -6.27046287e-01 7.51652181e-01
3.62798661e-01 -8.56296957e-01 -7.23002136e-01 -5.10481358e-01
-6.85503930e-02 3.27079952e-01 -8.57639372e-01 8.67134109e-02
-5.74794114e-01 -9.14555609e-01 8.80379319e-01 6.10238194e-01
8.40030909e-01 -1.01713586e+00 -3.45271975e-01 -2.67968148e-01
-8.29248190e-01 -1.50131536e+00 -5.82127333e-01 1.95119247e-01
-3.04190546e-01 -1.40997708e+00 -4.49808031e-01 -7.05905080e-01
7.88734436e-01 3.93844068e-01 1.18905818e+00 -9.29042622e-02
-6.13420188e-01 4.38739657e-02 -6.81701899e-01 -6.29623473e-01
-1.72177657e-01 2.57269621e-01 6.08360097e-02 -2.83735335e-01
5.08898020e-01 3.28138471e-01 -4.62569088e-01 8.87500226e-01
-8.76417398e-01 1.84114844e-01 3.98694366e-01 4.27575976e-01
1.15828538e+00 1.86987162e-01 1.97708338e-01 -8.03676426e-01
3.93608212e-01 -1.75180450e-01 -1.07402611e+00 3.84432644e-01
-6.01251841e-01 -3.25544745e-01 3.24846447e-01 -1.37357622e-01
-8.97483170e-01 4.10161048e-01 6.29340410e-02 -4.26667273e-01
6.88602552e-02 1.10123940e-01 -5.71819305e-01 5.90709686e-01
6.17481291e-01 8.28365311e-02 -5.20337045e-01 -4.82498765e-01
3.40556085e-01 9.20314968e-01 8.59448433e-01 -2.92746395e-01
9.42079604e-01 4.84530568e-01 -3.47396344e-01 -6.10337377e-01
-8.17620277e-01 -5.56478381e-01 -8.09936345e-01 -4.62177575e-01
1.03435636e+00 -1.18103075e+00 -4.66553032e-01 3.29345077e-01
-7.04331815e-01 -2.16042493e-02 -1.93631187e-01 4.09816146e-01
-3.78046572e-01 2.96220854e-02 -4.89436179e-01 -6.38040602e-01
-3.60428125e-01 -1.14119577e+00 1.42563295e+00 1.77299287e-02
-1.32385716e-01 -2.87370145e-01 -1.39999852e-01 8.39324594e-01
-2.54552573e-01 3.31021577e-01 5.28644741e-01 -5.93977153e-01
-4.86335754e-01 -5.55381417e-01 -2.98985124e-01 -3.12724262e-02
1.96204439e-01 2.27113232e-01 -7.82355249e-01 1.30020827e-01
-3.41825604e-01 -2.59773701e-01 7.37336040e-01 3.08856994e-01
1.48019922e+00 -7.62026608e-02 -2.84122974e-01 5.50321817e-01
1.18243074e+00 7.19999373e-01 1.14494610e+00 8.84075701e-01
9.08323586e-01 3.37279201e-01 1.37574291e+00 5.48608840e-01
7.17505395e-01 7.77570605e-01 2.30922446e-01 -7.69428164e-02
-1.03815652e-01 -3.01837295e-01 -1.47092585e-02 8.00462782e-01
-2.77423300e-02 -7.53060818e-01 -1.21637881e+00 3.02154869e-01
-1.52307200e+00 -9.34428751e-01 -3.75750899e-01 2.11778450e+00
5.35995245e-01 1.54736027e-01 2.19150037e-01 6.74467444e-01
1.01723874e+00 1.13812745e-01 -4.05622989e-01 -1.33324847e-01
-4.57505375e-01 -2.99664915e-01 7.36056447e-01 7.87956566e-02
-1.39032340e+00 8.59584570e-01 6.55539131e+00 9.47287440e-01
-5.76387763e-01 -2.95146227e-01 8.03820491e-01 1.04333244e-01
1.88421965e-01 -4.62447524e-01 -1.07020426e+00 2.57349938e-01
6.01041555e-01 -1.49882182e-01 2.92868853e-01 9.63924825e-01
6.64046854e-02 -1.96514145e-01 -7.84492493e-01 1.36549294e+00
3.52988988e-01 -1.12590086e+00 2.60932744e-01 1.93937391e-01
4.11713719e-01 -2.25942060e-01 1.87255904e-01 2.28183553e-01
-5.73284850e-02 -1.00739729e+00 7.58431911e-01 1.30258705e-02
1.35029471e+00 -8.00828278e-01 8.99075985e-01 -9.42695439e-02
-1.71100211e+00 1.44505203e-01 -4.43738133e-01 4.60719645e-01
-5.59406936e-01 2.62201101e-01 -1.03427875e+00 8.69914591e-01
1.31785643e+00 8.87841821e-01 -1.22550082e+00 6.77165449e-01
-8.93276334e-02 2.61632979e-01 -2.62930274e-01 4.21253473e-01
-4.42510068e-01 -3.97856265e-01 1.46162346e-01 9.61505949e-01
2.54545838e-01 -1.49989009e-01 1.14035182e-01 3.30366015e-01
-4.40932900e-01 2.89250106e-01 -7.63499439e-01 1.68690041e-01
8.80719006e-01 1.44286096e+00 -1.50951195e+00 -3.71537775e-01
-2.63864845e-01 6.86166167e-01 -2.45339632e-01 -3.00645293e-03
-1.17261791e+00 -4.95584875e-01 2.39236519e-01 5.97065166e-02
3.20016205e-01 1.22963205e-01 -5.17508805e-01 -1.42439711e+00
1.70067444e-01 -1.24296558e+00 9.32409763e-01 -1.08861375e+00
-9.16649878e-01 7.30952382e-01 3.54171604e-01 -1.67098951e+00
1.17012434e-01 -5.62776923e-01 9.36429128e-02 3.18292350e-01
-7.97037423e-01 -1.07165790e+00 -7.88197160e-01 6.08570755e-01
5.21157801e-01 -2.60722727e-01 7.40616441e-01 6.99639022e-01
-1.05137694e+00 9.01229262e-01 -2.11946890e-02 7.71882892e-01
9.05103207e-01 -1.08405495e+00 4.84970510e-01 9.41475332e-01
1.32498354e-01 3.35604548e-01 6.12427831e-01 -7.52547979e-01
-1.60250020e+00 -1.35708487e+00 4.96746778e-01 -9.76555884e-01
3.28158915e-01 -7.38166034e-01 -6.79925382e-01 9.23143148e-01
-6.04716614e-02 2.00650781e-01 5.80921233e-01 -3.78890812e-01
-3.22531164e-01 -4.65818912e-01 -1.33598816e+00 4.65902299e-01
1.19815230e+00 -4.42713767e-01 -2.27495834e-01 4.55134809e-01
3.80595088e-01 -8.46633971e-01 -1.15839446e+00 6.13821924e-01
5.87081373e-01 -5.33687949e-01 9.83032167e-01 4.74522740e-01
1.03552639e-01 -6.39505029e-01 -4.58309770e-01 -4.26832110e-01
-1.59324035e-02 -1.28810152e-01 -3.06370743e-02 1.36187339e+00
3.77039760e-01 -1.70831650e-01 8.25794876e-01 3.06867044e-02
-7.18735158e-02 -5.95212460e-01 -8.13171446e-01 -7.85854280e-01
-2.62094587e-01 -3.52756530e-01 7.28146672e-01 9.47057784e-01
-4.60752398e-01 1.68819994e-01 -4.43070561e-01 1.46807015e-01
4.51021075e-01 5.10440394e-02 1.03141928e+00 -7.98307657e-01
4.69629198e-01 -1.54071167e-01 -4.70773995e-01 -7.92290568e-01
-5.05751491e-01 -8.52402270e-01 -8.22177082e-02 -1.60118103e+00
3.69895965e-01 -1.95558190e-01 -1.06376052e-01 5.38827479e-01
-1.12893701e-01 1.01573074e+00 2.73369342e-01 5.04984856e-01
-7.83374369e-01 1.70982137e-01 1.35931313e+00 -1.97517663e-01
-5.94987348e-02 -4.61170197e-01 -7.19037056e-01 5.27311742e-01
8.25157821e-01 -6.40492558e-01 -4.13068295e-01 -1.61750138e-01
2.39872575e-01 -1.26385674e-01 -1.31582722e-01 -9.30962145e-01
-1.44187048e-01 1.58135258e-02 9.46062744e-01 -1.59141302e+00
2.04010174e-01 -7.01787651e-01 2.34750032e-01 5.27137160e-01
-2.16673642e-01 5.12211144e-01 1.44399181e-01 2.42197931e-01
-2.53130734e-01 2.41358012e-01 5.38302422e-01 -2.92641371e-02
-6.98369861e-01 4.19715852e-01 -4.48085010e-01 3.24827611e-01
1.16319585e+00 -4.86104578e-01 -5.23630083e-01 -8.74014571e-02
-3.41955990e-01 4.32925403e-01 5.57557702e-01 8.33412170e-01
7.17504263e-01 -1.46806645e+00 -6.16547287e-01 1.27739698e-01
7.41944551e-01 3.06175351e-01 9.48923528e-02 4.17423517e-01
-1.06997132e+00 5.40951312e-01 -2.15342477e-01 -7.41435587e-01
-1.36465669e+00 7.82116711e-01 1.37818679e-01 -1.64487794e-01
-9.99497652e-01 5.82285285e-01 8.81007537e-02 -5.85221171e-01
3.29254270e-01 -3.73520613e-01 -3.56848985e-01 4.27017003e-01
5.59742153e-01 3.46435726e-01 3.94661546e-01 -6.75526679e-01
-8.40219855e-01 5.77342331e-01 -1.70094982e-01 2.25413725e-01
1.09112751e+00 -2.19299868e-01 3.00189048e-01 4.08537179e-01
8.69808197e-01 1.42392069e-01 -6.73507869e-01 1.31184265e-01
9.77324992e-02 -7.79288888e-01 -4.01137382e-01 -8.35412621e-01
-1.07434189e+00 1.23395048e-01 8.13991964e-01 9.19958130e-02
1.37606466e+00 8.52631032e-02 4.89401579e-01 5.02978206e-01
2.60548145e-01 -1.19271600e+00 1.20429821e-01 2.50558019e-01
1.05607271e+00 -1.41013551e+00 4.88590240e-01 -7.51396954e-01
-9.99309361e-01 1.02088523e+00 8.41792107e-01 1.08215131e-01
2.69051611e-01 6.61615193e-01 4.06958163e-01 -2.56957471e-01
-4.54243124e-01 -1.15379632e-01 2.53147095e-01 7.11381674e-01
4.48256940e-01 -1.03385290e-02 -2.92569131e-01 4.43650305e-01
-7.18612492e-01 -1.75894499e-01 7.38519311e-01 9.76849020e-01
1.02960736e-01 -8.92250121e-01 -7.87503719e-01 4.92753685e-01
-5.08770406e-01 1.06078282e-01 -5.56804717e-01 1.23525107e+00
4.41245347e-01 9.65851009e-01 2.44107544e-01 -7.43874192e-01
4.93024439e-01 -2.32574552e-01 3.10096204e-01 -4.26450342e-01
-3.49275202e-01 1.69561386e-01 1.68222561e-01 -3.56039047e-01
-4.07384992e-01 -5.84194422e-01 -1.24032187e+00 -3.50610226e-01
-3.80302012e-01 -3.65916902e-04 4.74368870e-01 3.47930342e-01
3.57831746e-01 6.50086403e-01 6.16506755e-01 -3.60119373e-01
4.86195982e-02 -9.75534201e-01 -5.73540688e-01 2.57513970e-01
-1.50890604e-01 -7.56583989e-01 -2.37151086e-02 2.84229964e-01] | [11.697653770446777, 2.985502004623413] |
a2a35da1-0332-4e91-8c23-dbfe860a28fb | recognising-known-configurations-of-garments | 2205.00225 | null | https://arxiv.org/abs/2205.00225v2 | https://arxiv.org/pdf/2205.00225v2.pdf | Recognising Known Configurations of Garments For Dual-Arm Robotic Flattening | Robotic deformable-object manipulation is a challenge in the robotic industry because deformable objects have complicated and various object states. Predicting those object states and updating manipulation planning is time-consuming and computationally expensive. In this paper, we propose learning known configurations of garments to allow a robot to recognise garment states and choose a pre-designed manipulation plan for garment flattening. | ['Gerardo Argon-Camarasa', 'Li Duan'] | 2022-04-30 | null | null | null | null | ['deformable-object-manipulation'] | ['robots'] | [-5.72642609e-02 3.06956265e-02 -4.03461903e-01 -3.58481616e-01
-8.03826079e-02 -8.32264364e-01 9.36328024e-02 -5.55427074e-01
-2.61737686e-02 6.99930847e-01 -2.45422021e-01 2.46423900e-01
-5.13119459e-01 -7.49512792e-01 -8.05302739e-01 -6.18124902e-01
-3.53304476e-01 1.24670315e+00 5.15819073e-01 -4.64019567e-01
2.19685882e-01 1.29788828e+00 -1.37282360e+00 1.22167341e-01
3.45014364e-01 5.19804120e-01 8.65945816e-01 9.97126281e-01
4.72013295e-01 3.59184444e-01 -2.50378102e-01 -2.22238316e-03
6.53788507e-01 6.64235801e-02 -9.70263481e-01 3.78755212e-01
-6.32346980e-03 -4.73297209e-01 -6.43084884e-01 9.58747149e-01
1.27813697e-01 7.28895426e-01 6.68634355e-01 -1.03339136e+00
-4.84427899e-01 8.32597733e-01 2.27316022e-01 -1.38449907e-01
3.62982154e-01 4.47480589e-01 3.80332738e-01 -5.78348756e-01
9.72672582e-01 1.57148826e+00 2.87798494e-01 7.07022846e-01
-6.02234781e-01 -3.33720028e-01 3.58303279e-01 2.19583020e-01
-1.15797222e+00 -2.19215050e-01 6.82163477e-01 -3.96313101e-01
1.10945559e+00 5.22791564e-01 9.12220955e-01 7.39799678e-01
8.20601106e-01 5.45283198e-01 6.79653168e-01 -2.85471439e-01
2.02651322e-01 -5.81806660e-01 -4.17289644e-01 8.47993016e-01
3.09769273e-01 1.08194193e-02 2.04885736e-01 -2.64049377e-02
1.39845955e+00 4.28012460e-01 -4.26718742e-02 -6.15094662e-01
-1.27476203e+00 4.39971775e-01 4.88924056e-01 -1.72261558e-02
-7.11002111e-01 6.17217362e-01 8.13108981e-02 1.86019570e-01
-7.00525716e-02 8.00052762e-01 -5.78850865e-01 -2.09382832e-01
-1.40446156e-01 7.85391271e-01 1.09621179e+00 1.36255252e+00
4.14398789e-01 1.14089400e-01 -2.63126735e-02 2.69437134e-01
4.56116229e-01 6.19082570e-01 -6.87553510e-02 -8.97656441e-01
3.75916004e-01 5.56796134e-01 5.40425777e-01 -8.71952176e-01
-5.39850831e-01 5.82508922e-01 -4.36756223e-01 4.20861870e-01
6.89564645e-02 -6.58484399e-02 -1.37550199e+00 8.68258297e-01
5.40997088e-01 -1.59518152e-01 5.69712035e-02 1.28064525e+00
3.53866279e-01 5.46466708e-01 1.40650302e-01 1.06274642e-01
8.68797183e-01 -7.66021013e-01 -8.09368789e-01 -3.72502744e-01
-9.68645066e-02 -7.28773117e-01 5.80570400e-01 4.55137759e-01
-1.05378067e+00 -4.14451391e-01 -8.75510693e-01 -8.45560879e-02
-4.47525904e-02 2.77243197e-01 9.67981219e-01 -1.98765472e-01
-4.11473513e-01 1.18521035e+00 -1.67335904e+00 -1.28743380e-01
1.60851583e-01 7.54952252e-01 -2.42964804e-01 -1.44207045e-01
-6.56586409e-01 1.67714894e+00 9.09057796e-01 6.48453474e-01
-1.14798415e+00 9.52965990e-02 -6.93244159e-01 -3.89256537e-01
6.51262343e-01 -4.51388121e-01 1.41410732e+00 -2.32442185e-01
-2.16840005e+00 5.42279840e-01 4.17553186e-01 -5.18375337e-02
5.12957394e-01 -5.03977716e-01 -1.74635518e-02 1.14231529e-02
-5.45924604e-01 1.66386515e-01 1.12797403e+00 -1.28495920e+00
-4.89772595e-02 -2.23344132e-01 1.85349137e-01 5.59823573e-01
5.74877083e-01 -1.18172139e-01 -3.68878007e-01 -5.04808486e-01
6.96059167e-01 -1.46148610e+00 -6.68641567e-01 1.33357152e-01
-4.95306045e-01 -4.68017489e-01 1.08752120e+00 -6.98636711e-01
5.34175634e-01 -1.58471179e+00 9.10617471e-01 2.98031479e-01
-2.19300434e-01 2.33297095e-01 4.27139327e-02 6.18222892e-01
3.23062092e-01 -4.49526459e-01 2.85776079e-01 3.58699828e-01
7.08730519e-02 8.45149159e-01 -3.20769757e-01 5.23621976e-01
2.33364955e-01 1.10735869e+00 -1.03563333e+00 -2.28967786e-01
5.10829389e-01 1.64360762e-01 -5.40026367e-01 4.17048573e-01
-7.21923470e-01 4.88943368e-01 -1.20393515e+00 1.06992912e+00
4.87866580e-01 2.21120059e-01 4.71934229e-01 -3.14955384e-01
-2.49893084e-01 1.66847900e-01 -1.37311471e+00 1.52767122e+00
-1.87856674e-01 -1.09613240e-02 1.54044390e-01 -9.53659236e-01
1.07403064e+00 2.51961946e-01 4.51169789e-01 4.00991976e-01
6.50432467e-01 2.48167515e-01 -8.00000690e-03 -1.03126132e+00
6.91695571e-01 -1.04681909e-01 -1.67149857e-01 2.82485813e-01
-5.17088398e-02 -1.50977564e+00 -1.82472780e-01 -6.29528582e-01
1.08176804e+00 6.75933301e-01 3.91312897e-01 -2.02824682e-01
2.44849920e-03 3.36488456e-01 1.51713401e-01 3.80117059e-01
1.08985379e-01 2.63442844e-01 -2.11161062e-01 -8.49968672e-01
-9.90902781e-01 -1.10407305e+00 2.10443169e-01 8.38310361e-01
7.70653188e-01 2.64566064e-01 -2.84008056e-01 -2.85355568e-01
3.77126426e-01 4.50635433e-01 -1.19223073e-01 -2.18680546e-01
-1.52666211e+00 -2.59564996e-01 -2.51471639e-01 6.96815789e-01
1.24715753e-01 -1.50143027e+00 -1.04303718e+00 3.56963277e-01
5.17934468e-03 -9.08909559e-01 -3.44089448e-01 2.07961813e-01
-1.06754208e+00 -1.23255312e+00 -5.01213968e-01 -1.08136499e+00
9.93415833e-01 -6.81890920e-03 6.73360467e-01 7.50967637e-02
-6.83479488e-01 3.41760159e-01 -4.32349175e-01 -5.79558253e-01
-7.72502065e-01 3.85589860e-02 2.66732663e-01 -6.94971561e-01
-3.80567700e-01 -2.00600512e-02 -2.41860792e-01 6.91735983e-01
-5.33362746e-01 4.10213545e-02 7.93073356e-01 3.82575750e-01
9.85408664e-01 3.22049379e-01 -3.84184681e-02 -1.32009253e-01
3.73801589e-01 -4.62461449e-02 -8.57976139e-01 3.70125473e-01
1.61702618e-01 3.78143564e-02 3.03092599e-01 -1.17112482e+00
-1.04076588e+00 7.38948584e-01 1.59303978e-01 -8.04693639e-01
-8.27110466e-03 3.94600749e-01 -4.97173332e-03 -3.01263303e-01
4.25531268e-01 -2.32233018e-01 -2.25392114e-02 -6.51092231e-01
4.02311593e-01 3.16666037e-01 4.98555660e-01 -7.79224336e-01
1.08649170e+00 2.90384322e-01 2.06902832e-01 -3.89478326e-01
-3.98508281e-01 -1.38909236e-01 -1.07605016e+00 -5.75631320e-01
8.36553097e-01 -3.41580331e-01 -1.00190735e+00 5.46337605e-01
-1.21333349e+00 -7.44752467e-01 -5.20561278e-01 6.62456095e-01
-9.30287957e-01 3.74929421e-02 -5.77157676e-01 -8.42285335e-01
-3.47785592e-01 -1.29100156e+00 1.09958780e+00 1.52599663e-01
-2.51729995e-01 -2.94729024e-01 -6.78728148e-02 -9.84795168e-02
3.41217458e-01 6.77556336e-01 4.67682481e-01 -1.20847106e-01
-1.18330336e+00 -5.94855011e-01 2.64357924e-01 -5.76313809e-02
3.51408958e-01 1.39439985e-01 -1.22299068e-01 -4.63496059e-01
-1.24214150e-01 -3.87540668e-01 9.39258784e-02 3.36661398e-01
1.25177956e+00 -5.11044323e-01 -7.25826979e-01 8.06274861e-02
1.17335808e+00 4.91507500e-01 4.04555827e-01 1.05657406e-01
5.97051859e-01 4.35924172e-01 1.38561666e+00 4.73461211e-01
7.87850171e-02 8.64855707e-01 7.54986644e-01 5.44692039e-01
1.63796753e-01 5.68045750e-02 1.80983528e-01 5.59099913e-01
-9.72046971e-01 -3.27930450e-01 -9.81407344e-01 3.51769477e-01
-1.91656101e+00 -9.13735509e-01 1.07273273e-01 1.72468066e+00
1.06876814e+00 -2.44404115e-02 -3.71501327e-01 -8.90704766e-02
6.89916968e-01 -3.23986232e-01 -8.64725292e-01 -3.04627359e-01
5.67813516e-01 3.40925425e-01 6.96864128e-01 4.84769940e-01
-1.05249703e+00 1.28538394e+00 6.97229576e+00 2.73346245e-01
-1.21864438e+00 -3.17503572e-01 -5.68730116e-01 1.88521326e-01
1.06224768e-01 -8.61532390e-02 -7.95067430e-01 -1.27177350e-02
2.32135579e-01 -1.10308610e-01 8.12077105e-01 1.06126177e+00
-1.87514633e-01 -1.39603242e-01 -1.26834166e+00 5.90653419e-01
-1.32793620e-01 -1.04265487e+00 -5.32232784e-03 -3.38328302e-01
6.78641975e-01 -8.85282382e-02 -3.44826132e-01 4.04067189e-01
7.59071946e-01 -8.11329663e-01 1.10554481e+00 9.83168244e-01
3.85193974e-01 -3.63451183e-01 3.77894521e-01 4.59906936e-01
-1.24763966e+00 -1.81919217e-01 -6.79699719e-01 -2.79407203e-01
7.72489965e-01 5.09548821e-02 -1.13056195e+00 3.58551681e-01
6.20538235e-01 2.90693313e-01 2.77146339e-01 1.04237568e+00
-1.52252600e-01 -1.14230998e-02 -5.60019612e-01 -4.87851024e-01
-1.14137596e-02 -1.96225792e-01 9.50745046e-01 5.39103985e-01
3.80284071e-01 6.50980592e-01 8.86529505e-01 6.81406081e-01
2.47836351e-01 -5.35539925e-01 -4.93438810e-01 -4.08205152e-01
1.12253830e-01 1.04916942e+00 -1.02432632e+00 -1.74887255e-01
4.65266377e-01 8.38675678e-01 -1.26034962e-02 -7.96100423e-02
-9.19781387e-01 -2.90879846e-01 6.79687500e-01 8.64065364e-02
2.57284433e-01 -1.16282260e+00 3.57074767e-01 -9.50033903e-01
-1.97737012e-02 -3.02235276e-01 -1.62656233e-01 -8.34976971e-01
-8.53994012e-01 2.67854750e-01 6.13131940e-01 -1.35141635e+00
-3.91939104e-01 -7.66131341e-01 -3.53187695e-02 5.22629201e-01
-1.25507903e+00 -1.28200066e+00 -4.37854409e-01 3.91776383e-01
9.29778636e-01 1.91581324e-01 8.51587653e-01 -2.58171737e-01
-4.72014211e-02 -1.57360569e-01 -3.55528593e-01 -4.80504781e-02
2.32519627e-01 -8.66646886e-01 3.70736361e-01 3.41264665e-01
-2.83212692e-01 5.13340414e-01 9.90754366e-01 -1.17985094e+00
-2.27310038e+00 -1.11448658e+00 5.55371568e-02 -5.91573536e-01
5.62773407e-01 -9.38412175e-02 -8.29338849e-01 1.22557557e+00
-4.09224212e-01 2.77019054e-01 -3.85847002e-01 -5.15551567e-01
5.13430119e-01 4.82761443e-01 -1.29529047e+00 6.86735272e-01
1.45610702e+00 -2.21278612e-02 -8.19971561e-01 9.44622159e-01
4.75838453e-01 -1.75467479e+00 -1.11577404e+00 7.60093570e-01
6.69957042e-01 2.61141777e-01 1.17764854e+00 -8.44902337e-01
-3.31708007e-02 -2.14756951e-01 -8.34108051e-03 -1.15230620e+00
-4.41317409e-01 -7.36953855e-01 -4.79708135e-01 4.76823628e-01
-1.07970178e-01 -4.57139492e-01 8.44992101e-01 8.86069655e-01
-5.12023687e-01 -8.33516121e-01 -9.15890992e-01 -1.01458430e+00
-4.51800168e-01 4.88179326e-02 5.82587183e-01 6.30304396e-01
-1.30378425e-01 -5.33886671e-01 -3.35637629e-01 7.52623022e-01
6.44906685e-02 5.62350392e-01 8.25811446e-01 -1.16339827e+00
-2.11908221e-01 -3.47177722e-02 -6.69446409e-01 -9.27584171e-01
2.63247937e-01 -7.76086390e-01 7.99222648e-01 -1.67210150e+00
-8.18094313e-02 -9.72542286e-01 3.75020146e-01 8.05757403e-01
1.98929608e-01 -5.75963669e-02 2.41729364e-01 3.79145563e-01
-4.56234634e-01 3.25657517e-01 2.06108618e+00 -2.62281984e-01
-5.46332598e-01 4.95671362e-01 3.71900618e-01 9.22955215e-01
9.25894976e-01 -3.38636965e-01 -2.00720817e-01 -6.17698371e-01
1.92969963e-01 4.17797178e-01 3.79583240e-01 -8.58928919e-01
9.71547663e-02 -1.11987042e+00 2.15468943e-01 -8.10044706e-01
7.17614055e-01 -1.20549333e+00 6.46964073e-01 9.28915441e-01
-1.02316104e-01 6.88787401e-02 2.12360188e-01 5.86596787e-01
2.43381947e-01 -4.51841235e-01 6.33612990e-01 -3.11342925e-01
-8.83831561e-01 4.29958344e-01 -4.40151155e-01 -4.98889267e-01
1.55122936e+00 -6.91832229e-02 1.77853763e-01 2.67031521e-01
-1.09776258e+00 1.05447285e-01 4.36276019e-01 9.20587540e-01
8.53424191e-01 -9.55627263e-01 -2.93707073e-01 5.69331795e-02
-4.11080509e-01 6.77048922e-01 5.37127145e-02 2.63804793e-01
-9.93576229e-01 8.20127875e-03 -5.37060797e-01 -5.04748583e-01
-1.29302561e+00 3.14249724e-01 2.31434420e-01 -1.11510336e-01
-8.55273545e-01 9.35198724e-01 -7.32788086e-01 -6.23832107e-01
-9.71169993e-02 -1.10082686e+00 -2.42095590e-02 -5.76205432e-01
-1.08980477e-01 5.97624838e-01 1.31720770e-02 -4.36187238e-01
-2.35466346e-01 8.20634544e-01 2.29196578e-01 2.98953772e-01
1.59424818e+00 2.84789801e-01 -2.69176453e-01 1.81182936e-01
4.96190548e-01 -4.90678161e-01 -1.35545659e+00 1.36121616e-01
-7.30469972e-02 -6.48282826e-01 -3.08898926e-01 -5.34502387e-01
-7.82052398e-01 1.55716807e-01 2.58683592e-01 -1.86130375e-01
4.03913110e-01 3.29718798e-01 7.52075195e-01 1.21255982e+00
1.23814905e+00 -1.10854840e+00 2.78353691e-01 7.92605519e-01
1.82332146e+00 -8.07774246e-01 5.29394746e-01 -8.46811473e-01
-6.32548571e-01 1.39750695e+00 6.63167298e-01 -6.95199251e-01
7.35928535e-01 6.39054239e-01 -2.57924020e-01 -3.76533449e-01
-2.46025041e-01 2.00582847e-01 5.78102589e-01 4.95074570e-01
-3.15237463e-01 4.08325583e-01 -9.06675383e-02 2.40733400e-01
-2.75258660e-01 2.93112487e-01 2.71197736e-01 1.79045844e+00
-7.84886420e-01 -9.49746490e-01 -4.81968969e-01 3.23153019e-01
-6.29714802e-02 8.32976580e-01 -1.63938642e-01 9.76937413e-01
9.56758019e-03 3.13675314e-01 -2.73108240e-02 -4.71393839e-02
7.20764756e-01 -1.12847663e-01 1.23371089e+00 -9.96570349e-01
-4.04674530e-01 -2.93830812e-01 2.14302130e-02 -7.75618911e-01
-4.82021242e-01 -7.01860726e-01 -1.90105486e+00 5.85604087e-02
-7.56139457e-01 -4.23259676e-01 9.97602582e-01 9.35522199e-01
1.64320469e-01 4.32477891e-01 2.90957153e-01 -2.00465274e+00
-1.15767539e+00 -9.22086656e-01 -3.75743121e-01 1.64591089e-01
1.68531433e-01 -1.27841389e+00 9.29203257e-02 3.17963362e-01] | [5.021374225616455, 0.18460458517074585] |
acc4f64b-850a-4995-a979-6da51ef049e5 | unsupervised-learning-of-landmarks-by | 1908.06427 | null | https://arxiv.org/abs/1908.06427v1 | https://arxiv.org/pdf/1908.06427v1.pdf | Unsupervised Learning of Landmarks by Descriptor Vector Exchange | Equivariance to random image transformations is an effective method to learn landmarks of object categories, such as the eyes and the nose in faces, without manual supervision. However, this method does not explicitly guarantee that the learned landmarks are consistent with changes between different instances of the same object, such as different facial identities. In this paper, we develop a new perspective on the equivariance approach by noting that dense landmark detectors can be interpreted as local image descriptors equipped with invariance to intra-category variations. We then propose a direct method to enforce such an invariance in the standard equivariant loss. We do so by exchanging descriptor vectors between images of different object instances prior to matching them geometrically. In this manner, the same vectors must work regardless of the specific object identity considered. We use this approach to learn vectors that can simultaneously be interpreted as local descriptors and dense landmarks, combining the advantages of both. Experiments on standard benchmarks show that this approach can match, and in some cases surpass state-of-the-art performance amongst existing methods that learn landmarks without supervision. Code is available at www.robots.ox.ac.uk/~vgg/research/DVE/. | ['Hakan Bilen', 'Andrea Vedaldi', 'Samuel Albanie', 'James Thewlis'] | 2019-08-18 | unsupervised-learning-of-landmarks-by-1 | http://openaccess.thecvf.com/content_ICCV_2019/html/Thewlis_Unsupervised_Learning_of_Landmarks_by_Descriptor_Vector_Exchange_ICCV_2019_paper.html | http://openaccess.thecvf.com/content_ICCV_2019/papers/Thewlis_Unsupervised_Learning_of_Landmarks_by_Descriptor_Vector_Exchange_ICCV_2019_paper.pdf | iccv-2019-10 | ['unsupervised-facial-landmark-detection'] | ['computer-vision'] | [ 5.44876941e-02 7.80264363e-02 -1.22037388e-01 -4.94380385e-01
-3.98620993e-01 -8.17006171e-01 8.30165327e-01 -7.31218457e-02
-5.06616414e-01 4.12701488e-01 -1.24147478e-02 3.54063928e-01
-3.13639849e-01 -6.12116456e-01 -9.15293932e-01 -7.67179489e-01
1.54329106e-01 4.90177602e-01 1.60960168e-01 -2.43761614e-01
2.97744989e-01 7.86305010e-01 -1.63133180e+00 -5.75255305e-02
2.83248127e-01 8.23407590e-01 -1.04413040e-01 2.65440077e-01
4.54493575e-02 4.00022209e-01 -5.27781211e-02 -5.11749327e-01
6.78625882e-01 -2.48972163e-01 -8.27983499e-01 3.28639299e-01
1.15215802e+00 -5.36775962e-02 -2.20022723e-01 1.37152445e+00
3.29568118e-01 1.16737798e-01 6.52389765e-01 -1.35430503e+00
-8.35216820e-01 1.66906089e-01 -5.99550784e-01 -2.21088752e-02
4.47330177e-01 -3.10313851e-01 1.15437508e+00 -1.18648994e+00
8.86350274e-01 1.34228373e+00 7.17880368e-01 5.85536897e-01
-1.50597143e+00 -5.75073361e-01 1.74647674e-01 2.49781415e-01
-1.66000497e+00 -7.55837440e-01 9.47668910e-01 -4.06181902e-01
4.15453315e-01 1.68182403e-01 4.36704457e-01 8.09036136e-01
1.50955975e-01 4.71889526e-01 1.07212949e+00 -4.38672870e-01
1.44363835e-01 3.18602115e-01 -1.95690431e-02 9.43621457e-01
2.20120534e-01 9.16886926e-02 -4.34158713e-01 8.39690045e-02
8.89363408e-01 2.32295856e-01 -1.50060862e-01 -1.12605643e+00
-1.26688492e+00 9.31231737e-01 7.00848520e-01 3.12885821e-01
-1.46165982e-01 1.50394052e-01 3.10464591e-01 4.05652940e-01
1.76424995e-01 3.90547931e-01 -2.82891184e-01 3.27181578e-01
-4.66791064e-01 2.50353187e-01 6.65606558e-01 1.15601766e+00
1.18226671e+00 -1.26778275e-01 2.27321446e-01 7.06732988e-01
2.69776940e-01 4.49624658e-01 3.32506269e-01 -1.14278281e+00
6.13251552e-02 3.58330339e-01 1.00060552e-02 -1.20988727e+00
-2.68887460e-01 -1.24643408e-01 -7.14967608e-01 4.59405094e-01
4.75041628e-01 2.51844466e-01 -7.95406938e-01 1.95049143e+00
5.04540980e-01 4.19072807e-01 -1.42141124e-02 7.39276528e-01
5.14022052e-01 1.97724253e-01 -1.39518574e-01 1.04538448e-01
1.31135595e+00 -7.46759832e-01 -5.73721111e-01 -6.49209023e-02
2.82204121e-01 -8.33804071e-01 8.73399734e-01 1.07103601e-01
-1.01536345e+00 -5.84293485e-01 -1.05279911e+00 -1.69524506e-01
-5.48214436e-01 -1.00408278e-01 6.52106404e-01 5.43231726e-01
-1.26440048e+00 5.69562793e-01 -9.13518846e-01 -6.26873553e-01
2.51619041e-01 6.26785278e-01 -8.65715504e-01 -5.50868176e-02
-7.83778191e-01 9.19455528e-01 1.75897226e-01 3.20873708e-02
-7.75260329e-01 -7.15525389e-01 -1.19641471e+00 -2.56026983e-01
2.19184518e-01 -4.93287325e-01 1.03338265e+00 -1.25875628e+00
-1.44502246e+00 1.26022005e+00 -2.35258594e-01 -1.73840061e-01
6.15886271e-01 -8.07531178e-02 -1.71759337e-01 6.26237541e-02
1.96319908e-01 8.70302975e-01 1.09630704e+00 -1.26146567e+00
-6.11288249e-01 -6.50358975e-01 3.55148137e-01 1.61329836e-01
-1.55884564e-01 5.23694307e-02 -3.82663667e-01 -4.11747575e-01
3.24200869e-01 -1.21412957e+00 -5.92432581e-02 5.52074671e-01
-2.22485140e-01 -2.42332309e-01 8.76552522e-01 -2.09431559e-01
4.21949953e-01 -2.18241286e+00 3.68020087e-01 4.52857137e-01
1.23940691e-01 8.92176107e-03 -2.99734354e-01 1.28329888e-01
-2.44515046e-01 -1.60319418e-01 -2.29949027e-01 -4.53403056e-01
1.16157308e-01 4.51100439e-01 -2.23655656e-01 1.11814225e+00
3.17178547e-01 7.61539698e-01 -9.00099337e-01 -4.44013953e-01
3.49037617e-01 7.35879183e-01 -5.64954638e-01 -2.31164042e-02
2.91327208e-01 5.74452102e-01 -2.98906982e-01 5.37806451e-01
7.66935945e-01 1.12990022e-01 2.23528687e-02 -2.48542324e-01
-9.89277288e-02 4.96234745e-02 -1.53381491e+00 1.74135900e+00
-5.43448210e-01 5.51970005e-01 2.24761575e-01 -1.31146252e+00
8.24167609e-01 1.24941483e-01 4.67625499e-01 -5.13528049e-01
9.22524557e-02 1.75499648e-01 -1.54374942e-01 -3.29583324e-02
2.64711946e-01 -1.56040192e-01 3.55570018e-02 1.96596503e-01
3.47462565e-01 -1.94463357e-01 6.81762919e-02 -9.75936279e-02
6.37181401e-01 1.26640648e-01 4.33803022e-01 -5.11041224e-01
7.63578534e-01 -4.71948862e-01 5.90777874e-01 5.48082829e-01
-1.50858864e-01 6.43499196e-01 1.28194198e-01 -4.88138169e-01
-1.07594228e+00 -1.21364808e+00 -5.77948213e-01 9.48905766e-01
3.83243412e-01 -2.17277378e-01 -6.56566858e-01 -7.68971741e-01
1.31535068e-01 2.17545792e-01 -8.03316772e-01 -1.71467602e-01
-5.63405514e-01 -1.49123594e-01 2.87797958e-01 4.29880321e-01
3.87917966e-01 -8.54575634e-01 -4.25154239e-01 -3.76309268e-03
2.85734981e-01 -1.16975379e+00 -5.71622431e-01 7.29518682e-02
-6.07204735e-01 -1.13555336e+00 -6.01950705e-01 -9.73669946e-01
1.08861995e+00 3.89968932e-01 9.44210947e-01 1.24437079e-01
-3.72238249e-01 9.02179062e-01 -1.46497995e-01 -2.29304373e-01
-2.74931908e-01 -6.00324906e-02 4.53077018e-01 3.76324475e-01
2.92691916e-01 -6.21504605e-01 -3.85586619e-01 6.14881217e-01
-9.46020663e-01 -4.46114033e-01 3.39742094e-01 9.46244299e-01
8.83708715e-01 -2.34327883e-01 8.57622549e-02 -9.12164152e-01
4.45012972e-02 -2.68659294e-01 -8.67286861e-01 3.19839180e-01
-5.30102372e-01 2.85256773e-01 5.07894576e-01 -3.91009569e-01
-6.06541991e-01 5.62817216e-01 -3.07835918e-02 -5.79724252e-01
-3.19298059e-01 1.37146592e-01 -4.07175332e-01 -6.40756071e-01
4.35466021e-01 5.08825146e-02 3.75263505e-02 -3.38080198e-01
5.86721778e-01 3.33603024e-01 6.32556617e-01 -5.69854856e-01
1.21953940e+00 9.12761092e-01 3.28662068e-01 -8.19864631e-01
-6.56644285e-01 -6.80713534e-01 -1.25888574e+00 6.51583448e-02
7.87838697e-01 -7.26346195e-01 -4.88118201e-01 3.23759139e-01
-1.01206136e+00 1.57186482e-02 -3.85878205e-01 4.16880876e-01
-8.39418948e-01 4.14371252e-01 -1.77737772e-01 -2.82418072e-01
8.70288685e-02 -1.13209569e+00 1.17394865e+00 2.39155546e-01
-4.05794159e-02 -1.32986224e+00 2.63773743e-02 -1.38523698e-01
2.76137531e-01 2.58297145e-01 5.90376079e-01 -6.00963295e-01
-5.86109638e-01 -2.63383150e-01 -1.60331447e-02 3.49986285e-01
4.79913354e-01 1.08827017e-01 -9.28262293e-01 -4.87444133e-01
-1.40328165e-02 -1.06044404e-01 6.22653067e-01 8.45362619e-02
9.13706303e-01 -3.45431596e-01 -2.26931110e-01 8.46750975e-01
1.47221792e+00 -2.12949902e-01 5.63032389e-01 4.04641867e-01
8.65467489e-01 6.94932461e-01 5.69389522e-01 3.10523093e-01
2.23042756e-01 1.13171387e+00 4.57732975e-01 -2.23233581e-01
-6.69268146e-02 -2.36414030e-01 4.65142399e-01 5.22398591e-01
-1.72586516e-01 3.45482886e-01 -6.48863971e-01 5.78511536e-01
-1.61079752e+00 -9.71129000e-01 1.12079203e-01 2.27622628e+00
7.10276484e-01 -2.35074818e-01 -4.02075425e-02 4.48012352e-02
8.10783207e-01 1.46279171e-01 -4.93771583e-01 -3.87979329e-01
-3.29074333e-03 2.91249722e-01 6.81990027e-01 6.86371863e-01
-1.39561737e+00 1.14609408e+00 5.78671360e+00 4.29788977e-01
-1.18607020e+00 1.24584071e-01 1.45490810e-01 3.17089021e-01
-2.30949029e-01 1.28896311e-01 -9.70929563e-01 1.40712321e-01
5.10313630e-01 -2.70013750e-01 5.44447064e-01 9.82313275e-01
-1.57985777e-01 4.18969154e-01 -1.51560831e+00 1.04362643e+00
3.82223278e-01 -1.12831235e+00 7.69220898e-03 1.09921396e-01
7.65006959e-01 -2.06564769e-01 3.44459474e-01 -6.27446100e-02
1.59943327e-01 -9.11002815e-01 8.62557292e-01 4.95983154e-01
7.47286618e-01 -6.49866581e-01 5.59605122e-01 -8.35391283e-02
-1.36007178e+00 2.91359246e-01 -7.45903969e-01 1.77310899e-01
-1.78962499e-01 1.32984996e-01 -6.89615309e-01 4.64358568e-01
6.16744637e-01 1.08613825e+00 -6.30845547e-01 9.20837939e-01
-4.08491582e-01 5.78988455e-02 -3.55945557e-01 4.18607622e-01
1.51537120e-01 -3.66446286e-01 5.75337768e-01 1.13031340e+00
2.22936258e-01 -2.11576000e-01 1.85717180e-01 9.06433761e-01
-8.39879662e-02 1.04645431e-01 -1.01169562e+00 5.21537185e-01
4.46506739e-01 1.25950515e+00 -6.18825495e-01 -7.97082633e-02
-6.75294876e-01 1.11509538e+00 4.17363107e-01 3.07003140e-01
-7.66859412e-01 -3.28734636e-01 1.05193841e+00 1.50909845e-03
4.75919962e-01 -3.10474932e-01 1.55664325e-01 -1.15647709e+00
1.21094041e-01 -7.19707608e-01 2.96036690e-01 -4.19220418e-01
-1.38805795e+00 3.66275907e-01 6.37662560e-02 -1.33652711e+00
-2.43165091e-01 -8.97003293e-01 -5.17858803e-01 6.01773083e-01
-1.71178246e+00 -1.31340837e+00 -3.43221992e-01 1.04220748e+00
3.72007817e-01 -9.17797443e-03 8.37638795e-01 2.32654437e-01
-1.91858411e-01 9.00339663e-01 1.79279253e-01 2.96975464e-01
8.94876182e-01 -1.39501476e+00 2.36885518e-01 7.08278120e-01
6.70381725e-01 7.88292289e-01 4.88643169e-01 -1.70382738e-01
-1.62191308e+00 -1.24415791e+00 7.67422736e-01 -6.61054194e-01
7.67189622e-01 -5.01905322e-01 -9.13702488e-01 1.09047496e+00
1.84066240e-02 5.26908457e-01 4.11475748e-01 3.76195051e-02
-6.53100967e-01 -3.78789812e-01 -1.10886323e+00 5.27916193e-01
1.13778532e+00 -7.49419153e-01 -6.28950715e-01 4.62168217e-01
3.94647866e-01 -4.42092329e-01 -8.91635537e-01 2.03126028e-01
6.71905637e-01 -8.65055263e-01 1.22687638e+00 -6.10412598e-01
-3.31883579e-02 -5.31998515e-01 -3.63828450e-01 -1.16135311e+00
-3.40369076e-01 -5.09736180e-01 4.17292118e-01 1.35214639e+00
1.41864195e-01 -8.89052987e-01 4.23215240e-01 5.75802565e-01
-4.20467295e-02 -2.87682235e-01 -1.21210515e+00 -1.12525415e+00
1.57748476e-01 -1.22103892e-01 5.61674953e-01 1.18231475e+00
-8.58022943e-02 1.13319650e-01 -3.24608147e-01 4.49831367e-01
7.42630661e-01 -6.79047033e-02 9.12392259e-01 -1.25025094e+00
5.91242611e-02 -4.47490364e-01 -1.25379920e+00 -7.79065132e-01
8.12858403e-01 -1.13214278e+00 4.88154553e-02 -9.43222165e-01
1.05051205e-01 -5.43203413e-01 -4.08852786e-01 6.17709398e-01
1.00367546e-01 4.79254454e-01 2.04423606e-01 2.47097850e-01
-5.40335059e-01 5.50741374e-01 9.55836594e-01 -2.63651609e-01
3.20340507e-02 -4.27074656e-02 -5.50267458e-01 8.67088616e-01
7.17501342e-01 -5.23202658e-01 -3.18619162e-01 -3.52018535e-01
-1.64357901e-01 -5.69114208e-01 6.10676169e-01 -9.69135940e-01
2.80681968e-01 -1.10615283e-01 3.06272298e-01 1.37449220e-01
2.43578002e-01 -1.16003442e+00 1.96502402e-01 3.96091163e-01
-4.04475182e-01 2.40251139e-01 1.13399826e-01 5.73130786e-01
-2.84143388e-01 -3.09044838e-01 1.11562157e+00 -9.83546600e-02
-8.85292232e-01 4.84117329e-01 1.35674044e-01 -1.01729795e-01
1.11995327e+00 -2.38737807e-01 2.95115635e-02 -2.65690356e-01
-6.27586305e-01 -1.63668592e-03 8.33759189e-01 6.24869168e-01
5.47777236e-01 -1.54783869e+00 -6.86025023e-01 4.78594482e-01
2.70995975e-01 -1.60447776e-01 -2.25833833e-01 9.18459713e-01
-3.50028276e-01 3.44092518e-01 -4.85027671e-01 -8.55368912e-01
-1.47924352e+00 7.39321768e-01 6.21152401e-01 1.54332027e-01
-6.73514605e-01 6.97414458e-01 5.19864500e-01 -5.93482018e-01
2.00938791e-01 -2.66556561e-01 -1.64406791e-01 2.66061295e-02
4.43788677e-01 1.70008466e-01 1.68101937e-01 -1.26412272e+00
-5.28838396e-01 1.32353580e+00 -1.44931152e-01 -3.10795475e-02
1.16526496e+00 -2.52783984e-01 -1.98404089e-01 3.63300771e-01
1.59059644e+00 2.36542195e-01 -1.30416763e+00 -4.70981240e-01
1.07111633e-01 -7.76326537e-01 -1.79773271e-01 -2.35699251e-01
-1.30405438e+00 8.07324469e-01 8.25228512e-01 -1.49277344e-01
9.52342451e-01 2.53175139e-01 2.48818278e-01 4.87605304e-01
6.02180541e-01 -8.16155195e-01 -3.83106507e-02 4.05404299e-01
1.16293967e+00 -1.32364690e+00 8.61748960e-03 -4.29314673e-01
-4.67431217e-01 1.12010193e+00 3.89438689e-01 -6.13345683e-01
7.45386243e-01 -9.89091992e-02 1.78236261e-01 -1.72594905e-01
-2.39919230e-01 -2.80008107e-01 4.93012249e-01 7.53395021e-01
2.45617583e-01 1.30674802e-02 -1.79929614e-01 -8.01629797e-02
-2.24515051e-01 -4.33316141e-01 3.69006187e-01 9.05539811e-01
-1.51494130e-01 -1.26262116e+00 -3.70416850e-01 -9.91375968e-02
-3.29471827e-01 1.42334983e-01 -4.53229308e-01 1.05589318e+00
1.67862087e-01 6.10154748e-01 1.86768457e-01 -1.93616405e-01
5.27379274e-01 -3.14931199e-02 6.92404747e-01 -6.62753880e-01
-1.02508530e-01 -5.11684008e-02 -4.36913341e-01 -8.54495287e-01
-7.30338991e-01 -1.00424337e+00 -1.11837280e+00 -1.84147060e-01
-1.29676700e-01 1.56977028e-02 7.19916701e-01 6.98231161e-01
1.72624066e-01 -3.13354935e-03 8.04090798e-01 -1.11429012e+00
-6.41424119e-01 -5.75379133e-01 -6.93826139e-01 7.16289341e-01
5.82475543e-01 -1.01736844e+00 -5.28082371e-01 2.49947995e-01] | [8.115537643432617, -1.9960315227508545] |
5049ad36-f2a5-485f-be40-b8ea788651b5 | hulmona-the-universal-language-model-in | null | null | https://aclanthology.org/W19-4608 | https://aclanthology.org/W19-4608.pdf | hULMonA: The Universal Language Model in Arabic | Arabic is a complex language with limited resources which makes it challenging to produce accurate text classification tasks such as sentiment analysis. The utilization of transfer learning (TL) has recently shown promising results for advancing accuracy of text classification in English. TL models are pre-trained on large corpora, and then fine-tuned on task-specific datasets. In particular, universal language models (ULMs), such as recently developed BERT, have achieved state-of-the-art results in various NLP tasks in English. In this paper, we hypothesize that similar success can be achieved for Arabic. The work aims at supporting the hypothesis by developing the first Universal Language Model in Arabic (hULMonA - حلمنا meaning our dream), demonstrating its use for Arabic classifications tasks, and demonstrating how a pre-trained multi-lingual BERT can also be used for Arabic. We then conduct a benchmark study to evaluate both ULM successes with Arabic sentiment analysis. Experiment results show that the developed hULMonA and multi-lingual ULM are able to generalize well to multiple Arabic data sets and achieve new state of the art results in Arabic Sentiment Analysis for some of the tested sets. | ['Wassim El-Hajj', 'Obeida ElJundi', 'Nour El Droubi', 'Hazem Hajj', 'Wissam Antoun', 'Khaled Shaban'] | 2019-08-01 | null | null | null | ws-2019-8 | ['arabic-sentiment-analysis'] | ['natural-language-processing'] | [-3.50407451e-01 -2.32690394e-01 -1.67746976e-01 -5.37202954e-01
-8.74456644e-01 -5.92921376e-01 7.54678547e-01 3.57322007e-01
-5.88340580e-01 8.08577478e-01 -7.17196465e-02 -4.00651276e-01
1.65994972e-01 -7.34006941e-01 -4.04516816e-01 -4.15196478e-01
-1.04067750e-01 9.16836560e-01 1.34862691e-01 -1.14190185e+00
3.48316967e-01 3.18470448e-01 -1.33822668e+00 9.31856453e-01
9.60842192e-01 8.40898931e-01 5.98142557e-02 4.51519459e-01
-2.38404438e-01 8.77725840e-01 -7.44556129e-01 -7.68866360e-01
-1.91426113e-01 -2.88398474e-01 -1.04026890e+00 6.23711459e-02
2.17027944e-02 3.00305933e-02 6.92601740e-01 5.73180914e-01
3.51750761e-01 -7.20860437e-02 9.25472677e-01 -1.20011866e+00
-9.28348243e-01 7.75152326e-01 -4.37241048e-01 -2.92984694e-01
5.75595856e-01 -5.24812102e-01 9.63296056e-01 -1.17352104e+00
4.90069032e-01 1.29421389e+00 9.33406234e-01 4.54843014e-01
-5.77285826e-01 -5.53552866e-01 -1.22576155e-01 1.83400303e-01
-1.09806526e+00 -1.39686272e-01 4.56051528e-01 -9.58524644e-02
1.10854411e+00 1.03102764e-02 1.94733933e-01 8.64419758e-01
6.31966829e-01 1.09783208e+00 1.57890344e+00 -1.11721861e+00
9.79603361e-03 6.87332988e-01 9.39213336e-02 7.38893926e-01
-9.72613040e-03 -7.61203587e-01 -7.65186071e-01 3.46958369e-01
-5.53285331e-02 -4.76215541e-01 1.58163577e-01 2.87162483e-01
-1.01620817e+00 1.29891670e+00 4.01531845e-01 8.01087022e-01
-9.03141946e-02 -4.43219036e-01 7.98058927e-01 8.12697113e-01
9.08325970e-01 4.78163958e-01 -1.04058778e+00 -1.92611411e-01
-8.88789654e-01 -8.36851522e-02 1.06299174e+00 8.43366027e-01
6.03806853e-01 1.10573731e-02 3.55721891e-01 1.15695059e+00
5.45770645e-01 6.67707741e-01 1.29735374e+00 -4.25697193e-02
4.90965933e-01 8.62984359e-01 -7.45609775e-02 -9.55040812e-01
-8.22689891e-01 -7.99796581e-02 -5.48631668e-01 5.14302030e-02
5.20593405e-01 -3.27380836e-01 -7.21603096e-01 1.13153243e+00
-1.39678810e-02 -6.71821773e-01 7.22909689e-01 3.66099298e-01
7.49849737e-01 1.12114275e+00 1.36165693e-01 -1.09970160e-01
1.39701784e+00 -1.11317790e+00 -6.55937970e-01 -5.08616328e-01
1.35396111e+00 -1.12432623e+00 1.46228957e+00 7.48781502e-01
-7.16390371e-01 -5.49000859e-01 -1.15256369e+00 2.17970535e-02
-1.19305277e+00 5.17919421e-01 5.33267736e-01 1.28316927e+00
-9.85038698e-01 2.28375018e-01 -7.44913280e-01 -8.19667518e-01
1.11022979e-01 5.81064463e-01 -5.66736877e-01 -3.32442522e-01
-1.36139095e+00 1.44113386e+00 6.21475756e-01 7.78689757e-02
-4.04891700e-01 -3.75343091e-03 -9.91889477e-01 -3.55413139e-01
-5.19322082e-02 1.52654484e-01 1.06048608e+00 -1.49492705e+00
-1.64235044e+00 1.14986479e+00 4.62950207e-02 -4.84130055e-01
1.52531555e-02 -2.33127922e-01 -8.18278670e-01 -2.27881782e-02
2.13964939e-01 5.51221550e-01 6.67936802e-01 -1.33425033e+00
-8.14738810e-01 -5.72139800e-01 -3.83212306e-02 1.67469591e-01
-1.15019321e+00 4.62515980e-01 -1.06087513e-01 -6.18309915e-01
-8.16088542e-02 -1.18739426e+00 2.06886709e-01 -7.23022878e-01
2.33859003e-01 -4.92614657e-01 9.13294733e-01 -8.04265380e-01
1.26087332e+00 -1.61975336e+00 1.54305901e-02 1.19152404e-01
-6.52104199e-01 3.83131474e-01 -2.96340644e-01 7.52336740e-01
2.72084147e-01 4.79212403e-02 -1.80413246e-01 -4.02828187e-01
5.93517534e-02 4.93014842e-01 -9.18946117e-02 2.60919333e-01
4.50759381e-01 8.80145252e-01 -6.89831436e-01 -3.84235471e-01
6.21726550e-03 3.41897905e-01 -3.67707610e-01 -7.98313692e-02
-2.32505858e-01 1.74260378e-01 -2.04214126e-01 8.37483466e-01
3.55214566e-01 -7.07146898e-02 2.90940881e-01 -4.69049737e-02
-1.49667263e-01 1.20332964e-01 -7.90965915e-01 1.41673577e+00
-9.95933652e-01 5.79394817e-01 -2.52017587e-01 -9.79620516e-01
1.10623395e+00 4.23440814e-01 1.37366503e-01 -5.32867968e-01
5.30164003e-01 8.33150208e-01 1.24855697e-01 -4.54286963e-01
6.35310411e-01 -4.17072713e-01 -4.19114292e-01 6.24869049e-01
4.51319575e-01 -6.87552869e-01 5.74121594e-01 -1.65860504e-01
3.18209052e-01 2.83115000e-01 3.52329820e-01 -5.95873058e-01
1.30906916e+00 3.42513204e-01 -3.10274959e-01 1.06197290e-01
-1.10897057e-01 3.39145422e-01 3.26259404e-01 -4.18524534e-01
-6.65312529e-01 -2.75042653e-01 -3.10792416e-01 1.79127407e+00
-3.47405165e-01 -6.18601382e-01 -8.63400996e-01 -8.89997482e-01
-3.22674364e-01 8.52779090e-01 -7.30571687e-01 4.06002365e-02
-4.02937084e-01 -1.37288845e+00 6.11525595e-01 3.34368587e-01
4.32832032e-01 -1.46335816e+00 -3.15409452e-01 3.25119138e-01
-1.43482104e-01 -1.18649018e+00 1.68604702e-01 3.42688173e-01
-6.65060103e-01 -7.67177224e-01 -6.21602893e-01 -1.32530093e+00
3.80162328e-01 -2.39910200e-01 1.14866185e+00 2.72657834e-02
2.65079945e-01 3.64428878e-01 -1.16648436e+00 -1.03805137e+00
-9.83270645e-01 6.25225961e-01 1.58477664e-01 1.08060822e-01
7.88273156e-01 2.95048356e-01 2.62843996e-01 4.07248110e-01
-1.17391253e+00 -3.88019800e-01 4.82532114e-01 8.59280705e-01
-1.99803617e-02 1.58162112e-03 1.20859313e+00 -9.91790891e-01
9.39514935e-01 -4.83005315e-01 1.60328101e-03 3.41248542e-01
-7.35587835e-01 -1.11160852e-01 8.60952437e-01 -1.81651443e-01
-1.01766181e+00 -3.12783957e-01 -6.09432638e-01 6.34132504e-01
-2.14214791e-02 1.25533032e+00 8.26476887e-02 -1.11241333e-01
8.87570739e-01 6.01969659e-02 2.34944776e-01 -1.55038491e-01
2.14920849e-01 1.26573145e+00 -2.39766672e-01 -7.13907599e-01
1.60291046e-01 3.13891798e-01 -1.86821178e-01 -1.10246933e+00
-1.10205603e+00 -3.46440732e-01 -1.15888572e+00 -1.85446039e-01
8.95590901e-01 -1.01041186e+00 -1.87169760e-01 8.02194953e-01
-6.94068551e-01 -5.57595909e-01 1.77216947e-01 2.06555888e-01
-4.25470591e-01 1.77387774e-01 -9.34753895e-01 -6.82862699e-01
-4.76406157e-01 -1.19053924e+00 1.18184972e+00 2.35378463e-02
-3.80768210e-01 -1.64016008e+00 1.37938321e-01 4.70594734e-01
4.37361360e-01 -1.41803369e-01 9.14804280e-01 -9.68327582e-01
5.71415544e-01 -5.34783542e-01 5.10563292e-02 7.22004890e-01
2.77007043e-01 1.93479452e-02 -9.87911999e-01 -4.77643281e-01
-3.34356092e-02 -1.08980262e+00 4.03046876e-01 -4.51477692e-02
3.98452044e-01 1.25177681e-01 1.08842500e-01 -1.78220108e-01
1.31293011e+00 8.61473083e-02 5.85029483e-01 9.87510145e-01
3.31830680e-01 8.62593770e-01 1.09507108e+00 1.69019207e-01
6.98648095e-01 3.67516905e-01 1.79581434e-01 8.74798968e-02
1.68958634e-01 2.67511040e-01 1.02132487e+00 1.44890368e+00
-6.33403063e-02 -3.60154778e-01 -1.27323425e+00 4.91856426e-01
-1.71852744e+00 -2.98871875e-01 -3.01067740e-01 1.83988893e+00
1.07898819e+00 2.03298226e-01 1.32364437e-01 4.54836130e-01
6.50204122e-02 -2.85327464e-01 2.17774242e-01 -9.79964256e-01
-4.36165541e-01 6.19803131e-01 -2.85303928e-02 5.92843175e-01
-1.34373832e+00 1.57957542e+00 5.79829454e+00 9.23912942e-01
-1.19601762e+00 3.12865347e-01 6.76956892e-01 5.12480676e-01
1.63895115e-01 -4.43756610e-01 -9.48279679e-01 8.71475115e-02
1.58553898e+00 1.15427673e-01 -4.66650985e-02 6.44105852e-01
-2.22808830e-02 -6.87102303e-02 -6.56296372e-01 5.41173398e-01
7.49429882e-01 -8.53123009e-01 4.45222169e-01 -4.20761019e-01
9.22680557e-01 1.39699548e-01 7.17411786e-02 8.16838682e-01
1.49632305e-01 -1.17855799e+00 6.48801446e-01 -1.13536328e-01
7.03583956e-01 -1.05953348e+00 1.38919234e+00 3.32376480e-01
-8.00520122e-01 -1.44877270e-01 -5.59901059e-01 -1.78637981e-01
-1.65928513e-01 4.19394486e-02 -1.06455922e+00 7.22977638e-01
8.41549158e-01 1.01592290e+00 -1.06093657e+00 1.85357064e-01
-2.20969066e-01 7.36944199e-01 -1.76359922e-01 -5.34112751e-01
8.28176022e-01 -4.19212371e-01 -1.41982093e-01 1.66865909e+00
4.15957808e-01 -3.45381916e-01 2.01287001e-01 -1.50779620e-01
2.00829104e-01 1.12861586e+00 -4.29423153e-01 -1.24549493e-01
-2.89479196e-01 1.47005665e+00 -1.04606450e+00 -2.51112461e-01
-6.27799451e-01 1.04736733e+00 2.84107238e-01 -1.58324033e-01
-5.29312193e-01 -3.33106518e-01 -6.15653247e-02 -2.33919993e-01
7.49553442e-02 -2.46807024e-01 -2.63825804e-01 -1.15285635e+00
-3.80644381e-01 -1.41197956e+00 6.25853598e-01 -9.93774176e-01
-1.56053233e+00 1.23027205e+00 -3.54787946e-01 -1.30128014e+00
-4.21110451e-01 -1.60422444e+00 -4.64973338e-02 7.45421648e-01
-1.70586300e+00 -2.07687569e+00 5.80503568e-02 8.49053025e-01
8.87529433e-01 -8.05594623e-01 1.40608048e+00 2.70371199e-01
-3.14708114e-01 6.42101407e-01 2.36576274e-01 1.43992946e-01
1.32032073e+00 -1.33320737e+00 -2.75592178e-01 5.68792582e-01
3.25256735e-01 3.72140676e-01 4.71807718e-01 -3.53484392e-01
-1.27840304e+00 -8.85673344e-01 1.16814864e+00 -9.39069927e-01
1.13957024e+00 -2.84779638e-01 -8.44956338e-01 8.38315785e-01
5.64289808e-01 -5.25316358e-01 1.15346789e+00 3.03852499e-01
-3.36316049e-01 -6.85413405e-02 -1.13141561e+00 5.51151037e-01
-8.79565477e-02 -4.05475438e-01 -6.19026244e-01 7.10126340e-01
2.87951291e-01 -1.54913396e-01 -1.00089931e+00 3.14776689e-01
6.34291351e-01 -8.44615221e-01 5.54995894e-01 -8.99044991e-01
8.28238964e-01 4.24100608e-02 -4.27141368e-01 -1.67090833e+00
3.12304318e-01 -1.78370059e-01 3.03865284e-01 1.23942947e+00
1.00755668e+00 -7.02527225e-01 4.04432654e-01 1.30135804e-01
-1.90380827e-01 -6.04956329e-01 -5.67644835e-01 -4.90889579e-01
8.60977471e-01 -6.18129909e-01 1.49849653e-01 1.21499443e+00
5.26470959e-01 7.92715669e-01 -1.21264867e-01 -1.01247631e-01
-2.09827945e-01 -8.98437947e-03 6.91616833e-01 -1.27431381e+00
1.92207947e-01 -4.09164369e-01 -2.30870962e-01 -4.30212498e-01
5.44080138e-01 -1.14162564e+00 -2.18933374e-02 -1.32684720e+00
-3.78982663e-01 -5.83143950e-01 -5.48549332e-02 7.68112898e-01
-5.70978783e-02 8.42138529e-01 2.33343318e-01 3.48391421e-02
-4.95854825e-01 3.68578464e-01 1.10176969e+00 -5.30183837e-02
-2.75000423e-01 2.46677790e-02 -6.01987123e-01 8.25016081e-01
9.49735880e-01 -3.31626981e-01 -1.33429468e-01 -2.80683517e-01
6.91829801e-01 -4.65991348e-01 -5.64291954e-01 -9.18131709e-01
-1.43193588e-01 2.74184197e-01 3.57386410e-01 -2.95690924e-01
2.43954569e-01 -8.78858149e-01 -8.26464951e-01 4.58650172e-01
-8.48654062e-02 4.56015050e-01 7.03193069e-01 -2.16182411e-01
-5.98260343e-01 -7.18341768e-01 7.18326509e-01 -4.79793400e-02
-8.25682044e-01 -9.36291292e-02 -7.79240131e-01 -1.72171608e-01
8.59918833e-01 1.03452332e-01 -1.34735882e-01 -4.97155100e-01
-6.72918439e-01 1.14823297e-01 1.72227234e-01 8.22318852e-01
2.13237152e-01 -1.01559246e+00 -1.14709699e+00 2.35189363e-01
3.90758902e-01 -5.76927006e-01 -3.58260334e-01 7.50055909e-01
-1.11929488e+00 6.24999523e-01 -5.05911589e-01 -5.56311429e-01
-1.19755137e+00 3.32384467e-01 1.63058400e-01 -5.70911705e-01
1.29588187e-01 6.79503381e-01 -5.65243483e-01 -1.15597928e+00
-3.09732914e-01 -1.43817157e-01 -7.62625039e-01 6.75143838e-01
4.76251513e-01 7.63399899e-02 6.60513341e-01 -1.08435667e+00
-2.25813493e-01 5.58449090e-01 -1.99068993e-01 -3.40930074e-01
1.48278236e+00 -1.60652176e-01 -7.65519142e-01 8.59854698e-01
1.02884698e+00 4.23457593e-01 -3.15022290e-01 1.03456602e-01
3.79146278e-01 1.29148349e-01 -1.89827874e-01 -1.13299024e+00
-6.17159724e-01 9.49079514e-01 4.85356361e-01 5.53533494e-01
1.18714213e+00 -3.06422740e-01 5.86657643e-01 6.21305525e-01
6.05131924e-01 -1.14719033e+00 2.95612067e-01 1.06860626e+00
1.07141840e+00 -1.60394466e+00 -1.28563046e-01 -2.82776088e-01
-1.04279339e+00 1.60288143e+00 5.63532412e-01 -1.48739591e-01
1.00860918e+00 2.64718473e-01 7.25608170e-01 -9.62698609e-02
-3.79613549e-01 -1.95705235e-01 3.98160934e-01 5.96309066e-01
1.22073281e+00 1.38233975e-01 -6.33254707e-01 8.55935037e-01
-6.37698412e-01 -2.12406442e-01 8.53416026e-01 1.13023496e+00
-4.02091026e-01 -1.58924508e+00 -5.59649944e-01 3.30439985e-01
-6.26396954e-01 -2.27157488e-01 -4.53910977e-01 1.25961983e+00
-7.14046210e-02 1.27196062e+00 -1.46712035e-01 -2.37686723e-01
3.09444755e-01 5.74771285e-01 4.98751342e-01 -8.35226178e-01
-1.00164652e+00 -6.02542330e-03 2.42374286e-01 1.09376376e-02
-9.52327669e-01 -6.60601258e-01 -1.21206999e+00 -3.98807004e-02
-4.41400826e-01 4.39422280e-01 8.36378098e-01 1.19282150e+00
-1.19852826e-01 2.00435817e-01 3.60408962e-01 -8.98573160e-01
-4.20934737e-01 -1.56948221e+00 -6.67472720e-01 4.35129106e-01
-2.81006396e-02 -3.07881534e-01 -2.42828697e-01 3.57936829e-01] | [11.12940502166748, 7.066923141479492] |
deacced6-1ec6-487d-8c80-e7531238942b | 3d-line-mapping-revisited | 2303.17504 | null | https://arxiv.org/abs/2303.17504v1 | https://arxiv.org/pdf/2303.17504v1.pdf | 3D Line Mapping Revisited | In contrast to sparse keypoints, a handful of line segments can concisely encode the high-level scene layout, as they often delineate the main structural elements. In addition to offering strong geometric cues, they are also omnipresent in urban landscapes and indoor scenes. Despite their apparent advantages, current line-based reconstruction methods are far behind their point-based counterparts. In this paper we aim to close the gap by introducing LIMAP, a library for 3D line mapping that robustly and efficiently creates 3D line maps from multi-view imagery. This is achieved through revisiting the degeneracy problem of line triangulation, carefully crafted scoring and track building, and exploiting structural priors such as line coincidence, parallelism, and orthogonality. Our code integrates seamlessly with existing point-based Structure-from-Motion methods and can leverage their 3D points to further improve the line reconstruction. Furthermore, as a byproduct, the method is able to recover 3D association graphs between lines and points / vanishing points (VPs). In thorough experiments, we show that LIMAP significantly outperforms existing approaches for 3D line mapping. Our robust 3D line maps also open up new research directions. We show two example applications: visual localization and bundle adjustment, where integrating lines alongside points yields the best results. Code is available at https://github.com/cvg/limap. | ['Viktor Larsson', 'Marc Pollefeys', 'Rémi Pautrat', 'Yifan Yu', 'Shaohui Liu'] | 2023-03-30 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Liu_3D_Line_Mapping_Revisited_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Liu_3D_Line_Mapping_Revisited_CVPR_2023_paper.pdf | cvpr-2023-1 | ['visual-localization'] | ['computer-vision'] | [-6.98015317e-02 -3.41997057e-01 -2.76801497e-01 -9.89271849e-02
-7.33937383e-01 -9.16605175e-01 7.59290516e-01 2.61667579e-01
1.10113889e-01 3.54957134e-01 4.35539484e-02 -4.37823921e-01
-1.56250730e-01 -9.96846855e-01 -7.98631430e-01 -1.24583244e-01
-5.73836416e-02 3.94715548e-01 4.35843050e-01 -3.07192713e-01
4.39840198e-01 1.01313674e+00 -1.29760444e+00 -7.18259290e-02
7.86283553e-01 6.21257067e-01 2.44392455e-02 5.55623591e-01
-2.76341528e-01 2.14289978e-01 1.18161410e-01 -2.88695365e-01
4.34952110e-01 -1.08203404e-01 -4.52647537e-01 4.99395758e-01
8.46637487e-01 -4.23262626e-01 -2.79660791e-01 6.79950833e-01
3.61136734e-01 -1.34388492e-01 4.13789392e-01 -1.41278982e+00
-1.59275383e-01 -1.18856810e-01 -1.21123624e+00 -3.04291308e-01
8.34528744e-01 5.82825504e-02 1.10657406e+00 -1.34412408e+00
8.71784627e-01 9.52046096e-01 1.31304693e+00 -2.45381996e-01
-1.47639966e+00 -3.24237913e-01 7.83637352e-03 -1.02300942e-01
-1.68833387e+00 -6.08444035e-01 1.15716279e+00 -5.36227942e-01
9.69230354e-01 4.39718872e-01 9.01550710e-01 8.86850893e-01
-1.06777027e-01 7.33468890e-01 8.79436731e-01 -5.88363111e-01
-7.28986487e-02 -2.09948003e-01 -5.74015304e-02 8.30754936e-01
2.00889915e-01 -9.51777212e-03 -6.33923352e-01 -3.32546771e-01
1.43871367e+00 1.86937332e-01 -5.29623389e-01 -1.42231154e+00
-1.54324484e+00 5.72533429e-01 5.62782526e-01 -3.12589630e-02
-2.39732891e-01 2.12993294e-01 -4.54139672e-02 -1.18899591e-01
4.05085891e-01 2.14406312e-01 -7.33667091e-02 -8.44517574e-02
-1.10363472e+00 3.34518492e-01 5.30395210e-01 1.41493487e+00
1.21741331e+00 -3.54145825e-01 2.98025042e-01 8.13312411e-01
3.73999923e-01 7.13457942e-01 -3.48429054e-01 -1.33516502e+00
6.22650385e-01 7.16781378e-01 2.84558266e-01 -1.51922071e+00
-6.10102952e-01 -5.23083091e-01 -7.24109352e-01 2.97818363e-01
3.51840615e-01 3.22917372e-01 -5.32487094e-01 1.07337010e+00
4.88802999e-01 8.60228315e-02 -4.53744441e-01 6.11305833e-01
6.71010733e-01 4.82881606e-01 -6.42348766e-01 2.38551363e-01
1.17251766e+00 -9.16668177e-01 -3.34661931e-01 -3.51410657e-01
5.88368297e-01 -1.00665665e+00 1.08410490e+00 1.48733273e-01
-1.09791601e+00 -2.48872101e-01 -1.00703549e+00 -4.09610897e-01
-5.67747951e-02 1.73157379e-01 8.28439653e-01 3.22150737e-01
-1.01696837e+00 2.14049995e-01 -8.95696163e-01 -5.98765314e-01
5.35022855e-01 -7.98886716e-02 -5.93559623e-01 -2.33366802e-01
-3.31529528e-01 4.75148290e-01 -1.20290928e-02 9.55829024e-02
-1.58246532e-02 -9.67358172e-01 -1.05190170e+00 -2.13484630e-01
6.07691526e-01 -1.11091948e+00 1.06964540e+00 -2.00809896e-01
-1.12685382e+00 1.03263581e+00 -4.30265993e-01 -1.99693531e-01
6.48088932e-01 -2.27256507e-01 8.43313411e-02 2.93349117e-01
4.73371655e-01 6.26384556e-01 3.52675140e-01 -1.69476008e+00
-6.71681106e-01 -2.19690844e-01 9.64173954e-03 2.83570826e-01
2.32620344e-01 -5.58358908e-01 -8.67714882e-01 -4.80561316e-01
7.81342864e-01 -7.72199571e-01 -3.21484417e-01 7.01490581e-01
-6.99926496e-01 2.27727070e-01 7.64841557e-01 -2.24899232e-01
1.17632842e+00 -2.22830534e+00 -1.32854521e-01 5.94078243e-01
2.07468301e-01 -3.52949142e-01 1.01067342e-01 8.47194314e-01
1.49791062e-01 1.04437821e-01 -6.84783041e-01 -5.06729841e-01
-1.15547720e-02 1.57317162e-01 -3.31530333e-01 6.92345560e-01
7.76871368e-02 8.36815894e-01 -1.01378882e+00 -3.49717855e-01
6.57418847e-01 5.42418599e-01 -6.19622290e-01 -2.32714653e-01
2.16698162e-02 2.00586349e-01 -2.95000583e-01 8.36465478e-01
9.47929323e-01 -2.92908639e-01 -1.27654344e-01 -3.99715781e-01
-5.58388233e-01 5.91802585e-04 -1.44409776e+00 2.23376155e+00
-2.47107789e-01 7.48717725e-01 -5.25995530e-02 -4.09895450e-01
1.07214105e+00 -1.43195510e-01 6.49281383e-01 -6.01577997e-01
-3.82674694e-01 2.30155349e-01 -6.96661651e-01 -1.47174791e-01
5.86066008e-01 3.45728844e-01 -5.53272739e-02 2.41389439e-01
-2.77217656e-01 -9.13136721e-01 -1.54442772e-01 3.06214571e-01
9.52781498e-01 7.54199445e-01 8.37902486e-01 1.66426674e-02
1.41507551e-01 4.65725750e-01 3.97032768e-01 5.05937517e-01
1.70088768e-01 1.07505095e+00 3.11775953e-01 -2.87750542e-01
-1.41386402e+00 -1.29246533e+00 -3.69674563e-01 1.53349578e-01
3.62863958e-01 -7.86707282e-01 -2.26420909e-01 -2.52984017e-01
3.12040389e-01 4.69939321e-01 -3.09504926e-01 5.51827967e-01
-4.23950791e-01 -2.44145960e-01 3.30335766e-01 4.94501650e-01
6.07367933e-01 -1.72513962e-01 -6.29263461e-01 7.29837827e-03
-2.04900742e-01 -1.16127205e+00 -5.23639560e-01 -4.56928387e-02
-8.66164923e-01 -1.10526156e+00 -5.62050104e-01 -4.80513662e-01
7.59740412e-01 1.20512843e+00 1.14963186e+00 1.59508109e-01
-1.00083932e-01 7.21184313e-01 -2.20549420e-01 -3.61395441e-02
2.56607652e-01 -5.84540255e-02 -1.98039368e-01 -2.19000593e-01
1.96118236e-01 -9.10085559e-01 -6.68792307e-01 4.54918206e-01
-4.97335434e-01 6.06214225e-01 2.47397304e-01 6.18271530e-01
1.04899704e+00 -2.08071381e-01 -3.11080009e-01 -6.72764719e-01
1.74944833e-01 -3.85856628e-01 -8.13355923e-01 4.70474251e-02
-1.59310937e-01 -2.31350437e-01 1.01966135e-01 3.33267570e-01
-6.64186478e-01 4.44276512e-01 -1.12255484e-01 -4.05228853e-01
-2.91206121e-01 3.62368733e-01 3.16196904e-02 -3.27523708e-01
5.91191292e-01 4.13178885e-03 -2.40208536e-01 -4.65738982e-01
6.32929981e-01 2.73390085e-01 5.85985363e-01 -3.98547858e-01
1.20073879e+00 1.03237307e+00 3.18216383e-01 -1.09267390e+00
-8.04122865e-01 -8.24286520e-01 -1.17612326e+00 -2.19153255e-01
4.51370984e-01 -9.19270813e-01 -4.62502986e-01 2.04640701e-01
-1.18391156e+00 -3.46144706e-01 -2.15628386e-01 3.30333352e-01
-8.87383580e-01 6.59451306e-01 -2.61438519e-01 -6.38475358e-01
1.09471813e-01 -8.47352087e-01 1.43178737e+00 -8.66488889e-02
-2.82579333e-01 -1.19045305e+00 3.29288065e-01 1.56922773e-01
-7.90070444e-02 8.33411813e-01 6.86475158e-01 2.61900187e-01
-9.91329134e-01 -6.12025335e-02 -2.26118013e-01 -2.43836418e-01
1.20310456e-01 3.27984929e-01 -8.87582123e-01 -2.33845953e-02
-4.74286854e-01 2.19974965e-01 5.26676953e-01 4.17050660e-01
8.06825042e-01 3.71585693e-03 -6.33548737e-01 1.21594822e+00
1.65095222e+00 -2.48087376e-01 6.76620841e-01 8.25812519e-01
9.56929624e-01 5.85045159e-01 5.75074077e-01 5.66437185e-01
8.32759082e-01 1.05953300e+00 5.25414109e-01 -4.56992716e-01
-2.25672513e-01 -7.83272505e-01 -6.06443956e-02 6.12807393e-01
-1.40020624e-01 -2.24193726e-02 -1.24266076e+00 3.78957808e-01
-1.99108350e+00 -1.08646929e+00 -9.33592677e-01 2.31490111e+00
3.67171645e-01 7.06141302e-03 1.97713405e-01 2.29593456e-01
2.62055308e-01 1.70431733e-01 -3.09745282e-01 1.08037516e-01
-4.94096994e-01 -1.58522308e-01 8.12241077e-01 8.08088899e-01
-1.12792385e+00 9.98093545e-01 5.91610622e+00 6.53051436e-01
-7.93316007e-01 -2.34506145e-01 1.51917994e-01 2.36271486e-01
-7.25142121e-01 4.58081752e-01 -7.35751688e-01 -6.01548366e-02
5.99075258e-02 1.70724764e-02 2.14327246e-01 7.16525853e-01
4.22703534e-01 -3.21921945e-01 -8.34689617e-01 1.24874473e+00
1.05857529e-01 -1.64345312e+00 -4.99474630e-02 2.13733956e-01
7.38187611e-01 1.21210329e-01 -1.04112327e-01 -4.91931587e-01
1.39758572e-01 -7.79709220e-01 8.93790126e-01 5.54884136e-01
7.78400004e-01 -6.39732838e-01 1.63732871e-01 4.08589274e-01
-1.41172051e+00 4.16166276e-01 -2.54232883e-01 -1.19899914e-01
4.45665836e-01 7.31224000e-01 -9.53802288e-01 1.03017926e+00
5.95564842e-01 1.16002500e+00 -6.94576442e-01 1.56952167e+00
-3.29459131e-01 2.16262519e-01 -7.21240401e-01 4.99867529e-01
4.40696627e-01 -5.98652899e-01 5.97454548e-01 1.14840603e+00
5.22748888e-01 -1.03179373e-01 1.61908135e-01 8.98719966e-01
3.05838853e-01 1.58247739e-01 -1.21547914e+00 6.17066920e-01
7.98429847e-01 1.27012300e+00 -1.05131519e+00 5.15421294e-02
-6.24066830e-01 1.02909172e+00 3.37135941e-01 5.25569975e-01
-5.90935767e-01 -3.42897862e-01 4.50681895e-01 5.37102640e-01
3.07916969e-01 -1.12416375e+00 -6.96796417e-01 -1.15406013e+00
2.31818855e-01 -4.98417199e-01 -5.93399331e-02 -1.05268288e+00
-1.00953507e+00 1.33600786e-01 1.05266487e-02 -1.73296261e+00
-4.85009030e-02 -4.48107928e-01 -3.26204091e-01 6.48656905e-01
-1.67575312e+00 -1.38762867e+00 -6.29793704e-01 5.78183413e-01
3.77025992e-01 4.87772197e-01 7.65892386e-01 8.47129524e-02
-2.00321063e-01 5.01357056e-02 2.86395043e-01 -4.85106520e-02
5.57155848e-01 -1.21059716e+00 8.14991653e-01 9.73382473e-01
4.52159882e-01 6.21687233e-01 3.95410061e-01 -6.35431767e-01
-1.61388195e+00 -8.53776872e-01 5.88130295e-01 -7.09588170e-01
5.26832640e-01 -6.81165934e-01 -6.94697678e-01 8.68052781e-01
-1.38167053e-01 -3.68805416e-02 6.35043919e-01 8.02146643e-02
-4.47841555e-01 5.67830727e-02 -8.56527746e-01 8.86362970e-01
1.51551950e+00 -5.11590600e-01 -3.49152029e-01 3.23950350e-01
3.67240250e-01 -6.77562177e-01 -8.21922898e-01 2.48760745e-01
6.03902757e-01 -1.53573930e+00 1.61974788e+00 2.84777999e-01
1.93436071e-01 -7.34021842e-01 -2.11290672e-01 -1.27570593e+00
-4.42137808e-01 -7.41600931e-01 9.94389355e-02 1.23175967e+00
1.81841522e-01 -5.62862277e-01 7.30531871e-01 2.16747612e-01
-3.42475027e-01 -5.89439869e-01 -8.43527853e-01 -6.26124859e-01
-4.26405936e-01 -6.91240191e-01 7.34977543e-01 1.04998243e+00
-1.57600507e-01 2.71887541e-01 -3.02569568e-01 4.26557332e-01
9.96360362e-01 4.40896392e-01 1.43241501e+00 -1.05872607e+00
-6.73761070e-02 -6.02359891e-01 -4.79048014e-01 -1.61871874e+00
-2.58468777e-01 -9.42637682e-01 -2.08354011e-01 -1.94209266e+00
-2.15579435e-01 -9.19937074e-01 6.61704421e-01 4.58535671e-01
1.47441670e-01 4.90392625e-01 9.42944065e-02 4.07651335e-01
-3.67099524e-01 4.94626671e-01 1.11766195e+00 2.05477640e-01
-3.22652251e-01 -6.41905516e-02 -4.19927508e-01 1.00998962e+00
6.92503572e-01 -4.07575928e-02 -2.10209668e-01 -7.89816260e-01
4.40411389e-01 6.81922659e-02 6.92952216e-01 -8.78306210e-01
5.06192803e-01 -2.32029155e-01 4.13398236e-01 -1.19265676e+00
6.72791421e-01 -9.83026445e-01 5.12719214e-01 3.13977189e-02
2.16710836e-01 1.60861865e-01 2.41830826e-01 5.74198067e-01
6.49138987e-02 -1.09102495e-01 2.91047156e-01 -2.22405016e-01
-9.21924710e-01 4.10701275e-01 1.26806304e-01 -3.35980773e-01
1.03187919e+00 -8.73383760e-01 -3.41186374e-01 -3.94023001e-01
-3.30012590e-01 3.66347253e-01 1.29919648e+00 5.17685618e-03
9.45730329e-01 -1.53748012e+00 -5.37782133e-01 3.93368781e-01
5.01416683e-01 4.94646251e-01 1.09472699e-01 9.93531108e-01
-1.09081757e+00 3.49247724e-01 -1.94058970e-01 -1.14332366e+00
-1.06652105e+00 2.89502054e-01 2.55554408e-01 3.53138804e-01
-1.17885637e+00 5.12840390e-01 1.72868952e-01 -4.77748930e-01
-1.93104297e-02 -3.85769784e-01 2.72049963e-01 -1.26618862e-01
1.89388216e-01 5.92985153e-01 -3.67900878e-02 -7.56572068e-01
-3.84901583e-01 1.49607599e+00 4.72877800e-01 -1.69102445e-01
1.54612100e+00 -3.91565382e-01 5.06937644e-03 5.15008092e-01
9.82862532e-01 7.25468218e-01 -1.45161784e+00 -1.27721339e-01
6.31676018e-02 -9.53242600e-01 -4.97938171e-02 -2.39639357e-01
-5.19459844e-01 8.78480315e-01 8.76829550e-02 1.93031475e-01
8.18289936e-01 -5.44952862e-02 5.48226297e-01 1.82259351e-01
8.61997366e-01 -5.03523350e-01 -2.55760461e-01 3.64079863e-01
9.01112854e-01 -1.02864993e+00 3.36294889e-01 -9.99870420e-01
-2.83055663e-01 1.46753526e+00 1.68807238e-01 -2.31855780e-01
4.59029227e-01 3.69108975e-01 -1.08392186e-01 -4.30011809e-01
-1.30794525e-01 -1.82874173e-01 2.02605322e-01 8.95437360e-01
2.89977103e-01 3.40902582e-02 1.68360427e-01 -1.42774194e-01
-4.38363850e-01 -1.66684657e-01 5.42100012e-01 9.68513548e-01
-4.22475129e-01 -1.17233515e+00 -8.10507178e-01 1.09638005e-01
4.02676702e-01 -3.28924693e-02 -2.19424710e-01 9.86856222e-01
-2.18551591e-01 6.60262346e-01 2.01633215e-01 -1.93691507e-01
7.30325878e-01 -4.46204364e-01 5.39620280e-01 -5.15192151e-01
7.38208741e-02 1.88281789e-01 1.60041198e-01 -9.43025291e-01
-5.72869539e-01 -9.08942163e-01 -1.08195460e+00 -3.60460192e-01
-2.19705090e-01 -1.75110966e-01 5.74195147e-01 4.18424577e-01
5.68668485e-01 -1.18926920e-01 6.11262918e-01 -1.14497709e+00
7.59211776e-04 -2.08635792e-01 -4.30762082e-01 1.79635376e-01
4.39786553e-01 -8.76275003e-01 -2.20777288e-01 -1.01307547e-02] | [7.965060234069824, -2.276822328567505] |
2483abd4-c3f7-426f-915d-4059a71ff2c9 | a-parallel-algorithm-for-exact-bayesian | 1408.1664 | null | http://arxiv.org/abs/1408.1664v3 | http://arxiv.org/pdf/1408.1664v3.pdf | A Parallel Algorithm for Exact Bayesian Structure Discovery in Bayesian Networks | Exact Bayesian structure discovery in Bayesian networks requires exponential
time and space. Using dynamic programming (DP), the fastest known sequential
algorithm computes the exact posterior probabilities of structural features in
$O(2(d+1)n2^n)$ time and space, if the number of nodes (variables) in the
Bayesian network is $n$ and the in-degree (the number of parents) per node is
bounded by a constant $d$. Here we present a parallel algorithm capable of
computing the exact posterior probabilities for all $n(n-1)$ edges with optimal
parallel space efficiency and nearly optimal parallel time efficiency. That is,
if $p=2^k$ processors are used, the run-time reduces to
$O(5(d+1)n2^{n-k}+k(n-k)^d)$ and the space usage becomes $O(n2^{n-k})$ per
processor. Our algorithm is based the observation that the subproblems in the
sequential DP algorithm constitute a $n$-$D$ hypercube. We take a delicate way
to coordinate the computation of correlated DP procedures such that large
amount of data exchange is suppressed. Further, we develop parallel techniques
for two variants of the well-known \emph{zeta transform}, which have
applications outside the context of Bayesian networks. We demonstrate the
capability of our algorithm on datasets with up to 33 variables and its
scalability on up to 2048 processors. We apply our algorithm to a biological
data set for discovering the yeast pheromone response pathways. | ['Yetian Chen', 'Olga Nikolova', 'Jin Tian', 'Srinivas Aluru'] | 2014-08-07 | null | null | null | null | ['2048'] | ['playing-games'] | [ 2.57329136e-01 3.73552628e-02 7.06037804e-02 -2.60632634e-01
-4.92158890e-01 -6.27276480e-01 2.62466520e-02 3.45307469e-01
-8.08710337e-01 8.17674994e-01 -6.70340776e-01 -5.40104747e-01
-6.12574339e-01 -1.13015175e+00 -7.12701261e-01 -1.00977468e+00
-1.03865850e+00 1.03839231e+00 5.14169157e-01 2.49964416e-01
4.18714970e-01 6.00937486e-01 -1.55407882e+00 -5.31557918e-01
3.08467835e-01 9.76389885e-01 -5.01087904e-02 1.03610647e+00
-1.43737318e-02 -5.65165281e-02 -4.90962386e-01 -4.40270185e-01
4.87858444e-01 -3.27818960e-01 -7.42599487e-01 -3.23118120e-01
-1.57381326e-01 6.40390739e-02 -3.20116132e-01 1.33796763e+00
4.23154831e-01 1.92233860e-01 5.83019078e-01 -1.40189314e+00
2.66833991e-01 8.52705777e-01 -1.23276722e+00 2.20001310e-01
2.59064674e-01 -1.72770247e-01 1.00199878e+00 -2.23933578e-01
5.87975085e-01 1.44754863e+00 5.04235327e-01 6.15911596e-02
-1.59818113e+00 -1.05926967e+00 1.68954358e-01 7.18196109e-02
-1.73830664e+00 -1.18888311e-01 5.39670289e-01 -1.46938846e-01
9.97699261e-01 2.95396835e-01 7.02989042e-01 2.02292800e-01
1.94455042e-01 2.87080079e-01 1.16510963e+00 -4.68230963e-01
6.57549441e-01 -3.40407342e-01 3.35808814e-01 8.80380273e-01
7.43393064e-01 2.44960055e-01 -6.02040708e-01 -5.92654467e-01
5.83017290e-01 -1.27300739e-01 1.15248568e-01 -2.39690673e-02
-6.80450797e-01 7.95453370e-01 -1.19630016e-01 -1.84050146e-02
-1.37663186e-01 6.31532967e-01 2.28782341e-01 2.88629889e-01
9.70948413e-02 1.14668995e-01 -8.70187342e-01 -2.58062094e-01
-6.88759923e-01 3.77673686e-01 1.13157189e+00 1.02267885e+00
9.73299026e-01 -2.33422562e-01 5.92135012e-01 4.83536333e-01
4.07966942e-01 1.03876817e+00 -3.13683808e-01 -1.18382764e+00
2.43778020e-01 3.72871876e-01 1.53201342e-01 -1.18790019e+00
-4.10231322e-01 -6.92007765e-02 -1.05825448e+00 3.01483572e-01
6.68296754e-01 -3.70523661e-01 -6.06059551e-01 2.01233816e+00
5.60205102e-01 -2.26074591e-01 -2.62683243e-01 2.45905712e-01
-5.84285222e-02 7.30068624e-01 -1.62972510e-02 -7.36350954e-01
1.79138482e+00 -4.24271934e-02 -3.15030247e-01 -5.53312711e-02
3.49322975e-01 -6.42455757e-01 3.72625649e-01 7.81319976e-01
-1.21339154e+00 -1.92018431e-02 -9.26461518e-01 4.82357860e-01
-3.85533571e-01 -4.41054970e-01 9.99469459e-01 9.50496972e-01
-8.92692506e-01 5.48024356e-01 -1.00615358e+00 -3.18011679e-02
1.51823878e-01 9.65371668e-01 -1.44422263e-01 -2.77431965e-01
-7.72363544e-01 2.47085422e-01 3.40174556e-01 3.24353799e-02
-8.12317193e-01 -3.46995801e-01 -3.26660722e-01 4.29583825e-02
5.78895211e-01 -4.49801892e-01 6.19522691e-01 -2.83633202e-01
-1.26316202e+00 4.25602704e-01 -4.10115212e-01 -1.29525527e-01
7.92582706e-02 4.61773694e-01 -2.25403067e-02 2.00152859e-01
-1.71926007e-01 5.17724514e-01 3.64797920e-01 -8.25527430e-01
-8.29422474e-01 -1.00233626e+00 -5.70123903e-02 -1.30966336e-01
-8.54195748e-03 2.71400869e-01 -5.17050147e-01 -2.12347135e-01
6.34173393e-01 -1.24148679e+00 -5.88638842e-01 -8.99328738e-02
-3.82341862e-01 -3.37574869e-01 3.36127579e-01 -2.60908425e-01
9.92725909e-01 -2.05778170e+00 1.76312611e-01 9.87372994e-01
3.70918512e-01 -2.33794615e-01 8.74953270e-02 3.04744542e-01
6.96334466e-02 9.41505283e-02 -2.89364308e-01 1.54944330e-01
1.32996619e-01 4.00005639e-01 1.53357387e-01 6.86369121e-01
-4.73464400e-01 -5.98911494e-02 -5.46102881e-01 -4.68024492e-01
-4.95750606e-01 3.04987103e-01 -5.58241189e-01 -1.64447516e-01
-2.22250015e-01 -3.86040598e-01 -6.06220782e-01 6.05697453e-01
1.18486500e+00 -3.20988327e-01 8.33348811e-01 1.55732572e-01
-1.39812589e-01 -7.35556409e-02 -1.83226740e+00 1.30721796e+00
5.49019489e-04 2.10306585e-01 5.84474385e-01 -1.03883934e+00
7.83982277e-01 1.85183242e-01 6.79118872e-01 -4.07950073e-01
1.25677332e-01 2.01288357e-01 1.61888093e-01 3.75034921e-02
1.55185893e-01 -2.77304053e-01 -3.24693918e-01 7.80827105e-01
-2.19498485e-01 -7.35464320e-02 4.99416053e-01 1.17417000e-01
1.79624796e+00 -3.68905455e-01 1.24034740e-01 -3.84339541e-01
1.44741073e-01 7.41934180e-02 1.07438302e+00 8.16442668e-01
-1.73610538e-01 -4.05837119e-01 1.23597515e+00 -2.81398594e-01
-8.70484352e-01 -9.24586892e-01 -1.63060188e-01 1.08025241e+00
1.12393655e-01 -3.90984446e-01 -6.96005762e-01 -3.08623761e-01
5.68637773e-02 3.51890743e-01 -5.46881855e-01 4.24781621e-01
-4.90796953e-01 -1.55428958e+00 5.23287296e-01 4.30375278e-01
3.67462963e-01 -6.09406710e-01 -6.62457526e-01 3.96477431e-01
3.62162679e-01 -6.93186581e-01 -5.82227297e-03 8.73022616e-01
-1.04104745e+00 -1.28532839e+00 -1.10959359e-01 -5.72899938e-01
1.03254199e+00 1.56622946e-01 7.04523802e-01 -2.84623623e-01
-4.52799588e-01 7.39936084e-02 3.44937518e-02 -4.35906440e-01
4.52154092e-02 -3.19032252e-01 1.08329393e-01 -8.75102460e-01
6.85008943e-01 -9.38030839e-01 -5.45469224e-01 2.31426120e-01
-8.21405947e-01 -4.58083153e-01 5.50217628e-01 8.30046475e-01
8.10764313e-01 7.80579269e-01 -3.28595713e-02 -9.54806566e-01
2.78519183e-01 -3.04901540e-01 -1.61292207e+00 1.76450480e-02
-7.65030026e-01 3.10367942e-01 6.09811664e-01 -3.48023444e-01
-7.25926876e-01 5.46434104e-01 7.19004422e-02 2.60045715e-02
1.47987515e-01 6.19772375e-01 -1.51125669e-01 -6.02597706e-02
3.70595574e-01 4.60063927e-02 -8.99404660e-02 -3.85558695e-01
7.69790187e-02 2.92615086e-01 1.52565911e-01 -8.48780453e-01
5.43513298e-01 4.96243864e-01 9.16836560e-01 -6.57777309e-01
-3.72313820e-02 -4.91588823e-02 -1.29795879e-01 3.46070558e-01
4.42647457e-01 -5.20807981e-01 -1.80369937e+00 3.17589104e-01
-9.75599945e-01 -8.06472003e-02 5.20599186e-02 4.88385022e-01
-3.43777925e-01 4.19856489e-01 -5.07387400e-01 -1.17333639e+00
-2.15959117e-01 -1.14589381e+00 4.98504966e-01 -5.65839466e-03
1.31401662e-02 -4.41894203e-01 2.38698497e-01 -8.56124535e-02
6.41851723e-02 1.28172785e-01 1.36886919e+00 -5.68053901e-01
-7.25528121e-01 -3.37756127e-01 -4.22899604e-01 -2.28473544e-01
-3.26409042e-01 4.32202928e-02 -3.80532652e-01 -2.79094279e-01
-6.48575276e-02 5.90674430e-02 6.73627973e-01 3.96887928e-01
1.06277764e+00 -3.51636112e-01 -7.54990876e-01 3.72286677e-01
1.72538590e+00 6.19544089e-01 3.40660155e-01 -1.76795349e-01
8.31084624e-02 5.38784325e-01 3.63493174e-01 8.24333429e-01
1.46981597e-01 3.56275439e-01 4.13880140e-01 2.93694496e-01
3.93772095e-01 1.84967637e-01 1.55419484e-01 7.59456933e-01
-3.16333845e-02 -3.90311986e-01 -9.91955757e-01 4.86321598e-01
-1.51303852e+00 -6.98976576e-01 -1.63159534e-01 2.33929825e+00
1.12533402e+00 2.92830288e-01 -3.27199213e-02 4.02016312e-01
6.84036434e-01 -3.42448324e-01 -4.83815730e-01 -6.32618248e-01
1.35487735e-01 8.41801643e-01 9.30148304e-01 5.88479161e-01
-6.14381731e-01 4.91127431e-01 6.42266798e+00 1.14126813e+00
-4.52843189e-01 -5.95456958e-02 7.55459428e-01 -3.35519016e-01
-6.82125762e-02 3.46149862e-01 -1.08437264e+00 5.14173031e-01
1.28700364e+00 -9.64807868e-02 5.75425208e-01 9.16119516e-01
-6.52321726e-02 -8.04817736e-01 -1.14576352e+00 1.00141943e+00
-3.23454976e-01 -9.53575671e-01 -4.97365445e-01 5.73413849e-01
4.44911987e-01 -1.29177704e-01 -1.95399627e-01 -2.83600986e-01
1.13188469e+00 -8.57798517e-01 5.18306121e-02 -5.33268191e-02
6.95498645e-01 -1.15352106e+00 6.61526680e-01 3.49393576e-01
-1.26497054e+00 -9.65981260e-02 -5.92408121e-01 3.01154424e-02
4.18698676e-02 1.07925153e+00 -7.67787755e-01 2.97808886e-01
1.02673948e+00 -1.77086949e-01 1.70616433e-01 7.33104944e-01
6.37181476e-02 2.80477434e-01 -1.35942888e+00 -4.51513112e-01
2.08713058e-02 -3.09495449e-01 3.04076284e-01 9.45268333e-01
5.15651107e-01 7.35392034e-01 -4.27489262e-03 3.13226312e-01
-6.18189834e-02 -1.10974163e-01 -2.99312383e-01 -7.42849633e-02
8.86206985e-01 8.15130770e-01 -1.20571470e+00 -1.97103024e-01
1.56805979e-03 3.87754768e-01 1.03756197e-01 -1.21532585e-02
-5.53552270e-01 -9.56641555e-01 7.28029132e-01 -8.19628611e-02
7.05636263e-01 -5.35443246e-01 -2.18005642e-01 -3.60601664e-01
-2.32882813e-01 -7.89532840e-01 6.12312376e-01 -1.62622422e-01
-1.14552391e+00 3.61906111e-01 1.61793962e-01 -4.24219161e-01
-1.64264008e-01 -8.64661932e-01 1.26189619e-01 7.60407925e-01
-9.84468281e-01 -3.60008180e-01 2.46134788e-01 6.02654278e-01
-4.16862160e-01 7.89827853e-02 1.13028514e+00 8.64666328e-02
-6.40251637e-01 4.42696124e-01 4.63102162e-01 -1.40043288e-01
2.27390647e-01 -1.04349232e+00 1.26108676e-02 7.61620641e-01
-2.60309339e-01 7.47646034e-01 1.02710736e+00 -6.90420687e-01
-1.88948309e+00 -3.83410096e-01 9.48255658e-01 2.28331193e-01
5.86266041e-01 -3.25243980e-01 -3.64612609e-01 3.80770355e-01
-1.18451312e-01 -5.91154285e-02 1.08102012e+00 3.07086259e-01
-4.22987849e-01 -6.18233204e-01 -1.45932066e+00 4.90442127e-01
1.03878391e+00 -1.53361067e-01 1.13113120e-01 3.82645875e-01
2.62210876e-01 -2.10320324e-01 -1.07262349e+00 1.35299534e-01
5.48375010e-01 -7.75062978e-01 7.19909489e-01 -1.60670280e-01
-1.86250716e-01 -5.06675184e-01 -4.59262550e-01 -6.15458131e-01
-1.14767760e-01 -8.94483089e-01 1.51132241e-01 6.90225184e-01
4.35833514e-01 -7.37686515e-01 1.14797103e+00 6.27137661e-01
4.85617697e-01 -5.42432785e-01 -1.50555634e+00 -4.91261750e-01
-3.20186093e-02 -5.52254021e-01 5.55246294e-01 6.50341034e-01
2.19707802e-01 3.81285474e-02 -3.03494595e-02 4.02077377e-01
1.00926983e+00 8.63920152e-02 6.58750474e-01 -1.44273245e+00
-7.84850478e-01 -1.51797801e-01 -3.76329780e-01 -1.17701352e+00
-9.77857113e-02 -4.52962279e-01 1.12950429e-02 -7.29048669e-01
5.84225774e-01 -8.32461357e-01 -1.79190144e-01 5.54770529e-01
2.82392651e-01 -1.08905770e-01 -2.16387868e-01 -2.28695720e-01
-4.55841154e-01 1.56228170e-02 7.37636387e-01 1.36730611e-01
-8.19707438e-02 -5.10380566e-02 -5.41445136e-01 7.75509179e-01
4.98093605e-01 -1.03539467e+00 -2.67295271e-01 -4.23528641e-01
7.50881374e-01 7.06202328e-01 -1.39207035e-01 -6.34646356e-01
5.18410802e-01 -3.27493846e-01 4.04975802e-01 -8.64932418e-01
6.05144560e-01 -8.45249176e-01 5.00365853e-01 8.09930086e-01
-2.24712528e-02 4.84518111e-01 1.79848790e-01 9.14883137e-01
1.64366022e-01 -4.39799339e-01 6.80739522e-01 -3.52110595e-01
-2.21273899e-02 2.79139310e-01 -6.02243781e-01 -3.35756302e-01
1.13430858e+00 -1.09805062e-01 -3.47840399e-01 -6.66277185e-02
-7.04935491e-01 2.10208401e-01 2.55055457e-01 -6.83862031e-01
3.98251563e-01 -6.67447090e-01 -3.30627263e-01 1.11567527e-01
-4.90374357e-01 2.94643939e-01 3.32013547e-01 5.84118545e-01
-7.51401067e-01 3.45986813e-01 -1.21727727e-01 -5.80692768e-01
-1.41177440e+00 5.91561258e-01 -7.28888512e-02 -5.20090222e-01
-6.79420531e-02 1.27451408e+00 -3.86428833e-01 8.38926360e-02
2.47121036e-01 -3.41943651e-02 5.34115732e-01 -2.80835163e-02
4.91463512e-01 8.09225798e-01 -1.95852697e-01 1.69992700e-01
-5.89967310e-01 5.63656747e-01 -2.87694514e-01 -6.72782302e-01
1.64529216e+00 9.62664932e-02 -8.14542234e-01 -6.73624575e-02
1.02873313e+00 -8.46687332e-03 -1.05107033e+00 -1.48052827e-01
3.38750817e-02 -3.65604460e-01 3.88923436e-02 -4.45816427e-01
-1.05454445e+00 5.61364889e-01 7.43206620e-01 2.16135204e-01
1.34616065e+00 -1.99275371e-02 2.82671779e-01 6.38538420e-01
7.61503398e-01 -1.22105241e+00 -1.98714092e-01 4.21505183e-01
5.03609888e-02 -6.49724841e-01 5.21772981e-01 -5.64620137e-01
6.21882975e-02 1.11531532e+00 2.77357399e-01 -9.12536159e-02
1.11758828e+00 7.88316131e-01 -6.54268265e-01 -2.66944826e-01
-9.60766912e-01 1.53987244e-01 -6.98689997e-01 4.50190872e-01
1.29536659e-01 1.60865903e-01 -6.03755534e-01 6.38375819e-01
-2.30248049e-01 -7.10553452e-02 4.16058987e-01 1.26109338e+00
-6.76533103e-01 -1.56825423e+00 -4.25685704e-01 4.09396827e-01
-5.39615393e-01 -7.39939809e-02 -2.68072020e-02 6.67475164e-01
1.97386041e-01 9.81473565e-01 2.80440927e-01 -1.19818352e-01
-1.50937095e-01 1.79220244e-01 7.03833044e-01 -2.38243774e-01
-3.14471781e-01 4.45280641e-01 5.34697473e-02 -5.29822767e-01
-3.55548441e-01 -9.08479750e-01 -1.36047518e+00 -9.96111691e-01
-2.61118919e-01 4.28648055e-01 1.00963724e+00 5.57228923e-01
3.64431739e-01 1.05588011e-01 4.86983597e-01 -2.76674390e-01
-5.07909894e-01 -7.02886045e-01 -1.19388616e+00 -4.79725987e-01
-3.15622598e-01 -5.99812269e-01 -4.29928243e-01 -1.18112020e-01] | [6.728865623474121, 4.8850274085998535] |
ccac7b0f-18e2-4ef7-a0fa-5b47fe80aa6e | attention-based-cnn-lstm-and-xgboost-hybrid | 2204.02623 | null | https://arxiv.org/abs/2204.02623v2 | https://arxiv.org/pdf/2204.02623v2.pdf | Attention-based CNN-LSTM and XGBoost hybrid model for stock prediction | Stock market plays an important role in the economic development. Due to the complex volatility of the stock market, the research and prediction on the change of the stock price, can avoid the risk for the investors. The traditional time series model ARIMA can not describe the nonlinearity, and can not achieve satisfactory results in the stock prediction. As neural networks are with strong nonlinear generalization ability, this paper proposes an attention-based CNN-LSTM and XGBoost hybrid model to predict the stock price. The model constructed in this paper integrates the time series model, the Convolutional Neural Networks with Attention mechanism, the Long Short-Term Memory network, and XGBoost regressor in a non-linear relationship, and improves the prediction accuracy. The model can fully mine the historical information of the stock market in multiple periods. The stock data is first preprocessed through ARIMA. Then, the deep learning architecture formed in pretraining-finetuning framework is adopted. The pre-training model is the Attention-based CNN-LSTM model based on sequence-to-sequence framework. The model first uses convolution to extract the deep features of the original stock data, and then uses the Long Short-Term Memory networks to mine the long-term time series features. Finally, the XGBoost model is adopted for fine-tuning. The results show that the hybrid model is more effective and the prediction accuracy is relatively high, which can help investors or institutions to make decisions and achieve the purpose of expanding return and avoiding risk. Source code is available at https://github.com/zshicode/Attention-CLX-stock-prediction. | ['Jian Wu', 'Guangliang Mo', 'Yang Hu', 'Zhuangwei Shi'] | 2022-04-06 | null | null | null | null | ['stock-prediction'] | ['time-series'] | [-9.30188775e-01 -6.51288450e-01 -2.08344147e-01 -1.75737143e-01
6.35884553e-02 -2.17507064e-01 3.51624548e-01 -6.95784986e-01
-3.92259240e-01 5.54260969e-01 1.60948426e-01 -6.20753706e-01
-7.20548257e-02 -1.21972156e+00 -5.49920917e-01 -6.49879754e-01
-2.16044828e-01 -6.38163707e-04 2.58592218e-01 -4.02363986e-01
5.12355030e-01 3.54016870e-01 -1.40980160e+00 4.17348295e-02
6.56091809e-01 1.65750325e+00 3.52864206e-01 3.36120903e-01
-7.82072902e-01 1.20529699e+00 -4.56222683e-01 -1.53939143e-01
4.20088440e-01 -2.91838825e-01 -2.32789665e-01 -5.08981049e-01
-3.74447018e-01 -7.65653074e-01 -4.62696999e-01 1.08350921e+00
3.91543716e-01 1.85689013e-02 2.81949550e-01 -9.64913487e-01
-1.17123294e+00 8.76733840e-01 -2.87137687e-01 8.37971032e-01
-5.81254423e-01 2.86366343e-01 6.43448532e-01 -8.32455277e-01
-1.97889313e-01 1.11738479e+00 7.66352534e-01 2.81216353e-01
-5.12159348e-01 -1.17976022e+00 2.80629396e-01 2.67505795e-01
-9.76702034e-01 -4.24946398e-02 6.98928356e-01 -5.04209042e-01
1.18343508e+00 7.32962936e-02 1.13345468e+00 6.44925892e-01
7.99547136e-01 7.99981236e-01 1.01397192e+00 -3.14822830e-02
-5.56199290e-02 -1.26453545e-02 2.01889589e-01 2.40538225e-01
2.40711689e-01 6.40975237e-01 1.81182902e-02 1.58820197e-01
1.00676882e+00 8.41041803e-01 -9.02212188e-02 7.74063766e-01
-8.56818914e-01 9.33773279e-01 8.21754515e-01 8.34635377e-01
-7.73192883e-01 3.33201796e-01 4.37180012e-01 6.31112576e-01
7.04614043e-01 4.40404415e-01 -9.02426600e-01 -1.54923424e-01
-1.10493827e+00 2.52309799e-01 5.60745239e-01 4.81797814e-01
5.68019450e-01 1.02580976e+00 7.89578483e-02 4.70349610e-01
3.41422111e-01 5.74792027e-01 1.46470892e+00 -3.48892212e-01
1.60606518e-01 5.51645398e-01 -2.12083787e-01 -1.00621831e+00
-4.50810134e-01 -8.31229568e-01 -8.81788135e-01 4.94348735e-01
-1.89561218e-01 -5.05998611e-01 -9.63086128e-01 1.34629560e+00
-1.73444301e-01 5.07979095e-01 2.09036380e-01 6.90989554e-01
6.66941226e-01 1.34965742e+00 -3.03559843e-02 -4.53014582e-01
1.15568757e+00 -1.01251554e+00 -9.55770552e-01 -1.30776480e-01
1.80616006e-01 -6.33987129e-01 6.69704199e-01 2.27734387e-01
-1.01273584e+00 -8.22901189e-01 -1.07598400e+00 2.06837773e-01
-8.48645806e-01 -5.34216389e-02 4.63487923e-01 1.71842471e-01
-8.43529046e-01 9.22729313e-01 -7.26900637e-01 4.98225242e-01
2.08550826e-01 4.05101031e-01 3.25891584e-01 5.25322855e-01
-1.78488100e+00 1.02927399e+00 9.93843019e-01 5.95160544e-01
-3.19112986e-01 -7.62055337e-01 -4.74593788e-01 4.61967140e-01
2.43790485e-02 -3.24342966e-01 1.44077384e+00 -1.43037021e+00
-1.75007749e+00 -8.24304372e-02 4.49768305e-01 -6.39410079e-01
4.47507948e-01 -1.76641539e-01 -8.16993058e-01 -4.48405474e-01
-1.07541293e-01 1.01099849e-01 7.86974967e-01 -1.97243109e-01
-9.43478525e-01 -1.14077337e-01 -4.11245733e-01 -4.35801819e-02
-3.95791322e-01 1.26717851e-01 2.51046941e-02 -1.08641088e+00
-7.29142353e-02 -5.62664032e-01 -1.24506988e-01 -8.43482912e-01
1.53728440e-01 -2.68039107e-01 8.25354874e-01 -1.01550329e+00
1.49048746e+00 -2.06882977e+00 -5.43502867e-01 2.62797683e-01
-2.59367764e-01 5.74239790e-01 1.01794720e-01 2.12239161e-01
-5.01105964e-01 1.97538406e-01 1.20769538e-01 2.82227486e-01
3.51705775e-02 4.55399975e-02 -9.14123535e-01 -5.39089590e-02
1.97401717e-01 1.32015419e+00 -5.49291790e-01 -1.33364676e-02
9.55359042e-02 1.03461765e-01 9.71445516e-02 2.53658623e-01
-4.55449134e-01 -9.17196050e-02 -5.22282183e-01 4.04015809e-01
6.98953748e-01 -2.35031947e-01 -3.87617290e-01 1.22085072e-01
-7.49736786e-01 1.97749436e-01 -1.03752434e+00 8.78084660e-01
-3.25331956e-01 4.59821790e-01 -5.46255827e-01 -8.30686450e-01
1.34705997e+00 4.50092286e-01 2.75796235e-01 -1.08642316e+00
4.40445989e-01 7.36657619e-01 1.10534839e-01 -4.80023056e-01
2.83378899e-01 -4.67542738e-01 2.13817373e-01 5.74307084e-01
-1.42556399e-01 4.46398377e-01 -1.79645225e-01 -5.09230375e-01
4.72107619e-01 1.10956300e-02 1.39903650e-01 -3.24818909e-01
3.76687914e-01 -5.54499365e-02 7.78398633e-01 2.15880886e-01
2.70605117e-01 -2.98659597e-02 3.65353346e-01 -1.03200316e+00
-9.96717274e-01 -5.42688489e-01 -1.44302592e-01 1.00551414e+00
-1.95545793e-01 2.90389389e-01 -3.70515704e-01 -1.75575554e-01
2.27965117e-01 8.21668804e-01 -7.21999049e-01 -4.48928744e-01
-4.70122159e-01 -8.44397008e-01 3.45947117e-01 6.68054521e-01
9.99327540e-01 -1.77088392e+00 -3.79414141e-01 6.58048570e-01
3.26297730e-01 -8.06526914e-02 -4.19803500e-01 3.40995081e-02
-1.06233060e+00 -6.06069028e-01 -9.50151801e-01 -6.34690046e-01
2.48267828e-03 -1.63703158e-01 8.01064312e-01 3.30065668e-01
4.04612482e-01 -2.56911516e-01 -1.85775757e-01 -1.01978946e+00
5.35593368e-02 1.56068891e-01 -3.95658240e-02 8.42449963e-02
6.85974658e-01 -6.34506881e-01 -5.81001163e-01 -6.74888492e-02
-9.22699332e-01 -1.76550493e-01 8.39394093e-01 9.42443490e-01
4.10492480e-01 4.15001690e-01 9.89142299e-01 -5.06261766e-01
7.68061101e-01 -8.43147397e-01 -1.15980172e+00 4.52358164e-02
-1.28812587e+00 -5.34583181e-02 7.31492996e-01 -5.65037966e-01
-1.02908146e+00 -4.87524062e-01 -3.24396431e-01 -7.96601355e-01
4.80025142e-01 1.18193543e+00 1.91132456e-01 2.18981370e-01
8.93023461e-02 7.79717088e-01 1.97710365e-01 -5.91140985e-01
-2.46520624e-01 5.87945044e-01 2.48691767e-01 1.54416233e-01
5.59293389e-01 -1.35637194e-01 -3.59230876e-01 -2.35087588e-01
-5.51553249e-01 -3.25900712e-03 -3.05361152e-01 -1.70232937e-01
7.47384191e-01 -9.59833026e-01 -6.74901843e-01 1.08616281e+00
-1.10795748e+00 -3.83314162e-01 -1.14069976e-01 8.39076519e-01
-3.20568562e-01 -2.67199844e-01 -1.08871210e+00 -1.00807214e+00
-8.06844652e-01 -8.33208621e-01 1.42019823e-01 6.34871602e-01
3.12188238e-01 -1.19747722e+00 1.24017358e-01 -5.29152334e-01
8.67395937e-01 -7.22431242e-02 6.11468732e-01 -9.61200655e-01
-8.13212514e-01 -3.82386625e-01 -4.42345366e-02 6.93164766e-01
-2.63020638e-02 1.90001503e-01 -7.84188807e-01 -4.39396966e-03
4.43755478e-01 6.56864569e-02 1.15013099e+00 7.29318321e-01
9.83211517e-01 -6.73659921e-01 1.86752439e-01 8.80154729e-01
1.50114560e+00 9.53294039e-01 9.54036713e-01 8.81375253e-01
4.64075387e-01 2.72033691e-01 5.61268449e-01 2.09613979e-01
2.78335929e-01 -1.55029073e-01 2.78034091e-01 -2.98056565e-02
6.77640378e-01 -3.01384032e-01 4.60183293e-01 1.34759879e+00
-3.38281423e-01 1.99812785e-01 -8.58791351e-01 3.37842226e-01
-1.95545161e+00 -1.52421629e+00 8.47208351e-02 1.73352325e+00
7.11158276e-01 4.72387791e-01 5.76142296e-02 6.89840987e-02
5.79162121e-01 2.44225875e-01 -7.57663131e-01 -4.03594106e-01
-1.11556798e-01 8.85754451e-02 8.41327965e-01 2.09461287e-01
-1.04977000e+00 7.75821149e-01 5.69745588e+00 8.61744761e-01
-1.95651650e+00 -6.83645010e-02 8.28070939e-01 -2.53640823e-02
-2.84847885e-01 -2.23500982e-01 -9.08704221e-01 1.20973814e+00
1.22752166e+00 -5.77671468e-01 3.63735050e-01 1.12517905e+00
3.66673946e-01 4.65380639e-01 -3.21357310e-01 7.86378622e-01
-3.84532690e-01 -1.69068682e+00 -4.04896699e-02 1.88074782e-01
5.64038932e-01 3.19277197e-01 3.51062894e-01 7.67626286e-01
2.64633328e-01 -8.79411697e-01 9.67613697e-01 1.38792264e+00
2.89635390e-01 -9.81897414e-01 1.17062807e+00 5.35277426e-01
-1.38726640e+00 -6.00649834e-01 -6.92118883e-01 -4.79547560e-01
-9.84021798e-02 4.96756762e-01 -1.49539903e-01 5.20957232e-01
9.59791005e-01 8.76590371e-01 -2.14257836e-01 1.02166951e+00
6.01459779e-02 6.68843985e-01 -1.43452197e-01 -4.70048696e-01
6.11065030e-01 -4.46758598e-01 5.10547645e-02 9.71441209e-01
8.47136736e-01 2.79097736e-01 4.67985347e-02 7.66879320e-01
1.91155627e-01 1.92640662e-01 -3.46890420e-01 -2.20235676e-01
3.59107614e-01 7.92961061e-01 -4.05799180e-01 -6.74020112e-01
-4.87430960e-01 4.44261104e-01 3.06433719e-02 4.71976101e-01
-6.97677612e-01 -6.86464965e-01 4.66053128e-01 -9.00416635e-03
7.31433630e-01 8.70468095e-02 -4.13644880e-01 -1.14152575e+00
6.41507357e-02 -6.35153830e-01 4.42586660e-01 -8.81983697e-01
-1.48447311e+00 6.90288126e-01 -3.12227488e-01 -1.34077740e+00
-5.70798934e-01 -7.95243502e-01 -1.40857208e+00 1.49739194e+00
-1.85353136e+00 -8.74450326e-01 1.58936739e-01 4.69029218e-01
5.27740180e-01 -7.13047445e-01 3.92312884e-01 2.90869027e-01
-7.00814366e-01 1.39449134e-01 6.08281016e-01 4.43205774e-01
1.03525653e-01 -1.05933464e+00 6.19560897e-01 8.59225631e-01
-4.75264132e-01 6.09996200e-01 1.34773612e-01 -1.02979708e+00
-8.08468223e-01 -1.36190724e+00 8.20552051e-01 3.38054210e-01
1.02364230e+00 1.75521970e-01 -1.44658041e+00 9.58396137e-01
2.07725346e-01 -1.18746065e-01 3.18525285e-01 -2.87738919e-01
-1.32630423e-01 -5.42110622e-01 -5.96033216e-01 3.00949335e-01
2.20896944e-01 -3.73129010e-01 -9.87380147e-01 -8.61783549e-02
9.95288849e-01 -1.08216226e-01 -8.42390597e-01 2.39725262e-01
7.37811387e-01 -9.36485589e-01 6.89672530e-01 -6.50817692e-01
4.62099224e-01 -4.49202694e-02 2.47775510e-01 -1.43039584e+00
-1.04786956e+00 -2.42489412e-01 -1.80895552e-01 1.19750798e+00
5.53024769e-01 -1.23388922e+00 5.03396869e-01 6.30655885e-01
-3.43582481e-01 -7.97767520e-01 -7.35120595e-01 -9.73118603e-01
3.98153245e-01 -2.40362242e-01 1.37072301e+00 1.07627344e+00
-2.03932822e-01 6.77890182e-02 -5.28921723e-01 -6.64692419e-03
-5.33271618e-02 4.00428087e-01 3.46846819e-01 -1.17889285e+00
-1.71540320e-01 -1.07781911e+00 -2.14962214e-01 -7.97330558e-01
8.50301012e-02 -5.49153924e-01 -2.48398200e-01 -1.26760519e+00
-4.05948162e-01 -1.76993817e-01 -9.63475227e-01 4.46149081e-01
-2.90722847e-01 -3.01652521e-01 1.65896729e-01 6.09875083e-01
1.06305137e-01 7.11779058e-01 1.17728043e+00 -1.46453053e-01
-2.99629271e-01 4.55203205e-01 -4.44400340e-01 6.12122357e-01
1.22941124e+00 -1.44854516e-01 -1.43222004e-01 -3.89324903e-01
2.01413453e-01 1.83159769e-01 2.42343917e-02 -7.90919304e-01
3.90451580e-01 -3.94981623e-01 8.84305656e-01 -9.10810471e-01
1.00831069e-01 -8.50920141e-01 4.53822881e-01 1.10616887e+00
2.03752350e-02 6.76743448e-01 2.90159494e-01 1.73541173e-01
-5.63136041e-01 -3.95887643e-01 3.89180720e-01 -5.23349524e-01
-9.31853235e-01 8.81399274e-01 -4.12604868e-01 -3.15532118e-01
8.92037094e-01 -1.92391217e-01 -2.41716668e-01 -3.11753929e-01
-3.66921633e-01 4.62074101e-01 -1.50065184e-01 4.85353857e-01
6.00511670e-01 -1.77624846e+00 -6.54942393e-01 4.86593544e-01
-4.94627535e-01 -2.60157675e-01 4.18802500e-01 5.78843415e-01
-8.28972161e-01 5.19641042e-01 -5.35095692e-01 9.70714688e-02
-4.91196692e-01 8.88802826e-01 8.37360501e-01 -2.86797762e-01
-3.61103535e-01 6.22048497e-01 7.69579932e-02 -1.81275662e-02
-7.41725639e-02 -5.15076756e-01 -5.59477270e-01 3.75480562e-01
9.80665624e-01 3.93070698e-01 -7.99496099e-02 -3.69220108e-01
5.38753681e-02 4.74205554e-01 -3.13317254e-02 -1.29233534e-02
1.81567526e+00 2.35988256e-02 -4.54239815e-01 9.39754009e-01
1.24675751e+00 -3.51162195e-01 -1.14559627e+00 -2.58789003e-01
2.47363567e-01 -2.75232792e-01 3.15126151e-01 -5.62946320e-01
-1.41183567e+00 8.39587212e-01 5.96953154e-01 7.60570228e-01
1.15003479e+00 -6.28658473e-01 1.24194705e+00 2.88612545e-01
-1.11599930e-01 -1.08140886e+00 -2.49591619e-01 9.27164495e-01
9.27347541e-01 -9.81433809e-01 -2.89608032e-01 5.19046605e-01
-3.82759035e-01 1.47256672e+00 6.16777897e-01 -6.33286595e-01
1.31869018e+00 2.40186676e-01 4.06748146e-01 -9.28941146e-02
-9.42430139e-01 -2.14443030e-03 1.70489311e-01 -8.68590400e-02
1.22155540e-01 -1.35141276e-02 -1.24065876e-01 1.18062484e+00
-5.77465713e-01 2.89281070e-01 1.74087107e-01 6.38004720e-01
-7.58030891e-01 -6.61261201e-01 -4.96201932e-01 8.23419809e-01
-8.30761492e-01 -4.07749921e-01 1.17232658e-01 7.01102018e-01
1.36137575e-01 4.45412457e-01 8.13990057e-01 -6.26578629e-01
3.45389247e-01 2.06347436e-01 -3.91472459e-01 -1.71657383e-01
-7.85721302e-01 3.05927277e-01 -4.94015634e-01 -1.09777257e-01
-2.24013608e-02 -5.26803732e-01 -1.21043789e+00 -6.78956926e-01
-3.19163442e-01 3.26780856e-01 4.65958476e-01 9.44283843e-01
2.24513233e-01 6.83514535e-01 1.16039681e+00 -7.33310640e-01
-7.92384505e-01 -1.23313320e+00 -9.45070386e-01 -1.00410476e-01
5.23255169e-01 -4.50860292e-01 -3.56921285e-01 -1.81620061e-01] | [4.455032825469971, 4.234303951263428] |
0e94297b-0d66-4ed1-ac49-5ee20c4151d4 | a-benchmark-for-generalizable-and | 2201.05793 | null | https://arxiv.org/abs/2201.05793v1 | https://arxiv.org/pdf/2201.05793v1.pdf | A Benchmark for Generalizable and Interpretable Temporal Question Answering over Knowledge Bases | Knowledge Base Question Answering (KBQA) tasks that involve complex reasoning are emerging as an important research direction. However, most existing KBQA datasets focus primarily on generic multi-hop reasoning over explicit facts, largely ignoring other reasoning types such as temporal, spatial, and taxonomic reasoning. In this paper, we present a benchmark dataset for temporal reasoning, TempQA-WD, to encourage research in extending the present approaches to target a more challenging set of complex reasoning tasks. Specifically, our benchmark is a temporal question answering dataset with the following advantages: (a) it is based on Wikidata, which is the most frequently curated, openly available knowledge base, (b) it includes intermediate sparql queries to facilitate the evaluation of semantic parsing based approaches for KBQA, and (c) it generalizes to multiple knowledge bases: Freebase and Wikidata. The TempQA-WD dataset is available at https://github.com/IBM/tempqa-wd. | ['L Venkata Subramaniam', 'Francois Luus', 'Ryan Riegel', 'Guilherme Lima', 'Alexander Gray', 'Salim Roukos', 'Rosario Uceda-Sosa', 'Maria Chang', 'Sairam Gurajada', 'Srinivas Ravishankar', 'Dinesh Khandelwal', 'G P Shrivatsa Bhargav', 'Achille Fokoue', 'Dinesh Garg', 'Saswati Dana', 'Cezar Pendus', 'Santosh Srivastava', 'Young-suk Lee', 'Nandana Mihindukulasooriya', 'Ibrahim Abdelaziz', 'Pavan Kapanipathi', 'Shajith Ikbal', 'Hima Karanam', 'Udit Sharma', 'Sumit Neelam'] | 2022-01-15 | null | null | null | null | ['knowledge-base-question-answering'] | ['natural-language-processing'] | [-5.51776350e-01 2.79888064e-01 -4.43152666e-01 -4.15750831e-01
-9.11042333e-01 -9.07773614e-01 4.96004164e-01 4.17902559e-01
-2.30679929e-01 1.01973176e+00 5.44476211e-01 -4.03972715e-01
-5.44192016e-01 -1.26494646e+00 -6.74714029e-01 -1.60433531e-01
1.60302855e-02 9.82446492e-01 1.04435825e+00 -7.62192607e-01
-4.10601906e-02 2.56379664e-01 -1.52207804e+00 4.91094857e-01
9.06274617e-01 9.98122931e-01 -1.75962254e-01 3.92506927e-01
-3.79904956e-01 1.50279117e+00 -4.36242729e-01 -8.69135916e-01
-1.82674393e-01 -2.36884356e-01 -1.71925068e+00 -8.14593434e-01
4.09059376e-01 -1.25086054e-01 -5.03331661e-01 7.12166488e-01
3.01968992e-01 4.00266349e-01 1.02634452e-01 -1.49332130e+00
-8.09337318e-01 5.99571884e-01 1.38987660e-01 6.00690305e-01
7.15580523e-01 -1.30745163e-02 1.44639802e+00 -2.34943733e-01
1.11320651e+00 1.28028405e+00 7.32001781e-01 5.37829518e-01
-6.21711910e-01 -2.78273165e-01 1.39284849e-01 1.00223637e+00
-1.28217149e+00 -1.73393473e-01 4.52831358e-01 -2.97325939e-01
1.32453430e+00 4.75192577e-01 5.12230277e-01 7.76848435e-01
-2.75315285e-01 8.03219616e-01 9.14567411e-01 -3.61083746e-01
3.70732635e-01 -1.99259400e-01 5.41895270e-01 9.00069594e-01
-3.35269943e-02 -1.36903822e-01 -5.95081210e-01 -2.40700737e-01
4.10528749e-01 -4.35136706e-01 -2.30743334e-01 -4.16862607e-01
-1.18604338e+00 7.69328713e-01 5.89075685e-01 4.54513788e-01
-3.53546232e-01 3.52726161e-01 4.79769588e-01 3.30229551e-01
2.06694856e-01 5.35381615e-01 -8.01524520e-01 -4.97084141e-01
-2.80091047e-01 7.77335107e-01 1.15958619e+00 9.87330019e-01
7.24355161e-01 -6.12089872e-01 -1.13786325e-01 7.59269059e-01
4.31539156e-02 2.85979629e-01 2.65594363e-01 -1.37255323e+00
7.49303699e-01 8.96661580e-01 3.55360717e-01 -6.38263464e-01
-6.00331604e-01 2.83391744e-01 1.04719914e-01 -3.60810846e-01
5.20634115e-01 1.60818264e-01 -6.19214237e-01 1.72165239e+00
6.24174416e-01 -6.95770904e-02 4.61764514e-01 8.68101418e-01
1.34673870e+00 6.06543064e-01 2.56730348e-01 2.03933287e-02
1.70495605e+00 -1.14921117e+00 -8.34376574e-01 -1.37070864e-01
8.82423162e-01 -2.40714684e-01 1.13412941e+00 8.39018077e-02
-1.04484832e+00 2.27291249e-02 -7.82421291e-01 -6.93191051e-01
-1.07658184e+00 -5.26807308e-01 1.07044160e+00 2.50046581e-01
-1.03316355e+00 3.60623784e-02 -6.90501094e-01 -8.86787176e-01
2.42659435e-01 -1.79784656e-01 -2.74055183e-01 -4.95202869e-01
-1.86719072e+00 1.37564623e+00 8.75245631e-01 -2.36581117e-01
-6.32835567e-01 -9.73752856e-01 -8.34864199e-01 -1.00959398e-01
1.00755227e+00 -8.72766137e-01 1.57964647e+00 -2.47299612e-01
-1.05675459e+00 7.75538504e-01 -1.81786895e-01 -4.91152942e-01
2.64253646e-01 -2.19133645e-01 -8.93516123e-01 3.98824811e-01
5.04437327e-01 5.63377619e-01 -8.84137452e-02 -7.56356061e-01
-6.73419476e-01 -5.32457352e-01 8.56386721e-01 2.58608282e-01
-5.18579073e-02 8.60100538e-02 -7.46430814e-01 -3.34603369e-01
-8.58730674e-02 -6.40052438e-01 1.64755449e-01 -6.85335100e-02
-1.64510161e-01 -6.15448236e-01 7.15230584e-01 -7.82364666e-01
1.32798123e+00 -1.85052872e+00 8.45801681e-02 -2.19695508e-01
-7.96846822e-02 -1.91294011e-02 1.14256576e-01 8.16783845e-01
7.95365125e-02 7.09621683e-02 -2.09722623e-01 4.09047931e-01
1.31057680e-01 5.83359301e-01 -6.27278209e-01 -2.26245508e-01
8.44731331e-02 1.29861188e+00 -1.25889230e+00 -7.60103881e-01
-1.60279334e-01 -1.62944168e-01 -5.12922585e-01 -2.17978671e-01
-1.12580442e+00 7.33466877e-04 -8.06766510e-01 8.91633272e-01
2.00197056e-01 -3.93093139e-01 2.46908516e-01 -3.69541407e-01
-1.50937647e-01 6.68065190e-01 -6.44733310e-01 1.98449314e+00
-2.10425645e-01 5.20462096e-01 -4.17504013e-01 -6.81010067e-01
3.81075859e-01 6.08906746e-01 5.77661753e-01 -9.89292860e-01
-3.25840205e-01 2.37502038e-01 -2.97437876e-01 -8.71982872e-01
6.67901278e-01 -1.75932989e-01 -3.53371233e-01 3.92353714e-01
1.44129127e-01 -4.48887467e-01 8.27898026e-01 4.48775947e-01
1.49688089e+00 3.71575952e-01 3.82286638e-01 -2.26484880e-01
3.44412327e-01 7.99213827e-01 6.61153316e-01 4.03108597e-01
-3.85964543e-01 -4.54648063e-02 4.90299851e-01 -7.58496940e-01
-6.46809995e-01 -1.38973129e+00 -2.83104807e-01 1.13094461e+00
4.13748324e-01 -6.85371339e-01 -3.31404954e-01 -7.96333551e-01
1.97466716e-01 1.00181985e+00 -5.66662788e-01 1.69702500e-01
-5.88483572e-01 -4.52231437e-01 9.89837289e-01 6.74658954e-01
7.70455301e-01 -1.22695971e+00 -5.67549050e-01 2.33860523e-01
-1.00982535e+00 -1.21416938e+00 2.22718921e-02 -1.42891690e-01
-8.09334040e-01 -1.62815106e+00 -1.19398907e-01 -4.62043792e-01
4.85143773e-02 -7.24335015e-02 1.50643742e+00 -1.75218582e-02
-2.54894376e-01 8.55338156e-01 -6.47818327e-01 -5.60991108e-01
-1.39701804e-02 1.79229490e-02 -3.10694188e-01 -8.07657480e-01
6.11726701e-01 -3.97785485e-01 -3.13942313e-01 4.54453766e-01
-9.58659291e-01 -1.55743897e-01 -7.23283291e-02 4.00853157e-01
5.13067007e-01 4.98753935e-01 6.99244618e-01 -9.36869979e-01
5.82559526e-01 -8.11825871e-01 -5.34142971e-01 7.74752080e-01
-4.04230714e-01 9.99511108e-02 2.69627064e-01 7.75679648e-02
-1.62955344e+00 -6.88001275e-01 -1.81007192e-01 1.21943630e-01
-6.23875558e-02 9.33607697e-01 1.24137131e-02 3.84144872e-01
9.47064579e-01 -1.45605132e-02 -4.14277047e-01 -3.03890765e-01
8.18741500e-01 9.63410884e-02 6.80083334e-01 -1.16052008e+00
6.10605896e-01 6.14180923e-01 -1.62758976e-01 -4.18596834e-01
-1.19159782e+00 -5.77935696e-01 -4.93730009e-01 -1.34145051e-01
1.06200325e+00 -8.36945593e-01 -8.34606111e-01 8.10421556e-02
-1.02417457e+00 -6.66476130e-01 -4.30992424e-01 2.17423126e-01
-8.09877813e-01 2.24041462e-01 -7.67641842e-01 -4.89639580e-01
-2.55095214e-01 -5.25302410e-01 7.93120265e-01 1.62741929e-01
-3.88062209e-01 -1.22606468e+00 3.95167261e-01 1.04564655e+00
2.77627140e-01 3.52776080e-01 1.35086131e+00 -5.08180201e-01
-7.20817745e-01 5.03109433e-02 -2.00006098e-01 -2.35718936e-01
8.28257725e-02 -6.41174708e-03 -6.80177689e-01 2.25138858e-01
-5.37742138e-01 -8.34583640e-01 5.41436255e-01 2.91331317e-02
9.16848540e-01 -3.34681004e-01 -3.38098198e-01 7.67479166e-02
1.32899106e+00 4.46583301e-01 8.30075979e-01 7.90630519e-01
1.77414015e-01 6.33264542e-01 1.18303704e+00 1.99462399e-01
1.17139459e+00 7.46401489e-01 5.04891336e-01 4.35768545e-01
-4.08313386e-02 -1.85864732e-01 -5.63555174e-02 6.21447265e-01
-2.14124322e-01 -2.32037768e-01 -1.40736747e+00 9.98392940e-01
-2.37093711e+00 -1.22880542e+00 -5.95494248e-02 1.56732142e+00
1.14551592e+00 -2.24579230e-01 2.26603165e-01 -1.81925833e-01
4.01778579e-01 1.87122360e-01 -7.14275479e-01 -2.03532532e-01
-2.01077387e-01 2.18316451e-01 6.26641661e-02 3.34940732e-01
-9.48986709e-01 1.24381113e+00 5.97099161e+00 5.42417169e-01
-5.55879295e-01 3.08830827e-01 -8.24218839e-02 -1.43121421e-01
-4.30233508e-01 1.51632667e-01 -6.03636801e-01 1.47022441e-01
9.09203529e-01 -4.77132916e-01 5.29133499e-01 7.70733893e-01
-3.24798077e-01 -2.61801392e-01 -1.06326675e+00 5.26956260e-01
-2.60558784e-01 -1.62240708e+00 2.76436731e-02 -4.52898771e-01
3.64370316e-01 4.50583920e-02 -2.81792194e-01 7.60476172e-01
7.58901894e-01 -6.82632148e-01 8.08052957e-01 6.04337275e-01
3.83080244e-01 -5.19946277e-01 6.94693923e-01 1.21887103e-01
-1.25895596e+00 -2.32341483e-01 -1.12473480e-01 2.94579193e-02
5.02678335e-01 4.37659562e-01 -6.01441860e-01 1.10213602e+00
1.34997952e+00 5.67128897e-01 -5.37458420e-01 8.80749404e-01
-3.67079407e-01 4.24812824e-01 -2.40979686e-01 -3.11029498e-02
2.26104856e-01 4.19527851e-03 2.07544968e-01 8.31057370e-01
1.75887048e-01 7.09958494e-01 -2.14875445e-01 6.78250968e-01
-2.26716790e-02 2.85366159e-02 -5.22124410e-01 -2.97447979e-01
8.02686632e-01 8.26017141e-01 -4.56391752e-01 -3.10026139e-01
-5.23621798e-01 6.74466968e-01 6.75980628e-01 3.32395405e-01
-1.04851031e+00 -2.52160609e-01 8.44547808e-01 -2.59508379e-02
2.80749977e-01 3.90829407e-02 1.73155904e-01 -1.13596928e+00
2.73841053e-01 -7.80099869e-01 1.40960145e+00 -1.45503688e+00
-1.35508931e+00 4.39631879e-01 4.91117477e-01 -5.37158787e-01
-2.70613253e-01 -5.10317087e-01 -2.40922824e-01 4.92556632e-01
-1.67433619e+00 -1.22205293e+00 -6.17467523e-01 9.63667691e-01
1.31990731e-01 1.89553186e-01 9.39590216e-01 4.77472723e-01
-4.43483651e-01 1.91394631e-02 -4.45547163e-01 4.37473580e-02
6.96432590e-01 -1.19671333e+00 2.95175195e-01 5.88302672e-01
-5.16625121e-02 6.90109193e-01 5.65285325e-01 -6.34442687e-01
-1.31312013e+00 -1.13902271e+00 9.02692199e-01 -9.52367961e-01
1.12746203e+00 1.87329233e-01 -9.94118929e-01 1.30705106e+00
1.18934654e-01 1.72054470e-02 6.49879694e-01 4.57683265e-01
-8.22400630e-01 -1.67517856e-01 -1.24875104e+00 7.28037238e-01
1.37475383e+00 -7.04932809e-01 -1.04159093e+00 5.96981943e-01
1.09868324e+00 -6.34775579e-01 -1.49629021e+00 4.98867810e-01
3.12105626e-01 -7.61173010e-01 1.00529468e+00 -9.22262490e-01
3.83671463e-01 -5.44604599e-01 -3.82999957e-01 -9.41099226e-01
-2.42564172e-01 -1.15262650e-01 -5.99683464e-01 1.26813936e+00
6.19648278e-01 -7.77494371e-01 6.12408638e-01 7.80731559e-01
-1.64325312e-01 -6.30293608e-01 -1.11375940e+00 -9.18243229e-01
2.89045740e-02 -5.04645288e-01 8.11805308e-01 1.30935836e+00
3.99634749e-01 2.84295052e-01 5.22778742e-02 3.40086311e-01
2.00302377e-01 4.94648576e-01 4.40381289e-01 -1.20287919e+00
-9.64149758e-02 -2.36754835e-01 -3.08323145e-01 -8.25343251e-01
1.07873037e-01 -8.32430422e-01 -8.44563246e-02 -2.00468421e+00
-1.61465168e-01 -7.27047741e-01 -2.28970149e-03 1.04052806e+00
-3.77001688e-02 -6.10067248e-02 -2.10561037e-01 3.36209610e-02
-1.00073266e+00 3.81667376e-01 1.12786293e+00 -1.24523781e-01
1.32289603e-01 -3.09512913e-01 -7.93169320e-01 4.38355565e-01
8.34884107e-01 -1.89822137e-01 -8.17825258e-01 -6.05363548e-01
7.31604218e-01 3.51305366e-01 3.68444175e-01 -7.01604903e-01
4.16685790e-01 -5.20700753e-01 -3.84104043e-01 -5.51425278e-01
5.52211761e-01 -6.86708391e-01 3.52977157e-01 2.56366700e-01
-1.14601158e-01 2.75535583e-01 5.17320633e-01 4.23054039e-01
-5.66283643e-01 -2.35489652e-01 3.97932321e-01 -3.39574963e-01
-1.48226309e+00 3.10674548e-01 9.02704820e-02 7.23871768e-01
1.26587152e+00 1.99869752e-01 -1.12813795e+00 -2.85422981e-01
-6.61220253e-01 6.34622574e-01 2.58759975e-01 6.26164734e-01
3.18929315e-01 -1.31122398e+00 -3.33383143e-01 -7.95021534e-01
7.74790168e-01 1.77682415e-01 3.53857517e-01 7.45791435e-01
-8.06478083e-01 9.05965447e-01 -8.98972154e-02 -9.43545625e-02
-9.40763533e-01 9.01968896e-01 3.97772789e-01 -3.43220204e-01
-7.48550713e-01 7.60208368e-01 -2.36117423e-01 -8.36950481e-01
7.09664524e-02 -2.81124175e-01 -1.36426046e-01 1.45562962e-02
4.08161640e-01 4.53597903e-01 2.02371210e-01 -3.05883557e-01
-6.59505308e-01 1.69297442e-01 7.32123032e-02 -7.74261206e-02
1.35680425e+00 1.04911804e-01 -6.24586999e-01 5.06530643e-01
6.20831013e-01 -1.80780381e-01 -5.53638041e-01 -5.22931218e-01
5.06996393e-01 -1.84388861e-01 -2.43651375e-01 -1.28964448e+00
-7.25984931e-01 2.77749598e-01 -1.03796043e-01 3.99634451e-01
1.09878659e+00 5.65684915e-01 7.47789204e-01 7.88361788e-01
7.68123329e-01 -1.04991484e+00 7.80038387e-02 7.08113730e-01
1.09316087e+00 -8.91424537e-01 1.99057553e-02 -5.13643801e-01
-5.69335938e-01 8.75075281e-01 1.00335693e+00 6.26359105e-01
5.35166979e-01 -3.76509428e-02 1.13291748e-01 -8.79684091e-01
-1.18026412e+00 -5.72356105e-01 -2.04879437e-02 5.37702560e-01
3.22713584e-01 -1.27792448e-01 -2.71788180e-01 5.50754309e-01
-3.92721832e-01 3.12120736e-01 2.23915517e-01 1.19165850e+00
-1.99687749e-01 -9.92135048e-01 -1.37870535e-01 3.06409746e-01
-1.88643351e-01 -1.46242604e-01 -4.26463604e-01 1.19506025e+00
2.05195509e-02 9.76929784e-01 -1.04646955e-03 1.67207137e-01
5.90642154e-01 3.47645968e-01 6.66218221e-01 -4.97672021e-01
-3.08915943e-01 -1.12945771e+00 7.77250051e-01 -7.67701924e-01
-5.64266980e-01 -6.15081668e-01 -1.62298417e+00 -4.46111649e-01
1.41574414e-02 5.19838572e-01 3.33179146e-01 9.78692532e-01
3.60428244e-01 5.25773048e-01 -3.70918393e-01 2.42008984e-01
-1.29451007e-02 -6.13316715e-01 -3.58643323e-01 4.58987027e-01
-1.10956170e-01 -8.60456169e-01 6.11521117e-02 -4.51654941e-02] | [10.216981887817383, 7.967022895812988] |
ea056d51-ca20-4970-a1ba-d1140f7a5356 | you-only-evaluate-once-a-simple-baseline | 2110.02304 | null | https://arxiv.org/abs/2110.02304v1 | https://arxiv.org/pdf/2110.02304v1.pdf | You Only Evaluate Once: a Simple Baseline Algorithm for Offline RL | The goal of offline reinforcement learning (RL) is to find an optimal policy given prerecorded trajectories. Many current approaches customize existing off-policy RL algorithms, especially actor-critic algorithms in which policy evaluation and improvement are iterated. However, the convergence of such approaches is not guaranteed due to the use of complex non-linear function approximation and an intertwined optimization process. By contrast, we propose a simple baseline algorithm for offline RL that only performs the policy evaluation step once so that the algorithm does not require complex stabilization schemes. Since the proposed algorithm is not likely to converge to an optimal policy, it is an appropriate baseline for actor-critic algorithms that ought to be outperformed if there is indeed value in iterative optimization in the offline setting. Surprisingly, we empirically find that the proposed algorithm exhibits competitive and sometimes even state-of-the-art performance in a subset of the D4RL offline RL benchmark. This result suggests that future work is needed to fully exploit the potential advantages of iterative optimization in order to justify the reduced stability of such methods. | ['Scott Niekum', 'Wonjoon Goo'] | 2021-10-05 | null | null | null | null | ['d4rl'] | ['robots'] | [-1.63733736e-01 1.70101151e-01 -5.96945524e-01 -8.90836269e-02
-7.98043489e-01 -7.86274612e-01 8.50888908e-01 2.01841280e-01
-7.39486098e-01 1.03456986e+00 1.04974605e-01 -6.68713868e-01
-1.52690783e-01 -3.07891697e-01 -6.21766627e-01 -7.16296554e-01
-3.45924944e-01 3.88513863e-01 9.99684930e-02 -4.35347676e-01
3.25550318e-01 5.82496285e-01 -1.37190771e+00 -2.90257305e-01
7.25819111e-01 8.89273822e-01 2.91088875e-02 7.51564980e-01
1.57752663e-01 9.10579443e-01 -3.97617668e-01 -8.00210387e-02
5.84909856e-01 -7.71587431e-01 -6.55464113e-01 7.15729222e-02
1.65237039e-02 -6.86461091e-01 -8.30128640e-02 8.86627197e-01
6.11203074e-01 5.08191764e-01 3.50071341e-01 -9.20851946e-01
-1.55552194e-01 4.03581351e-01 -3.80122811e-01 2.17261061e-01
4.88234758e-01 4.93048638e-01 1.04212618e+00 -3.51462722e-01
4.53402311e-01 1.28936374e+00 4.51988727e-01 4.98799264e-01
-1.20765483e+00 -3.28994811e-01 6.11771762e-01 -1.15165263e-01
-7.84926772e-01 -4.91695225e-01 5.81342936e-01 1.19494302e-02
9.74823654e-01 2.72382563e-03 9.75853026e-01 1.08706176e+00
-1.54777430e-02 9.15281475e-01 1.37184954e+00 -4.26614493e-01
5.65578580e-01 1.09667554e-01 -3.37084591e-01 6.96548820e-01
1.59484938e-01 7.80934393e-01 -2.16878697e-01 -2.35728920e-01
7.93544710e-01 -4.01781470e-01 2.08373256e-02 -5.69889367e-01
-1.15998650e+00 7.82667875e-01 2.09838301e-01 1.85471117e-01
-6.42155886e-01 4.22318488e-01 5.35273910e-01 6.42268836e-01
2.37923756e-01 7.53904104e-01 -3.94494951e-01 -7.76830673e-01
-9.54030812e-01 7.88912833e-01 8.82351756e-01 4.76948738e-01
5.01024425e-01 3.37275922e-01 -1.66249245e-01 4.88232553e-01
2.16367245e-01 2.46467561e-01 4.61659789e-01 -1.56400633e+00
2.94034183e-01 2.59363830e-01 8.41156125e-01 -6.34049296e-01
-3.54606241e-01 -4.74185109e-01 -1.85396582e-01 7.12907314e-01
8.84349644e-01 -4.56917107e-01 -4.51748073e-01 1.66171002e+00
4.52166468e-01 2.52118528e-01 2.78675169e-01 9.53403890e-01
-8.56348649e-02 4.01186377e-01 -1.12902291e-01 -6.74936354e-01
5.75523794e-01 -9.66619670e-01 -7.12219834e-01 -1.17944323e-01
5.10014594e-01 -5.86627007e-01 1.28909421e+00 4.72009212e-01
-1.17636025e+00 -3.20536137e-01 -1.00664341e+00 4.79313076e-01
-2.35834531e-03 3.75020318e-02 6.31446481e-01 4.69695061e-01
-8.50950003e-01 1.03222764e+00 -1.04302967e+00 -3.39331836e-01
-1.64103694e-02 5.16981065e-01 1.46403000e-01 4.38852042e-01
-1.07908142e+00 1.20039797e+00 4.15220022e-01 5.53450808e-02
-1.19078171e+00 -3.73708695e-01 -4.98271555e-01 -2.52402186e-01
6.91232085e-01 -2.68799424e-01 1.91980410e+00 -1.30969465e+00
-2.31947112e+00 3.60329062e-01 -1.35317698e-01 -8.67412448e-01
8.71718168e-01 -3.35115314e-01 -1.37961537e-01 1.13702849e-01
-2.47324377e-01 4.28769320e-01 9.41226304e-01 -1.35752022e+00
-7.41235614e-01 5.53188920e-02 4.56159711e-01 5.35637856e-01
-4.73593064e-02 -4.76700328e-02 -1.01068512e-01 -3.64041775e-01
-3.89337927e-01 -1.11214960e+00 -5.75787127e-01 -2.79664665e-01
5.79749011e-02 -2.75495678e-01 7.80694902e-01 -3.30609471e-01
1.31187129e+00 -1.90603101e+00 -4.08669375e-02 2.04410017e-01
-2.47480661e-01 3.57339621e-01 -8.73064697e-02 7.17875123e-01
1.12158142e-01 -2.14689195e-01 1.10075995e-01 -9.55443755e-02
1.54798701e-01 3.02615792e-01 -6.13218725e-01 8.83549273e-01
-1.09386757e-01 9.01831865e-01 -1.29018939e+00 -2.60770589e-01
3.05112392e-01 -4.64457795e-02 -5.92664838e-01 2.42704839e-01
-6.38007760e-01 8.57807934e-01 -6.98618650e-01 4.47445393e-01
-2.39625238e-02 -1.35906905e-01 4.23480213e-01 3.31481278e-01
-5.30538440e-01 4.50136274e-01 -1.27587605e+00 1.52224398e+00
-2.58152813e-01 5.19846857e-01 1.32510334e-01 -1.31975865e+00
7.93830216e-01 5.13916075e-01 9.95406270e-01 -8.04002941e-01
1.31361902e-01 3.47049475e-01 2.16423392e-01 -2.90288508e-01
4.64837700e-01 -3.23660411e-02 1.90437570e-01 6.72737181e-01
-2.09723324e-01 -1.64316386e-01 3.46536756e-01 -1.31058767e-01
1.06146741e+00 7.33164072e-01 4.74173754e-01 -3.04043800e-01
5.10235786e-01 2.16152295e-01 4.33945775e-01 1.01419568e+00
-5.30296624e-01 -3.89631987e-02 5.97649932e-01 -4.33281004e-01
-1.14320886e+00 -6.74304545e-01 1.30449235e-01 1.16233194e+00
4.77257483e-02 -1.77751869e-01 -5.35682917e-01 -9.20309067e-01
1.64424464e-01 8.49869490e-01 -4.21019286e-01 5.20644337e-02
-7.52701759e-01 -4.75183219e-01 4.84480292e-01 2.32231855e-01
3.55221331e-01 -1.19436038e+00 -1.06575108e+00 6.61966085e-01
3.16636115e-01 -7.88456976e-01 -3.86962056e-01 1.31705940e-01
-1.12203741e+00 -1.03015220e+00 -7.18846679e-01 -3.03549677e-01
4.86357659e-01 3.07238698e-02 8.48563910e-01 -9.71908569e-02
4.63831544e-01 8.59448731e-01 -2.95671254e-01 -3.63620728e-01
-5.77587068e-01 1.25542069e-02 3.99149686e-01 -7.93699622e-02
-7.78120086e-02 -4.62729692e-01 -7.73002684e-01 3.96344692e-01
-4.54828024e-01 -3.31142813e-01 4.79134262e-01 9.08895493e-01
5.96505105e-01 -1.78287953e-01 8.65607858e-01 -5.37303507e-01
1.09536862e+00 -3.44623089e-01 -1.03951216e+00 2.99415261e-01
-1.00859666e+00 3.58634651e-01 1.01760304e+00 -7.06994474e-01
-1.11384988e+00 1.58650935e-01 -1.68021265e-02 -5.13615847e-01
2.37817150e-02 3.08384478e-01 5.59861362e-01 -2.02562094e-01
6.78340793e-01 2.55475700e-01 3.43581229e-01 -2.78884530e-01
4.07699436e-01 2.50188053e-01 2.38699436e-01 -9.25312102e-01
5.91306686e-01 3.39042366e-01 -2.96692108e-03 -6.02783084e-01
-7.74132013e-01 -3.35165769e-01 -2.25349620e-01 -4.94691908e-01
3.55751038e-01 -6.58063531e-01 -1.12403286e+00 -2.35441178e-02
-7.69705057e-01 -9.78348374e-01 -5.76415360e-01 6.28592789e-01
-1.12587643e+00 4.60269958e-01 -3.07237208e-01 -1.23303199e+00
-1.20957583e-01 -1.25944364e+00 6.99484587e-01 1.97232649e-01
-2.80510068e-01 -1.10477066e+00 4.75874960e-01 -1.87158704e-01
3.33373070e-01 2.50708312e-01 4.47896302e-01 -5.75656414e-01
-3.42487663e-01 8.20085313e-03 2.85226136e-01 2.63592064e-01
1.17373422e-01 8.88187904e-03 -6.55012727e-01 -8.14887583e-01
-1.58710152e-01 -8.08002532e-01 4.54929590e-01 4.28327680e-01
9.26039100e-01 -4.12678510e-01 1.63789168e-02 3.19858760e-01
1.36310458e+00 6.10214829e-01 2.81378716e-01 6.98704898e-01
1.36605352e-01 4.10926968e-01 1.08619618e+00 6.77746356e-01
2.39006460e-01 6.47711158e-01 4.75876600e-01 6.53906018e-02
2.71348029e-01 -4.98724282e-01 7.85789073e-01 4.60824907e-01
-2.68613189e-01 -7.68966153e-02 -7.07484782e-01 4.49228674e-01
-2.14802170e+00 -1.17942023e+00 3.82426500e-01 2.30990744e+00
8.95657957e-01 1.70502171e-01 5.80466092e-01 -3.55439968e-02
2.35745683e-01 2.45354578e-01 -9.81874287e-01 -7.37048090e-01
2.32852399e-01 2.11528674e-01 7.27816164e-01 4.77661461e-01
-8.43294084e-01 1.00002837e+00 7.21958971e+00 5.96397221e-01
-1.21601331e+00 -1.04564972e-01 4.40657496e-01 -1.57982618e-01
2.11485084e-02 2.84889966e-01 -7.37213850e-01 2.96587944e-01
1.05170405e+00 -2.29943648e-01 1.01725304e+00 8.36619437e-01
8.49556327e-01 -4.14951652e-01 -9.23428893e-01 6.55686021e-01
-5.12315929e-01 -1.05962944e+00 -5.82838714e-01 1.64084747e-01
8.67094100e-01 -6.30138367e-02 1.12979505e-02 6.06232524e-01
6.89509392e-01 -9.17818844e-01 7.27645576e-01 4.68415648e-01
4.52457577e-01 -9.24344063e-01 1.65379360e-01 6.59022629e-01
-9.60308492e-01 -4.07762259e-01 -2.61987984e-01 -2.68719107e-01
6.71187714e-02 -1.12806365e-01 -8.51518154e-01 3.09593558e-01
3.07836235e-01 6.96989059e-01 -1.32395715e-01 1.06430185e+00
-2.59021282e-01 8.75743568e-01 -4.08745140e-01 -3.38711292e-01
9.74616230e-01 -3.89636159e-01 6.92235529e-01 9.86117184e-01
1.40753880e-01 3.45706232e-02 4.89751130e-01 3.40563893e-01
3.12709570e-01 2.61317164e-01 -8.54237378e-01 -3.12200755e-01
1.97819218e-01 9.59473491e-01 -6.49212301e-01 -2.06657022e-01
-5.23290038e-01 6.24719679e-01 5.28254688e-01 5.96155822e-01
-8.93256366e-01 6.48769140e-02 6.74274325e-01 -9.14084390e-02
4.24910188e-01 -6.41801715e-01 -7.07491413e-02 -9.83608365e-01
-2.07902327e-01 -1.28551757e+00 3.66482645e-01 -1.95473105e-01
-8.57154429e-01 1.84942544e-01 5.51638640e-02 -1.34924626e+00
-7.82999337e-01 -2.37019300e-01 -5.29928148e-01 3.34343374e-01
-1.46802294e+00 -5.03244817e-01 2.48674393e-01 5.73603570e-01
6.69986427e-01 -2.10835397e-01 6.45011842e-01 -1.19336687e-01
-4.54957455e-01 5.02414882e-01 5.47431469e-01 -3.38437080e-01
7.26183474e-01 -1.39042711e+00 -1.68078374e-02 7.47603893e-01
5.46787828e-02 4.09542382e-01 1.03668880e+00 -6.56863332e-01
-1.57018852e+00 -7.75482774e-01 1.97816968e-01 -1.66292474e-01
8.68921041e-01 5.84259033e-02 -5.52432239e-01 7.48144746e-01
4.36235011e-01 -8.99584368e-02 9.72592756e-02 -4.26100120e-02
3.61939013e-01 -1.77228197e-01 -8.42705548e-01 9.19064224e-01
7.49036968e-01 -2.28575021e-01 -4.70731646e-01 2.42620707e-01
4.13698137e-01 -6.11304462e-01 -8.45603287e-01 3.07131588e-01
6.77040756e-01 -9.00053203e-01 8.59831452e-01 -7.67060399e-01
5.01266718e-02 -2.16778949e-01 1.73039183e-01 -1.40580106e+00
1.76689383e-02 -1.31766331e+00 -5.65500438e-01 7.85937250e-01
4.06912982e-01 -7.78955042e-01 7.77042866e-01 4.92251426e-01
1.76686868e-01 -1.11509037e+00 -8.65872324e-01 -1.05312145e+00
1.09738767e-01 -4.63430792e-01 3.65556300e-01 7.44681835e-01
4.71894182e-02 3.15554589e-02 -6.32048130e-01 -1.00788437e-01
6.09414697e-01 2.25915417e-01 9.24071431e-01 -6.29605591e-01
-6.34977818e-01 -5.40093780e-01 3.05063397e-01 -1.48325169e+00
4.43486661e-01 -5.50515652e-01 -2.08192109e-03 -1.17071414e+00
-4.04093444e-01 -6.99446619e-01 -4.00556773e-01 3.18342060e-01
-2.55841520e-02 -3.66851747e-01 1.72329098e-01 1.66581735e-01
-8.56232405e-01 8.25780272e-01 1.59732425e+00 2.67785698e-01
-8.13019156e-01 4.08718079e-01 -4.44014847e-01 5.70563257e-01
1.15695500e+00 -4.44542706e-01 -7.11721182e-01 3.24717984e-02
1.03226714e-01 4.99921918e-01 2.08284393e-01 -6.92576051e-01
2.18811288e-01 -6.64864540e-01 5.05615734e-02 -3.13130587e-01
2.06809178e-01 -7.69342184e-01 -1.43662602e-01 6.06070399e-01
-4.71459746e-01 3.72163534e-01 1.33976668e-01 7.19122767e-01
-6.32634908e-02 -2.90370405e-01 8.25369000e-01 -3.22339207e-01
-6.36344254e-01 2.68331230e-01 -5.93642533e-01 1.44744664e-01
1.07947707e+00 -1.98972970e-01 5.27823344e-02 -7.04135716e-01
-5.30496240e-01 3.28502774e-01 4.69926059e-01 2.28114873e-01
3.60652804e-01 -1.08531392e+00 -4.03856397e-01 -2.04182968e-01
-3.60454232e-01 -4.47780579e-01 -4.84986067e-01 9.49946046e-01
-3.25865060e-01 5.79779565e-01 -5.26984818e-02 -3.33687544e-01
-7.73152232e-01 5.38379848e-01 5.47633827e-01 -7.33049452e-01
-8.08441758e-01 2.54981071e-01 -3.84406924e-01 -2.51923591e-01
5.31928182e-01 -2.18931958e-01 -1.47417441e-01 -1.13245271e-01
2.58272558e-01 5.65884769e-01 -2.37222061e-01 -2.21453011e-01
-5.11793047e-02 2.20751643e-01 6.35786057e-02 -6.08197868e-01
1.18019915e+00 -6.58032438e-03 3.02465767e-01 4.39532012e-01
7.04770327e-01 1.38311628e-02 -1.88223755e+00 -3.70181128e-02
2.93418825e-01 -3.88285428e-01 1.26261473e-01 -7.34835625e-01
-7.52067208e-01 2.96112239e-01 5.01240015e-01 1.75322130e-01
9.97406602e-01 -6.30291402e-01 5.66522360e-01 7.83758402e-01
5.48001528e-01 -1.49271142e+00 1.45512089e-01 6.59025908e-01
8.84891272e-01 -1.23224652e+00 2.89728045e-01 5.25552928e-01
-7.90819407e-01 1.15639043e+00 5.47946692e-01 -5.21159887e-01
3.61016780e-01 1.03802621e-01 2.67359111e-02 8.60286355e-02
-1.03858805e+00 -5.26245773e-01 -1.61798485e-02 1.72332108e-01
2.01915547e-01 -3.88401039e-02 -6.47034109e-01 -9.20365602e-02
-3.07834614e-02 1.12858610e-02 2.14557648e-01 1.19862461e+00
-4.62334543e-01 -1.40762460e+00 -2.79226005e-01 3.15479755e-01
-5.38581073e-01 4.03686941e-01 -1.65229231e-01 1.10766876e+00
-4.07995492e-01 9.49467003e-01 -2.69009143e-01 2.57728398e-02
3.80617261e-01 9.39072296e-03 7.69345880e-01 -1.62394375e-01
-9.00356114e-01 2.98945159e-01 2.60534108e-01 -9.72194195e-01
-4.79485273e-01 -8.38035464e-01 -1.32310927e+00 -1.81558758e-01
-9.11603943e-02 1.58345550e-01 5.42624235e-01 9.95171547e-01
1.60107225e-01 2.23200858e-01 7.58794129e-01 -8.01099718e-01
-1.15286982e+00 -6.31937206e-01 -3.94408464e-01 1.63881004e-01
5.06253421e-01 -8.80275249e-01 -1.66898474e-01 -4.17873979e-01] | [4.110957622528076, 2.1513941287994385] |
1ea0bdcc-671e-4619-b2f5-2611b7a815d6 | revisiting-graph-neural-networks-for-link-1 | 2010.16103 | null | https://arxiv.org/abs/2010.16103v5 | https://arxiv.org/pdf/2010.16103v5.pdf | Labeling Trick: A Theory of Using Graph Neural Networks for Multi-Node Representation Learning | In this paper, we provide a theory of using graph neural networks (GNNs) for multi-node representation learning (where we are interested in learning a representation for a set of more than one node, such as link). We know that GNN is designed to learn single-node representations. When we want to learn a node set representation involving multiple nodes, a common practice in previous works is to directly aggregate the single-node representations obtained by a GNN into a joint node set representation. In this paper, we show a fundamental constraint of such an approach, namely the inability to capture the dependence between nodes in the node set, and argue that directly aggregating individual node representations does not lead to an effective joint representation for multiple nodes. Then, we notice that a few previous successful works for multi-node representation learning, including SEAL, Distance Encoding, and ID-GNN, all used node labeling. These methods first label nodes in the graph according to their relationships with the target node set before applying a GNN. Then, the node representations obtained in the labeled graph are aggregated into a node set representation. By investigating their inner mechanisms, we unify these node labeling techniques into a single and most general form -- labeling trick. We prove that with labeling trick a sufficiently expressive GNN learns the most expressive node set representations, thus in principle solves any joint learning tasks over node sets. Experiments on one important two-node representation learning task, link prediction, verified our theory. Our work explains the superior performance of previous node-labeling-based methods, and establishes a theoretical foundation of using GNNs for multi-node representation learning. | ['Long Jin', 'Kai Wang', 'Yinglong Xia', 'Pan Li', 'Muhan Zhang'] | 2020-10-30 | revisiting-graph-neural-networks-for-link | http://proceedings.neurips.cc/paper/2021/hash/4be49c79f233b4f4070794825c323733-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/4be49c79f233b4f4070794825c323733-Paper.pdf | neurips-2021-12 | ['link-property-prediction'] | ['graphs'] | [ 3.20870221e-01 7.47406065e-01 -5.30584395e-01 -2.72524357e-01
-3.55225623e-01 -6.16817296e-01 4.35324490e-01 4.94106710e-01
2.58945171e-02 5.47603846e-01 -1.17276445e-01 -6.17336571e-01
-4.77315307e-01 -1.43678880e+00 -7.14821994e-01 -6.90930426e-01
-4.49548632e-01 7.78327107e-01 4.14903834e-02 -5.08357465e-01
-2.13925555e-01 8.24120820e-01 -1.30520773e+00 -5.64012565e-02
4.59315747e-01 6.09863639e-01 -4.77553606e-02 6.72717571e-01
-3.87389868e-01 9.01013076e-01 -6.93270504e-01 -3.22214514e-01
3.18642944e-01 -3.40991974e-01 -1.11719751e+00 -2.34169483e-01
5.23777604e-01 2.75579747e-03 -5.31512380e-01 9.85467255e-01
2.01489016e-01 2.35441923e-01 1.04574943e+00 -1.69626820e+00
-7.54465640e-01 1.31978774e+00 -6.28086448e-01 -1.53650388e-01
3.26373100e-01 -5.74665606e-01 1.52192843e+00 -3.22579503e-01
6.32837474e-01 1.37971187e+00 1.05926144e+00 6.56778157e-01
-1.36238277e+00 -4.62560922e-01 9.93139297e-02 -1.40953377e-01
-1.52940285e+00 1.19346790e-01 9.16635633e-01 -2.69159406e-01
8.07592452e-01 2.45000184e-01 5.28791010e-01 8.42065632e-01
-9.07140896e-02 7.41616488e-01 5.90232313e-01 -6.57609761e-01
-2.31000692e-01 -1.13925554e-01 5.89334607e-01 1.01238954e+00
5.64144194e-01 -2.86358618e-03 5.41620217e-02 -2.59991735e-01
9.12689507e-01 1.88100711e-01 -2.15267539e-01 -6.08678579e-01
-7.44669497e-01 1.20715642e+00 1.10969329e+00 7.64183640e-01
-6.55687004e-02 6.89014375e-01 5.69801092e-01 6.63685143e-01
3.81981701e-01 3.78851682e-01 -2.78743953e-01 6.12609208e-01
-6.87410235e-01 -9.58274230e-02 1.04695928e+00 8.76721323e-01
1.24345219e+00 5.16075306e-02 -4.17601466e-02 6.08663440e-01
3.34291548e-01 1.78599805e-01 2.99184412e-01 -7.97403693e-01
2.26363301e-01 7.39717782e-01 -6.55415177e-01 -1.34214687e+00
-7.74394810e-01 -5.20379305e-01 -1.15756655e+00 4.64390814e-02
2.98102915e-01 -2.91247237e-02 -8.79853308e-01 2.32192922e+00
-6.21522628e-02 1.97298720e-01 2.36030102e-01 2.95015097e-01
1.08746886e+00 5.38736761e-01 1.62447438e-01 -1.95754290e-01
1.00138831e+00 -9.68582392e-01 -4.98086691e-01 5.82121313e-02
1.45169973e+00 1.69928316e-02 3.85934979e-01 -1.14764489e-01
-8.89586389e-01 -6.07370138e-01 -9.75192428e-01 2.71893591e-02
-7.81246245e-01 -3.71734947e-01 1.26466584e+00 5.90312302e-01
-1.49943459e+00 9.70688939e-01 -3.82520139e-01 -6.66852236e-01
3.70936126e-01 7.30021477e-01 -6.77068233e-01 -9.55354422e-02
-1.35130227e+00 9.55737591e-01 8.38708878e-01 -1.35889307e-01
-8.96256089e-01 -3.98981005e-01 -1.23955262e+00 3.79792124e-01
3.80157143e-01 -8.39824200e-01 9.22854900e-01 -9.80989754e-01
-6.88158393e-01 8.84919405e-01 3.32015641e-02 -4.74655002e-01
-1.41299427e-01 3.41419727e-01 -4.09672111e-01 2.42088690e-01
-4.40076366e-02 7.61899531e-01 5.65029383e-01 -1.52769005e+00
-1.79064363e-01 -2.00596213e-01 5.01656592e-01 9.87937022e-03
-5.99800527e-01 -3.38949382e-01 -8.51861909e-02 -4.86352056e-01
2.08018512e-01 -8.17368269e-01 -4.30266887e-01 2.31027212e-02
-6.00908697e-01 -8.28449190e-01 7.99704552e-01 -3.56998444e-02
1.25696063e+00 -1.90066063e+00 4.20560747e-01 8.05998087e-01
9.05943990e-01 3.37302238e-01 -6.24471307e-01 7.00749397e-01
-5.27404487e-01 5.94264448e-01 -1.91740558e-01 -3.82563800e-01
2.44896729e-02 8.56449187e-01 -9.31258202e-02 4.38875169e-01
-9.97474268e-02 1.19772410e+00 -1.08715582e+00 -6.62017524e-01
1.62367485e-02 5.47976375e-01 -3.25527281e-01 1.40942097e-01
-2.52239164e-02 -2.55087316e-02 -3.91600490e-01 4.82406676e-01
5.72096646e-01 -5.42288303e-01 5.15590787e-01 -3.58969659e-01
6.82103574e-01 1.36400789e-01 -1.13692999e+00 1.65763283e+00
-5.33241272e-01 5.52394509e-01 -4.22240533e-02 -1.48498976e+00
1.17157483e+00 3.27510029e-01 5.46059787e-01 -2.26256311e-01
6.92161098e-02 2.07322165e-01 1.03671394e-01 -1.95838451e-01
2.81241298e-01 -1.73295259e-01 -2.09284201e-01 7.66830146e-01
5.04365742e-01 6.99254051e-02 2.40719065e-01 7.59720743e-01
1.28051448e+00 -2.01526195e-01 5.90284348e-01 -1.97248504e-01
4.34072614e-01 -5.48582375e-01 2.71499395e-01 1.02520275e+00
1.18858382e-01 4.85754251e-01 8.41214359e-01 -5.14824092e-01
-7.59999454e-01 -1.00599825e+00 9.97615233e-02 1.42272282e+00
3.54636014e-02 -8.90298426e-01 -4.46585298e-01 -9.49442625e-01
9.52204093e-02 3.63773286e-01 -9.91597772e-01 -4.51667905e-01
-7.91704595e-01 -4.34148878e-01 9.18573022e-01 8.08041215e-01
-1.86223239e-02 -1.03361595e+00 -2.04560265e-01 1.06721424e-01
1.48473561e-01 -7.71095395e-01 -9.46474969e-02 6.84112668e-01
-9.33003128e-01 -1.16867423e+00 -5.68641841e-01 -1.20086849e+00
9.53934431e-01 2.13612244e-01 1.43978727e+00 9.39738095e-01
-5.42323180e-02 4.97407675e-01 -3.97102088e-01 1.07460342e-01
-6.89956725e-01 5.26739180e-01 9.15867984e-02 -3.36067200e-01
8.13705400e-02 -1.02901101e+00 1.44372433e-02 4.64804657e-02
-9.79280710e-01 -1.58865184e-01 6.85543299e-01 7.94881701e-01
3.05584729e-01 4.62070033e-02 6.92068994e-01 -1.43765926e+00
6.28892899e-01 -7.64477313e-01 -1.46012351e-01 6.92882836e-01
-5.57527363e-01 3.80941778e-01 7.47391284e-01 -4.63584632e-01
-3.44480246e-01 5.08687366e-03 -1.58793926e-01 -3.74966383e-01
4.62754928e-02 8.61560822e-01 -1.08966216e-01 -3.95604998e-01
7.62519479e-01 -2.48235259e-02 8.27893987e-02 -3.68930519e-01
7.42961586e-01 2.97667533e-01 4.64586735e-01 -8.43382299e-01
1.03350925e+00 1.34944215e-01 7.13258505e-01 -5.00786424e-01
-8.96644592e-01 -3.42236340e-01 -8.92230749e-01 5.64613640e-02
4.84182805e-01 -6.29644752e-01 -8.81692410e-01 -5.45644611e-02
-1.30383646e+00 -1.69968829e-01 -4.43491399e-01 2.27428973e-02
-4.25905913e-01 6.18236005e-01 -6.72407627e-01 -6.93677485e-01
-3.35994184e-01 -7.34180391e-01 7.40158319e-01 6.23982288e-02
-2.57779509e-01 -1.79251409e+00 1.55673370e-01 -2.71585375e-01
4.58821416e-01 3.90222222e-01 1.29256415e+00 -1.20969081e+00
-9.55623388e-02 -1.69967562e-01 -3.22459430e-01 3.66363227e-02
1.37266591e-01 -3.11396439e-02 -7.49214947e-01 -5.61350048e-01
-4.77425009e-01 -5.35641789e-01 1.14548564e+00 2.31704280e-01
1.18607485e+00 -3.05354238e-01 -8.71254146e-01 7.66796291e-01
1.64318001e+00 -2.55275488e-01 4.71034586e-01 -7.94004798e-02
1.15217674e+00 5.75664520e-01 -1.23238027e-01 2.25980338e-02
2.89526075e-01 5.17879963e-01 7.60195017e-01 -1.64514959e-01
-3.03890079e-01 -4.57596272e-01 1.18407294e-01 8.13707530e-01
-1.99744791e-01 -7.01036751e-01 -8.40930045e-01 5.73068380e-01
-1.85422182e+00 -7.27762282e-01 -2.01433718e-01 1.71172285e+00
5.37983954e-01 -4.25360985e-02 2.51084089e-01 2.00130105e-01
9.15538013e-01 3.77216667e-01 -2.97343850e-01 -3.82931381e-01
-2.22111911e-01 4.71020192e-01 4.39152062e-01 5.27938366e-01
-1.02466404e+00 8.53529274e-01 6.94472027e+00 8.48029673e-01
-6.97311699e-01 3.61234248e-02 3.03777009e-01 5.85061669e-01
-7.73716688e-01 1.11625940e-01 -6.93107724e-01 -1.51309177e-01
1.04186976e+00 -2.25277022e-01 3.05166662e-01 8.53737473e-01
-7.89217234e-01 5.00051975e-01 -1.52356529e+00 7.70189464e-01
8.50486532e-02 -1.44104850e+00 4.64723438e-01 1.38108701e-01
5.49546599e-01 -1.44851670e-01 -1.32696405e-01 5.60627401e-01
7.26147890e-01 -1.64029562e+00 7.76060075e-02 3.17710638e-01
7.61283755e-01 -6.60584688e-01 8.34965825e-01 2.75027514e-01
-1.72048330e+00 -1.11244708e-01 -5.45100749e-01 -1.77993938e-01
2.70576831e-02 3.85396004e-01 -7.02114046e-01 1.02719176e+00
1.59111604e-01 9.41802502e-01 -6.53952062e-01 7.59655476e-01
-3.42122018e-01 5.02731025e-01 -4.99106377e-01 5.11462614e-03
3.19514483e-01 8.47434159e-03 3.51417899e-01 1.17470598e+00
2.79134333e-01 -1.52275898e-02 4.44850117e-01 9.41659808e-01
-4.60111022e-01 -1.03371188e-01 -1.13778436e+00 6.30253972e-03
6.41050518e-01 1.49104226e+00 -7.40887165e-01 -2.10252106e-01
-3.95484954e-01 5.71163774e-01 9.37414348e-01 2.93650150e-01
-5.57280362e-01 -3.89438778e-01 2.98968196e-01 -4.22018245e-02
2.86957026e-01 -1.63600612e-02 -1.32016474e-02 -7.39839971e-01
-3.46243173e-01 -4.10791814e-01 8.37951779e-01 -5.41715980e-01
-1.47528040e+00 7.32600272e-01 2.70989090e-01 -9.86480415e-01
-4.09302682e-01 -7.35681057e-01 -8.06137502e-01 7.24040866e-01
-1.32969046e+00 -1.49606979e+00 -1.12727925e-01 6.63381517e-01
-1.53208300e-01 -1.59482092e-01 1.30146384e+00 1.40868500e-01
-3.26362342e-01 9.93558884e-01 -2.36786112e-01 5.19415557e-01
6.45742267e-02 -1.42541778e+00 2.68257499e-01 4.52281713e-01
6.39460146e-01 8.94406796e-01 3.92699271e-01 -4.37606037e-01
-1.18788612e+00 -1.02013242e+00 8.66241932e-01 -2.45483175e-01
6.66341782e-01 -2.69855022e-01 -1.12775707e+00 1.24021924e+00
5.26822247e-02 5.23289680e-01 7.78971136e-01 4.69874680e-01
-7.53951609e-01 2.28051133e-02 -1.11866176e+00 3.24234366e-01
1.33924794e+00 -6.83412790e-01 -6.08569026e-01 4.19489831e-01
9.58980978e-01 2.86254636e-03 -1.17021835e+00 4.77598071e-01
3.01175684e-01 -5.70658863e-01 1.21782875e+00 -8.95263135e-01
1.72982052e-01 -2.07343012e-01 -2.38804519e-01 -1.59753287e+00
-6.43134832e-01 -4.56456035e-01 -5.81215978e-01 1.20116889e+00
4.14495736e-01 -7.51776636e-01 7.66792238e-01 8.79886746e-02
-9.06705484e-02 -8.06610763e-01 -7.53024042e-01 -8.71085167e-01
3.34699988e-01 -1.64091453e-01 6.30626321e-01 1.08167589e+00
4.47730385e-02 7.06225097e-01 -3.50451589e-01 1.27425864e-01
7.12478995e-01 1.42083243e-01 5.06176472e-01 -1.89962983e+00
-3.98740232e-01 -5.43706417e-01 -7.99956620e-01 -1.12030768e+00
7.62181044e-01 -1.59691405e+00 -3.30760121e-01 -1.89903140e+00
3.21529299e-01 -9.21957254e-01 -5.80045879e-01 1.01833093e+00
5.28404377e-02 1.95835724e-01 2.57019997e-01 1.95596844e-01
-7.19495535e-01 1.77435011e-01 1.15413582e+00 -3.47814798e-01
9.72757041e-02 -6.65857047e-02 -1.02175128e+00 6.31260753e-01
6.57747507e-01 -6.94377244e-01 -5.54775894e-01 -2.45490730e-01
5.21782994e-01 2.60477122e-02 3.48781973e-01 -7.96329677e-01
3.89764339e-01 2.10864663e-01 1.55443445e-01 -3.74375463e-01
2.93252409e-01 -1.03441429e+00 2.67289490e-01 6.64308429e-01
-6.17362797e-01 -1.68979079e-01 -1.67399123e-01 5.04767478e-01
-2.70207506e-02 -7.21645832e-01 4.58947957e-01 -2.18705952e-01
-8.02289784e-01 4.12117660e-01 1.00164279e-01 6.51564403e-03
9.87892330e-01 -2.28076652e-01 -4.37609076e-01 -5.91209054e-01
-9.76041377e-01 3.54959726e-01 2.36359745e-01 9.71858054e-02
5.65971911e-01 -1.55161762e+00 -7.34550834e-01 9.69780833e-02
1.26720965e-01 7.99048021e-02 -9.20335278e-02 5.26623726e-01
-2.60009199e-01 9.90538374e-02 -1.24787457e-01 -4.80048269e-01
-1.21158803e+00 7.96754122e-01 4.71195102e-01 -6.14772677e-01
-6.16828263e-01 9.29469109e-01 9.96956229e-02 -7.41981447e-01
3.22680146e-01 -2.82762479e-02 -4.79937792e-01 1.13456853e-01
2.07708076e-01 2.14794084e-01 -3.01845223e-01 -8.58400583e-01
-3.08055878e-01 8.67657065e-01 -2.20804036e-01 4.92744476e-01
1.43409228e+00 1.54871985e-01 -4.60306942e-01 4.65433002e-01
1.68937600e+00 -3.69621009e-01 -4.77320492e-01 -4.55730945e-01
1.77773796e-02 4.05644663e-02 -2.99511999e-01 -3.19901943e-01
-1.42814136e+00 7.61467159e-01 1.71609059e-01 7.72808313e-01
9.39413786e-01 6.54639602e-01 5.85910738e-01 8.48565876e-01
3.96942586e-01 -3.66410524e-01 1.83909666e-02 4.75876123e-01
5.68741441e-01 -1.10276341e+00 1.27918303e-01 -7.48292327e-01
-6.51784837e-02 1.49117410e+00 5.90117037e-01 -2.68422097e-01
8.01086783e-01 2.44851962e-01 -3.39344114e-01 -5.53681552e-01
-4.90961224e-01 -4.13419873e-01 3.01556259e-01 1.02594721e+00
3.61073762e-01 3.35100442e-01 5.93693852e-02 3.40059906e-01
-1.05478987e-01 -4.39080566e-01 4.65186507e-01 6.94659293e-01
-4.50480580e-01 -1.40514529e+00 -2.20317505e-02 5.15323043e-01
-2.02437550e-01 -9.97447446e-02 -4.71223205e-01 1.06848168e+00
6.97963834e-02 6.87995076e-01 2.17207894e-01 -6.01679862e-01
-2.91721951e-02 -3.51284817e-02 6.51536345e-01 -9.20780778e-01
-7.05047488e-01 -6.27362430e-01 1.33846045e-01 -2.45218769e-01
-7.62498856e-01 -2.35435646e-02 -1.51639020e+00 -4.71191764e-01
-4.11949426e-01 3.80764127e-01 2.08710819e-01 8.29862356e-01
1.68862939e-02 6.27235174e-01 5.08094251e-01 -7.75546432e-01
-3.99489582e-01 -9.36358213e-01 -8.23769927e-01 6.63868308e-01
4.05700862e-01 -6.44412458e-01 -6.65708661e-01 -5.71949780e-01] | [7.042252063751221, 6.24869441986084] |
94f9a74a-e22e-4008-a6c1-066a392538fa | icn-interactive-convolutional-network-for | 2306.13897 | null | https://arxiv.org/abs/2306.13897v1 | https://arxiv.org/pdf/2306.13897v1.pdf | ICN: Interactive Convolutional Network for Forecasting Travel Demand of Shared Micromobility | Accurate shared micromobility demand predictions are essential for transportation planning and management. Although deep learning models provide powerful tools to deal with demand prediction problems, studies on forecasting highly-accurate spatiotemporal shared micromobility demand are still lacking. This paper proposes a deep learning model named Interactive Convolutional Network (ICN) to forecast spatiotemporal travel demand for shared micromobility. The proposed model develops a novel channel dilation method by utilizing multi-dimensional spatial information (i.e., demographics, functionality, and transportation supply) based on travel behavior knowledge for building the deep learning model. We use the convolution operation to process the dilated tensor to simultaneously capture temporal and spatial dependencies. Based on a binary-tree-structured architecture and interactive convolution, the ICN model extracts features at different temporal resolutions, and then generates predictions using a fully-connected layer. The proposed model is evaluated for two real-world case studies in Chicago, IL, and Austin, TX. The results show that the ICN model significantly outperforms all the selected benchmark models. The model predictions can help the micromobility operators develop optimal vehicle rebalancing schemes and guide cities to better manage the shared micromobility system. | ['Xilei Zhao', 'Xiaojian Zhang', 'Qian Ke', 'Yiming Xu'] | 2023-06-24 | null | null | null | null | ['management'] | ['miscellaneous'] | [-5.84339082e-01 -6.14769876e-01 -5.64347148e-01 -7.42204368e-01
-3.88480484e-01 -2.71871686e-01 4.53918934e-01 -1.53941587e-01
-2.19890088e-01 4.95497346e-01 4.62421507e-01 -8.36724997e-01
-3.85549188e-01 -1.24920893e+00 -4.16175812e-01 -5.40093780e-01
-4.02450651e-01 3.93182427e-01 1.79628238e-01 -5.86051106e-01
-1.15009844e-02 8.27722311e-01 -1.29656494e+00 4.50948000e-01
1.02576220e+00 1.33392465e+00 1.49061739e-01 4.96271133e-01
-2.55268425e-01 8.79047692e-01 -1.29779369e-01 -1.51719958e-01
3.10677081e-01 3.86838317e-01 -7.19496310e-01 -4.48478401e-01
5.10157458e-02 -8.52486730e-01 -9.79922593e-01 2.53048420e-01
3.72428954e-01 4.22830522e-01 3.78478587e-01 -1.62740886e+00
-6.79229975e-01 8.20795774e-01 -2.84289151e-01 7.93154776e-01
-2.79833585e-01 3.55682164e-01 8.27656388e-01 -6.71484888e-01
2.52226800e-01 1.08496785e+00 5.57394862e-01 -1.68317780e-01
-9.95778024e-01 -8.85817647e-01 6.99408054e-01 6.60563350e-01
-1.60713327e+00 -1.18237548e-01 6.37436509e-01 -4.44327265e-01
1.08741295e+00 3.04089844e-01 5.28239965e-01 8.56962323e-01
2.60323435e-01 9.68441844e-01 3.69595289e-01 4.72479463e-01
-8.65387693e-02 -6.68934407e-03 9.70017090e-02 4.38251644e-01
-2.05709711e-01 -2.86168884e-02 -1.57434191e-03 2.23300457e-01
4.38560843e-01 3.69130462e-01 1.00359827e-01 2.33674958e-01
-1.13128793e+00 5.94234169e-01 1.03603578e+00 2.21757665e-01
-6.09946251e-01 4.54402357e-01 4.73343223e-01 -1.88056640e-02
7.12872863e-01 -9.09511000e-02 -6.74973786e-01 -2.53872424e-01
-9.21422660e-01 3.37636530e-01 3.80763531e-01 1.01421022e+00
9.11799192e-01 3.48410457e-01 -2.26173878e-01 7.57137835e-01
5.69181144e-02 8.71925235e-01 2.05318823e-01 -6.36100948e-01
1.00413787e+00 6.32359505e-01 6.96163699e-02 -1.51962483e+00
-1.06137300e+00 -4.16036457e-01 -1.00210106e+00 -5.18525660e-01
2.04341765e-02 -6.50709689e-01 -7.79874980e-01 1.49704504e+00
2.29829997e-02 5.29363155e-01 -3.53469729e-01 1.03485072e+00
7.84015417e-01 1.08176172e+00 3.59582007e-01 4.53694075e-01
7.37911761e-01 -9.89802122e-01 -6.39917016e-01 1.52250871e-01
1.16570234e+00 -2.92323917e-01 7.31838584e-01 -1.07037306e-01
-8.39837909e-01 -5.79763949e-01 -6.07370555e-01 -3.79504040e-02
-8.92760873e-01 2.41964877e-01 6.53367937e-01 4.69293386e-01
-8.15643191e-01 1.52785063e-01 -8.21220875e-01 -2.77189463e-01
5.12040317e-01 6.28944278e-01 -1.07990010e-02 -1.28987640e-01
-1.55011201e+00 7.65320063e-01 6.85616909e-03 3.61400068e-01
-8.07263076e-01 -9.75163698e-01 -6.64246678e-01 3.47535819e-01
-2.77788956e-02 -4.78530258e-01 1.03250194e+00 -5.62479675e-01
-1.10125470e+00 -1.50657088e-01 6.55132309e-02 -4.14343566e-01
2.72728026e-01 2.99835522e-02 -1.26479328e+00 -3.81157875e-01
1.70977995e-01 6.61393464e-01 2.17773736e-01 -8.31712246e-01
-1.15253377e+00 -1.57049194e-01 2.07295895e-01 -1.38599604e-01
-5.65631449e-01 -3.80103625e-02 -5.43312073e-01 -5.48778176e-01
-1.03544094e-01 -9.37077641e-01 -5.33806801e-01 -5.65169632e-01
-4.72389758e-01 -1.85735226e-01 1.05392051e+00 -7.36188471e-01
1.81130111e+00 -1.98696387e+00 -3.93036008e-01 5.78725994e-01
1.55780122e-01 1.68919101e-01 -4.97050285e-01 6.84508502e-01
6.64378628e-02 1.70899421e-01 1.61817983e-01 -2.73027241e-01
1.42115667e-01 3.02052289e-01 -2.40571737e-01 3.28050762e-01
-1.48895970e-02 1.27336538e+00 -7.24842429e-01 -1.57191262e-01
5.11273205e-01 5.11928260e-01 -5.25904059e-01 -3.13960351e-02
-1.06679179e-01 5.53983033e-01 -6.10974848e-01 5.90502441e-01
1.15408468e+00 -1.87761784e-01 -1.50967808e-02 -2.68302172e-01
-7.51422584e-01 3.07321995e-01 -9.13356960e-01 1.19242060e+00
-1.01888704e+00 9.19857383e-01 -1.46964356e-01 -1.00618732e+00
1.01573932e+00 4.81426381e-02 7.71796882e-01 -1.40667284e+00
1.91165313e-01 2.83945084e-01 -1.21589050e-01 -7.49853492e-01
6.47720456e-01 5.83646417e-01 -5.09846687e-01 5.22442460e-01
-5.89487374e-01 5.79199195e-01 1.16593847e-02 8.27085972e-02
9.29176211e-01 -3.96608174e-01 -6.86943948e-01 -1.68206051e-01
5.38981497e-01 -3.20872255e-02 7.50624776e-01 2.74707109e-01
-1.86369300e-01 1.25652254e-01 2.38232732e-01 -1.12281394e+00
-9.64277267e-01 -9.16364133e-01 9.22264755e-02 1.25853658e+00
3.10281396e-01 -1.02076545e-01 -2.91141927e-01 -5.45521915e-01
3.54200453e-01 6.32406414e-01 -6.35003388e-01 -1.05425395e-01
-9.80733156e-01 -6.44473374e-01 6.13015473e-01 9.16212022e-01
8.21053088e-01 -6.85621679e-01 2.03124974e-02 4.29770529e-01
-4.66953397e-01 -1.40656662e+00 -5.28215230e-01 -3.15271050e-01
-3.10282379e-01 -5.33732235e-01 -5.34618974e-01 -7.61352956e-01
3.27330947e-01 6.87127948e-01 7.83489823e-01 2.49662939e-02
-2.86338739e-02 -5.75006530e-02 -1.10637154e-02 -8.92803445e-02
1.82421744e-01 8.30464125e-01 -7.11477622e-02 4.83868897e-01
5.58349729e-01 -7.35666096e-01 -9.80079412e-01 7.72877634e-01
-5.63958764e-01 2.32281506e-01 4.08701152e-01 2.75204450e-01
3.64340723e-01 2.37464964e-01 8.64252329e-01 -4.40494955e-01
8.47818077e-01 -1.34753442e+00 -5.59755981e-01 3.95873189e-02
-5.49254894e-01 -2.93434560e-01 9.51141775e-01 -2.25861982e-01
-1.15591288e+00 -1.47672310e-01 -3.61874908e-01 -1.68950036e-01
-2.71337062e-01 7.36560822e-01 -1.98903635e-01 9.35967490e-02
6.10713214e-02 2.69030303e-01 -6.15804970e-01 -5.70892751e-01
4.70351368e-01 8.42191100e-01 2.61131495e-01 -2.74266094e-01
6.40864611e-01 6.40977621e-01 -1.21512748e-01 -7.36954093e-01
-1.82294533e-01 -3.73030424e-01 -6.66058004e-01 -3.05754900e-01
7.74845004e-01 -9.18104291e-01 -8.59209657e-01 3.63544375e-01
-1.10154104e+00 -4.90178466e-01 2.24873438e-01 5.78412056e-01
-2.62593091e-01 -1.76418364e-01 -4.69773322e-01 -5.48337042e-01
4.09348346e-02 -1.16005933e+00 6.26440406e-01 1.63232476e-01
7.71722198e-02 -1.19069624e+00 -1.27769783e-01 3.49745661e-01
1.00365090e+00 7.08082318e-02 1.15562344e+00 -3.98560554e-01
-1.02512813e+00 -3.29881340e-01 -6.98093414e-01 8.40901434e-02
6.60840645e-02 1.07222907e-01 -5.90723336e-01 1.45008639e-01
-9.10814285e-01 3.55467469e-01 6.90421581e-01 5.82440913e-01
1.74175739e+00 -3.77024978e-01 -6.22060657e-01 8.32523286e-01
1.14413488e+00 4.15423363e-01 6.05236888e-01 2.30503529e-01
9.26842213e-01 7.31485486e-01 3.76349986e-01 7.49429822e-01
1.45333946e+00 6.46962464e-01 7.04505801e-01 -3.82003516e-01
6.73711896e-02 -1.36188254e-01 -1.20818108e-01 8.33489895e-01
-6.70031682e-02 -6.13098085e-01 -1.28504264e+00 9.96020913e-01
-2.03750348e+00 -9.99758482e-01 -5.93490422e-01 1.48059523e+00
-3.09651762e-01 5.01050912e-02 4.24388528e-01 -8.31042379e-02
4.41959471e-01 1.98647097e-01 -6.78138375e-01 -7.08329260e-01
-1.33471802e-01 -2.50384778e-01 1.09136152e+00 3.18652451e-01
-1.02602816e+00 9.11906898e-01 5.78210831e+00 6.03430688e-01
-1.52499533e+00 2.37016991e-01 7.99742579e-01 -2.36860901e-01
-6.14902914e-01 -4.18577969e-01 -6.72504663e-01 9.72840428e-01
1.24707448e+00 -1.94616392e-01 4.94535238e-01 5.95673084e-01
1.04590201e+00 3.12175781e-01 -6.67655230e-01 9.41954255e-01
-5.14158905e-01 -1.80123019e+00 1.51978657e-01 3.60874593e-01
1.07655966e+00 6.02390170e-01 5.86512625e-01 5.10040045e-01
2.82936484e-01 -1.13663721e+00 4.34769154e-01 8.89908850e-01
5.19085586e-01 -1.10840106e+00 8.54582787e-01 4.00195241e-01
-1.73539507e+00 -7.92797267e-01 2.94478908e-02 -7.75181800e-02
5.67715585e-01 2.94585675e-01 -5.34905970e-01 5.11779249e-01
6.54778957e-01 8.73928845e-01 -3.32617283e-01 9.84680176e-01
3.92146975e-01 6.14715815e-01 -3.83510441e-01 -8.26547667e-02
8.08009028e-01 6.53015003e-02 -7.05436617e-02 1.43232810e+00
4.70281452e-01 -6.42292276e-02 3.41775417e-01 6.54104829e-01
-1.33069888e-01 2.63854712e-01 -4.17827249e-01 1.10795155e-01
5.13780892e-01 1.15977657e+00 -4.47454572e-01 -1.36776909e-01
-9.44283307e-01 5.34455538e-01 -1.67788509e-02 7.78822839e-01
-1.35067904e+00 -5.26399910e-01 1.38524199e+00 3.18325371e-01
4.47776973e-01 -6.45486891e-01 -4.54874009e-01 -7.01182663e-01
-2.65861928e-01 -1.18829124e-01 2.36046433e-01 -4.85373229e-01
-1.09345663e+00 5.83930910e-01 -3.83130275e-02 -1.27198398e+00
7.21261799e-02 -3.75201732e-01 -8.36441219e-01 1.12931156e+00
-1.85523891e+00 -1.46007907e+00 -6.65639102e-01 8.53826642e-01
5.56031227e-01 -1.60560697e-01 4.39287841e-01 1.15029418e+00
-1.20174420e+00 6.96970582e-01 3.29393923e-01 2.18332171e-01
1.79318547e-01 -5.56908548e-01 1.02144408e+00 5.88553429e-01
-5.68208456e-01 2.81241357e-01 1.11978866e-01 -5.07539809e-01
-1.25689280e+00 -1.67920947e+00 1.06129539e+00 -1.39624089e-01
7.26507843e-01 -2.41013393e-01 -5.74268579e-01 6.70616925e-01
-8.23645666e-02 3.96549433e-01 8.54595900e-01 1.14837274e-01
-6.80812374e-02 -6.42096996e-01 -9.43233371e-01 4.53231424e-01
1.17150855e+00 -3.80475998e-01 3.66207242e-01 3.62977117e-01
8.85863125e-01 2.58149020e-02 -8.78841639e-01 2.35159308e-01
7.18350470e-01 -6.12330258e-01 7.51409292e-01 -8.22691739e-01
1.93832681e-01 -5.87060712e-02 -4.70201373e-01 -1.40326607e+00
-7.87595451e-01 -1.36505917e-01 -1.46379203e-01 8.92398477e-01
8.05028319e-01 -7.45652080e-01 7.87188888e-01 1.03414905e+00
-5.19435763e-01 -1.13438427e+00 -7.62245595e-01 -7.38108337e-01
1.30165100e-01 -1.08337927e+00 1.79427373e+00 8.01524222e-01
-4.17597264e-01 -6.51020780e-02 -4.68100339e-01 4.03811693e-01
-2.06049923e-02 1.09813720e-01 8.29014897e-01 -1.04875708e+00
5.32691598e-01 -6.77791059e-01 -4.65752870e-01 -1.63129401e+00
2.04116955e-01 -9.72140729e-01 -5.77625036e-01 -1.82560825e+00
-4.37723070e-01 -1.00762677e+00 -6.49595201e-01 2.82693177e-01
6.03461981e-01 1.22874022e-01 -8.13428499e-03 7.07626417e-02
-5.77492297e-01 6.03094578e-01 1.30961514e+00 -4.51208085e-01
-4.17872369e-01 2.06802309e-01 -6.14077866e-01 4.36871618e-01
1.02582335e+00 6.50777146e-02 -5.89215159e-01 -1.12167859e+00
2.72346884e-01 -7.81696290e-02 3.12655151e-01 -9.53129470e-01
4.71343011e-01 -6.49738550e-01 2.67042190e-01 -1.25877714e+00
2.37721860e-01 -9.54655170e-01 1.52402952e-01 2.57641584e-01
-2.52840877e-01 5.37718654e-01 3.57849181e-01 4.04373199e-01
-1.59619644e-01 8.44735861e-01 1.63875297e-01 3.19342822e-01
-1.23851848e+00 1.23003793e+00 -7.91080475e-01 -4.28845763e-01
1.16564190e+00 -1.92259356e-01 -4.75904107e-01 -5.46295106e-01
-7.38877475e-01 9.87033308e-01 -2.26368070e-01 8.38135958e-01
7.11435676e-01 -1.78523064e+00 -6.09022558e-01 3.55800182e-01
6.18994609e-02 -4.31936264e-01 9.11014676e-01 9.73393202e-01
-7.16620266e-01 1.02145016e+00 -3.45441967e-01 -3.34649831e-01
-6.29471600e-01 4.59638864e-01 3.40142578e-01 -6.11232109e-02
-4.11642224e-01 5.74492991e-01 7.39933699e-02 -5.96924603e-01
1.51828676e-01 -6.74739778e-01 -2.96255082e-01 4.74373959e-02
3.76516312e-01 1.03622389e+00 2.12644339e-01 -1.05035889e+00
-5.46116889e-01 2.97462374e-01 3.11893485e-02 1.44709304e-01
1.41708994e+00 -6.82316542e-01 2.91780025e-01 8.98088217e-02
1.58478951e+00 -4.37257051e-01 -1.06404495e+00 -2.37415031e-01
-2.57840365e-01 -6.54875934e-01 4.60699081e-01 -8.64915311e-01
-1.58714020e+00 1.10630488e+00 6.54715240e-01 4.90792006e-01
1.26254201e+00 -2.93807328e-01 1.80860829e+00 2.13900551e-01
2.56876826e-01 -1.31915522e+00 -3.46703798e-01 7.23409474e-01
6.51317000e-01 -1.18045831e+00 -6.06351852e-01 -1.22031815e-01
-4.58583683e-01 9.37528908e-01 8.56054187e-01 5.51282726e-02
1.26848495e+00 -3.71197574e-02 1.29050147e-02 -9.46095511e-02
-7.05930352e-01 -2.62933761e-01 1.63086534e-01 5.13261616e-01
1.77439693e-02 5.99317849e-01 1.49662927e-01 8.93104076e-01
-1.06034398e-01 -1.39202669e-01 1.90375730e-01 5.11668324e-01
-2.56484658e-01 -8.94434273e-01 2.90421154e-02 7.08592951e-01
1.67212322e-01 1.21533021e-01 2.76402950e-01 6.34376347e-01
7.46928811e-01 1.02291489e+00 5.82827985e-01 -1.11599636e+00
4.68293756e-01 -3.43486637e-01 -2.47518867e-01 -1.10899061e-01
-4.26436275e-01 -5.03329277e-01 9.35770497e-02 -7.60007679e-01
1.08772524e-01 -5.44451058e-01 -1.26630628e+00 -1.16753244e+00
1.19769521e-01 -2.06148684e-01 1.04484272e+00 8.35088968e-01
8.36850762e-01 7.39764333e-01 1.26756263e+00 -7.87210584e-01
2.30265751e-01 -5.90405703e-01 -5.09239435e-01 1.92930430e-01
6.42203212e-01 -8.11846375e-01 1.39979228e-01 -5.39713979e-01] | [6.435644149780273, 2.0472424030303955] |
d9749c36-a06b-458c-9ce3-1046f2062cfc | egfi-drug-drug-interaction-extraction-and | 2101.09914 | null | https://arxiv.org/abs/2101.09914v1 | https://arxiv.org/pdf/2101.09914v1.pdf | EGFI: Drug-Drug Interaction Extraction and Generation with Fusion of Enriched Entity and Sentence Information | The rapid growth in literature accumulates diverse and yet comprehensive biomedical knowledge hidden to be mined such as drug interactions. However, it is difficult to extract the heterogeneous knowledge to retrieve or even discover the latest and novel knowledge in an efficient manner. To address such a problem, we propose EGFI for extracting and consolidating drug interactions from large-scale medical literature text data. Specifically, EGFI consists of two parts: classification and generation. In the classification part, EGFI encompasses the language model BioBERT which has been comprehensively pre-trained on biomedical corpus. In particular, we propose the multi-head attention mechanism and pack BiGRU to fuse multiple semantic information for rigorous context modeling. In the generation part, EGFI utilizes another pre-trained language model BioGPT-2 where the generation sentences are selected based on filtering rules. We evaluated the classification part on "DDIs 2013" dataset and "DTIs" dataset, achieving the FI score of 0.842 and 0.720 respectively. Moreover, we applied the classification part to distinguish high-quality generated sentences and verified with the exiting growth truth to confirm the filtered sentences. The generated sentences that are not recorded in DrugBank and DDIs 2013 dataset also demonstrate the potential of EGFI to identify novel drug relationships. | ['Ka-Chun Wong', 'Linqi Song', 'Xiangtao Li', 'Jiecong Lin', 'Lei Huang'] | 2021-01-25 | null | null | null | null | ['drug-drug-interaction-extraction'] | ['natural-language-processing'] | [ 3.27766001e-01 9.42270160e-02 -3.11801791e-01 -2.42693216e-01
-1.06999433e+00 -2.84598261e-01 3.73691916e-01 4.74433959e-01
-4.00214851e-01 1.42111099e+00 3.71441841e-01 -4.87104356e-01
-3.82232696e-01 -6.80421472e-01 -8.90825391e-01 -6.07461452e-01
2.04548240e-01 5.15670419e-01 -7.20955655e-02 -3.08378842e-02
1.84906349e-01 1.51558936e-01 -1.04047477e+00 6.63399994e-01
1.43006408e+00 5.83184302e-01 5.23346424e-01 3.31148267e-01
-3.86523485e-01 6.47176027e-01 -6.34257257e-01 -5.71985304e-01
-4.17006910e-01 -5.90823829e-01 -9.28792059e-01 -3.95348072e-01
-3.23479950e-01 1.06386274e-01 -3.94213982e-02 1.06303453e+00
7.64563918e-01 -4.07394230e-01 5.70057094e-01 -6.93779945e-01
-6.82745218e-01 9.63528395e-01 -4.14681435e-01 3.36430728e-01
4.42535222e-01 1.57532603e-01 9.14075732e-01 -9.30601835e-01
1.01216912e+00 1.19094658e+00 3.45489562e-01 6.83267415e-01
-6.08265698e-01 -7.84957767e-01 9.32864961e-04 2.65106469e-01
-1.40554297e+00 -3.97345781e-01 3.65213126e-01 -4.87435758e-01
1.29181731e+00 2.64434397e-01 1.78783432e-01 1.27237320e+00
6.53771400e-01 7.41014004e-01 6.93322062e-01 -2.06835657e-01
8.77978280e-02 4.51631583e-02 4.08534735e-01 5.04001677e-01
4.19191480e-01 -9.46532935e-02 -4.42830771e-01 -3.67607921e-01
4.05183807e-02 -1.13141790e-01 -3.72567683e-01 9.14713740e-01
-1.24049354e+00 6.99925601e-01 3.63913253e-02 5.09366333e-01
-7.30089247e-01 -4.84889507e-01 5.24159849e-01 -2.13005953e-02
4.77190495e-01 5.56222916e-01 -8.70664537e-01 6.33968934e-02
-5.67871213e-01 -3.41367498e-02 7.54381716e-01 1.02799094e+00
3.31303716e-01 -4.71745312e-01 -5.58110178e-01 8.36436868e-01
3.62029016e-01 3.77255201e-01 8.74036312e-01 -2.54766166e-01
7.14245260e-01 7.81447530e-01 -1.57928735e-01 -8.24006677e-01
-4.01108265e-01 -7.60712087e-01 -9.37553227e-01 -7.51116216e-01
-1.26022100e-01 -4.43123311e-01 -9.97011185e-01 1.79099286e+00
3.63851815e-01 3.72531831e-01 4.95806992e-01 5.03893077e-01
1.52086496e+00 6.21979475e-01 7.20554054e-01 -4.15986747e-01
1.62401855e+00 -7.15291440e-01 -1.04048693e+00 -4.58169207e-02
6.84094548e-01 -7.40005910e-01 5.92846572e-01 2.35468403e-01
-9.97760594e-01 -3.42586935e-01 -1.16676140e+00 -2.03909263e-01
-4.69418705e-01 3.24768931e-01 3.64557952e-01 1.81393906e-01
-4.97529417e-01 6.05831981e-01 -6.03876293e-01 -1.31510630e-01
6.85349524e-01 3.85623276e-01 -3.78086329e-01 -4.49883677e-02
-1.77532399e+00 8.98821056e-01 7.24604070e-01 2.64345646e-01
-7.03602970e-01 -1.05185628e+00 -6.70449078e-01 1.13337860e-01
2.71932065e-01 -1.01579642e+00 8.97437811e-01 -3.54583710e-01
-1.31703663e+00 7.13365316e-01 -3.55710357e-01 -3.95786494e-01
2.29868054e-01 -1.55365467e-02 -5.87747097e-01 1.12285919e-01
5.08993268e-01 2.75109917e-01 8.12678561e-02 -7.19519377e-01
-6.44763410e-01 -6.58573329e-01 -2.60973692e-01 1.90316066e-01
-1.26788765e-01 1.28054932e-01 -5.77670515e-01 -5.96594453e-01
-3.76354396e-01 -6.17762029e-01 -3.23945552e-01 -6.64503932e-01
-7.96791852e-01 -3.78664672e-01 4.35429692e-01 -1.07250011e+00
1.59726262e+00 -1.78382492e+00 1.04223013e-01 1.75724730e-01
3.29629719e-01 5.08952558e-01 -2.07810491e-01 3.18772018e-01
-2.44389787e-01 3.93342465e-01 -2.17001140e-01 2.89424416e-02
-3.62563789e-01 1.53124273e-01 -4.59093451e-02 1.35448307e-01
4.60767716e-01 1.06724882e+00 -1.13756239e+00 -6.44818723e-01
-2.05010146e-01 3.92792046e-01 -4.72919673e-01 1.83097050e-01
-4.12934005e-01 6.00800633e-01 -1.04010689e+00 7.39058137e-01
5.79919636e-01 -5.67215681e-01 2.97354430e-01 -4.03417140e-01
7.11110085e-02 5.87476671e-01 -6.74361467e-01 1.71695960e+00
-3.23284417e-01 -2.50354975e-01 -2.49562219e-01 -9.75107849e-01
7.45574117e-01 7.22517073e-01 6.48722827e-01 -6.61864042e-01
2.18609244e-01 3.45779479e-01 1.34670153e-01 -9.56405342e-01
3.05771455e-02 -1.73920855e-01 5.26270224e-03 -8.27970505e-02
1.81862935e-01 3.94190460e-01 4.12663698e-01 2.47303158e-01
1.36981165e+00 -5.50985411e-02 5.09233832e-01 -3.06758493e-01
1.04557443e+00 1.32972628e-01 8.09893072e-01 4.74569052e-01
2.45809987e-01 2.05568254e-01 3.79972696e-01 -3.77672352e-02
-5.62898397e-01 -5.79329371e-01 -4.36489344e-01 3.38784635e-01
-1.08435072e-01 -5.60651302e-01 -7.34239101e-01 -9.76823688e-01
-1.52426243e-01 6.24756217e-01 -4.88236755e-01 -1.04865879e-01
-3.58290225e-01 -1.31623757e+00 5.77080607e-01 1.61893070e-01
4.85220879e-01 -1.26570368e+00 -1.91800240e-02 6.22970998e-01
-3.70993555e-01 -1.03075516e+00 -5.02194703e-01 7.42343739e-02
-5.09320140e-01 -1.36831546e+00 -5.25918126e-01 -5.80554366e-01
4.88544315e-01 -2.46153489e-01 1.05778086e+00 5.52453846e-02
-3.22572112e-01 -4.16876078e-01 -4.48567331e-01 -6.36704624e-01
-5.69532990e-01 2.82111347e-01 -2.67174572e-01 -1.86138511e-01
6.73054039e-01 -2.99779207e-01 -4.95367378e-01 -4.00048122e-02
-8.53118539e-01 2.03928471e-01 8.64452541e-01 1.05634153e+00
7.22684681e-01 1.65293943e-02 1.28017509e+00 -1.23960412e+00
8.06062698e-01 -1.04374814e+00 -3.07570338e-01 4.40546244e-01
-6.72960281e-01 2.35693052e-01 6.84393048e-01 -2.31690034e-01
-1.11462224e+00 1.10616097e-02 -7.71990597e-01 1.44383654e-01
-2.44163815e-02 1.23576462e+00 -5.81813157e-01 6.64414525e-01
3.83622527e-01 1.38110965e-01 -1.12628751e-01 -5.59625089e-01
6.86395392e-02 1.06091046e+00 3.33146423e-01 -4.96834159e-01
1.16943538e-01 -1.71567783e-01 -2.29519874e-01 -5.69722235e-01
-9.75314915e-01 -3.16647291e-01 -1.62482068e-01 5.12491107e-01
1.10899758e+00 -8.65770519e-01 -9.01415586e-01 2.11758137e-01
-1.48589635e+00 2.15200707e-01 3.53488252e-02 7.26432681e-01
7.36576831e-03 3.86617273e-01 -8.13708305e-01 -4.01298344e-01
-1.03891242e+00 -1.42497170e+00 1.06799376e+00 1.42985508e-01
-2.93617278e-01 -7.05835879e-01 1.21750586e-01 4.21679169e-01
-1.00779548e-01 1.27512455e-01 1.38934708e+00 -1.13170969e+00
-2.72305965e-01 -2.70751538e-03 -2.11028323e-01 1.12364687e-01
4.92225826e-01 -1.38154656e-01 -8.35524499e-01 2.07261831e-01
-2.59370059e-02 5.85144348e-02 7.63823867e-01 3.20413768e-01
1.26790738e+00 -4.91984546e-01 -7.37383246e-01 3.05668831e-01
1.23869061e+00 7.75597453e-01 7.32595265e-01 1.69128161e-02
5.58177948e-01 3.06214094e-01 4.75256354e-01 1.58632308e-01
2.69257277e-01 4.67398316e-01 -2.02557161e-01 -1.05822459e-01
-7.46731460e-02 -2.17338651e-01 8.13608617e-02 1.12114584e+00
7.60605559e-02 -3.97903711e-01 -8.98605227e-01 4.43018049e-01
-1.70487928e+00 -9.93234694e-01 -2.86543444e-02 1.83885729e+00
1.72521865e+00 8.07809979e-02 -4.21911329e-01 -2.66596287e-01
7.20029771e-01 -4.45656776e-01 -6.73187613e-01 -4.76991944e-02
-2.20070854e-01 6.08780980e-01 2.17462748e-01 5.64088345e-01
-8.53546679e-01 8.34782839e-01 5.22297144e+00 1.16816998e+00
-9.33024764e-01 1.52413025e-01 8.54526043e-01 2.69667089e-01
-5.03790259e-01 -1.88540623e-01 -1.27264023e+00 9.32734728e-01
1.21239698e+00 -5.36264181e-01 -2.01217681e-01 4.55280751e-01
4.81091827e-01 2.45343834e-01 -9.69280183e-01 7.64148891e-01
-1.24340087e-01 -1.77646911e+00 4.05619204e-01 2.56215651e-02
4.62847084e-01 2.52858344e-02 -3.29101682e-01 3.50296378e-01
1.57631040e-01 -1.08105576e+00 1.51652649e-01 7.43951499e-01
6.84430122e-01 -5.44134200e-01 1.01705933e+00 3.63059819e-01
-1.00507569e+00 9.45424736e-02 -2.93696553e-01 5.31327009e-01
-5.51813096e-03 9.86105800e-01 -1.07679367e+00 1.33938515e+00
2.96504170e-01 1.08534420e+00 -4.74066585e-01 9.68894839e-01
-1.99667856e-01 6.08342230e-01 -1.04022644e-01 -1.77837804e-01
1.65912032e-01 -6.64497241e-02 3.02547455e-01 1.36916566e+00
3.56762856e-01 3.44819933e-01 2.39448875e-01 8.59741569e-01
-4.16125894e-01 5.54128230e-01 -3.67864162e-01 -3.69361162e-01
5.47207773e-01 1.12644875e+00 -2.88629979e-01 -5.88439703e-01
-4.17490453e-01 6.82530224e-01 4.30093966e-02 2.73831457e-01
-9.12644744e-01 -4.87212241e-01 3.40914637e-01 -1.51207253e-01
-1.28181085e-01 4.25460458e-01 -2.46925473e-01 -1.20539522e+00
-1.70160472e-01 -1.06832063e+00 6.09348476e-01 -6.22062206e-01
-1.56758177e+00 9.33525920e-01 -1.29501596e-01 -9.12343502e-01
-1.98142558e-01 -4.43167269e-01 -3.82210732e-01 1.14583731e+00
-1.44111252e+00 -1.08297503e+00 4.66086455e-02 3.52229476e-01
7.16513455e-01 -1.73885792e-01 9.92076218e-01 8.58584940e-01
-9.49439704e-01 6.13809466e-01 -2.48719584e-02 1.43131278e-02
7.25255907e-01 -8.72718096e-01 2.03274533e-01 3.07377517e-01
-4.07583922e-01 1.04348803e+00 4.24123347e-01 -1.21977270e+00
-1.22842610e+00 -1.33039057e+00 1.23991466e+00 -2.44388893e-01
5.06921470e-01 1.23721622e-02 -1.15282142e+00 3.89631033e-01
1.50821000e-01 -4.88451183e-01 9.57782090e-01 -1.20530523e-01
1.00426972e-01 1.82995677e-01 -1.10018730e+00 3.54858547e-01
9.96814311e-01 -2.84817696e-01 -7.79811800e-01 5.70537388e-01
9.26207542e-01 -4.37931359e-01 -1.10270119e+00 7.04341769e-01
2.71769375e-01 -1.84329655e-02 1.01638687e+00 -9.72100139e-01
7.80185759e-01 -2.92889357e-01 2.88559962e-02 -1.04406798e+00
-1.22328557e-01 -4.58271950e-01 9.18576494e-02 1.27898490e+00
1.17550147e+00 -5.90206265e-01 5.09389877e-01 4.36002254e-01
-3.56734365e-01 -1.14019465e+00 -8.40218663e-01 -3.87065768e-01
1.02119692e-01 -1.70007199e-01 7.89869487e-01 9.48128104e-01
2.31339172e-01 9.55040157e-01 -1.23113602e-01 9.86806825e-02
2.43858933e-01 1.28810713e-02 2.05464646e-01 -1.14117944e+00
-2.99148142e-01 -2.21418440e-01 -2.01784093e-02 -6.97644949e-01
3.29792023e-01 -1.27428341e+00 -1.00591063e-01 -1.73503983e+00
6.73544168e-01 -3.18759203e-01 -5.32667279e-01 6.47353530e-01
-6.85021222e-01 -2.41029069e-01 -5.32084048e-01 1.85982287e-02
-3.27750862e-01 5.51676095e-01 1.23700535e+00 -4.15922493e-01
-2.74559051e-01 -1.42303616e-01 -1.03113639e+00 4.79449123e-01
6.27540648e-01 -4.41293657e-01 -4.65168476e-01 -1.22014605e-01
1.04247108e-01 2.36415312e-01 -5.12153469e-03 -4.21588898e-01
2.74258554e-01 -2.19934762e-01 3.41401637e-01 -7.84942567e-01
-1.44260868e-01 -3.84902298e-01 3.58873844e-01 6.28519535e-01
-4.10378098e-01 -9.68661606e-02 3.29578459e-01 5.22461832e-01
-2.09979489e-01 -3.48091841e-01 4.86130744e-01 -2.44076952e-01
-3.23547035e-01 5.28033197e-01 -1.28837153e-01 1.47546709e-01
7.27728844e-01 2.04222694e-01 -4.60166037e-01 2.38434926e-01
-8.39956701e-01 3.79652768e-01 -2.71294892e-01 4.27236825e-01
6.10703945e-01 -1.02635074e+00 -1.08136046e+00 1.56153962e-01
9.80026722e-02 -5.80593711e-03 4.86590296e-01 7.76527524e-01
-2.53370553e-01 6.16370678e-01 2.39231288e-01 -2.56954402e-01
-1.21971297e+00 7.01004744e-01 2.29282171e-01 -6.25014544e-01
-4.66905594e-01 6.56572402e-01 3.28257889e-01 -1.61154881e-01
-6.60461858e-02 -3.11355114e-01 -4.01209772e-01 1.46847870e-02
7.38236070e-01 -1.70422390e-01 3.67026627e-01 -3.65524799e-01
-6.74360693e-01 3.28460455e-01 -4.09722775e-01 2.42397353e-01
1.47050428e+00 1.34677172e-01 -4.89228636e-01 -8.03820789e-02
1.26980615e+00 4.33052294e-02 -3.74047101e-01 -1.34482905e-01
3.01857442e-01 6.34318665e-02 -1.38222501e-01 -1.32260513e+00
-8.86415243e-01 5.76373935e-01 2.53182679e-01 -4.33574557e-01
9.77007687e-01 4.84832898e-02 9.68643367e-01 1.78760350e-01
4.43957523e-02 -7.89282680e-01 -3.51946682e-01 4.80929822e-01
9.26446557e-01 -1.01026726e+00 4.03616019e-02 -5.26997387e-01
-4.79846001e-01 7.78846025e-01 4.19949621e-01 5.08618653e-01
7.72831619e-01 3.99152100e-01 -2.53927290e-01 -4.82322991e-01
-7.95465052e-01 1.07128895e-03 4.86750662e-01 1.34165421e-01
6.60077751e-01 -1.07927158e-01 -1.01095116e+00 1.33717227e+00
9.22597423e-02 4.92702961e-01 1.41614571e-01 6.08954906e-01
-1.09774835e-01 -1.33339357e+00 1.73430428e-01 6.03206754e-01
-9.71474886e-01 -6.22351289e-01 -2.63418794e-01 4.18117464e-01
3.84141654e-01 1.11924958e+00 -6.19228542e-01 -4.18322742e-01
3.03146422e-01 1.85420901e-01 1.71923652e-01 -8.49664092e-01
-7.57442653e-01 1.93174630e-01 3.49930286e-01 -2.49569610e-01
-4.33560759e-01 -2.36391872e-01 -1.53954315e+00 9.73868072e-02
-5.19347906e-01 7.01235950e-01 5.02352297e-01 1.19311678e+00
8.44798028e-01 9.12295640e-01 3.36029291e-01 5.59686422e-02
-3.61079007e-01 -1.02146089e+00 -1.35572374e-01 3.58595788e-01
-4.38623466e-02 -3.92334610e-01 1.17822133e-01 1.73644781e-01] | [8.46512508392334, 8.691789627075195] |
7908e0bc-81c7-46b0-b799-b899d0fba035 | human-art-a-versatile-human-centric-dataset | 2303.02760 | null | https://arxiv.org/abs/2303.02760v2 | https://arxiv.org/pdf/2303.02760v2.pdf | Human-Art: A Versatile Human-Centric Dataset Bridging Natural and Artificial Scenes | Humans have long been recorded in a variety of forms since antiquity. For example, sculptures and paintings were the primary media for depicting human beings before the invention of cameras. However, most current human-centric computer vision tasks like human pose estimation and human image generation focus exclusively on natural images in the real world. Artificial humans, such as those in sculptures, paintings, and cartoons, are commonly neglected, making existing models fail in these scenarios. As an abstraction of life, art incorporates humans in both natural and artificial scenes. We take advantage of it and introduce the Human-Art dataset to bridge related tasks in natural and artificial scenarios. Specifically, Human-Art contains 50k high-quality images with over 123k person instances from 5 natural and 15 artificial scenarios, which are annotated with bounding boxes, keypoints, self-contact points, and text information for humans represented in both 2D and 3D. It is, therefore, comprehensive and versatile for various downstream tasks. We also provide a rich set of baseline results and detailed analyses for related tasks, including human detection, 2D and 3D human pose estimation, image generation, and motion transfer. As a challenging dataset, we hope Human-Art can provide insights for relevant research and open up new research questions. | ['Lei Zhang', 'Qiang Xu', 'Jianan Wang', 'Ailing Zeng', 'Xuan Ju'] | 2023-03-05 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Ju_Human-Art_A_Versatile_Human-Centric_Dataset_Bridging_Natural_and_Artificial_Scenes_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Ju_Human-Art_A_Versatile_Human-Centric_Dataset_Bridging_Natural_and_Artificial_Scenes_CVPR_2023_paper.pdf | cvpr-2023-1 | ['3d-human-pose-estimation', 'human-detection'] | ['computer-vision', 'computer-vision'] | [ 8.58318731e-02 -2.27691203e-01 3.31593864e-02 -1.52240500e-01
-1.20801784e-01 -5.41621089e-01 1.04391348e+00 -3.74173701e-01
-4.37430322e-01 6.45753086e-01 3.97414178e-01 2.85398006e-01
4.11079079e-01 -6.21668458e-01 -5.54507017e-01 -3.93554568e-01
1.06262647e-01 7.92616308e-01 3.31897438e-01 -5.32386720e-01
-1.04178607e-01 5.94998062e-01 -1.57092607e+00 2.44197816e-01
3.09533477e-01 6.90217376e-01 4.79359776e-02 6.18713796e-01
3.50476682e-01 3.40430290e-01 -9.13390875e-01 -1.03850698e+00
3.55859816e-01 -3.52329224e-01 -4.69268858e-01 3.90428543e-01
5.49254298e-01 -4.55707282e-01 -8.92080188e-01 6.53497934e-01
7.25634873e-01 1.19825549e-01 7.73105443e-01 -1.45905793e+00
-6.46162689e-01 -1.13860909e-02 -9.47999537e-01 -7.13758543e-02
9.23220038e-01 6.52522385e-01 6.23587072e-01 -9.68097150e-01
9.53233182e-01 1.64966476e+00 6.12307310e-01 8.09265375e-01
-9.03012455e-01 -5.20245671e-01 -1.33062303e-01 1.09063834e-01
-1.22593832e+00 -3.00262749e-01 8.08473289e-01 -5.30007005e-01
6.27119184e-01 2.75161564e-01 1.26738954e+00 1.84364307e+00
6.55087233e-02 1.31412375e+00 7.86918938e-01 -3.52709502e-01
-1.36733606e-01 4.88046929e-02 -4.40990359e-01 6.90864146e-01
4.88893032e-01 3.13907325e-01 -6.23315513e-01 -1.33965481e-02
1.38179493e+00 1.12716824e-01 -3.29165459e-01 -6.42171741e-01
-1.84987819e+00 6.29224777e-01 5.52669227e-01 -3.88984345e-02
-3.36872280e-01 -6.72091469e-02 3.28310847e-01 -3.45193952e-01
9.60339755e-02 4.35714602e-01 -1.25829782e-02 4.75859903e-02
-6.13408566e-01 7.33262539e-01 5.00819504e-01 1.19156456e+00
2.18089476e-01 -8.81644115e-02 -2.30348125e-01 6.94025576e-01
5.99092096e-02 7.93063462e-01 2.43654996e-01 -1.00674701e+00
4.45292592e-01 7.21222162e-01 4.39291477e-01 -1.29848456e+00
-3.90355170e-01 -1.30945325e-01 -1.14514041e+00 3.59657183e-02
6.35069430e-01 -9.60765630e-02 -9.25757825e-01 1.27709389e+00
5.63578367e-01 -1.86435074e-01 -2.99668610e-01 1.34884238e+00
1.00511050e+00 6.79691315e-01 -1.52867017e-02 2.76576191e-01
1.70599115e+00 -1.21819568e+00 -6.08343124e-01 -5.64743638e-01
5.61635047e-02 -6.83643460e-01 1.02809572e+00 3.11435997e-01
-1.19092894e+00 -6.36324644e-01 -7.19500244e-01 -2.75573701e-01
-3.04307669e-01 2.01235697e-01 7.18010128e-01 4.66838449e-01
-4.22194570e-01 2.39274323e-01 -5.40674150e-01 -6.18602037e-01
5.51616967e-01 -1.87306806e-01 -6.98149741e-01 -6.05614781e-02
-1.10880923e+00 9.29115951e-01 2.49173835e-01 2.66064972e-01
-7.71218598e-01 -3.07514906e-01 -9.74917829e-01 -4.50785607e-01
3.65578413e-01 -1.19581485e+00 1.04703057e+00 -5.46265602e-01
-1.09516227e+00 1.37755203e+00 1.13102585e-01 -2.86951661e-01
1.09572184e+00 -4.82200861e-01 -3.98825854e-01 2.84487486e-01
3.55109647e-02 9.76010919e-01 8.74423683e-01 -1.31319392e+00
-4.89268690e-01 -4.94153321e-01 4.13336838e-03 2.90863186e-01
-3.14169638e-02 6.36523515e-02 -9.89580512e-01 -1.04512453e+00
-1.45684183e-01 -1.03315401e+00 -2.43532211e-01 4.13853258e-01
-6.81634426e-01 1.02859093e-02 7.36364901e-01 -7.29914963e-01
9.35035586e-01 -1.82061052e+00 4.85978663e-01 -1.67586654e-01
2.11347640e-01 3.24114531e-01 3.14727947e-02 3.17342311e-01
1.80451319e-01 -9.08210948e-02 -1.99294671e-01 -4.16736037e-01
1.47020951e-01 3.36879236e-03 -1.51137829e-01 5.00693023e-01
1.10627651e-01 1.40102351e+00 -8.91754031e-01 -7.80442357e-01
5.62226057e-01 7.21199572e-01 -2.30951026e-01 2.00498641e-01
-2.77792830e-02 7.66017258e-01 -4.32605505e-01 8.00713241e-01
3.56753081e-01 -2.02005550e-01 -2.42171735e-01 -3.01571518e-01
3.03011954e-01 -2.73199022e-01 -1.04485488e+00 1.71176112e+00
7.66874477e-02 7.28507817e-01 -8.46864656e-02 -2.90613502e-01
5.74444354e-01 1.80036783e-01 4.15238291e-01 -6.22313082e-01
1.44067138e-01 -2.45291889e-01 -3.38695109e-01 -5.74768424e-01
5.93440294e-01 2.75776777e-02 -2.85526723e-01 1.38797864e-01
-4.21078116e-01 -2.46576220e-01 2.84834385e-01 2.30044976e-01
8.29305828e-01 3.00429136e-01 3.41513574e-01 2.66394079e-01
2.66984522e-01 7.05579817e-02 3.10399383e-01 4.93298650e-01
-3.85624737e-01 1.16665924e+00 1.36651307e-01 -9.08021629e-01
-1.42878044e+00 -1.33141696e+00 1.53494487e-02 7.71114588e-01
2.97995865e-01 -4.07219678e-01 -7.21030414e-01 -4.74716812e-01
5.77635393e-02 2.72540540e-01 -5.85496247e-01 4.22984660e-02
-8.50695491e-01 -5.25036752e-01 7.11403251e-01 6.21267021e-01
8.10732305e-01 -1.55296087e+00 -1.04337704e+00 -1.26215935e-01
-5.09139478e-01 -1.33100295e+00 -6.29731357e-01 -5.99992871e-01
-4.68880802e-01 -1.24385703e+00 -1.39641511e+00 -6.80370033e-01
6.00779295e-01 3.20455194e-01 1.34571743e+00 -1.62474979e-02
-9.42098379e-01 4.08062279e-01 -1.92243353e-01 -3.93822670e-01
-1.23923458e-01 -2.98517644e-01 3.23945135e-01 -2.89229508e-02
2.11738497e-01 -2.12249786e-01 -9.12881672e-01 6.56810164e-01
-5.86574376e-01 4.26604986e-01 6.08791828e-01 6.90349638e-01
4.40781474e-01 -8.87781382e-02 -4.94625308e-02 -5.75977743e-01
1.95385158e-01 -1.22247808e-01 -2.63955653e-01 1.39139086e-01
2.76313692e-01 -3.84405315e-01 3.14885676e-01 -5.80121398e-01
-1.03605449e+00 2.50585139e-01 1.88627213e-01 -5.69910169e-01
-4.18254018e-01 -2.16398403e-01 -2.29881272e-01 3.10952067e-01
8.97999406e-01 2.64044970e-01 -1.20994084e-01 -4.03542995e-01
3.84156108e-01 4.55954134e-01 9.59179342e-01 -6.53190494e-01
1.01785374e+00 7.30897009e-01 -5.31642847e-02 -1.14292312e+00
-7.09517956e-01 -4.11924005e-01 -9.30101871e-01 -4.51771140e-01
1.08276200e+00 -9.82838869e-01 -8.76741529e-01 7.69646168e-01
-1.30111241e+00 -2.53366441e-01 -2.46968091e-01 2.24392086e-01
-5.95891118e-01 3.95147175e-01 -6.87920272e-01 -9.56640422e-01
-3.16776723e-01 -9.13161039e-01 1.64598751e+00 2.34695718e-01
-5.20223737e-01 -7.64290452e-01 -1.43837243e-01 6.63297474e-01
-8.55785385e-02 8.33456099e-01 4.46656495e-01 6.17375523e-02
-6.02859914e-01 -5.52080750e-01 -1.42238975e-01 -1.38019234e-01
-1.00251466e-01 -1.37204425e-02 -7.72698343e-01 -2.81349838e-01
-5.60917914e-01 -3.32767874e-01 6.83641136e-01 2.22718969e-01
8.44418108e-01 -2.76592940e-01 -6.03685915e-01 5.29180646e-01
7.54967511e-01 -9.07429382e-02 8.27967405e-01 3.18438530e-01
1.00277412e+00 9.81587172e-01 5.60857117e-01 5.21652281e-01
3.09174567e-01 1.01996100e+00 2.28432864e-01 -2.55790770e-01
-2.34208286e-01 -6.14037812e-01 2.04673126e-01 3.21091115e-01
-5.03953099e-01 -3.28527242e-01 -1.05033851e+00 5.36542058e-01
-1.70642710e+00 -1.24433124e+00 -2.20244855e-01 2.18830323e+00
6.34824753e-01 2.24372745e-01 6.62177861e-01 2.30900738e-02
8.75465393e-01 1.63057104e-01 -4.78211939e-01 5.27103662e-01
-3.68700385e-01 -2.11986095e-01 2.09917709e-01 -1.39630556e-01
-1.38451302e+00 8.46084476e-01 6.41842794e+00 4.70118731e-01
-5.20658791e-01 -2.34637365e-01 3.52526844e-01 -2.80659527e-01
2.10622400e-01 -3.55491459e-01 -7.82707214e-01 5.42429626e-01
1.84085280e-01 -3.76752466e-02 1.65361628e-01 8.54193032e-01
3.49401347e-02 -2.37268824e-02 -1.28471351e+00 1.59387660e+00
1.59198657e-01 -1.18451953e+00 2.89843827e-01 -1.56603637e-04
7.82331526e-01 -4.56434160e-01 1.54192433e-01 2.00081438e-01
1.52411446e-01 -1.18560350e+00 1.00846231e+00 6.65517390e-01
8.11408699e-01 -7.15533793e-01 5.70664346e-01 5.61509848e-01
-1.20503187e+00 9.73787829e-02 -3.38879883e-01 -1.60572216e-01
5.80280781e-01 1.90061286e-01 -3.89854938e-01 3.24274600e-01
8.58779252e-01 8.31761122e-01 -7.89940000e-01 1.10557282e+00
-3.07388276e-01 8.79785642e-02 -2.95047224e-01 -3.74946184e-02
-3.19361947e-02 -6.51116893e-02 6.20011270e-01 1.30292344e+00
9.34101939e-02 2.90674180e-01 2.13109940e-01 7.41520345e-01
2.09961571e-02 -1.94705874e-01 -7.50972033e-01 4.61672433e-02
4.00626302e-01 1.21646416e+00 -9.47962463e-01 -3.00523013e-01
-1.33221552e-01 1.35915542e+00 -8.78250226e-02 4.16897327e-01
-9.76620138e-01 -2.74069160e-01 6.44093990e-01 6.23087108e-01
-7.03088790e-02 -3.96103650e-01 -6.36206791e-02 -1.27054167e+00
1.76725864e-01 -8.95153701e-01 2.98858583e-01 -1.06237411e+00
-1.33156025e+00 4.74557966e-01 3.11643481e-01 -1.31812215e+00
-3.75626981e-01 -7.44563103e-01 -3.26717705e-01 5.31440496e-01
-6.56811476e-01 -1.31616616e+00 -7.46424794e-01 5.89403093e-01
6.94974184e-01 -1.97672203e-01 5.14758587e-01 2.24620283e-01
-6.24564469e-01 5.18958449e-01 -2.93814033e-01 6.78949773e-01
8.05741012e-01 -9.70694244e-01 1.00390375e+00 6.87168360e-01
3.10075969e-01 5.44585526e-01 7.96212614e-01 -8.04409981e-01
-1.46574700e+00 -1.07388008e+00 3.23295593e-01 -1.10285044e+00
1.94002151e-01 -6.40587807e-01 -5.70991635e-01 6.72826052e-01
-6.01876527e-02 -6.31957799e-02 -1.36602586e-02 -3.35937172e-01
-2.54548520e-01 3.31291527e-01 -9.28682327e-01 1.07888949e+00
1.53856301e+00 -1.85648441e-01 -5.56795955e-01 5.05609214e-01
5.03406942e-01 -6.70395613e-01 -6.41424954e-01 3.73743534e-01
8.18822384e-01 -1.02657938e+00 1.53455365e+00 -6.04588866e-01
5.95795929e-01 -1.95199460e-01 1.39629737e-01 -1.00492489e+00
-3.24469835e-01 -6.61862075e-01 -3.85261416e-01 7.25638568e-01
-8.55171457e-02 -1.39504641e-01 1.06450510e+00 5.97030520e-01
2.30919957e-01 -6.10383809e-01 -6.34355366e-01 -7.66777635e-01
-2.39471033e-01 -1.49205759e-01 4.69154000e-01 7.10869014e-01
-3.76018763e-01 4.61187512e-01 -7.99273670e-01 -2.09371224e-01
1.02272761e+00 -9.92910042e-02 1.35241997e+00 -1.23292017e+00
-1.67338789e-01 -4.50141162e-01 -7.18057811e-01 -1.16173148e+00
-1.07283056e-01 -5.04099190e-01 -1.28406435e-01 -1.52882206e+00
3.50878030e-01 -6.56262785e-02 6.21788919e-01 2.32347995e-01
-2.18136311e-01 6.21763647e-01 4.43471760e-01 4.87202644e-01
-5.23919284e-01 5.83009422e-01 1.67684197e+00 -2.14696720e-01
-2.06799582e-02 -3.45017798e-02 -1.93391606e-01 1.06722593e+00
3.74395460e-01 1.39231030e-02 -9.15146917e-02 -3.59458983e-01
5.22430502e-02 2.77084243e-02 9.87687290e-01 -1.17955720e+00
1.36566192e-01 -2.59055912e-01 1.06204927e+00 -8.20495903e-01
8.47213984e-01 -7.49255538e-01 5.15700340e-01 5.22290349e-01
-9.16806310e-02 9.71325040e-02 -3.16438898e-02 7.21240044e-01
9.21834707e-02 3.28852922e-01 9.13508236e-01 -5.27409256e-01
-9.61973071e-01 5.00135899e-01 1.76268220e-01 3.68431270e-01
1.26342142e+00 -5.41881979e-01 -2.89324015e-01 -6.29281163e-01
-6.86046958e-01 2.93014258e-01 8.40121031e-01 7.24069476e-01
9.19166088e-01 -1.52323163e+00 -9.56670284e-01 2.56969124e-01
2.55599797e-01 2.66867638e-01 3.25184375e-01 4.97427255e-01
-7.96948135e-01 4.26730871e-01 -5.36708772e-01 -6.75414205e-01
-1.22876072e+00 7.72550642e-01 1.64405510e-01 6.86483681e-02
-1.01282024e+00 7.57138491e-01 6.33972287e-01 -2.42849827e-01
3.28249961e-01 2.78631240e-01 2.48135366e-02 -1.42290890e-01
7.46003509e-01 6.47120476e-01 -4.46230650e-01 -1.20612895e+00
-3.81936699e-01 6.94018781e-01 1.51275605e-01 -2.14081898e-01
1.05288470e+00 -8.75103474e-03 2.52865523e-01 3.83446664e-01
8.01474571e-01 -1.59496903e-01 -1.44994116e+00 -1.31907940e-01
-4.77227271e-01 -7.78546214e-01 -7.61263609e-01 -6.83016896e-01
-9.69625235e-01 1.04044807e+00 3.06734055e-01 -2.61584103e-01
8.70770276e-01 2.47394338e-01 9.55476224e-01 3.55313987e-01
8.72682393e-01 -9.54586089e-01 6.60403430e-01 1.82316586e-01
1.26021612e+00 -1.22649658e+00 1.74892843e-01 -3.88094217e-01
-8.71357858e-01 7.79861569e-01 8.88852894e-01 5.17321117e-02
-7.58313527e-03 -6.91102259e-03 -2.09020779e-01 -2.11761907e-01
-3.12607974e-01 -1.52174547e-01 5.17962575e-01 8.67450416e-01
2.66846955e-01 1.31780475e-01 1.77914262e-01 5.29873312e-01
-4.96524483e-01 -3.63028616e-01 3.10553819e-01 8.87853980e-01
-2.75597245e-01 -5.96874475e-01 -8.86566997e-01 2.45490238e-01
-7.23398179e-02 3.34103078e-01 -7.81889260e-01 1.07061863e+00
1.66225359e-01 6.86937332e-01 -8.82476717e-02 -3.00984859e-01
7.61877835e-01 -1.23733863e-01 7.62691736e-01 -4.19902861e-01
-2.63573557e-01 -1.51326537e-01 -5.74480109e-02 -4.95051593e-01
-3.16450268e-01 -5.81343353e-01 -9.38206434e-01 -5.58912456e-01
-1.41620738e-02 -3.78683567e-01 3.29679847e-01 4.94592249e-01
2.28874087e-01 3.21872234e-01 7.19834343e-02 -1.59576392e+00
-1.66575670e-01 -9.09842193e-01 -5.64287007e-01 9.07524228e-01
1.06045254e-01 -9.13840532e-01 -8.29730704e-02 3.54481995e-01] | [7.234841346740723, -0.8071698546409607] |
166df0f1-041f-4774-b2c7-1bdcab30cef1 | camb-at-cwi-shared-task-2018-complex-word | null | null | https://aclanthology.org/W18-0520 | https://aclanthology.org/W18-0520.pdf | CAMB at CWI Shared Task 2018: Complex Word Identification with Ensemble-Based Voting | This paper presents the winning systems we submitted to the Complex Word Identification Shared Task 2018. We describe our best performing systems{'} implementations and discuss our key findings from this research. Our best-performing systems achieve an F1 score of 0.8792 on the NEWS, 0.8430 on the WIKINEWS and 0.8115 on the WIKIPEDIA test sets in the monolingual English binary classification track, and a mean absolute error of 0.0558 on the NEWS, 0.0674 on the WIKINEWS and 0.0739 on the WIKIPEDIA test sets in the probabilistic track. | ['Sian Gooding', 'Ekaterina Kochmar'] | 2018-06-01 | null | null | null | ws-2018-6 | ['complex-word-identification'] | ['natural-language-processing'] | [-0.46704742 0.10509726 -0.31144825 -0.14912735 -0.9989537 -0.82055056
0.78671753 0.2623794 -1.2306752 1.0890563 0.13030738 -0.4339785
-0.30762047 -0.41061085 -0.41613773 -0.38474357 -0.24123041 0.75750357
0.17366843 -0.47076482 0.10019413 -0.20689127 -1.1906964 0.11315796
1.0172603 0.94888055 0.14957097 0.8713519 0.05701904 0.38821495
-0.67513245 -0.8935212 0.08001247 0.365171 -0.70273983 -1.0160295
0.8226887 0.05294955 -0.6494509 1.330904 0.8573754 -0.06228754
0.9068079 -1.1958337 -0.7643256 1.2323583 -0.29730272 0.4333681
0.39728978 -0.07073188 1.4187859 -1.1274228 0.7931108 0.8570987
1.31063 0.5286855 -1.1029592 -1.063433 -0.10669 0.29579055
-1.7452385 -0.7530102 -0.31724343 -0.5392936 1.4503585 0.24557595
0.30802038 1.2165296 0.23771174 0.6837106 1.167053 -0.4665222
-0.25591636 0.17085347 0.83117914 1.0252118 0.7473054 0.09454569
-0.6124558 -0.01792735 -0.02677689 -0.8778792 -0.45640278 0.45981678
-1.3285116 0.63811636 0.20472614 0.21672258 -0.08518226 -0.02119994
0.46724793 0.34381533 0.6928493 0.7104714 -0.9187592 -0.37801477
-0.6617225 0.16201791 1.2534842 0.9701505 0.29941416 -0.10474351
-0.22944188 1.2823987 0.26893157 0.92571276 0.7108314 -0.34493926
0.76342714 -0.0299259 0.14582343 -0.6217198 -0.74914694 -0.7294486
-0.5190307 -0.2765836 0.39260268 -0.26307395 -0.85532445 1.8220899
-0.28397137 -0.08014867 0.09004778 0.35918358 1.3463869 0.6014006
0.53806555 0.20456913 1.464348 -1.2179648 -0.71283543 -0.21982887
1.0599353 -1.1162903 1.0390522 0.42760128 -0.7354471 -0.4639971
-1.3173764 0.16269854 -0.78100944 0.60270506 0.36747202 0.7912382
-0.95946103 0.4808276 -0.3279236 -0.35814074 0.24103999 0.2419682
-0.724593 -0.2521077 -1.5967232 1.4157697 0.5057348 -0.3115632
-0.6978342 -0.89795166 -0.69496065 -0.02914554 -0.03599687 -0.23904905
1.2857826 -0.1410303 -1.1124254 1.2731189 0.24467237 -0.59431136
0.5858364 -0.399678 -0.7919975 -0.29728124 0.46309254 0.65696174
-0.05254242 -0.8049493 -0.9408152 0.05561259 -0.22738889 0.15816066
-0.3476415 0.23682334 -0.38545108 -0.82050306 -0.34396645 -1.1107976
0.06546437 -0.3708239 -0.7602506 -0.6027779 -0.01343533 -1.1262549
1.3763393 -1.994679 -0.01450507 -0.02216535 0.24580562 0.3936964
-0.31589895 0.19011585 -0.17516695 0.4134853 0.02186994 -0.41167313
0.3450799 -0.14357354 -0.0979286 0.5612334 -0.06479151 0.8062191
-0.9761468 -0.3312027 -0.17304386 0.04059482 -0.29872808 0.13734023
0.21762194 -0.44223836 0.18827023 0.2775701 0.6077686 0.0765528
-0.11853179 -0.14391927 -0.45903623 0.8345691 -1.0369184 1.3841997
-0.5702952 0.92105085 0.00838414 -0.56298643 0.615544 0.28097895
0.15379477 -0.5614255 -0.03588735 0.5352884 0.522768 -0.27700937
0.61725783 0.22374025 -0.46219125 0.58113736 0.5056547 0.05325332
0.44842464 0.34847772 1.1441498 -0.12982646 0.4171775 -0.9946316
0.29770947 -0.03142723 0.45672545 0.8863621 -0.31432876 0.5397942
0.11840892 -0.4068116 -0.89846927 -1.4034722 -0.6772595 1.3013505
-0.10006873 -1.1016449 -0.6248601 -0.63179535 0.09741628 0.8743952
-0.57357293 -0.19389562 -0.16614957 -1.0057858 1.2438511 0.35027203
0.6584371 -0.782649 0.19037794 0.0766463 -0.32720977 -1.1989115
-0.6172574 0.43452004 0.07696967 -0.9666014 -0.7102285 -1.048424
0.06471855 -0.11076496 1.2312962 -0.16914126 -0.22847165 0.23662704
-0.388795 -0.5184801 -0.24009041 0.62407637 0.51104975 -0.54286337
0.32645747 0.26003107 0.06166659 0.22534476 -0.20748058 0.02485104
0.253662 1.2517996 0.1276293 -0.20989618 0.50994617 -0.87423617
0.72532505 -0.6985501 -0.5383268 0.61473805 -1.0281528 -0.11302373
0.32668862 -0.4312526 -0.7498799 -0.15842459 -0.32960752 0.38786614
0.1809518 0.7101556 0.05509017 -0.00711053 0.9033743 0.09584444
-0.4459929 -0.54764277 0.25893813 1.0553676 0.6417808 -0.66259503
0.6128961 -0.31761473 -0.73488927 -1.0058492 -0.9467049 -0.4365957
-0.4486386 0.10274918 0.97330004 -1.2036288 -0.44349498 0.78612375
-1.3040698 -0.3848659 0.04278453 0.52973986 -0.16521959 0.13610026
-0.6251872 -0.4742259 -0.72431856 -1.0456754 0.83475524 -0.01495617
-0.8991099 -1.1428111 0.25285915 0.16588558 0.5942064 -0.43032402
0.7272837 -1.0785387 0.13724403 -0.41801783 -0.4326464 0.3764526
-0.1678125 0.05606651 -0.8799897 -0.19903384 -0.49375778 -0.40881446
1.1044195 0.11948097 0.86922264 -0.09384575 -0.5720614 0.5481447
1.0739009 -0.11466631 0.3295543 0.47400516 0.6568605 0.47205308
0.3597058 0.07180189 0.8333949 0.8397547 -0.26370016 0.40182325
-0.5445301 -0.21572591 0.38810325 1.1504214 0.12385942 -0.40222368
-1.5565965 0.7376654 -1.5464106 -0.7408536 -0.38075444 2.0116146
1.1561527 0.2504646 -0.26915509 -0.09755721 0.88532746 -0.0623987
-0.00892372 -0.3990708 -0.67138225 0.7479228 0.68629247 0.73103744
-1.5597492 1.2729204 6.7588034 1.0502917 -0.57964295 0.37782547
0.2700403 0.10269 -0.05897597 -0.36716792 -1.2515447 0.6586108
1.3240637 -0.61761117 0.33672333 0.6884596 -0.71519643 -0.13571443
-0.8566994 1.1860312 0.23042437 -1.012182 -0.16696946 -0.06324226
0.8659543 0.87897897 0.17269728 0.7139803 0.686331 -1.3384662
0.8852393 0.35136712 1.079214 -0.5168642 0.9665131 0.14409842
-0.8922346 0.15569882 -0.24193068 -0.1066166 -0.23503786 0.58986026
-0.7150993 0.12686008 1.0225909 0.63834757 -0.57673347 1.2267447
-0.66072315 0.811839 -0.5751071 -0.33666435 0.10425094 0.31244257
0.5779009 1.7620746 0.06133125 -0.21428944 0.02536008 0.07791802
-0.47424445 0.26574543 -0.4074885 -0.07208396 0.70591575 1.2412499
-0.20243965 -0.4538534 -0.29089606 0.7284421 0.67760843 0.19730125
-0.8053855 -1.0950079 0.74182695 -0.4328313 0.09374755 -0.30065635
-0.4916881 -1.437927 -0.24680495 -0.6728855 0.67846495 -0.59049004
-1.763171 0.8950762 -0.16660914 -0.81654567 -0.07626473 -1.1424742
-0.30203912 1.1912607 -1.230231 -0.71263874 -0.11467881 0.16967905
0.11548989 -0.78500146 1.2836686 0.7307112 -0.729287 1.2697079
0.44879654 0.47008294 1.1805048 -1.3128568 0.76852334 0.6717156
0.3877632 0.78742814 0.48610103 -0.4641713 -0.91056347 -0.8869566
1.4985362 -0.55721724 1.0715029 -0.40948713 -0.6003459 0.72995776
0.45562607 -0.14207406 0.88105524 0.6088599 -0.82093775 0.09612393
-1.1717174 0.49329507 1.1802458 -0.86947745 -0.6500809 0.74409264
0.4989557 -0.5221987 -1.125122 0.35485542 0.9362588 -0.21413356
0.77257085 -0.85518837 0.3766833 0.05295001 -0.39628705 -1.9725138
-0.5095992 -0.01664317 0.18918386 1.1680574 1.1354585 -1.0622724
0.30844656 0.25060362 -0.48841852 -0.41035324 -1.3674557 -0.839863
0.39578155 -0.7563572 -0.00886621 1.3499198 0.27693325 0.4408367
-0.07796999 -0.10373078 0.671163 -0.4094078 0.46254992 -1.2135166
-0.21892917 -0.6423581 -0.32501888 -0.7696647 0.5184808 -1.5491285
0.13069165 -1.4895302 0.5172496 -0.62583345 -0.5314164 0.70977163
-0.47537267 0.6434949 0.16534911 0.16710794 -0.6116065 0.31411868
0.6401737 -0.42140812 0.5163724 -0.33107787 -0.84992725 0.5137218
1.0044215 -0.6129775 0.19263923 -0.87980855 0.1950059 -0.78581655
-0.16460696 -0.99515986 0.29181492 0.488144 0.24409229 -0.21555199
0.23484682 -0.06505401 -0.16604851 0.83444756 -0.3250325 0.22584866
0.55551594 0.06097071 0.1702317 -0.3296974 0.67467034 0.2754006
-0.7225413 0.07013932 -0.47964442 0.49080446 0.6967442 0.36073458
-0.7312788 -0.08005393 -0.66816366 0.46034032 -0.14872214 0.91688854
0.18851715 -1.4156712 -1.1298968 0.01166828 0.47501782 -0.63031334
-0.19956204 0.6093015 -0.6609962 0.7397144 -0.23962334 -0.3247195
-1.1711512 -0.2684176 0.36531538 -0.25223246 -0.02297607 1.3396598
-0.285772 -0.90729624 0.14003336 -0.01277205 -0.18211824 0.32664028
0.47576177 0.77104163 0.22415031 -0.8162589 -0.4546969 0.47059956
-0.41239923 -0.24730745 1.2578052 0.21849036 -0.30268717 0.57757705
1.3476057 0.10463087 -0.2915645 -0.33669248 0.35424107 0.15473339
0.27071002 -1.2188225 -0.6721276 0.52155346 0.7217297 0.1871929
0.33824778 -0.01756066 0.70054907 0.5168542 0.5367056 -1.3478813
-0.617674 1.2476673 0.8802397 -1.252005 0.011429 -0.4935068
-0.43736055 1.0232928 0.5130692 -0.02609858 1.0221987 0.5518043
0.05511676 0.04969562 -0.8179762 -0.14309472 0.60403305 0.44785252
0.68709284 0.48261824 -0.79863983 1.0841347 -0.8888839 -0.6291777
0.31952164 0.34783214 -0.24206273 -0.9060855 -0.03295476 0.6264589
-0.50197756 -0.67798835 -0.48051807 0.8899083 0.08971544 0.8914203
0.12377337 -0.6516669 0.46530947 0.5552303 0.42891416 -0.8630732
-0.6548773 -0.6045946 0.8191888 -0.3463046 0.16283463 -0.7098001
-1.1184407 -0.32820585 -0.41709182 0.29684603 0.81461674 0.66671556
0.20632218 0.37931767 -0.03624863 -0.4679283 -0.6965566 -1.3762884
-0.3687077 0.2769347 0.10374874 -0.6826685 -0.54046845 -0.3989629 ] | [10.492707252502441, 10.480818748474121] |
2a32d207-e13c-4fe8-b870-4b5b936d947e | multiview-deep-learning-for-predicting | 1712.08091 | null | http://arxiv.org/abs/1712.08091v1 | http://arxiv.org/pdf/1712.08091v1.pdf | Multiview Deep Learning for Predicting Twitter Users' Location | The problem of predicting the location of users on large social networks like
Twitter has emerged from real-life applications such as social unrest detection
and online marketing. Twitter user geolocation is a difficult and active
research topic with a vast literature. Most of the proposed methods follow
either a content-based or a network-based approach. The former exploits
user-generated content while the latter utilizes the connection or interaction
between Twitter users. In this paper, we introduce a novel method combining the
strength of both approaches. Concretely, we propose a multi-entry neural
network architecture named MENET leveraging the advances in deep learning and
multiview learning. The generalizability of MENET enables the integration of
multiple data representations. In the context of Twitter user geolocation, we
realize MENET with textual, network, and metadata features. Considering the
natural distribution of Twitter users across the concerned geographical area,
we subdivide the surface of the earth into multi-scale cells and train MENET
with the labels of the cells. We show that our method outperforms the state of
the art by a large margin on three benchmark datasets. | ['Bruno Cornelis', 'Tien Huu Do', 'Nikos Deligiannis', 'Duc Minh Nguyen', 'Evaggelia Tsiligianni'] | 2017-12-21 | null | null | null | null | ['multiview-learning'] | ['computer-vision'] | [-1.87193736e-01 -5.81659302e-02 -5.00494242e-01 -2.14577526e-01
-4.24286067e-01 -3.63998801e-01 9.83624101e-01 6.25784814e-01
-6.47955954e-01 6.31339967e-01 5.20471215e-01 -9.95110795e-02
-1.22972906e-01 -1.06471205e+00 -5.90780914e-01 -5.43594718e-01
2.05578953e-02 3.33202183e-01 1.17544793e-01 -3.30088258e-01
4.02971238e-01 4.00919139e-01 -1.48461854e+00 2.10913830e-02
6.60949409e-01 1.29362428e+00 -1.42377198e-01 2.86388807e-02
-3.86682749e-01 5.84866643e-01 -3.60575736e-01 -1.17809966e-01
-1.30454265e-02 7.32818544e-02 -6.62944138e-01 -6.79497942e-02
2.53379613e-01 5.00020869e-02 -4.67926711e-01 7.22458363e-01
5.60867786e-01 2.40662903e-01 5.63261747e-01 -1.35901093e+00
-3.82623672e-01 9.18359756e-01 -9.46743131e-01 3.09187710e-01
-3.48533085e-03 -4.09167439e-01 1.01181126e+00 -8.83930266e-01
4.98281151e-01 1.10423958e+00 7.33964026e-01 -1.99260518e-01
-7.79569030e-01 -7.33124316e-01 4.59999472e-01 3.72318663e-02
-1.65521240e+00 -3.83876652e-01 6.05924726e-01 -7.23468184e-01
5.74833870e-01 1.41915411e-01 4.72767025e-01 1.15755916e+00
1.16673812e-01 8.13030541e-01 8.65101635e-01 -1.22489110e-01
1.81912139e-01 9.18257013e-02 -4.45019081e-02 4.13651258e-01
2.22264230e-01 -4.63062167e-01 -5.18409967e-01 -2.73697644e-01
2.97011226e-01 4.77580249e-01 -1.63386688e-02 -3.23500074e-02
-1.43592882e+00 9.03470933e-01 8.02027881e-01 6.14070535e-01
-3.84570718e-01 8.47101584e-02 5.50509751e-01 -1.66362807e-01
1.02598691e+00 3.39455575e-01 -1.94548801e-01 2.10003808e-01
-1.20359838e+00 1.52751386e-01 6.98058784e-01 5.73570788e-01
9.94952977e-01 1.84728056e-02 -1.60507346e-03 7.46922672e-01
3.36332113e-01 2.92234063e-01 9.38826323e-01 -2.37566009e-01
5.48515379e-01 9.76493120e-01 2.45761573e-02 -1.69054413e+00
-7.76115835e-01 -7.21708834e-01 -1.04902124e+00 -5.52299440e-01
2.66124994e-01 -5.22689402e-01 -4.25107598e-01 1.57621646e+00
6.16393447e-01 6.01292908e-01 -1.90374389e-01 3.98087978e-01
9.02867258e-01 8.14236462e-01 2.03859270e-01 1.38149977e-01
1.25669205e+00 -9.60998237e-01 -5.54107845e-01 -2.56792277e-01
7.21645355e-01 -3.45616370e-01 5.71295381e-01 3.90996076e-02
-6.16863310e-01 -2.27477103e-01 -8.58057141e-01 8.44144076e-02
-1.11993241e+00 7.05219954e-02 5.40065110e-01 3.23126823e-01
-9.34248030e-01 4.74276721e-01 -4.67238367e-01 -5.25985420e-01
3.83205265e-01 2.83468783e-01 -3.90010536e-01 2.72606283e-01
-1.46790719e+00 4.96152431e-01 4.80569065e-01 1.91879660e-01
-2.77591616e-01 -4.22125757e-01 -8.03130209e-01 2.00482726e-01
3.45916480e-01 -4.64302510e-01 9.12622988e-01 -9.38822985e-01
-1.03849339e+00 7.80514538e-01 -3.03937756e-02 -4.83838886e-01
5.50849974e-01 -1.17528632e-01 -5.82705379e-01 -2.85537809e-01
3.37308109e-01 4.51888770e-01 6.76435411e-01 -1.03544450e+00
-1.03571546e+00 -4.06485111e-01 5.77232242e-03 9.37040001e-02
-8.51640701e-01 -1.52139455e-01 -3.55836749e-01 -6.31833017e-01
-5.11066429e-02 -1.00009334e+00 -1.36537716e-01 -4.04710084e-01
-7.74967372e-01 -6.48859560e-01 8.96533191e-01 -3.25985610e-01
1.52983677e+00 -2.27989531e+00 1.51899541e-02 2.46450722e-01
6.50273204e-01 3.16237509e-02 1.45380467e-01 8.88674676e-01
1.40502259e-01 2.61831611e-01 2.25497499e-01 -5.51403522e-01
1.19429342e-01 -1.75218076e-01 -3.99626464e-01 7.46807218e-01
-3.57449949e-01 7.36312270e-01 -9.98368263e-01 -4.14429933e-01
-5.55286147e-02 5.13335943e-01 -4.76720810e-01 -7.19626695e-02
-1.13658600e-01 3.85907292e-01 -7.69191265e-01 4.91089076e-01
5.21661758e-01 -7.18322456e-01 2.61410087e-01 -1.54782116e-01
-3.89775068e-01 -1.25685299e-03 -9.88940179e-01 1.28830075e+00
-5.83429337e-01 8.09640110e-01 1.21066375e-02 -9.04439092e-01
8.95017743e-01 1.48780510e-01 7.54751146e-01 -6.16347253e-01
3.38431180e-01 3.24949592e-01 -3.66556764e-01 -4.20472354e-01
8.67849588e-01 5.08568212e-02 -4.03238833e-01 6.39748335e-01
-1.64031595e-01 7.35401809e-01 -1.09514281e-01 1.20117381e-01
5.85062623e-01 -3.60663146e-01 7.01937497e-01 -2.91110873e-01
5.10643125e-01 -4.63360786e-01 3.37955981e-01 7.20086753e-01
1.18681677e-02 3.33834857e-01 4.34776813e-01 -4.48705912e-01
-6.27011776e-01 -5.02291560e-01 -9.70751196e-02 1.68796384e+00
2.05628917e-01 -4.06556696e-01 -5.25420070e-01 -7.16223598e-01
3.11906070e-01 3.78725678e-01 -9.29379702e-01 2.53853142e-01
-5.81744492e-01 -9.72862601e-01 5.49267948e-01 3.34964216e-01
6.15089357e-01 -9.15896177e-01 -1.49796292e-01 1.04295485e-01
-3.96998465e-01 -1.05856252e+00 -3.04132789e-01 -2.79343069e-01
-4.96744990e-01 -7.92536438e-01 -6.49704039e-01 -6.43098414e-01
4.27648664e-01 3.62828970e-01 8.84944439e-01 7.63077140e-02
2.38761976e-01 3.36884975e-01 -3.78339201e-01 -4.16615903e-01
1.56452745e-01 9.54451323e-01 2.45071918e-01 8.71818483e-01
2.43814722e-01 -6.52308285e-01 -7.77633369e-01 2.68883765e-01
-1.01803720e+00 -1.64063692e-01 2.68484086e-01 4.41582233e-01
1.84730053e-01 -1.01248577e-01 1.05818129e+00 -1.21421862e+00
5.81066430e-01 -1.52274179e+00 -3.35009962e-01 -4.73979115e-02
-5.19018888e-01 -1.98246673e-01 4.86523896e-01 -2.79612482e-01
-6.11430824e-01 -9.19670165e-02 -1.30187899e-01 3.61556374e-02
-1.13481700e-01 1.12237263e+00 -6.96025090e-03 2.13810652e-01
5.17431915e-01 1.35290563e-01 -3.43090028e-01 -6.22095108e-01
3.96262527e-01 1.01798534e+00 9.93025675e-02 -2.76553333e-01
6.07709050e-01 7.14564621e-01 -2.13025689e-01 -1.05718410e+00
-8.82044673e-01 -7.19047427e-01 -3.91684413e-01 -2.57269621e-01
7.05471218e-01 -1.06516612e+00 -9.75386143e-01 6.37338221e-01
-9.50344503e-01 -7.24925920e-02 8.44957605e-02 4.23765868e-01
-5.26156984e-02 1.49254084e-01 -3.72116566e-01 -6.05256319e-01
-2.04200909e-01 -9.04947042e-01 1.10311115e+00 2.00608119e-01
-1.12816148e-01 -1.43655431e+00 1.60915583e-01 1.73124582e-01
7.82749951e-01 5.59897661e-01 6.87534511e-01 -1.34981573e+00
-4.65495527e-01 -5.12632847e-01 -4.98640835e-01 -2.87709177e-01
2.53675401e-01 -2.90692031e-01 -1.13585627e+00 -4.49398637e-01
-3.51968408e-01 -2.29645327e-01 1.04706228e+00 2.56945938e-01
1.25714791e+00 -5.73841035e-01 -8.33489954e-01 5.14511883e-01
1.43242800e+00 -2.85350382e-01 2.84776568e-01 6.91049337e-01
1.05998635e+00 8.19603860e-01 2.24934086e-01 8.72447252e-01
8.98503304e-01 5.60579062e-01 7.13536859e-01 -2.17342049e-01
5.55807710e-01 -3.81663710e-01 -2.53942306e-03 8.35177004e-01
7.81257749e-02 -7.42899120e-01 -1.05877018e+00 7.62778938e-01
-2.03416419e+00 -9.39867854e-01 1.28001362e-01 1.87719727e+00
3.90398443e-01 -8.12372267e-02 3.03004026e-01 -1.00008547e-01
9.60067451e-01 7.33345866e-01 -4.56407487e-01 1.63583457e-01
-3.22299376e-02 -4.06834513e-01 8.00552726e-01 4.35726970e-01
-1.40179551e+00 6.87629163e-01 5.23706055e+00 8.31266999e-01
-1.55620039e+00 1.10522971e-01 7.76884079e-01 7.18223825e-02
-2.60546625e-01 -4.01925802e-01 -1.16046083e+00 7.07656503e-01
9.69099224e-01 -3.15380991e-01 3.11351102e-02 7.08992183e-01
1.60782561e-01 1.53269067e-01 -8.40466499e-01 9.25370812e-01
2.25233436e-01 -1.62509394e+00 -7.02032074e-02 4.10198480e-01
7.40893841e-01 5.80549359e-01 3.20675313e-01 3.33259493e-01
-2.36411160e-03 -1.02225828e+00 6.82942629e-01 6.67322993e-01
5.80320001e-01 -5.97414553e-01 8.35342228e-01 5.28097332e-01
-1.41077852e+00 -2.70496011e-01 1.81904837e-01 2.70009898e-02
1.09232008e-01 5.84547222e-01 -8.65173042e-01 5.14315486e-01
4.82699722e-01 1.16250980e+00 -7.11607337e-01 1.18160892e+00
3.46890926e-01 6.57043934e-01 -4.04347926e-01 -5.53872474e-02
6.93267643e-01 -1.54348910e-01 3.48435551e-01 1.34685040e+00
5.69053054e-01 -4.74074900e-01 3.92862767e-01 4.16940808e-01
-6.43768489e-01 5.02013206e-01 -8.96952748e-01 -4.40494483e-03
6.84775233e-01 1.45009398e+00 -8.12768936e-01 -2.07953036e-01
-5.36808729e-01 4.85206425e-01 4.16628242e-01 3.92393827e-01
-6.18032098e-01 -3.81682038e-01 3.09401155e-01 4.82532799e-01
1.65675431e-01 -1.87269822e-01 3.21477428e-02 -1.02290189e+00
-2.75722712e-01 -5.55809617e-01 3.89308035e-01 -2.24773183e-01
-1.54441726e+00 6.85774565e-01 -3.81835625e-02 -1.31651282e+00
-3.23907137e-01 -3.56918275e-01 -6.23770416e-01 5.52992404e-01
-1.66531467e+00 -1.34454811e+00 -4.34928805e-01 4.83928740e-01
3.09673399e-01 -3.29119295e-01 4.58234519e-01 5.68227828e-01
-7.15045869e-01 5.75129151e-01 5.46727419e-01 4.16618437e-01
6.16876245e-01 -1.11862457e+00 6.21542871e-01 3.86606127e-01
-1.32325023e-01 5.46613932e-01 3.16919118e-01 -5.53065181e-01
-1.00310624e+00 -1.42162430e+00 1.28257275e+00 -2.97305197e-01
1.01264381e+00 -3.55506390e-01 -6.40106797e-01 1.01283085e+00
3.77081752e-01 1.31387785e-01 9.43639219e-01 3.31695080e-01
-2.78432280e-01 -2.72198707e-01 -9.50250268e-01 6.00935161e-01
5.57386696e-01 -3.71029854e-01 -1.46086589e-01 4.50992614e-01
5.16414940e-01 -1.66389257e-01 -7.55765438e-01 1.03480764e-01
3.99085701e-01 -7.28691697e-01 8.50783288e-01 -4.08155203e-01
1.65116057e-01 -5.06523065e-02 -1.53795332e-01 -1.39158201e+00
-2.03100368e-01 -4.61082071e-01 8.43963400e-02 1.37649524e+00
4.26769912e-01 -1.05074120e+00 6.94209516e-01 1.40689701e-01
1.56574845e-01 -7.05807328e-01 -8.87210667e-01 -1.91772342e-01
1.37796476e-01 -4.01680022e-02 7.67108977e-01 1.28641117e+00
8.89562517e-02 4.92096901e-01 -6.53839231e-01 -5.54070957e-02
2.86652863e-01 1.06422558e-01 6.09215617e-01 -1.74897456e+00
1.60357371e-01 -4.58399981e-01 -2.50740558e-01 -1.06991816e+00
3.83856326e-01 -1.03241336e+00 -3.18279088e-01 -1.48004425e+00
-5.08638397e-02 -6.66253805e-01 -3.06972623e-01 3.37078989e-01
1.34919137e-01 3.60412151e-01 4.41982895e-02 3.72036576e-01
-9.72520649e-01 5.57598829e-01 7.45899141e-01 -3.10131282e-01
-2.25533366e-01 1.54770479e-01 -8.45692158e-01 8.71663213e-01
8.70596528e-01 -4.51604068e-01 -2.18307078e-01 -3.50654691e-01
9.52377379e-01 -1.98715657e-01 1.57175690e-01 -9.40607905e-01
5.52104652e-01 5.17786592e-02 1.74438030e-01 -5.66672921e-01
2.48867705e-01 -7.68305779e-01 -1.89951509e-01 -4.18850183e-02
-3.26541245e-01 2.72462606e-01 -5.98281547e-02 1.10828626e+00
-3.99250120e-01 1.96677998e-01 4.77011651e-01 -9.36898738e-02
-5.89120448e-01 5.21918476e-01 -5.72073936e-01 1.56197816e-01
8.69775593e-01 -1.59708317e-03 -4.78258133e-01 -5.16524553e-01
-6.42565668e-01 3.74890119e-01 1.79580227e-01 6.24432206e-01
2.74646252e-01 -1.34686625e+00 -6.15806937e-01 9.07918904e-03
2.94910699e-01 -1.81684688e-01 2.85134703e-01 1.02897775e+00
-1.92343026e-01 4.39333141e-01 2.88184304e-02 -4.65634525e-01
-7.48241246e-01 4.00886267e-01 3.52110267e-01 -2.86148816e-01
-3.89146000e-01 5.48173785e-01 2.04947174e-01 -5.22129118e-01
3.13174635e-01 -2.81817000e-02 -8.75472844e-01 9.69829202e-01
4.51818287e-01 4.97694343e-01 2.55495552e-02 -1.16871357e+00
-2.80818552e-01 3.60979915e-01 -1.02878742e-01 2.12750599e-01
1.43788862e+00 -3.97570968e-01 -1.76520303e-01 8.41813326e-01
1.45526206e+00 2.44473636e-01 -6.51824653e-01 -6.48973823e-01
2.46260524e-01 -2.31306985e-01 1.11603841e-01 -2.64840543e-01
-1.20089483e+00 7.96655357e-01 4.08770382e-01 7.70698845e-01
7.12165177e-01 5.91364242e-02 1.01068199e+00 4.12133485e-01
1.72741503e-01 -1.07278192e+00 -1.03298295e-02 6.04702115e-01
5.98855317e-01 -1.24583125e+00 -1.38581812e-01 -6.22034371e-02
-5.54503560e-01 9.00734186e-01 4.07326162e-01 1.57321896e-02
1.22314620e+00 -1.46119535e-01 -2.70624701e-02 -3.26706707e-01
-4.09698009e-01 -3.41511071e-01 2.51171440e-01 7.17545822e-02
5.31449378e-01 5.34959584e-02 -1.48259088e-01 4.97971773e-01
-2.19624229e-02 -1.91155091e-01 3.78212392e-01 7.10698128e-01
-5.11786699e-01 -8.39135289e-01 -1.67054936e-01 5.55970728e-01
-6.56439304e-01 -8.82608742e-02 -1.64298996e-01 7.77356982e-01
1.38445064e-01 8.35989237e-01 2.82449394e-01 -4.62036550e-01
3.02914623e-02 -6.32468238e-02 -4.27061975e-01 -4.55774963e-01
-8.14350665e-01 -6.72203526e-02 -5.71771860e-02 -2.36537680e-01
-5.71413457e-01 -5.23479402e-01 -9.20040905e-01 -5.84101439e-01
-1.81413978e-01 2.15922490e-01 7.26540208e-01 1.17081606e+00
6.78468287e-01 2.51331508e-01 8.77306998e-01 -1.19402719e+00
-1.51344672e-01 -1.07079613e+00 -6.92601621e-01 1.92698255e-01
7.36573339e-01 -7.29989111e-01 -4.35582846e-01 -4.38947678e-01] | [9.982928276062012, 6.805452823638916] |
2c955ae0-572b-43f3-b16b-60f51e1262e3 | example-guided-learning-of-stochastic-human | null | null | https://link.springer.com/article/10.1007/s00521-022-07947-2 | https://link.springer.com/article/10.1007/s00521-022-07947-2 | Example-guided learning of stochastic human driving policies using deep reinforcement learning | Deep reinforcement learning has been successfully applied to the generation of goal-directed behavior in artificial agents. However, existing algorithms are often not designed to reproduce human-like behavior, which may be desired in many environments, such as human–robot collaborations, social robotics and autonomous vehicles. Here we introduce a model-free and easy-to-implement deep reinforcement learning approach to mimic the stochastic behavior of a human expert by learning distributions of task variables from examples. As tractable use-cases, we study static and dynamic obstacle avoidance tasks for an autonomous vehicle on a highway road in simulation (Unity). Our control algorithm receives a feedback signal from two sources: a deterministic (handcrafted) part encoding basic task goals and a stochastic (data-driven) part that incorporates human expert knowledge. Gaussian processes are used to model human state distributions and to assess the similarity between machine and human behavior. Using this generic approach, we demonstrate that the learning agent acquires human-like driving skills and can generalize to new roads and obstacle distributions unseen during training. | ['Armin Biess', 'Avinoam Borowsky', 'Rotem Duffney', 'Ran Emuna'] | 2022-12-23 | null | null | null | neural-computing-and-applications-2022-12 | ['unity'] | ['computer-vision'] | [-7.24584563e-03 3.81681263e-01 3.09676290e-01 -3.61009389e-01
-4.01656896e-01 -2.36732751e-01 8.71571302e-01 5.51746376e-02
-7.13737071e-01 1.05626607e+00 -2.55275697e-01 -5.10335624e-01
-1.85579821e-01 -9.61332560e-01 -9.13830340e-01 -6.95593297e-01
-2.16260329e-01 9.60180283e-01 2.18535751e-01 -6.62136972e-01
4.21851724e-02 4.47648764e-01 -1.76464033e+00 -3.17665935e-01
9.81205523e-01 2.45247647e-01 5.92641592e-01 1.08215165e+00
1.98266909e-01 1.08956158e+00 -4.42867845e-01 -6.79500997e-02
3.11045617e-01 -4.51331079e-01 -5.91741145e-01 2.95570940e-01
-3.75401586e-01 -2.37069547e-01 -6.35890841e-01 8.68996263e-01
3.70190382e-01 4.06937957e-01 9.99321699e-01 -1.53943610e+00
-3.97927642e-01 2.52662033e-01 -5.02074845e-02 -4.39768881e-02
1.42393902e-01 9.81420517e-01 3.45917881e-01 -7.39257336e-02
5.18647850e-01 1.44001079e+00 2.06704885e-01 8.36996496e-01
-1.22956192e+00 -2.70902723e-01 2.80121285e-02 2.33408749e-01
-8.23224187e-01 -6.48907339e-03 5.15460134e-01 -7.10226774e-01
6.49289250e-01 -6.31486356e-01 6.81487620e-01 1.53398323e+00
6.23583019e-01 7.80823767e-01 1.34463906e+00 -2.00913593e-01
8.43039870e-01 2.57183522e-01 -2.86660194e-01 5.33648670e-01
7.24622458e-02 8.54884565e-01 -7.04990793e-03 -1.69710610e-02
6.06223524e-01 -1.58058643e-01 2.26550519e-01 -7.53680289e-01
-1.04183221e+00 9.16657209e-01 3.87415022e-01 -3.86395566e-02
-1.10944033e+00 5.54587960e-01 2.32792228e-01 6.60556734e-01
-2.37572506e-01 4.53285724e-01 -4.12637353e-01 -3.78298551e-01
-1.99563339e-01 9.77960527e-01 1.01884484e+00 9.77863491e-01
9.57801878e-01 4.04507160e-01 -2.36813039e-01 4.42977279e-01
9.57931206e-02 9.25294876e-01 4.61739957e-01 -1.29670525e+00
5.06458879e-02 1.40946761e-01 6.65191889e-01 -6.14301324e-01
-5.14460802e-01 -2.84669697e-01 -4.94239956e-01 8.53034139e-01
4.95385706e-01 -7.50319958e-01 -9.63112354e-01 1.83158553e+00
3.61879259e-01 1.43781574e-02 4.46100920e-01 8.51393402e-01
-2.62380522e-02 5.53272367e-01 4.15833741e-01 2.41540596e-01
8.30339491e-01 -7.21992552e-01 -3.28470588e-01 -5.00186265e-01
5.96744537e-01 -1.00111462e-01 8.91302824e-01 2.88809478e-01
-1.08954275e+00 -5.82123160e-01 -7.48026073e-01 4.89084691e-01
-4.50845808e-01 -4.13113385e-01 3.42021704e-01 4.19028997e-01
-1.09795082e+00 7.81247854e-01 -8.54234576e-01 -5.41925728e-01
4.63725239e-01 1.79703131e-01 -1.29115820e-01 -3.14413197e-02
-1.41246891e+00 1.34421444e+00 2.67978311e-01 -2.89436281e-01
-1.84525442e+00 -2.60742575e-01 -9.25566196e-01 5.72044924e-02
4.78321522e-01 -8.44241738e-01 1.72266996e+00 -9.95212913e-01
-1.90654993e+00 6.25216305e-01 5.79254776e-02 -8.67031217e-01
7.12138772e-01 1.20151028e-01 -1.09699577e-01 -4.62958068e-02
1.89902171e-01 7.70771742e-01 1.03273189e+00 -1.44055605e+00
-9.15419042e-01 -2.78617114e-01 2.04200074e-01 4.36726481e-01
3.49012405e-01 -5.03522456e-01 2.58845657e-01 -1.84771642e-01
-6.63191259e-01 -1.25602281e+00 -7.56288111e-01 -1.41031057e-01
-1.98046286e-02 -3.31121206e-01 5.07840931e-01 -2.36422628e-01
3.48332584e-01 -1.98826623e+00 2.79020160e-01 3.31585348e-01
-2.23731771e-02 1.82857960e-01 -3.27310354e-01 6.37355566e-01
3.34645778e-01 -3.27469110e-01 -2.95570731e-01 -1.22136824e-01
3.87543738e-01 5.82452178e-01 -1.43886814e-02 3.09842736e-01
3.30531329e-01 8.90568793e-01 -1.39773059e+00 -6.94710240e-02
2.39530325e-01 2.07993224e-01 -3.38784724e-01 5.02375662e-01
-5.74521244e-01 7.77647853e-01 -7.08422303e-01 -1.49855940e-02
1.45757779e-01 1.93401933e-01 1.39602600e-02 8.22327912e-01
-7.92565942e-02 -7.96667021e-03 -8.66230905e-01 1.28314471e+00
-7.75287807e-01 5.48559844e-01 6.98448867e-02 -1.08869874e+00
1.07264292e+00 1.57843590e-01 1.24240912e-01 -8.72191072e-01
4.50302660e-01 8.05930644e-02 5.14156342e-01 -7.73207963e-01
2.92107671e-01 -3.62086087e-01 -2.26598457e-01 5.42117178e-01
8.81244391e-02 -7.60574818e-01 2.46586695e-01 -1.87265389e-02
1.43367445e+00 1.49685338e-01 1.45167872e-01 -2.94494689e-01
3.90530676e-01 4.28893089e-01 2.57197201e-01 1.15159881e+00
-5.07140696e-01 2.78337337e-02 6.05436802e-01 -3.62493366e-01
-1.32538164e+00 -1.16575730e+00 5.39117873e-01 1.06648409e+00
2.22995356e-01 3.77052069e-01 -9.64404523e-01 -3.35390896e-01
1.42338276e-01 1.27402878e+00 -7.37748682e-01 -4.84380156e-01
-5.24366796e-01 -2.26196542e-01 1.11063525e-01 3.34984571e-01
2.76927829e-01 -1.74773502e+00 -1.12624252e+00 5.74372947e-01
4.82927233e-01 -9.71361518e-01 -5.10011278e-02 3.29332769e-01
-2.51657307e-01 -1.03303444e+00 -7.48572469e-01 -7.45917201e-01
4.11516368e-01 -4.36770394e-02 9.05919373e-01 -1.89374149e-01
-2.98310280e-01 8.22140634e-01 -3.63285281e-02 -8.65073919e-01
-9.82034564e-01 5.55741265e-02 2.60831028e-01 -6.04219846e-02
4.82453108e-01 -4.75554019e-01 -5.26228070e-01 2.83090740e-01
-7.63964415e-01 -1.13411814e-01 6.66144848e-01 8.80498469e-01
3.89551744e-02 2.82941878e-01 8.35365832e-01 -5.15721858e-01
1.12601900e+00 -7.58521140e-01 -8.03840101e-01 -1.08576007e-01
-4.84005094e-01 4.12762046e-01 9.81950641e-01 -6.41199112e-01
-1.05831516e+00 1.18411563e-01 9.64154024e-04 -1.43157721e-01
-7.86546052e-01 3.15653652e-01 3.81327048e-02 3.01943034e-01
8.56474221e-01 4.64139968e-01 5.34499347e-01 1.43511146e-01
4.30158645e-01 6.44090414e-01 5.28284013e-01 -9.73815858e-01
1.03852093e+00 2.59559512e-01 1.37714401e-01 -9.32093263e-01
-2.24686012e-01 6.85615391e-02 -4.14172828e-01 -4.44145083e-01
8.55213821e-01 -7.00707376e-01 -1.08908534e+00 7.31389821e-01
-9.10795093e-01 -1.32260585e+00 -6.75568044e-01 4.23606485e-01
-1.43807101e+00 -2.86770821e-01 -2.75788039e-01 -9.75049198e-01
2.91860729e-01 -1.23077190e+00 8.09362531e-01 5.97736716e-01
-3.11640110e-02 -1.10449541e+00 2.55173534e-01 -1.81951195e-01
7.39478350e-01 1.85553208e-01 9.97820675e-01 -5.83657205e-01
-4.91128027e-01 6.69331197e-03 2.04314530e-01 2.94387102e-01
-1.08355798e-01 -2.99638242e-01 -6.54241264e-01 -1.95169643e-01
-1.66592941e-01 -6.08634055e-01 2.07345933e-01 2.13684678e-01
6.80942476e-01 -7.84221217e-02 -2.34360680e-01 -1.36592805e-01
1.01833439e+00 5.09431720e-01 5.31542361e-01 4.70520407e-01
1.46058336e-01 1.10649228e+00 6.04572594e-01 4.22368497e-01
7.17562318e-01 4.07966733e-01 5.48768163e-01 6.86796159e-02
2.85266727e-01 -4.88134235e-01 4.41988677e-01 -1.05631806e-01
1.00341924e-01 -2.08665103e-01 -1.13611412e+00 7.62595952e-01
-2.09018517e+00 -1.05788970e+00 1.30052537e-01 2.11853528e+00
5.19896984e-01 3.38518977e-01 4.63919759e-01 -1.64076388e-01
6.34560406e-01 -3.75771821e-01 -9.19343352e-01 -5.58159590e-01
2.34139681e-01 1.51385581e-02 4.82131898e-01 5.96861303e-01
-6.82839036e-01 1.08173287e+00 5.92768717e+00 4.82338279e-01
-8.74793947e-01 -8.53312686e-02 4.45808470e-01 2.29234308e-01
-2.59515122e-02 -1.54199660e-01 -5.09140491e-01 4.19657916e-01
1.38919806e+00 -5.24022758e-01 7.48541474e-01 9.21728253e-01
7.19092607e-01 -1.83595210e-01 -1.10227430e+00 5.77527642e-01
-4.55316097e-01 -8.90038371e-01 -3.10488850e-01 1.97819754e-01
6.81230009e-01 1.10358611e-01 1.12149812e-01 9.89205599e-01
1.26838899e+00 -1.09887040e+00 8.14925075e-01 6.45183086e-01
1.87312961e-01 -8.98801267e-01 7.15634286e-01 8.90481532e-01
-5.45639157e-01 -5.39785564e-01 -1.20403089e-01 -3.83822113e-01
3.79369646e-01 -1.03178740e-01 -1.03288054e+00 5.14485501e-02
2.42532015e-01 2.44570777e-01 -2.01425314e-01 8.82727325e-01
-3.09764624e-01 4.50814813e-01 -2.62930840e-01 -6.61714613e-01
8.23991239e-01 -2.93525875e-01 6.67678237e-01 8.27870309e-01
1.66486621e-01 -5.52299470e-02 3.81070822e-01 9.01359677e-01
4.41573799e-01 -2.49643445e-01 -1.04747021e+00 8.76678601e-02
1.81590870e-01 9.25344288e-01 -4.86789435e-01 -3.04042935e-01
-1.15020670e-01 8.39662790e-01 4.99259293e-01 6.77375197e-01
-7.71184266e-01 -3.56806278e-01 8.51233482e-01 3.25708359e-01
2.93633580e-01 -4.32517380e-01 5.52713610e-02 -4.87347871e-01
-4.89641100e-01 -9.09380436e-01 -3.45156133e-01 -8.51944387e-01
-1.17008996e+00 4.57500309e-01 1.45914644e-01 -8.80325079e-01
-9.24842775e-01 -7.05500603e-01 -7.29408562e-01 9.40888762e-01
-1.34803605e+00 -6.78605914e-01 -3.37512612e-01 6.21578693e-01
5.45013785e-01 -5.94650030e-01 4.48088378e-01 -2.10188374e-01
-2.31119454e-01 1.03355851e-02 1.05474263e-01 -1.17348604e-01
2.65307426e-01 -1.42859268e+00 6.11250818e-01 3.86018306e-01
-4.56007242e-01 2.23043874e-01 1.20873392e+00 -4.93094146e-01
-1.25798059e+00 -1.20074534e+00 2.83440292e-01 -3.52399260e-01
8.59946132e-01 -2.92259306e-01 -8.50659430e-01 4.98869807e-01
2.13212222e-01 -2.23057866e-01 -6.95349351e-02 -3.41435939e-01
2.73136765e-01 1.53870538e-01 -1.24501884e+00 1.12665105e+00
8.02403688e-01 -2.02933356e-01 -5.98349154e-01 2.92846888e-01
4.95505750e-01 -2.13208094e-01 -2.81222165e-01 5.68538494e-02
2.66063124e-01 -7.04126716e-01 4.91648883e-01 -9.73296881e-01
3.56588125e-01 -1.86430335e-01 8.75518247e-02 -1.98118961e+00
-2.64811844e-01 -9.47112978e-01 7.89820030e-02 5.40682614e-01
3.37372690e-01 -7.56933987e-01 7.58441865e-01 5.08338273e-01
-3.49102691e-02 -5.94645262e-01 -8.30228388e-01 -9.48386848e-01
5.72588742e-01 -3.00098956e-01 5.10042965e-01 1.79820746e-01
-7.90187493e-02 3.84734720e-01 -2.60638028e-01 8.45457166e-02
7.60652363e-01 -4.60450113e-01 1.27645326e+00 -1.03631294e+00
-2.70249099e-01 -4.32008147e-01 -4.03216541e-01 -1.07673824e+00
6.49331629e-01 -5.04650474e-01 7.50804722e-01 -1.44808877e+00
-2.47466937e-01 -5.09199917e-01 1.97461277e-01 -3.01228408e-02
-1.18339241e-01 -5.37315071e-01 1.14889279e-01 -2.49513984e-01
-4.47659820e-01 8.16945314e-01 1.33532095e+00 -1.92115959e-02
-3.24390024e-01 3.65261853e-01 -5.08244097e-01 7.19339609e-01
1.05523157e+00 -5.22589207e-01 -6.41692102e-01 -9.51115638e-02
1.18011177e-01 3.07034999e-01 5.53212821e-01 -1.11219037e+00
3.17745447e-01 -6.37116849e-01 -1.43948629e-01 8.03303719e-03
1.21048324e-01 -8.51906717e-01 -1.01680912e-01 8.12453508e-01
-4.17008132e-01 6.47592172e-02 6.39156401e-02 9.57515359e-01
1.24901719e-01 -3.69317472e-01 9.65055764e-01 -4.17738974e-01
-7.84158289e-01 3.52974892e-01 -1.23195505e+00 1.20234966e-01
1.44324267e+00 9.80278328e-02 -1.17403030e-01 -1.02901804e+00
-8.63929927e-01 8.25854123e-01 4.71475333e-01 4.72450107e-01
4.48737979e-01 -7.66500533e-01 -8.70296001e-01 9.19908807e-02
4.76847962e-02 -1.05825923e-01 2.57318854e-01 3.57375443e-01
-5.07149339e-01 1.43722340e-01 -4.47922826e-01 -4.92126405e-01
-5.86292803e-01 7.30587542e-01 6.20602190e-01 -2.19766393e-01
-5.81012368e-01 2.08108917e-01 1.71480834e-01 -7.08269119e-01
1.35034919e-01 4.27183099e-02 -7.55771175e-02 -5.25250375e-01
3.96048307e-01 2.95490980e-01 -3.87772769e-01 -3.79154444e-01
2.26998478e-01 7.62130618e-02 4.20853868e-02 -5.77905357e-01
1.06271899e+00 -2.05189496e-01 7.42339373e-01 5.90495169e-01
6.25488281e-01 -8.44768941e-01 -1.79049599e+00 -1.15681060e-01
8.15962628e-02 -1.54769141e-02 -3.25111926e-01 -6.12983882e-01
-3.08289319e-01 1.04895377e+00 5.81898034e-01 4.26676422e-01
4.55991626e-01 -1.40511259e-01 6.00074172e-01 6.73235118e-01
8.29436362e-01 -1.28798819e+00 3.07723165e-01 8.17759752e-01
8.36287022e-01 -1.38107550e+00 -7.15266943e-01 3.09359223e-01
-1.16003919e+00 7.88818240e-01 7.98849285e-01 -5.31975269e-01
6.02046430e-01 3.15792561e-01 1.60976842e-01 -6.70364723e-02
-1.06859267e+00 -5.87389410e-01 -3.00966114e-01 1.34959292e+00
-3.32944244e-01 1.05682194e-01 8.83657560e-02 1.46171832e-02
-3.31952542e-01 2.62703151e-01 9.29584444e-01 1.03662562e+00
-7.41944253e-01 -7.95154154e-01 -1.71570435e-01 2.68179774e-01
4.16594855e-02 4.02521223e-01 4.91548777e-02 8.86877418e-01
-1.08582489e-01 9.13688958e-01 1.55351281e-01 3.22890580e-02
4.52935606e-01 1.00981325e-01 3.89331967e-01 -7.13030934e-01
-3.42589676e-01 -5.74725807e-01 -5.11253700e-02 -4.32552397e-01
-5.74216619e-03 -7.08215237e-01 -1.36499476e+00 -2.98510283e-01
5.18044233e-01 1.37160018e-01 8.40239167e-01 1.05923772e+00
2.62170374e-01 6.81979239e-01 6.85503185e-01 -1.11360705e+00
-9.60551202e-01 -8.27688515e-01 -6.25394702e-01 5.77361763e-01
4.84884620e-01 -8.67369294e-01 -4.94602561e-01 1.07557727e-02] | [4.403140068054199, 1.538069486618042] |
6d5fe98c-e354-4037-9573-5ab688b4872f | reference-based-variational-autoencoders | null | null | https://openreview.net/forum?id=H1e1XeXlP4 | https://openreview.net/pdf?id=H1e1XeXlP4 | Reference-based Variational Autoencoders | Learning disentangled representations from visual data, where different high-level generative factors are independently encoded, is of importance for many computer vision tasks. Solving this problem, however, typically requires to explicitly label all the factors of interest in training images. To alleviate the annotation cost, we introduce a learning setting which we refer to as \textit{reference-based disentangling}. Given a pool of unlabelled images, the goal is to learn a representation where a set of target factors are disentangled from others. The only supervision comes from an auxiliary \textit{reference set} containing images where the factors of interest are constant. To address this problem, we propose reference-based variational autoencoders, a novel deep generative model designed to exploit the weak-supervision provided by the reference set. By addressing tasks such as feature learning, conditional image generation or attribute transfer, we validate the ability of the proposed model to learn disentangled representations from this minimal form of supervision.
| ['Jakob Verbeek', 'Xavier Binefa', 'Oriol Martinez', 'Adrià Ruiz'] | 2019-03-08 | null | null | null | iclr-workshop-lld-2019 | ['conditional-image-generation'] | ['computer-vision'] | [ 3.90864104e-01 5.00483632e-01 -1.81475252e-01 -4.08645481e-01
-7.73231328e-01 -5.76246679e-01 9.58690107e-01 -4.03184533e-01
-3.06853801e-01 8.39604497e-01 1.67546615e-01 1.72630876e-01
-6.34700246e-03 -5.99676609e-01 -9.49439645e-01 -1.18536425e+00
2.80981779e-01 4.83056217e-01 -3.60112667e-01 2.07300168e-02
-2.36214086e-01 2.14694977e-01 -1.50184143e+00 1.56799763e-01
7.05711424e-01 7.59308040e-01 2.81760514e-01 5.08117437e-01
3.73820752e-01 7.19661176e-01 -5.26794910e-01 -4.43111092e-01
3.42634618e-01 -6.73268497e-01 -4.59551841e-01 5.14293194e-01
4.07150656e-01 -3.36375833e-01 -4.70364779e-01 9.56462204e-01
2.55048096e-01 1.60000890e-01 1.03294992e+00 -1.51420176e+00
-9.30949628e-01 5.09418249e-01 -5.51332474e-01 3.41952629e-02
-1.76855743e-01 4.64132503e-02 1.45100605e+00 -9.42494929e-01
6.91039562e-01 1.08100128e+00 -1.20637432e-01 9.11466062e-01
-1.95978487e+00 -7.02684402e-01 3.04798990e-01 -1.05030350e-01
-1.19061124e+00 -6.40469909e-01 9.27201629e-01 -7.40315437e-01
3.65512788e-01 7.50213787e-02 3.35738540e-01 1.55099559e+00
-2.09558174e-01 9.09058392e-01 1.17410707e+00 -3.72213870e-01
2.88174123e-01 3.63575190e-01 1.27070412e-01 7.70486832e-01
3.78149837e-01 1.63893253e-01 -7.49025762e-01 -1.22489855e-01
7.98913896e-01 1.19286165e-01 -6.55673504e-01 -9.89355266e-01
-1.27483642e+00 1.25126696e+00 5.14479816e-01 -4.87071760e-02
-3.08787555e-01 1.05069354e-01 -1.06767491e-02 6.23680688e-02
4.96869177e-01 3.35669965e-01 -4.14209247e-01 4.35872048e-01
-7.68109739e-01 1.10762224e-01 5.44151723e-01 9.40561473e-01
1.10649741e+00 1.96351230e-01 -3.38875532e-01 6.36450469e-01
5.07629931e-01 4.71498638e-01 2.29177177e-01 -8.52038085e-01
5.36439955e-01 5.03408253e-01 1.40795365e-01 -6.69142187e-01
1.89156771e-01 -5.01140296e-01 -9.64703441e-01 4.84309286e-01
3.84413332e-01 -8.77560452e-02 -1.20554554e+00 2.37276125e+00
1.23570390e-01 1.30981982e-01 2.27191523e-01 9.67598200e-01
6.36400044e-01 5.54647505e-01 -5.55685498e-02 -1.90861836e-01
1.30133903e+00 -7.86702335e-01 -5.68826020e-01 -2.03944832e-01
8.64164904e-02 -3.11262637e-01 8.91384006e-01 2.81976640e-01
-8.29967618e-01 -3.82551461e-01 -1.23600066e+00 -3.39477986e-01
-5.30728400e-02 5.05936265e-01 5.23929358e-01 4.17862177e-01
-7.48201370e-01 3.10063422e-01 -1.01138437e+00 1.03217006e-01
6.95470035e-01 3.84427696e-01 -7.17786729e-01 -1.39372960e-01
-9.14570510e-01 5.57739019e-01 1.87696919e-01 1.04463823e-01
-1.44138646e+00 -5.54698110e-01 -1.12041390e+00 2.56490707e-01
4.97690827e-01 -9.63509321e-01 7.98158646e-01 -9.83978808e-01
-1.20766711e+00 1.04045737e+00 -2.35596895e-01 -7.55731389e-02
4.46813911e-01 -2.56274521e-01 1.22189686e-01 1.76907599e-01
2.02414885e-01 6.45806670e-01 1.54095554e+00 -1.61498249e+00
3.97337507e-03 -6.76072538e-01 1.84582010e-01 7.46692419e-02
-2.81452298e-01 -3.67300779e-01 -2.20119387e-01 -5.71542680e-01
-1.12571575e-01 -1.07490778e+00 -1.05385311e-01 1.34509400e-01
-6.38449013e-01 -4.98558432e-02 7.84227371e-01 -5.27368963e-01
4.55050647e-01 -2.14561129e+00 8.85011137e-01 -2.99125854e-02
6.89654052e-01 -5.18129356e-02 -1.84733361e-01 1.66078895e-01
-1.54046580e-01 -1.57840010e-02 -3.38099778e-01 -7.71781802e-01
4.68136035e-02 5.93560100e-01 -3.93280149e-01 5.43999493e-01
7.11523175e-01 7.86893427e-01 -8.00353169e-01 -3.13603669e-01
7.54300952e-02 8.64449620e-01 -7.02594519e-01 5.73936880e-01
-5.24878383e-01 9.90523160e-01 -5.33829749e-01 1.18731573e-01
4.51826990e-01 -6.12490475e-01 1.45284921e-01 -3.09830099e-01
3.17032427e-01 2.51685567e-02 -9.83462512e-01 1.96321058e+00
-4.79978710e-01 5.67980766e-01 1.52078539e-01 -1.00822484e+00
7.66728580e-01 4.89553958e-01 2.78634071e-01 -1.47667885e-01
1.57595843e-01 -1.16793714e-01 2.00183578e-02 -4.59063053e-01
1.99050635e-01 -3.88789922e-01 -7.97245502e-02 6.47343099e-01
7.24062026e-01 5.02048656e-02 -5.21195643e-02 5.95702052e-01
8.65273833e-01 3.76219183e-01 4.17727828e-01 -1.06153794e-01
3.10913652e-01 -4.28664088e-01 8.13457608e-01 5.57052612e-01
-1.79383978e-02 7.12861598e-01 7.84417272e-01 -1.26305953e-01
-1.13269246e+00 -1.28690779e+00 4.04323600e-02 1.04732370e+00
-1.77051678e-01 -2.34566227e-01 -5.15344501e-01 -7.73662329e-01
-8.54357928e-02 5.77120006e-01 -1.07700717e+00 -1.96817279e-01
-2.25445271e-01 -6.84250414e-01 2.75091320e-01 4.21059251e-01
1.60717264e-01 -7.52953231e-01 -6.10227287e-01 -2.43801326e-01
-1.67420745e-01 -1.09584081e+00 -4.32871729e-01 3.87705743e-01
-6.37211263e-01 -1.07621002e+00 -8.49904537e-01 -3.54337513e-01
1.00165641e+00 1.98613107e-01 1.26955986e+00 -2.62852550e-01
-2.36358911e-01 4.05766189e-01 -9.65990648e-02 -1.56568617e-01
-2.39551634e-01 6.73955819e-03 -1.24642029e-01 5.73332250e-01
1.64619580e-01 -8.57106805e-01 -5.16929746e-01 9.58653912e-02
-1.06507969e+00 2.56176531e-01 7.29585111e-01 1.28256524e+00
5.66720068e-01 -5.16174138e-01 4.15866524e-01 -1.08059621e+00
2.67949641e-01 -5.04885972e-01 -5.53977489e-01 2.60840088e-01
-3.03445190e-01 7.29372382e-01 4.76688504e-01 -6.28428578e-01
-1.15530729e+00 2.72386134e-01 4.23205376e-01 -9.21822667e-01
-2.40276262e-01 3.14518690e-01 -7.94086576e-01 3.96180719e-01
6.31803989e-01 2.42570415e-01 1.91407487e-01 -4.37094688e-01
7.30670214e-01 2.89420485e-01 3.74234319e-01 -6.90914154e-01
8.12916756e-01 5.53844154e-01 7.62675330e-02 -6.31087601e-01
-1.24811602e+00 -1.39138773e-01 -8.42510462e-01 1.69242010e-01
1.08094430e+00 -1.16589856e+00 -6.12758577e-01 8.19831938e-02
-1.23964036e+00 -2.28218846e-02 -4.68456388e-01 4.75463629e-01
-7.43663549e-01 -5.26228100e-02 -2.59140313e-01 -7.67489374e-01
7.26097673e-02 -1.34586668e+00 1.28994918e+00 1.45249426e-01
-2.44828593e-02 -9.60677743e-01 2.43839890e-01 5.27657270e-01
-9.97252017e-02 5.99692583e-01 1.00702393e+00 -8.28074872e-01
-8.90871286e-01 6.63827732e-02 -1.12262689e-01 5.96895993e-01
2.78352052e-01 -3.71750027e-01 -1.29015875e+00 -3.34922463e-01
1.18628241e-01 -7.78604984e-01 1.18271768e+00 7.50514716e-02
8.93229067e-01 -4.33657467e-01 -2.32621983e-01 6.18497670e-01
1.14799809e+00 -2.42408484e-01 2.52631813e-01 -3.17368120e-01
9.70841289e-01 6.94958210e-01 8.47997959e-04 3.35048378e-01
3.31753999e-01 5.95162630e-01 4.67754066e-01 -4.97179441e-02
2.12860126e-02 -3.89900327e-01 2.73509175e-01 5.78285873e-01
-1.53719142e-01 -3.01948100e-01 -4.34746355e-01 4.73199725e-01
-1.84019780e+00 -9.61399198e-01 2.08501786e-01 2.14598680e+00
7.90671647e-01 -2.29537323e-01 -1.82103589e-01 5.04738837e-02
5.60480356e-01 4.56054181e-01 -9.16394591e-01 1.59113184e-01
-7.67839476e-02 1.84658721e-01 -5.71914874e-02 5.10389149e-01
-1.01569331e+00 5.88577390e-01 5.10433626e+00 3.79395217e-01
-1.01216090e+00 2.36176521e-01 5.94259739e-01 -1.51681364e-01
-6.99965417e-01 1.59513518e-01 -6.47970259e-01 2.51388729e-01
5.59226871e-01 7.31566362e-03 3.94538164e-01 6.65311277e-01
-1.22490190e-02 1.24319032e-01 -1.70094669e+00 8.56008947e-01
4.37124878e-01 -1.01358294e+00 2.56904572e-01 3.64276499e-01
8.04086864e-01 -2.67873466e-01 4.45619106e-01 2.41098344e-01
5.97460747e-01 -1.23488569e+00 6.05005503e-01 5.26647031e-01
8.42667282e-01 -5.72764754e-01 2.68229336e-01 4.59495276e-01
-7.13386655e-01 2.64198124e-01 -1.67462215e-01 7.80231580e-02
4.89953235e-02 6.30114675e-01 -5.70570529e-01 6.21813893e-01
1.88124448e-01 8.56594622e-01 -3.65629822e-01 3.07319015e-01
-6.69177890e-01 5.74300766e-01 2.94935033e-02 4.16949362e-01
-4.44040634e-02 -3.63987863e-01 5.09218037e-01 6.02134109e-01
5.82708642e-02 1.71422854e-01 5.03441170e-02 1.49378991e+00
-4.26962316e-01 -3.15786481e-01 -7.97266543e-01 -1.74106196e-01
-2.68844590e-02 1.33419299e+00 -3.89227033e-01 -1.75941616e-01
-4.88525689e-01 1.28617406e+00 6.14146590e-01 7.36511827e-01
-5.86870432e-01 8.59029815e-02 8.86758089e-01 -3.00270945e-01
5.14682651e-01 -3.14414948e-01 -2.80216597e-02 -1.69607413e+00
-1.51644116e-02 -5.69534957e-01 2.41319403e-01 -7.96454906e-01
-1.32354438e+00 5.73910773e-01 1.37665525e-01 -1.12116027e+00
-4.55914199e-01 -5.56545794e-01 -4.55427796e-01 1.18237340e+00
-1.33671379e+00 -1.48611248e+00 -2.93685317e-01 8.53534400e-01
4.02919590e-01 -1.81924582e-01 8.69629085e-01 -2.46863943e-02
-6.02466345e-01 4.28451926e-01 -5.82507513e-02 1.56609431e-01
5.02580106e-01 -1.38256073e+00 -1.96308926e-01 8.95479500e-01
7.18487620e-01 8.01356673e-01 7.86120355e-01 -2.78545499e-01
-1.23847616e+00 -1.05720878e+00 6.69780612e-01 -6.80350780e-01
5.23676455e-01 -9.34223533e-01 -7.59444952e-01 1.00705051e+00
2.89186537e-01 5.59126377e-01 1.00069118e+00 8.53056088e-02
-7.75810719e-01 6.34326860e-02 -9.31450248e-01 3.41453433e-01
9.09335792e-01 -7.60524631e-01 -5.91780722e-01 9.69537869e-02
9.34360623e-01 -1.83736429e-01 -5.80202043e-01 -5.34022227e-02
4.71526563e-01 -8.69995892e-01 8.82713079e-01 -8.82409632e-01
6.96191907e-01 -1.09017469e-01 -2.40856543e-01 -1.63090098e+00
-1.57351255e-01 -5.85067928e-01 -3.25762630e-01 1.20815921e+00
3.96162570e-01 -4.77701753e-01 8.47271204e-01 6.74132526e-01
1.61768556e-01 -5.29495001e-01 -8.25426936e-01 -4.40319300e-01
4.70630452e-02 -8.20072219e-02 2.89176106e-01 9.19399917e-01
-4.79398400e-01 9.54958200e-01 -8.07036161e-01 3.55376214e-01
9.95502293e-01 3.04453105e-01 6.82834327e-01 -1.34584641e+00
-7.69336939e-01 1.70619264e-01 -4.21644211e-01 -9.99450207e-01
5.41666746e-01 -1.05790114e+00 -1.01557898e-03 -1.18054628e+00
6.26514018e-01 -1.68909550e-01 -2.60004997e-01 6.06323183e-01
-3.51415694e-01 2.20653400e-01 2.95046926e-01 3.19392025e-01
-3.37468922e-01 1.09949410e+00 1.36176097e+00 -2.57281154e-01
1.57976493e-01 2.90878788e-02 -9.59925234e-01 4.80275571e-01
3.27778190e-01 -6.17659807e-01 -7.66378522e-01 -5.33335209e-01
2.40667999e-01 2.39350110e-01 7.22824693e-01 -3.22089493e-01
-1.90828964e-02 -2.17315972e-01 4.44366783e-01 -1.27072915e-01
6.53128564e-01 -8.68141115e-01 9.00488496e-02 -6.66731820e-02
-6.13000453e-01 -3.11157405e-01 -2.58242249e-01 8.20429265e-01
-2.44602636e-01 -2.04428717e-01 6.36266708e-01 -1.01659194e-01
-1.19041443e-01 5.07275164e-01 -8.21948871e-02 1.72765911e-01
1.06711221e+00 1.85269073e-01 -2.24477664e-01 -3.76398206e-01
-1.15451598e+00 -8.64040805e-04 3.89497250e-01 3.02534074e-01
6.75707281e-01 -1.39233637e+00 -6.78965390e-01 4.74257439e-01
4.76931989e-01 3.32037926e-01 6.78720698e-02 5.62947571e-01
2.56546557e-01 2.96583384e-01 -2.59084553e-01 -6.64936304e-01
-1.02249599e+00 7.25293994e-01 1.26052961e-01 -2.86416918e-01
-4.81014490e-01 9.32477832e-01 8.60450089e-01 -3.05211425e-01
-5.91008738e-02 -1.47928447e-01 -1.25223339e-01 1.31285265e-01
2.63148248e-01 -2.07919977e-04 -1.50940925e-01 -8.83211374e-01
-6.37734681e-02 1.97607696e-01 -2.13046208e-01 -4.78188276e-01
1.48880827e+00 -6.36385158e-02 4.93499159e-04 7.35685647e-01
1.38246477e+00 -6.23400956e-02 -1.86715913e+00 -3.37921143e-01
-3.07137638e-01 -4.75749463e-01 1.22844139e-02 -5.89432538e-01
-1.34195936e+00 1.22758853e+00 4.11229759e-01 -1.68892175e-01
9.65547144e-01 3.40928197e-01 1.71726182e-01 1.90127283e-01
2.41251856e-01 -3.52645189e-01 4.03467506e-01 7.69196153e-02
1.00344110e+00 -1.45590842e+00 -2.63154328e-01 -2.19499141e-01
-7.59648442e-01 7.42282212e-01 5.41514158e-01 -3.21303636e-01
6.14159107e-01 8.41788575e-02 -1.33480415e-01 -3.18765819e-01
-9.74133074e-01 -2.31329471e-01 5.19422770e-01 6.78854465e-01
2.68395215e-01 7.75502920e-02 1.57451287e-01 6.48589909e-01
-2.28903890e-02 -2.45795593e-01 4.44581747e-01 5.90714872e-01
8.40410814e-02 -1.15134156e+00 -1.84429333e-01 3.68896872e-01
-3.07484269e-01 -4.20205183e-02 -3.90653014e-01 6.62015915e-01
1.88890293e-01 6.96227491e-01 4.67817374e-02 -4.13658544e-02
4.48336871e-03 1.53653204e-01 7.17953801e-01 -9.14112747e-01
-7.89503679e-02 1.57153174e-01 -3.52811694e-01 -3.88575852e-01
-7.60349095e-01 -6.53417647e-01 -8.14660668e-01 3.48327041e-01
-2.44725168e-01 1.84766859e-01 4.26164776e-01 9.79965448e-01
2.77724385e-01 5.67524076e-01 6.19702458e-01 -9.01006341e-01
-7.20700681e-01 -9.39723730e-01 -7.12540150e-01 6.83393538e-01
8.53968203e-01 -9.70942080e-01 -5.08180141e-01 5.52193642e-01] | [10.960392951965332, 0.5820977091789246] |
8eda43de-30bf-422c-8fad-7a2e410f96a0 | differences-between-human-and-machine | 2011.14036 | null | https://arxiv.org/abs/2011.14036v1 | https://arxiv.org/pdf/2011.14036v1.pdf | Differences between human and machine perception in medical diagnosis | Deep neural networks (DNNs) show promise in image-based medical diagnosis, but cannot be fully trusted since their performance can be severely degraded by dataset shifts to which human perception remains invariant. If we can better understand the differences between human and machine perception, we can potentially characterize and mitigate this effect. We therefore propose a framework for comparing human and machine perception in medical diagnosis. The two are compared with respect to their sensitivity to the removal of clinically meaningful information, and to the regions of an image deemed most suspicious. Drawing inspiration from the natural image domain, we frame both comparisons in terms of perturbation robustness. The novelty of our framework is that separate analyses are performed for subgroups with clinically meaningful differences. We argue that this is necessary in order to avert Simpson's paradox and draw correct conclusions. We demonstrate our framework with a case study in breast cancer screening, and reveal significant differences between radiologists and DNNs. We compare the two with respect to their robustness to Gaussian low-pass filtering, performing a subgroup analysis on microcalcifications and soft tissue lesions. For microcalcifications, DNNs use a separate set of high frequency components than radiologists, some of which lie outside the image regions considered most suspicious by radiologists. These features run the risk of being spurious, but if not, could represent potential new biomarkers. For soft tissue lesions, the divergence between radiologists and DNNs is even starker, with DNNs relying heavily on spurious high frequency components ignored by radiologists. Importantly, this deviation in soft tissue lesions was only observable through subgroup analysis, which highlights the importance of incorporating medical domain knowledge into our comparison framework. | ['Krzysztof J. Geras', 'Kyunghyun Cho', 'Linda Moy', 'Laura Heacock', 'Daniel K. Sodickson', 'James Park', 'Alice Kim', 'Linda Du', 'Divya Awal', 'Hildegard Toth', 'Beatriu Reig', 'Kristine Pysarenko', 'Jiyon Lee', 'Eric Kim', 'Chloe Chhor', 'Naziya Samreen', 'Robin Ehrenpreis', 'Celin Chacko', 'Witold Oleszkiewicz', 'Stanislaw Jastrzebski', 'Taro Makino'] | 2020-11-28 | null | null | null | null | ['breast-cancer-detection', 'breast-cancer-detection'] | ['knowledge-base', 'medical'] | [ 8.42648327e-01 6.79438293e-01 -1.33971483e-01 -1.91044107e-01
-8.59211564e-01 -7.71788239e-01 5.76786697e-01 5.23218870e-01
-5.49173772e-01 2.83229649e-01 5.80420494e-01 -5.73230207e-01
-4.75742757e-01 -7.13484406e-01 -6.56936467e-01 -8.68944407e-01
-4.99003194e-02 1.38212457e-01 1.49394006e-01 1.66800637e-02
-1.23814017e-01 4.63986963e-01 -1.27873290e+00 5.94209194e-01
6.40829980e-01 7.18592525e-01 -5.99560253e-02 7.15809405e-01
4.32591289e-01 7.41481423e-01 -3.79344195e-01 -3.33448917e-01
3.42668205e-01 -3.52730066e-01 -7.53502429e-01 2.71445066e-01
4.88543481e-01 -3.16833586e-01 -3.16552550e-01 1.18505216e+00
8.27636838e-01 -1.23178303e-01 7.89937437e-01 -6.58248127e-01
-5.91664493e-01 5.88546097e-01 -3.82373899e-01 2.49809578e-01
1.60032317e-01 5.75247705e-01 7.73792922e-01 -3.89047205e-01
6.72495425e-01 1.33731329e+00 1.09218621e+00 3.83262247e-01
-1.38947475e+00 -1.30991191e-01 3.70419696e-02 -7.85969198e-02
-1.08664644e+00 -5.62344372e-01 2.20155746e-01 -7.03588724e-01
4.59675342e-01 5.70747018e-01 5.76061010e-01 1.18212306e+00
3.72525275e-01 4.34622765e-01 1.16917396e+00 -3.62970054e-01
1.85183778e-01 2.03921899e-01 1.54136896e-01 4.88805711e-01
6.20516360e-01 2.24790663e-01 -3.52539755e-02 -4.00858939e-01
6.94734812e-01 -1.55076876e-01 -6.15805566e-01 -2.96059400e-01
-1.36355329e+00 7.30367661e-01 3.64478946e-01 5.05703509e-01
-4.26429689e-01 -9.54346508e-02 4.95914936e-01 6.03138320e-02
1.35355636e-01 6.10908389e-01 -1.32709593e-01 1.36690885e-01
-7.91703105e-01 2.76961434e-03 6.27276361e-01 8.27731639e-02
1.59531221e-01 -3.37879330e-01 -4.02940065e-01 5.91917932e-01
-3.77803855e-02 3.62773150e-01 7.05427408e-01 -1.03283560e+00
-2.17342332e-01 4.01422203e-01 1.47191793e-01 -1.50813174e+00
-7.23788023e-01 -7.87322879e-01 -9.53240693e-01 1.95190534e-01
6.93069577e-01 2.48473436e-02 -9.50395107e-01 1.70316470e+00
1.41900450e-01 -1.33373961e-01 9.34666172e-02 1.04642153e+00
7.80858755e-01 -6.73635155e-02 1.12617448e-01 -1.79082323e-02
1.74536324e+00 -1.66017354e-01 -3.20560396e-01 -3.29595834e-01
7.53242671e-01 -5.11577487e-01 1.14571095e+00 3.76068383e-01
-1.20579362e+00 -4.74509358e-01 -1.06686854e+00 -2.35773604e-02
-1.31812602e-01 -9.55267400e-02 5.42751431e-01 7.66734838e-01
-1.12850928e+00 8.06883454e-01 -1.00161946e+00 -6.95336640e-01
5.13813436e-01 3.72339964e-01 -2.58006781e-01 6.12316430e-02
-1.13340378e+00 1.02498174e+00 1.12931408e-01 2.00528502e-01
-5.42057931e-01 -8.50288570e-01 -7.08327770e-01 2.15243146e-01
3.39618921e-01 -7.65420675e-01 1.40086854e+00 -1.24255443e+00
-7.76196063e-01 1.26182461e+00 -1.87358744e-02 -7.11084366e-01
9.61453676e-01 2.91392058e-01 -4.13980514e-01 3.95846605e-01
2.11439773e-01 5.13853192e-01 6.76113725e-01 -1.35402226e+00
-5.95070362e-01 -4.10904199e-01 4.30812649e-02 -4.18358669e-02
-4.92378287e-02 -1.82005137e-01 -1.68937609e-01 -5.64844370e-01
2.78652608e-01 -1.09636617e+00 -4.69998538e-01 1.39216736e-01
-3.88721257e-01 2.57951111e-01 2.42301807e-01 -6.46184742e-01
8.73835504e-01 -2.22823262e+00 -4.20738488e-01 4.49494153e-01
6.32667720e-01 9.92168635e-02 -1.33249894e-01 -9.19092521e-02
-2.74793565e-01 3.30357790e-01 -2.53382832e-01 3.33761901e-01
7.82670379e-02 9.14170668e-02 -8.90611857e-02 8.15549314e-01
4.79002863e-01 9.96730745e-01 -1.00166702e+00 -3.38122487e-01
1.28696188e-01 4.70059425e-01 -4.02061075e-01 -5.02115130e-01
1.94058314e-01 4.30046231e-01 -3.07967454e-01 6.21494353e-01
6.31378710e-01 -4.99466419e-01 3.97032410e-01 -4.64605719e-01
2.23133132e-01 1.98662415e-01 -9.42777872e-01 9.92290735e-01
-1.40586749e-01 7.60653675e-01 2.88134307e-01 -1.23375154e+00
4.48886395e-01 3.75296324e-01 5.44579744e-01 -6.05785847e-01
2.04443946e-01 2.68830746e-01 8.04892063e-01 -7.75843024e-01
1.60149738e-01 -6.11696899e-01 2.52201796e-01 4.44073796e-01
-3.58747691e-01 -7.73719102e-02 -8.89783055e-02 -5.54613173e-02
1.40807498e+00 -5.22387087e-01 3.63310814e-01 -4.23847049e-01
1.84202418e-01 7.79586583e-02 4.78131801e-01 1.23360789e+00
-5.68681240e-01 8.74693215e-01 8.42314363e-01 -3.13057005e-01
-8.54119658e-01 -1.37791741e+00 -5.24249256e-01 7.31875598e-01
4.64649610e-02 2.38806978e-01 -6.37548864e-01 -6.83351219e-01
2.89794534e-01 5.15665472e-01 -1.01787639e+00 -5.17792523e-01
-4.67069507e-01 -1.26847327e+00 7.35333443e-01 4.23819095e-01
2.51114964e-02 -7.84807801e-01 -1.05629444e+00 8.34321901e-02
-3.55846323e-02 -1.04344201e+00 -3.75932343e-02 1.41063407e-01
-7.08594561e-01 -1.29774904e+00 -8.95469546e-01 -4.37812716e-01
7.55063891e-01 1.72846466e-01 1.20612347e+00 7.48315826e-02
-4.57224816e-01 8.67647052e-01 -1.46281332e-01 -6.45202160e-01
-1.07507253e+00 -2.14408755e-01 -1.04670443e-01 -1.53709799e-01
4.26079631e-01 -4.40972596e-01 -9.26932991e-01 2.06935272e-01
-1.16650403e+00 -2.05762774e-01 9.34395134e-01 8.37881029e-01
3.32513124e-01 -5.00741005e-02 4.72212017e-01 -9.42554414e-01
7.08954692e-01 -5.40418744e-01 -7.95508400e-02 1.52385578e-01
-4.37937468e-01 -7.87399635e-02 2.31076911e-01 -5.48789263e-01
-8.85769963e-01 -2.47107327e-01 2.15127349e-01 -2.23145932e-02
-3.75456035e-01 5.93394876e-01 2.65774965e-01 -1.00093335e-01
1.20728576e+00 -1.22312240e-01 3.21574330e-01 -2.55189449e-01
2.62263212e-02 6.01913989e-01 7.10587084e-01 -3.64274919e-01
4.54376310e-01 9.69213009e-01 1.36852413e-01 -7.62929380e-01
-7.99321115e-01 -2.47807562e-01 -3.39263022e-01 -3.00521962e-02
8.10624480e-01 -7.25248158e-01 -7.38644481e-01 6.61224127e-03
-8.19138110e-01 -1.64346740e-01 -6.16997421e-01 6.71253979e-01
-3.31492752e-01 5.45075595e-01 -6.67072177e-01 -5.80090582e-01
-4.87259449e-03 -1.48334789e+00 8.62449348e-01 -3.97838913e-02
-6.21835947e-01 -1.04953229e+00 -3.26316506e-01 2.27619365e-01
3.31053138e-01 6.26574695e-01 1.18324411e+00 -9.89928722e-01
-1.37498319e-01 -4.11147416e-01 -4.58025157e-01 4.05952871e-01
3.77623707e-01 -6.93644062e-02 -1.22257483e+00 -3.17188680e-01
3.34328532e-01 -1.31691545e-02 1.07538152e+00 9.06519592e-01
1.11991775e+00 -1.33836657e-01 -4.59060580e-01 3.70232582e-01
1.07702160e+00 1.09802999e-01 4.65842813e-01 3.65779370e-01
4.68971729e-01 9.25175905e-01 1.38680428e-01 9.87587497e-02
8.74882564e-02 2.77833074e-01 3.84453952e-01 -5.56868970e-01
-2.35658571e-01 2.16995806e-01 2.82486640e-02 2.75904328e-01
-4.50766869e-02 -1.16693377e-01 -1.18638039e+00 7.76273370e-01
-1.58752823e+00 -8.30594480e-01 -1.17139392e-01 2.32873988e+00
7.28052855e-01 4.09746289e-01 1.20856605e-01 1.73823044e-01
6.85691297e-01 -6.72888905e-02 -5.83242178e-01 -2.45675072e-01
-3.04857790e-01 1.15235060e-01 7.36731112e-01 3.93364757e-01
-1.04430282e+00 3.57656665e-02 6.75259209e+00 5.78416288e-01
-1.25622940e+00 -7.26044476e-02 1.12455416e+00 -1.04296200e-01
-2.37102136e-01 -2.70101756e-01 -6.87444061e-02 3.02225292e-01
9.15417731e-01 -3.17456983e-02 -5.72633147e-02 5.48263431e-01
5.24395585e-01 -3.15307200e-01 -1.42915261e+00 5.52921355e-01
-1.15872122e-01 -1.23437333e+00 -1.54941574e-01 7.48858601e-02
5.34160078e-01 -1.93791336e-03 4.14686888e-01 -3.81567255e-02
2.45605588e-01 -1.51591074e+00 4.74503964e-01 5.12138486e-01
5.97547948e-01 -3.64169747e-01 9.17763591e-01 8.10379311e-02
-3.57167244e-01 -1.00808248e-01 -2.13530764e-01 5.70893511e-02
-3.86586666e-01 9.02713537e-01 -1.18080425e+00 2.36299276e-01
5.83651960e-01 3.41044426e-01 -5.68441153e-01 8.70896697e-01
2.61150688e-01 6.74883246e-01 -1.99928984e-01 3.40348125e-01
2.10623831e-01 1.99040323e-01 5.45658827e-01 1.51606059e+00
2.96996206e-01 2.66990904e-02 -7.07061291e-02 9.47302461e-01
2.83966184e-01 -8.68554041e-02 -4.50339705e-01 -1.20745623e-03
2.51850963e-01 1.14561498e+00 -1.04148269e+00 -4.55746502e-01
-5.37911296e-01 5.77705145e-01 -2.90450126e-01 3.63655835e-01
-5.69827020e-01 -6.38454333e-02 4.26232785e-01 3.53553116e-01
-1.73437931e-02 4.27491277e-01 -6.68088913e-01 -9.55765128e-01
6.13992140e-02 -1.28538775e+00 2.93163478e-01 -5.70160389e-01
-1.31871259e+00 1.52557790e-01 -4.05034125e-02 -1.09111714e+00
-1.35590896e-01 -8.59068036e-01 -3.43186170e-01 7.93373406e-01
-1.28102851e+00 -8.60140383e-01 -2.19597384e-01 8.17881599e-02
-5.97904585e-02 3.33329201e-01 5.99751532e-01 4.95445095e-02
-2.12754846e-01 6.90777421e-01 1.50799423e-01 1.33885279e-01
8.11603129e-01 -1.32338607e+00 2.40473390e-01 6.28951073e-01
-2.34887823e-01 6.63905740e-01 9.48388398e-01 -5.01797557e-01
-1.02591264e+00 -9.23782706e-01 4.76709634e-01 -6.88879013e-01
5.41991413e-01 1.07351445e-01 -1.02689731e+00 4.81394649e-01
-2.20765784e-01 5.86388744e-02 6.51877820e-01 -2.27959678e-02
-2.23547548e-01 1.95684299e-01 -1.42814982e+00 8.21632206e-01
8.30021024e-01 -3.89643222e-01 -5.59283435e-01 4.13537472e-01
4.47381854e-01 -2.36895904e-01 -1.01081121e+00 7.64845312e-01
7.92471588e-01 -1.36324096e+00 1.21976960e+00 -4.71020728e-01
5.71166456e-01 9.87213477e-03 -9.31471288e-02 -1.11147237e+00
-4.22220290e-01 -2.31223017e-01 4.16081458e-01 7.16403842e-01
3.13502520e-01 -7.05433130e-01 7.94912279e-01 6.17840409e-01
-6.22827373e-02 -8.26259494e-01 -7.52673507e-01 -5.89838862e-01
3.57520223e-01 -5.99281967e-01 3.53017658e-01 1.12278402e+00
-2.46369448e-02 -2.35507935e-01 2.09796980e-01 3.72298628e-01
2.97166228e-01 -2.06139982e-01 3.80704820e-01 -1.31531858e+00
-6.76346004e-01 -7.20953643e-01 -5.98707974e-01 -4.03672218e-01
-2.81535119e-01 -8.47644389e-01 8.57140645e-02 -1.30301714e+00
4.05245990e-01 -1.62106439e-01 -4.05715138e-01 4.00563926e-01
-2.65738308e-01 2.92364180e-01 1.37574375e-01 1.61733001e-01
5.63730709e-02 -2.93926179e-01 1.32499981e+00 -2.22078726e-01
-9.03915316e-02 3.09372619e-02 -1.10615933e+00 1.22149718e+00
6.52137399e-01 -2.62017459e-01 -2.35614002e-01 -1.46596432e-01
1.99647784e-01 -7.06621706e-02 9.61056948e-01 -8.95669818e-01
-1.65375039e-01 -1.06803373e-01 6.80453837e-01 -6.76044151e-02
-2.71962769e-02 -6.12870753e-01 2.67693195e-02 1.05120361e+00
-5.85928202e-01 -2.98023492e-01 3.02889407e-01 4.12811458e-01
4.82541835e-03 -1.75866991e-01 9.76746380e-01 -5.60254812e-01
-3.52544248e-01 -2.70243078e-01 -6.51885152e-01 7.43868575e-02
7.05657244e-01 -4.80972350e-01 -3.39913756e-01 -4.93063658e-01
-1.06090915e+00 -1.16516799e-01 5.44660985e-01 1.43635437e-01
2.61485904e-01 -9.27863121e-01 -7.61774957e-01 1.19327791e-01
4.29606140e-02 -9.52957273e-02 4.28374022e-01 1.34855545e+00
-3.30324471e-01 4.36946809e-01 1.45854577e-01 -7.91555524e-01
-1.14466107e+00 4.99104053e-01 7.03592539e-01 -2.35424966e-01
-4.91803527e-01 7.88418591e-01 5.94964862e-01 -1.89438790e-01
1.45190835e-01 -9.23515618e-01 1.65214151e-01 1.54547259e-01
3.34430307e-01 2.96973467e-01 2.44134203e-01 -5.28920472e-01
-4.81972396e-01 3.30458879e-01 -1.59634233e-01 4.71349955e-02
1.12796867e+00 -6.59154281e-02 -4.07585762e-02 2.26885200e-01
9.32851911e-01 2.47838214e-01 -1.01347518e+00 -3.97860333e-02
7.79131055e-02 -2.38696367e-01 4.18553576e-02 -1.02140582e+00
-7.64216959e-01 7.40750849e-01 9.43583012e-01 5.47244966e-01
1.13935709e+00 5.22021428e-02 3.66554201e-01 1.84350595e-01
-2.70741493e-01 -8.93368006e-01 -2.19087392e-01 -2.36052666e-02
6.63585126e-01 -1.43025315e+00 2.21850246e-01 -3.14938426e-01
-4.26417321e-01 1.04637229e+00 1.72192886e-01 -2.01215014e-01
4.19706076e-01 1.58372626e-01 2.21213683e-01 -2.66426176e-01
-4.84031677e-01 -1.22779191e-01 5.05461872e-01 8.47968042e-01
5.87744653e-01 1.56456530e-01 -4.08569604e-01 7.32729197e-01
-2.19026849e-01 -3.61639448e-02 7.19456792e-01 8.63997877e-01
-4.13388133e-01 -5.77068269e-01 -7.56343603e-01 8.30469608e-01
-7.88842440e-01 -6.48916885e-02 -2.90135056e-01 1.02848983e+00
4.75654989e-01 7.68147588e-01 2.13679418e-01 -7.09418282e-02
3.23211610e-01 -8.25253651e-02 3.88717085e-01 -6.22897744e-01
-6.21119320e-01 2.53593624e-01 1.30104005e-01 -4.87787962e-01
-3.18642348e-01 -8.03895295e-01 -9.49631989e-01 -4.65639308e-02
-1.04701549e-01 -2.77629495e-01 3.10242116e-01 8.31764102e-01
3.62398863e-01 7.25707650e-01 2.10957140e-01 -7.15693891e-01
-8.76005173e-01 -7.39963949e-01 -3.91385764e-01 6.09382153e-01
8.08849454e-01 -2.96680838e-01 -5.85552096e-01 4.88390923e-02] | [15.069869041442871, -2.318751811981201] |
a0111632-dee5-449d-9d27-f0839a819427 | conditional-generative-modeling-is-all-you | 2305.12569 | null | https://arxiv.org/abs/2305.12569v1 | https://arxiv.org/pdf/2305.12569v1.pdf | Conditional Generative Modeling is All You Need for Marked Temporal Point Processes | Recent advancements in generative modeling have made it possible to generate high-quality content from context information, but a key question remains: how to teach models to know when to generate content? To answer this question, this study proposes a novel event generative model that draws its statistical intuition from marked temporal point processes, and offers a clean, flexible, and computationally efficient solution for a wide range of applications involving multi-dimensional marks. We aim to capture the distribution of the point process without explicitly specifying the conditional intensity or probability density. Instead, we use a conditional generator that takes the history of events as input and generates the high-quality subsequent event that is likely to occur given the prior observations. The proposed framework offers a host of benefits, including exceptional efficiency in learning the model and generating samples, as well as considerable representational power to capture intricate dynamics in multi- or even high-dimensional event space. Our numerical results demonstrate superior performance compared to other state-of-the-art baselines. | ['Shixiang Zhu', 'Zekai Fan', 'Zheng Dong'] | 2023-05-21 | null | null | null | null | ['point-processes'] | ['methodology'] | [ 1.71066135e-01 -1.55262128e-01 -4.38069403e-02 -2.24025875e-01
-9.59200561e-01 -5.20794630e-01 1.02965879e+00 2.09912926e-01
-5.22209480e-02 9.08284485e-01 2.70644993e-01 -2.26210207e-01
-1.90803155e-01 -1.15284586e+00 -8.05111587e-01 -7.65865564e-01
-1.36302337e-01 7.04205334e-01 1.70844153e-01 1.10503733e-01
1.07257232e-01 4.38367367e-01 -1.71536076e+00 2.15706844e-02
9.22856569e-01 7.84501493e-01 3.14497381e-01 7.09401369e-01
-3.16133291e-01 8.21064591e-01 -6.98557019e-01 -4.35000509e-01
-1.26441494e-01 -6.67271078e-01 -3.73324126e-01 1.41724899e-01
-2.33598389e-02 -3.30046177e-01 -3.66742492e-01 6.53131664e-01
3.82441431e-01 2.36808434e-01 1.21894097e+00 -1.26762331e+00
-6.56727552e-01 4.80158865e-01 -4.85908538e-01 4.20130461e-01
3.55880290e-01 8.17427188e-02 9.16853547e-01 -7.26707816e-01
4.92203742e-01 1.20691812e+00 3.03215832e-01 3.10991198e-01
-1.26142502e+00 -6.04864240e-01 3.13281626e-01 6.11555530e-03
-1.25481462e+00 7.20623806e-02 7.28806615e-01 -4.78003651e-01
4.23647076e-01 1.94685217e-02 8.33369017e-01 1.59979510e+00
5.57966113e-01 1.12082446e+00 1.09735775e+00 -2.37353459e-01
5.76228678e-01 -1.54253840e-01 -1.87782019e-01 4.67380494e-01
1.20968640e-01 2.29685813e-01 -7.09278703e-01 -3.40611547e-01
1.17453659e+00 4.12676990e-01 5.93813788e-03 -8.26492384e-02
-1.04483604e+00 8.88857722e-01 -2.16542333e-02 -1.46805719e-01
-6.84911191e-01 3.80640656e-01 -1.61340818e-01 -1.77700505e-01
4.50624228e-01 -4.08375971e-02 -6.25407547e-02 -4.24223900e-01
-1.11584234e+00 6.08429790e-01 8.63470733e-01 1.01632226e+00
5.64308405e-01 1.40906706e-01 -5.37906110e-01 4.15114015e-01
3.54103178e-01 6.38315856e-01 9.63346660e-02 -6.31489515e-01
4.43610340e-01 2.48730525e-01 3.56519133e-01 -7.42132843e-01
6.86925650e-02 -4.41489309e-01 -9.68896508e-01 1.33577228e-01
3.98674518e-01 -3.19740713e-01 -1.25418723e+00 1.95326972e+00
4.99159902e-01 7.29404569e-01 -1.87565252e-01 5.63240170e-01
4.45154041e-01 1.12668025e+00 3.24930221e-01 -1.84120581e-01
1.34471416e+00 -3.53033394e-01 -7.69329011e-01 -1.47241771e-01
-2.27082804e-01 -7.69734561e-01 1.02764738e+00 2.15233117e-01
-1.35862470e+00 -4.84486848e-01 -7.66903996e-01 2.30163425e-01
-2.07540527e-01 -1.69299588e-01 8.90436709e-01 4.17143852e-01
-8.20251465e-01 4.63943392e-01 -1.17303467e+00 -1.15658775e-01
3.69311094e-01 -1.18600264e-01 2.09041342e-01 -8.19722712e-02
-1.26032674e+00 4.62131709e-01 2.79054552e-01 -2.64365077e-01
-1.40494156e+00 -9.33812559e-01 -6.98503256e-01 3.41841668e-01
3.91596824e-01 -9.53243375e-01 1.41405976e+00 -3.32569003e-01
-1.44694805e+00 2.49800101e-01 -3.77470255e-01 -4.01977301e-01
4.89710331e-01 -3.61270726e-01 -3.70078534e-01 1.75843418e-01
7.47011527e-02 3.99502754e-01 8.82961273e-01 -1.17012310e+00
-8.02792966e-01 1.33158445e-01 -1.57245278e-01 2.34863088e-01
6.51408508e-02 -2.33289838e-01 -5.39613247e-01 -9.75404143e-01
-7.27879107e-02 -8.49817514e-01 -2.26990238e-01 -1.47461623e-01
-3.53094608e-01 -1.76170275e-01 5.71897566e-01 -3.04615498e-01
1.20121026e+00 -1.98187530e+00 -4.69565317e-02 4.80864942e-02
-3.13689522e-02 -1.49565145e-01 2.11031109e-01 9.95247602e-01
1.18947215e-01 1.81979835e-01 -2.50264764e-01 -4.16569829e-01
9.46215317e-02 2.50109851e-01 -9.10880446e-01 5.90611398e-02
5.35955846e-01 7.96028912e-01 -1.08602691e+00 -6.25070572e-01
2.31767207e-01 6.77691281e-01 -4.80501771e-01 4.65115398e-01
-5.95603466e-01 3.94652039e-01 -6.13652408e-01 4.46856171e-01
3.40809822e-01 -6.24503851e-01 -1.34555653e-01 2.92691082e-01
7.46369511e-02 2.90053099e-01 -1.64040101e+00 1.40884197e+00
-2.24321663e-01 2.79167712e-01 -3.49489152e-01 -5.38265467e-01
8.08655381e-01 4.50263858e-01 4.47937578e-01 -3.13462853e-01
-6.23651110e-02 -3.09643358e-01 -4.46232796e-01 -1.07809313e-01
7.08293617e-01 -3.15689176e-01 -2.88400799e-01 7.57227361e-01
1.26679003e-01 -1.82512224e-01 3.22518796e-01 3.89062643e-01
1.15175545e+00 3.12537134e-01 3.18635523e-01 8.39519426e-02
-9.51344892e-03 -2.10896656e-01 5.44245243e-01 9.85890985e-01
2.59996891e-01 6.71983540e-01 5.75154185e-01 -1.02856934e-01
-1.01863968e+00 -1.82297647e+00 -8.20668563e-02 7.73151457e-01
9.10193995e-02 -3.63335788e-01 -4.60900068e-01 -2.42964357e-01
-1.17099136e-01 9.80110586e-01 -6.65648043e-01 -4.92099561e-02
-5.05929410e-01 -1.14177930e+00 1.03856564e-01 5.99733889e-01
2.62780309e-01 -1.08831501e+00 -3.29424083e-01 6.17636442e-01
-2.75726169e-01 -8.36751580e-01 -5.24913073e-01 1.37398973e-01
-9.25579607e-01 -8.52198839e-01 -6.91444278e-01 -4.89288479e-01
5.75130701e-01 3.38491537e-02 1.37265563e+00 -4.91253674e-01
-3.83266062e-01 4.64452475e-01 -1.52200162e-01 -7.30833828e-01
-3.98780316e-01 -4.02093632e-03 -3.69581193e-01 1.49511442e-01
1.60305947e-01 -7.08532393e-01 -7.34741509e-01 1.70137048e-01
-1.31599414e+00 1.10343553e-01 6.91316068e-01 6.98481321e-01
1.00051045e+00 3.37507576e-01 7.88228214e-01 -8.93645585e-01
8.16834986e-01 -8.70847702e-01 -5.67687392e-01 1.24026306e-01
-4.88572299e-01 2.95034021e-01 6.04839981e-01 -6.86362982e-01
-1.49423754e+00 -1.86066285e-01 -3.96937951e-02 -2.54250169e-01
-3.48325700e-01 4.72672999e-01 -3.54485773e-02 7.51717389e-01
4.28241551e-01 5.52483320e-01 -3.15608919e-01 -3.41241717e-01
5.17947435e-01 8.66911560e-02 6.73749506e-01 -1.06596935e+00
7.75608361e-01 6.25681698e-01 2.36967951e-01 -6.35213017e-01
-9.71976042e-01 -3.73078406e-01 -2.39955217e-01 -1.63771898e-01
7.87541807e-01 -1.10869539e+00 -4.99908060e-01 4.20681357e-01
-9.31540728e-01 -3.03638995e-01 -5.37093878e-01 4.21106398e-01
-7.83063769e-01 3.82181108e-02 -7.49612689e-01 -1.01686025e+00
-1.28417686e-01 -6.65402114e-01 1.26248050e+00 5.43674827e-01
-1.62029818e-01 -1.14980447e+00 3.65640521e-01 -3.30984473e-01
1.99637830e-01 4.89303142e-01 1.08184528e+00 -3.31694603e-01
-1.07964766e+00 -2.83813745e-01 1.67352110e-01 -1.33610547e-01
1.46187618e-01 2.09528059e-01 -6.06798351e-01 1.80964377e-02
-1.12796515e-01 5.70337474e-02 6.32781386e-01 5.94268799e-01
1.26974559e+00 -2.42581844e-01 -3.81665736e-01 3.95503789e-01
1.35769594e+00 3.16835523e-01 8.81860316e-01 -7.72090107e-02
2.47915387e-01 3.13418448e-01 5.03749013e-01 7.82615483e-01
4.98786747e-01 3.43037933e-01 2.69261599e-01 3.50841954e-02
-1.82987917e-02 -9.89644229e-01 1.81468010e-01 7.57503390e-01
7.40042403e-02 -7.04358816e-01 -6.46575689e-01 8.06834221e-01
-1.86264920e+00 -1.63704181e+00 1.90027524e-02 2.21548700e+00
1.03317130e+00 2.73517042e-01 1.52066678e-01 4.38177809e-02
7.03167677e-01 2.41421804e-01 -5.21574616e-01 1.11203097e-01
2.68246117e-03 3.92294616e-01 3.79699767e-02 2.89613932e-01
-9.99110818e-01 7.28325248e-01 7.34155130e+00 8.94451916e-01
-9.25058782e-01 -1.20712519e-01 6.70507193e-01 -2.22032532e-01
-4.27041233e-01 -4.85858805e-02 -1.18090534e+00 7.86743820e-01
1.16646445e+00 -4.17697817e-01 5.59858307e-02 6.69305325e-01
3.43638420e-01 -1.13243081e-01 -1.18033731e+00 7.76472867e-01
-1.06839508e-01 -1.56059480e+00 3.56141090e-01 7.07107484e-02
9.84401047e-01 -4.62543428e-01 2.75770664e-01 3.44697565e-01
9.72813070e-01 -9.47956324e-01 7.38187194e-01 1.03854549e+00
5.01505673e-01 -9.17764187e-01 2.68207252e-01 6.04861438e-01
-1.26764822e+00 2.96391308e-01 -1.53599530e-01 -2.32708137e-02
6.71564937e-01 7.73610950e-01 -1.03494680e+00 4.33757901e-01
4.43869054e-01 5.01504660e-01 -2.89957047e-01 1.27380109e+00
-4.23607528e-01 1.10101581e+00 -4.02417004e-01 -1.08135246e-01
1.82295993e-01 -1.58846572e-01 4.09742147e-01 1.21609712e+00
7.54082441e-01 2.45059803e-01 2.52760649e-01 1.10372388e+00
-5.63558610e-03 -2.00715840e-01 -4.76313204e-01 -1.08593136e-01
6.31735325e-01 1.17633808e+00 -8.41405392e-01 -4.29571956e-01
-2.49649838e-01 7.13085890e-01 2.47862160e-01 5.89288414e-01
-1.00318277e+00 -1.27711087e-01 6.29908204e-01 3.23608905e-01
5.95145702e-01 -3.51265818e-01 -1.74315840e-01 -1.06766260e+00
-2.00351566e-01 -5.81925631e-01 5.50893426e-01 -8.98901343e-01
-1.53079486e+00 3.26043606e-01 3.25327396e-01 -1.19112825e+00
-7.59128392e-01 -1.66065797e-01 -9.42422390e-01 1.05406809e+00
-1.39766312e+00 -8.68417382e-01 -1.59306943e-01 6.20877147e-01
5.88777542e-01 1.64637476e-01 6.02166414e-01 1.27581358e-01
-2.77778536e-01 2.00683400e-01 1.55932903e-01 -8.45944062e-02
5.20066321e-01 -1.54967165e+00 4.77586478e-01 9.07956243e-01
3.76061171e-01 5.28394997e-01 9.28637981e-01 -8.16633284e-01
-1.23122740e+00 -1.21072769e+00 6.17499948e-01 -6.93466246e-01
5.85945308e-01 -3.77181411e-01 -9.62539613e-01 6.77418888e-01
3.59337404e-02 -3.31713557e-01 7.11813450e-01 -2.81054564e-02
-7.85284117e-03 6.76007867e-02 -9.40623701e-01 8.92834544e-01
8.76940548e-01 -3.60932946e-01 -6.82166696e-01 1.82548150e-01
5.69412887e-01 -4.24344331e-01 -6.83269739e-01 -1.45083852e-02
1.48312375e-01 -6.32397950e-01 1.13054514e+00 -4.54524487e-01
8.08240473e-01 -3.10713053e-01 8.60443935e-02 -1.22653985e+00
-3.16838264e-01 -9.08292472e-01 -6.75169408e-01 1.40821981e+00
2.64679790e-01 -3.45409870e-01 7.80138791e-01 6.09853923e-01
3.66651230e-02 -8.34723175e-01 -6.85583949e-01 -5.20284116e-01
6.86911808e-04 -5.73986471e-01 8.32546532e-01 3.36091608e-01
-4.52175051e-01 3.47393066e-01 -6.21538937e-01 4.37687516e-01
8.90259922e-01 4.84504014e-01 7.92028487e-01 -1.24491715e+00
-3.55083674e-01 -1.63460195e-01 -6.65256232e-02 -1.37075913e+00
-1.31345868e-01 -5.88246107e-01 6.82919547e-02 -1.72757161e+00
2.07338765e-01 -3.26727390e-01 -1.50549531e-01 5.39284907e-02
-5.62113285e-01 -1.83421761e-01 4.31669541e-02 2.19433203e-01
-5.21413267e-01 8.77570033e-01 1.35533786e+00 3.21574360e-01
-8.73796046e-02 2.80604243e-01 -7.27280557e-01 6.34221077e-01
4.31607991e-01 -5.23945749e-01 -6.49163604e-01 -4.08158964e-03
3.08537334e-01 5.62477052e-01 6.25424147e-01 -8.97863448e-01
3.01422149e-01 -6.72064304e-01 6.91858530e-01 -1.10087132e+00
5.05013704e-01 -5.77818871e-01 5.82776725e-01 -1.86547767e-02
-4.10156220e-01 1.79667458e-01 8.36267695e-02 1.14367688e+00
-3.03079724e-01 -1.17837548e-01 5.43891251e-01 -2.45961890e-01
-4.15798813e-01 6.78770363e-01 -5.99346817e-01 2.75054514e-01
1.08079684e+00 1.12395063e-01 -1.92307860e-01 -8.17699194e-01
-6.19868636e-01 2.85285786e-02 2.35800788e-01 3.80638778e-01
5.15309632e-01 -1.53779030e+00 -6.02927208e-01 1.75519556e-01
-1.33140758e-01 1.19627520e-01 3.23590338e-01 1.08810008e-01
-1.27837628e-01 2.07219839e-01 1.28685638e-01 -6.59938097e-01
-5.96001327e-01 5.23950756e-01 -8.16945732e-02 -5.22305787e-01
-8.25404406e-01 4.92023766e-01 4.01130289e-01 5.28302975e-02
8.42100307e-02 -3.88143808e-01 1.02425724e-01 2.33450625e-02
6.96070254e-01 1.91870168e-01 -3.55706900e-01 -1.77387118e-01
1.59692362e-01 2.49350742e-01 -4.78600636e-02 -4.01097775e-01
1.28370631e+00 -8.72614309e-02 6.01612151e-01 8.47333312e-01
6.18914664e-01 -9.67456773e-02 -1.84181726e+00 -3.06259632e-01
-2.73785651e-01 -4.84550148e-01 -1.75557867e-01 -8.47311735e-01
-6.48136258e-01 8.39011729e-01 2.21927240e-01 5.22640109e-01
9.15108085e-01 9.70365778e-02 8.62833440e-01 -1.15810625e-01
3.98160875e-01 -9.15585816e-01 3.89230609e-01 4.28360134e-01
7.62511909e-01 -9.01150465e-01 -6.11340590e-02 -4.08471793e-01
-6.72314703e-01 7.58576274e-01 4.07111704e-01 -2.68484950e-01
7.62853742e-01 3.13130975e-01 -3.26794893e-01 -1.91544756e-01
-1.18184114e+00 -1.37166575e-01 4.13502395e-01 5.30624032e-01
2.99437135e-01 -1.20970868e-02 -3.23565193e-02 4.89820808e-01
-3.56742859e-01 6.80720881e-02 5.17818153e-01 1.00756800e+00
-4.48475659e-01 -9.89071369e-01 -3.93795937e-01 6.00783527e-01
-6.73495591e-01 3.09945340e-03 2.23254234e-01 7.58419156e-01
-2.43602395e-01 7.12452531e-01 2.28048697e-01 2.85750538e-01
1.82571918e-01 2.82622367e-01 2.79107988e-01 -6.80289805e-01
-2.54696012e-01 2.71888763e-01 -2.45304599e-01 -3.02747041e-01
-1.69671208e-01 -1.02509189e+00 -1.13418591e+00 -3.13895375e-01
-1.06915794e-02 1.90475836e-01 6.44125700e-01 8.51318717e-01
5.34888625e-01 7.32174397e-01 4.54497367e-01 -8.04982960e-01
-5.46175122e-01 -7.39820540e-01 -5.74274838e-01 5.40209055e-01
2.37561956e-01 -8.41571867e-01 -4.04102147e-01 3.43290120e-01] | [7.089450836181641, 3.54091477394104] |
c0be3a62-9eb1-44be-99d8-83e90ab5322e | 3d-bevis-birds-eye-view-instance-segmentation | 1904.02199 | null | https://arxiv.org/abs/1904.02199v3 | https://arxiv.org/pdf/1904.02199v3.pdf | 3D-BEVIS: Bird's-Eye-View Instance Segmentation | Recent deep learning models achieve impressive results on 3D scene analysis tasks by operating directly on unstructured point clouds. A lot of progress was made in the field of object classification and semantic segmentation. However, the task of instance segmentation is less explored. In this work, we present 3D-BEVIS, a deep learning framework for 3D semantic instance segmentation on point clouds. Following the idea of previous proposal-free instance segmentation approaches, our model learns a feature embedding and groups the obtained feature space into semantic instances. Current point-based methods scale linearly with the number of points by processing local sub-parts of a scene individually. However, to perform instance segmentation by clustering, globally consistent features are required. Therefore, we propose to combine local point geometry with global context information from an intermediate bird's-eye view representation. | ['Francis Engelmann', 'Cathrin Elich', 'Theodora Kontogianni', 'Bastian Leibe'] | 2019-04-03 | null | null | null | null | ['3d-instance-segmentation-1', '3d-semantic-instance-segmentation'] | ['computer-vision', 'computer-vision'] | [-1.08701818e-01 2.03508362e-02 3.67051549e-02 -7.65601814e-01
-4.74779516e-01 -6.49391055e-01 5.45235336e-01 4.66343850e-01
-2.08795384e-01 -1.45358071e-01 -4.23306853e-01 -1.42813949e-02
-1.07690450e-02 -1.11841750e+00 -8.98387015e-01 -3.69076639e-01
3.04798763e-02 9.73587930e-01 4.82617378e-01 -3.43360640e-02
5.14001489e-01 1.13248897e+00 -1.66271269e+00 1.67824119e-01
6.99490309e-01 1.02358186e+00 3.20966691e-01 3.77545595e-01
-9.25326824e-01 5.50192110e-02 -3.29046577e-01 8.73618200e-02
6.89236403e-01 -1.35858655e-01 -1.02716565e+00 7.26378918e-01
6.24246836e-01 -1.71870992e-01 1.90668941e-01 8.79310489e-01
1.77969351e-01 2.63285309e-01 6.47204041e-01 -1.23408508e+00
-2.62018681e-01 -2.59516407e-02 -6.49588585e-01 -1.44203231e-01
1.32706746e-01 -2.20252573e-02 9.33067799e-01 -1.04148746e+00
6.19927347e-01 1.32413101e+00 6.82138860e-01 2.67886877e-01
-1.20103550e+00 -2.06998348e-01 4.53731537e-01 1.71461906e-02
-1.32637775e+00 1.69571593e-01 1.21039355e+00 -7.02565372e-01
1.14538717e+00 1.25383800e-02 1.09109807e+00 2.95669079e-01
-2.76424170e-01 9.56519783e-01 1.04274118e+00 -2.01375932e-01
5.74316561e-01 -1.15300767e-01 3.31867725e-01 6.76452756e-01
2.99613420e-02 -2.22717747e-01 -1.32776260e-01 2.64179222e-02
1.13195550e+00 4.19843793e-01 2.61196047e-01 -1.02745032e+00
-9.87387598e-01 8.73993933e-01 1.00461483e+00 2.45902419e-01
-4.06709701e-01 2.26726204e-01 7.42025822e-02 1.43660322e-01
9.63352919e-01 4.42559391e-01 -7.88314283e-01 2.27048039e-01
-1.17634881e+00 3.51496100e-01 6.36209488e-01 1.03802347e+00
1.31789482e+00 -3.14077497e-01 2.96252668e-01 6.31570697e-01
4.34029728e-01 2.69420385e-01 -1.26815364e-01 -1.19400907e+00
2.56844312e-01 1.35903192e+00 -3.36168073e-02 -9.20314729e-01
-5.82484841e-01 -2.35831156e-01 -4.18164045e-01 5.29106259e-01
2.82146335e-01 3.40835214e-01 -1.31307638e+00 9.70544100e-01
8.05673361e-01 3.19056004e-01 -2.09444255e-01 1.02903605e+00
9.46911752e-01 7.33650029e-01 -1.20338798e-01 4.47423458e-01
1.05963361e+00 -8.45773697e-01 1.22485608e-01 -2.41329148e-01
5.61450362e-01 -5.88931322e-01 8.77255559e-01 2.47729018e-01
-1.06932878e+00 -7.42591202e-01 -7.73232400e-01 -4.51707035e-01
-5.99087834e-01 -1.35003284e-01 6.02851987e-01 3.34376782e-01
-1.09342527e+00 5.91719747e-01 -1.14643967e+00 -2.82834053e-01
9.34152782e-01 5.51318645e-01 -3.09859157e-01 -1.72931552e-01
-3.85263324e-01 5.07952332e-01 4.33483303e-01 -4.52100299e-02
-6.72702312e-01 -8.03691268e-01 -9.03518796e-01 -2.52441759e-03
2.57015675e-01 -9.20623064e-01 1.09910917e+00 -6.91819549e-01
-1.43170273e+00 1.35057461e+00 -1.45202026e-01 -2.55083174e-01
4.05777901e-01 -4.06174392e-01 5.95392346e-01 2.46819779e-01
1.27567321e-01 1.03795135e+00 8.22254419e-01 -1.68871868e+00
-7.01015174e-01 -9.07064438e-01 3.10122252e-01 3.55497718e-01
3.92224610e-01 -3.24103326e-01 -6.29309297e-01 -2.40378827e-03
8.56095493e-01 -8.67814422e-01 -7.60268390e-01 1.97053283e-01
-4.30093408e-01 -6.28924549e-01 1.09410310e+00 -1.38899773e-01
2.34206215e-01 -2.21688628e+00 3.21184635e-01 2.34838232e-01
4.40640092e-01 -6.05223980e-03 2.15480365e-02 2.85060227e-01
8.14874917e-02 2.38997474e-01 -5.49261451e-01 -5.98766923e-01
1.02331549e-01 3.00089091e-01 -1.09146439e-01 5.73054075e-01
3.53847086e-01 1.00258219e+00 -7.32347488e-01 -5.12910664e-01
8.10129344e-01 4.35122848e-01 -8.24642301e-01 3.12737554e-01
-7.55365968e-01 5.94489634e-01 -8.10203552e-01 6.55930877e-01
1.04233050e+00 -3.75763863e-01 -3.98811430e-01 2.11115107e-02
-3.89676154e-01 1.98538035e-01 -1.08013916e+00 2.22796917e+00
-4.02510941e-01 4.20033008e-01 1.19871080e-01 -1.29435098e+00
1.11201060e+00 2.89918836e-02 9.34154034e-01 -2.17034534e-01
9.30379704e-02 2.16089711e-01 -5.32630205e-01 -2.39630967e-01
2.87646830e-01 -6.92193359e-02 -1.10680409e-01 2.35835895e-01
5.60105853e-02 -1.10127532e+00 -2.43621275e-01 -3.54470573e-02
6.67146981e-01 3.84410024e-01 1.80210844e-01 -6.66321293e-02
4.28814381e-01 5.26850522e-01 2.47715473e-01 5.39588392e-01
1.26270456e-02 9.50139999e-01 2.17151821e-01 -7.18564987e-01
-1.09566343e+00 -1.13068843e+00 -4.14302051e-01 6.56209767e-01
7.60550797e-01 -3.43627512e-01 -8.92347336e-01 -7.56106138e-01
1.69296056e-01 4.00427669e-01 -3.50510895e-01 3.80479366e-01
-5.96598089e-01 -4.13135618e-01 -2.47409716e-02 5.50191760e-01
6.51698291e-01 -1.00038767e+00 -9.08452570e-01 2.07247406e-01
2.59004533e-01 -1.12800455e+00 3.61859463e-02 2.25255787e-01
-1.36132801e+00 -1.04790747e+00 -5.68895102e-01 -8.20116162e-01
8.10326993e-01 5.44535279e-01 1.28489614e+00 7.03224763e-02
-3.77736390e-01 6.90523565e-01 -3.06973994e-01 -5.02031744e-01
4.63778675e-02 2.70169318e-01 -2.91373461e-01 -1.63570315e-01
7.26186991e-01 -6.85236752e-01 -6.11719310e-01 1.55408189e-01
-7.96827555e-01 2.06753835e-01 3.85517150e-01 2.75300771e-01
1.18529356e+00 -6.49680421e-02 -1.15957499e-01 -7.69084930e-01
7.73488134e-02 -3.59583348e-01 -8.54554355e-01 -2.12067679e-01
-8.56634304e-02 -5.41610382e-02 3.16448569e-01 2.78597564e-01
-7.96982348e-01 4.88326758e-01 -4.16505098e-01 -7.98866153e-01
-1.03102148e+00 1.71626419e-01 -1.67475477e-01 -1.57244965e-01
3.37205917e-01 7.73019418e-02 -2.41176903e-01 -6.99475646e-01
5.97094476e-01 4.43860948e-01 5.63580282e-02 -4.81550843e-01
8.10325503e-01 9.10319209e-01 1.64176047e-01 -1.07354259e+00
-1.00253260e+00 -9.47541773e-01 -1.41001737e+00 -1.14274763e-01
1.29546332e+00 -8.21559668e-01 -4.28158969e-01 4.22162950e-01
-1.37606227e+00 -4.25755292e-01 -6.05292082e-01 2.73725390e-01
-9.63068783e-01 2.43580997e-01 -2.23900676e-01 -5.71068168e-01
2.62062671e-03 -1.15050399e+00 1.81721842e+00 1.05517298e-01
7.04225153e-02 -9.80704784e-01 1.74541205e-01 4.18992549e-01
-1.97649524e-01 4.40125197e-01 8.13128531e-01 -5.67836642e-01
-1.23811710e+00 -2.14977428e-01 -3.37016314e-01 3.55600923e-01
7.26434737e-02 -1.42800152e-01 -9.36803877e-01 1.31008983e-01
2.47690454e-01 -6.35545701e-02 7.58529186e-01 6.17724359e-01
1.47045064e+00 2.07019597e-01 -4.85090971e-01 9.39517260e-01
1.65408897e+00 -8.85197744e-02 2.53407687e-01 1.69667512e-01
1.12886822e+00 6.96465731e-01 4.92642194e-01 2.75556475e-01
3.75636667e-01 5.44139981e-01 8.11227798e-01 -2.62740344e-01
5.14545217e-02 -3.05393368e-01 -2.55498856e-01 5.49676478e-01
-1.30994976e-01 -3.62678766e-02 -1.20660710e+00 6.53094172e-01
-1.68297386e+00 -6.21457040e-01 -3.47467959e-01 1.92619014e+00
2.15841249e-01 1.09724112e-01 7.11278394e-02 2.12108389e-01
4.45793480e-01 2.24212319e-01 -6.45883381e-01 -3.26889634e-01
1.83853164e-01 2.97180772e-01 4.03127551e-01 4.53133374e-01
-1.22087157e+00 1.31091940e+00 5.30333090e+00 5.62779784e-01
-1.05325496e+00 -2.94478573e-02 4.53313768e-01 1.18617848e-01
-3.12792212e-01 7.35620558e-02 -8.14875245e-01 5.25344871e-02
2.08838031e-01 2.14371100e-01 2.08439697e-02 1.24120831e+00
9.82677042e-02 -2.12983484e-03 -1.25445151e+00 1.14185011e+00
-6.60191923e-02 -1.37639916e+00 1.77141011e-01 1.88661754e-01
7.62321115e-01 4.85587925e-01 -4.66249362e-02 8.09714422e-02
2.84736574e-01 -9.75874305e-01 5.00107288e-01 3.47117454e-01
4.03399825e-01 -6.84347451e-01 3.66402358e-01 7.19064832e-01
-1.19920635e+00 1.29943132e-01 -5.26723802e-01 -1.09725334e-01
2.20061332e-01 7.86131918e-01 -8.89038801e-01 5.57979465e-01
9.53049839e-01 1.07225347e+00 -4.95166540e-01 1.23893273e+00
-2.36551035e-02 3.80210787e-01 -8.34562123e-01 2.90235072e-01
7.99174011e-01 -6.30532444e-01 5.41569710e-01 9.09357071e-01
1.88915461e-01 2.64636040e-01 6.07970476e-01 1.22419763e+00
2.53450428e-03 1.16893217e-01 -7.56583273e-01 1.59244537e-01
9.02761146e-02 1.14724028e+00 -1.24292362e+00 -2.26991236e-01
-3.78449827e-01 9.36894417e-01 3.33242953e-01 2.40172237e-01
-3.73663753e-01 -9.88206714e-02 7.57803738e-01 2.46922269e-01
4.56464291e-01 -9.57813144e-01 -6.96506381e-01 -9.64341104e-01
-1.63722217e-01 -3.11728436e-02 -4.08850573e-02 -7.44462967e-01
-1.23557305e+00 2.51654655e-01 2.01856568e-01 -1.20263827e+00
9.92492884e-02 -6.39578998e-01 -5.18086612e-01 7.04086781e-01
-1.60011864e+00 -1.18086708e+00 -4.36454326e-01 6.18987143e-01
1.02718949e+00 3.12821299e-01 5.01188159e-01 -1.37811795e-01
-4.02070917e-02 -2.20284656e-01 -1.78819243e-02 1.40093878e-01
-1.01376422e-01 -1.46771562e+00 7.61974454e-01 3.91835600e-01
4.87755984e-01 4.14214432e-01 2.94083297e-01 -5.86667717e-01
-1.09956706e+00 -1.13692379e+00 5.67588985e-01 -7.27477729e-01
1.20649152e-01 -5.50261378e-01 -1.08209121e+00 5.64826369e-01
-1.50344595e-01 2.58326918e-01 4.71766263e-01 6.89278916e-02
-1.58869460e-01 1.65189743e-01 -1.22046959e+00 4.44689542e-01
1.38151801e+00 -4.71980810e-01 -8.09799850e-01 4.23081905e-01
8.47938240e-01 -6.83367431e-01 -8.36968124e-01 4.29050952e-01
-1.64632469e-01 -1.05393958e+00 1.37200141e+00 -5.45991540e-01
3.46052676e-01 -4.12919104e-01 -1.13849938e-01 -1.22895348e+00
-1.63103312e-01 -8.32881555e-02 1.09519131e-01 6.10312581e-01
-3.33012491e-02 -2.30378300e-01 1.21918857e+00 3.73821676e-01
-6.06578708e-01 -6.51994765e-01 -9.90908086e-01 -5.22966743e-01
2.77715653e-01 -6.74839139e-01 8.50313842e-01 7.39737272e-01
-6.54030085e-01 1.71067864e-01 4.54061300e-01 3.99443120e-01
8.91404688e-01 8.16462696e-01 1.08403134e+00 -1.84283304e+00
2.92934775e-01 -4.24329907e-01 -7.84606397e-01 -1.51530588e+00
2.86079228e-01 -1.03503537e+00 -5.25223501e-02 -1.98376667e+00
-3.18314254e-01 -7.55540252e-01 8.76417980e-02 2.98192233e-01
1.41879261e-01 2.19915688e-01 2.90844589e-01 3.45242411e-01
-7.54357100e-01 5.63622952e-01 1.23613799e+00 -1.40381023e-01
-5.33347368e-01 2.47539699e-01 -1.24618143e-01 8.71419489e-01
9.54143584e-01 -4.05861080e-01 -3.53061736e-01 -7.13646293e-01
7.50552937e-02 -1.75028831e-01 6.11885905e-01 -1.07131815e+00
2.15966001e-01 -2.79872805e-01 4.22193408e-01 -1.20770514e+00
6.75007999e-01 -1.05710709e+00 -9.60772038e-02 -1.45673221e-02
-1.06240273e-01 -3.35183769e-01 1.49342939e-01 6.18606448e-01
-2.71268636e-01 -2.00313479e-01 7.48773515e-01 -7.25233018e-01
-9.65005875e-01 7.85896480e-01 -7.62025565e-02 -9.27653983e-02
1.14691103e+00 -6.27686203e-01 3.40778172e-01 2.70402998e-01
-8.49145889e-01 2.42991388e-01 8.92098665e-01 4.19345260e-01
9.22805905e-01 -9.28294361e-01 -4.76735532e-01 3.44702005e-01
6.33366480e-02 1.14359736e+00 2.67012000e-01 3.44160616e-01
-8.94451201e-01 4.79025602e-01 -6.95296824e-02 -1.43186653e+00
-9.87997770e-01 5.04583538e-01 4.95612174e-01 3.18078935e-01
-1.09483957e+00 9.28358614e-01 4.66368437e-01 -8.98114979e-01
-4.73303720e-02 -5.61245084e-01 -2.34019816e-01 -1.97098032e-02
-1.16187684e-01 1.80772871e-01 1.08743131e-01 -6.72277451e-01
-3.14440250e-01 1.21643865e+00 1.55598164e-01 9.90491509e-02
1.59928584e+00 -1.03470616e-01 -2.25170046e-01 6.06951118e-01
1.40163255e+00 -3.85040343e-01 -1.41116846e+00 -2.87263066e-01
1.70268953e-01 -6.36919618e-01 2.79268742e-01 -1.67634383e-01
-1.14919567e+00 1.39327908e+00 4.86504436e-01 3.83298486e-01
7.73426652e-01 5.73343873e-01 6.76089406e-01 3.85915816e-01
5.71697593e-01 -8.93312871e-01 -2.06077304e-02 6.44194961e-01
6.29538178e-01 -1.40559483e+00 -1.67867858e-02 -6.26242995e-01
-1.80736318e-01 1.05245888e+00 5.83061218e-01 -5.63325167e-01
9.53271449e-01 -1.33900329e-01 9.22722835e-03 -8.00659180e-01
-1.32707924e-01 -4.52034593e-01 3.37464124e-01 6.60769641e-01
9.81249064e-02 3.57960686e-02 2.07192853e-01 2.04645738e-01
-2.68388510e-01 -1.08740896e-01 1.59110591e-01 8.64150703e-01
-7.64557719e-01 -9.23327327e-01 -4.64843929e-01 4.03747648e-01
2.89309286e-02 2.35669106e-01 -5.82816184e-01 7.64329135e-01
3.04541767e-01 5.59879959e-01 5.47772527e-01 -2.05841269e-02
4.38387305e-01 1.76298730e-02 4.85426664e-01 -1.20769107e+00
-3.73555899e-01 -1.49068469e-02 -6.06526196e-01 -7.26393759e-01
-6.32557631e-01 -7.28630960e-01 -1.63913441e+00 5.33774383e-02
-4.35794815e-02 8.93756375e-02 1.09512746e+00 9.81225193e-01
3.76056194e-01 1.44600645e-01 7.37966776e-01 -1.54806387e+00
3.01348995e-02 -3.79250944e-01 -6.04377270e-01 3.85934025e-01
3.33738476e-01 -7.10912049e-01 -4.48872209e-01 -3.21848020e-02] | [8.061105728149414, -3.20847225189209] |
54e58487-f66c-43e2-b671-0d1d7a1707f1 | discriminative-unsupervised-feature-learning | 1406.6909 | null | http://arxiv.org/abs/1406.6909v2 | http://arxiv.org/pdf/1406.6909v2.pdf | Discriminative Unsupervised Feature Learning with Exemplar Convolutional Neural Networks | Deep convolutional networks have proven to be very successful in learning
task specific features that allow for unprecedented performance on various
computer vision tasks. Training of such networks follows mostly the supervised
learning paradigm, where sufficiently many input-output pairs are required for
training. Acquisition of large training sets is one of the key challenges, when
approaching a new task. In this paper, we aim for generic feature learning and
present an approach for training a convolutional network using only unlabeled
data. To this end, we train the network to discriminate between a set of
surrogate classes. Each surrogate class is formed by applying a variety of
transformations to a randomly sampled 'seed' image patch. In contrast to
supervised network training, the resulting feature representation is not class
specific. It rather provides robustness to the transformations that have been
applied during training. This generic feature representation allows for
classification results that outperform the state of the art for unsupervised
learning on several popular datasets (STL-10, CIFAR-10, Caltech-101,
Caltech-256). While such generic features cannot compete with class specific
features from supervised training on a classification task, we show that they
are advantageous on geometric matching problems, where they also outperform the
SIFT descriptor. | ['Thomas Brox', 'Martin Riedmiller', 'Alexey Dosovitskiy', 'Philipp Fischer', 'Jost Tobias Springenberg'] | 2014-06-26 | null | null | null | null | ['geometric-matching'] | ['computer-vision'] | [ 4.17051047e-01 -1.85547128e-01 -2.73938388e-01 -6.28613174e-01
-6.44552410e-01 -5.11118233e-01 1.05097890e+00 3.49631280e-01
-8.02569687e-01 5.99397659e-01 -2.99888819e-01 4.61337231e-02
-4.44676071e-01 -7.87185490e-01 -7.46483147e-01 -7.80576229e-01
-1.32728651e-01 5.90282738e-01 2.34422132e-01 -2.67959356e-01
2.69454420e-01 9.47605729e-01 -1.96248996e+00 2.46312603e-01
4.50827390e-01 1.25449479e+00 1.18654974e-01 3.34167540e-01
-2.32094247e-03 4.98526871e-01 -4.54801261e-01 -1.07937098e-01
5.12980402e-01 -2.33989879e-01 -8.81724894e-01 8.18037987e-02
9.17101860e-01 1.11709006e-01 -3.42811644e-01 1.00323713e+00
2.60953397e-01 2.02974558e-01 1.09612632e+00 -1.29680717e+00
-3.32215816e-01 2.57885873e-01 -9.46298838e-02 5.97405806e-02
1.59113750e-01 -1.11049511e-01 1.01012540e+00 -9.27737772e-01
7.79566884e-01 9.51379597e-01 7.63402402e-01 4.03243840e-01
-1.25808358e+00 -4.80101407e-01 -2.84161955e-01 2.10495815e-01
-1.34881830e+00 -3.32407326e-01 8.39302897e-01 -4.97992098e-01
7.73796856e-01 4.77179959e-02 4.71944600e-01 9.87652481e-01
9.34463590e-02 7.83184946e-01 1.07762623e+00 -5.24010897e-01
1.86987787e-01 9.89287272e-02 1.07478224e-01 5.44323087e-01
2.89490670e-01 5.03221571e-01 -3.61270040e-01 1.41514853e-01
7.64843345e-01 2.14352772e-01 -1.47545025e-01 -9.22423422e-01
-1.19118083e+00 9.81918752e-01 9.37537968e-01 6.08230591e-01
-2.04161391e-01 -5.82967140e-03 4.67249572e-01 8.29642773e-01
1.73611403e-01 7.06076920e-01 -6.02523506e-01 2.63376534e-01
-1.15472782e+00 1.76258132e-01 6.29654348e-01 8.38189542e-01
1.14472497e+00 2.47662533e-02 -5.96185736e-02 7.09190249e-01
3.94639112e-02 4.18722183e-01 7.30565012e-01 -4.53690022e-01
1.36052206e-01 5.69398940e-01 -5.43708391e-02 -9.24027681e-01
-5.98937094e-01 -6.62033379e-01 -1.03679681e+00 5.11753798e-01
6.71961725e-01 1.42006800e-01 -1.06701589e+00 1.40201199e+00
8.14432427e-02 -5.58014326e-02 1.42838299e-01 6.66590929e-01
8.08307171e-01 2.99166083e-01 -1.73155650e-01 1.57867193e-01
9.52058554e-01 -8.40144515e-01 -1.95719987e-01 -2.10664779e-01
7.64561713e-01 -9.17366266e-01 6.72848642e-01 5.37901163e-01
-5.73273182e-01 -1.01781130e+00 -1.26522505e+00 2.40177244e-01
-7.86684394e-01 2.69317180e-01 7.46475160e-01 4.08098340e-01
-9.98632610e-01 9.10779655e-01 -4.43746239e-01 -5.40587425e-01
6.53113663e-01 6.93532228e-01 -7.74211228e-01 -2.56205589e-01
-9.70591664e-01 8.94065738e-01 5.79622269e-01 -8.41122568e-02
-7.03380346e-01 -3.78600389e-01 -9.30087268e-01 8.06313902e-02
9.63622183e-02 -2.26820797e-01 1.03289342e+00 -1.30920231e+00
-1.24978435e+00 1.03664708e+00 4.45878059e-01 -7.16839910e-01
4.84902143e-01 -8.21602941e-02 -3.61606389e-01 1.94088161e-01
-7.23607419e-03 8.87599349e-01 1.25015342e+00 -1.12228429e+00
-5.02131939e-01 -1.94092840e-01 -1.13793068e-01 -2.08061919e-01
-3.57062966e-01 -1.10807315e-01 3.33590396e-02 -6.79280162e-01
1.54077202e-01 -9.64960158e-01 -2.45189741e-01 1.47539601e-01
-2.14434862e-01 -2.04678550e-01 1.09367228e+00 1.98011622e-01
4.36850965e-01 -2.13189340e+00 -6.84484690e-02 4.93554533e-01
-2.94857714e-02 7.52142489e-01 -2.92980403e-01 6.24466538e-01
-3.73674631e-01 -2.51941502e-01 -2.47378200e-01 -7.05134645e-02
-4.90684062e-02 2.11220771e-01 -2.44250447e-01 7.24570572e-01
5.51338553e-01 1.03690231e+00 -7.91529000e-01 -2.55672306e-01
5.90273917e-01 5.19308031e-01 -3.46508801e-01 1.43384844e-01
-9.20191184e-02 4.52583343e-01 -3.45140904e-01 3.99498016e-01
6.49267137e-01 -1.18983909e-01 -1.00262664e-01 -2.02176616e-01
8.85140747e-02 -9.04490873e-02 -1.19780040e+00 1.65511799e+00
-5.91769099e-01 8.19076836e-01 -3.54610085e-01 -1.66070855e+00
1.34667408e+00 1.30516127e-01 4.30830330e-01 -6.50893986e-01
3.79481405e-01 4.09239322e-01 2.80969650e-01 -3.37674528e-01
1.45920500e-01 -5.44787906e-02 -1.20015509e-01 3.54534268e-01
7.73501158e-01 -3.18946421e-01 3.09140921e-01 -2.00927019e-01
9.73700166e-01 9.70735308e-03 3.15341055e-01 -4.53136295e-01
8.11162889e-01 -7.67021254e-03 2.20512643e-01 7.96765566e-01
-7.57625625e-02 8.27663302e-01 1.62965193e-01 -9.00329471e-01
-9.78045583e-01 -9.42204714e-01 -5.17456830e-01 8.82296979e-01
5.90101117e-03 -2.52168238e-01 -5.53658605e-01 -8.25571418e-01
1.77919477e-01 7.30828270e-02 -7.87527382e-01 -4.17240083e-01
-5.62139213e-01 -3.42937589e-01 4.49067265e-01 5.14792740e-01
6.69117093e-01 -1.24520707e+00 -7.20557630e-01 1.81667298e-01
5.69531977e-01 -1.09172821e+00 -3.97669114e-02 6.11741900e-01
-9.98573244e-01 -1.38717067e+00 -6.78638399e-01 -1.06943345e+00
8.65323544e-01 2.24487752e-01 1.13580465e+00 2.51870662e-01
-5.85675478e-01 3.48785073e-01 -5.01755893e-01 -3.37636828e-01
-2.25541741e-01 2.99837857e-01 1.56322747e-01 2.88077354e-01
4.41594779e-01 -5.71436942e-01 -4.36358184e-01 3.57726365e-01
-1.17365909e+00 -4.00603294e-01 9.55307901e-01 1.24159980e+00
4.57055390e-01 4.76007648e-02 5.30005872e-01 -8.39950740e-01
2.74990350e-01 -1.23885132e-01 -5.58105767e-01 1.18996985e-01
-3.79193068e-01 3.23187232e-01 1.05985069e+00 -4.77850944e-01
-5.00757694e-01 6.69658482e-01 -1.73993915e-01 -2.92157829e-01
-5.39863408e-01 4.56719041e-01 -3.83108631e-02 -6.73134625e-01
9.49300408e-01 2.98885435e-01 1.45223960e-01 -5.07417142e-01
1.97060674e-01 4.47206616e-01 6.02377355e-01 -5.04925430e-01
1.21271300e+00 3.84157538e-01 3.41387331e-01 -8.46285582e-01
-8.63001287e-01 -6.53037667e-01 -1.27314734e+00 -2.35727578e-02
5.53626597e-01 -6.03728414e-01 -2.91327596e-01 5.88710427e-01
-8.10564935e-01 -2.12894082e-01 -6.07600927e-01 5.85951984e-01
-7.31255889e-01 -1.49817644e-02 -5.23582399e-02 -2.30897516e-01
-8.42862576e-02 -1.08246183e+00 9.83817577e-01 3.62838656e-01
1.73682738e-02 -1.22301054e+00 6.93300366e-02 -3.04387242e-01
6.08326375e-01 3.51893067e-01 6.42726839e-01 -1.11346638e+00
-3.49391490e-01 -6.45491362e-01 -2.02291816e-01 5.57393610e-01
3.66657168e-01 5.31327538e-02 -1.03441739e+00 -7.16358185e-01
-2.65914798e-01 -7.44402647e-01 1.17072344e+00 7.41039589e-02
1.14238548e+00 1.71307698e-01 -2.67791450e-01 8.29375744e-01
1.50241840e+00 -4.11945805e-02 6.09601021e-01 5.49590170e-01
3.60868394e-01 4.15367603e-01 6.03638470e-01 -2.59464118e-03
-2.09834695e-01 6.92635238e-01 5.02950072e-01 -3.27512741e-01
-1.52710244e-01 -9.96063929e-03 -7.96514284e-03 2.65628636e-01
-3.84582430e-02 1.68737635e-01 -8.89658034e-01 4.27650273e-01
-1.66630411e+00 -8.00562561e-01 3.54801714e-02 2.49988580e+00
4.50191885e-01 3.44975740e-01 1.77935213e-01 5.96593976e-01
5.08437812e-01 4.88393605e-02 -3.08727413e-01 -2.05086678e-01
-1.98878884e-01 8.69884551e-01 5.49662352e-01 1.11693047e-01
-1.56176257e+00 9.90149140e-01 5.87740183e+00 7.51273930e-01
-1.54900360e+00 -1.77572891e-01 3.40042710e-01 4.65846717e-01
2.39458039e-01 -2.30217665e-01 -6.80037856e-01 3.33744511e-02
8.14332247e-01 1.53248593e-01 2.23887768e-02 8.64544630e-01
-3.89417559e-01 3.82010965e-03 -1.34843695e+00 1.07495725e+00
1.24673747e-01 -1.37872887e+00 1.57359332e-01 -5.04058339e-02
7.85734951e-01 1.22968979e-01 2.72977799e-01 4.04674470e-01
8.53947848e-02 -1.41528857e+00 3.59952569e-01 3.34613919e-01
8.38898063e-01 -8.19835603e-01 8.16952765e-01 2.09558636e-01
-1.14201939e+00 -6.91305101e-02 -7.97593236e-01 5.18107265e-02
-3.13990414e-01 4.14405644e-01 -6.45406485e-01 7.19641030e-01
5.48380256e-01 9.06313896e-01 -9.98470247e-01 1.48744881e+00
-2.24013835e-01 2.84040660e-01 -3.93654794e-01 9.01328474e-02
6.22923017e-01 3.95525210e-02 1.16476193e-01 1.21263182e+00
2.92318761e-01 -5.37420332e-01 1.22221775e-01 4.27806050e-01
-1.60589993e-01 2.59069085e-01 -1.04424679e+00 1.28446281e-01
7.60728493e-02 1.38898039e+00 -8.63095403e-01 -2.66327471e-01
-3.91069204e-01 8.88216019e-01 4.58873481e-01 1.22818455e-01
-3.80430222e-01 -6.19375169e-01 4.32044685e-01 -5.05281426e-02
6.98969781e-01 -2.83866942e-01 2.55971372e-01 -1.08144927e+00
-7.76624456e-02 -7.68486917e-01 2.99367160e-01 -3.14223856e-01
-1.26105070e+00 8.39427114e-01 -7.10903406e-02 -1.61603773e+00
-4.09303606e-01 -1.16341066e+00 -8.33808601e-01 5.88017702e-01
-1.80021644e+00 -1.25573885e+00 -3.87827337e-01 8.41288447e-01
4.23975140e-01 -5.69570482e-01 9.88174915e-01 1.97790042e-01
-7.41637722e-02 6.77587271e-01 3.79568934e-01 4.46817845e-01
9.33750808e-01 -1.26728892e+00 3.57336789e-01 5.73067427e-01
7.76557267e-01 4.45644051e-01 5.34982920e-01 -5.13551719e-02
-1.23435092e+00 -1.13871861e+00 7.27272570e-01 -1.17484562e-01
4.75201041e-01 -3.82627189e-01 -8.99008930e-01 4.44211453e-01
2.18285359e-02 8.54986548e-01 5.08993208e-01 -6.07054830e-02
-7.11653709e-01 -3.34331930e-01 -1.18513250e+00 5.47360145e-02
8.79911065e-01 -4.47585016e-01 -6.42032862e-01 3.94353777e-01
-2.57176068e-02 -1.14353634e-01 -8.39358330e-01 4.80669826e-01
5.66054463e-01 -1.14043832e+00 1.07772708e+00 -9.37209725e-01
3.07780117e-01 -2.09188953e-01 -1.18352450e-01 -1.51368964e+00
-1.90194309e-01 -4.72838283e-01 2.90082067e-01 9.24087763e-01
3.10053974e-01 -6.12692654e-01 8.93818557e-01 -4.94148061e-02
9.94704943e-03 -6.05778992e-01 -7.99957037e-01 -1.23993421e+00
3.58586192e-01 -1.25226110e-01 4.14683342e-01 1.00783968e+00
-4.60349560e-01 1.85647413e-01 -1.78172708e-01 -1.58279851e-01
6.28159165e-01 3.18875074e-01 9.18388724e-01 -1.84750807e+00
-1.33005194e-02 -4.78352904e-01 -1.05913150e+00 -7.91431010e-01
3.66892129e-01 -1.11098635e+00 -5.83823100e-02 -1.15167618e+00
-1.87627986e-01 -6.96035147e-01 -5.58267057e-01 5.49597383e-01
2.33032510e-01 5.21720290e-01 2.04416931e-01 1.87533319e-01
-5.32780170e-01 4.00816023e-01 1.09904194e+00 -2.83996254e-01
5.73050231e-03 2.59965658e-01 -3.45420122e-01 6.24540091e-01
8.43433917e-01 -4.57307547e-01 -3.18242282e-01 -1.19596526e-01
-1.33392408e-01 -4.19422776e-01 4.17034358e-01 -1.49404919e+00
2.42482573e-01 1.20900907e-01 7.03144670e-01 -3.64950418e-01
1.82105005e-01 -1.11892509e+00 -5.10124117e-02 6.23215139e-01
-3.95904183e-01 -1.78428054e-01 8.76525193e-02 3.62642854e-01
-6.35622621e-01 -5.42785764e-01 1.07433045e+00 -7.06461072e-02
-9.79123831e-01 4.57599759e-01 -1.91797942e-01 -8.65757540e-02
9.89649475e-01 -3.52302521e-01 -6.13333583e-02 -2.45918617e-01
-6.72988296e-01 -1.58701733e-01 4.97439057e-01 5.78665614e-01
6.52206600e-01 -1.46843147e+00 -5.82831860e-01 6.09862149e-01
4.72696930e-01 8.73480830e-03 -2.45324284e-01 6.64896965e-01
-5.36983430e-01 4.79865611e-01 -7.90865302e-01 -9.08408880e-01
-1.18668449e+00 7.77656198e-01 3.71865183e-01 -2.23724902e-01
-5.40807307e-01 7.43461132e-01 7.99547061e-02 -5.60698807e-01
1.46360099e-01 -2.75815338e-01 -3.65031809e-01 1.00836977e-01
3.60824913e-01 -2.23395541e-01 4.98665810e-01 -9.25317287e-01
-2.67152280e-01 7.99828529e-01 -1.97959423e-01 4.19973791e-01
1.73094082e+00 3.94804388e-01 2.52999693e-01 1.26955226e-01
1.74258804e+00 -3.25256348e-01 -1.25725329e+00 -5.71804821e-01
4.31026280e-01 -4.62865889e-01 -4.58918624e-02 -4.19734061e-01
-1.26724827e+00 9.90946651e-01 7.45623171e-01 2.51218647e-01
9.93606508e-01 -1.05671644e-01 4.68268633e-01 8.40990126e-01
4.92706984e-01 -9.60411251e-01 2.32694760e-01 7.04546094e-01
8.51664662e-01 -1.57361138e+00 -9.11560655e-02 -2.78020263e-01
-1.04590341e-01 1.65628958e+00 3.91468883e-01 -6.73660219e-01
7.38824189e-01 2.95132911e-03 1.43651649e-01 -2.94021666e-01
-2.93674201e-01 -6.65420771e-01 6.64010465e-01 8.41999352e-01
3.75903755e-01 -2.69225627e-01 -2.17911135e-02 -3.11415214e-02
-2.98681200e-01 -2.36848667e-01 1.54405206e-01 1.02623117e+00
-5.14928520e-01 -1.55921972e+00 -2.54332989e-01 6.71456516e-01
-3.31396282e-01 7.48441517e-02 -4.80405092e-01 1.27568018e+00
9.60398763e-02 4.96540993e-01 6.17718287e-02 -4.01724547e-01
3.44398588e-01 5.58574498e-02 8.01143527e-01 -6.65831029e-01
-7.70501614e-01 -2.25232691e-01 -1.35948241e-01 -3.85723859e-01
-6.55763209e-01 -6.23759389e-01 -8.63518059e-01 2.25536942e-01
-2.68172115e-01 1.92668900e-01 5.99781215e-01 1.10441673e+00
-7.91201442e-02 2.92288095e-01 9.12579656e-01 -1.24961019e+00
-7.55202115e-01 -7.91223288e-01 -5.62347889e-01 8.20044816e-01
4.72876400e-01 -1.10536230e+00 -3.21077973e-01 4.36134748e-02] | [9.503396987915039, 2.2626307010650635] |
b259d9ab-43fa-4e00-af8b-3768ed37bd0a | lamd-latent-motion-diffusion-for-video | 2304.11603 | null | https://arxiv.org/abs/2304.11603v1 | https://arxiv.org/pdf/2304.11603v1.pdf | LaMD: Latent Motion Diffusion for Video Generation | Generating coherent and natural movement is the key challenge in video generation. This research proposes to condense video generation into a problem of motion generation, to improve the expressiveness of motion and make video generation more manageable. This can be achieved by breaking down the video generation process into latent motion generation and video reconstruction. We present a latent motion diffusion (LaMD) framework, which consists of a motion-decomposed video autoencoder and a diffusion-based motion generator, to implement this idea. Through careful design, the motion-decomposed video autoencoder can compress patterns in movement into a concise latent motion representation. Meanwhile, the diffusion-based motion generator is able to efficiently generate realistic motion on a continuous latent space under multi-modal conditions, at a cost that is similar to that of image diffusion models. Results show that LaMD generates high-quality videos with a wide range of motions, from stochastic dynamics to highly controllable movements. It achieves new state-of-the-art performance on benchmark datasets, including BAIR, Landscape and CATER-GENs, for Image-to-Video (I2V) and Text-Image-to-Video (TI2V) generation. The source code of LaMD will be made available soon. | ['Chong Luo', 'Zhenzhong Chen', 'Yaosi Hu'] | 2023-04-23 | null | null | null | null | ['video-generation', 'video-reconstruction'] | ['computer-vision', 'computer-vision'] | [ 1.43319413e-01 -2.16966689e-01 -1.95648432e-01 2.23919779e-01
-5.53579152e-01 -4.68420208e-01 8.36341381e-01 -8.63252938e-01
1.23030499e-01 6.78536534e-01 6.73614085e-01 -6.46883175e-02
1.16282448e-01 -1.04702532e+00 -9.11778390e-01 -1.02524698e+00
-2.19756756e-02 6.38640746e-02 1.49575800e-01 -1.61663339e-01
-7.28532970e-02 2.15310857e-01 -1.62746012e+00 5.83649158e-01
6.21416748e-01 6.95394397e-01 6.12501383e-01 1.09996033e+00
1.33785278e-01 1.32227778e+00 -5.40467858e-01 -1.49528682e-01
1.85101256e-01 -9.48540330e-01 -6.86597109e-01 3.52812737e-01
1.78368777e-01 -8.86026144e-01 -7.17954934e-01 7.57937372e-01
4.89216000e-01 1.64282948e-01 8.38492274e-01 -1.32906187e+00
-1.13879550e+00 6.50294363e-01 -4.44194138e-01 2.68903263e-02
5.38809896e-01 5.40333509e-01 7.75628626e-01 -7.07917273e-01
1.22688437e+00 1.45543218e+00 2.47513786e-01 1.07510006e+00
-1.08773696e+00 -5.14699459e-01 3.56031694e-02 8.78415555e-02
-1.15843952e+00 -4.39017415e-01 6.23011649e-01 -7.38773346e-01
8.19857895e-01 3.30620170e-01 8.59983325e-01 1.67696416e+00
2.80712098e-01 1.03368437e+00 3.75284404e-01 -2.58888721e-01
1.84972212e-01 -4.41557497e-01 -6.74258888e-01 5.15737116e-01
7.66969845e-02 1.23834051e-01 -5.76590896e-01 1.02224268e-01
1.23163795e+00 -7.34299868e-02 -6.59016371e-01 -6.66200966e-02
-1.55744791e+00 8.32082212e-01 1.13033719e-01 2.46846244e-01
-8.14735532e-01 5.23151457e-01 -2.16011517e-02 1.01110376e-01
4.35529798e-01 1.22011736e-01 1.90028638e-01 -4.44471985e-01
-1.32835126e+00 6.85345888e-01 5.64627826e-01 1.16169822e+00
2.79154569e-01 7.56412446e-01 -6.57923281e-01 3.90702933e-01
3.43869299e-01 8.07701051e-01 8.63147438e-01 -1.45225072e+00
4.35305297e-01 8.46971348e-02 2.89994687e-01 -1.27565658e+00
2.22333834e-01 1.68263242e-01 -1.25211084e+00 -2.46163141e-02
-9.58388522e-02 -3.35577905e-01 -8.99386585e-01 1.71676147e+00
2.44512618e-01 3.25776964e-01 2.29535848e-01 9.90385532e-01
9.29312348e-01 1.53958571e+00 -9.61564258e-02 -5.29749572e-01
8.04759562e-01 -1.40039539e+00 -1.06810355e+00 2.01011330e-01
3.67595732e-01 -6.42215312e-01 8.33349407e-01 2.18532056e-01
-1.58727956e+00 -7.99278021e-01 -9.87554252e-01 4.34115864e-02
3.29621255e-01 -7.84454681e-03 3.93948913e-01 2.13073865e-01
-1.33132279e+00 4.99598920e-01 -8.90940130e-01 -1.25218645e-01
1.56987831e-01 4.25341092e-02 -4.49932158e-01 -1.12261593e-01
-1.23461747e+00 1.80292934e-01 4.64318752e-01 4.33077365e-02
-1.37022853e+00 -5.36141396e-01 -9.80980754e-01 -6.36232570e-02
5.41492589e-02 -1.41690338e+00 1.21782994e+00 -1.05018735e+00
-1.71502030e+00 2.62596518e-01 -2.72689462e-01 -5.08858740e-01
8.14236104e-01 -9.10567492e-02 -3.25668871e-01 5.20741522e-01
1.49420038e-01 1.27173412e+00 1.25449634e+00 -1.17040336e+00
-5.72017252e-01 3.42551202e-01 -1.65412903e-01 2.40647465e-01
-3.96068275e-01 -2.67614335e-01 -7.72366583e-01 -1.26801455e+00
-3.40715468e-01 -1.06880438e+00 -2.35637560e-01 -2.77143806e-01
-3.26175392e-01 2.27646139e-02 1.12282395e+00 -7.71375000e-01
1.58079576e+00 -1.96642733e+00 8.19863915e-01 -5.01951456e-01
2.91009367e-01 2.00685009e-01 -4.41344798e-01 5.34620404e-01
-1.29310135e-02 2.81515986e-01 -2.19233945e-01 -3.85715783e-01
-1.01128578e-01 1.00168481e-01 -5.84987044e-01 -3.69086536e-03
1.46593735e-01 1.26697421e+00 -9.19753015e-01 -2.81071156e-01
1.62774861e-01 4.75985974e-01 -9.50067937e-01 4.27355289e-01
-4.36622739e-01 4.87841219e-01 -4.54787850e-01 4.59527731e-01
4.94569957e-01 -3.58615398e-01 2.52811685e-02 -1.21991113e-01
-5.28784692e-02 -3.70718688e-01 -1.02157032e+00 1.67423844e+00
-3.25161248e-01 7.58459210e-01 -1.53169215e-01 -3.57292891e-01
6.19412839e-01 4.59052950e-01 7.11334705e-01 -4.75885183e-01
-1.18179657e-01 -1.05749648e-02 -3.88033360e-01 -8.09823632e-01
8.80906105e-01 -9.15638055e-04 8.12155530e-02 4.78233516e-01
3.71089429e-02 -1.28972074e-02 4.30113345e-01 3.22942585e-01
9.79825079e-01 3.37614447e-01 -1.94874763e-01 1.33093715e-01
3.01723123e-01 -1.21729232e-01 4.63882655e-01 5.31829000e-01
1.63220510e-01 9.59096372e-01 2.34840110e-01 -4.10342664e-01
-1.38720489e+00 -1.15220952e+00 5.94907701e-01 4.62220401e-01
6.29448295e-02 -4.25653309e-01 -1.07195950e+00 -2.88084269e-01
-3.43398303e-01 6.20896876e-01 -5.24305880e-01 -3.05067658e-01
-7.38894820e-01 -6.10537767e-01 4.64318305e-01 6.10935569e-01
6.35896862e-01 -1.34726405e+00 -4.55299705e-01 3.95657420e-01
-7.74160862e-01 -1.13005435e+00 -8.14681530e-01 -7.79468596e-01
-7.47535348e-01 -6.24135196e-01 -1.61962605e+00 -8.06738019e-01
4.84741062e-01 3.93000573e-01 9.50661182e-01 -5.97165562e-02
-1.63496599e-01 3.61980498e-01 -6.12248778e-01 2.13453874e-01
-8.84301484e-01 -1.84522584e-01 8.67687315e-02 2.70846665e-01
-2.39097640e-01 -5.55022478e-01 -9.10085618e-01 1.39025182e-01
-1.52354670e+00 6.28308296e-01 4.90749240e-01 9.78522718e-01
5.56314051e-01 2.17297450e-01 2.49410778e-01 -3.15453440e-01
7.20023811e-01 -7.52388716e-01 -3.81378263e-01 2.74206195e-02
-7.88141787e-02 4.14693309e-03 5.84193647e-01 -7.26444781e-01
-1.16314781e+00 -4.28075194e-02 -2.42369145e-01 -9.61792886e-01
4.23404947e-02 2.78812528e-01 -2.09921762e-01 4.22505468e-01
4.26203161e-01 8.30066144e-01 -1.40249431e-01 -1.68868050e-01
7.58381546e-01 6.10381365e-01 5.91071188e-01 -3.31447512e-01
8.04932714e-01 6.39901757e-01 -1.64824560e-01 -1.04934061e+00
-9.65221822e-02 5.00204675e-02 -3.53054851e-01 -4.27645236e-01
1.48142540e+00 -1.12536848e+00 -3.85147810e-01 9.43960905e-01
-1.30185068e+00 -8.16254616e-01 -3.74514878e-01 4.72769976e-01
-9.20390725e-01 5.16189873e-01 -1.07219386e+00 -3.98231536e-01
-4.05298859e-01 -1.47908723e+00 1.19739234e+00 9.35781971e-02
-2.03151122e-01 -9.77896392e-01 1.52642906e-01 2.65729815e-01
4.54888970e-01 5.68129778e-01 7.02470303e-01 3.09776545e-01
-1.22644377e+00 3.84449214e-02 2.35415921e-01 3.28944325e-01
1.61956578e-01 3.38297307e-01 -4.53368485e-01 -3.94013971e-01
-8.71943496e-03 -2.01262280e-01 1.03623652e+00 9.04894173e-01
8.89525592e-01 -7.62785554e-01 -2.02785641e-01 9.47731853e-01
1.22213697e+00 3.98616612e-01 9.89273071e-01 2.30363786e-01
9.25110281e-01 4.14701462e-01 3.56768310e-01 6.26818240e-01
4.86673027e-01 8.45069110e-01 2.39805102e-01 1.63636550e-01
-4.99350697e-01 -6.09477460e-01 7.65668988e-01 1.28032970e+00
-2.66800165e-01 -9.19079423e-01 -6.51005447e-01 6.08908594e-01
-1.93995714e+00 -1.55141997e+00 -2.33395934e-01 1.64902985e+00
5.37766814e-01 -2.10286379e-01 2.69963950e-01 9.14249122e-02
7.30632663e-01 4.75567907e-01 -3.12924206e-01 -1.05361603e-01
-3.31120253e-01 -4.11332846e-01 6.52933642e-02 5.37337840e-01
-8.86731446e-01 1.13218927e+00 6.57041025e+00 1.07927072e+00
-1.13818860e+00 -7.03632226e-03 7.09478080e-01 -1.38275608e-01
-7.60664582e-01 -1.96635932e-01 -7.40911186e-01 8.26350987e-01
9.07628238e-01 -2.74887621e-01 4.81117129e-01 7.19311059e-01
6.63907409e-01 2.34195188e-01 -8.04345608e-01 1.15146458e+00
1.20719604e-01 -2.02452588e+00 9.28681612e-01 1.22214660e-01
1.18455362e+00 -4.34956968e-01 3.15854698e-01 1.86605215e-01
1.94704279e-01 -1.04916918e+00 1.13960862e+00 6.75926805e-01
9.22820568e-01 -6.33814275e-01 3.29522192e-01 4.07908171e-01
-1.38055551e+00 -1.62424892e-01 -2.10546508e-01 1.35110676e-01
7.59735942e-01 3.23374689e-01 -1.66462809e-01 7.06118643e-01
6.33125663e-01 9.93885756e-01 -1.48401901e-01 6.47712767e-01
-7.63177425e-02 5.59484541e-01 1.75961494e-01 1.68890461e-01
3.74349505e-01 -1.96038693e-01 8.39084327e-01 1.10199583e+00
8.95560980e-01 7.53341615e-02 6.07906319e-02 1.12002313e+00
2.20623755e-04 -1.71969756e-01 -7.98101306e-01 -3.70769560e-01
3.98627996e-01 9.61192608e-01 -5.49825013e-01 -4.88144755e-01
-5.11377938e-02 1.45454192e+00 -2.83686012e-01 6.07033849e-01
-1.09619319e+00 -9.41613764e-02 4.17641759e-01 2.61830539e-01
5.92262149e-01 -3.65949959e-01 6.28057599e-01 -1.43664718e+00
5.07183955e-04 -1.01947355e+00 4.98263314e-02 -1.18701696e+00
-8.98857713e-01 9.39633071e-01 1.98208362e-01 -1.46318257e+00
-8.59151304e-01 -2.17930779e-01 -5.01730204e-01 6.20281518e-01
-1.07272017e+00 -1.34108162e+00 -5.65793335e-01 6.91027761e-01
1.04494560e+00 -4.28878248e-01 5.16394854e-01 4.03381288e-01
-4.44000602e-01 4.18579698e-01 1.05743065e-01 5.54362014e-02
3.12618852e-01 -8.72642875e-01 6.71168268e-01 1.10022891e+00
1.55987427e-01 3.46169204e-01 5.86969137e-01 -8.27190459e-01
-1.57939327e+00 -1.46755135e+00 4.87347752e-01 -2.88221210e-01
3.36706847e-01 -4.09930833e-02 -5.28299987e-01 5.70240438e-01
5.01524746e-01 -2.95544565e-01 3.22977632e-01 -1.03718436e+00
2.43710414e-01 3.48354161e-01 -6.49694204e-01 1.01738894e+00
1.20782840e+00 -2.21381903e-01 -2.73843676e-01 3.46063152e-02
1.24259853e+00 -4.61955786e-01 -7.17586279e-01 3.23418021e-01
4.20355380e-01 -9.27120626e-01 9.82270122e-01 -3.47984225e-01
1.16149068e+00 -4.66735214e-01 -2.07975507e-01 -1.17706978e+00
-5.73094666e-01 -1.22212827e+00 -6.96574628e-01 1.30037677e+00
2.04811688e-03 8.45077783e-02 7.51887977e-01 1.00802898e-01
-1.45000014e-02 -6.89320207e-01 -5.04866421e-01 -8.01747859e-01
-6.08452000e-02 -2.79579848e-01 6.14063919e-01 8.95687044e-01
-6.49127841e-01 1.21780425e-01 -1.03012717e+00 -9.97922570e-02
4.91126984e-01 1.20794110e-03 9.50623751e-01 -3.19037735e-01
-6.19240284e-01 -5.00045836e-01 -3.16576540e-01 -1.66657269e+00
1.07171848e-01 -7.21175313e-01 -5.68431132e-02 -1.71259320e+00
1.91021726e-01 3.02234709e-01 2.73542732e-01 -1.42077049e-02
-1.56315908e-01 2.75937676e-01 5.43774188e-01 5.33819616e-01
-4.04060870e-01 1.00402439e+00 1.82752943e+00 -2.11864799e-01
-3.51270497e-01 -2.49932304e-01 -4.05777812e-01 5.88095546e-01
4.17499721e-01 -2.91700333e-01 -7.57039905e-01 -6.60142481e-01
4.64202501e-02 6.47091806e-01 2.76628643e-01 -1.08964062e+00
1.43699065e-01 -3.78506362e-01 3.69418174e-01 -6.81481361e-01
3.90544176e-01 -2.83669323e-01 7.00766325e-01 5.81306934e-01
-2.99272090e-01 4.58343327e-01 2.96119917e-02 7.93067336e-01
-4.67485219e-01 1.23401865e-01 5.36018848e-01 -2.12995499e-01
-8.40445399e-01 6.67466283e-01 -7.30361581e-01 -2.36788094e-02
1.18686199e+00 -3.54105443e-01 -1.82181224e-01 -9.80827928e-01
-6.15048349e-01 3.68437842e-02 5.68243682e-01 8.29411328e-01
9.51392651e-01 -1.84842920e+00 -9.21513915e-01 7.43066147e-02
-3.52376193e-01 2.65742987e-02 7.51234770e-01 5.24422884e-01
-1.04433715e+00 3.85768801e-01 -3.01516026e-01 -6.83985293e-01
-1.01710927e+00 6.38511121e-01 4.45198826e-02 -2.14640245e-01
-8.09320569e-01 8.70365322e-01 5.47391355e-01 2.35649437e-01
-2.40219627e-02 -2.34379768e-01 -1.00972563e-01 -1.22079991e-01
8.79220426e-01 3.47766817e-01 -7.02336252e-01 -8.56808126e-01
1.45690009e-01 5.27880967e-01 3.14841807e-01 -3.01298171e-01
1.24558306e+00 -1.96713567e-01 1.09276876e-01 1.75819099e-01
1.09225178e+00 -3.38623254e-03 -1.61963642e+00 4.34213191e-01
-4.41889375e-01 -4.44702417e-01 -6.96290061e-02 -1.45691007e-01
-1.28223670e+00 8.41384888e-01 3.36511642e-01 1.02418959e-01
1.21198893e+00 -3.10932785e-01 1.35459614e+00 -4.52076979e-02
2.62014359e-01 -7.44273663e-01 6.53473377e-01 2.79882073e-01
1.06494176e+00 -7.87336051e-01 -2.19080940e-01 -1.15749180e-01
-9.40871537e-01 1.05692065e+00 5.06294847e-01 -3.25927958e-02
3.61585170e-01 2.10327178e-01 -2.11148113e-01 1.86287045e-01
-1.13166571e+00 1.61552459e-01 3.28191906e-01 6.94134772e-01
2.46192694e-01 -1.08296886e-01 -7.13794008e-02 4.34100568e-01
-1.22161321e-01 2.59862214e-01 8.00758123e-01 6.82499588e-01
-1.81390703e-01 -1.03602326e+00 -2.89277881e-01 -1.08136617e-01
-3.19734901e-01 -2.52903793e-02 2.10413020e-02 7.03307211e-01
6.37055635e-02 8.53643298e-01 1.21032506e-01 -5.93879700e-01
-8.67125914e-02 -3.75447065e-01 4.25139576e-01 -2.62123764e-01
3.18057910e-02 4.20030534e-01 -1.54416695e-01 -6.96637094e-01
-6.54212952e-01 -3.25855255e-01 -8.68691683e-01 -5.88092864e-01
1.46791022e-02 5.00917956e-02 2.13675305e-01 4.80758637e-01
5.20507395e-01 7.51916289e-01 5.92349589e-01 -1.15906346e+00
-3.21836680e-01 -7.33396649e-01 -4.42507297e-01 6.89852417e-01
3.32982928e-01 -3.42669874e-01 -2.41495252e-01 7.86354780e-01] | [10.837860107421875, -0.5500283241271973] |
be5aa763-1142-4a52-80af-299e1d80f2f0 | bargainnet-background-guided-domain | 2009.09169 | null | https://arxiv.org/abs/2009.09169v2 | https://arxiv.org/pdf/2009.09169v2.pdf | BargainNet: Background-Guided Domain Translation for Image Harmonization | Image composition is a fundamental operation in image editing field. However, unharmonious foreground and background downgrade the quality of composite image. Image harmonization, which adjusts the foreground to improve the consistency, is an essential yet challenging task. Previous deep learning based methods mainly focus on directly learning the mapping from composite image to real image, while ignoring the crucial guidance role that background plays. In this work, with the assumption that the foreground needs to be translated to the same domain as background, we formulate image harmonization task as background-guided domain translation. Therefore, we propose an image harmonization network with a novel domain code extractor and well-tailored triplet losses, which could capture the background domain information to guide the foreground harmonization. Extensive experiments on the existing image harmonization benchmark demonstrate the effectiveness of our proposed method. Code is available at https://github.com/bcmi/BargainNet. | ['Wenyan Cong', 'Li Niu', 'Jing Liang', 'Jianfu Zhang', 'Liqing Zhang'] | 2020-09-19 | null | null | null | null | ['image-harmonization'] | ['computer-vision'] | [ 4.34860200e-01 -2.04779446e-01 -8.34643170e-02 -3.07120472e-01
-5.02919495e-01 -4.78659570e-01 4.69262570e-01 -2.93641120e-01
-2.04204947e-01 6.31326199e-01 2.84993947e-01 -1.29702881e-01
1.27660543e-01 -6.85329199e-01 -8.38059664e-01 -1.01390183e+00
8.99135828e-01 2.11765748e-02 1.80089310e-01 -1.93534583e-01
1.16379410e-01 9.30844173e-02 -1.09911168e+00 2.27297187e-01
1.29460132e+00 8.63720298e-01 3.40080470e-01 3.00207585e-01
-2.11723030e-01 6.78769469e-01 -5.89677036e-01 -5.76429188e-01
4.50602442e-01 -7.12631226e-01 -4.22307968e-01 4.29943591e-01
5.20758688e-01 -2.57456750e-01 -3.71552795e-01 1.63191330e+00
6.60246491e-01 6.71071336e-02 3.64590108e-01 -1.41107690e+00
-9.86514032e-01 4.86148179e-01 -8.58413637e-01 1.04517817e-01
-6.54128864e-02 3.76692265e-01 8.13035607e-01 -7.25983143e-01
4.52740759e-01 1.08645284e+00 2.73582131e-01 4.77181971e-01
-1.20693052e+00 -9.80581045e-01 2.72442400e-01 3.16687852e-01
-1.31793666e+00 -5.11904895e-01 1.05269027e+00 -3.89658421e-01
4.00125608e-02 1.76028132e-01 5.83775699e-01 8.26231301e-01
2.77328845e-02 8.27974617e-01 1.06856656e+00 -2.57358104e-01
-2.33310714e-01 1.47762567e-01 -3.62717271e-01 5.34284115e-01
3.50633055e-01 -1.26753896e-01 -3.19408447e-01 3.17911327e-01
8.30427408e-01 4.76063713e-02 -6.86420977e-01 -6.56073034e-01
-1.24727559e+00 5.83067060e-01 4.78260875e-01 1.14340320e-01
-8.16876441e-02 4.28733155e-02 4.19648558e-01 1.11759789e-01
4.35244054e-01 1.04222029e-01 -2.47149959e-01 2.07994163e-01
-7.62898505e-01 1.52405173e-01 3.70611280e-01 1.11463261e+00
8.27072084e-01 1.49792865e-01 -4.92859215e-01 9.69971359e-01
1.62045822e-01 4.93546337e-01 2.77245492e-01 -1.00754178e+00
6.08410716e-01 7.22498894e-01 2.19426513e-01 -1.14785945e+00
4.87461910e-02 -6.34168446e-01 -1.12818563e+00 2.10907593e-01
2.83890128e-01 -1.75198913e-01 -7.75578558e-01 1.82882929e+00
5.66079736e-01 3.99079084e-01 -1.64867610e-01 1.09931064e+00
9.25034523e-01 7.20824718e-01 -8.31724331e-02 -2.72063285e-01
1.23963225e+00 -1.27344143e+00 -8.02753925e-01 -1.04515187e-01
-4.25740853e-02 -1.12813151e+00 1.21022248e+00 3.13558400e-01
-1.23182213e+00 -7.59618342e-01 -1.06041658e+00 -3.20612431e-01
1.13167122e-01 2.45483518e-01 4.83703166e-01 5.28109789e-01
-7.96915233e-01 1.31909311e-01 -4.28353518e-01 -1.40529558e-01
6.36458397e-01 1.97454214e-01 -1.53331250e-01 -1.15248986e-01
-1.16321540e+00 6.28052890e-01 5.87499201e-01 3.17374766e-01
-7.83736229e-01 -7.48386860e-01 -7.14983523e-01 -5.95659837e-02
5.32638788e-01 -8.02432835e-01 1.04174316e+00 -1.42587364e+00
-1.48516750e+00 1.09974563e+00 -7.96384439e-02 -1.74032167e-01
8.41030300e-01 -5.73146343e-02 -4.52475786e-01 -3.77307460e-02
1.95876315e-01 5.21901131e-01 1.06177676e+00 -1.50709248e+00
-9.47446465e-01 -9.27077383e-02 1.01649165e-01 4.39776212e-01
-2.68612057e-01 -9.75420922e-02 -1.13641942e+00 -9.89950657e-01
-2.10351452e-01 -7.82089412e-01 7.70895183e-02 2.12904871e-01
-4.22347248e-01 1.10678539e-01 7.03415215e-01 -9.67262089e-01
1.27299905e+00 -2.24342942e+00 3.63547802e-01 6.00751005e-02
3.27179044e-01 3.68505567e-01 -1.89841256e-01 -1.05974339e-01
-1.32549182e-01 -1.16064973e-01 -4.96451855e-01 -6.75223395e-02
-3.48359201e-04 -2.13463400e-02 -2.71312654e-01 4.92706150e-01
1.07899591e-01 9.06813920e-01 -9.02978957e-01 -7.37834811e-01
2.01574087e-01 5.56716979e-01 -5.93066037e-01 2.83858836e-01
-2.36682311e-01 8.00145149e-01 -3.33630830e-01 5.28333485e-01
1.13635755e+00 -2.91694790e-01 2.72279419e-02 -7.64371037e-01
4.51217741e-02 -2.33975291e-01 -1.12284219e+00 1.82856989e+00
-3.73541296e-01 4.99883205e-01 3.57395262e-01 -1.02381682e+00
8.64634037e-01 -3.27932154e-04 5.53909540e-01 -1.01288724e+00
2.08748832e-01 5.88635020e-02 6.03791280e-03 -3.19704324e-01
2.62526304e-01 -2.66760197e-02 1.39495060e-01 3.03249449e-01
-3.23685676e-01 -2.23527133e-01 2.82492377e-02 3.19706425e-02
4.46755677e-01 3.72064859e-01 1.86878502e-01 -1.70014799e-01
8.58206928e-01 -7.50983283e-02 1.06629050e+00 2.96157688e-01
-2.77473390e-01 7.76108682e-01 4.58009303e-01 -2.35947862e-01
-1.11975527e+00 -1.02506447e+00 7.36474916e-02 8.64379704e-01
8.55886459e-01 -1.42705943e-02 -9.92620647e-01 -3.22222024e-01
-2.70394444e-01 4.62358505e-01 -4.57424343e-01 -3.72475803e-01
-6.51743233e-01 -8.61188591e-01 3.24726522e-01 1.46207541e-01
1.03897035e+00 -9.15142536e-01 -1.58749998e-01 6.86794445e-02
-6.41485035e-01 -1.05507874e+00 -1.15735602e+00 -2.61790067e-01
-4.17938381e-01 -1.07221425e+00 -9.59876001e-01 -1.00596416e+00
7.49788821e-01 5.21719098e-01 1.09832060e+00 3.44877124e-01
-2.15394795e-01 6.36872500e-02 -1.96068496e-01 -2.51702458e-01
-3.85050118e-01 1.95199400e-02 -2.57810652e-01 4.66174185e-01
1.11516222e-01 -5.59328258e-01 -9.80880797e-01 4.17784303e-01
-1.15173936e+00 5.84004045e-01 5.65281034e-01 8.81515443e-01
8.41066062e-01 4.50762987e-01 2.76120991e-01 -9.49523211e-01
5.75578094e-01 -1.86063781e-01 -8.15867782e-01 4.79452640e-01
-4.98022079e-01 -2.11995080e-01 5.33754289e-01 -4.49790269e-01
-1.48188937e+00 1.54871687e-01 8.88513848e-02 -4.92405087e-01
1.24013789e-01 7.09026456e-02 -8.66580367e-01 -1.69565558e-01
3.01292539e-01 4.45833564e-01 -9.26256701e-02 -2.38020077e-01
4.72606272e-01 3.39442044e-01 8.41140926e-01 -6.88912511e-01
1.05691016e+00 6.62569582e-01 -2.79457211e-01 -2.54578799e-01
-7.98768222e-01 -2.19054192e-01 -5.13672650e-01 -1.48772210e-01
9.82468724e-01 -1.06917799e+00 -5.39303601e-01 6.51707232e-01
-1.18279350e+00 -5.63368082e-01 4.61980281e-03 4.10539880e-02
-4.16708738e-01 4.84683096e-01 -3.07094961e-01 -2.47728527e-01
-3.36989701e-01 -1.16443706e+00 9.69670475e-01 5.89699328e-01
3.72717589e-01 -7.99326420e-01 -7.34671876e-02 4.97307956e-01
3.23145777e-01 3.35662335e-01 8.23303223e-01 5.87651087e-03
-7.58503437e-01 2.67174721e-01 -7.40502119e-01 4.60158825e-01
4.59030688e-01 1.07936516e-01 -7.45219409e-01 -2.01578066e-01
1.32064167e-02 5.83115332e-02 8.72325420e-01 3.59058887e-01
1.27751076e+00 -2.94602364e-01 -5.97983375e-02 1.07038093e+00
1.46700191e+00 3.34764630e-01 7.47534394e-01 4.30541724e-01
1.03424108e+00 5.39760947e-01 7.86323965e-01 4.33449835e-01
4.79236335e-01 8.59666765e-01 3.78252953e-01 -5.65228641e-01
-5.13163507e-01 -2.57564008e-01 1.24399379e-01 7.62217224e-01
1.17033660e-01 -1.62364110e-01 -7.38986552e-01 3.65463853e-01
-2.02073169e+00 -9.16170776e-01 -5.46080694e-02 1.85918057e+00
1.26795685e+00 3.83404531e-02 -5.90479523e-02 -2.96584964e-01
1.28647065e+00 1.45375162e-01 -8.62797797e-01 2.29362413e-01
-4.05580699e-01 -8.31115022e-02 4.06710237e-01 5.61273754e-01
-1.24580622e+00 1.03609109e+00 4.54745007e+00 1.15222490e+00
-1.30876243e+00 3.27347577e-01 8.72982442e-01 -1.18464656e-01
-5.02355933e-01 -7.42984265e-02 -5.79395354e-01 8.89674425e-01
8.37699044e-03 -4.91248518e-01 7.95372844e-01 5.19019663e-01
4.19392020e-01 8.23759139e-02 -6.79422021e-01 1.23978579e+00
1.56586602e-01 -1.30341494e+00 1.79814577e-01 -2.47214094e-01
9.95350003e-01 -4.05800998e-01 2.25599259e-01 -4.25155321e-03
2.64619589e-01 -7.46590674e-01 9.06967878e-01 5.38044453e-01
8.61483872e-01 -7.30210423e-01 4.60086644e-01 -5.03339954e-02
-1.40333772e+00 7.67945349e-02 -4.44349825e-01 3.02156687e-01
1.59352794e-01 6.71574771e-01 -2.66086221e-01 6.28907681e-01
6.96582556e-01 9.30240393e-01 -6.48367524e-01 1.04063499e+00
-3.95732850e-01 3.13978732e-01 9.55767930e-02 6.26797080e-01
-1.78781748e-01 -6.52605295e-01 4.38716382e-01 1.06301200e+00
2.31916636e-01 4.56391051e-02 1.42659605e-01 1.12673557e+00
-2.89534360e-01 1.02232113e-01 -2.19304651e-01 1.38970897e-01
4.85091120e-01 1.43973100e+00 -7.21147418e-01 -1.97239876e-01
-4.57375228e-01 1.31870806e+00 -2.56866626e-02 5.46972990e-01
-1.34209335e+00 -4.25354570e-01 6.61235869e-01 -7.39350319e-02
2.81121880e-01 1.34771019e-01 -3.96527886e-01 -1.44450462e+00
2.30542496e-01 -1.26236296e+00 1.99448198e-01 -9.30969238e-01
-1.26094282e+00 4.14610028e-01 -2.23079294e-01 -1.54692221e+00
5.15215635e-01 -2.97910243e-01 -8.36458385e-01 8.47779036e-01
-1.69014561e+00 -1.46500504e+00 -8.40286553e-01 8.76570761e-01
3.54024976e-01 -9.72733647e-02 7.21249878e-02 7.29057431e-01
-8.94181132e-01 6.72497153e-01 8.63188803e-02 2.04779789e-01
1.13471293e+00 -9.64615941e-01 1.47492692e-01 1.13181853e+00
-1.79618821e-01 5.09826839e-01 7.14638829e-01 -5.31888723e-01
-1.08323240e+00 -1.43858683e+00 3.07500839e-01 -1.73174575e-01
5.69803715e-01 -2.68276483e-01 -1.01716948e+00 3.77932906e-01
4.22723591e-01 -3.70105952e-02 3.62231940e-01 -5.97601295e-01
-1.98230237e-01 -5.95013320e-01 -8.26580226e-01 8.39786172e-01
1.12084281e+00 -4.45941567e-01 -2.61966914e-01 3.69198591e-01
1.01763248e+00 -6.70629621e-01 -7.32634664e-01 3.75258923e-01
3.31668943e-01 -9.00874555e-01 1.13538778e+00 -1.70758367e-01
5.99748373e-01 -8.46114814e-01 8.61363262e-02 -1.16039050e+00
-3.19600224e-01 -6.62139356e-01 2.80653208e-01 1.61504495e+00
8.74760747e-02 -5.20745039e-01 5.85649192e-01 5.21052897e-01
-1.05969526e-01 -3.67541105e-01 -5.99468589e-01 -4.03644770e-01
6.03723526e-03 1.32505298e-01 8.71808112e-01 1.16601562e+00
-6.62821651e-01 2.10361406e-01 -5.92734873e-01 2.31263131e-01
7.01577425e-01 4.00816649e-01 1.13187528e+00 -7.71384895e-01
-3.26037139e-01 -6.55153453e-01 -2.09822040e-02 -8.70414853e-01
2.58930683e-01 -7.88246751e-01 1.01312891e-01 -1.50508308e+00
4.70694095e-01 -3.79125446e-01 -4.20971811e-01 3.53625506e-01
-5.41590750e-01 3.73684198e-01 2.50627011e-01 2.22830310e-01
-6.70669734e-01 7.52329528e-01 1.68337953e+00 -4.53207672e-01
-2.35726591e-02 -2.22036391e-01 -1.03854442e+00 6.06581986e-01
9.29628789e-01 -3.02898705e-01 -6.13718033e-01 -7.43390203e-01
1.97677612e-01 -1.21024460e-01 4.15288359e-01 -6.65726006e-01
3.03918451e-01 -4.85691279e-01 3.83047849e-01 -2.65412092e-01
-2.35591363e-02 -8.54803443e-01 4.54225332e-01 4.02126223e-01
-2.09103331e-01 1.33422390e-02 1.95018977e-01 6.10346615e-01
-3.92978847e-01 -3.93314399e-02 1.09877717e+00 -9.88160521e-02
-8.96679580e-01 6.66913927e-01 3.03298026e-01 3.27847153e-01
1.05012846e+00 -2.04166055e-01 -4.19875741e-01 -3.11853558e-01
-3.22837800e-01 3.14387530e-01 8.68677378e-01 4.24326003e-01
4.93132561e-01 -1.54278672e+00 -8.36753070e-01 7.86969960e-02
5.29011823e-02 1.66124836e-01 5.05566657e-01 8.96910548e-01
-6.91735625e-01 -9.16386545e-02 -4.48801190e-01 -4.27153140e-01
-1.24239504e+00 5.87187648e-01 6.75869882e-01 -4.77107838e-02
-5.28312743e-01 8.45779479e-01 8.55316460e-01 -2.98936963e-01
2.72776961e-01 -8.64130259e-02 -1.03368126e-01 1.21086286e-02
4.22377884e-01 1.59636751e-01 -1.31615430e-01 -6.62420928e-01
-2.38934234e-01 6.96254849e-01 -4.11133654e-02 1.85892209e-01
1.04072535e+00 -5.71947873e-01 -5.28622925e-01 1.84643213e-02
1.05971634e+00 6.74503967e-02 -1.50096977e+00 -4.54775125e-01
-3.54296297e-01 -6.91522360e-01 -3.68514634e-03 -6.30271852e-01
-1.55379879e+00 8.45459521e-01 7.05068946e-01 -2.04734311e-01
1.62121046e+00 -3.58559757e-01 8.99382353e-01 -6.20635562e-02
-8.28498232e-05 -1.15186226e+00 2.94986933e-01 1.06714413e-01
9.66923475e-01 -1.42990673e+00 1.95896961e-02 -3.86689901e-01
-7.62502849e-01 7.64718592e-01 1.00029480e+00 -8.04801509e-02
4.37400907e-01 7.68834352e-02 1.83732212e-01 9.00978222e-02
-2.52861440e-01 -1.23863555e-01 3.11014891e-01 6.16602838e-01
3.08727682e-01 -9.32013690e-02 -2.42403269e-01 5.21246791e-01
1.36787489e-01 -7.22047538e-02 3.63343447e-01 4.47277099e-01
-2.67477810e-01 -1.10639572e+00 -4.78436351e-01 1.22980542e-01
-3.69181156e-01 -2.91026592e-01 -2.47318372e-01 5.12630224e-01
4.93997246e-01 7.89121032e-01 -3.03256582e-03 -1.32618457e-01
3.01820666e-01 -5.28869092e-01 3.67449254e-01 -4.90926594e-01
-4.04917330e-01 3.48137110e-01 -4.04915452e-01 -5.73927641e-01
-7.01126814e-01 -3.56968760e-01 -1.18692875e+00 -4.73073155e-01
-1.53099731e-01 -1.15705743e-01 2.04536602e-01 6.36972427e-01
3.10078263e-01 8.14337134e-01 6.34380341e-01 -7.14173496e-01
-1.05486512e-01 -4.87975270e-01 -5.60780704e-01 6.98785663e-01
3.62935871e-01 -5.65042734e-01 -7.50119612e-02 5.37753403e-01] | [11.20981216430664, -1.2640913724899292] |
b93d6eb1-9dc0-4aff-8fca-cda121fdea1d | d3rlpy-an-offline-deep-reinforcement-learning | 2111.03788 | null | https://arxiv.org/abs/2111.03788v2 | https://arxiv.org/pdf/2111.03788v2.pdf | d3rlpy: An Offline Deep Reinforcement Learning Library | In this paper, we introduce d3rlpy, an open-sourced offline deep reinforcement learning (RL) library for Python. d3rlpy supports a set of offline deep RL algorithms as well as off-policy online algorithms via a fully documented plug-and-play API. To address a reproducibility issue, we conduct a large-scale benchmark with D4RL and Atari 2600 dataset to ensure implementation quality and provide experimental scripts and full tables of results. The d3rlpy source code can be found on GitHub: \url{https://github.com/takuseno/d3rlpy}. | ['Michita Imai', 'Takuma Seno'] | 2021-11-06 | null | null | null | null | ['d4rl'] | ['robots'] | [-1.05313432e+00 -1.62519097e-01 -5.49651146e-01 -2.47168019e-01
-9.69584048e-01 -8.97379994e-01 4.64147687e-01 -1.71826571e-01
-4.57632512e-01 1.04837573e+00 1.49599031e-01 -7.12892652e-01
2.12633729e-01 -7.48138964e-01 -7.75693774e-01 -5.17739415e-01
-2.62397975e-01 3.82537037e-01 2.05667987e-01 -1.36241332e-01
-4.83473763e-02 4.99212980e-01 -1.34886754e+00 2.66624987e-01
3.72135073e-01 7.61922538e-01 2.11902216e-01 8.70830417e-01
1.45898953e-01 1.08538210e+00 -3.90912265e-01 -1.05340108e-01
7.49337196e-01 -1.88812017e-01 -8.11918616e-01 -6.52695060e-01
-1.86931789e-01 -1.08861184e+00 -6.77838683e-01 8.15423608e-01
1.03774750e+00 8.06357637e-02 -1.91930626e-02 -1.35098767e+00
-4.19358492e-01 4.69109029e-01 -2.85695285e-01 3.21987987e-01
6.62123263e-01 1.09695339e+00 9.58260119e-01 -6.02465689e-01
5.50863326e-01 1.08400476e+00 3.41051221e-01 6.34314477e-01
-9.78334308e-01 -1.18226075e+00 -1.44799754e-01 1.19293042e-01
-9.77910817e-01 -4.54719931e-01 5.61148524e-01 -1.61252588e-01
1.08440506e+00 1.79518235e-03 7.05162823e-01 1.64566410e+00
7.55492598e-02 1.27482975e+00 1.53891218e+00 6.87925443e-02
4.44235682e-01 -4.17935848e-01 -6.58301264e-02 9.20526505e-01
-1.27977073e-01 6.50619268e-01 -4.52933252e-01 -2.35054508e-01
1.08573806e+00 -2.59353012e-01 1.40300229e-01 6.16187556e-03
-9.94815648e-01 9.48132038e-01 3.34832728e-01 -5.73489815e-02
-3.92367840e-01 4.81289834e-01 6.38212323e-01 7.41455793e-01
3.13016921e-01 1.78695947e-01 -7.45750606e-01 -7.61292338e-01
-3.83087099e-01 7.70480931e-01 9.68202829e-01 8.30329299e-01
8.24154735e-01 1.83739051e-01 -2.01941654e-01 6.79066181e-01
1.94258511e-01 3.93123746e-01 6.57638013e-01 -1.65512335e+00
2.33548120e-01 9.43574160e-02 2.96376407e-01 -2.50133127e-01
-4.57638711e-01 -8.38499591e-02 -2.22161323e-01 6.10165298e-01
5.43217480e-01 -4.76684451e-01 -5.50197005e-01 1.63181579e+00
5.40895581e-01 -1.27104465e-02 1.23033144e-01 1.20000088e+00
1.15582418e+00 5.37335813e-01 6.92021400e-02 1.31465212e-01
9.66838241e-01 -1.15905583e+00 -3.21859568e-01 -7.81558603e-02
6.23191714e-01 -5.98730326e-01 1.53184009e+00 4.32912827e-01
-1.09810197e+00 -3.83600414e-01 -8.80629361e-01 -3.70002896e-01
-4.01838005e-01 1.57478273e-01 9.82680500e-01 6.90649301e-02
-1.17377865e+00 7.07121313e-01 -1.28931201e+00 -3.15242529e-01
3.09226185e-01 2.52570480e-01 -1.47089332e-01 7.28199929e-02
-1.38186789e+00 7.38550246e-01 4.67743367e-01 -2.31775418e-01
-1.74207401e+00 -4.67177510e-01 -5.86921692e-01 -3.39313895e-01
5.63146412e-01 -2.48933643e-01 2.01042032e+00 -7.92736769e-01
-2.12945104e+00 8.53273511e-01 3.27964187e-01 -4.25509304e-01
8.76502156e-01 -5.63303471e-01 -9.73607600e-02 1.01971619e-01
2.31753122e-02 5.13229311e-01 3.77440035e-01 -9.89205956e-01
-3.10542911e-01 -2.07308292e-01 2.54899651e-01 1.84746325e-01
3.02714467e-01 4.92510684e-02 -6.02595508e-01 -5.70937514e-01
-6.74936056e-01 -8.98492098e-01 1.88453719e-02 -4.05600294e-02
-1.14435919e-01 -4.02897686e-01 6.33360684e-01 -8.10082257e-01
9.44921970e-01 -2.14635658e+00 -2.84695774e-01 -2.24999458e-01
5.81697114e-02 3.33896637e-01 -3.26328069e-01 7.20027447e-01
1.56411096e-01 -1.95150137e-01 -7.46085029e-03 -1.26225784e-01
4.17529017e-01 4.47956443e-01 -2.80435294e-01 5.38860738e-01
-2.03344643e-01 9.94588315e-01 -1.16914606e+00 -2.58585513e-01
3.53370130e-01 5.88160083e-02 -5.92634439e-01 5.36464691e-01
-6.72283590e-01 6.83899164e-01 -7.03642547e-01 9.10306334e-01
4.37120438e-01 -2.16745242e-01 3.76445681e-01 3.28548640e-01
-4.66008514e-01 6.26984894e-01 -8.26874435e-01 2.08075023e+00
-3.85761082e-01 4.16612923e-01 1.49656057e-01 -6.11691594e-01
6.67981267e-01 2.21758470e-01 7.13648617e-01 -1.01997614e+00
2.45830119e-01 2.36458972e-01 -3.62092316e-01 -4.43209678e-01
3.79484028e-01 5.23195207e-01 -1.79547384e-01 8.77475083e-01
1.90492734e-01 5.21346964e-02 2.62667805e-01 1.08138034e-02
1.48256207e+00 8.52357328e-01 3.71921659e-01 -2.62687862e-01
7.20223114e-02 2.25660115e-01 5.15533149e-01 9.35833931e-01
-6.50384068e-01 -2.49390434e-02 6.74622238e-01 -6.79010332e-01
-1.02397418e+00 -1.40289474e+00 7.52372015e-03 1.55078423e+00
-2.86244690e-01 -3.57612222e-01 -4.15479034e-01 -7.18628705e-01
5.04627109e-01 7.35559285e-01 -4.50973392e-01 2.21237689e-01
-3.67450565e-01 -4.37794328e-01 9.98001456e-01 3.26707453e-01
8.41337323e-01 -1.60862470e+00 -7.05400467e-01 1.74957961e-01
-7.62924775e-02 -9.57310200e-01 -6.23110384e-02 2.90777028e-01
-5.19670069e-01 -1.16077340e+00 -4.64542538e-01 -3.82760733e-01
4.67793755e-02 -6.12393692e-02 1.24740982e+00 6.75025880e-02
-2.00139642e-01 5.34280479e-01 -4.78797793e-01 -2.95207590e-01
-4.39380318e-01 9.59403887e-02 1.35467023e-01 -8.65617931e-01
4.42397624e-01 -6.79810047e-01 -8.43263328e-01 1.02837175e-01
-6.12909019e-01 1.62073538e-01 2.28301048e-01 5.09577632e-01
5.87621629e-01 -4.72724974e-01 4.78869170e-01 -3.87308687e-01
8.66375625e-01 -7.85447776e-01 -1.33821595e+00 -2.29670972e-01
-5.03109455e-01 1.27861217e-01 7.89877892e-01 -7.66184628e-02
-6.23010695e-01 -4.05970104e-02 -7.72072136e-01 -4.90230411e-01
-3.21163952e-01 2.01976389e-01 1.52696371e-01 2.76188612e-01
4.73146975e-01 4.43781972e-01 2.78137952e-01 -6.12851501e-01
4.84659165e-01 6.74446702e-01 3.79419059e-01 -1.05996978e+00
3.69548649e-01 2.89897889e-01 -5.91081798e-01 -3.94938052e-01
-8.76351595e-01 -1.66068092e-01 -1.52126312e-01 -4.07813400e-01
7.03615427e-01 -1.20661533e+00 -1.38640535e+00 5.33835948e-01
-6.32215977e-01 -1.66258204e+00 -2.52526283e-01 3.51749003e-01
-8.34191978e-01 1.40727624e-01 -1.01044917e+00 -5.60927451e-01
-6.01442516e-01 -1.27730656e+00 8.05451274e-01 2.68094003e-01
4.33576331e-02 -8.38140428e-01 5.79531968e-01 1.71369776e-01
3.04482698e-01 2.01198310e-01 4.06679213e-01 -7.29438543e-01
-4.42342639e-01 2.63370544e-01 -1.05145805e-01 4.85661656e-01
-1.60301208e-01 2.44611315e-03 -1.02268016e+00 -6.29362762e-01
-3.94937366e-01 -1.28804517e+00 5.41614473e-01 2.69984365e-01
1.58909225e+00 -4.01494056e-01 2.90563047e-01 7.99397349e-01
1.43901646e+00 1.19683839e-01 4.99211520e-01 8.53446424e-01
2.96873510e-01 -6.01956174e-02 7.41055429e-01 1.14185107e+00
5.86886406e-01 4.05468822e-01 5.25603354e-01 9.54670366e-03
1.53750584e-01 -6.07838035e-01 6.76414073e-01 6.37318552e-01
8.44247639e-02 -5.33132963e-02 -9.86092150e-01 2.09650233e-01
-1.96359777e+00 -1.01555884e+00 2.35314727e-01 2.04426956e+00
1.19347346e+00 -9.00054574e-02 6.60297215e-01 -3.11643600e-01
1.64874181e-01 2.38799930e-01 -1.05360019e+00 -4.79079366e-01
1.19377181e-01 4.27864879e-01 6.14700556e-01 5.74325681e-01
-8.31225693e-01 1.32999694e+00 6.09395695e+00 8.31110299e-01
-1.29393792e+00 4.18857425e-01 4.82334226e-01 -3.83618683e-01
-1.01911938e-02 6.69989139e-02 -7.30359614e-01 6.29251182e-01
9.91581798e-01 -3.08451295e-01 1.15348148e+00 1.09240270e+00
5.21266937e-01 -2.93376744e-01 -7.63335168e-01 6.66378021e-01
-7.06037819e-01 -1.28475404e+00 -4.44516838e-01 1.12006359e-01
2.65754938e-01 9.44820523e-01 3.42991501e-02 9.64903295e-01
1.29255056e+00 -7.52953887e-01 5.80982029e-01 1.17762215e-01
8.46503854e-01 -8.42225373e-01 1.66394234e-01 3.14420789e-01
-8.23230207e-01 -1.65374339e-01 -3.12283158e-01 -3.02697301e-01
-4.32562351e-01 2.22182110e-01 -7.19359338e-01 4.56221759e-01
1.09023094e+00 8.09145272e-01 -3.31091881e-01 6.59800231e-01
-7.15697229e-01 8.21097910e-01 -2.38207221e-01 4.23873542e-03
3.38895947e-01 -1.24875583e-01 2.92297512e-01 1.12825227e+00
-9.42905843e-02 3.03368241e-01 3.63377213e-01 8.83432567e-01
-1.85345709e-01 -1.30240858e-01 -5.90054452e-01 -2.76586056e-01
5.59250414e-01 1.34867239e+00 -3.25577974e-01 -2.47105628e-01
-3.45362574e-01 9.03696775e-01 6.91288948e-01 3.65058303e-01
-1.13373661e+00 -2.25991622e-01 1.16588056e+00 -1.61217585e-01
1.26505449e-01 -7.31987536e-01 2.14196205e-01 -1.16172981e+00
-4.22969609e-01 -1.28082097e+00 6.03918910e-01 -8.39979708e-01
-1.11347771e+00 2.66762584e-01 -6.02747612e-02 -1.10392642e+00
-2.19440922e-01 -6.34203851e-01 -5.27137220e-01 4.02106643e-01
-1.60567117e+00 -7.88905442e-01 -2.85512745e-01 9.31133270e-01
5.37275016e-01 -3.13518167e-01 9.02174652e-01 3.18418205e-01
-7.46081233e-01 5.53914249e-01 4.86904353e-01 2.87576407e-01
6.95328891e-01 -1.30489218e+00 4.50053513e-01 4.35705036e-01
-4.72339004e-01 4.98418510e-01 6.57386184e-01 -6.10527217e-01
-1.83711410e+00 -1.09334540e+00 -9.06777978e-02 -2.52885252e-01
1.05982709e+00 -3.65141809e-01 -6.88212276e-01 1.03134561e+00
4.19724464e-01 4.71991077e-02 6.15440786e-01 -1.36956990e-01
-1.28699183e-01 -4.30416316e-02 -1.00228989e+00 6.72432780e-01
8.22428167e-01 -5.45890749e-01 -2.13889942e-01 5.19991398e-01
6.97796941e-01 -8.12137306e-01 -1.11842656e+00 2.41155759e-01
7.23969996e-01 -1.19881225e+00 8.16545486e-01 -3.23060572e-01
5.30045092e-01 -1.50881112e-01 3.21052410e-02 -1.10489655e+00
2.61697657e-02 -9.34112310e-01 -5.12062371e-01 6.80089772e-01
6.59337789e-02 -8.72349203e-01 5.97367883e-01 -1.74325909e-02
-2.23961268e-02 -8.78703952e-01 -8.87350261e-01 -8.96421492e-01
3.11760217e-01 -6.14194214e-01 6.03156507e-01 8.32489550e-01
2.96247192e-02 -2.44543627e-01 -3.37900519e-01 2.07266724e-03
6.24068856e-01 1.70178458e-01 1.18037081e+00 -2.14991689e-01
-7.88984418e-01 -3.57248187e-01 2.59437293e-01 -1.11427820e+00
4.68858272e-01 -1.03264678e+00 -7.51852021e-02 -1.31083357e+00
-6.01130649e-02 -5.46653807e-01 -3.79101604e-01 1.16692364e+00
2.57467926e-01 1.47304852e-02 2.08666950e-01 3.80865127e-01
-9.37616765e-01 9.54179525e-01 1.32797682e+00 2.43460596e-01
-2.44755805e-01 -1.04715906e-01 -3.39023858e-01 4.25515354e-01
1.36986756e+00 -5.56377470e-01 -2.74792612e-01 -5.35552919e-01
-1.16261289e-01 4.58726943e-01 5.62304676e-01 -9.50016499e-01
-2.87830561e-01 -3.34365427e-01 4.55671608e-01 -4.94332671e-01
2.64694244e-01 -3.97219241e-01 -1.93837434e-01 8.09058428e-01
-3.75550210e-01 4.87646252e-01 6.41261876e-01 7.51838088e-02
2.31128082e-01 2.17049643e-01 7.74623275e-01 -4.19643611e-01
-8.65811288e-01 4.26058143e-01 -6.06243312e-01 3.05984467e-01
9.59429026e-01 3.52396548e-01 -6.65543437e-01 -5.31512380e-01
-5.91525257e-01 5.86773455e-01 6.69300079e-01 4.51259732e-01
4.39643234e-01 -1.18541276e+00 -5.14760494e-01 1.82538584e-01
-1.01050310e-01 -2.34557360e-01 -7.22747296e-02 7.99605846e-01
-8.88875186e-01 2.49529347e-01 -6.60148561e-01 -2.25964323e-01
-7.37366617e-01 3.61140817e-01 5.99908471e-01 -3.16082627e-01
-7.69306302e-01 4.66535211e-01 -3.19067597e-01 -1.03563499e+00
3.59211683e-01 -2.50173718e-01 2.74288476e-01 -5.60835600e-01
4.06038463e-01 4.01259065e-01 -2.68635601e-01 1.42834544e-01
-2.66629100e-01 -2.55521357e-01 1.91793963e-01 -2.36793742e-01
1.50399029e+00 2.04571605e-01 6.78126141e-02 4.38942313e-01
1.05352259e+00 -1.35480210e-01 -1.72545600e+00 1.08966883e-02
-1.54536277e-01 -2.55557358e-01 3.94976735e-02 -9.89171445e-01
-9.92016137e-01 3.38604182e-01 7.39131689e-01 -2.31819585e-01
1.05582225e+00 -2.03246385e-01 8.26432168e-01 4.42779452e-01
6.82464421e-01 -1.23986745e+00 1.82267427e-01 8.71509969e-01
9.80597019e-01 -1.28196955e+00 2.49999523e-01 6.67158663e-01
-7.53372192e-01 9.96796310e-01 8.63355160e-01 -4.83117551e-01
5.82359254e-01 4.50548798e-01 3.09754938e-01 -1.30391270e-01
-1.10319352e+00 -3.28904450e-01 -6.07867002e-01 2.18533218e-01
5.33613443e-01 1.52020827e-01 -3.95184010e-01 5.80590308e-01
-2.89592475e-01 5.67559063e-01 4.07939225e-01 1.28951192e+00
-1.59756601e-01 -1.34001553e+00 -1.55851349e-01 5.25596030e-02
-5.72522879e-01 8.05877000e-02 -2.12587655e-01 9.85432625e-01
-3.24593246e-01 6.23085856e-01 -9.84265953e-02 -4.46187854e-01
-2.86002061e-03 -2.31313184e-01 4.85430956e-01 -1.32193834e-01
-7.31402457e-01 -1.35856345e-01 2.04204470e-01 -1.15533698e+00
1.03980035e-01 -4.73935217e-01 -1.58059561e+00 -7.60452747e-01
5.17647624e-01 7.16558546e-02 7.07050920e-01 6.79777384e-01
5.53101361e-01 1.86800137e-01 6.12338424e-01 -1.00673330e+00
-7.79943645e-01 -7.78562903e-01 -4.80345219e-01 5.85550442e-02
6.76121488e-02 -5.54121733e-01 -1.77222759e-01 -5.63913345e-01] | [4.0651984214782715, 1.5559855699539185] |
4205edc5-b27e-47ec-97e5-75e3f4ee4906 | matched-sample-selection-with-gans-for | 2103.13455 | null | https://arxiv.org/abs/2103.13455v1 | https://arxiv.org/pdf/2103.13455v1.pdf | Matched sample selection with GANs for mitigating attribute confounding | Measuring biases of vision systems with respect to protected attributes like gender and age is critical as these systems gain widespread use in society. However, significant correlations between attributes in benchmark datasets make it difficult to separate algorithmic bias from dataset bias. To mitigate such attribute confounding during bias analysis, we propose a matching approach that selects a subset of images from the full dataset with balanced attribute distributions across protected attributes. Our matching approach first projects real images onto a generative adversarial network (GAN)'s latent space in a manner that preserves semantic attributes. It then finds image matches in this latent space across a chosen protected attribute, yielding a dataset where semantic and perceptual attributes are balanced across the protected attribute. We validate projection and matching strategies with qualitative, quantitative, and human annotation experiments. We demonstrate our work in the context of gender bias in multiple open-source facial-recognition classifiers and find that bias persists after removing key confounders via matching. Code and documentation to reproduce the results here and apply the methods to new data is available at https://github.com/csinva/matching-with-gans . | ['Pietro Perona', 'Guha Balakrishnan', 'Chandan Singh'] | 2021-03-24 | null | null | null | null | ['gender-bias-detection', 'gender-bias-detection'] | ['miscellaneous', 'natural-language-processing'] | [ 5.29827774e-01 1.63966432e-01 -2.86596477e-01 -1.02350593e+00
-6.78250730e-01 -6.77485764e-01 7.42490351e-01 4.81251851e-02
-4.89567935e-01 4.56819743e-01 5.53663075e-01 4.36846819e-03
-2.45547183e-02 -8.79541457e-01 -7.41682112e-01 -6.47616565e-01
2.94581652e-01 2.43335992e-01 -4.91770148e-01 2.26827249e-01
2.03104541e-01 3.60842019e-01 -1.62750030e+00 3.32689226e-01
6.87500656e-01 8.56742084e-01 -5.09912848e-01 3.71850133e-01
3.89989465e-01 2.99732298e-01 -5.48816442e-01 -9.66278911e-01
7.09064543e-01 -3.77191067e-01 -5.87405443e-01 -2.08325103e-01
1.29383707e+00 -4.08230841e-01 -2.60624558e-01 1.26445341e+00
7.84581423e-01 -3.36741418e-01 6.89931095e-01 -1.79660344e+00
-8.96376967e-01 3.63109440e-01 -4.94657248e-01 -8.27350169e-02
7.38532394e-02 4.38713253e-01 9.90325511e-01 -7.65399754e-01
6.60871565e-01 1.44165003e+00 8.38777125e-01 7.35838294e-01
-1.80494475e+00 -1.31922710e+00 -2.06338335e-02 -1.94250748e-01
-1.36129475e+00 -7.67810881e-01 6.78927422e-01 -7.26008415e-01
3.70680571e-01 3.30558240e-01 5.36441505e-01 1.61957085e+00
1.19451977e-01 3.54821146e-01 1.45659328e+00 -2.32143909e-01
3.74604017e-02 2.20478758e-01 3.46006267e-02 6.10423684e-01
4.26759452e-01 3.91780138e-01 -8.35493147e-01 -5.30680418e-01
3.96847159e-01 1.31244928e-01 1.16575353e-01 -4.51284856e-01
-1.28542745e+00 9.78339911e-01 3.91282827e-01 -3.49011123e-01
-3.32564205e-01 3.45178217e-01 2.06501737e-01 2.11946309e-01
6.14035726e-01 4.22277749e-01 -1.30979031e-01 1.93529606e-01
-9.47909653e-01 4.30982649e-01 3.40188742e-01 8.22778523e-01
8.51452470e-01 -1.26724005e-01 -4.11521047e-01 9.86730576e-01
1.31729156e-01 8.90250266e-01 2.06604272e-01 -1.39069831e+00
5.67431599e-02 5.83967328e-01 -1.94924235e-01 -1.22097540e+00
-1.21968322e-01 -1.48599759e-01 -6.97570860e-01 4.56875652e-01
5.64803481e-01 -1.12826757e-01 -1.04452646e+00 2.23459077e+00
2.89398372e-01 -3.15209061e-01 -3.09767067e-01 7.23635852e-01
8.16338122e-01 -1.05807018e-02 5.17820835e-01 2.56901503e-01
1.49387705e+00 -2.01904655e-01 -3.40459019e-01 -3.80534947e-01
1.74564347e-01 -7.01498926e-01 1.13030887e+00 9.97507572e-02
-8.57240796e-01 -4.20521498e-01 -9.27497625e-01 -1.87801197e-01
-3.59180391e-01 8.98647234e-02 6.47093475e-01 1.02942514e+00
-1.09182155e+00 3.66722703e-01 -3.94236743e-01 -5.23193836e-01
1.00693643e+00 5.41246176e-01 -5.59922516e-01 -7.24148750e-02
-9.18005466e-01 5.99212706e-01 -2.76732683e-01 -2.47122586e-01
-1.00202036e+00 -1.09336531e+00 -7.50203073e-01 -3.09950829e-01
1.98974218e-02 -8.55943203e-01 9.40867662e-01 -1.38680780e+00
-8.66921246e-01 1.44019759e+00 -1.99275583e-01 -1.96611509e-01
6.79456294e-01 -1.05853692e-01 -1.65116355e-01 -2.65799463e-01
3.27485204e-01 1.23069167e+00 1.06546104e+00 -1.17069507e+00
-2.77471036e-01 -7.27548778e-01 -5.87336496e-02 -1.21204130e-01
-4.98538643e-01 2.41064340e-01 6.67787939e-02 -7.52906144e-01
-1.36524409e-01 -1.22768950e+00 7.88984820e-03 3.24653476e-01
-3.35231632e-01 8.82494822e-02 2.66185939e-01 -6.66694522e-01
7.15885103e-01 -2.21429038e+00 9.34574157e-02 2.81223714e-01
4.70010519e-01 -2.63426423e-01 -2.48631060e-01 -2.98487097e-02
-1.96751907e-01 2.11195126e-01 -3.69381756e-01 -4.86802518e-01
1.00210011e-01 -4.17633541e-02 -3.42565149e-01 7.54657984e-01
1.00490674e-01 8.12552392e-01 -6.21925235e-01 -5.29336095e-01
-6.74172342e-02 5.16090512e-01 -8.31168950e-01 1.15886293e-01
8.18635300e-02 2.22624883e-01 -2.21880600e-02 8.93614948e-01
8.18850815e-01 2.26244792e-01 1.22046517e-02 -3.45562190e-01
1.06848344e-01 -2.17529647e-02 -7.71291018e-01 1.40943849e+00
-1.61621332e-01 7.43085504e-01 7.72489011e-02 -3.93586040e-01
1.10282242e+00 -1.16075978e-01 4.60522592e-01 -6.38920784e-01
1.69667810e-01 6.50626719e-02 6.41097203e-02 -7.74549916e-02
3.87164652e-01 -2.08195120e-01 -1.98353887e-01 4.09331113e-01
1.41165808e-01 -6.74004331e-02 -2.69374877e-01 1.93303466e-01
9.10148799e-01 -7.26757124e-02 4.51449454e-02 -3.40111077e-01
6.41715527e-02 -3.28446478e-02 7.51719236e-01 7.43939519e-01
-4.48814303e-01 9.19916213e-01 6.66709304e-01 -4.67341334e-01
-1.20978594e+00 -1.19977999e+00 -4.29296613e-01 1.18502235e+00
-4.18501437e-01 -2.62013972e-01 -7.96258569e-01 -7.35320091e-01
6.31245196e-01 5.42292416e-01 -1.22874117e+00 -4.28778261e-01
-1.36127964e-01 -8.78268600e-01 9.62398112e-01 5.76298475e-01
3.63384098e-01 -9.17533517e-01 -4.98999774e-01 -5.52088380e-01
-3.20545249e-02 -7.86628187e-01 -3.95855546e-01 -2.41490334e-01
-5.45911670e-01 -1.16062212e+00 -5.57612896e-01 -2.26887926e-01
9.54055667e-01 -1.90389320e-01 1.24224055e+00 6.99277520e-02
-5.25478482e-01 5.69159389e-01 1.62970364e-01 -8.58046830e-01
-4.59510624e-01 -7.00606359e-03 1.07870124e-01 2.97005564e-01
7.53962338e-01 -4.46499735e-01 -8.77089858e-01 4.44102645e-01
-6.42574430e-01 8.04966465e-02 4.86999243e-01 7.24239051e-01
4.59047347e-01 -6.17517471e-01 3.78942847e-01 -1.04911625e+00
4.57449019e-01 -5.76136291e-01 -6.47615731e-01 -2.01983005e-02
-9.82060730e-01 -1.89444646e-01 -1.48525005e-02 -4.24465537e-01
-8.53471816e-01 9.89214331e-02 2.85528034e-01 -3.47827256e-01
-3.47798198e-01 -1.32124260e-01 -4.90378380e-01 4.06053364e-02
8.62075031e-01 -1.66896969e-01 3.36668849e-01 -2.34045312e-01
3.16938549e-01 6.80338383e-01 5.04053175e-01 -6.83270633e-01
7.61907578e-01 7.96281874e-01 1.03168882e-01 -3.00802052e-01
-6.90954924e-01 -2.69833449e-02 -7.13260651e-01 -2.88022876e-01
9.12520587e-01 -1.01714909e+00 -6.72705591e-01 5.48937857e-01
-6.88938320e-01 -3.33050162e-01 -4.21587557e-01 3.91123682e-01
-4.50259089e-01 -1.91663370e-01 -1.13925450e-01 -5.58001459e-01
-2.71001905e-01 -1.08946157e+00 1.10819781e+00 1.31483346e-01
-7.87687361e-01 -6.20287538e-01 2.21348614e-01 5.60604036e-01
3.20578903e-01 4.32926238e-01 7.77623594e-01 -5.86463988e-01
-3.35859627e-01 -1.93862915e-01 -3.54380041e-01 5.33012331e-01
2.63897210e-01 1.52860463e-01 -1.39816904e+00 -3.23736101e-01
-4.65732247e-01 -3.88396561e-01 1.02695787e+00 5.12846470e-01
1.27677405e+00 -2.33737111e-01 -2.61753827e-01 9.46036041e-01
1.23413169e+00 -1.96428448e-01 5.53472579e-01 2.44116038e-01
7.77577639e-01 1.00156629e+00 3.30076605e-01 3.70764285e-01
2.77271807e-01 5.39117336e-01 3.34309638e-01 -2.94557095e-01
-2.36963868e-01 -4.36824292e-01 2.50090450e-01 -1.88860759e-01
-2.07899027e-02 2.17298359e-01 -1.08580720e+00 7.24029839e-01
-1.38418949e+00 -9.55582857e-01 1.06854118e-01 2.33897233e+00
9.91748333e-01 -1.08265363e-01 2.84145713e-01 -1.20820358e-01
7.43918121e-01 1.90177903e-01 -6.41319215e-01 -3.78009826e-01
-3.08957517e-01 1.24143064e-01 8.84870589e-01 4.73344058e-01
-9.28805113e-01 6.83088064e-01 6.54223728e+00 3.04711908e-01
-1.01067221e+00 1.28717199e-01 1.11594164e+00 -5.53907990e-01
-5.69101393e-01 3.82134765e-02 -6.23297572e-01 5.17355621e-01
7.96562850e-01 -1.92954719e-01 4.81724143e-01 7.46936500e-01
1.81751195e-02 5.53768873e-02 -1.38885093e+00 9.13314342e-01
1.81874141e-01 -9.94608223e-01 1.13550805e-01 4.36851352e-01
7.44268179e-01 -9.70239192e-02 6.93041205e-01 -7.66032636e-02
4.95614976e-01 -1.43854249e+00 8.27754498e-01 5.32993972e-01
1.23901534e+00 -6.16670012e-01 4.63303626e-01 -3.90367925e-01
-3.54868650e-01 -1.37915492e-01 -3.48544449e-01 8.63038450e-02
-5.63706219e-01 6.20696485e-01 -7.23706841e-01 -5.42726219e-02
9.85894859e-01 5.40505707e-01 -1.02534258e+00 5.01140952e-01
-6.97222278e-02 6.47876680e-01 -4.32279259e-02 4.63138670e-01
-3.93773645e-01 -1.72712341e-01 4.18358356e-01 1.13329160e+00
2.93820292e-01 -9.53511521e-02 -3.13732773e-01 1.11767030e+00
-4.13553655e-01 8.51217508e-02 -8.80026042e-01 1.48679847e-02
6.86830103e-01 1.25620437e+00 -4.08959895e-01 -1.39872581e-01
-4.02373374e-01 7.61763334e-01 6.43372238e-02 3.01404059e-01
-6.67787552e-01 -4.96954657e-02 1.35086679e+00 2.49199361e-01
-2.87330240e-01 1.42770573e-01 -7.26461530e-01 -8.36878479e-01
-6.00267760e-02 -1.22921467e+00 5.87066114e-01 -7.45903850e-01
-1.37044740e+00 2.00147375e-01 3.04756254e-01 -9.57772970e-01
3.08947898e-02 -5.04824817e-01 -2.31507361e-01 1.17393947e+00
-1.09328568e+00 -1.52701294e+00 -6.84167027e-01 6.62579358e-01
5.92761375e-02 -4.54328179e-01 9.05594945e-01 2.21682623e-01
-4.66658324e-01 9.31863427e-01 -3.58424723e-01 3.76965612e-01
1.41393709e+00 -1.14071321e+00 4.27126378e-01 7.65621424e-01
8.67131501e-02 8.72711658e-01 6.90178931e-01 -7.19584763e-01
-9.94228601e-01 -9.64013338e-01 7.12813675e-01 -9.77197945e-01
3.11849356e-01 -5.78284979e-01 -6.67944074e-01 9.64756548e-01
2.57576317e-01 1.07905522e-01 1.14684010e+00 3.24460149e-01
-1.00317740e+00 -4.87480342e-01 -1.42701507e+00 6.02315903e-01
1.33081639e+00 -6.17506623e-01 -2.81048536e-01 9.47911069e-02
3.00989777e-01 -2.53969491e-01 -8.82375836e-01 4.91003275e-01
1.01191235e+00 -9.89631295e-01 9.53172266e-01 -6.90649331e-01
6.42642736e-01 -1.37523144e-01 -4.31288987e-01 -1.36704600e+00
-3.08128536e-01 -4.18262362e-01 5.61490059e-01 1.57705188e+00
3.40680808e-01 -9.14555013e-01 7.49421716e-01 9.76443231e-01
3.27995539e-01 -4.12344456e-01 -8.12366307e-01 -4.79248643e-01
4.18357939e-01 -2.71409482e-01 1.06010044e+00 1.05064273e+00
-5.81891537e-01 -1.73639774e-01 -2.14041054e-01 2.10980773e-01
1.11012459e+00 1.07697099e-01 9.94222879e-01 -1.32673538e+00
1.10299766e-01 -4.78602141e-01 -5.31083465e-01 2.28719898e-02
4.49503273e-01 -1.01605988e+00 -2.50201136e-01 -9.89462137e-01
6.29738033e-01 -6.62185311e-01 -3.77069920e-01 8.27698410e-01
-1.06035650e-01 7.05375910e-01 1.95762560e-01 3.32104057e-01
5.32942079e-02 3.39028388e-01 7.71833181e-01 -3.34805399e-01
1.63239449e-01 -3.27464223e-01 -1.11313581e+00 5.79356372e-01
9.36106384e-01 -7.49297678e-01 -3.16046536e-01 -4.04184699e-01
2.29613259e-01 -5.91279924e-01 1.01042581e+00 -8.28168869e-01
-1.41401052e-01 -3.41570258e-01 9.29091811e-01 1.38216512e-02
3.31999183e-01 -7.19003201e-01 2.96418339e-01 3.52922589e-01
-7.22720563e-01 2.35986382e-01 2.85458833e-01 1.56892508e-01
8.92734081e-02 3.44193935e-01 9.20241952e-01 1.09408766e-01
-3.42528969e-01 3.28658998e-01 -1.90477986e-02 1.48811594e-01
8.58083069e-01 -1.39427349e-01 -4.92226273e-01 -4.15653765e-01
-4.75007683e-01 -1.50523065e-02 9.93012667e-01 6.10218942e-01
4.16765690e-01 -1.42135310e+00 -9.90997434e-01 5.11911511e-01
4.74617928e-01 -4.90984917e-01 2.45993994e-02 6.03279829e-01
-1.25895768e-01 -1.50295766e-02 -7.97426760e-01 -5.72718322e-01
-1.69394076e+00 3.54938954e-01 5.34009635e-01 5.27047336e-01
-1.02984823e-01 8.96586955e-01 4.03091013e-01 -4.96935248e-01
-1.21896425e-02 5.21614775e-02 2.25234941e-01 3.26753408e-01
2.52893656e-01 4.86140668e-01 -1.00148633e-01 -8.21565032e-01
-5.44801295e-01 5.00366211e-01 1.70634687e-01 -3.88809413e-01
1.25580013e+00 -6.84867520e-03 -2.49710768e-01 1.88271388e-01
1.16395926e+00 3.51687580e-01 -1.27315629e+00 -9.03274715e-02
-3.62931550e-01 -1.00342500e+00 -3.00013684e-02 -8.18226576e-01
-1.17571354e+00 6.30606174e-01 1.04151094e+00 -2.03557223e-01
1.15861475e+00 -3.03214733e-02 1.18270226e-01 -2.88377106e-01
1.34172395e-01 -8.99630904e-01 -2.33893558e-01 -1.88012213e-01
9.76094902e-01 -1.41825199e+00 3.09054792e-01 -2.97591627e-01
-5.84079623e-01 4.94228274e-01 7.48108327e-01 -4.80303876e-02
5.55186749e-01 1.50270820e-01 4.31765229e-01 -2.55809575e-01
-5.98364294e-01 1.10504970e-01 3.20663035e-01 7.77161598e-01
5.36314726e-01 2.81591237e-01 -1.43431306e-01 3.63482475e-01
-6.85009599e-01 -8.51698890e-02 3.31790477e-01 5.17034769e-01
3.08638275e-01 -1.29695296e+00 -4.84773993e-01 6.42982304e-01
-5.68070829e-01 -1.30011380e-01 -8.10005307e-01 6.54076397e-01
6.04511321e-01 6.42718494e-01 3.57737660e-01 -4.25396681e-01
2.94979066e-01 3.55744749e-01 5.77274561e-01 -4.64260012e-01
-6.32785082e-01 -4.50894207e-01 7.66533911e-02 -8.85110736e-01
-4.30394441e-01 -1.17579079e+00 -6.31423891e-01 -5.65646052e-01
3.28207344e-01 -3.53431463e-01 7.09830225e-01 2.84038991e-01
5.90969741e-01 5.44386543e-02 4.13402736e-01 -4.83736038e-01
-4.48140085e-01 -8.77939939e-01 -2.49157384e-01 8.73780966e-01
3.97253066e-01 -7.01296270e-01 -3.25481534e-01 2.74649382e-01] | [13.048284530639648, 1.1712429523468018] |
8793aaaf-a706-4bb3-bf40-55cb445a3303 | a-generative-modeling-approach-to-limited | 1802.06458 | null | http://arxiv.org/abs/1802.06458v3 | http://arxiv.org/pdf/1802.06458v3.pdf | A Generative Modeling Approach to Limited Channel ECG Classification | Processing temporal sequences is central to a variety of applications in
health care, and in particular multi-channel Electrocardiogram (ECG) is a
highly prevalent diagnostic modality that relies on robust sequence modeling.
While Recurrent Neural Networks (RNNs) have led to significant advances in
automated diagnosis with time-series data, they perform poorly when models are
trained using a limited set of channels. A crucial limitation of existing
solutions is that they rely solely on discriminative models, which tend to
generalize poorly in such scenarios. In order to combat this limitation, we
develop a generative modeling approach to limited channel ECG classification.
This approach first uses a Seq2Seq model to implicitly generate the missing
channel information, and then uses the latent representation to perform the
actual supervisory task. This decoupling enables the use of unsupervised data
and also provides highly robust metric spaces for subsequent discriminative
learning. Our experiments with the Physionet dataset clearly evidence the
effectiveness of our approach over standard RNNs in disease prediction. | ['Jayaraman J. Thiagarajan', 'Deepta Rajan'] | 2018-02-18 | null | null | null | null | ['ecg-classification'] | ['medical'] | [ 7.97717929e-01 -1.13448367e-01 2.88388748e-02 -2.88491666e-01
-8.73697281e-01 -4.10552710e-01 3.17834616e-01 9.68566760e-02
-2.85063386e-01 8.29183638e-01 3.59819233e-01 -3.97175997e-01
-3.11780870e-01 -4.11207318e-01 -2.88289309e-01 -8.47874463e-01
-1.09879814e-01 4.56926495e-01 -1.46250367e-01 -1.54516876e-01
-6.25798851e-02 4.87885654e-01 -1.12951958e+00 2.41261154e-01
6.74097240e-01 7.56172001e-01 1.10379271e-01 9.13603127e-01
2.85660237e-01 8.15130234e-01 -6.80986643e-01 2.71349493e-02
1.11259766e-01 -9.49966133e-01 -5.27434587e-01 -1.03586674e-01
-2.38539085e-01 -1.18041717e-01 -2.76641309e-01 6.48226917e-01
9.90529597e-01 1.42097902e-02 6.21647239e-01 -8.16542327e-01
1.20888494e-01 5.10226011e-01 -2.01679051e-01 3.52266192e-01
1.45343661e-01 -5.85533902e-02 8.58510852e-01 -4.56784427e-01
5.48492908e-01 7.50974357e-01 9.73965347e-01 6.48913503e-01
-1.57984209e+00 -5.01649857e-01 -2.33236045e-01 6.83884472e-02
-1.41907954e+00 -4.77929771e-01 9.69853222e-01 -3.33071768e-01
9.21715617e-01 2.34590262e-01 6.50086701e-01 1.49312687e+00
4.44862336e-01 5.95054924e-01 1.04507267e+00 -3.10675144e-01
3.40323031e-01 -1.92528367e-01 -5.69314398e-02 3.37099582e-01
-1.67511672e-01 8.96695256e-02 -4.17882562e-01 -4.09704387e-01
7.77430534e-01 3.84840965e-01 -2.66717494e-01 -3.52911681e-01
-1.32990134e+00 8.23703766e-01 -9.39863399e-02 5.56774616e-01
-6.48475826e-01 1.57743990e-01 6.15087867e-01 4.76083219e-01
3.66627634e-01 5.07462025e-01 -2.72825241e-01 -5.54544866e-01
-1.36049986e+00 1.20440975e-01 7.87931442e-01 4.98476774e-01
3.42848480e-01 1.59029424e-01 -2.02288911e-01 8.36368799e-01
-1.23286666e-02 2.20914215e-01 7.82637775e-01 -8.21424127e-01
3.08020622e-01 3.28487337e-01 -2.93665349e-01 -8.57572794e-01
-6.34769142e-01 -8.16311538e-01 -1.30803096e+00 -1.26684204e-01
1.98048294e-01 -1.63906276e-01 -1.04870975e+00 1.75856924e+00
1.17606698e-02 4.66247231e-01 1.16052695e-01 5.64628243e-01
2.59265929e-01 3.17007929e-01 -6.04767874e-02 -5.44357002e-01
7.86849797e-01 -3.70133877e-01 -6.56365573e-01 9.74178091e-02
8.34985077e-01 -3.91865164e-01 5.53473234e-01 5.84964335e-01
-8.37740898e-01 -3.79741043e-01 -9.79482234e-01 2.69790709e-01
3.00114248e-02 -1.78566337e-01 3.64447802e-01 7.50948548e-01
-1.10663724e+00 8.59761834e-01 -1.16495752e+00 -3.21985364e-01
4.30163532e-01 4.38086480e-01 -2.04534605e-01 1.02556750e-01
-1.17175257e+00 7.54738152e-01 4.01028872e-01 2.82288462e-01
-8.67608011e-01 -4.50089693e-01 -7.14283824e-01 -1.31823987e-01
1.43807858e-01 -7.30119646e-01 9.49075401e-01 -8.01836371e-01
-1.55163777e+00 4.43168283e-01 -1.30613163e-01 -8.15509439e-01
6.76775813e-01 -1.14774950e-01 -4.51685607e-01 2.36213639e-01
-9.68583450e-02 3.08797300e-01 1.01278687e+00 -7.19134867e-01
-2.93180287e-01 -1.93723634e-01 -4.68377411e-01 2.60079559e-02
-2.92953819e-01 -9.08180326e-02 -2.10908085e-01 -9.21640098e-01
2.57455498e-01 -1.19185352e+00 -6.33663356e-01 -5.90969145e-01
-5.08340180e-01 -3.31805125e-02 6.28843606e-01 -7.32023120e-01
1.50709879e+00 -2.24776220e+00 4.12105590e-01 5.40296018e-01
2.87673384e-01 2.25378379e-01 4.01121117e-02 5.21468937e-01
-2.77879268e-01 -4.19468097e-02 -5.73182344e-01 -3.37641746e-01
-3.76383543e-01 4.57925022e-01 -3.89625430e-01 4.85677540e-01
1.74245849e-01 8.89216900e-01 -7.87249386e-01 -3.50675434e-01
1.41024038e-01 6.15471542e-01 -3.95843744e-01 1.73685163e-01
-1.83387436e-02 8.99459600e-01 -4.36208695e-01 2.59918362e-01
-4.56070267e-02 -4.99267340e-01 5.82572103e-01 6.41188473e-02
1.79966077e-01 4.49806899e-01 -9.64067519e-01 2.04414797e+00
-3.13665211e-01 3.84402931e-01 -4.12283301e-01 -1.23994577e+00
7.00604200e-01 7.28988826e-01 1.03212619e+00 -4.06715393e-01
1.22732133e-01 3.24218631e-01 2.60462910e-01 -5.04253805e-01
-2.40782108e-02 -3.47846180e-01 -1.28724188e-01 6.57081068e-01
-3.75163406e-02 1.70161650e-01 1.63090639e-02 8.19823518e-02
1.47217023e+00 2.48610988e-01 4.22393411e-01 4.93519641e-02
2.92484134e-01 -3.07223260e-01 8.77795935e-01 7.91579783e-01
5.18266158e-03 9.66540694e-01 5.01724839e-01 -4.60521251e-01
-1.00237370e+00 -8.20684254e-01 -6.21978194e-02 5.69075227e-01
-3.39908391e-01 -5.22029161e-01 -4.94128227e-01 -7.65177965e-01
-2.40751609e-01 4.35098588e-01 -6.44204259e-01 -2.59360611e-01
-7.31477320e-01 -1.02571940e+00 8.92852962e-01 8.21085215e-01
-5.87741770e-02 -1.06986213e+00 -9.59008396e-01 7.91873515e-01
-3.18352193e-01 -9.05053675e-01 -1.21170536e-01 5.80011547e-01
-1.19019926e+00 -8.34725678e-01 -9.14218485e-01 -4.12567228e-01
2.96604782e-01 -2.76202917e-01 1.01054013e+00 -1.56651929e-01
-6.08418763e-01 3.92285079e-01 -3.10600668e-01 -3.13077569e-01
-6.57599032e-01 3.31683218e-01 1.50048211e-01 1.85870633e-01
1.89435020e-01 -1.01351690e+00 -5.51943243e-01 3.29447575e-02
-9.34424818e-01 -5.63027710e-02 7.14639902e-01 1.22191226e+00
6.48573995e-01 -9.87099186e-02 9.17416036e-01 -1.20985770e+00
7.16711164e-01 -5.21961212e-01 -1.64353698e-01 1.29500180e-01
-7.79107869e-01 1.07552536e-01 6.32509768e-01 -3.24363261e-01
-5.96065402e-01 2.43714750e-01 -4.02582198e-01 -6.36839449e-01
-2.19542190e-01 6.82366550e-01 6.72974586e-02 3.37457120e-01
6.53835952e-01 4.62833077e-01 1.86635703e-01 -4.81715590e-01
-8.63343775e-02 6.63859308e-01 4.16644245e-01 -2.46377543e-01
5.29288828e-01 4.00782764e-01 1.80400908e-01 -9.24068332e-01
-5.24897456e-01 -5.22491872e-01 -7.75736272e-01 4.69453931e-02
9.03159916e-01 -7.51786530e-01 -3.76781017e-01 2.14021772e-01
-8.71006489e-01 -2.76081204e-01 -3.45617533e-01 6.24415755e-01
-7.08775878e-01 4.88029093e-01 -7.19341934e-01 -9.18473065e-01
-4.79035228e-01 -9.26214814e-01 9.61254954e-01 -2.87040323e-01
-6.04649544e-01 -1.05843735e+00 3.42389315e-01 7.88687728e-03
4.04786110e-01 5.95055342e-01 9.95228887e-01 -7.96408534e-01
-2.86852598e-01 -2.60541320e-01 3.74068648e-01 5.38373470e-01
3.18118721e-01 -5.16535282e-01 -9.20592964e-01 -3.79725963e-01
2.76367635e-01 -1.99735001e-01 8.66000712e-01 3.61695200e-01
1.27223969e+00 7.22000152e-02 -4.18182135e-01 6.29576862e-01
1.21372342e+00 2.63659060e-01 6.49949253e-01 -1.43762589e-01
7.78554976e-01 3.42422038e-01 2.53917426e-01 4.74666923e-01
1.95695549e-01 5.00091195e-01 5.71544729e-02 -2.05893517e-01
1.72903374e-01 -2.77810022e-02 2.25275069e-01 1.23474371e+00
-2.51876205e-01 -1.06670648e-01 -1.06392455e+00 5.13062477e-01
-2.03471184e+00 -1.06976140e+00 9.60934162e-02 2.20124936e+00
8.89492989e-01 8.68606567e-02 2.82632589e-01 6.43141091e-01
4.32129711e-01 -9.22112167e-02 -6.25688314e-01 -1.41253874e-01
-1.84711590e-01 5.90722620e-01 2.54765958e-01 -3.87727246e-02
-1.01375675e+00 4.23576564e-01 6.80766010e+00 4.68005270e-01
-1.33871210e+00 8.51309225e-02 5.72231054e-01 -1.06353000e-01
-9.86559764e-02 -2.91200131e-01 -3.66431266e-01 5.36523759e-01
1.34824753e+00 1.42876387e-01 2.26559937e-01 3.63514811e-01
4.86947447e-01 1.93330243e-01 -1.23151231e+00 1.30877423e+00
1.61617845e-01 -1.15413952e+00 -1.14764132e-01 1.43065423e-01
4.86454308e-01 4.59131338e-02 4.81548496e-02 6.55935518e-03
-9.72483829e-02 -1.30629790e+00 1.52963996e-01 6.24907792e-01
9.46325362e-01 -8.25640559e-01 8.56629193e-01 4.20825630e-01
-7.65220821e-01 -8.38824064e-02 -7.22015575e-02 1.33891717e-01
2.23264724e-01 7.30295420e-01 -1.15298390e+00 6.62661791e-01
4.01034564e-01 1.01163864e+00 -3.55605006e-01 1.04734826e+00
-1.54398158e-01 1.01502621e+00 -3.06879401e-01 3.02006513e-01
-6.99935183e-02 -4.95050997e-02 5.93259573e-01 1.28574395e+00
3.89341801e-01 -1.28703505e-01 2.23719880e-01 4.63798404e-01
1.80399284e-01 5.34159914e-02 -7.43502438e-01 -1.79050192e-01
1.57043606e-01 9.91424859e-01 -7.67952025e-01 -2.80787528e-01
-2.96361446e-01 1.07850170e+00 8.19057524e-02 3.49682897e-01
-6.97854578e-01 -3.59961718e-01 3.27304304e-01 -3.65963131e-02
2.62795120e-01 -2.49245092e-01 -2.04446048e-01 -1.14211607e+00
2.64187306e-02 -1.22476447e+00 5.05455554e-01 -2.86907405e-01
-1.21786308e+00 6.87274039e-01 -3.07011455e-01 -1.48580885e+00
-9.51038718e-01 -1.77758917e-01 -2.63318062e-01 9.52006578e-01
-1.29598570e+00 -1.11879969e+00 1.43417552e-01 7.14476585e-01
4.01169509e-01 -8.61608908e-02 1.27171135e+00 4.84643161e-01
-4.10404503e-01 4.91802514e-01 9.81123224e-02 1.98922738e-01
7.09921002e-01 -1.25396848e+00 1.83277592e-01 9.04891789e-01
3.80984247e-01 7.28488445e-01 7.60372519e-01 -5.97210526e-01
-1.23390234e+00 -1.22143495e+00 8.51108551e-01 -4.66005951e-01
3.54304224e-01 -2.96265036e-01 -9.99025464e-01 5.66542685e-01
-1.75705075e-01 -2.50462983e-02 1.13305795e+00 1.69884533e-01
-1.46067426e-01 -7.10602775e-02 -6.50928378e-01 3.58481526e-01
9.16270137e-01 -8.20318639e-01 -5.87254584e-01 7.59487525e-02
2.18681961e-01 -2.79495835e-01 -1.05250514e+00 4.51858848e-01
6.82333291e-01 -8.96216452e-01 6.71115220e-01 -6.01132929e-01
2.40589157e-01 -3.22725713e-01 2.32552197e-02 -1.27852297e+00
-1.83583781e-01 -9.66386378e-01 -2.16913760e-01 9.08702016e-01
4.91395473e-01 -6.69389784e-01 8.32816780e-01 2.48677462e-01
-9.51785147e-02 -8.47629547e-01 -8.44996154e-01 -6.49222314e-01
-2.00871527e-01 -6.50466979e-01 2.79131711e-01 9.54484522e-01
-4.82007265e-02 3.34994704e-01 -7.84268558e-01 -1.31180296e-02
4.91135746e-01 1.57728106e-01 4.33582932e-01 -1.41662538e+00
-7.48017788e-01 -1.70408502e-01 -6.50550425e-01 -7.71363914e-01
1.00299641e-02 -9.51153040e-01 1.24660335e-01 -1.26864636e+00
4.50914055e-02 -5.60516179e-01 -8.36296499e-01 4.83309150e-01
-1.67733103e-01 4.77926701e-01 -1.56486824e-01 2.96604335e-01
-4.89213705e-01 2.38131389e-01 7.23742545e-01 8.76769647e-02
-5.37049949e-01 3.13638806e-01 -5.15688241e-01 5.28477788e-01
7.58977890e-01 -6.11989260e-01 -5.95385373e-01 -1.50946602e-01
3.44421417e-01 4.27537620e-01 2.73535609e-01 -1.08683848e+00
1.48478895e-01 3.19814622e-01 5.82014918e-01 -4.03965265e-01
3.78546596e-01 -6.39292359e-01 4.02840734e-01 5.63771248e-01
-4.65323061e-01 2.24529922e-01 -4.73712161e-02 7.67889857e-01
-3.85116428e-01 1.48168609e-01 5.44653475e-01 -1.20423190e-01
-2.40962967e-01 2.64331490e-01 -5.34692347e-01 9.14682150e-02
7.99082875e-01 -3.10459167e-01 4.87773925e-01 -4.81947839e-01
-1.08411694e+00 -7.67731667e-02 2.92437732e-01 3.48139107e-01
7.07710207e-01 -1.11588299e+00 -8.00076365e-01 3.83024871e-01
-2.97193006e-02 -1.12737998e-01 1.64878160e-01 1.21673453e+00
-2.74212390e-01 3.33492428e-01 -1.25584140e-01 -9.46465373e-01
-1.24225271e+00 3.11209679e-01 3.19438726e-01 -4.45936650e-01
-1.11357772e+00 5.17726898e-01 -4.68111970e-02 9.08960402e-02
1.14901200e-01 -1.74211487e-01 -2.75098622e-01 1.25089437e-01
4.09936219e-01 2.27404624e-01 3.16683352e-01 -5.30178785e-01
-2.62070745e-01 2.41047516e-01 -7.63031542e-02 -2.86540061e-01
1.58812976e+00 -1.83740053e-02 8.83911029e-02 8.69512081e-01
1.00046074e+00 -1.37550578e-01 -1.05172491e+00 -5.44057563e-02
3.26653838e-01 4.29358222e-02 -1.84441164e-01 -6.36103868e-01
-7.00232625e-01 1.14729249e+00 6.29260838e-01 2.37244338e-01
1.31252944e+00 -3.87873113e-01 8.33116353e-01 2.39451617e-01
4.13065374e-01 -9.78397012e-01 -8.20558369e-02 3.29576820e-01
4.85777676e-01 -8.34669888e-01 -1.50873140e-01 -6.05477728e-02
-5.01417577e-01 1.12619269e+00 -2.01283887e-01 -2.06682831e-02
5.31837642e-01 2.52804846e-01 2.75380224e-01 7.47936871e-03
-7.93224216e-01 1.84488785e-03 1.32612258e-01 4.09763783e-01
6.78458393e-01 -1.55449128e-02 -1.68396249e-01 4.00122017e-01
-6.74349219e-02 2.13988185e-01 4.52008039e-01 1.14292026e+00
1.15147434e-01 -1.48544765e+00 -1.20600030e-01 7.42956936e-01
-8.82651150e-01 -1.54147416e-01 -7.58566931e-02 3.74080449e-01
-4.53005582e-02 7.87052691e-01 -2.30941296e-01 -3.43425184e-01
6.31904677e-02 5.26068747e-01 4.85092402e-01 -7.44556665e-01
-6.99196160e-01 4.51641738e-01 -6.46249279e-02 -6.20788217e-01
-5.41002214e-01 -8.55916560e-01 -1.13083172e+00 4.05520588e-01
-5.07228486e-02 2.46505231e-01 5.21261394e-01 1.01758862e+00
6.26221001e-01 8.98090243e-01 5.94346344e-01 -5.19769609e-01
-6.52639747e-01 -9.74744856e-01 -6.80218816e-01 4.24837142e-01
6.28141701e-01 -2.69793540e-01 -1.21812187e-01 2.52373338e-01] | [14.248808860778809, 3.3317036628723145] |
38d78163-b6fb-4617-b787-69ec5b35ab83 | uav-images-dataset-for-moving-object | 2103.11460 | null | https://arxiv.org/abs/2103.11460v2 | https://arxiv.org/pdf/2103.11460v2.pdf | UAV Images Dataset for Moving Object Detection from Moving Cameras | This paper presents a new high resolution aerial images dataset in which moving objects are labelled manually. It aims to contribute to the evaluation of the moving object detection methods for moving cameras. The problem of recognizing moving objects from aerial images is one of the important issues in computer vision. The biggest problem in the images taken by UAV is that the background is constantly variable due to camera movement. There are various datasets in the literature in which proposed methods for motion detection are evaluated. Prepared dataset consists of challenging images containing small targets compared to other datasets. Two methods in the literature have been tested for the prepared dataset. In addition, a simpler method compared to these methods has been proposed for moving object object in this paper. | ['Ibrahim Delibasoglu'] | 2021-03-21 | null | null | null | null | ['motion-detection', 'moving-object-detection'] | ['computer-vision', 'computer-vision'] | [ 5.24553657e-01 -6.86944246e-01 1.82011843e-01 6.52113408e-02
2.71120165e-02 -9.15342927e-01 5.42838395e-01 -3.17134380e-01
-6.00848377e-01 7.50790358e-01 -2.39018083e-01 1.22119218e-01
-3.84909451e-01 -5.79941988e-01 -1.08029887e-01 -1.04896092e+00
-1.13105893e-01 -7.65654370e-02 9.91568446e-01 -1.18356667e-01
4.22110140e-01 6.29514873e-01 -1.77931595e+00 2.10229471e-01
1.33445546e-01 3.89131308e-01 7.44014382e-01 1.40811932e+00
3.69756460e-01 1.01787508e+00 -7.65631258e-01 3.37346584e-01
8.39714766e-01 -5.30882776e-01 -9.71310496e-01 6.40477598e-01
7.97850072e-01 -5.15521646e-01 9.10741556e-03 1.23868322e+00
3.42063695e-01 4.38743174e-01 7.78288186e-01 -1.21138752e+00
-3.56602132e-01 3.11963469e-01 -8.38714242e-01 1.26090562e+00
9.48913097e-02 6.49455711e-02 1.84543133e-01 -6.65230036e-01
5.15624166e-01 1.11704028e+00 4.02847975e-01 3.90079141e-01
-6.54433370e-01 -3.12455773e-01 2.23583341e-01 2.82467157e-01
-1.36070526e+00 -3.46645504e-01 6.73738837e-01 -7.39470243e-01
8.10256362e-01 5.14064372e-01 5.05008936e-01 4.65669662e-01
1.86917961e-01 3.37536454e-01 9.16186452e-01 -6.80417895e-01
-6.95685670e-02 2.24726230e-01 5.38402379e-01 4.88991261e-01
5.61675489e-01 1.42144680e-01 2.44653940e-01 1.49587438e-01
5.10712087e-01 1.35579601e-01 -5.13826668e-01 -2.37931043e-01
-1.11385202e+00 5.46300948e-01 5.44415321e-03 6.73471034e-01
-2.40757793e-01 -1.87999681e-01 2.15754494e-01 2.56829932e-02
2.98781302e-02 6.19555041e-02 -6.44229054e-01 7.03276321e-02
-1.20992625e+00 2.21011654e-01 3.77527058e-01 9.38348889e-01
9.55437645e-02 4.21949565e-01 1.81120813e-01 4.56964999e-01
3.32632899e-01 6.93570197e-01 6.94377780e-01 -5.76909900e-01
3.71533781e-01 5.64693749e-01 6.54218316e-01 -1.53972518e+00
-4.83009011e-01 1.69273708e-02 -6.56570435e-01 4.42044586e-01
5.16199768e-01 -8.67769346e-02 -1.16258180e+00 8.12774956e-01
7.15626240e-01 2.70485401e-01 5.12525499e-01 9.70705390e-01
1.24258649e+00 1.10000777e+00 -8.29795226e-02 -4.13505375e-01
1.16423643e+00 -1.08477592e+00 -7.08625555e-01 -1.20009556e-01
2.16723233e-01 -1.11042118e+00 3.14601094e-01 4.73266602e-01
-6.21506274e-01 -7.61266530e-01 -1.07399070e+00 5.97832859e-01
-5.48941255e-01 7.72460401e-01 3.94949019e-01 7.06175387e-01
-7.93879747e-01 4.04127270e-01 -7.73055673e-01 -9.64727938e-01
4.79041487e-02 5.72235405e-01 -3.60004991e-01 1.57012641e-01
-6.58918381e-01 1.00307310e+00 8.44699204e-01 5.14736652e-01
-1.10458648e+00 2.08504438e-01 -4.45900649e-01 -5.56888223e-01
3.33696932e-01 -2.46319324e-01 1.16245770e+00 -1.41401649e+00
-1.30051565e+00 8.55732441e-01 6.05025217e-02 -3.26930314e-01
6.14467382e-01 -1.38819247e-01 -7.24841058e-01 3.08607012e-01
2.94599701e-02 4.88182396e-01 1.13512695e+00 -9.78746414e-01
-1.26393771e+00 -3.05706263e-01 2.31865034e-01 2.15286672e-01
6.06321692e-02 3.47234875e-01 -2.74854316e-03 -6.49801195e-01
-1.67273924e-01 -1.05823183e+00 -1.47752106e-01 -5.36492169e-01
1.47588737e-02 -9.46399663e-03 1.57318580e+00 -5.40671885e-01
1.28066671e+00 -1.74230075e+00 1.82501718e-01 -4.18245703e-01
-2.68636376e-01 8.78712296e-01 -2.80132834e-02 3.65095258e-01
-1.27803937e-01 -1.03817709e-01 -2.08297297e-01 4.83824998e-01
-8.27462554e-01 -1.20047607e-01 -2.10854188e-01 7.70898700e-01
2.47679427e-01 5.22290543e-02 -7.03794658e-01 -6.11548662e-01
6.86865449e-01 1.31368309e-01 2.76779234e-01 -5.20359911e-02
2.95882702e-01 4.43189591e-01 -3.89662474e-01 1.03923416e+00
9.76525664e-01 3.20844829e-01 -3.17329496e-01 -4.18627352e-01
-5.72768331e-01 -6.90824807e-01 -1.62989426e+00 9.08522069e-01
3.28680009e-01 7.86787450e-01 1.25497878e-01 -1.00068748e+00
6.10001802e-01 4.07039046e-01 3.90959680e-01 2.33678207e-01
5.73241651e-01 -2.73573488e-01 2.05698654e-01 -9.07458425e-01
8.79801035e-01 5.90910316e-02 4.68377113e-01 -1.47572592e-01
-4.17808406e-02 -1.40219733e-01 7.49125898e-01 -1.90000623e-01
8.12384427e-01 4.02990967e-01 7.46884108e-01 -3.43016207e-01
1.01592696e+00 7.44353712e-01 4.67056930e-01 5.55375874e-01
-6.07877910e-01 3.32646102e-01 -5.25654197e-01 -7.86198080e-01
-8.80611420e-01 -4.79115874e-01 -2.66739458e-01 5.76435745e-01
4.46985811e-01 8.18722248e-02 -6.27503216e-01 -4.20976549e-01
-4.41017032e-01 2.64916699e-02 -5.53202391e-01 3.69034797e-01
-4.13923264e-01 -1.24320853e+00 3.27906728e-01 4.29679751e-01
8.69372606e-01 -8.45229626e-01 -1.50273073e+00 6.23719729e-02
1.10021658e-01 -1.26714420e+00 -6.28566369e-02 -1.35335863e-01
-1.30470085e+00 -1.47240913e+00 -7.77373552e-01 -8.42396975e-01
8.78874421e-01 1.28743601e+00 6.64557993e-01 4.34677601e-02
-8.27085495e-01 4.83118415e-01 -8.04464757e-01 -7.78563797e-01
-1.05080023e-01 -1.45429581e-01 1.13707371e-01 -4.91701439e-02
4.93195325e-01 1.44959167e-01 -5.75381637e-01 5.91720939e-01
-1.20111728e+00 -3.00523847e-01 5.61385989e-01 4.23856616e-01
2.06675217e-01 8.46386731e-01 -1.72196597e-01 -6.60148740e-01
1.24592809e-02 -3.60256672e-01 -1.24937284e+00 1.83744296e-01
1.71656832e-01 -4.77031559e-01 4.96314406e-01 -7.30718732e-01
-7.73608029e-01 7.28845716e-01 6.17459595e-01 7.65629229e-04
-6.05687976e-01 3.62738296e-02 4.44205366e-02 -4.83767003e-01
5.20888150e-01 1.38539225e-01 -5.32335401e-01 -2.83377003e-02
-8.73371735e-02 7.25504577e-01 5.02716362e-01 2.30696261e-01
9.65507507e-01 5.43702126e-01 2.17639595e-01 -1.58782399e+00
-5.79326391e-01 -1.07460356e+00 -1.14370465e+00 -5.52379251e-01
1.21304524e+00 -7.91881382e-01 -2.35370338e-01 8.06307197e-01
-1.05904460e+00 9.43290219e-02 2.67093956e-01 7.93346107e-01
-1.48970336e-01 3.62091869e-01 -1.79477230e-01 -1.41324604e+00
-3.43851775e-01 -1.08762741e+00 7.79158533e-01 6.11265063e-01
1.32323921e-01 -8.63391101e-01 -1.63132288e-02 2.10759073e-01
3.97836775e-01 6.10967934e-01 1.14201106e-01 -1.73666343e-01
-7.11517751e-01 -5.58178008e-01 1.83570191e-01 3.54977876e-01
6.41203642e-01 7.90056944e-01 -7.99982011e-01 -2.90723324e-01
2.59013534e-01 2.48849571e-01 8.08978736e-01 7.92215824e-01
1.74731940e-01 -2.71735359e-02 -5.53226292e-01 2.28308991e-01
1.87069464e+00 9.68863368e-01 5.72929919e-01 7.36635864e-01
5.45978904e-01 3.67702633e-01 1.24471271e+00 3.95288885e-01
-3.11154902e-01 4.86467540e-01 4.42776531e-01 1.60885796e-01
2.74449348e-01 5.28443813e-01 4.64120716e-01 4.37486231e-01
-6.35575771e-01 -4.53965992e-01 -8.83442521e-01 5.42631745e-01
-1.69051564e+00 -1.46201682e+00 -9.22288299e-01 2.09232116e+00
-1.14451393e-01 -2.16445178e-01 2.95654029e-01 5.31676292e-01
1.00915337e+00 1.66261494e-01 -5.92686683e-02 1.22391872e-01
-9.03331786e-02 -1.36376396e-01 9.22354341e-01 4.26954508e-01
-2.05594897e+00 1.03184724e+00 6.39801264e+00 6.00831956e-02
-1.27779496e+00 -7.21009076e-02 -1.20192222e-01 5.26550040e-02
1.16423142e+00 2.13506557e-02 -1.46265209e+00 5.39974332e-01
7.92477846e-01 -1.59119233e-01 1.32436156e-01 7.87042916e-01
4.38262254e-01 -7.48967528e-01 -4.79742289e-01 8.86777401e-01
4.88555580e-01 -7.47833133e-01 1.30909890e-01 -1.99936807e-01
8.12815189e-01 -4.00411785e-01 -9.35459659e-02 -1.77853972e-01
7.79643655e-02 -7.43212640e-01 5.60616374e-01 5.39297521e-01
8.37370753e-03 -8.50719690e-01 1.11254132e+00 5.15201569e-01
-1.57230783e+00 -3.84156734e-01 -1.04791665e+00 -4.98479933e-01
1.35287359e-01 2.07689102e-03 -8.53643179e-01 6.39755487e-01
8.21470797e-01 7.97205329e-01 -8.73320997e-01 1.38837266e+00
2.90656179e-01 5.23342729e-01 -1.50337219e-01 -3.74517776e-02
5.64629197e-01 -4.00773466e-01 7.15494156e-01 1.03624308e+00
6.10321939e-01 4.30043459e-01 2.18320876e-01 1.92103341e-01
6.60975933e-01 3.04789841e-01 -1.17761064e+00 -3.27995360e-01
6.85148239e-02 1.52808607e+00 -1.35146379e+00 -3.62986684e-01
-3.50894362e-01 1.04615510e+00 -5.03604054e-01 -3.99979483e-03
-7.35757947e-01 -3.61648053e-01 1.47978440e-01 2.20967177e-02
5.90047300e-01 -4.49206859e-01 9.51793730e-01 -8.81494224e-01
-3.62449318e-01 -6.16950929e-01 6.60088539e-01 -7.58265793e-01
-7.14562833e-01 8.22748840e-01 4.94759738e-01 -1.58453822e+00
1.02861375e-01 -1.17728937e+00 -5.13328075e-01 5.58405340e-01
-9.59720314e-01 -1.32488382e+00 -8.91337752e-01 4.78033096e-01
1.29137588e+00 -5.01845360e-01 5.98765850e-01 2.62199134e-01
-7.74527133e-01 -4.01090473e-01 5.78739271e-02 1.77065164e-01
5.02674162e-01 -9.98273373e-01 -1.36935845e-01 1.64498413e+00
1.45684093e-01 3.80887508e-01 8.56325924e-01 -7.31376231e-01
-1.31716859e+00 -1.26975358e+00 9.05533209e-02 -1.03882229e+00
3.59822661e-01 4.05667037e-01 -5.05712152e-01 7.32197285e-01
4.53383118e-01 2.19702199e-01 5.55439770e-01 -1.03993285e+00
7.95688033e-01 7.77020082e-02 -1.11687887e+00 -2.33625751e-02
4.43356961e-01 4.62061971e-01 -7.48873055e-01 2.35769197e-01
1.05838440e-01 -4.42628592e-01 -3.00482124e-01 5.11670291e-01
3.08912456e-01 -7.77958691e-01 8.57067406e-01 -7.49166548e-01
-2.52362192e-02 -1.23077333e+00 -4.35649455e-01 -9.62494731e-01
-3.72538567e-01 -2.81868547e-01 5.31551763e-02 1.46215641e+00
-1.02888793e-01 -1.71665356e-01 5.35331726e-01 1.88797936e-01
3.33912641e-01 1.28360674e-01 -3.66230756e-01 -8.03367376e-01
-4.11728650e-01 2.38685340e-01 -1.41558900e-01 9.25571740e-01
-7.66973853e-01 3.53359729e-01 -7.44623959e-01 8.53835523e-01
5.55340827e-01 3.25998962e-02 9.92826462e-01 -1.41066325e+00
-1.48083558e-02 1.46611780e-02 -1.17921281e+00 -5.03618002e-01
-3.10033977e-01 -3.07641238e-01 2.61365324e-02 -1.69044662e+00
3.19678992e-01 3.64129186e-01 1.96723059e-01 -1.79036409e-01
-2.19892561e-01 4.62439299e-01 2.86671370e-01 1.87701613e-01
-3.06768179e-01 -2.05443665e-01 8.85284126e-01 -4.00933564e-01
-2.27550939e-01 2.81399995e-01 -1.81844011e-02 1.17690396e+00
8.22055519e-01 -5.75699151e-01 -5.65321028e-01 -4.33832616e-01
-3.93917322e-01 -2.30892509e-01 2.84722030e-01 -1.42763555e+00
2.36028999e-01 -3.59563738e-01 7.53498375e-01 -1.21711600e+00
1.52046591e-01 -1.39490032e+00 3.44117224e-01 6.35668576e-01
1.72152430e-01 7.62644827e-01 3.62640113e-01 6.78663969e-01
-2.90298849e-01 -6.71533048e-01 1.25357473e+00 -7.75801241e-01
-1.46552777e+00 -5.72777726e-02 -8.48322153e-01 -1.90047815e-01
1.85754633e+00 -5.81152260e-01 -2.78444558e-01 7.29210824e-02
-5.67865312e-01 -1.00223996e-01 2.97598034e-01 4.10616517e-01
5.21295249e-01 -1.03500640e+00 -5.39173961e-01 -2.96013981e-01
1.47938021e-02 -6.78737387e-02 2.07497537e-01 9.49343622e-01
-1.26208854e+00 6.87231779e-01 -8.30090940e-01 -6.30615592e-01
-2.26723409e+00 9.81041074e-01 4.11901802e-01 2.83496022e-01
-4.57076550e-01 6.96237147e-01 -2.00448662e-01 3.58052075e-01
2.83772610e-02 -6.40011549e-01 -8.77036154e-01 1.12927206e-01
1.06981003e+00 8.29940498e-01 -2.08074078e-01 -1.17330873e+00
-5.48763096e-01 1.09062684e+00 1.47494659e-01 1.01308629e-01
1.27087033e+00 -2.07877040e-01 -9.08945873e-02 7.19258606e-01
4.74453032e-01 -4.20523174e-02 -8.53960514e-01 3.72225046e-01
1.90849930e-01 -1.00254238e+00 6.74417391e-02 -4.88354653e-01
-8.31506014e-01 7.75712311e-01 1.27499509e+00 2.98150003e-01
1.31039560e+00 -6.42357230e-01 1.08196154e-01 5.22912860e-01
6.05455518e-01 -1.08644843e+00 -1.08015738e-01 3.23291898e-01
6.24445736e-01 -1.53113627e+00 4.10950720e-01 -4.38320994e-01
-8.04467261e-01 1.36891496e+00 7.93127596e-01 -2.22585365e-01
3.21509480e-01 4.04456317e-01 3.33161622e-01 8.15456435e-02
-4.59731400e-01 -6.02155685e-01 2.09591731e-01 8.64358902e-01
4.54089969e-01 -2.51211315e-01 -5.05046010e-01 -1.16157584e-01
2.88516402e-01 1.37601450e-01 1.06916833e+00 1.39058065e+00
-8.59284759e-01 -7.69295633e-01 -9.45546091e-01 2.23008335e-01
-8.76621604e-01 2.68135786e-01 -8.11150491e-01 1.09038818e+00
3.10404599e-01 1.29461658e+00 -3.32688242e-01 -1.70599982e-01
2.07067117e-01 -1.69767916e-01 6.05159640e-01 -4.87703174e-01
-2.30084777e-01 -2.44403839e-01 -1.06048495e-01 -8.48372728e-02
-1.28846776e+00 -5.66924632e-01 -9.65928853e-01 2.04514593e-01
-7.32133269e-01 1.69614583e-01 6.14239931e-01 5.86131215e-01
-3.24209720e-01 6.28628612e-01 2.98225999e-01 -1.05147696e+00
-2.77795643e-01 -1.07577217e+00 -7.23332047e-01 2.49635532e-01
4.16114688e-01 -8.71357858e-01 -5.37792265e-01 6.69010460e-01] | [8.770750999450684, -0.9296712279319763] |
d810a6e1-1da3-410f-88c3-b8f63b63ab75 | few-shot-action-recognition-with-implicit | 2010.06215 | null | https://arxiv.org/abs/2010.06215v1 | https://arxiv.org/pdf/2010.06215v1.pdf | Few-shot Action Recognition with Implicit Temporal Alignment and Pair Similarity Optimization | Few-shot learning aims to recognize instances from novel classes with few labeled samples, which has great value in research and application. Although there has been a lot of work in this area recently, most of the existing work is based on image classification tasks. Video-based few-shot action recognition has not been explored well and remains challenging: 1) the differences of implementation details among different papers make a fair comparison difficult; 2) the wide variations and misalignment of temporal sequences make the video-level similarity comparison difficult; 3) the scarcity of labeled data makes the optimization difficult. To solve these problems, this paper presents 1) a specific setting to evaluate the performance of few-shot action recognition algorithms; 2) an implicit sequence-alignment algorithm for better video-level similarity comparison; 3) an advanced loss for few-shot learning to optimize pair similarity with limited data. Specifically, we propose a novel few-shot action recognition framework that uses long short-term memory following 3D convolutional layers for sequence modeling and alignment. Circle loss is introduced to maximize the within-class similarity and minimize the between-class similarity flexibly towards a more definite convergence target. Instead of using random or ambiguous experimental settings, we set a concrete criterion analogous to the standard image-based few-shot learning setting for few-shot action recognition evaluation. Extensive experiments on two datasets demonstrate the effectiveness of our proposed method. | ['Yanning Zhang', 'Peng Wang', 'Qinyi Lv', 'Yajuan Li', 'Congqi Cao'] | 2020-10-13 | null | null | null | null | ['few-shot-action-recognition'] | ['computer-vision'] | [ 5.51633358e-01 -5.26497960e-01 -5.41665554e-01 -4.32153314e-01
-7.24755406e-01 1.25744820e-01 6.41341031e-01 -3.56369317e-01
-4.49949801e-01 4.97083724e-01 1.11064389e-01 2.79025704e-01
-2.33670890e-01 -4.15933460e-01 -6.57884240e-01 -7.91434288e-01
-1.04865897e-02 3.65239978e-02 5.70013046e-01 -8.51786211e-02
5.75453103e-01 2.16480643e-01 -1.75812793e+00 1.63067758e-01
6.63645744e-01 1.21566641e+00 4.13895786e-01 4.74003106e-01
-1.35382101e-01 9.84582543e-01 -5.33753872e-01 -3.68111096e-02
6.13483548e-01 -9.92404640e-01 -6.47207439e-01 4.82220292e-01
5.29971123e-01 -2.73652226e-01 -3.05366516e-01 1.09019709e+00
7.07424164e-01 6.80107117e-01 5.09161115e-01 -1.50080454e+00
-5.43880939e-01 1.94406450e-01 -6.29170179e-01 5.70084095e-01
3.69189888e-01 5.61746776e-01 7.60965526e-01 -7.78316200e-01
5.45395851e-01 9.98194754e-01 6.12725258e-01 5.94712436e-01
-5.73261142e-01 -4.44322407e-01 1.55089557e-01 8.69108140e-01
-1.33555961e+00 -3.71216446e-01 7.92645454e-01 -2.89979398e-01
1.04716909e+00 -4.62615453e-02 7.07323313e-01 1.21112239e+00
6.23990744e-02 8.34618270e-01 8.23701322e-01 -4.61489916e-01
4.74216849e-01 -2.30041906e-01 1.76530436e-01 6.34946764e-01
-1.80596292e-01 1.04565457e-01 -4.71147478e-01 2.00951964e-01
5.37442863e-01 4.60746765e-01 -3.43328595e-01 -5.80672026e-01
-1.15145922e+00 7.20547616e-01 8.21618661e-02 5.96268892e-01
-3.19330156e-01 -2.73822155e-02 7.37720728e-01 4.82385486e-01
3.45982522e-01 1.52439058e-01 -1.36341751e-01 -6.33432090e-01
-9.77085888e-01 1.13457412e-01 5.37034512e-01 9.50286627e-01
6.42888188e-01 4.22152221e-01 -3.72570753e-01 1.08242679e+00
-8.82624611e-02 2.55804628e-01 1.12055337e+00 -8.48194122e-01
5.67840040e-01 3.80788594e-01 -1.95747819e-02 -1.09034944e+00
-3.88669111e-02 -1.11011028e-01 -7.04246104e-01 1.66858450e-01
4.61758107e-01 6.34790286e-02 -8.36081862e-01 1.53153670e+00
1.66446194e-01 9.38589811e-01 -1.95135757e-01 1.29903483e+00
6.03491127e-01 5.45389593e-01 -1.11839864e-02 -5.23144424e-01
9.96211052e-01 -1.31278098e+00 -6.27147317e-01 -2.52855301e-01
7.15038419e-01 -5.81634939e-01 1.26670384e+00 -1.39827346e-02
-9.24236000e-01 -8.24503899e-01 -1.25761318e+00 2.88925111e-01
-3.18892062e-01 -3.26641113e-01 3.98656279e-01 6.24195457e-01
-6.34993732e-01 6.67018354e-01 -8.86468828e-01 -6.29418373e-01
4.73116845e-01 5.61664850e-02 -2.57362306e-01 -3.14935207e-01
-1.18820918e+00 8.04779589e-01 5.64514279e-01 -8.44399706e-02
-7.03384101e-01 -4.95206356e-01 -1.04701269e+00 4.13055941e-02
8.11217546e-01 -5.07947266e-01 1.07305408e+00 -1.16132164e+00
-1.61436474e+00 8.13413680e-01 1.36413081e-02 -5.20253956e-01
5.58613181e-01 -1.66005909e-01 -4.06773955e-01 1.99985445e-01
6.52859062e-02 3.51111263e-01 9.79623616e-01 -6.07476473e-01
-6.45053446e-01 -4.05477077e-01 6.00476004e-02 4.20637101e-01
-4.64254856e-01 1.36566088e-01 -4.79213655e-01 -8.78024220e-01
-5.57553805e-02 -7.04813898e-01 -2.37827927e-01 3.99472602e-02
5.92711791e-02 -1.11251868e-01 8.84952307e-01 -1.83103636e-01
1.15023482e+00 -2.13521147e+00 8.23374316e-02 -3.55207145e-01
-1.96132064e-01 6.44058108e-01 -3.29592854e-01 4.00511175e-01
-1.60667971e-01 -2.73492008e-01 -3.06695908e-01 -7.40740672e-02
-4.33006883e-02 1.48126185e-01 -2.15400174e-01 6.33762836e-01
1.29376024e-01 8.34632456e-01 -9.73348558e-01 -5.84308743e-01
6.84848964e-01 2.05567256e-01 -3.46798688e-01 3.33504319e-01
4.42600213e-02 1.42520398e-01 -4.36203122e-01 9.05584395e-01
4.24701601e-01 -2.16940090e-01 -2.54787266e-01 -1.07285321e-01
5.85464854e-03 -2.30329737e-01 -1.45810151e+00 1.84199238e+00
-2.71137953e-01 4.44704592e-01 -3.82494181e-01 -1.68399429e+00
8.75441432e-01 1.82707518e-01 7.51300097e-01 -8.06232214e-01
2.40064517e-01 -4.20166105e-02 -1.31039500e-01 -7.94639170e-01
2.71293968e-01 -5.24401069e-01 1.72798559e-01 4.71394897e-01
1.49912119e-01 7.60923102e-02 3.56437176e-01 -1.35079846e-01
1.24490166e+00 2.75622755e-01 6.96375728e-01 2.70284534e-01
5.32922447e-01 -1.44626200e-01 8.99285734e-01 7.48100579e-01
-8.71622324e-01 9.13520873e-01 2.06674822e-02 -5.06752074e-01
-8.83588612e-01 -7.75497079e-01 1.29022509e-01 1.05617607e+00
5.09016335e-01 -2.84381717e-01 -5.24748325e-01 -7.24288821e-01
-2.06354886e-01 5.66434801e-01 -5.28750360e-01 -3.00013751e-01
-6.11842573e-01 -6.79146826e-01 3.39459300e-01 6.78035080e-01
8.08633447e-01 -1.28445804e+00 -8.13160837e-01 3.26840878e-02
-1.23448536e-01 -1.20916951e+00 -6.70735538e-01 -1.07672542e-01
-9.19589102e-01 -1.22167647e+00 -1.07115245e+00 -8.26331794e-01
3.62441629e-01 7.70679057e-01 6.31659508e-01 -1.63135692e-01
-6.13648534e-01 5.57044744e-01 -7.48273551e-01 -1.59153298e-01
3.78175527e-02 -2.85940170e-01 1.86072752e-01 3.00449967e-01
6.83524370e-01 -5.67140043e-01 -5.47893524e-01 5.95761776e-01
-9.70866084e-01 -2.45702088e-01 6.08848095e-01 1.05761659e+00
6.07117116e-01 -5.46617433e-02 5.23611784e-01 -3.98338705e-01
3.69155824e-01 -4.62362140e-01 -1.96808264e-01 3.87975514e-01
-5.24102569e-01 -2.28779227e-01 8.22401524e-01 -6.60363674e-01
-8.99249911e-01 -3.08929756e-03 9.70272049e-02 -9.41379666e-01
-2.42968962e-01 2.71119922e-01 -1.66012183e-01 -1.33173585e-01
4.66807872e-01 6.12615883e-01 1.95543677e-01 -3.11791033e-01
2.79033512e-01 7.35928774e-01 2.97743976e-01 -2.20676780e-01
5.66629469e-01 4.48907644e-01 -2.02493310e-01 -1.11739540e+00
-8.03505301e-01 -8.33176434e-01 -7.98940718e-01 -3.15666050e-01
8.49019349e-01 -7.04636693e-01 -3.83958817e-01 6.00020409e-01
-6.64891243e-01 -2.86932081e-01 -4.13750738e-01 8.67942631e-01
-1.03610039e+00 8.49666953e-01 -3.51800174e-01 -6.56149745e-01
-2.20100179e-01 -1.14403713e+00 9.28271294e-01 1.77681565e-01
-1.16535470e-01 -7.60863483e-01 2.40512162e-01 5.35972774e-01
2.08212182e-01 1.55411407e-01 4.89115894e-01 -6.14547968e-01
-5.17244160e-01 -2.84618974e-01 -1.56349316e-01 4.57696468e-01
2.98968911e-01 -1.99960008e-01 -6.10144615e-01 -2.95525223e-01
2.43840650e-01 -7.36574531e-01 9.00403917e-01 4.79195565e-01
1.17967272e+00 6.45696893e-02 -1.42735645e-01 7.15797246e-01
1.35847843e+00 4.25732225e-01 8.67831647e-01 3.52636188e-01
7.15868056e-01 2.89106697e-01 1.12711060e+00 6.85655415e-01
-6.98243305e-02 9.09570336e-01 2.82236874e-01 4.29167330e-01
-7.66175836e-02 2.17447616e-02 3.98515105e-01 8.56696606e-01
-3.20304483e-01 -3.46890166e-02 -5.98706722e-01 4.44071412e-01
-2.15243936e+00 -1.76547360e+00 4.79637533e-01 2.26517606e+00
4.25926268e-01 1.34017274e-01 2.65600562e-01 5.61467484e-02
9.46457982e-01 8.41705561e-01 -6.79507077e-01 -1.45661220e-01
-5.92255928e-02 1.32727195e-02 1.30827636e-01 4.19818126e-02
-1.24035764e+00 7.87416339e-01 5.56543589e+00 1.22494090e+00
-1.13095605e+00 4.78192903e-02 5.13507485e-01 -3.52613121e-01
3.86310637e-01 -1.12261556e-01 -7.88922429e-01 6.85295403e-01
5.54137766e-01 -2.11974666e-01 1.01757780e-01 1.01029325e+00
2.97953486e-01 5.70235103e-02 -1.04911935e+00 1.57704282e+00
6.01864994e-01 -1.22096670e+00 9.01961699e-02 -3.35189432e-01
7.17465818e-01 -1.07396714e-01 -1.25125065e-01 5.32270253e-01
-3.11772168e-01 -7.41955578e-01 4.00427461e-01 6.07299030e-01
5.92200935e-01 -7.16688633e-01 4.88232315e-01 4.68679577e-01
-1.28439009e+00 -4.32335198e-01 -6.25794411e-01 -3.98715496e-01
3.33749026e-01 1.90617427e-01 -4.39789027e-01 4.71213758e-01
5.38723767e-01 1.21797943e+00 -4.25157368e-01 1.27629638e+00
1.17630780e-01 4.81263459e-01 3.35136801e-02 -2.12687582e-01
5.99999547e-01 -3.65603238e-01 6.61253989e-01 1.00827706e+00
3.25039774e-01 2.77376473e-01 4.41939145e-01 4.78164554e-01
1.81628153e-01 3.32663685e-01 -7.75312841e-01 -1.54314101e-01
2.31738195e-01 9.16221201e-01 -8.33897293e-01 -3.72801274e-01
-7.61804521e-01 1.18266046e+00 2.19973162e-01 1.52022928e-01
-9.57872033e-01 -5.50760627e-01 6.30181730e-01 3.71200070e-02
5.77844024e-01 -3.32335494e-02 -7.34821483e-02 -1.37933528e+00
1.71432123e-01 -9.54502523e-01 6.29972756e-01 -5.92189908e-01
-1.29641902e+00 2.19139516e-01 7.46381059e-02 -1.80668855e+00
-4.56467867e-01 -3.01658481e-01 -8.89919519e-01 3.76522511e-01
-1.21462357e+00 -7.92935848e-01 -4.17882740e-01 6.04747355e-01
1.25939322e+00 -5.25059164e-01 5.48057735e-01 3.47284675e-01
-7.15148330e-01 6.70253813e-01 2.10818857e-01 8.06761906e-02
8.45114887e-01 -8.07981730e-01 2.26490095e-01 8.88956308e-01
2.22173855e-01 3.15710038e-01 5.59708297e-01 -5.61267734e-01
-1.43951643e+00 -9.46774960e-01 6.08545899e-01 -2.33247355e-01
6.42702222e-01 7.19841719e-02 -9.08418179e-01 4.84358966e-01
-2.44812425e-02 3.08001041e-01 7.62043178e-01 -2.60120958e-01
-1.77956715e-01 -2.07519591e-01 -9.01940286e-01 7.05354631e-01
1.37698984e+00 -3.56736362e-01 -7.70280361e-01 2.90721297e-01
4.22837675e-01 7.49300234e-04 -5.13969004e-01 4.76030201e-01
6.87106788e-01 -1.12267780e+00 9.62848902e-01 -8.27089489e-01
4.74304646e-01 -3.00587177e-01 -4.32364643e-01 -1.17566526e+00
-2.35552892e-01 -4.59719509e-01 -1.98904008e-01 9.12754297e-01
-2.93607652e-01 -3.27929497e-01 8.23675454e-01 4.51553136e-01
-2.52260596e-01 -9.39481378e-01 -8.67623687e-01 -1.21590257e+00
-2.92744547e-01 -4.76443768e-01 2.42184937e-01 9.82528448e-01
1.12147443e-01 2.15980768e-01 -7.65830338e-01 -2.71827966e-01
6.69546247e-01 4.20830309e-01 8.01294684e-01 -7.47143686e-01
-4.66371059e-01 -5.44283926e-01 -1.07422686e+00 -1.10476100e+00
2.52896249e-01 -5.70444524e-01 6.11021854e-02 -1.23900008e+00
4.34412450e-01 4.13834024e-03 -4.72202301e-01 3.17821711e-01
-3.06621164e-01 3.38714808e-01 3.81448597e-01 2.21383974e-01
-1.05573821e+00 9.69436944e-01 1.16944957e+00 -3.60075146e-01
-1.00882597e-01 7.70725012e-02 -2.45585084e-01 7.28103936e-01
5.43944359e-01 -2.07555369e-01 -7.70052016e-01 -2.46827975e-02
-4.29182410e-01 1.98121831e-01 3.01308483e-01 -1.33820295e+00
2.46027425e-01 -4.55510050e-01 1.78682551e-01 -4.11248565e-01
4.28670555e-01 -7.53142297e-01 -2.24794120e-01 5.02406001e-01
-3.27157140e-01 -3.65006298e-01 -2.87517011e-01 8.31925988e-01
-4.54275608e-01 -5.08205950e-01 9.32758451e-01 -2.95513332e-01
-1.54499602e+00 5.85228920e-01 -2.07567900e-01 3.49466056e-01
1.41214812e+00 -8.75542104e-01 6.81886822e-02 -4.09763068e-01
-5.92849135e-01 1.22127399e-01 5.40545344e-01 6.08728886e-01
8.64238441e-01 -1.59467947e+00 -5.20112514e-01 5.97817823e-02
5.06851196e-01 -5.80762804e-01 6.14199817e-01 8.71744454e-01
-2.33206019e-01 2.59774357e-01 -5.82487643e-01 -5.33469260e-01
-1.42513216e+00 7.50416458e-01 2.90994167e-01 4.75788713e-02
-8.66837144e-01 6.93248093e-01 1.14978679e-01 -1.70033693e-01
4.51049358e-01 5.95743805e-02 -2.46807098e-01 1.72971845e-01
9.14781630e-01 6.86145484e-01 -3.21139187e-01 -7.68265843e-01
-3.89222294e-01 6.98794365e-01 -7.04672411e-02 1.75240740e-01
1.28269207e+00 -2.09067494e-01 5.24525523e-01 8.20133090e-01
1.35603178e+00 -9.07067955e-01 -1.47854829e+00 -3.27559531e-01
-1.00482747e-01 -9.78276193e-01 -1.72440663e-01 -2.90450931e-01
-9.56602573e-01 1.02541828e+00 7.42421269e-01 -1.53307363e-01
1.05869484e+00 -4.29159284e-01 1.10999227e+00 5.17718196e-01
5.26861012e-01 -1.46438050e+00 5.52666008e-01 5.78908861e-01
5.17748713e-01 -1.68069267e+00 -3.17287371e-02 -9.69116762e-02
-7.78856575e-01 1.05736852e+00 7.63356805e-01 -1.01625882e-01
5.96632600e-01 -5.90475611e-02 1.71073154e-02 -1.30725175e-01
-5.63861072e-01 -3.31547678e-01 1.90125436e-01 6.61613107e-01
1.94313124e-01 -3.96987766e-01 -5.48733056e-01 5.16530812e-01
4.21588749e-01 4.24046874e-01 2.75223970e-01 1.10321820e+00
-6.96301639e-01 -7.79137671e-01 -1.30327463e-01 4.96676087e-01
-2.84522921e-01 1.44303888e-01 -1.21379882e-01 6.19629264e-01
-4.19475138e-02 8.23644161e-01 -9.24411044e-03 -4.28449541e-01
2.96931297e-01 1.66584089e-01 6.16220772e-01 -7.00621426e-01
-1.01311080e-01 -1.55793399e-01 -2.34863967e-01 -8.30211341e-01
-7.76421666e-01 -5.96065462e-01 -1.07448196e+00 -6.37471899e-02
-3.34545612e-01 -7.41861910e-02 2.87741959e-01 1.28718948e+00
2.64788061e-01 2.00343519e-01 8.64831984e-01 -9.51895475e-01
-1.00775146e+00 -9.47098851e-01 -9.43037987e-01 7.82796800e-01
-1.99050978e-02 -1.00105608e+00 -3.93971860e-01 1.23763382e-01] | [8.469121932983398, 0.779491662979126] |
5827ef00-2c48-4cc2-9995-47eb7fd82857 | tail-batch-sampling-approximating-global | 2210.12874 | null | https://arxiv.org/abs/2210.12874v4 | https://arxiv.org/pdf/2210.12874v4.pdf | Global Contrastive Batch Sampling via Optimization on Sample Permutations | Contrastive Learning has recently achieved state-of-the-art performance in a wide range of tasks. Many contrastive learning approaches use mined hard negatives to make batches more informative during training but these approaches are inefficient as they increase epoch length proportional to the number of mined negatives and require frequent updates of nearest neighbor indices or mining from recent batches. In this work, we provide an alternative to hard negative mining, Global Contrastive Batch Sampling (GCBS), an efficient approximation to the batch assignment problem that upper bounds the gap between the global and training losses, $\mathcal{L}^{Global} - \mathcal{L}^{Train}$, in contrastive learning settings. Through experimentation we find GCBS improves state-of-the-art performance in sentence embedding and code-search tasks. Additionally, GCBS is easy to implement as it requires only a few additional lines of code, does not maintain external data structures such as nearest neighbor indices, is more computationally efficient than the most minimal hard negative mining approaches, and makes no changes to the model being trained. | ['Chenguang Zhu', 'ZiYi Yang', 'Vin Sachidananda'] | 2022-10-23 | null | null | null | null | ['code-search', 'code-search'] | ['computer-code', 'computer-vision'] | [-3.77056897e-02 1.09928705e-01 -5.56014359e-01 -6.52551770e-01
-1.14266050e+00 -3.96877795e-01 4.13151801e-01 6.29192650e-01
-9.94461954e-01 7.85588443e-01 -4.32884514e-01 -7.21031964e-01
-2.13525921e-01 -6.80532694e-01 -7.69171119e-01 -3.70569199e-01
-3.73876154e-01 5.74218452e-01 5.00259638e-01 -2.68265992e-01
3.24141622e-01 1.74767494e-01 -1.67797172e+00 4.27073836e-01
6.76597059e-01 1.29698777e+00 1.43781990e-01 5.88643909e-01
-2.54407972e-01 6.94295168e-01 -5.31572104e-01 -7.23686635e-01
4.04451728e-01 -2.41008386e-01 -8.45561743e-01 -5.75845838e-01
6.53608859e-01 -1.44661486e-01 1.35649741e-02 9.77933764e-01
5.77176929e-01 2.34974638e-01 4.19401556e-01 -1.12945318e+00
-3.97404075e-01 6.16478741e-01 -7.87417769e-01 8.39619458e-01
-1.42380614e-02 -1.26295522e-01 1.21400154e+00 -1.10719192e+00
2.55814761e-01 8.78729999e-01 1.02487183e+00 7.15684950e-01
-1.23584235e+00 -7.75850177e-01 3.98216188e-01 1.98933184e-01
-1.14939713e+00 -5.17312765e-01 4.89770770e-01 -1.62042424e-01
1.66893852e+00 6.28431380e-01 7.77791202e-01 5.53920209e-01
-1.32613361e-01 9.58002090e-01 8.03442299e-01 -9.31657910e-01
3.28135937e-01 5.29064596e-01 1.84381679e-01 9.03243363e-01
2.12719902e-01 -6.97401613e-02 -6.46674752e-01 -4.38809514e-01
1.90199718e-01 1.09938890e-01 -3.08385864e-02 -1.65107712e-01
-4.38912362e-01 1.18542397e+00 -2.29635853e-02 1.33361727e-01
6.17238805e-02 2.02584848e-01 7.66750991e-01 7.48742938e-01
1.10370970e+00 4.11143482e-01 -8.96974683e-01 -4.94345069e-01
-1.20528865e+00 8.60809460e-02 9.01729286e-01 7.63901651e-01
8.42693508e-01 -1.44108146e-01 -2.10337583e-02 1.22591496e+00
2.70678638e-03 1.04601368e-01 5.96388578e-01 -6.50065184e-01
8.50268126e-01 4.28842485e-01 -1.50717780e-01 -6.80471420e-01
-2.08861336e-01 -5.82521439e-01 -5.80715001e-01 1.23249479e-01
2.10129723e-01 -1.42948866e-01 -5.56687236e-01 1.60637903e+00
2.63960034e-01 -2.14924291e-01 -3.15998822e-01 2.53397346e-01
5.23183465e-01 5.47210634e-01 -6.64186627e-02 -3.65830779e-01
8.44895184e-01 -1.14163637e+00 -3.09077799e-01 -6.80232584e-01
1.09216058e+00 -7.12711096e-01 1.42961717e+00 6.44175768e-01
-1.44629264e+00 -3.32283169e-01 -1.10254312e+00 -5.09454869e-02
-4.46894646e-01 -1.11328401e-01 9.28815901e-01 8.97668123e-01
-1.20944703e+00 9.24405158e-01 -8.41528952e-01 1.39651611e-01
6.01805151e-01 8.16421986e-01 -7.37533346e-02 -1.16790701e-02
-8.59752297e-01 8.51180196e-01 3.04755956e-01 -2.15941116e-01
-3.37723374e-01 -8.88080418e-01 -9.75783527e-01 5.35772778e-02
3.02364767e-01 -1.61232680e-01 1.34454727e+00 -8.60005140e-01
-1.05464089e+00 1.10954809e+00 -3.33137333e-01 -6.34278774e-01
4.90579575e-01 -3.87391150e-01 -1.80199906e-01 -1.33854464e-01
-8.88120234e-02 6.90122306e-01 6.22956812e-01 -6.31154895e-01
-6.82564080e-01 -4.36260104e-01 -1.75702870e-01 1.11103103e-01
-1.14625955e+00 8.40009227e-02 -5.94419062e-01 -5.70393145e-01
4.10537124e-02 -7.35868931e-01 -1.94033802e-01 1.51480719e-01
1.03337988e-02 -7.62374759e-01 6.95360780e-01 -2.88897336e-01
1.57415938e+00 -2.08992553e+00 -4.22846377e-01 4.07827884e-01
8.99696499e-02 4.79128331e-01 1.82495490e-02 3.25659990e-01
-4.61356975e-02 1.57796502e-01 -2.11371958e-01 -7.21733689e-01
4.42142524e-02 2.92576104e-01 -2.17132643e-01 4.73169565e-01
-1.84675902e-02 5.86113513e-01 -9.02995408e-01 -6.11930788e-01
-9.16420370e-02 1.17120303e-01 -8.02529216e-01 -1.97475590e-03
-2.31401354e-01 -5.91894865e-01 -1.79351494e-02 3.70761961e-01
4.80623215e-01 -3.14073741e-01 7.88856968e-02 3.69799018e-01
-5.49168848e-02 7.75368452e-01 -1.15449119e+00 1.26439226e+00
-5.11608601e-01 6.82534039e-01 2.75730770e-02 -1.35936594e+00
8.99479747e-01 5.84714720e-03 2.42752627e-01 -9.58766043e-01
-3.03291976e-01 2.96997011e-01 -3.95908087e-01 -3.10499221e-01
4.88819629e-01 -1.44045195e-02 1.60984978e-01 7.29507804e-01
-1.50883589e-02 -1.31712273e-01 3.76406282e-01 1.29748031e-03
1.22614241e+00 -1.87360987e-01 2.14248791e-01 -2.68138796e-01
4.08648700e-01 -1.86550245e-01 6.85746551e-01 9.29291844e-01
-1.16488844e-01 2.29875535e-01 5.27908742e-01 -7.07809091e-01
-9.07983363e-01 -9.37663496e-01 -8.78783241e-02 1.58106709e+00
-1.98234335e-01 -6.61329985e-01 -6.37815058e-01 -9.86191809e-01
-9.02806446e-02 6.00423753e-01 -5.21957397e-01 -2.46389881e-01
-8.43934119e-01 -1.02105558e+00 4.15957600e-01 6.64047658e-01
2.39237100e-01 -9.62301075e-01 -4.41359967e-01 9.56438109e-02
4.28602621e-02 -5.14981270e-01 -5.78090250e-01 7.89995015e-01
-1.11957276e+00 -8.78794193e-01 -2.91892320e-01 -1.07169020e+00
7.23231733e-01 7.16508403e-02 1.61299062e+00 3.26211601e-01
-4.72599834e-01 8.26129690e-02 -2.47626022e-01 -8.99835408e-01
-1.21281095e-01 2.06321985e-01 -2.06361581e-02 -5.90559483e-01
7.84361780e-01 -5.51961541e-01 -6.05338514e-01 -6.41735494e-02
-6.83679819e-01 -3.51639509e-01 4.18567270e-01 1.07202244e+00
8.93472552e-01 -8.40525329e-02 6.88655674e-01 -1.26703072e+00
5.96391141e-01 -4.43981141e-01 -6.43639505e-01 1.70038477e-01
-1.11460769e+00 9.43152755e-02 7.10180223e-01 -5.28416097e-01
-6.00912035e-01 -1.45046771e-01 -3.48974437e-01 -3.66052836e-01
3.37767601e-01 4.00317818e-01 4.48164463e-01 -5.41802570e-02
9.94922996e-01 2.32780993e-01 -3.29994336e-02 -5.45843005e-01
-1.13897189e-01 6.60427630e-01 1.91604972e-01 -4.01691258e-01
5.94505429e-01 2.93303818e-01 -3.16384017e-01 -6.82345569e-01
-1.12158835e+00 -6.94595695e-01 -2.42896989e-01 7.72233307e-02
2.00931132e-01 -7.54838407e-01 -6.65343463e-01 1.56930551e-01
-7.42465436e-01 -6.91766024e-01 -5.87068141e-01 4.26112145e-01
-4.19600904e-01 4.23792273e-01 -7.66941309e-01 -1.02127290e+00
-9.35528874e-01 -5.57723224e-01 6.67823851e-01 -8.39979947e-02
-3.80795628e-01 -1.06080246e+00 1.60156898e-02 2.34322354e-01
3.26859623e-01 -2.36245155e-01 1.17935443e+00 -8.38337839e-01
-1.29958421e-01 -5.16639352e-01 1.71373233e-01 7.95794547e-01
-1.42573833e-01 -2.69784391e-01 -9.34473991e-01 -6.58058584e-01
9.59411785e-02 -6.52821124e-01 1.02745605e+00 1.50627822e-01
1.67739415e+00 -6.08945787e-01 -2.89969325e-01 3.65757078e-01
1.41869593e+00 3.30849253e-02 5.08885562e-01 5.07283807e-01
2.25450560e-01 4.04848188e-01 9.22084510e-01 7.85719931e-01
1.13964871e-01 3.90998900e-01 4.38924402e-01 -3.33407521e-01
2.71728784e-01 4.88942750e-02 4.22029644e-01 8.65603030e-01
3.78624231e-01 -4.46047774e-03 -7.53516078e-01 8.67239654e-01
-1.59640992e+00 -1.17812347e+00 7.65404627e-02 2.44579697e+00
1.52407682e+00 6.85156226e-01 2.31227800e-01 5.04279315e-01
4.55839604e-01 2.31629964e-02 -4.45756853e-01 -8.61344695e-01
1.19161494e-01 7.69458711e-01 4.78336215e-01 3.99286300e-01
-1.05219257e+00 7.47578979e-01 6.12543869e+00 1.14128149e+00
-1.02892017e+00 3.38676363e-01 9.67557490e-01 -7.73923099e-01
-2.94306457e-01 -1.65821448e-01 -1.13162601e+00 4.03479755e-01
8.67467642e-01 3.43060583e-01 4.03047711e-01 1.28328168e+00
-1.07459970e-01 -1.96549341e-01 -1.45737088e+00 9.72234070e-01
2.08290547e-01 -1.54813671e+00 -4.38259661e-01 -8.52764398e-02
8.01580191e-01 5.14969945e-01 2.57516325e-01 6.30685568e-01
2.79044449e-01 -8.56702566e-01 7.16366947e-01 7.20995739e-02
7.99712181e-01 -9.85670745e-01 6.54827893e-01 7.42293000e-01
-8.26844931e-01 -3.72323006e-01 -5.63974321e-01 -1.75675094e-01
-3.00083309e-01 9.73153412e-01 -7.63763845e-01 -1.11028597e-01
1.31638432e+00 2.34452382e-01 -6.86938822e-01 9.12957549e-01
1.82211280e-01 8.12188864e-01 -6.37195289e-01 -4.31624055e-01
2.49355093e-01 1.63088292e-01 5.64784324e-03 1.42848718e+00
8.27602074e-02 -2.79002935e-01 2.16311172e-01 4.96190995e-01
-2.51197398e-01 3.55993569e-01 -4.19676542e-01 3.02045345e-01
7.00337529e-01 9.16938424e-01 -5.35807252e-01 -4.17417318e-01
-5.75031817e-01 6.36681795e-01 8.94807637e-01 -1.56681314e-01
-7.38332272e-01 -8.50521207e-01 4.00983810e-01 3.00116420e-01
6.44878983e-01 8.82286951e-02 -3.86252224e-01 -7.68656075e-01
5.74077845e-01 -8.33896101e-01 7.14207649e-01 -1.20841704e-01
-1.36001432e+00 5.10717869e-01 -6.77892268e-02 -1.03752291e+00
-4.54108804e-01 -4.89597529e-01 -6.64564252e-01 8.37060332e-01
-1.51250970e+00 -3.84105176e-01 -3.21021536e-03 5.44190645e-01
5.41257262e-01 -8.96596387e-02 1.06367528e+00 4.69834358e-01
-4.72283244e-01 1.34153163e+00 3.31756890e-01 2.28535179e-02
5.58778405e-01 -1.30401230e+00 3.91394675e-01 5.32087743e-01
2.28483588e-01 5.54280937e-01 5.70779443e-01 -4.16098297e-01
-1.01992667e+00 -1.00862551e+00 1.20300293e+00 -2.60716408e-01
3.57272774e-01 -5.39669931e-01 -7.15472102e-01 4.85695392e-01
-1.32729420e-02 4.95556027e-01 1.10846221e+00 4.55560356e-01
-4.72058922e-01 -5.58941424e-01 -1.31776106e+00 3.55920494e-01
9.33145165e-01 -5.92837155e-01 -1.76554665e-01 7.19514787e-01
4.43620086e-01 -2.33846948e-01 -8.63661408e-01 3.65889043e-01
2.94399589e-01 -1.04296446e+00 8.54362667e-01 -5.07696569e-01
2.02271402e-01 2.79350668e-01 7.02610379e-03 -8.41171324e-01
-1.35018537e-02 -7.45911956e-01 -5.34653246e-01 1.07616639e+00
9.58778679e-01 -6.72855794e-01 1.35412943e+00 5.27050018e-01
-2.15767995e-01 -1.55421925e+00 -1.01497185e+00 -9.71280932e-01
1.80777654e-01 -6.42193735e-01 1.93553671e-01 9.94145393e-01
1.42614275e-01 1.40239924e-01 -1.25254616e-01 -5.20779848e-01
4.20629978e-01 -1.02377146e-01 5.12287676e-01 -1.18119693e+00
-7.95224667e-01 -3.08581918e-01 -6.70853779e-02 -1.15649843e+00
6.07990399e-02 -1.12160301e+00 -1.49139296e-02 -1.29860044e+00
9.00774151e-02 -9.74878311e-01 -5.01288533e-01 9.15663660e-01
-9.15945619e-02 5.40989935e-01 -1.53051853e-01 1.00654274e-01
-8.22720289e-01 3.62613142e-01 6.10149145e-01 -2.55440176e-01
-3.53164345e-01 2.74212718e-01 -6.31369114e-01 6.71158671e-01
8.41687143e-01 -8.36846292e-01 -6.27068758e-01 -4.66441095e-01
6.60181284e-01 -1.63511455e-01 -6.12843595e-02 -7.57753670e-01
3.08115423e-01 1.24335438e-01 5.28860748e-01 -7.95992196e-01
5.08276284e-01 -3.49907875e-01 -6.47271633e-01 5.13622046e-01
-5.39911270e-01 5.01064003e-01 4.05595690e-01 4.66605216e-01
-4.23912928e-02 -7.88510263e-01 8.21073771e-01 -2.73351580e-01
-2.92252839e-01 2.02545956e-01 -3.32657516e-01 6.10477269e-01
8.96610796e-01 -5.43692112e-01 -6.78103194e-02 -4.04779576e-02
-3.22224379e-01 1.77098751e-01 1.12705685e-01 1.50747284e-01
6.21434212e-01 -1.00901318e+00 -3.72105926e-01 2.69840032e-01
2.69742496e-03 2.81667411e-01 -1.30863832e-02 9.17469859e-01
-4.56272840e-01 2.50480831e-01 4.59695995e-01 -4.49806154e-01
-1.54590249e+00 3.66273195e-01 1.46062016e-01 -5.69213986e-01
-5.23896992e-01 1.58251238e+00 -3.84780884e-01 -5.76848030e-01
7.29516625e-01 -4.88081947e-02 5.41513506e-03 1.18472785e-01
7.07461774e-01 5.53490877e-01 5.15724003e-01 2.10809514e-01
-2.59199202e-01 4.81433840e-03 -5.98989189e-01 -6.00346886e-02
1.66987634e+00 2.43573546e-01 -3.40232372e-01 4.98642892e-01
1.47786188e+00 -2.23348215e-02 -1.10438097e+00 -3.49623501e-01
3.00357074e-01 -3.51505935e-01 7.16384724e-02 -7.02598929e-01
-9.49033737e-01 8.80277812e-01 6.61629736e-01 3.52026045e-01
1.14109004e+00 1.45248864e-02 9.11581397e-01 8.19308817e-01
1.52900279e-01 -1.78384769e+00 2.48542219e-01 5.34998536e-01
3.71555299e-01 -1.19384766e+00 2.54119813e-01 -2.41086915e-01
-1.64546818e-01 9.29987073e-01 7.79956639e-01 -2.13371161e-02
8.14474583e-01 5.31614184e-01 -9.26698223e-02 -3.29892933e-02
-1.12815881e+00 2.33527839e-01 1.00497894e-01 3.83670449e-01
6.20358050e-01 -2.29252711e-01 -3.23317856e-01 4.34532672e-01
-3.72116059e-01 -2.66988039e-01 1.04150578e-01 1.23504305e+00
-5.53114533e-01 -1.30295169e+00 1.99452281e-01 1.07147300e+00
-8.78735900e-01 -5.90634882e-01 -1.73961937e-01 8.17319095e-01
1.02649264e-01 9.48051393e-01 3.15966278e-01 -2.61479080e-01
2.20967874e-01 1.49571925e-01 6.50059819e-01 -9.27063644e-01
-8.46211433e-01 -3.28263938e-01 2.23365828e-01 -4.00035650e-01
-3.16629931e-02 -6.53076828e-01 -1.32066607e+00 -4.54007387e-01
-8.30765247e-01 4.97517288e-01 6.40356004e-01 9.52318132e-01
2.96267599e-01 -8.78195688e-02 6.20364130e-01 -5.57067752e-01
-1.15047109e+00 -8.57607484e-01 -3.43108803e-01 1.85230508e-01
1.93783060e-01 -1.73326671e-01 -5.45660496e-01 -3.18771631e-01] | [8.700562477111816, 3.514206647872925] |
1f374b51-d18a-4a78-8f41-94e7c14e6a83 | intrinsic-normalization-and-extrinsic | 1609.05104 | null | http://arxiv.org/abs/1609.05104v2 | http://arxiv.org/pdf/1609.05104v2.pdf | Intrinsic normalization and extrinsic denormalization of formant data of vowels | Using a known speaker-intrinsic normalization procedure, formant data are
scaled by the reciprocal of the geometric mean of the first three formant
frequencies. This reduces the influence of the talker but results in a
distorted vowel space. The proposed speaker-extrinsic procedure re-scales the
normalized values by the mean formant values of vowels. When tested on the
formant data of vowels published by Peterson and Barney, the combined approach
leads to well separated clusters by reducing the spread due to talkers. The
proposed procedure performs better than two top-ranked normalization procedures
based on the accuracy of vowel classification as the objective measure. | ['A. G. Ramakrishnan', 'T. V. Ananthapadmanabha'] | 2016-09-16 | null | null | null | null | ['vowel-classification'] | ['audio'] | [ 6.25394434e-02 -9.95359793e-02 5.22410274e-02 -3.02427799e-01
-7.00880826e-01 -6.71609282e-01 5.39110720e-01 4.14114505e-01
-7.86625385e-01 4.58687484e-01 6.74546778e-01 -9.31793824e-02
-5.22324294e-02 -5.42584062e-01 -3.16280097e-01 -8.69238377e-01
2.77177423e-01 4.26125266e-02 1.20908238e-01 -2.74462789e-01
3.56137544e-01 4.40852880e-01 -1.67967796e+00 -1.14335850e-01
6.60964489e-01 4.90086854e-01 -1.44197196e-01 7.50215590e-01
-3.75232697e-01 1.26596302e-01 -1.08508205e+00 -9.16575566e-02
1.74573332e-01 -6.84942245e-01 -3.22470665e-01 6.34132028e-02
2.87735045e-01 1.20454937e-01 2.11300716e-01 1.13635159e+00
4.40760314e-01 3.66949618e-01 7.06126153e-01 -6.22952461e-01
-3.39483976e-01 8.61599147e-01 -1.94612220e-01 1.55470699e-01
3.17856073e-01 -4.41805452e-01 8.62988174e-01 -7.50278413e-01
4.27402943e-01 1.16994524e+00 3.18757743e-01 3.82864863e-01
-1.37788188e+00 -3.58096242e-01 -2.64110804e-01 -1.40851870e-01
-1.29591060e+00 -4.86652046e-01 8.00220847e-01 -5.33774495e-01
5.82755864e-01 4.81902391e-01 2.79436946e-01 6.01973534e-01
-1.34932801e-01 -2.23979548e-01 8.64542842e-01 -7.48848259e-01
2.26247951e-01 4.50375795e-01 3.39519054e-01 2.12663319e-02
1.20060883e-01 4.69310060e-02 -3.34931821e-01 -7.60858282e-02
1.28726318e-01 -3.55128348e-01 -3.77040684e-01 -2.37955600e-01
-1.03893805e+00 6.09292150e-01 2.99871385e-01 8.53490472e-01
-2.28888527e-01 -1.91059947e-01 4.25375074e-01 -8.72647110e-03
4.97349054e-01 2.19964892e-01 -7.47502670e-02 -1.27243057e-01
-9.25941885e-01 -9.37075261e-03 6.59199297e-01 1.65357873e-01
6.57283962e-01 4.98263896e-01 7.12850615e-02 8.97315681e-01
2.74410009e-01 4.22401547e-01 7.90825367e-01 -6.07390404e-01
2.28061944e-01 3.87684315e-01 8.21735188e-02 -7.38433242e-01
-3.35613757e-01 -6.74430311e-01 -3.17842603e-01 2.75776386e-01
8.78929675e-01 -1.03051282e-01 -5.65764666e-01 1.75645304e+00
3.22822690e-01 -5.48949122e-01 3.10449988e-01 7.36496449e-01
5.74262679e-01 7.06067085e-01 4.18797992e-02 -4.54678893e-01
1.00034511e+00 -4.87314701e-01 -9.88975346e-01 1.31554991e-01
2.76179075e-01 -1.07986593e+00 1.19204366e+00 3.46169621e-01
-5.29631495e-01 -6.63295686e-01 -1.25302029e+00 3.33910167e-01
-3.18795174e-01 2.31254205e-01 -1.37005836e-01 1.20661449e+00
-4.52804357e-01 4.77487475e-01 -5.65177381e-01 -1.32775009e-01
-3.77867311e-01 8.07450265e-02 -4.60867494e-01 6.10749900e-01
-7.68507838e-01 6.52963758e-01 2.97837168e-01 -1.51084764e-02
-3.22793663e-01 -5.07041037e-01 -6.07155621e-01 1.75691128e-01
-4.24853619e-03 1.41718090e-01 9.17853475e-01 -9.96543586e-01
-1.72059011e+00 3.84983629e-01 -2.29456946e-01 -3.57115775e-01
6.08674586e-01 -1.04859881e-01 -6.11071289e-01 5.49943559e-02
-1.45291567e-01 1.50902361e-01 8.74712050e-01 -1.17386985e+00
-2.43113980e-01 -3.63081843e-01 -6.14051163e-01 2.63268322e-01
-5.19549429e-01 3.32895853e-02 1.36275172e-01 -4.83742714e-01
3.79198551e-01 -6.47968888e-01 3.11013371e-01 -6.36682868e-01
-2.70856410e-01 -3.48027572e-02 5.54481149e-01 -7.89840937e-01
1.11300921e+00 -2.63895178e+00 -1.34654626e-01 3.08821172e-01
-1.86596617e-01 1.99603781e-01 7.14984834e-02 3.14211667e-01
-2.60344177e-01 -3.83782350e-02 -2.66199023e-01 -3.25123966e-01
-4.29898463e-02 1.16139784e-01 9.09642950e-02 5.91995716e-01
-2.11806551e-01 -1.51569337e-01 -7.03990281e-01 -2.30246037e-01
3.40735048e-01 8.99417579e-01 -2.71724135e-01 -5.93685582e-02
3.10459703e-01 1.77077115e-01 4.46595341e-01 -1.64903134e-01
6.28627658e-01 8.16709697e-01 -7.59216025e-03 -2.76931614e-01
-2.89766937e-01 5.60004354e-01 -1.25244975e+00 1.09375787e+00
-3.87274891e-01 7.31108665e-01 9.34335813e-02 -5.41881323e-01
1.52424514e+00 5.50312996e-01 3.33749950e-01 -1.74768001e-01
3.49115551e-01 4.48836744e-01 4.84896004e-01 -2.60493271e-02
5.52583337e-01 -3.73798579e-01 1.23203494e-01 -1.49169099e-02
1.09313585e-01 -2.15564147e-01 3.65425169e-01 8.71364176e-02
2.79856801e-01 -1.21945508e-01 2.36713678e-01 -5.89832962e-01
6.69750035e-01 -5.22007465e-01 3.48089516e-01 2.46311069e-01
-3.51942807e-01 7.09943414e-01 3.84431750e-01 2.88511422e-02
-7.40675807e-01 -1.23549664e+00 -2.79845089e-01 1.10725403e+00
-3.30739409e-01 -3.18470091e-01 -1.19372964e+00 -1.89013213e-01
-2.18032286e-01 9.12382960e-01 -5.29967129e-01 -3.07779431e-01
-4.41384077e-01 -5.65540910e-01 5.00353932e-01 3.16869430e-02
-1.82117391e-02 -8.60894561e-01 -4.19279873e-01 1.83193907e-01
-2.48285636e-01 -9.16302979e-01 -6.46995962e-01 3.04547608e-01
-6.36979818e-01 -5.89284778e-01 -7.55674779e-01 -5.97942114e-01
2.31679499e-01 -7.98016936e-02 3.74659985e-01 -4.69775319e-01
-1.18506104e-01 1.01965606e-01 -3.85284305e-01 -4.50914741e-01
-1.08613765e+00 2.00821161e-01 1.77448332e-01 4.14678276e-01
2.89615095e-01 -3.88779163e-01 -1.52336732e-01 1.71199337e-01
-5.22840202e-01 -7.23580062e-01 -1.78379510e-02 2.83391088e-01
2.43120030e-01 1.62621856e-01 7.69689858e-01 -5.59735596e-01
8.43866408e-01 9.42832977e-03 -5.36174059e-01 -2.40672916e-01
-5.29295743e-01 2.14540049e-01 7.26662636e-01 -3.84182274e-01
-1.12991881e+00 1.17492646e-01 -1.73388273e-01 1.66384667e-01
-3.93578738e-01 1.67517737e-01 -2.12996826e-01 5.63884117e-02
1.17982078e+00 1.63364992e-01 1.30761251e-01 -5.97708642e-01
3.66426349e-01 8.09975803e-01 9.18170750e-01 -4.02385779e-02
5.40096104e-01 7.52953216e-02 -1.56662494e-01 -1.31219816e+00
-3.43347281e-01 -5.30677676e-01 -4.84257013e-01 -3.52333069e-01
7.99152195e-01 -4.12841648e-01 -4.76748407e-01 4.62917745e-01
-9.19856668e-01 4.86241698e-01 -5.05203068e-01 8.49393129e-01
-2.51682311e-01 4.73154277e-01 -2.22725853e-01 -1.15684855e+00
-3.89242738e-01 -9.85329211e-01 3.93456519e-01 2.91295201e-01
-3.06284577e-01 -6.68932140e-01 1.49611205e-01 1.48437053e-01
3.37044805e-01 1.62439525e-01 8.18700373e-01 -7.25355089e-01
3.74311715e-01 -4.92582828e-01 3.18528682e-01 9.23347175e-01
6.37907505e-01 4.60526198e-01 -1.20469928e+00 8.93179998e-02
2.59461880e-01 4.79565173e-01 7.48827398e-01 3.48877043e-01
2.41118252e-01 -1.56791344e-01 2.96719491e-01 1.32466674e-01
1.24929786e+00 2.53492832e-01 2.55213380e-01 8.75607319e-03
4.78712589e-01 7.03568697e-01 1.74914092e-01 2.78983831e-01
-1.11744031e-01 5.29536426e-01 2.61672974e-01 1.43441528e-01
-3.34884167e-01 -6.44831806e-02 4.05365288e-01 1.04250872e+00
1.67726893e-02 -6.92469180e-02 -7.54452646e-01 7.96330452e-01
-1.10706496e+00 -8.79402459e-01 -4.52302814e-01 2.54899168e+00
5.65120399e-01 3.15851867e-01 4.74452406e-01 9.66782808e-01
9.34876323e-01 1.69318259e-01 8.44832137e-02 -1.07363498e+00
-1.61789730e-01 1.11621588e-01 6.26695037e-01 1.16571331e+00
-7.93182969e-01 3.83764327e-01 6.38415813e+00 5.91726780e-01
-1.07367265e+00 -3.58667970e-02 2.14132406e-02 -1.98754326e-01
-1.05964400e-01 -2.74420023e-01 -4.42416072e-01 3.49491507e-01
1.12606573e+00 -4.13023144e-01 5.93285620e-01 7.16365457e-01
2.70959526e-01 -2.94700444e-01 -6.97393596e-01 7.76655912e-01
5.75887337e-02 -5.06301880e-01 -8.51190761e-02 3.51371467e-02
4.18164194e-01 -1.33079126e-01 3.85849178e-02 -1.26766220e-01
-3.80950272e-01 -6.39598012e-01 1.07163179e+00 2.25187644e-01
1.32902220e-01 -9.58850801e-01 7.77931571e-01 1.57087460e-01
-1.12164676e+00 -5.13598230e-03 -1.67622775e-01 -2.14995205e-01
-3.45909968e-03 6.41307890e-01 -1.08926988e+00 3.13376695e-01
3.42775524e-01 -1.64445460e-01 -3.13084304e-01 9.53247666e-01
-4.07625251e-02 9.02399540e-01 -4.56782788e-01 -6.75524771e-02
-1.20026477e-01 -3.80086005e-01 7.64833629e-01 1.31874859e+00
2.50479639e-01 -4.85568851e-01 -4.50863570e-01 6.29510283e-01
1.36119500e-01 7.91634679e-01 -1.78878441e-01 -9.85916033e-02
5.96935451e-01 1.19249296e+00 -7.33007908e-01 -8.28838274e-02
-5.16300313e-02 6.38577104e-01 -1.94260553e-01 3.28964293e-01
-3.70458752e-01 -6.75454855e-01 5.19287467e-01 1.29247576e-01
7.65741337e-03 -2.50182182e-01 -3.45095873e-01 -3.56960684e-01
6.95780516e-02 -6.02332890e-01 1.55950218e-01 -1.75410494e-01
-7.37795472e-01 9.89053369e-01 -1.11205935e-01 -8.60494971e-01
-5.01299679e-01 -5.23848712e-01 -3.91988695e-01 1.19670022e+00
-6.29439890e-01 -2.69264638e-01 -6.26626611e-02 2.38684371e-01
1.52892604e-01 -5.77519275e-02 9.62981343e-01 5.00883698e-01
-2.65757203e-01 5.97180665e-01 2.66438484e-01 1.14965118e-01
6.99439704e-01 -1.25046277e+00 1.33381739e-01 8.23182583e-01
2.41183996e-01 4.65260208e-01 1.14242673e+00 -2.42505759e-01
-3.88020217e-01 -4.96824324e-01 1.13132930e+00 -2.36033439e-03
3.01655591e-01 -4.70662951e-01 -9.37847435e-01 8.94094482e-02
1.93185791e-01 -2.49916971e-01 7.37461567e-01 2.58459197e-03
-3.41631860e-01 -4.97567505e-01 -1.07688904e+00 2.84753203e-01
2.20585153e-01 -7.65335441e-01 -6.43884242e-01 -7.42408261e-02
5.55098355e-01 5.19343764e-02 -7.16763198e-01 -2.90731043e-02
4.97557163e-01 -1.24998021e+00 6.19377911e-01 -2.38369568e-03
-2.89296120e-01 -4.66018289e-01 -3.14767718e-01 -1.40002024e+00
-1.56250000e-01 -5.63598812e-01 5.80215693e-01 1.59615505e+00
4.35735524e-01 -6.50895476e-01 4.82635975e-01 1.24797933e-01
9.55809802e-02 1.11176439e-01 -1.03334343e+00 -7.30023682e-01
-1.61025777e-01 -1.96143046e-01 4.60852504e-01 4.71701115e-01
2.05156859e-02 3.69698107e-01 -3.31645310e-02 -1.00593626e-01
6.27909780e-01 -3.17221373e-01 4.33761716e-01 -1.30828452e+00
-1.21456601e-01 -5.31761169e-01 -6.02406502e-01 -3.46196182e-02
-2.10954443e-01 -6.28146112e-01 3.01311702e-01 -1.09909058e+00
-4.33414668e-01 1.92098722e-01 -4.76888567e-01 -7.60687366e-02
-1.59236580e-01 3.24232191e-01 4.53838080e-01 -1.81621045e-01
4.01140779e-01 4.36255187e-01 7.86557019e-01 1.99329481e-01
-5.93701482e-01 1.97246149e-01 -3.59878004e-01 7.27314234e-01
8.79468381e-01 -5.90699315e-01 -3.36189806e-01 -7.67377578e-03
-2.72580743e-01 -2.23855719e-01 -1.15379423e-01 -1.35974979e+00
-2.26543218e-01 2.09524766e-01 4.99031134e-02 -4.98836279e-01
3.17528635e-01 -6.92111194e-01 2.94081360e-01 3.93594384e-01
-3.21397752e-01 -5.56502678e-02 8.37950855e-02 1.83015868e-01
-3.28740060e-01 -5.27417898e-01 1.04910433e+00 2.00222865e-01
-1.02058966e-02 -6.78553343e-01 -7.87251592e-01 -2.53547639e-01
4.74425226e-01 -3.68528008e-01 -9.59364399e-02 -5.49542606e-01
-7.24523008e-01 -9.04886365e-01 1.32740006e-01 3.90649468e-01
-3.67822945e-02 -8.67063046e-01 -7.45415270e-01 1.49908960e-01
-1.30608231e-01 -7.00541258e-01 1.41146243e-01 4.79282320e-01
-4.79510367e-01 2.76430994e-01 -3.07612240e-01 -3.71562660e-01
-1.38686550e+00 4.99214739e-01 5.01821220e-01 2.95832217e-01
-3.14493403e-02 5.20913303e-01 8.79962742e-02 -4.98635530e-01
3.92760426e-01 -4.80863750e-01 -3.13667715e-01 4.25831974e-01
4.00029302e-01 8.11531365e-01 4.68630612e-01 -9.80395377e-01
-4.85549062e-01 5.55159688e-01 1.81909114e-01 -5.73308289e-01
9.22561824e-01 -2.09298939e-01 -2.97532082e-02 8.41277540e-01
1.36465800e+00 7.81338751e-01 -8.43423605e-01 1.07683413e-01
-3.83840241e-02 -2.52806038e-01 1.60172194e-01 -6.72240734e-01
-4.80383843e-01 7.31775343e-01 9.48689222e-01 4.96967882e-01
1.06143510e+00 -3.16810817e-01 2.33814958e-02 -1.23102479e-01
-1.45502165e-01 -1.13578582e+00 -3.30494851e-01 2.73044139e-01
7.13075340e-01 -5.93716800e-01 -1.81357026e-01 -5.11100113e-01
-4.04848307e-01 1.04786742e+00 1.91154853e-01 -1.20770939e-01
6.65795803e-01 1.12047732e-01 6.68521702e-01 5.62992156e-01
-1.31500075e-02 -1.68487757e-01 4.44521040e-01 6.95903480e-01
7.85270631e-01 3.42064261e-01 -9.39134181e-01 6.43671036e-01
-7.87879407e-01 -4.59009558e-01 6.31126761e-01 4.33406532e-01
-6.18622065e-01 -1.13780904e+00 -9.19924915e-01 -1.70902247e-04
-3.62805694e-01 -4.08551237e-03 -4.57687020e-01 5.69351435e-01
1.66151032e-01 1.30113280e+00 1.17596731e-01 -2.17003748e-01
3.74160290e-01 4.83536750e-01 3.16941440e-01 -3.30903023e-01
-8.17157447e-01 4.27179247e-01 -5.16609102e-02 8.17496702e-02
-1.75895631e-01 -6.65669203e-01 -1.35798609e+00 -1.09140657e-01
-3.97153437e-01 5.07933855e-01 1.24651980e+00 7.34328747e-01
1.26219178e-02 5.70709169e-01 6.89185321e-01 -5.78863502e-01
-7.46214211e-01 -1.22333157e+00 -7.75865972e-01 4.69344050e-01
4.70246077e-01 -2.37828344e-01 -7.74329245e-01 4.72982228e-03] | [14.973129272460938, 5.938873767852783] |
bf4eaf69-5461-4446-abca-52bc9db71712 | does-deep-machine-vision-have-just-noticeable | 2102.08168 | null | https://arxiv.org/abs/2102.08168v2 | https://arxiv.org/pdf/2102.08168v2.pdf | Just Noticeable Difference for Deep Machine Vision | As an important perceptual characteristic of the Human Visual System (HVS), the Just Noticeable Difference (JND) has been studied for decades with image and video processing (e.g., perceptual visual signal compression). However, there is little exploration on the existence of JND for the Deep Machine Vision (DMV), although the DMV has made great strides in many machine vision tasks. In this paper, we take an initial attempt, and demonstrate that the DMV has the JND, termed as the DMV-JND. We then propose a JND model for the image classification task in the DMV. It has been discovered that the DMV can tolerate distorted images with average PSNR of only 9.56dB (the lower the better), by generating JND via unsupervised learning with the proposed DMV-JND-NET. In particular, a semantic-guided redundancy assessment strategy is designed to restrain the magnitude and spatial distribution of the DMV-JND. Experimental results on image classification demonstrate that we successfully find the JND for deep machine vision. Our DMV-JND facilitates a possible direction for DMV-oriented image and video compression, watermarking, quality assessment, deep neural network security, and so on. | ['Yao Zhao', 'Jian Lou', 'Weisi Lin', 'huan zhang', 'Xin Fu', 'Xingxing Zhang', 'Jian Jin'] | 2021-02-16 | null | null | null | null | ['neural-network-security'] | ['miscellaneous'] | [ 3.84606749e-01 -1.95916086e-01 -2.17118502e-01 -1.41791198e-02
3.02530080e-02 -4.04299190e-03 4.44616348e-01 -3.82862752e-03
-2.29542926e-01 4.16352391e-01 3.08635712e-01 -3.94777596e-01
-1.50709271e-01 -5.61143875e-01 -5.36489069e-01 -8.20594370e-01
-1.00455202e-01 -5.87800324e-01 3.48102897e-01 -2.42956877e-01
5.76464474e-01 4.64106917e-01 -1.52229369e+00 1.91547722e-01
6.05961561e-01 1.53442967e+00 5.25865138e-01 6.43528640e-01
1.50924280e-01 9.74906564e-01 -9.24406767e-01 -3.68487597e-01
4.88834500e-01 -5.13903379e-01 -6.61884129e-01 3.14971864e-01
2.97812253e-01 -4.87665385e-01 -7.60365605e-01 1.47791612e+00
8.56328964e-01 2.21788045e-02 7.46154666e-01 -1.43820369e+00
-1.07003736e+00 2.96106547e-01 -5.84037662e-01 5.44070363e-01
-3.44512723e-02 2.95447201e-01 6.09954059e-01 -6.25427604e-01
4.44626391e-01 1.29843080e+00 6.10202730e-01 4.77779388e-01
-1.05663872e+00 -2.73387164e-01 -1.56073987e-01 8.35341871e-01
-1.46830380e+00 -2.19026953e-01 1.14045072e+00 -2.78549075e-01
8.80196750e-01 2.13951051e-01 5.87405860e-01 8.80125582e-01
7.60078073e-01 6.49447799e-01 1.09735703e+00 -4.57164556e-01
4.26404893e-01 -1.06095940e-01 -2.72396743e-01 3.64231646e-01
1.07854828e-01 3.34822416e-01 -3.45026135e-01 3.73108089e-01
8.85247350e-01 -7.58941174e-02 -4.46593374e-01 -2.09936097e-01
-8.13811541e-01 6.84842527e-01 4.25028920e-01 1.62802607e-01
-2.93908626e-01 5.88175654e-02 6.27569199e-01 8.22298169e-01
7.50807971e-02 2.52475768e-01 -2.59581327e-01 1.90336585e-01
-8.11079741e-01 -6.22062534e-02 3.55990946e-01 9.57657456e-01
2.53334582e-01 7.00044274e-01 -2.92699248e-01 1.15248978e+00
1.37578070e-01 3.80497158e-01 7.33240545e-01 -1.36791718e+00
2.13397622e-01 1.79219171e-01 -2.08917618e-01 -1.55431175e+00
-1.55692339e-01 -4.84077841e-01 -1.66025102e+00 6.98675811e-01
-2.49624848e-01 2.72067606e-01 -8.49882364e-01 1.49300730e+00
-1.50070816e-01 -8.62754434e-02 2.23928943e-01 1.01257062e+00
8.75292361e-01 7.91884184e-01 -8.98999199e-02 -6.21050656e-01
9.88372922e-01 -7.35687673e-01 -1.02166271e+00 -1.66934788e-01
-8.44862964e-03 -9.23896134e-01 8.74965906e-01 6.58270955e-01
-1.03992438e+00 -1.13118184e+00 -1.33750677e+00 -1.29875839e-01
-2.31392048e-02 -1.06502466e-01 3.87750715e-01 5.96632242e-01
-1.21510899e+00 4.82896268e-01 -2.64488459e-01 -1.61905631e-01
3.77243191e-01 1.24737144e-01 -1.77100062e-01 -6.63611069e-02
-1.37111950e+00 9.09120023e-01 5.45144320e-01 2.68328547e-01
-1.28781378e+00 -2.14936316e-01 -7.24735618e-01 -1.23151518e-01
1.24560729e-01 -5.19445777e-01 8.50785553e-01 -1.09230816e+00
-1.31157255e+00 7.37700760e-01 1.21183373e-01 -8.43906462e-01
4.46542799e-01 2.15733945e-01 -7.56781459e-01 3.89044225e-01
-2.08623037e-01 7.80985773e-01 1.34744775e+00 -1.45353329e+00
-3.14290434e-01 -1.49730921e-01 -5.99171743e-02 1.09700166e-01
-4.27960277e-01 -1.38542382e-02 -3.92275006e-01 -1.18967998e+00
3.60174984e-01 -4.78471875e-01 -3.09809409e-02 3.72111708e-01
-3.38515967e-01 -1.13352187e-01 9.37813640e-01 -9.52385008e-01
1.27173901e+00 -2.44201565e+00 1.82393745e-01 2.34306864e-02
2.46442944e-01 6.33625925e-01 -2.27646440e-01 1.93179086e-01
-9.34612378e-02 8.47625509e-02 -3.62402081e-01 8.72472897e-02
-2.38375366e-01 3.64053875e-01 -3.00056368e-01 5.37138104e-01
1.04174674e-01 5.40411592e-01 -5.68280280e-01 -5.59927404e-01
3.47041994e-01 4.88963664e-01 -6.18673265e-01 1.49401799e-01
8.30437988e-02 3.92616689e-01 2.47268817e-05 6.85801625e-01
1.09578919e+00 1.49768576e-01 -5.84169198e-03 -6.34593308e-01
-8.87893289e-02 -3.00394028e-01 -1.09575057e+00 1.32247865e+00
-7.99118355e-02 1.17122328e+00 1.52327061e-01 -1.44781244e+00
1.11683905e+00 1.68868035e-01 3.56077880e-01 -1.20936108e+00
3.40892822e-01 1.27765685e-01 2.67188996e-01 -8.15391541e-01
4.57110941e-01 -6.46731928e-02 3.51551652e-01 -1.78668693e-01
1.75542608e-01 -8.68694708e-02 -1.05661340e-01 4.19315025e-02
7.45817661e-01 -5.28011322e-01 4.22317743e-01 -2.48886704e-01
6.26814902e-01 -3.25327724e-01 6.49293482e-01 7.09736764e-01
-7.34014153e-01 7.06278622e-01 3.74722004e-01 -2.16723561e-01
-1.37300384e+00 -9.60425377e-01 -3.21988493e-01 5.15021205e-01
6.00458801e-01 7.78255761e-02 -7.33433723e-01 -4.58831489e-02
-2.01672226e-01 5.19683421e-01 -7.48651177e-02 -7.16312528e-01
-3.16001445e-01 -4.13863897e-01 7.15554416e-01 2.95860708e-01
1.18034399e+00 -1.11637950e+00 -3.14581901e-01 4.31410372e-02
-1.10711135e-01 -1.11902380e+00 -3.84310067e-01 7.69239664e-02
-8.33596706e-01 -6.70156538e-01 -9.32499051e-01 -1.33040845e+00
4.66042817e-01 4.90807742e-01 8.03885639e-01 3.89771573e-02
-2.28889912e-01 3.07162106e-01 -4.23679441e-01 -3.12794209e-01
-5.45030832e-01 -6.39905214e-01 1.78279608e-01 -1.90809034e-02
2.29408011e-01 -7.68529415e-01 -9.63839352e-01 4.83986288e-01
-1.14017701e+00 -1.97968185e-01 6.65874541e-01 8.88730884e-01
7.68059909e-01 8.45383525e-01 4.65876251e-01 -1.68126494e-01
6.98075235e-01 -1.93215474e-01 -4.30465311e-01 -5.15578613e-02
-8.12843502e-01 -2.56402552e-01 8.08116019e-01 -5.93432486e-01
-8.05104136e-01 -4.05538172e-01 -4.03582364e-01 -8.06449473e-01
-1.26462430e-01 3.91517669e-01 -6.19570851e-01 -3.56949151e-01
5.75528204e-01 7.98324764e-01 9.29846838e-02 -4.17719394e-01
2.53714800e-01 8.73283565e-01 8.06091189e-01 1.84306875e-02
7.59055674e-01 2.07425043e-01 1.68700531e-01 -1.26854181e+00
-1.48215339e-01 -1.62932396e-01 -1.62622064e-01 -3.01981509e-01
9.47435021e-01 -9.32803512e-01 -5.28426766e-01 8.32875848e-01
-1.33143497e+00 -6.11776523e-02 -2.32022241e-01 5.24660826e-01
-5.12759864e-01 8.60805511e-01 -7.11375594e-01 -6.97259784e-01
-1.96324870e-01 -1.21695447e+00 3.49802136e-01 -1.04000382e-02
1.15511023e-01 -8.82354498e-01 -3.60057563e-01 3.39125276e-01
4.39169496e-01 -1.41846696e-02 1.14699829e+00 -1.82663530e-01
-5.26781380e-01 8.57886374e-02 -3.30487967e-01 1.27319396e+00
4.04281691e-02 -3.77284467e-01 -8.89286399e-01 -3.47859055e-01
5.85678279e-01 -2.06026286e-01 1.12269676e+00 6.62276089e-01
1.67896652e+00 -5.86515427e-01 4.00286674e-01 1.00711405e+00
1.56131566e+00 7.50349939e-01 1.25505400e+00 4.02166367e-01
6.26382530e-01 4.65336442e-01 4.13996190e-01 3.32686126e-01
-1.54565126e-01 4.02623147e-01 8.59313011e-01 -2.56837189e-01
-7.27632940e-01 -1.75628901e-01 3.36023033e-01 1.06143010e+00
-1.60593376e-01 -3.48004937e-01 -4.38653320e-01 4.48501766e-01
-1.36222541e+00 -1.01402283e+00 -2.71938443e-02 1.89139736e+00
8.50042641e-01 3.29323351e-01 -3.91649097e-01 8.18332851e-01
7.74929583e-01 5.74363649e-01 -5.98400056e-01 -6.87092245e-01
-6.23867989e-01 -6.14070371e-02 6.23260856e-01 4.03659284e-01
-1.12393117e+00 6.08058929e-01 6.14861059e+00 1.30238652e+00
-1.20433450e+00 7.40221813e-02 7.25937545e-01 4.53129977e-01
-1.82210118e-01 -2.81088442e-01 -3.97651643e-01 6.37971282e-01
4.06427890e-01 8.09474885e-02 4.79867131e-01 9.04069245e-01
4.65368271e-01 1.10813044e-01 -6.91276670e-01 1.41697991e+00
1.42143533e-01 -1.19518507e+00 6.24929011e-01 2.01993007e-02
9.11229730e-01 -5.43445110e-01 3.62790436e-01 5.51675372e-02
-4.55285877e-01 -9.92163181e-01 8.37053418e-01 2.07478434e-01
1.06254232e+00 -8.46255541e-01 8.60364437e-01 3.25859278e-01
-9.13250029e-01 -3.15108955e-01 -9.43299532e-01 4.69978862e-02
-1.59685805e-01 6.77495182e-01 -3.42014819e-01 2.77181119e-01
6.83635652e-01 7.63637185e-01 -3.77757490e-01 1.17758524e+00
-2.52361670e-02 4.28458065e-01 3.56605262e-01 2.56419629e-01
-9.77757126e-02 -2.60979142e-02 8.91855597e-01 1.05894268e+00
4.00317818e-01 -6.37444621e-03 -1.70357316e-03 6.21242166e-01
-1.42228082e-01 -1.77105844e-01 -6.63664937e-01 1.11064285e-01
3.62855375e-01 6.41589999e-01 -5.77294230e-01 -8.51780623e-02
-1.87192142e-01 1.24377191e+00 -4.37322438e-01 5.42113662e-01
-6.93375230e-01 -3.87740582e-01 5.08977175e-01 -8.54272246e-02
4.96537089e-01 -1.13769569e-01 -5.14485717e-01 -8.13957453e-01
2.11014282e-02 -1.09134829e+00 -4.81169596e-02 -8.52759659e-01
-1.30817258e+00 5.64262092e-01 -1.50889143e-01 -1.67963803e+00
1.89380839e-01 -7.38846838e-01 -4.79789913e-01 6.05464578e-01
-1.65375566e+00 -5.70241749e-01 -1.73365638e-01 8.81229639e-01
7.94671595e-01 -6.73379481e-01 4.07624722e-01 4.70155358e-01
-2.68127650e-01 6.93688810e-01 2.70781696e-01 8.61508548e-02
4.96235937e-01 -8.39303732e-01 -1.57123655e-01 1.06691170e+00
-3.16615671e-01 2.47112051e-01 9.33036327e-01 -4.80733812e-01
-1.38301516e+00 -1.27007484e+00 5.51370978e-01 4.10193950e-01
3.30767334e-01 9.63607654e-02 -8.89913380e-01 8.15628767e-02
4.54817176e-01 -1.12124018e-01 2.51803696e-01 -6.97962046e-01
-3.41461420e-01 -5.40156245e-01 -1.37426531e+00 5.43279648e-01
1.06028986e+00 -6.98314846e-01 -6.22932315e-01 7.46916756e-02
1.02433240e+00 1.48353130e-01 -6.07034147e-01 4.87384319e-01
1.78007752e-01 -1.23803639e+00 1.32464767e+00 -1.79873720e-01
7.48681545e-01 -4.76155281e-01 -7.00881362e-01 -1.28114748e+00
-5.79206884e-01 -4.08119291e-01 -2.77159333e-01 1.17467546e+00
-5.62939644e-02 -3.11744124e-01 3.64589870e-01 -1.43223330e-01
-3.98385346e-01 -4.66993302e-01 -1.00321710e+00 -1.00566053e+00
-1.53857574e-01 -5.90677083e-01 1.95750222e-01 8.65582168e-01
-5.99159300e-01 -1.42349896e-03 -8.88243318e-01 3.77518296e-01
8.04686427e-01 -3.38786364e-01 2.58836240e-01 -8.25730860e-01
-4.77360100e-01 -5.47498882e-01 -9.55151856e-01 -1.34157562e+00
-2.15720743e-01 -6.77767813e-01 -1.93934422e-02 -1.56185317e+00
1.67883545e-01 1.58577889e-01 -6.49867177e-01 1.24549940e-02
2.30288148e-01 3.70809644e-01 3.00296605e-01 4.63022739e-01
-3.52928668e-01 7.25027978e-01 1.67046785e+00 -7.09787369e-01
-7.92414248e-02 -1.53367713e-01 -7.19935060e-01 8.40934992e-01
8.86537075e-01 -2.45808035e-01 -6.82178319e-01 -5.12134731e-01
-5.71504943e-02 9.27815512e-02 6.09015405e-01 -1.18362129e+00
3.14586610e-01 -2.46858507e-01 6.11711800e-01 -6.23589873e-01
2.93943852e-01 -9.30148721e-01 -7.62061477e-02 7.75703847e-01
-4.01719660e-01 -2.70345420e-01 -6.32549971e-02 6.99163079e-01
-6.71778083e-01 -2.06261218e-01 1.24234855e+00 -9.02342796e-02
-1.27573693e+00 8.43322575e-02 -5.41762412e-01 -1.33630589e-01
8.30589294e-01 -4.21955794e-01 -3.13820869e-01 -5.43046236e-01
-4.77123588e-01 -3.43995780e-01 1.25563532e-01 2.57120639e-01
1.44052482e+00 -1.36491191e+00 -6.64169848e-01 3.96099895e-01
-9.75434557e-02 -3.48863453e-01 5.19239962e-01 5.45117319e-01
-6.89399719e-01 1.36301756e-01 -6.60156429e-01 -6.48761630e-01
-1.26672912e+00 8.98386657e-01 2.84197628e-01 1.69892043e-01
-5.86011231e-01 7.73897827e-01 2.09267244e-01 2.94382662e-01
6.31559253e-01 -1.55823752e-01 -5.11171341e-01 -2.15943500e-01
6.71154916e-01 5.26325166e-01 -1.07876649e-02 -6.88345730e-01
-5.28800860e-02 5.11233449e-01 2.02109188e-01 1.51642919e-01
1.07789516e+00 -3.70614141e-01 -2.82502800e-01 1.43120810e-01
1.55685616e+00 -1.91643536e-01 -1.25706327e+00 -9.59884748e-03
-4.39731151e-01 -6.13736987e-01 3.93125534e-01 -5.68662167e-01
-1.41243851e+00 9.82881069e-01 1.18023717e+00 5.14090955e-01
1.77857113e+00 -2.76930988e-01 9.95348096e-01 2.55495012e-01
2.88713157e-01 -1.16824412e+00 4.24254924e-01 4.36199993e-01
1.23496735e+00 -1.08965755e+00 -4.12579924e-02 -2.09092528e-01
-7.41273344e-01 1.00861299e+00 3.14836919e-01 -2.30586901e-01
8.08979869e-01 4.63382564e-02 -4.78892140e-02 1.15289234e-01
-3.75516891e-01 1.38528839e-01 1.90871775e-01 1.02058554e+00
-1.15944616e-01 -1.03912368e-01 -5.25863588e-01 2.78492481e-01
-1.40086651e-01 -8.96265805e-02 4.88340318e-01 7.18490899e-01
-7.62427986e-01 -7.18778551e-01 -4.39563304e-01 5.87963946e-02
-5.31962216e-01 -2.07797542e-01 -1.36780769e-01 3.70603502e-01
6.81931317e-01 1.20136476e+00 3.66210975e-02 -7.87401974e-01
2.73351043e-01 -5.32226920e-01 3.12775701e-01 3.52095850e-02
2.92500257e-02 1.34641573e-01 -3.25361371e-01 -3.30055743e-01
-4.86719102e-01 8.13135505e-02 -9.31735873e-01 -4.87346947e-01
1.17889233e-01 -2.85984248e-01 5.85952520e-01 6.01342201e-01
4.64787520e-02 6.01237833e-01 1.00368297e+00 -6.76520467e-01
-5.12076318e-01 -7.15258360e-01 -8.83478880e-01 5.25338054e-01
6.46183491e-01 -4.12007391e-01 -5.93977749e-01 3.31054568e-01] | [11.425293922424316, -1.7933350801467896] |
5a679198-6bfa-43f8-ac26-61f93bcf1399 | faasta-a-fast-solver-for-total-variation | 1512.06999 | null | http://arxiv.org/abs/1512.06999v1 | http://arxiv.org/pdf/1512.06999v1.pdf | FAASTA: A fast solver for total-variation regularization of ill-conditioned problems with application to brain imaging | The total variation (TV) penalty, as many other analysis-sparsity problems,
does not lead to separable factors or a proximal operatorwith a closed-form
expression, such as soft thresholding for the $\ell\_1$ penalty. As a result,
in a variational formulation of an inverse problem or statisticallearning
estimation, it leads to challenging non-smooth optimization problemsthat are
often solved with elaborate single-step first-order methods. When thedata-fit
term arises from empirical measurements, as in brain imaging, it isoften very
ill-conditioned and without simple structure. In this situation, in proximal
splitting methods, the computation cost of thegradient step can easily dominate
each iteration. Thus it is beneficialto minimize the number of gradient
steps.We present fAASTA, a variant of FISTA, that relies on an internal solver
forthe TV proximal operator, and refines its tolerance to balance
computationalcost of the gradient and the proximal steps. We give benchmarks
andillustrations on "brain decoding": recovering brain maps from
noisymeasurements to predict observed behavior. The algorithm as well as
theempirical study of convergence speed are valuable for any non-exact
proximaloperator, in particular analysis-sparsity problems. | ['Michael Eickenberg', 'Gaël Varoquaux', 'Elvis Dohmatob', 'Bertand Thirion'] | 2015-12-22 | null | null | null | null | ['brain-decoding', 'brain-decoding'] | ['medical', 'miscellaneous'] | [ 9.19180140e-02 8.82688239e-02 1.37607992e-01 -4.51909065e-01
-9.93113637e-01 -2.56097198e-01 3.49400938e-01 -9.36873183e-02
-4.76220787e-01 9.00094390e-01 -1.73139438e-01 -2.97304660e-01
-2.01391548e-01 -3.16261768e-01 -8.33820939e-01 -1.12179005e+00
8.76725912e-02 4.16003346e-01 9.42457560e-03 -5.60617894e-02
6.72299862e-02 3.24502856e-01 -1.32147634e+00 3.53443436e-02
1.22301412e+00 1.13948023e+00 3.01608950e-01 2.04460546e-01
-1.29067302e-01 5.37978888e-01 -1.92392305e-01 -2.91418344e-01
2.02326044e-01 -7.88384497e-01 -6.66070044e-01 2.32977882e-01
4.97963488e-01 -7.27446154e-02 7.97918066e-02 1.35104001e+00
3.17501307e-01 2.65478253e-01 7.13800848e-01 -1.07424974e+00
-1.46716386e-01 5.02480209e-01 -9.24437404e-01 4.41721737e-01
1.61555871e-01 -1.66215971e-01 7.03613460e-01 -1.25050712e+00
5.12068212e-01 9.40765560e-01 8.72931480e-01 1.93505004e-01
-1.81255877e+00 -3.32352787e-01 9.65432003e-02 2.46064320e-01
-1.16120338e+00 -8.15723836e-01 7.40898967e-01 -6.86306775e-01
5.41531384e-01 4.58290905e-01 5.16349971e-01 9.66777682e-01
3.93282294e-01 8.25138211e-01 1.36994183e+00 -1.84266359e-01
2.98524529e-01 4.82537597e-02 3.10412943e-01 6.12243891e-01
1.56158820e-01 -3.61050665e-02 -4.37056154e-01 -2.48933807e-01
7.44496167e-01 -1.66126743e-01 -4.86017942e-01 -5.01767516e-01
-9.01745796e-01 7.89171875e-01 3.65245551e-01 3.36442560e-01
-6.39843464e-01 3.34119618e-01 1.82768896e-01 4.67682689e-01
6.14910305e-01 1.81810021e-01 -8.50886852e-02 -1.81101449e-02
-1.42709184e+00 4.36345190e-01 7.36149728e-01 6.40064180e-01
8.44174862e-01 2.81497508e-01 3.42305377e-02 9.64439094e-01
4.67109948e-01 5.92552006e-01 2.48393148e-01 -1.18155515e+00
1.83554426e-01 6.73876554e-02 4.69454639e-02 -1.08571208e+00
-5.15365720e-01 -5.43077350e-01 -9.95676398e-01 3.62607330e-01
8.35776389e-01 -2.14832798e-01 -1.00521016e+00 1.51659477e+00
4.36949849e-01 6.35484606e-02 -6.75055504e-01 1.32978904e+00
5.74206293e-01 4.02527839e-01 -2.00455442e-01 -5.98112285e-01
1.24249971e+00 -6.59209073e-01 -9.75842476e-01 -3.19591850e-01
6.12259567e-01 -7.69035459e-01 9.44501698e-01 7.54491329e-01
-1.52177048e+00 -1.86592624e-01 -7.52018750e-01 -1.46219462e-01
2.12276563e-01 -2.62421399e-01 5.91394544e-01 2.69991219e-01
-1.06248581e+00 8.35103333e-01 -1.19716370e+00 -5.62829226e-02
3.78637344e-01 1.81653082e-01 -2.72417128e-01 4.77841608e-02
-7.46293306e-01 1.01727068e+00 -6.93773851e-02 3.01172793e-01
-9.48445678e-01 -1.14497924e+00 -7.24203348e-01 -6.69756010e-02
4.88593698e-01 -3.96134526e-01 9.00527954e-01 -1.05898392e+00
-1.27276993e+00 8.44894409e-01 -3.41478020e-01 -3.41656834e-01
6.95334852e-01 -1.29260376e-01 1.45645991e-01 2.32037785e-03
1.90392971e-01 1.89327911e-01 1.19724178e+00 -1.05911231e+00
2.66923875e-01 -7.13169217e-01 -2.44298741e-01 8.87509286e-02
1.20336719e-01 2.91636847e-02 -3.20631921e-01 -9.53897238e-01
7.03571916e-01 -9.29357648e-01 -5.74157774e-01 2.02301800e-01
-4.35225397e-01 2.48541966e-01 2.71266758e-01 -1.00542843e+00
1.07146311e+00 -2.26846838e+00 5.91443121e-01 4.65820134e-01
3.41852307e-01 -1.83345214e-01 1.07658111e-01 1.30305842e-01
-3.96308213e-01 -2.03475758e-01 -8.68418336e-01 -4.72935617e-01
-4.39428277e-02 2.22486973e-01 -1.01703979e-01 9.21177030e-01
-4.37491164e-02 5.94394386e-01 -9.19217825e-01 -5.17555594e-01
1.29633144e-01 5.60069203e-01 -7.49582350e-01 4.39665727e-02
8.95649046e-02 6.64462686e-01 -3.99661809e-01 3.46234918e-01
8.07696283e-01 -2.81057656e-01 3.65625918e-02 -2.31849506e-01
-1.99770600e-01 9.50953364e-03 -1.41254520e+00 1.93084931e+00
-4.00091201e-01 5.45668602e-01 9.01351690e-01 -1.60822725e+00
5.62206864e-01 3.14611673e-01 7.35116899e-01 -4.14060742e-01
3.27065915e-01 6.66505277e-01 -1.50102869e-01 -5.82850218e-01
-2.36328728e-02 -3.66813809e-01 2.99372286e-01 2.38720700e-01
6.44078031e-02 -2.25575328e-01 3.35971117e-01 6.93500787e-02
1.16464841e+00 1.44456729e-01 1.37451634e-01 -9.33293343e-01
6.36452019e-01 -1.46746472e-01 5.04394412e-01 6.22289538e-01
-3.22853178e-02 7.93001115e-01 6.44131184e-01 -3.02943766e-01
-8.19224000e-01 -8.92909169e-01 -7.78133452e-01 8.29377711e-01
-2.89446652e-01 -2.07630500e-01 -1.04130650e+00 -3.07369798e-01
-1.07875198e-01 4.90996599e-01 -5.45848489e-01 1.14195406e-01
-4.06900585e-01 -1.14385700e+00 1.13807857e-01 3.85634333e-01
3.59947354e-01 -6.56297147e-01 -4.74164099e-01 2.71573901e-01
-2.47266993e-01 -8.99299920e-01 -5.44619799e-01 3.45069230e-01
-9.85537410e-01 -8.54715466e-01 -9.06452298e-01 -5.51004231e-01
8.10285568e-01 -1.02122128e-01 9.72422242e-01 1.33303225e-01
-2.00850651e-01 1.39859468e-01 -2.66789291e-02 -1.33114874e-01
4.42588702e-03 -3.67650390e-01 -5.56215681e-02 2.45762423e-01
2.03307986e-01 -8.91324461e-01 -3.56033981e-01 1.18120834e-01
-5.99359512e-01 -1.31490692e-01 3.76423091e-01 1.13509178e+00
8.70227516e-01 -3.13422471e-01 1.71654209e-01 -7.56298959e-01
6.07705295e-01 -4.88794297e-01 -8.74881506e-01 6.53524771e-02
-4.21729833e-01 2.39511684e-01 6.28555417e-01 -3.47059309e-01
-9.04915631e-01 1.47926569e-01 -1.82720393e-01 -6.71508312e-01
1.60588071e-01 8.38435292e-01 -2.99544148e-02 -2.21520349e-01
9.54827845e-01 2.64111072e-01 1.50555268e-01 -6.72227442e-01
2.31524706e-01 2.03955323e-01 3.76521289e-01 -6.14214420e-01
5.34107029e-01 4.92544234e-01 8.81711096e-02 -1.09774983e+00
-8.61277521e-01 -3.17509592e-01 -3.91618967e-01 -1.69873923e-01
8.58438253e-01 -5.70950568e-01 -4.86760199e-01 3.56453806e-01
-1.18341053e+00 -3.71034980e-01 -4.53260064e-01 5.96337497e-01
-6.60759330e-01 3.53543431e-01 -6.11776114e-01 -7.53013730e-01
-1.32168368e-01 -1.33871365e+00 7.29888976e-01 -1.37607083e-01
-1.22299679e-01 -1.10276520e+00 1.18119847e-02 1.92488000e-01
4.88001168e-01 3.04304630e-01 5.33430517e-01 -1.61787122e-01
-3.38047922e-01 -2.36844830e-02 -1.14880033e-01 3.04797918e-01
-4.70975637e-01 -3.02510530e-01 -8.95163655e-01 -3.25852841e-01
6.92190886e-01 -1.98250338e-01 9.16070938e-01 9.13051248e-01
1.35305703e+00 -3.69775146e-01 -5.30899875e-02 9.53539431e-01
1.32243240e+00 -2.26975530e-01 3.82193655e-01 -1.33560166e-01
5.74629664e-01 7.71445155e-01 1.64978579e-01 5.63063622e-01
1.18688293e-01 6.98988199e-01 3.00092310e-01 -4.04217616e-02
-1.26570627e-01 2.77934372e-01 3.61023217e-01 9.23469901e-01
-4.35426503e-01 2.35880971e-01 -9.51563358e-01 4.23306882e-01
-1.82611871e+00 -8.48488748e-01 -7.34187007e-01 2.35674787e+00
8.12085152e-01 -1.63000882e-01 4.00527492e-02 2.07166746e-01
5.25244832e-01 1.93567693e-01 -5.96294641e-01 -2.74539977e-01
1.40721440e-01 1.72714844e-01 4.55481619e-01 9.03857529e-01
-8.98740232e-01 4.91224140e-01 6.72692490e+00 9.07778442e-01
-9.70074058e-01 6.44093215e-01 5.33564270e-01 -2.00615451e-01
-4.11290318e-01 1.75328250e-03 -3.24201226e-01 5.67166865e-01
8.05027246e-01 -7.84770865e-03 9.63381231e-01 6.97913349e-01
3.27568203e-01 -2.87766397e-01 -9.24473405e-01 1.29891753e+00
3.21596526e-02 -8.59105885e-01 -5.52281260e-01 1.88062876e-01
4.62417811e-01 3.92721057e-01 -8.70920271e-02 -2.55023036e-02
1.40603865e-02 -9.68364120e-01 7.70110488e-01 4.06939328e-01
5.13317883e-01 -4.43049133e-01 4.84694541e-01 3.61893743e-01
-8.56117785e-01 1.12168610e-01 -5.02238095e-01 1.32091552e-01
4.23321575e-01 1.34026706e+00 -1.10179551e-01 1.83139697e-01
7.69107044e-01 7.58717775e-01 8.31243619e-02 1.02760518e+00
-2.78587550e-01 5.86686611e-01 -5.50063491e-01 4.20568347e-01
7.06015155e-02 -8.37530732e-01 1.02060437e+00 1.08418071e+00
3.03725809e-01 3.94311428e-01 3.33806485e-01 1.25678349e+00
1.89902991e-01 2.71941394e-01 -4.50072080e-01 4.26641107e-02
1.72958467e-02 1.29989958e+00 -7.75805771e-01 -1.69782192e-01
-4.16289896e-01 8.29185545e-01 3.52902234e-01 7.22735345e-01
-8.01513314e-01 -1.34823725e-01 4.14912045e-01 4.73502278e-01
1.84519961e-01 -2.97817379e-01 -6.61593854e-01 -1.42787302e+00
2.87378848e-01 -7.32631922e-01 2.34931067e-01 -5.25059819e-01
-1.10062551e+00 5.63639700e-01 1.57856457e-02 -7.14251697e-01
-1.39613897e-01 -4.86936361e-01 -4.73179758e-01 8.63511562e-01
-1.42600167e+00 -6.41261637e-01 -2.14439571e-01 7.48218894e-01
3.76885772e-01 3.11016411e-01 4.80515122e-01 5.17450452e-01
-7.36473262e-01 3.61145884e-01 2.44359106e-01 -2.81569481e-01
4.29798394e-01 -1.35159874e+00 -3.53389353e-01 1.09698486e+00
-2.67024487e-01 5.55919409e-01 1.26972926e+00 -5.06654918e-01
-1.28583372e+00 -4.29902792e-01 7.07782328e-01 -1.61737606e-01
9.34120953e-01 -4.13447410e-01 -1.35363984e+00 6.55110478e-01
9.38952789e-02 4.29356635e-01 3.64777803e-01 2.67164588e-01
1.37465268e-01 -1.17045224e-01 -1.05765629e+00 1.78667262e-01
1.00699937e+00 -3.74934077e-01 -4.07014787e-01 6.11408830e-01
2.16538459e-02 -4.98529345e-01 -1.05023098e+00 1.21346779e-01
3.26715827e-01 -1.04432106e+00 8.52697432e-01 -3.64287794e-01
3.39751422e-01 -1.12647135e-02 2.20815819e-02 -1.38305295e+00
-3.28233898e-01 -7.62727022e-01 -4.15761769e-01 6.70041561e-01
4.32460457e-01 -5.91379523e-01 7.96932757e-01 8.69169414e-01
-4.19714391e-01 -7.54671633e-01 -1.13663983e+00 -8.75742316e-01
1.62951380e-01 -5.78815281e-01 1.20822035e-01 9.90228176e-01
8.03179964e-02 9.05708820e-02 -3.27816844e-01 -8.21855068e-02
9.07002866e-01 -1.86319768e-01 4.47900742e-01 -1.19984913e+00
-5.88513911e-01 -4.87642467e-01 -2.76007742e-01 -1.05237567e+00
-6.53423890e-02 -9.59464312e-01 4.56339002e-01 -1.16764116e+00
1.84507947e-02 -4.47791040e-01 -3.20258826e-01 5.50039291e-01
-1.29908592e-01 8.05070624e-02 -2.74035502e-02 7.68031850e-02
-2.60871768e-01 5.49402475e-01 1.31134546e+00 4.97114509e-02
-2.07191229e-01 -2.20512733e-01 -2.38402382e-01 9.33088720e-01
3.75730157e-01 -7.95604765e-01 -6.39186025e-01 -5.55339634e-01
3.36021304e-01 3.60023975e-01 5.31491756e-01 -7.66554236e-01
2.62942553e-01 -2.10256293e-01 8.77887085e-02 -2.59779841e-01
5.25000870e-01 -6.54488623e-01 1.97889715e-01 4.61518109e-01
-9.31348559e-03 -1.65393557e-02 -1.71794474e-01 3.51692945e-01
-2.17830360e-01 -5.12287796e-01 9.33171093e-01 -2.06934065e-01
-1.05678163e-01 2.57437825e-01 -4.68511343e-01 3.53714824e-01
7.23491192e-01 -2.19478831e-01 2.56894648e-01 -3.68701816e-01
-1.28259981e+00 4.73361500e-02 4.11123097e-01 -2.68442571e-01
5.71617663e-01 -1.26936471e+00 -9.70688701e-01 2.61198968e-01
-4.48613018e-01 1.00515395e-01 4.55158323e-01 1.99282110e+00
-3.66559952e-01 1.29656643e-02 4.05923650e-02 -7.51048446e-01
-8.20551634e-01 4.98030216e-01 6.35619938e-01 -1.24013446e-01
-7.09158719e-01 1.10950780e+00 4.74692941e-01 -1.77470654e-01
5.88107668e-02 -2.64904112e-01 -1.28208205e-01 2.39224494e-01
7.30130136e-01 5.49874008e-01 2.95090079e-01 -5.25011182e-01
-5.11888742e-01 2.98421383e-01 1.05521694e-01 -2.90404320e-01
1.51337123e+00 -1.43229917e-01 -5.02875447e-01 4.07295793e-01
1.23222673e+00 -2.81037450e-01 -1.29817998e+00 -9.31151584e-02
-3.72791030e-02 -5.25511861e-01 5.97768664e-01 -3.10480565e-01
-1.46578813e+00 8.77747893e-01 5.82624137e-01 3.29056799e-01
1.00083983e+00 -2.98188180e-01 5.40184081e-01 1.91612750e-01
4.20452565e-01 -1.19076777e+00 -1.35240495e-01 3.66373658e-01
1.31774735e+00 -1.36084032e+00 1.52762532e-01 -5.20188332e-01
-2.34594017e-01 1.00579131e+00 2.62626708e-01 -4.75891173e-01
1.02386475e+00 3.11904877e-01 -1.87420577e-01 -3.27455938e-01
-2.84417242e-01 1.15545027e-01 5.03665268e-01 1.13585338e-01
5.33629894e-01 -5.08312806e-02 -8.63707304e-01 4.25462335e-01
-1.37537718e-01 -9.05806124e-02 3.22536498e-01 8.64684701e-01
-2.32090637e-01 -8.54841232e-01 -5.23417532e-01 5.93337357e-01
-4.56676751e-01 -1.58943430e-01 2.60241013e-02 4.17659461e-01
1.28727248e-02 7.58380532e-01 -1.18876144e-01 1.30342290e-01
1.66024014e-01 -2.11249188e-01 7.06005394e-01 -3.78149658e-01
-3.44653368e-01 5.36572397e-01 3.64129208e-02 -9.41867709e-01
-5.32374144e-01 -1.09702778e+00 -1.22255385e+00 -3.85538876e-01
-5.23625493e-01 1.61247522e-01 7.49006152e-01 1.16402471e+00
1.29797012e-02 3.74119520e-01 1.64845765e-01 -1.07919478e+00
-5.48268676e-01 -7.60879040e-01 -7.46898532e-01 4.45340395e-01
3.05642992e-01 -6.78264141e-01 -4.08372790e-01 -2.04372734e-01] | [6.969590663909912, 4.153221607208252] |
368525d8-f066-4783-a431-ef19d0090deb | gpatcher-a-simple-and-adaptive-mlp-model-for | 2306.14340 | null | https://arxiv.org/abs/2306.14340v1 | https://arxiv.org/pdf/2306.14340v1.pdf | GPatcher: A Simple and Adaptive MLP Model for Alleviating Graph Heterophily | While graph heterophily has been extensively studied in recent years, a fundamental research question largely remains nascent: How and to what extent will graph heterophily affect the prediction performance of graph neural networks (GNNs)? In this paper, we aim to demystify the impact of graph heterophily on GNN spectral filters. Our theoretical results show that it is essential to design adaptive polynomial filters that adapts different degrees of graph heterophily to guarantee the generalization performance of GNNs. Inspired by our theoretical findings, we propose a simple yet powerful GNN named GPatcher by leveraging the MLP-Mixer architectures. Our approach comprises two main components: (1) an adaptive patch extractor function that automatically transforms each node's non-Euclidean graph representations to Euclidean patch representations given different degrees of heterophily, and (2) an efficient patch mixer function that learns salient node representation from both the local context information and the global positional information. Through extensive experiments, the GPatcher model demonstrates outstanding performance on node classification compared with popular homophily GNNs and state-of-the-art heterophily GNNs. | ['Dawei Zhou', 'Si Zhang', 'Haohui Wang', 'Shuaicheng Zhang'] | 2023-06-25 | null | null | null | null | ['node-classification'] | ['graphs'] | [-1.02783866e-01 1.89781889e-01 -1.49901032e-01 -1.47407666e-01
3.08335394e-01 -3.34159642e-01 5.27711630e-01 1.48634598e-01
1.07838370e-01 2.94794858e-01 6.53401166e-02 -3.65793020e-01
-3.74190301e-01 -1.36922741e+00 -9.09962893e-01 -7.65321672e-01
-3.82790715e-01 3.65647405e-01 4.14578110e-01 -4.36205238e-01
5.07029854e-02 5.57786405e-01 -1.41020381e+00 6.06616847e-02
1.05595458e+00 8.99105847e-01 1.43334895e-01 7.48803079e-01
6.10280000e-02 7.68552065e-01 -4.37290996e-01 -3.58794481e-01
4.15210873e-01 -5.91649055e-01 -4.71957266e-01 -3.31836522e-01
5.90785503e-01 3.42327893e-01 -8.73955667e-01 1.35586596e+00
6.05514467e-01 2.89279580e-01 6.44014120e-01 -9.79894340e-01
-1.04328144e+00 1.11478138e+00 -5.91457427e-01 4.35699791e-01
-1.21207729e-01 6.20372035e-02 1.44281781e+00 -7.14469850e-01
6.92838013e-01 1.40943515e+00 1.08348060e+00 1.75612211e-01
-1.36271250e+00 -8.71083736e-01 4.48345572e-01 2.12810144e-01
-1.47209930e+00 6.21870086e-02 1.18783057e+00 -1.46566153e-01
8.65397871e-01 1.47013590e-01 1.00630784e+00 1.24579811e+00
3.68501931e-01 6.05444968e-01 6.80149734e-01 -2.67063498e-01
1.26956463e-01 -1.68428302e-01 2.67988667e-02 1.02758706e+00
4.61254090e-01 -1.97078247e-04 -5.79820931e-01 -1.80559888e-01
9.92585957e-01 1.81246370e-01 -4.87237304e-01 -5.23322701e-01
-6.85160637e-01 9.21315253e-01 1.22240090e+00 4.83549625e-01
-3.34947914e-01 3.66883725e-01 1.32985950e-01 7.84213424e-01
5.09496093e-01 5.37453651e-01 -5.68053909e-02 3.15335065e-01
-7.84643412e-01 -1.57024190e-01 1.11207592e+00 7.32228220e-01
9.71721649e-01 2.00157329e-01 1.92244858e-01 7.27412283e-01
3.10410082e-01 4.77135539e-01 3.47647130e-01 -2.83869412e-02
2.03756124e-01 9.46931124e-01 -7.44428873e-01 -1.93378544e+00
-6.23873234e-01 -1.34221172e+00 -1.23630333e+00 -2.98421115e-01
2.53247648e-01 3.82721983e-03 -8.49458218e-01 1.79626834e+00
2.04403967e-01 6.13063514e-01 -2.10429490e-01 7.15326905e-01
1.13964736e+00 5.60001493e-01 8.09267238e-02 2.59846628e-01
8.82668376e-01 -8.03928971e-01 -2.64996767e-01 -2.56690353e-01
5.40572941e-01 -4.16368753e-01 1.05030680e+00 -1.74314491e-02
-8.31445336e-01 -5.42956173e-01 -1.30740166e+00 2.40464598e-01
-5.41023612e-01 -7.83580914e-02 7.42272079e-01 8.35634708e-01
-1.18023288e+00 9.40995097e-01 -7.81702459e-01 -6.40745103e-01
2.40355283e-01 3.81343663e-01 -8.46773088e-02 3.71613838e-02
-1.44862843e+00 5.82439125e-01 4.41388875e-01 7.40730092e-02
-5.19449294e-01 -7.65856564e-01 -6.73161507e-01 4.51487213e-01
2.98045248e-01 -6.90498412e-01 5.43420911e-01 -1.08261216e+00
-1.30550683e+00 4.18753147e-01 2.62704998e-01 -6.47676647e-01
2.15252414e-01 3.39322418e-01 -5.81534445e-01 4.39117908e-01
-1.11794710e-01 3.31068099e-01 9.63126302e-01 -1.04122508e+00
-4.09025908e-01 -4.00934339e-01 -2.16999009e-01 2.78381675e-01
-5.97525060e-01 -6.70712829e-01 -4.06592280e-01 -8.71613204e-01
4.94411886e-01 -9.67932999e-01 -6.01650178e-02 -2.28532538e-01
-4.14995372e-01 -1.99769855e-01 8.11663210e-01 -2.54205197e-01
1.33770359e+00 -2.13852119e+00 1.47001371e-01 9.58828330e-01
8.17321718e-01 1.72920331e-01 -2.66934395e-01 5.42440057e-01
-7.02729151e-02 6.12993538e-02 2.74908096e-01 1.89615920e-01
-8.91355798e-02 1.46285370e-01 -3.76025975e-01 6.49289370e-01
3.16415094e-02 1.13773406e+00 -1.00996566e+00 -2.12566003e-01
1.19446851e-01 6.49981618e-01 -6.62034512e-01 -1.60435081e-01
-1.26452535e-01 -7.92370662e-02 -4.75368947e-01 6.44720614e-01
6.17094636e-01 -8.40833783e-01 4.68225598e-01 -2.30324060e-01
4.33081448e-01 1.52734518e-01 -9.63323176e-01 1.08717716e+00
-1.13667019e-01 7.19804525e-01 4.22558598e-02 -1.18046749e+00
1.38457036e+00 -2.29758117e-02 3.42722774e-01 -6.44211292e-01
1.63492173e-01 3.06589752e-01 2.19862550e-01 -7.76770636e-02
4.06481266e-01 7.70051330e-02 1.04225025e-01 2.37394229e-01
3.31653446e-01 4.01935846e-01 4.71450463e-02 4.27940786e-01
1.55242968e+00 -5.99803865e-01 1.12941571e-01 -6.78964436e-01
2.40375698e-01 -2.60817379e-01 4.83850598e-01 1.02579212e+00
-1.18322589e-01 3.72522205e-01 6.92730069e-01 -7.23352253e-01
-5.98766744e-01 -1.14635277e+00 1.24639131e-01 1.36462450e+00
5.62522054e-01 -6.56359315e-01 -4.77877468e-01 -6.62954092e-01
3.20026875e-01 2.85012007e-01 -8.44526112e-01 -4.53731418e-01
-6.54888570e-01 -9.78644013e-01 6.38946652e-01 5.26070476e-01
6.45314932e-01 -8.33680391e-01 -1.29113242e-01 8.57593417e-02
2.05233663e-01 -8.99358571e-01 -3.56723875e-01 3.38709205e-01
-7.97684431e-01 -1.19486284e+00 -4.87784296e-01 -9.48377430e-01
8.09033096e-01 6.20895028e-01 1.31779325e+00 4.66219753e-01
-5.29946648e-02 2.58208215e-01 -4.01871264e-01 -5.78789897e-02
-2.93873459e-01 6.59616172e-01 -6.43358454e-02 4.81219254e-02
3.96923751e-01 -1.20964515e+00 -8.34637463e-01 5.38172781e-01
-5.74978471e-01 -1.12441845e-01 9.24239695e-01 7.30313003e-01
3.82059932e-01 4.49034423e-01 4.61723745e-01 -1.27330649e+00
7.66549647e-01 -6.22150958e-01 -3.11011106e-01 3.28869194e-01
-8.73260379e-01 -2.41511501e-02 1.11857891e+00 -5.73355258e-01
-5.29512882e-01 -2.47230425e-01 -1.10701909e-02 -4.08415139e-01
2.63541937e-01 7.24878907e-01 -3.88883464e-02 -6.83187604e-01
1.15807843e+00 1.51212186e-01 -8.49574879e-02 -2.13183984e-01
4.18706357e-01 1.76503703e-01 5.46322763e-01 -4.12557900e-01
1.11187375e+00 5.55075228e-01 1.39786571e-01 -1.28751433e+00
-4.30580109e-01 -4.23588485e-01 -3.32266808e-01 -2.77466238e-01
4.30830419e-01 -9.33481336e-01 -6.46466017e-01 2.22086549e-01
-6.42087042e-01 -4.05044079e-01 -1.06154487e-01 1.09772861e-01
-9.39194486e-02 3.26940328e-01 -7.60134935e-01 -3.80820185e-01
-6.03550613e-01 -7.94310331e-01 5.98418713e-01 4.51092601e-01
3.26352417e-02 -1.28503394e+00 2.61871982e-02 -2.48773068e-01
6.73720956e-01 1.83602840e-01 1.07088470e+00 -8.20002854e-01
-8.17634106e-01 -2.78975010e-01 -3.98154706e-01 -1.88463479e-01
-2.22958565e-01 -1.58919729e-02 -6.13664389e-01 -5.37247181e-01
-2.65943587e-01 3.39233354e-02 1.18020928e+00 5.24061203e-01
7.43916094e-01 -3.39476556e-01 -5.86106122e-01 1.10606647e+00
1.37057245e+00 -1.82718545e-01 4.71166313e-01 2.20771700e-01
9.97946322e-01 3.18449646e-01 -2.71660775e-01 -5.48805371e-02
1.74304873e-01 2.33489022e-01 5.89907825e-01 -1.11197896e-01
-1.90980241e-01 -7.29345262e-01 -8.41477513e-03 9.60504711e-01
-5.53853102e-02 -4.69976842e-01 -8.47702742e-01 3.22315186e-01
-1.99973357e+00 -9.16585922e-01 8.52117911e-02 1.55863369e+00
2.22963735e-01 2.73175448e-01 8.54709372e-02 -7.28589520e-02
8.69972885e-01 7.09046423e-01 -6.32382095e-01 -2.68235821e-02
-4.32546169e-01 -8.10728520e-02 6.68367505e-01 2.65340984e-01
-1.05024791e+00 1.06068516e+00 5.98267794e+00 9.18382108e-01
-1.26127112e+00 -2.83655018e-01 5.60983002e-01 2.19422042e-01
-4.83620197e-01 -8.15512240e-02 -6.86873972e-01 3.52923244e-01
6.34462416e-01 -3.13228369e-02 4.86398578e-01 1.16576397e+00
-2.97210366e-01 5.69336891e-01 -7.36960292e-01 8.58128548e-01
1.43532589e-01 -1.65844357e+00 1.28001243e-01 2.61395816e-02
7.88301945e-01 4.03898686e-01 2.12101325e-01 3.69822770e-01
5.29535711e-01 -1.08769023e+00 3.13865602e-01 2.58891553e-01
3.45003277e-01 -7.17858851e-01 4.45399076e-01 1.93016574e-01
-1.68484521e+00 -2.13054195e-01 -7.22707748e-01 7.34286150e-03
-4.11086380e-01 6.93379164e-01 -9.21313643e-01 4.47759628e-01
6.65794075e-01 9.93003905e-01 -8.04512441e-01 1.07236481e+00
-4.26576674e-01 8.33370984e-01 -3.68704855e-01 -4.02718633e-01
3.59560043e-01 -1.78313226e-01 6.68032944e-01 1.28714979e+00
1.61032379e-01 -2.30481461e-01 1.45795196e-01 1.03994703e+00
-2.64105350e-01 1.88750044e-01 -5.82026601e-01 -1.71366453e-01
7.47222841e-01 1.30989516e+00 -1.25930774e+00 1.36969872e-02
-2.69116759e-01 7.01502442e-01 5.74964106e-01 6.74311876e-01
-5.18350482e-01 -5.96349955e-01 4.65913206e-01 3.05510461e-01
6.03343070e-01 -4.48855199e-02 -2.89698809e-01 -1.09190047e+00
-2.45597050e-01 -8.17118406e-01 5.99534750e-01 -2.35112742e-01
-1.63364708e+00 8.48308086e-01 -5.05500495e-01 -8.18194091e-01
4.56081405e-02 -5.70968151e-01 -1.02418327e+00 5.88366270e-01
-1.23506427e+00 -1.22962511e+00 -5.93724966e-01 5.97305059e-01
-5.94334863e-02 -3.76239598e-01 4.86061007e-01 1.42837048e-01
-5.98923862e-01 8.58118713e-01 7.37063885e-02 4.29328501e-01
4.37476814e-01 -1.33530080e+00 8.80872548e-01 9.00629878e-01
4.31474835e-01 8.46246183e-01 8.67662430e-01 -8.81152928e-01
-1.52648818e+00 -1.13526475e+00 5.19404352e-01 -2.15334054e-02
9.38331842e-01 -5.50614595e-01 -1.01034844e+00 5.74066281e-01
-2.49063700e-01 3.28295946e-01 4.55693424e-01 4.55560118e-01
-6.99186981e-01 -2.71308094e-01 -7.99659610e-01 9.28751707e-01
1.41772687e+00 -6.53781652e-01 -1.41044408e-01 2.09327266e-01
7.09216237e-01 -3.13176900e-01 -6.49597764e-01 4.97974604e-01
5.09438753e-01 -1.09719312e+00 1.12453520e+00 -4.33926463e-01
1.33628950e-01 -4.81706485e-02 -1.30169973e-01 -1.35625255e+00
-7.50800550e-01 -9.05368507e-01 -2.83421189e-01 5.93427956e-01
4.86676306e-01 -9.82298791e-01 1.30443966e+00 1.58828303e-01
-3.18974070e-02 -8.77283394e-01 -7.16577053e-01 -7.04172254e-01
-4.23316061e-02 -1.98370665e-02 5.19795597e-01 1.08552170e+00
1.16589768e-02 6.37836277e-01 -2.53776640e-01 2.95316666e-01
5.83743513e-01 3.65705073e-01 7.49686182e-01 -1.59262836e+00
-4.52525645e-01 -7.43226528e-01 -7.87204504e-01 -1.23145807e+00
1.34918854e-01 -1.26781571e+00 -3.62790227e-01 -1.21780992e+00
-1.42275414e-03 -3.72121185e-01 -3.68950695e-01 1.83154970e-01
-1.34092152e-01 2.25726143e-01 9.90140364e-02 2.70099252e-01
-7.52521753e-01 5.55477262e-01 1.17314219e+00 -6.88315332e-02
-5.22776008e-01 7.89775625e-02 -9.58007455e-01 6.87754154e-01
6.06709063e-01 -2.12656647e-01 -6.62871659e-01 -1.64042816e-01
6.99873745e-01 -3.29921633e-01 3.47923011e-01 -1.26811528e+00
5.79900026e-01 1.69557348e-01 4.72544551e-01 -4.85192508e-01
-3.35384756e-02 -6.83852077e-01 1.68887898e-01 5.58203220e-01
-2.02251121e-01 1.93926662e-01 -2.02442914e-01 1.22451317e+00
-7.30459811e-04 2.61659324e-01 8.23157251e-01 -1.23838827e-01
-5.45759499e-01 5.58697701e-01 -1.61051899e-01 1.43328682e-01
6.25894368e-01 -2.87838429e-01 -8.53058219e-01 -5.41430831e-01
-4.85008597e-01 7.11860135e-02 4.99122113e-01 2.56436199e-01
4.86439556e-01 -1.25752163e+00 -3.92277628e-01 4.53512222e-01
-1.62183270e-01 -3.32328498e-01 2.67598629e-01 7.35659897e-01
-6.10406041e-01 1.91801488e-01 -1.45122349e-01 -7.05881476e-01
-1.03066027e+00 4.73476231e-01 4.84548330e-01 -3.13660055e-01
-1.09862804e+00 1.20949197e+00 2.47873217e-01 -3.91119897e-01
3.75477463e-01 -1.65079162e-01 -7.41452053e-02 -8.66991505e-02
1.25652835e-01 2.92045325e-01 1.28836557e-01 -5.67093492e-01
-2.05330640e-01 4.77131099e-01 -1.77419230e-01 5.87927639e-01
1.30602634e+00 5.79183213e-02 -3.23930569e-02 1.08249873e-01
1.17415810e+00 -1.22043200e-01 -1.03023410e+00 -4.60211813e-01
4.05335240e-02 -2.94447273e-01 -7.58390278e-02 -3.52816820e-01
-1.34214830e+00 6.40533864e-01 2.83102542e-01 7.46724725e-01
1.10937870e+00 3.48023847e-02 8.72110844e-01 6.81661844e-01
1.80201679e-01 -1.01639366e+00 3.64013463e-01 6.51768684e-01
6.08309507e-01 -8.48004639e-01 -5.32700587e-03 -5.95869780e-01
-2.35917494e-01 1.00881577e+00 5.59386134e-01 -8.12351644e-01
1.08639240e+00 5.82621172e-02 -1.40915483e-01 -6.39732242e-01
-7.11126268e-01 -4.95799407e-02 5.37608147e-01 5.72000623e-01
1.82810202e-01 1.44955561e-01 1.43326167e-02 5.61700642e-01
-5.89488506e-01 -7.53422737e-01 8.57701302e-02 5.26119530e-01
-7.56605387e-01 -6.30314767e-01 -1.28443105e-04 7.15951562e-01
-2.07847267e-01 -1.32292226e-01 -7.48565257e-01 8.96154761e-01
-2.45086432e-01 5.52554131e-01 -6.14766078e-03 -1.01123512e+00
1.62224442e-01 -2.05085322e-01 8.84696990e-02 -4.33898717e-01
-7.05088675e-01 3.70733673e-03 -2.11898178e-01 -5.61158836e-01
-6.27211249e-03 -1.39722720e-01 -9.62106526e-01 -6.25528634e-01
-4.26803619e-01 1.67158678e-01 3.75216484e-01 7.16857374e-01
5.01065016e-01 5.22616565e-01 4.15971369e-01 -7.20828652e-01
-4.51781690e-01 -8.50457132e-01 -8.13756883e-01 3.59520972e-01
1.49583131e-01 -6.26123190e-01 -7.24767327e-01 -6.92785859e-01] | [6.9641900062561035, 6.131599426269531] |
3937a5ce-07a2-4ff2-aeef-c0aed54a21e1 | dreamtime-an-improved-optimization-strategy | 2306.12422 | null | https://arxiv.org/abs/2306.12422v1 | https://arxiv.org/pdf/2306.12422v1.pdf | DreamTime: An Improved Optimization Strategy for Text-to-3D Content Creation | Text-to-image diffusion models pre-trained on billions of image-text pairs have recently enabled text-to-3D content creation by optimizing a randomly initialized Neural Radiance Fields (NeRF) with score distillation. However, the resultant 3D models exhibit two limitations: (a) quality concerns such as saturated color and the Janus problem; (b) extremely low diversity comparing to text-guided image synthesis. In this paper, we show that the conflict between NeRF optimization process and uniform timestep sampling in score distillation is the main reason for these limitations. To resolve this conflict, we propose to prioritize timestep sampling with monotonically non-increasing functions, which aligns NeRF optimization with the sampling process of diffusion model. Extensive experiments show that our simple redesign significantly improves text-to-3D content creation with higher quality and diversity. | ['Lei Zhang', 'Zheng-Jun Zha', 'Xianbiao Qi', 'Yukai Shi', 'Jianan Wang', 'Yukun Huang'] | 2023-06-21 | null | null | null | null | ['text-to-3d'] | ['computer-vision'] | [ 4.55756247e-01 -5.27012460e-02 9.97799560e-02 -2.63875306e-01
-7.92236507e-01 -5.42308271e-01 1.04416752e+00 -2.47739494e-01
-2.44790986e-01 7.02355683e-01 6.67391300e-01 -3.05826098e-01
-3.14586200e-02 -8.13838124e-01 -5.59525371e-01 -7.16644108e-01
1.68896690e-01 2.11191684e-01 1.46236330e-01 -2.99285799e-01
6.38422608e-01 5.17934859e-01 -1.60052359e+00 2.79479057e-01
1.17312837e+00 8.67110014e-01 3.88781995e-01 1.12067890e+00
-4.32061970e-01 1.18440616e+00 -9.00453091e-01 -2.23228455e-01
6.42759979e-01 -9.51147735e-01 -4.79832023e-01 -4.59786244e-02
8.09032381e-01 -7.15420187e-01 -2.69191563e-01 1.00823104e+00
6.54337466e-01 2.63929337e-01 1.00990629e+00 -1.01007152e+00
-1.04702818e+00 2.38533258e-01 -8.35955977e-01 -1.51299283e-01
1.80371135e-01 2.95565724e-01 1.06553292e+00 -7.88018823e-01
8.58236849e-01 1.22392595e+00 5.95020354e-01 5.96253633e-01
-1.22697103e+00 -3.17643523e-01 -1.37135372e-01 -2.40215153e-01
-1.33041263e+00 -2.81746954e-01 5.99027932e-01 -3.63677293e-01
9.92289424e-01 5.98085284e-01 9.52945650e-01 9.19490159e-01
2.83052474e-01 8.39104056e-01 1.19820714e+00 -5.07372797e-01
2.70144254e-01 3.72309834e-01 -6.39032304e-01 7.72062182e-01
-2.86026788e-03 5.06205931e-02 -5.13651967e-01 -8.21591094e-02
1.15895653e+00 -4.05510217e-01 -1.19537756e-01 -2.17652008e-01
-1.20311284e+00 9.23509657e-01 4.92178649e-01 1.08722791e-01
-2.37499103e-01 6.00035667e-01 4.08004113e-02 2.77850598e-01
7.04797208e-01 9.03101027e-01 -1.20949596e-01 -1.50961146e-01
-1.28146040e+00 4.32201207e-01 4.80422288e-01 8.07138383e-01
7.62175858e-01 4.42846715e-01 -4.28876162e-01 1.02774370e+00
2.25282922e-01 8.88826609e-01 4.30035114e-01 -1.26980698e+00
2.44157150e-01 2.77572483e-01 1.36790723e-01 -7.45753586e-01
-6.21722862e-02 -3.24729592e-01 -6.69981062e-01 6.21684968e-01
3.16105574e-01 -2.19904646e-01 -1.22476602e+00 1.54024458e+00
2.24743143e-01 -2.13779032e-01 -1.26332387e-01 1.01216567e+00
5.16836524e-01 1.09604156e+00 -1.93070143e-01 1.33805543e-01
1.05574954e+00 -1.14440584e+00 -6.91991150e-01 -1.52576715e-02
7.39068925e-01 -9.84291434e-01 1.46027696e+00 1.23518459e-01
-1.27242470e+00 -2.82600760e-01 -1.03908408e+00 -2.41753235e-01
-1.73909038e-01 -9.92865264e-02 6.18970037e-01 8.59420776e-01
-1.37601149e+00 6.01480961e-01 -1.68441430e-01 -3.18082482e-01
2.40158677e-01 -3.55024934e-02 7.28830323e-02 8.53619352e-02
-1.02843845e+00 8.12413216e-01 -1.03874274e-01 -5.55894613e-01
-6.07140958e-01 -7.88807333e-01 -6.11910641e-01 -6.75122291e-02
-1.52990231e-02 -9.02103364e-01 1.22987664e+00 -1.15724218e+00
-1.88517189e+00 8.25423360e-01 -1.40352594e-02 -3.33983153e-01
7.08286047e-01 -1.37120485e-01 -5.42530743e-03 1.65611148e-01
7.19616339e-02 1.22434032e+00 9.03598011e-01 -1.39596093e+00
-6.95233047e-01 4.84803356e-02 -9.39831808e-02 6.81778431e-01
-5.61607599e-01 -2.73324192e-01 -4.53247160e-01 -9.13425505e-01
1.30642916e-03 -6.70094192e-01 -2.65923381e-01 3.36917281e-01
-3.40784341e-01 2.31557027e-01 6.74232841e-01 -4.00308788e-01
1.20994818e+00 -2.01347017e+00 -3.30438390e-02 1.14528254e-01
2.89881855e-01 -1.47332907e-01 -5.65175891e-01 2.51302123e-01
2.15473190e-01 3.56562734e-01 -2.09293917e-01 -3.54286671e-01
3.68553624e-02 -6.47826446e-03 -4.29425061e-01 2.58414298e-01
2.12698832e-01 1.00593507e+00 -9.35909688e-01 -6.33410215e-01
3.28577071e-01 4.99566555e-01 -8.71798635e-01 8.26934054e-02
-5.41836560e-01 3.11854631e-01 -3.24076235e-01 4.47143018e-01
8.10628951e-01 -2.03055397e-01 -2.81970352e-01 -3.75958420e-02
-4.43563968e-01 2.79551029e-01 -9.62175488e-01 1.70287442e+00
-4.40663099e-01 1.08725548e+00 -1.35424271e-01 -2.39249393e-01
8.88604224e-01 -1.08592428e-01 7.57621706e-01 -1.12925828e+00
-3.00234947e-02 2.09157810e-01 -3.79287452e-01 -3.25016230e-01
1.26090384e+00 -1.84311166e-01 1.00734517e-01 5.51634669e-01
-2.15469748e-01 -1.00835681e+00 1.87108144e-01 3.81239742e-01
9.81083989e-01 1.98016062e-01 -3.20454240e-01 -3.97327334e-01
-3.82544799e-03 2.66927749e-01 1.75736561e-01 1.03260362e+00
2.57357527e-02 1.17613947e+00 4.96815741e-01 -2.12097883e-01
-1.83156800e+00 -1.13339710e+00 -3.68620902e-02 8.68692279e-01
2.01842055e-01 -4.10739213e-01 -9.22407210e-01 -4.03988481e-01
-1.43093362e-01 1.03572953e+00 -6.40898824e-01 1.24451295e-01
-4.60128516e-01 -8.17291915e-01 6.74431384e-01 7.81461224e-02
6.45063758e-01 -4.80508834e-01 -6.64271474e-01 1.74363837e-01
-1.73822954e-01 -7.53531396e-01 -8.74017358e-01 1.38233051e-01
-8.28104138e-01 -3.94745260e-01 -1.50230956e+00 -5.19420028e-01
7.33301878e-01 6.05171084e-01 1.17381525e+00 -4.69668321e-02
-2.76642293e-01 2.58075565e-01 -2.50654846e-01 -3.26289445e-01
-6.15208745e-01 -4.38775793e-02 -2.70274013e-01 -2.27192223e-01
-2.33474925e-01 -2.54886389e-01 -1.00969946e+00 3.30180913e-01
-1.24330723e+00 4.88711208e-01 5.20303845e-01 7.14201808e-01
5.05167961e-01 -1.22982152e-01 2.17098475e-01 -4.40576345e-01
8.62613916e-01 -8.83857682e-02 -7.44042993e-01 7.67908022e-02
-8.76299798e-01 3.92878562e-01 5.28358579e-01 -5.65110207e-01
-1.25057256e+00 -7.44252726e-02 -1.04849465e-01 -3.45351756e-01
1.67616352e-01 6.45856485e-02 4.78947878e-01 -1.49433509e-01
1.03309476e+00 2.90514380e-01 -1.32477000e-01 -1.06038563e-01
7.42087662e-01 4.92486984e-01 2.88763493e-01 -5.87975323e-01
9.21074510e-01 6.13768876e-01 -1.34191528e-01 -9.05068398e-01
-4.57626522e-01 3.73055674e-02 -2.08667777e-02 -4.77575809e-01
1.08354020e+00 -9.49714303e-01 -1.53328478e-01 4.64133292e-01
-1.11726451e+00 -9.07119751e-01 -7.12556481e-01 3.05838317e-01
-5.49089491e-01 2.16421261e-01 -6.68925405e-01 -9.03922200e-01
-3.45831752e-01 -1.04503357e+00 1.27636254e+00 2.83592403e-01
-1.22359000e-01 -8.14957321e-01 2.85390258e-01 2.19206914e-01
9.30393636e-01 2.42562443e-01 9.61627781e-01 3.47274423e-01
-8.89745414e-01 1.97230905e-01 -5.20901799e-01 2.30659544e-01
-1.82555281e-02 1.57490417e-01 -9.81986761e-01 2.00999379e-02
-1.50617838e-01 -2.76049465e-01 8.21980536e-01 6.67299211e-01
8.20737183e-01 -2.68988937e-01 9.20597911e-02 8.39469552e-01
1.45568204e+00 9.90359411e-02 9.58844483e-01 4.74353462e-01
5.49933553e-01 5.38772047e-01 3.42003256e-01 5.56797683e-01
2.40288779e-01 5.69650590e-01 2.22386420e-01 -5.34782052e-01
-7.34148502e-01 -4.53565389e-01 4.03900057e-01 6.53170884e-01
1.25100598e-01 -6.05116904e-01 -5.26265740e-01 5.90706825e-01
-1.52135861e+00 -9.22234416e-01 -2.32301801e-01 2.13252521e+00
9.10496712e-01 8.26200098e-02 -1.07950859e-01 -2.93229252e-01
4.94769037e-01 5.04819632e-01 -5.38528681e-01 -4.58007216e-01
-6.07749283e-01 -2.66846921e-02 9.40438509e-01 7.86388755e-01
-4.77075100e-01 9.88785207e-01 6.80325794e+00 1.17006564e+00
-1.23425817e+00 5.41973896e-02 9.60058153e-01 -5.57256162e-01
-1.11617732e+00 -1.19273178e-01 -5.28070927e-01 2.60473013e-01
6.31986260e-01 -5.39521798e-02 7.27572560e-01 3.77175480e-01
4.58680987e-01 -2.86821842e-01 -6.11926377e-01 1.01128638e+00
1.43949687e-01 -1.52682817e+00 3.27072054e-01 1.93876311e-01
1.47490335e+00 1.69838890e-01 4.25971568e-01 2.23680492e-02
6.27541542e-01 -9.53601241e-01 1.13107562e+00 4.16551054e-01
1.13903201e+00 -5.56197882e-01 -7.93170929e-02 7.49127613e-03
-9.58269596e-01 1.08749039e-01 -2.99351871e-01 3.08041602e-01
3.27557892e-01 1.02253926e+00 -7.01310873e-01 1.59438998e-02
5.52705109e-01 2.11894274e-01 -4.10840154e-01 9.27414894e-01
-1.82642326e-01 3.24513286e-01 -4.16951537e-01 -4.13119733e-01
5.13048530e-01 -3.77503037e-01 7.02839077e-01 1.17301893e+00
8.26071620e-01 -6.96949437e-02 -4.65512842e-01 1.34061837e+00
-9.59349349e-02 1.26427829e-01 -8.17379236e-01 1.29435183e-02
3.13579708e-01 1.03155231e+00 -7.32412875e-01 -4.47400242e-01
-1.37547836e-01 1.40698409e+00 -9.95062813e-02 5.34394383e-01
-1.02742720e+00 -6.14297152e-01 6.21286452e-01 3.10428172e-01
2.15602368e-01 -3.64952922e-01 -6.70383811e-01 -1.10405493e+00
-2.40509287e-01 -7.71323979e-01 -2.26192579e-01 -1.46155798e+00
-9.08370018e-01 5.52121282e-01 -2.00662300e-01 -1.04016995e+00
-1.24985851e-01 -2.96126723e-01 -4.76089746e-01 1.04415774e+00
-1.69446313e+00 -8.82678449e-01 -3.28212619e-01 3.31480861e-01
7.83660650e-01 -3.45091037e-02 4.58815515e-01 2.00484082e-01
-2.25845948e-02 4.87271875e-01 5.56169212e-01 -2.84304202e-01
8.56622577e-01 -1.38947403e+00 8.72355461e-01 7.20188081e-01
-6.74189031e-02 1.37639627e-01 6.69010580e-01 -6.83135450e-01
-1.32310688e+00 -9.54130232e-01 8.46851766e-01 -4.70459074e-01
3.69217336e-01 -4.90125537e-01 -5.43876290e-01 -1.46658629e-01
4.88692909e-01 -7.03768194e-01 1.70058668e-01 -4.42791671e-01
-2.79565513e-01 4.02664430e-02 -1.18796623e+00 1.30789793e+00
9.70693648e-01 -5.31055272e-01 3.82032692e-02 3.89544755e-01
8.38051915e-01 -4.18517768e-01 -7.07580626e-01 1.33636460e-01
4.85784650e-01 -1.07364869e+00 9.49058950e-01 1.28147140e-01
8.55356157e-01 -3.67776990e-01 -2.85074353e-01 -1.35928941e+00
-4.25169677e-01 -9.04021859e-01 1.80432960e-01 9.06923592e-01
5.92771173e-01 -3.94730031e-01 9.40989077e-01 5.79433262e-01
1.11307041e-03 -5.61612785e-01 -6.41313314e-01 -5.94425261e-01
2.37365678e-01 -1.92031801e-01 7.29051948e-01 8.65982056e-01
-4.09711629e-01 3.67714137e-01 -5.61459839e-01 -3.03923130e-01
6.88790202e-01 -2.08241448e-01 8.09496582e-01 -7.68449068e-01
-2.86314845e-01 -9.11582172e-01 1.78757608e-01 -1.68312752e+00
-5.35119653e-01 -5.87733805e-01 2.65292972e-01 -1.62697911e+00
1.78568751e-01 -6.58152461e-01 2.35550344e-01 1.17631570e-01
-1.75501734e-01 3.66028845e-01 2.58389205e-01 3.12591344e-01
-2.50081390e-01 7.92842805e-01 1.81501222e+00 -3.78870994e-01
-4.33628678e-01 -6.61353707e-01 -7.75758386e-01 1.60989627e-01
7.64562666e-01 -4.66406345e-01 -5.83976686e-01 -9.46565866e-01
5.60667694e-01 3.22450362e-02 1.71417773e-01 -9.67896044e-01
2.04990488e-02 -3.86602402e-01 3.85898501e-01 -5.15016615e-01
4.45161223e-01 -4.74572062e-01 -2.99945865e-02 3.78369659e-01
-5.41001856e-01 2.56870687e-02 1.39348388e-01 3.35504621e-01
1.10746287e-01 -4.28293109e-01 8.59915316e-01 -1.72579855e-01
-4.76469725e-01 4.07077223e-02 -5.98835349e-01 -4.00138572e-02
6.35751963e-01 -5.99008501e-01 -4.77819264e-01 -8.09294820e-01
1.01822428e-01 -3.65882255e-02 8.40734601e-01 3.73820752e-01
5.36924243e-01 -1.48452997e+00 -9.38190222e-01 1.85176790e-01
4.53595817e-03 -9.55364406e-02 2.19265074e-01 5.56395710e-01
-8.27865362e-01 2.41823658e-01 4.00942229e-02 -6.69246674e-01
-9.84935820e-01 -8.07310939e-02 4.23628390e-01 -1.91230118e-01
-5.86892247e-01 9.29578304e-01 2.97843039e-01 -4.97466117e-01
2.48955831e-01 -2.83175439e-01 2.76240528e-01 -2.85307735e-01
3.53501916e-01 1.67329580e-01 -9.48120505e-02 -1.69701993e-01
1.54790238e-01 6.84237838e-01 5.02206795e-02 -7.54847050e-01
1.14781785e+00 -3.07782501e-01 1.37729555e-01 -1.04690880e-01
1.22091401e+00 1.31762847e-01 -1.74395549e+00 -7.56685063e-02
-5.36824524e-01 -7.25640893e-01 5.18294454e-01 -9.02826965e-01
-1.00731432e+00 8.03457201e-01 7.70484567e-01 2.48845160e-01
1.01342905e+00 -3.37285519e-01 1.05454910e+00 3.58038157e-01
-2.60328222e-02 -1.50269425e+00 4.84039515e-01 6.30169868e-01
7.45994806e-01 -9.06328142e-01 3.80003080e-02 1.26691028e-01
-7.96057284e-01 8.94374669e-01 4.34645534e-01 1.55844599e-01
1.79046795e-01 2.99889535e-01 2.49827445e-01 -8.40926170e-02
-7.37524390e-01 -1.58140600e-01 1.21082239e-01 5.99483669e-01
5.34015417e-01 -2.18867153e-01 -1.86062023e-01 -3.90634418e-01
-2.45124355e-01 -9.42053273e-02 5.87458193e-01 7.94874728e-01
-6.95833504e-01 -8.48241866e-01 -6.22705698e-01 3.90311509e-01
-1.30707681e-01 -5.40535212e-01 -3.94572377e-01 4.06809658e-01
-1.17133394e-01 8.45886827e-01 2.03430817e-01 -3.48962545e-01
1.57319054e-01 -2.40438908e-01 6.15641773e-01 -9.61812139e-02
-5.88949740e-01 3.28407764e-01 4.54633832e-02 -3.78586918e-01
-1.35824606e-01 -3.10831487e-01 -1.13307834e+00 -5.87909937e-01
-4.46533382e-01 -2.36539133e-02 1.13208425e+00 4.53928113e-01
5.55478334e-01 5.66699028e-01 8.51036429e-01 -8.16327572e-01
-4.81873333e-01 -7.97976494e-01 -6.35681033e-01 4.54668939e-01
2.21516207e-01 -1.28511846e-01 -4.44335669e-01 3.33394594e-02] | [11.323905944824219, -0.35568946599960327] |
bd207645-22c8-445e-b4d9-c2a6db3b59c4 | polarized-reflection-removal-with-perfect | 2003.12789 | null | https://arxiv.org/abs/2003.12789v1 | https://arxiv.org/pdf/2003.12789v1.pdf | Polarized Reflection Removal with Perfect Alignment in the Wild | We present a novel formulation to removing reflection from polarized images in the wild. We first identify the misalignment issues of existing reflection removal datasets where the collected reflection-free images are not perfectly aligned with input mixed images due to glass refraction. Then we build a new dataset with more than 100 types of glass in which obtained transmission images are perfectly aligned with input mixed images. Second, capitalizing on the special relationship between reflection and polarized light, we propose a polarized reflection removal model with a two-stage architecture. In addition, we design a novel perceptual NCC loss that can improve the performance of reflection removal and general image decomposition tasks. We conduct extensive experiments, and results suggest that our model outperforms state-of-the-art methods on reflection removal. | ['Xuhua Huang', 'Wenxiu Sun', 'Qifeng Chen', 'Mengdi Zhang', 'Qiong Yan', 'Chenyang Lei'] | 2020-03-28 | polarized-reflection-removal-with-perfect-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Lei_Polarized_Reflection_Removal_With_Perfect_Alignment_in_the_Wild_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Lei_Polarized_Reflection_Removal_With_Perfect_Alignment_in_the_Wild_CVPR_2020_paper.pdf | cvpr-2020-6 | ['reflection-removal'] | ['computer-vision'] | [ 8.71944010e-01 -6.68767467e-02 4.93957222e-01 -1.50128648e-01
-6.01520777e-01 -2.43200749e-01 3.30769718e-01 -1.03750587e+00
3.07372641e-02 3.14214885e-01 4.41378623e-01 -1.00534640e-01
9.72369537e-02 -7.82851577e-01 -7.53461897e-01 -1.18084884e+00
5.39459586e-01 -1.58832833e-01 1.40146613e-01 -3.79553437e-01
2.47931123e-01 1.25212267e-01 -1.42382753e+00 8.20221245e-01
9.16417837e-01 7.42163956e-01 2.29024366e-01 6.06547356e-01
1.03051774e-01 7.65910387e-01 -4.62920636e-01 -5.05383551e-01
8.20367277e-01 -2.97171891e-01 -6.15696013e-01 4.36682731e-01
7.35899746e-01 -6.51329458e-01 -4.12708461e-01 1.22657037e+00
4.65763271e-01 -1.14960864e-01 4.69059229e-01 -9.64914382e-01
-1.19873083e+00 3.66271168e-01 -9.41848278e-01 -1.32662416e-01
4.10402685e-01 2.17782736e-01 6.45861924e-01 -8.75936270e-01
4.80813801e-01 1.26138294e+00 6.52890563e-01 4.78443176e-01
-9.84675229e-01 -6.27709091e-01 4.03058171e-01 1.68318935e-02
-9.30980682e-01 -7.15194583e-01 1.01163745e+00 -1.09461449e-01
9.33200061e-01 4.72110063e-01 6.02911055e-01 1.05683005e+00
1.39289245e-01 5.11637390e-01 1.31010199e+00 -5.64370334e-01
-4.53833073e-01 -4.37524244e-02 6.53182417e-02 4.77385402e-01
5.67909658e-01 1.61280215e-01 -5.41277230e-01 2.30440840e-01
3.91846865e-01 1.46809489e-01 -6.91096306e-01 -8.58904719e-02
-1.08782411e+00 -1.15500605e-02 3.03279579e-01 -1.63652316e-01
-1.38136953e-01 -1.82159320e-01 -1.76138118e-01 3.48112255e-01
6.61363304e-01 3.88202906e-01 -1.07111290e-01 6.43795669e-01
-5.50392330e-01 -5.63254245e-02 5.95826328e-01 1.14446521e+00
7.50154853e-01 7.36788958e-02 -2.97976047e-01 1.13425899e+00
4.95230764e-01 1.07925522e+00 6.54505864e-02 -1.00816596e+00
8.10330689e-01 4.01256144e-01 2.08233058e-01 -1.07696950e+00
-2.65459388e-01 -4.11006659e-01 -9.88637686e-01 9.11011994e-02
9.68215987e-03 1.17163844e-01 -1.04951739e+00 1.33057237e+00
2.06309110e-01 2.54049242e-01 5.16492248e-01 9.91803885e-01
1.16067231e+00 7.29167581e-01 -5.35984755e-01 -4.74874139e-01
1.34727705e+00 -1.45521975e+00 -7.99825609e-01 -3.35653841e-01
-4.39463668e-02 -1.32849228e+00 9.61109161e-01 8.19991350e-01
-1.15988421e+00 -4.87124324e-01 -1.22333610e+00 -2.73060173e-01
1.64335415e-01 2.25468040e-01 7.18175650e-01 8.82298350e-01
-9.04488981e-01 2.61508465e-01 -4.53839928e-01 -3.04947775e-02
1.61637276e-01 6.45713657e-02 -1.35733500e-01 -6.19751930e-01
-6.74417973e-01 5.27659059e-01 -3.06942731e-01 6.76540494e-01
-6.24440491e-01 -5.45127332e-01 -5.29295385e-01 -3.53890628e-01
4.57766801e-01 -9.46886897e-01 8.24940085e-01 -6.46755874e-01
-1.73856747e+00 9.62317884e-01 -3.31749082e-01 9.81845632e-02
3.84372324e-01 -5.67948103e-01 -7.09766567e-01 2.58708388e-01
-2.28189513e-01 -4.05458175e-02 1.24933028e+00 -2.09546590e+00
-3.58371794e-01 -1.64012626e-01 2.17575729e-01 5.10722816e-01
-2.70508915e-01 1.45223113e-02 -8.94977510e-01 -4.06838357e-01
7.46850729e-01 -1.03587210e+00 -1.60576507e-01 -5.33861697e-01
-1.01953733e+00 6.79163098e-01 8.65499735e-01 -5.20212650e-01
8.84462953e-01 -2.20240831e+00 -2.62286086e-02 8.79672915e-02
5.53649724e-01 4.25921619e-01 -5.76214373e-01 3.96522433e-01
-3.24014574e-01 -1.82788730e-01 -2.34589025e-01 -7.06027269e-01
-1.82639137e-01 -5.31394966e-02 -7.26839423e-01 3.66521478e-01
-1.18528128e-01 5.37736714e-01 -6.02399111e-01 -8.65287930e-02
1.32200882e-01 6.59637630e-01 -7.62670994e-01 2.59328485e-01
1.73819795e-01 6.80118561e-01 -2.47174874e-01 8.55639756e-01
1.62119758e+00 -9.69938636e-02 2.82780170e-01 -1.03814995e+00
-1.56707734e-01 3.97283345e-01 -1.08354580e+00 1.25405645e+00
-4.37054068e-01 3.02657127e-01 6.22249991e-02 -7.64117539e-01
8.13665211e-01 1.20106824e-02 6.78276122e-01 -1.19286203e+00
-1.54192030e-01 1.83585837e-01 -9.59877223e-02 -6.79299772e-01
7.78985798e-01 -9.88376439e-02 4.49840933e-01 5.01640081e-01
-4.48778272e-01 -3.93725574e-01 3.43991071e-02 1.14704400e-01
1.02684474e+00 2.93925405e-01 -3.62034023e-01 1.59402400e-01
5.66002965e-01 -2.95425355e-01 7.16177642e-01 7.61741281e-01
1.27149701e-01 1.34027541e+00 1.66723520e-01 -4.23765898e-01
-8.97840083e-01 -1.25444329e+00 2.19950452e-03 7.10709870e-01
6.75562203e-01 -4.60774302e-01 -6.72956824e-01 -2.10494563e-01
-4.51619625e-01 1.71363756e-01 -6.50175437e-02 -1.38604522e-01
-7.93366432e-01 -1.37236989e+00 1.79289490e-01 8.19776356e-02
1.08519018e+00 -7.52540767e-01 -3.51546034e-02 -1.59050763e-01
-7.06145644e-01 -1.63301206e+00 -2.75331914e-01 -2.04889014e-01
-6.96168065e-01 -1.41348898e+00 -5.26260316e-01 -6.92263961e-01
1.02608693e+00 1.16007876e+00 1.39891398e+00 3.08311164e-01
-2.46271446e-01 5.51866949e-01 -3.95944864e-01 -3.60561520e-01
-1.38220370e-01 -4.11986560e-01 -8.55484456e-02 2.17776820e-01
2.36637875e-01 -6.63103998e-01 -1.05203497e+00 4.47287261e-01
-1.06553841e+00 5.11859417e-01 7.30538189e-01 4.99640673e-01
6.64072156e-01 3.54125053e-01 -2.54564762e-01 -1.17825258e+00
5.14478743e-01 -1.35023698e-01 -4.40824687e-01 2.75434583e-01
-4.75035310e-01 -1.99103773e-01 5.60692012e-01 -2.97665030e-01
-1.52700770e+00 -1.92919880e-01 -1.77606903e-02 -2.37649977e-01
1.65482506e-01 -6.13792762e-02 -3.01522911e-01 -3.38981390e-01
4.02085602e-01 2.03718126e-01 -3.15487891e-01 -5.40215015e-01
1.50186658e-01 5.58185518e-01 6.11490846e-01 -5.92222929e-01
1.11274421e+00 1.04280841e+00 2.64074326e-01 -9.33085740e-01
-1.14319861e+00 -3.48335743e-01 -3.17596346e-01 -1.10223368e-01
6.38771713e-01 -1.09089363e+00 -8.30276191e-01 1.02758563e+00
-1.11689258e+00 -2.25587353e-01 2.89883763e-02 4.82874483e-01
-3.69014174e-01 7.78517127e-01 -7.23886371e-01 -6.79345965e-01
-3.79607975e-01 -1.38481820e+00 1.12903118e+00 1.45682748e-02
4.80989665e-01 -3.97181243e-01 1.87422350e-01 9.83699262e-01
4.52849329e-01 -2.73374796e-01 7.30397403e-01 3.23952377e-01
-1.14161718e+00 3.22208881e-01 -5.59006691e-01 7.41494477e-01
2.89900929e-01 6.08248562e-02 -1.15968227e+00 -3.29663396e-01
4.26303387e-01 9.37443227e-02 1.38865304e+00 3.62798691e-01
8.74148309e-01 -3.04604173e-01 -2.42026344e-01 1.31588578e+00
1.39592218e+00 9.25888121e-02 1.38527226e+00 4.96510804e-01
1.05372441e+00 6.50491059e-01 5.52237451e-01 1.19305566e-01
4.45966601e-01 5.51183641e-01 6.53933167e-01 -4.82278675e-01
-6.26959503e-01 9.96745005e-02 4.11216825e-01 1.08010411e+00
-5.11561632e-01 -5.81227779e-01 -4.40532565e-01 2.14262798e-01
-1.62695014e+00 -1.00548589e+00 -6.05762124e-01 2.12320018e+00
5.13167858e-01 -2.22533420e-01 -4.28764403e-01 1.89617604e-01
4.95903820e-01 5.20900071e-01 -2.59992331e-01 -1.26670584e-01
-5.28208613e-01 1.28388762e-01 6.69354975e-01 5.85031807e-01
-9.74270225e-01 8.25040102e-01 7.22283268e+00 2.12338299e-01
-1.12501228e+00 -1.37172773e-01 4.31346983e-01 -9.10521522e-02
-6.84564948e-01 4.76154312e-02 -6.87366605e-01 3.80797744e-01
7.88803250e-02 5.81297576e-01 7.00337470e-01 1.22338980e-01
1.25620633e-01 -7.97436684e-02 -8.79439294e-01 1.23742938e+00
4.41857636e-01 -8.47912908e-01 1.92579284e-01 2.44870465e-02
9.47615564e-01 -2.85015292e-02 5.21165967e-01 -2.12223127e-01
2.80061871e-01 -8.87617886e-01 4.00510877e-01 7.45375037e-01
5.96407115e-01 -3.00714731e-01 3.83303195e-01 -1.55638799e-01
-9.75409508e-01 9.78730153e-03 -5.94800651e-01 1.53028905e-01
1.79175869e-01 1.04421008e+00 -2.87297904e-01 9.78247941e-01
1.01614904e+00 8.85168612e-01 -1.51577130e-01 7.62219548e-01
-4.91468757e-01 3.98265809e-01 -3.10253799e-01 7.57286310e-01
-2.30006263e-01 -8.79071772e-01 7.75738060e-01 1.00880289e+00
3.74330074e-01 2.66791046e-01 1.42689068e-02 7.81072676e-01
-4.21910584e-02 -2.05966368e-01 -7.31083035e-01 4.10630018e-01
8.24126452e-02 1.34846330e+00 -2.46887177e-01 -7.53054172e-02
-7.67923594e-01 1.12552488e+00 -7.60686919e-02 9.68667030e-01
-7.97922909e-01 -3.51579785e-02 8.84567440e-01 1.39515355e-01
2.67675787e-01 -2.05097124e-01 -1.36236310e-01 -1.60937405e+00
5.18107355e-01 -9.41852927e-01 4.22343193e-03 -1.26440728e+00
-1.35599411e+00 4.74303782e-01 -1.97506830e-01 -1.56570137e+00
4.47987854e-01 -8.53256941e-01 -5.22270918e-01 7.99170196e-01
-2.30371904e+00 -1.35473633e+00 -6.82855248e-01 7.86087155e-01
9.93860215e-02 6.64216205e-02 4.45866108e-01 5.76368570e-01
-6.26073778e-01 1.92583129e-01 2.20097870e-01 -3.10805254e-03
9.88472164e-01 -7.70967424e-01 2.49063551e-01 1.25669873e+00
-6.35484532e-02 9.07608747e-01 7.15029299e-01 -4.45142835e-01
-1.90056527e+00 -1.02382600e+00 2.59825587e-01 -1.71446875e-01
6.81284964e-02 -1.50177181e-01 -5.92678845e-01 9.10008430e-01
3.63038242e-01 -4.34464030e-02 7.75204360e-01 -9.12735146e-03
-7.84259796e-01 -4.90147561e-01 -1.04203713e+00 7.84621596e-01
1.45803368e+00 -4.66004640e-01 -2.76194125e-01 5.24494290e-01
6.92502856e-01 -6.92759633e-01 -3.78532141e-01 6.15736902e-01
6.71004117e-01 -1.59441972e+00 1.56499660e+00 -1.11053161e-01
7.23545909e-01 -5.11221349e-01 -3.38200629e-01 -1.22978771e+00
-4.39831913e-01 -8.00654948e-01 1.88697711e-01 1.15602744e+00
2.57789820e-01 -1.01235783e+00 5.89592755e-01 5.18262088e-01
-4.97383267e-01 -1.06116466e-01 -2.30963349e-01 -5.56544602e-01
-3.99570286e-01 -3.95325810e-01 6.42057300e-01 8.51930022e-01
-5.26138902e-01 3.69347364e-01 -8.62937689e-01 6.53769314e-01
1.08532608e+00 6.04521990e-01 9.87424254e-01 -7.95352280e-01
-5.11684477e-01 9.66605637e-03 5.68294525e-02 -1.44576800e+00
-1.09036930e-01 -4.73853827e-01 1.41048685e-01 -1.61590707e+00
6.09237254e-01 -3.32897693e-01 -3.51615608e-01 3.20525050e-01
-1.57128185e-01 7.29632318e-01 1.97885305e-01 4.73691225e-01
-5.48082054e-01 4.30362821e-01 1.79957211e+00 -3.34221095e-01
-1.99995637e-01 -1.16244383e-01 -1.20386517e+00 9.14606273e-01
4.04750884e-01 -3.09664696e-01 -5.58038890e-01 -1.09179032e+00
7.07488537e-01 -3.40249985e-01 3.76302034e-01 -9.39973652e-01
-8.62095878e-02 -2.58417845e-01 1.01194017e-01 -6.04396462e-01
6.23529792e-01 -1.01849556e+00 2.56756932e-01 -4.64009307e-02
1.57999367e-01 -4.11662817e-01 -1.80077523e-01 4.54418153e-01
-1.55788049e-01 8.42494443e-02 7.53695965e-01 -2.97018796e-01
-3.47296953e-01 1.77026019e-01 -1.74366683e-02 -1.23557739e-01
5.72996914e-01 -3.82465601e-01 -1.15147054e+00 -2.12340608e-01
-3.06836903e-01 -1.54386982e-01 6.29337966e-01 1.31315351e-01
7.86898851e-01 -9.95214820e-01 -5.99981368e-01 3.54830444e-01
-3.18474248e-02 2.13971421e-01 5.94951987e-01 1.03168213e+00
-7.30246127e-01 3.16074081e-02 -3.05408966e-02 -4.94890124e-01
-1.52372551e+00 4.99709994e-01 3.25666070e-01 -2.03757867e-01
-1.03301477e+00 7.22969234e-01 7.66185820e-01 -3.72388273e-01
-2.31491342e-01 -4.63742495e-01 -7.57388324e-02 -4.40358251e-01
6.41948521e-01 3.77913296e-01 2.66215146e-01 -5.89182377e-01
-2.36512646e-01 1.05616450e+00 -1.19895630e-01 1.15522616e-01
1.63870525e+00 -5.01130998e-01 -4.42319214e-01 -1.62872940e-01
1.05538118e+00 6.73110306e-01 -1.03681171e+00 -3.59446734e-01
-9.65353787e-01 -9.23120260e-01 -6.15314953e-02 -6.25827014e-01
-1.66336405e+00 6.13074720e-01 2.42576554e-01 1.80925742e-01
1.67572463e+00 -4.25154567e-01 1.00154638e+00 6.06085360e-01
1.01254053e-01 -9.95876312e-01 3.13227534e-01 5.72854459e-01
7.99432993e-01 -1.24511445e+00 4.36166465e-01 -1.24616170e+00
-5.41031420e-01 9.56812382e-01 5.83525717e-01 1.58856791e-02
4.90489781e-01 4.54973072e-01 5.01891851e-01 -4.04541463e-01
-4.84689057e-01 -1.27063379e-01 2.33915895e-01 1.01146531e+00
4.68303353e-01 -3.10386419e-01 -4.28356007e-02 2.85689414e-01
-2.43633896e-01 -2.39864171e-01 6.83125496e-01 8.03247631e-01
-1.99455336e-01 -1.16153049e+00 -6.76294148e-01 8.17436799e-02
-3.34931999e-01 -4.49865490e-01 -2.28371248e-01 3.50567520e-01
-1.08487263e-01 1.27642918e+00 -1.50938615e-01 -4.62753087e-01
4.67917651e-01 -4.90219712e-01 7.30547547e-01 -6.08304858e-01
-2.16940209e-01 2.31882721e-01 1.28515452e-01 -7.27435052e-01
-1.00002992e+00 -2.10223094e-01 -8.16886723e-01 -2.53709495e-01
-3.05141151e-01 -3.93292218e-01 3.79221499e-01 6.70153379e-01
4.39098924e-01 6.72245026e-01 8.08006763e-01 -1.04416478e+00
-1.02568358e-01 -7.88230300e-01 -7.25713134e-01 7.31289566e-01
5.16975880e-01 -4.58240628e-01 -5.21177709e-01 2.63684094e-01] | [10.482540130615234, -2.7876832485198975] |
629cfc83-d391-4250-bb9e-00d66fb2226a | stroke-extraction-of-chinese-character-based | 2307.04341 | null | https://arxiv.org/abs/2307.04341v1 | https://arxiv.org/pdf/2307.04341v1.pdf | Stroke Extraction of Chinese Character Based on Deep Structure Deformable Image Registration | Stroke extraction of Chinese characters plays an important role in the field of character recognition and generation. The most existing character stroke extraction methods focus on image morphological features. These methods usually lead to errors of cross strokes extraction and stroke matching due to rarely using stroke semantics and prior information. In this paper, we propose a deep learning-based character stroke extraction method that takes semantic features and prior information of strokes into consideration. This method consists of three parts: image registration-based stroke registration that establishes the rough registration of the reference strokes and the target as prior information; image semantic segmentation-based stroke segmentation that preliminarily separates target strokes into seven categories; and high-precision extraction of single strokes. In the stroke registration, we propose a structure deformable image registration network to achieve structure-deformable transformation while maintaining the stable morphology of single strokes for character images with complex structures. In order to verify the effectiveness of the method, we construct two datasets respectively for calligraphy characters and regular handwriting characters. The experimental results show that our method strongly outperforms the baselines. Code is available at https://github.com/MengLi-l1/StrokeExtraction. | ['Jian Wang', 'Guanghao Ren', 'Yi Yang', 'Yahan Yu', 'Meng Li'] | 2023-07-10 | null | null | null | null | ['image-registration'] | ['computer-vision'] | [ 3.44979227e-01 -6.48589194e-01 -1.87311649e-01 -3.08029950e-01
-3.78992170e-01 -8.24316859e-01 7.39774823e-01 -3.82772058e-01
-5.79686761e-01 3.05293024e-01 2.94962585e-01 -1.06929056e-01
3.97473089e-02 -1.06133199e+00 -2.95944393e-01 -6.72196090e-01
8.51270378e-01 4.57469016e-01 8.84473801e-01 -1.27297014e-01
7.86032915e-01 1.15319145e+00 -9.35791671e-01 2.36820057e-01
9.07906532e-01 4.70949382e-01 3.41673702e-01 7.62879193e-01
-6.62154317e-01 6.65142119e-01 -6.81143105e-01 -3.30553114e-01
2.62001514e-01 -6.89442098e-01 -6.70300901e-01 2.34626248e-01
2.59439468e-01 -7.62546301e-01 -6.71062827e-01 1.13117683e+00
7.24054158e-01 8.03734139e-02 8.90816450e-01 -1.01747489e+00
-6.78344011e-01 5.78306198e-01 -6.00350916e-01 -3.47413421e-02
-8.80564377e-02 -6.66799396e-02 4.63033944e-01 -7.09656775e-01
8.20513129e-01 9.47603881e-01 6.24464154e-01 5.54821074e-01
-5.15151858e-01 -5.44827104e-01 -5.24211116e-02 2.43175119e-01
-1.05813217e+00 -2.12273166e-01 9.40649152e-01 -3.96702349e-01
3.02306712e-01 3.45165789e-01 5.42237341e-01 6.71070874e-01
-1.80642143e-01 1.22801793e+00 1.13431048e+00 -3.54943544e-01
-1.25812888e-01 -1.75822660e-01 6.23891473e-01 6.42018795e-01
1.34940401e-01 -8.92357677e-02 -1.81805044e-01 3.19236308e-01
1.38845503e+00 2.65757471e-01 -2.87235290e-01 -3.20984870e-02
-1.23793852e+00 4.70394999e-01 2.79903054e-01 4.99289006e-01
-1.05802707e-01 7.78916478e-02 3.29512209e-01 5.65732410e-03
-7.41963834e-02 6.18900508e-02 -1.36705369e-01 -3.24179828e-01
-1.27484846e+00 2.00623110e-01 7.64244795e-01 1.23891211e+00
4.45421398e-01 1.33141994e-01 -5.94577074e-01 9.07157958e-01
-4.89059985e-02 6.73719883e-01 5.28679788e-01 -6.10840738e-01
3.97115797e-01 8.12791467e-01 -6.23374842e-02 -1.08031237e+00
-1.95390373e-01 1.34309933e-01 -9.86823559e-01 2.59606957e-01
6.34604931e-01 -1.90285280e-01 -1.25868583e+00 9.58173692e-01
-7.54214302e-02 2.15483055e-01 9.34446007e-02 1.05543637e+00
1.16800261e+00 7.29659021e-01 -2.47291207e-01 2.10872918e-01
1.34926999e+00 -1.18186498e+00 -8.52713168e-01 1.80639446e-01
4.18803133e-02 -1.24809027e+00 1.20043468e+00 6.27406836e-02
-1.01576960e+00 -5.93693435e-01 -1.13598621e+00 -3.63916725e-01
-3.97297710e-01 9.49623108e-01 4.20161188e-01 5.45706689e-01
-4.51608032e-01 5.55044055e-01 -9.02772188e-01 -2.68514067e-01
8.68908048e-01 4.20752391e-02 -4.04026918e-03 1.35831073e-01
-6.80754185e-01 6.79299712e-01 2.96479046e-01 3.24270844e-01
-4.73252684e-01 -4.11442250e-01 -3.30198467e-01 -9.31174755e-02
1.43885538e-01 -2.10373506e-01 9.09769535e-01 -7.98707962e-01
-1.85494685e+00 7.15904355e-01 -1.13302901e-01 1.89758778e-01
1.07547736e+00 -3.59342843e-01 -1.79812059e-01 2.48593628e-01
-1.29617319e-01 3.62842560e-01 5.94746292e-01 -1.15570581e+00
-5.76051593e-01 -7.91898221e-02 -4.77202863e-01 3.48372191e-01
-2.55309254e-01 3.02885771e-01 -1.10321999e+00 -1.08477807e+00
2.57606924e-01 -6.01883829e-01 -8.71221349e-03 1.36316448e-01
-7.51230299e-01 -1.49244040e-01 1.04761982e+00 -1.05831397e+00
1.07695222e+00 -1.64523697e+00 1.04028009e-01 4.83639926e-01
-6.55867532e-02 6.08189046e-01 -2.56427258e-01 3.62722963e-01
5.50889969e-01 -3.16733494e-02 -5.92800677e-01 8.82858410e-02
-2.35376544e-02 -8.80864263e-02 -5.64570487e-01 1.64386779e-01
1.09380506e-01 1.23871791e+00 -5.02713740e-01 -7.98923254e-01
1.84619620e-01 3.83281678e-01 3.38139720e-02 3.59167993e-01
1.22591846e-01 1.68603629e-01 -7.65152752e-01 7.24733114e-01
8.53003383e-01 1.23611286e-01 -1.25427336e-01 -4.96284872e-01
-4.27017897e-01 -2.44447112e-01 -1.25553834e+00 1.57720375e+00
1.23909235e-01 7.99154699e-01 -3.89130741e-01 -6.02581739e-01
1.21055508e+00 -1.99714959e-01 2.20100120e-01 -4.96315181e-01
4.41072762e-01 2.42689535e-01 -8.09979811e-02 -4.95698571e-01
5.62216163e-01 4.67020512e-01 8.68278220e-02 7.32083619e-01
-3.75574142e-01 -6.12818062e-01 3.31135929e-01 1.18308142e-01
5.65017343e-01 4.64257449e-01 -3.51300925e-01 -2.31066197e-01
7.16274440e-01 1.25031218e-01 7.18651772e-01 5.81708014e-01
2.99396757e-02 1.07280219e+00 3.28478724e-01 -2.40102351e-01
-1.11975718e+00 -1.01832080e+00 -8.76330137e-02 6.35385692e-01
6.69602752e-01 -4.43913192e-02 -1.22310758e+00 -5.74477792e-01
-2.18168825e-01 3.92136782e-01 -3.00910354e-01 2.34125584e-01
-1.16822422e+00 -8.31243813e-01 9.00421023e-01 8.37554336e-01
1.25211442e+00 -1.43477821e+00 -6.17793918e-01 -1.09223194e-01
-3.55023295e-02 -1.00203741e+00 -1.04682112e+00 -3.88559848e-01
-6.57903552e-01 -1.22775269e+00 -1.42041004e+00 -1.29137468e+00
1.01170802e+00 1.52442425e-01 4.34606493e-01 5.09424269e-01
-5.21377623e-01 6.83299303e-02 -3.12227011e-01 -2.19270498e-01
-2.49080911e-01 2.93059200e-01 -6.53959394e-01 -4.24217805e-02
3.22510153e-01 -2.42946714e-01 -7.09424078e-01 5.20855188e-01
-8.73386681e-01 3.53739381e-01 9.24072027e-01 5.60705960e-01
7.01466262e-01 -1.67726308e-01 9.90189463e-02 -9.63648677e-01
8.21979165e-01 3.49336177e-01 -6.41329587e-01 6.65262818e-01
-2.54199028e-01 -5.03393374e-02 3.30464840e-01 -6.33982062e-01
-1.29613233e+00 1.15521178e-01 3.54739130e-02 7.82973915e-02
-2.55870312e-01 1.46184072e-01 -4.99469221e-01 -1.84193969e-01
2.33718216e-01 8.18485916e-01 -3.43597233e-02 -7.00492561e-01
5.64413667e-01 8.99943650e-01 8.82641256e-01 -6.58384204e-01
9.37199652e-01 3.78957272e-01 -1.73254699e-01 -1.13357615e+00
-1.77272126e-01 -1.00711249e-01 -1.23118365e+00 -3.15039426e-01
1.09908795e+00 -3.78337443e-01 -3.66552591e-01 1.52974617e+00
-1.25570118e+00 -7.21676409e-01 -1.24425508e-01 2.59083152e-01
-4.53392446e-01 9.34426308e-01 -9.60032940e-01 -2.77683109e-01
-4.74546373e-01 -1.11198127e+00 8.95491421e-01 8.93371403e-01
1.66245654e-01 -6.95679605e-01 -4.62330170e-02 1.08797744e-01
3.34473759e-01 3.17619383e-01 9.09435451e-01 -4.34023559e-01
-7.33617604e-01 -3.08390766e-01 -7.73943365e-01 2.68047303e-01
4.84453738e-01 3.12597334e-01 -5.06053150e-01 5.51384240e-02
-4.58418131e-01 -4.03853692e-02 1.01057529e+00 1.52798429e-01
1.16275537e+00 3.54953632e-02 -2.73777813e-01 8.61983240e-01
1.26053429e+00 5.35192668e-01 9.94868457e-01 2.47699514e-01
1.23864126e+00 5.17980218e-01 4.18348491e-01 1.82519957e-01
1.96771637e-01 5.07523537e-01 -4.53399152e-01 -3.75203907e-01
-7.68832564e-01 -3.50699425e-01 1.20525472e-01 1.08491814e+00
-4.05181557e-01 -2.49001309e-01 -7.71167815e-01 2.86332458e-01
-1.70216417e+00 -9.40359592e-01 -4.69136566e-01 1.95736718e+00
1.08580649e+00 1.20193243e-01 9.41045582e-02 9.11471099e-02
1.04397976e+00 4.10429500e-02 -6.64794564e-01 -1.82262808e-01
-5.94677925e-01 2.10317239e-01 5.90108097e-01 5.07313371e-01
-1.13066649e+00 1.54290879e+00 4.94470739e+00 1.00622535e+00
-1.03332388e+00 -3.33487511e-01 4.64906245e-01 4.89087552e-01
-3.13134283e-01 -8.33610166e-03 -8.90486538e-01 6.12630188e-01
-2.33914346e-01 -1.54820159e-01 4.05074239e-01 2.58050889e-01
8.33928362e-02 4.99559529e-02 -5.65740407e-01 1.01140368e+00
1.20142855e-01 -1.46225429e+00 4.79890168e-01 -3.08246464e-01
8.61465812e-01 -3.28340083e-01 -3.19428504e-01 -2.89367497e-01
2.13753715e-01 -9.45547402e-01 7.94082403e-01 1.23162210e+00
7.50056803e-01 -6.82039678e-01 4.16473061e-01 2.07971722e-01
-1.35264373e+00 4.67233926e-01 -4.98168260e-01 2.42389768e-01
1.61702603e-01 2.10383847e-01 -8.15581307e-02 3.31042975e-01
2.50736833e-01 8.83855462e-01 -6.83071256e-01 1.28017914e+00
-6.67976320e-01 5.42398393e-01 -2.32854337e-01 -4.22961771e-01
7.14898333e-02 -7.41847217e-01 2.69927382e-01 1.50499487e+00
1.48035690e-01 4.72278893e-01 7.88641870e-02 1.03109109e+00
-4.24874425e-02 2.76494622e-01 -6.39199987e-02 -2.54532665e-01
6.95960164e-01 1.32070184e+00 -1.55725718e+00 -5.45719266e-01
-1.15206473e-01 1.50190425e+00 -1.05258331e-01 4.88998234e-01
-7.11417615e-01 -1.32783449e+00 6.81341141e-02 -1.85046673e-01
2.01994345e-01 -5.02683222e-01 -6.14445686e-01 -1.15957344e+00
1.07625024e-02 -7.11416423e-01 -1.99050438e-02 -6.53019845e-01
-1.04889262e+00 6.44459069e-01 -2.33609691e-01 -1.06446993e+00
3.29487979e-01 -6.97762072e-01 -1.31650352e+00 9.62745607e-01
-1.40386474e+00 -1.50414300e+00 -7.78227448e-01 5.94717324e-01
8.43421459e-01 -4.18644935e-01 4.17694241e-01 1.88993692e-01
-8.15933883e-01 8.34224999e-01 2.13634238e-01 1.01144671e+00
5.96247494e-01 -1.25242043e+00 7.11522698e-01 1.21266782e+00
4.08505872e-02 4.32929605e-01 5.98148368e-02 -1.04150486e+00
-1.06461978e+00 -9.52228725e-01 5.83463550e-01 -2.41979301e-01
3.07943374e-01 -2.80797243e-01 -8.18787932e-01 3.55385274e-01
1.29086122e-01 -4.06801134e-01 3.28414470e-01 -8.22434604e-01
-6.45495486e-03 1.90323710e-01 -8.23563397e-01 9.11875725e-01
1.06234837e+00 -3.76323313e-01 -7.16673136e-01 1.39908478e-01
1.13999344e-01 -5.52404940e-01 -5.02875805e-01 2.89788246e-01
8.51303101e-01 -6.23702347e-01 9.13225055e-01 -5.39219916e-01
4.70636398e-01 -4.91481930e-01 1.79022849e-01 -6.93514287e-01
-5.68086132e-02 -5.66354632e-01 3.13728839e-01 1.61777556e+00
1.06348597e-01 -3.48669708e-01 1.06386089e+00 5.52371323e-01
-1.30892366e-01 -5.83434463e-01 -2.43918449e-01 -6.25254571e-01
3.44985843e-01 1.35605618e-01 7.81768620e-01 8.03641260e-01
-5.52220464e-01 -8.20735246e-02 -2.58386940e-01 -5.90495132e-02
6.18421197e-01 6.27567828e-01 9.09132779e-01 -9.94242072e-01
8.06813389e-02 -1.08376014e+00 -2.30725273e-01 -1.45633447e+00
1.66604355e-01 -8.85023713e-01 1.46347880e-01 -1.79222965e+00
3.86913419e-01 -5.20877600e-01 3.24164778e-02 3.93723696e-01
-3.23388100e-01 2.68995583e-01 3.22480619e-01 5.76694012e-01
-2.13015601e-01 4.67012972e-01 1.62711644e+00 -9.43581909e-02
-4.05378014e-01 2.24026814e-01 -2.59014457e-01 8.65978718e-01
1.03962433e+00 -1.49342557e-02 -2.43257284e-01 -6.09164834e-01
-3.60475481e-01 -2.50024557e-01 3.35763693e-01 -7.37761319e-01
5.37801027e-01 -3.47326040e-01 6.33692324e-01 -9.26376462e-01
-7.38929734e-02 -3.72096658e-01 -2.83280581e-01 5.84545255e-01
-4.41568464e-01 7.34310895e-02 3.91826406e-02 1.79209635e-01
8.26044977e-02 -5.35861552e-01 6.85973346e-01 1.77232604e-02
-9.20031428e-01 3.68130386e-01 -4.36440170e-01 1.76517099e-01
6.89507365e-01 -4.68982548e-01 -4.91671592e-01 -8.97447616e-02
-3.53127807e-01 1.04713060e-01 4.79931563e-01 4.73617584e-01
9.22468960e-01 -1.41626847e+00 -8.76991987e-01 8.19094703e-02
-3.47732514e-01 -5.35200583e-03 1.73205771e-02 5.17267764e-01
-1.42591000e+00 -1.42428577e-02 -5.73673606e-01 -1.33228213e-01
-1.55422640e+00 4.70112637e-02 3.96114856e-01 2.55514115e-01
-7.19560683e-01 6.99901998e-01 -1.50048867e-01 -3.23456973e-01
3.37399662e-01 -4.66012388e-01 -4.91687864e-01 7.35102361e-03
4.47278649e-01 8.40031803e-01 -3.90277624e-01 -7.37936318e-01
-1.20759070e-01 1.34240794e+00 -1.71489120e-01 -1.65179566e-01
1.05637932e+00 2.46174023e-01 -3.65946859e-01 4.47233692e-02
9.52782035e-01 2.45976746e-01 -1.51751852e+00 -1.30194753e-01
1.22373469e-01 -5.16715765e-01 -1.08554311e-01 -8.65014017e-01
-1.37241399e+00 1.16337299e+00 6.72078907e-01 -3.70966673e-01
1.00921118e+00 -3.96601260e-01 1.22218442e+00 5.30830622e-01
-5.78186894e-03 -1.32563603e+00 8.68010968e-02 4.93032813e-01
9.49187219e-01 -8.35280478e-01 1.34559005e-01 -5.69992959e-01
-6.08736157e-01 1.65638459e+00 5.63562453e-01 -4.72774655e-01
6.72851026e-01 3.86825711e-01 4.53816354e-01 -6.35905638e-02
4.57756221e-01 -3.26557249e-01 5.53404450e-01 4.09622490e-01
1.60311341e-01 -4.91076931e-02 -5.10757387e-01 7.87115157e-01
-1.69930086e-01 -1.02998555e-01 5.73418260e-01 1.04431760e+00
-5.09029627e-01 -1.44500935e+00 -5.43810502e-02 4.15637732e-01
-1.73563272e-01 -1.26027659e-01 -9.80343342e-01 6.42608643e-01
-1.08827930e-02 6.67284489e-01 1.14972316e-01 -1.28384933e-01
4.33406025e-01 -3.20310950e-01 6.92984045e-01 -1.63572803e-01
-3.97316337e-01 1.76594660e-01 -2.66544193e-01 -1.91001609e-01
-3.86682451e-01 -7.26101637e-01 -1.68576491e+00 -1.69173703e-01
-3.74081641e-01 -5.20770364e-02 4.61192280e-01 7.52220631e-01
9.02004242e-02 4.48247522e-01 5.83881259e-01 -8.33790600e-01
-3.76758337e-01 -7.12329626e-01 -4.68714356e-01 5.85751951e-01
-2.05197573e-01 -2.01469332e-01 3.38846967e-02 4.29936051e-01] | [12.017559051513672, 2.2170488834381104] |
b12e81ff-621e-4998-a973-2a42f4569a35 | 3d-affordancenet-a-benchmark-for-visual | 2103.16397 | null | https://arxiv.org/abs/2103.16397v2 | https://arxiv.org/pdf/2103.16397v2.pdf | 3D AffordanceNet: A Benchmark for Visual Object Affordance Understanding | The ability to understand the ways to interact with objects from visual cues, a.k.a. visual affordance, is essential to vision-guided robotic research. This involves categorizing, segmenting and reasoning of visual affordance. Relevant studies in 2D and 2.5D image domains have been made previously, however, a truly functional understanding of object affordance requires learning and prediction in the 3D physical domain, which is still absent in the community. In this work, we present a 3D AffordanceNet dataset, a benchmark of 23k shapes from 23 semantic object categories, annotated with 18 visual affordance categories. Based on this dataset, we provide three benchmarking tasks for evaluating visual affordance understanding, including full-shape, partial-view and rotation-invariant affordance estimations. Three state-of-the-art point cloud deep learning networks are evaluated on all tasks. In addition we also investigate a semi-supervised learning setup to explore the possibility to benefit from unlabeled data. Comprehensive results on our contributed dataset show the promise of visual affordance understanding as a valuable yet challenging benchmark. | ['Kui Jia', 'Ke Chen', 'Chaozheng Wu', 'Xun Xu', 'Shengheng Deng'] | 2021-03-30 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Deng_3D_AffordanceNet_A_Benchmark_for_Visual_Object_Affordance_Understanding_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Deng_3D_AffordanceNet_A_Benchmark_for_Visual_Object_Affordance_Understanding_CVPR_2021_paper.pdf | cvpr-2021-1 | ['affordance-detection'] | ['computer-vision'] | [-1.38090402e-01 -4.61488105e-02 -3.22614461e-01 -4.59813207e-01
7.23935023e-04 -7.66312599e-01 8.10737491e-01 1.56055138e-01
-1.48110911e-01 5.78002408e-02 2.42716089e-01 -3.02326709e-01
-2.02353612e-01 -3.90244901e-01 -9.29939687e-01 -3.41357619e-01
-2.36884192e-01 7.60863245e-01 2.90988415e-01 -3.49814475e-01
5.12820542e-01 1.12143016e+00 -1.69118631e+00 4.66215521e-01
5.52901685e-01 1.14247108e+00 4.76269215e-01 5.92349470e-01
-6.98116943e-02 4.94088978e-01 -8.32183212e-02 2.19025433e-01
4.96280700e-01 4.58726048e-01 -9.40531492e-01 2.59876311e-01
8.62835526e-01 -6.60972953e-01 -4.23195392e-01 6.63617313e-01
2.05729365e-01 -1.19963358e-03 1.00026178e+00 -1.62403011e+00
-1.31388819e+00 1.37305021e-01 -1.31093219e-01 5.08692600e-02
7.53393054e-01 5.52543283e-01 1.30921698e+00 -1.01239777e+00
8.72527778e-01 1.54939771e+00 1.88667759e-01 6.65746927e-01
-1.34575975e+00 -1.15824327e-01 2.69102782e-01 6.80655390e-02
-9.62686241e-01 -1.99200898e-01 9.01883066e-01 -6.75515890e-01
1.57812333e+00 2.35096127e-01 9.85707581e-01 1.04060638e+00
-1.60916597e-01 1.10048807e+00 1.12228465e+00 -4.03879404e-01
1.24428891e-01 -3.60659540e-01 -4.08698432e-03 8.63093793e-01
1.78507075e-01 4.00387019e-01 -4.99425054e-01 -4.54666875e-02
1.25742841e+00 4.29407448e-01 -2.21821263e-01 -1.37079263e+00
-1.84955823e+00 4.16319102e-01 1.41728151e+00 6.61761537e-02
-1.32300407e-01 4.37138915e-01 9.32172388e-02 2.99518377e-01
3.23740900e-01 6.96033299e-01 -8.10363948e-01 3.09907705e-01
-2.33450651e-01 6.24945700e-01 7.35494852e-01 1.42083633e+00
7.03158379e-01 -1.17744066e-01 5.12721129e-02 4.95734245e-01
9.30503011e-01 6.65107608e-01 -1.27098605e-01 -1.03104568e+00
4.85288858e-01 1.00924432e+00 2.60361701e-01 -5.14635026e-01
-7.33185172e-01 3.44417155e-01 -3.72388124e-01 6.83109581e-01
6.40731752e-01 6.87187970e-01 -1.36415255e+00 1.27843213e+00
3.74680668e-01 -3.18447165e-02 -1.39655694e-01 1.50838852e+00
1.32031596e+00 1.31280184e-01 1.05699822e-01 6.03849649e-01
1.39814985e+00 -9.70004559e-01 -1.43449813e-01 -5.80234751e-02
4.31029052e-01 -6.63389742e-01 1.41268170e+00 1.28703937e-01
-6.48249328e-01 -5.64050019e-01 -1.14449823e+00 -7.89193213e-01
-6.08988941e-01 3.12691450e-01 1.30286944e+00 1.50249124e-01
-8.31256449e-01 6.49272025e-01 -1.02601981e+00 -5.66535890e-01
9.45614755e-01 3.49108636e-01 -8.26108456e-01 -2.02560082e-01
-4.89875019e-01 1.13064647e+00 3.22250634e-01 -5.68087921e-02
-1.31315958e+00 -8.29630077e-01 -1.10074031e+00 -1.66570410e-01
3.56323093e-01 -9.63552773e-01 1.25847471e+00 -3.60210389e-01
-1.10874748e+00 1.33027065e+00 1.10403515e-01 -2.28540629e-01
4.50985849e-01 -3.27691317e-01 1.13368444e-01 4.92864363e-02
1.80223644e-01 1.24259555e+00 9.23481226e-01 -1.60213625e+00
-1.62772655e-01 -6.86701298e-01 5.11792243e-01 2.49345481e-01
3.85865897e-01 -2.88322270e-01 4.37915465e-03 -4.22210246e-01
7.01236308e-01 -1.08046865e+00 -6.61025047e-02 8.52849722e-01
-5.96847594e-01 -7.57794976e-01 8.76298308e-01 -2.67428219e-01
1.09076113e-01 -1.93864274e+00 4.41502452e-01 -2.14612916e-01
5.40436208e-01 -8.92961919e-02 -2.10706919e-01 3.75058681e-01
-8.94987509e-02 7.20659420e-02 -1.94475204e-02 -2.70835966e-01
4.78677511e-01 4.74782526e-01 -4.81827646e-01 6.17502332e-01
6.35218441e-01 1.26105881e+00 -8.82380128e-01 -1.23393103e-01
6.98925197e-01 2.95089006e-01 -7.11560607e-01 4.57453907e-01
-8.62213790e-01 8.03542912e-01 -5.88218272e-01 1.27999210e+00
6.15194738e-01 -2.68062413e-01 -3.18319976e-01 -3.05924088e-01
-1.39908001e-01 4.31357503e-01 -7.88161159e-01 2.56482482e+00
-2.45351002e-01 5.80175042e-01 -3.14085066e-01 -7.61411190e-01
8.35106254e-01 5.37213124e-02 3.35340619e-01 -4.03343141e-01
1.60410479e-01 3.93240809e-01 1.04945160e-01 -7.52409935e-01
2.78133631e-01 1.37095571e-01 -7.93126673e-02 3.85948598e-01
5.48240960e-01 -8.92242789e-01 -1.29281566e-01 -5.30770123e-02
6.18779063e-01 9.11413789e-01 4.06545490e-01 -3.19742739e-01
7.76970610e-02 2.24048883e-01 -3.05594891e-01 4.24445093e-01
-4.24835473e-01 7.52914488e-01 3.31325650e-01 -1.07518029e+00
-1.35885000e+00 -1.43129408e+00 -4.26577419e-01 9.73414600e-01
4.58632410e-01 -5.18107265e-02 -2.52065808e-01 -5.43128610e-01
6.80025399e-01 3.92245382e-01 -7.42371023e-01 1.22757129e-01
-3.48925442e-01 -1.09591410e-01 -1.62875373e-02 6.89002156e-01
2.21856222e-01 -1.24570918e+00 -1.08474541e+00 -3.54049474e-01
3.53282630e-01 -1.32750547e+00 -5.36766276e-02 4.31280248e-02
-8.99402440e-01 -1.27987957e+00 -7.06216931e-01 -7.02158034e-01
5.65961361e-01 5.80224276e-01 1.19768620e+00 1.50712922e-01
-5.59422493e-01 7.74380326e-01 -2.60266036e-01 -7.97435701e-01
-1.89518452e-01 1.24907710e-01 2.23086193e-01 -4.78259832e-01
4.85161066e-01 -5.44644713e-01 -5.08212686e-01 2.33690694e-01
-5.19117713e-01 1.90246075e-01 5.52044034e-01 5.26737630e-01
7.42385209e-01 -1.06804192e+00 5.05194254e-02 -3.96957435e-02
1.87109500e-01 -2.80495882e-01 -5.38841367e-01 1.28900677e-01
-1.13025919e-01 2.42647350e-01 7.32993856e-02 -3.91878724e-01
-7.74526119e-01 2.85082519e-01 9.68739092e-02 -6.84236050e-01
-8.59033585e-01 -6.95259199e-02 -2.18049332e-01 -3.23952347e-01
9.03328657e-01 -2.94204712e-01 4.14757021e-02 -6.69171512e-01
1.00317204e+00 4.41713095e-01 2.10208908e-01 -6.99685395e-01
5.61711788e-01 9.34221208e-01 3.80469412e-01 -9.68241453e-01
-8.43746603e-01 -7.09715605e-01 -1.47345841e+00 -4.40646745e-02
1.12367904e+00 -9.40837622e-01 -9.27166700e-01 2.39629224e-01
-1.49914241e+00 -4.53033030e-01 -3.71277779e-01 5.34537017e-01
-1.13091433e+00 9.44364890e-02 -1.63082376e-01 -5.10736585e-01
6.13713749e-02 -1.24790823e+00 1.71849358e+00 -4.35252003e-02
-2.98633546e-01 -6.24875546e-01 -2.08411947e-01 3.12436849e-01
-1.59495369e-01 5.88614345e-01 8.72327328e-01 -4.58416522e-01
-1.39154601e+00 5.87071031e-02 -7.37747788e-01 4.46647517e-02
1.56813756e-01 -2.27272138e-01 -1.05522585e+00 -1.26393184e-01
-3.68284225e-01 -8.48697007e-01 9.61036444e-01 2.31355444e-01
1.09894156e+00 3.10778677e-01 -4.04924333e-01 6.83317602e-01
1.23863626e+00 -3.04134071e-01 2.37298086e-01 1.96388662e-01
1.23447359e+00 6.90850139e-01 6.20613515e-01 2.79386967e-01
5.47538996e-01 6.40096188e-01 1.24486744e+00 1.72773898e-01
-3.95711303e-01 -3.76956969e-01 -3.23663265e-01 1.54369608e-01
-6.87419951e-01 1.75307333e-01 -1.23004556e+00 3.86097789e-01
-1.68323660e+00 -4.29197788e-01 -3.20817471e-01 1.73184788e+00
3.55443984e-01 4.17269813e-03 4.76305149e-02 5.91890924e-02
-7.48020709e-02 1.61856592e-01 -1.05221128e+00 -2.16748834e-01
-2.19088309e-02 -4.93591130e-02 1.97224945e-01 1.26935467e-01
-1.36843395e+00 1.12148297e+00 5.90642881e+00 1.16824634e-01
-1.00412643e+00 5.19241653e-02 -8.95450264e-02 1.80167668e-02
-3.45905334e-01 1.52877420e-01 -3.92375648e-01 -3.07741314e-01
1.21248484e-01 6.37165129e-01 7.54650354e-01 1.12603056e+00
-2.64486343e-01 -1.07056901e-01 -1.74357784e+00 1.11239624e+00
1.92709073e-01 -1.28473902e+00 2.89033771e-01 4.40443419e-02
6.45726621e-01 5.46661496e-01 -1.14020236e-01 5.92234656e-02
-1.83062013e-02 -1.21834064e+00 1.09609151e+00 7.25752592e-01
8.14245224e-01 -1.11970723e-01 9.31456164e-02 3.65658998e-01
-1.16238940e+00 -1.68837577e-01 -6.54620349e-01 -3.77909899e-01
-1.24860704e-02 -1.74376275e-02 -7.78250039e-01 4.73337978e-01
1.05192161e+00 1.10626578e+00 -6.19967103e-01 9.05424654e-01
-4.68996495e-01 -3.77932310e-01 -3.87359828e-01 -2.36796156e-01
2.04571575e-01 3.00019402e-02 6.31745219e-01 5.68850279e-01
6.64061159e-02 1.35306288e-02 2.42865756e-01 1.34857821e+00
4.96076830e-02 -2.27477148e-01 -9.04065847e-01 -1.75541013e-01
7.58906975e-02 9.72467184e-01 -7.24056959e-01 -2.74385512e-02
-3.94352376e-01 8.44394147e-01 6.96670711e-01 2.69014478e-01
-4.57889706e-01 1.24203324e-01 1.03180051e+00 5.25068454e-02
1.48056164e-01 -8.35224032e-01 -5.30782163e-01 -1.24663842e+00
6.57675266e-02 -2.85755247e-01 -1.09694354e-01 -1.37045932e+00
-1.38496244e+00 2.25200266e-01 3.54702741e-01 -1.31678593e+00
2.09963143e-01 -1.61162281e+00 -6.17061108e-02 8.92562926e-01
-1.66418457e+00 -1.80765820e+00 -5.34486055e-01 5.09115398e-01
5.95642030e-01 -1.49885818e-01 9.39259887e-01 -3.12790930e-01
4.80774820e-01 -9.81389359e-02 -5.30341744e-01 -5.42985275e-02
4.15623069e-01 -1.49279082e+00 6.56330884e-01 8.11119899e-02
6.45508885e-01 6.67045951e-01 4.13394988e-01 -3.38176876e-01
-1.87071216e+00 -6.15808070e-01 5.41906714e-01 -1.47788870e+00
5.99303544e-01 -6.91727042e-01 -7.61439860e-01 8.03272069e-01
4.71915156e-02 6.26387477e-01 2.64336169e-01 6.75471947e-02
-7.15761781e-01 4.08985347e-01 -1.05669272e+00 7.32435584e-01
1.87375331e+00 -8.40500772e-01 -1.07420790e+00 5.44045269e-01
1.13099909e+00 -8.34684312e-01 -9.46049094e-01 5.47523081e-01
7.89336860e-01 -9.69806194e-01 1.51153231e+00 -1.07438993e+00
4.22984064e-01 -4.38319355e-01 -4.69080687e-01 -1.23426056e+00
-2.05733106e-01 -1.22577334e-02 -3.47418398e-01 4.00793582e-01
1.04180492e-01 -2.52945602e-01 5.41483223e-01 4.06612575e-01
-4.55439538e-01 -6.51255190e-01 -9.19213057e-01 -6.71760678e-01
1.80286001e-02 -7.56737947e-01 9.54298496e-01 7.30248809e-01
-1.10314555e-01 2.89877594e-01 7.49903172e-02 2.01015249e-01
6.73899949e-01 6.38807654e-01 1.01466823e+00 -1.68221521e+00
9.53475907e-02 -6.59884810e-01 -7.24920213e-01 -1.29548585e+00
3.83507103e-01 -1.22732341e+00 -1.31300956e-01 -1.85028505e+00
-1.33770645e-01 -5.95670938e-01 -8.28379393e-02 5.58361709e-01
2.48338789e-01 1.45086259e-01 4.53704357e-01 3.91492993e-01
-4.82002228e-01 6.81828916e-01 2.04507422e+00 -3.58792007e-01
3.40447500e-02 -2.24699393e-01 -5.23393229e-02 7.50157177e-01
6.85036719e-01 -1.02139860e-01 -4.16835517e-01 -7.55741537e-01
1.09442763e-01 -2.37990871e-01 1.11870384e+00 -6.79168344e-01
-2.17845902e-01 -2.17717350e-01 4.00841266e-01 -7.79449940e-01
5.98221481e-01 -1.08770084e+00 -4.05253649e-01 3.76205772e-01
-1.57801121e-01 -1.54953435e-01 1.47081047e-01 8.13714266e-01
1.93445504e-01 1.38602763e-01 3.72648388e-01 -4.72339720e-01
-1.31321263e+00 7.24284768e-01 2.14855865e-01 -1.91992819e-01
9.50118482e-01 -1.77208662e-01 -2.53051460e-01 2.33349636e-01
-9.85075235e-01 1.15408599e-01 7.50573695e-01 1.01051092e+00
9.18750823e-01 -1.55008054e+00 -4.28833902e-01 4.54849511e-01
8.86037230e-01 2.90489286e-01 1.97037533e-01 5.06445587e-01
-7.54866958e-01 6.71753407e-01 -6.35344803e-01 -1.27442825e+00
-8.18945050e-01 1.08287370e+00 2.06357196e-01 5.45132160e-01
-6.18345380e-01 8.94627154e-01 1.96713224e-01 -1.23273051e+00
1.45073593e-01 -1.31172931e+00 -2.60402888e-01 -4.01030749e-01
5.44771105e-02 1.62838608e-01 -1.91741183e-01 -8.49072993e-01
-4.16136771e-01 8.13488901e-01 5.65039694e-01 1.70644015e-01
1.30803084e+00 7.17615709e-02 -4.13675904e-01 7.72537410e-01
1.08416593e+00 -8.23718131e-01 -1.72240877e+00 -1.86294213e-01
2.14162067e-01 -6.95782721e-01 -3.44676197e-01 -6.33351982e-01
-4.09205496e-01 1.37160778e+00 4.95971829e-01 2.64350086e-01
5.75483918e-01 6.14824355e-01 1.51559502e-01 7.95946062e-01
7.78283060e-01 -5.74679434e-01 4.73843932e-01 6.71260953e-01
1.71467352e+00 -1.89248073e+00 4.80963774e-02 -6.01149917e-01
-5.06037354e-01 1.20665455e+00 6.81827843e-01 -5.66156626e-01
9.46533084e-01 -2.92432129e-01 -1.05254248e-01 -4.87440348e-01
-3.96597832e-01 -4.84870821e-01 8.93149793e-01 9.51148689e-01
2.15642989e-01 4.35478240e-01 3.70045602e-01 1.79559231e-01
-2.31836602e-01 -1.63093254e-01 7.89024532e-02 8.44299257e-01
-3.22919756e-01 -7.81208992e-01 -1.92319915e-01 3.25369686e-01
2.29562461e-01 1.43185437e-01 -4.84907478e-01 9.07458782e-01
3.51846248e-01 5.91557384e-01 4.05152440e-02 -9.34474636e-03
6.26145720e-01 -4.80597615e-02 1.24025381e+00 -8.66172075e-01
2.00902782e-02 -2.05488712e-01 -4.57803756e-02 -9.31985080e-01
-8.80892932e-01 -6.97739840e-01 -1.35397553e+00 8.94008204e-02
-1.09966144e-01 -6.93882167e-01 9.46020603e-01 1.05014563e+00
2.13724896e-01 4.21833128e-01 -1.27923548e-01 -1.81775200e+00
-4.89967376e-01 -1.08765221e+00 -5.49563408e-01 6.11090302e-01
6.87765956e-01 -1.29235399e+00 -6.87247962e-02 -1.78543061e-01] | [5.178782939910889, -0.13537679612636566] |
408a56bd-563f-4f49-8287-7aeabd9c484b | sentiment-perception-adversarial-attacks-on | 2305.01437 | null | https://arxiv.org/abs/2305.01437v2 | https://arxiv.org/pdf/2305.01437v2.pdf | Sentiment Perception Adversarial Attacks on Neural Machine Translation Systems | With the advent of deep learning methods, Neural Machine Translation (NMT) systems have become increasingly powerful. However, deep learning based systems are susceptible to adversarial attacks, where imperceptible changes to the input can cause undesirable changes at the output of the system. To date there has been little work investigating adversarial attacks on sequence-to-sequence systems, such as NMT models. Previous work in NMT has examined attacks with the aim of introducing target phrases in the output sequence. In this work, adversarial attacks for NMT systems are explored from an output perception perspective. Thus the aim of an attack is to change the perception of the output sequence, without altering the perception of the input sequence. For example, an adversary may distort the sentiment of translated reviews to have an exaggerated positive sentiment. In practice it is challenging to run extensive human perception experiments, so a proxy deep-learning classifier applied to the NMT output is used to measure perception changes. Experiments demonstrate that the sentiment perception of NMT systems' output sequences can be changed significantly with small imperceptible changes to input sequences. | ['Mark Gales', 'Vyas Raina'] | 2023-05-02 | null | null | null | null | ['nmt'] | ['computer-code'] | [ 8.46222878e-01 2.34323427e-01 3.47686023e-01 -1.81287691e-01
-6.80791795e-01 -1.29351532e+00 6.43022120e-01 -1.08663231e-01
-2.74884820e-01 4.14042026e-01 -1.09324031e-01 -8.49937916e-01
9.07157302e-01 -6.65460110e-01 -1.17091990e+00 -6.21272027e-01
2.64440656e-01 1.20456882e-01 -1.80741027e-01 -5.82840264e-01
1.32771626e-01 4.41253930e-01 -7.69260347e-01 6.94761217e-01
4.37269270e-01 5.68552077e-01 -4.13913727e-01 1.11576581e+00
2.05796078e-01 6.55855060e-01 -1.50285816e+00 -6.01740241e-01
7.83247828e-01 -6.25164688e-01 -3.96624029e-01 -1.32876769e-01
5.03948689e-01 -2.35141441e-01 -2.56878734e-01 1.54469645e+00
6.81937873e-01 -2.34914795e-01 4.78428870e-01 -1.32387877e+00
-1.08132792e+00 7.26740837e-01 -3.36758584e-01 7.11542666e-02
3.97854656e-01 7.61553466e-01 8.16630900e-01 -6.74673140e-01
3.41248155e-01 1.69702566e+00 3.77312392e-01 7.67995894e-01
-1.43149984e+00 -7.04126596e-01 8.11101049e-02 1.03674149e-02
-8.29293311e-01 -2.75014222e-01 6.01632416e-01 -2.62586445e-01
9.83543038e-01 5.07555604e-01 2.86304593e-01 1.70210886e+00
9.75048363e-01 6.40191555e-01 1.41016829e+00 -4.55146164e-01
3.35232973e-01 2.81905085e-01 -6.25370443e-01 2.01616079e-01
-7.92486519e-02 4.69327837e-01 -2.21108958e-01 -2.05623895e-01
2.88198233e-01 -4.90407228e-01 -1.00763896e-02 -1.80196762e-02
-1.16021800e+00 8.84243786e-01 4.67885256e-01 1.19293787e-01
-2.39456594e-01 3.24668616e-01 6.22666240e-01 1.08611441e+00
3.98052931e-01 1.06277013e+00 -5.16906977e-01 -1.81646466e-01
-3.40764254e-01 1.51236475e-01 7.85541832e-01 6.03662968e-01
3.16415012e-01 4.15591002e-01 -1.81532234e-01 3.40239614e-01
-1.34241477e-01 1.01054668e+00 4.97498840e-01 -6.35850489e-01
4.33816940e-01 1.14462189e-01 8.48189518e-02 -1.28271377e+00
8.06153715e-02 -3.09761047e-01 -5.23226917e-01 8.16137493e-01
3.65566969e-01 -7.05576718e-01 -1.13417327e+00 1.78339636e+00
-1.31594925e-03 -3.42635393e-01 4.81392950e-01 9.43170965e-01
1.18085362e-01 8.15847695e-01 -1.19251423e-01 2.56533790e-02
8.71328354e-01 -5.99447191e-01 -7.94098735e-01 -5.18711865e-01
6.69871330e-01 -1.33008444e+00 1.26490128e+00 3.59173268e-01
-9.65378642e-01 -4.44162786e-01 -1.47296894e+00 2.62271464e-01
-6.20026350e-01 -5.83824277e-01 9.60305054e-03 1.08173633e+00
-9.16960776e-01 3.96690637e-01 -5.75018644e-01 -4.15677093e-02
2.29290813e-01 5.99572122e-01 -2.08832756e-01 1.48051128e-01
-1.73673284e+00 1.39610195e+00 1.10452995e-03 1.39277279e-01
-9.89569008e-01 -4.20845717e-01 -6.42481744e-01 -1.20915145e-01
7.20271617e-02 -4.21620697e-01 1.75251567e+00 -1.85220397e+00
-1.50514245e+00 6.54182136e-01 2.46438071e-01 -7.56496310e-01
9.10901189e-01 -2.00724870e-01 -5.76909661e-01 -2.05509186e-01
-1.97358683e-01 5.66340208e-01 1.37685537e+00 -1.12359440e+00
-2.00537518e-01 -1.85017258e-01 2.90044844e-01 2.25188658e-01
-3.40467870e-01 1.69245198e-01 3.56301457e-01 -8.92165959e-01
-5.76623321e-01 -1.35040581e+00 -2.62945563e-01 -1.58242777e-01
-6.26203060e-01 2.98364013e-01 1.13595402e+00 -3.88895541e-01
5.99797904e-01 -2.23655105e+00 -1.23769425e-01 2.05914468e-01
-8.11121389e-02 7.12919056e-01 -3.93322617e-01 5.91262877e-01
-3.00917655e-01 3.52329433e-01 -1.03696555e-01 7.72426203e-02
1.49064764e-01 1.77894697e-01 -7.85621583e-01 5.37827492e-01
3.58601868e-01 1.05816031e+00 -8.82347286e-01 1.83636129e-01
1.31490543e-01 2.64572293e-01 -3.84723723e-01 2.92362303e-01
-5.28218687e-01 1.99059650e-01 -1.57466918e-01 2.58657992e-01
5.31160235e-01 4.34705138e-01 -3.56027521e-02 2.53265202e-01
1.59335062e-01 2.15358838e-01 -2.97680408e-01 8.86579096e-01
-5.59566617e-01 1.23202133e+00 -1.30967274e-01 -8.02961171e-01
9.25669551e-01 5.59313595e-01 -1.88859791e-01 -9.00392711e-01
3.33765447e-01 2.16941565e-01 7.36386120e-01 -2.35392943e-01
3.79396051e-01 -4.16522384e-01 -5.33966660e-01 5.77821910e-01
-5.86686850e-01 -3.12419087e-01 -4.14857090e-01 9.69424918e-02
1.22855210e+00 -3.24403137e-01 -5.28089851e-02 4.16494049e-02
1.08001143e-01 2.48448014e-01 2.81819224e-01 1.05724752e+00
-2.62793362e-01 3.99632990e-01 5.54071128e-01 -4.16792721e-01
-1.55227339e+00 -1.07517302e+00 4.02343035e-01 9.90851462e-01
-3.05238031e-02 1.88421175e-01 -1.18522930e+00 -6.62920713e-01
4.78410162e-02 1.00910640e+00 -7.23147810e-01 -8.34311187e-01
-6.58565223e-01 -5.29727519e-01 1.08941948e+00 5.15621006e-01
2.90341765e-01 -1.27192545e+00 -6.00531995e-01 1.70370787e-01
-1.19289104e-02 -1.08682466e+00 -8.21570992e-01 2.98640311e-01
-3.89330238e-01 -5.58008909e-01 -3.27804565e-01 -7.18438923e-01
7.72995830e-01 -1.27420574e-01 7.95347333e-01 -1.90551534e-01
-1.13474667e-01 5.08593442e-03 -2.91595817e-01 -9.76653636e-01
-1.43830609e+00 -5.92785329e-02 4.30177629e-01 -7.46820271e-02
5.67179322e-01 -3.49203974e-01 -3.48219991e-01 3.37603033e-01
-1.24312150e+00 -1.88295662e-01 6.48315847e-01 5.72408915e-01
-9.57920179e-02 1.54767305e-01 5.32478869e-01 -9.63135183e-01
1.20258915e+00 -1.65145829e-01 -2.69200921e-01 -4.56980430e-02
-4.11602855e-01 -8.33080038e-02 1.26810110e+00 -1.10671389e+00
-6.41185343e-01 -3.09731036e-01 -2.10149854e-01 -6.53546572e-01
-9.18483883e-02 2.96612501e-01 -4.30081099e-01 -2.44731214e-02
1.22606730e+00 1.99852347e-01 1.82280373e-02 9.98748094e-02
4.59814340e-01 8.64435077e-01 5.35616517e-01 -2.53848583e-01
1.34734178e+00 1.42646313e-01 -2.69966871e-01 -3.92606020e-01
-5.42703629e-01 3.97694230e-01 -2.45442674e-01 -1.46025553e-01
5.58230639e-01 -5.69438338e-01 -5.15754938e-01 7.24397242e-01
-1.28501546e+00 -4.54285771e-01 6.57413015e-03 -7.83681273e-02
-4.54800725e-01 1.41949728e-01 -6.27316356e-01 -5.66235363e-01
-3.61947477e-01 -1.39276934e+00 7.25328147e-01 -1.12921901e-01
-7.73178279e-01 -8.89970541e-01 -4.21080291e-02 2.39462927e-01
5.16887069e-01 4.64770913e-01 1.17011511e+00 -1.07460868e+00
-2.21083134e-01 -6.81864440e-01 3.46624076e-01 8.48207951e-01
3.06348383e-01 6.79256544e-02 -1.02395332e+00 -5.37286043e-01
3.60678971e-01 -1.90849513e-01 1.99300408e-01 -1.63786799e-01
6.61579192e-01 -7.50637710e-01 8.62543955e-02 2.24786535e-01
9.57917690e-01 6.82179570e-01 6.97972357e-01 5.67463875e-01
6.06831312e-01 3.96727264e-01 5.88552058e-01 -1.53273731e-01
-4.75877404e-01 6.25863194e-01 5.82220495e-01 -3.74790162e-01
1.92415029e-01 -3.32472116e-01 1.09861290e+00 4.46727157e-01
6.38439834e-01 -7.29782403e-01 -7.90118873e-01 3.27533454e-01
-1.52995014e+00 -8.47477138e-01 2.79711545e-01 1.93035448e+00
8.43421102e-01 7.53010631e-01 -2.77324826e-01 1.75783932e-01
8.44489336e-01 1.36448935e-01 -8.76807570e-01 -1.54071569e+00
-9.53370258e-02 1.13803864e-01 6.00291789e-01 5.00228167e-01
-1.04621339e+00 1.06171477e+00 6.44428205e+00 5.95613956e-01
-1.46669710e+00 -1.04264237e-01 6.17662728e-01 -1.39592275e-01
-2.98244715e-01 -2.12777868e-01 -2.08930865e-01 6.29059792e-01
1.12722862e+00 -2.87134826e-01 4.11178648e-01 7.47604191e-01
4.27384794e-01 4.89472806e-01 -1.24984419e+00 4.73464906e-01
1.46051738e-02 -8.65321875e-01 1.57948002e-01 2.04297677e-01
7.64602363e-01 -9.20434669e-02 8.30057085e-01 3.83381188e-01
6.64337695e-01 -1.31223524e+00 6.75186694e-01 1.42160818e-01
6.85598075e-01 -1.17255569e+00 8.25269461e-01 3.20448041e-01
-1.68290451e-01 2.02725325e-02 -2.53585935e-01 -4.37301785e-01
-1.82502732e-01 3.22764844e-01 -1.33360934e+00 -1.44119754e-01
2.94044971e-01 -5.51558137e-02 -2.64165670e-01 1.04783259e-01
-5.52301228e-01 7.52328277e-01 -7.74731413e-02 -2.90447414e-01
5.33265412e-01 1.93672031e-01 8.11509967e-01 1.22064459e+00
-3.95357534e-02 -2.98586279e-01 1.85871869e-01 7.07164109e-01
-4.42594230e-01 -8.20959136e-02 -1.07881284e+00 -4.21540290e-01
3.11771452e-01 7.67440200e-01 -2.96256959e-01 -2.28535801e-01
-1.38852194e-01 1.28882968e+00 -1.21905170e-01 3.84729177e-01
-9.30484235e-01 -4.87891167e-01 1.16695487e+00 -1.82245582e-01
2.07622260e-01 -1.59320533e-01 -4.80242193e-01 -7.74173379e-01
1.77929163e-01 -1.89327502e+00 -2.38226339e-01 -7.19887257e-01
-1.27231550e+00 6.57830179e-01 -6.38936341e-01 -1.11687624e+00
-3.95385474e-01 -4.67409492e-01 -6.95530176e-01 1.08778703e+00
-6.52346611e-01 -8.57861519e-01 4.21578974e-01 2.68392086e-01
7.36114502e-01 -3.43746245e-01 9.60861623e-01 -2.54633337e-01
-3.32526624e-01 1.14598000e+00 2.43205681e-01 4.66593146e-01
7.59831607e-01 -1.23904121e+00 1.28305900e+00 1.26562667e+00
1.05995923e-01 6.26573026e-01 1.46288979e+00 -6.61005676e-01
-1.48966765e+00 -1.19058752e+00 6.03021204e-01 -9.08961952e-01
8.26647162e-01 -6.73754632e-01 -8.52235317e-01 6.95309937e-01
5.42756379e-01 -3.54054689e-01 6.10410392e-01 -3.05657268e-01
-6.62089169e-01 2.68129259e-01 -1.23040128e+00 1.09599149e+00
4.97335345e-01 -9.94544506e-01 -5.11608422e-01 2.83082157e-01
1.02110207e+00 -6.37619019e-01 -6.16861165e-01 2.84582317e-01
6.27408981e-01 -5.63751400e-01 9.04568911e-01 -9.04601991e-01
5.64373076e-01 -4.41517293e-01 -2.04330191e-01 -1.89220667e+00
-1.08549476e-01 -8.54719698e-01 1.36606142e-01 7.08701730e-01
7.11604297e-01 -6.81639791e-01 5.43436408e-01 5.20457566e-01
-4.80509289e-02 -4.56910163e-01 -6.78710759e-01 -8.12265098e-01
5.80489993e-01 -3.22401702e-01 5.73411226e-01 9.54042614e-01
1.42721131e-01 4.68354911e-01 -6.25144780e-01 6.41572535e-01
1.24258779e-01 -9.95913520e-02 9.16800857e-01 -1.74868286e-01
-4.32948798e-01 -4.06358153e-01 -5.69332898e-01 -7.94386804e-01
3.75163481e-02 -6.61103368e-01 4.27997947e-01 -8.45829427e-01
-2.86654830e-01 1.58265576e-01 -2.89372385e-01 4.56594974e-01
-2.98251212e-01 5.03570855e-01 5.20258844e-01 3.69788073e-02
1.75605789e-02 3.43829036e-01 1.09291124e+00 -5.90663731e-01
1.58220544e-01 2.30377778e-01 -7.28392661e-01 5.54574013e-01
1.12314713e+00 -7.78099000e-01 -4.64062393e-01 -5.93502879e-01
3.87517184e-01 -3.19342643e-01 1.50218874e-01 -6.83784842e-01
-4.88192886e-02 -4.78259437e-02 4.21604306e-01 -4.43622246e-02
2.75333434e-01 -8.33795130e-01 -1.14866421e-01 7.08568215e-01
-7.18781471e-01 4.39174503e-01 4.45713192e-01 3.75415564e-01
-9.91701111e-02 -1.66418999e-02 9.84037101e-01 -1.35325834e-01
-1.78115025e-01 -5.26376925e-02 -1.14356256e+00 -1.36026964e-01
9.72678840e-01 -1.73594728e-01 -4.00255024e-01 -6.20667815e-01
-5.83307743e-01 -7.12823123e-02 7.92036116e-01 9.09267008e-01
4.78131741e-01 -1.09804595e+00 -6.72234893e-01 1.64750472e-01
-7.64921010e-02 -5.33794940e-01 -1.91350445e-01 2.35886082e-01
-4.30356115e-01 2.52874732e-01 -3.92198771e-01 -4.78796750e-01
-1.52552831e+00 9.37884510e-01 6.46395683e-01 1.69224575e-01
-3.24874222e-01 7.40878940e-01 2.60927528e-01 -5.12432039e-01
9.59768370e-02 -8.62613171e-02 3.31553996e-01 -3.58357370e-01
3.84681076e-01 -1.71805441e-01 2.07546532e-01 -3.54229957e-01
-6.37696162e-02 -6.50331676e-02 -5.17649829e-01 -2.55841345e-01
7.34225035e-01 1.45072162e-01 9.96430218e-02 3.50631982e-01
1.33721578e+00 -2.15719193e-02 -9.64371443e-01 -1.23654410e-01
-2.09696218e-01 -5.00179648e-01 -2.53451496e-01 -1.07119012e+00
-5.91494679e-01 1.00015116e+00 7.66707838e-01 5.39513886e-01
1.04983079e+00 -6.55243635e-01 1.02964926e+00 6.40893757e-01
1.55050635e-01 -9.51403916e-01 3.31788301e-01 5.31548142e-01
7.84562111e-01 -1.09013152e+00 -6.12731338e-01 4.95398156e-02
-7.67326355e-01 8.75625134e-01 6.60310209e-01 -1.46708965e-01
-5.78608215e-02 5.13728797e-01 8.92246127e-01 1.75350457e-01
-8.95822048e-01 5.82407594e-01 -6.48136511e-02 7.04555631e-01
2.32658535e-01 2.87454039e-01 1.73225030e-01 -3.26651841e-01
-7.44723141e-01 -5.70263326e-01 9.66701627e-01 9.27874863e-01
-3.07615548e-01 -1.28572917e+00 -7.82211602e-01 1.27113700e-01
-8.52374792e-01 -2.73216605e-01 -9.73414779e-01 5.13617754e-01
-1.32349888e-02 1.16129470e+00 -4.02011387e-02 -6.93316817e-01
4.54745054e-01 1.23670153e-01 2.71638542e-01 -7.80994833e-01
-9.39643264e-01 -2.26452589e-01 1.24973521e-01 -3.28939170e-01
1.09766558e-01 -6.04978085e-01 -9.62162614e-01 -5.15185416e-01
-1.17726391e-02 2.31607668e-02 8.14315438e-01 1.02335119e+00
1.55177698e-01 4.92559224e-01 1.08359957e+00 -4.73621011e-01
-9.58290994e-01 -9.81920540e-01 5.38610518e-02 6.27789974e-01
6.45407200e-01 8.88713375e-02 -4.74556327e-01 6.80449381e-02] | [6.027517795562744, 8.15935230255127] |
1c4a886d-8e09-4905-9d36-10195535ccd8 | despite-super-human-performance-current-llms | 2212.06295 | null | https://arxiv.org/abs/2212.06295v1 | https://arxiv.org/pdf/2212.06295v1.pdf | Despite "super-human" performance, current LLMs are unsuited for decisions about ethics and safety | Large language models (LLMs) have exploded in popularity in the past few years and have achieved undeniably impressive results on benchmarks as varied as question answering and text summarization. We provide a simple new prompting strategy that leads to yet another supposedly "super-human" result, this time outperforming humans at common sense ethical reasoning (as measured by accuracy on a subset of the ETHICS dataset). Unfortunately, we find that relying on average performance to judge capabilities can be highly misleading. LLM errors differ systematically from human errors in ways that make it easy to craft adversarial examples, or even perturb existing examples to flip the output label. We also observe signs of inverse scaling with model size on some examples, and show that prompting models to "explain their reasoning" often leads to alarming justifications of unethical actions. Our results highlight how human-like performance does not necessarily imply human-like understanding or reasoning. | ['Abraham J. Fetterman', 'Ellie Kitanidis', 'Joshua Albrecht'] | 2022-12-13 | null | null | null | null | ['common-sense-reasoning'] | ['reasoning'] | [ 3.80766660e-01 8.30642998e-01 -1.45245492e-02 -3.49287897e-01
-8.44956100e-01 -7.19565213e-01 1.01575661e+00 2.96813428e-01
-5.03765523e-01 9.36663628e-01 5.02230108e-01 -6.66324615e-01
1.00356698e-01 -4.54389423e-01 -6.93157017e-01 -2.77510434e-01
4.47866321e-01 4.25552368e-01 -1.83875605e-01 -3.65378886e-01
7.88407326e-01 3.81114818e-02 -1.00116587e+00 3.03198904e-01
1.08974743e+00 3.25031966e-01 -5.49025357e-01 8.81067753e-01
1.00959271e-01 1.72694147e+00 -1.16968942e+00 -1.09655964e+00
8.24329779e-02 -5.39138794e-01 -1.10602820e+00 -3.26450080e-01
7.49809206e-01 -4.16029841e-01 -1.21477328e-01 1.03553689e+00
6.26404047e-01 -4.64988947e-02 7.54438579e-01 -1.30213213e+00
-9.26218092e-01 8.88076007e-01 -2.71841735e-01 2.08368972e-01
6.44967318e-01 6.36903405e-01 1.14331996e+00 -2.21891448e-01
6.79465771e-01 1.51451552e+00 9.46984112e-01 8.26915264e-01
-1.32221019e+00 -5.35971642e-01 6.91734022e-03 7.79243112e-02
-7.82409012e-01 -5.65306723e-01 5.50450683e-01 -6.26146615e-01
1.00963032e+00 6.34272873e-01 1.93696067e-01 1.62768495e+00
4.54728246e-01 4.59896803e-01 1.30763924e+00 -2.11722657e-01
3.11312050e-01 4.07614022e-01 1.14504851e-01 4.38394547e-01
7.63355255e-01 -8.06070790e-02 -5.03398180e-01 -6.86029434e-01
1.80115059e-01 -5.21712005e-01 -2.11365938e-01 2.60756880e-01
-1.23293293e+00 9.15139437e-01 1.50142476e-01 1.84787676e-01
-4.57583219e-01 4.48682457e-01 5.06290913e-01 4.99054521e-01
4.41260964e-01 1.45349503e+00 -3.92914623e-01 -5.15483797e-01
-6.80744231e-01 6.30621731e-01 1.18545246e+00 4.11621898e-01
3.03250283e-01 1.14571042e-01 -3.09248179e-01 5.41710436e-01
-2.53720433e-01 4.17534500e-01 6.02028131e-01 -1.70348251e+00
4.26256984e-01 4.63468581e-01 5.85000992e-01 -1.46138251e+00
-3.25797468e-01 -3.65017623e-01 -7.26264656e-01 5.17263114e-02
6.23068988e-01 -2.86121815e-01 -2.64219910e-01 1.94621563e+00
-1.35763511e-01 -3.03731054e-01 1.71087131e-01 6.62368953e-01
4.05225128e-01 5.86489320e-01 3.86267334e-01 -1.30895421e-01
1.06110442e+00 -5.87151229e-01 -6.36584461e-01 -5.82016408e-01
9.80029762e-01 -5.39867103e-01 1.46263945e+00 4.65740681e-01
-1.09184718e+00 -1.36556604e-03 -7.33002901e-01 -1.30536094e-01
-1.53359666e-01 -4.76250172e-01 5.86398602e-01 6.92802429e-01
-8.46189082e-01 7.04181433e-01 -1.62075058e-01 -3.49024028e-01
4.64516491e-01 -9.53840837e-02 -2.90872186e-01 2.00795576e-01
-1.28725767e+00 1.42809772e+00 1.36085926e-02 -2.83585370e-01
-6.62453115e-01 -7.47106612e-01 -6.16317809e-01 1.11062257e-02
5.46464205e-01 -8.65907609e-01 1.46292436e+00 -1.13876569e+00
-1.13358510e+00 1.00010967e+00 -4.51500341e-03 -9.89231884e-01
1.06681156e+00 -5.88030219e-01 -1.28811732e-01 6.35442063e-02
2.01304734e-01 5.22426367e-01 7.13370979e-01 -1.28960657e+00
6.88023642e-02 -8.55021272e-03 3.34141374e-01 5.39423339e-02
-3.31427485e-01 2.77930558e-01 6.86724126e-01 -6.86769187e-01
-5.41094303e-01 -7.11181402e-01 -2.91804880e-01 -1.73887208e-01
-9.28083718e-01 -2.76782393e-01 3.96693736e-01 -8.60285580e-01
1.38027859e+00 -1.52738571e+00 -1.74567699e-01 -1.50536671e-01
4.82732207e-01 4.58770782e-01 -3.93890254e-02 5.92652380e-01
-2.12182458e-02 9.92692232e-01 -3.32255155e-01 -6.48189038e-02
3.59303027e-01 1.65770352e-02 -6.42993748e-01 2.23897263e-01
2.51399010e-01 1.14263248e+00 -8.88870537e-01 -7.07064807e-01
-1.03316464e-01 -5.15550375e-02 -6.86959743e-01 5.23295365e-02
-4.27021742e-01 1.97064921e-01 -2.44384810e-01 2.51519799e-01
1.41795605e-01 -4.93911117e-01 -6.90606609e-02 4.00236160e-01
3.53312612e-01 5.61142802e-01 -4.95089293e-01 8.68028343e-01
-2.79434651e-01 8.53805006e-01 -1.64518639e-01 -7.11244106e-01
5.59005558e-01 3.10754836e-01 -1.88977897e-01 -5.88065326e-01
7.96684995e-02 1.15041383e-01 2.94148296e-01 -8.04305255e-01
6.43823922e-01 -3.08227450e-01 -2.68239409e-01 8.86268437e-01
-5.91347754e-01 -5.17458975e-01 -2.10851595e-01 6.14839494e-01
1.35396230e+00 -3.37331861e-01 4.55924362e-01 -1.74391285e-01
4.40188766e-01 2.22920299e-01 5.01737893e-01 1.25490582e+00
-3.10166240e-01 3.36512715e-01 1.08143568e+00 -4.91737872e-01
-1.23681891e+00 -7.50308514e-01 2.91282892e-01 9.25698817e-01
-1.74862146e-01 -5.07953405e-01 -1.07097816e+00 -7.68034935e-01
1.64517000e-01 1.57015789e+00 -7.14854896e-01 -3.94906372e-01
-3.97208780e-01 -4.96688098e-01 1.08278131e+00 2.75117785e-01
5.27169228e-01 -1.17639196e+00 -9.05379713e-01 5.58340736e-02
-3.50718111e-01 -1.06382287e+00 -2.01335058e-01 -2.45288447e-01
-6.06670201e-01 -9.95292187e-01 -3.78616929e-01 4.78491187e-03
6.84331417e-01 -1.32753313e-01 1.41682231e+00 5.99520624e-01
1.76902488e-02 3.86279225e-01 -1.68352544e-01 -6.19483769e-01
-1.23410988e+00 -1.87385138e-02 1.18120812e-01 -4.53462452e-01
2.82507241e-01 -4.33267117e-01 -4.06025857e-01 1.30725950e-01
-8.51509869e-01 -6.99047297e-02 4.68558401e-01 6.26107812e-01
-2.98108608e-01 -4.05939579e-01 1.00571835e+00 -1.40691423e+00
1.24505043e+00 -4.28692788e-01 -1.70911312e-01 2.76684314e-01
-8.77907932e-01 4.20712344e-02 9.28704441e-01 -4.11198586e-01
-7.55700409e-01 -8.44772518e-01 5.74345179e-02 -1.52613446e-02
-2.30425879e-01 1.67338476e-01 1.94271043e-01 1.35450244e-01
1.20035923e+00 -8.21246728e-02 1.42094597e-01 -2.26062253e-01
3.18626374e-01 6.54417515e-01 6.04827046e-01 -7.84508705e-01
1.01293027e+00 1.65948495e-01 -2.41202995e-01 -5.12101948e-01
-1.06537116e+00 3.58690411e-01 1.02214150e-01 -8.53731409e-02
6.95268214e-01 -5.58565855e-01 -9.91418123e-01 3.80401134e-01
-1.42749023e+00 -5.15598834e-01 -2.48787045e-01 -5.99978082e-02
-4.94538695e-01 5.89621842e-01 -6.52741253e-01 -8.88166130e-01
-4.97874647e-01 -8.09814870e-01 6.93221271e-01 3.80035117e-02
-1.17782593e+00 -1.00238907e+00 -6.60753623e-02 7.33782589e-01
7.34431922e-01 4.49438542e-01 1.29605138e+00 -1.19693768e+00
-1.61448047e-01 -5.39063275e-01 3.06807924e-02 5.94786465e-01
-1.73498631e-01 8.05436373e-02 -9.10278261e-01 1.65939674e-01
1.47168323e-01 -5.68695009e-01 4.80538845e-01 -1.61079481e-01
9.75517094e-01 -1.07822692e+00 -4.23657037e-02 -1.18866928e-01
1.01000249e+00 -1.34107232e-01 7.19444752e-01 3.85044456e-01
6.05420768e-01 8.95553291e-01 3.07759136e-01 3.17955524e-01
4.64695096e-01 3.48347038e-01 2.06661582e-01 2.81429142e-01
3.00474852e-01 -5.96259654e-01 5.14820516e-01 3.02433997e-01
3.10941339e-01 -3.03854883e-01 -1.13782310e+00 3.83887261e-01
-1.68114197e+00 -1.27771533e+00 -1.79184735e-01 2.12503266e+00
1.03742850e+00 4.48110670e-01 -6.06446676e-02 -9.88324881e-02
6.87646389e-01 3.26825351e-01 -6.07611895e-01 -1.01191485e+00
-1.24702282e-01 -1.41770765e-01 3.25942963e-01 7.86590755e-01
-5.44193745e-01 9.60193396e-01 7.45681572e+00 6.66748226e-01
-7.73832619e-01 2.30604097e-01 1.03302634e+00 -1.06991522e-01
-8.99546921e-01 1.62515834e-01 -1.92518353e-01 6.75204754e-01
1.09566605e+00 -6.82550728e-01 3.17670614e-01 7.32855797e-01
3.83815676e-01 -2.64656037e-01 -1.25840104e+00 6.42164886e-01
1.41927734e-01 -1.30995858e+00 2.29411796e-01 -7.16720149e-02
7.74107933e-01 -4.34184462e-01 -3.13002951e-02 2.90570050e-01
7.63268411e-01 -1.51107883e+00 6.75373554e-01 5.26917517e-01
3.14679235e-01 -5.51470220e-01 7.18832493e-01 9.03281510e-01
1.80837587e-01 -2.11666688e-01 -3.71600300e-01 -6.56159699e-01
1.25769973e-01 4.80874866e-01 -8.41868699e-01 3.48693617e-02
2.60394365e-01 2.01355398e-01 -8.57409477e-01 4.51285869e-01
-5.23101628e-01 8.03340256e-01 -1.12349264e-01 -2.80889362e-01
1.41896188e-01 1.43131867e-01 6.81734264e-01 1.12698972e+00
-1.75336406e-01 1.13168143e-01 -4.05069232e-01 1.04152930e+00
-3.68604749e-01 1.64805874e-02 -9.91006613e-01 -3.70426685e-01
6.06346309e-01 1.00330794e+00 -1.64886862e-01 -7.00273097e-01
2.97089424e-02 7.54383266e-01 2.13680744e-01 3.23488593e-01
-8.74440491e-01 -2.63058752e-01 5.15246928e-01 1.03928424e-01
-4.41060424e-01 1.18463032e-01 -7.18163669e-01 -1.11896074e+00
1.09467931e-01 -1.42625678e+00 7.70041645e-02 -1.13579512e+00
-1.20652997e+00 3.23497295e-01 -2.12205410e-01 -6.32251024e-01
-5.07722557e-01 -2.29508877e-01 -8.82961035e-01 6.32492125e-01
-1.07118332e+00 -4.88696814e-01 -1.01252355e-01 3.70020270e-02
1.26031503e-01 9.18414891e-02 8.18548858e-01 -1.59693018e-01
-4.21601057e-01 7.46509850e-01 -4.10428405e-01 3.72898094e-02
9.80475187e-01 -1.21474445e+00 6.94513500e-01 7.21388161e-01
-7.21474886e-02 8.52582455e-01 1.39394224e+00 -6.44387186e-01
-9.96448517e-01 -6.59479737e-01 1.25158811e+00 -1.32874501e+00
8.05765331e-01 -1.34680688e-01 -9.26715791e-01 9.76572335e-01
3.88995886e-01 -6.68145478e-01 5.12669265e-01 -1.17398679e-01
-6.90676510e-01 2.48012438e-01 -1.35920870e+00 1.04815423e+00
9.71913576e-01 -7.59917438e-01 -8.96184146e-01 7.51622677e-01
1.04218149e+00 -5.72185069e-02 -5.90485275e-01 2.03300640e-01
3.35425854e-01 -1.16572726e+00 6.30650938e-01 -1.26422942e+00
9.68321979e-01 1.35041043e-01 -1.27589121e-01 -1.22433817e+00
-3.36700422e-03 -1.14105618e+00 7.59396926e-02 1.24954736e+00
4.69624281e-01 -9.16175544e-01 5.93774498e-01 1.25612545e+00
2.46793136e-01 -7.23523080e-01 -5.99680603e-01 -6.92789376e-01
5.76439202e-01 -4.23064739e-01 4.02203947e-01 1.16262627e+00
1.75306201e-01 3.33120555e-01 -5.02151251e-01 -2.18371172e-02
5.31630337e-01 -3.38945866e-01 1.07796919e+00 -1.14635766e+00
-2.79518008e-01 -6.72355831e-01 -3.67969304e-01 -6.09418213e-01
4.22860533e-01 -6.56379879e-01 -1.36904597e-01 -1.50357437e+00
3.21064830e-01 -4.00791556e-04 2.12731883e-01 4.22758847e-01
-5.27817547e-01 -5.90377264e-02 3.43945295e-01 2.16798604e-01
-6.96680009e-01 7.22714439e-02 9.43572998e-01 1.41394511e-03
2.37296358e-01 -2.35405743e-01 -1.52459681e+00 9.14935410e-01
7.76752770e-01 -5.32712340e-01 -2.35474706e-01 -3.40242058e-01
8.01304579e-01 -6.74147457e-02 9.08574283e-01 -9.27783668e-01
6.01490438e-02 -2.98857570e-01 3.23866270e-02 2.31409058e-01
5.15002124e-02 -4.69130754e-01 3.89344580e-02 7.63310134e-01
-1.09079742e+00 1.41512588e-01 2.90207621e-02 4.11358207e-01
1.61799043e-01 -3.46018404e-01 7.14974642e-01 -2.85415322e-01
-2.25906774e-01 -3.54240537e-01 -5.17304182e-01 6.09424829e-01
9.51282322e-01 -2.81238928e-02 -9.63849127e-01 -1.05262673e+00
-2.36505792e-01 2.85028249e-01 8.44771206e-01 3.22958082e-01
6.17719479e-02 -7.71185219e-01 -8.47259045e-01 -5.71062505e-01
-1.01269223e-03 -3.88869882e-01 9.39435959e-02 8.48621070e-01
-6.35377169e-01 3.66635829e-01 1.05893008e-01 -1.61871865e-01
-9.13091481e-01 5.15688717e-01 5.95339954e-01 -2.63234138e-01
-3.84712070e-01 8.93265963e-01 6.10564016e-02 -4.67394441e-01
1.49652466e-01 -7.24914148e-02 1.97275564e-01 -1.27533749e-01
4.96278197e-01 4.36642259e-01 -2.91161865e-01 -2.19767407e-01
-3.12695712e-01 1.56312659e-01 -3.72477770e-01 -1.48815334e-01
1.05573976e+00 6.10809140e-02 -1.78800538e-01 4.19537157e-01
8.42406094e-01 2.51970321e-01 -9.69746768e-01 -4.80963252e-02
2.60184199e-01 -4.30652469e-01 -5.37321508e-01 -1.30719650e+00
-2.52358049e-01 9.30353165e-01 -3.14324766e-01 5.98984241e-01
5.04387081e-01 -1.26451388e-01 7.64370143e-01 9.36923504e-01
4.34670210e-01 -1.16337359e+00 2.75356472e-01 2.81307220e-01
1.01124442e+00 -1.26851559e+00 2.17193410e-01 1.10837124e-01
-1.16640961e+00 6.30353153e-01 7.57035255e-01 -4.16620597e-02
-1.57708883e-01 1.87483784e-02 2.67040461e-01 -1.22055650e-01
-1.28694963e+00 6.55263960e-01 -3.72952461e-01 5.06151557e-01
4.25574213e-01 1.12632670e-01 -4.02973562e-01 5.78469336e-01
-6.90403879e-01 -1.01182528e-01 1.09510136e+00 5.51249266e-01
-6.29023969e-01 -8.13944876e-01 -4.85351324e-01 6.58660471e-01
-8.56431186e-01 -1.77890658e-01 -1.09715950e+00 8.72157395e-01
-4.55425858e-01 1.15738356e+00 -2.85896510e-01 -3.86768579e-01
1.35589257e-01 3.53511125e-01 5.30658923e-02 -4.95275170e-01
-9.31510746e-01 -8.04062545e-01 6.59751415e-01 -6.88508928e-01
8.24147835e-02 -4.10186917e-01 -1.03730822e+00 -9.76062477e-01
-7.52261207e-02 2.71348745e-01 3.49867105e-01 1.12239134e+00
5.84780335e-01 9.28934738e-02 3.36481303e-01 -4.31006551e-01
-1.28034675e+00 -1.08014381e+00 -1.60732251e-02 7.07939446e-01
3.37178081e-01 -1.53700650e-01 -8.29718173e-01 -2.20031872e-01] | [10.372310638427734, 7.607524394989014] |
7ca2142e-d6c8-48cd-9e27-2af2fa841c2f | strategize-before-teaching-a-conversational | 2302.13496 | null | https://arxiv.org/abs/2302.13496v1 | https://arxiv.org/pdf/2302.13496v1.pdf | Strategize Before Teaching: A Conversational Tutoring System with Pedagogy Self-Distillation | Conversational tutoring systems (CTSs) aim to help students master educational material with natural language interaction in the form of a dialog. CTSs have become a key pillar in educational data mining research. A key challenge in CTSs is to engage the student in the conversation while exposing them to a diverse set of teaching strategies, akin to a human teacher, thereby, helping them learn in the process. Different from previous work that generates responses given the strategies as input, we propose to jointly predict teaching strategies and generate tutor responses accordingly, which fits a more realistic application scenario. We benchmark several competitive models on three dialog tutoring datasets and propose a unified framework that combines teaching response generation and pedagogical strategy prediction, where a self-distillation mechanism is adopted to guide the teaching strategy learning and facilitate tutor response generation. Our experiments and analyses shed light on how teaching strategies affect dialog tutoring. | ['Kam-Fai Wong', 'Xingshan Zeng', 'Mrinmaya Sachan', 'Lingzhi Wang'] | 2023-02-27 | null | null | null | null | ['response-generation'] | ['natural-language-processing'] | [ 4.42405820e-01 4.95322406e-01 -2.32980818e-01 -5.57563424e-01
-1.95736617e-01 -8.13102782e-01 8.52771521e-01 3.42228711e-01
-2.21437916e-01 6.73577249e-01 2.80655831e-01 -6.82464004e-01
-3.73990946e-02 -1.05950880e+00 -2.04433948e-01 -4.02256787e-01
7.72406161e-01 6.32462561e-01 6.10318303e-01 -8.64095628e-01
3.83713871e-01 4.12410736e-01 -1.68498874e+00 4.02306706e-01
1.57248855e+00 3.72722238e-01 4.25064087e-01 8.34986866e-01
-5.50298333e-01 1.43490469e+00 -7.83925474e-01 -4.80447888e-01
-2.69798577e-01 -9.49902952e-01 -1.24185753e+00 2.77432334e-02
3.43376428e-01 -3.77665251e-01 -3.29215586e-01 7.89920390e-01
2.81926632e-01 5.52509546e-01 8.41162205e-01 -1.27533674e+00
-2.92851269e-01 7.85994053e-01 4.40209731e-02 3.32825514e-03
7.19745398e-01 1.00267135e-01 8.80184412e-01 -2.05819875e-01
3.70015979e-01 1.33300912e+00 -4.84325690e-03 1.36759400e+00
-1.13597882e+00 -3.64744812e-01 2.32064232e-01 3.61279249e-01
-3.02410156e-01 -2.54026383e-01 8.43632400e-01 -5.19596756e-01
6.90581128e-02 4.31258053e-01 7.97516227e-01 1.18211412e+00
1.60631388e-01 1.38024163e+00 1.13109231e+00 -5.25376856e-01
1.73956960e-01 6.81748569e-01 6.18549228e-01 4.47641194e-01
-3.48573208e-01 -8.82202759e-02 -8.09892774e-01 -1.82524443e-01
4.48797137e-01 1.06564753e-01 -5.12067139e-01 -4.02586520e-01
-7.39640296e-01 1.22744024e+00 3.76152657e-02 1.24692746e-01
-7.41278455e-02 -5.59927523e-01 2.66950130e-01 9.95912075e-01
-1.19403817e-01 9.11200941e-01 -6.38381422e-01 -5.13478577e-01
-4.43112046e-01 5.65344393e-01 1.31045520e+00 1.02359605e+00
3.35918128e-01 -1.56777546e-01 -6.06243670e-01 9.31842208e-01
2.63903290e-01 2.31847912e-01 7.72137463e-01 -9.12284791e-01
3.67806584e-01 1.30531991e+00 -1.05057016e-01 -3.69007528e-01
5.32515906e-02 2.58629024e-01 -2.84732521e-01 7.34569225e-03
8.00592899e-01 -5.54007828e-01 -3.33099991e-01 1.66706121e+00
6.39006615e-01 7.21772611e-02 2.92262793e-01 6.21692598e-01
1.37698913e+00 4.71815795e-01 2.00901046e-01 -2.04864964e-01
1.50136840e+00 -1.08439279e+00 -8.15972924e-01 8.28632638e-02
7.65119195e-01 -6.55414164e-01 1.25862253e+00 6.29457533e-01
-1.18728888e+00 -5.07158160e-01 -5.08123636e-01 -1.28583645e-03
-1.65406793e-01 -4.63295013e-01 3.60637993e-01 5.37750304e-01
-7.76100516e-01 4.07519788e-01 -2.71910191e-01 -2.26415232e-01
5.09849451e-02 2.39425525e-01 6.26134798e-02 3.25304627e-01
-1.24726593e+00 8.84555221e-01 -2.20628396e-01 -5.43809116e-01
-8.42368543e-01 -9.90670443e-01 -8.00384343e-01 3.75461310e-01
5.23382664e-01 -5.32218695e-01 2.04934764e+00 -6.30790114e-01
-2.38794684e+00 6.94186687e-01 6.28383532e-02 -3.24022055e-01
5.80767512e-01 -1.38566628e-01 4.15777683e-01 -2.25623958e-02
-3.40130150e-01 8.12330127e-01 3.48413259e-01 -8.95219684e-01
-7.10118294e-01 -2.74919629e-01 3.12177390e-01 6.72440469e-01
-5.11493683e-01 -1.77451044e-01 -5.11837639e-02 -2.43928820e-01
-3.98624577e-02 -8.58794808e-01 -2.65150636e-01 -6.84203446e-01
-2.56565064e-01 -9.33878005e-01 8.97698998e-01 2.93399990e-02
1.22271287e+00 -1.45872605e+00 4.61328298e-01 -2.65496410e-02
6.20133817e-01 4.22915488e-01 -7.42214769e-02 6.74748719e-01
1.82174921e-01 -2.05853671e-01 2.62575269e-01 2.37703393e-03
2.53728896e-01 2.94870697e-02 -5.26947141e-01 -3.79649103e-01
-1.69217929e-01 6.53865457e-01 -1.11411321e+00 -2.35476643e-01
4.65752929e-01 -9.91982445e-02 -7.83116579e-01 1.32613122e+00
-6.98834896e-01 8.71782064e-01 -8.37211967e-01 1.78344533e-01
1.65030792e-01 4.45366912e-02 1.59469664e-01 6.92833006e-01
-7.98353180e-02 8.14016044e-01 -7.02165544e-01 1.16078448e+00
-7.49560714e-01 5.62377393e-01 7.86465481e-02 -9.04714346e-01
1.23987412e+00 4.26354378e-01 3.18813115e-01 -7.94455290e-01
1.35042533e-01 -3.27812999e-01 5.19696832e-01 -6.30443633e-01
4.08668995e-01 1.99985951e-01 -1.34916410e-01 9.05129850e-01
1.72675729e-01 -4.77977931e-01 1.28079414e-01 5.35231352e-01
1.10299408e+00 2.93972380e-02 2.37608895e-01 -1.54247344e-01
8.77627671e-01 -1.18015811e-01 2.42286518e-01 9.22200203e-01
-3.10929358e-01 -1.17093891e-01 7.96810150e-01 -4.57552344e-01
-2.54665256e-01 -6.84010923e-01 1.06716134e-01 1.69882751e+00
-9.38652009e-02 -3.18495572e-01 -1.13588905e+00 -1.07024956e+00
-4.01858874e-02 9.02229488e-01 -4.13876683e-01 -4.12929922e-01
-6.23311043e-01 -7.28558749e-02 2.64870346e-01 2.53081143e-01
4.93368417e-01 -1.54585385e+00 -8.48620772e-01 3.96352649e-01
-1.84363514e-01 -8.48719716e-01 -5.08271515e-01 2.27730855e-01
-7.19455659e-01 -1.30775392e+00 -2.90182084e-01 -8.27221513e-01
5.91458201e-01 4.00845170e-01 1.30057895e+00 6.13761365e-01
8.32338184e-02 7.76407003e-01 -2.68397927e-01 -4.79608953e-01
-9.74720359e-01 2.58377969e-01 -2.44022417e-03 -4.06372398e-01
7.60332048e-01 -6.53983057e-01 -5.62551737e-01 5.06583214e-01
-6.68495774e-01 5.79057276e-01 2.54630417e-01 1.01030672e+00
-1.18312277e-01 -3.04523528e-01 8.72348547e-01 -1.53689814e+00
1.31855285e+00 -5.73615491e-01 -8.27561617e-01 4.61946666e-01
-7.81696260e-01 1.68378145e-01 9.20564413e-01 -5.74792743e-01
-1.34082830e+00 5.02686873e-02 -1.55201599e-01 -2.07128227e-01
-6.76674187e-01 2.68374085e-01 -3.23706061e-01 -2.00187773e-01
5.12603223e-01 4.40873832e-01 1.14262730e-01 -3.98684770e-01
2.51347572e-01 6.32124662e-01 1.49010807e-01 -1.21538556e+00
5.88026166e-01 -5.60756326e-01 -3.66793096e-01 -8.26085806e-01
-8.55731964e-01 -4.23993170e-01 -6.73714936e-01 -5.57899415e-01
6.29984558e-01 -5.02212465e-01 -1.69387853e+00 4.18855697e-01
-1.02590740e+00 -9.34523582e-01 -1.31929860e-01 -5.88921346e-02
-5.57904780e-01 -2.33527645e-01 -6.81906819e-01 -7.50674129e-01
-7.87290335e-02 -1.62152350e+00 3.96060348e-01 9.23031926e-01
-5.36342382e-01 -1.30659616e+00 2.57495940e-01 9.79288459e-01
4.58767205e-01 -2.42987528e-01 1.17878795e+00 -1.37748349e+00
-3.78280371e-01 4.31747854e-01 4.23960358e-01 -3.98207977e-02
1.92045256e-01 -1.03877001e-01 -9.79772687e-01 2.29615614e-01
3.40809315e-01 -8.10434580e-01 2.15308040e-01 -1.46092907e-01
1.20435011e+00 -5.54746866e-01 4.24576327e-02 4.57073897e-02
4.60007101e-01 5.37528813e-01 1.30794674e-01 1.94898292e-01
4.58250374e-01 1.32757342e+00 6.26731873e-01 7.06396550e-02
8.16881657e-01 7.09506750e-01 -1.91581938e-02 4.17631239e-01
7.87870586e-02 -3.51285011e-01 4.61773992e-01 9.75402474e-01
3.22131574e-01 -1.96237177e-01 -9.43118811e-01 4.07546103e-01
-1.91841459e+00 -7.33202934e-01 -2.96847403e-01 1.96165502e+00
1.63672245e+00 -1.07216731e-01 2.89692581e-01 -1.10334694e-01
2.38034889e-01 1.58172045e-02 -5.29112756e-01 -7.83594608e-01
6.59733832e-01 4.05355275e-01 -1.19811937e-01 8.86749208e-01
-3.66833240e-01 1.16978490e+00 5.58022070e+00 4.27501172e-01
-1.05781603e+00 -4.72654432e-01 5.25245607e-01 2.40837395e-01
-4.38935310e-01 -2.22197264e-01 -1.06625307e+00 3.66604626e-01
1.29543579e+00 -3.99860710e-01 3.25762510e-01 7.13619828e-01
3.58258009e-01 -1.20251678e-01 -1.49527144e+00 2.34995306e-01
-2.21157491e-01 -1.15462661e+00 1.42538801e-01 8.23346432e-03
5.73639870e-01 -7.20845044e-01 -8.96887947e-03 6.64263666e-01
1.08171189e+00 -8.80675077e-01 -1.60682469e-03 3.89579177e-01
-1.12056628e-01 -6.43127203e-01 3.86860311e-01 7.34531581e-01
-7.19140947e-01 -9.98089761e-02 -2.42552198e-02 -1.66498020e-01
-3.87919694e-01 -4.90456998e-01 -1.57425332e+00 9.12102312e-02
3.18090111e-01 1.76411867e-01 -5.36188960e-01 6.35375798e-01
-6.35606647e-01 9.62038517e-01 2.89218873e-01 -6.69950068e-01
4.55688201e-02 -5.75926542e-01 3.67821336e-01 7.53683448e-01
-1.27519503e-01 8.35634947e-01 4.52478021e-01 7.74252176e-01
-1.33936465e-01 2.65035152e-01 -5.95748663e-01 6.65323809e-02
9.61386979e-01 1.19727457e+00 -2.89898843e-01 -2.08283737e-01
-3.98544312e-01 5.01413465e-01 3.61787468e-01 1.90981716e-01
-2.56927282e-01 -1.63431361e-01 5.75098872e-01 2.37397984e-01
-2.63823986e-01 2.81453460e-01 -2.20345944e-01 -9.04760003e-01
-5.09710252e-01 -1.60839045e+00 2.49855980e-01 -5.37358999e-01
-9.97819364e-01 2.48336107e-01 1.14204705e-01 -9.93256450e-01
-6.51456714e-01 -4.79203492e-01 -1.36292839e+00 7.34720707e-01
-1.34614515e+00 -5.03372669e-01 -5.35658360e-01 7.02220976e-01
1.02283847e+00 -5.84366500e-01 9.61567283e-01 -1.70784771e-01
-7.62227356e-01 7.77422905e-01 -4.81287271e-01 -1.65508856e-04
9.55258489e-01 -1.79798210e+00 -7.09713176e-02 2.44301811e-01
-7.55571201e-02 8.19000065e-01 8.66037428e-01 -3.92067283e-01
-1.62801957e+00 -6.17056429e-01 8.37497294e-01 -4.53471869e-01
7.20062912e-01 -2.31758714e-01 -1.22407985e+00 3.34589273e-01
8.03505480e-01 -8.35712492e-01 1.13491333e+00 2.03082487e-01
9.54554453e-02 2.27468297e-01 -9.07523036e-01 9.78319108e-01
5.47148824e-01 -2.16447875e-01 -1.17120683e+00 2.29189828e-01
7.09256887e-01 -8.30598176e-01 -9.15859640e-01 5.73245669e-03
1.56182051e-01 -9.71803308e-01 5.53521752e-01 -1.01262832e+00
6.68483019e-01 2.24754587e-01 4.87828583e-01 -1.52606833e+00
3.39182436e-01 -9.48212683e-01 -2.91228414e-01 1.32917547e+00
9.64840874e-02 -4.72680211e-01 1.00731993e+00 1.14175522e+00
-2.05966488e-01 -9.39760506e-01 -2.60699779e-01 7.54910186e-02
4.44283038e-01 2.11732522e-01 6.88770831e-01 1.08166814e+00
5.04450381e-01 7.16666996e-01 -2.43438478e-03 -2.62424499e-01
2.30235249e-01 2.95310497e-01 1.24288499e+00 -1.71441519e+00
-1.04832448e-01 -7.56815910e-01 1.80331230e-01 -1.54990435e+00
5.03298759e-01 -5.18663168e-01 2.66306192e-01 -1.00592577e+00
-1.77516807e-02 -5.34960270e-01 1.46162853e-01 4.42299396e-01
-5.45421541e-01 -6.46142125e-01 4.14797701e-02 -2.17111096e-01
-6.63847327e-01 8.30406129e-01 1.67008781e+00 1.22192040e-01
-5.21332800e-01 7.81310856e-01 -9.70816016e-01 8.82042766e-01
5.94979703e-01 -2.74411768e-01 -1.05205905e+00 1.10134050e-01
-3.34685385e-01 7.96655774e-01 -1.71496928e-01 -4.78644311e-01
7.96801686e-01 -8.15565825e-01 -4.98573035e-02 -2.04861626e-01
5.70815941e-03 -6.37495697e-01 -7.57430017e-01 3.67151052e-01
-1.02943432e+00 7.96639249e-02 9.44527313e-02 1.89022750e-01
-2.07773671e-01 -6.42508090e-01 7.56908059e-01 -4.12753113e-02
-3.32759231e-01 1.42626897e-01 -8.65721107e-01 1.33095130e-01
1.09846520e+00 3.56681049e-02 -6.33025408e-01 -8.61329556e-01
-6.34436786e-01 9.76794124e-01 2.04205215e-01 6.69644237e-01
6.81593418e-01 -1.09988725e+00 -5.83725572e-01 1.58787072e-01
-4.23157209e-04 3.54110539e-01 3.02967839e-02 4.00860608e-01
-1.87155068e-01 5.66820204e-01 -2.20760599e-01 -5.64064562e-01
-1.71010256e+00 -4.42043245e-02 4.51231122e-01 -4.03018445e-01
-3.01856756e-01 9.06901479e-01 3.85118604e-01 -1.03482699e+00
6.98910415e-01 -1.61623552e-01 -9.64448452e-01 2.06411630e-01
8.10119987e-01 1.01658113e-01 -4.20396775e-03 6.87130168e-02
5.09619415e-01 -1.71836540e-01 -6.09559953e-01 -1.38187734e-03
1.25910521e+00 -3.45945582e-02 1.81334510e-01 3.18396807e-01
3.72914076e-01 1.25949651e-01 -1.19606733e+00 -4.88840580e-01
3.62043262e-01 -1.25717238e-01 -3.60682309e-01 -1.02390337e+00
-6.88341439e-01 8.73972237e-01 1.72306821e-01 6.40505373e-01
8.05415690e-01 -1.15546152e-01 5.06539524e-01 8.35815847e-01
-4.01015058e-02 -8.47429514e-01 5.96521735e-01 8.43596637e-01
6.50385737e-01 -1.43548107e+00 -3.20010334e-01 -3.73721808e-01
-8.05862606e-01 1.26686656e+00 1.64535153e+00 2.68434733e-01
2.31803492e-01 1.61103591e-01 2.94235647e-01 -2.71504372e-01
-1.49633586e+00 1.86244428e-01 5.04375041e-01 2.74942279e-01
1.14425874e+00 -1.14045411e-01 -2.41673172e-01 7.54119098e-01
-5.74426711e-01 -4.79150057e-01 8.17239821e-01 9.74070132e-01
-7.18263090e-01 -1.78359830e+00 -3.05097103e-01 4.06534165e-01
-1.69690549e-02 -5.54865748e-02 -1.07345068e+00 5.10234177e-01
-3.83365542e-01 1.28456712e+00 -2.77200758e-01 -4.02808458e-01
5.88519752e-01 4.61284220e-01 3.56833130e-01 -1.12743235e+00
-1.41449976e+00 -3.54909480e-01 -1.16276719e-01 -4.06774253e-01
-6.66827112e-02 -7.47832239e-01 -1.12866938e+00 -3.98413002e-01
-3.21470141e-01 7.90548444e-01 1.17885940e-01 1.10993063e+00
-4.42691222e-02 7.86173165e-01 9.45412695e-01 -1.78942576e-01
-8.79952669e-01 -1.03592086e+00 8.17519426e-02 2.67274469e-01
2.99896955e-01 -5.40107965e-01 -1.65249884e-01 -4.47146781e-02] | [12.228671073913574, 8.074029922485352] |
c1fff60c-bece-4426-b854-0653ccbf97b8 | linear-span-network-for-object-skeleton | 1807.09601 | null | http://arxiv.org/abs/1807.09601v1 | http://arxiv.org/pdf/1807.09601v1.pdf | Linear Span Network for Object Skeleton Detection | Robust object skeleton detection requires to explore rich representative
visual features and effective feature fusion strategies. In this paper, we
first re-visit the implementation of HED, the essential principle of which can
be ideally described with a linear reconstruction model. Hinted by this, we
formalize a Linear Span framework, and propose Linear Span Network (LSN)
modified by Linear Span Units (LSUs), which minimize the reconstruction error
of convolutional network. LSN further utilizes subspace linear span beside the
feature linear span to increase the independence of convolutional features and
the efficiency of feature integration, which enlarges the capability of fitting
complex ground-truth. As a result, LSN can effectively suppress the cluttered
backgrounds and reconstruct object skeletons. Experimental results validate the
state-of-the-art performance of the proposed LSN. | ['Chang Liu', 'Qixiang Ye', 'Fei Qin', 'Wei Ke'] | 2018-07-25 | linear-span-network-for-object-skeleton-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Chang_Liu_Linear_Span_Network_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Chang_Liu_Linear_Span_Network_ECCV_2018_paper.pdf | eccv-2018-9 | ['object-skeleton-detection'] | ['computer-vision'] | [ 5.22710271e-02 -3.28435972e-02 -1.64535865e-02 -1.11285314e-01
-2.69332290e-01 -3.16627771e-02 3.30348551e-01 -6.12230599e-01
-2.55099446e-01 4.57888633e-01 1.59737557e-01 1.42180845e-01
-2.88354546e-01 -7.11967170e-01 -8.06896329e-01 -7.56582439e-01
2.59618253e-01 -3.14799726e-01 4.85942185e-01 -9.77956131e-02
-1.21547721e-01 6.63758099e-01 -1.39415634e+00 6.03448935e-02
7.90580034e-01 1.15600145e+00 1.02817535e-01 5.09615652e-02
-2.78870106e-01 5.90415299e-01 -1.53109476e-01 -4.61085796e-01
3.47859651e-01 -2.01509312e-01 -2.21604511e-01 2.82157540e-01
4.60602641e-01 -6.31307364e-01 -8.35501313e-01 1.13759649e+00
6.39177561e-01 -8.23940933e-02 5.01951933e-01 -1.07379389e+00
-5.33969402e-01 5.56056261e-01 -6.44215822e-01 -2.80583538e-02
1.29486278e-01 3.27041477e-01 6.40229404e-01 -1.25146806e+00
3.47043157e-01 1.52109623e+00 9.11938727e-01 4.35964078e-01
-8.97000015e-01 -7.69716144e-01 3.01184595e-01 1.11361831e-01
-1.36464596e+00 -5.03953159e-01 8.65701973e-01 -1.27894238e-01
4.01752204e-01 1.79090738e-01 9.10997331e-01 1.02548742e+00
2.28979588e-02 1.16265595e+00 6.41851306e-01 -3.54323030e-01
-3.53608489e-01 -3.85844439e-01 2.21127674e-01 1.18376493e+00
6.45083725e-01 1.00098468e-01 -4.57322180e-01 1.87657848e-01
1.26761544e+00 3.32066745e-01 -5.61743021e-01 -8.44672799e-01
-1.28631091e+00 5.63832939e-01 6.91625476e-01 -3.75587605e-02
-1.46662757e-01 1.76451728e-01 4.52330887e-01 -2.71316051e-01
-1.63669750e-01 -1.49655983e-01 -1.25711054e-01 5.57287753e-01
-8.06555748e-01 -6.12045154e-02 2.55422950e-01 9.41122472e-01
7.23667860e-01 4.09052402e-01 -3.20401341e-01 7.07014859e-01
6.36430860e-01 6.01340055e-01 4.16836053e-01 -9.29927170e-01
2.94623166e-01 9.41522717e-01 -1.41271889e-01 -8.93807948e-01
-4.78522599e-01 -8.31435025e-01 -1.07076955e+00 1.80329368e-01
3.61914545e-01 -3.98474522e-02 -8.86021197e-01 1.47113693e+00
4.65147823e-01 3.91745418e-01 3.77676710e-02 9.45665479e-01
1.12448084e+00 4.16733533e-01 -1.10626176e-01 -3.28174502e-01
1.27697027e+00 -1.09873021e+00 -6.34148479e-01 -3.88115980e-02
-4.65120263e-02 -5.96529722e-01 8.89415920e-01 3.47865999e-01
-8.48585010e-01 -9.99781191e-01 -1.22273052e+00 -1.47470430e-01
5.92706427e-02 4.93711561e-01 7.28639364e-01 3.77334505e-01
-4.90321785e-01 6.00588739e-01 -9.95473325e-01 -2.53953606e-01
6.60980880e-01 2.99856663e-01 -4.18526620e-01 -1.36977449e-01
-9.10056591e-01 4.28466946e-01 6.63253486e-01 5.52292466e-01
-9.45794821e-01 -5.12209535e-01 -9.31893349e-01 1.13462053e-01
5.11365175e-01 -9.91941690e-01 1.12432659e+00 -6.50808096e-01
-1.37138510e+00 3.56654465e-01 1.48575634e-01 -2.91037500e-01
7.67241597e-01 -5.12075365e-01 -4.56893355e-01 2.60853261e-01
1.84243135e-02 4.76525664e-01 1.01523399e+00 -1.25919509e+00
-5.26335537e-01 -2.65618533e-01 -2.05835998e-01 4.63410579e-02
-5.20368636e-01 -1.04607277e-01 -7.75832236e-01 -5.77805221e-01
4.98143137e-01 -5.60253263e-01 -3.23985457e-01 5.83297610e-01
-3.26098502e-01 -1.29029274e-01 9.67879891e-01 -6.77355051e-01
1.35243428e+00 -2.30269384e+00 -1.42955482e-01 9.29806381e-02
3.37922692e-01 4.61067051e-01 -8.03915560e-02 1.10101141e-01
-1.11270964e-01 -3.77797335e-01 -2.81655490e-01 -2.74075568e-01
6.17154054e-02 2.92306811e-01 -1.23180702e-01 6.44341052e-01
3.83118480e-01 1.02582586e+00 -5.79294562e-01 -9.72686946e-01
4.72946346e-01 5.18126726e-01 -2.21110955e-01 4.02991027e-02
1.20361466e-02 2.55132526e-01 -8.09424043e-01 8.91156614e-01
9.03650820e-01 -1.57822177e-01 -2.19359934e-01 -8.13542545e-01
-1.76668122e-01 -2.82832682e-01 -1.63237357e+00 1.82786524e+00
5.26106283e-02 1.68209150e-01 1.95090130e-01 -9.16836262e-01
9.72476661e-01 -1.11101829e-01 6.14169300e-01 -4.03968126e-01
1.89189136e-01 1.64773270e-01 -2.24686667e-01 -5.75057030e-01
-1.25165090e-01 1.76123958e-02 2.13473365e-01 -1.76119447e-01
1.08437784e-01 5.38855135e-01 -6.86917454e-02 -1.12518705e-02
7.40350306e-01 5.71688294e-01 5.92922032e-01 -1.73398271e-01
7.28044450e-01 -5.01229823e-01 8.98343921e-01 4.63788748e-01
-2.62348831e-01 5.59758723e-01 2.79591203e-01 -5.73095679e-01
-7.12848842e-01 -1.30802023e+00 -3.95074606e-01 7.45394289e-01
4.86048907e-01 -3.62247795e-01 -6.22888863e-01 -8.39194894e-01
-2.62557324e-02 2.23099530e-01 -5.69187164e-01 -2.58274585e-01
-8.88224542e-01 -5.05548358e-01 5.29075801e-01 7.60032296e-01
1.05613863e+00 -9.56791580e-01 -9.44422483e-01 1.63567498e-01
-1.20474420e-01 -8.65817189e-01 -4.09968406e-01 6.42749593e-02
-9.24270511e-01 -1.15720296e+00 -7.75120974e-01 -7.34977841e-01
6.11101747e-01 5.25753856e-01 8.56476665e-01 2.75886416e-01
-5.16948879e-01 3.41129124e-01 -2.78387934e-01 -2.35688508e-01
1.27588734e-01 -2.41844684e-01 6.07814454e-02 2.11871281e-01
1.10837109e-01 -7.20999300e-01 -8.77163470e-01 2.85060316e-01
-9.63990927e-01 1.70600563e-01 1.04848039e+00 8.77517343e-01
6.74027503e-01 -1.32767722e-01 4.85440522e-01 -2.59001106e-01
-5.42208143e-02 -2.02383831e-01 -4.04814571e-01 5.47950804e-01
-4.93414074e-01 6.44667679e-03 6.45470500e-01 -4.21105534e-01
-1.16558135e+00 3.77335995e-01 -1.51930988e-01 -8.42923820e-01
-7.42549971e-02 1.82125509e-01 -6.59120381e-01 -3.68036121e-01
3.71994168e-01 4.22244549e-01 1.57507867e-01 -6.54181719e-01
4.37451631e-01 4.93522078e-01 7.69751668e-01 -4.96837467e-01
8.70842099e-01 5.57304919e-01 2.96958864e-01 -7.33721375e-01
-9.94291246e-01 -3.98373008e-01 -7.05553710e-01 -4.46448356e-01
7.36646354e-01 -9.81852829e-01 -6.46236897e-01 4.56470251e-01
-1.03302014e+00 3.20326030e-01 -3.55947733e-01 5.02087057e-01
-5.50956845e-01 9.12043154e-01 -4.89727825e-01 -7.04920232e-01
-3.91223550e-01 -9.65547800e-01 1.12492537e+00 5.68919301e-01
4.80084270e-01 -4.94620681e-01 -3.21418762e-01 1.87011451e-01
1.21217027e-01 4.98394191e-01 5.16438961e-01 -2.04953909e-01
-7.12667406e-01 -4.10259217e-02 -5.18009543e-01 5.78999400e-01
-7.00255930e-02 5.49287573e-02 -1.01227379e+00 -3.53567928e-01
3.16729359e-02 -2.33703911e-01 1.26605248e+00 4.88679707e-01
1.15802896e+00 -2.56510496e-01 -5.55944324e-01 9.83867824e-01
1.51352119e+00 -2.47682437e-01 7.52503693e-01 2.55762964e-01
8.84734571e-01 3.72040123e-01 5.81099629e-01 5.93999624e-01
8.88388827e-02 4.47570533e-01 5.96692204e-01 -4.48343366e-01
-5.44275343e-01 -4.04682726e-01 4.12784845e-01 6.47176027e-01
-1.89607114e-01 2.36990064e-01 -4.03226614e-01 1.99160084e-01
-2.09113789e+00 -9.45575833e-01 -2.14491189e-01 1.92984605e+00
5.14256597e-01 2.03404173e-01 1.83371469e-01 7.77020231e-02
7.65744627e-01 4.46465537e-02 -7.69350946e-01 3.94480765e-01
-2.66659200e-01 -2.03037560e-02 4.58860546e-01 -6.98509673e-03
-1.23259377e+00 6.66200042e-01 5.90229559e+00 1.10386622e+00
-8.07562470e-01 -1.43910751e-01 1.54133037e-01 6.60571903e-02
5.82818389e-02 -1.89968169e-01 -9.33646500e-01 3.36519122e-01
1.65365994e-01 5.48558943e-02 2.06161201e-01 1.09013104e+00
-3.59348841e-02 2.43118420e-01 -9.01678026e-01 1.14529526e+00
5.41662835e-02 -1.10171282e+00 3.02018225e-01 -1.19371310e-01
4.63343322e-01 -1.65280953e-01 -1.25372007e-01 5.19607484e-01
-1.55004337e-02 -8.71971786e-01 8.09543133e-01 8.52785170e-01
7.90363133e-01 -6.89414680e-01 6.62103057e-01 4.74525213e-01
-1.59820187e+00 -3.80158335e-01 -7.30675697e-01 1.57745853e-01
2.73336351e-01 5.52262902e-01 -2.85413146e-01 1.03860879e+00
8.83812428e-01 9.32590663e-01 -7.72465289e-01 1.39451492e+00
-1.31379768e-01 4.21504170e-01 -4.50420380e-01 3.20321292e-01
1.12001561e-01 -2.45735422e-01 5.19589901e-01 1.40768421e+00
3.82765621e-01 7.08254129e-02 5.27172923e-01 6.89595044e-01
1.49976403e-01 1.21209867e-01 -4.34705049e-01 2.80878067e-01
1.54927194e-01 1.39878786e+00 -7.84127653e-01 -1.77471131e-01
-6.29218757e-01 8.10379207e-01 2.64080048e-01 2.85547912e-01
-1.07189333e+00 -2.54771709e-01 3.04976523e-01 2.28619650e-01
6.01393700e-01 -1.08334586e-01 4.05741669e-03 -1.29696441e+00
2.23391980e-01 -6.11566842e-01 3.36118698e-01 -7.05193698e-01
-1.31562304e+00 3.33862036e-01 -7.08745746e-03 -1.58796287e+00
3.72013628e-01 -8.50971043e-01 -5.85332394e-01 5.08375704e-01
-1.54060888e+00 -1.83157253e+00 -6.36493325e-01 8.23340237e-01
6.46619260e-01 -1.02855578e-01 5.29532194e-01 3.50814939e-01
-8.37757587e-01 5.40346682e-01 1.95682161e-02 3.47573757e-01
5.38471580e-01 -7.84246743e-01 -4.98199128e-02 1.07475138e+00
2.92766821e-02 8.73288989e-01 3.95816147e-01 -6.89003587e-01
-1.31183434e+00 -1.20456576e+00 7.34392256e-02 -2.32021119e-02
4.18141335e-01 1.79299642e-03 -8.57323647e-01 5.90772688e-01
-1.22170657e-01 4.56775665e-01 4.26555037e-01 -3.03680986e-01
-3.74664456e-01 -3.81468594e-01 -8.46089602e-01 6.12607658e-01
1.39939439e+00 -2.55914610e-02 -6.94694221e-01 1.61178321e-01
8.37685823e-01 -2.78260469e-01 -7.34438717e-01 9.16021109e-01
8.96964490e-01 -1.20739329e+00 1.39910185e+00 -4.91444379e-01
1.90211043e-01 -7.11230099e-01 -2.04919234e-01 -5.10896266e-01
-7.36931801e-01 -3.19837481e-01 -4.22558069e-01 1.34797788e+00
-2.00893104e-01 -2.86186039e-01 6.12190425e-01 1.02252916e-01
-4.20962363e-01 -9.52998281e-01 -9.76121664e-01 -8.57421517e-01
-2.86750644e-01 -3.00387412e-01 4.76548254e-01 5.13229012e-01
-3.55181247e-01 3.17205161e-01 -4.34722900e-01 2.38531560e-01
9.79200780e-01 2.45659053e-01 7.02747881e-01 -1.38999414e+00
-3.25225443e-01 -4.36227977e-01 -4.99678820e-01 -1.32482541e+00
7.77382478e-02 -7.52311349e-01 -2.97880378e-02 -1.64005637e+00
4.94515449e-01 -8.04718733e-02 -5.62386155e-01 5.56521654e-01
-3.40563625e-01 2.64307529e-01 1.88125044e-01 3.13745409e-01
-9.68805909e-01 9.95214701e-01 1.44148064e+00 -4.99415733e-02
1.63070753e-01 2.26549685e-01 -6.32878900e-01 1.00168741e+00
6.01323664e-01 -1.32267550e-01 -2.82703996e-01 -3.31737399e-01
-2.16891125e-01 -1.26019403e-01 7.44974971e-01 -1.29949725e+00
2.30910063e-01 -1.40964225e-01 1.03774881e+00 -7.64395773e-01
2.37939283e-01 -1.03425658e+00 1.57909662e-01 5.83461463e-01
1.20259218e-01 -3.63270521e-01 4.19601165e-02 8.50613952e-01
-7.01121986e-02 -3.29197943e-01 9.98691201e-01 -1.99118063e-01
-7.92628765e-01 3.94063294e-01 2.14570120e-01 -5.11246383e-01
1.09495914e+00 -6.77806973e-01 -4.40367535e-02 1.37421861e-01
-5.37468255e-01 3.22103649e-01 3.46920162e-01 1.73232570e-01
9.19006884e-01 -1.58941686e+00 -6.25937104e-01 5.15184104e-01
-1.43807046e-02 2.77139097e-01 4.03524399e-01 9.10279989e-01
-5.02267957e-01 2.18875006e-01 -3.59616518e-01 -5.77107966e-01
-1.00708151e+00 7.78843880e-01 3.42658699e-01 -1.76704988e-01
-9.69105065e-01 6.62304103e-01 5.33526003e-01 -3.10413361e-01
4.47815031e-01 -2.68405646e-01 -3.12675923e-01 -1.08216695e-01
4.83668357e-01 6.27870560e-01 -2.36099243e-01 -4.89302337e-01
-3.16419721e-01 9.41369593e-01 1.03188850e-01 2.91956067e-01
1.23578668e+00 -1.24426492e-01 -4.82505299e-02 2.07382500e-01
9.21301305e-01 -2.04578340e-02 -1.62198508e+00 -3.14404368e-01
-3.94890308e-01 -3.56391966e-01 1.85135147e-03 -4.69173789e-01
-1.18079233e+00 8.42843890e-01 7.96922445e-01 -1.75147861e-01
1.26060951e+00 -8.89167339e-02 8.79780650e-01 3.75074983e-01
2.93087363e-01 -8.10760379e-01 3.02274704e-01 2.05119908e-01
9.22100365e-01 -1.09085333e+00 2.57053256e-01 -5.55892348e-01
-2.28690267e-01 1.47236335e+00 9.44720566e-01 -2.56215155e-01
5.42826831e-01 1.19695462e-01 -2.43418664e-01 -1.28667861e-01
-5.06968573e-02 -3.78867269e-01 3.57099205e-01 6.37595713e-01
-2.03052014e-02 -2.09170878e-01 -4.43059027e-01 1.10857058e+00
1.30520776e-01 3.33428979e-02 -2.49862727e-02 7.22035229e-01
-7.75945067e-01 -6.99721813e-01 -5.50978482e-01 3.20634991e-01
-2.52787977e-01 9.68713537e-02 -3.75632644e-02 8.83797824e-01
6.34229124e-01 5.52697062e-01 -3.79335791e-01 -2.89307654e-01
5.10667920e-01 -9.48372707e-02 4.87598509e-01 -2.22106010e-01
-1.43718466e-01 5.20564079e-01 -2.70339727e-01 -6.83587968e-01
-4.97868896e-01 -6.74081087e-01 -1.30342996e+00 2.01127872e-01
-7.16195107e-01 -2.99491286e-01 2.89752156e-01 7.48514652e-01
5.95023297e-02 7.56291330e-01 3.73356134e-01 -8.85353267e-01
-1.02107239e+00 -9.42491114e-01 -5.83024263e-01 1.41437352e-01
2.91113377e-01 -9.74912524e-01 -1.57537803e-01 1.71001479e-01] | [9.468149185180664, -0.6149161458015442] |
3f4f1fda-c592-4751-96e0-11926c712127 | definition-extraction-from-mathematical-texts | null | null | https://aclanthology.org/2021.konvens-1.9 | https://aclanthology.org/2021.konvens-1.9.pdf | Definition Extraction from Mathematical Texts on Graph Theory in German and English | null | ['Fritz Kliche', 'Theresa Kruse'] | null | null | null | null | konvens-ws-2021-9 | ['definition-extraction'] | ['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] | [-7.485416889190674, 3.7858314514160156] |
c8317dd6-31ca-415c-900b-f74fae6eab78 | deep-parametric-3d-filters-for-joint-video | 2207.01797 | null | https://arxiv.org/abs/2207.01797v2 | https://arxiv.org/pdf/2207.01797v2.pdf | Deep Parametric 3D Filters for Joint Video Denoising and Illumination Enhancement in Video Super Resolution | Despite the quality improvement brought by the recent methods, video super-resolution (SR) is still very challenging, especially for videos that are low-light and noisy. The current best solution is to subsequently employ best models of video SR, denoising, and illumination enhancement, but doing so often lowers the image quality, due to the inconsistency between the models. This paper presents a new parametric representation called the Deep Parametric 3D Filters (DP3DF), which incorporates local spatiotemporal information to enable simultaneous denoising, illumination enhancement, and SR efficiently in a single encoder-and-decoder network. Also, a dynamic residual frame is jointly learned with the DP3DF via a shared backbone to further boost the SR quality. We performed extensive experiments, including a large-scale user study, to show our method's effectiveness. Our method consistently surpasses the best state-of-the-art methods on all the challenging real datasets with top PSNR and user ratings, yet having a very fast run time. | ['Jiaya Jia', 'Chi-Wing Fu', 'RuiXing Wang', 'Xiaogang Xu'] | 2022-07-05 | null | null | null | null | ['video-super-resolution', 'video-denoising'] | ['computer-vision', 'computer-vision'] | [ 2.82256633e-01 -5.45014858e-01 3.00467084e-03 -2.57193595e-01
-9.39272463e-01 -1.98373049e-01 1.59976333e-01 -7.10021377e-01
-4.63800207e-02 6.79804862e-01 6.46218836e-01 2.01897636e-01
9.86595526e-02 -3.79906207e-01 -8.53674054e-01 -6.31679773e-01
8.27741399e-02 -4.20734614e-01 4.51428920e-01 -2.45402798e-01
1.76009342e-01 2.89370298e-01 -1.68118072e+00 4.63664651e-01
1.05122089e+00 1.05681527e+00 5.10991335e-01 6.47958398e-01
2.32124597e-01 8.94226611e-01 -2.17554346e-01 -2.17076272e-01
4.26897585e-01 -4.64762211e-01 -3.21907461e-01 2.21581951e-01
6.74553335e-01 -8.04986596e-01 -5.83828509e-01 1.07096517e+00
9.06965256e-01 2.30560675e-01 -5.67444004e-02 -7.14822114e-01
-1.04570425e+00 8.40076804e-02 -8.80915701e-01 2.85955966e-01
6.60026431e-01 1.80989578e-01 7.81944096e-01 -1.11222970e+00
6.73995137e-01 1.49856341e+00 6.43392980e-01 4.86415923e-01
-1.31414998e+00 -7.45599687e-01 -1.34592624e-02 3.95994425e-01
-1.53266418e+00 -8.09737623e-01 6.76493347e-01 -4.30912748e-02
7.13546276e-01 -3.99885811e-02 4.74483877e-01 1.12640035e+00
1.10735610e-01 6.74956441e-01 1.32328081e+00 2.18278617e-02
3.59331071e-02 -2.05008551e-01 -3.66361141e-01 4.38279539e-01
-1.64059609e-01 2.55566806e-01 -9.72706497e-01 1.39340982e-01
1.38589084e+00 -1.44302160e-01 -7.12024271e-01 -1.37503251e-01
-1.08410692e+00 3.51219296e-01 2.58030474e-01 1.33931696e-01
-3.65850866e-01 2.56090701e-01 1.62804887e-01 3.64341468e-01
6.76581562e-01 -1.14150075e-02 -3.71714413e-01 -1.34745583e-01
-1.08070862e+00 2.38753587e-01 3.61087710e-01 7.78769016e-01
4.69630569e-01 2.11206317e-01 -4.18992519e-01 1.12021029e+00
2.71897227e-01 3.44262451e-01 -3.88260297e-02 -1.63099980e+00
3.30581278e-01 -6.09737262e-02 5.73646069e-01 -1.00394213e+00
4.77850698e-02 -6.73763931e-01 -1.01586711e+00 2.06654474e-01
-1.39280409e-01 1.13741882e-01 -7.26778030e-01 1.76246440e+00
1.28486648e-01 7.39231229e-01 -1.96006089e-01 1.28085315e+00
8.53204966e-01 8.11221004e-01 -1.43209979e-01 -4.42136228e-01
1.09630132e+00 -1.04949462e+00 -1.01771307e+00 -9.43299681e-02
-2.72406004e-02 -8.99252236e-01 9.68325853e-01 7.25772858e-01
-1.52709687e+00 -1.09656298e+00 -9.92036045e-01 -4.39601183e-01
4.75732446e-01 6.23093210e-02 4.06541586e-01 4.84625667e-01
-1.52320039e+00 7.65696228e-01 -7.19022453e-01 -1.40065268e-01
5.19070387e-01 1.32695973e-01 -2.18434572e-01 -6.15501285e-01
-1.20022571e+00 8.90425980e-01 -1.32437512e-01 1.39207318e-01
-1.04463661e+00 -7.06519663e-01 -8.28695357e-01 -1.48958340e-02
3.82272035e-01 -9.02814448e-01 1.02257848e+00 -9.63212669e-01
-1.85349381e+00 6.10881925e-01 -4.71517622e-01 -1.73291340e-01
4.61331457e-01 -5.66726744e-01 -6.26263499e-01 2.56021380e-01
-1.19172297e-01 6.18572474e-01 1.13366151e+00 -1.54276013e+00
-6.02673113e-01 -2.69481003e-01 5.26137725e-02 4.64969993e-01
-2.20362797e-01 4.77220148e-01 -8.67191613e-01 -7.20847607e-01
1.28174439e-01 -4.70625162e-01 -2.23203555e-01 2.28195533e-01
8.83269235e-02 1.07393555e-01 7.41893709e-01 -1.15205419e+00
1.32070827e+00 -2.46248984e+00 4.41146284e-01 -3.62205327e-01
2.37725988e-01 2.26338208e-01 -3.39730948e-01 -2.66447738e-02
-3.72055732e-02 2.56270077e-02 -1.69771641e-01 -5.99604845e-01
-2.02043548e-01 1.44250631e-01 -9.29843560e-02 5.45787573e-01
2.93849260e-02 5.73711395e-01 -8.31818819e-01 -4.30573732e-01
1.92228749e-01 1.13669920e+00 -6.55192375e-01 3.14766735e-01
6.75467402e-02 7.52161264e-01 -1.10983059e-01 6.08857214e-01
9.92763102e-01 -2.51601726e-01 -8.39437693e-02 -5.75196147e-01
-2.05652073e-01 4.41330206e-03 -1.54470563e+00 2.14802027e+00
-4.86535549e-01 5.76722801e-01 5.00970125e-01 -5.32921851e-01
7.86341250e-01 4.22946602e-01 4.99100566e-01 -1.01335955e+00
-1.92975476e-01 3.26861322e-01 -4.63525265e-01 -6.09234989e-01
5.68029463e-01 -6.41783059e-04 6.09912574e-01 1.21029997e-02
-8.75469446e-02 1.08569779e-01 1.44679323e-01 1.35986671e-01
9.56843197e-01 7.27894187e-01 -3.98862772e-02 -3.95038500e-02
5.68511128e-01 -8.47173870e-01 8.05775285e-01 3.86242360e-01
-1.26257718e-01 1.04880381e+00 1.66380718e-01 -2.46791899e-01
-1.12636650e+00 -1.26812577e+00 -2.51124762e-02 9.51907635e-01
5.39019346e-01 -4.42796826e-01 -5.48650503e-01 -1.28426582e-01
-4.15938467e-01 5.04917860e-01 -2.17064053e-01 8.16565380e-02
-4.84534889e-01 -5.09942591e-01 1.02687076e-01 4.17266548e-01
8.97066355e-01 -6.61234975e-01 -3.38463008e-01 2.36273617e-01
-6.15376949e-01 -1.47601163e+00 -6.14401460e-01 -4.05993551e-01
-8.20286155e-01 -6.83967769e-01 -9.77607489e-01 -5.55904329e-01
2.09482223e-01 7.44118750e-01 1.13555622e+00 1.43985733e-01
-6.65740520e-02 2.98798103e-02 -3.87249202e-01 2.47795746e-01
-1.16514668e-01 -5.67593992e-01 1.79487597e-02 2.22927421e-01
-2.94102281e-01 -7.38827884e-01 -1.02894497e+00 4.72136378e-01
-1.02300274e+00 2.03161180e-01 5.09018302e-01 7.60511100e-01
7.28925586e-01 4.77774918e-01 4.25446242e-01 -4.37171608e-01
3.39744657e-01 -3.28096658e-01 -4.71142203e-01 -2.04876866e-02
-4.24944967e-01 -2.33641207e-01 5.49903691e-01 -2.84361839e-01
-1.66102922e+00 -1.81149706e-01 -3.17228168e-01 -4.85141128e-01
4.39461321e-03 2.09519148e-01 -3.10365677e-01 -2.59653032e-01
5.17799139e-01 2.65024126e-01 -7.89949968e-02 -9.07176256e-01
3.49546760e-01 5.40750325e-01 7.31714368e-01 -2.11787164e-01
8.49800825e-01 6.70503199e-01 -5.51481098e-02 -7.14656711e-01
-1.02938855e+00 -3.90272826e-01 -4.13418412e-01 -1.88248813e-01
7.32185304e-01 -1.52520037e+00 -3.54265094e-01 6.15717649e-01
-1.17241263e+00 -2.77210891e-01 -8.44692811e-02 4.36702371e-01
-6.24722600e-01 7.13645339e-01 -9.76710200e-01 -6.21554136e-01
-2.38666445e-01 -1.30442548e+00 1.03058314e+00 2.77295232e-01
3.63192976e-01 -5.69329262e-01 -1.56837106e-01 5.29820859e-01
9.10529017e-01 5.34648891e-04 4.79880899e-01 4.33485806e-01
-8.63416433e-01 3.02282721e-01 -6.82152569e-01 7.74451554e-01
9.10618082e-02 -3.45754586e-02 -9.40614462e-01 -5.44089079e-01
1.20980181e-01 -5.33433333e-02 1.01677847e+00 5.93980372e-01
1.39279294e+00 1.35568669e-02 1.08179338e-01 1.01423049e+00
1.69873273e+00 -5.76041490e-02 1.11371028e+00 2.63555855e-01
6.46938980e-01 1.54285580e-01 6.95789695e-01 5.74540257e-01
4.51036334e-01 8.49542558e-01 5.95092237e-01 -4.19011176e-01
-5.93456507e-01 -3.36744171e-03 7.87579000e-01 6.93627357e-01
-5.84828019e-01 -1.69894591e-01 -1.45346984e-01 3.57356727e-01
-1.76378798e+00 -1.13673663e+00 -7.36244097e-02 2.17595458e+00
9.27413940e-01 -4.26531546e-02 -4.95442934e-02 2.04415414e-02
5.96675813e-01 5.82203031e-01 -3.58813316e-01 -5.09965532e-02
-4.72256720e-01 3.25285196e-01 4.03284013e-01 4.38887030e-01
-9.11720872e-01 7.12582290e-01 6.52111149e+00 8.69214118e-01
-8.47715020e-01 3.32506090e-01 5.99135697e-01 -2.08799243e-01
-2.43221074e-01 -1.78555459e-01 -6.47531986e-01 5.97029209e-01
8.72136652e-01 1.42823100e-01 8.43034267e-01 5.20966470e-01
8.72350574e-01 -1.99359804e-01 -8.75019312e-01 1.25747573e+00
1.81659862e-01 -1.29296327e+00 -6.37865216e-02 -1.14897579e-01
9.85855460e-01 1.02638807e-02 2.36402065e-01 7.09174648e-02
-4.25963476e-03 -1.04025328e+00 6.43541753e-01 7.60777771e-01
1.11650383e+00 -8.11778367e-01 6.77242279e-01 5.09265214e-02
-1.23885047e+00 -2.63084233e-01 -4.24421310e-01 2.12082043e-01
4.75405633e-01 6.15087688e-01 2.74852723e-01 7.12965190e-01
1.19312871e+00 1.17407572e+00 -3.45395446e-01 1.13924420e+00
-3.91789198e-01 3.74397993e-01 -4.17882800e-02 8.93139422e-01
-2.70146161e-01 -2.27203026e-01 5.86519182e-01 1.13715744e+00
6.23921156e-01 3.65688413e-01 1.00305311e-01 7.45415211e-01
-1.35296881e-01 -1.80624202e-01 -2.68825859e-01 4.54457074e-01
3.06645572e-01 1.20658660e+00 -1.41357407e-01 -2.57301569e-01
-6.47338271e-01 1.39883161e+00 -1.55177906e-01 6.38348877e-01
-1.02420235e+00 7.76903033e-02 8.84370029e-01 2.79510230e-01
4.42834973e-01 -4.17935640e-01 -1.50875255e-01 -1.35020399e+00
1.49437800e-01 -1.10828948e+00 4.01500240e-02 -1.18751895e+00
-1.12267375e+00 4.67564821e-01 -3.08903873e-01 -1.36948860e+00
1.33729473e-01 -2.42418170e-01 -1.47160456e-01 9.03829098e-01
-2.15458226e+00 -9.75502610e-01 -5.62714815e-01 8.79214287e-01
8.01970363e-01 2.22026497e-01 4.20639187e-01 9.28180039e-01
-5.91113687e-01 3.43186647e-01 2.60197878e-01 -3.27344209e-01
1.06220269e+00 -8.78568113e-01 1.94856942e-01 1.19859838e+00
-2.80557781e-01 2.59773999e-01 8.15011263e-01 -4.90237981e-01
-1.67284286e+00 -9.74962234e-01 5.06036878e-01 -9.64794457e-02
3.73559326e-01 -8.66095498e-02 -1.00320435e+00 3.54991853e-01
3.59089673e-01 3.73330712e-01 2.29348868e-01 -2.16896951e-01
-4.07521874e-01 -3.26635301e-01 -1.12987304e+00 6.15526199e-01
1.48754334e+00 -5.12272835e-01 -8.72514099e-02 8.42116550e-02
9.31154609e-01 -5.96712708e-01 -1.06702006e+00 4.62378204e-01
5.05164444e-01 -1.37557685e+00 1.33631384e+00 2.06013247e-02
6.99273407e-01 -7.30489552e-01 -4.92186159e-01 -1.33503234e+00
-6.87598705e-01 -8.46863985e-01 -5.82598388e-01 1.13493729e+00
-1.26514137e-02 -2.33990490e-01 3.57253760e-01 3.94099146e-01
-1.66360900e-01 -5.90553880e-01 -7.38839507e-01 -7.71928012e-01
-6.06945455e-01 -3.46653640e-01 4.04755652e-01 7.13335037e-01
-6.45293593e-01 2.41726965e-01 -9.78617668e-01 3.64140362e-01
1.14881217e+00 -6.43787608e-02 6.59353316e-01 -8.45762789e-01
-4.40011084e-01 -9.49407071e-02 -2.38447174e-01 -1.52483666e+00
-1.57735750e-01 -3.34930658e-01 1.16079167e-01 -1.67777121e+00
4.43209887e-01 6.46077329e-03 -3.69607866e-01 9.81176794e-02
-3.08095515e-01 5.62129021e-01 1.45580694e-01 1.15109310e-01
-8.41430604e-01 6.39461160e-01 1.58982706e+00 3.15972865e-01
-2.76535958e-01 -2.94455677e-01 -7.64883995e-01 6.79499090e-01
4.12061632e-01 -8.48724321e-02 -3.66189986e-01 -8.46137822e-01
2.20873252e-01 4.31552410e-01 4.62039471e-01 -9.52330947e-01
2.35972092e-01 -1.67980060e-01 5.87884009e-01 -5.87514997e-01
6.45666003e-01 -8.28034282e-01 3.54244769e-01 -7.06327567e-03
-1.42664343e-01 -1.42998517e-01 -4.72545139e-02 6.98904157e-01
-2.81016141e-01 5.16181588e-01 1.29491925e+00 -6.40775114e-02
-9.98044610e-01 5.52700460e-01 2.33539697e-02 -7.76718929e-02
6.31610155e-01 -3.82569581e-01 -2.35108972e-01 -6.07865512e-01
-6.00512445e-01 1.21677980e-01 6.96303487e-01 3.59374970e-01
1.03246069e+00 -1.27225816e+00 -1.10676098e+00 1.13346241e-01
-3.57103080e-01 -6.04066104e-02 7.66637683e-01 9.69009161e-01
-3.45397025e-01 -1.27579510e-01 -2.95103848e-01 -6.29102290e-01
-1.29722786e+00 5.33040226e-01 3.63714784e-01 -1.05374344e-01
-8.79126072e-01 8.49580169e-01 2.23901168e-01 2.24112093e-01
3.83255839e-01 5.27564958e-02 -2.06858397e-01 -2.73358881e-01
8.86372566e-01 5.89014530e-01 -2.22002178e-01 -6.60129368e-01
-7.21226335e-02 8.05668950e-01 1.35830149e-01 1.26665039e-02
1.65188134e+00 -8.52347136e-01 -7.29762092e-02 6.59550056e-02
1.05146635e+00 -2.70657185e-02 -1.81698489e+00 -5.87336779e-01
-4.52085137e-01 -1.13757992e+00 3.81711274e-01 -7.87452281e-01
-1.45929801e+00 6.23810887e-01 7.28468359e-01 -2.38649309e-01
1.68825018e+00 -3.44501376e-01 1.12377334e+00 -2.11614922e-01
5.91982007e-01 -1.10159647e+00 2.10527271e-01 3.25631768e-01
9.76526797e-01 -1.33549762e+00 1.98944464e-01 -5.86880684e-01
-6.02088988e-01 9.12652910e-01 4.55131233e-01 -1.79290056e-01
5.31738222e-01 3.78855884e-01 -9.95092168e-02 2.22839549e-01
-9.02334094e-01 -4.94300835e-02 3.07558984e-01 5.38509667e-01
5.61814964e-01 -3.34550977e-01 -1.84935898e-01 3.93951088e-01
4.05584216e-01 5.20812690e-01 5.34828305e-01 4.69824523e-01
-4.50230598e-01 -8.56589854e-01 -3.25584084e-01 1.14623398e-01
-8.06440532e-01 -2.31798396e-01 3.67573142e-01 4.03553367e-01
1.22262828e-01 1.20693886e+00 -3.15891236e-01 -3.47488850e-01
4.03575152e-01 -5.56222618e-01 6.06526792e-01 -2.97242701e-01
-2.19504073e-01 4.60554957e-01 6.37899116e-02 -1.26486886e+00
-7.96555221e-01 -4.68472034e-01 -1.00939202e+00 -4.62338716e-01
-4.04419452e-02 -3.87877971e-01 5.69651484e-01 6.49144053e-01
5.28290331e-01 6.00754499e-01 9.52539265e-01 -1.14973831e+00
-4.10021663e-01 -7.28471577e-01 -8.06708276e-01 4.57363188e-01
5.06768286e-01 -6.53974116e-01 -4.54499722e-01 2.19789967e-01] | [11.111259460449219, -2.0631000995635986] |
1d19c046-7162-4063-9287-1c6c6d9d9a55 | learning-personalized-high-quality-volumetric | 2304.01436 | null | https://arxiv.org/abs/2304.01436v1 | https://arxiv.org/pdf/2304.01436v1.pdf | Learning Personalized High Quality Volumetric Head Avatars from Monocular RGB Videos | We propose a method to learn a high-quality implicit 3D head avatar from a monocular RGB video captured in the wild. The learnt avatar is driven by a parametric face model to achieve user-controlled facial expressions and head poses. Our hybrid pipeline combines the geometry prior and dynamic tracking of a 3DMM with a neural radiance field to achieve fine-grained control and photorealism. To reduce over-smoothing and improve out-of-model expressions synthesis, we propose to predict local features anchored on the 3DMM geometry. These learnt features are driven by 3DMM deformation and interpolated in 3D space to yield the volumetric radiance at a designated query point. We further show that using a Convolutional Neural Network in the UV space is critical in incorporating spatial context and producing representative local features. Extensive experiments show that we are able to reconstruct high-quality avatars, with more accurate expression-dependent details, good generalization to out-of-training expressions, and quantitatively superior renderings compared to other state-of-the-art approaches. | ['yinda zhang', 'Sean Fanello', 'Thabo Beeler', 'Ping Tan', 'Rohit Pandey', 'Sergio Orts-Escolano', 'Mingsong Dou', 'Ruofei Du', 'Abhimitra Meka', 'Di Qiu', 'Danhang Tang', 'Kripasindhu Sarkar', 'Zeng Huang', 'Feitong Tan', 'Ziqian Bai'] | 2023-04-04 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Bai_Learning_Personalized_High_Quality_Volumetric_Head_Avatars_From_Monocular_RGB_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Bai_Learning_Personalized_High_Quality_Volumetric_Head_Avatars_From_Monocular_RGB_CVPR_2023_paper.pdf | cvpr-2023-1 | ['face-model'] | ['computer-vision'] | [-1.30927349e-02 2.56769359e-01 3.29868555e-01 -7.15139568e-01
-7.87041664e-01 -5.77774763e-01 6.86262727e-01 -5.11064351e-01
4.14042697e-05 4.51638013e-01 2.59224623e-01 4.94549811e-01
3.64296287e-01 -6.55117512e-01 -8.37561488e-01 -5.45507908e-01
-1.74767226e-02 5.20440996e-01 -8.55101645e-02 -3.30461472e-01
-9.46608931e-02 1.10910773e+00 -1.95754325e+00 2.65965700e-01
4.69669461e-01 1.33908629e+00 -3.30543071e-01 8.66344571e-01
4.42861505e-02 7.53641665e-01 -6.01751089e-01 -3.58242959e-01
5.23481727e-01 -4.14779902e-01 -5.07217109e-01 3.14464271e-01
9.27700281e-01 -5.84350526e-01 -2.41468713e-01 4.58178252e-01
6.36719227e-01 1.47118762e-01 6.46328747e-01 -1.11079919e+00
-3.33385289e-01 -5.58854222e-01 -4.64276373e-01 -7.92153299e-01
8.97832036e-01 3.06083292e-01 7.05622554e-01 -8.71950209e-01
1.04338074e+00 1.51877069e+00 6.41684353e-01 9.66350198e-01
-1.44788122e+00 -6.04938686e-01 4.98971790e-02 -3.50750446e-01
-1.46037650e+00 -6.10293746e-01 9.43982422e-01 -4.29467112e-01
5.70891976e-01 6.45309687e-01 1.29863811e+00 9.99723971e-01
-3.07734068e-02 4.71992314e-01 1.14902806e+00 -2.78835177e-01
2.87212938e-01 -9.28549916e-02 -7.68205941e-01 1.02482963e+00
-8.15976083e-01 2.46592849e-01 -9.22492743e-01 -2.39444911e-01
1.41976774e+00 -3.55079412e-01 -3.92854720e-01 -7.02446938e-01
-8.50751758e-01 5.92926025e-01 6.28220022e-01 -2.71494687e-01
-3.41822386e-01 4.87098366e-01 -1.08708367e-01 -2.78048888e-02
6.73245013e-01 3.16242993e-01 -6.71662688e-01 -2.76139081e-01
-9.38986599e-01 7.15557933e-01 6.28672898e-01 8.72524679e-01
1.06702530e+00 1.57564029e-01 -9.87400785e-02 4.99791056e-01
3.01534772e-01 7.33603299e-01 -1.53788120e-01 -1.79977906e+00
-2.76698321e-01 5.43823123e-01 2.00071469e-01 -6.06635153e-01
-4.33056831e-01 -1.39205664e-01 -3.76376331e-01 9.52730298e-01
3.41107517e-01 -5.76812364e-02 -1.05338418e+00 1.80185294e+00
9.19943750e-01 5.60410544e-02 -2.21332833e-01 1.31490350e+00
8.35790098e-01 6.35362685e-01 -8.75387117e-02 -1.07048522e-03
1.16895342e+00 -5.12373269e-01 -4.31001067e-01 5.63780330e-02
3.72677356e-01 -6.00530207e-01 1.30756998e+00 4.47613090e-01
-1.32852423e+00 -2.92062253e-01 -6.28368914e-01 -4.84835178e-01
1.01535492e-01 -1.24041110e-01 7.02854931e-01 3.62849236e-01
-1.22921622e+00 6.43561721e-01 -8.05605352e-01 -2.98923969e-01
4.04036015e-01 4.79781836e-01 -5.29669881e-01 2.01058298e-01
-6.64991677e-01 8.12661231e-01 -5.39420962e-01 -4.28155400e-02
-9.41596687e-01 -1.26025665e+00 -1.02145541e+00 -3.27262551e-01
-1.27871528e-01 -1.08708262e+00 1.37257767e+00 -1.25423670e+00
-2.23587322e+00 1.23331010e+00 -1.12491988e-01 2.00776726e-01
7.85452068e-01 -1.29825354e-01 1.54992938e-01 3.03523004e-01
-2.91558295e-01 1.01728833e+00 8.62425566e-01 -1.52125514e+00
-1.02113806e-01 -5.87833643e-01 1.46187872e-01 5.93791664e-01
1.16252013e-01 7.78519064e-02 -5.87032139e-01 -3.77899498e-01
1.44209161e-01 -8.84775817e-01 -2.90393811e-02 1.06395519e+00
-1.18622474e-01 2.40532890e-01 9.95831609e-01 -7.27380872e-01
3.31812739e-01 -1.90530574e+00 4.85528827e-01 2.32521534e-01
1.05112039e-01 -1.90578863e-01 2.48414632e-02 -3.69867943e-02
4.01466005e-02 -3.48608047e-01 -1.21426344e-01 -8.72893870e-01
3.42777781e-02 2.54973769e-01 2.25539710e-02 6.71017587e-01
2.63080984e-01 9.51042116e-01 -7.02721715e-01 -4.42436963e-01
4.26043749e-01 1.27763522e+00 -7.44965136e-01 7.64134586e-01
-6.86762631e-01 1.19072437e+00 -3.06536436e-01 7.94175923e-01
6.71579421e-01 1.76635325e-01 -9.58001092e-02 -2.26710409e-01
-1.52658761e-01 3.18800285e-02 -8.40219021e-01 2.34210229e+00
-7.26732373e-01 4.81358767e-01 5.76111436e-01 -1.84108004e-01
9.73021924e-01 2.43542776e-01 5.80472827e-01 -8.57195556e-01
2.88402468e-01 1.74007583e-02 -7.58842409e-01 -3.84569556e-01
2.74365723e-01 -4.47810739e-01 -6.14441605e-03 9.57720354e-02
1.19537283e-02 -9.25191820e-01 -7.59067774e-01 -1.16719648e-01
6.43288553e-01 9.16762412e-01 -2.39069581e-01 -1.56564966e-01
1.99388117e-01 -2.83594012e-01 3.67099583e-01 -1.75980479e-01
2.14057192e-01 1.10526419e+00 4.99357194e-01 -6.26197278e-01
-1.05113769e+00 -1.16713536e+00 -1.51605889e-01 1.09641385e+00
-1.41790271e-01 -2.78936833e-01 -9.23315227e-01 -1.93744391e-01
4.81619649e-02 5.43342292e-01 -9.73113894e-01 1.43936083e-01
-6.09540701e-01 -2.50771254e-01 3.16779286e-01 3.09927732e-01
3.00815612e-01 -6.57118976e-01 -1.13161707e+00 -2.02653304e-01
1.64608344e-01 -1.10276556e+00 -3.87928873e-01 -6.02660589e-02
-5.84496439e-01 -8.11407506e-01 -8.62895906e-01 -4.70142484e-01
6.71707094e-01 -5.04590273e-01 1.19100213e+00 9.18934867e-02
-4.83678669e-01 5.23814619e-01 2.54833582e-03 -2.21036404e-01
-3.94906878e-01 -3.92271698e-01 -4.12650816e-02 1.41926765e-01
-2.98985213e-01 -9.67988610e-01 -8.33593369e-01 2.38483205e-01
-6.66195691e-01 4.76613849e-01 -1.02127135e-01 5.15735567e-01
6.64246976e-01 -8.78381610e-01 -3.56060833e-01 -4.01482791e-01
3.81230004e-02 2.72976041e-01 -9.79793608e-01 -7.59873465e-02
-1.33858234e-01 1.83366373e-01 2.80200869e-01 -5.12664497e-01
-1.19109476e+00 4.71656770e-01 -2.27194518e-01 -7.80651987e-01
-1.34270832e-01 -3.05502206e-01 -4.73518759e-01 -5.89084268e-01
7.31196523e-01 -1.84889421e-01 2.52990246e-01 -3.49687129e-01
6.27471209e-01 2.85671890e-01 6.82189465e-01 -9.28104341e-01
7.35719502e-01 7.70762146e-01 5.23020804e-01 -9.75818813e-01
-6.88085973e-01 1.31608382e-01 -9.81878877e-01 -6.09305203e-01
1.00873184e+00 -1.10001850e+00 -1.21138537e+00 5.31984508e-01
-1.31589019e+00 -9.65263903e-01 -5.69671571e-01 9.09571573e-02
-9.98025119e-01 -2.49805883e-01 -6.41812444e-01 -1.01192296e+00
-2.64322788e-01 -1.13246143e+00 1.95731246e+00 2.25603670e-01
-3.97393078e-01 -7.63014913e-01 1.41877770e-01 2.38282919e-01
2.43464053e-01 1.06154716e+00 5.51589489e-01 5.88608146e-01
-8.73056948e-01 -3.19257961e-03 6.24938868e-02 -2.95812543e-02
-1.65511176e-01 3.62885296e-01 -1.63468730e+00 4.26641703e-02
-3.55835170e-01 -4.80065912e-01 2.22695097e-01 3.97250861e-01
1.01183653e+00 -2.69253552e-01 -3.22129461e-03 1.46534061e+00
1.21318102e+00 -2.51178980e-01 7.11602747e-01 -9.47790891e-02
9.25338089e-01 8.42528343e-01 3.95749837e-01 7.31558025e-01
5.74840188e-01 1.14510071e+00 6.78874552e-01 -3.01676005e-01
-3.00644249e-01 -4.60067928e-01 2.44878963e-01 2.06884965e-01
-3.67137551e-01 3.95498335e-01 -5.49333811e-01 -1.95885040e-02
-1.47773635e+00 -4.79761332e-01 9.48749185e-02 2.17858195e+00
9.00948465e-01 -5.03570139e-01 3.36696267e-01 -1.68345004e-01
-4.38765101e-02 9.70510021e-02 -5.67117989e-01 -5.55717707e-01
-3.18404377e-01 5.15346646e-01 1.94534898e-01 8.86672020e-01
-6.47772968e-01 1.13239694e+00 6.26517916e+00 2.01674268e-01
-1.38416219e+00 2.04889439e-02 8.47072661e-01 -5.64918160e-01
-7.86032200e-01 -2.02524111e-01 -3.39588642e-01 -2.06270725e-01
7.26647735e-01 3.75411958e-01 5.32671154e-01 6.37945950e-01
3.24482739e-01 5.91792315e-02 -1.09556317e+00 1.21495914e+00
1.58334345e-01 -1.31532300e+00 -1.44620508e-01 1.91922128e-01
7.51015365e-01 -2.96668798e-01 8.72771740e-02 -8.97203237e-02
2.13543564e-01 -1.25535357e+00 1.26658916e+00 9.06065524e-01
1.33444309e+00 -8.40816021e-01 -1.93649784e-01 1.77865520e-01
-9.33482051e-01 4.57469016e-01 -1.08702831e-01 -6.28612638e-02
2.27072358e-01 1.67131752e-01 -5.60763478e-01 2.13067070e-01
8.26382995e-01 4.06638861e-01 -1.10669851e-01 4.56556499e-01
-2.65001476e-01 -5.27373925e-02 -5.25818348e-01 7.07518011e-02
-1.53927654e-01 -2.80341625e-01 3.28649104e-01 8.94295216e-01
2.36178100e-01 4.23558623e-01 -3.79114188e-02 1.10889411e+00
-1.36727735e-01 1.95560843e-01 -4.42819268e-01 3.81803602e-01
-1.59549177e-01 1.39148879e+00 -1.80552065e-01 8.01524520e-02
5.29988520e-02 1.46383667e+00 3.61109167e-01 5.38468003e-01
-7.31082082e-01 1.66160062e-01 1.31732702e+00 5.46549916e-01
7.96087757e-02 -4.72027153e-01 -1.74003735e-01 -1.19782543e+00
-2.68408544e-02 -5.07056475e-01 -1.53263256e-01 -1.55667841e+00
-7.21660018e-01 8.27724576e-01 -6.63155094e-02 -8.93720627e-01
-5.33457518e-01 -6.70258105e-01 -3.11760396e-01 1.07859504e+00
-1.37224352e+00 -1.77894604e+00 -7.19730020e-01 7.01038539e-01
2.54450478e-02 3.61175776e-01 1.27222741e+00 -5.37948869e-02
3.10304016e-02 5.57525218e-01 -5.60053229e-01 -6.28870577e-02
5.45293987e-01 -1.10737562e+00 3.53482157e-01 2.95833886e-01
-2.45970991e-02 5.73026501e-02 8.46168697e-01 -2.79257834e-01
-1.77745450e+00 -7.96274841e-01 4.35571134e-01 -8.08780611e-01
1.10640444e-01 -8.39625895e-01 -5.84799111e-01 6.92270219e-01
-8.18325877e-02 5.71386814e-01 3.46120805e-01 -2.08994344e-01
-6.41453564e-01 -3.27107549e-01 -1.60255706e+00 6.80867493e-01
1.20178235e+00 -6.69482529e-01 1.13723809e-02 1.94323391e-01
6.31640971e-01 -1.13041425e+00 -1.02964008e+00 2.29848817e-01
1.01531708e+00 -1.19120908e+00 9.40807581e-01 -5.00117779e-01
3.30998391e-01 -3.54105979e-01 -3.86215419e-01 -1.33327258e+00
1.41027570e-01 -1.22297370e+00 -4.58505712e-02 9.45772648e-01
-1.47651732e-02 -5.44835217e-02 1.24102938e+00 1.14229059e+00
3.42979208e-02 -7.89037824e-01 -1.08270311e+00 -2.84408361e-01
1.21863171e-01 -5.48233926e-01 8.44755530e-01 6.27778411e-01
-2.03651458e-01 1.06820896e-01 -4.37282532e-01 2.98949987e-01
7.29007304e-01 2.60897428e-01 1.20396900e+00 -1.20815289e+00
-2.23711193e-01 -1.14195690e-01 -6.71985567e-01 -1.15210712e+00
3.61557901e-01 -5.67127407e-01 -1.86063964e-02 -1.22078204e+00
-3.00022364e-01 -3.86581212e-01 6.19765997e-01 3.55090797e-01
3.74950081e-01 5.97834051e-01 1.75166994e-01 -2.00323030e-01
-1.32253334e-01 8.24773908e-01 1.63927329e+00 3.40960354e-01
-2.70585418e-01 -2.69774497e-01 -3.17063898e-01 6.91886246e-01
3.12801510e-01 -2.44992245e-02 -2.16056764e-01 -4.61368024e-01
3.17283496e-02 3.37395102e-01 6.04687691e-01 -8.49747002e-01
-1.36534572e-01 -3.00675422e-01 8.73781323e-01 -2.87629932e-01
1.31769145e+00 -7.37764537e-01 5.31558335e-01 2.83011273e-02
-2.76092321e-01 -8.46386775e-02 2.33781993e-01 -6.72851875e-02
2.91729361e-01 6.03192210e-01 1.06188774e+00 -1.65616319e-01
-3.14995825e-01 6.97869658e-01 2.07746804e-01 -4.13245112e-02
7.50034094e-01 -1.90531105e-01 2.64490336e-01 -9.09041584e-01
-7.02456415e-01 -1.81838542e-01 1.18095529e+00 2.36773700e-01
7.23521590e-01 -1.47585762e+00 -6.29665971e-01 7.03020930e-01
2.24199340e-01 5.03240526e-01 1.90625921e-01 3.50855082e-01
-9.86999094e-01 -2.47311249e-01 -9.71707553e-02 -8.91051352e-01
-1.34056616e+00 7.13323727e-02 1.04438984e+00 4.04713452e-01
-7.36859381e-01 1.12276733e+00 3.72683018e-01 -7.91175604e-01
2.62833476e-01 -3.58094364e-01 4.38308507e-01 -2.90501416e-01
4.99378353e-01 -5.41948304e-02 2.48821210e-02 -1.13018739e+00
-3.05440754e-01 1.19439089e+00 8.49868834e-01 -5.12249827e-01
1.48975611e+00 -2.14390885e-02 -5.90076707e-02 4.24816877e-01
1.42129815e+00 2.36589655e-01 -2.18115115e+00 1.10709980e-01
-6.42635703e-01 -7.19665945e-01 8.20864588e-02 -8.62628579e-01
-1.38240337e+00 8.66200626e-01 5.32221377e-01 -6.09724164e-01
1.21284258e+00 9.84447673e-02 5.33793271e-01 -1.07454002e-01
6.15433633e-01 -7.65152454e-01 2.77824551e-02 2.44601920e-01
1.18927097e+00 -9.07626510e-01 -1.04496852e-02 -5.59504688e-01
-5.55383742e-01 1.22009122e+00 5.05168319e-01 -1.16295986e-01
7.38825977e-01 6.31795585e-01 5.93316972e-01 -2.67553836e-01
-5.07665813e-01 1.14876300e-01 5.11700690e-01 7.64129519e-01
4.03040081e-01 7.74922147e-02 6.60495520e-01 1.36713162e-01
-6.61769867e-01 -4.24525477e-02 1.91731453e-01 7.34589159e-01
5.79717616e-03 -9.46501315e-01 -4.32513207e-01 -2.77179271e-01
-1.52155310e-01 3.28158528e-01 -4.16544408e-01 8.49566221e-01
1.68708190e-01 3.67583394e-01 3.62510651e-01 -2.74832487e-01
6.84106827e-01 6.79548457e-02 1.17898726e+00 -3.25219631e-01
-5.86887121e-01 3.34419072e-01 6.30826652e-02 -1.07406914e+00
-3.69263887e-01 -6.77230537e-01 -1.40991759e+00 -4.53460962e-01
2.69853920e-01 -3.66873711e-01 9.63369310e-01 5.47137797e-01
4.52079952e-01 1.64491311e-01 6.29717946e-01 -1.58139586e+00
-4.59584780e-03 -5.90046644e-01 -6.93892241e-01 6.35327041e-01
5.31926572e-01 -5.88225126e-01 -3.19014341e-01 8.76142010e-02] | [12.767753601074219, -0.37280645966529846] |
69261a8a-e162-4ea0-b141-beccceac4d15 | camouflaged-chinese-spam-content-detection | null | null | https://aclanthology.org/2020.acl-main.279 | https://aclanthology.org/2020.acl-main.279.pdf | Camouflaged Chinese Spam Content Detection with Semi-supervised Generative Active Learning | We propose a Semi-supervIsed GeNerative Active Learning (SIGNAL) model to address the imbalance, efficiency, and text camouflage problems of Chinese text spam detection task. A {``}self-diversity{''} criterion is proposed for measuring the {``}worthiness{''} of a candidate for annotation. A semi-supervised variational autoencoder with masked attention learning approach and a character variation graph-enhanced augmentation procedure are proposed for data augmentation. The preliminary experiment demonstrates the proposed SIGNAL model is not only sensitive to spam sample selection, but also can improve the performance of a series of conventional active learning models for Chinese spam detection task. To the best of our knowledge, this is the first work to integrate active learning and semi-supervised generative learning for text spam detection. | ['Zhuoren Jiang', 'Zhe Gao', 'Yu Duan', 'Yangyang Kang', 'Xiaozhong Liu', 'Qiong Zhang', 'Changlong Sun'] | 2020-07-01 | null | null | null | acl-2020-6 | ['spam-detection'] | ['natural-language-processing'] | [ 3.01487535e-01 -9.25027765e-03 -1.15586765e-01 -3.95066351e-01
-1.02478743e+00 -3.20666105e-01 7.24802911e-01 1.06345780e-01
-4.12043720e-01 6.31843984e-01 1.81332082e-01 -4.99207526e-01
8.57800469e-02 -5.84020495e-01 -3.57261807e-01 -9.84990239e-01
2.81137615e-01 6.56266034e-01 3.53867412e-01 -2.70762652e-01
3.13980609e-01 2.20643878e-01 -1.14721704e+00 4.15099442e-01
1.47022653e+00 7.77668893e-01 1.83603197e-01 6.22690618e-01
-4.13480312e-01 1.05868018e+00 -1.00149691e+00 -6.62969589e-01
-1.24757320e-01 -5.35774887e-01 -7.11585224e-01 4.34852868e-01
2.62875110e-01 -3.62511724e-01 -2.69864261e-01 1.32367432e+00
5.09126067e-01 1.78036660e-01 8.49808395e-01 -1.20602202e+00
-1.15517962e+00 7.77273297e-01 -3.67763698e-01 2.45623410e-01
-2.19115511e-01 2.36864865e-01 1.12016678e+00 -8.53485465e-01
3.88472021e-01 1.13782191e+00 5.25288582e-01 7.04630494e-01
-1.13206792e+00 -5.17596304e-01 4.34802294e-01 4.46614623e-02
-7.81879246e-01 -2.65244216e-01 1.00806427e+00 -3.42998236e-01
6.30299211e-01 3.47945273e-01 3.64375740e-01 1.51502609e+00
-2.57758647e-01 1.48605609e+00 7.81523705e-01 -4.26756978e-01
4.84979570e-01 5.82720697e-01 5.94415307e-01 9.25692499e-01
3.71288151e-01 -1.19548611e-01 -5.29704273e-01 -6.84097826e-01
3.01054090e-01 5.44091910e-02 -4.68277745e-02 -3.90457779e-01
-5.30598104e-01 1.34352565e+00 2.81071752e-01 2.05830663e-01
-4.11058068e-02 6.08634651e-02 4.52804387e-01 2.82127321e-01
1.14417112e+00 5.25058806e-01 -2.51128584e-01 1.30609632e-01
-1.15956020e+00 -1.89668983e-01 7.53147364e-01 8.98146749e-01
4.47594672e-01 6.31835639e-01 -4.25940275e-01 8.63705397e-01
7.54033089e-01 4.78471249e-01 5.53587615e-01 -4.55773145e-01
3.94599915e-01 9.83343780e-01 1.36075884e-01 -5.60769796e-01
-5.08717541e-03 -6.15123391e-01 -5.45781732e-01 1.75330937e-01
2.20994741e-01 -2.08685011e-01 -8.98185849e-01 1.23558187e+00
-2.54154787e-03 1.07677840e-01 -2.80465484e-01 7.50886083e-01
1.04583406e+00 5.38970172e-01 3.27967793e-01 -4.49361861e-01
8.78684521e-01 -9.13076699e-01 -8.12923670e-01 -2.52971560e-01
7.01548874e-01 -7.46288300e-01 1.21350718e+00 1.16099097e-01
-8.25727105e-01 -3.18441689e-01 -8.80640030e-01 1.90471448e-02
-2.69285023e-01 4.31623966e-01 7.20570147e-01 1.22052193e+00
-6.12860978e-01 2.41159409e-01 -8.77199173e-01 2.54898006e-03
9.31086719e-01 3.10602158e-01 2.00914353e-01 5.30906677e-01
-1.13137543e+00 5.61776817e-01 1.88113734e-01 -2.89215241e-02
-9.41179037e-01 -4.23198313e-01 -4.13549155e-01 3.13120961e-01
3.79870504e-01 -1.91349849e-01 1.01886988e+00 -1.38659477e+00
-1.19583428e+00 8.61147106e-01 -3.70969437e-02 -7.46175289e-01
5.52225053e-01 -3.36221427e-01 -4.21625078e-01 3.23480368e-02
-1.34862363e-01 1.81567773e-01 1.49325502e+00 -1.65930736e+00
-1.72586486e-01 -4.54815328e-01 -5.51095188e-01 6.27936646e-02
-8.07010591e-01 2.52313107e-01 -1.74289539e-01 -1.12918389e+00
1.55010864e-01 -7.87900031e-01 -1.05416767e-01 3.99451219e-02
-3.88667673e-01 -3.48515362e-01 1.44959855e+00 -8.28767657e-01
1.38965857e+00 -2.10173535e+00 -3.81009318e-02 9.16068405e-02
3.95274132e-01 8.49034071e-01 -5.56906164e-02 3.66802275e-01
2.98671454e-01 1.55554876e-01 -7.13053226e-01 -5.40760040e-01
1.02984607e-01 -1.54617513e-02 -6.92068577e-01 6.61114633e-01
5.43135464e-01 1.07786179e+00 -9.48048890e-01 -5.26389122e-01
1.17213093e-01 4.55557883e-01 -3.10640544e-01 2.60329574e-01
-5.96181810e-01 2.82348484e-01 -4.98050183e-01 6.64947271e-01
5.96269488e-01 -4.77457762e-01 7.93543234e-02 3.07008296e-01
2.55995572e-01 2.14568675e-01 -8.65807116e-01 1.07314718e+00
7.06295967e-02 7.34446049e-01 6.39854446e-02 -1.23153102e+00
1.19016778e+00 2.03058496e-01 3.35458159e-01 -3.21675509e-01
4.29603010e-01 2.16526926e-01 2.82153562e-02 -4.05050009e-01
6.88944221e-01 1.87112123e-01 -7.30307102e-02 7.56774366e-01
1.58636436e-01 1.76715463e-01 1.57717809e-01 6.87478721e-01
9.81796622e-01 1.74894318e-01 -3.00583273e-01 -3.63334745e-01
6.89110577e-01 -1.18109854e-02 3.93203437e-01 1.14399946e+00
-5.55297196e-01 3.99338990e-01 7.53753483e-01 7.61005282e-02
-1.12138486e+00 -9.87514377e-01 -5.85657451e-03 1.46964407e+00
7.71740004e-02 -1.12691246e-01 -8.62690687e-01 -1.34161794e+00
5.27230017e-02 1.02950382e+00 -6.07141614e-01 -3.72325838e-01
-5.60958862e-01 -1.52366579e+00 5.50589800e-01 5.72140992e-01
4.70588416e-01 -1.08326519e+00 4.15587500e-02 3.63228992e-02
-8.12576935e-02 -5.91911376e-01 -4.09691989e-01 2.83547908e-01
-1.02534723e+00 -9.22415376e-01 -5.65005004e-01 -1.08738852e+00
1.01462817e+00 2.55226284e-01 9.34024572e-01 4.20380950e-01
-2.42898524e-01 2.28618056e-01 -8.85208309e-01 -4.66400415e-01
-9.44395363e-01 6.72122315e-02 -3.32434326e-01 2.13404968e-01
4.29971874e-01 6.92063048e-02 -4.67435211e-01 2.80934840e-01
-8.88032615e-01 -4.32575613e-01 5.76592028e-01 1.35259807e+00
1.95244178e-01 -5.47915921e-02 8.64027262e-01 -1.39872050e+00
1.05146933e+00 -3.11812788e-01 -5.76865733e-01 3.51555198e-01
-7.59281516e-01 2.19765324e-02 5.86756349e-01 -6.65319383e-01
-1.58617473e+00 7.66417682e-02 -1.84796140e-01 3.03059146e-02
1.85654834e-02 -8.48303884e-02 -2.25847408e-01 1.22019805e-01
9.85671639e-01 3.68694365e-01 -8.94658919e-03 -2.89248019e-01
3.57229054e-01 1.10343742e+00 8.71070921e-02 -3.16775590e-02
8.69164348e-01 5.33351362e-01 -5.55301785e-01 -7.88974404e-01
-7.61481047e-01 -6.59375727e-01 -6.95462704e-01 3.00729685e-02
4.14025873e-01 -7.42000461e-01 -5.05519748e-01 6.41136527e-01
-7.21190155e-01 -2.24751770e-01 -3.39419484e-01 1.97931886e-01
-4.14536566e-01 1.00191736e+00 -7.16587484e-01 -1.33925045e+00
-9.40504372e-01 -1.02494872e+00 1.10963476e+00 -1.90764274e-02
-3.20147164e-02 -1.08142340e+00 -2.06930652e-01 7.17683852e-01
3.24176878e-01 -4.72239733e-01 6.90104127e-01 -1.19479656e+00
-5.54017067e-01 -6.07460141e-01 -2.55006582e-01 6.03097439e-01
-1.15136959e-01 -4.76630256e-02 -1.03059745e+00 -5.90890944e-01
1.77051291e-01 -3.49452436e-01 1.41128886e+00 2.30526209e-01
1.03332877e+00 -5.38563371e-01 -2.08551601e-01 1.81798741e-01
1.12391770e+00 2.44259328e-01 6.56665683e-01 1.16198743e-02
5.69414437e-01 2.22922713e-01 4.08205479e-01 3.51164669e-01
-4.25153553e-01 2.03182027e-01 3.58429044e-01 -1.97780237e-01
-1.35400176e-01 -2.34572709e-01 4.65152234e-01 7.20533729e-01
5.29033899e-01 -7.86108613e-01 -7.13032901e-01 3.31085473e-01
-1.91755879e+00 -1.13139379e+00 -4.63105619e-01 2.03769875e+00
6.52188122e-01 2.90803283e-01 7.75009766e-02 2.58843660e-01
9.56521749e-01 1.71814278e-01 -5.15583694e-01 -1.98286310e-01
-4.32707429e-01 1.46156311e-01 3.67661804e-01 5.52327037e-01
-1.59005415e+00 1.31177497e+00 5.51691437e+00 1.09846210e+00
-8.15116048e-01 4.09870893e-01 6.45617068e-01 1.95005223e-01
-2.82248378e-01 1.64244175e-01 -7.70538270e-01 9.47213471e-01
4.79417205e-01 2.68496066e-01 -7.94757903e-02 9.19143558e-01
1.05060272e-01 -1.22965261e-01 -3.09113741e-01 4.90631431e-01
6.34808838e-01 -1.18109465e+00 1.22967616e-01 -1.01902358e-01
8.76129627e-01 -1.16375461e-01 3.04649055e-01 2.27907375e-01
5.48230410e-01 -6.01600528e-01 4.79452163e-01 4.71444398e-01
2.26810545e-01 -5.87957382e-01 8.28092575e-01 6.24587476e-01
-4.65237468e-01 -4.58536029e-01 -2.65177786e-01 5.41847050e-01
2.71901309e-01 4.76170331e-01 -9.79999006e-01 4.18206081e-02
3.72919589e-01 4.00899351e-01 -9.59069610e-01 1.04999971e+00
-3.90106827e-01 1.35783470e+00 1.16989076e-01 -7.76497066e-01
6.49313703e-02 -2.08866403e-01 8.93254757e-01 1.20815086e+00
-8.88673291e-02 -3.42061043e-01 1.37552889e-02 1.08238900e+00
-2.17496559e-01 2.34968781e-01 -9.00910795e-02 -5.70636511e-01
3.96558881e-01 1.02252758e+00 -1.10273910e+00 -4.61840391e-01
-2.40769520e-01 1.33892727e+00 2.27266282e-01 4.05547291e-01
-7.05801010e-01 -3.86409938e-01 -1.48786427e-02 5.69168553e-02
2.21029162e-01 -1.57854840e-01 -6.94502175e-01 -1.33756769e+00
-3.41519982e-01 -6.41890824e-01 3.63435298e-01 -6.48372591e-01
-1.64318740e+00 3.58466923e-01 -5.19518137e-01 -1.19902742e+00
7.09473565e-02 -5.91924071e-01 -7.40242541e-01 7.19418049e-01
-9.38960910e-01 -1.30436420e+00 -4.89512198e-02 2.05673084e-01
8.58929098e-01 -9.12658870e-01 6.63306892e-01 2.01276485e-02
-6.95937574e-01 8.06214631e-01 7.27454603e-01 4.38329428e-01
5.24344444e-01 -1.22793519e+00 6.29977167e-01 1.10385609e+00
2.69484252e-01 7.62493610e-01 7.46954322e-01 -1.19662952e+00
-1.14123297e+00 -1.07009912e+00 8.09384465e-01 -7.47674704e-01
6.26887262e-01 -6.74474299e-01 -1.14489877e+00 4.89507705e-01
1.62779763e-02 -5.32829165e-01 7.90436327e-01 7.90496394e-02
-2.40767702e-01 1.98029831e-01 -1.08599901e+00 3.62633586e-01
6.78923130e-01 -4.45186108e-01 -4.75045860e-01 6.29565060e-01
5.15537798e-01 4.11930121e-02 -2.55273551e-01 2.38313079e-01
9.65341628e-02 -6.85010612e-01 9.61718321e-01 -8.34852993e-01
1.30479813e-01 1.22950852e-01 3.57232600e-01 -9.88560438e-01
-5.67130089e-01 -5.29519975e-01 -3.97065341e-01 1.23634374e+00
6.55337036e-01 -3.77648830e-01 1.29511237e+00 1.10360615e-01
-2.58187950e-01 -3.75364035e-01 -7.19805002e-01 -3.87632936e-01
2.57124782e-01 -2.86395550e-01 1.32488817e-01 9.40225124e-01
-2.80559920e-02 3.99800181e-01 -6.71299160e-01 -1.54545113e-01
8.03185523e-01 -1.08115852e-01 4.31312621e-01 -1.46474564e+00
-2.14632392e-01 -5.03673971e-01 -1.05191067e-01 -1.07081354e+00
3.39069068e-01 -1.03761029e+00 -4.07705382e-02 -1.41749728e+00
3.76858413e-01 -2.66870290e-01 5.28056659e-02 1.51111171e-01
-6.56788290e-01 1.01403534e-01 -2.39809807e-02 3.46007347e-01
-7.21317351e-01 7.79299915e-01 1.00275385e+00 -3.73105943e-01
-3.64903808e-01 5.37068605e-01 -7.68167734e-01 4.83802646e-01
1.04113400e+00 -3.40715468e-01 -3.78890812e-01 -2.49914482e-01
2.88517803e-01 -3.45902890e-01 4.85765398e-01 -2.50402987e-01
2.15644106e-01 1.72723413e-01 4.83564079e-01 -6.45935655e-01
2.86125362e-01 -4.34638411e-01 -6.14761710e-01 5.64346850e-01
-6.76698685e-01 -4.31983769e-01 -1.89177260e-01 8.66125762e-01
1.84785888e-01 -5.53873122e-01 9.88807023e-01 -2.02042997e-01
-5.16728222e-01 1.13664418e-01 -7.66136050e-01 1.21382184e-01
5.63573182e-01 -6.12214394e-02 -8.16033661e-01 -3.85322511e-01
-6.70645058e-01 1.28133819e-01 1.77837372e-01 5.33649564e-01
6.92498684e-01 -9.10995781e-01 -8.86266649e-01 5.05968809e-01
1.00019559e-01 -5.56817591e-01 9.25736651e-02 6.71042621e-01
-5.35215795e-01 1.50125191e-01 1.96379587e-01 -4.13280487e-01
-1.41565895e+00 3.82923871e-01 1.53470591e-01 -1.35636494e-01
-5.04403651e-01 1.15394306e+00 -1.61437064e-01 -4.11507159e-01
5.24837077e-01 5.90011060e-01 -2.78171748e-01 9.35524926e-02
1.23252831e-01 7.17405736e-01 1.41795263e-01 -7.56917715e-01
1.06766030e-01 -4.02149349e-01 -3.05596232e-01 -4.19474542e-02
1.01335084e+00 -3.60740423e-02 -1.41696841e-01 1.68698341e-01
9.64778483e-01 -1.23756528e-01 -9.61041570e-01 -4.16664362e-01
1.10477410e-01 -4.58760202e-01 1.72016263e-01 -9.75076556e-01
-7.63764739e-01 8.59287858e-01 8.39301527e-01 4.17081654e-01
7.12804317e-01 8.56972300e-03 7.34746039e-01 3.99980992e-01
-1.17382154e-01 -1.58890915e+00 5.91406822e-01 4.60355073e-01
5.81317425e-01 -1.48306000e+00 8.64770934e-02 -5.08553088e-01
-9.71651971e-01 7.09957838e-01 6.89120829e-01 -2.07099512e-01
5.15445590e-01 2.39205584e-01 7.44836852e-02 -1.56832948e-01
-6.41792476e-01 -3.73688251e-01 3.35479349e-01 4.72999722e-01
5.63005686e-01 -5.76432906e-02 -6.00696921e-01 3.69421154e-01
1.85183629e-01 -4.72731262e-01 2.05898196e-01 9.17430758e-01
-1.00325441e+00 -9.65991378e-01 -2.92265445e-01 8.45922410e-01
-2.85268515e-01 -1.15979403e-01 -1.06956422e+00 2.56817788e-01
9.04796645e-03 1.14896488e+00 2.23681584e-01 -1.44935817e-01
-1.14806421e-01 2.85380691e-01 1.88291788e-01 -6.01156175e-01
-9.72719073e-01 4.79309171e-01 -7.21120462e-02 3.71933162e-01
-1.85526088e-01 -6.31061614e-01 -1.06464362e+00 -1.23376518e-01
-1.19629920e+00 3.45465630e-01 3.20073307e-01 6.40466273e-01
1.86816573e-01 1.72316179e-01 6.66832328e-01 7.46173188e-02
-9.51630414e-01 -1.18111014e+00 -6.20339155e-01 7.60820270e-01
-4.78836745e-02 -3.11986804e-01 -8.08757901e-01 -4.95976210e-03] | [7.881422519683838, 9.956777572631836] |
e9274e45-37bb-4aee-98d1-3f00ece0b57e | efficient-fine-grained-road-segmentation | 2207.02844 | null | https://arxiv.org/abs/2207.02844v1 | https://arxiv.org/pdf/2207.02844v1.pdf | Efficient fine-grained road segmentation using superpixel-based CNN and CRF models | Towards a safe and comfortable driving, road scene segmentation is a rudimentary problem in camera-based advance driver assistance systems (ADAS). Despite of the great achievement of Convolutional Neural Networks (CNN) for semantic segmentation task, the high computational efforts of CNN based methods is still a challenging area. In recent work, we proposed a novel approach to utilise the advantages of CNNs for the task of road segmentation at reasonable computational effort. The runtime benefits from using irregular super pixels as basis for the input for the CNN rather than the image grid, which tremendously reduces the input size. Although, this method achieved remarkable low computational time in both training and testing phases, the lower resolution of the super pixel domain yields naturally lower accuracy compared to high cost state of the art methods. In this work, we focus on a refinement of the road segmentation utilising a Conditional Random Field (CRF).The refinement procedure is limited to the super pixels touching the predicted road boundary to keep the additional computational effort low. Reducing the input to the super pixel domain allows the CNNs structure to stay small and efficient to compute while keeping the advantage of convolutional layers and makes them eligible for ADAS. Applying CRF compensate the trade off between accuracy and computational efficiency. The proposed system obtained comparable performance among the top performing algorithms on the KITTI road benchmark and its fast inference makes it particularly suitable for realtime applications. | ['Josef Pauli', 'Mirko Meuter', 'Jan Siegemund', 'Farnoush Zohourian'] | 2022-06-22 | null | null | null | null | ['scene-segmentation', 'road-segementation'] | ['computer-vision', 'computer-vision'] | [ 3.74224156e-01 4.27406192e-01 3.48819122e-02 -5.51815391e-01
-3.74171466e-01 -9.42811444e-02 7.28485763e-01 -9.16922912e-02
-9.75842774e-01 7.30884075e-01 -5.12237966e-01 -5.63010275e-01
-1.99262220e-02 -1.14117599e+00 -5.05614996e-01 -6.09885812e-01
5.09825468e-01 5.43346465e-01 8.77672315e-01 -2.38749623e-01
2.74294138e-01 8.50089908e-01 -2.17780304e+00 -1.04263999e-01
1.07029188e+00 1.07003987e+00 5.30630052e-01 4.96285290e-01
-4.15568411e-01 5.45781314e-01 -2.19802439e-01 -1.96919620e-01
4.57167923e-01 -5.23645766e-02 -7.46820807e-01 1.32941324e-02
5.25467873e-01 -1.75860792e-01 2.42080726e-02 9.24692452e-01
3.17905009e-01 2.65446931e-01 4.64545697e-01 -9.66869056e-01
3.28714937e-01 2.03866269e-02 -4.23888266e-01 5.55698983e-02
-2.41604328e-01 6.92442879e-02 4.48661149e-01 -6.07987821e-01
5.67505598e-01 7.11010218e-01 7.92869687e-01 3.57018471e-01
-8.30979407e-01 -4.35743719e-01 -2.84366943e-02 3.31504881e-01
-1.47922671e+00 -3.84735197e-01 5.63372791e-01 -4.00798470e-01
9.20652986e-01 1.06001645e-01 6.89371347e-01 3.37211251e-01
7.65111595e-02 4.44760025e-01 1.40108848e+00 -3.82125407e-01
3.61980468e-01 4.31444377e-01 3.12175781e-01 6.91706836e-01
7.43595660e-02 1.12629063e-01 -1.76036000e-01 4.21304524e-01
7.04264462e-01 -1.45087600e-01 2.12387577e-01 -1.98693082e-01
-6.89053655e-01 8.09456527e-01 4.96076792e-01 2.82560289e-01
-5.44322908e-01 1.62414964e-02 2.69587904e-01 -5.73383458e-02
3.78540248e-01 -1.54637862e-02 -4.41703200e-01 -2.37229541e-01
-1.28582990e+00 2.91884184e-01 5.46335757e-01 8.19690108e-01
1.16333187e+00 -6.64724316e-03 6.78355917e-02 6.76677644e-01
8.63216668e-02 3.36939961e-01 7.94385299e-02 -7.66314328e-01
3.54394019e-01 8.21337759e-01 1.98829398e-02 -8.04026425e-01
-5.45206010e-01 -5.06048501e-01 -7.43721783e-01 7.31100321e-01
7.90344238e-01 -2.63189785e-02 -1.37977254e+00 1.04263043e+00
4.95831341e-01 6.22240640e-02 1.07716664e-03 7.95212924e-01
6.05746925e-01 5.17567813e-01 2.64120221e-01 8.42414945e-02
1.39410257e+00 -1.04503918e+00 -5.71576178e-01 -3.46057147e-01
8.18820417e-01 -7.91712284e-01 7.28055656e-01 3.07474107e-01
-8.98591101e-01 -8.96738291e-01 -9.87894475e-01 -2.62796700e-01
-7.53261626e-01 3.93336177e-01 5.96080363e-01 8.80722880e-01
-1.22222745e+00 4.64627624e-01 -6.28892899e-01 -4.18531299e-01
6.42272115e-01 5.01981258e-01 -3.50976765e-01 -9.68372375e-02
-9.65288043e-01 1.21741796e+00 5.22610009e-01 4.80858654e-01
-2.71768779e-01 -5.14841080e-01 -8.30225527e-01 -6.47807568e-02
5.16466975e-01 -4.43774581e-01 9.31088448e-01 -8.62179220e-01
-1.74031317e+00 1.02792430e+00 -2.38970712e-01 -9.31675732e-01
8.96096230e-01 -2.41658643e-01 -8.80181193e-02 2.26696134e-02
-5.79698198e-02 1.04334283e+00 8.92919540e-01 -1.03088117e+00
-1.02908862e+00 -9.73079130e-02 7.99744797e-04 1.96033686e-01
1.62927449e-01 -1.91488281e-01 -5.23904204e-01 -1.90759934e-02
1.56478092e-01 -1.02542186e+00 -5.96356452e-01 -2.48586178e-01
-1.02071591e-01 -3.86721581e-01 1.15560186e+00 -4.94006425e-01
8.81149411e-01 -2.04464078e+00 -3.60443175e-01 3.01318437e-01
-5.41307814e-02 7.59038687e-01 4.37835485e-01 2.47329418e-02
1.55067414e-01 -3.13468039e-01 -6.62898898e-01 -3.23606312e-01
-3.48323494e-01 4.23440725e-01 1.45763475e-02 4.10539359e-01
3.89977336e-01 8.82565379e-01 -4.46207583e-01 -7.37208068e-01
8.32245529e-01 6.79089487e-01 -2.58869171e-01 1.57916918e-01
-3.15515175e-02 5.08308053e-01 -4.03975904e-01 2.98091948e-01
1.05838120e+00 3.49485457e-01 -3.96235734e-01 -1.57844067e-01
-7.97693849e-01 1.56988323e-01 -1.31969607e+00 1.42951167e+00
-6.25540376e-01 7.30441272e-01 8.46968591e-02 -1.22705221e+00
1.31693792e+00 1.29812341e-02 1.98516935e-01 -9.56238985e-01
2.29013830e-01 3.05663943e-01 -5.04032671e-02 -3.50887984e-01
8.71655405e-01 2.78455783e-02 2.53250331e-01 -1.24391347e-01
-1.96555510e-01 -4.86026615e-01 7.43351579e-02 -1.98354185e-01
4.82313156e-01 6.11094296e-01 2.31422260e-01 -5.72540641e-01
7.92687237e-01 4.36127454e-01 3.57591093e-01 6.09943449e-01
-1.83501795e-01 6.05021477e-01 2.22291142e-01 -7.49024451e-01
-1.15550399e+00 -5.81997275e-01 -4.56344336e-01 6.12364829e-01
2.32096732e-01 -1.54301515e-02 -1.13964474e+00 -4.38845813e-01
-2.98004597e-01 5.75707614e-01 -3.68890256e-01 3.17865759e-01
-7.76554406e-01 -6.87268555e-01 4.09736782e-01 4.86877650e-01
1.05396497e+00 -1.14642692e+00 -1.22870421e+00 2.39229724e-01
9.66277719e-02 -1.43971252e+00 1.75265551e-01 3.09318036e-01
-9.57824469e-01 -9.32288408e-01 -6.95698202e-01 -7.38499165e-01
5.59483647e-01 1.94954216e-01 8.45476151e-01 5.50878309e-02
-4.77658242e-01 -1.00296527e-01 -1.82539597e-01 -6.31419063e-01
-2.84827203e-01 3.67084175e-01 -4.75413293e-01 1.28661677e-01
4.68766749e-01 -2.79950112e-01 -6.34939253e-01 3.06185156e-01
-7.38784432e-01 4.77817595e-01 7.05956757e-01 5.25842845e-01
6.53426290e-01 3.48814398e-01 2.74242222e-01 -1.14725673e+00
1.32887540e-02 -1.02563947e-01 -1.01501155e+00 -1.79393202e-01
-8.30241859e-01 -8.06132704e-02 5.88259280e-01 2.15706199e-01
-1.32977235e+00 5.55260718e-01 -5.86345732e-01 1.13929257e-01
-7.34455705e-01 1.74416646e-01 8.00897777e-02 -2.06724241e-01
4.50059235e-01 4.47715074e-03 1.23516098e-01 -4.05191392e-01
2.20642716e-01 6.65262461e-01 6.05599165e-01 -1.10719636e-01
6.82289779e-01 6.99166358e-01 4.46667522e-01 -1.18158329e+00
-6.37903988e-01 -7.05129623e-01 -1.02865171e+00 -3.79632115e-01
1.13982511e+00 -7.45878935e-01 -5.30688345e-01 5.70322812e-01
-1.09781086e+00 -4.20963824e-01 -2.84079850e-01 2.92181700e-01
-5.34542263e-01 2.48468429e-01 -1.10789977e-01 -1.02187538e+00
-1.76606178e-01 -1.16430044e+00 9.31221008e-01 5.41673899e-01
5.40227853e-02 -7.75499642e-01 -4.05593544e-01 7.19528377e-01
6.31724119e-01 3.87780458e-01 6.02325857e-01 -1.75723240e-01
-8.05743575e-01 -4.14465755e-01 -6.36801183e-01 3.98327857e-01
-2.33003020e-01 3.58931050e-02 -1.36660671e+00 3.25839669e-01
-1.38443619e-01 4.76413332e-02 1.03851092e+00 4.85582799e-01
1.12293053e+00 3.63436848e-01 -2.55334407e-01 5.03736615e-01
1.57006621e+00 1.72811449e-01 1.15606034e+00 6.21218324e-01
7.77337074e-01 1.04018545e+00 9.25907910e-01 7.92738423e-02
4.03030068e-01 7.77066469e-01 4.96158451e-01 -6.16918981e-01
-3.12149316e-01 4.74427156e-02 -2.89123971e-02 3.63160670e-01
-4.33365881e-01 6.20900355e-02 -1.07867539e+00 7.88957000e-01
-1.86240339e+00 -7.33464062e-01 -7.84160674e-01 2.21651053e+00
5.19296706e-01 5.15397608e-01 5.41506521e-02 4.30341989e-01
5.42686701e-01 -1.19148865e-02 -6.96678311e-02 -8.56632411e-01
-6.74556792e-02 5.91955125e-01 1.09168184e+00 6.89519763e-01
-1.10341012e+00 1.11773825e+00 5.30368662e+00 1.10229027e+00
-1.11704206e+00 4.19241376e-02 6.81633294e-01 4.66622531e-01
3.10754597e-01 1.65116657e-02 -1.29412270e+00 3.27327073e-01
1.25939989e+00 5.13653398e-01 7.11444914e-02 9.59420264e-01
4.40787673e-01 -9.50904965e-01 -5.12722969e-01 6.97354972e-01
-2.91682869e-01 -1.29676151e+00 -3.52812469e-01 5.29549569e-02
6.38777316e-01 3.06730330e-01 -1.33671671e-01 -1.98889337e-02
-8.03919882e-02 -1.19178152e+00 6.32018685e-01 5.41663408e-01
7.61538684e-01 -9.95226324e-01 1.20903409e+00 5.28870225e-01
-1.18831408e+00 1.11099996e-01 -4.47247416e-01 -2.69861698e-01
2.64116555e-01 6.47437513e-01 -1.11107612e+00 5.87561607e-01
7.36796737e-01 2.77374506e-01 -4.96180773e-01 9.20415640e-01
6.38944954e-02 5.61075866e-01 -4.78960305e-01 1.44767568e-01
7.34868109e-01 -4.76507038e-01 2.03302443e-01 1.46739471e+00
8.37153941e-02 -1.66743949e-01 4.58068624e-02 7.10420549e-01
4.70020473e-01 9.62971225e-02 -6.83552504e-01 4.31755304e-01
-5.48947342e-02 1.40394747e+00 -1.15859747e+00 -3.82163793e-01
-3.93366188e-01 6.40739977e-01 2.59434372e-01 7.44100884e-02
-7.21924543e-01 -5.00629902e-01 3.45478505e-01 6.66444182e-01
4.24675971e-01 -4.66918766e-01 -8.85264635e-01 -3.36740404e-01
9.89987254e-02 -1.97157815e-01 -6.13800660e-02 -4.69528526e-01
-5.33528209e-01 8.45513344e-01 1.79204494e-01 -1.08595741e+00
-9.70731378e-02 -6.87320590e-01 -4.95856553e-01 1.24209368e+00
-2.16322732e+00 -1.25594640e+00 -6.88655972e-01 5.24896204e-01
6.29081130e-01 1.50357872e-01 7.28946149e-01 3.37909162e-01
-3.05264175e-01 2.75693953e-01 -2.01214314e-01 -5.77493533e-02
2.30651826e-01 -1.08394659e+00 4.01334047e-01 9.68257785e-01
-3.73406619e-01 2.18829870e-01 6.49338305e-01 -5.11975050e-01
-7.14440763e-01 -1.14132702e+00 1.14377213e+00 -2.30590269e-01
2.57330775e-01 -2.22584516e-01 -8.59432697e-01 6.46927357e-02
1.47438794e-01 2.92931087e-02 5.26801795e-02 -1.95872799e-01
1.40928537e-01 -2.55739003e-01 -1.12474096e+00 4.55525249e-01
7.86673844e-01 -4.61140245e-01 -3.57734174e-01 2.09587272e-02
2.89359689e-01 -4.77293938e-01 -5.17262638e-01 4.81367588e-01
3.74872983e-01 -1.31695676e+00 7.18116045e-01 1.58008441e-01
1.94122076e-01 -6.20455861e-01 3.39868575e-01 -8.10550690e-01
5.65258488e-02 -3.93887639e-01 4.33289766e-01 1.01521063e+00
4.95151252e-01 -7.10649371e-01 1.18158746e+00 7.59707212e-01
-4.86334622e-01 -7.46847689e-01 -1.07751536e+00 -5.15970588e-01
-1.24274321e-01 -5.56841850e-01 2.57787377e-01 4.25254494e-01
-6.50053918e-01 5.62428944e-02 -1.50043651e-01 1.47012174e-01
5.76074958e-01 -2.85734292e-02 8.50134969e-01 -1.48278928e+00
1.23691186e-01 -2.55963236e-01 -6.52748883e-01 -9.86057937e-01
5.89421317e-02 -6.04204535e-01 2.63895959e-01 -1.57540250e+00
-3.25979590e-01 -9.57222879e-01 1.18707746e-01 3.37010205e-01
-1.01058781e-02 6.39911592e-01 -9.41105708e-02 4.80935210e-03
-1.96781129e-01 9.37593803e-02 1.14886713e+00 1.78373128e-01
-3.35622966e-01 3.00008535e-01 -1.25308216e-01 8.03964615e-01
9.61201310e-01 -3.83361995e-01 -4.22138363e-01 -1.91672772e-01
2.27692653e-03 -1.74412027e-01 5.05804241e-01 -1.16472554e+00
5.73127031e-01 2.26204805e-02 8.02548826e-02 -1.03425086e+00
3.52638364e-01 -1.05819726e+00 1.00822128e-01 5.03949225e-01
1.40883043e-01 -2.56142408e-01 3.98157537e-01 3.31241339e-01
-3.95966172e-01 -4.17672962e-01 1.22098732e+00 -9.82819274e-02
-1.08576429e+00 2.00951114e-01 -5.22202134e-01 -4.27951574e-01
1.19429719e+00 -1.01161408e+00 1.98381588e-01 5.13943434e-02
-5.50826490e-01 -3.47941108e-02 3.59888643e-01 2.17103034e-01
3.98982823e-01 -6.15254045e-01 -5.30463815e-01 3.30242336e-01
-1.17252758e-02 6.50049746e-01 4.06707972e-01 9.61479604e-01
-9.83863533e-01 7.03174114e-01 -4.63311046e-01 -8.57688546e-01
-1.23707104e+00 2.22456366e-01 3.52782995e-01 -2.25645885e-01
-9.71290410e-01 6.85816646e-01 3.95872369e-02 -1.80593088e-01
4.11632173e-02 -3.47345233e-01 -6.52382016e-01 6.60016611e-02
2.65198559e-01 5.53120613e-01 4.87452954e-01 -8.78645897e-01
-2.99758315e-01 7.45823801e-01 -5.89264221e-02 8.89219567e-02
1.25929117e+00 -2.60129988e-01 -6.55515958e-03 1.17657229e-01
9.55195248e-01 -1.66140541e-01 -1.52435887e+00 8.45581479e-03
2.09832832e-01 -4.16002005e-01 5.78827441e-01 -5.60359418e-01
-9.82297182e-01 9.46743667e-01 6.62120998e-01 1.41977325e-01
9.88598347e-01 -3.48176956e-01 8.47727656e-01 1.75505042e-01
4.03004020e-01 -1.43852592e+00 -8.26947510e-01 5.66725254e-01
2.45380491e-01 -1.36255252e+00 1.27919938e-03 -8.91042888e-01
-5.61326981e-01 1.29590642e+00 4.80141252e-01 -2.96643525e-01
6.75340652e-01 4.08195257e-01 1.58267662e-01 -2.09404245e-01
-5.68755306e-02 -6.67094707e-01 2.56677210e-01 6.79334641e-01
2.15805754e-01 1.05702117e-01 -5.46378851e-01 5.17349094e-02
-2.67154694e-01 1.68408960e-01 4.22726095e-01 8.03824484e-01
-5.92666268e-01 -1.01763868e+00 -1.82205528e-01 4.18466091e-01
-3.57543379e-01 -6.30498379e-02 1.47277135e-02 9.22537446e-01
7.41482615e-01 9.90098536e-01 2.23950267e-01 1.08175417e-02
4.32768285e-01 2.99525857e-02 1.85911536e-01 -3.97132188e-01
-5.68967402e-01 -2.23245174e-01 2.06761435e-01 -6.28408909e-01
-7.68983603e-01 -6.65384829e-01 -1.43861878e+00 -1.78225666e-01
-3.91765952e-01 5.41193336e-02 1.09778464e+00 1.22288644e+00
2.06002265e-01 3.98212761e-01 4.95155811e-01 -8.76008391e-01
3.50659378e-02 -7.51951098e-01 -4.36508000e-01 3.62640359e-02
1.79887310e-01 -7.59997547e-01 -9.14535597e-02 2.09097773e-01] | [8.8145170211792, -1.4199206829071045] |
a13a6f71-6443-44f6-b952-a889b715f2cc | online-sequence-clustering-algorithm-for | 2305.08418 | null | https://arxiv.org/abs/2305.08418v1 | https://arxiv.org/pdf/2305.08418v1.pdf | Online Sequence Clustering Algorithm for Video Trajectory Analysis | Target tracking and trajectory modeling have important applications in surveillance video analysis and have received great attention in the fields of road safety and community security. In this work, we propose a lightweight real-time video analysis scheme that uses a model learned from motion patterns to monitor the behavior of objects, which can be used for applications such as real-time representation and prediction. The proposed sequence clustering algorithm based on discrete sequences makes the system have continuous online learning ability. The intrinsic repeatability of the target object trajectory is used to automatically construct the behavioral model in the three processes of feature extraction, cluster learning, and model application. In addition to the discretization of trajectory features and simple model applications, this paper focuses on online clustering algorithms and their incremental learning processes. Finally, through the learning of the trajectory model of the actual surveillance video image, the feasibility of the algorithm is verified. And the characteristics and performance of the clustering algorithm are discussed in the analysis. This scheme has real-time online learning and processing of motion models while avoiding a large number of arithmetic operations, which is more in line with the application scenarios of front-end intelligent perception. | ['Zhitang Song', 'Shunfeng Li', 'Longfei Liang', 'Xingyu Qian', 'Xiaogang Chen', 'Aximu Yuemaier'] | 2023-05-15 | null | null | null | null | ['online-clustering', 'incremental-learning', 'trajectory-modeling'] | ['computer-vision', 'methodology', 'time-series'] | [ 2.17502832e-01 -4.36225981e-01 -3.50248486e-01 -3.85638714e-01
-1.05200969e-01 -4.50711340e-01 3.96462142e-01 1.73213199e-01
-3.52799058e-01 2.59238817e-02 -3.96669298e-01 -4.34438348e-01
-3.83392662e-01 -6.42075598e-01 -5.90216875e-01 -9.42578018e-01
-5.60569823e-01 1.32444846e-02 7.74253607e-01 2.36303002e-01
1.80598691e-01 6.85140848e-01 -1.97541749e+00 3.43166083e-01
6.45729959e-01 1.30021620e+00 3.84131014e-01 1.04598129e+00
8.49623457e-02 1.08821940e+00 -3.10164124e-01 -4.67286855e-02
2.89456844e-01 -1.35004878e-01 -1.27536640e-01 5.88453531e-01
1.62382275e-01 -4.84073490e-01 -4.82133120e-01 8.72560382e-01
1.25314116e-01 3.55055213e-01 4.15060222e-01 -1.61903560e+00
2.41133749e-01 1.21337056e-01 -3.71679693e-01 1.15744457e-01
2.02800184e-01 1.47708073e-01 3.70352656e-01 -4.47541565e-01
3.44987720e-01 9.50534821e-01 8.38137865e-01 3.15700978e-01
-6.28059149e-01 -5.26687682e-01 4.04129803e-01 1.01943004e+00
-1.47056365e+00 -3.20759207e-01 7.39087164e-01 -6.77989960e-01
6.30196273e-01 3.81734341e-01 1.01660049e+00 5.78406870e-01
3.14041197e-01 9.93403852e-01 3.62466455e-01 -3.28821212e-01
6.02064610e-01 7.68794790e-02 3.37067097e-01 6.50409222e-01
1.58087745e-01 2.95809776e-01 7.34259561e-02 -2.64948994e-01
5.51145434e-01 6.21596158e-01 -2.26414144e-01 -5.08046985e-01
-1.01506317e+00 6.47222221e-01 9.38878134e-02 9.28077325e-02
-4.32297826e-01 9.30684060e-02 5.10358095e-01 2.14734331e-01
3.10975406e-02 -3.51867735e-01 -3.43281686e-01 -1.76613837e-01
-1.08588886e+00 -5.90962823e-03 6.42465830e-01 1.30614805e+00
6.85748160e-01 5.40153533e-02 1.56436991e-02 2.68348217e-01
2.42456466e-01 7.29352951e-01 4.63962406e-01 -1.07732582e+00
-1.83349792e-02 7.72773385e-01 2.25670710e-01 -1.32521164e+00
-5.52284658e-01 1.27886117e-01 -9.51158226e-01 2.74652660e-01
3.82829279e-01 -1.42571986e-01 -4.90548611e-01 1.23801601e+00
7.84498692e-01 7.33542204e-01 4.77673225e-02 5.82484663e-01
2.61072069e-01 1.00975323e+00 -3.20018157e-02 -9.55138206e-01
9.49091792e-01 -7.87468672e-01 -8.04121792e-01 4.48641598e-01
7.30084419e-01 -3.23152930e-01 4.70437586e-01 4.79413956e-01
-6.55878663e-01 -8.88703346e-01 -7.12466121e-01 6.15413487e-01
-2.25294322e-01 3.36051255e-01 5.49514413e-01 6.97599173e-01
-1.02810359e+00 3.30830067e-01 -1.27337670e+00 -6.03850186e-01
6.93110302e-02 6.12360179e-01 1.06134981e-01 -4.50390950e-02
-7.75108159e-01 5.21968722e-01 7.24236250e-01 1.90318301e-01
-8.86013985e-01 -5.50504327e-01 -6.86631024e-01 -1.23507045e-02
4.89987791e-01 -2.55590826e-01 1.12139094e+00 -1.20747530e+00
-1.42083573e+00 2.56470501e-01 -3.81933123e-01 -5.27172148e-01
5.32436848e-01 3.65370587e-02 -6.58947229e-01 5.07865250e-01
-3.85616750e-01 3.03475618e-01 1.22317660e+00 -8.42306137e-01
-1.29472959e+00 -2.25955367e-01 -1.60286590e-01 -5.65051176e-02
-5.87118268e-01 2.45471057e-02 -4.53911275e-01 -2.78518647e-01
-1.97532445e-01 -9.57931340e-01 -4.14007843e-01 2.17461750e-01
3.16432148e-01 -1.28253177e-01 1.42162788e+00 -6.67290509e-01
1.67216229e+00 -2.20592403e+00 -2.77597129e-01 5.57511866e-01
-3.61420549e-02 4.68624949e-01 1.61005214e-01 4.56997216e-01
1.29607290e-01 -5.73073030e-01 -2.42156219e-02 2.84728289e-01
-4.99019831e-01 -2.71822196e-02 -2.44773284e-01 5.36958158e-01
-1.20444022e-01 4.66364801e-01 -9.47915316e-01 -7.06620276e-01
5.11267364e-01 1.87323451e-01 -3.16872835e-01 4.18544590e-01
-1.82187110e-01 2.66718745e-01 -4.61892098e-01 4.17424321e-01
5.77726960e-01 -1.04989283e-01 4.29966092e-01 -3.58068973e-01
-4.17916983e-01 -6.26194715e-01 -1.30176103e+00 1.07462788e+00
-1.37997732e-01 6.19902372e-01 4.49951850e-02 -1.09590697e+00
8.52746725e-01 4.34665561e-01 1.02870250e+00 -3.02136421e-01
1.78368151e-01 -2.36088529e-01 -1.30296277e-04 -9.94448304e-01
4.28355336e-01 5.61822116e-01 2.85778493e-01 3.60444129e-01
-2.94226289e-01 3.91807228e-01 2.55994141e-01 1.69470608e-02
1.10265303e+00 -3.50162722e-02 4.11045671e-01 -3.20724957e-02
7.95634747e-01 4.36689913e-01 6.27139330e-01 5.77399492e-01
-2.60897070e-01 -1.40584692e-01 -2.97916401e-03 -7.55344927e-01
-1.03017509e+00 -8.46299648e-01 1.81612834e-01 8.24365735e-01
3.78976107e-01 -3.35002691e-01 -8.41190577e-01 -4.25705642e-01
-3.87146056e-01 4.51642185e-01 -3.52815896e-01 -3.49853903e-01
-7.77147889e-01 -5.91891825e-01 1.51596338e-01 5.19445360e-01
3.49560589e-01 -8.73695791e-01 -1.24061549e+00 4.03201938e-01
-1.34877235e-01 -1.20227671e+00 -3.84707302e-01 -5.62588453e-01
-1.00116873e+00 -1.43434167e+00 -1.68880045e-01 -8.66193116e-01
7.52183139e-01 8.79269004e-01 3.46861631e-01 1.47018775e-01
-3.31638873e-01 1.12576115e+00 -5.28228939e-01 -4.41739202e-01
-5.41321456e-01 -2.49075413e-01 2.80217737e-01 6.93670869e-01
5.64601898e-01 -2.13284135e-01 -5.06464660e-01 5.08908510e-01
-8.79367292e-01 4.97860797e-02 4.02324319e-01 5.56042612e-01
5.06559849e-01 6.38929009e-01 4.11026150e-01 -3.07505727e-01
1.65773287e-01 -4.87950802e-01 -1.23883164e+00 6.32874429e-01
-2.67705470e-01 -5.20039260e-01 8.49261045e-01 -9.51468349e-01
-1.01187193e+00 6.52597785e-01 3.52016121e-01 -8.76519799e-01
-2.00362042e-01 1.99736863e-01 -2.45363563e-01 -6.50619790e-02
3.40274245e-01 6.64396822e-01 5.73955059e-01 2.13770583e-01
1.20627411e-01 7.28060246e-01 3.88820857e-01 -7.44661167e-02
6.93606257e-01 6.83056533e-01 1.08346276e-01 -1.44307280e+00
-2.55809695e-01 -8.65974128e-01 -8.23736429e-01 -9.86012578e-01
9.54702616e-01 -8.47874165e-01 -1.33112693e+00 6.93307996e-01
-1.17778945e+00 -3.51333618e-01 1.37978988e-02 9.01468515e-01
-8.67624104e-01 6.49568558e-01 -4.25200611e-01 -1.29208708e+00
-1.17267303e-01 -9.04569089e-01 5.97372711e-01 8.33038762e-02
8.93924311e-02 -1.11194432e+00 -1.73502937e-01 -1.60275444e-01
3.22446674e-02 3.40293616e-01 5.29927015e-01 -4.14578110e-01
-9.55992043e-01 -4.22317088e-01 1.18743628e-01 1.60867527e-01
1.43714532e-01 3.65016550e-01 -7.40678549e-01 -5.80629826e-01
3.46370310e-01 2.44817376e-01 4.44214314e-01 6.29347444e-01
1.21261811e+00 -5.19203901e-01 -6.95595980e-01 4.78608549e-01
1.34139347e+00 8.51792216e-01 5.02057374e-01 1.10914513e-01
6.18442535e-01 8.66197944e-01 1.18506157e+00 6.81621075e-01
1.43149257e-01 5.93541205e-01 4.70970839e-01 3.12506229e-01
3.91711801e-01 1.26749024e-01 7.05273330e-01 8.13635528e-01
-1.18746042e-01 -5.06335832e-02 -8.51858199e-01 4.82993156e-01
-2.37686253e+00 -1.76535845e+00 -2.62769550e-01 2.41683912e+00
1.11374505e-01 -1.48690507e-01 5.00087380e-01 2.36552313e-01
1.06816757e+00 -1.90205380e-01 -6.31493211e-01 2.47100387e-02
3.71181518e-01 -6.99096084e-01 5.36506772e-01 3.65409404e-01
-1.35197163e+00 6.66988313e-01 6.04836082e+00 8.84801269e-01
-1.20681751e+00 -1.05215691e-01 3.46520036e-01 1.85596347e-01
5.09533823e-01 -2.21981809e-01 -8.73228610e-01 6.82242930e-01
1.19784153e+00 -2.10493743e-01 3.39654505e-01 1.00815427e+00
7.84759700e-01 -8.07007551e-02 -1.16577315e+00 1.08507466e+00
2.15111803e-02 -1.22849309e+00 1.66548431e-01 -8.25058669e-02
4.56347018e-01 -2.64101923e-01 -2.42362082e-01 -3.93444449e-02
-1.09743491e-01 -4.62119758e-01 7.43763030e-01 6.94310308e-01
4.37556684e-01 -8.72291982e-01 4.71283019e-01 9.21914458e-01
-1.43913984e+00 -7.90849864e-01 -5.22866726e-01 -1.28562257e-01
2.54804164e-01 3.30779850e-01 -9.07865107e-01 5.18816471e-01
7.16604888e-01 8.88962269e-01 -2.71724343e-01 1.34210193e+00
4.00515139e-01 8.09310853e-01 -3.32867533e-01 -2.69732267e-01
1.45116478e-01 -3.50968868e-01 5.29899299e-01 1.35183823e+00
4.94873762e-01 4.02926147e-01 5.62282979e-01 1.12703465e-01
7.20256567e-01 -5.00109931e-03 -7.38516033e-01 2.07070827e-01
5.99833727e-01 1.18092918e+00 -7.28386223e-01 -5.99895000e-01
-5.94347954e-01 5.15933216e-01 -4.57199253e-02 1.86960548e-01
-1.21306086e+00 -2.72561282e-01 5.51457226e-01 2.99229592e-01
6.45710886e-01 -6.60123825e-01 2.29190588e-01 -9.38607395e-01
1.46178557e-02 -6.15987837e-01 3.82452428e-01 -1.69762641e-01
-8.92487586e-01 4.01194811e-01 3.55989307e-01 -1.98710692e+00
-4.84182268e-01 -7.14948595e-01 -9.40729499e-01 -1.93500053e-02
-1.09711123e+00 -1.09765601e+00 -5.73197544e-01 1.03902280e+00
6.92810595e-01 -3.42925370e-01 4.43072855e-01 2.52128333e-01
-5.02177417e-01 4.23535436e-01 3.12422872e-01 9.15678442e-02
1.92628220e-01 -5.20295858e-01 -2.94052754e-02 1.00681531e+00
-2.19430879e-01 1.93284139e-01 4.98854369e-01 -6.71467960e-01
-1.56706917e+00 -1.62445366e+00 2.69919902e-01 -1.91898853e-01
6.58903241e-01 -2.22343519e-01 -8.06163132e-01 4.88983959e-01
-1.99556813e-01 -8.06862265e-02 6.57707393e-01 -6.57669961e-01
1.09696209e-01 -5.19318044e-01 -1.01253557e+00 5.01543224e-01
9.13151205e-01 7.36679584e-02 -1.39750853e-01 3.75504613e-01
6.29327416e-01 -4.88575399e-02 -6.26956046e-01 3.25592756e-01
6.59591615e-01 -8.16996098e-01 8.49396706e-01 -5.49316883e-01
-3.76427382e-01 -6.38643801e-01 -2.02389657e-02 -6.19980752e-01
-6.36792123e-01 -6.98019087e-01 -6.11129403e-01 9.84830379e-01
-1.17675118e-01 -3.69180918e-01 7.40252674e-01 5.44830918e-01
-2.17744354e-02 -5.26249349e-01 -9.12000775e-01 -1.04349649e+00
-7.07820594e-01 -5.31085432e-01 2.98821628e-01 7.18764842e-01
8.24235603e-02 -1.58841178e-01 -5.20245850e-01 6.76255882e-01
9.41457629e-01 1.72255993e-01 9.09178019e-01 -1.13415110e+00
-2.20265508e-01 -1.82054803e-01 -9.35303926e-01 -1.14544857e+00
9.21524540e-02 -3.80509347e-01 7.42678568e-02 -1.10336137e+00
6.13710098e-02 -2.95984089e-01 -3.72633599e-02 -1.03283472e-01
1.60610273e-01 -4.62513417e-01 1.82184622e-01 4.10445005e-01
-8.96369755e-01 3.68891239e-01 6.01355314e-01 -2.54398942e-01
-4.13114637e-01 6.69519126e-01 2.59975970e-01 8.29961300e-01
7.78062224e-01 -1.56930566e-01 -8.46264601e-01 -1.34934247e-01
-4.25772369e-01 2.20836073e-01 4.39125240e-01 -1.33893013e+00
8.24400663e-01 -4.54641700e-01 2.95434773e-01 -8.02849412e-01
3.91009808e-01 -1.77222490e+00 2.73606300e-01 9.79962826e-01
-7.93415904e-02 1.76688850e-01 3.30128707e-02 9.23119247e-01
-9.53927189e-02 -1.19386546e-01 7.05199420e-01 2.36300230e-02
-1.23356056e+00 5.13094664e-01 -9.47143734e-01 -5.73532462e-01
1.90063500e+00 -7.84769237e-01 1.00683965e-01 -5.10207891e-01
-7.79684901e-01 4.47765917e-01 2.61675000e-01 3.56698662e-01
8.61157238e-01 -1.12119186e+00 -3.68888944e-01 2.55555391e-01
-1.08031169e-01 -1.97990224e-01 5.08271217e-01 9.14832115e-01
-5.04540682e-01 2.76737928e-01 -1.30404651e-01 -1.05270588e+00
-1.51796699e+00 1.36604047e+00 1.46393225e-01 1.69646487e-01
-5.30042708e-01 -2.16146335e-01 5.86919151e-02 9.77092460e-02
5.56896210e-01 -3.04173112e-01 -3.16401571e-01 -2.82265767e-02
1.06398392e+00 9.22065139e-01 -3.49028468e-01 -6.38066113e-01
-2.12638512e-01 7.07696140e-01 1.39194354e-01 2.90086478e-01
1.07766259e+00 -4.02611166e-01 -1.48503371e-02 5.38745165e-01
9.88262057e-01 -5.02648175e-01 -1.67587757e+00 7.85800070e-03
2.86203861e-01 -3.53707343e-01 -3.49365830e-01 -7.03720152e-02
-7.63771355e-01 7.45384812e-01 8.97969007e-01 2.61402220e-01
1.32844973e+00 -3.95626694e-01 7.50541031e-01 7.73811936e-01
7.56551802e-01 -1.16429329e+00 4.02370803e-02 3.25373292e-01
3.52364659e-01 -1.07147968e+00 -2.41642758e-01 -5.77084661e-01
-5.66416562e-01 1.30821586e+00 6.97709441e-01 -6.43530637e-02
9.74980295e-01 2.89456010e-01 -5.54389581e-02 2.22590029e-01
-8.02586317e-01 -4.62273322e-02 4.90459353e-02 9.44869697e-01
-2.67693549e-01 -9.54284295e-02 -4.81020510e-02 2.24925354e-01
3.67109090e-01 1.51459187e-01 4.92656916e-01 9.78945196e-01
-9.45187271e-01 -5.28254330e-01 -6.30749047e-01 2.23891363e-01
1.01117035e-02 5.21553457e-01 1.46913618e-01 6.49644613e-01
2.91082468e-02 1.24900842e+00 2.36247450e-01 -6.34510219e-01
1.45788535e-01 -1.52541414e-01 3.06119353e-01 -1.11026485e-02
-1.76371872e-01 -4.98655587e-02 -3.51535201e-01 -6.99360013e-01
-6.69422448e-01 -8.44253063e-01 -1.32331836e+00 -1.69901520e-01
-1.84749275e-01 2.12244347e-01 5.86348414e-01 7.00420558e-01
4.23432648e-01 1.34111911e-01 1.11594963e+00 -8.95701349e-01
-5.17247558e-01 -5.76584280e-01 -2.42508128e-01 2.10185155e-01
3.37973505e-01 -4.38544095e-01 -1.87384500e-03 6.46261096e-01] | [8.357586860656738, -0.7898454666137695] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.