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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
f38c22b9-af7e-4d70-a966-a5b9f36d837d | extrinsic-factors-affecting-the-accuracy-of | 2305.18152 | null | https://arxiv.org/abs/2305.18152v1 | https://arxiv.org/pdf/2305.18152v1.pdf | Extrinsic Factors Affecting the Accuracy of Biomedical NER | Biomedical named entity recognition (NER) is a critial task that aims to identify structured information in clinical text, which is often replete with complex, technical terms and a high degree of variability. Accurate and reliable NER can facilitate the extraction and analysis of important biomedical information, which can be used to improve downstream applications including the healthcare system. However, NER in the biomedical domain is challenging due to limited data availability, as the high expertise, time, and expenses are required to annotate its data. In this paper, by using the limited data, we explore various extrinsic factors including the corpus annotation scheme, data augmentation techniques, semi-supervised learning and Brill transformation, to improve the performance of a NER model on a clinical text dataset (i2b2 2012, \citet{sun-rumshisky-uzuner:2013}). Our experiments demonstrate that these approaches can significantly improve the model's F1 score from original 73.74 to 77.55. Our findings suggest that considering different extrinsic factors and combining these techniques is a promising approach for improving NER performance in the biomedical domain where the size of data is limited. | ['Jungyeul Park', 'Yujie Song', 'Shengjie Zhang', 'Zhiyi Li'] | 2023-05-29 | null | null | null | null | ['named-entity-recognition-ner'] | ['natural-language-processing'] | [ 4.01876308e-02 1.70629755e-01 -1.29189461e-01 -2.22268343e-01
-9.53549147e-01 -4.46634263e-01 1.41590595e-01 7.32450426e-01
-8.79412651e-01 1.00575829e+00 4.12134081e-01 -4.28292155e-01
-6.32318929e-02 -3.84285480e-01 -1.16341069e-01 -5.00303984e-01
2.47827485e-01 3.31676960e-01 -1.05874576e-01 -1.90503284e-01
1.52801916e-01 5.44324636e-01 -8.95829260e-01 2.88644433e-01
1.10982502e+00 5.10083675e-01 1.88299462e-01 6.53596938e-01
-4.27906126e-01 8.92627537e-01 -7.87900925e-01 -4.52712089e-01
-1.98168769e-01 -4.46391642e-01 -1.03769147e+00 -5.48208117e-01
-4.30276215e-01 1.50312200e-01 -1.50546432e-02 1.01952744e+00
9.68018115e-01 9.64563489e-02 5.87029397e-01 -5.68655849e-01
-3.96936387e-01 7.46463180e-01 -2.24879146e-01 3.58899415e-01
2.76581347e-01 -1.99180886e-01 5.73502958e-01 -7.31128812e-01
8.26085389e-01 5.41523397e-01 1.03717089e+00 9.04123724e-01
-8.08961987e-01 -6.26118720e-01 -5.08563757e-01 -3.72179076e-02
-1.49531829e+00 -5.33055484e-01 2.99189210e-01 -4.09981072e-01
8.99309814e-01 3.15820515e-01 8.55084881e-02 1.04926491e+00
2.14833885e-01 3.99617970e-01 8.78737867e-01 -6.73483789e-01
2.31952950e-01 2.96760798e-01 3.15381944e-01 5.26175618e-01
3.96164417e-01 -2.74818242e-01 -2.52460718e-01 -4.97304916e-01
3.39564204e-01 6.85134828e-02 -3.17380428e-01 4.71169442e-01
-1.21472692e+00 4.67273653e-01 1.21304877e-01 8.57794166e-01
-5.47566473e-01 -5.36645114e-01 6.88269436e-01 5.38412742e-02
3.13425004e-01 1.01239061e+00 -8.60340953e-01 -3.37300390e-01
-9.31085706e-01 -3.60704154e-01 1.04509032e+00 9.83809114e-01
1.08493604e-01 -2.38390341e-01 -3.56231958e-01 9.62518990e-01
4.94035743e-02 2.77959198e-01 8.27006340e-01 -5.55932403e-01
4.96092051e-01 7.89617360e-01 2.84814760e-02 -8.09775233e-01
-9.19631898e-01 -3.09800327e-01 -1.18084037e+00 -5.91550052e-01
3.40375453e-01 -5.34076393e-01 -1.13165903e+00 1.49827158e+00
3.94698977e-01 -2.45166458e-02 4.36997116e-01 6.02285981e-01
1.39894485e+00 2.96136767e-01 8.45234334e-01 -2.81826794e-01
1.69499493e+00 -7.60913730e-01 -1.17728376e+00 2.89747212e-03
1.27628505e+00 -1.01626444e+00 5.33556044e-01 1.38312846e-01
-7.11667478e-01 -2.25105852e-01 -7.11670995e-01 -2.54485190e-01
-6.26492441e-01 5.56420684e-01 4.86367375e-01 6.95208848e-01
-6.93435013e-01 6.12496793e-01 -1.02124357e+00 -6.03464425e-01
5.16532719e-01 4.20394868e-01 -6.91329122e-01 -1.35877162e-01
-1.46263194e+00 1.03701568e+00 6.21407509e-01 2.58818716e-01
-2.73345232e-01 -7.49553084e-01 -9.11099672e-01 4.15342040e-02
1.49036080e-01 -6.53947830e-01 9.30853009e-01 -7.16388151e-02
-1.03760338e+00 7.16645420e-01 -7.47409984e-02 -2.49721110e-01
2.30920598e-01 7.72172911e-03 -8.08751881e-01 8.43474343e-02
2.17465699e-01 2.89512396e-01 -1.20304026e-01 -6.54156148e-01
-5.82528412e-01 -5.41269600e-01 -4.34198052e-01 8.07661340e-02
-3.92227113e-01 3.87213230e-01 -2.88440198e-01 -6.75233901e-01
-6.33277446e-02 -8.62485290e-01 -7.38822639e-01 -5.86948156e-01
-6.15879476e-01 -3.00599337e-01 1.58573210e-01 -1.02255332e+00
1.64661038e+00 -2.11951327e+00 -3.24519992e-01 1.08821474e-01
3.05768907e-01 5.28883934e-01 2.22201958e-01 4.12190229e-01
-2.13759750e-01 8.04266095e-01 -2.12094009e-01 -1.47840306e-02
-4.53075379e-01 7.86030572e-03 4.51890886e-01 2.27715358e-01
1.90746590e-01 8.82867694e-01 -8.52507532e-01 -7.37613142e-01
-1.68522656e-01 6.36160672e-01 -1.28205419e-01 3.27082306e-01
3.92311633e-01 5.98338068e-01 -7.89007425e-01 7.47951865e-01
3.62417102e-01 -4.26523924e-01 1.74095199e-01 -4.89938080e-01
-2.21821219e-02 2.02525914e-01 -1.03082275e+00 1.68177533e+00
-1.49861231e-01 2.31301427e-01 -9.31893215e-02 -7.64658749e-01
7.84087420e-01 8.06959093e-01 6.59797549e-01 -3.33616525e-01
4.82106745e-01 2.91858554e-01 -5.36069237e-02 -8.95368516e-01
3.96024555e-01 -2.95353562e-01 -2.14717224e-01 -3.39235063e-03
3.69205549e-02 3.82532239e-01 2.98361510e-01 1.24563403e-01
1.47045195e+00 -3.26742917e-01 8.04863989e-01 -1.72589362e-01
5.56605637e-01 2.74393708e-01 9.22833741e-01 5.92301250e-01
-4.00809944e-01 4.75878358e-01 2.72639483e-01 -8.51501897e-02
-9.43602443e-01 -1.61850408e-01 -4.92749244e-01 6.47556901e-01
-3.69768679e-01 -4.91265833e-01 -8.25010478e-01 -9.54400420e-01
-4.69089419e-01 6.77373886e-01 -5.91284811e-01 -1.22178659e-01
-4.83566940e-01 -1.14989221e+00 1.15959370e+00 5.64677835e-01
2.73276985e-01 -1.10311902e+00 -3.45075101e-01 4.48190212e-01
-5.23795724e-01 -1.31893039e+00 -5.57924867e-01 3.43725890e-01
-9.61251795e-01 -1.20385420e+00 -7.32450306e-01 -8.40563893e-01
9.49661911e-01 -2.81213611e-01 8.49611700e-01 1.38300538e-01
-5.27272224e-01 -2.33897921e-02 -5.35475314e-01 -7.30516911e-01
-6.87166333e-01 3.85607690e-01 -1.26532838e-01 -5.29539943e-01
7.41187334e-01 6.94939345e-02 -5.84915996e-01 2.13121653e-01
-1.10570300e+00 -7.34886900e-02 8.20333600e-01 1.18713248e+00
6.45429492e-01 1.09019175e-01 7.62101412e-01 -1.43056726e+00
6.58978939e-01 -5.58289528e-01 5.26327677e-02 4.48659748e-01
-8.51495504e-01 1.33141264e-01 5.92194378e-01 -3.87365103e-01
-1.16943228e+00 5.78048229e-02 -5.13318062e-01 1.76481187e-01
-4.36009824e-01 9.32551324e-01 -1.41601428e-01 1.32578477e-01
9.24834847e-01 -1.25660896e-01 -1.02888152e-01 -7.98672915e-01
-8.81649852e-02 1.15505302e+00 3.21885467e-01 -1.84355408e-01
2.03603312e-01 5.50567321e-02 -1.39122652e-02 -6.03914142e-01
-9.67178524e-01 -8.72843504e-01 -6.29935384e-01 4.24515903e-01
1.08144665e+00 -8.93607259e-01 -4.41583216e-01 1.47347286e-01
-1.07240152e+00 8.88993070e-02 -5.89963906e-02 8.73551786e-01
1.71126962e-01 4.29696172e-01 -9.25129533e-01 -5.99032700e-01
-9.25475419e-01 -9.29233968e-01 7.23744690e-01 3.81646365e-01
-4.47586596e-01 -9.52986479e-01 1.84536204e-02 6.50110900e-01
2.14378655e-01 3.66831779e-01 9.65120256e-01 -1.45233107e+00
4.84121352e-01 -3.27846378e-01 -7.13374168e-02 2.26560816e-01
4.50902879e-01 -3.73010427e-01 -6.91199601e-01 -5.22761680e-02
1.57646209e-01 7.13774655e-03 4.69416171e-01 3.51861082e-02
1.17451346e+00 -3.25169951e-01 -5.47545850e-01 3.01604569e-01
1.29486442e+00 3.90172839e-01 6.18527830e-01 3.42713952e-01
7.62827277e-01 5.39292693e-01 6.94350362e-01 3.35734576e-01
3.78416538e-01 1.53565913e-01 -2.44373038e-01 -4.01321888e-01
1.83457136e-02 1.47272078e-02 -1.62475914e-01 9.89582121e-01
-1.81080610e-01 -1.58170983e-01 -1.23501301e+00 6.80430830e-01
-1.56922901e+00 -7.19896913e-01 -3.28020245e-01 1.97600317e+00
1.36407018e+00 -8.89285058e-02 -3.10062140e-01 8.58406276e-02
8.85030389e-01 -6.86147988e-01 -3.71526480e-01 -2.83848047e-01
-1.79517549e-02 2.21555740e-01 6.59872890e-01 -6.57390133e-02
-1.02241504e+00 7.23332524e-01 6.28529167e+00 7.05736458e-01
-8.55112195e-01 1.37094155e-01 7.12814808e-01 3.89610678e-01
1.16660990e-01 -3.51138562e-01 -1.00152147e+00 5.38728774e-01
1.47979164e+00 -7.18102977e-02 -3.61169390e-02 5.97344697e-01
3.55765462e-01 1.57366037e-01 -1.01423323e+00 8.69455099e-01
1.08528614e-01 -1.28130198e+00 -2.42654815e-01 7.58394524e-02
5.59011400e-01 -6.41589100e-03 -5.13200343e-01 4.43656951e-01
1.67471662e-01 -1.13064408e+00 -9.30535123e-02 5.93876004e-01
8.52113426e-01 -5.59618294e-01 1.50579381e+00 3.95789087e-01
-8.74436021e-01 7.47602358e-02 -1.93614706e-01 6.23086572e-01
1.51433259e-01 8.23333323e-01 -1.19687653e+00 8.24401975e-01
7.01120079e-01 3.49295408e-01 -4.03953552e-01 1.05117154e+00
-1.01204157e-01 7.34553218e-01 -1.89880013e-01 -7.25606084e-02
-1.85279399e-01 2.38574237e-01 2.64992058e-01 1.57666659e+00
3.80902439e-01 6.60026371e-01 6.65579960e-02 5.55517375e-01
-3.95771831e-01 6.63889587e-01 -3.01317096e-01 -4.87275273e-01
5.35574377e-01 1.48087800e+00 -7.98892975e-01 -2.62283593e-01
-3.09423327e-01 7.00421214e-01 2.22255394e-01 1.14553042e-01
-6.77440286e-01 -7.64257431e-01 1.50095716e-01 -8.67002234e-02
-4.44680899e-02 1.32850990e-01 -3.95454913e-01 -1.12312222e+00
-2.55633205e-01 -9.45051610e-01 8.56349289e-01 -4.30164874e-01
-1.31707883e+00 9.19279337e-01 -3.90745103e-01 -1.13016009e+00
-1.53743848e-02 -5.69791019e-01 -1.12628013e-01 7.82614708e-01
-1.43498051e+00 -1.05905461e+00 -1.70832232e-01 3.75714660e-01
2.56796986e-01 -2.09952831e-01 1.22361887e+00 8.17867041e-01
-8.14460278e-01 8.65318775e-01 3.09965461e-01 6.20364487e-01
1.06271803e+00 -1.18184459e+00 -2.61967164e-02 5.13083577e-01
-2.55870134e-01 9.99967575e-01 4.60846484e-01 -7.65185118e-01
-1.08957624e+00 -1.06937242e+00 1.46276748e+00 -5.37725627e-01
3.85430872e-01 3.40243056e-02 -9.82279062e-01 3.25113207e-01
-2.01215625e-01 -3.57487537e-02 1.51986480e+00 1.19417906e-01
1.22612402e-01 1.50548548e-01 -1.56668568e+00 5.69645345e-01
6.24056578e-01 -3.07238668e-01 -7.71794558e-01 1.76981941e-01
6.20743394e-01 -5.01638114e-01 -1.65692985e+00 4.62118626e-01
2.53251433e-01 -1.41921431e-01 6.92330301e-01 -1.02543306e+00
1.24152623e-01 -2.40361840e-01 1.04115225e-01 -1.15567243e+00
-2.41172925e-01 -5.02094805e-01 1.24286503e-01 1.47950840e+00
8.71051490e-01 -4.41809118e-01 5.75155258e-01 1.12967479e+00
-1.33488446e-01 -7.60536611e-01 -8.34328234e-01 -3.17587197e-01
-9.06411186e-02 -1.07247785e-01 5.67082345e-01 1.28658462e+00
4.83123600e-01 4.87250179e-01 -1.40453011e-01 1.40268132e-01
1.93777621e-01 -5.48077524e-01 2.37075195e-01 -1.39365113e+00
2.61307001e-01 2.12086756e-02 -3.69217724e-01 -3.60378295e-01
-1.40791267e-01 -8.97665620e-01 1.77465469e-01 -1.71735013e+00
3.74303311e-01 -6.90697849e-01 -7.00908720e-01 8.52766454e-01
-5.09688914e-01 4.25120033e-02 -1.74966216e-01 4.52943951e-01
-5.34454048e-01 5.54997735e-02 1.01860225e+00 3.39621976e-02
-2.85742432e-01 -1.34369075e-01 -1.13436902e+00 6.48276746e-01
7.36352682e-01 -9.72115338e-01 4.27234322e-02 -2.31106460e-01
9.98196304e-02 8.89641717e-02 -2.56288648e-01 -6.17823005e-01
6.34765863e-01 -8.03648308e-02 5.15806019e-01 -2.96316624e-01
-2.43795514e-01 -8.07087004e-01 3.08237344e-01 4.68060583e-01
-5.70337355e-01 -5.87974414e-02 3.13145876e-01 5.40332139e-01
-1.37012079e-01 -4.99235749e-01 5.90492070e-01 -2.41764650e-01
-3.95112634e-01 1.37304828e-01 -3.31264824e-01 2.51646549e-01
9.15357769e-01 -3.27708982e-02 -1.90634564e-01 2.60992169e-01
-9.71918702e-01 2.53331661e-01 1.25487559e-02 1.88411921e-01
3.44725341e-01 -1.14291906e+00 -7.80405819e-01 -1.31665692e-01
2.25070655e-01 1.02496512e-01 3.98771942e-01 1.02719808e+00
-4.74464774e-01 6.02018297e-01 2.23443490e-02 -1.83326483e-01
-1.45155692e+00 4.07655925e-01 9.52939093e-02 -8.26042652e-01
-5.30411959e-01 4.77541268e-01 -3.36749643e-01 -5.90101361e-01
1.43082067e-01 -2.39937454e-01 -8.97542357e-01 1.72738016e-01
6.14435673e-01 5.19781888e-01 4.97875988e-01 -5.94402611e-01
-5.26933968e-01 1.16190046e-01 -2.78592408e-01 2.70272851e-01
1.44246221e+00 -1.03767373e-01 -2.27628902e-01 1.25631943e-01
1.08989668e+00 1.20825827e-01 -1.46524608e-01 -1.82320461e-01
5.65526843e-01 1.26475319e-01 1.33350179e-01 -1.14656544e+00
-8.22524905e-01 5.08985341e-01 6.66771889e-01 -1.03587352e-01
1.03972328e+00 -7.61680007e-02 8.62085938e-01 6.25463188e-01
1.73114270e-01 -1.22953713e+00 -4.67137843e-01 4.30983841e-01
3.77127767e-01 -1.42593300e+00 4.56099622e-02 -4.24945086e-01
-8.00256491e-01 1.09671235e+00 3.63420099e-01 7.60766745e-01
7.72745132e-01 4.03707176e-01 2.34854788e-01 -2.21358523e-01
-3.70221764e-01 1.30994311e-02 3.95203829e-01 3.46707165e-01
8.15996945e-01 -1.14034250e-01 -9.42316711e-01 1.26627326e+00
4.06089462e-02 2.21417755e-01 4.10052717e-01 9.94486094e-01
-2.12084167e-02 -1.28503680e+00 -2.47273728e-01 6.89368427e-01
-1.47727513e+00 -4.10840392e-01 -2.30115727e-01 5.45889080e-01
1.26591757e-01 1.25611842e+00 -6.86146379e-01 -1.65525123e-01
6.10698223e-01 1.80778131e-01 -1.53067082e-01 -8.52457285e-01
-8.49430621e-01 5.45726791e-02 4.51901644e-01 -1.77299842e-01
-5.70913315e-01 -3.72703522e-01 -1.72338760e+00 2.91067399e-02
-7.17257440e-01 7.20288813e-01 8.01331758e-01 1.09278703e+00
7.42448568e-01 6.88950181e-01 2.89314538e-01 1.05939910e-01
-5.66995084e-01 -1.13772595e+00 -3.80790085e-01 4.40648675e-01
4.09770310e-02 -2.83466667e-01 -2.70175695e-01 3.32748085e-01] | [8.409204483032227, 8.833733558654785] |
2204caac-402b-426f-be12-4fe998416148 | user-friendly-image-editing-with-minimal-text | 2306.02717 | null | https://arxiv.org/abs/2306.02717v1 | https://arxiv.org/pdf/2306.02717v1.pdf | User-friendly Image Editing with Minimal Text Input: Leveraging Captioning and Injection Techniques | Recent text-driven image editing in diffusion models has shown remarkable success. However, the existing methods assume that the user's description sufficiently grounds the contexts in the source image, such as objects, background, style, and their relations. This assumption is unsuitable for real-world applications because users have to manually engineer text prompts to find optimal descriptions for different images. From the users' standpoint, prompt engineering is a labor-intensive process, and users prefer to provide a target word for editing instead of a full sentence. To address this problem, we first demonstrate the importance of a detailed text description of the source image, by dividing prompts into three categories based on the level of semantic details. Then, we propose simple yet effective methods by combining prompt generation frameworks, thereby making the prompt engineering process more user-friendly. Extensive qualitative and quantitative experiments demonstrate the importance of prompts in text-driven image editing and our method is comparable to ground-truth prompts. | ['Gayeong Lee', 'Seungryong Kim', 'Yunjey Choi', 'Junho Kim', 'Hyunsu Kim', 'Wooseok Jang', 'Sunwoo Kim'] | 2023-06-05 | null | null | null | null | ['prompt-engineering'] | ['natural-language-processing'] | [ 4.58051592e-01 -1.05715275e-01 -1.57214906e-02 -6.57454193e-01
-5.27617276e-01 -6.42024994e-01 8.22326720e-01 2.44798273e-01
-4.93355066e-01 3.16678613e-01 4.22254175e-01 -2.50673473e-01
-1.80837594e-03 -5.05611897e-01 -3.09341073e-01 -2.79534280e-01
5.57666957e-01 2.07909733e-01 3.33461285e-01 -3.87835056e-01
5.64292371e-01 2.05047280e-01 -1.33867192e+00 4.01136965e-01
1.10480177e+00 5.12543797e-01 8.63473415e-01 4.24467027e-01
-5.57339489e-01 7.25288928e-01 -6.61359429e-01 -4.11921263e-01
6.26044795e-02 -7.55972683e-01 -7.25047946e-01 6.54405296e-01
4.75947499e-01 -7.54855573e-01 -8.60320628e-02 1.43033552e+00
3.19959939e-01 8.77990425e-02 7.40808904e-01 -1.11366594e+00
-1.11793828e+00 5.19661784e-01 -4.57057804e-01 4.27981950e-02
5.26762247e-01 1.75182343e-01 9.30279851e-01 -1.00632071e+00
8.27287555e-01 1.05756629e+00 1.22919887e-01 7.09484160e-01
-1.09406018e+00 -3.54283333e-01 6.37357414e-01 1.94115385e-01
-1.11179435e+00 -4.49285835e-01 9.06973839e-01 -7.54092634e-01
4.71599102e-01 3.34417731e-01 5.72513342e-01 1.20979571e+00
-3.23888939e-03 6.32898867e-01 1.20127594e+00 -5.51414311e-01
3.08457136e-01 6.52272284e-01 3.45372967e-02 6.09070361e-01
1.67273298e-01 -1.64958149e-01 -6.08374178e-01 -2.59520710e-02
1.03447330e+00 -2.03071944e-02 -3.29374611e-01 -3.88934106e-01
-1.37290239e+00 7.21427619e-01 6.05569892e-02 3.30936968e-01
-4.88812804e-01 -2.23484561e-01 1.18917145e-01 1.11331649e-01
6.13789022e-01 4.12988961e-01 -1.45034373e-01 3.17807905e-02
-1.02480876e+00 3.27676982e-01 5.73726356e-01 1.36249340e+00
6.77484930e-01 -2.26095110e-01 -5.67025304e-01 9.02213216e-01
4.49565709e-01 5.12542307e-01 2.47765973e-01 -9.82512772e-01
2.78255403e-01 7.07090378e-01 4.08439040e-01 -1.23828959e+00
1.59448236e-01 -1.37153730e-01 -6.76097751e-01 5.31594604e-02
3.80017370e-01 -1.09369308e-01 -8.13966870e-01 1.66449130e+00
1.86199084e-01 -3.54327172e-01 -2.17448875e-01 1.04120278e+00
7.73441434e-01 5.78620374e-01 2.70714462e-01 -2.85651326e-01
1.42011154e+00 -1.03344810e+00 -9.43106771e-01 -4.15938616e-01
4.55649942e-01 -8.70600224e-01 1.56198704e+00 1.73079386e-01
-1.03139043e+00 -4.88558799e-01 -8.21223319e-01 -1.86631307e-01
-2.23271474e-01 3.18375617e-01 3.87364894e-01 4.96788055e-01
-1.04861796e+00 2.48017997e-01 -2.79459924e-01 -6.78013086e-01
2.34945118e-01 -8.34322423e-02 -1.04175098e-01 -6.99471235e-02
-1.04850996e+00 8.20855319e-01 9.08635855e-02 -1.43578276e-01
-8.04742754e-01 -6.19553804e-01 -7.09186912e-01 -2.55828514e-03
5.35695970e-01 -8.88722479e-01 1.55722749e+00 -1.45904028e+00
-1.46653068e+00 8.29728007e-01 -4.11272049e-01 1.99180655e-02
7.24585772e-01 -1.33158848e-01 -1.30324259e-01 2.41515368e-01
2.92090386e-01 8.14150929e-01 1.10183465e+00 -1.55901408e+00
-8.14023554e-01 -5.55625632e-02 5.11207521e-01 4.06194121e-01
-6.09028757e-01 2.89896578e-01 -7.84235537e-01 -8.03947330e-01
-9.31880623e-02 -6.99336410e-01 -5.05627036e-01 3.36454272e-01
-4.33115005e-01 5.07275015e-03 7.07993269e-01 -6.06831193e-01
1.25311410e+00 -2.32287765e+00 -9.81968641e-02 9.13443416e-02
4.22314763e-01 -1.23703051e-02 -2.67113358e-01 4.57766712e-01
1.59991741e-01 3.22714925e-01 -1.23046011e-01 -3.79127532e-01
-4.81216349e-02 3.58100571e-02 -3.04258734e-01 2.22267173e-02
7.31666237e-02 7.35904157e-01 -1.08304858e+00 -7.11007476e-01
3.06655407e-01 4.56992149e-01 -4.76476043e-01 2.86874831e-01
-4.45518374e-01 5.50884068e-01 -6.39067769e-01 3.77325952e-01
4.60152417e-01 -4.65036869e-01 -1.83278080e-02 -2.54056633e-01
-9.56294909e-02 1.20815851e-01 -9.53512132e-01 1.68301880e+00
-4.19674069e-01 6.31943703e-01 -3.15984488e-02 -4.19045150e-01
7.31945813e-01 2.31833801e-01 3.08926225e-01 -7.64314294e-01
-2.79214866e-02 -2.13456303e-02 -2.24900916e-01 -7.05143571e-01
7.19700575e-01 -6.72246590e-02 2.02112153e-01 7.58623183e-01
-2.89242893e-01 -3.09556186e-01 4.00651902e-01 6.68724477e-01
7.47918606e-01 1.11653037e-01 3.44742298e-01 -2.79079229e-01
3.37187618e-01 2.02366695e-01 2.66491771e-01 9.22093809e-01
-1.72293916e-01 7.85518765e-01 3.44298601e-01 -1.69168502e-01
-9.41272616e-01 -6.52075171e-01 4.24711913e-01 1.09734666e+00
4.55809832e-01 -6.48709893e-01 -1.12830341e+00 -7.83726215e-01
-5.16436696e-01 8.29721928e-01 -5.97207665e-01 3.46787348e-02
-1.53428063e-01 -3.30042005e-01 -8.27009380e-02 3.92493695e-01
5.50741374e-01 -1.08961427e+00 -8.40451896e-01 5.12717552e-02
-4.64727014e-01 -1.26047719e+00 -1.10390627e+00 -3.99431318e-01
-7.50515044e-01 -8.20794225e-01 -8.68934691e-01 -6.94792747e-01
1.37321591e+00 7.77078509e-01 9.86839235e-01 3.44068766e-01
-5.20745777e-02 5.56163549e-01 -6.11495316e-01 -2.92783111e-01
-4.97890264e-01 -2.09829777e-01 -2.26918653e-01 6.84724972e-02
5.11730649e-02 -2.03663006e-01 -6.91328943e-01 4.01363552e-01
-1.15813768e+00 6.82353199e-01 5.83959281e-01 5.49967945e-01
4.24545437e-01 1.79541051e-01 1.55314282e-01 -1.14236152e+00
1.07893813e+00 9.46922321e-03 -4.10657704e-01 4.96086895e-01
-6.48056030e-01 1.00469701e-02 4.79370207e-01 -5.87051332e-01
-1.40763545e+00 1.30740508e-01 1.61754772e-01 -6.99035078e-02
-2.66953260e-01 4.43723172e-01 -8.68575647e-02 8.45134184e-02
7.07990527e-01 3.61137122e-01 -2.10564449e-01 -2.70991623e-01
5.95618546e-01 5.78032255e-01 2.89567441e-01 -6.92095399e-01
7.54753232e-01 4.44794029e-01 -6.88607633e-01 -8.89439821e-01
-8.84357870e-01 -2.78040916e-01 -6.42644644e-01 -4.39210355e-01
7.08086848e-01 -6.36507452e-01 -1.36818007e-01 2.36333802e-01
-1.59893036e+00 -3.38547528e-01 -2.32624307e-01 1.17231630e-01
-4.24737394e-01 4.34874535e-01 -3.55968386e-01 -7.18060136e-01
-4.66745533e-02 -1.31943965e+00 9.54481363e-01 2.76943028e-01
-4.03887123e-01 -1.08595014e+00 -3.68947595e-01 3.21597844e-01
5.67331314e-01 -6.49799332e-02 1.15152109e+00 -3.88209611e-01
-6.07044399e-01 4.02995385e-03 -5.21637619e-01 2.10978419e-01
4.43176270e-01 1.76515043e-01 -7.19656467e-01 9.60032120e-02
1.54318929e-01 8.70935917e-02 3.70871276e-01 2.31454462e-01
1.11519837e+00 -4.86874312e-01 -2.58469492e-01 1.59795731e-01
1.25177252e+00 2.82557368e-01 3.83478016e-01 2.83751160e-01
6.35316133e-01 9.48058426e-01 7.09369361e-01 3.88759822e-01
6.62515819e-01 5.81665695e-01 1.38053164e-01 -4.01018441e-01
-1.65498778e-01 -4.16189194e-01 2.69255608e-01 6.93994999e-01
1.64329529e-01 -3.94034594e-01 -7.08690166e-01 6.03875041e-01
-1.86738718e+00 -7.20563829e-01 -1.33853778e-01 2.05974460e+00
8.69772077e-01 -1.25972405e-02 -1.24318324e-01 -2.84253359e-01
7.20021248e-01 1.90919697e-01 -4.22213584e-01 -9.88687053e-02
9.50357765e-02 -5.10463536e-01 1.77375436e-01 6.66956127e-01
-6.18566155e-01 1.03905582e+00 6.82769823e+00 6.27548277e-01
-1.11873853e+00 2.38805950e-01 7.15035081e-01 5.85494898e-02
-8.10276926e-01 2.79607207e-01 -6.61966562e-01 3.71642351e-01
3.60992283e-01 -3.42932284e-01 4.20830041e-01 5.35530508e-01
8.37870419e-01 -3.40545893e-01 -1.21796310e+00 1.01443768e+00
1.91870540e-01 -1.19393182e+00 5.20721912e-01 -4.89177108e-02
6.53400183e-01 -5.60386062e-01 5.54650910e-02 -2.14012805e-02
3.20500493e-01 -7.15281010e-01 9.88096714e-01 4.84421283e-01
8.87056351e-01 -2.95394301e-01 2.38633499e-01 5.51270545e-01
-8.36679637e-01 2.79626757e-01 -1.13362789e-01 -9.88264978e-02
3.40004802e-01 6.18903458e-01 -6.16667032e-01 2.49655470e-01
4.60424930e-01 5.32247782e-01 -7.55344033e-01 5.80020845e-01
-4.18675572e-01 3.25683296e-01 3.48755047e-02 -1.36747122e-01
6.92257583e-02 -1.94498196e-01 3.64453703e-01 1.31056821e+00
2.54634976e-01 4.02872950e-01 3.00502568e-01 1.10943878e+00
1.24061428e-01 4.19408411e-01 -5.28094888e-01 -3.65685642e-01
4.44774836e-01 1.29052019e+00 -9.11956072e-01 -5.35808027e-01
-4.92864370e-01 1.33285439e+00 1.17763765e-01 6.34104550e-01
-7.75189579e-01 -2.55887330e-01 2.54595786e-01 3.67429435e-01
-3.05547509e-02 -3.42710137e-01 -4.50897574e-01 -1.13841617e+00
1.20373219e-01 -1.07636380e+00 -9.49084386e-02 -1.15632296e+00
-1.22018278e+00 6.40068233e-01 1.93182349e-01 -1.13602507e+00
-4.42255698e-02 -3.96933228e-01 -5.88826597e-01 8.36689472e-01
-1.36349571e+00 -1.01966202e+00 -6.16596937e-01 5.73564410e-01
1.03703594e+00 2.63116658e-01 5.27541876e-01 3.04835588e-01
-4.38332617e-01 1.83487147e-01 -3.44699502e-01 -4.43454832e-02
9.90246594e-01 -1.13530135e+00 4.54573452e-01 9.92982209e-01
2.47921482e-01 9.83852148e-01 1.01466703e+00 -8.22501004e-01
-1.05699408e+00 -8.82806838e-01 1.14453816e+00 -4.03962940e-01
4.19677645e-01 -2.90368915e-01 -8.99157166e-01 4.34446871e-01
4.60623473e-01 -3.63264740e-01 5.04319787e-01 -1.01032861e-01
-1.46507204e-01 1.05226956e-01 -8.92822027e-01 1.11752844e+00
1.11004770e+00 -6.11810863e-01 -3.55252355e-01 4.94022459e-01
5.84404945e-01 -3.93833846e-01 -3.33612025e-01 6.15985096e-02
3.36479932e-01 -9.94888961e-01 6.78102314e-01 -2.03831136e-01
5.36452055e-01 -3.81214291e-01 3.57678346e-02 -1.31421900e+00
-3.64475101e-01 -6.42370224e-01 3.61086577e-01 1.41134846e+00
4.50990796e-01 -2.86704659e-01 3.90819728e-01 1.03441131e+00
2.62766838e-01 -4.76243645e-01 -2.25235626e-01 -4.68027443e-01
-3.79395783e-01 -4.62671340e-01 4.85637724e-01 9.12380278e-01
2.95845699e-02 5.05798042e-01 -3.85087967e-01 3.69943641e-02
3.72106254e-01 -8.61041620e-02 7.63158083e-01 -1.03950989e+00
-6.71794564e-02 -5.84669709e-01 1.95565134e-01 -1.33616066e+00
-1.63478762e-01 -3.73975992e-01 4.16921049e-01 -1.96320987e+00
5.19847155e-01 -4.40996528e-01 2.61871070e-02 4.04410422e-01
-4.67723042e-01 -1.42595889e-02 2.17310354e-01 5.44642866e-01
-6.77333415e-01 3.35246980e-01 1.52191544e+00 -1.84574708e-01
-1.98507801e-01 -2.72230297e-01 -1.13750935e+00 7.96537638e-01
6.06489003e-01 -4.22209531e-01 -8.71062160e-01 -7.96510816e-01
2.94836193e-01 -1.19505055e-01 3.54897410e-01 -4.97375637e-01
4.13778216e-01 -5.72218835e-01 9.96730328e-02 -2.32121006e-01
3.41299251e-02 -8.86428893e-01 -4.27762270e-02 1.85265765e-01
-6.43802166e-01 8.95675048e-02 -4.06809449e-02 5.34506917e-01
-1.24829769e-01 -3.46346498e-01 6.20853543e-01 -2.84142017e-01
-8.29312980e-01 2.99398154e-01 -5.79547286e-01 -8.29688013e-02
9.76384878e-01 -5.71806669e-01 -1.52544275e-01 -8.12664032e-01
-4.74855185e-01 1.28325716e-01 7.57859588e-01 6.00358963e-01
8.08546007e-01 -1.11255765e+00 -6.76950455e-01 2.31143937e-01
3.56113881e-01 -9.84770954e-02 2.62293011e-01 6.79138660e-01
-2.78082907e-01 7.05132335e-02 -2.16891561e-02 -6.18660808e-01
-1.28855431e+00 5.03985226e-01 2.08795011e-01 1.51110336e-01
-6.94243610e-01 5.76223969e-01 8.11298847e-01 -5.37664555e-02
3.06153089e-01 -3.51598978e-01 -2.03458026e-01 -1.22950776e-02
7.28004098e-01 -1.16203651e-01 -2.89593697e-01 -5.34541845e-01
-1.09786145e-01 5.92694163e-01 -4.04066145e-01 -6.15374029e-01
9.00984883e-01 -6.19646490e-01 1.23701675e-03 2.90665209e-01
5.58249056e-01 1.16455130e-01 -1.51462495e+00 -3.94884229e-01
3.82780423e-03 -7.24427819e-01 1.63369123e-02 -9.19736862e-01
-7.93906093e-01 7.73083210e-01 2.67888814e-01 1.93867803e-01
1.09477997e+00 -8.96082968e-02 5.67641139e-01 2.89588183e-01
2.90254056e-01 -1.20435965e+00 4.27685738e-01 3.46930623e-01
1.08071923e+00 -1.27454996e+00 -5.84707372e-02 -6.52929246e-01
-9.90690410e-01 9.08720195e-01 7.07788706e-01 2.96382189e-01
3.52843523e-01 1.14373557e-01 5.60819209e-01 -2.90575027e-01
-6.96039438e-01 -3.56120616e-02 3.75167727e-01 6.36331320e-01
6.05229318e-01 -2.29204416e-01 -4.36920404e-01 4.75790411e-01
1.00774772e-01 6.57393858e-02 5.81156194e-01 9.15154159e-01
-5.36084592e-01 -1.15677810e+00 -2.77464598e-01 2.90510923e-01
-1.54205978e-01 -2.79711336e-01 -7.38781273e-01 5.10582268e-01
-4.14413549e-02 1.14585674e+00 -3.27917606e-01 -8.28849524e-02
4.03505266e-01 -2.00009122e-01 4.29517537e-01 -9.85017359e-01
-3.93440992e-01 1.74916983e-01 -1.23681650e-01 -3.21008325e-01
-4.67588365e-01 -4.72051859e-01 -1.06366885e+00 -1.38691649e-01
-3.71255249e-01 3.58565673e-02 7.68928587e-01 1.19968235e+00
5.01937509e-01 4.57974881e-01 4.66089606e-01 -6.58002555e-01
-3.72845292e-01 -8.76172721e-01 -3.24688226e-01 6.90710604e-01
1.82071388e-01 -4.87471014e-01 -1.65929437e-01 6.79247737e-01] | [11.343694686889648, -0.2582493722438812] |
07376350-dc1e-421d-a207-ce9f42d79707 | a-novel-membership-inference-attack-against | 2210.08956 | null | https://arxiv.org/abs/2210.08956v1 | https://arxiv.org/pdf/2210.08956v1.pdf | A Novel Membership Inference Attack against Dynamic Neural Networks by Utilizing Policy Networks Information | Unlike traditional static deep neural networks (DNNs), dynamic neural networks (NNs) adjust their structures or parameters to different inputs to guarantee accuracy and computational efficiency. Meanwhile, it has been an emerging research area in deep learning recently. Although traditional static DNNs are vulnerable to the membership inference attack (MIA) , which aims to infer whether a particular point was used to train the model, little is known about how such an attack performs on the dynamic NNs. In this paper, we propose a novel MI attack against dynamic NNs, leveraging the unique policy networks mechanism of dynamic NNs to increase the effectiveness of membership inference. We conducted extensive experiments using two dynamic NNs, i.e., GaterNet, BlockDrop, on four mainstream image classification tasks, i.e., CIFAR-10, CIFAR-100, STL-10, and GTSRB. The evaluation results demonstrate that the control-flow information can significantly promote the MIA. Based on backbone-finetuning and information-fusion, our method achieves better results than baseline attack and traditional attack using intermediate information. | ['Kai Chen', 'Ruigang Liang', 'Shenchen Zhu', 'Peizhuo Lv', 'Pan Li'] | 2022-10-17 | null | null | null | null | ['inference-attack', 'membership-inference-attack'] | ['adversarial', 'computer-vision'] | [-1.95270732e-01 -3.13721895e-01 -3.22279155e-01 -5.57399988e-01
-1.61674067e-01 -1.06397033e+00 4.27718282e-01 -3.94511998e-01
-7.39771843e-01 7.33586788e-01 -3.62171382e-01 -1.00960839e+00
-1.27522826e-01 -7.07902670e-01 -9.53445256e-01 -9.64776754e-01
-1.38856605e-01 2.09874995e-02 7.26851225e-01 -5.32202758e-02
1.40275180e-01 9.46316659e-01 -8.58367920e-01 2.34888643e-01
3.38352680e-01 1.28181827e+00 -2.01724008e-01 5.91657519e-01
7.77442977e-02 9.17152405e-01 -1.18125296e+00 -5.10699987e-01
6.15413427e-01 2.04807147e-01 -5.30993581e-01 -9.02762294e-01
6.32254779e-01 -7.57672906e-01 -8.50803196e-01 1.55367517e+00
6.28967047e-01 3.15380581e-02 -3.44754313e-04 -1.50066471e+00
-6.09603465e-01 1.24032080e+00 -3.39883715e-01 6.42032266e-01
-6.46519482e-01 4.17459011e-01 7.29575872e-01 -1.73256844e-01
1.48370340e-01 1.59505808e+00 5.09148121e-01 9.18472707e-01
-9.61810172e-01 -1.60320055e+00 7.15921879e-01 4.58614558e-01
-1.03973162e+00 -5.20143569e-01 5.55943191e-01 -2.39151753e-02
6.06108308e-01 1.90042660e-01 2.63713598e-01 1.53296077e+00
2.60381728e-01 7.51527965e-01 9.04040396e-01 8.99394155e-02
4.55585271e-01 -2.47964814e-01 5.57494521e-01 3.53462517e-01
3.95217299e-01 6.00470781e-01 -1.78035706e-01 -3.41070056e-01
8.06542933e-01 1.98435202e-01 -3.36879909e-01 3.18636559e-02
-8.83458436e-01 8.38037848e-01 5.69018126e-01 -1.09607898e-01
-7.95169100e-02 5.97259164e-01 8.85996997e-01 6.91029191e-01
-1.16776012e-01 5.53038791e-02 -9.38044786e-01 -9.20997337e-02
-2.81531900e-01 2.32467830e-01 8.88232112e-01 6.85769320e-01
3.31914067e-01 4.82036263e-01 -1.72976762e-01 3.10174495e-01
3.79153818e-01 7.19084620e-01 4.72372323e-01 -9.31299150e-01
6.35892510e-01 2.81218410e-01 -3.80859613e-01 -1.00393975e+00
-2.55417675e-01 -7.36042500e-01 -1.15105963e+00 4.54142690e-01
3.18159014e-01 -7.53452897e-01 -1.02850246e+00 2.07319736e+00
3.65463376e-01 8.51259172e-01 3.29735309e-01 6.44746065e-01
4.99616057e-01 5.96231282e-01 7.58030564e-02 5.61167151e-02
1.06232381e+00 -7.38224924e-01 -5.46516240e-01 1.13008820e-01
4.82570469e-01 -2.91622877e-01 6.41966522e-01 4.34832960e-01
-6.16945744e-01 -4.83142793e-01 -1.30161381e+00 4.71740931e-01
-4.58132058e-01 -3.87332290e-01 5.21371245e-01 1.01708448e+00
-9.76283967e-01 4.40148532e-01 -1.16798103e+00 4.71673876e-01
9.51106131e-01 7.42200196e-01 6.34527206e-02 2.86306024e-01
-1.72506106e+00 3.48436266e-01 7.51739502e-01 2.14244410e-01
-1.22082973e+00 -9.37425494e-01 -2.84610361e-01 3.00533384e-01
5.00400364e-01 -2.65540332e-01 1.29778528e+00 -7.40859210e-01
-1.46463728e+00 1.04348689e-01 3.90690923e-01 -1.12980080e+00
5.49367845e-01 -3.75897378e-01 -7.00614512e-01 2.21856982e-01
-4.79593843e-01 4.54779238e-01 8.20482314e-01 -1.08072007e+00
-7.16894329e-01 -2.89037347e-01 4.27335441e-01 -3.59067142e-01
-6.91581309e-01 1.27980456e-01 -2.99333185e-02 -7.38449335e-01
-1.36689037e-01 -8.75925541e-01 -2.80017525e-01 -6.28427267e-02
-8.13312173e-01 -3.39222163e-01 1.67060506e+00 -6.06390461e-02
1.42821074e+00 -2.21243525e+00 -4.36460584e-01 7.02118158e-01
2.71203190e-01 1.08507979e+00 -7.78167471e-02 -2.36043528e-01
6.42692624e-03 3.66456926e-01 3.33132714e-01 8.19650814e-02
-5.58104813e-02 6.12200022e-01 -1.01381397e+00 3.47050607e-01
-3.15382600e-01 8.11000407e-01 -6.05042279e-01 -1.16187043e-01
8.16569999e-02 2.92643577e-01 -5.52639365e-01 -1.14026792e-01
-2.44517505e-01 -8.68308358e-03 -6.64971709e-01 5.37568331e-01
7.79627740e-01 -1.12741217e-01 -1.24310423e-02 -3.29907805e-01
1.81500703e-01 1.98597275e-02 -1.05256391e+00 9.52702224e-01
-6.50918707e-02 6.20204866e-01 -4.03868444e-02 -1.04929805e+00
7.65287161e-01 3.43346387e-01 1.61781207e-01 -5.29410601e-01
3.07247281e-01 2.02323357e-03 4.65204358e-01 5.20097241e-02
-5.24583943e-02 4.02627379e-01 -7.44942529e-03 3.32471877e-01
-1.88279212e-01 8.88940752e-01 -1.46182567e-01 1.25528827e-01
1.30118334e+00 -5.69826722e-01 -1.99876115e-01 -1.85613200e-01
5.46211481e-01 -4.19517964e-01 7.71094322e-01 1.47935355e+00
-6.71200156e-01 -9.52642411e-02 6.83120966e-01 -6.76590383e-01
-5.74300587e-01 -1.18023455e+00 -2.68066585e-01 1.03979957e+00
1.64479300e-01 -3.76935601e-02 -8.56284320e-01 -1.06929255e+00
6.43200278e-02 6.03936732e-01 -5.37616253e-01 -7.41760373e-01
-6.84623599e-01 -6.60560608e-01 1.36816251e+00 5.55556893e-01
1.22858131e+00 -9.61040437e-01 -5.60500681e-01 2.18765080e-01
1.75285667e-01 -1.18468225e+00 -5.01326442e-01 1.36677504e-01
-8.57094347e-01 -1.03597903e+00 -1.46452919e-01 -6.08632743e-01
4.19257551e-01 1.89896137e-01 6.73989475e-01 5.53374067e-02
1.85410336e-01 -1.79109555e-02 -1.46916881e-01 -5.27955115e-01
-4.87313062e-01 3.31076026e-01 4.16703731e-01 -6.54423088e-02
3.76881480e-01 -7.57023573e-01 -6.18256390e-01 7.43786573e-01
-9.51321125e-01 -4.52594012e-01 5.23593903e-01 7.33740687e-01
3.36575449e-01 3.27089071e-01 8.07294071e-01 -1.09618032e+00
4.85193342e-01 -3.00815701e-01 -9.51458573e-01 5.05564928e-01
-5.36092937e-01 6.77224100e-02 1.03162909e+00 -1.10473752e+00
-7.22652972e-01 -3.23585123e-01 -3.65995109e-01 -1.12042093e+00
-4.04811874e-02 1.65058851e-01 -5.34386456e-01 -4.53745633e-01
6.78823113e-01 6.37449697e-02 -1.75348088e-01 -3.19183856e-01
1.19744226e-01 5.81192791e-01 7.66591191e-01 -6.64738417e-01
1.11508822e+00 4.56511915e-01 9.51089431e-03 -2.45507285e-01
-8.13761473e-01 1.77106470e-01 2.19650306e-02 -1.00574896e-01
5.25696218e-01 -7.77600825e-01 -1.37417901e+00 1.04456830e+00
-1.16233373e+00 -4.36723948e-01 1.39275238e-01 3.22598815e-01
1.57023996e-01 3.27589288e-02 -8.33216906e-01 -4.55328166e-01
-5.89208424e-01 -1.37558961e+00 6.69703782e-02 4.47168678e-01
4.33167100e-01 -9.68669415e-01 -3.77562076e-01 1.49607556e-02
6.58877552e-01 2.72848904e-01 8.73453498e-01 -1.18034410e+00
-7.99573720e-01 -7.46346191e-02 -3.57908875e-01 6.68477774e-01
6.19342551e-03 2.45305777e-01 -1.05874527e+00 -5.21049261e-01
1.25201523e-01 -3.09625745e-01 8.07008684e-01 3.71240377e-01
1.77255249e+00 -1.03176928e+00 -2.71590710e-01 9.32414114e-01
1.37167490e+00 8.69673491e-01 6.02629960e-01 5.59504449e-01
9.57911193e-01 -5.38514033e-02 2.35152096e-01 1.75482020e-01
-9.06241462e-02 2.73603648e-01 9.76775348e-01 2.79610872e-01
3.28792959e-01 -6.70600906e-02 6.08066380e-01 2.87332088e-01
2.88539678e-01 -3.78653556e-01 -6.95360601e-01 -1.14970587e-01
-1.71412218e+00 -1.03784394e+00 4.03347820e-01 1.66231871e+00
9.40425396e-01 8.81700993e-01 -3.56041342e-01 9.62666795e-02
7.43291140e-01 3.70524228e-01 -1.19338882e+00 -3.13681722e-01
-1.40278310e-01 -2.20531672e-02 1.03614914e+00 1.45361900e-01
-1.15347445e+00 1.08913827e+00 5.63082933e+00 1.13207674e+00
-1.51794255e+00 1.38961628e-01 9.71350491e-01 -2.41123036e-01
3.08313351e-02 -2.36887783e-01 -1.33630526e+00 6.93369329e-01
1.11127841e+00 1.70411170e-02 3.46601993e-01 1.12112236e+00
-6.81224391e-02 5.99121094e-01 -7.41716564e-01 6.54786110e-01
-4.50671434e-01 -1.58998501e+00 3.19459200e-01 8.09675530e-02
6.25277460e-01 4.33676541e-01 3.35967362e-01 6.37391448e-01
9.49902654e-01 -8.05267215e-01 6.99742854e-01 1.57645643e-01
6.40352190e-01 -1.22761834e+00 7.15106606e-01 3.09779644e-01
-9.22987640e-01 -5.44347703e-01 -5.19474685e-01 4.32587713e-01
-6.99980482e-02 3.19070041e-01 -2.38843352e-01 2.04803094e-01
9.44734693e-01 3.52260143e-01 -4.86412257e-01 6.60435677e-01
-2.18759686e-01 1.12135732e+00 -3.96985292e-01 -8.14053863e-02
4.60858315e-01 2.24208266e-01 6.50571465e-01 6.89433455e-01
-2.64705896e-01 -1.44891754e-01 9.39592570e-02 6.10267997e-01
-5.09552419e-01 -6.64275587e-01 -3.17595154e-01 6.41357154e-02
9.28971231e-01 1.08909118e+00 -4.14332986e-01 -3.38288456e-01
-1.15707606e-01 2.69706279e-01 3.36747855e-01 4.97825801e-01
-1.13897884e+00 -4.98333514e-01 1.09655130e+00 -4.63935852e-01
3.53836179e-01 -1.98142335e-01 -2.53144920e-01 -9.03082848e-01
-2.59546280e-01 -1.14991879e+00 5.04933953e-01 -3.00158322e-01
-1.20973206e+00 6.77444279e-01 4.04572487e-03 -8.15611184e-01
6.24587871e-02 -7.57948637e-01 -8.01809907e-01 5.16535819e-01
-1.37019014e+00 -7.78537035e-01 2.10407391e-01 1.03941882e+00
1.35248303e-01 -5.98627448e-01 4.92891371e-01 5.37612379e-01
-1.07883048e+00 1.40496898e+00 2.70791978e-01 9.02767539e-01
5.47925293e-01 -8.40278447e-01 7.07995355e-01 1.26905429e+00
8.26782361e-02 9.31118548e-01 4.54125464e-01 -4.16971833e-01
-1.19320607e+00 -1.48491359e+00 -1.25053808e-01 2.13557798e-02
9.51833069e-01 -3.95413637e-01 -1.06291354e+00 7.95450330e-01
-1.52414054e-01 3.75918329e-01 3.53704751e-01 -1.49311572e-01
-7.15026438e-01 -6.32673025e-01 -1.14812148e+00 8.28440428e-01
1.07102847e+00 -3.26172262e-01 -2.23230466e-01 6.26266971e-02
1.53136623e+00 -6.88419282e-01 -7.65584707e-01 5.04285812e-01
5.92595041e-01 -8.15815568e-01 1.24442470e+00 -9.99729693e-01
-7.17514753e-02 -1.31699309e-01 -1.83244273e-01 -1.07449257e+00
-1.76602855e-01 -8.14331830e-01 -6.39017403e-01 1.32014942e+00
2.00134739e-01 -1.29665220e+00 9.87586379e-01 4.46383893e-01
-1.84748664e-01 -7.96220124e-01 -1.18805575e+00 -9.13354576e-01
2.28356883e-01 -3.27973217e-01 1.07630599e+00 8.37769449e-01
-1.01893079e+00 -3.74833085e-02 -1.57169938e-01 6.40587807e-01
6.56467497e-01 -2.78324306e-01 6.42045200e-01 -1.09568775e+00
-2.47353509e-01 -6.11157179e-01 -4.76826161e-01 -1.01069546e+00
4.59127188e-01 -4.67991143e-01 -4.06643569e-01 -5.61303675e-01
-4.86219078e-01 -6.66979849e-01 -1.02349043e+00 9.06077921e-01
1.25682414e-01 -1.45085141e-01 1.81795388e-01 2.63266116e-01
-5.58880746e-01 1.34159341e-01 1.11634457e+00 -1.52815059e-01
-1.03887297e-01 3.82592827e-01 -8.40803504e-01 6.82704389e-01
1.28838396e+00 -6.70634151e-01 -7.90101707e-01 -3.99494112e-01
1.13884374e-01 -1.62628472e-01 6.17734969e-01 -1.12924278e+00
7.11723566e-01 -3.22275937e-01 2.46745154e-01 -7.28959918e-01
-2.64014214e-01 -1.05268013e+00 -3.71654294e-02 8.58430564e-01
-5.23736119e-01 1.77153185e-01 3.72360259e-01 7.91468084e-01
8.45105723e-02 3.18049155e-02 9.54571486e-01 1.10014990e-01
-6.81352019e-01 9.71701503e-01 -1.41426638e-01 2.10863248e-01
1.00700891e+00 8.33398011e-03 -1.00432956e+00 1.21381953e-01
-4.41849500e-01 5.52604735e-01 -1.63616136e-01 4.90884155e-01
5.99501908e-01 -1.23637629e+00 -2.46123850e-01 2.86134094e-01
-4.26779628e-01 3.58600736e-01 3.18435788e-01 5.26982784e-01
-4.13426042e-01 2.41886333e-01 -2.03454107e-01 -5.54886520e-01
-1.34865499e+00 5.13202310e-01 7.96029329e-01 -3.42533052e-01
-7.10345626e-01 9.34009612e-01 1.54545203e-01 -3.57581437e-01
9.12319243e-01 -2.98829317e-01 -1.96217850e-01 -3.15507531e-01
7.70237625e-01 2.80110419e-01 -8.60576052e-03 -3.05265673e-02
-2.80336738e-01 -1.63834810e-01 -7.47389555e-01 6.79943636e-02
9.17264998e-01 3.68174911e-02 6.60746545e-02 3.03594563e-02
1.33835292e+00 -5.38074315e-01 -1.62959325e+00 -5.60464859e-01
-1.41835451e-01 -2.46939361e-01 2.63299018e-01 -7.84916580e-01
-1.77501714e+00 7.52664745e-01 7.42223144e-01 1.85621917e-01
1.05114257e+00 -4.98656482e-01 1.19021499e+00 1.08640313e+00
3.36849749e-01 -8.60123038e-01 1.25294831e-02 8.08158457e-01
3.62337291e-01 -8.54906440e-01 -3.36134672e-01 1.53882742e-01
-1.80086464e-01 1.11257374e+00 1.00229597e+00 -1.78160340e-01
1.14023018e+00 6.26067579e-01 2.76942372e-01 1.23130076e-01
-9.95406151e-01 6.24404371e-01 -2.44410932e-01 2.17216626e-01
-6.76478863e-01 -1.77270472e-01 5.21854699e-01 4.70944315e-01
-2.14353845e-01 -1.53961003e-01 3.95758599e-01 1.00515616e+00
-5.43682873e-01 -8.85415435e-01 -1.80580899e-01 5.92227042e-01
-9.30179775e-01 -2.64088064e-02 -2.07397304e-02 7.22067297e-01
-1.38880238e-01 7.30614603e-01 1.48062661e-01 -8.38984966e-01
1.80195257e-01 -2.58476913e-01 -5.76976463e-02 -1.12531252e-01
-7.55361319e-01 -4.12933618e-01 -3.02720457e-01 -6.37386143e-01
-1.81692645e-01 -2.43799046e-01 -1.35525107e+00 -7.04815924e-01
-3.50838095e-01 8.85583758e-02 5.60557485e-01 8.09818506e-01
2.38501936e-01 7.79717445e-01 1.04297948e+00 -1.58580333e-01
-1.32327628e+00 -5.14948785e-01 -3.83625567e-01 -3.90494354e-02
5.88185966e-01 -6.45757198e-01 -7.90740311e-01 -4.12320882e-01] | [5.592456817626953, 7.815441608428955] |
ef64ad20-5f99-40eb-9aed-2b23d7039ece | distance-guided-ga-based-approach-to | 1901.05564 | null | http://arxiv.org/abs/1901.05564v1 | http://arxiv.org/pdf/1901.05564v1.pdf | Distance-Guided GA-Based Approach to Distributed Data-Intensive Web Service Composition | Distributed computing which uses Web services as fundamental elements,
enables high-speed development of software applications through composing many
interoperating, distributed, re-usable, and autonomous services. As a
fundamental challenge for service developers, service composition must fulfil
functional requirements and optimise Quality of Service (QoS) attributes,
simultaneously. On the other hand, huge amounts of data have been created by
advances in technologies, which may be exchanged between services.
Data-intensive Web services are of great interest to implement data-intensive
processes. However, current approaches to Web service composition have omitted
either the effect of data, or the distribution of services. Evolutionary
Computing (EC) techniques allow for the creation of compositions that meet all
the above factors. In this paper, we will develop Genetic Algorithm (GA)-based
approach for solving the problem of distributed data-intensive Web service
composition (DWSC). In particular, we will introduce two new heuristics, i.e.
Longest Common Subsequence(LCS) distance of services, in designing crossover
operators. Additionally, a new local search technique incorporating distance of
services will be proposed. | ['Hui Ma', 'Soheila Sadeghiram', 'Gang Chen'] | 2019-01-16 | null | null | null | null | ['service-composition'] | ['miscellaneous'] | [ 2.01779500e-01 -5.20824790e-01 1.51183814e-01 -5.49318612e-01
-2.35235438e-01 -7.13157117e-01 3.78913671e-01 -1.80968925e-01
4.66283923e-03 5.35659015e-01 5.36045134e-02 -1.63987920e-01
-6.72564685e-01 -1.01025403e+00 9.67405513e-02 -1.06323326e+00
-1.28680140e-01 6.45032465e-01 5.09519100e-01 -4.43201423e-01
4.74748880e-01 7.67962635e-01 -2.29372525e+00 3.21964771e-01
1.06470919e+00 9.83977795e-01 4.84483749e-01 2.72346467e-01
-9.79008555e-01 -8.49098861e-02 -5.43765903e-01 -4.30696368e-01
2.34542981e-01 -8.94919872e-01 -4.50935841e-01 3.33714813e-01
-7.77062893e-01 2.05806479e-01 4.75352287e-01 1.13371265e+00
6.47159219e-01 1.64280623e-01 1.96858868e-01 -1.85398161e+00
-4.82848465e-01 7.90992618e-01 -4.68871474e-01 -1.01646192e-01
4.99899328e-01 -1.36295989e-01 8.09076548e-01 -4.65327233e-01
7.79415488e-01 9.26385760e-01 6.43894225e-02 7.00325847e-01
-7.77204335e-01 -3.05007756e-01 2.37278193e-01 6.02866232e-01
-1.28239560e+00 -2.13579535e-01 6.84020698e-01 6.08466901e-02
1.07826316e+00 8.98769200e-01 9.42986429e-01 5.06038129e-01
1.80200890e-01 4.41757053e-01 6.17043316e-01 -5.62846839e-01
7.06400156e-01 1.30721807e-01 -4.18013811e-01 -3.74483764e-02
4.82799292e-01 -2.29009435e-01 -1.34683862e-01 -4.37204301e-01
-7.93315992e-02 3.49279881e-01 -2.56072521e-01 -6.27980232e-01
-6.94811106e-01 5.43262541e-01 -1.78499684e-01 7.77915478e-01
-5.16922772e-01 -3.09010744e-01 4.23644871e-01 3.16109866e-01
-1.51283413e-01 2.41244331e-01 -7.87876546e-01 -6.81178093e-01
-1.72910422e-01 2.75311440e-01 1.22462475e+00 1.12320340e+00
3.57162625e-01 -9.22614150e-03 4.18209791e-01 7.43996143e-01
4.19885129e-01 1.46930933e-01 7.86293924e-01 -7.52929747e-01
-1.24037221e-01 1.07932472e+00 -2.58017704e-02 -8.90000880e-01
-6.60640821e-02 -9.42626148e-02 -5.81542850e-01 2.19965801e-01
-3.69364560e-01 1.34912401e-01 -4.78789687e-01 1.29009736e+00
7.95283496e-01 -1.38324201e-01 3.06439430e-01 1.01772237e+00
2.40765750e-01 5.73192656e-01 -4.16581593e-02 -6.28726363e-01
1.19573784e+00 -7.91174173e-01 -8.18921626e-01 6.29734397e-01
2.18104988e-01 -7.64956892e-01 6.72858775e-01 4.43785876e-01
-9.82321799e-01 1.30347237e-01 -8.11893821e-01 4.80258912e-01
-5.32299101e-01 -7.84242392e-01 8.78265917e-01 8.78973961e-01
-1.07890296e+00 3.70898664e-01 -5.31549990e-01 -5.15669227e-01
-2.06475526e-01 4.42072421e-01 -9.23971087e-02 -1.31888255e-01
-9.33274686e-01 6.16973042e-01 7.16306567e-02 4.45693582e-02
-6.11313581e-01 -2.64082313e-01 -3.44782919e-01 3.34915847e-01
5.26339948e-01 -4.43320483e-01 9.89062905e-01 -1.53302908e+00
-1.54134715e+00 3.95020664e-01 2.87086033e-04 -3.44892442e-02
2.21725494e-01 7.66614079e-01 -1.08325613e+00 -2.03207210e-01
-3.07277024e-01 -1.97049603e-01 3.22090566e-01 -1.35562956e+00
-1.37391591e+00 -3.88820142e-01 -4.29101214e-02 2.62781054e-01
-4.28537190e-01 5.50937593e-01 -3.33541393e-01 1.03825759e-02
3.53031009e-01 -7.18783975e-01 -4.32460129e-01 -2.69760817e-01
2.59447426e-01 -2.97759086e-01 9.31939423e-01 -4.58335966e-01
1.21690822e+00 -2.14303231e+00 4.87690657e-01 7.30171084e-01
-1.01488866e-01 1.05430678e-01 -1.15473308e-01 8.35244954e-01
2.13509291e-01 7.93895051e-02 -5.17040968e-01 3.37203920e-01
2.98541129e-01 6.46703243e-01 4.00337696e-01 5.93425967e-02
-1.23277619e-01 1.65198579e-01 -5.84570527e-01 -2.60374337e-01
-3.59433830e-01 4.82936829e-01 -6.10726416e-01 1.66772991e-01
-4.22000915e-01 2.46126667e-01 -5.97275972e-01 1.00679004e+00
6.65501714e-01 1.38954595e-01 7.22710431e-01 2.84593403e-01
-4.87188697e-01 -1.29318058e-01 -1.44886136e+00 1.60129559e+00
-5.13928175e-01 -2.48246297e-01 4.31429863e-01 -1.02766061e+00
1.13358629e+00 7.84013271e-01 1.12302816e+00 -8.50330532e-01
2.03444064e-01 7.04753518e-01 5.62214255e-01 -7.37425745e-01
1.41309887e-01 6.83678910e-02 1.91895336e-01 7.56888568e-01
-4.87426996e-01 -5.96460560e-03 4.84630734e-01 -1.12136237e-01
1.13167465e+00 2.09511593e-01 1.16719879e-01 -1.14178762e-01
8.94735932e-01 -9.21798199e-02 1.02700412e+00 -1.14619866e-01
-2.04488769e-01 1.91299409e-01 5.59857011e-01 -4.95850086e-01
-1.17422235e+00 -6.22392178e-01 1.52656138e-01 9.66568530e-01
2.68106818e-01 -2.52435803e-01 -6.54926956e-01 -3.41923445e-01
6.57623559e-02 6.42273247e-01 1.25227511e-01 -1.73106357e-01
-4.99478996e-01 -6.38886571e-01 2.60735244e-01 -2.53988337e-02
2.03417987e-01 -1.12761939e+00 -7.24548876e-01 6.11099660e-01
9.38287154e-02 -7.69195676e-01 -1.64746895e-01 -8.06805193e-02
-6.83094084e-01 -8.43988717e-01 -3.77879739e-01 -7.37385392e-01
6.57341480e-01 4.58299130e-01 7.95654416e-01 5.68209529e-01
-4.38391954e-01 2.25085765e-01 -8.64967287e-01 -2.04746947e-01
-2.43179426e-01 -2.37059891e-02 -4.63579968e-02 2.82897711e-01
5.67795277e-01 -1.10714412e+00 -3.63964796e-01 5.39110243e-01
-1.17072320e+00 -9.96789560e-02 4.81187642e-01 4.50144291e-01
5.25067210e-01 7.08144844e-01 8.61416399e-01 -4.68728215e-01
8.71232331e-01 -7.96758950e-01 -7.42787123e-01 5.14063120e-01
-9.34268594e-01 -6.29682466e-02 7.79247642e-01 -3.02861392e-01
-1.19533026e+00 -1.31254986e-01 -3.53669733e-01 3.29269260e-01
-1.72706231e-01 5.22795439e-01 -9.27854836e-01 -7.70882964e-02
1.12122460e-03 2.94650227e-01 1.78763345e-01 -6.56810999e-01
1.21226162e-01 1.15943944e+00 -1.24270104e-01 -8.10134590e-01
3.94759774e-01 3.92696291e-01 -5.93779609e-02 -3.36064994e-01
4.92929608e-01 -5.44777334e-01 -9.21344981e-02 -1.18643232e-01
4.76668060e-01 2.45737229e-02 -8.57203603e-01 2.74979055e-01
-8.60492587e-01 4.25183326e-01 -1.19683050e-01 1.61377087e-01
-3.74157667e-01 6.55603036e-02 -4.55990322e-02 -1.02036071e+00
-4.09712136e-01 -1.46111417e+00 5.40817142e-01 4.20402259e-01
4.76345643e-02 -4.30589497e-01 -5.76058291e-02 1.20685712e-01
8.50158155e-01 2.00543612e-01 1.03440356e+00 -8.66193354e-01
-7.26418018e-01 -3.41167957e-01 3.78319882e-02 1.31041452e-01
3.85957658e-01 4.02447581e-01 -1.09871171e-01 -1.24655351e-01
7.21896738e-02 4.78826106e-01 -2.86494285e-01 -2.11982995e-01
1.04371750e+00 -1.58027515e-01 -2.34232351e-01 5.30737817e-01
1.72757649e+00 1.16622126e+00 7.30730951e-01 4.83637363e-01
2.90811121e-01 8.77246141e-01 5.44811308e-01 8.96813631e-01
2.55099416e-01 8.19815576e-01 4.52867180e-01 4.31203693e-01
1.64815232e-01 3.53008986e-01 -1.88941527e-02 1.13867450e+00
-3.70288938e-01 -3.25716585e-01 -1.01488698e+00 3.47506046e-01
-1.92202675e+00 -9.25900817e-01 -2.59152949e-01 1.95333719e+00
7.29488313e-01 -3.50766093e-01 6.80025816e-02 5.25861681e-01
7.84571826e-01 -6.76754177e-01 -6.03235602e-01 -9.17808950e-01
-5.43629155e-02 2.62814555e-02 1.94572687e-01 1.03817791e-01
-1.76490963e-01 4.48215842e-01 4.63982677e+00 4.96801227e-01
-1.07954204e+00 2.13639125e-01 -1.60761562e-03 -1.90168561e-03
-1.15941119e+00 1.82805702e-01 -4.79368031e-01 8.36738646e-01
9.84271049e-01 -6.83113277e-01 9.54871953e-01 8.84673119e-01
3.15953493e-01 -6.81478716e-03 -4.50747728e-01 6.10858619e-01
9.97317880e-02 -9.90942478e-01 -1.05563059e-01 6.54253513e-02
1.07800496e+00 -3.51129502e-01 -4.23464239e-01 -2.88434476e-01
1.89584419e-01 -3.14877748e-01 5.95952332e-01 3.36560458e-01
3.06154877e-01 -1.11917245e+00 8.47471416e-01 2.40879923e-01
-1.13877487e+00 -4.18187559e-01 -1.56437740e-01 2.49908134e-01
4.52169538e-01 4.52699095e-01 -7.60411769e-02 9.43006217e-01
1.07323468e+00 -9.11236852e-02 2.22070683e-02 1.52836764e+00
1.68903008e-01 -1.70252174e-01 -2.11166352e-01 -6.16499603e-01
8.25010389e-02 -6.34621263e-01 4.46046501e-01 6.45551860e-01
6.74003780e-01 2.42048442e-01 6.26948625e-02 4.18301433e-01
2.87395537e-01 5.66338301e-01 -3.15286785e-01 -2.87441671e-01
5.40818930e-01 1.16094816e+00 -8.52799535e-01 -1.28440306e-01
-5.90085983e-01 8.07499468e-01 -2.31660590e-01 3.00888926e-01
-7.42917597e-01 -6.37044549e-01 1.25578964e+00 6.63797781e-02
2.38584191e-01 -3.48581523e-01 -1.67005956e-01 -7.84475088e-01
1.52513862e-01 -1.45618331e+00 2.62019634e-01 -3.88119936e-01
-1.05711806e+00 7.05700874e-01 -3.12639296e-01 -9.91140664e-01
-2.68199295e-01 -8.72193277e-02 -2.36109123e-01 8.95102203e-01
-1.34538829e+00 -8.14809442e-01 -3.39478970e-01 6.73236310e-01
4.82625097e-01 -4.39945251e-01 9.33437586e-01 7.23096073e-01
-3.03546607e-01 -2.24163267e-03 4.05674160e-01 -6.77148163e-01
2.12340906e-01 -6.37512088e-01 1.51468679e-01 8.84849250e-01
-1.38371959e-01 4.93447214e-01 7.67881930e-01 -5.19754589e-01
-1.99160659e+00 -2.78797895e-01 1.10259759e+00 4.28870052e-01
5.59842825e-01 -1.44619361e-01 -6.91235662e-01 1.09450497e-01
1.17235169e-01 -3.92217040e-01 8.42152655e-01 -2.04166725e-01
-3.04165874e-02 -5.38617492e-01 -1.65754592e+00 6.88700378e-01
1.14621067e+00 1.04125082e-01 1.24183930e-01 2.03073815e-01
8.08022380e-01 5.34138344e-02 -6.42727017e-01 2.20552459e-01
6.51832640e-01 -1.04739666e+00 6.34168565e-01 -6.13732696e-01
-2.97629293e-02 -5.72708070e-01 -4.40674812e-01 -1.04621065e+00
-1.69356093e-01 -6.26030385e-01 9.05822963e-02 1.46516776e+00
4.18796718e-01 -8.68858635e-01 6.47655249e-01 9.45838332e-01
-3.17527741e-01 -6.46996140e-01 -8.79313171e-01 -8.66966963e-01
-6.32304072e-01 -2.13273868e-01 1.63824248e+00 8.77701700e-01
2.12264806e-01 -4.92449671e-01 -2.82266252e-02 5.01480028e-02
3.17605317e-01 2.90878832e-01 3.09860766e-01 -1.22433412e+00
-4.28723514e-01 -7.50016987e-01 -6.52719438e-01 -1.52580738e-01
-1.12044051e-01 -6.64507806e-01 -1.55552346e-02 -1.80499351e+00
-7.42233396e-02 -8.08970690e-01 -1.64640889e-01 2.21599951e-01
3.34253699e-01 -3.25927556e-01 1.11285545e-01 2.37895280e-01
-2.93110430e-01 2.46061131e-01 1.08068001e+00 2.40438536e-01
-2.27786347e-01 1.80950344e-01 -6.37217045e-01 2.60031939e-01
9.91391540e-01 -5.43892443e-01 -6.46505833e-01 -4.60733294e-01
5.03034115e-01 4.83631268e-02 -5.13093114e-01 -5.62142670e-01
4.16492432e-01 -8.01683903e-01 -3.73736024e-01 -6.87071458e-02
-3.03500630e-02 -1.66287446e+00 1.00037014e+00 5.69756508e-01
-1.46281630e-01 5.46517193e-01 -3.89359415e-01 2.96108156e-01
-3.82391781e-01 -5.65514922e-01 3.70821834e-01 -2.19449848e-01
-1.00638843e+00 3.48247379e-01 -5.44149339e-01 -5.55758238e-01
1.58222544e+00 -3.49438787e-01 1.11765884e-01 -1.18084833e-01
-4.27201211e-01 1.95833892e-01 6.36875927e-01 7.26749599e-01
5.78608334e-01 -1.10631263e+00 -4.44660664e-01 3.80343109e-01
4.81848884e-03 -4.15850103e-01 1.49697512e-01 5.21231592e-01
-9.08017933e-01 1.78424269e-01 -5.39524317e-01 -1.71991978e-02
-1.30990338e+00 7.23786175e-01 2.38399096e-02 3.55346680e-01
-4.78149205e-01 6.44425929e-01 -7.49769628e-01 -5.36469400e-01
2.80895710e-01 2.32742161e-01 -1.33048430e-01 1.05689280e-02
7.03828573e-01 4.79134738e-01 3.52854073e-01 -3.85515183e-01
-5.89762628e-01 2.62453079e-01 4.54424262e-01 -4.33562696e-01
1.39392829e+00 -2.48954907e-01 -8.27853620e-01 1.84907783e-02
8.80789280e-01 -6.65087923e-02 -5.40708065e-01 2.47350469e-01
4.45497394e-01 -7.96968222e-01 -2.87094951e-01 -9.90764737e-01
-1.26686466e+00 2.71739036e-01 4.92540717e-01 6.52299464e-01
1.55018353e+00 -2.52232283e-01 8.31840038e-01 -1.10778436e-01
8.63367796e-01 -1.09844470e+00 -5.07670939e-01 -6.13760203e-02
6.20585620e-01 -5.52279294e-01 -4.01111573e-01 -3.87466818e-01
-5.54370403e-01 1.08205092e+00 3.84920150e-01 3.48283619e-01
5.54708421e-01 7.13622749e-01 -3.03132534e-02 9.52882692e-02
-1.04718697e+00 -3.04507494e-01 -2.39284843e-01 8.36698532e-01
2.40184858e-01 9.14447457e-02 -1.37450421e+00 6.46235883e-01
2.89719105e-01 -1.23974517e-01 3.19141477e-01 1.46641827e+00
-5.59126914e-01 -1.91044056e+00 -3.75266671e-01 7.58564100e-02
-3.76522034e-01 -2.49738097e-02 -1.87540278e-01 3.17402780e-01
5.61305404e-01 1.09271622e+00 6.85120076e-02 -1.95888534e-01
3.43358874e-01 7.28650987e-02 1.42482057e-01 -1.71855852e-01
-6.91874862e-01 4.83284444e-02 1.22807734e-01 -4.71733004e-01
-3.50990444e-01 -8.86122167e-01 -1.59736311e+00 -5.80465436e-01
-2.82738060e-01 8.12711000e-01 1.62688911e+00 7.08307981e-01
6.75707102e-01 6.62568629e-01 8.46271574e-01 -2.35696167e-01
-4.76289898e-01 -2.92541444e-01 -7.69183517e-01 2.94890434e-01
-6.33470237e-01 -3.97031069e-01 -3.07126820e-01 -2.83687897e-02] | [8.587454795837402, 6.943497657775879] |
14a8fff7-9549-40ec-8a41-ac597d7600ca | aggregating-long-term-context-for-learning | 2009.00681 | null | https://arxiv.org/abs/2009.00681v4 | https://arxiv.org/pdf/2009.00681v4.pdf | Aggregating Long-Term Context for Learning Laparoscopic and Robot-Assisted Surgical Workflows | Analyzing surgical workflow is crucial for surgical assistance robots to understand surgeries. With the understanding of the complete surgical workflow, the robots are able to assist the surgeons in intra-operative events, such as by giving a warning when the surgeon is entering specific keys or high-risk phases. Deep learning techniques have recently been widely applied to recognizing surgical workflows. Many of the existing temporal neural network models are limited in their capability to handle long-term dependencies in the data, instead, relying upon the strong performance of the underlying per-frame visual models. We propose a new temporal network structure that leverages task-specific network representation to collect long-term sufficient statistics that are propagated by a sufficient statistics model (SSM). We implement our approach within an LSTM backbone for the task of surgical phase recognition and explore several choices for propagated statistics. We demonstrate superior results over existing and novel state-of-the-art segmentation techniques on two laparoscopic cholecystectomy datasets: the publicly available Cholec80 dataset and MGH100, a novel dataset with more challenging and clinically meaningful segment labels. | ['Hidekazu Iwaki', 'Yutong Ban', 'Thomas Ward', 'Taisei Kondo', 'Guy Rosman', 'Daniela Rus', 'Ozanan Meireles', 'Daniel Hashimoto'] | 2020-09-01 | null | null | null | null | ['surgical-phase-recognition'] | ['computer-vision'] | [ 1.61348134e-01 2.19424218e-01 -6.98743165e-01 -5.66821218e-01
-3.06921571e-01 -5.97290456e-01 3.24842513e-01 4.17271733e-01
-7.01223075e-01 5.09291403e-02 3.27503949e-01 -6.46866977e-01
-4.41297889e-01 -2.50711590e-01 -5.87750018e-01 -6.29021227e-01
-5.29981554e-01 5.09912848e-01 -9.77991745e-02 1.14338443e-01
8.47741142e-02 5.73063970e-01 -9.73238409e-01 3.65020186e-01
4.13733661e-01 1.04535532e+00 2.38652304e-01 7.61303008e-01
-3.57772294e-03 1.02035773e+00 -2.14854881e-01 3.39359850e-01
3.41797829e-01 -2.82878339e-01 -8.48658621e-01 -2.23906144e-01
3.74337472e-02 -2.71128714e-01 -5.24309039e-01 9.00028944e-01
1.95851013e-01 -2.70452164e-02 3.99663866e-01 -8.79510522e-01
-1.73596404e-02 9.02393639e-01 -1.71044081e-01 3.69720727e-01
6.84871525e-02 3.26086193e-01 3.62163126e-01 -1.73350155e-01
9.68236506e-01 7.19560146e-01 8.66852462e-01 5.68726778e-01
-8.77026677e-01 -4.40625370e-01 2.26551086e-01 -1.49138615e-01
-6.78143978e-01 -2.78871685e-01 4.98672426e-01 -8.00674319e-01
8.52043986e-01 3.68307121e-02 1.00895417e+00 1.38429236e+00
8.37716043e-01 7.69112766e-01 7.73283541e-01 -1.44357145e-01
2.95095108e-02 -3.15035552e-01 2.58797407e-01 1.16816580e+00
2.09490180e-01 5.19637346e-01 -6.09038293e-01 2.19229966e-01
1.07879305e+00 4.58013952e-01 -4.30579394e-01 -7.97196150e-01
-1.95097113e+00 6.17698967e-01 7.96743751e-01 4.32750911e-01
-5.05725861e-01 5.46275437e-01 7.61191785e-01 4.27900478e-02
1.64820533e-03 5.74994862e-01 -5.42855501e-01 -4.07368898e-01
-8.33746970e-01 -3.31240058e-01 1.10138762e+00 1.03728020e+00
2.50142425e-01 -1.95468292e-01 -6.09972358e-01 2.60973990e-01
2.21774518e-01 -1.36165082e-01 6.36884391e-01 -1.20750880e+00
7.63532966e-02 5.57010293e-01 8.42622444e-02 -5.28797865e-01
-1.13612652e+00 -4.79423761e-01 -1.03980637e+00 1.58677652e-01
3.20883036e-01 -8.65939781e-02 -1.67963302e+00 1.54024518e+00
-6.49285316e-02 2.30952531e-01 1.12130679e-01 9.11605835e-01
1.06423903e+00 6.59016669e-02 3.39782387e-01 -1.94027007e-01
1.27084565e+00 -1.10051417e+00 -8.23320150e-01 -4.35619712e-01
9.16123629e-01 -5.17020345e-01 7.12872684e-01 3.60692352e-01
-7.93040037e-01 -3.19024891e-01 -8.93862128e-01 -9.56148058e-02
-2.38679871e-01 3.82711619e-01 1.19649816e+00 1.10049658e-01
-1.07558548e+00 9.00619090e-01 -1.63481593e+00 -5.07692933e-01
6.12100244e-01 6.92717791e-01 -4.76762831e-01 -1.22758793e-02
-6.38858676e-01 1.10121047e+00 5.55776894e-01 6.24085367e-01
-1.32744038e+00 -7.90744483e-01 -1.30450690e+00 -5.85583933e-02
3.64280343e-01 -8.53610694e-01 1.63010657e+00 -7.09751010e-01
-1.46907640e+00 9.53054488e-01 1.79333892e-02 -5.66537559e-01
7.38328636e-01 -3.24314743e-01 1.86151415e-01 2.14020938e-01
-1.97265029e-01 7.07111895e-01 4.53881532e-01 -1.10250235e+00
-3.35346371e-01 -2.46088058e-01 2.80279890e-02 1.64723888e-01
-1.16066210e-01 -2.45754942e-01 -6.21116638e-01 -6.05840325e-01
2.49413744e-01 -1.31003511e+00 -9.07539904e-01 4.96221602e-01
-6.05786026e-01 3.01572680e-01 4.17127192e-01 -6.75041080e-01
9.14938867e-01 -2.14644504e+00 1.79029927e-01 1.38492122e-01
3.33389223e-01 1.34973526e-01 7.14238733e-02 4.36422043e-02
-7.56309107e-02 -1.45632848e-01 -2.10902423e-01 -4.70571905e-01
-3.46008033e-01 6.66052639e-01 6.03604242e-02 6.21955216e-01
-2.14956850e-01 1.05598724e+00 -1.20606768e+00 -6.51876569e-01
4.67729986e-01 2.75311291e-01 -3.68304789e-01 3.86515349e-01
-1.19650826e-01 9.20532942e-01 -3.92335892e-01 5.31857431e-01
9.01406854e-02 -5.62663555e-01 4.50187683e-01 -3.85547906e-01
-1.11389510e-01 1.19263560e-01 -3.86494964e-01 2.66611338e+00
-5.79948545e-01 6.08162165e-01 2.06184071e-02 -9.73093688e-01
3.68865341e-01 4.32644248e-01 9.97338533e-01 -5.70310116e-01
5.19620419e-01 1.51971802e-01 1.74615189e-01 -7.16122925e-01
3.35650623e-01 5.59992827e-02 -3.16702455e-01 1.99382395e-01
2.95547366e-01 1.39915124e-01 3.10469955e-01 1.34681761e-01
1.28029501e+00 3.88199270e-01 3.74409914e-01 -3.11520129e-01
-6.17770627e-02 3.08432937e-01 5.18527389e-01 1.00324643e+00
-6.09146774e-01 5.62561154e-01 6.02695048e-01 -8.67195249e-01
-7.15669513e-01 -9.83220577e-01 1.90071359e-01 7.35973895e-01
4.21075046e-01 -7.75265545e-02 -2.88788587e-01 -9.95548129e-01
-2.29790285e-02 3.45411867e-01 -1.10294569e+00 -4.06870872e-01
-8.31947029e-01 -3.88668090e-01 1.77918687e-01 8.37199748e-01
-1.03667751e-01 -1.49054635e+00 -1.19281244e+00 3.73002619e-01
-1.31488182e-02 -1.31581700e+00 -3.82047296e-01 8.80678892e-01
-1.23259306e+00 -1.26358521e+00 -6.12655759e-01 -1.09155762e+00
8.62592936e-01 3.92595530e-02 1.09301877e+00 -7.60083199e-02
-7.40857840e-01 4.52440262e-01 -7.87126198e-02 -4.60302144e-01
-5.36838293e-01 1.73781812e-01 -1.07973769e-01 -6.11452818e-01
-2.79895514e-02 -2.62015194e-01 -9.05385077e-01 -1.68734729e-01
-6.24114454e-01 4.55879629e-01 1.02702248e+00 8.11156392e-01
4.17316169e-01 -7.05629468e-01 -6.22245744e-02 -1.05587411e+00
3.91904473e-01 -4.60895866e-01 -4.59481120e-01 1.98029220e-01
-5.47343910e-01 4.91232276e-01 4.69363451e-01 -5.67742109e-01
-9.36843395e-01 3.70927304e-01 9.72470865e-02 -8.03702772e-01
-1.90291867e-01 8.79297912e-01 7.32261002e-01 -1.38044553e-02
4.15269345e-01 -1.06778387e-02 1.49618611e-01 -2.29730427e-01
2.71103680e-01 -9.14842915e-03 9.35607255e-01 -3.11195433e-01
1.02705322e-01 6.68334663e-01 1.53630331e-01 -1.58070147e-01
-1.10343444e+00 -9.49231446e-01 -7.64370739e-01 -3.29309106e-01
9.66119707e-01 -6.85517848e-01 -1.19239974e+00 3.68737012e-01
-1.13389838e+00 -7.92513371e-01 -4.64224875e-01 8.33192050e-01
-7.54774094e-01 1.97160453e-01 -1.04812288e+00 -2.97882140e-01
-4.49957728e-01 -1.42146444e+00 1.04349363e+00 2.47117117e-01
-2.84240276e-01 -1.14536285e+00 1.40007526e-01 -2.36357048e-01
4.55762118e-01 6.55094028e-01 8.58415842e-01 -6.23006761e-01
-6.28037512e-01 -1.76612109e-01 -8.28474686e-02 1.04209408e-02
2.44278818e-01 -1.26230404e-01 -5.10442495e-01 -3.13955307e-01
-1.23454854e-01 -1.04577258e-01 9.73386347e-01 9.44120765e-01
1.47268021e+00 -2.19123259e-01 -7.92645931e-01 1.11050200e+00
1.28608978e+00 1.66686535e-01 2.35124290e-01 2.48006925e-01
7.42426753e-01 4.99934018e-01 5.82090795e-01 3.62379581e-01
3.60092878e-01 2.27156252e-01 6.94195211e-01 -4.08661157e-01
-2.34934352e-02 6.27246574e-02 4.26823050e-02 1.00665128e+00
-4.34584804e-02 3.65663506e-02 -1.24419928e+00 5.75771093e-01
-2.20085096e+00 -5.21067798e-01 7.87974298e-02 1.95601320e+00
8.42556357e-01 1.63169667e-01 -5.87120295e-01 -4.92532283e-01
2.31065109e-01 2.81534433e-01 -5.83817184e-01 -2.62817442e-01
6.12023234e-01 1.23105817e-01 8.80458832e-01 2.48672426e-01
-1.42449129e+00 7.88766325e-01 6.44826126e+00 1.89099059e-01
-1.38084745e+00 -2.26205718e-02 5.33596218e-01 -1.31360173e-01
2.17115432e-01 -8.86327550e-02 -2.31233850e-01 1.43447772e-01
8.80003452e-01 1.85486585e-01 1.54925287e-01 8.36163104e-01
3.35058719e-01 -4.15234238e-01 -1.55495250e+00 1.07932222e+00
1.14187337e-02 -1.51696622e+00 -2.32878432e-01 -2.29759857e-01
5.16481280e-01 3.66073251e-01 -9.79842991e-02 1.57854289e-01
5.39258838e-01 -1.23968422e+00 5.57026148e-01 8.84714782e-01
8.32691252e-01 -1.86424896e-01 7.75545359e-01 1.91692427e-01
-1.02787077e+00 -1.89981923e-01 3.00402075e-01 2.96556860e-01
2.34622285e-01 1.55660093e-01 -1.18767262e+00 5.19550860e-01
8.22667181e-01 1.08331740e+00 -2.72110939e-01 1.29924166e+00
-2.46531874e-01 1.97603032e-01 -2.01036081e-01 2.83418119e-01
5.25139511e-01 3.78148295e-02 4.10572201e-01 1.41012800e+00
3.24867427e-01 -2.48642072e-01 3.99142236e-01 5.85354388e-01
-5.99681623e-02 -4.20019418e-01 -6.32612705e-01 -2.07139805e-01
-9.47631001e-02 1.42653632e+00 -1.17755234e+00 -3.03402722e-01
-2.18549132e-01 1.00577998e+00 3.24080020e-01 3.88968885e-01
-6.32808745e-01 -5.20217009e-02 5.10701239e-01 -4.49117154e-01
8.55638534e-02 -5.89921832e-01 -5.22172987e-01 -9.36289072e-01
1.33416103e-02 -4.90853339e-01 5.82477033e-01 -6.57596707e-01
-7.68052280e-01 6.33075416e-01 -1.67082325e-01 -1.44328392e+00
-3.68029952e-01 -9.29708600e-01 -3.62917781e-01 4.80696917e-01
-1.69287956e+00 -1.31335843e+00 -8.86716187e-01 4.53591734e-01
5.13341844e-01 2.47808039e-01 9.85145211e-01 1.51488045e-02
-3.65175515e-01 -4.52257842e-02 -1.73769385e-01 4.36678320e-01
8.33254337e-01 -1.24742997e+00 1.20104834e-01 7.03669071e-01
-6.39342815e-02 8.70278835e-01 6.58146441e-01 -5.29085398e-01
-1.35500085e+00 -1.15762055e+00 5.05856395e-01 -4.94121283e-01
4.43359524e-01 -3.16268653e-01 -4.41479415e-01 1.24606657e+00
1.88723698e-01 3.74606520e-01 6.30948305e-01 -1.13190431e-03
8.31602439e-02 6.73117340e-02 -7.23602355e-01 5.24465621e-01
1.09041882e+00 -2.47974813e-01 -5.66268384e-01 4.21081901e-01
6.93015039e-01 -1.10967803e+00 -9.43758667e-01 8.96711409e-01
8.29226434e-01 -8.33977878e-01 7.71841407e-01 -6.57589495e-01
4.72977191e-01 -6.06301427e-03 6.44700766e-01 -1.16170549e+00
-3.26077677e-02 -7.43198335e-01 -2.67159671e-01 1.59042105e-01
1.89197093e-01 -2.33596802e-01 9.85796094e-01 4.84253824e-01
-8.12717319e-01 -9.01029348e-01 -6.11014545e-01 -4.62889850e-01
-4.23507184e-01 -5.05397260e-01 -1.06001273e-01 9.38358366e-01
1.78961158e-01 -3.39391291e-01 -1.71076670e-01 2.63976127e-01
3.49267960e-01 2.44920164e-01 3.73857349e-01 -1.09857380e+00
3.12557481e-02 -5.46131492e-01 -2.46277839e-01 -8.83683443e-01
3.79879139e-02 -1.03685462e+00 6.90103054e-01 -2.14350486e+00
4.06291395e-01 -6.01402164e-01 -6.90164626e-01 8.59277666e-01
9.87572670e-02 -3.58806215e-02 7.43760914e-02 3.32307994e-01
-8.37338030e-01 2.34953523e-01 1.35858834e+00 -9.27567557e-02
-4.48250771e-01 2.79607642e-02 -1.97892293e-01 9.33985114e-01
4.96014088e-01 -7.27309108e-01 -2.56077230e-01 -5.19716561e-01
1.85460284e-01 4.35645223e-01 4.07580972e-01 -9.82154310e-01
7.91258514e-01 -9.02596936e-02 3.59121531e-01 -5.21068454e-01
9.49400887e-02 -9.17460918e-01 3.38966139e-02 1.04484248e+00
-4.42499220e-01 3.94943021e-02 4.04745162e-01 5.60300529e-01
-3.52235794e-01 1.72888458e-01 5.70317924e-01 -4.93158400e-01
-9.48790312e-01 6.81222260e-01 -4.45965558e-01 -1.56890303e-01
1.05102515e+00 -1.09445140e-01 -2.46545330e-01 -8.32182020e-02
-1.13330507e+00 3.23488027e-01 2.95425087e-01 7.37996161e-01
7.20891237e-01 -8.44559371e-01 -2.34016165e-01 1.03478007e-01
2.75014222e-01 3.87827367e-01 4.48023021e-01 1.39199424e+00
-9.51576769e-01 6.85681522e-01 -3.89518470e-01 -9.76782501e-01
-8.67031395e-01 7.09965825e-01 5.15387535e-01 -6.89284503e-01
-1.04043198e+00 8.00847113e-01 3.93269002e-01 -4.00572360e-01
5.77526391e-01 -1.15917528e+00 -2.73820430e-01 -1.55671611e-01
-1.18207902e-01 -3.01481962e-01 1.60292499e-02 -1.11428164e-01
-5.04059553e-01 3.04581672e-01 -1.02736004e-01 1.98057100e-01
1.40708363e+00 2.21192747e-01 -8.03759787e-04 6.51273012e-01
9.86861467e-01 -5.25025189e-01 -1.42023003e+00 -5.98427281e-02
2.23791122e-01 6.19411990e-02 6.96938112e-02 -1.07124281e+00
-1.28260219e+00 7.72770166e-01 7.04530656e-01 -3.71655077e-01
9.01364625e-01 -3.32995355e-02 7.57281125e-01 3.95167619e-01
5.31546831e-01 -7.35196829e-01 1.12016678e-01 6.65928423e-01
6.13185465e-01 -1.39410031e+00 -4.62315269e-02 -2.20150948e-01
-7.35527039e-01 1.33847356e+00 4.83734190e-01 -4.41583544e-02
7.33920932e-01 6.29709184e-01 5.69842935e-01 -5.76680839e-01
-5.13061821e-01 -5.59086651e-02 4.12558675e-01 2.31491834e-01
4.34638560e-01 4.29179966e-02 -4.27215882e-02 4.32455093e-01
2.18664091e-02 4.27790225e-01 4.77316469e-01 1.38921750e+00
7.61732683e-02 -6.98265135e-01 4.39350903e-01 6.90705240e-01
-5.27187765e-01 -1.25234857e-01 2.53094167e-01 8.81343305e-01
-1.05778210e-01 3.87454093e-01 1.14476673e-01 6.93859458e-02
1.86031446e-01 -1.51104823e-01 5.68233788e-01 -6.92583799e-01
-8.60013485e-01 -1.36194900e-01 1.29983604e-01 -1.26836300e+00
-6.79688573e-01 -3.67501974e-01 -1.48399830e+00 4.23878789e-01
1.20724060e-01 -1.30597427e-01 9.33134317e-01 8.89614105e-01
3.54568422e-01 1.22770977e+00 -4.32988927e-02 -1.21029460e+00
-2.33905509e-01 -8.35347235e-01 -2.26877987e-01 4.61721838e-01
8.51863742e-01 -5.75943291e-01 4.97383848e-02 2.80287623e-01] | [14.068340301513672, -3.3631186485290527] |
d9f34b89-7a7f-4857-a753-34976e573eb9 | assisted-rtf-vector-based-binaural-direction | 2211.17202 | null | https://arxiv.org/abs/2211.17202v1 | https://arxiv.org/pdf/2211.17202v1.pdf | Assisted RTF-Vector-Based Binaural Direction of Arrival Estimation Exploiting a Calibrated External Microphone Array | Recently, a relative transfer function (RTF)-vector-based method has been proposed to estimate the direction of arrival (DOA) of a target speaker for a binaural hearing aid setup, assuming the availability of external microphones. This method exploits the external microphones to estimate the RTF vector corresponding to the binaural hearing aid and constructs a one-dimensional spatial spectrum by comparing the estimated RTF vector against a database of anechoic prototype RTF vectors for several directions. In this paper we assume the availability of a calibrated array of external microphones, which is characterized by a second database of anechoic prototype RTF vectors. We propose a method, where the external microphones are not only exploited to estimate the RTF vector corresponding to the binaural hearing aid but also assist in estimating the DOA of the target speaker. Based on the estimated RTF vector for all microphones and both prototype databases, a two-dimensional spatial spectrum is constructed from which the DOA is estimated. Experimental results for a reverberant environment with diffuse-like noise show that assisted DOA estimation outperforms DOA estimation where the prototype database characterizing the array of external microphones is not used. | ['Simon Doclo', 'Daniel Fejgin'] | 2022-11-30 | null | null | null | null | ['direction-of-arrival-estimation'] | ['audio'] | [-2.59587973e-01 -4.27937865e-01 8.15794408e-01 1.62689596e-01
-1.06682062e+00 -6.32225096e-01 2.29808301e-01 6.05408438e-02
-2.07668096e-01 2.22737148e-01 5.62544107e-01 -2.94386327e-01
-2.32544661e-01 -4.81372446e-01 -5.17932773e-01 -9.66435432e-01
-1.87453657e-01 -4.98920958e-03 1.93078190e-01 -3.96520272e-02
-1.53328208e-02 4.59567547e-01 -1.85948551e+00 -1.54028147e-01
3.44955176e-01 1.33188105e+00 5.37775457e-01 1.02368367e+00
8.76430497e-02 2.20538095e-01 -1.21408045e+00 1.83527261e-01
5.83830714e-01 -3.37147683e-01 3.37215930e-01 -1.99736640e-01
4.20440674e-01 -3.33073765e-01 -8.26025382e-02 1.06708109e+00
1.09868443e+00 2.32930616e-01 6.28297269e-01 -8.77955079e-01
3.02820057e-01 1.26765519e-01 -7.72088170e-02 1.97640389e-01
6.38510168e-01 -8.20479617e-02 6.20926857e-01 -1.08031321e+00
2.13815272e-01 9.85936880e-01 7.35935032e-01 -7.23558217e-02
-7.10456729e-01 -6.89790666e-01 -3.99135888e-01 4.97828424e-03
-1.81705868e+00 -6.01249993e-01 1.07427478e+00 -4.36752707e-01
4.76108968e-01 2.39369497e-01 8.93791080e-01 3.20350796e-01
1.16504394e-02 5.50035127e-02 1.00564063e+00 -8.60002518e-01
6.62881374e-01 2.50421852e-01 4.80354838e-02 3.79964471e-01
-8.90040621e-02 2.99799323e-01 -4.76916701e-01 -6.20267272e-01
4.93270576e-01 -5.44735968e-01 -8.69204462e-01 -3.52671713e-01
-9.58052278e-01 4.93373126e-01 1.14758819e-01 7.34199584e-01
-7.45388746e-01 -1.62200481e-01 -3.85796800e-02 3.68455470e-01
3.20300162e-01 2.75812209e-01 -5.44758961e-02 -1.02967948e-01
-8.13503444e-01 -1.04010366e-01 1.12973070e+00 6.88151956e-01
6.85859561e-01 5.53142428e-01 1.81508645e-01 1.29073584e+00
6.23553514e-01 1.13571000e+00 4.76153582e-01 -6.86943948e-01
4.20742363e-01 -3.42574149e-01 4.32260692e-01 -1.01360738e+00
-2.49835327e-01 -9.89386857e-01 -5.05092144e-01 3.19859147e-01
5.44916749e-01 -4.51762050e-01 -5.86151779e-01 1.58382595e+00
8.22178960e-01 5.48285902e-01 2.59932786e-01 9.02232528e-01
3.58875930e-01 7.20596611e-01 -8.19484890e-01 -4.73661274e-01
1.06010008e+00 -4.89105999e-01 -7.84320295e-01 -5.72929084e-02
2.53565282e-01 -1.34282231e+00 6.51565373e-01 8.58346105e-01
-8.48742306e-01 -8.05573404e-01 -1.09105265e+00 7.67642796e-01
-5.59998788e-02 3.42447050e-02 -2.51632363e-01 1.18392813e+00
-1.27506721e+00 -1.99573681e-01 -4.53864396e-01 2.06837263e-02
-7.23463416e-01 -9.24761146e-02 -1.85981721e-01 -7.38037527e-02
-9.28899407e-01 6.14013314e-01 -3.59863847e-01 2.82711178e-01
-9.21495974e-01 -7.74141014e-01 -7.55291462e-01 -1.46276783e-02
-2.18191385e-01 -4.00989175e-01 1.16079307e+00 -6.04335308e-01
-1.62527728e+00 9.40428749e-02 -5.08942068e-01 -1.76153153e-01
2.01906264e-01 -1.79993808e-02 -1.00576341e+00 2.29982346e-01
2.73770452e-01 -3.28414887e-01 1.03800631e+00 -1.37758577e+00
-6.16822660e-01 -2.27377445e-01 -5.59008300e-01 1.89202756e-01
-2.67505318e-01 -2.64912188e-01 -1.96612552e-01 -5.05732834e-01
6.17192984e-01 -5.97917199e-01 5.78931309e-02 -2.18104988e-01
-4.25775528e-01 1.30009860e-01 3.16821843e-01 -6.97860897e-01
1.34311426e+00 -2.54462552e+00 -4.84625518e-01 6.40589952e-01
-1.79196764e-02 2.07221925e-01 -2.04127997e-01 4.92315501e-01
-2.93054640e-01 -7.32146204e-01 1.99101135e-01 -3.52620274e-01
-2.36759901e-01 -3.22874486e-01 -3.32971305e-01 7.48132348e-01
-6.15983367e-01 -3.23163390e-01 -7.41024077e-01 -7.80582651e-02
3.79677713e-01 8.00321996e-01 -5.33285379e-01 7.17836022e-01
5.43306887e-01 4.32535976e-01 -2.50682324e-01 3.21839809e-01
1.12396145e+00 6.15759015e-01 -2.61849552e-01 -2.41342038e-01
-4.47049320e-01 2.47132733e-01 -1.92849267e+00 1.12026536e+00
-1.04620147e+00 6.87186599e-01 6.58837855e-01 -7.87278891e-01
1.49461889e+00 6.77263737e-01 2.43463397e-01 -5.43964207e-01
-2.74552777e-02 7.24283278e-01 2.25359112e-01 -3.56477767e-01
-3.24260406e-02 -2.55678684e-01 2.96001822e-01 3.64665270e-01
1.60995394e-01 -3.48592043e-01 -3.15890759e-01 -1.65858597e-01
1.09815955e+00 -5.24711311e-01 4.38527435e-01 -2.06054002e-01
8.78188550e-01 -7.50465035e-01 3.10337692e-01 7.34922707e-01
-1.54833436e-01 6.53146744e-01 -2.95883358e-01 -7.72920847e-02
-9.11685348e-01 -1.34942782e+00 -1.68547332e-01 6.91555321e-01
-1.51936054e-01 -1.30904183e-01 -7.41306305e-01 1.43736944e-01
-1.91039756e-01 9.03685391e-01 -9.42607969e-02 1.82647616e-01
-5.74518204e-01 7.99631998e-02 3.97546709e-01 -6.51328564e-02
4.80275214e-01 -1.27023146e-01 -1.95847467e-01 4.91735041e-01
-3.32283974e-01 -8.72949362e-01 -5.41446805e-01 -2.00026147e-02
-5.03786087e-01 -7.75849044e-01 -1.05040586e+00 -6.92109287e-01
3.95280004e-01 6.11053765e-01 6.51946485e-01 -4.66140985e-01
4.26685577e-03 1.08675158e+00 -8.46600160e-02 -7.25807905e-01
-5.97443163e-01 -1.08309174e+00 4.14609879e-01 5.45440435e-01
1.24265943e-02 -1.02269197e+00 -8.40339243e-01 6.71862066e-01
-4.46861923e-01 -6.89322948e-01 9.54273045e-02 4.52195227e-01
4.77570772e-01 3.45941454e-01 7.11264968e-01 1.97139502e-01
6.22147560e-01 -4.73167956e-01 -9.11820710e-01 -2.12048948e-01
-2.07072228e-01 -3.50607753e-01 8.66660118e-01 -5.72052360e-01
-1.16358817e+00 -9.65709165e-02 -4.93069500e-01 -4.21827853e-01
-2.92629600e-01 2.81457901e-01 -4.14247245e-01 -1.37700047e-02
9.24022675e-01 3.87966096e-01 -2.04292789e-01 -7.82852829e-01
8.40451848e-03 1.24928200e+00 6.42024219e-01 -1.05850779e-01
9.05776143e-01 2.92717874e-01 2.84722373e-02 -1.36010396e+00
-1.48342937e-01 -1.13007295e+00 -3.88463475e-02 -4.97521847e-01
4.41737592e-01 -1.08418190e+00 -4.50799465e-01 7.41997421e-01
-1.16042137e+00 6.83886632e-02 -1.26874849e-01 1.42125690e+00
-4.79019105e-01 3.29108536e-01 5.86805381e-02 -1.35398006e+00
-1.14550672e-01 -1.05327165e+00 8.33426595e-01 -1.35586172e-01
-1.17961725e-04 -9.27989244e-01 4.27453101e-01 3.31323177e-01
5.70720017e-01 -1.74301341e-01 5.40585935e-01 -5.50070643e-01
-4.01763111e-01 -5.75578630e-01 5.72228074e-01 6.56119704e-01
4.00235802e-01 -6.62092984e-01 -1.27645946e+00 -3.00318003e-01
7.85779655e-01 4.83346432e-01 5.76263890e-02 8.99131536e-01
4.21287984e-01 -2.04359785e-01 -6.48975596e-02 5.35363138e-01
1.36462462e+00 6.55233085e-01 3.86429071e-01 -9.84909665e-03
1.58861250e-01 4.69225824e-01 7.01065660e-01 4.97123480e-01
9.79173854e-02 8.00969362e-01 4.11422968e-01 3.70455831e-02
-3.99406850e-01 -1.89971462e-01 1.04928106e-01 1.13021815e+00
1.87257305e-01 -4.67042208e-01 -6.64144516e-01 7.90863574e-01
-9.27238762e-01 -5.31549215e-01 -4.06683534e-02 2.92318821e+00
3.76099855e-01 -2.62721419e-01 1.41899645e-01 6.03790939e-01
8.87481570e-01 -4.33383472e-02 -3.04763347e-01 -3.01430285e-01
1.42255038e-01 2.25565389e-01 2.95921296e-01 1.05263853e+00
-4.94495332e-01 -4.87148436e-03 5.55730629e+00 6.51738286e-01
-1.35681748e+00 1.24926515e-01 -1.29141331e-01 2.81527340e-01
-3.49827886e-01 -2.58251935e-01 -7.24533617e-01 4.44607049e-01
1.00966394e+00 -1.63426936e-01 4.01528627e-01 7.76445150e-01
4.25001264e-01 -2.88696676e-01 -9.74327385e-01 1.11559331e+00
1.52355671e-01 -6.52148545e-01 -5.14209151e-01 1.47692904e-01
3.63752723e-01 -8.56870860e-02 1.48771212e-01 -1.44868419e-01
-8.58878866e-02 -4.29639667e-01 6.72691643e-01 4.60467726e-01
5.95772088e-01 -6.11370504e-01 7.12069094e-01 5.11594534e-01
-1.43474519e+00 -2.02062786e-01 -2.67763674e-01 -3.64587829e-02
1.68808296e-01 1.04476619e+00 -1.61173248e+00 2.64366120e-01
6.75166547e-01 -1.26077980e-01 1.48251802e-01 1.67946279e+00
-1.47093266e-01 1.00554538e+00 -6.77621841e-01 -9.48610753e-02
-1.28315436e-02 -2.14718208e-02 1.25807536e+00 1.13176060e+00
1.10111487e+00 -6.75096316e-03 -1.89173356e-01 2.59444743e-01
3.08544427e-01 4.64717567e-01 -6.77716255e-01 5.70002198e-01
8.38810384e-01 7.94707119e-01 -2.39597023e-01 -9.82590169e-02
-1.42979845e-01 5.06436706e-01 -6.22778356e-01 8.68748069e-01
-4.11464483e-01 -6.19876921e-01 6.68542802e-01 2.59310246e-01
5.67075670e-01 -2.56228387e-01 3.77535760e-01 -5.31987607e-01
9.81696770e-02 -7.54580736e-01 3.45847346e-02 -1.07926428e+00
-9.74273682e-01 6.41633093e-01 -2.02805012e-01 -1.88166976e+00
-7.08097756e-01 -4.13513720e-01 -7.12649345e-01 1.55392122e+00
-1.19854772e+00 -5.19199312e-01 -2.16551013e-02 8.74147356e-01
2.66729206e-01 -3.72927159e-01 9.68260825e-01 3.35253865e-01
2.41484076e-01 5.91641486e-01 6.66290879e-01 -3.11275423e-02
7.34945595e-01 -1.01537240e+00 -2.37132996e-01 6.08100235e-01
-1.09981865e-01 7.30342090e-01 1.21131289e+00 -2.30011731e-01
-1.38981009e+00 -7.41572917e-01 8.50848138e-01 1.10713832e-01
5.09892881e-01 -3.82515609e-01 -6.87685072e-01 3.06617111e-01
-1.86603114e-01 1.67309180e-01 1.07172811e+00 -1.76305294e-01
-3.29737633e-01 -7.13758409e-01 -1.13937664e+00 1.40455201e-01
3.62123191e-01 -6.82768703e-01 -5.26970148e-01 2.23104000e-01
4.21504617e-01 -2.59075582e-01 -6.95654094e-01 2.85770036e-02
5.82358897e-01 -1.17926919e+00 1.20805287e+00 4.68202978e-01
-4.72377300e-01 -7.27164984e-01 -6.73412979e-01 -1.73421836e+00
7.51301600e-03 -5.56663454e-01 2.75975496e-01 1.17563796e+00
8.47766101e-02 -1.24257791e+00 2.98966616e-01 -1.37815803e-01
-2.07934231e-01 -1.55446589e-01 -1.38603735e+00 -9.99240816e-01
-3.16671342e-01 -9.20094788e-01 4.54522252e-01 4.84049320e-01
-3.95373315e-01 2.71643490e-01 -2.82874048e-01 9.06960487e-01
9.46936429e-01 -3.89253721e-02 9.29104805e-01 -1.06115627e+00
-7.00152099e-01 3.79869379e-02 -6.59459472e-01 -1.48296046e+00
-2.88274765e-01 -1.96695611e-01 4.11868870e-01 -1.39320195e+00
-8.45892310e-01 -4.93094504e-01 -3.38485271e-01 -4.69445854e-01
1.98838085e-01 8.74605700e-02 -5.62709048e-02 -3.83270383e-02
2.53474593e-01 1.16988003e-01 9.17018712e-01 3.79217410e-04
-5.18323779e-01 9.19078767e-01 -2.14318931e-01 9.55642700e-01
2.85353124e-01 -4.91814077e-01 -5.17054915e-01 -1.97801620e-01
-5.90036958e-02 7.70556450e-01 1.75878763e-01 -1.43556607e+00
5.53578556e-01 4.20674592e-01 2.18498185e-01 -8.42267454e-01
8.30573380e-01 -1.10495079e+00 3.96773040e-01 4.34915900e-01
5.62440716e-02 -5.02627015e-01 2.00844437e-01 7.99108982e-01
-5.15609205e-01 -2.72978991e-01 7.34240592e-01 2.32490852e-01
-1.61447510e-01 -2.55497605e-01 -7.37378418e-01 -5.25553167e-01
5.52777827e-01 -3.81079495e-01 9.46700722e-02 -1.32298183e+00
-6.56363487e-01 -5.24215817e-01 -1.91582218e-01 -1.38600782e-01
7.28013694e-01 -1.19527602e+00 -7.18349159e-01 5.86848319e-01
-7.33219907e-02 -3.53807032e-01 4.73417878e-01 4.45179969e-01
-9.70145315e-02 4.97503221e-01 1.64010003e-01 -8.26077342e-01
-1.44578731e+00 4.44942027e-01 6.51772738e-01 2.18373522e-01
-5.19430339e-02 9.68345582e-01 5.09579659e-01 -2.82882750e-01
3.83472562e-01 -1.21494964e-01 -4.00917649e-01 -1.45435840e-01
7.87574828e-01 5.49023092e-01 3.70228887e-01 -8.37800145e-01
-2.42343888e-01 1.01277399e+00 7.26014316e-01 -7.82896638e-01
1.03971636e+00 -6.13306105e-01 9.29228961e-02 5.23889840e-01
1.71524060e+00 1.17806065e+00 -6.72342300e-01 -4.78881538e-01
-5.59359550e-01 -7.56420970e-01 2.49758050e-01 -8.58435273e-01
-7.32754648e-01 1.06446481e+00 1.26305330e+00 2.95099199e-01
1.71289599e+00 -1.65850610e-01 3.71223688e-01 3.76576453e-01
6.54918909e-01 -6.71621382e-01 2.22773403e-01 2.06336349e-01
9.14173067e-01 -6.08484983e-01 -6.07058048e-01 -3.12501758e-01
-8.85125529e-03 1.17559862e+00 9.90913715e-03 6.46411553e-02
1.17424834e+00 3.18622887e-01 6.50157690e-01 2.22295552e-01
-8.44044760e-02 -5.28336056e-02 2.43871033e-01 9.49756444e-01
2.71159381e-01 6.91838115e-02 6.18089996e-02 3.79099578e-01
-6.34417951e-01 -4.62907702e-01 5.97765982e-01 4.87302631e-01
-8.10501397e-01 -8.36881280e-01 -1.28799939e+00 -7.56834671e-02
-3.21216643e-01 -2.80114233e-01 2.26169080e-01 1.56303138e-01
1.51399434e-01 1.51983142e+00 1.79000080e-01 -2.89236009e-01
7.79130876e-01 1.88672051e-01 4.59206337e-03 -2.56316185e-01
-3.10639977e-01 8.33536327e-01 1.48954973e-01 -2.38874689e-01
-8.64153951e-02 -4.53424245e-01 -7.56316900e-01 2.01316774e-01
-6.21882737e-01 6.82567477e-01 1.35100591e+00 4.86419469e-01
7.47262388e-02 3.04505140e-01 1.40642536e+00 -4.85104799e-01
-4.74815667e-01 -9.32303667e-01 -1.02809632e+00 -1.78538293e-01
8.91961157e-01 -3.83205384e-01 -1.04342473e+00 -1.59283817e-01] | [15.136013984680176, 5.773771286010742] |
ad8fdb9b-d017-43f5-b0d2-dd7cc9c59805 | spatiotemporal-self-supervised-learning-for | 2303.16235 | null | https://arxiv.org/abs/2303.16235v1 | https://arxiv.org/pdf/2303.16235v1.pdf | Spatiotemporal Self-supervised Learning for Point Clouds in the Wild | Self-supervised learning (SSL) has the potential to benefit many applications, particularly those where manually annotating data is cumbersome. One such situation is the semantic segmentation of point clouds. In this context, existing methods employ contrastive learning strategies and define positive pairs by performing various augmentation of point clusters in a single frame. As such, these methods do not exploit the temporal nature of LiDAR data. In this paper, we introduce an SSL strategy that leverages positive pairs in both the spatial and temporal domain. To this end, we design (i) a point-to-cluster learning strategy that aggregates spatial information to distinguish objects; and (ii) a cluster-to-cluster learning strategy based on unsupervised object tracking that exploits temporal correspondences. We demonstrate the benefits of our approach via extensive experiments performed by self-supervised training on two large-scale LiDAR datasets and transferring the resulting models to other point cloud segmentation benchmarks. Our results evidence that our method outperforms the state-of-the-art point cloud SSL methods. | ['Mathieu Salzmann', 'Sabine Süsstrunk', 'Wei Ke', 'Tong Zhang', 'Yanhao Wu'] | 2023-03-28 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Wu_Spatiotemporal_Self-Supervised_Learning_for_Point_Clouds_in_the_Wild_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Wu_Spatiotemporal_Self-Supervised_Learning_for_Point_Clouds_in_the_Wild_CVPR_2023_paper.pdf | cvpr-2023-1 | ['point-cloud-segmentation'] | ['computer-vision'] | [ 2.09392533e-01 -2.15832323e-01 -3.41791391e-01 -4.62833524e-01
-8.22178245e-01 -7.12677598e-01 8.59833777e-01 4.99769598e-01
-3.93359900e-01 3.69193643e-01 -5.97774506e-01 -3.69702429e-01
-1.65399566e-01 -7.44645715e-01 -8.66644800e-01 -4.48910147e-01
-2.53913730e-01 7.73948193e-01 8.52730870e-01 -2.62835566e-02
3.78217489e-01 8.63010406e-01 -1.66403449e+00 5.81921339e-02
9.47998345e-01 1.00771654e+00 1.61116719e-01 2.20157370e-01
-5.99726319e-01 2.91634262e-01 -7.94839785e-02 1.51147088e-02
5.85938811e-01 -6.23715706e-02 -7.09109664e-01 2.65397519e-01
6.12363875e-01 2.42639840e-01 3.39328021e-01 8.95911753e-01
4.31401730e-02 -7.86741544e-03 5.12401104e-01 -1.58119369e+00
6.04886971e-02 1.99479267e-01 -8.91619861e-01 1.37904584e-01
7.62124360e-03 1.73500806e-01 9.16333675e-01 -9.85246897e-01
5.22776902e-01 1.12533724e+00 9.01503503e-01 7.90982619e-02
-1.46998978e+00 -7.19237328e-01 3.16551685e-01 -3.95503407e-03
-1.32095218e+00 -1.86458454e-01 1.04488158e+00 -8.00475478e-01
5.75786889e-01 5.97748086e-02 7.11068034e-01 5.29546857e-01
-3.81434113e-01 8.15592766e-01 1.38753033e+00 -4.48789686e-01
5.07944226e-01 1.65627676e-03 1.58208534e-01 5.07139087e-01
1.22002982e-01 2.63120383e-01 -4.68564570e-01 -1.28025144e-01
6.39913440e-01 1.75922796e-01 1.70003504e-01 -1.11602211e+00
-1.26315844e+00 7.76543021e-01 5.01836419e-01 1.79402918e-01
-1.23118334e-01 1.21594332e-01 3.97962034e-01 3.75669524e-02
7.01979280e-01 2.14912802e-01 -5.93206763e-01 1.02178492e-01
-1.37710178e+00 2.06229866e-01 6.26193464e-01 1.04385197e+00
1.19456410e+00 -3.99695367e-01 2.53277451e-01 6.64736927e-01
3.64921838e-01 3.63336742e-01 7.94843435e-02 -9.31482852e-01
2.97585368e-01 7.72769749e-01 1.27765194e-01 -6.95138276e-01
-3.24784935e-01 -2.28265345e-01 -3.21443707e-01 5.03009915e-01
3.35506678e-01 1.53774694e-01 -9.46386993e-01 1.41776609e+00
6.79143906e-01 6.95541024e-01 -1.37011290e-01 6.75793529e-01
5.04152417e-01 3.59384686e-01 1.65993869e-01 -1.52601704e-01
8.92190039e-01 -9.21571195e-01 -2.94422179e-01 -2.39083648e-01
6.16664410e-01 -6.94471598e-01 1.08284473e+00 1.33980677e-01
-9.23801243e-01 -6.90680444e-01 -9.03699636e-01 2.47704133e-01
-5.72381854e-01 1.04638331e-01 6.43857360e-01 4.16332036e-01
-8.69949639e-01 8.58970821e-01 -1.21929133e+00 -5.67861259e-01
7.15880692e-01 4.22984153e-01 -1.21013403e-01 1.26008317e-01
-5.97990334e-01 5.38508415e-01 5.31527698e-01 -1.30348504e-01
-3.22358757e-01 -9.75153387e-01 -8.45588624e-01 -3.18759024e-01
4.33195919e-01 -4.42749619e-01 1.18464577e+00 -8.30433846e-01
-1.06718040e+00 1.21795535e+00 -1.59324020e-01 -5.97251952e-01
8.11420441e-01 -4.03542548e-01 -6.53116405e-02 1.62935123e-01
5.36754549e-01 9.20050502e-01 8.44997585e-01 -1.75760603e+00
-1.09999228e+00 -5.06783605e-01 -1.95374295e-01 -5.98338852e-03
-2.60570012e-02 -1.72442883e-01 -6.04293942e-01 -5.07580161e-01
4.60785121e-01 -1.16740620e+00 -3.48586559e-01 3.45873147e-01
-4.07826841e-01 -4.47844654e-01 1.42566741e+00 1.58392265e-02
8.42351317e-01 -2.20766783e+00 -1.14996292e-01 3.78952622e-01
1.32441595e-01 3.41241419e-01 1.30647108e-01 3.60318631e-01
-9.42579657e-02 3.39966640e-02 -6.17397189e-01 -7.58170068e-01
-6.78526834e-02 3.40267956e-01 -4.62720811e-01 5.24510205e-01
4.42914069e-01 8.03016901e-01 -1.17496204e+00 -9.67418492e-01
5.56393087e-01 1.11926652e-01 -4.26240414e-01 1.32937253e-01
-4.87879455e-01 6.21921062e-01 -3.03417921e-01 7.26086795e-01
7.26635277e-01 -2.50027418e-01 -9.48319957e-02 -3.04330260e-01
-4.22573447e-01 5.80479167e-02 -1.11539221e+00 1.89685202e+00
-2.10518032e-01 4.89433020e-01 -1.22712895e-01 -9.26966488e-01
8.27351093e-01 -1.06121592e-01 1.03276837e+00 -3.10924709e-01
-1.92507550e-01 3.80844086e-01 -2.98556536e-01 -3.15666169e-01
3.99319917e-01 -2.82065094e-01 1.28944010e-01 3.74552578e-01
-5.90818487e-02 -4.46388930e-01 2.10997209e-01 1.60415098e-01
9.53017950e-01 5.80030680e-01 1.70617685e-01 -3.00856590e-01
4.39119905e-01 4.30895656e-01 5.78310132e-01 6.07412398e-01
-2.87140936e-01 6.75182283e-01 1.35022551e-01 -2.98000395e-01
-9.25634801e-01 -1.25876510e+00 -2.94464231e-01 7.94331133e-01
4.67842221e-01 -3.55254322e-01 -5.05800843e-01 -8.74594629e-01
3.44166756e-01 4.69916970e-01 -2.87048727e-01 1.87062070e-01
-6.58265173e-01 -2.46785164e-01 2.64028460e-01 7.98200905e-01
3.82645816e-01 -9.10150290e-01 -8.21203709e-01 7.27712363e-03
7.32028782e-02 -1.40329325e+00 -2.18420774e-01 2.91722566e-01
-1.22491217e+00 -1.16042268e+00 -2.94429123e-01 -8.16408098e-01
6.06573641e-01 5.36929071e-01 1.16109705e+00 5.20094037e-02
-1.51374698e-01 5.91081738e-01 -2.92115659e-01 -5.49613535e-01
-5.71239516e-02 2.34447911e-01 -8.98855645e-03 -8.21960121e-02
5.87340891e-01 -7.96259820e-01 -3.73444468e-01 3.93793344e-01
-8.23665321e-01 1.28228702e-02 5.11144102e-01 4.06049877e-01
8.86913240e-01 -6.39696792e-02 3.53729218e-01 -1.02839696e+00
9.99269634e-02 -4.47728634e-01 -8.92276406e-01 1.86437313e-02
-6.08893514e-01 -1.38050318e-01 2.25615948e-01 -2.22131595e-01
-7.82582939e-01 6.77025676e-01 1.79013744e-01 -7.69563377e-01
-4.69056994e-01 3.33245039e-01 2.38391664e-03 -2.56526172e-01
3.78189147e-01 5.04610725e-02 -7.09435120e-02 -4.75228488e-01
5.34553885e-01 4.82848525e-01 8.03217471e-01 -8.61732602e-01
1.27711952e+00 9.45017159e-01 8.64696950e-02 -7.13326693e-01
-8.57862651e-01 -1.08277678e+00 -1.30773973e+00 -2.35846967e-01
8.51093829e-01 -8.11617494e-01 -3.66331905e-01 1.89182550e-01
-1.04332387e+00 -4.10166979e-01 -5.97712159e-01 3.91390860e-01
-7.39410400e-01 4.96590376e-01 -2.63018668e-01 -8.87025535e-01
8.42338279e-02 -9.63460326e-01 1.55198765e+00 -2.90801423e-03
-1.46414831e-01 -8.90169144e-01 1.73577845e-01 2.03875065e-01
-7.89180398e-02 5.05045831e-01 6.96790457e-01 -7.47980297e-01
-8.56360793e-01 1.22538442e-02 -3.01755816e-01 2.45939508e-01
2.49996573e-01 1.24666385e-01 -1.00364780e+00 -2.38333821e-01
-2.06166461e-01 -2.66246825e-01 8.78031194e-01 1.95710942e-01
1.25464153e+00 2.24526718e-01 -6.99485600e-01 5.16441941e-01
1.43838978e+00 7.79439211e-02 3.58551472e-01 4.43949819e-01
9.05140877e-01 7.90565848e-01 1.08270741e+00 2.36290812e-01
4.52086598e-01 7.42833495e-01 5.65720856e-01 -4.45778430e-01
3.29005606e-02 -2.96610802e-01 -1.20017968e-01 5.14898837e-01
6.32982701e-02 2.45569929e-01 -1.22040296e+00 8.95807862e-01
-2.16787004e+00 -7.03955114e-01 -3.57493937e-01 2.21487665e+00
7.63772190e-01 3.84655714e-01 3.69077295e-01 3.47547680e-01
7.25242496e-01 1.15509935e-01 -5.12632787e-01 2.34569013e-01
9.91346017e-02 3.99634689e-01 5.66691101e-01 2.36338645e-01
-1.50650263e+00 1.25633419e+00 5.39043045e+00 6.16434753e-01
-1.05434895e+00 1.00484729e-01 2.07729429e-01 1.79872721e-01
7.74356425e-02 3.60062987e-01 -6.54130638e-01 4.91175443e-01
4.25422519e-01 2.78211594e-01 -1.16795920e-01 1.02275181e+00
1.83138683e-01 -6.12175129e-02 -1.43481326e+00 9.79773700e-01
-3.16896200e-01 -1.34362662e+00 -3.62096250e-01 1.48218662e-01
6.37082934e-01 2.27865338e-01 -1.83688879e-01 9.77481380e-02
3.88986439e-01 -6.92042291e-01 8.09220016e-01 2.82576472e-01
5.96020103e-01 -5.02469718e-01 3.79984379e-01 4.51640308e-01
-1.57691360e+00 1.09070040e-01 -2.40122396e-02 8.19580480e-02
3.20892632e-01 6.67398632e-01 -8.28335702e-01 6.18316650e-01
8.88532400e-01 1.11295378e+00 -7.30090082e-01 1.27223933e+00
-1.20464198e-01 6.53072357e-01 -6.47856772e-01 5.47524810e-01
3.92734081e-01 -4.26113188e-01 5.37355423e-01 1.38924837e+00
1.38230011e-01 -1.34589851e-01 6.49384081e-01 9.46046650e-01
1.75059646e-01 -6.07474446e-02 -7.68809855e-01 2.02356949e-01
7.26997852e-01 1.25562084e+00 -1.22124457e+00 -1.77038133e-01
-4.20213461e-01 7.34867096e-01 2.74277896e-01 6.98945820e-02
-6.69270337e-01 -3.60489450e-02 4.98512805e-01 4.11521137e-01
4.84734893e-01 -5.27612805e-01 -5.91129422e-01 -7.70687103e-01
1.26158580e-01 -3.02778244e-01 3.53833288e-01 -6.82772577e-01
-1.53435874e+00 2.42930070e-01 2.97365904e-01 -1.63638461e+00
-1.03871971e-01 -4.01047349e-01 -6.43895805e-01 4.05365855e-01
-1.97332788e+00 -1.51784432e+00 -5.27024090e-01 4.87410396e-01
5.42153895e-01 5.89201711e-02 3.15278113e-01 2.58772165e-01
-3.26795220e-01 1.13998130e-01 -2.23695040e-01 8.06188360e-02
5.97897351e-01 -1.50443721e+00 4.52673137e-01 8.13177645e-01
3.95486206e-01 5.43830693e-01 5.10569453e-01 -6.96940601e-01
-1.02876890e+00 -1.45297551e+00 5.99149942e-01 -5.53428113e-01
6.71052575e-01 -3.69756490e-01 -1.06275117e+00 7.95864344e-01
-1.55567527e-01 6.32089078e-01 6.05478466e-01 -6.40821643e-03
-2.17597961e-01 -1.80931196e-01 -1.12307358e+00 3.61367255e-01
1.25753009e+00 -4.73002523e-01 -5.96828341e-01 4.82262671e-01
7.31955588e-01 -5.18219292e-01 -7.81066835e-01 7.44346976e-01
3.02572429e-01 -1.01045573e+00 1.19037640e+00 -4.86526638e-01
3.82164329e-01 -7.60408521e-01 -1.39492124e-01 -9.78172362e-01
-2.89560901e-03 -4.61102724e-01 -5.56327105e-02 1.42418849e+00
1.11476861e-01 -5.58280826e-01 1.02536845e+00 3.02004665e-01
-3.19738805e-01 -6.27061486e-01 -9.03855920e-01 -1.18441737e+00
6.60311133e-02 -6.73724234e-01 5.34657478e-01 1.11988759e+00
-3.80472213e-01 1.14634439e-01 1.96753100e-01 3.87334108e-01
9.01484489e-01 5.31985700e-01 1.04130185e+00 -1.78479719e+00
7.22569302e-02 -3.96648794e-01 -5.54251552e-01 -1.00572586e+00
4.56917405e-01 -8.22321475e-01 1.37038022e-01 -1.32448375e+00
-1.05856240e-01 -1.15090466e+00 -8.76598284e-02 5.90290368e-01
-3.02485600e-02 3.09805334e-01 3.55008185e-01 6.32832527e-01
-8.08382452e-01 3.86776865e-01 6.55809641e-01 -4.45882864e-02
-3.69395852e-01 1.01796895e-01 -2.13637084e-01 1.06249070e+00
8.07508409e-01 -5.54154098e-01 -3.99206817e-01 -1.77734703e-01
-1.53439894e-01 -3.18358123e-01 6.28296673e-01 -1.23025882e+00
4.33291793e-01 -2.98917413e-01 4.46115695e-02 -1.17067492e+00
4.03998643e-01 -1.10476947e+00 -6.74116164e-02 3.06270689e-01
-8.22060257e-02 8.34875181e-02 2.94725418e-01 7.57867038e-01
-2.37745479e-01 -3.86132463e-03 9.14486289e-01 -1.66681394e-01
-9.12921071e-01 2.64557391e-01 3.08782369e-01 -3.16187106e-02
1.32825732e+00 -4.95565146e-01 7.84138367e-02 1.82097889e-02
-5.03887475e-01 4.43238616e-01 8.98326218e-01 4.11592394e-01
4.56995159e-01 -1.22657359e+00 -2.44576588e-01 1.21569417e-01
4.90393072e-01 5.85073233e-01 -2.72684723e-01 9.64634418e-01
-3.59072059e-01 2.36172676e-01 1.93440303e-01 -1.47762287e+00
-1.30900431e+00 5.58880389e-01 1.73105940e-01 -1.20597020e-01
-7.60303259e-01 5.14856756e-01 9.17892605e-02 -6.87961757e-01
2.01448500e-01 -4.17631447e-01 -9.00618508e-02 1.05738878e-01
-9.93493199e-02 2.57874370e-01 2.26918295e-01 -7.42380798e-01
-5.04954815e-01 9.27787960e-01 9.07070041e-02 -7.93635324e-02
1.40326393e+00 5.55567704e-02 -1.03998885e-01 7.75523067e-01
1.00357437e+00 -7.25068897e-02 -1.49272013e+00 -5.74360430e-01
6.70160055e-01 -6.38668358e-01 1.39610553e-02 -4.14499015e-01
-1.12832201e+00 7.87184060e-01 7.72297382e-01 2.93453783e-01
9.30465758e-01 2.54061937e-01 8.25701773e-01 2.43140280e-01
5.67531884e-01 -1.07111871e+00 9.27178264e-02 3.03254664e-01
3.45781952e-01 -1.61004961e+00 5.81892096e-02 -8.59762192e-01
-3.69533181e-01 9.46699321e-01 5.33030689e-01 -2.42115110e-01
7.80423820e-01 2.79859781e-01 -1.25324894e-02 -4.28517312e-01
-3.02901536e-01 -6.69154048e-01 2.53736675e-01 7.18237281e-01
1.92152768e-01 -2.66431272e-02 -1.44979224e-01 1.25161767e-01
-1.53211936e-01 1.97914213e-01 7.04970211e-02 1.44457448e+00
-3.35208356e-01 -1.31882870e+00 -4.69329894e-01 3.77142459e-01
9.00172889e-02 1.52966827e-01 -4.94342774e-01 1.00739193e+00
3.48775625e-01 7.44491875e-01 2.74961114e-01 -1.93817884e-01
3.35331023e-01 -3.98876593e-02 3.36985767e-01 -8.67925704e-01
-4.19075340e-01 1.53098300e-01 -3.69136691e-01 -6.03218973e-01
-9.87621844e-01 -1.00274813e+00 -1.50141084e+00 1.75488904e-01
-2.88943619e-01 -3.82466353e-02 6.86317265e-01 1.02737617e+00
3.90258789e-01 2.61608422e-01 6.01704180e-01 -1.33601880e+00
-2.94714689e-01 -5.34006655e-01 -4.16172296e-01 5.80575526e-01
3.42126429e-01 -9.48493779e-01 -2.10299969e-01 1.92769736e-01] | [8.043432235717773, -2.953274965286255] |
8491e467-23b7-473f-9c96-45be4e4f0fbc | eben-extreme-bandwidth-extension-network-1 | 2210.1409 | null | https://arxiv.org/abs/2210.14090v2 | https://arxiv.org/pdf/2210.14090v2.pdf | EBEN: Extreme bandwidth extension network applied to speech signals captured with noise-resilient body-conduction microphones | In this paper, we present Extreme Bandwidth Extension Network (EBEN), a Generative Adversarial network (GAN) that enhances audio measured with body-conduction microphones. This type of capture equipment suppresses ambient noise at the expense of speech bandwidth, thereby requiring signal enhancement techniques to recover the wideband speech signal. EBEN leverages a multiband decomposition of the raw captured speech to decrease the data time-domain dimensions, and give better control over the full-band signal. This multiband representation is fed to a U-Net-like model, which adopts a combination of feature and adversarial losses to recover an enhanced audio signal. We also benefit from this original representation in the proposed discriminator architecture. Our approach can achieve state-of-the-art results with a lightweight generator and real-time compatible operation. | ['Éric Bavu', 'Véronique Zimpfer', 'Thomas Joubaud', 'Julien Hauret'] | 2022-10-25 | null | null | null | null | ['bandwidth-extension', 'bandwidth-extension'] | ['audio', 'speech'] | [ 5.22982657e-01 4.91486758e-01 1.85950473e-03 1.19887553e-01
-1.22048569e+00 -5.25524259e-01 5.98412044e-02 -7.34779894e-01
-4.51238714e-02 5.96331537e-01 6.21922910e-01 -1.33841947e-01
3.15947011e-02 -7.87830949e-01 -5.82978427e-01 -8.10002446e-01
-1.57553509e-01 -4.93156910e-01 -2.41611078e-01 -3.10682714e-01
-6.65810764e-01 2.28198186e-01 -1.19413126e+00 4.37211961e-01
6.81295693e-01 1.34004688e+00 -2.78832197e-01 1.03427148e+00
4.14496571e-01 5.93961000e-01 -1.09347200e+00 -4.74507332e-01
4.69582587e-01 -5.66103816e-01 -1.43267378e-01 -3.46222281e-01
2.35605389e-01 -6.24765992e-01 -8.22652638e-01 9.43705738e-01
1.28954816e+00 9.81564298e-02 4.84829575e-01 -1.23328745e+00
-6.26634240e-01 8.06437492e-01 -3.42101485e-01 1.92638785e-02
5.24688423e-01 4.38058197e-01 7.71307111e-01 -5.02547741e-01
9.35997292e-02 9.99157667e-01 1.07557619e+00 9.81198549e-01
-1.33389449e+00 -1.14405918e+00 -3.91613632e-01 -3.08014959e-01
-1.37296164e+00 -7.13364661e-01 1.41421282e+00 1.04802966e-01
6.10430479e-01 6.08192801e-01 6.70617759e-01 1.77739465e+00
-5.25865797e-03 4.98511195e-01 9.05585706e-01 -2.62436807e-01
1.88592151e-01 -5.82263805e-02 -5.90045929e-01 2.17905477e-01
-3.84137422e-01 5.76707661e-01 -7.30909407e-01 -3.76551002e-01
9.10631239e-01 5.52268103e-02 -7.30404496e-01 2.15463564e-01
-7.08144963e-01 5.63071728e-01 3.73607963e-01 2.65374571e-01
-5.91926575e-01 4.67430621e-01 2.85292387e-01 5.18123209e-01
4.93765414e-01 3.39516073e-01 -4.46242541e-02 -3.69090110e-01
-9.80014145e-01 6.25841618e-02 7.90833533e-01 7.70796895e-01
4.32091653e-02 8.90820563e-01 -3.09036791e-01 1.04104805e+00
2.22482607e-01 6.83808684e-01 7.57310629e-01 -1.06482816e+00
4.80621904e-01 -2.74021000e-01 -3.45980883e-01 -8.03813040e-01
-5.16914949e-02 -9.28270102e-01 -1.05620396e+00 1.39976472e-01
1.64974690e-03 -7.42927611e-01 -7.74436116e-01 2.15185142e+00
2.16868401e-01 5.71340919e-01 3.13564628e-01 9.25354004e-01
6.53746367e-01 6.19927704e-01 -1.21836200e-01 -6.97983131e-02
9.38124001e-01 -8.90755475e-01 -9.16692436e-01 -1.89032182e-01
-3.49887490e-01 -5.74497759e-01 9.47255313e-01 4.99579251e-01
-1.41834390e+00 -7.73861468e-01 -1.22282016e+00 1.61311641e-01
1.65068742e-03 -3.18640053e-01 4.28361028e-01 1.24684453e+00
-1.04422259e+00 4.78246391e-01 -5.05162895e-01 4.98829931e-01
1.91169620e-01 5.48259974e-01 -1.88085824e-01 3.18449616e-01
-1.63743913e+00 6.15410246e-02 -9.85686854e-02 1.74755845e-02
-1.04570687e+00 -1.21065319e+00 -8.94937336e-01 2.32447520e-01
1.23262629e-01 -1.07944489e+00 1.21173465e+00 -1.00685585e+00
-2.03431273e+00 2.49201700e-01 4.91528988e-01 -7.03106105e-01
6.63627923e-01 -1.68273345e-01 -1.14705777e+00 3.06961834e-01
-3.84741724e-01 3.52946192e-01 1.69605184e+00 -9.01672482e-01
-2.10641548e-01 -1.08141437e-01 -1.43281937e-01 -9.60189924e-02
-6.29537880e-01 -3.09375256e-01 3.05215865e-02 -1.53170347e+00
-4.07691859e-02 -7.15624094e-01 -1.18167780e-01 -1.54395774e-01
-2.76148885e-01 7.08638012e-01 9.78128791e-01 -1.10195887e+00
1.19072592e+00 -2.51973367e+00 -6.66561816e-03 1.92714006e-01
3.23619485e-01 3.39746475e-01 -1.08408920e-01 3.17847371e-01
-2.23046660e-01 1.27325416e-01 -2.42303908e-01 -4.67576116e-01
1.09021753e-01 -1.52451664e-01 -6.35430396e-01 4.70386773e-01
-4.56795394e-02 6.58961236e-01 -5.25358379e-01 -2.72788089e-02
1.83042452e-01 1.01376939e+00 -9.51364517e-01 5.07008016e-01
1.22873075e-01 5.98481178e-01 -2.54396766e-01 4.93277580e-01
6.80363357e-01 6.72076523e-01 -7.67216384e-02 -2.93228805e-01
4.46130633e-01 2.82378644e-01 -1.15770268e+00 1.99805999e+00
-1.04240787e+00 2.61852980e-01 7.33073533e-01 -7.49160767e-01
1.05465007e+00 8.31663549e-01 4.73083109e-01 -5.86918652e-01
3.25191468e-01 1.80223972e-01 -1.02350846e-01 -4.17349786e-01
-3.41050909e-03 -3.38053942e-01 -3.06516021e-01 3.31844956e-01
3.61613184e-01 -1.27640784e-01 -8.97931814e-01 -5.88350818e-02
1.20961142e+00 -1.03532732e-01 1.62590235e-01 5.09320013e-02
4.74174082e-01 -9.04459298e-01 5.16017556e-01 7.23318160e-01
-1.70442075e-01 6.58523679e-01 -5.91551252e-02 1.07474558e-01
-9.34620321e-01 -1.32720149e+00 2.83138603e-01 9.22620356e-01
-2.74462044e-01 -1.24843039e-01 -9.26798642e-01 -3.40155542e-01
-1.23181082e-02 6.56446755e-01 -4.37979668e-01 -8.35557044e-01
-5.03322363e-01 -6.48837909e-02 1.41153240e+00 6.03838384e-01
5.67921042e-01 -6.72086120e-01 -3.24367464e-01 4.31199431e-01
-1.27865940e-01 -7.53191531e-01 -9.30582702e-01 2.92302579e-01
-7.41833031e-01 -2.56336004e-01 -9.12386298e-01 -5.70923030e-01
1.16480023e-01 -2.65576273e-01 7.48976946e-01 -4.09961671e-01
-1.73053980e-01 5.26371777e-01 -2.02033475e-01 -4.97459888e-01
-5.98636210e-01 -1.90412447e-01 3.32954556e-01 4.06722844e-01
-2.39697725e-01 -1.42089784e+00 -8.73304009e-01 4.21348996e-02
-9.68288839e-01 -4.10577506e-01 4.38886523e-01 1.04619408e+00
2.82646000e-01 2.57263660e-01 1.01290691e+00 -2.64469653e-01
9.22971547e-01 -3.99479389e-01 1.98384728e-02 -2.81036228e-01
-6.15732148e-02 -1.88131362e-01 1.09539509e+00 -9.76513624e-01
-1.16500747e+00 -3.06176543e-01 -7.60263383e-01 -7.28401601e-01
1.06252380e-01 2.83255987e-02 -5.19135594e-01 -2.27924436e-01
5.87267876e-01 3.07388246e-01 1.17803469e-01 -4.20642227e-01
5.92866480e-01 1.31857610e+00 1.21776760e+00 -4.25935715e-01
7.53194749e-01 3.76672477e-01 -3.74614120e-01 -5.85234404e-01
-3.56677175e-01 -1.95495471e-01 2.53982604e-01 -2.02077731e-01
5.43617785e-01 -1.10954654e+00 -6.42761707e-01 5.37187278e-01
-8.21757495e-01 -3.71310055e-01 -6.59107387e-01 6.81263506e-01
-7.74689019e-01 -1.13301918e-01 -9.67640579e-01 -1.07809913e+00
-9.34980869e-01 -7.14920700e-01 9.88334537e-01 5.25712855e-02
-3.28399241e-01 -7.67407835e-01 -1.34230657e-02 4.12188053e-01
7.20268786e-01 5.35261869e-01 4.29432631e-01 -3.26318532e-01
-1.45388529e-01 -2.69317895e-01 7.45496809e-01 6.90911293e-01
1.00179143e-01 -4.67599779e-01 -1.52564037e+00 -5.70799828e-01
7.55346656e-01 -1.76390722e-01 6.57601476e-01 4.05979931e-01
1.21074820e+00 -6.35813594e-01 2.19216198e-02 1.22843552e+00
1.08484089e+00 2.29022071e-01 7.62103796e-01 -1.97269037e-01
3.96556884e-01 1.08198456e-01 1.29158944e-01 4.40273732e-01
-4.54601347e-01 5.32119870e-01 4.05870587e-01 -4.15803313e-01
-6.72371805e-01 -5.54230511e-01 6.02128267e-01 9.30426836e-01
1.87845886e-01 -1.69066697e-01 -3.35122555e-01 4.50503528e-01
-1.14021361e+00 -1.10257614e+00 5.65430343e-01 1.92025733e+00
9.44332004e-01 2.41035298e-02 2.77957529e-01 6.57837093e-01
8.18687201e-01 2.72413462e-01 -6.71436131e-01 -2.96495080e-01
1.41856536e-01 6.25552595e-01 5.16950727e-01 3.65891218e-01
-7.07153022e-01 2.34942943e-01 6.62843752e+00 1.08580530e+00
-1.49692094e+00 4.54874814e-01 4.12401170e-01 -4.81953770e-01
-3.73879403e-01 -7.91066468e-01 -1.35163218e-01 5.73192120e-01
1.30328000e+00 -3.17919552e-01 9.06285942e-01 9.76863205e-01
3.77973616e-02 8.40028942e-01 -9.56374824e-01 1.08261347e+00
7.95384943e-02 -8.64105999e-01 -2.05363974e-01 4.15424034e-02
4.43794310e-01 -3.81367922e-01 7.66388237e-01 5.20138681e-01
9.80310962e-02 -1.13526416e+00 8.80074143e-01 4.09660399e-01
1.47788513e+00 -1.15561986e+00 4.27140743e-01 4.23329085e-01
-1.08752286e+00 -4.73920196e-01 -1.17513025e-02 -3.23437825e-02
8.27999786e-02 7.17445672e-01 -7.86425829e-01 4.85886753e-01
5.41409552e-01 -1.76513270e-01 2.13042796e-01 5.57519436e-01
-1.32639647e-01 1.03869402e+00 -3.85911077e-01 4.06173974e-01
-7.47135729e-02 1.68593243e-01 9.29731131e-01 9.91668701e-01
5.22147417e-01 1.03717640e-01 -1.41824976e-01 1.03427017e+00
-4.69806403e-01 -4.33682084e-01 -7.16758668e-01 2.24910781e-01
6.79126322e-01 1.02748585e+00 7.17560425e-02 7.81360269e-02
-2.24637881e-01 9.51874077e-01 -1.98028475e-01 3.50331187e-01
-1.18389225e+00 -8.15428257e-01 8.55527341e-01 4.98574926e-04
1.87860191e-01 1.08612530e-01 -9.59522575e-02 -7.24764764e-01
7.92531669e-02 -1.37916017e+00 1.74822375e-01 -6.78739607e-01
-1.19208741e+00 6.35758102e-01 -5.99164844e-01 -1.32449245e+00
-5.33958793e-01 2.22164933e-02 -6.06662512e-01 1.16530609e+00
-1.05802894e+00 -1.12666273e+00 -8.01003575e-02 9.88952816e-01
3.29419315e-01 -2.18942091e-01 1.06850123e+00 5.73339820e-01
-9.94174480e-02 1.14218557e+00 8.17472190e-02 2.59809643e-01
5.57320833e-01 -1.13141978e+00 5.79904795e-01 8.21090043e-01
-7.99598247e-02 6.19490027e-01 7.00236201e-01 -3.70340019e-01
-1.44229841e+00 -1.40214920e+00 5.29635847e-02 3.61357555e-02
4.24417704e-01 -6.40840471e-01 -5.71037710e-01 3.99417430e-01
3.40318739e-01 1.19452424e-01 1.09348905e+00 -4.27015156e-01
-3.80605072e-01 -4.96681064e-01 -1.71362805e+00 5.80252171e-01
9.81544554e-01 -1.09190214e+00 -7.19572723e-01 -2.47333601e-01
1.28821516e+00 -5.14263570e-01 -1.19594705e+00 1.58144996e-01
8.21646750e-01 -7.79969931e-01 1.28543937e+00 -3.85422140e-01
4.14146096e-01 4.28355150e-02 -2.76817918e-01 -1.79349780e+00
-2.03243017e-01 -1.43398595e+00 -4.88881499e-01 1.56729102e+00
2.92513788e-01 -7.34169424e-01 6.34578764e-01 2.43294507e-01
-2.31827810e-01 -6.04483843e-01 -1.13628197e+00 -9.99408722e-01
-7.64879882e-02 -7.55426288e-01 1.02867186e+00 6.66238308e-01
2.23091125e-01 1.36202037e-01 -7.87132382e-01 3.70049894e-01
6.84044600e-01 -4.35099214e-01 4.64076966e-01 -5.99293888e-01
-9.17645156e-01 -6.89416528e-02 -3.96709144e-01 -9.75681245e-01
1.00503247e-02 -6.66009545e-01 -4.52869423e-02 -6.92866266e-01
-4.99097496e-01 -1.27324194e-01 -4.97955054e-01 2.74422709e-05
-7.35905766e-03 6.65533781e-01 2.30967194e-01 -3.61190647e-01
3.34151462e-02 6.69044197e-01 1.07907224e+00 -4.75363880e-01
-2.82747805e-01 1.99858919e-01 -8.44028473e-01 7.24710286e-01
7.72999883e-01 -5.77264309e-01 -5.90314865e-01 -2.35583514e-01
-4.19556409e-01 6.64059758e-01 4.54015344e-01 -1.43186271e+00
2.22167939e-01 4.30202454e-01 3.80139321e-01 1.95978750e-02
7.87667871e-01 -1.19652319e+00 5.27547836e-01 6.07980192e-01
-5.62116146e-01 -4.85332936e-01 1.01323962e-01 7.23025560e-01
-3.85238409e-01 2.69564003e-01 9.04829800e-01 1.44417822e-01
1.93166301e-01 2.88412213e-01 -1.67056277e-01 6.20286688e-02
8.56685519e-01 -3.46453898e-02 -1.65049270e-01 -1.00433803e+00
-8.86537254e-01 -4.22930062e-01 5.08462824e-02 3.12554330e-01
5.75817227e-01 -1.70189345e+00 -5.82040310e-01 5.60697675e-01
-4.43391263e-01 -2.01185495e-01 5.90290189e-01 5.04179299e-01
-7.17872381e-02 9.76637602e-02 -1.26399502e-01 -3.02450120e-01
-1.09205925e+00 5.49699545e-01 6.44141614e-01 -1.80504501e-01
-7.70316839e-01 8.84946108e-01 -6.64891973e-02 -1.91345185e-01
4.73686069e-01 6.00201115e-02 1.03761986e-01 -3.61174315e-01
7.24720895e-01 5.17310679e-01 1.70855448e-01 -2.34268755e-01
-1.23577856e-01 2.20110893e-01 5.97204030e-01 -5.90862691e-01
1.23649442e+00 -3.44410017e-02 3.49502802e-01 4.03377376e-02
1.38236022e+00 5.95491469e-01 -1.25715327e+00 -5.51560894e-03
-8.93189013e-01 -4.19727266e-01 4.84368294e-01 -8.87476444e-01
-1.58024621e+00 5.77576220e-01 7.36602545e-01 3.54444712e-01
1.70750344e+00 -3.53189111e-01 1.45466971e+00 -1.72292039e-01
3.58193189e-01 -7.96857238e-01 1.58629999e-01 -1.30441830e-01
9.48014319e-01 -2.97333360e-01 -4.69911486e-01 -2.93747008e-01
-2.17080846e-01 7.67065883e-01 1.34499177e-01 -2.43884474e-01
5.99229276e-01 9.68441010e-01 1.89163759e-01 3.48671138e-01
-4.59038258e-01 3.72879356e-01 -2.09641550e-02 1.25182152e+00
2.18172073e-02 1.45496368e-01 9.16180089e-02 1.46013439e+00
-6.83074534e-01 -1.24302611e-01 2.74429590e-01 5.56990147e-01
1.24966558e-02 -8.67695630e-01 -8.05938840e-01 3.77312303e-01
-1.14126396e+00 -2.31915087e-01 -2.20275313e-01 3.08043391e-01
1.72239378e-01 1.24222684e+00 -5.63660190e-02 -7.23253965e-01
5.08589447e-01 4.98234667e-02 3.02850157e-01 -4.33278382e-02
-9.24953818e-01 4.09879029e-01 6.05026353e-03 -5.73600531e-01
-1.73150338e-02 -2.87234336e-01 -8.31413388e-01 -4.38434899e-01
-3.26043427e-01 1.72069237e-01 6.66462481e-01 3.09347570e-01
4.33665752e-01 1.20980072e+00 1.09826124e+00 -7.92251110e-01
-1.19326723e+00 -9.13440347e-01 -9.01440203e-01 4.11790639e-01
9.10775304e-01 -1.17124967e-01 -6.12447381e-01 -1.13097979e-02] | [15.35477352142334, 6.078392028808594] |
bef2e5d2-07a1-4b41-bc37-081304740e51 | a-comparison-of-deep-saliency-map-generators | 2108.11767 | null | https://arxiv.org/abs/2108.11767v1 | https://arxiv.org/pdf/2108.11767v1.pdf | A Comparison of Deep Saliency Map Generators on Multispectral Data in Object Detection | Deep neural networks, especially convolutional deep neural networks, are state-of-the-art methods to classify, segment or even generate images, movies, or sounds. However, these methods lack of a good semantic understanding of what happens internally. The question, why a COVID-19 detector has classified a stack of lung-ct images as positive, is sometimes more interesting than the overall specificity and sensitivity. Especially when human domain expert knowledge disagrees with the given output. This way, human domain experts could also be advised to reconsider their choice, regarding the information pointed out by the system. In addition, the deep learning model can be controlled, and a present dataset bias can be found. Currently, most explainable AI methods in the computer vision domain are purely used on image classification, where the images are ordinary images in the visible spectrum. As a result, there is no comparison on how the methods behave with multimodal image data, as well as most methods have not been investigated on how they behave when used for object detection. This work tries to close the gaps. Firstly, investigating three saliency map generator methods on how their maps differ across the different spectra. This is achieved via accurate and systematic training. Secondly, we examine how they behave when used for object detection. As a practical problem, we chose object detection in the infrared and visual spectrum for autonomous driving. The dataset used in this work is the Multispectral Object Detection Dataset, where each scene is available in the FIR, MIR and NIR as well as visual spectrum. The results show that there are differences between the infrared and visual activation maps. Further, an advanced training with both, the infrared and visual data not only improves the network's output, it also leads to more focused spots in the saliency maps. | ['Michael Arens', 'David Münch', 'Jens Bayer'] | 2021-08-26 | null | null | null | null | ['multispectral-object-detection'] | ['computer-vision'] | [ 4.73080248e-01 2.35365741e-02 -1.49261728e-01 -1.02563925e-01
-9.61712152e-02 -6.20079875e-01 4.96478379e-01 -8.06140453e-02
-3.03739667e-01 7.30325162e-01 -1.64908066e-01 -4.26784992e-01
-2.70028800e-01 -9.82939839e-01 -5.59339285e-01 -8.59263122e-01
5.48803449e-01 2.38419652e-01 6.61764324e-01 -4.03777599e-01
2.26894021e-01 4.91966248e-01 -2.05024648e+00 5.38950205e-01
7.99397051e-01 9.45686936e-01 5.15485406e-01 4.38281626e-01
-2.17234448e-01 1.02683175e+00 -7.47879863e-01 -1.82433560e-01
1.72010407e-01 -7.01133788e-01 -7.73108244e-01 4.87721637e-02
2.61291623e-01 -1.47489980e-01 1.38324108e-02 1.38086545e+00
5.73206186e-01 -1.50946438e-01 7.35368133e-01 -1.35409009e+00
-6.98499799e-01 4.98187155e-01 -3.81723255e-01 3.45239371e-01
2.19219640e-01 4.11478072e-01 5.74269593e-01 -5.92491090e-01
5.63622653e-01 1.13113678e+00 4.16019827e-01 5.58612287e-01
-1.07303238e+00 -5.26539564e-01 -1.39548779e-01 4.72634017e-01
-1.11145341e+00 5.16664162e-02 1.00586367e+00 -6.26920462e-01
4.08151001e-01 4.97848928e-01 7.43092239e-01 1.41211224e+00
1.42848715e-01 4.10549045e-01 1.56647551e+00 -6.51437104e-01
2.25848019e-01 7.76317716e-01 -1.37981519e-01 4.27208513e-01
3.46333832e-01 4.32120174e-01 -3.51040155e-01 2.65793204e-01
4.94531125e-01 -1.69777021e-01 -3.70139897e-01 -1.91488370e-01
-1.19028914e+00 7.65898228e-01 7.40342379e-01 7.39595890e-01
-3.23983133e-01 -5.02642766e-02 1.05467707e-01 1.94312572e-01
2.39697963e-01 6.82011485e-01 -2.54903674e-01 4.80791658e-01
-8.27354193e-01 9.50806886e-02 3.60676378e-01 5.00893295e-01
7.13085711e-01 1.77262723e-01 -2.69799441e-01 7.06727862e-01
2.26892650e-01 5.77982426e-01 6.87194586e-01 -7.44430184e-01
3.06390077e-02 6.66909456e-01 1.32885473e-02 -1.13913524e+00
-6.95361972e-01 -4.24441099e-01 -6.99751496e-01 9.05323327e-01
5.95307887e-01 -2.54967302e-01 -1.05641854e+00 1.54810905e+00
1.22225709e-01 -1.21374473e-01 2.56607026e-01 1.36438215e+00
1.26224041e+00 4.64237511e-01 8.46655145e-02 2.48220526e-02
1.71071637e+00 -7.19379723e-01 -5.92727065e-01 -6.06092215e-01
1.72363907e-01 -8.53096604e-01 9.53471124e-01 4.52424973e-01
-7.21204281e-01 -9.48415160e-01 -1.10822988e+00 2.21951813e-01
-7.63776779e-01 4.96475041e-01 5.12754798e-01 7.00153589e-01
-8.88955355e-01 5.35388052e-01 -3.46291035e-01 -6.44517004e-01
1.87414184e-01 8.74045417e-02 -1.27392948e-01 1.87202021e-01
-1.53082514e+00 1.28915119e+00 8.42085958e-01 3.84620428e-02
-6.99729681e-01 -4.23299223e-01 -5.45879841e-01 1.18933670e-01
3.74193251e-01 -5.74556053e-01 1.07384491e+00 -1.72482538e+00
-1.26134562e+00 1.09011221e+00 2.66337365e-01 -4.86862779e-01
5.96725821e-01 2.84544706e-01 -5.84983647e-01 2.61427850e-01
1.65967864e-03 8.35467041e-01 8.49364281e-01 -1.49099720e+00
-6.51848018e-01 -8.65895972e-02 3.07653725e-01 1.20021947e-01
6.94863200e-02 2.03391254e-01 9.40449089e-02 -5.39927542e-01
-1.53372204e-02 -9.05640185e-01 -1.88176826e-01 -4.77056243e-02
-5.81435978e-01 -1.21690974e-01 6.69509530e-01 -3.25170577e-01
8.34961772e-01 -2.18177152e+00 -3.04608017e-01 2.64082670e-01
-6.84568658e-02 3.26724380e-01 5.67471087e-02 1.08380876e-01
-4.43592280e-01 2.79782236e-01 -2.77576268e-01 5.33079326e-01
-2.14336216e-01 -3.79097499e-02 -3.23262453e-01 2.26022959e-01
3.82431388e-01 5.15717268e-01 -8.07709575e-01 -4.89639193e-01
3.97306204e-01 3.56243938e-01 -8.83404389e-02 -4.47348255e-04
-1.42135099e-01 5.13914108e-01 -4.96931195e-01 4.24209416e-01
6.85276866e-01 -1.35876864e-01 1.04140751e-02 -5.33415139e-01
-2.64049798e-01 5.00721019e-03 -1.02717781e+00 1.29345012e+00
-1.08521782e-01 9.93734181e-01 -2.79299498e-01 -1.15667796e+00
1.07802570e+00 2.99359590e-01 3.71225327e-01 -9.69007850e-01
2.30409175e-01 4.41668481e-01 3.31500560e-01 -7.54253089e-01
2.75049567e-01 -3.22037905e-01 3.79415929e-01 2.19668120e-01
-1.45911306e-01 -3.96285832e-01 1.15536369e-01 -2.15864196e-01
6.93683624e-01 8.62973705e-02 1.43648148e-01 -1.62530273e-01
6.23352826e-01 4.49226588e-01 2.31022656e-01 7.66559541e-01
-2.46407315e-02 9.71767187e-01 6.12662017e-01 -4.02397841e-01
-8.95228624e-01 -8.59551251e-01 -3.40861022e-01 9.30506229e-01
5.12523770e-01 2.93675691e-01 -9.53357637e-01 -4.71040368e-01
-3.00565034e-01 9.83366370e-01 -5.80884039e-01 -3.83534193e-01
-2.14378029e-01 -8.57832849e-01 2.95369744e-01 2.32923180e-01
6.39633119e-01 -1.70367777e+00 -1.20377803e+00 -2.26630680e-02
-2.14729398e-01 -9.86856163e-01 2.63693213e-01 3.55723768e-01
-6.68563604e-01 -1.16298413e+00 -7.44216561e-01 -6.50535583e-01
5.98423898e-01 3.93481910e-01 1.10947239e+00 1.75123155e-01
-3.56306881e-01 1.56219572e-01 -3.99924725e-01 -8.75087261e-01
-6.65957391e-01 -3.03040240e-02 -2.94583559e-01 2.41279285e-02
4.37276423e-01 -1.72090620e-01 -5.92115402e-01 4.97534603e-01
-1.07314551e+00 2.21775785e-01 7.64044285e-01 6.23262465e-01
3.39151740e-01 1.83503956e-01 5.35211086e-01 -7.84529626e-01
5.56596160e-01 -3.56767565e-01 -4.00905430e-01 2.93092906e-01
-5.71747065e-01 5.67811690e-02 3.97806734e-01 -3.69470477e-01
-1.14472687e+00 1.37775019e-01 -6.57480508e-02 -3.09083521e-01
-6.58382595e-01 3.74343097e-01 1.41122133e-01 4.58725430e-02
1.43002307e+00 1.23132527e-01 1.12249099e-01 -1.26524940e-01
7.81951249e-02 7.21774518e-01 4.73072022e-01 -1.17015325e-01
8.62338185e-01 4.56625611e-01 -8.40114504e-02 -6.67806327e-01
-9.01951849e-01 -2.79005975e-01 -4.83514518e-01 -5.48479378e-01
1.28328443e+00 -6.60827994e-01 -4.00270283e-01 2.85783321e-01
-1.29964983e+00 -2.42191717e-01 -3.04542810e-01 5.70780337e-01
-4.96889949e-01 1.69494096e-02 -9.48520973e-02 -9.04398382e-01
-9.38759893e-02 -1.27042949e+00 9.10233140e-01 4.90479797e-01
-1.29717231e-01 -7.72197366e-01 -2.30962947e-01 3.15229028e-01
5.04180491e-01 3.26736361e-01 9.49002624e-01 -4.99000192e-01
-4.57566679e-01 7.87439290e-03 -5.40048778e-01 4.25992310e-01
1.17306799e-01 1.44082919e-01 -1.29436409e+00 1.31287292e-01
1.53850257e-01 -2.26675436e-01 9.20802951e-01 4.35389429e-01
1.29775226e+00 -7.52311051e-02 -2.52972305e-01 1.62823930e-01
1.38108969e+00 4.78580773e-01 9.47444439e-01 7.21256018e-01
4.70583916e-01 1.12841284e+00 7.27585196e-01 -7.58273378e-02
-1.79329708e-01 8.38473976e-01 8.30672503e-01 -6.04874849e-01
-4.41834360e-01 9.65679288e-02 3.76534969e-01 -1.29586235e-02
-2.75305599e-01 -1.97750390e-01 -8.53316247e-01 3.52172345e-01
-1.70814979e+00 -1.34243941e+00 -5.19475162e-01 2.18366313e+00
7.13827193e-01 1.88761652e-01 1.87822357e-01 3.92559528e-01
9.05711889e-01 2.10708249e-02 -4.06035662e-01 -3.71323884e-01
-3.68982971e-01 1.22809015e-01 6.80503130e-01 8.31864476e-02
-1.17978323e+00 5.81361949e-01 6.06568050e+00 7.90267944e-01
-1.63453794e+00 -2.11260598e-02 8.04032981e-01 1.07989319e-01
-3.55112493e-01 -1.92832172e-01 -3.55670899e-01 6.17311358e-01
8.29000950e-01 1.62688583e-01 2.51647919e-01 9.14574623e-01
2.87997127e-01 -4.27715510e-01 -8.26117039e-01 1.02742696e+00
1.58060521e-01 -1.03450406e+00 -1.21503100e-01 -2.59892702e-01
5.60864806e-01 -4.35702018e-02 2.12333322e-01 5.33746220e-02
-1.37516618e-01 -1.27914977e+00 1.14716864e+00 6.68388307e-01
5.30736744e-01 -6.11083388e-01 9.34559882e-01 3.31091702e-01
-6.90032899e-01 -2.11289093e-01 -5.05478382e-01 2.28736892e-01
-1.46084964e-01 4.60043907e-01 -8.53289247e-01 4.51651335e-01
9.50673163e-01 1.99979499e-01 -9.24974620e-01 1.17217124e+00
-4.95384336e-01 6.27862811e-01 -7.69348294e-02 -2.20333099e-01
1.80969059e-01 -1.12223305e-01 4.65092987e-01 1.22817028e+00
4.56545889e-01 -1.26960576e-01 1.96218770e-03 1.34073424e+00
5.23918033e-01 1.33152921e-02 -8.80381227e-01 2.36393571e-01
5.75964488e-02 1.36642742e+00 -9.10238564e-01 -2.90216565e-01
-2.89933592e-01 6.40974879e-01 -3.75160903e-01 6.01777554e-01
-1.03739130e+00 -4.01129574e-01 2.54959732e-01 3.07548523e-01
8.66307318e-02 3.55701834e-01 -4.96350616e-01 -7.00219810e-01
-1.94948222e-02 -8.75152946e-01 1.23616040e-01 -1.47091603e+00
-1.13799620e+00 7.13842154e-01 1.55299515e-01 -1.50814676e+00
-2.01930135e-01 -1.00788236e+00 -6.74314380e-01 8.59622419e-01
-1.60899615e+00 -8.85836422e-01 -6.54207051e-01 4.60252106e-01
3.91722828e-01 -8.14255774e-02 4.27776039e-01 1.99628830e-01
-3.11673254e-01 4.46409956e-02 -2.33041123e-01 1.60993397e-01
5.70706248e-01 -1.16105175e+00 -2.61565596e-01 6.60194218e-01
2.27012157e-01 4.63260487e-02 9.69582558e-01 -2.66594350e-01
-7.11938858e-01 -9.36679542e-01 4.17558432e-01 -4.47719842e-01
4.74354088e-01 1.75100826e-02 -1.07318556e+00 8.19756538e-02
4.26105529e-01 -1.12331174e-01 2.52656430e-01 -3.79105568e-01
-1.40515557e-02 -2.61437923e-01 -1.18011749e+00 5.10812163e-01
7.01171637e-01 -4.40255314e-01 -6.79368854e-01 5.63103497e-01
5.07148087e-01 -3.59994262e-01 -3.36073667e-01 4.76270407e-01
3.27859789e-01 -1.43071651e+00 1.10348821e+00 -5.90708137e-01
6.28955841e-01 -5.53192496e-01 1.72393933e-01 -1.30985856e+00
-3.25022697e-01 1.84677258e-01 5.94869614e-01 1.11971176e+00
7.71270514e-01 -6.28108382e-01 6.96202040e-01 2.60370761e-01
-2.76310798e-02 -4.67641503e-01 -7.43701994e-01 -6.33112431e-01
-1.33820817e-01 -4.45492834e-01 3.38868678e-01 9.85628903e-01
-5.03386676e-01 3.22236687e-01 -1.18698724e-01 1.34432301e-01
2.51365960e-01 1.06510013e-01 5.68312287e-01 -1.45289600e+00
-1.54850289e-01 -7.28213966e-01 -5.32988369e-01 -1.72076553e-01
-3.80736776e-02 -8.06593120e-01 1.44998252e-01 -1.68895972e+00
1.00158006e-01 -2.18830556e-01 -3.06758404e-01 5.73248506e-01
-1.64792493e-01 3.61135453e-01 3.58912766e-01 1.81994706e-01
-5.11128716e-02 7.25116804e-02 1.40633225e+00 -3.89703125e-01
-1.21321782e-01 1.34919174e-02 -7.01048791e-01 8.31158400e-01
1.15457046e+00 -4.55924809e-01 -4.55107540e-01 -1.33928522e-01
3.46834600e-01 -2.71002233e-01 8.65591943e-01 -1.26464462e+00
2.17039045e-02 -3.73184025e-01 4.89398748e-01 -3.62711281e-01
1.49535790e-01 -1.06313324e+00 3.15472811e-01 5.54223478e-01
-3.07838529e-01 -1.74018040e-01 2.53462881e-01 2.62951702e-01
-3.44818383e-01 -7.06883013e-01 1.12794447e+00 -4.13132906e-01
-9.17122424e-01 -2.34786212e-01 -5.28575718e-01 -1.63320556e-01
1.13915193e+00 -6.79644704e-01 -5.72434127e-01 -3.25540215e-01
-5.20729125e-01 -1.53705582e-01 3.02173734e-01 5.06397903e-01
3.99817914e-01 -1.28206980e+00 -6.37470484e-01 -6.78812414e-02
1.86706066e-01 -9.58701372e-02 2.61969477e-01 1.04017961e+00
-5.75371861e-01 2.48310626e-01 -5.08256316e-01 -7.63540268e-01
-1.18947029e+00 8.14751089e-01 7.25309968e-01 1.79277703e-01
-2.63182640e-01 5.22183776e-01 3.83991629e-01 -2.74580836e-01
-4.34053317e-02 -4.73984033e-01 -7.85256803e-01 3.08706790e-01
3.29752922e-01 1.23998880e-01 3.77215482e-02 -6.23513460e-01
-2.82966971e-01 7.22650528e-01 4.11972404e-01 -2.52133068e-02
9.43976521e-01 1.42471790e-01 -1.23365723e-01 5.43671310e-01
7.59354413e-01 -7.81663433e-02 -7.53303170e-01 2.03276262e-01
-1.18254438e-01 -3.27110797e-01 1.51096908e-02 -1.05464840e+00
-1.22726858e+00 9.86231446e-01 1.06150770e+00 8.68664801e-01
1.34482241e+00 5.55750318e-02 4.14060950e-02 2.01069996e-01
9.89863053e-02 -1.23835611e+00 1.49165630e-01 2.20572829e-01
1.14052188e+00 -1.50621736e+00 -1.14483170e-01 -4.58804309e-01
-9.99727428e-01 1.33859038e+00 7.34764278e-01 1.15370311e-01
2.00484782e-01 -1.27363726e-01 3.46847504e-01 -3.05931985e-01
-3.65316272e-01 -6.38661683e-01 5.76646686e-01 8.16622376e-01
5.41609704e-01 1.89659894e-02 -2.71515697e-01 3.57192963e-01
-3.13491404e-01 -7.97999576e-02 6.82543397e-01 4.30648983e-01
-7.34783053e-01 -7.40828156e-01 -8.32991838e-01 3.99831057e-01
-4.67164159e-01 7.12169781e-02 -7.48756230e-01 8.19658756e-01
5.93864620e-01 1.10266900e+00 1.21965865e-02 -3.79679918e-01
3.49321395e-01 -3.67311950e-05 2.08226368e-01 -6.30367219e-01
-5.16587675e-01 -2.12647468e-01 7.06380904e-02 -2.79661000e-01
-7.45510161e-01 -3.63300413e-01 -1.24027622e+00 1.18306965e-01
-4.78476107e-01 -1.17453160e-02 8.80861461e-01 8.23567092e-01
-8.39478001e-02 8.82138312e-01 5.69991231e-01 -8.34540129e-01
-3.63231540e-01 -9.38654006e-01 -3.82669121e-01 5.56003332e-01
1.88680589e-01 -8.55379283e-01 -6.02447867e-01 -4.86835092e-02] | [10.004846572875977, 1.8953365087509155] |
ad71ca70-1d96-4a81-8be8-beea89a2e643 | stress-rules-from-surface-forms-experiments | null | null | https://aclanthology.org/2021.icon-main.76 | https://aclanthology.org/2021.icon-main.76.pdf | Stress Rules from Surface Forms: Experiments with Program Synthesis | Learning linguistic generalizations from only a few examples is a challenging task. Recent work has shown that program synthesis – a method to learn rules from data in the form of programs in a domain-specific language – can be used to learn phonological rules in highly data-constrained settings. In this paper, we use the problem of phonological stress placement as a case to study how the design of the domain-specific language influences the generalization ability when using the same learning algorithm. We find that encoding the distinction between consonants and vowels results in much better performance, and providing syllable-level information further improves generalization. Program synthesis, thus, provides a way to investigate how access to explicit linguistic information influences what can be learnt from a small number of examples. | ['Dipti Sharma', 'Monojit Choudhury', 'Partho Sarthi', 'Saujas Vaduguru'] | null | null | null | null | icon-2021-12 | ['program-synthesis'] | ['computer-code'] | [ 3.04139763e-01 -1.15936190e-01 -5.52156925e-01 -6.88852251e-01
-5.71127415e-01 -8.52013111e-01 4.29106474e-01 3.81475210e-01
-6.07837915e-01 6.03112221e-01 2.65614390e-01 -8.74322474e-01
1.52341828e-01 -8.62459123e-01 -9.17344689e-01 -3.31979305e-01
2.73535401e-03 2.44032487e-01 2.73488998e-01 -2.85605401e-01
4.91107315e-01 5.03783524e-01 -1.87859988e+00 4.82484818e-01
1.04625261e+00 3.93455416e-01 5.95555127e-01 6.89694405e-01
-3.55318218e-01 3.67988199e-01 -6.33309662e-01 -1.39021844e-01
2.18197018e-01 -5.60874462e-01 -6.31748259e-01 -5.80935143e-02
6.75984323e-01 1.56360250e-02 1.33469805e-01 9.17381465e-01
3.25056672e-01 4.68295813e-01 5.68892837e-01 -5.30432045e-01
-8.77862096e-01 9.55084264e-01 4.33249176e-02 4.34971541e-01
3.94267738e-01 3.79495770e-01 1.12439561e+00 -8.35982442e-01
5.99681675e-01 1.27503216e+00 6.61102474e-01 6.62527382e-01
-1.73371387e+00 -2.94133246e-01 2.54879981e-01 -8.87005031e-02
-1.24718237e+00 -3.72213215e-01 7.91970909e-01 -4.82022792e-01
1.19957447e+00 2.98027843e-01 5.90753019e-01 6.60367727e-01
-1.64480712e-02 4.88231272e-01 1.27460206e+00 -8.47257018e-01
2.30955645e-01 2.60902047e-01 5.10384381e-01 7.11382210e-01
6.35658085e-01 4.51327115e-01 -5.63622057e-01 -1.84773266e-01
5.80741167e-01 -3.91741097e-01 -2.48616770e-01 -2.04726398e-01
-8.30231547e-01 9.98758793e-01 4.44581717e-01 5.44035912e-01
8.11945125e-02 7.39716813e-02 6.51072562e-01 6.21425271e-01
2.05263011e-02 1.09232116e+00 -7.26718545e-01 -2.28934169e-01
-8.54401410e-01 2.71262109e-01 1.01967764e+00 7.79809654e-01
1.08184290e+00 4.63162094e-01 1.16132021e-01 1.11616707e+00
4.34479564e-02 3.72906834e-01 6.46724343e-01 -8.90066266e-01
4.52206045e-01 2.82988697e-01 -1.85805500e-01 -6.59017861e-01
-2.54676759e-01 -6.60223514e-02 9.79426056e-02 2.99850464e-01
7.99171090e-01 -6.63990319e-01 -8.44060659e-01 2.28143954e+00
-1.65495262e-01 -3.39358896e-01 7.83424750e-02 4.96613562e-01
4.06006306e-01 8.96726489e-01 2.31309935e-01 -4.00834799e-01
1.06250858e+00 -6.07254028e-01 -3.29380214e-01 -8.35597098e-01
1.10045171e+00 -6.34868443e-01 1.98397934e+00 2.76558459e-01
-1.15842652e+00 -6.33205593e-01 -1.03364885e+00 -2.56286830e-01
-5.60323298e-01 -1.30741522e-01 8.69076490e-01 8.29170167e-01
-9.53748882e-01 7.25923419e-01 -6.60321236e-01 -5.61708391e-01
5.40620312e-02 3.54893625e-01 -3.50000858e-02 -1.85481712e-01
-1.06118333e+00 1.10640514e+00 8.52172017e-01 -3.02501619e-01
-4.38818395e-01 -8.17063808e-01 -1.18479192e+00 1.91767052e-01
4.51414317e-01 -4.66491997e-01 1.31443131e+00 -1.21179855e+00
-1.22660863e+00 9.07009661e-01 -3.07867706e-01 -1.07213564e-01
-3.09940040e-01 1.21799521e-01 -1.84002534e-01 -2.81505495e-01
-5.22741787e-02 4.82717514e-01 5.88815808e-01 -1.08406937e+00
-5.96731484e-01 -3.60937804e-01 1.58924073e-01 8.84364620e-02
-3.15420836e-01 1.04887187e-01 5.10106012e-02 -9.22737896e-01
-7.12454244e-02 -9.41603005e-01 -7.93732926e-02 -3.53431314e-01
1.98967699e-02 -5.42906106e-01 2.63199866e-01 -5.07719576e-01
1.56016135e+00 -2.43668628e+00 3.01506549e-01 3.13651383e-01
-5.35811603e-01 2.64666647e-01 -1.13612682e-01 2.80542761e-01
-1.54306842e-02 4.51969534e-01 -4.62474287e-01 1.27851173e-01
8.29738975e-02 5.92499912e-01 -2.19623432e-01 -1.10912256e-01
3.07000428e-01 6.89595520e-01 -8.62393498e-01 -3.06995124e-01
-1.75772518e-01 3.02747972e-02 -1.03364670e+00 1.04755543e-01
-4.49317992e-01 -2.65852865e-02 -1.18753396e-01 2.17236459e-01
1.59849182e-01 2.27505773e-01 3.71997982e-01 3.07890534e-01
-1.25000998e-01 8.35919857e-01 -1.07436025e+00 1.69660485e+00
-8.86257172e-01 7.05764472e-01 -3.06584090e-01 -9.03205097e-01
9.05655444e-01 -1.30751105e-02 -4.14724380e-01 -5.23604810e-01
-1.62693694e-01 5.05176961e-01 7.35659838e-01 -7.80151486e-01
2.54750937e-01 -6.71892226e-01 -2.68584698e-01 5.53300381e-01
2.05605868e-02 -6.20377898e-01 6.24743879e-01 -1.61305919e-01
7.80362725e-01 -7.62702385e-03 3.77288699e-01 -6.40144765e-01
3.18761826e-01 2.63349801e-01 8.02928448e-01 9.32026923e-01
1.71085432e-01 1.52270436e-01 4.16619748e-01 -4.41927195e-01
-9.98731077e-01 -1.19355130e+00 -1.95560887e-01 1.66719615e+00
-6.54231966e-01 -5.51939309e-01 -7.22831905e-01 -3.78928363e-01
1.63289860e-01 1.31621242e+00 -4.12898481e-01 -4.21612471e-01
-1.02133012e+00 -5.45743287e-01 3.77130985e-01 8.17733586e-01
9.07522347e-03 -1.28087699e+00 -7.26050675e-01 1.57091185e-01
3.83393407e-01 -6.74858749e-01 -7.29061961e-01 6.19061649e-01
-1.21288121e+00 -7.51951575e-01 -2.86554962e-01 -1.30980814e+00
6.95226312e-01 -9.57580209e-02 1.24929547e+00 2.80590206e-01
5.31363562e-02 2.48640224e-01 -1.91613197e-01 -4.53661978e-01
-7.29134500e-01 2.56973565e-01 1.41937122e-01 -5.06083906e-01
5.55015504e-01 -6.06796265e-01 2.57753730e-01 -2.29820609e-01
-1.00954151e+00 -2.82511979e-01 3.88925642e-01 8.10261071e-01
2.54562825e-01 2.40006242e-02 6.51247859e-01 -1.44920588e+00
1.07160544e+00 -2.46678323e-01 -4.76377815e-01 2.75315315e-01
-4.91877496e-01 5.30863583e-01 9.52560306e-01 -7.77395308e-01
-1.14677262e+00 2.37702057e-01 7.33368918e-02 2.27252707e-01
-3.49737138e-01 1.14667404e+00 -2.97441095e-01 1.13073975e-01
1.20066559e+00 2.12212607e-01 -1.05227642e-01 -4.80847567e-01
5.21309257e-01 4.23475683e-01 4.02702272e-01 -1.31679010e+00
5.68565130e-01 -3.65132898e-01 -2.07609698e-01 -9.77341592e-01
-6.34796858e-01 1.15986846e-01 -7.59823203e-01 3.47220778e-01
5.62992096e-01 -6.08317256e-01 -6.49711210e-03 5.10767922e-02
-1.07434547e+00 -8.89952600e-01 -4.74144548e-01 4.06582892e-01
-7.23396122e-01 9.79652256e-02 -5.80537736e-01 -4.25691336e-01
3.62982154e-01 -1.30032027e+00 3.01604956e-01 1.72037169e-01
-6.42286599e-01 -1.26633430e+00 -3.71077284e-02 -1.98300302e-01
3.86544138e-01 1.52732655e-02 1.72228885e+00 -8.60365272e-01
-4.23080534e-01 1.88634768e-01 4.16258425e-01 6.40914917e-01
3.81547689e-01 1.05450928e-01 -8.05462062e-01 -1.00485720e-01
2.65751064e-01 -3.33919615e-01 7.24489748e-01 2.64549553e-01
1.20003724e+00 -5.07909060e-01 2.88085081e-02 6.31323695e-01
1.50101769e+00 2.97812968e-01 3.08754236e-01 1.55076683e-01
7.14470088e-01 7.41050184e-01 3.50681990e-01 -1.41539738e-01
1.67952284e-01 5.20934463e-01 -5.02982318e-01 3.57023954e-01
-2.89550871e-01 -3.24448615e-01 5.87596536e-01 9.56427872e-01
2.15751871e-01 2.80203968e-01 -1.31177306e+00 6.10154450e-01
-1.41876042e+00 -8.77436578e-01 3.10167968e-01 2.32204580e+00
1.45980370e+00 3.75086963e-01 3.07944983e-01 1.53835431e-01
6.24705493e-01 8.76384825e-02 -4.08102900e-01 -1.23677611e+00
-5.05216271e-02 5.40733695e-01 2.15252429e-01 1.00731707e+00
-5.66834688e-01 1.07397878e+00 7.44534445e+00 6.34373188e-01
-1.35324061e+00 -2.07812130e-01 3.44911754e-01 7.41260424e-02
-5.79839289e-01 2.81802826e-02 -6.81267560e-01 5.18120170e-01
1.31516898e+00 -5.14417648e-01 8.40604484e-01 7.54031777e-01
2.54690647e-01 -1.51019588e-01 -1.83073521e+00 5.32611251e-01
1.02661148e-01 -1.36003828e+00 3.39187384e-01 -1.39774621e-01
6.63161099e-01 -2.80756146e-01 2.16605097e-01 5.36495924e-01
3.90161514e-01 -1.05909061e+00 6.50765181e-01 1.37257546e-01
6.27177477e-01 -8.67664218e-01 1.08181112e-01 7.02611446e-01
-8.57475162e-01 -3.57090622e-01 -4.49933052e-01 -5.50218284e-01
-1.54845968e-01 2.25141943e-01 -8.77339005e-01 -2.02684164e-01
4.03851539e-01 3.82359177e-01 -9.33274448e-01 8.57442617e-01
-4.89139110e-01 1.01979053e+00 -3.53830725e-01 -2.64492482e-01
8.92809108e-02 2.09316760e-02 2.98762649e-01 1.49400604e+00
2.33824790e-01 3.33650768e-01 2.94717103e-01 9.30834293e-01
1.33967742e-01 2.70885199e-01 -1.12105656e+00 -9.87068415e-02
6.76082194e-01 4.29808557e-01 -5.14418542e-01 -3.05365533e-01
-5.50564647e-01 2.91018277e-01 6.52739942e-01 4.35587108e-01
-1.82149753e-01 -5.56479633e-01 5.82085967e-01 1.09188184e-01
4.20044601e-01 -5.18581629e-01 -6.88250840e-01 -1.12010884e+00
-3.93277220e-02 -1.25518882e+00 3.86312544e-01 -8.03408027e-01
-1.16806865e+00 1.44125387e-01 3.54971260e-01 -5.01944005e-01
-4.91652519e-01 -8.58291209e-01 -8.81742120e-01 1.11458349e+00
-1.07404530e+00 -4.32627738e-01 3.80248994e-01 2.21807569e-01
7.83681691e-01 -1.54186979e-01 7.54029453e-01 -8.74189585e-02
-3.99973422e-01 7.43915617e-01 -4.04038914e-02 1.67118147e-01
4.73536044e-01 -1.44581962e+00 3.90737861e-01 9.57945943e-01
4.74612653e-01 1.25424647e+00 8.20187688e-01 -5.73852360e-01
-1.30687559e+00 -6.90662384e-01 1.02318168e+00 -6.65827632e-01
6.34423494e-01 -3.99638385e-01 -1.38047516e+00 8.69596243e-01
1.38213739e-01 -1.67438105e-01 8.99951816e-01 4.84656215e-01
-5.13567626e-01 -1.22890629e-01 -9.18598950e-01 8.58549893e-01
1.04678893e+00 -9.81904089e-01 -1.23437655e+00 6.44213632e-02
1.00461054e+00 -2.43741900e-01 -8.24493110e-01 3.41853984e-02
1.08922027e-01 -7.06139147e-01 7.64376879e-01 -9.81584787e-01
3.50604743e-01 -2.51548231e-01 -3.94220769e-01 -1.79638875e+00
-4.11713511e-01 -4.48581815e-01 1.31834283e-01 1.28456807e+00
8.52098525e-01 -6.13438964e-01 5.16761243e-01 7.48681605e-01
-3.41436148e-01 -4.45407063e-01 -4.33419853e-01 -1.00743413e+00
9.21636820e-01 -4.91270423e-01 6.22992694e-01 8.00191104e-01
3.31546426e-01 5.97639620e-01 3.59438479e-01 -2.36713095e-03
1.67408511e-01 9.84192193e-02 4.83966351e-01 -1.19535923e+00
-4.04605865e-01 -4.82187063e-01 -1.81660667e-01 -9.39968765e-01
5.96331179e-01 -1.37754667e+00 1.54392749e-01 -8.94308627e-01
-1.17079102e-01 -6.71620488e-01 -6.63564876e-02 5.10537207e-01
-3.70820314e-01 -3.13352436e-01 5.86134255e-01 -2.00864986e-01
1.35735333e-01 -1.22617021e-01 1.08473623e+00 2.70762164e-02
-4.24623638e-01 6.54179379e-02 -9.33018625e-01 7.96807766e-01
9.88089442e-01 -5.34433126e-01 -7.00836062e-01 -8.45470786e-01
3.58675748e-01 -1.97144017e-01 -5.09279221e-02 -8.54806602e-01
2.97672480e-01 -5.53333700e-01 7.25015700e-02 8.06677621e-03
-5.93694262e-02 -6.87017977e-01 -3.04078072e-01 5.12172937e-01
-8.56815159e-01 5.43851793e-01 7.22783029e-01 3.02143544e-01
-4.11909074e-02 -1.00937283e+00 6.47753954e-01 -4.93730545e-01
-1.04037237e+00 -2.51067966e-01 -6.45912051e-01 6.45499825e-01
5.58425903e-01 -2.40808249e-01 -8.45478773e-02 3.55221517e-02
-8.74681711e-01 -1.01783700e-01 6.88412189e-01 2.44785517e-01
2.27594152e-01 -1.40476012e+00 -5.86292148e-01 4.81514961e-01
6.39076233e-02 -1.38374224e-01 -4.57477450e-01 1.87477827e-01
-4.05216336e-01 4.11475599e-01 -1.80650771e-01 -5.80969334e-01
-1.05984044e+00 7.52700090e-01 3.83859307e-01 2.94429958e-01
-2.57925034e-01 8.10026228e-01 2.40492504e-02 -6.50791049e-01
1.72781110e-01 -9.28406775e-01 2.16764033e-01 -1.57837048e-02
4.05702502e-01 -2.71769315e-02 -9.24580321e-02 -2.80715555e-01
-1.48460880e-01 5.77940464e-01 -5.59461713e-02 -3.25783432e-01
1.26640391e+00 1.60235211e-01 9.71590634e-03 1.02984929e+00
1.10693228e+00 1.76126584e-01 -1.18571055e+00 -3.99873108e-01
5.56867361e-01 -5.36930680e-01 -1.02822848e-01 -8.16451907e-01
-6.51696682e-01 1.02831376e+00 1.01378098e-01 3.46029460e-01
9.30569828e-01 -5.11511788e-02 3.48768651e-01 8.36317360e-01
3.36668342e-01 -1.28710854e+00 2.98254322e-02 8.14045846e-01
7.98664391e-01 -1.17374432e+00 -1.71655521e-01 -3.09974700e-01
-5.40223777e-01 1.09421539e+00 7.60727406e-01 -2.77753890e-01
6.40597761e-01 3.77171755e-01 2.83054952e-02 1.68404594e-01
-7.85977721e-01 -1.81257054e-01 1.97852716e-01 9.55796599e-01
7.17345655e-01 5.13671264e-02 -3.66881490e-01 4.47761685e-01
-6.25349224e-01 -1.56100243e-01 7.31064558e-01 1.10205150e+00
-8.06347251e-01 -1.34445190e+00 -4.13954645e-01 6.13833845e-01
-1.72572061e-01 -4.15747523e-01 -5.16084731e-01 1.00467038e+00
3.04971904e-01 6.86629295e-01 1.25523999e-01 -4.95499559e-02
4.49665308e-01 6.73012853e-01 7.61190236e-01 -1.70593619e+00
-8.45363200e-01 -3.37689459e-01 1.91360429e-01 -1.39562875e-01
-1.62204430e-01 -9.60673809e-01 -1.51185417e+00 -2.52420753e-01
-6.70856237e-02 2.52723426e-01 2.87306249e-01 9.74581122e-01
7.18765929e-02 2.09731743e-01 3.52473944e-01 -5.82517385e-01
-8.39711547e-01 -5.66406369e-01 -5.12383163e-01 3.78565639e-01
5.18177748e-01 -3.72482657e-01 -5.56061625e-01 1.89657301e-01] | [10.663908004760742, 9.260122299194336] |
a4759fde-8384-4ca9-ad65-ebaf307a528a | flood-prediction-using-machine-learning-1 | 2208.01234 | null | https://arxiv.org/abs/2208.01234v1 | https://arxiv.org/pdf/2208.01234v1.pdf | Flood Prediction Using Machine Learning Models | Floods are one of nature's most catastrophic calamities which cause irreversible and immense damage to human life, agriculture, infrastructure and socio-economic system. Several studies on flood catastrophe management and flood forecasting systems have been conducted. The accurate prediction of the onset and progression of floods in real time is challenging. To estimate water levels and velocities across a large area, it is necessary to combine data with computationally demanding flood propagation models. This paper aims to reduce the extreme risks of this natural disaster and also contributes to policy suggestions by providing a prediction for floods using different machine learning models. This research will use Binary Logistic Regression, K-Nearest Neighbor (KNN), Support Vector Classifier (SVC) and Decision tree Classifier to provide an accurate prediction. With the outcome, a comparative analysis will be conducted to understand which model delivers a better accuracy. | ['Tanvir Rahman', 'Meherin Hossain Nushra', 'Ipshita Ishrar', 'Ishadie Namir', 'Maisha Farzana', 'Miah Mohammad Asif Syeed'] | 2022-08-02 | null | null | null | null | ['machine-learning', 'machine-learning'] | ['methodology', 'miscellaneous'] | [-7.15370430e-03 -2.35856146e-01 8.25101361e-02 -4.12164241e-01
1.03449516e-01 -3.40131313e-01 2.19748512e-01 9.29476678e-01
-3.00481528e-01 1.06605148e+00 3.31641972e-01 -1.00049961e+00
-3.25654030e-01 -1.40934229e+00 7.52990544e-02 -6.56450093e-01
-6.07205331e-01 2.54977465e-01 1.62875131e-01 -6.62914991e-01
5.96845567e-01 8.97322297e-01 -1.34799480e+00 8.11678544e-03
8.46571565e-01 4.36273456e-01 1.75590098e-01 8.74553740e-01
-8.44926462e-02 2.92072892e-01 -3.86011988e-01 3.04731458e-01
7.74559304e-02 -1.60642207e-01 -6.53576434e-01 -6.29562438e-01
-5.66438735e-01 -5.20743370e-01 2.11411297e-01 2.96740085e-01
6.50962472e-01 1.29075378e-01 1.12138033e+00 -1.26334298e+00
-2.29888663e-01 4.59563613e-01 -7.34208465e-01 5.24084866e-01
4.37523812e-01 2.85930317e-02 2.18836013e-02 -6.21367872e-01
-8.37658197e-02 1.21308517e+00 7.85202026e-01 -1.92021072e-01
-9.94225264e-01 -8.52882922e-01 -4.61334497e-01 1.82329834e-01
-1.29985726e+00 -6.96674883e-02 2.70209908e-01 -5.68544328e-01
1.27394176e+00 3.89867961e-01 6.73166990e-01 1.79790556e-01
7.19197989e-01 -3.93369831e-02 1.12538302e+00 -6.41007841e-01
4.48415756e-01 2.91702989e-02 3.23888697e-02 1.38941869e-01
4.82788682e-01 3.84204894e-01 -1.83324382e-01 -3.05001289e-01
4.28658307e-01 1.90748647e-01 -1.14852540e-01 5.14396727e-01
-5.46483934e-01 1.04713285e+00 7.31150866e-01 4.91465181e-01
-7.87139654e-01 -5.98746091e-02 2.46389300e-01 3.27408612e-01
5.33518493e-01 2.27101266e-01 -7.51778603e-01 1.85540438e-01
-1.27071083e+00 3.41357499e-01 8.23658049e-01 3.61942321e-01
7.97607601e-01 1.06058724e-01 2.11103395e-01 3.34453553e-01
5.52282810e-01 1.07453489e+00 8.70573297e-02 -2.93603390e-01
3.19150597e-01 5.63523591e-01 2.71319032e-01 -1.53296173e+00
-8.82369816e-01 -8.72920230e-02 -1.21940911e+00 3.47414404e-01
2.65163541e-01 -5.19152701e-01 -7.36022413e-01 1.02945518e+00
2.47592747e-01 2.15378374e-01 2.74130374e-01 4.66690183e-01
5.00788867e-01 1.25086510e+00 7.21868098e-01 -2.25450039e-01
1.21566689e+00 1.46870345e-01 -5.76563776e-01 -9.17974412e-02
7.85219431e-01 -7.86538541e-01 6.10556066e-01 -1.28915533e-01
-4.27906483e-01 -1.47146240e-01 -6.96800470e-01 5.67875206e-01
-8.92178118e-01 -1.28793210e-01 7.09339201e-01 7.84767330e-01
-6.86248541e-01 5.50235808e-01 -9.92059946e-01 -7.06446469e-01
3.07857513e-01 2.05348030e-01 -2.46988714e-01 1.20346457e-01
-1.40980697e+00 1.33124447e+00 3.70680898e-01 3.72325122e-01
-2.62438416e-01 -8.21700513e-01 -9.17289555e-01 1.13143831e-01
-5.50382793e-01 -2.34410986e-01 4.21664834e-01 1.78818181e-01
-6.77801549e-01 3.54913682e-01 -2.32522830e-01 -4.39417273e-01
2.42690861e-01 -1.01407059e-01 -1.84768394e-01 -1.99094549e-01
2.97747344e-01 4.12528723e-01 2.96735149e-02 -1.00426257e+00
-8.01084280e-01 -5.41424394e-01 -5.18014967e-01 1.38764724e-01
-2.57832915e-01 2.34349430e-01 1.17229235e+00 -5.37614822e-01
4.06521976e-01 -5.66343486e-01 -7.33871281e-01 -3.85855645e-01
-2.43479714e-01 -7.26294816e-02 9.37478304e-01 -9.00166929e-01
1.44475222e+00 -1.64906728e+00 -4.00493026e-01 3.08222294e-01
-3.78173053e-01 4.37693208e-01 3.84000599e-01 1.06621635e+00
8.03289264e-02 3.74496847e-01 -7.12691545e-01 3.80206674e-01
-4.90141898e-01 2.23742157e-01 -4.81879145e-01 5.25437415e-01
3.34793687e-01 3.85243922e-01 -6.81874037e-01 -2.71452963e-01
6.16322517e-01 6.13532484e-01 8.43243673e-02 2.07639590e-01
3.59220088e-01 4.88667518e-01 -5.05516589e-01 6.44586205e-01
1.18599272e+00 3.99201751e-01 -2.47087911e-01 1.32482350e-01
-5.98305285e-01 -3.76286328e-01 -1.12726831e+00 4.91777331e-01
-4.46830720e-01 5.69653630e-01 -3.02561641e-01 -1.22413170e+00
1.54989040e+00 5.18128395e-01 3.48750085e-01 -4.36732858e-01
-9.79802534e-02 2.81707108e-01 -3.99068743e-01 -9.84013379e-01
4.95796412e-01 -4.36290622e-01 1.61700379e-02 4.82506067e-01
-7.28632331e-01 -8.52127746e-02 -1.60084084e-01 -1.51429549e-01
7.66742766e-01 -2.69243777e-01 5.51031590e-01 -6.67571962e-01
4.76538748e-01 3.51772994e-01 3.58924001e-01 4.26152140e-01
-3.11850637e-01 1.76210687e-01 1.81276888e-01 -8.74808729e-01
-8.60174179e-01 -8.92452955e-01 -4.54611003e-01 7.79659688e-01
1.97395496e-02 5.66672683e-01 -3.29421431e-01 1.68254122e-01
2.52534389e-01 1.05223465e+00 -1.81889445e-01 3.15266736e-02
-6.42046332e-01 -1.58211064e+00 6.11221492e-01 5.29148340e-01
6.45285606e-01 -1.30972469e+00 -1.15986311e+00 4.68188882e-01
-7.99788982e-02 -5.00272870e-01 5.67783535e-01 3.64698023e-01
-1.22569585e+00 -9.60756302e-01 -6.39210761e-01 -7.47981489e-01
5.57550788e-01 2.26153091e-01 8.69799852e-01 3.38941872e-01
-6.86060071e-01 -3.47622782e-01 -5.10606885e-01 -6.49718881e-01
-3.31616431e-01 1.97757676e-01 -1.03777841e-01 -7.16017008e-01
8.10415447e-01 -8.17824602e-01 -7.98139155e-01 2.22776845e-01
-8.18816543e-01 -3.16390008e-01 4.35614944e-01 1.48355827e-01
2.38802601e-02 5.26425540e-01 1.07237387e+00 -4.82964247e-01
7.38564670e-01 -1.20419133e+00 -5.11126876e-01 2.86106199e-01
-7.43609846e-01 -1.65672123e-01 3.77616286e-01 -5.92338759e-03
-1.07776105e+00 2.79218405e-01 -1.70406535e-01 7.54200220e-01
-7.83818662e-01 8.02791953e-01 3.00618082e-01 2.60529537e-02
8.75052512e-01 1.48540035e-01 -4.28724498e-01 -4.43290353e-01
-9.69387591e-02 1.06526625e+00 4.60314184e-01 -7.16163963e-02
8.65464926e-01 3.57754499e-01 4.45368260e-01 -1.52636039e+00
-1.17440604e-01 -6.59848154e-01 -9.73902464e-01 -4.43814546e-01
6.68447912e-01 -7.30302274e-01 -3.88112247e-01 7.15946138e-01
-1.14696395e+00 -2.49680266e-01 3.94151419e-01 6.44410491e-01
4.12560068e-02 4.61161025e-02 -1.78712472e-01 -1.40367770e+00
-5.78675807e-01 -5.34050047e-01 4.38466668e-01 6.90917194e-01
-1.74506679e-01 -1.05557585e+00 2.31232539e-01 -1.65252864e-01
1.03351939e+00 9.60450888e-01 8.87814164e-01 -2.69040018e-01
-2.67273616e-02 -2.29531467e-01 -3.94566804e-01 -3.43168259e-01
4.67120826e-01 2.66084015e-01 -6.43553317e-01 -9.77975428e-02
-2.25521609e-01 1.09235998e-02 9.60763872e-01 9.12697256e-01
2.14828104e-01 -3.31513554e-01 -5.35969019e-01 4.24581349e-01
1.72427547e+00 3.86238784e-01 8.91137183e-01 5.77898264e-01
-5.44544347e-02 1.17302895e+00 7.46227622e-01 8.40252340e-01
4.71582919e-01 -2.99686622e-02 4.42689538e-01 -2.11845174e-01
2.53505588e-01 2.36448478e-02 3.87478434e-02 1.07323676e-01
-8.18227530e-02 -2.64216810e-01 -1.52086747e+00 7.54898787e-01
-1.45561779e+00 -1.16656768e+00 -8.17666590e-01 2.13696384e+00
4.73414898e-01 -2.70324737e-01 -1.51031967e-02 7.10739911e-01
6.15308762e-01 -1.10105947e-01 -7.32307658e-02 -7.28802323e-01
3.49716702e-03 2.55310595e-01 1.01346016e+00 7.24599302e-01
-1.14055204e+00 8.60215008e-01 6.44402742e+00 2.16194764e-01
-1.27516031e+00 -3.88450176e-01 7.19739854e-01 5.33029974e-01
-6.28471896e-02 2.77695030e-01 -8.31462085e-01 4.27701861e-01
1.13928556e+00 -6.52090758e-02 -1.13789560e-02 2.15045795e-01
9.67275381e-01 -9.39345956e-01 -1.85995355e-01 1.66653872e-01
-6.37949526e-01 -1.00371277e+00 -2.93235809e-01 -1.10814065e-01
4.06901091e-01 -1.08404845e-01 -4.66501236e-01 8.85823090e-03
5.17600954e-01 -1.22940207e+00 -1.43847719e-01 7.28993773e-01
2.29570255e-01 -8.36770475e-01 1.04415011e+00 6.08627677e-01
-1.35268962e+00 -1.09878302e-01 -4.82221335e-01 -7.46693373e-01
6.51957333e-01 8.86638999e-01 -8.07353079e-01 2.67586738e-01
8.96817029e-01 4.09953624e-01 -4.19065684e-01 1.28904235e+00
-1.46418869e-01 9.27547097e-01 -7.74973929e-01 -1.01040341e-01
1.90458730e-01 -8.82655084e-02 2.53404975e-01 1.30215263e+00
4.54765409e-01 7.26312995e-01 3.75932790e-02 4.64978814e-01
8.43884885e-01 3.72906983e-01 -7.82554209e-01 5.27110934e-01
8.08772087e-01 7.88921356e-01 -9.35847819e-01 1.03909686e-01
3.66543271e-02 4.16763037e-01 -3.67395699e-01 1.49321556e-01
-4.01610583e-01 -7.39197791e-01 4.01758552e-01 3.64875019e-01
-4.27534878e-01 -4.39718246e-01 -6.89812601e-01 -4.18884873e-01
-3.09519589e-01 1.66879267e-01 3.52933407e-01 -4.37071890e-01
-9.44921672e-01 4.45237249e-01 4.71551210e-01 -7.88974345e-01
-1.97194576e-01 -2.35780418e-01 -1.20091426e+00 1.29120684e+00
-2.02326655e+00 -1.13125944e+00 -4.52700794e-01 7.75147006e-02
1.11594684e-01 9.15120170e-03 1.08458316e+00 2.58632153e-01
-4.71808523e-01 7.07995333e-03 2.35470071e-01 -1.28436565e-01
3.46013546e-01 -9.36190188e-01 3.88862640e-01 5.36850572e-01
-9.01249468e-01 -1.08638883e-01 1.05020118e+00 -1.11754882e+00
-6.28863692e-01 -1.14278424e+00 1.52947152e+00 2.72827536e-01
3.43351871e-01 3.49659026e-01 -1.17185807e+00 1.98510081e-01
-1.65887758e-01 -2.21861377e-01 6.28341436e-01 -1.83950379e-01
2.70775974e-01 -1.07782751e-01 -1.72661030e+00 1.55252397e-01
8.75929594e-02 9.54120383e-02 -3.09270263e-01 3.13803703e-01
2.37453997e-01 2.51222312e-01 -1.04629278e+00 4.52205211e-01
8.27138484e-01 -1.03433549e+00 1.00047755e+00 -5.11796355e-01
4.38106149e-01 -9.62063745e-02 -2.66773343e-01 -1.10407305e+00
-1.70875534e-01 8.08809549e-02 6.05681956e-01 1.34372866e+00
5.16523004e-01 -6.96055055e-01 7.84213245e-01 9.16760206e-01
4.83637989e-01 -7.09700644e-01 -8.10384393e-01 -5.06296098e-01
7.28556216e-01 -1.88867703e-01 8.43276024e-01 8.85584593e-01
6.59195259e-02 -1.99144498e-01 -4.28511053e-01 7.59021223e-01
6.81377351e-01 -1.86338678e-01 3.65107626e-01 -1.49478114e+00
7.75186300e-01 -1.73050553e-01 -5.15246153e-01 9.55399498e-02
-2.97046900e-01 -2.94538945e-01 -1.40414611e-01 -1.73801982e+00
-1.26850650e-01 -9.42637861e-01 1.36531100e-01 8.41199338e-01
-1.26639307e-01 1.90873425e-02 -1.87315032e-01 2.41320744e-01
7.18681216e-01 4.59006220e-01 5.27028799e-01 8.07765722e-02
-8.45213950e-01 4.07929510e-01 -2.75848776e-01 4.85818416e-01
1.40407336e+00 -8.75028551e-01 -2.56809354e-01 -3.28510880e-01
3.33235472e-01 5.05746067e-01 3.27407092e-01 -1.04867697e+00
2.87762463e-01 -8.98883998e-01 4.09622669e-01 -9.50617194e-01
-2.48980701e-01 -8.27773511e-01 2.47797310e-01 1.10948193e+00
-1.62807181e-01 1.33153379e-01 4.45726871e-01 3.73435110e-01
-1.59245282e-02 -4.22848105e-01 8.75036836e-01 1.72763355e-02
-3.86393547e-01 9.13248137e-02 -8.84846926e-01 -4.39277172e-01
1.40566683e+00 -2.18813837e-01 -1.94805712e-01 -2.38649979e-01
-5.91906965e-01 6.89225018e-01 6.60096928e-02 1.02412373e-01
5.92124701e-01 -8.20401430e-01 -1.31168211e+00 2.08786950e-01
-2.01770261e-01 -1.49308741e-01 1.34654835e-01 3.62032920e-01
-1.27006805e+00 4.81639177e-01 -5.29167533e-01 -1.15972728e-01
-1.07901525e+00 2.98371632e-02 1.60610825e-01 -2.64992863e-01
-4.23460990e-01 5.51726758e-01 -6.99102700e-01 -4.14886683e-01
-1.38960734e-01 -1.59559503e-01 -9.13291514e-01 2.49962658e-01
6.17776752e-01 9.41700339e-01 -1.37315020e-01 -8.62266898e-01
-6.16544783e-01 6.83425963e-01 3.71288329e-01 -1.04680084e-01
1.64617085e+00 -4.41356301e-01 -3.15697163e-01 1.99591771e-01
8.71165216e-01 -6.29166961e-01 -7.54712343e-01 1.93092436e-01
2.81001270e-01 -4.13064957e-01 5.08021951e-01 -7.30924964e-01
-8.77673090e-01 1.05275190e+00 8.83812666e-01 3.02116513e-01
1.25970924e+00 -3.98729593e-01 8.16922367e-01 3.95206273e-01
1.19518526e-01 -7.76056528e-01 -6.78109229e-01 2.22606748e-01
9.40536559e-01 -1.45671070e+00 3.01981091e-01 -3.37048560e-01
-1.93902239e-01 1.33674991e+00 1.60061479e-01 -1.03757553e-01
1.45665371e+00 5.76680005e-01 5.38520254e-02 7.34865665e-04
-2.71013409e-01 -2.29288880e-02 -6.11989796e-01 1.02177572e+00
5.58095992e-01 5.12093127e-01 -6.40547097e-01 3.26315045e-01
-1.66933820e-01 3.32667232e-01 5.92903078e-01 1.26574564e+00
-1.17836678e+00 -9.73378956e-01 -7.93918073e-01 7.37530768e-01
-2.42427409e-01 -2.43511453e-01 2.36311853e-02 3.66100311e-01
-1.89781323e-01 1.51264572e+00 5.81049221e-03 -9.12251025e-02
8.19303766e-02 -2.28852972e-01 -4.15019691e-01 -2.30438858e-01
-5.83472610e-01 -5.68985939e-01 -3.65494996e-01 -4.82569598e-02
-4.07541066e-01 -6.80937350e-01 -1.54873741e+00 -9.52405870e-01
-4.65471178e-01 2.58926868e-01 1.11826503e+00 1.02570570e+00
5.10268770e-02 3.60303670e-02 8.81967545e-01 -9.54142034e-01
-3.78165156e-01 -1.18412292e+00 -8.93589020e-01 -1.94399893e-01
3.33899915e-01 -5.76450229e-01 -5.14331102e-01 -3.29302587e-02] | [9.311513900756836, -1.2035691738128662] |
06c8cccd-16da-40b5-bf0e-dd15f7ab40d1 | ranking-vs-classifying-measuring-knowledge | 2102.06145 | null | https://arxiv.org/abs/2102.06145v1 | https://arxiv.org/pdf/2102.06145v1.pdf | Ranking vs. Classifying: Measuring Knowledge Base Completion Quality | Knowledge base completion (KBC) methods aim at inferring missing facts from the information present in a knowledge base (KB) by estimating the likelihood of candidate facts. In the prevailing evaluation paradigm, models do not actually decide whether a new fact should be accepted or not but are solely judged on the position of true facts in a likelihood ranking with other candidates. We argue that consideration of binary predictions is essential to reflect the actual KBC quality, and propose a novel evaluation paradigm, designed to provide more transparent model selection criteria for a realistic scenario. We construct the data set FB14k-QAQ where instead of single facts, we use KB queries, i.e., facts where one entity is replaced with a variable, and construct corresponding sets of entities that are correct answers. We randomly remove some of these correct answers from the data set, simulating the realistic scenario of real-world entities missing from a KB. This way, we can explicitly measure a model's ability to handle queries that have more correct answers in the real world than in the KB, including the special case of queries without any valid answer. The latter especially contrasts the ranking setting. We evaluate a number of state-of-the-art KB embeddings models on our new benchmark. The differences in relative performance between ranking-based and classification-based evaluation that we observe in our experiments confirm our hypothesis that good performance on the ranking task does not necessarily translate to good performance on the actual completion task. Our results motivate future work on KB embedding models with better prediction separability and, as a first step in that direction, we propose a simple variant of TransE that encourages thresholding and achieves a significant improvement in classification F1 score relative to the original TransE. | ['Benjamin Roth', 'Martin Schmitt', 'Marina Speranskaya'] | 2021-02-02 | null | https://openreview.net/forum?id=3pcecaCEK- | https://openreview.net/pdf?id=3pcecaCEK- | akbc-2020-6 | ['knowledge-base-completion', 'knowledge-base-completion'] | ['graphs', 'knowledge-base'] | [-1.04744770e-01 4.72169608e-01 -4.85020787e-01 -4.58215147e-01
-9.43778157e-01 -5.84541321e-01 7.48415530e-01 5.05109191e-01
-7.81871259e-01 1.24313450e+00 4.77252752e-01 -2.70379096e-01
-3.36139143e-01 -1.16522169e+00 -9.85032856e-01 -3.28475296e-01
5.97397313e-02 1.04981363e+00 4.08897460e-01 -3.70982498e-01
-1.64832547e-01 1.28787428e-01 -1.55802190e+00 3.98943454e-01
9.09312546e-01 9.00659382e-01 -3.15436013e-02 4.06732738e-01
-9.02637914e-02 9.69666243e-01 -6.48142517e-01 -1.18796849e+00
-7.45984912e-02 8.47760811e-02 -1.12245274e+00 -5.68943322e-01
6.74686193e-01 -1.31385833e-01 -1.78797066e-01 9.75846946e-01
2.98981249e-01 -1.31151125e-01 8.58125448e-01 -1.06136978e+00
-7.29327798e-01 8.08484852e-01 2.60070443e-01 1.88919261e-01
6.40151381e-01 -2.77605295e-01 1.58241677e+00 -1.12285113e+00
8.99090230e-01 1.03325820e+00 6.42215788e-01 4.43161398e-01
-1.36431253e+00 -2.29773268e-01 5.47813736e-02 5.94361663e-01
-1.35660946e+00 -3.11783373e-01 4.66098875e-01 -3.38406175e-01
1.04997754e+00 5.24340749e-01 4.55712885e-01 1.03380275e+00
-8.26030076e-02 5.63027322e-01 1.03672147e+00 -4.95387971e-01
3.64247680e-01 6.29323661e-01 4.86772388e-01 4.87682641e-01
7.22777128e-01 -9.49069411e-02 -5.23605585e-01 -3.66362333e-01
1.47299156e-01 -4.38397110e-01 -6.37865126e-01 -5.80518782e-01
-1.30440176e+00 8.20061088e-01 3.66332471e-01 1.60484686e-01
-4.01987314e-01 1.05611645e-01 2.13760599e-01 3.46229583e-01
3.65718454e-01 7.58176863e-01 -9.66395557e-01 -5.83954304e-02
-7.52605081e-01 6.21100426e-01 1.09525490e+00 7.25371718e-01
8.40166330e-01 -4.30231273e-01 -2.73809224e-01 7.20522940e-01
6.96708709e-02 3.87095034e-01 3.37355107e-01 -7.62573838e-01
6.14883840e-01 7.42127180e-01 5.74273527e-01 -1.00832915e+00
-2.43013650e-01 -5.88830531e-01 -3.42940897e-01 -1.19108193e-01
3.21827382e-01 1.15360379e-01 -7.10096419e-01 1.95514548e+00
3.15757155e-01 1.35344133e-01 5.72959661e-01 7.55570471e-01
8.93310547e-01 4.29842561e-01 1.54760450e-01 -3.06682419e-02
1.46709955e+00 -5.15668511e-01 -5.90783656e-01 -5.84949851e-02
7.85330296e-01 -3.79111618e-01 1.14803267e+00 3.13864321e-01
-7.19057322e-01 -4.10373300e-01 -1.05129397e+00 -1.42928049e-01
-8.44385862e-01 4.07144159e-01 8.67636204e-01 7.48817563e-01
-9.27940011e-01 4.29026127e-01 -4.20592070e-01 -1.16248488e-01
-2.29559597e-02 2.23486498e-01 -6.49792910e-01 -1.68698043e-01
-1.82213235e+00 1.37120235e+00 7.36061931e-01 3.60544994e-02
-6.48722649e-01 -7.48228133e-01 -8.56481194e-01 3.84772658e-01
8.48366499e-01 -8.42268109e-01 9.95917499e-01 -6.08589709e-01
-8.84597540e-01 7.44342566e-01 -1.29018739e-01 -5.95933259e-01
4.28297848e-01 -3.46173227e-01 -7.75128961e-01 -2.16918066e-01
1.82765111e-01 5.91391444e-01 4.17656720e-01 -1.50540590e+00
-8.16308558e-01 -1.58740044e-01 6.94749713e-01 1.22496679e-01
-2.01485798e-01 -3.47566098e-01 -5.00812471e-01 -3.04950893e-01
-1.44572586e-01 -8.55482042e-01 1.03424095e-01 -3.96882236e-01
-4.11755562e-01 -3.27360749e-01 2.64955997e-01 -5.18488109e-01
1.41125703e+00 -1.73666131e+00 1.19084284e-01 2.11478710e-01
1.87542737e-01 3.92000228e-01 -1.38133541e-02 5.01645565e-01
-1.72552168e-01 2.54660219e-01 -1.47199975e-02 -3.20303917e-01
2.71470040e-01 4.72788870e-01 -7.16734231e-01 -8.72409437e-03
3.37120175e-01 9.91114259e-01 -9.29983556e-01 -4.65289116e-01
-1.78338960e-01 2.14038819e-01 -6.54046416e-01 -2.43728235e-01
-5.60779095e-01 -2.09049150e-01 -3.82550806e-01 4.68719840e-01
5.16072810e-01 -3.82095993e-01 3.54113996e-01 -4.01758820e-01
2.89713055e-01 8.12999189e-01 -1.43124843e+00 1.17811179e+00
-4.90324646e-01 2.69318074e-01 -4.23526555e-01 -8.12463045e-01
6.00921214e-01 4.33212370e-01 -2.49202736e-02 -4.10066903e-01
-4.80137467e-01 3.05785775e-01 -4.58400361e-02 -4.34353352e-01
1.04743719e+00 -2.91104704e-01 -1.40751317e-01 2.18201969e-02
3.34266394e-01 2.05646548e-03 4.04109061e-01 4.89360064e-01
1.07789099e+00 6.01934828e-02 3.76221746e-01 -1.57410055e-01
5.14159977e-01 2.41832927e-01 7.28722632e-01 8.87677193e-01
1.87776461e-01 2.97043294e-01 7.37227440e-01 -3.77712935e-01
-7.06129432e-01 -1.09637892e+00 -2.27404594e-01 8.22408795e-01
1.54610500e-01 -8.03907931e-01 -1.35533705e-01 -1.19995868e+00
3.62182587e-01 1.08748233e+00 -8.83698642e-01 -1.08249843e-01
-2.95571536e-01 -7.51495123e-01 5.45933723e-01 4.10056293e-01
1.14578113e-01 -9.37182009e-01 -4.21846598e-01 2.77394593e-01
-4.59656298e-01 -1.06151307e+00 7.01203048e-02 3.21246654e-01
-5.84659457e-01 -1.28019774e+00 -2.57162690e-01 -5.72666883e-01
4.09748405e-01 -2.85681874e-01 1.67563915e+00 -2.83940919e-02
2.30136350e-01 4.83701020e-01 -4.30028260e-01 -2.57435858e-01
-3.49770546e-01 3.39800417e-02 1.46142334e-01 -1.09986603e-01
4.93744254e-01 -3.21063370e-01 -3.42464924e-01 1.06572822e-01
-1.02935350e+00 -1.12217292e-01 5.73167503e-01 9.98543143e-01
5.26932061e-01 3.33138764e-01 7.66474366e-01 -1.22830999e+00
7.28881896e-01 -5.47009408e-01 -4.68727082e-01 8.53518188e-01
-8.31441998e-01 4.73655671e-01 6.67378247e-01 -2.91426986e-01
-1.08742833e+00 -2.69354731e-01 -1.48584634e-01 -3.72247659e-02
5.51244430e-02 1.14713240e+00 -2.26909578e-01 3.15178066e-01
7.45905042e-01 1.93053126e-01 -6.17159963e-01 -3.83654147e-01
5.78084648e-01 3.46921355e-01 4.88853574e-01 -8.86508226e-01
6.50615215e-01 2.75920898e-01 -1.56642705e-01 -3.54539841e-01
-9.78749275e-01 -4.65313733e-01 -3.69963795e-01 5.65513521e-02
4.04940307e-01 -1.00221860e+00 -5.82687795e-01 -1.67565033e-01
-1.24794030e+00 -2.18483415e-02 -4.31842387e-01 5.79385638e-01
-3.20729554e-01 2.97052175e-01 -3.63565207e-01 -6.49814546e-01
-2.10601799e-02 -9.22726691e-01 9.85682309e-01 -1.43994659e-01
-2.32191548e-01 -1.05569267e+00 3.72746140e-01 3.04956079e-01
2.44389191e-01 -4.88649532e-02 1.44199216e+00 -8.92387629e-01
-6.49862587e-01 -3.66818041e-01 -2.18070596e-02 4.75856781e-01
-8.97647440e-03 -1.12018630e-01 -9.64776218e-01 -1.04042171e-02
-3.32293004e-01 -5.09757340e-01 1.16526484e+00 -1.24127291e-01
5.91383338e-01 -3.51148635e-01 -4.20164883e-01 -7.92389177e-03
1.72903490e+00 -1.56398728e-01 8.64919782e-01 2.00834617e-01
3.69819015e-01 6.41383171e-01 7.82970965e-01 1.17302619e-01
6.92672670e-01 9.36910927e-01 2.67515540e-01 2.01707751e-01
-1.11903518e-01 -5.05307794e-01 2.95981646e-01 5.71250618e-01
-1.33847624e-01 -4.35316354e-01 -8.29295218e-01 8.21269989e-01
-1.72577929e+00 -1.00760818e+00 -7.07658976e-02 2.50612879e+00
1.33528006e+00 3.72613996e-01 -1.61114633e-01 1.46619767e-01
3.18770826e-01 -1.02442406e-01 -4.99694683e-02 1.23514812e-02
-3.17850381e-01 2.88798183e-01 3.99710178e-01 8.89240444e-01
-1.07059002e+00 9.75235581e-01 5.87443447e+00 9.45541322e-01
-9.37126040e-01 2.29634717e-02 4.34883326e-01 2.08353490e-01
-9.25975144e-01 3.05529445e-01 -1.05618131e+00 4.85806674e-01
8.36880445e-01 -5.73142320e-02 7.44443759e-02 8.28890860e-01
-4.08264607e-01 -2.92492509e-01 -1.40167856e+00 5.85110486e-01
1.74359959e-02 -1.40622699e+00 2.72249609e-01 -3.53459194e-02
4.63891208e-01 -1.78945884e-01 -6.14099428e-02 9.54525292e-01
4.76812273e-01 -1.00368094e+00 8.74281287e-01 7.38402963e-01
7.49746740e-01 -4.57919151e-01 1.03040540e+00 3.58165592e-01
-8.50385606e-01 2.74042189e-02 -5.39938509e-01 -1.51795531e-02
3.53330672e-02 9.19384062e-01 -1.18540668e+00 9.66318965e-01
5.09316027e-01 3.34451199e-01 -8.41885090e-01 9.26078439e-01
-4.60642099e-01 6.87569618e-01 -2.28636801e-01 -1.42855152e-01
-9.03910995e-02 4.18691039e-02 4.52437967e-01 1.28571177e+00
8.17446187e-02 -8.22174922e-02 -1.49722204e-01 8.25793266e-01
-2.87582636e-01 1.28506750e-01 -4.57432479e-01 1.59695745e-01
5.65392911e-01 1.03491426e+00 -3.62782955e-01 -6.47783041e-01
-4.77681369e-01 7.03267634e-01 8.07359934e-01 4.12891656e-01
-9.05120134e-01 -1.84934318e-01 5.23707032e-01 -1.96079001e-01
5.28037786e-01 1.21399455e-01 -4.46141064e-02 -1.49268568e+00
5.16307771e-01 -7.79971063e-01 4.53422993e-01 -6.94677174e-01
-1.36437654e+00 5.57913840e-01 1.93282351e-01 -8.47097099e-01
-3.29797834e-01 -6.41850352e-01 -1.98589683e-01 8.78816903e-01
-1.87477028e+00 -9.41629469e-01 -3.57121676e-02 3.80458653e-01
-2.81064138e-02 2.24372849e-01 1.08290327e+00 3.88928056e-01
-2.33432367e-01 5.06763935e-01 -5.32463845e-03 -1.36853037e-02
8.44211340e-01 -1.70621049e+00 1.95830360e-01 7.59742677e-01
5.01158535e-01 8.82932305e-01 8.92734468e-01 -8.90546203e-01
-1.05813408e+00 -9.60652888e-01 1.49152648e+00 -1.00480056e+00
5.56080639e-01 -4.66052920e-01 -9.88382578e-01 7.96973407e-01
-2.38511533e-01 7.61809200e-02 6.56876385e-01 7.16898263e-01
-6.98611200e-01 -2.52170324e-01 -9.75934386e-01 6.57018125e-01
6.93249345e-01 -5.73587835e-01 -1.02180243e+00 1.10858910e-01
7.92254627e-01 -1.20860435e-01 -8.28919351e-01 6.90879345e-01
5.33723533e-01 -8.61320019e-01 1.05288196e+00 -9.35071409e-01
4.05985802e-01 -4.94337887e-01 -4.25696790e-01 -1.28684759e+00
-1.95830375e-01 3.59607232e-03 -2.77378440e-01 1.17183685e+00
1.01920986e+00 -6.42829716e-01 9.62766767e-01 7.32309580e-01
5.23064807e-02 -9.18943703e-01 -1.13794065e+00 -8.34494054e-01
4.17663194e-02 -4.98017073e-01 6.81639016e-01 1.00266349e+00
8.25723633e-02 4.62115079e-01 -3.51375073e-01 3.30234021e-01
1.77141219e-01 2.08344474e-01 6.42412961e-01 -1.36195958e+00
-5.22586346e-01 -6.19097874e-02 -5.45504928e-01 -9.18470383e-01
1.85117468e-01 -7.90364087e-01 -1.34493545e-01 -1.62200630e+00
3.08572322e-01 -8.32713306e-01 -3.40513468e-01 5.96720040e-01
-5.53802669e-01 2.89585348e-02 1.90523285e-02 3.95015441e-02
-7.76941836e-01 6.13196373e-01 8.16798866e-01 -2.51532644e-01
1.01897106e-01 -2.10242108e-01 -8.25387716e-01 4.23805118e-01
4.13359463e-01 -6.15555227e-01 -5.17870486e-01 -2.28631228e-01
9.28471923e-01 1.45276144e-01 6.18927777e-01 -8.06810915e-01
2.12089598e-01 -5.96258789e-02 1.81969896e-01 -1.70617998e-01
6.34199142e-01 -9.00216997e-01 2.37928405e-01 1.47307664e-01
-4.67169732e-01 -1.09825552e-01 -6.27539679e-03 7.65507162e-01
-4.39233869e-01 -5.00077069e-01 5.72999753e-02 1.07555777e-01
-8.67285490e-01 -1.66000739e-01 -2.55308375e-02 1.55948892e-01
6.41778409e-01 -8.96989331e-02 -6.15357876e-01 -5.15038252e-01
-8.43474925e-01 7.91583434e-02 3.28766614e-01 1.98861063e-01
6.08556628e-01 -1.32988048e+00 -8.31065416e-01 -1.07300617e-01
6.06168568e-01 -2.35512719e-01 -2.40368918e-02 7.38452673e-01
-2.03196898e-01 7.99475253e-01 2.94404477e-01 -6.25292864e-03
-1.11679626e+00 6.10762537e-01 2.33776912e-01 -7.99443841e-01
-1.12931378e-01 7.32427776e-01 6.42237440e-02 -5.99236727e-01
1.70255020e-01 -6.49741769e-01 -4.00007695e-01 7.79153332e-02
1.80667758e-01 1.52997896e-01 5.49487293e-01 -4.42165762e-01
-4.20134872e-01 4.47630882e-03 -3.23647141e-01 -1.23528801e-01
1.15585566e+00 2.61814028e-01 2.32678670e-02 4.13914233e-01
8.54937553e-01 6.18396759e-01 -6.54003322e-01 -3.18549216e-01
2.84615546e-01 -4.70514446e-01 -1.62435994e-01 -1.38443077e+00
-6.49745882e-01 4.18312401e-01 2.26304159e-01 2.45417401e-01
7.94345140e-01 2.37359583e-01 2.73393303e-01 7.27229595e-01
6.60433352e-01 -8.76585662e-01 -2.57836819e-01 4.90864217e-01
9.15143013e-01 -1.25737429e+00 4.42174636e-02 -4.84133095e-01
-7.12810874e-01 8.43473315e-01 5.11266112e-01 1.66149601e-01
3.76148134e-01 -4.42964248e-02 -1.43854260e-01 -3.83954853e-01
-1.07590950e+00 -3.21164012e-01 4.91384715e-01 4.73026067e-01
4.00346637e-01 2.68296242e-01 -4.67251927e-01 9.66654480e-01
-3.48101258e-01 -5.37446588e-02 5.59277773e-01 6.34466648e-01
-2.98506588e-01 -1.23884547e+00 -1.14322431e-01 5.84347963e-01
-4.37791467e-01 -4.56019640e-01 -4.70324069e-01 9.74956691e-01
2.34379530e-01 8.67614746e-01 -3.26904297e-01 -4.33452994e-01
4.64068860e-01 3.75982970e-01 3.66056293e-01 -7.60273516e-01
-3.01867843e-01 -6.10215306e-01 6.93534374e-01 -2.85733849e-01
-3.74572933e-01 -4.44735378e-01 -8.81901026e-01 2.21293494e-02
-6.75817013e-01 5.58265865e-01 4.48361665e-01 8.94585788e-01
4.53228861e-01 3.05988431e-01 -3.94650511e-02 1.00945681e-02
-6.44283593e-01 -8.40993226e-01 -5.17941475e-01 5.86391389e-01
2.69929945e-01 -9.47741508e-01 -2.62201190e-01 -6.29191697e-02] | [9.390646934509277, 8.341694831848145] |
e94cd481-1b20-4b6c-8382-7fd527514b76 | variable-selection-for-nonlinear-cox | 2211.09287 | null | https://arxiv.org/abs/2211.09287v1 | https://arxiv.org/pdf/2211.09287v1.pdf | Variable selection for nonlinear Cox regression model via deep learning | Variable selection problem for the nonlinear Cox regression model is considered. In survival analysis, one main objective is to identify the covariates that are associated with the risk of experiencing the event of interest. The Cox proportional hazard model is being used extensively in survival analysis in studying the relationship between survival times and covariates, where the model assumes that the covariate has a log-linear effect on the hazard function. However, this linearity assumption may not be satisfied in practice. In order to extract a representative subset of features, various variable selection approaches have been proposed for survival data under the linear Cox model. However, there exists little literature on variable selection for the nonlinear Cox model. To break this gap, we extend the recently developed deep learning-based variable selection model LassoNet to survival data. Simulations are provided to demonstrate the validity and effectiveness of the proposed method. Finally, we apply the proposed methodology to analyze a real data set on diffuse large B-cell lymphoma. | ['Kexuan Li'] | 2022-11-17 | null | null | null | null | ['variable-selection', 'survival-analysis'] | ['methodology', 'miscellaneous'] | [ 1.24494240e-01 -4.73297745e-01 -6.82933629e-01 -7.08392859e-01
-8.38155389e-01 2.71653742e-01 1.69054836e-01 6.27099216e-01
-5.14039397e-01 1.10288894e+00 1.19415475e-02 -2.83638388e-01
-2.51049966e-01 -8.81241143e-01 -2.41886958e-01 -1.15447855e+00
-3.66177708e-01 5.42510808e-01 -4.79333520e-01 -3.96421887e-02
1.56648532e-02 3.43496799e-01 -8.47518206e-01 -3.80128086e-01
7.09385097e-01 6.46919072e-01 -4.95315827e-02 2.60226697e-01
8.41961280e-02 4.56906140e-01 -3.23112041e-01 6.89424500e-02
-1.46944642e-01 -5.38167715e-01 -3.68683994e-01 -1.37206987e-01
-9.77617204e-02 -2.93254912e-01 -5.20703435e-01 6.79793894e-01
5.22815883e-01 3.07394750e-02 9.24787104e-01 -1.49452734e+00
-9.25158188e-02 6.46749258e-01 -6.70112133e-01 2.14649215e-01
-1.74861386e-01 -2.64925718e-01 9.54297602e-01 -7.38121450e-01
4.09688592e-01 1.09077513e+00 7.29371846e-01 2.98145026e-01
-1.41597283e+00 -9.45728540e-01 -6.41945004e-02 9.50418413e-02
-1.57621181e+00 -1.85983554e-01 6.74715459e-01 -4.51973796e-01
4.88091767e-01 1.28179446e-01 6.23323560e-01 9.08999324e-01
7.95366943e-01 6.26114666e-01 1.24787557e+00 -3.23865294e-01
1.71525061e-01 8.01912546e-02 6.20471537e-01 5.61406136e-01
6.35442436e-02 5.71276665e-01 -4.28055763e-01 -5.44869125e-01
6.85942471e-01 2.47229114e-01 -1.75170422e-01 -3.91608387e-01
-8.99496436e-01 1.37278581e+00 3.25944722e-01 8.68427828e-02
-3.94498110e-01 2.31803909e-01 5.90467572e-01 4.24417019e-01
5.29821277e-01 -6.56686053e-02 -2.92162001e-01 2.52656013e-01
-1.00776994e+00 3.34848315e-01 6.11311793e-01 7.80891836e-01
2.05647409e-01 -1.28578013e-02 -4.36007380e-01 6.08836412e-01
3.24723542e-01 3.72653663e-01 2.91120529e-01 -1.47935703e-01
4.01560292e-02 3.25498432e-01 -5.13890237e-02 -6.81449711e-01
-9.31370258e-01 -7.98437893e-01 -1.41596770e+00 2.16678843e-01
6.20244801e-01 -2.99669337e-02 -6.54664218e-01 1.83941686e+00
3.87380391e-01 1.65648803e-01 -8.83244798e-02 8.45550239e-01
6.77414894e-01 3.77488583e-01 4.60837007e-01 -8.70415807e-01
1.28968012e+00 -4.84227687e-01 -9.04959261e-01 2.98102468e-01
6.99343920e-01 -3.12857449e-01 7.36826241e-01 2.59865850e-01
-6.47462666e-01 -1.34586558e-01 -9.10908818e-01 5.06012402e-02
4.67001274e-02 -1.32437244e-01 9.98554170e-01 5.30773997e-01
-5.33636689e-01 3.62601548e-01 -1.00376832e+00 -3.46570253e-01
4.97576445e-01 5.43829679e-01 -2.67638326e-01 -1.18790001e-01
-1.59411752e+00 6.35214508e-01 1.70731694e-01 6.35073334e-02
-1.13857138e+00 -5.69359004e-01 -6.30372107e-01 2.24454388e-01
1.85684204e-01 -1.14609814e+00 8.53484154e-01 -9.10038114e-01
-9.31111991e-01 6.60797536e-01 -4.47172642e-01 -5.92086494e-01
7.58987367e-01 8.67861882e-02 -1.24572724e-01 -4.34949636e-01
-1.19027086e-01 -2.74259262e-02 6.76838040e-01 -6.90018415e-01
-6.72918200e-01 -7.45668769e-01 -2.26002842e-01 1.42816976e-01
-2.78992057e-01 1.96017355e-01 -1.23951495e-01 -7.94619918e-01
7.75984824e-02 -8.24799061e-01 -4.64344054e-01 9.74470470e-03
-3.61386776e-01 -3.32160473e-01 3.95894825e-01 -6.79526389e-01
1.39822304e+00 -2.17069936e+00 1.17235042e-01 2.69336402e-01
3.12034607e-01 -4.61624712e-01 2.65209734e-01 3.91260475e-01
-2.13915095e-01 3.89813492e-03 -4.10386622e-01 -4.10452634e-01
-4.65704560e-01 -1.66707292e-01 -7.32777268e-02 1.00862610e+00
-2.43417323e-02 5.97625732e-01 -6.31014645e-01 -7.61208117e-01
-5.65613993e-03 7.10333288e-01 -2.81016231e-01 3.30026150e-01
2.61720002e-01 7.76009738e-01 -5.86888671e-01 6.55135691e-01
6.15778804e-01 -1.61067829e-01 1.55670702e-01 2.10863397e-01
-1.39078721e-02 -1.51644602e-01 -9.00047898e-01 1.15262508e+00
-3.35191309e-01 5.21120310e-01 -1.09394036e-01 -1.25769079e+00
1.02056980e+00 3.92495155e-01 8.77552629e-01 -3.75660598e-01
3.38526666e-01 5.21554090e-02 5.49990907e-02 -4.39749986e-01
-8.44967440e-02 -9.23239350e-01 -1.88296646e-01 1.32607639e-01
-3.16285282e-01 1.80275977e-01 -3.15789767e-02 -2.04561979e-01
1.04579997e+00 -2.13903740e-01 9.83927965e-01 -1.31567419e-01
6.52155161e-01 2.33057246e-01 1.03306401e+00 7.80011654e-01
-2.88695782e-01 4.29227591e-01 6.48123384e-01 -4.50160563e-01
-8.38720739e-01 -8.70124996e-01 -8.75764072e-01 7.70461500e-01
-1.61434114e-01 3.54974568e-01 -1.96709439e-01 -5.71723759e-01
1.59455180e-01 6.96887136e-01 -8.36112797e-01 -9.07206535e-02
-3.82653147e-01 -1.30552351e+00 4.03577387e-01 4.05283034e-01
1.02979593e-01 -8.12767148e-01 -3.34979743e-01 2.36785233e-01
-2.22521555e-02 -3.34073156e-01 -2.76447624e-01 4.73415524e-01
-1.01066422e+00 -1.12620378e+00 -6.79892361e-01 -8.99905562e-01
6.69388771e-01 -7.33743757e-02 8.35750699e-01 1.83796674e-01
-1.62978590e-01 -2.28576973e-01 -1.97309539e-01 -3.13829064e-01
-3.58385444e-01 1.09458014e-01 -1.34550199e-01 -1.35032222e-01
5.66043556e-01 -1.31738812e-01 -4.91478235e-01 2.61312783e-01
-7.57620931e-01 -3.34630199e-02 4.53055978e-01 1.40829551e+00
8.81453276e-01 5.77830851e-01 9.29982305e-01 -1.19414711e+00
3.83516729e-01 -7.97540784e-01 -6.18537605e-01 1.82458773e-01
-8.72033238e-01 -2.39371598e-01 6.55448973e-01 -4.44770813e-01
-8.79811168e-01 1.72902554e-01 4.04780097e-02 -5.34293391e-02
1.34968536e-03 1.05354249e+00 -3.10906649e-01 7.36424625e-02
6.48534968e-02 2.04051271e-01 -1.83200017e-02 -1.70967817e-01
-2.07755476e-01 5.59953809e-01 7.10300580e-02 -5.13862595e-02
5.80568135e-01 6.59204304e-01 6.53005004e-01 -6.51630998e-01
-5.11976302e-01 -4.89692658e-01 -4.92269784e-01 1.67111889e-01
7.52323508e-01 -8.92144263e-01 -7.76856244e-01 5.19174695e-01
-7.10164011e-01 -7.87921548e-02 7.35454038e-02 8.71514261e-01
-7.91723907e-01 1.04259271e-02 -5.14210939e-01 -8.12215269e-01
-4.43327695e-01 -1.19176173e+00 6.60655856e-01 1.19183324e-01
-2.85442024e-01 -1.31259441e+00 2.48360977e-01 1.69792101e-02
2.79144734e-01 5.02129376e-01 1.46396303e+00 -8.46047997e-01
-1.12977810e-01 -6.26473069e-01 2.50395834e-02 -2.07175225e-01
1.65296942e-01 -1.52480334e-01 -4.78608340e-01 -7.08540380e-01
2.10281447e-01 -7.37344548e-02 8.50276113e-01 9.80155587e-01
1.11960173e+00 1.59687042e-01 -6.79735303e-01 9.58162665e-01
1.54768765e+00 4.00257915e-01 3.81160945e-01 4.46712762e-01
4.96927321e-01 6.78118348e-01 9.50050056e-01 7.24372923e-01
3.21176529e-01 7.89143205e-01 2.16161832e-01 -4.10015643e-01
4.79821414e-01 -9.58902612e-02 -5.47432341e-02 3.09601873e-01
3.11971724e-01 -4.48919117e-01 -9.56012130e-01 3.76102418e-01
-1.66301024e+00 -6.52715743e-01 -7.50156283e-01 2.50727153e+00
6.64646924e-01 6.75765052e-02 2.17739999e-01 1.59364894e-01
7.75412202e-01 -2.67664760e-01 -7.83401906e-01 -3.52087170e-02
-2.44242117e-01 6.92409724e-02 5.84758997e-01 4.76006061e-01
-1.37997997e+00 4.77632642e-01 6.31906319e+00 7.13769019e-01
-1.20647669e+00 1.18855890e-02 9.49169695e-01 9.76231247e-02
-1.20052166e-01 2.19834819e-01 -7.97762275e-01 2.63021946e-01
6.90589726e-01 -5.38967252e-01 -7.13903084e-02 5.13881028e-01
8.84905279e-01 1.54238949e-02 -1.20747900e+00 6.82939529e-01
-2.22562224e-01 -6.12693489e-01 -3.55437070e-01 1.21694915e-01
3.94766033e-01 -6.27217770e-01 1.61361322e-01 5.07573128e-01
1.26959741e-01 -1.33391166e+00 4.85859253e-02 6.45057380e-01
8.10368717e-01 -9.22629476e-01 1.07429683e+00 3.81454140e-01
-8.55024159e-01 -2.15182155e-01 -2.98565835e-01 8.37961957e-02
4.32978243e-01 7.11108327e-01 -1.06487167e+00 3.45814943e-01
2.96847612e-01 5.91101468e-01 -3.18232745e-01 1.30697083e+00
-6.81491047e-02 9.73230898e-01 2.60947761e-03 8.35397914e-02
-1.44096777e-01 -1.06746532e-01 4.65639710e-01 9.22114193e-01
3.47863406e-01 1.42982081e-01 9.89141688e-02 6.29638135e-01
2.17936769e-01 5.96186042e-01 -3.48316878e-01 1.07296698e-01
3.63685101e-01 9.77102220e-01 -7.15655088e-01 -7.36680254e-02
-6.55355513e-01 4.21270967e-01 8.16654488e-02 3.20884496e-01
-9.25816238e-01 -1.52391657e-01 3.75801533e-01 4.07763511e-01
-6.29052818e-02 -1.33403195e-02 -6.89938843e-01 -8.39625835e-01
-5.81499517e-01 -8.53472948e-01 8.46513569e-01 -1.57789245e-01
-1.44751787e+00 3.47993553e-01 -8.78035370e-03 -1.10915947e+00
-1.09075028e-02 4.06332267e-03 -5.74820936e-01 1.08735216e+00
-1.34950781e+00 -1.26167107e+00 -3.35345685e-01 5.73219895e-01
5.73686481e-01 -1.28390133e-01 8.49373281e-01 2.16167882e-01
-8.07855725e-01 7.36917436e-01 4.61665452e-01 -1.09801017e-01
9.00390804e-01 -1.16620576e+00 2.27431804e-02 3.43039274e-01
-5.66630006e-01 6.40364468e-01 9.46154118e-01 -9.21447694e-01
-1.27988029e+00 -1.08645487e+00 1.01726079e+00 -7.59973079e-02
5.28523803e-01 -1.96602061e-01 -1.16315579e+00 4.92232084e-01
-3.53503317e-01 -6.91933483e-02 8.84829879e-01 2.68365204e-01
2.12507680e-01 -1.81801245e-01 -1.32171893e+00 2.60863751e-01
3.99814069e-01 -4.47268151e-02 -1.85856566e-01 2.20371082e-01
4.76330727e-01 -1.07261911e-01 -9.98107731e-01 4.66038138e-01
7.53953755e-01 -4.52630609e-01 8.65595639e-01 -9.22090948e-01
4.32724476e-01 -5.76500073e-02 2.52451077e-02 -1.27498960e+00
-4.86866176e-01 -2.56667942e-01 1.73734099e-01 1.19858479e+00
2.52947152e-01 -6.13109946e-01 8.07876885e-01 6.29859149e-01
4.23460156e-01 -1.03459013e+00 -1.19122589e+00 -5.79230666e-01
4.44270015e-01 -1.85829461e-01 5.00199199e-01 1.05616498e+00
-5.93160242e-02 3.31304133e-01 -6.15109801e-01 7.35636353e-02
8.71578097e-01 1.40267059e-01 7.19938755e-01 -1.23937392e+00
-2.59788036e-01 -2.29252115e-01 -4.43651855e-01 -4.88997459e-01
4.05251384e-01 -8.46369565e-01 -2.71064471e-02 -1.30758691e+00
8.41730356e-01 -8.02104473e-01 -7.90649176e-01 1.00141346e-01
-5.93591392e-01 -3.50678772e-01 -3.12337816e-01 2.41122574e-01
2.65557580e-02 6.31374121e-01 9.13651049e-01 -1.93329960e-01
-4.07056987e-01 6.78020835e-01 -4.87492144e-01 4.47291195e-01
6.99028850e-01 -8.83860946e-01 -3.36526573e-01 1.69574872e-01
-4.85160500e-02 9.45414662e-01 2.32431501e-01 -3.99758518e-01
5.13474531e-02 -6.46230102e-01 3.67378980e-01 -6.10730112e-01
1.63827941e-01 -1.02082860e+00 4.49838370e-01 6.84862137e-01
-6.26057386e-01 1.01325348e-01 -2.76640058e-01 7.87172914e-01
-3.92895401e-01 -1.69319008e-02 9.21731651e-01 2.95979619e-01
-2.64537573e-01 6.19684279e-01 -3.76517862e-01 -2.67003566e-01
1.20378149e+00 8.24316144e-02 1.02664150e-01 -5.59872031e-01
-8.95660400e-01 5.41479111e-01 3.17403853e-01 1.75929651e-01
8.07307839e-01 -1.34583020e+00 -1.17047024e+00 2.20221072e-01
2.70032883e-01 -2.88660139e-01 2.70007759e-01 1.31982768e+00
-2.64996082e-01 4.65288103e-01 -5.20864613e-02 -4.20449436e-01
-1.61124563e+00 8.42161596e-01 3.48639876e-01 -5.26106358e-01
-4.80397493e-01 7.06357300e-01 5.09563565e-01 1.44868745e-02
2.66444474e-01 1.82622343e-01 -5.30382752e-01 1.09128304e-01
3.69060963e-01 5.50934494e-01 -1.71957120e-01 -5.77208996e-01
-4.31365013e-01 1.50746077e-01 -7.96919242e-02 1.55969150e-02
1.47023487e+00 -2.57779598e-01 -1.39810488e-01 6.92297518e-01
1.30491459e+00 -3.14936906e-01 -9.58891690e-01 -2.20055059e-01
3.55140083e-02 -5.37768066e-01 3.43582630e-01 -3.33553046e-01
-9.09508109e-01 7.72680342e-01 9.71274555e-01 -7.18167722e-02
1.29031324e+00 -2.91922331e-01 5.73731959e-01 -1.10911481e-01
4.31046128e-01 -6.87690318e-01 -5.37844300e-01 2.14963302e-01
6.75912738e-01 -1.44929469e+00 2.11215615e-01 -3.43017787e-01
-5.29392481e-01 1.14338100e+00 4.62181002e-01 -7.82165527e-02
8.75955284e-01 2.19181314e-01 7.87132904e-02 -1.55599162e-01
-7.41165876e-01 1.33079782e-01 -1.08269155e-01 4.37882870e-01
9.32160795e-01 3.27048331e-01 -1.08337367e+00 6.24917090e-01
-5.37133627e-02 2.15556115e-01 4.35260028e-01 5.96001208e-01
-8.31200033e-02 -1.08872235e+00 -5.68027496e-01 6.10814571e-01
-7.22055495e-01 -7.77557865e-02 -3.01902164e-02 9.92638648e-01
-1.88789964e-01 9.49317813e-01 -9.22469869e-02 1.87268376e-01
2.46053845e-01 6.52944110e-03 9.29894447e-02 -3.43452662e-01
-4.39201683e-01 3.00395727e-01 -1.68321982e-01 -5.85453361e-02
-6.85309544e-02 -9.29916441e-01 -1.28654039e+00 -3.27277929e-01
-5.85775256e-01 8.39366093e-02 5.89329839e-01 8.15145135e-01
-4.37056810e-01 6.44391179e-01 9.41033423e-01 -1.73919983e-02
-8.68955016e-01 -8.65648270e-01 -9.48777616e-01 3.44366968e-01
5.41541874e-01 -8.98304105e-01 -4.21151817e-01 -1.33369029e-01] | [7.775826454162598, 5.407729148864746] |
22616729-4dbd-4f4f-89f2-55746b8cae6a | counterfactual-learning-with-multioutput-deep | 2211.11119 | null | https://arxiv.org/abs/2211.11119v1 | https://arxiv.org/pdf/2211.11119v1.pdf | Counterfactual Learning with Multioutput Deep Kernels | In this paper, we address the challenge of performing counterfactual inference with observational data via Bayesian nonparametric regression adjustment, with a focus on high-dimensional settings featuring multiple actions and multiple correlated outcomes. We present a general class of counterfactual multi-task deep kernels models that estimate causal effects and learn policies proficiently thanks to their sample efficiency gains, while scaling well with high dimensions. In the first part of the work, we rely on Structural Causal Models (SCM) to formally introduce the setup and the problem of identifying counterfactual quantities under observed confounding. We then discuss the benefits of tackling the task of causal effects estimation via stacked coregionalized Gaussian Processes and Deep Kernels. Finally, we demonstrate the use of the proposed methods on simulated experiments that span individual causal effects estimation, off-policy evaluation and optimization. | ['Ioanna Manolopoulou', 'Gianluca Baio', 'Alberto Caron'] | 2022-11-20 | null | null | null | null | ['counterfactual-inference'] | ['miscellaneous'] | [ 8.84863734e-02 1.11251399e-01 -4.65929598e-01 -2.11057156e-01
-7.49217629e-01 -1.33537203e-01 9.46004272e-01 -1.62209481e-01
-3.92944574e-01 1.31769621e+00 7.33140767e-01 -5.68510175e-01
-6.68865561e-01 -6.72426820e-01 -9.29794431e-01 -6.41026318e-01
-5.76620817e-01 6.17830813e-01 -5.04262269e-01 5.30004203e-01
1.96135432e-01 6.67579770e-02 -9.52553391e-01 -2.16075063e-01
1.30392587e+00 3.19026887e-01 -2.72134900e-01 6.93433225e-01
4.82533306e-01 8.66973162e-01 -3.38409513e-01 -7.95821667e-01
1.30440906e-01 7.46801868e-02 -3.74557436e-01 -4.31903720e-01
7.21725702e-01 -6.21200502e-01 -5.40069699e-01 8.81661057e-01
7.59843469e-01 2.87837982e-01 1.22459042e+00 -1.29221559e+00
-9.12516952e-01 7.37702787e-01 -7.49019444e-01 2.03346744e-01
-2.32389331e-01 4.14759845e-01 8.38358939e-01 -5.89614630e-01
-1.39583098e-02 1.89653420e+00 6.98989451e-01 4.06204045e-01
-1.69128382e+00 -7.94268906e-01 4.64296371e-01 -6.02169223e-02
-7.44729698e-01 -3.34797800e-01 3.55578065e-01 -8.45911324e-01
5.61969101e-01 -1.59217611e-01 1.95104256e-01 1.87005377e+00
4.35551375e-01 6.98938131e-01 1.56052732e+00 -9.57416296e-02
5.14416635e-01 -4.15418178e-01 4.21489596e-01 4.86008286e-01
6.50905013e-01 9.06782508e-01 -3.88861388e-01 -9.08054411e-01
9.77908731e-01 1.16532013e-01 -2.32559323e-01 -4.65052575e-01
-1.20029044e+00 1.21394897e+00 -4.82533127e-03 -4.29621309e-01
-9.31870699e-01 8.38482440e-01 3.80943984e-01 -1.19623989e-01
7.58366883e-01 1.80508003e-01 -5.68755269e-01 1.86705831e-02
-6.25244021e-01 7.77482450e-01 6.86155856e-01 6.18180215e-01
1.30343810e-01 1.82652667e-01 -9.44223166e-01 6.06290460e-01
9.07303691e-02 9.88241732e-01 2.54464988e-02 -1.04282320e+00
5.88695347e-01 -1.75281361e-01 7.49376714e-01 -4.25575584e-01
-4.17474866e-01 -2.02647775e-01 -1.05532479e+00 1.04563400e-01
6.31645858e-01 -9.14345980e-01 -9.20428514e-01 2.20787644e+00
5.26913702e-01 9.92023349e-01 -7.98672140e-02 6.57865107e-01
1.41313389e-01 3.17997098e-01 8.49544346e-01 -3.16623896e-01
1.31894040e+00 -4.71956342e-01 -9.11821544e-01 9.46263224e-02
2.38153160e-01 -3.23600054e-01 1.04232812e+00 1.29720852e-01
-9.76902068e-01 -3.27912331e-01 -5.20857036e-01 2.01243341e-01
-1.10596493e-01 2.11479366e-01 1.18758357e+00 7.08481550e-01
-8.17699611e-01 7.13962555e-01 -8.42946589e-01 1.60732437e-02
7.17171311e-01 1.90018758e-01 1.98723793e-01 -1.56859104e-02
-1.51105988e+00 7.19886363e-01 1.43576860e-01 -2.09653646e-01
-1.56117368e+00 -1.78600943e+00 -5.91817439e-01 4.02503371e-01
6.54078722e-01 -1.35090113e+00 1.37004566e+00 -4.86040205e-01
-1.49489188e+00 1.96445867e-01 6.88288435e-02 -7.53152728e-01
8.48535955e-01 -6.10194743e-01 -3.13502222e-01 -3.34698200e-01
2.92788178e-01 6.98830262e-02 9.64500487e-01 -7.89082289e-01
-5.85883975e-01 -5.73365271e-01 8.86021778e-02 9.56875533e-02
5.84512874e-02 7.16700591e-03 1.90669358e-01 -9.24790621e-01
-1.05728805e+00 -8.99219334e-01 -5.02171278e-01 -5.10738313e-01
-5.79636633e-01 -3.03250343e-01 4.07772124e-01 -7.11397052e-01
1.01157963e+00 -1.96155834e+00 7.90691823e-02 -1.67114124e-01
3.22995245e-01 -9.62201059e-02 -4.55742516e-02 7.99623281e-02
-3.52819949e-01 3.01660337e-02 -3.37595999e-01 -1.58457890e-01
3.34126353e-01 -1.81110725e-01 -7.37690747e-01 5.86838841e-01
1.58783615e-01 1.14083052e+00 -9.27979887e-01 -1.27035126e-01
1.74819753e-01 3.03691089e-01 -5.08230507e-01 2.55616099e-01
-2.06903040e-01 3.35619241e-01 -7.06666946e-01 8.37561339e-02
7.45687187e-01 -2.20787138e-01 2.81703889e-01 2.18699221e-02
-1.45463303e-01 9.34890658e-02 -1.25467598e+00 1.23186409e+00
-6.72334492e-01 1.36139497e-01 -8.94137472e-02 -1.08959579e+00
-5.18752895e-02 5.37175536e-01 3.53066176e-01 -2.27832019e-01
-3.72912176e-02 -1.94821268e-01 -8.02228414e-03 -3.75626057e-01
5.15203699e-02 -5.79822540e-01 -2.88193911e-01 5.42418778e-01
1.58521712e-01 2.84607172e-01 -9.59897786e-02 -1.08086178e-02
9.97672379e-01 2.71515518e-01 5.89146018e-01 -5.42488575e-01
-8.05456713e-02 -3.18562835e-01 3.77376497e-01 1.37668371e+00
-1.40122816e-01 1.94996595e-02 9.01428223e-01 -3.22086632e-01
-9.43671584e-01 -1.68685174e+00 -2.87116230e-01 1.01844692e+00
-5.69942474e-01 3.82620871e-01 -5.72213233e-01 -7.41801858e-01
7.03129053e-01 1.17546439e+00 -1.27195239e+00 -1.54420406e-01
-3.29309732e-01 -1.69960988e+00 5.54947555e-01 5.77862740e-01
2.68450707e-01 -6.66362584e-01 -4.24377948e-01 5.03605939e-02
1.03323393e-01 -5.07620335e-01 -4.49356347e-01 -2.45214865e-01
-7.83971190e-01 -1.19657016e+00 -1.05714548e+00 2.86725670e-01
9.22057591e-03 -1.02203064e-01 1.10776842e+00 -9.49333787e-01
-1.69186950e-01 5.44560015e-01 5.10621667e-01 -8.15281510e-01
-1.48777753e-01 -4.21713293e-01 4.31523979e-01 7.79799148e-02
2.10360691e-01 -7.34896541e-01 -9.17438745e-01 -2.12033018e-01
-6.84439421e-01 7.03181848e-02 6.89319491e-01 8.94563556e-01
1.64841712e-01 -4.03759211e-01 8.84250164e-01 -1.22868848e+00
9.59459662e-01 -7.93229163e-01 -1.06736219e+00 4.36425477e-01
-5.38688898e-01 3.66266787e-01 4.49439406e-01 -7.13407815e-01
-1.83280230e+00 -5.22792220e-01 5.80397844e-01 -4.18543488e-01
-2.51365930e-01 3.61197501e-01 -1.58229798e-01 7.34987319e-01
5.25982916e-01 -4.12962139e-01 -1.23457707e-01 -3.88914078e-01
9.16971445e-01 3.26423645e-01 3.77117008e-01 -9.12865460e-01
4.28735405e-01 8.38950098e-01 3.00686270e-01 -2.56310135e-01
-1.12767422e+00 5.60906529e-02 -3.62518102e-01 1.99996367e-01
1.04746079e+00 -1.21069014e+00 -1.22429514e+00 4.01472867e-01
-9.74569738e-01 -9.12281096e-01 -1.80266246e-01 1.01039147e+00
-1.06480002e+00 5.81760937e-03 -5.84807098e-01 -1.18025076e+00
-1.27296478e-01 -8.19466054e-01 1.01704800e+00 -4.51312922e-02
2.97552757e-02 -1.45345855e+00 6.12285793e-01 5.04010506e-02
1.16299011e-01 2.55598158e-01 1.21746254e+00 -2.86371380e-01
-3.51255476e-01 2.40467727e-01 -3.94710392e-01 -8.47653598e-02
9.13975835e-02 -2.17395395e-01 -1.09043479e+00 -9.00484324e-02
-1.56906277e-01 -9.07684565e-02 1.10938168e+00 1.54238009e+00
1.40271497e+00 -3.88267577e-01 -4.80706304e-01 4.87492025e-01
1.29061997e+00 -1.89592056e-02 3.50799680e-01 -2.41717249e-01
7.68272102e-01 7.62985051e-01 5.17944813e-01 6.36362135e-01
3.99535149e-01 4.21253592e-01 2.06400871e-01 3.97954844e-02
1.82563022e-01 -6.12628758e-01 2.76245266e-01 -1.84402019e-01
-2.08333522e-01 -2.05861464e-01 -6.70107722e-01 6.69528961e-01
-2.21205068e+00 -1.28310597e+00 -4.64432538e-01 2.47307444e+00
7.38038719e-01 -2.97786057e-01 3.63290250e-01 -9.49042857e-01
9.42538679e-01 -6.70934841e-02 -8.29225004e-01 -9.08417925e-02
2.16876660e-02 3.75891477e-01 1.00075412e+00 5.19990265e-01
-1.45887673e+00 6.39514983e-01 7.06294632e+00 8.48801136e-01
-5.48791647e-01 5.52635789e-01 1.00216281e+00 -3.75138491e-01
-1.53736517e-01 -1.64028313e-02 -5.83091199e-01 6.39673889e-01
1.46361172e+00 -2.90003836e-01 2.61018574e-01 5.09826660e-01
8.43035161e-01 -2.50289679e-01 -1.17282498e+00 5.42862773e-01
-5.24094641e-01 -1.31142747e+00 -3.10228835e-03 1.53617039e-01
1.19343472e+00 3.27673256e-02 6.04508221e-01 5.68590403e-01
1.60098624e+00 -1.07754052e+00 5.04570663e-01 7.75708973e-01
8.66431236e-01 -7.83681810e-01 3.79588485e-01 5.90001494e-02
-2.54409194e-01 -4.08708483e-01 -4.21438098e-01 -3.05279613e-01
3.34050298e-01 8.85395586e-01 -4.28532451e-01 5.99723339e-01
4.77052957e-01 5.34114480e-01 1.47095934e-01 9.47851956e-01
-3.97576272e-01 8.86643112e-01 -8.72722175e-03 1.58828765e-01
1.19336449e-01 -2.60576725e-01 3.48615825e-01 1.18101227e+00
4.67160374e-01 1.16990402e-01 -2.57639289e-01 1.20675790e+00
-9.78277996e-02 -2.69536972e-01 -6.40602708e-01 2.08244100e-01
3.10712487e-01 7.19294786e-01 -5.91768324e-03 -7.51862228e-01
-5.98250270e-01 7.06749201e-01 5.13656735e-01 9.45638239e-01
-1.20431280e+00 3.49165976e-01 1.12099683e+00 -2.34474897e-01
9.85487103e-02 8.91054794e-02 -3.73778492e-01 -1.41799569e+00
-4.92801100e-01 -5.83540976e-01 7.65049219e-01 -4.51331377e-01
-1.84683132e+00 -7.54003584e-01 6.30233645e-01 -4.09183890e-01
-1.05014168e-01 -7.10416794e-01 -6.78719163e-01 1.32752931e+00
-1.45810211e+00 -9.22510922e-01 4.49929148e-01 4.35619950e-01
3.35627526e-01 3.22961628e-01 6.04336798e-01 7.47995973e-02
-7.31981158e-01 1.97263643e-01 4.64998871e-01 -2.16211557e-01
8.49933147e-01 -1.75334895e+00 4.66342241e-01 6.73682213e-01
-3.46734136e-01 8.34330261e-01 6.67119920e-01 -1.06296182e+00
-1.05111730e+00 -1.39101303e+00 1.86147436e-01 -6.84626400e-01
1.12874067e+00 -3.69384259e-01 -4.43635970e-01 9.04863715e-01
1.65805653e-01 -2.98657715e-01 7.89546788e-01 6.71050072e-01
-3.59257847e-01 2.88731873e-01 -1.11700511e+00 8.70562196e-01
1.01015842e+00 -2.70327061e-01 -6.64687812e-01 4.06019092e-01
8.08855951e-01 2.67551630e-03 -7.35024035e-01 3.35606545e-01
5.78293324e-01 -5.25713861e-01 1.23718226e+00 -1.42483318e+00
6.51580453e-01 1.08630426e-01 2.12131903e-01 -1.64795971e+00
-5.08858860e-01 -7.63669193e-01 -3.39689761e-01 8.05900991e-01
2.68223494e-01 -6.76358998e-01 3.58791858e-01 7.20461726e-01
3.20457786e-01 -1.47358656e-01 -1.01956105e+00 -5.53305924e-01
5.69944680e-01 -6.05428517e-01 6.08158350e-01 1.13392270e+00
-3.73659313e-01 3.25176477e-01 -8.03878486e-01 5.80688894e-01
1.26637781e+00 2.33290404e-01 6.60310447e-01 -1.29191947e+00
-7.11373150e-01 -3.85454804e-01 3.16405028e-01 -6.29351437e-01
5.22991478e-01 -2.98681170e-01 -1.78158417e-01 -9.95013654e-01
8.69626760e-01 -1.22001261e-01 -3.85011315e-01 -1.75208542e-02
-1.04355311e+00 -4.43154335e-01 -1.47935241e-01 -2.66451865e-01
-2.29201213e-01 7.72869408e-01 9.70250249e-01 -1.22037694e-01
-5.40498458e-03 3.04148942e-01 -6.58407152e-01 6.50557816e-01
6.80771172e-01 -3.88543397e-01 -6.95602119e-01 -2.17381015e-01
-6.04303647e-03 4.91535932e-01 1.05175209e+00 -2.45695308e-01
-3.97275090e-01 -7.28879929e-01 4.04400796e-01 -2.43190736e-01
9.57589224e-02 -3.61550629e-01 1.30045190e-01 5.62325358e-01
-6.62790895e-01 -7.00010806e-02 3.33791435e-01 1.21497607e+00
4.48117286e-01 3.31622154e-01 5.48896611e-01 -1.25002071e-01
-3.14487100e-01 5.22287607e-01 -3.85106236e-01 2.08105847e-01
1.05056632e+00 6.57896876e-01 -4.48932618e-01 -3.21948320e-01
-9.07633841e-01 4.26151246e-01 -5.37062101e-02 2.63100266e-01
2.96969473e-01 -1.27796948e+00 -1.07998979e+00 -4.17708218e-01
-3.20056193e-02 -6.30943537e-01 9.34413373e-01 1.09195018e+00
3.46770674e-01 6.11169279e-01 7.80953541e-02 -4.71976101e-02
-7.68041313e-01 1.10425127e+00 3.93723071e-01 -5.86528957e-01
-3.61133069e-01 5.24391592e-01 1.30427873e+00 -4.22002494e-01
-7.66316056e-02 -4.04150605e-01 -9.38872099e-02 -6.24288246e-02
6.06093526e-01 8.07393253e-01 -5.27491868e-01 1.44158721e-01
2.24692244e-02 -1.52332261e-01 1.32152930e-01 -4.09694433e-01
1.30646622e+00 -1.24832965e-01 -3.61257941e-02 6.31024361e-01
7.18539059e-01 -1.73243910e-01 -1.83353508e+00 8.08741245e-03
5.06523103e-02 -4.58028406e-01 1.02505058e-01 -1.03980887e+00
-4.34756130e-01 8.51913631e-01 7.13721991e-01 -6.52008429e-02
6.82817519e-01 -1.30796105e-01 -5.59482761e-02 -1.26464173e-01
1.25748262e-01 -8.55454862e-01 -2.63584197e-01 1.35988429e-01
6.49390697e-01 -1.12428486e+00 -1.58075184e-01 -9.85178724e-02
-4.72432524e-01 5.39888561e-01 3.02094847e-01 -3.90795648e-01
8.52893651e-01 1.80768982e-01 -4.41364437e-01 -3.14370096e-01
-1.02613342e+00 -2.21280634e-01 4.11702365e-01 8.59576523e-01
4.01795775e-01 7.55812824e-01 -3.84018093e-01 5.59353352e-01
1.32704526e-01 1.68684870e-01 5.69036722e-01 1.52001590e-01
1.40220076e-01 -6.92946672e-01 -5.36946952e-01 8.84796858e-01
-8.13617527e-01 -4.00981218e-01 5.35177141e-02 9.88519788e-01
-2.26080343e-01 6.60192966e-01 1.88717499e-01 5.23868024e-01
2.87253737e-01 1.04984917e-01 4.37358052e-01 -4.13232565e-01
-8.22032988e-02 2.79609524e-02 -3.49886082e-02 -5.49114585e-01
-4.51338679e-01 -9.72532034e-01 -4.71859813e-01 -4.96286601e-01
1.43657148e-01 -7.65955523e-02 1.99583963e-01 7.90908456e-01
3.23733270e-01 7.68332481e-01 2.29323894e-01 -5.53555310e-01
-1.21993661e+00 -1.19924116e+00 -6.46079719e-01 3.13402951e-01
6.43920243e-01 -1.12522197e+00 -2.86265671e-01 -1.43263713e-01] | [8.029699325561523, 5.375897407531738] |
9ec48bff-37f5-48f8-a050-42c31e2330ed | et-bert-a-contextualized-datagram | 2202.06335 | null | https://arxiv.org/abs/2202.06335v2 | https://arxiv.org/pdf/2202.06335v2.pdf | ET-BERT: A Contextualized Datagram Representation with Pre-training Transformers for Encrypted Traffic Classification | Encrypted traffic classification requires discriminative and robust traffic representation captured from content-invisible and imbalanced traffic data for accurate classification, which is challenging but indispensable to achieve network security and network management. The major limitation of existing solutions is that they highly rely on the deep features, which are overly dependent on data size and hard to generalize on unseen data. How to leverage the open-domain unlabeled traffic data to learn representation with strong generalization ability remains a key challenge. In this paper,we propose a new traffic representation model called Encrypted Traffic Bidirectional Encoder Representations from Transformer (ET-BERT), which pre-trains deep contextualized datagram-level representation from large-scale unlabeled data. The pre-trained model can be fine-tuned on a small number of task-specific labeled data and achieves state-of-the-art performance across five encrypted traffic classification tasks, remarkably pushing the F1 of ISCX-Tor to 99.2% (4.4% absolute improvement), ISCX-VPN-Service to 98.9% (5.2% absolute improvement), Cross-Platform (Android) to 92.5% (5.4% absolute improvement), CSTNET-TLS 1.3 to 97.4% (10.0% absolute improvement). Notably, we provide explanation of the empirically powerful pre-training model by analyzing the randomness of ciphers. It gives us insights in understanding the boundary of classification ability over encrypted traffic. The code is available at: https://github.com/linwhitehat/ET-BERT. | ['Jing Yu', 'Junzheng Shi', 'Zhen Li', 'Gaopeng Gou', 'Gang Xiong', 'Xinjie Lin'] | 2022-02-13 | null | null | null | null | ['traffic-classification'] | ['miscellaneous'] | [ 1.69242844e-01 -4.73089010e-01 -6.94801867e-01 -3.56626958e-01
-9.88420129e-01 -7.55526483e-01 4.36939031e-01 -3.60051364e-01
-4.03729826e-02 7.83382416e-01 -1.15710631e-01 -1.02659738e+00
-5.74143194e-02 -8.43333781e-01 -7.93593466e-01 -5.96125782e-01
-4.20903899e-02 4.19352919e-01 2.37294853e-01 -3.30786675e-01
1.10684931e-02 5.73190272e-01 -1.35247386e+00 5.47857523e-01
8.71666849e-01 1.42101717e+00 -1.38317078e-01 5.93451977e-01
-2.64321595e-01 7.48233318e-01 -6.83594048e-01 -6.67295814e-01
3.95504177e-01 1.73416138e-01 -7.14148462e-01 -2.84361511e-01
2.75988638e-01 -4.01005059e-01 -7.26748943e-01 7.42211998e-01
2.22820625e-01 -4.26320732e-01 4.53140587e-01 -1.65965879e+00
-6.54157579e-01 3.55546713e-01 -6.89353943e-01 4.69186574e-01
-3.35525602e-01 4.55222160e-01 1.03590572e+00 -4.48407710e-01
3.43897521e-01 8.98158669e-01 5.05253017e-01 5.83163977e-01
-1.23321426e+00 -1.33070290e+00 -9.58524719e-02 5.43596327e-01
-1.23374820e+00 -5.46156883e-01 5.27366757e-01 -3.44471186e-01
7.35410392e-01 4.64121819e-01 1.81519285e-01 1.60656703e+00
-1.30044118e-01 5.76832592e-01 1.06418157e+00 1.29621878e-01
-1.97583228e-01 4.44858432e-01 1.41905174e-01 3.51748854e-01
3.66447777e-01 3.06260586e-01 -1.09149188e-01 -4.99276854e-02
2.49980554e-01 1.98707849e-01 -1.58237338e-01 2.92946119e-02
-7.43064344e-01 7.40427256e-01 5.21797836e-01 1.45885035e-01
-2.01349959e-01 3.00579280e-01 7.82480776e-01 4.80296671e-01
4.03501421e-01 1.26002282e-01 -8.37869883e-01 -6.71487987e-01
-7.83813596e-01 -3.09126705e-01 6.93755746e-01 8.90307069e-01
8.54211628e-01 4.65780377e-01 7.27935210e-02 6.85949802e-01
-1.80593401e-01 8.89960766e-01 2.21799999e-01 -6.93969131e-01
8.80112767e-01 4.42952663e-01 -4.16072011e-01 -6.32974803e-01
8.75747502e-02 -7.24528909e-01 -6.56463742e-01 -1.13436036e-01
3.61648470e-01 -1.29647911e-01 -7.83550441e-01 1.75288010e+00
-5.87801412e-02 4.02843833e-01 1.11625455e-02 4.01752740e-01
4.84060854e-01 7.21030474e-01 1.61024272e-01 1.27276525e-01
1.27831459e+00 -6.31396592e-01 -2.84599513e-01 9.37459394e-02
6.68026388e-01 -7.35577643e-01 1.18366110e+00 1.26404956e-01
-5.73045254e-01 -5.70007443e-01 -8.23240876e-01 1.28160924e-01
-5.91609597e-01 1.37189239e-01 4.62133586e-01 1.11354113e+00
-6.94968939e-01 2.49249473e-01 -4.45978612e-01 -2.79193044e-01
9.98988271e-01 6.25983119e-01 -3.72181177e-01 -2.85462260e-01
-1.25358212e+00 3.40249330e-01 1.74056247e-01 -3.45322132e-01
-1.16506624e+00 -1.02952445e+00 -3.84741634e-01 2.42507383e-01
5.45538425e-01 -2.36690417e-01 1.00724816e+00 -7.91643381e-01
-1.35180402e+00 6.50999904e-01 -2.00205460e-01 -6.06612682e-01
3.29544216e-01 -7.79370740e-02 -8.75267029e-01 1.41287804e-01
1.26825616e-01 5.27697086e-01 8.64940524e-01 -1.27979851e+00
-7.01570511e-01 -1.87904269e-01 1.13024972e-01 -6.19590461e-01
-6.75248802e-01 3.31440493e-02 -3.28271538e-01 -5.84669650e-01
-6.79888785e-01 -9.89357173e-01 2.31272712e-01 -3.31587315e-01
-3.25156629e-01 -6.60393760e-02 1.63558376e+00 -5.59498549e-01
1.24870861e+00 -2.11357045e+00 -6.87979460e-01 2.15730190e-01
3.32517117e-01 8.69231701e-01 -4.05538499e-01 3.63031238e-01
-3.19563448e-01 5.09591341e-01 9.62104201e-02 7.45293349e-02
1.30156651e-01 4.83499020e-02 -8.19702744e-01 2.11977422e-01
4.97946471e-01 1.10002637e+00 -5.79297125e-01 -1.29107580e-01
3.88004720e-01 5.52730680e-01 -6.35096371e-01 1.55308977e-01
-4.38364074e-02 5.58822691e-01 -5.04820168e-01 7.45646894e-01
9.54225183e-01 -4.43191946e-01 8.09131414e-02 -5.03847897e-01
1.08416818e-01 5.71318746e-01 -5.93606055e-01 9.47757542e-01
-8.86294186e-01 1.05881929e+00 -1.80026069e-01 -1.20391023e+00
9.45701718e-01 2.20862731e-01 7.34177828e-01 -1.12881339e+00
2.29413807e-01 3.19807589e-01 1.48333862e-01 -3.72071207e-01
2.63796419e-01 7.24080652e-02 -5.44857085e-02 5.20522714e-01
2.01668255e-02 5.67356467e-01 1.51111931e-01 2.55666107e-01
1.15490282e+00 -4.92505312e-01 -2.19069436e-01 -5.65384142e-02
7.45926261e-01 -2.87292719e-01 6.17254138e-01 4.21173185e-01
-3.31345707e-01 2.62034476e-01 7.57597089e-01 -4.95442897e-01
-1.06308758e+00 -1.11063254e+00 -2.22169518e-01 1.12028551e+00
-3.04452237e-02 -6.78713739e-01 -5.29749990e-01 -1.03288770e+00
5.19999340e-02 5.43417931e-01 -4.52115923e-01 -3.01495641e-01
-7.80181527e-01 -5.82040489e-01 9.50664461e-01 5.29655874e-01
7.89191723e-01 -5.03167927e-01 8.80114958e-02 -4.25892919e-02
-3.34994555e-01 -1.82876468e+00 -2.82009095e-01 3.34109664e-02
-5.27090788e-01 -1.22447062e+00 -1.92985296e-01 -2.85369068e-01
2.45112240e-01 4.94742125e-01 1.07605696e+00 1.71562284e-01
-1.65062025e-01 2.04855636e-01 -2.56032735e-01 -1.84943303e-01
-4.83411372e-01 4.43051070e-01 1.40104637e-01 3.76139551e-01
5.26893139e-01 -7.03788817e-01 -5.20473778e-01 7.59797215e-01
-6.43836915e-01 -3.26509565e-01 6.52698934e-01 8.30321312e-01
2.73740739e-01 2.33198404e-01 6.73334002e-01 -7.79160082e-01
3.21940243e-01 -8.45340133e-01 -4.94361222e-01 1.54468447e-01
-5.18018425e-01 -1.43433899e-01 1.01107574e+00 -4.07190531e-01
-7.77846694e-01 -5.35709381e-01 -2.86655426e-01 -7.91318834e-01
-7.45477378e-02 -1.22061171e-01 -3.19291085e-01 -1.95139736e-01
4.34737176e-01 2.40268365e-01 -7.91632198e-03 -1.86678573e-01
2.21137270e-01 1.03571296e+00 2.22606793e-01 -8.61137688e-01
1.04766202e+00 6.34899616e-01 -5.53903319e-02 -7.42496848e-01
-5.41341782e-01 -1.81084991e-01 -1.49090886e-01 -5.68102784e-02
4.72243696e-01 -8.90801072e-01 -1.09184647e+00 1.97630659e-01
-7.56355107e-01 -4.43782628e-01 -3.56042504e-01 2.87709862e-01
-1.91438377e-01 4.13715601e-01 -6.61778390e-01 -5.68338513e-01
-4.72207040e-01 -1.52864039e+00 9.62747633e-01 -1.23989239e-01
3.81187983e-02 -9.13407683e-01 -4.19503480e-01 8.50290298e-01
8.40541184e-01 -4.51269485e-02 9.68234837e-01 -8.98496628e-01
-8.90888810e-01 -2.92075098e-01 -7.51447320e-01 8.66168797e-01
2.44382441e-01 2.47583032e-01 -1.14390266e+00 -3.22228253e-01
-3.13992679e-01 -3.88232708e-01 8.42828929e-01 7.21008331e-02
1.84311891e+00 -2.94850618e-01 -3.20073456e-01 8.06973577e-01
1.13006306e+00 7.87349939e-02 8.41428220e-01 1.67763278e-01
7.37802029e-01 4.31881428e-01 3.56608748e-01 2.71290660e-01
3.42657655e-01 7.16467202e-01 5.86933911e-01 7.34099820e-02
-2.59144127e-01 -3.15126717e-01 4.84563619e-01 6.79422557e-01
1.94529057e-01 -5.30399501e-01 -9.58622634e-01 3.25158507e-01
-1.16476607e+00 -1.11489975e+00 -2.42891237e-01 2.04779387e+00
3.93419892e-01 5.69804668e-01 3.12093079e-01 4.00081456e-01
8.39384437e-01 2.35099435e-01 -4.64044124e-01 -4.11840320e-01
-2.18087640e-02 5.97888470e-01 8.65139306e-01 2.54230976e-01
-9.38497305e-01 1.01160777e+00 5.14559984e+00 1.48588276e+00
-1.49679017e+00 1.78392559e-01 8.68135631e-01 2.06281245e-02
-3.23081374e-01 5.36492020e-02 -8.77465963e-01 1.11220920e+00
1.57508266e+00 -8.34564716e-02 4.24520820e-01 7.27660060e-01
2.17042252e-01 4.50912297e-01 -7.78328061e-01 1.18433332e+00
-3.24152678e-01 -1.45865500e+00 2.51550317e-01 6.47219956e-01
5.97259402e-01 3.73156339e-01 5.85147917e-01 7.50579536e-01
2.02930912e-01 -9.90610123e-01 2.50680357e-01 -1.27262492e-02
1.49100554e+00 -7.01289117e-01 6.13906801e-01 1.49306944e-02
-1.40807021e+00 -5.36950588e-01 -8.13142061e-02 3.98661822e-01
1.98674034e-02 5.87210774e-01 -8.53247702e-01 5.64546347e-01
6.95838392e-01 7.77851522e-01 -5.08912444e-01 5.32910049e-01
9.11548063e-02 1.21170390e+00 -3.19532454e-01 2.63647884e-01
3.03076684e-01 -3.30982208e-02 3.76400590e-01 1.16527987e+00
3.61416221e-01 -3.11911434e-01 -4.00832202e-03 7.67407358e-01
-6.16938710e-01 -1.77272141e-01 -4.85582620e-01 -2.06977293e-01
7.28134215e-01 1.33397377e+00 -2.26801366e-01 -3.33370566e-01
-3.84828478e-01 6.28686726e-01 1.28614813e-01 3.87664974e-01
-1.33679366e+00 -3.92630070e-01 1.18222928e+00 4.01782066e-01
5.35245895e-01 -1.76433511e-02 -1.98970601e-01 -1.09754670e+00
5.68849258e-02 -1.08584225e+00 3.07663828e-01 -4.03143376e-01
-1.38006043e+00 7.26557195e-01 -2.46206880e-01 -1.52246594e+00
1.06363609e-01 -8.79225969e-01 -7.40092158e-01 6.29839718e-01
-1.73265457e+00 -1.29537416e+00 -4.85867590e-01 7.20181584e-01
3.86982679e-01 -6.83739960e-01 5.04450023e-01 1.04359961e+00
-9.32643712e-01 1.21629965e+00 2.77773321e-01 3.02850574e-01
6.76095366e-01 -7.39289761e-01 5.20024598e-01 5.82759678e-01
4.63996781e-03 4.14609164e-01 1.25724405e-01 -2.29889065e-01
-1.32275879e+00 -1.38649857e+00 5.93078256e-01 -6.20817125e-01
9.91043627e-01 -4.79647249e-01 -8.67011428e-01 5.46367407e-01
-5.84545061e-02 4.94596958e-01 1.00117815e+00 -1.20925894e-02
-1.08652031e+00 -9.01614487e-01 -1.16036403e+00 3.70934933e-01
9.91351366e-01 -9.21011925e-01 1.94599688e-01 1.96961433e-01
7.51111031e-01 -2.96616573e-02 -8.59577477e-01 6.02751553e-01
5.28677106e-01 -9.22267556e-01 1.00384057e+00 -9.21439290e-01
9.51735452e-02 -6.66324347e-02 -5.36809444e-01 -8.25801790e-01
-1.86158746e-01 -7.88949132e-01 -3.72412980e-01 1.42071331e+00
3.61119598e-01 -1.01391542e+00 1.05241847e+00 1.37320980e-01
-7.70750046e-02 -9.15931463e-01 -9.21330810e-01 -1.04483402e+00
1.92529917e-01 -8.18780720e-01 9.49534774e-01 9.39592481e-01
-4.51994568e-01 3.07522357e-01 -3.53378952e-01 6.01646909e-03
4.87926155e-01 -1.08310379e-01 1.09465730e+00 -1.10486937e+00
1.55059602e-02 -4.18693662e-01 -6.81934178e-01 -1.14063931e+00
5.45071304e-01 -9.88001525e-01 -8.82294536e-01 -7.58575320e-01
5.90898022e-02 -8.24501514e-01 -5.93568504e-01 3.93971056e-01
1.66770890e-01 3.87160838e-01 2.56176621e-01 2.16247514e-01
-6.78484321e-01 6.86290145e-01 8.90636981e-01 -3.94616097e-01
2.97701836e-01 2.31948614e-01 -9.93547022e-01 9.42115188e-02
1.39435208e+00 -4.85337585e-01 -5.79303265e-01 -3.31512392e-01
-2.48428702e-01 -3.24219376e-01 5.11771500e-01 -1.08801746e+00
-1.80510849e-01 -4.86864969e-02 1.14635430e-01 -4.97939587e-01
3.79612923e-01 -1.02602422e+00 -1.12288250e-02 3.88972461e-01
-2.47014724e-02 7.13550672e-02 3.28505516e-01 5.35012305e-01
-1.46385342e-01 3.29200923e-01 6.91728652e-01 3.46731514e-01
-6.72420025e-01 7.96027422e-01 -2.03324407e-01 5.02077937e-01
1.13362217e+00 -3.71860385e-01 -8.15446973e-01 -5.35247266e-01
-1.07488468e-01 1.34578198e-01 2.22081780e-01 5.82773626e-01
3.88141870e-01 -1.39832127e+00 -6.38581395e-01 4.16234344e-01
2.15761706e-01 -7.44277894e-01 3.63062024e-01 8.07999194e-01
-3.28722209e-01 7.57823825e-01 -2.82346755e-01 -7.45211303e-01
-1.16715229e+00 5.25891721e-01 3.36874366e-01 -3.73832911e-01
-1.38717040e-01 4.01022971e-01 1.30250871e-01 -4.11804825e-01
6.07493147e-02 -6.76573440e-02 2.96192557e-01 -3.01199377e-01
3.90139520e-01 6.86610579e-01 2.59172469e-01 -8.23246360e-01
-4.48095739e-01 5.33459127e-01 -2.89699465e-01 3.82114679e-01
1.09876072e+00 3.64465341e-02 1.23837277e-01 -1.93982914e-01
1.94015324e+00 1.46578431e-01 -1.07867408e+00 -2.40060970e-01
-3.11618924e-01 -7.07304955e-01 -2.21272901e-01 -6.22886956e-01
-1.65134454e+00 1.31962311e+00 6.96948588e-01 2.74293065e-01
1.05280864e+00 -1.80240974e-01 1.31757677e+00 3.53850782e-01
3.12433988e-01 -6.20894015e-01 2.95544446e-01 5.28383791e-01
3.62149715e-01 -1.36312056e+00 -3.86660486e-01 -4.81368065e-01
-5.53183317e-01 1.06068873e+00 7.60742426e-01 1.08055316e-01
8.25839221e-01 3.15985858e-01 -8.39133859e-02 -8.82135630e-02
-9.95422542e-01 -2.92844743e-01 1.06160399e-02 5.86308122e-01
2.13615298e-01 1.90679014e-01 2.93127745e-01 3.91551465e-01
-1.99281350e-01 -1.51157543e-01 1.55774474e-01 5.46476364e-01
-5.31184524e-02 -1.43016100e+00 -2.27698069e-02 6.18091941e-01
-7.13992178e-01 -7.89288282e-02 -1.44983545e-01 8.39109123e-01
1.97480008e-01 1.09609556e+00 1.06018424e-01 -1.04533851e+00
2.40165979e-01 -1.77484348e-01 5.45882583e-02 -9.58498716e-02
-4.20493543e-01 -4.75499600e-01 2.83861101e-01 -5.50079167e-01
8.89892131e-02 -2.21561283e-01 -8.58640790e-01 -1.16868520e+00
-2.90265083e-01 2.34446615e-01 4.69625831e-01 5.08380949e-01
9.48562205e-01 4.96045351e-01 1.15982842e+00 -2.63032317e-01
-4.13452536e-01 -5.09336174e-01 -1.52044490e-01 5.30935705e-01
3.83460313e-01 -6.59125686e-01 -4.36914086e-01 -3.05552632e-01] | [5.0724992752075195, 7.244496822357178] |
1f25364d-dedc-41fa-baf1-f173991a3353 | structure-guided-multi-modal-pre-trained | 2307.03591 | null | https://arxiv.org/abs/2307.03591v1 | https://arxiv.org/pdf/2307.03591v1.pdf | Structure Guided Multi-modal Pre-trained Transformer for Knowledge Graph Reasoning | Multimodal knowledge graphs (MKGs), which intuitively organize information in various modalities, can benefit multiple practical downstream tasks, such as recommendation systems, and visual question answering. However, most MKGs are still far from complete, which motivates the flourishing of MKG reasoning models. Recently, with the development of general artificial architectures, the pretrained transformer models have drawn increasing attention, especially for multimodal scenarios. However, the research of multimodal pretrained transformer (MPT) for knowledge graph reasoning (KGR) is still at an early stage. As the biggest difference between MKG and other multimodal data, the rich structural information underlying the MKG still cannot be fully leveraged in existing MPT models. Most of them only utilize the graph structure as a retrieval map for matching images and texts connected with the same entity. This manner hinders their reasoning performances. To this end, we propose the graph Structure Guided Multimodal Pretrained Transformer for knowledge graph reasoning, termed SGMPT. Specifically, the graph structure encoder is adopted for structural feature encoding. Then, a structure-guided fusion module with two different strategies, i.e., weighted summation and alignment constraint, is first designed to inject the structural information into both the textual and visual features. To the best of our knowledge, SGMPT is the first MPT model for multimodal KGR, which mines the structural information underlying the knowledge graph. Extensive experiments on FB15k-237-IMG and WN18-IMG, demonstrate that our SGMPT outperforms existing state-of-the-art models, and prove the effectiveness of the designed strategies. | ['Xinwang Liu', 'Meng Liu', 'Lingyuan Meng', 'Yue Liu', 'Sihang Zhou', 'Ke Liang'] | 2023-07-06 | null | null | null | null | ['visual-question-answering-1', 'knowledge-graphs', 'recommendation-systems', 'question-answering'] | ['computer-vision', 'knowledge-base', 'miscellaneous', 'natural-language-processing'] | [-6.35684803e-02 2.49745086e-01 -3.29924107e-01 -1.19392946e-01
-3.50266039e-01 -4.03729826e-01 5.78613758e-01 1.34994105e-01
5.08946879e-03 3.53637397e-01 4.00838673e-01 -2.72788316e-01
-2.81729668e-01 -8.61113667e-01 -6.96543515e-01 -7.62417555e-01
5.25810003e-01 2.78872252e-01 2.06045166e-01 -4.67233777e-01
5.83812483e-02 -2.59179566e-02 -1.29050958e+00 5.60742199e-01
9.06498611e-01 1.23542058e+00 3.28427851e-01 2.19681770e-01
-6.54523313e-01 9.81644452e-01 -7.43057504e-02 -1.03414798e+00
-2.44032502e-01 -3.69603395e-01 -7.69496560e-01 1.32770568e-01
3.11363429e-01 -2.99859971e-01 -9.90198255e-01 1.04234171e+00
4.16913480e-01 1.02854431e-01 5.22889316e-01 -1.29686749e+00
-1.16810107e+00 9.83394444e-01 -6.37368739e-01 -1.33747831e-01
4.54206556e-01 1.04821719e-01 1.32266963e+00 -9.79050875e-01
6.03156090e-01 1.46532953e+00 3.20071220e-01 3.26978087e-01
-5.80750346e-01 -4.11530286e-01 5.22078454e-01 5.00679255e-01
-1.38490391e+00 -2.18411505e-01 9.81406212e-01 -1.24989755e-01
7.35240042e-01 2.16535792e-01 7.42967963e-01 9.15303528e-01
-2.70338327e-01 1.39538717e+00 8.12503159e-01 -2.13258237e-01
-2.28217915e-01 1.59739349e-02 5.46144247e-02 1.19500315e+00
7.23159835e-02 -4.65320230e-01 -7.60582149e-01 1.88054547e-01
7.40489483e-01 1.84735894e-01 -5.54197490e-01 -4.55292225e-01
-1.33993161e+00 5.70295751e-01 7.56006062e-01 2.70162761e-01
-2.48923883e-01 7.06249028e-02 3.41599584e-01 2.46048853e-01
1.23054527e-01 -6.73841238e-02 -2.15197846e-01 2.23197356e-01
-4.66870517e-01 -1.01244047e-01 5.38552999e-01 1.20857751e+00
7.84608424e-01 -2.08673686e-01 -3.14386278e-01 8.76270354e-01
7.67966628e-01 7.20073164e-01 3.46763551e-01 -5.87941647e-01
8.54422390e-01 1.37819171e+00 -3.31305355e-01 -1.50198066e+00
-2.55760223e-01 -3.16863924e-01 -9.00659382e-01 -6.15828693e-01
1.67718008e-01 1.74238071e-01 -9.95882571e-01 1.60094428e+00
3.60116750e-01 8.04257542e-02 2.92219609e-01 1.11157799e+00
1.62108517e+00 5.66638470e-01 -3.36425975e-02 -2.11176053e-02
1.43788731e+00 -1.16326106e+00 -7.82245696e-01 -1.88008070e-01
4.40037906e-01 -7.47068226e-01 1.04525554e+00 3.60747166e-02
-8.51719677e-01 -3.98285925e-01 -8.28374207e-01 -3.26136738e-01
-6.30461454e-01 3.08351159e-01 8.80797029e-01 3.33937883e-01
-7.80965269e-01 -5.18808179e-02 -4.45869386e-01 -5.12647927e-01
4.92982864e-01 9.79546923e-04 -5.07186353e-01 -6.71346188e-01
-1.38085699e+00 6.05396211e-01 7.12586820e-01 6.85149729e-01
-6.84494615e-01 -2.47681811e-01 -9.03610349e-01 8.52100104e-02
7.69734502e-01 -9.39695597e-01 8.08827341e-01 -6.17265165e-01
-1.21251917e+00 6.87019646e-01 -1.74436662e-02 6.03758078e-03
3.22687477e-01 -1.37203317e-02 -6.37954116e-01 3.04721326e-01
-1.87450632e-01 6.25889838e-01 9.51608062e-01 -1.44493222e+00
-4.67063129e-01 -4.52379972e-01 4.69081670e-01 6.18371129e-01
-4.51842576e-01 -3.89116943e-01 -1.20607650e+00 -4.58000243e-01
2.57007480e-01 -6.18908823e-01 1.59213737e-01 -3.55054915e-01
-6.89331114e-01 -2.91191667e-01 6.97290897e-01 -8.51698399e-01
1.40256405e+00 -2.04006553e+00 6.43858552e-01 3.32975656e-01
5.09549975e-01 2.09204152e-01 -2.50106424e-01 7.16334879e-01
1.94260314e-01 -1.97516312e-03 -7.01152533e-02 -3.15321803e-01
3.46249253e-01 4.49253619e-01 -3.72240901e-01 5.97292483e-02
1.68333594e-02 1.54954624e+00 -9.80462790e-01 -8.49314809e-01
1.92334056e-01 4.11400378e-01 -1.92550033e-01 2.57320732e-01
-3.99714768e-01 2.12243080e-01 -7.85099208e-01 1.03588808e+00
6.28990233e-01 -5.97561538e-01 2.51394480e-01 -9.35124099e-01
2.34394625e-01 -3.01459163e-01 -8.71746123e-01 2.06417155e+00
-6.93734214e-02 2.04120651e-01 -8.50778446e-02 -1.07635355e+00
8.50357115e-01 1.63608864e-01 3.13839287e-01 -9.85720873e-01
2.21097380e-01 1.97484568e-01 -2.65388370e-01 -7.25721240e-01
5.62765956e-01 1.95891391e-02 -6.36688471e-02 1.25487342e-01
2.46791825e-01 2.36277640e-01 9.77297127e-02 7.42071390e-01
9.66417253e-01 2.70268202e-01 -3.38574313e-02 2.59002000e-01
7.30470359e-01 -1.31557435e-01 2.71473318e-01 4.56146508e-01
8.97941068e-02 3.41903746e-01 4.90797490e-01 -5.82937784e-02
-5.52278876e-01 -9.86663699e-01 4.06593353e-01 1.01294971e+00
7.66239285e-01 -6.24345481e-01 -6.04322493e-01 -8.88708651e-01
1.18748516e-01 3.35401982e-01 -6.01628602e-01 -4.72804636e-01
-4.28803951e-01 -5.94980121e-01 6.18723691e-01 4.95848775e-01
8.18387389e-01 -1.05241954e+00 -3.79517972e-02 -8.74667168e-02
-5.15370429e-01 -1.36490488e+00 -4.43397313e-01 -3.90475214e-01
-6.50490880e-01 -1.28572679e+00 -7.44965851e-01 -7.80550361e-01
7.62938082e-01 6.01821303e-01 1.11727846e+00 4.23878849e-01
7.13486746e-02 9.76560056e-01 -7.85708725e-01 -5.88490069e-03
6.68101981e-02 9.11321566e-02 -4.17011797e-01 3.57085824e-01
2.86758095e-01 -4.87117201e-01 -7.66092300e-01 3.84552151e-01
-1.02335763e+00 5.60620487e-01 1.18016779e+00 7.30252504e-01
6.71747625e-01 1.39970779e-01 4.60124552e-01 -7.39739180e-01
5.60300291e-01 -5.57443798e-01 -2.15898305e-01 9.57166731e-01
-4.00621086e-01 1.45379186e-01 5.09754956e-01 -3.06694627e-01
-1.20892835e+00 -3.64832193e-01 1.79826133e-02 -7.67128527e-01
1.66591734e-01 1.09887636e+00 -6.59616649e-01 -1.50584742e-01
-5.71726076e-02 4.99018967e-01 -1.59257844e-01 -5.31398773e-01
7.99498141e-01 4.25074220e-01 5.48787892e-01 -6.97488368e-01
8.58961403e-01 3.83247316e-01 2.03674622e-02 -5.88042259e-01
-9.15299833e-01 -4.02400851e-01 -2.19077766e-01 -3.50824982e-01
8.32225442e-01 -8.80807102e-01 -9.20849800e-01 5.37123799e-01
-9.60674524e-01 -6.23706281e-02 1.48681775e-01 2.97426492e-01
-2.89057732e-01 7.02107012e-01 -5.33162713e-01 -7.65104711e-01
-3.51671666e-01 -1.03324938e+00 1.29977632e+00 4.91636515e-01
6.05242252e-01 -9.86697853e-01 -4.12257433e-01 9.92259800e-01
2.37616867e-01 6.41830638e-02 1.11589730e+00 -4.44501966e-01
-1.01895463e+00 1.67137273e-02 -7.98465669e-01 8.13955590e-02
-1.09596169e-02 -2.53401875e-01 -7.63861060e-01 -1.43896013e-01
-4.66172397e-01 -5.86553931e-01 1.19270527e+00 -5.94354570e-02
1.07515633e+00 -1.54397696e-01 -4.73283678e-01 4.75737214e-01
1.26969516e+00 6.33857772e-02 6.96881115e-01 1.12649612e-01
1.46120763e+00 5.33018410e-01 6.62175596e-01 1.89087451e-01
9.85650599e-01 5.04898369e-01 7.55076110e-01 -2.27297880e-02
-2.77620941e-01 -7.24251032e-01 2.19881162e-01 1.30092216e+00
-2.84269363e-01 -5.08697391e-01 -7.81865001e-01 4.41134006e-01
-2.30485988e+00 -7.78296828e-01 1.65290590e-02 1.77698421e+00
4.77797240e-01 -2.03524068e-01 -2.93903798e-01 -1.64016128e-01
6.83571100e-01 4.04314339e-01 -6.70018017e-01 2.93969840e-01
-3.20189387e-01 -3.47911209e-01 1.58645168e-01 1.02172300e-01
-8.94982398e-01 1.00100040e+00 3.99783921e+00 1.17692637e+00
-6.26869619e-01 -1.68013826e-01 1.89909324e-01 3.89181316e-01
-7.14210153e-01 1.16457179e-01 -6.62277520e-01 3.78952146e-01
2.16126427e-01 -2.47269981e-02 7.00134158e-01 4.36775744e-01
-2.69850969e-01 1.81666434e-01 -7.83245146e-01 1.42461705e+00
3.59215796e-01 -1.12917733e+00 5.33492327e-01 2.13515535e-02
3.62880856e-01 -2.94547021e-01 1.21666282e-01 6.43514812e-01
8.27861950e-02 -1.02614987e+00 6.19413495e-01 8.79991055e-01
5.61745644e-01 -7.44277358e-01 8.78252804e-01 2.06561968e-01
-1.73748624e+00 1.18640950e-02 -4.33905691e-01 5.35538256e-01
2.77755380e-01 5.15810549e-01 -3.66103142e-01 1.42923725e+00
5.53714395e-01 9.85725343e-01 -8.80029142e-01 7.56217420e-01
-3.85683864e-01 5.25534570e-01 -1.16221040e-01 5.93452118e-02
3.67088765e-01 -2.35005468e-01 3.64441127e-01 8.95176530e-01
3.08256298e-01 1.60172239e-01 2.75144249e-01 6.86521292e-01
-2.82531291e-01 3.60328078e-01 -5.09788275e-01 -5.38913548e-01
2.97162354e-01 1.52597642e+00 -5.94939530e-01 -1.45854399e-01
-7.44576931e-01 9.55817461e-01 6.51013076e-01 6.10565484e-01
-9.36286569e-01 -2.28379488e-01 1.73094407e-01 -3.53107840e-01
4.70547467e-01 -1.37747437e-01 2.59171218e-01 -1.61636889e+00
1.15759566e-01 -9.81768310e-01 8.33013535e-01 -1.05107808e+00
-1.51912308e+00 4.32305843e-01 2.20111702e-02 -1.08050168e+00
3.20044398e-01 -6.41626894e-01 -2.43212521e-01 6.41181171e-01
-1.70730078e+00 -1.88239014e+00 -6.29300177e-01 9.71488833e-01
1.44804180e-01 -1.84939921e-01 4.48470145e-01 3.76765013e-01
-7.09162235e-01 6.52675688e-01 -2.25272089e-01 4.22670752e-01
5.13052940e-01 -1.12760079e+00 -5.61733246e-02 6.03892922e-01
3.19773316e-01 8.28517139e-01 1.98227912e-01 -6.81394458e-01
-2.17581034e+00 -1.03798461e+00 3.52405518e-01 -2.06610978e-01
8.22838068e-01 -1.46004096e-01 -1.05659020e+00 6.80899918e-01
2.98748851e-01 -1.39427379e-01 6.00217283e-01 2.04789326e-01
-6.46969438e-01 -1.49004221e-01 -6.34673119e-01 7.79263914e-01
1.36531293e+00 -7.13603914e-01 -6.65523827e-01 2.44259968e-01
9.61058259e-01 -5.26257932e-01 -1.16219985e+00 5.58489978e-01
5.19541860e-01 -7.46650696e-01 1.06649733e+00 -5.88060737e-01
4.25335258e-01 -5.67788601e-01 -4.16582525e-01 -1.13969731e+00
-8.42502266e-02 -2.94348866e-01 -5.19142389e-01 1.60193169e+00
2.91850328e-01 -5.71673393e-01 5.93801022e-01 2.90789723e-01
-3.01982641e-01 -8.01372111e-01 -6.94428146e-01 -3.54286164e-01
-4.49662328e-01 -2.81508148e-01 6.97959185e-01 1.10495532e+00
7.35730007e-02 7.62975752e-01 -5.03482282e-01 2.46313766e-01
5.59453011e-01 6.04968786e-01 7.29210436e-01 -9.28708136e-01
-3.47818911e-01 -4.86438543e-01 -4.86692101e-01 -1.35313225e+00
2.15177342e-01 -1.01690614e+00 -2.87815213e-01 -2.02583337e+00
4.76569414e-01 -1.52870208e-01 -4.06616122e-01 7.61858582e-01
-5.23636639e-01 -2.90261488e-02 2.76104271e-01 7.64152631e-02
-1.12267900e+00 8.71429086e-01 1.70054960e+00 -4.14645076e-01
2.52948910e-01 -6.91022098e-01 -8.31016421e-01 5.98781466e-01
5.17834008e-01 1.84124798e-01 -9.33648825e-01 -6.06005847e-01
6.84877336e-01 2.16585159e-01 5.72172165e-01 -5.92614949e-01
5.23270965e-01 -2.69508243e-01 2.38691226e-01 -7.87639856e-01
4.02252585e-01 -1.00056279e+00 2.08686426e-01 -1.10123321e-01
-6.66955113e-02 -4.29739766e-02 2.50260010e-02 8.66239309e-01
-5.30647218e-01 1.05208486e-01 1.56209424e-01 -1.82863399e-01
-1.19562817e+00 6.59316897e-01 3.38166207e-01 1.04641378e-01
5.92454910e-01 -1.67765114e-02 -8.58191252e-01 -3.27273458e-01
-5.15302539e-01 7.09858775e-01 2.45978802e-01 5.05368114e-01
9.64388490e-01 -1.63570523e+00 -4.45623636e-01 -1.59303963e-01
5.88661313e-01 1.48454875e-01 8.01243782e-01 1.12147212e+00
-7.66606554e-02 4.01933968e-01 1.60872996e-01 -3.26663285e-01
-1.18034065e+00 8.44102859e-01 1.19506367e-01 -3.92206132e-01
-6.31778657e-01 6.85962021e-01 3.92723382e-01 -4.33294654e-01
2.93203164e-02 -1.13165341e-01 -5.66104591e-01 1.25552297e-01
3.40481132e-01 5.52635416e-02 -1.74551278e-01 -8.69790196e-01
-2.53746331e-01 7.76973963e-01 -1.26254454e-01 1.71656147e-01
9.45339680e-01 -4.27880019e-01 -3.16526741e-01 4.06616688e-01
1.03882420e+00 -6.85102865e-02 -6.61075294e-01 -5.35554647e-01
-2.47177124e-01 -4.04666156e-01 1.45095848e-02 -7.77138174e-01
-1.43643665e+00 9.65014338e-01 1.06275961e-01 1.23609506e-01
1.26376879e+00 2.88652599e-01 1.06227160e+00 6.28837287e-01
4.79335129e-01 -8.14841807e-01 2.68895924e-01 3.52825761e-01
7.69988239e-01 -1.21592033e+00 -7.39041045e-02 -6.54763341e-01
-7.98479438e-01 9.51915324e-01 7.54272938e-01 5.09156644e-01
3.34664762e-01 -4.49986249e-01 -1.47028446e-01 -6.73941255e-01
-5.49982011e-01 -6.62787497e-01 7.53141284e-01 3.65175098e-01
1.96907252e-01 9.06592757e-02 -1.87307417e-01 8.11470628e-01
1.91566452e-01 -2.08046585e-02 -1.70695826e-01 7.64632940e-01
-3.97262633e-01 -9.57979262e-01 -7.93941244e-02 4.18492645e-01
-2.33283993e-02 -2.48313159e-01 -5.43208361e-01 7.62593746e-01
1.84215903e-01 9.57529068e-01 -5.81579566e-01 -6.47178710e-01
4.19179618e-01 -9.14030671e-02 4.94641751e-01 -1.72776222e-01
-1.55503258e-01 -3.17810699e-02 1.91840604e-01 -5.15801966e-01
-7.64063776e-01 -2.52028286e-01 -1.27271819e+00 -3.35547596e-01
-3.44625324e-01 2.15899825e-01 1.59099638e-01 1.02248585e+00
2.64042467e-01 5.34040928e-01 2.74976790e-01 -3.36758435e-01
-1.88095812e-02 -7.80005097e-01 -5.66091001e-01 5.56340575e-01
-9.92124379e-02 -8.78567457e-01 -1.04621492e-01 -1.51124701e-01] | [10.63102912902832, 1.8644695281982422] |
90ce792f-2794-46f7-8b5d-e00112d92a2a | chatgpt-needs-spade-sustainability-privacy | 2305.03123 | null | https://arxiv.org/abs/2305.03123v1 | https://arxiv.org/pdf/2305.03123v1.pdf | ChatGPT Needs SPADE (Sustainability, PrivAcy, Digital divide, and Ethics) Evaluation: A Review | ChatGPT is another large language model (LLM) inline but due to its performance and ability to converse effectively, it has gained a huge popularity amongst research as well as industrial community. Recently, many studies have been published to show the effectiveness, efficiency, integration, and sentiments of chatGPT and other LLMs. In contrast, this study focuses on the important aspects that are mostly overlooked, i.e. sustainability, privacy, digital divide, and ethics and suggests that not only chatGPT but every subsequent entry in the category of conversational bots should undergo Sustainability, PrivAcy, Digital divide, and Ethics (SPADE) evaluation. This paper discusses in detail about the issues and concerns raised over chatGPT in line with aforementioned characteristics. We support our hypothesis by some preliminary data collection and visualizations along with hypothesized facts. We also suggest mitigations and recommendations for each of the concerns. Furthermore, we also suggest some policies and recommendations for AI policy act, if designed by the governments. | ['Kapal Dev', 'Parus Khuwaja', 'Sunder Ali Khowaja'] | 2023-04-13 | null | null | null | null | ['ethics'] | ['miscellaneous'] | [-3.21408093e-01 4.53560591e-01 -1.39661446e-01 -1.05623029e-01
-5.43463156e-02 -6.22680008e-01 9.35825646e-01 2.39557907e-01
-3.61495554e-01 8.38686466e-01 5.92708290e-01 -4.70962316e-01
-1.90795198e-01 -3.72148812e-01 -1.30554382e-02 -4.21474665e-01
2.29472533e-01 1.86119094e-01 9.79599282e-02 -2.50516385e-01
6.55943513e-01 2.65302658e-01 -1.11237729e+00 -2.47182120e-02
1.18561924e+00 3.73424888e-01 7.94885308e-02 3.07326317e-01
-3.33015025e-01 1.23249340e+00 -7.94050038e-01 -8.77889514e-01
9.41473618e-02 -2.75387585e-01 -1.04433417e+00 -1.33876756e-01
9.64326039e-02 -7.02010810e-01 -1.13372177e-01 7.94542134e-01
4.66387123e-01 1.43201128e-02 3.17045271e-01 -1.80844247e+00
-7.13831902e-01 6.85507596e-01 -4.28265572e-01 -1.51228160e-01
3.02003294e-01 5.32801092e-01 8.27675819e-01 -2.52044231e-01
7.74450541e-01 1.08484030e+00 5.95233440e-01 3.73156875e-01
-6.94946826e-01 -8.69332075e-01 1.20042399e-01 2.00409126e-02
-9.22852755e-01 -3.60246420e-01 2.30229035e-01 -5.18127501e-01
1.07926440e+00 4.37133491e-01 7.41699100e-01 1.11867857e+00
2.30660677e-01 3.94262552e-01 1.40319264e+00 -3.56065691e-01
1.28922433e-01 9.78742182e-01 1.08579583e-01 2.90799767e-01
4.89878207e-01 -4.97084647e-01 -3.37379336e-01 -3.80681694e-01
3.66204798e-01 -2.59452641e-01 3.40346177e-03 6.56226724e-02
-8.79558146e-01 7.98598707e-01 -4.68279868e-02 7.95751929e-01
-3.66982728e-01 5.11195660e-02 6.57288790e-01 1.82877213e-01
6.74332321e-01 5.24100304e-01 -1.40030831e-01 -8.60044777e-01
-5.71802974e-01 -5.70750386e-02 1.23594594e+00 8.32796395e-01
3.97126436e-01 -1.26392037e-01 -1.82527855e-01 6.80522203e-01
3.47124785e-01 7.76332840e-02 9.36023444e-02 -1.15582621e+00
4.13479716e-01 8.57024252e-01 8.99785012e-02 -1.16807640e+00
-1.56754956e-01 -2.23403201e-01 -4.52798128e-01 7.82965496e-03
3.92486364e-01 -5.32614648e-01 1.43397048e-01 1.54911160e+00
2.77024191e-02 -1.64624900e-01 -1.39584556e-01 4.15667266e-01
7.18673348e-01 5.44489443e-01 3.14503700e-01 -2.21793577e-01
1.22134686e+00 -7.37907469e-01 -9.99043822e-01 -2.90371299e-01
8.42474818e-01 -7.16509998e-01 1.20817935e+00 1.04535766e-01
-7.75180101e-01 1.89609364e-01 -4.87507433e-01 1.00371785e-01
-4.79556859e-01 1.61691457e-01 6.01268470e-01 1.16501260e+00
-1.10918760e+00 3.10450435e-01 -4.10383284e-01 -1.20020759e+00
9.23508406e-02 2.02154920e-01 -1.61913782e-01 4.22312438e-01
-9.13436532e-01 1.26156092e+00 1.26125431e-02 -9.59974229e-02
-2.21348494e-01 -1.32150248e-01 -1.45159200e-01 -1.12043038e-01
3.87844622e-01 -2.11840034e-01 1.00914347e+00 -5.32546818e-01
-1.60799229e+00 7.21787632e-01 5.63346334e-02 -3.98911446e-01
6.45347357e-01 -1.33699208e-01 -1.65788338e-01 3.13530043e-02
2.91696906e-01 4.40026939e-01 8.96629915e-02 -1.15495849e+00
-7.65966594e-01 -9.15709510e-02 4.72950041e-01 1.80956319e-01
-7.88972199e-01 5.85448146e-01 6.59896135e-02 -5.87686412e-02
-4.58042979e-01 -8.65382135e-01 -9.53282118e-02 -3.25028360e-01
-4.42134887e-01 -3.63833278e-01 1.18865371e+00 -8.80975366e-01
1.43791115e+00 -1.93380499e+00 -4.62205589e-01 7.34869763e-02
2.47271582e-01 5.01345575e-01 7.27286339e-02 1.08069968e+00
5.46637475e-01 7.64048219e-01 3.00813943e-01 -1.84232980e-01
2.07360685e-01 -5.18895760e-02 1.11634791e-01 3.56931746e-01
-1.04939252e-01 7.27907777e-01 -7.31587470e-01 -4.01025981e-01
2.13025048e-01 3.93722832e-01 -5.02522886e-01 -7.54361600e-02
-3.60570401e-02 5.98476231e-01 -4.45841640e-01 4.27120000e-01
6.01581395e-01 -1.14900321e-01 4.49322909e-01 2.36431703e-01
-8.27852070e-01 5.48455060e-01 -5.22856712e-01 7.25955248e-01
-3.69780689e-01 9.57378030e-01 3.72204214e-01 -5.60128987e-01
9.01676357e-01 6.27905190e-01 4.94396985e-01 -4.31323677e-01
4.21540231e-01 3.83236170e-01 2.18204558e-01 -8.11431587e-01
7.94694185e-01 3.80234659e-01 8.49233791e-02 9.32448804e-01
-2.89804101e-01 -1.73479483e-01 -5.82239293e-02 4.82169390e-01
8.91267776e-01 5.90563975e-02 3.45739901e-01 -2.64244676e-01
3.26280594e-01 -5.70491329e-02 1.61656782e-01 5.88926136e-01
-7.16930330e-01 -2.30435818e-01 9.71794605e-01 2.61286885e-01
-8.02675128e-01 3.82518774e-04 4.08046216e-01 9.59804177e-01
1.52948266e-02 -7.68669367e-01 -1.01383924e+00 -6.01160467e-01
-4.57578659e-01 1.05411124e+00 -1.19081713e-01 1.15412846e-01
-3.59475255e-01 -4.33532566e-01 6.89142883e-01 -3.58029418e-02
9.86370087e-01 -9.97425795e-01 -8.01592171e-01 1.21006621e-02
-5.76027155e-01 -1.28320813e+00 -1.85777649e-01 -2.87147045e-01
-5.64056277e-01 -9.77552772e-01 -4.04686838e-01 -5.65745234e-01
3.80119026e-01 5.44438720e-01 5.24307191e-01 2.90064722e-01
2.33482286e-01 6.80759251e-01 -5.75019002e-01 -5.02775013e-01
-7.43564427e-01 1.76796466e-01 -4.74734828e-02 -3.23358119e-01
4.82215911e-01 -6.46369576e-01 -2.49178693e-01 3.95657092e-01
-4.86988038e-01 1.82480719e-02 4.68124330e-01 1.99382991e-01
-4.64448363e-01 1.35432616e-01 6.77043438e-01 -8.33020151e-01
1.28643250e+00 -6.25982523e-01 -1.13228522e-01 3.25126231e-01
-5.11799514e-01 -6.55902565e-01 3.59451234e-01 -1.48060933e-01
-1.22487485e+00 -6.35288894e-01 -8.84531438e-02 4.01560187e-01
-2.94536710e-01 3.82931858e-01 -1.05257027e-01 -2.48946175e-01
4.86297965e-01 -4.46147732e-02 3.18718225e-01 -2.43269429e-01
9.53748599e-02 1.12262821e+00 -2.73592591e-01 -3.99965018e-01
6.86644793e-01 3.84758413e-01 -5.08551300e-01 -1.30292225e+00
-1.97325855e-01 -2.14736685e-01 -3.21650535e-01 -6.83659554e-01
7.46657014e-01 -5.56476772e-01 -1.23970592e+00 6.12176001e-01
-1.07280934e+00 -2.92824894e-01 1.85057998e-01 5.54127634e-01
-2.03287110e-01 6.66554213e-01 -7.28934705e-01 -1.44795334e+00
-4.87453938e-01 -8.61390650e-01 1.37327597e-01 3.27841789e-01
-7.76536882e-01 -9.61801231e-01 -3.47982943e-01 8.99125278e-01
9.34967935e-01 1.21720158e-01 9.78099227e-01 -8.37840259e-01
-4.39176321e-01 -2.46258393e-01 -3.30471575e-01 2.95485735e-01
2.84974962e-01 1.97828680e-01 -7.11374223e-01 -2.39498079e-01
2.14634374e-01 -4.76059556e-01 -1.19879588e-01 8.99912417e-02
4.24883336e-01 -9.43627775e-01 -2.53203779e-01 -2.61943251e-01
1.09042001e+00 5.02769530e-01 4.69150364e-01 6.21857941e-01
2.13312954e-01 1.14476264e+00 8.18244874e-01 6.73910499e-01
6.29568398e-01 3.73900652e-01 5.03420770e-01 2.35449046e-01
3.20951432e-01 -3.78375530e-01 7.06499517e-01 9.33727443e-01
-2.36884981e-01 -2.92040646e-01 -8.43824565e-01 3.55159342e-01
-1.74368000e+00 -8.57865810e-01 -4.18144256e-01 1.88068449e+00
2.64287442e-01 1.34754628e-01 5.54051399e-01 -1.83140486e-01
7.11219013e-01 1.57815993e-01 -1.20375618e-01 -7.74884164e-01
7.06962356e-03 -4.52603549e-01 5.03561199e-01 2.84343094e-01
-4.40288275e-01 9.77939188e-01 6.63611126e+00 5.06857157e-01
-1.14844251e+00 2.67866433e-01 7.88833499e-01 2.20360324e-01
-3.76927882e-01 5.31167090e-01 -5.43030500e-01 3.58141273e-01
8.33919704e-01 -5.53918898e-01 5.46508551e-01 6.15671933e-01
7.76983559e-01 -4.54979360e-01 -5.07861853e-01 4.44875270e-01
-8.85147825e-02 -1.14237642e+00 -4.54245657e-01 5.17424226e-01
4.27592814e-01 1.36074238e-02 7.53508434e-02 2.39354074e-01
3.78672749e-01 -7.48744607e-01 6.03267908e-01 -2.49360800e-02
4.25859034e-01 -6.17768288e-01 9.17899728e-01 6.84389830e-01
-7.66553402e-01 -2.72945762e-01 -1.81509331e-01 -6.56402707e-01
1.25262171e-01 -3.46391485e-03 -9.31136131e-01 5.15947878e-01
6.43593669e-01 4.70932513e-01 -5.40888906e-01 8.56629252e-01
-1.71054870e-01 8.77636194e-01 -2.11951777e-01 -5.87368250e-01
5.11400640e-01 -7.11488545e-01 5.54707587e-01 1.14943647e+00
2.76601166e-01 7.14279413e-02 -2.02206030e-01 8.17420006e-01
2.48729110e-01 3.35364491e-01 -7.92137206e-01 -6.82327509e-01
1.01823044e+00 1.22628653e+00 -8.27257991e-01 -3.89927141e-02
-5.05262256e-01 5.18681884e-01 1.32347390e-01 2.50061095e-01
-6.93856835e-01 -2.15030774e-01 5.43316126e-01 2.72017807e-01
-3.35563332e-01 -1.84641913e-01 -4.90363061e-01 -8.14697683e-01
-1.22399718e-01 -1.04428232e+00 -4.20415401e-02 -4.75402385e-01
-1.05040264e+00 5.28627932e-01 1.19592078e-01 -1.05606520e+00
2.00764351e-02 -2.13995814e-01 -8.92207444e-01 5.72513759e-01
-9.63542759e-01 -1.28524053e+00 -6.37893379e-02 1.17017418e-01
2.12101132e-01 -1.39749989e-01 6.72668040e-01 2.41653338e-01
-7.40381360e-01 3.80876273e-01 1.25371248e-01 1.66546162e-02
7.21295774e-01 -5.95489979e-01 1.63843200e-01 5.11692464e-01
-5.88652611e-01 8.56966197e-01 8.50162864e-01 -7.09050179e-01
-1.04569864e+00 -6.86784267e-01 1.41584826e+00 -3.67904663e-01
7.75337636e-01 -4.38587993e-01 -4.78381306e-01 7.90509939e-01
8.36428285e-01 -1.18909407e+00 6.98421180e-01 1.26599386e-01
1.73752978e-01 1.65600851e-01 -1.45193434e+00 1.03754008e+00
7.27227211e-01 -5.10807097e-01 -1.55336425e-01 5.01562893e-01
5.13363898e-01 1.76852010e-02 -6.36947632e-01 -4.11371291e-01
4.93470162e-01 -1.36065137e+00 3.51643860e-01 -1.46028325e-01
3.14230710e-01 1.81614518e-01 5.37258089e-02 -1.01618004e+00
-1.07430801e-01 -1.03342402e+00 4.61220503e-01 1.86305559e+00
2.07404435e-01 -1.23854399e+00 7.83234358e-01 1.25051272e+00
1.10080272e-01 -5.29165804e-01 -7.00784326e-01 -6.66315377e-01
9.76718143e-02 -3.00627589e-01 3.81837279e-01 1.13367510e+00
6.36672139e-01 3.10064673e-01 -8.26184332e-01 -2.32732952e-01
1.25053197e-01 -3.30544531e-01 1.06385493e+00 -9.86715615e-01
6.01106919e-02 -3.90205711e-01 -5.84217831e-02 -6.69902742e-01
-4.76264767e-02 -5.36567152e-01 -4.08607304e-01 -2.02170610e+00
2.99254239e-01 -3.48246813e-01 4.44963098e-01 3.94265890e-01
2.57508695e-01 -1.86955675e-01 5.77001870e-01 3.18160266e-01
-4.77180004e-01 2.13259473e-01 1.08687806e+00 3.31699044e-01
-3.40873688e-01 3.38509306e-02 -1.16463542e+00 6.59438729e-01
1.21596563e+00 -3.05355459e-01 -4.39152837e-01 -1.27526850e-01
1.20628953e-01 4.36963625e-02 2.06993476e-01 -7.45268643e-01
2.72840738e-01 -5.80618262e-01 -4.40817535e-01 -4.25193012e-01
5.11693478e-01 -9.86884892e-01 2.50467747e-01 5.60162067e-01
-2.21512690e-01 -1.15972422e-02 1.61196198e-02 5.59512749e-02
2.93456223e-02 -2.83511698e-01 4.40386474e-01 -2.13500068e-01
-1.54436857e-01 -1.63522705e-01 -1.13305390e+00 -7.98314884e-02
1.19334471e+00 -6.70388401e-01 -7.16886640e-01 -1.17482257e+00
7.56481616e-03 3.95989537e-01 5.70094407e-01 3.92598003e-01
1.00509949e-01 -7.27081180e-01 -5.75602412e-01 -2.66456425e-01
-1.44938827e-01 -7.93526113e-01 1.93612590e-01 1.24031770e+00
-5.27120888e-01 6.25283420e-01 -4.24471438e-01 -1.46645503e-02
-1.46206474e+00 7.69434904e-04 -9.44467168e-03 -1.06163740e-01
-4.68067259e-01 4.05634105e-01 -1.73705816e-01 -5.00952959e-01
4.87702608e-01 1.82759315e-02 -4.97283489e-01 1.22258104e-01
2.16154665e-01 8.80002737e-01 -4.50888366e-01 -7.66571462e-01
-4.10124809e-01 4.33277935e-02 7.10841417e-02 -2.93316305e-01
1.27268255e+00 -5.31855762e-01 -6.53726637e-01 3.57416660e-01
7.30283141e-01 4.55227584e-01 -6.87066197e-01 5.13411313e-02
3.80692631e-01 -7.54961431e-01 -6.29173458e-01 -9.99360144e-01
-7.61478484e-01 6.94457710e-01 7.61065781e-02 6.78824842e-01
7.11881757e-01 -1.78699091e-01 5.86012840e-01 2.92332351e-01
3.19874018e-01 -1.35259998e+00 1.10225819e-01 7.68987834e-01
7.27910936e-01 -8.07485163e-01 -1.58763587e-01 -5.47240555e-01
-1.04482675e+00 9.03054953e-01 9.43010867e-01 3.65612954e-01
5.03204703e-01 1.67648301e-01 2.57538170e-01 -1.01480097e-01
-9.20815229e-01 2.85519302e-01 -6.93518817e-01 8.27302337e-01
8.67726266e-01 -5.23862094e-02 -1.05188489e+00 3.63426179e-01
-3.21297407e-01 3.91553789e-02 9.12248135e-01 9.03800845e-01
-3.81562889e-01 -1.22414458e+00 -2.60867506e-01 5.42681038e-01
-5.13870299e-01 7.53458887e-02 -1.14365232e+00 9.93672788e-01
2.41575502e-02 1.60805428e+00 -4.25451040e-01 -6.74890459e-01
9.17908475e-02 2.21473770e-03 -4.33555424e-01 -3.77856195e-01
-1.27769446e+00 -2.01131493e-01 7.67476916e-01 -8.73545408e-02
-3.82847995e-01 -5.44597507e-01 -8.71635199e-01 -1.10153198e+00
-6.75423384e-01 5.54206669e-01 7.78774500e-01 9.31703389e-01
4.86534595e-01 1.03939690e-01 4.39752877e-01 -2.11982593e-01
-4.11567241e-01 -1.23689747e+00 -5.61229289e-01 1.02626055e-03
-2.41908878e-01 -1.48283347e-01 -4.13745195e-01 -6.24588966e-01] | [10.320940017700195, 7.263548851013184] |
df935477-aa1c-4964-9e19-017938b60da0 | mapping-and-cleaning-open-commonsense | 2306.12766 | null | https://arxiv.org/abs/2306.12766v1 | https://arxiv.org/pdf/2306.12766v1.pdf | Mapping and Cleaning Open Commonsense Knowledge Bases with Generative Translation | Structured knowledge bases (KBs) are the backbone of many know\-ledge-intensive applications, and their automated construction has received considerable attention. In particular, open information extraction (OpenIE) is often used to induce structure from a text. However, although it allows high recall, the extracted knowledge tends to inherit noise from the sources and the OpenIE algorithm. Besides, OpenIE tuples contain an open-ended, non-canonicalized set of relations, making the extracted knowledge's downstream exploitation harder. In this paper, we study the problem of mapping an open KB into the fixed schema of an existing KB, specifically for the case of commonsense knowledge. We propose approaching the problem by generative translation, i.e., by training a language model to generate fixed-schema assertions from open ones. Experiments show that this approach occupies a sweet spot between traditional manual, rule-based, or classification-based canonicalization and purely generative KB construction like COMET. Moreover, it produces higher mapping accuracy than the former while avoiding the association-based noise of the latter. | ['Simon Razniewski', 'Julien Romero'] | 2023-06-22 | null | null | null | null | ['open-information-extraction'] | ['natural-language-processing'] | [ 8.36505815e-02 9.28857386e-01 -3.85644644e-01 -1.26672044e-01
-7.51741290e-01 -6.74037337e-01 5.41545153e-01 2.86556423e-01
-1.20815888e-01 1.43909276e+00 1.17191695e-01 -4.21538681e-01
-2.19802380e-01 -1.28320205e+00 -8.88111413e-01 -4.36499238e-01
4.18249547e-01 8.69618297e-01 9.17163119e-02 -3.33934277e-01
-1.07326008e-01 8.94927680e-02 -1.37295401e+00 3.05124909e-01
1.31707013e+00 8.45357656e-01 8.13280568e-02 -1.03820816e-01
-7.82534242e-01 9.73351657e-01 -4.12604898e-01 -1.14010477e+00
7.28668198e-02 -6.03853405e-01 -1.20714295e+00 -1.07339539e-01
-1.04627371e-01 4.74908561e-01 -3.36927883e-02 1.20063090e+00
1.22348601e-02 -9.86959413e-02 6.55232191e-01 -1.19775665e+00
-8.18540096e-01 1.23233712e+00 -1.81379408e-01 -1.83508471e-01
5.73128641e-01 -3.19090694e-01 1.20530915e+00 -6.62492335e-01
1.12003207e+00 8.85983527e-01 6.84953094e-01 4.14517015e-01
-1.53067851e+00 -4.54997897e-01 -2.60605782e-01 3.56959105e-01
-1.71754265e+00 -4.78601962e-01 6.28485441e-01 -4.04890090e-01
7.03996539e-01 3.70754004e-01 6.62600160e-01 1.09936476e+00
-1.01781942e-01 5.61593950e-01 1.28151953e+00 -8.50354671e-01
3.13977301e-01 7.90849805e-01 1.62524939e-01 5.24973392e-01
9.17504072e-01 -2.40417868e-01 -4.29922193e-01 -2.50525326e-01
4.91288006e-01 -3.99348915e-01 -5.48901021e-01 -4.60140079e-01
-1.01117182e+00 5.83637118e-01 2.77254432e-01 4.43748325e-01
-3.90107185e-01 -2.44075686e-01 2.89251119e-01 2.37117231e-01
3.37331206e-01 7.39755273e-01 -6.03766620e-01 -4.30502780e-02
-7.41725445e-01 4.18755531e-01 1.42658389e+00 1.38931012e+00
1.05760455e+00 -5.06226540e-01 1.33472487e-01 6.45950198e-01
8.57660100e-02 3.03976178e-01 4.33816075e-01 -5.08740842e-01
6.12769485e-01 8.76955569e-01 1.84545979e-01 -9.70242918e-01
-5.11270985e-02 -2.63153851e-01 -7.01783597e-01 -5.09928465e-01
4.15839076e-01 -8.94242376e-02 -6.82420075e-01 1.69291258e+00
3.62227798e-01 5.92608657e-03 4.90637451e-01 5.31979561e-01
7.77764320e-01 5.26286185e-01 5.13072032e-03 -5.47519982e-01
1.35635161e+00 -5.35546064e-01 -9.33047056e-01 -1.09646574e-01
8.08857799e-01 -2.11998791e-01 7.24669337e-01 2.99339026e-01
-6.00519240e-01 -1.58983618e-01 -9.58765745e-01 -9.58332568e-02
-7.96570539e-01 -2.11417973e-01 6.33653760e-01 6.47620618e-01
-5.56643307e-01 3.75108272e-01 -3.29231232e-01 -3.50140631e-01
3.51984233e-01 -6.01673350e-02 -5.84144115e-01 -2.99304694e-01
-1.60324538e+00 1.12113476e+00 1.14837837e+00 -9.51943621e-02
-1.47553191e-01 -7.51885653e-01 -8.79770577e-01 1.64548963e-01
1.30837929e+00 -9.15392458e-01 8.73783469e-01 -5.87136209e-01
-1.11864471e+00 7.01930344e-01 -1.49094433e-01 -5.45789897e-01
2.43686348e-01 -1.50202468e-01 -5.51505327e-01 -2.53001511e-01
3.07939202e-01 1.75844684e-01 5.29335797e-01 -1.53496134e+00
-3.52453887e-01 -4.28507775e-01 3.23872179e-01 -1.70722201e-01
-1.71955168e-01 -2.21449703e-01 -2.74167597e-01 -5.02455652e-01
2.16048598e-01 -8.61407399e-01 -1.40433609e-01 -5.52000403e-01
-8.86141896e-01 -2.36513630e-01 2.90272355e-01 -5.51038027e-01
1.60386550e+00 -1.81935430e+00 3.12753201e-01 5.94901025e-01
2.50631154e-01 2.25012392e-01 3.70942593e-01 6.96842849e-01
-1.16678789e-01 4.57473129e-01 -3.82441252e-01 3.00780684e-01
1.61687285e-02 6.77557945e-01 -5.85846901e-01 -3.13804865e-01
1.16321169e-01 1.13017142e+00 -9.53168631e-01 -7.81630814e-01
-4.00479704e-01 -7.38660023e-02 -5.13097882e-01 1.98649578e-02
-7.65687406e-01 1.48963168e-01 -4.22875732e-01 4.65697020e-01
3.02988201e-01 -2.81171530e-01 7.04651833e-01 -2.26382598e-01
3.69093269e-02 3.79422039e-01 -1.24472392e+00 1.62744617e+00
-4.90591645e-01 1.61645800e-01 -4.23631161e-01 -8.00850093e-01
1.09737408e+00 3.60209465e-01 1.23528197e-01 -1.21221781e-01
3.81121263e-02 3.58504683e-01 -4.35784757e-02 -7.35166252e-01
6.27990603e-01 -5.39745986e-01 -2.52995282e-01 3.72622490e-01
2.61140198e-01 -1.13670088e-01 4.03800726e-01 2.98443943e-01
9.56677675e-01 2.33140692e-01 8.80110085e-01 -1.33155346e-01
4.83616173e-01 2.67316908e-01 8.77224982e-01 4.87675250e-01
5.33582866e-01 1.09172858e-01 8.23524356e-01 -2.72304088e-01
-8.43572915e-01 -8.04349065e-01 -2.07048237e-01 2.89572537e-01
1.33985817e-01 -9.60112035e-01 -7.71510243e-01 -9.42930579e-01
8.41663107e-02 1.09050679e+00 -6.02160156e-01 -1.88076541e-01
-2.94334769e-01 -6.05696380e-01 7.86242127e-01 3.57794136e-01
3.32534730e-01 -8.32947314e-01 -3.56061965e-01 3.13749343e-01
-9.03424144e-01 -1.24419570e+00 2.12221920e-01 3.16358149e-01
-6.02763653e-01 -1.15462661e+00 -1.48148313e-01 -3.76698822e-01
6.74297333e-01 -3.12402964e-01 1.25041473e+00 -2.36536160e-01
6.79995343e-02 -7.46852309e-02 -7.99496353e-01 -5.73049843e-01
-7.25537598e-01 3.59944105e-01 -1.81365162e-02 5.04602939e-02
7.28130877e-01 -6.17549896e-01 2.68925369e-01 1.03153005e-01
-1.08302426e+00 2.96884954e-01 6.69041455e-01 9.63839710e-01
6.50046766e-01 3.47698480e-01 8.61366034e-01 -1.38771117e+00
7.17979908e-01 -5.59143186e-01 -3.88243377e-01 7.02317059e-01
-8.95181715e-01 4.42045927e-01 5.32638073e-01 -1.14684328e-01
-1.33351099e+00 -1.22638531e-01 2.92046573e-02 -1.90989763e-01
-1.64922774e-01 1.10973418e+00 -6.09761834e-01 2.35390022e-01
9.19586122e-01 2.12858856e-01 -2.14896187e-01 -5.13786137e-01
7.45366931e-01 7.31994271e-01 4.80748892e-01 -8.65730047e-01
8.49083841e-01 2.34729365e-01 -1.25970080e-01 -6.13728404e-01
-1.17152452e+00 -1.48933172e-01 -8.03938508e-01 1.43625200e-01
7.35576689e-01 -5.73451579e-01 -2.01386154e-01 -2.10509911e-01
-1.19791138e+00 5.10124154e-02 -7.36170590e-01 1.70554101e-01
-5.77739239e-01 2.22866446e-01 -3.27588141e-01 -7.20525026e-01
-1.61815703e-01 -6.80611253e-01 6.40888095e-01 1.69264767e-02
-5.74995041e-01 -7.71146595e-01 2.99043208e-01 3.81727725e-01
-1.56926543e-01 2.41525203e-01 1.41980398e+00 -1.06138849e+00
-5.68942606e-01 -2.17300594e-01 -2.48303324e-01 2.87439585e-01
1.51316643e-01 -3.60344470e-01 -8.62927794e-01 2.79540002e-01
-1.58472463e-01 -2.73717761e-01 4.93847460e-01 -3.60436231e-01
7.83163846e-01 -5.88088453e-01 -4.74675298e-01 1.98456630e-01
1.61175966e+00 1.73065022e-01 8.69963765e-01 3.20011020e-01
6.08459532e-01 7.53450692e-01 6.41227663e-01 2.97075808e-01
4.44843113e-01 6.16281986e-01 -1.44595876e-01 4.00352746e-01
2.01416358e-01 -6.39020801e-01 -8.79173726e-03 6.74983621e-01
-3.73141915e-01 -4.77298424e-02 -1.04772544e+00 4.51441199e-01
-1.84213519e+00 -1.06927764e+00 -7.80156255e-02 1.91982269e+00
1.45427346e+00 1.88638419e-01 -1.59070477e-01 4.09642190e-01
4.80045766e-01 -3.44101518e-01 -1.00073770e-01 -2.12393180e-01
-2.19350621e-01 2.29456440e-01 2.08722264e-01 2.86731958e-01
-6.50020242e-01 1.00981820e+00 5.15574646e+00 8.70171547e-01
-6.45071507e-01 1.44165799e-01 1.14529386e-01 3.24332774e-01
-6.51230574e-01 4.55864131e-01 -1.01382101e+00 3.66060287e-01
8.75371933e-01 -5.00636220e-01 1.94369689e-01 8.34374547e-01
-3.03782314e-01 -2.68834710e-01 -1.30168176e+00 7.83686519e-01
1.43270299e-01 -1.44585621e+00 2.79717565e-01 3.36322099e-01
6.34967387e-01 -4.89259928e-01 -7.00360000e-01 4.63544667e-01
2.89103299e-01 -7.36127853e-01 8.24498594e-01 6.90324485e-01
8.78668189e-01 -7.54104257e-01 9.94179547e-01 5.75390279e-01
-8.79705191e-01 -8.17206725e-02 -3.13988060e-01 1.25406697e-01
2.17054218e-01 1.08871448e+00 -9.14587736e-01 1.19953620e+00
3.36628646e-01 3.49885911e-01 -5.05094051e-01 6.27409101e-01
-5.53476214e-01 3.93855959e-01 -1.02611922e-01 -3.60284075e-02
-1.09759219e-01 -4.69657868e-01 4.83659655e-01 1.16107595e+00
2.49613374e-01 3.55414748e-01 9.59485173e-02 1.17368567e+00
-2.28173450e-01 2.20757961e-01 -1.00361311e+00 -2.56869435e-01
4.55980897e-01 1.00727499e+00 -5.12996256e-01 -6.72610044e-01
-3.07246447e-01 6.27541065e-01 5.53613544e-01 2.28598505e-01
-6.41718030e-01 -6.18737876e-01 3.97820532e-01 1.49534181e-01
2.83336282e-01 1.00212090e-01 -3.18698287e-01 -1.36937356e+00
3.23451072e-01 -7.63004243e-01 4.07883286e-01 -7.95205891e-01
-1.23150980e+00 7.04065323e-01 3.31541091e-01 -9.78019774e-01
-5.87819576e-01 -2.55325168e-01 -8.52383822e-02 8.26812088e-01
-1.15568352e+00 -1.08672929e+00 -5.36365472e-02 3.73544484e-01
1.67260751e-01 2.39144415e-01 1.01548111e+00 2.66496330e-01
-5.53135931e-01 4.60169882e-01 -2.01622501e-01 3.79106075e-01
5.49836338e-01 -1.27230859e+00 4.01638597e-02 8.89754951e-01
3.25607985e-01 9.10111308e-01 7.69188464e-01 -1.09186292e+00
-1.34569967e+00 -1.13485217e+00 1.35917354e+00 -6.68121457e-01
8.67916107e-01 -3.33041579e-01 -1.17417836e+00 1.01989460e+00
6.61175884e-03 -2.19029576e-01 9.41155672e-01 5.21082580e-01
-7.30055034e-01 -1.42965466e-01 -9.43014085e-01 5.80512822e-01
1.09517133e+00 -5.45747757e-01 -1.18144190e+00 3.63217406e-02
8.02500010e-01 -2.48490989e-01 -1.12994957e+00 2.88824409e-01
4.29490179e-01 -6.04623020e-01 6.20549858e-01 -8.36965084e-01
6.12520456e-01 -3.63935590e-01 -1.50011942e-01 -1.49417746e+00
-1.69790611e-01 -4.78515238e-01 -3.11774552e-01 1.59666991e+00
9.03956771e-01 -6.01748407e-01 4.76485997e-01 7.20482469e-01
-3.27017605e-02 -7.44624019e-01 -8.24869931e-01 -1.02517593e+00
-1.42112613e-01 -5.31415522e-01 7.74316907e-01 1.22629762e+00
7.62405276e-01 7.89725244e-01 -1.82887822e-01 -4.32935990e-02
4.45399255e-01 4.36006904e-01 8.24392676e-01 -1.43606949e+00
-3.31734359e-01 2.48652021e-03 -4.27308649e-01 -5.80389380e-01
2.34410584e-01 -1.23157072e+00 1.74797252e-01 -1.42945874e+00
1.81711808e-01 -9.58932698e-01 1.29027143e-01 6.80225611e-01
-4.64637503e-02 -1.15161076e-01 6.04103729e-02 9.16737691e-02
-3.75965357e-01 3.86812389e-01 1.08034503e+00 -4.37939279e-02
-3.63387585e-01 -3.42097878e-01 -1.13064086e+00 7.03336656e-01
6.24274254e-01 -6.36231244e-01 -5.52677393e-01 6.37583807e-02
9.55098808e-01 1.00797839e-01 2.38411993e-01 -6.67444110e-01
1.70696452e-01 -2.74414361e-01 -1.01732440e-01 -1.87196285e-01
7.88674727e-02 -7.88070798e-01 6.46273673e-01 6.26742169e-02
-3.08635622e-01 -3.35557014e-01 -1.83595736e-02 6.96232915e-01
-2.85016686e-01 -5.47730446e-01 3.33296537e-01 -3.42866451e-01
-5.94504416e-01 -1.28357410e-01 -1.36206612e-01 4.02108878e-01
9.51459765e-01 -5.45317680e-02 -2.67078698e-01 1.89027525e-02
-7.80393541e-01 -1.83614030e-01 3.17058951e-01 2.43194506e-01
4.95277077e-01 -1.17572916e+00 -5.64824760e-01 5.54520860e-02
5.95905304e-01 3.69783729e-01 -2.03148201e-01 9.52546477e-01
-1.09198742e-01 4.97038066e-01 -3.87872756e-02 -9.60693359e-02
-8.03297639e-01 8.22248042e-01 -5.38354516e-02 -5.70176542e-01
-7.73182929e-01 4.71013039e-01 -4.90976200e-02 -3.71961147e-01
-1.48813322e-01 -3.56963426e-01 -2.71196991e-01 2.24194646e-01
3.64895135e-01 1.95245042e-01 2.55344927e-01 -4.75047797e-01
-3.83275002e-02 -8.03540573e-02 -6.00609966e-02 -1.20330602e-01
1.14920938e+00 -7.11564645e-02 -6.55082762e-01 6.37090683e-01
6.61355495e-01 1.99423730e-01 -1.23220049e-01 -5.51084518e-01
5.50556600e-01 -3.81507218e-01 -3.82811278e-01 -7.40416527e-01
-4.85302716e-01 3.22504342e-01 -3.75305206e-01 4.66708869e-01
7.69012332e-01 4.45867032e-01 7.21266747e-01 7.42659330e-01
8.94819975e-01 -9.17098761e-01 -6.10020041e-01 4.76253897e-01
9.36378300e-01 -9.40251291e-01 -7.31226131e-02 -1.02918708e+00
-4.98088270e-01 8.37394595e-01 5.53229332e-01 4.19037968e-01
5.54991901e-01 3.65578145e-01 -1.74188837e-01 -2.60627270e-01
-7.81081200e-01 -4.32744652e-01 -2.32640486e-02 5.70322633e-01
1.14006758e-01 2.18844950e-01 -4.35355663e-01 1.04948127e+00
-4.73413885e-01 2.80617774e-01 5.87498784e-01 9.31379855e-01
-2.85282433e-01 -1.16997981e+00 -3.12795639e-01 6.94315732e-01
-2.10494056e-01 -3.44997913e-01 -6.59376562e-01 8.38079572e-01
5.28800011e-01 7.27448761e-01 -3.71017694e-01 -4.94525880e-01
3.74393106e-01 6.30102575e-01 4.79529887e-01 -8.62471759e-01
-1.69610292e-01 -3.76655757e-01 6.59447908e-01 -2.12806627e-01
-4.74972099e-01 -5.23216784e-01 -1.11270678e+00 -1.67631388e-01
-7.68460751e-01 6.95942283e-01 4.29459661e-01 1.30562365e+00
3.15310985e-01 3.72104883e-01 1.81028694e-01 -3.95200215e-02
-4.15049970e-01 -8.47280622e-01 -7.10114062e-01 4.29215461e-01
-2.99111158e-01 -8.16723228e-01 -6.78371266e-02 3.38188618e-01] | [9.371479988098145, 8.412247657775879] |
6a1f322e-2798-45e4-87fd-a46843234343 | formulating-neural-sentence-ordering-as-the | null | null | https://aclanthology.org/2021.inlg-1.13 | https://aclanthology.org/2021.inlg-1.13.pdf | Formulating Neural Sentence Ordering as the Asymmetric Traveling Salesman Problem | The task of Sentence Ordering refers to rearranging a set of given sentences in a coherent ordering. Prior work (Prabhumoye et al., 2020) models this as an optimal graph traversal (with sentences as nodes, and edges as local constraints) using topological sorting. However, such an approach has major limitations – it cannot handle the presence of cycles in the resulting graphs and considers only the binary presence/absence of edges rather than a more granular score. In this work, we propose an alternate formulation of this task as a classic combinatorial optimization problem popular as the Traveling Salesman Problem (or TSP in short). Compared to the previous approach of using topological sorting, our proposed technique gracefully handles the presence of cycles and is more expressive since it takes into account real-valued constraint/edge scores rather than just the presence/absence of edges. Our experiments demonstrate improved handling of such cyclic cases in resulting graphs. Additionally, we highlight how model accuracy can be sensitive to the ordering of input sentences when using such graphs-based formulations. Finally, we note that our approach requires only lightweight fine-tuning of a classification layer built on pretrained BERT sentence encoder to identify local relationships. | ['Harsh Jhamtani', 'Vishal Keswani'] | null | null | null | null | inlg-acl-2021-8 | ['sentence-ordering'] | ['natural-language-processing'] | [ 4.90679175e-01 2.89310277e-01 -1.33735970e-01 -5.81689477e-01
-3.23870540e-01 -7.55148530e-01 5.25885224e-01 7.02502668e-01
-4.20084059e-01 6.35163724e-01 9.73149016e-02 -6.87568009e-01
-5.78288794e-01 -1.00384963e+00 -7.82252610e-01 -2.39285529e-01
-3.97875726e-01 6.93804860e-01 5.25069356e-01 -4.03375357e-01
4.94186014e-01 5.20748675e-01 -1.40428245e+00 3.55331302e-01
7.89272487e-01 8.21765542e-01 2.58733928e-01 7.45818138e-01
-2.27788940e-01 6.59545124e-01 -6.68630540e-01 -5.54625750e-01
1.96513146e-01 -2.50635475e-01 -1.18691051e+00 1.42387778e-01
7.43024051e-01 -7.02713355e-02 3.99579220e-02 1.00496209e+00
1.18412226e-01 6.17122874e-02 4.23172265e-01 -1.21995950e+00
-5.71871936e-01 9.03102994e-01 -4.81060565e-01 2.43996531e-01
6.94465399e-01 -8.52321908e-02 1.58747685e+00 -4.36694592e-01
6.64521933e-01 1.06180525e+00 4.80238169e-01 3.25145200e-02
-1.20652950e+00 -4.11113352e-02 5.57362258e-01 3.40863615e-01
-1.15696394e+00 -1.84233055e-01 7.83024490e-01 -2.51494825e-01
1.35742724e+00 7.64791310e-01 5.13681769e-01 4.61512625e-01
2.81753421e-01 5.80251157e-01 8.04873407e-01 -4.45667148e-01
2.09801704e-01 -1.76321119e-01 7.28758633e-01 7.18921661e-01
2.51402378e-01 -4.16617602e-01 -3.43218148e-01 5.69350570e-02
1.33404866e-01 -2.54773945e-01 6.69318251e-03 -3.95838469e-01
-9.73579466e-01 7.36214340e-01 4.85332102e-01 3.98700327e-01
-2.04381272e-01 2.63818085e-01 5.47094226e-01 4.56077456e-01
2.28268936e-01 5.33676982e-01 -4.64285672e-01 3.38789285e-03
-1.08178258e+00 1.88928023e-01 8.14914942e-01 1.05890596e+00
8.16570640e-01 -4.08188760e-01 -1.95962742e-01 5.99683225e-01
1.32335335e-01 -9.31847468e-02 3.55237871e-02 -6.32089376e-01
7.94291317e-01 8.22142959e-01 -1.44558221e-01 -1.43037438e+00
-6.00456357e-01 -5.71620405e-01 -5.84674716e-01 -2.91258752e-01
3.38197410e-01 2.14362770e-01 -8.74805689e-01 1.82615268e+00
8.88079330e-02 -1.30461484e-01 -2.22550333e-01 7.77212024e-01
6.33905411e-01 4.82267648e-01 -1.19336314e-01 -3.24247748e-01
1.32644784e+00 -9.64773059e-01 -7.20223188e-01 -4.18036401e-01
5.85975528e-01 -5.42081714e-01 1.09505856e+00 3.46742690e-01
-1.34686506e+00 -1.53030902e-01 -1.17497587e+00 -3.30363601e-01
-6.42853439e-01 -2.03748867e-01 8.26060534e-01 7.03261137e-01
-1.55645633e+00 6.83723688e-01 -5.78592956e-01 -5.15599370e-01
6.22719452e-02 7.90359557e-01 -2.16315076e-01 4.43790574e-03
-1.30033875e+00 8.96879196e-01 3.87997925e-01 2.30619773e-01
6.17324514e-03 -3.23788971e-01 -1.01429856e+00 3.41818720e-01
7.14675665e-01 -6.49327099e-01 1.02728331e+00 -7.42846429e-01
-1.25300372e+00 8.36933017e-01 -4.35888588e-01 -6.05529130e-01
4.06231195e-01 1.38419479e-01 -2.53211319e-01 1.41470522e-01
3.34478654e-02 4.58105475e-01 2.94874936e-01 -9.79807913e-01
-4.31244075e-01 -3.30046773e-01 8.29287469e-01 2.61231214e-01
-2.04979107e-01 1.73775226e-01 -5.50456762e-01 -3.20424736e-01
3.77238393e-01 -8.29828680e-01 -2.16464207e-01 -3.67758006e-01
-6.85096860e-01 -3.25844914e-01 4.08288985e-01 -4.39346492e-01
1.67975581e+00 -1.85850310e+00 2.28386343e-01 3.50392848e-01
2.63676971e-01 -1.09531909e-01 -5.64904399e-02 9.88861322e-01
-1.47225901e-01 4.63369697e-01 -5.98427832e-01 -2.89313883e-01
2.32351556e-01 3.05814445e-01 -1.74820155e-03 2.39006460e-01
3.78017396e-01 8.32301021e-01 -9.81287718e-01 -4.97416586e-01
4.36206348e-02 4.28757407e-02 -6.33325577e-01 -5.01385391e-01
-2.69435823e-01 -7.39523694e-02 -7.46991411e-02 3.59835744e-01
7.47974098e-01 -2.04383880e-01 5.25165439e-01 -6.89701885e-02
-1.44395918e-01 5.99165678e-01 -1.43275821e+00 1.21422637e+00
-4.80255097e-01 5.14345169e-01 3.69863361e-02 -1.12358141e+00
7.56977737e-01 -1.45543441e-01 3.22459668e-01 -5.84218204e-01
6.92954212e-02 1.59437940e-01 8.96604359e-02 -4.10678208e-01
8.99569154e-01 1.13051422e-01 -4.07041460e-01 3.15858841e-01
-3.08518589e-01 -3.66276279e-02 7.37879753e-01 5.83518863e-01
1.30910754e+00 -2.94520050e-01 2.73525834e-01 -3.95827651e-01
6.29504859e-01 -2.84074526e-02 5.04013956e-01 8.01865816e-01
4.33083288e-02 6.00867331e-01 1.05317318e+00 -3.41935188e-01
-7.87674963e-01 -8.49591970e-01 -9.01822224e-02 8.96043658e-01
3.11967373e-01 -6.77759945e-01 -4.66817558e-01 -6.32801712e-01
4.03869189e-02 7.00824440e-01 -4.86872077e-01 -1.22224949e-02
-9.04088199e-01 -7.65191078e-01 2.18347684e-01 2.81684279e-01
1.84037939e-01 -9.71701801e-01 -6.15809381e-01 2.07045510e-01
-2.11417958e-01 -1.23755014e+00 -6.88558161e-01 3.19305599e-01
-7.45635211e-01 -1.09504402e+00 -1.50777092e-02 -1.12103987e+00
9.11213458e-01 1.16425179e-01 1.18116105e+00 2.36215502e-01
-1.11432537e-01 1.90795481e-01 -4.71657127e-01 3.42290476e-02
-1.66741818e-01 1.29692629e-01 -2.31063366e-01 8.49083140e-02
1.20202541e-01 -4.10027981e-01 -3.15973729e-01 9.52059925e-02
-1.02859914e+00 8.92215967e-02 3.20001930e-01 6.79252028e-01
4.23810184e-01 4.72571105e-01 5.78454435e-01 -1.37162256e+00
9.32481647e-01 -2.77249962e-01 -3.89354706e-01 5.47583401e-01
-6.14808261e-01 1.45411626e-01 8.02767932e-01 2.96009541e-03
-5.11688173e-01 1.31878033e-01 -1.57681461e-02 8.03762376e-02
8.38192999e-02 9.39977825e-01 -8.60498175e-02 9.42277461e-02
3.16573121e-02 1.38935387e-01 -3.16556096e-01 -1.13241814e-01
1.96943954e-01 4.39310759e-01 2.42119983e-01 -4.01609838e-01
5.23389041e-01 1.64160982e-01 4.43572223e-01 -6.52102351e-01
-2.87906796e-01 -4.07704353e-01 -9.23578262e-01 -1.46804959e-01
6.93510175e-01 -2.84041256e-01 -7.49971151e-01 1.32522494e-01
-1.28222990e+00 3.75589519e-03 6.49530888e-02 -1.91729926e-02
-3.96288127e-01 7.48828888e-01 -5.58340311e-01 -9.18469250e-01
1.97986830e-02 -1.20192134e+00 9.96898413e-01 -1.83328882e-01
-3.66223752e-01 -1.00066209e+00 -4.09069061e-01 3.35078925e-01
3.61610323e-01 4.13820297e-01 1.24620831e+00 -6.41874731e-01
-7.37276614e-01 -4.02952552e-01 -8.07784274e-02 -1.21058188e-01
6.03975207e-02 2.31531724e-01 -3.54786098e-01 -2.03002051e-01
-3.90618891e-02 6.45353198e-02 8.18132937e-01 1.49515972e-01
9.61764514e-01 -5.66445053e-01 -5.75584322e-02 1.38987631e-01
1.49865854e+00 1.24937579e-01 4.11760122e-01 3.72350246e-01
5.85985899e-01 8.90863776e-01 4.78975952e-01 2.29474083e-01
7.79553950e-01 7.09264636e-01 5.48185349e-01 4.37447242e-02
-1.27760076e-03 -6.41919225e-02 1.86384216e-01 8.60345960e-01
3.90677243e-01 -6.27217531e-01 -9.43033576e-01 5.42293429e-01
-1.83095825e+00 -9.01302338e-01 -5.05674183e-01 2.28954959e+00
5.86221635e-01 4.98041332e-01 1.14555955e-01 5.50729632e-01
9.54219759e-01 3.00131679e-01 -1.27064735e-01 -1.12444866e+00
-1.46908134e-01 3.18940692e-02 6.04855299e-01 1.05250061e+00
-1.04516554e+00 7.80521274e-01 5.95188189e+00 4.51885283e-01
-9.86973405e-01 -2.81987458e-01 5.63693881e-01 -1.45979464e-01
-4.88632828e-01 1.45002812e-01 -5.93321145e-01 5.27793586e-01
7.07211316e-01 -6.21672487e-04 5.36542356e-01 1.90456942e-01
2.65660286e-01 -3.49784017e-01 -1.28507304e+00 5.56216955e-01
1.30111560e-01 -1.14889324e+00 -8.12667832e-02 -1.11684360e-01
4.21472937e-01 -2.28719011e-01 7.18802185e-05 -9.34519991e-03
5.69555201e-02 -8.62994373e-01 9.02399719e-01 3.08202624e-01
5.40496290e-01 -6.36249542e-01 7.15875685e-01 2.86611497e-01
-1.32455242e+00 -1.41200796e-01 1.51753888e-01 -3.10317397e-01
3.19538295e-01 5.61932862e-01 -8.98577988e-01 8.61549795e-01
4.42129016e-01 4.66339946e-01 -7.10777044e-01 1.02542436e+00
-6.68973401e-02 3.27304155e-01 -4.87567455e-01 -2.92126268e-01
3.76414418e-01 -2.00036004e-01 6.06795490e-01 1.35863185e+00
-5.01035228e-02 -7.35594109e-02 3.40064198e-01 4.18454111e-01
9.94535759e-02 -1.06013846e-02 -5.10524988e-01 1.60823852e-01
4.29635555e-01 1.01279521e+00 -1.08075392e+00 -1.96822077e-01
-3.94358456e-01 7.77231932e-01 5.78335226e-01 2.23491043e-01
-7.39748895e-01 -4.37268913e-01 4.27710898e-02 2.93429255e-01
3.05246949e-01 -4.67701107e-01 -5.75459421e-01 -8.86472583e-01
5.15874565e-01 -6.22982919e-01 5.09130895e-01 -4.35336202e-01
-1.00323951e+00 7.63299584e-01 1.48388669e-01 -1.01450074e+00
9.40904953e-03 -5.09098053e-01 -7.24194586e-01 6.35694146e-01
-1.49596226e+00 -1.02235925e+00 -7.79068284e-03 9.68362167e-02
5.17672658e-01 5.79173863e-01 4.62115914e-01 3.39681119e-01
-7.53817976e-01 6.17132306e-01 -2.65238822e-01 -1.28298149e-01
3.17793161e-01 -1.53512931e+00 3.96594375e-01 1.07499909e+00
2.67627761e-02 8.01684797e-01 9.41760957e-01 -6.44993544e-01
-1.40173721e+00 -8.13854098e-01 1.74832273e+00 -1.62792683e-01
7.23968089e-01 -6.00183249e-01 -6.99606836e-01 7.50899851e-01
3.35689992e-01 -3.28476220e-01 3.43841344e-01 2.67509639e-01
-2.19748504e-02 -1.50278032e-01 -1.02832794e+00 6.77224398e-01
1.32327545e+00 -3.93346161e-01 -4.18212116e-01 5.21882474e-01
7.42386699e-01 -4.64785963e-01 -6.53412640e-01 3.88670623e-01
1.74760535e-01 -1.03929245e+00 7.09406853e-01 -6.17413640e-01
4.61027473e-01 -4.08677429e-01 -5.25209047e-02 -1.00953555e+00
-4.22961533e-01 -5.19801140e-01 1.84466749e-01 1.02161467e+00
7.75085270e-01 -6.29725277e-01 7.43314385e-01 6.73831344e-01
-3.80948961e-01 -1.07685649e+00 -1.05753779e+00 -8.13879907e-01
-2.40526378e-01 -2.87681699e-01 6.76342964e-01 1.00757074e+00
4.48091149e-01 3.16174835e-01 -1.71929270e-01 2.90198267e-01
2.57119864e-01 3.20255965e-01 2.97144204e-01 -1.10597265e+00
-3.35311621e-01 -6.70335948e-01 -5.41981876e-01 -1.00530040e+00
6.86269850e-02 -1.19817924e+00 2.23301165e-02 -1.94848597e+00
-2.09441744e-02 -4.45723355e-01 -1.68125659e-01 4.65288430e-01
1.72298066e-02 3.91622856e-02 3.28769565e-01 -1.60626844e-01
-8.57583225e-01 1.58296779e-01 1.17421412e+00 -3.20421785e-01
-1.45980552e-01 -1.06292702e-02 -7.64819324e-01 3.51876587e-01
7.45729685e-01 -3.74587029e-01 -4.59689766e-01 -7.92116165e-01
7.56431699e-01 2.36683518e-01 2.80671895e-01 -6.63910568e-01
7.15001285e-01 -2.67158002e-01 -2.61492610e-01 -6.98849380e-01
2.00022355e-01 -9.78009045e-01 1.49332583e-01 4.82283533e-01
-5.67241788e-01 5.82140923e-01 6.40977621e-02 4.77479130e-01
-3.15882862e-01 -3.54874611e-01 3.84003192e-01 9.18051042e-03
-5.70874035e-01 -2.02103779e-02 -2.95765191e-01 -2.17061356e-01
1.12237597e+00 -6.52664125e-01 -4.17514741e-01 -3.21925223e-01
-8.44580352e-01 6.53634310e-01 3.52650464e-01 3.75092059e-01
5.21406054e-01 -8.74305367e-01 -5.21151483e-01 1.32672623e-01
-1.27979249e-01 1.12849966e-01 3.32031921e-02 9.57529068e-01
-7.71343768e-01 5.56260705e-01 5.54838702e-02 -3.40267271e-01
-1.39166033e+00 5.95762491e-01 1.39525741e-01 -5.21659613e-01
-4.00992602e-01 6.63479924e-01 -1.45783469e-01 -4.01169509e-01
2.70663083e-01 -6.77752435e-01 -4.41715926e-01 2.16386035e-01
-3.90867926e-02 3.19790095e-01 4.81374055e-01 -6.56097114e-01
-7.77619600e-01 5.64397812e-01 -1.61882445e-01 5.44415712e-02
1.24065602e+00 -2.44000718e-01 -5.81737280e-01 3.75080556e-01
1.14546204e+00 2.03406602e-01 -5.80703735e-01 -6.64718151e-02
4.34030712e-01 -3.32240075e-01 -1.85138449e-01 -8.14072788e-01
-7.65713751e-01 5.50522447e-01 -1.40699521e-01 7.55555391e-01
1.20838797e+00 -1.39447540e-01 7.41729558e-01 3.94963056e-01
3.52526158e-01 -1.05761898e+00 -1.40818372e-01 8.23820651e-01
8.27149689e-01 -9.67672467e-01 1.24147542e-01 -8.98392260e-01
-5.71568310e-01 1.18316269e+00 4.43665057e-01 -1.31026819e-01
4.16046411e-01 1.93162054e-01 -2.84516245e-01 -1.55046552e-01
-8.06465685e-01 -1.99390322e-01 2.37693548e-01 1.21926732e-01
4.15428251e-01 1.75480619e-01 -8.93847227e-01 4.18719947e-01
-4.79237258e-01 -3.46721858e-01 8.23603392e-01 1.03844368e+00
-4.64517236e-01 -1.27109528e+00 -8.49057063e-02 6.83496535e-01
-3.30602080e-01 -3.77214402e-01 -6.50251627e-01 6.70821190e-01
7.07288235e-02 1.22325075e+00 3.29968125e-01 -4.51802254e-01
5.21743059e-01 -1.37593105e-01 4.16891038e-01 -8.78728509e-01
-8.75719607e-01 -2.09191397e-01 5.06351829e-01 -3.04824322e-01
-1.64110109e-01 -8.30922186e-01 -1.33052564e+00 -2.74686188e-01
-3.95352572e-01 1.21226281e-01 5.20450354e-01 9.14210618e-01
3.53353053e-01 7.67279685e-01 6.49428487e-01 -4.04839009e-01
-4.56484795e-01 -5.82031846e-01 -4.68651205e-01 3.55445236e-01
4.62916315e-01 -4.41261321e-01 -3.68294597e-01 -3.09121698e-01] | [10.973098754882812, 8.867088317871094] |
54942f83-75a5-4a13-8ecb-195254d5d2ba | when-machine-unlearning-jeopardizes-privacy | 2005.02205 | null | https://arxiv.org/abs/2005.02205v2 | https://arxiv.org/pdf/2005.02205v2.pdf | When Machine Unlearning Jeopardizes Privacy | The right to be forgotten states that a data owner has the right to erase their data from an entity storing it. In the context of machine learning (ML), the right to be forgotten requires an ML model owner to remove the data owner's data from the training set used to build the ML model, a process known as machine unlearning. While originally designed to protect the privacy of the data owner, we argue that machine unlearning may leave some imprint of the data in the ML model and thus create unintended privacy risks. In this paper, we perform the first study on investigating the unintended information leakage caused by machine unlearning. We propose a novel membership inference attack that leverages the different outputs of an ML model's two versions to infer whether a target sample is part of the training set of the original model but out of the training set of the corresponding unlearned model. Our experiments demonstrate that the proposed membership inference attack achieves strong performance. More importantly, we show that our attack in multiple cases outperforms the classical membership inference attack on the original ML model, which indicates that machine unlearning can have counterproductive effects on privacy. We notice that the privacy degradation is especially significant for well-generalized ML models where classical membership inference does not perform well. We further investigate four mechanisms to mitigate the newly discovered privacy risks and show that releasing the predicted label only, temperature scaling, and differential privacy are effective. We believe that our results can help improve privacy protection in practical implementations of machine unlearning. Our code is available at https://github.com/MinChen00/UnlearningLeaks. | ['Zhikun Zhang', 'Yang Zhang', 'Tianhao Wang', 'Min Chen', 'Michael Backes', 'Mathias Humbert'] | 2020-05-05 | null | null | null | null | ['membership-inference-attack'] | ['computer-vision'] | [ 4.35567260e-01 4.65061754e-01 -4.19201761e-01 -2.00733870e-01
-7.29848504e-01 -1.31373072e+00 2.74348557e-01 2.61306226e-01
-3.66344959e-01 7.91241050e-01 -3.29861879e-01 -8.99132729e-01
2.40622610e-01 -9.23185349e-01 -1.32823217e+00 -9.50057924e-01
5.51680103e-02 2.21249089e-02 -8.73802826e-02 5.33805847e-01
1.31939471e-01 7.31564999e-01 -1.25491774e+00 2.86801010e-01
5.44786572e-01 6.97571814e-01 -5.29347599e-01 5.58392465e-01
3.28139275e-01 3.12399447e-01 -6.68730080e-01 -7.75335133e-01
8.96409750e-01 -3.29468369e-01 -6.74581766e-01 -4.35971528e-01
6.29729569e-01 -5.68306267e-01 -3.84535462e-01 1.32810521e+00
1.76093921e-01 -3.43400747e-01 1.95774809e-01 -1.80488563e+00
-3.30741316e-01 6.97531700e-01 -3.00776690e-01 -2.27451131e-01
-2.59765182e-02 2.66411692e-01 8.41826022e-01 -3.63417476e-01
5.27605474e-01 6.52954400e-01 3.67088705e-01 7.01226950e-01
-1.32717025e+00 -1.28890443e+00 -3.62504721e-01 -1.90423101e-01
-1.45998788e+00 -6.94026887e-01 3.67008090e-01 -2.86436915e-01
5.30619800e-01 8.85564864e-01 1.14606045e-01 9.63172615e-01
3.44333738e-01 6.40532553e-01 1.35364199e+00 -4.17459369e-01
3.37771088e-01 7.45175064e-01 2.08826214e-01 6.76607311e-01
1.01609397e+00 5.77725530e-01 -4.54862952e-01 -9.94587243e-01
3.65586996e-01 1.98919624e-01 -4.87139970e-01 -7.64650583e-01
-6.70787692e-01 7.28142917e-01 2.29867861e-01 -9.59805921e-02
2.03636691e-01 2.38696769e-01 2.99388409e-01 5.20027816e-01
1.01344019e-01 5.60591459e-01 -8.35212350e-01 2.30462819e-01
-5.08659661e-01 1.21024631e-01 1.25775564e+00 9.84940052e-01
1.01720667e+00 -5.59733570e-01 2.18339399e-01 -6.52221888e-02
1.59216017e-01 4.59541023e-01 2.44573668e-01 -7.81870961e-01
3.36167365e-01 3.55521142e-01 2.03023165e-01 -8.62046301e-01
2.34451160e-01 -2.54935443e-01 -5.43578625e-01 3.16929609e-01
6.53198123e-01 -3.77251059e-01 -4.77311671e-01 2.02388382e+00
4.36645776e-01 1.94478974e-01 3.49363714e-01 5.41265905e-01
1.48977458e-01 4.34650004e-01 -1.20025743e-02 -1.95009902e-01
1.26798701e+00 -6.13359213e-01 -5.09003878e-01 -6.93797544e-02
9.06742990e-01 -2.05843240e-01 7.11906433e-01 3.85207593e-01
-7.64397621e-01 4.76883575e-02 -1.42584574e+00 -2.54350882e-02
-5.73394775e-01 -3.58320951e-01 8.75632524e-01 1.18371272e+00
-7.36603320e-01 6.13128722e-01 -8.53338838e-01 -1.22286208e-01
7.62539029e-01 6.08846843e-01 -8.40816677e-01 -6.03132322e-02
-1.28482378e+00 3.83136153e-01 3.84761631e-01 -1.98971570e-01
-8.32250118e-01 -7.00298071e-01 -7.65364170e-01 3.02526541e-03
5.61699390e-01 -5.72703660e-01 9.39913154e-01 -7.31190205e-01
-9.38751101e-01 1.08568370e+00 -2.93314364e-02 -8.48854244e-01
7.78625309e-01 -2.72919890e-02 -4.07048076e-01 -8.58252943e-02
-4.59489465e-01 -4.21892339e-03 1.08806968e+00 -1.41791928e+00
-6.19412899e-01 -9.63711619e-01 -1.08057559e-01 -2.18724146e-01
-4.99673277e-01 -2.62542069e-01 -1.76606774e-01 -4.98763531e-01
-1.28324151e-01 -1.08257079e+00 -1.69938281e-01 1.62784413e-01
-6.49172544e-01 4.95838106e-01 1.19652474e+00 -4.02594149e-01
1.27136016e+00 -2.48104787e+00 -5.74874282e-01 5.27364552e-01
2.55884618e-01 6.20789587e-01 2.48940557e-01 3.96358967e-01
6.02442250e-02 6.25783741e-01 -5.06846368e-01 -3.18688601e-01
1.73477586e-02 3.98205042e-01 -7.89529026e-01 9.03252602e-01
-4.52858478e-01 8.84545743e-01 -5.40795147e-01 1.35698784e-02
-1.56115353e-01 3.01476687e-01 -5.17547667e-01 2.92899102e-01
-3.09843302e-01 3.07169557e-01 -4.26792711e-01 4.28436279e-01
1.19207096e+00 -2.00112998e-01 5.53568602e-01 1.51569650e-01
3.85298282e-01 3.12763214e-01 -9.57396567e-01 1.14776587e+00
-2.15527508e-02 2.71591038e-01 4.28930908e-01 -4.02388036e-01
5.72710395e-01 2.84281641e-01 -1.06306016e-01 -1.09557472e-01
1.45720886e-02 1.17387898e-01 -1.37563050e-01 -3.48237157e-01
3.82099181e-01 -2.31547058e-01 -3.05023223e-01 1.01211631e+00
-2.29786381e-01 4.88830656e-01 -8.88828158e-01 6.62526786e-02
1.16638613e+00 -1.75433606e-01 3.39538723e-01 -2.09202394e-01
2.43814617e-01 -4.97726202e-01 8.16571057e-01 1.02506459e+00
-3.16099584e-01 7.79065043e-02 4.94605720e-01 -1.44616991e-01
-8.47568035e-01 -1.05663431e+00 -2.01584518e-01 1.00397146e+00
9.70445946e-02 -5.03735125e-01 -8.30928028e-01 -1.37015021e+00
7.78128445e-01 8.10051322e-01 -6.30181015e-01 -3.61433864e-01
-2.33986557e-01 -4.58537012e-01 1.14058220e+00 8.94137621e-02
4.44222212e-01 -4.66344327e-01 -4.14437890e-01 -3.14145416e-01
2.85311520e-01 -7.66927660e-01 -5.63468575e-01 3.38463992e-01
-8.70251477e-01 -1.20691180e+00 2.62666196e-01 -2.78267056e-01
1.03056824e+00 1.24563470e-01 2.81050354e-01 4.54473376e-01
-6.76141530e-02 2.52645701e-01 3.72894034e-02 -6.03896499e-01
-8.24229956e-01 1.86141655e-01 1.94802837e-04 2.74513394e-01
6.55232072e-01 -5.70789278e-01 -3.96675527e-01 1.36287704e-01
-1.28412163e+00 -2.46782348e-01 2.52223134e-01 5.12680113e-01
3.14443082e-01 3.13632131e-01 2.92179048e-01 -1.77142715e+00
3.15108180e-01 -4.94525135e-01 -8.82844925e-01 3.56778681e-01
-1.00043321e+00 3.43373865e-01 7.48791695e-01 -4.87404019e-01
-7.49829113e-01 1.98010817e-01 8.43045264e-02 -4.95159239e-01
-1.88563123e-01 9.62005521e-04 -7.98714221e-01 -3.31277221e-01
3.53757441e-01 1.46090895e-01 3.51930022e-01 -6.28486753e-01
3.37779731e-01 8.57700408e-01 4.55727875e-01 -5.37187874e-01
1.09322762e+00 6.20949924e-01 2.36799151e-01 -3.52862656e-01
-4.49575275e-01 -7.31746554e-02 -2.67896265e-01 2.83838660e-01
3.17682534e-01 -6.07032955e-01 -1.01503897e+00 6.66973293e-01
-7.19709098e-01 -3.13902259e-01 -2.40384731e-02 2.90198233e-02
-1.96500555e-01 6.91378474e-01 -5.22763073e-01 -8.85440886e-01
-4.82131779e-01 -1.07160962e+00 4.91934359e-01 1.79546237e-01
-9.46800485e-02 -9.33748603e-01 -2.54787266e-01 6.05453551e-01
1.21956654e-01 4.76435989e-01 9.13053691e-01 -1.31132531e+00
-8.14136088e-01 -6.45395994e-01 2.79885978e-01 4.23290938e-01
1.53884947e-01 -1.63287967e-01 -1.33368933e+00 -7.70458758e-01
4.36890423e-01 -1.30946606e-01 7.95651674e-01 -3.85846704e-01
1.41694880e+00 -1.15935302e+00 -4.76825982e-01 9.10354853e-01
1.46767497e+00 4.18676622e-02 7.09592044e-01 1.11433037e-01
6.20888174e-01 2.97902226e-01 3.24705899e-01 6.37135208e-01
-7.08189327e-03 3.30737978e-01 4.90589350e-01 5.78134647e-03
7.12024927e-01 -8.69010925e-01 4.73734409e-01 2.99217664e-02
6.61346316e-01 -1.64801344e-01 -3.35835397e-01 1.51090547e-01
-1.72530925e+00 -8.12059164e-01 2.09049694e-02 2.93999505e+00
1.13253129e+00 8.19318444e-02 -1.75962538e-01 -9.06583201e-03
5.29516935e-01 9.22260247e-03 -7.77103961e-01 -8.04140568e-01
4.67476211e-02 2.98991203e-01 1.35201025e+00 6.53842211e-01
-1.20018613e+00 8.47973764e-01 5.35050011e+00 6.02926254e-01
-1.06033421e+00 2.02858761e-01 7.19500959e-01 -1.45447806e-01
-5.96767664e-01 5.09964824e-01 -8.70521247e-01 5.61496913e-01
1.04033732e+00 -5.24850845e-01 6.08654737e-01 9.53296125e-01
-4.66695368e-01 -9.63737443e-03 -1.57686770e+00 5.46104729e-01
-1.23245262e-01 -1.15292311e+00 -2.52982289e-01 8.01447272e-01
2.18827263e-01 -3.71152520e-01 3.47329915e-01 1.89128697e-01
5.28777003e-01 -1.00289822e+00 3.50611061e-01 5.22899508e-01
8.13308299e-01 -1.12590110e+00 3.60734582e-01 7.09699214e-01
-4.28545386e-01 -1.28696606e-01 -3.48803073e-01 -3.02273761e-02
-5.21012902e-01 5.27545571e-01 -1.26204407e+00 4.38608497e-01
3.69955361e-01 3.04364618e-02 -6.54493093e-01 5.36590397e-01
-5.05339503e-01 8.13669980e-01 -4.58495378e-01 1.45149931e-01
-1.24922201e-01 -1.28858849e-01 6.52850568e-01 9.87707853e-01
1.10947333e-01 2.53284797e-02 -1.82348460e-01 1.02986097e+00
-6.31277859e-01 -1.60930738e-01 -9.94173944e-01 -1.75752595e-01
8.35663199e-01 9.19431686e-01 -5.98861352e-02 -7.03217909e-02
-7.43609667e-02 1.17424381e+00 3.16438884e-01 2.80880690e-01
-5.24577081e-01 -2.77104259e-01 1.11654782e+00 3.15905139e-02
3.15895557e-01 4.67344522e-02 -4.35919195e-01 -1.14712501e+00
6.24041073e-02 -9.29590404e-01 8.82667422e-01 -1.86499014e-01
-1.18381584e+00 -5.67898266e-02 -4.00521934e-01 -8.10706973e-01
-3.48180085e-02 -2.37214223e-01 -3.27271312e-01 9.58327115e-01
-1.05888557e+00 -9.99450505e-01 4.11363810e-01 4.63223547e-01
-5.91639519e-01 9.86463279e-02 1.18600750e+00 -2.53330290e-01
-6.36390865e-01 1.42546594e+00 4.41159934e-01 3.07353377e-01
7.49742031e-01 -9.62502480e-01 4.22914535e-01 1.23798132e+00
1.76772684e-01 1.32124519e+00 5.99459231e-01 -7.61002243e-01
-2.01806927e+00 -1.25776744e+00 9.38824296e-01 -7.28513896e-01
2.84219980e-01 -9.94686127e-01 -1.24556839e+00 1.26583791e+00
-1.22752026e-01 6.95587099e-02 1.29903293e+00 -1.73996091e-01
-9.71005738e-01 -2.14859873e-01 -1.82697117e+00 4.31291312e-01
4.67353076e-01 -8.66006076e-01 -1.58537388e-01 -6.36755005e-02
9.64512289e-01 -2.04759642e-01 -7.94854343e-01 2.78013676e-01
8.22883070e-01 -7.88108766e-01 6.92002356e-01 -8.69450510e-01
-2.73084283e-01 -3.14372748e-01 -3.03318113e-01 -6.41696990e-01
2.67254442e-01 -1.05534172e+00 -6.36464357e-01 1.36983335e+00
5.77108502e-01 -1.26585186e+00 1.04761887e+00 1.36965680e+00
8.27700317e-01 -2.76495993e-01 -1.08778811e+00 -6.80594385e-01
4.24741060e-01 -3.83881599e-01 1.19586706e+00 1.31546950e+00
1.46530405e-01 -1.42513245e-01 -5.89803755e-01 7.90035784e-01
8.90647829e-01 1.88362330e-01 1.15578449e+00 -8.76457512e-01
-6.26204073e-01 1.39055699e-01 -1.68575600e-01 -7.07592309e-01
2.48856470e-01 -1.16566467e+00 -3.48621041e-01 -4.60464060e-01
3.14701468e-01 -6.75559580e-01 -5.88809788e-01 8.50657821e-01
-1.91137120e-01 -1.12527162e-01 2.80348301e-01 1.84748396e-01
-1.04414806e-01 1.65460184e-02 8.69523525e-01 1.51666924e-01
2.60552932e-02 4.70912576e-01 -1.16618764e+00 2.72490710e-01
1.00610387e+00 -8.58797431e-01 -3.92291278e-01 2.30756015e-01
1.65959209e-01 1.30166626e-02 4.27500725e-01 -7.05973089e-01
3.56173128e-01 -2.19347447e-01 1.15672395e-01 -2.00008407e-01
-1.99515745e-03 -1.31956840e+00 7.19105959e-01 6.03674591e-01
-4.83626425e-01 -4.65663522e-01 1.87105194e-01 7.99628854e-01
4.66600090e-01 -2.37504676e-01 7.81969726e-01 8.93936865e-03
-1.82057098e-01 4.94759858e-01 -8.74480829e-02 -1.90718964e-01
1.25852287e+00 -2.30996199e-02 -5.89929521e-01 -1.57255262e-01
-6.29005671e-01 1.48117855e-01 1.31744134e+00 6.72991425e-02
4.30762142e-01 -8.30584764e-01 -2.08191231e-01 8.39954317e-01
1.84185401e-01 -2.32818872e-01 -4.50869128e-02 4.58559066e-01
-2.12159172e-01 1.63681716e-01 1.71477064e-01 -7.44777638e-03
-1.74450088e+00 1.10708594e+00 3.32512110e-01 -1.79036066e-01
-5.13664246e-01 6.53652608e-01 1.98896974e-01 -5.60844004e-01
4.77943629e-01 -5.83694577e-02 8.10683370e-01 -3.70417506e-01
6.67816460e-01 2.32138664e-01 7.70900026e-02 -3.02613258e-01
-2.32293710e-01 -3.10970306e-01 -4.06225920e-01 -8.07516128e-02
8.88461947e-01 -9.40283388e-02 -3.99716258e-01 4.73249741e-02
1.64886796e+00 4.73390311e-01 -1.11387038e+00 -3.50431591e-01
-2.91756876e-02 -8.34000230e-01 -1.61735788e-01 -8.14668834e-01
-9.75917935e-01 7.74159789e-01 4.79597420e-01 5.87938353e-02
1.07123947e+00 -1.34204134e-01 8.78979862e-01 4.66127485e-01
7.92541146e-01 -7.70976901e-01 -7.89128423e-01 -1.58734813e-01
3.40661883e-01 -9.06357765e-01 1.88921824e-01 -3.08965355e-01
-4.74394470e-01 6.48480177e-01 4.62714136e-01 4.03487384e-01
8.06253970e-01 5.22305310e-01 -7.17059299e-02 2.84591079e-01
-8.40154886e-01 5.83199143e-01 -2.82393515e-01 5.36454201e-01
-4.27463055e-01 4.71383661e-01 -3.48125100e-02 1.00044334e+00
-1.88533053e-01 8.67620856e-02 6.00021124e-01 1.37559617e+00
-2.00285867e-01 -1.59486926e+00 -4.47302401e-01 4.25828695e-01
-8.86746466e-01 1.88053418e-02 -7.51418114e-01 5.66032231e-01
1.64134484e-02 6.80635154e-01 -3.02295208e-01 -4.92772549e-01
-2.89813764e-02 3.78273249e-01 2.33381957e-01 -4.76038128e-01
-8.29212606e-01 -3.17683518e-01 9.72418636e-02 -5.61060131e-01
1.27302855e-01 -6.14362478e-01 -1.31170821e+00 -7.38152146e-01
-4.39741582e-01 2.90720642e-01 5.61135411e-01 4.24618542e-01
8.00428867e-01 -4.35857385e-01 1.01714051e+00 2.34389395e-01
-8.81669700e-01 -3.60685617e-01 -9.21799183e-01 3.78407955e-01
6.35641277e-01 -5.04936427e-02 -7.63503730e-01 4.33887728e-02] | [5.897848129272461, 7.106816291809082] |
e6ab5cc8-2455-4717-9940-0171b0912a7a | biomedical-ner-using-novel-schema-and-distant | null | null | https://aclanthology.org/2022.bionlp-1.15 | https://aclanthology.org/2022.bionlp-1.15.pdf | Biomedical NER using Novel Schema and Distant Supervision | Biomedical Named Entity Recognition (BMNER) is one of the most important tasks in the field of biomedical text mining. Most work so far on this task has not focused on identification of discontinuous and overlapping entities, even though they are present in significant fractions in real-life biomedical datasets. In this paper, we introduce a novel annotation schema to capture complex entities, and explore the effects of distant supervision on our deep-learning sequence labelling model. For BMNER task, our annotation schema outperforms other BIO-based annotation schemes on the same model. We also achieve higher F1-scores than state-of-the-art models on multiple corpora without fine-tuning embeddings, highlighting the efficacy of neural feature extraction using our model. | ['Kamalakar Karlapalem', 'Veera Raghavendra Chikka', 'Alok Kar', 'Anshita Khandelwal'] | null | null | null | null | bionlp-acl-2022-5 | ['medical-named-entity-recognition'] | ['natural-language-processing'] | [ 2.08883733e-01 5.33081651e-01 -2.82925963e-01 -3.90127271e-01
-7.07967281e-01 -2.10899487e-01 4.49032784e-01 8.30663562e-01
-1.09609509e+00 1.10485148e+00 3.84668022e-01 -1.64317921e-01
-3.76856737e-02 -5.04026532e-01 -6.41580939e-01 -5.43807685e-01
-8.40691626e-02 6.58102334e-01 2.82807916e-01 1.71648003e-02
-1.32726595e-01 4.14245039e-01 -7.46405661e-01 2.12974265e-01
7.50801682e-01 5.26218116e-01 -2.50824112e-02 3.31895471e-01
-2.00343460e-01 4.45520908e-01 -6.73384011e-01 -8.36105824e-01
-3.94880921e-01 -2.34630957e-01 -1.06323040e+00 -2.94755012e-01
3.35877156e-03 1.04493253e-01 -3.48487437e-01 9.92216110e-01
8.91494632e-01 -1.78064361e-01 7.23388851e-01 -6.63922727e-01
-4.89077359e-01 6.87757432e-01 -3.42165977e-01 4.07628924e-01
1.03567630e-01 -2.29141891e-01 1.10714412e+00 -6.09896958e-01
1.16942811e+00 5.69260597e-01 8.80789757e-01 6.56197131e-01
-1.33624685e+00 -6.70281947e-01 -1.42106488e-01 1.48335889e-01
-1.40368736e+00 -3.96064222e-01 2.80199140e-01 -4.12130862e-01
1.11090469e+00 -7.80680329e-02 4.00837988e-01 1.24330401e+00
4.12033260e-01 8.08636963e-01 5.29610038e-01 -3.53179187e-01
1.43107012e-01 -3.20734875e-03 3.50931674e-01 6.45577133e-01
7.01516747e-01 -3.61944079e-01 -3.45074236e-01 -4.05663460e-01
5.24351895e-01 2.07198039e-03 -2.70408690e-01 -1.91106826e-01
-1.28354228e+00 7.82150865e-01 2.85866261e-01 9.54606235e-01
-4.23437476e-01 4.77615464e-03 6.88446462e-01 -1.72304496e-01
6.18957937e-01 7.94343174e-01 -8.78201425e-01 1.85542239e-03
-8.90930831e-01 -1.85891762e-01 7.96242237e-01 7.61229694e-01
3.47460449e-01 -4.37142611e-01 -3.20138395e-01 7.55921245e-01
6.29179599e-03 -2.54464358e-01 7.55931675e-01 -1.32533461e-01
2.09893808e-01 8.00852418e-01 -7.71739148e-03 -5.07428765e-01
-9.76787210e-01 -8.06113183e-01 -8.58843446e-01 -4.96423960e-01
4.32758659e-01 -5.15145838e-01 -8.42886329e-01 1.63391590e+00
4.42818046e-01 2.44152591e-01 1.24201864e-01 5.03489375e-01
9.19025064e-01 1.62036493e-01 6.36290014e-01 7.27052242e-03
1.67585135e+00 -7.77296901e-01 -1.04340565e+00 -7.40200803e-02
1.05577695e+00 -3.50003660e-01 2.15318650e-01 1.43227234e-01
-5.55920243e-01 -1.51529029e-01 -8.42036605e-01 -2.19361767e-01
-7.15563297e-01 2.91819006e-01 7.65349269e-01 5.91791868e-01
-6.70121610e-01 6.23922527e-01 -9.69791472e-01 -5.13226211e-01
7.37840652e-01 4.66644406e-01 -8.55849862e-01 5.63715398e-02
-1.46075654e+00 1.10704458e+00 6.55806005e-01 -3.27792242e-02
-4.76578027e-01 -9.03003454e-01 -8.54963839e-01 2.35732958e-01
2.16474712e-01 -6.88908517e-01 9.87513483e-01 -4.49683994e-01
-8.12187493e-01 1.20108044e+00 -1.14735097e-01 -6.23706639e-01
2.51488149e-01 -3.78179640e-01 -6.61477268e-01 7.38403276e-02
1.52311489e-01 7.60771930e-01 -6.65652603e-02 -5.83681345e-01
-3.60838443e-01 -4.75059777e-01 -2.51717925e-01 -1.57689363e-01
-5.74135661e-01 8.10614079e-02 -4.11312580e-01 -6.55903518e-01
-4.60340112e-01 -7.68161833e-01 -6.84372663e-01 -5.26885130e-02
-3.13649178e-01 -5.33829868e-01 2.42710456e-01 -6.37843609e-01
1.31344557e+00 -2.14845967e+00 -4.13723812e-02 -1.57748207e-01
4.14037406e-01 6.02543533e-01 8.18006992e-02 6.17981970e-01
-2.10838601e-01 3.36540341e-01 -1.51858807e-01 -2.20846146e-01
-1.67083982e-02 3.15408185e-02 1.61759138e-01 5.06412029e-01
5.73942959e-01 1.12365198e+00 -9.61465061e-01 -5.65252304e-01
-1.76859438e-01 6.41605496e-01 -3.28060716e-01 -9.23932865e-02
-1.02660239e-01 4.85199332e-01 -5.45498371e-01 3.56166095e-01
2.87392288e-01 -6.57346427e-01 7.59876430e-01 -2.91609317e-01
2.49949202e-01 3.08010846e-01 -6.07445836e-01 1.95095658e+00
-1.98657334e-01 5.31713486e-01 -3.39860767e-01 -1.17175436e+00
6.11247718e-01 7.38857150e-01 7.74845123e-01 -5.60605168e-01
1.96547091e-01 1.55424654e-01 1.15208283e-01 -5.57837903e-01
3.27353656e-01 -2.06906691e-01 -1.64107010e-01 5.24852499e-02
5.01023412e-01 7.96475708e-01 2.54473895e-01 8.20912868e-02
1.40534151e+00 4.95853014e-02 7.71315873e-01 -2.20494047e-01
3.62345666e-01 4.02000453e-03 8.44238937e-01 4.83929724e-01
-2.99733758e-01 3.49025309e-01 5.84923744e-01 -4.70116764e-01
-8.13746333e-01 -4.14591074e-01 -7.34441996e-01 9.53140914e-01
-1.42972246e-01 -5.16530454e-01 -6.62967443e-01 -1.04701650e+00
-9.39723551e-02 4.79327887e-01 -8.73236358e-01 -2.32951194e-02
-4.18822080e-01 -1.32717681e+00 1.08509219e+00 5.57237625e-01
1.33685082e-01 -1.01588225e+00 -4.98847336e-01 4.75910336e-01
-1.17270991e-01 -1.47106957e+00 -4.12078530e-01 4.57966685e-01
-6.60760045e-01 -1.15657616e+00 -8.55401814e-01 -8.02786469e-01
5.77582657e-01 -6.42898858e-01 1.21998990e+00 -1.43023714e-01
-6.17069840e-01 -4.05857228e-02 -3.04152757e-01 -6.73075438e-01
-2.44875506e-01 5.02252042e-01 -3.18822831e-01 -9.10589546e-02
7.92227149e-01 -1.63325116e-01 -6.64556503e-01 1.72246099e-01
-9.21312630e-01 1.17702025e-03 9.02118862e-01 1.01651382e+00
5.33173382e-01 -2.76410908e-01 9.77792740e-01 -1.54484832e+00
3.93477499e-01 -5.87972641e-01 -2.06509843e-01 3.38685960e-01
-5.53747237e-01 2.46679828e-01 3.84501845e-01 -4.39792275e-01
-8.26487303e-01 1.80348873e-01 -5.44717431e-01 1.93163261e-01
-4.67455089e-01 7.24125504e-01 -1.37921602e-01 3.20115685e-01
5.52870214e-01 -1.06700428e-01 -4.99444842e-01 -5.61643183e-01
1.77124754e-01 6.70810401e-01 3.29741180e-01 -2.18871281e-01
-6.10800311e-02 2.83382297e-01 4.93475981e-02 -8.36900711e-01
-1.06424725e+00 -8.54643226e-01 -7.14386940e-01 4.78935301e-01
1.32876086e+00 -9.31240559e-01 -5.72242141e-01 2.03776971e-01
-1.12119365e+00 -6.33101091e-02 -8.49727243e-02 6.18857265e-01
-5.13222627e-02 1.25870615e-01 -8.63867223e-01 -4.54281121e-01
-4.66799498e-01 -9.12767887e-01 1.25283015e+00 7.00890794e-02
-5.59699893e-01 -1.14167917e+00 4.05685753e-01 3.28416526e-01
1.26914129e-01 4.14391965e-01 1.03820848e+00 -1.45893312e+00
-5.18363975e-02 -2.75480002e-01 -2.58032531e-01 -2.28630900e-01
1.61530346e-01 -4.06154633e-01 -9.24579859e-01 -1.98408216e-02
-4.86772090e-01 -1.37391657e-01 1.07544434e+00 2.10325554e-01
1.01015496e+00 -6.09152801e-02 -9.74168837e-01 3.09970319e-01
1.34255540e+00 8.55427012e-02 4.28485096e-01 2.82774150e-01
5.99339128e-01 5.31158984e-01 2.42523536e-01 1.97766945e-01
2.64028877e-01 5.33556461e-01 1.65347740e-01 -4.80403841e-01
1.70492619e-01 -3.42176408e-02 -1.04024895e-01 4.68487144e-01
2.29984909e-01 -4.43921328e-01 -1.18305814e+00 9.25869465e-01
-1.82773960e+00 -7.35450447e-01 -1.28788695e-01 1.76959562e+00
1.26302016e+00 1.22899383e-01 -9.73947048e-02 -1.33686960e-01
8.01924884e-01 -1.55503511e-01 -4.52575058e-01 -6.26736209e-02
-6.23374954e-02 5.84241807e-01 5.71995676e-01 -7.58773834e-02
-1.36753213e+00 7.23509789e-01 5.98910379e+00 6.66557729e-01
-8.94280314e-01 2.55286634e-01 7.65136480e-01 2.28793368e-01
1.82008538e-02 -3.15344363e-01 -1.05048692e+00 4.46105689e-01
1.20896733e+00 4.30146651e-03 -4.11715150e-01 6.27765059e-01
-9.33615416e-02 1.87920704e-01 -1.24751544e+00 7.30004489e-01
6.07452989e-02 -1.48547292e+00 -2.62806237e-01 2.24241495e-01
6.81531131e-01 3.23770612e-01 -4.02822256e-01 2.87352294e-01
3.84680778e-01 -1.15401340e+00 1.92547083e-01 5.11832297e-01
7.98649013e-01 -5.73708117e-01 1.35617840e+00 1.36145711e-01
-7.09670782e-01 1.39596865e-01 -3.01950902e-01 5.37519574e-01
2.54802823e-01 1.03228903e+00 -1.01275849e+00 7.36820042e-01
5.10004878e-01 6.67292237e-01 -5.69035292e-01 1.27942848e+00
-2.27653086e-01 7.13193297e-01 -1.56831354e-01 -1.40613452e-01
2.33060587e-02 2.63136446e-01 1.31915778e-01 1.67075050e+00
-1.27966991e-02 -4.57731746e-02 2.37926710e-02 5.40364623e-01
-7.66541600e-01 3.95361185e-01 -2.99839020e-01 -4.14348990e-01
2.46971712e-01 1.32529891e+00 -8.42418432e-01 -2.92902082e-01
-3.92366946e-01 8.60302508e-01 6.53160810e-01 -1.78999687e-03
-7.66766667e-01 -5.00679970e-01 5.04975855e-01 -2.48390008e-02
4.93739933e-01 -4.57881726e-02 -7.92266503e-02 -1.11112237e+00
-3.43788773e-01 -5.58215320e-01 6.90179646e-01 -2.33135849e-01
-1.58446205e+00 5.51181555e-01 -4.11798298e-01 -7.22800076e-01
7.01657534e-02 -6.77474856e-01 -2.22465783e-01 5.48491895e-01
-1.51787281e+00 -1.06492054e+00 1.83692858e-01 8.73589739e-02
1.35223866e-01 1.59427240e-01 1.27962363e+00 7.66323268e-01
-8.16837609e-01 6.90798938e-01 3.68776828e-01 6.57650769e-01
1.08369398e+00 -1.19296300e+00 5.04827201e-01 3.71070683e-01
3.05499136e-01 8.34249258e-01 3.56004268e-01 -6.48386478e-01
-8.13147247e-01 -1.18089581e+00 1.36262918e+00 -5.06914437e-01
7.47513056e-01 -4.71073508e-01 -9.72262442e-01 7.36992836e-01
2.37641394e-01 1.77934647e-01 1.27755654e+00 5.22347987e-01
-2.44505316e-01 3.93740207e-01 -1.04196179e+00 1.92357317e-01
1.01718307e+00 -3.99555892e-01 -6.67304516e-01 3.66641670e-01
6.58852220e-01 -3.38223964e-01 -1.30318999e+00 4.94116366e-01
5.44673443e-01 -4.09727186e-01 8.50530326e-01 -1.11931252e+00
3.69526327e-01 -6.19846508e-02 3.64179403e-01 -1.08437085e+00
-2.70722926e-01 -2.08578497e-01 3.05969100e-02 1.22734737e+00
7.57022381e-01 -6.18042409e-01 7.56107211e-01 3.84972274e-01
-9.86580774e-02 -9.44246173e-01 -9.23980474e-01 -4.32880521e-01
2.07809612e-01 -3.64283584e-02 4.55525964e-01 1.33385229e+00
2.23639548e-01 6.02697730e-01 -6.58967197e-02 1.27736360e-01
1.82103425e-01 -1.16933353e-01 2.38971770e-01 -1.39644170e+00
-1.83047354e-01 -2.96665698e-01 -6.30180717e-01 -4.49737191e-01
3.21121067e-01 -1.02079821e+00 -1.18741661e-01 -1.76871789e+00
4.38316405e-01 -3.27663124e-01 -7.71822393e-01 7.88613677e-01
-4.91035581e-01 3.49105984e-01 -3.15272242e-01 -1.34943381e-01
-8.90267968e-01 3.35959524e-01 6.76875830e-01 -1.13940589e-01
8.98247436e-02 -4.33683008e-01 -7.34811366e-01 5.96203864e-01
7.68256247e-01 -9.00107086e-01 1.52320072e-01 -4.31698561e-01
3.31706166e-01 -1.15979910e-01 1.36689069e-02 -8.31306756e-01
2.09488809e-01 2.01960191e-01 5.67768335e-01 -2.96378195e-01
-6.89265504e-02 -7.61238694e-01 2.58437753e-01 3.86174232e-01
-5.44318676e-01 -1.09522559e-01 3.42367291e-01 7.25464880e-01
-1.13347121e-01 -4.04104620e-01 4.73784566e-01 6.80889636e-02
-4.70391214e-01 2.89443046e-01 -5.20174742e-01 2.41557166e-01
9.53815699e-01 2.21891329e-01 -4.34581459e-01 3.56202751e-01
-1.07367802e+00 2.37932593e-01 1.79741278e-01 2.71845818e-01
5.84970452e-02 -9.05517817e-01 -7.15311229e-01 -2.43400544e-01
4.17368889e-01 1.29321509e-03 2.61388898e-01 8.93897057e-01
-4.25109118e-01 6.99322820e-01 -5.03895655e-02 -3.93510014e-01
-1.17848182e+00 5.28073430e-01 3.02576959e-01 -1.00569296e+00
-4.44744587e-01 7.27896988e-01 4.15295899e-01 -5.82023382e-01
7.00234026e-02 2.75540445e-03 -5.59616208e-01 2.18028918e-01
3.24438840e-01 9.45567936e-02 3.70166451e-01 -3.65425795e-01
-5.87366760e-01 7.18330368e-02 -4.11782771e-01 3.22480917e-01
1.57800138e+00 3.46015453e-01 -1.40044149e-02 2.94613421e-01
1.11069489e+00 -4.26606461e-02 -4.71508801e-01 -2.40565762e-01
6.46827877e-01 2.16096535e-01 -1.26455128e-01 -9.38582718e-01
-7.91663289e-01 7.67022550e-01 5.40910661e-01 -2.24318877e-01
6.82597637e-01 1.52205765e-01 7.48883069e-01 4.72684324e-01
1.36960879e-01 -8.29322815e-01 -3.25341612e-01 3.30500782e-01
2.79699773e-01 -1.13148367e+00 4.66594705e-03 -3.79679948e-01
-5.54515779e-01 7.28892446e-01 3.53051066e-01 2.39366040e-01
5.73407054e-01 2.42518932e-01 -1.38696790e-01 -3.74094129e-01
-7.33234763e-01 -3.22853833e-01 3.24350625e-01 4.39493567e-01
9.44603622e-01 -1.55090913e-01 -6.43355608e-01 9.17099953e-01
2.10874662e-01 2.78797150e-01 2.25941300e-01 8.21932554e-01
-4.59175296e-02 -1.31546366e+00 2.69463420e-01 6.63475990e-01
-1.46926248e+00 -3.25101882e-01 -3.61389399e-01 7.15455353e-01
2.65477777e-01 6.35577798e-01 -7.11636096e-02 9.12452117e-02
2.21371531e-01 3.89271945e-01 3.70284915e-01 -7.65421510e-01
-8.46860468e-01 2.11902093e-02 4.09186840e-01 -2.10038543e-01
-5.90649307e-01 -5.25685549e-01 -1.48533154e+00 2.02194378e-01
-7.70712256e-01 4.35899079e-01 5.37090242e-01 9.69137967e-01
7.12822080e-01 8.73793900e-01 -1.02486938e-01 -2.28694290e-01
-1.71482250e-01 -1.08022523e+00 -4.28875715e-01 5.97344697e-01
1.04310038e-02 -4.75751907e-01 2.52685666e-01 1.13686785e-01] | [8.566058158874512, 8.791389465332031] |
6c0e7527-9753-4e80-a2af-a4f0e02fb590 | deep-attention-recurrent-q-network | 1512.01693 | null | http://arxiv.org/abs/1512.01693v1 | http://arxiv.org/pdf/1512.01693v1.pdf | Deep Attention Recurrent Q-Network | A deep learning approach to reinforcement learning led to a general learner
able to train on visual input to play a variety of arcade games at the human
and superhuman levels. Its creators at the Google DeepMind's team called the
approach: Deep Q-Network (DQN). We present an extension of DQN by "soft" and
"hard" attention mechanisms. Tests of the proposed Deep Attention Recurrent
Q-Network (DARQN) algorithm on multiple Atari 2600 games show level of
performance superior to that of DQN. Moreover, built-in attention mechanisms
allow a direct online monitoring of the training process by highlighting the
regions of the game screen the agent is focusing on when making decisions. | ['Anastasiia Ignateva', 'Aleksandr Fedorov', 'Mikhail Pavlov', 'Alexey Seleznev', 'Ivan Sorokin'] | 2015-12-05 | null | null | null | null | ['deep-attention', 'hard-attention', 'deep-attention'] | ['computer-vision', 'methodology', 'natural-language-processing'] | [-5.20179272e-01 1.30921677e-01 5.84977381e-02 1.95839763e-01
-6.47069886e-02 -4.77040052e-01 6.54747486e-01 -1.99008554e-01
-8.30955744e-01 7.13520229e-01 -1.41239971e-01 -5.06985307e-01
-2.88326830e-01 -8.42465401e-01 -5.01924217e-01 -4.61081445e-01
-2.83840299e-01 5.99491000e-01 3.50074410e-01 -8.65999281e-01
5.73411405e-01 6.26139343e-01 -1.53613782e+00 1.13742292e-01
5.10637403e-01 6.74350202e-01 3.93704057e-01 1.22734797e+00
5.34054711e-02 1.52307749e+00 -1.06742716e+00 -3.09258699e-01
5.28454363e-01 -6.12124324e-01 -7.55618393e-01 -3.05325121e-01
-3.13869938e-02 -4.99273747e-01 -5.23290575e-01 1.01184809e+00
7.27744699e-01 5.51583946e-01 3.43581766e-01 -1.44325030e+00
-7.39077449e-01 2.42942080e-01 -4.30956900e-01 1.06584132e+00
4.53306377e-01 6.30489349e-01 8.20987880e-01 -5.25489032e-01
4.89028931e-01 1.21957469e+00 4.31447208e-01 6.80720448e-01
-7.08177865e-01 -5.24524033e-01 -7.49035180e-02 4.90074009e-01
-7.54271805e-01 8.03087130e-02 5.40285647e-01 -3.42621058e-01
1.43630219e+00 -1.61601707e-01 9.57299173e-01 1.10120344e+00
4.75291193e-01 8.86314213e-01 1.31122243e+00 -4.03501302e-01
5.31253040e-01 -7.04474971e-02 -3.04213703e-01 6.85466349e-01
-6.36588112e-02 9.47842598e-01 -5.47098279e-01 1.65620923e-01
1.41721702e+00 8.07401165e-02 3.22796106e-01 -3.82400990e-01
-7.59337962e-01 1.07757163e+00 7.25769639e-01 1.87381372e-01
-8.47860813e-01 4.70334291e-01 5.70238531e-01 6.49910212e-01
1.51584491e-01 6.07348859e-01 -3.10757071e-01 -6.35866463e-01
-4.60956901e-01 5.04132032e-01 4.71665472e-01 7.09967911e-01
4.51113522e-01 5.76124310e-01 -3.07120264e-01 4.45199579e-01
7.18720108e-02 1.68687865e-01 8.63083601e-01 -1.08335650e+00
3.42699409e-01 5.89936137e-01 2.96925902e-01 -6.69531345e-01
-6.03188932e-01 -4.19656307e-01 -4.44263428e-01 1.35360897e+00
3.58880758e-01 -6.54633939e-01 -9.97443497e-01 1.30166626e+00
9.18106064e-02 -8.57286230e-02 2.92712390e-01 1.08484495e+00
8.48033309e-01 6.51726007e-01 3.82798880e-01 2.99174130e-01
1.10566521e+00 -1.08940589e+00 -5.32421529e-01 -1.34510040e-01
3.65532070e-01 -2.43802831e-01 1.06310081e+00 7.16441393e-01
-1.31053722e+00 -1.03453517e+00 -1.10443449e+00 2.60504872e-01
-4.43852246e-01 -4.49815303e-01 7.55655110e-01 5.65375865e-01
-1.35669136e+00 9.19653118e-01 -6.11027181e-01 -3.33679557e-01
3.84926945e-01 5.86751997e-01 -1.12574153e-01 3.10042799e-01
-1.30765462e+00 1.25104010e+00 4.84715998e-01 -7.99763650e-02
-1.86177897e+00 -2.01103970e-01 -5.49289525e-01 7.10024238e-02
4.28768426e-01 -5.60979605e-01 1.57528877e+00 -1.47884250e+00
-2.04096794e+00 7.88701653e-01 7.48031497e-01 -7.11435616e-01
7.66048312e-01 -4.60171938e-01 -1.43766522e-01 3.42242211e-01
-6.87840283e-02 7.88092554e-01 8.22762430e-01 -7.28046656e-01
-1.02496994e+00 -2.86447048e-01 7.51006663e-01 7.60925293e-01
2.68964946e-01 1.85457826e-01 -7.56178126e-02 -6.31222308e-01
-7.25109696e-01 -6.85339332e-01 -4.63479668e-01 -2.37407818e-01
1.10353999e-01 -7.47077048e-01 7.27098584e-01 -3.41688842e-01
7.52716184e-01 -2.01191044e+00 1.99814335e-01 -5.24873883e-02
3.76830161e-01 5.53415358e-01 -2.37259254e-01 6.59582913e-01
-2.35207811e-01 -3.14709961e-01 3.76087278e-01 1.95759043e-01
9.28247496e-02 4.88751233e-01 1.46547258e-01 2.49969169e-01
-1.12245798e-01 1.11727524e+00 -1.15826261e+00 -1.40445262e-01
3.70535135e-01 8.09109285e-02 -5.13395667e-01 7.25144506e-01
-2.24278465e-01 3.52043778e-01 -2.98766255e-01 4.38459843e-01
8.28078464e-02 -6.41498268e-02 -1.41941696e-01 6.48929119e-01
-1.88020244e-01 1.33820787e-01 -9.48813200e-01 1.69941473e+00
-1.82966322e-01 6.32210135e-01 -1.16743051e-01 -8.54557335e-01
8.84233356e-01 4.99720812e-01 3.12310219e-01 -1.43049526e+00
2.21258894e-01 -2.26035014e-01 3.63762677e-01 -6.91649735e-01
6.30495071e-01 -2.28647172e-01 1.74309522e-01 5.56857109e-01
4.58752453e-01 1.85849950e-01 2.18416288e-01 2.73238510e-01
1.23560262e+00 5.53078353e-01 3.04427683e-01 -3.84045929e-01
3.09063971e-01 1.49380267e-01 2.08832607e-01 1.13641214e+00
-7.41180182e-01 -4.93882224e-02 8.58576298e-01 -8.23719382e-01
-1.20909238e+00 -9.05910969e-01 5.50273836e-01 1.91228557e+00
2.69809868e-02 -1.88492686e-01 -9.10852253e-01 -8.37922692e-01
-2.08340749e-01 5.30536652e-01 -1.12682855e+00 -2.49933571e-01
-4.93262082e-01 -3.40704948e-01 3.93370152e-01 6.96030438e-01
6.04320645e-01 -2.16246033e+00 -1.49357450e+00 3.12409908e-01
6.41240656e-01 -4.81267840e-01 -3.46896462e-02 6.90222383e-01
-7.24252284e-01 -1.05635822e+00 -8.78994465e-01 -5.58889210e-01
1.58344045e-01 -9.66503546e-02 1.42098975e+00 1.26099931e-02
-2.51307100e-01 4.34238732e-01 -6.42556012e-01 -5.51648140e-01
-3.88356239e-01 -3.17058712e-02 -4.81323414e-02 -7.33140528e-01
4.00165111e-01 -5.08543849e-01 -8.28278363e-01 1.67585433e-01
-5.78057349e-01 -1.05400227e-01 6.27736390e-01 7.57023752e-01
2.59842374e-03 -1.20829776e-01 5.00369370e-01 -7.90436447e-01
1.17111504e+00 -5.30991316e-01 -9.46924031e-01 -1.55305699e-01
-5.53932548e-01 1.24912366e-01 6.51714981e-01 -3.11655492e-01
-8.14852595e-01 -3.65538329e-01 -2.79445082e-01 -3.93525422e-01
-7.41232261e-02 7.10955262e-02 2.67579705e-01 -8.95785093e-02
9.36711371e-01 8.59118849e-02 3.32863368e-02 -1.26151651e-01
3.01040977e-01 3.08448046e-01 5.82326353e-01 -3.53296906e-01
5.19384980e-01 -5.65746129e-02 -9.35361534e-02 -3.32087427e-01
-3.87157559e-01 -8.43393952e-02 -3.38966042e-01 -5.69702804e-01
1.01657879e+00 -6.48912847e-01 -1.36046386e+00 4.37899709e-01
-7.51202822e-01 -8.64640236e-01 -6.19957268e-01 2.16528520e-01
-9.10079658e-01 -2.24883646e-01 -8.19863558e-01 -7.94742882e-01
-3.63861024e-01 -8.99734199e-01 3.54683489e-01 7.96299458e-01
4.87396121e-03 -1.11162293e+00 7.11959898e-01 -2.39115041e-02
4.76815104e-01 4.78404500e-02 5.35430193e-01 -6.33204579e-01
-2.44717941e-01 1.66705012e-01 -7.15956837e-02 2.87997425e-01
-1.34250671e-01 -1.07836582e-01 -9.48762476e-01 -3.19482416e-01
-2.97705084e-01 -6.66649759e-01 3.88778359e-01 4.50448334e-01
7.86574244e-01 4.54516237e-04 2.66823649e-01 4.85512555e-01
1.54317880e+00 7.64825940e-01 8.69139671e-01 8.83146286e-01
3.20135683e-01 2.79292941e-01 5.29457748e-01 6.60079718e-01
1.16832323e-01 5.45713425e-01 1.01711059e+00 -3.78570169e-01
1.23185560e-01 -3.44599545e-01 4.81610626e-01 5.23397140e-03
-7.57932425e-01 -1.90360084e-01 -8.20264816e-01 3.61436814e-01
-1.75111926e+00 -1.22747779e+00 2.31350332e-01 1.80658257e+00
4.25663888e-01 5.67941606e-01 8.23472500e-01 -2.26009917e-02
5.55373013e-01 -8.10275525e-02 -7.87259936e-01 -1.13537717e+00
1.81002736e-01 6.50424063e-01 4.50222045e-01 4.24481869e-01
-9.22842085e-01 1.15847313e+00 7.49945593e+00 6.74145162e-01
-8.53823602e-01 9.02486071e-02 3.27897578e-01 -9.41922963e-02
2.22401321e-01 -3.23510081e-01 -4.76615220e-01 1.83230355e-01
1.20447898e+00 -2.89547414e-01 6.52183771e-01 1.05558181e+00
3.45693082e-01 -4.01126891e-01 -6.68991148e-01 9.40101981e-01
-2.16365889e-01 -1.18858051e+00 -2.51488477e-01 1.43144026e-01
6.49703443e-01 3.04868013e-01 1.91447526e-01 9.89554048e-01
1.26015019e+00 -1.47965872e+00 5.83275080e-01 5.62930524e-01
5.62531650e-01 -1.25828063e+00 6.86506271e-01 4.12712336e-01
-7.85154939e-01 -5.83157241e-01 -5.52898467e-01 -6.11674786e-01
-3.35218191e-01 -5.96849084e-01 -8.62607598e-01 1.76613316e-01
9.06410038e-01 5.33114314e-01 -7.50964940e-01 8.90069783e-01
-6.02206230e-01 3.76138031e-01 1.69311002e-01 -3.57788712e-01
9.30305004e-01 -1.46537572e-01 3.18481594e-01 1.04424608e+00
-8.97624642e-02 3.64622563e-01 5.73277473e-04 5.84491611e-01
9.44374725e-02 -1.33457929e-01 -7.76854813e-01 6.44209832e-02
-3.53696160e-02 1.25044966e+00 -7.23815620e-01 -4.60015446e-01
-1.92193165e-01 1.04302549e+00 5.93354583e-01 2.90550351e-01
-8.22019577e-01 -5.18132925e-01 7.37016439e-01 -1.05980290e-02
4.38848287e-01 -8.00712556e-02 2.61626929e-01 -4.68486339e-01
-4.90982682e-01 -1.08828032e+00 4.97459769e-01 -1.12341714e+00
-1.05323637e+00 9.30124104e-01 -1.89506844e-01 -1.02775919e+00
-6.07296944e-01 -8.48499298e-01 -7.92923510e-01 9.39896762e-01
-1.24952924e+00 -5.11315465e-01 -2.77479649e-01 1.05190957e+00
6.84787393e-01 -7.22385824e-01 7.23092973e-01 -7.27120340e-02
-2.54375577e-01 2.59590298e-01 -7.27196410e-02 2.94680268e-01
1.88521042e-01 -1.76899099e+00 5.17207265e-01 6.01316631e-01
5.41309267e-02 1.57514542e-01 8.27337623e-01 -3.06774020e-01
-9.60345447e-01 -5.21033883e-01 -7.18680676e-03 -2.53010035e-01
4.95173603e-01 -1.17093936e-01 -6.81056499e-01 6.38974786e-01
9.44593251e-01 -1.61490262e-01 3.66193712e-01 -1.36911735e-01
1.60556629e-01 -6.58228099e-02 -9.42251980e-01 5.03768623e-01
7.82531977e-01 -3.44907045e-01 -9.04449701e-01 2.39319324e-01
3.52456033e-01 -4.39347744e-01 -3.70065480e-01 -8.71546417e-02
2.39810973e-01 -1.55197048e+00 6.86676443e-01 -1.04233456e+00
6.13078058e-01 -6.06108159e-02 3.53226155e-01 -1.57266748e+00
-5.21930158e-01 -9.47348893e-01 -9.69310626e-02 2.88859695e-01
-4.87971343e-02 -2.59663641e-01 1.20806301e+00 1.78048953e-01
-1.67189822e-01 -5.37332356e-01 -7.52394795e-01 -5.09967148e-01
1.53767139e-01 -1.85960501e-01 2.53711879e-01 4.73715305e-01
1.74518123e-01 2.91476160e-01 -4.89918739e-01 -2.77775139e-01
2.71649003e-01 -4.86291647e-01 7.67609179e-01 -1.02345431e+00
-5.44672668e-01 -4.84224409e-01 -6.92280054e-01 -9.45664525e-01
-3.07662249e-01 -5.10185838e-01 -1.75859123e-01 -1.51404822e+00
-1.28715560e-02 -1.02589475e-02 -7.05035985e-01 4.63957161e-01
-8.59316438e-02 1.28836751e-01 3.84968400e-01 -5.07971086e-02
-9.65777040e-01 5.09818554e-01 1.55381560e+00 2.25088269e-01
-2.26997212e-01 5.86506426e-02 -5.76975405e-01 5.62186480e-01
9.64838088e-01 -5.38984239e-01 -5.56769729e-01 -1.93825111e-01
5.99249780e-01 5.92380166e-01 4.51116353e-01 -1.33329737e+00
4.35717046e-01 -2.06348255e-01 6.13930404e-01 -3.70342553e-01
4.50890735e-02 -6.50323212e-01 -1.48972824e-01 9.24195290e-01
-3.51240665e-01 9.80081916e-01 3.74683887e-01 2.82522798e-01
-1.58488780e-01 -3.02537471e-01 8.97611082e-01 -6.89707935e-01
-1.05088329e+00 1.53992549e-01 -8.96706760e-01 2.17894569e-01
1.31821918e+00 -3.89776021e-01 -2.62091309e-01 -6.41479492e-01
-9.16330159e-01 2.91179061e-01 3.65191475e-02 3.79947960e-01
6.92414522e-01 -1.08496904e+00 -7.02417254e-01 1.74203172e-01
-2.90063798e-01 -4.24091220e-01 2.38333046e-01 3.60585928e-01
-1.11075413e+00 3.81352246e-01 -1.19371462e+00 -3.01700532e-01
-9.98244286e-01 7.97718287e-01 9.71769214e-01 -5.42515039e-01
-7.95241296e-01 9.75661099e-01 7.31130689e-02 -2.71650314e-01
3.31503451e-01 1.16006494e-01 -7.65738726e-01 -1.86415777e-01
6.03399456e-01 4.65338826e-01 -6.16538525e-02 -1.65244102e-01
-1.02335364e-01 1.42983139e-01 -7.33307302e-02 -3.18068892e-01
1.43282640e+00 3.27699393e-01 3.93281370e-01 2.47269467e-01
4.83436525e-01 -5.32705784e-01 -1.86336839e+00 2.05358893e-01
4.57644537e-02 -1.91296250e-01 -6.94023296e-02 -1.16864955e+00
-1.03463769e+00 9.55190718e-01 1.01295328e+00 4.44583505e-01
1.01176870e+00 -4.53787118e-01 1.63061529e-01 4.65257883e-01
2.94526845e-01 -1.36572480e+00 6.58357203e-01 6.87617421e-01
8.35978627e-01 -1.13241255e+00 -1.69821724e-01 8.83368790e-01
-1.11867273e+00 1.17525959e+00 9.27470326e-01 -8.50432277e-01
3.33417892e-01 2.53803611e-01 3.80380362e-01 -6.23860240e-01
-1.07724166e+00 -2.75815755e-01 -8.14056396e-02 9.12909865e-01
1.58083320e-01 -8.89425129e-02 1.21415868e-01 1.41450182e-01
-2.40133479e-01 2.75569469e-01 6.00744426e-01 9.92007315e-01
-7.99030185e-01 -6.98269129e-01 -2.79813617e-01 4.34144400e-02
-4.56191659e-01 -3.08578555e-02 -1.69725031e-01 1.15975535e+00
1.60852790e-01 7.18852103e-01 1.39021799e-01 -1.73541859e-01
6.05531037e-01 -3.18777353e-01 4.34188634e-01 -5.43633640e-01
-1.40554690e+00 -2.18104988e-01 -2.16155484e-01 -8.41499150e-01
-1.93647861e-01 -3.31750333e-01 -1.08747005e+00 -4.38143462e-01
2.96310484e-01 4.17411089e-01 4.31858152e-01 7.98242927e-01
4.13702093e-02 8.99664640e-01 5.16963959e-01 -8.73316348e-01
-5.34026861e-01 -1.11865795e+00 -8.63516569e-01 3.35795581e-01
3.24258655e-01 -5.41353822e-01 -7.13533834e-02 -4.16107595e-01] | [3.737582206726074, 1.4860949516296387] |
173e5804-c59f-4ee4-b353-b66ff57fcd31 | neural-video-portrait-relighting-in-real-time | 2104.00484 | null | https://arxiv.org/abs/2104.00484v1 | https://arxiv.org/pdf/2104.00484v1.pdf | Neural Video Portrait Relighting in Real-time via Consistency Modeling | Video portraits relighting is critical in user-facing human photography, especially for immersive VR/AR experience. Recent advances still fail to recover consistent relit result under dynamic illuminations from monocular RGB stream, suffering from the lack of video consistency supervision. In this paper, we propose a neural approach for real-time, high-quality and coherent video portrait relighting, which jointly models the semantic, temporal and lighting consistency using a new dynamic OLAT dataset. We propose a hybrid structure and lighting disentanglement in an encoder-decoder architecture, which combines a multi-task and adversarial training strategy for semantic-aware consistency modeling. We adopt a temporal modeling scheme via flow-based supervision to encode the conjugated temporal consistency in a cross manner. We also propose a lighting sampling strategy to model the illumination consistency and mutation for natural portrait light manipulation in real-world. Extensive experiments demonstrate the effectiveness of our approach for consistent video portrait light-editing and relighting, even using mobile computing. | ['Lan Xu', 'Jingyi Yu', 'Minye Wu', 'Qixuan Zhang', 'Longwen Zhang'] | 2021-04-01 | null | http://openaccess.thecvf.com//content/ICCV2021/html/Zhang_Neural_Video_Portrait_Relighting_in_Real-Time_via_Consistency_Modeling_ICCV_2021_paper.html | http://openaccess.thecvf.com//content/ICCV2021/papers/Zhang_Neural_Video_Portrait_Relighting_in_Real-Time_via_Consistency_Modeling_ICCV_2021_paper.pdf | iccv-2021-1 | ['single-image-portrait-relighting'] | ['computer-code'] | [ 3.89344990e-01 -5.55883825e-01 -2.69495286e-02 -5.41376829e-01
-5.32256961e-01 -4.90588814e-01 3.75311852e-01 -8.03163052e-01
3.09526408e-03 6.93499625e-01 1.26712859e-01 1.11019410e-01
1.73034862e-01 -5.47907889e-01 -1.23456359e+00 -5.80416322e-01
4.47117269e-01 -1.87773392e-01 -1.47545293e-01 -2.15384632e-01
-1.55879557e-01 5.06357610e-01 -1.61981332e+00 2.36664101e-01
8.59087288e-01 1.02415216e+00 1.64420113e-01 1.09458303e+00
3.16062808e-01 1.12676179e+00 -4.47919637e-01 -7.98030436e-01
5.09279490e-01 -4.50767130e-01 -3.66740912e-01 4.12996829e-01
1.11043227e+00 -8.92537057e-01 -7.71670580e-01 9.76965368e-01
5.08709610e-01 8.58533606e-02 1.24833263e-01 -1.64908218e+00
-9.99344349e-01 -9.18350220e-02 -6.69838846e-01 -1.94147334e-01
7.46767700e-01 4.62326944e-01 6.49866402e-01 -5.44343531e-01
1.12270617e+00 1.23162007e+00 8.94800007e-01 6.18825972e-01
-1.19712448e+00 -8.19497108e-01 4.04459178e-01 3.59497637e-01
-1.43767953e+00 -8.11992347e-01 1.04376531e+00 4.35506292e-02
6.89642370e-01 5.72417676e-01 1.06712830e+00 1.44278467e+00
4.17411059e-01 7.55314231e-01 1.26200783e+00 3.13745067e-02
1.74997617e-02 -1.51176509e-02 -6.75946176e-01 7.63075352e-01
-1.53345451e-01 2.17437714e-01 -8.63504469e-01 3.55504788e-02
1.10707760e+00 2.51829475e-01 -5.63262224e-01 -3.20081621e-01
-1.05513906e+00 2.31132388e-01 4.39489305e-01 -3.05549856e-02
-4.11449485e-02 6.46786273e-01 3.98395509e-01 5.41336238e-01
4.87865329e-01 -4.82331216e-02 -2.59266585e-01 5.43606617e-02
-1.06645906e+00 2.69204676e-01 4.52170789e-01 1.37385237e+00
3.67035478e-01 4.61600661e-01 -1.19642496e-01 6.66425169e-01
2.98419029e-01 8.55399430e-01 4.97878008e-02 -1.14089882e+00
4.60714191e-01 7.36183161e-03 5.51359691e-02 -1.13361919e+00
3.05143725e-02 -3.15992497e-02 -1.09825754e+00 3.54227304e-01
-1.51502624e-01 1.63809851e-01 -7.99734533e-01 1.87348330e+00
4.71935213e-01 8.42332721e-01 2.73392745e-03 1.14855087e+00
6.98083699e-01 7.80362964e-01 -1.42360464e-01 -2.99010575e-01
1.08755541e+00 -1.00983894e+00 -1.26374900e+00 5.87824658e-02
3.20697390e-02 -7.38865435e-01 1.14645147e+00 5.15829504e-01
-1.33033335e+00 -5.56648374e-01 -1.17025912e+00 -7.44930506e-01
1.59761403e-02 1.76925529e-02 6.19781017e-01 5.43720901e-01
-1.08964241e+00 4.13911998e-01 -7.09716499e-01 -6.34848997e-02
4.85580593e-01 8.92218500e-02 -4.94419515e-01 -4.77040023e-01
-1.21246910e+00 6.16470098e-01 -1.05900176e-01 4.90736544e-01
-1.11691809e+00 -8.17795157e-01 -1.11997640e+00 -2.12645426e-01
3.90718520e-01 -1.08989406e+00 8.59668195e-01 -1.40227270e+00
-1.95908809e+00 8.06517720e-01 -2.40389213e-01 -1.73849583e-01
8.88355374e-01 -2.69432545e-01 -5.17323315e-01 1.07225537e-01
-1.76177621e-01 7.48674691e-01 1.31458414e+00 -1.59469187e+00
-2.79934317e-01 -1.65123194e-01 2.85712332e-02 5.65243959e-01
-2.17403308e-01 -1.22426689e-01 -6.72797382e-01 -8.59030128e-01
-1.47063836e-01 -8.37751746e-01 9.63280350e-02 8.57285619e-01
-3.63460600e-01 5.61433554e-01 1.03740871e+00 -1.06433713e+00
9.01228547e-01 -1.96977043e+00 4.75091159e-01 -1.33704960e-01
9.18786526e-02 7.24683031e-02 -1.15592010e-01 1.48758799e-01
2.38647461e-02 -2.15498656e-01 -2.48829424e-01 -9.84883249e-01
-3.11858971e-02 3.39190781e-01 -6.32567763e-01 6.04062200e-01
1.15304418e-01 1.08497417e+00 -9.00041342e-01 -5.13161361e-01
5.91736495e-01 1.08161700e+00 -6.02648079e-01 4.35409248e-01
-3.05869341e-01 5.52061141e-01 6.95127100e-02 9.81741309e-01
1.04859805e+00 2.00019926e-01 3.74291837e-02 -6.37187481e-01
1.61132842e-01 -2.65080422e-01 -1.23539853e+00 2.51704144e+00
-8.71658087e-01 9.42694843e-01 2.76116878e-01 -1.25306219e-01
6.30518436e-01 7.09927380e-02 2.79506266e-01 -9.72252488e-01
5.05891927e-02 -9.59588364e-02 -7.40120590e-01 -6.24690831e-01
9.93856490e-01 -1.14094496e-01 9.95817930e-02 1.67437524e-01
-2.13779479e-01 -5.22359729e-01 -3.68213236e-01 9.15351287e-02
7.71255970e-01 8.01398933e-01 -1.00211337e-01 3.68590295e-01
3.31445962e-01 -5.19995391e-01 6.51111066e-01 1.48394808e-01
-1.20586909e-01 1.01986444e+00 2.14125052e-01 -5.24552822e-01
-1.29785526e+00 -1.05917430e+00 -4.18635905e-02 6.35837257e-01
5.60150206e-01 -2.15887710e-01 -5.95101416e-01 -2.30616316e-01
-1.42713219e-01 6.50664330e-01 -5.30498981e-01 -1.09920427e-01
-5.73362529e-01 -3.23664755e-01 5.63542485e-01 2.17414886e-01
8.93076956e-01 -6.00504994e-01 -4.57210988e-01 -1.94187015e-01
-3.54039639e-01 -1.49159896e+00 -9.43208158e-01 -2.98394650e-01
-5.72765946e-01 -7.13716626e-01 -6.35487616e-01 -6.02477431e-01
4.30493057e-01 5.89705288e-01 9.33002591e-01 1.20225802e-01
-5.03530681e-01 5.78584909e-01 -9.65263993e-02 1.84908658e-01
-4.05083120e-01 -4.63193506e-01 -8.44129473e-02 3.79082084e-01
-5.69888592e-01 -7.62384057e-01 -7.90612578e-01 3.27814102e-01
-1.28973746e+00 5.61145663e-01 1.19949326e-01 8.22894692e-01
6.57570302e-01 -8.10292643e-03 2.24339932e-01 -5.96419990e-01
2.35388175e-01 -2.35334173e-01 -4.66559827e-01 5.86912930e-01
-3.00432205e-01 -3.43854159e-01 9.02509332e-01 -5.86764991e-01
-1.39009547e+00 4.34008762e-02 -1.72470629e-01 -1.23125243e+00
7.72098601e-02 -3.71427894e-01 -6.37101531e-01 -4.57018107e-01
1.15633510e-01 4.67556030e-01 -9.82329845e-02 1.43492624e-01
6.97054982e-01 2.90133059e-01 8.50806594e-01 -4.45798278e-01
1.03661990e+00 8.32222164e-01 8.65286812e-02 -5.67429066e-01
-3.52474004e-01 1.89259440e-01 -3.77634734e-01 -5.93629420e-01
8.50040495e-01 -1.32816708e+00 -9.69638884e-01 7.18464077e-01
-1.15528429e+00 -4.31149513e-01 -3.15615147e-01 4.89285178e-02
-8.24327767e-01 5.70274651e-01 -6.24537885e-01 -5.47061741e-01
-3.62180680e-01 -1.24206281e+00 1.58485246e+00 2.05882400e-01
1.75242081e-01 -8.76173496e-01 2.65514366e-02 5.33667028e-01
3.89446288e-01 5.42910695e-01 4.10627186e-01 5.48112869e-01
-1.13269055e+00 2.75307186e-02 -3.13076824e-01 3.51542383e-01
1.32661670e-01 8.54034871e-02 -1.14671946e+00 -2.69743681e-01
6.00805972e-03 -4.82215941e-01 5.62861562e-01 2.66801536e-01
1.33365858e+00 -3.72342497e-01 1.21410854e-01 1.48898888e+00
1.68295336e+00 1.12109613e-02 1.00617969e+00 7.22459257e-02
1.38322473e+00 2.65364528e-01 6.10171974e-01 4.39006865e-01
6.23661101e-01 8.98123860e-01 5.97225666e-01 -1.67949438e-01
-4.79114860e-01 -6.35671139e-01 5.36484599e-01 6.51967824e-01
8.27712491e-02 -5.51019609e-01 -1.29993379e-01 1.06888972e-01
-1.65334117e+00 -1.10741353e+00 5.70053719e-02 2.00956082e+00
9.05003786e-01 -2.89475352e-01 -3.24251354e-01 -4.87089604e-02
5.77636719e-01 5.65051258e-01 -8.49119365e-01 -4.61591691e-01
-4.23148155e-01 -1.96659878e-01 6.07923627e-01 6.86766565e-01
-7.50132978e-01 9.14455533e-01 5.65847111e+00 8.73586237e-01
-1.26024985e+00 2.59880215e-01 8.39306593e-01 -4.73390460e-01
-9.08418179e-01 -1.24197595e-01 -3.82673353e-01 5.55825531e-01
5.16361415e-01 2.31138259e-01 9.33985472e-01 4.96248096e-01
3.22896957e-01 -1.41205803e-01 -1.01455724e+00 1.49887848e+00
7.68581331e-01 -1.35882521e+00 1.16719462e-01 -1.40470490e-01
9.52011347e-01 -6.21997595e-01 3.75479549e-01 -2.20621392e-01
1.95712727e-02 -8.86719048e-01 1.12809420e+00 8.05393934e-01
1.40270102e+00 -6.72020972e-01 7.29539841e-02 -1.94611564e-01
-1.20061874e+00 7.17874020e-02 -8.61793309e-02 3.29661638e-01
6.49116337e-01 2.43937090e-01 -2.31431141e-01 8.33185911e-01
6.01809144e-01 1.22620690e+00 -4.33279186e-01 5.22045612e-01
-9.99468341e-02 -9.37385559e-02 -1.51913524e-01 3.92157614e-01
-2.38681808e-01 -3.00574541e-01 5.55445790e-01 8.38578820e-01
2.67443627e-01 4.27552834e-02 -1.84267506e-01 8.91101837e-01
-2.91287303e-01 -3.08160454e-01 -8.03891182e-01 3.31899732e-01
4.18971956e-01 1.13683379e+00 -4.49544132e-01 -1.49801895e-01
-1.94279239e-01 1.87011743e+00 1.51025832e-01 4.44393039e-01
-1.36831868e+00 1.18448809e-02 8.12178493e-01 -4.94532622e-02
5.86567521e-02 -3.66216600e-01 -2.64648825e-01 -1.60080409e+00
3.65808040e-01 -8.76872718e-01 -2.11999565e-01 -1.55442870e+00
-1.09686446e+00 6.27844751e-01 -7.23672509e-02 -1.46376944e+00
5.42121753e-03 -2.32766092e-01 -5.39693296e-01 4.41933632e-01
-1.90601420e+00 -1.69468379e+00 -7.04915226e-01 9.92524326e-01
8.43454719e-01 8.51909295e-02 4.36116964e-01 5.87747693e-01
-7.51632750e-01 7.44934142e-01 8.74779448e-02 -4.16026592e-01
1.03336513e+00 -9.25357640e-01 4.46269363e-01 1.10225534e+00
-1.79689261e-04 2.34551698e-01 7.14797974e-01 -6.58351421e-01
-2.16429329e+00 -1.40991759e+00 3.34868997e-01 -3.32989663e-01
2.66179293e-01 -5.51084161e-01 -6.96616590e-01 7.04411626e-01
5.09169102e-01 2.04822212e-01 4.27016795e-01 -5.48272312e-01
-5.59291184e-01 -5.59994340e-01 -1.32460439e+00 9.29599822e-01
1.44308114e+00 -7.52902329e-01 -1.82882585e-02 3.10839862e-01
1.07226467e+00 -8.49005342e-01 -9.03742969e-01 2.09642686e-02
8.59998882e-01 -1.08559060e+00 1.19789135e+00 1.20505039e-03
7.39914417e-01 -4.33555990e-01 -3.17838222e-01 -1.17788124e+00
1.54364288e-01 -1.17973566e+00 -1.01735920e-01 1.38149154e+00
-4.45222616e-01 -3.56412560e-01 4.55043375e-01 8.34840894e-01
-1.69823915e-01 -5.81444323e-01 -9.98006165e-01 -5.38097560e-01
-4.96734649e-01 -5.08119762e-01 7.69622445e-01 9.94886756e-01
-6.68739736e-01 -7.56347477e-02 -1.31448591e+00 2.34270588e-01
7.87838459e-01 2.29616258e-02 9.67040300e-01 -3.81706476e-01
-3.04530412e-01 2.70896964e-02 -4.07404572e-01 -9.65327442e-01
3.73225302e-01 -6.33388758e-01 1.72951259e-02 -1.15999877e+00
5.30043319e-02 -3.47201794e-01 2.77780414e-01 1.22765981e-01
-5.28624840e-02 5.00158548e-01 4.82294291e-01 1.41965985e-01
-6.55022979e-01 9.22820270e-01 1.47455025e+00 -1.30061373e-01
-7.52815455e-02 -6.40868425e-01 -3.56001794e-01 4.55230653e-01
2.57103115e-01 -2.48986945e-01 -8.48316371e-01 -9.57446098e-01
4.95615989e-01 3.44817460e-01 9.03701127e-01 -9.64845240e-01
2.70566851e-01 -3.53931040e-01 6.13452435e-01 -2.94246197e-01
8.55401397e-01 -1.21494091e+00 8.46915722e-01 2.07160264e-01
-3.61354679e-01 3.35104853e-01 1.48272097e-01 9.02618289e-01
-5.11286892e-02 4.10675555e-01 7.43999362e-01 1.82499126e-01
-7.07127988e-01 7.42114007e-01 2.37418517e-01 -9.66996625e-02
1.18689704e+00 -3.73985291e-01 -2.22667396e-01 -5.61292887e-01
-3.05061698e-01 1.78761438e-01 9.72051203e-01 5.37023187e-01
1.08533442e+00 -1.70678830e+00 -5.57688773e-01 6.29270911e-01
-3.53997722e-02 -6.84222132e-02 9.53024685e-01 4.62570548e-01
-9.15934086e-01 -2.26133555e-01 -4.56765622e-01 -5.96756995e-01
-1.33854854e+00 5.23938179e-01 4.00483429e-01 6.66044205e-02
-8.93854260e-01 6.95066810e-01 2.61268392e-02 -1.01494767e-01
2.83219486e-01 -4.00909424e-01 4.76414025e-01 -2.85037935e-01
5.24258912e-01 2.02550143e-01 -1.88752815e-01 -6.39171600e-01
-1.26520768e-01 9.19600248e-01 8.60547647e-02 -2.59758860e-01
1.20245337e+00 -7.47061789e-01 9.13532078e-02 4.33739036e-01
1.39737868e+00 -1.18408762e-01 -1.77292848e+00 -2.80357543e-02
-1.04377282e+00 -1.20351315e+00 1.34507731e-01 -6.17336512e-01
-1.28933406e+00 6.65767789e-01 6.83656216e-01 -4.09180373e-01
1.52121675e+00 -7.35323787e-01 1.31381178e+00 1.45067662e-01
4.18307453e-01 -8.99367929e-01 6.21457957e-02 -1.56136185e-01
1.06147468e+00 -1.17776263e+00 1.14177890e-01 -4.11073267e-01
-9.71823275e-01 1.05941427e+00 5.38419783e-01 -2.25881547e-01
3.27408195e-01 4.42892164e-01 -2.10155919e-01 1.58694655e-01
-9.46548522e-01 3.30447525e-01 2.91304905e-02 6.54203892e-01
-8.05742070e-02 -1.82727396e-01 3.11842769e-01 1.26095135e-02
1.45956976e-02 3.02811444e-01 5.83470643e-01 7.06561565e-01
3.82374287e-01 -7.41944194e-01 -4.03311223e-01 -2.08854526e-01
-4.38915603e-02 -1.91465840e-01 2.10666526e-02 5.98984480e-01
3.08933020e-01 6.64864957e-01 5.19339144e-02 -4.17222440e-01
2.56376266e-01 -3.10859919e-01 9.06914771e-01 9.50278342e-02
-5.38218796e-01 3.65513489e-02 4.15734714e-03 -9.79397893e-01
-6.11133337e-01 -4.66499329e-01 -6.57413244e-01 -8.04453373e-01
8.43369365e-02 -5.19243479e-01 7.31721282e-01 6.18650079e-01
3.92803758e-01 6.84366524e-01 1.02586424e+00 -1.14948094e+00
-1.15712704e-02 -2.51428097e-01 -5.82676947e-01 5.26913643e-01
6.88521385e-01 -4.13767636e-01 -3.00880224e-01 4.82717365e-01] | [11.408105850219727, -1.1214675903320312] |
d8a7486a-da7d-45d1-9a5e-bf41f8d5cdf9 | extractive-text-summarization-using-neural | 1802.10137 | null | http://arxiv.org/abs/1802.10137v1 | http://arxiv.org/pdf/1802.10137v1.pdf | Extractive Text Summarization using Neural Networks | Text Summarization has been an extensively studied problem. Traditional
approaches to text summarization rely heavily on feature engineering. In
contrast to this, we propose a fully data-driven approach using feedforward
neural networks for single document summarization. We train and evaluate the
model on standard DUC 2002 dataset which shows results comparable to the state
of the art models. The proposed model is scalable and is able to produce the
summary of arbitrarily sized documents by breaking the original document into
fixed sized parts and then feeding it recursively to the network. | ['Aakash Sinha', 'Akshay Gahlot', 'Abhishek Yadav'] | 2018-02-27 | null | null | null | null | ['extractive-document-summarization'] | ['natural-language-processing'] | [ 4.81544226e-01 3.68755937e-01 -1.44159883e-01 -4.84829158e-01
-8.20736527e-01 -4.59553659e-01 6.27207816e-01 6.06271327e-01
-5.37068725e-01 1.01461339e+00 9.39648330e-01 -1.21346608e-01
7.22077163e-03 -5.46225905e-01 -6.31843328e-01 -2.52483368e-01
7.43504763e-02 5.50308228e-01 1.74249545e-01 -2.23165810e-01
9.60221171e-01 2.32496291e-01 -1.42698920e+00 4.50322360e-01
1.18923652e+00 5.29861033e-01 9.71776694e-02 1.27313507e+00
-3.73711288e-01 8.20120573e-01 -1.27070820e+00 -1.90896913e-01
-1.05746999e-01 -9.00811851e-01 -1.22055960e+00 7.60166124e-02
8.47110629e-01 -5.84174752e-01 -3.82883847e-01 8.65974009e-01
6.25633836e-01 4.55103248e-01 8.31584513e-01 -6.08162999e-01
-7.86543906e-01 1.12313461e+00 -3.76570851e-01 2.65403777e-01
6.04636967e-01 -5.74400604e-01 1.14037311e+00 -6.18388772e-01
7.22856224e-01 1.05279887e+00 6.57188058e-01 6.05380177e-01
-9.81890619e-01 -1.32278353e-01 1.86663628e-01 -7.29207993e-02
-5.14906168e-01 -5.58028460e-01 6.92145109e-01 -3.75495367e-02
1.49090981e+00 4.99497414e-01 6.78344607e-01 6.83500171e-01
4.69864905e-01 1.20268965e+00 2.44471580e-01 -5.95492899e-01
3.04668278e-01 -2.15486616e-01 8.08368266e-01 5.80060601e-01
7.80249596e-01 -6.64799809e-01 -5.15048027e-01 -1.19591072e-01
2.44439811e-01 -1.36269227e-01 -2.37427413e-01 1.62502453e-01
-1.16388690e+00 8.85396004e-01 2.93255687e-01 3.32646161e-01
-6.39916182e-01 2.77125567e-01 9.81384397e-01 3.23581696e-01
7.13059843e-01 7.87914991e-01 -2.84145296e-01 -1.61141098e-01
-1.70789421e+00 6.14292681e-01 1.25563800e+00 9.67423320e-01
1.69340432e-01 3.30649823e-01 -5.97709656e-01 8.46917808e-01
-1.42332897e-01 1.84638396e-01 9.66047704e-01 -9.25681949e-01
8.27265084e-01 6.49983346e-01 1.26787320e-01 -8.35190713e-01
-5.34103453e-01 -1.35954127e-01 -1.15686810e+00 -2.89658070e-01
-1.56979710e-01 -5.31152487e-01 -9.85028803e-01 1.00426936e+00
-2.06665814e-01 -4.58178461e-01 4.24499720e-01 3.32090229e-01
1.37470698e+00 1.04304564e+00 -3.84769559e-01 -5.05651832e-01
7.80585647e-01 -1.41306806e+00 -9.47212338e-01 1.42077068e-02
4.04158443e-01 -5.77110350e-01 6.10277176e-01 5.14428675e-01
-1.55516112e+00 -4.11699235e-01 -1.34655559e+00 -4.32041109e-01
-2.13813752e-01 4.61076885e-01 4.68923718e-01 4.68494594e-01
-1.32575488e+00 9.95378375e-01 -6.29639089e-01 -7.34279573e-01
4.96861815e-01 4.95713174e-01 -3.26897353e-01 3.00367802e-01
-8.08358133e-01 8.51521432e-01 1.00614369e+00 4.33721282e-02
-3.09355795e-01 -2.46279612e-01 -8.11580122e-01 4.42643523e-01
1.03105672e-01 -1.20190024e+00 1.83564138e+00 -8.41124654e-01
-2.07165575e+00 9.52580273e-02 -3.22964013e-01 -9.86179769e-01
2.48470277e-01 -6.62667811e-01 -1.28891235e-02 3.63325566e-01
-8.96898955e-02 6.02432966e-01 7.37979650e-01 -8.95527005e-01
-7.89007783e-01 -4.52094302e-02 -1.65726900e-01 3.22950065e-01
-4.94413972e-01 1.59866989e-01 -2.16727536e-02 -7.41204977e-01
-1.77665412e-01 -4.45555896e-01 -1.55320317e-01 -8.53482187e-01
-9.12003279e-01 -4.28148776e-01 7.96652853e-01 -8.82669210e-01
1.58629906e+00 -1.29037118e+00 3.10800195e-01 -3.69535863e-01
8.45898464e-02 5.37304044e-01 -2.02761278e-01 9.35212255e-01
8.97525698e-02 1.65175289e-01 -2.45169461e-01 -5.27099609e-01
5.80378808e-03 -1.30376399e-01 -4.77974743e-01 1.28167555e-01
1.79982439e-01 9.80602562e-01 -7.15209305e-01 -4.79577422e-01
-3.60148624e-02 -2.97101699e-02 -6.24575853e-01 2.91679591e-01
-3.23401093e-01 -1.89538807e-01 -4.62500840e-01 3.22321057e-01
3.78377318e-01 7.08309337e-02 -5.31970523e-03 2.14179605e-03
-2.59173334e-01 4.37831521e-01 -9.39225554e-01 1.69103932e+00
-1.43240988e-01 1.01662254e+00 -2.85337299e-01 -1.15759730e+00
9.30436969e-01 3.12058777e-01 1.15822315e-01 -1.91106826e-01
3.83801013e-01 2.04597786e-01 -3.19923043e-01 -4.83894348e-01
1.48625124e+00 1.50486559e-01 -2.41973892e-01 8.76156449e-01
3.82642776e-01 -4.80451941e-01 8.46577525e-01 5.11171758e-01
1.13761437e+00 -2.69132461e-02 7.49771059e-01 -2.72997588e-01
6.38258219e-01 2.98213661e-01 -1.86657906e-02 9.79985535e-01
2.65928477e-01 8.62764597e-01 6.91946030e-01 -4.02880609e-01
-1.21860826e+00 -4.95813400e-01 4.10774827e-01 1.11231434e+00
-1.53363124e-01 -6.64625466e-01 -1.26265848e+00 -7.54959404e-01
-1.97227925e-01 1.14516830e+00 -7.17583418e-01 -1.38100967e-01
-6.86997652e-01 -4.86229122e-01 6.78643823e-01 6.47853136e-01
6.08724654e-01 -1.43400943e+00 -7.75916636e-01 4.93618429e-01
-1.59257382e-01 -5.12155950e-01 -7.38601327e-01 3.35826278e-01
-1.28795743e+00 -7.21246600e-01 -8.10689211e-01 -9.11029100e-01
6.14444017e-01 3.00897270e-01 1.11842895e+00 -9.41112712e-02
-3.02910805e-03 7.96519592e-02 -3.80276740e-01 -8.32570910e-01
-7.67949939e-01 7.54001200e-01 -9.31898281e-02 -4.16905403e-01
8.55564252e-02 -2.74298608e-01 -4.31649983e-01 -4.70600575e-01
-1.21648955e+00 1.25196770e-01 7.59224892e-01 8.96475852e-01
2.94070303e-01 -2.84263361e-02 1.07244289e+00 -1.28385079e+00
1.60986459e+00 -1.93783462e-01 -2.21192747e-01 3.79972100e-01
-4.68837112e-01 2.83127755e-01 1.01679826e+00 -2.29982674e-01
-1.12150598e+00 -5.48856631e-02 -8.48377869e-02 2.57710665e-01
-1.82906151e-01 7.46245444e-01 2.50674248e-01 4.43748832e-01
6.36474252e-01 4.47769433e-01 -8.29597265e-02 -5.10855079e-01
4.54195917e-01 9.89043176e-01 6.76444411e-01 -1.00840621e-01
4.39664364e-01 1.01920351e-01 -4.15250957e-01 -1.16023672e+00
-7.54609346e-01 -4.25949097e-01 -8.24870706e-01 8.22864324e-02
4.83429968e-01 -3.89850289e-01 -2.17190504e-01 4.04196739e-01
-1.50273454e+00 3.13545540e-02 -6.93688810e-01 1.94305047e-01
-6.65334582e-01 5.89207649e-01 -5.84917486e-01 -4.29774076e-01
-1.35080743e+00 -5.37137151e-01 1.02256417e+00 6.50104165e-01
-6.49022043e-01 -9.48807418e-01 5.18558443e-01 -4.32813093e-02
5.73373497e-01 8.49729329e-02 8.16560328e-01 -1.19110632e+00
1.07782319e-01 -7.53869176e-01 -1.47755072e-01 5.81421018e-01
2.19882488e-01 2.52304912e-01 -5.56977212e-01 -2.06794813e-01
-6.36369409e-03 -3.44449997e-01 1.34351814e+00 8.05620849e-01
9.10881639e-01 -6.44072592e-01 -1.78240925e-01 1.55699164e-01
1.13752449e+00 7.81051489e-03 4.60034579e-01 3.89073968e-01
4.15832102e-01 5.42272687e-01 3.66976082e-01 4.22285289e-01
3.02901804e-01 -1.06606886e-01 1.08893877e-02 1.02270335e-01
-7.29491040e-02 -2.02366427e-01 4.12946403e-01 1.17260385e+00
-8.48818719e-02 -8.93967569e-01 -4.29452896e-01 5.88369429e-01
-2.11002707e+00 -1.20254624e+00 -2.75185764e-01 1.68177915e+00
7.69655526e-01 1.70689508e-01 1.93324864e-01 7.88699090e-02
6.43764913e-01 4.79402244e-01 -4.56161559e-01 -1.26208878e+00
3.95032996e-03 8.31060261e-02 2.89468229e-01 3.27452719e-01
-1.08284175e+00 9.33754265e-01 7.88710451e+00 4.78393555e-01
-1.03734791e+00 -4.10008281e-01 3.11027110e-01 -3.52356851e-01
-1.02648571e-01 -3.49760890e-01 -7.95816958e-01 1.17128968e-01
1.33128858e+00 -8.95120502e-01 -3.07805426e-02 6.75072789e-01
4.25170392e-01 -5.08941710e-01 -1.22394645e+00 5.04410744e-01
6.12749040e-01 -1.72554290e+00 6.20409727e-01 -3.48041981e-01
1.06189609e+00 -1.51972234e-01 -4.54685032e-01 3.44669402e-01
2.68463880e-01 -9.55221117e-01 5.37205696e-01 6.39832258e-01
5.14265656e-01 -8.98109436e-01 1.03682339e+00 4.78539944e-01
-5.47482550e-01 2.40703914e-02 -6.71676040e-01 -1.08760722e-01
2.51913905e-01 3.29414368e-01 -9.10637259e-01 8.96601975e-01
2.67564923e-01 8.24460268e-01 -7.60788798e-01 1.21172702e+00
-1.10901780e-02 7.13351607e-01 -1.20476738e-01 -5.45502365e-01
3.46839249e-01 -1.01779010e-02 6.73301160e-01 1.59059131e+00
2.61328012e-01 -9.47447568e-02 3.59814502e-02 4.25812751e-01
-4.05057609e-01 4.39799786e-01 -6.45738721e-01 -2.40068749e-01
-8.03701673e-03 1.03937256e+00 -7.09355116e-01 -8.48353326e-01
-1.46372661e-01 1.18553305e+00 4.51534063e-01 2.88556993e-01
-3.16249043e-01 -1.17126572e+00 -2.80495405e-01 -2.97720551e-01
6.53237700e-01 -8.20344761e-02 -4.45722282e-01 -1.23411751e+00
3.47047262e-02 -8.51346135e-01 2.95116693e-01 -5.04567981e-01
-8.52586448e-01 8.88699770e-01 2.10686892e-01 -9.60337102e-01
-6.26822233e-01 -1.05944604e-01 -1.12557912e+00 6.87007546e-01
-1.25558293e+00 -8.57445657e-01 -1.17032237e-01 -3.09341270e-02
1.18522155e+00 -3.69202554e-01 1.03843200e+00 -4.19918418e-01
-6.56682670e-01 3.04849237e-01 7.06740081e-01 -1.66551933e-01
6.43644035e-01 -1.56283975e+00 8.79873216e-01 9.97664809e-01
6.07933216e-02 6.67819083e-01 1.06132054e+00 -7.47658968e-01
-1.18056440e+00 -1.06892729e+00 1.31719673e+00 -1.40231594e-01
4.14044321e-01 3.02384458e-02 -9.94888067e-01 6.81178331e-01
1.03917778e+00 -7.45828032e-01 5.68596601e-01 -1.44975156e-01
1.31213009e-01 1.28564298e-01 -9.31963980e-01 4.66404617e-01
5.94687045e-01 2.10014045e-01 -1.40410972e+00 3.33951503e-01
8.00407529e-01 -3.45845014e-01 -4.91634369e-01 -5.72937494e-03
5.56228399e-01 -8.39466035e-01 3.30509663e-01 -8.86743963e-01
9.73289728e-01 7.10817203e-02 2.17592716e-01 -1.82160497e+00
-2.08734170e-01 -7.40041733e-01 -4.58966404e-01 1.23958588e+00
4.34511989e-01 -4.14087445e-01 8.45700860e-01 3.08323562e-01
-3.82193357e-01 -6.63969040e-01 -5.53150177e-01 -4.94947881e-01
3.05287033e-01 2.88937539e-01 4.13398117e-01 2.16055363e-01
3.13520938e-01 8.22512269e-01 -2.96298563e-01 -4.57001090e-01
3.18868488e-01 3.51185948e-01 8.40512693e-01 -1.35731840e+00
8.75764713e-02 -8.34945261e-01 6.75833747e-02 -1.19249344e+00
2.68450558e-01 -8.44777346e-01 3.76185775e-01 -2.46907449e+00
3.55491787e-01 5.85664272e-01 7.73337036e-02 1.41171008e-01
-2.50109345e-01 4.76705562e-03 3.43800858e-02 8.83115977e-02
-9.57459807e-01 6.88311160e-01 9.64818954e-01 -3.96086484e-01
-5.87010264e-01 2.43624598e-01 -1.04899073e+00 6.23729050e-01
1.06558943e+00 -4.01627868e-01 -4.03384745e-01 -4.30946320e-01
6.27439916e-02 1.21875964e-01 -3.30840260e-01 -9.59102511e-01
6.52297258e-01 9.69256982e-02 4.13630694e-01 -1.36453366e+00
-2.99951673e-01 -2.90308893e-01 -4.07668442e-01 4.60203618e-01
-8.70702028e-01 2.06431568e-01 3.43416482e-01 6.60889387e-01
-4.59511787e-01 -9.14366305e-01 4.05890465e-01 -2.01784000e-01
-1.03173956e-01 -2.29188502e-01 -7.57438600e-01 -6.03997009e-03
6.73854709e-01 -3.43489587e-01 -4.08269405e-01 -5.75931013e-01
-3.43706340e-01 2.94815183e-01 2.43111819e-01 3.65753174e-01
6.31638885e-01 -7.52932549e-01 -1.07542586e+00 -6.24437965e-02
-3.07628512e-01 1.84027463e-01 6.45136684e-02 3.05928797e-01
-1.00143933e+00 1.01968634e+00 -1.68848410e-01 -1.95633948e-01
-1.34500909e+00 3.01287383e-01 1.19317092e-01 -5.55134535e-01
-8.12593758e-01 4.61060584e-01 -4.38765019e-01 -2.76580215e-01
2.64402479e-01 -5.59948564e-01 -7.05345869e-01 3.72135460e-01
9.07271981e-01 6.80520713e-01 3.38191181e-01 -3.09613079e-01
4.83793952e-02 1.79464117e-01 -6.85239077e-01 -1.71882771e-02
1.69711125e+00 -2.16825716e-02 -2.92825043e-01 5.58907568e-01
1.04830909e+00 -8.84672701e-02 -7.42497504e-01 -5.22568785e-02
2.36406252e-01 8.89406577e-02 -5.59254214e-02 -7.14737356e-01
-5.57532072e-01 6.58429146e-01 -1.99259102e-01 6.16059840e-01
9.00766075e-01 -2.97869861e-01 9.93085980e-01 1.12967002e+00
-2.70145774e-01 -1.46737790e+00 1.24369033e-01 8.26231897e-01
1.33232868e+00 -8.85718107e-01 6.43979430e-01 1.26693666e-01
-6.77501082e-01 1.66793966e+00 3.75859588e-01 -7.31950104e-01
-5.20014996e-03 6.32459223e-02 -2.34522611e-01 -1.61248833e-01
-1.01330423e+00 4.29307878e-01 3.96599263e-01 3.56331319e-01
7.35765696e-01 -2.25662977e-01 -7.40714014e-01 6.26473665e-01
-6.11966074e-01 1.80110782e-01 1.18007362e+00 1.13834894e+00
-9.11985099e-01 -9.00450647e-01 -2.82255560e-01 1.04304779e+00
-5.77524126e-01 -1.06256083e-01 -8.56148660e-01 5.50281942e-01
-8.40829492e-01 9.63769972e-01 1.26614384e-02 -1.17218746e-02
7.32742906e-01 1.82068571e-01 4.10565525e-01 -9.73846376e-01
-1.10741973e+00 -6.44447505e-02 3.60309243e-01 -1.18592381e-03
-4.26396489e-01 -7.96889603e-01 -1.15661895e+00 -3.02285463e-01
-5.30761063e-01 3.69875014e-01 7.41198778e-01 7.47154236e-01
4.14180398e-01 9.10985887e-01 5.57315767e-01 -1.27547181e+00
-8.39925468e-01 -1.52290273e+00 -4.46141839e-01 -1.62537415e-02
5.60964644e-01 1.29998058e-01 -2.13354111e-01 2.99319953e-01] | [12.520356178283691, 9.506135940551758] |
646fd3d1-5f7b-4b79-9a6e-621c399ee198 | hierarchical-neural-representation-of-dreamed | 1611.0952 | null | http://arxiv.org/abs/1611.09520v2 | http://arxiv.org/pdf/1611.09520v2.pdf | Hierarchical Neural Representation of Dreamed Objects Revealed by Brain Decoding with Deep Neural Network Features | Dreaming is generally thought to be generated by spontaneous brain activity
during sleep with patterns common to waking experience. This view is supported
by a recent study demonstrating that dreamed objects can be predicted from
brain activity during sleep using statistical decoders trained with
stimulus-induced brain activity. However, it remains unclear whether and how
visual image features associated with dreamed objects are represented in the
brain. In this study, we used a deep neural network (DNN) model for object
recognition as a proxy for hierarchical visual feature representation, and DNN
features for dreamed objects were analyzed with brain decoding of fMRI data
collected during dreaming. The decoders were first trained with
stimulus-induced brain activity labeled with the feature values of the stimulus
image from multiple DNN layers. The decoders were then used to decode DNN
features from the dream fMRI data, and the decoded features were compared with
the averaged features of each object category calculated from a large-scale
image database. We found that the feature values decoded from the dream fMRI
data positively correlated with those associated with dreamed object categories
at mid- to high-level DNN layers. Using the decoded features, the dreamed
object category could be identified at above-chance levels by matching them to
the averaged features for candidate categories. The results suggest that
dreaming recruits hierarchical visual feature representations associated with
objects, which may support phenomenal aspects of dream experience. | [] | 2017-01-23 | null | null | null | null | ['brain-decoding', 'brain-decoding'] | ['medical', 'miscellaneous'] | [-6.40551895e-02 -9.66672450e-02 1.49044335e-01 -6.85401440e-01
-3.24682623e-01 -2.44951501e-01 6.24729514e-01 -3.75011981e-01
-5.99402547e-01 5.45605958e-01 9.35915470e-01 3.51113677e-01
-1.18087389e-01 -5.37720382e-01 -4.78368074e-01 -7.58033812e-01
-1.85267314e-01 4.28041071e-01 -3.54605615e-01 -2.30206195e-02
3.30894083e-01 5.02955675e-01 -1.94717276e+00 1.05136430e+00
3.57306898e-01 1.17459905e+00 7.88707316e-01 2.07347199e-01
-2.27791056e-01 5.78967273e-01 -5.53838491e-01 1.96796194e-01
3.18520397e-01 -9.62139130e-01 -4.67466861e-01 1.71292007e-01
2.98226148e-01 -3.24807942e-01 -6.16445303e-01 9.88415360e-01
2.57460058e-01 3.41536105e-01 5.04013479e-01 -8.50208342e-01
-9.09244418e-01 2.11940840e-01 -5.51073663e-02 1.03514957e+00
2.81663597e-01 5.30898154e-01 1.08383012e+00 -1.59660113e+00
5.23201942e-01 1.20171082e+00 6.76305667e-02 5.99192739e-01
-1.41347003e+00 -7.23790050e-01 -3.47984225e-01 5.53795457e-01
-1.53319383e+00 -7.64401495e-01 5.18632710e-01 -5.17952859e-01
1.30405307e+00 8.72036293e-02 1.56449878e+00 9.47371781e-01
5.83127618e-01 4.05574560e-01 1.40080786e+00 1.31251663e-01
3.36156577e-01 1.50714561e-01 3.39424312e-02 6.40303254e-01
9.61286128e-02 2.12373048e-01 -8.84889722e-01 5.36672808e-02
6.72013760e-01 3.34910035e-01 -3.69103283e-01 1.97380528e-01
-1.37231159e+00 9.90664244e-01 9.50523674e-01 6.94702804e-01
-8.46578836e-01 -1.29771486e-01 -4.19233143e-02 3.01536191e-02
2.09370926e-01 6.45061672e-01 -1.97325841e-01 9.59705338e-02
-1.23714888e+00 -3.83932114e-01 2.78355390e-01 1.32909551e-01
1.03907025e+00 3.54944319e-01 -2.31983811e-01 9.06685889e-01
3.69728208e-01 2.36345068e-01 1.10755038e+00 -9.02279437e-01
-4.52529266e-02 7.58225143e-01 -3.01356167e-01 -7.61550665e-01
-8.25372279e-01 -3.14243883e-02 -4.94938791e-01 1.90948725e-01
3.05257887e-02 4.31017935e-01 -9.84510601e-01 1.66994977e+00
-3.83684307e-01 -1.26696810e-01 2.39053950e-01 1.25469625e+00
1.13309753e+00 6.02597833e-01 1.64197460e-01 -4.90152240e-01
1.74185443e+00 -4.55671489e-01 -7.03210413e-01 -8.70678961e-01
3.24359149e-01 -9.03669894e-02 1.32263577e+00 2.21447155e-01
-8.29178035e-01 -9.07466650e-01 -9.33917880e-01 -2.54797369e-01
-1.32571176e-01 -6.33523986e-02 6.76942945e-01 5.39323874e-02
-1.40411115e+00 4.52948779e-01 -6.86653435e-01 -7.25528240e-01
9.65724230e-01 5.56741834e-01 -6.87003076e-01 2.29014978e-01
-9.77805734e-01 1.25506449e+00 5.48061430e-01 1.45237178e-01
-1.56508827e+00 -2.01907322e-01 -8.32874060e-01 3.04766595e-01
-4.07523990e-01 -6.81040883e-01 9.05128896e-01 -1.42872751e+00
-1.04668653e+00 1.26940572e+00 -5.14934123e-01 -3.76535475e-01
-7.11404979e-01 2.93048471e-01 -6.94569826e-01 4.82490540e-01
2.32468680e-01 9.25937533e-01 7.13148296e-01 -1.17242897e+00
-1.93325505e-01 -7.04855144e-01 -4.96495843e-01 4.21071738e-01
-2.35938236e-01 -8.87709111e-02 3.50446515e-02 -1.61929846e-01
4.12787080e-01 -8.02147090e-01 1.10477637e-02 1.56428844e-01
1.41477838e-01 -2.99220294e-01 4.12073940e-01 -5.76602399e-01
4.27227527e-01 -2.71007919e+00 6.95766415e-03 1.62085320e-03
5.21556497e-01 -2.18612567e-01 -3.18226010e-01 -7.17089046e-03
-2.54664063e-01 -1.64987624e-01 -1.67222679e-01 7.02850819e-02
-2.85340428e-01 5.14210343e-01 -2.46325880e-01 9.05024529e-01
9.39152911e-02 1.19294071e+00 -8.82929444e-01 -3.81227344e-01
2.84630030e-01 3.11147600e-01 -5.83071530e-01 5.27420223e-01
6.32131621e-02 5.18385351e-01 1.55426458e-01 4.53382462e-01
2.92293638e-01 -3.16523939e-01 3.75589103e-01 -3.54316503e-01
-1.07619248e-01 7.58021057e-01 -2.48542354e-01 1.85139406e+00
-3.57861966e-01 1.24781835e+00 -1.87505484e-01 -9.75914299e-01
9.74157453e-01 3.32711846e-01 1.56441763e-01 -1.28633296e+00
2.40863666e-01 -9.94217545e-02 7.90087759e-01 -7.11243570e-01
1.29752025e-01 -9.18623030e-01 1.45952180e-01 3.91356289e-01
7.20529974e-01 2.58878507e-02 -2.37336338e-01 6.48627579e-02
8.93540084e-01 -4.44429159e-01 5.53793192e-01 -4.62972105e-01
-1.88583672e-01 -1.85436934e-01 3.49243969e-01 5.70073485e-01
3.42771821e-02 5.90072334e-01 3.79400790e-01 -5.85279882e-01
-6.32940173e-01 -1.54951179e+00 -2.94789463e-01 9.58591998e-01
-2.51983851e-02 -3.58222246e-01 -1.87074319e-01 -2.67675817e-01
-3.18284005e-01 1.21391416e+00 -1.04533875e+00 -6.43492639e-01
-5.01048081e-02 -1.04671156e+00 6.83592781e-02 4.05407339e-01
3.15170705e-01 -1.70624316e+00 -8.17936301e-01 1.69224426e-01
-6.18395172e-02 -8.92315388e-01 -8.02849978e-02 6.18211269e-01
-8.93001676e-01 -1.12888825e+00 -1.51919290e-01 -9.32266772e-01
9.03655291e-01 4.50553358e-01 9.67817724e-01 5.56138866e-02
-7.45582879e-01 3.38285148e-01 1.48457184e-01 8.71956497e-02
-1.79112315e-01 -7.29885340e-01 5.26834488e-01 2.13135004e-01
7.59245455e-01 -9.70170677e-01 -9.74750936e-01 5.85860945e-02
-8.81091058e-01 1.37724996e-01 8.62706006e-01 6.43578827e-01
6.76313877e-01 -3.76156747e-01 4.80845362e-01 -8.01284164e-02
6.06309056e-01 -9.21353579e-01 -2.15313569e-01 -2.64409125e-01
-3.67835343e-01 1.43545628e-01 5.19977570e-01 -4.25928384e-01
-9.41891313e-01 -3.04277577e-02 -1.54814020e-01 -4.28000927e-01
-3.26180518e-01 3.10823858e-01 -1.44478351e-01 1.21506467e-01
8.82470191e-01 8.20842326e-01 -8.75526592e-02 -3.53411734e-01
2.13388845e-01 6.61021411e-01 4.98014092e-01 9.63004529e-02
2.55736291e-01 7.91242480e-01 -2.80936420e-01 -9.34127510e-01
-1.08767331e+00 -2.20538765e-01 -5.57310939e-01 -2.32582763e-01
1.27276039e+00 -1.12465930e+00 -8.70479286e-01 -2.03707650e-01
-9.46090162e-01 -2.76541114e-01 -4.19509143e-01 1.01793873e+00
-5.68565488e-01 2.57658027e-03 -4.52849925e-01 -5.03994763e-01
-2.55958676e-01 -1.09035659e+00 5.44832051e-01 1.99338540e-01
-4.62321192e-01 -5.72220743e-01 1.78833768e-01 1.89661011e-01
1.44561380e-01 -3.46324176e-01 1.01966846e+00 -8.53456378e-01
-1.72335207e-01 1.98901832e-01 -2.91271389e-01 2.77181298e-01
3.97494614e-01 -7.49333680e-01 -1.19012022e+00 4.59774435e-02
7.01702356e-01 -4.13758874e-01 1.11232722e+00 3.56739789e-01
1.40617204e+00 -5.68055689e-01 -1.82326764e-01 7.83841491e-01
1.35651398e+00 1.70300797e-01 7.72034168e-01 4.53076661e-02
3.64838183e-01 4.86565232e-01 -1.26556695e-01 3.92497718e-01
2.58913577e-01 5.16559839e-01 4.37946022e-01 5.47526367e-02
-1.33645833e-01 -8.08413327e-02 6.67452812e-01 6.23459995e-01
5.36096767e-02 2.78816730e-01 -7.61359334e-01 7.89867103e-01
-1.24196541e+00 -1.26381004e+00 -1.05419859e-01 1.92930532e+00
6.56891108e-01 1.37642939e-02 -1.87914014e-01 -4.75858063e-01
2.89650559e-01 1.96772963e-01 -6.02234125e-01 -3.81124735e-01
-3.58466834e-01 3.59734327e-01 -7.71549419e-02 -3.04111019e-02
-4.84789252e-01 7.93719769e-01 7.15818310e+00 1.84437633e-01
-9.44126427e-01 5.27070582e-01 1.96311116e-01 -9.57830429e-01
-2.80861914e-01 3.80728170e-02 -4.20495600e-01 4.78651226e-01
1.45715654e+00 -3.37713718e-01 1.08819664e+00 9.18091059e-01
4.77228105e-01 -4.02191818e-01 -1.32537854e+00 1.25853646e+00
5.10379195e-01 -1.56566620e+00 1.33686289e-01 3.07788908e-01
2.86011040e-01 4.01433200e-01 1.15526162e-01 2.83675283e-01
8.38374570e-02 -1.25556457e+00 5.88312566e-01 7.43651927e-01
7.38882601e-01 -3.35497439e-01 3.01898420e-01 2.20327154e-01
-7.29728103e-01 -2.52421916e-01 -6.40309811e-01 -3.97264093e-01
6.74534217e-02 1.87916175e-01 -1.26873386e+00 -6.52983189e-01
9.52007890e-01 7.40706623e-01 -6.43098831e-01 8.56664002e-01
-1.83177114e-01 7.51081765e-01 9.39091071e-02 9.40711945e-02
1.43022671e-01 5.82670942e-02 2.67504632e-01 1.01132011e+00
5.22737026e-01 4.44878906e-01 -3.37311238e-01 1.47415721e+00
-1.82611078e-01 -1.01918645e-01 -7.61110902e-01 -1.09181374e-01
9.13853571e-02 1.55135250e+00 -9.92890358e-01 -4.80115414e-01
-5.71164787e-01 1.10281301e+00 4.51346546e-01 3.05008143e-01
-3.59631807e-01 2.31600448e-01 8.94544661e-01 3.57713550e-01
1.66520998e-01 -3.02298605e-01 -2.17422888e-01 -1.20959175e+00
-4.29227561e-01 -3.40590805e-01 2.06131041e-01 -1.39743578e+00
-1.31838751e+00 9.68150795e-01 -1.18785739e-01 -1.04838407e+00
-1.74961425e-02 -5.18735766e-01 -6.46300375e-01 9.08901155e-01
-9.02742803e-01 -4.38559681e-01 -4.19579893e-01 1.06112063e+00
6.19624615e-01 -1.29809633e-01 1.10894847e+00 -4.57870141e-02
-1.44855574e-01 -2.18017519e-01 2.25339376e-04 1.73367351e-01
4.06554252e-01 -8.72813761e-01 -7.67782703e-02 4.28296149e-01
8.88447046e-01 7.92271018e-01 4.00402486e-01 -5.89588225e-01
-1.25227475e+00 -8.86084378e-01 5.95513403e-01 -4.62708384e-01
4.87390786e-01 -7.30677843e-01 -1.05329692e+00 7.62504697e-01
3.52364302e-01 2.57056385e-01 1.25025392e+00 -1.65305063e-01
-9.83769223e-02 1.06454100e-02 -1.11056256e+00 3.43348235e-01
1.05859709e+00 -9.07666385e-01 -1.46824372e+00 5.01651406e-01
1.66422307e-01 2.56909132e-01 -4.43262994e-01 -3.35341156e-01
3.85748088e-01 -7.95748115e-01 9.04030800e-01 -6.10831916e-01
1.12070292e-01 -2.39013612e-01 -4.59152102e-01 -1.49244428e+00
-7.85805345e-01 2.49214441e-01 9.45310518e-02 5.15628874e-01
1.19332127e-01 -2.89365709e-01 3.95458728e-01 3.88144821e-01
-4.13073123e-01 -8.64562541e-02 -1.45382154e+00 -6.08072579e-01
-5.47650158e-01 -1.85071334e-01 9.72836688e-02 6.54919505e-01
1.46650478e-01 5.66059172e-01 -1.77509785e-01 7.45458975e-02
3.31465602e-01 2.30020732e-01 -3.62712368e-02 -1.29601347e+00
1.94718376e-01 -5.88025153e-02 -5.12432456e-01 -3.90404612e-01
6.00015521e-01 -1.76141143e+00 3.51344019e-01 -2.16164351e+00
6.84449971e-01 2.56710708e-01 -6.83108032e-01 8.23919296e-01
3.09103340e-01 6.80409908e-01 1.97123021e-01 5.00843465e-01
-5.55543065e-01 7.08618283e-01 9.31771576e-01 7.72748888e-02
-3.21701974e-01 -4.54918712e-01 -9.62489486e-01 9.18406844e-01
8.02454114e-01 -8.48990023e-01 -8.75521526e-02 -2.39557579e-01
-1.99026661e-03 8.71606618e-02 9.55474198e-01 -1.31119215e+00
-1.70223594e-01 -1.43738866e-01 1.25026405e+00 -2.83876598e-01
9.19177353e-01 -7.19579756e-01 2.78852969e-01 5.39441943e-01
-1.46107748e-01 -1.98098645e-01 3.13205063e-01 3.38187575e-01
3.13604437e-02 6.80420399e-02 9.96152818e-01 -6.54508471e-01
-1.17731380e+00 -1.66404303e-02 -1.10959983e+00 -1.52542427e-01
7.71296859e-01 -5.85783720e-01 -1.53842628e-01 -3.00956577e-01
-1.12474597e+00 -4.00723338e-01 2.14011893e-01 5.33186555e-01
1.26880765e+00 -1.52288663e+00 -4.76594746e-01 5.65449119e-01
1.05006628e-01 -6.60925984e-01 3.50927353e-01 9.58792925e-01
1.36287630e-01 4.24490213e-01 -8.77311289e-01 -4.96729106e-01
-7.29754806e-01 9.04798985e-01 3.37323606e-01 4.84825015e-01
-8.97856832e-01 7.32661903e-01 7.98119605e-01 2.75932670e-01
-5.30474484e-01 -1.85106814e-01 -3.10421765e-01 4.21880722e-01
7.49413669e-01 -2.29575410e-01 -1.75538342e-02 -1.07765508e+00
-6.53751969e-01 6.45505115e-02 1.80226430e-01 -2.02285796e-01
1.76481068e+00 5.18989284e-03 -4.20996994e-01 8.19556475e-01
1.31760287e+00 -3.01588953e-01 -1.14512050e+00 -5.49568683e-02
-1.11261055e-01 -3.95010442e-01 3.33736926e-01 -8.13454032e-01
-1.27665246e+00 8.96193266e-01 1.06099403e+00 -2.43511334e-01
1.15794432e+00 8.02969992e-01 3.21702451e-01 4.14753318e-01
3.43146682e-01 -8.22350144e-01 4.01057154e-01 6.54601231e-02
1.17178261e+00 -1.10171449e+00 -2.99816951e-02 6.15234554e-01
-1.00751245e+00 1.00685918e+00 6.95668459e-01 -4.88417059e-01
6.42330408e-01 -2.36912280e-01 -3.00688684e-01 -7.83325076e-01
-7.92246580e-01 -3.51817638e-01 5.28230369e-01 3.74025971e-01
2.21976995e-01 2.07133353e-01 -2.18805626e-01 1.07285309e+00
-3.57463866e-01 -2.65497179e-03 4.47214156e-01 3.21087986e-01
-9.44420576e-01 -2.99084723e-01 -4.79103625e-01 1.27414417e+00
-2.58306772e-01 -5.41633546e-01 -2.01480150e-01 4.83937919e-01
4.03637081e-01 7.89706886e-01 7.96259999e-01 -4.07655478e-01
-1.53928110e-03 4.04053658e-01 5.79175889e-01 -1.15147102e+00
-6.10048652e-01 -1.68703660e-01 -1.27334476e-01 -7.66367793e-01
-4.66812193e-01 -7.09372580e-01 -1.64418674e+00 3.03326756e-01
2.38904446e-01 2.93289602e-01 4.35722560e-01 1.04675901e+00
3.70130301e-01 1.78375706e-01 2.33853981e-01 -1.01374853e+00
3.32951903e-01 -1.13296902e+00 -9.47980285e-01 3.16696167e-01
4.15730953e-01 -8.74900520e-01 -5.96980870e-01 1.17987476e-01] | [10.636307716369629, 2.489307165145874] |
508078e9-b738-4d37-ac1c-ad3938fd9858 | probabilistic-deep-learning-with-generalised | null | null | https://openreview.net/forum?id=L_jGauvvbu0 | https://openreview.net/pdf?id=L_jGauvvbu0 | Probabilistic Deep Learning with Generalised Variational Inference | We study probabilistic Deep Learning methods through the lens of Approximate Bayesian Inference. In particular, we examine Bayesian Neural Networks (BNNs), which usually suffer from multiple ill-posed assumptions such as prior and likelihood misspecification. In this direction, we investigate a recently proposed approximate inference framework called Generalised Variational Inference (GVI) in comparison to state-of-the-art methods including standard Variational Inference, Monte-Carlo Dropout, Stochastic gradient Langevin dynamics and Deep Ensembles. Also, we expand the original research around GVI by exploring a broader set of model architectures and mathematical settings on both real and synthetic data. Our experiments demonstrate that approximate posterior distributions derived from such a method offer attractive properties with respect to uncertainty quantification, prior specification robustness and predictive performance, especially in the case of BNNs. | ['Brooks Paige', 'Theo Damoulas', 'Giorgos Felekis'] | 2021-11-22 | null | null | null | pproximateinference-aabi-symposium-2022-2 | ['probabilistic-deep-learning'] | ['computer-vision'] | [ 9.88421496e-03 1.80540860e-01 3.04556876e-01 -4.85527426e-01
-7.00240195e-01 -2.42508367e-01 9.08967793e-01 -3.34363610e-01
-4.93650436e-01 1.26109707e+00 -6.10359572e-02 -2.16795683e-01
-4.87754673e-01 -7.39744067e-01 -1.01235390e+00 -8.08858752e-01
1.02145746e-01 6.90971613e-01 -7.14289024e-02 4.70561147e-01
1.02495700e-01 4.45241034e-01 -1.19713807e+00 -4.48052675e-01
9.76962924e-01 8.12144279e-01 -1.76109985e-01 2.61709124e-01
4.98399921e-02 4.20353860e-01 -4.22292531e-01 -7.35909045e-01
-3.29197675e-01 -8.63702297e-02 -2.72006571e-01 -3.49310964e-01
5.00524938e-01 -6.59600973e-01 -5.59749842e-01 1.54462636e+00
3.66740286e-01 4.17050660e-01 1.10221338e+00 -1.24307573e+00
-5.46439826e-01 9.17282403e-01 -3.44588578e-01 -7.52962977e-02
-3.33899468e-01 2.24846125e-01 7.59527087e-01 -9.40678120e-01
2.41268367e-01 1.59687531e+00 9.24764335e-01 4.98514414e-01
-1.71138787e+00 -6.87332511e-01 2.35106945e-01 -2.90520303e-02
-1.61993945e+00 -3.72278303e-01 5.33849418e-01 -5.74278772e-01
5.25827467e-01 -2.63252944e-01 1.84565037e-01 1.81503904e+00
3.42280179e-01 8.97270799e-01 1.09739316e+00 -1.19756319e-01
7.97690511e-01 -1.33176325e-02 4.01048452e-01 6.03351057e-01
5.82983255e-01 5.59700370e-01 -5.00094175e-01 -3.38271439e-01
8.96175802e-01 1.16155045e-02 -1.92193106e-01 -3.47885072e-01
-1.00769556e+00 9.66805041e-01 -2.93802768e-05 -2.71218210e-01
-2.77431965e-01 7.75268793e-01 1.30524874e-01 -3.00144702e-01
4.25843954e-01 -9.80755612e-02 -2.38306060e-01 -6.11738823e-02
-1.22052801e+00 8.51851285e-01 1.04772556e+00 9.85537529e-01
6.90117776e-01 4.61152256e-01 -5.92861533e-01 5.69668293e-01
9.16924596e-01 7.32922375e-01 -2.21346080e-01 -1.32903969e+00
2.07264021e-01 -3.51025224e-01 4.01979864e-01 -4.84749556e-01
-8.53438079e-02 -7.59421468e-01 -1.28232110e+00 3.06670964e-01
5.96928596e-01 -3.26301277e-01 -1.02803600e+00 2.29160666e+00
1.74160585e-01 6.22321010e-01 -9.31943655e-02 6.41737103e-01
5.04029214e-01 5.33488512e-01 1.17497437e-01 -2.20798209e-01
1.05573714e+00 -4.38917398e-01 -7.40147233e-01 8.46043155e-02
-1.11249581e-01 -8.91248509e-02 7.33494282e-01 5.90219796e-01
-1.10812020e+00 -4.26772118e-01 -9.17304873e-01 1.73411980e-01
-1.43182904e-01 -5.58085293e-02 3.61646801e-01 7.71656096e-01
-1.01984167e+00 9.47782338e-01 -1.47422016e+00 8.42741877e-02
6.04998052e-01 2.38467958e-02 2.12662056e-01 -1.89279482e-01
-1.20194554e+00 7.07593143e-01 4.80849653e-01 4.89072531e-01
-1.50251961e+00 -9.15352106e-01 -7.52638400e-01 3.01895767e-01
4.80757713e-01 -1.08851302e+00 1.42703485e+00 -1.19749554e-01
-1.85943758e+00 1.53209284e-01 -2.29837269e-01 -6.80331767e-01
7.63974249e-01 -4.27957177e-01 1.10411234e-01 -2.48055831e-01
-3.85186881e-01 4.54040945e-01 9.43706512e-01 -1.01703238e+00
-1.57642350e-01 -4.08182472e-01 -5.39702773e-02 -4.22123879e-01
1.22500896e-01 -3.66979033e-01 -2.17686906e-01 -5.69311798e-01
2.10327096e-02 -8.23788464e-01 -1.90272242e-01 5.83467409e-02
-7.55546153e-01 -4.17202860e-01 2.88807541e-01 -3.46066356e-01
9.71564412e-01 -1.79580200e+00 4.09268051e-01 2.83922970e-01
2.42549658e-01 6.34737834e-02 1.88328534e-01 2.80634046e-01
4.91878808e-01 2.11991683e-01 -6.32030904e-01 -6.76166713e-01
7.95168400e-01 4.16221112e-01 -4.98593807e-01 5.31111181e-01
1.71696648e-01 7.93846667e-01 -8.88663173e-01 -4.27289456e-01
1.66773215e-01 9.48996007e-01 -7.12240458e-01 3.70208696e-02
-6.10116184e-01 4.08032775e-01 -5.72601020e-01 3.91251534e-01
7.62904644e-01 -2.73362458e-01 1.43834382e-01 -2.23919228e-01
2.06029247e-02 9.67274830e-02 -1.41733122e+00 1.46879900e+00
-2.50331461e-01 4.80661631e-01 1.58174783e-01 -9.42306817e-01
5.41507244e-01 2.49994934e-01 -1.30497470e-01 2.57114738e-01
1.86943486e-01 1.28783926e-01 -2.34860361e-01 3.58799361e-02
3.98531526e-01 -1.14702992e-01 1.49944440e-01 2.94806749e-01
4.92326975e-01 -5.72266802e-02 2.02131569e-01 2.02033803e-01
7.35251665e-01 6.70112669e-01 -2.31806003e-02 -4.55073804e-01
1.89791676e-02 -5.93820333e-01 6.23026311e-01 1.66310227e+00
-6.23014756e-02 4.56785679e-01 6.00309193e-01 9.05384645e-02
-1.09738624e+00 -1.74943781e+00 -7.71935821e-01 4.72025573e-01
-2.80887663e-01 -1.31697923e-01 -9.32936370e-01 -1.68610916e-01
1.56810641e-01 1.11981595e+00 -4.25995171e-01 -2.39490539e-01
-3.44951488e-02 -1.17534590e+00 6.39457583e-01 5.17940283e-01
4.75076467e-01 -6.02884591e-01 -3.05699944e-01 3.46209139e-01
1.07190706e-01 -9.85111892e-01 1.93916336e-02 1.01185165e-01
-9.88686621e-01 -5.23781240e-01 -7.76614428e-01 5.62617667e-02
2.28073046e-01 -5.98087609e-01 1.19045377e+00 -7.51439095e-01
5.53251468e-02 3.55644792e-01 1.90585613e-01 -2.82332301e-01
-4.76274043e-01 -5.48263974e-02 3.76686633e-01 -8.13133121e-02
1.52949750e-01 -9.52865005e-01 -4.45011199e-01 3.09605785e-02
-9.21607435e-01 -2.58448292e-02 5.61596036e-01 9.69579458e-01
6.78577244e-01 1.32508680e-01 3.14435065e-01 -8.10947776e-01
5.60889304e-01 -5.36723495e-01 -1.11402643e+00 3.38353753e-01
-6.77614868e-01 5.53470850e-01 2.48280063e-01 -4.68277544e-01
-1.54834104e+00 -4.97498542e-01 -8.92307907e-02 -5.37286818e-01
-3.17440629e-01 7.52828538e-01 -1.21099748e-01 2.84853667e-01
4.20392513e-01 1.31177440e-01 -1.01113051e-01 -6.79339111e-01
4.10355955e-01 2.74539888e-01 7.32910991e-01 -9.91193295e-01
6.31920576e-01 5.10311544e-01 4.05767173e-01 -6.69475317e-01
-9.92206931e-01 3.94132555e-01 -3.79190832e-01 -5.85824773e-02
7.13880181e-01 -8.38875711e-01 -1.12156892e+00 8.08666646e-01
-1.27443457e+00 -3.90876025e-01 -4.42191996e-02 8.13197553e-01
-6.35503948e-01 2.79723018e-01 -7.80180335e-01 -1.42329812e+00
-2.82317866e-02 -1.34826791e+00 9.62863505e-01 3.48472446e-01
-2.15219110e-02 -1.20805526e+00 2.02677920e-01 -1.54543877e-01
4.28795576e-01 2.76807457e-01 8.74430776e-01 -3.71652573e-01
-7.23746896e-01 6.88200742e-02 -3.59329492e-01 4.07947272e-01
-3.84324104e-01 4.05213356e-01 -1.11694109e+00 -1.71886936e-01
-7.97223598e-02 -2.22668782e-01 1.26043797e+00 1.00799799e+00
1.11143708e+00 -2.23572105e-01 -3.04752022e-01 6.93844557e-01
1.52482009e+00 -1.95134401e-01 4.08076167e-01 -2.49951899e-01
5.90428710e-01 4.94972289e-01 1.21791745e-02 6.68624520e-01
2.31852040e-01 4.11951482e-01 5.71027577e-01 6.94486141e-01
3.04818243e-01 -4.92089421e-01 1.55112326e-01 3.76720428e-01
7.25575313e-02 -4.18011397e-01 -8.77612472e-01 3.35632026e-01
-2.06490588e+00 -9.68764424e-01 2.55628824e-02 2.23198891e+00
1.08470619e+00 2.71023422e-01 -1.07988618e-01 -2.89093256e-01
7.19612002e-01 1.35148996e-02 -9.31449950e-01 1.59731492e-01
4.38481085e-02 2.15411887e-01 3.10555011e-01 5.92653394e-01
-9.45202768e-01 7.25621998e-01 6.76785707e+00 1.10685110e+00
-4.32810724e-01 2.42433861e-01 8.06713223e-01 -2.84319967e-01
-4.50143516e-01 -3.10148578e-02 -1.26842213e+00 6.32539928e-01
1.24238360e+00 2.03581065e-01 6.13172650e-01 6.46441460e-01
1.95338324e-01 -1.68114454e-01 -1.29827714e+00 8.19908738e-01
-3.95616293e-01 -1.43502390e+00 8.39075353e-03 1.82092756e-01
8.37503374e-01 2.08778739e-01 3.03124905e-01 5.27410865e-01
9.49835539e-01 -1.10194468e+00 8.55607867e-01 1.27648485e+00
4.99351263e-01 -8.62496793e-01 6.77762508e-01 3.04087520e-01
-3.96663785e-01 3.19604069e-01 -5.35259187e-01 2.79630516e-02
4.18245554e-01 1.15173662e+00 -2.19734907e-01 2.42501631e-01
6.74647629e-01 5.50050795e-01 4.86659706e-02 9.48838770e-01
-3.33303124e-01 1.01641142e+00 -7.80961692e-01 -1.97532237e-01
2.99673468e-01 -5.47850609e-01 7.61527240e-01 9.82815087e-01
3.68718565e-01 -1.59073651e-01 -2.86312521e-01 1.94502747e+00
-7.31450990e-02 -8.27835560e-01 -3.84199381e-01 -2.25185156e-02
7.44200528e-01 8.98312211e-01 -3.16417277e-01 -3.63242239e-01
-1.30242318e-01 4.35653538e-01 4.15030718e-01 9.88913655e-01
-9.51662600e-01 -2.01589361e-01 7.09710240e-01 -4.31934714e-01
5.38332760e-01 -3.46388847e-01 -1.49872392e-01 -1.35978055e+00
-1.41435325e-01 -4.05302048e-01 -7.03285588e-03 -7.85949588e-01
-1.56737006e+00 2.44645253e-01 6.19776785e-01 -4.12460387e-01
-6.03799343e-01 -1.04657483e+00 -4.92228359e-01 9.92992222e-01
-1.36399221e+00 -8.73332500e-01 1.23974606e-01 1.65135592e-01
6.44561201e-02 -1.43633541e-02 4.38504517e-01 -6.75863177e-02
-8.86100233e-01 5.87465167e-01 9.15028453e-01 -5.70782125e-02
1.87238604e-01 -1.24021614e+00 3.57165396e-01 8.11310768e-01
2.60732949e-01 8.57996583e-01 1.10796249e+00 -4.17739868e-01
-1.36631560e+00 -1.01179433e+00 2.76450068e-01 -5.21981299e-01
7.80087113e-01 -4.93422598e-01 -8.42949510e-01 7.14868963e-01
-1.48290470e-01 -1.13018632e-01 2.73551702e-01 3.01648736e-01
-3.60745430e-01 -2.32749265e-02 -1.20716512e+00 7.00665236e-01
9.11989212e-01 -5.27914464e-01 -3.96312922e-01 1.35572582e-01
6.79218352e-01 -2.50343144e-01 -9.56992745e-01 4.86126393e-01
8.30229402e-01 -1.01628566e+00 1.06369281e+00 -6.04154706e-01
3.79977435e-01 -9.14541483e-02 -4.19147760e-01 -1.17006969e+00
-1.29724726e-01 -8.64754438e-01 -6.15365326e-01 1.44063663e+00
1.77295327e-01 -6.49465144e-01 5.83469391e-01 8.01224351e-01
-4.23829742e-02 -5.85464418e-01 -1.14874852e+00 -8.67241144e-01
5.04418015e-01 -9.45510507e-01 6.13562167e-01 2.34811395e-01
-5.53496361e-01 -2.93902140e-02 -3.17898542e-01 2.79483378e-01
1.50241399e+00 -1.54877186e-01 4.50688720e-01 -1.46006906e+00
-5.85851371e-01 -5.26238501e-01 -1.82827964e-01 -1.13597381e+00
5.89610517e-01 -4.80982900e-01 2.42565751e-01 -1.15388739e+00
3.96864474e-01 4.75471914e-02 -2.58813858e-01 -7.73284808e-02
-1.96480885e-01 2.91374307e-02 -1.77516446e-01 -3.80773284e-02
-6.94731891e-01 1.07822037e+00 7.24744260e-01 -1.04161672e-01
2.72290766e-01 2.06730574e-01 -3.36793035e-01 8.93669367e-01
4.55947697e-01 -6.59727395e-01 -4.77537453e-01 -4.45473433e-01
2.98316866e-01 2.02445745e-01 9.09541547e-01 -8.51967692e-01
2.19642267e-01 -1.99906468e-01 2.60547876e-01 -8.27254474e-01
4.35270369e-01 -2.61684716e-01 3.14854354e-01 2.52246350e-01
-3.46389830e-01 -5.17560840e-01 1.98181659e-01 1.09037066e+00
-5.55803953e-03 -3.90028149e-01 7.85311759e-01 -5.10210060e-02
-2.12203547e-01 5.38648903e-01 -5.10891378e-01 3.37011814e-01
2.80883729e-01 1.29536152e-01 -2.46073470e-01 -3.63130540e-01
-9.09830332e-01 1.39456570e-01 5.17830923e-02 -3.78061801e-01
5.51753283e-01 -1.14873850e+00 -9.02447522e-01 -1.75265700e-01
-2.44129211e-01 3.03737521e-01 3.87828857e-01 8.62048626e-01
-7.12503493e-02 3.31118077e-01 2.55062729e-01 -8.62608552e-01
-2.90308565e-01 1.71341717e-01 5.49099743e-01 -2.13930875e-01
-3.92020881e-01 9.22664404e-01 3.51733804e-01 -5.56514382e-01
6.04469895e-01 -6.26328528e-01 2.94559211e-01 -4.16668802e-01
2.88430899e-01 6.27994537e-01 -4.23626125e-01 -1.16778528e-02
-1.70070693e-01 5.17054908e-02 2.53166687e-02 -6.19560301e-01
1.20612383e+00 -1.99227706e-01 3.28424647e-02 8.42447817e-01
8.37254524e-01 -7.23727465e-01 -2.07574224e+00 -5.51300168e-01
-5.58859147e-02 -7.97064975e-02 4.52777892e-01 -6.99986875e-01
-7.95381844e-01 1.21789134e+00 5.66927791e-01 -1.57791942e-01
4.23471212e-01 -4.73431125e-02 2.20790744e-01 6.58109903e-01
3.58979076e-01 -8.22871029e-01 -2.96744645e-01 6.14007235e-01
5.16973734e-01 -1.15602195e+00 -1.53447026e-02 3.33748162e-01
-1.00770794e-01 9.13184047e-01 3.06969374e-01 -3.28910381e-01
1.06898201e+00 3.25122565e-01 -7.13669062e-01 4.56766747e-02
-8.37458551e-01 1.77133590e-01 3.43745142e-01 6.50295496e-01
2.49874517e-01 -2.90556066e-02 1.32414356e-01 9.22301471e-01
-4.57057729e-03 1.21217705e-01 4.76221174e-01 4.55763876e-01
-1.52740970e-01 -5.99653363e-01 -1.02509685e-01 4.26923633e-01
-6.11012340e-01 -3.90751243e-01 2.79154122e-01 7.54994631e-01
-4.02446389e-01 6.29728138e-01 1.76721320e-01 1.77661121e-01
-1.79791540e-01 1.42893791e-01 7.04523027e-01 -2.90653974e-01
1.88888833e-01 1.66353285e-01 -1.19637132e-01 -4.43903655e-01
-2.98857272e-01 -9.49450195e-01 -7.54680097e-01 -4.81738031e-01
-2.82329530e-01 -6.21732995e-02 8.36941361e-01 1.32675409e+00
3.25759888e-01 5.61767399e-01 -2.30983615e-01 -1.05818915e+00
-1.44972622e+00 -1.29904878e+00 -8.99840534e-01 -8.13293234e-02
4.39068258e-01 -9.80439365e-01 -6.05960369e-01 -2.59750187e-01] | [7.0425333976745605, 3.882415294647217] |
99c25dd2-aedf-4859-a13f-dcfddb066030 | automated-machine-learning-for-remaining | 2306.12215 | null | https://arxiv.org/abs/2306.12215v1 | https://arxiv.org/pdf/2306.12215v1.pdf | Automated Machine Learning for Remaining Useful Life Predictions | Being able to predict the remaining useful life (RUL) of an engineering system is an important task in prognostics and health management. Recently, data-driven approaches to RUL predictions are becoming prevalent over model-based approaches since no underlying physical knowledge of the engineering system is required. Yet, this just replaces required expertise of the underlying physics with machine learning (ML) expertise, which is often also not available. Automated machine learning (AutoML) promises to build end-to-end ML pipelines automatically enabling domain experts without ML expertise to create their own models. This paper introduces AutoRUL, an AutoML-driven end-to-end approach for automatic RUL predictions. AutoRUL combines fine-tuned standard regression methods to an ensemble with high predictive power. By evaluating the proposed method on eight real-world and synthetic datasets against state-of-the-art hand-crafted models, we show that AutoML provides a viable alternative to hand-crafted data-driven RUL predictions. Consequently, creating RUL predictions can be made more accessible for domain experts using AutoML by eliminating ML expertise from data-driven model construction. | ['Marco F. Huber', 'Marius Lindauer', 'Peter Zeiler', 'Fabian Mauthe', 'Marc-André Zöller'] | 2023-06-21 | null | null | null | null | ['automl', 'management'] | ['methodology', 'miscellaneous'] | [-2.28751986e-03 2.81564087e-01 1.86091930e-01 -3.70517820e-01
-9.77510393e-01 -2.23675177e-01 2.86956877e-01 6.64922655e-01
9.63609517e-02 1.00862026e+00 -4.50407892e-01 -7.57732272e-01
-5.09582639e-01 -8.19653571e-01 -5.49339294e-01 -5.89440703e-01
7.83300698e-02 9.72093046e-01 4.76308942e-01 -2.42553473e-01
1.92003697e-01 9.60229516e-01 -1.50909209e+00 4.02817577e-01
1.02461267e+00 1.24880219e+00 -1.24403767e-01 7.14048684e-01
2.10007727e-01 1.09291971e+00 -2.86481410e-01 1.34283796e-01
3.91447321e-02 -3.54421765e-01 -4.64290261e-01 -2.76506245e-01
1.76026430e-02 -1.34660341e-02 2.41007842e-02 5.72583452e-02
4.66259211e-01 -1.61654189e-01 8.34119081e-01 -1.17629993e+00
2.47747861e-02 1.89293817e-01 3.57711352e-02 -2.49114752e-01
3.41391295e-01 4.51981902e-01 7.08636284e-01 -7.33100951e-01
3.26139569e-01 9.59553182e-01 1.05480480e+00 2.56056219e-01
-1.51376176e+00 -2.68113732e-01 -1.88503444e-01 1.56903431e-01
-1.28380370e+00 -2.19338343e-01 9.20876801e-01 -7.47012973e-01
1.12450469e+00 4.25360203e-01 6.61188126e-01 8.47253561e-01
8.77742469e-01 3.98413986e-01 1.40899420e+00 -5.53115785e-01
7.17191219e-01 2.56121963e-01 8.29237029e-02 6.12372518e-01
3.26948762e-01 6.36021256e-01 -4.67153043e-01 -3.90944123e-01
6.31929874e-01 -6.65862411e-02 -9.51295197e-02 -5.80913067e-01
-9.99071836e-01 5.51309884e-01 9.45619345e-02 -4.85902466e-02
-5.41771412e-01 1.03754260e-01 3.27379972e-01 3.41491491e-01
5.08612752e-01 8.55898619e-01 -1.07540536e+00 1.30708411e-01
-1.17958999e+00 4.89338100e-01 9.08197820e-01 5.65925360e-01
6.51240051e-01 1.88634858e-01 -3.38634014e-01 4.85726506e-01
3.38317156e-01 5.65813839e-01 5.23966504e-03 -8.67127061e-01
-4.36900780e-02 7.57145047e-01 4.64482844e-01 -4.94255453e-01
-6.06064022e-01 -6.38614237e-01 -5.78244269e-01 8.09323907e-01
2.17994794e-01 -1.77676976e-01 -8.92098546e-01 9.51194108e-01
2.70159543e-01 1.39305498e-02 1.12584203e-01 5.49404085e-01
4.33535486e-01 5.56500137e-01 1.00790016e-01 -3.96149665e-01
7.99662530e-01 -6.42949462e-01 -3.21230203e-01 -4.71822843e-02
6.71006739e-01 -3.96895200e-01 1.05936158e+00 9.28875744e-01
-7.62512743e-01 -5.67546785e-01 -1.04012716e+00 3.48277479e-01
-2.37861186e-01 1.71292543e-01 4.56773460e-01 4.65452641e-01
-5.87336540e-01 1.33598030e+00 -9.60435569e-01 -7.10309371e-02
3.01399410e-01 2.70248801e-01 -1.90191194e-01 5.48919663e-02
-1.19730520e+00 1.34637690e+00 3.17816585e-01 -7.00559542e-02
-1.25442338e+00 -1.12817001e+00 -6.20959044e-01 -1.54621154e-01
4.80791926e-01 -8.36355746e-01 1.38864386e+00 -3.35464239e-01
-1.65839577e+00 1.66966215e-01 1.33829907e-01 -6.55751824e-01
8.64671290e-01 -4.15672660e-01 -6.18595958e-01 -6.05135188e-02
-2.93510258e-01 -1.19555360e-02 1.02505457e+00 -1.47438514e+00
-4.05551851e-01 1.74366981e-01 -3.43928099e-01 -4.52198356e-01
1.32602185e-01 -3.97304118e-01 2.10815594e-01 -3.04624408e-01
-9.61720869e-02 -6.34009421e-01 -4.27095860e-01 3.64602245e-02
-5.13247073e-01 -9.25370082e-02 8.07006419e-01 -8.73595178e-01
1.37580562e+00 -1.40357673e+00 -1.99541867e-01 4.33172107e-01
6.96587712e-02 1.57020271e-01 3.89237761e-01 9.85247076e-01
-9.89776328e-02 -3.11575741e-01 -6.32032633e-01 -2.61240620e-02
-1.27430484e-01 1.92789286e-01 -2.36168042e-01 2.91082829e-01
4.74584967e-01 6.96565568e-01 -8.52091551e-01 -5.55102348e-01
6.71412766e-01 3.59022796e-01 -3.31694424e-01 4.51159239e-01
-6.84360802e-01 1.06402719e+00 -5.08056879e-01 8.28463316e-01
2.32370570e-01 -1.80583283e-01 1.49624795e-01 -2.27290750e-01
-3.05198524e-02 -8.27904418e-02 -8.70272040e-01 1.25012815e+00
-1.07768989e+00 8.99442509e-02 -4.52061355e-01 -9.30560768e-01
1.19382048e+00 4.96982217e-01 6.48810506e-01 -3.85569453e-01
9.96221006e-02 4.32502151e-01 -2.06557959e-01 -5.20470083e-01
-1.03684418e-01 -5.25434315e-01 -3.03951740e-01 2.14707464e-01
1.73415408e-01 -6.82423830e-01 -6.21144585e-02 -2.30385065e-01
1.47936857e+00 5.00994444e-01 5.88472426e-01 -3.99761319e-01
8.01047564e-01 2.82724231e-01 6.06991351e-01 4.41910505e-01
3.08570385e-01 3.49477559e-01 3.73205483e-01 -7.16231823e-01
-1.17244780e+00 -9.26271915e-01 -2.52319872e-01 5.21330297e-01
-3.84172797e-01 -3.99590224e-01 -6.20475471e-01 -7.39226341e-01
3.75794619e-01 1.13661695e+00 -5.38592160e-01 -2.14906767e-01
-3.23957503e-01 -3.53659630e-01 2.26242334e-01 5.06139755e-01
-1.78683102e-01 -1.07488513e+00 -6.41482651e-01 5.68840861e-01
5.39058208e-01 -9.04836535e-01 3.64746958e-01 2.42617488e-01
-1.10972726e+00 -1.24191749e+00 -3.00395519e-01 1.09195620e-01
7.46524930e-01 -7.32865214e-01 1.37993407e+00 8.40737000e-02
-6.72513843e-01 2.07021907e-01 -2.40317181e-01 -7.37207830e-01
-1.01087022e+00 -1.43048346e-01 1.61426902e-01 -2.33039051e-01
-1.81232363e-01 -7.97346234e-01 -5.43313861e-01 3.63735110e-01
-6.73741877e-01 2.87288576e-01 7.77456820e-01 7.14379430e-01
7.99569547e-01 3.75020415e-01 8.84630680e-01 -1.18483591e+00
5.52481592e-01 -2.78939575e-01 -8.10683429e-01 5.08467376e-01
-1.35039234e+00 5.11267304e-01 1.12557304e+00 -2.49378294e-01
-1.06524873e+00 3.52650821e-01 -2.72625327e-01 -4.90348309e-01
-3.93059254e-01 6.39596581e-01 -2.53163487e-01 -1.92364492e-02
8.83035004e-01 -1.35376021e-01 -1.35200366e-01 -8.67724359e-01
1.43103167e-01 5.95403910e-01 4.05674517e-01 -8.99107635e-01
8.01895738e-01 4.46331650e-02 5.93216002e-01 -4.18303937e-01
-9.25288320e-01 -4.69618849e-02 -7.87423551e-01 -5.12518466e-01
3.66263300e-01 -6.01109624e-01 -7.07681894e-01 3.83206844e-01
-8.41554344e-01 -5.75531960e-01 -7.34727204e-01 1.57385424e-01
-8.90238822e-01 1.20918803e-01 -2.50722021e-01 -1.17174160e+00
-6.60769820e-01 -8.53583634e-01 1.05412209e+00 -1.35041237e-01
-5.66840947e-01 -1.03450847e+00 1.97034299e-01 2.84780860e-01
4.62363929e-01 8.55017900e-01 1.17768645e+00 -7.22308338e-01
-1.21844538e-01 -9.38191473e-01 6.89811409e-02 6.43738925e-01
3.53345014e-02 1.70277834e-01 -1.15172851e+00 -2.20194653e-01
-7.53722042e-02 -2.16181323e-01 4.94917125e-01 9.63912383e-02
1.26836288e+00 -1.53956860e-01 -5.24930954e-01 -1.05174772e-01
1.48470581e+00 1.90525260e-02 4.77448761e-01 8.40422213e-02
5.05755603e-01 6.04716837e-01 9.28062141e-01 6.10089958e-01
5.03847264e-02 5.81623673e-01 2.12104931e-01 -9.89544690e-02
6.99354410e-02 -3.44723642e-01 1.66303620e-01 4.25997585e-01
-2.87994653e-01 -4.37398665e-02 -1.44295514e+00 1.83128163e-01
-1.89455688e+00 -5.21725476e-01 -2.95076966e-01 2.33939457e+00
9.28895652e-01 6.29491925e-01 -1.35135502e-01 6.73960626e-01
2.47251406e-01 -5.10980129e-01 -8.74518991e-01 -2.63946414e-01
4.14363712e-01 5.28637528e-01 5.66728532e-01 5.11086345e-01
-8.78778875e-01 6.17008030e-01 6.37616634e+00 8.93413603e-01
-1.00418711e+00 6.00559153e-02 5.41208923e-01 1.22995004e-01
-9.98999998e-02 3.74969363e-01 -7.09291637e-01 2.03403354e-01
1.50046110e+00 -1.42750889e-01 2.31363654e-01 9.84734416e-01
7.08462059e-01 -2.25543231e-01 -1.48003089e+00 3.84517103e-01
-4.34090078e-01 -1.32550442e+00 -1.80880532e-01 -1.19072080e-01
4.91930336e-01 -5.24609208e-01 -6.05279565e-01 3.30515087e-01
3.33451301e-01 -1.00677133e+00 6.86217487e-01 1.36541402e+00
9.57038403e-01 -7.68823028e-01 9.65260565e-01 6.35547042e-01
-9.34308231e-01 -2.60685652e-01 -1.36246132e-02 -9.14883763e-02
4.08400625e-01 1.38983369e+00 -1.11967874e+00 9.57476199e-01
3.31267595e-01 4.99001235e-01 -7.10561693e-01 9.18029726e-01
-2.67305613e-01 8.70705009e-01 -2.70132422e-01 2.99039960e-01
-4.70155120e-01 3.79895836e-01 3.03703487e-01 9.44570780e-01
5.32012224e-01 -2.29716808e-01 2.18664557e-01 8.12867880e-01
4.69035238e-01 -1.88118502e-01 -3.98578554e-01 -5.18783592e-02
2.27734894e-01 1.13596094e+00 -5.37902236e-01 -2.68414408e-01
2.56831590e-02 5.85217893e-01 -8.28965083e-02 -1.95733439e-02
-7.62743533e-01 -3.02061409e-01 2.14929894e-01 9.81109500e-01
-1.13566048e-01 -1.36412904e-01 -4.19041812e-01 -5.13817012e-01
-3.41868609e-01 -7.79874384e-01 7.14067593e-02 -9.63325679e-01
-1.55036867e+00 5.75463772e-01 3.23504657e-01 -1.53020489e+00
-5.96072733e-01 -8.17237616e-01 -6.54077053e-01 9.90164220e-01
-1.27285409e+00 -1.24227655e+00 -3.35778564e-01 1.20231166e-01
4.34700429e-01 -1.81430921e-01 1.06964111e+00 2.13919766e-02
-3.18974227e-01 1.25713587e-01 1.69850633e-01 -3.99638385e-01
6.96292102e-01 -1.25855994e+00 2.41420031e-01 4.78111297e-01
-4.29299623e-01 2.09419839e-02 1.20000148e+00 -9.85841453e-01
-1.39432836e+00 -1.27913332e+00 5.48865139e-01 -7.45553970e-01
7.40776896e-01 -8.98398459e-02 -9.67128932e-01 8.93997699e-02
-3.82996649e-01 3.73791993e-01 6.29014134e-01 -7.68619403e-02
1.23323113e-01 -3.80877346e-01 -1.33097446e+00 2.16778480e-02
6.16460502e-01 -4.11637187e-01 -4.33187842e-01 3.82683724e-01
3.75883430e-01 -3.61900359e-01 -1.34505785e+00 9.00623143e-01
5.63218832e-01 -8.93744290e-01 9.17462230e-01 -6.21202886e-01
4.33327019e-01 -4.85384196e-01 8.30154270e-02 -1.23182809e+00
-7.51444995e-02 -3.84497166e-01 -8.09788346e-01 1.03571439e+00
6.61756396e-01 -5.67278087e-01 5.28859019e-01 1.01525831e+00
-2.76944846e-01 -1.19898486e+00 -7.81977296e-01 -9.79539454e-01
1.94622353e-01 -7.89130926e-01 6.53539836e-01 4.91506606e-01
-1.40876740e-01 9.29303169e-02 -3.27113301e-01 3.38152468e-01
6.03163421e-01 1.56979829e-01 6.64029360e-01 -1.97953939e+00
-5.83025455e-01 1.30536824e-01 -4.03181016e-01 -6.96872920e-02
9.41923484e-02 -4.85900223e-01 1.64855808e-01 -1.67245603e+00
-2.52045274e-01 -6.98442578e-01 -4.83661771e-01 7.75587857e-01
1.65914103e-01 -8.42551515e-02 -1.38133347e-01 2.42500201e-01
-2.14319050e-01 4.12631750e-01 1.22747409e+00 7.62035251e-02
-2.42127195e-01 2.89438069e-01 -3.13493013e-01 6.79276288e-01
9.09332752e-01 -5.74485540e-01 -3.65690976e-01 4.54844534e-01
3.22683513e-01 4.84366655e-01 5.92928588e-01 -1.55324066e+00
-9.38773826e-02 -3.50381374e-01 5.26111662e-01 -5.06924093e-01
1.37473479e-01 -1.02244794e+00 6.30822897e-01 5.44758081e-01
-1.13658905e-02 -4.38858032e-01 3.68236393e-01 6.40667975e-01
-4.83595617e-02 -1.07186802e-01 7.94652760e-01 -1.76621545e-02
-4.04532105e-01 1.45992145e-01 -3.01685959e-01 -4.66386467e-01
1.23153687e+00 -6.80904388e-02 -4.06634659e-02 -1.65790349e-01
-9.42568719e-01 9.73243788e-02 4.22409117e-01 1.04272887e-01
5.07127523e-01 -8.82421970e-01 -6.29031479e-01 2.21132532e-01
3.14975232e-01 1.36554345e-01 3.39757025e-01 7.56199956e-01
-7.05301404e-01 5.71429074e-01 -1.11580938e-01 -3.93890977e-01
-9.63143349e-01 6.39776230e-01 5.12556016e-01 -7.76845694e-01
-7.44489253e-01 4.82717663e-01 -3.66328508e-01 -6.77113473e-01
-2.72404104e-01 -5.30817807e-01 8.75027180e-02 -3.41692686e-01
2.02778667e-01 5.36824107e-01 5.97407460e-01 -1.22432813e-01
-2.64797837e-01 4.10994232e-01 3.09217125e-01 1.14369892e-01
1.53364873e+00 3.84746730e-01 2.06241861e-01 8.07276189e-01
4.60017085e-01 -1.70905411e-01 -1.47804415e+00 1.16792418e-01
2.82719225e-01 -8.02940503e-02 3.93915743e-01 -1.37167430e+00
-7.77226150e-01 8.88600886e-01 5.35176575e-01 1.06165990e-01
1.16986561e+00 -1.31779730e-01 7.44011939e-01 5.96877813e-01
7.73687303e-01 -1.19216931e+00 -1.84305906e-01 4.26074453e-02
1.26507342e+00 -1.04195261e+00 4.33401823e-01 -4.08012629e-01
-7.24510968e-01 1.19556272e+00 4.37063962e-01 -8.97934958e-02
1.03223169e+00 3.68929178e-01 -8.93103704e-02 -1.14445724e-01
-9.33226764e-01 2.97135890e-01 4.50705469e-01 4.89393443e-01
3.23397279e-01 6.08064383e-02 -6.74715117e-02 1.09706199e+00
-1.06786378e-02 4.87537444e-01 1.77728698e-01 1.38754439e+00
-5.38743556e-01 -1.62486160e+00 -6.09711230e-01 9.28221583e-01
-1.72864556e-01 9.03585851e-02 -1.65919080e-01 6.62520349e-01
2.88857430e-01 9.08965588e-01 -5.73708415e-01 -4.55716908e-01
5.95871627e-01 3.89300406e-01 4.05390322e-01 -6.40988588e-01
-2.99837649e-01 -3.49289745e-01 3.74519438e-01 -5.06081104e-01
1.82946116e-01 -4.62400168e-01 -1.33699572e+00 -1.14902683e-01
-3.21782231e-01 1.47722691e-01 6.17948771e-01 1.16137242e+00
4.42058682e-01 1.00812268e+00 6.03817880e-01 -1.08960783e+00
-5.11359155e-01 -9.46140349e-01 -6.74166799e-01 -5.33400662e-03
2.38846466e-01 -9.58549738e-01 -2.27777928e-01 1.63960591e-01] | [6.7730937004089355, 2.666893243789673] |
ea2c4870-1d1a-4fc7-b4ca-e13d644c3023 | a-whisper-transformer-for-audio-captioning | 2305.0969 | null | https://arxiv.org/abs/2305.09690v1 | https://arxiv.org/pdf/2305.09690v1.pdf | A Whisper transformer for audio captioning trained with synthetic captions and transfer learning | The field of audio captioning has seen significant advancements in recent years, driven by the availability of large-scale audio datasets and advancements in deep learning techniques. In this technical report, we present our approach to audio captioning, focusing on the use of a pretrained speech-to-text Whisper model and pretraining on synthetic captions. We discuss our training procedures and present our experiments' results, which include model size variations, dataset mixtures, and other hyperparameters. Our findings demonstrate the impact of different training strategies on the performance of the audio captioning model. Our code and trained models are publicly available on GitHub and Hugging Face Hub. | ['Radosław Winiecki', 'Jürgen Kieslich', 'Adam Hájek', 'Marek Kadlčík'] | 2023-05-15 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 4.07599211e-01 2.59692639e-01 1.83818787e-02 -5.84430456e-01
-1.58691430e+00 -4.97174501e-01 4.62915629e-01 -1.78910613e-01
1.24330848e-01 6.00236833e-01 8.43878329e-01 7.25399032e-02
3.09982568e-01 -5.71584851e-02 -7.72930682e-01 -2.72667587e-01
-2.54163593e-01 6.73589647e-01 1.55282277e-03 -1.71431810e-01
-8.67389664e-02 8.73229578e-02 -1.55358672e+00 8.45018029e-01
4.89259623e-02 1.09745502e+00 -1.85304433e-01 1.18116677e+00
-8.25499371e-02 6.87518656e-01 -9.75082934e-01 -4.84148830e-01
-8.41276869e-02 -3.56681436e-01 -8.49448442e-01 -1.05257213e-01
5.40293813e-01 -2.63145834e-01 -4.62321043e-01 5.75107813e-01
1.07778609e+00 -1.17395289e-01 5.07457852e-01 -1.60315847e+00
-6.09225273e-01 1.02911961e+00 -1.67334422e-01 3.36975455e-01
6.79406643e-01 -7.83015788e-02 1.04214394e+00 -7.53569603e-01
2.76131481e-01 1.35697997e+00 7.82940328e-01 9.18596745e-01
-1.21380937e+00 -9.56176221e-01 -1.01436138e-01 2.26576731e-01
-1.61313808e+00 -1.08568037e+00 9.07939851e-01 -4.23561960e-01
7.78181016e-01 3.15638155e-01 2.85499215e-01 1.73175955e+00
-4.94554460e-01 7.81290114e-01 5.73936999e-01 -6.68975770e-01
2.72705674e-01 2.60614544e-01 -1.96034864e-01 1.44548491e-01
-3.85863334e-01 -7.63608515e-02 -7.83904850e-01 -4.88632739e-01
6.50432348e-01 -9.06934440e-01 -1.44820705e-01 2.54005883e-02
-9.81077492e-01 7.45313406e-01 -5.61373830e-02 1.13264315e-01
-1.50996432e-01 6.09114826e-01 6.34354472e-01 1.08156137e-01
6.28678143e-01 3.75723749e-01 -4.30922538e-01 -6.21263206e-01
-1.03045273e+00 2.96157539e-01 7.79150784e-01 1.09411860e+00
6.69807792e-02 2.75462717e-01 -3.47420722e-01 1.30130148e+00
3.88032764e-01 4.29272562e-01 6.03697777e-01 -1.09453797e+00
6.70779049e-01 -3.37917745e-01 1.93032026e-01 -6.33957505e-01
-1.78745046e-01 -1.19514845e-01 -3.19243610e-01 -5.36527157e-01
1.07350893e-01 -5.11958838e-01 -1.07569408e+00 1.69053352e+00
-6.16706647e-02 6.20172620e-01 2.86464784e-02 6.63569987e-01
1.09120858e+00 8.39384556e-01 2.20755756e-01 9.04124230e-02
1.23198295e+00 -9.82972503e-01 -8.91450286e-01 -2.03767315e-01
3.68969172e-01 -1.08788645e+00 1.23603630e+00 1.68761879e-01
-1.17620134e+00 -5.97396612e-01 -8.23899627e-01 5.06956279e-02
-4.09744456e-02 2.31364638e-01 5.59461474e-01 1.07269764e+00
-1.29369032e+00 1.23220660e-01 -7.22349226e-01 -4.07251328e-01
4.49716002e-01 4.46509659e-01 -2.35279843e-01 1.75819635e-01
-1.10111678e+00 3.11493486e-01 2.49809369e-01 -8.24163854e-02
-1.19146466e+00 -7.55657494e-01 -8.06637049e-01 3.43170434e-01
1.71798300e-02 -4.16890383e-01 2.07606888e+00 -1.01364470e+00
-1.66550410e+00 5.87735772e-01 -1.31021008e-01 -5.88444710e-01
3.78503799e-01 -2.92483419e-01 -7.56651521e-01 3.23293686e-01
-1.43126816e-01 1.27937818e+00 8.87424588e-01 -1.19278765e+00
-2.60360241e-01 2.75616378e-01 -1.51623204e-01 1.43092945e-01
-4.32451338e-01 5.94788015e-01 -3.63304466e-01 -7.36496150e-01
-4.83356804e-01 -8.91367137e-01 5.60987405e-02 -3.60650986e-01
-3.93232167e-01 1.18692681e-01 6.31762087e-01 -6.14247501e-01
1.24122405e+00 -2.41470504e+00 -3.54541183e-01 -2.22758889e-01
-5.01686215e-01 1.87369138e-01 -5.32044351e-01 6.28900707e-01
-3.72011513e-01 3.16707551e-01 -4.29433249e-02 -7.80830443e-01
1.52946532e-01 -1.42340556e-01 -7.76274621e-01 -1.08091138e-01
3.43159497e-01 5.90456188e-01 -7.63523340e-01 -5.48476279e-01
-5.76284081e-02 7.85599470e-01 -8.13868940e-01 4.16983366e-01
-3.73403877e-01 3.26058239e-01 -4.30009104e-02 6.84244215e-01
3.01585436e-01 -1.20985478e-01 -2.09371790e-01 3.25014405e-02
3.18058521e-01 7.44285524e-01 -9.70117867e-01 1.67612302e+00
-4.87811387e-01 9.17244375e-01 2.17755258e-01 -6.25449479e-01
8.84481430e-01 1.06903923e+00 5.38809121e-01 -1.99949995e-01
2.43108422e-01 2.75247425e-01 -2.44967759e-01 -5.58293700e-01
3.99276584e-01 -8.30521211e-02 1.91943645e-02 4.12280113e-01
3.63616794e-01 -2.46348053e-01 1.07725337e-01 1.61718845e-01
8.59729648e-01 2.36180332e-02 -2.42827654e-01 2.08364263e-01
2.41920367e-01 -1.52120426e-01 1.29304854e-02 6.45200133e-01
-2.67027497e-01 1.27859521e+00 4.78620917e-01 -1.20103143e-01
-1.12048221e+00 -1.01857841e+00 -1.96663797e-01 1.31244004e+00
-5.50845861e-01 -6.24220848e-01 -9.32202578e-01 -3.18400770e-01
-3.69608432e-01 7.11159825e-01 -6.40831053e-01 -5.74628077e-02
-5.13777375e-01 -6.51243508e-01 1.11308026e+00 5.06986499e-01
7.77501091e-02 -1.38492107e+00 -3.49193513e-02 2.47244373e-01
-5.10485291e-01 -1.29475081e+00 -5.80380857e-01 -1.64048150e-01
-5.60679913e-01 -5.58574140e-01 -9.51888740e-01 -8.84673774e-01
3.27446572e-02 1.09999906e-02 1.11981511e+00 -4.64940637e-01
-1.88990787e-01 6.96036398e-01 -4.72280174e-01 -1.01399469e+00
-7.76462018e-01 4.93690640e-01 -2.06687190e-02 5.38262688e-02
2.34095082e-01 -7.77913094e-01 -1.95186883e-01 1.62059501e-01
-8.26491416e-01 3.03575080e-02 3.87399554e-01 4.30366963e-01
3.40768486e-01 -4.30822849e-01 9.35472131e-01 -5.93409419e-01
9.64931786e-01 -6.45294070e-01 -3.15866798e-01 7.13531375e-02
-7.74814561e-02 -2.15510398e-01 3.54468524e-01 -6.00034356e-01
-9.67116892e-01 1.50015503e-01 -5.59289813e-01 -6.55916393e-01
-5.57708442e-01 3.89311969e-01 -1.07297361e-01 1.83028519e-01
8.13061953e-01 -1.53465355e-02 2.16844659e-02 -6.28717661e-01
3.56750697e-01 1.36645448e+00 7.01934934e-01 -7.02035844e-01
4.84259456e-01 2.40291983e-01 -5.64868808e-01 -9.69201326e-01
-6.87756658e-01 -3.29331249e-01 -1.69997275e-01 -3.16841990e-01
5.46030700e-01 -1.21307695e+00 -4.14483726e-01 2.48940438e-01
-1.30882478e+00 -3.04244846e-01 -3.13138425e-01 5.64480901e-01
-8.04665148e-01 1.06287830e-01 -7.55705953e-01 -8.99418652e-01
-4.62555081e-01 -1.08259606e+00 1.41924608e+00 1.23662539e-01
-4.94500130e-01 -7.42155373e-01 5.73617041e-01 5.44416726e-01
6.26329064e-01 -7.06524327e-02 6.08048022e-01 -1.05178285e+00
-2.65029967e-01 -3.39792401e-01 -1.14445671e-01 3.73260140e-01
-9.50917676e-02 1.45091191e-01 -1.57564151e+00 -4.84047197e-02
-4.51026827e-01 -6.16342545e-01 3.74203295e-01 4.83457118e-01
1.23525298e+00 -3.50281626e-01 -2.26677746e-01 2.96358258e-01
8.40923071e-01 1.60672128e-01 5.69631875e-01 1.44136414e-01
3.54991674e-01 5.29460073e-01 3.65773708e-01 4.80780125e-01
9.52198207e-02 8.18463564e-01 3.57637197e-01 1.22556113e-01
-3.41473877e-01 -5.29691100e-01 4.46418077e-01 7.84009516e-01
2.89572835e-01 -3.95371974e-01 -9.52076733e-01 7.28107333e-01
-1.49875224e+00 -9.61278975e-01 2.95746326e-01 1.94550693e+00
8.60406697e-01 7.97876157e-03 4.66352403e-01 2.24917069e-01
9.25505400e-01 9.23844203e-02 -8.93165618e-02 -5.87949038e-01
1.90928593e-01 3.12211365e-01 -3.57818231e-02 5.86411953e-01
-1.15422225e+00 7.86663711e-01 8.04033566e+00 6.90671742e-01
-1.12647402e+00 9.02176946e-02 6.70901477e-01 -4.16165054e-01
-2.45207816e-01 -5.09522021e-01 -7.89123833e-01 5.54522514e-01
1.68266511e+00 -2.97911137e-01 4.64922279e-01 8.53246748e-01
3.49943429e-01 4.39940780e-01 -1.24810386e+00 1.13515389e+00
2.45829985e-01 -1.32633841e+00 1.55254602e-01 -1.82007939e-01
4.37356263e-01 3.41238290e-01 2.80821979e-01 4.77066845e-01
1.07796915e-01 -1.01119041e+00 8.31136882e-01 -1.11442745e-01
9.09087002e-01 -4.80819702e-01 7.24839211e-01 -2.77852297e-01
-9.68773961e-01 -1.31076857e-01 -1.00558043e-01 1.03007350e-02
4.19539422e-01 1.85239509e-01 -1.41382694e+00 2.92260736e-01
7.28755951e-01 2.82460421e-01 -4.30561095e-01 1.41161788e+00
5.89621477e-02 1.17209077e+00 -5.25354564e-01 -9.30615962e-02
-8.04150850e-02 4.23433155e-01 4.36007351e-01 1.49070418e+00
3.75286341e-01 -3.78829725e-02 -1.69482365e-01 5.73853195e-01
-2.73807734e-01 3.11223507e-01 -3.57095391e-01 -4.13196087e-01
7.17195451e-01 1.08141899e+00 -3.78939599e-01 -1.68581367e-01
-3.08067322e-01 5.51367879e-01 -1.33252963e-01 4.11658913e-01
-1.04946291e+00 -4.45029140e-01 6.63218021e-01 2.81033397e-01
9.73799005e-02 -1.96725614e-02 -3.69402347e-03 -8.93275559e-01
2.56125312e-02 -9.36593533e-01 4.36508775e-01 -1.15759087e+00
-1.08735383e+00 9.48631942e-01 3.92004073e-01 -1.03842270e+00
-4.71694410e-01 -2.49909595e-01 -6.49384439e-01 6.28800631e-01
-1.17695808e+00 -1.01154912e+00 -1.98545918e-01 3.20711970e-01
6.52149618e-01 -2.85496563e-01 1.15389216e+00 6.84216738e-01
-4.30224925e-01 8.32881272e-01 -5.38651347e-02 2.51910120e-01
9.90133643e-01 -7.64920712e-01 6.88799858e-01 3.66023719e-01
5.33848345e-01 2.56023467e-01 1.14709020e+00 -4.53204438e-02
-6.46283865e-01 -9.61929739e-01 8.75884056e-01 -5.54412246e-01
8.23076308e-01 -8.42777789e-01 -7.13727832e-01 8.90359163e-01
4.30850118e-01 -2.33107373e-01 1.16594779e+00 1.32775918e-01
-4.75127310e-01 -1.15994737e-01 -9.20599997e-01 4.51395839e-01
6.17680550e-01 -7.45470762e-01 -4.30987388e-01 5.62967598e-01
1.04938185e+00 -4.69501436e-01 -6.55996263e-01 2.75604725e-01
5.87353528e-01 -6.72817469e-01 9.15707827e-01 -8.06963027e-01
3.56212944e-01 2.88792849e-02 -1.87554792e-01 -1.14798427e+00
6.25934750e-02 -1.13426030e+00 -7.57699534e-02 1.52411747e+00
6.80743575e-01 -1.75268427e-01 9.95873153e-01 5.93614876e-01
-3.98981959e-01 -4.32090282e-01 -1.06614161e+00 -6.39613807e-01
-1.55787960e-01 -6.70453548e-01 6.49824440e-01 6.69611275e-01
1.71321690e-01 6.46019042e-01 -4.61497605e-01 9.81125012e-02
1.80547208e-01 -3.56225908e-01 7.74935484e-01 -1.00727069e+00
-5.41489422e-01 -1.39235795e-01 -3.08198184e-01 -8.17510307e-01
7.38329515e-02 -5.89152694e-01 1.63612559e-01 -1.25733817e+00
-9.28859189e-02 -2.74828881e-01 -9.92050171e-02 6.90450728e-01
7.81166777e-02 5.51140368e-01 2.67549992e-01 7.24463463e-02
-5.06498396e-01 5.73107064e-01 5.13167024e-01 -3.04910213e-01
-3.92656326e-01 1.35468259e-01 -6.83257937e-01 3.62794548e-01
1.00410306e+00 -6.03965223e-01 -4.96586770e-01 -5.69448054e-01
4.45725732e-02 2.04672907e-02 2.26904437e-01 -1.38329136e+00
-1.43943176e-01 3.15551013e-01 7.72686526e-02 -2.23671392e-01
9.64100838e-01 -5.89677393e-01 1.92416087e-01 -3.21964361e-02
-8.25577974e-01 -9.80058238e-02 7.61083543e-01 3.86406988e-01
-4.54133034e-01 -2.45167330e-01 5.89409113e-01 2.58988529e-01
-5.05821072e-02 1.71009883e-01 -3.93838376e-01 1.03846595e-01
6.48238599e-01 9.22430232e-02 -2.68575400e-01 -1.04432940e+00
-1.07874441e+00 -4.42861430e-02 3.16768400e-02 8.60104918e-01
3.57336432e-01 -1.36684227e+00 -9.89575863e-01 1.13656916e-01
1.86919376e-01 -3.46142858e-01 9.31141451e-02 4.57147926e-01
-3.40942532e-01 7.59477675e-01 -1.77329794e-01 -5.26029408e-01
-1.40546632e+00 3.42967182e-01 2.52359182e-01 2.44459152e-01
-1.42429814e-01 1.05465782e+00 -1.19577132e-01 -1.51913479e-01
6.64608419e-01 1.22518558e-02 -2.45558217e-01 -9.95614305e-02
7.35271454e-01 1.91795975e-01 1.49777770e-01 -5.31284750e-01
-1.92334697e-01 1.73310544e-02 -1.58653319e-01 -6.76233649e-01
1.13447523e+00 -4.75352025e-03 3.74223560e-01 5.27909219e-01
1.23055279e+00 1.43028202e-03 -9.33655202e-01 1.66635022e-01
-2.59726018e-01 -1.69782996e-01 -8.91991928e-02 -7.53943563e-01
-6.57954872e-01 1.07512796e+00 6.68183804e-01 5.39754391e-01
9.96812463e-01 2.01125279e-01 7.34919071e-01 1.98717907e-01
1.65765956e-01 -8.40992033e-01 2.19788209e-01 3.60089988e-01
9.34487462e-01 -8.91963601e-01 -5.21801233e-01 -5.29018700e-01
-7.21049309e-01 9.37048316e-01 4.86972272e-01 2.54131436e-01
6.25899851e-01 3.46961528e-01 4.80482817e-01 1.49388045e-01
-9.68752623e-01 1.27065390e-01 3.07515766e-02 8.81760836e-01
7.47046471e-01 -1.59828410e-01 4.19444442e-01 6.92959130e-01
-6.40891790e-01 1.82112679e-01 6.80833757e-01 6.68558598e-01
-2.55250722e-01 -1.30896068e+00 -5.24750054e-01 2.24998146e-01
-7.74997294e-01 -1.63830489e-01 -6.73912585e-01 4.26162064e-01
-3.53980392e-01 1.10205865e+00 2.36541361e-01 -3.86829227e-01
2.38562018e-01 3.71846706e-01 3.42578292e-01 -8.83490503e-01
-4.96367306e-01 2.37003028e-01 4.92688984e-01 -2.42789835e-01
-3.17661345e-01 -6.10069394e-01 -9.61530209e-01 6.35154098e-02
-2.84282953e-01 6.93277061e-01 8.87145042e-01 5.42234898e-01
7.78240621e-01 5.31952977e-01 4.81140018e-01 -1.15395617e+00
-4.19088811e-01 -1.35992610e+00 -2.42632180e-01 4.80408520e-02
2.94671774e-01 -3.53557616e-01 -5.31107485e-01 4.34388548e-01] | [15.285457611083984, 4.852954864501953] |
ced4b667-2576-4099-893f-2d7db6389499 | exploration-by-random-network-distillation | 1810.12894 | null | http://arxiv.org/abs/1810.12894v1 | http://arxiv.org/pdf/1810.12894v1.pdf | Exploration by Random Network Distillation | We introduce an exploration bonus for deep reinforcement learning methods
that is easy to implement and adds minimal overhead to the computation
performed. The bonus is the error of a neural network predicting features of
the observations given by a fixed randomly initialized neural network. We also
introduce a method to flexibly combine intrinsic and extrinsic rewards. We find
that the random network distillation (RND) bonus combined with this increased
flexibility enables significant progress on several hard exploration Atari
games. In particular we establish state of the art performance on Montezuma's
Revenge, a game famously difficult for deep reinforcement learning methods. To
the best of our knowledge, this is the first method that achieves better than
average human performance on this game without using demonstrations or having
access to the underlying state of the game, and occasionally completes the
first level. | ['Yuri Burda', 'Amos Storkey', 'Oleg Klimov', 'Harrison Edwards'] | 2018-10-30 | null | https://openreview.net/forum?id=H1lJJnR5Ym | https://openreview.net/pdf?id=H1lJJnR5Ym | iclr-2019 | ['montezumas-revenge'] | ['playing-games'] | [-2.08648145e-01 5.69039762e-01 -8.51731971e-02 -6.47318875e-03
-5.27494967e-01 -5.17818511e-01 6.15085304e-01 -2.85876095e-01
-9.70297039e-01 1.16491163e+00 -3.83538097e-01 -3.27368677e-01
-1.22134872e-01 -7.87834466e-01 -8.22795391e-01 -6.47918284e-01
-6.11368716e-01 6.59519851e-01 6.06519245e-02 -7.56822467e-01
2.91423202e-01 3.01764071e-01 -1.31841397e+00 -1.97011501e-01
4.21066701e-01 9.08441007e-01 3.25706720e-01 8.74402106e-01
5.51265657e-01 1.16236711e+00 -5.31016171e-01 -2.52891123e-01
7.21253455e-01 -3.82121623e-01 -9.23664331e-01 -2.23407879e-01
-7.11059347e-02 -8.73255849e-01 -6.11560225e-01 1.10061121e+00
3.76225919e-01 4.01225090e-01 2.78951257e-01 -1.18695021e+00
-2.83752263e-01 9.93470132e-01 -3.70175242e-01 3.40105265e-01
6.85341423e-03 8.26672375e-01 1.17997730e+00 -3.31429571e-01
5.78585386e-01 1.03473699e+00 4.07223433e-01 7.91112542e-01
-1.23445988e+00 -6.33721054e-01 -3.38999182e-02 1.21695392e-01
-9.30989325e-01 -2.11793855e-01 6.33344293e-01 -2.12698057e-01
1.03662682e+00 -1.55711368e-01 9.03735459e-01 1.25844181e+00
1.14253171e-01 9.57005024e-01 1.29356360e+00 -3.26627433e-01
7.29310095e-01 -1.77362144e-01 -2.49725208e-01 8.99400771e-01
1.00850835e-01 9.45456803e-01 -3.76641184e-01 7.81429708e-02
9.99263525e-01 -1.94984883e-01 -6.37252778e-02 -7.15130329e-01
-9.33461785e-01 1.01013041e+00 6.87141657e-01 -7.53322467e-02
-5.15965164e-01 8.45377505e-01 2.92898595e-01 6.68483198e-01
-3.08664471e-01 1.15816510e+00 -6.62459671e-01 -7.64257014e-01
-5.97406566e-01 5.29533505e-01 8.83040667e-01 5.13545334e-01
7.00887978e-01 3.87000144e-01 2.60012120e-01 3.63507241e-01
1.98433958e-02 3.01334739e-01 6.73486412e-01 -1.67446041e+00
2.45933443e-01 6.44816086e-02 3.20914537e-01 -5.59571505e-01
-4.57595766e-01 -4.46722150e-01 -3.83805364e-01 1.20118225e+00
5.85353076e-01 -7.40865052e-01 -7.05223143e-01 2.07073522e+00
6.65146206e-03 9.58650839e-03 2.40815237e-01 8.43263865e-01
1.90775365e-01 3.20999652e-01 -3.39712203e-01 1.25975043e-01
9.04646575e-01 -1.08079147e+00 -2.37039298e-01 -4.25712168e-01
5.79360306e-01 -6.98527321e-02 1.12785363e+00 8.19528043e-01
-1.36401606e+00 -2.14400798e-01 -1.39669347e+00 1.91627204e-01
-2.69086301e-01 -2.55867064e-01 1.09613204e+00 6.10243857e-01
-1.14329720e+00 1.23764598e+00 -1.15356386e+00 5.25428429e-02
5.03332078e-01 6.63677514e-01 -3.43485534e-01 4.55036223e-01
-1.32002044e+00 1.23716223e+00 6.71873927e-01 -1.60491824e-01
-1.62088406e+00 -2.78997988e-01 -6.54734075e-01 1.75382957e-01
9.41641748e-01 -3.50486219e-01 1.93397498e+00 -9.92282212e-01
-2.09667850e+00 4.65678543e-01 6.78273976e-01 -9.49242651e-01
6.14423394e-01 -2.64443249e-01 2.54326075e-01 3.82659305e-03
-2.38293976e-01 1.13630843e+00 6.48749292e-01 -8.67872715e-01
-4.04300094e-01 -1.85402170e-01 6.01345718e-01 3.53948176e-01
-1.80566162e-01 -3.90500605e-01 9.99037772e-02 -6.14269912e-01
-3.09900671e-01 -1.15900970e+00 -6.66407347e-01 -1.75783113e-02
-1.70678914e-01 -2.16135412e-01 4.21267122e-01 -2.28527740e-01
7.14236200e-01 -1.89294183e+00 2.70236462e-01 9.69596729e-02
4.20808047e-01 1.78915322e-01 -2.97932684e-01 2.97926784e-01
-5.53880706e-02 -2.40861741e-03 -2.13778447e-02 -8.51243585e-02
3.50431830e-01 4.26644504e-01 -2.47510612e-01 1.74989358e-01
-1.93633407e-01 1.19533134e+00 -1.20716584e+00 8.03153403e-03
3.80466580e-02 -2.39145815e-01 -8.36601794e-01 2.30258405e-01
-4.69187975e-01 1.11626118e-01 -2.63742477e-01 7.68083036e-02
1.73107207e-01 -2.11413011e-01 3.44992459e-01 4.62494791e-01
1.01230487e-01 5.50801933e-01 -1.18029118e+00 1.74752712e+00
-1.74198493e-01 6.86587095e-01 2.11241432e-02 -9.91909623e-01
6.21199846e-01 1.77898481e-01 2.77245402e-01 -5.99509180e-01
3.38952631e-01 2.53746361e-01 4.24904853e-01 -4.47226427e-02
6.73939347e-01 -2.09039301e-01 -1.35989442e-01 8.61107171e-01
4.09234852e-01 -3.67828876e-01 2.13629082e-01 2.43050948e-01
1.33344769e+00 3.12671572e-01 4.24079001e-01 -2.59056449e-01
-2.02339515e-01 6.51174560e-02 3.70031565e-01 1.27054250e+00
-2.60105491e-01 1.16260856e-01 9.06220019e-01 -6.15659118e-01
-1.21225059e+00 -1.10934794e+00 4.83323097e-01 1.13562405e+00
-1.22316509e-01 -4.62575763e-01 -8.88025701e-01 -7.57087827e-01
-1.94739759e-01 6.97943747e-01 -9.43545699e-01 -2.33880341e-01
-4.21674937e-01 -3.98968786e-01 7.84282565e-01 6.30339622e-01
5.99288046e-01 -1.45288384e+00 -1.17245579e+00 1.52911827e-01
3.05502325e-01 -6.18865311e-01 -1.14630833e-01 7.84792900e-01
-9.40980971e-01 -9.66699421e-01 -4.08961922e-01 -4.88453090e-01
1.77740067e-01 -1.71778962e-01 1.10414219e+00 -1.83329023e-02
-2.39976853e-01 4.86922376e-02 -6.39576390e-02 -2.36252055e-01
-2.52881080e-01 1.94655091e-01 2.75267214e-01 -1.06482756e+00
1.56511217e-01 -7.86596835e-01 -5.11660278e-01 2.68775746e-02
-6.00336611e-01 4.65186387e-02 4.85383689e-01 1.27731788e+00
9.18060169e-03 2.07594618e-01 2.45400056e-01 -6.05573475e-01
9.98720229e-01 -4.03526038e-01 -1.06575227e+00 -3.44029605e-01
-7.36722350e-01 5.63783884e-01 6.16906762e-01 -4.93041456e-01
-6.25401556e-01 2.12706234e-02 2.00334247e-02 -3.37147295e-01
-1.77229960e-02 3.95088017e-01 3.59821886e-01 -1.67446300e-01
1.09117162e+00 2.80305743e-02 3.44548225e-01 -2.88553685e-01
5.07655084e-01 1.25322998e-01 6.76699281e-01 -7.11739421e-01
5.75473905e-01 3.33828814e-02 1.78815201e-01 -1.73748240e-01
-6.74084187e-01 2.43855640e-01 -2.83454031e-01 -6.21408522e-02
5.84271073e-01 -6.74910605e-01 -1.44097650e+00 4.66147214e-01
-7.65354216e-01 -1.06154346e+00 -7.42460907e-01 4.45901662e-01
-1.15576923e+00 1.17620289e-01 -8.83236885e-01 -7.38135815e-01
-2.10226839e-03 -1.32983279e+00 3.56223553e-01 2.60050565e-01
-7.28438348e-02 -6.57670021e-01 3.18136334e-01 8.79093707e-02
4.91018683e-01 1.38088390e-01 3.88158232e-01 -4.45656091e-01
-7.65189111e-01 2.12533906e-01 8.55037794e-02 3.57654542e-01
-1.26444161e-01 -4.34465647e-01 -9.23871934e-01 -3.91526669e-01
1.66562200e-02 -1.00963128e+00 9.20405388e-01 2.86721945e-01
1.26369214e+00 -4.03132439e-01 1.81994393e-01 4.21363413e-01
1.21108449e+00 2.18444690e-01 7.39926040e-01 7.47069895e-01
2.33359307e-01 2.42789105e-01 5.68418145e-01 7.20853269e-01
1.52902305e-01 7.13023305e-01 8.69167030e-01 1.79868400e-01
3.32379401e-01 -3.81044805e-01 5.11778891e-01 1.09026112e-01
-2.34201252e-01 2.35055871e-02 -7.94134140e-01 3.85357231e-01
-2.00745702e+00 -1.21969628e+00 6.43169284e-01 2.05891109e+00
1.00000310e+00 5.97107589e-01 3.05384010e-01 -8.21068734e-02
2.39072502e-01 1.05164923e-01 -9.83507752e-01 -6.35809958e-01
2.12878212e-01 6.42286122e-01 6.87601626e-01 8.35087836e-01
-1.02527559e+00 1.25375783e+00 7.55700254e+00 9.15996432e-01
-7.95871258e-01 -5.38278148e-02 5.19518971e-01 -4.43355918e-01
1.62394140e-02 7.34851137e-02 -3.45682561e-01 2.85736948e-01
8.98873210e-01 -1.12069510e-01 1.27540648e+00 1.24774992e+00
-7.27730840e-02 -2.63127089e-01 -1.23433959e+00 8.31874371e-01
-4.36953872e-01 -1.39591193e+00 -6.12507820e-01 5.28544962e-01
6.12342298e-01 5.10130048e-01 2.54297495e-01 7.26541936e-01
1.32610798e+00 -1.35336757e+00 6.24800563e-01 2.54886031e-01
4.30525213e-01 -9.36266005e-01 5.57006955e-01 5.84561884e-01
-4.28920567e-01 -3.21965784e-01 -4.23600048e-01 -6.87374413e-01
-1.86625391e-01 -5.14299236e-02 -1.07945085e+00 -2.12663487e-01
4.10047561e-01 2.79645443e-01 -3.84488374e-01 7.45557725e-01
-5.92580199e-01 6.25539720e-01 -4.75108773e-01 -4.84161794e-01
7.39386797e-01 5.65395430e-02 5.09977877e-01 7.16587901e-01
1.12518646e-01 5.99791370e-02 3.05243760e-01 8.79652023e-01
-2.91074812e-01 -4.57256496e-01 -7.89610922e-01 -1.18601017e-01
2.89650917e-01 1.06579530e+00 -4.98560727e-01 -1.99780583e-01
2.90242553e-01 7.67137527e-01 8.83750439e-01 5.90006225e-02
-7.21249282e-01 -5.02831161e-01 6.55091345e-01 -2.84912527e-01
5.01868308e-01 -3.94599676e-01 -2.22082838e-01 -1.20996761e+00
-1.87553748e-01 -1.09659755e+00 2.93207943e-01 -8.90307963e-01
-8.92531991e-01 4.77037817e-01 -2.15732262e-01 -8.44678462e-01
-8.65435958e-01 -7.48735249e-01 -4.19146359e-01 4.97854441e-01
-1.16458857e+00 -3.10202420e-01 2.66800597e-02 4.67378587e-01
3.82572502e-01 -5.70808113e-01 9.60411191e-01 -3.94082636e-01
-3.22094142e-01 6.90301180e-01 1.48061872e-01 9.35611799e-02
2.08771288e-01 -1.74959755e+00 5.86175382e-01 5.51543951e-01
1.16888650e-01 4.71871853e-01 8.92879367e-01 -3.01286638e-01
-1.22328651e+00 -3.76422226e-01 -2.79801320e-02 -4.11869496e-01
1.02042699e+00 -3.82509202e-01 -4.12859619e-01 8.13369393e-01
4.10532236e-01 -1.18297882e-01 2.62493908e-01 3.98467034e-01
-3.19557369e-01 1.24486566e-01 -1.09978628e+00 9.76181030e-01
8.95800412e-01 -4.08869654e-01 -8.99859428e-01 2.27867305e-01
6.44424558e-01 -7.18218982e-01 -6.22715533e-01 2.86452863e-02
5.86323082e-01 -9.94310975e-01 9.44300294e-01 -8.95302653e-01
7.27449656e-01 -4.74215001e-02 -1.75844669e-01 -1.58176589e+00
-3.65868956e-01 -1.08724952e+00 -3.71938586e-01 3.33079785e-01
4.89005595e-01 -5.74137092e-01 1.34674811e+00 6.65549099e-01
8.44562501e-02 -8.30381155e-01 -8.66155803e-01 -9.42559838e-01
3.92577946e-01 -3.01608086e-01 5.20735323e-01 8.05996358e-01
5.64653277e-01 1.64103791e-01 -6.64852262e-01 -2.08345711e-01
7.08977938e-01 -6.86063170e-02 8.35439324e-01 -9.12612557e-01
-1.00345123e+00 -5.99475205e-01 -3.42978835e-01 -1.19822383e+00
6.81044385e-02 -6.87006056e-01 2.05865204e-01 -8.56877029e-01
1.11282997e-01 -4.53570694e-01 -5.16617894e-01 8.21318328e-01
1.96107611e-01 4.62475047e-02 4.51217055e-01 -7.70940483e-02
-9.08845961e-01 7.65009999e-01 1.29923582e+00 3.65858153e-02
-2.49225348e-01 -1.10092700e-01 -9.44520831e-01 8.99370611e-01
1.37931252e+00 -5.99035800e-01 -5.49847126e-01 -3.21112722e-01
4.39638644e-01 2.29597256e-01 5.28904796e-01 -1.02324152e+00
2.35748693e-01 -4.09032792e-01 2.63428092e-01 -5.40059134e-02
5.39967895e-01 -3.94602150e-01 -2.69577831e-01 8.90946686e-01
-7.09383011e-01 1.75818279e-01 2.23256990e-01 5.23500144e-01
1.50946170e-01 -5.03587902e-01 9.45706367e-01 -5.91124415e-01
-6.79070771e-01 1.27589330e-01 -7.70067573e-01 3.12073499e-01
9.80524123e-01 -2.47282758e-02 -3.85014474e-01 -6.78550720e-01
-8.43390226e-01 3.11270446e-01 5.21106303e-01 1.37115225e-01
4.22076285e-01 -1.10310674e+00 -3.54093045e-01 -3.33979949e-02
-3.39379579e-01 -2.67342031e-01 -9.49875563e-02 3.08622777e-01
-5.21908998e-01 2.13715047e-01 -7.89379895e-01 -9.59324464e-02
-8.90455186e-01 6.22869968e-01 7.15016782e-01 -7.43969262e-01
-6.64638758e-01 8.37291956e-01 -4.55889367e-02 -5.63663363e-01
2.26776049e-01 -2.00122476e-01 9.24398676e-02 -4.68078285e-01
4.55235153e-01 3.20225358e-01 -1.48589104e-01 2.44345650e-01
-2.50137784e-02 -2.31580526e-01 -3.85781646e-01 -7.90962219e-01
1.34657836e+00 3.77641290e-01 2.29945794e-01 1.69639707e-01
7.17165887e-01 -4.14227068e-01 -1.80286002e+00 3.42131741e-02
-2.90984929e-01 -3.18854481e-01 3.13417941e-01 -9.71214294e-01
-1.03603697e+00 9.22967255e-01 5.65441370e-01 1.63366273e-01
7.96868622e-01 -3.29094887e-01 7.06960678e-01 1.26604819e+00
7.47943103e-01 -1.36292994e+00 5.26760757e-01 7.91758299e-01
7.72450149e-01 -1.20693684e+00 -7.18910992e-02 5.66822648e-01
-8.03812265e-01 1.06900191e+00 9.23616230e-01 -7.07840204e-01
4.83339727e-01 3.74077469e-01 -2.87080079e-01 -2.87427992e-01
-1.13947392e+00 -1.83240727e-01 -4.42690402e-01 6.16066933e-01
-2.72683442e-01 2.38027439e-01 -3.96903269e-02 4.21983302e-01
-6.84789062e-01 2.05953434e-01 8.26983452e-01 9.25260365e-01
-7.30379641e-01 -1.20698297e+00 -8.79082829e-03 3.46591711e-01
-4.85655606e-01 -1.27062052e-01 -1.64573535e-01 8.78559887e-01
-2.42568389e-01 6.53997123e-01 -1.95238829e-01 -4.06537592e-01
-3.95883769e-02 -2.73176312e-01 7.53474355e-01 -3.87186527e-01
-6.60240233e-01 -3.59624207e-01 3.42155397e-02 -8.79507899e-01
1.63480878e-01 -5.81789434e-01 -1.23434591e+00 -6.15343809e-01
-1.71903610e-01 2.60212898e-01 6.04032338e-01 7.98441887e-01
2.74706393e-01 3.62582594e-01 6.03769720e-01 -1.06612229e+00
-1.14692628e+00 -7.36091733e-01 -7.09746957e-01 4.81813550e-02
4.45811629e-01 -8.54984581e-01 -3.20752323e-01 -5.21814883e-01] | [3.838038682937622, 1.6538687944412231] |
5dd85be7-19b8-4dc3-83b4-a643100e8ef4 | m-text-kg-a-library-for-multi-source | 2207.11442 | null | https://arxiv.org/abs/2207.11442v2 | https://arxiv.org/pdf/2207.11442v2.pdf | $μ\text{KG}$: A Library for Multi-source Knowledge Graph Embeddings and Applications | This paper presents $\mu\text{KG}$, an open-source Python library for representation learning over knowledge graphs. $\mu\text{KG}$ supports joint representation learning over multi-source knowledge graphs (and also a single knowledge graph), multiple deep learning libraries (PyTorch and TensorFlow2), multiple embedding tasks (link prediction, entity alignment, entity typing, and multi-source link prediction), and multiple parallel computing modes (multi-process and multi-GPU computing). It currently implements 26 popular knowledge graph embedding models and supports 16 benchmark datasets. $\mu\text{KG}$ provides advanced implementations of embedding techniques with simplified pipelines of different tasks. It also comes with high-quality documentation for ease of use. $\mu\text{KG}$ is more comprehensive than existing knowledge graph embedding libraries. It is useful for a thorough comparison and analysis of various embedding models and tasks. We show that the jointly learned embeddings can greatly help knowledge-powered downstream tasks, such as multi-hop knowledge graph question answering. We will stay abreast of the latest developments in the related fields and incorporate them into $\mu\text{KG}$. | ['Wei Hu', 'Zequn Sun', 'Xindi Luo'] | 2022-07-23 | null | null | null | null | ['graph-question-answering', 'knowledge-graph-embeddings', 'knowledge-graph-embeddings', 'entity-typing'] | ['graphs', 'graphs', 'methodology', 'natural-language-processing'] | [-3.87195796e-01 3.91347796e-01 -5.24493456e-01 -1.31194547e-01
-6.07919157e-01 -5.70943236e-01 8.32389966e-02 8.98209691e-01
-3.38516682e-01 6.15231514e-01 7.42693394e-02 -6.19177341e-01
-7.81950057e-01 -1.38360786e+00 -5.80862999e-01 -3.27912867e-01
-6.42262578e-01 6.66456342e-01 4.11600798e-01 -2.88080841e-01
4.50720489e-02 1.44680724e-01 -1.49830461e+00 2.33264640e-01
3.87840271e-01 8.02612603e-01 -8.26470777e-02 9.54878807e-01
-3.36649895e-01 7.98654020e-01 -2.41482392e-01 -1.16183817e+00
-1.14161544e-01 2.95249015e-01 -1.15692127e+00 -9.11335766e-01
4.42719698e-01 -1.66019931e-01 -8.70536864e-01 9.84217703e-01
6.14263892e-01 1.91987053e-01 4.37903523e-01 -1.45176196e+00
-1.14082551e+00 7.08085835e-01 -1.86070651e-01 6.29966259e-01
5.28687775e-01 1.66154969e-02 1.76779377e+00 -7.64871478e-01
8.66821289e-01 8.37266564e-01 8.92452061e-01 2.89223026e-02
-7.67798722e-01 -5.26762545e-01 1.70299932e-02 8.62143517e-01
-1.55695570e+00 -4.95945960e-02 5.48253894e-01 -5.31694889e-01
1.79290998e+00 2.13820025e-01 6.26965225e-01 7.09477961e-01
-7.01468587e-02 7.81430483e-01 2.84931868e-01 -4.79489446e-01
-1.95895374e-01 -3.29315588e-02 7.75466561e-01 1.47887373e+00
6.73809767e-01 -1.39724791e-01 -6.45094931e-01 -6.79049611e-01
5.94152987e-01 -9.18117166e-02 -1.54948905e-01 -3.91261965e-01
-1.30639470e+00 8.77742052e-01 4.90513682e-01 2.70866066e-01
-8.49165246e-02 5.88168979e-01 6.30374968e-01 4.70521867e-01
2.03565344e-01 1.94661945e-01 -5.95996499e-01 -3.69154513e-01
-4.28545892e-01 1.70325860e-01 1.27839565e+00 1.40008712e+00
1.24691439e+00 2.74217166e-02 -4.90913764e-02 8.95491958e-01
2.80438870e-01 1.52253389e-01 1.32771775e-01 -8.00316155e-01
5.36421478e-01 9.31182504e-01 -1.95730880e-01 -1.15299261e+00
-3.99090856e-01 -2.07979739e-01 -5.23852587e-01 -3.31327975e-01
1.96258098e-01 -1.08261071e-01 -5.19836724e-01 1.42053056e+00
4.59266603e-01 2.13191122e-01 1.09611064e-01 5.49392998e-01
1.56723106e+00 3.82804513e-01 2.16843680e-01 2.93470472e-01
1.64513659e+00 -1.22085917e+00 -5.69622338e-01 -5.87249361e-02
1.17985177e+00 -6.81638122e-01 6.30261481e-01 -1.46409363e-01
-1.01670361e+00 -2.62229711e-01 -8.99226069e-01 -5.02206922e-01
-1.24701214e+00 -8.05051476e-02 1.38896167e+00 5.36754787e-01
-1.31183338e+00 6.98707759e-01 -8.38628113e-01 -7.18301237e-01
2.82359451e-01 5.12967229e-01 -5.78346491e-01 -3.44578862e-01
-1.58462548e+00 8.38089645e-01 7.72061110e-01 -1.77624062e-01
-4.11530167e-01 -7.78120637e-01 -1.21558237e+00 1.09740555e-01
4.11603659e-01 -1.10545099e+00 5.85246146e-01 -1.42002910e-01
-8.61710668e-01 9.37801421e-01 6.58467561e-02 -9.69514847e-02
-1.57658055e-01 -2.15682704e-02 -7.99051940e-01 1.61253348e-01
-3.34216952e-02 2.96899229e-01 3.76864910e-01 -6.40927553e-01
-3.58266264e-01 -4.63188529e-01 2.62160629e-01 7.60838091e-02
-6.97107077e-01 1.03731342e-02 -1.01026726e+00 -4.01383042e-01
-2.93361694e-01 -7.51470149e-01 2.01239139e-01 -9.22527909e-02
-5.45683920e-01 -4.60426837e-01 6.81085467e-01 -7.79611051e-01
1.49908125e+00 -1.86886859e+00 2.37642199e-01 3.55774164e-01
6.49360001e-01 3.67515445e-01 -4.61820960e-02 9.56935227e-01
-1.81947380e-01 4.33377437e-02 1.67905074e-03 -7.54036978e-02
4.09372270e-01 4.70022559e-01 8.91188160e-02 2.50779778e-01
-4.07860428e-02 1.44754744e+00 -1.08509481e+00 -5.43856621e-01
1.20275483e-01 6.31132424e-01 -4.35762852e-01 -8.78107548e-02
-3.43741924e-01 -3.91951412e-01 -5.26914477e-01 9.83069956e-01
3.09805572e-01 -8.35083902e-01 5.11444271e-01 -4.32487607e-01
1.99453592e-01 2.45049801e-02 -1.27694929e+00 1.68629205e+00
-4.71268333e-02 4.41303849e-01 1.21405549e-01 -9.58198071e-01
7.42058396e-01 2.28026643e-01 5.66225350e-01 -1.57566294e-01
-7.92359635e-02 1.04548246e-01 -1.55959532e-01 -6.34353399e-01
7.59991229e-01 4.56350237e-01 1.26078546e-01 6.90157235e-01
5.52953184e-01 1.70700043e-01 3.86155844e-01 7.81711519e-01
1.67307556e+00 7.42276758e-02 1.96675956e-01 -1.69471398e-01
2.01749638e-01 1.63429186e-01 1.82214528e-01 6.43977702e-01
-1.47105336e-01 -2.25960210e-01 6.16491556e-01 -5.65612197e-01
-9.10239220e-01 -1.20779097e+00 -5.22745885e-02 1.66771519e+00
3.82079482e-02 -1.19159722e+00 -1.67142987e-01 -6.73394740e-01
6.39015198e-01 4.87874836e-01 -6.43446267e-01 -1.93015680e-01
-2.64987797e-01 -8.42329204e-01 1.04659390e+00 8.60379875e-01
2.27695823e-01 -9.66384530e-01 3.18082660e-01 9.62511078e-02
-6.19375482e-02 -9.43241417e-01 -1.46837071e-01 1.81964532e-01
-6.07259750e-01 -1.51705015e+00 -2.57290810e-01 -1.08937347e+00
3.84844363e-01 8.91763791e-02 1.32825410e+00 5.75910389e-01
-7.30454087e-01 1.12612498e+00 -4.62153494e-01 -1.93112716e-01
7.18742535e-02 2.18711734e-01 2.90696211e-02 -5.91527462e-01
8.67057562e-01 -7.52115130e-01 -5.69293261e-01 7.69857392e-02
-9.11130071e-01 -3.81755769e-01 3.00822556e-01 7.69108653e-01
7.76102126e-01 1.90230399e-01 4.76213783e-01 -9.85273659e-01
7.08530307e-01 -7.59106934e-01 -4.37391579e-01 4.18866903e-01
-8.48063767e-01 1.92208305e-01 3.77515286e-01 7.39197358e-02
-4.46516007e-01 -4.44507807e-01 -3.85857910e-01 -4.60719138e-01
2.05010489e-01 8.38505805e-01 2.41167456e-01 -2.51598626e-01
6.27122343e-01 1.34120017e-01 -9.73013118e-02 -6.07879579e-01
8.18676949e-01 3.64953190e-01 3.80241722e-01 -7.08973825e-01
7.05998838e-01 2.57161409e-01 -1.43980533e-01 -7.73083210e-01
-4.81799364e-01 -7.10891366e-01 -6.00648224e-01 1.88015878e-01
7.56959021e-01 -8.62713337e-01 -1.05403352e+00 2.01285947e-02
-9.39468741e-01 -3.99044067e-01 -1.26314104e-01 4.94169265e-01
-4.28500980e-01 6.92458808e-01 -8.88417721e-01 -3.03387016e-01
-7.08603501e-01 -8.30222666e-01 8.14296007e-01 6.11294471e-02
-8.46046805e-02 -1.71522820e+00 2.16120645e-01 6.98822916e-01
3.89650017e-01 1.05997041e-01 1.38478363e+00 -9.67800617e-01
-7.62374640e-01 -4.60495919e-01 -5.08750737e-01 1.10240564e-01
-7.52815306e-02 1.55471653e-01 -7.48332381e-01 -3.56822014e-01
-1.16427958e+00 -4.77235436e-01 1.03021896e+00 -1.44954458e-01
1.24085987e+00 -2.25560203e-01 -7.91794181e-01 6.05817258e-01
1.50402117e+00 -5.33607781e-01 5.22631764e-01 3.84389937e-01
1.08352673e+00 3.50201458e-01 1.92037851e-01 2.76246846e-01
1.08506966e+00 5.06548405e-01 4.19110209e-01 4.45441693e-01
-2.84181058e-01 -2.04566792e-01 1.70545474e-01 1.15251386e+00
-3.82088691e-01 -1.12412952e-01 -1.10465229e+00 8.92495751e-01
-1.99326575e+00 -1.13469076e+00 -4.43341523e-01 1.95737195e+00
7.52088368e-01 -2.45972201e-01 7.61621147e-02 -3.85821462e-02
6.88499987e-01 1.90516740e-01 -5.34625292e-01 -5.69850385e-01
-1.26962453e-01 6.65869474e-01 6.75748169e-01 6.97482407e-01
-1.02054203e+00 1.17666030e+00 5.97549057e+00 7.76301146e-01
-5.26424885e-01 3.43085051e-01 -1.45236552e-01 5.73421903e-02
-6.31934762e-01 1.36962101e-01 -9.50175762e-01 1.61307663e-01
8.87445211e-01 -3.74985456e-01 4.36061174e-01 9.85599101e-01
-7.76495695e-01 1.07915498e-01 -1.12582266e+00 1.06584954e+00
2.40279734e-02 -1.91897476e+00 -2.93207973e-01 -1.03865057e-01
4.34010863e-01 6.20437741e-01 -2.25127771e-01 7.49936521e-01
7.89771080e-01 -8.97899866e-01 -1.51172861e-01 4.55310911e-01
9.83602345e-01 -5.42979240e-01 6.60568357e-01 -1.15478359e-01
-1.70874989e+00 -4.78664190e-02 -4.39663142e-01 6.04476072e-02
6.82051331e-02 7.57449508e-01 -8.00080836e-01 1.24748373e+00
8.92639875e-01 8.53448987e-01 -6.46966636e-01 6.95366859e-01
-3.42983872e-01 4.35088098e-01 -3.34204137e-01 -2.29670063e-01
-1.98084757e-01 -9.14139487e-03 4.82239515e-01 1.61854339e+00
8.46844763e-02 3.85027900e-02 2.20311750e-02 6.00550830e-01
-3.25925797e-01 2.51413167e-01 -8.09653938e-01 -4.78829861e-01
5.45669079e-01 1.28804493e+00 -3.66300881e-01 -3.95170182e-01
-7.00904667e-01 9.77374792e-01 8.31008315e-01 3.86964709e-01
-6.35298729e-01 -1.12140882e+00 1.07454276e+00 -1.31433770e-01
4.39309537e-01 -4.90934283e-01 1.75295010e-01 -1.35096645e+00
9.68745071e-03 -2.41741955e-01 1.23000562e+00 -7.36239672e-01
-1.60271287e+00 2.96343654e-01 -6.03850633e-02 -4.53872323e-01
5.52830473e-02 -8.68395388e-01 -3.45329851e-01 8.12899888e-01
-1.76333070e+00 -1.25099051e+00 -3.65264326e-01 1.03078830e+00
-3.83316278e-01 -2.60972172e-01 1.35380447e+00 6.11586332e-01
-7.08721220e-01 8.24634075e-01 2.15758339e-01 6.69587433e-01
6.05223358e-01 -1.33219314e+00 3.78126681e-01 1.74039394e-01
2.20611751e-01 7.59746134e-01 1.54070958e-01 -7.78227866e-01
-2.16033530e+00 -1.32405102e+00 9.75614071e-01 -8.15135360e-01
1.39056146e+00 -1.92504779e-01 -9.70728397e-01 1.34951389e+00
9.83053818e-02 3.72093976e-01 1.29004407e+00 6.63634956e-01
-5.92005789e-01 -3.43076736e-02 -1.01231647e+00 3.47546786e-01
1.22138298e+00 -7.88340509e-01 -5.82564056e-01 7.27084994e-01
9.28388000e-01 -3.25010747e-01 -1.92263269e+00 2.07992032e-01
3.83067012e-01 -6.15158081e-01 1.16238582e+00 -9.36988890e-01
-5.04949782e-03 -1.53301090e-01 -2.76509106e-01 -8.35387766e-01
-6.34028077e-01 -5.15395164e-01 -1.04090691e+00 9.13008451e-01
4.63966012e-01 -9.61433232e-01 9.66681838e-01 4.58946347e-01
-1.24815315e-01 -8.69115174e-01 -8.38423550e-01 -7.62362123e-01
-8.14266875e-02 -5.11854053e-01 4.21948403e-01 1.43376565e+00
4.12892848e-01 2.89112419e-01 2.28483379e-02 4.73895311e-01
7.11312711e-01 2.23820776e-01 7.68453658e-01 -1.02711856e+00
-4.80290234e-01 -3.59720707e-01 -9.96984065e-01 -7.28266895e-01
2.73501664e-01 -1.60016811e+00 -1.02364826e+00 -2.16650391e+00
6.91955090e-02 -5.66549301e-01 -4.90112066e-01 1.15147460e+00
-1.08852662e-01 2.62659341e-01 -2.65726984e-01 4.82872725e-02
-8.69116366e-01 3.92832875e-01 8.91671479e-01 -2.50120848e-01
1.62897751e-01 -4.36861724e-01 -7.55333841e-01 3.55107129e-01
6.55137122e-01 -2.81686097e-01 -2.83994377e-01 -6.72238350e-01
8.39590669e-01 -2.28564873e-01 4.65091050e-01 -6.18943870e-01
5.94339848e-01 7.90760443e-02 1.50650114e-01 -2.46817082e-01
5.67019522e-01 -4.68304962e-01 -4.02011350e-02 2.21908107e-01
1.39330074e-01 2.08894119e-01 4.14590299e-01 8.33255827e-01
-1.36460423e-01 -2.14963377e-01 2.76441813e-01 -1.66749194e-01
-1.21136951e+00 6.64437532e-01 2.31840029e-01 1.36094570e-01
1.17499864e+00 -1.56030416e-01 -8.35211813e-01 -2.35833794e-01
-1.15224683e+00 6.54673636e-01 2.40503281e-01 5.39912522e-01
5.34933448e-01 -1.38429475e+00 -5.15134096e-01 -9.30842012e-02
6.16362095e-01 -1.03102453e-01 2.95953244e-01 9.63287175e-01
-7.05842197e-01 4.81920391e-01 2.33308122e-01 -4.23555262e-02
-1.07951498e+00 6.86373532e-01 3.42197977e-02 -2.81072140e-01
-5.21903396e-01 1.11031783e+00 -8.12563539e-01 -8.06894839e-01
1.55568346e-01 1.27787918e-01 -2.46733055e-02 5.59362508e-02
4.34812039e-01 8.10514271e-01 2.82621622e-01 -1.98017538e-01
-5.99320292e-01 4.10954475e-01 -3.63512374e-02 5.44221044e-01
1.41433215e+00 1.60652474e-01 -7.15591192e-01 3.41783822e-01
1.46682024e+00 1.65839046e-02 -3.07433218e-01 -3.96133065e-01
-1.05341114e-01 -1.65758118e-01 -1.00390539e-01 -7.61232316e-01
-9.05894935e-01 6.83604360e-01 2.10643396e-01 3.28906745e-01
5.52565277e-01 4.05486077e-01 8.09352994e-01 7.88107693e-01
4.89153862e-01 -1.16170371e+00 6.33468106e-02 6.72609627e-01
3.39292496e-01 -1.15777409e+00 4.04296905e-01 -4.38303202e-01
-4.15560246e-01 1.21807289e+00 5.73534846e-01 1.45164862e-01
1.09410477e+00 6.27689287e-02 -3.13676804e-01 -8.30935419e-01
-8.13204050e-01 -4.55412388e-01 1.46019906e-01 8.32652807e-01
4.44404423e-01 1.72140121e-01 -9.47793052e-02 6.99931800e-01
-2.51771957e-01 -1.38960332e-01 2.26133004e-01 1.14276087e+00
-3.06590140e-01 -1.42164397e+00 7.16740191e-02 8.86456907e-01
-2.29071289e-01 -3.84694159e-01 -1.81797251e-01 8.91155601e-01
1.14097364e-01 7.61238694e-01 -1.11987688e-01 -5.41485071e-01
1.43561706e-01 4.53821152e-01 5.54573953e-01 -7.82584310e-01
-5.23093522e-01 -7.99245179e-01 5.29829204e-01 -5.51883996e-01
-5.87273873e-02 -2.67499417e-01 -1.36575413e+00 -7.72991598e-01
-2.16212362e-01 2.91532546e-01 5.85084260e-01 4.28310931e-01
9.25610542e-01 4.11870211e-01 -2.04526424e-01 -2.66777396e-01
-5.29681891e-02 -7.73803830e-01 -7.26313055e-01 2.15516284e-01
-3.72180372e-01 -7.02988625e-01 -2.09344447e-01 -2.37147629e-01] | [8.777631759643555, 7.902659893035889] |
817e28ca-35f7-4977-bb6a-52c17ec89edd | broad-context-language-modeling-as-reading | 1610.08431 | null | http://arxiv.org/abs/1610.08431v3 | http://arxiv.org/pdf/1610.08431v3.pdf | Broad Context Language Modeling as Reading Comprehension | Progress in text understanding has been driven by large datasets that test
particular capabilities, like recent datasets for reading comprehension
(Hermann et al., 2015). We focus here on the LAMBADA dataset (Paperno et al.,
2016), a word prediction task requiring broader context than the immediate
sentence. We view LAMBADA as a reading comprehension problem and apply
comprehension models based on neural networks. Though these models are
constrained to choose a word from the context, they improve the state of the
art on LAMBADA from 7.3% to 49%. We analyze 100 instances, finding that neural
network readers perform well in cases that involve selecting a name from the
context based on dialogue or discourse cues but struggle when coreference
resolution or external knowledge is needed. | ['David Mcallester', 'Zewei Chu', 'Kevin Gimpel', 'Hai Wang'] | 2016-10-26 | broad-context-language-modeling-as-reading-1 | https://aclanthology.org/E17-2009 | https://aclanthology.org/E17-2009.pdf | eacl-2017-4 | ['lambada'] | ['natural-language-processing'] | [ 6.62697017e-01 6.60518765e-01 -1.89678699e-01 -4.31750238e-01
-9.07327712e-01 -7.41033554e-01 8.68089974e-01 6.28515959e-01
-6.96235001e-01 5.33158123e-01 8.70025814e-01 -8.23945999e-01
-2.31301948e-01 -7.75009930e-01 -5.33045530e-01 -1.24069475e-01
3.34872514e-01 7.94757843e-01 1.72955304e-01 -6.52297616e-01
6.89151466e-01 -1.57250345e-01 -1.37627399e+00 6.39816403e-01
8.98979247e-01 6.92204654e-01 2.68195957e-01 7.97597170e-01
-2.69656003e-01 8.63411129e-01 -7.37031102e-01 -3.56686622e-01
-3.68523866e-01 -4.85791266e-01 -1.64397371e+00 -6.67140603e-01
7.89912701e-01 -4.65864092e-01 2.98013538e-02 5.53111970e-01
6.35505438e-01 3.31308305e-01 6.55759513e-01 -5.77205658e-01
-6.84113562e-01 1.21671963e+00 -7.98719078e-02 5.51425278e-01
1.09773397e+00 2.70678978e-02 1.40774536e+00 -6.65993869e-01
6.76403821e-01 1.31456673e+00 6.54980421e-01 7.30154097e-01
-1.27817011e+00 -2.56720960e-01 4.49954867e-01 7.29654431e-01
-6.74464703e-01 -6.87147081e-01 4.61137056e-01 -2.09221140e-01
1.68771660e+00 3.04646432e-01 3.03482264e-01 1.62585342e+00
9.12577137e-02 8.19573402e-01 1.03537905e+00 -9.58992541e-01
1.15966924e-01 -6.74025536e-01 8.92281711e-01 2.85370022e-01
1.09768228e-03 -9.85865220e-02 -9.08037961e-01 -1.36962339e-01
-3.08905169e-03 -5.80130279e-01 -6.86147630e-01 1.76186636e-01
-1.27238059e+00 9.40616667e-01 2.80725330e-01 2.32672304e-01
-3.09510320e-01 -3.00954640e-01 3.31124932e-01 5.41013002e-01
2.03432664e-01 8.93226385e-01 -8.15890431e-01 -4.27357048e-01
-6.53794944e-01 4.51516539e-01 1.47756815e+00 5.67922235e-01
2.38974482e-01 -5.35484135e-01 -4.26246554e-01 9.18672860e-01
1.44014418e-01 4.61843908e-01 5.35045147e-01 -9.93036807e-01
8.98847282e-01 3.10602963e-01 1.04836430e-02 -9.57897186e-01
-8.19368780e-01 -1.01853140e-01 -3.88961494e-01 -2.65248954e-01
9.77098465e-01 -1.57001689e-01 -6.33504808e-01 2.00147009e+00
-1.07953340e-01 -3.03503722e-01 3.33632439e-01 6.28117085e-01
1.17780578e+00 5.48171282e-01 4.92930084e-01 -2.15673804e-01
1.34380746e+00 -9.09394503e-01 -7.38482416e-01 -8.22276473e-01
9.92224574e-01 -5.83818316e-01 1.17481422e+00 4.89660561e-01
-1.31670308e+00 -2.90208250e-01 -9.69220042e-01 -4.13884193e-01
-3.21788013e-01 -2.91560680e-01 3.23935002e-01 1.66960448e-01
-1.08919013e+00 3.74713778e-01 -4.32866722e-01 -6.45330071e-01
2.43266284e-01 1.46597043e-01 -2.46805623e-01 -2.58329839e-01
-1.61147130e+00 1.55576456e+00 5.46213448e-01 -9.63918194e-02
-5.92279851e-01 -6.73047900e-01 -8.00715923e-01 3.76871496e-01
6.73464835e-01 -8.12266707e-01 1.69588637e+00 -1.06566429e+00
-1.44162643e+00 1.12533653e+00 -3.23635519e-01 -7.55889475e-01
2.45909914e-01 -7.06630349e-01 -3.44975889e-01 3.01334262e-01
1.11469544e-01 5.98396361e-01 5.94473541e-01 -7.58577704e-01
-4.49856371e-01 -3.63956839e-01 2.96227187e-01 3.47807348e-01
1.74892694e-01 2.13289455e-01 2.12462679e-01 -4.12584990e-01
4.81837913e-02 -6.06527150e-01 3.76970917e-01 -5.24568856e-01
-5.34840107e-01 -8.93198133e-01 4.00379986e-01 -9.69293058e-01
1.29063451e+00 -1.59452057e+00 3.99132401e-01 -2.08232984e-01
3.01701903e-01 2.25554720e-01 -2.68595517e-01 8.16862643e-01
-2.73879319e-01 4.15218413e-01 -2.51628309e-01 -8.35654289e-02
1.43122017e-01 -9.40701291e-02 -5.39105654e-01 -2.76086181e-02
2.09921747e-01 9.58558321e-01 -7.32259989e-01 -1.16072237e-01
-4.63346951e-02 4.06123437e-02 -4.58100110e-01 3.12979221e-01
-6.25419319e-01 -3.11114769e-02 -2.92932659e-01 3.25914681e-01
3.59228671e-01 -2.77547300e-01 9.27174389e-02 4.01141644e-02
5.09932339e-02 1.15039921e+00 -5.59498608e-01 1.61666262e+00
-3.29503775e-01 8.41863990e-01 2.69367062e-02 -9.25018907e-01
6.62195325e-01 2.80411273e-01 -4.15560067e-01 -8.28173280e-01
3.31784040e-01 9.00266990e-02 5.53142548e-01 -7.87051320e-01
3.03745151e-01 -8.79649296e-02 -9.83594637e-03 8.08430672e-01
-1.88559871e-02 1.05683366e-02 1.29520237e-01 4.29822326e-01
1.12764525e+00 1.68897048e-01 5.79560101e-01 -3.65520895e-01
5.39270043e-01 1.77323759e-01 1.55546200e-02 1.18879306e+00
-1.76736847e-01 6.31639004e-01 7.33272731e-01 -2.60042638e-01
-7.83804715e-01 -9.06709075e-01 -1.50340259e-01 1.67686510e+00
-2.32200876e-01 -5.61194003e-01 -1.02104199e+00 -6.64761484e-01
-2.07179010e-01 1.53964317e+00 -8.45688581e-01 -3.34418744e-01
-7.46997893e-01 -3.74826461e-01 6.67496443e-01 6.68977976e-01
5.30536652e-01 -1.37484896e+00 -8.99056852e-01 2.75948346e-01
-5.47420621e-01 -8.61621141e-01 -1.17464602e-01 2.63407946e-01
-6.11504555e-01 -1.27910101e+00 -3.40036482e-01 -6.84385240e-01
5.89198135e-02 4.74320129e-02 1.68550837e+00 3.86325002e-01
9.90737230e-02 6.35221124e-01 -6.40720069e-01 -5.57931840e-01
-5.00691831e-01 6.88699424e-01 -3.70207250e-01 -6.88112557e-01
8.44154179e-01 -2.60016888e-01 -3.36103499e-01 2.83182207e-02
-5.76668918e-01 1.87968954e-01 2.30920434e-01 1.24792802e+00
9.53426119e-03 -9.12899852e-01 1.05248082e+00 -1.02445889e+00
1.16154444e+00 -8.68177712e-01 -2.61070654e-02 3.93305153e-01
-6.10441804e-01 -1.53768733e-02 4.41818684e-01 -3.00019354e-01
-1.20266056e+00 -5.62696278e-01 -4.26477373e-01 4.70024258e-01
-6.49290204e-01 8.69792163e-01 -1.12390313e-02 5.53265810e-01
9.59947586e-01 9.76613015e-02 8.00100192e-02 -5.08073270e-01
2.89003462e-01 4.76631671e-01 5.33826590e-01 -9.16957676e-01
2.53538132e-01 -3.10302973e-01 -4.92284447e-01 -5.87614655e-01
-1.54501951e+00 -1.35659829e-01 -7.29924202e-01 1.18328497e-01
1.14929104e+00 -6.54912591e-01 -8.28867614e-01 2.56848782e-01
-1.42067349e+00 -6.15955412e-01 7.90058449e-02 1.71333522e-01
-6.02853239e-01 1.29606903e-01 -5.52221715e-01 -5.13354838e-01
-4.45327669e-01 -7.80795336e-01 7.79752970e-01 2.47034654e-01
-1.00403285e+00 -1.14262712e+00 -1.02825209e-01 6.03689849e-01
6.36246622e-01 -5.68061657e-02 1.44452786e+00 -1.47024775e+00
-1.73428684e-01 7.29587376e-02 -8.78380239e-02 -8.80826339e-02
-2.54640132e-01 -3.29720289e-01 -1.03535366e+00 -9.35467482e-02
3.54427993e-02 -8.68616045e-01 1.10436440e+00 1.47442192e-01
1.13711023e+00 -2.48717710e-01 -2.40973160e-01 1.74605802e-01
8.81845295e-01 6.84180409e-02 4.80165780e-01 6.81375444e-01
3.59461397e-01 1.03177476e+00 3.30752850e-01 -1.15145937e-01
7.20138073e-01 3.29890221e-01 3.59774768e-01 5.66556275e-01
-1.28024727e-01 -4.06088501e-01 1.96373165e-01 5.76840758e-01
7.80374929e-02 -7.32301891e-01 -1.41714907e+00 5.98467708e-01
-1.74610794e+00 -9.26087499e-01 -3.16191941e-01 1.90059924e+00
1.04605806e+00 3.08275491e-01 -2.19302505e-01 -6.19155169e-02
4.68937188e-01 3.27999622e-01 -5.66079974e-01 -7.81368852e-01
-3.81620258e-01 4.15575802e-01 -2.08484098e-01 9.10358191e-01
-8.51632416e-01 1.06753170e+00 6.98406219e+00 4.85822976e-01
-6.43814325e-01 -1.55278578e-01 6.86003745e-01 9.95220765e-02
-3.69598269e-01 -9.69971046e-02 -7.72605121e-01 5.61046228e-02
1.23031056e+00 1.30924769e-02 2.87279427e-01 2.93231606e-01
-1.47119418e-01 -5.93397737e-01 -1.49650919e+00 5.37478268e-01
6.45335972e-01 -1.21383262e+00 1.58042789e-01 -5.15623450e-01
2.23260269e-01 8.86158049e-02 -2.01477915e-01 4.58617359e-01
2.16522664e-01 -1.67677259e+00 3.62275153e-01 5.94282389e-01
2.86373675e-01 -4.69648480e-01 7.58973777e-01 7.84141183e-01
-2.01218590e-01 -8.44199359e-02 -1.72295168e-01 -5.94548345e-01
2.55110145e-01 -1.46922320e-01 -8.54411900e-01 7.37828165e-02
7.16584980e-01 5.88349581e-01 -8.78046572e-01 5.62010646e-01
-7.12332547e-01 1.18924880e+00 -3.33944291e-01 -3.05518448e-01
1.67412207e-01 2.34953135e-01 5.79066038e-01 1.21979451e+00
-7.96822160e-02 5.85255146e-01 -9.53597948e-04 7.63220489e-01
-1.84386313e-01 3.31824303e-01 -6.00033045e-01 1.58511072e-01
6.24834836e-01 7.78438568e-01 -2.47952402e-01 -3.39040995e-01
-3.73518705e-01 6.31751835e-01 7.72671461e-01 4.23339397e-01
-2.64694870e-01 -3.31443042e-01 1.76800400e-01 -1.90366700e-01
2.06269532e-01 1.14510274e-02 -4.67009306e-01 -1.02759778e+00
-2.35009808e-02 -1.27374530e+00 6.43967211e-01 -1.08627141e+00
-1.34013534e+00 3.66520315e-01 1.96021914e-01 -3.08022946e-01
-5.05300343e-01 -7.81175971e-01 -9.69770968e-01 1.02263629e+00
-1.34508002e+00 -1.01505542e+00 -2.66748428e-01 3.25105369e-01
8.42480242e-01 -8.74246433e-02 1.25284600e+00 -4.78515714e-01
-4.69652325e-01 5.44321179e-01 -2.87405610e-01 1.59847379e-01
9.39733565e-01 -1.32765770e+00 4.59289849e-01 3.99541080e-01
-9.52337459e-02 8.63722682e-01 8.62155914e-01 -5.13947546e-01
-1.19841051e+00 -1.14676833e-01 1.60208166e+00 -9.92005885e-01
7.47815609e-01 -1.95039734e-01 -1.37134957e+00 1.08495271e+00
1.15151286e+00 -6.91658854e-01 9.08499897e-01 7.16059685e-01
-6.58447206e-01 5.21812022e-01 -8.82563591e-01 6.97308779e-01
1.20863652e+00 -6.00583494e-01 -1.75790262e+00 2.26723313e-01
8.17323446e-01 -5.60988069e-01 -8.80340993e-01 3.10071319e-01
3.21068913e-01 -1.08166039e+00 8.60649884e-01 -1.21416974e+00
8.75658810e-01 2.36183017e-01 -6.58999756e-02 -1.56907284e+00
-1.58375911e-02 -4.30068165e-01 -2.17789501e-01 1.14081848e+00
8.06922674e-01 -5.16627014e-01 1.46924555e-01 8.23349774e-01
-1.21693768e-01 -6.39627695e-01 -9.09216464e-01 -1.42719254e-01
7.61571109e-01 -4.34934676e-01 4.40127015e-01 9.89320159e-01
4.95992929e-01 1.05891478e+00 1.68175921e-01 -2.66597718e-01
1.67274222e-01 2.31499057e-02 3.99840951e-01 -1.42654622e+00
-2.33345151e-01 -7.11943865e-01 3.76722395e-01 -1.34004545e+00
5.96817136e-01 -8.72977555e-01 1.83443263e-01 -1.73151469e+00
1.65930554e-01 1.21170543e-01 2.13966072e-02 6.71972454e-01
-5.25062799e-01 -4.01268005e-01 3.48570675e-01 -3.97497453e-02
-6.92722619e-01 3.47276837e-01 1.02660429e+00 -1.93679839e-01
1.11907437e-01 -2.00748116e-01 -1.03305137e+00 8.23965967e-01
1.02478361e+00 2.22024973e-02 -3.51274639e-01 -8.85584891e-01
5.30655086e-01 1.90802798e-01 3.33810210e-01 -4.32908416e-01
4.52009499e-01 2.68217977e-02 3.93181264e-01 -4.70585108e-01
1.88646525e-01 -2.92347848e-01 -6.91946864e-01 2.70711094e-01
-1.23130584e+00 3.71714562e-01 3.43078494e-01 4.27089214e-01
-2.03184083e-01 -6.47998333e-01 3.13864440e-01 -9.59547982e-02
-6.45922422e-01 -4.66648966e-01 -6.09331608e-01 7.73178995e-01
4.90593642e-01 3.96006480e-02 -9.44695592e-01 -7.77974665e-01
-1.00896740e+00 4.09192711e-01 -1.63305849e-02 6.21523380e-01
4.95033175e-01 -7.00812936e-01 -1.02052641e+00 -7.70195723e-02
2.74619818e-01 2.27350473e-01 -8.57682079e-02 7.28502333e-01
-2.39189625e-01 5.96177995e-01 3.16726789e-02 -3.54240119e-01
-1.13874125e+00 4.05627251e-01 3.60525966e-01 -1.02508351e-01
-5.58615088e-01 8.55507731e-01 -8.37601647e-02 -4.74968255e-01
2.55719841e-01 -3.00205022e-01 -8.00122201e-01 3.56773108e-01
8.38619232e-01 3.55044276e-01 -2.42523625e-02 -3.95003259e-01
-1.99209824e-01 3.97568941e-01 -4.11815554e-01 -4.53744084e-02
1.18824637e+00 -2.54711717e-01 -1.02334566e-01 6.17419362e-01
8.76167119e-01 -1.30237058e-01 -7.79214621e-01 -4.00257796e-01
5.13262272e-01 3.32139097e-02 -5.38201816e-02 -1.55930817e+00
-3.16312999e-01 9.72303391e-01 1.26115680e-01 4.35470879e-01
8.19629610e-01 2.32605010e-01 6.91171885e-01 1.04016531e+00
2.65426040e-02 -1.15602219e+00 -1.23652175e-01 1.25550115e+00
1.22323930e+00 -1.44420826e+00 -1.75080165e-01 -1.69141114e-01
-5.32293618e-01 1.42075193e+00 9.97728944e-01 1.25565650e-02
4.22434837e-01 -6.18190132e-02 1.74136445e-01 -2.40898088e-01
-1.23864329e+00 -5.40156402e-02 2.66992718e-01 6.14656210e-01
8.16333413e-01 -3.66937406e-02 -5.56763351e-01 8.80848587e-01
-6.15922034e-01 -3.64997774e-01 5.92520595e-01 6.81959689e-01
-6.36505365e-01 -9.98931706e-01 -3.78212124e-01 6.90307379e-01
-3.67886037e-01 -5.75534999e-01 -8.79046977e-01 8.49203289e-01
-3.83622766e-01 1.28246486e+00 2.57781982e-01 4.55413871e-02
5.01752675e-01 6.58338010e-01 4.13920820e-01 -7.70263195e-01
-9.39663053e-01 -3.01280349e-01 8.70290041e-01 -3.86370838e-01
-3.51949066e-01 -9.12299633e-01 -1.29830933e+00 -2.05703661e-01
-2.66695350e-01 6.18307590e-02 9.68888029e-02 1.39712048e+00
2.81771719e-01 3.60004038e-01 -1.22266255e-01 -3.24376434e-01
-7.36951232e-01 -1.38119721e+00 1.56756029e-01 4.68172014e-01
4.25610512e-01 -4.72383142e-01 -4.60217535e-01 -7.77370334e-02] | [11.152833938598633, 8.199649810791016] |
44e602a5-46e9-436b-97b8-5c215e8dd03c | temporal-sub-sampling-of-audio-feature | 2007.02676 | null | https://arxiv.org/abs/2007.02676v1 | https://arxiv.org/pdf/2007.02676v1.pdf | Temporal Sub-sampling of Audio Feature Sequences for Automated Audio Captioning | Audio captioning is the task of automatically creating a textual description for the contents of a general audio signal. Typical audio captioning methods rely on deep neural networks (DNNs), where the target of the DNN is to map the input audio sequence to an output sequence of words, i.e. the caption. Though, the length of the textual description is considerably less than the length of the audio signal, for example 10 words versus some thousands of audio feature vectors. This clearly indicates that an output word corresponds to multiple input feature vectors. In this work we present an approach that focuses on explicitly taking advantage of this difference of lengths between sequences, by applying a temporal sub-sampling to the audio input sequence. We employ a sequence-to-sequence method, which uses a fixed-length vector as an output from the encoder, and we apply temporal sub-sampling between the RNNs of the encoder. We evaluate the benefit of our approach by employing the freely available dataset Clotho and we evaluate the impact of different factors of temporal sub-sampling. Our results show an improvement to all considered metrics. | ['Tuomas Virtanen', 'Khoa Nguyen', 'Konstantinos Drossos'] | 2020-07-06 | null | null | null | null | ['audio-captioning'] | ['audio'] | [ 6.52685642e-01 1.84557319e-01 3.01986188e-01 -3.03069919e-01
-8.83423686e-01 -6.33175135e-01 8.31076980e-01 3.43176544e-01
-5.19975245e-01 7.46625364e-01 5.98957539e-01 -4.89888750e-02
1.36992440e-01 -5.23714304e-01 -1.03848445e+00 -4.94959593e-01
-1.53116316e-01 4.55608010e-01 2.09926814e-01 -9.63318273e-02
1.34844795e-01 1.84119627e-01 -1.86107385e+00 6.03893340e-01
1.42262101e-01 9.73385453e-01 3.65591347e-01 1.03382647e+00
-2.58295387e-01 4.98817861e-01 -9.20001626e-01 -5.10322601e-02
1.14152350e-01 -7.24886239e-01 -7.70430028e-01 7.19510019e-02
3.93442392e-01 -1.61903709e-01 -3.96958329e-02 9.18373823e-01
4.96404469e-01 2.32190073e-01 6.05478346e-01 -1.08305717e+00
-9.20111388e-02 1.01663220e+00 -3.16573353e-03 1.21554375e-01
6.78428829e-01 7.69363157e-03 1.21715128e+00 -6.07965589e-01
7.32937932e-01 8.54241967e-01 4.16283011e-01 4.03071433e-01
-1.20902729e+00 -3.91042262e-01 -1.29243508e-01 2.46305749e-01
-1.31973481e+00 -5.76517522e-01 6.50748610e-01 -7.15541899e-01
7.10432887e-01 2.19168976e-01 5.93589604e-01 1.20785284e+00
-9.77743566e-02 3.20892423e-01 5.75081766e-01 -5.91833174e-01
4.18642431e-01 1.28230631e-01 -1.44422442e-01 -1.45562366e-01
-2.47147143e-01 -4.36109565e-02 -5.93740344e-01 -1.53701380e-01
4.65441406e-01 -3.05651963e-01 -3.48554641e-01 -6.07805587e-02
-1.17729473e+00 6.54128909e-01 1.80453956e-01 6.80967093e-01
-6.91797316e-01 4.74285811e-01 8.24251831e-01 2.18721747e-01
2.74253696e-01 2.68953651e-01 -1.47745252e-01 -4.41010624e-01
-1.18891203e+00 4.48397338e-01 7.87089050e-01 7.50896096e-01
5.34494698e-01 1.13500608e-02 -3.33219558e-01 6.50299668e-01
-4.56971116e-02 2.19108820e-01 6.80526912e-01 -6.70059741e-01
5.44712484e-01 -1.78140029e-03 2.76726514e-01 -7.81618237e-01
-1.45860836e-01 -4.06276137e-01 -5.31641722e-01 7.16222748e-02
5.38092196e-01 -4.63541955e-01 -6.29444361e-01 1.83204520e+00
8.00597668e-02 2.89463043e-01 2.08183408e-01 1.10559797e+00
4.56923574e-01 1.07443738e+00 2.14528628e-02 -2.21132502e-01
1.43016529e+00 -6.83066607e-01 -6.97103560e-01 -1.53270811e-01
4.48527575e-01 -8.72652829e-01 9.48937058e-01 4.60159570e-01
-9.86339450e-01 -6.83595717e-01 -1.03374088e+00 1.57029212e-01
-2.49347419e-01 2.04757556e-01 -2.22563058e-01 2.21109986e-01
-8.96264017e-01 7.79386520e-01 -3.24441344e-01 -2.16768727e-01
-2.17032611e-01 1.79068521e-01 -5.14281929e-01 3.12125951e-01
-1.38672853e+00 4.74326372e-01 7.92773306e-01 2.72298069e-03
-9.73235428e-01 -6.33541822e-01 -7.06711769e-01 4.49203700e-01
2.17326015e-01 -3.93556684e-01 1.42583072e+00 -1.48495054e+00
-1.33252215e+00 4.91630077e-01 -1.21648908e-01 -8.75442505e-01
6.02080703e-01 -2.51172900e-01 -3.04376155e-01 3.63497436e-01
-1.67593718e-01 8.33200157e-01 9.81844604e-01 -1.13928533e+00
-5.65723181e-01 1.12625696e-01 3.08630727e-02 1.30349934e-01
-2.85258591e-01 6.61027953e-02 -1.71656668e-01 -6.48863912e-01
-4.34300542e-01 -9.57870483e-01 8.11740849e-03 -3.13416451e-01
-4.80117708e-01 -8.80450159e-02 6.68973625e-01 -8.00402641e-01
1.31174517e+00 -2.45074534e+00 1.41022876e-01 6.94824830e-02
-1.46918342e-01 2.33386829e-01 -2.40198091e-01 8.23793650e-01
-4.22961533e-01 -1.16244227e-01 -3.96115154e-01 -4.40125763e-01
1.33853897e-01 2.33333614e-02 -4.82657850e-01 2.36069500e-01
2.14224726e-01 4.73896176e-01 -9.57213938e-01 -3.17235291e-01
2.83890516e-02 7.41919160e-01 -3.91492933e-01 3.97186607e-01
-5.92254579e-01 2.42587388e-01 -4.36838605e-02 -3.73485059e-01
2.41186142e-01 1.97042108e-01 5.06844260e-02 -1.62559375e-01
-3.23936462e-01 6.97474539e-01 -1.20218837e+00 1.85220981e+00
-6.22632325e-01 1.00823975e+00 -1.99001148e-01 -9.28474367e-01
1.14171720e+00 9.56030607e-01 3.66239905e-01 -2.81889260e-01
2.00748995e-01 4.38354433e-01 1.00892916e-01 -5.04964530e-01
3.93433958e-01 -2.82482296e-01 -6.08475693e-02 6.56747401e-01
1.17416896e-01 1.58857107e-01 3.51800978e-01 -1.38867302e-02
1.02416658e+00 -5.49127348e-03 2.98116624e-01 2.76984787e-03
6.99415088e-01 -3.56912404e-01 -1.67984754e-01 4.66389269e-01
2.41505012e-01 1.01749074e+00 7.89365292e-01 -2.67673045e-01
-1.48986280e+00 -6.97221160e-01 2.28154078e-01 9.87140954e-01
-4.85601336e-01 -4.79579836e-01 -1.04774547e+00 -3.20128143e-01
-2.99963802e-01 7.25037694e-01 -7.54076242e-01 -5.22242785e-02
-6.27840281e-01 -1.89365417e-01 6.68222487e-01 3.55753243e-01
-8.59425664e-02 -1.52879536e+00 -9.16001737e-01 5.55044353e-01
-2.48324573e-01 -1.20313334e+00 -5.90355039e-01 3.22146267e-01
-5.88300169e-01 -4.47483093e-01 -9.74240720e-01 -6.40567839e-01
2.75870472e-01 -3.54523867e-01 9.86198545e-01 -2.56367832e-01
-5.64965010e-02 1.04734123e-01 -6.34559214e-01 -3.42527032e-01
-7.97147334e-01 3.35809350e-01 2.35701762e-02 5.40079951e-01
6.60464391e-02 -8.23067129e-01 -4.11893040e-01 -2.16783375e-01
-1.37826920e+00 1.08227149e-01 6.52555168e-01 4.76158082e-01
2.63904691e-01 -3.44513506e-01 6.16470933e-01 -5.20114958e-01
8.69118035e-01 -4.59801137e-01 -4.43398237e-01 -1.46029666e-01
1.66512758e-01 3.51619065e-01 1.05746889e+00 -7.47780740e-01
-4.97495741e-01 4.77198720e-01 -3.88290942e-01 -4.81392980e-01
-3.99326056e-01 4.42838609e-01 -2.50063203e-02 4.35044765e-01
5.35624146e-01 3.94356102e-01 -1.11338794e-01 -5.71918309e-01
3.40792179e-01 9.45249736e-01 5.82746089e-01 -3.08420420e-01
6.14261985e-01 8.35833251e-02 -1.65429711e-01 -8.82574022e-01
-5.18825829e-01 -3.62938762e-01 -4.86154407e-01 -3.68556798e-01
7.69989431e-01 -5.38799047e-01 -6.22660279e-01 -2.71802489e-02
-1.62445390e+00 -7.78014958e-02 -4.92659986e-01 6.11569941e-01
-7.75199175e-01 7.98616335e-02 -3.23632747e-01 -8.51062596e-01
-3.83668542e-01 -1.07366347e+00 1.10718465e+00 -5.83022982e-02
-5.40427446e-01 -6.25168741e-01 2.25191131e-01 -1.47218838e-01
3.01934332e-01 3.52639258e-01 8.26346636e-01 -1.12913215e+00
-1.95131481e-01 -3.24383885e-01 -9.15699080e-02 5.64407647e-01
3.93505394e-02 -1.74616858e-01 -1.20227480e+00 1.18764695e-02
-1.35118380e-01 -1.98572874e-01 5.15851498e-01 9.95893925e-02
9.89118874e-01 -4.67552990e-01 9.03298184e-02 -8.45128596e-02
1.49913847e+00 3.08999717e-01 6.49412453e-01 1.87793761e-01
3.45994055e-01 8.74385476e-01 4.44866955e-01 5.91475487e-01
-1.75552517e-01 9.83963549e-01 5.54985344e-01 1.60067290e-01
-9.12879184e-02 -2.96852261e-01 5.48950493e-01 6.56372905e-01
1.58408538e-01 -3.47758055e-01 -8.86658549e-01 9.72475171e-01
-1.53824174e+00 -9.48951066e-01 -6.84142336e-02 2.27415204e+00
8.20140600e-01 3.31930459e-01 3.23365539e-01 5.58418989e-01
8.44493568e-01 1.60404697e-01 -1.75042778e-01 -7.17635393e-01
3.54613245e-01 1.43513769e-01 2.20726803e-01 6.36964917e-01
-7.26623952e-01 2.92173594e-01 5.68769503e+00 7.85884142e-01
-1.33915532e+00 6.03826717e-02 1.85176298e-01 -3.26818168e-01
-2.00368345e-01 -2.63096094e-01 -5.44503927e-01 6.68775618e-01
1.49620974e+00 -3.64648938e-01 4.19777781e-01 4.15045619e-01
5.25998652e-01 1.30070150e-01 -1.45295489e+00 8.43300879e-01
2.11036224e-02 -1.12755561e+00 3.18577588e-01 -7.29567781e-02
3.06266278e-01 -1.07649185e-01 -2.52042077e-02 1.47847226e-02
-4.93209749e-01 -1.04276156e+00 1.09081817e+00 4.56994385e-01
8.32396626e-01 -7.58893132e-01 8.39347422e-01 2.81228244e-01
-1.29269183e+00 8.32384229e-02 -1.10283658e-01 -1.26218140e-01
5.21079600e-01 5.73859811e-01 -1.23755562e+00 4.83717382e-01
2.83093870e-01 3.34060967e-01 -9.84609127e-02 1.14040315e+00
4.71131764e-02 7.04075098e-01 -3.50389391e-01 -2.01067552e-01
5.56339741e-01 6.40875474e-02 7.30953991e-01 1.69392359e+00
6.23221457e-01 -3.61415565e-01 -2.23646313e-01 8.10956955e-01
-1.44168481e-01 3.21582973e-01 -6.47354186e-01 -2.26584226e-01
3.84146839e-01 8.97533834e-01 -4.88637060e-01 -3.99186522e-01
-1.50572374e-01 9.39881444e-01 -1.14557199e-01 2.08995014e-01
-7.93111503e-01 -7.59243846e-01 5.17125547e-01 3.22604060e-01
6.53869689e-01 -2.81555369e-03 3.04976225e-01 -5.43712974e-01
3.11297089e-01 -7.80372560e-01 -1.14865109e-01 -9.95716989e-01
-7.10089922e-01 1.01605034e+00 8.63337368e-02 -1.36407959e+00
-8.71869445e-01 -2.37096906e-01 -5.09889185e-01 1.00344753e+00
-1.01073420e+00 -6.52042627e-01 -6.06480725e-02 1.72024921e-01
6.57406032e-01 1.07186303e-01 8.14698458e-01 5.62233925e-01
7.53161833e-02 2.43521079e-01 5.95573001e-02 1.24484360e-01
5.56993008e-01 -1.13241553e+00 5.43259263e-01 6.07112646e-01
5.21032870e-01 2.19541222e-01 1.45600951e+00 -2.80798554e-01
-8.54218483e-01 -9.84602153e-01 1.37950957e+00 -5.58025651e-02
8.54628682e-01 -7.52060235e-01 -9.84234035e-01 5.56558371e-01
5.22741258e-01 -1.80020958e-01 5.04156351e-01 -4.40668821e-01
-1.60919353e-01 -1.99845478e-01 -6.38189256e-01 2.88516760e-01
6.30920291e-01 -7.31963813e-01 -7.56111205e-01 5.22225089e-02
9.99886751e-01 -2.09379420e-01 -6.74772561e-01 -1.89992562e-02
6.80901945e-01 -9.01971459e-01 7.36344934e-01 -3.75145763e-01
6.95165098e-01 -3.74742448e-01 -1.12134755e-01 -1.35612261e+00
1.46860868e-01 -6.92792237e-01 1.18915103e-01 1.35512209e+00
4.76283491e-01 -6.58031031e-02 5.31590998e-01 -1.13783151e-01
1.01063363e-02 -4.31539714e-01 -1.07203078e+00 -7.21304834e-01
-3.11407208e-01 -5.41015744e-01 5.93039155e-01 4.34169412e-01
-1.25812152e-02 5.00384450e-01 -6.14477217e-01 6.36092871e-02
2.26314515e-01 -1.43316403e-01 4.50807929e-01 -1.11400104e+00
-4.67719078e-01 -2.75778204e-01 -4.80633438e-01 -9.03275073e-01
8.65005255e-02 -6.22728348e-01 4.31845605e-01 -1.20448232e+00
-2.54180372e-01 4.95850183e-02 -1.69336572e-01 2.61926353e-01
2.49105319e-01 4.22281593e-01 4.43808556e-01 -1.38251036e-01
-2.32073963e-01 3.82820338e-01 7.35458851e-01 -9.13013667e-02
-1.14316300e-01 5.09433821e-02 -1.38293996e-01 2.41045401e-01
7.25169957e-01 -7.42806375e-01 -3.24869066e-01 -3.39529186e-01
2.82491565e-01 2.73197740e-01 4.13605303e-01 -1.24118507e+00
2.43288085e-01 2.99644411e-01 -8.00671875e-02 -4.87990171e-01
4.91358995e-01 -9.96469975e-01 5.10180533e-01 5.73247612e-01
-7.52484441e-01 1.16905048e-01 3.95803660e-01 3.61346692e-01
-5.28711021e-01 -8.42167377e-01 4.90569293e-01 2.39028707e-02
-3.07795912e-01 -1.95870385e-01 -6.15033269e-01 -1.81692138e-01
8.06378365e-01 -5.23671918e-02 2.78079450e-01 -7.13944018e-01
-9.48098540e-01 -3.48438025e-01 1.03196166e-01 3.79446238e-01
4.46670979e-01 -1.34055853e+00 -9.07090843e-01 1.01864208e-02
1.04577489e-01 -3.23130369e-01 1.70196984e-02 5.38547218e-01
-4.08128500e-01 6.64412022e-01 -3.24251086e-01 -4.98875856e-01
-1.43422735e+00 5.12725890e-01 3.59406590e-01 -7.07215667e-02
-6.11661136e-01 5.34609199e-01 1.56952962e-02 1.60593763e-01
4.88365591e-01 -4.03352708e-01 -4.29982543e-01 3.16955358e-01
6.90866768e-01 5.93481027e-03 1.69761375e-01 -6.58245504e-01
-1.47474930e-01 3.88987988e-01 1.50655612e-01 -7.25669324e-01
1.29611468e+00 7.71454051e-02 1.00975171e-01 8.51476610e-01
1.59177339e+00 -3.93820070e-02 -1.09845591e+00 4.47094487e-03
1.50028333e-01 6.89038262e-03 -1.90212086e-01 -4.28327650e-01
-6.47367775e-01 1.07624936e+00 5.12373269e-01 5.02283931e-01
1.14336944e+00 9.91931930e-03 7.92655528e-01 5.98989278e-02
9.24551263e-02 -6.44952834e-01 -1.11377891e-02 3.99061024e-01
1.04918957e+00 -5.74177682e-01 -4.64541972e-01 -5.06673083e-02
-4.93237853e-01 1.35041034e+00 3.00798845e-02 -2.07151771e-01
1.84226960e-01 2.04645693e-01 6.72976300e-02 1.11616112e-01
-8.55877280e-01 -2.51597017e-01 2.61422366e-01 2.94931293e-01
4.30223048e-01 -1.16293401e-01 -4.81134385e-01 4.74211365e-01
-4.78791744e-01 1.89686418e-01 6.98560774e-01 5.49098730e-01
-3.57888699e-01 -1.29311752e+00 -5.40340066e-01 1.39975220e-01
-5.29797852e-01 -1.87304869e-01 -4.43346202e-01 3.83583367e-01
1.09274477e-01 7.04831481e-01 2.67074227e-01 -4.00912255e-01
3.77239406e-01 4.54832554e-01 1.65928602e-01 -6.56804621e-01
-8.54272485e-01 4.00162712e-02 8.33079517e-02 -1.91819221e-01
-2.82459676e-01 -6.84005320e-01 -1.13938844e+00 1.68368056e-01
2.66855843e-02 6.73451126e-01 1.00701761e+00 9.58816230e-01
1.47228748e-01 6.81782663e-01 5.44049203e-01 -1.13708389e+00
-4.71694469e-01 -1.27813804e+00 -3.82737547e-01 5.57128727e-01
8.06844890e-01 -1.84631139e-01 -5.15407264e-01 4.58188415e-01] | [15.284660339355469, 4.910427093505859] |
2d213761-7a97-4017-a6f6-a2bc549d2651 | comparison-of-uncertainty-quantification-with | 2211.06233 | null | https://arxiv.org/abs/2211.06233v1 | https://arxiv.org/pdf/2211.06233v1.pdf | Comparison of Uncertainty Quantification with Deep Learning in Time Series Regression | Increasingly high-stakes decisions are made using neural networks in order to make predictions. Specifically, meteorologists and hedge funds apply these techniques to time series data. When it comes to prediction, there are certain limitations for machine learning models (such as lack of expressiveness, vulnerability of domain shifts and overconfidence) which can be solved using uncertainty estimation. There is a set of expectations regarding how uncertainty should ``behave". For instance, a wider prediction horizon should lead to more uncertainty or the model's confidence should be proportional to its accuracy. In this paper, different uncertainty estimation methods are compared to forecast meteorological time series data and evaluate these expectations. The results show how each uncertainty estimation method performs on the forecasting task, which partially evaluates the robustness of predicted uncertainty. | ['Matias Valdenegro-Toro', 'Levente Foldesi'] | 2022-11-11 | null | null | null | null | ['time-series-regression'] | ['time-series'] | [-1.04110010e-01 5.07202804e-01 -3.03194135e-01 -7.12458253e-01
-2.32716039e-01 -5.31307161e-01 7.47895777e-01 1.96638212e-01
-4.54736799e-01 1.36569619e+00 4.21991587e-01 -8.39938104e-01
-4.83726114e-01 -8.80574226e-01 -5.88491917e-01 -3.82718295e-01
3.43667232e-02 1.99603528e-01 -1.35451883e-01 7.95901567e-02
4.61307794e-01 3.44137788e-01 -1.30866432e+00 8.19109157e-02
8.30972075e-01 1.43402553e+00 -2.25026980e-01 2.05100700e-01
-1.15843460e-01 9.60563898e-01 -7.16707349e-01 -3.56156409e-01
7.79382586e-02 2.25403413e-01 -5.59806883e-01 -5.71919799e-01
-2.10629314e-01 -6.29031420e-01 5.09691983e-02 1.19185936e+00
1.39153987e-01 2.71683335e-01 1.07096899e+00 -1.23100340e+00
-4.64522243e-01 1.25539947e+00 -9.74475965e-02 4.15275723e-01
5.21758199e-02 1.52827263e-01 8.07853520e-01 -2.37942174e-01
3.02724540e-01 1.30377758e+00 8.38295341e-01 1.72754899e-01
-1.13103926e+00 -8.22059691e-01 3.93243253e-01 -1.28074408e-01
-1.00065851e+00 -2.96123415e-01 5.82934618e-01 -7.10520089e-01
8.96407604e-01 2.58136034e-01 1.99271590e-01 1.06174326e+00
5.57979286e-01 2.28992850e-01 1.45629764e+00 -2.98117191e-01
7.35559583e-01 4.69906986e-01 -8.72238204e-02 -4.03090805e-01
4.65573996e-01 9.44382012e-01 -2.00247094e-01 -7.63751864e-02
4.00978982e-01 -1.45794034e-01 -2.01937437e-01 3.56423557e-01
-7.87970960e-01 1.09949708e+00 3.81235957e-01 4.05509084e-01
-3.76586288e-01 1.73380077e-02 4.69434083e-01 3.23675036e-01
9.71783698e-01 8.55034292e-01 -6.95037603e-01 -1.66975588e-01
-8.63587320e-01 4.56731409e-01 9.33896124e-01 3.75756919e-01
2.03931913e-01 4.64457870e-01 -1.64432049e-01 2.21211433e-01
4.67643946e-01 5.00924528e-01 6.54351294e-01 -1.18735588e+00
3.22206378e-01 4.56014350e-02 5.17765760e-01 -1.36085045e+00
-6.76011562e-01 -3.58937681e-01 -7.62884438e-01 5.12906969e-01
4.51006055e-01 -7.46335924e-01 -5.49064279e-01 1.83156896e+00
-7.38165826e-02 1.60862327e-01 3.74048114e-01 6.24688208e-01
4.62672979e-01 8.00017893e-01 2.40947992e-01 -3.62089247e-01
8.25920224e-01 -2.56070286e-01 -8.10889721e-01 -2.86502630e-01
5.27501762e-01 -3.17217886e-01 5.89335263e-01 3.03792149e-01
-7.02147961e-01 -4.26417410e-01 -9.65440869e-01 7.21920311e-01
-4.22955841e-01 -5.48352301e-01 5.15472114e-01 6.99778199e-01
-4.94773239e-01 1.29805338e+00 -6.52252078e-01 1.87016845e-01
1.27146691e-01 1.54911482e-03 5.29371426e-02 5.62032640e-01
-1.87579000e+00 1.47344208e+00 1.02726543e+00 2.38881201e-01
-3.69461507e-01 -7.15064287e-01 -8.12119484e-01 3.52249146e-01
8.80238116e-02 -3.02866727e-01 1.49672747e+00 -1.20996833e+00
-1.46202290e+00 1.17307283e-01 6.46857083e-01 -9.88016486e-01
7.13730454e-01 -1.08010650e-01 -7.54445851e-01 -4.21747774e-01
-6.59292191e-02 2.86260277e-01 6.55573845e-01 -8.48414302e-01
-6.72227383e-01 -1.83363751e-01 -5.45429438e-03 3.88972349e-02
5.79461791e-02 2.55314469e-01 1.01437771e+00 -8.78169835e-01
-1.66461453e-01 -7.24911034e-01 -4.66272742e-01 -5.24736702e-01
-1.56922102e-01 -6.29510432e-02 3.31387073e-01 -5.75860083e-01
1.63217628e+00 -1.83091378e+00 -5.14227688e-01 5.32957315e-01
-1.52298793e-01 -4.02560495e-02 4.19574708e-01 3.18900853e-01
-2.51041770e-01 5.33335328e-01 -2.38266647e-01 2.54295230e-01
4.47023422e-01 2.39507928e-01 -8.61698627e-01 1.82098955e-01
2.15044111e-01 5.24887204e-01 -6.02811277e-01 -7.58268014e-02
2.86146164e-01 3.98523808e-02 -3.24915111e-01 -1.12108871e-01
-4.95021313e-01 3.56662691e-01 -4.19569165e-01 3.41585785e-01
5.60417652e-01 -1.66048840e-01 1.94743544e-01 2.57424444e-01
-4.66221809e-01 3.76842290e-01 -1.13769484e+00 7.13133335e-01
-2.80792147e-01 6.61723912e-01 -5.10461569e-01 -1.00453615e+00
9.67894554e-01 5.26186585e-01 -6.41409531e-02 -2.99149543e-01
2.46311799e-01 2.37595618e-01 1.79393083e-01 -3.00647050e-01
3.83358181e-01 -4.85229194e-01 -2.21651986e-01 6.16371095e-01
-2.73620158e-01 -4.83652681e-01 -2.14367077e-01 -4.50447619e-01
5.15681267e-01 1.15058161e-01 4.06592190e-01 -5.33799350e-01
2.80412361e-02 3.65977436e-02 8.88123930e-01 6.49490654e-01
-1.13547951e-01 1.93366125e-01 7.54433036e-01 -6.53598666e-01
-8.99948716e-01 -8.19040895e-01 -6.10193551e-01 7.59465933e-01
-5.66285074e-01 1.74818322e-01 -4.40116256e-01 -4.14944112e-01
4.13687289e-01 1.67340982e+00 -7.76759863e-01 -2.65613884e-01
1.52974263e-01 -7.62413144e-01 3.67525339e-01 7.93859184e-01
3.31500411e-01 -9.93473887e-01 -1.02859628e+00 2.63909996e-01
1.48858592e-01 -6.18877411e-01 1.97215993e-02 2.59408295e-01
-8.93279731e-01 -5.99241614e-01 -4.37088609e-01 3.25928479e-01
3.19829583e-01 -5.68397462e-01 1.15730739e+00 -4.72176999e-01
5.00882208e-01 9.37150717e-02 -6.61166012e-02 -1.12255836e+00
-5.64799130e-01 -3.67832989e-01 4.55545217e-01 -4.18098003e-01
4.14966762e-01 -6.42993808e-01 -3.99937063e-01 1.98650718e-01
-8.93944442e-01 -5.05803764e-01 3.35648775e-01 8.72334182e-01
1.77057788e-01 5.66761792e-01 9.42427397e-01 -6.51604354e-01
1.11479926e+00 -7.10764229e-01 -1.06413770e+00 3.58270258e-01
-1.30584633e+00 2.03861117e-01 2.82764673e-01 -4.95494634e-01
-1.50843298e+00 -3.38385940e-01 2.76455820e-01 -2.31389090e-01
-1.89333066e-01 1.28860378e+00 3.45240057e-01 2.56468832e-01
8.18203688e-01 -3.96371931e-01 5.02114743e-02 -2.20926568e-01
1.26069039e-01 5.73678970e-01 2.79993325e-01 -7.85741031e-01
5.27908981e-01 -1.36374846e-01 -1.06686853e-01 -3.61268848e-01
-9.91386473e-01 3.41495812e-01 -3.41777533e-01 -5.10990322e-01
5.48107088e-01 -7.84159005e-01 -5.60217083e-01 1.61124244e-01
-9.46743131e-01 -2.81329393e-01 -4.64917809e-01 1.07273662e+00
-5.87075472e-01 -3.34081426e-02 -1.50520936e-01 -1.30861270e+00
-1.39522776e-01 -9.73976672e-01 1.77476630e-01 4.11018163e-01
-7.02644408e-01 -1.32059944e+00 2.10854590e-01 -1.37932450e-01
8.95879447e-01 5.88066757e-01 8.98519993e-01 -1.04932678e+00
1.31104246e-01 -2.37370148e-01 -8.35309401e-02 6.26062095e-01
-1.22408554e-01 3.72831672e-01 -1.14042354e+00 1.65719301e-01
2.59723723e-01 -3.79197091e-01 7.18496919e-01 9.13577020e-01
1.08410752e+00 -8.21312785e-01 -9.30521563e-02 1.69488013e-01
1.08441031e+00 5.45966685e-01 4.08308089e-01 5.99391282e-01
-3.07212144e-01 1.14554048e+00 6.56227291e-01 8.08218002e-01
1.24824889e-01 2.04616472e-01 4.11659151e-01 5.80770671e-01
8.81899118e-01 -3.40994179e-01 2.82869667e-01 2.04052091e-01
-2.10658729e-01 -2.12468281e-01 -1.18352211e+00 1.57563314e-01
-1.88318062e+00 -1.26241660e+00 2.83656478e-01 2.40431237e+00
7.95722544e-01 5.91183245e-01 -4.57627000e-04 1.14014909e-01
5.84155500e-01 1.38175607e-01 -6.35500431e-01 -7.93203354e-01
-1.32249951e-01 -2.10944980e-01 5.76392233e-01 5.18935323e-01
-1.10475004e+00 5.91920495e-01 7.40464306e+00 6.35139585e-01
-1.10455441e+00 -3.92528355e-01 1.16752923e+00 -7.70563483e-02
-6.98467791e-01 -1.17734345e-02 -4.90934730e-01 7.34832644e-01
1.41895485e+00 -6.70337439e-01 1.24468133e-01 9.18220937e-01
2.54558533e-01 -2.47530356e-01 -9.50910449e-01 2.25917116e-01
-6.48038566e-01 -1.34623909e+00 -2.77550220e-01 -2.74964571e-01
7.78175950e-01 -1.39080048e-01 2.52589464e-01 5.06261766e-01
8.37223113e-01 -1.41647089e+00 7.48117507e-01 9.06898379e-01
4.03839052e-01 -9.86937284e-01 1.21875715e+00 5.07290184e-01
-2.71809161e-01 -3.75421524e-01 -6.45464420e-01 -7.32128501e-01
8.11129957e-02 1.01763856e+00 -8.21781039e-01 2.18008935e-01
9.36532378e-01 2.29764149e-01 -6.64747432e-02 6.29725099e-01
-3.34479779e-01 7.27281034e-01 -5.72269320e-01 -2.88418561e-01
4.54685420e-01 -2.33023435e-01 2.79646456e-01 8.94206822e-01
5.91686010e-01 2.49839291e-01 -3.16226721e-01 1.10206282e+00
3.77013028e-01 -1.61143139e-01 -8.83397579e-01 -3.43818605e-01
7.03034937e-01 4.98439431e-01 -3.38736802e-01 -2.07189918e-01
-3.15121561e-01 1.00306921e-01 -1.28892213e-01 4.02207375e-01
-5.45080066e-01 -1.75941736e-01 5.00940323e-01 -1.59651145e-01
-1.67655632e-01 2.72620857e-01 -5.85117400e-01 -8.37526739e-01
-2.96534121e-01 -7.39051521e-01 5.46467721e-01 -8.71683896e-01
-1.35975635e+00 4.78214979e-01 2.94688493e-01 -1.12156606e+00
-8.25053930e-01 -7.91913867e-01 -7.67834008e-01 1.01241302e+00
-1.30967534e+00 -6.23721957e-01 3.91962141e-01 3.81923653e-02
2.44310349e-02 -1.75939903e-01 9.06663835e-01 -4.41909999e-01
-3.69974077e-01 1.54279143e-01 4.87589926e-01 -1.25127763e-01
6.89799547e-01 -1.21871781e+00 1.46407351e-01 5.93583465e-01
-4.88472819e-01 3.39366823e-01 1.17454696e+00 -7.62646198e-01
-3.65717560e-01 -7.74710715e-01 1.11653161e+00 -4.78906333e-01
1.03648078e+00 4.04374093e-01 -8.69449973e-01 9.27083790e-01
-1.39174536e-02 -2.28905991e-01 7.93169260e-01 4.13807780e-01
-3.51270050e-01 -3.82083617e-02 -1.51387501e+00 4.13501322e-01
1.23720244e-01 -3.25914413e-01 -1.08514678e+00 1.44865140e-01
8.74257565e-01 -2.42595062e-01 -1.25706112e+00 7.76196718e-01
8.80407333e-01 -9.28292036e-01 5.68041861e-01 -1.05769932e+00
5.41740954e-01 4.42074575e-02 -4.70790118e-01 -1.71236467e+00
-3.43352795e-01 -2.67174393e-01 1.37341365e-01 1.12998354e+00
9.51295674e-01 -1.00228119e+00 3.48445207e-01 1.51590729e+00
2.36182809e-01 -4.57300067e-01 -1.12075257e+00 -8.75057518e-01
6.65491343e-01 -9.11942124e-01 1.04925203e+00 1.31870401e+00
3.56352240e-01 -1.40834153e-01 -4.04369444e-01 2.57544518e-01
4.19259578e-01 1.65231079e-01 1.20095491e-01 -1.66557860e+00
-1.98372707e-01 -6.38211966e-01 -2.75308341e-01 -1.68392196e-01
3.46968412e-01 -2.84619331e-01 -1.32813379e-01 -8.81422698e-01
-3.96905601e-01 -3.64760339e-01 -6.28731072e-01 2.10935101e-01
-1.14418836e-02 -5.69775164e-01 1.38751760e-01 -3.14906836e-02
1.03586294e-01 5.03693461e-01 6.64381742e-01 -8.06696042e-02
-2.10377157e-01 6.49075627e-01 -6.39585137e-01 1.07015002e+00
1.22732306e+00 -2.93378830e-01 -3.80049109e-01 -1.73829392e-01
7.37172544e-01 5.60778797e-01 1.78025141e-01 -7.45052993e-01
-2.39802413e-02 -7.51377463e-01 5.45443535e-01 -3.58946681e-01
-1.11584015e-01 -9.51461136e-01 6.19696379e-01 5.76255441e-01
-6.83712125e-01 2.46949896e-01 5.40806174e-01 5.14983833e-01
-3.46573710e-01 -4.82104838e-01 4.89331096e-01 -1.66166902e-01
-4.82122779e-01 -1.01804540e-01 -5.60591996e-01 -1.13716245e-01
9.44426417e-01 1.38323577e-02 -2.32107654e-01 -7.26575792e-01
-8.55237246e-01 4.14123088e-01 1.15683831e-01 3.39373261e-01
1.86689407e-01 -1.31475520e+00 -9.23884511e-01 -2.06620768e-01
-1.63074094e-03 -1.98747650e-01 3.01762491e-01 5.79871714e-01
-2.37123474e-01 8.83331716e-01 -2.35066935e-01 -1.38691142e-01
-4.07618731e-01 6.24877393e-01 5.58550596e-01 -3.99479061e-01
-1.36793971e-01 7.56922543e-01 8.67093652e-02 -5.28342903e-01
1.75198033e-01 -7.03926742e-01 -3.48245054e-01 3.36519778e-01
5.57568610e-01 6.32308543e-01 -2.00443566e-01 -1.54379845e-01
-1.70013696e-01 7.31732324e-02 2.04551354e-01 -4.32827353e-01
1.32243299e+00 -2.54750270e-02 3.82478908e-02 7.74439216e-01
3.56362134e-01 -4.22338605e-01 -1.52846313e+00 -2.45139107e-01
5.18672228e-01 -3.94783020e-01 4.45987910e-01 -1.33503032e+00
-7.12569356e-01 8.10075998e-01 4.92195368e-01 2.96287388e-01
8.71814430e-01 -5.25547862e-01 -2.10840747e-01 5.28313994e-01
1.86830416e-01 -1.68031335e+00 -6.44778669e-01 3.28424752e-01
1.28540492e+00 -1.49386752e+00 4.41678874e-02 4.62552309e-01
-9.29507911e-01 1.21556973e+00 3.60183924e-01 2.63690680e-01
1.14320886e+00 3.23257238e-01 -8.12571272e-02 2.68560767e-01
-1.00780237e+00 2.18029439e-01 3.71970117e-01 5.37119269e-01
3.46963197e-01 5.55411875e-01 -3.53575349e-01 1.15322518e+00
-4.98786062e-01 2.10706159e-01 4.50218409e-01 5.21538675e-01
-5.91845691e-01 -3.99652719e-01 -6.54004037e-01 6.78925157e-01
-7.10849404e-01 -1.30709648e-01 -2.25614175e-01 7.72809923e-01
-1.66571483e-01 1.17252266e+00 2.75831878e-01 -4.06697720e-01
1.51257440e-01 2.40187064e-01 -3.19309562e-01 -2.39052549e-01
-3.75260472e-01 -1.56077370e-01 4.18724447e-01 -4.40096438e-01
-3.37903529e-01 -7.80041158e-01 -8.82714152e-01 -5.26443660e-01
-3.74529511e-01 3.93534720e-01 6.57904625e-01 1.03726161e+00
1.94913954e-01 4.55347151e-01 5.75268567e-01 -4.78653729e-01
-1.45087194e+00 -1.27915144e+00 -7.33276129e-01 -1.90903559e-01
1.53576061e-01 -6.48108900e-01 -8.88427734e-01 -4.66798037e-01] | [7.320659160614014, 4.055183410644531] |
a762c775-45e6-4faa-9e0c-943c2ac640c1 | light-weight-deep-extreme-multilabel | 2304.11045 | null | https://arxiv.org/abs/2304.11045v1 | https://arxiv.org/pdf/2304.11045v1.pdf | Light-weight Deep Extreme Multilabel Classification | Extreme multi-label (XML) classification refers to the task of supervised multi-label learning that involves a large number of labels. Hence, scalability of the classifier with increasing label dimension is an important consideration. In this paper, we develop a method called LightDXML which modifies the recently developed deep learning based XML framework by using label embeddings instead of feature embedding for negative sampling and iterating cyclically through three major phases: (1) proxy training of label embeddings (2) shortlisting of labels for negative sampling and (3) final classifier training using the negative samples. Consequently, LightDXML also removes the requirement of a re-ranker module, thereby, leading to further savings on time and memory requirements. The proposed method achieves the best of both worlds: while the training time, model size and prediction times are on par or better compared to the tree-based methods, it attains much better prediction accuracy that is on par with the deep learning based methods. Moreover, the proposed approach achieves the best tail-label prediction accuracy over most state-of-the-art XML methods on some of the large datasets\footnote{accepted in IJCNN 2023, partial funding from MAPG grant and IIIT Seed grant at IIIT, Hyderabad, India. Code: \url{https://github.com/misterpawan/LightDXML} | ['Pawan Kumar', 'Bamdev Mishra', 'Pratik Jawanpuria', 'Arpan Dasgupta', 'Istasis Mishra'] | 2023-04-20 | null | null | null | null | ['multi-label-learning'] | ['methodology'] | [ 2.39264607e-01 -6.81021884e-02 -4.31882173e-01 -6.10593259e-01
-9.44493175e-01 -4.45312232e-01 5.04977465e-01 6.11782312e-01
-5.66078484e-01 7.73327351e-01 9.59474780e-03 -5.09223342e-01
-3.12227011e-01 -7.56017327e-01 -2.15923786e-01 -8.03196251e-01
1.32858664e-01 6.57976329e-01 1.58089213e-02 3.09047818e-01
1.70651719e-01 4.80453908e-01 -1.56227815e+00 2.00419903e-01
5.19081533e-01 1.47839200e+00 -8.60788822e-02 4.95540470e-01
-2.96547949e-01 9.09803152e-01 -1.42157674e-01 -4.98812884e-01
3.00827175e-01 -9.03029740e-02 -9.43977654e-01 -2.87152710e-03
5.67517698e-01 -3.41309816e-01 1.65539876e-01 1.01805627e+00
6.61854386e-01 5.01977354e-02 8.44369888e-01 -1.35586703e+00
-4.21368539e-01 4.90818679e-01 -9.46757376e-01 -6.00936785e-02
1.73640177e-02 -4.71514195e-01 1.42851794e+00 -9.66105938e-01
3.84190798e-01 1.01234317e+00 8.65228236e-01 5.07967114e-01
-1.05886519e+00 -9.47793961e-01 -1.24744363e-02 2.32438013e-01
-1.28709102e+00 -1.09191455e-01 5.65230668e-01 -4.38210130e-01
6.80961907e-01 4.54783052e-01 2.82160193e-01 8.86273265e-01
1.94168594e-02 8.18301320e-01 1.44446301e+00 -4.90326345e-01
2.22927392e-01 3.12057167e-01 5.15335381e-01 8.98257077e-01
5.27688153e-02 -1.11074686e-01 -2.07745343e-01 -4.47726637e-01
1.50890410e-01 1.52914315e-01 1.57298073e-01 -3.77372324e-01
-7.78101087e-01 1.11788809e+00 1.04549840e-01 2.83277363e-01
-3.59749466e-01 2.06496626e-01 7.68387139e-01 2.26946220e-01
8.25339735e-01 -8.07814226e-02 -8.14665675e-01 6.41487830e-04
-9.80085969e-01 9.24140215e-02 6.41110897e-01 7.60994077e-01
5.66286147e-01 -1.72004700e-01 1.11578599e-01 1.21831012e+00
4.67904210e-01 7.65054524e-02 5.98004222e-01 -7.75318861e-01
3.97421062e-01 6.55359805e-01 -1.40601829e-01 -8.00062895e-01
-7.67140269e-01 -6.65556073e-01 -8.99132013e-01 1.16469450e-01
3.57833773e-01 -1.56959161e-01 -7.73507118e-01 1.60099554e+00
6.15986168e-01 -1.23121571e-02 1.89620145e-02 3.48955452e-01
7.42271185e-01 6.18553400e-01 5.05971670e-01 -4.01229382e-01
1.46400666e+00 -1.06069362e+00 -5.17490327e-01 -6.41810969e-02
1.18868697e+00 -7.45125949e-01 9.10628736e-01 5.71094453e-01
-6.23708069e-01 -4.26207244e-01 -1.07167733e+00 4.48204093e-02
-5.53357542e-01 2.00095996e-01 7.88538635e-01 7.98230469e-01
-1.00116634e+00 5.14747620e-01 -3.90230000e-01 -4.05360490e-01
5.40605843e-01 5.94187796e-01 -4.43473577e-01 -4.13210094e-01
-1.19546878e+00 6.49791420e-01 4.99483138e-01 -1.96391702e-01
-4.46805269e-01 -5.66250801e-01 -7.26415694e-01 -1.49915442e-01
3.03659678e-01 -9.30138826e-02 1.13519394e+00 -7.13606417e-01
-1.08820164e+00 8.72214496e-01 1.30121514e-01 -1.99366584e-01
6.09895647e-01 -1.26167700e-01 -5.08117855e-01 -2.24345326e-02
1.79731041e-01 8.83468509e-01 4.61889356e-01 -1.15869653e+00
-1.07770145e+00 -4.03093785e-01 -3.75299789e-02 1.26855910e-01
-8.73342395e-01 9.02328640e-02 -2.36584455e-01 -4.93168682e-01
-1.08210575e-02 -1.03865528e+00 -1.58261091e-01 -8.80485326e-02
-3.54953825e-01 -8.18581164e-01 8.53267312e-01 -5.60081303e-01
1.47770846e+00 -2.02319431e+00 -2.64069557e-01 8.33682865e-02
2.95861393e-01 3.24589312e-01 -7.73507506e-02 5.25651038e-01
-3.36139321e-01 2.31758758e-01 3.44769917e-02 -4.25973028e-01
1.60210788e-01 1.31281102e-02 1.49140343e-01 6.21423542e-01
-1.66453451e-01 6.04303598e-01 -8.47982764e-01 -8.47653568e-01
1.83171317e-01 3.35417658e-01 -3.11333627e-01 -1.92299247e-01
-1.61862358e-01 2.64632106e-01 -3.36769789e-01 8.20291340e-01
6.37143135e-01 -4.36153889e-01 4.28935945e-01 -1.71948224e-01
-7.11160749e-02 -5.93220741e-02 -1.37873673e+00 1.27551985e+00
-5.15951335e-01 8.84920582e-02 -2.47309372e-01 -1.15015471e+00
8.13930035e-01 5.16970217e-01 8.33177745e-01 -6.11963451e-01
4.14484084e-01 4.33050305e-01 -3.05831283e-01 -2.73082614e-01
2.26221427e-01 -2.91310430e-01 -1.89047024e-01 6.82242692e-01
5.39279133e-02 4.74694043e-01 2.67491043e-01 5.03962860e-03
9.32440519e-01 -3.85359935e-02 5.45354187e-01 -2.03743577e-01
5.55632293e-01 -2.76040345e-01 6.65915906e-01 4.03868705e-01
-1.23395853e-01 1.24189639e-02 4.56294477e-01 -5.66953123e-01
-9.33109999e-01 -6.17353499e-01 -4.85817343e-01 1.49405301e+00
-3.12823772e-01 -3.33538592e-01 -5.04967272e-01 -1.20546198e+00
1.83458462e-01 7.23630369e-01 -7.20638573e-01 1.77414536e-01
-3.93368721e-01 -8.54821205e-01 6.15978062e-01 4.98144865e-01
4.28615689e-01 -8.82317007e-01 -3.04100454e-01 1.74117059e-01
5.32080159e-02 -8.54348063e-01 -2.77192771e-01 7.43893325e-01
-9.37581122e-01 -9.80618000e-01 -7.32625186e-01 -8.26893389e-01
4.55426186e-01 -1.37867138e-01 8.46278310e-01 4.19307016e-02
-1.41546115e-01 -8.73892307e-02 -4.77784604e-01 -3.24904025e-01
-3.25664520e-01 3.24368924e-01 6.16110004e-02 -4.62788679e-02
6.10561490e-01 -3.89563739e-01 -5.30741394e-01 1.23457171e-01
-9.99226749e-01 5.80820208e-03 6.62179112e-01 8.94359469e-01
6.82085514e-01 3.19070697e-01 1.00315702e+00 -1.35842490e+00
4.94823933e-01 -6.62482560e-01 -4.58687782e-01 3.13739151e-01
-1.28065658e+00 -1.04208075e-01 7.74753630e-01 -3.89810115e-01
-7.85018086e-01 1.48427054e-01 -4.18242484e-01 -1.62290279e-02
-1.06708482e-01 3.90836418e-01 -1.21612497e-01 4.95550819e-02
4.14753109e-01 4.52181920e-02 -1.58972532e-01 -7.54522502e-01
3.17399442e-01 1.14765322e+00 -1.82219192e-01 -2.67021179e-01
3.52240294e-01 2.44646683e-01 3.05005789e-01 -4.34786260e-01
-1.12697959e+00 -5.44858217e-01 -8.51984799e-01 -2.86764026e-01
8.29676807e-01 -6.90581322e-01 -6.37632310e-01 5.96480131e-01
-5.58890820e-01 -5.70642054e-02 -3.86193730e-02 4.20653909e-01
-2.73510665e-01 5.11206746e-01 -8.12950730e-01 -6.47703886e-01
-5.93193531e-01 -1.00728440e+00 8.06991339e-01 2.01517288e-02
-2.95411497e-01 -1.18601894e+00 -1.14907034e-01 5.37041545e-01
2.27182046e-01 3.70171666e-01 1.45346391e+00 -1.27335477e+00
-8.72227848e-02 -6.02332115e-01 -4.08877909e-01 6.40828848e-01
1.34503007e-01 -3.46693188e-01 -1.01063073e+00 -5.90663195e-01
-1.52376100e-01 -7.13630021e-01 6.07442200e-01 1.74944028e-01
1.18504798e+00 -2.13116109e-01 -3.57797801e-01 2.48984545e-01
1.80017471e+00 2.81919122e-01 -1.29610170e-02 4.46709186e-01
7.25883424e-01 6.21360362e-01 8.95042956e-01 7.26920605e-01
3.94844025e-01 5.89144051e-01 4.81045306e-01 -1.82541698e-01
-1.74124949e-02 -7.58763263e-03 -4.67417613e-02 8.86852860e-01
5.19934952e-01 -3.97821844e-01 -9.89752948e-01 4.40709412e-01
-1.71179581e+00 -4.72570598e-01 -3.58068317e-01 2.09247851e+00
9.34883416e-01 1.10209160e-01 2.71419853e-01 5.30814171e-01
5.99059403e-01 1.63594335e-01 -5.70733488e-01 -5.83427250e-01
3.20121616e-01 3.10463727e-01 7.46270418e-01 3.28279644e-01
-1.30814302e+00 7.09274292e-01 5.12159824e+00 1.18979311e+00
-1.11892378e+00 5.22609651e-01 6.44221365e-01 -4.69018519e-02
-1.51008414e-02 -2.28001237e-01 -1.12600434e+00 3.51329476e-01
1.15341449e+00 1.10252514e-01 5.37289344e-02 9.93185639e-01
-1.20520204e-01 -4.47420068e-02 -1.07820451e+00 7.88769305e-01
2.21233983e-02 -1.01073563e+00 -1.23890579e-01 3.74980509e-01
4.69975740e-01 8.50429013e-03 5.66272326e-02 3.72422934e-01
4.25219536e-01 -7.72133052e-01 6.07264698e-01 1.74617961e-01
1.05774546e+00 -1.04197049e+00 9.48521256e-01 4.02612656e-01
-1.13351977e+00 -5.16348600e-01 -1.89752653e-01 1.87307239e-01
3.76097858e-02 9.00757074e-01 -7.81195700e-01 4.87388670e-01
5.26523113e-01 5.13546050e-01 -7.52636433e-01 8.45004559e-01
7.10301176e-02 7.47339249e-01 -2.00263813e-01 -1.28070056e-01
4.10507351e-01 2.32219398e-02 -4.45885770e-02 1.17381775e+00
1.94335356e-01 -2.19828025e-01 3.64387393e-01 8.59779306e-03
-1.90089405e-01 5.92698991e-01 -5.75194538e-01 -1.83582343e-02
5.09443581e-01 1.44794393e+00 -9.40703809e-01 -4.07136381e-01
-7.30725706e-01 6.59250021e-01 4.06835496e-01 -1.13198452e-01
-9.56438303e-01 -2.85844594e-01 1.59605369e-01 -1.85632315e-02
2.62126833e-01 2.49459013e-01 -3.02017003e-01 -4.98069108e-01
-1.81684926e-01 -7.98772752e-01 6.65207565e-01 -1.52275965e-01
-1.25123477e+00 5.54707408e-01 -1.35137260e-01 -1.21805024e+00
-1.67079851e-01 -5.64765811e-01 1.01946361e-01 6.56861484e-01
-1.30510128e+00 -1.39332211e+00 1.63602363e-02 2.63513535e-01
5.53745985e-01 -1.09777234e-01 1.12101018e+00 8.25756133e-01
-5.36917269e-01 9.78251815e-01 5.61577082e-01 -1.12529248e-02
6.29236042e-01 -1.28610289e+00 -5.94024919e-02 1.88608468e-01
3.88202593e-02 2.25538760e-01 2.57048011e-01 -4.77808893e-01
-7.57551789e-01 -1.36674798e+00 1.22968721e+00 -1.71060994e-01
6.25295103e-01 -4.40289080e-01 -5.78480542e-01 5.25581479e-01
7.20747113e-02 -8.39300570e-04 1.33089614e+00 1.06303491e-01
-3.51046830e-01 -2.73368329e-01 -1.35285187e+00 1.95659876e-01
5.84529281e-01 -2.92521805e-01 -3.08378041e-02 6.07743680e-01
4.57113355e-01 4.09512520e-02 -1.16727543e+00 4.84420896e-01
7.97341168e-01 -7.48614430e-01 7.20891237e-01 -3.56275350e-01
3.12888324e-01 -8.84857494e-03 -4.48258132e-01 -8.94403756e-01
-4.84071761e-01 2.15643540e-01 -3.20354849e-02 1.59430683e+00
5.02522528e-01 -6.92447960e-01 9.09444451e-01 4.71109241e-01
1.22508265e-01 -1.34745145e+00 -8.83836567e-01 -6.75087392e-01
1.80425078e-01 -4.48748529e-01 3.76428396e-01 1.23814929e+00
-3.09841275e-01 4.05607998e-01 -4.69886184e-01 -1.65427700e-01
8.22196007e-01 -6.86850473e-02 2.21791849e-01 -1.54613078e+00
-1.44303247e-01 -1.27956763e-01 -1.84332699e-01 -6.81443870e-01
2.94200748e-01 -1.36884487e+00 -3.91882390e-01 -1.59878421e+00
3.21499288e-01 -9.19903100e-01 -6.38531864e-01 7.96582401e-01
1.04521230e-01 4.21898007e-01 1.30621076e-01 2.15636805e-01
-7.00098336e-01 1.68550074e-01 9.33563888e-01 -3.47574688e-02
1.97809696e-01 2.04862595e-01 -7.73903489e-01 6.38605058e-01
8.95263076e-01 -8.28032613e-01 -4.84371245e-01 -2.53614187e-01
2.01374695e-01 1.01863772e-01 -5.60051873e-02 -9.93132770e-01
-7.83529598e-03 -2.66888868e-02 4.22858238e-01 -7.34322667e-01
1.71197295e-01 -9.04069185e-01 2.20489353e-01 5.42061329e-01
-6.30743861e-01 1.73198089e-01 1.52211320e-02 6.31437778e-01
8.11490491e-02 -6.42209649e-01 9.79773700e-01 -1.13040537e-01
-6.13384962e-01 3.71062100e-01 -3.48719299e-01 -1.06695913e-01
1.28676701e+00 -1.96210191e-01 -1.60071358e-01 1.22556232e-01
-7.43300676e-01 3.33600879e-01 2.66877860e-01 4.35670853e-01
2.38813102e-01 -1.52072704e+00 -4.45487857e-01 -7.46899769e-02
1.29393265e-01 -1.84345081e-01 1.55693114e-01 8.48116577e-01
-4.14135516e-01 5.68515599e-01 1.40845776e-01 -3.16662580e-01
-1.66846347e+00 5.17644346e-01 -6.99167550e-02 -7.55524933e-01
-4.47450131e-01 9.34771776e-01 -7.94417933e-02 -7.44792998e-01
5.62123358e-01 6.37140647e-02 -4.79632527e-01 6.77202523e-01
3.43940347e-01 5.34482300e-01 2.74351627e-01 -6.23593748e-01
-3.53447706e-01 4.67180699e-01 -4.85541224e-01 1.59032822e-01
1.46546805e+00 9.75171383e-03 -1.55705214e-01 5.83767116e-01
1.72028434e+00 -2.10909575e-01 -7.57865489e-01 -2.95383126e-01
4.11839992e-01 -2.22339094e-01 2.17330739e-01 -1.05341566e+00
-1.12302196e+00 7.81363249e-01 1.08652663e+00 1.70960903e-01
1.10581028e+00 -2.60794144e-02 9.38276052e-01 7.27423504e-02
3.15777272e-01 -1.25027561e+00 4.90067229e-02 2.95146316e-01
3.99445266e-01 -1.11858296e+00 1.05727687e-01 -1.18402667e-01
-5.10174394e-01 9.41243887e-01 4.31740642e-01 3.21662098e-01
9.83076811e-01 1.16392083e-01 2.76033700e-01 -2.40595564e-01
-8.38077843e-01 1.49338329e-02 -3.05079278e-02 2.61315048e-01
6.54105604e-01 1.40390962e-01 -7.37956882e-01 3.97828668e-01
1.41441822e-01 1.87922150e-01 1.18704893e-01 9.81771111e-01
-4.32717055e-01 -1.58925807e+00 4.14077146e-03 1.01800370e+00
-9.51300085e-01 -2.48537466e-01 -1.70862004e-01 7.77338803e-01
4.33268696e-01 8.91659260e-01 -2.51064956e-01 -4.99782503e-01
1.90779716e-01 4.18502122e-01 1.55525565e-01 -6.12295985e-01
-5.25255144e-01 -3.27201486e-02 3.29437047e-01 -2.19376951e-01
-3.30471367e-01 -7.66303599e-01 -1.26201737e+00 -2.92272717e-01
-6.84178233e-01 2.27318808e-01 7.87335336e-01 8.95054698e-01
1.47708058e-01 3.59097987e-01 8.70404184e-01 -4.65617210e-01
-7.53103256e-01 -1.16906655e+00 -8.40716958e-01 2.56817043e-01
-1.25127407e-02 -7.99746871e-01 -3.50924373e-01 -3.84716749e-01] | [9.518860816955566, 4.382863521575928] |
efee30ae-ab21-46c7-9772-2d1b75c59c1c | spcnet-stepwise-point-cloud-completion | 2209.01746 | null | https://arxiv.org/abs/2209.01746v1 | https://arxiv.org/pdf/2209.01746v1.pdf | SPCNet: Stepwise Point Cloud Completion Network | How will you repair a physical object with large missings? You may first recover its global yet coarse shape and stepwise increase its local details. We are motivated to imitate the above physical repair procedure to address the point cloud completion task. We propose a novel stepwise point cloud completion network (SPCNet) for various 3D models with large missings. SPCNet has a hierarchical bottom-to-up network architecture. It fulfills shape completion in an iterative manner, which 1) first infers the global feature of the coarse result; 2) then infers the local feature with the aid of global feature; and 3) finally infers the detailed result with the help of local feature and coarse result. Beyond the wisdom of simulating the physical repair, we newly design a cycle loss %based training strategy to enhance the generalization and robustness of SPCNet. Extensive experiments clearly show the superiority of our SPCNet over the state-of-the-art methods on 3D point clouds with large missings. | ['Mingqiang Wei', 'Fu Lee Wang', 'Weiming Wang', 'Jun Wang', 'Zhe Zhu', 'Xuequan Lu', 'Honghua Chen', 'Fei Hu'] | 2022-09-05 | null | null | null | null | ['point-cloud-completion'] | ['computer-vision'] | [-1.42534629e-01 1.36200354e-01 1.59104660e-01 -1.32241979e-01
-5.32294273e-01 -2.30813906e-01 2.66965896e-01 1.18568614e-02
1.35740250e-01 3.12371165e-01 -1.81372330e-01 -4.09230053e-01
-2.26412207e-01 -7.86130011e-01 -1.20509529e+00 -5.03549755e-01
-8.48334432e-02 5.31222463e-01 3.39870185e-01 -6.79768100e-02
3.85603011e-01 1.18207848e+00 -1.60671139e+00 -3.85596156e-02
1.02919853e+00 1.17903745e+00 4.25196379e-01 3.14439356e-01
-5.95458634e-02 3.89436066e-01 -1.92209229e-01 -8.25371221e-02
3.92118126e-01 4.88358140e-01 -8.14988792e-01 3.50101411e-01
3.14775050e-01 -5.28768420e-01 -1.63720250e-01 8.30425203e-01
2.82248557e-01 1.68814003e-01 4.94433790e-01 -1.18397081e+00
-9.31905210e-01 -5.30695654e-02 -6.00799561e-01 -3.07521224e-01
1.33896887e-01 2.64438063e-01 7.27149367e-01 -1.39841938e+00
5.03794551e-01 1.18694115e+00 8.61482620e-01 4.21808004e-01
-9.88209844e-01 -5.65950096e-01 4.42003936e-01 -2.21924126e-01
-1.55900192e+00 -8.36648941e-02 9.27319348e-01 -3.26253951e-01
9.67020690e-01 5.48271351e-02 5.23180962e-01 5.13174295e-01
2.00689331e-01 7.51719296e-01 4.72329259e-01 -9.07993689e-02
2.68270314e-01 -3.98791701e-01 7.50809833e-02 8.47843587e-01
1.81858152e-01 4.12259102e-01 -3.06638926e-01 -2.30740368e-01
1.04420042e+00 3.63858432e-01 -2.66672462e-01 -6.14306867e-01
-1.02509916e+00 5.03557861e-01 8.40411782e-01 -1.32072300e-01
-5.52571237e-01 3.61574560e-01 1.71515718e-02 1.93204865e-01
6.12777650e-01 2.26841360e-01 -6.38956845e-01 3.58124077e-01
-7.37020850e-01 4.11923677e-01 5.35242260e-01 1.39988661e+00
1.08124280e+00 1.80443481e-01 1.28675103e-01 7.02773035e-01
1.51295155e-01 2.69783586e-01 -2.40157530e-01 -1.18500447e+00
4.58696425e-01 8.35165381e-01 4.74343598e-01 -8.11883926e-01
-3.31892699e-01 -4.62157607e-01 -1.16840458e+00 6.58051074e-01
-1.91080928e-01 1.57678291e-01 -1.21235931e+00 1.46162879e+00
6.31370604e-01 3.91564220e-01 -8.02749246e-02 7.68148124e-01
7.90394068e-01 5.56758463e-01 -8.87003615e-02 -1.17657809e-02
1.04554892e+00 -8.41378689e-01 -1.70045242e-01 3.04707065e-02
2.91511089e-01 -4.97092545e-01 1.01143813e+00 3.92449379e-01
-1.27633262e+00 -8.63635778e-01 -1.17204821e+00 -2.08579391e-01
-1.30902350e-01 2.47974694e-01 7.47033417e-01 -5.83526939e-02
-1.27511942e+00 1.13286567e+00 -8.81537139e-01 -4.66601513e-02
4.64960515e-01 4.83837306e-01 -4.80296463e-01 -2.90892750e-01
-5.80985606e-01 6.24654651e-01 2.22417399e-01 5.55751860e-01
-1.03525734e+00 -1.04304862e+00 -7.40282178e-01 2.37212002e-01
4.38804686e-01 -1.32014894e+00 1.23732686e+00 -2.04270810e-01
-1.17649817e+00 4.95980263e-01 -2.23570332e-01 -7.84651190e-02
3.98764551e-01 -3.03968638e-01 1.76609084e-02 1.37305006e-01
2.39626095e-01 9.21227634e-01 1.06789541e+00 -1.88988996e+00
-5.98328173e-01 -3.83848459e-01 -3.14637534e-02 1.16837688e-01
2.53753096e-01 -4.60034788e-01 -6.63904846e-01 -5.14493406e-01
8.22613895e-01 -7.74495244e-01 -1.74577132e-01 5.31184733e-01
-5.13399005e-01 -5.96218586e-01 8.42028260e-01 -5.76987088e-01
7.20841050e-01 -2.18166780e+00 8.61294642e-02 1.88175499e-01
5.47128499e-01 5.23793958e-02 -1.43961817e-01 4.50735420e-01
-3.03822368e-01 2.27017805e-01 -3.86689425e-01 -8.35531890e-01
8.55431368e-04 3.55349064e-01 -4.88699079e-01 2.65088707e-01
3.80115956e-01 7.35247850e-01 -7.18322754e-01 -1.62794739e-01
1.71857998e-01 3.50566447e-01 -6.72634244e-01 2.80949086e-01
-2.84553379e-01 2.61690021e-01 -5.93562841e-01 9.80960608e-01
1.28589427e+00 -4.31720853e-01 -3.15135628e-01 -3.06373805e-01
-2.17812881e-01 -9.72841904e-02 -1.12018383e+00 1.75826955e+00
-2.76593059e-01 -7.89011270e-02 4.08727556e-01 -6.39498830e-01
1.07335591e+00 2.26803735e-01 4.96944398e-01 -3.00628483e-01
-2.00093359e-01 1.80937722e-01 -4.62930471e-01 -3.64460170e-01
7.51256526e-01 -3.61091495e-01 1.62681475e-01 1.31578356e-01
-2.12998882e-01 -3.55241865e-01 -4.73760366e-01 3.30025911e-01
1.10843980e+00 4.04353648e-01 -2.00244382e-01 -5.56922145e-02
4.39288527e-01 -1.75085783e-01 6.70714676e-01 6.39508963e-01
9.21511501e-02 8.69070530e-01 3.70222718e-01 -6.81575835e-01
-1.06831193e+00 -1.33161175e+00 3.77124287e-02 5.04068792e-01
4.85602617e-01 -3.28626111e-02 -3.42536718e-01 -5.85415125e-01
4.64344382e-01 7.10248053e-01 -5.25791883e-01 -1.87326029e-01
-5.02409339e-01 -1.04441985e-01 2.32156768e-01 7.80131876e-01
6.28088474e-01 -1.22909415e+00 -1.89162597e-01 2.31297687e-01
-6.69716671e-02 -9.84914660e-01 -3.58999610e-01 5.68004064e-02
-1.37347853e+00 -9.37001228e-01 -3.30450475e-01 -8.59968603e-01
9.62530136e-01 6.58846796e-01 1.11258602e+00 7.98241496e-01
2.16098040e-01 2.20376685e-01 -4.20628995e-01 -2.26313502e-01
-1.60676867e-01 -2.80413358e-03 2.09020004e-01 -1.28747776e-01
-1.53376132e-01 -1.14146554e+00 -5.91691732e-01 2.85166264e-01
-1.05571508e+00 2.67808020e-01 6.52535379e-01 6.11399233e-01
8.26988757e-01 3.12662780e-01 4.82871562e-01 -4.42340285e-01
3.69506061e-01 -2.93174267e-01 -5.79239547e-01 1.11924030e-01
-4.73148048e-01 -6.02151558e-04 5.44591367e-01 -2.49813065e-01
-9.92708802e-01 1.86976939e-01 -3.76029462e-01 -1.12093806e+00
-1.50078893e-01 3.66540700e-01 -2.01581717e-01 -2.26822525e-01
2.68389583e-01 2.52177775e-01 -1.87613249e-01 -9.56379890e-01
1.23364151e-01 1.09811433e-01 9.11141872e-01 -9.33611035e-01
1.16265368e+00 6.81724608e-01 1.48786962e-01 -5.13353646e-01
-7.61273563e-01 -4.00943249e-01 -7.55049050e-01 -8.59519988e-02
7.26837218e-01 -9.53449726e-01 -1.32595110e+00 6.90904200e-01
-1.53325546e+00 -1.87250078e-01 -2.85270035e-01 7.00574964e-02
-4.72612292e-01 5.44788837e-01 -7.14342296e-01 -6.41375065e-01
-4.79033828e-01 -9.16698992e-01 1.39507997e+00 -8.57983902e-02
3.53014708e-01 -5.97814977e-01 -4.27829921e-01 7.85047784e-02
3.94425988e-02 4.44706112e-01 1.15777683e+00 -8.80783349e-02
-1.29188836e+00 -2.10795209e-01 -4.62972462e-01 4.49108064e-01
4.02790420e-02 1.04680425e-02 -6.49273038e-01 -4.63947117e-01
2.57654250e-01 5.14658019e-02 7.37694085e-01 3.29396963e-01
1.50974250e+00 -1.91417292e-01 -3.29157084e-01 9.15744305e-01
1.62358713e+00 -2.33527690e-01 8.01478863e-01 1.14034839e-01
9.15526927e-01 2.73644984e-01 5.79111814e-01 3.97500157e-01
4.39788818e-01 2.04481110e-01 1.14149785e+00 -9.18977335e-02
-9.22685489e-02 -6.61160767e-01 3.42318378e-02 9.83165264e-01
-2.04546154e-01 -1.02122732e-01 -8.18632603e-01 8.01615894e-01
-1.82234991e+00 -6.02697611e-01 -1.02737300e-01 1.95280445e+00
3.53732586e-01 2.73135841e-01 -4.83764321e-01 -4.16370966e-02
7.48675585e-01 6.01671934e-02 -6.82280660e-01 -1.24923937e-01
2.31307134e-01 4.71949168e-02 2.64971823e-01 4.83243853e-01
-5.85341573e-01 9.51489270e-01 5.41724110e+00 8.20453405e-01
-8.59695733e-01 -1.00449510e-01 2.36428007e-01 2.35531375e-01
-4.61115122e-01 3.14743698e-01 -7.16043770e-01 3.09650928e-01
1.43498763e-01 2.96774566e-01 5.17449081e-01 1.02678621e+00
5.10418452e-02 -1.53917223e-01 -1.27748835e+00 8.17194104e-01
-2.54503340e-01 -1.65492225e+00 4.59251076e-01 -1.20746279e-02
6.64514065e-01 2.71456242e-01 -1.72986001e-01 2.99627990e-01
3.93687874e-01 -8.27912569e-01 8.92119050e-01 8.30377579e-01
9.78948057e-01 -7.71320999e-01 3.80419910e-01 8.45131516e-01
-1.46838057e+00 -3.80444489e-02 -5.90122044e-01 8.52725655e-03
1.69697553e-01 5.69764733e-01 -7.52907693e-01 9.98979449e-01
7.93789566e-01 6.61355019e-01 -4.39614236e-01 1.10576463e+00
-2.85855889e-01 1.58021986e-01 -5.21084368e-01 4.35917050e-01
1.03436776e-01 -2.40563259e-01 7.12414563e-01 5.08059859e-01
5.37805438e-01 3.09708208e-01 2.14999944e-01 1.30699027e+00
-2.70791024e-01 -4.94211584e-01 -5.09143949e-01 5.72778702e-01
7.07494617e-01 1.01646733e+00 -5.11300981e-01 -3.06973487e-01
-1.49011180e-01 8.93540442e-01 6.15225911e-01 5.09212255e-01
-4.81083840e-01 -2.43788287e-01 5.23633480e-01 2.93282330e-01
6.87267840e-01 -3.94844115e-01 -6.01171553e-01 -9.03079331e-01
3.85392666e-01 -2.66204327e-01 -2.00069118e-02 -1.56716669e+00
-1.41784966e+00 5.92430234e-01 -2.07369076e-03 -1.50504470e+00
3.05424899e-01 -3.37068468e-01 -9.45468783e-01 8.45920086e-01
-1.60480213e+00 -1.38203657e+00 -5.96151233e-01 6.00388706e-01
5.95211446e-01 1.85445100e-01 5.70904315e-01 5.32116368e-02
-2.38333255e-01 3.30603778e-01 -2.26251245e-01 -1.99888095e-01
1.15350716e-01 -9.49650586e-01 8.05444658e-01 8.37154210e-01
-6.35931671e-01 7.09467530e-01 4.60248917e-01 -1.01846361e+00
-1.38800108e+00 -1.28175926e+00 8.23954821e-01 -3.48992884e-01
3.10623497e-01 -1.65819079e-01 -1.21399117e+00 7.09850132e-01
-2.68331379e-01 1.34675443e-01 -3.56447101e-01 1.62046123e-02
-3.98579866e-01 -1.82974502e-01 -1.35093379e+00 5.31435907e-01
1.27062273e+00 -3.49571824e-01 -5.97635567e-01 2.72696555e-01
1.39603329e+00 -6.51477098e-01 -8.85397017e-01 8.43126893e-01
1.21705323e-01 -9.07208800e-01 1.07442081e+00 -2.12643877e-01
4.55202669e-01 -6.51640356e-01 -1.07646056e-01 -1.15037155e+00
-6.63087964e-01 -6.19997621e-01 -1.39844730e-01 9.38829303e-01
1.02420069e-01 -6.35193288e-01 8.56143236e-01 5.97268760e-01
-9.69106078e-01 -9.63431954e-01 -1.01674306e+00 -7.57134318e-01
8.79358798e-02 -6.61572576e-01 1.08803773e+00 7.60325015e-01
-5.09679556e-01 2.82431208e-02 -2.47103974e-01 7.79733539e-01
6.99101567e-01 3.53812814e-01 7.43880033e-01 -1.54161024e+00
1.15481772e-01 3.00815497e-02 -4.74980520e-03 -1.45522416e+00
-5.01675084e-02 -7.30198562e-01 2.79608250e-01 -1.65764081e+00
5.23583852e-02 -7.82545507e-01 7.00734649e-03 6.07910395e-01
-7.63796344e-02 -2.86426060e-02 1.27084866e-01 5.27894914e-01
-4.37607706e-01 8.30664277e-01 1.67040598e+00 1.03755608e-01
-2.09984511e-01 1.96699008e-01 -6.78611100e-01 6.86457634e-01
5.90476871e-01 -5.78979194e-01 -2.70572960e-01 -6.49239182e-01
2.79892266e-01 5.01792610e-01 9.23500180e-01 -1.10130906e+00
4.03600335e-01 -1.23500600e-01 4.04947400e-01 -1.43694353e+00
5.27160287e-01 -9.82313573e-01 2.04101562e-01 2.45484829e-01
2.97824264e-01 2.45637953e-01 1.97255298e-01 7.39214242e-01
-8.53516906e-03 -1.16782457e-01 6.32486224e-01 -3.01503927e-01
-5.28809726e-01 9.37598825e-01 5.68183325e-02 -5.51712811e-01
6.17167354e-01 -4.86874998e-01 -2.05505595e-01 -9.88623351e-02
-9.75311697e-01 4.10492361e-01 7.47098267e-01 4.52379793e-01
1.04820716e+00 -1.43948162e+00 -4.58078921e-01 4.97726142e-01
7.59923533e-02 8.35461020e-01 5.18464625e-01 4.95371103e-01
-5.88493466e-01 9.18482095e-02 5.00337258e-02 -8.11421394e-01
-8.32809567e-01 9.37387943e-01 3.14101458e-01 -1.42471939e-01
-9.41022694e-01 8.51749241e-01 4.65222001e-01 -7.96039164e-01
2.74416029e-01 -6.68286026e-01 3.27078670e-01 -5.38494647e-01
3.44261080e-02 2.81488568e-01 3.01380068e-01 -2.39030972e-01
-8.90277103e-02 7.30539441e-01 -1.49933234e-01 3.06906879e-01
1.64414287e+00 -1.12433143e-01 -3.27066898e-01 -7.41008893e-02
1.08435905e+00 -3.16208601e-01 -1.59635186e+00 -2.53519058e-01
-3.97563964e-01 -5.11289477e-01 -7.73832574e-02 -6.01467550e-01
-1.14960122e+00 7.30890095e-01 2.57315040e-01 7.83969555e-03
1.13614154e+00 2.63668913e-02 9.63532686e-01 3.75236154e-01
5.62797904e-01 -8.70107055e-01 9.56402197e-02 6.07469082e-01
1.33071423e+00 -8.14160109e-01 1.19403355e-01 -6.58105195e-01
-2.54911453e-01 9.18438077e-01 7.64322221e-01 -4.03868884e-01
7.72248626e-01 6.83021694e-02 -5.11683166e-01 -7.10591137e-01
-7.92505920e-01 6.17515258e-02 1.72631502e-01 6.32966042e-01
-5.57623267e-01 -1.34001538e-01 3.73206586e-01 4.80984658e-01
-1.20584376e-01 2.20760018e-01 2.45968223e-01 8.74278545e-01
-5.46091557e-01 -7.64124215e-01 -4.90156859e-01 2.58426160e-01
1.37161583e-01 -8.02144557e-02 -4.94221710e-02 9.14012849e-01
2.61149317e-01 6.11636698e-01 1.11274086e-01 -6.21752858e-01
5.37553489e-01 -1.31352827e-01 2.23192826e-01 -6.09763384e-01
-5.02014995e-01 -7.34688565e-02 -2.27179974e-01 -5.19272923e-01
-7.66721517e-02 -5.04004419e-01 -1.43392289e+00 -6.90781534e-01
-2.95632213e-01 4.54264581e-02 6.13666058e-01 8.82817328e-01
6.62751198e-01 5.37942469e-01 7.83245206e-01 -1.42606008e+00
-4.90214944e-01 -8.69806468e-01 -4.63122368e-01 2.79484421e-01
6.29144311e-01 -7.63913274e-01 -4.57360506e-01 -1.88391745e-01] | [8.420385360717773, -3.6807920932769775] |
58e265b0-3067-4113-a328-734f2b129624 | condenseunet-a-memory-efficient-condensely | 2004.02249 | null | https://arxiv.org/abs/2004.02249v1 | https://arxiv.org/pdf/2004.02249v1.pdf | CondenseUNet: A Memory-Efficient Condensely-Connected Architecture for Bi-ventricular Blood Pool and Myocardium Segmentation | With the advent of Cardiac Cine Magnetic Resonance (CMR) Imaging, there has been a paradigm shift in medical technology, thanks to its capability of imaging different structures within the heart without ionizing radiation. However, it is very challenging to conduct pre-operative planning of minimally invasive cardiac procedures without accurate segmentation and identification of the left ventricle (LV), right ventricle (RV) blood-pool, and LV-myocardium. Manual segmentation of those structures, nevertheless, is time-consuming and often prone to error and biased outcomes. Hence, automatic and computationally efficient segmentation techniques are paramount. In this work, we propose a novel memory-efficient Convolutional Neural Network (CNN) architecture as a modification of both CondenseNet, as well as DenseNet for ventricular blood-pool segmentation by introducing a bottleneck block and an upsampling path. Our experiments show that the proposed architecture runs on the Automated Cardiac Diagnosis Challenge (ACDC) dataset using half (50%) the memory requirement of DenseNet and one-twelfth (~ 8%) of the memory requirements of U-Net, while still maintaining excellent accuracy of cardiac segmentation. We validated the framework on the ACDC dataset featuring one healthy and four pathology groups whose heart images were acquired throughout the cardiac cycle and achieved the mean dice scores of 96.78% (LV blood-pool), 93.46% (RV blood-pool) and 90.1% (LV-Myocardium). These results are promising and promote the proposed methods as a competitive tool for cardiac image segmentation and clinical parameter estimation that has the potential to provide fast and accurate results, as needed for pre-procedural planning and/or pre-operative applications. | ['Cristian A. Linte', 'S. M. Kamrul Hasan'] | 2020-04-05 | null | null | null | null | ['cardiac-segmentation'] | ['medical'] | [ 2.39513367e-01 1.08527869e-01 1.62837952e-01 -2.13950202e-01
-3.22346538e-01 -5.98327458e-01 1.14630848e-01 3.28441858e-01
-5.95262885e-01 6.91292882e-01 -2.37771302e-01 -6.84777915e-01
3.61752138e-02 -7.03134239e-01 -9.29293111e-02 -7.37699807e-01
-3.44544828e-01 7.80445158e-01 2.45092154e-01 3.62764537e-01
3.72097199e-03 8.90754580e-01 -8.81706297e-01 -3.47303301e-02
9.27747905e-01 9.68778908e-01 1.59145907e-01 8.83622468e-01
1.26662463e-01 7.54430532e-01 -3.30725878e-01 -1.27735406e-01
4.10053939e-01 -5.26555538e-01 -7.35841215e-01 6.67148307e-02
2.91025400e-01 -6.02121055e-01 -2.06056722e-02 7.02578187e-01
8.50056112e-01 -7.59384930e-02 4.57214028e-01 -5.40927291e-01
1.31771341e-01 5.18707454e-01 -4.33971345e-01 6.52720630e-01
-5.59599400e-01 2.43060753e-01 3.23360324e-01 -5.93320310e-01
6.12267137e-01 4.02759045e-01 8.63080025e-01 4.44068730e-01
-8.66798162e-01 -4.80235070e-01 -4.79041457e-01 -2.52797604e-01
-1.16887629e+00 -2.08180428e-01 5.28190076e-01 -6.04533076e-01
7.25779772e-01 2.87273288e-01 8.77396822e-01 2.78899968e-01
4.65961963e-01 2.75619566e-01 9.49976623e-01 -2.31473207e-01
1.79143980e-01 -1.24584675e-01 1.82113379e-01 7.99555600e-01
4.23507601e-01 -2.13456620e-02 1.45626590e-01 -1.47535279e-01
1.13168895e+00 1.09305918e-01 -4.42480683e-01 -4.07198399e-01
-1.37841237e+00 6.08721852e-01 4.30392116e-01 6.52766049e-01
-4.87941056e-01 2.05088302e-01 7.91210473e-01 -2.34665230e-01
2.16538340e-01 2.19739929e-01 -2.86076128e-01 -1.42929524e-01
-1.20193148e+00 -7.01464829e-04 5.94943106e-01 4.77497846e-01
2.53524721e-01 8.90346989e-02 -1.45631030e-01 7.81525373e-01
4.40670364e-02 3.09423774e-01 5.76050282e-01 -1.00903594e+00
3.24450195e-01 4.81287658e-01 -1.13803782e-01 -1.05439436e+00
-7.43650913e-01 -8.46649706e-01 -1.51098609e+00 1.23628244e-01
5.42638004e-01 -3.23981196e-01 -1.04391015e+00 1.29799485e+00
5.27558088e-01 8.05275366e-02 -2.80541539e-01 1.23900843e+00
7.37373531e-01 2.34441131e-01 2.52149493e-01 -4.71047074e-01
1.44307792e+00 -7.67376065e-01 -5.05096078e-01 3.11608985e-02
9.64148700e-01 -5.89811683e-01 6.54786170e-01 1.11807965e-01
-1.26740706e+00 -4.53975260e-01 -1.07824433e+00 2.22524151e-01
1.72835365e-01 2.30144799e-01 7.57787406e-01 7.80778587e-01
-1.00289845e+00 9.15018380e-01 -1.16765058e+00 -4.33029272e-02
7.82156408e-01 4.43683565e-01 -3.05843383e-01 8.97821132e-03
-9.26034212e-01 6.77070558e-01 3.84968698e-01 3.77068728e-01
-7.53200710e-01 -1.15035272e+00 -5.36687553e-01 3.43349501e-02
1.35354817e-01 -9.42940414e-01 8.80732775e-01 -7.53093719e-01
-1.33813083e+00 1.02201188e+00 3.92249346e-01 -6.17244661e-01
7.33491600e-01 4.18057032e-02 1.01201339e-02 5.87755382e-01
-9.87292305e-02 5.04029930e-01 3.86415958e-01 -8.19572687e-01
-3.74593019e-01 -5.42468309e-01 -2.71213591e-01 1.37845352e-01
9.40419734e-02 7.86428079e-02 -1.64334431e-01 -6.85658991e-01
4.70412701e-01 -9.44323480e-01 -4.68498051e-01 1.82137698e-01
-2.43269786e-01 4.97076452e-01 6.54320061e-01 -1.06575036e+00
1.15691388e+00 -1.94504273e+00 -2.12565750e-01 2.53343344e-01
6.89562023e-01 6.87514126e-01 4.30584699e-01 -2.29812041e-01
-1.60755038e-01 2.78416485e-01 -6.94664955e-01 2.68725511e-02
-6.51478946e-01 9.25702900e-02 2.57341981e-01 7.19247520e-01
-1.21488631e-01 8.97151589e-01 -7.41189241e-01 -7.91120529e-01
4.21176225e-01 6.40768528e-01 -3.63587916e-01 2.76805833e-02
3.90011758e-01 8.54860544e-01 -4.33638632e-01 6.58549607e-01
6.69598341e-01 -2.88133174e-01 5.35190225e-01 -2.49254987e-01
-1.04223400e-01 -1.08955689e-01 -9.11393285e-01 1.58632219e+00
-2.60801226e-01 3.20583910e-01 4.07192260e-01 -1.19038165e+00
7.65122712e-01 6.16170228e-01 9.67637539e-01 -7.93046057e-01
4.91326690e-01 4.70317185e-01 4.85194176e-01 -3.59511644e-01
-1.02369577e-01 -4.79154825e-01 2.83858955e-01 5.43414712e-01
-1.67672545e-01 7.08265975e-03 2.48721406e-01 2.35728398e-02
1.06638491e+00 -9.64216664e-02 2.34326422e-01 -5.58408558e-01
7.10349321e-01 -1.20584918e-02 7.54258454e-01 4.58603621e-01
-6.12715304e-01 9.03097093e-01 3.70163053e-01 -1.01492107e+00
-1.07556021e+00 -8.79751146e-01 -2.58222222e-01 1.80721492e-01
-1.08924381e-01 8.63331407e-02 -9.47748661e-01 -5.89427173e-01
-4.12837178e-01 1.15580700e-01 -2.33115360e-01 2.36316338e-01
-1.12967312e+00 -8.62180471e-01 5.83404481e-01 6.93250239e-01
6.19763732e-01 -1.10161853e+00 -1.40794206e+00 5.34353018e-01
-2.29018942e-01 -1.05317140e+00 -2.06789359e-01 7.03774244e-02
-1.58185983e+00 -1.14629197e+00 -1.06039393e+00 -6.37945473e-01
7.74231493e-01 -1.04602441e-01 1.39821959e+00 4.81806070e-01
-7.67351687e-01 -2.64125347e-01 -6.65544048e-02 -3.85152772e-02
-4.52951998e-01 6.76349103e-02 -2.97338098e-01 -3.24549764e-01
-3.03422838e-01 -7.59574413e-01 -1.06501591e+00 1.83893561e-01
-6.93952799e-01 3.09902221e-01 5.35215318e-01 7.80095756e-01
7.91245043e-01 -2.91607320e-01 4.21957970e-01 -1.07375681e+00
2.85114825e-01 -2.63866037e-01 -6.43225431e-01 9.06550139e-02
-6.86672747e-01 -4.71439511e-01 6.84480131e-01 -6.00999445e-02
-7.76474714e-01 1.93308592e-01 -1.28357112e-01 -3.62519115e-01
-2.13249281e-01 4.43850011e-01 2.84954041e-01 -1.06945887e-01
4.08664644e-01 7.51982555e-02 1.07701845e-01 -1.95335522e-01
2.89532822e-02 3.56649429e-01 7.00800955e-01 -4.13105339e-01
4.36679423e-01 5.34255266e-01 3.45836967e-01 -5.64423323e-01
-3.74152482e-01 -3.47785771e-01 -9.78445351e-01 -3.53555113e-01
1.10348344e+00 -5.41316569e-01 -5.42630732e-01 3.78012449e-01
-1.01785254e+00 -1.95865065e-01 -3.81871074e-01 5.58746636e-01
-3.60948414e-01 6.22492552e-01 -8.36242855e-01 -4.87152606e-01
-1.15974486e+00 -1.28986919e+00 5.53233624e-01 1.76698714e-01
-2.14884385e-01 -1.05987036e+00 -9.78847966e-02 5.38117468e-01
7.90156960e-01 7.96532631e-01 1.18280780e+00 -5.10589600e-01
-6.40727997e-01 -2.51931787e-01 -3.16044748e-01 4.48334813e-01
6.57189190e-02 -2.01126277e-01 -6.10816836e-01 -2.79104412e-01
6.43825904e-02 4.08021547e-02 4.64337319e-01 9.46996748e-01
1.19728553e+00 1.16632201e-01 -1.64721474e-01 8.46648276e-01
1.50393093e+00 5.79035044e-01 5.96777201e-01 -4.24400810e-03
7.24486113e-01 2.66646385e-01 3.84436160e-01 5.58088303e-01
1.55892923e-01 2.87599385e-01 4.05982018e-01 -4.88824785e-01
-1.81869715e-01 4.19220865e-01 -3.39504659e-01 1.12310553e+00
-3.98310065e-01 8.23075026e-02 -1.40295351e+00 6.67505026e-01
-1.36092341e+00 -6.71835482e-01 -3.56187761e-01 2.19309044e+00
7.93237984e-01 1.36954486e-02 6.08426929e-02 2.05084980e-01
6.91950858e-01 -1.87478676e-01 -3.91327083e-01 -4.66580808e-01
3.55500877e-01 6.41036868e-01 4.84314203e-01 3.19463432e-01
-1.24042594e+00 3.90412331e-01 5.86606407e+00 2.79226780e-01
-1.37782681e+00 2.40497231e-01 1.29357255e+00 2.99682282e-02
2.00137094e-01 -8.95822197e-02 -3.80151957e-01 3.65727633e-01
9.11330462e-01 1.70300797e-01 9.81200263e-02 5.53635240e-01
2.21302614e-01 -2.73310065e-01 -7.95949161e-01 8.68678510e-01
-2.20715448e-01 -1.58854032e+00 -3.39516908e-01 -7.89917037e-02
4.93074417e-01 -4.98290658e-02 -4.07925934e-01 -2.49106307e-02
-4.55059201e-01 -1.04056704e+00 2.48846516e-01 3.58346522e-01
1.01723087e+00 -7.57744670e-01 1.20559561e+00 4.19020891e-01
-1.03044045e+00 2.05198795e-01 -1.22585535e-01 1.45225465e-01
2.10256338e-01 8.92481327e-01 -9.93730724e-01 4.09501523e-01
6.11349642e-01 4.12325203e-01 -2.40633681e-01 9.74675834e-01
3.45886976e-01 6.14953220e-01 -2.95660704e-01 3.91081691e-01
5.76124862e-02 -3.37836027e-01 3.67468208e-01 1.01926315e+00
2.56180853e-01 2.59728432e-01 2.04928562e-01 9.60195243e-01
-1.44629195e-01 2.98927575e-01 -2.19674423e-01 8.60423222e-02
1.49361700e-01 1.57378626e+00 -1.46500659e+00 -5.68433583e-01
-1.59112021e-01 5.49347103e-01 -1.08047843e-01 2.32467218e-03
-9.29800808e-01 -2.69101888e-01 1.25818700e-01 4.55684513e-01
-1.12258792e-02 -9.12306607e-02 -7.48923242e-01 -8.14629912e-01
3.79654467e-02 -7.37022817e-01 2.38918781e-01 -2.42158726e-01
-7.33740032e-01 7.37193644e-01 -2.43029848e-01 -1.08828866e+00
-1.66803986e-01 -3.48438740e-01 -4.98988032e-01 1.10421133e+00
-1.36091340e+00 -8.56186986e-01 -5.88568151e-01 2.15610385e-01
1.84142366e-01 4.80817929e-02 8.91752183e-01 6.57245696e-01
-3.93857658e-01 2.41336599e-01 -2.27256030e-01 3.64842266e-01
2.41663560e-01 -1.23626649e+00 1.47648960e-01 7.32251287e-01
-4.41324264e-01 6.65443897e-01 1.50272816e-01 -5.77334702e-01
-9.36013877e-01 -1.08926952e+00 7.11364031e-01 9.47201625e-02
1.72089294e-01 2.64189448e-02 -6.93621278e-01 2.83301800e-01
2.94081848e-02 6.66747689e-01 7.95881629e-01 -5.01147747e-01
4.66429025e-01 -1.14700839e-01 -1.37927377e+00 2.88355708e-01
5.50189078e-01 -6.21192195e-02 -1.45451084e-01 2.22311094e-01
2.67547965e-01 -7.93630302e-01 -1.24918997e+00 7.36679614e-01
5.30788541e-01 -1.17397249e+00 9.56990957e-01 -1.45925418e-01
4.49692637e-01 -2.56053776e-01 3.34095478e-01 -6.96620882e-01
-1.26109004e-01 -4.02156174e-01 -1.53795794e-01 7.74026573e-01
1.43980861e-01 -4.63197798e-01 1.01031876e+00 6.92387342e-01
-4.78579551e-01 -1.18140674e+00 -1.00398731e+00 -1.82149634e-01
2.08550125e-01 -2.90665835e-01 8.55113119e-02 9.24822569e-01
-5.73976755e-01 -7.33197704e-02 4.90625808e-03 -1.03873745e-01
6.87453508e-01 6.80138096e-02 3.32425952e-01 -1.31904471e+00
-1.86067730e-01 -4.52793628e-01 -3.57513607e-01 -4.70661223e-01
-3.78614485e-01 -9.62345123e-01 -1.15829274e-01 -1.59970224e+00
3.66183639e-01 -7.70856678e-01 -4.04767811e-01 2.82481074e-01
-1.03121653e-01 4.45983797e-01 1.41438350e-01 2.36041591e-01
-1.31286711e-01 -7.19061196e-02 1.69522154e+00 1.20610617e-01
-1.26011580e-01 1.56722944e-02 -2.36918882e-01 7.48717070e-01
9.53874409e-01 -4.27259594e-01 -4.02822822e-01 -2.55877703e-01
-1.45058915e-01 6.22531354e-01 4.46130067e-01 -1.26489794e+00
3.56722884e-02 3.75970334e-01 6.32482171e-01 -7.31308699e-01
-1.26503825e-01 -8.51262748e-01 2.75471300e-01 9.92047727e-01
-3.89926247e-02 2.09323362e-01 1.93205655e-01 -6.23903908e-02
-2.67456353e-01 -1.45601943e-01 1.26910484e+00 -5.93989074e-01
-2.52733707e-01 5.15908003e-01 -4.80481356e-01 2.24502787e-01
1.19735897e+00 -4.44905758e-01 9.76638347e-02 7.64982253e-02
-9.71671939e-01 -9.82746631e-02 9.25977826e-02 -1.68035805e-01
6.96101189e-01 -8.78705025e-01 -6.60162866e-01 1.99075460e-01
-5.43477356e-01 3.95116508e-01 7.23806143e-01 1.47162533e+00
-1.50470066e+00 4.68515962e-01 -3.41444582e-01 -9.28018332e-01
-1.12735939e+00 1.18470997e-01 8.34231317e-01 -7.11174130e-01
-9.70953047e-01 7.29295909e-01 4.79214005e-02 -3.19483548e-01
4.35910597e-02 -5.44701278e-01 -1.13783933e-01 -2.28533864e-01
3.32323909e-01 4.86462206e-01 2.86222577e-01 -5.08970559e-01
-4.36660498e-01 5.57855785e-01 5.77357076e-02 3.48978221e-01
1.17446387e+00 -3.40041122e-03 -3.89300466e-01 8.79349336e-02
9.58695114e-01 -4.13276076e-01 -8.88077617e-01 1.14869699e-01
-5.57942651e-02 -2.44358361e-01 5.03309786e-01 -1.05292022e+00
-1.58184278e+00 1.18402529e+00 1.09071469e+00 -4.59127091e-02
1.23696268e+00 -4.87528861e-01 1.14513731e+00 -2.57111043e-01
4.04160321e-01 -7.89669991e-01 -3.42583120e-01 1.53777376e-01
4.37548429e-01 -1.03676224e+00 1.83354154e-01 -4.66503292e-01
-5.24471760e-01 1.29747558e+00 4.56217885e-01 -2.06373006e-01
7.58637547e-01 4.41441596e-01 3.74479115e-01 -2.84168929e-01
-2.85279989e-01 3.41194779e-01 1.76954508e-01 3.37231427e-01
9.56056774e-01 8.92209187e-02 -4.27632809e-01 3.52014989e-01
2.51982510e-01 2.03174949e-01 3.53521258e-01 1.04426658e+00
-1.68964922e-01 -7.79073656e-01 -7.34402388e-02 5.58232963e-01
-1.06884563e+00 -2.27837518e-01 1.05081595e-01 8.62307191e-01
2.93480575e-01 4.50062722e-01 1.27297128e-02 1.96129769e-01
7.95329064e-02 1.30894165e-02 3.47862184e-01 -4.96091992e-01
-1.04619181e+00 2.94030875e-01 -2.37058267e-01 -5.72303474e-01
-3.98688406e-01 -4.97930318e-01 -1.63277173e+00 -2.48284206e-01
-1.91847399e-01 -5.38754873e-02 7.23582506e-01 8.25521052e-01
3.51929724e-01 9.12495017e-01 3.47224206e-01 -6.95505798e-01
-3.05383414e-01 -8.90636742e-01 -6.43993616e-01 2.31092155e-01
1.00300640e-01 -3.53256792e-01 -2.50673443e-02 2.05727488e-01] | [14.205362319946289, -2.4651436805725098] |
915e9abd-84a8-4eef-93b3-9bc5e1664fff | seqnet-learning-descriptors-for-sequence | 2102.11603 | null | https://arxiv.org/abs/2102.11603v2 | https://arxiv.org/pdf/2102.11603v2.pdf | SeqNet: Learning Descriptors for Sequence-based Hierarchical Place Recognition | Visual Place Recognition (VPR) is the task of matching current visual imagery from a camera to images stored in a reference map of the environment. While initial VPR systems used simple direct image methods or hand-crafted visual features, recent work has focused on learning more powerful visual features and further improving performance through either some form of sequential matcher / filter or a hierarchical matching process. In both cases the performance of the initial single-image based system is still far from perfect, putting significant pressure on the sequence matching or (in the case of hierarchical systems) pose refinement stages. In this paper we present a novel hybrid system that creates a high performance initial match hypothesis generator using short learnt sequential descriptors, which enable selective control sequential score aggregation using single image learnt descriptors. Sequential descriptors are generated using a temporal convolutional network dubbed SeqNet, encoding short image sequences using 1-D convolutions, which are then matched against the corresponding temporal descriptors from the reference dataset to provide an ordered list of place match hypotheses. We then perform selective sequential score aggregation using shortlisted single image learnt descriptors from a separate pipeline to produce an overall place match hypothesis. Comprehensive experiments on challenging benchmark datasets demonstrate the proposed method outperforming recent state-of-the-art methods using the same amount of sequential information. Source code and supplementary material can be found at https://github.com/oravus/seqNet. | ['Michael Milford', 'Sourav Garg'] | 2021-02-23 | null | null | null | null | ['sequential-place-learning', 'sequential-place-recognition'] | ['robots', 'robots'] | [ 3.98007989e-01 -4.64762956e-01 -9.90277305e-02 -3.73922437e-01
-7.74353266e-01 -7.14098036e-01 9.24811244e-01 3.66626769e-01
-6.59163177e-01 3.18618715e-01 4.66002971e-02 1.05588436e-01
-1.18132465e-01 -8.12990665e-01 -7.60751069e-01 -5.38501263e-01
-2.95188963e-01 3.39194387e-01 7.61032462e-01 -3.50772530e-01
6.03195488e-01 7.63486087e-01 -2.08608222e+00 4.24219698e-01
4.58197296e-01 1.06484509e+00 8.24861944e-01 6.81495428e-01
5.02076447e-02 7.65555918e-01 -4.53569621e-01 1.35530541e-02
6.33366823e-01 -3.46893191e-01 -7.41493762e-01 -1.44697666e-01
7.51512468e-01 -1.77879736e-01 -4.88377303e-01 9.33649063e-01
5.12349606e-01 5.40630102e-01 2.72847325e-01 -1.21715617e+00
-6.29063129e-01 1.76298022e-01 -2.64949650e-01 2.45638102e-01
8.13932002e-01 3.96332234e-01 1.08463120e+00 -9.61437523e-01
9.98957753e-01 9.54444110e-01 6.20532870e-01 2.95641422e-01
-1.37492037e+00 -6.51974499e-01 7.13460967e-02 5.97619832e-01
-1.68846834e+00 -4.06989545e-01 5.22647142e-01 -5.04951239e-01
1.47570908e+00 2.41466612e-01 8.90886009e-01 8.55296552e-01
6.78409860e-02 6.81274652e-01 1.02395999e+00 -7.61777535e-02
1.90057546e-01 -2.06699893e-01 -2.56794959e-01 7.84276843e-01
-3.11399847e-01 4.66330975e-01 -8.06951582e-01 -2.11498011e-02
7.61244714e-01 3.91344130e-01 -2.41557822e-01 -4.58137065e-01
-1.45025790e+00 6.25021517e-01 1.06703937e+00 3.85451555e-01
-6.50256932e-01 5.34414947e-02 3.42542708e-01 4.06999022e-01
-5.97933009e-02 4.70041454e-01 -1.48715138e-01 1.67812705e-01
-1.41247690e+00 5.12193203e-01 4.27666396e-01 8.22039723e-01
1.29059494e+00 -2.73251504e-01 -5.15062690e-01 7.11480677e-01
5.01318201e-02 4.39340472e-01 5.07950306e-01 -7.67398715e-01
2.61309087e-01 6.18676603e-01 4.90085222e-03 -1.17638087e+00
-3.06448519e-01 -1.78612337e-01 -5.39006233e-01 5.54252326e-01
2.47607648e-01 4.79895473e-01 -1.36133492e+00 1.36848187e+00
1.01216130e-01 3.36524457e-01 -3.28083411e-02 1.06611538e+00
7.92201161e-01 8.97403896e-01 -1.37791246e-01 4.19283241e-01
1.16036010e+00 -1.07417095e+00 -1.23928107e-01 -3.40517938e-01
2.83924609e-01 -7.81396687e-01 7.31998980e-01 1.70777634e-01
-7.47509301e-01 -7.58281767e-01 -1.13902748e+00 -2.09084198e-01
-6.85516238e-01 -2.79311407e-02 2.83025026e-01 1.98522881e-02
-1.43438935e+00 8.41185093e-01 -7.08646059e-01 -5.88372231e-01
3.10049623e-01 5.58065712e-01 -8.77857089e-01 -1.86556712e-01
-1.02203441e+00 9.89957273e-01 5.10157168e-01 6.07525967e-02
-1.30067050e+00 -6.64850354e-01 -1.19328272e+00 -1.19783312e-01
8.43594670e-02 -4.60273832e-01 1.11399603e+00 -9.28443134e-01
-1.18510056e+00 9.90815580e-01 -1.66299075e-01 -6.88548028e-01
5.40439248e-01 2.62381602e-02 -9.32645053e-02 3.25462580e-01
3.61774355e-01 1.13108790e+00 8.27342451e-01 -9.73148108e-01
-1.00560677e+00 -2.47855589e-01 -3.78073123e-03 2.13336766e-01
1.87243789e-01 1.16662182e-01 -5.55875659e-01 -4.65218395e-01
-1.50898506e-03 -9.81115103e-01 -5.10831833e-01 1.15337335e-01
-8.50744918e-02 -1.17803477e-01 7.31600344e-01 -6.55341029e-01
9.19470429e-01 -2.11450148e+00 6.18147142e-02 3.99179041e-01
5.34530170e-02 4.15520996e-01 -4.64463025e-01 5.90708494e-01
-2.33344793e-01 -2.89245993e-01 -2.64529228e-01 -2.18208194e-01
-1.12180889e-01 5.32612763e-02 -1.92764863e-01 5.57978630e-01
3.42915952e-01 8.99260342e-01 -1.16159654e+00 -3.35183024e-01
7.15094447e-01 4.44356322e-01 -4.60852176e-01 3.66989017e-01
-1.10010728e-01 4.62445110e-01 -1.83477610e-01 6.25027239e-01
4.33103383e-01 2.87239309e-02 -2.51890182e-01 -1.42472476e-01
-6.21455431e-01 2.87682444e-01 -1.16936874e+00 2.00930405e+00
-4.60591882e-01 8.29915106e-01 -4.57562149e-01 -8.59635174e-01
1.30521691e+00 -2.04754192e-02 5.32861531e-01 -1.15999484e+00
-1.92507699e-01 2.40246728e-01 -1.94238961e-01 -3.44882488e-01
7.10193276e-01 1.31108627e-01 -1.17472626e-01 -1.10056652e-02
4.22941685e-01 -1.46929070e-01 2.41082266e-01 5.68655990e-02
1.43770218e+00 2.52836317e-01 3.72942477e-01 1.98006816e-02
5.65234125e-01 4.15501863e-01 4.16698217e-01 8.82513762e-01
-2.55452186e-01 1.00486994e+00 -3.55588049e-02 -8.13736379e-01
-1.37633049e+00 -1.01930296e+00 -6.86442032e-02 8.41751873e-01
3.66009921e-01 -6.46281838e-01 -3.10077161e-01 -4.07461196e-01
1.21449351e-01 3.51941764e-01 -8.60296369e-01 -3.30130979e-02
-5.88936567e-01 -2.46327326e-01 4.83728945e-01 5.25878251e-01
7.23560274e-01 -1.43456435e+00 -1.07489657e+00 1.78154692e-01
-9.90408957e-02 -9.18262720e-01 -3.95258605e-01 4.25815403e-01
-3.98077667e-01 -7.86591828e-01 -8.37155700e-01 -9.45940375e-01
5.49382508e-01 3.77118140e-01 8.51715505e-01 3.81613597e-02
-5.76162279e-01 1.31061912e-01 -5.23794234e-01 -5.38576534e-03
-5.50599024e-02 9.99969766e-02 -1.39877722e-01 1.03307508e-01
2.21379086e-01 -5.19725561e-01 -8.58838916e-01 3.02529722e-01
-8.72521162e-01 5.43947779e-02 6.95236564e-01 9.66178834e-01
8.47562969e-01 -3.39512199e-01 -6.74487054e-02 -2.45169669e-01
2.37970978e-01 -2.98056632e-01 -9.56449628e-01 1.52104691e-01
-2.89915651e-01 3.67726058e-01 7.04986453e-01 -4.47168887e-01
-6.65035605e-01 6.28418386e-01 -3.36908281e-01 -7.66806006e-01
-3.97326976e-01 3.74388576e-01 2.70260572e-01 -3.90257090e-01
6.97031915e-01 6.17738485e-01 -1.01149045e-01 -2.85432190e-01
3.29379767e-01 4.49693561e-01 8.09276879e-01 -2.78659374e-01
9.91218328e-01 5.14564276e-01 -9.42167416e-02 -6.19313478e-01
-5.10338426e-01 -9.05251682e-01 -8.76612067e-01 -3.12258124e-01
1.03333282e+00 -7.66016424e-01 -5.19309044e-01 3.74217004e-01
-9.71708417e-01 -5.95853150e-01 -1.75712720e-01 2.76499510e-01
-6.39291584e-01 -1.92582153e-03 -2.40062281e-01 -3.81627679e-01
-2.09056050e-01 -1.25801921e+00 1.40365946e+00 3.42071503e-01
-1.52871624e-01 -6.83648825e-01 4.08212095e-01 -3.49195860e-03
3.84339243e-01 3.65562677e-01 2.62910515e-01 -6.68930829e-01
-9.16720927e-01 -3.92113328e-01 -1.50946140e-01 1.48556635e-01
-2.01113790e-01 -5.52935898e-02 -8.11015487e-01 -3.26824725e-01
-5.66871166e-01 -2.90596306e-01 1.02734482e+00 8.62015486e-02
7.94457972e-01 -1.31070897e-01 -4.12392974e-01 9.26118314e-01
1.79643738e+00 2.35869572e-01 7.85188138e-01 8.34192753e-01
5.64980686e-01 4.41338509e-01 8.21184099e-01 5.62516093e-01
3.59060287e-01 1.07766712e+00 4.93021309e-01 -1.93660766e-01
-1.95912227e-01 -4.25530612e-01 3.96908551e-01 2.49151215e-01
-1.26122534e-01 2.89741069e-01 -1.26912594e+00 8.20077658e-01
-1.97040486e+00 -1.38531172e+00 1.62389308e-01 2.35082960e+00
5.07175624e-01 -1.29112571e-01 1.84075367e-02 -3.34124193e-02
7.16728926e-01 2.75048763e-01 -3.10216635e-01 -1.92254260e-01
-5.36489636e-02 3.82895023e-01 7.87913024e-01 4.25616235e-01
-1.20485187e+00 1.05368114e+00 5.47677517e+00 7.05693245e-01
-1.41268003e+00 -6.09938949e-02 1.33273959e-01 -3.30935828e-02
1.27392337e-01 2.13033468e-01 -7.46556044e-01 2.08943158e-01
7.08781719e-01 -2.42154494e-01 5.14305592e-01 8.60622108e-01
2.41646054e-03 -2.44209468e-01 -1.11651993e+00 1.25473368e+00
1.48866490e-01 -1.49866736e+00 6.70644119e-02 1.75215546e-02
6.66773379e-01 6.06738329e-01 -1.59040780e-03 1.59266219e-01
1.61705047e-01 -9.70155478e-01 1.07438648e+00 6.50436938e-01
7.18012571e-01 -6.36107266e-01 6.43631995e-01 1.00031987e-01
-1.65091753e+00 -3.33740562e-01 -4.25914884e-01 -2.68999897e-02
2.49898266e-02 3.19062546e-02 -9.48554873e-01 6.24394119e-01
1.05521739e+00 1.01853478e+00 -1.08151329e+00 1.66718709e+00
-1.37086436e-01 -3.66694033e-02 -3.46665174e-01 5.71534559e-02
6.92978859e-01 1.40394673e-01 5.34246624e-01 1.38332653e+00
3.57160896e-01 -1.08854853e-01 3.73392344e-01 7.12555587e-01
2.23582625e-01 -2.76017189e-02 -1.06336832e+00 3.54978204e-01
4.45153356e-01 1.51981056e+00 -8.11045647e-01 -3.92041892e-01
-4.82476205e-01 1.28967333e+00 4.40858662e-01 3.79018784e-02
-6.25199795e-01 -6.25148237e-01 7.10989952e-01 1.02108821e-01
6.47129774e-01 -3.37551028e-01 1.68892384e-01 -1.13697743e+00
-6.73899129e-02 -6.95176184e-01 4.25948322e-01 -8.14520299e-01
-9.23024833e-01 8.16858768e-01 -2.23917770e-03 -1.44497013e+00
-3.48237813e-01 -5.37610948e-01 -5.98780692e-01 8.67822111e-01
-1.47308075e+00 -1.10237825e+00 -6.52969003e-01 9.15313184e-01
6.85073793e-01 -1.27594516e-01 8.11589003e-01 3.15237850e-01
-1.90187335e-01 4.27442670e-01 -7.45058209e-02 3.19528759e-01
7.26412296e-01 -1.17056370e+00 7.78516591e-01 9.16769147e-01
3.58902395e-01 4.31667477e-01 5.74020803e-01 -5.91244102e-01
-1.22125888e+00 -1.24817252e+00 8.87773454e-01 -4.14791822e-01
6.08736753e-01 -5.96550763e-01 -8.00290644e-01 4.76036847e-01
2.92792737e-01 4.14075881e-01 3.02860886e-01 -4.37070787e-01
-4.86340433e-01 -2.79877067e-01 -9.93038833e-01 6.20467484e-01
1.10603893e+00 -7.12265909e-01 -5.87277591e-01 8.92819613e-02
4.76332814e-01 -4.53813076e-01 -7.73020744e-01 3.97913307e-01
6.84195757e-01 -1.04671097e+00 1.11798203e+00 -1.16687916e-01
2.75826693e-01 -8.82414103e-01 -2.45986179e-01 -1.06816232e+00
-5.02157688e-01 -3.68632108e-01 5.16363144e-01 9.49382842e-01
3.08455497e-01 -3.30702752e-01 5.80607533e-01 2.84150481e-01
-1.76729351e-01 -5.26681662e-01 -8.49793434e-01 -9.40766335e-01
-5.07360518e-01 -3.16262573e-01 6.00798845e-01 7.50375569e-01
-2.21633390e-01 9.09978375e-02 -1.76545769e-01 3.30314934e-01
6.15560710e-01 3.19473833e-01 6.89711630e-01 -9.39006329e-01
-8.84560049e-02 -4.89442617e-01 -1.11672664e+00 -6.71038985e-01
1.20540619e-01 -1.19342375e+00 4.10581768e-01 -1.62925732e+00
6.49713278e-02 -3.36159736e-01 -3.36137593e-01 8.53174865e-01
1.16264895e-01 5.29130340e-01 4.52218980e-01 3.36288124e-01
-8.08714032e-01 4.59862471e-01 8.31704915e-01 -2.36921772e-01
-2.42550075e-01 -3.15270364e-01 1.60409641e-02 2.05471367e-01
7.17895746e-01 -5.99229991e-01 -2.59492606e-01 -3.01262319e-01
-2.53833979e-02 -2.21572863e-03 8.17684531e-01 -1.46457028e+00
5.95132947e-01 -8.09711367e-02 6.07148349e-01 -6.64686382e-01
2.52961367e-01 -8.23800266e-01 3.93213958e-01 5.29274166e-01
-3.85886431e-01 4.86364543e-01 2.73769796e-01 2.77054667e-01
-4.95209903e-01 -1.27772063e-01 8.46416771e-01 -3.98952812e-01
-1.49128473e+00 4.73544329e-01 -2.28555650e-01 -3.29241633e-01
1.14866245e+00 -5.39868414e-01 -7.44197071e-02 -5.05627394e-02
-7.13547289e-01 2.35179171e-01 8.11850786e-01 7.12995231e-01
9.36539710e-01 -1.35416627e+00 -6.99293733e-01 2.35231131e-01
4.86997783e-01 -9.03978497e-02 2.16166809e-01 6.96294129e-01
-7.56940722e-01 3.78318250e-01 -7.55671263e-01 -7.55882919e-01
-1.17514157e+00 6.50731683e-01 4.54935789e-01 -8.93200114e-02
-9.33448851e-01 7.50935197e-01 1.31586626e-01 -4.66097653e-01
2.09415495e-01 -9.83663499e-02 -2.38047302e-01 2.02579439e-01
7.68846333e-01 7.78195336e-02 2.86248058e-01 -9.51049387e-01
-7.09607005e-01 6.20409250e-01 -5.73721901e-02 -3.15760314e-01
1.60140073e+00 -1.41904643e-02 -8.93822536e-02 1.65593237e-01
1.45623147e+00 -5.36178350e-01 -1.42722714e+00 -3.29533070e-01
1.76727667e-01 -7.06048548e-01 -6.40739948e-02 -4.99224007e-01
-9.36789334e-01 7.38963485e-01 8.07983935e-01 -1.42303541e-01
1.32297230e+00 1.03054650e-01 4.97552693e-01 5.08503258e-01
5.83217740e-01 -8.75762522e-01 1.70628086e-01 6.15559757e-01
9.64199603e-01 -1.10846496e+00 -3.66480857e-01 1.84309557e-01
-6.63364053e-01 1.16976583e+00 5.56533813e-01 -5.37213206e-01
4.12745118e-01 2.44693607e-01 8.00376385e-02 -1.94681689e-01
-7.82577157e-01 -7.51616359e-01 3.84275377e-01 7.65321732e-01
1.62007004e-01 -1.97527409e-01 1.43897846e-01 5.40811345e-02
-2.38619685e-01 -4.95313993e-03 2.01866418e-01 9.65138137e-01
-3.81178677e-01 -1.06762993e+00 -4.01561230e-01 2.81825095e-01
-9.30811316e-02 -2.57908344e-01 -2.56020188e-01 5.17211139e-01
3.47593158e-01 5.50533473e-01 2.25506425e-01 -6.76824450e-01
5.04452169e-01 -1.67138308e-01 4.12644565e-01 -5.94803035e-01
-9.43271756e-01 -3.02896261e-01 -1.62717998e-01 -1.09783864e+00
-3.47024947e-01 -1.03091848e+00 -1.05167091e+00 -1.13950700e-01
1.81278419e-02 -1.64107695e-01 5.44531226e-01 6.98570251e-01
3.59337449e-01 2.17949152e-01 5.05200326e-01 -1.55081677e+00
-3.60384658e-02 -6.34489596e-01 -2.08053187e-01 6.13182664e-01
5.84147453e-01 -6.11471653e-01 5.34010641e-02 7.66240433e-02] | [7.7754364013671875, -1.8471144437789917] |
ed3ba94b-c61e-45f6-8025-a3673263ca43 | the-spectacl-of-nonconvex-clustering-a | 1907.0068 | null | https://arxiv.org/abs/1907.00680v1 | https://arxiv.org/pdf/1907.00680v1.pdf | The SpectACl of Nonconvex Clustering: A Spectral Approach to Density-Based Clustering | When it comes to clustering nonconvex shapes, two paradigms are used to find the most suitable clustering: minimum cut and maximum density. The most popular algorithms incorporating these paradigms are Spectral Clustering and DBSCAN. Both paradigms have their pros and cons. While minimum cut clusterings are sensitive to noise, density-based clusterings have trouble handling clusters with varying densities. In this paper, we propose \textsc{SpectACl}: a method combining the advantages of both approaches, while solving the two mentioned drawbacks. Our method is easy to implement, such as spectral clustering, and theoretically founded to optimize a proposed density criterion of clusterings. Through experiments on synthetic and real-world data, we demonstrate that our approach provides robust and reliable clusterings. | ['Sibylle Hess', 'Wouter Duivesteijn', 'Katharina Morik', 'Philipp Honysz'] | 2019-07-01 | null | null | null | null | ['clustering-algorithms-evaluation'] | ['methodology'] | [-3.51660311e-01 -3.70174080e-01 -2.42654830e-01 -2.21695676e-01
-6.99053645e-01 -6.65973604e-01 3.70618820e-01 7.93427378e-02
-2.75535792e-01 6.21924758e-01 -1.71332911e-01 -1.64193645e-01
-5.67752063e-01 -8.14463913e-01 -2.33646438e-01 -9.81351733e-01
-1.89645104e-02 8.26628923e-01 4.51727957e-01 2.27478430e-01
3.73274714e-01 6.59900486e-01 -1.67113590e+00 -1.50157869e-01
1.45656431e+00 8.25864434e-01 2.85520464e-01 8.25528726e-02
-4.32897449e-01 9.09612998e-02 -4.14265186e-01 -9.33960229e-02
3.32807422e-01 -3.93188477e-01 -4.69251901e-01 5.34588039e-01
-3.69955540e-01 3.27364057e-01 2.71384895e-01 1.19589758e+00
3.86016458e-01 2.70584375e-01 1.14410484e+00 -1.35137796e+00
-5.65076053e-01 5.82100034e-01 -1.29551041e+00 1.28283113e-01
3.59022200e-01 -1.36036888e-01 5.22176266e-01 -8.46467733e-01
2.15552673e-01 1.23934448e+00 7.84142196e-01 3.15681756e-01
-1.31378579e+00 -4.76695746e-01 -1.69816930e-02 1.29019007e-01
-2.06109214e+00 -4.55612272e-01 8.03593576e-01 -3.64785254e-01
4.70116258e-01 3.66011709e-01 5.35876989e-01 4.60104227e-01
-1.95000038e-01 7.12258101e-01 1.21181226e+00 -3.45248580e-01
7.61827826e-01 2.06360698e-01 1.34293465e-02 2.89478362e-01
6.05725229e-01 -3.18562418e-01 1.11476645e-01 -5.01086652e-01
5.24335146e-01 4.28739004e-02 -3.51901323e-01 -7.01352060e-01
-9.12036240e-01 8.59227598e-01 1.39974475e-01 6.08825386e-01
-2.67678767e-01 -8.59723017e-02 -8.64375569e-03 -1.98023513e-01
4.38060105e-01 1.06300637e-01 1.09755464e-01 -9.13329720e-02
-1.26812565e+00 4.20762822e-02 7.58240044e-01 1.21494949e+00
7.11692393e-01 -1.68369189e-01 4.02784735e-01 1.03101134e+00
6.76018775e-01 6.15288556e-01 4.27602947e-01 -9.32949007e-01
2.57509768e-01 7.53661215e-01 1.04732461e-01 -1.36902761e+00
-4.36284900e-01 9.29640085e-02 -1.04232121e+00 6.12941980e-02
4.47816759e-01 -7.06565082e-02 -8.01114559e-01 1.27859354e+00
4.71349537e-01 3.97748828e-01 9.50647518e-03 8.13051462e-01
5.00349104e-01 7.56453276e-01 -1.21719323e-01 -7.82269895e-01
7.33339369e-01 -3.32233429e-01 -9.21071172e-01 2.82827646e-01
2.54177690e-01 -8.65795672e-01 8.34693909e-01 4.31485593e-01
-1.24967122e+00 -1.76838785e-01 -8.52490664e-01 4.88893956e-01
-4.05976981e-01 -7.58433193e-02 3.52432549e-01 9.59684432e-01
-1.31499970e+00 3.40711564e-01 -1.03226662e+00 -4.67079669e-01
4.31569181e-02 5.74026823e-01 1.58081204e-01 4.95321862e-02
-5.27137756e-01 5.83701670e-01 5.59283555e-01 9.75636542e-02
-9.99629796e-02 -2.65304297e-01 -5.19945085e-01 3.03923991e-02
3.46541613e-01 -4.42406148e-01 6.14800036e-01 -6.36159658e-01
-1.30536032e+00 5.98134100e-01 -2.66888201e-01 -2.36571789e-01
4.94761854e-01 2.08719432e-01 -3.79273921e-01 1.57375038e-01
2.01327741e-01 3.10741425e-01 5.66543818e-01 -1.71794367e+00
-5.18238127e-01 -3.53281736e-01 -6.82791829e-01 2.00685978e-01
-3.73200178e-01 -2.47480609e-02 -6.55658662e-01 -4.32441264e-01
5.68930030e-01 -7.98833728e-01 -5.31813025e-01 -4.37710017e-01
-5.53942382e-01 -4.04746354e-01 1.12884259e+00 -1.41447544e-01
1.56451452e+00 -2.19390178e+00 1.84349298e-01 9.05610979e-01
8.01008195e-02 1.17757037e-01 3.66630644e-01 4.55569416e-01
1.77509964e-01 2.90749103e-01 -7.98868120e-01 -2.27740198e-01
-2.61215214e-02 2.59703130e-01 1.79814145e-01 7.97003150e-01
-2.19075829e-01 3.69460136e-01 -8.27372551e-01 -1.04622197e+00
4.83597010e-01 5.18380702e-01 -4.80051816e-01 -1.14469722e-01
-9.41410102e-03 1.11750536e-01 -4.46301162e-01 7.90207803e-01
9.35093403e-01 1.11065535e-02 3.80033404e-01 1.18161872e-01
-3.14023674e-01 -4.52786058e-01 -2.01281738e+00 1.19641554e+00
1.70567274e-01 1.57342955e-01 4.58675444e-01 -1.17973912e+00
1.19171250e+00 2.57592648e-01 9.77001429e-01 -7.84133151e-02
2.96399683e-01 5.37097156e-01 -1.27608538e-01 -4.22674090e-01
3.54923487e-01 -2.86444753e-01 2.90195271e-02 3.92691702e-01
-2.51834810e-01 -2.18355656e-01 5.58167696e-01 2.05467552e-01
8.05722117e-01 -2.77773470e-01 3.83541554e-01 -8.42124283e-01
5.11963427e-01 6.86438009e-02 6.26208901e-01 3.83032352e-01
-1.80896789e-01 8.04166973e-01 1.93339720e-01 -1.57626477e-04
-8.37429702e-01 -1.25787449e+00 -2.88717151e-01 3.92272890e-01
5.37439823e-01 -2.56994098e-01 -1.02426565e+00 -3.31790477e-01
5.97391389e-02 5.11327922e-01 -3.01999092e-01 1.60968751e-01
-5.59298277e-01 -1.18175578e+00 2.30827555e-01 3.10552090e-01
3.30768555e-01 -5.50797224e-01 -4.52315331e-01 1.84327736e-01
-2.57837445e-01 -6.48669899e-01 -4.09805417e-01 7.51720965e-02
-9.16213810e-01 -1.29117775e+00 -7.81910419e-01 -8.76521111e-01
7.71690965e-01 5.98163784e-01 9.29645360e-01 1.44159719e-01
-2.50431389e-01 5.50377488e-01 -5.33797801e-01 -1.10514045e-01
-2.20381618e-01 7.31099695e-02 2.75878042e-01 2.48769075e-01
5.10063052e-01 -7.49992311e-01 -4.14094657e-01 6.48665428e-01
-9.23296332e-01 -5.32570302e-01 4.50784415e-01 3.43236655e-01
9.36906099e-01 6.67612433e-01 6.25715256e-01 -7.21329272e-01
9.60144341e-01 -7.99185276e-01 -5.97974062e-01 2.32106939e-01
-6.64099932e-01 -2.98205882e-01 6.23044968e-01 -2.72337407e-01
-9.86624956e-01 4.03248459e-01 3.65209393e-02 -7.28211462e-01
-3.42866421e-01 3.77130538e-01 -3.38673264e-01 2.26297826e-02
4.96851444e-01 1.38354689e-01 7.97088891e-02 -5.60507476e-01
5.13996840e-01 8.25707078e-01 5.69258034e-01 -8.41200948e-01
8.49805057e-01 7.29376197e-01 -1.21861026e-01 -1.14722836e+00
-2.35632151e-01 -1.05780542e+00 -8.37629855e-01 -4.34485584e-01
8.32465708e-01 -3.61153662e-01 -7.28196025e-01 1.81449920e-01
-7.68529892e-01 1.23723716e-01 -1.95862427e-01 3.73866290e-01
-7.12868392e-01 7.24583805e-01 -2.59488821e-01 -1.21596420e+00
9.44841877e-02 -1.09902418e+00 8.36078525e-01 3.99037957e-01
-5.50215878e-03 -1.22959518e+00 9.73759219e-02 -4.74686734e-02
2.26691991e-01 6.63562298e-01 7.14310884e-01 -4.40640569e-01
-3.31125081e-01 -1.36071108e-02 -9.40268189e-02 -2.05247030e-02
3.62326115e-01 5.53729832e-01 -4.55253601e-01 -4.02371734e-01
2.17948511e-01 1.47599459e-01 5.68608940e-01 7.30835617e-01
1.17460001e+00 -2.49187514e-01 -7.84958422e-01 4.75014150e-01
1.72734725e+00 6.31221116e-01 7.17083752e-01 7.75132626e-02
5.45193732e-01 5.75598359e-01 7.06420660e-01 4.85716045e-01
3.83279115e-01 5.94014466e-01 2.73153216e-01 -2.11236820e-01
2.30016783e-01 1.52134404e-01 3.48159522e-02 1.20654035e+00
-2.54036844e-01 -2.26999968e-01 -1.26887918e+00 6.38452947e-01
-2.13322306e+00 -1.20058513e+00 -6.63458228e-01 2.15442348e+00
7.25962043e-01 3.96458013e-03 6.03580058e-01 5.23622751e-01
1.16806090e+00 -3.49885732e-01 -3.12019855e-01 -2.36066282e-01
-2.45807186e-01 -1.38293862e-01 4.18594867e-01 4.66896325e-01
-1.20454884e+00 5.95532000e-01 7.14233065e+00 1.11224508e+00
-6.90264106e-01 -1.24769926e-01 6.72475100e-01 -7.76300579e-02
-3.87740225e-01 -1.55677140e-01 -5.61054051e-01 8.12556446e-01
6.86061144e-01 -3.91554296e-01 3.02906007e-01 7.70385683e-01
3.97698402e-01 -3.53512079e-01 -7.39707649e-01 1.08405685e+00
5.78964502e-02 -1.14171505e+00 -1.73969150e-01 2.62009591e-01
8.01067531e-01 -3.71658534e-01 -7.56815746e-02 -9.66725200e-02
6.05315983e-01 -1.07119405e+00 6.00385666e-01 4.76980895e-01
5.86682081e-01 -1.13288176e+00 7.19117999e-01 5.48497319e-01
-1.54031444e+00 2.69725006e-02 -5.40316582e-01 2.23482698e-01
3.33873481e-01 1.06906533e+00 -5.99314451e-01 6.30670249e-01
8.59289765e-01 5.74093103e-01 -3.55884045e-01 1.54038167e+00
3.10080051e-01 4.63301301e-01 -7.43945837e-01 -1.82506919e-01
3.43666166e-01 -9.21262503e-01 4.75938916e-01 1.28388894e+00
5.55252910e-01 2.87778765e-01 6.01617754e-01 1.09661829e+00
4.38453525e-01 3.58254433e-01 -8.16678524e-01 4.16524142e-01
1.01764309e+00 1.12429285e+00 -1.57959175e+00 -2.01586872e-01
-2.36867070e-01 4.90568936e-01 5.74156567e-02 3.12538415e-01
-9.24435616e-01 -2.75252223e-01 1.82384267e-01 3.98597628e-01
1.11536115e-01 -3.68077844e-01 -6.39246047e-01 -7.39741623e-01
-1.14697330e-01 -6.32430196e-01 5.61699510e-01 -2.95801967e-01
-1.41682053e+00 2.46685490e-01 5.26908875e-01 -1.35639536e+00
-2.06164923e-03 -4.58976656e-01 -6.62566960e-01 3.71523649e-01
-9.58750904e-01 -7.13495791e-01 -1.54486522e-01 8.87492478e-01
4.15483564e-01 4.10384219e-03 4.32616919e-01 4.44513232e-01
-5.89048803e-01 1.39331207e-01 5.90494514e-01 -1.94687143e-01
3.90619487e-01 -1.52550232e+00 -2.52525479e-01 8.47969711e-01
-5.52206859e-02 6.93753660e-01 6.97001278e-01 -5.84283948e-01
-1.25137651e+00 -9.21559572e-01 4.92031544e-01 -1.08750537e-01
4.68905330e-01 -2.31298238e-01 -9.13340569e-01 2.53203481e-01
1.37038529e-01 -4.69564527e-01 8.92713726e-01 3.51515152e-02
3.68717723e-02 8.26576874e-02 -1.45328772e+00 5.16585410e-01
8.37283134e-01 1.05370633e-01 -5.51116168e-01 2.06737921e-01
2.15387061e-01 4.11695335e-03 -1.03079557e+00 4.20932025e-01
1.33166507e-01 -1.40496969e+00 1.01413667e+00 3.31841946e-01
-1.82454929e-01 -6.70360923e-01 -1.22483149e-01 -1.39858997e+00
-3.03560853e-01 -6.22285903e-01 1.33209020e-01 1.54659319e+00
3.70035201e-01 -6.38720214e-01 7.81724811e-01 3.24512273e-01
-1.52252272e-01 -7.30978906e-01 -1.12950349e+00 -1.10579872e+00
1.97347596e-01 -2.58259773e-01 5.96097589e-01 1.28730738e+00
4.18408573e-01 8.30153748e-02 -4.50542085e-02 1.77215450e-02
8.93390834e-01 9.53673124e-02 6.04041398e-01 -1.57987177e+00
1.03604287e-01 -7.93719232e-01 -2.51646399e-01 -6.83569431e-01
-1.07995868e-02 -6.93754673e-01 2.26757631e-01 -1.70985067e+00
2.56011546e-01 -9.18801069e-01 1.27414882e-01 5.90027347e-02
2.55834125e-02 1.37344584e-01 9.13778022e-02 5.40757835e-01
-6.50457442e-01 4.59395856e-01 8.76396596e-01 -3.09037399e-02
-4.90973920e-01 8.12201425e-02 -4.36193854e-01 9.50357437e-01
1.01846838e+00 -3.72894496e-01 -6.32171631e-01 -6.68227226e-02
-1.89137533e-01 -1.11845717e-01 -1.80172160e-01 -1.17208552e+00
4.67828721e-01 -3.60656470e-01 7.65381530e-02 -9.81425583e-01
1.81658402e-01 -1.13636887e+00 4.86963809e-01 2.58651584e-01
2.22717389e-01 1.96259722e-01 4.74914499e-02 6.92705572e-01
-2.91256040e-01 -1.96186781e-01 1.23555267e+00 -2.18382925e-01
-4.27613020e-01 5.31056896e-02 -4.93293911e-01 6.27536140e-03
1.62527442e+00 -7.38054812e-01 4.72649001e-02 -3.36728066e-01
-8.77594709e-01 5.05231619e-01 7.61386693e-01 7.55630881e-02
6.12197995e-01 -1.44401145e+00 -4.05574709e-01 -4.48327251e-02
-2.83180386e-01 1.79908887e-01 -1.57039359e-01 1.06034470e+00
-6.50380015e-01 2.97982067e-01 1.05623432e-01 -9.90294695e-01
-1.11306798e+00 9.94591892e-01 2.64798373e-01 1.87172577e-01
-3.99874121e-01 4.84886676e-01 -8.74328315e-02 -4.09206182e-01
3.68136644e-01 -2.10820869e-01 -2.37364396e-01 1.48549959e-01
2.38988400e-01 9.19590056e-01 -5.56335114e-02 -7.86456525e-01
-6.20589375e-01 8.12058389e-01 3.68474722e-01 -1.78374007e-01
1.09037983e+00 -3.55695218e-01 -3.39302003e-01 5.33247411e-01
8.89270008e-01 4.84722219e-02 -8.86098206e-01 2.66071588e-01
5.29050350e-01 -6.17585421e-01 -3.06841731e-01 -2.93202519e-01
-1.16509819e+00 4.97042239e-01 3.79332483e-01 1.07220161e+00
1.27652347e+00 7.10615292e-02 4.70047504e-01 4.89070006e-02
5.54799199e-01 -1.42830729e+00 -8.61866623e-02 1.36156216e-01
4.78806257e-01 -8.81682515e-01 4.65871543e-02 -8.87865901e-01
-4.99264479e-01 8.94275308e-01 6.27055705e-01 -2.25626022e-01
1.01601517e+00 6.02603614e-01 -5.79751842e-02 -2.85212576e-01
-3.45039487e-01 -3.89157295e-01 5.70421293e-02 8.50136817e-01
3.54796141e-01 1.66677356e-01 -7.21921861e-01 2.60694683e-01
-1.19886473e-01 -3.82953554e-01 5.06044745e-01 8.61787796e-01
-8.22462201e-01 -9.17197764e-01 -1.03134954e+00 4.15576518e-01
-2.63002068e-01 3.64529788e-01 -6.42782569e-01 9.63101983e-01
2.03019828e-01 1.29715395e+00 1.43663272e-01 -2.46448740e-01
2.42122501e-01 -1.29469097e-01 1.33330211e-01 -3.81794989e-01
-2.05436140e-01 6.21529996e-01 -2.91646004e-01 -3.39025855e-01
-9.53196228e-01 -9.44009185e-01 -1.45745468e+00 -6.44885957e-01
-6.80107236e-01 5.02091646e-01 4.20005709e-01 6.96267486e-01
1.33983657e-01 6.88963085e-02 7.23107278e-01 -8.14901948e-01
-1.76657230e-01 -5.72016180e-01 -9.96370256e-01 3.21521848e-01
-2.51671582e-01 -8.77501249e-01 -4.71120596e-01 1.15084469e-01] | [7.539121627807617, 4.52930212020874] |
f507ba6f-a197-40d7-8c26-6247725c8789 | scibertsum-extractive-summarization-for | 2201.08495 | null | https://arxiv.org/abs/2201.08495v1 | https://arxiv.org/pdf/2201.08495v1.pdf | SciBERTSUM: Extractive Summarization for Scientific Documents | The summarization literature focuses on the summarization of news articles. The news articles in the CNN-DailyMail are relatively short documents with about 30 sentences per document on average. We introduce SciBERTSUM, our summarization framework designed for the summarization of long documents like scientific papers with more than 500 sentences. SciBERTSUM extends BERTSUM to long documents by 1) adding a section embedding layer to include section information in the sentence vector and 2) applying a sparse attention mechanism where each sentences will attend locally to nearby sentences and only a small number of sentences attend globally to all other sentences. We used slides generated by the authors of scientific papers as reference summaries since they contain the technical details from the paper. The results show the superiority of our model in terms of ROUGE scores. | ['C Lee Giles', 'Athar Sefid'] | 2022-01-21 | null | null | null | null | ['extractive-summarization'] | ['natural-language-processing'] | [-1.47015467e-01 4.77412164e-01 -2.35900789e-01 -1.68119624e-01
-1.18421042e+00 -5.82621455e-01 7.22596288e-01 7.20915079e-01
-3.69588554e-01 9.79446054e-01 1.36317265e+00 -4.43455055e-02
5.94967455e-02 -5.16043901e-01 -8.79816175e-01 -4.45996255e-01
1.28452899e-02 1.01296656e-01 6.18424118e-02 -1.37590319e-01
7.70299196e-01 1.98065355e-01 -7.86662877e-01 5.01269877e-01
8.79374802e-01 3.87017459e-01 3.88705283e-01 1.11715662e+00
-5.07277489e-01 8.63708496e-01 -1.30602181e+00 -4.56321180e-01
-1.06255300e-01 -5.03310561e-01 -8.89131308e-01 5.61096631e-02
9.03484762e-01 -5.36240995e-01 -6.04880393e-01 8.12054396e-01
6.83439791e-01 3.33493203e-01 6.14471614e-01 -5.41080236e-01
-7.84649730e-01 1.32479954e+00 -7.60207593e-01 7.77189434e-01
4.32690680e-01 -2.19760165e-01 1.37741184e+00 -7.37361550e-01
9.75976944e-01 1.12199342e+00 4.38251436e-01 5.51886082e-01
-8.93130422e-01 -4.76826467e-02 3.65468621e-01 8.23162496e-02
-8.07972431e-01 -6.13853335e-01 7.29228854e-01 -1.73098713e-01
1.40425766e+00 6.07661366e-01 5.82263291e-01 1.16440415e+00
9.28292990e-01 9.97095406e-01 -2.18468443e-01 -1.08666718e-02
3.07152867e-01 -9.16676968e-02 8.08192074e-01 2.18250006e-01
7.29966462e-01 -9.21445727e-01 -5.47234714e-01 -5.04827872e-02
3.23924333e-01 -7.25986362e-02 -2.78098404e-01 4.21273828e-01
-1.49419832e+00 8.92021477e-01 5.27873158e-01 2.73740023e-01
-7.17635870e-01 3.09507877e-01 1.02143013e+00 2.10350886e-01
8.43822420e-01 9.20001626e-01 -2.78582275e-02 1.41958013e-01
-1.37957764e+00 4.87347424e-01 8.32311869e-01 1.11941612e+00
3.03922981e-01 2.13785008e-01 -8.87302458e-01 7.36824691e-01
-2.09951520e-01 2.25240514e-01 5.00418723e-01 -1.01501131e+00
9.32740211e-01 3.97041231e-01 7.79734552e-02 -1.03491271e+00
-2.16142535e-01 -8.52623641e-01 -1.03482842e+00 -6.38867378e-01
-4.60919261e-01 -6.43802941e-01 -6.01444066e-01 1.17546344e+00
-1.54358253e-01 -1.93265885e-01 2.59836137e-01 6.26910150e-01
1.72730231e+00 1.48649263e+00 -1.94665119e-01 -4.87430513e-01
1.16755855e+00 -1.37551546e+00 -1.05983961e+00 -1.31186560e-01
4.48969394e-01 -7.69775748e-01 5.55911899e-01 -1.23881614e-02
-1.56152034e+00 -3.83515358e-01 -1.21188593e+00 -5.14365673e-01
-2.70992190e-01 1.83560029e-01 3.36345792e-01 -2.62034833e-01
-1.25387192e+00 6.52830124e-01 -5.42661548e-01 -4.59034204e-01
6.27372444e-01 -4.08300869e-02 -3.53943497e-01 8.91034305e-02
-8.13315868e-01 7.35555887e-01 3.22285861e-01 3.70150874e-03
-5.84462106e-01 -7.62022853e-01 -9.52961326e-01 7.29369819e-01
6.12521768e-02 -1.11856282e+00 1.32366312e+00 -2.85381436e-01
-1.35881054e+00 3.22707057e-01 -5.03135026e-01 -7.31113970e-01
2.98603624e-01 -4.67034847e-01 3.10937092e-02 5.84291458e-01
4.19840544e-01 5.39365947e-01 3.14969599e-01 -9.32443440e-01
-4.26260054e-01 -9.28304642e-02 9.73905548e-02 4.72138792e-01
-3.00723404e-01 -1.77828353e-02 -5.30018747e-01 -8.00638795e-01
-1.38029024e-01 -2.67769337e-01 -3.99581343e-01 -7.13983774e-01
-1.10840034e+00 -3.56257081e-01 5.76551318e-01 -9.33744907e-01
1.54352319e+00 -1.86057174e+00 4.46881056e-01 -4.09161061e-01
2.56295472e-01 -6.15527406e-02 -4.81587619e-01 1.02711797e+00
9.10422802e-02 2.91412294e-01 7.66670927e-02 -5.98709702e-01
-1.69111952e-01 -2.16269001e-01 -6.47054076e-01 2.62653023e-01
1.06764443e-01 1.13140798e+00 -9.99375939e-01 -4.86327559e-01
-1.45647332e-01 8.55371729e-02 -4.14835542e-01 4.92278822e-02
-1.07932612e-01 3.28785628e-02 -7.45418191e-01 1.94281161e-01
5.20678878e-01 -2.07633659e-01 -3.79659057e-01 -7.24749416e-02
-3.89248669e-01 6.95131600e-01 -5.16326129e-01 1.77496767e+00
-2.42844060e-01 1.10837960e+00 3.56562831e-03 -8.54894102e-01
5.83856583e-01 4.60866123e-01 3.55561525e-01 -9.43462830e-03
1.44286990e-01 -2.91729689e-01 -3.48536819e-01 -4.90073442e-01
1.08366406e+00 4.13274884e-01 -3.00534457e-01 6.42836511e-01
2.21549913e-01 -4.93991792e-01 9.24287260e-01 1.13608134e+00
1.21621323e+00 -4.23112810e-01 3.15028906e-01 -3.67743343e-01
3.01191062e-01 3.44268233e-02 3.03674519e-01 1.18919694e+00
1.63188770e-01 9.67189074e-01 9.75908279e-01 -2.28076279e-01
-1.32903171e+00 -8.90205085e-01 1.85227558e-01 9.04811203e-01
-7.16571733e-02 -8.57972324e-01 -8.03079069e-01 -5.83988786e-01
-8.28396454e-02 1.05029106e+00 -6.27156973e-01 5.55006824e-02
-5.22537351e-01 -5.85221171e-01 1.88432083e-01 5.39675295e-01
2.98379451e-01 -1.16261816e+00 -3.49790305e-01 2.41383493e-01
-2.21837297e-01 -7.60742545e-01 -1.04517543e+00 3.89192067e-02
-9.83437836e-01 -5.84328830e-01 -1.20881641e+00 -8.58862936e-01
7.27925062e-01 3.36078584e-01 8.92416358e-01 -3.23302418e-01
5.38311489e-02 2.25574732e-01 -3.25888276e-01 -6.89679086e-01
-2.98557758e-01 6.03756487e-01 -2.45620340e-01 -3.34024757e-01
-1.04299232e-01 -2.48573214e-01 -4.65763807e-01 -6.79653406e-01
-8.46699655e-01 1.28939033e-01 6.15670502e-01 7.02518225e-01
2.49600902e-01 -3.85042965e-01 9.91875768e-01 -1.04329813e+00
1.26335561e+00 -5.97067773e-01 -4.30532880e-02 1.75711185e-01
2.80762743e-03 -7.32864738e-02 6.55246615e-01 -1.72856241e-01
-9.71416891e-01 -5.41974664e-01 -1.71457946e-01 1.52229950e-01
2.72806346e-01 9.63713646e-01 2.02342197e-01 6.12710178e-01
5.44097602e-01 4.93747711e-01 -2.13773757e-01 -4.56798553e-01
3.09381843e-01 6.38940811e-01 6.13752306e-01 -4.30505909e-02
3.78745556e-01 1.92754269e-01 -3.66304874e-01 -1.28383148e+00
-1.13608253e+00 -7.19024956e-01 -3.82365316e-01 -1.43753543e-01
6.62211180e-01 -8.86582375e-01 -5.46799541e-01 4.57632765e-02
-1.71711290e+00 2.92084455e-01 -5.88553727e-01 5.66754460e-01
-2.60571510e-01 7.42352009e-01 -1.02821541e+00 -3.50571603e-01
-1.17714465e+00 -7.64034212e-01 1.14807010e+00 5.81626296e-01
-6.19971454e-01 -8.10279965e-01 1.55586213e-01 1.04866378e-01
4.86341655e-01 7.17370585e-02 6.43923223e-01 -8.30965579e-01
-3.17117363e-01 -5.76074779e-01 -1.57033846e-01 3.71924996e-01
6.51569515e-02 3.12004983e-01 -5.46255827e-01 -2.94607252e-01
-8.94715413e-02 2.57169511e-02 1.64804232e+00 1.03184855e+00
1.04984152e+00 -7.89908290e-01 -3.15233797e-01 1.58644199e-01
1.09854889e+00 -1.29251897e-01 6.69556618e-01 2.73150355e-01
5.90114534e-01 4.97945666e-01 -2.84947362e-02 4.96948868e-01
1.70950070e-01 -9.31501109e-03 9.95813757e-02 9.70731825e-02
-2.18841910e-01 -6.28437474e-02 4.86484289e-01 1.58934355e+00
2.25485116e-01 -8.20888638e-01 -3.75248671e-01 9.74177122e-01
-1.86013830e+00 -1.40658319e+00 -2.54510999e-01 1.70073843e+00
7.92501688e-01 1.19328767e-01 -2.52387702e-01 -3.97668093e-01
6.25316739e-01 8.75819027e-01 -3.66146773e-01 -8.21880877e-01
-3.99129659e-01 -1.83399290e-01 3.93232793e-01 6.46630168e-01
-1.00090647e+00 7.26547778e-01 6.95772362e+00 6.41070426e-01
-9.12764132e-01 -2.62501538e-01 4.98387963e-01 -5.96739650e-01
-5.66202402e-01 -1.52618378e-01 -7.59493291e-01 4.99573708e-01
9.15302157e-01 -1.00284243e+00 -2.66906679e-01 6.29500210e-01
5.24899900e-01 -4.12904382e-01 -8.82282376e-01 5.25887370e-01
4.44883406e-01 -2.09649730e+00 7.43496776e-01 -3.62949729e-01
1.26293623e+00 2.13997632e-01 -2.29833558e-01 2.35087380e-01
5.14640100e-03 -7.16272891e-01 4.54377681e-01 7.10171402e-01
2.96159863e-01 -8.24365914e-01 1.19022906e+00 3.65496844e-01
-5.60184121e-01 2.24198624e-01 -8.00973952e-01 1.27394684e-02
3.57471257e-01 8.90260816e-01 -8.04250956e-01 8.45880449e-01
4.56826836e-01 1.17690003e+00 -4.70919549e-01 1.30950534e+00
-1.96957797e-01 7.94555962e-01 -1.01896808e-01 -4.40507323e-01
5.68391383e-01 -1.91978484e-01 1.05951917e+00 1.63426542e+00
4.03436244e-01 -7.70959929e-02 -1.89677060e-01 5.07756650e-01
-6.04304790e-01 2.91371137e-01 -5.69138110e-01 -2.74238408e-01
2.69163042e-01 1.13761568e+00 -6.27809227e-01 -7.43369341e-01
-1.74483642e-01 1.00005949e+00 7.80581534e-02 6.57517135e-01
-3.23853016e-01 -1.14594138e+00 1.05847202e-01 -1.02104716e-01
5.33390284e-01 -9.65550262e-03 -6.27182066e-01 -1.29485226e+00
1.34391725e-01 -2.75548756e-01 1.89816400e-01 -9.38052714e-01
-1.13518035e+00 6.81739688e-01 -1.19801797e-01 -8.18886697e-01
-6.80811480e-02 8.30762461e-02 -1.39140141e+00 7.27165997e-01
-1.21284485e+00 -7.51032412e-01 1.29682627e-02 -3.35782170e-01
1.18441486e+00 -2.39832908e-01 6.20469391e-01 -3.00295055e-01
-9.10389006e-01 2.24320978e-01 7.24496186e-01 4.90031317e-02
7.39757836e-01 -1.29373503e+00 6.76940441e-01 7.36578166e-01
-2.00677827e-01 8.43743443e-01 1.08277667e+00 -8.12808752e-01
-1.14598238e+00 -1.11510777e+00 1.56241930e+00 -3.41648459e-01
5.03856540e-01 -1.36066914e-01 -9.60677922e-01 8.41908038e-01
1.10379052e+00 -7.21886337e-01 5.36593676e-01 1.35890037e-01
2.03549951e-01 -1.09944455e-01 -7.03222096e-01 7.44291842e-01
4.45438564e-01 -9.65184346e-02 -1.13526976e+00 6.73686743e-01
9.63605165e-01 -3.47437590e-01 -5.78100264e-01 8.03295337e-03
3.01196486e-01 -3.71437460e-01 6.88581109e-01 -7.66586304e-01
1.27700496e+00 -4.94830385e-02 2.97502339e-01 -1.72933626e+00
-5.40257514e-01 -6.22582912e-01 -2.79846758e-01 1.32024181e+00
5.47552466e-01 -3.05311620e-01 6.47889137e-01 -2.68132351e-02
-6.12816215e-01 -7.64459729e-01 -7.75243640e-01 -3.73814583e-01
2.01781839e-01 1.68891847e-01 2.39673615e-01 4.97035652e-01
4.85712320e-01 8.67903113e-01 -2.54425436e-01 -2.66416281e-01
3.46941948e-01 1.22123674e-01 5.95229626e-01 -9.88222837e-01
1.96657196e-01 -7.32781291e-01 -1.10175133e-01 -1.33191192e+00
2.77495801e-01 -9.10464108e-01 6.88386261e-02 -2.51791072e+00
7.58996665e-01 7.57054746e-01 4.74462435e-02 1.66173875e-02
-4.63040233e-01 -2.13748306e-01 6.22113496e-02 2.37410609e-02
-9.89842653e-01 8.64683688e-01 1.35003936e+00 -7.06539869e-01
-9.30984691e-02 9.54365805e-02 -1.15637434e+00 3.29185098e-01
6.11858845e-01 -3.54732990e-01 -1.82838470e-01 -6.81707740e-01
1.39256969e-01 8.99459571e-02 -1.57484680e-01 -7.70473719e-01
6.05819046e-01 1.36290923e-01 4.43582654e-01 -1.22166443e+00
-3.25415209e-02 8.16735104e-02 -4.29152519e-01 2.40915135e-01
-9.84854221e-01 2.21953634e-02 3.64192933e-01 6.26470268e-01
-4.02567625e-01 -5.01650453e-01 3.93041015e-01 -2.17161924e-01
-1.39350891e-01 5.76286726e-02 -8.31898570e-01 -7.23615661e-02
7.93094575e-01 2.81908214e-02 -6.45934343e-01 -8.47890854e-01
-4.87198174e-01 6.59377754e-01 1.12568095e-01 3.36569458e-01
7.10618734e-01 -9.27931190e-01 -1.43749511e+00 -2.97301620e-01
-1.62500367e-01 3.12609702e-01 5.48513949e-01 5.49928129e-01
-8.15867186e-01 8.60486984e-01 2.33515296e-02 -1.76940948e-01
-1.14937413e+00 3.75497371e-01 -2.71444678e-01 -2.16400936e-01
-8.88812125e-01 9.62047160e-01 2.42308319e-01 4.83614989e-02
3.59795064e-01 -4.15004581e-01 -6.20939434e-01 4.98211920e-01
1.04835939e+00 6.32798016e-01 8.45457539e-02 -3.16013902e-01
-5.11059873e-02 3.02883834e-01 -7.25320041e-01 1.59788560e-02
1.73313129e+00 -2.48510823e-01 -3.68215650e-01 5.92698991e-01
1.36129761e+00 2.99273878e-01 -8.12229514e-01 -6.94011524e-02
-9.68230292e-02 1.77672952e-01 2.55000621e-01 -6.23828769e-01
-7.22564280e-01 6.28383815e-01 -5.70854604e-01 2.64564842e-01
6.10700965e-01 3.13947886e-01 8.93966377e-01 6.89421415e-01
-4.19618249e-01 -1.22163343e+00 -2.08448339e-02 8.53303432e-01
1.49067664e+00 -9.30145860e-01 6.55922532e-01 6.94376752e-02
-7.34670401e-01 1.27695513e+00 1.77258506e-01 -4.93519545e-01
1.60890862e-01 5.54795749e-03 -1.22556604e-01 -2.38735855e-01
-1.14873862e+00 4.70125556e-01 4.25967693e-01 1.28106251e-01
5.82722604e-01 -2.10786477e-01 -6.87983632e-01 7.27713287e-01
-3.37262213e-01 -3.66509318e-01 1.16575837e+00 6.96228027e-01
-7.27495134e-01 -3.32345903e-01 -1.59111276e-01 8.81766200e-01
-7.00042784e-01 -2.33761653e-01 -6.28758729e-01 2.88643599e-01
-8.44053626e-01 9.36706185e-01 3.13608855e-01 1.72110349e-01
5.01105130e-01 -1.26157105e-01 7.09146634e-02 -1.00574660e+00
-7.97511816e-01 1.77267075e-01 1.86054274e-01 -1.66166455e-01
-6.00286685e-02 -7.04706073e-01 -1.12323213e+00 -5.74782670e-01
-7.82592967e-02 5.62614918e-01 7.46583641e-01 5.93263149e-01
6.80595040e-01 1.16241610e+00 6.80664182e-01 -1.09745252e+00
-6.52955890e-01 -1.47957444e+00 -6.73976779e-01 7.90889189e-02
7.05582321e-01 3.75971168e-01 -4.42077458e-01 4.08468843e-02] | [12.557252883911133, 9.520431518554688] |
f92237e4-52f4-4a89-94c7-76829eb2d622 | xtreme-s-evaluating-cross-lingual-speech | 2203.10752 | null | https://arxiv.org/abs/2203.10752v3 | https://arxiv.org/pdf/2203.10752v3.pdf | XTREME-S: Evaluating Cross-lingual Speech Representations | We introduce XTREME-S, a new benchmark to evaluate universal cross-lingual speech representations in many languages. XTREME-S covers four task families: speech recognition, classification, speech-to-text translation and retrieval. Covering 102 languages from 10+ language families, 3 different domains and 4 task families, XTREME-S aims to simplify multilingual speech representation evaluation, as well as catalyze research in "universal" speech representation learning. This paper describes the new benchmark and establishes the first speech-only and speech-text baselines using XLS-R and mSLAM on all downstream tasks. We motivate the design choices and detail how to use the benchmark. Datasets and fine-tuning scripts are made easily accessible at https://hf.co/datasets/google/xtreme_s. | ['Michael Auli', 'Melvin Johnson', 'Jason Riesa', 'Sebastian Ruder', 'Orhan Firat', 'Jonathan H. Clark', 'Simran Khanuja', 'Vera Axelrod', 'Daan van Esch', 'Mihir Kale', 'Clara Rivera', 'Ye Jia', 'Colin Cherry', 'Anton Lozhkov', 'Patrick von Platen', 'Min Ma', 'Yu Zhang', 'Ankur Bapna', 'Alexis Conneau'] | 2022-03-21 | null | null | null | null | ['speech-to-text-translation'] | ['natural-language-processing'] | [ 1.54320508e-01 -2.18986962e-02 -4.49413836e-01 -5.60812354e-01
-1.64541972e+00 -7.72236049e-01 7.07357764e-01 -2.57826984e-01
-3.80036205e-01 6.24433994e-01 6.08710647e-01 -1.00795007e+00
5.04007399e-01 -1.74481377e-01 -6.00174487e-01 -4.04259801e-01
3.90648752e-01 7.12540209e-01 -1.27371162e-01 -2.74142504e-01
-4.57195431e-01 2.07810923e-01 -1.05273783e+00 7.81899929e-01
5.41679144e-01 5.60109198e-01 4.24012750e-01 8.37007582e-01
-1.62617013e-01 8.29936385e-01 -8.29624891e-01 -6.33423686e-01
-7.75668994e-02 -2.70172685e-01 -9.94167626e-01 -3.12016368e-01
3.99372250e-01 9.84792784e-02 -8.20052743e-01 7.75221407e-01
1.04118121e+00 3.61130327e-01 5.95466733e-01 -9.07131314e-01
-1.17443144e+00 1.12936401e+00 -6.77466616e-02 6.81415379e-01
4.54280972e-01 1.71356082e-01 1.04543614e+00 -1.16456521e+00
5.67906618e-01 1.55992723e+00 4.28679138e-01 8.79539788e-01
-1.18202615e+00 -6.35699749e-01 2.06249446e-01 8.23116079e-02
-1.52021599e+00 -1.53264046e+00 2.87135929e-01 -2.20323741e-01
1.79011190e+00 6.19484186e-01 -1.09949887e-01 1.99367821e+00
-3.70234221e-01 1.39563465e+00 8.67861450e-01 -4.77991015e-01
-1.08630463e-01 1.83505595e-01 8.54636654e-02 4.69998300e-01
-2.52977341e-01 1.99105486e-01 -8.31417739e-01 1.97341032e-02
3.61152351e-01 -6.66265011e-01 -3.06057751e-01 4.55852240e-01
-1.43335927e+00 7.07260132e-01 -5.47229312e-02 4.71741974e-01
-7.81786740e-02 1.12771794e-01 1.01601982e+00 6.71553671e-01
7.81570792e-01 2.39101499e-01 -9.54710066e-01 -4.92090017e-01
-7.45678544e-01 -1.99748293e-01 7.09131181e-01 1.25167859e+00
1.91985026e-01 8.37200344e-01 -2.62884140e-01 1.60798395e+00
4.30844799e-02 1.10490572e+00 9.51556385e-01 -5.94274402e-01
1.11661303e+00 -2.28801146e-01 -5.91212213e-01 -1.83887869e-01
-1.17980391e-01 -5.52264690e-01 -6.16746664e-01 -6.39058888e-01
-3.24590281e-02 -3.92788410e-01 -8.97579312e-01 1.68939638e+00
-1.36685997e-01 -2.80416131e-01 5.24063706e-01 5.11238873e-01
1.48764873e+00 1.08729243e+00 1.69931278e-01 -1.86587855e-01
1.38202941e+00 -1.23227966e+00 -7.90086329e-01 -5.41741490e-01
9.78745818e-01 -1.03141356e+00 1.50193000e+00 -2.11001076e-02
-1.21288836e+00 -5.50528169e-01 -6.56467676e-01 -4.98270959e-01
-7.58092761e-01 7.92184174e-01 3.73183548e-01 6.55387998e-01
-1.31441951e+00 6.89705014e-02 -1.00800776e+00 -6.31761551e-01
-3.42239328e-02 3.29017644e-06 -5.04776835e-01 2.17902556e-01
-1.48549724e+00 1.03041756e+00 2.05214947e-01 -3.52218807e-01
-1.04417801e+00 -5.10486782e-01 -1.07154894e+00 -9.83985215e-02
1.30344853e-01 -4.80958045e-01 1.79073751e+00 -7.79596865e-01
-1.68583798e+00 1.27660739e+00 -4.36377615e-01 -4.87293035e-01
1.94335610e-01 -2.27545395e-01 -9.72797215e-01 -4.01546538e-01
1.61110923e-01 5.44032216e-01 4.79047656e-01 -7.63883650e-01
-4.18050766e-01 -2.99051195e-01 -3.44771683e-01 3.85146677e-01
-1.62434831e-01 6.99967980e-01 -6.13981843e-01 -1.05869699e+00
-2.73409814e-01 -9.50564384e-01 -6.23147003e-03 -7.46306658e-01
-5.85915923e-01 -3.71794373e-01 4.29176599e-01 -1.08836675e+00
1.27773130e+00 -2.16207409e+00 1.38548657e-01 -4.01226401e-01
-4.92478102e-01 5.42336762e-01 -5.68585098e-01 7.89073944e-01
-3.85002941e-01 1.95197105e-01 -1.53976202e-01 -1.02617526e+00
2.04670414e-01 3.49175066e-01 -6.81132078e-01 3.11235398e-01
2.67073333e-01 1.04869258e+00 -6.77321196e-01 -4.71951440e-02
3.52421105e-01 6.25580549e-01 -8.76448974e-02 1.62462965e-01
4.79984023e-02 4.09939080e-01 -2.11707309e-01 7.41523445e-01
3.03068906e-01 1.89188436e-01 1.82491139e-01 1.33084387e-01
-2.33780831e-01 1.49605715e+00 -3.50382030e-01 1.84685540e+00
-9.74551320e-01 7.72144318e-01 1.55657113e-01 -8.43570292e-01
9.00443316e-01 6.97672367e-01 1.12720720e-01 -9.25434649e-01
6.31672814e-02 5.17686486e-01 -3.73810410e-01 -2.68222660e-01
5.57465553e-01 9.46097523e-02 -2.92912602e-01 1.72097310e-01
5.02770364e-01 -2.47192428e-01 -8.73661228e-03 9.11670923e-02
1.12181246e+00 -1.54045731e-01 4.98801440e-01 -2.11409599e-01
5.51972270e-01 -2.81786889e-01 3.36835563e-01 5.58573246e-01
-7.14896023e-02 5.34679353e-01 1.40559360e-01 -6.67421892e-02
-7.83577085e-01 -1.35250449e+00 -2.33614609e-01 1.84806311e+00
-7.68199325e-01 -6.29704952e-01 -6.17754340e-01 -7.24568486e-01
9.42000374e-02 1.07068884e+00 -3.40459466e-01 -1.37258559e-01
-5.38950622e-01 -4.92898196e-01 1.29924607e+00 5.01605928e-01
-3.34781036e-02 -1.33169198e+00 5.32025218e-01 4.76938710e-02
-4.41297531e-01 -1.34842074e+00 -9.09595191e-01 4.79132921e-01
-4.81215864e-01 -4.79060501e-01 -9.35701668e-01 -9.68878627e-01
4.10673060e-02 2.57982701e-01 1.29068339e+00 -4.13883984e-01
7.76642784e-02 4.58215624e-01 -4.96479243e-01 -2.22842738e-01
-9.46922362e-01 7.23885477e-01 3.82297784e-01 -2.97426432e-01
3.18788856e-01 -3.93303752e-01 1.27411380e-01 2.05959026e-02
-3.96038353e-01 -1.38437301e-01 4.39016342e-01 7.33886003e-01
6.00238740e-01 -6.45936251e-01 7.84465253e-01 -7.58573711e-01
6.93985283e-01 -6.15317225e-01 -5.26984870e-01 4.01870459e-01
3.24358279e-03 -8.72189458e-03 5.75635433e-01 -2.97388375e-01
-1.04209507e+00 1.19521357e-01 -9.04332936e-01 -4.26896811e-01
-1.49339154e-01 5.88237226e-01 -3.98912549e-01 4.93067801e-01
7.08945334e-01 6.54998899e-01 -3.27953696e-01 -9.53956366e-01
7.77569413e-01 1.24133897e+00 6.77319765e-01 -7.29601562e-01
3.41355145e-01 -2.25451052e-01 -9.84986544e-01 -1.39992607e+00
-5.47430217e-01 -5.99468827e-01 -3.99778426e-01 2.43960902e-01
7.83757925e-01 -1.29418015e+00 -5.07082283e-01 2.57767379e-01
-1.54368412e+00 -5.91121137e-01 -2.44450107e-01 4.91406709e-01
-5.83856285e-01 1.27140597e-01 -8.52818727e-01 -8.13215017e-01
-6.43733263e-01 -1.39780557e+00 1.47947264e+00 -4.86763239e-01
-2.47318313e-01 -1.06248903e+00 1.75607920e-01 5.13079107e-01
5.34347713e-01 -6.63939714e-01 5.58525622e-01 -1.04782295e+00
-1.25673767e-02 3.18446755e-01 6.10051230e-02 8.55128825e-01
2.64122158e-01 -4.45866175e-02 -1.33800411e+00 -4.97369200e-01
-4.49809134e-01 -4.18387085e-01 9.87906814e-01 3.29143733e-01
1.01264071e+00 -4.72252399e-01 -1.66179150e-01 8.37758541e-01
6.06554329e-01 2.91906774e-01 3.16546470e-01 1.04945019e-01
5.65008819e-01 5.88547707e-01 1.32067114e-01 2.80311592e-02
5.00656843e-01 8.60962451e-01 -2.88599312e-01 -8.89737681e-02
-7.24606454e-01 -3.47166181e-01 1.15955973e+00 1.96910846e+00
4.71517444e-01 -5.68101466e-01 -1.15155590e+00 7.21929371e-01
-1.50616050e+00 -7.13815749e-01 9.08206105e-02 2.02794623e+00
1.00286412e+00 -2.49360174e-01 1.78772300e-01 -2.90436357e-01
7.43571103e-01 3.58586699e-01 -3.50139916e-01 -5.40113151e-01
-6.55868828e-01 3.44163954e-01 4.81510878e-01 8.20540011e-01
-1.19168568e+00 1.83832753e+00 6.48865509e+00 1.21611559e+00
-1.32224357e+00 6.88173771e-01 6.23548925e-01 -2.07379237e-01
-3.69075924e-01 -1.85080111e-01 -1.13501990e+00 2.64437824e-01
1.55736232e+00 -5.53640723e-01 6.87318563e-01 9.25425053e-01
1.61784410e-01 7.01990366e-01 -1.01591003e+00 1.10665810e+00
-4.20866609e-02 -1.26641417e+00 3.29085082e-01 -3.46407354e-01
4.64013577e-01 1.17818165e+00 3.07046592e-01 7.03261197e-01
6.43377125e-01 -9.96309698e-01 8.76396060e-01 -3.51454988e-02
1.51789463e+00 -5.96819222e-01 2.96859264e-01 1.50665082e-02
-1.29641914e+00 2.91122496e-01 -2.69742459e-01 4.24539566e-01
3.35845947e-01 2.05154464e-01 -1.08843553e+00 5.99630415e-01
4.54786003e-01 9.74470854e-01 -4.57640082e-01 3.51216644e-01
-2.61923641e-01 1.06744516e+00 -2.00445041e-01 -6.24850160e-03
2.32250705e-01 1.40286997e-01 8.11669827e-01 1.95297587e+00
3.59727353e-01 -2.97479451e-01 1.43209308e-01 3.97800058e-01
-3.74703497e-01 5.54890692e-01 -8.55453014e-01 -6.54585004e-01
8.53025317e-01 8.18832695e-01 -2.17995241e-01 -4.77976441e-01
-5.62463760e-01 1.07912409e+00 5.60027599e-01 5.28784573e-01
-5.62120199e-01 -3.08273256e-01 1.22985649e+00 -1.27528319e-02
-1.44270435e-02 -5.15832961e-01 4.36769016e-02 -1.27249515e+00
-3.91515791e-02 -1.39449239e+00 3.90813231e-01 -8.22641611e-01
-1.47292531e+00 1.26043165e+00 -3.44605356e-01 -6.88039184e-01
-5.23632526e-01 -7.79648662e-01 -1.74077004e-01 1.01368976e+00
-1.46012115e+00 -1.41471720e+00 4.44613218e-01 6.99606895e-01
1.22402823e+00 -1.00271499e+00 1.16359234e+00 6.94145977e-01
-7.82795489e-01 1.07143939e+00 4.41839278e-01 4.15379465e-01
8.59447896e-01 -1.04577219e+00 1.45394933e+00 6.07829154e-01
5.90903997e-01 6.44938588e-01 2.98244834e-01 -5.24015367e-01
-1.26675332e+00 -1.27754176e+00 1.13588822e+00 -8.10367107e-01
1.21273100e+00 -7.51333416e-01 -7.82119393e-01 1.23816121e+00
4.36257392e-01 -6.53087348e-02 5.75652122e-01 3.49496067e-01
-4.96309400e-01 -1.41238272e-01 -4.41673428e-01 6.80388272e-01
1.32755625e+00 -1.40578926e+00 -4.95254844e-01 7.54287720e-01
1.26485896e+00 -3.34004104e-01 -8.04096460e-01 2.12808132e-01
2.95273781e-01 -2.23174214e-01 9.87745464e-01 -9.38125074e-01
-8.75652283e-02 1.30541027e-01 -7.50245273e-01 -1.64537907e+00
-1.15013622e-01 -9.76610005e-01 2.10374221e-01 1.56206310e+00
9.43351686e-01 -9.98438716e-01 1.71089232e-01 -3.39094460e-01
-9.46951389e-01 -3.85069609e-01 -1.50448012e+00 -1.23943591e+00
3.83155078e-01 -8.37976277e-01 7.06591308e-01 1.16513741e+00
-6.40793070e-02 6.12361670e-01 -1.37845427e-01 2.01490715e-01
1.20039783e-01 -5.34979045e-01 6.11379504e-01 -5.33778012e-01
-4.06826794e-01 -6.34510100e-01 -1.71353728e-01 -1.27853227e+00
6.76911116e-01 -1.53172565e+00 4.96496670e-02 -1.24784982e+00
-3.01677763e-01 -4.75611091e-01 -3.21656674e-01 9.86325145e-01
5.71183637e-02 -1.09930776e-01 8.12225789e-02 6.65954649e-02
-3.81562978e-01 8.37468922e-01 7.72533417e-01 -3.56260628e-01
-3.95070277e-02 1.09161906e-01 -6.03338599e-01 2.59473890e-01
8.85990739e-01 -4.29211110e-01 -4.12381917e-01 -1.02741015e+00
-2.63047338e-01 1.79636031e-01 -1.22278571e-01 -7.98993766e-01
3.20427865e-02 1.22369014e-01 -2.05136016e-01 -5.21257639e-01
7.48878837e-01 -1.52330965e-01 -7.63117969e-02 -1.79481264e-02
-5.66146076e-01 4.16003793e-01 4.81509030e-01 -1.61008105e-01
-3.76094341e-01 1.38034865e-01 6.64946020e-01 -8.56994018e-02
-5.81371248e-01 2.81095058e-01 -4.63942736e-01 4.45773125e-01
3.36236030e-01 5.10626435e-01 -7.04791486e-01 -4.98197347e-01
-9.17217016e-01 -4.97665852e-02 -4.61045876e-02 1.15399826e+00
3.72211874e-01 -1.35343587e+00 -1.21255457e+00 3.77012759e-01
4.29196894e-01 -5.84572136e-01 -1.46122081e-02 6.33092225e-01
-1.57770619e-01 9.90125060e-01 3.96408945e-01 -3.19648236e-01
-1.38404357e+00 2.75988668e-01 5.60885549e-01 7.45861530e-02
-2.98282772e-01 1.23162580e+00 3.73923779e-01 -1.10953701e+00
3.45325917e-01 -3.25137019e-01 1.26139551e-01 -1.62180558e-01
4.74746764e-01 3.18737298e-01 5.23722649e-01 -1.12945366e+00
-7.67376542e-01 -5.68423308e-02 -1.43068373e-01 -4.34486449e-01
1.16419971e+00 -2.91955143e-01 1.47463411e-01 8.21499050e-01
1.36163890e+00 2.57818669e-01 -5.88408411e-01 -4.46498990e-01
2.28724718e-01 6.11089580e-02 1.90197587e-01 -9.94944453e-01
-1.10825515e+00 1.10019755e+00 2.46261835e-01 4.82085645e-02
7.75487721e-01 4.52953130e-01 8.80966306e-01 5.94939053e-01
1.33757174e-01 -1.11491477e+00 -2.88445383e-01 1.24293029e+00
1.42879009e+00 -1.20457053e+00 -5.87017596e-01 -2.93154091e-01
-1.01708591e+00 5.83142340e-01 3.26377749e-01 4.62458014e-01
5.04072249e-01 5.07945240e-01 4.47821409e-01 1.01827584e-01
-1.07406080e+00 -4.33893472e-01 4.64385986e-01 6.86144292e-01
1.15946853e+00 5.10930240e-01 1.38460964e-01 7.92649269e-01
-7.64916778e-01 -4.86916929e-01 -3.89211113e-03 4.96060461e-01
-5.82421683e-02 -1.33429253e+00 -2.02089801e-01 1.25038520e-01
-4.83842343e-01 -7.88375556e-01 -6.34836197e-01 5.18242955e-01
-3.79807055e-01 1.11988747e+00 -5.55279292e-02 -4.13526505e-01
5.62400043e-01 3.65966558e-01 1.61545202e-01 -8.76432598e-01
-5.51368296e-01 9.68987346e-02 8.70105803e-01 -6.03530228e-01
2.54104048e-01 -7.68591106e-01 -1.06813633e+00 -3.34754109e-01
-2.76215952e-02 2.01063976e-01 1.02261662e+00 5.45067251e-01
5.50048292e-01 6.62740469e-01 5.15181959e-01 -6.31044328e-01
-5.89614391e-01 -1.28305876e+00 -3.77731055e-01 -1.41929448e-01
4.68777180e-01 -3.91945630e-01 -4.69051987e-01 -2.92635188e-02] | [14.39675521850586, 6.9865217208862305] |
8decee2c-5128-4e11-9b63-dfc2cbef93f2 | exploring-softly-masked-language-modelling | 2305.0353 | null | https://arxiv.org/abs/2305.03530v2 | https://arxiv.org/pdf/2305.03530v2.pdf | Exploring Softly Masked Language Modelling for Controllable Symbolic Music Generation | This document presents some early explorations of applying Softly Masked Language Modelling (SMLM) to symbolic music generation. SMLM can be seen as a generalisation of masked language modelling (MLM), where instead of each element of the input set being either known or unknown, each element can be known, unknown or partly known. We demonstrate some results of applying SMLM to constrained symbolic music generation using a transformer encoder architecture. Several audio examples are available at https://erl-j.github.io/smlm-web-supplement/ | ['Bob L. T. Sturm', 'Nicolas Jonason'] | 2023-05-05 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 5.63145459e-01 5.34144819e-01 -3.71118218e-01 -2.22156495e-01
-9.84760761e-01 -6.00911081e-01 7.85870135e-01 -6.15208328e-01
1.48366675e-01 8.38781714e-01 4.96588945e-01 -2.95288473e-01
1.24998584e-01 -4.13990200e-01 -8.50538850e-01 -3.54097307e-01
-1.09003022e-01 4.26007658e-01 -1.25473440e-01 -2.47319683e-01
-2.52207726e-01 -6.20294400e-02 -1.52109480e+00 8.85038495e-01
-1.82246253e-01 5.83851099e-01 1.73908561e-01 9.42211032e-01
-1.08077049e-01 1.25179470e+00 -8.07628512e-01 -3.70840341e-01
2.03266054e-01 -7.95935035e-01 -8.26374173e-01 -1.28571570e-01
7.51799867e-02 7.61493221e-02 -4.08532977e-01 9.11099017e-01
5.22464931e-01 -1.59224615e-01 3.54406416e-01 -1.53035510e+00
-4.74485695e-01 1.57993424e+00 1.08139142e-01 -1.18400261e-01
9.17789698e-01 -4.09245156e-02 9.04180944e-01 -1.06983089e+00
5.14260530e-01 1.40526628e+00 4.69552219e-01 6.71176851e-01
-1.60987198e+00 -1.11476266e+00 2.36482471e-02 1.39808282e-01
-1.70345151e+00 -1.15363657e+00 6.99073374e-01 -4.96391654e-01
7.51758575e-01 6.42914712e-01 3.28306794e-01 1.44065678e+00
-1.41455650e-01 1.05414367e+00 1.02131724e+00 -8.85460973e-01
1.71350867e-01 3.65568936e-01 -3.55812699e-01 4.10730004e-01
-3.71801078e-01 5.70312381e-01 -1.37759852e+00 -4.91929293e-01
6.43551290e-01 -6.61438763e-01 -2.78768808e-01 1.20754102e-02
-1.63418531e+00 4.96500075e-01 -3.08241576e-01 1.70608684e-01
1.02128208e-01 4.36204702e-01 3.76531720e-01 7.55025804e-01
3.14316720e-01 4.49586332e-01 -2.78741866e-01 -3.16983610e-01
-1.65819371e+00 2.55873144e-01 7.73364246e-01 1.21246552e+00
5.16136408e-01 7.72442758e-01 -8.92819688e-02 5.59974790e-01
4.26547676e-01 4.81528044e-01 6.00249469e-01 -1.23753810e+00
1.68774635e-01 6.75004628e-03 -2.32836939e-02 -4.37531501e-01
3.24331298e-02 -2.88106620e-01 -6.99211895e-01 2.78170705e-01
1.53946117e-01 -3.50004882e-02 -8.57875168e-01 2.07903862e+00
-2.87863493e-01 8.18326771e-01 2.39883393e-01 6.17830634e-01
1.04521871e+00 6.83794856e-01 -2.44680896e-01 -2.89736956e-01
9.97029662e-01 -6.83871806e-01 -8.77351403e-01 -2.62665451e-01
1.35341465e-01 -1.13268626e+00 8.51433933e-01 7.10641623e-01
-1.47849417e+00 -7.27122724e-01 -1.14274251e+00 -2.22202502e-02
-2.32219607e-01 2.17218444e-01 4.70040917e-01 6.03366852e-01
-1.19614029e+00 4.33659047e-01 -7.02843547e-01 3.65830123e-01
4.59265597e-02 4.73658502e-01 -2.40386337e-01 2.70741224e-01
-1.61966324e+00 5.02008915e-01 4.48934287e-01 1.83624625e-01
-1.30966449e+00 -5.16392112e-01 -9.98597503e-01 -2.86539763e-01
3.60085428e-01 -4.68195707e-01 1.80410254e+00 -1.13850164e+00
-1.74768221e+00 8.74771178e-01 -5.88038683e-01 -5.42931437e-01
2.93727487e-01 -2.30303556e-02 -7.83054471e-01 3.54410559e-02
1.74047984e-02 5.91454089e-01 1.16640937e+00 -1.22079730e+00
-8.89058635e-02 4.42300826e-01 -7.08375946e-02 -1.72787100e-01
4.56433505e-01 6.76466167e-01 -7.23130256e-02 -1.16334057e+00
-2.20751781e-02 -9.29161191e-01 -3.63553911e-02 -2.54107833e-01
-7.19641447e-01 4.17056739e-01 4.60537165e-01 -6.26136363e-01
1.54660296e+00 -2.24127936e+00 3.96949768e-01 1.18875191e-01
-6.17860109e-02 5.58722131e-02 -6.26878291e-02 8.64605188e-01
-5.63046694e-01 1.76233187e-01 -4.97825027e-01 -7.47713327e-01
4.24760818e-01 1.91241160e-01 -1.08517385e+00 1.02953799e-01
1.72463611e-01 9.43689346e-01 -8.39331150e-01 -3.64747226e-01
1.32915512e-01 5.93961060e-01 -2.24632487e-01 1.80172094e-03
-5.84119797e-01 6.79204762e-01 2.30432346e-01 7.43724763e-01
1.79889351e-01 -6.59375414e-02 9.92859453e-02 3.60502630e-01
-8.54589194e-02 6.53820693e-01 -1.76292515e+00 1.69796276e+00
-4.86986846e-01 8.34671199e-01 2.16798425e-01 -4.09350961e-01
7.09693849e-01 1.08299828e+00 8.62617046e-02 -9.88350064e-02
-3.99724506e-02 1.42516702e-01 -7.60838240e-02 1.37650669e-01
5.74686468e-01 -4.90799934e-01 -3.53267461e-01 6.82894468e-01
1.84467703e-01 -3.48953426e-01 1.03515625e-01 4.15050745e-01
6.80046201e-01 5.79077005e-01 4.58353102e-01 1.33219376e-01
5.35633624e-01 -2.81268865e-01 4.28486198e-01 7.86078513e-01
3.94829541e-01 7.33740449e-01 3.07316899e-01 1.75894141e-01
-7.45805025e-01 -1.31913793e+00 5.84775070e-03 1.18376064e+00
-4.67312694e-01 -1.09928942e+00 -3.69350553e-01 2.50484258e-01
-8.39467943e-02 9.76254404e-01 -1.93157047e-01 -2.65068114e-01
-3.60711068e-01 -1.66010603e-01 1.17854989e+00 4.38682228e-01
-5.10044992e-02 -1.44184709e+00 -3.69679153e-01 9.31719616e-02
-1.30218551e-01 -8.58155072e-01 -5.51550090e-01 4.74296361e-01
-4.83205736e-01 -5.43120801e-01 -3.49358141e-01 -6.46417141e-01
2.17422307e-01 -2.34000981e-01 1.40119648e+00 -1.54754013e-01
-8.17232952e-02 2.55439848e-01 -2.17081487e-01 -5.95152497e-01
-1.08480597e+00 -1.68449059e-01 4.43242431e-01 1.63421631e-01
1.15071557e-01 -8.30631852e-01 1.63890317e-01 2.01602608e-01
-8.51673245e-01 4.84721810e-01 2.24877715e-01 6.22784317e-01
6.25905752e-01 -3.14743854e-02 4.10437524e-01 -8.46482515e-01
4.81111139e-01 -4.35861707e-01 -4.41829622e-01 2.84630992e-02
-3.57978940e-01 -2.28312723e-02 4.41257805e-01 -8.96942794e-01
-5.22137523e-01 3.51325065e-01 5.43980510e-04 -6.00987613e-01
-8.08421671e-02 6.75478697e-01 -3.78510267e-01 2.42934123e-01
4.44922060e-01 6.26343131e-01 -9.52439085e-02 -5.99072576e-01
5.39332569e-01 9.19750631e-01 8.47846985e-01 -6.94512486e-01
9.03923392e-01 3.95775229e-01 -1.39903337e-01 -7.82400548e-01
-4.86061543e-01 -1.00917771e-01 -7.64932156e-01 -1.81374833e-01
3.18013579e-01 -1.22664332e+00 -5.99698484e-01 3.62892359e-01
-8.43552828e-01 -7.29069352e-01 -8.88509154e-01 3.45657468e-01
-9.09187555e-01 2.01632246e-01 -5.56158841e-01 -1.08709764e+00
5.49742021e-02 -9.06322420e-01 1.15809131e+00 -4.64836985e-01
-1.00951052e+00 -7.99238145e-01 -1.61455140e-01 1.20564125e-01
3.75155807e-01 1.44063141e-02 7.77398944e-01 -5.91888428e-01
-4.13356036e-01 -3.74411553e-01 6.27543628e-01 4.48066264e-01
2.57216357e-02 -9.06437486e-02 -1.35877538e+00 -2.51701623e-01
3.36570619e-03 -3.19310278e-01 6.51350856e-01 1.47309408e-01
7.44607985e-01 -5.26056707e-01 -9.82677266e-02 5.40969729e-01
8.48879457e-01 -1.22719951e-01 4.84003067e-01 5.18539324e-02
2.38616645e-01 -3.23021524e-02 1.33280903e-01 5.95520675e-01
1.81047797e-01 9.74999905e-01 1.32364780e-01 6.95057437e-02
-5.51792383e-01 -7.38241136e-01 9.75085735e-01 1.10857999e+00
1.51292473e-01 -1.14405425e-02 -7.23274648e-01 3.90902549e-01
-1.57080698e+00 -1.45364392e+00 1.21673152e-01 2.31589842e+00
1.49791455e+00 2.38624394e-01 3.09819490e-01 6.24434352e-01
6.84987664e-01 2.44603977e-01 -2.93182313e-01 -4.38401043e-01
-5.23162603e-01 6.99109972e-01 1.25026196e-01 1.13504136e+00
-8.04023504e-01 8.78408670e-01 7.47574997e+00 8.88601542e-01
-9.51013446e-01 2.92256624e-01 -1.12771489e-01 -6.73565984e-01
-4.64369684e-01 3.85559529e-01 -7.74125397e-01 6.70109034e-01
1.38004506e+00 -5.72582722e-01 9.79455829e-01 3.19111764e-01
4.08072978e-01 1.48188680e-01 -1.38297451e+00 1.00576627e+00
-6.59092888e-02 -1.28483129e+00 2.49709398e-01 -1.46004006e-01
5.15120685e-01 -1.10418007e-01 2.64413714e-01 4.34924841e-01
4.23839241e-01 -1.27490282e+00 1.46798742e+00 7.53387630e-01
1.26114559e+00 -4.33799952e-01 2.10911766e-01 5.66051066e-01
-1.22934568e+00 -1.72964260e-02 1.05408527e-01 -4.66864049e-01
3.77228081e-01 2.79756635e-01 -8.07137966e-01 4.94282633e-01
2.22393677e-01 6.37891173e-01 -2.65407503e-01 5.35292447e-01
-5.11156440e-01 1.11574090e+00 -3.16783786e-01 5.29064536e-01
-2.05446675e-01 2.55007207e-01 8.60470951e-01 1.35239542e+00
1.13813266e-01 -1.97512686e-01 3.14493150e-01 1.15668797e+00
1.72527060e-01 -3.60764533e-01 -5.69767118e-01 -3.07198137e-01
6.38246715e-01 6.38495743e-01 -1.29757434e-01 -4.76623237e-01
-2.26665437e-01 9.55322623e-01 -4.52591687e-01 4.96340245e-01
-6.57162607e-01 -4.58784215e-02 5.46849191e-01 3.59214902e-01
5.03008738e-02 -3.69571120e-01 5.24193570e-02 -1.22151566e+00
-2.34688535e-01 -1.27108788e+00 3.28474641e-01 -1.21310115e+00
-7.95343220e-01 6.78335369e-01 3.16997528e-01 -1.64041090e+00
-9.87316787e-01 -3.24887037e-01 -3.49702567e-01 1.19918978e+00
-1.06929541e+00 -1.33849120e+00 2.94787765e-01 4.82122511e-01
5.59155583e-01 -6.05293810e-01 1.40726566e+00 7.85360783e-02
1.66233312e-02 6.29486620e-01 -1.77781299e-01 9.53995530e-03
5.66432238e-01 -1.12341428e+00 7.36737728e-01 8.14512908e-01
7.71910310e-01 8.18234265e-01 8.66261244e-01 -5.90486765e-01
-1.11955988e+00 -8.89324725e-01 1.10979593e+00 -7.11185217e-01
7.95734525e-01 -8.70231807e-01 -6.16192102e-01 1.23173869e+00
3.25237662e-01 -3.82514238e-01 1.00601971e+00 -2.97659457e-01
-3.83321166e-01 4.46482241e-01 -7.65372753e-01 7.43602633e-01
9.79152322e-01 -1.16887033e+00 -6.47934616e-01 2.22873420e-01
8.34061444e-01 -7.79349208e-01 -7.84296095e-01 2.03043878e-01
6.11754835e-01 -7.06331968e-01 8.69665504e-01 -5.76233745e-01
2.45019972e-01 -7.60943949e-01 -7.27015793e-01 -1.00202131e+00
-2.19908848e-01 -1.50038624e+00 -6.93519235e-01 1.08000076e+00
5.80228090e-01 -5.01066089e-01 4.66090977e-01 1.73226878e-01
-5.24217263e-02 -4.12791222e-01 -1.29097462e+00 -9.36832547e-01
3.08487955e-02 -1.15630615e+00 7.57486522e-01 8.61081004e-01
3.66302937e-01 1.72430530e-01 -8.35619569e-01 2.54541993e-01
3.45051527e-01 1.78692997e-01 7.11039305e-01 -7.44023263e-01
-1.03307390e+00 -2.95406431e-01 -4.39009428e-01 -8.30031931e-01
4.01138932e-01 -1.50151944e+00 -1.47288784e-01 -1.30824101e+00
-1.48380950e-01 1.09224729e-01 -3.66489857e-01 9.40869808e-01
5.45284808e-01 6.42903090e-01 5.28092206e-01 2.38788381e-01
-3.24440271e-01 1.44807369e-01 6.59814119e-01 -2.80328304e-01
-9.60837230e-02 5.36222935e-01 -7.42184341e-01 5.06840825e-01
6.90281630e-01 -5.70849597e-01 -3.99717152e-01 -2.78253049e-01
6.66732073e-01 3.85461748e-01 6.69017136e-01 -9.44274068e-01
2.11603314e-01 -5.31020984e-02 1.63991764e-01 -4.76310730e-01
9.89142478e-01 -5.10569215e-01 9.79144156e-01 3.34578305e-01
-7.20481455e-01 -1.35775413e-02 5.89864969e-01 -7.43632158e-03
-4.88253295e-01 -2.34628737e-01 2.84242421e-01 -2.89780378e-01
-5.46284437e-01 -2.50014961e-01 -6.10984445e-01 -3.08285095e-03
4.68328357e-01 -4.37078685e-01 1.35760643e-02 -8.25851381e-01
-1.32369637e+00 -1.26524583e-01 3.30914050e-01 5.23863435e-01
7.69751430e-01 -1.47671664e+00 -7.91289508e-01 5.11096895e-01
-9.36592836e-03 -2.27749497e-01 -2.09131420e-01 7.46164799e-01
1.49119407e-01 4.27060813e-01 2.78307498e-01 -1.35428756e-01
-1.41151321e+00 5.64940929e-01 3.28850985e-01 3.44046652e-01
-5.84351480e-01 9.62813675e-01 -2.30646461e-01 -4.56376582e-01
3.61268997e-01 -1.82371110e-01 4.95186180e-01 -2.66972363e-01
7.34121561e-01 1.08907171e-01 -4.16631728e-01 -8.48350346e-01
-5.03019273e-01 3.19229871e-01 5.48492253e-01 -8.55012953e-01
1.08344448e+00 3.77075672e-02 -3.80812109e-01 1.33003628e+00
6.79479003e-01 5.08388281e-01 -6.84217393e-01 -3.90949130e-01
2.80691218e-02 -1.10438362e-01 -1.65826436e-02 -1.06645727e+00
-5.32996655e-01 8.37518752e-01 2.27104634e-01 -1.01997077e-01
1.10380638e+00 1.72284216e-01 3.87178481e-01 4.96548302e-02
4.55271214e-01 -5.61346710e-01 -4.19995397e-01 4.23639566e-01
1.21838605e+00 -5.63702703e-01 8.95513892e-02 -3.13987657e-02
-8.61916244e-01 9.78893042e-01 -1.26436934e-01 1.50420502e-01
8.34191918e-01 9.77762282e-01 2.57448465e-01 6.08702637e-02
-1.12264752e+00 -1.86555654e-01 2.73320943e-01 4.77541000e-01
5.72252870e-01 2.41579786e-01 3.36561084e-01 1.03324926e+00
-9.87965345e-01 3.38243604e-01 6.84469163e-01 8.99928510e-01
-4.51721437e-02 -1.42442930e+00 -6.05902135e-01 2.11805061e-01
-5.84105194e-01 -5.29580355e-01 -6.28231168e-01 4.53084081e-01
3.36361349e-01 1.12989044e+00 -2.08123431e-01 -6.50460124e-01
-4.52546813e-02 5.89665890e-01 5.80324113e-01 -1.14637208e+00
-4.90764529e-01 2.14879394e-01 1.83825105e-01 -2.98338205e-01
-4.12609220e-01 -7.96045899e-01 -1.26565325e+00 -4.24365848e-02
-1.17822327e-01 9.64226052e-02 4.12689328e-01 5.39670289e-01
1.17253169e-01 5.23765683e-01 4.56712812e-01 -1.04919827e+00
-5.36118686e-01 -1.02043259e+00 -5.25021791e-01 -2.53631234e-01
7.21006036e-01 -3.78823519e-01 -4.92045075e-01 6.62965596e-01] | [15.693034172058105, 5.7549519538879395] |
2d4b1647-1d0b-44f3-9840-4dd52d8ffe31 | incorporating-external-knowledge-into-machine | 1909.02745 | null | https://arxiv.org/abs/1909.02745v1 | https://arxiv.org/pdf/1909.02745v1.pdf | Incorporating External Knowledge into Machine Reading for Generative Question Answering | Commonsense and background knowledge is required for a QA model to answer many nontrivial questions. Different from existing work on knowledge-aware QA, we focus on a more challenging task of leveraging external knowledge to generate answers in natural language for a given question with context. In this paper, we propose a new neural model, Knowledge-Enriched Answer Generator (KEAG), which is able to compose a natural answer by exploiting and aggregating evidence from all four information sources available: question, passage, vocabulary and knowledge. During the process of answer generation, KEAG adaptively determines when to utilize symbolic knowledge and which fact from the knowledge is useful. This allows the model to exploit external knowledge that is not explicitly stated in the given text, but that is relevant for generating an answer. The empirical study on public benchmark of answer generation demonstrates that KEAG improves answer quality over models without knowledge and existing knowledge-aware models, confirming its effectiveness in leveraging knowledge. | ['Ming Yan', 'Jiangnan Xia', 'Bin Bi', 'Chen Wu', 'Chenliang Li', 'Wei Wang'] | 2019-09-06 | incorporating-external-knowledge-into-machine-1 | https://aclanthology.org/D19-1255 | https://aclanthology.org/D19-1255.pdf | ijcnlp-2019-11 | ['generative-question-answering'] | ['natural-language-processing'] | [ 0.27264056 0.6811645 -0.0829123 -0.14374126 -1.0907495 -0.84546375
0.70775545 0.3415755 -0.39602652 1.2839886 0.66903895 -0.49102792
-0.09988374 -1.3975536 -0.80639714 0.0225477 0.6347822 0.71248937
0.57716274 -0.7562965 0.38919726 -0.02134978 -1.3930149 0.6250641
1.5080576 0.720081 0.26867205 0.6724764 -0.7670168 1.580148
-0.87353796 -0.6949183 -0.22256362 -0.9723229 -1.5740862 -0.5837681
0.28052008 -0.29035115 -0.19316958 0.8018186 0.40836793 0.40401143
0.7512906 -0.774207 -1.1525129 1.0666963 0.5437158 0.38942268
0.9494821 0.33491826 1.1199242 -0.7139049 0.93493015 1.2523284
0.371745 0.7650952 -0.9684766 -0.17800811 -0.0786033 0.6797181
-0.8006373 -0.3265708 0.8955019 -0.25802562 1.1003067 0.31171182
0.4797404 1.0115436 -0.10473621 0.85582423 1.2505928 -0.82606477
0.3620807 0.2793421 0.5165767 0.5681741 0.04300527 -0.11612626
-0.77353066 -0.39995047 0.3419067 -0.71118957 -0.47865742 0.09679116
-1.0502236 1.045157 0.44161436 0.32455623 -0.6633914 0.01626587
0.15121946 0.5283395 -0.09883515 1.2453834 -0.64590144 -0.26460317
-0.67058533 0.6031414 1.369784 0.7565418 0.87102425 -0.19009343
-0.95804393 0.5229223 0.17234103 0.66232115 0.66985625 -1.3843119
0.7436563 1.3064758 0.3791162 -0.92058367 -0.02709368 -0.24741223
-0.24343075 -0.36084536 0.47941944 -0.38237676 -0.73822623 1.7524049
0.52687925 -0.02358382 0.4979068 0.61837107 1.6621137 0.624098
0.1193335 -0.02232885 1.2708026 -0.92548245 -0.8385624 -0.38368326
0.5379571 -0.28875077 1.2766156 -0.03603448 -1.0929841 -0.46472183
-0.87496024 -0.44341052 -0.44918144 -0.1278756 0.39516628 0.31997567
-0.9448298 0.22452198 0.15380573 -0.15309097 0.20399415 -0.08908916
-0.02159547 -0.38247073 -1.9672297 1.347257 0.67819303 -0.17039421
-0.85955596 -0.62672055 -1.0180023 0.29499218 0.75587153 -1.3933579
1.4677043 -0.725275 -1.652472 0.4956231 -0.6590796 -0.71643496
0.14245625 -0.15676549 -0.23604284 0.39805555 0.25513843 0.5969923
0.68309426 -1.2701198 -0.30330488 -0.1347306 0.8100916 0.29887938
-0.03108756 -0.09607171 -0.24640092 -0.18009186 -0.14590888 -0.40765917
-0.11914033 -0.7095172 -0.38388005 -0.66698307 0.2416663 -0.8763754
1.3450097 -1.2433203 0.1762603 0.0808975 0.12315171 0.32692832
-0.26053536 0.6153424 0.522162 0.14518476 -0.22070725 0.46991134
0.25019658 0.3095707 -0.65671927 -0.7526835 0.32267535 1.6069881
-1.2876393 -0.5754115 -0.21410689 -0.03624246 -0.60284185 0.41666326
-0.9673976 0.37902996 -0.79014957 0.45034188 0.1972763 -0.49044198
-0.07711414 -0.11823156 0.5137379 0.8787351 -0.9618896 1.4901938
-0.71656775 0.31493145 -0.40786472 -0.5131987 0.8094861 0.45623532
-0.5004777 -0.7986762 0.09818809 0.28934628 -0.04672914 -1.0453912
0.70840096 -0.255148 -0.02175152 0.61451316 0.4068418 -0.64945334
0.6402818 0.7209695 1.3150575 0.15521547 0.45005995 0.10476369
0.93339086 0.43560812 0.2602003 1.0696219 0.0624918 0.12421855
0.45037434 0.03552949 -0.4234746 -0.9056899 0.38020957 0.9717942
-0.03570862 -0.4631343 -0.73977214 -1.0981784 -0.10433546 1.4513466
-0.73913616 -0.24289036 -0.6208843 -0.13052984 0.58633584 0.440059
0.6793419 -1.6733277 -0.6820746 0.35282803 -0.9764937 -0.9952971
-0.09724848 -0.17089127 -0.7947209 -1.2571882 -0.3284284 -0.3731837
0.48066643 -0.13999332 1.740724 0.18434112 0.25932205 0.7443247
-0.549249 -0.41147852 -0.7369523 0.11480108 -0.6710861 -0.29218426
0.6408682 -0.36419535 -0.4517266 -0.12998234 -0.89821076 -0.08763288
0.3546594 0.8413332 0.29548338 -0.27621433 1.2888609 -1.0553277
1.3165003 -0.9287569 -0.18055761 0.67800206 -0.25014728 0.4223142
0.6192577 -0.11231375 -1.5703497 -0.42842293 -0.26481462 0.36283407
-0.1342737 1.0011932 -0.2839365 0.19517647 1.1297575 0.40708512
-0.38392588 -0.15357083 0.95382893 0.32183963 0.42811912 -1.0882947
0.62975174 0.11912411 -0.21108128 -0.3574567 -1.424352 -0.43637517
-0.36574495 -0.10447607 0.6597862 -0.5104892 -0.55931115 -0.05357933
-1.2441208 -0.38980126 -0.6927575 0.07348161 -0.59830517 0.29778183
-0.3594283 -0.67761153 -0.5900429 -0.70320505 0.41958404 0.4989245
-0.46736968 -1.195823 0.28424585 0.9202313 0.6237858 0.12735093
1.3174524 -0.9220081 -0.7899339 -0.10005482 -0.05527785 0.35315362
0.0477513 -0.4465825 -0.9109455 0.46278954 0.36001492 -0.9357626
0.8737971 -0.02989879 0.74054146 -0.7682903 0.07139488 -0.11934468
1.2037033 -0.04832341 0.6333634 0.4507377 0.22381856 0.8847619
0.41539216 -0.03358686 0.9751737 0.3619486 0.26885667 0.517569
-0.40005693 -0.53687173 0.2055424 0.6269418 0.05382854 -0.13370158
-0.8654002 1.0906278 -1.753102 -1.1795081 -0.19761576 1.7513005
1.6038853 -0.18219131 -0.32934487 -0.1692212 0.22471695 -0.06436992
-0.708916 -0.29439142 -0.42017385 0.58862233 -0.27656686 0.81990147
-0.5111947 1.1303729 5.874549 0.79498357 -0.36946043 0.16071089
0.07901961 0.27300355 -0.99814594 0.24695668 -0.83551985 0.17262669
0.94857246 -0.58403504 0.3271104 0.5415049 -0.10391016 -0.41521642
-1.0238208 0.15741293 0.44191593 -1.6722033 0.65463734 -0.56817275
0.96124965 -0.27672875 -0.29686865 0.8232568 0.8899996 -1.0111926
0.41360635 0.86524016 0.17733672 -0.52276087 0.94158965 0.76277375
-0.6691396 -0.15868664 -0.20295142 -0.18925393 0.44307473 0.42763
-1.1150205 0.87534946 0.29606536 0.15433885 -0.8704286 0.7969138
-1.2085335 1.0222971 -0.0974982 -0.2634412 0.29388335 0.23470622
0.47300595 0.90787846 0.22401884 0.50364137 -0.07656423 1.3160998
-0.3273068 0.2291269 -0.5656435 0.10293998 0.73663616 1.0183337
-0.04769595 -0.7353513 -0.14308758 0.72409195 0.632704 0.5704303
-0.29911178 -0.44045278 -0.19719501 -0.17123139 0.18178014 0.15064041
-0.26254112 -1.1713331 0.1329611 -1.1140935 0.65178883 -1.2121
-1.3237983 0.5652776 -0.0117886 -0.5302695 -0.71601635 -0.39744818
-0.6636115 1.2470013 -1.9359859 -1.1750306 -0.20084229 0.6652578
0.50075305 -0.07485317 0.9788419 -0.32949147 0.1274578 0.28544363
-0.5426926 0.13919818 0.5658043 -1.4076523 0.20959045 0.8762346
0.19657914 1.0495541 0.6200862 -0.8599786 -1.3711976 -0.84552354
1.4635948 -1.1781325 0.6604094 0.39871743 -1.2036492 0.61622417
0.5880221 -0.5114108 0.9298059 0.09804039 -0.54758024 0.3444437
-1.3674402 0.6496275 0.70092684 -0.59822696 -1.7073584 0.13192259
1.1769248 -0.38381815 -0.60415536 0.285652 -0.03798584 -0.66039306
0.8612122 -1.0031968 0.7721665 -0.5348172 -0.24572532 -1.5274531
-0.05110103 -0.45688263 -0.88662416 1.2580969 0.81391805 -0.5518433
0.44066104 0.8296861 -0.13114141 -0.59375525 -0.94985425 -0.5037484
0.32133126 -0.22302745 0.8945765 0.8732912 0.27717638 0.8784301
0.13903283 -0.1293111 0.29660308 0.35176006 0.7767355 -1.0347455
-0.38025767 -0.2004214 0.33716547 -1.0971173 0.38799202 -1.1589124
0.0517283 -2.3171747 0.19684713 -0.14952052 -0.01891898 0.42121074
-0.8550159 -0.15219426 0.24751757 -0.25316513 -0.77204895 0.6917209
1.6171033 -0.19592956 -0.11058728 -0.21425048 -1.1260422 0.73340786
0.741968 -0.24261422 -0.6601481 -0.4738591 0.88059497 0.22433603
0.5340142 -0.67844206 0.51701003 -0.31514868 0.09482188 -0.42931545
0.16979401 -0.46742162 -0.41292855 0.16466035 -0.71769834 -0.05408252
0.19271071 0.4172674 -0.4983482 -0.81797904 0.16056253 -0.5927842
-0.7920643 -0.2086212 -0.22876 0.9004029 0.40540242 0.03559444
-0.82446915 -0.62628615 -0.66245615 0.5094269 0.0818168 0.2848466
0.7929444 -1.3025682 -1.0150151 -0.42048702 0.3547145 -0.17518564
0.40188894 0.34642735 -0.16846262 0.6560404 0.17161523 -0.01644415
-0.7333054 0.4290755 0.34253344 -0.75677264 -0.20412716 0.9246581
-0.33928666 -0.81705254 -0.11773009 -0.29415804 -0.892883 0.05220101
0.7023161 0.26609993 0.0978332 -0.2865588 0.09123746 0.2323868
0.11100452 -0.40655646 0.94684047 -0.02042808 -0.45786825 0.4162294
0.36231834 0.3169997 -0.6524027 -0.71035045 0.17690329 -0.19950072
-0.4284277 -1.6479965 -0.32993343 0.7544744 -0.21814635 0.16805466
0.8916227 0.17118596 0.96168476 1.0146898 0.49761984 -0.98657703
0.43649706 1.1833056 1.3057003 -1.0832598 -0.37731445 -0.1652138
-0.7379484 0.85073876 0.9565321 0.12836644 0.15230168 -0.29548258
0.1645965 -0.5211341 -0.93020636 -0.44940186 0.5302592 0.66675705
0.30655316 -0.14165187 -0.34209514 0.9386768 -0.5856856 0.27270555
0.5961694 0.92399246 -0.71955705 -1.0850078 -0.32944566 0.6023641
-0.24619117 -0.44503176 -0.8668164 0.31366885 0.05304552 1.5081903
-0.6968551 0.11261171 0.38252497 0.6453711 0.6224866 -0.98413414
-0.9940698 -0.97053325 0.68001246 -0.27320787 -0.4469269 -0.21697265
-1.3342909 0.04722967 -0.35759985 0.68477994 0.22983547 1.3435467
0.5186991 0.62712735 0.07401592 0.05272858 -0.6125869 -1.1219493
0.18420899 0.4989088 0.27859935 -0.41953486 -0.2791177 0.15257475] | [10.952467918395996, 8.005026817321777] |
38f18a19-cc1b-403e-bf28-013011418539 | provably-consistent-partial-label-learning | 2007.08929 | null | https://arxiv.org/abs/2007.08929v2 | https://arxiv.org/pdf/2007.08929v2.pdf | Provably Consistent Partial-Label Learning | Partial-label learning (PLL) is a multi-class classification problem, where each training example is associated with a set of candidate labels. Even though many practical PLL methods have been proposed in the last two decades, there lacks a theoretical understanding of the consistency of those methods-none of the PLL methods hitherto possesses a generation process of candidate label sets, and then it is still unclear why such a method works on a specific dataset and when it may fail given a different dataset. In this paper, we propose the first generation model of candidate label sets, and develop two novel PLL methods that are guaranteed to be provably consistent, i.e., one is risk-consistent and the other is classifier-consistent. Our methods are advantageous, since they are compatible with any deep network or stochastic optimizer. Furthermore, thanks to the generation model, we would be able to answer the two questions above by testing if the generation model matches given candidate label sets. Experiments on benchmark and real-world datasets validate the effectiveness of the proposed generation model and two PLL methods. | ['Masashi Sugiyama', 'Xin Geng', 'Lei Feng', 'Jiaqi Lv', 'Bo Han', 'Bo An', 'Miao Xu', 'Gang Niu'] | 2020-07-17 | null | http://proceedings.neurips.cc/paper/2020/hash/7bd28f15a49d5e5848d6ec70e584e625-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/7bd28f15a49d5e5848d6ec70e584e625-Paper.pdf | neurips-2020-12 | ['partial-label-learning'] | ['methodology'] | [ 3.00495207e-01 1.70141205e-01 -2.88883835e-01 -5.79214275e-01
-7.36721933e-01 -4.99159575e-01 4.67209995e-01 3.61785650e-01
-1.88284650e-01 9.65946198e-01 -4.82875258e-01 -1.37913048e-01
-5.73223829e-01 -7.07976639e-01 -5.70582688e-01 -7.34342813e-01
2.07063094e-01 7.81778038e-01 4.20534134e-01 2.05026239e-01
2.80374765e-01 2.79510915e-01 -2.00592613e+00 1.37209207e-01
9.26637709e-01 1.40958464e+00 7.19673093e-03 3.38597357e-01
1.18905241e-02 7.71993577e-01 -2.86973059e-01 -7.46290863e-01
3.60985994e-01 -6.14440501e-01 -1.11915076e+00 -3.09481118e-02
5.03677666e-01 2.00625926e-01 5.24567902e-01 1.17055273e+00
2.26406798e-01 -9.99812558e-02 9.02861893e-01 -1.64301217e+00
-4.47274894e-01 4.46105957e-01 -2.43874177e-01 -3.33891660e-01
1.01846926e-01 -2.44737625e-01 1.31022108e+00 -6.80093288e-01
6.34039342e-01 9.42319393e-01 9.95453894e-01 7.33110070e-01
-1.13083923e+00 -4.75358993e-01 1.34536535e-01 6.79901689e-02
-1.30306470e+00 -7.33990371e-02 7.95974851e-01 -5.25572121e-01
3.83352220e-01 2.42491752e-01 4.40665632e-01 9.41960156e-01
1.90867841e-01 7.89745450e-01 1.61295140e+00 -5.42023361e-01
3.71518373e-01 7.53147662e-01 3.67550462e-01 8.80061090e-01
2.88398892e-01 3.02214265e-01 -4.38161671e-01 -2.12138355e-01
6.18590266e-02 -3.38197649e-01 -2.43452579e-01 -8.06803167e-01
-1.02944422e+00 9.09972310e-01 -2.51513217e-02 5.24241745e-01
5.50159141e-02 -4.18370999e-02 3.60944122e-01 3.41186106e-01
5.07786691e-01 6.37869775e-01 -4.96052712e-01 3.72108966e-01
-1.15461671e+00 3.81411225e-01 9.57797706e-01 8.47612083e-01
7.91291952e-01 -2.82153428e-01 4.62062992e-02 6.29323304e-01
4.19900507e-01 2.52905875e-01 5.93699872e-01 -8.02213728e-01
1.05749950e-01 4.03173864e-01 1.03107579e-01 -9.65770841e-01
-6.89199805e-01 -5.25801957e-01 -7.13907540e-01 2.34674707e-01
7.16054022e-01 1.45738810e-01 -3.86078656e-01 2.09464145e+00
3.50591153e-01 2.14442462e-01 1.49073660e-01 5.96905053e-01
4.99302804e-01 3.27611566e-01 -8.55010524e-02 -4.70579207e-01
9.63934362e-01 -9.14720535e-01 -5.73647559e-01 -3.52467746e-02
7.64296055e-01 -6.91176236e-01 9.55835283e-01 6.62348568e-01
-8.41966569e-01 -5.76562524e-01 -1.16961586e+00 3.15168738e-01
-3.25966060e-01 7.49778226e-02 6.35769308e-01 1.12895477e+00
-1.02208257e+00 8.91148865e-01 -5.46204984e-01 -3.28379899e-01
8.17883983e-02 4.18243676e-01 -3.42042685e-01 3.19992095e-01
-1.24284506e+00 1.06706989e+00 5.81483603e-01 5.50958663e-02
-6.80023909e-01 -2.92193830e-01 -5.73363602e-01 -3.11438236e-02
2.95075059e-01 -5.57421088e-01 1.29418242e+00 -1.27085650e+00
-1.41071320e+00 9.44022119e-01 -6.55105338e-03 -4.48176920e-01
8.36943865e-01 8.67424533e-02 -2.92508900e-01 -1.06829397e-01
1.68805301e-01 5.78194857e-01 6.67080998e-01 -1.58147061e+00
-8.22695434e-01 -2.30515093e-01 -4.77797687e-02 -1.32609364e-02
-3.25897396e-01 -2.19994336e-01 1.12134658e-01 -3.59765798e-01
2.62585700e-01 -1.08233953e+00 -1.49615839e-01 -3.63163650e-02
-5.61622500e-01 -4.92286950e-01 5.04117131e-01 -1.57302320e-02
1.01806939e+00 -1.89999115e+00 -1.10255487e-01 3.55111808e-01
-9.70012695e-02 8.80393982e-02 -5.75719727e-03 2.27319613e-01
-1.70057699e-01 4.45912540e-01 -4.63504553e-01 -5.29676735e-01
1.90609187e-01 1.21076725e-01 -3.96160185e-01 6.41579568e-01
2.35146448e-01 4.61981505e-01 -9.16643679e-01 -6.78971767e-01
1.13322223e-02 1.60713941e-01 -4.45774287e-01 1.68101802e-01
-4.62363601e-01 4.49727505e-01 -4.50978875e-01 5.81599295e-01
6.22169077e-01 -3.26752871e-01 3.40736210e-01 -2.88500953e-02
9.21946093e-02 6.88845515e-02 -1.61611664e+00 1.28564537e+00
-4.04933691e-01 1.25971064e-01 -4.52008367e-01 -1.27634752e+00
1.02737749e+00 3.97478044e-01 6.17378235e-01 -3.61967355e-01
1.52458638e-01 7.57775724e-01 -2.54790992e-01 -3.33630890e-01
4.34756011e-01 -5.76449990e-01 -5.78794666e-02 7.44409144e-01
1.59940615e-01 -1.06000908e-01 2.55329341e-01 -3.45138967e-01
7.21973300e-01 2.49851763e-01 2.60135293e-01 -4.57460076e-01
7.20621943e-01 -2.08231583e-01 8.53791833e-01 9.77711082e-01
-2.89549112e-01 7.75109768e-01 5.66579461e-01 -6.42005920e-01
-8.78470957e-01 -9.29130554e-01 -4.79267627e-01 6.63115203e-01
3.82934928e-01 -1.37004003e-01 -7.45873749e-01 -1.06380105e+00
-1.07434817e-01 7.32481062e-01 -5.17656863e-01 -8.98085237e-02
-3.62431973e-01 -9.12836134e-01 5.46774030e-01 1.59633219e-01
4.05939311e-01 -9.89258170e-01 -6.54477715e-01 2.45117664e-01
-1.62312374e-01 -8.02131414e-01 -7.97206014e-02 4.71841544e-01
-7.92712450e-01 -1.35479558e+00 -2.02047378e-01 -9.08063054e-01
5.10414779e-01 -1.41188368e-01 1.29995131e+00 2.26236805e-01
1.37517050e-01 5.76357841e-02 -3.27020228e-01 -2.86881387e-01
-8.51840854e-01 2.74604648e-01 2.05984861e-01 2.85820901e-01
1.88192934e-01 -4.60456014e-01 -1.96979135e-01 4.41468686e-01
-8.72641981e-01 1.24254242e-01 3.94220620e-01 1.05541170e+00
8.05405676e-01 2.35776111e-01 9.32922363e-01 -1.18508041e+00
4.61976856e-01 -3.94039094e-01 -7.67120659e-01 7.37370193e-01
-1.31360829e+00 3.70427519e-01 8.66537571e-01 -2.91316152e-01
-7.74859607e-01 3.88404131e-01 -3.39678764e-01 -3.11131738e-02
-1.81273744e-01 3.87299478e-01 -2.84750521e-01 1.19243478e-02
6.16010010e-01 -1.05721485e-02 -1.74061954e-01 -4.19671953e-01
2.53905326e-01 4.68105495e-01 2.09967151e-01 -7.88970113e-01
5.40682197e-01 4.27769542e-01 3.10379833e-01 -1.70218214e-01
-1.17271888e+00 -3.24659854e-01 -7.12865591e-01 -2.81978041e-01
8.45430613e-01 -2.66546786e-01 -6.54538214e-01 4.23788667e-01
-1.12269080e+00 -1.98570266e-02 -1.82623133e-01 3.32331896e-01
-9.62279201e-01 3.88823062e-01 -3.10919523e-01 -8.11614275e-01
-2.48223767e-01 -1.41380930e+00 8.53086531e-01 8.19957554e-02
-1.39623359e-01 -1.02403629e+00 6.73055649e-02 8.20583031e-02
8.90783668e-02 2.27430254e-01 1.11814117e+00 -1.01414967e+00
-3.55644792e-01 -2.92164475e-01 4.77646366e-02 5.19095719e-01
4.01979424e-02 3.31848674e-02 -1.09904695e+00 -3.65795374e-01
3.61504406e-01 -6.57188296e-01 8.32215965e-01 3.09472978e-01
1.18878341e+00 -2.76006669e-01 -4.16127741e-01 3.71772587e-01
1.68520665e+00 2.73068547e-01 3.39180827e-01 3.52632523e-01
2.96123832e-01 6.74611688e-01 6.45645261e-01 5.33595383e-02
2.83671528e-01 7.99889624e-01 5.91426075e-01 1.91123709e-01
9.20467377e-02 -2.72924662e-01 1.18129075e-01 8.27477217e-01
2.17799008e-01 -5.04132152e-01 -8.20233643e-01 4.41326648e-01
-2.00171185e+00 -9.35739160e-01 -1.81885183e-01 2.43590355e+00
8.66691887e-01 3.06627065e-01 9.79354084e-02 4.55389291e-01
8.04266095e-01 -1.00895844e-01 -4.60986942e-01 -3.16181600e-01
-2.17873484e-01 2.00379431e-01 3.04459929e-01 5.04171073e-01
-1.33379257e+00 5.21557391e-01 6.35099173e+00 8.40386450e-01
-1.05285239e+00 2.51883149e-01 8.75364542e-01 3.63813668e-01
-4.12476331e-01 2.64239311e-01 -1.08435535e+00 3.94903570e-01
7.94122696e-01 -9.51221064e-02 4.95232157e-02 1.08125532e+00
-3.59562546e-01 -1.83122724e-01 -1.43866646e+00 6.47183001e-01
1.77867785e-01 -1.18824387e+00 -1.45076498e-01 -1.66680366e-01
8.39628994e-01 -4.72930968e-01 1.16958886e-01 2.33279333e-01
2.82867670e-01 -1.04302168e+00 1.02668929e+00 5.39094687e-01
4.18380618e-01 -8.91619921e-01 1.00033772e+00 6.40756965e-01
-8.54482889e-01 -8.66185799e-02 -3.05446059e-01 2.76298940e-01
6.04789890e-02 7.88064480e-01 -6.85571373e-01 8.69431257e-01
3.31759125e-01 4.04517353e-01 -7.96488464e-01 1.06631958e+00
-2.35693291e-01 5.86465299e-01 -2.93916762e-01 -1.20420784e-01
1.08904429e-01 -5.88491298e-02 3.85281891e-01 9.31269169e-01
5.58314383e-01 -5.09360552e-01 2.71079838e-01 8.39800775e-01
1.19017184e-01 3.07973444e-01 -4.98731583e-01 2.45631203e-01
3.69567186e-01 1.20607984e+00 -9.81704473e-01 -1.81059670e-02
-2.97511220e-01 6.41757429e-01 4.86361533e-01 -1.80175141e-01
-9.17549133e-01 -5.51454239e-02 7.34519660e-02 -1.27123594e-01
-9.55527183e-03 2.38615200e-01 -5.10832489e-01 -1.05915701e+00
2.12464750e-01 -8.56065392e-01 5.18633306e-01 -3.95004779e-01
-1.67138147e+00 8.01398516e-01 -1.49237096e-01 -1.34959304e+00
-3.74183655e-01 -5.94411492e-01 -1.30106807e-01 6.03587687e-01
-1.46938932e+00 -1.04414225e+00 4.06423137e-02 3.04676890e-01
2.17218414e-01 -8.77364278e-02 1.01255310e+00 2.61192560e-01
-2.69248515e-01 7.42153406e-01 1.88787043e-01 -2.63286442e-01
6.78292930e-01 -1.41591191e+00 -1.92785367e-01 7.65173733e-01
4.18189406e-01 4.69213277e-01 7.71304369e-01 -3.61103773e-01
-7.23821223e-01 -1.08959877e+00 1.24830914e+00 -5.26580393e-01
4.48711932e-01 -2.29979604e-01 -9.65873837e-01 3.87590170e-01
-4.77541760e-02 1.84782162e-01 6.04382038e-01 1.40777230e-01
-4.25647944e-01 -3.42609257e-01 -1.19882095e+00 1.97441414e-01
8.53202283e-01 -2.46943384e-01 -5.17882049e-01 6.60355031e-01
3.93547565e-01 -3.68720800e-01 -6.36146009e-01 8.71217370e-01
6.68901861e-01 -1.35877168e+00 6.07939541e-01 -4.53766078e-01
4.18044209e-01 -3.82409632e-01 -2.45241582e-01 -1.21398020e+00
-7.08849775e-03 -6.67521060e-02 1.79987878e-01 1.36271083e+00
6.95674598e-01 -7.44908094e-01 6.77796602e-01 4.44878250e-01
-1.77060768e-01 -1.09985709e+00 -1.08986437e+00 -1.34780169e+00
2.52399296e-01 -5.03620982e-01 7.96676874e-01 1.04991066e+00
-2.38411173e-01 1.70858920e-01 -5.71137369e-01 1.46825328e-01
5.21701455e-01 5.73097765e-01 3.76190603e-01 -1.54802251e+00
-5.39602578e-01 -6.23871148e-01 -2.60656416e-01 -5.24213850e-01
5.28622389e-01 -1.25450432e+00 2.13321865e-01 -1.46531880e+00
2.22545743e-01 -1.24487412e+00 -4.48774397e-01 5.84713399e-01
-1.20707467e-01 1.29333153e-01 1.18063778e-01 2.51680940e-01
-5.92514098e-01 4.67125624e-01 7.99737275e-01 -1.31284148e-01
-3.99922468e-02 4.07047987e-01 -5.27616739e-01 8.26660514e-01
8.75569820e-01 -9.11742151e-01 -5.90190589e-01 4.18499665e-04
5.45683444e-01 -7.90362507e-02 3.33222032e-01 -1.09795022e+00
1.76797039e-03 -1.98555529e-01 -5.02589084e-02 -3.83950382e-01
2.53836140e-02 -8.55394423e-01 5.38512766e-01 5.68180799e-01
-6.49071991e-01 -5.35510778e-02 -2.44130448e-01 5.43227851e-01
-1.41024604e-01 -8.70330751e-01 1.12233365e+00 -1.32189080e-01
-5.46313226e-01 1.63781330e-01 1.05659319e-02 1.30630672e-01
1.16117167e+00 -1.19006447e-01 -1.90903559e-01 -8.50516036e-02
-5.25376439e-01 -1.65630772e-03 5.83705246e-01 3.77396524e-01
2.00212181e-01 -1.34912479e+00 -6.13212287e-01 8.27829763e-02
1.49572060e-01 -1.26591071e-01 -1.67659581e-01 6.87970638e-01
-3.39020729e-01 3.14623326e-01 6.96888864e-02 -6.79393232e-01
-1.06993115e+00 7.96557367e-01 5.69313526e-01 -5.92529774e-01
-3.00886095e-01 8.37034702e-01 1.88654289e-02 -7.00452924e-01
2.63021439e-01 -9.85372663e-02 -1.57839388e-01 5.74195106e-03
1.37939721e-01 7.13795498e-02 2.79431075e-01 -6.40179098e-01
-4.15590703e-01 6.75973892e-01 1.19207114e-01 -3.42216045e-02
1.11599505e+00 7.38944113e-02 -2.09874526e-01 7.61271417e-01
1.13674176e+00 -1.78855926e-01 -9.01905119e-01 -1.13708459e-01
5.52199006e-01 -2.74875432e-01 -2.71987021e-01 -6.92647576e-01
-1.08285165e+00 7.40931213e-01 6.61475718e-01 5.21720469e-01
1.19459140e+00 -2.29127824e-01 5.14620125e-01 3.19193661e-01
6.96384430e-01 -1.09661412e+00 2.29850076e-02 3.78850013e-01
5.76170206e-01 -1.15928054e+00 -3.94458473e-02 -5.43191612e-01
-5.50991535e-01 1.16518986e+00 5.48226535e-01 1.37919724e-01
5.99902391e-01 -4.49316762e-02 -9.97118503e-02 1.77546050e-02
-8.47213030e-01 -1.69462562e-01 2.47569725e-01 3.98288310e-01
3.64138842e-01 -7.09820241e-02 -7.70907283e-01 7.15055227e-01
-2.60806292e-01 1.18664421e-01 4.08798635e-01 6.07787788e-01
-3.86899889e-01 -1.45384240e+00 -1.02825768e-01 4.32536721e-01
-4.19281989e-01 1.87026426e-01 -2.32116774e-01 8.95482779e-01
6.46456242e-01 1.00824296e+00 -3.55989188e-01 -4.64372367e-01
8.22186098e-02 4.09378469e-01 5.09086549e-01 -5.26544511e-01
-4.89074707e-01 -2.96662122e-01 1.26925126e-01 -2.67398298e-01
-7.40148067e-01 -8.52917492e-01 -1.03672409e+00 -3.44203576e-03
-7.25439131e-01 3.21378946e-01 5.39531112e-01 1.17315495e+00
-2.28805374e-02 1.57007650e-02 8.22038293e-01 -2.78980613e-01
-8.91602576e-01 -6.05722070e-01 -7.58368194e-01 4.12736535e-01
5.18183447e-02 -9.24319446e-01 -5.15365541e-01 6.48386180e-02] | [9.11936092376709, 4.114560127258301] |
50343b90-6f76-4c5d-a375-2c6f877c3b8b | document-embedding-with-paragraph-vectors | 1507.07998 | null | http://arxiv.org/abs/1507.07998v1 | http://arxiv.org/pdf/1507.07998v1.pdf | Document Embedding with Paragraph Vectors | Paragraph Vectors has been recently proposed as an unsupervised method for
learning distributed representations for pieces of texts. In their work, the
authors showed that the method can learn an embedding of movie review texts
which can be leveraged for sentiment analysis. That proof of concept, while
encouraging, was rather narrow. Here we consider tasks other than sentiment
analysis, provide a more thorough comparison of Paragraph Vectors to other
document modelling algorithms such as Latent Dirichlet Allocation, and evaluate
performance of the method as we vary the dimensionality of the learned
representation. We benchmarked the models on two document similarity data sets,
one from Wikipedia, one from arXiv. We observe that the Paragraph Vector method
performs significantly better than other methods, and propose a simple
improvement to enhance embedding quality. Somewhat surprisingly, we also show
that much like word embeddings, vector operations on Paragraph Vectors can
perform useful semantic results. | ['Christopher Olah', 'Quoc V. Le', 'Andrew M. Dai'] | 2015-07-29 | null | null | null | null | ['document-embedding'] | ['methodology'] | [-2.07918063e-01 2.17272088e-01 -5.20612061e-01 -5.14511645e-01
-7.33298063e-01 -8.21764827e-01 1.05808485e+00 7.75737464e-01
-5.40052712e-01 3.44737262e-01 9.80593145e-01 -2.98173726e-01
-3.86331975e-02 -7.19066560e-01 -3.49639863e-01 -7.13973820e-01
2.60660708e-01 3.44416618e-01 1.19063750e-01 -2.44600713e-01
1.01855707e+00 1.96711421e-01 -1.43851435e+00 3.47257376e-01
2.24290237e-01 5.08476019e-01 -9.86604765e-02 5.95276117e-01
-7.05025256e-01 5.17361999e-01 -6.17590904e-01 -6.63082838e-01
-2.02478841e-02 5.42244874e-02 -7.59606600e-01 -2.67349817e-02
4.77123946e-01 -2.23438323e-01 -3.14409345e-01 9.03372765e-01
3.09363991e-01 5.64014435e-01 1.41255379e+00 -1.04011559e+00
-1.14034772e+00 8.11436176e-01 -5.77493489e-01 2.91648984e-01
2.50088394e-01 -5.41935265e-01 1.66745794e+00 -1.16009867e+00
7.66096413e-01 1.05043530e+00 6.99596822e-01 5.10752261e-01
-1.08235216e+00 -1.70986921e-01 5.08548990e-02 1.35073036e-01
-8.57433677e-01 -2.20784411e-01 8.40300798e-01 -6.07997417e-01
1.18628621e+00 -5.26334867e-02 4.00524795e-01 9.64609444e-01
2.98386693e-01 7.73057163e-01 5.47101021e-01 -7.10457861e-01
4.62092876e-01 6.05913401e-01 6.21722996e-01 5.38613141e-01
5.48404872e-01 -6.05803907e-01 -5.52725613e-01 -6.86571956e-01
2.58484930e-01 3.78472358e-01 -1.17225967e-01 -8.69084835e-01
-1.05439436e+00 1.60013402e+00 9.28367153e-02 4.98967946e-01
-1.53745100e-01 2.47414082e-01 6.89389706e-01 2.66510189e-01
9.51172411e-01 7.07920372e-01 -4.55600351e-01 -3.19849163e-01
-9.13247168e-01 2.88671225e-01 1.00303125e+00 6.56980574e-01
7.05082476e-01 -2.63471812e-01 -5.92775345e-02 1.03923154e+00
5.72626591e-01 1.18010245e-01 8.64752710e-01 -9.21771288e-01
3.50221097e-01 2.40429163e-01 8.99329036e-02 -1.46534264e+00
-1.12689883e-01 1.74037069e-01 -2.25425854e-01 -2.25661561e-01
2.00609773e-01 -1.88429028e-01 -4.17184204e-01 1.34970641e+00
-1.93608794e-02 -3.13919425e-01 8.16151723e-02 5.94593346e-01
6.38834476e-01 9.11070645e-01 -5.35764657e-02 1.22551555e-02
1.35193491e+00 -1.03830314e+00 -7.40481198e-01 7.79431909e-02
1.11979628e+00 -9.00762975e-01 1.04697382e+00 3.48715872e-01
-9.86975968e-01 -2.73842365e-01 -1.16299176e+00 -3.30387145e-01
-7.44384408e-01 -1.04590118e-01 1.13692272e+00 7.21234143e-01
-9.33930874e-01 7.26192534e-01 -6.26467228e-01 -7.92789936e-01
3.17233741e-01 -1.75307747e-02 -4.45707083e-01 -1.06397063e-01
-1.05352426e+00 7.51760125e-01 -1.53002828e-01 -6.30308867e-01
-5.26709139e-01 -7.10666180e-01 -9.19576466e-01 2.20271155e-01
-2.96619028e-01 -4.37126547e-01 1.17942107e+00 -6.70781374e-01
-1.36633229e+00 9.36273694e-01 -2.89324492e-01 -4.37839806e-01
-3.94371152e-01 -1.04141839e-01 -2.19769821e-01 5.06671548e-01
1.20101437e-01 4.09063518e-01 1.00924206e+00 -8.79107535e-01
-1.29365787e-01 -3.31485868e-01 2.17014208e-01 3.27069424e-02
-1.29204059e+00 1.04696043e-01 -1.93430632e-01 -7.06145704e-01
-1.10109404e-01 -7.83263922e-01 -3.20589930e-01 8.57212096e-02
-1.93149224e-02 -7.40801156e-01 3.90434355e-01 -3.61960441e-01
1.05617750e+00 -2.11361074e+00 1.74381614e-01 1.08162351e-01
4.86450195e-01 -9.75628644e-02 -1.37423605e-01 9.83976841e-01
-1.90121159e-01 3.69974405e-01 -1.64027795e-01 -6.19551897e-01
3.85245532e-01 1.31904274e-01 -6.31508529e-01 6.40107393e-01
-3.70205902e-02 9.10303295e-01 -8.37603211e-01 -2.49140292e-01
-1.68569297e-01 5.02540350e-01 -8.17497194e-01 -1.96869522e-01
-1.10674798e-01 -5.06567001e-01 -2.95249909e-01 1.57043442e-01
3.12631845e-01 -2.88069993e-01 2.61658996e-01 -2.06352174e-01
6.98703602e-02 4.28447306e-01 -7.15740204e-01 1.83966887e+00
-3.63156915e-01 1.25276816e+00 -3.83258402e-01 -1.28069186e+00
9.89609897e-01 3.29241931e-01 5.62127471e-01 -2.19887719e-01
2.70157307e-02 -2.88030088e-01 -4.60526586e-01 -5.39893091e-01
1.17217827e+00 -3.95260006e-01 -3.32055271e-01 1.02991796e+00
2.62365103e-01 -3.23173761e-01 3.32533531e-02 8.87923241e-01
1.11884296e+00 -3.62580448e-01 6.81453496e-02 -4.42494988e-01
-7.69159105e-03 2.01690182e-01 -2.06319422e-01 6.82246625e-01
1.01390295e-03 6.40995026e-01 8.80184770e-01 -3.05510670e-01
-1.31622970e+00 -1.00087535e+00 -3.07810962e-01 1.24550724e+00
-1.28707901e-01 -1.00811112e+00 -5.00346303e-01 -9.27953362e-01
4.62533385e-01 7.78544307e-01 -9.88196552e-01 -1.95956722e-01
-5.99637106e-02 -7.74958670e-01 3.61708760e-01 8.06894422e-01
-4.96819288e-01 -5.85455537e-01 5.56994528e-02 -4.37455587e-02
2.30278790e-01 -7.17638850e-01 -4.47997481e-01 9.82552320e-02
-9.92104173e-01 -8.33771646e-01 -9.64312792e-01 -8.78451884e-01
7.93363690e-01 6.80271268e-01 1.23570848e+00 -1.29091054e-01
-2.48477831e-01 1.12907422e+00 -6.23628974e-01 -4.04461771e-01
-1.93419605e-01 6.20082282e-02 3.80078048e-01 -3.50639999e-01
1.07175839e+00 -4.34420317e-01 -4.58077490e-01 3.71050946e-02
-1.20405626e+00 -7.70798504e-01 -2.56564468e-03 7.95300066e-01
1.63823172e-01 -2.87923664e-01 6.88893020e-01 -1.16165924e+00
1.20273483e+00 -8.48664761e-01 -1.10364988e-01 -1.18066929e-01
-8.53496194e-01 2.96055496e-01 4.01010871e-01 -2.92265356e-01
-6.43555403e-01 -4.75847930e-01 2.68436670e-02 -2.15769246e-01
1.93206206e-01 6.30295336e-01 4.48851109e-01 3.14485013e-01
7.08562732e-01 -1.11104712e-01 1.81509435e-01 -6.95204377e-01
8.89784336e-01 9.60228384e-01 -6.24783374e-02 -4.79439139e-01
6.50478184e-01 6.50502086e-01 -4.34277177e-01 -8.85715604e-01
-8.05068374e-01 -9.51058030e-01 -6.02957487e-01 3.42649937e-01
9.72402930e-01 -8.26203525e-01 -2.91814297e-01 -2.42390141e-01
-1.09525764e+00 1.54585674e-01 -5.11125326e-01 6.87034011e-01
-4.44238901e-01 7.32134700e-01 -6.46427095e-01 -3.32109898e-01
-1.13438800e-01 -4.81441885e-01 8.97871792e-01 1.21461917e-02
-4.69728440e-01 -1.64366639e+00 7.16243982e-01 2.21540332e-01
5.20144463e-01 -3.01606327e-01 1.22953808e+00 -1.16202581e+00
1.54571757e-01 -4.83942896e-01 -1.63968399e-01 5.44366419e-01
1.75028667e-01 9.48039070e-02 -9.00902271e-01 -3.12288880e-01
1.51437232e-02 -2.82710999e-01 1.17429328e+00 1.86924651e-01
9.95456159e-01 -1.69452533e-01 -3.02595139e-01 2.14396983e-01
1.42763913e+00 -5.56932330e-01 5.70541799e-01 4.42720473e-01
6.85749412e-01 6.86673522e-01 3.62249464e-01 5.72487175e-01
4.78866965e-01 4.68141168e-01 2.21602302e-02 2.49378443e-01
1.18927807e-01 -1.74751446e-01 4.48375136e-01 1.38931811e+00
1.51554495e-01 -1.33459508e-01 -6.37368739e-01 9.33964431e-01
-1.71677828e+00 -1.01024687e+00 -1.46547463e-02 1.93834126e+00
5.84982276e-01 -3.01875956e-02 8.69783908e-02 -2.87786443e-02
4.59812731e-01 5.27789354e-01 -9.40572247e-02 -1.09804201e+00
-2.27421597e-02 3.44090641e-01 4.91167247e-01 4.00011480e-01
-8.48092496e-01 7.40373015e-01 7.48404217e+00 5.73091626e-01
-6.32356703e-01 1.92392200e-01 8.48845765e-02 -2.76394635e-01
-9.83762085e-01 3.64684127e-02 -8.48594904e-01 3.11683685e-01
1.19483316e+00 -3.96994442e-01 -1.26563743e-01 1.03566241e+00
-2.83700041e-02 1.24655426e-01 -1.20626056e+00 7.64518440e-01
7.35645652e-01 -1.62170911e+00 2.58333832e-01 6.70758411e-02
1.01321530e+00 -1.13659417e-02 1.82744384e-01 4.37228918e-01
1.19076826e-01 -8.66019070e-01 -9.63910446e-02 4.89059985e-01
1.73549339e-01 -8.93253267e-01 7.35612929e-01 -4.38223295e-02
-6.56677246e-01 1.49046276e-02 -1.14621413e+00 -1.15164302e-01
3.09131686e-02 7.95055568e-01 -8.47861648e-01 1.30598187e-01
3.73684913e-01 1.19888306e+00 -7.58633614e-01 7.73206472e-01
-2.00606719e-01 7.96325922e-01 6.63186759e-02 -5.35633266e-01
3.94878179e-01 -1.11954048e-01 5.77390552e-01 1.34900951e+00
1.52058914e-01 -3.54766101e-01 -2.90268838e-01 4.49886441e-01
-2.57399768e-01 4.72696662e-01 -1.09830189e+00 -6.69466853e-01
3.65028352e-01 1.42900002e+00 -7.42650270e-01 -4.85516906e-01
-6.80051148e-01 1.00477004e+00 3.86550546e-01 2.78515100e-01
-5.43361902e-01 -8.25252771e-01 8.69960845e-01 -7.61463568e-02
8.99676919e-01 -5.73179722e-01 -2.40014866e-02 -1.51407695e+00
-1.45201251e-01 -5.05910695e-01 1.47292018e-01 -7.10330606e-01
-1.72681916e+00 1.47063836e-01 -9.73077267e-02 -1.06901598e+00
-3.81316423e-01 -8.21730256e-01 -6.23330534e-01 6.14010036e-01
-1.41115868e+00 -5.57927608e-01 3.35976124e-01 3.00764203e-01
5.99170625e-01 -4.66630101e-01 1.07692254e+00 9.12984908e-02
-1.41430542e-01 6.01972580e-01 8.93887520e-01 9.97280777e-02
1.05020583e+00 -1.47331548e+00 3.76151949e-01 3.01389277e-01
6.59230530e-01 1.10865390e+00 7.39143133e-01 -3.92973691e-01
-1.54603601e+00 -5.45527160e-01 1.11896586e+00 -9.21988487e-01
1.10409987e+00 -5.16862571e-01 -8.18532825e-01 6.55258596e-01
5.33956230e-01 -3.01312625e-01 1.41447031e+00 6.17909312e-01
-7.64387906e-01 3.92789245e-01 -9.07422900e-01 4.09916699e-01
4.18561578e-01 -8.17944884e-01 -1.16746068e+00 6.19000375e-01
1.06911910e+00 4.31604117e-01 -1.05920458e+00 -4.68897372e-01
6.55829906e-01 -7.52454340e-01 9.42919016e-01 -1.19440460e+00
9.51593220e-01 1.02943689e-01 -6.27608120e-01 -1.41351032e+00
-4.14386243e-01 -2.70422012e-01 -3.26694489e-01 1.35643125e+00
4.02847111e-01 -5.30632794e-01 8.90425146e-01 6.54244304e-01
1.35943860e-01 -7.46550322e-01 -4.60515022e-01 -6.28888667e-01
6.01414621e-01 -3.32408398e-01 3.16775799e-01 1.19613695e+00
5.28673589e-01 5.66929400e-01 -3.72691192e-02 -2.01993480e-01
5.00596523e-01 7.52674788e-02 7.85435915e-01 -1.37726307e+00
-3.13921452e-01 -4.47212607e-01 -6.76824272e-01 -1.20022082e+00
4.93417829e-01 -1.19056392e+00 -2.75282085e-01 -1.72428310e+00
2.89727271e-01 -6.90781372e-03 -4.06345695e-01 1.50004238e-01
1.45435389e-02 4.28858936e-01 1.17229395e-01 3.90491009e-01
-6.18302107e-01 5.87162018e-01 8.03457260e-01 -2.68838435e-01
1.64584428e-01 -3.59109670e-01 -1.23499441e+00 7.75649130e-01
6.89584255e-01 -7.02314377e-01 -3.61123264e-01 -5.72667778e-01
7.43557155e-01 -4.12331641e-01 -6.21445030e-02 -3.24668825e-01
2.28129029e-01 3.27236652e-01 3.39167386e-01 -4.55129266e-01
3.63561302e-01 -4.31639642e-01 -9.76519644e-01 -1.04148552e-01
-8.39097023e-01 1.88255578e-01 2.20311005e-02 9.79684472e-01
-2.11038381e-01 -7.96825945e-01 2.98460186e-01 -7.72016793e-02
-5.20338356e-01 4.09414954e-02 -7.28038907e-01 4.08784449e-02
7.16056883e-01 -5.95698990e-02 -3.08190286e-01 -5.76127410e-01
-5.79493165e-01 -1.14229366e-01 6.63047314e-01 5.68432093e-01
6.25424981e-01 -1.41410458e+00 -4.82338697e-01 -8.20904598e-02
3.85029852e-01 -6.43265188e-01 -6.85672462e-02 6.95753336e-01
-1.40669793e-01 5.48316360e-01 1.67264864e-01 -3.66731018e-01
-1.10225654e+00 7.88444102e-01 -3.45777631e-01 -5.03422134e-02
-6.98386788e-01 7.79177129e-01 5.78767583e-02 -4.99180049e-01
2.11631149e-01 -1.77487403e-01 -6.50323629e-01 7.84682453e-01
8.58769536e-01 1.73470348e-01 -8.39888118e-03 -2.99379706e-01
-3.27556133e-01 8.43069673e-01 -4.97132152e-01 -3.26148421e-01
1.79181266e+00 -3.10474962e-01 -2.41709441e-01 9.07512128e-01
1.88685369e+00 3.01034659e-01 -6.33416414e-01 -2.59065092e-01
1.41382888e-02 -4.68354344e-01 1.56437114e-01 -8.04015845e-02
-6.98453069e-01 1.12648821e+00 1.50674582e-01 7.78022468e-01
3.56261492e-01 2.23047271e-01 8.56162012e-01 8.51839364e-01
-2.78238337e-02 -1.10564446e+00 3.30596089e-01 5.52453279e-01
4.42088872e-01 -1.18985879e+00 4.96126801e-01 2.73390442e-01
-8.91190648e-01 1.37896228e+00 -1.23646706e-01 -6.00195229e-01
9.16891515e-01 -1.02372669e-01 -1.62510648e-01 -5.89496732e-01
-9.28297460e-01 1.63635060e-01 3.30744088e-01 6.39558196e-01
7.12955832e-01 -2.97781020e-01 -4.44879234e-01 7.74476767e-01
-1.31627470e-01 -3.37795675e-01 9.12165225e-01 9.88149524e-01
-8.31673801e-01 -1.16260445e+00 -7.56089538e-02 6.45720840e-01
-6.41827166e-01 -3.85289878e-01 -4.93382692e-01 4.18044776e-01
-5.85888565e-01 8.39287758e-01 3.39047760e-01 -1.34701803e-01
-1.40902728e-01 3.44996303e-01 4.33367550e-01 -1.02812099e+00
-3.95369679e-01 -3.56681943e-01 1.35278493e-01 -3.18464637e-01
-5.04605830e-01 -7.50020146e-01 -1.05154145e+00 -3.54975700e-01
-3.25044900e-01 6.66417003e-01 1.20952892e+00 9.44110990e-01
4.73259330e-01 2.22594216e-01 8.35021436e-01 -8.25962365e-01
-7.01705158e-01 -8.72366369e-01 -9.46079195e-01 4.72939104e-01
1.17998153e-01 -4.99405533e-01 -9.56559360e-01 1.45591628e-02] | [10.492572784423828, 8.619662284851074] |
d66ab527-8644-4d77-b9db-4e263035db9b | speech-enhancement-and-dereverberation-with | 2208.0583 | null | https://arxiv.org/abs/2208.05830v2 | https://arxiv.org/pdf/2208.05830v2.pdf | Speech Enhancement and Dereverberation with Diffusion-based Generative Models | In this work, we build upon our previous publication and use diffusion-based generative models for speech enhancement. We present a detailed overview of the diffusion process that is based on a stochastic differential equation and delve into an extensive theoretical examination of its implications. Opposed to usual conditional generation tasks, we do not start the reverse process from pure Gaussian noise but from a mixture of noisy speech and Gaussian noise. This matches our forward process which moves from clean speech to noisy speech by including a drift term. We show that this procedure enables using only 30 diffusion steps to generate high-quality clean speech estimates. By adapting the network architecture, we are able to significantly improve the speech enhancement performance, indicating that the network, rather than the formalism, was the main limitation of our original approach. In an extensive cross-dataset evaluation, we show that the improved method can compete with recent discriminative models and achieves better generalization when evaluating on a different corpus than used for training. We complement the results with an instrumental evaluation using real-world noisy recordings and a listening experiment, in which our proposed method is rated best. Examining different sampler configurations for solving the reverse process allows us to balance the performance and computational speed of the proposed method. Moreover, we show that the proposed method is also suitable for dereverberation and thus not limited to additive background noise removal. Code and audio examples are available online, see https://github.com/sp-uhh/sgmse | ['Timo Gerkmann', 'Bunlong Lay', 'Jean-Marie Lemercier', 'Simon Welker', 'Julius Richter'] | 2022-08-11 | null | null | null | null | ['speech-dereverberation'] | ['speech'] | [ 2.66058594e-01 1.48598075e-01 5.12094557e-01 -1.41291603e-01
-1.03222442e+00 -4.12063628e-01 7.09782779e-01 -2.26542249e-01
-6.18714929e-01 7.73788393e-01 2.42186308e-01 -3.09934258e-01
-1.13293469e-01 -5.13537467e-01 -3.84921193e-01 -1.03054512e+00
-1.87595077e-02 1.90157354e-01 3.85417879e-01 -3.45445842e-01
-1.24888599e-01 3.66862267e-01 -1.58041155e+00 8.11815485e-02
8.42006087e-01 6.86260879e-01 4.73773509e-01 1.08417439e+00
1.70371473e-01 4.09381717e-01 -1.04295754e+00 -4.54745233e-01
1.46666661e-01 -8.06650996e-01 -5.29549181e-01 8.37210417e-02
1.74422562e-01 -9.29290652e-02 -2.08003104e-01 1.21867204e+00
1.23999524e+00 3.57411057e-01 5.63152134e-01 -7.23477662e-01
-4.95367944e-01 6.61657274e-01 -1.02019794e-01 3.66126716e-01
2.58041650e-01 6.52745962e-02 6.95460975e-01 -8.36040318e-01
5.26954949e-01 1.01025891e+00 8.07802379e-01 7.65798330e-01
-1.24795437e+00 -4.72543955e-01 -4.39523868e-02 1.14170969e-01
-1.28002739e+00 -9.24508631e-01 8.81622612e-01 -1.40977889e-01
9.14212525e-01 4.08703446e-01 5.62744737e-01 1.54566050e+00
-2.28913963e-01 7.71016479e-01 1.25652409e+00 -6.91225350e-01
3.67890090e-01 3.08344543e-01 8.93765036e-03 2.00134739e-01
2.39266064e-02 3.29049766e-01 -4.32024270e-01 -6.93466514e-02
4.15861130e-01 -6.53527319e-01 -6.55415893e-01 -1.60129100e-01
-9.03200746e-01 7.10287035e-01 4.32175249e-02 8.37881863e-01
-3.45244944e-01 7.51144141e-02 2.94018626e-01 3.59962553e-01
7.04271555e-01 1.47916272e-01 -1.97505549e-01 -4.50341702e-01
-1.43340790e+00 2.59429783e-01 1.08742738e+00 7.39886701e-01
3.85595143e-01 5.20924211e-01 -9.58274677e-02 1.17696011e+00
2.53617227e-01 7.27689803e-01 5.54994941e-01 -9.73524094e-01
2.11062685e-01 -5.38983047e-01 -4.65575643e-02 -5.35474539e-01
-2.13700324e-01 -8.48189831e-01 -9.35719132e-01 2.41271272e-01
5.30542791e-01 -5.54834664e-01 -8.85932326e-01 1.77441573e+00
7.72107616e-02 1.08047530e-01 1.33176014e-01 7.19935715e-01
4.78272110e-01 6.76856995e-01 -2.00878099e-01 -5.91902435e-01
9.77529347e-01 -7.42335856e-01 -1.14504957e+00 4.56857793e-02
3.19123149e-01 -1.09403610e+00 7.32406020e-01 7.67334878e-01
-1.34970415e+00 -5.90412199e-01 -1.02285647e+00 2.96399295e-01
-2.48109207e-01 1.37919769e-01 2.83226550e-01 1.24441397e+00
-1.53712749e+00 9.08527672e-01 -7.65748024e-01 -4.50556248e-01
6.48357719e-02 2.01490343e-01 -4.31793779e-02 1.85908765e-01
-1.35875177e+00 7.75960743e-01 9.91555527e-02 2.48445004e-01
-7.24452436e-01 -3.42330486e-01 -7.42090642e-01 -6.81135431e-02
1.64396048e-01 -5.86415648e-01 1.49073946e+00 -9.19298589e-01
-2.00147724e+00 4.58430767e-01 -3.39734912e-01 -7.32387543e-01
6.99157238e-01 -2.93864667e-01 -7.10311532e-01 1.27124742e-01
-3.62453848e-01 2.49395207e-01 1.04788876e+00 -1.33133209e+00
-3.39281827e-01 5.86918592e-02 -3.23173761e-01 6.20779507e-02
-2.22293302e-01 1.72144808e-02 -3.32442403e-01 -8.69061291e-01
-5.50351627e-02 -9.00801718e-01 -1.56575561e-01 -5.29125690e-01
-1.95093498e-01 9.42045972e-02 6.13592207e-01 -7.93207288e-01
1.28854740e+00 -2.28398681e+00 1.13389291e-01 2.30201975e-01
-1.33273706e-01 6.32914484e-01 6.30436931e-03 6.46757424e-01
-2.67994702e-01 1.71431042e-02 -4.93196607e-01 -9.70723391e-01
1.08415052e-01 1.65559545e-01 -3.60257357e-01 5.69106400e-01
-2.10132189e-02 4.70069587e-01 -8.12631547e-01 -2.45758623e-01
1.34379700e-01 9.42697585e-01 -5.43695569e-01 -4.37323935e-03
1.85837492e-01 3.84320468e-01 1.64761275e-01 2.05714539e-01
8.04932356e-01 3.93715352e-01 5.19148782e-02 -2.91652028e-02
-1.21405341e-01 3.44627172e-01 -1.57116199e+00 1.70820594e+00
-6.19641006e-01 8.22523534e-01 4.56604064e-01 -9.36749578e-01
9.68304217e-01 6.57783628e-01 1.43872827e-01 -4.70579028e-01
4.67411652e-02 4.47036952e-01 2.50804722e-01 -3.32139909e-01
4.03543293e-01 -4.65074658e-01 3.98383856e-01 2.63043940e-01
4.76353526e-01 -4.40757930e-01 2.39876255e-01 2.14493111e-01
9.25650895e-01 1.77106962e-01 1.39843956e-01 -3.27887535e-01
4.90717173e-01 -5.91688454e-01 3.07938963e-01 8.89885962e-01
-2.86681294e-01 8.93001437e-01 2.94944614e-01 3.50090146e-01
-8.49279284e-01 -1.20812321e+00 -3.09280455e-01 7.67680705e-01
-2.48113140e-01 -4.59899396e-01 -1.08592308e+00 -2.37970486e-01
-5.82302451e-01 9.61160541e-01 -4.38303441e-01 -2.02309410e-03
-4.99487191e-01 -1.05850339e+00 7.93232560e-01 1.80396229e-01
3.22362572e-01 -1.02119339e+00 -8.10138360e-02 2.44381875e-01
-2.04261646e-01 -8.61946166e-01 -3.82747263e-01 5.19190550e-01
-7.33347952e-01 -4.93204921e-01 -1.13522661e+00 -5.41466534e-01
2.80424178e-01 -5.55930696e-02 8.61009896e-01 -1.96151704e-01
9.49155018e-02 5.15906692e-01 -4.49372798e-01 -4.00005490e-01
-1.06244981e+00 6.18856028e-02 1.69143409e-01 3.85501720e-02
4.53641675e-02 -9.07654405e-01 -5.51153123e-01 3.45724784e-02
-9.69298482e-01 -3.09626460e-01 5.70956349e-01 9.09426391e-01
1.90955579e-01 2.65367955e-01 6.33541286e-01 -7.34119892e-01
9.36279118e-01 -2.85336882e-01 -4.75831538e-01 -3.32669675e-01
-6.88600183e-01 1.51555449e-01 4.63981092e-01 -5.17190218e-01
-1.30826294e+00 -7.55842477e-02 -9.67853606e-01 -2.36171067e-01
-4.09356058e-01 1.33830175e-01 -1.93953931e-01 1.65140778e-01
7.92395234e-01 3.64122927e-01 -1.97112951e-02 -7.75458992e-01
3.68567854e-01 7.95839787e-01 3.81572932e-01 -3.30871552e-01
8.12607467e-01 2.94942111e-01 -2.20667899e-01 -1.18203759e+00
-3.11410278e-01 -4.23833132e-01 -4.15220380e-01 -1.93019032e-01
6.45928204e-01 -7.20026255e-01 -3.78055632e-01 6.56407416e-01
-1.23519135e+00 -4.56430525e-01 -6.09963477e-01 6.54337704e-01
-6.58165753e-01 5.16929626e-01 -8.33551228e-01 -1.36064291e+00
-2.25070998e-01 -1.07575142e+00 8.83623123e-01 7.81263877e-03
-9.14619341e-02 -1.19933414e+00 3.78631681e-01 2.09594779e-02
7.38913834e-01 -3.29300493e-01 3.31682742e-01 -8.23872924e-01
-1.50555477e-01 -5.02515323e-02 3.18913251e-01 8.82269263e-01
2.26342931e-01 -4.55181450e-02 -1.52681577e+00 -3.45627517e-01
5.48381150e-01 2.27921665e-01 1.08199096e+00 4.66078520e-01
6.33328915e-01 -7.77488798e-02 -7.33922468e-03 5.39672434e-01
1.25509369e+00 2.40697965e-01 9.49455738e-01 8.90977755e-02
2.16068849e-01 6.10159218e-01 2.30297849e-01 3.12645555e-01
-1.94301620e-01 7.51763582e-01 1.37291789e-01 -2.91991889e-01
-6.23468637e-01 -1.34415999e-01 7.16882646e-01 1.18214667e+00
-2.97329068e-01 -4.75117922e-01 -6.19972229e-01 6.21308327e-01
-1.45268679e+00 -1.27584445e+00 -1.72476768e-01 2.32902503e+00
8.43619049e-01 2.85592914e-01 2.01354727e-01 6.50096595e-01
6.39548779e-01 1.49038285e-01 4.73152176e-02 -4.69089001e-01
-3.37413996e-01 6.46948874e-01 3.69138181e-01 1.03121507e+00
-9.26428020e-01 6.43646300e-01 6.75616980e+00 1.10945404e+00
-1.13576210e+00 4.41955507e-01 3.73204827e-01 -2.43704975e-01
-2.10598990e-01 -2.31566146e-01 -6.39157534e-01 4.13461924e-01
1.28790176e+00 3.32574956e-02 4.37734783e-01 5.42727828e-01
4.75496978e-01 -1.56121820e-01 -7.38201916e-01 8.03221703e-01
8.49420205e-02 -8.46404433e-01 -2.88249493e-01 1.04896583e-01
5.11385560e-01 1.69985324e-01 1.21178150e-01 1.86776623e-01
1.62568256e-01 -7.79417813e-01 8.06511819e-01 4.23650771e-01
4.16840076e-01 -5.68538010e-01 7.41846740e-01 4.06553864e-01
-8.69881332e-01 1.38095513e-01 -1.22275218e-01 7.27336034e-02
3.77427667e-01 8.26337039e-01 -7.40894914e-01 7.53721416e-01
5.86650133e-01 3.33516210e-01 -2.76315987e-01 1.12795317e+00
-3.95452350e-01 9.78903949e-01 -3.79925996e-01 -2.80188024e-02
-8.67368933e-03 -1.66323200e-01 1.08558917e+00 1.73829722e+00
7.23032594e-01 -3.41153890e-01 -5.52668452e-01 7.26923406e-01
3.36122572e-01 1.16361059e-01 -5.35277545e-01 2.72980928e-01
3.39638963e-02 1.10280037e+00 -5.66851377e-01 -2.80464053e-01
-2.33888835e-01 1.19423962e+00 -1.64640304e-02 7.19531775e-01
-8.48092735e-01 -5.43636680e-01 5.10793388e-01 7.15503395e-02
6.96020186e-01 -3.21843237e-01 -6.86018320e-04 -1.06739330e+00
3.47096403e-03 -9.71603751e-01 -3.76096033e-02 -7.16817975e-01
-9.51152861e-01 1.03650284e+00 3.88718545e-02 -1.08730888e+00
-5.31039655e-01 -4.45136726e-01 -4.78315771e-01 1.32596648e+00
-1.38093877e+00 -5.80837905e-01 1.13191508e-01 4.52297300e-01
5.46111882e-01 2.98860129e-02 8.41498911e-01 6.00896001e-01
-4.07243907e-01 4.96214688e-01 2.98306853e-01 -2.15397850e-01
7.53866732e-01 -1.42534208e+00 5.17682433e-01 1.23322511e+00
5.21285772e-01 6.32215798e-01 1.23564088e+00 -3.86739701e-01
-7.65901029e-01 -6.58099651e-01 8.81862640e-01 -2.14028239e-01
7.13244319e-01 -6.60184443e-01 -9.71230328e-01 8.62058252e-02
7.06163466e-01 -3.57589036e-01 5.80743432e-01 9.88572184e-03
7.25091398e-02 -5.71768619e-02 -9.61433649e-01 4.78268385e-01
1.11442506e+00 -4.12620068e-01 -5.39086819e-01 1.08264267e-01
5.28037250e-01 -4.10952777e-01 -5.43798029e-01 1.22820809e-01
2.16557637e-01 -1.29602206e+00 7.29957581e-01 -3.84677202e-02
-8.87714028e-02 -3.21025223e-01 -2.07523584e-01 -1.66571558e+00
-6.73448369e-02 -1.19835508e+00 -1.25378268e-02 1.51650512e+00
6.08210206e-01 -7.88267672e-01 3.54057997e-01 1.38039857e-01
-2.46720374e-01 -5.42004168e-01 -1.06439567e+00 -9.34658051e-01
1.41889900e-01 -9.24840391e-01 2.05788955e-01 4.38595682e-01
-1.76611096e-01 2.53856242e-01 -5.22390187e-01 2.35101908e-01
5.70137501e-01 -3.80560607e-01 4.97100919e-01 -8.23569059e-01
-8.11276734e-01 -5.50817847e-01 -2.17089117e-01 -1.25966883e+00
-1.32089198e-01 -7.10308313e-01 3.22342157e-01 -1.38426757e+00
-3.33228350e-01 -1.94990531e-01 -1.61072090e-01 -2.80574225e-02
-1.71074659e-01 2.74911046e-01 2.48859555e-01 -9.35149342e-02
-7.64487460e-02 6.14687562e-01 1.06183934e+00 7.61529058e-02
-3.46770674e-01 4.79251862e-01 -6.19315028e-01 5.76868474e-01
8.52710605e-01 -5.17869890e-01 -4.61329669e-01 -2.13058233e-01
1.62215214e-02 -1.14293478e-01 2.69742221e-01 -1.18706214e+00
2.71790594e-01 5.32247066e-01 7.90137500e-02 -6.37068376e-02
6.39214575e-01 -6.46905184e-01 2.07952797e-01 4.49131012e-01
-1.84416428e-01 -2.77716100e-01 3.99162471e-01 5.13425052e-01
-3.21963161e-01 -6.07205808e-01 9.03578758e-01 1.78109095e-01
-2.60768175e-01 -1.46788552e-01 -6.69886470e-01 -1.58069208e-01
4.85697508e-01 -1.89226680e-02 -1.04153603e-02 -8.65110755e-01
-1.24033546e+00 -5.44906497e-01 2.34043390e-01 1.90421790e-01
3.45977873e-01 -1.04338920e+00 -8.15756559e-01 2.34348059e-01
-4.65137959e-01 -4.23407942e-01 4.28311914e-01 1.02281344e+00
-1.98629975e-01 2.34897599e-01 3.69190603e-01 -4.75284368e-01
-1.34019816e+00 4.59471971e-01 6.49135709e-01 -1.43689722e-01
-4.72397923e-01 7.75408328e-01 3.93337151e-03 -1.07057340e-01
3.56437206e-01 -3.38794202e-01 -8.06247890e-02 4.30244952e-02
4.87404197e-01 3.50462317e-01 3.93480301e-01 -6.51141167e-01
-1.75698951e-01 2.91504681e-01 2.31178209e-01 -7.27534115e-01
1.20909154e+00 -2.70569384e-01 1.84032053e-01 6.19350910e-01
1.07909989e+00 6.60395086e-01 -1.06068194e+00 -6.67909682e-02
-2.95802444e-01 -1.20251700e-01 1.94962233e-01 -7.30363846e-01
-9.73772287e-01 9.66138303e-01 9.49928820e-01 6.71471775e-01
1.44799852e+00 -4.91599776e-02 4.69561547e-01 1.20390020e-01
1.34573579e-01 -1.03935087e+00 -2.12153241e-01 3.91659856e-01
9.67805088e-01 -9.10469055e-01 -2.92670518e-01 -4.07785028e-01
-4.31649357e-01 9.95972514e-01 -2.66045332e-02 6.09743148e-02
7.95324802e-01 7.12269127e-01 2.78151542e-01 2.05828398e-01
-6.07120395e-01 -5.81641793e-01 1.45088255e-01 9.76852357e-01
6.07405961e-01 -1.24455698e-01 -3.43410939e-01 4.90506649e-01
-4.12641406e-01 -1.37354091e-01 5.48034072e-01 6.91119909e-01
-3.21285576e-01 -1.55909657e+00 -5.21997869e-01 2.54915468e-02
-6.48284316e-01 -4.89203930e-01 -2.59930313e-01 7.14757860e-01
5.53764068e-02 1.18940997e+00 -2.47836411e-01 -1.78986222e-01
4.36900705e-01 3.36895227e-01 5.28850138e-01 -4.08607036e-01
-6.27096176e-01 5.46636999e-01 2.19530270e-01 -2.45647728e-01
-6.57390237e-01 -9.03520882e-01 -7.30614245e-01 -1.10873543e-01
-5.00677705e-01 3.96567345e-01 8.95473242e-01 6.97901726e-01
1.56237766e-01 8.20207834e-01 4.00680721e-01 -8.85332525e-01
-5.06693304e-01 -1.33463490e+00 -7.27623284e-01 2.47444570e-01
5.36220849e-01 -3.64129841e-01 -8.73556852e-01 1.79430544e-01] | [15.07546615600586, 5.960217475891113] |
b0416241-4377-4426-8f14-0cdefa16c6f9 | unifying-event-detection-and-captioning-as | 2207.08625 | null | https://arxiv.org/abs/2207.08625v1 | https://arxiv.org/pdf/2207.08625v1.pdf | Unifying Event Detection and Captioning as Sequence Generation via Pre-Training | Dense video captioning aims to generate corresponding text descriptions for a series of events in the untrimmed video, which can be divided into two sub-tasks, event detection and event captioning. Unlike previous works that tackle the two sub-tasks separately, recent works have focused on enhancing the inter-task association between the two sub-tasks. However, designing inter-task interactions for event detection and captioning is not trivial due to the large differences in their task specific solutions. Besides, previous event detection methods normally ignore temporal dependencies between events, leading to event redundancy or inconsistency problems. To tackle above the two defects, in this paper, we define event detection as a sequence generation task and propose a unified pre-training and fine-tuning framework to naturally enhance the inter-task association between event detection and captioning. Since the model predicts each event with previous events as context, the inter-dependency between events is fully exploited and thus our model can detect more diverse and consistent events in the video. Experiments on the ActivityNet dataset show that our model outperforms the state-of-the-art methods, and can be further boosted when pre-trained on extra large-scale video-text data. Code is available at \url{https://github.com/QiQAng/UEDVC}. | ['Qin Jin', 'Yuqing Song', 'Qi Zhang'] | 2022-07-18 | null | null | null | null | ['dense-video-captioning'] | ['computer-vision'] | [ 4.06960875e-01 -1.24583971e-02 -5.59411198e-02 -4.19405758e-01
-8.90069544e-01 -4.46286827e-01 6.82667255e-01 3.80580984e-02
-2.90003628e-01 6.86019838e-01 6.46320820e-01 1.64545834e-01
2.47694641e-01 -4.52467710e-01 -7.69901335e-01 -3.26981664e-01
-1.57655790e-01 2.35561490e-01 5.13390422e-01 8.28767195e-02
-7.21409991e-02 -1.23911984e-01 -1.54286849e+00 7.44032323e-01
6.56552732e-01 9.77909625e-01 4.91613358e-01 6.85129285e-01
-1.03444502e-01 1.09139478e+00 -5.81319094e-01 -2.25336269e-01
-1.67143971e-01 -8.58817279e-01 -6.14458561e-01 2.80977011e-01
1.76307365e-01 -4.52422053e-01 -6.92389429e-01 7.63213217e-01
5.21229982e-01 2.31891811e-01 3.70415211e-01 -1.59467137e+00
-4.76726323e-01 6.05801702e-01 -4.04334486e-01 4.97524053e-01
5.60657561e-01 9.20787752e-02 9.60004747e-01 -8.27456713e-01
6.14135623e-01 9.47110474e-01 4.24086392e-01 7.01782227e-01
-8.82458031e-01 -6.86954975e-01 4.04330522e-01 6.98840976e-01
-1.27060032e+00 -4.65510428e-01 7.36291230e-01 -5.05980432e-01
9.46048081e-01 1.63752183e-01 5.85600674e-01 1.64459538e+00
-5.95387742e-02 1.08258510e+00 4.35046792e-01 -6.83555752e-02
1.16011895e-01 -2.14808375e-01 -1.05212063e-01 2.37153128e-01
-2.17977744e-02 -1.61627337e-01 -7.14307725e-01 -1.02313636e-02
7.85203338e-01 1.09937772e-01 -5.57521284e-01 -2.93840673e-02
-1.69925714e+00 5.73996544e-01 1.14941135e-01 3.89168888e-01
-6.29434764e-01 1.66370451e-01 8.32786679e-01 -6.97768200e-03
4.81887311e-01 1.13672271e-01 -4.33063537e-01 -4.39092606e-01
-9.49268937e-01 2.83180714e-01 6.25980496e-01 1.08633852e+00
4.16846782e-01 -7.51818195e-02 -8.34218562e-01 7.72861421e-01
1.13663428e-01 1.54275239e-01 4.76889580e-01 -6.91135824e-01
7.69456267e-01 3.53021145e-01 2.39573687e-01 -8.92161191e-01
-2.61376470e-01 -3.11701864e-01 -6.83401763e-01 -5.91837168e-01
3.25069487e-01 -3.73851091e-01 -9.14399564e-01 2.06679535e+00
2.33307689e-01 9.11635876e-01 -7.47724175e-02 1.07574189e+00
1.00070250e+00 1.05462670e+00 4.28537101e-01 -4.97401655e-01
1.65166652e+00 -1.18490112e+00 -1.10229778e+00 -5.54303646e-01
3.45245987e-01 -7.49730587e-01 8.07768106e-01 3.44888633e-03
-1.09623539e+00 -7.12368190e-01 -8.31891537e-01 -2.02777516e-02
8.73182062e-03 2.12523252e-01 4.43195194e-01 -1.56000763e-01
-7.40158796e-01 1.62420675e-01 -8.49334121e-01 -4.37416911e-01
3.26504767e-01 -4.81387600e-02 -2.36988142e-01 -1.28867840e-02
-1.59518087e+00 5.07586181e-01 7.59381115e-01 1.10464104e-01
-1.10712218e+00 -6.46317720e-01 -1.05099940e+00 1.49279982e-01
7.64315426e-01 -6.33433044e-01 1.48814082e+00 -1.04017091e+00
-1.13795412e+00 5.89819729e-01 -4.98733252e-01 -3.78130168e-01
4.06635970e-01 -2.24088266e-01 -5.88389575e-01 3.26465636e-01
2.58259982e-01 8.58513653e-01 7.44777799e-01 -9.78473604e-01
-9.42126393e-01 4.43504192e-02 4.01812308e-02 4.49689358e-01
-2.81096697e-01 3.08511019e-01 -1.08233345e+00 -1.01476848e+00
-2.46673673e-01 -7.42214322e-01 -4.58723791e-02 -2.64220297e-01
-1.89641282e-01 -3.63983840e-01 9.19062912e-01 -7.21064866e-01
1.41839814e+00 -2.26637220e+00 8.33215266e-02 -4.53084767e-01
5.20153344e-02 3.27239096e-01 -2.91208714e-01 4.73928154e-01
-2.78828889e-01 -7.34802485e-02 -5.04598506e-02 -5.88064909e-01
-8.49393606e-02 2.03513384e-01 -3.75233531e-01 1.21045023e-01
4.69776809e-01 8.86849225e-01 -1.23886728e+00 -6.84874594e-01
2.17535719e-01 6.86141133e-01 -2.31301591e-01 4.97969747e-01
-4.77366865e-01 7.25477397e-01 -5.97860456e-01 3.28420103e-01
3.19988251e-01 -5.01705527e-01 1.20169409e-01 -2.68808484e-01
8.99273157e-02 4.03245449e-01 -1.15951085e+00 1.93286312e+00
-3.16104144e-01 7.67528534e-01 -2.47324660e-01 -1.02806139e+00
4.98291641e-01 9.74024296e-01 6.82751954e-01 -6.95206523e-01
9.17326957e-02 -1.13670183e-02 -2.18470886e-01 -8.01459432e-01
4.73288238e-01 9.22325626e-02 -1.80690199e-01 3.87357831e-01
2.08734989e-01 5.47029376e-01 6.35242164e-01 3.48893851e-01
1.08001077e+00 4.28698987e-01 3.46071005e-01 3.43236893e-01
5.18448412e-01 -8.35693404e-02 8.26768577e-01 7.27630794e-01
-3.61862421e-01 9.72757518e-01 5.53536654e-01 -3.38234872e-01
-8.92749846e-01 -9.23281610e-01 1.12977363e-01 1.20329821e+00
3.77121806e-01 -6.75844014e-01 -6.43041790e-01 -8.57017457e-01
-4.78355318e-01 6.85259938e-01 -6.73925638e-01 -5.92275262e-02
-7.30420709e-01 -7.65777826e-01 4.44873244e-01 6.85848415e-01
5.32939076e-01 -1.39670706e+00 -5.83442807e-01 5.68970621e-01
-9.80456233e-01 -1.66962385e+00 -8.75008047e-01 -1.33030459e-01
-4.19097841e-01 -1.01732516e+00 -8.38391244e-01 -8.74200463e-01
3.67369503e-01 2.31815442e-01 1.33649778e+00 -1.11762889e-01
-1.07886598e-01 3.59307021e-01 -7.54101455e-01 -3.01162362e-01
-3.39184582e-01 8.39869608e-04 -3.06450635e-01 3.01056474e-01
5.02099574e-01 -5.53265572e-01 -7.77444780e-01 4.62257594e-01
-1.08350706e+00 6.26802504e-01 4.86734241e-01 5.73006988e-01
6.69329405e-01 -1.76677316e-01 7.78918982e-01 -5.28187871e-01
2.20619053e-01 -7.61870980e-01 -2.05357820e-01 3.02008539e-01
-4.46985252e-02 -1.87847987e-01 4.97681499e-01 -6.57985806e-01
-1.28710032e+00 1.78997517e-01 -2.60617733e-02 -5.88824093e-01
-3.80559117e-01 4.12883610e-01 -3.07934314e-01 6.44233286e-01
2.30917856e-01 4.98154908e-01 -4.57011402e-01 -2.05931127e-01
6.32603392e-02 4.65949774e-01 7.36893177e-01 -3.46265882e-01
5.38096607e-01 3.76084507e-01 -3.05378437e-01 -5.00229180e-01
-1.11261785e+00 -5.64765573e-01 -2.74100214e-01 -4.15035456e-01
1.15723312e+00 -1.25005114e+00 -3.41731071e-01 4.59626913e-01
-1.44925952e+00 -2.97837138e-01 -2.05742911e-01 5.94522536e-01
-6.71983004e-01 5.04126072e-01 -6.42555773e-01 -5.37968755e-01
-2.52017498e-01 -1.06869900e+00 1.24899805e+00 3.15302640e-01
-2.44918600e-01 -8.90824318e-01 1.23362519e-01 2.70568430e-01
1.21969931e-01 3.27848107e-01 3.93678427e-01 -8.51141274e-01
-5.07062316e-01 -1.54501319e-01 -3.14359546e-01 1.31272972e-01
1.00634009e-01 -1.95212260e-01 -8.01274478e-01 -9.07405391e-02
-2.71583479e-02 -2.17840672e-01 8.40503633e-01 3.91682416e-01
1.22616303e+00 -2.58854121e-01 -3.61901522e-01 4.00983810e-01
1.07436728e+00 3.14336687e-01 8.34878683e-01 2.53921330e-01
7.58657038e-01 5.28096497e-01 8.14531147e-01 6.74737513e-01
6.45559907e-01 9.59623218e-01 3.37012053e-01 -7.65704587e-02
-2.14513198e-01 -3.64970326e-01 4.71299738e-01 4.56756413e-01
5.04046585e-03 -7.79072404e-01 -7.16961563e-01 8.05690527e-01
-2.40620732e+00 -1.36344337e+00 -2.11523816e-01 1.98097134e+00
8.89034390e-01 7.83110317e-03 1.78884059e-01 -1.28056005e-01
1.07747948e+00 4.16274548e-01 -4.74487185e-01 1.65085062e-01
-5.16929552e-02 -2.38231391e-01 3.14147845e-02 1.17812268e-01
-1.40466380e+00 8.16550910e-01 5.26438618e+00 8.15120101e-01
-1.00530088e+00 3.38774681e-01 5.86545885e-01 -4.54215139e-01
9.51349176e-03 -5.73607497e-02 -7.09383547e-01 9.14461553e-01
9.98564243e-01 -2.60937005e-01 7.37074539e-02 5.21472216e-01
5.40047586e-01 4.37314995e-02 -1.36492658e+00 1.16153991e+00
2.37638652e-01 -1.20438993e+00 2.02796072e-01 -3.29989254e-01
5.98925531e-01 -7.70524442e-02 -3.02316755e-01 4.97416347e-01
-1.75595000e-01 -6.04677022e-01 9.78120565e-01 3.25447917e-01
6.20639145e-01 -4.56426054e-01 6.98074937e-01 1.31600276e-01
-1.53634059e+00 1.22320630e-01 1.12183683e-01 -1.07017919e-01
8.64690185e-01 5.13741195e-01 -5.97735345e-01 7.01074839e-01
6.82446897e-01 9.61515427e-01 -3.81211668e-01 1.13877940e+00
-3.10074627e-01 6.68179214e-01 -1.64881870e-01 2.92242527e-01
1.70881033e-01 1.21168576e-01 7.13843942e-01 1.47007501e+00
3.94171715e-01 1.49312928e-01 3.66795719e-01 6.89436674e-01
-1.08439505e-01 -1.24423437e-01 -1.86338350e-01 -7.96559229e-02
4.35372978e-01 1.23891532e+00 -7.73767531e-01 -6.54509962e-01
-6.68017387e-01 1.19469726e+00 1.26357168e-01 4.73705143e-01
-1.55229616e+00 -3.38268578e-01 5.40359557e-01 9.18353796e-02
5.43552160e-01 -3.91670465e-02 2.08059907e-01 -1.52072489e+00
3.09997886e-01 -7.12605000e-01 6.59985006e-01 -9.62352514e-01
-1.08831763e+00 6.69669867e-01 2.13561267e-01 -1.36477733e+00
-4.67860013e-01 -1.24740556e-01 -7.19542146e-01 5.22834718e-01
-1.41930914e+00 -1.13033247e+00 -5.52291036e-01 6.96011126e-01
1.05691528e+00 3.38701516e-01 4.35956836e-01 7.56281793e-01
-8.16995859e-01 4.04304475e-01 -3.41181129e-01 2.37402216e-01
9.21997845e-01 -1.01185763e+00 3.66869152e-01 1.19539785e+00
1.57714397e-01 -5.77699281e-02 7.18296945e-01 -8.29890430e-01
-9.48340118e-01 -1.52093828e+00 1.08799672e+00 -3.74185681e-01
6.65147364e-01 -5.08828700e-01 -1.02521348e+00 8.94126713e-01
2.71708608e-01 -3.24054505e-04 5.74032843e-01 -2.25717559e-01
-2.63944000e-01 1.45036742e-01 -5.35225809e-01 5.64818561e-01
1.32273555e+00 -5.27529120e-01 -6.55365646e-01 5.50110996e-01
7.99678683e-01 -5.09924293e-01 -6.16653860e-01 4.64589953e-01
2.23599300e-01 -7.69062161e-01 8.79657924e-01 -5.21607876e-01
6.48741782e-01 -4.06828523e-01 1.52212515e-01 -1.03215802e+00
-3.05044532e-01 -7.79214740e-01 -4.79324162e-01 1.57020748e+00
2.94125944e-01 -1.41309023e-01 4.93119419e-01 3.72470737e-01
-3.85954976e-01 -4.50380862e-01 -6.83845818e-01 -6.25486135e-01
-6.84687078e-01 -5.34143031e-01 4.13804829e-01 9.06488419e-01
5.84875271e-02 6.35528564e-01 -7.81810105e-01 3.07031959e-01
3.50150704e-01 5.57472222e-02 5.26192129e-01 -8.64707589e-01
-3.01220566e-01 -2.25773767e-01 -2.21673667e-01 -1.11748457e+00
9.24130678e-02 -6.08779848e-01 4.50628191e-01 -1.73962128e+00
4.65730280e-01 1.34655789e-01 -2.97567487e-01 6.49460971e-01
-6.70348585e-01 2.89596885e-01 1.50324628e-01 3.07416111e-01
-1.30300808e+00 7.24413633e-01 1.11194050e+00 1.79462895e-01
-2.51696497e-01 -1.90346971e-01 -4.33108866e-01 5.74519575e-01
6.00372553e-01 -4.39171076e-01 -5.81343651e-01 -5.74495733e-01
-1.00220572e-02 4.30264115e-01 5.55616081e-01 -1.14523757e+00
1.78263590e-01 -1.60298273e-01 2.45568573e-01 -6.23115718e-01
3.33356857e-01 -6.57946229e-01 3.55944782e-01 2.42304243e-02
-4.08575654e-01 -3.07830386e-02 2.88979083e-01 8.32374275e-01
-5.04144907e-01 2.56262552e-02 4.42325354e-01 -1.10543624e-01
-1.07934415e+00 6.21729910e-01 -5.36055565e-01 3.27791184e-01
1.28901958e+00 -8.09040070e-02 -2.98393935e-01 -6.20111763e-01
-8.57416391e-01 5.43227673e-01 1.69575334e-01 7.49257326e-01
4.63504970e-01 -1.37303436e+00 -9.66354430e-01 -1.92943916e-01
2.75848150e-01 1.01088628e-01 7.06648767e-01 9.18237805e-01
-1.00245081e-01 3.57639074e-01 -1.43353283e-01 -6.36461377e-01
-1.16962135e+00 6.85424030e-01 1.10607415e-01 -4.14607555e-01
-5.76839209e-01 6.56004965e-01 5.64482629e-01 3.41977030e-01
4.88475412e-01 -1.15020528e-01 -3.60260159e-01 1.70062393e-01
8.79230142e-01 7.55505562e-02 -2.17549577e-01 -7.19445109e-01
-3.13590884e-01 2.32421741e-01 -2.27728903e-01 -6.39107497e-03
1.22691417e+00 -4.18741792e-01 1.89154908e-01 2.50310659e-01
1.11797690e+00 -4.64643747e-01 -1.63575745e+00 -2.71879673e-01
-3.36936265e-02 -3.05250287e-01 -1.48004085e-01 -6.30706310e-01
-1.01745880e+00 6.31036460e-01 2.39377558e-01 2.19183922e-01
1.43443501e+00 2.33920798e-01 1.06944633e+00 1.64581314e-02
-7.72009743e-03 -8.31907392e-01 2.98740000e-01 5.37181497e-01
9.17448521e-01 -1.25618565e+00 -3.50638032e-01 -5.38700283e-01
-9.99602795e-01 8.05799007e-01 8.03024292e-01 1.43039972e-01
2.74205059e-01 7.48488382e-02 -1.60989031e-01 -6.66235611e-02
-9.61714864e-01 -3.83339286e-01 4.13754731e-01 4.08850938e-01
4.64580327e-01 -3.61580491e-01 -3.96590710e-01 8.02426040e-01
3.24940741e-01 4.05554593e-01 3.67233217e-01 9.48164582e-01
-1.35401800e-01 -9.51309383e-01 -1.68868914e-01 1.84252307e-01
-5.89714885e-01 -1.09156437e-01 -3.56902815e-02 5.51148891e-01
2.75567859e-01 1.03769445e+00 3.39201510e-01 -2.15099081e-01
4.16228980e-01 1.90058306e-01 2.56242871e-01 -5.95535636e-01
-4.35144573e-01 2.57502317e-01 3.36895496e-01 -7.40496695e-01
-6.54317319e-01 -6.91815317e-01 -1.23483372e+00 8.96095335e-02
-3.72631140e-02 2.50265002e-01 1.45172700e-01 8.99289787e-01
7.00749457e-01 8.16255689e-01 4.29002076e-01 -9.67946708e-01
-1.06093742e-01 -9.22573686e-01 -1.90484792e-01 8.27479541e-01
1.97726354e-01 -5.94588757e-01 -2.25282356e-01 5.79615235e-01] | [10.374902725219727, 0.6835125684738159] |
ec0f3ddd-d9e1-4f8c-aad9-a061fc2cfed9 | cntn-cyclic-noise-tolerant-network-for-gait | 2210.0691 | null | https://arxiv.org/abs/2210.06910v1 | https://arxiv.org/pdf/2210.06910v1.pdf | CNTN: Cyclic Noise-tolerant Network for Gait Recognition | Gait recognition aims to identify individuals by recognizing their walking patterns. However, an observation is made that most of the previous gait recognition methods degenerate significantly due to two memorization effects, namely appearance memorization and label noise memorization. To address the problem, for the first time noisy gait recognition is studied, and a cyclic noise-tolerant network (CNTN) is proposed with a cyclic training algorithm, which equips the two parallel networks with explicitly different abilities, namely one forgetting network and one memorizing network. The overall model will not memorize the pattern unless the two different networks both memorize it. Further, a more refined co-teaching constraint is imposed to help the model learn intrinsic patterns which are less influenced by memorization. Also, to address label noise memorization, an adaptive noise detection module is proposed to rule out the samples with high possibility to be noisy from updating the model. Experiments are conducted on the three most popular benchmarks and CNTN achieves state-of-the-art performances. We also reconstruct two noisy gait recognition datasets, and CNTN gains significant improvements (especially 6% improvements on CL setting). CNTN is also compatible with any off-the-shelf backbones and improves them consistently. | ['Liang Wang', 'Chunshui Cao', 'Yan Huang', 'Hongyuan Yu', 'Weichen Yu'] | 2022-10-13 | null | null | null | null | ['gait-recognition'] | ['computer-vision'] | [ 3.22321773e-01 -2.71700889e-01 -1.22404940e-01 -2.03183472e-01
2.15217341e-02 2.03654140e-01 1.39535233e-01 -1.42469302e-01
-5.04261732e-01 8.51729751e-01 -6.11222014e-02 3.33088487e-01
-2.81993505e-02 -1.06852710e+00 -4.87643540e-01 -1.08838487e+00
-8.80901664e-02 4.20812011e-01 3.77731442e-01 -2.75678456e-01
-5.04889749e-02 4.47135279e-03 -1.89341319e+00 3.15373801e-02
1.05265582e+00 9.62678909e-01 1.63477287e-01 2.86876261e-01
-5.96467592e-02 6.99617684e-01 -6.01894438e-01 -3.51612031e-01
-3.29694757e-03 -4.60729539e-01 -4.81294513e-01 2.46673360e-01
3.79483998e-01 1.21468298e-01 -4.42655116e-01 1.08206427e+00
8.07131886e-01 3.17643881e-01 4.21528310e-01 -1.13978982e+00
-7.60635138e-01 6.75422311e-01 -6.43297970e-01 3.11908007e-01
1.68694407e-01 6.08147979e-01 6.38067067e-01 -7.06753254e-01
4.09226894e-01 1.29538834e+00 7.88625479e-01 9.90764737e-01
-1.07647479e+00 -7.49963939e-01 4.85855907e-01 5.79013228e-01
-1.27451766e+00 -2.41772190e-01 9.41876471e-01 -1.24299072e-01
7.25836813e-01 1.86870247e-01 9.32507157e-01 1.59255099e+00
2.67015338e-01 8.24629784e-01 9.80585933e-01 -2.45073631e-01
2.27231488e-01 -3.11095387e-01 5.06097972e-01 8.22524488e-01
6.17302477e-01 -1.56686455e-01 -6.52187109e-01 7.33916387e-02
4.89513844e-01 3.82544808e-02 -3.02872628e-01 -2.26954818e-01
-8.16636026e-01 2.96019167e-01 2.01254219e-01 4.94690508e-01
-1.40193805e-01 1.71632007e-01 6.31202877e-01 3.21551085e-01
2.24072397e-01 -7.66809881e-02 -1.74080148e-01 -3.02020073e-01
-7.41232932e-01 -6.66230395e-02 4.59618628e-01 5.67717791e-01
7.22803831e-01 6.44090295e-01 -2.46841997e-01 1.10566223e+00
1.33372635e-01 6.48864567e-01 1.09355211e+00 -4.34833825e-01
5.30020237e-01 7.02215612e-01 -3.68455976e-01 -1.34860075e+00
-5.54843545e-01 -9.39003289e-01 -1.57682073e+00 -1.79133534e-01
1.58519164e-01 -6.25738204e-02 -1.18542576e+00 1.94359720e+00
-1.33141857e-02 6.85057878e-01 -2.03645557e-01 7.94596732e-01
7.83025503e-01 3.18937778e-01 2.22897813e-01 -1.81032389e-01
1.28728700e+00 -1.04978681e+00 -6.81950450e-01 -4.81520444e-01
4.87663925e-01 -4.56629544e-01 1.10260010e+00 6.37147248e-01
-8.12233508e-01 -1.12787485e+00 -1.37798703e+00 4.76515144e-01
-2.38779202e-01 2.72795945e-01 5.11045098e-01 8.56263936e-01
-8.22344959e-01 8.82025957e-01 -1.01086128e+00 -5.94440162e-01
4.15022075e-01 3.17649782e-01 -3.00252378e-01 -1.92331478e-01
-1.23873353e+00 5.92425883e-01 3.95245135e-01 5.13864636e-01
-5.40323496e-01 -2.45215163e-01 -9.34161246e-01 -5.62434718e-02
1.70648187e-01 -9.70292091e-01 4.86756682e-01 -9.89681900e-01
-1.47206461e+00 6.10957444e-01 -1.36796534e-02 -3.60824436e-01
6.85696244e-01 -1.40071198e-01 -8.53508115e-01 -1.18904121e-01
1.76729470e-01 4.07324851e-01 1.32835841e+00 -1.10892403e+00
-5.70761561e-01 -3.04160148e-01 -5.26522458e-01 -7.14570954e-02
-6.71515942e-01 -5.19008636e-01 -5.19039869e-01 -9.85934258e-01
3.64679426e-01 -7.71465838e-01 -1.38710484e-01 -2.62718767e-01
-4.71013069e-01 -7.19445944e-02 1.05643845e+00 -4.97275144e-01
1.38385046e+00 -2.01417613e+00 1.06783293e-01 4.73526716e-01
2.24083677e-01 5.73998332e-01 -3.53307724e-01 4.18020561e-02
-3.13512623e-01 -1.23704009e-01 -3.42549086e-01 -7.89317191e-01
2.02540997e-02 7.72046924e-01 -5.38390838e-02 3.37777078e-01
3.04747641e-01 5.85959315e-01 -8.40165973e-01 -4.40460026e-01
1.68813393e-01 3.07486862e-01 -4.66595232e-01 -1.16716906e-01
2.63517082e-01 6.21728897e-01 -3.17502290e-01 8.04882586e-01
1.07585800e+00 -1.89797461e-01 3.71080309e-01 -1.91864625e-01
3.24587554e-01 -2.60640264e-01 -1.57035136e+00 1.52636790e+00
-1.24348886e-01 -7.30907321e-02 -2.19103113e-01 -1.44135273e+00
1.15386105e+00 -1.94773991e-02 3.64408791e-01 -9.36766982e-01
1.53966725e-01 1.18932478e-01 3.62009518e-02 -6.41894996e-01
4.15440172e-01 2.54334062e-01 -2.10094806e-02 3.14263821e-01
3.36662233e-01 6.86717808e-01 4.69673723e-01 -1.39043838e-01
1.22254956e+00 1.42727569e-01 -1.45487100e-01 -3.15702647e-01
8.35089803e-01 -4.64483231e-01 1.22821808e+00 1.07908285e+00
-5.03713787e-01 6.20553076e-01 1.19988255e-01 -6.91865444e-01
-6.26759529e-01 -1.08016479e+00 7.63446316e-02 8.84441793e-01
3.91095251e-01 -5.13502657e-01 -5.92201531e-01 -6.53948188e-01
-1.67096809e-01 9.47457924e-02 -8.37688625e-01 -6.26284778e-01
-9.58353162e-01 -1.27230406e+00 8.02819729e-01 7.94779301e-01
1.33837938e+00 -1.13578522e+00 -5.73643446e-01 3.94168377e-01
-4.34424400e-01 -8.05683434e-01 -4.41942930e-01 2.28710651e-01
-8.53061140e-01 -1.03473675e+00 -7.26156116e-01 -9.89158452e-01
7.49453008e-01 5.43706939e-02 1.06174850e+00 6.37914598e-01
-3.74954164e-01 3.17694575e-01 -4.23858583e-01 1.68433145e-01
-9.75656807e-02 2.05842882e-01 5.27631879e-01 2.83735365e-01
3.10778379e-01 -1.05957413e+00 -6.27058864e-01 4.96320188e-01
-7.00831831e-01 -1.52999505e-01 5.95807910e-01 1.33679140e+00
5.22377670e-01 3.37029010e-01 5.41304469e-01 -5.68215609e-01
4.93910939e-01 -2.86501527e-01 1.53705791e-01 1.77093014e-01
-4.58837748e-01 1.52125284e-01 7.44243801e-01 -7.60333002e-01
-1.00896895e+00 -3.56125943e-02 -3.78683805e-01 -2.91851014e-01
-3.14075947e-01 2.97757000e-01 -3.41320723e-01 -1.11145295e-01
4.50614274e-01 6.36702001e-01 -1.20767921e-01 -6.30939960e-01
5.68638928e-03 1.21988133e-01 7.07950890e-01 -8.71828198e-01
9.65364277e-01 3.42188179e-01 2.30186470e-02 -9.51720297e-01
-4.44049776e-01 -6.63860887e-02 -6.15876555e-01 -3.10601979e-01
5.68256915e-01 -7.03339875e-01 -7.86758244e-01 1.09093869e+00
-7.32116640e-01 -1.41294181e-01 -4.58694220e-01 2.09159419e-01
-3.73277426e-01 8.36651623e-01 -8.44039619e-01 -5.06206274e-01
-4.14621085e-01 -9.95067537e-01 5.49282193e-01 4.63226438e-01
-1.80300653e-01 -7.79078066e-01 -2.43773796e-02 1.73769504e-01
5.83132863e-01 2.25168929e-01 9.07030106e-01 -3.32969666e-01
-2.07241654e-01 1.47815168e-01 -2.64418833e-02 4.88455474e-01
3.59009951e-01 -6.30938485e-02 -9.24667180e-01 -4.59447265e-01
-4.28560898e-02 -3.21160972e-01 1.42996132e+00 2.90063977e-01
1.06266868e+00 -4.15873118e-02 -4.37793821e-01 4.86978769e-01
1.19441307e+00 1.66120648e-01 9.20170844e-01 4.89066184e-01
7.42162287e-01 2.59046853e-01 2.51410693e-01 4.23658252e-01
2.49730662e-01 5.29649854e-01 2.49639854e-01 3.46409669e-03
-3.74461055e-01 -2.55039841e-01 3.39372814e-01 1.29148483e+00
-2.65122592e-01 -2.50868350e-01 -6.30289912e-01 3.19035321e-01
-2.18511009e+00 -9.54538167e-01 -6.77507967e-02 1.85337555e+00
6.14708424e-01 4.92864192e-01 1.13151379e-01 5.74366271e-01
8.98281932e-01 3.37660074e-01 -5.64259708e-01 2.42621209e-02
-5.36227703e-01 4.96226698e-01 8.78283903e-02 6.41478747e-02
-1.32680070e+00 7.55814493e-01 5.66179514e+00 1.23916483e+00
-1.06328142e+00 -5.14110271e-03 6.47208869e-01 3.09187353e-01
2.95273125e-01 -4.18337733e-01 -7.02990353e-01 8.33427250e-01
4.10254776e-01 2.99456030e-01 5.44575192e-02 7.88735569e-01
1.85787797e-01 -1.60256013e-01 -7.16645539e-01 1.25340891e+00
1.34679139e-01 -9.43482697e-01 4.52333033e-01 -3.41692418e-01
4.54459071e-01 -5.66449344e-01 2.40244433e-01 4.65055615e-01
-3.51324379e-02 -7.52769172e-01 5.04608333e-01 8.92446160e-01
4.50348616e-01 -8.27900529e-01 9.88780975e-01 2.29184896e-01
-1.75424886e+00 -2.18605652e-01 -6.29208505e-01 -2.93254912e-01
-5.72171174e-02 8.45833838e-01 1.52026936e-01 8.79967690e-01
9.28889394e-01 1.04416871e+00 -1.01783633e+00 1.15731335e+00
-2.15004191e-01 8.58817995e-01 -2.24201962e-01 6.75308779e-02
-1.39886275e-01 -2.16925889e-01 5.49123466e-01 1.10403705e+00
4.07963187e-01 -2.69500613e-01 4.30912763e-01 6.02833867e-01
-9.34168696e-03 1.81690212e-02 -2.07136780e-01 4.23274785e-01
2.21429467e-01 9.60800350e-01 -9.37381387e-01 -4.46778387e-01
4.56618704e-03 1.22856009e+00 2.39865556e-01 1.42837718e-01
-7.51865208e-01 -3.75100821e-01 6.05189502e-01 -1.81542635e-02
3.10640663e-01 -1.54494494e-01 -2.49294460e-01 -1.46501303e+00
2.96606094e-01 -8.05424511e-01 4.75954384e-01 -4.18193966e-01
-1.64274442e+00 4.77819204e-01 -4.34641123e-01 -1.48018754e+00
4.16024812e-02 -4.25171524e-01 -8.68790269e-01 5.46444774e-01
-1.30925548e+00 -1.08413649e+00 -5.70887864e-01 6.82399511e-01
5.48157752e-01 -1.73723817e-01 8.04752886e-01 8.80195796e-01
-1.10322154e+00 1.07735050e+00 -2.13416249e-01 3.07783186e-01
7.91493475e-01 -9.91243780e-01 2.83113807e-01 1.09061575e+00
-1.75767004e-01 6.66871667e-01 6.35619998e-01 -9.93231773e-01
-1.27460778e+00 -1.36579776e+00 6.38219893e-01 1.52799897e-02
5.02789676e-01 -2.64236569e-01 -1.14790285e+00 3.67202073e-01
-4.31498326e-02 1.28164276e-01 4.59088475e-01 1.03617106e-02
-2.02950239e-01 -2.09761798e-01 -1.12029684e+00 5.84473968e-01
1.59916711e+00 2.76363064e-02 -6.19048178e-01 -5.92041425e-02
3.13114613e-01 -2.18412980e-01 -4.92864698e-01 5.37200332e-01
5.19651711e-01 -8.74674022e-01 9.15400207e-01 -4.40773219e-01
1.25472173e-01 -5.18072426e-01 5.45236915e-02 -1.36261809e+00
-6.33034348e-01 -3.71876180e-01 -2.45260596e-01 1.49970913e+00
-2.12627314e-02 -7.98399568e-01 1.17679167e+00 5.39943539e-02
-8.53099152e-02 -6.18545949e-01 -1.12688518e+00 -1.19464004e+00
-2.49786511e-01 -3.91603053e-01 5.02715409e-01 9.89125848e-01
-1.52313009e-01 2.36485377e-01 -9.75130439e-01 7.58781610e-03
7.37566352e-01 -1.48563862e-01 6.06609523e-01 -1.38476741e+00
-4.01458621e-01 -3.41816276e-01 -8.11874270e-01 -1.18540728e+00
1.42987920e-02 -6.56574726e-01 1.05899209e-02 -1.05873430e+00
1.55544326e-01 -3.54479849e-01 -5.17981350e-01 7.43305445e-01
-3.53100479e-01 4.41135168e-01 7.16079250e-02 1.40557840e-01
-8.52228403e-01 9.54348147e-01 1.14668190e+00 -5.59188426e-01
-2.13007659e-01 6.21319264e-02 -4.96436596e-01 8.68461728e-01
8.76605809e-01 -4.07195121e-01 -4.79904622e-01 -2.57858753e-01
-2.70845354e-01 -4.39691037e-01 2.99993783e-01 -1.79842365e+00
3.58127058e-01 2.68999904e-01 7.65237689e-01 -6.32477939e-01
3.20303857e-01 -5.85345984e-01 1.86011717e-01 9.24852729e-01
1.25934198e-01 1.86785311e-01 1.65037259e-01 7.30524719e-01
-8.64702165e-02 -1.50465190e-01 9.29740012e-01 -6.94978833e-02
-1.02461004e+00 3.81901979e-01 -7.95440078e-01 1.03800334e-01
6.16471887e-01 -5.86239576e-01 -3.63288552e-01 -7.06322789e-02
-1.16408169e+00 3.18059891e-01 -1.18925143e-02 4.19742614e-01
8.58148873e-01 -1.54187882e+00 -4.62441564e-01 5.87458074e-01
5.08958399e-02 -3.86994213e-01 7.83196330e-01 6.60930157e-01
-1.84485555e-01 -1.16077438e-01 -4.94030237e-01 -6.96120858e-01
-1.14918530e+00 2.63163000e-01 5.06060362e-01 -5.96979141e-01
-7.60719180e-01 7.67556965e-01 -2.94696122e-01 -4.03640568e-01
3.78608704e-01 -3.49327028e-01 -3.23263347e-01 4.28322703e-03
5.23977935e-01 6.76696539e-01 1.51149645e-01 -6.72565043e-01
-3.94631952e-01 9.70935762e-01 -9.40464363e-02 4.50234503e-01
1.16161776e+00 -1.74662679e-01 3.52798100e-03 2.99964756e-01
6.64873540e-01 -4.74807262e-01 -1.10996854e+00 -2.62829453e-01
3.42836641e-02 -3.08372110e-01 -6.08566463e-01 -5.90138733e-01
-1.46882987e+00 7.92822719e-01 1.11775219e+00 -3.02312523e-02
1.34045517e+00 -7.31500447e-01 1.00778461e+00 6.20849371e-01
5.68037212e-01 -1.43247545e+00 5.54595172e-01 7.55935848e-01
4.70104039e-01 -9.82897639e-01 -2.44519070e-01 -4.84413356e-01
-1.71093330e-01 1.11950886e+00 1.10083532e+00 -3.91516060e-01
6.83742523e-01 2.94719845e-01 -2.33316764e-01 7.34738037e-02
-4.83955413e-01 -2.13857815e-01 7.96177518e-03 9.24270928e-01
-7.49387220e-02 -5.27701452e-02 -5.16537488e-01 1.03100443e+00
-1.20547134e-04 3.70370522e-02 3.15025061e-01 9.86158013e-01
-4.99487013e-01 -1.27947092e+00 -3.77344877e-01 5.10345399e-01
-6.24174550e-02 1.50745884e-01 1.27197295e-01 6.64846063e-01
7.59545326e-01 9.07157362e-01 -2.86501441e-02 -8.79812896e-01
4.71398056e-01 2.37962797e-01 3.03727567e-01 -2.75005519e-01
-4.81213719e-01 -1.83537602e-01 -9.53114331e-02 -5.14831066e-01
-5.26039422e-01 -4.67670590e-01 -1.01630342e+00 -4.97847021e-01
-1.24986961e-01 1.19602054e-01 -1.65041581e-01 9.58492041e-01
2.00875476e-01 8.88237476e-01 4.35146958e-01 -8.15183818e-01
-3.64792943e-01 -8.31875682e-01 -7.53810525e-01 5.04999101e-01
-5.38467839e-02 -9.45409060e-01 -2.85145998e-01 2.41896547e-02] | [14.305439949035645, 1.4132035970687866] |
cfe4e510-f1c3-438d-950c-9dd9db962446 | multi-modal-visual-tracking-review-and | 2012.04176 | null | https://arxiv.org/abs/2012.04176v1 | https://arxiv.org/pdf/2012.04176v1.pdf | Multi-modal Visual Tracking: Review and Experimental Comparison | Visual object tracking, as a fundamental task in computer vision, has drawn much attention in recent years. To extend trackers to a wider range of applications, researchers have introduced information from multiple modalities to handle specific scenes, which is a promising research prospect with emerging methods and benchmarks. To provide a thorough review of multi-modal track-ing, we summarize the multi-modal tracking algorithms, especially visible-depth (RGB-D) tracking and visible-thermal (RGB-T) tracking in a unified taxonomy from different aspects. Second, we provide a detailed description of the related benchmarks and challenges. Furthermore, we conduct extensive experiments to analyze the effectiveness of trackers on five datasets: PTB, VOT19-RGBD, GTOT, RGBT234, and VOT19-RGBT. Finally, we discuss various future directions from different perspectives, including model design and dataset construction for further research. | ['Huchuan Lu', 'Dong Wang', 'Pengyu Zhang'] | 2020-12-08 | null | null | null | null | ['rgb-t-tracking'] | ['computer-vision'] | [-1.55234262e-01 -7.79669046e-01 -3.97929758e-01 -2.33347621e-02
-5.91037393e-01 -8.75460863e-01 4.25802618e-01 -3.85667920e-01
-3.19440573e-01 2.98315257e-01 -1.14778928e-01 -1.40480384e-01
1.93693534e-01 -2.91605622e-01 -4.52870965e-01 -8.39368403e-01
1.51483014e-01 1.44110739e-01 8.46347034e-01 6.86321482e-02
-1.49893433e-01 6.06300235e-01 -1.73439574e+00 -3.54798585e-02
3.56232792e-01 1.18230224e+00 2.20690429e-01 5.24158001e-01
1.12553149e-01 3.90120298e-01 -4.19262379e-01 -7.08205998e-01
4.03721273e-01 -1.23617567e-01 -1.56421930e-01 9.17619690e-02
9.99146521e-01 -9.49962959e-02 -6.24527514e-01 1.24268568e+00
5.84206581e-01 2.21567620e-02 3.82318974e-01 -1.75931621e+00
-6.00663781e-01 3.08500025e-02 -8.80419493e-01 4.05992627e-01
3.86295855e-01 5.88359475e-01 5.56366205e-01 -8.10639262e-01
5.65039337e-01 1.34762132e+00 8.84210765e-01 7.14194357e-01
-7.49478877e-01 -8.35181177e-01 5.72198331e-01 2.63348520e-01
-1.11426222e+00 -3.02805956e-02 4.81355429e-01 -6.57192767e-01
5.61416805e-01 4.10398036e-01 1.05170619e+00 1.20163536e+00
2.12979451e-01 1.06214261e+00 1.11137378e+00 -1.59608424e-01
-2.16418609e-01 3.52990851e-02 4.34873030e-02 6.53242528e-01
6.85589314e-01 7.15696216e-01 -8.00853491e-01 -5.89667298e-02
6.91598177e-01 4.45878468e-02 -2.40229011e-01 -8.85596633e-01
-1.60397100e+00 5.11901796e-01 5.05457222e-01 1.47233568e-02
4.60483544e-02 5.07941604e-01 5.19753158e-01 -1.36812121e-01
4.07555133e-01 -2.00421810e-01 -4.61865783e-01 -2.27530420e-01
-5.93529880e-01 2.17185199e-01 1.33634299e-01 1.44066405e+00
3.86341721e-01 3.37476939e-01 -3.83864939e-01 3.55249852e-01
8.38139772e-01 1.24184072e+00 7.58459121e-02 -7.94507027e-01
3.45804274e-01 4.09854174e-01 2.89005697e-01 -7.01380730e-01
-4.24956530e-01 -3.52186531e-01 -4.02919441e-01 3.58541459e-01
4.59429562e-01 -1.70701265e-01 -8.50325346e-01 1.35890925e+00
8.44007373e-01 3.62356305e-01 -2.35860616e-01 1.28429937e+00
1.35319567e+00 2.49382168e-01 2.55227059e-01 -1.26987368e-01
1.72952402e+00 -1.17316246e+00 -7.96378076e-01 -2.98531830e-01
2.86605120e-01 -9.80804503e-01 5.57971954e-01 8.38469863e-02
-6.67321861e-01 -7.55053163e-01 -7.50167668e-01 4.38311137e-02
-4.52415556e-01 2.83339500e-01 7.84721434e-01 9.36689198e-01
-8.80184889e-01 -5.93713820e-02 -1.01313460e+00 -7.74681985e-01
3.19741666e-01 1.71232492e-01 -7.71663710e-02 -4.82501946e-02
-1.05708778e+00 9.62310195e-01 2.71264225e-01 1.54267341e-01
-1.12954128e+00 -6.28411949e-01 -7.28235424e-01 -5.62505722e-01
4.08603698e-01 -7.44926512e-01 1.22071612e+00 -5.94825745e-02
-1.27758467e+00 9.23453927e-01 -2.60861546e-01 -9.77997854e-02
6.73689783e-01 -2.61421233e-01 -5.92328429e-01 -1.33231223e-01
-6.11022338e-02 6.43122673e-01 7.84328520e-01 -1.37409556e+00
-1.08647621e+00 -4.80531365e-01 -6.12933710e-02 1.95836902e-01
-3.32798243e-01 5.56957722e-01 -9.97354448e-01 -6.18931413e-01
6.53403392e-03 -1.22053969e+00 -6.22829841e-03 5.35361469e-01
-4.44060773e-01 -2.98900396e-01 1.24535728e+00 -1.47104383e-01
9.64526832e-01 -2.07436657e+00 1.37520671e-01 -3.57030928e-01
3.05463761e-01 3.86219561e-01 1.00917168e-01 1.11681066e-01
3.38983387e-01 -3.95330369e-01 3.40734273e-01 -5.55053651e-01
9.76795703e-02 7.17964694e-02 -2.85793751e-01 8.05111051e-01
-4.93975222e-01 1.26566660e+00 -1.05404019e+00 -7.91984379e-01
7.02684402e-01 4.90769058e-01 1.02180742e-01 -1.96712883e-03
-3.12267721e-01 6.85209632e-01 -6.89555645e-01 1.34546280e+00
8.04026425e-01 -2.43568510e-01 -2.75724351e-01 -4.46085334e-01
-5.90780258e-01 -3.86915594e-01 -1.07920623e+00 1.62650001e+00
8.79625604e-02 9.21128988e-01 -4.52520438e-02 -2.62619972e-01
5.79331875e-01 2.39519671e-01 9.25770581e-01 -6.53022349e-01
3.04610789e-01 9.21777338e-02 -2.53381938e-01 -3.54169816e-01
8.04528534e-01 1.59401491e-01 -7.49697760e-02 1.56623319e-01
-1.73965350e-01 3.30807835e-01 2.15534508e-01 3.93969268e-02
7.52996266e-01 5.24513006e-01 -2.69101318e-02 2.40445405e-01
3.28292072e-01 4.39682811e-01 6.77441359e-01 7.68785894e-01
-9.04867828e-01 4.50992137e-01 -3.23164672e-01 -5.02277017e-01
-5.76834738e-01 -1.15673947e+00 -2.39448965e-01 1.09104848e+00
9.06797707e-01 -3.59974325e-01 -2.35433936e-01 -7.55490482e-01
4.51387018e-01 7.88174197e-02 -6.95984662e-01 2.06184201e-02
-5.62144041e-01 -7.16142774e-01 7.06824362e-01 7.67363131e-01
6.13168001e-01 -9.31075573e-01 -9.24765110e-01 -1.66191429e-01
-3.41123134e-01 -1.47383034e+00 -5.61113477e-01 1.58160031e-02
-1.04184127e+00 -1.07857323e+00 -9.26204145e-01 -5.24915099e-01
2.08959773e-01 1.03438461e+00 9.37091708e-01 -7.51163587e-02
-3.43078673e-01 9.77074265e-01 -3.30434859e-01 -6.32549584e-01
1.52434230e-01 -1.63326293e-01 3.75489950e-01 -2.55761415e-01
5.61593473e-01 3.07042032e-01 -5.36429644e-01 8.47250104e-01
-3.86656404e-01 -1.01627097e-01 3.28925341e-01 1.44431293e-01
7.93950200e-01 -2.87985563e-01 -2.47103676e-01 -8.59524161e-02
-1.40410349e-01 -1.21612005e-01 -1.20707297e+00 4.94875699e-01
-2.15588495e-01 -4.98030812e-01 1.66043174e-02 -6.78294241e-01
-1.09257102e+00 3.29359919e-01 2.56886363e-01 -9.79041159e-01
2.61898674e-02 -1.38364479e-01 -2.73286790e-01 -9.21367884e-01
4.43725407e-01 2.61730939e-01 -2.35272840e-01 -6.10679209e-01
5.96561909e-01 2.42425695e-01 5.32786608e-01 -4.58730459e-01
1.34654689e+00 9.28948045e-01 -9.01557356e-02 -6.14990234e-01
-9.48399484e-01 -9.01044667e-01 -6.53959751e-01 -9.51395810e-01
1.04853570e+00 -9.90954161e-01 -1.13397229e+00 7.47027636e-01
-1.05199850e+00 -2.20964223e-01 -4.20629866e-02 8.20364475e-01
-3.18018138e-01 3.28829139e-01 -4.53502685e-01 -8.38243067e-01
-1.48823768e-01 -1.31619060e+00 1.54920244e+00 7.92338252e-01
4.15759742e-01 -1.08083975e+00 1.99854165e-01 3.84980828e-01
2.55652249e-01 3.06534410e-01 -3.35206985e-02 -1.05817184e-01
-1.17129219e+00 -1.59685150e-01 -4.21851516e-01 -2.39072397e-01
6.07094616e-02 1.07738286e-01 -1.00591552e+00 -5.13520777e-01
-3.37504208e-01 -1.86157569e-01 8.31935823e-01 6.40739262e-01
8.18572998e-01 5.37799597e-01 -1.12239504e+00 8.90243709e-01
1.38036692e+00 2.83784240e-01 2.12378323e-01 7.36470163e-01
1.12826610e+00 5.16243838e-02 1.22525787e+00 2.14109436e-01
5.12070894e-01 1.21713305e+00 7.28117049e-01 -5.77744059e-02
-5.43741047e-01 1.51091237e-02 5.81912935e-01 5.20733654e-01
-2.56987631e-01 -2.27448225e-01 -6.78327441e-01 2.86295384e-01
-1.85805237e+00 -1.01705849e+00 -5.93597233e-01 2.06167912e+00
3.20869058e-01 -4.05589342e-02 5.57359457e-01 -3.63386542e-01
9.28715825e-01 2.09538385e-01 -8.34070563e-01 4.42019701e-01
-2.31987342e-01 -4.65169251e-01 9.09105062e-01 -3.80271412e-02
-1.32922077e+00 1.06946206e+00 7.08371258e+00 6.86275840e-01
-1.10659313e+00 3.87269199e-01 -2.22647861e-01 -1.30094603e-01
2.24054545e-01 -5.25008962e-02 -1.49215162e+00 6.49843752e-01
5.55697918e-01 9.96191055e-03 9.99327004e-02 8.92290711e-01
-1.08039834e-01 -2.93898731e-01 -8.98591936e-01 1.19388437e+00
1.44010969e-02 -1.03270829e+00 -3.35307181e-01 1.68569714e-01
6.63849652e-01 4.78974968e-01 3.51737380e-01 3.00026476e-01
3.06307584e-01 -3.59729737e-01 9.91170049e-01 3.71475041e-01
6.15876198e-01 -1.96327105e-01 4.29139942e-01 4.97751758e-02
-2.16591930e+00 1.66154757e-01 -5.02048969e-01 3.82126957e-01
4.95926738e-01 2.34829411e-01 -1.13729768e-01 9.79604542e-01
1.25828230e+00 1.16364217e+00 -8.56007516e-01 1.74582517e+00
-7.81447664e-02 2.09411874e-01 -3.26739937e-01 -9.40814838e-02
2.82700304e-02 2.78160591e-02 6.50933385e-01 1.06750762e+00
2.45570719e-01 -1.36437327e-01 4.77969408e-01 4.38903630e-01
2.76296288e-01 -5.53061306e-01 -3.81970346e-01 2.29776606e-01
5.65600276e-01 1.47206652e+00 -8.85130346e-01 -2.75091112e-01
-7.87648618e-01 5.49479246e-01 -1.44711331e-01 3.60129505e-01
-1.52776492e+00 1.33162513e-01 9.82219160e-01 -2.58545667e-01
4.41892862e-01 -3.86884928e-01 -2.12710304e-03 -1.33447313e+00
-7.55127668e-02 -5.87763071e-01 5.55091739e-01 -1.06224716e+00
-1.16621697e+00 3.43439847e-01 3.03190917e-01 -1.96035135e+00
3.61893147e-01 -8.75872910e-01 -1.65252000e-01 6.26829028e-01
-1.68795931e+00 -1.42207551e+00 -8.31106842e-01 6.72807395e-01
3.72790396e-01 -6.01325445e-02 2.13118210e-01 7.22546697e-01
-6.06672704e-01 4.59908217e-01 1.57054067e-01 2.96186507e-01
8.81674945e-01 -9.70893085e-01 5.74616671e-01 9.33838904e-01
1.82899848e-01 4.99872327e-01 6.49907887e-01 -7.66019464e-01
-2.01155138e+00 -1.32445765e+00 1.98304862e-01 -1.00160658e+00
6.60154939e-01 -2.74746060e-01 -5.39752483e-01 9.61742282e-01
2.47296095e-01 5.69592118e-01 3.75523984e-01 -2.11187631e-01
-2.86166698e-01 7.76464213e-03 -9.25273359e-01 3.66287827e-01
1.39324343e+00 -2.37229988e-01 -2.30468854e-01 3.73431474e-01
6.68445647e-01 -1.08446944e+00 -9.94415939e-01 4.18814480e-01
7.89039195e-01 -8.64437580e-01 1.34821534e+00 -1.64269999e-01
-6.69676185e-01 -1.01115787e+00 -7.43650496e-02 -9.67299759e-01
-3.01358193e-01 -4.38032120e-01 -6.15352273e-01 1.12551463e+00
-3.10160249e-01 -5.92471838e-01 8.97585511e-01 2.12550491e-01
-2.62871176e-01 -4.36119318e-01 -9.73684847e-01 -1.24013174e+00
-3.46123964e-01 -5.54268777e-01 3.98939103e-01 6.58700705e-01
-6.30664587e-01 -2.68464893e-01 -6.88268006e-01 3.01974505e-01
1.35165286e+00 3.95969898e-01 1.04891801e+00 -1.28182423e+00
1.63988203e-01 -4.13156182e-01 -5.27958274e-01 -1.47679305e+00
-2.75363684e-01 -7.11136281e-01 7.34411702e-02 -1.61059868e+00
3.66205603e-01 -4.56906259e-01 -1.79723904e-01 3.62603992e-01
-2.26652056e-01 6.58964038e-01 4.96505439e-01 3.54211062e-01
-1.20114172e+00 6.03750467e-01 1.61315954e+00 -2.40305975e-01
2.34407932e-01 1.64313927e-01 -3.82438034e-01 5.71245968e-01
2.27236241e-01 -5.65896630e-01 1.17029855e-02 -5.89693427e-01
-1.02777682e-01 -4.75461558e-02 5.16085327e-01 -1.12368190e+00
5.80745697e-01 -2.94151634e-01 5.69239020e-01 -1.35138917e+00
6.66930377e-01 -1.18321490e+00 3.43527019e-01 6.72729552e-01
2.77817875e-01 3.81441355e-01 4.64908093e-01 7.36299813e-01
-3.65485139e-02 3.84440482e-01 8.56107950e-01 8.76004472e-02
-1.31600821e+00 8.42618823e-01 -1.60821244e-01 3.19941975e-02
1.46079528e+00 -6.68661535e-01 -6.97868645e-01 1.43382162e-01
-4.91203547e-01 6.87404633e-01 6.23081923e-01 9.66626585e-01
5.37547886e-01 -1.82525837e+00 -3.21614742e-01 -1.42244190e-01
5.36059558e-01 -3.56709838e-01 3.29103500e-01 1.15599287e+00
-1.55507594e-01 7.92817593e-01 -2.02849001e-01 -1.30957079e+00
-1.59829950e+00 7.47079313e-01 4.86281753e-01 1.79401293e-01
-7.19703496e-01 7.18521774e-01 2.73679376e-01 -2.43728012e-01
6.00632370e-01 -3.73093486e-01 -2.20317103e-04 -8.35590437e-02
5.09870112e-01 5.76471448e-01 -7.41633922e-02 -1.03506756e+00
-9.12180960e-01 1.17005086e+00 2.97908843e-01 1.84241012e-01
7.37453818e-01 -3.44946295e-01 3.09254885e-01 3.59246045e-01
7.46064901e-01 -1.89827636e-01 -1.48672247e+00 -3.37195605e-01
-8.06902768e-04 -7.72355139e-01 -1.32013127e-01 -5.87082922e-01
-1.55283284e+00 7.20585823e-01 1.09723914e+00 1.89940572e-01
8.80213916e-01 2.78730482e-01 7.33202279e-01 -2.53856322e-03
8.67702425e-01 -7.50143588e-01 3.18724662e-01 5.51316917e-01
2.96352386e-01 -1.42365074e+00 2.34153956e-01 -4.16895509e-01
-5.02664924e-01 9.58089948e-01 1.13238394e+00 3.02652746e-01
5.50478160e-01 3.04872096e-01 3.55119139e-01 -2.92847991e-01
-3.83327782e-01 -7.04482019e-01 4.78188157e-01 9.34728920e-01
3.44406515e-01 -1.01859301e-01 3.10377806e-01 -3.19715999e-02
2.01880470e-01 -1.71120003e-01 -1.20505570e-02 1.06525755e+00
-4.35663909e-01 -1.09628415e+00 -9.52897489e-01 1.29867822e-01
-2.47901920e-02 3.80119741e-01 -3.77595544e-01 1.13649964e+00
1.72589928e-01 9.70824540e-01 -1.74785301e-01 -4.40621763e-01
5.01134813e-01 -4.19427723e-01 8.67463529e-01 -3.70422810e-01
-6.64913833e-01 2.02274218e-01 -3.01664382e-01 -7.33382761e-01
-1.04898608e+00 -9.96118426e-01 -9.32806492e-01 -4.92101431e-01
-7.05666244e-01 -1.87024355e-01 7.90463984e-01 8.33242536e-01
1.74842447e-01 7.35889137e-01 9.76496041e-02 -1.10651505e+00
-9.47279856e-02 -6.56514168e-01 -4.87387717e-01 6.18207790e-02
2.94200450e-01 -1.42070878e+00 -1.39944032e-01 -5.11562862e-02] | [6.448585033416748, -2.112894296646118] |
ba9a8208-a68d-4e47-adca-18ed0c32093d | optimal-copula-transport-for-clustering | 1509.08144 | null | http://arxiv.org/abs/1509.08144v2 | http://arxiv.org/pdf/1509.08144v2.pdf | Optimal Copula Transport for Clustering Multivariate Time Series | This paper presents a new methodology for clustering multivariate time series
leveraging optimal transport between copulas. Copulas are used to encode both
(i) intra-dependence of a multivariate time series, and (ii) inter-dependence
between two time series. Then, optimal copula transport allows us to define two
distances between multivariate time series: (i) one for measuring
intra-dependence dissimilarity, (ii) another one for measuring inter-dependence
dissimilarity based on a new multivariate dependence coefficient which is
robust to noise, deterministic, and which can target specified dependencies. | ['Philippe Donnat', 'Gautier Marti', 'Frank Nielsen'] | 2015-09-27 | null | null | null | null | ['clustering-multivariate-time-series'] | ['time-series'] | [-1.02384582e-01 -7.28778005e-01 1.18463649e-03 -2.39217162e-01
-4.31563586e-01 -1.09163141e+00 5.33063054e-01 6.19480908e-01
-2.83318788e-01 6.07208550e-01 -2.33523846e-01 -4.34188038e-01
-7.38869190e-01 -8.04711699e-01 -4.29582685e-01 -8.00169170e-01
-8.16887319e-01 4.22576725e-01 3.28831375e-01 -2.24269480e-02
8.84199515e-02 7.93213785e-01 -1.43905830e+00 -1.62145227e-01
9.81346250e-01 8.66645873e-01 1.22012667e-01 7.87680507e-01
-5.83952591e-02 1.80222213e-01 -8.85181189e-01 -8.74114931e-02
-6.07714173e-05 -3.73376429e-01 -1.26995489e-01 -1.24611050e-01
-8.29515040e-01 6.58302367e-01 3.14054281e-01 1.08888972e+00
-1.08645111e-01 1.91320270e-01 1.26720762e+00 -1.71881998e+00
-6.51591659e-01 5.71420610e-01 -5.99924147e-01 4.10711169e-01
2.91656494e-01 -3.05697590e-01 7.22082496e-01 -7.99279511e-02
2.61793822e-01 1.03101277e+00 6.17343187e-01 -1.23618267e-01
-1.35104179e+00 -2.16105402e-01 -4.11809906e-02 -1.79254577e-01
-1.31640816e+00 3.66429478e-01 9.36388254e-01 -5.60331047e-01
6.94509864e-01 4.65954661e-01 7.31175780e-01 8.27217817e-01
9.97484744e-01 2.48296380e-01 1.35141075e+00 -3.82340074e-01
4.71667886e-01 -1.95511244e-02 3.20224315e-01 -2.23305970e-01
1.26773313e-01 9.02694911e-02 2.90935487e-01 -3.49579185e-01
5.00114322e-01 2.82167673e-01 -5.84530272e-02 -1.06466310e-02
-1.02744281e+00 7.41458952e-01 -1.79831475e-01 7.65491903e-01
-2.09650055e-01 -1.81948945e-01 2.50943691e-01 5.88863552e-01
2.58735329e-01 2.78500646e-01 -5.80575228e-01 -2.90835142e-01
-7.07506120e-01 6.28289357e-02 8.87734473e-01 9.55092013e-01
5.77874303e-01 -2.77273238e-01 -6.89917505e-02 5.35770833e-01
3.67339730e-01 1.06798589e+00 3.90172362e-01 -3.30342561e-01
5.68132937e-01 3.12829673e-01 1.47095099e-01 -1.09968936e+00
-6.11800015e-01 -2.42426440e-01 -7.80927718e-01 -3.88933942e-02
4.35799479e-01 -3.53707105e-01 -5.48820317e-01 1.97536719e+00
3.75192277e-02 2.80159414e-01 6.60401583e-02 3.19832772e-01
1.45409137e-01 8.56061399e-01 -1.17375731e-01 -8.06193829e-01
1.07215452e+00 -4.44709718e-01 -9.20797944e-01 4.10463244e-01
2.22863212e-01 -9.85710621e-01 5.98174632e-01 2.19444141e-01
-1.04353964e+00 -4.27628219e-01 -1.00115955e+00 4.42735791e-01
-7.93437064e-01 -4.96242464e-01 5.03255129e-01 9.05331790e-01
-1.00990868e+00 6.38764739e-01 -8.34446728e-01 -1.19596139e-01
-4.60156977e-01 4.10576016e-01 -3.04359466e-01 4.72842664e-01
-1.27905202e+00 5.32409847e-01 2.57170022e-01 2.95422357e-02
-3.38816822e-01 -5.19262373e-01 -6.74521565e-01 1.33748680e-01
-1.56700432e-01 -2.43014872e-01 6.11340821e-01 -6.68888450e-01
-1.37111509e+00 3.98328751e-01 -7.59528652e-02 -2.18794391e-01
5.30153573e-01 3.39327723e-01 -1.16484690e+00 1.68956995e-01
2.06946909e-01 -1.71569675e-01 1.12922049e+00 -1.24070597e+00
-4.72675353e-01 -4.71217901e-01 -2.96298742e-01 -2.34948173e-01
-7.02583715e-02 2.18062475e-01 -3.03574950e-01 -7.74242461e-01
1.55537620e-01 -9.19119656e-01 -7.17736408e-02 -5.60579121e-01
-5.08467913e-01 -2.15894490e-01 8.79303575e-01 -5.90206802e-01
1.38973117e+00 -2.22486019e+00 1.23371184e-01 8.02770317e-01
-3.09567422e-01 -7.46082217e-02 -1.21360376e-01 9.08798277e-01
-5.40339828e-01 2.60521710e-01 -6.27467275e-01 -4.18386579e-01
7.65377209e-02 5.50372601e-01 -4.00862694e-01 6.48051858e-01
5.85991263e-01 2.77815729e-01 -1.02787900e+00 -3.86776119e-01
4.31521088e-02 4.55127507e-01 -5.91138713e-02 1.99438781e-01
1.23114012e-01 5.78343630e-01 -2.85411417e-01 5.69419920e-01
1.14931142e+00 2.46390134e-01 5.78155927e-02 -4.45726141e-02
-4.43460017e-01 -3.45168084e-01 -1.27489769e+00 8.90469551e-01
-1.41525820e-01 4.12896007e-01 -1.02152377e-01 -8.84018421e-01
1.36400044e+00 3.40915114e-01 7.72240818e-01 -2.42812693e-01
2.92569727e-01 1.15940817e-01 5.29344529e-02 -3.61775994e-01
3.78258824e-01 -1.37066007e-01 -3.23501438e-01 6.33302271e-01
-8.10832605e-02 -3.23429286e-01 6.47619188e-01 -2.66032457e-01
9.10229862e-01 -1.33829609e-01 -4.18003909e-02 -5.85607588e-01
6.76983178e-01 -6.02058887e-01 5.48814476e-01 4.46869969e-01
-1.18707493e-01 6.51227176e-01 1.04507935e+00 8.28995332e-02
-9.51039016e-01 -1.65691769e+00 -3.75798047e-01 5.94724119e-01
1.80853233e-01 5.09698540e-02 -3.83628309e-01 -2.97872961e-01
2.69813359e-01 7.41330624e-01 -9.20853257e-01 -4.97360760e-03
-3.58776122e-01 -6.40005589e-01 4.49940771e-01 6.72482789e-01
-4.64839004e-02 -5.87061644e-01 -3.28766942e-01 9.11596045e-02
-5.73164672e-02 -7.63677299e-01 -8.98377836e-01 7.36525834e-01
-1.03688896e+00 -1.02885783e+00 -6.41455472e-01 -5.24827480e-01
5.50971568e-01 2.03766659e-01 9.35470879e-01 -9.55443457e-02
3.37335467e-02 7.01288879e-01 -5.12271285e-01 -5.16586721e-01
-3.47833335e-01 -4.89567310e-01 1.96928754e-02 2.22534671e-01
1.94464117e-01 -8.36780727e-01 -2.65286088e-01 4.90533531e-01
-1.36389971e+00 -8.96078348e-01 1.51905641e-01 5.93880296e-01
6.45221949e-01 4.39961225e-01 5.02112210e-01 -1.19614944e-01
1.01251125e+00 -9.07062232e-01 -6.77876949e-01 4.97738212e-01
-3.83936018e-01 -1.11488728e-02 9.05208349e-01 -8.69000554e-01
-7.28209913e-01 -4.71525252e-01 2.91630507e-01 -4.52366173e-01
-1.01412937e-01 8.46509218e-01 -8.98995250e-02 3.92951101e-01
2.42802069e-01 1.76367924e-01 -7.90526941e-02 -2.59732723e-01
3.58215749e-01 5.05082369e-01 5.26707828e-01 -8.39995146e-01
6.30504966e-01 6.55579984e-01 3.11952919e-01 -9.55678642e-01
1.41312340e-02 -5.88137269e-01 -1.04068315e+00 -9.54137444e-02
1.05647218e+00 -3.05992454e-01 -6.11711085e-01 3.93010646e-01
-1.14469564e+00 9.73242894e-02 -2.68926531e-01 7.84395456e-01
-5.70845962e-01 4.85630482e-01 -5.42737901e-01 -1.01961243e+00
1.36319296e-02 -8.97526860e-01 7.50381589e-01 2.19094619e-01
-1.13251284e-01 -1.77177441e+00 4.89003420e-01 -4.10142332e-01
-7.51519576e-02 7.15364873e-01 1.06702793e+00 -9.07290697e-01
8.94346461e-02 -6.98137939e-01 -7.43401349e-02 4.89135474e-01
3.12358797e-01 8.99350584e-01 -3.41477603e-01 -3.63113970e-01
2.41744652e-01 5.27787149e-01 1.98320374e-01 5.20428300e-01
8.32462192e-01 5.57155274e-02 -4.50757802e-01 4.58762318e-01
1.30329752e+00 7.92361975e-01 6.60582542e-01 1.43700272e-01
1.42762840e-01 5.16602576e-01 6.28615916e-01 5.13231397e-01
4.47933733e-01 2.85625428e-01 3.18984926e-01 3.83400694e-02
6.78461313e-01 -6.97536841e-02 2.45686829e-01 1.20650458e+00
-2.63836175e-01 -2.87824392e-01 -7.86082625e-01 7.72281885e-01
-1.86701846e+00 -1.09558344e+00 -5.49769461e-01 2.47237706e+00
3.78786534e-01 1.65061474e-01 3.78423631e-01 1.75013185e-01
1.02642536e+00 7.70895854e-02 -3.66675884e-01 -9.71635103e-01
-4.89188135e-01 2.38756537e-01 4.32561696e-01 2.58619934e-01
-1.06652105e+00 6.29810393e-02 6.94287825e+00 5.34825802e-01
-1.18596268e+00 -1.85391247e-01 1.34403110e-01 6.23054981e-01
-6.45477116e-01 1.13395959e-01 -3.76826137e-01 1.00349879e+00
9.71609592e-01 -5.55465341e-01 7.54579678e-02 4.99115318e-01
2.73446321e-01 -1.00036249e-01 -1.17492819e+00 6.77117407e-01
-1.22580692e-01 -5.69727719e-01 -2.56924897e-01 1.14295170e-01
8.03602695e-01 -3.61737221e-01 1.85830519e-01 4.16330323e-02
1.19188830e-01 -9.79372442e-01 5.98671675e-01 8.36407304e-01
1.60129413e-01 -1.06650305e+00 9.98921156e-01 2.98791319e-01
-1.62242353e+00 3.69439162e-02 -2.79173374e-01 4.17652400e-03
3.70774001e-01 8.41384768e-01 -1.01894498e-01 1.00596189e+00
5.81020117e-01 8.28434110e-01 -3.44121873e-01 1.27866650e+00
-2.47099072e-01 2.55581081e-01 -4.37954128e-01 -1.96920663e-01
1.32578060e-01 -9.80025053e-01 7.43126392e-01 1.30349827e+00
7.75316238e-01 -3.76700938e-01 9.47639197e-02 9.13747609e-01
8.16692591e-01 -3.07198856e-02 -5.39146900e-01 -2.10778192e-01
7.60009646e-01 9.67336476e-01 -1.03577638e+00 -7.75760561e-02
-2.58974105e-01 6.82880819e-01 -3.57376814e-01 5.21714687e-01
-1.23343921e+00 -8.37723076e-01 6.53271437e-01 -4.00605679e-01
5.35613477e-01 -6.84292674e-01 -4.62778002e-01 -9.31566060e-01
1.59338325e-01 -2.53064722e-01 6.81039631e-01 -5.68652153e-01
-1.80815864e+00 6.72173500e-01 3.49031538e-01 -1.60266137e+00
-4.33554530e-01 -5.08643448e-01 -1.01182473e+00 1.14040911e+00
-1.15904140e+00 -1.14143491e+00 2.08938986e-01 1.03376532e+00
-2.46198580e-01 8.84483978e-02 7.74156868e-01 1.91244379e-01
-3.30094725e-01 4.53404784e-01 3.56487364e-01 -1.69898853e-01
6.20106816e-01 -1.40115082e+00 -1.61312506e-01 7.38773704e-01
-3.34903747e-01 8.10371816e-01 8.93594265e-01 -5.84213912e-01
-1.28730881e+00 -8.12603712e-01 9.03162181e-01 -3.91004115e-01
1.18569481e+00 -1.44907087e-01 -8.81192505e-01 6.58746541e-01
1.03799365e-02 -4.29648548e-01 9.91739571e-01 9.12665278e-02
-4.71989542e-01 -3.44382763e-01 -1.43378174e+00 6.00657046e-01
2.26841971e-01 -4.63655055e-01 -8.10819566e-01 1.78413883e-01
8.39632213e-01 4.05247137e-02 -1.51961899e+00 3.61597031e-01
6.20835721e-01 -1.18066633e+00 8.02143872e-01 -2.93974012e-01
2.18821257e-01 -5.70555568e-01 -3.46446007e-01 -1.37048531e+00
-2.18307197e-01 -8.16015303e-01 2.56562084e-01 1.31450117e+00
5.06255507e-01 -1.16535842e+00 -4.22129072e-02 4.76098835e-01
-4.90077287e-02 -2.79470891e-01 -1.01996553e+00 -1.64037800e+00
4.42329884e-01 -7.86381841e-01 9.71640646e-01 1.02548063e+00
1.94350928e-01 -1.87841594e-01 -1.40924789e-02 4.54739660e-01
6.30854368e-01 2.78139800e-01 4.52264756e-01 -1.17500293e+00
-3.65581185e-01 -7.85841286e-01 -5.34254909e-01 -6.45981312e-01
5.34807071e-02 -3.85220438e-01 -3.16439685e-03 -1.13059890e+00
-1.07129775e-01 -2.95442492e-01 -4.52450097e-01 -4.65464145e-02
1.44013464e-01 -2.09521592e-01 1.48073182e-01 2.20574781e-01
-2.04334900e-01 4.38449055e-01 1.05164063e+00 2.09705830e-01
-5.47943115e-01 4.92739916e-01 -9.71681401e-02 5.71555495e-01
7.07933366e-01 -5.27983069e-01 -5.28978944e-01 -4.98267170e-03
-1.40858833e-02 5.12078643e-01 8.86251852e-02 -9.22657371e-01
2.53139406e-01 -3.69152248e-01 5.54439090e-02 -9.66155469e-01
2.84694403e-01 -1.16003430e+00 3.72763723e-01 3.61392140e-01
2.00054839e-01 7.65830874e-01 2.67920494e-01 7.45888054e-01
-5.84527910e-01 -3.49314988e-01 5.28340101e-01 4.53425378e-01
-1.62238494e-01 9.07422975e-02 -3.97568971e-01 -1.24071240e-01
1.28622437e+00 -2.47416079e-01 -3.83247972e-01 -5.04482031e-01
-9.02158737e-01 1.62285700e-01 3.08714330e-01 4.56758738e-01
3.59590322e-01 -1.47148836e+00 -3.88402402e-01 1.98992133e-01
-8.63194987e-02 -5.59613705e-01 1.84792608e-01 1.15319908e+00
-3.07030946e-01 2.23005295e-01 -2.38523707e-01 -9.26130474e-01
-1.21838915e+00 9.04642642e-01 -1.21184714e-01 -5.70649087e-01
-1.10730059e-01 5.02552867e-01 -1.41553864e-01 -3.62855822e-01
-5.55078089e-02 -7.83353865e-01 -2.36178845e-01 3.46313417e-01
4.22747463e-01 5.44109166e-01 -9.94672179e-02 -6.40616357e-01
-4.61880565e-01 9.44861770e-01 4.32409108e-01 -2.23740026e-01
1.18856871e+00 -2.78881609e-01 -7.71649182e-01 1.21436548e+00
1.33045542e+00 6.03136048e-02 -1.14111161e+00 4.65432078e-01
3.48491907e-01 -3.90584350e-01 -5.70193410e-01 -5.12051523e-01
-4.82552648e-01 7.04389274e-01 2.78446227e-01 1.27156138e+00
1.53651893e+00 -3.02257627e-01 5.05013883e-01 -2.61238366e-01
4.42398757e-01 -1.09023035e+00 1.30440444e-01 7.66552866e-01
8.79207790e-01 -6.23867273e-01 -2.01322734e-01 -3.53267372e-01
-6.13288462e-01 1.52471113e+00 -1.79310188e-01 -3.70656997e-01
1.33775616e+00 4.20221746e-01 -1.55556694e-01 1.97597325e-01
-6.91886961e-01 -3.84002626e-01 5.17946720e-01 1.16955996e+00
3.47506493e-01 4.56842244e-01 -7.17230856e-01 7.35125065e-01
-2.26684794e-01 -3.69037539e-01 4.35184956e-01 6.86235487e-01
-3.41466330e-02 -1.04929841e+00 -6.99241817e-01 1.04868347e-02
-3.26300830e-01 2.35184789e-01 6.66947737e-02 1.21714175e+00
1.44394562e-01 1.44105637e+00 5.55329859e-01 -3.27499598e-01
4.47981358e-01 -2.13517770e-01 3.17003638e-01 -1.06340677e-01
-4.61845577e-01 4.81305867e-01 -2.12961525e-01 -1.03247434e-01
-4.19938415e-01 -7.86918819e-01 -1.07056034e+00 -2.05837354e-01
-2.67981142e-01 3.93836766e-01 9.56249118e-01 9.67677832e-01
1.01068072e-01 3.34157735e-01 1.15529084e+00 -5.91702759e-01
-5.41517019e-01 -7.41083324e-01 -1.28542626e+00 4.63000000e-01
1.86827585e-01 -5.84728003e-01 -7.61160910e-01 4.21923259e-03] | [7.187014102935791, 3.4126603603363037] |
9f0c6942-7f8c-4e95-bf34-6c63cfdd3758 | data-augmentation-for-diverse-voice | 2305.10684 | null | https://arxiv.org/abs/2305.10684v1 | https://arxiv.org/pdf/2305.10684v1.pdf | Data Augmentation for Diverse Voice Conversion in Noisy Environments | Voice conversion (VC) models have demonstrated impressive few-shot conversion quality on the clean, native speech populations they're trained on. However, when source or target speech accents, background noise conditions, or microphone characteristics differ from training, quality voice conversion is not guaranteed. These problems are often left unexamined in VC research, giving rise to frustration in users trying to use pretrained VC models on their own data. We are interested in accent-preserving voice conversion for name pronunciation from self-recorded examples, a domain in which all three of the aforementioned conditions are present, and posit that demonstrating higher performance in this domain correlates with creating VC models that are more usable by otherwise frustrated users. We demonstrate that existing SOTA encoder-decoder VC models can be made robust to these variations and endowed with natural denoising capabilities using more diverse data and simple data augmentation techniques in pretraining. | ['William Yang Wang', 'Amr El Abbadi', 'Michael Saxon', 'Avani Tanna'] | 2023-05-18 | null | null | null | null | ['voice-conversion', 'voice-conversion'] | ['audio', 'speech'] | [ 2.19772175e-01 -8.39975774e-02 2.34294668e-01 -2.49240786e-01
-1.01401877e+00 -6.92010283e-01 4.63076681e-01 -3.87448072e-01
-4.67102975e-01 7.49023259e-01 6.39670491e-01 -5.42303026e-01
2.49214709e-01 -2.76538163e-01 -4.58755255e-01 -4.35756505e-01
2.36015230e-01 1.02650084e-01 -1.97363257e-01 -4.81975585e-01
-2.36863598e-01 2.96044677e-01 -1.56373918e+00 9.93724912e-02
9.12655175e-01 5.41202247e-01 3.84251237e-01 1.09626448e+00
-1.19194783e-01 5.57033002e-01 -9.98324454e-01 -5.41126132e-01
1.00380443e-01 -8.12913895e-01 -5.48003137e-01 -7.17290118e-02
5.93384624e-01 -7.18004853e-02 -1.50556758e-01 9.79908764e-01
9.04817283e-01 3.25425267e-01 5.20072401e-01 -6.55797720e-01
-8.87939453e-01 6.11437678e-01 2.21538380e-01 3.90266001e-01
4.85353470e-01 3.48099291e-01 9.66479361e-01 -9.01115477e-01
6.48103714e-01 1.03554118e+00 7.39735007e-01 9.93510246e-01
-1.54135931e+00 -5.56435227e-01 -6.50706440e-02 -5.05311005e-02
-1.13593853e+00 -1.27334774e+00 6.93878651e-01 -1.12989992e-01
1.02922750e+00 5.70045590e-01 5.49018383e-01 1.68909144e+00
-1.65438652e-01 4.00327593e-01 9.84959006e-01 -4.80310827e-01
2.50520289e-01 6.12415910e-01 -3.80784959e-01 8.42518434e-02
-2.11480021e-01 2.94949174e-01 -6.79305017e-01 1.04698017e-01
5.11428893e-01 -5.42532027e-01 -7.58401155e-01 -3.32355276e-02
-1.04839349e+00 5.71734309e-01 8.39014873e-02 4.90947336e-01
-2.61195958e-01 -8.91942605e-02 4.24379408e-01 7.80181706e-01
3.14887822e-01 9.10200357e-01 -5.14948845e-01 -7.78495252e-01
-1.07761967e+00 -4.71245497e-02 9.86727715e-01 1.13803244e+00
1.16574019e-01 8.24323297e-01 2.49229074e-02 1.35975552e+00
-1.81666344e-01 4.00828719e-01 8.50837886e-01 -1.04208922e+00
2.65016913e-01 -2.59586483e-01 -1.10199496e-01 -4.82553154e-01
4.51037548e-02 -8.54553878e-01 -5.74232519e-01 2.91694343e-01
2.28738800e-01 -3.40198606e-01 -8.73383999e-01 1.96051443e+00
-1.61350310e-01 -2.97684204e-02 4.72455829e-01 6.99752033e-01
6.45420909e-01 5.30866742e-01 1.05591781e-01 -4.74190980e-01
9.65902209e-01 -6.99317575e-01 -1.10449469e+00 -4.30846423e-01
2.71425545e-01 -1.04681265e+00 1.81824410e+00 5.40108860e-01
-1.14490426e+00 -7.81798005e-01 -1.19658375e+00 4.79968451e-02
-2.90867805e-01 -2.54111022e-01 4.77847487e-01 1.27947319e+00
-1.16474855e+00 7.60632694e-01 -6.88663363e-01 -3.03767502e-01
2.56628513e-01 1.75825521e-01 -4.92932439e-01 4.96745855e-02
-1.04286289e+00 1.09827065e+00 -2.45860860e-01 -4.58202362e-02
-7.84827709e-01 -8.48775923e-01 -7.39464164e-01 1.03590377e-01
3.43585238e-02 -4.08863097e-01 1.53907287e+00 -1.06564677e+00
-1.64513028e+00 4.53012109e-01 -1.23022057e-01 -1.87597349e-01
5.89949250e-01 -4.47214752e-01 -1.13046861e+00 -1.91396371e-01
-4.65473533e-02 4.11572903e-01 1.16989732e+00 -1.29483545e+00
-3.93726736e-01 -3.10351886e-02 -3.74301344e-01 3.23445201e-01
-6.53551221e-01 1.54732078e-01 -7.34567866e-02 -6.28194332e-01
-2.56857336e-01 -5.16240835e-01 5.20922877e-02 -3.25706929e-01
-4.00371104e-01 2.15090603e-01 8.42599273e-01 -9.98483956e-01
1.14515233e+00 -2.34815955e+00 -1.24865107e-01 -1.20217912e-01
-1.77195519e-01 6.63857043e-01 -1.88033313e-01 4.24469829e-01
-8.37279558e-02 4.57675695e-01 -1.38950258e-01 -4.56022620e-01
7.86822140e-02 1.04643255e-01 -2.23001495e-01 1.69803843e-01
3.63341749e-01 3.67831379e-01 -7.16794848e-01 -3.51970762e-01
6.87684640e-02 7.31616557e-01 -6.59908831e-01 5.28512061e-01
8.10977146e-02 5.62510967e-01 2.89012641e-01 5.44394314e-01
2.81565726e-01 4.81965363e-01 1.80937685e-02 1.24014869e-01
-1.40959144e-01 6.67810678e-01 -1.16588616e+00 1.53899407e+00
-7.65037715e-01 9.48638082e-01 4.99099910e-01 -4.81102914e-01
8.12596083e-01 8.25916588e-01 -3.02932616e-02 -7.19477236e-01
-1.71107352e-02 4.90703821e-01 4.05251682e-01 -6.08101487e-01
5.29006898e-01 -6.17800355e-01 2.97071397e-01 8.12957883e-02
2.46155560e-01 -6.13707244e-01 9.31409560e-03 -8.39033276e-02
1.11401331e+00 -2.87951957e-02 -6.34701923e-02 -1.14655092e-01
1.00620896e-01 -3.83397751e-02 5.25814235e-01 7.10395515e-01
-3.43690604e-01 9.76249874e-01 5.08679710e-02 4.31366056e-01
-1.04707325e+00 -1.42109489e+00 -2.39982486e-01 1.02205336e+00
-6.16112471e-01 -3.30502421e-01 -8.02742541e-01 -1.08476035e-01
-4.33404922e-01 1.11777782e+00 1.82155520e-02 -2.45886773e-01
-5.50940752e-01 -2.32221112e-01 7.98908532e-01 4.55494583e-01
1.25852570e-01 -1.05757105e+00 -1.98868528e-01 6.25614464e-01
-2.37082034e-01 -1.02409184e+00 -6.70746982e-01 5.38094938e-01
-6.92246974e-01 -5.98958194e-01 -8.34457815e-01 -9.81997609e-01
8.15505162e-02 7.81283677e-02 1.17298412e+00 -1.08563073e-01
-1.25297144e-01 6.51115119e-01 -3.72754157e-01 -5.77639878e-01
-1.11251700e+00 1.48453936e-01 4.42465991e-01 -3.19589645e-01
2.44791001e-01 -9.68520045e-01 -2.56933033e-01 -8.29308853e-03
-7.58674741e-01 -3.15174669e-01 3.45288754e-01 9.29386020e-01
1.97746649e-01 -4.67015766e-02 1.04677355e+00 -7.72130609e-01
1.19483125e+00 -1.42090037e-01 -1.14539243e-01 -3.50617990e-02
-5.83200097e-01 -1.59584448e-01 8.34114373e-01 -8.09482098e-01
-1.17977774e+00 -1.92495048e-01 -6.70793951e-01 -3.63953650e-01
-4.82507169e-01 3.63557339e-01 -4.76557761e-01 1.75341949e-01
1.02271593e+00 1.72210991e-01 6.12728968e-02 -6.62760794e-01
3.89867634e-01 1.15207946e+00 1.01577389e+00 -3.74993414e-01
8.15376103e-01 -2.81580448e-01 -8.11256289e-01 -1.32835531e+00
-2.23220035e-01 -1.36719525e-01 -4.35435742e-01 -7.96260387e-02
5.48004329e-01 -8.95530164e-01 -1.32857889e-01 2.36420810e-01
-1.00811303e+00 -3.03684205e-01 -5.28635561e-01 6.22517765e-01
-5.10213196e-01 1.66817948e-01 -4.41216886e-01 -9.99710560e-01
-1.34559900e-01 -1.14934266e+00 3.82974565e-01 1.02920175e-01
-6.46787405e-01 -9.40746009e-01 9.97640267e-02 2.15754405e-01
1.00252521e+00 -1.71800911e-01 9.93042588e-01 -7.18577206e-01
-5.79598732e-02 -1.77705362e-01 5.15196025e-01 1.02120626e+00
6.50122881e-01 3.54404710e-02 -1.47545290e+00 -2.36309320e-01
2.05190957e-01 -2.44880885e-01 4.92617995e-01 1.37106866e-01
6.82592273e-01 -3.56212586e-01 2.66144544e-01 5.97311676e-01
1.07328141e+00 2.88208097e-01 5.17083168e-01 -4.00206409e-02
4.01606321e-01 5.63948154e-01 7.95913339e-02 -1.37751877e-01
-1.51581779e-01 6.53239608e-01 1.08141646e-01 -1.99745059e-01
-5.74166954e-01 -3.58282626e-01 6.79672241e-01 1.27129841e+00
-5.13932705e-02 -2.89038330e-01 -5.76368332e-01 8.38935375e-01
-9.72545862e-01 -1.05532742e+00 9.41139646e-03 2.17651415e+00
1.33357084e+00 2.37701893e-01 8.66266340e-02 3.70035529e-01
6.87897682e-01 2.73510069e-01 -4.48444337e-01 -8.35866213e-01
-2.92035639e-01 6.65484548e-01 6.47873506e-02 5.43252409e-01
-8.54097605e-01 9.68424618e-01 6.82980776e+00 5.58747828e-01
-1.30365121e+00 1.01068608e-01 3.66833657e-01 -2.87085056e-01
-3.88337672e-01 -2.20437914e-01 -2.97360897e-01 2.28606135e-01
1.38397312e+00 -4.34632190e-02 9.14533079e-01 7.49313414e-01
3.59122187e-01 4.00524169e-01 -1.29086125e+00 8.94143343e-01
2.00805301e-03 -1.05676889e+00 -3.56952697e-01 -2.82618254e-02
4.15252358e-01 2.02051122e-02 4.33913320e-01 4.85210061e-01
3.13639760e-01 -1.38899720e+00 7.52959371e-01 8.57921019e-02
1.21304369e+00 -6.05241179e-01 4.02461231e-01 1.84803069e-01
-7.18828857e-01 4.47303168e-02 -3.16027790e-01 -2.34283577e-03
1.57897174e-01 2.69984633e-01 -1.15019751e+00 9.17106196e-02
4.89339769e-01 1.23760670e-01 -3.92494649e-01 1.00551510e+00
-8.21752101e-02 1.14991796e+00 -1.79038897e-01 7.50760585e-02
-9.45971534e-02 -1.18490672e-02 6.65201068e-01 1.50642371e+00
4.74240154e-01 -8.85802582e-02 -5.62551260e-01 8.09570134e-01
-2.27522880e-01 2.56572127e-01 -7.64218509e-01 -4.62117761e-01
6.01899803e-01 9.54807997e-01 -1.48942277e-01 6.97338209e-02
-5.21127641e-01 1.22589934e+00 1.38795421e-01 5.60540080e-01
-4.20570612e-01 -4.83339399e-01 1.28884053e+00 2.34271716e-02
2.98486352e-01 -3.37638497e-01 -3.35953176e-01 -1.11384428e+00
-6.84970617e-02 -1.34932709e+00 -1.25558138e-01 -6.69783115e-01
-1.23640013e+00 9.76933241e-01 -5.70254087e-01 -1.14676249e+00
-5.30818820e-01 -5.67124605e-01 -8.06836963e-01 1.02302849e+00
-1.33324671e+00 -7.55560219e-01 6.16041720e-02 6.36912346e-01
7.61921763e-01 -3.07539999e-01 1.29957211e+00 4.27913725e-01
-3.34012628e-01 8.15653741e-01 2.01958314e-01 6.48322254e-02
8.39657545e-01 -1.39375484e+00 4.99601215e-01 1.06501317e+00
5.90880871e-01 8.31106365e-01 1.07801938e+00 -3.52758616e-01
-1.18215239e+00 -8.64625633e-01 8.93274963e-01 -4.38148171e-01
3.24884474e-01 -6.02226317e-01 -1.03127849e+00 5.27822971e-01
4.02147472e-01 -3.11994493e-01 8.37876976e-01 3.39826822e-01
-2.97734588e-01 -1.34807020e-01 -9.50523674e-01 7.94691741e-01
1.15341079e+00 -1.01410496e+00 -8.40796292e-01 8.20933841e-03
8.85539472e-01 -2.77688205e-01 -9.20481384e-01 -8.68206844e-02
3.52987915e-01 -8.66645336e-01 8.99031818e-01 -6.95825398e-01
2.18723148e-01 4.21791039e-02 -3.36401016e-01 -1.91995656e+00
-1.48319125e-01 -8.52712929e-01 3.28685045e-01 1.87928176e+00
8.51551473e-01 -3.40552062e-01 4.21446830e-01 7.04015672e-01
-5.67213297e-01 3.55672964e-04 -1.07552636e+00 -9.45026577e-01
3.50841284e-01 -7.60303915e-01 5.71137905e-01 1.04853630e+00
1.87628549e-02 4.96191770e-01 -4.58054811e-01 -1.41086774e-02
1.52943537e-01 -4.38620865e-01 6.67174995e-01 -1.12491333e+00
-3.60326797e-01 -4.02829230e-01 -5.83940111e-02 -6.46425128e-01
-1.19517505e-01 -6.80641413e-01 3.32126021e-01 -1.37441850e+00
-5.53873360e-01 -2.89981872e-01 -8.94304067e-02 1.89628541e-01
-1.45784527e-01 1.32169694e-01 2.98617840e-01 -1.16529781e-02
1.51071236e-01 6.40929759e-01 1.10233593e+00 4.02471423e-02
-4.53510463e-01 3.08663130e-01 -1.09356940e+00 4.36690271e-01
7.79961646e-01 -3.91863436e-01 -6.13935351e-01 -3.98604423e-01
-2.85051852e-01 1.44155994e-01 1.20314136e-01 -1.42296898e+00
1.58166006e-01 5.70892915e-02 4.07929838e-01 1.31222025e-01
7.39360094e-01 -8.40292990e-01 3.03601444e-01 1.37732819e-01
-4.90569949e-01 -8.44461173e-02 3.49667937e-01 2.92099625e-01
-3.41059297e-01 -2.52668023e-01 9.03267086e-01 -3.16615254e-01
-4.52729285e-01 -1.23269573e-01 -6.44634902e-01 3.18547100e-01
1.78598866e-01 -4.11024958e-01 -1.00507922e-01 -8.68038297e-01
-9.30918515e-01 -4.07147735e-01 4.71944422e-01 6.60341620e-01
5.39422870e-01 -1.20322835e+00 -8.51141453e-01 4.75685954e-01
-9.57499668e-02 -3.09370309e-01 5.69310114e-02 5.86672664e-01
-1.88528880e-01 1.09359719e-01 -1.35177523e-01 -1.93884134e-01
-1.17382812e+00 2.76362926e-01 5.56258678e-01 4.17907536e-01
-6.89191401e-01 9.89364266e-01 -3.49840462e-01 -5.80422819e-01
4.77239460e-01 -1.73866823e-01 -3.90207916e-02 2.26243362e-02
3.62982154e-01 7.64318034e-02 4.21618223e-01 -4.75479960e-01
-1.59218714e-01 -7.22149834e-02 -2.11523473e-02 -5.94786704e-01
1.33589828e+00 -1.29811153e-01 5.65780699e-01 5.32254457e-01
1.12779093e+00 5.70302546e-01 -1.21059763e+00 -1.23973094e-01
-1.12332851e-01 -4.23423231e-01 1.17232442e-01 -1.01880753e+00
-8.17939878e-01 1.04172885e+00 7.12570131e-01 3.05193007e-01
1.09133089e+00 -1.94130197e-01 8.09577584e-01 3.21387291e-01
3.58680673e-02 -1.27461267e+00 -1.44849226e-01 4.56401348e-01
9.26667154e-01 -9.10036027e-01 -5.40308893e-01 -1.38242722e-01
-8.86801302e-01 9.17047322e-01 5.43240130e-01 1.85099557e-01
3.87707889e-01 4.15457040e-01 6.10204935e-01 2.73356080e-01
-7.65984774e-01 -2.84183770e-01 -1.02966562e-01 1.29414439e+00
6.79473460e-01 -1.07715400e-02 4.14651632e-02 6.67819500e-01
-1.02982593e+00 -4.42492962e-01 7.26319671e-01 6.67761862e-01
-3.91730666e-01 -1.12504733e+00 -4.83638585e-01 3.54885161e-01
-6.55046880e-01 -5.04141152e-01 -4.84669954e-01 8.25229824e-01
-6.26710951e-02 1.41091573e+00 5.27921617e-02 -4.48431462e-01
6.03427410e-01 6.57145083e-01 1.93277374e-01 -8.03675234e-01
-7.85089374e-01 2.19862163e-01 5.44529319e-01 -1.34699330e-01
2.53879353e-02 -8.70871484e-01 -9.79687572e-01 -2.04982981e-01
-4.28370833e-01 2.28372321e-01 9.03866112e-01 7.69568264e-01
1.69505402e-01 7.88107097e-01 4.70630676e-01 -5.32372296e-01
-8.53822470e-01 -1.25927055e+00 -6.63637459e-01 3.62936258e-01
7.13676393e-01 -1.86635360e-01 -5.53486288e-01 3.28937858e-01] | [14.893218040466309, 6.430174827575684] |
8ecb12b5-61b0-4316-abb5-bfc92dd7823f | taming-contrast-maximization-for-learning | 2303.05214 | null | https://arxiv.org/abs/2303.05214v1 | https://arxiv.org/pdf/2303.05214v1.pdf | Taming Contrast Maximization for Learning Sequential, Low-latency, Event-based Optical Flow | Event cameras have recently gained significant traction since they open up new avenues for low-latency and low-power solutions to complex computer vision problems. To unlock these solutions, it is necessary to develop algorithms that can leverage the unique nature of event data. However, the current state-of-the-art is still highly influenced by the frame-based literature, and usually fails to deliver on these promises. In this work, we take this into consideration and propose a novel self-supervised learning pipeline for the sequential estimation of event-based optical flow that allows for the scaling of the models to high inference frequencies. At its core, we have a continuously-running stateful neural model that is trained using a novel formulation of contrast maximization that makes it robust to nonlinearities and varying statistics in the input events. Results across multiple datasets confirm the effectiveness of our method, which establishes a new state of the art in terms of accuracy for approaches trained or optimized without ground truth. | ['Guido C. H. E. de Croon', 'Christophe De Wagter', 'Kirk Y. W. Scheper', 'Federico Paredes-Vallés'] | 2023-03-09 | null | null | null | null | ['event-based-optical-flow'] | ['computer-vision'] | [ 4.17766809e-01 -3.78374845e-01 -1.52371913e-01 -3.32694888e-01
-6.15348577e-01 -4.91606534e-01 7.55397379e-01 -1.31136496e-02
-5.04697025e-01 5.17634273e-01 2.15930864e-01 -2.08536893e-01
-1.36682943e-01 -5.19274056e-01 -6.92748845e-01 -4.51928794e-01
-7.63391703e-02 1.50554866e-01 6.36644065e-01 2.06497595e-01
2.76457995e-01 5.26342809e-01 -1.87262833e+00 8.28418955e-02
5.44515789e-01 9.93155658e-01 1.42311111e-01 9.38182592e-01
2.07846940e-01 1.16922307e+00 -2.22391948e-01 -1.04994141e-01
3.36102247e-01 -3.86819750e-01 -5.94699800e-01 -3.26737538e-02
8.96081269e-01 -7.74821758e-01 -4.80094612e-01 5.76378167e-01
4.68446463e-01 1.60524368e-01 1.55404016e-01 -1.22088432e+00
-1.86992750e-01 -1.35317430e-01 -3.49340469e-01 7.54475176e-01
2.30065525e-01 5.46353698e-01 1.02539206e+00 -5.77642381e-01
8.51022542e-01 9.26326513e-01 7.34276116e-01 4.65024143e-01
-1.29545069e+00 -2.99082100e-01 1.17581896e-01 6.68669224e-01
-9.14994240e-01 -7.52146900e-01 5.94863176e-01 -4.70846564e-01
1.25408947e+00 -1.04950354e-01 7.89339781e-01 1.20537186e+00
2.73506463e-01 8.00489008e-01 1.08365726e+00 -4.46301371e-01
3.06369752e-01 -1.66675046e-01 -9.94660985e-03 7.22220838e-01
3.10786277e-01 3.25460672e-01 -1.05278075e+00 1.69413716e-01
8.70214641e-01 -9.50821936e-02 -2.06289977e-01 -4.63711292e-01
-1.34781301e+00 7.45454609e-01 1.27716392e-01 8.81874785e-02
-3.62748057e-01 4.51757431e-01 3.81483883e-01 -1.96185410e-01
4.44430947e-01 3.98488283e-01 -4.09143239e-01 -5.58884263e-01
-1.29886079e+00 3.19972903e-01 8.65327001e-01 5.46830535e-01
6.69924200e-01 -1.24473393e-01 -1.10514618e-01 2.03623787e-01
3.95749688e-01 4.43757892e-01 1.23960689e-01 -1.36377823e+00
5.65398522e-02 4.46052939e-01 1.04343213e-01 -9.19051409e-01
-5.47419965e-01 -3.66949886e-01 -3.08535516e-01 4.00361389e-01
6.07631922e-01 2.85291057e-02 -7.27187395e-01 1.72441757e+00
5.77146232e-01 8.57116282e-01 -6.54160976e-03 9.31981564e-01
4.96870309e-01 6.55225039e-01 -6.58222884e-02 -2.62636662e-01
1.05446565e+00 -1.02624524e+00 -5.77668488e-01 -3.47847462e-01
1.99643821e-01 -6.98720753e-01 9.69804943e-01 3.83444130e-01
-1.12751019e+00 -3.86391997e-01 -1.14413512e+00 -2.98331589e-01
-1.14883386e-01 -2.36065239e-01 1.03996491e+00 4.93048996e-01
-1.07310915e+00 6.48776412e-01 -1.41620159e+00 -6.50108278e-01
5.81981838e-01 2.67283320e-01 -1.81188568e-01 -1.77455917e-01
-6.47354722e-01 9.84646857e-01 2.04390094e-01 9.98252854e-02
-8.16185415e-01 -8.80863905e-01 -7.75813520e-01 -7.47763216e-02
3.87953669e-01 -8.50380242e-01 1.38070476e+00 -9.19326782e-01
-1.69451344e+00 7.07079887e-01 -3.29145819e-01 -6.19052410e-01
5.89307845e-01 -2.13807985e-01 -1.31707549e-01 6.11812651e-01
9.39784423e-02 7.68231750e-01 8.31390440e-01 -8.08728874e-01
-9.02125239e-01 -1.17179267e-01 7.29264915e-02 -4.60094139e-02
-3.41868967e-01 1.23989902e-01 -4.53284353e-01 -3.48354101e-01
-2.35060394e-01 -9.42081690e-01 -1.41531065e-01 3.40132177e-01
1.31801618e-02 9.30705369e-02 7.05408394e-01 -3.39929312e-01
9.42539454e-01 -2.14345694e+00 -7.40164667e-02 -1.83732048e-01
1.64648011e-01 3.16714585e-01 1.75938666e-01 3.07260424e-01
3.37724447e-01 -3.55260044e-01 -7.58938864e-02 -5.74554801e-01
-2.74547283e-02 3.52308303e-01 -3.66075248e-01 6.48271799e-01
5.57928920e-01 8.80905449e-01 -1.18670452e+00 -5.20606697e-01
7.68217266e-01 7.65392601e-01 -7.26739526e-01 3.20210665e-01
-2.54299015e-01 5.66128194e-01 -2.96752423e-01 4.47831750e-01
3.29345405e-01 -6.22741818e-01 1.02883466e-01 -3.10093313e-01
-5.01704574e-01 3.84302378e-01 -1.39140940e+00 1.87584591e+00
-6.49283752e-02 1.00002325e+00 -1.04574673e-01 -8.30860138e-01
3.14736307e-01 1.33715764e-01 8.91597033e-01 -6.62079334e-01
2.56082863e-01 2.26713613e-01 -1.39218971e-01 -7.24486828e-01
5.83915114e-01 2.80283000e-02 5.05547822e-01 4.00424957e-01
2.29430899e-01 7.11761937e-02 3.95159036e-01 1.00596450e-01
1.40018201e+00 6.63245916e-01 2.74371207e-01 -3.05040702e-02
2.70259649e-01 1.89363971e-01 5.17325401e-01 8.48934174e-01
-4.10685331e-01 7.44500995e-01 1.48915216e-01 -5.91695189e-01
-1.02507102e+00 -1.16878271e+00 -3.03711057e-01 7.30864525e-01
1.29878283e-01 -3.08826566e-01 -6.07120216e-01 -3.61447245e-01
-1.38588756e-01 3.13604385e-01 -3.44093829e-01 2.28368528e-02
-5.70174396e-01 -8.71942341e-01 3.57800812e-01 5.78057408e-01
4.92930591e-01 -8.74212325e-01 -1.51663756e+00 5.01924396e-01
-1.52372688e-01 -1.83885789e+00 -1.49833038e-01 -2.76851561e-02
-9.33622956e-01 -1.21527433e+00 -2.91796476e-01 -2.23784894e-01
3.26677740e-01 3.45747918e-01 1.15358818e+00 -1.53458536e-01
-6.15351021e-01 7.40198433e-01 -3.23400408e-01 -4.90206510e-01
-1.26602560e-01 1.19635738e-01 -7.08870366e-02 2.29769945e-01
3.84792030e-01 -7.01141417e-01 -8.10843945e-01 1.63820721e-02
-1.01581335e+00 1.67846635e-01 4.58380640e-01 6.50448501e-01
5.43762863e-01 -3.12854618e-01 2.86928773e-01 -4.30385739e-01
2.90678721e-02 -3.07118058e-01 -8.67754281e-01 -7.79257447e-04
-7.91114807e-01 1.63070470e-01 4.42872256e-01 -4.01670367e-01
-1.05896807e+00 3.07627976e-01 1.41150460e-01 -4.53573972e-01
-1.12335145e-01 2.33950719e-01 4.21092838e-01 -2.95912653e-01
5.61012447e-01 -4.81899902e-02 1.61288857e-01 -1.12737656e-01
4.73279536e-01 3.31371456e-01 9.89914238e-01 -4.39486593e-01
4.86074597e-01 1.21760392e+00 3.87784690e-01 -8.31362486e-01
-1.03168297e+00 -6.29771113e-01 -5.33376455e-01 -4.81869131e-01
1.02279294e+00 -9.47989583e-01 -8.49540412e-01 6.23011351e-01
-1.15775836e+00 -4.10014480e-01 -5.13127089e-01 7.01430321e-01
-7.16137171e-01 3.97461206e-01 -5.03674984e-01 -8.81969869e-01
-1.71636268e-01 -1.18455637e+00 1.07455504e+00 5.09397209e-01
-4.19225022e-02 -1.04621649e+00 3.49301755e-01 3.21535200e-01
5.63395381e-01 4.59433645e-01 1.52145311e-01 -2.34818593e-01
-1.30245483e+00 8.36582109e-03 -4.35450882e-01 3.77585739e-01
-1.71430528e-01 1.97772533e-01 -1.32383287e+00 -2.46934384e-01
-6.80800751e-02 -2.65021980e-01 8.70690882e-01 7.23153889e-01
7.36301005e-01 1.61396608e-01 -1.80047095e-01 1.03966963e+00
1.56364322e+00 -2.21527219e-01 5.83386302e-01 5.69908023e-01
5.85309684e-01 3.97929430e-01 2.31495097e-01 6.31581843e-01
7.53614485e-01 5.97739160e-01 6.33973122e-01 -6.63950890e-02
-3.54091913e-01 -9.96943861e-02 3.22004110e-01 3.91740382e-01
-1.79595396e-01 -1.74380511e-01 -7.75494635e-01 6.30205989e-01
-2.01372623e+00 -1.23730338e+00 -2.30998844e-01 2.22161174e+00
6.05851412e-01 3.13320041e-01 -1.11832854e-03 6.30384684e-02
2.61799127e-01 4.00978565e-01 -6.73195481e-01 -5.36398888e-02
-4.66570184e-02 3.96758735e-01 5.87787688e-01 4.39185977e-01
-1.17812955e+00 8.35509956e-01 6.95208502e+00 1.83522120e-01
-1.30983818e+00 1.02822430e-01 2.70413876e-01 -4.46610898e-01
1.64726824e-01 2.31015921e-01 -9.43684399e-01 3.63774449e-01
1.18012965e+00 3.99983600e-02 6.64947689e-01 5.17338753e-01
5.32815099e-01 -5.93840361e-01 -1.22015011e+00 1.04514182e+00
3.12014490e-01 -1.65168858e+00 -3.40564728e-01 2.36439891e-02
7.68161237e-01 5.77539921e-01 -1.22571029e-01 -2.11354181e-01
1.58728495e-01 -7.08733022e-01 6.70091212e-01 6.75828815e-01
5.53867638e-01 -2.92569280e-01 4.28161711e-01 1.21897832e-01
-1.03114069e+00 1.23390304e-02 -3.22124921e-02 -4.88776684e-01
5.01042366e-01 6.85449958e-01 -7.48970807e-01 4.52775747e-01
7.70472944e-01 1.15350032e+00 -5.02282262e-01 1.11810040e+00
-8.18172377e-03 6.99421167e-01 -6.61673307e-01 1.65221587e-01
1.61159486e-01 9.63856801e-02 5.04583955e-01 1.17967570e+00
2.64464393e-02 -1.66656792e-01 3.14775497e-01 7.30134308e-01
1.42395765e-01 -2.27701500e-01 -5.00137687e-01 3.58999297e-02
3.08490664e-01 1.36875880e+00 -8.81855845e-01 -2.03014836e-01
-7.45796144e-01 8.27600837e-01 3.47181469e-01 2.12298647e-01
-8.31131458e-01 9.24531668e-02 6.70634687e-01 1.92452408e-02
4.29544091e-01 -6.77488208e-01 -2.71217734e-01 -1.46725309e+00
2.18598232e-01 -6.42100573e-01 3.97195160e-01 -7.53152788e-01
-1.12537301e+00 2.34663248e-01 9.61251929e-03 -1.08138061e+00
-4.98117477e-01 -6.44042790e-01 -3.53909433e-01 3.63496631e-01
-2.12293124e+00 -1.24464881e+00 -5.84722698e-01 5.88807583e-01
3.99191499e-01 1.22871786e-01 6.15623891e-01 4.17008877e-01
-5.70406854e-01 1.62917659e-01 3.63205113e-02 -9.21368301e-02
7.99690485e-01 -1.01532304e+00 2.87808895e-01 1.38217819e+00
3.68061572e-01 3.46948236e-01 8.42904985e-01 -2.89075315e-01
-1.73753011e+00 -8.28136742e-01 8.38257670e-01 -8.20208967e-01
8.85499597e-01 -1.49157807e-01 -5.78253210e-01 8.19517791e-01
7.94450380e-03 5.51280797e-01 3.84409845e-01 -1.14712588e-01
-3.25250208e-01 -3.44067991e-01 -8.00156415e-01 3.92475903e-01
9.84496117e-01 -5.58971763e-01 -4.06555176e-01 2.22748056e-01
3.95890445e-01 -6.39853239e-01 -5.72234452e-01 3.01539123e-01
6.34955525e-01 -1.33514059e+00 1.05322933e+00 -4.02155340e-01
4.90877658e-01 -3.96323889e-01 -1.67960465e-01 -7.00500786e-01
-5.01607284e-02 -8.84525716e-01 -5.97437978e-01 1.02824318e+00
-5.35026193e-02 -5.56226075e-01 7.84401357e-01 6.72626674e-01
-1.49919584e-01 -6.62581444e-01 -1.15723586e+00 -6.96907520e-01
-5.73605835e-01 -7.55149007e-01 1.72455385e-02 6.63998485e-01
-3.54063392e-01 2.91985333e-01 -3.77709329e-01 1.65882811e-01
9.24547017e-01 -4.10005674e-02 8.99616957e-01 -1.20260096e+00
-5.11541665e-01 -3.59052420e-01 -7.76697278e-01 -1.04977512e+00
7.99393207e-02 -5.49458385e-01 3.56370248e-02 -1.49734700e+00
2.23975614e-01 -1.79619968e-01 -1.73887998e-01 3.59155864e-01
-1.93737611e-01 4.33078855e-01 2.57434458e-01 2.86146879e-01
-8.55665922e-01 2.83433318e-01 7.74267256e-01 3.26288968e-01
-9.71738696e-02 -3.00022900e-01 -4.17460769e-01 7.41281629e-01
4.00980800e-01 -3.11895072e-01 -3.94044459e-01 -6.85707033e-01
2.60381430e-01 -2.25796402e-01 9.03120697e-01 -1.29451740e+00
5.53326130e-01 -1.41570047e-01 1.98335916e-01 -3.32636237e-01
3.40421200e-01 -8.38851631e-01 -4.56701964e-02 2.47601569e-01
-1.23285681e-01 1.60620332e-01 2.25043669e-01 7.19935775e-01
3.53236310e-02 8.67229179e-02 8.69915009e-01 4.99948300e-02
-1.01704264e+00 3.43063086e-01 -1.30839765e-01 3.96074891e-01
9.56168413e-01 -2.72308171e-01 -3.42393279e-01 -2.01673687e-01
-2.25176230e-01 1.28300218e-02 4.18294877e-01 3.74237537e-01
2.74126589e-01 -8.80258143e-01 -6.57550573e-01 2.09779948e-01
1.20361093e-02 5.36785908e-02 1.70322359e-01 1.01221299e+00
-5.35525084e-01 2.81330019e-01 -2.38526478e-01 -1.05741668e+00
-9.58548725e-01 3.94386262e-01 3.75601083e-01 -2.36057520e-01
-8.51845562e-01 5.39926589e-01 -3.77987921e-01 1.67843819e-01
1.51858151e-01 -3.98685843e-01 8.38223845e-02 -9.56939161e-02
7.06247449e-01 5.20283699e-01 9.34038535e-02 -4.25661683e-01
-3.84548664e-01 4.27273899e-01 7.33483285e-02 -2.60710716e-01
1.51041150e+00 -1.08678393e-01 9.78249162e-02 5.62545896e-01
9.22274053e-01 -1.72266752e-01 -1.96603513e+00 -1.08919553e-01
-8.47915933e-02 -6.65316761e-01 4.38195765e-01 -5.53205550e-01
-9.09993827e-01 7.28256643e-01 7.02284813e-01 1.24307126e-01
1.19049871e+00 -1.68546245e-01 9.37092602e-01 2.84016043e-01
2.74676651e-01 -9.83255565e-01 8.09062347e-02 5.52024722e-01
4.88613881e-02 -1.39534521e+00 2.63558179e-01 -9.51773971e-02
-9.05532464e-02 1.21240890e+00 3.69762778e-01 -3.02081078e-01
5.48649073e-01 5.13291240e-01 1.60715058e-02 -1.39440075e-01
-9.15085375e-01 -3.85569125e-01 1.35719404e-01 5.84926188e-01
3.05911601e-01 -4.07483399e-01 -1.76528543e-01 -2.02398807e-01
1.50949493e-01 5.27552485e-01 5.40809035e-01 1.10997546e+00
-3.67075235e-01 -9.73307252e-01 -1.02784917e-01 2.76155740e-01
-5.33708692e-01 -7.48329759e-02 1.44120365e-01 7.06709802e-01
1.48059279e-02 1.04405105e+00 1.25957191e-01 -1.27008617e-01
3.53138387e-01 -3.23898792e-02 7.42627680e-01 -3.23377639e-01
-4.49624896e-01 -1.42309919e-01 -8.49876404e-02 -1.04936886e+00
-1.01288664e+00 -9.33233142e-01 -1.12397778e+00 -3.92293423e-01
-1.98741674e-01 -4.97007996e-01 8.07150304e-01 1.23778236e+00
7.52832055e-01 5.33972085e-01 4.31354105e-01 -1.25231779e+00
-3.82761866e-01 -3.90208900e-01 -6.11473583e-02 3.71405542e-01
5.25252104e-01 -6.59256279e-01 -2.98562437e-01 4.89714473e-01] | [8.596031188964844, -1.2110975980758667] |
084f425b-76b2-45f6-a4ce-048df0d2b0b3 | quantum-annealing-for-single-image-super | 2304.08924 | null | https://arxiv.org/abs/2304.08924v1 | https://arxiv.org/pdf/2304.08924v1.pdf | Quantum Annealing for Single Image Super-Resolution | This paper proposes a quantum computing-based algorithm to solve the single image super-resolution (SISR) problem. One of the well-known classical approaches for SISR relies on the well-established patch-wise sparse modeling of the problem. Yet, this field's current state of affairs is that deep neural networks (DNNs) have demonstrated far superior results than traditional approaches. Nevertheless, quantum computing is expected to become increasingly prominent for machine learning problems soon. As a result, in this work, we take the privilege to perform an early exploration of applying a quantum computing algorithm to this important image enhancement problem, i.e., SISR. Among the two paradigms of quantum computing, namely universal gate quantum computing and adiabatic quantum computing (AQC), the latter has been successfully applied to practical computer vision problems, in which quantum parallelism has been exploited to solve combinatorial optimization efficiently. This work demonstrates formulating quantum SISR as a sparse coding optimization problem, which is solved using quantum annealers accessed via the D-Wave Leap platform. The proposed AQC-based algorithm is demonstrated to achieve improved speed-up over a classical analog while maintaining comparable SISR accuracy. | ['Luc van Gool', 'Suryansh Kumar', 'Han Yao Choong'] | 2023-04-18 | null | null | null | null | ['image-super-resolution', 'image-enhancement', 'combinatorial-optimization'] | ['computer-vision', 'computer-vision', 'methodology'] | [ 8.40748131e-01 -1.89722255e-01 1.76683038e-01 -1.31650001e-01
-8.80026460e-01 -5.48905171e-02 5.00201404e-01 -1.42989233e-01
-5.89981437e-01 6.33287787e-01 -2.12452784e-01 -8.52761939e-02
-2.65536338e-01 -1.04969716e+00 -4.20541644e-01 -1.11079431e+00
1.84188709e-01 3.69687676e-01 2.19746120e-02 -8.14206302e-01
5.79066515e-01 5.07633924e-01 -1.80848801e+00 7.49264732e-02
9.25533950e-01 8.71647835e-01 1.45301357e-01 5.13735592e-01
1.46131769e-01 7.73620605e-01 -2.72558965e-02 -4.25825000e-01
4.44739670e-01 -6.65363073e-01 -8.21801484e-01 -4.83404458e-01
3.56448591e-01 -1.59669220e-01 -9.52748895e-01 1.67565739e+00
4.64417279e-01 2.77298361e-01 2.85227209e-01 -5.89798331e-01
-8.53818059e-01 2.76812822e-01 -4.13218230e-01 3.95981282e-01
1.12293594e-01 6.88707829e-02 1.21019804e+00 -5.08245587e-01
7.63163805e-01 7.88424075e-01 2.80982465e-01 7.24404275e-01
-1.37971330e+00 -5.17820418e-01 -1.03836668e+00 9.27756727e-01
-1.89034045e+00 -4.10303116e-01 7.52933264e-01 2.07517967e-01
1.17868686e+00 1.88117683e-01 5.29501677e-01 4.34133023e-01
4.22952712e-01 2.28678986e-01 1.54832637e+00 -7.44251847e-01
5.76783121e-01 4.60503176e-02 1.09016426e-01 8.39667261e-01
1.93870798e-01 5.32088101e-01 -1.07392132e+00 8.15955624e-02
6.94256842e-01 -1.97790951e-01 -1.11929394e-01 2.63222575e-01
-1.06294513e+00 1.01223791e+00 6.53525889e-01 4.24301445e-01
-4.00700063e-01 1.35684043e-01 1.75792843e-01 1.27325490e-01
3.11968140e-02 4.70077574e-01 3.28855008e-01 -2.29753196e-01
-9.95097756e-01 3.23449969e-01 4.78884071e-01 6.52021766e-01
1.01557839e+00 9.81765091e-02 7.62694851e-02 4.31307226e-01
-3.95372882e-02 6.65505350e-01 2.98495173e-01 -1.33850741e+00
-2.88011849e-01 5.48379198e-02 -6.02494515e-02 -9.94256854e-01
-1.73579752e-01 -3.38081151e-01 -1.30126548e+00 3.90830785e-01
1.14292189e-01 4.69893843e-01 -6.71830475e-01 1.49829352e+00
1.68793589e-01 2.32492790e-01 5.41026771e-01 1.19166183e+00
8.09219003e-01 8.48776996e-01 -4.20544118e-01 -3.32387716e-01
1.45162642e+00 -6.36768699e-01 -9.20938492e-01 3.50237995e-01
3.62067640e-01 -7.47957051e-01 4.01707083e-01 5.77987552e-01
-8.93164873e-01 -3.90893847e-01 -1.27893853e+00 -4.99341398e-01
-2.69468874e-01 -4.13854867e-01 9.50424433e-01 7.37547457e-01
-1.19854236e+00 1.11171281e+00 -6.62438393e-01 -3.12549531e-01
2.99298346e-01 3.96889061e-01 -4.56770718e-01 -3.54803234e-01
-1.50960803e+00 9.75778878e-01 5.97835183e-01 8.08912292e-02
-5.41988790e-01 -3.22958946e-01 -3.68432879e-01 1.07412726e-01
1.33905023e-01 -7.26536155e-01 7.84686863e-01 -5.46815097e-01
-1.93959069e+00 1.06365550e+00 -3.27604145e-01 -7.73858964e-01
-3.00250620e-01 4.44739968e-01 -4.51083332e-01 5.89166999e-01
-1.75787657e-01 5.63858449e-01 8.34653199e-01 -5.08222580e-01
-3.18873316e-01 -5.95764935e-01 1.85485780e-01 2.08718300e-01
-2.14882791e-01 2.06665117e-02 1.13547131e-01 5.29730134e-03
4.10454810e-01 -1.07656908e+00 -3.99817884e-01 -3.52473587e-01
-9.51762199e-02 -1.33901849e-01 2.99162716e-01 -3.98958743e-01
1.20636880e+00 -2.10189509e+00 4.77030426e-01 1.50878178e-02
3.75210077e-01 3.48500222e-01 -4.71857004e-03 4.63631362e-01
6.00973144e-02 -2.28281438e-01 -4.84756380e-01 -7.49883242e-03
-7.85484537e-02 2.83838719e-01 -1.95767269e-01 6.43543839e-01
2.22314686e-01 1.01075721e+00 -7.50211954e-01 -5.08933485e-01
2.15834171e-01 8.53895962e-01 -8.74594748e-01 -3.53737682e-01
8.51436034e-02 5.62093854e-01 -2.97113359e-01 5.62172472e-01
9.41663980e-01 -4.10917073e-01 9.22748893e-02 -5.98170877e-01
-6.72776520e-01 8.80033523e-02 -1.22078204e+00 1.87030458e+00
-1.06131233e-01 5.43101132e-01 -5.07160835e-03 -1.23214614e+00
8.25758994e-01 7.07419738e-02 5.38836360e-01 -1.29598820e+00
6.05416782e-02 6.83338821e-01 7.96289295e-02 -2.90979952e-01
9.88140166e-01 -5.83157241e-01 2.69362122e-01 1.59675866e-01
2.40734950e-01 -3.94853055e-01 2.50522703e-01 3.34576458e-01
1.10489094e+00 8.79497230e-02 4.70852375e-01 -4.29717988e-01
7.17486560e-01 3.05703640e-01 2.62752473e-01 8.36153209e-01
-4.00942922e-01 5.47577858e-01 -3.45236436e-02 -5.98030686e-01
-1.27813327e+00 -8.26648831e-01 -5.88015616e-01 4.52596933e-01
5.70146322e-01 -6.16495132e-01 -7.81551063e-01 4.30184186e-01
-6.57625258e-01 5.78238487e-01 -2.09725514e-01 -1.59819350e-01
-5.69779575e-01 -1.24552941e+00 4.12984401e-01 -1.83967799e-01
9.99156892e-01 -8.67758036e-01 -8.85885596e-01 3.63932788e-01
-1.23707227e-01 -1.37452662e+00 1.97043464e-01 3.32760036e-01
-8.66964877e-01 -7.31096089e-01 -5.74815392e-01 -5.11008143e-01
2.43241206e-01 5.42298615e-01 7.86513925e-01 -1.04036212e-01
-8.03343892e-01 2.25186661e-01 -4.16348368e-01 1.37293369e-01
-4.69092637e-01 -8.49411450e-03 2.44251192e-01 1.81620643e-01
6.75309956e-01 -6.84151292e-01 -6.41639113e-01 -3.50929260e-01
-1.13693511e+00 1.86532721e-01 6.77482069e-01 1.01501918e+00
1.09040332e+00 3.09654683e-01 1.31508216e-01 -5.63706636e-01
2.12977752e-01 -4.93698604e-02 -7.27348506e-01 1.63104147e-01
-6.90336287e-01 5.26220560e-01 7.76711166e-01 2.17729032e-01
-1.05423164e+00 1.21433415e-01 -4.64514554e-01 -8.30296278e-02
-9.06974152e-02 6.52717710e-01 3.83740216e-01 -8.54067743e-01
7.61060476e-01 7.75160134e-01 2.88197212e-02 1.12572256e-02
2.72463351e-01 7.45376766e-01 6.71264708e-01 -3.17962646e-01
9.81464267e-01 9.71215487e-01 9.76672113e-01 -1.44657314e+00
-9.39393401e-01 -5.09394586e-01 -4.62798446e-01 -5.83884008e-02
1.36273086e+00 -8.32630217e-01 -8.87366235e-01 4.31074649e-01
-1.21995425e+00 3.02228481e-01 -2.13379562e-01 5.95218658e-01
-6.02310956e-01 5.71561038e-01 -6.25357330e-01 -9.32120621e-01
-4.16376889e-01 -1.38707030e+00 1.04443622e+00 5.86795390e-01
5.23151934e-01 -6.99407041e-01 6.45726100e-02 5.39298713e-01
7.21628189e-01 2.00896412e-01 6.86317980e-01 7.81226531e-02
-1.20989156e+00 -3.79431807e-02 -4.77146983e-01 2.41319075e-01
-4.25969303e-01 -3.58851999e-01 -1.10208845e+00 -2.73852438e-01
3.25639457e-01 -4.29480404e-01 9.44997251e-01 1.83715850e-01
9.38444495e-01 3.42054486e-01 2.78984606e-01 8.73950958e-01
1.97003794e+00 1.20206922e-01 1.10945821e+00 3.15479398e-01
4.47382420e-01 1.35112569e-01 2.65064836e-01 3.72347027e-01
2.65945643e-01 9.95568216e-01 6.25996172e-01 3.45744520e-01
-1.13252148e-01 8.02453533e-02 1.59875810e-01 1.26353562e+00
-7.41295934e-01 5.00101984e-01 -6.61358297e-01 -9.69283506e-02
-1.47478187e+00 -1.39244354e+00 -4.20255035e-01 2.23351336e+00
6.25496686e-01 -1.72929153e-01 -5.44348180e-01 7.13320971e-02
6.04909658e-01 2.60626376e-01 -4.80608940e-01 -4.85409796e-01
-4.42284971e-01 9.24011290e-01 5.75384557e-01 3.77232760e-01
-9.37295973e-01 1.01403368e+00 5.26529074e+00 1.16557670e+00
-1.18874371e+00 3.94217610e-01 2.60863185e-01 2.49596030e-01
-2.95290440e-01 2.29684129e-01 -8.02257538e-01 1.19442843e-01
1.20981169e+00 -3.03209782e-01 1.42098165e+00 4.61182088e-01
-9.93441492e-02 -3.09404820e-01 -6.65837646e-01 1.78122580e+00
1.22719534e-01 -1.78373122e+00 6.56105056e-02 1.99206591e-01
9.08361435e-01 3.47269416e-01 2.72154748e-01 2.02660769e-01
-5.09160936e-01 -1.07781029e+00 3.76984149e-01 3.88497859e-01
1.10085428e+00 -8.16549897e-01 7.09839106e-01 2.63965458e-01
-1.08767188e+00 -4.90158685e-02 -7.24333584e-01 -3.75353754e-01
1.81422740e-01 5.15562236e-01 1.41530171e-01 8.98703396e-01
6.91173792e-01 5.99703491e-01 -3.08915943e-01 7.62842417e-01
1.21053740e-01 2.99561888e-01 -3.25339586e-01 -1.13422982e-01
4.29705232e-01 -9.07901645e-01 5.93054831e-01 5.86997628e-01
4.93487716e-01 6.26530826e-01 -3.88493478e-01 1.23657453e+00
-4.96784300e-02 7.18116760e-02 -6.17700160e-01 -3.33826721e-01
1.26602218e-01 1.44130480e+00 -6.81079745e-01 -1.96401671e-01
-3.51054251e-01 1.18916273e+00 1.22790895e-01 1.04691871e-01
-7.51141310e-01 -5.81889391e-01 3.06469083e-01 -2.23835513e-01
1.50533870e-01 -3.27600956e-01 -3.65583360e-01 -1.56835628e+00
-2.86990196e-01 -6.87921405e-01 -7.39456639e-02 -5.56995511e-01
-8.33024323e-01 6.14363134e-01 -3.36736560e-01 -1.06910908e+00
2.71183699e-01 -7.57831335e-01 -2.36218750e-01 1.08331406e+00
-1.95447755e+00 -7.75145829e-01 -2.76433408e-01 5.11617422e-01
-4.61859908e-03 -2.05382511e-01 1.28142345e+00 4.58856046e-01
-3.92535299e-01 1.72887534e-01 6.34981453e-01 -2.62950659e-01
1.73415884e-01 -1.05766582e+00 1.03996605e-01 9.25574601e-01
5.04621923e-01 7.68829226e-01 8.95254552e-01 -1.94566980e-01
-2.17588711e+00 -5.73354185e-01 8.82732928e-01 5.75015880e-02
7.35016108e-01 -8.72413591e-02 -8.22398484e-01 -5.37971631e-02
2.51925141e-01 2.59188622e-01 5.83468437e-01 -3.00717413e-01
-3.71094912e-01 -2.25323111e-01 -1.18939650e+00 3.00691158e-01
9.77912605e-01 -1.12110651e+00 -3.21649879e-01 4.95061487e-01
3.72470289e-01 -3.87631088e-01 -8.40709507e-01 1.29269361e-01
5.00271201e-01 -1.32294691e+00 9.19230163e-01 -1.13328286e-01
7.02116430e-01 -4.69116688e-01 -6.78894937e-01 -1.00506461e+00
-3.80440265e-01 -9.28825736e-01 -5.92201874e-02 3.57757002e-01
-6.28222153e-02 -7.10731685e-01 7.04595327e-01 2.40601256e-01
-1.61595255e-01 -5.20059049e-01 -1.63370621e+00 -6.39427423e-01
3.16220410e-02 -2.33111888e-01 1.69167653e-01 7.81039834e-01
-1.26240820e-01 2.59780705e-01 -4.23504651e-01 3.28333169e-01
1.25703323e+00 2.24527076e-01 7.74155930e-02 -1.10762048e+00
-5.06860316e-01 -2.78172165e-01 -7.57608473e-01 -7.65461087e-01
5.68134114e-02 -1.32376325e+00 -1.74063385e-01 -1.14747846e+00
5.35542548e-01 -9.47270617e-02 -4.48925465e-01 -2.41784140e-01
1.72527224e-01 8.37450206e-01 3.13909531e-01 3.16157162e-01
-7.04339445e-01 6.28306329e-01 1.22926319e+00 -2.21293524e-01
4.34045009e-02 -4.92120862e-01 -4.15629268e-01 2.50230402e-01
7.34630704e-01 -4.65920180e-01 1.45315044e-02 -3.30071062e-01
7.09644854e-01 1.67606324e-01 6.18465960e-01 -1.29542542e+00
6.40276611e-01 6.12275451e-02 -3.22197020e-01 -2.61924595e-01
5.34651935e-01 -5.07435679e-01 2.76068181e-01 6.32791162e-01
-1.43150076e-01 -3.20461005e-01 -2.96250224e-01 4.21757489e-01
-5.64808547e-01 -5.72751641e-01 1.32216847e+00 -3.99049312e-01
-1.10713649e+00 3.55455250e-01 -4.79673035e-02 -1.67458549e-01
8.02594781e-01 -1.02199033e-01 -3.68933558e-01 -6.23206384e-02
-5.53399920e-01 -5.20137012e-01 4.02601302e-01 -2.41831556e-01
7.99686193e-01 -1.19201529e+00 -6.67900622e-01 2.33862504e-01
9.84330475e-02 -2.23245621e-01 7.92595327e-01 1.02347386e+00
-9.53351319e-01 6.78924680e-01 -4.30911064e-01 -5.83797336e-01
-9.74822104e-01 5.38178265e-01 3.74091059e-01 -1.47518143e-01
-7.59939551e-01 8.38989139e-01 -2.89321899e-01 -1.12720519e-01
-3.47476900e-01 4.08712029e-01 6.98475912e-02 -3.60244930e-01
8.06303561e-01 5.27819216e-01 1.87623471e-01 -9.52169061e-01
-3.72711122e-01 7.76685596e-01 8.70555714e-02 -3.05827633e-02
1.44020104e+00 -1.44543555e-02 -6.89030647e-01 5.98518364e-02
1.34292495e+00 -4.38303173e-01 -7.41144359e-01 -2.67632753e-01
-1.67691901e-01 -2.50045985e-01 6.40013278e-01 -3.45960289e-01
-8.00165236e-01 1.21862149e+00 9.80143011e-01 2.25038454e-01
1.24818683e+00 -1.10462196e-01 1.05419552e+00 7.86518872e-01
1.18376660e+00 -1.17234755e+00 -2.82869428e-01 5.44894397e-01
3.96772236e-01 -1.51634681e+00 1.57598436e-01 -1.28258094e-01
-2.28372127e-01 1.51475120e+00 -8.32531452e-02 -2.24326417e-01
3.75325561e-01 -2.54909471e-02 -4.48606551e-01 -2.15098888e-01
-3.46249044e-01 -4.43173915e-01 -1.28434390e-01 2.46464565e-01
2.03609720e-01 3.08191597e-01 -4.07991529e-01 -1.07910648e-01
-1.18234903e-01 2.84178287e-01 8.47719133e-01 7.94836581e-01
-4.09492612e-01 -1.11215413e+00 -3.43362153e-01 2.46811971e-01
-4.69431341e-01 -4.16375428e-01 2.77105868e-01 3.25031459e-01
1.83164299e-01 5.77557743e-01 -3.54276270e-01 -4.31796521e-01
-5.83614111e-02 -4.94366735e-02 9.88688886e-01 -3.57243806e-01
-2.11019173e-01 -2.97690243e-01 -4.32604283e-01 -7.56035566e-01
-8.44776034e-01 -5.68503261e-01 -1.40091288e+00 -5.96621454e-01
-3.18439424e-01 1.02852292e-01 9.42510843e-01 8.90007019e-01
3.18324596e-01 3.39102864e-01 5.47214627e-01 -9.16177988e-01
-9.54746068e-01 -3.70964706e-01 -9.68798578e-01 1.20362997e-01
1.66303217e-01 -5.55217981e-01 -2.49412522e-01 -1.62907764e-01] | [5.574599742889404, 4.959182262420654] |
d5f0d699-c168-41d7-ac0f-544cccaba33b | credit-card-fraud-detection-using-asexual | 2306.01008 | null | https://arxiv.org/abs/2306.01008v1 | https://arxiv.org/pdf/2306.01008v1.pdf | Credit Card Fraud Detection Using Asexual Reproduction Optimization | As the number of credit card users has increased, detecting fraud in this domain has become a vital issue. Previous literature has applied various supervised and unsupervised machine learning methods to find an effective fraud detection system. However, some of these methods require an enormous amount of time to achieve reasonable accuracy. In this paper, an Asexual Reproduction Optimization (ARO) approach was employed, which is a supervised method to detect credit card fraud. ARO refers to a kind of production in which one parent produces some offspring. By applying this method and sampling just from the majority class, the effectiveness of the classification is increased. A comparison to Artificial Immune Systems (AIS), which is one of the best methods implemented on current datasets, has shown that the proposed method is able to remarkably reduce the required training time and at the same time increase the recall that is important in fraud detection problems. The obtained results show that ARO achieves the best cost in a short time, and consequently, it can be considered a real-time fraud detection system. | ['Mohammadreza Fani Sani', 'Ramin Yavari', 'Nila Bahrambeik', 'Mohammad Reza Sadeghi Moghadam', 'Taha Mansouri', 'Anahita Farhang Ghahfarokhi'] | 2023-05-31 | null | null | null | null | ['fraud-detection'] | ['miscellaneous'] | [ 2.80313641e-01 -1.76949829e-01 -1.10804755e-02 -1.97498336e-01
2.14637965e-01 -4.57027815e-02 3.37725669e-01 6.62058175e-01
-6.95358276e-01 9.40427899e-01 -6.18659854e-01 -6.47334382e-02
-1.59500733e-01 -1.30786502e+00 -3.31688970e-01 -6.43170118e-01
1.39208645e-01 6.40939236e-01 4.94286232e-02 -1.57273009e-01
7.85838962e-01 5.62992156e-01 -1.87490559e+00 2.30080578e-02
1.35501146e+00 9.40383255e-01 1.30255744e-02 2.31246620e-01
-3.88017833e-01 7.77149320e-01 -8.40265274e-01 -7.23991275e-01
3.39589752e-02 -7.63623595e-01 -4.44441348e-01 3.97460647e-02
-4.43542778e-01 -2.46795207e-01 4.49054629e-01 1.08351099e+00
3.64700943e-01 -7.48421475e-02 8.96705091e-01 -1.18402553e+00
-1.78701833e-01 2.35855505e-01 -7.18202889e-01 9.31313112e-02
2.99936503e-01 -3.38979572e-01 5.21050453e-01 -6.79797590e-01
4.13670540e-01 8.74915898e-01 5.98297715e-01 1.69169828e-01
-1.05147409e+00 -8.85496259e-01 -4.16833520e-01 7.81023875e-02
-1.09824514e+00 -2.90515926e-02 8.44172835e-01 -3.20278347e-01
8.44151437e-01 2.22538680e-01 9.92769718e-01 4.27822888e-01
3.20210755e-01 8.67700100e-01 1.01452112e+00 -8.75461161e-01
5.90632915e-01 4.92586762e-01 1.81740403e-01 4.67917234e-01
1.10688829e+00 6.18005963e-03 -3.83277863e-01 -4.20112342e-01
5.82732260e-01 3.47537458e-01 8.31298158e-02 -3.57747823e-01
-5.29551148e-01 1.01048124e+00 1.67301238e-01 6.14513457e-01
-5.33315480e-01 -3.18423837e-01 4.33284789e-01 3.06763023e-01
2.94427931e-01 6.25683665e-01 -1.03930831e-01 -2.12303236e-01
-1.01691246e+00 3.02199543e-01 9.29582179e-01 2.93570012e-01
5.18228412e-01 -2.65670773e-02 2.19983876e-01 8.34170461e-01
2.09110230e-01 3.20837528e-01 7.77661443e-01 -4.56666380e-01
4.41117883e-01 1.46211267e+00 3.26571286e-01 -1.19802976e+00
-1.77934900e-01 -4.46029782e-01 -8.37257504e-01 4.37183887e-01
4.22945261e-01 -6.70150965e-02 -6.71302378e-01 1.31190240e+00
1.60287440e-01 -3.50033492e-02 2.82117635e-01 5.56690812e-01
2.38044366e-01 5.24583697e-01 1.21627510e-01 -5.33441663e-01
1.11317134e+00 -5.01949012e-01 -9.04365718e-01 -5.59425168e-02
4.98435110e-01 -8.26968431e-01 5.93501270e-01 8.40894163e-01
-8.96962106e-01 -2.74106711e-01 -1.18120778e+00 5.45993924e-01
-5.66451490e-01 2.46082202e-01 9.38838065e-01 1.00696027e+00
-3.96299154e-01 6.54958010e-01 -8.08582723e-01 -6.14938676e-01
3.66490066e-01 4.73747730e-01 -2.02267602e-01 -9.56656486e-02
-9.55923319e-01 8.92609954e-01 5.07429898e-01 1.84981272e-01
-9.00301784e-02 -3.16193327e-02 -5.12091637e-01 -1.29091460e-02
6.67064860e-02 -2.72897035e-01 8.29186261e-01 -1.33226299e+00
-1.16416705e+00 7.15909660e-01 -8.54473561e-04 -7.61052847e-01
8.40317667e-01 -1.52317196e-01 -2.22409710e-01 2.32042104e-01
-7.10965768e-02 2.54700631e-01 6.85042620e-01 -8.48447204e-01
-8.23312163e-01 -6.91859126e-01 -3.57343286e-01 1.48869768e-01
-7.19069242e-01 1.59237072e-01 3.74828354e-02 -6.55680954e-01
3.30387324e-01 -6.69659495e-01 6.26452342e-02 -3.01029772e-01
1.18147340e-02 -1.87335685e-01 7.26644099e-01 -6.72561526e-01
1.37864661e+00 -1.79952991e+00 -6.26799688e-02 5.21114409e-01
-3.97281706e-01 6.96037948e-01 6.69291854e-01 6.17232263e-01
2.59699792e-01 -2.24077292e-02 -6.39797747e-01 -3.44038159e-02
-3.61827075e-01 2.85480827e-01 1.29902914e-01 1.44372433e-01
4.09729928e-01 3.23874801e-01 -7.08760023e-01 -7.07888544e-01
2.27697983e-01 2.02364028e-01 -4.61088002e-01 4.29454148e-01
1.53904229e-01 1.75125934e-02 -4.76498067e-01 9.35788631e-01
8.02126408e-01 1.21065654e-01 4.21220422e-01 3.05667013e-01
6.19530454e-02 -3.07341903e-01 -1.32138491e+00 9.62756217e-01
-1.55792132e-01 2.45196715e-01 -3.74686420e-01 -1.52136254e+00
1.51072168e+00 3.12876552e-01 4.59190577e-01 -7.95823395e-01
2.32336983e-01 7.29444385e-01 -1.70661490e-02 -7.23535299e-01
3.35929275e-01 -4.80549559e-02 1.31048486e-01 3.93471926e-01
-2.87486106e-01 7.82008842e-02 4.89658415e-01 -2.78045952e-01
8.95523429e-01 2.42208671e-02 6.62999570e-01 -5.34492508e-02
7.78411627e-01 9.85983908e-02 7.39770055e-01 6.59090042e-01
-1.52292177e-01 1.38160467e-01 5.48807442e-01 -4.86599594e-01
-9.02610362e-01 -6.09332681e-01 -2.08295763e-01 2.37862229e-01
2.20149755e-01 3.03174883e-01 -8.49142671e-01 -3.86821032e-01
3.72990817e-01 6.14751875e-01 -3.91418517e-01 -2.31215209e-01
-4.65876073e-01 -1.34490788e+00 5.27570188e-01 1.58379406e-01
9.39287424e-01 -1.44426239e+00 -1.18318093e+00 5.20830214e-01
-1.00679956e-02 -4.03130293e-01 5.14733195e-01 1.56781524e-01
-1.54357231e+00 -1.24314713e+00 -9.38375056e-01 -8.03690851e-01
1.01300776e+00 -9.19555500e-02 8.05170298e-01 5.79398334e-01
-5.72212756e-01 -5.51514864e-01 -7.39151061e-01 -7.42918432e-01
-5.70616126e-01 -9.06193443e-03 -1.94852874e-01 2.41540641e-01
8.31819415e-01 -4.16638136e-01 -4.47002530e-01 1.09657064e-01
-1.05433464e+00 -3.31857562e-01 8.22637022e-01 1.12372732e+00
2.03312874e-01 4.18212295e-01 1.07526875e+00 -1.16992962e+00
8.58247995e-01 -4.63822573e-01 -7.36801744e-01 3.29503268e-01
-1.13671553e+00 -1.95473246e-02 5.90098321e-01 -1.32078588e-01
-1.09649241e+00 -1.47173107e-02 1.25954106e-01 1.68834820e-01
-1.04439303e-01 4.32745844e-01 8.94479528e-02 -1.61127865e-01
4.20566797e-01 2.95909494e-01 3.50435555e-01 -6.43947124e-01
-6.50850773e-01 1.07028317e+00 6.28712447e-03 -1.64457664e-01
1.84107453e-01 2.54702538e-01 2.41687313e-01 -6.32324815e-01
-2.32555106e-01 -5.55041373e-01 -3.25224489e-01 -1.23612083e-01
4.04707462e-01 -4.18871135e-01 -7.40821064e-01 1.03217351e+00
-9.52695310e-01 4.19095933e-01 1.22787885e-01 6.16223872e-01
-1.80826917e-01 4.09238726e-01 -4.65650916e-01 -1.57845294e+00
-5.47457278e-01 -8.09338093e-01 3.80259305e-01 5.25196254e-01
-1.00502998e-01 -6.12960577e-01 -1.42206892e-01 3.75038326e-01
3.05455506e-01 5.74534297e-01 9.00213420e-01 -7.52500653e-01
-1.84863240e-01 -7.70487726e-01 -1.00257307e-01 4.33506876e-01
2.37122774e-01 4.90510166e-02 -4.55096126e-01 -1.53100953e-01
1.89901039e-01 -2.66458571e-01 7.05698907e-01 1.35223031e-01
8.45141828e-01 -1.63076937e-01 -2.88052082e-01 2.29811400e-01
1.67304027e+00 9.57034945e-01 8.62728953e-01 9.84647214e-01
-6.49968907e-02 7.55273163e-01 1.24583161e+00 6.04298711e-01
-1.20072737e-02 4.16720986e-01 4.75208372e-01 -2.86184549e-01
5.86001456e-01 -8.97398219e-02 9.10490900e-02 4.32992667e-01
-2.47038603e-01 -1.47682339e-01 -7.27469683e-01 4.23629463e-01
-1.96568930e+00 -1.06802833e+00 -6.13496043e-02 2.52545047e+00
6.77442670e-01 3.08160722e-01 2.40200371e-01 1.00638092e+00
8.90897334e-01 -5.63692212e-01 -5.62538028e-01 -8.26496124e-01
-9.67791677e-03 3.21423560e-01 3.52166325e-01 -1.08025931e-02
-8.97443831e-01 3.96183521e-01 5.79026842e+00 5.71438730e-01
-9.68945324e-01 -2.30404124e-01 7.21629679e-01 2.35522911e-01
1.26817793e-01 -2.21675802e-02 -3.81523848e-01 8.21814358e-01
5.21726787e-01 -1.21033795e-01 2.34943822e-01 8.06295037e-01
1.69532195e-01 -6.82530105e-01 -6.18130922e-01 7.03228176e-01
1.30282551e-01 -7.38632560e-01 1.85290858e-01 5.46812192e-02
5.47547519e-01 -8.29263508e-01 -4.80463326e-01 -1.38335600e-02
-3.83817047e-01 -9.16929364e-01 2.44262874e-01 7.18137860e-01
1.02336541e-01 -1.35029209e+00 1.68399119e+00 5.73900044e-01
-7.52149045e-01 -3.37398320e-01 -4.39394087e-01 -3.71335685e-01
-3.40613760e-02 7.98258364e-01 -7.99905300e-01 5.61044395e-01
7.80200362e-01 2.84980416e-01 -5.47355294e-01 1.54080856e+00
-4.94029261e-02 5.07981241e-01 -3.64742905e-01 -6.59144759e-01
-7.53196925e-02 -6.61163032e-01 3.19839120e-01 9.51399505e-01
5.95624983e-01 -1.10926010e-01 -2.47122020e-01 7.27389693e-01
1.52249500e-01 5.54351866e-01 -6.79685295e-01 -8.70316476e-02
4.55643594e-01 8.28722775e-01 -1.01271152e+00 -3.74885082e-01
-7.80406371e-02 1.01332498e+00 1.53748780e-01 -1.46172389e-01
-6.10530138e-01 -9.29529786e-01 -1.68252122e-02 1.00289449e-01
1.02851681e-01 3.29849541e-01 -3.52671266e-01 -8.47212136e-01
2.18072027e-01 -7.44389534e-01 5.73732913e-01 -4.37562525e-01
-9.45913851e-01 2.84744710e-01 -2.29742616e-01 -1.34246838e+00
-3.87062639e-01 -5.35895944e-01 -4.89925623e-01 5.50349534e-01
-1.37270951e+00 -5.91294885e-01 -2.19289139e-01 2.11416960e-01
2.81490386e-01 -4.07420695e-01 9.20892954e-01 4.07883763e-01
-7.15957284e-01 6.37082160e-01 3.99003327e-01 1.17549393e-02
4.34701800e-01 -1.07919514e+00 -2.98230529e-01 9.23469126e-01
-2.08732054e-01 6.65267289e-01 7.13004947e-01 -8.52946222e-01
-1.09552860e+00 -5.42287648e-01 1.20365024e+00 2.97859758e-01
-9.55399498e-03 4.14140411e-02 -8.32161367e-01 1.10905848e-01
-1.51723534e-01 -3.82297128e-01 4.96613771e-01 -1.29603043e-01
2.56012022e-01 -4.60124314e-01 -1.81447279e+00 1.95336282e-01
4.64250326e-01 1.18568234e-01 -7.78925776e-01 4.41993847e-02
-9.08097550e-02 2.50535328e-02 -7.92465210e-01 4.94228274e-01
6.30984783e-01 -1.42143798e+00 7.05485344e-01 -2.80952692e-01
4.60148722e-01 -1.63822249e-01 4.94868428e-01 -8.62359047e-01
1.69527337e-01 -1.67548612e-01 -1.22051358e-01 1.49332070e+00
2.84755677e-01 -1.06282020e+00 9.35407996e-01 4.39666092e-01
5.96302629e-01 -8.49239171e-01 -7.30129719e-01 -7.58460462e-01
-3.58229965e-01 2.23064929e-01 6.79742575e-01 9.85061526e-01
9.45135355e-02 -2.92958081e-01 -2.56363571e-01 -3.20072412e-01
8.58907163e-01 2.99506307e-01 5.63491404e-01 -1.57698643e+00
-5.45821600e-02 -2.88039535e-01 -7.12396860e-01 9.02447551e-02
-4.09663945e-01 -4.39078122e-01 -3.26044887e-01 -1.22695386e+00
2.27071136e-01 -6.11563325e-01 -5.05497694e-01 2.11080804e-01
-3.15459281e-01 2.00436547e-01 -6.42816424e-02 3.85645866e-01
1.21377960e-01 1.39190271e-01 8.36810172e-01 2.01715723e-01
-3.89752716e-01 4.15104002e-01 -5.95377564e-01 8.05157185e-01
1.03355503e+00 -7.06095278e-01 -2.07743824e-01 5.74428104e-02
2.58055478e-01 2.73707509e-02 8.10766872e-03 -9.81414557e-01
6.22024909e-02 -1.99460879e-01 6.43311143e-01 -4.93626744e-01
1.61309943e-01 -8.19444895e-01 2.68625647e-01 1.09314728e+00
-5.02972305e-02 7.12710842e-02 -5.94077408e-02 4.65073943e-01
-5.30121744e-01 -9.40801978e-01 7.51627862e-01 -2.31701061e-01
-5.14926374e-01 -2.96633959e-01 -3.47485423e-01 -4.42838937e-01
1.34118271e+00 -6.65847182e-01 -2.86794994e-02 -4.01346236e-02
-1.80896297e-01 5.17811924e-02 6.35524690e-01 9.37958285e-02
5.21978438e-01 -9.49682295e-01 -5.30426681e-01 1.57564119e-01
1.21748596e-01 -2.56827831e-01 -1.43717736e-01 6.96472168e-01
-1.14790416e+00 2.38563761e-01 -6.53693140e-01 -3.58701468e-01
-1.30817246e+00 2.54681200e-01 8.34608003e-02 -5.94838858e-01
-3.27213109e-01 5.44022799e-01 -6.81641638e-01 -1.25828460e-01
1.80423424e-01 3.77787873e-02 -8.21685851e-01 2.20048562e-01
2.75025159e-01 6.44135654e-01 2.05401614e-01 -3.89431477e-01
-4.28942949e-01 5.15621662e-01 5.34304082e-02 3.01060956e-02
1.32720113e+00 3.65904212e-01 -4.21598047e-01 3.08231860e-01
6.79519117e-01 -4.92879786e-02 -5.90009153e-01 3.81397367e-01
4.43967432e-01 -8.10886264e-01 -1.96253747e-01 -8.18902433e-01
-8.53926182e-01 6.37816548e-01 7.39210129e-01 5.60637295e-01
1.54703522e+00 -8.67157698e-01 6.58764780e-01 4.77395892e-01
5.80671430e-01 -1.54186368e+00 -1.67993367e-01 -7.73513168e-02
5.33899605e-01 -1.37933290e+00 2.45364457e-01 -4.21343476e-01
-5.65706015e-01 1.29420078e+00 6.15393877e-01 -2.98810571e-01
2.56866336e-01 4.85915914e-02 -2.30307981e-01 3.06744780e-02
-3.02777797e-01 3.54208723e-02 -3.43835682e-01 3.26824278e-01
5.30561626e-01 4.10718843e-02 -1.33219457e+00 4.93153691e-01
1.39796790e-02 5.06826878e-01 4.45339978e-01 1.30750167e+00
-5.51432252e-01 -1.38287497e+00 -6.37911022e-01 7.60729432e-01
-8.21853697e-01 2.76860207e-01 -4.87833530e-01 8.52603674e-01
3.58600527e-01 1.17880988e+00 4.88861240e-02 6.91525545e-03
3.01587522e-01 6.97292760e-02 3.84389222e-01 -2.59583026e-01
-8.88837874e-01 -3.05723637e-01 6.00396423e-03 -1.68922976e-01
-6.48718536e-01 -6.51494026e-01 -1.24325824e+00 -2.10532025e-01
-7.85310328e-01 5.48733950e-01 8.16896439e-01 8.63065302e-01
1.51223034e-01 2.18130141e-01 7.88577914e-01 -2.93606818e-01
-5.80349147e-01 -9.55894411e-01 -8.42008114e-01 6.88812256e-01
-1.82391346e-01 -8.31256568e-01 -4.16243166e-01 -1.58733875e-01] | [8.145827293395996, 4.739491939544678] |
05a1777c-0f77-444a-b73f-563f588fc97f | unfolding-the-alternating-optimization-for | 2010.02631 | null | https://arxiv.org/abs/2010.02631v4 | https://arxiv.org/pdf/2010.02631v4.pdf | Unfolding the Alternating Optimization for Blind Super Resolution | Previous methods decompose blind super resolution (SR) problem into two sequential steps: \textit{i}) estimating blur kernel from given low-resolution (LR) image and \textit{ii}) restoring SR image based on estimated kernel. This two-step solution involves two independently trained models, which may not be well compatible with each other. Small estimation error of the first step could cause severe performance drop of the second one. While on the other hand, the first step can only utilize limited information from LR image, which makes it difficult to predict highly accurate blur kernel. Towards these issues, instead of considering these two steps separately, we adopt an alternating optimization algorithm, which can estimate blur kernel and restore SR image in a single model. Specifically, we design two convolutional neural modules, namely \textit{Restorer} and \textit{Estimator}. \textit{Restorer} restores SR image based on predicted kernel, and \textit{Estimator} estimates blur kernel with the help of restored SR image. We alternate these two modules repeatedly and unfold this process to form an end-to-end trainable network. In this way, \textit{Estimator} utilizes information from both LR and SR images, which makes the estimation of blur kernel easier. More importantly, \textit{Restorer} is trained with the kernel estimated by \textit{Estimator}, instead of ground-truth kernel, thus \textit{Restorer} could be more tolerant to the estimation error of \textit{Estimator}. Extensive experiments on synthetic datasets and real-world images show that our model can largely outperform state-of-the-art methods and produce more visually favorable results at much higher speed. The source code is available at https://github.com/greatlog/DAN.git. | ['Tieniu Tan', 'Liang Wang', 'Shang Li', 'Yan Huang', 'Zhengxiong Luo'] | 2020-10-06 | null | http://proceedings.neurips.cc/paper/2020/hash/3d2d8ccb37df977cb6d9da15b76c3f3a-Abstract.html | http://proceedings.neurips.cc/paper/2020/file/3d2d8ccb37df977cb6d9da15b76c3f3a-Paper.pdf | neurips-2020-12 | ['burst-image-super-resolution'] | ['computer-vision'] | [ 1.94919154e-01 -3.17803651e-01 2.13240102e-01 -2.08674669e-01
-9.02383864e-01 -5.05568862e-01 1.38883218e-01 -6.61065519e-01
-2.43354470e-01 8.17645371e-01 1.97471797e-01 -3.01544756e-01
-1.25139989e-02 -5.38696170e-01 -7.40176558e-01 -7.69940972e-01
4.26234424e-01 -2.06991121e-01 2.03651935e-01 1.41640872e-01
3.09886694e-01 1.75614059e-01 -1.21706605e+00 1.88518003e-01
1.37374818e+00 9.89634395e-01 7.89204299e-01 7.39760756e-01
1.27917215e-01 1.09062541e+00 -3.30929935e-01 -5.46643995e-02
2.37526625e-01 -5.38422346e-01 -5.26827037e-01 6.06346689e-02
2.95292944e-01 -8.67067218e-01 -7.21583247e-01 1.44198930e+00
5.80580056e-01 7.14969784e-02 4.43436146e-01 -6.09017789e-01
-1.17564404e+00 4.78506386e-01 -1.04019964e+00 3.18384260e-01
1.80586472e-01 3.87577534e-01 4.92244244e-01 -9.64691997e-01
1.73292115e-01 9.51400697e-01 5.69422126e-01 4.47505414e-01
-1.22331095e+00 -7.13125765e-01 -4.06345874e-02 1.02259330e-01
-1.43461597e+00 -6.91958487e-01 7.13890910e-01 -3.16529334e-01
3.02326620e-01 3.41882169e-01 9.04246420e-02 8.01210940e-01
1.01202846e-01 8.21731508e-01 1.40037680e+00 1.48750395e-02
9.57873911e-02 9.75890085e-02 2.05391988e-01 4.90872860e-01
1.10533565e-01 2.00092360e-01 -1.68377951e-01 1.04673579e-01
1.35946465e+00 1.40799478e-01 -1.01103330e+00 -1.72205456e-02
-1.34277916e+00 3.61221015e-01 6.34407103e-01 2.19237879e-01
-3.56332362e-01 1.40407503e-01 -1.99667625e-02 2.55033821e-01
5.10666132e-01 3.57738510e-03 -2.33409643e-01 -1.58326942e-02
-1.05240691e+00 -1.05662875e-01 3.84229779e-01 8.48971248e-01
8.91841233e-01 1.40347257e-01 -2.35739589e-01 1.25708401e+00
3.25024903e-01 4.98476267e-01 5.28295338e-01 -1.12642217e+00
3.69515002e-01 1.79256842e-01 4.96962726e-01 -6.58121705e-01
-1.15472004e-01 -5.72713256e-01 -1.17439210e+00 3.68170530e-01
5.22344112e-01 -1.41034946e-01 -1.10363710e+00 1.41995680e+00
2.08649728e-02 4.35795158e-01 -1.89281779e-03 1.47504961e+00
8.19567263e-01 6.61160469e-01 -3.47726464e-01 -3.72551203e-01
1.24057555e+00 -1.05372536e+00 -5.56562364e-01 -4.15178090e-01
1.04054853e-01 -1.06387603e+00 1.03037345e+00 3.12857598e-01
-1.18336487e+00 -7.99595952e-01 -9.84572887e-01 -2.13151917e-01
2.04382151e-01 8.14657032e-01 3.50621939e-01 2.49038041e-01
-1.22030675e+00 6.33317113e-01 -6.90053999e-01 -5.69011159e-02
3.48320991e-01 1.13434941e-01 -1.43484488e-01 -3.21226031e-01
-9.26620007e-01 8.28094959e-01 7.77877197e-02 5.94370306e-01
-8.78700435e-01 -4.91762906e-01 -7.81193018e-01 3.93523276e-02
1.40118018e-01 -8.17653060e-01 1.23188972e+00 -9.84991848e-01
-1.57412493e+00 6.09871924e-01 -3.58630091e-01 -1.02792852e-01
6.72599971e-01 -4.67328995e-01 -3.81956786e-01 1.26718685e-01
1.05496220e-01 1.42605230e-01 1.16280627e+00 -1.58853960e+00
-3.50807995e-01 -2.84056008e-01 -6.50569946e-02 1.33368120e-01
-2.98792105e-02 1.19568422e-01 -6.26526535e-01 -7.55035579e-01
3.63757432e-01 -4.83283907e-01 2.61593722e-02 5.77475764e-02
-4.79226470e-01 2.21554011e-01 6.28840089e-01 -1.14344370e+00
1.25230956e+00 -2.26182437e+00 8.75300169e-02 -3.38107824e-01
4.41305012e-01 4.27175224e-01 -6.64510280e-02 1.91611145e-02
-2.47320011e-01 -1.22737460e-01 -4.67955530e-01 -2.80194789e-01
-2.43926927e-01 -2.15331748e-01 -4.45625782e-01 5.41163385e-01
-2.69388165e-02 5.64416945e-01 -8.20027113e-01 -3.04383159e-01
4.31064516e-01 6.14984989e-01 -7.61525780e-02 5.08882701e-01
1.11157015e-01 6.06820941e-01 -4.75664884e-01 4.69019741e-01
1.19186628e+00 -4.30461735e-01 -1.23224422e-01 -6.25755906e-01
-2.48995885e-01 1.72179006e-02 -1.24677193e+00 1.43075502e+00
-4.79022175e-01 5.54202557e-01 1.66525230e-01 -8.28810513e-01
9.38290238e-01 3.13534170e-01 1.96574673e-01 -6.31146431e-01
7.09831566e-02 2.94268936e-01 -2.11792096e-01 -5.43430030e-01
2.84736067e-01 -3.41350213e-02 4.88141716e-01 5.58386087e-01
-1.67867765e-01 4.58541624e-02 -2.72948623e-01 9.65595841e-02
9.51358140e-01 4.62001890e-01 6.63859174e-02 -2.95930374e-02
7.80145526e-01 -5.10272801e-01 5.30300736e-01 8.01293910e-01
-3.35470885e-01 1.12249362e+00 1.12089418e-01 -4.38753903e-01
-1.08888078e+00 -1.35414231e+00 -1.47397637e-01 5.96234798e-01
6.03859425e-01 5.24959452e-02 -9.00903165e-01 -5.23389637e-01
-2.71631211e-01 6.22608662e-01 -4.75351214e-01 8.27844720e-03
-6.71582103e-01 -8.92390609e-01 3.72644961e-01 3.19276035e-01
1.07055497e+00 -8.56690586e-01 -2.78758317e-01 2.73784529e-02
-4.37330931e-01 -9.78898823e-01 -8.21375072e-01 -3.39835644e-01
-7.28506386e-01 -1.00980651e+00 -1.02663958e+00 -6.20965421e-01
8.26014578e-01 7.90751815e-01 7.84424305e-01 3.46092224e-01
-9.16065946e-02 -2.62832791e-02 -2.06540212e-01 4.76148278e-02
-2.87587702e-01 -3.99841070e-01 -2.41930149e-02 2.51753092e-01
3.41659747e-02 -6.20399892e-01 -9.87678289e-01 4.21316445e-01
-9.61799622e-01 4.29983675e-01 8.27164769e-01 8.01538348e-01
5.57443440e-01 1.00499190e-01 5.43151140e-01 -5.20192921e-01
5.16771019e-01 -2.86631823e-01 -5.84760725e-01 2.05187410e-01
-5.79017639e-01 1.21078398e-02 8.30107808e-01 -5.24283767e-01
-1.41179204e+00 -3.78769003e-02 -5.89637756e-02 -8.03353667e-01
-1.51992902e-01 1.22633383e-01 -1.61178056e-02 3.45684737e-02
5.52612484e-01 7.56605923e-01 2.10144054e-02 -8.03020060e-01
1.99322999e-01 1.00084472e+00 8.47366154e-01 -2.45228186e-01
1.05743062e+00 4.08177316e-01 -4.90570277e-01 -6.17339849e-01
-1.02272201e+00 -2.19434410e-01 -3.02905053e-01 -1.41201586e-01
8.30691457e-01 -1.14434624e+00 -6.41245723e-01 9.15059865e-01
-1.06770992e+00 -4.57318604e-01 5.19151092e-02 6.38079643e-01
-5.12254119e-01 6.13647521e-01 -9.65349793e-01 -9.18644667e-01
-4.04105574e-01 -1.26183176e+00 1.11402535e+00 7.26994514e-01
3.40409875e-01 -7.26864815e-01 -3.40308487e-01 6.50886536e-01
5.66815913e-01 -1.42736524e-01 3.57058018e-01 2.64371447e-02
-9.68700469e-01 5.30063435e-02 -9.44024324e-01 6.99587882e-01
3.87793183e-01 -2.25605085e-01 -1.28001034e+00 -4.70859081e-01
4.82477784e-01 -1.37856632e-01 1.04260731e+00 4.81936514e-01
1.33549106e+00 -2.76811808e-01 4.55937237e-02 8.40542912e-01
1.60964608e+00 -2.30709724e-02 1.02234280e+00 3.88064593e-01
7.36259639e-01 4.51448150e-02 4.99124199e-01 8.94694254e-02
4.88508523e-01 6.27415001e-01 2.35399812e-01 -2.55922675e-01
-5.04616439e-01 -1.09396011e-01 5.89803815e-01 5.96909225e-01
-3.39561671e-01 4.45129462e-02 -4.66811389e-01 2.96297401e-01
-1.83641016e+00 -9.96482372e-01 -1.10856362e-01 2.29841709e+00
1.03223002e+00 -1.40329823e-01 -2.72051007e-01 -2.61986196e-01
9.33427513e-01 3.15395445e-01 -7.26660848e-01 1.20539501e-01
-6.59084097e-02 -1.33846268e-01 4.50029671e-01 5.88136613e-01
-8.50807190e-01 7.49639690e-01 4.84266901e+00 8.67989004e-01
-1.14364481e+00 5.93567453e-02 9.22870815e-01 4.77005579e-02
-2.07852989e-01 1.45215407e-01 -4.69745040e-01 9.49101567e-01
4.99847263e-01 1.29488800e-02 1.04576635e+00 5.38372695e-01
5.51599622e-01 -3.88098866e-01 -7.62840033e-01 1.10129309e+00
1.56358201e-02 -1.10627711e+00 -1.80523366e-01 -5.62620051e-02
4.99975324e-01 1.51108339e-01 -7.57060423e-02 1.94324374e-01
2.64743567e-01 -1.03928113e+00 6.48389757e-01 9.71581459e-01
1.03046727e+00 -4.53758240e-01 7.75472343e-01 5.36534548e-01
-1.24037588e+00 4.39136066e-02 -6.07909679e-01 1.06813684e-01
1.18932419e-01 9.11317110e-01 -1.23880327e-01 8.28087687e-01
9.95735228e-01 6.48149610e-01 -5.09513676e-01 1.11562872e+00
-3.72976273e-01 6.14515126e-01 1.47488162e-01 5.79047859e-01
-3.08646590e-01 -5.67394197e-01 7.22748935e-01 1.04731345e+00
4.22194570e-01 3.30212921e-01 1.17413262e-02 1.20600843e+00
-1.18612066e-01 -4.30370361e-01 -1.95938066e-01 3.97055596e-01
4.40616488e-01 1.44919324e+00 -4.16635603e-01 -3.59362572e-01
-4.29525822e-01 1.44708920e+00 4.06415164e-01 8.85869920e-01
-9.86401200e-01 -5.46850502e-01 6.17399275e-01 3.87052726e-03
2.33826622e-01 -1.21352963e-01 -4.04255241e-01 -1.65218437e+00
2.47305140e-01 -7.52399206e-01 8.59267637e-02 -1.45223606e+00
-1.33966911e+00 7.16185391e-01 -4.14403111e-01 -1.40051305e+00
1.70953244e-01 -3.35371107e-01 -7.30113387e-01 1.40488493e+00
-1.83956182e+00 -1.02784908e+00 -7.55347371e-01 4.89607066e-01
6.66152596e-01 2.35226691e-01 3.17566365e-01 1.95537016e-01
-8.32827806e-01 4.38665360e-01 3.97053182e-01 2.22392589e-01
1.08280206e+00 -1.28468120e+00 2.40258291e-01 1.17004943e+00
-3.88061285e-01 6.93844914e-01 7.03698158e-01 -5.80474675e-01
-1.24967599e+00 -1.06049716e+00 4.31228161e-01 -3.38644207e-01
5.77953160e-01 2.86420044e-02 -1.16161966e+00 5.68894088e-01
2.02549472e-01 3.09227481e-02 5.26686944e-02 -4.70813990e-01
-3.29342246e-01 -3.89596403e-01 -1.12794912e+00 5.31385124e-01
7.93198228e-01 -5.85641146e-01 -3.82987440e-01 1.49189562e-01
7.47819483e-01 -6.83220387e-01 -7.99782217e-01 3.28195184e-01
3.20203424e-01 -1.40230799e+00 1.09949851e+00 1.79679304e-01
6.50633931e-01 -8.87138367e-01 7.05856457e-02 -1.29855609e+00
-4.19221461e-01 -6.14838123e-01 -3.63472790e-01 1.23160005e+00
2.02624336e-01 -8.54018986e-01 2.91230559e-01 5.31511188e-01
-1.48573890e-01 -7.32158184e-01 -5.27122676e-01 -6.40135527e-01
-1.90196842e-01 -3.40990443e-03 4.33498085e-01 8.58596504e-01
-4.88365501e-01 1.50920242e-01 -6.29495025e-01 5.34738123e-01
8.98664355e-01 3.84394616e-01 6.96453035e-01 -7.27033496e-01
-4.88184750e-01 -2.84519076e-01 1.86737955e-01 -1.47908020e+00
-3.06065977e-01 -4.55744207e-01 2.57286996e-01 -1.78567731e+00
6.05162084e-01 -4.23248529e-01 -3.32398325e-01 3.52735490e-01
-7.32603669e-01 4.20117438e-01 9.45303515e-02 6.67187274e-01
-3.09618443e-01 4.88669604e-01 1.57431293e+00 1.63268685e-01
-8.99005160e-02 2.04693936e-02 -9.09239769e-01 7.00444460e-01
6.80828154e-01 -1.64926052e-02 -2.98265636e-01 -7.75431693e-01
-2.64839500e-01 3.58897597e-01 7.84635484e-01 -8.57006788e-01
2.77625203e-01 -1.33658737e-01 8.70168269e-01 -1.87361419e-01
2.20054463e-01 -4.20978636e-01 8.08965191e-02 3.00028175e-02
-3.61126778e-03 -3.09835970e-01 6.66068122e-02 4.33238238e-01
-5.46852797e-02 -1.38744861e-01 1.16723275e+00 -1.62488505e-01
-4.88726258e-01 3.97563159e-01 -6.66858815e-03 -2.79947251e-01
6.00728452e-01 -3.53880495e-01 -7.94741511e-01 -5.40056884e-01
-4.93475616e-01 1.87184513e-01 7.85205245e-01 2.90195495e-01
7.75079310e-01 -1.06142366e+00 -8.60067070e-01 1.73979029e-01
-3.04667056e-01 1.85027003e-01 6.81457222e-01 1.01479280e+00
-4.33293134e-01 -8.56428146e-02 1.06682517e-01 -4.84359533e-01
-9.29749370e-01 6.06980443e-01 6.53529584e-01 -1.22119367e-01
-6.46093786e-01 7.86192119e-01 5.23472130e-01 -2.33303174e-01
-6.68842718e-02 -7.00418502e-02 -1.19985521e-01 -4.99679893e-01
9.03103352e-01 2.71201521e-01 -3.20752114e-01 -5.42156219e-01
-1.34347662e-01 6.91630602e-01 -1.76546663e-01 -2.16001961e-02
1.32158482e+00 -6.52162194e-01 -3.55604708e-01 9.72346365e-02
1.02875972e+00 6.71964586e-02 -1.66684556e+00 -3.60350430e-01
-3.76480430e-01 -7.98602164e-01 2.83483267e-01 -9.78770077e-01
-1.18888867e+00 6.66377723e-01 7.20404744e-01 -1.79423299e-02
1.70250559e+00 -1.97706744e-01 9.05205011e-01 -5.40019386e-02
2.64610827e-01 -8.04689467e-01 -3.40134501e-02 1.98070601e-01
9.02667105e-01 -1.27222228e+00 1.38673857e-01 -3.79809439e-01
-5.78788519e-01 1.21508527e+00 7.56054521e-01 -2.38946706e-01
3.04243743e-01 3.28048207e-02 1.83337450e-01 2.52796382e-01
-4.13124204e-01 -1.31226838e-01 2.80779332e-01 3.68984014e-01
3.57034266e-01 -1.60127595e-01 -2.13176653e-01 5.24648309e-01
5.78935407e-02 2.82922953e-01 6.95286214e-01 5.63847721e-01
-5.70819557e-01 -8.33894432e-01 -5.19414723e-01 4.21535641e-01
-3.36924642e-01 -3.12430531e-01 5.70378900e-02 2.85022676e-01
2.46230997e-02 1.03935850e+00 -2.13659048e-01 -3.58190745e-01
5.63709475e-02 -3.91698420e-01 3.03708702e-01 -2.41202861e-01
-1.62479550e-01 2.79820830e-01 -1.27652541e-01 -5.65437138e-01
-5.90056442e-02 -5.13820827e-01 -1.06693459e+00 -2.94851691e-01
-3.64641309e-01 6.12767460e-03 3.31435055e-01 9.28406596e-01
2.58456528e-01 5.67313075e-01 8.87653291e-01 -8.11276376e-01
-6.53390825e-01 -1.17146826e+00 -6.32473886e-01 2.15196609e-01
7.19706416e-01 -2.57536918e-01 -5.43379188e-01 2.67233104e-01] | [11.414156913757324, -2.5814127922058105] |
ac62ae3e-90f1-49a3-aad7-8dc01dacb988 | vidm-video-implicit-diffusion-models | 2212.00235 | null | https://arxiv.org/abs/2212.00235v1 | https://arxiv.org/pdf/2212.00235v1.pdf | VIDM: Video Implicit Diffusion Models | Diffusion models have emerged as a powerful generative method for synthesizing high-quality and diverse set of images. In this paper, we propose a video generation method based on diffusion models, where the effects of motion are modeled in an implicit condition manner, i.e. one can sample plausible video motions according to the latent feature of frames. We improve the quality of the generated videos by proposing multiple strategies such as sampling space truncation, robustness penalty, and positional group normalization. Various experiments are conducted on datasets consisting of videos with different resolutions and different number of frames. Results show that the proposed method outperforms the state-of-the-art generative adversarial network-based methods by a significant margin in terms of FVD scores as well as perceptible visual quality. | ['Vishal M. Patel', 'Kangfu Mei'] | 2022-12-01 | null | null | null | null | ['video-generation'] | ['computer-vision'] | [ 2.04485282e-01 -2.57550418e-01 -3.26904766e-02 1.23164080e-01
-3.43045384e-01 -5.82431555e-01 8.88626873e-01 -6.20827496e-01
-9.09726843e-02 8.34984303e-01 4.49106067e-01 1.21044025e-01
3.64961028e-02 -9.14870739e-01 -6.46004677e-01 -9.93907273e-01
1.77136227e-01 -1.27140999e-01 3.63642961e-01 -7.01482967e-02
2.47456416e-01 4.86615896e-01 -1.27581525e+00 2.04072461e-01
8.05735648e-01 8.42400432e-01 -3.04214694e-02 6.82556808e-01
2.06116110e-01 9.47189450e-01 -8.32240164e-01 -5.50007761e-01
4.23420578e-01 -9.06041980e-01 -2.05072597e-01 5.09479403e-01
2.60089338e-01 -5.14219105e-01 -7.20990121e-01 1.16348851e+00
6.47427440e-01 2.75636107e-01 8.54907453e-01 -1.13573468e+00
-9.52423096e-01 2.56287754e-01 -6.45094931e-01 2.46063337e-01
4.55975294e-01 4.17235285e-01 4.49983984e-01 -7.35712171e-01
8.45871568e-01 1.50619674e+00 1.99643433e-01 7.93759823e-01
-1.30236280e+00 -7.53196001e-01 -3.47768851e-02 1.81685552e-01
-1.37563944e+00 -6.00466490e-01 1.08222449e+00 -5.15236080e-01
4.45003480e-01 2.17891514e-01 5.58522463e-01 1.36881733e+00
4.77320045e-01 6.04924738e-01 9.00992572e-01 -3.83208334e-01
3.67311507e-01 -3.00359786e-01 -6.82216048e-01 4.96900171e-01
2.67782807e-01 3.32864583e-01 -4.03043985e-01 -8.03707689e-02
1.25266695e+00 -1.08734004e-01 -3.98815811e-01 -3.65893245e-01
-1.22991300e+00 1.01412380e+00 1.13676459e-01 8.97986218e-02
-5.96484542e-01 1.25073105e-01 1.67286679e-01 3.76829766e-02
5.81917882e-01 1.51788950e-01 2.15815619e-01 1.02641873e-01
-9.90145922e-01 4.71245408e-01 3.87711138e-01 1.05657482e+00
1.76517487e-01 7.53606141e-01 -4.92090374e-01 6.48693681e-01
4.15723890e-01 6.22263014e-01 5.60413957e-01 -1.23128426e+00
3.17962736e-01 1.90394208e-01 2.29469702e-01 -1.25799060e+00
3.10497522e-01 -1.83680221e-01 -1.19788456e+00 4.18071836e-01
5.40404841e-02 -4.33728874e-01 -1.14250147e+00 1.63889611e+00
2.97026783e-01 3.78465056e-01 1.75521255e-01 8.18396389e-01
7.07988203e-01 1.00955427e+00 3.71599980e-02 -6.61893189e-01
8.35669756e-01 -1.02948701e+00 -1.06828654e+00 1.87172696e-01
-2.50398517e-01 -9.32223141e-01 6.85858250e-01 5.29931784e-01
-1.32563984e+00 -8.96570325e-01 -1.10222232e+00 5.04797816e-01
2.11236790e-01 -7.62678161e-02 3.47641200e-01 8.38823915e-01
-1.03960502e+00 4.63862836e-01 -6.69926167e-01 4.72047776e-02
4.72202986e-01 1.36667460e-01 -1.75408483e-01 -1.34049326e-01
-9.68628109e-01 3.72949541e-01 3.31055522e-01 -1.54280618e-01
-1.29543984e+00 -3.00182819e-01 -8.24028313e-01 -6.83699250e-02
2.59464860e-01 -9.42955852e-01 7.68937826e-01 -1.13891268e+00
-1.74745870e+00 3.15314710e-01 4.77499813e-02 -3.27445656e-01
8.20010483e-01 -5.59497550e-02 -5.38135946e-01 3.28171104e-01
-1.31870568e-01 7.85164535e-01 1.38073349e+00 -1.33966660e+00
-3.94530565e-01 7.13770688e-02 2.19026413e-02 1.78327873e-01
-1.24186747e-01 2.90017370e-02 -5.20849407e-01 -1.15987384e+00
-1.22546017e-01 -1.03524220e+00 -2.52021551e-01 -1.81244593e-02
-3.12534988e-01 1.64638102e-01 1.09542227e+00 -8.05990160e-01
1.20853639e+00 -2.14054632e+00 3.65208805e-01 1.03923008e-01
1.99799627e-01 4.43406463e-01 -1.29644588e-01 3.38829726e-01
5.28342165e-02 3.61938149e-01 -5.35921343e-02 -2.17214644e-01
-2.73132265e-01 1.35629594e-01 -3.26556444e-01 3.45895916e-01
1.67248189e-01 7.54355729e-01 -7.46125340e-01 -4.33802783e-01
3.81188482e-01 7.10386038e-01 -6.51554704e-01 4.51789021e-01
-1.77168161e-01 7.11591542e-01 -4.98732835e-01 5.59391081e-01
6.21769547e-01 1.63774826e-02 3.03360708e-02 -8.56953561e-02
2.13565379e-01 -3.12032551e-01 -1.20985138e+00 1.42098117e+00
-1.36256754e-01 6.06047809e-01 -2.90422380e-01 -4.14818883e-01
9.29201007e-01 5.74901879e-01 2.69890964e-01 -3.89128804e-01
1.31534353e-01 -1.54392645e-01 8.75334144e-02 -4.91673797e-01
4.04371560e-01 -2.37410009e-01 2.22592518e-01 1.37095172e-02
-1.06995188e-01 -1.44313648e-01 4.01112527e-01 9.16146934e-02
8.63799274e-01 2.67966568e-01 1.59595385e-01 -2.65702698e-02
5.68186820e-01 -4.72721577e-01 6.51750028e-01 4.44008708e-01
-5.81497848e-02 8.43104720e-01 4.17600960e-01 -1.49784043e-01
-1.56436336e+00 -1.07210231e+00 2.81900853e-01 2.94040918e-01
2.75418311e-01 -9.58822519e-02 -1.00242138e+00 -4.36061978e-01
-3.66491467e-01 6.18718863e-01 -6.42024219e-01 -2.10709035e-01
-5.65976381e-01 -5.96178412e-01 3.84051025e-01 5.07567286e-01
7.86343038e-01 -1.27104592e+00 -2.70755351e-01 1.56412318e-01
-1.94934875e-01 -1.28189218e+00 -6.43981159e-01 -7.76181459e-01
-9.12891269e-01 -8.02993298e-01 -1.09090066e+00 -6.83698952e-01
7.72076428e-01 2.90748537e-01 7.55287945e-01 -7.63119534e-02
4.34266999e-02 5.13125844e-02 -4.81923819e-01 -2.80688237e-02
-8.13160002e-01 -4.24917430e-01 3.86144370e-02 3.46401900e-01
-2.82779813e-01 -5.95332861e-01 -9.07986283e-01 3.27488035e-01
-1.36059856e+00 2.18011141e-01 5.11271536e-01 8.04530859e-01
5.83776832e-01 4.51872468e-01 6.04458809e-01 -6.21129930e-01
8.92987192e-01 -4.23890650e-01 -5.64587951e-01 2.66611506e-03
-4.70824599e-01 -5.71226403e-02 8.07462990e-01 -8.40367913e-01
-1.30721104e+00 -3.07109863e-01 3.68415266e-02 -8.76844645e-01
-1.21373765e-01 8.70743021e-02 -4.83193994e-01 -1.76804245e-01
3.84433895e-01 5.36780477e-01 -8.93402100e-03 -1.15219854e-01
4.35190678e-01 3.72840166e-01 4.53736901e-01 -4.28233176e-01
1.07141042e+00 4.54088151e-01 9.10292193e-02 -8.07059288e-01
-9.72915366e-02 1.59744889e-01 -2.82117456e-01 -4.58543777e-01
1.15137851e+00 -9.15895820e-01 -2.68399715e-01 8.18731189e-01
-1.05027306e+00 -1.16069630e-01 -1.04258761e-01 5.72693348e-01
-6.11047149e-01 5.07154107e-01 -7.90764272e-01 -6.98448658e-01
-3.34835440e-01 -1.42248273e+00 8.02709222e-01 3.73971701e-01
-1.13358058e-01 -8.48156214e-01 -1.63363032e-02 3.79685670e-01
4.62818831e-01 7.73992419e-01 6.88227534e-01 -1.30330175e-01
-9.59002137e-01 -3.58802006e-02 1.41280606e-01 6.83489442e-01
2.70967394e-01 3.50646675e-01 -5.14784694e-01 -4.05934393e-01
9.00218412e-02 8.08822438e-02 5.28432488e-01 6.08493567e-01
9.00804281e-01 -6.16738319e-01 -1.27677873e-01 6.45864248e-01
1.52131438e+00 8.15825105e-01 1.26653290e+00 1.30243808e-01
7.41122305e-01 1.77091956e-01 4.43829566e-01 4.80110973e-01
-2.72347599e-01 6.22544825e-01 3.44687760e-01 9.77022573e-02
-4.63933647e-01 -3.12965870e-01 3.39816809e-01 6.98331773e-01
-3.87380749e-01 -8.50529313e-01 -3.64538759e-01 4.43535835e-01
-1.66683841e+00 -1.45271003e+00 1.01887174e-01 2.05376530e+00
5.55116296e-01 2.87107319e-01 1.64446965e-01 2.86147684e-01
1.01628137e+00 4.10469890e-01 -4.48243648e-01 -1.53810233e-01
-2.28575960e-01 -1.17208008e-02 2.79484600e-01 3.76429349e-01
-9.16549385e-01 9.08410788e-01 6.78826475e+00 1.08246744e+00
-1.10592747e+00 -5.53940870e-02 8.36756766e-01 -1.59799039e-01
-3.21292609e-01 -1.49479926e-01 -4.99337286e-01 8.62648964e-01
6.04773939e-01 -2.37025887e-01 5.01583815e-01 5.99720895e-01
4.06058997e-01 8.57822448e-02 -5.58109939e-01 9.62865710e-01
3.32688510e-01 -1.50422144e+00 4.80928063e-01 1.97946072e-01
1.21694803e+00 -6.51332259e-01 3.28429967e-01 -8.41803551e-02
1.12924777e-01 -9.76659238e-01 9.41572428e-01 7.44552910e-01
7.42595375e-01 -9.12601173e-01 5.08336008e-01 1.81797475e-01
-9.30818141e-01 1.29330620e-01 -1.82727173e-01 1.44534990e-01
4.44917530e-01 2.62235343e-01 -1.17996462e-01 5.07577837e-01
3.90123218e-01 5.33417523e-01 -3.13456178e-01 1.03036296e+00
-4.29726779e-01 7.19924867e-01 1.38894200e-01 4.81256098e-02
1.44170344e-01 -4.23784047e-01 7.37028420e-01 8.45433772e-01
6.50103629e-01 2.63629526e-01 -1.41159505e-01 8.92229259e-01
-8.53626505e-02 -2.08832454e-02 -7.48649418e-01 3.37764546e-02
5.22170961e-01 9.70397234e-01 -6.72281981e-01 -4.21997577e-01
-2.49819383e-01 1.09585965e+00 -3.85840386e-01 5.40712535e-01
-1.22947669e+00 -1.20492302e-01 3.83351505e-01 2.31894195e-01
6.89639986e-01 -3.15244794e-01 1.60471082e-01 -1.12624097e+00
-8.17998312e-03 -1.07154107e+00 -5.20531610e-02 -9.86176431e-01
-1.08110690e+00 7.88490832e-01 6.27260357e-02 -1.43454027e+00
-5.19472480e-01 -1.55872494e-01 -6.34315670e-01 7.49441922e-01
-1.07484448e+00 -1.06676853e+00 -3.20178539e-01 6.43033266e-01
8.57072413e-01 -5.58408499e-01 4.35802937e-01 2.94035017e-01
-5.78420997e-01 4.26063001e-01 2.05734447e-01 5.50267100e-02
5.05738139e-01 -6.50402963e-01 3.96547168e-01 1.29088044e+00
9.06818211e-02 4.00534868e-01 8.93339097e-01 -6.93814039e-01
-1.21657133e+00 -1.12896132e+00 3.96442831e-01 -6.65664151e-02
2.38037288e-01 -7.36901537e-02 -6.50153697e-01 4.54328746e-01
5.92968285e-01 -7.54486546e-02 4.49477702e-01 -8.62703800e-01
9.50200558e-02 6.79812506e-02 -1.38410509e+00 8.69114101e-01
1.02886927e+00 -1.47017881e-01 -3.49969566e-01 -5.47761060e-02
7.47378886e-01 -4.70792025e-01 -8.23697984e-01 4.04209346e-01
3.91603053e-01 -1.13055754e+00 1.05461061e+00 -3.02978277e-01
7.68082201e-01 -4.79612261e-01 -2.34869018e-01 -1.24194455e+00
-6.98805094e-01 -9.00541008e-01 -3.10629755e-01 1.44866085e+00
-1.89752057e-02 -2.68371493e-01 6.63776278e-01 4.17673558e-01
1.28482342e-01 -5.73672950e-01 -6.63580716e-01 -7.98958778e-01
-3.02237719e-01 -2.74458770e-02 5.95726490e-01 8.05225670e-01
-5.07576227e-01 3.65298927e-01 -8.62776816e-01 3.91173884e-02
7.22640395e-01 -1.88381881e-01 8.77163947e-01 -5.77469587e-01
-3.97452921e-01 -3.82426292e-01 -6.47772968e-01 -9.35409844e-01
9.16544627e-03 -1.45791873e-01 -2.41336003e-01 -1.40879571e+00
1.27424449e-01 3.39326374e-02 -5.87505586e-02 -4.74934429e-02
-2.25405529e-01 3.53779376e-01 4.01600897e-01 3.51489216e-01
-2.19163433e-01 7.75753617e-01 1.69215524e+00 -1.27881125e-01
-1.08039781e-01 -9.44213644e-02 -4.73726571e-01 7.11870074e-01
7.89096832e-01 -3.58382791e-01 -8.40949655e-01 -3.35438848e-01
-1.63495660e-01 4.59014595e-01 2.90105790e-01 -1.24962306e+00
-2.16424733e-01 -4.61818367e-01 7.37432659e-01 -3.48883361e-01
4.89787877e-01 -5.40223002e-01 6.49478078e-01 5.00720382e-01
-2.57620126e-01 1.58036843e-01 1.38936207e-01 7.50101924e-01
-3.87022465e-01 -3.50801200e-02 8.75029922e-01 1.88044850e-02
-6.65534079e-01 4.01306361e-01 -3.59976530e-01 -2.64436062e-02
1.34839070e+00 -3.46686184e-01 -3.07761252e-01 -7.58368313e-01
-4.59717900e-01 -4.28617090e-01 6.49628341e-01 6.29409313e-01
9.00486052e-01 -1.67834795e+00 -8.76052976e-01 2.97000557e-01
-3.25322211e-01 -2.59574920e-01 3.57912987e-01 3.97983044e-01
-6.46424532e-01 1.15751453e-01 -4.97268319e-01 -3.28267485e-01
-1.26731658e+00 7.70856023e-01 3.92545685e-02 -1.41502723e-01
-4.73583132e-01 6.06839001e-01 4.11571085e-01 4.03915286e-01
-4.92251478e-02 -1.96213312e-02 -2.04433993e-01 -3.78973126e-01
4.51453626e-01 3.98674965e-01 -4.90819484e-01 -9.08248305e-01
-6.88543096e-02 5.35104036e-01 1.40349537e-01 -2.69303560e-01
1.07947409e+00 -9.32534859e-02 3.22739631e-01 -1.06099896e-01
8.66364598e-01 3.42009902e-01 -1.74252498e+00 1.67395443e-01
-7.17291117e-01 -9.04974639e-01 -2.72790194e-02 -5.35286903e-01
-1.47965670e+00 4.87876832e-01 6.80435181e-01 1.37460366e-01
1.27785838e+00 -3.74561071e-01 8.67013395e-01 -4.76737231e-01
3.69472295e-01 -6.52821779e-01 4.56544966e-01 4.10276987e-02
9.37140167e-01 -9.79858398e-01 4.98410948e-02 -4.03640091e-01
-7.25015700e-01 7.30520010e-01 5.45542181e-01 -4.83645827e-01
3.10249329e-01 1.18537530e-01 -5.52168190e-02 2.42050827e-01
-6.94313228e-01 2.04346731e-01 5.51402926e-01 7.75142789e-01
3.33241582e-01 -1.18608087e-01 -5.40629625e-01 2.18851134e-01
7.07378983e-02 2.79615782e-02 6.34705544e-01 7.02983499e-01
-2.58691281e-01 -1.07565165e+00 -6.34015739e-01 9.74621028e-02
-6.18004024e-01 -7.14118779e-02 -6.25214204e-02 7.47781515e-01
2.46112406e-01 1.07469344e+00 -2.31557712e-01 -5.94242990e-01
8.18269551e-02 -2.78468668e-01 6.67138875e-01 -1.03116587e-01
-1.23915002e-01 4.79367137e-01 -2.33749412e-02 -3.84325773e-01
-7.00612366e-01 -5.31416535e-01 -8.58639598e-01 -4.49688733e-01
-3.11494619e-01 -4.92574312e-02 3.30681950e-01 6.74579144e-01
4.11748260e-01 7.76422083e-01 7.78786421e-01 -9.63045180e-01
-4.86166120e-01 -9.64973629e-01 -5.23430347e-01 6.58029377e-01
7.92859644e-02 -5.72437465e-01 -3.08829695e-01 5.71558475e-01] | [11.000083923339844, -0.5838242769241333] |
aec3bbc1-e929-4ccb-8bd3-9b33b592bf17 | image-to-image-translation-with-multi-path | 1905.12498 | null | https://arxiv.org/abs/1905.12498v1 | https://arxiv.org/pdf/1905.12498v1.pdf | Image-to-Image Translation with Multi-Path Consistency Regularization | Image translation across different domains has attracted much attention in both machine learning and computer vision communities. Taking the translation from source domain $\mathcal{D}_s$ to target domain $\mathcal{D}_t$ as an example, existing algorithms mainly rely on two kinds of loss for training: One is the discrimination loss, which is used to differentiate images generated by the models and natural images; the other is the reconstruction loss, which measures the difference between an original image and the reconstructed version through $\mathcal{D}_s\to\mathcal{D}_t\to\mathcal{D}_s$ translation. In this work, we introduce a new kind of loss, multi-path consistency loss, which evaluates the differences between direct translation $\mathcal{D}_s\to\mathcal{D}_t$ and indirect translation $\mathcal{D}_s\to\mathcal{D}_a\to\mathcal{D}_t$ with $\mathcal{D}_a$ as an auxiliary domain, to regularize training. For multi-domain translation (at least, three) which focuses on building translation models between any two domains, at each training iteration, we randomly select three domains, set them respectively as the source, auxiliary and target domains, build the multi-path consistency loss and optimize the network. For two-domain translation, we need to introduce an additional auxiliary domain and construct the multi-path consistency loss. We conduct various experiments to demonstrate the effectiveness of our proposed methods, including face-to-face translation, paint-to-photo translation, and de-raining/de-noising translation. | ['Yijun Wang', 'Zhibo Chen', 'Jianxin Lin', 'Tao Qin', 'Yingce Xia'] | 2019-05-29 | null | null | null | null | ['face-to-face-translation'] | ['computer-vision'] | [ 2.47636676e-01 -1.55514166e-01 -2.13961340e-02 -4.66292083e-01
-8.17780495e-01 -5.91322184e-01 3.26897860e-01 -4.36778933e-01
-2.39758268e-01 7.61156261e-01 -2.37980694e-01 -3.64169598e-01
-4.42032591e-02 -1.08330512e+00 -8.85903418e-01 -7.35376537e-01
2.91784585e-01 5.01149595e-01 1.99496314e-01 -2.19073623e-01
-8.58218297e-02 3.29265893e-01 -1.27661264e+00 2.49309778e-01
1.08737314e+00 1.26297474e+00 1.35374129e-01 9.41521004e-02
-3.04889262e-01 2.79585481e-01 -7.09526181e-01 -6.70499206e-01
4.62415755e-01 -9.80682611e-01 -5.10021448e-01 8.71321931e-02
4.84030128e-01 -2.66481727e-01 -1.76342592e-01 1.38512385e+00
4.81296480e-01 1.06522664e-02 8.32784474e-01 -1.19223738e+00
-1.06045961e+00 -3.64349666e-03 -1.00162554e+00 1.27167790e-03
3.42685312e-01 4.55450803e-01 4.76894349e-01 -8.76604080e-01
6.87905848e-01 1.52351308e+00 4.40332204e-01 5.09946108e-01
-1.20574427e+00 -1.28849709e+00 2.05936700e-01 -8.24778453e-02
-1.30657339e+00 -2.07648292e-01 6.86719716e-01 -5.39887607e-01
2.93669045e-01 7.70414062e-03 1.98709011e-01 8.39528382e-01
1.38571084e-01 5.67807376e-01 1.58148301e+00 -4.25215364e-01
-1.15541838e-01 7.32506588e-02 -2.48765618e-01 7.93951452e-01
-1.14826530e-01 2.18320191e-01 -2.47103095e-01 1.59778327e-01
1.19028306e+00 -1.27983004e-01 -2.75105804e-01 8.80945195e-03
-1.00790846e+00 7.71803737e-01 4.02077526e-01 1.66328609e-01
-2.29013152e-02 -1.42913818e-01 1.96510553e-01 5.66907465e-01
4.46560562e-01 -4.18454558e-02 -4.02356654e-01 2.07666993e-01
-8.54688346e-01 2.21372932e-01 2.77686030e-01 1.28240764e+00
1.15203297e+00 8.53200927e-02 -1.79049790e-01 1.02692652e+00
1.84242606e-01 8.91050875e-01 2.63616383e-01 -1.07555723e+00
8.40128779e-01 7.30210483e-01 1.70804322e-01 -7.63728321e-01
3.39151099e-02 -1.89245313e-01 -1.17081809e+00 3.28181982e-01
3.62697005e-01 -2.17353165e-01 -1.02550125e+00 2.04307437e+00
3.82995933e-01 -8.74897349e-04 -1.02970414e-01 1.06305671e+00
8.87821376e-01 7.93853343e-01 -1.19006597e-01 -5.96453726e-01
1.33610260e+00 -9.86168623e-01 -3.34507406e-01 -2.36043632e-01
3.33884776e-01 -1.28732991e+00 1.57110274e+00 1.65354028e-01
-1.38176978e+00 -7.34481871e-01 -8.64747286e-01 -5.29492125e-02
-1.48294091e-01 1.48376703e-01 5.65842912e-02 4.75127220e-01
-7.49067724e-01 4.37492698e-01 -2.85137773e-01 -9.33755487e-02
4.55309451e-01 3.47156167e-01 -4.59154993e-01 -5.71831048e-01
-1.30038667e+00 7.19135106e-01 2.18015701e-01 -1.82698831e-01
-8.43107581e-01 -5.36766827e-01 -7.30866611e-01 -1.55890077e-01
3.17479461e-01 -6.10747039e-01 8.42956543e-01 -1.26374233e+00
-1.26864052e+00 1.33244836e+00 -1.60457090e-01 8.21677074e-02
6.05335832e-01 2.87350476e-01 -6.83357716e-01 1.82879865e-02
6.23163044e-01 8.62461150e-01 8.58418584e-01 -1.27081418e+00
-5.42879164e-01 -6.88925445e-01 8.81630555e-02 2.82526612e-01
-1.57671735e-01 2.52183735e-01 -7.03211010e-01 -8.37393224e-01
3.06287915e-01 -9.41941559e-01 3.47383618e-01 1.46554634e-01
-3.45754683e-01 -2.58009341e-02 9.48801339e-01 -8.85657132e-01
1.07942080e+00 -2.23424649e+00 1.47081688e-01 1.77589715e-01
7.82457739e-02 4.33923036e-01 -2.30220214e-01 1.52492478e-01
-1.29738584e-01 3.76822539e-02 -5.50501108e-01 -2.84506112e-01
-1.94796994e-01 2.52105564e-01 -2.17035607e-01 1.56479374e-01
5.02298996e-02 5.48952997e-01 -6.02976203e-01 -4.89527255e-01
2.92921085e-02 4.36683297e-01 -2.28060126e-01 2.61146516e-01
-2.84394026e-01 8.19672287e-01 -5.80652416e-01 5.14370739e-01
1.24702334e+00 -1.15296789e-01 -1.10640228e-01 -1.44917607e-01
-3.06220422e-03 1.18405364e-01 -1.20932114e+00 1.62555289e+00
-2.23234668e-01 1.41141517e-02 3.26096863e-01 -1.08763039e+00
1.21298003e+00 1.57920256e-01 4.67138290e-01 -1.22448051e+00
1.45841643e-01 3.42008948e-01 -5.10841757e-02 -2.33742297e-01
-6.98155910e-02 -6.54977918e-01 7.02677369e-02 6.26169741e-01
-2.15808779e-01 -4.40966010e-01 2.14527547e-01 -1.61792129e-01
7.50373065e-01 3.60079646e-01 -3.55064958e-01 -1.75254479e-01
9.02157545e-01 -6.88771298e-03 6.95045173e-01 2.25225329e-01
-1.37331843e-01 7.89730489e-01 6.98957324e-01 -2.68948168e-01
-1.17412746e+00 -1.26584971e+00 -4.31950018e-02 1.02079713e+00
5.67705929e-01 1.61254674e-01 -1.04095972e+00 -7.57330060e-01
-2.96045482e-01 5.71818531e-01 -3.79005551e-01 -4.86206651e-01
-6.59747839e-01 -4.71086383e-01 6.07957602e-01 3.76973093e-01
1.17609286e+00 -1.12433672e+00 2.10764054e-02 -1.39971003e-01
-3.56243581e-01 -8.87132764e-01 -9.56314027e-01 -2.39834249e-01
-7.68796444e-01 -9.29286242e-01 -9.66976881e-01 -9.47185516e-01
9.70731258e-01 1.30635023e-01 1.01931524e+00 1.60628244e-01
1.89258382e-02 6.15409128e-02 -2.95036435e-01 -3.39381188e-01
-3.70590001e-01 -2.04034030e-01 -6.92441612e-02 2.00071791e-03
2.74298072e-01 -7.74319708e-01 -7.96314895e-01 8.97129238e-01
-1.13752675e+00 1.22422330e-01 6.55343592e-01 7.65884697e-01
8.36662412e-01 6.36839643e-02 2.51139581e-01 -6.32805347e-01
5.75765312e-01 -1.72562033e-01 -6.57778561e-01 1.10154249e-01
-6.55957997e-01 -7.91477039e-02 8.66724610e-01 -5.96422315e-01
-9.23320949e-01 -2.54305750e-01 -9.44312010e-03 -8.96284878e-01
-8.24558809e-02 2.58593410e-01 -5.61959088e-01 6.13449328e-02
5.44989049e-01 7.47791290e-01 1.03386633e-01 -5.00509202e-01
3.07022572e-01 4.39061910e-01 6.41806126e-01 -7.84719408e-01
1.07426298e+00 4.28390533e-01 -1.10166959e-01 -3.31891567e-01
-6.23240173e-01 1.98525697e-01 -3.80775392e-01 1.71579067e-02
9.41020489e-01 -7.72598922e-01 -5.56493938e-01 5.56014776e-01
-1.06109691e+00 -2.15037569e-01 -1.56858131e-01 4.24821705e-01
-4.83847231e-01 3.37548077e-01 -3.31068695e-01 -3.92930478e-01
-3.52545142e-01 -1.44449317e+00 1.12669373e+00 2.96226948e-01
1.34162396e-01 -7.57246315e-01 -3.05265576e-01 5.56420147e-01
2.25974590e-01 1.78253502e-01 1.26046503e+00 -9.07698348e-02
-6.64558709e-01 -1.08704060e-01 -7.92161763e-01 8.79302323e-01
2.44840100e-01 -2.25323796e-01 -4.58329499e-01 -4.70370650e-01
1.98973328e-01 -3.33695531e-01 6.14178121e-01 2.38683358e-01
1.07424486e+00 -4.34012622e-01 -2.32206285e-01 6.17483079e-01
1.24501383e+00 5.29156625e-01 9.98734236e-01 2.00912148e-01
3.79410118e-01 4.05968338e-01 6.29808426e-01 1.15936175e-01
3.94333363e-01 6.86198592e-01 2.95313895e-01 -2.61899561e-01
-3.28502834e-01 -2.75222868e-01 4.03438717e-01 4.96995211e-01
2.04594042e-02 -6.37392551e-02 -5.61400473e-01 3.26874584e-01
-1.33377111e+00 -6.94139957e-01 5.69630191e-02 2.43825388e+00
1.03468966e+00 1.58173904e-01 -5.94240502e-02 -2.79064745e-01
1.10015166e+00 2.03649104e-01 -6.92125261e-01 -2.90008575e-01
-2.17864826e-01 6.18515670e-01 1.04135975e-01 3.19077581e-01
-7.15115905e-01 9.71447706e-01 4.01295471e+00 1.26256514e+00
-1.41472352e+00 1.77751988e-01 7.97827601e-01 -2.66609411e-03
-4.90105093e-01 2.57282704e-01 -6.23722136e-01 1.06793344e+00
2.58093208e-01 1.27018303e-01 5.10124862e-01 6.92861676e-01
2.19482809e-01 -1.14407964e-01 -1.12840104e+00 1.02534783e+00
6.20145984e-02 -1.06385911e+00 1.01112559e-01 -4.53642420e-02
8.84695530e-01 -5.13635576e-01 4.42018867e-01 3.56948674e-01
3.95440787e-01 -1.16338527e+00 7.62294412e-01 1.03948526e-02
1.47945487e+00 -6.12235129e-01 3.10742974e-01 5.64602733e-01
-1.40379393e+00 2.14998007e-01 -4.07718867e-01 1.21157303e-01
-5.20091541e-02 6.21288121e-01 -3.04769892e-02 7.21784830e-01
7.44253099e-01 3.93405080e-01 -1.11761913e-01 4.53268349e-01
-3.01796287e-01 2.18246937e-01 -2.14569882e-01 4.76385206e-01
2.15219572e-01 -7.21425474e-01 3.47550839e-01 7.30028272e-01
6.09365404e-01 2.65697449e-01 3.84527415e-01 1.18439233e+00
-4.51686412e-01 -2.26819552e-02 -4.25014108e-01 3.06373477e-01
6.07654333e-01 8.36051345e-01 -4.94895190e-01 -2.57781595e-01
-3.64912689e-01 1.15758681e+00 -2.14718562e-02 4.49199438e-01
-1.20189404e+00 -5.75889289e-01 6.73738599e-01 3.92009914e-01
1.45173833e-01 -8.12377632e-02 -2.61892676e-01 -1.05533838e+00
4.62503046e-01 -9.48228359e-01 2.32964754e-01 -9.47328568e-01
-1.30068779e+00 5.87254167e-01 5.31541891e-02 -1.49195135e+00
1.99958384e-02 -3.62043023e-01 -6.07193470e-01 1.42140603e+00
-1.47574294e+00 -1.27696002e+00 -3.81115645e-01 9.81365979e-01
1.83337465e-01 -4.11351562e-01 5.56906819e-01 7.00464010e-01
-3.25662076e-01 8.95809412e-01 2.61221200e-01 3.76942873e-01
1.01424408e+00 -5.25530517e-01 5.60966991e-02 6.47318244e-01
-3.36018115e-01 4.22048867e-01 3.90904307e-01 -4.31250483e-01
-1.10508120e+00 -1.16205418e+00 6.17291510e-01 -1.41840622e-01
2.09404320e-01 -1.48781538e-01 -9.34811890e-01 9.41049755e-01
1.85141131e-01 -8.12933128e-03 2.57631958e-01 -5.73327720e-01
-4.90074128e-01 -5.05141854e-01 -1.56983066e+00 6.16952062e-01
1.00906754e+00 -3.95543367e-01 -3.06360155e-01 3.27424258e-01
7.06411600e-01 -6.58386469e-01 -9.06654835e-01 5.04558623e-01
2.14029089e-01 -1.16944253e+00 9.07927334e-01 -2.48130217e-01
8.57698143e-01 -3.90131921e-01 -9.63791460e-02 -1.05316424e+00
-7.08310902e-02 -4.27572072e-01 5.39318204e-01 1.47799456e+00
2.52566308e-01 -9.19470608e-01 8.65810990e-01 4.23126549e-01
-2.29970589e-01 -6.71149731e-01 -1.16730881e+00 -8.88925672e-01
8.13788176e-01 -8.38058069e-02 7.84709871e-01 1.11623371e+00
-7.88747728e-01 2.96456605e-01 -3.82107466e-01 -4.62045483e-02
4.51654851e-01 2.99053341e-01 8.58021975e-01 -9.16443288e-01
-9.76129845e-02 -4.47634995e-01 8.12220126e-02 -1.40388846e+00
1.28241211e-01 -6.50874138e-01 -2.37652645e-01 -1.58699715e+00
1.25544295e-01 -7.74545610e-01 -3.87124792e-02 5.05738020e-01
1.21094309e-01 3.32832277e-01 9.50176567e-02 4.20644641e-01
-5.08265011e-02 5.87942958e-01 1.98286402e+00 -3.16584110e-01
8.58614817e-02 -1.46657392e-01 -7.94615865e-01 6.75347030e-01
3.87487173e-01 -4.46783930e-01 -6.47608995e-01 -8.27967644e-01
-1.43502966e-01 4.59500313e-01 2.97275901e-01 -8.19840312e-01
-7.15492219e-02 -6.25824153e-01 4.42175955e-01 -4.27666575e-01
4.83323723e-01 -6.84440136e-01 2.36781493e-01 2.59810388e-01
-1.03006050e-01 6.88479915e-02 1.24777360e-02 1.19349651e-01
-4.31083113e-01 2.08169427e-02 1.36882257e+00 -3.50797564e-01
-3.57020169e-01 5.83483696e-01 2.62651294e-01 3.07741016e-01
1.17592883e+00 -4.64322120e-01 -3.46178710e-01 -3.28672528e-01
-6.16336703e-01 2.99142629e-01 6.92901373e-01 3.64124835e-01
5.69060862e-01 -1.76980650e+00 -9.13946152e-01 3.67720574e-01
1.20368926e-02 4.88812774e-01 3.94265443e-01 8.79216909e-01
-5.25853038e-01 9.35668424e-02 -3.49502981e-01 -5.08944750e-01
-1.12076187e+00 4.56267297e-01 4.89410818e-01 -3.26171756e-01
-1.66194946e-01 1.20253003e+00 5.21017909e-01 -7.19448030e-01
-5.62579408e-02 -5.46319261e-02 3.62200797e-01 -2.07321718e-01
2.32093986e-02 4.01395351e-01 -1.02129877e-01 -8.34691465e-01
-3.01018029e-01 9.24880922e-01 -4.83461283e-02 -2.50527084e-01
8.09781134e-01 -2.20122114e-01 -5.33760548e-01 -7.80707225e-02
1.44659507e+00 -1.95857778e-01 -1.16572154e+00 -4.02414501e-01
-7.07691669e-01 -7.31306493e-01 -5.41857123e-01 -1.03949714e+00
-1.29298759e+00 1.24765897e+00 9.76016402e-01 -2.07717210e-01
1.46288991e+00 -3.81260179e-02 1.09987819e+00 -2.13522479e-01
4.49276865e-01 -1.18993819e+00 5.26925147e-01 5.70627153e-01
8.96547973e-01 -1.18983245e+00 -1.28715649e-01 -2.96969682e-01
-6.59150600e-01 7.83541381e-01 8.20979118e-01 -8.51489156e-02
3.60626042e-01 -1.08798876e-01 5.38333803e-02 2.91118510e-02
-8.17025676e-02 7.80391544e-02 2.36521482e-01 5.85596979e-01
3.10750335e-01 -1.63058061e-02 -3.95286828e-01 4.82667804e-01
-1.12268679e-01 9.76031348e-02 1.35426685e-01 7.56112039e-01
-3.40654820e-01 -1.60220778e+00 -6.14084303e-01 3.64553571e-01
-1.37039125e-01 -9.35565382e-02 -3.19976985e-01 7.59452581e-01
7.31481135e-01 7.86576033e-01 1.04444835e-03 -2.47882411e-01
4.84183341e-01 1.63919270e-01 5.95891893e-01 -5.51281571e-01
-3.06911170e-01 3.36692184e-01 -4.28707302e-01 -2.36026198e-01
-2.01282814e-01 -4.88056213e-01 -1.50490987e+00 -6.39543295e-01
-8.31258148e-02 6.09723032e-02 2.94499665e-01 8.87938857e-01
3.67507368e-01 1.92008644e-01 8.76721025e-01 -3.56508881e-01
-5.04121780e-01 -8.72680604e-01 -6.40038133e-01 7.44898081e-01
-5.28430715e-02 -7.45932639e-01 -7.24025071e-02 1.68894991e-01] | [11.754684448242188, -0.3893105387687683] |
ea9e01eb-d2ca-4d64-9127-c1667bb94262 | performance-analysis-of-empirical-open | 2306.16575 | null | https://arxiv.org/abs/2306.16575v1 | https://arxiv.org/pdf/2306.16575v1.pdf | Performance Analysis of Empirical Open-Circuit Voltage Modeling in Lithium Ion Batteries, Part-3: Experimental Results | This paper is the third part of a series of papers about empirical approaches to open circuit voltage (OCV) modeling of lithium-ion batteries. The first part of the series proposed models to quantify various sources of uncertainties in the OCV models; and, the second part of the series presented systematic data collection approaches to compute the uncertainties in the OCV-SOC models. This paper uses data collected from 28 OCV characterization experiments, performed according to the data collection plan presented, to compute and analyze the following three different OCV uncertainty metrics: cell-to-cell variations, cycle-rate error, and curve fitting error. From the computed metrics, it was observed that a lower C-Rate showed smaller errors in the OCV-SOC model and vice versa. The results reported in this paper establish a relationship between the C-Rate and the uncertainty of the OCV-SOC model. This research can be thus useful to battery researchers for quantifying the tradeoff between the time taken to complete the OCV characterization test and the corresponding uncertainty in the OCV-SOC modeling. Further, quantified uncertainty model parameters can be used to accurately characterize the uncertainty in various battery management functionalities, such as state of charge and state of health estimation. | ['Balakumar Balasingam', 'James Nguyen', 'Prarthana Pillai'] | 2023-06-28 | null | null | null | null | ['management'] | ['miscellaneous'] | [-4.90641952e-01 -6.74704611e-01 -3.11124951e-01 -2.67703116e-01
-4.92140472e-01 -5.82213640e-01 5.81600845e-01 8.91000271e-01
-2.57839411e-01 1.28055418e+00 -2.64629334e-01 -5.74977815e-01
-3.61199200e-01 -8.62986743e-01 -7.07701504e-01 -6.32294536e-01
2.07114935e-01 5.25082529e-01 4.37988639e-01 -5.18975891e-02
5.47381520e-01 8.53846252e-01 -1.72194147e+00 -3.65160197e-01
1.03250706e+00 1.41784704e+00 -1.79839581e-02 3.87652844e-01
-2.22497992e-02 2.11289376e-01 -5.32665193e-01 -3.53253596e-02
-4.51260835e-01 -1.59614146e-01 -1.66825399e-01 -6.91609502e-01
-5.86780190e-01 -1.84180155e-01 -1.62024826e-01 9.22356367e-01
5.02330005e-01 -2.14664906e-01 9.70609009e-01 -1.75808609e+00
1.77978486e-01 6.58822298e-01 1.92926645e-01 3.84188354e-01
2.93726236e-01 -3.36440429e-02 3.80791515e-01 -6.57779574e-01
4.13834937e-02 8.89085174e-01 3.99580747e-01 1.14083029e-01
-1.10196304e+00 -8.55361044e-01 -4.46819633e-01 2.71331847e-01
-1.70690978e+00 -1.97628275e-01 4.96142954e-01 -4.55170155e-01
1.10149622e+00 5.54686248e-01 1.05255270e+00 3.17791253e-01
9.95427370e-01 1.05416961e-01 1.31995094e+00 -3.84014040e-01
7.55036175e-01 5.94661355e-01 6.53565586e-01 -3.69140208e-01
1.27438760e+00 2.02839151e-01 -1.27928555e-01 -2.71912307e-01
7.36666471e-02 -4.01638865e-01 -2.61121690e-01 -1.82229459e-01
-1.04295544e-01 3.89126033e-01 3.97404544e-02 4.76903647e-01
1.86087847e-01 4.51138854e-01 5.89155555e-01 -2.51906663e-01
-1.07141994e-01 1.86952334e-02 -1.87664136e-01 -3.98498416e-01
-7.37442374e-01 3.95777881e-01 1.01220191e+00 1.15880394e+00
4.89849120e-01 1.60686955e-01 -2.75712907e-01 1.75292403e-01
7.94362366e-01 8.04232001e-01 3.47879410e-01 -5.10869205e-01
2.33768806e-01 4.47108835e-01 5.77392101e-01 -1.05887935e-01
-3.34739089e-01 -1.19685531e-01 -2.13398442e-01 -1.83542565e-01
8.06395486e-02 -9.46134701e-02 -8.18589807e-01 8.47104907e-01
-1.86871573e-01 -4.43118483e-01 1.43182874e-01 3.21291566e-01
8.17054808e-01 6.16978526e-01 2.55192906e-01 -4.45593774e-01
1.11610639e+00 -5.65662421e-02 -1.19264042e+00 1.80823848e-01
3.66593003e-01 -2.87591457e-01 5.00568390e-01 2.79190834e-03
-1.32906282e+00 -1.29475236e-01 -1.67649555e+00 3.69245768e-01
-6.58899486e-01 -1.31433293e-01 1.16266981e-01 1.11031282e+00
-7.90062726e-01 6.66648030e-01 -6.50161624e-01 -2.43349493e-01
-2.14599296e-02 3.53070974e-01 3.36612791e-01 1.29176706e-01
-1.33355296e+00 1.27478743e+00 3.94527256e-01 1.93605661e-01
-9.78910267e-01 -8.10270548e-01 -7.09226787e-01 6.73293769e-02
-7.61899212e-03 1.38322935e-01 1.04421639e+00 1.26893967e-01
-9.52365935e-01 2.55899757e-01 -3.39073271e-01 -5.25585234e-01
5.10576189e-01 8.75108540e-02 -6.50706172e-01 -3.12365055e-01
-2.56564945e-01 -6.27512932e-02 7.29078650e-02 -1.60679924e+00
-2.88999796e-01 -3.52665722e-01 -4.69869763e-01 -2.75856107e-01
8.35596472e-02 -2.33304068e-01 -2.86656886e-01 -8.37283954e-02
-2.77503766e-02 -9.27995682e-01 1.38330206e-01 -5.22950470e-01
-1.20011270e-02 -3.86937886e-01 6.59982085e-01 -4.04878706e-01
1.51226926e+00 -1.76966798e+00 -3.31029564e-01 5.51217079e-01
-5.67087829e-01 5.60708642e-02 7.79120386e-01 6.59414828e-01
2.03106895e-01 4.47962254e-01 -2.61796176e-01 4.27963994e-02
1.52426571e-01 3.20086390e-01 -1.48477048e-01 5.86435735e-01
4.52532573e-03 8.07616532e-01 -6.89417362e-01 -6.17637150e-02
1.90013558e-01 2.32132003e-01 1.89247012e-01 6.52325945e-03
-6.88743591e-02 -9.56198946e-02 -1.88812777e-01 7.44998455e-01
8.91431093e-01 5.41045427e-01 2.60312706e-01 -4.42779481e-01
-5.52937210e-01 1.65412933e-01 -9.63021755e-01 6.53979838e-01
-3.34189028e-01 5.37402630e-01 -1.40561998e-01 -5.57090044e-01
1.16383410e+00 1.69581383e-01 2.93003768e-01 -9.20217395e-01
6.81406915e-01 8.47821951e-01 -1.38369411e-01 -2.91620702e-01
7.58597374e-01 -2.26744160e-01 -6.70800880e-02 -6.35730475e-02
-5.26741184e-02 -7.41747022e-01 4.09396380e-01 6.51660562e-02
2.47217387e-01 -1.09360360e-01 -1.32036299e-01 -1.17698514e+00
7.58486927e-01 -2.10657716e-01 2.94988632e-01 2.44353831e-01
-2.36600235e-01 6.71788305e-02 7.77152061e-01 3.69763553e-01
-1.15701807e+00 -9.19815481e-01 -9.35278714e-01 1.41224666e-02
7.22999096e-01 -2.06260741e-01 -6.10556543e-01 1.45415768e-01
6.46851957e-01 1.27715421e+00 -2.94100910e-01 -4.86375123e-01
-5.69147952e-02 -8.17300797e-01 5.28001726e-01 1.04070842e+00
2.19616324e-01 -2.51858681e-01 -3.76832157e-01 3.29260454e-02
1.52889237e-01 -1.04525232e+00 9.35721993e-02 6.42883658e-01
-1.12946117e+00 -1.02800989e+00 -2.68411696e-01 -1.73608199e-01
5.59080601e-01 -1.97118491e-01 9.53835249e-01 1.65684626e-01
-1.55774564e-01 5.10287821e-01 -9.63687822e-02 -1.10555768e+00
-5.35221815e-01 -3.99573565e-01 2.03207985e-01 -5.27483821e-01
5.71864367e-01 4.73920070e-02 -4.24340069e-01 6.47822142e-01
-8.87086451e-01 -9.74540353e-01 -9.14437249e-02 9.98320654e-02
8.45210075e-01 3.12733352e-01 8.89144957e-01 -1.18845731e-01
8.36020529e-01 -5.74339688e-01 -1.03156090e+00 2.50270694e-01
-1.44076896e+00 1.79671839e-01 1.55734271e-01 -4.15713247e-03
-8.60369086e-01 -4.80922639e-01 -2.26877078e-01 -1.27123520e-01
2.84975111e-01 3.58809203e-01 -5.44870377e-01 -2.47590914e-01
1.00140519e-01 1.57576233e-01 4.68274541e-02 -3.31747949e-01
-4.44907546e-01 9.17185366e-01 5.25207877e-01 -6.18338048e-01
4.03988689e-01 2.54969925e-01 3.78269732e-01 -6.84266269e-01
1.51353493e-01 -3.93817335e-01 -2.47908667e-01 -5.53987086e-01
5.68425596e-01 -1.16867232e+00 -9.81206298e-01 6.98743343e-01
-5.23895681e-01 -2.96514928e-01 -4.84367721e-02 4.39454466e-01
-1.29763231e-01 2.25945562e-01 -2.20169023e-01 -1.46965373e+00
-5.47192216e-01 -1.48512661e+00 5.41680276e-01 5.34153461e-01
-1.31308123e-01 -1.02149975e+00 -2.20076442e-01 1.30560948e-02
4.80215877e-01 6.82311580e-02 9.46074486e-01 -2.80183911e-01
-3.56295705e-01 -5.28542519e-01 -8.63407254e-02 4.86309618e-01
-2.69718558e-01 2.25796968e-01 -8.70913863e-01 -6.93709671e-01
4.80055958e-02 -4.45172265e-02 4.56044823e-01 3.98622066e-01
1.08820426e+00 5.11961699e-01 -5.45257449e-01 1.03809349e-02
2.11866665e+00 7.93083966e-01 1.28828514e+00 2.87091017e-01
4.44017313e-02 1.05324112e-01 8.96862805e-01 5.85796535e-01
3.55223686e-01 5.11210203e-01 6.62792087e-01 6.30618155e-01
3.40735495e-01 -2.77351350e-01 3.07946146e-01 5.34012079e-01
1.72410905e-02 -5.72778463e-01 -8.03992450e-01 6.34402514e-01
-1.26207960e+00 -3.12992811e-01 -8.14378262e-01 2.72223711e+00
4.03416693e-01 4.19927448e-01 5.06956652e-02 6.88201249e-01
5.81076562e-01 -4.60102260e-01 -4.39710140e-01 -7.08630323e-01
-7.68652782e-02 1.95969287e-02 1.16859257e+00 4.30691361e-01
-2.10211605e-01 -3.92061733e-02 7.14919472e+00 7.09806859e-01
-1.07180941e+00 -8.01005960e-02 4.12625611e-01 -2.21465919e-02
-9.44464684e-01 3.04211915e-01 -1.20408690e+00 9.07662392e-01
1.59057295e+00 -6.33048832e-01 3.12703215e-02 3.32295775e-01
3.58336926e-01 -1.07500517e+00 -1.13476002e+00 7.67930090e-01
-9.37215984e-02 -1.05701363e+00 -8.44952390e-02 9.29126069e-02
6.01268947e-01 -2.57614315e-01 -5.55611610e-01 2.44286463e-01
-4.67154205e-01 -8.23063195e-01 1.12314939e+00 9.60037291e-01
1.14366543e+00 -9.99353230e-01 1.28983951e+00 7.40935951e-02
-1.25000322e+00 -3.26440036e-01 -2.96784937e-01 9.45112258e-02
2.99318343e-01 6.35719657e-01 -4.73070204e-01 6.77717626e-01
8.79073083e-01 -1.70128629e-01 -5.70730567e-01 1.28068006e+00
1.92887902e-01 4.83489901e-01 -4.57070172e-01 -8.61163497e-01
-2.19762310e-01 -4.19647455e-01 1.51623815e-01 8.97040606e-01
5.80056667e-01 -5.88775938e-03 -6.69692576e-01 9.15671051e-01
5.15647642e-02 -1.51716337e-01 -2.84236789e-01 -2.15480030e-01
9.67105269e-01 8.23587954e-01 -5.52211821e-01 -1.94286972e-01
-9.36006904e-02 3.62904593e-02 -5.62966168e-01 3.09924394e-01
-1.03343725e+00 -5.54823518e-01 3.74827325e-01 7.79148281e-01
-2.16516014e-02 -1.52616113e-01 -8.86929512e-01 -4.29583520e-01
2.24161837e-02 -4.77621034e-02 -1.88836418e-02 -9.43636954e-01
-6.52365804e-01 -1.07003450e-01 8.86375070e-01 -8.21488321e-01
1.21456079e-01 -7.31627822e-01 -8.14307034e-01 1.10821950e+00
-1.25189126e+00 -3.03421199e-01 -4.79910105e-01 4.99347039e-02
3.87014598e-02 1.06807925e-01 4.49216306e-01 4.04768467e-01
-5.41624367e-01 5.92099190e-01 8.41173351e-01 -6.11472130e-01
2.32175738e-01 -1.00519788e+00 -1.87484100e-01 5.09776175e-01
-1.11794567e+00 2.27916330e-01 1.07752669e+00 -1.08170342e+00
-1.82173181e+00 -6.36378646e-01 6.26654029e-01 -2.93080330e-01
7.14886487e-01 -2.43732616e-01 -7.68079519e-01 2.04812810e-01
1.60060823e-02 -2.67523319e-01 3.03351253e-01 -6.29633188e-01
5.41589677e-01 -6.21888697e-01 -1.35213864e+00 1.99012682e-01
2.12758869e-01 -5.88495791e-01 -9.11114067e-02 -2.92920828e-01
4.13445622e-01 -3.00039500e-01 -1.45354044e+00 7.11831927e-01
8.02787662e-01 -5.67436039e-01 5.37050307e-01 -4.52728607e-02
-2.82572806e-01 -3.91677529e-01 -3.52490097e-01 -9.59573865e-01
1.56949431e-01 -4.45123948e-02 -1.39005184e-01 1.59457898e+00
6.25773966e-01 -7.35428751e-01 2.74936616e-01 1.04607999e+00
-2.17279017e-01 -8.72624218e-01 -1.23509681e+00 -9.31951344e-01
4.19302523e-01 -7.73709476e-01 8.24393213e-01 -1.52170405e-01
4.42259572e-02 -2.55567789e-01 3.29740882e-01 6.61183670e-02
5.48933744e-01 -6.63861156e-01 3.10536414e-01 -1.30306268e+00
4.65699047e-01 -9.21984911e-02 -4.46348339e-01 -2.30465014e-03
-2.56907493e-01 -4.56219465e-01 6.90209419e-02 -1.78991616e+00
2.37942308e-01 -7.35235929e-01 -4.79237020e-01 -3.27032298e-01
6.70775995e-02 1.74993295e-02 2.48420045e-01 8.33833367e-02
1.26499906e-01 5.88016391e-01 6.61094069e-01 -2.19564930e-01
-2.67179579e-01 1.94356151e-04 -1.50273830e-01 2.62980998e-01
6.45222187e-01 -3.93108696e-01 -5.79747558e-01 2.39749238e-01
4.56289381e-01 1.18766971e-01 3.96224409e-01 -1.33506978e+00
6.09097034e-02 -2.99930684e-02 4.35695976e-01 -1.29269981e+00
9.00790915e-02 -1.16804898e+00 8.98091555e-01 6.54202163e-01
4.35853571e-01 1.81491123e-04 7.50579059e-01 5.64284027e-01
1.89754650e-01 -9.90146637e-01 7.02795386e-01 3.46912086e-01
-4.99722719e-01 -2.28807017e-01 -6.37249827e-01 -2.63008535e-01
1.53010809e+00 -3.80716741e-01 -4.11675900e-01 -2.50221461e-01
-4.76268917e-01 5.60291708e-01 6.50231540e-01 1.98093116e-01
4.23829556e-01 -1.41149747e+00 -1.32648930e-01 -1.30009465e-02
2.32386082e-01 -2.11476088e-01 2.02397943e-01 8.40209544e-01
-7.06015527e-01 7.63674617e-01 -1.93041377e-02 -6.52092516e-01
-1.03555453e+00 4.75181907e-01 6.97161496e-01 -3.51446867e-02
2.79651701e-01 2.10546598e-01 -7.30857253e-01 2.53087342e-01
1.66040286e-01 -5.25547326e-01 -1.41164362e-01 3.99089992e-01
2.55971581e-01 1.07396293e+00 5.87749898e-01 -6.06532097e-01
-7.07927048e-01 4.73990440e-01 5.69649458e-01 -3.12693954e-01
6.98925734e-01 -2.58160353e-01 -3.15852404e-01 1.09409928e+00
1.08067942e+00 -1.63426116e-01 -8.33560288e-01 4.75200266e-01
8.65314305e-02 -1.51839778e-01 2.64338613e-01 -8.19950461e-01
-5.68507910e-01 5.01225173e-01 9.75756526e-01 2.28829592e-01
9.79970038e-01 -1.80025786e-01 3.02684933e-01 -2.21057795e-02
7.56948352e-01 -1.59237242e+00 -5.26907146e-01 3.19481403e-01
9.65031028e-01 -7.60179996e-01 5.51980257e-01 -2.44242743e-01
-3.96981657e-01 1.04464591e+00 5.91732681e-01 2.67057061e-01
1.04560733e+00 7.21668959e-01 -4.98192847e-01 1.65746421e-01
-3.95515800e-01 3.05378616e-01 6.13352694e-02 2.79146016e-01
3.56660396e-01 2.90545493e-01 -1.17900932e+00 9.96888697e-01
2.44833857e-01 2.10479230e-01 7.89526165e-01 1.15288234e+00
-5.45759320e-01 -1.00692475e+00 -6.43871248e-01 6.63913727e-01
-2.09702402e-01 4.28600997e-01 -7.79902115e-02 8.44991446e-01
-9.90812085e-04 1.05675709e+00 3.47265899e-01 -2.62172759e-01
8.26909184e-01 1.48521364e-01 4.64473307e-01 2.66055584e-01
-2.52468169e-01 -1.91816062e-01 4.55012918e-01 -1.17326140e-01
-6.97371215e-02 -5.47137499e-01 -1.72559285e+00 -3.69922131e-01
-8.54574442e-01 6.52581036e-01 1.46504319e+00 8.45052123e-01
-9.27348286e-02 8.08374763e-01 5.88240564e-01 -7.51798868e-01
-5.50150812e-01 -1.03835118e+00 -9.86868262e-01 -5.64135388e-02
-9.42149758e-02 -9.21657741e-01 -6.81479633e-01 -7.39763260e-01] | [6.326366424560547, 2.764009952545166] |
f2384093-09da-42ab-a588-314405185580 | contextualized-spatio-temporal-contrastive | 2112.05181 | null | https://arxiv.org/abs/2112.05181v2 | https://arxiv.org/pdf/2112.05181v2.pdf | Contextualized Spatio-Temporal Contrastive Learning with Self-Supervision | Modern self-supervised learning algorithms typically enforce persistency of instance representations across views. While being very effective on learning holistic image and video representations, such an objective becomes sub-optimal for learning spatio-temporally fine-grained features in videos, where scenes and instances evolve through space and time. In this paper, we present Contextualized Spatio-Temporal Contrastive Learning (ConST-CL) to effectively learn spatio-temporally fine-grained video representations via self-supervision. We first design a region-based pretext task which requires the model to transform in-stance representations from one view to another, guided by context features. Further, we introduce a simple network design that successfully reconciles the simultaneous learning process of both holistic and local representations. We evaluate our learned representations on a variety of downstream tasks and show that ConST-CL achieves competitive results on 6 datasets, including Kinetics, UCF, HMDB, AVA-Kinetics, AVA and OTB. | ['Ting Liu', 'Hartwig Adam', 'Ming-Hsuan Yang', 'Florian Schroff', 'Boqing Gong', 'Yin Cui', 'Rui Qian', 'Liangzhe Yuan'] | 2021-12-09 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Yuan_Contextualized_Spatio-Temporal_Contrastive_Learning_With_Self-Supervision_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Yuan_Contextualized_Spatio-Temporal_Contrastive_Learning_With_Self-Supervision_CVPR_2022_paper.pdf | cvpr-2022-1 | ['action-localization', 'spatio-temporal-action-localization'] | ['computer-vision', 'computer-vision'] | [ 1.87627688e-01 -2.77755737e-01 -5.55835187e-01 -5.67022562e-01
-6.76285446e-01 -5.54209232e-01 1.00245655e+00 -2.64838561e-02
-3.75699401e-01 6.60550535e-01 5.65457821e-01 -7.11777583e-02
-2.45140553e-01 -5.97878397e-01 -1.08066726e+00 -5.11609435e-01
-3.17984641e-01 1.06049227e-02 1.63705349e-01 -1.24525778e-01
4.98863943e-02 4.84556466e-01 -1.64302325e+00 9.82782960e-01
3.79405379e-01 8.72192562e-01 8.41362476e-02 7.95452178e-01
1.47131160e-01 1.31022000e+00 -1.48771018e-01 -8.14129133e-03
2.33631924e-01 -6.43144131e-01 -1.03687298e+00 4.24874216e-01
8.81775856e-01 -1.37164891e-01 -7.65852630e-01 5.90618908e-01
3.94902714e-02 4.56191391e-01 9.19103980e-01 -1.39807665e+00
-9.08483446e-01 3.94863576e-01 -7.97526717e-01 7.90877819e-01
3.52648169e-01 2.63434470e-01 1.09320402e+00 -7.76937723e-01
1.00035059e+00 1.24685812e+00 5.18266022e-01 5.51757991e-01
-1.27577460e+00 -4.43987489e-01 8.13683510e-01 4.67228204e-01
-1.08458400e+00 -4.18536931e-01 5.90383887e-01 -5.76385796e-01
1.20086086e+00 8.64314809e-02 7.64779866e-01 1.53419983e+00
2.34689474e-01 9.89009440e-01 1.18693948e+00 -2.59886622e-01
8.23132768e-02 -2.99476653e-01 -5.85907102e-02 6.71361685e-01
-2.94098169e-01 4.16322291e-01 -1.04077137e+00 2.92978197e-01
1.02739787e+00 4.55678463e-01 -2.15414867e-01 -5.91906965e-01
-1.37388694e+00 7.02002704e-01 6.76991403e-01 4.27889556e-01
-3.60705823e-01 4.13686305e-01 5.98501444e-01 6.35718405e-01
6.00399673e-01 2.58447498e-01 -6.30635083e-01 -4.08479944e-02
-1.00408113e+00 1.00366727e-01 2.82905072e-01 9.70442474e-01
6.71177328e-01 1.28146231e-01 -3.98233235e-01 5.62937617e-01
1.67334024e-02 1.18949696e-01 7.08626330e-01 -1.05634356e+00
3.48245919e-01 3.08065683e-01 -1.38616055e-01 -8.28167796e-01
-2.03194141e-01 -2.87231892e-01 -8.85493159e-01 1.34179145e-01
9.31286588e-02 6.90854862e-02 -1.08511257e+00 1.88193500e+00
2.06492662e-01 5.81144750e-01 7.28750750e-02 8.48225594e-01
8.58648896e-01 7.97210693e-01 2.53247082e-01 -3.64408314e-01
9.27885950e-01 -1.44977117e+00 -5.93954325e-01 -9.67571288e-02
4.50565457e-01 -7.30039775e-02 1.02664077e+00 2.03222230e-01
-1.09084284e+00 -8.34667087e-01 -8.55852902e-01 -1.59357548e-01
-4.67574626e-01 -2.39742443e-01 5.53169250e-01 -1.34886295e-01
-1.27486110e+00 6.62519157e-01 -8.76756012e-01 -5.57721853e-01
6.94366395e-01 8.85458365e-02 -8.56606781e-01 -2.33014375e-01
-8.26332986e-01 5.77906549e-01 3.69431138e-01 -2.88982481e-01
-1.39044249e+00 -8.01542938e-01 -9.95618224e-01 -3.66066061e-02
3.40016633e-01 -8.11992347e-01 1.11473095e+00 -1.47826207e+00
-1.09740460e+00 1.07949436e+00 -2.81727284e-01 -6.95335090e-01
4.84689415e-01 -4.24677044e-01 -3.81723106e-01 4.63289320e-01
1.92612007e-01 9.08974171e-01 1.08084810e+00 -1.31264699e+00
-7.38335311e-01 -3.64868253e-01 3.71428102e-01 4.51327145e-01
-2.44249031e-01 -6.28603548e-02 -3.91449064e-01 -9.30444598e-01
-1.02829896e-01 -5.99701703e-01 -1.95943296e-01 1.95991158e-01
-3.59292291e-02 -3.17318738e-01 1.04742098e+00 -4.27824557e-01
9.13178563e-01 -2.35318351e+00 4.06469613e-01 -2.16442421e-01
1.83970153e-01 -5.45565709e-02 -3.80564213e-01 5.00554442e-01
-4.97101814e-01 -1.55719640e-02 -6.03311472e-02 -4.10695970e-01
-3.55891526e-01 4.29144472e-01 -4.44708049e-01 5.26834249e-01
3.23289663e-01 1.16365957e+00 -1.17923677e+00 -5.23655355e-01
2.58794129e-01 4.79500532e-01 -6.59925938e-01 3.18661839e-01
-3.42696816e-01 5.95483840e-01 -2.94905931e-01 5.68899870e-01
8.19803849e-02 -5.51928937e-01 1.10261396e-01 -2.30953872e-01
-8.33938643e-02 1.15534656e-01 -6.84397995e-01 2.13007116e+00
-5.01011789e-01 7.33633637e-01 -1.51198044e-01 -1.32450974e+00
4.49369460e-01 3.82877380e-01 6.73785627e-01 -1.00393975e+00
-2.61839598e-01 -3.66696745e-01 -5.61773002e-01 -6.64206803e-01
4.08681035e-01 -2.01936066e-01 -2.83598881e-02 4.31837648e-01
5.65639138e-01 3.19527298e-01 2.30708316e-01 4.60387588e-01
1.07183397e+00 7.07123339e-01 4.14589018e-01 -1.41671002e-01
3.64394158e-01 3.90869528e-02 5.89380264e-01 6.51450455e-01
-3.22069436e-01 7.86695719e-01 3.24766129e-01 -6.69308126e-01
-8.18334043e-01 -1.21755195e+00 1.32591933e-01 1.63803697e+00
1.69782132e-01 -5.31164408e-01 -3.45241994e-01 -9.93705809e-01
-1.54900849e-02 4.08942580e-01 -9.88528550e-01 -2.49045402e-01
-5.02094805e-01 -2.12010682e-01 1.47047639e-01 8.48396480e-01
5.49336255e-01 -1.23357546e+00 -6.86996400e-01 8.97089988e-02
-6.46129772e-02 -1.11369514e+00 -5.63511312e-01 3.41399282e-01
-9.60983753e-01 -9.99053061e-01 -5.07813692e-01 -7.87414730e-01
4.72943306e-01 6.31589115e-01 1.29741931e+00 1.09216357e-02
-3.74187112e-01 8.22315335e-01 -6.06362939e-01 1.30718172e-01
-5.61561435e-02 -2.97456272e-02 4.56368253e-02 1.67828828e-01
7.06577823e-02 -7.74119437e-01 -6.36339724e-01 1.99816361e-01
-9.06336367e-01 1.94787353e-01 3.62031043e-01 9.44895506e-01
8.16733420e-01 -2.87888408e-01 6.42720580e-01 -9.62925017e-01
1.28727406e-01 -6.91308737e-01 -1.05265766e-01 3.14119041e-01
-2.42513120e-01 -4.68464009e-02 5.93154788e-01 -4.27472800e-01
-1.10983455e+00 9.39059407e-02 2.89495409e-01 -1.09346437e+00
-3.70900184e-01 4.62478101e-01 9.35570225e-02 2.66655207e-01
7.75030077e-01 3.70261490e-01 -1.53660178e-01 -2.93154955e-01
7.29426384e-01 2.18705714e-01 7.19739556e-01 -6.57867491e-01
6.10205889e-01 7.56775856e-01 -2.83141524e-01 -6.39035702e-01
-1.16717601e+00 -5.73300719e-01 -1.04762352e+00 -2.56202310e-01
1.09390783e+00 -1.25021577e+00 -3.32401395e-01 1.08096220e-01
-6.52378201e-01 -6.37743890e-01 -6.87949896e-01 2.14168161e-01
-9.59985614e-01 1.20339036e-01 -5.61823308e-01 -2.63820082e-01
2.38028206e-02 -7.77962446e-01 1.12745965e+00 1.63000420e-01
-1.69329926e-01 -1.27847970e+00 2.21377969e-01 3.31487507e-01
1.62314147e-01 5.12560725e-01 6.04575098e-01 -3.00298214e-01
-5.82920849e-01 2.18083918e-01 -2.23169431e-01 1.62253976e-01
3.46554697e-01 7.43934363e-02 -9.87530828e-01 -5.96927583e-01
-2.91133404e-01 -8.14243019e-01 1.21692109e+00 4.33190852e-01
1.58799088e+00 -3.93874735e-01 -3.67894232e-01 9.72810090e-01
1.45035684e+00 -5.24345376e-02 3.35952103e-01 4.33130413e-01
8.14760566e-01 5.87066472e-01 6.63772583e-01 3.32884640e-01
4.30622578e-01 5.07026136e-01 4.37461615e-01 -1.00754917e-01
-3.80093127e-01 -3.69658649e-01 5.78388989e-01 5.73682904e-01
-3.13277125e-01 -6.50565773e-02 -5.55189252e-01 8.86208773e-01
-2.12121344e+00 -1.44447470e+00 3.46393228e-01 1.76034284e+00
7.05190539e-01 -1.49844989e-01 3.35310191e-01 -1.97550386e-01
4.64093953e-01 6.69750631e-01 -7.28404403e-01 -2.49869451e-01
-2.45069087e-01 7.52964690e-02 1.48758322e-01 1.83010414e-01
-1.44438267e+00 1.00478983e+00 6.84173632e+00 4.69612330e-01
-1.22605312e+00 3.92801106e-01 7.90041029e-01 -6.50663614e-01
-2.19408110e-01 -2.35075474e-01 -2.37881854e-01 2.24656016e-01
8.78762007e-01 -1.15977809e-01 3.19580555e-01 6.74061000e-01
8.13248158e-02 2.32823357e-01 -1.50648797e+00 9.86508369e-01
1.86680079e-01 -1.73215508e+00 2.86811471e-01 -1.10539153e-01
1.14502823e+00 2.57853419e-01 2.59945303e-01 4.68871742e-01
5.01106918e-01 -1.28610444e+00 6.94570303e-01 5.58987200e-01
9.58874941e-01 -6.04641676e-01 1.27378836e-01 1.83345318e-01
-1.28877056e+00 -2.35630721e-01 -1.44622937e-01 -1.40189584e-02
5.16960286e-02 -4.74797469e-03 -1.43342465e-01 7.42208660e-01
1.03190482e+00 1.56149936e+00 -4.86290365e-01 6.90467775e-01
-1.70551255e-01 4.10334527e-01 1.35680810e-01 6.12656355e-01
5.99200904e-01 4.17767540e-02 1.62474588e-01 1.37620449e+00
-2.68106312e-02 2.46198341e-01 3.57690662e-01 6.13462567e-01
-1.92572400e-01 -2.08528623e-01 -8.79920065e-01 7.82125979e-04
2.52432197e-01 1.16761672e+00 -5.41164577e-01 -5.28689027e-01
-7.05015540e-01 1.11677980e+00 7.00048864e-01 6.54706120e-01
-8.05864334e-01 4.05726314e-01 8.16619575e-01 2.17796341e-01
6.08725309e-01 -1.57248735e-01 1.81822479e-01 -1.42202032e+00
-1.38847947e-01 -9.68070090e-01 8.47621799e-01 -8.72115791e-01
-1.46511018e+00 6.11396790e-01 2.24331617e-01 -1.31066942e+00
-4.33985651e-01 -5.42159081e-01 -5.54884613e-01 4.52210814e-01
-1.66382444e+00 -1.35169446e+00 -3.74285430e-01 1.18904138e+00
1.02120388e+00 -2.71204591e-01 6.67392612e-01 1.15707681e-01
-4.74249095e-01 4.61751342e-01 3.28404387e-03 1.17987759e-01
7.08160579e-01 -1.21725678e+00 3.17019343e-01 7.62904048e-01
6.38041139e-01 4.07407939e-01 4.42520857e-01 -4.17333871e-01
-1.32945907e+00 -1.43737841e+00 4.41183418e-01 -4.05491650e-01
6.80390656e-01 -1.97505087e-01 -8.60703170e-01 1.11191189e+00
5.50963640e-01 8.01145434e-01 5.98853588e-01 2.03860641e-01
-8.12444329e-01 -7.03352392e-02 -9.52001512e-01 4.45870042e-01
1.58589470e+00 -9.69750881e-01 -7.64025569e-01 4.79900151e-01
7.97603667e-01 -2.99615145e-01 -9.32150245e-01 3.70970547e-01
4.47787285e-01 -1.11881232e+00 1.13005197e+00 -1.24345136e+00
7.90324569e-01 -1.44963592e-01 -4.06353354e-01 -1.35915506e+00
-5.98370790e-01 -5.07597685e-01 -4.57715839e-01 9.70391810e-01
9.61112976e-02 -2.43774727e-01 7.06487298e-01 5.86033473e-03
-8.45107958e-02 -8.88184071e-01 -8.05527687e-01 -7.84876227e-01
9.09074992e-02 -2.49225616e-01 2.47756019e-01 1.29295158e+00
-1.07384883e-02 4.02707964e-01 -4.39740479e-01 5.37544303e-03
4.91821527e-01 4.08929914e-01 5.37325025e-01 -8.47012639e-01
-4.05320406e-01 -3.17752391e-01 -5.13470888e-01 -1.10416818e+00
3.15482765e-01 -9.50568736e-01 -1.82694107e-01 -1.40559733e+00
5.16822875e-01 -9.82275531e-02 -6.79393172e-01 6.36754274e-01
-1.13807209e-01 2.78882951e-01 1.90015495e-01 3.43955278e-01
-1.10116756e+00 6.03203714e-01 1.32841599e+00 -1.98144242e-01
-9.76196490e-03 -3.76272678e-01 -5.87548614e-01 5.23486137e-01
5.47025800e-01 -2.25695193e-01 -8.19165885e-01 -5.28868914e-01
-1.29599869e-02 1.35427743e-01 5.56129396e-01 -7.91177332e-01
1.10855162e-01 -3.94235134e-01 6.90793216e-01 -5.69703043e-01
3.85505676e-01 -6.33251846e-01 -1.42918080e-01 4.95920479e-02
-7.19158351e-01 2.87754208e-01 7.62113035e-02 9.29132342e-01
-5.58564246e-01 3.49435210e-01 7.20042408e-01 -4.10052538e-01
-1.21782947e+00 7.32720673e-01 -1.41017988e-01 3.01638186e-01
1.21278024e+00 -3.72667611e-01 -3.33555639e-01 -4.48546916e-01
-1.09791911e+00 3.19222808e-01 4.96864676e-01 7.30240643e-01
7.52642930e-01 -1.48937905e+00 -6.37858927e-01 2.56820500e-01
4.00367647e-01 2.01443713e-02 6.13219678e-01 6.72490239e-01
-3.38301599e-01 3.36092919e-01 -5.28626680e-01 -9.65765893e-01
-1.24371672e+00 8.91325951e-01 4.24309641e-01 -1.85898766e-01
-1.00259614e+00 9.85780299e-01 7.69227982e-01 -3.40695173e-01
2.96485543e-01 -1.61722526e-01 -3.06495011e-01 -2.24436577e-02
5.71823657e-01 -3.08650900e-02 -2.49762326e-01 -8.41382504e-01
-1.95336178e-01 4.58752424e-01 -2.66259700e-01 -3.80814783e-02
1.59499443e+00 -2.90267795e-01 2.11179137e-01 7.93643057e-01
1.39420724e+00 -7.05448985e-01 -2.00107908e+00 -4.74237233e-01
-9.08083096e-02 -4.63639259e-01 -3.46871763e-02 -5.85328639e-01
-1.30187905e+00 8.68397534e-01 5.01748741e-01 -1.14544518e-02
1.15271974e+00 2.84939706e-01 3.00431818e-01 2.43253052e-01
3.36363643e-01 -9.37784255e-01 6.96993589e-01 5.13450801e-01
1.06842268e+00 -1.31767464e+00 4.99062985e-02 1.36775756e-03
-8.40485394e-01 8.81831527e-01 8.61018836e-01 -4.32794958e-01
6.01129055e-01 3.10738254e-02 -1.27724931e-01 -3.63858014e-01
-1.34492135e+00 -1.97756425e-01 4.39315081e-01 6.61616504e-01
5.20157337e-01 -1.15301467e-01 3.52460206e-01 1.79921776e-01
1.09939761e-01 -5.17452620e-02 2.44508490e-01 1.16828418e+00
-1.71946302e-01 -7.61315882e-01 1.90630853e-02 3.50681812e-01
-3.67070764e-01 6.01668991e-02 -2.93126047e-01 1.01368773e+00
1.13795526e-01 6.41310692e-01 4.60553795e-01 -2.88916737e-01
1.78705826e-01 4.00555618e-02 7.27806091e-01 -7.67346859e-01
-5.61134934e-01 2.54224911e-02 -4.15296257e-02 -1.04226077e+00
-1.00876701e+00 -8.83401990e-01 -1.11794364e+00 -1.18240476e-01
2.14845240e-01 -1.74690187e-01 1.22381620e-01 1.11417413e+00
4.21721786e-01 8.85520875e-01 7.30022967e-01 -9.46803629e-01
-1.94391623e-01 -7.56108224e-01 -5.40215731e-01 6.94666684e-01
6.74866796e-01 -7.03414977e-01 -2.05078006e-01 4.77907836e-01] | [8.783595085144043, 0.7800199389457703] |
6ff4c43a-3fe4-4b12-9d3a-2cd4821553b2 | inpaint-anything-segment-anything-meets-image | 2304.0679 | null | https://arxiv.org/abs/2304.06790v1 | https://arxiv.org/pdf/2304.06790v1.pdf | Inpaint Anything: Segment Anything Meets Image Inpainting | Modern image inpainting systems, despite the significant progress, often struggle with mask selection and holes filling. Based on Segment-Anything Model (SAM), we make the first attempt to the mask-free image inpainting and propose a new paradigm of ``clicking and filling'', which is named as Inpaint Anything (IA). The core idea behind IA is to combine the strengths of different models in order to build a very powerful and user-friendly pipeline for solving inpainting-related problems. IA supports three main features: (i) Remove Anything: users could click on an object and IA will remove it and smooth the ``hole'' with the context; (ii) Fill Anything: after certain objects removal, users could provide text-based prompts to IA, and then it will fill the hole with the corresponding generative content via driving AIGC models like Stable Diffusion; (iii) Replace Anything: with IA, users have another option to retain the click-selected object and replace the remaining background with the newly generated scenes. We are also very willing to help everyone share and promote new projects based on our Inpaint Anything (IA). Our codes are available at https://github.com/geekyutao/Inpaint-Anything. | ['Zhibo Chen', 'Wenjun Zeng', 'Xin Jin', 'Jinming Liu', 'Ruoyu Feng', 'Runseng Feng', 'Tao Yu'] | 2023-04-13 | null | null | null | null | ['image-inpainting'] | ['computer-vision'] | [ 1.93543717e-01 -1.05223887e-01 7.61208981e-02 -6.27471432e-02
-7.15173364e-01 -5.01422107e-01 2.91233271e-01 -1.55334651e-01
-2.38495305e-01 6.44157052e-01 1.67797297e-01 -2.45408654e-01
3.25586140e-01 -5.18754959e-01 -7.73601532e-01 -5.19632041e-01
3.98462296e-01 2.53988534e-01 5.49693942e-01 -3.64504486e-01
3.77457529e-01 3.69490474e-01 -1.45355034e+00 4.78777945e-01
9.72177207e-01 6.98511958e-01 8.44195604e-01 6.07656419e-01
-4.71311718e-01 8.68215978e-01 -4.92628872e-01 -3.54534388e-01
5.37825823e-01 -7.88546085e-01 -7.16536224e-01 2.39421487e-01
4.11402583e-01 -5.12763977e-01 -2.09632277e-01 9.91357267e-01
4.04759705e-01 1.26937747e-01 1.09421179e-01 -1.24317586e+00
-6.93807542e-01 3.53527904e-01 -8.94234180e-01 1.90578088e-01
3.02612782e-01 5.41505516e-01 5.89159369e-01 -1.27792668e+00
9.87760723e-01 1.11382592e+00 6.77668095e-01 7.53023744e-01
-1.09616506e+00 -4.83530611e-01 2.98281670e-01 5.64321950e-02
-1.10442662e+00 -5.48202693e-01 6.70935214e-01 -3.54324639e-01
7.43393838e-01 6.88094497e-01 9.64246452e-01 8.47471714e-01
7.76855275e-02 1.02216792e+00 1.00719309e+00 -5.45377910e-01
7.73131996e-02 1.19984701e-01 -1.27789125e-01 8.02637994e-01
-2.49915645e-01 -2.16096178e-01 -5.81299961e-01 2.20440384e-02
9.87717628e-01 2.70398438e-01 -2.75189400e-01 -1.85514446e-02
-1.32284582e+00 5.82111597e-01 4.39503133e-01 2.55334914e-01
-5.53799689e-01 2.51052618e-01 1.09764688e-01 2.85800099e-01
6.66128337e-01 4.57045227e-01 -3.32353294e-01 -1.75215185e-01
-1.27553046e+00 4.37977284e-01 4.30883467e-01 1.07905650e+00
1.00045133e+00 -1.09884612e-01 -4.39351976e-01 1.13925755e+00
8.46889615e-02 1.43003941e-01 2.03174353e-01 -1.28577650e+00
2.26363674e-01 3.48795444e-01 2.24324569e-01 -7.61959076e-01
-9.07137021e-02 -2.44581819e-01 -6.19325638e-01 5.78026593e-01
9.52445567e-02 -2.06774727e-01 -1.04513538e+00 1.36921990e+00
3.37069005e-01 2.02561021e-01 -6.05969429e-01 8.45182657e-01
8.13818157e-01 8.54869545e-01 2.07163356e-02 -1.53447121e-01
1.43559575e+00 -1.39826882e+00 -8.78641367e-01 -5.80335736e-01
1.70675695e-01 -1.19332743e+00 1.31427729e+00 3.59292954e-01
-1.50564814e+00 -5.67137420e-01 -8.33450198e-01 -4.66286659e-01
-1.65645942e-01 -7.64120445e-02 6.10602140e-01 1.25860438e-01
-1.35072052e+00 6.47734165e-01 -7.13910520e-01 -4.37867284e-01
4.85087752e-01 -1.22103300e-02 -2.04948649e-01 -1.93139121e-01
-8.28164756e-01 8.24557543e-01 1.33013241e-02 1.47218823e-01
-7.96691597e-01 -8.09844613e-01 -6.47628963e-01 -1.86571419e-01
5.40247500e-01 -9.35211480e-01 1.23101687e+00 -1.11965179e+00
-1.25730920e+00 8.24070275e-01 -5.29296458e-01 -1.18891552e-01
9.23028409e-01 -5.99784613e-01 3.94780450e-02 7.44910762e-02
3.09130073e-01 1.24843979e+00 1.23548508e+00 -1.50155473e+00
-5.52215457e-01 -8.74588713e-02 -2.05788717e-01 3.66751969e-01
-7.37516284e-02 2.62316883e-01 -1.10822332e+00 -1.08192897e+00
1.70882970e-01 -8.69076431e-01 -2.27965310e-01 4.64605659e-01
-5.45284748e-01 1.21999517e-01 1.19719195e+00 -1.15894830e+00
1.46958184e+00 -2.42743063e+00 1.57092690e-01 -3.58936220e-01
3.60512674e-01 2.75759399e-01 -2.90323615e-01 7.15692639e-01
-8.67258459e-02 1.13785945e-01 -4.21241164e-01 -7.50433803e-01
-3.13322514e-01 6.63459972e-02 -2.97007084e-01 9.10567343e-02
3.20702493e-01 9.61324155e-01 -8.64322901e-01 -3.92378569e-01
4.46579993e-01 4.05413598e-01 -6.45660460e-01 1.62824258e-01
-5.45516908e-01 4.53150302e-01 -1.04101896e-01 8.17138135e-01
9.98693585e-01 -3.03984862e-02 -2.64072597e-01 -7.37025812e-02
-4.33151186e-01 1.43267393e-01 -1.19101560e+00 1.83402789e+00
-1.76056668e-01 9.30853665e-01 5.17273128e-01 -5.39665461e-01
7.02005208e-01 1.60336029e-02 3.77697766e-01 -5.13281167e-01
-1.87030181e-01 1.52185336e-01 -3.65946800e-01 -5.29271841e-01
7.01309979e-01 9.55656394e-02 3.78030956e-01 5.45206308e-01
-2.24381700e-01 -4.16532099e-01 2.20414519e-01 5.91001272e-01
1.22099078e+00 5.59135437e-01 -1.38062224e-01 -4.33427207e-02
8.38305876e-02 2.60345414e-02 5.14286220e-01 8.81190240e-01
-2.19073236e-01 1.14546454e+00 4.37001020e-01 -3.76225859e-01
-1.06195867e+00 -9.74008381e-01 2.35223174e-01 1.06943834e+00
2.07754895e-01 -3.92455995e-01 -9.15893078e-01 -4.89206493e-01
-1.13656946e-01 8.13667178e-01 -6.79353058e-01 2.99384654e-01
-6.62631214e-01 -2.79195040e-01 -4.01511602e-02 2.24487290e-01
7.48807251e-01 -1.51001000e+00 -5.40957928e-01 2.52758414e-01
-5.13992131e-01 -6.13337159e-01 -9.10678506e-01 1.52830973e-01
-8.08806598e-01 -6.58851266e-01 -1.11083317e+00 -7.64660537e-01
7.35715628e-01 7.60574520e-01 9.10834432e-01 5.38854003e-01
-4.02857304e-01 2.11537942e-01 -4.39549237e-01 -2.48599842e-01
-3.71292502e-01 -1.61241591e-01 -5.61150968e-01 -1.68759242e-01
-1.12689838e-01 -4.91796136e-01 -9.29580331e-01 3.01544756e-01
-1.15203309e+00 6.74119949e-01 4.36992347e-01 4.51662213e-01
7.42260993e-01 -3.23963612e-02 1.99860558e-01 -8.66940022e-01
7.83127129e-01 -2.89095432e-01 -2.86822468e-01 2.37021577e-02
-2.81142265e-01 -2.07044423e-01 1.72861993e-01 -4.81117338e-01
-8.92443001e-01 -5.98923527e-02 -3.84104699e-01 -5.93443811e-01
7.90047050e-02 2.27956668e-01 -1.07654370e-01 1.58546031e-01
4.42446411e-01 2.17324406e-01 -5.34500219e-02 -5.93923271e-01
5.71534574e-01 3.04622501e-01 6.01903915e-01 -2.75978893e-01
7.37118959e-01 5.76088011e-01 -5.36837876e-01 -5.64729035e-01
-5.37533879e-01 -3.10314745e-01 -3.85138005e-01 -4.26687062e-01
9.35916722e-01 -6.68607712e-01 -1.57755092e-01 7.09821403e-01
-1.24416327e+00 -7.25920677e-01 -5.36053836e-01 6.36340380e-02
-3.93861920e-01 4.31137085e-01 -8.25546682e-01 -5.47436357e-01
-1.95132568e-01 -1.21797121e+00 9.16515470e-01 4.89800662e-01
-3.37264657e-01 -6.98285520e-01 -1.43380582e-01 3.20276767e-01
7.62341201e-01 4.24856134e-02 8.18017960e-01 2.14572176e-02
-8.46289933e-01 -5.13170287e-02 -4.13111687e-01 3.68413478e-01
2.12912828e-01 2.66524851e-01 -7.25913823e-01 -1.81033164e-01
4.28554006e-02 5.51446564e-02 9.35491800e-01 5.40272951e-01
1.29969358e+00 -2.50942498e-01 -4.29443717e-01 6.23256743e-01
1.22477794e+00 2.64667630e-01 1.06092095e+00 6.92751288e-01
5.56789815e-01 4.71074194e-01 5.93707502e-01 4.29860026e-01
1.68475062e-01 6.32717192e-01 5.30270457e-01 -4.58257198e-01
-6.13596141e-01 -5.58607817e-01 3.22556674e-01 7.37993181e-01
-2.99327411e-02 -2.33144775e-01 -4.67664540e-01 4.76380527e-01
-1.97950983e+00 -1.00299323e+00 -3.29383731e-01 2.01068497e+00
9.75211740e-01 1.39966041e-01 -2.34423997e-03 -2.48399034e-01
8.11664760e-01 3.63740593e-01 -4.91500348e-01 -3.39575231e-01
-1.31599620e-01 1.81977853e-01 2.30009288e-01 8.82286072e-01
-8.60993445e-01 1.23523951e+00 5.83486223e+00 9.54442143e-01
-1.04932249e+00 2.15405554e-01 8.10216129e-01 -2.03990892e-01
-5.84727705e-01 4.41432178e-01 -7.57995844e-01 7.24353254e-01
2.69558519e-01 2.23446980e-01 7.43491948e-01 5.48366666e-01
5.54674923e-01 -7.03664243e-01 -6.24718249e-01 8.67853284e-01
6.93415403e-02 -1.59351754e+00 2.82124206e-02 -1.08735919e-01
6.67589724e-01 -1.25580758e-01 8.22937489e-03 9.77671817e-02
2.49648824e-01 -7.41204083e-01 9.66908872e-01 7.04747438e-01
7.25983024e-01 -3.42826277e-01 1.95656598e-01 4.16843683e-01
-8.78570318e-01 1.65609092e-01 -3.75317246e-01 -5.81568442e-02
4.96186912e-01 7.68667638e-01 -4.93290484e-01 2.03950658e-01
8.85924518e-01 6.32795155e-01 -6.11177027e-01 1.32516479e+00
-4.00472224e-01 3.83202434e-01 -3.53932351e-01 3.85800302e-01
-5.18493205e-02 -4.00712252e-01 6.20060086e-01 1.32529175e+00
5.25225341e-01 7.32054114e-02 -1.50602207e-01 9.66429949e-01
-3.34965661e-02 -1.06035720e-03 -2.96755970e-01 1.75975844e-01
4.56787229e-01 1.22683704e+00 -1.09449577e+00 -3.44424486e-01
-4.35510367e-01 1.58098745e+00 1.09132975e-01 4.25672948e-01
-8.73871326e-01 -4.10881251e-01 4.10069197e-01 5.36656857e-01
4.07488912e-01 -1.66566417e-01 -4.43183005e-01 -1.00525379e+00
9.55288932e-02 -9.92709517e-01 2.73120441e-02 -1.41683209e+00
-1.01233184e+00 6.14189804e-01 -6.89814240e-02 -1.06414759e+00
3.80901277e-01 -3.29857469e-02 -8.03106546e-01 9.26700234e-01
-1.19611144e+00 -1.08062971e+00 -5.48240304e-01 6.66896582e-01
1.07186735e+00 3.25822473e-01 4.13511813e-01 4.01330888e-01
-5.85427701e-01 1.08884320e-01 -6.57329336e-02 -1.85885936e-01
8.93977463e-01 -9.69198525e-01 5.99659741e-01 9.72623348e-01
1.02076389e-01 7.83283949e-01 8.14758480e-01 -9.73392785e-01
-1.09678555e+00 -7.63838470e-01 9.16655481e-01 -3.83630574e-01
4.32170421e-01 -3.90309364e-01 -1.01303589e+00 7.61754096e-01
6.57687783e-01 -7.15495646e-02 1.90423176e-01 -4.11823004e-01
-9.81039032e-02 9.21764821e-02 -1.03004384e+00 8.63110423e-01
1.09105802e+00 -1.82053924e-01 -2.98865676e-01 4.61755931e-01
7.77948976e-01 -4.40808237e-01 -7.51532093e-02 8.33249018e-02
3.38117778e-01 -1.31232715e+00 8.62000763e-01 -1.69335276e-01
6.78563893e-01 -5.93208015e-01 1.05718508e-01 -1.23538542e+00
-4.78951275e-01 -1.07719731e+00 1.19639397e-01 1.27651572e+00
3.47778708e-01 -5.29279888e-01 7.18365252e-01 7.23737836e-01
-3.28424782e-01 -8.02408099e-01 -7.05347061e-01 -3.80899936e-01
-1.26810968e-01 -4.91479248e-01 4.11056876e-01 8.63051414e-01
-2.81962007e-01 -7.47555122e-02 -6.97129011e-01 -1.89064756e-01
3.46874058e-01 -8.85157511e-02 8.54755998e-01 -7.64694154e-01
-3.95094693e-01 -4.52229947e-01 3.39391470e-01 -1.41463685e+00
-5.23378313e-01 -5.17412186e-01 1.04888842e-01 -1.95839691e+00
3.90063226e-01 -6.78601921e-01 8.88444334e-02 7.05835223e-01
-5.34766495e-01 5.25039077e-01 5.59484482e-01 6.65455163e-01
-4.89044487e-01 2.97150791e-01 1.47093475e+00 6.09990731e-02
-3.58293414e-01 6.99397996e-02 -9.06267285e-01 6.99561536e-01
6.59682512e-01 -5.02433419e-01 -1.90421060e-01 -6.13751829e-01
2.69340813e-01 1.37849525e-01 5.67598164e-01 -7.92866290e-01
5.83386384e-02 -2.50707746e-01 4.90436405e-01 -6.95859134e-01
4.72121924e-01 -6.16530180e-01 4.70956177e-01 3.64824831e-01
-9.74469110e-02 1.97318494e-01 4.48377341e-01 8.50920305e-02
-8.11779797e-02 -5.05455434e-01 7.89697289e-01 -4.97541219e-01
-7.10185111e-01 2.57021904e-01 -4.34386671e-01 -1.84109852e-01
1.03439248e+00 -4.37195659e-01 -4.15219545e-01 -7.28498399e-01
-9.35898006e-01 1.61967307e-01 8.93715382e-01 4.71772045e-01
6.52907789e-01 -1.03456581e+00 -6.52731538e-01 3.03286135e-01
-2.89221436e-01 -6.28771959e-03 4.77246791e-01 8.38402569e-01
-8.90679240e-01 -1.07689304e-02 -4.87334132e-02 -3.80855739e-01
-1.26448047e+00 5.09856462e-01 1.67780831e-01 -4.02979460e-03
-8.54966462e-01 1.31396902e+00 3.00500363e-01 -5.60701787e-02
1.27363369e-01 -1.34081259e-01 3.00038099e-01 2.20406353e-02
6.59442425e-01 2.85541326e-01 -1.66189820e-02 -2.38965034e-01
-1.50365904e-01 2.78827101e-01 -3.18347305e-01 -4.41141129e-01
1.47446239e+00 -3.06874305e-01 -3.72973472e-01 1.31048262e-01
7.75070906e-01 3.90694827e-01 -1.80525970e+00 -2.18001474e-02
-4.97639686e-01 -9.93340790e-01 -1.60379261e-01 -8.60814333e-01
-1.13528192e+00 7.19384611e-01 4.69311386e-01 3.24492961e-01
1.18843639e+00 1.69666946e-01 8.85752141e-01 -4.60980356e-01
1.60359889e-01 -9.19682980e-01 2.97097623e-01 2.97644079e-01
1.25514901e+00 -8.33776832e-01 2.17447951e-01 -5.18369794e-01
-8.54652226e-01 8.06571603e-01 6.19854331e-01 -1.10868849e-01
5.37766933e-01 3.48889023e-01 3.33756924e-01 -1.09292291e-01
-7.38347411e-01 -2.56038290e-02 -1.51380956e-01 7.14141726e-01
3.59253377e-01 -3.24846119e-01 -4.65715259e-01 1.32669494e-01
-9.69571769e-02 1.56551108e-01 4.13475156e-01 9.90135849e-01
-6.32049501e-01 -1.42149508e+00 -5.01997828e-01 3.96986097e-01
-2.22015575e-01 -3.52993459e-01 -3.80119115e-01 5.99451959e-01
3.29185307e-01 8.38793159e-01 -1.96381554e-01 -8.74475017e-02
2.10384667e-01 -7.35753402e-02 3.53125274e-01 -6.40402019e-01
-7.27810204e-01 3.52020234e-01 -2.40130827e-01 -5.55101335e-01
-1.01407260e-01 -6.61300600e-01 -9.20948923e-01 -4.22194064e-01
-2.23355904e-01 -1.00755982e-01 5.43596625e-01 6.50467992e-01
6.47100449e-01 4.17345196e-01 3.03622544e-01 -1.16810846e+00
9.43299830e-02 -9.85024571e-01 -3.20958287e-01 3.10901135e-01
2.80965388e-01 -3.21645051e-01 -4.01777416e-01 2.09658056e-01] | [11.273577690124512, -0.6605263352394104] |
a2f5a488-6d42-42e3-9227-d2d2fb55233b | pac-prediction-sets-for-large-language-models | 2302.08703 | null | https://arxiv.org/abs/2302.08703v2 | https://arxiv.org/pdf/2302.08703v2.pdf | PAC Prediction Sets for Large Language Models of Code | Prediction sets have recently been shown to be a promising strategy for quantifying the uncertainty of deep neural networks in a way that provides theoretical guarantees. However, existing techniques have largely targeted settings where the space of labels is simple, so prediction sets can be arbitrary subsets of labels. For structured prediction problems where the space of labels is exponential in size, even prediction sets containing a small fraction of all labels can be exponentially large. In the context of code generation, we propose a solution that considers a restricted set of prediction sets that can compactly be represented as partial programs, which are programs with portions replaced with holes. Given a trained code generation model, our algorithm leverages a programming language's abstract syntax tree to generate a set of programs such that the correct program is in the set with high-confidence. Valuable applications of our algorithm include a Codex-style code generator with holes in uncertain parts of the generated code, which provides a partial program with theoretical guarantees. We evaluate our approach on PICARD (a T5 model for SQL semantic parsing) and Codex (a GPT model for over a dozen programming languages, including Python), demonstrating that our approach generates compact PAC prediction sets. This is the first research contribution that generates PAC prediction sets for generative code models. | ['Osbert Bastani', 'Stephen Mell', 'Adam Khakhar'] | 2023-02-17 | null | null | null | null | ['semantic-parsing'] | ['natural-language-processing'] | [ 3.03832471e-01 9.73363340e-01 -5.34717381e-01 -7.66648710e-01
-1.13467336e+00 -8.42599750e-01 2.11962268e-01 2.56723352e-03
2.34160602e-01 3.59164387e-01 -1.19200066e-01 -6.60937905e-01
1.50589466e-01 -1.19594991e+00 -1.57927132e+00 -3.89241338e-01
-2.69182384e-01 7.95309782e-01 2.77456224e-01 1.46679223e-01
3.24937478e-02 -1.24622487e-01 -1.72840810e+00 5.43390512e-01
9.13994610e-01 8.53036582e-01 1.58371657e-01 7.22736537e-01
-3.48942876e-01 9.65651989e-01 -5.23745537e-01 -7.84724295e-01
4.49296623e-01 -3.63088548e-01 -6.96968317e-01 -4.69843179e-01
3.16462040e-01 -2.67294854e-01 5.34967519e-03 1.44210410e+00
-1.54553235e-01 -3.72867644e-01 4.11130607e-01 -1.54469848e+00
-5.53553104e-01 1.83449137e+00 -1.06970243e-01 -5.22275209e-01
-1.74320918e-02 -3.44076939e-02 1.55839109e+00 -4.64795023e-01
3.81141722e-01 1.20165956e+00 7.75363684e-01 9.11381781e-01
-1.62298083e+00 -7.88904190e-01 -2.08631903e-01 -5.13974845e-01
-1.11990106e+00 -1.99288636e-01 4.67732847e-01 -6.53897762e-01
1.09033394e+00 2.92430580e-01 3.16094458e-01 9.09700811e-01
7.47063458e-02 1.04527044e+00 6.98621571e-01 -5.20799756e-01
7.01143503e-01 4.38335210e-01 4.66491073e-01 9.82825994e-01
3.48149419e-01 2.05775663e-01 -1.89845741e-01 -7.11240530e-01
1.62681699e-01 1.43171651e-02 1.28085315e-02 -5.71487069e-01
-8.40533495e-01 1.38792181e+00 2.79582858e-01 -1.72349632e-01
1.03876412e-01 7.08927989e-01 3.30768973e-01 -2.67832093e-02
1.40822217e-01 6.59268856e-01 -6.21296406e-01 -2.98558116e-01
-1.11178911e+00 5.16907454e-01 1.39424348e+00 1.54678023e+00
7.62982130e-01 -1.02712642e-02 -1.01513684e-01 6.14282787e-01
4.36379761e-01 5.64326763e-01 4.80248839e-01 -1.12937248e+00
6.71492994e-01 5.17968714e-01 -1.63885653e-01 -5.64730942e-01
1.15006000e-01 -6.54122680e-02 -2.70666569e-01 2.08785146e-01
3.15946341e-01 -3.56998593e-01 -8.18105459e-01 2.09190702e+00
-2.77223289e-01 -2.27848947e-01 1.70765787e-01 3.90805036e-01
3.44309330e-01 7.50112653e-01 -1.52329905e-02 2.65787423e-01
1.00302947e+00 -9.64096367e-01 1.48057207e-01 -5.42946756e-01
1.09154820e+00 -1.17752030e-02 1.18262601e+00 3.28943729e-01
-9.63565230e-01 -2.00184733e-01 -1.05993557e+00 1.26991987e-01
-2.76669916e-02 4.00223136e-02 9.41411316e-01 1.28602076e+00
-1.22296524e+00 5.67646980e-01 -9.19642329e-01 1.87741190e-01
6.87055349e-01 3.24706167e-01 5.92434034e-02 -2.43011862e-04
-7.22983539e-01 3.20706397e-01 9.69528913e-01 -6.45593166e-01
-1.05124354e+00 -6.73537731e-01 -9.54690337e-01 4.67234254e-01
3.72479975e-01 -4.99969840e-01 1.59181142e+00 -1.02357984e+00
-1.23474526e+00 6.34463847e-01 -1.93342403e-01 -1.09696364e+00
4.07228172e-01 3.77718732e-02 1.56682611e-01 -2.50076354e-01
3.74891274e-02 8.74184489e-01 6.40791357e-01 -1.20573807e+00
-6.15112245e-01 -1.97595298e-01 3.36523622e-01 -1.49675488e-01
-1.33615196e-01 -1.02176666e-01 -1.26786053e-01 -3.87976229e-01
4.65149879e-02 -1.22423279e+00 -3.68104637e-01 -2.71208525e-01
-6.70924783e-01 -2.23243386e-01 4.10350442e-01 -3.92054081e-01
1.08623397e+00 -2.22782707e+00 -5.80424704e-02 3.52662921e-01
2.65292346e-01 -1.28787458e-01 -1.43498741e-03 -3.44696976e-02
1.98107779e-01 7.15248525e-01 -6.16287708e-01 -2.71883130e-01
6.01467371e-01 4.70153511e-01 -1.11192060e+00 -1.42842740e-01
2.58735418e-01 1.03439498e+00 -6.64358258e-01 -4.01060969e-01
-4.00628030e-01 -1.89178735e-01 -1.16325235e+00 3.01318765e-01
-1.22585607e+00 -4.69366103e-01 -4.21212018e-01 5.20431519e-01
5.14603376e-01 -4.52412665e-01 3.72589305e-02 5.93788385e-01
4.60629463e-01 4.97803539e-01 -8.93311977e-01 1.27157164e+00
-4.47368503e-01 4.90702689e-01 -3.74202341e-01 -7.39026546e-01
1.11636734e+00 -8.73311982e-02 -4.31584194e-05 1.37880489e-01
-9.45185572e-02 3.57172132e-01 -7.10859746e-02 -1.94763452e-01
5.96724570e-01 -1.92451268e-01 -5.14406443e-01 7.45885193e-01
1.30170986e-01 -5.04300058e-01 4.57006842e-02 4.13766265e-01
1.31144333e+00 1.68444425e-01 1.70809552e-02 -1.71823829e-01
-2.21026540e-01 1.62181631e-01 5.02629161e-01 1.38857174e+00
1.83345545e-02 8.41332138e-01 1.11200893e+00 -1.12372823e-01
-1.30796123e+00 -9.47902799e-01 -1.63546175e-01 1.16617239e+00
-2.16976106e-01 -6.82224274e-01 -1.00464261e+00 -8.87744367e-01
-8.50705653e-02 1.08758116e+00 -5.80484271e-01 -1.34256646e-01
-4.32249874e-01 -8.06897521e-01 9.88348722e-01 6.38092637e-01
1.89971313e-01 -9.83220518e-01 -8.06549013e-01 6.96985424e-02
-1.36118382e-02 -8.80239189e-01 -2.81700730e-01 8.28883648e-01
-1.00080466e+00 -1.15781283e+00 -1.20982371e-01 -8.84871244e-01
6.49135232e-01 -4.90323752e-01 1.43771839e+00 -6.09411672e-02
8.15770477e-02 8.93614888e-02 -1.77554727e-01 -5.60026288e-01
-1.28716373e+00 5.54057099e-02 -3.96336436e-01 -6.04398012e-01
5.46139359e-01 -5.10147929e-01 3.52968499e-02 -2.27813348e-01
-8.66011798e-01 2.39378422e-01 3.75425547e-01 9.35784936e-01
6.37288928e-01 3.67540956e-01 1.86588094e-01 -1.54389393e+00
4.04598445e-01 -9.41186070e-01 -1.00092888e+00 4.16881025e-01
-6.56594157e-01 5.42417049e-01 9.44531560e-01 -3.79650205e-01
-8.27616394e-01 3.27695310e-01 -9.08149406e-02 -2.94789374e-01
-7.03783706e-02 6.43948793e-01 -2.09724844e-01 1.55547559e-01
1.16227806e+00 3.66564691e-01 -9.55454409e-02 -9.93707627e-02
5.34955144e-01 5.23783863e-01 4.55162287e-01 -1.17218089e+00
6.49711728e-01 5.92978671e-04 -5.95653020e-02 -6.92888349e-03
-7.41235197e-01 2.08019286e-01 -3.67860198e-01 3.83703351e-01
4.47636873e-01 -9.61281538e-01 -3.54815394e-01 1.14016473e-01
-1.00049353e+00 -6.68956280e-01 -5.13396382e-01 2.85312831e-02
-8.56320679e-01 1.81914255e-01 -6.69569135e-01 -8.65166485e-01
-2.60066032e-01 -1.50470638e+00 1.03133881e+00 1.03193127e-01
-3.23832691e-01 -7.02809930e-01 -9.90723893e-02 5.47025539e-02
2.03764960e-01 1.74969405e-01 1.63353539e+00 -1.34006810e+00
-1.06002998e+00 -2.26019964e-01 1.27220392e-01 6.17511988e-01
-5.11477172e-01 1.55209124e-01 -9.99871433e-01 1.32065833e-01
1.85543969e-02 -6.66192949e-01 7.99622953e-01 3.27183783e-01
1.58570075e+00 -7.63455391e-01 -4.15051222e-01 7.04578578e-01
1.41900206e+00 3.10551077e-01 4.65816945e-01 3.09931859e-02
5.74151754e-01 4.70125169e-01 3.13674837e-01 4.81542468e-01
2.92535424e-01 2.22272843e-01 5.68508983e-01 8.61778617e-01
3.55318397e-01 -8.10865641e-01 4.76231992e-01 4.67716783e-01
6.77469432e-01 -6.48004711e-02 -1.16794312e+00 4.68566179e-01
-1.59740281e+00 -8.82387996e-01 5.46712391e-02 2.22140479e+00
1.29300106e+00 3.80696416e-01 6.09250888e-02 -7.96367377e-02
5.55575788e-01 -6.07137717e-02 -6.41493320e-01 -5.04220963e-01
-7.28615224e-02 4.52458680e-01 6.96076870e-01 3.54278326e-01
-1.12587285e+00 7.72164941e-01 6.70011663e+00 9.11915541e-01
-7.58370996e-01 2.14404501e-02 1.03751230e+00 1.28679439e-01
-1.03129315e+00 3.78868401e-01 -1.05150461e+00 7.26171076e-01
1.47441351e+00 -4.64507878e-01 7.30067551e-01 1.92126369e+00
-8.05133879e-01 -9.82820801e-03 -1.69309592e+00 6.47310615e-01
-6.16593175e-02 -1.43143988e+00 -1.80658743e-01 1.79617777e-02
9.20176625e-01 2.40493327e-01 3.48363459e-01 8.93917143e-01
1.17392802e+00 -1.25004935e+00 1.22154653e+00 2.64866669e-02
7.05916762e-01 -7.18711019e-01 4.87773865e-01 7.04699159e-01
-5.11927068e-01 -5.04471540e-01 -5.60680270e-01 2.60975182e-01
-3.02991539e-01 8.54715943e-01 -1.17079365e+00 -2.97114313e-01
7.14858413e-01 3.06535810e-01 -6.41499996e-01 8.12310159e-01
-1.92427531e-01 1.03903604e+00 -5.60488105e-01 -3.14724952e-01
3.84186730e-02 1.33460790e-01 3.05436015e-01 1.13785660e+00
7.20720112e-01 -1.50956631e-01 -2.89527681e-02 1.78212190e+00
-4.13534313e-01 -2.14011610e-01 -9.06296968e-01 -3.09300333e-01
7.99199760e-01 7.31340528e-01 -5.28377891e-01 -3.56579423e-01
-2.43227407e-01 3.21376562e-01 4.63942975e-01 1.80414785e-02
-1.00782502e+00 -3.32497895e-01 3.69627446e-01 -1.56052351e-01
4.55099940e-01 1.55897334e-01 -7.52846479e-01 -1.05303025e+00
2.74158001e-01 -1.04183614e+00 4.09067631e-01 -7.23810256e-01
-9.37064648e-01 5.20951450e-01 1.28198743e-01 -7.42993236e-01
-6.58172667e-01 -7.13621259e-01 -3.32693338e-01 6.90450370e-01
-7.37302780e-01 -8.39457333e-01 1.08826958e-01 1.51943574e-02
2.61007041e-01 -4.22360152e-01 9.74971175e-01 -3.57722878e-01
-1.25953227e-01 9.70624626e-01 2.28037536e-01 8.59996527e-02
-1.18160896e-01 -1.57180500e+00 5.35992026e-01 1.07647669e+00
3.97351176e-01 7.56769538e-01 7.23217905e-01 -5.55566251e-01
-1.42034686e+00 -1.46881151e+00 7.11936355e-01 -8.26696277e-01
6.02456391e-01 -5.35166025e-01 -7.64365375e-01 1.29495549e+00
-4.82351303e-01 4.76416294e-03 6.68291926e-01 4.06448394e-02
-9.15890217e-01 2.06293747e-01 -1.19381011e+00 6.59246802e-01
8.23418379e-01 -6.06486678e-01 -5.04728854e-01 4.01970923e-01
1.29493141e+00 -7.69568682e-01 -7.84565747e-01 2.80276716e-01
3.55030388e-01 -9.98956919e-01 5.34312487e-01 -5.80662727e-01
9.71923947e-01 -2.14799196e-02 -6.87094092e-01 -1.03657436e+00
1.86570048e-01 -5.71597040e-01 -5.29374242e-01 1.13743651e+00
7.32370973e-01 -5.56937695e-01 1.08400130e+00 1.22092009e+00
-2.02317983e-01 -6.91154063e-01 -7.29476333e-01 -6.54584289e-01
3.93853664e-01 -9.34753418e-01 1.01922512e+00 3.54718328e-01
2.23159909e-01 -1.23180553e-01 -3.05725485e-02 1.61534816e-01
4.91298765e-01 3.99882674e-01 6.53365672e-01 -1.19813037e+00
-1.04561806e+00 -4.89293158e-01 -2.99256861e-01 -9.02265966e-01
6.00759685e-01 -1.35217237e+00 5.44828296e-01 -6.79506838e-01
6.43236041e-01 -1.10731578e+00 1.77678406e-01 8.20541799e-01
6.37460575e-02 -3.37151349e-01 1.36490852e-01 1.36622593e-01
-3.55766386e-01 2.36798346e-01 3.61750990e-01 -2.59040952e-01
-1.09392203e-01 3.76736343e-01 -1.19646752e+00 5.24988890e-01
8.78665328e-01 -8.36076379e-01 -6.99397087e-01 -2.82819241e-01
6.46681309e-01 2.03205228e-01 3.81859958e-01 -1.05517578e+00
-3.30835916e-02 -6.96081109e-03 -3.37871879e-01 -2.62571096e-01
-2.13772766e-02 -7.11917281e-01 4.10581350e-01 4.73838329e-01
-1.15269518e+00 -3.19822967e-01 1.28342867e-01 6.33516788e-01
-2.38524843e-02 -8.95170033e-01 7.53867388e-01 -3.24920177e-01
-3.31383675e-01 3.16387296e-01 -4.15778086e-02 5.26182055e-01
1.00243545e+00 1.25246048e-01 -4.68676209e-01 -2.84328163e-01
-4.33004618e-01 -7.82267302e-02 5.93225121e-01 3.77027571e-01
4.86221462e-01 -1.00992632e+00 -3.17685664e-01 2.90982395e-01
3.18707794e-01 4.42123771e-01 -2.40629911e-01 8.08638334e-02
-6.10738575e-01 5.09217739e-01 1.86248675e-01 -6.84643507e-01
-1.01565289e+00 6.09082818e-01 4.11574453e-01 -1.80030644e-01
-6.10444069e-01 1.19759595e+00 2.33684942e-01 -8.17630410e-01
3.87615144e-01 -7.25971878e-01 3.96362245e-01 -6.66794598e-01
3.55267227e-01 -1.90143138e-01 -2.87287176e-01 -1.84496462e-01
1.25985965e-02 -1.70276672e-01 4.78166863e-02 -7.52031505e-02
1.35540342e+00 5.00666618e-01 -4.39066291e-01 4.14979756e-01
1.23525977e+00 9.04776007e-02 -1.30949879e+00 -2.26075068e-01
3.16663504e-01 -4.33468521e-01 -4.48094845e-01 -7.48046160e-01
-9.33719814e-01 1.00646687e+00 2.18091071e-01 2.14694470e-01
8.49725127e-01 4.57237124e-01 6.14959419e-01 6.96379721e-01
9.11460757e-01 -5.35418153e-01 -4.40368876e-02 5.92773199e-01
4.16134924e-01 -1.15974951e+00 -6.89827323e-01 -2.54037261e-01
-6.91952229e-01 1.20280313e+00 7.72975445e-01 -2.42844261e-02
6.15503430e-01 8.31030130e-01 -6.00630045e-01 1.02942914e-01
-1.00859869e+00 4.44263071e-01 -1.36629909e-01 7.23913848e-01
1.89832300e-01 6.36785507e-01 3.05862576e-01 1.45509505e+00
-8.83830786e-01 4.18146551e-02 7.74307430e-01 8.50593984e-01
-6.73974276e-01 -1.07294083e+00 -1.35673553e-01 1.02087545e+00
-5.47028899e-01 -5.31374037e-01 8.06850344e-02 3.80471349e-01
1.37360364e-01 5.88480592e-01 -6.17079213e-02 -7.90390909e-01
-3.61945868e-01 4.23421264e-01 2.95523822e-01 -1.03920460e+00
-3.28358054e-01 -5.82352877e-01 6.23052344e-02 -3.88103336e-01
2.69044727e-01 -5.90962291e-01 -1.49238467e+00 -1.61006391e-01
-1.13664441e-01 1.93942979e-01 4.94222969e-01 6.18101954e-01
4.09799516e-01 2.28017345e-02 3.69001269e-01 -2.48711959e-01
-1.09775865e+00 -5.91900587e-01 -6.35449409e-01 1.09360412e-01
1.22165158e-01 -9.76219401e-02 -4.07088071e-01 2.89281338e-01] | [8.00947380065918, 7.612479209899902] |
830224fc-ccad-4cbb-9755-35a9dd4f574b | visual-abstraction-and-reasoning-through | 2303.04091 | null | https://arxiv.org/abs/2303.04091v3 | https://arxiv.org/pdf/2303.04091v3.pdf | Abstract Visual Reasoning Enabled by Language | While artificial intelligence (AI) models have achieved human or even superhuman performance in many well-defined applications, they still struggle to show signs of broad and flexible intelligence. The Abstraction and Reasoning Corpus (ARC), a visual intelligence benchmark introduced by Fran\c{c}ois Chollet, aims to assess how close AI systems are to human-like cognitive abilities. Most current approaches rely on carefully handcrafted domain-specific program searches to brute-force solutions for the tasks present in ARC. In this work, we propose a general learning-based framework for solving ARC. It is centered on transforming tasks from the vision to the language domain. This composition of language and vision allows for pre-trained models to be leveraged at each stage, enabling a shift from handcrafted priors towards the learned priors of the models. While not yet beating state-of-the-art models on ARC, we demonstrate the potential of our approach, for instance, by solving some ARC tasks that have not been solved previously. | ['Roger Wattenhofer', 'Joël Mathys', 'Benjamin Estermann', 'Loic Houmard', 'Giacomo Camposampiero'] | 2023-03-07 | null | null | null | null | ['visual-reasoning', 'visual-reasoning'] | ['computer-vision', 'reasoning'] | [ 2.27876961e-01 3.50983977e-01 -5.51139005e-02 -2.19264373e-01
-2.59717643e-01 -7.17479587e-01 1.27707827e+00 -7.04989657e-02
-3.91778916e-01 3.57093155e-01 2.88888037e-01 -4.53829229e-01
-2.55935967e-01 -5.15751958e-01 -6.35987997e-01 -2.31608510e-01
1.12760983e-01 8.55435014e-01 2.00423211e-01 -3.12521279e-01
4.02865082e-01 4.97708499e-01 -1.59298372e+00 5.47471702e-01
8.60947430e-01 6.53840005e-01 1.36782706e-01 7.94397831e-01
-2.14586705e-01 1.34644604e+00 -3.94368887e-01 -5.17315030e-01
3.13916296e-01 -2.29185626e-01 -9.54360962e-01 8.30811784e-02
5.78689039e-01 -7.39178201e-03 -1.23953030e-01 1.02823389e+00
-3.07690203e-02 1.14838115e-03 8.10745478e-01 -1.47722018e+00
-8.85380328e-01 5.47441363e-01 -5.10148644e-01 6.97470680e-02
5.87681890e-01 5.68900406e-01 1.11273193e+00 -7.97288299e-01
6.86825335e-01 1.42319322e+00 5.41178703e-01 6.53797269e-01
-1.34597528e+00 -3.56921464e-01 2.46732637e-01 4.26129729e-01
-1.04055500e+00 -3.89193475e-01 7.45400012e-01 -8.88464510e-01
1.32927608e+00 3.63971926e-02 7.99715102e-01 9.77401793e-01
-9.30688977e-02 9.74401414e-01 1.12558663e+00 -6.41182184e-01
2.65600264e-01 1.40712440e-01 1.75150603e-01 9.08370078e-01
1.58823833e-01 1.13639086e-01 -4.85908777e-01 4.24322188e-02
7.92151868e-01 -6.43908978e-02 -2.50241637e-01 -9.58666384e-01
-1.38758934e+00 7.73057222e-01 4.91039246e-01 5.03482223e-01
-5.24804235e-01 2.01068297e-01 2.91111857e-01 1.72824666e-01
-1.36017278e-01 1.01864624e+00 -4.35308993e-01 6.84601348e-03
-1.01890373e+00 4.42983419e-01 9.50544059e-01 9.30832267e-01
5.95125020e-01 -2.03520656e-02 -4.24293220e-01 4.04439658e-01
3.94567251e-01 7.06214979e-02 3.10234368e-01 -1.21851587e+00
3.03191930e-01 1.06839502e+00 1.07477263e-01 -8.04377973e-01
-4.47325796e-01 -3.24900597e-01 -3.25761169e-01 6.04458630e-01
6.28592372e-01 2.33733542e-02 -1.10909736e+00 1.64386046e+00
7.25296559e-03 -1.94440573e-03 3.95187020e-01 8.09158742e-01
6.11979306e-01 5.74146390e-01 3.14621836e-01 2.28043109e-01
1.50310552e+00 -1.34020841e+00 -9.49604362e-02 -7.51693189e-01
2.60933042e-01 -4.12698507e-01 1.36663747e+00 7.62849927e-01
-1.35777354e+00 -4.87097472e-01 -1.16255438e+00 -3.12821388e-01
-3.19638491e-01 -1.75830916e-01 7.78409421e-01 4.49170798e-01
-1.25239348e+00 2.73946822e-01 -8.27319324e-01 -4.85752583e-01
5.06830871e-01 1.08751878e-01 -2.63965458e-01 -2.76368469e-01
-7.24583864e-01 1.38851213e+00 5.44129789e-01 -3.20244193e-01
-1.25888002e+00 -7.88766265e-01 -8.59394491e-01 2.10394040e-01
6.56179368e-01 -1.00901604e+00 1.40269744e+00 -1.17629778e+00
-1.29345238e+00 1.14105213e+00 7.99843669e-02 -7.12040842e-01
5.75335920e-01 -2.12387189e-01 -1.04497343e-01 2.92499691e-01
-9.72613394e-02 9.45793927e-01 8.64999473e-01 -1.46232033e+00
-6.23848319e-01 -3.73447776e-01 5.40925682e-01 1.38965845e-01
-2.75670588e-01 1.93304405e-01 -7.61646807e-01 -5.59374571e-01
-4.34219211e-01 -8.56500804e-01 -1.84743032e-01 1.41207397e-01
-1.28100410e-01 -3.85501593e-01 5.33085823e-01 -5.79059899e-01
9.29001570e-01 -1.95724869e+00 7.23307014e-01 -2.94311605e-02
4.28881139e-01 4.75752622e-01 -2.44360775e-01 2.59098321e-01
7.87621662e-02 -9.99998152e-02 -3.45307052e-01 -3.20724636e-01
5.04266262e-01 3.30249518e-01 -2.69624859e-01 2.36085951e-01
2.89056838e-01 9.83460903e-01 -9.55437064e-01 -5.14622808e-01
2.17448249e-01 5.24999976e-01 -7.64812350e-01 2.62195826e-01
-7.16071844e-01 4.26793247e-01 -3.40516120e-01 5.16287625e-01
1.01907887e-01 -3.35946739e-01 1.89668670e-01 7.32371584e-02
-3.08025512e-03 1.08504863e-02 -8.15364122e-01 1.85338950e+00
-4.52623814e-01 5.62029481e-01 1.56605020e-01 -1.16144645e+00
7.08018661e-01 2.32366264e-01 1.52538225e-01 -9.02098596e-01
-8.86842087e-02 -1.38311610e-01 2.57277012e-01 -6.65451467e-01
1.67990118e-01 -1.27170786e-01 -6.96657225e-02 3.61220956e-01
3.44381869e-01 -5.60044885e-01 5.08507967e-01 5.72617091e-02
1.14315224e+00 4.57933277e-01 6.19481981e-01 -2.81639636e-01
7.97178090e-01 4.33701426e-01 4.41270679e-01 7.80714631e-01
-1.87010840e-01 4.83399093e-01 5.88461816e-01 -6.78381503e-01
-1.10872066e+00 -1.01537180e+00 3.10119271e-01 1.34465182e+00
-2.19387487e-01 -4.17383730e-01 -9.58549380e-01 -6.29550874e-01
-8.21173266e-02 1.02816832e+00 -6.13125265e-01 -7.69830793e-02
-5.29942632e-01 -3.01392049e-01 6.06533170e-01 6.55423820e-01
5.76264977e-01 -1.44994009e+00 -1.23754740e+00 1.15733825e-01
3.27814579e-01 -1.24085712e+00 1.02871649e-01 2.87412778e-02
-4.64716434e-01 -1.03494442e+00 -6.68028057e-01 -6.27415895e-01
4.21264946e-01 -8.69992301e-02 1.72099030e+00 1.45751104e-01
-4.16635424e-01 8.45418394e-01 -1.62706047e-01 -7.70229638e-01
-3.33008766e-01 -1.05222138e-02 -2.31106788e-01 -3.13980311e-01
5.09118915e-01 -4.93446827e-01 -4.35627609e-01 5.55791706e-02
-8.44852567e-01 2.10622028e-01 7.18754351e-01 6.93167269e-01
2.42959470e-01 -7.85968974e-02 1.81851491e-01 -7.36223757e-01
7.80417144e-01 -3.17772567e-01 -7.07504988e-01 5.15092373e-01
-5.48559487e-01 2.31863946e-01 7.07836747e-01 -3.49812090e-01
-1.02203131e+00 2.02733546e-01 2.56613404e-01 -5.21446347e-01
-5.21011949e-01 5.20065308e-01 -1.56317636e-01 -7.83941709e-03
8.03849697e-01 1.03276931e-01 -1.68578252e-01 -2.53207862e-01
6.54425800e-01 1.27948910e-01 9.94467437e-01 -1.01946664e+00
8.02667797e-01 2.91103482e-01 -2.01393306e-01 -5.55660367e-01
-8.29202056e-01 -1.21596701e-01 -6.22046173e-01 4.29111421e-02
1.27964759e+00 -7.08983302e-01 -7.36860752e-01 9.26743373e-02
-1.10743642e+00 -5.09304345e-01 -2.59280354e-01 6.28632456e-02
-9.03270602e-01 2.17712879e-01 -2.52430588e-01 -6.45655215e-01
-1.21703342e-01 -1.28237939e+00 7.28067040e-01 3.76442075e-01
-4.82412428e-01 -1.06633747e+00 2.38249570e-01 6.30584598e-01
4.91433203e-01 2.16649294e-01 1.21304011e+00 -8.36111546e-01
-5.69308460e-01 -7.04356357e-02 -3.47404987e-01 1.39877439e-01
-4.18790519e-01 -6.65694922e-02 -1.03958023e+00 -2.29617581e-01
3.90829220e-02 -5.17129600e-01 8.77063453e-01 4.42791777e-03
1.07204080e+00 -2.29718745e-01 -2.65878588e-01 5.94007969e-01
1.25801325e+00 3.89960974e-01 5.79765201e-01 6.68546975e-01
6.18152320e-01 6.66591704e-01 2.47051165e-01 2.09579065e-01
6.57831907e-01 5.76974988e-01 6.55697763e-01 -8.94212201e-02
-1.28967613e-01 -1.58841074e-01 2.38434568e-01 -7.81831369e-02
-4.24478590e-01 1.98536040e-03 -1.58930302e+00 7.19731212e-01
-2.05771041e+00 -9.85336840e-01 4.36494023e-01 1.77461684e+00
8.07743013e-01 4.08551395e-01 1.42799973e-01 4.58625890e-02
1.58561543e-01 7.91505724e-02 -6.53505504e-01 -5.63776731e-01
2.61806637e-01 3.82277578e-01 -9.14720148e-02 4.51965392e-01
-1.04003298e+00 1.14077210e+00 6.56586123e+00 3.22136760e-01
-8.95158172e-01 -1.20396391e-01 2.09674269e-01 -1.68473157e-03
-3.36161792e-01 1.67292029e-01 -5.25555551e-01 -6.66327551e-02
6.76735699e-01 -1.33116797e-01 9.53315198e-01 9.94826555e-01
-4.01051909e-01 6.87879324e-02 -1.61533713e+00 8.11231494e-01
3.63964438e-01 -1.30709291e+00 6.79197609e-02 -8.45073164e-02
5.70396364e-01 -9.19503625e-03 9.77777094e-02 8.13285053e-01
7.87282586e-01 -1.58702755e+00 9.00403976e-01 6.86141193e-01
2.66687989e-01 -5.30522764e-01 2.73059279e-01 5.71753561e-01
-8.44562829e-01 -3.46162826e-01 -1.72699198e-01 -3.33648533e-01
-2.61380911e-01 1.21729858e-01 -8.86864662e-01 2.92650670e-01
6.91537201e-01 5.65642893e-01 -9.49575245e-01 9.33154285e-01
-5.41079283e-01 2.19841391e-01 2.88941748e-02 5.51128574e-02
4.73789394e-01 -3.54383588e-02 5.12692034e-01 1.33548295e+00
7.07746716e-04 8.87740999e-02 3.35600644e-01 1.28281915e+00
2.06941515e-01 -3.33056062e-01 -7.65673101e-01 -3.42199713e-01
2.05873176e-01 1.15373588e+00 -4.64110434e-01 -3.03824633e-01
-7.74475217e-01 6.71394587e-01 6.00421369e-01 4.79633719e-01
-7.59893060e-01 -1.68568537e-01 5.37362158e-01 9.55751073e-03
3.57725054e-01 -4.62514281e-01 -3.95565093e-01 -1.27053595e+00
-2.22359687e-01 -1.43340409e+00 5.27376175e-01 -1.01390421e+00
-1.23813355e+00 4.98717159e-01 1.67465612e-01 -6.48502290e-01
-6.28674805e-01 -9.72841084e-01 -6.66934192e-01 7.09464312e-01
-1.49517822e+00 -1.40733576e+00 -3.32887203e-01 9.23670769e-01
5.09900510e-01 -5.32684922e-01 9.22754407e-01 -1.70535192e-01
-5.05447030e-01 2.56279439e-01 -6.40866399e-01 1.60646737e-01
4.31189120e-01 -1.46434867e+00 4.77298200e-01 9.75519836e-01
4.18573409e-01 6.14037573e-01 1.00180781e+00 -1.43017247e-01
-1.58426368e+00 -5.96984684e-01 6.20524526e-01 -8.13356221e-01
8.01876366e-01 -3.11883539e-01 -8.30297709e-01 1.03704321e+00
5.71685016e-01 -5.22912443e-02 2.51902431e-01 2.61897445e-02
-8.11557591e-01 1.47562459e-01 -1.06206596e+00 8.63910913e-01
1.00608492e+00 -6.34045243e-01 -1.30410111e+00 1.87932000e-01
4.80522603e-01 -2.52815306e-01 -6.52349114e-01 3.58399451e-01
4.38784391e-01 -1.16004825e+00 1.39034843e+00 -9.33420658e-01
7.84882307e-01 -4.38529015e-01 -2.82373875e-01 -1.13159239e+00
-4.48494971e-01 -4.59223062e-01 -3.44542354e-01 1.21995187e+00
3.68650377e-01 -3.19923997e-01 6.93711340e-01 8.55614066e-01
-1.96757212e-01 -5.13036609e-01 -5.61222196e-01 -5.34313440e-01
2.40134388e-01 -4.70057309e-01 5.23184001e-01 1.01954043e+00
2.01887950e-01 5.36264122e-01 6.26977906e-02 9.27596614e-02
6.08470619e-01 1.41302064e-01 9.89414632e-01 -1.47051942e+00
-6.08814180e-01 -8.87995720e-01 -3.76797855e-01 -5.39938688e-01
4.06267762e-01 -8.93727124e-01 -5.09873293e-02 -1.80021536e+00
3.12505573e-01 -1.81137249e-02 -2.05229148e-01 8.65713239e-01
-1.28935859e-01 2.39142869e-02 5.25865257e-01 5.82571588e-02
-7.49791622e-01 9.90018770e-02 1.14850426e+00 -4.21054661e-01
-7.62127638e-02 -3.43394011e-01 -1.01530397e+00 1.22044325e+00
7.27366030e-01 5.21889739e-02 -5.99819362e-01 -8.04602623e-01
3.75110745e-01 -9.21265129e-03 6.17734075e-01 -1.38955629e+00
5.75783789e-01 -3.88110727e-01 3.35358769e-01 -6.96186647e-02
2.82445729e-01 -9.06840503e-01 -6.76261783e-02 4.37134147e-01
-3.79145890e-01 2.17540085e-01 2.83577114e-01 3.32293034e-01
-2.06960410e-01 -2.44281471e-01 8.39237034e-01 -6.14256918e-01
-1.21868539e+00 -1.71499141e-02 -1.99470013e-01 2.65254617e-01
1.24186170e+00 -1.43965572e-01 -5.66474855e-01 -2.49069482e-01
-7.24609017e-01 4.43599045e-01 7.65513837e-01 4.29660678e-01
4.17264491e-01 -7.14408517e-01 -5.53592145e-01 5.44986911e-02
3.93031061e-01 -2.33265553e-02 -2.10198626e-01 5.33072293e-01
-5.79868793e-01 5.54667950e-01 -6.70851290e-01 -5.29313624e-01
-1.01403868e+00 1.13152790e+00 4.31103230e-01 -5.15568256e-01
-7.41136670e-01 8.46394062e-01 2.93509632e-01 -5.21146655e-01
4.86473083e-01 -2.85401076e-01 -3.92114669e-01 -1.74934968e-01
5.88305891e-01 6.00846484e-02 -1.21236004e-01 -2.85358369e-01
-4.09513891e-01 5.95306158e-01 -5.38327508e-02 -2.63575643e-01
1.45408010e+00 3.02726269e-01 -1.28800616e-01 1.25083789e-01
4.53901440e-01 -3.05215716e-01 -1.48662126e+00 -2.11709842e-01
4.45564628e-01 -1.17093667e-01 -1.30810114e-02 -1.16361713e+00
-8.61795068e-01 1.22258067e+00 2.96825439e-01 1.30334347e-01
1.10510492e+00 1.15792230e-01 2.50846982e-01 6.94462776e-01
3.18319112e-01 -7.68651307e-01 3.25687617e-01 7.11884737e-01
1.09829783e+00 -1.11703730e+00 1.94623508e-02 5.42252026e-02
-9.72288847e-01 9.77778792e-01 7.63749123e-01 -2.72675157e-01
1.81528866e-01 3.87390792e-01 -9.41803828e-02 -4.19325680e-01
-9.34276521e-01 -2.71878749e-01 4.48157817e-01 9.58948612e-01
3.72849554e-01 -1.42311156e-01 1.58838362e-01 5.87061286e-01
-3.23883861e-01 2.45699495e-01 2.52338231e-01 1.02534592e+00
-4.55066502e-01 -8.71929646e-01 -4.59481776e-01 -7.35789239e-02
-2.62543291e-01 -8.69557336e-02 -7.86264122e-01 9.96146202e-01
2.24139512e-01 6.75367594e-01 -2.42882315e-02 -3.08170356e-02
4.84147787e-01 3.69880080e-01 8.66368055e-01 -6.80816352e-01
-7.53937423e-01 -3.74965221e-01 7.40140826e-02 -6.82793081e-01
-4.56643999e-01 -5.22615254e-01 -1.18370509e+00 -7.84625933e-02
5.47958314e-01 -1.35038361e-01 5.34467459e-01 1.06026673e+00
2.05642387e-01 5.57551086e-01 -3.16765934e-01 -1.20508516e+00
-3.86232585e-01 -5.73230505e-01 -8.17478597e-02 5.10254562e-01
2.26881802e-01 -6.62382782e-01 -2.58922074e-02 1.92271128e-01] | [10.586837768554688, 2.1928224563598633] |
7bec0dd0-b647-4bda-a8b3-3e62e2b54304 | autokge-searching-scoring-functions-for | 1904.11682 | null | https://arxiv.org/abs/1904.11682v3 | https://arxiv.org/pdf/1904.11682v3.pdf | AutoSF: Searching Scoring Functions for Knowledge Graph Embedding | Scoring functions (SFs), which measure the plausibility of triplets in knowledge graph (KG), have become the crux of KG embedding. Lots of SFs, which target at capturing different kinds of relations in KGs, have been designed by humans in recent years. However, as relations can exhibit complex patterns that are hard to infer before training, none of them can consistently perform better than others on existing benchmark data sets. In this paper, inspired by the recent success of automated machine learning (AutoML), we propose to automatically design SFs (AutoSF) for distinct KGs by the AutoML techniques. However, it is non-trivial to explore domain-specific information here to make AutoSF efficient and effective. We firstly identify a unified representation over popularly used SFs, which helps to set up a search space for AutoSF. Then, we propose a greedy algorithm to search in such a space efficiently. The algorithm is further sped up by a filter and a predictor, which can avoid repeatedly training SFs with same expressive ability and help removing bad candidates during the search before model training. Finally, we perform extensive experiments on benchmark data sets. Results on link prediction and triplets classification show that the searched SFs by AutoSF, are KG dependent, new to the literature, and outperform the state-of-the-art SFs designed by humans. | ['Yongqi Zhang', 'Quanming Yao', 'Lei Chen', 'Wenyuan Dai'] | 2019-04-26 | null | null | null | null | ['link-property-prediction'] | ['graphs'] | [-1.87199712e-01 2.61314601e-01 -6.56322539e-01 -2.24290743e-01
-8.11623260e-02 -3.40383321e-01 4.49272811e-01 2.49293089e-01
1.07437529e-01 9.14598167e-01 1.58009365e-01 -3.36180329e-01
-7.70019889e-01 -1.17894816e+00 -6.59742117e-01 -5.17542124e-01
-2.66662419e-01 6.59492791e-01 5.57491601e-01 -3.21857125e-01
-6.28235340e-02 3.01537007e-01 -1.38157582e+00 1.55697241e-01
1.12249339e+00 1.01060367e+00 6.82970211e-02 -1.52837604e-01
-2.96494037e-01 6.54204786e-01 -4.47765350e-01 -8.11334074e-01
9.56588387e-02 -3.61123860e-01 -1.04712319e+00 -2.94175267e-01
6.05156012e-02 2.30837911e-01 -5.01648605e-01 9.61511493e-01
3.17772448e-01 -1.99875280e-01 8.07125866e-01 -1.51548493e+00
-8.74895215e-01 1.01617658e+00 -2.88767159e-01 -1.67497888e-01
5.52343905e-01 -1.83612749e-01 1.55445004e+00 -7.41400301e-01
7.11801410e-01 1.30831444e+00 6.01267219e-01 2.81116366e-01
-1.24010205e+00 -7.56249607e-01 8.85958672e-02 7.69893825e-01
-1.58403718e+00 -1.28333405e-01 9.75350559e-01 -3.04733276e-01
7.26412952e-01 5.14443278e-01 1.01036727e+00 1.03242254e+00
-1.22982621e-01 1.04757059e+00 8.21705699e-01 -6.06536329e-01
-3.33141722e-02 2.23505259e-01 2.88978308e-01 8.92834485e-01
6.72508597e-01 4.23184671e-02 -6.81436956e-01 -3.05809766e-01
5.85419893e-01 -2.85541177e-01 -6.00199461e-01 -6.71137452e-01
-1.22635186e+00 7.72040188e-01 6.11371398e-01 3.36014509e-01
-5.22457436e-02 -2.59504944e-01 1.81401059e-01 3.65229815e-01
1.60191983e-01 6.81483686e-01 -6.10802531e-01 3.74642134e-01
-4.74127531e-01 1.97462827e-01 8.84073257e-01 9.58678246e-01
6.99842095e-01 -4.40945625e-01 -3.05043042e-01 8.65253329e-01
4.21137720e-01 1.13456495e-01 3.47186476e-01 -2.75098234e-01
5.17486572e-01 1.25296795e+00 -2.05248281e-01 -1.35967255e+00
-3.83057266e-01 -7.67875195e-01 -9.42166567e-01 -3.59447747e-01
2.46564716e-01 2.86974221e-01 -4.20025498e-01 1.62935364e+00
4.61965799e-01 2.35122383e-01 1.06845107e-02 7.15296745e-01
7.98749685e-01 4.00250524e-01 -1.98974386e-02 -3.85109097e-01
1.15606391e+00 -9.81356025e-01 -6.01053476e-01 1.13558233e-01
9.47726011e-01 -4.25431997e-01 1.12218308e+00 4.72751379e-01
-5.76112449e-01 -3.99680883e-01 -1.20901513e+00 1.94852531e-01
-5.94572365e-01 1.56133011e-01 1.10373330e+00 4.74782407e-01
-7.33839869e-01 7.88786113e-01 -3.93159360e-01 -3.60324204e-01
4.12151664e-01 4.22130287e-01 -4.36848283e-01 -6.91802055e-02
-1.95726132e+00 1.03119183e+00 9.64692593e-01 6.19684868e-02
-3.09185803e-01 -5.41370571e-01 -6.61547482e-01 1.99551359e-01
9.31642652e-01 -9.36922550e-01 5.12045443e-01 -4.81927872e-01
-1.28839517e+00 5.13194859e-01 1.71585709e-01 -4.82019186e-01
2.52863079e-01 -1.51094511e-01 -9.67254639e-01 -1.69879347e-02
-1.34678744e-02 2.46430144e-01 6.83919251e-01 -1.22398496e+00
-4.18659896e-01 -5.70564829e-02 1.75265431e-01 7.91770071e-02
-7.86626279e-01 -1.57254919e-01 -4.69157010e-01 -5.26760340e-01
2.25559905e-01 -7.03748643e-01 5.48615418e-02 -2.06664145e-01
-8.35315883e-01 -7.35306025e-01 5.00490606e-01 -3.36141586e-01
1.95062065e+00 -1.80265784e+00 3.30405980e-01 7.12441325e-01
3.06967944e-01 6.15054250e-01 3.13684009e-02 6.43657148e-01
-7.88506791e-02 2.42198810e-01 1.29858986e-01 1.63644567e-01
2.05032527e-01 4.42979634e-01 -2.17540458e-01 6.78531677e-02
7.20862672e-02 1.08623266e+00 -1.04221106e+00 -9.43005741e-01
-1.94274914e-03 -4.90102023e-02 -3.40893149e-01 3.76448631e-02
-3.89797211e-01 4.36647236e-02 -7.11447179e-01 8.42172205e-01
4.43512648e-01 -5.62062621e-01 5.44256151e-01 -6.68471336e-01
2.99199849e-01 3.46454948e-01 -1.32018197e+00 1.37338352e+00
8.23509544e-02 -7.53245875e-02 -5.99809170e-01 -1.23169947e+00
1.13069713e+00 1.40409961e-01 3.40361893e-01 -3.48322809e-01
-8.03822204e-02 3.74658048e-01 1.91030324e-01 -4.29755270e-01
1.60339087e-01 1.36341661e-01 9.09692273e-02 7.43014440e-02
7.92408660e-02 3.74166727e-01 5.03997743e-01 3.56340498e-01
1.18108094e+00 1.82515401e-02 5.79554975e-01 -4.89779189e-02
8.11042845e-01 -2.33433247e-02 6.86224341e-01 4.93082941e-01
1.53789222e-01 8.76914635e-02 6.41276777e-01 -5.29408991e-01
-4.61106867e-01 -9.22058105e-01 -2.08650634e-01 6.81950867e-01
3.64513457e-01 -1.05124986e+00 -1.90072253e-01 -1.23142862e+00
4.14621174e-01 3.30581993e-01 -3.99740785e-01 -4.67830688e-01
-3.77985805e-01 -6.09746277e-01 6.33474827e-01 3.06254596e-01
5.11014163e-01 -9.62181687e-01 -9.90192592e-03 3.18650961e-01
-1.18736081e-01 -1.01875854e+00 -1.34494051e-01 7.45329708e-02
-6.11042023e-01 -1.31841803e+00 -2.41471469e-01 -6.24733984e-01
5.51452935e-01 1.14164554e-01 1.30554152e+00 4.04653162e-01
-9.61720720e-02 -1.62709638e-01 -7.09396660e-01 -8.16172585e-02
-6.06018081e-02 3.68460596e-01 2.07536489e-01 8.85771364e-02
3.59965891e-01 -7.36525118e-01 -4.76059288e-01 5.65226018e-01
-6.12896562e-01 2.38808706e-01 1.06468987e+00 9.43748534e-01
5.50082505e-01 4.82728034e-01 6.95304275e-01 -1.09829640e+00
6.50013566e-01 -4.16105062e-01 -3.89719367e-01 7.92284489e-01
-1.13527620e+00 4.04365212e-01 7.48984456e-01 -2.21497327e-01
-5.70279360e-01 -2.47671306e-01 1.52863204e-01 -3.97435993e-01
1.10806286e-01 9.57856655e-01 -4.77955699e-01 -3.60973209e-01
6.28492236e-01 1.61795437e-01 -3.57973009e-01 -6.53824151e-01
3.45767409e-01 4.31127846e-01 2.12241307e-01 -7.64639378e-01
1.05675626e+00 -6.61317036e-02 3.01989108e-01 -3.28071207e-01
-1.03234065e+00 -4.12155002e-01 -3.82835239e-01 -1.15677774e-01
3.64393532e-01 -5.34258485e-01 -7.80609012e-01 -7.37567768e-02
-9.59837973e-01 2.80026108e-01 -1.17123924e-01 3.53367180e-01
-2.02180415e-01 5.06750643e-01 -2.86935955e-01 -5.61582744e-01
-3.33324581e-01 -6.79615974e-01 6.77033484e-01 1.41763300e-01
-2.72532970e-01 -8.70862246e-01 2.10447796e-02 4.04572368e-01
6.33366182e-02 1.84192702e-01 1.30134678e+00 -7.68895864e-01
-7.75140047e-01 -8.29448700e-02 -1.92013994e-01 1.74961880e-01
2.70599782e-01 -7.71806166e-02 -4.39386517e-01 -5.33052981e-02
-6.22158706e-01 -2.19718531e-01 9.11186695e-01 -1.80156901e-01
1.27855086e+00 -5.43472230e-01 -9.65130627e-01 5.49013317e-01
1.27620661e+00 -3.75698134e-03 5.98432958e-01 4.87089545e-01
7.25974143e-01 3.74544591e-01 9.08988059e-01 9.04957876e-02
5.55875659e-01 9.38475907e-01 3.42590660e-01 9.12274718e-02
-7.40374327e-02 -7.57887781e-01 1.42302424e-01 9.38171625e-01
-3.41839582e-01 -1.31341010e-01 -7.58766949e-01 4.73363578e-01
-2.15408444e+00 -8.32147956e-01 -2.38086194e-01 1.99964762e+00
1.18590677e+00 4.34798151e-01 1.12961881e-01 6.98819339e-01
6.02867723e-01 4.12730984e-02 -2.44953230e-01 8.26798156e-02
-3.01218927e-01 1.67343542e-01 3.97250563e-01 2.47055426e-01
-1.05061817e+00 1.04790568e+00 5.17717791e+00 1.33053029e+00
-7.03384936e-01 -2.40814313e-01 1.24366194e-01 3.28349978e-01
-4.95332748e-01 3.53919685e-01 -9.42091703e-01 4.99615490e-01
4.87825006e-01 -2.76812077e-01 2.96642274e-01 8.55896771e-01
-2.60133088e-01 2.18634307e-01 -1.12435997e+00 8.78743351e-01
-1.86364189e-01 -1.33981109e+00 4.51944143e-01 -1.02154739e-01
5.58819473e-01 -5.46847224e-01 -3.10255587e-01 4.87948745e-01
3.18769604e-01 -9.30535853e-01 3.50326210e-01 5.27627587e-01
4.46500182e-01 -6.30135059e-01 6.48362100e-01 4.56042200e-01
-1.35103464e+00 -8.98441020e-03 -6.03027999e-01 1.15626745e-01
-6.81219026e-02 9.66502845e-01 -8.62782598e-01 1.39807379e+00
4.60526496e-01 8.93163204e-01 -7.97996819e-01 1.15593517e+00
-7.41007924e-01 5.39625645e-01 -2.85778672e-01 -3.83902282e-01
-3.99822816e-02 -1.05189130e-01 4.71889526e-01 1.02877939e+00
1.53056875e-01 1.21909417e-01 2.77201563e-01 6.39612377e-01
-6.32705390e-02 3.53354484e-01 -5.71823418e-01 -1.02037825e-01
6.66819334e-01 1.31953609e+00 -4.74011362e-01 -1.68578461e-01
-2.21307606e-01 5.45292616e-01 6.18460596e-01 3.01111609e-01
-7.46239603e-01 -3.92084986e-01 2.86882877e-01 1.81732401e-01
2.03421459e-01 5.14026992e-02 -7.67308101e-02 -1.23395038e+00
1.75037280e-01 -7.27327466e-01 7.86053300e-01 -5.66206932e-01
-1.61641848e+00 5.54488361e-01 1.16542190e-01 -1.31238115e+00
-2.58019362e-02 -5.42998910e-01 -2.95959324e-01 5.59015989e-01
-1.62302613e+00 -1.20361781e+00 -2.41147682e-01 7.92047620e-01
-5.82981855e-02 -2.47914165e-01 8.09827089e-01 3.79978150e-01
-6.23311877e-01 7.69198358e-01 -1.35228083e-01 8.06649402e-02
6.52322471e-01 -1.16753829e+00 1.89253055e-02 5.27984619e-01
6.77747071e-01 9.59903359e-01 5.44291794e-01 -7.57938981e-01
-1.36622226e+00 -9.44322526e-01 1.13210273e+00 -1.62832975e-01
7.76941717e-01 -1.35656685e-01 -1.07956576e+00 4.72230464e-01
-2.33004928e-01 1.57657683e-01 7.24903464e-01 7.48184204e-01
-4.94285375e-01 -5.01055658e-01 -8.58535111e-01 5.77899337e-01
1.66095316e+00 -2.85991251e-01 -6.94859266e-01 4.83180821e-01
6.35710061e-01 -1.06119849e-01 -9.77330625e-01 9.92462814e-01
6.19353056e-01 -8.47866714e-01 1.10869777e+00 -8.72307718e-01
2.75810242e-01 -5.27161837e-01 1.17285617e-01 -1.43231368e+00
-5.98280132e-01 -5.31273484e-01 -7.91445017e-01 1.38306701e+00
5.10636270e-01 -8.09692562e-01 8.40300083e-01 2.01842993e-01
-5.05000502e-02 -1.28389227e+00 -7.92279840e-01 -1.15435576e+00
-5.36472321e-01 -1.63597882e-01 8.83312404e-01 1.20388937e+00
4.13278878e-01 5.61490595e-01 -3.72269332e-01 1.61469638e-01
5.60470462e-01 5.27096629e-01 7.21731722e-01 -1.68168867e+00
-5.82313120e-01 -5.56157410e-01 -7.26118922e-01 -9.03512716e-01
1.82687566e-01 -1.19902444e+00 -5.59298277e-01 -1.75668526e+00
1.23301700e-01 -6.87993348e-01 -5.69900334e-01 8.62744510e-01
-2.95698881e-01 -1.20638393e-01 -3.99960577e-02 4.34913963e-01
-7.20411241e-01 7.07186639e-01 1.34581983e+00 -2.28707358e-01
-1.49161778e-02 4.28435616e-02 -7.75976121e-01 6.81922376e-01
5.37650228e-01 -4.01195347e-01 -6.70999229e-01 -1.91129800e-02
6.22686505e-01 -9.64733586e-02 2.73233831e-01 -8.37492585e-01
3.10317010e-01 -2.65836418e-01 2.94826508e-01 -5.60139894e-01
1.14599973e-01 -7.98767090e-01 4.62651253e-01 3.84253711e-01
-1.79978222e-01 -5.61426938e-01 -3.14747155e-01 6.61722004e-01
-3.47004890e-01 -2.00952873e-01 2.86961406e-01 6.71560168e-02
-7.54253626e-01 4.39633161e-01 5.27058184e-01 -1.19649008e-01
1.00227010e+00 -1.06155284e-01 -4.20423836e-01 -7.29820728e-02
-7.69004345e-01 4.48861212e-01 6.30850047e-02 3.88119280e-01
6.02968335e-01 -1.71515274e+00 -4.40929323e-01 5.73030077e-02
3.77376169e-01 -5.36106974e-02 -1.34139791e-01 7.07806349e-01
-1.68841258e-01 5.97585082e-01 3.07382923e-02 -2.82795995e-01
-1.26162493e+00 7.39119589e-01 -2.17139404e-02 -7.73531616e-01
-5.40737927e-01 8.21436107e-01 -1.90900430e-01 -6.02768622e-02
2.19446734e-01 -3.61787200e-01 -5.00038743e-01 4.03444767e-02
6.78228810e-02 1.60651058e-01 1.38339758e-01 -1.97517037e-01
-5.36020756e-01 5.49344301e-01 -1.08043902e-01 4.32246268e-01
1.32151449e+00 1.58050492e-01 -3.12035620e-01 2.02353835e-01
8.83870602e-01 1.04331277e-01 -4.88029957e-01 -4.62490320e-01
4.46905285e-01 -5.99349797e-01 -3.03352475e-01 -9.26880121e-01
-1.02522480e+00 4.47109997e-01 2.03599464e-02 6.18753612e-01
1.09569478e+00 2.37927109e-01 7.50412822e-01 5.59787035e-01
7.23964810e-01 -8.83214712e-01 -1.51213231e-02 3.20924431e-01
8.76154423e-01 -9.39152837e-01 1.57013074e-01 -9.99105036e-01
-4.34592128e-01 1.01286376e+00 7.08326757e-01 1.75256774e-01
5.90842783e-01 -1.65864736e-01 -6.35387659e-01 -3.02422673e-01
-8.08205247e-01 -3.94718021e-01 7.03494608e-01 4.58205432e-01
1.40101686e-01 8.38682055e-02 -8.73715222e-01 1.02122915e+00
-2.83142537e-01 7.45533481e-02 -9.64642540e-02 5.98765016e-01
-4.48077559e-01 -1.65380764e+00 -6.93487078e-02 6.32592738e-01
2.42995974e-02 -7.49366730e-02 -6.69080138e-01 7.79990673e-01
4.70640182e-01 6.84295475e-01 -8.23761225e-01 -8.23362529e-01
4.05180603e-01 2.13535413e-01 5.56556344e-01 -5.06940782e-01
-2.73106605e-01 -4.46983308e-01 5.65995216e-01 -3.96576405e-01
-3.34279448e-01 -3.49271476e-01 -9.86602485e-01 -1.91581100e-01
-7.92547464e-01 4.64429110e-01 3.63970399e-02 1.01418841e+00
2.80272752e-01 3.44347954e-01 6.12861872e-01 5.40495478e-02
-4.87624645e-01 -9.26234007e-01 -7.18969762e-01 4.75181729e-01
-3.63323450e-01 -1.27013946e+00 -1.31313935e-01 -3.78976166e-01] | [8.805641174316406, 7.883284568786621] |
3780563e-01f5-425d-8336-8c3cea03669d | lcd-learned-cross-domain-descriptors-for-2d | 1911.09326 | null | https://arxiv.org/abs/1911.09326v1 | https://arxiv.org/pdf/1911.09326v1.pdf | LCD: Learned Cross-Domain Descriptors for 2D-3D Matching | In this work, we present a novel method to learn a local cross-domain descriptor for 2D image and 3D point cloud matching. Our proposed method is a dual auto-encoder neural network that maps 2D and 3D input into a shared latent space representation. We show that such local cross-domain descriptors in the shared embedding are more discriminative than those obtained from individual training in 2D and 3D domains. To facilitate the training process, we built a new dataset by collecting $\approx 1.4$ millions of 2D-3D correspondences with various lighting conditions and settings from publicly available RGB-D scenes. Our descriptor is evaluated in three main experiments: 2D-3D matching, cross-domain retrieval, and sparse-to-dense depth estimation. Experimental results confirm the robustness of our approach as well as its competitive performance not only in solving cross-domain tasks but also in being able to generalize to solve sole 2D and 3D tasks. Our dataset and code are released publicly at \url{https://hkust-vgd.github.io/lcd}. | ['Sai-Kit Yeung', 'Binh-Son Hua', 'Quang-Hieu Pham', 'Gemma Roig', 'Mikaela Angelina Uy', 'Duc Thanh Nguyen'] | 2019-11-21 | null | null | null | null | ['3d-point-cloud-matching'] | ['computer-vision'] | [-1.33493185e-01 -3.43507230e-01 -2.39788294e-01 -4.80562299e-01
-1.28078282e+00 -6.54507041e-01 5.68497241e-01 -1.77660391e-01
-1.96846575e-01 2.77301311e-01 8.88627470e-02 7.45621845e-02
-2.08108842e-01 -8.12948644e-01 -9.29010808e-01 -5.95374703e-01
1.00743338e-01 5.97973228e-01 -1.89444683e-02 3.82797141e-03
1.88459650e-01 9.46319520e-01 -1.67736042e+00 7.88728967e-02
3.52430582e-01 1.39575434e+00 2.24546239e-01 4.48878109e-01
-3.20661142e-02 2.38543600e-01 -2.03108266e-01 -2.12812364e-01
9.21457171e-01 -1.82821881e-02 -5.27929187e-01 1.00480607e-02
8.56427491e-01 -5.72747707e-01 -7.00496495e-01 9.01656508e-01
7.65337646e-01 1.03641942e-01 5.86439371e-01 -1.30216146e+00
-8.76790226e-01 -4.39735919e-01 -5.34568965e-01 -1.82779312e-01
6.34883523e-01 8.18516761e-02 1.04390490e+00 -1.31077170e+00
8.42403889e-01 1.24739122e+00 7.31735229e-01 5.22626698e-01
-1.06356478e+00 -8.37455630e-01 -1.72495082e-01 -1.28306538e-01
-1.73456776e+00 -3.62258524e-01 1.05909431e+00 -4.86113638e-01
1.19940054e+00 -1.47109225e-01 5.61460912e-01 1.12755835e+00
3.95126343e-02 7.06691265e-01 1.02312005e+00 -2.47913912e-01
-1.11972401e-02 -9.75269452e-02 -3.04201216e-01 8.75187099e-01
1.09591268e-01 2.92896658e-01 -8.26619744e-01 -3.23063999e-01
1.36473644e+00 2.28623450e-01 -1.05824418e-01 -9.51376557e-01
-1.26852572e+00 8.73232365e-01 6.48482502e-01 1.80849642e-01
-2.46496841e-01 2.75734693e-01 1.59606442e-01 3.02337199e-01
7.87375629e-01 2.07222357e-01 -4.12796110e-01 -1.87464520e-01
-6.07092917e-01 2.92683542e-01 5.50429761e-01 1.34429896e+00
1.15585589e+00 -2.83569068e-01 3.88023078e-01 9.09988821e-01
1.63818896e-01 9.87104475e-01 3.86225194e-01 -1.22709703e+00
4.78854835e-01 5.93256891e-01 2.82535385e-02 -1.03177071e+00
-1.72259912e-01 4.44802754e-02 -9.67387974e-01 3.62083942e-01
4.88844700e-02 3.53688151e-01 -7.95738876e-01 1.58293116e+00
2.77644336e-01 3.37508649e-01 -6.29549697e-02 9.07624066e-01
8.96468043e-01 4.50180322e-01 -4.99684393e-01 3.90967786e-01
9.63434458e-01 -7.55018711e-01 -2.90474206e-01 -1.70851350e-01
5.01556635e-01 -9.01843309e-01 7.56968379e-01 1.52298650e-02
-1.24194264e+00 -7.68553257e-01 -1.00045633e+00 -6.39034450e-01
-4.01721746e-01 6.39784113e-02 5.52373052e-01 7.22240433e-02
-1.03910565e+00 6.03519261e-01 -8.16890895e-01 -4.44721609e-01
5.19648433e-01 4.31415766e-01 -8.53886247e-01 -3.96275908e-01
-9.23902631e-01 7.57684827e-01 -1.26475682e-02 -3.02292496e-01
-7.03018904e-01 -6.20636165e-01 -1.05546796e+00 -2.97688127e-01
-4.97418381e-02 -9.35106575e-01 9.26749527e-01 -3.22746396e-01
-1.16116834e+00 1.60539544e+00 -3.69024903e-01 -1.99592616e-02
2.75751233e-01 -3.40295553e-01 5.13995513e-02 2.36431912e-01
4.94917154e-01 7.74640977e-01 7.34587550e-01 -1.27710176e+00
-4.05521005e-01 -7.62583017e-01 1.56865660e-02 2.18327895e-01
-4.54450585e-02 -3.45907480e-01 -7.37223983e-01 -4.08685237e-01
7.00322330e-01 -9.78398919e-01 -3.11339013e-02 5.63256383e-01
-1.61846966e-01 -2.63064563e-01 6.75460875e-01 -3.38701129e-01
5.49833119e-01 -2.32878947e+00 3.39287877e-01 2.01828539e-01
2.62650311e-01 -1.33226663e-01 -2.30577230e-01 4.42322761e-01
-9.43223387e-02 -1.50127441e-01 -8.35220292e-02 -8.77615094e-01
3.50178838e-01 3.48344564e-01 -3.01004201e-01 7.73403108e-01
2.76286781e-01 9.30452168e-01 -7.70230412e-01 -3.69686186e-01
4.87492144e-01 6.64982557e-01 -3.15818191e-01 4.10907805e-01
1.56518698e-01 5.16127050e-01 -5.27921081e-01 1.01318431e+00
9.47833955e-01 -4.51865613e-01 -4.10465300e-01 -2.81227857e-01
-9.70793292e-02 2.98468769e-01 -1.09823215e+00 2.50316048e+00
-5.99079669e-01 5.34855545e-01 -1.31898746e-01 -9.11148489e-01
1.24796975e+00 7.16859326e-02 8.66750419e-01 -1.01598275e+00
8.83651059e-03 5.93899310e-01 -7.69366264e-01 -2.48783246e-01
3.90618771e-01 -1.06274292e-01 -1.96101382e-01 5.06365895e-01
3.02716345e-01 -7.21978247e-01 -3.04914355e-01 -3.93253528e-02
9.70137894e-01 2.77035415e-01 2.67853379e-01 3.97764798e-03
3.98709357e-01 -2.40665495e-01 2.57914126e-01 5.07102847e-01
-4.52650376e-02 8.29585850e-01 2.11626723e-01 -6.29171133e-01
-1.34665775e+00 -1.34470701e+00 -3.12055826e-01 5.24824321e-01
4.21631724e-01 -2.31282249e-01 -1.02968149e-01 -3.39115739e-01
5.35151184e-01 2.68698111e-02 -5.81158102e-01 -1.35658458e-01
-4.81019258e-01 6.50328174e-02 5.10280788e-01 5.92914104e-01
5.70721030e-01 -5.37396610e-01 -5.10441184e-01 -2.46922269e-01
-2.04439834e-02 -1.33480525e+00 -4.38721240e-01 3.46542329e-01
-9.28014457e-01 -1.01863158e+00 -9.37053800e-01 -9.80794787e-01
6.80466890e-01 5.78340232e-01 1.17653251e+00 -1.93479002e-01
-3.77933025e-01 7.84491181e-01 -2.35589758e-01 -2.50354558e-01
6.42302483e-02 -8.13970491e-02 2.35207751e-01 -3.01968545e-01
7.76964128e-01 -8.63605738e-01 -6.49079561e-01 4.38296109e-01
-7.40902066e-01 -1.56441003e-01 5.33248961e-01 8.62984121e-01
1.02344584e+00 -4.07603472e-01 -1.73862651e-01 -1.75915703e-01
3.06052357e-01 -2.96711624e-01 -7.45506406e-01 -6.02627806e-02
-4.38782483e-01 1.33245036e-01 9.57326069e-02 -1.47642583e-01
-4.35951293e-01 3.31098199e-01 -2.03127235e-01 -1.19214666e+00
-3.62230390e-01 -7.05726538e-03 -7.78360516e-02 -3.52597028e-01
4.70298797e-01 3.64314556e-01 1.64999247e-01 -7.02420056e-01
2.77035594e-01 5.81086755e-01 4.37950671e-01 -7.21659601e-01
1.20342064e+00 7.63289988e-01 1.95830598e-01 -5.31101823e-01
-8.21469247e-01 -8.19840968e-01 -9.61104631e-01 1.15526855e-01
8.26872110e-01 -1.41869879e+00 -4.80272233e-01 4.79628772e-01
-1.26047671e+00 -1.40545964e-01 -4.32992876e-01 4.81780946e-01
-9.26883340e-01 3.83261651e-01 -3.22720766e-01 -4.13466007e-01
-2.06426427e-01 -1.21339989e+00 1.87058687e+00 -1.90248683e-01
1.23073243e-01 -9.86614823e-01 4.72078413e-01 1.03573032e-01
1.61050990e-01 4.10145909e-01 5.58680534e-01 -2.83166915e-01
-9.02339518e-01 -3.04555923e-01 -5.44865012e-01 5.23335218e-01
2.70295888e-01 -3.45627993e-01 -1.07067788e+00 -4.73950773e-01
-1.34237453e-01 -5.77206254e-01 6.96577907e-01 2.02122867e-01
1.19644594e+00 3.44591975e-01 -3.14641118e-01 1.19234073e+00
1.64822602e+00 -1.03140399e-01 4.46450442e-01 4.28491771e-01
5.72988093e-01 3.80813986e-01 5.54049730e-01 6.13185287e-01
4.05337900e-01 9.82724488e-01 4.73528475e-01 -1.81441307e-01
-2.19763801e-01 -4.45813298e-01 -6.97165204e-04 7.70603359e-01
-7.54926950e-02 1.04105540e-01 -1.01811826e+00 5.44061899e-01
-1.67691493e+00 -7.28375494e-01 2.50895917e-01 2.24552083e+00
5.40385127e-01 -1.61032453e-01 -1.65905774e-01 -1.99978709e-01
4.53094691e-01 2.29847416e-01 -7.45372295e-01 -7.52717033e-02
-4.37834531e-01 7.03677416e-01 5.73220491e-01 4.32736158e-01
-1.08860052e+00 7.58329809e-01 5.40760231e+00 5.66820860e-01
-1.02536249e+00 9.87367779e-02 2.41312400e-01 -2.36377120e-01
-4.96555209e-01 -1.87873110e-01 -7.51511097e-01 1.64459929e-01
3.89492303e-01 -1.95314035e-01 2.13851228e-01 8.32224369e-01
-3.74414831e-01 1.37656584e-01 -1.33052278e+00 1.74529409e+00
1.99766889e-01 -1.40914965e+00 -3.16399932e-02 2.91931868e-01
8.10180008e-01 4.23043758e-01 2.21789509e-01 8.12486857e-02
9.10280123e-02 -9.31303203e-01 5.70721924e-01 4.90234703e-01
1.35297644e+00 -5.91051877e-01 5.26651323e-01 1.22145705e-01
-1.20673299e+00 2.12984502e-01 -6.43270135e-01 1.02053471e-02
-4.89775091e-02 5.68690717e-01 -4.52443540e-01 6.28450871e-01
9.43281949e-01 1.20976651e+00 -3.04021627e-01 8.01850319e-01
-7.61247426e-02 -3.78349513e-01 -5.94664752e-01 2.44016111e-01
2.66987979e-01 -2.47402027e-01 4.80503947e-01 8.59844148e-01
7.03253746e-01 7.92027190e-02 -8.59097093e-02 9.54066515e-01
-1.94637984e-01 -1.78779036e-01 -1.33365226e+00 1.74020261e-01
5.27500570e-01 8.11151624e-01 -1.95837140e-01 -1.09155372e-01
-4.17987436e-01 1.24973226e+00 4.21515882e-01 3.18846196e-01
-7.77920783e-01 -4.91941810e-01 1.16143215e+00 -7.28978664e-02
5.17037213e-01 -6.86244547e-01 -2.37244368e-01 -1.29671323e+00
3.32003802e-01 -3.18399757e-01 8.79571885e-02 -1.01151037e+00
-1.68581676e+00 5.84232509e-01 9.24524739e-02 -1.68954682e+00
-2.08132863e-01 -9.72229242e-01 -8.54118764e-02 1.10690284e+00
-1.69479251e+00 -1.19823062e+00 -6.11903727e-01 9.95954692e-01
1.71330333e-01 -2.27751374e-01 1.04405177e+00 5.28490901e-01
-1.48565434e-02 6.18170917e-01 3.39423746e-01 1.74939975e-01
8.90939355e-01 -1.00438392e+00 6.20589614e-01 3.65399748e-01
2.25166991e-01 6.19187415e-01 1.66792452e-01 -2.35383689e-01
-1.63088429e+00 -7.10340738e-01 9.01361704e-01 -6.41117632e-01
5.95536351e-01 -5.95573545e-01 -5.65039814e-01 6.74095273e-01
-1.23227723e-01 3.44842345e-01 7.60506868e-01 -4.91123982e-02
-7.48048902e-01 -2.34709844e-01 -1.20667875e+00 6.64585084e-02
1.50696445e+00 -1.08994269e+00 -5.24724483e-01 5.34156084e-01
7.27620423e-01 -8.88663530e-01 -1.18063104e+00 3.59866112e-01
6.67211115e-01 -1.13260138e+00 1.49664950e+00 -3.07941377e-01
6.34798348e-01 -3.93016003e-02 -8.82600427e-01 -9.42732573e-01
-1.99820414e-01 -2.10197404e-01 -8.62282217e-02 7.72826731e-01
-1.51055737e-03 -7.83719242e-01 8.02843869e-01 4.17693526e-01
-1.01416975e-01 -8.79474401e-01 -1.30910933e+00 -9.27613735e-01
2.47474477e-01 -5.29642284e-01 7.72663593e-01 8.30614865e-01
-5.50151348e-01 -7.36681223e-02 -2.22149014e-01 3.85175496e-01
9.01342154e-01 6.93492770e-01 1.06894684e+00 -1.12472486e+00
-6.29753172e-02 -3.16658378e-01 -8.63131881e-01 -1.66023767e+00
3.54362726e-01 -1.01287603e+00 -2.55126655e-01 -1.47062707e+00
6.76075891e-02 -5.73607743e-01 -3.46869409e-01 3.88331503e-01
4.88058954e-01 5.00317514e-01 1.31749079e-01 4.37191129e-01
-5.48708856e-01 7.18040347e-01 1.25510871e+00 -6.24335743e-02
1.11162320e-01 -2.58611858e-01 -3.29625785e-01 3.16488832e-01
6.02132320e-01 -3.51341575e-01 -7.55249634e-02 -9.48301911e-01
7.40349814e-02 -9.78091266e-03 8.27467144e-01 -1.04651964e+00
2.45144695e-01 7.91096985e-02 5.58225751e-01 -8.41723919e-01
9.20949459e-01 -9.41835344e-01 -1.11619629e-04 8.56168717e-02
-1.15446374e-02 2.07559958e-01 2.63109714e-01 3.95859838e-01
-4.57996309e-01 1.47908434e-01 6.67970240e-01 -2.61973262e-01
-8.79606843e-01 8.79922152e-01 4.88367379e-01 -7.87926689e-02
9.07571375e-01 -4.57984596e-01 -7.01060370e-02 -4.39154565e-01
-6.29410386e-01 7.89226219e-02 8.97903800e-01 5.81933379e-01
1.07260787e+00 -1.82953572e+00 -5.23873031e-01 6.91909015e-01
5.65642953e-01 3.65048647e-01 2.21050099e-01 5.27359843e-01
-6.12887621e-01 6.50884748e-01 -4.67980564e-01 -1.06658125e+00
-1.00817358e+00 3.39198589e-01 4.50151742e-01 4.57389913e-02
-7.18656182e-01 1.07915127e+00 1.54851153e-01 -8.11844409e-01
3.33491176e-01 -2.20793709e-01 5.75544298e-01 -2.45047048e-01
2.44192213e-01 8.07887092e-02 2.04815678e-02 -7.34478891e-01
-4.22760546e-01 1.31573689e+00 3.06554615e-01 1.00798994e-01
1.62935114e+00 -1.10049836e-01 -1.73488528e-01 5.16411304e-01
2.02459121e+00 -2.38278523e-01 -1.31977046e+00 -5.89413941e-01
-4.07339185e-01 -1.01120150e+00 3.66737843e-02 -2.68201232e-01
-1.02837598e+00 1.00480735e+00 8.23432684e-01 -1.85393527e-01
1.19716287e+00 4.24127638e-01 9.75060761e-01 4.07079399e-01
7.49219477e-01 -6.88839376e-01 3.32520932e-01 6.31033242e-01
9.59189415e-01 -1.46629035e+00 1.37575388e-01 -3.56387794e-02
-1.92918345e-01 1.11796141e+00 4.75364894e-01 -4.33516979e-01
7.74829090e-01 5.13676507e-03 6.34664297e-02 -5.78915775e-01
-4.77139920e-01 -2.48056129e-01 3.04537088e-01 6.03548944e-01
2.05484062e-01 -1.85745835e-01 3.93985152e-01 -9.42747518e-02
-1.59374088e-01 -9.44069326e-02 -5.08631580e-02 1.08734775e+00
-3.94315831e-02 -1.30742955e+00 -2.49888733e-01 3.79895940e-02
1.80578128e-01 2.26620629e-01 -3.69110435e-01 9.64932382e-01
2.47398436e-01 3.03233832e-01 2.69958049e-01 -4.66302782e-01
5.32146513e-01 4.29313108e-02 7.71156907e-01 -4.80869889e-01
9.03680846e-02 -5.95173500e-02 -4.89509910e-01 -1.08596623e+00
-5.95520675e-01 -6.83378816e-01 -1.00442410e+00 -3.03815484e-01
4.08047289e-02 -1.95456386e-01 9.26860094e-01 3.68755728e-01
6.38970017e-01 -3.88078429e-02 9.13946390e-01 -1.19004858e+00
-4.23976421e-01 -5.90791285e-01 -6.72993779e-01 6.36585176e-01
6.09373152e-01 -1.02538371e+00 -5.52415967e-01 -2.82406181e-01] | [7.97990083694458, -2.949850082397461] |
c895c24b-352f-44ad-a396-3c831331e34e | sups-a-simulated-underground-parking-scenario | 2302.12966 | null | https://arxiv.org/abs/2302.12966v1 | https://arxiv.org/pdf/2302.12966v1.pdf | SUPS: A Simulated Underground Parking Scenario Dataset for Autonomous Driving | Automatic underground parking has attracted considerable attention as the scope of autonomous driving expands. The auto-vehicle is supposed to obtain the environmental information, track its location, and build a reliable map of the scenario. Mainstream solutions consist of well-trained neural networks and simultaneous localization and mapping (SLAM) methods, which need numerous carefully labeled images and multiple sensor estimations. However, there is a lack of underground parking scenario datasets with multiple sensors and well-labeled images that support both SLAM tasks and perception tasks, such as semantic segmentation and parking slot detection. In this paper, we present SUPS, a simulated dataset for underground automatic parking, which supports multiple tasks with multiple sensors and multiple semantic labels aligned with successive images according to timestamps. We intend to cover the defect of existing datasets with the variability of environments and the diversity and accessibility of sensors in the virtual scene. Specifically, the dataset records frames from four surrounding fisheye cameras, two forward pinhole cameras, a depth camera, and data from LiDAR, inertial measurement unit (IMU), GNSS. Pixel-level semantic labels are provided for objects, especially ground signs such as arrows, parking lines, lanes, and speed bumps. Perception, 3D reconstruction, depth estimation, and SLAM, and other relative tasks are supported by our dataset. We also evaluate the state-of-the-art SLAM algorithms and perception models on our dataset. Finally, we open source our virtual 3D scene built based on Unity Engine and release our dataset at https://github.com/jarvishou829/SUPS. | ['Jian Pu', 'Taiping Zeng', 'xiangyang xue', 'Guang Chen', 'Yurong Cheng', 'Qi Chen', 'Jiawei Hou'] | 2023-02-25 | null | null | null | null | ['simultaneous-localization-and-mapping', 'unity'] | ['computer-vision', 'computer-vision'] | [-1.67131588e-01 -2.67468214e-01 -1.87027678e-01 -8.05265367e-01
-5.80945432e-01 -4.56387132e-01 4.65464085e-01 -1.56515595e-02
-8.06647241e-01 7.82824457e-01 -3.57280344e-01 -4.31590319e-01
2.48860300e-01 -9.38972175e-01 -9.17942107e-01 -3.29421937e-01
1.03475429e-01 8.25564921e-01 7.19064236e-01 -3.56339693e-01
4.44688290e-01 4.69505131e-01 -1.83218801e+00 -4.16548401e-01
8.53118539e-01 1.18602347e+00 1.01600516e+00 4.92555082e-01
-1.32790148e-01 3.40035915e-01 3.84734310e-02 2.85358541e-02
2.84800529e-01 3.19265157e-01 -3.23829919e-01 1.10416904e-01
3.98520023e-01 -3.47411543e-01 -4.15178001e-01 9.68040586e-01
4.48406965e-01 -7.40531757e-02 2.59476840e-01 -1.65342414e+00
-2.28462785e-01 5.99208288e-02 -4.70288277e-01 -1.11191034e-01
1.25146389e-01 2.50100255e-01 5.11968553e-01 -1.06926966e+00
6.42714560e-01 1.03776026e+00 7.70842612e-01 1.61401600e-01
-6.22070670e-01 -7.54991770e-01 -1.26950949e-01 6.35123730e-01
-1.70780659e+00 -5.49852431e-01 6.46750093e-01 -3.76986980e-01
6.95256948e-01 -8.69104937e-02 5.59507966e-01 9.81122971e-01
2.16819763e-01 6.04462266e-01 1.04799807e+00 -1.47029757e-01
3.73930752e-01 3.69154900e-01 -2.71988176e-02 8.24651182e-01
3.22628200e-01 7.03323409e-02 -6.35717392e-01 4.17221904e-01
5.75214088e-01 2.83352226e-01 6.79411665e-02 -8.30483913e-01
-1.54410768e+00 7.45813847e-01 4.80594963e-01 -2.42188439e-01
-3.75981063e-01 2.17559844e-01 2.27170303e-01 -1.70480430e-01
-4.61738780e-02 -2.25166515e-01 -4.88382041e-01 -1.60456911e-01
-7.91241050e-01 1.07239783e-01 4.12630588e-01 1.47110021e+00
1.52170551e+00 3.89023162e-02 4.95476961e-01 4.74503219e-01
5.34590423e-01 1.21428835e+00 5.32709777e-01 -1.36543107e+00
6.86472714e-01 4.91860271e-01 3.93284798e-01 -8.87187064e-01
-5.77443600e-01 1.41716674e-01 -4.51789588e-01 1.65205821e-01
6.72062784e-02 1.71101347e-01 -9.96116757e-01 1.41913259e+00
3.91722381e-01 2.71185815e-01 1.51185572e-01 9.11060691e-01
7.80319929e-01 2.84890622e-01 -1.31219059e-01 2.13931948e-01
1.21707106e+00 -9.45507467e-01 -5.75641036e-01 -9.97370243e-01
7.39724159e-01 -6.03018939e-01 1.01710582e+00 -1.32439169e-03
-5.05586028e-01 -7.48118818e-01 -1.16008735e+00 -1.38208911e-01
-8.20913613e-01 3.52069139e-01 5.54355621e-01 3.57366294e-01
-1.12942684e+00 1.53716952e-01 -8.08938086e-01 -6.39753580e-01
1.73145548e-01 1.69984967e-01 -5.60188055e-01 -3.47936839e-01
-1.38155580e+00 1.14200664e+00 4.68386650e-01 2.67874062e-01
-1.00121653e+00 -9.88952368e-02 -1.28289270e+00 -5.53648293e-01
3.61799449e-01 -4.68777895e-01 1.04029644e+00 -7.85867646e-02
-9.76262808e-01 1.10598719e+00 -4.37957525e-01 -7.00636029e-01
4.84322906e-01 -1.58536583e-01 -3.92614394e-01 -1.96779281e-01
8.70123327e-01 9.65281546e-01 2.80337960e-01 -1.44272661e+00
-1.14333951e+00 -6.71348274e-01 -2.76189089e-01 3.14263225e-01
3.87822658e-01 -4.86750662e-01 -7.30477631e-01 5.40531516e-01
6.29649997e-01 -1.09906566e+00 -4.01421398e-01 -6.36470085e-03
-4.71772313e-01 1.03384264e-01 1.01796770e+00 -4.30219144e-01
6.14973307e-01 -2.27662563e+00 -4.91640151e-01 -1.61440998e-01
-1.03868201e-01 -3.21315855e-01 1.93628028e-01 1.84953168e-01
4.47817296e-01 -3.41385096e-01 -1.59062326e-01 -8.50902617e-01
3.31977010e-01 8.88080716e-01 -3.83172423e-01 6.52896523e-01
-4.32661086e-01 8.97460401e-01 -8.12058032e-01 -8.48882973e-01
8.55788231e-01 1.05786934e-01 -1.17716186e-01 -1.05350003e-01
-7.99709633e-02 4.27493870e-01 -4.06191558e-01 1.11740339e+00
1.11706436e+00 1.10535743e-02 -1.87455341e-01 -1.32071555e-01
-6.81390345e-01 2.28760689e-01 -1.46192503e+00 2.10479331e+00
-4.46393579e-01 7.16046095e-01 2.38342002e-01 -8.35817516e-01
1.19600725e+00 -2.77306020e-01 3.17023486e-01 -1.27747357e+00
5.90276718e-02 6.94329143e-01 -8.24667633e-01 -4.44493741e-01
1.17128766e+00 2.53632963e-01 -4.52817380e-01 -1.99421093e-01
-2.97242671e-01 -4.48074609e-01 1.29894882e-01 1.92403436e-01
8.02530766e-01 3.34789187e-01 -2.26740390e-02 9.61171165e-02
4.16582733e-01 6.54777706e-01 6.75680161e-01 7.32680917e-01
-5.15291393e-01 4.83438462e-01 -2.94481784e-01 -3.71987671e-01
-1.23179436e+00 -1.28474033e+00 -2.23376825e-01 4.24926430e-01
1.10533428e+00 -1.75568357e-01 -1.93328708e-01 -1.69400021e-01
3.90196443e-01 6.57622516e-01 -2.51466990e-01 1.68936148e-01
-3.93472493e-01 -2.46169850e-01 3.28749061e-01 5.35559058e-01
9.98578966e-01 -8.43183637e-01 -9.96973515e-01 1.82595506e-01
-4.77477044e-01 -1.71398735e+00 1.83157012e-01 3.58319938e-01
-6.56702101e-01 -1.18550837e+00 -1.75648858e-03 -6.47968471e-01
3.42056394e-01 8.97558689e-01 9.52863157e-01 -3.72324377e-01
-5.19947708e-02 3.76814067e-01 -9.03409123e-02 -3.70155573e-01
-3.50542516e-02 -2.06426650e-01 4.34074223e-01 -4.25511301e-01
5.21320701e-01 -3.45139861e-01 -6.50832474e-01 8.54645610e-01
-3.03705454e-01 3.66518617e-01 4.74678218e-01 3.08558047e-01
1.05232084e+00 2.27902271e-03 -5.85830882e-02 -2.60877788e-01
-1.35021850e-01 -6.32811069e-01 -8.03351998e-01 -2.55941510e-01
-7.52928197e-01 -3.44585150e-01 1.58377051e-01 2.55882472e-01
-6.90247118e-01 4.74882960e-01 2.15164199e-02 -4.94311571e-01
-5.40962517e-01 2.35657424e-01 -2.09084317e-01 -1.01247244e-01
5.74672461e-01 5.58713675e-01 7.11899996e-02 -1.59745768e-01
4.74010974e-01 9.57451344e-01 1.00144982e+00 -2.59731412e-01
7.03553796e-01 9.84397113e-01 1.33103758e-01 -7.97142386e-01
-5.93923807e-01 -7.91559875e-01 -8.71029973e-01 -3.52149457e-01
7.60159075e-01 -1.49105990e+00 -6.53464615e-01 4.90718901e-01
-9.77366686e-01 -3.84513259e-01 -6.52658520e-03 5.58168292e-01
-8.85727108e-01 4.84964967e-01 -1.08744353e-01 -8.42970371e-01
-5.86535875e-03 -1.33020878e+00 1.34895360e+00 3.63671958e-01
2.23419666e-01 -6.78172350e-01 -1.12223946e-01 7.37973273e-01
3.25089782e-01 1.84853494e-01 -4.72927168e-02 -1.60180613e-01
-1.05171990e+00 -2.91339219e-01 -3.12138557e-01 -3.83966719e-03
-1.47099689e-01 -4.15032744e-01 -8.95957112e-01 1.09790318e-01
-3.08521509e-01 -2.63718903e-01 8.90248358e-01 4.46651787e-01
7.84185767e-01 3.16042632e-01 -7.53555417e-01 7.09594131e-01
1.49273348e+00 1.53950796e-01 5.98163068e-01 1.01969004e+00
7.50687659e-01 4.74682689e-01 1.38990009e+00 4.09098059e-01
1.24465168e+00 6.49871349e-01 1.11214244e+00 4.26199585e-02
2.46261641e-01 -4.39938098e-01 2.55187839e-01 4.75306660e-01
2.18573317e-01 4.38784771e-02 -1.03767872e+00 7.83735931e-01
-2.05409455e+00 -6.35401964e-01 -5.79757392e-01 2.21148038e+00
1.22839101e-01 2.25952193e-01 -3.08011234e-01 -1.78221576e-02
7.76582241e-01 1.84337780e-01 -6.70562625e-01 1.39187300e-03
-3.62610042e-01 -5.13900578e-01 1.40639150e+00 5.67540944e-01
-1.12212086e+00 1.22405851e+00 5.02095938e+00 6.47217572e-01
-1.00622237e+00 2.45576650e-01 1.85476225e-02 3.01927775e-01
-2.49066815e-01 3.42333853e-01 -1.29137146e+00 8.25753272e-01
1.00663185e+00 2.32754126e-01 1.89942494e-01 1.34677613e+00
5.52061498e-01 -7.73107231e-01 -6.86258018e-01 1.32271707e+00
-1.44167123e-02 -1.49045992e+00 -6.03450119e-01 2.60340571e-01
4.55714792e-01 9.09689188e-01 -1.62889674e-01 2.97687143e-01
3.50598782e-01 -5.87305725e-01 1.16214097e+00 3.80235076e-01
8.73599589e-01 -4.04944807e-01 7.82218814e-01 6.96866274e-01
-1.45004880e+00 -1.04996860e-01 -4.85178351e-01 -1.52007252e-01
5.55697620e-01 4.83976960e-01 -1.15470612e+00 5.77681720e-01
8.37209463e-01 9.78003979e-01 -5.96946418e-01 9.71917033e-01
-3.55120778e-01 -8.49839374e-02 -6.10742271e-01 2.18216348e-02
4.17460114e-01 -4.54117656e-01 1.36252463e-01 7.81873226e-01
6.27076328e-01 -3.83098781e-01 4.81986851e-01 7.31868446e-01
3.41662496e-01 -3.31003547e-01 -9.12789047e-01 5.32165408e-01
9.46061492e-01 1.29135811e+00 -6.21118903e-01 -3.77772063e-01
-3.90144944e-01 8.34845245e-01 5.10744867e-04 9.84245315e-02
-1.01697552e+00 -1.08830273e-01 8.17934215e-01 3.55623454e-01
6.30062744e-02 -8.98022115e-01 -4.57625389e-01 -9.34848189e-01
8.59408677e-02 8.19317251e-02 -7.15498105e-02 -1.37591529e+00
-4.82269853e-01 1.60047889e-01 -1.80247724e-01 -1.40070200e+00
-2.99853049e-02 -5.72413623e-01 -1.97422028e-01 7.26303399e-01
-2.09974694e+00 -1.26316059e+00 -1.06663477e+00 7.51024604e-01
6.29685223e-01 2.48215590e-02 4.45635915e-01 5.93479097e-01
-3.07900548e-01 -2.86781698e-01 2.04929635e-01 -1.55605674e-01
7.83560753e-01 -9.74381804e-01 4.63817537e-01 8.72883618e-01
-1.99396729e-01 -8.36655404e-03 7.96575844e-01 -7.70017624e-01
-1.61212611e+00 -1.21553218e+00 1.06063521e+00 -5.36944091e-01
6.88317716e-01 -3.42801034e-01 -3.81104559e-01 9.24719810e-01
-5.08273423e-01 1.34676084e-01 -5.49256243e-02 -3.76265675e-01
1.82772756e-01 -3.03264558e-01 -1.07579327e+00 3.66514891e-01
1.15110648e+00 -4.91597772e-01 -1.70180231e-01 4.22752947e-01
6.03335261e-01 -8.80123913e-01 -2.67577708e-01 5.37234008e-01
3.59632403e-01 -1.19402134e+00 1.16757166e+00 4.34889913e-01
1.68532029e-01 -8.03142905e-01 -8.54257822e-01 -7.52255917e-01
-3.77224549e-03 2.57264942e-01 2.38015443e-01 9.96553481e-01
2.06493989e-01 -7.95701742e-01 1.10006833e+00 5.25947452e-01
-6.69212222e-01 -1.40064120e-01 -1.26132405e+00 -7.02112734e-01
-6.63967073e-01 -8.98423553e-01 5.67844629e-01 7.06937432e-01
-4.46792454e-01 1.28443614e-01 -5.14397204e-01 6.43290699e-01
8.86972189e-01 2.54012287e-01 1.28349996e+00 -1.19756651e+00
4.60835695e-01 5.34246713e-02 -8.30283463e-01 -1.23384321e+00
1.39659062e-01 -6.90675974e-01 3.70061666e-01 -1.98952854e+00
-1.97325870e-01 -8.46230567e-01 1.21040568e-01 4.18370098e-01
4.26863968e-01 1.81965351e-01 -1.83257446e-01 6.27112806e-01
-9.50851262e-01 6.60313308e-01 8.47111642e-01 -1.55778036e-01
9.40615162e-02 -7.27465376e-02 -1.25107124e-01 8.07019651e-01
8.71507645e-01 -2.66719311e-01 -2.27084026e-01 -7.39574313e-01
1.57576740e-01 1.11444145e-01 7.47145355e-01 -1.39635134e+00
6.23740494e-01 -2.28464544e-01 3.26202571e-01 -1.45855427e+00
8.63405764e-01 -9.38518286e-01 2.61739701e-01 6.23618066e-01
6.07192516e-01 1.09697104e-01 5.35119623e-02 5.84910035e-01
-2.44716033e-01 -2.81584766e-02 6.44906044e-01 -4.30332661e-01
-2.00104642e+00 5.18841445e-01 -1.50428802e-01 -3.67539197e-01
1.27913880e+00 -8.35254252e-01 -5.07179618e-01 -1.21269546e-01
-3.66926044e-01 7.54464984e-01 8.51117849e-01 5.51928759e-01
8.63735437e-01 -1.38304472e+00 -2.25744054e-01 4.44512457e-01
4.85776067e-01 6.25995576e-01 4.50834006e-01 9.74312305e-01
-7.56302714e-01 6.45363748e-01 -2.86651701e-01 -1.16172421e+00
-9.84655857e-01 4.08051193e-01 1.84772357e-01 5.34348369e-01
-5.56363344e-01 4.34110850e-01 -1.53257132e-01 -8.67562354e-01
-7.69491643e-02 -2.45740473e-01 -1.28769670e-02 -1.30404457e-01
1.11992918e-01 2.81681597e-01 3.87610048e-02 -1.06393719e+00
-7.52559781e-01 8.31993461e-01 5.04475415e-01 7.28649572e-02
1.00625896e+00 -1.05736005e+00 5.46972826e-02 5.72310925e-01
8.70548189e-01 -2.06814229e-01 -1.30546665e+00 -3.50658983e-01
-1.79433301e-01 -5.95179915e-01 -4.97052819e-02 -2.81197727e-01
-5.87054968e-01 8.13130617e-01 8.39407861e-01 -1.63051695e-01
6.26215935e-01 1.40618637e-01 8.92922699e-01 3.64826620e-01
1.19510806e+00 -1.48224819e+00 -2.67315209e-01 8.95774007e-01
2.41448760e-01 -1.84123719e+00 -2.05638319e-01 -2.46630207e-01
-7.32889175e-01 7.55766928e-01 7.49904394e-01 1.50840387e-01
4.43796992e-01 2.33887658e-01 3.68851125e-01 -2.60247082e-01
-2.15098292e-01 -4.61056888e-01 -6.74019217e-01 7.82158732e-01
-4.51595724e-01 4.05083656e-01 1.49835870e-01 1.94826037e-01
-4.47936952e-01 -4.26621288e-02 5.55671394e-01 1.13909459e+00
-1.10595024e+00 -6.09178126e-01 -5.67878366e-01 -3.27630900e-04
4.78406221e-01 7.55442157e-02 2.77042240e-01 8.70683432e-01
5.05390942e-01 8.66612196e-01 2.34706029e-01 -5.68609595e-01
4.17709500e-01 -1.68249607e-01 2.38858499e-02 -2.65094578e-01
6.64086163e-01 -4.93206859e-01 1.77827567e-01 -8.36159766e-01
-2.76681960e-01 -6.87008262e-01 -1.59506333e+00 -4.73792464e-01
-1.50345579e-01 4.69464697e-02 1.60890651e+00 9.42774057e-01
5.40626526e-01 1.56921018e-02 7.56235301e-01 -1.29038298e+00
-1.26100674e-01 -6.68985844e-01 -8.37505400e-01 6.29677698e-02
1.91535980e-01 -1.03949940e+00 -4.00155634e-01 -1.35849670e-01] | [7.540064811706543, -2.143901824951172] |
31837529-047c-4f3c-a5f1-bafb90b4445b | partial-adversarial-domain-adaptation | 1808.04205 | null | http://arxiv.org/abs/1808.04205v1 | http://arxiv.org/pdf/1808.04205v1.pdf | Partial Adversarial Domain Adaptation | Domain adversarial learning aligns the feature distributions across the
source and target domains in a two-player minimax game. Existing domain
adversarial networks generally assume identical label space across different
domains. In the presence of big data, there is strong motivation of
transferring deep models from existing big domains to unknown small domains.
This paper introduces partial domain adaptation as a new domain adaptation
scenario, which relaxes the fully shared label space assumption to that the
source label space subsumes the target label space. Previous methods typically
match the whole source domain to the target domain, which are vulnerable to
negative transfer for the partial domain adaptation problem due to the large
mismatch between label spaces. We present Partial Adversarial Domain Adaptation
(PADA), which simultaneously alleviates negative transfer by down-weighing the
data of outlier source classes for training both source classifier and domain
adversary, and promotes positive transfer by matching the feature distributions
in the shared label space. Experiments show that PADA exceeds state-of-the-art
results for partial domain adaptation tasks on several datasets. | ['Jian-Min Wang', 'Zhangjie Cao', 'Mingsheng Long', 'Lijia Ma'] | 2018-08-10 | partial-adversarial-domain-adaptation-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Zhangjie_Cao_Partial_Adversarial_Domain_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Zhangjie_Cao_Partial_Adversarial_Domain_ECCV_2018_paper.pdf | eccv-2018-9 | ['partial-domain-adaptation'] | ['methodology'] | [ 1.64562374e-01 1.25001565e-01 -2.76438802e-01 -4.97490525e-01
-9.67755675e-01 -1.23320413e+00 5.89278340e-01 -1.61024600e-01
-5.99146485e-01 1.12648845e+00 -1.55101065e-02 5.01618125e-02
2.50565499e-01 -1.05189776e+00 -9.03800547e-01 -6.97587252e-01
3.85956019e-01 8.51660728e-01 3.47303450e-01 -4.56896156e-01
-2.40597695e-01 1.79578483e-01 -8.91142905e-01 3.03240508e-01
8.67623031e-01 8.70429695e-01 -3.46219957e-01 1.43877611e-01
-3.30492169e-01 6.03928030e-01 -9.61330295e-01 -6.45197928e-01
7.66987562e-01 -4.83533710e-01 -6.74956381e-01 -3.00119370e-01
4.66342121e-01 -5.17073750e-01 -3.88467282e-01 1.39291203e+00
6.90793157e-01 1.59484059e-01 8.21479499e-01 -1.87810659e+00
-9.91650343e-01 2.82634437e-01 -5.29936254e-01 -5.52227162e-02
9.12489370e-02 1.41271383e-01 7.37865746e-01 -6.03372335e-01
7.20772088e-01 1.24092889e+00 7.20396757e-01 1.12026453e+00
-1.31908751e+00 -1.36253238e+00 2.92717576e-01 -1.51937008e-01
-1.30325258e+00 -3.30776884e-03 8.77477527e-01 -6.33271575e-01
5.32390177e-01 -2.06028536e-01 4.12031040e-02 1.77664948e+00
-1.05846003e-01 5.61683059e-01 1.14515448e+00 -7.90686086e-02
5.95802724e-01 4.40324008e-01 -1.12685665e-01 -1.32494047e-01
5.65526821e-02 6.50807023e-01 -1.91019937e-01 -4.86242890e-01
6.49253607e-01 1.40380934e-01 1.19380012e-01 -1.03001893e+00
-9.26846743e-01 1.14330292e+00 3.96528542e-01 -1.04813665e-01
-1.76516756e-01 -2.22783595e-01 7.47201622e-01 7.48629808e-01
7.44373620e-01 5.35660863e-01 -7.23066211e-01 2.04300061e-01
-4.24703926e-01 6.51910901e-01 7.58973479e-01 1.24591660e+00
6.74857676e-01 -7.66407177e-02 -1.86082467e-01 9.21768844e-01
1.65863410e-02 7.30798006e-01 6.00487530e-01 -8.35897565e-01
7.48106718e-01 5.42218864e-01 3.35073739e-01 -3.40370297e-01
-1.45918712e-01 -3.23085964e-01 -7.66893446e-01 6.63655639e-01
9.41945910e-01 -6.28753066e-01 -8.77093256e-01 2.39608788e+00
5.43386519e-01 8.50796551e-02 3.47286999e-01 8.62363577e-01
6.70505404e-01 4.04762149e-01 4.78477538e-01 2.98876792e-01
9.78573084e-01 -8.78588438e-01 -4.57765132e-01 -7.49246299e-01
5.71724296e-01 -3.86857510e-01 1.32662702e+00 7.29393959e-02
-7.58837938e-01 -5.27314544e-01 -1.10877812e+00 6.42431825e-02
-7.13414252e-01 -5.21368027e-01 3.48305047e-01 6.47964716e-01
-5.10245025e-01 4.85577464e-01 -3.50025535e-01 -1.75429881e-01
8.47149491e-01 5.09932876e-01 -7.18960166e-01 -3.40707779e-01
-1.82522845e+00 7.76876032e-01 4.83064950e-01 -8.23808372e-01
-1.05977893e+00 -1.06280470e+00 -1.01802564e+00 -4.85926904e-02
2.77275056e-01 -5.08702517e-01 1.27635872e+00 -1.58341610e+00
-1.36841941e+00 1.28675759e+00 3.56651008e-01 -4.73113537e-01
8.46679032e-01 -1.95056841e-01 -5.58565855e-01 -4.23438430e-01
3.09041083e-01 7.19837487e-01 1.00454152e+00 -1.17145300e+00
-6.04525864e-01 -5.15305877e-01 1.05674252e-01 3.31106126e-01
-3.20052594e-01 -1.51584089e-01 7.53258392e-02 -9.60989416e-01
-4.86681521e-01 -8.97188306e-01 -1.61815777e-01 1.68184027e-01
-1.41469881e-01 -1.10158376e-01 7.89744079e-01 -3.22289646e-01
7.89461017e-01 -2.34224606e+00 1.14106938e-01 1.20128229e-01
1.85299620e-01 3.77905875e-01 -4.14280057e-01 8.02190825e-02
-5.41995287e-01 -8.45177546e-02 -4.19542968e-01 -1.69575557e-01
2.77612865e-01 4.16666776e-01 -7.64269769e-01 4.64570701e-01
3.21926862e-01 8.13283205e-01 -1.05865550e+00 -1.73744708e-01
-1.10159844e-01 4.70148474e-02 -5.69295764e-01 5.23561656e-01
-2.27158323e-01 6.26174867e-01 -3.62609327e-01 3.94685894e-01
1.20163226e+00 9.13622081e-02 2.56879237e-02 3.38081867e-01
7.12303162e-01 9.84628946e-02 -1.19751847e+00 1.67982376e+00
-1.37494475e-01 1.83329210e-01 1.43716425e-01 -1.10567164e+00
1.05081797e+00 2.51271516e-01 4.88354623e-01 -8.67023528e-01
5.29965721e-02 2.26770326e-01 -9.88579243e-02 2.92539205e-02
1.15405433e-01 -1.06988096e+00 -7.15689719e-01 4.38111454e-01
3.56909752e-01 -1.74365669e-01 -3.67432803e-01 5.25233001e-02
9.50971186e-01 2.45719776e-01 3.76484662e-01 -1.62654966e-01
3.35895121e-01 1.44860506e-01 8.68743598e-01 6.58514023e-01
-8.10316145e-01 7.42756784e-01 7.24252403e-01 -4.31057036e-01
-1.21740508e+00 -1.64339566e+00 1.93145964e-02 1.63460016e+00
4.06706005e-01 3.80730689e-01 -8.40523779e-01 -1.43489468e+00
3.66586328e-01 5.86873949e-01 -9.65021014e-01 -6.97651803e-01
-4.39896643e-01 -4.98333335e-01 8.79253507e-01 6.88511491e-01
3.17455739e-01 -1.17560124e+00 1.55988842e-01 1.13475017e-01
-2.86608338e-02 -9.32611823e-01 -6.55472875e-01 4.45448726e-01
-5.48891604e-01 -9.60938990e-01 -1.05002701e+00 -1.01531792e+00
4.70087647e-01 -2.12387472e-01 1.45772958e+00 -6.77713096e-01
2.84357578e-01 -1.53805502e-02 -2.12788209e-01 -6.92270160e-01
-6.93622828e-01 1.37563169e-01 3.62374991e-01 -2.65275836e-01
9.33877289e-01 -6.04691029e-01 -2.84512073e-01 5.80061316e-01
-8.95810366e-01 -5.58504522e-01 1.86221570e-01 1.05929959e+00
5.96254230e-01 -4.89197522e-02 1.13885498e+00 -1.46747231e+00
6.95543706e-01 -9.88096178e-01 -5.25140107e-01 7.00183213e-02
-3.07888985e-01 -2.73253415e-02 1.04580200e+00 -1.10935831e+00
-1.05132270e+00 -5.86659983e-02 -6.01801313e-02 -5.07380843e-01
-4.98089015e-01 -2.92382061e-01 -8.70076478e-01 1.78364381e-01
1.19441164e+00 -1.03164449e-01 1.83405727e-01 -3.82751018e-01
4.52280581e-01 5.53586662e-01 6.43561602e-01 -8.43048394e-01
1.16796994e+00 5.51388919e-01 -4.25101221e-01 1.59292668e-01
-9.37512040e-01 -3.22109640e-01 -7.36373901e-01 5.05649984e-01
6.14743650e-01 -1.24575603e+00 -3.96174453e-02 7.59389639e-01
-8.56819570e-01 -6.47532821e-01 -1.03406358e+00 1.58055440e-01
-7.08446205e-01 1.95508003e-01 -3.41413826e-01 -5.38630523e-02
-1.41977564e-01 -9.34557498e-01 9.20247138e-01 7.58823454e-02
-3.42856556e-01 -1.25465477e+00 4.68578666e-01 1.22384578e-01
3.10619652e-01 4.55897957e-01 8.02419543e-01 -1.42264128e+00
2.39469454e-01 -3.55409205e-01 -1.43524513e-01 7.93847978e-01
2.03885704e-01 -8.89002740e-01 -1.09614599e+00 -6.10611022e-01
-3.15468945e-02 -7.37561166e-01 6.11058176e-01 1.81463793e-01
8.29117775e-01 -1.15870677e-01 -2.57993907e-01 7.08560050e-01
1.29948950e+00 1.65127054e-01 5.74618936e-01 6.14088476e-01
6.39573693e-01 6.02523208e-01 8.52455497e-01 4.28453326e-01
1.31818518e-01 4.90654647e-01 4.72939938e-01 -4.15210903e-01
-6.16606735e-02 -4.73421305e-01 5.33348441e-01 -2.53974169e-01
8.72350931e-01 -3.26769352e-01 -8.75770271e-01 8.05901229e-01
-1.59706962e+00 -8.61548722e-01 6.01346850e-01 2.19882989e+00
1.17207301e+00 2.93698996e-01 4.12710488e-01 -3.70391339e-01
9.49819148e-01 6.12539053e-02 -1.21077228e+00 -3.84952396e-01
-1.91346064e-01 3.78185272e-01 7.65562713e-01 4.31993932e-01
-1.46159720e+00 1.04421794e+00 6.06697130e+00 1.06481659e+00
-1.01155233e+00 3.48265648e-01 2.46139348e-01 -2.78987765e-01
-2.37516254e-01 -2.48146623e-01 -5.96073210e-01 7.21185386e-01
6.81403697e-01 -3.60262126e-01 3.19883406e-01 1.33647406e+00
-5.13754785e-01 5.85557342e-01 -1.38775313e+00 7.24413633e-01
-2.51685083e-01 -7.70428658e-01 5.04796319e-02 3.04052860e-01
1.01633811e+00 1.90505281e-01 3.82359862e-01 9.14138436e-01
1.07769668e+00 -1.13332558e+00 5.45182884e-01 -2.07099348e-01
1.32513666e+00 -9.19287920e-01 9.83264506e-01 5.24798691e-01
-7.00992823e-01 -1.88331157e-01 -5.19856751e-01 -1.38778031e-01
-6.24309778e-02 1.14524439e-01 -7.38781989e-01 1.08132035e-01
5.57480752e-01 5.59002578e-01 -3.32051694e-01 4.11711484e-01
-1.15597725e-01 3.55459869e-01 -2.93277681e-01 6.21067345e-01
2.29311287e-01 2.21288335e-02 4.55001175e-01 8.81605029e-01
4.26329561e-02 -1.56299189e-01 3.92934650e-01 9.52695727e-01
-5.65943599e-01 -1.65651545e-01 -1.00620770e+00 1.72787696e-01
7.96920896e-01 6.74641550e-01 -1.04761999e-02 -4.00483966e-01
-5.14687121e-01 1.13816249e+00 4.70942199e-01 4.36131924e-01
-1.02989018e+00 -5.67482054e-01 1.41683650e+00 2.09002495e-01
7.94862062e-02 4.50317383e-01 -5.77824235e-01 -1.24840844e+00
-8.74948204e-02 -1.13321257e+00 7.65045464e-01 -2.92130351e-01
-1.97619998e+00 3.56866300e-01 -1.21450558e-01 -1.68875682e+00
-3.08845103e-01 -4.93946522e-01 -5.60537636e-01 1.25901794e+00
-1.50240469e+00 -1.22228575e+00 -3.38725112e-02 1.13449907e+00
4.23846394e-01 -5.49358547e-01 1.12329078e+00 4.81900871e-01
-1.94127738e-01 1.22254014e+00 6.33203804e-01 5.13825476e-01
1.56703484e+00 -1.47520721e+00 7.19562471e-01 6.14203751e-01
-5.14052510e-01 3.58279526e-01 5.46710014e-01 -5.50236166e-01
-4.00140464e-01 -1.42008448e+00 5.04718721e-01 -9.20764625e-01
5.80895782e-01 -5.96800506e-01 -1.26157546e+00 8.65731299e-01
-1.42502606e-01 4.76578087e-01 1.06207454e+00 3.71278077e-02
-9.77873564e-01 -4.33319472e-02 -1.87348855e+00 3.19378644e-01
9.50025976e-01 -6.61406457e-01 -8.59389544e-01 2.27372676e-01
7.93263793e-01 -4.35905665e-01 -9.00194824e-01 2.28951961e-01
2.46783257e-01 -5.48972845e-01 1.32214272e+00 -1.22717702e+00
4.46063995e-01 -8.86835754e-02 -2.36627042e-01 -1.69675422e+00
-3.13728988e-01 -2.08651394e-01 -9.35523584e-03 1.36158431e+00
1.43341005e-01 -7.64382243e-01 1.06947911e+00 7.77555466e-01
3.15481812e-01 -1.63353443e-01 -1.06723022e+00 -1.12422287e+00
1.27780676e+00 -7.12572634e-02 9.61494803e-01 1.38482141e+00
-7.91325942e-02 4.30548817e-01 -3.63314033e-01 1.74435318e-01
7.74860799e-01 4.11608405e-02 8.75113606e-01 -1.34853804e+00
-2.18793735e-01 -2.90296376e-01 -3.29357326e-01 -7.42077291e-01
6.84306145e-01 -8.83265197e-01 2.92857513e-02 -9.47644949e-01
1.29626885e-01 -5.90444088e-01 -6.29990041e-01 5.26063204e-01
-2.91419774e-01 2.75517434e-01 1.09982222e-01 1.07361354e-01
-5.48388064e-01 6.21047676e-01 1.24956894e+00 -3.05071235e-01
-1.65384188e-01 4.17535752e-02 -1.14000297e+00 7.26121187e-01
7.83582807e-01 -8.28770220e-01 -6.11176550e-01 -3.58511508e-01
-1.87516317e-01 -3.71899158e-01 1.59711361e-01 -8.86287272e-01
-1.72626361e-01 -4.22376335e-01 5.78143299e-01 1.21768467e-01
1.16262510e-01 -1.09372652e+00 -1.80032551e-01 2.60641843e-01
-5.36335707e-01 -4.94933933e-01 4.95421201e-01 6.58547342e-01
-3.26514065e-01 7.51817524e-02 1.45225203e+00 -1.14620410e-01
-9.85534370e-01 4.01012778e-01 1.50363415e-01 9.21439588e-01
1.36865556e+00 -1.94621593e-01 -4.01195586e-01 -2.85773903e-01
-7.91922808e-01 4.35856968e-01 7.98957527e-01 6.64094090e-01
2.13790998e-01 -1.71584761e+00 -7.89861917e-01 5.34706831e-01
3.90359432e-01 2.74004906e-01 3.79799783e-01 -1.12538688e-01
-5.66074811e-02 -1.90264404e-01 -8.11088562e-01 -2.39079252e-01
-1.00934875e+00 1.11261106e+00 5.79854846e-01 -6.13175988e-01
-3.16006333e-01 1.19230163e+00 1.06170762e+00 -1.26943946e+00
2.24776849e-01 3.05159628e-01 3.85539904e-02 -3.07333842e-03
5.27139604e-01 1.82390839e-01 -1.91124439e-01 -5.74255168e-01
-4.12376910e-01 4.24196094e-01 -3.19009364e-01 1.72693823e-02
1.08018994e+00 -3.82024646e-02 4.07740146e-01 -2.48203315e-02
1.37712538e+00 -1.38918489e-01 -1.76016438e+00 -7.04989076e-01
-2.79890567e-01 -4.25047547e-01 -5.76193810e-01 -1.18060863e+00
-8.42698932e-01 1.10024619e+00 8.34847093e-01 -1.71801671e-01
1.09037733e+00 1.03733078e-01 9.06305432e-01 2.80362293e-02
2.96995848e-01 -1.25868547e+00 1.10401951e-01 7.27277339e-01
6.81299329e-01 -1.55039358e+00 -3.64392281e-01 -1.01394862e-01
-1.10650611e+00 3.97127479e-01 1.17001224e+00 -5.11067152e-01
5.41503906e-01 3.28819156e-01 2.82981306e-01 2.13437080e-01
-4.04384404e-01 6.37284145e-02 -3.32275964e-02 1.41214812e+00
1.75411496e-02 2.85598248e-01 2.27633923e-01 1.29794240e+00
-8.67800713e-02 -5.55442013e-02 1.77154511e-01 7.60883451e-01
-1.53402984e-01 -1.56893921e+00 -5.48249960e-01 1.02713533e-01
-6.06742918e-01 7.86026195e-03 -7.09625542e-01 9.21853006e-01
4.77499813e-01 5.66489697e-01 1.70325786e-01 -3.39957297e-01
7.84774899e-01 3.48649085e-01 1.55625105e-01 -7.91083932e-01
-5.99873662e-01 -3.44260961e-01 -3.64191413e-01 -4.33471203e-01
2.19122812e-01 -3.45673770e-01 -1.22045445e+00 -3.15203041e-01
-2.92933211e-02 1.10806815e-01 8.76005217e-02 6.66720629e-01
4.08119857e-01 3.74652714e-01 4.73007023e-01 -5.30347824e-01
-1.29259670e+00 -9.51640069e-01 -9.68782842e-01 1.22892976e+00
5.79201996e-01 -8.83159697e-01 -3.15277427e-01 -2.37250142e-02] | [10.347396850585938, 3.128276824951172] |
56efd82b-76d4-40e5-8cfe-5e73a0d62ecf | monte-carlo-dropout-ensembles-for-robust | 2007.10114 | null | https://arxiv.org/abs/2007.10114v1 | https://arxiv.org/pdf/2007.10114v1.pdf | Monte Carlo Dropout Ensembles for Robust Illumination Estimation | Computational color constancy is a preprocessing step used in many camera systems. The main aim is to discount the effect of the illumination on the colors in the scene and restore the original colors of the objects. Recently, several deep learning-based approaches have been proposed to solve this problem and they often led to state-of-the-art performance in terms of average errors. However, for extreme samples, these methods fail and lead to high errors. In this paper, we address this limitation by proposing to aggregate different deep learning methods according to their output uncertainty. We estimate the relative uncertainty of each approach using Monte Carlo dropout and the final illumination estimate is obtained as the sum of the different model estimates weighted by the log-inverse of their corresponding uncertainties. The proposed framework leads to state-of-the-art performance on INTEL-TAU dataset. | ['Alexandros Iosifidis', 'Moncef Gabbouj', 'Jenni Raitoharju', 'Jarno Nikkanen', 'Firas Laakom'] | 2020-07-20 | null | null | null | null | ['color-constancy'] | ['computer-vision'] | [-9.13218334e-02 -3.35483491e-01 3.49473268e-01 -5.47342837e-01
-8.57204139e-01 -4.30826694e-01 6.26953423e-01 7.42745697e-02
-7.31255770e-01 7.88286984e-01 -3.75132084e-01 1.96166277e-01
4.38085459e-02 -4.78918821e-01 -9.62447107e-01 -9.69052494e-01
3.66201788e-01 3.57872546e-01 2.82967567e-01 3.83291274e-01
3.83897036e-01 2.91143656e-01 -1.57861674e+00 5.02401497e-03
1.29491627e+00 1.27325475e+00 -5.91323944e-03 5.65637350e-01
-1.02646247e-01 5.15679240e-01 -5.79092741e-01 -5.11807919e-01
3.88457894e-01 -4.97273386e-01 -1.83784097e-01 1.05280317e-01
7.06888974e-01 -4.54446852e-01 -2.42981032e-01 1.58291495e+00
3.67317557e-01 1.56650096e-01 6.07560456e-01 -1.15553570e+00
-4.90954489e-01 1.31947249e-01 -8.78910005e-01 -1.57524630e-01
-3.78864892e-02 2.07652509e-01 6.35758281e-01 -7.78365016e-01
2.43681312e-01 1.01806545e+00 4.47478205e-01 3.38320822e-01
-1.40089929e+00 -4.77712035e-01 1.48514047e-01 3.43794554e-01
-1.31215644e+00 -3.71652812e-01 6.00459039e-01 -2.83494353e-01
5.06131053e-01 -1.53225765e-01 3.59134793e-01 8.11138988e-01
3.93359601e-01 5.23229420e-01 1.54163206e+00 -6.52152658e-01
4.79292423e-01 3.11036825e-01 1.45567805e-01 6.39089346e-01
5.54975510e-01 1.08963981e-01 -4.88675952e-01 1.34824723e-01
5.70679367e-01 -6.60078377e-02 -2.39989936e-01 -2.74526715e-01
-6.99276686e-01 5.61965525e-01 5.73844552e-01 -2.96167154e-02
-2.27779612e-01 5.09416819e-01 9.48241949e-02 -1.46314934e-01
6.04815960e-01 4.40968946e-02 -2.10404947e-01 -1.61537509e-02
-9.93030608e-01 1.68188661e-01 4.80262160e-01 6.40446723e-01
7.18185425e-01 2.04673260e-02 -2.88764030e-01 6.13770783e-01
4.00596350e-01 5.05925953e-01 -2.95564416e-03 -9.37088192e-01
2.01181591e-01 3.79430354e-01 4.23571140e-01 -7.43673325e-01
-3.25574011e-01 -5.75250089e-01 -7.49270439e-01 8.53030562e-01
8.81923020e-01 -1.70170322e-01 -1.22263169e+00 1.80294204e+00
2.18675584e-01 1.26396716e-01 -1.87921196e-01 1.00177097e+00
2.65832573e-01 5.85500360e-01 7.75202736e-03 1.68074016e-03
1.09957910e+00 -1.00259018e+00 -7.74154842e-01 -1.58170268e-01
-4.27439474e-02 -8.65626633e-01 1.06628704e+00 8.03867102e-01
-1.00571716e+00 -5.71183681e-01 -1.28457558e+00 -2.58519389e-02
-9.06568095e-02 3.55311036e-01 3.70315820e-01 1.05750930e+00
-8.90959620e-01 9.15132403e-01 -1.01975918e+00 -5.13258874e-02
4.77670550e-01 1.00483112e-01 9.82266739e-02 -1.91693947e-01
-5.90375006e-01 8.46054256e-01 3.19640398e-01 2.41698101e-01
-9.32577431e-01 -6.38454795e-01 -2.82310665e-01 1.18546307e-01
3.93456489e-01 -4.49918866e-01 1.11505222e+00 -1.17676938e+00
-1.63789165e+00 6.03975594e-01 -1.72623862e-02 -4.17006284e-01
9.84952688e-01 -5.36934316e-01 -8.03221092e-02 -9.06292796e-02
-3.49634707e-01 4.56408292e-01 1.00998163e+00 -1.47824705e+00
-6.67048275e-01 -4.45602775e-01 1.23277687e-01 7.79327825e-02
-1.26212239e-01 -2.53802091e-01 -7.35698462e-01 -4.07645583e-01
4.21459414e-02 -1.14086592e+00 -2.71784002e-03 2.87340194e-01
-2.94242084e-01 1.15542680e-01 3.17464232e-01 -5.37105381e-01
7.92287469e-01 -2.28943205e+00 -9.99411270e-02 -1.34626135e-01
4.66472842e-02 2.17976242e-01 8.06745440e-02 -1.20601237e-01
1.73258990e-01 -2.80810446e-01 -3.60316902e-01 -6.50995970e-01
1.28634468e-01 -5.56333810e-02 -1.83674127e-01 8.03756535e-01
1.95182711e-01 3.65105927e-01 -7.63801455e-01 -2.98712939e-01
5.43638945e-01 7.93516874e-01 -3.85112256e-01 2.43637800e-01
-4.07140046e-01 4.26887959e-01 -1.49384839e-02 2.98110485e-01
1.18416154e+00 7.54940212e-02 1.76583342e-02 -3.35818827e-01
-7.54495040e-02 -2.73206439e-02 -1.41168499e+00 1.60515964e+00
-3.91141564e-01 6.56096220e-01 -4.88940962e-02 -4.60722923e-01
6.54320061e-01 -7.90138394e-02 1.20828293e-01 -6.38769925e-01
2.01645821e-01 3.95874530e-01 1.12088978e-01 5.40099666e-02
5.65243900e-01 -3.28057706e-01 3.63076150e-01 1.80148646e-01
6.85126036e-02 -2.00325936e-01 1.59884661e-01 -8.22750106e-02
6.08088136e-01 4.45283830e-01 -1.49842147e-02 -1.83939144e-01
3.54643375e-01 -3.45324308e-01 5.49570858e-01 7.57578611e-01
-2.24528864e-01 9.81622159e-01 6.73282146e-01 -3.07635933e-01
-1.07320964e+00 -1.19345474e+00 -1.90020412e-01 6.14257038e-01
2.36982167e-01 8.70687738e-02 -1.03297698e+00 -5.50813735e-01
-1.76785171e-01 1.08593631e+00 -5.68081379e-01 -1.96811274e-01
-2.06821263e-01 -1.12970281e+00 2.65832782e-01 3.84617984e-01
6.57307744e-01 -5.90751827e-01 -8.93312335e-01 2.17662081e-02
4.36195880e-02 -1.21141803e+00 -3.62362355e-01 2.21450940e-01
-7.37204492e-01 -1.03651822e+00 -8.40138495e-01 -1.44711003e-01
7.51528144e-01 1.97964251e-01 9.57866788e-01 -1.34433135e-01
-2.39676490e-01 1.70086265e-01 -1.64474919e-01 -3.46872896e-01
-2.53722996e-01 -2.03797236e-01 9.36342329e-02 4.02794868e-01
-5.66091202e-02 -1.71882927e-01 -7.44056821e-01 2.60317940e-02
-1.09015989e+00 -1.70944437e-01 5.10149002e-01 6.63635850e-01
7.35217631e-01 3.31713855e-01 5.83244599e-02 -8.37975740e-01
4.19818461e-01 6.69169351e-02 -1.22444975e+00 3.24331850e-01
-8.90604258e-01 4.98895347e-01 7.56492734e-01 -2.89822996e-01
-1.52588356e+00 1.23354226e-01 1.43988639e-01 -4.57330763e-01
-5.61877042e-02 1.16514467e-01 -5.03641404e-02 -1.83038190e-01
4.10107344e-01 5.16597815e-02 -4.18986291e-01 -3.63225430e-01
4.43168998e-01 2.25591913e-01 6.42095685e-01 -5.87349296e-01
5.81847131e-01 5.32186151e-01 3.09570462e-01 -5.29471517e-01
-8.77880871e-01 1.45007391e-02 -3.47885728e-01 -4.01229680e-01
9.02800441e-01 -7.87607789e-01 -6.51305854e-01 9.87278819e-01
-1.15097940e+00 -2.51220763e-01 -2.10862949e-01 5.24855375e-01
-3.81902933e-01 4.53375101e-01 -3.25596273e-01 -1.16827738e+00
-2.03365356e-01 -1.33192468e+00 6.50403976e-01 7.96373963e-01
4.73305643e-01 -6.47613406e-01 -4.92548645e-02 -2.07105428e-01
3.92597705e-01 2.35625461e-01 8.99539173e-01 -1.20659797e-02
-6.38415933e-01 -2.71343380e-01 -5.82141459e-01 6.54289782e-01
-2.26993244e-02 3.46219450e-01 -1.37396622e+00 -2.21354142e-01
-7.41123036e-03 -1.77592784e-01 1.09854364e+00 7.84778178e-01
1.03919804e+00 2.19665363e-01 1.11592963e-01 6.60825074e-01
1.96310532e+00 7.05709755e-02 8.45443726e-01 2.75143981e-01
5.56622386e-01 4.42335814e-01 5.06060183e-01 6.10214233e-01
-2.02077441e-02 6.65686548e-01 6.92249596e-01 7.19765052e-02
-3.62531692e-01 3.07406671e-02 4.14015293e-01 2.93521374e-01
1.36930132e-02 -4.43995833e-01 -6.90473378e-01 3.73970658e-01
-1.77732325e+00 -6.38604999e-01 -5.29896319e-01 2.77080178e+00
6.23335123e-01 5.36356747e-01 -9.86087918e-02 -1.89441428e-01
6.49866104e-01 1.18077807e-01 -7.41813421e-01 -2.32296020e-01
-1.63129240e-01 1.00499138e-01 7.98765898e-01 5.18682241e-01
-9.14530694e-01 7.19646335e-01 6.22809219e+00 8.71182978e-01
-1.11406064e+00 6.68447763e-02 8.77440155e-01 -2.18237177e-01
-6.20471053e-02 1.18950665e-01 -6.46148801e-01 6.57828212e-01
8.02248061e-01 -3.93921696e-02 7.01466084e-01 6.84771597e-01
-3.53842489e-02 -8.66935849e-01 -9.75652695e-01 1.04756594e+00
1.25429511e-01 -7.83569038e-01 -2.25522071e-01 -8.99255872e-02
9.86982167e-01 4.16877270e-02 4.52403575e-01 -3.70387807e-02
1.88379958e-01 -7.53867447e-01 9.77420211e-01 8.86197090e-01
7.33473718e-01 -7.87465870e-01 8.37561905e-01 8.52137059e-02
-8.82907093e-01 -1.69396885e-02 -4.46272641e-01 9.60355811e-03
1.69761494e-01 1.14948690e+00 -3.96056414e-01 5.19359887e-01
7.38112569e-01 4.75460514e-02 -7.14618921e-01 1.33330595e+00
-4.29941326e-01 4.02736336e-01 -4.92306352e-01 4.42927405e-02
5.26766367e-02 -6.31177902e-01 2.91013867e-01 9.34516907e-01
3.22624534e-01 -2.62916237e-01 -2.33603850e-01 1.07652760e+00
-2.31213912e-01 -3.45092326e-01 6.69356510e-02 1.11845173e-01
2.52104551e-01 1.28209829e+00 -1.03572059e+00 -3.89526099e-01
-5.02188683e-01 1.13083982e+00 3.53361785e-01 3.85491639e-01
-1.09904873e+00 -4.36475545e-01 5.91355145e-01 -4.69743423e-02
2.62089759e-01 -3.18085849e-01 -4.72052544e-01 -1.07968450e+00
2.83198029e-01 -6.25353038e-01 6.35338575e-02 -8.22118342e-01
-1.16361463e+00 5.73355734e-01 -6.21897206e-02 -1.04148078e+00
2.86979172e-02 -7.98734069e-01 -4.06428039e-01 8.69089127e-01
-1.59420145e+00 -7.51142800e-01 -5.35208046e-01 2.02762961e-01
5.13819635e-01 9.59417447e-02 5.40108979e-01 4.88857210e-01
-6.70226872e-01 5.24751425e-01 6.07222974e-01 -2.11944059e-01
1.04854977e+00 -1.37217319e+00 1.91140234e-01 1.16662014e+00
1.01771265e-01 3.19320351e-01 1.07678461e+00 -3.74657661e-01
-1.14326572e+00 -7.64548779e-01 2.76252419e-01 -3.73866618e-01
4.00556862e-01 -1.20230347e-01 -8.19332957e-01 2.61776090e-01
3.53798419e-01 3.38563859e-01 6.15316629e-02 -1.05093159e-01
-3.83072436e-01 -5.10900974e-01 -1.27296686e+00 4.36577290e-01
4.73385721e-01 -4.48303908e-01 -1.54040307e-01 3.33875865e-02
3.88621926e-01 -4.74731117e-01 -2.14103296e-01 7.68396184e-02
6.66906416e-01 -1.52544022e+00 4.84161526e-01 -7.34029114e-02
4.92573738e-01 -5.13705611e-01 -1.73315763e-01 -1.51662242e+00
-1.14921592e-01 -1.29929036e-01 3.06539480e-02 1.42321098e+00
1.46096572e-01 -4.93540496e-01 6.51368976e-01 8.39540184e-01
-1.41135613e-02 -3.27743024e-01 -9.99033332e-01 -6.08924031e-01
-1.24212734e-01 -3.97721678e-01 3.25297713e-01 4.60212022e-01
-7.63076782e-01 -1.73608270e-02 -4.35995311e-01 2.59116888e-01
1.13828254e+00 3.78459902e-03 6.45195544e-01 -1.10052204e+00
-3.97441596e-01 -4.84494120e-01 -1.84134558e-01 -6.39807761e-01
-7.57034719e-02 -3.02453578e-01 3.41455847e-01 -1.35497868e+00
4.45701957e-01 -2.11297512e-01 -4.73810256e-01 4.69452478e-02
-3.70002419e-01 2.55991340e-01 3.80027860e-01 -1.43888414e-01
-5.91939032e-01 5.56062877e-01 7.54615068e-01 -1.99701145e-01
-7.97170103e-02 -7.48390108e-02 -4.02125090e-01 7.76629567e-01
8.20587516e-01 -7.19747841e-01 -9.46078673e-02 -8.05693865e-01
3.20180029e-01 -5.27315915e-01 4.22238082e-01 -1.38780963e+00
2.11676046e-01 -2.69407965e-02 6.82318270e-01 -6.10415459e-01
2.80781329e-01 -9.95816231e-01 2.02006131e-01 3.69293928e-01
-9.09772441e-02 -9.63146538e-02 3.19111466e-01 6.97272420e-01
-9.78600308e-02 -3.83193165e-01 1.33702290e+00 -6.05347147e-03
-5.11789858e-01 -3.69532295e-02 -8.73048529e-02 -7.03373551e-02
8.05988610e-01 1.63545981e-02 -3.25541228e-01 -3.52943897e-01
-2.03702852e-01 -1.81863844e-01 6.48950100e-01 3.37755717e-02
3.91391903e-01 -1.17178059e+00 -6.51720941e-01 -4.54783365e-02
7.79307261e-02 -1.71703458e-01 4.15472955e-01 8.58839512e-01
-6.65888250e-01 1.14614464e-01 -2.92058200e-01 -5.46789348e-01
-9.11623359e-01 5.41773438e-01 5.01946151e-01 -1.23108625e-01
-2.38579184e-01 8.75949800e-01 2.83609211e-01 1.58089325e-02
4.13801163e-01 -5.04617810e-01 2.96906233e-01 5.71602210e-03
4.89243358e-01 5.12169957e-01 2.38960445e-01 -2.84143567e-01
-3.36413592e-01 7.42723286e-01 -2.12824550e-02 -3.26681018e-01
1.19335759e+00 -2.29758918e-01 -4.46212888e-02 4.85846758e-01
1.00468504e+00 -1.35523849e-03 -1.85771286e+00 3.48604430e-04
-1.66944250e-01 -7.92561054e-01 5.74082613e-01 -9.39183474e-01
-1.27707803e+00 9.76844490e-01 1.23806453e+00 -9.21175927e-02
1.33203924e+00 -4.45786417e-01 3.93450737e-01 9.59134847e-02
3.33783031e-01 -1.34631753e+00 -1.00775674e-01 2.95455247e-01
4.60993826e-01 -1.50112581e+00 3.82824272e-01 -1.71540290e-01
-5.60480118e-01 1.18429542e+00 4.95798707e-01 -3.54290068e-01
3.20624650e-01 2.06673831e-01 4.66706157e-02 2.99047798e-01
-4.45163548e-01 -2.15980574e-01 1.96945012e-01 4.42680448e-01
5.32880247e-01 5.93849123e-02 -4.61263269e-01 3.52177799e-01
4.55330521e-01 8.73130038e-02 6.00535154e-01 5.46318233e-01
-2.42079720e-01 -9.21536446e-01 -5.95829070e-01 2.50840485e-01
-4.94318873e-01 -6.62817806e-02 -1.67767048e-01 6.16748512e-01
1.25203475e-01 9.14910018e-01 -5.80346724e-03 -1.04965851e-01
3.05474877e-01 -4.06125002e-02 9.11005795e-01 -2.03284118e-02
-2.40848765e-01 2.51189023e-01 -1.85057342e-01 -5.71725547e-01
-2.97427386e-01 -5.50684214e-01 -1.15913439e+00 -3.07967454e-01
-4.68839943e-01 -1.45908177e-01 8.69444549e-01 6.43914938e-01
8.87599438e-02 6.15976989e-01 5.96940756e-01 -9.20184672e-01
-8.39355826e-01 -7.65694678e-01 -8.88599157e-01 4.32115316e-01
3.18955779e-01 -7.24542320e-01 -4.53242123e-01 -1.32037789e-01] | [10.440155029296875, -2.4881718158721924] |
e4950f99-56b9-4263-86c2-992a9a63d5dd | enhanced-knowledge-graphs-using-typed | null | null | https://openreview.net/forum?id=W0-o-iIrHdf | https://openreview.net/pdf?id=W0-o-iIrHdf | Enhanced Knowledge Graphs Using Typed Entailment Graphs | Constructing knowledge graphs from open-domain corpora is a crucial stage in question answering. Most previous works are based on open information extraction methods, which extract relations by parsing sentences into triples <e1, r, e2>. These methods lack inference ability and are limited by corpus. When the query is different from the relations in the text-based knowledge graph, it is hard to return correct answers. In this paper, we propose a method to enhance knowledge graphs by using typed entailment graphs to add missing links. We construct the enhanced knowledge graph in both dynamical and offline ways. The experiment shows that our method outperforms the pre-trained language models in zero-shot cloze-style question answering. Furthermore, we find entailment graphs can significantly improve the recall and F-score of knowledge graphs. | ['Anonymous'] | 2022-01-20 | null | null | null | acl-arr-january-2022-1 | ['open-information-extraction'] | ['natural-language-processing'] | [-1.54316559e-01 5.97812474e-01 -2.49944851e-01 -1.73723862e-01
-7.74404466e-01 -9.15012777e-01 2.76088655e-01 5.66264510e-01
-8.85730460e-02 8.81150782e-01 2.47513756e-01 -6.83487296e-01
-4.32487756e-01 -1.45507884e+00 -7.91435361e-01 1.38450578e-01
1.41481265e-01 6.11202836e-01 1.02028286e+00 -6.66388929e-01
1.11838102e-01 -1.38415638e-02 -1.37183630e+00 5.21808088e-01
1.07112777e+00 7.02764511e-01 -2.41425298e-02 7.00726986e-01
-1.00087106e+00 1.33674037e+00 -5.12814164e-01 -1.02845407e+00
-8.02058503e-02 -4.89212990e-01 -1.76999974e+00 -3.45639139e-01
2.69301176e-01 5.33167506e-03 -4.63198096e-01 1.07589710e+00
1.85771525e-01 1.67715043e-01 3.40750128e-01 -1.00283206e+00
-7.89009452e-01 9.50850487e-01 -9.52727273e-02 3.20331424e-01
1.03474057e+00 -3.77604634e-01 1.49373031e+00 -5.23278594e-01
7.34197259e-01 1.13301921e+00 4.69079524e-01 4.08191442e-01
-7.59865463e-01 -2.64802694e-01 1.57814890e-01 7.44854987e-01
-1.30728626e+00 -1.05252586e-01 5.83235621e-01 -1.99862421e-01
1.29558969e+00 4.28144276e-01 3.44640434e-01 4.96331811e-01
-2.34942541e-01 5.37794530e-01 8.91523659e-01 -6.94125772e-01
-8.63579735e-02 3.16235833e-02 8.46609116e-01 1.04204524e+00
4.17965174e-01 -4.60780114e-01 -5.33529520e-01 -3.10909003e-01
3.86999756e-01 -2.13710025e-01 -3.12301815e-01 -8.39233473e-02
-8.17636371e-01 8.31334591e-01 3.22910935e-01 4.11979586e-01
-7.93210641e-02 -1.72961876e-01 2.66634375e-01 7.18622744e-01
3.76944602e-01 5.85995317e-01 -7.65255034e-01 2.39100251e-02
-2.30701208e-01 1.26552448e-01 1.58443117e+00 1.13660276e+00
1.04843879e+00 -8.47552240e-01 -2.58524120e-01 7.38517225e-01
1.80856884e-01 2.05275968e-01 3.03563297e-01 -7.90232241e-01
7.02802002e-01 1.22544563e+00 1.27990216e-01 -1.13552213e+00
-2.75073946e-01 -4.38899435e-02 -1.76499531e-01 -8.57027948e-01
5.61707854e-01 -3.53260189e-01 -5.74810266e-01 1.37422848e+00
7.92014599e-01 2.41167545e-02 3.79634887e-01 6.14562273e-01
1.43053722e+00 5.99279344e-01 -1.19185835e-01 -2.10916311e-01
1.70978534e+00 -1.20752954e+00 -1.01395857e+00 -3.28895003e-01
1.01145601e+00 -6.40908420e-01 9.54595804e-01 4.64807600e-02
-8.33301306e-01 -1.88494563e-01 -7.68418312e-01 -5.15344679e-01
-6.92240596e-01 -3.11877936e-01 7.11032331e-01 4.54113394e-01
-6.34195149e-01 3.61016691e-01 -3.56118232e-01 -3.07712078e-01
7.95881152e-02 2.14378685e-01 -3.54685783e-01 -3.62089664e-01
-1.78076065e+00 1.03216124e+00 7.33357072e-01 -1.25205323e-01
-2.54260033e-01 -4.14035469e-01 -1.04381061e+00 2.33662456e-01
1.21088064e+00 -8.27705085e-01 1.27131104e+00 -4.55994546e-01
-1.35308909e+00 7.75184095e-01 -3.99859160e-01 -3.37627679e-01
2.50360053e-02 -3.04970086e-01 -3.54502380e-01 3.84456009e-01
8.35006535e-02 6.56833453e-03 2.93776721e-01 -1.01129413e+00
-5.61646938e-01 -4.50508118e-01 8.40973854e-01 1.75877884e-01
-2.57916242e-01 1.88593134e-01 -6.81234956e-01 5.99294044e-02
2.68610626e-01 -4.80730236e-01 -1.08696409e-01 -5.40284216e-01
-4.67362016e-01 -9.42250192e-01 6.07118130e-01 -9.02593255e-01
1.72758520e+00 -1.45458400e+00 -9.19520259e-02 3.81008863e-01
6.13826573e-01 4.14250702e-01 6.74849823e-02 7.28482306e-01
1.72025129e-01 2.74364114e-01 -4.96208519e-02 5.97879589e-01
1.05331868e-01 6.44573271e-01 -2.80928969e-01 -3.77188623e-01
1.27615724e-02 1.27093363e+00 -1.15958023e+00 -9.30332124e-01
-4.25003886e-01 -2.60907948e-01 -4.52618837e-01 3.41851562e-01
-8.72488678e-01 3.47044691e-02 -7.87754714e-01 4.31776434e-01
4.51058000e-01 -5.37278175e-01 5.86417615e-01 -2.34612823e-02
3.05819422e-01 5.73020816e-01 -1.12724733e+00 1.51255858e+00
-5.18864274e-01 2.55616546e-01 -2.28616595e-01 -9.31417584e-01
9.36963975e-01 2.42298484e-01 -7.38730878e-02 -5.63175440e-01
1.10174134e-01 1.54481724e-01 -9.64051485e-02 -1.21686411e+00
4.90678191e-01 -3.49136442e-02 -4.46832478e-02 4.08749342e-01
2.91529566e-01 -1.36963248e-01 7.02270627e-01 6.96775675e-01
1.46250629e+00 6.38627559e-02 3.36550474e-01 5.64700477e-02
7.23532319e-01 2.75989473e-01 4.33566064e-01 7.45431662e-01
2.14280918e-01 -3.67491469e-02 9.78759944e-01 -2.37478003e-01
-5.53319812e-01 -9.53994393e-01 6.87653497e-02 9.78406668e-01
2.10511625e-01 -9.96931195e-01 -8.40864599e-01 -1.20410991e+00
-1.73011824e-01 4.77109849e-01 -1.90387934e-01 -1.41210690e-01
-6.29228830e-01 -1.83753371e-01 6.67014956e-01 3.26444507e-01
5.93695641e-01 -7.57174671e-01 -3.59868221e-02 1.66098848e-01
-1.00331163e+00 -1.43449461e+00 -2.36534536e-01 -1.21903829e-01
-7.73570895e-01 -1.56728065e+00 -4.45938446e-02 -1.12050354e+00
7.28884280e-01 5.55709787e-02 1.39605510e+00 5.82819462e-01
1.58856317e-01 4.67796057e-01 -9.04184997e-01 -2.81236410e-01
-2.20745325e-01 2.95200199e-01 -4.12803888e-01 -1.24056242e-01
8.84305060e-01 -4.10210937e-01 -3.36858362e-01 3.60114574e-01
-9.34109509e-01 -1.49532050e-01 2.40649313e-01 6.69490397e-01
5.09781897e-01 1.56281099e-01 7.43827701e-01 -1.25527287e+00
9.91506517e-01 -6.62713945e-01 -4.76428181e-01 8.62093508e-01
-5.78534663e-01 5.27558744e-01 6.64039314e-01 -1.40610442e-01
-1.20390761e+00 -2.22815812e-01 -2.97944963e-01 1.35369316e-01
1.64899249e-02 9.24704015e-01 -1.57735065e-01 -8.79446343e-02
8.51238310e-01 -1.45330846e-01 -2.64844954e-01 -5.25780797e-01
6.51902199e-01 6.74870729e-01 3.82886201e-01 -9.27846730e-01
7.71565735e-01 4.41870578e-02 -5.79464361e-02 -7.19226122e-01
-1.50708115e+00 -7.42952585e-01 -5.73954165e-01 5.92199154e-02
6.95518374e-01 -6.04877830e-01 -8.62100661e-01 -4.64065857e-02
-1.31324840e+00 -1.03999726e-01 -2.05922350e-01 2.99995899e-01
-1.07636042e-01 8.60328794e-01 -6.50317848e-01 -6.37077808e-01
-3.74164104e-01 -4.73579317e-01 5.39900243e-01 3.52236658e-01
-1.64514259e-01 -1.12165320e+00 2.89609969e-01 9.14803982e-01
1.49531448e-02 -5.02566360e-02 1.20587444e+00 -1.14516532e+00
-7.19861805e-01 -2.71077752e-01 -2.82018155e-01 1.90920085e-01
-7.34987035e-02 -2.70454645e-01 -5.16064584e-01 4.14434165e-01
-2.53589571e-01 -5.72282135e-01 6.05520487e-01 -3.65714967e-01
8.10572326e-01 -6.60047412e-01 -3.15323681e-01 -6.17311448e-02
1.23485160e+00 -1.65403783e-01 8.73037100e-01 1.20356336e-01
7.16969967e-01 8.76811743e-01 5.02248406e-01 -4.76215184e-02
9.02270734e-01 2.28673533e-01 -1.15585424e-01 4.83358979e-01
-1.36319235e-01 -7.31770992e-01 4.40221466e-02 1.12984931e+00
-9.03814659e-02 -1.19468711e-01 -9.78332698e-01 8.36882055e-01
-2.05496716e+00 -9.79098678e-01 -6.79017544e-01 1.97651386e+00
1.35229337e+00 5.97312823e-02 -2.60133445e-01 -3.61272246e-02
6.51677251e-01 -1.33068621e-01 -2.07529992e-01 -3.41498703e-01
2.26768292e-02 6.66266978e-01 1.25549555e-01 9.72009897e-01
-5.52245975e-01 1.30950987e+00 5.99089050e+00 8.46505225e-01
-1.48180872e-01 1.79280773e-01 -8.99366736e-02 3.87064487e-01
-5.62258244e-01 6.42646253e-01 -9.14398372e-01 5.13304882e-02
8.13650072e-01 -2.96225935e-01 3.29524755e-01 6.09676957e-01
-5.01349866e-01 -4.09420192e-01 -1.10494149e+00 6.89509332e-01
2.25434333e-01 -1.23055840e+00 1.42155945e-01 -4.25785005e-01
5.79949677e-01 -3.51476640e-01 -8.82508934e-01 7.32047081e-01
5.92436254e-01 -7.62035608e-01 6.22946396e-03 6.07739985e-01
3.14369053e-01 -5.52593410e-01 8.91479671e-01 7.10030138e-01
-1.30733442e+00 3.38856243e-02 -3.96502376e-01 -4.30866808e-01
1.96471989e-01 5.62849402e-01 -7.24108756e-01 1.13566685e+00
5.54889321e-01 1.01382479e-01 -6.93855584e-01 7.18833625e-01
-9.81270492e-01 7.51863778e-01 -4.01686251e-01 -6.50853753e-01
-6.80462718e-02 -2.56756902e-01 2.64116317e-01 9.21185076e-01
-8.57912824e-02 6.70187056e-01 2.21659631e-01 6.50146127e-01
-4.33918983e-01 3.65950733e-01 -6.66751742e-01 -1.42141879e-01
3.82560551e-01 1.17424977e+00 -5.14636099e-01 -6.78207934e-01
-7.51733124e-01 8.68588924e-01 9.03961480e-01 3.24457854e-01
-5.76309502e-01 -9.77405906e-01 3.77171487e-02 1.07054219e-01
2.83029288e-01 -1.99918181e-01 1.66560709e-01 -1.49616349e+00
4.21688795e-01 -8.11521769e-01 9.87798154e-01 -9.91869152e-01
-1.35331559e+00 4.54789490e-01 8.69835243e-02 -4.99765456e-01
-1.52223334e-01 -4.48390216e-01 -4.29746002e-01 7.18792200e-01
-1.53737855e+00 -8.22813332e-01 -1.98692352e-01 7.86971033e-01
1.99881002e-01 3.82611036e-01 8.93058896e-01 3.37425053e-01
-4.42821264e-01 3.93124253e-01 -3.05078238e-01 4.83129442e-01
4.03953552e-01 -1.16563475e+00 1.66534394e-01 9.05875802e-01
4.31697518e-01 8.35969508e-01 4.70167458e-01 -8.57836604e-01
-1.46975052e+00 -7.08849609e-01 1.60862923e+00 -8.35552335e-01
9.91873920e-01 -8.72817338e-02 -1.21736121e+00 8.70285213e-01
3.52645487e-01 -7.24194050e-02 8.43920290e-01 5.45408249e-01
-5.87418795e-01 1.50732733e-02 -9.64229345e-01 4.82916951e-01
1.31862819e+00 -7.12366283e-01 -1.45909154e+00 5.07994652e-01
1.18481338e+00 -5.68995655e-01 -9.91392732e-01 3.26233178e-01
1.12879626e-01 -5.52033722e-01 7.74114370e-01 -1.19490004e+00
4.81897384e-01 -3.52723360e-01 5.35634719e-02 -1.09137011e+00
-2.78996397e-02 -7.27886617e-01 -6.43125176e-01 1.21535206e+00
8.60943913e-01 -7.99426615e-01 6.62541866e-01 7.47339308e-01
1.74250022e-01 -6.78328812e-01 -7.11491823e-01 -7.22137988e-01
-8.95960331e-02 -3.44655693e-01 6.12472117e-01 9.76960897e-01
7.93764293e-01 1.17555714e+00 6.84335008e-02 2.51297355e-01
3.81297678e-01 4.97969955e-01 7.09081173e-01 -1.34344041e+00
-3.39252621e-01 5.51863052e-02 -7.53107890e-02 -1.38001692e+00
1.85371354e-01 -1.12806535e+00 -2.72191495e-01 -2.19022012e+00
1.96080804e-01 -2.78194249e-01 1.75219521e-01 6.71479046e-01
-4.43403423e-01 -2.94336498e-01 -1.98402762e-01 -1.57720894e-01
-1.11634099e+00 2.64616579e-01 1.48286760e+00 -2.68375780e-02
-4.99764793e-02 -1.80697873e-01 -8.51108491e-01 7.94782162e-01
8.37641776e-01 -4.75057870e-01 -5.53673267e-01 -3.41336250e-01
9.64013934e-01 3.55050683e-01 9.12986845e-02 -4.86670494e-01
5.51108062e-01 -2.57760465e-01 -3.38259608e-01 -3.83120805e-01
-8.14716220e-02 -5.94378114e-01 -2.74894893e-01 2.47894004e-01
-2.08751678e-01 -4.38868776e-02 -1.51777804e-01 6.24381065e-01
-5.03342927e-01 -7.13610649e-01 5.38835526e-02 -4.39111024e-01
-7.42584884e-01 2.97589868e-01 -1.70825180e-02 8.30635071e-01
7.56063104e-01 2.52111435e-01 -6.93086624e-01 -3.78411978e-01
-7.18975663e-01 6.98222876e-01 -2.14241162e-01 3.89829665e-01
6.44209623e-01 -1.12275290e+00 -4.83112574e-01 -4.19937521e-01
3.00745815e-01 3.34139258e-01 1.90866113e-01 6.77590430e-01
-7.04869390e-01 4.27992970e-01 4.05812770e-01 8.05904344e-02
-1.33327651e+00 7.23535776e-01 2.39128634e-01 -7.87990272e-01
-3.23057562e-01 7.13758886e-01 -3.65145594e-01 -8.10527563e-01
8.85314494e-02 -2.40598753e-01 -5.86067975e-01 3.84587049e-02
5.38162708e-01 3.33141059e-01 1.92376927e-01 -1.75946042e-01
-1.88204303e-01 5.89127362e-01 -2.19711848e-03 6.54649809e-02
9.07018125e-01 -1.20613217e-01 -6.51520729e-01 7.02335685e-02
1.07862854e+00 3.14585030e-01 -2.21656874e-01 -7.02976644e-01
3.43368262e-01 -2.89164394e-01 -3.70338053e-01 -6.53193712e-01
-5.59087098e-01 6.22438252e-01 -2.57636935e-01 7.35320032e-01
6.82365716e-01 5.96188843e-01 1.17537224e+00 1.14518023e+00
3.98098469e-01 -1.16801345e+00 1.21493950e-01 9.93385196e-01
6.20653927e-01 -1.16571653e+00 -5.09459898e-02 -9.14551914e-01
-4.11965817e-01 1.04157877e+00 8.73578668e-01 3.08039874e-01
5.40616572e-01 3.69077735e-02 4.54693660e-02 -7.17281103e-01
-8.58861268e-01 -7.89291501e-01 2.65944719e-01 3.58503520e-01
2.71107674e-01 -1.12468839e-01 -8.50159824e-01 9.10550356e-01
-4.47782338e-01 -6.07736968e-02 4.04397935e-01 1.09624600e+00
-8.69165301e-01 -1.30817139e+00 -2.98078537e-01 5.25914192e-01
-5.42291462e-01 -2.79133230e-01 -7.95447826e-01 4.96225566e-01
-5.65804504e-02 1.50073540e+00 -3.14939022e-01 -3.16233128e-01
5.42257786e-01 5.27356863e-01 8.41868103e-01 -7.26717651e-01
-4.94892955e-01 -7.95317352e-01 8.32140684e-01 -3.03243905e-01
-2.71384954e-01 1.67942084e-02 -1.66349769e+00 -3.22158158e-01
-8.39100957e-01 7.70837605e-01 1.42090693e-01 1.26556826e+00
4.23131347e-01 3.52159113e-01 3.56249303e-01 6.02949023e-01
-4.38808858e-01 -8.25930834e-01 -2.58326441e-01 5.46579599e-01
-3.95868309e-02 -4.25555795e-01 -1.80096403e-01 5.83481155e-02] | [10.473465919494629, 7.9604291915893555] |
69f8d708-e98e-424e-8b9c-54e4a20501d0 | group-invariant-tensor-train-networks-for | 2206.15051 | null | https://arxiv.org/abs/2206.15051v1 | https://arxiv.org/pdf/2206.15051v1.pdf | Group-invariant tensor train networks for supervised learning | Invariance has recently proven to be a powerful inductive bias in machine learning models. One such class of predictive or generative models are tensor networks. We introduce a new numerical algorithm to construct a basis of tensors that are invariant under the action of normal matrix representations of an arbitrary discrete group. This method can be up to several orders of magnitude faster than previous approaches. The group-invariant tensors are then combined into a group-invariant tensor train network, which can be used as a supervised machine learning model. We applied this model to a protein binding classification problem, taking into account problem-specific invariances, and obtained prediction accuracy in line with state-of-the-art deep learning approaches. | ['Nick Vannieuwenhoven', 'Brent Sprangers'] | 2022-06-30 | null | null | null | null | ['tensor-networks'] | ['methodology'] | [ 3.21248591e-01 2.08135438e-03 -2.69616067e-01 -4.91985142e-01
-2.17204496e-01 -6.56969786e-01 7.95860529e-01 1.00534506e-01
-3.47937435e-01 5.99411249e-01 2.46646047e-01 -3.63766849e-01
-3.30270469e-01 -8.55889797e-01 -7.08189368e-01 -8.84369552e-01
-3.22609544e-01 7.78429747e-01 1.65134192e-01 -5.47105432e-01
3.07429314e-01 9.95693505e-01 -1.22880864e+00 4.64104474e-01
6.00271821e-01 8.91778648e-01 -5.62400162e-01 6.56061769e-01
4.18689707e-03 7.65307784e-01 3.68509106e-02 -2.95099616e-01
1.96836039e-01 -7.42909983e-02 -1.20451283e+00 -3.55045721e-02
4.77734447e-01 -8.27790499e-02 -3.83211166e-01 1.00150156e+00
1.14968725e-01 3.01847637e-01 9.01517928e-01 -1.05748153e+00
-7.29144216e-01 4.91523027e-01 1.05896720e-03 2.51717001e-01
1.68461233e-01 -1.66799068e-01 1.45955098e+00 -9.16098416e-01
7.40332425e-01 1.33166969e+00 4.45016056e-01 3.76834542e-01
-1.89879620e+00 -2.81462669e-01 -1.04624957e-01 2.06919119e-01
-9.56418693e-01 -1.86889037e-01 1.04638362e+00 -8.99101794e-01
9.46286201e-01 3.97141516e-01 6.09372437e-01 8.71695280e-01
5.65363646e-01 7.07336664e-01 1.04643738e+00 -3.05907995e-01
2.53473192e-01 -3.84105951e-01 4.40931529e-01 9.42366540e-01
8.31809565e-02 -9.19008031e-02 -4.28676039e-01 -5.54674029e-01
8.91184568e-01 3.33855838e-01 -7.29547516e-02 -8.34620953e-01
-1.58636832e+00 1.33801579e+00 7.76126921e-01 6.14875555e-01
-5.29477596e-01 3.04609627e-01 4.81536657e-01 2.54280597e-01
6.94165885e-01 5.91275692e-01 -4.86134022e-01 2.13560775e-01
-5.13811409e-01 4.44317460e-01 6.20921731e-01 2.33614832e-01
8.13892722e-01 4.26621027e-02 -6.93329051e-02 7.34273434e-01
4.41291571e-01 2.78064072e-01 4.48096871e-01 -7.63446927e-01
-5.29067293e-02 8.41358006e-01 -1.14362173e-01 -1.08125818e+00
-6.93792284e-01 -6.34812653e-01 -1.07728267e+00 5.27284481e-02
4.49742198e-01 5.33730686e-01 -7.46857285e-01 1.59088182e+00
3.46058786e-01 -2.03662202e-01 -1.75272286e-01 7.95518994e-01
3.27550024e-01 4.12414819e-01 -2.10273743e-01 2.46221982e-02
1.22206807e+00 -4.70563978e-01 -5.30357063e-01 4.13763613e-01
9.74822104e-01 -7.96252787e-01 5.85790515e-01 4.50829118e-01
-7.21519411e-01 -3.09312701e-01 -8.54019165e-01 -1.48765534e-01
-4.81746614e-01 4.09959368e-02 1.35499954e+00 2.98373371e-01
-9.73504782e-01 1.04411483e+00 -1.01517451e+00 -3.42090040e-01
2.38842949e-01 6.84921145e-01 -6.37526810e-01 -1.05057143e-01
-1.00381172e+00 7.66728878e-01 4.50558364e-01 1.40094951e-01
-6.90421164e-01 -7.66839623e-01 -8.18190038e-01 -2.61253506e-01
-1.59208745e-01 -6.68733716e-01 1.08807623e+00 -7.42477953e-01
-1.47093248e+00 8.20150971e-01 -8.64464864e-02 -3.01562041e-01
9.97723043e-02 1.34572675e-02 -2.75725156e-01 6.93946257e-02
-4.13012579e-02 4.34479475e-01 9.64283466e-01 -7.80297458e-01
4.96846251e-03 -5.47608852e-01 2.76320517e-01 -2.37351656e-01
-3.58596146e-01 1.01594619e-01 2.20892847e-01 -5.43944001e-01
6.58220947e-01 -1.24077165e+00 -4.70514476e-01 -1.53361723e-01
-5.12796402e-01 -5.75438261e-01 6.27081633e-01 -5.90823293e-01
8.23315561e-01 -1.76899803e+00 9.55918491e-01 4.77507651e-01
5.22371829e-01 3.57930847e-02 -2.80730367e-01 5.73631227e-01
-6.75924778e-01 -5.58589920e-02 -6.02783114e-02 3.08145862e-02
5.25234714e-02 4.02794540e-01 -4.42797452e-01 6.94654942e-01
3.30353379e-01 8.36781144e-01 -9.29977834e-01 6.52181916e-04
2.35528335e-01 5.20743906e-01 -8.13573778e-01 1.57779351e-01
-2.95814782e-01 5.36590934e-01 -5.98021150e-01 4.68780339e-01
4.61661637e-01 -2.96805114e-01 2.29576603e-01 -7.23298073e-01
3.99090573e-02 4.68130589e-01 -8.19733024e-01 1.49019730e+00
-1.54871061e-01 4.44320649e-01 -5.14470875e-01 -1.53731346e+00
9.97923255e-01 2.30864897e-01 7.78897405e-01 -1.97917387e-01
1.86861917e-01 3.33896577e-01 4.40928698e-01 -2.60884315e-01
1.59139633e-01 -4.07269955e-01 1.39797494e-01 4.67578679e-01
5.50889790e-01 -4.49031107e-02 2.68795967e-01 2.75774211e-01
1.05528641e+00 2.45565996e-02 1.93444416e-01 -5.39198160e-01
5.59626341e-01 -3.14940512e-01 2.56148309e-01 3.24804574e-01
1.63447425e-01 2.50035316e-01 6.28682733e-01 -1.24986708e+00
-1.12611639e+00 -7.96517193e-01 -3.91296148e-01 1.40150750e+00
-5.37325442e-01 -4.18478191e-01 -6.22047067e-01 -5.29202521e-01
1.19445525e-01 2.14588746e-01 -9.64220881e-01 -2.73813426e-01
-4.45555210e-01 -1.01829922e+00 4.13032353e-01 4.85766321e-01
1.73215717e-01 -8.09203684e-01 1.30207136e-01 2.03053370e-01
6.31562248e-02 -9.86706495e-01 -2.41829641e-02 3.67921650e-01
-1.17133951e+00 -1.06773460e+00 -4.38528657e-01 -3.10230166e-01
8.51422668e-01 -5.69888167e-02 9.24211919e-01 1.28857881e-01
-2.12334737e-01 3.90701860e-01 -5.43846227e-02 -2.47159123e-01
-6.35860443e-01 1.82094932e-01 5.78981161e-01 5.28922617e-01
2.19960392e-01 -5.87131202e-01 -3.52725983e-01 1.76832676e-01
-1.21154320e+00 1.95203777e-02 4.23990995e-01 1.05899620e+00
4.92831796e-01 -3.63303691e-01 1.86535910e-01 -6.93680584e-01
4.60603952e-01 -3.04233104e-01 -6.67824805e-01 4.43457160e-03
-1.64856970e-01 6.78113699e-01 5.74920595e-01 -5.30451953e-01
-4.08606380e-01 2.91478664e-01 -1.78558454e-01 -2.98329055e-01
1.29610151e-01 8.78965795e-01 1.10916816e-01 -7.13581324e-01
6.50009811e-01 1.22943543e-01 1.05327390e-01 -5.89186072e-01
3.42134237e-01 2.15800434e-01 1.89574718e-01 -7.87546694e-01
1.00396502e+00 5.34724712e-01 7.62724280e-01 -9.49371099e-01
-1.08177447e+00 -5.26442528e-01 -1.31351924e+00 -1.85549557e-01
8.59916627e-01 -4.19380337e-01 -7.05279589e-01 4.74489361e-01
-1.29869676e+00 -1.45031631e-01 -1.83090158e-02 6.92990184e-01
-6.25342011e-01 2.66658902e-01 -6.21362984e-01 -3.99863929e-01
-5.32089472e-02 -9.62205946e-01 9.04522777e-01 -3.22566450e-01
-1.36066407e-01 -1.40193832e+00 3.97853434e-01 2.80278951e-01
4.29010242e-01 3.81976187e-01 1.26514304e+00 -1.01399875e+00
-5.80009103e-01 -5.04526913e-01 -7.56734833e-02 4.80795532e-01
-7.26869032e-02 1.79146796e-01 -7.56965458e-01 -3.74325484e-01
-1.51102155e-01 -1.54358193e-01 9.28432167e-01 1.29019499e-01
1.22260082e+00 -3.35059196e-01 -3.15458596e-01 6.19703889e-01
1.20907044e+00 -1.55881420e-01 4.34311599e-01 2.16651440e-01
9.77207363e-01 2.10402071e-01 5.44306561e-02 1.13445170e-01
6.88185692e-02 1.05221784e+00 3.74377310e-01 -1.39033437e-01
3.55241150e-01 -6.21691942e-02 1.18226990e-01 1.21124732e+00
-8.25450182e-01 3.83232325e-01 -9.04801726e-01 2.46835157e-01
-1.94660485e+00 -1.01168144e+00 -1.98676795e-01 1.93870521e+00
6.92971051e-01 2.64427718e-02 6.02409132e-02 3.34847838e-01
2.49132350e-01 7.56927431e-02 -2.53443658e-01 -5.77579021e-01
4.31662723e-02 4.81141031e-01 4.35258627e-01 4.40698415e-01
-1.32292414e+00 7.63167500e-01 7.51519680e+00 3.79465193e-01
-1.35777414e+00 2.52821520e-02 2.23935112e-01 4.13732827e-01
-3.64123642e-01 2.95758285e-02 -4.50831026e-01 -1.80554739e-03
8.34540427e-01 4.24743444e-02 4.95118171e-01 7.69907951e-01
-1.04237683e-01 4.67656136e-01 -1.55269945e+00 6.28609478e-01
1.01499520e-01 -1.55074656e+00 3.68373632e-01 4.05641258e-01
6.73015714e-01 2.24440038e-01 1.00670442e-01 2.79400289e-01
3.93515587e-01 -9.81187046e-01 2.02340916e-01 8.53252590e-01
3.46763968e-01 -5.56184649e-01 7.61681259e-01 1.30289301e-01
-8.67781162e-01 9.62324664e-02 -6.71779096e-01 -3.63125592e-01
-1.71296179e-01 7.98387170e-01 -7.17041016e-01 5.70041001e-01
5.13470113e-01 8.89946520e-01 -5.69180548e-01 6.88237071e-01
2.83403434e-02 7.87825525e-01 -3.86236548e-01 -1.26713023e-01
3.91402453e-01 -4.66310591e-01 5.81927717e-01 9.70348060e-01
2.86250245e-02 1.60980836e-01 1.42086387e-01 8.19215536e-01
-1.33204192e-01 1.38139755e-01 -8.38710546e-01 -4.23095793e-01
-3.39395493e-01 1.34155560e+00 -7.62660444e-01 -3.48821104e-01
-2.68929005e-01 7.94130147e-01 7.39233732e-01 4.17120934e-01
-5.52627921e-01 -2.13189051e-01 8.88795674e-01 -1.01882800e-01
4.04840052e-01 -6.24895871e-01 5.47957644e-02 -1.54373062e+00
-1.62378177e-01 -7.85783947e-01 2.19441712e-01 -5.36997139e-01
-1.44977295e+00 3.48607540e-01 1.68193635e-02 -9.57239449e-01
-2.20022008e-01 -1.35376000e+00 -3.44818771e-01 6.40395820e-01
-1.09766340e+00 -1.30282903e+00 8.70200619e-02 6.22733176e-01
-1.55261993e-01 -4.55155998e-01 1.25173223e+00 1.97322458e-01
-4.43675488e-01 2.14013427e-01 3.15926164e-01 3.43630075e-01
2.63450265e-01 -1.28231835e+00 2.01679453e-01 6.20679557e-01
3.93754989e-01 1.02759016e+00 6.79326177e-01 -2.69419551e-01
-1.75782990e+00 -1.05530095e+00 9.46577787e-01 -7.14671016e-01
1.15893722e+00 -5.41862845e-01 -1.01745045e+00 9.27703619e-01
-2.81647712e-01 6.02328658e-01 8.88069808e-01 5.33510208e-01
-8.23247910e-01 -1.72864288e-01 -7.45001793e-01 4.80262041e-01
8.93011987e-01 -8.16345632e-01 -4.68987495e-01 8.48075747e-01
6.41020119e-01 -3.66348863e-01 -1.19207489e+00 4.06544119e-01
5.62873602e-01 -8.25537682e-01 1.10834217e+00 -1.40117633e+00
3.49215776e-01 -1.89118564e-01 -3.58246118e-01 -1.37381792e+00
-8.38001370e-01 -6.60864294e-01 -2.93527264e-02 5.46484590e-01
1.67378023e-01 -7.44292855e-01 2.89736062e-01 4.67263222e-01
-7.26060718e-02 -6.95458472e-01 -1.03766227e+00 -8.06541681e-01
1.72681034e-01 -4.61575806e-01 4.60498303e-01 1.12921882e+00
1.81553766e-01 5.88046491e-01 -1.63179725e-01 -7.94419497e-02
5.26738346e-01 1.13604516e-01 5.78822315e-01 -1.71302462e+00
-9.18922722e-02 -4.32081968e-01 -1.20248389e+00 -7.11441815e-01
4.44606274e-01 -1.50376379e+00 -4.33845669e-01 -1.19845176e+00
2.70777375e-01 -3.10658902e-01 -5.62354565e-01 8.18223715e-01
2.45116696e-01 5.70694208e-01 -1.24054411e-02 1.88093379e-01
-4.10744101e-01 4.77074951e-01 1.21160245e+00 -2.48622730e-01
1.27621531e-01 2.82101110e-02 -3.98211062e-01 8.27958286e-01
7.03172982e-01 -3.33449036e-01 1.87987629e-02 3.05668656e-02
4.40744370e-01 -4.18313384e-01 4.60834920e-01 -8.05070281e-01
-1.34471506e-01 -3.58082950e-01 3.91033590e-01 -2.48243853e-01
2.57424474e-01 -7.04900682e-01 9.33583155e-02 6.02951288e-01
-4.53284711e-01 6.56072982e-03 1.10118590e-01 2.24027634e-01
-1.33541241e-01 -1.25556946e-01 6.61121964e-01 2.50323743e-01
-3.70037138e-01 5.25426924e-01 -3.98792118e-01 -2.96720296e-01
5.63537359e-01 2.05360353e-01 -7.47801363e-02 -1.05835490e-01
-7.98275113e-01 -6.00200713e-01 1.35375977e-01 3.79947662e-01
3.99236798e-01 -1.70617807e+00 -5.93861341e-01 3.33460152e-01
2.21584901e-01 -4.61076260e-01 -4.59072813e-02 9.36172903e-01
-5.15639842e-01 7.86298752e-01 -4.56161499e-01 -9.72685635e-01
-1.11383224e+00 7.23121762e-01 5.58536589e-01 -4.31606442e-01
-3.09938222e-01 5.18691361e-01 4.08899225e-02 -7.87874579e-01
-2.80325890e-01 -7.30347812e-01 -1.50774091e-01 -1.70838609e-01
3.37178260e-01 1.68595135e-01 3.19954962e-01 -9.51836050e-01
-4.56566900e-01 6.03406072e-01 -2.56141089e-02 9.36402306e-02
1.56502175e+00 6.35723293e-01 -7.74584234e-01 5.97528458e-01
1.57301533e+00 -2.49311194e-01 -6.11737132e-01 -4.55770969e-01
-2.71987561e-02 -1.63841739e-01 1.35650784e-01 -2.85619646e-01
-9.02194500e-01 9.48673725e-01 3.50833803e-01 4.49342579e-01
7.23970294e-01 -1.15841404e-02 1.66753113e-01 8.98471773e-01
3.83721173e-01 -6.70053601e-01 2.35964075e-01 8.77579689e-01
1.26611948e+00 -1.31439018e+00 1.66952327e-01 -4.29231733e-01
7.33887870e-03 1.51087534e+00 7.87788779e-02 -4.52180445e-01
8.70886862e-01 -9.73562747e-02 -2.69998968e-01 -3.71021420e-01
-6.95141912e-01 -1.17102884e-01 1.00610042e+00 6.18502855e-01
8.76021028e-01 3.68049115e-01 -3.36047143e-01 1.80377573e-01
-3.42229664e-01 3.52081284e-02 2.75943935e-01 5.80083847e-01
-2.99717933e-01 -1.20592225e+00 -3.04520458e-01 4.43539917e-01
-3.28372836e-01 7.75026083e-02 -2.94651121e-01 4.22247082e-01
-7.54673854e-02 4.20374304e-01 -2.32104980e-03 -4.48383868e-01
5.03671356e-02 4.10778612e-01 7.97649086e-01 -5.26631653e-01
-1.57812849e-01 -2.57959634e-01 -1.25787869e-01 -7.27194071e-01
-6.04948163e-01 -6.47603989e-01 -1.07737446e+00 -3.06468695e-01
-1.54225230e-01 2.17069779e-02 7.12421000e-01 1.25343800e+00
1.68464959e-01 4.71837908e-01 6.05263531e-01 -1.24938345e+00
-6.26405835e-01 -1.03319561e+00 -4.31787312e-01 6.87940955e-01
3.58208716e-01 -9.20408964e-01 -3.47619444e-01 7.48888105e-02] | [5.976195812225342, 5.038966655731201] |
e90f5619-b135-41d4-bb36-204f6aa986e4 | keyphrase-generation-beyond-the-boundaries-of | 2112.06776 | null | https://arxiv.org/abs/2112.06776v2 | https://arxiv.org/pdf/2112.06776v2.pdf | Keyphrase Generation Beyond the Boundaries of Title and Abstract | Keyphrase generation aims at generating important phrases (keyphrases) that best describe a given document. In scholarly domains, current approaches have largely used only the title and abstract of the articles to generate keyphrases. In this paper, we comprehensively explore whether the integration of additional information from the full text of a given article or from semantically similar articles can be helpful for a neural keyphrase generation model or not. We discover that adding sentences from the full text, particularly in the form of the extractive summary of the article can significantly improve the generation of both types of keyphrases that are either present or absent from the text. Experimental results with three widely used models for keyphrase generation along with one of the latest transformer models suitable for longer documents, Longformer Encoder-Decoder (LED) validate the observation. We also present a new large-scale scholarly dataset FullTextKP for keyphrase generation. Unlike prior large-scale datasets, FullTextKP includes the full text of the articles along with the title and abstract. We release the source code at https://github.com/kgarg8/FullTextKP. | ['Cornelia Caragea', 'Jishnu Ray Chowdhury', 'Krishna Garg'] | 2021-12-13 | null | null | null | null | ['keyphrase-generation'] | ['natural-language-processing'] | [ 1.53616190e-01 2.72952527e-01 -4.52884197e-01 2.79452175e-01
-1.13793862e+00 -8.09556007e-01 1.14748693e+00 4.40580815e-01
-1.89430609e-01 1.06253493e+00 1.02617025e+00 -1.89498454e-01
-1.03465892e-01 -8.63067210e-01 -1.12413704e+00 -3.52245152e-01
3.79368871e-01 2.37617806e-01 -1.75368667e-01 -3.73089284e-01
7.70635486e-01 2.25661188e-01 -1.32527673e+00 6.03113115e-01
8.97011101e-01 5.19268453e-01 6.68050349e-01 7.53963768e-01
-5.29181778e-01 7.92102993e-01 -8.70716453e-01 -4.06159461e-01
1.71344846e-01 -5.84844351e-01 -7.99014032e-01 -1.71430573e-01
8.75617564e-01 -5.01144230e-01 -7.53455400e-01 7.83644974e-01
5.00264823e-01 -8.34167153e-02 6.70090973e-01 -1.21557796e+00
-9.59802091e-01 1.54741395e+00 -4.27873671e-01 2.15200752e-01
5.29497564e-01 -3.19104731e-01 1.48423398e+00 -1.14188278e+00
9.62868214e-01 1.09643042e+00 3.76081496e-01 3.32832485e-01
-8.47844422e-01 -6.83962762e-01 -1.45095751e-01 2.06535995e-01
-1.27866554e+00 -5.36445320e-01 1.02720058e+00 -6.97604269e-02
8.28969419e-01 2.68159777e-01 7.05441177e-01 1.50113189e+00
5.45406580e-01 1.21016073e+00 8.02391469e-01 -5.47033131e-01
-1.49627239e-01 9.42335054e-02 1.14724055e-01 4.62519556e-01
3.64992112e-01 -2.19793916e-01 -9.21814442e-01 -3.02216649e-01
6.13682389e-01 -1.05878279e-01 -2.59640276e-01 3.44145179e-01
-1.75537884e+00 7.08946466e-01 3.98965599e-03 2.53529280e-01
-6.39116287e-01 3.41463566e-01 2.69590139e-01 3.02194387e-01
2.65199840e-01 9.75122392e-01 -3.33961487e-01 -4.21494812e-01
-1.24275744e+00 8.97392869e-01 1.16480756e+00 1.46292508e+00
6.48381650e-01 -1.91249430e-01 -7.88564861e-01 6.48087561e-01
4.30678437e-03 6.11055732e-01 4.82816964e-01 -8.37298393e-01
7.79395342e-01 5.73712707e-01 1.23560220e-01 -1.07092762e+00
8.29849690e-02 -4.26284790e-01 -6.77163601e-01 -6.90321565e-01
-6.01580329e-02 -3.34898591e-01 -8.80038440e-01 1.25602365e+00
-1.43253222e-01 -2.24806182e-02 1.22694813e-01 1.77083611e-01
1.43092442e+00 1.24435759e+00 -3.97272617e-01 -1.57544121e-01
1.42033505e+00 -1.02646744e+00 -8.39264095e-01 -1.57933921e-01
3.09523225e-01 -1.16505039e+00 9.11312401e-01 2.52377182e-01
-1.13153648e+00 -5.07210195e-01 -9.77702916e-01 -5.13209999e-01
-6.51018023e-01 3.99542481e-01 2.98135668e-01 -9.60510820e-02
-9.50040340e-01 5.82384229e-01 -1.63665891e-01 -1.97884127e-01
3.10047835e-01 -3.17418188e-01 -1.41801327e-01 -1.75044954e-01
-1.42292440e+00 5.86657941e-01 6.08250022e-01 -3.34405571e-01
-7.71685243e-01 -1.24235499e+00 -5.81560254e-01 1.11406617e-01
7.45863914e-01 -9.49665904e-01 1.41061521e+00 -1.70637950e-01
-1.05552316e+00 4.98509794e-01 -3.77850443e-01 -3.88263404e-01
3.17536891e-01 -5.74957848e-01 -1.53728798e-01 4.54215020e-01
3.75060618e-01 7.92081475e-01 9.71523583e-01 -1.11858881e+00
-8.19292367e-01 6.41443282e-02 1.90031052e-01 1.60736918e-01
-5.19710302e-01 -3.64116952e-02 -7.31655955e-01 -1.42650199e+00
-3.71745199e-01 -9.38137114e-01 1.43899873e-01 -3.15792501e-01
-1.03776503e+00 -3.53265464e-01 7.39760876e-01 -8.74428809e-01
1.59247184e+00 -1.72303700e+00 2.13086516e-01 -5.62922992e-02
2.28838742e-01 -3.22750121e-01 -1.97285444e-01 1.32632446e+00
1.50195241e-01 4.37322766e-01 -2.50349361e-02 -7.71513730e-02
1.69078320e-01 8.90350193e-02 -1.08705008e+00 -2.08077744e-01
-4.16190252e-02 1.30536759e+00 -9.06562030e-01 -5.98906517e-01
-2.83674061e-01 2.05775797e-01 -2.01869264e-01 1.34419352e-01
-5.57960033e-01 -1.78577323e-02 -7.09671795e-01 4.90356356e-01
1.32582143e-01 -2.44137719e-01 -5.10751247e-01 -4.31062520e-01
-5.37980199e-02 6.19239151e-01 -9.92236972e-01 1.81872141e+00
-4.52671021e-01 8.18651438e-01 -4.12941188e-01 -3.39108855e-01
6.62564337e-01 5.16355157e-01 5.79356670e-01 -4.58384544e-01
-2.75505453e-01 3.20858836e-01 -4.06834811e-01 -8.61989483e-02
1.27966797e+00 1.77207187e-01 -2.85156071e-01 7.66760528e-01
1.45684570e-01 -6.59606099e-01 9.17246282e-01 9.83780622e-01
1.04524815e+00 -2.15460807e-02 2.52813160e-01 -1.69834331e-01
2.24486202e-01 8.64561275e-02 -1.23422079e-01 1.27799344e+00
7.91983604e-01 7.49201596e-01 6.21345699e-01 9.33241919e-02
-1.38328767e+00 -7.39063203e-01 9.52885672e-02 7.87559092e-01
-5.77360354e-02 -1.28364444e+00 -3.12510937e-01 -6.92890763e-01
1.71938553e-01 1.10012031e+00 -6.37288392e-01 -2.49350294e-01
-6.24331415e-01 -4.81035888e-01 8.01102579e-01 4.48175043e-01
2.82843858e-01 -1.09578085e+00 -1.70227170e-01 2.46599689e-01
-5.09985089e-01 -1.14544463e+00 -7.72321939e-01 7.22828731e-02
-6.37226224e-01 -7.46120334e-01 -1.27897000e+00 -5.72929442e-01
6.76679373e-01 2.51372457e-01 1.30092871e+00 -1.99412536e-02
-1.29391238e-01 5.54622531e-01 -7.70960152e-01 -5.15683949e-01
-7.16534972e-01 3.60306293e-01 -2.25407809e-01 -3.87995213e-01
1.00600207e-02 -4.39362496e-01 -2.06680998e-01 -4.64646250e-01
-1.15883672e+00 4.64398652e-01 9.63503003e-01 6.21310413e-01
4.91889626e-01 3.33368212e-01 4.25795555e-01 -8.87350857e-01
1.35466290e+00 -7.10110247e-01 -4.20765132e-02 3.41242194e-01
-7.85524130e-01 4.71875310e-01 8.05784583e-01 -3.94038618e-01
-9.34067726e-01 -6.72149777e-01 6.14143955e-03 -1.96880773e-01
1.80690482e-01 8.82321298e-01 1.42957479e-01 6.24059618e-01
4.77400422e-01 8.54015648e-01 -4.94728208e-01 -7.61152804e-01
7.46046424e-01 5.78470111e-01 5.20207584e-01 -8.54505181e-01
1.07265747e+00 1.57473475e-01 -2.78405752e-02 -9.08349991e-01
-7.76281416e-01 -4.61062223e-01 -5.40318787e-01 -1.25322044e-01
2.84502268e-01 -1.06198609e+00 1.06169306e-01 3.97977978e-01
-1.30041659e+00 1.24331720e-01 -6.60799205e-01 1.27100900e-01
-2.62607634e-01 5.29375076e-01 -6.29614711e-01 -2.87239030e-02
-9.06995535e-01 -7.55020380e-01 1.35193491e+00 3.00788850e-01
-4.97672141e-01 -8.06321561e-01 -1.18985198e-01 1.62462696e-01
1.92084759e-01 1.46897491e-02 1.14888370e+00 -8.32267702e-01
-5.35162032e-01 -5.30681670e-01 -1.27840504e-01 2.11914688e-01
4.00480479e-01 3.80789310e-01 -5.01585960e-01 -7.50070363e-02
-3.66636306e-01 -2.03908131e-01 1.25424600e+00 2.35786885e-01
1.04169965e+00 -9.61052895e-01 -3.96569997e-01 2.45085597e-01
1.11681926e+00 2.06136964e-02 5.13975084e-01 3.66949230e-01
9.81266499e-01 4.61511940e-01 4.01541531e-01 6.47175074e-01
6.92420125e-01 3.83100867e-01 -1.16140306e-01 2.56081045e-01
-4.12996083e-01 -8.78955185e-01 6.51523829e-01 1.19662857e+00
8.46331790e-02 -6.16988182e-01 -6.96603417e-01 7.01367736e-01
-1.50420475e+00 -1.15036881e+00 -8.75123292e-02 1.72933340e+00
1.43212497e+00 1.34164914e-01 -1.94999114e-01 2.92735342e-02
2.41280258e-01 5.80954313e-01 -2.69944072e-01 -1.12367131e-01
-3.51495385e-01 1.73000976e-01 5.05406201e-01 4.26810890e-01
-6.93326771e-01 8.73704493e-01 5.89034557e+00 1.10817695e+00
-7.73451209e-01 -4.68122602e-01 2.39368021e-01 -1.37346715e-01
-9.32772815e-01 1.26263827e-01 -1.31691718e+00 5.32688618e-01
9.86617208e-01 -1.00187314e+00 1.64888114e-01 6.28426433e-01
2.24660635e-01 -1.51836142e-01 -1.26022756e+00 8.12827110e-01
2.45738015e-01 -1.83951366e+00 8.11162889e-01 7.72983804e-02
8.51102114e-01 -2.48947054e-01 1.03594512e-02 2.30528176e-01
3.38737816e-01 -7.63244152e-01 9.45284486e-01 7.49920070e-01
5.95535517e-01 -5.91130257e-01 4.58523244e-01 3.06902051e-01
-1.03329372e+00 1.58491865e-01 -2.70344079e-01 3.04712921e-01
7.55641386e-02 7.74309397e-01 -1.03433204e+00 6.85678303e-01
4.63295400e-01 1.25800574e+00 -9.56003428e-01 7.70946562e-01
-5.34118772e-01 6.07092500e-01 -2.44399115e-01 -2.78767109e-01
4.07237053e-01 1.11350566e-01 9.62037802e-01 1.38913035e+00
6.22152269e-01 -3.37184370e-02 -4.58738171e-02 1.12587845e+00
-5.48442721e-01 2.10645780e-01 -5.63656509e-01 -8.63686979e-01
7.85798907e-01 1.19188988e+00 -5.79232752e-01 -8.51720035e-01
-2.91103661e-01 1.03729868e+00 -2.21024558e-01 4.62332785e-01
-3.44398111e-01 -7.58571148e-01 1.75830558e-01 9.47603732e-02
3.62319887e-01 -6.14749342e-02 -1.89175695e-01 -1.37216687e+00
3.58877391e-01 -1.15229356e+00 2.64858484e-01 -1.09335744e+00
-1.08263457e+00 3.15541983e-01 4.36513275e-01 -1.06086421e+00
-4.75498557e-01 -1.91629663e-01 -6.21423244e-01 8.80825400e-01
-1.33249021e+00 -1.34851813e+00 -4.74463925e-02 2.30319634e-01
9.60432708e-01 -2.72988617e-01 6.67516530e-01 -6.80744201e-02
-2.41660252e-01 3.08786303e-01 3.34198713e-01 2.44867653e-01
7.88219094e-01 -1.58770323e+00 7.38665760e-01 1.08143878e+00
4.66111988e-01 1.12735069e+00 8.64696681e-01 -1.09912920e+00
-1.82512820e+00 -8.05899203e-01 1.23561323e+00 -5.04970253e-01
8.42653751e-01 -3.41518879e-01 -7.01637685e-01 6.26660705e-01
7.41889656e-01 -8.04368377e-01 4.44678128e-01 -2.14503974e-01
-2.02833071e-01 3.39014158e-02 -4.12180394e-01 1.11824381e+00
6.27727270e-01 -5.99632323e-01 -1.01536751e+00 3.73366356e-01
1.21849263e+00 -4.32578713e-01 -9.96026337e-01 2.14558065e-01
6.17390633e-01 -2.35713899e-01 1.04808199e+00 -3.44143242e-01
1.30670953e+00 -1.01858877e-01 -5.99778853e-02 -1.47769690e+00
-2.49895707e-01 -8.65747273e-01 -9.43774819e-01 1.36545670e+00
5.61349034e-01 -1.03259884e-01 5.17425776e-01 3.16562861e-01
-1.84954226e-01 -7.14380622e-01 -4.88227963e-01 -4.30037856e-01
2.67283618e-01 -3.33449602e-01 5.92333257e-01 7.35982597e-01
5.27395271e-02 6.88975692e-01 -5.20121574e-01 -3.30579072e-01
5.64640880e-01 3.83962393e-01 8.68780673e-01 -1.03829861e+00
-3.64278182e-02 -5.52395880e-01 2.97076762e-01 -1.27721953e+00
-4.52173315e-02 -1.03411329e+00 -2.15657324e-01 -2.15225387e+00
4.40588593e-01 6.39145300e-02 7.38737360e-02 6.38718307e-01
-4.38353926e-01 -2.14832336e-01 3.07449549e-01 4.39480603e-01
-1.33917958e-01 6.57283545e-01 1.59504068e+00 -4.16108489e-01
-9.50894952e-02 -9.13137347e-02 -1.32404649e+00 3.27888578e-01
5.45730710e-01 -5.81457496e-01 -5.74359119e-01 -1.31655216e-01
7.39864826e-01 3.81377824e-02 3.27624142e-01 -6.35860085e-01
4.31106061e-01 -2.70918250e-01 4.30953860e-01 -9.46309686e-01
1.94145843e-01 -5.66093445e-01 -8.15086067e-02 2.33941823e-01
-8.63217652e-01 2.75603354e-01 3.44371140e-01 4.70799774e-01
-1.96308836e-01 -4.99154717e-01 -1.99949406e-02 -5.12482107e-01
-6.82401299e-01 2.42676049e-01 -5.52363098e-01 3.65174860e-01
4.47776943e-01 2.51610614e-02 -5.14562964e-01 -6.61981881e-01
-1.23184249e-01 1.01239897e-01 2.27118000e-01 8.86221826e-01
9.51385558e-01 -1.27029204e+00 -1.32224154e+00 -1.37554392e-01
2.87133694e-01 -1.23295456e-01 -1.84492469e-01 5.06480455e-01
-2.99624920e-01 8.61855507e-01 3.37574631e-02 2.52767384e-01
-1.10020936e+00 3.01545143e-01 -2.65430003e-01 -3.89004678e-01
-8.97980213e-01 7.65487313e-01 -6.96759969e-02 -1.31537274e-01
8.30872282e-02 -7.54455447e-01 -3.02873939e-01 6.01804018e-01
8.07354689e-01 2.92418003e-01 -1.00598723e-01 -5.69862723e-01
2.30297014e-01 2.36991391e-01 -6.10066950e-01 -3.73263210e-01
1.43507946e+00 -1.75094381e-01 -3.84264201e-01 5.87504208e-01
1.21708834e+00 5.96904397e-01 -6.18858695e-01 -5.29382944e-01
-3.04460363e-03 -1.21016465e-01 2.48951748e-01 -9.02774274e-01
-7.69623637e-01 4.54116434e-01 -4.13024694e-01 1.35233290e-02
8.31866086e-01 2.22276151e-01 1.21397817e+00 6.73105657e-01
1.29191339e-01 -1.18917131e+00 2.17123136e-01 5.59510291e-01
1.39708364e+00 -8.53039086e-01 6.70428038e-01 8.08371045e-03
-6.66466177e-01 1.32528841e+00 2.74996489e-01 2.11777538e-01
5.03340125e-01 1.47959456e-01 -2.74037421e-01 -3.38072032e-01
-1.04002666e+00 1.67621270e-01 8.05151045e-01 8.46873894e-02
3.94681245e-01 -2.63306886e-01 -2.69900739e-01 7.13525176e-01
-7.82262743e-01 -1.12117931e-01 1.11668324e+00 1.10977805e+00
-3.59322220e-01 -1.05838597e+00 -3.41565669e-01 9.63465691e-01
-7.61059582e-01 -6.85834408e-01 -9.31287527e-01 4.89776045e-01
-2.61488825e-01 7.22115517e-01 -2.81090885e-01 -3.05163682e-01
8.43575299e-02 1.75250918e-01 2.40101591e-01 -8.24193180e-01
-5.03153920e-01 1.35387003e-01 1.83168739e-01 -1.33380875e-01
-2.36802146e-01 -4.40589666e-01 -1.18215728e+00 -5.73623300e-01
-8.98346230e-02 4.29277480e-01 4.57629055e-01 7.80184031e-01
5.23263693e-01 6.35600269e-01 4.35156673e-01 -5.90335190e-01
-3.45749259e-01 -1.30527890e+00 -5.39624214e-01 2.20789254e-01
3.85785937e-01 -5.40716760e-02 -1.93942055e-01 6.32340312e-01] | [12.29856014251709, 8.948966026306152] |
c78b6f65-a235-4327-b5b0-3b2cc28dd940 | batch-incremental-triplet-sampling-for | 2007.0561 | null | https://arxiv.org/abs/2007.05610v2 | https://arxiv.org/pdf/2007.05610v2.pdf | Batch-Incremental Triplet Sampling for Training Triplet Networks Using Bayesian Updating Theorem | Variants of Triplet networks are robust entities for learning a discriminative embedding subspace. There exist different triplet mining approaches for selecting the most suitable training triplets. Some of these mining methods rely on the extreme distances between instances, and some others make use of sampling. However, sampling from stochastic distributions of data rather than sampling merely from the existing embedding instances can provide more discriminative information. In this work, we sample triplets from distributions of data rather than from existing instances. We consider a multivariate normal distribution for the embedding of each class. Using Bayesian updating and conjugate priors, we update the distributions of classes dynamically by receiving the new mini-batches of training data. The proposed triplet mining with Bayesian updating can be used with any triplet-based loss function, e.g., triplet-loss or Neighborhood Component Analysis (NCA) loss. Accordingly, Our triplet mining approaches are called Bayesian Updating Triplet (BUT) and Bayesian Updating NCA (BUNCA), depending on which loss function is being used. Experimental results on two public datasets, namely MNIST and histopathology colorectal cancer (CRC), substantiate the effectiveness of the proposed triplet mining method. | ['Fakhri Karray', 'Milad Sikaroudi', 'Mark Crowley', 'H. R. Tizhoosh', 'Benyamin Ghojogh'] | 2020-07-10 | null | null | null | null | ['histopathological-image-classification'] | ['medical'] | [ 2.54596531e-01 -2.02118888e-01 -2.89454609e-01 -5.71001232e-01
-7.13129103e-01 -2.18298301e-01 5.33496082e-01 3.46900105e-01
-4.43393528e-01 8.46571088e-01 -1.21604584e-01 -2.47042164e-01
-8.70881319e-01 -9.64751005e-01 -4.99960274e-01 -1.23462713e+00
3.09024360e-02 6.88336372e-01 1.48653015e-01 9.86478105e-02
2.11419225e-01 3.10522527e-01 -1.36088657e+00 3.86051297e-01
1.09723568e+00 9.21098411e-01 -4.25165927e-04 2.17504054e-01
-2.43926764e-01 2.89843768e-01 -4.48702157e-01 -4.48371261e-01
5.38171291e-01 -3.28254610e-01 -2.47385815e-01 7.52702951e-02
5.10907918e-02 -7.10955728e-03 -1.58516869e-01 1.07075965e+00
4.46380138e-01 1.57805026e-01 1.03483653e+00 -1.35319710e+00
-2.03676715e-01 7.05496848e-01 -9.38432574e-01 2.46587694e-01
5.65978736e-02 -2.32799619e-01 1.00776911e+00 -1.05410278e+00
3.10284346e-01 1.14805794e+00 6.41073823e-01 2.14473680e-01
-1.28152096e+00 -9.27205563e-01 2.30089590e-01 5.22932768e-01
-1.64625752e+00 -4.01834309e-01 1.05217063e+00 -2.88251191e-01
4.01212722e-01 4.05966818e-01 6.67770386e-01 1.13893223e+00
1.56254902e-01 8.16190362e-01 1.22045147e+00 -5.66074312e-01
3.39243233e-01 5.65305889e-01 4.38203901e-01 5.10282397e-01
7.68998742e-01 1.04351014e-01 -6.13073111e-01 -7.87532747e-01
2.20297664e-01 3.14322770e-01 -1.69154957e-01 -6.76334143e-01
-1.11886668e+00 8.56678009e-01 -9.57393795e-02 6.50967807e-02
-2.63538301e-01 -2.04604119e-01 3.57248336e-01 3.08140725e-01
4.15758014e-01 -2.81029314e-01 -4.16843861e-01 1.99408472e-01
-8.02652419e-01 -1.04014456e-01 5.24524689e-01 9.06499803e-01
9.77544844e-01 -2.36693576e-01 -1.25673756e-01 8.71460497e-01
6.41210735e-01 2.74808705e-01 6.21890545e-01 -2.36934319e-01
5.79143941e-01 6.52571499e-01 -1.47092879e-01 -1.14111054e+00
-1.85148910e-01 -3.26742470e-01 -8.48105073e-01 -2.12728575e-01
3.24511260e-01 -2.35226110e-01 -8.55143070e-01 1.55263627e+00
6.72187209e-01 4.97354597e-01 -3.20846513e-02 3.82011592e-01
5.36062300e-01 3.82078558e-01 -2.05415651e-01 -3.57908547e-01
1.20857835e+00 -2.32826903e-01 -5.92780888e-01 3.94734353e-01
5.42353928e-01 -5.45908928e-01 6.99207008e-01 6.32597029e-01
-4.66617048e-01 -1.97133437e-01 -1.05199099e+00 4.83834624e-01
-3.30234647e-01 2.40752220e-01 6.31091774e-01 8.36551368e-01
-6.57773614e-01 5.23896098e-01 -1.14159024e+00 -2.46973023e-01
5.06196260e-01 4.68563616e-01 -3.81853461e-01 -1.46456689e-01
-9.92089093e-01 3.64345431e-01 6.28216624e-01 1.72658235e-01
-7.09746420e-01 -6.98960245e-01 -6.73464715e-01 -1.72589734e-01
3.66375178e-01 -7.09370315e-01 4.26515669e-01 -6.40965760e-01
-1.20145226e+00 4.75718528e-01 -9.20276344e-02 -6.75308585e-01
5.17500341e-01 6.60295561e-02 -4.38542694e-01 6.89363033e-02
-2.24536508e-02 2.12745443e-01 1.06898880e+00 -1.16636288e+00
-7.24021316e-01 -5.29257953e-01 -9.89331827e-02 1.63255051e-01
-8.72944474e-01 -3.17127615e-01 -2.96586275e-01 -4.28922206e-01
5.01254618e-01 -1.00255156e+00 -1.25314191e-01 -5.35912998e-02
-8.01575780e-01 -5.48010468e-01 9.86736834e-01 -3.26557189e-01
1.25426006e+00 -2.29212022e+00 -1.34484619e-01 8.14386010e-01
1.38017938e-01 -2.44752839e-01 1.24765299e-01 3.58454287e-01
-6.50820956e-02 2.34953463e-02 -3.37223440e-01 -3.78160954e-01
-1.37658641e-01 4.05716062e-01 -1.42479509e-01 7.46614039e-01
6.76675439e-02 4.19654921e-02 -7.97601283e-01 -8.11839342e-01
1.25926837e-01 4.26946789e-01 -5.42075276e-01 -1.61023617e-01
2.13264301e-02 1.23155467e-01 -6.90730333e-01 7.35400975e-01
6.66073263e-01 -1.38282359e-01 3.85775298e-01 -5.66155434e-01
2.29520485e-01 -1.02240600e-01 -1.43713117e+00 1.34472775e+00
-2.92144537e-01 1.91530988e-01 -3.92036200e-01 -1.27147055e+00
1.09207237e+00 2.09865957e-01 7.74722993e-01 -6.34601489e-02
-9.13326815e-03 1.86806351e-01 2.25346163e-01 -3.34766358e-01
1.70616284e-01 -1.66317493e-01 2.00847164e-01 5.22300959e-01
-4.55958508e-02 5.17271399e-01 4.34234321e-01 5.97558245e-02
9.49832976e-01 9.34608430e-02 3.74355674e-01 -1.93718120e-01
4.71063942e-01 -1.20505407e-01 1.12028599e+00 6.95715487e-01
-1.58320926e-02 4.70757902e-01 6.28243566e-01 -5.28041780e-01
-6.07146144e-01 -1.23447883e+00 -7.16240346e-01 6.61924958e-01
1.26285180e-01 -3.99727643e-01 -4.05527800e-02 -1.25633049e+00
1.78572699e-01 5.59479833e-01 -8.33414376e-01 -3.66016835e-01
-3.31740797e-01 -1.45877492e+00 2.30822161e-01 2.15919733e-01
3.06233883e-01 -6.51594996e-01 -2.79053658e-01 4.79745641e-02
-1.28116444e-01 -7.45315909e-01 -2.42467478e-01 5.08536935e-01
-9.81467128e-01 -1.27623653e+00 -3.04553092e-01 -4.80387390e-01
1.09204924e+00 1.44011304e-01 7.88489699e-01 -1.92016497e-01
-2.65906274e-01 1.91099420e-01 -4.03717071e-01 -2.70808071e-01
-1.05714373e-01 -1.90965366e-02 4.26983416e-01 5.21976054e-01
5.58229089e-01 -8.12290907e-01 -4.64036942e-01 2.59263605e-01
-9.21292782e-01 -3.07248622e-01 8.22863996e-01 1.04340637e+00
9.77596462e-01 4.24165219e-01 5.77829182e-01 -1.24339676e+00
5.05987942e-01 -7.92033672e-01 -3.60613465e-01 4.85705107e-01
-7.59855568e-01 1.14861988e-01 4.45411801e-01 -5.77326357e-01
-9.55161870e-01 6.20738380e-02 3.08518887e-01 -6.38787091e-01
-9.17500705e-02 7.44423568e-01 -2.08236709e-01 2.90578246e-01
5.65932512e-01 2.91589260e-01 -5.57115749e-02 -3.63202393e-01
-5.29499550e-04 6.94390833e-01 3.44596580e-02 -7.50630736e-01
9.41886663e-01 7.13510990e-01 2.82528531e-02 -7.31259823e-01
-6.89891696e-01 -5.23756981e-01 -6.23731732e-01 -2.47194376e-02
5.12298167e-01 -6.22526765e-01 -3.33079189e-01 1.48878902e-01
-6.32147372e-01 3.55530292e-01 -2.98184931e-01 7.69307613e-01
-2.54389018e-01 5.33337653e-01 -2.83785779e-02 -9.13278282e-01
-2.77610093e-01 -1.13879585e+00 8.09847116e-01 1.35631546e-01
1.57169804e-01 -1.04451072e+00 2.35957310e-01 1.67428523e-01
-3.05485539e-02 6.09612942e-01 1.10543990e+00 -9.80092704e-01
-2.21845552e-01 -5.58018506e-01 1.64217483e-02 3.94562393e-01
6.71523273e-01 2.99468458e-01 -8.93900931e-01 -5.79664469e-01
-9.72107798e-02 -1.55698434e-01 8.84901822e-01 3.04401845e-01
1.32792485e+00 -3.67237538e-01 -7.62681723e-01 5.81511080e-01
1.35031903e+00 2.38003522e-01 3.94969583e-01 2.49818712e-01
5.30697525e-01 5.58719814e-01 8.84954572e-01 9.26670372e-01
3.27476144e-01 4.66943204e-01 3.58639956e-01 2.96962768e-01
3.89716238e-01 -1.79933712e-01 4.47465152e-01 8.41629207e-01
2.54439294e-01 -3.42562348e-02 -7.89089143e-01 6.48551166e-01
-1.77041698e+00 -1.02194786e+00 1.30021647e-01 2.55139971e+00
1.11594629e+00 3.50109547e-01 2.74738818e-01 2.80210674e-01
9.26955700e-01 -2.41412237e-01 -8.46093476e-01 1.74641348e-02
-5.44795655e-02 -6.77165091e-02 2.98936814e-01 6.80437833e-02
-1.03384924e+00 2.06216365e-01 4.49275208e+00 1.05567920e+00
-8.57122600e-01 -1.66768372e-01 6.50593340e-01 -7.18896613e-02
-5.02536118e-01 2.56197929e-01 -1.06892180e+00 6.30956531e-01
7.54307508e-01 -2.27443576e-01 -1.27078637e-01 5.51968515e-01
1.05466008e-01 4.33100248e-03 -1.23047578e+00 9.72192883e-01
-2.07240768e-02 -1.01538169e+00 3.82707894e-01 2.20284641e-01
6.27080619e-01 -1.62010461e-01 2.49666318e-01 2.73015022e-01
3.27056795e-01 -5.53113163e-01 4.09306675e-01 7.27291763e-01
4.05801117e-01 -1.00419998e+00 8.44971120e-01 4.46996629e-01
-9.85938072e-01 -6.21217228e-02 -5.08736551e-01 6.34167612e-01
-3.13201070e-01 1.26097763e+00 -9.56370234e-01 9.34828162e-01
8.09759438e-01 9.15211499e-01 -7.03195095e-01 1.19827855e+00
3.14461440e-01 7.70422637e-01 -8.79532456e-01 -1.12421595e-01
-2.07386483e-02 -5.49156427e-01 5.75681329e-01 9.39034820e-01
4.26138908e-01 -2.41314441e-01 3.29501897e-01 6.35155678e-01
2.02602401e-01 1.78071707e-01 -5.62216163e-01 1.80258006e-01
8.09423923e-01 1.38009262e+00 -6.92609966e-01 -1.08633518e-01
-3.42501640e-01 4.22879517e-01 1.91132292e-01 2.02456981e-01
-8.54112744e-01 -5.82400501e-01 3.82168680e-01 -1.13962084e-01
4.20583606e-01 1.22285210e-01 -5.98486997e-02 -1.08683228e+00
2.63699085e-01 -7.25492537e-01 8.31636310e-01 2.41691247e-02
-1.71150935e+00 4.50599313e-01 5.05372822e-01 -1.81641102e+00
-1.21539645e-01 -4.62654352e-01 -7.12307632e-01 4.47009146e-01
-1.42500019e+00 -8.64097476e-01 -1.39659002e-01 7.16645777e-01
-4.68561761e-02 -2.24537238e-01 6.04192734e-01 2.82556117e-01
-8.17623198e-01 9.01097059e-01 3.99167061e-01 7.52890781e-02
6.83981836e-01 -1.27502370e+00 -5.21254122e-01 5.49591482e-01
4.50902760e-01 7.22241640e-01 6.16221726e-01 -3.38796705e-01
-1.27212596e+00 -1.11771083e+00 3.81755590e-01 -3.44038278e-01
5.35655379e-01 -3.09440225e-01 -7.48800397e-01 4.88457680e-01
-6.40319586e-02 2.78896242e-01 1.27029037e+00 2.18154341e-01
-3.39078009e-01 -6.24189079e-01 -1.49460411e+00 4.87289160e-01
6.22355342e-01 -1.93339258e-01 -4.35735017e-01 5.76893091e-01
3.69710386e-01 -2.47068871e-02 -1.09400606e+00 6.83000326e-01
6.02844238e-01 -1.03269017e+00 9.12236989e-01 -6.36302710e-01
-4.85638417e-02 -7.14201272e-01 -2.87854433e-01 -1.20530415e+00
-8.18913281e-02 -4.23777610e-01 -2.38355562e-01 1.30008507e+00
6.22139752e-01 -8.51081133e-01 9.22661245e-01 3.16518396e-01
-5.08253165e-02 -1.03572917e+00 -1.19917476e+00 -8.12718809e-01
-9.89973694e-02 -3.99643749e-01 6.75185144e-01 1.02447236e+00
6.57459348e-03 9.03977901e-02 -1.28415316e-01 3.30900162e-01
1.07224178e+00 2.98380345e-01 6.27775371e-01 -1.39821959e+00
-5.01025438e-01 -3.52378078e-02 -5.39105296e-01 -5.59773326e-01
8.10915008e-02 -1.04303849e+00 -2.10080653e-01 -1.22705173e+00
3.96777302e-01 -8.04555595e-01 -6.84837937e-01 5.52724779e-01
-2.70913634e-02 -4.18187529e-02 -2.97694117e-01 2.86179155e-01
-4.85811919e-01 8.59639406e-01 8.44073892e-01 -2.89943844e-01
-2.41187617e-01 3.44671696e-01 -6.26229644e-01 7.48265028e-01
8.87138188e-01 -6.90621316e-01 -6.54176295e-01 -1.23826280e-01
1.50700703e-01 -6.98830336e-02 1.50294036e-01 -8.38848710e-01
3.99342448e-01 -2.30708733e-01 4.61286634e-01 -1.05555463e+00
3.02351683e-01 -1.03079867e+00 2.45527253e-01 4.89387035e-01
-1.93177834e-01 -1.98056293e-03 2.78980546e-02 1.07927799e+00
-2.19644770e-01 -3.80444109e-01 8.55162263e-01 1.84997037e-01
-1.84952796e-01 4.99983579e-01 -5.51482029e-02 -7.20637143e-02
1.16257024e+00 -5.59924543e-01 -9.43329483e-02 8.90400633e-02
-8.61523330e-01 2.21445814e-01 -1.58080030e-02 1.52183726e-01
9.86284375e-01 -1.56709754e+00 -8.62127185e-01 3.57230037e-01
4.50991839e-01 1.36218548e-01 9.12373979e-03 1.18057668e+00
-8.28020051e-02 6.95781782e-02 2.10194588e-01 -9.03953075e-01
-1.14515364e+00 2.99000174e-01 1.74520269e-01 -2.73776501e-01
-4.52134103e-01 6.01193488e-01 1.51832357e-01 -6.31870925e-01
2.28712291e-01 -2.09532931e-01 -3.81893307e-01 2.45133713e-01
2.11639583e-01 4.09110427e-01 1.22935019e-01 -2.12661415e-01
-5.54127812e-01 3.42090249e-01 -5.94773471e-01 8.52693468e-02
1.34534943e+00 1.09007232e-01 -2.27374375e-01 6.25368237e-01
1.06239402e+00 -1.07762199e-02 -9.48763192e-01 -4.30869877e-01
1.81677341e-01 -4.91229385e-01 -4.20455225e-02 -4.71850038e-01
-1.15215206e+00 4.26501364e-01 6.85681999e-01 2.04284117e-01
1.04319561e+00 -1.27281591e-01 4.56411093e-01 5.03168166e-01
5.04618824e-01 -9.86435115e-01 -1.09085985e-01 8.12075436e-02
5.80323875e-01 -1.03643048e+00 2.60232240e-01 -3.27791035e-01
-2.91159660e-01 1.16948509e+00 6.95247054e-01 -1.39526010e-01
1.24816978e+00 3.25626694e-02 -2.62124121e-01 -7.24559426e-02
-9.59490657e-01 7.79155493e-02 3.41993362e-01 6.28285408e-01
1.80433586e-01 3.37363034e-02 -3.11095655e-01 4.89462644e-01
-2.42110845e-02 -2.76150942e-01 3.52952927e-01 6.35264814e-01
-1.93732381e-01 -1.17900491e+00 -3.98691773e-01 1.17556739e+00
-2.12731346e-01 -1.42863438e-01 -2.50649363e-01 7.97189772e-01
2.56852984e-01 8.67073417e-01 5.57262078e-03 -4.96323079e-01
2.78674483e-01 -4.74087149e-02 4.99482274e-01 -6.83750749e-01
-2.50921249e-01 -4.42936234e-02 -4.95727882e-02 9.45157185e-03
-4.69428778e-01 -1.39549470e+00 -1.07682383e+00 -6.08859248e-02
-8.99369836e-01 3.38936061e-01 3.61104965e-01 9.02946115e-01
2.16553703e-01 9.45869163e-02 1.11344564e+00 -3.19858372e-01
-9.50513840e-01 -9.24850643e-01 -7.23333955e-01 1.14436619e-01
1.19849093e-01 -8.88250172e-01 -7.55607367e-01 -2.06296325e-01] | [9.36251449584961, 3.44171404838562] |
11e98fa6-f83d-426f-8798-4eb04500de3e | husp-sp-faster-utility-mining-on-sequence | 2212.14255 | null | https://arxiv.org/abs/2212.14255v1 | https://arxiv.org/pdf/2212.14255v1.pdf | HUSP-SP: Faster Utility Mining on Sequence Data | High-utility sequential pattern mining (HUSPM) has emerged as an important topic due to its wide application and considerable popularity. However, due to the combinatorial explosion of the search space when the HUSPM problem encounters a low utility threshold or large-scale data, it may be time-consuming and memory-costly to address the HUSPM problem. Several algorithms have been proposed for addressing this problem, but they still cost a lot in terms of running time and memory usage. In this paper, to further solve this problem efficiently, we design a compact structure called sequence projection (seqPro) and propose an efficient algorithm, namely discovering high-utility sequential patterns with the seqPro structure (HUSP-SP). HUSP-SP utilizes the compact seq-array to store the necessary information in a sequence database. The seqPro structure is designed to efficiently calculate candidate patterns' utilities and upper bound values. Furthermore, a new upper bound on utility, namely tighter reduced sequence utility (TRSU) and two pruning strategies in search space, are utilized to improve the mining performance of HUSP-SP. Experimental results on both synthetic and real-life datasets show that HUSP-SP can significantly outperform the state-of-the-art algorithms in terms of running time, memory usage, search space pruning efficiency, and scalability. | ['Philip S. Yu', 'Wensheng Gan', 'Zilin Du', 'Yuting Yang', 'Chunkai Zhang'] | 2022-12-29 | null | null | null | null | ['sequential-pattern-mining'] | ['natural-language-processing'] | [ 4.11952466e-01 -3.24540764e-01 -4.08149362e-01 -3.45874717e-03
-2.45360553e-01 -3.19074132e-02 -1.44111484e-01 2.27366641e-01
-4.10470217e-01 9.25754011e-01 -5.67185357e-02 -2.81022727e-01
-3.59643757e-01 -1.20439756e+00 -1.27421424e-01 -7.79536128e-01
-1.29359350e-01 2.94036210e-01 9.11070883e-01 9.75067690e-02
4.21843648e-01 1.66960865e-01 -1.66130066e+00 2.16932997e-01
1.06127906e+00 1.16842461e+00 6.78697705e-01 -1.39136732e-01
-3.98065299e-01 4.07061577e-01 -4.21108454e-01 -1.42835096e-01
5.43124557e-01 -5.87680817e-01 -7.60240138e-01 9.33490042e-03
-9.86337066e-01 1.24425841e-02 -9.92304180e-03 1.01415706e+00
3.01229537e-01 2.10684955e-01 5.81218116e-02 -1.28447628e+00
-1.01130180e-01 6.35345221e-01 -1.13646340e+00 1.16902098e-01
5.30562401e-01 -1.73680261e-01 1.04577208e+00 -6.02879643e-01
6.14813089e-01 9.61583972e-01 6.26359105e-01 3.91612053e-02
-1.04790294e+00 -8.24928105e-01 1.68782175e-01 8.07886422e-01
-1.61398959e+00 1.68734372e-01 4.71349984e-01 -3.24603505e-02
1.06074989e+00 6.34605289e-01 8.48782241e-01 3.95402551e-01
-1.15410596e-01 9.32702005e-01 9.19495404e-01 -4.88570422e-01
4.24125463e-01 -7.01036826e-02 2.16305748e-01 4.17264104e-01
5.02320051e-01 2.08822489e-02 -4.95925158e-01 -5.33377886e-01
4.42454964e-01 3.34794253e-01 -3.04224521e-01 -2.76328623e-01
-8.66733789e-01 7.98566341e-01 -2.70596951e-01 2.55122602e-01
-4.82487142e-01 -6.48554862e-01 5.28400719e-01 3.93751740e-01
1.88464988e-02 2.68701345e-01 -4.53605860e-01 -5.83170235e-01
-6.61109865e-01 3.01375508e-01 8.15822303e-01 9.55520451e-01
7.26091564e-01 -3.24780703e-01 6.88487291e-02 9.67217147e-01
-1.33029997e-01 6.97385669e-02 6.99654460e-01 -4.77037042e-01
3.88684303e-01 1.25400102e+00 3.13109979e-02 -1.01318634e+00
-3.93312961e-01 -8.99355188e-02 -9.19202626e-01 -4.30860072e-01
4.65904996e-02 5.51288091e-02 -4.01142776e-01 1.40938175e+00
6.16039991e-01 1.18553363e-01 -9.74668190e-02 8.57087791e-01
2.47410551e-01 1.09230769e+00 -1.59830466e-01 -1.10499537e+00
1.30787385e+00 -7.50049949e-01 -4.20163810e-01 4.00105380e-02
6.23559892e-01 -5.01052082e-01 1.07376504e+00 5.05993009e-01
-8.54965389e-01 -2.44025752e-01 -1.08158422e+00 5.02347171e-01
2.84252018e-02 -1.49902195e-01 4.82450902e-01 5.94483554e-01
-3.18277538e-01 5.55500865e-01 -7.19682813e-01 -3.77666771e-01
3.62690508e-01 3.13936055e-01 -1.00820646e-01 -2.10859612e-01
-1.08698380e+00 5.52395642e-01 9.41031814e-01 -4.23577249e-01
-1.99651361e-01 -6.20048702e-01 -6.39926851e-01 2.95809984e-01
9.02523696e-01 -2.25759551e-01 1.08952427e+00 -2.32098728e-01
-1.18725026e+00 3.09481323e-01 -2.47409403e-01 -5.61872602e-01
2.84101158e-01 1.24740727e-01 -4.31094766e-01 7.37230331e-02
9.34205726e-02 -1.17640607e-01 2.60169387e-01 -6.32369280e-01
-1.12371826e+00 -2.90701717e-01 -4.55942869e-01 1.77346528e-01
-6.96715593e-01 1.82110533e-01 -4.92783576e-01 -5.84497571e-01
2.32124627e-01 -7.34766483e-01 -6.55549586e-01 -5.63441575e-01
-2.11544260e-01 -4.06838804e-01 9.33015525e-01 -5.33315957e-01
2.07322216e+00 -2.09853578e+00 1.99822299e-02 7.06586480e-01
-1.85773477e-01 4.97208685e-01 1.15251407e-01 7.49151766e-01
2.50506669e-01 -1.01846091e-01 -5.24446547e-01 5.56571186e-01
-3.41651440e-01 6.40309274e-01 -1.44676134e-01 1.06935509e-01
-2.86578566e-01 4.65044290e-01 -9.59256411e-01 -5.74778795e-01
1.82619020e-02 -3.14508051e-01 -5.82392216e-01 2.68558472e-01
-2.12039709e-01 -9.94345993e-02 -5.88397980e-01 6.70066237e-01
8.47089350e-01 -2.93198556e-01 7.55326033e-01 3.26778919e-01
-3.75417650e-01 1.11482382e-01 -1.52960420e+00 1.04326773e+00
-8.04652795e-02 -1.82486430e-01 -2.69300073e-01 -1.33652890e+00
1.18560970e+00 2.03475840e-02 7.47467816e-01 -8.97756636e-01
-4.42433618e-02 2.99873739e-01 -2.52002459e-02 -5.04105031e-01
3.86285514e-01 -1.85333058e-01 -8.98285210e-02 7.12289214e-01
-6.09977722e-01 6.12112522e-01 7.77401149e-01 -1.49468973e-01
1.53623068e+00 -2.14813665e-01 9.91166115e-01 -3.23547542e-01
9.32772398e-01 1.29343003e-01 1.47201002e+00 3.18548441e-01
1.58420727e-01 1.15671560e-01 6.15222394e-01 -5.21777511e-01
-9.01896834e-01 -6.17734969e-01 5.78489490e-02 8.34010780e-01
4.57484901e-01 -7.40094244e-01 -4.69074309e-01 -4.80441719e-01
1.00230172e-01 5.95532715e-01 -2.10171565e-01 -1.06579252e-01
-7.31668830e-01 -1.13252604e+00 1.08996272e-01 5.07462502e-01
5.59786022e-01 -1.18884718e+00 -1.11061442e+00 5.58924437e-01
-1.69174224e-01 -9.78309214e-01 -4.40490246e-01 1.25370935e-01
-8.05219531e-01 -1.28189528e+00 -2.42254481e-01 -7.53600180e-01
5.78833759e-01 6.10675871e-01 5.89773536e-01 2.23945677e-01
-4.05343503e-01 -3.90466630e-01 -9.15774345e-01 -4.40328985e-01
-1.74265429e-01 9.04049054e-02 -1.25882775e-02 -1.38793346e-02
7.28800714e-01 -8.35126102e-01 -5.12483537e-01 8.33159447e-01
-9.63097274e-01 1.69899419e-01 9.01180983e-01 9.87528026e-01
8.29549432e-01 7.09634781e-01 7.25079238e-01 -8.28384519e-01
8.78258169e-01 -5.50128043e-01 -7.41441071e-01 3.94931346e-01
-1.08270109e+00 -3.85351628e-02 8.14630032e-01 -3.82686466e-01
-1.10338759e+00 7.72862062e-02 -1.01915836e-01 -2.07791969e-01
3.23997378e-01 8.82602215e-01 -3.91521543e-01 3.61668676e-01
3.38498801e-01 8.47300351e-01 2.39391506e-01 -6.79509401e-01
-3.01099628e-01 6.62378430e-01 2.67520875e-01 -4.84104395e-01
6.36042356e-01 2.43362829e-01 2.46942580e-01 -8.58338654e-01
-5.28362393e-01 -8.90232325e-01 -2.45427504e-01 3.34581047e-01
1.88663334e-01 -3.24459434e-01 -9.14376795e-01 2.23320946e-01
-7.44724095e-01 2.11765185e-01 -2.03594700e-01 1.65169567e-01
-3.66747350e-01 9.23767030e-01 -3.67133319e-01 -1.03223979e+00
-8.29207718e-01 -8.79765272e-01 2.11102292e-01 1.56620204e-01
-3.53433967e-01 -1.98317721e-01 7.35420734e-02 -5.81883937e-02
1.28650684e-02 3.24750721e-01 1.11136746e+00 -6.61612868e-01
-5.60447097e-01 -1.31671458e-01 -2.73824811e-01 7.56293461e-02
7.60892183e-02 -4.71574098e-01 -2.66327057e-02 -1.68961868e-01
1.43579906e-02 -2.23684311e-02 4.77100492e-01 4.07735184e-02
1.53137958e+00 -5.49696445e-01 -6.75806105e-01 4.19545144e-01
1.28527129e+00 8.40748489e-01 5.85337460e-01 7.28120506e-01
-3.14399623e-03 5.15793443e-01 1.46942294e+00 1.09786093e+00
3.54874246e-02 7.34600544e-01 1.66909218e-01 3.68210465e-01
5.63789964e-01 -4.17736501e-01 2.30938658e-01 8.47675323e-01
-1.19701974e-01 2.55323127e-02 -6.02584541e-01 5.46703219e-01
-2.30757236e+00 -1.21387827e+00 -1.00781418e-01 2.25014257e+00
1.03107536e+00 1.98293209e-01 5.75709283e-01 8.94916058e-01
5.41052878e-01 -1.74021184e-01 -5.31492531e-01 -5.91441989e-01
-7.60314763e-02 3.00759654e-02 3.42016935e-01 2.57899761e-02
-5.99429488e-01 4.59440768e-01 5.35159254e+00 1.23702884e+00
-6.42315090e-01 -9.12301987e-02 2.89579809e-01 -1.33825973e-01
-1.21631950e-01 8.29076171e-02 -9.03340876e-01 8.59735429e-01
6.79844022e-01 -6.85849905e-01 2.40336731e-01 1.27048576e+00
1.48261830e-01 -3.49401057e-01 -7.31603086e-01 1.16605127e+00
-3.14477861e-01 -1.38145411e+00 -1.22435419e-02 2.23737508e-01
4.95613933e-01 -4.71061230e-01 -4.74225461e-01 1.54192716e-01
1.03116699e-01 -5.38317859e-01 2.56048381e-01 -1.49485186e-01
6.97721064e-01 -1.10228455e+00 8.37854147e-01 8.22493732e-01
-1.54081905e+00 -4.72841531e-01 -7.14972794e-01 -3.62543315e-01
4.20557827e-01 9.56471741e-01 -9.07764912e-01 8.18457186e-01
9.79643762e-01 5.16722500e-01 7.02142343e-02 1.20325255e+00
-1.00602753e-01 4.67910409e-01 -6.03750527e-01 -5.01586318e-01
1.82259232e-01 -6.09225512e-01 4.01318610e-01 1.02944958e+00
5.99515676e-01 5.44484019e-01 5.18894494e-01 4.01195258e-01
2.54473835e-01 4.51480269e-01 -1.72980919e-01 -1.96099356e-02
8.90231729e-01 1.01663578e+00 -6.72583282e-01 -1.80490360e-01
-5.37816405e-01 4.73743796e-01 1.13515064e-01 -2.19109640e-01
-6.79852247e-01 -6.35776103e-01 6.31786883e-01 4.20273900e-01
5.25907934e-01 2.89036543e-04 -4.76289779e-01 -7.13250935e-01
3.89539897e-01 -9.32802916e-01 9.74375188e-01 5.41523471e-03
-9.70624924e-01 3.88943285e-01 4.99118045e-02 -1.32272995e+00
-1.03112042e-01 -1.93064630e-01 -6.03006184e-01 6.76552892e-01
-1.13123560e+00 -5.60293794e-01 -1.93721876e-01 4.59780484e-01
6.22520089e-01 -1.78039148e-01 6.36442721e-01 1.80291370e-01
-7.41911530e-01 6.84472919e-01 1.00328438e-01 -4.32789743e-01
-5.95913716e-02 -6.30251288e-01 1.32295251e-01 8.76826167e-01
-2.82912910e-01 6.28378272e-01 6.18558884e-01 -7.59403408e-01
-1.44575524e+00 -8.42254043e-01 7.57340074e-01 3.22477311e-01
4.50294495e-01 -8.81389081e-02 -1.07343304e+00 1.98087767e-01
-1.90790609e-01 -4.07175720e-01 1.05543840e+00 1.16112053e-01
-4.85935621e-03 -2.03552350e-01 -1.18262088e+00 6.71029210e-01
1.39302504e+00 2.96075016e-01 -6.47573888e-01 -5.56669645e-02
6.63283944e-01 -1.31693125e-01 -9.04525042e-01 7.17566133e-01
6.28135800e-01 -1.06652582e+00 8.37339938e-01 -7.94307515e-02
1.79679126e-01 -4.77616131e-01 4.39145863e-02 -9.03548360e-01
-5.41554451e-01 -7.57236719e-01 -3.26440215e-01 9.37101901e-01
2.70854592e-01 -6.48169696e-01 1.00876546e+00 3.00865948e-01
4.95328680e-02 -1.56006932e+00 -9.61015999e-01 -1.08893490e+00
-6.69515431e-01 -2.88313836e-01 1.07512236e+00 6.96890473e-01
7.86265969e-01 3.10214490e-01 -6.23623073e-01 -2.80749768e-01
3.87194276e-01 8.17540228e-01 7.89134264e-01 -1.25380218e+00
-4.97285068e-01 -3.80874068e-01 -2.92035937e-01 -1.07573020e+00
-4.29338902e-01 -6.97582662e-01 -2.60212630e-01 -1.44409049e+00
4.82753843e-01 -5.15692115e-01 -4.05215085e-01 5.78993022e-01
-2.78714061e-01 -2.87616640e-01 -7.55911926e-03 3.88184845e-01
-6.43990934e-01 5.80506504e-01 1.01304626e+00 2.90421098e-01
-6.13848627e-01 2.09220082e-01 -6.25059545e-01 5.63916028e-01
7.88429976e-01 -4.27664012e-01 -5.97563446e-01 2.50854254e-01
1.83531851e-01 2.82719821e-01 -5.75490415e-01 -6.90270245e-01
1.15448609e-01 -7.11111903e-01 1.77450433e-01 -1.18611741e+00
3.08783036e-02 -7.93208957e-01 5.53541303e-01 1.11236322e+00
1.41776890e-01 6.45005703e-02 1.28009804e-02 6.10120654e-01
-1.81989208e-01 -3.46335322e-01 6.94556773e-01 -5.30457869e-02
-9.35368776e-01 3.37721020e-01 -2.86847234e-01 -3.32747281e-01
1.54256356e+00 -6.73687637e-01 3.17977704e-02 1.93918064e-01
-2.36996457e-01 6.21983647e-01 4.41722691e-01 2.21838653e-01
9.90844369e-01 -1.24156952e+00 -3.51521283e-01 3.43767792e-01
2.42859975e-01 1.02902554e-01 2.74517447e-01 8.60495389e-01
-4.60273206e-01 4.37319547e-01 -3.13342094e-01 -3.57758611e-01
-1.61716914e+00 9.23521936e-01 -5.56835294e-01 -7.81377196e-01
-9.83389199e-01 5.84702313e-01 2.90485676e-02 -1.30409703e-01
3.95636968e-02 3.76071930e-02 -7.31836632e-02 -6.96421117e-02
8.61911118e-01 8.31829846e-01 -5.98403141e-02 6.13713227e-02
-3.70092273e-01 2.53071517e-01 -2.47138828e-01 3.36293727e-01
1.42710316e+00 -3.87633592e-02 -4.95333731e-01 -1.20437905e-01
7.64401078e-01 -1.62520900e-01 -7.54597902e-01 -5.33268511e-01
6.10201180e-01 -8.66757989e-01 -5.26848376e-01 -4.54023033e-01
-7.31335700e-01 2.74662942e-01 2.28424426e-02 2.13624850e-01
1.55732918e+00 -2.83388883e-01 1.46232390e+00 4.60332334e-01
7.53730297e-01 -1.38574493e+00 -1.36342734e-01 3.86051834e-01
5.73839545e-01 -6.62904322e-01 3.17567915e-01 -8.24150503e-01
-6.96174562e-01 7.72940397e-01 7.96342611e-01 3.13352704e-01
4.92775232e-01 3.77298653e-01 -5.36457181e-01 3.10371995e-01
-8.35116506e-01 -6.18279763e-02 -2.48136118e-01 4.38024849e-01
-3.32702585e-02 1.59071013e-01 -1.31013584e+00 1.04366648e+00
-2.99330086e-01 9.71101895e-02 3.07481378e-01 1.39906037e+00
-8.94630253e-01 -1.57814348e+00 -4.10880655e-01 8.40035498e-01
-3.29992563e-01 2.30700433e-01 -1.98092759e-01 4.18884158e-01
8.26228261e-02 8.29773724e-01 -2.24485934e-01 -7.26121783e-01
4.01898265e-01 -7.12667629e-02 2.59277195e-01 -5.06111383e-01
-1.83991298e-01 1.39333203e-01 2.00441256e-01 -5.85001945e-01
-2.59503070e-02 -7.29902983e-01 -1.56523848e+00 -3.85749757e-01
-3.41654778e-01 6.36963785e-01 7.99944624e-02 7.61309505e-01
4.30117041e-01 1.21577233e-01 7.36978590e-01 3.79758440e-02
-8.70533109e-01 -7.82158732e-01 -8.33207786e-01 3.31143528e-01
-5.33918381e-01 -7.12662458e-01 3.67847770e-01 -4.50558156e-01] | [8.299981117248535, 6.30908203125] |
b2c945aa-bd50-481c-9a9b-c62891e87f77 | rethinking-generalization-performance-of | 2110.11626 | null | https://arxiv.org/abs/2110.11626v1 | https://arxiv.org/pdf/2110.11626v1.pdf | Rethinking Generalization Performance of Surgical Phase Recognition with Expert-Generated Annotations | As the area of application of deep neural networks expands to areas requiring expertise, e.g., in medicine and law, more exquisite annotation processes for expert knowledge training are required. In particular, it is difficult to guarantee generalization performance in the clinical field in the case of expert knowledge training where opinions may differ even among experts on annotations. To raise the issue of the annotation generation process for expertise training of CNNs, we verified the annotations for surgical phase recognition of laparoscopic cholecystectomy and subtotal gastrectomy for gastric cancer. We produce calibrated annotations for the seven phases of cholecystectomy by analyzing the discrepancies of previously annotated labels and by discussing the criteria of surgical phases. For gastrectomy for gastric cancer has more complex twenty-one surgical phases, we generate consensus annotation by the revision process with five specialists. By training the CNN-based surgical phase recognition networks with revised annotations, we achieved improved generalization performance over models trained with original annotation under the same cross-validation settings. We showed that the expertise data annotation pipeline for deep neural networks should be more rigorous based on the type of problem to apply clinical field. | ['Min-Kook Choi', 'Woo Jin Hyung', 'Sunghyun Park', 'Anwar H. Alfadhel', 'Ahmed A. Alwusaibie', 'Bokyung Park', 'Jiwon Lee', 'Seungbum Hong'] | 2021-10-22 | null | null | null | null | ['surgical-phase-recognition'] | ['computer-vision'] | [-2.80863456e-02 1.01175416e+00 -3.18215400e-01 -3.83529842e-01
-6.07490003e-01 -8.86323392e-01 -1.71267465e-01 3.51542711e-01
-6.96747899e-01 7.04303265e-01 1.86010256e-01 -8.36493552e-01
-3.20726484e-01 -4.95575398e-01 -6.31323576e-01 -5.46250701e-01
1.24664884e-02 6.77621901e-01 -1.88675076e-01 -9.94619876e-02
-1.50327474e-01 2.42772698e-01 -9.71495569e-01 5.38185477e-01
9.01626289e-01 1.05583143e+00 -1.83715239e-01 3.57542515e-01
1.63850188e-01 5.71677506e-01 -6.13156140e-01 -5.03379285e-01
6.51222646e-01 -2.89836049e-01 -9.86309648e-01 -1.85614049e-01
2.41949826e-01 -7.76743889e-02 -3.70740891e-02 1.07392395e+00
5.66904366e-01 -3.80537719e-01 5.11005104e-01 -9.82520521e-01
-8.20300937e-01 1.11386979e+00 -2.42225230e-02 1.19285800e-01
4.87405062e-02 1.68934092e-01 7.01192975e-01 -5.73862731e-01
6.61839366e-01 5.36768734e-01 1.26575065e+00 8.46917391e-01
-8.10138404e-01 -6.70720994e-01 2.34521423e-02 -1.09586619e-01
-1.29901743e+00 1.44038856e-01 7.16741383e-02 -8.13710988e-01
5.53010046e-01 -8.55548382e-02 1.02418602e+00 1.18856061e+00
3.01419169e-01 4.20372784e-01 6.85039282e-01 -2.46046230e-01
1.57034203e-01 4.11385357e-01 -1.18377440e-01 9.08904612e-01
9.71114516e-01 1.16143197e-01 -9.17915925e-02 -1.29768839e-02
8.77423763e-01 3.43904039e-03 -5.69927454e-01 -2.99954981e-01
-1.73090518e+00 8.46516848e-01 4.87428069e-01 5.70306480e-01
-1.92456663e-01 -7.22716600e-02 8.54810774e-01 3.54072958e-01
1.20477140e-01 1.17369485e+00 -1.00501072e+00 2.79736072e-01
-9.50824797e-01 -4.00035948e-01 1.32572138e+00 1.03420222e+00
4.71163183e-01 -1.79426506e-01 -1.28108516e-01 3.73943388e-01
1.79707140e-01 -1.03926212e-01 7.45958388e-01 -1.08152759e+00
2.83588096e-02 8.25570881e-01 1.56203568e-01 -6.16504490e-01
-7.13447690e-01 -8.60794127e-01 -1.14669025e+00 1.54902846e-01
6.59157395e-01 -4.93339449e-01 -1.07747519e+00 1.35087895e+00
-1.22786932e-01 -1.75056428e-01 3.41399848e-01 9.60127592e-01
1.22615433e+00 -3.36517185e-01 4.38241303e-01 -4.60824631e-02
1.56199419e+00 -9.49727952e-01 -7.41648555e-01 -1.20345280e-01
1.29818070e+00 -4.16755170e-01 7.27005303e-01 4.49920982e-01
-8.12774122e-01 -5.27303219e-01 -1.07313037e+00 1.52249798e-01
-3.93214524e-01 4.86844331e-01 8.70449364e-01 5.69440842e-01
-1.15096486e+00 5.07980168e-01 -8.63812804e-01 -4.81565684e-01
3.76756459e-01 6.42082453e-01 -5.78148425e-01 4.46001619e-01
-1.60587406e+00 1.28469121e+00 9.24419224e-01 4.76490200e-01
-1.03416431e+00 -1.06201577e+00 -9.28596914e-01 1.67997405e-01
2.00031400e-01 -1.12387586e+00 1.39961588e+00 -1.53266799e+00
-1.08327127e+00 1.47903228e+00 6.79153800e-01 -5.17422795e-01
5.40202796e-01 7.60030523e-02 -2.98096120e-01 -2.84684096e-02
8.30798224e-02 7.36356020e-01 1.08472779e-01 -1.08949888e+00
-7.47691870e-01 -2.12896630e-01 3.14653873e-01 8.66555199e-02
-1.93814501e-01 -4.88880336e-01 -1.98150888e-01 -3.28708947e-01
-7.42842108e-02 -1.09131384e+00 -6.04843676e-01 2.40599290e-01
-7.48957545e-02 -1.31377906e-01 2.49939620e-01 -5.52502096e-01
1.09117472e+00 -2.26940894e+00 -1.51894316e-01 3.47966731e-01
4.40406561e-01 2.86529303e-01 2.56759435e-01 -4.07053642e-02
-3.73447865e-01 5.33220053e-01 1.43313482e-01 -2.11640894e-02
-7.59640485e-02 4.03503448e-01 7.80951930e-03 4.17245626e-01
-3.03157675e-03 8.49545479e-01 -1.19248092e+00 -6.61410332e-01
-1.81130350e-01 2.29835257e-01 -7.22480118e-01 5.58095686e-02
-4.84626507e-03 7.72691131e-01 -3.58197540e-01 7.74010003e-01
5.45222349e-02 -8.76554370e-01 3.49280179e-01 -5.77058375e-01
2.37490326e-01 -1.13478109e-01 -7.65579700e-01 2.30166507e+00
-5.71672261e-01 5.61671972e-01 1.95533603e-01 -1.01058137e+00
6.55271947e-01 1.01425385e+00 7.44059682e-01 -6.48772269e-02
3.46222460e-01 6.15375340e-01 5.17060757e-01 -7.70128846e-01
2.66615808e-01 -1.01849966e-01 -1.85010776e-01 2.05939054e-01
3.90473068e-01 -1.18292302e-01 7.10158423e-02 -6.59172535e-02
9.13909316e-01 2.81343330e-02 6.77231312e-01 -6.70902431e-01
4.01897818e-01 5.97141683e-01 7.32872605e-01 7.64420092e-01
-7.15266109e-01 5.01862824e-01 4.68027353e-01 -9.92158771e-01
-1.12517309e+00 -6.56188011e-01 -3.74812216e-01 7.54009843e-01
-6.75456971e-02 -1.00862481e-01 -7.43937135e-01 -1.03100395e+00
-9.42986757e-02 3.00004631e-01 -1.26756120e+00 -2.00112522e-01
-3.60915035e-01 -5.76797545e-01 8.30237091e-01 6.06983125e-01
2.72532284e-01 -9.26581502e-01 -7.16051459e-01 2.09034279e-01
-1.43109664e-01 -1.00988972e+00 -2.22383603e-01 5.33273339e-01
-8.13408971e-01 -1.55259097e+00 -8.35328162e-01 -1.17903113e+00
1.09466612e+00 -3.95239651e-01 1.28846478e+00 4.38746750e-01
-3.56733531e-01 7.05050051e-01 -3.76223445e-01 -7.70330310e-01
-6.66114390e-01 5.76948106e-01 -9.22971293e-02 -5.75684786e-01
4.91092175e-01 -1.60010412e-01 -1.15047979e+00 4.19840038e-01
-9.88064349e-01 3.66383791e-02 9.71534073e-01 1.12407982e+00
3.10174376e-01 -2.75738269e-01 5.95147669e-01 -1.07905662e+00
5.98065555e-01 -6.08196318e-01 -2.84381926e-01 4.16428536e-01
-8.68669569e-01 5.14918705e-03 3.09551567e-01 -5.41955590e-01
-8.53023052e-01 1.94600910e-01 7.18547180e-02 -3.72876287e-01
-2.77219981e-01 8.51870418e-01 4.66789424e-01 -2.77411379e-02
1.33726144e+00 -4.78298694e-01 2.93658189e-02 1.84554979e-01
2.78137196e-02 5.19694269e-01 5.28313458e-01 -6.30643010e-01
3.32730561e-01 2.62694538e-01 -2.33738706e-01 9.01672915e-02
-1.22449255e+00 -4.38322574e-01 -6.17250741e-01 -4.19303998e-02
1.28492415e+00 -1.14177239e+00 -8.05148959e-01 -1.31558225e-01
-1.14295661e+00 -5.89540124e-01 -3.98281068e-01 8.49004209e-01
-4.13382858e-01 2.80875564e-01 -5.88833809e-01 -1.13978945e-01
-6.71721578e-01 -1.42615974e+00 9.23081219e-01 3.04235518e-01
-6.12623572e-01 -1.56565261e+00 9.00268033e-02 1.42013028e-01
4.37581092e-01 3.65768373e-01 8.75597060e-01 -1.29310119e+00
-2.51470715e-01 -3.10500622e-01 5.33433408e-02 2.02913836e-01
2.79863805e-01 -3.31599228e-02 -8.36142898e-01 -2.12336719e-01
-9.25738066e-02 -3.21408063e-01 3.78356636e-01 2.04771101e-01
1.39056742e+00 -2.20062017e-01 -5.41875362e-01 7.19662488e-01
1.24370670e+00 2.58708507e-01 4.29952532e-01 7.17049479e-01
3.57034355e-01 5.65536857e-01 4.22389418e-01 1.44247919e-01
6.30448237e-02 8.43647048e-02 5.45953393e-01 -4.50685024e-01
1.61054254e-01 1.20783493e-01 -2.03503922e-01 4.66325760e-01
-2.84879714e-01 -1.43671289e-01 -1.43635702e+00 7.86208928e-01
-1.69805551e+00 -6.25665247e-01 5.37298992e-02 1.97541380e+00
1.15736365e+00 -6.88465387e-02 -4.83360261e-01 -1.79918930e-01
5.95478892e-01 -5.57139814e-01 -3.40455770e-01 -1.36343151e-01
2.35706687e-01 1.27540797e-01 7.29158521e-01 1.73423395e-01
-9.06634688e-01 5.77281892e-01 6.83162737e+00 2.72238135e-01
-1.16667867e+00 1.86218590e-01 7.29225755e-01 3.06802481e-01
-1.24835126e-01 -1.15242682e-01 -5.98535359e-01 1.98274657e-01
7.18136907e-01 -1.28502399e-02 -1.45894065e-01 1.07143152e+00
-9.75409374e-02 1.48184858e-02 -1.58761013e+00 6.85514569e-01
-5.96471094e-02 -1.28530049e+00 -2.50478148e-01 -1.72629416e-01
1.02423656e+00 -1.53160682e-02 -2.86811255e-02 4.19248044e-01
5.19344509e-01 -1.24915564e+00 1.52415633e-01 4.57400650e-01
9.49553668e-01 2.61533633e-02 1.44325614e+00 1.51982576e-01
-6.22348845e-01 -1.26710877e-01 -5.92325069e-02 3.43040138e-01
-1.50762245e-01 4.12856400e-01 -1.42025232e+00 5.81972003e-01
8.34676683e-01 4.49477702e-01 -2.76898295e-01 1.15291357e+00
-2.86613196e-01 3.97428751e-01 -4.21549268e-02 2.57444173e-01
2.57192492e-01 3.00603479e-01 -6.08122088e-02 1.24949813e+00
5.49620271e-01 2.75563560e-02 -6.59286045e-03 7.42541254e-01
-3.46148908e-02 -3.47342528e-02 -5.57288766e-01 -6.19791672e-02
2.20031932e-01 1.62711787e+00 -8.36601853e-01 -3.58887792e-01
-3.90363365e-01 6.79911673e-01 -2.53240224e-02 3.69676113e-01
-7.27550566e-01 -1.73600495e-01 1.37848422e-01 1.31893456e-01
-4.54233922e-02 2.10685223e-01 -5.81245184e-01 -1.02298844e+00
-4.08187866e-01 -8.75357389e-01 7.83691704e-01 -7.91359246e-01
-1.32819152e+00 7.41717696e-01 -7.29325116e-02 -1.47759426e+00
-1.78496078e-01 -1.15858424e+00 -5.20440698e-01 7.47619331e-01
-1.20448637e+00 -1.19514978e+00 -8.40181231e-01 3.71181339e-01
2.17208236e-01 -1.44660503e-01 1.18927455e+00 3.95080119e-01
-1.40692651e-01 7.11053193e-01 -4.01904434e-01 6.74241483e-01
1.24113262e+00 -1.18039501e+00 -3.53092074e-01 1.20100610e-01
-5.03792286e-01 9.26073134e-01 5.92896342e-01 -6.43265128e-01
-7.57742584e-01 -1.07401133e+00 5.95104814e-01 -5.06848454e-01
8.10142756e-01 2.27804720e-01 -7.28287935e-01 1.22405291e+00
4.61032748e-01 -3.75094675e-02 1.35880649e+00 1.22886784e-01
-1.70243546e-01 1.01457670e-01 -1.06881428e+00 4.18558896e-01
9.42047238e-01 -2.88309455e-01 -7.14501321e-01 5.64109623e-01
4.94515568e-01 -9.35803652e-01 -1.60175359e+00 8.47478628e-01
5.59700787e-01 -4.42166120e-01 6.23431981e-01 -7.67740965e-01
4.27948326e-01 -3.26450378e-01 3.81449193e-01 -1.37348950e+00
-1.55626714e-01 -2.49342829e-01 4.01678383e-01 3.72368872e-01
1.02215672e+00 -5.62618494e-01 8.92690480e-01 8.48166704e-01
-6.59447372e-01 -8.65785360e-01 -6.65955245e-01 -3.34974647e-01
2.85021752e-01 1.19364239e-01 2.78724611e-01 1.48186326e+00
4.26614732e-01 -8.90691355e-02 1.93578064e-01 2.15138286e-01
5.88258393e-02 -2.85017252e-01 5.93288600e-01 -1.50146878e+00
-2.39198402e-01 -5.67928851e-01 -4.16381419e-01 -4.47417587e-01
1.23473205e-01 -1.14090359e+00 1.36278182e-01 -1.88084328e+00
1.43875211e-01 -4.79604304e-01 -4.76652265e-01 9.16621983e-01
-1.78018153e-01 2.15345949e-01 -2.72058129e-01 4.00065660e-01
-6.71685517e-01 -3.20800573e-01 1.44455981e+00 -1.42473966e-01
-2.69234069e-02 -8.11048225e-02 -1.01563060e+00 9.82978165e-01
8.33915770e-01 -4.47503716e-01 -3.06151867e-01 -5.79692781e-01
1.00040913e+00 -8.94401371e-02 2.06384018e-01 -9.14703667e-01
5.81305623e-01 1.57833591e-01 5.19788265e-01 -1.61815152e-01
-3.24938208e-01 -1.18086445e+00 5.31177938e-01 9.50552821e-01
-4.82457578e-01 -8.73842612e-02 4.33675885e-01 2.08434090e-01
-3.65704715e-01 -4.08278793e-01 4.92436707e-01 -7.67413795e-01
-6.02406263e-01 3.62513155e-01 -3.61218065e-01 4.14451472e-02
1.10583889e+00 -4.81139600e-01 -4.20191139e-01 -2.13847682e-01
-1.50194907e+00 4.51349795e-01 4.74193424e-01 2.07548752e-01
1.83148548e-01 -1.12510145e+00 -6.97306275e-01 -9.50048864e-02
3.96584272e-01 5.81842661e-01 3.28254372e-01 1.41045737e+00
-9.40415800e-01 5.79174578e-01 -2.96571463e-01 -7.05550671e-01
-6.35349154e-01 6.04730785e-01 1.00731611e+00 -5.46626687e-01
-2.55124301e-01 5.63992620e-01 3.97889197e-01 -6.60993099e-01
5.44322431e-01 -6.54546559e-01 -3.24257612e-01 1.27372751e-02
7.59073645e-02 -9.59340632e-02 2.35800773e-01 1.16077475e-01
-1.62409589e-01 2.62176573e-01 -1.59244418e-01 3.65659952e-01
9.33164060e-01 3.04136753e-01 4.39656042e-02 2.96658158e-01
7.09138632e-01 -2.85959244e-01 -8.72184634e-01 6.10705279e-02
1.93234496e-02 2.93114692e-01 -2.42251128e-01 -1.26615822e+00
-1.20819473e+00 5.87432027e-01 6.97408199e-01 4.96057831e-02
9.12547231e-01 8.67180619e-03 2.11494654e-01 5.62971294e-01
4.74953443e-01 -1.14961004e+00 -4.35762443e-02 5.09033561e-01
7.86439896e-01 -1.49477971e+00 -1.12295002e-01 -2.22779855e-01
-5.95733583e-01 1.41946864e+00 9.14794922e-01 1.79427207e-01
7.51958072e-01 2.57734060e-01 6.26789331e-01 -4.73583966e-01
-3.75170767e-01 2.70270973e-01 4.10285413e-01 4.56907541e-01
8.23084533e-01 4.56377193e-02 -6.83886290e-01 9.28406060e-01
-1.69445708e-01 3.91239941e-01 4.33919042e-01 7.16356099e-01
-1.62151039e-01 -7.40053654e-01 -1.25598773e-01 4.07980323e-01
-9.42467928e-01 -2.97405630e-01 -2.78991044e-01 1.20005870e+00
5.09860396e-01 4.90429342e-01 1.22313812e-01 -1.03762127e-01
1.05154544e-01 5.21328375e-02 2.51563668e-01 -9.83079553e-01
-1.16807640e+00 -2.86697209e-01 2.24161819e-01 -2.57021278e-01
-6.76609814e-01 -1.52940005e-01 -1.39145398e+00 2.53276914e-01
-3.21212560e-01 7.05535889e-01 6.29180372e-01 7.83040762e-01
3.71686339e-01 8.07507694e-01 -1.36504948e-01 -4.88404393e-01
-5.87418795e-01 -1.17985606e+00 -2.32668638e-01 4.89739388e-01
1.14497095e-01 -5.50470591e-01 -6.35187149e-01 2.79867142e-01] | [14.938945770263672, -2.7776315212249756] |
7cdbffbc-3f0c-4398-85b6-8d3385d23879 | few-shot-class-incremental-learning | 2004.10956 | null | https://arxiv.org/abs/2004.10956v2 | https://arxiv.org/pdf/2004.10956v2.pdf | Few-Shot Class-Incremental Learning | The ability to incrementally learn new classes is crucial to the development of real-world artificial intelligence systems. In this paper, we focus on a challenging but practical few-shot class-incremental learning (FSCIL) problem. FSCIL requires CNN models to incrementally learn new classes from very few labelled samples, without forgetting the previously learned ones. To address this problem, we represent the knowledge using a neural gas (NG) network, which can learn and preserve the topology of the feature manifold formed by different classes. On this basis, we propose the TOpology-Preserving knowledge InCrementer (TOPIC) framework. TOPIC mitigates the forgetting of the old classes by stabilizing NG's topology and improves the representation learning for few-shot new classes by growing and adapting NG to new training samples. Comprehensive experimental results demonstrate that our proposed method significantly outperforms other state-of-the-art class-incremental learning methods on CIFAR100, miniImageNet, and CUB200 datasets. | ['Xiaoyu Tao', 'Xiaopeng Hong', 'Songlin Dong', 'Xing Wei', 'Yihong Gong', 'Xinyuan Chang'] | 2020-04-23 | few-shot-class-incremental-learning-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Tao_Few-Shot_Class-Incremental_Learning_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Tao_Few-Shot_Class-Incremental_Learning_CVPR_2020_paper.pdf | cvpr-2020-6 | ['few-shot-class-incremental-learning'] | ['methodology'] | [ 2.56907016e-01 1.14451274e-01 -1.67862087e-01 -3.44831854e-01
-3.49143967e-02 -4.09076303e-01 5.63738227e-01 4.02711052e-03
-4.90562558e-01 1.07810867e+00 -1.01645410e-01 1.80132866e-01
-3.26786131e-01 -1.04076123e+00 -1.01113939e+00 -8.15795898e-01
-3.16850282e-02 3.40688109e-01 7.83527672e-01 1.16618304e-02
1.27682269e-01 5.58146298e-01 -1.85284042e+00 2.27515072e-01
9.39911187e-01 9.96113420e-01 2.14182436e-01 2.23919064e-01
-3.12241256e-01 9.53748763e-01 -3.22669744e-01 -1.54223666e-01
3.09334069e-01 -4.08505797e-01 -8.06118309e-01 4.58583310e-02
7.31266797e-01 -1.82448432e-01 -4.74006295e-01 1.10447955e+00
2.59764582e-01 5.45035183e-01 4.38808262e-01 -1.27312326e+00
-8.81972075e-01 8.46596062e-01 -4.26321000e-01 3.03605378e-01
-4.89347696e-01 8.45357105e-02 5.00044942e-01 -1.19922876e+00
9.54651356e-01 1.08845723e+00 7.87290037e-01 8.45648527e-01
-9.39941466e-01 -8.36961031e-01 5.64418793e-01 7.23797023e-01
-1.49472916e+00 -4.27690685e-01 1.01160395e+00 -1.55217692e-01
5.44520855e-01 -1.40132904e-01 7.74919033e-01 6.90008640e-01
-1.57054126e-01 9.94181693e-01 8.60607207e-01 -3.97034138e-01
6.01885557e-01 1.64979398e-01 3.32150757e-01 9.02170002e-01
5.02951562e-01 -4.15871106e-02 -6.53177261e-01 1.32360741e-01
6.12112105e-01 3.93937945e-01 -3.31238568e-01 -9.84744251e-01
-8.58987749e-01 6.75696731e-01 6.89624012e-01 3.84018660e-01
-7.04422072e-02 7.03966245e-02 4.14467275e-01 4.24066693e-01
5.03030181e-01 2.06283450e-01 -5.97357273e-01 1.29735246e-01
-8.39571893e-01 -1.52998701e-01 4.17835772e-01 1.12504148e+00
1.16203034e+00 2.85389364e-01 -1.49283394e-01 9.15077507e-01
-2.00000703e-01 1.52543321e-01 6.97956085e-01 -6.10068679e-01
1.21062793e-01 9.77208376e-01 -2.54399627e-01 -7.31043100e-01
-2.52985120e-01 -8.64122450e-01 -9.04666424e-01 1.16204597e-01
1.10446289e-02 -9.15154293e-02 -1.23111570e+00 1.64323092e+00
5.19327879e-01 8.54197264e-01 1.35225192e-01 3.00797760e-01
7.92410135e-01 6.80211186e-01 3.27752642e-02 -4.72329170e-01
6.46139741e-01 -1.19428957e+00 -4.01394933e-01 -1.19735695e-01
4.10763562e-01 -1.99875515e-02 9.17374134e-01 2.89245039e-01
-7.55589902e-01 -9.77765977e-01 -1.15343654e+00 2.43044317e-01
-6.07180655e-01 -2.75618583e-02 4.38147366e-01 3.85806680e-01
-8.84997785e-01 8.05726945e-01 -7.64721096e-01 -2.33924374e-01
1.15251565e+00 1.30648628e-01 -2.83530504e-01 -4.91051942e-01
-1.00173330e+00 5.14400005e-01 9.83579695e-01 -2.12448090e-02
-1.06193733e+00 -1.09764874e+00 -6.23816252e-01 1.93161711e-01
6.89706445e-01 -3.92788351e-01 1.14041638e+00 -1.17119205e+00
-1.48134494e+00 3.60437393e-01 5.52490465e-02 -7.01634526e-01
4.15045172e-01 -1.53716892e-01 -3.37070316e-01 2.16949031e-01
-5.92334494e-02 7.97620058e-01 1.16583705e+00 -1.33054256e+00
-9.00465250e-01 -2.62330920e-01 -2.20036492e-01 1.76797852e-01
-7.73405969e-01 -7.65886843e-01 -2.84291536e-01 -5.73810399e-01
1.55023456e-01 -7.30127811e-01 -1.43562937e-02 2.30155751e-01
1.33322790e-01 -4.59701449e-01 1.30641508e+00 -1.66036323e-01
1.10633254e+00 -2.09010124e+00 9.23781991e-02 -1.27392083e-01
2.75115132e-01 8.37994397e-01 -2.06097037e-01 -1.13400228e-01
-2.80089956e-02 -2.28582725e-01 -4.12899464e-01 -2.68237442e-01
-4.02019739e-01 4.36384290e-01 -3.49411786e-01 8.77809078e-02
6.54494092e-02 1.09951007e+00 -1.16229367e+00 -2.99436063e-01
2.35098362e-01 4.12665009e-01 -4.18820947e-01 -1.76207826e-01
-4.31196302e-01 2.51656085e-01 -8.52765813e-02 6.48312986e-01
7.68275619e-01 -2.33005673e-01 -5.34231178e-02 -8.88710767e-02
1.55987609e-02 -4.72998589e-01 -1.14614165e+00 1.96205962e+00
-2.34700546e-01 5.04815340e-01 -5.03360450e-01 -1.33161044e+00
1.03702068e+00 2.55205818e-02 2.29356825e-01 -5.94662905e-01
-1.32328281e-02 1.23030424e-01 -2.20673576e-01 -2.38273501e-01
2.66910285e-01 -1.04705602e-01 4.09379184e-01 2.86223441e-01
5.38998663e-01 1.78812072e-01 2.37219319e-01 1.62387282e-01
7.67395616e-01 -4.36063157e-03 3.03314030e-01 -1.16486490e-01
4.72726732e-01 -1.29244924e-01 8.81239295e-01 9.40015078e-01
-4.06152189e-01 2.90781230e-01 -2.96003912e-02 -9.49074388e-01
-7.58990526e-01 -1.26389909e+00 -5.78007586e-02 1.09055507e+00
8.20394456e-02 5.58452122e-02 -5.70430517e-01 -1.12909842e+00
5.98355010e-02 9.03560579e-01 -7.70639718e-01 -8.65858853e-01
-5.99352241e-01 -5.57040632e-01 6.92096725e-02 5.99580288e-01
9.93559599e-01 -1.13672793e+00 -6.58328414e-01 5.79252720e-01
1.26968771e-01 -7.47688472e-01 -3.29779625e-01 2.04168171e-01
-1.28443754e+00 -1.32601917e+00 -7.54078388e-01 -1.18132508e+00
8.60687733e-01 4.33816642e-01 7.43911743e-01 -6.93264380e-02
-5.00923753e-01 5.63295543e-01 -5.98178625e-01 -5.25472701e-01
3.59933525e-02 3.12218398e-01 1.29201531e-01 4.09381896e-01
3.47153395e-01 -7.35376120e-01 -5.28293848e-01 1.27568975e-01
-9.67730641e-01 6.77355900e-02 5.26062071e-01 9.94893014e-01
9.30115044e-01 5.97937644e-01 1.04949629e+00 -1.18442416e+00
2.44613081e-01 -4.54086185e-01 -4.27673727e-01 6.11868143e-01
-7.71985233e-01 1.71294212e-01 9.46939170e-01 -9.65532541e-01
-1.33692658e+00 2.60723710e-01 4.47533637e-01 -6.86254740e-01
-1.07051849e-01 4.99553710e-01 -5.66349775e-02 -4.22634989e-01
6.73289061e-01 6.04213774e-01 -2.57992357e-01 -6.44810557e-01
7.50348568e-01 3.46291304e-01 8.28756273e-01 -4.66611892e-01
7.49483466e-01 7.62367845e-01 -2.72073317e-02 -7.92863846e-01
-1.16525435e+00 -3.84560704e-01 -1.07638514e+00 -2.93882728e-01
3.90239805e-01 -7.70045638e-01 -1.38059407e-01 8.73479605e-01
-7.58472204e-01 -4.47535247e-01 -9.34683859e-01 2.17358857e-01
-4.43244398e-01 5.52534945e-02 -1.77311331e-01 -4.34704661e-01
-3.82845908e-01 -4.71732557e-01 3.87789845e-01 6.04186177e-01
4.81332600e-01 -9.65974271e-01 1.92162320e-01 -1.59110740e-01
6.21007323e-01 2.52777010e-01 9.96899247e-01 -5.58153093e-01
-6.55724764e-01 -1.94554403e-01 -1.59746096e-01 4.48901355e-01
3.46990287e-01 -3.84577453e-01 -9.88239825e-01 -7.34351635e-01
-9.06142145e-02 -4.62363392e-01 1.49340045e+00 1.43075824e-01
1.42341220e+00 -3.45930129e-01 -4.34259951e-01 7.47342706e-01
1.43423915e+00 4.57098961e-01 5.45971513e-01 1.74499810e-01
8.13811779e-01 1.60039827e-01 4.13520724e-01 3.81866306e-01
2.52207786e-01 1.48932934e-01 3.11071366e-01 2.24429414e-01
-6.74343884e-01 -4.62807685e-01 -5.40563613e-02 9.95931149e-01
7.83529598e-03 2.07085803e-01 -7.36689627e-01 7.33580649e-01
-1.95870686e+00 -9.38988090e-01 6.24704063e-01 2.06862450e+00
1.00549674e+00 2.45005012e-01 -3.70834112e-01 1.24487616e-01
8.44491899e-01 -3.84091772e-02 -1.17619658e+00 2.20515743e-01
-1.21515885e-01 4.80176479e-01 1.91177502e-01 2.70539045e-01
-1.22039354e+00 1.21711528e+00 5.27982044e+00 1.01523209e+00
-1.34773302e+00 2.57111281e-01 6.87842727e-01 -2.28308260e-01
7.82074258e-02 -3.84004824e-02 -9.60782766e-01 2.47167319e-01
6.87212765e-01 -4.08398509e-01 4.22774732e-01 1.22462916e+00
-4.44399744e-01 5.49421348e-02 -9.67024565e-01 1.05685484e+00
4.13104773e-01 -1.70140791e+00 3.95900816e-01 -5.90819955e-01
1.36108816e+00 -2.20905561e-02 3.88843454e-02 6.71876073e-01
2.53169596e-01 -6.03627443e-01 5.69337666e-01 1.00103474e+00
9.55135047e-01 -9.78753030e-01 5.59407830e-01 3.45827967e-01
-1.44142628e+00 -4.60307628e-01 -7.00680554e-01 1.18270017e-01
-2.77444690e-01 5.58978617e-01 -8.32191467e-01 3.37371528e-01
8.47089767e-01 1.11032069e+00 -9.34850335e-01 1.48753941e+00
-2.53097951e-01 7.65634596e-01 -1.31658778e-01 -3.34093045e-03
9.88685489e-02 1.58045366e-01 3.94076496e-01 7.16024637e-01
2.33989835e-01 2.48347580e-01 2.36137182e-01 7.59364009e-01
-5.21492422e-01 -1.78890035e-01 -5.19808650e-01 4.07276973e-02
7.30111063e-01 9.20724511e-01 -9.74473596e-01 -7.22414255e-01
-2.47360587e-01 1.08459961e+00 6.60987794e-01 4.20333624e-01
-3.71800780e-01 -8.23047042e-01 2.03717962e-01 -9.14271176e-02
6.69901431e-01 -6.06661811e-02 -1.49514265e-02 -1.16241038e+00
1.46886975e-01 -3.45403582e-01 5.35524666e-01 -5.77681601e-01
-1.20709682e+00 5.13426602e-01 7.61958770e-03 -1.14261746e+00
1.07745707e-01 -3.70409012e-01 -7.54511595e-01 1.91362917e-01
-1.90946376e+00 -1.23184526e+00 -5.57702363e-01 8.67961943e-01
8.15365136e-01 -5.61649799e-01 6.09161377e-01 2.31880486e-01
-3.75592500e-01 7.47193992e-01 4.61595446e-01 -1.81668289e-02
4.95505899e-01 -9.57585573e-01 2.90812343e-01 7.93705881e-01
2.91883320e-01 4.52484131e-01 7.06385970e-02 -7.29457319e-01
-1.02842724e+00 -1.69503820e+00 5.26306033e-01 -5.18792048e-02
4.39550251e-01 -2.68466562e-01 -1.22113609e+00 6.05486214e-01
-2.08925664e-01 4.53292429e-01 2.83619195e-01 -1.64534420e-01
-6.01834357e-01 -5.61301649e-01 -1.13072467e+00 3.06314379e-01
1.25184762e+00 -3.05060536e-01 -7.16330290e-01 2.24994391e-01
1.04267180e+00 1.04295351e-01 -4.19565350e-01 4.79033917e-01
3.81917983e-01 -5.67563295e-01 9.38326597e-01 -7.98869073e-01
-2.40486071e-01 -3.49513233e-01 1.67562217e-01 -1.63512361e+00
-5.23693621e-01 -2.90145069e-01 -5.65374017e-01 1.05841815e+00
9.87462625e-02 -6.70197010e-01 1.01193929e+00 1.21447116e-01
-3.21767956e-01 -7.11095452e-01 -1.06558955e+00 -1.14629912e+00
3.17510635e-01 1.36838795e-03 7.99145937e-01 1.24568820e+00
-3.04262072e-01 1.92938194e-01 -3.28389198e-01 -1.38968423e-01
8.27978671e-01 1.50256038e-01 4.60228533e-01 -1.66592121e+00
-3.18841189e-02 -1.81666642e-01 -5.51957488e-01 -6.84979379e-01
4.27669473e-02 -1.02445400e+00 -1.02704808e-01 -1.36648762e+00
4.10601348e-01 -5.63383460e-01 -7.60495663e-01 9.62112665e-01
-1.76371574e-01 1.41965672e-01 2.42474973e-01 2.12677658e-01
-1.19215572e+00 9.79139924e-01 1.26024961e+00 -3.96044075e-01
-4.44793016e-01 -2.12665871e-02 -5.80275059e-01 7.57708132e-01
9.25646544e-01 -5.83657384e-01 -8.28041434e-01 -3.86314839e-01
-9.28191189e-03 -6.68399513e-01 2.35862955e-01 -1.45650578e+00
6.72563314e-01 -1.64795816e-01 5.95339298e-01 -6.74388528e-01
2.49604359e-02 -7.91614830e-01 -1.76484257e-01 7.80438781e-01
-2.04114437e-01 -4.95484531e-01 2.88629234e-01 8.89648736e-01
-6.89497292e-02 -2.87613064e-01 1.15158319e+00 -4.17579889e-01
-1.33042467e+00 7.44758606e-01 8.43466371e-02 2.74369150e-01
1.22889650e+00 -2.50190198e-01 -4.98069614e-01 1.04139395e-01
-8.80992591e-01 1.82166427e-01 1.44262403e-01 5.23787916e-01
1.02143586e+00 -1.47495234e+00 -5.26071906e-01 4.20899898e-01
3.11295152e-01 3.66560817e-01 3.77622038e-01 2.41004035e-01
-3.68955463e-01 1.67752534e-01 -4.57348943e-01 -3.72767866e-01
-9.94893789e-01 8.76763761e-01 2.92667806e-01 7.78568238e-02
-7.11793542e-01 1.15959048e+00 -6.25629351e-02 -6.28940940e-01
5.68935633e-01 -1.31301716e-01 -2.96361893e-01 1.31226450e-01
8.58988523e-01 6.19156957e-01 6.48336411e-02 -1.55349717e-01
-3.71969230e-02 4.12475437e-01 -7.00814486e-01 4.39813495e-01
1.71799505e+00 -1.53095499e-01 -2.21564155e-02 7.69976199e-01
1.13635790e+00 -6.83141351e-01 -1.68258870e+00 -8.68646264e-01
-5.27750254e-02 -3.27986002e-01 -1.38852000e-02 -9.13095832e-01
-1.22869682e+00 9.20319319e-01 1.03612161e+00 -3.79771918e-01
1.10329270e+00 -1.32575184e-01 8.00938725e-01 9.48895276e-01
6.78282559e-01 -1.52099109e+00 5.33162236e-01 6.60985410e-01
6.87104166e-01 -1.04707003e+00 -8.53651464e-02 -1.47913992e-01
-3.14717144e-01 1.24396944e+00 9.27486420e-01 -8.35107639e-02
1.05595613e+00 -4.72724080e-01 -2.88881332e-01 -1.00586608e-01
-6.52357697e-01 -1.27613842e-01 2.46993601e-01 7.89472342e-01
-2.81837851e-01 -4.06143293e-02 1.26799166e-01 4.34115380e-01
5.34740053e-02 3.36049199e-01 3.94992828e-01 1.12817371e+00
-9.88685310e-01 -7.66664863e-01 7.53504783e-02 6.05798244e-01
3.57062161e-01 -6.33099452e-02 -3.13544542e-01 6.94543600e-01
6.38823271e-01 5.49449384e-01 2.03716099e-01 -3.49878609e-01
1.76504120e-01 3.53315234e-01 4.87049758e-01 -7.93830037e-01
9.46431905e-02 -4.25782472e-01 -5.99256456e-01 -2.16052815e-01
-3.47358823e-01 -5.75587094e-01 -1.21971583e+00 -1.32970354e-02
-4.22124475e-01 1.85621276e-01 4.17905897e-01 9.34346974e-01
4.91591871e-01 6.35528803e-01 6.59617901e-01 -5.13198495e-01
-4.63041574e-01 -8.11158359e-01 -6.14975154e-01 3.41303796e-01
2.63597399e-01 -9.92687702e-01 -2.60214657e-01 2.34908566e-01] | [9.824529647827148, 3.3752622604370117] |
f2b74072-a88f-4dc8-869d-13bb9902faa8 | just-noticeable-defocus-blur-detection-and | null | null | http://openaccess.thecvf.com/content_cvpr_2015/html/Shi_Just_Noticeable_Defocus_2015_CVPR_paper.html | http://openaccess.thecvf.com/content_cvpr_2015/papers/Shi_Just_Noticeable_Defocus_2015_CVPR_paper.pdf | Just Noticeable Defocus Blur Detection and Estimation | We tackle a fundamental problem to detect and estimate just noticeable blur (JNB) caused by defocus that spans a small number of pixels in images. This type of blur is common during photo taking. Although it is not strong, the slight edge blurriness contains informative clues related to depth. We found existing blur descriptors based on local information cannot distinguish this type of small blur reliably from unblurred structures. We propose a simple yet effective blur feature via sparse representation and image decomposition. It directly establishes correspondence between sparse edge representation and blur strength estimation. Extensive experiments manifest the generality and robustness of this feature. | ['Jiaya Jia', 'Li Xu', 'Jianping Shi'] | 2015-06-01 | null | null | null | cvpr-2015-6 | ['defocus-blur-detection'] | ['computer-vision'] | [ 8.75066817e-02 -5.93600333e-01 9.29759070e-02 -4.13618147e-01
-2.00985238e-01 -5.55245757e-01 3.80554527e-01 -4.26468760e-01
1.20485783e-01 1.10883713e+00 9.22732472e-01 1.81146830e-01
-2.33577177e-01 -2.00591609e-01 -6.07332528e-01 -7.33606994e-01
-2.83083916e-02 -5.73241711e-01 2.81578302e-01 1.21613853e-01
6.91182196e-01 6.25259578e-01 -1.36800826e+00 2.38109216e-01
1.21089983e+00 9.29272294e-01 5.48036516e-01 7.72434235e-01
3.52166533e-01 1.32762969e+00 -7.84558475e-01 7.83341303e-02
3.13725471e-01 -3.96185219e-01 -6.55857205e-01 3.73775244e-01
9.93008256e-01 -9.20367539e-01 -7.18934119e-01 1.46151781e+00
4.66564149e-01 -1.68880150e-01 5.91756523e-01 -8.06723058e-01
-1.00570941e+00 1.75688237e-01 -9.12401855e-01 8.70438635e-01
6.81442976e-01 2.12944776e-01 4.46578622e-01 -1.00857186e+00
5.01200199e-01 1.06126738e+00 7.70511687e-01 1.12655103e-01
-9.11610484e-01 -1.16798893e-01 -4.68411773e-01 4.60264325e-01
-1.30358148e+00 -6.51548207e-01 9.19657409e-01 -5.18568397e-01
4.53452438e-01 5.04242361e-01 3.86705279e-01 9.09806132e-01
5.72308302e-01 5.00004113e-01 1.54239118e+00 -3.00872117e-01
1.26841128e-01 3.60424817e-02 3.16327542e-01 6.33728027e-01
8.28637838e-01 3.40514570e-01 -5.53374708e-01 -7.83298910e-02
1.15199697e+00 1.49195582e-01 -1.29646599e+00 -2.57394195e-01
-1.25375962e+00 2.78115541e-01 6.56331599e-01 3.10365766e-01
-2.82727540e-01 2.25950927e-01 -5.04440330e-02 2.24376336e-01
2.59305447e-01 6.48048580e-01 -2.12940142e-01 -5.50708532e-01
-8.64846468e-01 -2.02651948e-01 5.60379863e-01 8.99351954e-01
9.14195895e-01 -7.22664669e-02 -4.04930115e-01 9.29833055e-01
-7.24158287e-02 5.30310273e-01 3.77424359e-01 -1.33998847e+00
-7.69361854e-02 2.10279897e-01 6.65521920e-01 -1.14443851e+00
4.41777892e-03 -2.51025528e-01 -9.07116055e-01 3.64560395e-01
4.15589392e-01 -1.89602170e-02 -6.67777896e-01 1.24101484e+00
3.43757309e-02 5.37309825e-01 -3.79219800e-01 1.56916881e+00
8.11653137e-01 1.64738953e-01 -7.78502584e-01 -4.42236334e-01
1.34005713e+00 -1.00236666e+00 -1.10305417e+00 -3.27561110e-01
-1.55464754e-01 -1.21714258e+00 6.73921287e-01 3.01053196e-01
-1.16827178e+00 -5.78563690e-01 -9.06917810e-01 -3.34407896e-01
1.02434859e-01 1.91135809e-01 8.19448054e-01 5.78921676e-01
-1.14628553e+00 4.42804754e-01 -2.77147412e-01 -2.15430811e-01
3.16941112e-01 -1.43045083e-01 -3.61164898e-01 -5.96332014e-01
-8.98777604e-01 1.02955794e+00 -3.60761762e-01 4.53327119e-01
-6.97592318e-01 -6.08403742e-01 -9.31713402e-01 -8.06929544e-02
4.13997620e-02 -8.67550910e-01 1.13811457e+00 -5.85949004e-01
-1.14045775e+00 6.02183700e-01 -7.12805927e-01 -4.20483768e-01
6.03319824e-01 -6.67771220e-01 -3.20658088e-01 4.43319887e-01
-4.01481502e-02 1.07346818e-01 1.46350813e+00 -1.39469302e+00
-2.03726426e-01 -9.08402130e-02 5.99055141e-02 2.51525581e-01
2.27751270e-01 5.89876659e-02 -5.06368559e-03 -7.02698469e-01
4.10736986e-02 -3.27182531e-01 4.74977717e-02 1.34298384e-01
-3.05337042e-01 3.19121808e-01 9.76114333e-01 -6.21909738e-01
1.40021777e+00 -2.09591603e+00 -3.39715838e-01 -4.02049363e-01
5.20718455e-01 5.95415086e-02 1.18310317e-01 4.31039810e-01
-1.25657376e-02 -3.77023786e-01 -3.03583831e-01 2.41277173e-01
-3.30026537e-01 4.13151868e-02 -4.45929706e-01 8.67699802e-01
9.06452686e-02 8.33936870e-01 -1.22040462e+00 -2.43059516e-01
4.09738749e-01 4.41378683e-01 -1.83551311e-01 4.37831759e-01
4.65085775e-01 2.57138848e-01 -1.23599850e-01 8.45493972e-01
1.35534501e+00 -2.50139892e-01 -5.28232455e-01 -8.43777716e-01
-3.69755715e-01 8.34674388e-02 -9.03577030e-01 1.65764737e+00
-1.07117489e-01 1.21347642e+00 1.73776016e-01 -5.05826116e-01
7.18348086e-01 6.92074969e-02 1.29914775e-01 -7.60170147e-02
2.37309374e-02 2.17308924e-01 -2.14298338e-01 -1.15112519e+00
7.05420315e-01 -3.85018773e-02 4.42876548e-01 7.99573436e-02
-3.01383913e-01 -4.09158558e-01 -4.77871075e-02 3.46557796e-02
1.45726061e+00 -9.40679163e-02 6.80652857e-01 -5.56879938e-01
4.78164911e-01 -4.42986220e-01 3.15579325e-01 8.28022659e-01
-6.82273149e-01 1.23104930e+00 1.01798676e-01 -3.36110830e-01
-6.24980807e-01 -1.01796985e+00 -4.05253172e-01 2.13253319e-01
8.39551270e-01 -2.03929305e-01 -7.13857949e-01 -4.98883903e-01
2.06376091e-01 2.31881738e-01 -5.99539518e-01 -2.14501277e-01
-2.29985699e-01 -4.82107341e-01 3.31353284e-02 2.90994734e-01
8.13174725e-01 -5.90490878e-01 -6.68717802e-01 -2.90865183e-01
-3.94488931e-01 -1.25033855e+00 -1.15712702e+00 -2.23614439e-01
-7.70946264e-01 -1.33627319e+00 -1.04332817e+00 -8.07936847e-01
8.95023525e-01 1.19422185e+00 1.04509842e+00 -1.56585991e-01
-6.08710051e-01 5.25648773e-01 -4.01006043e-02 1.30303293e-01
1.19928448e-02 -9.76173460e-01 5.62416315e-02 2.13360205e-01
4.24801677e-01 -4.01240677e-01 -1.20517433e+00 3.84831041e-01
-5.80897689e-01 -1.84751332e-01 7.10560322e-01 8.63212526e-01
-1.36825711e-01 2.80614346e-01 -7.33744130e-02 -3.78434330e-01
9.51310217e-01 -2.39166230e-01 -3.02011549e-01 -2.55930852e-02
-3.54063541e-01 3.10294218e-02 9.42065343e-02 -2.22044095e-01
-1.37234986e+00 -4.31083590e-01 5.88686705e-01 -7.33815074e-01
-4.96098965e-01 1.82659402e-01 2.49440283e-01 -5.62848270e-01
8.16876590e-01 1.22972779e-01 -3.28473076e-02 -6.49489105e-01
1.54682457e-01 1.05584812e+00 8.39508891e-01 -1.61398828e-01
5.82132816e-01 7.18685985e-01 5.33025637e-02 -1.14909232e+00
-8.27354312e-01 -8.84615660e-01 -4.23898190e-01 -2.17466503e-01
4.74934489e-01 -1.14235151e+00 -7.87492573e-01 7.54282773e-01
-1.32250869e+00 7.73648173e-02 -2.69937128e-01 6.67479038e-01
-2.96168238e-01 1.14653158e+00 -1.01668942e+00 -8.69514108e-01
1.47452012e-01 -8.58907759e-01 1.07233989e+00 3.11148971e-01
-1.27017111e-01 -9.90294755e-01 -6.64371066e-03 4.57439989e-01
6.62505448e-01 1.62265018e-01 6.30460158e-02 6.20012224e-01
-9.79958832e-01 2.03196816e-02 -8.06841791e-01 5.41108429e-01
8.63477647e-01 -1.02479011e-01 -1.19584072e+00 -2.57197917e-01
7.61160195e-01 6.99325204e-02 1.06311655e+00 8.61393571e-01
1.00621152e+00 -3.65250319e-01 -4.97717001e-02 7.02582657e-01
1.68936670e+00 5.55976890e-02 9.24788356e-01 3.05509996e-02
6.83852315e-01 1.93016604e-01 5.84101796e-01 3.57923090e-01
-3.31989042e-02 5.55010021e-01 2.56154150e-01 -2.50945449e-01
-7.96195865e-01 3.63302678e-02 3.90617788e-01 7.00103104e-01
-2.55073339e-01 2.45693222e-01 -4.82998967e-01 6.04803741e-01
-1.68580878e+00 -1.15507114e+00 -4.71034586e-01 2.14714670e+00
1.05480468e+00 -1.28787547e-01 -4.30021942e-01 -1.36845306e-01
9.50725377e-01 3.32120568e-01 -2.66086876e-01 -2.62047917e-01
-4.64593202e-01 -1.14300020e-01 6.63590491e-01 8.96464169e-01
-8.99108648e-01 5.58925629e-01 7.48342323e+00 5.98164082e-01
-1.01456070e+00 -1.27739459e-01 3.86084735e-01 2.42393896e-01
-2.82213211e-01 -3.53673808e-02 -4.19797033e-01 9.63145316e-01
3.37814003e-01 -4.87084389e-02 6.87971950e-01 6.53316975e-01
6.22431219e-01 -8.18026245e-01 -1.00894392e+00 1.62261569e+00
3.57222021e-01 -1.13532460e+00 -1.98589504e-01 -1.13130808e-01
9.21557248e-01 -6.39458746e-02 1.59529019e-02 -6.21568263e-01
-6.21230900e-02 -9.54633951e-01 4.35494334e-01 1.01288378e+00
8.46619010e-01 -1.74751848e-01 7.85054803e-01 9.63792298e-03
-8.20672989e-01 -7.94403479e-02 -7.81101763e-01 -5.62853277e-01
1.14223190e-01 1.34109652e+00 -5.80088139e-01 2.50476778e-01
7.32565463e-01 1.16303194e+00 -6.11368775e-01 1.74707401e+00
-4.17203307e-01 3.64480525e-01 1.16344407e-01 3.35320443e-01
-1.21111423e-01 -4.99268144e-01 8.90259802e-01 1.39279401e+00
6.48633301e-01 9.48178396e-02 -4.98410136e-01 1.04395413e+00
1.18786573e-01 -7.25978017e-01 -1.01880181e+00 3.42167467e-01
3.88625741e-01 1.22544861e+00 -2.59574145e-01 -2.51287609e-01
-5.69796801e-01 1.61651921e+00 -9.87826139e-02 6.94399059e-01
-6.47224307e-01 -5.08101046e-01 1.08554935e+00 -9.18765441e-02
4.19172853e-01 -1.64344609e-01 -4.17901188e-01 -1.72296095e+00
8.03476870e-02 -5.68849027e-01 -2.07406193e-01 -1.49871457e+00
-1.57953131e+00 3.24893653e-01 -2.22407520e-01 -1.51337671e+00
2.11497396e-01 -7.07789361e-01 -7.12777495e-01 9.55530345e-01
-1.80017829e+00 -6.91436470e-01 -1.07661450e+00 6.00572765e-01
6.46737933e-01 3.65583926e-01 3.89872938e-01 -4.11241129e-02
-2.75890589e-01 -4.62059267e-02 1.89135119e-01 -1.83003306e-01
1.10767984e+00 -1.51490176e+00 -1.91784296e-02 1.25356078e+00
-2.72882342e-01 8.47244561e-01 1.20565510e+00 -5.27871549e-01
-1.41210818e+00 -6.01494074e-01 8.04786623e-01 -5.77878773e-01
6.13067925e-01 8.92753899e-02 -8.19807112e-01 2.34076962e-01
7.43639588e-01 3.69613707e-01 5.81656583e-02 -2.71464676e-01
-3.36916357e-01 -2.17575446e-01 -1.07464325e+00 1.89458683e-01
1.13591099e+00 -8.18428874e-01 -1.01398575e+00 1.88464969e-01
3.95271540e-01 -5.94964802e-01 -6.28868937e-01 4.28877771e-01
3.17842901e-01 -1.66220510e+00 1.12107611e+00 1.65879622e-01
5.42974412e-01 -5.92507005e-01 7.91152790e-02 -1.48959017e+00
-7.49222219e-01 -1.19353509e+00 -5.13890862e-01 9.02914107e-01
-3.90797257e-01 -6.61356091e-01 4.26959187e-01 1.84102997e-01
-3.17227691e-01 -2.22575232e-01 -4.99655575e-01 -8.96266937e-01
-6.92699611e-01 1.54278502e-01 3.24634314e-01 1.19356120e+00
2.82977134e-01 2.74693340e-01 -7.14900136e-01 3.48869383e-01
9.63116467e-01 4.15396571e-01 4.81061310e-01 -9.60508585e-01
-1.09483115e-01 -1.64528146e-01 -5.85836709e-01 -1.61277235e+00
-5.10009468e-01 -1.85536325e-01 1.73351541e-01 -1.57510817e+00
4.80926633e-01 7.53737846e-03 -3.97327662e-01 -2.09673673e-01
-4.18645889e-01 2.80628890e-01 -3.71416122e-01 4.31715637e-01
-2.25834668e-01 3.60285372e-01 1.66258383e+00 -9.70226601e-02
2.80105501e-01 -1.52330488e-01 -7.76757896e-01 7.75272012e-01
4.44877833e-01 1.07600249e-01 -4.57028359e-01 -5.23479700e-01
-1.05428174e-01 2.39427779e-02 6.59779191e-01 -1.03341055e+00
2.76649803e-01 -3.24365050e-01 5.78433990e-01 -3.94638628e-01
3.31390262e-01 -8.31736922e-01 -4.24222425e-02 3.73376876e-01
1.57177627e-01 -3.53708148e-01 -5.98860048e-02 7.71554828e-01
-5.74796021e-01 -4.37730223e-01 8.48576069e-01 -2.32268482e-01
-9.94812608e-01 -6.38642013e-02 -1.14721395e-01 -7.43804947e-02
7.50023365e-01 -5.48315585e-01 -9.79337633e-01 -6.50113225e-01
-9.26023796e-02 -4.57910925e-01 8.52546096e-01 1.33234203e-01
9.38760698e-01 -1.23267341e+00 -4.46616024e-01 4.27893907e-01
-1.26980647e-01 -3.19630206e-01 3.00056905e-01 1.29668963e+00
-8.80374789e-01 3.82655710e-01 -2.88910091e-01 -6.30080402e-01
-1.33075857e+00 6.27561390e-01 4.12499547e-01 6.36703014e-01
-5.20216286e-01 1.13871467e+00 4.55172032e-01 7.46037900e-01
9.62932482e-02 -8.92628312e-01 -2.01374255e-02 -3.36316347e-01
7.76295066e-01 7.09584713e-01 -1.67585522e-01 -4.49447483e-01
-3.69360715e-01 8.31838667e-01 1.03215925e-01 4.68174189e-01
1.00667822e+00 -7.90054440e-01 -5.99283576e-01 4.56399292e-01
1.17659867e+00 6.81744695e-01 -1.71474552e+00 5.43320738e-02
-3.07819068e-01 -1.41362751e+00 2.10880488e-01 -7.14886427e-01
-8.04377913e-01 7.61204064e-01 5.24181783e-01 4.50733811e-01
1.37027621e+00 2.97250021e-02 8.01858008e-01 3.93966697e-02
4.39730972e-01 -7.39928365e-01 2.17119575e-01 3.10416371e-01
1.24198210e+00 -1.42852926e+00 2.97600091e-01 -8.37770760e-01
-2.93816507e-01 1.22112370e+00 4.03540194e-01 -2.72836804e-01
6.36005998e-01 3.85072142e-01 6.76810071e-02 -2.90288925e-01
-5.22906899e-01 -3.25052649e-01 5.26917815e-01 7.65947819e-01
4.08771425e-01 -2.90179461e-01 -1.61040232e-01 -1.41497329e-01
6.41461611e-02 2.37947762e-01 1.05221188e+00 8.49148631e-01
-8.99606824e-01 -3.42590511e-01 -6.72252774e-01 1.21343434e-01
-3.45183581e-01 -1.99409053e-01 -3.19036454e-01 3.53057116e-01
9.12711993e-02 1.00224435e+00 1.22444192e-02 -1.04025774e-01
-2.59962045e-02 -5.95101178e-01 8.54646146e-01 -2.29226947e-01
1.34743422e-01 1.34500548e-01 4.95668612e-02 -8.26298058e-01
-6.96292520e-01 -5.11302054e-01 -4.31156039e-01 -2.27989510e-01
-4.40281689e-01 2.01674160e-02 3.04065138e-01 4.47931856e-01
3.40572596e-01 2.01993033e-01 7.67435849e-01 -7.45508134e-01
-3.31067443e-01 -1.22541690e+00 -1.02651334e+00 7.03824997e-01
1.24235201e+00 -6.52082980e-01 -1.07965922e+00 3.32496524e-01] | [11.598541259765625, -2.750579357147217] |
62c1de37-13d7-4371-a445-a8485f325770 | a-simple-joint-model-for-improved-contextual | 1904.02306 | null | https://arxiv.org/abs/1904.02306v4 | https://arxiv.org/pdf/1904.02306v4.pdf | A Simple Joint Model for Improved Contextual Neural Lemmatization | English verbs have multiple forms. For instance, talk may also appear as talks, talked or talking, depending on the context. The NLP task of lemmatization seeks to map these diverse forms back to a canonical one, known as the lemma. We present a simple joint neural model for lemmatization and morphological tagging that achieves state-of-the-art results on 20 languages from the Universal Dependencies corpora. Our paper describes the model in addition to training and decoding procedures. Error analysis indicates that joint morphological tagging and lemmatization is especially helpful in low-resource lemmatization and languages that display a larger degree of morphological complexity. Code and pre-trained models are available at https://sigmorphon.github.io/sharedtasks/2019/task2/. | ['Shijie Wu', 'Ryan Cotterell', 'Chaitanya Malaviya'] | 2019-04-04 | a-simple-joint-model-for-improved-contextual-1 | https://aclanthology.org/N19-1155 | https://aclanthology.org/N19-1155.pdf | naacl-2019-6 | ['morphological-tagging'] | ['natural-language-processing'] | [-2.02733666e-01 2.05790550e-01 -4.27850336e-01 -6.00281596e-01
-1.09203720e+00 -9.04921830e-01 4.11391497e-01 2.47688159e-01
-7.19025075e-01 8.08809996e-01 6.45831108e-01 -7.21646965e-01
4.36501950e-01 -5.33621550e-01 -6.26963139e-01 -4.38647091e-01
9.24155414e-02 7.20321119e-01 -2.15097025e-01 2.38797367e-02
-4.10209149e-01 4.28117454e-01 -6.96043372e-01 5.85627913e-01
6.00627542e-01 4.19640243e-01 3.35022122e-01 2.90152848e-01
-6.44289196e-01 2.15647846e-01 -6.22088730e-01 -9.78461742e-01
1.08486928e-01 -1.56005502e-01 -1.20336711e+00 -1.04060635e-01
3.76769215e-01 1.11430511e-01 -2.35306352e-01 1.11827934e+00
3.36278617e-01 7.84789398e-02 5.38741052e-01 -5.44083357e-01
-7.68308342e-01 1.46759188e+00 -3.14667463e-01 3.16298634e-01
2.44596615e-01 -2.99860518e-02 1.40622163e+00 -9.21798825e-01
8.30427289e-01 1.34765196e+00 7.88212359e-01 5.83813071e-01
-1.39639533e+00 -7.60480404e-01 2.24106163e-01 -9.49710459e-02
-1.46943998e+00 -4.92276043e-01 4.23884869e-01 -2.73931444e-01
1.47559178e+00 -4.89417389e-02 4.90060359e-01 1.25539243e+00
2.27169201e-01 8.48596334e-01 1.07808399e+00 -4.50344145e-01
-2.12009296e-01 5.98036759e-02 3.37348342e-01 6.56783700e-01
5.77612817e-01 -4.65429932e-01 -2.52921402e-01 -8.17173999e-03
5.97644687e-01 -4.19672966e-01 -1.07501969e-01 1.12910539e-01
-1.24828720e+00 8.58779430e-01 1.86509728e-01 7.76089787e-01
-4.00922835e-01 6.85284436e-02 5.21930218e-01 2.31811449e-01
5.82183242e-01 5.53118229e-01 -1.01732421e+00 -1.36018351e-01
-6.48277402e-01 -1.55309513e-01 1.01237583e+00 1.05335021e+00
8.28589380e-01 1.14945330e-01 1.92892328e-01 1.02564156e+00
1.37331516e-01 3.74987930e-01 4.64078188e-01 -7.73823202e-01
7.34419346e-01 3.26579750e-01 -2.28190392e-01 -2.86437184e-01
-5.01354337e-01 -1.85170457e-01 -4.33252126e-01 -2.82344550e-01
4.92121845e-01 -4.43396568e-01 -1.13372386e+00 1.95352137e+00
2.48581827e-01 -3.21218878e-01 1.98089913e-01 3.73310030e-01
1.13863277e+00 5.61096668e-01 7.10268021e-01 -2.79686719e-01
1.76413643e+00 -6.74826741e-01 -1.04732168e+00 -6.48319125e-01
9.92069125e-01 -1.11965489e+00 1.15355182e+00 1.95782080e-01
-1.34577835e+00 -2.74792135e-01 -7.41103768e-01 -4.09976155e-01
-6.80415988e-01 2.77184665e-01 1.10254884e+00 6.02978766e-01
-9.38172400e-01 4.21346962e-01 -1.13796401e+00 -7.73056865e-01
2.56310314e-01 4.26260591e-01 -7.21902490e-01 1.74035743e-01
-1.19739425e+00 1.15864491e+00 7.27502525e-01 -2.17651248e-01
-4.63570833e-01 -2.98683643e-01 -9.91205215e-01 -1.75469220e-01
7.45187178e-02 -4.12891150e-01 1.42775512e+00 -1.02395308e+00
-1.35390258e+00 1.48359287e+00 -1.37658760e-01 -3.53389740e-01
1.56837795e-02 -4.64697659e-01 -4.87516850e-01 -1.66960120e-01
1.94607958e-01 8.01461518e-01 3.49757910e-01 -1.04824579e+00
-5.64553022e-01 -1.17088601e-01 5.43775894e-02 2.66394913e-01
-2.62644231e-01 6.06149673e-01 -5.30347824e-01 -6.62273169e-01
1.69574365e-01 -7.37369597e-01 5.72317466e-03 -6.80796325e-01
-5.04856706e-01 -6.53333902e-01 1.27162546e-01 -9.52765644e-01
1.29627550e+00 -2.17032170e+00 1.09684251e-01 -3.22592735e-01
-9.46464166e-02 3.66332620e-01 -5.77874593e-02 5.84552884e-01
-4.20448154e-01 5.44872642e-01 -2.37039074e-01 -7.55204141e-01
7.25143552e-02 6.66587174e-01 -2.48602461e-02 5.28311849e-01
2.76332796e-01 9.64424253e-01 -6.76082373e-01 -7.36115336e-01
-5.03152572e-02 4.03444320e-01 -2.48713017e-01 4.42907140e-02
-1.00116909e-01 1.69567183e-01 -1.84339762e-01 6.87032700e-01
6.52778745e-01 1.70682976e-03 7.46355712e-01 -3.09242636e-01
-2.12160274e-01 1.15488863e+00 -9.74043190e-01 1.90205181e+00
-5.82325816e-01 3.04040283e-01 3.21198583e-01 -7.89803684e-01
5.02911866e-01 5.83949506e-01 1.90385133e-01 -4.48755324e-02
6.61910236e-01 2.43769705e-01 8.42813328e-02 -3.74643743e-01
4.39437121e-01 -3.81075680e-01 -5.19002855e-01 2.43335187e-01
6.38354540e-01 -1.94698006e-01 5.41794300e-01 1.12332836e-01
1.08643019e+00 2.62634158e-01 8.42263401e-01 -3.29062372e-01
-9.11046751e-03 1.16776168e-01 8.40032399e-01 2.62403667e-01
-2.54454724e-02 2.88078696e-01 3.60074103e-01 -1.52099133e-01
-8.06243122e-01 -1.51228297e+00 -4.66885895e-01 1.18981373e+00
-4.89226639e-01 -4.55388486e-01 -6.32441223e-01 -8.32184017e-01
-1.23531580e-01 1.09234130e+00 -4.83793795e-01 1.96007416e-01
-1.01199138e+00 -7.15346932e-01 9.08149362e-01 5.97912729e-01
1.80253893e-01 -1.42179191e+00 -1.67549681e-03 3.01294625e-01
-5.08279681e-01 -1.31513131e+00 -4.35073107e-01 9.16680813e-01
-8.27950776e-01 -6.47780716e-01 -1.93639830e-01 -1.32699752e+00
5.38888991e-01 -1.34219810e-01 1.35562694e+00 -1.92931518e-01
9.71042216e-02 7.76727870e-02 -4.55256611e-01 -3.25119436e-01
-6.45179391e-01 3.83749664e-01 -2.60750130e-02 -3.92002046e-01
6.59380198e-01 -8.00519943e-01 8.16414207e-02 -3.17114502e-01
-8.85212541e-01 -1.58057734e-01 8.10727715e-01 5.23242891e-01
6.47017896e-01 -4.06975597e-01 8.57046172e-02 -1.17728353e+00
5.98313630e-01 -4.39973205e-01 -4.26117659e-01 1.85879171e-01
-1.74302816e-01 1.13014318e-01 4.67984855e-01 -6.30270600e-01
-1.22973907e+00 1.99550673e-01 -4.70528871e-01 -4.78326008e-02
-6.98435545e-01 7.36942172e-01 -5.42372823e-01 4.73152161e-01
5.46238184e-01 -2.08141848e-01 -4.37748939e-01 -9.98537779e-01
7.52845347e-01 5.95793843e-01 5.85812628e-01 -6.44001126e-01
7.29178369e-01 3.26620162e-01 -4.00233597e-01 -7.46583998e-01
-1.04265702e+00 -5.92100918e-01 -1.04255223e+00 3.34005326e-01
1.31435966e+00 -9.98676419e-01 -2.30236292e-01 1.90389350e-01
-1.50032890e+00 -6.06748581e-01 -4.69957978e-01 7.31829047e-01
-2.81320453e-01 4.47806478e-01 -1.30873656e+00 -2.48436630e-01
-3.74323130e-01 -7.80150592e-01 7.76897430e-01 8.98555480e-03
-5.72574019e-01 -1.34212935e+00 7.91766793e-02 2.04790249e-01
1.05670886e-02 -8.76915231e-02 1.17349529e+00 -1.15449595e+00
-8.88135210e-02 1.06992805e-02 -9.14203972e-02 4.40295547e-01
3.89819622e-01 -6.81542680e-02 -6.64426446e-01 -1.12673886e-01
-1.79164514e-01 -4.23759192e-01 9.25386906e-01 3.17496061e-01
3.99576157e-01 -3.97271037e-01 -2.95436382e-01 8.24107230e-01
1.36980093e+00 1.36380836e-01 3.22486609e-01 3.41838956e-01
6.63081348e-01 3.87913674e-01 2.88779199e-01 4.33441736e-02
4.83651817e-01 3.72682393e-01 1.01998247e-01 -2.95985546e-02
-4.84557450e-01 -1.01433881e-01 6.96164370e-01 1.27977240e+00
2.63368249e-01 -3.54516715e-01 -1.15307164e+00 7.89271772e-01
-1.38412797e+00 -6.72181964e-01 -3.22851628e-01 1.84354126e+00
1.50141203e+00 1.74842775e-01 -2.70103782e-01 -3.42315704e-01
7.72967100e-01 2.22724020e-01 1.14931762e-01 -8.21628034e-01
-3.31345648e-01 6.96468890e-01 6.75586164e-01 7.67992616e-01
-1.39185071e+00 1.83854842e+00 6.42806292e+00 1.08545041e+00
-1.01689041e+00 5.94557047e-01 1.76203355e-01 1.96188182e-01
-3.22895110e-01 1.84325770e-01 -1.20681870e+00 -3.41320746e-02
1.04658580e+00 -4.07449491e-02 4.51988399e-01 5.88636160e-01
-8.86195060e-03 -9.16493386e-02 -9.20676172e-01 5.37566364e-01
-3.81683670e-02 -9.36694086e-01 1.76076949e-01 -3.15503776e-02
4.64054674e-01 5.13778806e-01 -1.83872506e-01 4.25635576e-01
9.83330905e-01 -6.68382406e-01 7.66842306e-01 -1.20933708e-02
1.04745579e+00 -5.71734369e-01 7.38685787e-01 -4.22042906e-02
-1.26259136e+00 4.14800018e-01 -2.14701459e-01 -4.49020416e-02
5.54677486e-01 4.55057323e-01 -9.24208283e-01 4.78438377e-01
4.83284950e-01 5.26013672e-01 -6.18360877e-01 6.89952612e-01
-8.71968567e-01 1.07800186e+00 -5.77505469e-01 6.04939982e-02
2.84871668e-01 -2.65843064e-01 5.97166538e-01 1.92220092e+00
1.02606647e-01 1.06771044e-01 2.96122342e-01 4.84939486e-01
-2.80701756e-01 5.51831126e-01 -4.73526567e-01 -3.79277468e-01
4.99518365e-01 1.43401361e+00 -1.09722066e+00 -4.96662468e-01
-5.35724103e-01 8.49764049e-01 7.71768630e-01 3.49695623e-01
-9.03340459e-01 -3.06851625e-01 7.79712558e-01 -1.14650548e-01
4.21626240e-01 -7.15831876e-01 -3.71445447e-01 -1.16881490e+00
-1.93113998e-01 -7.09547758e-01 6.03394687e-01 -4.89334375e-01
-1.28224695e+00 7.43230343e-01 1.93644479e-01 -5.94879270e-01
-7.18653798e-02 -7.68709421e-01 -4.13131088e-01 6.51771367e-01
-1.00086033e+00 -1.42947006e+00 2.40023196e-01 5.16683817e-01
5.53521156e-01 7.14164302e-02 1.14156294e+00 4.10380751e-01
-4.95921582e-01 6.07521117e-01 -8.20423737e-02 6.88897252e-01
9.04420257e-01 -1.39547884e+00 6.82908714e-01 8.94157112e-01
6.10288262e-01 7.50289559e-01 7.16855586e-01 -7.81768799e-01
-1.02647722e+00 -1.03733265e+00 1.65607238e+00 -4.81308162e-01
1.15696800e+00 -5.39010346e-01 -8.31818879e-01 1.40221524e+00
7.02308595e-01 -3.42078209e-01 8.40408683e-01 5.67999125e-01
-5.47386289e-01 1.85801387e-01 -7.85530388e-01 7.50062287e-01
1.29375076e+00 -5.84685564e-01 -1.17288589e+00 5.96386909e-01
8.85481298e-01 -3.47397894e-01 -8.19189727e-01 1.00169666e-01
1.95883378e-01 -3.60941678e-01 5.59145451e-01 -7.50063360e-01
-4.73769521e-03 7.15332478e-02 -3.15929651e-01 -1.21896040e+00
-5.15937448e-01 -6.77420735e-01 3.81008625e-01 1.58626795e+00
8.70139241e-01 -6.41480029e-01 5.08919179e-01 1.26980513e-01
-4.66220468e-01 -3.89262050e-01 -1.16392064e+00 -9.76548076e-01
5.14175892e-01 -6.45841837e-01 1.35883644e-01 1.24038625e+00
1.73310339e-01 7.87120104e-01 9.52757075e-02 7.96727166e-02
2.31965616e-01 4.48928215e-04 3.15367669e-01 -9.30164456e-01
-2.29672939e-01 -3.44464153e-01 -1.42043844e-01 -1.01862395e+00
7.04315066e-01 -1.44118547e+00 8.92046466e-02 -1.84666181e+00
-5.16493097e-02 -2.78994292e-01 -6.00407682e-02 1.10236156e+00
-7.44281858e-02 2.72979975e-01 6.33548796e-02 3.35542373e-02
-5.02014637e-01 2.25258395e-01 7.47911394e-01 3.29377167e-02
-2.53312349e-01 -1.65843353e-01 -5.85388362e-01 8.80509734e-01
1.25179219e+00 -9.46762204e-01 2.77780831e-01 -8.97627652e-01
1.80248335e-01 -1.75928816e-01 -2.61131346e-01 -8.46022725e-01
-9.31791216e-02 -1.16164006e-01 -2.44059730e-02 -4.71315414e-01
6.12880290e-01 -5.86627245e-01 4.18436602e-02 2.60145128e-01
-1.19805515e-01 5.13445139e-01 3.23879242e-01 -5.24063446e-02
-1.20414488e-01 -3.51268500e-01 7.38400221e-01 -3.49898458e-01
-3.93207073e-01 8.28290656e-02 -9.27664280e-01 4.16326165e-01
6.35500908e-01 6.71377331e-02 -5.21539040e-02 -1.34538159e-01
-1.24404037e+00 -9.86084640e-02 3.01711768e-01 2.78239489e-01
1.16938487e-01 -1.10531390e+00 -6.75603867e-01 -1.06963620e-01
-2.76363313e-01 -2.16385543e-01 -3.18217844e-01 8.14202189e-01
-4.76750970e-01 3.64640862e-01 -8.51596221e-02 -1.12314813e-01
-1.39704192e+00 5.77865720e-01 1.36310726e-01 -6.15478992e-01
-5.00567019e-01 1.19778502e+00 1.14038043e-01 -5.43323874e-01
8.88785347e-02 -6.47273242e-01 -9.90291163e-02 4.00880218e-01
-1.93617139e-02 -1.98880896e-01 1.63322315e-01 -9.12259400e-01
-5.21691859e-01 2.20679700e-01 -1.65212706e-01 -3.50428462e-01
1.30502164e+00 -1.48139792e-02 -3.81719261e-01 7.20752835e-01
1.16446888e+00 5.08735120e-01 -6.61729634e-01 -3.60473782e-01
3.01562786e-01 2.32673973e-01 -2.06923127e-01 -6.85980320e-01
-8.15142393e-01 7.89593756e-01 8.75644460e-02 2.41094623e-02
8.36664557e-01 4.71622556e-01 8.38003457e-01 6.30043566e-01
2.17628062e-01 -1.24339747e+00 -3.64603430e-01 1.04034901e+00
5.93284726e-01 -8.46362054e-01 8.15477893e-02 -6.66661203e-01
-6.48498476e-01 9.57321346e-01 2.95827985e-01 -1.86398089e-01
8.91784847e-01 7.22971618e-01 4.91201699e-01 -1.20380521e-01
-6.69404745e-01 -6.38433933e-01 -7.71903023e-02 5.19316435e-01
1.01742721e+00 5.01174390e-01 -8.42091024e-01 8.07092547e-01
-5.85682034e-01 -6.12773418e-01 2.61933923e-01 9.26360130e-01
-2.52766490e-01 -1.49857533e+00 -1.04029857e-01 2.95895100e-01
-1.08827078e+00 -7.64582157e-01 -6.11897767e-01 1.17020416e+00
3.02955806e-01 8.23084772e-01 2.05880299e-01 -1.17624544e-01
2.52728015e-01 6.78523481e-01 5.01734018e-01 -1.09678817e+00
-9.35642958e-01 2.78749436e-01 8.03430319e-01 -3.27348441e-01
-4.86077189e-01 -1.01450324e+00 -1.64957738e+00 -2.52739221e-01
-3.47123623e-01 3.05631816e-01 5.87528110e-01 9.73967493e-01
2.80444305e-02 3.53902221e-01 -5.35319895e-02 -7.89823055e-01
-3.77176911e-01 -1.34456944e+00 -5.19127727e-01 4.44762141e-01
-2.45593488e-01 -3.90854061e-01 -4.44595903e-01 2.74384737e-01] | [10.452674865722656, 10.068629264831543] |
cc7bbf74-fb6a-4410-9501-f636074ef708 | a-new-persian-text-summarization-approach | null | null | https://jipm.irandoc.ac.ir/browse.php?a_id=2842&sid=1&slc_lang=en | https://jipm.irandoc.ac.ir/article-1-2842-en.pdf | A New Persian Text Summarization Approach Based on Natural Language Processing and Graph Similarity | Abstract: A significant amount of available information is stored in textual
databases which contain a large collection of documents from different
sources (such as news, articles, books, emails and web pages). The
increasing visibility and importance of this class of information motivates
us to work on having better automatic evaluation tools for textual
resources.
The automatic summarization of text is one of the ways to prevent
the waste of users’ time. The extractive text summarization consists
of the extraction of the more important sentences with the purpose of
shortening input text while maintaining the topics covered and the
subjects discussed.
In this paper, we have tried to improve the accuracy of the extracted
summaries by combining natural language processing and text mining
techniques. By modifying the mentioned algorithms and sentence
scoring measures, accuracy is increased as compared to the previously
used techniques.
Part of speech tagging is used for calculating coefficient of words’
importance. Using this approach will in turn help us with picking the more
meaningful words and phrases that will result in better accuracy of the
system.
Graph similarity’s methods are used to select sentences. Changing weight
of the selected sentences in each step leads to solve the redundancy
problem. | ['Azadeh Mohebi', 'Abbas Ahmadi', 'Tayyebeh Hosseinikhah'] | 2017-02-01 | null | null | null | iranian-journal-of-onformation-processing-and | ['graph-similarity', 'extractive-document-summarization'] | ['graphs', 'natural-language-processing'] | [ 3.01985502e-01 2.02574432e-01 -1.78716972e-01 -1.03502296e-01
-5.25290072e-01 -5.18531442e-01 5.45464694e-01 1.19115615e+00
-5.86804211e-01 1.12389684e+00 8.80436420e-01 -2.07355712e-02
-3.38564932e-01 -8.22681248e-01 9.07797366e-02 -3.71529728e-01
1.23092219e-01 4.18977946e-01 6.32321775e-01 -5.18207312e-01
1.12696266e+00 3.59973788e-01 -1.68022358e+00 2.61430472e-01
1.14348996e+00 4.35425669e-01 5.64722359e-01 7.40177155e-01
-9.55859184e-01 6.97276592e-01 -8.13790321e-01 -2.09294215e-01
-1.98497221e-01 -6.84072316e-01 -1.06971419e+00 1.47791758e-01
-2.59730928e-02 8.57780948e-02 3.24463278e-01 9.72743750e-01
2.93278605e-01 2.77779251e-01 7.08767951e-01 -5.99135101e-01
5.33235781e-02 7.97434449e-01 -7.90187836e-01 4.68330473e-01
7.89453208e-01 -6.58285975e-01 9.94773448e-01 -6.00193441e-01
6.86051786e-01 1.20329380e+00 2.46303722e-01 -5.19671403e-02
-6.30179524e-01 -8.40133056e-02 -2.27446079e-01 1.98651403e-01
-1.09377825e+00 -3.40648204e-01 6.63199365e-01 -2.39922151e-01
1.16282833e+00 6.10433996e-01 5.62817454e-01 3.56729209e-01
3.04121435e-01 4.03415710e-01 6.92153752e-01 -1.04168439e+00
9.30557102e-02 5.87244511e-01 7.60231853e-01 5.43512404e-01
7.68853843e-01 -8.89503598e-01 -4.09609914e-01 -2.59614736e-01
-2.81327426e-01 1.20208133e-02 -2.37550691e-01 1.67465240e-01
-8.75670373e-01 8.24042022e-01 -2.36085996e-01 1.07648802e+00
-6.46850646e-01 -4.47309881e-01 8.36960495e-01 9.60330516e-02
5.45852304e-01 6.11246943e-01 -3.12504500e-01 -3.03650439e-01
-1.15786147e+00 2.33945757e-01 9.78941441e-01 7.01607704e-01
6.03968084e-01 -2.59221196e-01 -1.89863995e-01 8.52354348e-01
2.53435522e-01 2.00337380e-01 7.76788890e-01 -4.83695298e-01
7.13732958e-01 1.34711456e+00 8.43704503e-04 -1.39121163e+00
-2.87535191e-01 -3.69510084e-01 -6.28657579e-01 -1.83239028e-01
-3.51318829e-02 -2.18220532e-01 -6.05512619e-01 1.07215965e+00
2.55478323e-01 -7.26418614e-01 1.92161113e-01 2.51128614e-01
7.79978156e-01 1.04478562e+00 9.87271741e-02 -6.88474596e-01
1.65950418e+00 -5.84421515e-01 -9.77095068e-01 2.37348601e-01
5.55497587e-01 -1.37222493e+00 6.12232327e-01 3.89423192e-01
-1.00881982e+00 -4.26542729e-01 -1.10394764e+00 1.06740139e-01
-6.80827081e-01 2.53947556e-01 2.18540564e-01 5.28273880e-01
-9.25524056e-01 7.66970217e-01 -4.30421263e-01 -8.82484198e-01
1.09078281e-01 2.68986940e-01 -2.60844290e-01 1.33533463e-01
-9.85407293e-01 1.06343091e+00 6.81245983e-01 -3.78325552e-01
2.67534882e-01 -1.57301575e-01 -4.74861503e-01 1.32407382e-01
5.22646427e-01 -5.13598144e-01 8.94586623e-01 -9.43590760e-01
-8.91070962e-01 5.62643886e-01 -3.87051791e-01 -4.75149572e-01
9.48965698e-02 -1.37083337e-01 -2.10630134e-01 5.35432577e-01
2.65609622e-01 1.67801693e-01 4.08085972e-01 -8.30575705e-01
-9.73070443e-01 -5.03523350e-01 -3.46081436e-01 3.84206057e-01
-5.22162497e-01 3.40602130e-01 -1.34306952e-01 -4.37084645e-01
1.20658889e-01 -4.91286516e-01 -1.14972122e-01 -9.32287574e-01
-4.46513355e-01 -6.79970205e-01 9.74160790e-01 -1.07377112e+00
1.66126597e+00 -1.65644658e+00 1.53043434e-01 2.80919850e-01
1.96083590e-01 2.96422929e-01 4.43173766e-01 1.16128159e+00
2.19171703e-01 4.34092969e-01 -1.53136536e-01 -3.31179276e-02
-3.24358791e-01 3.64172645e-02 -5.69825843e-02 -1.16641402e-01
-1.42645776e-01 1.87020600e-01 -6.23610735e-01 -1.05032158e+00
1.00491062e-01 8.22890177e-02 -2.26642075e-03 -1.03924893e-01
-6.10420182e-02 2.22988874e-02 -8.41322482e-01 2.78722018e-01
1.93048388e-01 1.96745291e-01 -5.03532924e-02 -1.10062569e-01
-4.53075141e-01 5.66654861e-01 -1.02822924e+00 1.14868486e+00
-2.44014025e-01 6.98069394e-01 -3.46734434e-01 -1.13923681e+00
1.02363515e+00 4.25395250e-01 5.67761481e-01 -4.98590440e-01
4.50664550e-01 1.52381957e-01 -1.21855006e-01 -8.51088107e-01
1.06200922e+00 1.78865314e-01 -3.93208154e-02 5.26218414e-01
-9.38466862e-02 1.76439881e-02 9.70281482e-01 7.86385000e-01
8.07942271e-01 -1.70797735e-01 9.36199129e-01 -2.81958044e-01
8.55405986e-01 4.01222497e-01 6.72510266e-02 4.13289607e-01
1.52026325e-01 2.35522483e-02 5.20269752e-01 -7.09774196e-02
-1.19496965e+00 -1.26838878e-01 1.34272635e-01 7.67978966e-01
-3.04224312e-01 -5.77734828e-01 -7.84469545e-01 -3.54648948e-01
-2.60761827e-01 8.98019135e-01 -3.82369131e-01 9.46112946e-02
-5.32583475e-01 -4.90301877e-01 -1.20810617e-03 -9.68294516e-02
4.19092804e-01 -1.18002117e+00 -7.37305760e-01 3.23817998e-01
-3.15084815e-01 -6.91668868e-01 -1.68171078e-01 5.47217485e-03
-1.19869828e+00 -1.03588593e+00 -6.09713197e-01 -6.84843659e-01
7.73912430e-01 3.91175359e-01 7.93088436e-01 3.00285608e-01
-2.00723395e-01 5.62001169e-02 -1.02552807e+00 -7.41729081e-01
-7.68244982e-01 5.28971553e-01 -3.20445538e-01 -4.28006649e-01
3.45200777e-01 -3.89064372e-01 -2.35595867e-01 -4.46756303e-01
-1.00843751e+00 -4.37664501e-02 6.46456838e-01 3.65266860e-01
1.43518984e-01 3.28144610e-01 7.92529762e-01 -1.13839710e+00
1.31332278e+00 -3.87785345e-01 -3.40508908e-01 3.54660153e-01
-8.21641684e-01 5.40104151e-01 5.96494317e-01 -3.29444818e-02
-9.44069088e-01 -3.80949169e-01 -3.86989824e-02 7.10412800e-01
-4.69041392e-02 6.43517554e-01 7.32997879e-02 1.72232658e-01
4.97635543e-01 3.65168452e-01 -8.54844227e-02 -4.78113681e-01
-1.48050129e-01 1.00424266e+00 -4.20068502e-02 -6.67451173e-02
3.29963267e-01 -9.69171375e-02 4.31909934e-02 -1.14864123e+00
-6.29430354e-01 -1.06938207e+00 -5.57951093e-01 -5.03774226e-01
7.55704761e-01 -1.60610780e-01 -3.41354966e-01 -1.62168667e-01
-1.21217167e+00 7.72481680e-01 -1.07415728e-01 5.59573472e-01
1.00480905e-02 8.68221700e-01 -2.05399822e-02 -1.11292410e+00
-9.06774402e-01 -6.92109227e-01 6.63638055e-01 4.96625304e-01
-6.37913704e-01 -7.01534629e-01 1.86160296e-01 3.37357283e-01
3.27670157e-01 2.35854208e-01 1.13791859e+00 -1.04741967e+00
-1.48207068e-01 -5.95374942e-01 9.96327922e-02 3.48620474e-01
3.77186000e-01 2.03834444e-01 -3.04812759e-01 2.17779279e-01
1.04578800e-01 1.59968778e-01 9.67781126e-01 3.31141025e-01
5.63772380e-01 -6.30015969e-01 -4.94159371e-01 -4.45708364e-01
1.50975096e+00 5.65204680e-01 7.05682635e-01 4.65756476e-01
3.28497648e-01 8.41801226e-01 7.98750699e-01 5.14822483e-01
7.02902973e-02 3.99826437e-01 1.05185539e-01 3.06625575e-01
1.11912534e-01 -3.26698236e-02 1.19498894e-01 1.20403862e+00
-8.21608528e-02 -4.56486642e-01 -6.33846223e-01 6.50739551e-01
-1.75190556e+00 -1.31160438e+00 -5.03451407e-01 2.28069615e+00
7.06283748e-01 3.61368328e-01 2.96056151e-01 6.78808033e-01
7.28347182e-01 1.34049997e-01 2.33678728e-01 -9.37545717e-01
1.25706062e-01 7.74874240e-02 5.13297677e-01 5.87663114e-01
-6.46507740e-01 6.60634160e-01 5.00797844e+00 7.09903717e-01
-7.88507700e-01 -2.13140398e-01 3.16578925e-01 2.58045673e-01
-2.13302150e-01 2.00636432e-01 -8.74155343e-01 5.64777374e-01
9.84797537e-01 -6.84186161e-01 -8.96109045e-02 6.14050806e-01
6.37963235e-01 -1.09052134e+00 -3.62319201e-01 4.38026935e-01
4.10368353e-01 -1.24226856e+00 2.52754420e-01 -8.41767620e-03
6.50075197e-01 -4.39484626e-01 -6.87926829e-01 -4.01990451e-02
-1.70762345e-01 -3.70987028e-01 3.48995864e-01 6.30690157e-01
-7.32503459e-02 -1.00150812e+00 1.27287626e+00 5.02071679e-01
-9.30811107e-01 1.98813826e-01 -4.40218180e-01 -1.79453075e-01
1.26960546e-01 7.04020858e-01 -1.28351581e+00 7.21558273e-01
3.73974293e-01 2.80492574e-01 -6.20039105e-01 1.15347457e+00
5.44903167e-02 6.45283222e-01 -2.22508758e-01 -9.51277196e-01
2.69492954e-01 -3.98534000e-01 7.15298057e-01 1.38713431e+00
3.42160314e-01 2.27962956e-01 -7.66021833e-02 1.43025294e-01
5.56372181e-02 9.78672683e-01 -7.72748172e-01 -1.88812494e-01
3.80417854e-01 1.24918699e+00 -1.16467142e+00 -7.10726321e-01
-2.36748531e-01 7.01691389e-01 -4.18191031e-02 -1.65686920e-01
-1.15652397e-01 -9.31207478e-01 -2.46365353e-01 4.30806339e-01
6.05824441e-02 -2.69761682e-01 -2.12555185e-01 -5.51138997e-01
1.69340372e-02 -6.71513677e-01 3.64028871e-01 -5.47445595e-01
-6.06284976e-01 6.28397882e-01 3.46306473e-01 -1.02044606e+00
-2.99279660e-01 -9.67639610e-02 -7.29123712e-01 8.24994862e-01
-1.02921045e+00 -3.98636132e-01 -8.06566700e-02 1.26932159e-01
1.08234751e+00 -2.62322515e-01 7.02164173e-01 8.48814920e-02
-2.75241703e-01 -1.34118050e-01 1.20854333e-01 -8.65400210e-02
6.06175363e-01 -1.29679072e+00 -7.01086521e-02 9.54229474e-01
-4.38551269e-02 4.88996148e-01 1.12141252e+00 -9.21588480e-01
-8.90286326e-01 -4.46104825e-01 1.67440557e+00 6.48159906e-02
5.63397825e-01 2.74083167e-01 -8.56382012e-01 -4.52138670e-02
7.74501801e-01 -1.17884147e+00 7.76525080e-01 -7.43367523e-02
3.67101878e-01 -1.76507279e-01 -1.05552983e+00 5.75764179e-01
3.24419767e-01 1.41161844e-01 -1.19581687e+00 4.25523996e-01
5.34977853e-01 1.49319127e-01 -4.11947995e-01 -1.66648224e-01
2.42253393e-01 -7.90313780e-01 4.76377666e-01 -5.06749272e-01
5.55527866e-01 -5.54550067e-02 2.16380775e-01 -1.29536831e+00
-1.04076630e-02 -4.06999916e-01 3.09873164e-01 1.64076269e+00
6.43085957e-01 -4.44272190e-01 4.58914161e-01 2.63510048e-01
-7.85692409e-02 -5.46565175e-01 -4.88117248e-01 -1.24871731e-01
-5.78671873e-01 1.15941711e-01 8.45169798e-02 6.92590237e-01
6.08773708e-01 8.37131262e-01 -1.66056350e-01 -4.46478635e-01
5.17136276e-01 -2.71825250e-02 5.51071346e-01 -1.57703054e+00
2.36953169e-01 -4.13270533e-01 -3.73851240e-01 -2.82989979e-01
-3.39436203e-01 -6.12747133e-01 -2.25461379e-01 -2.21569371e+00
4.15128976e-01 5.20782880e-02 2.56706595e-01 1.50096938e-01
-2.11397052e-01 -3.85615438e-01 1.22264656e-03 2.83574164e-01
-5.71486473e-01 1.45857781e-01 1.18250096e+00 9.91492495e-02
-4.16362315e-01 1.56741530e-01 -9.58687663e-01 6.01445556e-01
9.74653900e-01 -7.17428982e-01 -4.53398734e-01 9.05267224e-02
4.29991812e-01 6.11824766e-02 -4.28777397e-01 -8.63994062e-01
3.51642102e-01 -2.45012790e-01 1.89117283e-01 -1.08522058e+00
-1.98665962e-01 -8.09498429e-01 -2.76311524e-02 5.30641973e-01
-4.34349805e-01 4.95464206e-01 4.85802926e-02 3.99003237e-01
-3.84622008e-01 -1.00453627e+00 4.81495708e-01 -3.80113304e-01
-4.48879987e-01 -3.99682641e-01 -6.96862578e-01 -1.53006777e-01
9.22476470e-01 -5.53854883e-01 -2.10950911e-01 -4.99174803e-01
-2.99059451e-01 1.97866306e-01 8.61323625e-02 2.71644801e-01
4.00445908e-01 -6.05293989e-01 -7.25164413e-01 -3.77493173e-01
-6.30168840e-02 -3.05369705e-01 4.27184924e-02 8.81111085e-01
-8.51094544e-01 7.21000135e-01 -3.44055414e-01 1.18012521e-02
-1.95535481e+00 2.52324343e-01 -4.22512591e-01 -6.89111590e-01
-5.58651984e-01 2.81477034e-01 -6.68507516e-01 5.33236742e-01
1.07696749e-01 -3.74720901e-01 -1.21386850e+00 5.85458338e-01
6.19090497e-01 7.03447163e-01 2.73838550e-01 -6.09218121e-01
-1.67533010e-01 4.18130904e-01 -3.56840521e-01 -2.27158248e-01
1.50758648e+00 -3.70466530e-01 -5.23156166e-01 5.27918816e-01
1.05985367e+00 6.34141386e-01 -3.73444743e-02 1.49260521e-01
6.24099791e-01 -2.35700667e-01 8.25716257e-02 -6.41839623e-01
-3.17372322e-01 4.93808359e-01 1.16613626e-01 8.36400807e-01
1.19341207e+00 -3.51093560e-02 6.70658112e-01 5.77822804e-01
1.23802751e-01 -1.49219382e+00 -2.04852283e-01 3.79919678e-01
8.00011814e-01 -1.11198342e+00 5.21185398e-01 -3.62747699e-01
-6.71176970e-01 1.34258032e+00 7.39144441e-03 8.26622993e-02
3.45211357e-01 8.17247331e-02 -2.49644652e-01 -3.16120684e-01
-6.02861702e-01 -3.51846576e-01 3.17601681e-01 1.80480510e-01
8.29555452e-01 -1.47587597e-01 -1.65121365e+00 3.24944228e-01
-3.02054197e-01 -2.35829011e-01 8.48995984e-01 1.03771758e+00
-1.39994907e+00 -1.34479928e+00 -5.87532997e-01 8.80865872e-01
-8.77980471e-01 2.98215356e-02 -9.41817701e-01 7.22227871e-01
-1.67231724e-01 1.32806528e+00 -2.48013213e-01 -2.16222610e-02
2.99444288e-01 2.30303556e-01 2.66096979e-01 -7.09051311e-01
-7.19807267e-01 1.12369262e-01 5.17694354e-01 1.72444984e-01
-5.94211519e-01 -6.84013784e-01 -1.43670464e+00 -2.31286272e-01
-4.97750938e-01 9.53278542e-01 1.20195401e+00 1.10739398e+00
8.51732641e-02 6.46693707e-01 6.48365021e-01 -4.65567917e-01
-3.12098116e-01 -1.29057860e+00 -3.09627175e-01 3.65271866e-01
-1.88464131e-02 -3.98720622e-01 -1.82481751e-01 6.87891394e-02] | [12.302277565002441, 9.53227424621582] |
1fb3311b-43c5-4efd-9b5b-10de519f8e64 | k-means-for-unsupervised-instance | null | null | https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4251338 | https://papers.ssrn.com/sol3/Delivery.cfm/456a55bb-5b72-49b6-be69-b5f39b85c44c-MECA.pdf?abstractid=4251338&mirid=1 | K-means for unsupervised instance segmentation using a self-supervised transformer | Instance segmentation is a fundamental task in computer vision that assigns every pixel to an
appropriate class and localizes objects into bounding boxes. However, collecting pixel-level segmentation labels is more resource- and time-consuming than collecting classification and detection
labels. Herein, we present a novel approach, iterative mask refinement using a self-supervised
transformer (IMST), which performs class agnostic unsupervised instance segmentation using simple K-means clustering and a self-supervised vision transformer. IMST generates pseudo-ground-truth labels that can be used to train an off-the-shelf instance segmentation model. The pseudo labels
demonstrate improved performance on multiple datasets. The instance segmentation model trained
on the pseudo labels outperforms state-of-the-art unsupervised instance segmentation methods on
COCO20k (+4.0 average precision (AP)) and COCO val2017(+2.6 AP) without modifications to
the training loss or architecture. We demonstrate that our method can be extended to tasks such as
single/multiple object discovery and supervised fine-tuning instance segmentation while outperforming previous methods. | ['Lee HongChul', 'Lee MinYoung', 'Park JaeEon', 'Lim SeongTaek'] | 2022-10-04 | null | null | null | pattern-recognition-2022-10 | ['single-object-discovery', 'object-discovery'] | ['computer-vision', 'computer-vision'] | [ 7.70776033e-01 4.20428783e-01 -4.15144622e-01 -6.51470900e-01
-1.23008239e+00 -8.43949616e-01 4.48062867e-01 1.86611339e-01
-6.17791116e-01 4.52422917e-01 -6.54839754e-01 -1.45430677e-02
1.90912113e-01 -5.97757101e-01 -8.97227764e-01 -6.67709053e-01
2.63585746e-01 1.15338266e+00 7.39357650e-01 5.94822049e-01
2.87611216e-01 2.83223271e-01 -1.59328902e+00 4.17456388e-01
9.29391861e-01 1.28339255e+00 3.00650597e-01 6.04554474e-01
-3.25298190e-01 4.13438141e-01 -5.88549972e-01 -2.26191774e-01
4.05434638e-01 -1.77595660e-01 -1.20796585e+00 5.96966863e-01
8.10863554e-01 7.93114156e-02 4.54046071e-01 1.08184004e+00
7.25978762e-02 -7.39073828e-02 8.53540838e-01 -1.09945655e+00
-4.06321615e-01 5.85334003e-01 -8.67390513e-01 6.41611367e-02
-4.61935133e-01 2.01870292e-01 9.11940157e-01 -7.46903956e-01
5.16354501e-01 9.28716838e-01 8.04397166e-01 5.57205856e-01
-1.63144755e+00 -6.10658109e-01 3.27395767e-01 -2.81710532e-02
-1.40984547e+00 -2.07165763e-01 4.92330134e-01 -5.66182196e-01
8.06645930e-01 1.82813525e-01 4.65312928e-01 5.29613435e-01
-4.61007744e-01 1.21026695e+00 1.33439732e+00 -4.52225268e-01
5.07954895e-01 4.20327216e-01 4.36466962e-01 8.46834242e-01
9.50323716e-02 -3.12402010e-01 -1.47019729e-01 2.71226555e-01
7.17000961e-01 -2.07854629e-01 1.33262411e-01 -3.98472160e-01
-1.24503589e+00 5.45850515e-01 5.76502264e-01 6.23789504e-02
-2.17492580e-01 1.46741688e-01 3.65758479e-01 -1.84830308e-01
6.45667911e-01 6.57429636e-01 -9.50178206e-01 1.57431558e-01
-1.50837004e+00 -1.34194687e-01 6.66022599e-01 1.09159195e+00
1.12282670e+00 -3.48741204e-01 -3.94570589e-01 1.00218821e+00
2.21004352e-01 4.03380901e-01 3.05417955e-01 -1.36945653e+00
1.41199335e-01 9.03803349e-01 -5.07019609e-02 -4.30277437e-01
-3.92017126e-01 -7.56103158e-01 -4.80578542e-01 1.91426903e-01
6.69495344e-01 9.77608655e-03 -1.65487456e+00 1.34100556e+00
5.29343069e-01 4.09567624e-01 -1.27811551e-01 8.01857233e-01
8.91606748e-01 3.99837285e-01 2.56599247e-01 -8.40690136e-02
1.34898281e+00 -1.42673671e+00 -4.03300494e-01 -6.59888268e-01
4.64110762e-01 -5.23086071e-01 1.07157874e+00 3.64111692e-01
-9.25392628e-01 -6.56634986e-01 -8.45101655e-01 1.80537440e-03
-4.75866139e-01 4.41818476e-01 6.30111992e-01 7.02873707e-01
-8.98490369e-01 4.38755274e-01 -9.88401473e-01 -1.89903826e-01
1.10787988e+00 5.57593226e-01 -1.55717850e-01 8.05518106e-02
-4.93861169e-01 4.32466179e-01 6.55595243e-01 -1.19059168e-01
-9.73796010e-01 -1.01377344e+00 -7.73432732e-01 -2.92036593e-01
6.86697006e-01 -3.36783379e-01 1.42870450e+00 -1.29015362e+00
-1.27268016e+00 1.45314121e+00 -1.63756207e-01 -7.34988570e-01
4.10807163e-01 -5.50581105e-02 1.11376703e-01 2.64883012e-01
6.22397661e-01 1.26867676e+00 1.03825605e+00 -1.49827826e+00
-1.15424538e+00 -5.01119971e-01 -2.92943090e-01 1.14676781e-01
1.23348549e-01 -2.99022466e-01 -1.00346518e+00 -5.13599753e-01
3.85753959e-01 -9.75792825e-01 -4.50047344e-01 -1.00140706e-01
-8.66333246e-01 -2.95703769e-01 9.61899102e-01 -3.20327431e-01
7.29945183e-01 -1.90108931e+00 -2.74053607e-02 2.09352404e-01
7.70341009e-02 3.52146566e-01 7.52357021e-02 -5.16981184e-01
1.29677802e-01 1.22426882e-01 -9.01760578e-01 -6.64501429e-01
-3.21642943e-02 4.13497150e-01 5.40119149e-02 3.31592381e-01
3.88614684e-01 1.07326853e+00 -7.98803627e-01 -8.25532913e-01
3.83604705e-01 2.69010007e-01 -3.88684809e-01 7.15431347e-02
-7.50440419e-01 6.06390834e-01 -2.63011605e-01 8.54009628e-01
5.37150860e-01 -6.14560604e-01 -8.57519507e-02 -1.83214471e-01
-4.68609622e-03 7.56236836e-02 -1.19930089e+00 1.69003749e+00
-1.32321581e-01 4.92715269e-01 5.75906150e-02 -1.12992311e+00
6.35398328e-01 -3.35935280e-02 4.90207553e-01 -6.48931146e-01
5.12503535e-02 1.45342618e-01 -4.91061777e-01 -2.92870581e-01
9.31458399e-02 7.97851831e-02 -7.48142973e-02 3.83927047e-01
3.07983041e-01 -4.91628170e-01 4.86420095e-01 8.35993066e-02
8.30244005e-01 3.05899739e-01 -4.51359991e-03 -3.44732821e-01
4.03003722e-01 6.20099127e-01 6.64461493e-01 8.52525890e-01
-2.65665352e-01 9.25672472e-01 2.62096286e-01 -2.51501828e-01
-8.21717858e-01 -9.50172842e-01 -5.00107586e-01 1.14760697e+00
1.66450113e-01 -7.27619529e-02 -1.40029156e+00 -1.08403432e+00
9.84528102e-03 5.83592653e-01 -6.98390782e-01 2.88058996e-01
-1.80577695e-01 -1.00715852e+00 5.78769088e-01 6.65547967e-01
7.75935173e-01 -1.26562238e+00 -3.95452589e-01 1.14830732e-01
-1.43091425e-01 -1.37952447e+00 -5.21922231e-01 5.70459485e-01
-9.02463436e-01 -1.17339432e+00 -5.79511642e-01 -1.06320131e+00
1.10022354e+00 -1.25907352e-02 1.31696332e+00 -1.45776272e-01
-6.63360834e-01 4.30839300e-01 -1.55141607e-01 -4.96289879e-01
-2.24781141e-01 3.29557598e-01 -2.78861910e-01 2.29675323e-01
5.13216436e-01 -1.97315052e-01 -6.23899519e-01 4.68570232e-01
-8.45703065e-01 1.73015565e-01 5.62275946e-01 5.08396745e-01
1.51466846e+00 1.72591433e-01 4.52074677e-01 -1.57030213e+00
-4.94566411e-02 -4.03069407e-02 -8.23752999e-01 2.25822583e-01
-7.58602440e-01 -7.31051527e-03 2.87066966e-01 -3.62608314e-01
-1.02646255e+00 8.30143631e-01 8.86311755e-02 -3.54451716e-01
-6.37531936e-01 6.24019280e-02 -2.14583039e-01 -1.64713770e-01
6.93877041e-01 1.12039499e-01 -2.16288581e-01 -5.02022207e-01
5.70534647e-01 6.42126322e-01 9.00355339e-01 -4.00446177e-01
7.13692486e-01 6.45684540e-01 -2.15303123e-01 -5.47577262e-01
-1.57276845e+00 -8.39184046e-01 -1.15882003e+00 -8.88978988e-02
1.23812389e+00 -8.40055645e-01 -2.55289376e-01 6.21353924e-01
-7.53644824e-01 -9.32872534e-01 -4.73426193e-01 -3.77992019e-02
-6.12233460e-01 5.66711165e-02 -5.80744863e-01 -5.54114401e-01
-3.91545326e-01 -1.15382254e+00 1.51120102e+00 4.61076289e-01
-1.68724030e-01 -8.64904404e-01 -2.03925982e-01 9.97484148e-01
1.77748036e-02 1.27078876e-01 4.61376548e-01 -6.80412531e-01
-7.12370932e-01 -1.59196883e-01 -5.55652440e-01 5.22681534e-01
1.13801092e-01 -1.91435181e-02 -1.28476846e+00 -1.70793533e-02
-4.31470007e-01 -4.56264049e-01 1.15726197e+00 6.54578507e-01
1.61393046e+00 -1.96839720e-01 -5.58075845e-01 7.48862445e-01
1.37623858e+00 5.22065572e-02 4.06340152e-01 4.08429623e-01
1.00811541e+00 5.45266569e-01 7.79367387e-01 3.55393961e-02
4.15554613e-01 4.97844875e-01 4.11145210e-01 -4.28003639e-01
-3.16307932e-01 5.28927557e-02 -1.85238302e-01 5.92216179e-02
2.35097587e-01 2.07410440e-01 -1.02412403e+00 8.58461738e-01
-1.84981322e+00 -5.96596718e-01 -2.75989622e-01 2.03578734e+00
1.16722810e+00 4.43676203e-01 2.26778090e-01 2.20210508e-01
7.61339366e-01 -2.98969150e-01 -8.36852491e-01 7.36267269e-02
1.11339957e-01 3.70674968e-01 8.12759876e-01 4.80345339e-01
-1.69115937e+00 1.43296599e+00 5.97063255e+00 9.11859632e-01
-9.15888011e-01 2.15763494e-01 1.29865718e+00 9.77834240e-02
4.03431505e-01 -2.53042430e-01 -9.57908988e-01 4.08835441e-01
5.91699421e-01 5.29904425e-01 1.49758816e-01 1.07084262e+00
-6.34929314e-02 -3.44639808e-01 -1.05196404e+00 8.12282681e-01
-2.00444460e-03 -1.37397814e+00 -2.39063278e-01 -1.92159355e-01
1.12202144e+00 2.47132674e-01 5.01720095e-03 2.85302371e-01
3.56079310e-01 -9.73544896e-01 6.38209939e-01 1.82393402e-01
9.15287673e-01 -5.73156655e-01 5.30648232e-01 3.95028353e-01
-1.05994320e+00 1.17822580e-01 -1.40249535e-01 3.28954548e-01
-1.13323219e-01 7.70418763e-01 -9.62514341e-01 9.71227437e-02
9.97878909e-01 6.37632668e-01 -7.89807081e-01 1.14040446e+00
-3.47330064e-01 9.79824185e-01 -5.22193432e-01 4.26126391e-01
3.71954203e-01 -1.47201195e-01 2.01360062e-01 1.45870888e+00
-3.66084605e-01 3.99692208e-02 5.94244242e-01 9.82688427e-01
-2.00766876e-01 -1.58261389e-01 9.86456200e-02 9.85265449e-02
3.20389062e-01 1.56090939e+00 -1.58817160e+00 -6.29901826e-01
-5.24505302e-02 1.22056174e+00 2.85792053e-01 3.04209828e-01
-7.06222057e-01 -1.42399907e-01 3.79436135e-01 1.31464854e-01
6.02839768e-01 1.38798788e-01 -8.33025873e-01 -6.93091452e-01
-7.16730580e-02 -5.46739042e-01 3.64434630e-01 -4.85104144e-01
-1.25871265e+00 4.49183971e-01 -1.14546187e-01 -8.09086084e-01
3.38085815e-02 -7.18236804e-01 -3.13417196e-01 4.35230136e-01
-1.62612522e+00 -1.31599677e+00 -4.19718742e-01 2.92002082e-01
7.92107403e-01 1.27389237e-01 6.58358395e-01 2.24307731e-01
-8.14453423e-01 5.02780080e-01 -1.37623072e-01 3.83636028e-01
5.61623394e-01 -1.69455171e+00 3.26209098e-01 7.27125704e-01
4.09599572e-01 1.68747678e-01 4.59675372e-01 -5.88394701e-01
-8.39144230e-01 -1.78860664e+00 4.56953734e-01 -6.72434747e-01
3.93870711e-01 -3.34211230e-01 -8.19591701e-01 7.79736459e-01
-1.81630582e-01 5.78683257e-01 5.67203999e-01 -9.60036665e-02
-3.75447989e-01 -1.59348339e-01 -1.49309599e+00 2.43054599e-01
9.47429955e-01 -3.15810472e-01 -3.32047582e-01 6.36687458e-01
7.93807983e-01 -5.86137831e-01 -7.04311967e-01 5.05030274e-01
1.61686629e-01 -6.79443657e-01 1.04408836e+00 -2.51325071e-01
2.14384407e-01 -6.25617921e-01 6.97124898e-02 -8.94262612e-01
-2.06712529e-01 -3.70668501e-01 5.93827963e-02 1.34469497e+00
9.25984383e-01 -2.65071422e-01 1.26607943e+00 7.37284958e-01
-1.47963703e-01 -6.99141562e-01 -7.80983627e-01 -5.90074658e-01
-1.50610030e-01 -6.88589692e-01 2.90461123e-01 9.94135499e-01
-5.61347067e-01 2.92006969e-01 3.58520240e-01 4.74337578e-01
1.13031256e+00 2.15424627e-01 6.15041912e-01 -1.33876765e+00
-2.61194348e-01 -4.46448892e-01 -3.27440441e-01 -9.74538803e-01
2.57898629e-01 -9.93036628e-01 3.94504488e-01 -1.70779526e+00
2.78898925e-01 -9.02278662e-01 -3.58300030e-01 9.67671096e-01
-2.20170990e-01 1.06284022e+00 -5.04247732e-02 2.77246654e-01
-1.17059338e+00 1.81596354e-02 1.07460749e+00 -4.46202904e-01
-3.84733647e-01 2.39652246e-01 -6.18742228e-01 7.86566615e-01
7.72318423e-01 -5.82840085e-01 -3.46986055e-01 -3.65253121e-01
-2.90305972e-01 -4.07401413e-01 4.06352341e-01 -1.20828187e+00
2.77133673e-01 7.04023056e-03 4.21909928e-01 -7.70701826e-01
1.89794376e-01 -8.51074278e-01 -2.09874421e-01 1.66583002e-01
-3.27550560e-01 -6.81985676e-01 3.48828584e-01 5.86702049e-01
-4.69412655e-02 -3.22339177e-01 1.03009462e+00 -4.19493198e-01
-9.98553097e-01 3.71272504e-01 -1.84521109e-01 2.90903777e-01
1.20081234e+00 -4.89380240e-01 -1.35292634e-01 3.25465560e-01
-9.38382685e-01 4.61288452e-01 4.46008623e-01 1.99938297e-01
2.98572212e-01 -7.76422143e-01 -4.24540251e-01 1.27713427e-01
2.65108287e-01 8.74698281e-01 -6.03682585e-02 6.67095840e-01
-4.74753380e-01 2.78117716e-01 1.75038457e-01 -1.23743379e+00
-1.26347649e+00 3.79834205e-01 5.08044720e-01 -2.83036768e-01
-5.81015527e-01 1.19807446e+00 3.01750213e-01 -8.83761227e-01
5.15922308e-01 -2.75093585e-01 -1.41735017e-01 6.27580881e-02
4.27151263e-01 1.03834741e-01 1.88809842e-01 -4.80572402e-01
-3.97794545e-01 7.16037929e-01 -2.04357386e-01 -5.96887358e-02
1.30912733e+00 2.34094411e-02 -1.97915167e-01 4.70668018e-01
9.68610823e-01 -5.99917769e-01 -1.73956943e+00 -3.46537769e-01
3.62675637e-01 -1.11727655e-01 3.03304642e-01 -1.24860716e+00
-1.43242252e+00 5.70466638e-01 6.67339504e-01 6.74063191e-02
1.10404778e+00 3.53804082e-01 5.83903909e-01 3.64898950e-01
3.52090925e-01 -1.36463749e+00 1.07370513e-02 3.83865774e-01
2.86415130e-01 -1.59737599e+00 -2.02797279e-01 -8.76680017e-01
-6.72897577e-01 5.92071056e-01 8.16456079e-01 -1.31108776e-01
5.81582487e-01 4.55988467e-01 2.68643677e-01 -2.46110186e-01
-2.64208466e-01 -5.28021395e-01 5.92656434e-01 7.09159851e-01
2.18892232e-01 1.91975802e-01 2.09647015e-01 5.39186954e-01
-1.58588048e-02 -8.77778605e-02 7.38115907e-02 6.96097612e-01
-5.28838754e-01 -9.71979678e-01 -4.50451046e-01 7.29591131e-01
-5.88214755e-01 -7.06143975e-02 -5.70738256e-01 5.74433863e-01
3.98778349e-01 8.87338877e-01 4.07694519e-01 9.15848836e-03
1.34264931e-01 2.13867962e-01 4.34506625e-01 -1.05228484e+00
-6.18593812e-01 1.90688625e-01 -5.10462001e-02 -5.52768171e-01
-7.03895807e-01 -6.37653887e-01 -1.72806919e+00 4.26939338e-01
-4.07990128e-01 -6.76106960e-02 7.50116050e-01 1.18440318e+00
2.99104601e-01 6.35740221e-01 1.96085185e-01 -1.03772485e+00
1.77716032e-01 -8.55947793e-01 -3.89953047e-01 4.98700649e-01
1.88937619e-01 -4.32749689e-01 -2.25661322e-01 5.11333942e-01] | [9.499966621398926, 0.6302962899208069] |
5ef6f792-f667-4a15-a6a4-c12cf063c936 | moving-towards-a-functional-approach-in-the | null | null | https://aclanthology.org/2022.signlang-1.4 | https://aclanthology.org/2022.signlang-1.4.pdf | Moving towards a Functional Approach in the Flemish Sign Language Dictionary Making Process | This presentation will outline the dictionary making process of the new online Flemish Sign Language dictionary launched in 2019. First some necessary background information is provided, consisting of a brief history of Flemish Sign Language (VGT) lexicography. Then three phases in the development of the renewed dictionary of VGT will be explored: (i) user research, (ii) data-cleaning and modeling, and (iii) innovations. More than wanting to project a report of lexicographic research on a website, the goal was to make the new dictionary a practical, user-friendly reference tool that meets the needs, expectations, and skills of the dictionary users. To gain a better understanding of who the users were, several sources were consulted: the user research by Joni Oyserman (2013), the quantitative data from Google Analytics and VGTC’s own user profiles. Since 2017, VGTC has been using Signbank, an electronic database specifically developed to compile and manage lexicographic data for sign languages. Bringing together all this raw data inadvertently led to inconsistencies and small mistakes, therefore the data had to be manually revised and complemented. The VGT dictionary was mainly formally modernized, but there are also several substantive differences regarding the previous dictionary: for instance, search options were expanded, and semantic categories were added as well as a new feedback feature. In addition, the new website is also structurally different, it is now responsive to all screen sizes. Lastly, possible future innovations will briefly be discussed. VGTC aims to continuously improve both the user-based interface and the content of the current dictionary. Future goals include, but are not limited to, adding definitions and sample sentences (preferably extracted from the corpus), as well as information on the etymology and common use of signs. | ['Hannes De Durpel', 'Thijs Vandamme', 'Sam Verstraete', 'Margot Janssens', 'Caro Brosens'] | null | null | null | null | signlang-lrec-2022-6 | ['instance-search'] | ['computer-vision'] | [-9.07090902e-02 1.04252204e-01 -5.10296285e-01 -1.35612860e-01
-5.89507341e-01 -8.58689070e-01 4.70914423e-01 1.57328218e-01
-7.52024472e-01 4.67248738e-01 1.16724360e+00 -5.19773543e-01
-1.76585764e-01 -1.74876943e-01 1.31172851e-01 -1.96170732e-01
2.93097526e-01 4.19381678e-01 5.01056649e-02 -5.35804033e-01
6.37019992e-01 2.63765931e-01 -1.65679765e+00 7.05050007e-02
8.36403668e-01 4.57902133e-01 1.93916366e-01 3.07127327e-01
-3.94770741e-01 8.39291871e-01 -5.99786103e-01 -6.15596771e-01
2.68231928e-01 -5.54586232e-01 -8.09762895e-01 6.40472099e-02
5.50715446e-01 -5.50754905e-01 -1.58003435e-01 9.05386508e-01
7.91921854e-01 -2.06066057e-01 1.75620764e-01 -8.62153172e-01
-7.82728970e-01 7.36198485e-01 2.41618603e-01 -1.03715464e-01
6.87835515e-01 5.88985741e-01 1.02539539e+00 -8.64457369e-01
1.36534953e+00 9.25563872e-01 6.29917741e-01 6.05494797e-01
-5.30678093e-01 -7.02417135e-01 2.82177955e-01 1.21198833e-01
-1.08693898e+00 -4.46246654e-01 3.80739093e-01 -5.93342185e-01
1.12156594e+00 2.37164527e-01 1.60263658e+00 1.05877018e+00
-2.27156058e-01 6.82805657e-01 1.27723026e+00 -7.90630758e-01
-6.21999567e-03 3.70432794e-01 2.95543261e-02 4.56178755e-01
4.15539354e-01 6.92320243e-02 -6.43018246e-01 -1.41026467e-01
6.98391855e-01 -5.72487056e-01 -2.52246588e-01 -3.33614409e-01
-1.29259145e+00 5.51453769e-01 -2.26212859e-01 9.33247149e-01
-1.50406912e-01 -2.57131875e-01 8.05210292e-01 5.49860656e-01
-4.75905761e-02 4.38284904e-01 -4.65327114e-01 -8.06913197e-01
-6.96840405e-01 3.36007774e-01 1.03501284e+00 9.70392227e-01
-1.25275582e-01 4.37559485e-02 2.83499390e-01 9.42376494e-01
7.72262096e-01 4.49225485e-01 6.62086010e-01 -6.46977663e-01
2.75999904e-01 6.15074813e-01 -2.07647067e-02 -8.47528875e-01
-3.31026137e-01 -2.54956871e-01 -1.70774385e-01 3.39733541e-01
7.51244426e-01 -5.94717786e-02 -1.18790400e+00 1.53404498e+00
-1.24947168e-01 -9.29777443e-01 -2.68537551e-01 1.06961405e+00
1.25480676e+00 -1.54525533e-01 4.63289529e-01 -8.07576105e-02
1.28317738e+00 -2.95147240e-01 -1.05387008e+00 -9.31518376e-02
9.16014791e-01 -1.05170643e+00 1.23182023e+00 5.04843891e-01
-1.11204064e+00 2.65164152e-02 -9.60537851e-01 -2.43699640e-01
-6.47110105e-01 7.64477178e-02 7.19000280e-01 1.01127768e+00
-1.02624357e+00 1.57815069e-01 -6.67801917e-01 -1.15290558e+00
1.54536992e-01 1.22530935e-02 -2.94223547e-01 1.39470128e-02
-1.18013465e+00 1.50489497e+00 3.03531766e-01 -1.06838467e-02
1.73492447e-01 -3.90274286e-01 -8.93058062e-01 -7.52109468e-01
1.66436285e-01 -5.01755774e-01 1.19375181e+00 -9.20070887e-01
-1.45657790e+00 1.35461342e+00 -2.83216164e-02 7.74351358e-02
7.45069861e-01 1.32264182e-01 -7.02685595e-01 -2.27768034e-01
4.86782908e-01 3.13085377e-01 9.42848027e-02 -8.65411818e-01
-1.02392507e+00 -1.37193099e-01 -9.34891850e-02 3.53892177e-01
3.57442535e-02 6.13011122e-01 -5.98675966e-01 -9.81370807e-01
2.36443236e-01 -5.31409383e-01 1.31289929e-01 2.36838564e-01
2.92522132e-01 -1.10811114e-01 2.13597640e-01 -1.22584498e+00
1.67196941e+00 -2.10559201e+00 -1.71312019e-01 4.84509885e-01
1.82659999e-01 2.45078608e-01 3.37299667e-02 8.79842043e-01
1.83126479e-01 2.50245959e-01 -6.90871384e-03 5.42239845e-02
2.79387742e-01 4.89024490e-01 5.64679392e-02 6.75616503e-01
-5.48106670e-01 8.41809511e-01 -1.26305723e+00 -3.55554074e-01
4.17706907e-01 4.59828228e-01 -4.13462013e-01 -6.79081798e-01
1.75292432e-01 5.30079246e-01 1.40341356e-01 1.15260065e+00
2.33900487e-01 3.05915694e-03 4.55998152e-01 -5.25362380e-02
-7.52168477e-01 6.75155938e-01 -1.25838172e+00 1.77276325e+00
-3.04955631e-01 8.77200782e-01 -2.00451352e-02 -2.37200454e-01
7.48201311e-01 5.14867842e-01 5.94609082e-01 -9.57097948e-01
3.33917975e-01 9.90054607e-01 5.43346703e-02 -8.24664116e-01
5.40881395e-01 -2.18314901e-01 -6.02590591e-02 5.04485369e-01
3.40081193e-02 2.21214369e-02 6.33846879e-01 1.26903161e-01
7.32170939e-01 2.77928978e-01 5.98849833e-01 -1.50722638e-01
1.01097047e-01 5.39632916e-01 4.67580557e-01 4.86444950e-01
-2.13525534e-01 2.44517848e-01 5.34656756e-02 -1.59037948e-01
-9.98251855e-01 -6.67142391e-01 -3.70561659e-01 7.92742252e-01
-6.57649040e-02 -8.88499498e-01 -4.18536574e-01 -3.29599679e-01
-4.54582041e-04 7.55007148e-01 -3.09848517e-01 2.84898221e-01
-6.74458623e-01 -2.34370261e-01 6.21907413e-01 4.28769410e-01
4.62133080e-01 -1.18921995e+00 -1.08347845e+00 2.26738140e-01
-4.08386976e-01 -6.17750347e-01 -5.42629421e-01 -2.26478428e-01
-5.29798687e-01 -1.19332206e+00 -8.02085757e-01 -1.05383956e+00
5.25605977e-01 -1.70162678e-01 6.63886607e-01 3.19431901e-01
-2.05020562e-01 7.77175128e-01 -8.24823260e-01 -4.43991274e-01
-6.16306722e-01 -1.13602534e-01 -1.03803255e-01 -6.31307602e-01
6.81541443e-01 -2.45003849e-01 -2.50770867e-01 1.21679790e-01
-8.53644192e-01 -5.52733848e-03 7.53677845e-01 7.49345064e-01
2.00612456e-01 -4.00435776e-01 3.25282961e-01 -5.88370025e-01
9.21801984e-01 -1.48906261e-01 -2.50794917e-01 1.19372129e-01
-1.09153330e+00 -2.97843367e-01 -1.33448675e-01 -4.17303443e-01
-7.42503226e-01 -2.90521801e-01 -4.97199327e-01 4.45489675e-01
-8.52346495e-02 8.95325720e-01 -1.74835742e-01 -2.55149633e-01
5.60918212e-01 1.23729214e-01 2.07478166e-01 -6.72140539e-01
3.51071775e-01 1.01758540e+00 5.06798327e-01 -2.33501315e-01
3.95224392e-01 3.10960501e-01 -6.66815341e-01 -7.76046872e-01
1.02651715e-01 -6.38843894e-01 -7.60007441e-01 -6.43377125e-01
3.85067642e-01 -6.83443546e-01 -5.03582358e-01 7.63326287e-01
-6.83735907e-01 -4.64535743e-01 -6.30426168e-01 7.68904269e-01
-5.39937437e-01 3.04122061e-01 -3.16312134e-01 -6.77157104e-01
-7.82901719e-02 -9.44568098e-01 4.78148669e-01 -3.21922079e-02
-9.98051345e-01 -1.10117853e+00 9.08190161e-02 3.47769350e-01
6.41391575e-01 4.52130646e-01 8.90101552e-01 -6.04940236e-01
-1.50807649e-01 -4.55965042e-01 -9.68639180e-02 2.18881115e-01
5.03935277e-01 -9.80241746e-02 -3.03393364e-01 -3.26191425e-01
-2.84876823e-01 -7.64044747e-02 1.82789400e-01 1.03121534e-01
3.72207686e-02 -5.55706441e-01 -9.13621262e-02 2.77111292e-01
1.40820396e+00 5.75422883e-01 5.33031225e-01 8.91580462e-01
3.25924069e-01 6.75077319e-01 6.81725502e-01 3.95213872e-01
5.91393769e-01 7.30125964e-01 -1.88091576e-01 1.97551057e-01
-6.55849457e-01 -6.06824756e-01 3.70356470e-01 1.06270409e+00
-4.72637624e-01 6.54230341e-02 -9.82243121e-01 1.00906718e+00
-1.64812839e+00 -9.86711621e-01 -2.84443736e-01 2.14560151e+00
6.31971240e-01 2.67907977e-02 5.14697313e-01 2.82842010e-01
1.35250136e-01 2.70349421e-02 -3.95327896e-01 -3.19695652e-01
-4.06255275e-01 3.09374601e-01 8.01962852e-01 6.07952952e-01
-5.88513494e-01 1.03624189e+00 7.09704971e+00 2.34853908e-01
-1.34008801e+00 8.42022821e-02 -4.69954997e-01 -2.06517175e-01
-4.46159750e-01 3.04533571e-01 -5.38340628e-01 5.78666270e-01
2.80096620e-01 -4.52988178e-01 3.10358077e-01 5.88402271e-01
3.53474021e-01 -2.30130583e-01 -4.96167928e-01 1.15602791e+00
2.80933380e-01 -1.33654487e+00 1.50217175e-01 3.65159690e-01
3.33295226e-01 2.92703956e-01 -1.15316451e-01 1.36004910e-01
2.88060397e-01 -6.14549935e-01 1.35137010e+00 4.36961234e-01
1.25364101e+00 -1.71152338e-01 7.19107389e-01 5.79604842e-02
-1.09767091e+00 -4.91042212e-02 3.13642323e-01 -3.25303257e-01
7.16563880e-01 -1.16511330e-01 -5.21054387e-01 1.57101586e-01
6.45694256e-01 8.65528822e-01 -4.19135034e-01 1.30215335e+00
-4.55335826e-01 4.06263053e-01 -3.66464049e-01 -3.63804162e-01
9.21802446e-02 -6.96058944e-02 8.73725593e-01 1.29802990e+00
1.25598043e-01 1.75762132e-01 -1.91040233e-01 2.10048109e-01
4.71053213e-01 6.23433709e-01 -4.11960810e-01 -2.43202806e-01
4.78904873e-01 6.18423998e-01 -6.63606167e-01 -2.71186382e-01
-7.14786053e-01 7.75110424e-01 -3.23684305e-01 2.98047662e-01
-2.32310578e-01 -4.60426092e-01 7.71612287e-01 6.46881640e-01
-2.87288870e-03 -3.44992340e-01 -5.79930842e-01 -1.03882539e+00
2.54373729e-01 -1.31785965e+00 4.41859037e-01 -5.61666608e-01
-9.50263321e-01 1.61469415e-01 2.16280118e-01 -1.28786004e+00
-2.50845611e-01 -6.43484473e-01 8.10975134e-02 8.82462502e-01
-9.25360978e-01 -1.25405264e+00 -2.24110574e-01 2.75140703e-01
3.82701486e-01 -4.83830720e-02 8.35467339e-01 7.15131819e-01
8.26849639e-02 6.89755797e-01 -2.44087558e-02 3.96402270e-01
9.28138494e-01 -1.02253711e+00 4.44129378e-01 6.63808763e-01
-8.22732076e-02 9.69339907e-01 6.49683058e-01 -1.23756838e+00
-9.97338891e-01 -2.89927244e-01 1.66342437e+00 -5.81410944e-01
9.06781197e-01 -4.10950230e-03 -3.58970702e-01 8.05943966e-01
1.67983100e-01 -9.20290053e-01 8.14782023e-01 -2.20580310e-01
-1.72281172e-02 4.09175515e-01 -1.28119183e+00 8.65334094e-01
1.56003499e+00 -4.88465101e-01 -8.93389225e-01 2.68803388e-01
-1.67865217e-01 -7.04814017e-01 -9.38000739e-01 1.57347426e-01
1.41861284e+00 -5.49076259e-01 4.18997198e-01 -4.16122109e-01
-2.05718786e-01 -3.17216009e-01 -6.38914108e-02 -9.97192860e-01
-2.11651132e-01 -7.55982697e-01 4.71527189e-01 1.07917190e+00
4.25165981e-01 -1.00118935e+00 4.40013349e-01 9.86386836e-01
-3.61683190e-01 -5.31298459e-01 -9.59520280e-01 -7.51943052e-01
-4.78732847e-02 -7.56266832e-01 4.05822545e-01 1.04998541e+00
8.14021647e-01 -1.13471702e-01 -2.77196079e-01 -5.16343892e-01
3.63959283e-01 -6.03445530e-01 7.75030375e-01 -1.25585783e+00
-4.48001251e-02 -1.01224446e+00 -5.27916372e-01 -1.21118629e+00
-7.22327292e-01 -1.11214638e+00 -3.28892708e-01 -2.18612599e+00
-2.77716100e-01 -3.81334513e-01 3.18495005e-01 6.76976442e-01
4.23470646e-01 5.07992841e-02 3.87962818e-01 4.50568855e-01
2.28708953e-01 -1.46193013e-01 1.21712685e+00 8.81538913e-02
-6.30189538e-01 -2.85266519e-01 -1.13099110e+00 6.56806827e-01
6.16450608e-01 -2.43080556e-01 -1.37097880e-01 -4.75974590e-01
5.07302701e-01 -5.48698306e-01 3.00094485e-01 -6.60959840e-01
4.06688809e-01 -1.46184713e-01 -5.56727201e-02 -7.16927290e-01
-5.90937920e-02 -7.51396477e-01 5.31699836e-01 5.85532784e-01
-8.40984955e-02 -2.38311384e-02 3.14123660e-01 -9.43769440e-02
-1.50669590e-01 -1.96033731e-01 5.54741740e-01 -3.82068276e-01
-1.14038885e+00 -1.16260767e-01 -9.05006170e-01 -2.68425141e-03
7.35076964e-01 -1.02647257e+00 -2.09670871e-01 -5.87536216e-01
-9.36935365e-01 2.18460098e-01 8.85285020e-01 6.11036897e-01
4.32750553e-01 -1.35830998e+00 -4.43198562e-01 2.53677040e-01
5.22115111e-01 -4.53299969e-01 1.03688374e-01 1.10084116e+00
-8.05729091e-01 6.47892237e-01 -4.89553303e-01 9.36729982e-02
-1.43036520e+00 -5.37877716e-02 1.05064727e-01 8.81945863e-02
-1.03989077e+00 5.05841732e-01 -7.30908871e-01 -3.89724523e-01
5.26348174e-01 -6.94858372e-01 -3.30481768e-01 4.69101816e-01
4.94771600e-01 4.16982055e-01 -1.29876822e-01 -1.18673301e+00
-4.70103413e-01 7.81227410e-01 1.19346328e-01 -6.34841561e-01
1.21573865e+00 -4.18451697e-01 -7.22367987e-02 6.33895397e-01
7.25655138e-01 4.10719633e-01 -3.96223843e-01 -1.78879201e-01
2.05478996e-01 -5.75717866e-01 -3.70464951e-01 -1.48761857e+00
-6.73210979e-01 2.23737627e-01 6.35722458e-01 -1.57090768e-01
1.07532823e+00 1.54722914e-01 8.90682459e-01 -1.48142278e-01
6.42672479e-01 -1.53116214e+00 -6.35469139e-01 4.99745369e-01
1.18752611e+00 -6.80387437e-01 -5.58647811e-02 -1.06837749e-01
-7.66133726e-01 1.11974001e+00 5.63348606e-02 3.67817819e-01
8.07905734e-01 1.80138275e-01 7.47717440e-01 -4.40089017e-01
-3.35511304e-02 -5.78198910e-01 1.47839472e-01 8.80024493e-01
5.31562805e-01 8.58543534e-03 -1.46146441e+00 5.87095380e-01
-6.90110922e-01 4.83730584e-01 3.64162296e-01 1.13030481e+00
-2.06936315e-01 -1.53030705e+00 -3.51036519e-01 5.71203172e-01
-1.78612351e-01 -1.07729077e-01 -7.60737360e-01 1.27488959e+00
4.28129435e-01 7.82718003e-01 -3.71121854e-01 -3.54386300e-01
8.12060058e-01 9.92434099e-02 5.33441961e-01 -5.20749867e-01
-7.97781706e-01 1.32002637e-01 5.68028927e-01 -3.32101613e-01
-6.08985782e-01 -1.15841472e+00 -1.14678204e+00 -3.03580463e-01
-1.34253785e-01 -1.26985222e-01 9.06605005e-01 1.08707988e+00
1.32521819e-02 3.79653531e-03 -4.70625281e-01 -2.70119220e-01
-2.28254944e-01 -1.11586249e+00 -6.64185762e-01 3.41057479e-01
1.02918580e-01 -6.59723938e-01 -3.95821959e-01 1.28637673e-02] | [9.139941215515137, -6.44369649887085] |
4b01517d-9abd-4fb5-ab47-7a3897b4e237 | hmd-amp-protein-language-powered-hierarchical | 2111.06023 | null | https://arxiv.org/abs/2111.06023v1 | https://arxiv.org/pdf/2111.06023v1.pdf | HMD-AMP: Protein Language-Powered Hierarchical Multi-label Deep Forest for Annotating Antimicrobial Peptides | Identifying the targets of an antimicrobial peptide is a fundamental step in studying the innate immune response and combating antibiotic resistance, and more broadly, precision medicine and public health. There have been extensive studies on the statistical and computational approaches to identify (i) whether a peptide is an antimicrobial peptide (AMP) or a non-AMP and (ii) which targets are these sequences effective to (Gram-positive, Gram-negative, etc.). Despite the existing deep learning methods on this problem, most of them are unable to handle the small AMP classes (anti-insect, anti-parasite, etc.). And more importantly, some AMPs can have multiple targets, which the previous methods fail to consider. In this study, we build a diverse and comprehensive multi-label protein sequence database by collecting and cleaning amino acids from various AMP databases. To generate efficient representations and features for the small classes dataset, we take advantage of a protein language model trained on 250 million protein sequences. Based on that, we develop an end-to-end hierarchical multi-label deep forest framework, HMD-AMP, to annotate AMP comprehensively. After identifying an AMP, it further predicts what targets the AMP can effectively kill from eleven available classes. Extensive experiments suggest that our framework outperforms state-of-the-art models in both the binary classification task and the multi-label classification task, especially on the minor classes.The model is robust against reduced features and small perturbations and produces promising results. We believe HMD-AMP contributes to both the future wet-lab investigations of the innate structural properties of different antimicrobial peptides and build promising empirical underpinnings for precise medicine with antibiotics. | ['Yu Li', 'Licheng Zong', 'Xingyu Fan', 'Zhihang Dong', 'Qinze Yu'] | 2021-11-11 | null | null | null | null | ['protein-language-model'] | ['medical'] | [ 7.77234554e-01 -6.15620434e-01 -4.28312927e-01 -2.73419559e-01
-7.48579323e-01 -9.46637094e-01 3.41539443e-01 6.14594698e-01
-3.48060131e-01 1.19348633e+00 -1.50361657e-01 -6.45127892e-01
7.87458792e-02 -6.12987995e-01 -8.69102299e-01 -1.16644609e+00
-1.10217638e-01 7.13219523e-01 -6.83611110e-02 3.63676362e-02
3.08494627e-01 6.36514664e-01 -1.38735509e+00 7.36855626e-01
6.90451920e-01 7.56204724e-01 2.97370404e-01 6.44548416e-01
-2.71414310e-01 5.75056911e-01 -3.25657398e-01 -9.76623595e-02
2.33131200e-02 -4.04827744e-01 -8.48383248e-01 -5.45768738e-01
1.83636751e-02 -1.14402637e-01 4.97582495e-01 7.66087651e-01
8.01082909e-01 -2.82586247e-01 1.04557908e+00 -9.76091206e-01
-4.21092182e-01 2.42927810e-03 -5.49822330e-01 -1.55155018e-01
4.73991543e-01 3.61133963e-01 8.42455029e-01 -7.75125027e-01
6.30585134e-01 1.30892301e+00 7.29398131e-01 4.98475164e-01
-1.28101110e+00 -7.23010778e-01 -6.30991831e-02 1.68796778e-01
-1.12845337e+00 -1.99074179e-01 1.31630853e-01 -9.83714223e-01
1.38128531e+00 1.65697873e-01 1.89873815e-01 1.47943926e+00
6.65957391e-01 5.73910415e-01 1.44122183e+00 -2.36404866e-01
2.30883300e-01 -2.29208723e-01 6.69642761e-02 6.41928852e-01
-5.28874807e-02 2.08204702e-01 -2.58556753e-01 -8.46736431e-01
3.21346909e-01 3.40148985e-01 -1.61333382e-01 -3.73438656e-01
-1.18360806e+00 1.13309312e+00 1.13600641e-01 2.94458903e-02
-6.33367538e-01 -4.34119999e-01 5.29654741e-01 1.29395828e-03
1.59064606e-01 6.19012713e-01 -1.17062986e+00 1.98490977e-01
-3.96129429e-01 1.19733594e-01 8.77700150e-01 4.77456152e-01
8.63924325e-01 -5.56951344e-01 7.53795058e-02 1.08140397e+00
1.91535771e-01 5.78921616e-01 3.96462619e-01 -3.56489331e-01
-7.86719918e-02 5.59652448e-01 1.25256419e-01 -5.93170822e-01
-7.90552437e-01 -4.52487692e-02 -8.63184154e-01 2.37431645e-01
4.10360157e-01 -1.58969741e-02 -1.11956501e+00 1.80123079e+00
4.93435621e-01 -3.27918641e-02 1.10507853e-01 6.42243624e-01
6.52126670e-01 8.30505490e-01 5.28989494e-01 -4.80463296e-01
1.76358819e+00 -8.01742554e-01 -3.31875086e-01 -2.17836678e-01
7.40686476e-01 -9.91894066e-01 7.14463890e-01 5.39698541e-01
-2.16385052e-01 -3.29976678e-01 -1.00717819e+00 1.39520317e-01
-7.03683972e-01 -1.48421019e-01 6.00689530e-01 3.76652300e-01
-4.75516379e-01 5.73958457e-01 -6.22662842e-01 -4.31662410e-01
4.18732882e-01 5.21352947e-01 -5.52072227e-01 -4.57676888e-01
-1.31118083e+00 1.18982768e+00 5.32701313e-01 -3.60681117e-01
-1.11776912e+00 -4.44872797e-01 -6.49726748e-01 -4.43789810e-01
8.57313275e-02 -7.83489943e-01 8.68938625e-01 -7.47030377e-01
-1.34828210e+00 8.62529099e-01 -3.19077402e-01 -2.73457110e-01
-4.49199229e-02 -3.56323928e-01 -2.28520393e-01 6.97876140e-02
3.64584148e-01 7.21440077e-01 5.00957787e-01 -9.76545811e-01
-7.06724048e-01 -5.90198219e-01 -5.12052439e-02 9.22668353e-02
2.88811624e-01 1.76437840e-01 3.90153259e-01 -6.16716087e-01
-1.79615855e-01 -1.18339121e+00 -5.15768170e-01 -3.23878855e-01
-5.18322647e-01 -4.01202798e-01 4.83179718e-01 -4.84283775e-01
6.49403870e-01 -1.85491943e+00 2.77283251e-01 -8.21795501e-03
-1.32195493e-02 4.39793348e-01 -4.04993773e-01 8.77752960e-01
-2.12387398e-01 2.21867114e-01 -3.64626795e-01 5.39231837e-01
-4.52597857e-01 2.22140282e-01 -5.48516393e-01 6.99210942e-01
5.85355103e-01 6.86529636e-01 -9.18409050e-01 -3.11992496e-01
2.53255486e-01 7.59271801e-01 -3.28155845e-01 3.74363095e-01
-6.33147717e-01 8.72904718e-01 -6.51098907e-01 1.08697820e+00
6.49102628e-01 -5.35830617e-01 4.22822684e-01 -2.37593994e-01
-5.74557260e-02 2.64080316e-01 -6.60857737e-01 1.38995624e+00
-1.78399161e-01 -1.06754504e-01 -1.30243286e-01 -8.85114193e-01
8.60039413e-01 1.99713394e-01 6.90105617e-01 -3.75561297e-01
-5.26734181e-02 5.20195842e-01 1.33324504e-01 -4.04128760e-01
-3.28949839e-01 -3.99737805e-01 1.15114734e-01 3.81951660e-01
-1.04818165e-01 3.06649297e-01 2.78358581e-03 -2.08367079e-01
1.22336423e+00 4.17798042e-01 8.44853282e-01 -2.89996028e-01
6.73482776e-01 2.21456885e-01 7.89704621e-01 6.42420113e-01
-2.04094544e-01 3.41849178e-01 2.13399932e-01 -8.70918572e-01
-9.93699491e-01 -7.05606163e-01 -4.46264535e-01 1.70389009e+00
-5.84081337e-02 -2.01032497e-02 -4.75949705e-01 -5.90272188e-01
8.30090195e-02 2.22745672e-01 -4.43776548e-01 1.13213561e-01
-5.52149355e-01 -1.35527945e+00 6.44885361e-01 1.49760380e-01
2.27713529e-02 -1.36609173e+00 -5.39286554e-01 5.79058886e-01
-3.93231839e-01 -7.59772241e-01 -1.11712046e-01 1.03267252e+00
-3.58535856e-01 -1.55417848e+00 -5.90870023e-01 -9.67017770e-01
4.51622866e-02 -5.75995110e-02 9.61725891e-01 -1.91041991e-01
-4.61326331e-01 -3.39690149e-01 -4.45351273e-01 -6.91577017e-01
-5.76450527e-01 4.78051491e-02 3.42206091e-01 -2.90377915e-01
7.92583108e-01 -4.12604928e-01 -6.89832330e-01 4.27153736e-01
-7.57003665e-01 -5.04173450e-02 1.01598024e+00 1.06441915e+00
1.12748849e+00 -2.67457843e-01 8.82315338e-01 -9.52108085e-01
4.00279403e-01 -7.63179600e-01 -2.19522685e-01 3.44938099e-01
-5.47842324e-01 2.37987831e-01 8.39149475e-01 -6.37826264e-01
-4.87013191e-01 7.34442830e-01 -5.88082910e-01 1.93262801e-01
-5.21487117e-01 4.21324581e-01 -2.32425004e-01 -6.60000667e-02
6.57459259e-01 3.68891895e-01 -4.06658724e-02 -4.56791192e-01
2.22386152e-01 9.15086865e-01 9.84789282e-02 -6.12244010e-01
7.25017339e-02 3.41390342e-01 2.89948523e-01 -7.60264456e-01
-9.09232199e-01 -7.96440661e-01 -5.36644578e-01 4.27987427e-01
1.11246514e+00 -1.02286661e+00 -8.44873726e-01 6.65842652e-01
-1.25486088e+00 -2.08888575e-01 4.78246957e-01 3.14379930e-01
-7.21421540e-01 5.39268196e-01 -9.33001637e-01 -6.01303756e-01
-6.50028527e-01 -1.60066664e+00 1.27985692e+00 -1.52810320e-01
-4.39845175e-01 -5.45335293e-01 5.90142131e-01 3.18856001e-01
4.70770076e-02 5.93074143e-01 1.49875498e+00 -9.97096896e-01
-2.27252185e-01 1.83390573e-01 -1.27041861e-01 1.98881596e-01
4.10609305e-01 -3.50908712e-02 -8.78814518e-01 -3.68158937e-01
-2.19373256e-02 -9.46793139e-01 1.00434470e+00 4.86940295e-01
8.43665779e-01 -2.95158625e-01 -7.71258116e-01 6.88519955e-01
1.44470179e+00 6.31515563e-01 5.89152575e-01 6.47159874e-01
7.03504026e-01 7.79558301e-01 9.17440236e-01 2.87556857e-01
5.23417350e-03 5.52629411e-01 5.67024350e-01 -1.36086017e-01
4.74942148e-01 5.86341210e-02 2.99038738e-01 4.55887288e-01
1.14600398e-01 -4.68172878e-01 -1.04482722e+00 3.20756942e-01
-1.66817796e+00 -8.24550986e-01 -1.78519979e-01 1.89577973e+00
1.40374112e+00 -2.55202919e-01 7.78793916e-02 -1.50954142e-01
6.58231080e-01 -3.74395512e-02 -1.03511310e+00 -5.19737899e-01
-5.63036621e-01 4.96642888e-01 4.68480706e-01 2.66720712e-01
-1.69967687e+00 7.41083145e-01 6.57759762e+00 8.66366446e-01
-1.13079000e+00 -3.39213103e-01 6.29300594e-01 3.92080486e-01
2.81451851e-01 -1.20564349e-01 -1.02921045e+00 4.52384800e-01
1.13917983e+00 4.37762320e-01 3.09479415e-01 9.36945140e-01
-1.35112166e-01 1.24769270e-01 -1.27596307e+00 6.49785340e-01
-2.20383644e-01 -1.28720379e+00 2.85303652e-01 2.83316225e-01
4.91695672e-01 4.15449232e-01 -2.49827594e-01 1.46248981e-01
5.89186907e-01 -1.49754262e+00 9.76407900e-02 2.99452335e-01
9.23000872e-01 -7.44841456e-01 1.00878632e+00 6.09121323e-01
-9.10064697e-01 1.33686615e-02 -5.18272996e-01 5.44349477e-02
-4.57286648e-02 5.92978060e-01 -1.04028153e+00 3.52710962e-01
4.95890677e-01 6.14002049e-01 -2.65823513e-01 5.80714107e-01
-3.09351459e-02 3.05369049e-01 -2.16894507e-01 -1.50553450e-01
1.22307755e-01 -1.23354137e-01 9.84308496e-02 1.35197270e+00
3.62768248e-02 3.55935603e-01 7.23765135e-01 3.22990835e-01
2.41873875e-01 4.38063711e-01 -5.35845518e-01 -4.22451645e-01
3.18669587e-01 1.11824071e+00 -5.19512653e-01 -3.04753542e-01
-3.43136966e-01 8.10812116e-01 2.50596315e-01 2.61799574e-01
-5.68075061e-01 -2.69334286e-01 1.05509531e+00 -2.51686692e-01
3.32900345e-01 9.33998749e-02 -1.10060103e-01 -7.32182562e-01
-5.51320195e-01 -1.55777633e+00 5.65659046e-01 -4.37850147e-01
-1.86250091e+00 4.84530598e-01 -5.34166634e-01 -9.21415687e-01
-3.18100125e-01 -1.07044256e+00 6.54989630e-02 1.04358160e+00
-1.48448980e+00 -1.22696853e+00 2.36230895e-01 2.96552241e-01
4.63708699e-01 -9.40770879e-02 1.55041599e+00 8.63921419e-02
-3.71160954e-01 1.80158585e-01 4.23093170e-01 -1.35032728e-01
1.01658952e+00 -9.85731483e-01 1.69302061e-01 2.17976972e-01
-1.77124247e-01 8.33059609e-01 6.98006392e-01 -9.33765292e-01
-1.17273927e+00 -1.37429857e+00 6.37141705e-01 -6.52002156e-01
4.95325267e-01 -2.76556373e-01 -1.03983998e+00 3.82464260e-01
-5.53689385e-03 -4.75567222e-01 1.48261118e+00 -8.71448889e-02
-6.72579110e-01 5.87085903e-01 -1.19751716e+00 2.33233631e-01
7.04661846e-01 -4.59891230e-01 -4.56633806e-01 8.89105201e-01
8.99533391e-01 -9.98037755e-02 -1.02650011e+00 9.88743305e-01
9.32402551e-01 -8.39682102e-01 1.22710669e+00 -1.04824197e+00
5.39402723e-01 -5.04270673e-01 -4.14568573e-01 -1.15129113e+00
-6.33748472e-01 1.75895151e-02 6.05925657e-02 7.43392408e-01
4.15168613e-01 -7.50878215e-01 5.33479869e-01 -2.01860160e-01
-8.06058347e-02 -1.20336711e+00 -8.14468026e-01 -6.38677597e-01
3.71480227e-01 1.55103758e-01 6.09012723e-01 1.19054377e+00
-2.20116481e-01 6.25548780e-01 -5.69549859e-01 1.26864538e-01
3.87344271e-01 4.88991410e-01 6.05847597e-01 -1.47145402e+00
-4.12451237e-01 -2.48681843e-01 -2.34040782e-01 -7.66391218e-01
1.54628977e-01 -7.61385322e-01 1.67590305e-01 -1.43351233e+00
4.96850878e-01 -3.15983415e-01 -6.62771821e-01 7.09312081e-01
-3.55365634e-01 1.24293782e-01 -5.62985182e-01 3.63341033e-01
-4.13974047e-01 1.31990045e-01 8.14545929e-01 -3.28564733e-01
-4.78093401e-02 -1.00436725e-01 -6.90199256e-01 6.43023551e-01
8.84361386e-01 -7.05867231e-01 6.20702095e-02 1.57677695e-01
2.54994214e-01 -2.75857508e-01 4.31770310e-02 -4.07149643e-01
-2.87874907e-01 -7.74148107e-01 5.48768878e-01 -7.57288098e-01
2.99186558e-01 -4.85431403e-01 3.78289580e-01 1.02491140e+00
-1.72532558e-01 -1.77786909e-02 2.02669576e-01 7.48943090e-01
7.21584409e-02 4.52953316e-02 1.04523015e+00 -3.41843009e-01
-5.20501614e-01 2.95947701e-01 -6.31971478e-01 -7.37729818e-02
1.02009094e+00 -1.20067084e-02 -5.80729961e-01 1.88354805e-01
-4.20401633e-01 7.37812072e-02 7.17835546e-01 3.62994522e-01
2.98967302e-01 -9.47206378e-01 -7.82040358e-01 2.10458398e-01
5.56771934e-01 -1.53901711e-01 -7.83612952e-02 5.59454143e-01
-8.33514869e-01 8.65241647e-01 -3.75268608e-01 -8.00801754e-01
-1.37470448e+00 9.92772698e-01 2.26874664e-01 -4.19935077e-01
5.89356869e-02 9.04202044e-01 7.29983807e-01 -6.97909653e-01
2.56425347e-02 6.90013766e-02 -3.88254791e-01 7.53343105e-03
6.74832225e-01 2.57885680e-02 9.32313055e-02 -8.89083028e-01
-8.36291015e-01 5.92056990e-01 -2.89729476e-01 6.99447095e-01
1.46045303e+00 2.97884345e-01 -5.43018579e-01 3.92357230e-01
1.21841180e+00 -2.03547090e-01 -8.12038302e-01 -7.09105432e-02
7.92778805e-02 -2.21456029e-02 -7.06811607e-01 -1.21622598e+00
-8.58278945e-02 9.17152882e-01 9.28136706e-01 -2.57454008e-01
9.16741490e-01 -2.81191431e-02 7.63383389e-01 5.60496211e-01
6.22638047e-01 -7.03890383e-01 -6.55482635e-02 5.69905639e-01
6.85718417e-01 -1.33583498e+00 1.31293423e-02 -3.18507344e-01
-3.74959141e-01 1.01758063e+00 3.46196413e-01 3.34753931e-01
2.54568607e-01 2.26371124e-01 3.07162046e-01 -1.99064136e-01
-8.39519799e-01 6.16907552e-02 -2.74099014e-03 9.05296504e-01
7.98400342e-01 3.63709003e-01 -4.07296419e-01 3.15275699e-01
1.19303741e-01 -9.02174264e-02 1.48138165e-01 1.00140715e+00
-7.70489991e-01 -1.47035253e+00 -3.44924331e-01 5.54589391e-01
-8.80864978e-01 -4.46597874e-01 -8.95260394e-01 2.70914227e-01
3.36376250e-01 1.09775746e+00 -5.66949368e-01 -4.25710618e-01
1.23323262e-01 4.64476734e-01 2.80125111e-01 -7.17574298e-01
-3.70461196e-01 9.94548574e-02 5.09531833e-02 -2.94421017e-01
-5.21134257e-01 -5.13943374e-01 -1.35742867e+00 6.74090162e-02
-4.84011889e-01 5.69277257e-02 6.21504784e-01 8.65184486e-01
5.85525393e-01 1.12516358e-01 5.37933826e-01 -6.62358582e-01
-6.37965262e-01 -9.92839813e-01 -4.45678711e-01 3.92764628e-01
2.86652178e-01 -8.11129630e-01 -2.28515506e-01 6.77649677e-02] | [4.807767868041992, 5.59494686126709] |
aa536d33-fc77-4d00-b580-e27d05ddc1f3 | a-report-on-the-automatic-evaluation-of | null | null | https://aclanthology.org/W16-0506 | https://aclanthology.org/W16-0506.pdf | A Report on the Automatic Evaluation of Scientific Writing Shared Task | null | ['Vidas Daudaravicius', 'Rafael E. Banchs', 'Elena Volodina', 'Courtney Napoles'] | 2016-06-01 | null | null | null | ws-2016-6 | ['grammatical-error-detection'] | ['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] | [-7.242435932159424, 3.730689764022827] |
d6105bc5-e59f-45cc-9dee-e42e80d8d4b1 | instructabsa-instruction-learning-for-aspect | 2302.08624 | null | https://arxiv.org/abs/2302.08624v5 | https://arxiv.org/pdf/2302.08624v5.pdf | InstructABSA: Instruction Learning for Aspect Based Sentiment Analysis | In this paper, we present InstructABSA, Aspect Based Sentiment Analysis (ABSA) using the instruction learning paradigm for the ABSA subtasks: Aspect Term Extraction (ATE), Aspect Term Sentiment Classification (ATSC), and Joint Task modeling. Our method introduces positive, negative, and neutral examples to each training sample, and instruction tunes the model (Tk-Instruct) the ABSA subtasks, yielding significant performance improvements. Experimental results on the Sem Eval 2014, 15, and 16 datasets demonstrate that InstructABSA outperforms the previous state-of-the-art (SOTA) approaches on the three ABSA subtasks (ATE, ATSC, and Joint Task) by a significant margin, outperforming 7x larger models. In particular, InstructABSA surpasses the SOTA on the Rest14 ATE subtask by 5.69% points, Rest15 ATSC subtask by 9.59% points, and on the Lapt14 Joint Task by 3.37% points. Our results also suggest a strong generalization ability to new domains across all three subtasks | ['Chitta Baral', 'Swaroop Mishra', 'Siddharth Goyal', 'Saurabh Arjun Sawant', 'Himanshu Gupta', 'Kevin Scaria'] | 2023-02-16 | null | null | null | null | ['term-extraction', 'aspect-extraction', 'aspect-based-sentiment-analysis'] | ['natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [ 9.56456177e-03 1.84172675e-01 -4.70619202e-01 -4.70805705e-01
-1.13973010e+00 -5.95873177e-01 8.68712723e-01 2.40058929e-01
-2.13542148e-01 6.23995900e-01 1.72889724e-01 -5.28209865e-01
1.87831864e-01 -5.93361676e-01 -7.76475847e-01 -5.68491459e-01
2.77330637e-01 5.77819407e-01 1.96288973e-01 -5.23294389e-01
1.99287876e-01 -3.03047866e-01 -1.54506838e+00 6.47800982e-01
9.54851985e-01 1.39758074e+00 -7.30336189e-01 3.19254190e-01
-4.51897293e-01 6.22575104e-01 -6.72459960e-01 -9.98933375e-01
-1.01172447e-01 6.72627166e-02 -7.37789094e-01 -2.99719602e-01
3.89548212e-01 2.45234206e-01 5.80574989e-01 7.30525613e-01
1.33431524e-01 1.40310019e-01 8.22659969e-01 -1.49919307e+00
-5.78526914e-01 8.11271012e-01 -8.38126183e-01 -2.87326545e-01
2.88860768e-01 5.29353023e-02 1.44359183e+00 -1.10358381e+00
3.65899652e-01 1.11506855e+00 6.52055204e-01 5.67899644e-01
-8.83475184e-01 -9.03477550e-01 6.69116676e-01 -2.02284589e-01
-5.88817298e-01 -2.38936003e-02 6.83509231e-01 -4.22093183e-01
1.68950546e+00 5.05100906e-01 8.18898201e-01 1.25658774e+00
6.89517319e-01 1.49547505e+00 1.59293365e+00 -7.31880590e-02
5.52039742e-01 2.40900263e-01 9.15094078e-01 3.73739690e-01
4.16924775e-01 -6.76887631e-02 -7.77969062e-01 -3.11692744e-01
-1.84838429e-01 -3.08735222e-01 1.25886515e-01 -1.78462222e-01
-9.82391536e-01 7.47092187e-01 1.51475057e-01 8.75143856e-02
-5.09928942e-01 4.95885313e-03 7.90737033e-01 5.81361711e-01
8.92154694e-01 7.88861752e-01 -1.41738951e+00 -5.27512193e-01
-3.52633804e-01 4.11487311e-01 1.11324000e+00 1.04061759e+00
6.20098233e-01 4.79217350e-01 -3.67973119e-01 7.14687467e-01
1.66803971e-01 9.37102139e-01 6.90536380e-01 -5.41860104e-01
3.13426346e-01 9.56216812e-01 -3.72702360e-01 -3.19127262e-01
-3.54962260e-01 -8.35900307e-01 -3.79060596e-01 2.09055662e-01
4.63945009e-02 -2.33071595e-01 -1.37111104e+00 1.45568430e+00
3.89896721e-01 -3.49168330e-01 6.16297163e-02 3.02721709e-01
1.14586723e+00 5.07873952e-01 4.40778643e-01 6.18227087e-02
1.87276328e+00 -1.34325957e+00 -8.64199519e-01 -9.27055478e-01
8.84541810e-01 -1.04120409e+00 1.47790241e+00 7.08206177e-01
-9.56335187e-01 -2.56494135e-01 -1.28564477e+00 2.13570312e-01
-7.86522865e-01 1.12351231e-01 1.13650048e+00 6.33238792e-01
-1.07485306e+00 -1.08656563e-01 -7.57815540e-01 -1.70466989e-01
3.02917480e-01 4.59672481e-01 -9.16551203e-02 1.25170842e-01
-9.94102061e-01 6.13004148e-01 1.25063911e-01 -5.46426594e-01
-7.98891366e-01 -1.12128568e+00 -1.29560840e+00 -2.02377755e-02
6.28902137e-01 -8.32288444e-01 1.48657036e+00 -9.37313199e-01
-1.59499872e+00 9.86656845e-01 -3.89344931e-01 -3.07138175e-01
-1.29332736e-01 -7.01883912e-01 -4.07073021e-01 -5.53437293e-01
3.09376299e-01 4.91567433e-01 9.08802986e-01 -1.21039045e+00
-6.11160934e-01 -4.95782614e-01 4.32800263e-01 1.08038552e-01
-3.78864855e-01 -9.27488133e-02 -2.85298437e-01 -7.40649879e-01
-2.59810418e-01 -1.05791926e+00 -8.09650272e-02 -9.15677905e-01
-6.82010949e-01 -5.82242310e-01 6.93009079e-01 -3.83695602e-01
1.22459698e+00 -2.25073791e+00 -1.68374076e-03 -1.13823637e-01
1.82663545e-01 1.69789955e-01 -2.49484032e-01 -9.63665247e-02
-3.16402376e-01 -3.58508201e-03 -1.12938762e-01 -4.69231993e-01
1.63685590e-01 2.97872089e-02 -3.84311616e-01 -4.19825882e-01
4.62201506e-01 1.18532121e+00 -6.96746171e-01 -1.06037073e-01
-1.71115756e-01 1.36006325e-01 -8.65742803e-01 1.68497533e-01
-5.35322011e-01 -2.15788051e-01 -7.27136195e-01 1.05690491e+00
4.72533792e-01 -2.81926304e-01 -1.24932729e-01 -4.20035161e-02
-1.56525243e-02 9.64064181e-01 -4.47743118e-01 1.36112845e+00
-6.81989670e-01 3.57932866e-01 -1.56910658e-01 -6.06594384e-01
1.02188706e+00 2.09253818e-01 3.60881478e-01 -9.36809421e-01
3.10550690e-01 2.69329607e-01 8.55767075e-03 -2.21274927e-01
7.06724048e-01 -2.96704501e-01 -7.04583704e-01 5.81466675e-01
3.22255284e-01 -4.86686170e-01 2.16435283e-01 2.79077560e-01
1.04477155e+00 1.50578842e-01 4.59239662e-01 -5.44695020e-01
5.62683225e-01 9.74597931e-02 4.60538536e-01 3.24788660e-01
-3.60104769e-01 2.56249279e-01 8.01694989e-01 -6.03299797e-01
-3.77080768e-01 -9.19789016e-01 -7.55639467e-03 1.58948576e+00
-3.12056959e-01 -7.90204644e-01 -6.95199668e-01 -1.34293818e+00
5.12325056e-02 1.24084556e+00 -1.15563774e+00 -3.10227364e-01
-3.61983091e-01 -9.15255547e-01 1.78116225e-02 6.36605680e-01
5.13588607e-01 -1.11732864e+00 -9.84641537e-02 -1.59996241e-01
-7.61094736e-03 -1.15726829e+00 -5.34469962e-01 6.10763967e-01
-7.01813400e-01 -8.28396261e-01 -3.22557032e-01 -7.09746897e-01
2.40821093e-01 7.84832835e-02 1.69915390e+00 -3.70479077e-01
2.71032691e-01 6.05830193e-01 -3.24482441e-01 -8.80037725e-01
-4.10969853e-02 4.99373019e-01 -1.68357402e-01 -2.44120136e-01
1.07071412e+00 -4.30814415e-01 -2.56979674e-01 -1.00699462e-01
-8.73299479e-01 -1.10892951e-02 8.42114985e-01 8.37706685e-01
7.63006687e-01 -3.86632949e-01 7.07628608e-01 -1.33987439e+00
6.72692716e-01 -5.11231482e-01 -2.38348722e-01 2.91052330e-02
-1.24754262e+00 -4.46953699e-02 3.28233093e-01 -2.16525674e-01
-1.20457006e+00 -4.47656363e-01 -1.26711816e-01 5.21185882e-02
-1.59774750e-01 6.39623046e-01 -2.58764803e-01 2.42298439e-01
5.91031611e-01 8.84639695e-02 -1.69351503e-01 -9.55279395e-02
8.72239918e-02 4.81481254e-01 -8.69608670e-02 -6.76843584e-01
6.80062175e-01 3.48446220e-01 -2.79468834e-01 -4.27839220e-01
-1.38671398e+00 -3.84291679e-01 -1.41025141e-01 -9.83971730e-02
1.04683316e+00 -1.29297554e+00 -5.04484534e-01 5.76402068e-01
-7.17284620e-01 -3.65806371e-01 -5.29070437e-01 2.36086771e-01
-2.88217008e-01 -2.75580019e-01 -6.49391174e-01 -5.45657992e-01
-9.77480769e-01 -1.37010610e+00 1.61254811e+00 3.05121928e-01
-7.89203465e-01 -1.08240342e+00 1.12921834e-01 8.63655567e-01
4.71812785e-01 1.17693372e-01 1.24322808e+00 -1.19071281e+00
-3.83341879e-01 -3.08438122e-01 1.43941730e-01 4.64179367e-01
9.39395949e-02 -5.21889329e-02 -1.27773476e+00 -7.19250888e-02
2.88149327e-01 -6.05595589e-01 8.57532620e-01 5.63747585e-01
7.68815577e-01 1.01334319e-01 -5.76429442e-02 4.97045249e-01
8.48056376e-01 4.87525910e-01 3.20365041e-01 8.86566520e-01
5.95373333e-01 3.10639441e-01 1.14567459e+00 1.25431821e-01
8.91937912e-01 3.69615972e-01 4.16087061e-01 1.08979471e-01
-1.54076740e-01 5.56831621e-02 1.00477827e+00 1.32787514e+00
7.74354786e-02 7.99955241e-03 -6.69429362e-01 7.59284317e-01
-1.52256942e+00 -1.19012408e-01 -2.68310130e-01 1.87346268e+00
1.08714426e+00 5.93437672e-01 -1.08396605e-01 9.80634242e-04
-2.71057516e-01 5.24289250e-01 -6.82187915e-01 -1.14747536e+00
-1.65845498e-01 6.28947616e-01 -1.24087870e-01 3.49203795e-01
-1.25723457e+00 1.13805759e+00 5.71557283e+00 9.77384448e-01
-9.23192620e-01 3.12110465e-02 9.88642037e-01 -2.09863633e-01
-7.58796692e-01 2.70290468e-02 -8.48950207e-01 6.95002750e-02
9.16595995e-01 -3.69047850e-01 2.02095404e-01 1.35084069e+00
-5.47248423e-01 -1.64205775e-01 -1.07871974e+00 5.03946722e-01
4.60601419e-01 -7.69042969e-01 4.18513000e-01 -2.35666707e-01
1.19831872e+00 1.71269998e-01 6.46281660e-01 1.43066847e+00
4.54322129e-01 -7.91583061e-01 3.60592991e-01 -1.60313472e-01
3.97355139e-01 -7.72854269e-01 1.16351295e+00 -1.60240963e-01
-1.01549971e+00 2.75135279e-01 1.41194403e-01 -1.07345052e-01
-2.78048098e-01 7.85130382e-01 -6.19821429e-01 5.35454214e-01
9.55635726e-01 8.56476843e-01 -8.49512041e-01 3.52483898e-01
-4.30470198e-01 1.02226162e+00 3.21638659e-02 -3.50329548e-01
4.75350618e-01 -1.33880928e-01 7.47960925e-01 9.45197403e-01
-1.69643499e-02 -2.10065484e-01 1.29132271e-01 3.86318266e-01
-2.41488591e-01 2.90467203e-01 -3.18305701e-01 -2.81379670e-01
-9.51288790e-02 1.43541563e+00 -2.79232979e-01 -6.64850891e-01
-5.44929206e-01 5.29530644e-01 1.47301331e-01 6.29075885e-01
-7.80332088e-01 -4.54311222e-01 9.41181481e-01 -1.72126397e-01
4.22110409e-01 4.30858165e-01 -8.92689168e-01 -1.33543754e+00
1.98830366e-01 -1.42643416e+00 6.38019681e-01 -6.72290802e-01
-1.30039430e+00 8.97462904e-01 -1.19571269e-01 -1.12751341e+00
-4.47383463e-01 -9.43138897e-01 -7.63118088e-01 4.36729193e-01
-1.62901759e+00 -1.38701141e+00 -7.04418421e-02 3.92891794e-01
8.07527661e-01 -4.30938870e-01 8.55080724e-01 3.27635072e-02
-4.67220515e-01 8.51381481e-01 -3.90693188e-01 -1.69926539e-01
8.71189415e-01 -1.67796874e+00 7.13159680e-01 3.36696297e-01
-2.18125209e-01 7.89971530e-01 7.66326964e-01 -7.11045742e-01
-1.61003315e+00 -9.96172607e-01 9.17325079e-01 -8.62247765e-01
1.08631825e+00 -4.25940782e-01 -8.46225321e-01 1.26234031e+00
6.56931579e-01 -4.60535705e-01 1.08321238e+00 9.58177447e-01
-8.22529972e-01 -1.11833863e-01 -8.26676786e-01 8.11603010e-01
4.42896038e-01 -5.34474254e-01 -7.34301865e-01 2.29678318e-01
1.15271080e+00 -4.35422271e-01 -1.04083073e+00 8.57339442e-01
4.56039250e-01 -8.32333684e-01 7.15410352e-01 -8.26753676e-01
8.21898878e-01 6.87536374e-02 -2.29766101e-01 -1.49616170e+00
-1.28544822e-01 -4.12093580e-01 -3.37299347e-01 1.24094093e+00
9.99285161e-01 -8.57624173e-01 7.38259137e-01 7.83725262e-01
-3.69101316e-01 -1.39726937e+00 -8.42191517e-01 -5.73566735e-01
5.54265380e-01 -8.58631909e-01 7.87506759e-01 9.35832858e-01
1.34599850e-01 1.04479444e+00 8.16634521e-02 -2.62955993e-01
4.40809488e-01 5.97321510e-01 9.44172204e-01 -7.77499318e-01
-4.67222691e-01 -5.86622775e-01 -1.34502605e-01 -8.19087863e-01
1.63399622e-01 -8.31151903e-01 -2.04783171e-01 -1.22386682e+00
4.94700164e-01 1.90368992e-05 -5.90023398e-01 6.76469684e-01
-9.04022157e-01 6.20586686e-02 -1.21443935e-01 -4.04955357e-01
-9.51757312e-01 9.40925896e-01 1.21999168e+00 -5.36986172e-01
-7.39659788e-03 2.66403019e-01 -1.18942845e+00 9.39114571e-01
7.75037646e-01 -2.73442030e-01 -4.42527115e-01 -2.90618569e-01
4.73972827e-01 -6.74867511e-01 -2.63550431e-01 -6.62189186e-01
-2.67854154e-01 9.36073065e-02 1.39319286e-01 -6.33525431e-01
5.51462948e-01 -5.61844409e-01 -5.09899259e-01 3.69645208e-01
-2.05083266e-01 2.21750528e-01 8.88392985e-01 3.38029593e-01
-4.87599224e-01 2.29792088e-01 1.77295417e-01 2.65913785e-01
-4.25218254e-01 1.33312359e-01 -4.28048611e-01 5.03673017e-01
1.01889503e+00 2.87140697e-01 -5.66760361e-01 -4.03908879e-01
-4.80744392e-01 4.20006424e-01 1.43626958e-01 7.71392822e-01
4.78137076e-01 -1.29553449e+00 -3.25912237e-01 3.73099387e-01
5.73590815e-01 1.07941050e-02 2.65842974e-01 9.91052985e-01
2.02580661e-01 6.08400226e-01 2.22208858e-01 -6.26385033e-01
-1.22664237e+00 2.84493715e-01 -4.64063212e-02 -9.22711551e-01
-2.45657731e-02 9.95608032e-01 7.65684426e-01 -1.02053285e+00
-5.77462930e-03 -5.97093582e-01 -4.01116818e-01 4.50942554e-02
3.63896102e-01 -3.86978313e-02 3.40588301e-01 -3.04878294e-01
-2.94265926e-01 5.98450065e-01 -7.08049655e-01 6.67445138e-02
1.25447643e+00 3.51037294e-01 -1.76459283e-01 7.97254562e-01
1.00534761e+00 3.12078804e-01 -5.03167748e-01 -2.39764944e-01
1.10430427e-01 3.06948368e-02 -8.11085552e-02 -1.45409596e+00
-1.21643043e+00 7.13000298e-01 7.81388953e-02 2.63401479e-01
1.21518719e+00 3.74798961e-02 9.89433289e-01 3.36010069e-01
3.82459641e-01 -9.51253355e-01 1.25127882e-01 9.10163343e-01
8.36664319e-01 -1.50684357e+00 1.19691975e-01 -3.49185556e-01
-1.24286914e+00 6.02993011e-01 1.08843875e+00 -2.13125478e-02
9.14784014e-01 4.92151022e-01 3.69121701e-01 -6.18844569e-01
-1.51505852e+00 1.03308000e-01 6.75864875e-01 1.99518815e-01
6.91195905e-01 5.03883719e-01 -2.05675453e-01 1.24433243e+00
-6.44417107e-01 -4.22662795e-01 1.32725194e-01 9.81770575e-01
-5.88284619e-02 -9.82768357e-01 4.40384559e-02 8.30771446e-01
-7.58204341e-01 -5.45059562e-01 -7.14119077e-01 1.02855897e+00
-3.10963482e-01 8.94717634e-01 -2.44709998e-01 -6.96441233e-01
5.59959114e-01 4.59199280e-01 -1.48346871e-01 -6.27961636e-01
-1.25935328e+00 1.88680142e-01 5.77385604e-01 -6.72303379e-01
-1.31873474e-01 -8.08042526e-01 -1.19215989e+00 7.69432634e-04
-3.16132396e-01 3.26580733e-01 6.50583446e-01 1.11592841e+00
5.55302441e-01 1.06155765e+00 3.74924630e-01 -1.44388333e-01
-2.49404326e-01 -1.14615524e+00 -1.90990552e-01 3.14466029e-01
1.26849577e-01 -4.70555961e-01 -6.15708768e-01 -3.79206508e-01] | [11.46821403503418, 6.685230731964111] |
74e895d3-9342-4491-aa72-0707f20925d4 | regularizing-towards-soft-equivariance-under | 2306.00356 | null | https://arxiv.org/abs/2306.00356v1 | https://arxiv.org/pdf/2306.00356v1.pdf | Regularizing Towards Soft Equivariance Under Mixed Symmetries | Datasets often have their intrinsic symmetries, and particular deep-learning models called equivariant or invariant models have been developed to exploit these symmetries. However, if some or all of these symmetries are only approximate, which frequently happens in practice, these models may be suboptimal due to the architectural restrictions imposed on them. We tackle this issue of approximate symmetries in a setup where symmetries are mixed, i.e., they are symmetries of not single but multiple different types and the degree of approximation varies across these types. Instead of proposing a new architectural restriction as in most of the previous approaches, we present a regularizer-based method for building a model for a dataset with mixed approximate symmetries. The key component of our method is what we call equivariance regularizer for a given type of symmetries, which measures how much a model is equivariant with respect to the symmetries of the type. Our method is trained with these regularizers, one per each symmetry type, and the strength of the regularizers is automatically tuned during training, leading to the discovery of the approximation levels of some candidate symmetry types without explicit supervision. Using synthetic function approximation and motion forecasting tasks, we demonstrate that our method achieves better accuracy than prior approaches while discovering the approximate symmetry levels correctly. | ['Juho Lee', 'Hongseok Yang', 'Hyungi Lee', 'Hyunsu Kim'] | 2023-06-01 | null | null | null | null | ['motion-forecasting'] | ['computer-vision'] | [-1.16163723e-01 2.64597535e-01 -4.50895429e-01 -1.83199465e-01
-4.39016908e-01 -7.24071980e-01 8.82357895e-01 -3.66028398e-01
3.00145030e-01 5.31097233e-01 6.84777260e-01 3.39261852e-02
-1.87402844e-01 -8.21038783e-01 -1.17327738e+00 -5.24781168e-01
7.18478784e-02 6.64317846e-01 3.02195907e-01 -3.31728816e-01
2.46346980e-01 5.47368288e-01 -1.56610107e+00 6.35821164e-01
7.77659655e-01 7.08563626e-01 -2.24030346e-01 4.25921679e-01
-2.89025344e-02 6.64210200e-01 -2.53862113e-01 -2.71531492e-01
6.78687513e-01 -6.57205880e-01 -1.01593494e+00 3.40610772e-01
7.87404537e-01 -2.07917079e-01 -3.92790258e-01 1.02809846e+00
-1.99306712e-01 1.80831086e-02 7.51607597e-01 -1.11951613e+00
-2.72325993e-01 5.35548389e-01 -2.67391950e-01 -6.53573945e-02
1.79848760e-01 2.59639416e-02 1.42968130e+00 -8.03539455e-01
9.66029763e-01 1.20964670e+00 8.92888665e-01 5.19088447e-01
-1.59060764e+00 -3.37256789e-01 2.77832031e-01 1.27034351e-01
-1.32031035e+00 -5.88796079e-01 8.98993015e-01 -5.81068516e-01
7.56179690e-01 1.85903907e-01 7.14714825e-01 9.90581810e-01
5.89123480e-02 7.67843723e-01 9.91620719e-01 -9.15295854e-02
3.64869475e-01 -7.22601190e-02 -5.04788570e-03 7.96435773e-01
1.00351743e-01 -1.22349560e-01 -6.20303392e-01 -2.72379696e-01
8.08045745e-01 -4.85864840e-03 -2.80519843e-01 -8.21169555e-01
-1.26709223e+00 8.48618448e-01 6.06695235e-01 3.22014838e-01
-3.50292981e-01 3.21534425e-02 2.68231422e-01 2.58614570e-01
2.69735277e-01 8.53837669e-01 -5.25390923e-01 -5.18360659e-02
-1.00352025e+00 6.25563741e-01 6.52204573e-01 9.22261477e-01
8.72166753e-01 -1.61729947e-01 6.36984631e-02 7.06580281e-01
-2.35638455e-01 -9.29565579e-02 4.56554949e-01 -1.05016601e+00
2.37435281e-01 1.00131762e+00 1.17940389e-01 -9.58251357e-01
-3.53071034e-01 -5.36593258e-01 -9.34804142e-01 2.59856701e-01
5.27919650e-01 4.77021068e-01 -8.68189991e-01 2.00227141e+00
3.52429807e-01 1.95731685e-01 -1.41587093e-01 7.06424057e-01
9.57328305e-02 5.54183722e-01 -5.82068622e-01 2.84096867e-01
1.17324638e+00 -9.05272961e-01 -3.44103277e-01 8.48611817e-03
6.91921890e-01 -7.91765988e-01 1.08412969e+00 3.16045493e-01
-1.04665101e+00 -3.48934889e-01 -1.09751952e+00 1.56328119e-02
-1.21760838e-01 -1.74774408e-01 5.51402330e-01 1.93618014e-01
-9.24481928e-01 9.45779562e-01 -8.75323951e-01 -2.47807741e-01
2.53610224e-01 2.61563063e-01 -5.09489894e-01 -7.16715381e-02
-7.68177390e-01 8.11019897e-01 1.69294462e-01 -2.04358861e-01
-6.40872538e-01 -8.60187352e-01 -7.11144865e-01 9.29130539e-02
3.00387830e-01 -8.27477276e-01 1.17900026e+00 -1.37569451e+00
-1.14176023e+00 5.66543281e-01 -3.58806014e-01 -3.14870268e-01
6.85085535e-01 1.39515817e-01 -6.64747804e-02 -9.76503789e-02
2.00264066e-01 6.09404147e-01 8.52993190e-01 -1.15805984e+00
-5.31483114e-01 -1.18665740e-01 5.21157265e-01 -4.47072685e-02
-1.78971857e-01 -2.60746330e-01 -2.61843383e-01 -6.95446551e-01
5.01124144e-01 -1.15682817e+00 -8.84911790e-02 3.61740351e-01
-4.23376471e-01 -1.65099189e-01 9.67726946e-01 -4.87136245e-01
9.88787711e-01 -1.99992907e+00 6.22329712e-01 5.10925651e-01
2.78872937e-01 1.50776142e-02 -1.10301852e-01 4.08020616e-01
-2.04815090e-01 1.54821098e-01 -3.20750862e-01 -2.23780200e-01
8.89144614e-02 5.36778748e-01 -5.32980680e-01 4.01388347e-01
3.21522444e-01 5.43582082e-01 -7.51166999e-01 9.48614180e-02
2.72730254e-02 4.14412975e-01 -1.17459178e+00 6.69928864e-02
-4.38254565e-01 5.11936903e-01 -4.28646445e-01 4.54381108e-01
7.50102282e-01 -2.39012152e-01 1.89638749e-01 -4.33117658e-01
-1.46149054e-01 7.65237451e-01 -1.56946504e+00 1.32306957e+00
-5.02459466e-01 3.04491311e-01 -1.92823663e-01 -1.18477952e+00
5.97655177e-01 1.62303656e-01 5.13013780e-01 -4.56486613e-01
-3.61661464e-01 4.99800622e-01 2.08070397e-01 -1.95243135e-01
3.18062633e-01 -1.13689631e-01 2.13516667e-01 3.87866735e-01
2.07759365e-02 -4.56832675e-03 2.61625618e-01 -6.34724423e-02
1.33190703e+00 2.57700980e-01 4.82396424e-01 -4.96075690e-01
4.44995314e-01 -1.27938375e-01 9.49784696e-01 5.97597063e-01
1.09887429e-01 9.99332011e-01 7.98388302e-01 -1.09022343e+00
-1.70746148e+00 -8.82425368e-01 -2.88999885e-01 5.92902780e-01
-1.02317929e-01 -7.26071417e-01 -8.13350201e-01 -6.29368842e-01
6.90167248e-02 3.35817546e-01 -8.04429710e-01 -3.13495666e-01
-8.07809353e-01 -3.53841156e-01 3.41299117e-01 4.76202577e-01
6.17955208e-01 -7.72962868e-01 -3.67644489e-01 4.62515429e-02
-1.94859073e-01 -1.10823214e+00 -5.99082053e-01 -1.70532223e-02
-1.00317550e+00 -1.34210396e+00 -3.60466450e-01 -4.98563200e-01
1.02016926e+00 6.28671423e-02 1.08530545e+00 2.22889245e-01
9.08904150e-02 -2.18319031e-03 -1.48267984e-01 1.72246397e-01
-4.87104207e-01 1.15679704e-01 1.73945487e-01 3.28429967e-01
-4.56932373e-03 -9.08564925e-01 -4.92258966e-01 4.54302162e-01
-9.78739500e-01 2.86677986e-01 4.73420292e-01 1.01680744e+00
6.51162207e-01 5.54781361e-03 9.78595614e-02 -8.56149614e-01
1.09496541e-01 -3.31350416e-01 -5.58923483e-01 -1.89003833e-02
-1.76860034e-01 6.68284833e-01 9.54492092e-01 -4.99238878e-01
-4.62722033e-01 2.29392067e-01 1.79545414e-02 -7.79497683e-01
-4.72292537e-03 4.97978956e-01 -2.52941519e-01 -1.44192174e-01
7.67900169e-01 2.19993293e-01 -6.71037734e-02 -6.00564957e-01
3.02714333e-02 1.63138568e-01 3.06737065e-01 -9.00982380e-01
1.05903101e+00 6.83027804e-01 2.17831746e-01 -8.02266836e-01
-9.19859469e-01 -1.40557855e-01 -7.73383200e-01 1.57088712e-01
6.22719586e-01 -6.81187749e-01 -2.60002077e-01 4.11379457e-01
-1.25313699e+00 -1.88330933e-01 -4.60475028e-01 2.65096009e-01
-7.07947671e-01 2.99459219e-01 -2.49611184e-01 -3.13154727e-01
2.15747178e-01 -1.44418752e+00 1.06053448e+00 -2.27716774e-01
-6.33124352e-01 -7.84550309e-01 1.86126679e-01 5.83437867e-02
4.81744617e-01 3.14795941e-01 1.26458669e+00 -6.22344494e-01
-9.30919826e-01 -3.16024333e-01 6.70283064e-02 2.82601088e-01
2.69909412e-01 1.08252212e-01 -8.14491749e-01 -2.19573036e-01
-5.81736080e-02 -2.12603271e-01 7.65066981e-01 2.22045079e-01
1.33695972e+00 -7.48786867e-01 -1.02804415e-01 9.11372423e-01
9.33427989e-01 -2.13861018e-01 5.54982841e-01 4.92521226e-01
7.86345959e-01 5.46587825e-01 -6.89618438e-02 1.65795714e-01
3.00980508e-01 1.14244461e+00 4.30853456e-01 3.76164019e-02
-8.83820653e-02 -5.22425354e-01 4.85291123e-01 4.32113767e-01
-3.62741321e-01 3.06821853e-01 -7.01877832e-01 4.40537959e-01
-1.89364576e+00 -1.01316178e+00 3.17503005e-01 2.36127543e+00
7.03649104e-01 2.40030095e-01 2.76658386e-01 1.97455868e-01
4.71712142e-01 3.46819580e-01 -5.39318919e-01 -4.62130070e-01
-2.26590887e-01 8.51734281e-02 2.50557542e-01 5.62247455e-01
-1.06922615e+00 7.30566025e-01 6.21010828e+00 6.20448828e-01
-1.11890340e+00 -2.68136948e-01 3.65962654e-01 1.32705877e-03
-6.18228078e-01 4.81626570e-01 -5.66371083e-01 2.70833611e-01
4.81292188e-01 8.06754753e-02 8.59890163e-01 8.44753027e-01
7.38545088e-03 4.00629699e-01 -1.51320279e+00 5.15283346e-01
-5.12967817e-02 -1.77406764e+00 5.96756339e-01 2.94194788e-01
8.86171043e-01 -1.05885483e-01 -6.81332946e-02 1.29359514e-01
1.35199398e-01 -9.04030681e-01 1.04357445e+00 4.28813875e-01
3.66036594e-01 -7.33671248e-01 4.10115212e-01 4.70414847e-01
-1.23332751e+00 -8.78999159e-02 -4.99641031e-01 -7.75199234e-02
-1.30015299e-01 4.23159003e-01 -3.20572585e-01 4.31163639e-01
5.14733136e-01 8.08510959e-01 -3.23333085e-01 8.48359108e-01
-2.49231830e-01 4.52594489e-01 -5.59489965e-01 5.25347233e-01
3.10738295e-01 -3.50457996e-01 8.75079572e-01 7.43515730e-01
4.70149159e-01 -1.19109347e-03 2.35180482e-01 1.21122634e+00
-1.63915008e-01 -2.00936913e-01 -8.57457042e-01 1.44979730e-01
3.17352563e-01 9.75512683e-01 -6.27771735e-01 -2.05923408e-01
-4.98408079e-01 7.92215645e-01 5.37661612e-01 2.79739857e-01
-6.20863676e-01 1.13289535e-01 1.10450613e+00 5.74985206e-01
5.54749727e-01 -3.13166738e-01 -1.48913056e-01 -1.60523880e+00
4.89380360e-01 -1.14760494e+00 3.94090295e-01 -4.59389865e-01
-1.05951786e+00 3.81143212e-01 1.95085711e-03 -1.57259834e+00
-3.54118615e-01 -6.92427099e-01 -9.15170789e-01 5.77315927e-01
-1.19698179e+00 -1.38381684e+00 2.67657395e-02 5.21083891e-01
4.72504526e-01 -2.09346980e-01 8.72613311e-01 1.03042997e-01
-2.94982433e-01 6.65119410e-01 6.12643734e-02 2.56812703e-02
3.96911830e-01 -1.23123753e+00 4.91759270e-01 9.85315681e-01
4.78257656e-01 7.61565328e-01 8.96802783e-01 -4.54203665e-01
-1.41608775e+00 -1.10845208e+00 1.01319218e+00 -5.44806600e-01
6.96969867e-01 -6.47944331e-01 -1.02877080e+00 1.11877930e+00
-1.72700658e-01 3.37217748e-01 1.76677421e-01 1.86520725e-01
-8.79897654e-01 -5.79006746e-02 -9.40834463e-01 9.66477394e-01
1.38698733e+00 -5.92555046e-01 -4.34906274e-01 5.24140179e-01
4.91191179e-01 -5.69905877e-01 -5.99386811e-01 5.04731894e-01
4.75138366e-01 -1.19012642e+00 8.64620924e-01 -9.23866153e-01
7.21203327e-01 -6.18266106e-01 -2.13909492e-01 -1.32842469e+00
-3.94250095e-01 -6.13364935e-01 -2.74714410e-01 9.09844160e-01
3.48402679e-01 -5.72235525e-01 7.75415659e-01 5.08896470e-01
-4.33026791e-01 -7.70218551e-01 -1.28899634e+00 -9.97818947e-01
1.55795470e-01 -1.71283856e-01 8.57071877e-01 1.22842133e+00
-1.23645380e-01 6.77657947e-02 -4.07955885e-01 2.45872676e-01
2.96228260e-01 2.93669134e-01 1.09707630e+00 -1.08436775e+00
-4.92107272e-01 -6.82323098e-01 -8.78293693e-01 -9.69177604e-01
3.79380018e-01 -1.19036937e+00 -3.02990675e-01 -1.01057935e+00
2.57915527e-01 -2.29851440e-01 -2.94000469e-02 6.29482329e-01
1.65386990e-01 1.10360071e-01 -3.67198102e-02 4.43391979e-01
-1.90136790e-01 7.44276464e-01 1.10928154e+00 -1.75528929e-01
-1.04802020e-01 7.80678242e-02 -6.08832538e-01 1.00717747e+00
4.68905419e-01 -3.94784123e-01 -2.19256490e-01 -3.43950987e-01
3.20573509e-01 -3.63332480e-01 6.17538929e-01 -1.10525048e+00
8.89206678e-02 -2.22080693e-01 3.39204639e-01 -1.66771606e-01
1.79268420e-01 -7.65388489e-01 3.98004949e-01 4.34410095e-01
-3.81599605e-01 1.87141478e-01 -2.62229200e-02 3.43580961e-01
-2.92029530e-01 2.42936146e-02 9.68099952e-01 -4.03764285e-02
-4.24499452e-01 4.08131629e-01 -1.33072540e-01 8.39138106e-02
6.69908106e-01 -2.35626474e-01 -1.14676602e-01 -4.34696108e-01
-6.08276606e-01 -5.28124459e-02 8.52420747e-01 4.39470202e-01
3.22398812e-01 -1.63512313e+00 -6.14282072e-01 3.68666917e-01
3.80033702e-01 2.41721958e-01 -9.09113139e-02 7.51033306e-01
-3.29086363e-01 9.42234918e-02 -4.37742949e-01 -6.83159113e-01
-9.07942533e-01 3.83551776e-01 8.09254110e-01 -3.92214537e-01
-9.94840145e-01 2.75981218e-01 5.18890500e-01 -7.74397254e-01
-3.08346208e-02 -5.37539244e-01 1.86891817e-02 -2.59337664e-01
2.34163165e-01 2.12720588e-01 1.62703156e-01 -9.12784278e-01
-2.50202119e-01 8.76657486e-01 -1.39249146e-01 5.18715428e-03
1.35761487e+00 3.26093376e-01 -2.71263093e-01 3.28976512e-01
1.13095844e+00 2.80231684e-01 -1.29800427e+00 -4.30322349e-01
-1.37717780e-02 -4.82803404e-01 -3.77949655e-01 -2.37380549e-01
-1.00480080e+00 6.42864287e-01 1.56411843e-03 1.91203542e-02
8.37842286e-01 3.73714082e-02 5.59294939e-01 6.35325670e-01
3.14571291e-01 -8.54946017e-01 -3.37348543e-02 7.66336620e-01
1.11495733e+00 -9.62039173e-01 -4.31943797e-02 -2.17243910e-01
-2.90227324e-01 1.29257667e+00 5.57934582e-01 -6.78763211e-01
5.68737328e-01 1.01300739e-01 -2.85204351e-01 -2.13905796e-01
-7.28612006e-01 1.32174313e-01 6.97740018e-01 2.46371821e-01
3.22172403e-01 -6.98074996e-02 -1.29868984e-01 3.31742883e-01
-6.70733392e-01 -2.80525535e-01 4.92173225e-01 6.86602056e-01
-2.58240104e-01 -1.24602437e+00 -1.89975068e-01 3.43750358e-01
-1.73918039e-01 1.69446558e-01 -6.30741596e-01 9.10596073e-01
1.46581113e-01 2.24014699e-01 1.85703069e-01 -4.87973839e-01
6.24349236e-01 1.48756400e-01 4.05914277e-01 -4.49332088e-01
-3.07689965e-01 -1.71015397e-01 -5.26443794e-02 -8.86291444e-01
-2.64732271e-01 -8.51523578e-01 -1.10992336e+00 -3.62364113e-01
2.45205715e-01 5.70162721e-02 3.07612807e-01 1.23393404e+00
4.03615177e-01 1.68681443e-01 8.12329233e-01 -9.83183086e-01
-6.96167707e-01 -5.92073262e-01 -3.39270741e-01 8.63655269e-01
6.23412788e-01 -8.67847800e-01 -5.68077862e-01 -8.90930817e-02] | [8.985289573669434, 2.38551664352417] |
16cfedef-1dc7-49e7-b0c3-1ec2c7756bf3 | a-machine-transliteration-tool-between-uzbek | 2205.09578 | null | https://arxiv.org/abs/2205.09578v1 | https://arxiv.org/pdf/2205.09578v1.pdf | A machine transliteration tool between Uzbek alphabets | Machine transliteration, as defined in this paper, is a process of automatically transforming written script of words from a source alphabet into words of another target alphabet within the same language, while preserving their meaning, as well as pronunciation. The main goal of this paper is to present a machine transliteration tool between three common scripts used in low-resource Uzbek language: the old Cyrillic, currently official Latin, and newly announced New Latin alphabets. The tool has been created using a combination of rule-based and fine-tuning approaches. The created tool is available as an open-source Python package, as well as a web-based application including a public API. To our knowledge, this is the first machine transliteration tool that supports the newly announced Latin alphabet of the Uzbek language. | ['Carlos Gómez-Rodríguez', 'Elmurod Kuriyozov', 'Ulugbek Salaev'] | 2022-05-19 | null | null | null | null | ['transliteration'] | ['natural-language-processing'] | [ 2.20664725e-01 3.37652825e-02 1.13358758e-01 -2.73804367e-01
-3.60873073e-01 -1.15956366e+00 1.05721772e+00 -2.52119023e-02
-5.13781190e-01 1.09635079e+00 1.67782471e-01 -1.03124154e+00
-3.94364111e-02 -7.29131162e-01 -3.58745039e-01 -3.08977932e-01
5.21615744e-01 8.16809416e-01 1.87653765e-01 -5.66776156e-01
2.86416471e-01 7.90373564e-01 -1.43493545e+00 1.14220232e-01
1.13306069e+00 1.64873704e-01 3.41347754e-01 8.02533090e-01
-8.93675089e-02 6.40329301e-01 -4.96894568e-01 -4.61575717e-01
2.64036566e-01 -8.23381245e-01 -1.13700259e+00 -3.31503332e-01
3.43997717e-01 -9.96466726e-02 1.09929450e-01 7.49195755e-01
3.02261055e-01 1.89514440e-02 5.41949093e-01 -4.95472103e-01
-6.14290714e-01 7.80528784e-01 3.34247619e-01 -4.32317220e-02
4.61087048e-01 2.20867500e-01 7.92404890e-01 -7.46426761e-01
8.07806194e-01 1.14986634e+00 5.67311108e-01 3.34069401e-01
-1.08516705e+00 -6.47195578e-01 -5.06689310e-01 -1.97498292e-01
-1.35479748e+00 -4.17727470e-01 4.79895547e-02 -8.13642204e-01
1.24899745e+00 4.42098320e-01 6.79660738e-01 8.14050853e-01
4.05769110e-01 -3.14534940e-02 1.72517228e+00 -9.30449545e-01
7.49035925e-02 4.75549340e-01 -2.65403651e-03 7.63110697e-01
4.51325744e-01 2.91111190e-02 -4.63188261e-01 -2.51159936e-01
6.26256347e-01 -7.65030026e-01 1.71903566e-01 2.92980433e-01
-1.18119681e+00 7.05924451e-01 -4.49188143e-01 7.33778000e-01
-1.69434935e-01 -1.94092080e-01 3.00576538e-01 4.11544800e-01
5.43484449e-01 6.18022263e-01 -6.94072604e-01 -7.35013008e-01
-9.72187281e-01 3.74117345e-01 1.07543123e+00 8.93763244e-01
8.96364868e-01 2.20291778e-01 1.34937704e-01 7.52491832e-01
3.85247052e-01 7.33588636e-01 7.32214212e-01 -4.33096588e-01
2.13962838e-01 5.17269254e-01 1.05997600e-01 -3.43944520e-01
-1.93229124e-01 -1.73697785e-01 -1.57452762e-01 2.58616656e-01
7.03541040e-01 -2.91363329e-01 -8.93516064e-01 1.46965897e+00
4.45149332e-01 -4.70312923e-01 3.34524244e-01 4.67557102e-01
7.72888541e-01 5.89157760e-01 -2.09977522e-01 -2.58654296e-01
1.45168364e+00 -7.00058579e-01 -6.58922195e-01 1.12334423e-01
5.31758666e-01 -1.44589710e+00 1.08175159e+00 4.14322138e-01
-8.95862877e-01 -4.97063786e-01 -9.95600164e-01 -5.98536469e-02
-7.74164021e-01 2.88316905e-01 5.17516613e-01 1.38071609e+00
-1.11368465e+00 5.09590805e-01 -5.57794213e-01 -9.14832592e-01
-2.70999163e-01 1.56480208e-01 -4.75161910e-01 5.26233137e-01
-1.23035669e+00 1.21036863e+00 7.42267489e-01 -3.39774251e-01
-5.56880951e-01 -4.73541826e-01 -7.90858626e-01 -5.00642419e-01
1.47620961e-01 -2.10821226e-01 1.38258874e+00 -1.19766998e+00
-2.24391484e+00 1.38593519e+00 -2.94730384e-02 -1.03798606e-01
8.47481489e-01 -2.72096962e-01 -7.04436421e-01 -3.71209890e-01
2.42412791e-01 -1.34154171e-01 6.56886935e-01 -8.29893887e-01
-6.80952370e-01 -1.71788797e-01 -2.17100948e-01 -1.78323522e-01
1.58187240e-01 7.06603825e-01 -2.99273193e-01 -9.93510962e-01
-3.42793107e-01 -9.92657423e-01 1.50656089e-01 -6.20893478e-01
-1.50131091e-01 -1.16232857e-01 3.58669311e-01 -1.42111433e+00
1.37160957e+00 -1.75600016e+00 1.86701734e-02 3.76008958e-01
-5.10946870e-01 6.86536670e-01 2.34845221e-01 8.78229558e-01
-1.19543776e-01 6.48493022e-02 -4.61458772e-01 1.39290288e-01
3.38785760e-02 4.12000090e-01 -3.99936765e-01 5.64687371e-01
1.19985923e-01 6.08138084e-01 -1.21194160e+00 -2.00200647e-01
3.84360462e-01 2.55460113e-01 -9.60047469e-02 -3.07295527e-02
-3.60196114e-01 6.82096660e-01 -3.10553070e-02 4.12178606e-01
6.75753474e-01 5.98483324e-01 4.87698615e-01 4.95537221e-01
-1.04229283e+00 6.40876412e-01 -1.18110895e+00 1.49399745e+00
-6.27717495e-01 4.69915003e-01 -3.79551172e-01 -3.05994928e-01
1.30061650e+00 3.71912748e-01 -1.10466443e-01 -4.06230420e-01
2.64404863e-01 9.41819370e-01 -1.08698895e-02 -4.20343250e-01
1.03124273e+00 -1.87290743e-01 -9.05744210e-02 7.88368762e-01
1.00642800e-01 -4.97682273e-01 6.07407749e-01 -2.20946744e-01
6.11998200e-01 8.46397519e-01 1.03952837e+00 -6.93057120e-01
8.43442440e-01 1.24859020e-01 3.73829901e-01 8.43138754e-01
2.49611408e-01 3.54811072e-01 3.68679017e-01 -4.63512868e-01
-1.45559621e+00 -9.62332129e-01 -4.61046010e-01 9.29854691e-01
-7.16069221e-01 -5.58393002e-01 -9.83722746e-01 -3.40503097e-01
-1.84087440e-01 9.02612865e-01 -3.28431278e-01 2.98346311e-01
-8.39726448e-01 -3.52574199e-01 1.22042549e+00 -3.01190674e-01
4.37451690e-01 -1.14769769e+00 -5.91242254e-01 3.49156886e-01
5.67965629e-03 -9.03093457e-01 -3.36273074e-01 -7.66290501e-02
-4.06983674e-01 -8.33229184e-01 -6.21236265e-01 -7.32164204e-01
3.79481614e-01 -4.91377681e-01 8.52189243e-01 -1.19224161e-01
2.85624675e-02 4.92531713e-03 -6.04976594e-01 -6.48158073e-01
-1.23130131e+00 4.28120613e-01 2.30360508e-01 6.62802532e-02
2.32945636e-01 -3.68012160e-01 2.51928777e-01 -4.47666571e-02
-1.08117807e+00 1.87756851e-01 3.18121314e-01 3.13466668e-01
2.07106262e-01 -3.32203358e-01 5.22418767e-02 -1.01688170e+00
5.87729812e-01 -2.93042600e-01 -1.03107905e+00 3.16903710e-01
-6.61276758e-01 2.31120288e-01 9.31833267e-01 -1.85783908e-01
-1.04917431e+00 1.78402215e-01 -5.24453521e-01 4.98149723e-01
-2.50060678e-01 5.30515015e-01 -2.71186411e-01 -7.06150010e-02
6.33550882e-01 4.84635234e-01 -1.58852354e-01 -8.58094156e-01
5.20465374e-01 1.15600109e+00 7.31375277e-01 -7.15240717e-01
9.06625569e-01 3.03069711e-01 -2.83292979e-01 -1.01447141e+00
-1.69270813e-01 -3.00759047e-01 -9.27252352e-01 -1.87176019e-01
6.14438891e-01 -5.80214620e-01 -1.47241116e-01 9.11343873e-01
-1.43511307e+00 -5.41997135e-01 -2.19357133e-01 4.50157404e-01
-5.21606743e-01 3.85776192e-01 -1.32378474e-01 -6.30741537e-01
-5.31355619e-01 -9.34439421e-01 7.90217936e-01 2.15462506e-01
-7.83706844e-01 -9.24813211e-01 8.47595215e-01 -1.93157658e-01
2.01315090e-01 1.99882135e-01 8.75128806e-01 -7.41945863e-01
8.61532912e-02 1.49218859e-02 6.13294281e-02 6.42137110e-01
6.13184154e-01 6.19790435e-01 -7.52066851e-01 -4.70839366e-02
-3.42992365e-01 1.76994666e-01 3.07941943e-01 -3.82933766e-01
8.89231190e-02 -4.92973596e-01 2.25487947e-01 6.58338368e-01
1.37275612e+00 3.40717375e-01 6.05774462e-01 7.61991084e-01
2.93905050e-01 5.16252995e-01 3.90699863e-01 5.22389591e-01
1.88516513e-01 7.46624112e-01 -4.21011031e-01 9.78618413e-02
-3.82332853e-03 -2.81653404e-01 7.25958943e-01 8.80262434e-01
-2.71068335e-01 5.50663657e-02 -1.40661347e+00 5.01593411e-01
-1.72215855e+00 -7.70076394e-01 -3.98962557e-01 2.47235990e+00
1.19124103e+00 -1.80320263e-01 5.64651191e-02 1.17576279e-01
4.31431860e-01 -2.73990724e-02 2.55350977e-01 -1.39539337e+00
-2.55892515e-01 1.12779737e+00 6.72306180e-01 1.22705042e+00
-1.01996720e+00 1.65126681e+00 6.49261475e+00 6.86829627e-01
-1.44218504e+00 2.46313423e-01 -4.64278013e-02 3.36768270e-01
-3.57733488e-01 4.95696455e-01 -8.64381909e-01 3.04147363e-01
1.14275825e+00 -5.37350833e-01 6.78623676e-01 6.10646009e-01
7.00251758e-01 -1.90288778e-02 -7.89983988e-01 5.70668876e-01
2.06391692e-01 -1.29842281e+00 1.14209011e-01 -8.61927494e-02
7.85985172e-01 1.00446448e-01 -3.26125383e-01 1.26709193e-01
5.96920133e-01 -9.67915893e-01 1.26473105e+00 6.84459448e-01
1.01627564e+00 -7.65325785e-01 5.72430551e-01 3.40040736e-02
-9.62863803e-01 4.95582938e-01 -1.95136480e-02 -5.34716845e-01
-5.37717575e-03 5.13973892e-01 -8.88870239e-01 7.33086109e-01
3.98084760e-01 3.71244997e-01 -7.09641635e-01 6.30299568e-01
-5.77693939e-01 8.37744474e-01 -1.25972614e-01 -1.59056231e-01
3.08549225e-01 -6.55460119e-01 6.37005210e-01 1.71707630e+00
4.64455992e-01 -2.83482552e-01 -5.94278984e-02 4.90265548e-01
3.69529903e-01 7.91543603e-01 -5.71895361e-01 -2.66097486e-01
1.30541071e-01 9.99546111e-01 -6.56750500e-01 -5.65148294e-01
-1.83186501e-01 1.20175874e+00 1.58880919e-01 2.39396751e-01
-7.43905544e-01 -7.05860257e-01 8.82950008e-01 1.18327290e-01
1.33254334e-01 -4.86304402e-01 -4.26611662e-01 -9.90853190e-01
2.72703450e-02 -1.38792729e+00 2.06485540e-01 -5.82185090e-01
-5.76233506e-01 7.21311629e-01 1.61040127e-01 -9.58413899e-01
-4.50807869e-01 -8.30270946e-01 -2.54359305e-01 1.36524057e+00
-8.85506153e-01 -1.25149238e+00 9.88733023e-02 1.25445575e-01
2.14647174e-01 -5.36157191e-01 1.15984917e+00 1.99064121e-01
-2.73561627e-01 4.90674853e-01 7.95970500e-01 2.98839584e-02
7.55919516e-01 -1.02442229e+00 7.20385015e-01 1.13286090e+00
-2.86135897e-02 8.95212770e-01 9.88420129e-01 -8.82166028e-01
-9.08196390e-01 -1.09296572e+00 1.90828168e+00 -3.71992856e-01
9.10044134e-01 -6.05527699e-01 -5.48074424e-01 9.14228678e-01
4.10325646e-01 -7.90016472e-01 6.68185472e-01 -3.30070496e-01
-3.56021851e-01 1.09163076e-01 -1.03253591e+00 8.71535063e-01
6.57436490e-01 -7.30412304e-01 -6.32462025e-01 5.86034119e-01
4.65545654e-01 -5.61809242e-01 -8.52369905e-01 -7.92508647e-02
9.97212768e-01 -8.39669466e-01 4.19855684e-01 -5.47826827e-01
2.35680506e-01 -4.90483880e-01 -1.24853551e-02 -1.29180360e+00
-1.55119523e-01 -1.41342044e+00 4.51899022e-01 1.54659998e+00
5.25516987e-01 -1.21517372e+00 7.59002641e-02 7.25110844e-02
-1.75850883e-01 2.19306070e-02 -1.26488519e+00 -9.48676765e-01
3.32212687e-01 -5.58606625e-01 7.88071990e-01 9.87208605e-01
1.70574337e-02 1.35073543e-01 -5.04921854e-01 6.62883744e-02
-5.25257774e-02 -3.03877741e-01 1.13101125e+00 -9.61417139e-01
-5.25566220e-01 -5.43139100e-01 -4.39189672e-01 -3.65907282e-01
1.90200374e-01 -1.17712975e+00 -4.18546349e-02 -1.23692429e+00
-2.96613514e-01 -3.57634276e-01 2.64580607e-01 6.01041198e-01
5.14961854e-02 3.44010860e-01 6.07092567e-02 1.06294356e-01
4.05727804e-01 4.10803854e-02 8.19148719e-01 1.28394991e-01
-4.97588009e-01 -3.28342989e-02 -4.85068172e-01 5.28119981e-01
1.08498490e+00 -7.09131539e-01 -4.53660227e-02 -3.01755905e-01
3.80407691e-01 -7.99503565e-01 -1.48636550e-01 -8.12804997e-01
-3.49673063e-01 -5.59972763e-01 -2.24552572e-01 -2.54147679e-01
-2.37530485e-01 -3.83767843e-01 6.30005181e-01 8.14850211e-01
7.37948790e-02 4.22934204e-01 4.25661951e-01 -4.00114268e-01
1.07410893e-01 -4.66407210e-01 1.02727079e+00 -2.81775832e-01
-7.87474394e-01 -1.71966463e-01 -1.18388259e+00 -1.09467126e-01
1.11629093e+00 -2.58067816e-01 -2.66041458e-01 3.67360339e-02
-2.60954797e-01 -5.04172504e-01 8.11799109e-01 6.25311315e-01
1.34424135e-01 -9.56793904e-01 -1.15851760e+00 5.08901715e-01
3.21162552e-01 -7.47141063e-01 -4.12828833e-01 6.07088268e-01
-1.64640808e+00 8.22881579e-01 -5.74917614e-01 -8.49299971e-03
-1.39894521e+00 1.15734264e-01 5.25495410e-01 -2.80265033e-01
-4.74051654e-01 3.51736009e-01 -4.40497190e-01 -1.23927498e+00
-3.71362984e-01 -3.08023095e-01 -9.43997651e-02 -2.48109102e-01
6.24689698e-01 3.60010415e-01 2.79634416e-01 -1.18206131e+00
-5.07120311e-01 4.69168782e-01 1.86628014e-01 -6.40289426e-01
8.63328159e-01 9.66971293e-02 -8.18289876e-01 6.93208694e-01
6.66528165e-01 6.59557581e-01 -2.95519859e-01 -6.69292212e-02
1.94282606e-01 -4.00858372e-01 -4.52607036e-01 -7.38018930e-01
-8.70194808e-02 2.82580763e-01 4.39100206e-01 -2.26430416e-01
7.14154005e-01 -1.84028253e-01 3.40646029e-01 4.25992936e-01
3.18159312e-01 -1.32241809e+00 -8.27248096e-01 1.32318664e+00
7.92980909e-01 -5.09900689e-01 4.68793558e-03 -2.36976624e-01
-4.01198059e-01 1.30073237e+00 1.44066930e-01 1.93318292e-01
4.17904139e-01 2.19424233e-01 4.97727096e-01 4.13998991e-01
-3.50594312e-01 -4.32976872e-01 7.66848102e-02 7.44980216e-01
9.36532080e-01 3.72103482e-01 -1.35122657e+00 3.72210294e-01
-9.42487299e-01 1.01504199e-01 6.44361556e-01 9.77339029e-01
-2.87258327e-01 -1.76617873e+00 -7.55566716e-01 1.94228396e-01
-5.18383145e-01 -6.00668669e-01 -9.23659980e-01 8.68647397e-01
6.28433883e-01 9.10971761e-01 -1.79937277e-02 -3.04447919e-01
1.74068674e-01 2.59619921e-01 5.69376826e-01 -6.86979711e-01
-1.15908945e+00 -3.44444722e-01 5.53358495e-01 -1.61060430e-02
-3.32332402e-02 -7.99728215e-01 -1.11650264e+00 -7.26262867e-01
2.39774480e-01 3.07470828e-01 8.51390898e-01 1.25994647e+00
-8.74912366e-02 8.78751576e-02 4.00225163e-01 -4.89592761e-01
-2.93641567e-01 -1.16230547e+00 -5.50113261e-01 8.84364918e-02
-8.38763118e-02 -6.73500588e-03 4.45969403e-02 2.23513290e-01] | [10.60546588897705, 10.427474975585938] |
690fb0a9-fde2-4e4b-bb22-ee870902fe89 | cluster-induced-mask-transformers-for | 2307.04525 | null | https://arxiv.org/abs/2307.04525v1 | https://arxiv.org/pdf/2307.04525v1.pdf | Cluster-Induced Mask Transformers for Effective Opportunistic Gastric Cancer Screening on Non-contrast CT Scans | Gastric cancer is the third leading cause of cancer-related mortality worldwide, but no guideline-recommended screening test exists. Existing methods can be invasive, expensive, and lack sensitivity to identify early-stage gastric cancer. In this study, we explore the feasibility of using a deep learning approach on non-contrast CT scans for gastric cancer detection. We propose a novel cluster-induced Mask Transformer that jointly segments the tumor and classifies abnormality in a multi-task manner. Our model incorporates learnable clusters that encode the texture and shape prototypes of gastric cancer, utilizing self- and cross-attention to interact with convolutional features. In our experiments, the proposed method achieves a sensitivity of 85.0% and specificity of 92.6% for detecting gastric tumors on a hold-out test set consisting of 100 patients with cancer and 148 normal. In comparison, two radiologists have an average sensitivity of 73.5% and specificity of 84.3%. We also obtain a specificity of 97.7% on an external test set with 903 normal cases. Our approach performs comparably to established state-of-the-art gastric cancer screening tools like blood testing and endoscopy, while also being more sensitive in detecting early-stage cancer. This demonstrates the potential of our approach as a novel, non-invasive, low-cost, and accurate method for opportunistic gastric cancer screening. | ['Ling Zhang', 'Zaiyi Liu', 'Li Zhang', 'Le Lu', 'Bin Dong', 'Jingren Zhou', 'Hexin Dong', 'Mingyan Qiu', 'Junli Wang', 'Jiawen Yao', 'Xin Chen', 'Yingda Xia', 'Mingze Yuan'] | 2023-07-10 | null | null | null | null | ['specificity'] | ['natural-language-processing'] | [-6.01117834e-02 2.49303635e-02 -3.14965427e-01 -6.18555993e-02
-1.24216366e+00 -3.41282636e-01 1.73857734e-01 5.01273692e-01
-5.41472495e-01 2.13764414e-01 -1.94013447e-01 -4.91590619e-01
3.39548290e-01 -9.90774691e-01 -4.85574275e-01 -9.42004502e-01
-3.49010736e-01 5.36362648e-01 4.54949796e-01 1.65702567e-01
-4.12249118e-01 1.92262217e-01 -1.17594790e+00 6.54046297e-01
9.99332845e-01 1.06241059e+00 2.17822745e-01 7.08665192e-01
3.73834461e-01 4.61987585e-01 -2.91589618e-01 -3.12816381e-01
1.29032433e-01 -5.77006400e-01 -6.45062506e-01 1.51242957e-01
2.54962146e-02 -4.55235422e-01 -4.53070216e-02 9.81018066e-01
3.95342022e-01 -4.48643178e-01 6.49190366e-01 -4.61709887e-01
-4.71728086e-01 2.67039418e-01 -7.62275636e-01 2.29858279e-01
1.62729174e-01 4.29043639e-03 6.93403304e-01 -6.38434350e-01
2.73122549e-01 5.15617013e-01 1.03989172e+00 6.53005123e-01
-1.09586453e+00 -5.33783019e-01 -2.01254606e-01 -4.46899273e-02
-1.25753391e+00 -1.63574219e-02 2.25786403e-01 -1.38765320e-01
7.00228751e-01 4.42907274e-01 1.13446426e+00 7.79276431e-01
4.45545107e-01 8.48218858e-01 8.41724873e-01 -4.00550216e-01
7.24958330e-02 2.75322080e-01 -1.97942927e-01 1.24954629e+00
8.93334866e-01 2.22656112e-02 1.65579230e-01 -5.55756748e-01
8.08050096e-01 4.51330572e-01 -2.66465038e-01 -2.47594491e-01
-1.30534935e+00 1.16677785e+00 5.32054245e-01 5.76671779e-01
-4.23494607e-01 3.93569544e-02 5.40453196e-01 2.48715077e-02
3.34910512e-01 6.00806512e-02 -2.66264081e-01 4.38751817e-01
-8.15693557e-01 -3.71778876e-01 1.02355754e+00 3.20461899e-01
2.78216660e-01 -4.49505329e-01 -9.20435116e-02 7.05915451e-01
5.15104353e-01 5.04885435e-01 1.06212199e+00 -3.40665668e-01
-2.22057328e-01 8.55722189e-01 -2.74087302e-02 -7.92875528e-01
-7.49022543e-01 -6.05503023e-01 -1.12608111e+00 -1.88842416e-01
3.05139780e-01 -1.22845665e-01 -9.32956338e-01 1.27042079e+00
3.24447572e-01 1.31900266e-01 6.94452301e-02 9.17492986e-01
8.01030397e-01 2.06062362e-01 2.41772234e-01 5.07374778e-02
1.90830898e+00 -8.76200616e-01 -2.55093426e-01 -1.11628130e-01
1.10083687e+00 -4.71275419e-01 7.71557391e-01 2.36104622e-01
-8.18702638e-01 -2.46661484e-01 -9.08797443e-01 1.55156106e-01
-1.10946491e-01 5.80596507e-01 7.54628658e-01 1.04325116e+00
-1.16115391e+00 -1.31522873e-02 -1.42056596e+00 -6.85131788e-01
3.87953609e-01 5.46350121e-01 -3.12835962e-01 -2.15649962e-01
-9.12844360e-01 5.07107496e-01 5.11180222e-01 -1.62962675e-01
-8.84107113e-01 -5.96454084e-01 -8.06253731e-01 2.18998846e-02
2.81172156e-01 -1.01290667e+00 1.19853127e+00 -9.38588858e-01
-1.36009562e+00 1.22808659e+00 -5.87028079e-02 -5.98245382e-01
5.77910244e-01 5.52336648e-02 -3.30239564e-01 5.33623815e-01
2.19694883e-01 6.07191801e-01 2.03048766e-01 -7.07938075e-01
-1.08976781e+00 -3.04986268e-01 -3.33086759e-01 9.69824046e-02
-4.54625279e-01 -1.73125207e-01 -8.30315411e-01 -5.11394501e-01
1.86219156e-01 -1.13795316e+00 -6.30298853e-01 3.00128400e-01
-4.82594341e-01 1.95560157e-02 5.42510092e-01 -6.38391674e-01
1.02393150e+00 -2.06298566e+00 -4.29673731e-01 5.20761847e-01
3.47794890e-01 2.93657124e-01 1.37863636e-01 -1.29236817e-01
3.57084513e-01 5.09053290e-01 -6.33062348e-02 -3.66379209e-02
-3.83364379e-01 -1.66155428e-01 6.62005782e-01 7.28717864e-01
3.02827537e-01 1.10751998e+00 -1.06255829e+00 -6.73136771e-01
2.98436224e-01 7.22707391e-01 -5.73695779e-01 6.04438856e-02
6.70629740e-02 2.66571313e-01 -6.56708002e-01 1.08167374e+00
2.64315188e-01 -9.96594489e-01 5.93428195e-01 -2.12169737e-01
2.08996058e-01 -6.87330514e-02 -7.94489324e-01 1.44754493e+00
-2.74146080e-01 2.62165874e-01 5.67757040e-02 -9.21133459e-01
6.57672882e-01 6.12925470e-01 6.25202179e-01 -6.83089375e-01
1.48206562e-01 6.26558483e-01 1.56807259e-01 -6.69180930e-01
-1.79104730e-01 -1.82375953e-01 -4.04864624e-02 2.54821211e-01
-2.03661144e-01 2.94507056e-01 2.00095654e-01 -1.41853884e-01
1.25161314e+00 -3.73875022e-01 7.56774306e-01 -5.55890143e-01
6.36721432e-01 2.25110278e-01 4.80060697e-01 8.67902517e-01
-1.94237962e-01 5.71817756e-01 3.00581664e-01 -6.53970838e-01
-6.08253598e-01 -8.45770717e-01 -3.38658661e-01 6.36340022e-01
4.83694412e-02 5.36775105e-02 -5.47104478e-01 -7.31392622e-01
8.45480487e-02 1.13016151e-01 -9.51495230e-01 -1.05333380e-01
-5.56425452e-01 -1.37822819e+00 5.11075020e-01 4.74813968e-01
7.25726724e-01 -6.11458957e-01 -8.15848470e-01 4.59541231e-01
-2.83200562e-01 -7.29768395e-01 -3.24162394e-01 2.42197216e-01
-1.00486588e+00 -1.62049365e+00 -1.17014539e+00 -1.31320107e+00
8.70715082e-01 4.07855034e-01 1.23981273e+00 5.93755782e-01
-6.31590009e-01 7.63818473e-02 -2.07784101e-01 -2.95788378e-01
-6.04940653e-01 1.74293995e-01 -4.52060550e-01 -2.16995608e-02
6.68407619e-01 1.99956283e-01 -1.23229849e+00 5.22160113e-01
-7.65759706e-01 7.14615136e-02 1.03007698e+00 9.96745586e-01
1.04443920e+00 -7.94734359e-02 3.78978252e-01 -9.99750197e-01
2.39241496e-01 -7.48597980e-01 -2.85353541e-01 2.52360463e-01
-4.54424322e-01 -1.89570367e-01 4.72121149e-01 -5.07607162e-01
-7.45473087e-01 3.89643162e-01 -9.65184793e-02 -8.20012856e-03
-1.73249543e-01 6.83382273e-01 4.46488827e-01 -2.69775838e-01
6.65692568e-01 1.30277440e-01 4.00908917e-01 -1.72860354e-01
-2.96906859e-01 4.45609659e-01 3.94069761e-01 -2.61775777e-02
3.57929021e-01 8.67309749e-01 7.39013106e-02 -6.12942278e-01
-6.84068918e-01 -1.01879585e+00 -2.84471422e-01 2.38580495e-01
1.00615263e+00 -1.31198394e+00 -3.84766042e-01 5.12315094e-01
-4.18505788e-01 -3.92544061e-01 1.52381837e-01 6.77775085e-01
-1.49909869e-01 4.46862847e-01 -1.06746233e+00 -3.44123870e-01
-8.22147012e-01 -1.20315623e+00 1.23772144e+00 1.26632303e-01
-1.44949734e-01 -1.35567760e+00 3.61805111e-02 5.07546514e-02
5.85677624e-01 6.38618886e-01 7.04895258e-01 -9.24219668e-01
-4.24114645e-01 -6.81335866e-01 -4.41118062e-01 -8.94281790e-02
4.46369797e-01 -2.06989020e-01 -6.89693511e-01 -5.55153608e-01
-7.86320791e-02 -2.20898882e-01 1.25179839e+00 7.05931723e-01
1.11622715e+00 -5.29646575e-02 -9.92678881e-01 8.22895527e-01
1.69059420e+00 4.81705278e-01 3.87231737e-01 5.57739675e-01
4.75899220e-01 1.40788555e-01 3.57808560e-01 2.98036873e-01
4.86461252e-01 2.75382966e-01 4.45283890e-01 -7.17450738e-01
1.35935405e-02 1.33955747e-01 -7.40351379e-02 3.75469595e-01
-1.09814107e-02 -3.27380359e-01 -1.12654448e+00 6.71291173e-01
-1.44542861e+00 -8.10060799e-01 -1.98170707e-01 2.11310172e+00
6.66915774e-01 -1.08097330e-01 2.49838084e-01 -2.33369600e-02
7.56493390e-01 -5.92676818e-01 -4.14200783e-01 1.71536937e-01
5.02359644e-02 1.25467032e-01 5.22490501e-01 9.07276571e-03
-1.61644506e+00 2.43621215e-01 6.21916008e+00 4.46707428e-01
-1.29410648e+00 1.01008385e-01 1.20618045e+00 5.14222644e-02
-1.14127800e-01 -6.23388648e-01 -5.79554975e-01 2.91230142e-01
9.66592908e-01 1.50802687e-01 -3.02151889e-01 7.75101781e-01
4.95912693e-03 -2.08091646e-01 -1.08162010e+00 6.14042103e-01
2.17106923e-01 -1.33423519e+00 -1.96991250e-01 2.96295404e-01
8.26361716e-01 2.35481650e-01 1.17232390e-01 1.87310755e-01
8.67644921e-02 -8.93962502e-01 2.98803478e-01 2.74780512e-01
1.07440877e+00 -5.54755330e-01 1.34943998e+00 1.66836843e-01
-1.27586782e+00 -1.72235277e-02 2.64465809e-02 4.96348530e-01
-1.04833454e-01 5.10788023e-01 -1.08054376e+00 1.93665385e-01
8.87930036e-01 4.51852262e-01 -7.34593868e-01 1.41552353e+00
1.48617893e-01 6.21899247e-01 -6.27924681e-01 -4.63037223e-01
3.63952309e-01 2.34496787e-01 -4.21954542e-02 1.49662817e+00
6.48178756e-01 5.17677777e-02 4.89043742e-01 5.18640220e-01
1.12153128e-01 3.86896044e-01 -1.93229616e-01 2.44693056e-01
1.73388422e-01 1.31608093e+00 -1.27364969e+00 -4.99625236e-01
-8.15505564e-01 1.05127156e+00 -9.56123099e-02 -2.72883847e-02
-9.32179868e-01 -3.01204979e-01 -1.17156498e-01 -1.90144945e-02
4.84303713e-01 5.99753559e-01 -1.01541042e-01 -1.17242396e+00
-1.85931936e-01 -7.98550785e-01 7.79949427e-01 1.06455080e-01
-1.11681390e+00 5.39979815e-01 -2.95575202e-01 -1.40685475e+00
-2.60873199e-01 -8.07239354e-01 -7.70414174e-01 5.45523584e-01
-1.69347739e+00 -1.28371823e+00 -7.75414944e-01 3.71453136e-01
1.74199924e-01 -8.20562020e-02 1.25081813e+00 1.62144884e-01
-6.59595609e-01 7.54862309e-01 2.59909391e-01 3.34451586e-01
5.35396755e-01 -1.27054465e+00 4.45687734e-02 3.56003881e-01
-3.88439655e-01 5.65337837e-01 4.53367978e-02 -5.80964148e-01
-1.32742763e+00 -1.32333231e+00 5.81806004e-01 1.85458511e-02
4.86093521e-01 3.82330231e-02 -9.80236351e-01 3.91218305e-01
-2.76848704e-01 3.97593647e-01 1.24605250e+00 -2.45956615e-01
-8.72001946e-02 1.06278703e-01 -1.50350988e+00 3.07496160e-01
3.42247695e-01 -2.36796126e-01 -6.49693236e-02 4.90366876e-01
2.72071928e-01 -5.29829562e-01 -1.07574177e+00 5.06625652e-01
8.01553428e-01 -1.06118989e+00 1.06100166e+00 4.06106226e-02
2.84904819e-02 2.56332010e-02 2.09217414e-01 -8.54472518e-01
-7.31803894e-01 -1.57694757e-01 1.00178488e-01 3.78755480e-01
6.92735374e-01 -7.76255906e-01 1.18261766e+00 4.26056176e-01
-2.04862699e-01 -1.06011009e+00 -6.80281699e-01 -3.28936607e-01
1.68962136e-01 9.35948789e-02 1.32317290e-01 7.72228956e-01
-5.46411350e-02 -4.32901859e-01 2.94916838e-01 1.60784408e-01
5.98812461e-01 1.42362699e-01 2.59982735e-01 -1.30553627e+00
-3.58070582e-01 -5.96812427e-01 -3.04828376e-01 -4.78391111e-01
-3.19276214e-01 -8.53863120e-01 -1.94535807e-01 -1.66747093e+00
7.77729750e-01 -2.65947253e-01 -6.71501935e-01 5.14948964e-01
-4.58603233e-01 5.90045929e-01 -3.35429013e-01 3.90088677e-01
-5.55775285e-01 -2.67065972e-01 1.10406148e+00 -2.46575072e-01
-3.23756933e-01 2.12583661e-01 -7.87010193e-01 7.37958372e-01
1.06642282e+00 -3.23313087e-01 2.19522789e-02 -1.13178097e-01
-2.10301980e-01 2.87527293e-02 5.53172708e-01 -9.64522898e-01
6.54961616e-02 1.21207438e-01 7.11505234e-01 -4.42918211e-01
-7.74857998e-02 -7.04165757e-01 2.95148313e-01 1.50750887e+00
4.90010018e-04 -2.42642358e-01 3.03611249e-01 6.58919871e-01
-2.94333488e-01 -5.37317954e-02 7.87621319e-01 -5.68078756e-01
-7.22206116e-01 2.45879456e-01 -6.06578827e-01 -2.60198683e-01
1.30106974e+00 -3.47080737e-01 -1.93691894e-01 9.89864394e-02
-3.32467884e-01 2.47370273e-01 4.77378249e-01 1.32615697e-02
5.91158271e-01 -1.03609788e+00 -9.71027970e-01 4.61803705e-01
2.85318196e-01 7.67462626e-02 2.10230872e-01 1.33034050e+00
-1.01826930e+00 6.24993145e-01 1.31147683e-01 -9.03532803e-01
-1.20215690e+00 3.71782601e-01 7.38140702e-01 -6.25764847e-01
-8.06205034e-01 8.85785460e-01 2.47313365e-01 -2.94164661e-02
1.20809712e-01 -6.83312535e-01 -6.67633936e-02 -1.38200715e-01
6.45605803e-01 8.87042806e-02 1.19377330e-01 -2.57456273e-01
-4.81040269e-01 4.85086501e-01 -3.31067175e-01 3.93545687e-01
8.60528350e-01 2.10339025e-01 -4.26861718e-02 7.23209977e-02
1.29200530e+00 -2.02358872e-01 -7.54325151e-01 -5.82951633e-03
3.59321944e-03 -1.62532330e-02 2.08042979e-01 -7.52029240e-01
-1.26014292e+00 4.01323020e-01 1.07602131e+00 3.39776218e-01
1.16646123e+00 2.09205270e-01 8.03290546e-01 3.00827593e-01
3.79745603e-01 -4.76315409e-01 1.11465938e-01 -1.25801668e-01
3.69562447e-01 -1.75453258e+00 -9.87097695e-02 -5.26547372e-01
-4.42990452e-01 1.18736327e+00 5.02638161e-01 -2.39605770e-01
5.91309726e-01 3.32008898e-01 2.87229776e-01 -3.15863222e-01
-6.95677161e-01 -1.32219270e-01 3.85516733e-01 5.32800972e-01
8.49147081e-01 3.56759578e-01 -1.79699957e-01 4.49573547e-01
3.22901934e-01 1.72034726e-01 1.02754712e-01 9.19575095e-01
-6.50727212e-01 -7.39183903e-01 -3.45914602e-01 8.69514823e-01
-1.06849360e+00 1.63197087e-03 -7.60561824e-02 1.18413556e+00
2.26821215e-03 7.24813461e-01 2.51686782e-01 -1.19722791e-01
2.14924756e-02 -1.98063746e-01 1.01696469e-01 -6.48515940e-01
-7.44303286e-01 6.92233682e-01 7.50026554e-02 -4.10316944e-01
-5.49046278e-01 -7.35861361e-01 -1.28849459e+00 2.11036444e-01
-4.58219886e-01 1.22842908e-01 3.62108469e-01 4.96355325e-01
2.52927750e-01 4.24928188e-01 3.91998678e-01 -5.28524816e-01
-4.28234577e-01 -7.44468451e-01 -4.62720901e-01 5.10075450e-01
4.43005860e-01 -3.15033674e-01 -4.69109833e-01 -2.56124083e-02] | [15.06147575378418, -2.670104503631592] |
04924c78-2d1c-4551-b095-66f28ad3f6c9 | federated-online-clustering-of-bandits | 2208.14865 | null | https://arxiv.org/abs/2208.14865v1 | https://arxiv.org/pdf/2208.14865v1.pdf | Federated Online Clustering of Bandits | Contextual multi-armed bandit (MAB) is an important sequential decision-making problem in recommendation systems. A line of works, called the clustering of bandits (CLUB), utilize the collaborative effect over users and dramatically improve the recommendation quality. Owing to the increasing application scale and public concerns about privacy, there is a growing demand to keep user data decentralized and push bandit learning to the local server side. Existing CLUB algorithms, however, are designed under the centralized setting where data are available at a central server. We focus on studying the federated online clustering of bandit (FCLUB) problem, which aims to minimize the total regret while satisfying privacy and communication considerations. We design a new phase-based scheme for cluster detection and a novel asynchronous communication protocol for cooperative bandit learning for this problem. To protect users' privacy, previous differential privacy (DP) definitions are not very suitable, and we propose a new DP notion that acts on the user cluster level. We provide rigorous proofs to show that our algorithm simultaneously achieves (clustered) DP, sublinear communication complexity and sublinear regret. Finally, experimental evaluations show our superior performance compared with benchmark algorithms. | ['John C. S. Lui', 'Shuai Li', 'Tong Yu', 'Haoru Zhao', 'Xutong Liu'] | 2022-08-31 | null | null | null | null | ['online-clustering'] | ['computer-vision'] | [-2.70561159e-01 -1.96116194e-01 -6.84895694e-01 -3.99647564e-01
-1.18474126e+00 -9.71315563e-01 9.83538851e-02 7.76498169e-02
-2.75551468e-01 8.91897738e-01 2.57850617e-01 -5.77053249e-01
-5.90979159e-01 -6.27083182e-01 -9.22577560e-01 -1.26391351e+00
1.11723915e-01 4.25495446e-01 -1.68275565e-01 2.97216177e-01
-4.42262366e-02 3.03419918e-01 -7.73907423e-01 5.31423151e-01
9.57393289e-01 1.09220243e+00 -1.33348331e-01 3.27456415e-01
-2.02768724e-02 7.56231964e-01 -1.42866736e-02 -7.77809024e-01
7.89626896e-01 -4.09253746e-01 -8.73361170e-01 9.70971957e-02
8.10154825e-02 -5.03574014e-01 -3.44457030e-01 1.27935445e+00
4.78785425e-01 1.68034688e-01 1.38371840e-01 -1.37884212e+00
-6.92400396e-01 1.20936716e+00 -8.12101126e-01 -9.72940847e-02
-1.01015419e-01 -6.52438641e-01 1.28123844e+00 -8.62114877e-02
1.96663320e-01 9.22622502e-01 7.75245607e-01 5.63748479e-01
-1.44974005e+00 -6.94887102e-01 4.22625571e-01 6.50037155e-02
-1.10254610e+00 -2.90580660e-01 6.45095646e-01 -1.90114230e-01
-1.18264399e-01 1.17487562e+00 6.49328232e-01 7.67936945e-01
-6.01739049e-01 1.33268440e+00 1.38823819e+00 -4.33620840e-01
6.62198424e-01 4.07697707e-01 4.69475448e-01 1.24210007e-01
6.00185275e-01 -3.87458727e-02 -2.46545553e-01 -7.92754710e-01
5.96136510e-01 6.45775735e-01 -4.95145440e-01 -9.04451132e-01
-6.57980561e-01 9.99741733e-01 4.08130080e-01 1.10434785e-01
-3.66078347e-01 3.20157677e-01 1.22845747e-01 5.73177636e-01
9.19257343e-01 1.96068501e-03 -4.96338874e-01 2.80283630e-01
-8.49838972e-01 2.58967876e-01 1.07750368e+00 1.18553805e+00
6.62321091e-01 -6.72714114e-01 -6.05734766e-01 7.35479474e-01
1.37180567e-01 4.97020662e-01 -4.25543897e-02 -1.06002510e+00
5.16746640e-01 1.10014327e-01 6.42888606e-01 -8.65911543e-01
-5.43354675e-02 -7.73089468e-01 -9.48777735e-01 -3.58899355e-01
5.32502890e-01 -5.52456141e-01 -1.18892267e-01 1.55149639e+00
6.09678566e-01 1.17107183e-01 -1.90635979e-01 1.19535089e+00
3.08039282e-02 6.57776296e-01 -4.10110772e-01 -9.17608261e-01
9.60731149e-01 -1.06625140e+00 -6.72505617e-01 3.02678943e-01
8.47110569e-01 -3.58729243e-01 4.23147917e-01 5.22924781e-01
-1.12476730e+00 4.69149530e-01 -4.43836659e-01 1.86634675e-01
-5.12904078e-02 -2.53544658e-01 9.62589085e-01 1.40033805e+00
-8.85581493e-01 3.44546854e-01 -7.32370198e-01 -3.82353216e-01
7.58879721e-01 4.78778541e-01 9.23648775e-02 -2.83446908e-01
-5.28851748e-01 -2.11468190e-02 -1.25643820e-01 3.64438444e-02
-3.83788437e-01 -7.96651244e-01 -4.85737398e-02 5.59605248e-02
6.50468171e-01 -8.36205065e-01 1.28963649e+00 -1.22163427e+00
-1.44883156e+00 5.09891152e-01 -4.76529971e-02 -7.59139299e-01
9.69288826e-01 -7.58912265e-02 2.16989592e-01 -1.14397101e-01
-1.48725212e-01 -3.80390912e-01 4.18293417e-01 -1.50329173e+00
-1.10131288e+00 -7.29034483e-01 2.82249153e-01 1.52688414e-01
-6.36442900e-01 5.91129623e-02 -4.85719800e-01 -4.49991077e-01
9.48890671e-02 -1.10230172e+00 -5.88894606e-01 -1.30628973e-01
-4.81427222e-01 -9.95001271e-02 6.02934778e-01 -4.40384179e-01
1.18053067e+00 -2.12269568e+00 -1.80731609e-01 6.77415371e-01
1.26636207e-01 1.31381020e-01 2.19160497e-01 4.28836375e-01
3.40370864e-01 2.10073426e-01 1.12095512e-01 -5.66877246e-01
2.83418357e-01 1.20949879e-01 -5.68096459e-01 9.37310576e-01
-1.24239385e+00 7.11313069e-01 -7.45898962e-01 -1.25791997e-01
-1.43472090e-01 -2.07331851e-01 -1.11000037e+00 1.28416300e-01
-2.40852296e-01 4.87427592e-01 -1.03255928e+00 4.58276898e-01
1.26641583e+00 -2.89044231e-01 4.92781222e-01 1.11886777e-01
-2.36879826e-01 -6.15247376e-02 -1.23283815e+00 1.41438973e+00
-4.23358023e-01 -5.64529598e-02 8.44064951e-01 -1.41639662e+00
2.80394524e-01 5.06100118e-01 7.00322688e-01 -3.40187699e-01
4.24747393e-02 1.94001615e-01 -5.10424435e-01 -2.81142265e-01
1.00420393e-01 -5.72798438e-02 -5.81334205e-03 8.98312807e-01
-6.54551268e-01 5.49435318e-01 -4.98416573e-01 2.34359995e-01
9.46931839e-01 -4.75295246e-01 -6.58934116e-02 -4.23902154e-01
1.75579742e-01 -6.31760806e-02 7.72796035e-01 1.27765512e+00
-1.65358588e-01 3.97246480e-01 5.12951732e-01 -4.48616832e-01
-7.44116783e-01 -5.04302025e-01 1.34787589e-01 1.58195651e+00
2.51302183e-01 -1.51870012e-01 -8.68625581e-01 -7.98954546e-01
4.02772367e-01 5.83989561e-01 -6.40486836e-01 3.12971056e-01
-3.28984968e-02 -9.28055942e-01 6.38135597e-02 -8.71102512e-03
5.75454891e-01 -8.51501599e-02 1.28162593e-01 2.42121235e-01
-3.24482501e-01 -7.14685798e-01 -9.12411094e-01 1.04095666e-02
-7.58179963e-01 -8.01935971e-01 -7.15374768e-01 -5.14905214e-01
6.86152279e-01 9.62407112e-01 5.54079413e-01 -3.17954570e-02
2.80012280e-01 7.68940389e-01 -5.76167583e-01 -4.26762670e-01
1.85643271e-01 1.60873905e-01 1.18493522e-03 7.81586587e-01
1.84463724e-01 -7.91696489e-01 -1.00310707e+00 4.57116663e-01
-8.34471881e-01 1.35688320e-01 2.40034238e-01 7.08266377e-01
5.26337326e-01 -1.19728362e-02 3.64501923e-01 -1.43671823e+00
5.60496867e-01 -5.65762401e-01 -8.49695086e-01 5.56963444e-01
-5.90778887e-01 -2.91095346e-01 6.68497682e-01 -2.02572018e-01
-1.12491751e+00 2.28545964e-01 3.14419687e-01 -2.67649323e-01
2.38480657e-01 2.72544652e-01 -1.62544906e-01 -3.20021987e-01
2.83873737e-01 1.73179746e-01 -1.50443435e-01 -7.61362135e-01
5.80065489e-01 1.07605958e+00 -2.85055898e-02 -8.31112981e-01
5.47239900e-01 9.66309369e-01 -2.31195211e-01 -4.26647663e-01
-1.35119689e+00 -6.79529369e-01 -5.70181832e-02 6.18757382e-02
2.90346831e-01 -1.04538882e+00 -1.23898256e+00 9.78061482e-02
-8.94270658e-01 -1.36259690e-01 -1.69902816e-02 5.78149796e-01
-7.00718582e-01 4.82333809e-01 -4.45465595e-01 -1.51579630e+00
-5.32399356e-01 -5.29205322e-01 4.21096116e-01 -1.20744437e-01
4.25468415e-01 -7.46026039e-01 2.32425053e-02 9.03656960e-01
3.04959029e-01 3.57094854e-02 4.50639546e-01 -7.86093831e-01
-8.80751371e-01 -1.32602185e-01 -2.85652786e-01 1.02871604e-01
1.71442274e-02 -6.61446035e-01 -8.96346331e-01 -7.74469018e-01
1.84171960e-01 -8.59315395e-02 7.54867554e-01 6.20031893e-01
1.70189548e+00 -1.20500553e+00 -5.41840971e-01 7.55088270e-01
1.35168743e+00 6.36112019e-02 2.15977594e-01 5.17645329e-02
6.04239464e-01 4.35379624e-01 4.69038457e-01 9.96562183e-01
3.12620312e-01 5.10830402e-01 4.23918724e-01 5.39597645e-02
6.77009881e-01 -3.12710464e-01 8.97622481e-02 5.18802881e-01
-9.50914621e-02 -2.04304695e-01 -3.06046188e-01 5.83535552e-01
-2.70863843e+00 -1.05096805e+00 -3.14231783e-01 2.64476395e+00
9.18620765e-01 -4.95049715e-01 6.08250678e-01 -9.63180140e-02
9.53266025e-01 -2.59761214e-01 -5.84889472e-01 -2.15288445e-01
-4.48495187e-02 -3.55266809e-01 1.31390977e+00 3.84233594e-01
-1.08919108e+00 5.86316943e-01 5.29037666e+00 1.06752789e+00
-8.16687405e-01 5.57006180e-01 1.05705214e+00 -6.27815366e-01
-4.49886709e-01 -4.23185527e-03 -4.38557565e-01 5.63326657e-01
6.54608548e-01 -4.06673402e-01 1.03990412e+00 1.03521836e+00
4.42403764e-01 1.93886295e-01 -9.98393893e-01 1.19808066e+00
-5.14403284e-01 -1.58155954e+00 -3.09781641e-01 5.38347185e-01
1.27865934e+00 -4.42810804e-02 2.34836593e-01 -2.54288800e-02
9.26469088e-01 -3.69132847e-01 5.29103518e-01 3.60746831e-01
3.54855359e-01 -1.23803091e+00 3.66262972e-01 7.81217813e-01
-6.46596074e-01 -7.19731390e-01 -6.46781266e-01 -1.27383873e-01
-1.30400658e-01 8.26125801e-01 -1.23220915e-02 6.96513593e-01
8.72428954e-01 5.10724485e-01 2.30014190e-01 1.44230330e+00
3.23964834e-01 9.49851811e-01 -5.21462321e-01 -3.02061379e-01
3.34271103e-01 -7.58630216e-01 3.51463228e-01 9.32233453e-01
4.10899073e-01 6.54243112e-01 2.71284342e-01 4.73943889e-01
-3.27916116e-01 5.14830947e-01 -3.61252993e-01 1.50487736e-01
5.81082761e-01 1.19962680e+00 -3.29047889e-01 -1.29155546e-01
-5.45550644e-01 1.07231247e+00 4.45958614e-01 3.20981354e-01
-6.66590989e-01 2.00593308e-01 8.81100476e-01 -2.67432444e-02
6.69633210e-01 1.64288878e-01 -6.45510972e-01 -1.14896834e+00
-1.09441229e-04 -6.59681261e-01 8.35874498e-01 -1.26906186e-01
-1.61363101e+00 -2.88767487e-01 -4.39061016e-01 -1.05224764e+00
4.85403299e-01 -1.13354381e-02 -4.59578782e-01 4.15290207e-01
-1.23178518e+00 -1.21311426e+00 2.42320299e-01 9.42543328e-01
-1.61034063e-01 1.64844841e-01 6.91067576e-01 5.41509449e-01
-7.07114875e-01 1.09058034e+00 1.47183275e+00 -6.74000904e-02
5.52561402e-01 -1.17031384e+00 -6.17256761e-01 6.16170406e-01
9.91354287e-02 6.43833041e-01 6.54009700e-01 -1.28850713e-01
-1.93415737e+00 -1.40161514e+00 5.78445733e-01 -2.22573206e-01
5.92879057e-01 -6.81287289e-01 -3.38876784e-01 9.35420275e-01
-4.52625938e-02 1.81382358e-01 1.13781703e+00 4.26799506e-01
-4.14454132e-01 -8.57875824e-01 -1.55652916e+00 4.85758990e-01
1.22576368e+00 -1.61708862e-01 3.73485163e-02 7.89679468e-01
8.03264201e-01 -1.48797512e-01 -7.03577936e-01 -2.26758569e-01
6.14929855e-01 -1.01085174e+00 5.44007301e-01 -8.08268726e-01
-3.27289402e-01 -5.81214167e-02 -2.40473494e-01 -1.27194929e+00
-5.80378413e-01 -1.42480052e+00 -3.56309451e-02 1.33643544e+00
3.78570288e-01 -9.12624300e-01 1.23913169e+00 1.27323377e+00
3.84828359e-01 -4.02006477e-01 -1.12056327e+00 -6.79000378e-01
1.84475303e-01 -2.72619694e-01 7.51421392e-01 1.21232343e+00
4.40591007e-01 -2.67307162e-02 -9.44446504e-01 4.50796574e-01
1.00401545e+00 8.35870743e-01 9.68888164e-01 -1.12232172e+00
-5.95503747e-01 -1.15719669e-01 4.00674641e-01 -1.38824213e+00
-2.02351570e-01 -8.97845209e-01 -2.97501266e-01 -1.28885031e+00
6.26013339e-01 -7.30071306e-01 -5.67843258e-01 3.73024613e-01
3.48693877e-01 -2.92205047e-02 2.39361107e-01 3.29897553e-01
-1.18224978e+00 2.96527952e-01 1.05075395e+00 -8.18349496e-02
-2.16075823e-01 9.36896682e-01 -1.39428329e+00 2.79164284e-01
6.35382831e-01 -5.58834374e-01 -2.19769329e-01 -4.18567210e-01
2.90089488e-01 2.99924076e-01 7.60638788e-02 -4.09873188e-01
5.64816892e-01 -4.85926539e-01 -1.17362492e-01 -6.80336118e-01
1.86899509e-02 -1.47741652e+00 4.08956647e-01 4.64258313e-01
-5.90806425e-01 -8.43801498e-01 -4.96332526e-01 1.18500662e+00
3.47708255e-01 4.10652421e-02 7.09592879e-01 -2.51019746e-02
3.17725718e-01 8.13448370e-01 -2.34311596e-01 -3.30598474e-01
1.18463969e+00 1.58164129e-01 -1.38243765e-01 -8.94460440e-01
-7.55254209e-01 6.84765756e-01 2.36361206e-01 -1.91101879e-01
-1.27293125e-01 -1.32844937e+00 -7.16884792e-01 -2.76066780e-01
-5.51592782e-02 -1.36880100e-01 3.78372580e-01 1.09810352e+00
-2.51826830e-04 5.90318441e-01 5.53748608e-01 -2.24823713e-01
-1.38943386e+00 9.25419331e-01 2.57084578e-01 -1.38669670e-01
-1.94549128e-01 8.41343284e-01 2.53071457e-01 -3.37320805e-01
6.23441577e-01 -7.97462314e-02 4.15391207e-01 2.43411288e-02
5.84009886e-01 6.10781550e-01 -7.09246546e-02 3.31584290e-02
2.02108789e-02 -1.16590783e-01 -3.30099493e-01 -7.88769033e-03
1.68219531e+00 -6.12380028e-01 -4.74924862e-01 1.20677575e-01
1.16041958e+00 2.86613494e-01 -1.08275676e+00 -8.29435825e-01
-2.59950142e-02 -8.18527162e-01 2.22112060e-01 -6.31882548e-01
-1.36222720e+00 2.77882040e-01 4.48440224e-01 8.84449720e-01
9.74240065e-01 -4.04489003e-02 8.39222610e-01 4.93689239e-01
9.02767360e-01 -1.26121259e+00 -7.42735922e-01 7.61861280e-02
3.63154382e-01 -1.03971231e+00 6.73653278e-03 -5.14094651e-01
-3.17735851e-01 6.24078393e-01 7.27481842e-02 2.17609257e-02
9.98219311e-01 -1.35210603e-01 -3.10605824e-01 3.52897704e-01
-6.91003263e-01 -2.66274717e-02 -2.02988625e-01 2.96349585e-01
7.46182203e-02 5.27912438e-01 -7.79142857e-01 1.32735097e+00
1.11223407e-01 4.93591204e-02 2.18538970e-01 8.95781457e-01
-4.90979284e-01 -1.18340909e+00 -6.31147265e-01 7.17489004e-01
-1.02300322e+00 1.45622090e-01 -4.77683723e-01 1.00147553e-01
-2.45625466e-01 1.19723427e+00 -1.01149507e-01 -1.04080383e-02
-3.21849249e-02 -3.61892492e-01 2.46407777e-01 -2.27829114e-01
-6.39723837e-01 3.80503953e-01 3.30171995e-02 -4.45328444e-01
-5.49259245e-01 -7.39799440e-01 -4.96482641e-01 -7.52112865e-01
-6.46421790e-01 8.28223586e-01 5.26031315e-01 7.94591129e-01
6.21901274e-01 -3.06090742e-01 1.65312469e+00 -2.13462830e-01
-8.86887312e-01 -5.04345059e-01 -1.17224431e+00 2.51478851e-01
4.09270257e-01 7.32131973e-02 -4.21025187e-01 -2.98236787e-01] | [4.600452899932861, 3.4438791275024414] |
dd2ce228-3d5b-4120-adb5-d32ec5bc1829 | clothes-invariant-feature-learning-by-causal | 2305.06145 | null | https://arxiv.org/abs/2305.06145v1 | https://arxiv.org/pdf/2305.06145v1.pdf | Clothes-Invariant Feature Learning by Causal Intervention for Clothes-Changing Person Re-identification | Clothes-invariant feature extraction is critical to the clothes-changing person re-identification (CC-ReID). It can provide discriminative identity features and eliminate the negative effects caused by the confounder--clothing changes. But we argue that there exists a strong spurious correlation between clothes and human identity, that restricts the common likelihood-based ReID method P(Y|X) to extract clothes-irrelevant features. In this paper, we propose a new Causal Clothes-Invariant Learning (CCIL) method to achieve clothes-invariant feature learning by modeling causal intervention P(Y|do(X)). This new causality-based model is inherently invariant to the confounder in the causal view, which can achieve the clothes-invariant features and avoid the barrier faced by the likelihood-based methods. Extensive experiments on three CC-ReID benchmarks, including PRCC, LTCC, and VC-Clothes, demonstrate the effectiveness of our approach, which achieves a new state of the art. | ['Nenghai Yu', 'Wanli Ouyang', 'Qi Chu', 'Yating Liu', 'Yuenan Hou', 'Bin Liu', 'Yan Lu', 'Xulin Li'] | 2023-05-10 | null | null | null | null | ['person-re-identification'] | ['computer-vision'] | [-3.74988653e-02 -3.86413991e-01 -1.17674388e-01 -5.19392908e-01
-2.93866664e-01 -2.84960955e-01 7.70073593e-01 -3.37807715e-01
-2.73497373e-01 6.71562135e-01 5.91455102e-01 3.87689501e-01
-2.24142179e-01 -6.19454145e-01 -8.05824161e-01 -9.05477107e-01
1.31991394e-02 -2.59574026e-01 -3.56569469e-01 -1.06506832e-01
-9.16144550e-02 3.16954911e-01 -1.56908381e+00 -2.16277480e-01
6.51267171e-01 7.10902810e-01 -3.17667663e-01 3.99777532e-01
1.26657695e-01 3.95611733e-01 -4.35688972e-01 -8.11122835e-01
5.11147201e-01 -5.00129819e-01 -2.27271438e-01 -1.58607781e-01
1.00259995e+00 -5.01317859e-01 -6.57768309e-01 1.11250412e+00
5.17064750e-01 -1.47599176e-01 6.73248529e-01 -1.48270059e+00
-1.33124876e+00 3.75690192e-01 -9.54527557e-01 -1.76493824e-01
6.81747735e-01 3.32020283e-01 7.88474083e-01 -9.56087291e-01
4.95395631e-01 1.96587658e+00 1.09649277e+00 9.17619824e-01
-1.34373486e+00 -1.13117242e+00 6.59599960e-01 1.58336431e-01
-1.54854500e+00 -3.83584142e-01 9.52990115e-01 -4.71666664e-01
5.48003502e-02 6.24990344e-01 6.76271141e-01 1.68514347e+00
1.50774315e-01 1.03686476e+00 1.63627815e+00 -2.86750704e-01
-3.13265681e-01 8.86265337e-02 4.14964139e-01 8.07868183e-01
5.02622128e-01 5.28448880e-01 -7.48089910e-01 -3.53425324e-01
7.79872239e-01 4.06281859e-01 -2.41822585e-01 -7.14220926e-02
-1.28119695e+00 6.44209385e-01 2.79098928e-01 -2.17833936e-01
-1.61758721e-01 2.57349491e-01 2.25135356e-01 2.18426570e-01
3.84763777e-01 1.37641311e-01 -4.97202277e-01 2.31305450e-01
-3.85136455e-01 5.95121920e-01 5.77369630e-01 1.02804887e+00
5.02830267e-01 -1.61869694e-02 -7.16875315e-01 6.00775182e-01
2.96078116e-01 1.13821554e+00 -7.22517371e-02 -5.47492623e-01
-1.49991348e-01 5.22562563e-01 2.20634446e-01 -1.28363883e+00
-3.00817579e-01 -2.16135278e-01 -1.06849205e+00 -1.41009197e-01
3.62133801e-01 -6.25343099e-02 -6.72012269e-01 2.14122343e+00
4.73491997e-01 3.61902148e-01 -4.42875564e-01 1.04351735e+00
9.29531097e-01 1.04728773e-01 3.89519900e-01 -4.72164780e-01
1.38839829e+00 -6.61496818e-01 -9.70444262e-01 8.17827433e-02
-1.67300895e-01 -7.82752216e-01 1.17408693e+00 1.97225139e-01
-4.98496652e-01 -7.92225778e-01 -8.75609934e-01 -1.50913056e-02
-3.00238460e-01 1.32924020e-01 9.86991048e-01 7.86731839e-01
-6.40022635e-01 6.71036780e-01 -2.83904165e-01 -3.86824369e-01
4.02871430e-01 4.49874550e-01 -4.89116848e-01 -2.66102761e-01
-1.37369180e+00 6.43904388e-01 -3.02117109e-01 2.14800686e-01
-8.06216061e-01 -8.81881475e-01 -8.30534697e-01 -3.45086366e-01
4.06134039e-01 -8.16036522e-01 5.19194961e-01 -8.96955252e-01
-1.34735584e+00 8.58986259e-01 -4.35456246e-01 2.68929452e-01
8.29287052e-01 -5.51841617e-01 -6.53685510e-01 -4.65471566e-01
2.91337550e-01 4.33631390e-01 1.39122093e+00 -1.64897072e+00
-4.43008393e-01 -5.38556218e-01 -2.70435005e-01 -9.76390839e-02
-4.79805946e-01 1.18355520e-01 -5.10928392e-01 -1.07247603e+00
-1.45458758e-01 -1.12880623e+00 8.95794556e-02 2.55718738e-01
-5.04737914e-01 -6.28207564e-01 6.38484478e-01 -8.79294395e-01
1.05365551e+00 -2.15139723e+00 1.05309203e-01 5.60079217e-02
4.04411167e-01 -1.36941582e-01 -2.70654440e-01 1.14113003e-01
-1.97872110e-02 2.60894895e-01 7.87796006e-02 -5.43675244e-01
1.39598936e-01 9.82435569e-02 -2.31931508e-01 9.19405520e-01
2.53481835e-01 1.07527053e+00 -1.10037673e+00 -6.08854651e-01
2.47571096e-01 4.53161597e-01 -1.53476223e-01 2.41292462e-01
2.83587158e-01 5.45253575e-01 -3.36230159e-01 9.23629701e-01
1.04848588e+00 3.05876046e-01 1.06428593e-01 -5.14180243e-01
-7.44728073e-02 -3.15200299e-01 -1.18027675e+00 1.38348401e+00
-2.19960333e-04 3.27694327e-01 -1.16816640e-01 -2.25420997e-01
8.96314383e-01 -8.68459195e-02 5.27805030e-01 -6.23903334e-01
7.08228350e-02 -2.03415141e-01 -4.22478676e-01 -4.35835600e-01
2.07086846e-01 -1.12808332e-01 -3.53743702e-01 2.58751195e-02
-9.83762667e-02 6.57293677e-01 -4.39179093e-01 2.13040002e-02
7.34368622e-01 4.28141594e-01 1.85920626e-01 -5.81597328e-01
4.47646588e-01 -4.66412961e-01 9.91768539e-01 1.07259905e+00
-4.43306446e-01 4.08881724e-01 9.95713845e-02 -5.40045381e-01
-8.31490874e-01 -1.42663491e+00 -2.28926718e-01 9.32851076e-01
7.60236323e-01 -3.16228837e-01 -5.56508303e-01 -9.83066380e-01
6.32414222e-01 4.38174427e-01 -1.21357894e+00 -3.87696713e-01
-6.28520668e-01 -9.37975645e-01 4.52615231e-01 4.77031082e-01
4.07592267e-01 -3.02796066e-01 3.59988034e-01 1.17308348e-02
-2.02013880e-01 -7.73901761e-01 -1.06778443e+00 -5.91195881e-01
-1.67150885e-01 -1.07083774e+00 -6.58326745e-01 -2.64392495e-01
7.92532563e-01 4.85193908e-01 6.79577649e-01 1.40552536e-01
-5.78683496e-01 6.22754931e-01 -2.40496337e-01 -6.19392276e-01
1.50730461e-02 -3.59172165e-01 6.31848872e-01 5.88108957e-01
6.58633292e-01 -2.87018567e-01 -8.83314490e-01 5.65537512e-01
-1.37341842e-01 -5.23664951e-02 4.12612110e-01 1.09276080e+00
4.72181827e-01 -1.39167830e-01 7.89155841e-01 -4.82604623e-01
5.58400214e-01 -3.32895637e-01 -3.69729400e-01 3.79659653e-01
-8.80222201e-01 -2.05709815e-01 4.29043770e-01 -1.04289067e+00
-1.11560261e+00 1.18664503e-01 2.33512431e-01 -5.87713897e-01
-3.01073734e-02 -1.30126148e-01 -6.27553523e-01 -1.86643332e-01
3.13070118e-01 1.77595228e-01 -4.36562188e-02 -6.71246648e-01
4.44207013e-01 3.68230969e-01 7.59716213e-01 -7.56888270e-01
1.30026305e+00 5.94006717e-01 2.53379107e-01 -4.74997163e-01
-9.13220584e-01 -4.66022700e-01 -7.17634678e-01 -5.13398111e-01
9.83019292e-01 -1.16189766e+00 -1.30654585e+00 6.68659985e-01
-1.12593305e+00 3.16996872e-01 -1.96501657e-01 5.11983156e-01
1.97149366e-02 4.07314330e-01 -4.70963717e-01 -1.07098460e+00
-2.87988752e-01 -6.41304851e-01 1.08316207e+00 5.20777643e-01
-2.35454738e-01 -7.44011283e-01 -2.58589839e-03 1.57659501e-01
9.65134948e-02 6.39044642e-01 6.19342268e-01 -1.00494049e-01
-4.64895606e-01 -3.13627839e-01 -4.18759406e-01 1.94715917e-01
5.90606928e-01 -8.79671052e-02 -1.24628079e+00 -4.04224753e-01
-1.87417805e-01 -5.31674065e-02 1.00077951e+00 1.43012151e-01
1.11205769e+00 -5.10793567e-01 -4.37765509e-01 7.33722210e-01
1.30817294e+00 -2.42785901e-01 4.00810659e-01 -6.70626312e-02
1.28558028e+00 5.57575881e-01 6.62755132e-01 3.99577588e-01
7.34559715e-01 8.27761292e-01 1.23182289e-01 -4.92035836e-01
-5.12477100e-01 -7.83714175e-01 5.21760881e-01 7.46947527e-01
-2.96089888e-01 1.87982216e-01 -1.99139833e-01 4.51598108e-01
-2.05955625e+00 -9.84272420e-01 -5.90327144e-01 2.23926830e+00
1.04215515e+00 -2.83933669e-01 2.73392171e-01 -3.16544056e-01
7.60653913e-01 2.05307472e-02 -7.88866580e-01 5.19903228e-02
-3.81644636e-01 -1.24208979e-01 7.28741884e-01 3.47311765e-01
-1.41095853e+00 8.71654093e-01 6.57834482e+00 8.48745525e-01
-8.12076032e-01 2.39405155e-01 2.65282393e-01 1.66063532e-02
-2.51448572e-01 -2.37342924e-01 -1.06489468e+00 6.68948054e-01
2.70323694e-01 -1.32293046e-01 6.52494907e-01 8.97361636e-01
2.65596181e-01 1.94689453e-01 -1.36877882e+00 1.37091124e+00
3.60787213e-01 -8.01810682e-01 -1.52383884e-02 2.05179989e-01
7.75980353e-01 -7.30753005e-01 1.18669353e-01 3.29402864e-01
4.35135156e-01 -8.93168926e-01 8.56929719e-01 9.55808640e-01
1.13092542e+00 -5.83851874e-01 4.92717832e-01 -2.59551972e-01
-1.32145941e+00 -4.38829102e-02 -4.40909624e-01 -1.05643257e-01
-1.09896183e-01 6.02932692e-01 -3.25211495e-01 8.60034466e-01
8.33949864e-01 8.52427721e-01 -8.28462005e-01 5.31786859e-01
-3.04498613e-01 5.03277719e-01 -1.10774204e-01 -8.29265937e-02
-4.70688671e-01 -1.41203739e-02 6.96463823e-01 1.33464038e+00
1.24868955e-02 1.47090569e-01 3.59456629e-01 1.20184278e+00
-2.32896715e-01 1.29614389e-02 -4.75982457e-01 4.22758818e-01
3.77116084e-01 1.14036524e+00 -2.23033741e-01 -1.05972625e-01
-5.07394314e-01 1.33406150e+00 4.25976776e-02 3.09896916e-01
-1.07881844e+00 6.37920275e-02 1.02633464e+00 1.04137115e-01
-2.14329921e-02 -1.23824574e-01 -2.60135621e-01 -1.52716446e+00
1.10922232e-01 -1.01349485e+00 2.09415346e-01 -2.87872642e-01
-2.17804408e+00 -8.64110962e-02 1.51706219e-01 -1.08067572e+00
5.61470687e-01 -4.93154198e-01 -4.28770512e-01 1.04627013e+00
-1.53054154e+00 -1.89911866e+00 -4.91773248e-01 8.70237112e-01
4.32489693e-01 -2.41415184e-02 7.86472082e-01 4.35960859e-01
-7.28819370e-01 1.17091262e+00 -7.09951147e-02 2.49690175e-01
1.23759234e+00 -1.31928921e+00 3.36655617e-01 8.77897382e-01
-3.09127599e-01 1.15581036e+00 6.80771530e-01 -1.13987565e+00
-2.13486862e+00 -1.23234117e+00 6.99374259e-01 -8.19265902e-01
3.84169042e-01 -6.38902307e-01 -6.53640151e-01 4.90423054e-01
-1.36020839e-01 1.51649490e-02 7.73122370e-01 5.33126235e-01
-9.03860927e-01 -5.82953691e-01 -1.21552277e+00 7.62021601e-01
1.63189328e+00 -5.87128282e-01 -6.63174689e-01 1.57540724e-01
7.38654613e-01 1.29218951e-01 -1.13169217e+00 5.14063597e-01
1.09876263e+00 -4.77515966e-01 1.26876009e+00 -6.32612705e-01
-2.93195294e-03 -3.88363421e-01 2.18268186e-02 -1.12033272e+00
-8.09569478e-01 -8.68723333e-01 -1.51594177e-01 1.77446115e+00
-1.92076638e-01 -6.75078213e-01 2.48837918e-01 8.80501926e-01
4.42195386e-01 -2.23459288e-01 -8.03241909e-01 -1.17133009e+00
3.72203859e-03 -7.48758540e-02 1.07649696e+00 1.47344387e+00
-4.68906075e-01 2.03765765e-01 -1.15690386e+00 3.30286860e-01
1.30720544e+00 1.44066736e-01 1.10286534e+00 -1.39867139e+00
-1.70230940e-01 -2.97852695e-01 -1.47676736e-01 -6.65107846e-01
2.98655987e-01 -6.18877530e-01 -1.46393478e-01 -1.10984075e+00
8.65045846e-01 -4.29805487e-01 -4.76806313e-01 5.43870270e-01
-9.38397050e-01 1.60534177e-02 2.81207263e-01 2.35622779e-01
-4.08711106e-01 5.62648356e-01 1.50189745e+00 -4.48094219e-01
-3.70192640e-02 -2.68579870e-01 -8.91580105e-01 8.35419238e-01
3.66563082e-01 -5.77952683e-01 -9.32507664e-02 -8.88201818e-02
-4.04702164e-02 -5.69574535e-01 7.90901303e-01 -3.30970347e-01
4.54010554e-02 -5.53566635e-01 9.56256568e-01 -3.32114398e-01
1.70617551e-01 -6.77762210e-01 3.05454135e-01 3.13622445e-01
-1.02598794e-01 -1.17460087e-01 7.10269064e-02 8.21202993e-01
3.57695192e-01 3.66698265e-01 5.86380303e-01 -7.48872459e-02
-5.48453808e-01 5.74071884e-01 3.10975999e-01 -2.07824305e-01
6.81234658e-01 -3.52642015e-02 -3.99599582e-01 -1.80168226e-01
-3.30907494e-01 2.02089489e-01 5.39511919e-01 8.66712630e-01
5.96128941e-01 -1.76186812e+00 -1.05736125e+00 1.99095830e-01
2.62397140e-01 -6.37203991e-01 3.04046720e-01 7.19081700e-01
3.69730353e-01 -1.22770905e-01 -1.74869820e-02 -4.24556226e-01
-1.51161599e+00 9.07285094e-01 1.30130365e-01 8.64412710e-02
-5.61826587e-01 8.72001827e-01 6.06732428e-01 -3.65469366e-01
2.03755870e-01 1.22643292e-01 8.26823637e-02 -9.86500233e-02
6.28888369e-01 4.86520082e-01 -5.65418065e-01 -7.09898889e-01
-5.96578538e-01 8.19820642e-01 -7.08405450e-02 3.02911252e-01
1.06899703e+00 -3.64310145e-01 -3.27994108e-01 3.90965194e-01
1.12975633e+00 4.05887127e-01 -1.45725882e+00 -1.87915772e-01
-7.24162683e-02 -9.96004105e-01 -1.76913187e-01 -1.01761472e+00
-7.60295987e-01 3.51425916e-01 1.02591324e+00 -5.36057539e-02
9.44426894e-01 -1.48668617e-01 8.00002217e-01 -5.43868775e-03
5.36262393e-01 -9.23199773e-01 9.72356368e-03 -1.19759284e-01
1.12489569e+00 -1.44810653e+00 4.68071043e-01 -5.45762241e-01
-3.14379215e-01 7.22558737e-01 6.40758932e-01 3.39512853e-03
6.64699197e-01 5.41867614e-02 -1.42681241e-01 6.56976923e-02
-1.86007649e-01 -2.65309542e-01 5.85026562e-01 7.64834940e-01
2.86948606e-02 5.35062373e-01 -4.73352611e-01 1.03222108e+00
-1.90806445e-02 -1.85225233e-01 -7.28609115e-02 4.12930548e-01
2.60497451e-01 -1.22441006e+00 -6.34906471e-01 2.64645040e-01
-3.21903616e-01 -3.25070620e-02 -8.47805440e-01 1.04261899e+00
6.93742394e-01 1.03655326e+00 -2.28361309e-01 -7.50909269e-01
4.64394361e-01 -6.28544539e-02 7.97429383e-01 -8.51518139e-02
-4.91505593e-01 1.87190488e-01 2.67916042e-02 -6.29873455e-01
-4.76577312e-01 -1.04879642e+00 -7.80276358e-01 -7.47797370e-01
-4.44773972e-01 -4.89759147e-01 4.54938501e-01 6.75546169e-01
2.32679859e-01 3.47357333e-01 1.08658183e+00 -5.81088483e-01
-5.34861028e-01 -7.40554571e-01 -5.32833338e-01 1.03350520e+00
4.73907173e-01 -1.00772226e+00 -4.13945794e-01 2.47074515e-01] | [14.716497421264648, 0.9601019620895386] |
180ca778-fff4-4651-ad22-2589db6fd6cd | forget-me-not-learning-to-forget-in-text-to | 2303.17591 | null | https://arxiv.org/abs/2303.17591v1 | https://arxiv.org/pdf/2303.17591v1.pdf | Forget-Me-Not: Learning to Forget in Text-to-Image Diffusion Models | The unlearning problem of deep learning models, once primarily an academic concern, has become a prevalent issue in the industry. The significant advances in text-to-image generation techniques have prompted global discussions on privacy, copyright, and safety, as numerous unauthorized personal IDs, content, artistic creations, and potentially harmful materials have been learned by these models and later utilized to generate and distribute uncontrolled content. To address this challenge, we propose \textbf{Forget-Me-Not}, an efficient and low-cost solution designed to safely remove specified IDs, objects, or styles from a well-configured text-to-image model in as little as 30 seconds, without impairing its ability to generate other content. Alongside our method, we introduce the \textbf{Memorization Score (M-Score)} and \textbf{ConceptBench} to measure the models' capacity to generate general concepts, grouped into three primary categories: ID, object, and style. Using M-Score and ConceptBench, we demonstrate that Forget-Me-Not can effectively eliminate targeted concepts while maintaining the model's performance on other concepts. Furthermore, Forget-Me-Not offers two practical extensions: a) removal of potentially harmful or NSFW content, and b) enhancement of model accuracy, inclusion and diversity through \textbf{concept correction and disentanglement}. It can also be adapted as a lightweight model patch for Stable Diffusion, allowing for concept manipulation and convenient distribution. To encourage future research in this critical area and promote the development of safe and inclusive generative models, we will open-source our code and ConceptBench at \href{https://github.com/SHI-Labs/Forget-Me-Not}{https://github.com/SHI-Labs/Forget-Me-Not}. | ['Humphrey Shi', 'Zhangyang Wang', 'Xingqian Xu', 'Kai Wang', 'Eric Zhang'] | 2023-03-30 | null | null | null | null | ['memorization'] | ['natural-language-processing'] | [ 1.58159301e-01 -1.00933336e-01 1.69686422e-01 -2.16051668e-01
-6.73644066e-01 -8.55892599e-01 5.99959850e-01 1.68587074e-01
-3.42829913e-01 6.81339204e-01 -9.22109038e-02 -4.29269582e-01
4.50843051e-02 -8.91487420e-01 -7.99156785e-01 -3.66929650e-01
2.32768565e-01 1.36621073e-01 6.42799139e-02 -3.52107324e-02
3.38053226e-01 5.04018366e-01 -1.70954669e+00 2.69728452e-01
9.93701339e-01 8.17314982e-01 2.38278434e-01 5.57968140e-01
-1.22758381e-01 6.40656710e-01 -9.19207036e-01 -7.02741504e-01
3.29061121e-01 -3.94590974e-01 -5.02725959e-01 -1.95990860e-01
7.48197496e-01 -6.89691126e-01 -2.10567906e-01 1.03762221e+00
6.29877985e-01 5.02816327e-02 5.83475292e-01 -1.52996409e+00
-1.01599860e+00 4.13259357e-01 -5.92950642e-01 7.90287778e-02
2.64130473e-01 3.25224400e-01 7.15609372e-01 -8.65829170e-01
6.05573773e-01 1.10430121e+00 4.52904075e-01 8.72187376e-01
-1.31389284e+00 -1.47592294e+00 1.38372242e-01 -7.04251751e-02
-1.44138253e+00 -4.72700119e-01 5.32319069e-01 -3.89818311e-01
6.29466593e-01 6.34322524e-01 4.95952517e-01 1.44075632e+00
2.28998229e-01 8.28476548e-01 9.25162971e-01 -3.31117839e-01
3.83417994e-01 5.25172532e-01 1.43996784e-02 7.01921582e-01
5.94762325e-01 3.08453813e-02 -6.30550861e-01 -2.43969247e-01
8.56237590e-01 1.23892307e-01 -2.90836900e-01 -2.47126251e-01
-8.14754426e-01 7.94133306e-01 1.86970606e-02 1.72417745e-01
1.35068567e-02 1.85469255e-01 -1.81083903e-02 7.77332708e-02
4.96116638e-01 6.41714275e-01 -1.80953860e-01 -5.87944798e-02
-1.11509061e+00 4.95312959e-01 7.46023953e-01 1.44749808e+00
7.32667863e-01 1.22448079e-01 -3.90653461e-01 7.95349300e-01
-1.03702262e-01 7.05421686e-01 3.54965210e-01 -9.04860139e-01
2.88144410e-01 4.27417517e-01 3.56771834e-02 -9.60508525e-01
-9.88458097e-02 -6.43496633e-01 -8.58395040e-01 1.33285761e-01
-3.05362400e-02 -2.22866073e-01 -1.00783932e+00 1.93406284e+00
-8.94163176e-03 7.80455247e-02 -5.35783052e-01 4.94663328e-01
6.47740126e-01 4.86192226e-01 1.63173944e-01 -3.20932120e-02
1.09917653e+00 -6.40016615e-01 -3.86683047e-01 -1.00901276e-01
3.56672049e-01 -8.16178858e-01 1.28375041e+00 5.79674304e-01
-1.17786288e+00 -4.51879859e-01 -1.04273760e+00 2.36517619e-02
-5.76398611e-01 4.10175510e-02 5.98157346e-01 1.00156581e+00
-1.22365749e+00 4.84630138e-01 -3.96176606e-01 -3.61024559e-01
8.53505731e-01 2.53224075e-01 -1.60278112e-01 -1.61541641e-01
-9.23247874e-01 5.29436767e-01 6.14743680e-02 -3.49646866e-01
-1.03390324e+00 -8.92956853e-01 -6.12870038e-01 2.13790089e-01
3.91004533e-01 -9.38814521e-01 1.01229620e+00 -6.79437220e-01
-1.11229110e+00 5.84854424e-01 1.28263477e-02 -2.13425115e-01
6.12556040e-01 -3.84950429e-01 -4.00616258e-01 4.72788438e-02
1.95776597e-01 8.47442925e-01 1.14902949e+00 -1.69853377e+00
-5.38460970e-01 -1.70061290e-01 -6.92234784e-02 9.70816240e-03
-9.09696639e-01 -9.15241241e-02 -5.92548192e-01 -1.03599691e+00
-2.41134465e-01 -9.60327089e-01 5.51943034e-02 1.44839324e-02
-6.64361775e-01 2.12412536e-01 7.80382514e-01 -5.40740132e-01
1.40274763e+00 -2.19361448e+00 -3.43288660e-01 2.60689199e-01
3.73266667e-01 4.04359162e-01 -4.26835001e-01 4.17858660e-01
4.25509997e-02 7.29739726e-01 -8.28153938e-02 -5.11868715e-01
3.97272371e-02 -1.79054141e-01 -4.82000232e-01 4.12487462e-02
5.35336696e-02 7.13719547e-01 -6.53349996e-01 -1.55129984e-01
1.92150727e-01 5.70406020e-01 -7.75661707e-01 3.94923277e-02
-2.68875211e-01 1.59450695e-01 -2.65535682e-01 7.37039864e-01
9.31616008e-01 -1.75460696e-01 -2.92454828e-02 7.95685351e-02
1.49465621e-01 -8.13416392e-02 -1.05476093e+00 1.33504784e+00
-3.76439035e-01 5.58063149e-01 -2.05361331e-03 -2.10685015e-01
8.10609996e-01 -3.78722176e-02 2.95128077e-01 -8.51403475e-01
8.22914690e-02 1.62123159e-01 -4.43951577e-01 -1.66607335e-01
8.57255638e-01 5.98291121e-02 -1.27948551e-02 6.14843071e-01
8.14568717e-03 -1.16345346e-01 1.95135847e-01 6.03454351e-01
1.03184772e+00 2.81497668e-02 -1.97076961e-01 -2.39283055e-01
1.03933848e-01 -3.39518011e-01 1.93199605e-01 1.11718631e+00
-1.04734134e-02 6.96530998e-01 2.05945477e-01 -8.06318969e-02
-9.59333003e-01 -1.32063532e+00 4.53355163e-02 1.34189749e+00
2.10753769e-01 -5.76285422e-01 -9.28319216e-01 -7.21231997e-01
7.97918737e-02 1.17244935e+00 -4.25494015e-01 -4.04223859e-01
-3.10537219e-01 -6.42643392e-01 7.72380888e-01 2.82544613e-01
4.73420054e-01 -1.03707826e+00 -4.60097849e-01 3.49720903e-02
-2.02563360e-01 -5.74554384e-01 -7.89194882e-01 -1.53298527e-01
-6.54795110e-01 -7.94217467e-01 -7.94775307e-01 -4.64956701e-01
8.94265056e-01 4.21319842e-01 9.77814794e-01 3.69443148e-01
-5.27326345e-01 6.39168084e-01 -3.10718775e-01 -7.45914876e-01
-2.53331125e-01 8.03035349e-02 -6.00905046e-02 -5.73764034e-02
2.28942946e-01 -6.74528003e-01 -8.36680949e-01 3.23883504e-01
-1.27849698e+00 6.80111200e-02 4.96074885e-01 5.57243466e-01
3.51687640e-01 2.17163473e-01 6.24202609e-01 -9.73382890e-01
1.00081444e+00 -3.81819755e-01 -3.69891554e-01 1.53287426e-01
-9.28898275e-01 -1.62501410e-01 4.03095156e-01 -5.80767751e-01
-9.62494075e-01 -3.56338590e-01 -9.66092199e-02 -5.15232086e-01
-1.21450879e-01 8.75110999e-02 -2.56654143e-01 6.64080959e-03
7.91658223e-01 5.31830788e-01 -1.30314648e-01 -4.19815421e-01
4.52133834e-01 5.64099312e-01 3.62795413e-01 -7.01379716e-01
1.04597604e+00 2.98735768e-01 -4.10185546e-01 -6.77540600e-01
-5.04921556e-01 -1.95938185e-01 -6.84801787e-02 -1.18605033e-01
4.40631777e-01 -8.63648295e-01 -6.79027915e-01 6.07216716e-01
-9.42956388e-01 -3.89274061e-01 -2.41197139e-01 2.29904857e-02
-2.58086503e-01 3.19029301e-01 -5.68244517e-01 -8.65127087e-01
-6.00467205e-01 -9.38961506e-01 8.02990675e-01 2.77393907e-01
-4.33437437e-01 -7.70331204e-01 -3.86874259e-01 5.20717442e-01
6.89807832e-01 4.67683822e-02 1.05312145e+00 -6.24175191e-01
-9.67728972e-01 -3.07924032e-01 -2.20399648e-01 5.12390792e-01
6.55478463e-02 -6.66444823e-02 -1.00938272e+00 -7.21508265e-01
-3.80187370e-02 -2.53974885e-01 8.62441301e-01 1.53532326e-01
1.58960986e+00 -7.09589303e-01 -3.87587935e-01 5.63040972e-01
1.32994783e+00 3.91821355e-01 8.62294436e-01 2.89224714e-01
5.77830374e-01 4.08350438e-01 3.41796339e-01 6.67104065e-01
1.90441877e-01 3.35907459e-01 2.58351803e-01 -1.24886394e-01
-2.30621830e-01 -5.08505583e-01 2.17529014e-01 4.80647296e-01
1.24116674e-01 -7.01605618e-01 -6.74528837e-01 4.83137339e-01
-1.51059496e+00 -1.11162758e+00 1.95826024e-01 2.35253263e+00
9.61820066e-01 2.01418698e-01 7.01118261e-03 -1.44337893e-01
6.48124099e-01 -3.97500955e-02 -6.06328547e-01 -2.77067542e-01
3.75197604e-02 4.56272870e-01 5.46149015e-01 3.22538346e-01
-7.92946577e-01 8.99977982e-01 6.04945040e+00 1.25768113e+00
-1.18923950e+00 1.95627257e-01 9.74437952e-01 -5.51765680e-01
-8.45848978e-01 -7.87078589e-02 -1.03472841e+00 7.09881544e-01
6.04000270e-01 -3.79985601e-01 4.87178206e-01 9.44800258e-01
1.24473743e-01 -1.20599553e-01 -1.01556587e+00 9.12219107e-01
2.69156814e-01 -1.47444546e+00 5.92889190e-01 2.49165833e-01
6.38141215e-01 -5.51366627e-01 6.52958930e-01 4.75357920e-01
3.53193700e-01 -9.84771490e-01 8.84713352e-01 4.69943911e-01
1.20284784e+00 -8.37185621e-01 3.61574858e-01 2.29923084e-01
-8.66301537e-01 -1.49842978e-01 -1.68209508e-01 2.76957452e-01
-1.09032944e-01 6.54902220e-01 -6.29757524e-01 3.50412697e-01
7.46629775e-01 1.89689115e-01 -8.42217088e-01 9.30241942e-01
-2.03128718e-02 6.17189407e-01 -4.36974429e-02 -5.49064279e-02
-2.60150820e-01 1.80465043e-01 5.59726238e-01 1.26163185e+00
6.63773417e-01 -7.16736689e-02 -1.55904979e-01 1.23691511e+00
-3.29577297e-01 4.72497977e-02 -4.50282604e-01 -7.39662722e-02
7.94474006e-01 1.07703996e+00 -6.95925951e-01 -2.67618060e-01
-3.96906994e-02 1.09051859e+00 7.46779740e-02 5.66811502e-01
-8.88200939e-01 -5.20824850e-01 5.87317705e-01 7.07797587e-01
1.40310198e-01 -1.12428054e-01 -5.74805796e-01 -9.95907009e-01
6.21935911e-02 -1.10605860e+00 1.72440469e-01 -8.86990488e-01
-1.28817081e+00 5.79369009e-01 1.09061427e-01 -1.07746470e+00
1.35248527e-01 -2.78120637e-01 -5.15672565e-01 9.42567229e-01
-1.08054876e+00 -1.32255471e+00 -4.87184882e-01 6.85994744e-01
5.07024884e-01 -4.58451360e-01 7.61025786e-01 4.05215174e-01
-5.56030452e-01 1.20921719e+00 4.31107916e-02 -6.44178391e-02
9.93174314e-01 -9.55934227e-01 4.01024550e-01 8.98720026e-01
5.67692146e-02 1.19740534e+00 6.07271135e-01 -8.70363355e-01
-1.01979637e+00 -1.26514649e+00 5.77046514e-01 -6.62384450e-01
2.79405802e-01 -7.73254871e-01 -6.65198565e-01 6.16684318e-01
1.31726325e-01 -6.14517629e-01 8.72719467e-01 -5.32067148e-03
-5.66875160e-01 -3.20171505e-01 -1.32560134e+00 9.32491124e-01
1.09048474e+00 -4.58128512e-01 2.42760647e-02 1.66354924e-01
8.15277159e-01 -2.32502028e-01 -5.16682506e-01 1.52159393e-01
5.72388768e-01 -1.22669923e+00 8.84999037e-01 -6.90274267e-03
4.99793351e-01 -1.66292617e-03 1.97750423e-02 -9.75689411e-01
-5.49485505e-01 -9.61907923e-01 -1.49586409e-01 1.57256043e+00
6.00000322e-01 -6.75143242e-01 9.26226497e-01 9.48037684e-01
7.99785927e-03 -5.95092773e-01 -5.27654886e-01 -8.56938303e-01
1.68970883e-01 -4.36479330e-01 7.50014424e-01 9.21456337e-01
-3.17409724e-01 2.32627019e-02 -6.30649626e-01 -1.26765907e-01
7.43545532e-01 -2.83513665e-01 8.37868571e-01 -9.57637489e-01
-3.84003907e-01 -6.01720512e-01 -1.12044215e-02 -9.23771322e-01
-2.78026044e-01 -7.85460055e-01 -2.12389395e-01 -1.52249694e+00
4.43178594e-01 -6.53371274e-01 -2.71316558e-01 7.88761079e-01
-1.34872332e-01 3.55603248e-01 5.30737221e-01 4.24349129e-01
-4.28466707e-01 3.61367881e-01 1.20072818e+00 -1.37559101e-01
-2.48223975e-01 2.51594149e-02 -1.25091767e+00 2.64377505e-01
9.26396132e-01 -5.64243495e-01 -7.96496630e-01 -2.41931304e-01
1.69512033e-01 -3.12086254e-01 4.68861729e-01 -1.07958913e+00
1.15650550e-01 -1.75516367e-01 6.50880516e-01 -5.26171565e-01
3.34397584e-01 -5.26303411e-01 3.40746492e-01 2.77819961e-01
-3.69878858e-01 -8.01606476e-03 5.62830925e-01 5.26186407e-01
2.24070176e-01 -2.37020046e-01 5.86478472e-01 -2.54929692e-01
-4.01849151e-01 4.52172369e-01 -4.54201490e-01 -5.87193705e-02
1.14763105e+00 -3.09994966e-01 -7.38815606e-01 -6.14513636e-01
-5.29033184e-01 1.43473437e-02 6.87110186e-01 6.60302162e-01
8.58875334e-01 -1.11469138e+00 -5.14998436e-01 4.93747026e-01
1.44073635e-01 -1.55366078e-01 5.06563187e-01 3.09697568e-01
-4.28437322e-01 2.66524702e-01 -1.59522623e-01 -1.77498296e-01
-1.26610529e+00 6.25652313e-01 2.78478675e-02 -7.65173808e-02
-3.08035761e-01 9.50650334e-01 4.58724380e-01 -1.49769157e-01
3.87655944e-01 1.66132197e-01 3.01095992e-01 -1.13160357e-01
5.97755432e-01 3.90869379e-01 -1.53031617e-01 -1.78937480e-01
-1.25028402e-01 3.46055776e-02 -6.32814288e-01 -2.14558750e-01
1.06256998e+00 -7.31915236e-02 -4.52184863e-02 -9.70313102e-02
9.47039068e-01 3.18234324e-01 -1.24566400e+00 1.87980890e-01
-3.74876797e-01 -6.58401191e-01 -2.13274896e-01 -1.18645525e+00
-9.71200645e-01 5.84280312e-01 7.03619838e-01 1.97465450e-01
1.24322748e+00 -2.26160005e-01 9.03317094e-01 -1.64243076e-02
5.28390765e-01 -1.04812860e+00 5.58098435e-01 1.88399047e-01
1.06223249e+00 -7.30715513e-01 8.17699656e-02 -3.62286091e-01
-7.11938918e-01 6.15811586e-01 8.86033952e-01 2.89158225e-01
5.34193635e-01 3.14918458e-01 -4.12882492e-03 6.00252338e-02
-6.09881759e-01 1.83235899e-01 1.51291505e-01 9.03298795e-01
3.06810975e-01 -1.94422957e-02 -1.49746969e-01 7.47576594e-01
-2.89942563e-01 1.39257491e-01 6.56516731e-01 1.04315138e+00
-3.01513821e-01 -1.26119876e+00 -2.62148917e-01 7.86825895e-01
-3.98649246e-01 -4.03086722e-01 -4.13068235e-01 5.80559134e-01
4.57560599e-01 1.07376909e+00 -9.40712020e-02 -5.84213078e-01
1.77993491e-01 1.31903298e-03 3.49273354e-01 -7.82765508e-01
-6.92690909e-01 6.00551851e-02 2.12257518e-03 -4.03850019e-01
2.21039385e-01 -4.38870639e-01 -8.99035573e-01 -7.77157128e-01
-3.34778368e-01 -1.43139735e-01 6.24816597e-01 4.36527014e-01
7.82323241e-01 3.26070279e-01 4.68400538e-01 -6.23848140e-01
-5.20730555e-01 -7.97981501e-01 -6.11452103e-01 5.32549620e-01
1.68150961e-02 -4.43441510e-01 -4.07349646e-01 1.06325872e-01] | [11.493905067443848, -0.258380264043808] |
039fde06-f4fb-4745-8d24-47a79b2d2ff4 | look-across-elapse-disentangled | 1809.00338 | null | http://arxiv.org/abs/1809.00338v2 | http://arxiv.org/pdf/1809.00338v2.pdf | Look Across Elapse: Disentangled Representation Learning and Photorealistic Cross-Age Face Synthesis for Age-Invariant Face Recognition | Despite the remarkable progress in face recognition related technologies,
reliably recognizing faces across ages still remains a big challenge. The
appearance of a human face changes substantially over time, resulting in
significant intra-class variations. As opposed to current techniques for
age-invariant face recognition, which either directly extract age-invariant
features for recognition, or first synthesize a face that matches target age
before feature extraction, we argue that it is more desirable to perform both
tasks jointly so that they can leverage each other. To this end, we propose a
deep Age-Invariant Model (AIM) for face recognition in the wild with three
distinct novelties. First, AIM presents a novel unified deep architecture
jointly performing cross-age face synthesis and recognition in a mutual
boosting way. Second, AIM achieves continuous face rejuvenation/aging with
remarkable photorealistic and identity-preserving properties, avoiding the
requirement of paired data and the true age of testing samples. Third, we
develop effective and novel training strategies for end-to-end learning the
whole deep architecture, which generates powerful age-invariant face
representations explicitly disentangled from the age variation. Moreover, we
propose a new large-scale Cross-Age Face Recognition (CAFR) benchmark dataset
to facilitate existing efforts and push the frontiers of age-invariant face
recognition research. Extensive experiments on both our CAFR and several other
cross-age datasets (MORPH, CACD and FG-NET) demonstrate the superiority of the
proposed AIM model over the state-of-the-arts. Benchmarking our model on one of
the most popular unconstrained face recognition datasets IJB-C additionally
verifies the promising generalizability of AIM in recognizing faces in the
wild. | ['ShengMei Shen', 'Yan Xu', 'Lin Xiong', 'Junliang Xing', 'Jian Zhao', 'Yu Cheng', 'Sugiri Pranata', 'Jianshu Li', 'Hengzhu Liu', 'Fang Zhao', 'Yi Cheng', 'Yang Yang', 'Jiashi Feng', 'Shuicheng Yan', 'Haochong Lan'] | 2018-09-02 | null | null | null | null | ['age-invariant-face-recognition'] | ['computer-vision'] | [ 2.47569740e-01 -2.85344303e-01 1.80488229e-01 -8.64651442e-01
-5.73400080e-01 -5.25027633e-01 6.93420470e-01 -7.00900137e-01
-1.65576816e-01 5.54960787e-01 3.63258012e-02 -2.74480209e-02
1.15554419e-03 -6.76084697e-01 -6.87520087e-01 -9.16130722e-01
-2.50639394e-02 2.52895325e-01 -6.63042724e-01 -1.36294305e-01
-1.60187826e-01 9.60460544e-01 -1.99571180e+00 4.66843508e-02
7.60015905e-01 1.46478879e+00 -5.59670389e-01 4.16434616e-01
3.22108299e-01 2.97887027e-01 -4.45522279e-01 -9.67496693e-01
7.15399683e-01 -3.06467950e-01 -4.23475444e-01 1.76636502e-01
1.41126382e+00 -5.96642017e-01 -6.46556616e-01 8.75962377e-01
8.64958227e-01 -1.67285845e-01 7.52973795e-01 -1.64720643e+00
-1.16010475e+00 9.00796652e-02 -8.05100858e-01 -1.98179558e-01
1.78003192e-01 2.57245600e-01 5.83576381e-01 -1.02512515e+00
3.00576717e-01 1.57037294e+00 7.46783435e-01 1.18940008e+00
-1.14363408e+00 -1.27311373e+00 2.54112184e-01 -3.57628451e-03
-1.35476911e+00 -9.88364220e-01 7.42008626e-01 -3.63429040e-01
3.23841155e-01 2.48599142e-01 3.31579506e-01 1.61837363e+00
-1.81317776e-01 5.81297100e-01 1.42998028e+00 -2.21690491e-01
-1.86328497e-02 -3.68151873e-01 -9.27076787e-02 8.27407122e-01
2.27974936e-01 3.69595885e-01 -8.51362407e-01 3.70057710e-02
7.01788664e-01 1.48669690e-01 -2.64389604e-01 -3.47105652e-01
-9.25438404e-01 3.75419527e-01 1.86222240e-01 1.91372018e-02
-2.69546509e-01 3.41827832e-02 3.23259234e-01 5.58501005e-01
6.17176235e-01 1.04633071e-01 -5.20481408e-01 1.84456259e-01
-9.72530901e-01 2.91072100e-01 3.93339604e-01 7.06149280e-01
6.29261792e-01 3.22493047e-01 -3.47718745e-01 9.37928498e-01
2.74348766e-01 8.37266922e-01 2.04720184e-01 -8.34606707e-01
1.12697817e-01 5.46646237e-01 -1.59686580e-01 -7.86391377e-01
-5.13634831e-02 -3.52465272e-01 -1.03960669e+00 6.31343007e-01
6.87809765e-01 5.72282821e-02 -1.21179056e+00 2.34665799e+00
4.31370258e-01 2.01689124e-01 -9.86551493e-03 5.40090024e-01
7.19500124e-01 2.34830692e-01 2.50064135e-01 -2.66340375e-01
1.52519095e+00 -6.14922464e-01 -5.02552688e-01 -1.85049877e-01
2.19484000e-03 -5.78396678e-01 8.45696151e-01 2.93020546e-01
-1.02107811e+00 -6.60076499e-01 -1.05607665e+00 -9.94952172e-02
-4.11261976e-01 2.72271723e-01 9.06376064e-01 1.04535174e+00
-1.32943881e+00 5.46508312e-01 -5.74250698e-01 -2.13138059e-01
9.58706439e-01 6.54456437e-01 -1.03329945e+00 -3.54756832e-01
-8.29619825e-01 5.54082155e-01 -1.86797753e-01 3.53966862e-01
-1.14195657e+00 -1.30007970e+00 -8.49303186e-01 -1.19565532e-01
-1.64688360e-02 -8.77676427e-01 9.18063045e-01 -1.31828856e+00
-1.42407095e+00 1.36999571e+00 -2.17749044e-01 -1.36370376e-01
7.03868568e-01 -3.49433511e-01 -6.53144062e-01 -5.64733483e-02
-2.01246575e-01 8.04964721e-01 1.41953230e+00 -1.32357311e+00
-3.00201982e-01 -1.19994640e+00 -4.24781442e-02 -2.59253144e-01
-8.11620116e-01 2.58984536e-01 -2.76850611e-01 -7.39519119e-01
-2.75701523e-01 -7.79152334e-01 4.34246391e-01 6.74528003e-01
7.38257244e-02 -2.93401241e-01 8.28796208e-01 -9.08072352e-01
8.66340935e-01 -2.12680626e+00 3.34337533e-01 -1.39990956e-01
3.77144396e-01 3.33282024e-01 -6.07836485e-01 -6.75565749e-02
-5.79461455e-01 -1.38695389e-01 -1.50294349e-01 -7.02120304e-01
3.53583023e-02 1.39279440e-02 -3.82234484e-01 6.28828049e-01
5.52487969e-01 9.20773089e-01 -5.57068169e-01 -2.34195277e-01
-1.64584249e-01 8.63064706e-01 -4.36527580e-01 3.52688581e-01
8.97320285e-02 5.31420112e-01 -7.53934383e-02 1.16928458e+00
1.13601851e+00 2.84619510e-01 -2.72095725e-02 -4.62839901e-01
2.42215395e-01 -4.81470078e-01 -7.58933902e-01 1.46871591e+00
-4.90122855e-01 2.90448368e-01 3.27322364e-01 -8.20725203e-01
1.08884740e+00 3.42529744e-01 4.96890455e-01 -8.15093637e-01
1.85364231e-01 2.61223167e-01 -2.12619528e-01 -9.14527476e-03
1.01685021e-02 -1.58158034e-01 2.88723677e-01 3.65365982e-01
3.94507259e-01 4.05902296e-01 -4.99766245e-02 -1.05615616e-01
9.18976188e-01 1.66975185e-01 -9.85436216e-02 -3.17836255e-01
6.92257166e-01 -1.11307228e+00 7.69094110e-01 3.03212196e-01
-5.77234685e-01 7.52593160e-01 3.04232299e-01 -6.28695369e-01
-9.14091110e-01 -1.33659029e+00 -1.70914650e-01 1.17818177e+00
-3.36212009e-01 -5.27678616e-02 -7.86850095e-01 -1.09250975e+00
2.71308720e-01 5.02410121e-02 -1.23930609e+00 -3.05428326e-01
-5.46150804e-01 -6.89040720e-01 6.99094415e-01 6.49072409e-01
6.38402343e-01 -8.06377232e-01 9.93690863e-02 -4.22615975e-01
2.46827722e-01 -1.11218786e+00 -7.07495928e-01 -5.20377576e-01
-6.25096858e-01 -1.13406277e+00 -1.01527941e+00 -7.36520350e-01
9.30306256e-01 2.19295084e-01 1.17413497e+00 1.02185138e-01
-5.50247490e-01 6.48906410e-01 -1.06102638e-01 -3.46821845e-01
-2.63313651e-01 -1.28284231e-01 5.39659858e-01 6.51270151e-01
3.17988575e-01 -9.64887977e-01 -9.23392653e-01 3.93946260e-01
-6.46515429e-01 -2.09896475e-01 4.75444764e-01 8.79930139e-01
2.18976691e-01 -3.18534076e-01 1.03619730e+00 -5.95579386e-01
1.76575288e-01 -1.48133963e-01 -5.88809729e-01 5.65240562e-01
-6.71572566e-01 -7.11859763e-02 3.63035709e-01 -4.75954622e-01
-1.21383321e+00 8.39895234e-02 -1.65349245e-01 -6.50032699e-01
-1.23886384e-01 -2.38765106e-01 -7.46381700e-01 -3.53753328e-01
5.76520741e-01 1.58417180e-01 4.46702778e-01 -4.30607408e-01
3.27145010e-01 6.18519723e-01 8.64676118e-01 -9.02475119e-01
1.29114771e+00 5.94595432e-01 2.22783208e-01 -6.66345537e-01
-7.95541167e-01 5.19992635e-02 -6.30315542e-01 -3.25210333e-01
6.05364442e-01 -1.28145397e+00 -1.04609728e+00 1.20110917e+00
-9.28356588e-01 -1.14111803e-01 -1.44552067e-01 3.88363888e-03
-3.14063936e-01 3.06566149e-01 -4.16505098e-01 -9.58902776e-01
-7.29545891e-01 -8.41269493e-01 1.46567690e+00 3.80909204e-01
1.30082980e-01 -7.10830927e-01 -1.20852441e-01 5.36020815e-01
4.85206097e-01 4.28280085e-01 6.36293709e-01 -2.58122623e-01
-4.20869887e-01 -2.59671479e-01 -4.26001996e-01 6.59389079e-01
5.23921788e-01 3.24142724e-01 -1.50901365e+00 -7.00709224e-01
-2.97050714e-01 -6.08914912e-01 9.80498910e-01 -7.22438022e-02
1.39163780e+00 -2.37037748e-01 -2.43878104e-02 9.51732993e-01
1.06329250e+00 -5.69593459e-02 8.58629882e-01 -2.67696142e-01
7.68531919e-01 8.17570508e-01 3.07023793e-01 4.38654870e-01
2.15224013e-01 6.52643561e-01 4.07510906e-01 -1.87527522e-01
-3.66191983e-01 -1.92403555e-01 5.83708286e-01 4.84654367e-01
-2.39688113e-01 1.41830221e-01 -5.55505514e-01 4.34591234e-01
-1.37396455e+00 -9.58011210e-01 4.92963046e-01 2.27174854e+00
9.75037992e-01 -4.70607013e-01 2.60777444e-01 1.29308775e-01
6.04997218e-01 1.94029078e-01 -7.54959047e-01 -1.15493901e-01
-3.62052530e-01 6.25085890e-01 1.06758572e-01 1.30624786e-01
-1.11352599e+00 6.34948134e-01 5.58442259e+00 6.94485426e-01
-1.31182539e+00 -4.85332273e-02 1.09556484e+00 -1.88066110e-01
-5.60037754e-02 -5.15043437e-01 -9.07561243e-01 3.23099226e-01
7.51748443e-01 -2.03477144e-01 7.67480373e-01 8.33888352e-01
-3.16522509e-01 5.50564349e-01 -1.50293934e+00 1.35294652e+00
4.07527357e-01 -9.32248771e-01 1.97190955e-01 2.08321810e-01
7.97345757e-01 -4.65334594e-01 7.25566089e-01 4.20927584e-01
7.99172446e-02 -1.29885459e+00 6.41555727e-01 5.21886051e-01
1.34381580e+00 -7.66948760e-01 2.33040124e-01 -3.34517777e-01
-1.21228266e+00 -2.13556796e-01 -1.89685583e-01 1.31629720e-01
-3.95108610e-01 5.24801910e-01 -2.25337163e-01 7.37614870e-01
7.00639069e-01 7.27830529e-01 -1.00315309e+00 5.00280201e-01
9.49812531e-02 2.42664501e-01 -3.75316888e-02 7.48848259e-01
-3.88838351e-01 -1.13060303e-01 1.05111003e-01 7.18343675e-01
5.11243403e-01 -1.70593292e-01 -3.25533211e-01 7.74035215e-01
-7.56685436e-01 -8.40605274e-02 -6.36897206e-01 -2.80676395e-01
4.76008624e-01 1.49236345e+00 -1.83397859e-01 1.14821678e-03
-3.72342318e-01 1.14019215e+00 4.97929364e-01 2.42808923e-01
-7.35027909e-01 3.99245173e-02 1.30813336e+00 -5.85196838e-02
2.70565003e-01 -1.09421693e-01 -4.72849011e-02 -1.22278106e+00
3.76397133e-01 -1.38748205e+00 3.46629143e-01 -3.85820210e-01
-1.86978352e+00 6.89178407e-01 -3.42746228e-01 -9.22513664e-01
-2.14197463e-03 -9.24828351e-01 -5.68966269e-01 9.85249937e-01
-1.61338866e+00 -1.84568942e+00 -4.58694458e-01 7.88263202e-01
3.52040052e-01 -4.75007832e-01 9.07528341e-01 6.74500406e-01
-8.38316500e-01 1.39917719e+00 -2.02307329e-01 2.50226676e-01
1.14367735e+00 -1.00984776e+00 6.27277970e-01 9.14899349e-01
2.74460781e-02 8.00886333e-01 2.99901575e-01 -3.85100037e-01
-1.77042079e+00 -1.21251249e+00 5.79718471e-01 -7.66144335e-01
2.99221754e-01 -7.17850983e-01 -9.74009991e-01 5.72830796e-01
3.77897508e-02 4.37622249e-01 6.66817665e-01 2.22796738e-01
-1.31229043e+00 -8.99237990e-01 -1.26379633e+00 6.91191614e-01
1.57517207e+00 -7.62901962e-01 -1.25172064e-01 1.29059017e-01
5.44076085e-01 1.81307457e-02 -1.06475210e+00 8.29698205e-01
1.14016449e+00 -1.00958192e+00 1.23230577e+00 -6.39506996e-01
3.07910383e-01 -1.13682270e-01 -2.60814756e-01 -1.08272433e+00
-1.60273597e-01 -8.32392871e-01 -3.59951854e-01 1.99350429e+00
2.27411068e-03 -6.90714478e-01 9.08106625e-01 7.46522725e-01
1.39685094e-01 -7.20260978e-01 -1.02602744e+00 -9.72747326e-01
3.39813143e-01 5.40637523e-02 9.70098615e-01 9.09885585e-01
-7.86808848e-01 8.90640914e-02 -5.32887876e-01 2.28671595e-01
9.42623317e-01 1.65124573e-02 9.03023183e-01 -1.48613894e+00
-9.42800045e-02 -4.37990189e-01 -5.21238863e-01 -4.75256354e-01
7.34586298e-01 -6.81416035e-01 -1.49698228e-01 -7.70710647e-01
2.88218051e-01 -3.19357127e-01 -3.84792268e-01 6.77719653e-01
-3.26187611e-01 5.99993885e-01 8.12946036e-02 -1.97453395e-01
-2.14094281e-01 9.72863674e-01 1.26244819e+00 -3.09272438e-01
3.95821661e-01 -1.47454932e-01 -9.54251409e-01 3.24735224e-01
5.20115793e-01 6.52931817e-03 -3.95259470e-01 -3.38742912e-01
-1.13488063e-01 -3.14745784e-01 4.41840261e-01 -1.01990187e+00
-2.86819842e-02 -7.17856586e-02 8.52444053e-01 -1.65045168e-02
4.84134555e-01 -7.15183675e-01 1.35202199e-01 2.30063155e-01
-2.98289508e-02 1.91740349e-01 2.74825305e-01 4.70305800e-01
-3.65430005e-02 4.62594032e-01 1.12241900e+00 3.10042024e-01
-4.36019808e-01 1.16035831e+00 3.91047180e-01 1.40039712e-01
9.38725889e-01 -2.51633018e-01 -4.54079092e-01 -1.47432804e-01
-2.85637170e-01 3.64018604e-02 5.87488830e-01 8.37244987e-01
5.33685982e-01 -1.59419370e+00 -1.05511200e+00 6.23538315e-01
3.74695271e-01 -4.39462990e-01 6.13512754e-01 5.41968465e-01
7.81943090e-03 -6.75463900e-02 -6.85604692e-01 -3.12131196e-01
-1.60671473e+00 7.56060183e-01 4.70678538e-01 4.76179123e-02
-6.33488223e-02 1.03030574e+00 6.15908980e-01 -4.37176377e-01
2.81905860e-01 4.34542358e-01 8.28056708e-02 1.85533300e-01
8.88258159e-01 2.00584620e-01 1.65247023e-01 -8.18783700e-01
-3.74667197e-01 7.13383853e-01 -3.70327562e-01 2.38841280e-01
1.44871569e+00 6.64634109e-02 -3.66474241e-01 -6.85853884e-03
1.24226594e+00 -7.91618899e-02 -1.62580657e+00 -2.17531532e-01
-3.62432688e-01 -7.65135646e-01 -2.25985706e-01 -8.71674776e-01
-1.60626626e+00 9.20325339e-01 9.99389112e-01 -3.34274739e-01
1.43639934e+00 -2.60152310e-01 5.97939610e-01 1.17603555e-01
3.69998962e-01 -7.11871326e-01 2.82673419e-01 1.06688462e-01
1.30111909e+00 -1.33622479e+00 -6.81495070e-02 -2.87396401e-01
-1.81414932e-01 9.62897360e-01 9.71365929e-01 2.26585478e-01
6.47668898e-01 1.04696088e-01 1.37981504e-01 3.88498195e-02
-7.36122608e-01 3.35393324e-02 4.62478280e-01 9.44608927e-01
4.86962497e-01 -9.57474858e-02 1.42503500e-01 6.33332014e-01
-1.93321526e-01 1.03122434e-02 -1.09901384e-01 6.00726664e-01
1.92330271e-01 -1.44210958e+00 -2.57922828e-01 2.88914680e-01
-5.05925179e-01 1.47251129e-01 -5.88411152e-01 6.16423905e-01
1.34430766e-01 6.37841523e-01 -1.64605435e-02 -3.95575255e-01
2.64495015e-01 3.36768657e-01 1.01726711e+00 -2.79819459e-01
-5.46590030e-01 -5.87656379e-01 -1.92680448e-01 -6.74814224e-01
-2.17148975e-01 -7.25992441e-01 -4.65829939e-01 -5.91691792e-01
7.87967145e-02 -4.02155608e-01 6.45923615e-01 5.98966599e-01
6.73210800e-01 2.75764644e-01 1.05828142e+00 -9.12081301e-01
-7.65885890e-01 -8.44735801e-01 -5.67394674e-01 7.00952411e-01
4.25873071e-01 -8.35507214e-01 -3.68950039e-01 9.63498726e-02] | [13.32457447052002, 0.6653309464454651] |
42054a3c-615a-41f6-a82d-3b269ac7551d | hybrid-relation-guided-set-matching-for-few | 2204.13423 | null | https://arxiv.org/abs/2204.13423v1 | https://arxiv.org/pdf/2204.13423v1.pdf | Hybrid Relation Guided Set Matching for Few-shot Action Recognition | Current few-shot action recognition methods reach impressive performance by learning discriminative features for each video via episodic training and designing various temporal alignment strategies. Nevertheless, they are limited in that (a) learning individual features without considering the entire task may lose the most relevant information in the current episode, and (b) these alignment strategies may fail in misaligned instances. To overcome the two limitations, we propose a novel Hybrid Relation guided Set Matching (HyRSM) approach that incorporates two key components: hybrid relation module and set matching metric. The purpose of the hybrid relation module is to learn task-specific embeddings by fully exploiting associated relations within and cross videos in an episode. Built upon the task-specific features, we reformulate distance measure between query and support videos as a set matching problem and further design a bidirectional Mean Hausdorff Metric to improve the resilience to misaligned instances. By this means, the proposed HyRSM can be highly informative and flexible to predict query categories under the few-shot settings. We evaluate HyRSM on six challenging benchmarks, and the experimental results show its superiority over the state-of-the-art methods by a convincing margin. Project page: https://hyrsm-cvpr2022.github.io/. | ['Nong Sang', 'Rong Jin', 'Changxin Gao', 'Zhengrong Zuo', 'Mingqian Tang', 'Zhiwu Qing', 'Shiwei Zhang', 'Xiang Wang'] | 2022-04-28 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Wang_Hybrid_Relation_Guided_Set_Matching_for_Few-Shot_Action_Recognition_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Wang_Hybrid_Relation_Guided_Set_Matching_for_Few-Shot_Action_Recognition_CVPR_2022_paper.pdf | cvpr-2022-1 | ['few-shot-action-recognition', 'set-matching'] | ['computer-vision', 'computer-vision'] | [ 2.04398304e-01 -4.09317404e-01 -4.66502577e-01 -3.36057872e-01
-8.24565113e-01 -1.98944837e-01 6.11257195e-01 -1.71712965e-01
-3.09445620e-01 4.00114447e-01 4.10415053e-01 3.94125015e-01
-4.42534983e-01 -4.93271202e-01 -5.19077599e-01 -7.73391962e-01
-1.54304951e-01 1.65909499e-01 5.84601581e-01 -1.89003170e-01
2.97415495e-01 2.35475451e-01 -1.67563581e+00 2.70674914e-01
7.50973344e-01 1.23795724e+00 2.11140037e-01 2.31445223e-01
8.27591270e-02 9.27796364e-01 -2.27769732e-01 -3.71873617e-01
3.76868725e-01 -5.43291330e-01 -7.11382747e-01 2.81670719e-01
2.76666880e-01 -3.99911404e-01 -7.96992600e-01 8.89732659e-01
4.48570490e-01 4.91053879e-01 5.14468074e-01 -1.53459811e+00
-5.70785224e-01 2.96222478e-01 -5.76791883e-01 5.22479892e-01
5.15539289e-01 3.09286863e-01 1.15851665e+00 -1.04558372e+00
5.44154644e-01 1.05724001e+00 6.13441885e-01 5.07180333e-01
-7.68716693e-01 -5.70135355e-01 3.55636358e-01 7.01476634e-01
-1.45152974e+00 -5.97131670e-01 9.18655992e-01 -5.29863119e-01
8.68960738e-01 2.64911175e-01 7.90097713e-01 1.18111801e+00
-9.35298800e-02 1.00979924e+00 7.04197407e-01 -8.13916773e-02
2.15520501e-01 -2.00723678e-01 1.35912836e-01 7.92281210e-01
-5.83808199e-02 -1.18115190e-02 -7.12808251e-01 2.86881998e-02
6.69219851e-01 5.42258143e-01 -4.53006744e-01 -6.48422599e-01
-1.45495927e+00 8.14541638e-01 4.46617663e-01 4.47733283e-01
-2.09459662e-01 3.00140362e-02 7.11829662e-01 2.77687937e-01
2.36937210e-01 3.41528624e-01 -1.44694343e-01 -4.19698089e-01
-7.61792421e-01 2.22161308e-01 3.74858022e-01 1.11595893e+00
7.55928457e-01 -3.80872428e-01 -4.16002333e-01 8.82918239e-01
1.78584427e-01 1.26644969e-01 8.99859071e-01 -8.74587178e-01
7.03309298e-01 7.46160686e-01 -3.42125222e-02 -1.28111541e+00
-1.94825590e-01 -2.40832269e-01 -7.49169409e-01 -3.12856436e-01
2.87229717e-01 3.28935564e-01 -6.98967457e-01 1.64232051e+00
3.54520917e-01 5.69286227e-01 -6.87151402e-02 9.92697954e-01
8.53281200e-01 4.30768788e-01 -1.79450363e-01 -3.05901796e-01
1.00753975e+00 -1.40826249e+00 -7.90837228e-01 -1.94597587e-01
9.01152730e-01 -4.89805311e-01 1.17648721e+00 -8.12074244e-02
-8.76736581e-01 -7.35050559e-01 -1.19781113e+00 7.48447776e-02
-2.00735033e-01 7.80446827e-02 4.82558995e-01 5.36444373e-02
-4.66150045e-01 7.80559421e-01 -8.52987766e-01 -4.30934966e-01
5.31353116e-01 1.79382816e-01 -3.77655238e-01 -3.38535607e-01
-1.28680634e+00 6.50800765e-01 2.15908587e-01 1.22072801e-01
-7.64925957e-01 -6.43579721e-01 -1.01550400e+00 -1.19240023e-01
7.17342734e-01 -4.67528343e-01 1.01152956e+00 -6.49666011e-01
-1.31846845e+00 7.29900420e-01 -1.98248439e-02 -3.82707000e-01
7.08447278e-01 -4.83825594e-01 -5.11180758e-01 3.56029153e-01
2.53516912e-01 3.22124124e-01 7.73663402e-01 -7.89871335e-01
-7.09418774e-01 -3.76646370e-01 7.50725418e-02 4.07517552e-01
-6.13818645e-01 -1.46315247e-01 -7.23344266e-01 -7.26933002e-01
1.28999889e-01 -8.02533507e-01 4.94318940e-02 9.05931443e-02
-1.27877831e-01 -3.52628738e-01 9.61436391e-01 -4.17848706e-01
1.46252525e+00 -2.32344341e+00 3.32309246e-01 -1.78495437e-01
5.15099987e-03 4.76348877e-01 -2.31379524e-01 6.14309192e-01
6.95575178e-02 -2.36200139e-01 -1.81552559e-01 -3.53173673e-01
6.35598749e-02 2.57579803e-01 -3.29163432e-01 6.72950625e-01
1.87278956e-01 1.05394912e+00 -1.10805082e+00 -7.26498604e-01
3.17589968e-01 3.42358708e-01 -4.11928415e-01 4.78047132e-01
2.75769904e-02 2.88350165e-01 -5.78157783e-01 7.48615205e-01
2.61223882e-01 -3.43106598e-01 -3.20085213e-02 -4.08455729e-01
1.02414876e-01 9.61794481e-02 -1.31729829e+00 2.05300641e+00
-1.15235940e-01 4.46476877e-01 -4.66234148e-01 -1.39981675e+00
9.50995386e-01 1.57106191e-01 8.63733590e-01 -6.78553700e-01
-7.79621452e-02 1.75934300e-01 -2.41293162e-01 -9.38970447e-01
2.74679214e-01 1.28354132e-01 -5.74859157e-02 1.83041930e-01
1.96889788e-01 2.57969588e-01 2.17813715e-01 1.36438280e-01
1.29530191e+00 3.69060516e-01 4.02057379e-01 4.68123928e-02
6.33764029e-01 -2.39711344e-01 1.08916342e+00 5.06957591e-01
-8.12797546e-01 6.23361230e-01 3.70066851e-01 -5.48502803e-01
-8.05340230e-01 -8.52093279e-01 2.67015900e-02 9.73827839e-01
6.61626518e-01 -6.26222014e-01 -4.13441896e-01 -1.02007389e+00
-5.29409498e-02 2.25657791e-01 -8.04278612e-01 -4.87616956e-01
-6.81750953e-01 -5.19601822e-01 2.69970655e-01 7.64794827e-01
8.57752204e-01 -9.42763448e-01 -6.50029600e-01 1.20858014e-01
-3.90335739e-01 -1.22973657e+00 -8.02296937e-01 -2.63724089e-01
-7.49949813e-01 -1.26375937e+00 -7.11794019e-01 -7.38479376e-01
2.77020603e-01 6.86275721e-01 8.60222399e-01 1.25536039e-01
-3.64417613e-01 5.20323992e-01 -7.91639328e-01 1.33167148e-01
2.64167905e-01 -7.94411600e-02 1.17859095e-01 4.11838412e-01
5.34187257e-01 -6.23567522e-01 -9.21670437e-01 7.17976213e-01
-7.68149257e-01 -6.10016286e-02 5.58553815e-01 1.01219618e+00
7.58011699e-01 -3.29942852e-02 6.08505309e-01 -4.73450124e-01
1.59719154e-01 -6.19250655e-01 -1.03798151e-01 4.54459548e-01
-5.37058949e-01 -2.17082828e-01 5.25046051e-01 -5.67285657e-01
-7.49812961e-01 3.93643789e-02 2.25766897e-01 -8.32921565e-01
8.03016573e-02 3.19288492e-01 -4.54945117e-01 1.05703309e-01
2.52646148e-01 5.01983106e-01 4.60073724e-02 -4.52858448e-01
1.96910515e-01 5.03341496e-01 5.20630777e-01 -3.56775194e-01
7.51373827e-01 5.24423003e-01 -5.82615957e-02 -6.83489203e-01
-1.08934355e+00 -8.63461912e-01 -8.64906311e-01 -3.84003013e-01
8.29285026e-01 -9.17050302e-01 -3.52482826e-01 4.95071590e-01
-6.41149521e-01 -3.04716408e-01 -4.65815216e-01 6.69941425e-01
-9.29237247e-01 5.85529685e-01 -3.61283094e-01 -5.21801353e-01
-6.42348975e-02 -1.16359830e+00 1.07531023e+00 2.54268259e-01
-3.86494696e-02 -7.88080454e-01 1.89864218e-01 5.41518211e-01
6.85699359e-02 3.25309724e-01 4.35560346e-01 -8.51071656e-01
-5.65024853e-01 -2.41805971e-01 -1.77497104e-01 2.67721266e-01
2.85145819e-01 -3.08051240e-02 -7.23795950e-01 -3.74636948e-01
8.32775012e-02 -5.32530069e-01 1.01289511e+00 7.05181211e-02
1.18365860e+00 -2.54120708e-01 -3.65556955e-01 7.45784760e-01
1.35539103e+00 2.25096896e-01 7.24748611e-01 5.81293344e-01
7.11577594e-01 4.32858825e-01 1.22817588e+00 6.15700245e-01
3.84374857e-01 1.05481935e+00 3.70138109e-01 2.99734563e-01
-1.48471788e-01 -3.45988452e-01 5.82877159e-01 9.28928494e-01
-9.39180702e-02 5.36602437e-02 -5.18554747e-01 6.26154065e-01
-2.35216379e+00 -1.36221671e+00 2.34166861e-01 2.11768794e+00
6.02443218e-01 9.66372043e-02 4.35660511e-01 3.09627593e-01
7.60266662e-01 5.23515999e-01 -7.08067060e-01 2.29057208e-01
-5.17766923e-02 -2.88193733e-01 6.34293035e-02 6.60899933e-03
-1.34558451e+00 7.22985387e-01 4.80973530e+00 9.55705881e-01
-7.96401203e-01 1.93381757e-01 2.61410356e-01 -2.46186629e-01
1.23622596e-01 -3.55960801e-02 -7.26603389e-01 7.14075089e-01
4.21979845e-01 -1.75987110e-01 2.30597898e-01 8.73554349e-01
-3.18750329e-02 1.34088695e-01 -1.23760676e+00 1.25489712e+00
3.23060870e-01 -1.25168395e+00 -5.52340113e-02 -9.39609110e-02
6.35417163e-01 -1.82357952e-01 5.05225584e-02 5.15806675e-01
-2.81047106e-01 -7.97567248e-01 5.18381000e-01 7.62269855e-01
4.96391535e-01 -6.06698573e-01 6.34659171e-01 1.47806108e-01
-1.53256357e+00 -4.07493353e-01 -4.86510962e-01 -2.97284685e-03
1.50192022e-01 3.76081884e-01 -2.35321328e-01 7.32186556e-01
8.39040458e-01 1.38339400e+00 -6.37395978e-01 1.20916057e+00
2.16900222e-02 3.34983408e-01 4.07409482e-03 1.01350754e-01
3.00399899e-01 -2.85856307e-01 6.75103247e-01 1.01676989e+00
2.69385338e-01 1.21530190e-01 3.75027299e-01 4.47873324e-01
6.74228072e-02 1.92266017e-01 -7.08959639e-01 -1.90354921e-02
6.13797903e-01 1.15054393e+00 -4.89761889e-01 -2.69022286e-01
-6.72015786e-01 1.21553493e+00 4.88947958e-01 2.39612609e-01
-1.09341764e+00 -4.04025763e-01 8.95322084e-01 -2.14760825e-02
5.72179019e-01 -3.36998608e-04 1.59885570e-01 -1.47321224e+00
4.85639244e-01 -8.10867906e-01 6.99238658e-01 -5.70803404e-01
-1.24667323e+00 4.04424191e-01 -7.52954185e-02 -1.82114816e+00
-4.72751856e-02 -3.30968499e-01 -6.06746674e-01 1.61851436e-01
-1.40137410e+00 -1.18046749e+00 -4.97454077e-01 7.57399797e-01
8.25657129e-01 -3.47206712e-01 6.74327374e-01 4.95771080e-01
-9.56569076e-01 7.70843923e-01 1.56071097e-01 2.47656286e-01
8.36760521e-01 -1.05115175e+00 3.34533900e-02 7.26102352e-01
2.31084079e-01 3.24159682e-01 4.02986586e-01 -4.13805395e-01
-1.47648180e+00 -1.28248441e+00 6.47139490e-01 -3.54781270e-01
7.20551848e-01 -2.56645083e-01 -1.00150967e+00 7.01454580e-01
-2.15964496e-01 4.32889044e-01 7.50144005e-01 -9.76929814e-02
-5.15167952e-01 -3.71435702e-01 -8.21584702e-01 4.86176670e-01
1.52445900e+00 -5.93982816e-01 -8.20434928e-01 5.01735806e-01
7.24716902e-01 -1.96644679e-01 -1.17065358e+00 6.75347984e-01
5.70496142e-01 -1.18654644e+00 9.27309692e-01 -7.33041286e-01
4.70431358e-01 -3.19414705e-01 -4.37911153e-01 -1.05072451e+00
-3.66136998e-01 -7.10879028e-01 -5.80021918e-01 1.46919107e+00
-7.58455247e-02 -4.43177283e-01 5.04853666e-01 4.00324017e-01
-2.28582591e-01 -1.25892210e+00 -9.62527633e-01 -1.17346883e+00
-3.41213822e-01 -2.42609650e-01 6.82455063e-01 9.74823356e-01
2.26193964e-01 2.10750431e-01 -6.39541507e-01 -7.71675855e-02
5.32978535e-01 4.36853677e-01 7.99949288e-01 -8.21400166e-01
-3.68924975e-01 -4.76849556e-01 -8.82587314e-01 -1.17092180e+00
6.11383766e-02 -7.05587983e-01 -4.04517800e-02 -1.33976626e+00
4.50843900e-01 -3.79418492e-01 -5.89501500e-01 3.64391446e-01
-3.58306378e-01 1.58926040e-01 2.95849621e-01 5.69554448e-01
-1.20946574e+00 1.03133607e+00 1.17823088e+00 -1.08613193e-01
-8.63940120e-02 -4.13938798e-02 -4.46356475e-01 6.61884427e-01
6.48496985e-01 -2.82442242e-01 -5.61499059e-01 -2.54770935e-01
-2.36332849e-01 -7.97766596e-02 4.77926254e-01 -1.28553319e+00
1.96829140e-01 -2.88449526e-01 2.06676051e-01 -4.62466836e-01
4.39457864e-01 -7.07989454e-01 5.20503800e-03 3.03550184e-01
-4.73324895e-01 -1.34463519e-01 -2.64820874e-01 8.68606806e-01
-4.14482266e-01 -8.48706067e-02 6.88254118e-01 1.21593356e-01
-1.18009472e+00 7.23552823e-01 2.92789936e-01 3.30121696e-01
1.39239490e+00 -4.36819226e-01 -2.88043499e-01 -1.93816900e-01
-5.64011395e-01 4.37737823e-01 5.50591111e-01 7.11245358e-01
7.60769606e-01 -1.76606703e+00 -4.12086546e-01 8.63980204e-02
5.29839516e-01 -1.27294123e-01 5.97570479e-01 1.24316943e+00
2.17173286e-02 2.48990148e-01 -2.17668414e-01 -6.06640816e-01
-1.21334529e+00 7.38143384e-01 2.86627322e-01 -2.77655393e-01
-9.41325724e-01 7.84476638e-01 2.68259466e-01 -9.18552577e-02
4.90930855e-01 4.51043211e-02 -3.30421835e-01 2.04752028e-01
7.36040235e-01 4.96052712e-01 -1.99398607e-01 -7.55614519e-01
-5.44832230e-01 8.71912122e-01 -1.11270443e-01 2.74287701e-01
1.42553008e+00 -2.52952814e-01 3.76417696e-01 6.46077991e-01
1.52449286e+00 -4.56654817e-01 -1.57326162e+00 -5.58461785e-01
-3.18250582e-02 -8.82851958e-01 -2.51826227e-01 -3.33285272e-01
-1.15567279e+00 7.95284927e-01 6.14159465e-01 1.01971731e-01
1.27342486e+00 7.85147846e-02 1.05573428e+00 2.86949635e-01
1.65764943e-01 -1.24599910e+00 6.34440899e-01 2.33947635e-01
9.27601099e-01 -1.34198141e+00 7.09154382e-02 -2.83632964e-01
-9.65986252e-01 9.95437324e-01 7.71670938e-01 -2.07965121e-01
6.68546915e-01 -2.92043149e-01 -2.03469992e-01 -3.72499973e-01
-6.73805773e-01 -3.44763637e-01 4.04475391e-01 5.44370294e-01
1.36589497e-01 -2.11927474e-01 -4.28743899e-01 6.58555329e-01
2.47424856e-01 1.48019075e-01 1.37705162e-01 1.06117201e+00
-4.30908173e-01 -8.35550547e-01 9.02416259e-02 5.13783336e-01
-1.81544170e-01 3.48706603e-01 -1.74388587e-01 8.75073195e-01
5.67275807e-02 7.88895845e-01 8.91900733e-02 -8.25100422e-01
5.00965834e-01 -6.74471855e-02 4.08257961e-01 -4.49121386e-01
-4.72375900e-02 -9.55485404e-02 -6.79663420e-02 -1.01264524e+00
-7.52727509e-01 -9.78180230e-01 -1.13065922e+00 -6.27413392e-02
-3.36664438e-01 2.23452747e-02 5.58471307e-03 1.07789028e+00
5.37142873e-01 4.04015452e-01 9.76579547e-01 -7.65927792e-01
-9.05240774e-01 -9.24262047e-01 -6.27886713e-01 8.18445444e-01
2.30496973e-01 -1.20297360e+00 -3.93381119e-01 6.79086223e-02] | [8.523598670959473, 0.803330659866333] |
202490b8-ebf7-4159-ade2-51d1bc6ce0bd | unsupervised-continual-learning-and-self-1 | null | null | https://openreview.net/forum?id=SJxakiC4u4 | https://openreview.net/pdf?id=SJxakiC4u4 | Unsupervised Continual Learning and Self-Taught Associative Memory Hierarchies | We first pose the Unsupervised Continual Learning (UCL) problem: learning salient representations from a non-stationary stream of unlabeled data in which the number of object classes varies with time. Given limited labeled data just before inference, those representations can also be associated with specific object types to perform classification. To solve the UCL problem, we propose an architecture that involves a single module, called Self-Taught Associative Memory (STAM), which loosely models the function of a cortical column in the mammalian brain. Hierarchies of STAM modules learn based on a combination of Hebbian learning, online clustering, detection of novel patterns and forgetting outliers, and top-down predictions. We illustrate the operation of STAMs in the context of learning handwritten digits in a continual manner with only 3-12 labeled examples per class. STAMs suggest a promising direction to solve the UCL problem without catastrophic forgetting. | ['Constantine Dovrolis', 'Zsolt Kira', 'Seth Baer', 'James Smith'] | 2019-03-24 | null | null | null | iclr-workshop-lld-2019 | ['online-clustering'] | ['computer-vision'] | [ 2.93097407e-01 2.19161451e-01 3.37765515e-02 -4.42594141e-01
-9.77645516e-02 -2.59403259e-01 4.22858417e-01 5.53496003e-01
-3.97383153e-01 7.59134531e-01 -8.90163034e-02 -2.16637403e-02
-2.29783222e-01 -6.20173097e-01 -1.12341511e+00 -4.94889051e-01
-4.40907359e-01 7.41232514e-01 6.46610200e-01 -5.71150668e-02
5.35750449e-01 4.62156296e-01 -1.93198764e+00 6.91679418e-01
8.53194833e-01 8.91222596e-01 2.91979015e-01 5.64165711e-01
-4.81516212e-01 9.93882239e-01 -4.98860389e-01 -1.00739673e-01
1.67278126e-02 -5.91614544e-01 -9.18743610e-01 3.44405949e-01
3.19511801e-01 2.25458235e-01 -5.57451211e-02 1.01260662e+00
-7.94801489e-02 4.52233642e-01 1.02614903e+00 -1.29463267e+00
-9.39198852e-01 5.55633068e-01 -4.98315245e-01 4.98940349e-01
-7.30586722e-02 -4.60631549e-02 6.57008946e-01 -1.38833952e+00
5.55462658e-01 1.33036840e+00 6.93031907e-01 9.30030346e-01
-1.45413363e+00 -4.43829983e-01 5.46261311e-01 5.55759728e-01
-1.13481343e+00 -2.64251590e-01 5.55675447e-01 -7.10143387e-01
9.26230252e-01 1.60234481e-01 8.42494369e-01 1.07506192e+00
5.77460170e-01 8.63245487e-01 8.44071925e-01 -4.20449138e-01
9.14632738e-01 3.00956294e-02 6.47439182e-01 7.02345967e-01
3.88979256e-01 -1.35946527e-01 -1.19629025e+00 -1.07713431e-01
4.96141672e-01 5.82417846e-01 1.15058154e-01 -6.58691347e-01
-9.89434838e-01 6.29536331e-01 5.64282000e-01 2.47528359e-01
-2.17206120e-01 8.06505140e-03 1.98950008e-01 5.38013279e-01
4.38039213e-01 4.01746243e-01 -6.93107486e-01 4.90800768e-01
-9.49392140e-01 -2.75424093e-01 4.15455788e-01 9.30595398e-01
1.04936719e+00 2.20125780e-01 1.26971811e-01 7.92019308e-01
2.05348760e-01 1.35216981e-01 1.23226666e+00 -6.59102738e-01
-2.58973271e-01 8.62973571e-01 -1.44390866e-01 -4.58193988e-01
-2.84661114e-01 -4.91186380e-01 -9.60966051e-01 4.34117526e-01
1.82895854e-01 2.19227046e-01 -1.32065952e+00 1.79430425e+00
1.41354248e-01 4.84052747e-01 1.55836999e-04 4.29947138e-01
3.64820153e-01 6.27904415e-01 2.70279586e-01 -5.29274464e-01
7.09782958e-01 -9.54901576e-01 -4.06435937e-01 -3.91424477e-01
1.84627339e-01 -1.27649471e-01 1.17941785e+00 5.62061131e-01
-8.56354117e-01 -6.87349737e-01 -1.08597028e+00 1.58007607e-01
-7.09625423e-01 -4.70322073e-01 4.06588346e-01 -5.79012632e-02
-1.09560513e+00 1.01996505e+00 -9.77875829e-01 -5.66404164e-01
4.72186118e-01 3.11912775e-01 -1.06062412e-01 -8.77957791e-03
-5.66581130e-01 8.08771551e-01 6.05292499e-01 -2.30767608e-01
-1.20788991e+00 -6.17764533e-01 -6.85288370e-01 1.57487303e-01
5.98060861e-02 -5.11461735e-01 1.05259013e+00 -1.34638500e+00
-9.84872222e-01 8.05573344e-01 -3.44596684e-01 -7.49125123e-01
1.84892491e-01 -4.63845491e-01 -4.89733040e-01 -4.85113487e-02
1.16897978e-01 7.78385162e-01 1.42004871e+00 -1.39477408e+00
-6.75473571e-01 -5.10157108e-01 -7.12712646e-01 -7.90632814e-02
-5.22142887e-01 -6.53151691e-01 1.85943753e-01 -8.33334863e-01
5.76968968e-01 -6.68406129e-01 -2.16477826e-01 -2.32378423e-01
-3.24082002e-02 -3.68484885e-01 8.74351561e-01 -3.76629412e-01
1.07189631e+00 -2.45078611e+00 2.94899285e-01 1.61364134e-02
1.48117915e-01 -1.26449645e-01 -2.51786411e-01 2.38035306e-01
-2.95094848e-01 -1.57474354e-01 -4.79473859e-01 -1.20985799e-01
-5.16158879e-01 3.86558294e-01 -8.29742074e-01 1.44630179e-01
3.37051839e-01 8.08087230e-01 -1.21344602e+00 -3.30547869e-01
-2.12712407e-01 -7.17578083e-02 -5.54024518e-01 2.29371488e-01
-4.67091262e-01 4.41321820e-01 1.89979494e-01 7.18864918e-01
2.88078398e-01 -5.86303532e-01 9.24233273e-02 2.46120229e-01
5.31940050e-02 -6.07034005e-02 -1.09152865e+00 1.66665685e+00
3.67462635e-02 5.59241116e-01 -4.51329678e-01 -1.07472408e+00
1.00661778e+00 4.37327251e-02 2.45340571e-01 -3.64663869e-01
-3.71348411e-01 6.27410635e-02 3.29064950e-02 -2.67603904e-01
3.03729475e-01 -1.40913114e-01 1.79508731e-01 4.97982115e-01
8.55760992e-01 2.67453104e-01 1.71037301e-01 3.72382939e-01
1.32370460e+00 -8.12518150e-02 3.94316286e-01 -5.26539266e-01
1.26792446e-01 1.05358414e-01 8.89139950e-01 1.12082028e+00
-1.84598327e-01 5.41644275e-01 -1.57395261e-03 -9.03500617e-01
-1.00784743e+00 -1.61795259e+00 -4.68627363e-02 1.48835123e+00
1.83070719e-01 -2.04359010e-01 -3.92704993e-01 -6.20588243e-01
2.21626431e-01 7.61890590e-01 -8.51697147e-01 -6.98226929e-01
-2.49107420e-01 -8.05194378e-01 -2.37950742e-01 4.29336101e-01
3.01942110e-01 -1.57584977e+00 -8.05878997e-01 3.02632183e-01
3.24197948e-01 -1.25124574e-01 -2.33611107e-01 8.98485243e-01
-1.49287724e+00 -1.12065279e+00 -4.11782444e-01 -1.12897801e+00
1.11732447e+00 3.55474144e-01 1.14799380e+00 1.43678337e-01
-6.15789592e-01 6.47528291e-01 -2.86488205e-01 -4.74171609e-01
-3.26684356e-01 -3.65084484e-02 6.56557620e-01 2.37403661e-01
3.70951861e-01 -9.79285955e-01 -4.50178713e-01 -2.66474839e-02
-9.59175766e-01 5.93394972e-02 6.77125812e-01 1.14182425e+00
7.14715064e-01 6.32784590e-02 9.46480811e-01 -1.12128317e+00
2.55476028e-01 -7.93171942e-01 -3.85003418e-01 5.30693650e-01
-9.21985269e-01 3.55375499e-01 5.79916596e-01 -8.87657642e-01
-9.35869157e-01 3.61543745e-01 5.18755674e-01 -5.42497635e-01
-1.66201577e-01 2.22525120e-01 1.88564301e-01 1.92054376e-01
9.52713251e-01 6.38411820e-01 -8.77796561e-02 -6.47748113e-01
4.34237242e-01 2.28991792e-01 8.33280325e-01 -4.01559740e-01
7.37531602e-01 3.75303596e-01 -2.91697949e-01 -7.56048262e-01
-1.16847706e+00 -2.96626210e-01 -1.09536588e+00 -4.53365654e-01
4.36796755e-01 -8.66989911e-01 -4.42095041e-01 3.38444293e-01
-8.83024275e-01 -4.82607067e-01 -9.95302260e-01 2.32501447e-01
-5.82001328e-01 2.20443875e-01 -8.52294743e-01 -8.34050655e-01
-2.28818312e-01 -3.56357604e-01 4.69273776e-01 3.58761549e-01
-4.45436180e-01 -6.05884552e-01 1.08102053e-01 -9.01365802e-02
2.03090623e-01 -2.11607888e-01 1.27733457e+00 -8.74288797e-01
-5.59587836e-01 2.19132110e-01 3.70841235e-01 2.22778320e-01
2.59846717e-01 -2.54458576e-01 -1.09129548e+00 -7.13284433e-01
6.05398789e-02 -5.23474872e-01 1.47669017e+00 1.87483624e-01
1.40679860e+00 -3.73897433e-01 -3.18393648e-01 3.40264887e-01
1.22442663e+00 4.45910752e-01 2.25117639e-01 2.92085022e-01
3.45359504e-01 4.92474347e-01 4.55895603e-01 6.04321301e-01
-1.98899075e-01 -4.58642721e-01 4.01097119e-01 2.31648937e-01
-1.06068797e-01 -4.76992995e-01 3.80272150e-01 1.18393910e+00
2.02648491e-01 2.40521058e-01 -9.22476888e-01 6.69775784e-01
-2.07683015e+00 -1.01669538e+00 2.93619335e-01 2.29327703e+00
8.94946575e-01 4.93184716e-01 -1.89630568e-01 1.95117161e-01
8.46779287e-01 -3.74121279e-01 -1.26409960e+00 -8.80880952e-02
-2.39713207e-01 7.94296712e-02 1.31005958e-01 3.45888346e-01
-1.05441761e+00 8.93683255e-01 7.14050102e+00 4.90433306e-01
-9.20653939e-01 1.28371835e-01 7.67455637e-01 1.95325464e-02
-1.14764005e-01 2.08085254e-02 -6.37709022e-01 4.17174578e-01
9.83353257e-01 -3.32939863e-01 5.26583016e-01 1.06149447e+00
-4.29246038e-01 -2.32759789e-01 -1.36801004e+00 9.44640756e-01
2.53614247e-01 -1.33693123e+00 5.40064216e-01 -3.79876792e-01
9.28594768e-01 -6.17543086e-02 3.02693009e-01 5.84608495e-01
5.91060579e-01 -8.04689467e-01 6.18963659e-01 9.03629661e-01
4.28987175e-01 -5.18649161e-01 9.62119102e-02 7.12643504e-01
-8.08579981e-01 -6.15044236e-01 -7.20123887e-01 -5.10122553e-02
-4.48757589e-01 4.69317466e-01 -8.36023986e-01 -1.14662781e-01
1.04480183e+00 9.59413290e-01 -1.05896533e+00 1.39385450e+00
-7.78541490e-02 7.32172430e-01 6.18498474e-02 1.21691942e-01
-1.63343653e-01 -2.37756968e-02 4.17134672e-01 9.65988934e-01
4.22351420e-01 6.41059130e-02 8.14784095e-02 7.77262092e-01
-1.16652787e-01 -7.75463507e-02 -6.87625825e-01 3.25615197e-01
6.64183319e-01 9.25112367e-01 -1.19919074e+00 -6.51328564e-01
-1.28299788e-01 1.29589283e+00 8.25760543e-01 3.97290766e-01
-1.62221104e-01 5.70388464e-03 3.09410423e-01 2.41111070e-01
2.18941271e-01 -1.29740700e-01 -4.01735634e-01 -1.32567716e+00
-2.89109200e-01 -5.22435546e-01 7.93681026e-01 -7.67012715e-01
-1.84224439e+00 4.26500112e-01 -4.18628871e-01 -1.32909358e+00
-4.55478169e-02 -4.73487973e-01 -6.21196747e-01 1.22710153e-01
-1.07498932e+00 -6.10503137e-01 -1.68365330e-01 8.50166380e-01
1.06739938e+00 -6.34174287e-01 1.09852397e+00 -9.25961956e-02
-4.27253544e-01 2.17444137e-01 5.58291376e-01 -1.48912907e-01
8.36021602e-01 -1.46769726e+00 2.07770467e-01 8.01805556e-01
5.09130716e-01 7.15534389e-01 5.45659423e-01 -9.02593672e-01
-1.04637563e+00 -1.49527872e+00 7.12661564e-01 -5.09547710e-01
6.12665832e-01 -5.14159560e-01 -1.41981804e+00 9.24711227e-01
-8.81350115e-02 1.82483476e-02 8.40767562e-01 2.08219320e-01
-5.16703844e-01 -3.89811397e-01 -9.15766537e-01 4.96748477e-01
9.98071492e-01 -2.50894964e-01 -1.20434022e+00 4.50853139e-01
9.16473866e-01 1.80736765e-01 -2.71205932e-01 2.94463485e-01
8.84945393e-02 -8.14518332e-01 7.34871328e-01 -1.13939953e+00
1.97852224e-01 -4.40304548e-01 -6.73063844e-02 -1.46191120e+00
-8.03758860e-01 -2.32321143e-01 -6.01492345e-01 9.49671805e-01
3.57590705e-01 -3.13998133e-01 7.44562328e-01 3.85930687e-01
-1.90146595e-01 -3.85747671e-01 -9.94889200e-01 -1.00454986e+00
-1.26769751e-01 -2.27288287e-02 8.64003301e-02 1.06969750e+00
1.92468628e-01 7.48478770e-01 -4.61677164e-01 1.04030788e-01
9.35064733e-01 4.00113761e-01 2.26213202e-01 -1.74879420e+00
-2.49818861e-01 -1.37405068e-01 -4.41660672e-01 -7.14121461e-01
1.28428623e-01 -1.09703994e+00 2.30746269e-01 -9.61290956e-01
5.86665869e-01 -1.01079278e-01 -9.02669132e-01 6.29392266e-01
-2.01073050e-01 -9.11693126e-02 8.36263448e-02 6.58404827e-01
-1.18359768e+00 6.38050973e-01 6.14594042e-01 -4.30717051e-01
-3.07494462e-01 -4.91373949e-02 -3.05593491e-01 9.34156895e-01
6.61806464e-01 -6.69384837e-01 -3.55626643e-01 -2.84857541e-01
-2.38127187e-02 -2.22451523e-01 1.97028920e-01 -1.35749054e+00
6.19871914e-01 -1.48587376e-01 1.11891592e+00 -7.70147264e-01
1.77068084e-01 -8.12736988e-01 1.87108461e-02 8.79278600e-01
-6.17099464e-01 1.35673508e-01 -1.56745631e-02 1.09014881e+00
-2.78704137e-01 -2.27331996e-01 1.06351399e+00 -4.78213102e-01
-1.27276099e+00 2.38699958e-01 -7.92284608e-01 -1.80150922e-02
1.03931439e+00 7.17375576e-02 -1.84058622e-01 -4.66463603e-02
-1.49430859e+00 2.47286916e-01 2.08514884e-01 4.92007762e-01
1.07516038e+00 -1.30739391e+00 -3.32027406e-01 5.48796058e-01
1.92284584e-01 2.72715297e-02 1.69034585e-01 4.12741378e-02
-2.24282816e-02 -8.45500305e-02 -6.23544812e-01 -7.55373776e-01
-9.13609326e-01 1.19742966e+00 1.13510668e-01 1.40553087e-01
-5.60651600e-01 8.78783822e-01 2.45632261e-01 -2.31154248e-01
5.48704147e-01 -1.74634997e-02 -2.16004908e-01 1.99984461e-01
6.86202049e-01 2.53047019e-01 3.78625095e-02 4.03712094e-02
-2.11791292e-01 5.56363910e-02 -2.44203076e-01 3.11640445e-02
1.56804645e+00 -1.85865760e-01 -3.04423809e-01 1.51870477e+00
5.15546322e-01 -6.92874491e-01 -1.35919476e+00 -4.15734589e-01
6.91001177e-01 -1.00282386e-01 -6.78404987e-01 -6.97653711e-01
-6.52728975e-01 8.53823721e-01 9.32654440e-01 -9.01120231e-02
1.03731143e+00 9.62777585e-02 3.51105988e-01 8.91942561e-01
6.30662918e-01 -1.42935979e+00 9.05895472e-01 8.68389487e-01
9.06164765e-01 -1.33851671e+00 -2.96337247e-01 2.97873795e-01
-4.09279436e-01 1.25552654e+00 1.05902708e+00 -3.65985125e-01
9.39250946e-01 8.09343308e-02 -1.96687654e-01 3.38868616e-04
-1.28063452e+00 5.31996116e-02 1.65164515e-01 5.56164265e-01
2.41297767e-01 -5.82529791e-02 3.26647758e-02 6.06857955e-01
2.99011786e-02 -2.35342860e-01 4.18661207e-01 1.16616571e+00
-1.20291853e+00 -6.87557757e-01 -3.16030741e-01 7.34583676e-01
9.17187780e-02 -1.08004674e-01 -3.18958819e-01 3.18870038e-01
2.69336998e-01 5.73420286e-01 3.49285036e-01 -4.78625923e-01
5.51270209e-02 8.51371765e-01 2.93631464e-01 -1.04767728e+00
-2.82625794e-01 -1.36802047e-01 -9.49551404e-01 -3.40176880e-01
-2.08519921e-01 -7.27325916e-01 -1.29060328e+00 1.29907668e-01
1.96140483e-01 2.30606347e-01 6.78697228e-02 7.78098702e-01
4.15676862e-01 4.99602109e-01 5.49167216e-01 -6.39046073e-01
-4.90305126e-01 -1.06374228e+00 -9.69897389e-01 7.57683158e-01
5.79536378e-01 -5.70849359e-01 -2.66515046e-01 6.94673419e-01] | [9.824429512023926, 3.3538100719451904] |
dbdc2e61-4689-4526-acf4-8ac0d0954260 | on-the-n-gram-approximation-of-pre-trained | 2306.06892 | null | https://arxiv.org/abs/2306.06892v1 | https://arxiv.org/pdf/2306.06892v1.pdf | On the N-gram Approximation of Pre-trained Language Models | Large pre-trained language models (PLMs) have shown remarkable performance across various natural language understanding (NLU) tasks, particularly in low-resource settings. Nevertheless, their potential in Automatic Speech Recognition (ASR) remains largely unexplored. This study investigates the potential usage of PLMs for language modelling in ASR. We compare the application of large-scale text sampling and probability conversion for approximating GPT-2 into an n-gram model. Furthermore, we introduce a vocabulary-restricted decoding method for random sampling, and evaluate the effects of domain difficulty and data size on the usability of generated text. Our findings across eight domain-specific corpora support the use of sampling-based approximation and show that interpolating with a large sampled corpus improves test perplexity over a baseline trigram by 15%. Our vocabulary-restricted decoding method pushes this improvement further by 5% in domain-specific settings. | ['Dietrich Klakow', 'Jesujoba Alabi', 'Aravind Krishnan'] | 2023-06-12 | null | null | null | null | ['automatic-speech-recognition'] | ['speech'] | [ 4.00615275e-01 3.55471730e-01 -3.50298166e-01 -4.64870542e-01
-1.48237145e+00 -3.26745689e-01 8.27001750e-01 1.20841272e-01
-6.12011194e-01 7.42416382e-01 6.22328877e-01 -9.32478845e-01
3.40641797e-01 -2.84356385e-01 -6.98633134e-01 -1.03091955e-01
4.24953014e-01 8.66439342e-01 1.70412660e-01 -4.89727519e-02
8.36172327e-02 3.15724641e-01 -1.24728298e+00 4.77136940e-01
9.74740565e-01 3.99010450e-01 6.77298665e-01 8.22674632e-01
-4.22872096e-01 6.71003938e-01 -1.06706619e+00 -2.78061450e-01
-1.14965819e-01 -2.74537832e-01 -7.22191453e-01 2.00761542e-01
4.41234410e-01 -5.45446813e-01 -5.02484202e-01 5.63077033e-01
5.70198953e-01 3.11452389e-01 6.59976959e-01 -6.35448039e-01
-5.67971706e-01 9.81535792e-01 1.88365038e-02 2.25049183e-01
4.63986009e-01 1.24378309e-01 7.99453557e-01 -7.85341501e-01
3.58582109e-01 1.52510965e+00 4.85682368e-01 7.11922169e-01
-1.43290889e+00 -5.37923872e-01 7.63994232e-02 -3.52562398e-01
-1.64785802e+00 -9.42582846e-01 1.78108364e-01 -1.86235711e-01
1.74451375e+00 1.99321151e-01 1.65281191e-01 1.26871204e+00
-2.46087134e-01 1.00209308e+00 9.54216003e-01 -9.34742630e-01
2.65992403e-01 4.99421090e-01 1.20226249e-01 1.51948363e-01
2.13264257e-01 -2.70185620e-01 -4.58761007e-01 -2.82748252e-01
6.62290931e-01 -5.30044556e-01 -2.38447085e-01 3.56456846e-01
-1.07214689e+00 9.06493008e-01 -5.55377781e-01 2.08193526e-01
-2.50455618e-01 -1.13428138e-01 4.05879885e-01 2.62337685e-01
7.80816019e-01 5.68652689e-01 -5.45426965e-01 -6.71777964e-01
-1.25817883e+00 1.28923208e-01 9.15158093e-01 1.27460146e+00
3.20357412e-01 4.67782736e-01 -3.40181857e-01 1.62207782e+00
2.76779801e-01 7.85427868e-01 1.00013494e+00 -5.84467173e-01
6.89040780e-01 2.65434712e-01 1.82778358e-01 -8.41890052e-02
-8.05158634e-03 -2.50681311e-01 -5.59951216e-02 -4.13426369e-01
5.03076196e-01 -1.63130715e-01 -1.12777746e+00 1.60945201e+00
-6.92205727e-02 -7.88231865e-02 1.92708433e-01 3.96860451e-01
3.35545897e-01 1.09871936e+00 4.48068202e-01 -3.27708215e-01
1.21052730e+00 -9.02041137e-01 -6.79025471e-01 -6.68509245e-01
1.11478209e+00 -8.60755265e-01 1.66656995e+00 4.64645356e-01
-1.18422830e+00 -4.86452669e-01 -9.16301310e-01 -1.55574441e-01
-3.77202667e-02 5.18686771e-01 1.66767642e-01 1.21032548e+00
-1.07157743e+00 2.36696884e-01 -8.18942964e-01 -5.25682092e-01
6.73501566e-02 2.08774284e-01 -1.16330944e-01 -3.76427770e-01
-9.55452740e-01 9.32483852e-01 3.71156365e-01 -4.95062798e-01
-7.95372486e-01 -6.00519300e-01 -8.28195393e-01 8.62117391e-03
1.54994965e-01 -3.28017145e-01 1.73859751e+00 -6.12642229e-01
-1.58388782e+00 7.03372598e-01 -6.72504902e-01 -8.37617099e-01
3.18418592e-01 -3.53049397e-01 -6.77871287e-01 -1.27894625e-01
-2.29821950e-01 6.43796742e-01 5.21895230e-01 -8.36497903e-01
-4.52357471e-01 -5.60947396e-02 -3.45610231e-01 4.16108429e-01
-4.90161777e-01 3.09394002e-01 -5.13666034e-01 -7.53324330e-01
-2.26083875e-01 -7.53405035e-01 -8.61718431e-02 -8.15783620e-01
2.31556706e-02 -4.98401642e-01 3.06915611e-01 -1.11476982e+00
1.72425413e+00 -2.05352187e+00 -3.88520896e-01 -3.58662643e-02
-2.99714595e-01 7.12650657e-01 -2.14638710e-01 5.15816092e-01
2.75236368e-01 3.42215151e-01 -5.25751635e-02 -5.79347670e-01
9.54127908e-02 3.67038399e-01 -4.30143237e-01 -4.61919084e-02
1.71651348e-01 8.47171843e-01 -6.05534256e-01 -2.65834361e-01
4.02805567e-01 4.88779187e-01 -5.44212341e-01 2.43860781e-01
-4.86710757e-01 -3.22099142e-02 1.34897036e-02 3.14513236e-01
3.24114621e-01 1.17797270e-01 3.07512820e-01 4.04955119e-01
6.14362061e-02 1.33578968e+00 -8.00841391e-01 1.40256929e+00
-9.63560998e-01 7.53466785e-01 -1.14462420e-01 -5.29962361e-01
1.11991954e+00 4.50901359e-01 -3.55396718e-01 -6.91746116e-01
-7.87176415e-02 5.08675396e-01 1.94982216e-01 -2.11397082e-01
8.02303791e-01 -3.04373413e-01 3.55371803e-01 4.94415522e-01
5.17345034e-02 -2.56197721e-01 5.46716712e-02 1.17280141e-01
1.07059169e+00 -1.33705199e-01 3.03377628e-01 -1.92368969e-01
3.55815440e-01 1.75912946e-01 -1.35361115e-02 9.42636490e-01
9.98164266e-02 5.60369551e-01 1.19922392e-01 1.43769413e-01
-1.24513125e+00 -8.48498523e-01 -2.67289639e-01 1.21383405e+00
-7.23555505e-01 -5.10520339e-01 -9.65813756e-01 -5.55740416e-01
-2.33484898e-02 1.80642688e+00 2.33591616e-01 -9.67079550e-02
-6.90477490e-01 -4.87925410e-01 9.84430254e-01 4.98499244e-01
-1.15662418e-01 -1.01290905e+00 1.63520604e-01 4.72383112e-01
-2.93302149e-01 -1.54121399e+00 -6.16595805e-01 2.83190366e-02
-9.54383254e-01 -2.03819901e-01 -9.35269296e-01 -7.88409054e-01
4.64454412e-01 2.52626121e-01 1.09231317e+00 -2.66303450e-01
1.96522966e-01 5.08340776e-01 -4.86223876e-01 -4.47498828e-01
-1.30741596e+00 2.79216856e-01 2.49371156e-01 -3.94513130e-01
9.63952124e-01 -1.38932198e-01 -2.39590718e-03 5.24275303e-01
-8.10469449e-01 -1.06647827e-01 6.27006710e-01 7.35565186e-01
2.86952287e-01 -3.98015440e-01 7.92503834e-01 -7.80776322e-01
1.12513554e+00 -2.35373914e-01 -3.19926262e-01 3.76009434e-01
-6.61509454e-01 2.43186951e-01 5.99018633e-01 -9.96785700e-01
-1.19459450e+00 -5.09805202e-01 -4.89560843e-01 -2.32715383e-01
-3.57898206e-01 4.16997373e-01 -1.77292019e-01 1.83744267e-01
8.54076743e-01 4.99571025e-01 1.60778359e-01 -5.42014599e-01
2.68639117e-01 1.45769167e+00 6.60946816e-02 -7.30143189e-01
2.23264500e-01 -2.15397298e-01 -7.84902155e-01 -1.47022557e+00
-4.57923472e-01 -5.61105967e-01 -2.54391551e-01 1.79489300e-01
4.79414731e-01 -1.18967569e+00 -8.38761777e-02 1.45512417e-01
-1.02282155e+00 -8.99515808e-01 -3.25967491e-01 9.59904492e-01
-4.05736417e-01 4.88533348e-01 -5.13502479e-01 -1.30850959e+00
-2.43143171e-01 -1.25311470e+00 1.24333668e+00 -2.22491935e-01
-7.42648244e-01 -9.29051340e-01 2.82912981e-02 7.05236256e-01
4.39438164e-01 -9.74905372e-01 9.32303488e-01 -1.17250264e+00
-2.26831555e-01 -2.92042911e-01 -6.65581748e-02 7.60388732e-01
-1.38670906e-01 -3.79654050e-01 -1.10018575e+00 -3.95254791e-01
-1.55760169e-01 -4.89240140e-01 4.29172009e-01 5.47588110e-01
7.40509450e-01 -3.70532066e-01 -2.66559273e-01 1.23074181e-01
9.24283981e-01 3.85966688e-01 4.83358175e-01 5.14181852e-02
5.51760137e-01 5.19525766e-01 5.50025880e-01 2.28691310e-01
2.15971783e-01 7.38485098e-01 -5.36372244e-01 4.44795817e-01
-4.52204585e-01 -7.70680785e-01 8.10402334e-01 1.18148518e+00
4.91608381e-01 -5.53812444e-01 -1.16770387e+00 7.79457748e-01
-1.19557440e+00 -4.29527074e-01 7.86604136e-02 2.58015800e+00
1.02349889e+00 3.49609941e-01 1.92556381e-01 -2.00592995e-01
6.71104789e-01 -4.83564511e-02 -2.35136569e-01 -7.45378554e-01
1.52558014e-02 3.08653921e-01 6.73749149e-01 8.44472528e-01
-3.70441735e-01 1.16393149e+00 6.78082323e+00 1.38369846e+00
-9.89782572e-01 1.99443609e-01 6.86435223e-01 -1.14417046e-01
-5.63587010e-01 -8.38831663e-02 -1.47927690e+00 4.08423752e-01
1.93031263e+00 -2.13934094e-01 5.70712984e-01 8.40379119e-01
5.22433519e-01 -6.18279539e-02 -1.08864868e+00 8.75665724e-01
1.73159674e-01 -9.80911016e-01 3.55891436e-01 1.67643383e-01
4.51801956e-01 3.24485987e-01 -1.34987697e-01 7.68748999e-01
3.34458917e-01 -1.23270583e+00 6.13083899e-01 -8.34281892e-02
1.17215383e+00 -5.81796706e-01 4.96861249e-01 6.93695724e-01
-7.54864991e-01 9.80060622e-02 -5.85629940e-01 -7.02192113e-02
3.65203321e-01 2.59928077e-01 -1.79466414e+00 -5.46352826e-02
-1.37671297e-02 -1.28098831e-01 -4.37640369e-01 5.92687368e-01
-1.20562948e-01 1.53048742e+00 -6.25045180e-01 -4.44899172e-01
3.05863351e-01 -8.23491067e-02 5.76092899e-01 1.72540414e+00
3.05947393e-01 -1.49592727e-01 -3.53888795e-02 5.68789661e-01
-3.14156152e-02 4.17905301e-01 -5.03862798e-01 -6.10378563e-01
9.13610399e-01 6.11671865e-01 -3.34647059e-01 -5.41949391e-01
-4.94752377e-01 8.32670808e-01 3.93597066e-01 4.19576347e-01
-4.71503466e-01 -1.60454899e-01 5.33434927e-01 4.97660756e-01
1.76310256e-01 -6.34877026e-01 -4.08881962e-01 -9.34132874e-01
1.95723176e-01 -1.33330917e+00 -4.61217873e-02 -8.11436832e-01
-8.39295745e-01 6.69330060e-01 2.38666639e-01 -8.07048857e-01
-5.75783789e-01 -6.23748422e-01 -1.81684613e-01 1.32748926e+00
-1.24081326e+00 -8.73402774e-01 2.89525777e-01 6.33670390e-03
1.34558141e+00 -4.58209664e-01 9.34807718e-01 2.97851235e-01
-4.32650894e-01 1.03513503e+00 2.80894697e-01 -6.53872266e-03
7.04982340e-01 -1.01920986e+00 9.97238934e-01 7.78489828e-01
3.47535521e-01 9.66853678e-01 6.42715096e-01 -8.46220374e-01
-1.13029957e+00 -9.31555390e-01 1.29180801e+00 -7.05257595e-01
6.97405517e-01 -7.14240193e-01 -1.14690673e+00 8.31872523e-01
9.51456930e-03 -5.73303044e-01 6.48626387e-01 2.38379896e-01
-2.86146939e-01 2.07832828e-01 -1.09207845e+00 8.75920475e-01
8.91298711e-01 -8.60374451e-01 -6.80161297e-01 3.16667378e-01
1.03485990e+00 -2.41719887e-01 -8.77961636e-01 1.02961488e-01
5.16892374e-01 -3.50604236e-01 7.48205602e-01 -5.47025025e-01
1.05947759e-02 1.59891874e-01 -3.79580200e-01 -1.39776278e+00
-3.08984164e-02 -8.42060447e-01 -5.20288236e-02 1.36877966e+00
8.04440796e-01 -7.23074138e-01 6.73435807e-01 8.97988379e-01
-3.41472596e-01 -4.93168086e-01 -7.96372235e-01 -9.47878897e-01
2.49087334e-01 -8.51935089e-01 3.75271261e-01 4.64954615e-01
1.55283839e-01 5.14753580e-01 -4.82218573e-03 2.40989067e-02
2.03092858e-01 -9.33602214e-01 8.00854146e-01 -7.16192484e-01
-6.00814164e-01 -2.65905946e-01 -1.72259703e-01 -1.80289876e+00
3.32296938e-01 -7.84340262e-01 1.02463104e-01 -1.32468009e+00
-1.16977349e-01 -5.44894993e-01 3.63425046e-01 2.47913316e-01
-1.90880492e-01 -2.78375179e-01 6.28909320e-02 9.53219011e-02
-1.36947736e-01 4.24105674e-01 6.86470091e-01 4.04414833e-02
-3.59565556e-01 1.18261501e-01 -5.60391247e-01 4.25767124e-01
7.92997956e-01 -3.99935097e-01 -6.89010680e-01 -5.04282534e-01
-3.90456825e-01 1.76150486e-01 -3.16134661e-01 -9.46485758e-01
3.14790234e-02 -1.44489240e-02 -4.73686866e-02 -4.20291156e-01
3.57404143e-01 -4.11145508e-01 -1.40697643e-01 2.20586509e-02
-6.56950772e-01 -2.69303210e-02 6.89433992e-01 3.16737890e-01
-6.35308698e-02 -6.31119788e-01 6.22000515e-01 -4.56651822e-02
-4.48104829e-01 -2.99724907e-01 -8.13665032e-01 4.10240918e-01
5.13836741e-01 -4.49158967e-01 -1.21535808e-01 -6.32715642e-01
-3.27302486e-01 -1.57219514e-01 2.85848945e-01 4.55730528e-01
4.71744359e-01 -8.25775325e-01 -7.44072974e-01 4.94098574e-01
1.87278956e-01 -1.57458246e-01 -9.24082547e-02 5.66524148e-01
-4.15253729e-01 8.84692550e-01 5.24246395e-01 -4.38807786e-01
-1.39821172e+00 8.11894760e-02 1.29580677e-01 -2.01933548e-01
-4.87799674e-01 8.39910269e-01 2.72332821e-02 -6.22905552e-01
5.46708941e-01 -4.70346957e-01 1.88963220e-01 -4.68193561e-01
7.33724117e-01 4.64381933e-01 3.41991812e-01 -4.76183921e-01
-3.95259298e-02 -2.10632816e-01 -4.66682553e-01 -7.86053061e-01
8.22538137e-01 -2.81818569e-01 4.35406238e-01 7.06969380e-01
1.01828468e+00 1.36151537e-01 -1.07490230e+00 -3.31513852e-01
1.52552843e-01 -1.21188246e-01 2.02535942e-01 -8.23982596e-01
4.03080210e-02 8.96025300e-01 9.77797285e-02 1.32745519e-01
6.05996609e-01 -5.43518737e-02 9.23640311e-01 5.50531089e-01
3.82597476e-01 -1.14869654e+00 -3.65548164e-01 8.22284222e-01
6.12777531e-01 -1.04635799e+00 -3.02035481e-01 -1.75045654e-01
-9.33795571e-01 8.74978185e-01 3.90076101e-01 3.01997215e-01
3.34323823e-01 3.17554772e-01 1.21069640e-01 3.16697419e-01
-8.49011898e-01 -1.00265034e-01 3.26717675e-01 7.98817158e-01
8.45745444e-01 3.73146057e-01 -3.54800433e-01 4.11525220e-01
-4.79367018e-01 -5.47506511e-02 5.15324652e-01 6.41241491e-01
-7.08199382e-01 -1.22448945e+00 -4.45142448e-01 6.78264737e-01
-4.56819654e-01 -6.64407730e-01 -2.68899679e-01 7.14574873e-01
-5.89403093e-01 1.26282668e+00 2.34391630e-01 -2.23580733e-01
1.00877807e-01 8.04482341e-01 3.65611255e-01 -1.12124145e+00
-3.49743098e-01 3.75592202e-01 8.09591353e-01 -5.23462705e-02
1.53926387e-01 -8.00053656e-01 -9.63652492e-01 -1.14144899e-01
-6.13061965e-01 3.05444717e-01 8.88136506e-01 1.02281117e+00
3.90948176e-01 3.09588939e-01 1.62442982e-01 -3.88543755e-01
-1.04073370e+00 -1.62260425e+00 -4.25165057e-01 4.91334051e-02
4.21017110e-02 -2.75417715e-01 -4.25519705e-01 -1.91458676e-03] | [14.37478256225586, 6.877290725708008] |
c6e4bc2a-4e1b-4080-80dd-be7175dd4791 | the-naughtyformer-a-transformer-understands | 2211.14369 | null | https://arxiv.org/abs/2211.14369v1 | https://arxiv.org/pdf/2211.14369v1.pdf | The Naughtyformer: A Transformer Understands Offensive Humor | Jokes are intentionally written to be funny, but not all jokes are created the same. Some jokes may be fit for a classroom of kindergarteners, but others are best reserved for a more mature audience. While recent work has shown impressive results on humor detection in text, here we instead investigate the more nuanced task of detecting humor subtypes, especially of the less innocent variety. To that end, we introduce a novel jokes dataset filtered from Reddit and solve the subtype classification task using a finetuned Transformer dubbed the Naughtyformer. Moreover, we show that our model is significantly better at detecting offensiveness in jokes compared to state-of-the-art methods. | ['Jason Wang', 'Steve Li', 'Alexander Cai', 'Leonard Tang'] | 2022-11-25 | null | null | null | null | ['humor-detection'] | ['natural-language-processing'] | [-3.73732150e-01 2.71779180e-01 -3.00652713e-01 1.15981713e-01
-1.61863565e-01 -5.28444827e-01 5.32511771e-01 1.13822654e-01
5.47272861e-02 5.22520244e-01 6.56649649e-01 -9.77792516e-02
6.08785544e-03 -5.74097633e-01 -2.62472630e-01 -4.74596471e-01
7.66304314e-01 1.22591473e-01 2.88694769e-01 -6.32286251e-01
5.74273050e-01 4.75796402e-01 -1.62929368e+00 1.40298545e-01
1.05216968e+00 2.72782683e-01 -2.93954968e-01 7.06354976e-01
-1.14008315e-01 1.94912171e+00 -9.47991550e-01 -9.28817749e-01
-3.72295707e-01 -7.39799023e-01 -1.01570153e+00 -3.14542390e-02
1.19094253e+00 -2.45793402e-01 -5.15381098e-01 1.02454162e+00
5.42322397e-01 3.94764900e-01 5.66639364e-01 -1.25948060e+00
-7.95773208e-01 1.15494251e+00 -2.83469945e-01 2.73906350e-01
6.12119555e-01 2.17051029e-01 1.08479106e+00 -7.83759177e-01
4.52223271e-01 1.12146294e+00 6.40658379e-01 6.93388343e-01
-1.09304631e+00 -8.45874727e-01 -4.84944403e-01 5.84559381e-01
-8.50951433e-01 -3.50060105e-01 1.29428828e+00 -7.09645212e-01
6.30425096e-01 5.84088922e-01 5.28904557e-01 1.43637753e+00
-1.83014110e-01 1.11308110e+00 1.16205013e+00 -4.07184929e-01
-1.02547839e-01 1.30230561e-01 6.31428599e-01 7.68166423e-01
9.15761590e-02 -5.90438664e-01 -8.54106843e-01 -1.74774691e-01
1.25751114e-02 -2.92257577e-01 -4.56721812e-01 -6.90371171e-02
-8.78858447e-01 1.10226178e+00 1.15140356e-01 5.15497923e-01
1.22239709e-01 5.82866482e-02 6.42143130e-01 4.98050749e-01
2.20022231e-01 9.52681363e-01 1.90163821e-01 -6.67196631e-01
-1.10554862e+00 5.60909331e-01 1.25229466e+00 7.96815693e-01
1.78562641e-01 1.92530110e-01 -2.79085606e-01 1.18553197e+00
-3.47856045e-01 2.30573684e-01 5.49763381e-01 -8.47725868e-01
1.31545231e-01 6.69199407e-01 -2.61929065e-01 -1.17843592e+00
-5.23607314e-01 -5.23589551e-01 -4.89761949e-01 1.40954077e-01
7.14125276e-01 4.14487302e-01 -2.70202935e-01 1.37130141e+00
3.33139598e-01 -9.87582654e-02 -4.61814016e-01 8.33418310e-01
1.56729507e+00 2.64233083e-01 -1.12486772e-01 -3.15976329e-02
1.19365442e+00 -1.45356107e+00 -8.49156857e-01 -1.63319483e-02
7.39195704e-01 -1.30188572e+00 1.55871892e+00 8.36325943e-01
-1.24507856e+00 -6.48401156e-02 -1.01316631e+00 -6.60165608e-01
-3.53496760e-01 -1.22325972e-01 2.94671476e-01 1.09273112e+00
-3.40066016e-01 7.01371074e-01 -1.83530655e-02 -1.62372738e-01
3.85447323e-01 -5.65520346e-01 -1.48568898e-01 2.47684628e-01
-1.11570048e+00 1.33793294e+00 -2.48957515e-01 -6.11096978e-01
-9.23235774e-01 -1.03433728e+00 -6.24710917e-01 1.54591069e-01
4.35300499e-01 -3.87403399e-01 1.46683896e+00 -3.57097775e-01
-1.49561453e+00 1.38546515e+00 2.87574500e-01 -1.82125062e-01
7.18705535e-01 -3.55774522e-01 -2.26907656e-01 1.74594313e-01
1.45101592e-01 -2.70746529e-01 1.02062047e+00 -8.88012111e-01
1.59942344e-01 2.78087389e-02 7.10508600e-02 -2.05451399e-01
-7.53387630e-01 4.73630637e-01 2.89339811e-01 -8.56564224e-01
-2.75286555e-01 -6.53948843e-01 5.23455739e-01 -3.92692864e-01
-8.29563558e-01 -5.56030571e-01 8.50653946e-01 -6.02861106e-01
2.01555490e+00 -1.86664760e+00 -1.38665751e-01 -1.95265993e-01
7.22463608e-01 3.16254467e-01 2.40854993e-01 7.19541550e-01
-1.13513492e-01 1.52508825e-01 3.30672771e-01 -3.06942850e-01
4.03438270e-01 -3.73902433e-02 -5.68021357e-01 4.32374835e-01
-5.10301173e-01 8.18899930e-01 -1.07902896e+00 -4.78208512e-01
1.18674226e-01 -8.86042193e-02 -3.40739518e-01 3.33460122e-01
4.90233228e-02 2.01435201e-02 -3.67636412e-01 5.42887449e-01
3.29645753e-01 -2.94221044e-01 -3.37858588e-01 3.12388599e-01
-2.73416281e-01 1.04995251e+00 -5.58045983e-01 1.02330017e+00
-3.95643920e-01 1.09132087e+00 -3.05650961e-02 -4.34487194e-01
9.99867260e-01 1.04287170e-01 6.40971884e-02 -3.42667788e-01
2.58417964e-01 1.09233171e-01 -3.80664580e-02 -8.19553852e-01
9.74491775e-01 -4.92843181e-01 -2.45920002e-01 7.62873948e-01
-7.21567450e-03 -6.96313143e-01 2.95430243e-01 6.54751003e-01
1.34646022e+00 -4.31771189e-01 5.37745953e-01 -3.04586440e-01
5.42596400e-01 1.33018112e-02 3.17269787e-02 1.03766131e+00
-4.57360774e-01 5.38301885e-01 8.99668157e-01 -3.26727360e-01
-9.85861182e-01 -9.53074217e-01 -2.73277402e-01 1.66876245e+00
-3.03952903e-01 -6.52944148e-01 -7.85835683e-01 -8.29271317e-01
6.57453388e-02 1.23809326e+00 -5.23690879e-01 -2.33826801e-01
-3.97887409e-01 -2.33219475e-01 1.17063725e+00 -7.26954937e-02
4.29366142e-01 -1.05821371e+00 -7.96756804e-01 -4.90867766e-03
-2.07392469e-01 -1.10206318e+00 -5.95270753e-01 2.95050945e-02
-2.28964850e-01 -1.15311229e+00 -3.53098154e-01 -6.50652945e-01
5.58452047e-02 5.10101914e-01 1.54100955e+00 5.85474193e-01
-2.60085493e-01 3.88372689e-01 -6.25229776e-01 -5.17618120e-01
-8.81210923e-01 7.03746751e-02 -1.62818626e-01 -5.56859612e-01
6.83475137e-01 -7.79043078e-01 -1.25274703e-01 1.15796112e-01
-7.44759202e-01 1.03340536e-01 -2.48204648e-01 8.96036148e-01
-4.80426431e-01 -1.75313056e-01 5.81111372e-01 -1.19140899e+00
8.84456575e-01 -4.14553195e-01 1.72836393e-01 -1.01570562e-01
-5.52748501e-01 -2.76700199e-01 1.07018602e+00 -7.46367812e-01
-7.66735256e-01 -5.13944983e-01 -2.27605492e-01 -4.53632802e-01
-1.60587549e-01 2.92054806e-02 1.14935890e-01 -2.93045342e-01
1.07261980e+00 5.86699024e-02 -2.10182726e-01 -6.27645493e-01
3.35152119e-01 6.90406978e-01 7.46326983e-01 -6.22250140e-01
9.15713072e-01 1.27082318e-01 3.95752713e-02 -1.01553905e+00
-1.42013204e+00 -6.67957127e-01 -3.99643213e-01 -6.18175864e-01
5.04031360e-01 -6.56137466e-01 -1.07891738e+00 6.40952766e-01
-1.07318950e+00 -3.89107287e-01 -8.73320028e-02 8.42002109e-02
-5.52718401e-01 6.40835047e-01 -9.55451250e-01 -6.88689351e-01
-3.88646305e-01 -6.38487577e-01 4.19837505e-01 3.02192122e-01
-8.01724732e-01 -9.36835110e-01 3.76034230e-01 1.05590308e+00
1.82853654e-01 2.80802965e-01 1.09242165e+00 -1.14701009e+00
-6.83900341e-02 -1.06279396e-01 1.25876993e-01 2.99996883e-01
-3.96144301e-01 2.47689202e-01 -1.20085704e+00 3.62171710e-01
8.22962224e-02 -9.54431772e-01 9.38568950e-01 -3.88317883e-01
9.11049426e-01 -8.51540267e-01 2.87673950e-01 6.83445260e-02
8.40055287e-01 -4.02349353e-01 5.65415740e-01 5.67508042e-01
7.86878288e-01 5.16210735e-01 2.94766784e-01 7.51396060e-01
5.16236961e-01 4.98151332e-01 4.40287977e-01 3.66869241e-01
-5.01822412e-01 -6.81278050e-01 4.49727416e-01 1.12177956e+00
3.48166168e-01 4.32072394e-02 -9.88937199e-01 8.27494919e-01
-1.56159711e+00 -1.59612286e+00 -8.11316252e-01 1.67934036e+00
1.23426044e+00 1.66625589e-01 5.41357636e-01 4.02188778e-01
4.51961219e-01 5.19881606e-01 1.59213431e-02 -1.05775762e+00
-2.63694152e-02 2.69769371e-01 -1.59530774e-01 4.27568048e-01
-7.90462971e-01 1.01061010e+00 6.67804193e+00 1.26163089e+00
-9.52607512e-01 1.29970446e-01 2.71724343e-01 -3.68266523e-01
-4.02632803e-01 -2.39076436e-01 -5.95257759e-01 6.58746362e-01
5.24909556e-01 -1.46845430e-01 5.09336352e-01 1.19085777e+00
2.95295537e-01 -2.49690294e-01 -9.83686805e-01 9.91165340e-01
5.98529100e-01 -1.15126693e+00 -3.32252413e-01 -5.24245620e-01
8.63264084e-01 -3.78265589e-01 -6.19975924e-02 5.23386896e-01
4.48154718e-01 -1.28921688e+00 9.01764452e-01 3.75828207e-01
-5.54266423e-02 -5.89871943e-01 2.68257171e-01 6.90200210e-01
-1.53072655e-01 -2.47558877e-02 -2.91333377e-01 -6.41481698e-01
-4.02735591e-01 8.63849878e-01 -8.21700096e-01 -2.73929447e-01
6.40151441e-01 6.23813987e-01 -1.10866153e+00 9.83564436e-01
-6.89107478e-01 9.94478345e-01 -4.84185927e-02 -6.88417315e-01
1.52692199e-01 1.00648262e-01 9.80997741e-01 1.45687890e+00
8.51339251e-02 1.99053451e-01 -5.82877621e-02 1.14669764e+00
-3.38103622e-01 1.89701319e-01 -7.31914878e-01 -1.84241965e-01
5.38056850e-01 1.73766148e+00 -2.71392614e-01 -2.83091366e-01
-1.44925490e-01 6.84402347e-01 5.68493545e-01 -1.50896668e-01
-7.62328148e-01 -4.07849520e-01 6.51227593e-01 1.20513268e-01
-3.02328099e-03 8.59108195e-02 -8.46383035e-01 -1.23877215e+00
-3.37699592e-01 -1.00349903e+00 4.20805246e-01 -1.05626667e+00
-1.64870524e+00 -1.10533282e-01 -2.74952382e-01 -9.42686558e-01
-1.49244875e-01 -5.29406250e-01 -1.27272975e+00 2.83546537e-01
-1.00181377e+00 -1.08184099e+00 -3.64307970e-01 3.99378985e-01
4.00382161e-01 4.59431186e-02 4.11144167e-01 5.65666407e-02
-5.68869174e-01 7.72180974e-01 -9.30764750e-02 2.61664212e-01
1.00264442e+00 -1.31066322e+00 -2.55681008e-01 6.22985661e-01
1.45378768e-01 6.61878347e-01 1.48409462e+00 -2.90384561e-01
-1.02368557e+00 -3.75847369e-01 1.42228961e+00 -9.73620713e-01
1.42430425e+00 -1.85266376e-01 -1.19580996e+00 2.97421604e-01
4.65292424e-01 -6.31396055e-01 7.75599778e-01 3.80239427e-01
-1.16067851e+00 4.77173567e-01 -1.02201104e+00 5.79671860e-01
8.60238314e-01 -6.29985690e-01 -1.14521408e+00 3.69224191e-01
3.33394438e-01 -3.92630398e-01 -7.61524439e-01 -6.58193454e-02
4.37253624e-01 -1.41325521e+00 7.11070180e-01 -8.37243319e-01
1.49517536e+00 6.41124994e-02 3.54791105e-01 -1.38762367e+00
-4.26483721e-01 -7.46572852e-01 -3.11855644e-01 1.54797518e+00
-1.49665296e-01 -2.75023803e-02 7.44618654e-01 6.21094525e-01
-3.82154375e-01 -5.83424509e-01 -8.55531633e-01 -8.31916988e-01
6.43972874e-01 -3.11948061e-01 3.96027416e-01 1.68352294e+00
9.06612456e-01 8.32727015e-01 -7.48475850e-01 -5.62914908e-01
5.74628592e-01 4.89257604e-01 1.04688799e+00 -8.93607795e-01
-2.50291586e-01 -1.18119037e+00 -3.21048468e-01 -7.10407913e-01
5.06357729e-01 -9.81391609e-01 -1.49749324e-01 -1.00673687e+00
7.45130360e-01 1.50169194e-01 3.31595808e-01 6.26443386e-01
-3.14112276e-01 4.96734619e-01 5.04116237e-01 -1.10062055e-01
-7.31100440e-01 3.69302511e-01 1.60179222e+00 -1.76836267e-01
-6.70191348e-02 -1.55204445e-01 -7.13022649e-01 8.46871436e-01
7.08128691e-01 -4.34864193e-01 -1.50109977e-01 2.44430333e-01
6.72658145e-01 -2.15033233e-01 6.63385510e-01 -7.59194255e-01
2.49328405e-01 -5.10079145e-01 -1.43671751e-01 -4.87039953e-01
5.41798323e-02 -3.43335807e-01 -2.65700758e-01 1.33033708e-01
-6.01872981e-01 -5.66461027e-01 -1.68433934e-01 -1.22634239e-01
-3.30056101e-02 -9.45847571e-01 1.16158819e+00 3.15187452e-03
-1.75536275e-01 -4.32938188e-01 -6.78679824e-01 4.77460891e-01
6.79348350e-01 -8.41150209e-02 -1.25411904e+00 -8.92447233e-01
-4.40406024e-01 -4.69890200e-02 6.88938916e-01 4.05386090e-01
3.18465739e-01 -1.21383512e+00 -5.30493140e-01 -2.68610954e-01
2.91129798e-01 -6.52933300e-01 2.02802852e-01 1.14278162e+00
-4.16905940e-01 7.69058615e-02 -1.76336467e-01 5.11374846e-02
-1.70749307e+00 7.28789866e-01 4.87268716e-02 -2.12118879e-01
-6.82284594e-01 7.78745532e-01 -5.58294691e-02 -5.44217229e-01
1.70464385e-02 1.21029779e-01 -4.46814060e-01 3.62792373e-01
6.83802545e-01 1.10542381e+00 8.03725235e-03 -7.86624134e-01
-7.23171979e-02 -3.14027518e-02 6.99725747e-02 5.51830113e-01
1.15020871e+00 -4.73843589e-02 -4.75781858e-01 7.43662477e-01
9.96302903e-01 8.22238922e-01 -4.12387073e-01 2.11639013e-02
-4.27522808e-02 -7.19459116e-01 -1.72419399e-01 -9.54323947e-01
-4.52645272e-01 9.76669133e-01 -5.34757555e-01 8.13258648e-01
8.65956306e-01 2.12290019e-01 9.41574275e-01 3.40820789e-01
-8.16333890e-02 -1.31519508e+00 7.48055816e-01 1.07305872e+00
8.58466506e-01 -9.25077319e-01 3.63582551e-01 -4.62414205e-01
-7.53316462e-01 1.36806786e+00 6.98725462e-01 -4.59945574e-02
2.11193621e-01 -1.66988112e-02 -2.76064388e-02 -4.90604132e-01
-9.66231883e-01 -7.27339834e-02 4.95669246e-01 8.12954307e-02
8.18988979e-01 1.49410471e-01 -6.81850016e-01 7.73931503e-01
-9.91625786e-01 -3.56100112e-01 1.39810872e+00 2.87476361e-01
-8.88817012e-01 -4.60780084e-01 -7.18268931e-01 4.60821241e-01
-6.63188875e-01 -1.55771866e-01 -1.24374127e+00 6.56148255e-01
-2.05090255e-01 1.22491622e+00 -5.39999902e-01 -5.38228154e-01
1.29179418e-01 4.89475191e-01 5.96464813e-01 -5.28499067e-01
-1.29340780e+00 -3.91108155e-01 6.26413405e-01 -3.06704670e-01
-1.26487643e-01 -4.04857725e-01 -8.95369649e-01 -1.11869073e+00
-1.52019486e-01 -1.17194490e-03 1.57898992e-01 9.19197381e-01
-3.21975619e-01 1.01718932e-01 6.43371761e-01 -3.12059641e-01
-7.09817171e-01 -9.34118629e-01 -6.88126445e-01 9.88530099e-01
3.99983943e-01 -6.23171508e-01 -6.35470092e-01 -2.03588918e-01] | [8.886015892028809, 11.054946899414062] |
3236121b-3fad-4f64-9fe5-2b30008a8897 | exploring-optimal-voting-in-native-language | null | null | https://aclanthology.org/W17-5041 | https://aclanthology.org/W17-5041.pdf | Exploring Optimal Voting in Native Language Identification | We describe the submissions entered by the National Research Council Canada in the NLI-2017 evaluation. We mainly explored the use of voting, and various ways to optimize the choice and number of voting systems. We also explored the use of features that rely on no linguistic preprocessing. Long ngrams of characters obtained from raw text turned out to yield the best performance on all textual input (written essays and speech transcripts). Voting ensembles turned out to produce small performance gains, with little difference between the various optimization strategies we tried. Our top systems achieved accuracies of 87{\%} on the essay track, 84{\%} on the speech track, and close to 92{\%} by combining essays, speech and i-vectors in the fusion track. | ["Serge L{\\'e}ger", 'Cyril Goutte'] | 2017-09-01 | null | null | null | ws-2017-9 | ['native-language-identification'] | ['natural-language-processing'] | [ 1.30249947e-01 1.62669569e-01 -3.34652513e-02 -5.46288848e-01
-1.30725169e+00 -9.22190070e-01 1.15012586e+00 2.31370673e-01
-8.50099504e-01 7.45917618e-01 4.35158074e-01 -5.16189158e-01
-4.18541431e-02 -4.32003915e-01 -1.20434023e-01 -5.43511033e-01
5.47250152e-01 5.95803618e-01 -3.42618264e-02 -3.04746240e-01
4.80475157e-01 7.72838518e-02 -1.32324469e+00 4.29089546e-01
6.76941454e-01 7.57827461e-01 -1.35227785e-01 1.02815175e+00
-2.85880446e-01 6.65919125e-01 -9.51882720e-01 -7.54430652e-01
2.29319647e-01 -2.73778915e-01 -6.83634758e-01 -2.56075710e-01
7.39181101e-01 -9.89500135e-02 -4.06097859e-01 8.05812418e-01
6.10343397e-01 1.83199689e-01 9.24116790e-01 -8.20800960e-01
-7.21874774e-01 1.12210548e+00 -3.26431423e-01 3.24323416e-01
5.97759962e-01 7.29987174e-02 1.46235681e+00 -1.19194210e+00
6.97246253e-01 9.91505325e-01 6.46612644e-01 4.31894630e-01
-1.16115808e+00 -7.31267869e-01 -1.20054476e-01 3.49615440e-02
-1.07121670e+00 -8.69506419e-01 2.17361555e-01 -5.71328938e-01
1.58763015e+00 3.39473546e-01 4.41101462e-01 1.04447436e+00
1.56577557e-01 8.52416813e-01 1.37547088e+00 -8.83830249e-01
-1.15679406e-01 2.52957702e-01 8.66405904e-01 6.40854001e-01
1.92439213e-01 -5.51902573e-04 -6.95187986e-01 -5.33869445e-01
1.02137670e-01 -6.09817386e-01 -2.39662945e-01 6.90116405e-01
-1.18238997e+00 1.06345153e+00 -3.89665455e-01 4.06772166e-01
-1.71999559e-01 -6.59079328e-02 3.88803422e-01 4.74357069e-01
4.67739284e-01 7.32095122e-01 -5.98576903e-01 -4.27612990e-01
-1.34630346e+00 4.10812676e-01 1.21835494e+00 7.90574193e-01
4.79008436e-01 -6.38520271e-02 -4.70996350e-01 1.06940591e+00
1.69438258e-01 6.17875755e-01 4.11136836e-01 -9.62795019e-01
9.20333385e-01 3.99708748e-01 1.23693988e-01 -7.45175600e-01
-6.06038094e-01 -1.48996562e-01 -4.18800801e-01 -2.09096581e-01
6.87521458e-01 -5.55368364e-01 -1.14873874e+00 1.46137881e+00
-2.00331956e-01 -6.42176509e-01 -2.44201664e-02 4.97007370e-01
1.00631130e+00 8.27829838e-01 7.44763911e-02 -4.03655767e-01
1.31713259e+00 -9.76941168e-01 -7.12946832e-01 -1.51402742e-01
8.71014535e-01 -1.32388937e+00 9.38644111e-01 4.64186847e-01
-1.21002471e+00 -4.12620872e-01 -1.07248962e+00 8.17128867e-02
-1.75490290e-01 3.72916102e-01 4.51463580e-01 9.78263795e-01
-1.25885594e+00 5.30149996e-01 -5.16580105e-01 -2.41734087e-01
-3.31692882e-02 4.58716750e-01 -2.49368176e-01 1.00910164e-01
-1.18748891e+00 1.30377471e+00 9.99005735e-02 -1.13231294e-01
-2.02487454e-01 -4.86883759e-01 -6.75082505e-01 -1.04752779e-01
7.97031820e-02 -2.78372407e-01 1.41191328e+00 -6.65224433e-01
-1.65697587e+00 1.00175869e+00 -2.74305761e-01 -3.79631966e-01
3.04672986e-01 -2.84981094e-02 -4.83947754e-01 -1.98453352e-01
-2.86730696e-02 4.70043629e-01 3.05170834e-01 -4.71029669e-01
-8.37603211e-01 -2.07847208e-01 -2.05033243e-01 2.25585505e-01
-4.09388602e-01 5.15561819e-01 -2.30877742e-01 -5.27034581e-01
-7.36044347e-02 -1.24431598e+00 -8.18946958e-02 -9.00799215e-01
-4.56265002e-01 -5.78020155e-01 3.64488065e-01 -1.05907917e+00
1.45566487e+00 -1.63136530e+00 -2.43083499e-02 1.88425988e-01
-8.39364454e-02 5.13100505e-01 -1.73224866e-01 6.51359141e-01
1.76326334e-01 4.11779702e-01 8.80208239e-02 -2.74682611e-01
2.01411054e-01 3.12047713e-02 -3.53862405e-01 2.82807827e-01
2.03599781e-02 8.22182834e-01 -5.47585189e-01 -4.15676683e-01
1.55653894e-01 1.86598957e-01 -3.10464770e-01 -1.77609503e-01
6.24346174e-02 1.70906056e-02 -1.05038188e-01 4.46521789e-01
1.99575126e-01 -1.17678627e-01 1.53875947e-01 2.19306797e-01
-3.22471410e-01 9.56762791e-01 -9.93464768e-01 1.22194362e+00
-3.84884089e-01 9.80653405e-01 1.27321571e-01 -3.74794215e-01
1.01233220e+00 4.95945305e-01 1.30696893e-01 -4.83178943e-01
1.83921978e-01 5.29937565e-01 3.53497595e-01 -2.54340142e-01
7.82271683e-01 -1.49870291e-01 -3.16647172e-01 7.01486766e-01
3.21590304e-01 -4.41558540e-01 4.80217040e-01 2.11678550e-01
1.20165002e+00 -2.06986293e-01 5.06258488e-01 -4.97465879e-01
4.42273974e-01 3.21416944e-01 2.42549926e-01 9.48763609e-01
-1.93801466e-02 7.39636719e-01 2.99871415e-01 -3.84522259e-01
-1.23124242e+00 -4.89942312e-01 -2.28735611e-01 1.31585193e+00
-4.54944283e-01 -6.47366107e-01 -7.24289179e-01 -6.73124194e-01
-2.95772761e-01 1.07265902e+00 -4.39514369e-01 2.66730905e-01
-5.96029103e-01 -9.35057998e-01 1.06045663e+00 4.47167665e-01
6.57382002e-03 -1.07295585e+00 -2.58189291e-01 3.29675406e-01
-3.98723066e-01 -1.02729285e+00 -7.36748755e-01 7.02998161e-01
-6.13292456e-01 -8.10333371e-01 -6.98032737e-01 -4.83682424e-01
2.00327754e-01 -9.94587243e-02 1.04939449e+00 -3.71368718e-03
-1.10406270e-02 1.08110525e-01 -4.04665411e-01 -4.73430514e-01
-6.32766962e-01 5.82148850e-01 3.86308879e-02 -4.88144130e-01
5.05039394e-01 -1.31589442e-01 -6.25330061e-02 1.87266022e-01
-4.71035898e-01 -2.52517611e-01 4.23216462e-01 9.07087505e-01
9.08369720e-02 -4.98056501e-01 4.13252532e-01 -1.18803430e+00
1.05106020e+00 -2.29929149e-01 -2.80656725e-01 4.05721068e-01
-8.50251496e-01 1.29863694e-01 5.18225968e-01 -3.02087784e-01
-9.59848762e-01 -9.84578729e-02 -5.03513873e-01 8.34806561e-02
-9.31219831e-02 5.64275742e-01 1.79855913e-01 7.70213008e-02
8.30833793e-01 -6.18522353e-02 -3.18763047e-01 -3.63470972e-01
2.90614098e-01 1.18823361e+00 2.90960014e-01 -5.94005585e-01
2.70022094e-01 -6.93764165e-02 -4.75663632e-01 -9.42099333e-01
-8.28800201e-01 -5.63294172e-01 -6.60877526e-01 1.16801947e-01
6.57076776e-01 -6.78493261e-01 -3.88826907e-01 4.86314774e-01
-1.24406290e+00 -3.13946426e-01 -1.73308060e-01 5.74632168e-01
-2.76626855e-01 2.88478315e-01 -8.43422294e-01 -9.31675792e-01
-6.90229058e-01 -1.08342397e+00 9.85067546e-01 3.11979443e-01
-9.79170322e-01 -9.01810348e-01 2.96246946e-01 5.77377021e-01
4.16104943e-01 -3.52448821e-01 9.66345549e-01 -1.29228592e+00
1.13608278e-01 -2.71660775e-01 -7.75864273e-02 3.03987294e-01
3.81262004e-02 5.50440490e-01 -1.17425931e+00 -1.27263099e-01
-2.91372567e-01 -3.39525968e-01 1.04282570e+00 2.23299503e-01
5.47867119e-01 -3.99595112e-01 -1.78482667e-01 1.94289401e-01
9.18802261e-01 2.90781498e-01 4.81881320e-01 2.51965761e-01
3.52535784e-01 6.58757687e-01 3.56909096e-01 2.50358790e-01
3.38264167e-01 8.61265540e-01 -1.79924667e-01 4.68004495e-01
-1.15056373e-01 -1.09233186e-02 5.91746032e-01 1.28350687e+00
-1.50181472e-01 -6.01473391e-01 -1.15890217e+00 6.14147842e-01
-1.66226220e+00 -9.08249676e-01 -5.73056281e-01 1.96504378e+00
8.86442423e-01 2.32677817e-01 3.48710939e-02 1.86243862e-01
6.12701535e-01 1.97769687e-01 8.92687440e-02 -8.89001429e-01
-3.55267853e-01 6.79313838e-01 7.18812585e-01 7.65500605e-01
-1.10563207e+00 9.51390862e-01 7.62151814e+00 8.44324887e-01
-9.77722049e-01 1.42319083e-01 3.34176332e-01 -3.03594261e-01
-4.99923229e-01 3.11329290e-02 -1.39508200e+00 5.63236296e-01
1.51484811e+00 -1.59422085e-01 2.22408146e-01 4.09339845e-01
-1.73023090e-01 -2.76442587e-01 -8.84674072e-01 6.44320965e-01
7.15256259e-02 -1.34456158e+00 -2.16612726e-01 6.48621097e-02
9.47986364e-01 4.27232385e-01 -1.64779827e-01 4.81405139e-01
7.23137558e-01 -1.23018122e+00 8.61770153e-01 2.69903541e-01
7.05836177e-01 -6.18973792e-01 8.88567567e-01 6.17449224e-01
-6.95260882e-01 9.16946903e-02 -3.19227427e-01 -2.59499639e-01
7.37373754e-02 6.32670105e-01 -9.60582078e-01 4.98517811e-01
4.65375364e-01 3.68044227e-01 -5.40853024e-01 7.06237495e-01
-4.66542035e-01 1.23997700e+00 -6.65595293e-01 -5.04613221e-01
4.53589261e-01 1.28204352e-03 5.07442772e-01 1.83061397e+00
2.10275203e-01 2.48071119e-01 1.61610082e-01 1.11370891e-01
-1.89267993e-01 3.48002255e-01 -4.60929275e-01 -1.41380072e-01
5.47795236e-01 1.25175548e+00 -6.29586577e-01 -3.77434611e-01
-4.67917472e-01 5.91240585e-01 3.26502323e-01 7.67644048e-02
-5.62810242e-01 -5.84535062e-01 3.95006359e-01 -1.06125973e-01
1.99225277e-01 -2.72559494e-01 -6.06927156e-01 -1.00070977e+00
-1.04061149e-01 -1.11147738e+00 4.70567495e-01 -4.28689122e-01
-1.15539742e+00 8.20338368e-01 -1.14146776e-01 -5.55329442e-01
-5.79020381e-01 -6.91659510e-01 -5.42636573e-01 1.15752280e+00
-9.52902138e-01 -6.68305516e-01 8.28996524e-02 1.55385256e-01
5.43965161e-01 -3.88017058e-01 1.02706015e+00 2.82005936e-01
-3.96310031e-01 8.83104086e-01 1.28560111e-01 1.75186619e-01
9.36507106e-01 -1.16454673e+00 6.31606698e-01 5.58919251e-01
6.92357957e-01 7.63191342e-01 7.48643696e-01 -5.19942462e-01
-1.06162333e+00 -7.16566324e-01 1.52412152e+00 -7.64787734e-01
7.80501664e-01 -1.41376272e-01 -7.31944978e-01 7.37785697e-01
6.53129637e-01 -5.26790202e-01 7.91789174e-01 4.40437943e-01
-3.01613778e-01 4.22255009e-01 -8.52571547e-01 4.97488856e-01
6.67143047e-01 -6.98997378e-01 -7.64186502e-01 1.47655606e-01
4.97785538e-01 -5.52170455e-01 -8.42324197e-01 2.88401544e-01
7.49362409e-01 -5.97306669e-01 4.64304626e-01 -5.63003063e-01
3.82655591e-01 1.27517432e-01 -3.27418685e-01 -1.39984405e+00
-3.70151907e-01 -6.65497899e-01 8.55621397e-02 1.40482640e+00
1.10341859e+00 -6.13310099e-01 6.81952894e-01 6.83179617e-01
-1.59687042e-01 -7.93147027e-01 -8.73745084e-01 -4.99329865e-01
3.44228089e-01 -5.53046286e-01 2.82965451e-01 8.42416644e-01
2.03273803e-01 6.12674236e-01 -1.94132864e-01 -4.15147483e-01
2.80818999e-01 -1.15644686e-01 5.24332643e-01 -1.13813651e+00
-2.44229183e-01 -8.62170756e-01 -1.56543061e-01 -8.21795046e-01
2.19326913e-01 -1.09720659e+00 1.12756379e-01 -1.55133426e+00
3.02829593e-01 -3.52284104e-01 -1.07863806e-01 6.85644031e-01
-3.46925437e-01 4.54291373e-01 3.84725302e-01 4.43225801e-01
-2.03738570e-01 6.99616447e-02 7.75807559e-01 -2.05090463e-01
-1.55349582e-01 1.58312485e-01 -1.02222002e+00 8.06321561e-01
7.31152952e-01 -5.18155515e-01 4.15962860e-02 -5.76279342e-01
3.37196857e-01 9.54169780e-02 -2.09000319e-01 -5.82756937e-01
4.36539769e-01 -6.76812604e-02 2.71623254e-01 -6.43609881e-01
3.70265156e-01 -2.34618559e-01 5.18587306e-02 1.43356025e-01
-6.74424469e-01 6.80494010e-02 2.02190280e-01 3.22441645e-02
-1.93974540e-01 -6.35884106e-01 7.06468701e-01 -8.72503966e-02
-2.34239280e-01 -2.22973093e-01 -6.49137378e-01 2.24179789e-01
6.72968149e-01 -2.53731489e-01 -2.76919901e-01 -4.09703046e-01
-7.59256482e-01 1.91977352e-01 3.20614785e-01 4.68477577e-01
2.29492709e-02 -9.82248902e-01 -1.03011847e+00 -4.53366786e-02
-1.48824632e-01 -5.66388249e-01 -3.55640471e-01 9.42017972e-01
-3.97922844e-01 9.10325170e-01 2.26259023e-01 -3.32840592e-01
-1.51345992e+00 -2.88655370e-01 1.37730464e-01 -7.26693630e-01
-1.47895083e-01 1.10163617e+00 -3.70387763e-01 -8.65548015e-01
2.55089432e-01 -1.73962861e-01 -1.85480371e-01 5.95040619e-01
5.50897181e-01 6.99687600e-01 6.40956521e-01 -6.74656451e-01
-4.47952867e-01 3.58627498e-01 -2.00193375e-01 -6.19869947e-01
1.21181738e+00 -2.56036278e-02 6.00315742e-02 7.02593327e-01
9.79933500e-01 3.59027117e-01 -6.51703835e-01 -9.34923738e-02
3.35912704e-01 -4.87334020e-02 1.02389334e-02 -9.94725347e-01
-6.86172068e-01 9.41502929e-01 -5.18903434e-02 3.46942246e-01
4.46516812e-01 -3.11858177e-01 7.26927042e-01 5.94333827e-01
4.70497370e-01 -1.34285295e+00 -5.93892813e-01 1.05542016e+00
5.87021053e-01 -1.07605374e+00 2.19880298e-01 -1.34096080e-02
-7.83176661e-01 1.39632201e+00 1.68982327e-01 -6.61411583e-02
5.70982754e-01 6.07541025e-01 2.41649061e-01 -6.13086000e-02
-9.98398244e-01 -3.47672179e-02 5.01515269e-01 1.44065246e-01
9.56852734e-01 2.37451375e-01 -6.52061105e-01 7.11720824e-01
-6.34198785e-01 -2.21907943e-01 6.29788399e-01 6.25787437e-01
-5.71091413e-01 -1.26626873e+00 -3.51174384e-01 9.11112010e-01
-8.32046926e-01 -3.65447015e-01 -9.43944871e-01 8.78217638e-01
-1.20272256e-01 1.31025648e+00 -1.72316357e-01 -4.92126465e-01
2.63138533e-01 6.93358481e-01 5.46935737e-01 -9.28815305e-01
-1.30070043e+00 4.39829426e-03 7.78083384e-01 -1.94597080e-01
-1.96981981e-01 -1.12866449e+00 -1.08766198e+00 -4.92231727e-01
-8.04942608e-01 3.75468642e-01 7.63822734e-01 1.20759594e+00
-3.67486328e-02 2.18664229e-01 3.08029801e-01 -7.03880250e-01
-9.75906074e-01 -1.35922945e+00 -4.46269929e-01 -1.26228735e-01
7.22657815e-02 -1.72314838e-01 -4.97367531e-01 -1.12949654e-01] | [10.467229843139648, 10.41737174987793] |
c6f2c95d-dea8-4cac-aaef-42df256a61bc | sinusoidal-flow-a-fast-invertible | 2110.13344 | null | https://arxiv.org/abs/2110.13344v1 | https://arxiv.org/pdf/2110.13344v1.pdf | Sinusoidal Flow: A Fast Invertible Autoregressive Flow | Normalising flows offer a flexible way of modelling continuous probability distributions. We consider expressiveness, fast inversion and exact Jacobian determinant as three desirable properties a normalising flow should possess. However, few flow models have been able to strike a good balance among all these properties. Realising that the integral of a convex sum of sinusoidal functions squared leads to a bijective residual transformation, we propose Sinusoidal Flow, a new type of normalising flows that inherits the expressive power and triangular Jacobian from fully autoregressive flows while guaranteed by Banach fixed-point theorem to remain fast invertible and thereby obviate the need for sequential inversion typically required in fully autoregressive flows. Experiments show that our Sinusoidal Flow is not only able to model complex distributions, but can also be reliably inverted to generate realistic-looking samples even with many layers of transformations stacked. | ['Yumou Wei'] | 2021-10-26 | null | null | null | null | ['normalising-flows'] | ['methodology'] | [-9.22271460e-02 2.93798655e-01 7.57296979e-02 2.81485505e-02
-3.96341026e-01 -8.15932214e-01 1.01991630e+00 -6.52718008e-01
-3.67688164e-02 1.03182101e+00 3.27053100e-01 -5.42350173e-01
-4.58986282e-01 -6.94676340e-01 -6.09175742e-01 -6.48577452e-01
-3.41177464e-01 4.21752781e-01 -1.40873000e-01 -1.31389216e-01
8.03699940e-02 7.24963605e-01 -1.35260355e+00 -4.45003390e-01
1.11582136e+00 8.02793145e-01 -3.41027170e-01 8.95991862e-01
-1.54117793e-01 1.01944768e+00 -2.92057633e-01 -5.92224002e-01
6.01711750e-01 -6.33559704e-01 -5.99101424e-01 -1.36326879e-01
6.13066256e-01 -1.96104199e-01 -2.60982692e-01 8.55091274e-01
2.10813031e-01 1.87570676e-01 1.09078133e+00 -1.47139275e+00
-7.62536228e-01 3.44436646e-01 -2.15176389e-01 2.50490516e-01
3.12378705e-01 3.26951802e-01 1.12178528e+00 -9.25534070e-01
4.57755119e-01 1.23430288e+00 8.79328489e-01 3.65004987e-01
-1.58025169e+00 -3.70022148e-01 -1.95338994e-01 -3.53381097e-01
-1.12645936e+00 -6.35788023e-01 6.20266497e-01 -6.05505645e-01
6.78527832e-01 6.38032079e-01 1.02437949e+00 9.02101815e-01
2.81302661e-01 4.74692255e-01 9.13260281e-01 -1.17623441e-01
1.02790967e-01 2.62390494e-01 -2.75747001e-01 6.60258830e-01
2.01055110e-01 2.92045146e-01 -3.39962721e-01 -1.10085011e-01
1.23194396e+00 -2.54648536e-01 -5.09884238e-01 -6.57585740e-01
-1.43990242e+00 8.43781173e-01 2.83145964e-01 1.95193484e-01
-4.09731299e-01 4.21607912e-01 2.84178168e-01 3.07693899e-01
4.29071486e-01 6.01087153e-01 3.42547633e-02 -4.73368615e-01
-8.86702597e-01 3.43590111e-01 1.01931679e+00 9.11082566e-01
7.17556655e-01 5.66991866e-01 -8.06262046e-02 4.02186424e-01
3.55246484e-01 8.90868783e-01 3.13861787e-01 -1.44134355e+00
2.48407826e-01 3.19098890e-01 3.36809047e-02 -1.17185247e+00
-1.78400695e-01 -5.82515836e-01 -1.13880658e+00 2.42030069e-01
8.42718422e-01 -2.31694505e-01 -4.89143878e-01 1.75387204e+00
4.00873087e-02 2.64858782e-01 -1.09159142e-01 7.69596040e-01
1.59956053e-01 8.27523232e-01 -1.12170897e-01 -1.36835605e-01
9.73905087e-01 -4.08864290e-01 -5.36824703e-01 3.59923393e-03
3.17785561e-01 -6.42035306e-01 1.25682747e+00 3.64372522e-01
-1.45769715e+00 -2.99603850e-01 -9.57345724e-01 -1.20989703e-01
-9.23341215e-02 -4.05265659e-01 9.30105805e-01 1.09533358e+00
-1.19265890e+00 6.32848799e-01 -7.68489003e-01 1.23764269e-01
4.27018315e-01 2.10679010e-01 -5.36728084e-01 4.11785781e-01
-1.14303291e+00 9.53715980e-01 -1.38550252e-01 2.29106829e-01
-6.90110266e-01 -1.43606937e+00 -9.72334981e-01 4.94312763e-01
-9.33301002e-02 -9.00390267e-01 9.43175614e-01 -9.12750781e-01
-1.94207621e+00 4.39672261e-01 -1.35348678e-01 -4.56603914e-01
1.11644769e+00 -9.33528766e-02 -6.15320951e-02 1.12625167e-01
5.13032228e-02 5.27546525e-01 1.02312136e+00 -7.42329597e-01
-2.75216073e-01 -9.65916514e-02 -2.17426363e-02 3.64248194e-02
-4.68815982e-01 -2.56173134e-01 1.62757710e-01 -6.87862992e-01
-3.20459425e-01 -7.46645570e-01 -2.19067350e-01 1.90111622e-01
-2.46602088e-01 1.53920770e-01 5.62974572e-01 -4.19466704e-01
1.02642512e+00 -1.89724672e+00 7.71173686e-02 4.26131874e-01
1.09162509e-01 1.15912385e-01 -1.32274136e-01 3.29603225e-01
-1.41346321e-01 3.10901493e-01 -4.57377821e-01 -1.24839544e-01
3.51400882e-01 1.13464803e-01 -7.12445021e-01 6.94273770e-01
4.12233561e-01 1.14740956e+00 -9.37440813e-01 -2.88564354e-01
3.55067790e-01 9.10523534e-01 -8.46267521e-01 -2.47331560e-02
9.38459411e-02 5.26623845e-01 -1.70923084e-01 -9.18235630e-02
5.62482655e-01 -7.56044239e-02 -1.54387742e-01 9.97685865e-02
-9.26574469e-02 3.15236300e-01 -1.58665860e+00 1.09621882e+00
-8.44462037e-01 7.95930803e-01 2.27276608e-01 -7.81882584e-01
8.31279278e-01 3.00508559e-01 5.81055164e-01 -5.19356191e-01
-6.81800619e-02 3.56080234e-01 -1.26569971e-01 -9.02846828e-02
3.36861759e-01 -7.97269344e-01 7.93678463e-02 3.83872539e-01
7.20236301e-02 -6.19207442e-01 3.77591193e-01 1.88907832e-01
7.67968714e-01 3.20597619e-01 1.23265378e-01 -8.38677764e-01
8.91312301e-01 -3.31213206e-01 5.27725041e-01 5.82312703e-01
4.46185768e-02 7.67235100e-01 9.54078257e-01 -5.04253805e-01
-1.22475004e+00 -1.40547431e+00 -1.63023472e-01 5.97924888e-01
-2.11168900e-01 -6.47192895e-02 -4.39221144e-01 -3.51454616e-01
1.09884322e-01 6.86427772e-01 -7.24676907e-01 8.24564248e-02
-6.44683301e-01 -5.64477265e-01 7.72990048e-01 4.41186219e-01
4.85914022e-01 -8.41076791e-01 -6.71376109e-01 1.56378761e-01
6.09222874e-02 -7.30719924e-01 -8.13489735e-01 -1.02985185e-02
-7.59787023e-01 -8.88789117e-01 -1.07870221e+00 -4.35104728e-01
4.85305607e-01 -5.97584099e-02 1.11892593e+00 -3.08990240e-01
-7.77331740e-02 4.23792034e-01 3.74019146e-01 -2.32235074e-01
-4.77324247e-01 -6.17918931e-03 -8.87744576e-02 2.29082555e-01
-2.59927332e-01 -9.47144270e-01 -5.07675886e-01 3.73088688e-01
-9.67348218e-01 -2.74065942e-01 3.11149359e-01 8.55104744e-01
1.50267348e-01 -1.20800838e-01 7.95671642e-01 -8.31697106e-01
7.38650918e-01 -2.57739693e-01 -7.85941482e-01 1.97653230e-02
-3.99116933e-01 3.01668763e-01 1.01764429e+00 -2.61633694e-01
-1.19855666e+00 -6.73544928e-02 -7.97044188e-02 -2.72948205e-01
1.33549139e-01 3.47484142e-01 6.20680526e-02 -6.57482371e-02
8.26630354e-01 2.67575324e-01 4.18400943e-01 -2.80060917e-02
7.92820156e-01 2.22738028e-01 6.80789948e-01 -6.31701827e-01
1.02342165e+00 7.20600128e-01 4.63702321e-01 -1.02351248e+00
-4.83059496e-01 -2.19883874e-01 -5.15989184e-01 -2.61105567e-01
5.14736414e-01 -5.54401815e-01 -1.08895814e+00 2.04118311e-01
-9.07883406e-01 -4.14962441e-01 -9.14033532e-01 4.35622990e-01
-1.03275728e+00 4.62927073e-01 -5.52348852e-01 -9.89611506e-01
-8.66630003e-02 -8.32735240e-01 7.24133730e-01 -8.67774785e-02
-4.11488235e-01 -1.39601350e+00 2.09525749e-01 -2.86710650e-01
9.62569892e-01 2.35360354e-01 6.64492011e-01 -1.15812086e-01
-5.02490401e-01 -2.26003185e-01 -1.30995512e-01 4.01814520e-01
1.53574109e-01 3.32523108e-01 -8.86631846e-01 3.45640369e-02
1.73788354e-01 1.08105741e-01 6.77679837e-01 4.61778015e-01
5.83910167e-01 -7.23819911e-01 1.27645344e-01 8.80485952e-01
1.09405065e+00 -1.58735067e-01 8.83707166e-01 -2.78584252e-04
5.71119726e-01 5.70900917e-01 -1.11999616e-01 2.32847095e-01
1.63632110e-01 3.92740160e-01 1.65812597e-01 1.35262370e-01
-8.48813131e-02 -3.38962793e-01 4.74922776e-01 9.09436643e-01
-1.01813816e-01 1.02901593e-01 -7.64101684e-01 5.60197055e-01
-1.42399299e+00 -1.23907506e+00 -2.56905347e-01 2.28268337e+00
6.78738832e-01 2.14940667e-01 3.20423186e-01 1.92162722e-01
2.48409063e-01 1.92931131e-01 -1.85511768e-01 -5.72571754e-01
-9.43008587e-02 3.23882729e-01 5.95805705e-01 9.49645698e-01
-9.40383077e-01 4.28734869e-01 7.39568329e+00 6.17968023e-01
-1.04284430e+00 -2.17732161e-01 3.35419923e-01 2.29798984e-02
-8.76735389e-01 3.90629210e-02 -4.27378476e-01 5.40842950e-01
9.23007369e-01 -5.85410416e-01 5.28886259e-01 5.67909658e-01
2.08657086e-01 2.35743761e-01 -9.99715924e-01 6.77286863e-01
-1.48698017e-01 -1.35412073e+00 2.15119317e-01 2.34993666e-01
5.83961248e-01 -4.02477324e-01 4.05566722e-01 2.30981335e-01
3.41781944e-01 -1.32043421e+00 7.93442309e-01 7.24078178e-01
6.19080603e-01 -7.59301126e-01 4.33680743e-01 3.95422101e-01
-9.98471022e-01 7.18653798e-02 -2.22638622e-01 -3.08655471e-01
5.83096504e-01 6.78490281e-01 -6.51944578e-01 4.88169909e-01
3.79346274e-02 7.95020938e-01 -6.20850537e-04 1.16270220e+00
-1.17418766e-01 5.17642200e-01 -6.77069306e-01 1.53834850e-01
3.27620506e-01 -6.36489511e-01 7.23769009e-01 1.29633451e+00
6.19019628e-01 -2.76443303e-01 -2.98761100e-01 1.07236362e+00
2.42070153e-01 1.11680001e-01 -9.03662860e-01 7.52780959e-02
1.82838067e-01 1.05654776e+00 -5.78423023e-01 -1.62623540e-01
-2.91449696e-01 5.63076556e-01 -2.44973414e-02 3.46144259e-01
-6.96308374e-01 -3.50067794e-01 8.32827091e-01 2.80457228e-01
1.83087915e-01 -5.63459277e-01 -6.00684524e-01 -1.20184696e+00
7.69811273e-02 -6.12183094e-01 1.65585428e-01 -6.29821181e-01
-1.34404290e+00 5.17119110e-01 -1.69771425e-02 -1.29245257e+00
-5.69860637e-01 -6.75090492e-01 -7.41054893e-01 1.00452149e+00
-1.51099086e+00 -1.04268503e+00 -1.73431262e-01 4.64148283e-01
8.75556171e-02 4.26125713e-02 7.02044487e-01 3.86700690e-01
-2.04213381e-01 4.13027108e-01 -1.55758455e-01 -1.48491889e-01
3.01870257e-01 -1.57232189e+00 4.35436904e-01 9.19927239e-01
2.91301399e-01 8.26858819e-01 8.59091401e-01 -3.84673029e-01
-1.34495389e+00 -7.44819522e-01 7.84887612e-01 -6.05564952e-01
9.01942253e-01 -2.08303124e-01 -9.23243344e-01 9.70420241e-01
7.40640834e-02 8.67683515e-02 1.57829151e-01 -8.13050494e-02
-4.22974169e-01 -1.81788653e-01 -8.76341879e-01 6.03199780e-01
9.83508468e-01 -3.75739008e-01 -4.25450623e-01 5.51658906e-02
1.86088040e-01 -3.04662496e-01 -1.05446041e+00 2.52226323e-01
6.78924084e-01 -1.18702602e+00 1.18263018e+00 -3.23711574e-01
3.59377265e-01 -2.39723727e-01 2.93100387e-01 -1.37006497e+00
-2.30879828e-01 -1.28648329e+00 3.24703790e-02 1.10542798e+00
4.84941393e-01 -1.18441117e+00 7.94875979e-01 7.43704259e-01
4.96823415e-02 -3.29058081e-01 -1.04270172e+00 -1.15135789e+00
5.42911351e-01 -5.54855645e-01 6.87066257e-01 7.99346924e-01
3.24584424e-01 2.17468172e-01 -5.60205877e-01 -2.59590983e-01
6.78569198e-01 -2.66203135e-01 9.56348896e-01 -1.31894410e+00
-2.94191986e-01 -8.88246357e-01 -4.87810969e-01 -1.39344680e+00
1.92008659e-01 -1.07264721e+00 1.47186667e-02 -1.23895717e+00
-1.81740671e-01 -6.50967538e-01 2.35666782e-01 6.28696010e-02
1.13057327e-02 4.15808409e-01 2.63178438e-01 3.58592384e-02
6.97570518e-02 7.58680820e-01 1.45164371e+00 1.04786970e-01
-2.76756227e-01 1.51212782e-01 -7.61332095e-01 7.88505256e-01
5.64380646e-01 -1.50905386e-01 -6.88097358e-01 -1.72946051e-01
5.72708488e-01 1.04795732e-01 5.13490438e-01 -9.11870599e-01
2.77602285e-01 -7.08976239e-02 2.32122436e-01 5.85323907e-02
3.82506788e-01 -7.52526700e-01 4.84554827e-01 2.85793841e-01
-3.96712810e-01 4.74649295e-02 1.30271792e-01 4.21553344e-01
-2.36382395e-01 -1.67754650e-01 8.04546058e-01 -5.78062087e-02
-2.03500971e-01 2.67894804e-01 -4.98531133e-01 3.53872508e-01
7.92363405e-01 -3.91408592e-01 -1.79697067e-01 -7.78773665e-01
-6.80920005e-01 -1.62082147e-02 3.44061583e-01 1.50389910e-01
3.46400470e-01 -1.32478905e+00 -8.49555314e-01 6.75148249e-01
-5.82081914e-01 -3.05623859e-02 1.10958785e-01 9.95805681e-01
-8.25573325e-01 6.42831326e-01 -3.57285947e-01 -5.25279880e-01
-5.78830898e-01 3.84880573e-01 5.14339089e-01 -2.82831520e-01
-9.17870343e-01 5.76793790e-01 4.20240194e-01 -4.98877257e-01
6.19085552e-03 -2.68760622e-01 1.01076365e-01 1.62173942e-01
4.84361976e-01 6.84912384e-01 -9.37255099e-02 -6.05486453e-01
-2.57716745e-01 6.03795886e-01 4.83429193e-01 -4.14168388e-01
1.17643452e+00 -1.31221429e-01 -9.63864326e-02 2.87399024e-01
1.19829178e+00 3.12598497e-01 -1.65774095e+00 2.81523705e-01
-1.71636000e-01 -6.31331146e-01 -2.29918331e-01 -3.91585946e-01
-1.10122156e+00 9.77225304e-01 -9.03042406e-02 8.01972270e-01
6.68807089e-01 -4.81402725e-01 5.71014404e-01 4.58664820e-02
4.03112732e-02 -5.15138447e-01 -9.00308266e-02 5.44310033e-01
1.07030487e+00 -5.55966437e-01 -1.40919790e-01 -5.25922000e-01
-5.53068161e-01 1.19950712e+00 1.90848961e-01 -6.05594635e-01
8.57291281e-01 5.02167761e-01 6.25223145e-02 9.48170274e-02
-5.68091154e-01 1.36798844e-02 3.23574990e-01 7.05174208e-01
3.22252423e-01 -7.42940307e-02 -1.86342336e-02 -1.92146480e-01
-6.28778219e-01 4.44260836e-02 8.31173301e-01 2.68456250e-01
-2.20015079e-01 -7.38123953e-01 -3.08794051e-01 3.57063502e-01
-4.79973167e-01 5.94224036e-03 1.05166659e-01 1.00409019e+00
-5.34065723e-01 5.67734838e-01 3.24737489e-01 3.63485724e-01
4.93639201e-01 1.04071759e-01 4.76565242e-01 -1.77953675e-01
-4.13435698e-01 -5.56588992e-02 -8.16371292e-02 -4.93375182e-01
-3.11803490e-01 -6.82863712e-01 -8.01265419e-01 -6.01887405e-01
-1.36254385e-01 2.60917813e-01 3.35278869e-01 7.47714996e-01
1.82292819e-01 4.51472789e-01 5.84934473e-01 -9.69131589e-01
-7.92371392e-01 -7.84766674e-01 -7.28586137e-01 3.13966215e-01
7.05394566e-01 -6.04012489e-01 -7.96415150e-01 1.48868293e-01] | [7.171891689300537, 3.8037006855010986] |
bc93e4e9-490e-426e-9b60-89a0e2727f97 | an-efficient-framework-for-few-shot-skeleton | 2207.09925 | null | https://arxiv.org/abs/2207.09925v1 | https://arxiv.org/pdf/2207.09925v1.pdf | An Efficient Framework for Few-shot Skeleton-based Temporal Action Segmentation | Temporal action segmentation (TAS) aims to classify and locate actions in the long untrimmed action sequence. With the success of deep learning, many deep models for action segmentation have emerged. However, few-shot TAS is still a challenging problem. This study proposes an efficient framework for the few-shot skeleton-based TAS, including a data augmentation method and an improved model. The data augmentation approach based on motion interpolation is presented here to solve the problem of insufficient data, and can increase the number of samples significantly by synthesizing action sequences. Besides, we concatenate a Connectionist Temporal Classification (CTC) layer with a network designed for skeleton-based TAS to obtain an optimized model. Leveraging CTC can enhance the temporal alignment between prediction and ground truth and further improve the segment-wise metrics of segmentation results. Extensive experiments on both public and self-constructed datasets, including two small-scale datasets and one large-scale dataset, show the effectiveness of two proposed methods in improving the performance of the few-shot skeleton-based TAS task. | ['Lin Yuan', 'Xiaotian Lin', 'Qiang Wang', 'Leiyang Xu'] | 2022-07-20 | null | null | null | null | ['action-segmentation'] | ['computer-vision'] | [ 6.33564413e-01 -1.90472230e-01 -4.95508254e-01 -3.34408998e-01
-9.82360840e-01 8.54295492e-02 3.06585729e-01 -4.22815084e-01
-4.26853836e-01 3.85969281e-01 4.90088195e-01 2.20290482e-01
2.13962212e-01 -5.92319787e-01 -4.95671839e-01 -7.26628125e-01
1.93173334e-01 7.46878088e-02 9.44529593e-01 -1.92954820e-02
2.18923554e-01 5.87527491e-02 -1.33837891e+00 4.58100259e-01
1.12598610e+00 1.24331689e+00 1.31019145e-01 4.75316375e-01
-2.41280034e-01 1.01454520e+00 -4.07416284e-01 -1.86455324e-01
4.30768043e-01 -8.53272915e-01 -8.54064822e-01 4.97953653e-01
2.11660698e-01 -5.04887879e-01 -3.69538128e-01 6.90283716e-01
3.87971193e-01 5.83056152e-01 3.63277286e-01 -1.32929742e+00
-2.94849873e-01 6.19684994e-01 -7.53982008e-01 2.17171907e-01
1.10392787e-01 5.84980607e-01 9.41751540e-01 -4.34621751e-01
6.43007576e-01 1.08658218e+00 7.72771657e-01 7.05089033e-01
-8.39801967e-01 -4.94824499e-01 3.09570640e-01 5.72762489e-01
-9.53133881e-01 -3.30542743e-01 8.35391819e-01 -4.18292314e-01
7.78579593e-01 -1.07817844e-01 1.02756429e+00 1.32960582e+00
-8.05266574e-02 1.40888846e+00 6.79552555e-01 -5.84609993e-02
1.88340321e-01 -7.30766535e-01 -3.30101922e-02 6.88807487e-01
-3.69657159e-01 -3.13103124e-02 -5.23455799e-01 3.07433575e-01
8.31277490e-01 2.76449531e-01 3.03048585e-02 -3.41360062e-01
-1.35815167e+00 4.57624704e-01 4.24182415e-01 4.03453320e-01
-4.13596720e-01 2.66110659e-01 7.15924442e-01 -3.13646525e-01
4.60790783e-01 1.39186367e-01 -4.69432861e-01 -6.84407830e-01
-1.15672982e+00 1.77362412e-02 2.05465183e-01 8.12583148e-01
3.75427037e-01 2.61571586e-01 -7.05913007e-01 8.30742300e-01
-1.17523056e-02 1.88551694e-01 7.83896446e-01 -1.21070206e+00
6.89964533e-01 9.38847005e-01 3.08743734e-02 -7.74468124e-01
-2.94157565e-01 -9.15265977e-02 -6.11559927e-01 -1.84935972e-01
5.61555207e-01 -1.30698085e-01 -1.22523189e+00 1.57743227e+00
5.83637774e-01 8.47963035e-01 -2.37396926e-01 9.87813175e-01
6.04159355e-01 5.57426929e-01 2.86276311e-01 -1.48492366e-01
9.42771316e-01 -1.61643362e+00 -7.89007425e-01 -3.00056189e-01
8.53210092e-01 -3.31051379e-01 1.18332195e+00 1.33737519e-01
-8.89400184e-01 -8.69231462e-01 -9.63772714e-01 -2.55207002e-01
-1.07453810e-02 2.91840076e-01 5.36039233e-01 3.59948218e-01
-3.68596971e-01 8.02036405e-01 -1.30829847e+00 -3.50539565e-01
1.03476954e+00 1.33364405e-02 -2.27643445e-01 -5.82662635e-02
-8.83902550e-01 4.76816595e-01 5.66016316e-01 2.66471207e-01
-1.08593309e+00 -4.62961406e-01 -9.17621136e-01 -8.63202885e-02
7.23424554e-01 -3.17454338e-01 1.36350930e+00 -1.11102164e+00
-1.58774531e+00 3.98822516e-01 -1.90857619e-01 -7.53502309e-01
7.79441595e-01 -2.83494562e-01 -2.53452390e-01 3.14964086e-01
4.54442859e-01 8.25971127e-01 6.62609696e-01 -6.40120685e-01
-8.97789657e-01 -3.10886681e-01 -1.92246139e-01 6.98109865e-02
-3.18643779e-01 -1.24110445e-01 -7.73729384e-01 -8.27330947e-01
1.63446665e-01 -8.67784321e-01 -6.33732021e-01 5.91182560e-02
-3.13137919e-01 -8.04037452e-02 8.25198650e-01 -1.01801229e+00
1.21408749e+00 -2.20939589e+00 1.96377486e-01 -3.41200978e-01
-1.41336799e-01 4.20371473e-01 -3.47159892e-01 1.53680563e-01
6.40952438e-02 -1.19166985e-01 -7.57099450e-01 -4.12172794e-01
-2.07707509e-01 4.51869130e-01 -3.56315598e-02 3.10019225e-01
3.02137882e-01 1.25560021e+00 -9.64321017e-01 -9.20196116e-01
3.63640457e-01 6.87698051e-02 -3.91196758e-01 1.57870844e-01
-5.95218658e-01 9.17983413e-01 -6.72416151e-01 7.57278502e-01
2.63025254e-01 -2.51855135e-01 -1.66206002e-01 -8.45588520e-02
-7.70989582e-02 7.25003630e-02 -1.01673484e+00 2.32000589e+00
-7.76117593e-02 4.27809447e-01 -4.91741866e-01 -1.03667009e+00
6.23294234e-01 2.22817048e-01 1.08891082e+00 -9.55993235e-01
3.37063104e-01 -4.72937003e-02 -1.21536545e-01 -8.60610008e-01
4.69014227e-01 -6.97947443e-02 1.21239126e-01 5.31968892e-01
-5.39827086e-02 2.73802519e-01 4.59501624e-01 2.26326305e-02
1.14089692e+00 7.94226885e-01 -2.67175622e-02 3.94414991e-01
4.53114212e-01 2.56748587e-01 1.13107288e+00 2.71165967e-01
-7.38866150e-01 8.34886253e-01 4.17246222e-01 -4.00086641e-01
-9.69737589e-01 -5.74403465e-01 3.49298388e-01 9.82484698e-01
3.70459914e-01 -3.82681280e-01 -1.02428758e+00 -9.76515293e-01
-3.62273574e-01 6.72135592e-01 -6.94846869e-01 -3.34814966e-01
-7.71482646e-01 -6.64544523e-01 6.07772052e-01 1.13315129e+00
1.08046246e+00 -1.15330720e+00 -7.33893275e-01 3.23094189e-01
-7.83646047e-01 -1.34904325e+00 -6.42700911e-01 -2.76580364e-01
-1.12890112e+00 -1.16365123e+00 -8.78455043e-01 -5.65727651e-01
5.11622906e-01 2.85699040e-01 5.20202339e-01 -9.25122425e-02
-2.35431850e-01 1.61711961e-01 -6.57866240e-01 -1.34703800e-01
-1.50655434e-01 5.41959852e-02 -3.95024896e-01 4.32227761e-01
3.31307560e-01 -4.51412648e-01 -8.22620451e-01 5.21415472e-01
-8.22552502e-01 3.18068683e-01 6.12092018e-01 5.23616612e-01
7.80595303e-01 -1.79150611e-01 5.21022499e-01 -5.29405892e-01
1.05096526e-01 -1.74231321e-01 -1.91126719e-01 2.22217917e-01
-3.80978316e-01 -8.22465718e-02 3.72704268e-01 -5.51275730e-01
-1.37352669e+00 4.56261545e-01 -1.30569398e-01 -6.12240911e-01
-1.13821574e-01 3.09088707e-01 -1.63123176e-01 1.52812511e-01
4.91494924e-01 4.70476061e-01 2.24056337e-02 -6.50450766e-01
5.24272621e-01 4.50357556e-01 6.37883842e-01 -3.48933190e-01
4.35910106e-01 8.01726997e-01 -1.34224147e-01 -5.26445448e-01
-1.10698450e+00 -5.80683291e-01 -1.12638128e+00 -4.05449212e-01
1.37828970e+00 -7.80765057e-01 -1.64611340e-01 7.84871340e-01
-9.90737796e-01 -7.31719613e-01 -5.54220080e-01 4.58057016e-01
-8.63679051e-01 6.91833615e-01 -5.73485672e-01 -6.94329619e-01
-1.98720992e-01 -1.06078506e+00 1.18397832e+00 2.98563719e-01
-1.27351195e-01 -6.10524476e-01 1.88692868e-01 8.33112180e-01
-1.47587821e-01 4.72903073e-01 4.38635230e-01 -4.63044167e-01
-6.68849051e-01 -3.16430002e-01 -1.60249889e-01 5.26610076e-01
1.27786636e-01 1.85719609e-01 -7.02548265e-01 1.52375266e-01
-8.98231491e-02 -3.50716263e-01 1.07429814e+00 5.81945956e-01
1.36905968e+00 2.16584839e-02 -2.95237511e-01 4.76595730e-01
9.45429146e-01 3.73754799e-01 8.36363137e-01 3.04159462e-01
1.01816118e+00 3.99295419e-01 1.29238105e+00 5.07217407e-01
3.23478341e-01 6.64782882e-01 3.50442737e-01 2.85137631e-02
-2.74634778e-01 -4.44598973e-01 2.97539949e-01 7.45664716e-01
-3.74160945e-01 7.88909718e-02 -8.45862269e-01 8.18927944e-01
-2.30007577e+00 -1.14122558e+00 -1.94317296e-01 1.92273104e+00
7.30726719e-01 2.62572199e-01 4.71645981e-01 1.52164549e-01
6.78198814e-01 4.65134770e-01 -8.15260410e-01 1.11927688e-01
4.02524509e-02 -9.12382379e-02 3.26211572e-01 2.71090027e-02
-1.43506277e+00 1.24701107e+00 5.57232618e+00 1.11344850e+00
-9.07383919e-01 2.32047006e-01 6.12672865e-01 -9.85512808e-02
3.61761987e-01 2.63414420e-02 -5.61554730e-01 6.67704403e-01
7.96882749e-01 -1.54716559e-02 5.02047949e-02 9.70690966e-01
7.18344927e-01 -2.16162145e-01 -9.70943928e-01 8.23850214e-01
5.16905747e-02 -1.23709524e+00 -8.88958126e-02 -2.60202497e-01
9.98454094e-01 -2.30748318e-02 -2.25723609e-01 4.20084417e-01
2.89408676e-02 -7.36891627e-01 7.01292753e-01 6.07516706e-01
5.31722248e-01 -5.69856703e-01 5.48864365e-01 3.44676703e-01
-1.58154213e+00 -3.26023132e-01 -6.13590926e-02 -4.62090448e-02
5.35160005e-01 3.77345495e-02 -2.86082387e-01 6.58382297e-01
5.93843162e-01 1.50143886e+00 -7.22900152e-01 1.34659922e+00
-3.08703989e-01 6.37200654e-01 -7.23902835e-03 1.88756540e-01
5.93203247e-01 -3.20026547e-01 3.40129077e-01 8.61428559e-01
7.12470114e-02 2.55908757e-01 3.87363821e-01 7.09124684e-01
2.37775281e-01 -4.76548634e-03 -1.54477820e-01 -4.19129550e-01
3.96843851e-02 1.20584452e+00 -9.75139678e-01 -5.20853460e-01
-4.02154028e-01 1.16760170e+00 -4.87557612e-03 2.89890885e-01
-1.33149052e+00 -2.03468069e-01 4.89221364e-01 -6.37999922e-02
4.86422867e-01 -3.68342161e-01 -5.30049026e-01 -1.17082632e+00
1.27749026e-01 -8.93063426e-01 4.80557203e-01 -7.53839552e-01
-8.79116774e-01 1.20356552e-01 -7.08498955e-02 -1.72373581e+00
-8.36636499e-02 -1.24710388e-01 -7.28248298e-01 3.29601347e-01
-1.10542083e+00 -1.24185359e+00 -5.04574239e-01 5.48919857e-01
1.12625873e+00 1.48005009e-01 2.88425386e-01 4.73329395e-01
-1.08026218e+00 2.10570201e-01 -1.36874199e-01 4.26823527e-01
4.15642500e-01 -8.66971791e-01 5.72884142e-01 1.20977545e+00
1.85426369e-01 1.03294693e-01 2.62066364e-01 -9.04358149e-01
-9.14896727e-01 -1.47394943e+00 4.11629170e-01 -2.99982101e-01
6.89468384e-01 3.15250792e-02 -9.31692481e-01 7.17202127e-01
-2.04847500e-01 -1.12404162e-02 4.14804369e-01 -4.31294262e-01
-3.58881848e-03 -3.90218198e-02 -9.17222083e-01 8.42426538e-01
1.41886926e+00 -2.67111272e-01 -5.98411143e-01 3.54256332e-01
8.95439386e-01 -4.19816941e-01 -7.32406318e-01 4.16815549e-01
6.60120666e-01 -9.61316347e-01 8.76393139e-01 -8.48784983e-01
8.26292336e-01 -2.44778231e-01 7.74777606e-02 -8.39216948e-01
-8.40656087e-02 -4.78345990e-01 -2.22727135e-01 1.16187537e+00
8.13223049e-02 -1.45132989e-01 1.15979862e+00 6.91212952e-01
-4.39112633e-01 -9.39193010e-01 -9.24752653e-01 -9.73452449e-01
-3.50957572e-01 -6.93727493e-01 4.52911854e-01 6.74238205e-01
-1.80131331e-01 1.14862993e-01 -5.42669773e-01 -2.41293609e-01
4.64239866e-01 6.84742033e-02 9.01351035e-01 -8.52121472e-01
-1.12986028e-01 -3.63733143e-01 -4.66341078e-01 -1.34116805e+00
9.33140740e-02 -5.96274912e-01 3.54373962e-01 -1.72062290e+00
1.98615551e-01 -1.00650236e-01 -2.78903842e-01 6.97261870e-01
-3.13038975e-01 1.44897729e-01 1.19687416e-01 1.11643016e-01
-9.42347884e-01 1.07185292e+00 1.65950930e+00 -1.86334774e-01
-3.34429801e-01 8.74295905e-02 -1.06518969e-01 9.20124650e-01
6.93389952e-01 -4.32129413e-01 -5.47221363e-01 -3.30402225e-01
-4.65136588e-01 1.65945977e-01 3.88394386e-01 -1.38249159e+00
1.64947405e-01 -5.21046102e-01 2.33939439e-01 -8.14429045e-01
4.58669573e-01 -7.23699093e-01 -4.14872319e-02 5.87789059e-01
-4.81340557e-01 -4.74432230e-01 1.01705911e-02 9.43647265e-01
-2.35503182e-01 -9.19924974e-02 8.92083406e-01 -1.92443117e-01
-1.21484840e+00 8.11741054e-01 -3.83614302e-02 2.28595302e-01
1.29322350e+00 -6.50241971e-01 -7.29662329e-02 -9.52629521e-02
-7.09607899e-01 4.90094721e-01 2.43120685e-01 5.76297343e-01
6.17281795e-01 -1.38428485e+00 -3.32883358e-01 -1.00171350e-01
1.13906078e-01 1.69606924e-01 5.47116697e-01 1.38603139e+00
-4.48857397e-01 1.41026840e-01 -4.59423512e-01 -6.90289319e-01
-1.27745211e+00 4.13319021e-01 3.97369325e-01 -1.63973525e-01
-9.70629394e-01 8.66654098e-01 5.52822556e-03 -9.13213789e-02
2.84978449e-01 -5.51137924e-01 -3.40552837e-01 7.19810873e-02
4.96125877e-01 7.28895366e-01 -3.47720385e-01 -7.30034113e-01
-2.13098720e-01 5.82011461e-01 1.70049325e-01 -1.92392111e-01
1.36033392e+00 -1.20541371e-01 1.49771035e-01 6.74117148e-01
9.71863508e-01 -6.99082136e-01 -1.85413146e+00 -2.03044057e-01
8.41321424e-02 -7.30431974e-01 -2.01464772e-01 -5.46480536e-01
-1.36696577e+00 1.15737641e+00 4.45079952e-01 -1.64191678e-01
1.10788572e+00 -3.51002336e-01 1.55942988e+00 1.22594036e-01
2.46878803e-01 -1.42680824e+00 6.70659661e-01 3.91662776e-01
5.66170812e-01 -1.35268497e+00 -1.59501031e-01 -3.74390811e-01
-1.08488405e+00 9.40819442e-01 9.36052144e-01 3.74239385e-02
1.17939197e-01 -2.65691459e-01 -6.54312372e-02 -1.94757318e-04
-3.77562016e-01 -5.20672023e-01 2.58757085e-01 5.35979450e-01
1.09306572e-03 -1.77294701e-01 -4.54479188e-01 8.42824578e-01
2.70803958e-01 4.01606619e-01 3.13828379e-01 1.10839319e+00
-5.48974812e-01 -1.00297439e+00 5.13198301e-02 4.55714852e-01
-2.75623858e-01 2.06107944e-01 -4.12007093e-01 4.72734153e-01
2.45194972e-01 9.05088723e-01 -8.41041207e-02 -5.31455934e-01
2.23572657e-01 1.78112030e-01 3.50344568e-01 -5.21743119e-01
-3.39469433e-01 1.49508998e-01 7.11402148e-02 -9.76847649e-01
-8.05566609e-01 -8.45283449e-01 -1.59243321e+00 1.21797204e-01
-1.88721761e-01 -2.75144666e-01 2.33256221e-01 1.29385769e+00
3.87944102e-01 7.42255628e-01 4.43183035e-01 -7.73221970e-01
-3.41762036e-01 -9.64118302e-01 -4.66658890e-01 4.59969461e-01
-8.11140053e-03 -7.30590284e-01 2.95907389e-02 4.53363687e-01] | [8.421177864074707, 0.5094887614250183] |
4bbe569e-1b12-4ec4-91ff-10e5d1466bd1 | thompson-sampling-for-improved-exploration-in | 2306.17693 | null | https://arxiv.org/abs/2306.17693v1 | https://arxiv.org/pdf/2306.17693v1.pdf | Thompson sampling for improved exploration in GFlowNets | Generative flow networks (GFlowNets) are amortized variational inference algorithms that treat sampling from a distribution over compositional objects as a sequential decision-making problem with a learnable action policy. Unlike other algorithms for hierarchical sampling that optimize a variational bound, GFlowNet algorithms can stably run off-policy, which can be advantageous for discovering modes of the target distribution. Despite this flexibility in the choice of behaviour policy, the optimal way of efficiently selecting trajectories for training has not yet been systematically explored. In this paper, we view the choice of trajectories for training as an active learning problem and approach it using Bayesian techniques inspired by methods for multi-armed bandits. The proposed algorithm, Thompson sampling GFlowNets (TS-GFN), maintains an approximate posterior distribution over policies and samples trajectories from this posterior for training. We show in two domains that TS-GFN yields improved exploration and thus faster convergence to the target distribution than the off-policy exploration strategies used in past work. | ['Yoshua Bengio', 'Nikolay Malkin', 'Sarath Chandar', 'Cheng-Hao Liu', 'Maksym Korablyov', 'Moksh Jain', 'Kanika Madan', 'Jarrid Rector-Brooks'] | 2023-06-30 | null | null | null | null | ['thompson-sampling', 'active-learning', 'multi-armed-bandits', 'active-learning', 'decision-making'] | ['methodology', 'methodology', 'miscellaneous', 'natural-language-processing', 'reasoning'] | [ 1.27561586e-02 3.97556484e-01 -9.05931175e-01 -1.70534790e-01
-1.06047344e+00 -4.56485540e-01 9.30596292e-01 -3.12244982e-01
-5.52616835e-01 1.45945859e+00 1.71890706e-01 -4.96934444e-01
-5.51610291e-01 -9.78094280e-01 -8.78486335e-01 -1.00684679e+00
-4.66084927e-02 1.16408253e+00 2.74799943e-01 4.91047233e-01
2.16294721e-01 3.86030912e-01 -1.49417639e+00 -7.05299452e-02
1.04615533e+00 5.85644424e-01 2.21005842e-01 7.29401231e-01
-3.87007892e-01 7.41809428e-01 -4.24526036e-01 -2.06438959e-01
1.97804123e-01 -7.33777761e-01 -8.41247559e-01 -4.08756062e-02
6.63492531e-02 -4.23432857e-01 1.42853707e-01 1.19707251e+00
7.74816945e-02 7.51821220e-01 1.06561220e+00 -1.18408346e+00
-1.79667562e-01 9.82331753e-01 -2.43284866e-01 2.23597199e-01
-1.57283604e-01 2.89607227e-01 9.92009461e-01 -5.01316547e-01
7.11570680e-01 1.53823471e+00 1.53570130e-01 5.55016220e-01
-1.70733547e+00 -6.28309429e-01 5.64491689e-01 1.43173456e-01
-1.08001816e+00 -4.70315903e-01 5.46462297e-01 -3.51587892e-01
6.70839369e-01 3.16412002e-01 1.13677001e+00 1.45114934e+00
-3.82972099e-02 1.30663574e+00 9.49893057e-01 -5.45428813e-01
1.11263919e+00 2.76820779e-01 -2.89280117e-01 5.69647491e-01
4.41105127e-01 3.56650531e-01 -6.18638217e-01 -6.05656803e-01
1.01562381e+00 1.02479041e-01 -7.34222457e-02 -8.53503048e-01
-8.12607646e-01 1.30865598e+00 1.86159760e-02 -1.27687469e-01
-5.85474610e-01 5.57520986e-01 2.08543941e-01 1.88001975e-01
6.09112561e-01 3.36992651e-01 -2.94151336e-01 -4.50170010e-01
-1.32823849e+00 8.02922249e-01 9.53021348e-01 8.10597301e-01
9.02194917e-01 2.32398182e-01 -5.58178544e-01 5.30954599e-01
6.96559906e-01 5.36208272e-01 5.20590916e-02 -1.42930686e+00
2.30204135e-01 1.68527439e-01 6.15493596e-01 -1.24221377e-01
3.63317221e-01 -3.38100106e-01 -3.34918290e-01 5.22377193e-01
7.20740139e-01 -4.65819776e-01 -9.54802990e-01 1.69536889e+00
5.36384225e-01 -7.28140920e-02 -2.15503097e-01 5.48277557e-01
-1.74039006e-01 7.65652597e-01 2.95456469e-01 -6.41682029e-01
6.34455264e-01 -7.54597902e-01 -5.99489808e-01 1.07905433e-01
2.81503350e-01 -1.65183678e-01 8.00963104e-01 7.18114793e-01
-1.30882716e+00 -2.11618200e-01 -5.74138105e-01 4.85493034e-01
-3.14110070e-02 -1.46996230e-01 7.53140271e-01 7.98467577e-01
-1.02038634e+00 9.49572384e-01 -1.35585403e+00 -1.37422934e-01
9.56981421e-01 2.27206469e-01 4.05108422e-01 2.52374530e-01
-8.74062479e-01 6.43997252e-01 6.49465740e-01 -1.61436602e-01
-1.69354522e+00 -7.44226873e-01 -3.23094934e-01 1.90299839e-01
8.42420340e-01 -6.34944320e-01 1.54270124e+00 -9.56737757e-01
-2.05640507e+00 2.83163749e-02 -3.43266696e-01 -8.59000564e-01
8.83540452e-01 -1.79440945e-01 2.50814885e-01 5.21123111e-02
6.45232722e-02 6.41099513e-01 1.27459669e+00 -1.09477556e+00
-6.32730663e-01 -1.31729513e-01 -3.93837094e-02 2.72723019e-01
8.01123306e-02 -4.89393771e-01 -1.48106396e-01 -2.36163422e-01
-2.21150845e-01 -9.21705008e-01 -6.01278424e-01 -7.67129511e-02
-4.56381112e-01 -6.44323170e-01 6.15837812e-01 2.61987397e-03
1.21958184e+00 -1.68923748e+00 2.96524614e-01 5.36752880e-01
-2.22340990e-02 1.15536638e-01 4.38190013e-01 5.48917592e-01
5.42656898e-01 2.25932404e-01 -2.42680222e-01 -4.54331696e-01
2.58680820e-01 5.35402596e-01 -6.75782144e-01 5.51391304e-01
-1.52195543e-01 6.45541251e-01 -1.28668737e+00 -6.49389148e-01
2.44769096e-01 2.91607287e-02 -8.13773274e-01 1.73011586e-01
-9.30749416e-01 4.77863014e-01 -8.15027833e-01 3.55504662e-01
3.69682074e-01 -1.87033653e-01 4.98780340e-01 5.49632907e-01
-2.35599965e-01 9.39662084e-02 -1.12950230e+00 1.48225474e+00
-3.11148018e-01 4.50693876e-01 -5.96963502e-02 -1.02119589e+00
8.18697691e-01 2.89788544e-01 4.20067012e-01 -8.53430107e-02
-2.70458125e-02 1.72754705e-01 -1.26068249e-01 -2.91034639e-01
3.81108999e-01 -1.22769162e-01 2.67530411e-01 6.39506221e-01
1.87954932e-01 -6.19553849e-02 4.44524437e-01 1.76056802e-01
8.01403999e-01 7.62161076e-01 2.05302253e-01 -5.97181141e-01
3.29396874e-02 2.27139533e-01 6.33680046e-01 1.52937698e+00
-1.51017353e-01 7.02740699e-02 7.07103014e-01 -3.47166389e-01
-1.08261931e+00 -1.39951706e+00 -1.75610214e-01 1.08689678e+00
-2.41019174e-01 -1.75053194e-01 -7.45812833e-01 -7.35183954e-01
3.23168710e-02 1.12950647e+00 -6.83450460e-01 2.15309799e-01
-5.04067719e-01 -6.06860161e-01 1.01488911e-01 2.93196589e-01
3.57923985e-01 -1.17946970e+00 -9.96425331e-01 4.28374827e-01
1.31728172e-01 -2.81898946e-01 -2.05175772e-01 3.12243909e-01
-1.16592395e+00 -9.04244125e-01 -7.95687795e-01 3.93777387e-03
5.70199728e-01 -4.87316310e-01 1.17712510e+00 -5.16212404e-01
6.95236921e-02 2.63709903e-01 -3.68272588e-02 -6.09074116e-01
-4.46509629e-01 2.99995840e-01 -1.59639150e-01 -3.93234752e-02
2.14801028e-01 -5.05331933e-01 -6.81779325e-01 -2.84410664e-03
-6.43267155e-01 -2.07958780e-02 3.43320072e-01 8.83901238e-01
7.01501489e-01 1.74723566e-02 5.13105869e-01 -1.11732244e+00
7.38958597e-01 -6.40826523e-01 -1.06201637e+00 2.51735538e-01
-7.17631280e-01 5.45486271e-01 4.95976269e-01 -6.68462694e-01
-1.40023732e+00 -6.45804480e-02 3.06759417e-01 -9.18869078e-01
-1.63957804e-01 5.73396124e-02 1.61726773e-01 5.45418680e-01
6.10912144e-01 1.33410409e-01 1.29210800e-01 -4.08257872e-01
4.74136263e-01 2.48142451e-01 -8.42666328e-02 -1.07450962e+00
3.74590248e-01 4.94176507e-01 5.78834116e-02 -8.19574118e-01
-9.40160096e-01 -1.07372150e-01 -3.36962551e-01 -4.29132342e-01
8.02602470e-01 -5.83255649e-01 -8.45150411e-01 -3.13581713e-02
-7.72649407e-01 -7.19106793e-01 -8.80821645e-01 6.60277009e-01
-1.12707853e+00 -2.20494404e-01 -7.77594298e-02 -1.78383100e+00
-3.42565551e-02 -1.11810684e+00 7.83023417e-01 3.97247344e-01
-5.11599183e-01 -1.15792274e+00 5.57590663e-01 -6.93169087e-02
4.39645320e-01 2.14658380e-01 7.72525370e-01 -4.14152354e-01
-1.14652014e+00 3.52023095e-01 4.30661649e-01 -7.66715854e-02
-1.08885661e-01 3.46834004e-01 -7.69770086e-01 -5.21998644e-01
-1.30484670e-01 -4.76475984e-01 1.03664768e+00 1.11597371e+00
1.18070853e+00 -5.58344603e-01 -6.01144612e-01 4.84420091e-01
1.38144636e+00 5.65838635e-01 4.90070313e-01 2.30805710e-01
2.47879341e-01 3.38089913e-01 5.19502282e-01 5.38723767e-01
-2.94165939e-01 2.54268587e-01 4.18788195e-01 4.51353818e-01
5.38391888e-01 -6.52283490e-01 2.77691394e-01 -1.10207796e-01
-1.59244820e-01 -1.97371826e-01 -7.26583660e-01 7.75138140e-01
-2.16541052e+00 -1.29431105e+00 4.95144397e-01 2.39722133e+00
1.03768504e+00 3.84323835e-01 6.63035095e-01 -1.87384754e-01
5.76682150e-01 9.09131393e-02 -1.08469164e+00 -4.70247954e-01
4.96206850e-01 3.03099066e-01 7.13166058e-01 6.41674757e-01
-8.43402445e-01 7.90211499e-01 6.79509735e+00 1.12668490e+00
-7.38046169e-01 1.50679827e-01 7.06730127e-01 -5.07499397e-01
-5.41826785e-01 4.06610698e-01 -1.11807930e+00 4.29003388e-01
1.05653107e+00 -9.67426375e-02 5.41164279e-01 9.60160434e-01
3.35736215e-01 -3.90938759e-01 -1.05524182e+00 5.77426791e-01
-8.53117824e-01 -1.52358079e+00 2.21522301e-02 3.02219987e-01
9.87966537e-01 -7.63622299e-02 1.14418931e-01 3.76835823e-01
1.22502935e+00 -9.69502032e-01 1.00251436e+00 7.77924180e-01
5.67082822e-01 -1.11475646e+00 -1.66234076e-02 6.65653646e-01
-6.57553732e-01 -2.71875918e-01 -5.03106177e-01 1.29556298e-01
2.02202231e-01 4.77451682e-01 -9.78609324e-01 1.21374369e-01
6.19084895e-01 4.68072236e-01 2.10598662e-01 1.08356464e+00
-2.43694887e-01 1.18898833e+00 -5.16791046e-01 -5.74903548e-01
6.23553634e-01 -5.91878831e-01 9.01582658e-01 8.02109063e-01
2.11020887e-01 -4.14073914e-01 3.70514899e-01 1.44467878e+00
2.56315351e-01 -2.38268331e-01 -4.73401248e-01 -4.20871228e-01
7.76083946e-01 7.34536469e-01 -9.92488086e-01 -4.59471732e-01
2.66010463e-01 2.76036620e-01 4.71421540e-01 6.97181880e-01
-6.93265676e-01 -1.21615015e-01 6.31955385e-01 4.53761220e-02
8.30600441e-01 -7.31839985e-02 1.18309502e-02 -1.00790679e+00
-5.50542474e-01 -6.16340995e-01 3.09867978e-01 -1.95111737e-01
-9.43806112e-01 2.52113163e-01 8.01046729e-01 -9.19222116e-01
-7.94148207e-01 -2.12049171e-01 -7.78275609e-01 9.03162003e-01
-8.29805315e-01 -4.27238554e-01 5.44563890e-01 3.87594491e-01
8.28029394e-01 -1.86365873e-01 5.88567078e-01 -5.44652998e-01
-5.92341125e-01 1.37607083e-01 6.82913482e-01 -2.20253944e-01
-2.40796469e-02 -1.56127560e+00 1.58894211e-01 5.59696794e-01
1.95283473e-01 6.36918545e-01 9.76053894e-01 -8.14474463e-01
-1.15798831e+00 -8.43283176e-01 1.29612654e-01 -1.87727138e-01
5.21896064e-01 -4.16321188e-01 -7.58170128e-01 8.36029649e-01
3.85904819e-01 -3.97325069e-01 3.33770335e-01 3.06614101e-01
1.95721924e-01 1.08916700e-01 -1.17609811e+00 7.26149440e-01
8.87773693e-01 -1.26758829e-01 -2.25665659e-01 2.36463651e-01
4.47696477e-01 -2.16049120e-01 -5.51493704e-01 -1.66884750e-01
4.75345016e-01 -1.10890210e+00 7.44081736e-01 -7.92285264e-01
7.85738081e-02 1.08505726e-01 1.30941480e-01 -1.45004654e+00
-2.09398016e-01 -1.25733602e+00 -5.68954825e-01 9.07143116e-01
4.02613103e-01 -6.34033382e-01 1.20393467e+00 2.71217376e-01
2.89435178e-01 -1.15108585e+00 -9.77715015e-01 -6.84792519e-01
3.59019130e-01 -3.20593894e-01 6.33714199e-01 3.11920077e-01
-3.25334489e-01 2.41960548e-02 -2.97294229e-01 -2.93965578e-01
1.42474270e+00 1.94500253e-01 7.07764149e-01 -1.16845322e+00
-6.09105408e-01 -4.69131231e-01 3.66581678e-01 -1.19263303e+00
2.44738907e-01 -5.39445519e-01 2.14452267e-01 -1.37547231e+00
5.34875952e-02 -5.31542420e-01 -4.30246443e-03 1.42056420e-01
7.81199932e-02 -5.31486094e-01 5.21593578e-02 2.96939015e-01
-6.15528405e-01 8.77732694e-01 1.24091935e+00 5.45007251e-02
-4.91447419e-01 3.59542489e-01 -4.40357447e-01 6.19258225e-01
6.65705681e-01 -6.18180931e-01 -1.03248131e+00 7.26761669e-02
2.50701696e-01 5.14400184e-01 -1.52722076e-02 -6.01293921e-01
8.39204639e-02 -7.07817554e-01 4.46746677e-01 -9.54303443e-01
3.59791964e-01 -3.63756418e-01 5.40123463e-01 5.49391150e-01
-7.74003088e-01 -4.49212462e-01 -1.82051182e-01 1.04199338e+00
2.59278536e-01 -5.64981937e-01 8.89714241e-01 -5.52829623e-01
-1.98019132e-01 3.31720948e-01 -9.20125902e-01 3.24843377e-01
9.09906507e-01 -2.52776355e-01 1.18121222e-01 -3.66951257e-01
-1.02853346e+00 2.70299017e-01 2.28317410e-01 -2.28991911e-01
4.29531068e-01 -1.19196665e+00 -4.23515201e-01 -4.13085287e-03
-5.58186471e-01 4.92258698e-01 -8.35605785e-02 5.83701074e-01
-3.32930148e-01 4.05677229e-01 3.65124829e-02 -7.59766161e-01
-6.64625943e-01 2.99969733e-01 5.13088107e-01 -5.30579448e-01
-5.49674273e-01 8.04613829e-01 -1.40836254e-01 -1.60198003e-01
3.28582644e-01 -2.68410116e-01 -2.41654031e-02 1.28306523e-01
4.13264096e-01 8.00011337e-01 -4.48370129e-01 1.68856427e-01
1.62048012e-01 -1.19494699e-01 -2.32430190e-01 -6.40785515e-01
1.35040116e+00 -1.17738962e-01 1.36982612e-02 8.19476426e-01
8.44677389e-01 -3.36158782e-01 -1.87981904e+00 -1.83986217e-01
8.91696438e-02 -6.57294869e-01 3.01727861e-01 -3.53983283e-01
-8.07693481e-01 3.99672300e-01 3.51314515e-01 6.38480425e-01
4.26125675e-01 -7.85848498e-02 2.50756294e-01 3.71049136e-01
4.58450824e-01 -1.24068892e+00 -7.59444386e-02 2.20098987e-01
4.35449660e-01 -9.08618689e-01 5.27695231e-02 1.16002105e-01
-5.40191889e-01 9.95556772e-01 4.28400636e-01 -4.39821333e-01
7.42109239e-01 3.09802163e-02 -4.04140770e-01 -3.44473235e-02
-1.15498042e+00 -2.17389137e-01 1.66583821e-01 4.96472895e-01
-4.60572280e-02 1.71763510e-01 -9.45823714e-02 -1.55920789e-01
-1.15071587e-01 2.05791727e-01 1.82336956e-01 9.83031332e-01
-6.56677723e-01 -1.06609249e+00 -4.06347543e-01 7.72453725e-01
-3.94199044e-01 3.58228952e-01 7.08665624e-02 7.59693384e-01
-9.35821980e-02 6.69802725e-01 3.47864509e-01 3.23656231e-01
-2.14350834e-01 2.82085747e-01 7.99633980e-01 -5.10504186e-01
-2.79957265e-01 3.12800974e-01 1.22756362e-01 -5.43575585e-01
-4.23522115e-01 -1.12959063e+00 -7.87021399e-01 -2.42014512e-01
-3.40928286e-01 6.95717037e-01 4.10560191e-01 1.03210127e+00
4.83800694e-02 3.11478436e-01 3.89172822e-01 -9.21741247e-01
-1.15510011e+00 -8.19755673e-01 -8.03854346e-01 -4.30405587e-02
3.36078733e-01 -8.95806849e-01 -4.56138611e-01 -3.85931343e-01] | [4.369206428527832, 2.527585506439209] |
a6bd91be-4fec-476b-9e44-a720c71f09c1 | john-ate-5-apples-john-ate-some-apples-self | 2206.08263 | null | https://arxiv.org/abs/2206.08263v1 | https://arxiv.org/pdf/2206.08263v1.pdf | 'John ate 5 apples' != 'John ate some apples': Self-Supervised Paraphrase Quality Detection for Algebraic Word Problems | This paper introduces the novel task of scoring paraphrases for Algebraic Word Problems (AWP) and presents a self-supervised method for doing so. In the current online pedagogical setting, paraphrasing these problems is helpful for academicians to generate multiple syntactically diverse questions for assessments. It also helps induce variation to ensure that the student has understood the problem instead of just memorizing it or using unfair means to solve it. The current state-of-the-art paraphrase generation models often cannot effectively paraphrase word problems, losing a critical piece of information (such as numbers or units) which renders the question unsolvable. There is a need for paraphrase scoring methods in the context of AWP to enable the training of good paraphrasers. Thus, we propose ParaQD, a self-supervised paraphrase quality detection method using novel data augmentations that can learn latent representations to separate a high-quality paraphrase of an algebraic question from a poor one by a wide margin. Through extensive experimentation, we demonstrate that our method outperforms existing state-of-the-art self-supervised methods by up to 32% while also demonstrating impressive zero-shot performance. | ['Vikram Goyal', 'Mukesh Mohania', 'Venktesh V', 'Rishabh Gupta'] | 2022-06-16 | null | null | null | null | ['paraphrase-generation', 'paraphrase-generation'] | ['computer-code', 'natural-language-processing'] | [ 2.99446702e-01 1.44406080e-01 -3.40554357e-01 -3.57460797e-01
-1.31534457e+00 -9.74476576e-01 2.60903209e-01 5.90901136e-01
-1.38805076e-01 6.53576434e-01 2.10609838e-01 -7.69755840e-01
-7.47025758e-02 -8.80309820e-01 -8.63172293e-01 -2.42460549e-01
6.69809043e-01 4.13150668e-01 1.56832546e-01 -5.15999675e-01
6.01597369e-01 3.18666607e-01 -1.65665317e+00 4.22823995e-01
1.63062239e+00 4.38615859e-01 1.60006151e-01 7.20361173e-01
-4.71728265e-01 1.06449592e+00 -9.61022437e-01 -6.18319631e-01
1.81689397e-01 -7.96091437e-01 -1.22529280e+00 -4.47300486e-02
1.08596957e+00 -2.42037043e-01 -2.22166091e-01 1.04245424e+00
1.96017295e-01 3.46955776e-01 5.48789859e-01 -1.14858866e+00
-1.03475201e+00 6.67284489e-01 -3.86032939e-01 4.24170703e-01
7.05210567e-01 1.86956212e-01 1.43656051e+00 -9.55384433e-01
4.00484115e-01 9.51727033e-01 3.93609405e-01 5.27839422e-01
-1.39402092e+00 -6.93216681e-01 -1.53089866e-01 5.52750528e-01
-9.46844161e-01 -3.49365711e-01 1.01421952e+00 -2.90193975e-01
6.52629435e-01 2.25478619e-01 8.48022163e-01 1.03012395e+00
1.07307591e-01 1.02725565e+00 1.24695420e+00 -6.33444846e-01
3.22630733e-01 2.92675704e-01 5.88596642e-01 8.12025368e-01
1.57819182e-01 -5.15927076e-01 -7.51489878e-01 -1.70497552e-01
5.26783943e-01 -1.54448271e-01 -1.77255526e-01 -5.80709279e-01
-8.39083076e-01 9.77614582e-01 2.48549730e-01 1.17170289e-01
5.54386973e-02 -4.09767240e-01 2.08365321e-01 9.79036510e-01
1.52957663e-01 1.16010284e+00 -2.52211899e-01 -4.91194069e-01
-1.21606171e+00 5.04938424e-01 9.06833827e-01 9.95685756e-01
5.65974474e-01 -8.76419470e-02 -3.02089721e-01 1.10208726e+00
-1.44674987e-01 2.21171245e-01 8.49293590e-01 -1.08488321e+00
6.85004652e-01 1.08600330e+00 8.71004835e-02 -7.14067161e-01
2.91140079e-01 -2.71023393e-01 -3.76129150e-01 3.04255813e-01
6.06716573e-01 1.74121454e-01 -6.85413539e-01 1.68525362e+00
7.75328055e-02 2.77576953e-01 1.79763824e-01 6.68960690e-01
9.01505649e-01 7.45689511e-01 -2.00070620e-01 -3.12643461e-02
1.34104800e+00 -1.29740417e+00 -5.64901650e-01 -3.77499014e-01
6.66251123e-01 -9.33499694e-01 1.68335330e+00 4.86069411e-01
-1.73887444e+00 -6.37352109e-01 -1.19375491e+00 -5.33101678e-01
-1.55618355e-01 -9.27752033e-02 1.31685868e-01 7.12584853e-01
-7.50866234e-01 7.06903815e-01 -2.00114459e-01 8.66028946e-03
2.76187718e-01 -4.14588526e-02 -1.98495403e-01 -3.19298655e-01
-1.15906918e+00 9.30305660e-01 -1.50776893e-01 -5.37368000e-01
-6.81106329e-01 -1.28206694e+00 -1.04522061e+00 5.58986783e-01
3.30150336e-01 -4.60619569e-01 1.42291260e+00 -6.45467639e-01
-1.76878452e+00 1.13672972e+00 -1.67078957e-01 -2.27203086e-01
3.33015710e-01 -2.85835683e-01 1.81399986e-01 2.08130255e-01
3.32875341e-01 3.82750422e-01 8.57056022e-01 -7.47574091e-01
-1.74208701e-01 -1.81099683e-01 3.14927906e-01 5.84387720e-01
-5.02363324e-01 -1.27345234e-01 -7.18235001e-02 -7.56241202e-01
2.53211617e-01 -8.88896704e-01 -1.16717899e-02 -6.92306533e-02
-1.16299488e-01 -6.63080931e-01 4.60068822e-01 -4.84682351e-01
1.24579787e+00 -1.73582971e+00 4.12839532e-01 2.44731605e-02
3.74679267e-01 5.55412352e-01 -4.27984476e-01 3.04784060e-01
-4.21553701e-01 6.88999519e-02 -1.76225454e-02 -3.70866686e-01
-6.02780543e-02 1.93595812e-02 -6.13479137e-01 1.03807747e-01
2.07793131e-01 9.50675845e-01 -1.26045907e+00 -4.57089663e-01
1.53087586e-01 -3.22329432e-01 -6.40071929e-01 7.49379694e-01
-4.46068719e-02 -7.00631589e-02 -1.97284356e-01 4.69196975e-01
4.27310050e-01 -2.51929581e-01 -1.85912415e-01 4.18345958e-01
2.05101684e-01 9.12961483e-01 -1.13373327e+00 1.90892112e+00
-8.66871834e-01 6.19358599e-01 -2.64999568e-01 -1.16104746e+00
1.01073790e+00 -8.64468329e-03 3.95616442e-02 -6.10019863e-01
-1.71379998e-01 1.13480225e-01 -2.57595628e-01 -4.61392701e-01
8.19503367e-01 -4.57706302e-02 -1.80295289e-01 1.00307512e+00
2.26536900e-01 -8.57639611e-01 4.55074489e-01 7.30224490e-01
1.22559869e+00 -1.12294324e-01 1.15631886e-01 -2.91934520e-01
7.33263254e-01 -1.13779716e-02 2.11017147e-01 1.04162169e+00
-2.58952200e-01 6.51258111e-01 5.20241499e-01 -4.10584249e-02
-1.03983498e+00 -1.41017330e+00 1.58844456e-01 1.40700781e+00
9.14793462e-03 -5.85775018e-01 -7.12890506e-01 -7.36844063e-01
1.70134827e-02 9.07555878e-01 -3.24612737e-01 -6.68262601e-01
-5.11965036e-01 6.67414535e-03 3.64957213e-01 4.88059878e-01
3.14534575e-01 -1.02964747e+00 -1.76378787e-01 1.57496795e-01
-4.13426101e-01 -1.04877603e+00 -6.05570436e-01 2.05284189e-02
-7.17623115e-01 -1.07381094e+00 -7.47497320e-01 -1.07307649e+00
8.50240350e-01 7.92148530e-01 1.38156724e+00 2.72509009e-01
-1.34154350e-01 3.14877003e-01 -3.58717263e-01 -2.92481780e-01
-8.34392607e-01 6.72655404e-02 -9.89209637e-02 -5.66295505e-01
6.28564954e-01 -6.01811767e-01 -5.45676351e-01 4.44155261e-02
-8.44393313e-01 1.61094606e-01 3.83759230e-01 8.73741388e-01
1.72682256e-01 -3.25709999e-01 7.32705176e-01 -8.39871764e-01
1.31660962e+00 -4.17023391e-01 -4.61737394e-01 5.19854009e-01
-6.11387670e-01 2.47921988e-01 1.08033347e+00 -5.95237732e-01
-7.14994252e-01 -4.74716038e-01 -3.44712883e-02 -5.50276220e-01
-3.27408388e-02 3.36870730e-01 2.37193540e-01 -2.03154445e-01
9.97922421e-01 4.68893826e-01 1.81307927e-01 -2.43987009e-01
4.43722904e-01 6.96205139e-01 4.50616628e-01 -7.87113547e-01
1.04789484e+00 -2.25251555e-01 -2.01788440e-01 -6.47995472e-01
-1.55104566e+00 -8.12012613e-01 -4.28926557e-01 -1.24637567e-01
2.99563259e-01 -9.74423528e-01 -4.89473879e-01 2.88340092e-01
-1.05554569e+00 -1.61022127e-01 -6.40250623e-01 -6.90080523e-02
-4.90538985e-01 7.23464191e-01 -8.42049420e-01 -4.06296343e-01
-1.63787410e-01 -1.37392068e+00 7.86556602e-01 4.59674686e-01
-6.56653225e-01 -7.68634856e-01 3.62872630e-01 1.32562101e+00
4.24079329e-01 -4.43601191e-01 1.20877635e+00 -8.90144110e-01
-5.19836903e-01 -1.51773572e-01 -2.67747939e-02 5.26118457e-01
1.28340036e-01 -1.72073439e-01 -7.23288238e-01 -1.50673822e-01
2.11574703e-01 -1.07912803e+00 6.39832616e-01 -1.98688000e-01
1.24347174e+00 -4.93011981e-01 4.30306166e-01 3.55330259e-01
8.38433981e-01 -4.91639555e-01 4.62592393e-01 3.53153050e-01
5.88944972e-01 5.95063090e-01 3.91373545e-01 1.21596195e-01
4.18147594e-01 5.46623826e-01 -7.85882622e-02 1.32633790e-01
-1.64502203e-01 -5.19990444e-01 4.60805535e-01 1.25648403e+00
5.51571727e-01 6.86486512e-02 -6.68724537e-01 8.95472467e-01
-1.60333228e+00 -9.38735723e-01 -1.51525423e-01 2.27999687e+00
1.51648211e+00 9.86999571e-02 8.59912187e-02 3.25540513e-01
4.55733448e-01 1.48131371e-01 -3.68981779e-01 -9.40183401e-01
3.12864661e-01 8.37458253e-01 -8.96175858e-03 6.00943267e-01
-5.40057421e-01 1.19814360e+00 5.97511625e+00 9.52485144e-01
-6.43420756e-01 -2.53772885e-01 5.55514872e-01 -1.33431684e-02
-5.66274464e-01 -4.72418368e-02 -7.03396618e-01 4.69003111e-01
8.85320425e-01 -5.49910903e-01 5.04166424e-01 1.09645081e+00
1.69176951e-01 -9.45176929e-02 -1.10489011e+00 8.54787111e-01
4.79591370e-01 -1.23650289e+00 4.71831411e-01 -4.59172845e-01
1.10526240e+00 -5.88702619e-01 1.23638026e-01 6.63035929e-01
6.11523628e-01 -1.08553421e+00 2.65309781e-01 7.09562451e-02
4.00160879e-01 -7.98794746e-01 4.59408492e-01 4.27003771e-01
-7.81603098e-01 -1.17439434e-01 -6.48842037e-01 -3.48397166e-01
-2.89983928e-01 5.25710583e-01 -9.03110862e-01 2.82690018e-01
2.36073434e-01 3.88947845e-01 -9.78695989e-01 9.77305293e-01
-7.46059239e-01 8.68446112e-01 6.71867728e-02 -3.20783317e-01
3.27582777e-01 -3.73178929e-01 4.36115921e-01 7.72469401e-01
2.68527359e-01 2.71030039e-01 6.12678230e-02 1.17826200e+00
-4.22630310e-01 3.80305856e-01 -5.79899848e-01 -3.98600101e-02
6.67466462e-01 1.14736927e+00 -1.54767394e-01 -4.84646410e-01
-2.72418171e-01 1.17284536e+00 7.51241565e-01 1.60905510e-01
-5.22206187e-01 -6.32166564e-01 6.20342195e-01 1.97947547e-02
-5.78581952e-02 -1.87873498e-01 -6.05431795e-01 -1.59459674e+00
1.26837760e-01 -1.27753127e+00 2.15166152e-01 -1.03194046e+00
-1.55625665e+00 6.60481900e-02 -2.76924431e-01 -1.15909004e+00
-3.34211439e-01 -4.56023484e-01 -1.21538329e+00 1.03263307e+00
-1.53659213e+00 -6.74485743e-01 -5.84874153e-01 4.52175021e-01
1.01929116e+00 -3.74833286e-01 7.50355601e-01 -2.78690338e-01
-4.33946073e-01 9.73277092e-01 6.71986043e-02 -4.50654101e-04
1.08779788e+00 -1.62410736e+00 4.77842152e-01 9.18170691e-01
3.97605240e-01 9.20352876e-01 8.94744635e-01 -3.87893140e-01
-1.33404315e+00 -6.62163019e-01 1.26935697e+00 -5.28506398e-01
9.76604819e-01 -2.34995306e-01 -1.38801014e+00 1.99730635e-01
3.32699835e-01 -4.08272654e-01 1.00948477e+00 1.94468677e-01
-5.30236483e-01 2.42544021e-02 -1.02160168e+00 9.68410671e-01
7.06004918e-01 -8.35185051e-01 -1.56691337e+00 6.92481220e-01
6.22596502e-01 -4.42091763e-01 -6.56259239e-01 -4.23377380e-02
2.11845741e-01 -9.12605762e-01 1.03538620e+00 -1.02919197e+00
1.15248728e+00 2.76459336e-01 5.30798972e-01 -1.68977404e+00
-2.08498657e-01 -7.44619012e-01 -2.32412875e-01 1.10822904e+00
1.75660476e-01 -1.25505581e-01 1.29333258e+00 8.26004922e-01
-6.96278214e-02 -7.71852612e-01 -7.76644707e-01 -7.48625517e-01
5.49022615e-01 -1.09887972e-01 3.18983972e-01 1.32246435e+00
7.34045208e-01 7.30137408e-01 -2.22141240e-02 -2.16406837e-01
6.48356318e-01 4.90038276e-01 9.83966112e-01 -1.00145435e+00
-1.53867781e-01 -6.44971371e-01 -2.21091345e-01 -1.48032081e+00
6.71157718e-01 -1.07491601e+00 -1.38926700e-01 -1.30847871e+00
2.59429753e-01 -3.08647335e-01 -1.64112195e-01 3.30474943e-01
-7.31000602e-01 -1.45256272e-04 1.26769111e-01 -9.04170498e-02
-6.08900845e-01 4.47062790e-01 1.31084931e+00 -1.67656958e-01
-2.21422628e-01 2.09121734e-01 -1.02344167e+00 4.98793632e-01
8.39000285e-01 -3.74655008e-01 -6.74880862e-01 -2.44395554e-01
2.34832764e-01 2.37182468e-01 2.14955285e-01 -8.65558922e-01
4.61456120e-01 -3.59602630e-01 9.00418982e-02 -2.61499822e-01
8.84354189e-02 -4.87499624e-01 -9.31993127e-01 1.82063818e-01
-9.72423494e-01 2.67477393e-01 1.23395875e-01 2.72500396e-01
-2.26090178e-01 -1.01928329e+00 8.51862311e-01 -2.99824357e-01
-2.44632125e-01 -2.24164918e-01 -5.02243757e-01 6.97297275e-01
8.99265647e-01 -1.52336940e-01 -4.45021600e-01 -6.70525849e-01
-3.99023294e-01 4.81818736e-01 4.32220995e-01 5.89633644e-01
7.90955484e-01 -1.27126920e+00 -7.11825252e-01 3.65824848e-01
2.69949377e-01 2.65761055e-02 1.81488082e-01 5.00338967e-04
-5.26999116e-01 4.64740932e-01 -3.30207080e-01 -2.74304658e-01
-1.62582517e+00 3.64535540e-01 -6.33496195e-02 -7.36884534e-01
-3.55583251e-01 1.00711322e+00 -1.43988684e-01 -5.89697421e-01
4.13854003e-01 -4.55199391e-01 -2.30531543e-01 8.31612796e-02
6.44342542e-01 1.89842865e-01 1.94627732e-01 -8.33329931e-03
3.05390656e-01 3.47279161e-01 -4.80534285e-01 6.66324869e-02
1.17764127e+00 7.49400258e-02 2.63605535e-01 3.21794748e-01
1.07885230e+00 2.32506335e-01 -6.77819312e-01 -3.56633037e-01
-8.66336972e-02 -6.08201623e-01 -2.24854052e-01 -7.84987330e-01
-4.32621211e-01 9.29595828e-01 -7.29812607e-02 2.11399153e-01
7.79472470e-01 -1.89365327e-01 1.05139053e+00 8.17489505e-01
1.30482629e-01 -1.03237367e+00 7.98620760e-01 7.35014856e-01
8.47461760e-01 -1.36738360e+00 1.39424652e-01 -3.97019506e-01
-7.81697512e-01 1.15107906e+00 9.77185428e-01 -3.17368805e-01
-9.24831778e-02 -1.36127502e-01 3.56494784e-02 -4.51511936e-03
-8.94483984e-01 2.28829116e-01 5.22694111e-01 3.65347229e-02
4.35345232e-01 -2.70845383e-01 -2.12674260e-01 5.29652834e-01
-6.37619197e-01 -1.86879143e-01 1.17168725e+00 1.04624093e+00
-6.97243631e-01 -1.31462669e+00 -3.05372179e-01 5.30382991e-01
-3.08695048e-01 -3.85967761e-01 -7.58203030e-01 2.99438924e-01
-6.16387248e-01 9.21932399e-01 -1.13216415e-01 -8.37751776e-02
3.86236399e-01 5.73486090e-01 6.50359929e-01 -1.08585691e+00
-9.51402545e-01 -7.31077433e-01 -1.59198210e-01 -2.82897413e-01
8.47736299e-02 -3.95574838e-01 -9.19347167e-01 -3.91463995e-01
-3.49999607e-01 4.95895833e-01 3.44620198e-01 1.13645673e+00
2.03409448e-01 3.36654395e-01 9.29902852e-01 -2.93105066e-01
-1.30507827e+00 -8.91700685e-01 -1.81164026e-01 9.00982797e-01
2.18084902e-01 -3.87910992e-01 -6.94857478e-01 -1.15661837e-01] | [11.532612800598145, 9.201079368591309] |
c0e78fe1-1a5a-4d13-9979-132b42b1e50e | patchnet-short-range-template-matching-for | 2103.07371 | null | https://arxiv.org/abs/2103.07371v1 | https://arxiv.org/pdf/2103.07371v1.pdf | PatchNet -- Short-range Template Matching for Efficient Video Processing | Object recognition is a fundamental problem in many video processing tasks, accurately locating seen objects at low computation cost paves the way for on-device video recognition. We propose PatchNet, an efficient convolutional neural network to match objects in adjacent video frames. It learns the patchwise correlation features instead of pixel features. PatchNet is very compact, running at just 58MFLOPs, $5\times$ simpler than MobileNetV2. We demonstrate its application on two tasks, video object detection and visual object tracking. On ImageNet VID, PatchNet reduces the flops of R-FCN ResNet-101 by 5x and EfficientDet-D0 by 3.4x with less than 1% mAP loss. On OTB2015, PatchNet reduces SiamFC and SiamRPN by 2.5x with no accuracy loss. Experiments on Jetson Nano further demonstrate 2.8x to 4.3x speed-ups associated with flops reduction. Code is open sourced at https://github.com/RalphMao/PatchNet. | ['William J. Dally', 'Song Han', 'Sibo Zhu', 'Huizi Mao'] | 2021-03-10 | null | null | null | null | ['template-matching'] | ['computer-vision'] | [-2.58497715e-01 -5.02178609e-01 -2.18156680e-01 -1.63006693e-01
-6.04676545e-01 -5.70742548e-01 1.81159288e-01 -1.95033476e-01
-6.42062128e-01 5.19139707e-01 -3.50100398e-01 -4.32461590e-01
1.02348775e-01 -4.68142748e-01 -1.15761507e+00 -3.00591409e-01
-2.78751999e-01 -1.53439835e-01 5.64912200e-01 1.59650937e-01
1.21405780e-01 6.67278171e-01 -1.39269996e+00 3.30239087e-01
2.16463834e-01 1.64439535e+00 2.53543198e-01 1.10874677e+00
3.28616291e-01 1.13368905e+00 -3.81612837e-01 -1.87120810e-01
6.02266312e-01 1.46956608e-01 -5.55437326e-01 -9.14444774e-02
1.08597219e+00 -6.10330284e-01 -1.03988194e+00 1.10001075e+00
4.50967133e-01 -3.05157248e-02 1.98438168e-01 -1.24350548e+00
-3.52657050e-01 2.91141063e-01 -5.83784521e-01 8.89271080e-01
-1.10235684e-01 3.66278559e-01 7.29718268e-01 -1.18780017e+00
3.64827603e-01 1.02181220e+00 9.08317089e-01 3.15090775e-01
-8.06731761e-01 -8.42767000e-01 -2.09339485e-01 4.45502609e-01
-1.65411294e+00 -6.36546850e-01 7.70725384e-02 -2.76725531e-01
1.16604829e+00 2.78971016e-01 6.06645942e-01 5.11954248e-01
2.91336209e-01 7.46216118e-01 6.15301430e-01 2.16698814e-02
-7.02131689e-02 -2.50240862e-01 1.87213212e-01 1.05036283e+00
4.52728003e-01 1.41937584e-02 -4.99583453e-01 1.09046929e-01
1.13993049e+00 4.83817816e-01 -3.37838203e-01 6.01464771e-02
-1.01964760e+00 4.20709282e-01 6.45738363e-01 2.88570225e-02
-1.85515076e-01 1.02167559e+00 5.12882471e-01 4.14722621e-01
2.22880945e-01 6.93966001e-02 -4.54314113e-01 -3.86881083e-01
-9.06372488e-01 1.84130564e-01 3.72941673e-01 1.39518404e+00
6.71831548e-01 2.67563611e-01 3.95869128e-02 5.94118416e-01
2.39193588e-01 9.50189829e-01 2.04901963e-01 -1.23767281e+00
4.80953604e-01 3.65651757e-01 5.95211349e-02 -1.17273772e+00
-3.43029261e-01 -3.20093662e-01 -7.80492127e-01 2.64085114e-01
4.27862108e-01 -2.85061061e-01 -1.03816915e+00 1.16091013e+00
8.64169523e-02 4.63048011e-01 -3.10882509e-01 1.07560742e+00
1.03514099e+00 6.68750644e-01 -1.81866899e-01 3.17524910e-01
1.50335252e+00 -1.15475714e+00 -2.79566377e-01 -2.11658329e-01
5.74522674e-01 -9.33760881e-01 6.32307231e-01 2.71923900e-01
-1.33102214e+00 -8.21522951e-01 -1.20186806e+00 -2.23555952e-01
-9.27174464e-02 4.00746822e-01 6.39168799e-01 5.24353623e-01
-1.25014460e+00 6.69264436e-01 -1.06546998e+00 -1.47699609e-01
9.32093740e-01 8.14976573e-01 -1.27564058e-01 -1.56011537e-01
-6.93010926e-01 3.36462647e-01 6.89657182e-02 -5.14548086e-03
-1.06409895e+00 -1.18453741e+00 -6.48967445e-01 3.36795539e-01
4.35038805e-01 -4.17922944e-01 1.30061805e+00 -8.62470448e-01
-1.05406511e+00 5.57467639e-01 -5.01128547e-02 -8.90298426e-01
3.92543435e-01 -4.22663480e-01 -5.71947873e-01 3.57449204e-01
1.42275304e-01 9.01892722e-01 7.90626943e-01 -4.13790852e-01
-9.41080153e-01 -1.38032451e-01 9.09313112e-02 2.51490995e-03
-2.03648105e-01 2.46288866e-01 -1.06975520e+00 -6.20879173e-01
-8.76225680e-02 -1.13097811e+00 -2.25977272e-01 5.79817772e-01
-5.84944002e-02 -3.39591429e-02 1.13440430e+00 -4.43424463e-01
6.16922438e-01 -2.27173781e+00 -6.29025698e-01 -2.64950804e-02
7.17633903e-01 6.77293122e-01 -1.66955382e-01 -2.44667679e-01
1.52290046e-01 -5.06491028e-02 2.96179205e-01 -6.94087446e-02
-2.67125726e-01 -2.54392117e-01 -1.43024907e-01 7.47573435e-01
7.84834251e-02 1.21927404e+00 -5.66993654e-01 -1.55103669e-01
3.37423027e-01 7.12293625e-01 -6.76083922e-01 -2.26556957e-01
1.00725882e-01 -5.12609743e-02 -1.70253277e-01 8.64909291e-01
8.63367498e-01 -7.09478855e-01 1.17091529e-01 -5.05824089e-01
-1.88179776e-01 2.36823902e-01 -1.00518596e+00 1.57386339e+00
-3.22824836e-01 1.47651327e+00 2.35299692e-01 -8.66763234e-01
6.41020715e-01 1.19802672e-02 5.98007619e-01 -7.80964911e-01
4.35915917e-01 1.46905914e-01 -3.80545296e-02 -2.39634544e-01
6.07480824e-01 5.12245238e-01 7.15953335e-02 1.57493904e-01
1.46352276e-01 5.41808963e-01 1.50617555e-01 1.45471215e-01
1.42674077e+00 -2.73164898e-01 1.16167152e-02 -4.97693300e-01
2.20546246e-01 1.10563394e-02 5.03295124e-01 8.65266800e-01
-2.96258509e-01 5.38280845e-01 1.33239165e-01 -9.58431184e-01
-9.84506547e-01 -1.02434599e+00 -8.81599411e-02 9.66226518e-01
5.58404982e-01 -6.43598676e-01 -5.44626236e-01 -5.67603469e-01
1.43058851e-01 -2.36843541e-01 -2.54864037e-01 1.93238288e-01
-8.43307793e-01 -5.09635687e-01 7.05577314e-01 9.32764411e-01
8.36559653e-01 -7.89303482e-01 -9.50897336e-01 4.52946611e-02
2.15193033e-01 -1.35019934e+00 -8.58262420e-01 -8.97961557e-02
-8.66064250e-01 -1.24803543e+00 -8.12080979e-01 -1.02864575e+00
6.26681209e-01 7.17509508e-01 1.26500583e+00 3.84701967e-01
-6.89932346e-01 2.87717015e-01 -1.31927341e-01 -1.88414395e-01
2.05534026e-01 1.61485933e-02 1.64044052e-01 -1.78069487e-01
5.43702424e-01 -3.31403255e-01 -1.07564473e+00 5.62140226e-01
-5.28825104e-01 -2.02642322e-01 4.88986522e-01 4.23430532e-01
5.95161974e-01 -1.47472486e-01 9.86151993e-02 -5.00527740e-01
-9.83483940e-02 -3.02251786e-01 -1.02128041e+00 -1.17203787e-01
-2.77524173e-01 -3.77101868e-01 4.15252954e-01 -4.88402694e-01
-4.07429188e-01 1.26678646e-01 -6.19693287e-03 -1.05345547e+00
6.47791401e-02 -6.17961250e-02 2.76646644e-01 -5.47843277e-01
4.81796354e-01 1.32851124e-01 1.68663874e-01 -2.07306460e-01
1.18426599e-01 4.47475880e-01 5.06938219e-01 -8.72031748e-02
5.05826235e-01 7.17423499e-01 -1.24289058e-01 -9.78602111e-01
-3.42153937e-01 -4.60879356e-01 -7.69728720e-02 -3.20804149e-01
8.09436500e-01 -1.44052851e+00 -1.31853509e+00 3.28931987e-01
-1.03686953e+00 -4.95632410e-01 -1.21021196e-01 6.25222862e-01
-3.17331791e-01 7.10753947e-02 -8.02015007e-01 -2.01062039e-01
-6.90156400e-01 -1.23698163e+00 8.70252848e-01 3.45712095e-01
4.74079400e-02 -5.05097210e-01 -6.28962040e-01 2.03817099e-01
6.70974731e-01 -1.97592914e-01 1.48451328e-01 -1.80995256e-01
-1.32143259e+00 -1.51482731e-01 -8.50133657e-01 3.86238366e-01
-7.24419951e-02 -1.90852463e-01 -6.33503318e-01 -6.06150746e-01
-1.99226245e-01 -7.06539303e-02 1.01748657e+00 7.11690426e-01
1.39325058e+00 -3.62628847e-01 -5.08835435e-01 9.44497943e-01
1.60520291e+00 4.45910364e-01 6.96547270e-01 4.87517267e-02
8.95864189e-01 -2.58712679e-01 5.61231792e-01 3.75403881e-01
1.49837673e-01 7.97300279e-01 4.39495116e-01 -1.68567419e-01
-6.49745584e-01 1.78029463e-01 4.28520262e-01 6.77399158e-01
-9.72676426e-02 -2.13700607e-01 -9.94848967e-01 5.14516473e-01
-1.80468249e+00 -8.33408654e-01 -1.24223486e-01 1.84939516e+00
1.76821843e-01 3.07037950e-01 1.41775653e-01 -2.79558361e-01
8.57301235e-01 8.39369148e-02 -5.92064083e-01 6.26343116e-02
4.72716838e-02 2.48856410e-01 1.05756450e+00 4.09040034e-01
-1.33683538e+00 9.19436693e-01 5.25896788e+00 9.71057236e-01
-1.35824585e+00 2.26033345e-01 8.79670262e-01 -6.23745620e-01
5.56867599e-01 -3.86693805e-01 -1.07717478e+00 6.35368824e-01
9.15967226e-01 1.56921253e-01 3.93050194e-01 1.13986754e+00
-6.30896911e-02 1.96492430e-02 -1.03944778e+00 1.66475308e+00
1.37821510e-01 -2.01357913e+00 -2.44321346e-01 1.77947953e-01
6.87818527e-01 7.09055185e-01 2.31471121e-01 1.57892317e-01
-2.63341377e-03 -9.66214299e-01 7.03581393e-01 2.25376680e-01
1.12731636e+00 -8.34619284e-01 6.58752263e-01 -2.30087727e-01
-1.54174328e+00 -3.43533978e-02 -6.81920230e-01 -2.39011012e-02
6.00937642e-02 2.98625559e-01 -8.56985569e-01 -2.16732740e-01
1.40107810e+00 8.42370808e-01 -4.59690005e-01 1.23918355e+00
3.31151396e-01 5.30582845e-01 -4.28761870e-01 -1.59908265e-01
2.43919611e-01 1.78600073e-01 4.15276676e-01 1.32378399e+00
5.34265041e-01 2.63353497e-01 6.69240057e-02 4.77942795e-01
-6.29764736e-01 -1.97060108e-01 -3.57125193e-01 1.54666111e-01
5.02456129e-01 1.23053372e+00 -9.83575225e-01 -5.88141799e-01
-4.37895656e-01 1.14017999e+00 2.18426622e-02 1.21809557e-01
-1.28631890e+00 -6.85089111e-01 1.08017695e+00 4.01554406e-01
1.05243850e+00 -3.57602656e-01 1.64360162e-02 -9.44297731e-01
1.56767800e-01 -7.57957399e-01 -2.25991178e-02 -5.76507032e-01
-8.31856191e-01 7.06400990e-01 -4.79936391e-01 -1.39069402e+00
3.07711750e-01 -9.88480031e-01 -3.31216067e-01 4.08535689e-01
-1.24903774e+00 -5.46167731e-01 -5.89011133e-01 5.49591780e-01
7.76467264e-01 -3.54036212e-01 2.20371038e-01 8.58354509e-01
-4.43078011e-01 7.85589516e-01 2.29271591e-01 6.10483348e-01
3.88212711e-01 -7.03950047e-01 9.02852774e-01 7.38979638e-01
1.74348399e-01 5.16706645e-01 3.14656436e-01 -4.92810041e-01
-1.63544858e+00 -1.50670362e+00 4.67474371e-01 -3.13601017e-01
6.30488515e-01 -3.85758311e-01 -6.97369218e-01 6.45228207e-01
2.45256618e-01 8.44408810e-01 3.23731601e-01 -3.42325717e-01
-4.14671719e-01 -4.72860247e-01 -9.69604850e-01 5.19776702e-01
1.24946988e+00 -4.65910912e-01 2.29199380e-01 4.93853033e-01
7.24238276e-01 -7.10547626e-01 -8.26293468e-01 3.87843579e-01
6.12471282e-01 -8.63076270e-01 1.16132057e+00 -3.57073188e-01
1.04006827e-01 -5.88006258e-01 -3.31220150e-01 -5.19090891e-01
-4.28618133e-01 -6.67328656e-01 -3.62701476e-01 6.01924479e-01
2.86727637e-01 -3.39466184e-01 1.17067242e+00 3.41610372e-01
-1.62479058e-01 -8.31595302e-01 -9.66918886e-01 -1.02962506e+00
-4.11472827e-01 -6.64531946e-01 3.52826774e-01 5.09693980e-01
-4.90371943e-01 1.69951245e-01 -4.12910521e-01 3.51049960e-01
5.84265351e-01 -1.30068690e-01 7.32790172e-01 -7.33416021e-01
-3.23392212e-01 -4.88508105e-01 -9.72570002e-01 -1.68650842e+00
-2.29163542e-01 -6.22089744e-01 -1.23853080e-01 -9.48109388e-01
7.40808696e-02 -5.53964317e-01 -3.89489770e-01 3.97579700e-01
2.89598614e-01 1.05125856e+00 5.72812438e-01 1.02190755e-01
-1.18737674e+00 1.14611514e-01 7.02174425e-01 -2.85789281e-01
1.74251609e-02 -1.99032724e-01 -3.25220972e-01 7.59748101e-01
1.10076165e+00 -4.51752692e-01 -2.78480351e-01 -8.22511196e-01
-7.87276030e-02 2.11873576e-02 5.70963502e-01 -1.67196131e+00
4.35080379e-01 3.47007543e-01 6.82549894e-01 -7.40775287e-01
5.94833374e-01 -7.28192389e-01 1.67524427e-01 6.87437356e-01
-2.33635716e-02 4.14596200e-01 6.74610555e-01 4.24974799e-01
2.05981955e-02 7.39020184e-02 7.37752140e-01 -1.86633728e-02
-1.04857659e+00 6.77585483e-01 -4.43182200e-01 -8.16590618e-04
1.12232876e+00 -2.27634460e-01 -4.20772165e-01 -2.53781885e-01
-4.20471162e-01 6.88947439e-02 2.07627341e-01 4.32778925e-01
9.22553062e-01 -1.43857276e+00 -4.61100876e-01 1.49615645e-01
-3.82084399e-01 -2.29961425e-01 6.21175408e-01 8.13553870e-01
-1.02612031e+00 8.10443938e-01 -2.39540577e-01 -9.11766052e-01
-1.76684463e+00 3.86947989e-01 4.81165469e-01 3.79273087e-01
-6.68471336e-01 1.21284580e+00 2.16331571e-01 2.49536201e-01
3.16794813e-01 -4.88970101e-01 1.91722602e-01 -4.31320995e-01
9.70992744e-01 6.69236898e-01 1.68404683e-01 -5.15580297e-01
-6.14728153e-01 5.56983829e-01 -3.49096566e-01 2.78896123e-01
1.13301969e+00 1.72386635e-02 1.04682155e-01 -1.20625794e-01
1.52848876e+00 -2.09060758e-01 -1.65961432e+00 -2.17138708e-01
-2.16096207e-01 -5.52961111e-01 2.73068696e-01 -3.67504269e-01
-1.51238894e+00 6.62444651e-01 1.10878205e+00 -7.21056685e-02
1.04588294e+00 1.96572587e-01 1.01683509e+00 5.08788466e-01
3.89796793e-01 -8.38078320e-01 2.28798404e-01 5.22847652e-01
6.72718704e-01 -1.31312799e+00 1.25893265e-01 -4.46304053e-01
-2.57346839e-01 9.99615073e-01 7.99767077e-01 -5.11646211e-01
8.85923624e-01 4.88344252e-01 4.65521067e-02 -3.05785149e-01
-6.20791435e-01 -1.05320802e-02 3.25730026e-01 1.33069813e-01
1.80863559e-01 1.05972819e-01 3.58429939e-01 1.72263846e-01
5.91243953e-02 -2.49227583e-02 3.46844852e-01 9.29036498e-01
-4.38061923e-01 -5.54556727e-01 -1.86618179e-01 6.18029535e-01
-7.31213391e-01 -3.44169408e-01 2.98536777e-01 7.88366914e-01
1.65258631e-01 6.95616484e-01 7.21315980e-01 -6.57039404e-01
1.80368677e-01 -6.89644277e-01 3.42344671e-01 -3.10687125e-01
-4.08662915e-01 4.85785343e-02 -1.01936854e-01 -9.60068047e-01
-3.71584207e-01 -5.14581859e-01 -1.22431338e+00 -6.73124731e-01
-1.19084843e-01 -1.57564521e-01 6.13333344e-01 3.88245255e-01
9.43125069e-01 4.30245519e-01 3.52233529e-01 -9.58650351e-01
-3.52660976e-02 -5.39398551e-01 -2.32146308e-01 -5.22068702e-02
5.64701617e-01 -3.68224859e-01 -1.84859838e-02 5.26168235e-02] | [8.836750030517578, -0.15914317965507507] |
544110fb-68cb-4963-b16d-92b9953f08f4 | diverse-few-shot-text-classification-with | 1805.07513 | null | http://arxiv.org/abs/1805.07513v1 | http://arxiv.org/pdf/1805.07513v1.pdf | Diverse Few-Shot Text Classification with Multiple Metrics | We study few-shot learning in natural language domains. Compared to many
existing works that apply either metric-based or optimization-based
meta-learning to image domain with low inter-task variance, we consider a more
realistic setting, where tasks are diverse. However, it imposes tremendous
difficulties to existing state-of-the-art metric-based algorithms since a
single metric is insufficient to capture complex task variations in natural
language domain. To alleviate the problem, we propose an adaptive metric
learning approach that automatically determines the best weighted combination
from a set of metrics obtained from meta-training tasks for a newly seen
few-shot task. Extensive quantitative evaluations on real-world sentiment
analysis and dialog intent classification datasets demonstrate that the
proposed method performs favorably against state-of-the-art few shot learning
algorithms in terms of predictive accuracy. We make our code and data available
for further study. | ['Jin-Feng Yi', 'Bo-Wen Zhou', 'Shiyu Chang', 'Xiaoxiao Guo', 'Yu Cheng', 'Saloni Potdar', 'Mo Yu', 'Haoyu Wang', 'Gerald Tesauro'] | 2018-05-19 | diverse-few-shot-text-classification-with-1 | https://aclanthology.org/N18-1109 | https://aclanthology.org/N18-1109.pdf | naacl-2018-6 | ['few-shot-text-classification'] | ['natural-language-processing'] | [ 3.51061702e-01 -5.33935010e-01 -4.17523414e-01 -6.06657863e-01
-1.00719440e+00 -7.18421936e-02 9.03679073e-01 6.85843155e-02
-7.96642661e-01 6.00802183e-01 3.35068017e-01 1.80363730e-01
-2.33009517e-01 -5.69228411e-01 -1.61138475e-01 -5.60699463e-01
2.18403444e-01 6.00842655e-01 5.46778917e-01 -4.81977820e-01
7.66365826e-01 -3.22316676e-01 -1.70588720e+00 3.47626746e-01
8.96099210e-01 8.65943134e-01 5.15301406e-01 4.39657360e-01
-4.28878218e-01 1.02950597e+00 -5.51229239e-01 -4.43449557e-01
2.11586386e-01 -6.80970550e-01 -6.92347288e-01 1.54208794e-01
4.30092365e-01 -4.06503111e-01 -1.95676401e-01 1.17687082e+00
7.03097463e-01 5.69785297e-01 9.11875784e-01 -1.29385066e+00
-7.93238461e-01 2.82558173e-01 -5.33425093e-01 5.37007987e-01
3.57473671e-01 2.05374405e-01 1.14980161e+00 -9.86262977e-01
7.67550349e-01 1.13780880e+00 4.11638588e-01 7.90502489e-01
-1.15185499e+00 -5.42561173e-01 5.69196492e-02 7.21851289e-01
-1.06024992e+00 -4.92542952e-01 9.56723750e-01 -5.57345808e-01
1.12811947e+00 -3.43561888e-01 2.40468770e-01 1.06736779e+00
3.36297929e-01 8.16356301e-01 1.18112743e+00 -5.49046397e-01
8.10362577e-01 1.11664951e-01 4.19302791e-01 7.62643337e-01
-8.68446231e-02 -1.23689607e-01 -7.76265442e-01 -3.45377415e-01
2.57185847e-02 3.32179397e-01 3.65259149e-03 -8.62772405e-01
-1.19148672e+00 1.21750700e+00 -1.32650100e-02 4.61314052e-01
-1.77819952e-01 -9.98514593e-02 6.23297155e-01 6.20953202e-01
6.41940236e-01 5.52271068e-01 -3.97271395e-01 -4.53903764e-01
-1.03526616e+00 3.16061109e-01 7.77040660e-01 8.69594634e-01
1.05149019e+00 -1.44638956e-01 -3.98458928e-01 1.27853155e+00
7.25297257e-02 1.50815025e-01 1.11248338e+00 -9.09956932e-01
4.16374564e-01 5.88849425e-01 1.01866178e-01 -7.40901887e-01
-2.24022165e-01 2.78311223e-01 -3.89210403e-01 3.00644219e-01
2.65280753e-01 -1.78906366e-01 -8.81871283e-01 1.55867600e+00
2.41945907e-01 4.08271372e-01 -3.02668158e-02 9.16023314e-01
8.07482064e-01 5.03953040e-01 3.18250954e-02 -3.87344271e-01
1.20106244e+00 -1.24426687e+00 -8.09103787e-01 -3.40530217e-01
8.41627836e-01 -7.07993567e-01 1.48628080e+00 2.20282391e-01
-6.39596581e-01 -6.79125488e-01 -1.34853041e+00 1.34862080e-01
-5.02367258e-01 -4.71040606e-01 4.66827899e-01 7.34797359e-01
-6.29619360e-01 5.93470991e-01 -5.28424919e-01 -6.68330312e-01
2.92677104e-01 -1.06607392e-01 -8.08253959e-02 -4.04329330e-01
-1.19736648e+00 1.20100331e+00 2.19519451e-01 -8.62630963e-01
-1.20790017e+00 -8.02665114e-01 -1.07466590e+00 -2.34376580e-01
6.67046010e-01 -5.35035789e-01 1.70216691e+00 -6.04325473e-01
-1.49387014e+00 8.95847082e-01 -1.58465073e-01 -5.24298429e-01
3.78238976e-01 -1.65325269e-01 -2.03595415e-01 -1.15247063e-01
1.05345875e-01 4.91244793e-01 9.54028428e-01 -8.78163815e-01
-7.04003513e-01 -3.01838517e-01 3.15912008e-01 4.04519677e-01
-8.43134344e-01 3.36367935e-02 -2.30230868e-01 -4.73267317e-01
-4.50389981e-01 -7.76861906e-01 -3.76754463e-01 1.51045963e-01
2.59314388e-01 -3.53467524e-01 9.01633084e-01 4.17003259e-02
1.37182271e+00 -1.94307649e+00 1.53206766e-01 -6.00774169e-01
1.03196748e-01 4.27794039e-01 -3.20911497e-01 3.77761483e-01
2.88450420e-01 -2.03294620e-01 -4.37112004e-01 -5.51731169e-01
9.63048935e-02 1.52763724e-01 -1.09855188e-02 4.54926997e-01
-1.14522442e-01 6.55472338e-01 -1.27500153e+00 -8.68127227e-01
6.14262462e-01 3.22234631e-02 -2.64208883e-01 4.19111401e-01
-2.34919071e-01 -1.24146722e-01 -5.62498391e-01 6.56445920e-01
3.65882367e-01 -1.11138828e-01 -1.14420839e-01 6.47504255e-02
1.93822712e-01 -1.74213329e-03 -9.92121339e-01 2.32510662e+00
-9.68372464e-01 5.34629941e-01 -4.38912690e-01 -1.23742962e+00
9.97968256e-01 2.09672600e-01 7.41520286e-01 -9.16078389e-01
2.33020306e-01 -3.57382037e-02 2.04610631e-01 -7.46481121e-01
4.50836301e-01 -5.16938388e-01 -2.55157709e-01 7.87622333e-01
6.44951463e-01 -3.39093417e-01 5.94570994e-01 1.70308784e-01
1.34455371e+00 1.35213872e-02 8.18346441e-01 -1.67977795e-01
5.37405610e-01 1.22109234e-01 4.87748742e-01 7.31464505e-01
-1.01078331e+00 7.00374305e-01 1.06010392e-01 -5.60310960e-01
-1.11883712e+00 -9.58196580e-01 -1.00340813e-01 1.80371332e+00
2.09828064e-01 -5.23294747e-01 -6.27136588e-01 -7.40915596e-01
-1.53420135e-01 9.55145180e-01 -7.49695599e-01 -3.99738580e-01
-1.66852549e-01 -8.34662080e-01 3.09478994e-02 3.41710001e-01
5.82319260e-01 -1.15517712e+00 -9.49479759e-01 3.70381504e-01
-7.63353631e-02 -1.11473775e+00 -5.39716721e-01 -1.11476821e-03
-8.93495023e-01 -1.18471777e+00 -9.30989623e-01 -7.96955347e-01
1.49056301e-01 8.12477350e-01 1.26950276e+00 -2.88501918e-01
-5.51733971e-01 6.38923705e-01 -7.08548784e-01 -6.08658075e-01
-2.73585171e-01 -6.34135753e-02 5.05510606e-02 -4.24445346e-02
1.04641581e+00 -4.24133807e-01 -4.63605911e-01 3.49228889e-01
-7.54348576e-01 -2.00489432e-01 2.57297814e-01 1.19648123e+00
4.14445817e-01 -1.29392728e-01 6.59886897e-01 -1.04190624e+00
9.25145686e-01 -6.67678356e-01 -2.81697780e-01 4.29411650e-01
-9.10643518e-01 9.26585272e-02 6.08639777e-01 -6.15707934e-01
-1.22240412e+00 5.14049530e-02 4.10583675e-01 -3.18163961e-01
-9.69931334e-02 3.30836266e-01 1.27351344e-01 -3.11167371e-02
1.05992436e+00 2.44979337e-01 -8.98859836e-03 -2.01331362e-01
5.61652124e-01 8.31049144e-01 1.69417560e-01 -4.36628491e-01
5.12540281e-01 6.24998689e-01 -3.66261393e-01 -8.15892577e-01
-1.16404843e+00 -1.15747237e+00 -7.57165432e-01 -2.81757474e-01
8.67893875e-01 -7.30450749e-01 -6.08108826e-02 4.13014799e-01
-9.99172091e-01 -2.18602210e-01 -2.47247458e-01 4.97138023e-01
-1.05514884e+00 3.02534223e-01 -3.10190767e-01 -7.00704038e-01
-3.52079123e-01 -1.12774706e+00 9.60847139e-01 1.95003763e-01
-3.30403239e-01 -1.19414663e+00 7.14456260e-01 4.97535437e-01
5.54881811e-01 7.80205131e-02 7.35074341e-01 -8.28654587e-01
5.15460745e-02 -1.22479320e-01 -2.56654527e-02 2.30168208e-01
4.30105478e-01 -5.24959981e-01 -9.92424071e-01 -1.73949927e-01
4.57622319e-01 -9.36171293e-01 1.03695607e+00 2.91218489e-01
8.92396927e-01 1.04216993e-01 -1.35259166e-01 2.51674056e-01
1.60056877e+00 1.97417095e-01 4.46997315e-01 5.97050905e-01
3.81389707e-01 3.90034437e-01 1.17614448e+00 8.82574856e-01
2.80825615e-01 6.78124964e-01 1.71257690e-01 4.06854272e-01
-1.63571298e-01 1.73716575e-01 3.53042513e-01 6.48359478e-01
1.83427766e-01 -3.89197916e-02 -1.03715849e+00 7.64849305e-01
-2.20735860e+00 -1.24756873e+00 5.58214009e-01 2.07796073e+00
8.69725764e-01 2.95867950e-01 2.38124162e-01 -4.47982140e-02
7.99188435e-01 7.78932571e-01 -6.96219325e-01 -5.31631231e-01
3.40700030e-01 1.87489077e-01 7.34890625e-02 2.21140534e-01
-1.30855978e+00 1.07848430e+00 5.98042488e+00 9.65081394e-01
-7.21543729e-01 6.31126642e-01 3.09800267e-01 -4.66028184e-01
-5.01300730e-02 -1.65763453e-01 -6.29534304e-01 3.70206684e-01
9.49258864e-01 -6.98242605e-01 3.29781413e-01 1.24474132e+00
-9.06566565e-04 -1.80098817e-01 -1.19133818e+00 1.26223254e+00
6.53533697e-01 -1.25123906e+00 -8.45714360e-02 -2.72747129e-01
1.13052797e+00 3.06880414e-01 4.10217904e-02 6.81471586e-01
5.16700923e-01 -6.25177979e-01 2.26479813e-01 3.81320983e-01
5.92330217e-01 -6.22872293e-01 7.39697635e-01 6.29883111e-01
-1.02616954e+00 -2.97266930e-01 -9.07048941e-01 -3.33763927e-01
1.96684167e-01 4.34125245e-01 -9.28175926e-01 8.30185637e-02
5.11135459e-01 9.24642920e-01 -5.80107629e-01 1.09322178e+00
5.42875640e-02 3.70145857e-01 3.46448928e-01 -4.25884724e-01
5.24947047e-01 5.24143036e-03 3.93639088e-01 1.21772873e+00
1.66514501e-01 1.66730136e-02 5.10581374e-01 4.78346527e-01
-9.45585370e-02 3.19556475e-01 -1.11838758e+00 9.10540149e-02
4.16328698e-01 1.39945149e+00 -5.74894190e-01 -4.95269120e-01
-9.31836903e-01 1.06344819e+00 5.22482395e-01 8.71851295e-03
-7.31099606e-01 -5.29038489e-01 7.61539578e-01 -1.53399274e-01
3.31672370e-01 -2.26737931e-01 -3.21300298e-01 -1.26047790e+00
-1.35832697e-01 -8.04426730e-01 6.37317777e-01 -4.44227844e-01
-1.70801878e+00 3.63680542e-01 1.30112618e-01 -1.60659003e+00
-6.35072708e-01 -4.62276369e-01 -8.57694328e-01 3.54294181e-01
-1.67381489e+00 -9.07009661e-01 -4.01385188e-01 6.78377688e-01
1.43763888e+00 -7.43781447e-01 1.13146472e+00 4.34178896e-02
-1.95333615e-01 3.90463829e-01 3.20735455e-01 -7.12955743e-02
1.16811264e+00 -1.22390556e+00 4.19751972e-01 6.32678270e-01
3.63007009e-01 2.15613529e-01 7.03254402e-01 -4.06350136e-01
-1.32352197e+00 -9.00761545e-01 7.04333842e-01 -5.78595102e-01
9.04280663e-01 -2.79401183e-01 -7.66640663e-01 1.26920834e-01
4.56040621e-01 6.71713650e-02 1.14184904e+00 2.08449200e-01
-5.83341122e-01 -1.05113320e-01 -1.14015114e+00 5.62805533e-01
1.11797953e+00 -5.87197363e-01 -1.17103469e+00 5.36192298e-01
6.14172757e-01 5.41230924e-02 -6.85722888e-01 1.99004352e-01
4.54058439e-01 -9.96055186e-01 8.24045837e-01 -7.64083922e-01
7.13261187e-01 -4.71145399e-02 -5.26423156e-01 -1.61806774e+00
-4.62971687e-01 -4.00800198e-01 -3.54414284e-01 1.01710308e+00
1.92025810e-01 -8.15611333e-03 7.73405790e-01 5.60756028e-01
-6.95291162e-02 -6.83794796e-01 -9.72880602e-01 -1.05721772e+00
1.22929029e-01 -4.55777854e-01 3.14841121e-01 1.03358817e+00
3.93572003e-01 6.35366619e-01 -6.36066318e-01 -6.30105019e-01
8.37782383e-01 2.55590111e-01 8.89254987e-01 -1.24764156e+00
-2.83316284e-01 -3.68697345e-01 -5.88332117e-01 -4.66208547e-01
5.41373074e-01 -6.80824280e-01 3.34000945e-01 -1.56604016e+00
6.15231931e-01 6.79685846e-02 -7.94210196e-01 1.54561788e-01
-2.70885289e-01 2.76283979e-01 3.66207719e-01 1.65633634e-01
-1.30973816e+00 8.89680743e-01 1.09003377e+00 -5.32952547e-01
-1.70096993e-01 -1.56264484e-01 -5.98065555e-01 8.01196992e-01
7.46149123e-01 -6.75140619e-01 -9.74000573e-01 -3.61261934e-01
-2.99161375e-01 -1.60672128e-01 -1.34869665e-01 -1.30594099e+00
4.08327430e-01 -5.83624244e-01 3.47731635e-02 -2.36726224e-01
4.39822048e-01 -6.20187044e-01 -7.34530687e-01 4.83684152e-01
-4.62683767e-01 -1.56308457e-01 -3.04607511e-01 7.03933179e-01
-2.91159481e-01 -7.06309378e-01 1.05315161e+00 -4.54484999e-01
-1.50592279e+00 4.05073971e-01 -1.94316000e-01 4.83087212e-01
1.26787770e+00 -2.57322848e-01 -4.17103678e-01 -4.58813608e-01
-3.16718102e-01 1.67125151e-01 5.35754800e-01 7.74511814e-01
8.06372225e-01 -1.43112385e+00 -7.00063705e-01 -2.73768127e-01
8.75641882e-01 -6.09354258e-01 3.37867171e-01 4.49645132e-01
-2.83011235e-02 1.85687482e-01 -5.02846420e-01 -4.76282656e-01
-1.08073235e+00 8.72545540e-01 9.93420482e-02 -4.57525522e-01
-3.74641597e-01 6.45598054e-01 -1.55993970e-02 -3.83257270e-01
2.69212753e-01 1.81391388e-01 -3.50103796e-01 4.50750500e-01
1.01203275e+00 5.80241323e-01 5.43314219e-02 -4.27534252e-01
-3.42729449e-01 5.03912449e-01 -4.34329629e-01 -3.48439574e-01
1.41088307e+00 -1.76239803e-01 5.47880054e-01 9.94032800e-01
1.31991160e+00 -8.30119610e-01 -1.18383777e+00 -7.83410311e-01
1.45085424e-01 -8.23333979e-01 9.31266397e-02 -4.84042525e-01
-5.88993609e-01 1.10313761e+00 8.63140762e-01 -1.29795045e-01
9.93058622e-01 -1.66906878e-01 8.67666841e-01 8.23774934e-01
7.07870483e-01 -1.81955624e+00 8.08570087e-01 6.84707701e-01
5.52546024e-01 -1.89612365e+00 2.50311315e-01 1.92705676e-01
-1.04514658e+00 9.19692755e-01 9.51636672e-01 -3.16302031e-01
9.18389976e-01 2.47803479e-02 1.22270368e-01 -1.67744700e-02
-1.05506301e+00 -4.69083041e-01 2.02710807e-01 9.67522681e-01
5.47809660e-01 -2.07924277e-01 -5.86171269e-01 3.81678283e-01
1.04766011e-01 1.87176272e-01 5.99782705e-01 1.26745021e+00
-8.82309675e-01 -1.11257505e+00 1.40530951e-02 5.72890103e-01
-1.85276970e-01 -6.42121360e-02 -2.54411280e-01 4.49396938e-01
-5.11379242e-02 1.19000697e+00 -4.78585921e-02 -4.94131804e-01
2.15363339e-01 4.26687509e-01 4.35134530e-01 -1.16176260e+00
-2.96651572e-01 -2.60731012e-01 -5.56133799e-02 -6.33680463e-01
-6.56788945e-01 -6.35646045e-01 -9.56880033e-01 -1.28965035e-01
-2.57872850e-01 -2.36535192e-01 5.12390375e-01 1.18008149e+00
1.66060075e-01 3.08200419e-01 7.72014558e-01 -7.84625471e-01
-1.00570059e+00 -1.28358388e+00 -6.47098243e-01 8.26408982e-01
7.01231956e-02 -1.07845473e+00 -1.66083992e-01 1.17954127e-01] | [10.02521800994873, 3.078815221786499] |
65c6e5bb-b23e-468c-bf84-b3c2ad64a435 | coloured-noise-time-series-as-appropriate | 2006.16204 | null | https://arxiv.org/abs/2006.16204v1 | https://arxiv.org/pdf/2006.16204v1.pdf | Coloured noise time series as appropriate models for environmental variation in artificial evolutionary systems | Ecological, environmental and geophysical time series consistently exhibit the characteristics of coloured (1/f^\b{eta}) noise. Here we briefly survey the literature on coloured noise, population persistence and related evolutionary dynamics, before introducing coloured noise as an appropriate model for environmental variation in artificial evolutionary systems. To illustrate and explore the effects of different noise colours, a simple evolutionary model that examines the trade-off between specialism and generalism in fluctuating environments is applied. The results of the model clearly demonstrate a need for greater generalism as environmental variability becomes `whiter', whilst specialisation is favoured as environmental variability becomes `redder'. Pink noise, sitting midway between white and red noise, is shown to be the point at which the pressures for generalism and specialism balance, providing some insight in to why `pinker' noise is increasingly being seen as an appropriate model of typical environmental variability. We go on to discuss how the results presented here feed in to a wider discussion on evolutionary responses to fluctuating environments. Ultimately we argue that Artificial Life as a field should embrace the use of coloured noise to produce models of environmental variability. | ['James M. Borg', 'Matt Grove', 'Fiona Polack'] | 2020-06-29 | null | null | null | null | ['artificial-life'] | ['miscellaneous'] | [ 3.76617312e-01 -5.71204007e-01 7.37003982e-01 4.66522910e-02
3.44525158e-01 -5.60212851e-01 8.55413914e-01 -1.16622843e-01
-7.34771669e-01 7.79924631e-01 2.27488205e-01 -6.01773620e-01
-4.32468951e-01 -7.47322321e-01 -2.90905982e-01 -1.23005712e+00
-5.01822054e-01 -9.20279548e-02 2.65177935e-01 -7.56482482e-01
2.07757980e-01 4.36540186e-01 -1.89760280e+00 -3.16914916e-01
4.92174625e-01 2.26308733e-01 3.04850161e-01 1.13905108e+00
-7.97895342e-02 3.91387455e-02 -1.16354156e+00 -1.53097823e-01
3.05867374e-01 -9.01028514e-01 -3.24665278e-01 -1.10996686e-01
-1.58207476e-01 5.42379498e-01 -4.55326624e-02 8.73035431e-01
9.77583766e-01 1.77744672e-01 6.20370746e-01 -9.02759969e-01
-8.85339320e-01 5.16393423e-01 -4.57608968e-01 6.57270253e-01
-2.05435473e-02 9.70952034e-01 5.91163278e-01 -3.88050258e-01
5.96350670e-01 1.42668760e+00 9.11038280e-01 4.14672971e-01
-1.60536373e+00 -3.22376907e-01 3.26779455e-01 -3.26665878e-01
-1.32950485e+00 -4.12061214e-01 3.79013777e-01 -4.71422940e-01
1.02666700e+00 7.54290104e-01 1.44371092e+00 8.54510903e-01
6.31973445e-01 -3.54355685e-02 1.54316211e+00 -4.45327193e-01
4.24973816e-01 -1.00943126e-01 -5.68451822e-01 -1.51547253e-01
6.90506101e-01 8.15800667e-01 -5.53164959e-01 -3.27378035e-01
7.99918532e-01 -3.82640272e-01 -4.42076892e-01 2.24973559e-01
-7.30427265e-01 4.13096309e-01 1.59364790e-01 3.61368358e-01
-2.01630771e-01 5.25329828e-01 2.86584228e-01 5.62389910e-01
4.22486246e-01 7.99396873e-01 -5.37828624e-01 -6.46656036e-01
-4.28438604e-01 2.64978200e-01 6.79815233e-01 2.97586948e-01
4.96087044e-01 5.25251269e-01 4.84739780e-01 1.25628841e+00
6.09787822e-01 1.02901316e+00 1.42337590e-01 -8.51194203e-01
-3.89140248e-01 1.30973011e-01 1.67243332e-01 -8.68466616e-01
-4.92077023e-01 -5.43813050e-01 -5.79574704e-01 6.78772569e-01
5.39061069e-01 -5.18957078e-01 -7.81550109e-01 1.97300625e+00
1.40559971e-01 -6.21039085e-02 1.19232237e-01 5.54361999e-01
2.77584732e-01 6.35407150e-01 2.63067871e-01 -5.41497231e-01
1.29669154e+00 1.83069408e-01 -5.86548209e-01 -4.06765491e-01
2.97285050e-01 -7.10209310e-01 1.10689366e+00 9.85417888e-02
-6.85251772e-01 -9.65267941e-02 -8.12652230e-01 8.14299703e-01
-4.38919842e-01 -8.73799145e-01 4.38966781e-01 1.16295171e+00
-1.32803559e+00 5.62390745e-01 -8.99822116e-01 -8.03897858e-01
-2.35092845e-02 -2.17494392e-03 3.08525134e-02 4.93899524e-01
-1.07386565e+00 9.44569707e-01 -1.34180978e-01 6.69180572e-01
-2.86102325e-01 -4.40300077e-01 -5.17426789e-01 -4.49585676e-01
-9.60940421e-02 -4.28738534e-01 9.26411331e-01 -1.07310426e+00
-1.46249688e+00 8.49953413e-01 1.78959921e-01 -6.04918860e-02
3.35424811e-01 9.73265246e-02 -6.64109826e-01 -3.55599523e-01
-1.70201674e-01 1.85347974e-01 2.37034917e-01 -1.71636140e+00
-4.73036557e-01 -4.07420039e-01 -3.33107978e-01 1.78573057e-01
2.84474522e-01 4.65500951e-01 3.38035345e-01 -7.66523123e-01
-6.72615245e-02 -1.17324972e+00 -5.69127321e-01 -3.22562195e-02
1.12251908e-01 -5.06005734e-02 6.34461820e-01 -6.86040223e-02
1.33803999e+00 -2.21875525e+00 -6.94086552e-02 3.98942977e-01
1.35515640e-02 1.70407072e-01 -7.75014386e-02 8.95464182e-01
8.77603143e-02 3.84088248e-01 -6.26270235e-01 3.30855429e-01
2.41305027e-02 5.65375328e-01 1.37811869e-01 6.79391921e-01
3.38670790e-01 4.61851031e-01 -1.09917259e+00 7.99368396e-02
2.99951085e-03 7.44648397e-01 -2.86547065e-01 -2.55611598e-01
2.56439120e-01 5.09688854e-01 2.17894524e-01 4.59288597e-01
5.46853364e-01 3.32878023e-01 1.50646612e-01 7.16389596e-01
-9.71114635e-01 -6.64495751e-02 -1.20863521e+00 5.91553569e-01
-1.43522710e-01 1.11527002e+00 3.92430425e-01 -3.82259429e-01
1.29861462e+00 -7.60937333e-02 -2.05990560e-02 -6.92529857e-01
1.88374892e-01 6.30943418e-01 8.91158998e-01 -5.75402498e-01
3.57721120e-01 -7.62942433e-01 -2.37454849e-04 4.99739915e-01
-6.31341100e-01 -7.45163918e-01 -1.24377780e-01 -2.11596787e-01
8.28500390e-01 3.24610382e-01 3.97880152e-02 -7.90129364e-01
1.53393328e-01 -2.02435806e-01 7.68631577e-01 9.60739434e-01
-5.80629587e-01 1.08552128e-01 4.92880970e-01 -4.51076984e-01
-1.12907720e+00 -7.73283422e-01 -5.46191394e-01 1.29355490e+00
1.79811344e-01 -3.04829478e-01 -3.37382615e-01 5.36652505e-01
7.64897838e-02 7.29543507e-01 -8.85386646e-01 -3.71328413e-01
-6.22033954e-01 -1.64259326e+00 6.54929996e-01 1.70135856e-01
2.29104087e-01 -1.36720943e+00 -1.51361549e+00 2.11252496e-01
4.02155161e-01 -5.13331592e-02 1.41269267e-01 8.05389941e-01
-6.13101602e-01 -5.96172452e-01 -6.24224544e-01 -3.19186568e-01
3.74364048e-01 1.37935072e-01 1.12601733e+00 4.58749652e-01
-5.59802592e-01 8.07426870e-01 -3.66413355e-01 -6.68770671e-01
-6.61166489e-01 -6.71192288e-01 2.39041194e-01 -3.76866460e-01
3.57641697e-01 -8.96658897e-01 -7.51702309e-01 2.41614461e-01
-1.09894574e+00 -6.16463840e-01 1.34284854e-01 9.63833094e-01
1.64689645e-01 7.80050382e-02 4.86461580e-01 -5.22855401e-01
7.73138762e-01 -4.87404555e-01 -4.54984844e-01 -4.84287068e-02
-4.72578853e-01 -1.89070255e-01 2.49973640e-01 -5.62422335e-01
-1.03884602e+00 -2.75191486e-01 -1.71167940e-01 6.21363342e-01
-2.02084541e-01 4.82893020e-01 1.13449723e-01 -2.01821506e-01
1.01331067e+00 3.61686796e-01 4.21743467e-02 -1.02047399e-01
1.86237261e-01 6.89728677e-01 2.66509056e-01 -6.10158563e-01
8.26417565e-01 4.08756047e-01 1.44047320e-01 -1.52217400e+00
3.32769752e-01 1.63053229e-01 -3.40284199e-01 -7.24482596e-01
5.04384279e-01 -4.50250596e-01 -6.73220694e-01 8.56782854e-01
-4.58651662e-01 -7.96150446e-01 -6.29099131e-01 3.50956976e-01
-5.04868925e-01 4.28074710e-02 -3.19654256e-01 -1.62878680e+00
2.52895862e-01 -8.50297272e-01 6.79422617e-01 4.16202992e-01
-3.30976844e-01 -1.11861908e+00 6.43674552e-01 -5.38604200e-01
9.54927623e-01 5.14503717e-01 7.83671856e-01 -2.08376244e-01
5.89718968e-02 2.11462319e-01 3.10972124e-01 -1.91617489e-01
2.98688263e-01 8.06811094e-01 -8.09921324e-01 -1.11163244e-01
2.33178824e-01 1.55006349e-01 6.09918356e-01 3.98971617e-01
9.05355960e-02 -9.55476984e-02 -6.97526559e-02 6.77305579e-01
1.56243801e+00 7.74615288e-01 5.95026135e-01 1.01794147e+00
6.20491430e-02 1.04869378e+00 1.21828519e-01 5.42259395e-01
1.42160371e-01 3.96940500e-01 3.43585640e-01 -1.21961325e-01
1.42775550e-01 3.28963310e-01 3.93953174e-01 7.48098016e-01
-3.14417094e-01 -1.73868746e-01 -1.26207924e+00 5.73668003e-01
-1.24766469e+00 -1.20107138e+00 -5.89788556e-01 2.45555496e+00
7.98556387e-01 2.13592917e-01 4.53139544e-01 -4.35436517e-02
7.14695036e-01 1.27733648e-01 -4.25737470e-01 -8.33882630e-01
-7.91613579e-01 2.07917131e-02 6.66000307e-01 6.25927627e-01
-7.17172384e-01 6.42704070e-01 8.22675133e+00 8.77504423e-02
-1.10961199e+00 -3.81429553e-01 4.74016577e-01 -1.06219731e-01
-5.95560908e-01 1.48239285e-01 -3.04318905e-01 5.75253665e-01
1.29846740e+00 -2.79097706e-01 3.21950883e-01 9.84272808e-02
7.17347682e-01 -5.77030361e-01 -2.92478740e-01 4.40045655e-01
-4.36807752e-01 -5.27386546e-01 -4.10926908e-01 2.41456881e-01
4.55005705e-01 2.21582406e-04 8.52010176e-02 2.21169945e-02
8.74393344e-01 -9.90683258e-01 6.64999247e-01 5.91756463e-01
5.92972636e-01 -6.21032834e-01 6.08508527e-01 1.02288453e-02
-1.20798099e+00 -2.44839981e-01 -4.91164029e-01 -6.41731739e-01
3.30880731e-01 2.61909604e-01 -2.06511900e-01 -4.45363596e-02
1.07154238e+00 1.28380850e-01 -6.93418860e-01 1.33620298e+00
1.05032183e-01 5.94912827e-01 -6.87164247e-01 -3.12252313e-01
1.72291130e-01 -4.40813094e-01 8.52731943e-01 1.45597148e+00
3.33413363e-01 1.80474922e-01 -4.41671580e-01 4.90990162e-01
8.88661981e-01 -3.90107781e-02 -9.49214101e-01 -1.56444430e-01
7.48148680e-01 5.78885078e-01 -1.06914890e+00 -1.13713473e-01
-2.78306276e-01 5.68359613e-01 -4.11245406e-01 4.41295564e-01
-4.12253678e-01 -5.17700911e-01 1.00864255e+00 -2.68216431e-02
1.12902567e-01 -3.86542857e-01 -3.34800035e-01 -5.73864818e-01
-5.75726211e-01 -9.22516525e-01 -6.28573773e-03 -6.36164248e-01
-1.18720257e+00 6.34141386e-01 1.54137403e-01 -8.92776489e-01
2.51458958e-03 -5.38511693e-01 -6.41423106e-01 1.16065156e+00
-6.91537738e-01 -4.53257769e-01 -3.48009914e-02 -1.64186373e-01
2.80814767e-01 1.87426344e-01 6.95303202e-01 -3.12573344e-01
-5.88145971e-01 1.85055017e-01 7.77923465e-01 -4.93930221e-01
3.91424119e-01 -1.36577010e+00 8.57391179e-01 8.15345883e-01
-3.91512603e-01 8.74218941e-01 1.39379179e+00 -1.03536475e+00
-1.19561946e+00 -4.91376460e-01 3.81500214e-01 -5.03929317e-01
9.18616593e-01 -3.02622348e-01 -9.20704067e-01 9.82467905e-02
3.60137939e-01 -4.36078876e-01 1.06544662e+00 -4.22914810e-02
-5.74575886e-02 5.98867238e-02 -1.04373503e+00 1.07796454e+00
1.17052758e+00 -3.18671495e-01 -3.03384721e-01 -6.34864345e-02
4.67029154e-01 2.81406671e-01 -9.19175208e-01 1.79409280e-01
9.48294580e-01 -9.84856606e-01 6.93813086e-01 -4.01224911e-01
-2.39273205e-01 -5.74472189e-01 -1.43616363e-01 -1.62445605e+00
-6.79640889e-01 -8.59891534e-01 9.33915973e-01 1.13350272e+00
4.16885942e-01 -1.18322361e+00 1.18602738e-01 4.94280040e-01
4.47863229e-02 -3.98865849e-01 -1.26405895e+00 -9.87985969e-01
5.21694422e-01 -2.69034863e-01 4.22665983e-01 8.08914840e-01
-9.76418406e-02 -7.01526180e-02 -1.88018054e-01 3.98086384e-02
5.09038985e-01 -6.22966111e-01 6.43779457e-01 -1.44796336e+00
-2.71822065e-01 -9.74650979e-01 -7.26052344e-01 -2.50474095e-01
-6.81394994e-01 -1.02306582e-01 5.55063128e-01 -1.19740796e+00
-2.22099125e-01 -3.98963153e-01 -7.95159042e-02 -1.11972362e-01
-4.17721659e-01 3.79446298e-01 2.59114057e-01 2.40281403e-01
2.50268996e-01 4.65115994e-01 9.22634840e-01 3.58264983e-01
-6.91414535e-01 -1.65307507e-01 -9.65781689e-01 4.47460741e-01
7.72312462e-01 -3.54249716e-01 -2.19561607e-01 -1.43865749e-01
6.41188920e-01 -3.52106303e-01 5.77634461e-02 -8.13492835e-01
-5.71680926e-02 -6.51519418e-01 2.31081292e-01 1.05684176e-01
2.88811266e-01 -7.96009243e-01 8.41208756e-01 9.21398282e-01
-1.07555784e-01 6.31082356e-01 6.90843165e-01 6.83317363e-01
4.64723140e-01 -9.02881026e-02 1.01176882e+00 -3.62495452e-01
-4.81315970e-01 -4.48128164e-01 -1.33234942e+00 3.06085125e-02
8.69904935e-01 -8.03298652e-01 -6.67770684e-01 -2.11524770e-01
-9.08432961e-01 7.04289824e-02 1.10928822e+00 2.64117777e-01
2.40411997e-01 -9.49870586e-01 -9.02015448e-01 2.51840234e-01
7.29200849e-03 -4.93685395e-01 2.26703152e-01 8.12033534e-01
-9.87687111e-01 -2.06988096e-01 -3.18641692e-01 -5.56210816e-01
-1.23270440e+00 2.72318512e-01 8.75697911e-01 6.01413369e-01
-6.84061646e-01 1.28269601e+00 1.80629194e-01 -3.02240044e-01
-1.31887808e-01 1.12983592e-01 -5.93105927e-02 5.40703386e-02
4.58951205e-01 5.35078406e-01 -5.93813360e-01 -8.79207850e-01
-6.17887259e-01 7.49269664e-01 6.08227313e-01 -5.27040064e-01
1.48445237e+00 -5.36764503e-01 -5.13033271e-01 1.20432031e+00
3.65692645e-01 1.86934397e-01 -1.53139901e+00 1.30099282e-01
1.34764880e-01 -6.44553423e-01 -4.09456462e-01 -8.32229078e-01
-4.54874486e-01 5.71231365e-01 8.85604680e-01 8.91527057e-01
1.25530720e+00 -1.11366577e-01 -1.90896824e-01 2.19732493e-01
1.25446603e-01 -1.14276445e+00 -3.27873856e-01 4.54179019e-01
9.96608377e-01 -4.62245882e-01 -2.68694200e-02 6.66377414e-03
-5.89132845e-01 9.09416676e-01 5.27959883e-01 -1.34048715e-01
6.88084364e-01 5.38051188e-01 4.41935211e-01 -1.75186977e-01
-1.11122966e+00 -4.62215751e-01 -5.81313789e-01 1.13388598e+00
4.97845143e-01 2.86467999e-01 -6.40753150e-01 -2.72544712e-01
-6.93637013e-01 -6.65842593e-01 9.11132932e-01 1.09797561e+00
-9.61302698e-01 -1.08888054e+00 -1.09015799e+00 4.00268257e-01
-4.46428686e-01 -1.73651129e-01 -9.63813424e-01 8.39780271e-01
3.77450734e-01 9.27659392e-01 2.31532976e-01 -2.70790488e-01
4.28947300e-01 1.72948807e-01 2.33165920e-01 -3.41030210e-01
-1.01229465e+00 5.47271788e-01 1.38917372e-01 2.06111431e-01
-4.99588996e-01 -9.92769241e-01 -5.90935051e-01 -6.49839401e-01
-5.52505791e-01 3.33543092e-01 6.16173208e-01 5.22196352e-01
3.34526673e-02 5.99389434e-01 4.20264393e-01 -6.71313047e-01
-1.27299413e-01 -9.06411350e-01 -8.34997833e-01 5.77819161e-02
3.47277492e-01 -6.32303417e-01 -9.92299795e-01 -2.12491632e-01] | [5.592878341674805, 4.124729156494141] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.