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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
5eded2ee-2bfe-4878-a6d4-69e652c583f9 | enforcing-temporal-consistency-in-deep | 1906.07160 | null | https://arxiv.org/abs/1906.07160v1 | https://arxiv.org/pdf/1906.07160v1.pdf | Enforcing temporal consistency in Deep Learning segmentation of brain MR images | Longitudinal analysis has great potential to reveal developmental trajectories and monitor disease progression in medical imaging. This process relies on consistent and robust joint 4D segmentation. Traditional techniques are dependent on the similarity of images over time and the use of subject-specific priors to reduce random variation and improve the robustness and sensitivity of the overall longitudinal analysis. This is however slow and computationally intensive as subject-specific templates need to be rebuilt every time. The focus of this work to accelerate this analysis with the use of deep learning. The proposed approach is based on deep CNNs and incorporates semantic segmentation and provides a longitudinal relationship for the same subject. The proposed approach is based on deep CNNs and incorporates semantic segmentation and provides a longitudinal relationship for the same subject. The state of art using 3D patches as inputs to modified Unet provides results around ${0.91 \pm 0.5}$ Dice and using multi-view atlas in CNNs provide around the same results. In this work, different models are explored, each offers better accuracy and fast results while increasing the segmentation quality. These methods are evaluated on 135 scans from the EADC-ADNI Harmonized Hippocampus Protocol. Proposed CNN based segmentation approaches demonstrate how 2D segmentation using prior slices can provide similar results to 3D segmentation while maintaining good continuity in the 3D dimension and improved speed. Just using 2D modified sagittal slices provide us a better Dice and longitudinal analysis for a given subject. For the ADNI dataset, using the simple UNet CNN technique gives us ${0.84 \pm 0.5}$ and while using modified CNN techniques on the same input yields ${0.89 \pm 0.5}$. Rate of atrophy and RMS error are calculated for several test cases using various methods and analyzed. | ['Malav Bateriwala', 'Pierrick Bourgeat'] | 2019-06-13 | null | null | null | null | ['4d-spatio-temporal-semantic-segmentation', '3d-medical-imaging-segmentation', 'brain-image-segmentation'] | ['computer-vision', 'medical', 'medical'] | [-2.49498814e-01 5.09174205e-02 3.83200459e-02 -6.13910198e-01
-5.97073913e-01 -4.44333076e-01 2.18371719e-01 4.28972661e-01
-6.59347475e-01 8.14960420e-01 1.17125176e-02 1.65183201e-01
-5.11187851e-01 -7.02742338e-01 -5.08547366e-01 -5.75166464e-01
-3.77064824e-01 7.45170176e-01 5.52124560e-01 -7.43989646e-02
7.26502687e-02 5.32253146e-01 -1.10869575e+00 -4.41356078e-02
7.73804724e-01 7.89495528e-01 9.94245932e-02 4.41037178e-01
2.10091636e-01 -1.08937427e-01 -3.26488495e-01 -3.11261088e-01
4.41405237e-01 -1.87116578e-01 -9.05506194e-01 -8.90305638e-02
7.53326833e-01 -4.13420945e-01 1.37056321e-01 9.09320295e-01
8.06777120e-01 2.20199004e-02 7.89381802e-01 -8.57201636e-01
-3.59203190e-01 5.89007795e-01 -8.13859463e-01 7.40739584e-01
-1.79909728e-06 -4.04520631e-02 3.55160892e-01 -3.04486632e-01
6.51918888e-01 8.57417703e-01 1.10024619e+00 3.41964722e-01
-1.35966659e+00 -8.41940045e-01 -2.40184352e-01 4.66890559e-02
-1.11287081e+00 -8.10651667e-03 5.41427732e-01 -6.90612376e-01
1.00537670e+00 3.77440527e-02 9.30658519e-01 5.16820371e-01
5.28202236e-01 1.73427343e-01 1.37639272e+00 -1.35790348e-01
2.48850267e-02 -2.56450832e-01 3.84732932e-01 6.00364804e-01
3.11886579e-01 -1.05193861e-01 1.37239262e-01 1.92970559e-01
7.79221058e-01 -8.05709213e-02 -1.33178920e-01 -1.54142961e-01
-1.09150577e+00 7.32089162e-01 4.73151833e-01 7.35457540e-01
-4.11788851e-01 1.29028901e-01 6.71943665e-01 5.88439479e-02
6.02981448e-01 1.22733898e-01 -3.28241318e-01 4.28313762e-02
-1.69926417e+00 4.39131767e-01 1.46434661e-02 3.73195648e-01
3.18194002e-01 -1.19812321e-02 -8.83208960e-03 9.58835602e-01
2.73962319e-01 3.64174366e-01 6.43115640e-01 -1.13933897e+00
2.00765580e-01 6.92703009e-01 -3.34956765e-01 -9.69033420e-01
-9.99465287e-01 -7.50038922e-01 -8.88685286e-01 4.76218909e-01
5.68134069e-01 -1.99606791e-01 -1.48300898e+00 1.83897054e+00
4.71777439e-01 -3.88529927e-01 -3.04241508e-01 8.03683877e-01
5.29699981e-01 2.20849767e-01 7.61459768e-02 -1.48307651e-01
1.37064207e+00 -5.69506109e-01 -5.67822754e-01 1.96918413e-01
5.65660000e-01 -7.96794236e-01 6.93722248e-01 3.96306455e-01
-1.50674343e+00 -3.84362102e-01 -1.21928596e+00 1.28764391e-01
-2.74828523e-01 9.01720822e-02 2.00174227e-01 8.81305218e-01
-1.55296373e+00 9.21586812e-01 -1.28042495e+00 -5.43683827e-01
6.99759841e-01 8.06500852e-01 -6.15149260e-01 4.60272431e-02
-9.16259527e-01 9.74638164e-01 3.20802003e-01 5.80567159e-02
-6.34385347e-01 -9.06762779e-01 -5.77316642e-01 -1.57259271e-01
-2.18954161e-01 -5.55310845e-01 8.86961162e-01 -7.32801616e-01
-9.36973393e-01 1.06465685e+00 1.52270138e-01 -6.15710855e-01
7.90734053e-01 -6.11115843e-02 -8.92809257e-02 4.57756281e-01
5.17060280e-01 9.78887081e-01 3.06613088e-01 -8.63723695e-01
-2.30970755e-01 -1.00810885e+00 -2.91075885e-01 -2.29411535e-02
8.50234833e-03 1.20844044e-01 -1.87387064e-01 -7.43009508e-01
5.44190109e-01 -9.65182304e-01 -2.02650189e-01 -2.50635808e-03
-1.30364567e-01 2.07976829e-02 7.11667418e-01 -1.27965462e+00
9.01179850e-01 -1.55993283e+00 -5.40195033e-02 2.96490073e-01
4.36685771e-01 1.56390086e-01 4.20743823e-01 -1.03321970e-01
-4.05087531e-01 2.45037585e-01 -7.02646673e-01 -2.93946236e-01
-5.31115949e-01 -9.37556550e-02 8.23413134e-01 8.34634781e-01
-2.49029726e-01 6.04638338e-01 -4.57673788e-01 -5.76390803e-01
-1.28824254e-02 5.99845886e-01 -6.16653204e-01 -6.64729998e-02
3.28833908e-01 4.44595903e-01 -1.10271744e-01 3.59658837e-01
8.97468448e-01 1.35528624e-01 -3.10938079e-02 -3.57671112e-01
-1.15668178e-01 -3.61903965e-01 -1.01206648e+00 1.81322336e+00
-2.32291132e-01 5.06711960e-01 2.35111699e-01 -1.46141171e+00
9.38128173e-01 6.01984978e-01 7.53491223e-01 -8.26854169e-01
3.46569121e-01 4.31303561e-01 4.72715855e-01 -3.76818061e-01
-1.69257134e-01 -3.96714419e-01 3.22758943e-01 5.98965049e-01
2.30670944e-01 2.40599569e-02 2.43023738e-01 2.10039709e-02
9.20511305e-01 1.54206380e-01 -4.38086949e-02 -6.49401844e-01
4.47739065e-01 -7.26118982e-02 5.37907183e-01 2.82253921e-01
-4.85531777e-01 9.59430933e-01 4.57083017e-01 -4.92590606e-01
-1.32657099e+00 -1.12094307e+00 -4.53353196e-01 1.71377346e-01
-4.99645680e-01 1.18139237e-01 -1.05823481e+00 -5.63666880e-01
-3.65020752e-01 3.52593541e-01 -9.18663085e-01 1.53954357e-01
-7.44588435e-01 -1.14140344e+00 5.87488055e-01 5.59126079e-01
7.10341275e-01 -9.52380657e-01 -9.21998918e-01 2.58750260e-01
9.25592184e-02 -7.78050184e-01 -2.89322704e-01 5.86923696e-02
-1.50075448e+00 -8.86247814e-01 -1.40970457e+00 -6.02691710e-01
7.41617799e-01 -2.78021336e-01 8.33000481e-01 -3.18362080e-02
-3.74868363e-01 1.52133077e-01 -1.49553448e-01 -1.86420500e-01
-1.48478821e-01 1.32474869e-01 -5.44773266e-02 -5.04295647e-01
-1.35475621e-01 -1.23928881e+00 -1.08668399e+00 1.04612242e-02
-7.04999089e-01 -8.57529268e-02 4.02506977e-01 6.39044285e-01
6.38641119e-01 -8.54642019e-02 6.96825206e-01 -7.35832810e-01
5.03285706e-01 -4.43191737e-01 -4.68457222e-01 2.15143133e-02
-1.11662209e+00 -1.45559818e-01 4.30469960e-01 -1.44117072e-01
-8.51509631e-01 -1.07348748e-01 -2.11876467e-01 -2.67680496e-01
-3.86245519e-01 2.50614017e-01 2.11303622e-01 4.23996849e-03
5.46039283e-01 7.40249529e-02 3.31569672e-01 -6.16698682e-01
-1.44279078e-02 1.79561093e-01 4.29667681e-01 -3.13447535e-01
1.38579831e-01 5.76777816e-01 2.90805370e-01 -3.36637527e-01
-3.34447473e-01 -2.44902760e-01 -1.16807246e+00 -3.59458953e-01
1.27000523e+00 -4.79470879e-01 -2.40280211e-01 6.06610596e-01
-9.86114860e-01 -2.60840148e-01 1.03123426e-01 6.37966216e-01
-5.34826577e-01 2.83283561e-01 -4.32135820e-01 -3.68524164e-01
-1.04777765e+00 -1.65094316e+00 6.79389954e-01 1.88810617e-01
-4.23744321e-01 -1.10763299e+00 1.93464741e-01 5.67265689e-01
4.78438616e-01 6.92145705e-01 1.08186615e+00 -6.45385385e-01
-3.41108721e-03 -3.64035606e-01 -2.89726824e-01 4.90118504e-01
1.69696376e-01 -1.53564617e-01 -6.40700161e-01 -1.94524720e-01
-2.32829098e-02 2.22843260e-01 6.74590051e-01 9.82831299e-01
8.46520901e-01 -3.21121402e-02 -1.69256106e-01 4.89498287e-01
1.57303846e+00 6.46341383e-01 4.65667218e-01 4.57582951e-01
6.47556543e-01 9.15086985e-01 8.06546807e-02 -5.76422177e-02
3.79978031e-01 5.11527598e-01 2.47558385e-01 -2.24770382e-01
-3.76941383e-01 4.91635561e-01 -1.63705021e-01 6.88437939e-01
-3.09358895e-01 2.47977853e-01 -1.34716165e+00 1.07206738e+00
-1.36730242e+00 -7.08222210e-01 -4.56867874e-01 2.08334279e+00
7.80381382e-01 2.44922563e-01 3.62551212e-01 2.66442746e-01
7.01636136e-01 -1.95925340e-01 -3.04271340e-01 -6.15651667e-01
1.13842383e-01 8.12896967e-01 5.78501582e-01 4.43477929e-01
-7.84246325e-01 5.17967701e-01 5.65548086e+00 5.51647663e-01
-1.36122715e+00 6.75670266e-01 9.06148314e-01 -3.85729253e-01
1.83100715e-01 -1.98775545e-01 -4.35599744e-01 5.37268758e-01
9.74493504e-01 7.19584227e-02 -5.25949225e-02 3.73112798e-01
6.18025064e-01 -6.07238710e-01 -7.34208405e-01 7.11726069e-01
-1.36422902e-01 -1.19361436e+00 -2.23905683e-01 7.08636120e-02
7.54261136e-01 2.10645422e-01 -5.82082905e-02 -2.91381121e-01
-2.22942486e-01 -1.22301280e+00 5.55617630e-01 6.19288266e-01
7.29055107e-01 -9.03580785e-01 1.07818675e+00 -5.73971309e-02
-8.62289906e-01 3.01549017e-01 1.14642940e-01 1.35799393e-01
2.31749505e-01 5.77080429e-01 -8.78355086e-01 5.10567904e-01
1.17762864e+00 4.36913818e-01 -6.06341481e-01 1.27509189e+00
2.93303639e-01 6.91835403e-01 -4.03948992e-01 4.33694810e-01
4.68952775e-01 -3.11173290e-01 6.32205725e-01 1.23062027e+00
4.81198102e-01 8.85838196e-02 -3.50219131e-01 8.40220988e-01
2.21219331e-01 4.56758261e-01 -3.98809522e-01 4.71190125e-01
-1.12161599e-01 1.19298851e+00 -1.37690711e+00 -2.32788280e-01
-2.79807627e-01 6.93007708e-01 1.09548450e-01 -2.66213063e-03
-4.79371488e-01 -1.30272046e-01 1.40718952e-01 7.20727444e-01
1.76618055e-01 -2.40201607e-01 -6.53932393e-01 -4.93575126e-01
-2.00338867e-02 -6.66026533e-01 4.69562024e-01 -6.63268805e-01
-9.43417430e-01 6.90156281e-01 4.37036127e-01 -6.95151091e-01
-2.43660603e-02 -2.67358720e-01 -7.52943635e-01 1.10165894e+00
-7.15877950e-01 -1.10266960e+00 -1.58393726e-01 4.64813948e-01
4.19899553e-01 -3.22217010e-02 5.60998201e-01 5.49690962e-01
-5.97643912e-01 5.28343201e-01 -5.49578667e-02 9.12147686e-02
6.27197385e-01 -1.35166264e+00 -3.69114205e-02 8.65649998e-01
-4.34638411e-01 7.29564250e-01 5.81786454e-01 -9.92315888e-01
-3.42293262e-01 -9.09081519e-01 5.95274270e-01 -6.73290715e-02
2.47466356e-01 1.57546133e-01 -9.67758298e-01 5.33590317e-01
4.74586725e-01 -8.66444856e-02 4.53449547e-01 -2.12107196e-01
2.31146723e-01 -2.63169378e-01 -1.65453351e+00 2.14203924e-01
6.15771115e-01 -2.20457707e-02 -5.38102806e-01 1.05177239e-01
2.97569811e-01 -3.72306794e-01 -1.30702949e+00 5.52856028e-01
8.34828496e-01 -1.26245511e+00 8.17378402e-01 -1.66901276e-01
3.35966796e-01 -2.20267922e-02 1.39113307e-01 -1.12730694e+00
-5.67210242e-02 -4.04412225e-02 5.26206493e-01 1.12395918e+00
5.66352069e-01 -5.37820637e-01 8.67385209e-01 6.26618385e-01
-2.47092217e-01 -1.13985634e+00 -1.21773136e+00 -6.14582002e-01
5.66778183e-01 -4.33037639e-01 1.84065938e-01 8.47232580e-01
-4.70293611e-01 -9.65314656e-02 1.89273909e-01 6.97325841e-02
8.46960664e-01 -3.31430644e-01 8.97436514e-02 -1.22669458e+00
1.86767861e-01 -5.37603796e-01 -6.68570399e-01 1.21987961e-01
-2.23633602e-01 -9.68252003e-01 -4.60866183e-01 -1.78985715e+00
4.43124413e-01 -5.05914450e-01 -2.10300535e-01 4.85819191e-01
2.26106018e-01 5.89162052e-01 -7.57173300e-02 4.33939174e-02
1.61558107e-01 1.01084165e-01 1.36845803e+00 1.15691744e-01
-1.86077356e-01 -3.12837400e-02 -2.53830850e-01 7.81732917e-01
1.07288325e+00 -6.00540698e-01 -5.54804981e-01 -4.15238142e-01
1.80346165e-02 7.22712353e-02 5.21346331e-01 -1.30134082e+00
-2.32585967e-02 5.87330282e-01 8.15811217e-01 -8.88202310e-01
1.90765917e-01 -6.82635188e-01 2.73425639e-01 7.47538865e-01
7.29837939e-02 6.24547482e-01 4.31986421e-01 8.50996897e-02
-5.87932579e-02 -3.41936082e-01 1.19910264e+00 -4.94121850e-01
-6.15826026e-02 3.46390396e-01 -1.60241440e-01 -1.37636125e-01
9.54905629e-01 -5.99875093e-01 2.30800986e-01 -3.11283290e-01
-1.34079099e+00 1.57659754e-01 2.96630740e-01 -3.25689614e-02
3.31967890e-01 -1.07535517e+00 -6.07103348e-01 -3.11624587e-01
-5.76465845e-01 9.56647247e-02 5.11470556e-01 1.47605979e+00
-8.89379799e-01 5.44600308e-01 -8.54192376e-01 -7.29913652e-01
-1.48994768e+00 5.69867641e-02 7.12291479e-01 -4.27071661e-01
-5.36467552e-01 7.92617559e-01 1.17610320e-02 -2.44614586e-01
-1.83422327e-01 -4.26179558e-01 -4.65867102e-01 4.26556677e-01
4.04346101e-02 6.44681692e-01 2.65286297e-01 -8.08618069e-01
-4.67103928e-01 9.27137434e-01 -2.90371060e-01 -5.18927395e-01
1.72289765e+00 -2.92574346e-01 -3.79206508e-01 1.93496957e-01
1.38173306e+00 -2.60865062e-01 -1.04362857e+00 2.60021359e-01
-1.64046828e-02 4.64687981e-02 5.15839696e-01 -1.06438756e+00
-1.51286519e+00 1.03272212e+00 1.51169789e+00 -1.73735619e-01
1.17091346e+00 -7.85233006e-02 9.88260865e-01 -6.06194854e-01
1.13298349e-01 -9.41466749e-01 -2.23992333e-01 -9.34355892e-03
9.04553771e-01 -1.07809997e+00 1.77865505e-01 -1.07378624e-01
-3.39491308e-01 1.23556542e+00 6.60295129e-01 -3.42198402e-01
7.61933804e-01 1.71264783e-01 7.35327974e-02 -6.45047665e-01
5.33104353e-02 1.31868914e-01 4.76827204e-01 6.45323515e-01
7.03967035e-01 1.86648332e-02 -1.06297421e+00 6.49585366e-01
-3.28525513e-01 -6.95595592e-02 3.94504845e-01 8.10855091e-01
-6.61977455e-02 -1.13504052e+00 -3.23275030e-01 6.34602010e-01
-1.19795203e+00 2.97257546e-02 -2.24933084e-02 1.06732476e+00
6.46355271e-01 4.11074758e-01 1.02052219e-01 6.18995316e-02
2.01743647e-01 1.37241438e-01 7.47420371e-01 -4.79750812e-01
-8.75417411e-01 2.61179775e-01 -7.47599499e-03 -5.11208892e-01
-6.81552827e-01 -1.08940721e+00 -1.42871702e+00 -1.06437718e-02
1.06661975e-01 -1.39801681e-01 8.75120699e-01 9.56326783e-01
1.05757646e-01 4.28403139e-01 8.99998844e-02 -6.34990275e-01
8.72995108e-02 -1.10154033e+00 -4.77402359e-01 1.68458819e-01
1.93910480e-01 -8.88591468e-01 7.03696907e-02 8.45592991e-02] | [14.17313003540039, -2.321460247039795] |
a24f6474-f6b1-4d42-ac0e-92297051a3b2 | segmentation-evaluation-metrics-a-comparison | null | null | https://aclanthology.org/L14-1709 | https://aclanthology.org/L14-1709.pdf | Segmentation evaluation metrics, a comparison grounded on prosodic and discourse units | Knowledge on evaluation metrics and best practices of using them have improved fast in the recent years Fort et al. (2012). However, the advances concern mostly evaluation of classification related tasks. Segmentation tasks have received less attention. Nevertheless, there are crucial in a large number of linguistic studies. A range of metrics is available (F-score on boundaries, F-score on units, WindowDiff ((WD), Boundary Similarity (BS) but it is still relatively difficult to interpret these metrics on various linguistic segmentation tasks, such as prosodic and discourse segmentation. In this paper, we consider real segmented datasets (introduced in Peshkov et al. (2012)) as references which we deteriorate in different ways (random addition of boundaries, random removal boundaries, near-miss errors introduction). This provide us with various measures on controlled datasets and with an interesting benchmark for various linguistic segmentation tasks. | ["Laurent Pr{\\'e}vot", 'Klim Peshkov'] | 2014-05-01 | null | null | null | lrec-2014-5 | ['discourse-segmentation'] | ['natural-language-processing'] | [ 1.57357633e-01 3.28178704e-01 -1.46731153e-01 -4.40076798e-01
-9.05795753e-01 -9.33480799e-01 7.25321233e-01 7.84471631e-01
-7.13575184e-01 1.00005829e+00 2.26462260e-01 -1.66508213e-01
-1.27850398e-01 -4.06926274e-01 -3.01797271e-01 -4.75631714e-01
-9.13373530e-02 5.03022611e-01 7.16952860e-01 -1.82866812e-01
6.56352103e-01 2.78909504e-01 -1.71386278e+00 3.60349357e-01
1.02591336e+00 8.09468925e-01 2.17689171e-01 5.83311141e-01
-2.13589579e-01 3.22341919e-01 -8.92532587e-01 -5.07556081e-01
-1.13060631e-01 -4.54678863e-01 -1.06289184e+00 -1.06016777e-01
1.47765920e-01 4.47780222e-01 5.65549731e-01 1.07273924e+00
4.90365386e-01 2.91327238e-01 8.00203264e-01 -7.36704886e-01
-3.94302040e-01 1.21813107e+00 -4.36683118e-01 4.95901585e-01
5.65475702e-01 -1.73543379e-01 1.20871484e+00 -7.43599594e-01
7.55721450e-01 1.15541959e+00 7.89697707e-01 3.48875910e-01
-1.21284556e+00 -1.04990967e-01 -1.79416597e-01 1.77948162e-01
-1.03184295e+00 -2.89661884e-01 4.74090487e-01 -7.33031571e-01
7.77027607e-01 4.95918900e-01 2.77649909e-01 9.36941922e-01
-1.48228288e-01 7.51348495e-01 1.43108833e+00 -8.29951584e-01
-6.70580380e-03 4.15946722e-01 9.13338363e-01 4.75327782e-02
1.48003995e-01 -4.06255238e-02 -2.48593152e-01 3.10100108e-01
1.66510299e-01 -7.91231215e-01 -2.71632820e-01 1.67374775e-01
-1.22099936e+00 8.05031776e-01 3.71660888e-02 1.04725945e+00
-1.23596624e-01 -5.13037205e-01 8.01871598e-01 2.64416486e-01
5.60370326e-01 5.07228553e-01 -5.46122909e-01 -4.74092126e-01
-1.13194609e+00 4.33665574e-01 9.24842536e-01 8.09833229e-01
5.54390669e-01 -2.93391347e-01 -4.81063724e-01 1.05514467e+00
2.42454521e-02 1.17433198e-01 8.03380907e-01 -7.48257041e-01
4.13713068e-01 3.77679050e-01 4.66355868e-02 -9.51040387e-01
-7.34880149e-01 4.50228341e-02 -7.42492259e-01 -5.87471128e-02
8.83287609e-01 -3.13821018e-01 -5.46276271e-01 1.53312051e+00
1.93631858e-01 -3.04229319e-01 -1.70504913e-01 7.64217496e-01
9.89342391e-01 6.59807980e-01 2.55075991e-01 -5.78031778e-01
1.48202705e+00 -9.07339215e-01 -1.01625252e+00 -1.47422627e-02
6.73573077e-01 -1.30079639e+00 1.29183543e+00 6.19917214e-01
-1.19602633e+00 -7.71994889e-01 -7.90330768e-01 -1.44120067e-01
-6.87208712e-01 1.48945570e-01 4.58393037e-01 9.34871256e-01
-1.00664926e+00 7.55710840e-01 -6.11268401e-01 -5.46238899e-01
-8.57538804e-02 2.47992545e-01 -2.12306663e-01 5.55199325e-01
-1.22118151e+00 9.01246965e-01 6.92077339e-01 -1.20637335e-01
-1.01833217e-01 -5.33711374e-01 -9.31570590e-01 -3.05582285e-01
4.32063401e-01 9.73365828e-02 1.31408930e+00 -8.84204686e-01
-1.54348433e+00 1.31551826e+00 1.52875468e-01 -6.41338468e-01
6.99098945e-01 -2.22766355e-01 -3.42178047e-01 -1.44621566e-01
1.44009203e-01 6.34025276e-01 1.86317787e-01 -1.13638425e+00
-8.06705058e-01 -2.51185417e-01 1.48641202e-03 7.21388683e-02
3.80582139e-02 4.90884811e-01 -1.14855699e-01 -8.26975644e-01
-3.21231447e-02 -6.97346807e-01 2.11063866e-03 -7.24843204e-01
-6.44139230e-01 -5.87348282e-01 4.30812508e-01 -9.15595651e-01
1.74004185e+00 -2.08778000e+00 2.29363725e-01 -2.44935438e-01
-2.53179908e-01 6.20914578e-01 2.67865270e-01 3.65456969e-01
-3.83478217e-02 6.60447657e-01 -6.63137197e-01 -4.54737455e-01
2.17173114e-01 1.20111138e-01 -1.22418506e-02 2.17674926e-01
2.87032843e-01 6.26486123e-01 -6.81250155e-01 -7.08500147e-01
2.67076790e-01 3.19870234e-01 -3.28444809e-01 5.72706386e-02
-1.92083746e-01 5.72326362e-01 -1.72853902e-01 4.43387270e-01
4.18780833e-01 5.09402871e-01 -2.06207514e-01 -6.92345202e-02
-6.72442973e-01 4.69894081e-01 -1.33011019e+00 1.47469652e+00
-2.77807683e-01 7.94134259e-01 -3.07963882e-02 -1.14153540e+00
8.95309806e-01 5.13167679e-01 1.69679016e-01 -6.01280093e-01
4.20763582e-01 2.61019319e-01 2.18348786e-01 -5.75177312e-01
8.35316956e-01 -1.27687961e-01 -3.08730692e-01 2.33318880e-01
-4.25423235e-02 -1.26676574e-01 8.73281002e-01 -3.99061084e-01
6.45214319e-01 2.10598558e-01 4.45531189e-01 -6.26675069e-01
8.93299580e-01 1.13041297e-01 3.93461704e-01 4.31894422e-01
-4.77327049e-01 9.80814457e-01 8.41644108e-01 9.75844786e-02
-9.56172049e-01 -7.67890275e-01 -6.98676407e-01 1.10859811e+00
-1.83763415e-01 -2.69372761e-01 -1.16839361e+00 -4.00215864e-01
-3.33129495e-01 7.99428940e-01 -7.13628948e-01 2.79897451e-01
-8.42932165e-01 -8.08998048e-01 7.53033400e-01 2.79682785e-01
4.26403672e-01 -1.62447190e+00 -6.64590061e-01 2.60113150e-01
-1.34473637e-01 -1.09034407e+00 -3.46564054e-01 1.48205400e-01
-8.30591619e-01 -1.04310560e+00 -7.15072393e-01 -9.22817647e-01
-1.35974986e-02 -3.24073553e-01 1.37686563e+00 -3.39497149e-01
-1.17982917e-01 5.31457402e-02 -6.42850399e-01 -3.41792017e-01
-7.42347956e-01 3.82042140e-01 -3.44148993e-01 -2.62896150e-01
4.12301660e-01 -1.96899205e-01 -5.74601628e-02 3.06304187e-01
-9.53596592e-01 -4.48353976e-01 1.58642188e-01 5.64439654e-01
7.22669065e-01 -2.22311154e-01 6.87587738e-01 -1.22287524e+00
1.03969407e+00 -3.65642160e-01 -6.05975389e-01 2.23479465e-01
-5.02223134e-01 -2.05759332e-01 5.53682387e-01 -2.19074503e-01
-8.65985274e-01 -6.39459014e-01 -4.10091788e-01 1.79991305e-01
-6.27773166e-01 5.62320590e-01 -1.39206201e-01 3.46698076e-01
5.70226610e-01 -2.85383165e-01 -3.58104646e-01 -7.76730895e-01
2.43992969e-01 7.90705025e-01 3.13770980e-01 -5.68359852e-01
9.53231528e-02 -1.19027101e-01 -3.31781209e-01 -9.84580874e-01
-8.66322219e-01 -5.31022787e-01 -9.79694009e-01 -9.20762122e-02
1.22955072e+00 -2.26740837e-01 -5.56893587e-01 3.23594332e-01
-1.06063318e+00 -4.16125476e-01 -3.49397570e-01 4.65220839e-01
-4.45839852e-01 5.20659447e-01 -6.02102697e-01 -9.87848103e-01
-1.95161432e-01 -1.23755896e+00 8.37827206e-01 5.90854406e-01
-6.31055117e-01 -1.25917482e+00 1.26884490e-01 2.75464624e-01
2.31278822e-01 5.76228023e-01 8.83817077e-01 -8.83763433e-01
6.48851991e-02 7.78774992e-02 -2.36377083e-02 6.03027523e-01
-1.23408879e-03 4.43048716e-01 -1.05601597e+00 1.39868230e-01
3.23917419e-02 -1.16174750e-01 9.20266151e-01 7.35154390e-01
8.39474857e-01 1.11142866e-01 -1.02621995e-01 5.40018082e-03
1.25424337e+00 3.59422714e-01 5.18800199e-01 4.09324437e-01
2.31226861e-01 1.08801639e+00 9.48239207e-01 1.82598770e-01
1.08839646e-01 7.22365081e-01 -9.21179429e-02 2.22966820e-01
-3.37893695e-01 3.75623167e-01 4.19265419e-01 1.08039176e+00
-2.63548911e-01 -4.04770285e-01 -1.18329692e+00 6.62479937e-01
-1.72378349e+00 -8.04900944e-01 -5.17522812e-01 2.39791965e+00
9.82273400e-01 6.30729616e-01 5.39609134e-01 6.97197795e-01
1.16280091e+00 1.28725246e-01 2.26166233e-01 -8.68033409e-01
-3.85748774e-01 3.61363530e-01 3.55149470e-02 7.68773854e-01
-1.41809285e+00 9.34482992e-01 6.12962437e+00 1.01016366e+00
-9.42380369e-01 1.44499034e-01 8.59798551e-01 4.29457933e-01
-8.39130282e-02 -1.43610582e-01 -8.57250392e-01 6.15009010e-01
1.09836137e+00 -1.73725188e-01 4.71059009e-02 4.13839847e-01
1.45757303e-01 -5.82345426e-01 -1.06525385e+00 7.52203941e-01
1.20833717e-01 -8.73770773e-01 -3.65805060e-01 -2.56607294e-01
5.96080661e-01 -1.21118970e-01 -4.13687117e-02 5.14061213e-01
-1.98982939e-01 -9.82797086e-01 8.25113535e-01 4.23218608e-01
3.98136824e-01 -6.91956639e-01 1.01013350e+00 1.90039560e-01
-9.72667813e-01 3.39785844e-01 -2.82051384e-01 -1.20721869e-01
2.42482856e-01 7.94828236e-01 -5.71766675e-01 7.87183404e-01
5.67965746e-01 5.60222626e-01 -6.59563661e-01 1.19243884e+00
-2.75991321e-01 9.64794993e-01 -3.70632231e-01 -3.11489135e-01
3.88284683e-01 -5.40448666e-01 6.23638451e-01 1.52292025e+00
4.06785458e-02 -1.44082248e-01 5.13387769e-02 7.97676206e-01
2.64535397e-01 7.29036272e-01 -3.44607323e-01 -3.01870741e-02
2.50949472e-01 1.38137114e+00 -1.30389810e+00 -1.12529188e-01
-2.78808296e-01 3.86691362e-01 1.37926981e-01 1.24477362e-02
-8.38484943e-01 -4.59716290e-01 3.61718476e-01 -3.94785255e-02
1.00903697e-01 -4.30986322e-02 -5.75875819e-01 -6.60879910e-01
4.87325266e-02 -6.68499291e-01 4.85720128e-01 -3.18398029e-01
-1.26701581e+00 6.86715961e-01 3.41840714e-01 -8.55976760e-01
-6.48802370e-02 -5.71681678e-01 -4.17963475e-01 7.09471405e-01
-1.30712903e+00 -5.58310151e-01 -6.55986965e-02 -1.37579031e-02
7.57194638e-01 1.03955001e-01 6.71985269e-01 4.67797846e-01
-6.94538534e-01 4.14196998e-01 9.60718840e-03 2.19566807e-01
8.26972246e-01 -1.62017953e+00 1.74265355e-01 5.60338736e-01
1.04742691e-01 1.76802114e-01 8.06726277e-01 -3.06741863e-01
-3.23336333e-01 -6.25695050e-01 1.30436409e+00 -4.83880222e-01
6.93765879e-01 -1.62757695e-01 -9.84294057e-01 3.91250700e-01
5.57674825e-01 -4.46977407e-01 7.17872500e-01 1.04088113e-01
1.50436148e-01 1.35121971e-01 -1.08868742e+00 5.68258405e-01
6.71028256e-01 -1.59218520e-01 -9.55248177e-01 2.02954575e-01
5.33145607e-01 -3.63782257e-01 -1.02690530e+00 3.86421233e-01
1.74805775e-01 -1.62031901e+00 5.71128726e-01 -3.29493731e-01
2.21195221e-01 -6.45279512e-02 4.53846119e-02 -1.15617847e+00
2.10425302e-01 -5.86364806e-01 5.34301877e-01 1.86554432e+00
6.76560819e-01 -4.41679716e-01 3.24631840e-01 3.86798412e-01
-4.44618016e-01 -6.38787925e-01 -9.48415339e-01 -6.07404411e-01
4.34595108e-01 -4.02393073e-01 3.36978078e-01 9.28955495e-01
1.32437676e-01 4.17386562e-01 2.56529003e-01 -2.75700510e-01
1.77212507e-01 1.60242524e-02 3.59009683e-01 -1.50237834e+00
5.88879436e-02 -1.13577867e+00 -4.09760863e-01 -5.21127284e-01
1.60377517e-01 -8.78974199e-01 -2.03567911e-02 -1.41358304e+00
-2.05465958e-01 -5.01916409e-01 -2.30092049e-01 1.84812099e-01
-2.44995683e-01 2.06426337e-01 2.16366306e-01 -2.89852377e-02
-4.78808671e-01 1.72160387e-01 1.15971839e+00 1.65319860e-01
-2.71051705e-01 1.87529549e-01 -1.38431057e-01 8.79236042e-01
9.94027138e-01 -3.55138332e-01 -1.32746071e-01 -2.07546592e-01
1.48520544e-01 -8.36405717e-03 -7.18024001e-02 -1.10347402e+00
-1.42429963e-01 1.32344738e-01 -1.85029715e-01 -5.41111887e-01
-6.86271936e-02 -4.67763364e-01 -1.20025180e-01 2.91380465e-01
-3.88689458e-01 1.99742645e-01 2.64119476e-01 6.74959645e-02
-6.56370759e-01 -9.04785514e-01 8.25625777e-01 -1.11518480e-01
-6.02678478e-01 -2.81066298e-01 -1.70509994e-01 5.48638761e-01
1.22111166e+00 -4.55553293e-01 2.70450655e-02 8.44228938e-02
-1.17473865e+00 1.46199586e-02 2.56520420e-01 2.73017466e-01
-1.33010715e-01 -9.16122496e-01 -6.99613273e-01 -1.01354271e-01
-1.59081355e-01 1.97789654e-01 -4.71136905e-03 1.24695945e+00
-8.29973400e-01 6.48869574e-01 -1.66257173e-01 -5.96183717e-01
-1.24150944e+00 4.14084852e-01 9.66641158e-02 -5.63093305e-01
-1.79010555e-01 9.24612999e-01 -9.53864008e-02 -3.00689161e-01
3.86651278e-01 -7.54027665e-01 -8.26158583e-01 1.00261891e+00
1.67158037e-01 7.06831813e-01 1.49475008e-01 -8.98392677e-01
-3.06814134e-01 7.37081468e-01 1.10908464e-01 -2.52765208e-01
1.00497389e+00 -1.86565384e-01 -1.86708868e-01 1.17876315e+00
8.61234486e-01 7.89565593e-02 -7.67935574e-01 1.09825306e-01
7.90811121e-01 -2.73086224e-02 -2.43658587e-01 -7.13081896e-01
-5.83990037e-01 1.04128253e+00 4.09800231e-01 1.22735918e+00
8.69314432e-01 -5.88816442e-02 5.66478670e-01 7.98463523e-02
2.65647382e-01 -1.32691860e+00 -3.31125110e-01 7.36506522e-01
8.59932065e-01 -1.29057610e+00 -3.12405407e-01 -6.87710285e-01
-7.02929318e-01 1.16459179e+00 1.95563212e-01 -1.17379390e-02
8.29499066e-01 2.34197810e-01 2.38370635e-02 9.42476839e-02
-3.26169938e-01 -6.28868520e-01 2.54774153e-01 6.20436609e-01
1.07702482e+00 1.82303503e-01 -1.25736988e+00 8.01909745e-01
-4.76589322e-01 -3.73820692e-01 4.45571721e-01 6.36878014e-01
-6.07556164e-01 -1.32222724e+00 -5.06309271e-01 2.05410153e-01
-8.27039957e-01 -9.65332240e-02 -4.71450984e-01 1.21559286e+00
2.48800218e-01 1.18713427e+00 5.26889525e-02 -6.05742447e-03
5.02477407e-01 3.33578765e-01 3.91056210e-01 -9.31103766e-01
-1.18051887e+00 1.04028493e-01 4.00680721e-01 -1.02727078e-01
-8.26054215e-01 -1.09964919e+00 -1.21924698e+00 2.43381374e-02
-4.97652173e-01 3.99665534e-01 4.87304270e-01 9.52774465e-01
-2.63806432e-01 6.78138256e-01 1.98988587e-01 -7.55042791e-01
-2.77553558e-01 -1.53421581e+00 -7.67148376e-01 6.42052174e-01
-4.06601578e-02 -7.24847555e-01 -5.15093029e-01 3.75729591e-01] | [10.64479923248291, 9.82609748840332] |
ba721f20-e5c4-4563-aae8-c4966fea39cb | alignve-visual-entailment-recognition-based | 2211.08736 | null | https://arxiv.org/abs/2211.08736v1 | https://arxiv.org/pdf/2211.08736v1.pdf | AlignVE: Visual Entailment Recognition Based on Alignment Relations | Visual entailment (VE) is to recognize whether the semantics of a hypothesis text can be inferred from the given premise image, which is one special task among recent emerged vision and language understanding tasks. Currently, most of the existing VE approaches are derived from the methods of visual question answering. They recognize visual entailment by quantifying the similarity between the hypothesis and premise in the content semantic features from multi modalities. Such approaches, however, ignore the VE's unique nature of relation inference between the premise and hypothesis. Therefore, in this paper, a new architecture called AlignVE is proposed to solve the visual entailment problem with a relation interaction method. It models the relation between the premise and hypothesis as an alignment matrix. Then it introduces a pooling operation to get feature vectors with a fixed size. Finally, it goes through the fully-connected layer and normalization layer to complete the classification. Experiments show that our alignment-based architecture reaches 72.45\% accuracy on SNLI-VE dataset, outperforming previous content-based models under the same settings. | ['James Tin-Yau Kwok', 'Yuan Yan Tang', 'Lei He', 'Bo Liu', 'Jiayun Shen', 'Jie Gui', 'Jiuxin Cao', 'Biwei Cao'] | 2022-11-16 | null | null | null | null | ['visual-entailment'] | ['reasoning'] | [ 2.56125689e-01 -2.30868921e-01 -2.43216604e-02 -6.26326442e-01
-1.25659272e-01 -3.70828301e-01 1.02531660e+00 -5.63238785e-02
-3.47999871e-01 2.03851268e-01 2.81541139e-01 -5.09413958e-01
2.57738054e-01 -7.41524041e-01 -9.24204826e-01 -2.46578380e-01
5.89636922e-01 3.38652683e-03 3.24973315e-01 2.53015086e-02
4.92815286e-01 -1.85163859e-02 -1.84716809e+00 9.69784260e-01
4.98487562e-01 1.40142095e+00 3.86487633e-01 5.29459774e-01
-6.63801849e-01 1.22608960e+00 -3.44498217e-01 -4.76063728e-01
-6.25086427e-02 -7.19359040e-01 -1.17544019e+00 2.02277213e-01
8.55661988e-01 -4.41773593e-01 -3.57132405e-01 1.04851472e+00
9.82018635e-02 -1.76381111e-01 9.35471594e-01 -1.71336329e+00
-1.29048073e+00 2.92931288e-01 -6.25118613e-01 4.86279309e-01
5.48314869e-01 4.87720743e-02 1.34849334e+00 -1.39158738e+00
5.34086823e-01 1.38497102e+00 4.43862110e-01 2.55530566e-01
-8.32214773e-01 -4.24811929e-01 4.50270265e-01 9.55838621e-01
-1.50428069e+00 -3.90599698e-01 8.39366674e-01 -4.38992262e-01
1.21742165e+00 4.49638873e-01 6.65616751e-01 7.99710393e-01
5.40642627e-02 9.54770029e-01 1.20186675e+00 -5.38150907e-01
-1.36064887e-01 1.90559447e-01 1.66121960e-01 9.52048063e-01
1.40733607e-02 -4.76345509e-01 -5.32715678e-01 3.93316001e-01
4.22070682e-01 1.78898349e-01 -4.67196435e-01 -2.75540054e-01
-1.33007789e+00 6.46462560e-01 7.99786448e-01 4.21610653e-01
-2.85495788e-01 1.55896991e-01 3.06335121e-01 1.89533249e-01
1.48091659e-01 -3.59598324e-02 -6.15128241e-02 5.75770915e-01
-7.97030389e-01 1.40117660e-01 5.57827353e-01 9.27639484e-01
7.71023631e-01 -2.02756017e-01 -4.44737375e-01 4.16737497e-01
7.37625182e-01 3.56044680e-01 5.04625201e-01 -3.83670062e-01
7.11206317e-01 1.19416308e+00 -4.08370122e-02 -1.39322090e+00
-2.68585503e-01 -2.24674493e-01 -7.41428912e-01 1.69317886e-01
2.49021918e-01 2.05011383e-01 -8.62112105e-01 1.59113109e+00
4.96450990e-01 4.42434251e-01 2.65309602e-01 1.07837093e+00
1.58100379e+00 7.64166892e-01 2.23768905e-01 -8.07895735e-02
1.84337246e+00 -1.03516662e+00 -9.84474957e-01 -5.32775402e-01
3.20847094e-01 -9.72834289e-01 1.49997020e+00 -1.03369959e-01
-8.10729563e-01 -7.64015257e-01 -1.33315718e+00 -5.41965425e-01
-6.54108405e-01 3.53827238e-01 3.84862900e-01 1.99916825e-01
-7.99008071e-01 1.97748616e-02 -6.53830767e-02 -3.19276780e-01
6.15201950e-01 -5.79820499e-02 -4.53394294e-01 -4.20491874e-01
-1.14177501e+00 1.03120649e+00 5.67549288e-01 5.56018114e-01
-6.81463599e-01 -2.86716282e-01 -9.91383553e-01 1.39204100e-01
4.29532319e-01 -9.83225286e-01 1.00734925e+00 -1.18168497e+00
-9.54393446e-01 1.15161920e+00 -5.83933890e-01 -1.23069070e-01
3.85642678e-01 -1.15467288e-01 -4.98499066e-01 1.00164443e-01
1.40475810e-01 3.39856535e-01 1.05223584e+00 -1.33564270e+00
-6.26315713e-01 -4.33970124e-01 3.28477025e-01 4.99808103e-01
-1.15474015e-01 3.43710966e-02 -6.00843906e-01 -1.39121473e-01
5.24457633e-01 -2.89193362e-01 3.86143267e-01 1.95459768e-01
-5.46583593e-01 -6.54309332e-01 1.14633453e+00 -6.48315728e-01
1.25493908e+00 -2.08392930e+00 9.14080665e-02 4.97599468e-02
5.88082850e-01 8.66476223e-02 -1.04079299e-01 2.62755990e-01
-2.19364583e-01 6.54856954e-03 -9.44216251e-02 -2.75560170e-01
2.12122992e-01 1.73435524e-01 -4.67425704e-01 4.33994740e-01
3.86595786e-01 1.21639347e+00 -4.50542659e-01 -8.47341716e-01
2.41803393e-01 1.92159101e-01 -2.19204426e-01 4.30570751e-01
-2.84568757e-01 -1.21886343e-01 -1.97261468e-01 5.63394547e-01
8.96606028e-01 -6.42890215e-01 1.85668766e-01 -8.08842719e-01
-1.19730666e-01 -5.07236645e-02 -1.33017826e+00 1.63532257e+00
-1.70424506e-01 7.96622336e-01 -2.24495873e-01 -1.28166103e+00
8.84213686e-01 2.53906578e-01 -1.69981107e-01 -9.26838100e-01
3.30758631e-01 -1.29688188e-01 -2.36084722e-02 -1.13822556e+00
2.76798278e-01 -3.75231653e-01 1.56157970e-01 2.93673515e-01
-4.19425964e-03 9.57872123e-02 -1.53576970e-01 3.64847690e-01
4.21492457e-01 2.20503032e-01 4.58020836e-01 -1.90395536e-03
9.60072458e-01 -1.48641944e-01 2.38953501e-01 5.75137019e-01
-3.67245436e-01 4.39577788e-01 5.58587730e-01 -5.73200166e-01
-9.50724602e-01 -1.21342516e+00 2.87263468e-02 8.88521671e-01
5.71811199e-01 -4.99678046e-01 -6.40411079e-01 -7.66693890e-01
-2.75147520e-02 6.09953105e-01 -7.95822024e-01 -8.99842903e-02
-2.75577188e-01 -4.02693689e-01 4.00439113e-01 5.88685095e-01
9.61552501e-01 -1.14439225e+00 -5.57753921e-01 -3.29206854e-01
-6.66844249e-01 -1.32050526e+00 -2.36416429e-01 -3.28636408e-01
-3.00260603e-01 -1.30002105e+00 -3.07920098e-01 -1.11893117e+00
6.43604577e-01 2.93619245e-01 9.73601878e-01 3.60614300e-01
-3.19089472e-01 2.96179503e-01 -1.00069292e-01 -4.12610739e-01
5.63929863e-02 -5.67186236e-01 -2.07964867e-01 4.35849369e-01
7.87358999e-01 -8.22050571e-02 -7.02988744e-01 1.41953990e-01
-8.66401255e-01 4.51804787e-01 7.10893095e-01 5.94383359e-01
8.16115320e-01 -2.75180459e-01 4.20737594e-01 -5.19713938e-01
5.22785366e-01 -4.56683636e-01 -5.56535535e-02 7.83730805e-01
-4.11277950e-01 1.75856680e-01 5.90006828e-01 -2.89978713e-01
-1.23955750e+00 -1.30109519e-01 -7.82314967e-03 -4.08836603e-01
-2.67219305e-01 5.43677092e-01 -5.77314854e-01 2.54531443e-01
4.06704426e-01 3.87901485e-01 -6.27531111e-02 -1.48439899e-01
7.85803795e-01 8.59366775e-01 6.74020410e-01 -1.89226046e-01
5.74610054e-01 6.01140559e-01 5.62998615e-02 -7.17103302e-01
-1.16244328e+00 -5.25420666e-01 -5.95072389e-01 -4.10095960e-01
1.11462677e+00 -7.92313099e-01 -1.04769444e+00 3.43096614e-01
-1.55400658e+00 2.46514693e-01 1.71238616e-01 3.98531139e-01
-4.15327400e-01 6.63443089e-01 -3.17013711e-01 -6.59089088e-01
-3.74296844e-01 -8.88511598e-01 7.70122528e-01 3.58076930e-01
-8.14551562e-02 -6.69598341e-01 -2.81342447e-01 5.38693964e-01
1.69243366e-01 4.81496602e-02 1.14172709e+00 -5.47857642e-01
-5.88699818e-01 3.06455791e-03 -1.04596913e+00 2.91177154e-01
-2.00288892e-01 5.39069623e-03 -9.98865783e-01 1.44533977e-01
1.97422311e-01 -4.29072946e-01 8.98618579e-01 1.44259989e-01
1.17798460e+00 -3.63913924e-01 -1.34981364e-01 5.03727674e-01
1.66278422e+00 -6.02697320e-02 6.61245346e-01 1.80656582e-01
1.03461063e+00 6.18745506e-01 5.56465030e-01 2.08518535e-01
7.16487765e-01 3.15527022e-01 7.32426047e-01 -1.02812566e-01
-3.72919738e-01 -4.22527432e-01 7.31744543e-02 6.57457530e-01
-7.79090822e-02 -2.31160313e-01 -9.25143421e-01 3.63470405e-01
-2.00515532e+00 -1.18645966e+00 -5.36004007e-01 1.71854794e+00
5.58492661e-01 -1.72855165e-02 -3.16481709e-01 1.03727281e-01
6.91354990e-01 2.76098967e-01 -5.27035356e-01 -4.32487488e-01
-2.23881334e-01 -1.42555535e-01 -2.04099029e-01 3.75849873e-01
-1.15975404e+00 8.88669491e-01 5.71669483e+00 6.22102082e-01
-8.81856740e-01 1.65967811e-02 4.12522495e-01 1.70726836e-01
-4.48826432e-01 1.31353080e-01 -6.34149075e-01 1.44458681e-01
3.14246655e-01 1.44123420e-01 7.48792589e-02 5.89277923e-01
-1.01261251e-01 -1.49170503e-01 -1.27313781e+00 1.31319106e+00
8.71355534e-01 -1.41044807e+00 4.31497216e-01 -3.81781399e-01
2.97895581e-01 -3.70101184e-01 8.73590782e-02 1.77159816e-01
-5.00959873e-01 -1.36321354e+00 9.41052616e-01 8.82757246e-01
7.16912746e-01 -6.13667786e-01 8.27913165e-01 3.53147238e-01
-1.68234849e+00 3.75096910e-02 -5.08745551e-01 -4.21970785e-01
5.21880994e-03 1.53205302e-02 -6.49417579e-01 5.46354473e-01
7.75009632e-01 8.06522965e-01 -8.07676494e-01 8.01281393e-01
-5.02379656e-01 1.59573525e-01 -3.11491266e-03 -2.37129703e-01
2.32413203e-01 4.34762761e-02 1.88439012e-01 1.06629729e+00
-1.12878659e-03 -7.95030072e-02 4.76964526e-02 1.23526537e+00
-1.94551617e-01 3.47865760e-01 -6.67972028e-01 2.13173568e-01
2.49719515e-01 1.20205748e+00 -5.17453015e-01 -4.82571810e-01
-8.27069759e-01 1.37112546e+00 6.74851835e-01 4.56604987e-01
-1.01292849e+00 -5.68364859e-01 3.42966735e-01 -2.54378915e-01
4.66648877e-01 5.46005256e-02 -2.26811528e-01 -1.10368741e+00
3.93724948e-01 -6.25619769e-01 5.67919016e-01 -1.28100359e+00
-1.33785915e+00 4.37076956e-01 -1.79970369e-01 -1.23709095e+00
1.54189080e-01 -1.00124884e+00 -8.52770090e-01 9.49157298e-01
-1.73497844e+00 -1.66311288e+00 -6.14770949e-01 8.75994802e-01
4.91622359e-01 1.68216899e-01 6.60107076e-01 8.10331404e-02
-6.89463317e-01 4.35244113e-01 -4.30117488e-01 2.72079021e-01
5.61551034e-01 -1.07270730e+00 4.45591621e-02 1.19318914e+00
4.52833503e-01 6.44263268e-01 6.51676834e-01 -4.34163690e-01
-1.26345265e+00 -7.34501719e-01 1.31882715e+00 -5.24658740e-01
5.18367767e-01 -2.97912329e-01 -9.73471403e-01 8.09844971e-01
6.10946715e-01 2.44445801e-01 7.84489334e-01 -1.38587266e-01
-8.71982098e-01 2.75967177e-02 -8.50342870e-01 6.80189669e-01
1.15587449e+00 -9.71747458e-01 -1.22248328e+00 1.26196027e-01
6.31803632e-01 -1.62207261e-01 -5.96412241e-01 1.35663971e-01
5.90932310e-01 -9.21816409e-01 1.01897979e+00 -1.04514575e+00
8.65340710e-01 -8.50935638e-01 -4.22715962e-01 -7.32337773e-01
-2.73114204e-01 2.68857002e-01 -8.44282061e-02 1.20579767e+00
4.60185558e-01 -1.83754072e-01 3.83918762e-01 3.90808493e-01
1.69161253e-03 -6.37305975e-01 -8.76197696e-01 -3.77224535e-01
-3.05834293e-01 -4.41252261e-01 6.00158572e-01 1.07630026e+00
1.39087290e-01 9.92094338e-01 -3.38260174e-01 2.84336835e-01
4.52036142e-01 6.77015066e-01 5.98711789e-01 -8.99990261e-01
-1.38506547e-01 -4.05916721e-01 -5.24461329e-01 -1.08941972e+00
1.52420834e-01 -1.15057516e+00 5.15629770e-03 -2.09382963e+00
5.07014275e-01 2.47021794e-01 -2.38062114e-01 3.58535767e-01
-4.53303427e-01 2.51635104e-01 2.93035537e-01 9.04554054e-02
-8.88402998e-01 5.89343131e-01 1.39620018e+00 -2.26220265e-01
2.73318797e-01 -6.84729755e-01 -5.69962621e-01 1.00098324e+00
6.77828848e-01 7.88352937e-02 -5.49201310e-01 -6.16976917e-01
3.99674416e-01 -3.09698761e-01 9.02986228e-01 -7.88581312e-01
3.66168231e-01 -2.61421472e-01 7.06269681e-01 -7.75005519e-01
1.83824420e-01 -9.71047580e-01 -1.73812121e-01 3.37484539e-01
-2.00228393e-01 3.38223249e-01 1.09538622e-02 7.74867177e-01
-4.14748698e-01 -1.47665203e-01 3.66754323e-01 -8.53366256e-02
-1.45325232e+00 2.70465195e-01 -4.67896275e-02 3.51080507e-01
1.22157323e+00 -3.13564003e-01 -7.25927413e-01 -1.06302686e-01
-4.55129653e-01 2.27755636e-01 -4.47628684e-02 3.97834539e-01
1.32058787e+00 -1.45134699e+00 -6.52848244e-01 1.86535329e-01
6.33157909e-01 -1.12619989e-01 5.79682946e-01 8.65154982e-01
-4.45235193e-01 2.50029713e-01 -3.24183464e-01 -5.86623013e-01
-1.54183686e+00 7.64117122e-01 3.96239877e-01 2.32455552e-01
-5.81892312e-01 9.81950104e-01 3.34542811e-01 -3.45492549e-02
1.33465469e-01 -3.62179697e-01 -7.18977869e-01 1.05096444e-01
8.17620575e-01 -4.99688014e-02 -3.94004673e-01 -1.11457634e+00
-6.39926434e-01 6.95248365e-01 3.34655680e-02 7.43397698e-02
7.83129752e-01 -3.46325338e-01 -3.58429343e-01 6.05331123e-01
1.46698272e+00 -3.20912331e-01 -9.35502172e-01 -5.08422256e-01
-1.45507142e-01 -4.21347946e-01 -8.22150335e-02 -5.33245862e-01
-7.82261789e-01 1.19315743e+00 5.02743244e-01 8.15387592e-02
8.90142500e-01 5.16309738e-01 4.40200746e-01 2.86090285e-01
-4.61803347e-01 -9.36803818e-01 2.96760917e-01 3.72315645e-01
1.14310634e+00 -1.43769312e+00 -6.03148388e-03 -5.25078654e-01
-8.11850727e-01 1.10805583e+00 9.75074053e-01 4.93643694e-02
6.04709625e-01 -2.41524935e-01 3.13889794e-02 -5.62888920e-01
-5.36484420e-01 -5.92277646e-01 7.96298504e-01 5.07395089e-01
4.26417053e-01 8.62612352e-02 -3.20366412e-01 7.42651165e-01
-1.66120101e-02 -1.74761102e-01 1.20289534e-01 6.18232608e-01
-5.68717062e-01 -4.74686235e-01 -1.33079767e-01 4.58339095e-01
-1.13375753e-01 -3.63237798e-01 -5.03707826e-01 7.08557427e-01
4.98298138e-01 1.10270000e+00 4.11243826e-01 -4.63542789e-01
3.74933034e-01 4.14984465e-01 5.19753456e-01 -2.86541015e-01
-2.10676014e-01 -4.20794040e-01 8.00380632e-02 -6.43591285e-01
-5.78465223e-01 -3.37989181e-01 -1.56861770e+00 -2.33111084e-01
-2.60685861e-01 -3.37968439e-01 4.86132085e-01 1.50649369e+00
7.56687298e-02 5.94548464e-01 2.54094571e-01 -1.26071036e-01
-3.22739094e-01 -7.84243226e-01 -3.58364969e-01 9.36255515e-01
2.88741142e-01 -6.00203335e-01 -3.90916049e-01 1.36542723e-01] | [10.727478981018066, 1.6515034437179565] |
b29d0ed7-a0af-4957-98da-82326ebed86c | privacy-preserving-deep-learning-model-for | 2209.04445 | null | https://arxiv.org/abs/2209.04445v2 | https://arxiv.org/pdf/2209.04445v2.pdf | Privacy-Preserving Deep Learning Model for Covid-19 Disease Detection | Recent studies demonstrated that X-ray radiography showed higher accuracy than Polymerase Chain Reaction (PCR) testing for COVID-19 detection. Therefore, applying deep learning models to X-rays and radiography images increases the speed and accuracy of determining COVID-19 cases. However, due to Health Insurance Portability and Accountability (HIPAA) compliance, the hospitals were unwilling to share patient data due to privacy concerns. To maintain privacy, we propose differential private deep learning models to secure the patients' private information. The dataset from the Kaggle website is used to evaluate the designed model for COVID-19 detection. The EfficientNet model version was selected according to its highest test accuracy. The injection of differential privacy constraints into the best-obtained model was made to evaluate performance. The accuracy is noted by varying the trainable layers, privacy loss, and limiting information from each sample. We obtained 84\% accuracy with a privacy loss of 10 during the fine-tuning process. | ['Xiali Hei', 'Sonya Hsu', 'Vijay Srinivas Tida Sai Venkatesh Chilukoti'] | 2022-09-07 | null | null | null | null | ['privacy-preserving-deep-learning', 'privacy-preserving-deep-learning'] | ['methodology', 'natural-language-processing'] | [-8.30622464e-02 -4.46086526e-02 -1.38195261e-01 -5.84032655e-01
-1.02701259e+00 -6.56351388e-01 -6.81086034e-02 2.92522401e-01
-9.38814342e-01 8.99691224e-01 -9.95809212e-02 -6.88735366e-01
-9.91522819e-02 -7.51861095e-01 -7.62687743e-01 -6.45782948e-01
-1.49176091e-01 3.71991426e-01 -3.68264288e-01 5.62519968e-01
-1.19686574e-01 7.85164475e-01 -7.09298909e-01 4.96312112e-01
5.92284679e-01 1.05181670e+00 -1.41568288e-01 6.53097093e-01
5.23462534e-01 5.51790297e-01 -5.35916388e-01 -7.65989065e-01
8.34895432e-01 -2.38748342e-01 -5.65543771e-01 -4.48418438e-01
2.95691639e-01 -1.03848171e+00 -3.05980504e-01 7.86189139e-01
6.42494440e-01 -2.37325445e-01 3.33372504e-01 -1.17921281e+00
-5.38125813e-01 3.69340837e-01 -4.81689185e-01 2.05221072e-01
9.56871957e-02 3.22395325e-01 8.46465766e-01 -3.80554676e-01
8.33753109e-01 6.42561138e-01 7.54686713e-01 7.03111649e-01
-1.10940659e+00 -1.09095621e+00 -4.02085990e-01 -1.33784130e-01
-1.45505023e+00 -6.44896599e-03 2.07329065e-01 -5.61092377e-01
8.34947228e-01 3.32054645e-01 8.36047411e-01 1.35335577e+00
8.65014970e-01 4.19541299e-01 1.10228813e+00 1.81542456e-01
2.39001125e-01 4.14895624e-01 2.36086950e-01 6.15354478e-01
8.13869119e-01 1.95006862e-01 -8.00744146e-02 -8.02872539e-01
7.98655033e-01 4.67215925e-01 -2.03300536e-01 -3.45226079e-01
-7.71688223e-01 9.43844855e-01 1.70655638e-01 -1.17274016e-01
-3.74289185e-01 1.08761631e-01 6.12522781e-01 2.75917977e-01
1.12187788e-01 6.20086133e-01 -6.22506976e-01 1.65299565e-01
-5.79847276e-01 4.35708374e-01 5.71013808e-01 7.80606031e-01
2.86748856e-01 -2.48529151e-01 -3.60409498e-01 4.83019829e-01
2.17558339e-01 4.33044046e-01 3.27553362e-01 -7.04709888e-01
5.46742857e-01 4.73889083e-01 2.46054739e-01 -8.44372272e-01
-4.55184549e-01 -6.01711035e-01 -8.33952010e-01 -1.10701062e-01
3.72605443e-01 -6.71918929e-01 -1.04648328e+00 1.66313148e+00
-8.84730369e-02 -1.48965633e-02 9.73047540e-02 9.87543344e-01
4.64946032e-01 3.33434939e-01 1.87370256e-01 3.03924717e-02
1.67367971e+00 -1.59578964e-01 -6.46532059e-01 1.05029391e-02
8.02588940e-01 -4.21118736e-01 7.02851653e-01 2.82990843e-01
-8.62170398e-01 9.48443711e-02 -1.00193882e+00 6.33873120e-02
-2.22204328e-01 1.28248036e-01 6.41205430e-01 1.09258783e+00
-7.46447086e-01 2.41923362e-01 -1.10117638e+00 -2.85835147e-01
9.75746572e-01 5.59686899e-01 -6.00003958e-01 -8.25053006e-02
-1.35319710e+00 6.09252870e-01 1.00834504e-01 5.66362292e-02
-1.06302059e+00 -7.14749813e-01 -5.55439651e-01 2.03520060e-01
1.21437401e-01 -8.58411551e-01 8.30237091e-01 -2.63397664e-01
-7.49191999e-01 1.10091877e+00 3.34902316e-01 -1.03151953e+00
8.59779477e-01 -1.90150693e-01 -2.54722059e-01 2.78773516e-01
-6.08284250e-02 4.79964465e-01 5.12782276e-01 -9.19773221e-01
-5.38194776e-01 -6.43581510e-01 -9.16064307e-02 -3.63877654e-01
-1.50140658e-01 8.25728178e-02 -1.39508888e-01 -5.75240493e-01
-2.30328009e-01 -1.14466572e+00 -2.59028703e-01 -1.64120436e-01
-4.09599125e-01 3.98918390e-01 7.44475067e-01 -8.28605175e-01
9.08198476e-01 -2.41619635e+00 -8.73239338e-01 5.49521565e-01
4.99025792e-01 3.32643121e-01 -6.27190843e-02 1.21772766e-01
-4.96317036e-02 4.98763800e-01 -2.38293409e-01 -8.14740732e-02
-1.60090342e-01 -5.53634912e-02 -9.29098502e-02 8.27902317e-01
2.61342704e-01 7.93490171e-01 -3.90170425e-01 -4.72931564e-01
-1.24866292e-01 5.20740390e-01 -9.33177054e-01 2.77787715e-01
1.31453484e-01 3.61069977e-01 -6.91529870e-01 6.82327211e-01
7.95061290e-01 -4.99879599e-01 3.43952537e-01 1.02352262e-01
4.74490166e-01 -3.21048908e-02 -7.91462183e-01 1.07065713e+00
-1.27611607e-02 4.24173415e-01 1.35665044e-01 -2.18846321e-01
9.08967793e-01 4.90820765e-01 6.73870027e-01 -4.02472377e-01
4.29096311e-01 5.08708768e-02 1.57497525e-01 -5.81993699e-01
3.65112275e-01 -1.26070529e-01 -2.22826470e-02 3.84618998e-01
-6.35244071e-01 3.35104227e-01 -4.42265421e-01 3.96656953e-02
1.20031881e+00 -3.82047772e-01 4.79167737e-02 -3.86221223e-02
3.17325965e-02 2.04482581e-02 7.67163217e-01 8.25288653e-01
-4.96276706e-01 7.38801181e-01 5.96550882e-01 -5.22162795e-01
-8.62118781e-01 -1.09029627e+00 -4.78305966e-01 5.14798522e-01
-2.95382380e-01 6.40321448e-02 -7.10057735e-01 -6.33745253e-01
2.16584146e-01 5.90748906e-01 -6.96783364e-01 -5.41893616e-02
-5.97242892e-01 -8.31222773e-01 8.47938061e-01 5.53481340e-01
4.11156744e-01 -6.22695744e-01 -1.19189882e+00 -5.50897829e-02
-5.26718199e-02 -9.16050971e-01 -5.44729233e-01 2.99988985e-01
-5.83849728e-01 -1.41998363e+00 -7.54105747e-01 -5.36914766e-01
5.79236865e-01 -2.27833435e-01 4.20279980e-01 -4.29029912e-02
-6.49077892e-01 2.31727347e-01 7.32041001e-02 -6.10021174e-01
-3.78831863e-01 2.50829011e-01 -1.44043788e-01 -2.30774105e-01
6.76015258e-01 7.48185292e-02 -9.16456401e-01 3.81151997e-02
-8.72658849e-01 -5.01867056e-01 2.99031854e-01 8.43569040e-01
7.12577879e-01 -1.26052767e-01 4.50875610e-01 -1.13556123e+00
7.78396606e-01 -5.28936148e-01 -8.14250946e-01 -4.37331997e-04
-8.37157607e-01 -3.08203310e-01 4.18620080e-01 -1.25590473e-01
-8.41008067e-01 1.02600172e-01 -8.02970231e-02 -4.95687068e-01
-1.40671358e-02 2.03247115e-01 1.24517009e-01 3.97141092e-03
3.44991148e-01 -1.24940753e-01 3.53754550e-01 -2.96184301e-01
-3.57136101e-01 7.13101745e-01 2.42888436e-01 -1.68525755e-01
1.66314662e-01 3.74760509e-01 -3.10166068e-02 -3.87907624e-01
-4.71418709e-01 -1.36507198e-01 3.66364159e-02 3.33297104e-01
1.31926310e+00 -1.12178934e+00 -1.29099858e+00 3.39034259e-01
-1.01122296e+00 3.22106868e-01 -4.53202575e-02 1.01083255e+00
-4.53905106e-01 7.33704120e-02 -1.03804302e+00 -6.03438675e-01
-5.40306270e-01 -1.39584970e+00 7.07504034e-01 -1.98236763e-01
-3.75778466e-01 -4.95580286e-01 -8.37162044e-03 4.24082905e-01
3.14943373e-01 6.06804907e-01 1.02891529e+00 -1.12753594e+00
-8.51887047e-01 -7.87149370e-01 -4.13101166e-01 4.76163208e-01
-4.73760627e-02 -1.35261580e-01 -1.06455684e+00 -5.92377961e-01
3.82711262e-01 -3.35696459e-01 5.06676316e-01 6.23567283e-01
1.45229733e+00 -4.76738095e-01 -4.05265659e-01 9.37485874e-01
1.59697616e+00 6.73542023e-01 6.57870591e-01 4.44567233e-01
4.01912481e-01 1.83258653e-01 2.98717648e-01 8.61717701e-01
1.58156216e-01 8.83509964e-02 5.20059347e-01 -1.61313981e-01
5.74287057e-01 -1.29025176e-01 -1.22259684e-01 -4.95998412e-02
1.39794856e-01 -3.34539026e-01 -1.23728597e+00 4.18893933e-01
-1.27407408e+00 -7.69320905e-01 -4.30225097e-02 2.12986541e+00
5.01172125e-01 -5.81550375e-02 -6.83460310e-02 -4.43543732e-01
6.17888391e-01 -5.85542507e-02 -6.64872527e-01 -4.79709685e-01
-1.55075505e-01 -5.75035177e-02 1.04978251e+00 1.04597926e-01
-9.04875636e-01 2.97726244e-01 7.02348995e+00 2.46458039e-01
-1.04858756e+00 1.13770403e-01 1.15657079e+00 -5.00284791e-01
-2.34343588e-01 -5.23369253e-01 -7.39735246e-01 3.92946422e-01
8.29639077e-01 -2.65039355e-01 1.02777192e-02 9.00669754e-01
1.26051471e-01 1.42926857e-01 -1.13937569e+00 8.81123483e-01
-1.88004270e-01 -1.63166583e+00 -1.62260279e-01 4.89635617e-01
4.78500098e-01 3.24121475e-01 3.63390863e-01 -6.95608556e-02
3.67019743e-01 -1.08519256e+00 2.08777875e-01 4.93701637e-01
9.56665337e-01 -1.14494538e+00 1.27640045e+00 -2.62774993e-02
-5.44069648e-01 -3.16262484e-01 -1.53030008e-01 3.75287741e-01
-7.71282390e-02 9.29063708e-02 -1.41405404e+00 1.63820341e-01
7.74836481e-01 1.11979671e-01 -3.48492235e-01 7.95891941e-01
2.84277320e-01 6.09053552e-01 -3.58828932e-01 4.61591408e-02
3.70097280e-01 -5.63742705e-02 3.38363498e-01 1.14883745e+00
3.33358288e-01 9.03987810e-02 -1.04573697e-01 7.33148515e-01
-2.72779137e-01 1.06185213e-01 -6.83020651e-01 -1.07379526e-01
3.93651128e-01 6.57214940e-01 -2.93429136e-01 3.28772850e-02
-2.81619996e-01 5.16419709e-01 -6.66233227e-02 2.38366842e-01
-9.77182746e-01 -3.76191974e-01 9.93152142e-01 4.18670297e-01
2.69816130e-01 2.20624804e-01 -5.67842603e-01 -7.09702551e-01
-1.11952402e-01 -1.01112604e+00 8.31763744e-01 -4.88913357e-01
-1.33213592e+00 6.55985236e-01 -2.76387315e-02 -1.15192330e+00
-4.54796195e-01 -4.78376657e-01 9.02497321e-02 1.05686879e+00
-1.06185567e+00 -8.27818692e-01 5.48851490e-02 3.99206251e-01
-2.59937078e-01 -4.47143942e-01 9.35403407e-01 2.46248469e-01
-6.64179862e-01 1.16268313e+00 3.56122583e-01 4.50128078e-01
5.51313460e-01 -6.34905040e-01 3.07829112e-01 7.70574450e-01
-6.97732389e-01 9.75154638e-01 4.15936232e-01 -7.94230998e-01
-1.30533826e+00 -1.25318098e+00 8.58627677e-01 -6.83443248e-01
4.46352512e-01 -4.69000578e-01 -8.42713475e-01 1.07644129e+00
-5.58639914e-02 6.48426823e-03 1.16907835e+00 -2.89920598e-01
-3.38549107e-01 -1.44711137e-01 -1.92039835e+00 4.80620861e-01
5.05310416e-01 -7.08702564e-01 -2.52372324e-01 2.44230539e-01
9.08324063e-01 -4.47374225e-01 -1.17522418e+00 1.16164364e-01
8.19266200e-01 -7.65181005e-01 8.17329228e-01 -8.15537274e-01
3.35462749e-01 1.17054336e-01 -2.69798756e-01 -6.06965482e-01
-2.38875568e-01 -4.28270131e-01 4.13208306e-01 6.85802817e-01
5.50630391e-01 -1.05759180e+00 1.19283009e+00 1.11687922e+00
5.22315383e-01 -9.29330289e-01 -9.37793195e-01 -5.89054406e-01
-8.34132582e-02 -4.41392928e-01 1.19100916e+00 1.10911286e+00
-2.76435226e-01 -2.63730973e-01 -6.42402053e-01 3.86885285e-01
6.05409682e-01 -4.39034477e-02 6.44736409e-01 -6.80908442e-01
-4.73494530e-01 1.80593118e-01 -4.69633609e-01 -2.69578338e-01
-2.20772699e-01 -8.38860691e-01 -3.31706554e-01 -1.09311843e+00
4.10966367e-01 -4.82252568e-01 -6.38006747e-01 6.45102859e-01
-2.96089053e-02 2.07793236e-01 3.29567909e-01 5.24320686e-03
1.32098272e-01 2.33979803e-02 1.17230785e+00 -2.54218113e-02
-1.16867356e-01 1.93225220e-01 -7.85613537e-01 3.55401099e-01
1.10363579e+00 -8.12150836e-01 -3.34438860e-01 -5.88488102e-01
2.43580818e-01 4.77956206e-01 5.14589548e-01 -6.32475674e-01
-1.09305814e-01 -1.64972872e-01 6.76534295e-01 -8.50064278e-01
3.17001194e-01 -1.26331866e+00 5.52733123e-01 1.00367653e+00
-4.51008767e-01 2.21200734e-01 3.51785392e-01 5.27428448e-01
1.01100750e-01 7.85800442e-02 8.38583410e-01 -6.84535205e-02
4.34230547e-03 5.93304098e-01 -4.54734147e-01 -1.64277911e-01
1.10849202e+00 -2.86747068e-01 -3.99969935e-01 -1.42678574e-01
-6.12652481e-01 2.07767099e-01 6.27887428e-01 9.92130041e-02
7.83266425e-01 -7.54850805e-01 -7.48532534e-01 5.94388247e-01
2.09860474e-01 -2.10236460e-01 3.64397198e-01 5.11687279e-01
-8.88783038e-01 4.35946941e-01 -3.61054420e-01 -4.94468987e-01
-1.43960512e+00 6.56978548e-01 3.75301689e-01 -3.42675418e-01
-7.29086220e-01 6.61774755e-01 2.37807840e-01 -1.89316228e-01
3.15578252e-01 -3.13902140e-01 3.02276134e-01 -3.08810204e-01
4.87334430e-01 4.62527126e-01 8.89548361e-02 -1.04966238e-01
-7.09401548e-01 -9.70696583e-02 -5.63812494e-01 1.86206609e-01
1.35819101e+00 5.29364534e-02 3.61414254e-02 -1.15484171e-01
1.62559271e+00 -1.16567083e-01 -1.05152678e+00 3.37310374e-01
-3.82118255e-01 -7.87781894e-01 -1.46575451e-01 -8.53160441e-01
-1.17084754e+00 5.52916050e-01 1.16346157e+00 -1.72955021e-01
9.97362494e-01 -2.54410446e-01 8.46823215e-01 5.28209567e-01
4.39197510e-01 -8.00567806e-01 -7.24113584e-01 2.04797521e-01
3.72370839e-01 -1.17252529e+00 1.59110561e-01 -5.67479022e-02
-6.34059846e-01 4.98745441e-01 5.28189778e-01 -2.68305745e-02
6.19635701e-01 5.08406639e-01 3.27145964e-01 -2.34803885e-01
-7.36535490e-01 8.19466054e-01 -3.01900148e-01 7.53613949e-01
2.56147683e-01 3.25191855e-01 -3.80066603e-01 7.01994777e-01
-2.30266556e-01 4.39023197e-01 5.21434426e-01 1.10050118e+00
2.06597000e-01 -7.78283119e-01 -3.84245008e-01 9.99099493e-01
-1.02795470e+00 9.63776484e-02 -2.51027822e-01 9.46416557e-01
1.37912109e-01 4.26506460e-01 3.56130540e-01 -2.75215179e-01
2.08840430e-01 -1.05561465e-01 9.28912833e-02 -2.82944083e-01
-9.92947638e-01 -1.07546344e-01 6.77692667e-02 -5.25826097e-01
1.38219655e-01 -6.19239688e-01 -1.24797237e+00 -4.34080958e-01
-4.51811962e-02 1.24352388e-02 5.97117960e-01 4.59265977e-01
6.71803594e-01 3.82803023e-01 6.06221378e-01 3.91135991e-01
-7.40226924e-01 -3.29831779e-01 -7.97280252e-01 3.57603341e-01
5.60753703e-01 8.16531181e-02 -1.31309435e-01 -1.73652738e-01] | [6.15775728225708, 6.663617134094238] |
27a51490-35f4-4295-a804-b3dd6086ab43 | using-self-supervised-co-training-to-improve | 2105.06421 | null | https://arxiv.org/abs/2105.06421v3 | https://arxiv.org/pdf/2105.06421v3.pdf | Using Self-Supervised Auxiliary Tasks to Improve Fine-Grained Facial Representation | In this paper, at first, the impact of ImageNet pre-training on fine-grained Facial Emotion Recognition (FER) is investigated which shows that when enough augmentations on images are applied, training from scratch provides better result than fine-tuning on ImageNet pre-training. Next, we propose a method to improve fine-grained and in-the-wild FER, called Hybrid Multi-Task Learning (HMTL). HMTL uses Self-Supervised Learning (SSL) as an auxiliary task during classical Supervised Learning (SL) in the form of Multi-Task Learning (MTL). Leveraging SSL during training can gain additional information from images for the primary fine-grained SL task. We investigate how proposed HMTL can be used in the FER domain by designing two customized version of common pre-text task techniques, puzzling and in-painting. We achieve state-of-the-art results on the AffectNet benchmark via two types of HMTL, without utilizing pre-training on additional data. Experimental results on the common SSL pre-training and proposed HMTL demonstrate the difference and superiority of our work. However, HMTL is not only limited to FER domain. Experiments on two types of fine-grained facial tasks, i.e., head pose estimation and gender recognition, reveals the potential of using HMTL to improve fine-grained facial representation. | ['Farzaneh Esmaili', 'Gholam Ali Montazer', 'Mahdi Pourmirzaei'] | 2021-05-13 | null | null | null | null | ['facial-emotion-recognition', 'head-pose-estimation'] | ['computer-vision', 'computer-vision'] | [ 3.04195791e-01 1.84831619e-01 3.30081843e-02 -7.87934780e-01
-6.96741462e-01 -1.34070829e-01 4.51674879e-01 -4.14478093e-01
-4.95080203e-01 5.73876441e-01 1.36014931e-02 2.59454072e-01
-2.75856033e-02 -4.36019540e-01 -7.70420909e-01 -7.51019061e-01
1.50392547e-01 2.76540756e-01 -2.58774310e-01 -4.61550742e-01
-1.62090912e-01 6.69899285e-01 -2.01336360e+00 5.17975152e-01
5.52065790e-01 1.49429500e+00 -2.94964798e-02 3.42228562e-01
-9.09144133e-02 5.42635202e-01 -6.56654537e-01 -6.89163327e-01
1.96713746e-01 -2.06991449e-01 -7.36456931e-01 1.72829762e-01
8.04280341e-01 -2.20924407e-01 3.12260956e-01 8.09858084e-01
8.90186012e-01 -3.93323041e-02 5.71653783e-01 -1.63544774e+00
-2.56860137e-01 2.14588955e-01 -9.07710791e-01 -2.44398370e-01
1.90003589e-01 -2.92508751e-02 5.63327193e-01 -1.20447481e+00
3.47884834e-01 1.70286191e+00 8.56900215e-01 1.03119946e+00
-9.17616189e-01 -1.21170366e+00 3.00479740e-01 7.76381642e-02
-1.56938601e+00 -7.57326782e-01 8.55109215e-01 -1.97636992e-01
8.50493968e-01 5.46618551e-02 2.75360823e-01 1.31887054e+00
3.43656018e-02 7.77674079e-01 1.63182485e+00 -4.80174452e-01
-7.78596895e-03 1.31054565e-01 -1.13648556e-01 1.10457826e+00
-7.88950175e-02 1.42527714e-01 -8.96460354e-01 -2.04523534e-01
6.36688828e-01 -2.26604059e-01 -5.86594455e-02 -5.22152297e-02
-4.21626598e-01 8.08815598e-01 3.69390137e-02 2.26692259e-01
-3.06905836e-01 1.44601539e-01 6.78791940e-01 3.64043564e-01
9.60042655e-01 3.82153764e-02 -7.92022943e-01 6.57268008e-03
-1.20261538e+00 -6.56537265e-02 5.67290187e-01 5.90422928e-01
1.09687841e+00 2.58078456e-01 -3.48530114e-01 1.10350013e+00
2.61875361e-01 5.19642115e-01 7.20280111e-01 -6.07077360e-01
4.38062027e-02 4.39441353e-01 -4.42925930e-01 -6.28471017e-01
-4.90724146e-01 -2.78718203e-01 -8.36936951e-01 3.42799634e-01
2.25286797e-01 -4.59642589e-01 -1.23031056e+00 2.07729149e+00
3.94052327e-01 3.23689669e-01 -1.64036199e-01 6.75066352e-01
1.14697587e+00 2.71640480e-01 3.05218786e-01 -1.17503576e-01
1.36866701e+00 -1.06892157e+00 -6.88569367e-01 -1.76502034e-01
6.27184212e-01 -6.69017553e-01 1.12515962e+00 5.84816635e-01
-9.13237453e-01 -8.39354753e-01 -7.75132358e-01 1.01950616e-01
-5.38939893e-01 5.49817681e-01 8.57991397e-01 1.12880492e+00
-1.27103388e+00 4.19441134e-01 -5.20360172e-01 -3.46645057e-01
8.59133661e-01 6.09862506e-01 -7.13774681e-01 5.03199473e-02
-1.07012236e+00 7.06753373e-01 9.73304436e-02 2.38602161e-01
-9.96782720e-01 -8.95020664e-01 -1.01253486e+00 -6.28387090e-03
4.51597840e-01 -6.36112571e-01 1.16600740e+00 -1.47997749e+00
-1.78378856e+00 1.12710619e+00 -3.20509642e-01 -7.96392113e-02
3.20710301e-01 -3.57653439e-01 -3.82580787e-01 1.64583698e-01
-1.23030119e-01 9.60347533e-01 1.59931135e+00 -1.27269852e+00
-3.85278702e-01 -6.57982647e-01 -9.65554081e-03 -4.16522585e-02
-5.32084882e-01 1.22474298e-01 -4.44197655e-01 -6.27965331e-01
-5.27913511e-01 -9.07307327e-01 7.42477849e-02 7.02738613e-02
-1.81485727e-01 -4.65071082e-01 1.16375387e+00 -4.03348207e-01
1.04089177e+00 -2.21997881e+00 -1.15889944e-01 1.87126726e-01
1.17656186e-01 4.46926832e-01 -5.68243802e-01 -8.28861520e-02
-5.35170555e-01 2.00415313e-01 1.17043212e-01 -9.43006635e-01
1.68341249e-01 1.79958686e-01 9.17757973e-02 3.80701840e-01
5.12247264e-01 1.01202202e+00 -4.06669080e-01 -6.55091584e-01
1.42443702e-01 6.69327974e-01 -6.28031790e-01 1.92070574e-01
-5.17754667e-02 5.92627466e-01 -2.37175092e-01 9.34966624e-01
9.27225173e-01 -7.20649958e-02 -1.39634892e-01 -7.40664244e-01
3.42621118e-01 -4.82518554e-01 -9.79124248e-01 1.77558553e+00
-8.79523993e-01 2.50932127e-01 1.84506834e-01 -9.72843289e-01
9.42792892e-01 3.21095139e-01 4.20746624e-01 -8.41012716e-01
3.31206173e-01 -2.22655442e-02 -2.08881751e-01 -4.91476625e-01
3.73452097e-01 -4.75877881e-01 -1.29377961e-01 3.84758085e-01
7.52425849e-01 1.66936681e-01 -1.49537297e-02 -1.68823689e-01
6.48214996e-01 3.35298717e-01 2.20022812e-01 -2.02536598e-01
5.28371811e-01 -6.44197404e-01 5.60099542e-01 4.00218725e-01
-2.52407849e-01 3.74213189e-01 3.13802809e-01 -1.51676118e-01
-5.66058576e-01 -5.40400445e-01 -1.87195390e-01 1.84074903e+00
-2.91555375e-01 -4.56550986e-01 -1.02168322e+00 -1.09711134e+00
8.40784088e-02 1.36066243e-01 -1.19728732e+00 -2.50735670e-01
-2.53028661e-01 -1.02656198e+00 9.78474021e-01 5.53074002e-01
8.79790187e-01 -1.09601605e+00 -6.16784394e-01 -2.66210496e-01
-6.26141503e-02 -1.36149025e+00 -3.38152200e-01 3.45377684e-01
-5.41051686e-01 -7.77569115e-01 -9.15494561e-01 -4.89603549e-01
6.62210524e-01 -5.41144237e-02 1.14494693e+00 1.36725932e-01
-3.77343327e-01 5.62492967e-01 -3.57787549e-01 -6.19851947e-01
6.40159920e-02 -2.95787882e-02 -1.71204377e-02 6.21003985e-01
2.69389778e-01 -6.08730853e-01 -5.02148867e-01 3.59914899e-01
-8.09618711e-01 -1.15465924e-01 9.63644385e-01 1.06318438e+00
5.28645217e-01 -2.29382906e-02 7.00300336e-01 -1.11179042e+00
4.39688146e-01 -1.85490116e-01 -6.21091537e-02 3.50466907e-01
-3.45633775e-01 1.22411206e-01 3.84639680e-01 -5.39722383e-01
-1.42429733e+00 5.74199185e-02 -4.05256420e-01 -9.09112871e-01
-4.82433707e-01 2.75464386e-01 -4.06136423e-01 -5.25013506e-01
3.83445799e-01 -1.61111712e-01 7.90461451e-02 -5.68979740e-01
2.55778134e-01 5.21922410e-01 2.28109300e-01 -7.68527150e-01
3.94250691e-01 5.56571722e-01 2.26870790e-01 -7.79788792e-01
-1.03501713e+00 -3.33658487e-01 -7.09914327e-01 -3.25350672e-01
8.05461943e-01 -1.00651801e+00 -9.03739870e-01 7.72874296e-01
-8.02633703e-01 -5.70247412e-01 -2.00655267e-01 -2.78691370e-02
-4.91002977e-01 4.75532487e-02 -2.94350892e-01 -8.76968682e-01
-4.89528149e-01 -1.04172659e+00 1.87769258e+00 2.21669257e-01
2.90056802e-02 -1.06386614e+00 -1.43211439e-01 3.78917217e-01
3.90939236e-01 2.62146115e-01 6.32037580e-01 -5.66380858e-01
1.40981421e-01 2.41238847e-01 -2.47946411e-01 5.27880132e-01
8.01218376e-02 -1.87003627e-01 -1.62241852e+00 -4.67692226e-01
-7.16744140e-02 -1.05229175e+00 1.06298935e+00 3.42987210e-01
1.54363203e+00 -3.71819511e-02 -1.93642914e-01 8.95057321e-01
1.37825620e+00 -1.93530217e-01 6.55843496e-01 1.15697980e-01
7.71677494e-01 7.11328685e-01 8.27825963e-01 7.11558402e-01
4.36454654e-01 8.21024299e-01 3.11764687e-01 -6.41588986e-01
-4.06684726e-01 -1.00977384e-01 4.12725270e-01 1.96570650e-01
-3.01543385e-01 1.80768609e-01 -5.21064699e-01 1.22385882e-01
-1.56849360e+00 -7.40703225e-01 4.53550756e-01 1.80933559e+00
8.73728096e-01 -2.70494163e-01 2.41833165e-01 2.03065112e-01
5.10641515e-01 9.26222801e-02 -6.26781404e-01 -6.58959150e-01
-2.18770325e-01 1.01904345e+00 3.79629970e-01 1.70855001e-01
-1.28229344e+00 1.29283094e+00 5.23290825e+00 1.42306983e+00
-1.65033984e+00 5.46688378e-01 9.46534932e-01 -3.69244814e-02
1.99501514e-01 -5.80313087e-01 -1.15708101e+00 1.16974689e-01
8.87256742e-01 3.12868625e-01 2.59361863e-01 9.02360201e-01
9.59883034e-02 -1.89305693e-01 -9.68772531e-01 1.42986298e+00
4.78727549e-01 -8.58043671e-01 2.01272160e-01 -2.14241996e-01
5.41176200e-01 -3.49122465e-01 3.29130620e-01 6.43900454e-01
-3.25482376e-02 -1.22697198e+00 6.81656480e-01 4.14091647e-01
1.35859776e+00 -8.57416511e-01 7.68096864e-01 -3.48999910e-02
-1.31725359e+00 -1.08549614e-02 -2.09918678e-01 2.27054149e-01
-1.07658371e-01 5.19785285e-01 -6.02763355e-01 6.53352976e-01
1.09253502e+00 5.82847118e-01 -9.52448249e-01 5.03043890e-01
-2.04240456e-01 4.54460412e-01 -2.80702084e-01 2.71243602e-01
6.63576573e-02 2.88968533e-01 1.30930245e-02 1.38539088e+00
1.13335192e-01 -8.27273801e-02 8.18803012e-02 4.35984135e-01
-1.71069056e-01 1.84749812e-01 -3.21456701e-01 3.34692001e-02
1.45782689e-02 1.79027641e+00 -4.42587823e-01 -2.68059880e-01
-3.13277036e-01 1.20905912e+00 3.80393714e-01 3.30561966e-01
-9.27000225e-01 -1.60629198e-01 8.15741658e-01 -8.58914033e-02
5.17250180e-01 1.89611927e-01 -1.66980416e-01 -8.25138330e-01
-2.43857995e-01 -1.02515888e+00 4.14014488e-01 -7.12587476e-01
-1.24710143e+00 9.44073439e-01 1.05934367e-02 -7.70516217e-01
-3.03504348e-01 -7.59664834e-01 -4.38570291e-01 7.80147851e-01
-1.93873310e+00 -1.95024848e+00 -5.33524573e-01 1.22989917e+00
4.90582764e-01 -2.12832183e-01 9.71122265e-01 5.47738254e-01
-6.55514181e-01 1.40456307e+00 -5.96382916e-01 -2.30456796e-02
1.11729956e+00 -1.03465199e+00 -1.75649196e-01 3.63824815e-01
3.20144789e-03 2.40876421e-01 2.99767673e-01 -4.53475565e-01
-1.26409495e+00 -1.44114566e+00 4.58186209e-01 -4.73415619e-03
2.54653782e-01 -6.73530042e-01 -6.29366040e-01 5.54504156e-01
9.17693302e-02 2.16355577e-01 8.22131097e-01 4.09609824e-01
-4.80839700e-01 -3.27883154e-01 -1.49546099e+00 2.42759764e-01
1.02338123e+00 -6.42103970e-01 -1.57956824e-01 3.94658260e-02
3.81158024e-01 -3.11394602e-01 -8.99933636e-01 8.22084129e-01
7.85506248e-01 -9.62629735e-01 9.13618803e-01 -4.84397471e-01
3.87040615e-01 1.75746053e-01 -1.29502773e-01 -1.49820781e+00
-6.11701123e-02 -6.99721336e-01 1.01168260e-01 1.35203099e+00
1.84937373e-01 -5.53093553e-01 8.48541021e-01 2.53013283e-01
-1.79630920e-01 -1.05695486e+00 -8.22584331e-01 -6.88210845e-01
8.60679522e-02 -4.91729945e-01 7.13288009e-01 8.63393486e-01
-5.23040891e-01 3.88727367e-01 -5.26460528e-01 2.03692038e-02
5.68611920e-01 -9.08324495e-02 9.04469550e-01 -1.30806792e+00
-1.75330237e-01 -1.87563881e-01 -2.77903765e-01 -6.54672682e-01
6.75499737e-01 -6.76650763e-01 -1.05436869e-01 -8.82461846e-01
3.48360926e-01 -3.70428234e-01 -2.21496567e-01 1.13343656e+00
-1.64484084e-01 8.59077513e-01 3.62552553e-01 -2.08399549e-01
-6.58485889e-01 7.70122826e-01 1.38745177e+00 6.92965910e-02
1.01305425e-01 -3.36778551e-01 -7.80155957e-01 1.00146210e+00
6.04562640e-01 -3.17737877e-01 -4.21949148e-01 -8.82659629e-02
-1.33976877e-01 -2.56248087e-01 4.72161084e-01 -8.26475143e-01
-3.89303342e-02 1.76340148e-01 5.74787676e-01 -3.53124440e-01
6.95068836e-01 -6.54804289e-01 -1.48508430e-01 -7.45770186e-02
3.71902213e-02 -1.36614144e-01 7.91840017e-01 1.54262573e-01
-3.73979300e-01 2.71846056e-02 9.71483648e-01 -4.50624786e-02
-9.87387896e-01 5.29642820e-01 -6.76752031e-02 2.62944810e-02
8.67700696e-01 -3.27807546e-01 -1.32313833e-01 -3.11247289e-01
-1.07108104e+00 -1.07613660e-01 4.67086583e-03 3.72130752e-01
5.69055498e-01 -1.33137798e+00 -5.44323504e-01 4.87924486e-01
1.74819946e-01 -4.44539458e-01 5.98445177e-01 9.29332376e-01
3.79375964e-01 3.79348546e-01 -5.04970372e-01 -6.59210503e-01
-1.79564238e+00 1.72396347e-01 4.63496506e-01 -4.72510159e-01
1.83593705e-02 1.14763641e+00 6.44630432e-01 -5.68738937e-01
2.70381033e-01 -4.92098294e-02 -3.41720730e-01 3.37320149e-01
3.95552427e-01 8.44731554e-03 4.39024478e-01 -8.67914617e-01
-3.30557048e-01 8.67627740e-01 -1.88927814e-01 -8.66224915e-02
1.38383615e+00 -1.08999051e-01 2.37376173e-03 3.11123282e-01
1.45670557e+00 -1.42888024e-01 -1.17094302e+00 -2.79928483e-02
-2.15088889e-01 -4.47397143e-01 3.50805640e-01 -1.14904535e+00
-1.57115221e+00 1.01749313e+00 9.88415897e-01 -5.77454984e-01
1.80365562e+00 6.67877793e-02 6.19994223e-01 1.53748631e-01
5.62196374e-01 -1.06848121e+00 5.63384771e-01 4.59961265e-01
1.02841997e+00 -1.35362530e+00 -2.75983274e-01 -5.77283323e-01
-7.40871966e-01 1.00008500e+00 9.04977739e-01 1.18267581e-01
1.02268374e+00 4.77604032e-01 6.34218231e-02 -4.08341467e-01
-6.90905988e-01 -5.55473089e-01 5.00700116e-01 4.96394664e-01
5.45049489e-01 -2.20894173e-01 3.23840715e-02 9.83087301e-01
-2.62779683e-01 3.14122945e-01 -1.43591851e-01 8.95114899e-01
-9.89907328e-03 -1.26934159e+00 -3.70318234e-01 4.13243830e-01
-6.61432862e-01 -1.37113720e-01 -5.12255251e-01 8.81667674e-01
6.98894322e-01 8.02038133e-01 3.24554369e-03 -5.13658345e-01
3.16851467e-01 2.41517842e-01 1.11978555e+00 -5.51995099e-01
-7.86678493e-01 1.90604225e-01 2.45459288e-01 -9.48880732e-01
-7.77972460e-01 -4.46610302e-01 -8.89837742e-01 -1.26998469e-01
-3.85940969e-01 -1.09078415e-01 6.79109573e-01 1.13192487e+00
4.34471190e-01 6.00839198e-01 6.39897704e-01 -1.20009232e+00
-2.43911371e-01 -1.09484792e+00 -7.01828599e-01 5.53490937e-01
2.94781178e-01 -1.19015253e+00 -2.64316082e-01 1.48029590e-03] | [13.508712768554688, 1.3637762069702148] |
ef335c69-d6e0-4f9b-815b-93f2190a56fe | large-scale-privacy-preserving-network | 2205.14440 | null | https://arxiv.org/abs/2205.14440v1 | https://arxiv.org/pdf/2205.14440v1.pdf | Large-Scale Privacy-Preserving Network Embedding against Private Link Inference Attacks | Network embedding represents network nodes by a low-dimensional informative vector. While it is generally effective for various downstream tasks, it may leak some private information of networks, such as hidden private links. In this work, we address a novel problem of privacy-preserving network embedding against private link inference attacks. Basically, we propose to perturb the original network by adding or removing links, and expect the embedding generated on the perturbed network can leak little information about private links but hold high utility for various downstream tasks. Towards this goal, we first propose general measurements to quantify privacy gain and utility loss incurred by candidate network perturbations; we then design a PPNE framework to identify the optimal perturbation solution with the best privacy-utility trade-off in an iterative way. Furthermore, we propose many techniques to accelerate PPNE and ensure its scalability. For instance, as the skip-gram embedding methods including DeepWalk and LINE can be seen as matrix factorization with closed form embedding results, we devise efficient privacy gain and utility loss approximation methods to avoid the repetitive time-consuming embedding training for every candidate network perturbation in each iteration. Experiments on real-life network datasets (with up to millions of nodes) verify that PPNE outperforms baselines by sacrificing less utility and obtaining higher privacy protection. | ['Yuncong Yang', 'Junjie Wu', 'Leye Wang', 'Xiao Han'] | 2022-05-28 | null | null | null | null | ['network-embedding'] | ['methodology'] | [-1.05767976e-03 5.29354692e-01 -5.17966092e-01 -1.56160459e-01
-5.78477800e-01 -1.08893561e+00 3.08536917e-01 1.36674404e-01
-3.08575749e-01 7.00154960e-01 4.17065918e-01 -5.46072900e-01
-2.87530512e-01 -1.05738926e+00 -7.14941800e-01 -8.96624446e-01
-2.59431303e-01 1.72518000e-01 5.43231796e-03 8.15748945e-02
-1.88324183e-01 6.62955463e-01 -5.62046409e-01 -1.43638283e-01
5.00415564e-01 7.09770918e-01 -5.85094512e-01 3.32941264e-01
3.37319165e-01 3.19277525e-01 -3.08450729e-01 -1.13280559e+00
8.67913246e-01 -1.85566202e-01 -7.24585772e-01 -2.80332714e-01
-3.18890736e-02 -7.15407193e-01 -1.14406300e+00 1.31775582e+00
5.91489494e-01 7.45244548e-02 3.81975055e-01 -1.69631493e+00
-6.36575341e-01 1.01153171e+00 -5.10755718e-01 3.92355677e-03
1.28921866e-01 2.60810941e-01 1.25266397e+00 -3.36233258e-01
5.81258714e-01 1.15503109e+00 7.82636404e-01 6.38904989e-01
-1.70688450e+00 -9.75682676e-01 1.28121346e-01 -1.50242448e-01
-1.43052161e+00 -5.47416031e-01 8.34797084e-01 -4.87868190e-02
3.85901779e-01 8.13316703e-01 3.56791496e-01 1.31903005e+00
-7.57913515e-02 7.08526790e-01 4.48876500e-01 2.11437330e-01
1.33227393e-01 5.60100734e-01 -1.67638287e-01 4.84735578e-01
6.56453192e-01 8.35289881e-02 -2.83195436e-01 -8.46805096e-01
4.89309400e-01 4.09032732e-01 -7.73640871e-01 -7.13740349e-01
-1.04732180e+00 8.89866233e-01 5.74050128e-01 -1.08678386e-01
-1.33632079e-01 5.22644222e-01 6.85142100e-01 6.95558429e-01
3.30791354e-01 3.01860422e-01 -4.72131997e-01 1.76505208e-01
-4.48680431e-01 2.53693581e-01 1.12386525e+00 8.95431697e-01
7.42661774e-01 -5.61125278e-01 -5.39520323e-01 4.14059579e-01
3.27517778e-01 1.44623160e-01 1.62482530e-01 -1.08068562e+00
8.55593443e-01 3.78514022e-01 3.97005714e-02 -1.60265362e+00
2.47564584e-01 -3.47273707e-01 -1.14359725e+00 -2.05872655e-01
2.64068753e-01 -5.30426443e-01 -1.18650883e-01 1.99980724e+00
4.67344165e-01 3.64268839e-01 1.50794372e-01 6.47713363e-01
2.71327585e-01 7.43282676e-01 -7.27945715e-02 -2.23700225e-01
1.15562487e+00 -9.26171660e-01 -7.65473843e-01 -1.05439022e-01
8.75391901e-01 -1.31235167e-01 7.80929387e-01 -1.14467837e-01
-8.74164283e-01 3.16442221e-01 -8.97038579e-01 -1.45064503e-01
-2.72484958e-01 -3.28171730e-01 7.07598209e-01 9.82526600e-01
-1.02593565e+00 8.68261278e-01 -7.48346806e-01 -1.96410030e-01
7.44311988e-01 5.41619599e-01 -8.86385560e-01 -1.67808190e-01
-1.42782760e+00 1.45675465e-01 5.14423013e-01 5.31470440e-02
-8.35366905e-01 -1.08007824e+00 -7.76927888e-01 5.11867642e-01
6.19544387e-01 -6.80300832e-01 7.69073129e-01 -2.06117481e-01
-1.28050816e+00 4.48832810e-01 4.80496809e-02 -7.63201237e-01
8.95333588e-01 5.06711639e-02 -3.40398818e-01 2.71659940e-01
-1.51628375e-01 1.55821919e-01 6.68270886e-01 -1.30314541e+00
-2.67248780e-01 -4.44571227e-01 4.06071633e-01 -3.42905754e-03
-1.30593336e+00 -2.01755747e-01 -5.66924572e-01 -6.93127394e-01
1.43207060e-02 -9.36755002e-01 -5.62975228e-01 7.52997637e-01
-7.62063384e-01 1.60328165e-01 1.15959132e+00 -6.03487611e-01
1.38924181e+00 -2.18491459e+00 4.25319634e-02 7.35330760e-01
8.33522379e-01 3.13820720e-01 -2.18918428e-01 6.34578466e-01
7.72298872e-02 5.41957915e-01 -2.35676393e-01 -5.69259644e-01
3.25984448e-01 3.89298141e-01 -4.57506835e-01 7.14736700e-01
-2.01699793e-01 1.06123781e+00 -9.89044249e-01 -2.25105554e-01
-1.34639934e-01 6.38978839e-01 -7.85043240e-01 1.22619107e-01
1.47252500e-01 1.12493388e-01 -6.26899123e-01 3.91831279e-01
8.46513689e-01 -2.07847148e-01 4.42146748e-01 -3.08154345e-01
6.76188290e-01 2.39710789e-02 -1.07566726e+00 1.26334798e+00
-2.61595637e-01 5.10365665e-01 4.13307309e-01 -8.98974836e-01
6.47876203e-01 4.65937406e-01 4.48630393e-01 7.06194714e-02
1.58721700e-01 -1.51677385e-01 -2.98029542e-01 -3.07863414e-01
3.90651673e-02 2.88461167e-02 -6.71378523e-02 8.15930903e-01
-3.07855159e-01 5.99566877e-01 -2.86795199e-01 5.20744801e-01
1.49517035e+00 -7.95871198e-01 7.01543614e-02 5.01016453e-02
4.79451120e-01 -7.07720637e-01 7.56478727e-01 6.16810739e-01
-4.17482823e-01 2.18111858e-01 1.15888357e+00 -1.41270176e-01
-8.01032186e-01 -1.03552806e+00 2.06478342e-01 7.70236433e-01
1.13143787e-01 -7.37203062e-01 -5.78361988e-01 -1.11795282e+00
5.10601938e-01 3.98444027e-01 -6.14701092e-01 -6.42511487e-01
-2.14966938e-01 -7.72915721e-01 9.10355031e-01 2.51527548e-01
4.47960824e-01 -4.33192313e-01 3.92893583e-01 -7.99828991e-02
-8.21257830e-02 -9.63731587e-01 -9.10893202e-01 -4.91292626e-02
-7.83833742e-01 -9.26442444e-01 -4.00860548e-01 -5.30168116e-01
1.09585071e+00 3.57406884e-01 5.38124144e-01 6.06922954e-02
-9.14524682e-03 1.79778263e-01 -7.68689737e-02 1.54027984e-01
-1.71404049e-01 2.94079363e-01 1.20770812e-01 2.82219172e-01
3.64319354e-01 -1.06004870e+00 -8.50447655e-01 9.15092006e-02
-1.02298903e+00 -5.70251882e-01 5.81746459e-01 8.08280170e-01
4.24270868e-01 5.01244485e-01 2.10816532e-01 -1.27030098e+00
1.14242971e+00 -8.16119492e-01 -3.62006754e-01 2.01438993e-01
-7.89690971e-01 1.27141133e-01 8.28428030e-01 -4.72150862e-01
-5.39489388e-01 -2.75621377e-02 1.10926405e-02 -6.62272632e-01
4.24599588e-01 2.19229832e-01 -7.50369012e-01 -4.99391168e-01
4.52119529e-01 1.41621187e-01 2.28328317e-01 -4.57632273e-01
6.17597520e-01 5.61260283e-01 3.06731820e-01 -6.36586487e-01
1.59472680e+00 6.21770561e-01 2.76958287e-01 -2.86257088e-01
-5.25959671e-01 -1.91839665e-01 -2.14767992e-01 3.11062694e-01
2.82601804e-01 -7.29272723e-01 -1.19818079e+00 2.24022884e-02
-8.75994146e-01 4.60158698e-02 -3.88185471e-01 1.56320274e-01
-1.32470310e-01 8.60575855e-01 -6.99326217e-01 -6.60334706e-01
-5.79741597e-01 -9.05025482e-01 5.22978067e-01 -1.74590617e-01
-5.09419590e-02 -1.18796134e+00 -4.19014366e-03 2.07797989e-01
2.86148727e-01 3.92617613e-01 7.12423623e-01 -7.38009214e-01
-7.00256348e-01 -7.89753377e-01 -3.66017610e-01 4.83094275e-01
9.34122130e-02 -2.57010937e-01 -8.00601184e-01 -6.89483523e-01
-1.48813173e-01 -4.61829528e-02 7.15820312e-01 -1.97753727e-01
1.77714729e+00 -1.21126235e+00 -6.36365592e-01 1.18408453e+00
1.43657243e+00 -4.48381543e-01 6.13997281e-01 9.45144668e-02
8.72648001e-01 6.08194649e-01 2.01781809e-01 5.90267539e-01
2.18012497e-01 2.87727386e-01 6.30206227e-01 1.19324654e-01
4.91250664e-01 -8.98286343e-01 4.02262181e-01 3.33166122e-01
4.29197013e-01 -4.83881265e-01 -3.91256303e-01 5.36925614e-01
-1.94577181e+00 -8.43516529e-01 2.06665710e-01 2.24442840e+00
8.02300692e-01 1.12047404e-01 1.17919013e-01 2.51495123e-01
5.23867548e-01 4.56044286e-01 -5.79737008e-01 -1.63819268e-01
1.81477834e-02 -2.99501359e-01 1.10182738e+00 5.63707173e-01
-8.93418789e-01 7.21341312e-01 5.99490499e+00 8.23326230e-01
-6.93752825e-01 1.20833956e-01 7.27003276e-01 -3.18573892e-01
-9.04169798e-01 1.75669566e-01 -6.18962646e-01 6.04546189e-01
8.12858820e-01 -7.44437635e-01 7.34199047e-01 8.99673522e-01
7.63411596e-02 8.57582867e-01 -1.21422005e+00 8.18106711e-01
-3.72678190e-01 -1.34704745e+00 1.56814933e-01 4.61592734e-01
5.49301207e-01 -2.74862498e-01 2.25232020e-01 -2.04681549e-02
6.20704830e-01 -8.67667854e-01 -4.52473201e-02 1.60922676e-01
7.30451167e-01 -1.08439493e+00 5.67606032e-01 2.79648572e-01
-9.67478156e-01 -2.50630140e-01 -7.37512708e-01 2.82270610e-01
7.93590322e-02 7.51318395e-01 -6.18349731e-01 4.73916084e-01
5.18873155e-01 7.15880215e-01 -2.25383580e-01 7.34030247e-01
-3.57143193e-01 5.24535120e-01 -5.32610953e-01 7.15284422e-03
2.52582520e-01 -5.02638757e-01 8.89001727e-01 9.18487191e-01
3.14439535e-01 1.81455940e-01 -2.96903253e-02 8.33661497e-01
-8.04330468e-01 5.39948642e-02 -9.50537801e-01 -2.94587493e-01
1.03683949e+00 1.29636848e+00 -1.20536707e-01 8.40060562e-02
-1.14895284e-01 1.28498924e+00 4.49207067e-01 6.28985226e-01
-6.92086995e-01 -6.38478935e-01 1.48581576e+00 1.44787207e-01
2.08289519e-01 1.60082802e-01 -4.62660529e-02 -1.20944619e+00
3.31753135e-01 -5.05012155e-01 5.20345867e-01 -3.77381407e-02
-1.48464787e+00 2.17559308e-01 -3.90703142e-01 -1.06444740e+00
1.03147015e-01 -2.42063105e-01 -8.32584620e-01 8.21573734e-01
-1.46340454e+00 -1.00343478e+00 1.27103657e-01 6.56235695e-01
-4.40442234e-01 -9.70509499e-02 8.25799406e-01 4.36268449e-01
-9.73013937e-01 1.51382434e+00 6.31804764e-01 4.76786822e-01
5.33276498e-01 -1.01434326e+00 4.21023637e-01 9.91455078e-01
-1.29097909e-01 8.87745380e-01 6.78974986e-01 -4.55600083e-01
-1.69418740e+00 -1.47927201e+00 8.67364705e-01 -1.93072215e-01
8.59102368e-01 -6.10757470e-01 -8.36452246e-01 1.11366618e+00
-2.75811434e-01 5.64549446e-01 9.19947267e-01 -8.29657987e-02
-6.46661282e-01 -5.62086523e-01 -1.86274266e+00 1.01915741e+00
1.27202320e+00 -7.32532322e-01 5.90521172e-02 6.15873516e-01
1.26108813e+00 -3.97343338e-02 -1.13331866e+00 3.27593647e-02
4.46108192e-01 -7.09686458e-01 1.25247049e+00 -8.71803105e-01
1.55816257e-01 -2.17356145e-01 -7.14101121e-02 -1.21043944e+00
-4.77218807e-01 -1.31005132e+00 -7.70910382e-01 1.56628323e+00
4.74657238e-01 -1.05303621e+00 1.42295063e+00 1.01988554e+00
7.04183757e-01 -7.64255226e-01 -9.03451681e-01 -7.58875012e-01
-1.56143624e-02 -9.84715372e-02 1.06119299e+00 1.12417674e+00
3.62912603e-02 -1.35157868e-01 -6.53289080e-01 5.81129014e-01
9.76528764e-01 -3.49160999e-01 9.16143239e-01 -1.08864081e+00
-3.67598265e-01 -2.18338162e-01 -6.04119182e-01 -9.59327340e-01
6.32357538e-01 -1.08140624e+00 -5.71208656e-01 -1.13788390e+00
1.81375891e-01 -5.02269208e-01 -3.71472538e-01 5.96339703e-01
-1.73234403e-01 1.17017411e-01 -3.96680646e-02 1.24840721e-01
-3.44640046e-01 7.22846508e-01 9.40728724e-01 -1.40418813e-01
-8.55247825e-02 2.03794718e-01 -1.35361063e+00 2.86749035e-01
7.81492591e-01 -6.52310729e-01 -8.50508392e-01 -2.61359245e-01
3.67600113e-01 -5.40959537e-02 3.44592154e-01 -6.06807590e-01
2.69869119e-01 8.45429078e-02 1.51373781e-02 -2.21307486e-01
2.78728902e-01 -1.45005631e+00 3.53210986e-01 6.19140744e-01
-4.21928346e-01 -3.29263061e-01 -2.29699075e-01 1.21453166e+00
1.25672027e-01 -8.31211880e-02 5.37357211e-01 1.63722143e-01
-1.05887197e-01 1.03031862e+00 7.79821351e-02 -5.01947030e-02
1.42807853e+00 -7.81451687e-02 -2.91756243e-01 -6.03788733e-01
-7.37751663e-01 7.17439890e-01 6.66309595e-01 1.16775528e-01
5.09269476e-01 -1.56208968e+00 -6.04258001e-01 2.30387852e-01
8.10443535e-02 -1.40937686e-01 1.08086735e-01 5.25101006e-01
-3.06166649e-01 1.51293814e-01 9.99467224e-02 8.00822452e-02
-1.36019433e+00 8.01559746e-01 2.15937525e-01 -3.97985876e-01
-7.27325320e-01 1.02269995e+00 -1.47025147e-02 -6.28122866e-01
5.70871353e-01 1.25641227e-01 1.92474470e-01 -5.49681932e-02
6.36532307e-01 5.74403703e-01 -2.84805238e-01 -1.61666960e-01
-9.89842340e-02 5.77624515e-02 -2.16772377e-01 8.64353850e-02
1.35960793e+00 -4.44621295e-01 -4.49355960e-01 -1.96609527e-01
1.90191960e+00 1.16241254e-01 -1.29688478e+00 -4.61811274e-01
-2.65370786e-01 -9.88209069e-01 1.18666857e-01 -1.63554564e-01
-1.58043969e+00 5.93751311e-01 2.41264701e-01 3.96907002e-01
1.04389596e+00 -1.51269317e-01 1.17580032e+00 6.30503714e-01
1.22609444e-01 -6.26430154e-01 -4.64182645e-01 -8.44416767e-02
4.83770818e-01 -1.04242134e+00 3.39392126e-02 -6.79027736e-01
-3.51458758e-01 7.89397359e-01 5.21003962e-01 -3.79500492e-03
9.94025111e-01 9.18265730e-02 -3.93634021e-01 2.60282448e-03
-7.89304376e-01 5.39790094e-01 -1.28234670e-01 6.84785306e-01
-8.67150128e-02 1.47597283e-01 -2.47251660e-01 6.85068905e-01
-1.30544245e-01 -3.63397986e-01 4.40157712e-01 6.29120052e-01
-3.71866603e-03 -1.49289429e+00 -7.41621330e-02 7.49538243e-01
-6.33613229e-01 8.79762769e-02 -5.52367687e-01 4.40531790e-01
-3.84020746e-01 4.81598586e-01 -2.85188168e-01 -6.06707096e-01
1.83957875e-01 -2.59172976e-01 -1.54805645e-01 -3.59486014e-01
-5.51055431e-01 -4.44090664e-01 1.05796918e-01 -7.87586987e-01
2.34326825e-01 -5.01948476e-01 -7.90052474e-01 -1.14203954e+00
-2.04952672e-01 3.38123202e-01 1.61732987e-01 4.04866040e-01
6.28024936e-01 6.92832023e-02 1.19244099e+00 -2.94519842e-01
-9.65960562e-01 -4.70262945e-01 -8.90210450e-01 4.60530400e-01
4.85477775e-01 -1.56491712e-01 -9.26589668e-01 -5.41107953e-01] | [6.018126010894775, 6.891495704650879] |
061c386c-bc52-4fb2-bae4-51829215e87d | i2-sdf-intrinsic-indoor-scene-reconstruction | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Zhu_I2-SDF_Intrinsic_Indoor_Scene_Reconstruction_and_Editing_via_Raytracing_in_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Zhu_I2-SDF_Intrinsic_Indoor_Scene_Reconstruction_and_Editing_via_Raytracing_in_CVPR_2023_paper.pdf | I2-SDF: Intrinsic Indoor Scene Reconstruction and Editing via Raytracing in Neural SDFs | In this work, we present I^2-SDF, a new method for intrinsic indoor scene reconstruction and editing using differentiable Monte Carlo raytracing on neural signed distance fields (SDFs). Our holistic neural SDF-based framework jointly recovers the underlying shapes, incident radiance and materials from multi-view images. We introduce a novel bubble loss for fine-grained small objects and error-guided adaptive sampling scheme to largely improve the reconstruction quality on large-scale indoor scenes. Further, we propose to decompose the neural radiance field into spatially-varying material of the scene as a neural field through surface-based, differentiable Monte Carlo raytracing and emitter semantic segmentations, which enables physically based and photorealistic scene relighting and editing applications. Through a number of qualitative and quantitative experiments, we demonstrate the superior quality of our method on indoor scene reconstruction, novel view synthesis, and scene editing compared to state-of-the-art baselines. Our project page is at https://jingsenzhu.github.io/i2-sdf. | ['Rui Wang', 'Hujun Bao', 'Wei Hua', 'Rui Tang', 'Lisha Wang', 'Dianbing Xi', 'Jifan Li', 'Fujun Luan', 'Qi Ye', 'Yuchi Huo', 'Jingsen Zhu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['indoor-scene-reconstruction', 'novel-view-synthesis'] | ['computer-vision', 'computer-vision'] | [ 4.85916704e-01 -3.48615646e-01 5.89692295e-01 -6.00135624e-01
-6.88013852e-01 -6.47065401e-01 5.69165885e-01 -4.06222314e-01
3.53827961e-02 9.59862769e-01 3.87006402e-01 2.15661786e-02
-1.49442554e-02 -1.29286349e+00 -1.16065562e+00 -4.81840163e-01
4.17005658e-01 2.41192281e-01 -8.89257900e-03 -1.15695857e-01
7.24862069e-02 5.99523902e-01 -1.45531452e+00 2.66382486e-01
1.26912808e+00 8.07641089e-01 3.58133495e-01 9.06145275e-01
-8.56789500e-02 9.16690052e-01 -1.40714183e-01 -2.88593560e-01
3.88315439e-01 -4.68393594e-01 -6.44579053e-01 -9.52362642e-02
9.80367661e-01 -7.00533628e-01 -5.06789386e-01 8.95379543e-01
4.83310133e-01 5.01208007e-01 7.32140303e-01 -7.10855603e-01
-8.40389669e-01 -1.94256127e-01 -4.53799278e-01 -2.64132470e-01
7.23271430e-01 1.62486285e-01 5.52422523e-01 -1.17774999e+00
7.50700533e-01 1.24635088e+00 8.91916633e-01 4.30635005e-01
-1.35053253e+00 -5.19742072e-01 1.96884796e-01 -4.24466640e-01
-1.27717221e+00 -4.15780663e-01 1.11252117e+00 -5.95799923e-01
7.05293477e-01 5.21164715e-01 8.04421127e-01 1.07311070e+00
2.77834803e-01 4.78815734e-01 1.27216887e+00 -1.91291451e-01
3.61426651e-01 -1.82098612e-01 -4.14751232e-01 1.10563183e+00
-1.25813797e-01 4.86475646e-01 -6.09290004e-01 -1.19809799e-01
1.37904716e+00 3.46194625e-01 -5.17439604e-01 -4.78588432e-01
-1.44161499e+00 4.85312134e-01 7.53652215e-01 -2.45038465e-01
-2.55156159e-01 6.95019484e-01 -8.49876404e-02 -2.18108237e-01
1.03658402e+00 1.25806525e-01 -3.29341710e-01 2.11226419e-01
-8.73374224e-01 4.91307050e-01 4.81311023e-01 1.02701843e+00
9.12724078e-01 1.75485790e-01 -1.59077108e-01 8.12046587e-01
5.04971027e-01 1.05047071e+00 -5.54326594e-01 -1.68144977e+00
2.11037576e-01 1.46006525e-01 2.55030036e-01 -8.43855321e-01
8.56973529e-02 -2.77882844e-01 -9.82244134e-01 3.98573577e-01
1.28202945e-01 5.32663278e-02 -1.03014147e+00 1.41118371e+00
6.87648058e-01 4.81051564e-01 -2.52967000e-01 7.83523202e-01
1.10439873e+00 9.34518516e-01 -3.95681262e-01 1.52301669e-01
1.00526834e+00 -1.19845152e+00 -5.59209347e-01 2.48525403e-02
6.71254024e-02 -6.73955023e-01 1.28570855e+00 3.73887688e-01
-1.37400961e+00 -3.66059631e-01 -7.66769052e-01 -5.65453708e-01
4.54750024e-02 -1.10198326e-01 7.75516331e-01 4.26172793e-01
-9.19134855e-01 6.50220931e-01 -8.62974823e-01 6.97879568e-02
7.89234757e-01 -1.23217687e-01 1.06408829e-02 -5.02651989e-01
-7.17631757e-01 1.01549029e-01 -3.59527528e-01 -1.02013722e-02
-1.21233940e+00 -1.52228057e+00 -9.92574453e-01 -2.88024604e-01
1.76480740e-01 -1.57881784e+00 1.09719014e+00 -6.22145593e-01
-1.56763637e+00 7.86740899e-01 -3.80209595e-01 1.41608819e-01
8.35502267e-01 -2.05987811e-01 9.37859043e-02 2.07895879e-02
2.99068779e-01 5.62156439e-01 6.49417460e-01 -1.94458175e+00
-2.26292863e-01 -3.52133840e-01 1.70061782e-01 4.26794469e-01
4.30172384e-01 -4.48317438e-01 -3.17696869e-01 -9.45758641e-01
1.46113798e-01 -5.57056844e-01 -3.84827465e-01 6.65059507e-01
-5.18890560e-01 5.12109637e-01 3.98661971e-01 -8.10106933e-01
7.11699605e-01 -2.00362659e+00 2.46529832e-01 9.83533412e-02
2.40333840e-01 -5.76006293e-01 9.68009010e-02 2.49416992e-01
3.36504430e-01 -2.70992629e-02 -8.12735736e-01 -7.96750367e-01
-5.77035435e-02 1.19204007e-01 -4.51596558e-01 5.45215726e-01
-3.94940048e-01 9.56862092e-01 -1.12984502e+00 -1.15280166e-01
6.20413661e-01 1.19572854e+00 -9.22172248e-01 1.72302797e-01
-4.52978343e-01 1.20623720e+00 -4.57698256e-01 6.92332327e-01
1.11289477e+00 -1.60771921e-01 -3.69202167e-01 -3.26415211e-01
-3.40281397e-01 1.29028037e-01 -1.13561559e+00 2.56164336e+00
-9.71574605e-01 5.00898838e-01 3.85658681e-01 -1.03852004e-01
5.59061408e-01 -1.34504616e-01 5.15599966e-01 -7.90906668e-01
-1.08287089e-01 1.85956687e-01 -1.10843885e+00 9.89759900e-03
6.18613064e-01 -2.03573823e-01 2.45162502e-01 2.85022885e-01
-3.27264994e-01 -9.01879489e-01 -3.85210037e-01 2.63617635e-01
7.49208331e-01 7.23456204e-01 -1.30978137e-01 -3.82225215e-01
4.00611907e-01 -1.88915029e-01 3.77355069e-01 4.82339203e-01
3.74996364e-01 1.15096855e+00 -3.14417899e-01 -3.88773203e-01
-1.21425021e+00 -1.63367796e+00 -3.22475016e-01 7.53195524e-01
3.79225016e-01 -1.33190617e-01 -1.00490534e+00 -3.02604824e-01
1.27487928e-02 9.79383051e-01 -5.00931084e-01 3.37399960e-01
-6.06028974e-01 -3.69858146e-01 7.94722065e-02 4.49438304e-01
8.44230533e-01 -7.14641869e-01 -5.38111329e-01 7.45810270e-02
-3.17928582e-01 -9.62080300e-01 -7.74357319e-01 -2.99634218e-01
-8.51280391e-01 -9.36187446e-01 -8.98950338e-01 -5.31604528e-01
9.26342607e-01 5.33994675e-01 1.38847971e+00 -1.68861449e-02
-4.65306550e-01 7.90471494e-01 1.60442993e-01 -1.11002393e-01
-2.43070021e-01 -5.72702885e-01 -2.69677162e-01 -1.26857147e-01
-8.09150398e-01 -9.71000850e-01 -1.12826204e+00 3.43363285e-01
-9.45084512e-01 7.77949691e-01 -2.27512971e-01 5.24709105e-01
1.00127101e+00 -1.80467427e-01 -2.72138596e-01 -1.01524556e+00
1.06858194e-01 -1.82709634e-01 -7.74082243e-01 1.48253858e-01
-1.92683384e-01 -7.96263516e-02 5.25785148e-01 1.16046108e-01
-1.71669388e+00 -5.29909171e-02 -3.13830137e-01 -3.85283947e-01
-2.63325065e-01 -2.25344151e-01 -2.40533337e-01 -3.40935588e-01
6.30755663e-01 3.17548394e-01 -5.69093704e-01 -3.98879111e-01
6.26790524e-01 7.27264956e-02 6.27595484e-01 -9.18040991e-01
8.04294646e-01 1.30863214e+00 1.23779275e-01 -8.35051239e-01
-1.07286370e+00 -1.51629865e-01 -5.76247036e-01 -4.56994802e-01
9.69945431e-01 -1.20825493e+00 -6.19702995e-01 8.57187688e-01
-1.41082859e+00 -9.92723167e-01 -6.06337309e-01 1.63582265e-01
-7.90754914e-01 1.63308591e-01 -5.65386117e-01 -7.17750251e-01
-2.03690097e-01 -1.10145020e+00 1.62329102e+00 4.61404957e-02
1.80148020e-01 -1.20878935e+00 3.52777392e-01 7.06878364e-01
3.35582972e-01 7.28233457e-01 5.54894447e-01 8.85486662e-01
-1.27938521e+00 6.85656369e-01 -3.31915855e-01 3.01521599e-01
1.92801818e-01 5.30896150e-02 -1.42479861e+00 -1.38755828e-01
-6.58856239e-04 2.54789274e-02 8.58505726e-01 7.88928390e-01
1.48966849e+00 -1.81144670e-01 -2.34048471e-01 1.53240204e+00
1.71221328e+00 7.45949000e-02 7.83084452e-01 -9.96237546e-02
1.31332660e+00 5.21486700e-01 2.75753617e-01 5.42604566e-01
7.42770374e-01 5.11786878e-01 5.48414290e-01 -2.39729747e-01
-6.26016974e-01 -6.03815615e-01 1.43694177e-01 8.03534031e-01
-3.02390844e-01 -3.74409646e-01 -6.29912972e-01 3.33309323e-01
-1.39864516e+00 -8.56019437e-01 -5.67630470e-01 2.19975567e+00
6.28660917e-01 -4.03826773e-01 -4.52169567e-01 -3.29945683e-01
3.13630044e-01 4.30225760e-01 -6.65814519e-01 -2.66408212e-02
-1.25813454e-01 2.51276642e-01 5.58707178e-01 1.18945920e+00
-7.87604272e-01 7.17164397e-01 5.61205626e+00 8.31790566e-01
-7.91088700e-01 4.09309864e-01 8.86388600e-01 -2.33503729e-01
-1.35688424e+00 -1.82460323e-01 -4.18431759e-01 2.74644554e-01
3.83658677e-01 2.02121511e-01 9.63177800e-01 5.06658554e-01
4.20035362e-01 -3.09949815e-01 -9.34565604e-01 1.17322052e+00
4.76721562e-02 -1.85211933e+00 -1.31872119e-02 7.36333057e-02
1.54156160e+00 1.02565527e-01 2.48312987e-02 -2.64797956e-01
6.25452638e-01 -9.57913697e-01 1.11747837e+00 1.00905359e+00
1.23544681e+00 -6.99906290e-01 -1.56428069e-01 1.46950975e-01
-1.29278755e+00 3.58471751e-01 -1.20460801e-02 1.90228716e-01
5.25540769e-01 1.01830864e+00 -6.29673824e-02 8.60059142e-01
7.68203795e-01 9.23366070e-01 -6.44050613e-02 6.75983965e-01
-2.49647290e-01 2.85006940e-01 -2.90725023e-01 4.70428467e-01
-1.46421790e-01 -6.26831055e-01 5.50688922e-01 8.94109488e-01
3.81897867e-01 3.46055210e-01 -6.06074482e-02 1.61557269e+00
-3.03203732e-01 -2.37347767e-01 -7.16713905e-01 5.35238564e-01
2.22385928e-01 1.07137895e+00 -6.43441916e-01 -2.14480698e-01
-9.05261263e-02 1.47220719e+00 8.78959000e-02 7.62197971e-01
-9.58590150e-01 -1.16584174e-01 6.86045885e-01 4.81356770e-01
7.44334608e-02 -4.58192170e-01 -6.89826369e-01 -1.41841233e+00
1.26296788e-01 -2.45154962e-01 -2.60089189e-01 -1.34164190e+00
-1.24316204e+00 3.12700510e-01 3.06405630e-02 -9.39512968e-01
3.03623319e-01 -3.49310428e-01 -6.06400251e-01 8.33713472e-01
-1.71200407e+00 -1.37823188e+00 -7.46807516e-01 6.62005961e-01
6.11872196e-01 4.86582905e-01 5.85434139e-01 3.86934161e-01
-1.45904228e-01 6.71829563e-03 5.76735914e-01 -2.19285831e-01
4.38726753e-01 -1.24133635e+00 7.69257069e-01 8.39882135e-01
-1.05974503e-01 4.50210929e-01 4.18613285e-01 -8.05154502e-01
-1.46871936e+00 -1.41995227e+00 1.49953112e-01 -5.78204155e-01
3.92054394e-03 -6.34927213e-01 -5.91793716e-01 5.27734697e-01
7.78583586e-02 2.66708344e-01 3.81807894e-01 -5.49219906e-01
-2.27999449e-01 -5.43117449e-02 -1.62580442e+00 5.85032046e-01
1.76868808e+00 -6.54591084e-01 1.23960160e-01 5.44095099e-01
1.00538647e+00 -8.35169435e-01 -7.51199007e-01 3.62995327e-01
5.08100152e-01 -1.36459279e+00 1.52205932e+00 1.13796093e-01
8.06429327e-01 -4.76188391e-01 -5.37745833e-01 -1.43421173e+00
-1.10225126e-01 -7.82797277e-01 -1.08603135e-01 1.05002213e+00
-9.77687985e-02 -6.62523925e-01 6.43991470e-01 4.97410059e-01
-5.83102763e-01 -7.00426400e-01 -7.32599854e-01 -3.39243829e-01
2.65550483e-02 -6.77861392e-01 7.77976453e-01 9.22919691e-01
-1.21167946e+00 -1.77542418e-01 -1.72484934e-01 2.42611840e-01
1.29630554e+00 2.66846210e-01 8.63155365e-01 -9.13965106e-01
-3.65819037e-01 -7.66466036e-02 3.32601249e-01 -1.44774699e+00
5.33815101e-03 -8.67343724e-01 1.64106399e-01 -2.12170625e+00
9.43474844e-02 -5.14826715e-01 4.43317920e-01 -1.62730396e-01
1.13117702e-01 3.97655576e-01 -1.49609983e-01 -8.39257315e-02
-2.69454151e-01 1.00830805e+00 1.87482464e+00 -5.21921776e-02
-1.04216561e-01 -1.97767273e-01 -4.19851094e-01 8.37860942e-01
3.92312378e-01 -3.34451646e-01 -5.12770474e-01 -9.30435181e-01
5.61966896e-01 2.51238495e-02 9.52859759e-01 -9.25483406e-01
-9.39599797e-02 -4.14912671e-01 5.53562224e-01 -7.35065997e-01
7.36928105e-01 -6.68558657e-01 6.25873268e-01 1.30451858e-01
-1.24896713e-01 -2.71137297e-01 1.16068989e-01 7.43085921e-01
3.01397890e-01 2.87325144e-01 8.61660480e-01 -5.06136119e-01
-3.36174726e-01 6.44508064e-01 1.57629102e-01 2.55737841e-01
7.62393236e-01 -4.09725964e-01 -1.53210893e-01 -3.30539823e-01
-4.00266171e-01 -5.54774031e-02 1.07729578e+00 -8.63170773e-02
9.49949503e-01 -1.55444503e+00 -6.00771546e-01 3.92480314e-01
-8.64116699e-02 8.58244896e-01 8.31253231e-01 4.57301557e-01
-1.14588821e+00 -3.68057340e-02 1.86531514e-01 -6.22568667e-01
-1.02678537e+00 -1.09735705e-01 6.18962765e-01 1.21756218e-01
-8.37531030e-01 1.24776113e+00 8.34624708e-01 -1.14972532e+00
-8.52838904e-02 -7.34576941e-01 6.14593923e-01 -5.85304439e-01
3.22802603e-01 7.59618104e-01 -7.92556107e-02 -5.06876528e-01
-2.04935431e-01 9.51855123e-01 5.85421741e-01 -3.01891893e-01
1.56115091e+00 -4.10104185e-01 -2.61860013e-01 5.27826309e-01
1.14953077e+00 4.94949609e-01 -1.87224507e+00 -3.22622396e-02
-1.06029737e+00 -1.04597521e+00 3.95506501e-01 -8.46436799e-01
-1.08217394e+00 8.85059357e-01 3.44298363e-01 -3.95767570e-01
8.70918512e-01 -1.46709546e-01 1.07260716e+00 -1.17196785e-02
7.03940868e-01 -6.89501047e-01 2.76796194e-03 4.64498848e-01
1.13162506e+00 -9.76328373e-01 2.76200503e-01 -8.02457154e-01
-3.01192075e-01 8.75775754e-01 2.71154404e-01 -2.98089474e-01
8.18854511e-01 4.20424372e-01 -3.21320683e-01 -4.43289638e-01
-1.60015568e-01 3.88059974e-01 5.05149066e-01 6.07173860e-01
3.20223212e-01 1.73283488e-01 4.93795484e-01 3.62831429e-02
-2.91992605e-01 6.79367632e-02 4.23908710e-01 8.40505421e-01
-9.87893194e-02 -6.53747737e-01 -4.64554787e-01 2.18048200e-01
7.10232556e-02 -3.64562511e-01 3.68586481e-02 1.32941604e-01
1.93093136e-01 6.02917135e-01 1.37551844e-01 5.11018671e-02
4.92373884e-01 -5.55273950e-01 1.00004685e+00 -6.33038223e-01
-3.97186905e-01 5.32352105e-02 -1.43901417e-02 -1.00069427e+00
-5.11686146e-01 -6.19929910e-01 -1.38016593e+00 -6.19478226e-01
1.03063077e-01 -4.19082850e-01 8.66612077e-01 4.95570332e-01
4.29116130e-01 8.46460819e-01 7.44442821e-01 -1.35634017e+00
2.70843923e-01 -3.28528792e-01 -5.18468678e-01 2.01541945e-01
5.68683028e-01 -6.52396083e-01 -3.65092337e-01 2.19004840e-01] | [9.618024826049805, -3.1217360496520996] |
5fc77b47-b9b9-4573-84f9-7485401ce63d | weighted-combination-of-bert-and-n-gram | null | null | https://aclanthology.org/2020.wanlp-1.27 | https://aclanthology.org/2020.wanlp-1.27.pdf | Weighted combination of BERT and N-GRAM features for Nuanced Arabic Dialect Identification | Around the Arab world, different Arabic dialects are spoken by more than 300M persons, and are increasingly popular in social media texts. However, Arabic dialects are considered to be low-resource languages, limiting the development of machine-learning based systems for these dialects. In this paper, we investigate the Arabic dialect identification task, from two perspectives: country-level dialect identification from 21 Arab countries, and province-level dialect identification from 100 provinces. We introduce an unified pipeline of state-of-the-art models, that can handle the two subtasks. Our experimental studies applied to the NADI shared task, show promising results both at the country-level (F1-score of 25.99%) and the province-level (F1-score of 6.39%), and thus allow us to be ranked 2nd for the country-level subtask, and 1st in the province-level subtask. | ['Ismail Berrada', 'Ahmed Khoumsi', 'Hamza Alami', 'Ahmed Alami', 'Abdellah El Mekki'] | null | null | null | null | coling-wanlp-2020-12 | ['dialect-identification'] | ['natural-language-processing'] | [-6.91040218e-01 -2.44660407e-01 -1.31895363e-01 -4.48567420e-01
-9.73190248e-01 -1.02215123e+00 1.28949690e+00 -1.54547524e-02
-6.00754023e-01 4.00319248e-01 3.35092634e-01 -4.02484208e-01
1.85342073e-01 -7.43828237e-01 -1.93168968e-01 -6.29249871e-01
-1.33617476e-01 9.99143779e-01 2.50814878e-03 -8.44832659e-01
1.74232662e-01 4.80966121e-01 -1.24593818e+00 4.97734338e-01
1.52546322e+00 4.28916335e-01 -1.39113590e-01 4.06117320e-01
-6.62436038e-02 5.94795763e-01 -4.00768638e-01 -8.54064643e-01
6.03203662e-02 -3.20507884e-01 -1.18320143e+00 -1.86682194e-01
6.14903688e-01 -4.48399007e-01 -4.87554595e-02 1.11781251e+00
2.81285137e-01 -3.87978643e-01 1.01114428e+00 -6.75393760e-01
-6.71542764e-01 8.90914142e-01 -5.64197898e-01 -2.72265166e-01
4.65368718e-01 -1.57989636e-01 1.13719869e+00 -1.23998737e+00
6.24034464e-01 1.50615847e+00 5.14283419e-01 2.43637666e-01
-9.50843394e-01 -6.66482925e-01 2.27295727e-01 3.29037637e-01
-1.56432557e+00 -5.91398895e-01 3.40163469e-01 -6.01387918e-01
8.95911336e-01 1.71516240e-01 3.56889397e-01 7.76519597e-01
-1.78400800e-01 1.02851892e+00 1.50648785e+00 -5.00114024e-01
-2.52480149e-01 2.23374933e-01 6.44098520e-02 7.63194084e-01
-2.79713478e-02 -3.70689631e-01 -4.67102796e-01 1.22837527e-02
6.13960922e-01 -4.81983215e-01 -5.92420399e-02 2.68060893e-01
-1.37781370e+00 1.08019388e+00 4.28004652e-01 5.16385019e-01
-3.33454311e-01 -7.00835347e-01 2.88353503e-01 5.56628764e-01
7.23038435e-01 1.82896793e-01 -5.41080832e-01 -2.42535949e-01
-8.59838247e-01 5.88065028e-01 7.92397201e-01 7.81666398e-01
6.10061467e-01 -6.57934323e-02 1.88819408e-01 1.55032682e+00
4.89469647e-01 9.65405405e-01 2.81660944e-01 -2.73023337e-01
6.30239248e-01 6.91508293e-01 4.49222401e-02 -6.45368934e-01
-5.60995936e-01 -3.91776413e-02 -9.20582533e-01 -7.87743647e-03
1.12421417e+00 -3.19058925e-01 -8.76295507e-01 1.63278782e+00
9.98486057e-02 -5.13659537e-01 -1.02609955e-01 1.08194423e+00
6.76685572e-01 8.57161462e-01 -3.02864939e-01 2.31632411e-01
1.66382122e+00 -1.01115537e+00 -3.46262366e-01 -1.89737290e-01
6.00176692e-01 -1.11491895e+00 1.04597962e+00 4.95102882e-01
-9.61732268e-01 -3.37922633e-01 -8.16345692e-01 2.32225463e-01
-6.63177907e-01 3.49110603e-01 3.14861953e-01 9.64558363e-01
-1.25136817e+00 -2.18282968e-01 -4.78448868e-01 -6.47969961e-01
6.87268004e-02 1.17485896e-01 -7.07137525e-01 -5.07468104e-01
-1.45674562e+00 1.16632593e+00 1.24418169e-01 1.96997657e-01
-7.51203358e-01 -3.99062335e-01 -7.61286497e-01 -1.85979605e-01
-1.96575060e-01 9.42758322e-02 8.19948196e-01 -8.74644756e-01
-1.51411402e+00 1.35209012e+00 -1.73717737e-01 -8.96710008e-02
4.90954995e-01 -2.25715682e-01 -9.23774898e-01 -2.10654780e-01
2.73347616e-01 4.28067088e-01 4.42622960e-01 -1.06893504e+00
-1.14887452e+00 -5.63471317e-01 1.26352370e-01 4.93757516e-01
-3.26467901e-01 5.58327258e-01 -5.67616999e-01 -6.44126534e-01
2.12103590e-01 -1.14250147e+00 2.23746151e-02 -8.68719876e-01
-2.95539498e-01 -3.85189265e-01 3.74204010e-01 -1.29582345e+00
1.14888656e+00 -2.01332259e+00 3.27176988e-01 6.07692003e-01
-5.82735948e-02 1.76786348e-01 -2.30414838e-01 4.92538989e-01
9.33877900e-02 -1.92839485e-02 -2.85922527e-01 -1.65931150e-01
5.08620292e-02 -2.44319409e-01 -3.32120545e-02 5.78603327e-01
4.47929054e-01 5.34271359e-01 -9.93993819e-01 9.76703539e-02
-6.56138500e-03 2.81621069e-01 -4.72975969e-01 1.31925195e-02
1.54639423e-01 3.42569411e-01 -1.63330864e-02 1.03874969e+00
1.05924368e+00 3.37137550e-01 4.99329954e-01 2.04945803e-01
-5.00175357e-01 5.12002468e-01 -9.86963034e-01 1.28603315e+00
-5.39924085e-01 6.99454308e-01 4.80071068e-01 -4.89787102e-01
9.47655439e-01 1.42519787e-01 8.98415372e-02 -7.62482464e-01
-1.09943412e-01 8.28416646e-01 3.63049746e-01 -2.99781971e-02
6.35495245e-01 1.01336658e-01 -5.21108806e-01 4.97251928e-01
8.18273723e-02 -4.33169715e-02 3.89863312e-01 -3.58413309e-02
3.73780549e-01 -1.18744314e-01 2.95754701e-01 -9.24705327e-01
1.02408898e+00 7.27490634e-02 -4.70220484e-02 6.74999058e-01
-1.32480785e-01 7.51860499e-01 5.05515635e-01 -6.37936532e-01
-6.70651793e-01 -1.23032224e+00 -4.24531490e-01 1.52995694e+00
-3.13828349e-01 -1.99412107e-01 -9.48563159e-01 -8.42636526e-01
-9.75143239e-02 3.90380889e-01 -4.06672210e-01 3.37749124e-01
-8.20775270e-01 -1.24445593e+00 9.54815805e-01 4.87096384e-02
7.62320340e-01 -8.84589076e-01 2.46208623e-01 2.13962197e-01
-6.58232391e-01 -9.93658006e-01 -4.56782639e-01 -2.92048484e-01
-2.44835347e-01 -9.89465475e-01 -1.35864711e+00 -1.23365605e+00
3.42838347e-01 -5.30679114e-02 1.40090740e+00 -1.21180184e-01
4.26785499e-01 -1.35316253e-01 -3.86606961e-01 -3.89097154e-01
-5.74882329e-01 6.14830554e-01 1.87681898e-01 3.35882962e-01
6.01957560e-01 1.73386171e-01 -1.37184858e-01 5.20704985e-01
-4.34437990e-01 -1.97038189e-01 4.59150106e-01 6.75849259e-01
1.65753625e-02 -2.23903045e-01 6.59178615e-01 -1.01370609e+00
4.72601116e-01 -4.96810764e-01 -8.07320952e-01 3.22185248e-01
-2.24553540e-01 -3.40918869e-01 6.09872758e-01 -5.79567775e-02
-1.16616988e+00 3.03840190e-02 -6.69558644e-01 7.24070787e-01
-4.09390688e-01 8.90894949e-01 -3.60460997e-01 2.04196990e-01
6.43930614e-01 3.04224700e-01 -7.38139972e-02 -3.31519842e-01
5.45083702e-01 1.30868888e+00 4.26445872e-01 -5.49569845e-01
6.17982388e-01 2.37207025e-01 -5.37569344e-01 -1.23371494e+00
-5.31061411e-01 -6.23966932e-01 -1.01897705e+00 -2.77842373e-01
8.21165085e-01 -1.40582144e+00 -4.31978583e-01 1.24129307e+00
-9.18443263e-01 -3.38457614e-01 7.54278898e-01 3.03225428e-01
-1.96341887e-01 1.31326810e-01 -7.77834535e-01 -5.72837472e-01
-1.91060036e-01 -1.38423884e+00 8.29336524e-01 -7.44623393e-02
-3.71711463e-01 -1.11781335e+00 1.75737098e-01 3.12566161e-01
2.99456686e-01 -6.86187968e-02 1.22338533e+00 -7.96970546e-01
-1.41070709e-01 -1.22184850e-01 -4.08365220e-01 5.05230203e-02
4.58251119e-01 -1.45788565e-01 -1.29258513e+00 -3.65353316e-01
-5.96381009e-01 -2.33663633e-01 8.47387612e-01 3.29995543e-01
3.17562252e-01 5.03269723e-03 6.51329085e-02 3.00358742e-01
1.05724359e+00 -5.99626265e-02 4.90029275e-01 4.98550147e-01
7.95206010e-01 9.38865721e-01 5.69238901e-01 1.74175024e-01
9.91636693e-01 7.93284416e-01 3.39223891e-02 -1.85781494e-01
8.12850147e-03 8.98868665e-02 8.37723792e-01 1.19694746e+00
-4.47054029e-01 1.64803341e-01 -1.61883640e+00 9.91272449e-01
-1.76086020e+00 -7.51093209e-01 -4.94846195e-01 2.40822148e+00
9.21158552e-01 -1.34749502e-01 7.57410467e-01 -6.04878999e-02
6.46264911e-01 4.96678688e-02 9.04140472e-02 -4.90088284e-01
-5.64204812e-01 -2.05588937e-01 2.85142034e-01 8.81687105e-01
-1.48238027e+00 1.55457008e+00 6.35613680e+00 8.73884320e-01
-1.18027747e+00 -1.46776095e-01 7.06997752e-01 3.57193351e-01
-8.46396610e-02 -3.83717477e-01 -9.80930984e-01 4.33942407e-01
1.05954242e+00 1.19252399e-01 7.22046614e-01 3.24791342e-01
-3.51239555e-02 4.49070036e-02 -8.86799812e-01 8.67504656e-01
1.86797142e-01 -9.41296637e-01 3.21977407e-01 1.77508235e-01
8.52020919e-01 5.30745029e-01 1.91862524e-01 4.29598361e-01
5.54827511e-01 -1.21466076e+00 9.17399228e-01 -1.42235458e-02
9.24882650e-01 -1.18097317e+00 1.00260949e+00 2.44370401e-01
-1.21264637e+00 1.09221235e-01 -2.67482132e-01 4.60162349e-02
3.40273976e-02 3.60615969e-01 -8.78655195e-01 7.70291507e-01
9.10311162e-01 8.78398418e-01 -7.32251585e-01 6.66124165e-01
-3.77308071e-01 9.51573670e-01 -2.67894685e-01 -2.33760979e-02
7.22412109e-01 -6.81220591e-01 3.06698442e-01 1.56459463e+00
2.99427658e-01 -4.78113353e-01 2.11753294e-01 1.60046384e-01
7.21074045e-02 4.73181427e-01 -5.20922422e-01 6.99328706e-02
2.55113304e-01 1.20104504e+00 -5.43581307e-01 -8.58791620e-02
-7.68268943e-01 1.02793980e+00 4.33392048e-01 3.02874923e-01
-5.76927364e-01 -4.22987908e-01 9.16846931e-01 -1.86866209e-01
-1.04674697e-01 -2.20173568e-01 -1.22961476e-01 -1.24560094e+00
-2.72663563e-01 -1.35640311e+00 6.86934173e-01 -1.40947238e-01
-1.39606094e+00 7.77671933e-01 -2.90818840e-01 -1.04346609e+00
-6.69884145e-01 -1.07193506e+00 -2.04219013e-01 1.23341680e+00
-1.63831508e+00 -1.79287720e+00 1.26594365e-01 7.63071060e-01
4.36998427e-01 -8.47267747e-01 1.22833192e+00 5.48346162e-01
-3.20729077e-01 7.43181109e-01 5.05964935e-01 7.74698019e-01
1.13284695e+00 -1.52604890e+00 6.18614495e-01 8.25665772e-01
2.75502861e-01 5.99368989e-01 1.81624308e-01 -3.16158265e-01
-1.02095556e+00 -9.19016540e-01 1.64660811e+00 -7.28878140e-01
8.83589864e-01 -7.28252530e-01 -7.50122070e-01 5.59649229e-01
2.68286437e-01 -7.80078292e-01 7.07529545e-01 8.50994170e-01
-7.08931863e-01 4.03171480e-02 -1.13755560e+00 8.42248082e-01
5.46505570e-01 -8.70023727e-01 -3.10090363e-01 3.71189207e-01
-1.20049469e-01 -1.85916662e-01 -9.74674881e-01 -7.14178905e-02
9.35911834e-01 -8.56079876e-01 9.45693791e-01 -3.53494197e-01
3.64502728e-01 -1.47925526e-01 -2.17837483e-01 -1.89433110e+00
-5.58045685e-01 -2.51554459e-01 5.48503041e-01 1.34583426e+00
8.98342907e-01 -8.04085433e-01 4.27542031e-01 2.05899514e-02
-1.01873562e-01 -1.09871253e-01 -9.06286299e-01 -6.93175316e-01
8.53877902e-01 -8.17822069e-02 8.07777464e-01 1.27140343e+00
1.36848614e-01 5.16899228e-01 -3.47257048e-01 4.00659859e-01
1.79016873e-01 1.10556379e-01 6.59058332e-01 -1.29310572e+00
2.72543907e-01 -1.15846109e+00 -3.78476419e-02 -7.50040531e-01
3.64391774e-01 -1.06348789e+00 1.95509717e-02 -1.45017719e+00
-6.19321596e-03 -3.76116097e-01 -2.64734053e-03 4.04324651e-01
-3.09419453e-01 4.30183679e-01 2.30096117e-01 2.25108683e-01
-2.74608463e-01 1.01118244e-01 8.20538342e-01 -3.68590444e-01
-1.54622555e-01 8.60634074e-02 -5.96752763e-01 7.43282676e-01
7.92022645e-01 6.75986931e-02 4.69844341e-02 -7.57977426e-01
2.28167862e-01 -5.02138913e-01 -3.89551014e-01 -6.11326277e-01
-1.10559829e-01 -1.32995665e-01 3.16300660e-01 -5.85520148e-01
1.17089674e-01 -5.32986403e-01 -5.12815356e-01 5.79550147e-01
2.48071089e-01 3.11653554e-01 2.64447421e-01 -3.27368468e-01
-5.83852708e-01 3.60676609e-02 9.40405726e-01 -3.12493108e-02
-8.22692692e-01 2.38105133e-01 -1.01141250e+00 8.46338645e-02
4.25333649e-01 1.38775796e-01 -2.36442313e-01 -5.72330177e-01
-1.77840129e-01 2.82829612e-01 4.05960858e-01 6.50435507e-01
2.04810306e-01 -1.21594429e+00 -1.70893121e+00 5.58868349e-01
6.44620836e-01 -3.46306890e-01 9.02527198e-03 6.61633611e-01
-9.28075016e-01 5.85493207e-01 -4.47690994e-01 -5.56047976e-01
-1.13354933e+00 -2.67300814e-01 3.23266417e-01 -4.02773261e-01
1.71633005e-01 8.21953118e-01 1.40741140e-01 -1.55610693e+00
-1.87542632e-01 2.47025043e-01 -7.71563590e-01 7.11476505e-01
7.88914859e-01 5.11037886e-01 2.93525994e-01 -1.29500282e+00
-7.00942218e-01 7.00798213e-01 -4.10872161e-01 -4.82401371e-01
1.10260320e+00 -7.55364746e-02 -6.75612450e-01 3.69683594e-01
8.82084310e-01 5.49318731e-01 -5.15322506e-01 -1.05825216e-01
3.80613476e-01 -2.19538599e-01 -2.57518232e-01 -1.26052892e+00
-6.81962490e-01 6.61738455e-01 4.23352867e-01 3.29351246e-01
8.17430556e-01 -5.55376336e-02 3.88156116e-01 3.47338706e-01
4.73675340e-01 -1.12029278e+00 -6.45011365e-01 1.58757579e+00
8.11938643e-01 -1.50972724e+00 -3.51225853e-01 -4.25353527e-01
-6.67012036e-01 9.18416798e-01 4.21893001e-01 1.44834638e-01
8.29367936e-01 -6.68379813e-02 8.30561876e-01 1.50317997e-01
-8.38835016e-02 -4.77235168e-01 6.67031050e-01 7.73536563e-01
1.06364202e+00 4.91039872e-01 -2.10705802e-01 5.68205714e-01
-5.16392648e-01 -7.33584821e-01 4.24015582e-01 2.76491284e-01
-2.63090819e-01 -1.19618523e+00 -5.80824852e-01 3.03976387e-01
-4.71951336e-01 -4.14632469e-01 -7.67239153e-01 9.77072001e-01
-1.55030757e-01 1.25073075e+00 3.23304087e-01 -3.57114077e-01
2.23734692e-01 6.75312653e-02 3.99893999e-01 -3.29492629e-01
-9.51147437e-01 1.47929072e-01 4.47166473e-01 -8.91968384e-02
-2.51917183e-01 -8.75183582e-01 -1.00059426e+00 -7.08128810e-01
1.17687918e-01 5.52350320e-02 4.54511404e-01 9.45543885e-01
-1.56879872e-01 -3.03325385e-01 8.18497419e-01 -7.09353387e-01
-1.83659390e-01 -1.24504948e+00 -6.95988357e-01 4.02968496e-01
2.69994348e-01 -2.42457479e-01 3.06383763e-02 -1.34603336e-01] | [10.183980941772461, 10.770488739013672] |
fadaf665-7a8a-4947-84bf-043107374857 | multimodal-information-bottleneck-learning | 2210.17444 | null | https://arxiv.org/abs/2210.17444v3 | https://arxiv.org/pdf/2210.17444v3.pdf | Multimodal Information Bottleneck: Learning Minimal Sufficient Unimodal and Multimodal Representations | Learning effective joint embedding for cross-modal data has always been a focus in the field of multimodal machine learning. We argue that during multimodal fusion, the generated multimodal embedding may be redundant, and the discriminative unimodal information may be ignored, which often interferes with accurate prediction and leads to a higher risk of overfitting. Moreover, unimodal representations also contain noisy information that negatively influences the learning of cross-modal dynamics. To this end, we introduce the multimodal information bottleneck (MIB), aiming to learn a powerful and sufficient multimodal representation that is free of redundancy and to filter out noisy information in unimodal representations. Specifically, inheriting from the general information bottleneck (IB), MIB aims to learn the minimal sufficient representation for a given task by maximizing the mutual information between the representation and the target and simultaneously constraining the mutual information between the representation and the input data. Different from general IB, our MIB regularizes both the multimodal and unimodal representations, which is a comprehensive and flexible framework that is compatible with any fusion methods. We develop three MIB variants, namely, early-fusion MIB, late-fusion MIB, and complete MIB, to focus on different perspectives of information constraints. Experimental results suggest that the proposed method reaches state-of-the-art performance on the tasks of multimodal sentiment analysis and multimodal emotion recognition across three widely used datasets. The codes are available at \url{https://github.com/TmacMai/Multimodal-Information-Bottleneck}. | ['Haifeng Hu', 'Ying Zeng', 'Sijie Mai'] | 2022-10-31 | null | null | null | null | ['multimodal-sentiment-analysis', 'multimodal-emotion-recognition', 'multimodal-sentiment-analysis', 'multimodal-emotion-recognition'] | ['computer-vision', 'computer-vision', 'natural-language-processing', 'speech'] | [ 6.51834533e-02 -1.37435064e-01 -2.03652591e-01 -2.97428906e-01
-9.39848244e-01 -4.73456830e-01 4.63126779e-01 2.36163497e-01
-2.87036151e-01 4.60065037e-01 1.42172650e-01 6.02741130e-02
-3.86960745e-01 -4.37819332e-01 -4.55075145e-01 -1.13072622e+00
2.25508869e-01 5.29116951e-02 -3.80992651e-01 -4.77984130e-01
-4.86400947e-02 7.99264386e-02 -1.59129655e+00 3.19674402e-01
9.27227676e-01 1.11930633e+00 1.31625205e-01 2.02211052e-01
-2.14151308e-01 3.33950788e-01 -1.63635314e-01 -5.95386922e-01
-1.16552033e-01 -2.86060244e-01 -5.44656456e-01 -7.16621354e-02
-1.42436936e-01 1.91568553e-01 -4.09031719e-01 1.10530055e+00
6.13443673e-01 2.27045685e-01 8.01162720e-01 -1.57652771e+00
-2.76964575e-01 6.39144301e-01 -7.73202062e-01 -1.95247650e-01
3.32551867e-01 -1.01050392e-01 1.17462611e+00 -1.08112013e+00
3.29238206e-01 1.26536775e+00 3.36121082e-01 6.81843877e-01
-1.38822603e+00 -5.41765392e-01 4.97442216e-01 -6.36470914e-02
-1.42530239e+00 -3.97686481e-01 1.13161445e+00 -4.62202579e-01
4.27998632e-01 5.07118523e-01 4.27760333e-01 1.28163195e+00
9.22037140e-02 1.12574649e+00 8.23754311e-01 -2.17589095e-01
-2.70113144e-02 2.56446898e-01 3.50435138e-01 6.02487862e-01
-4.93771881e-02 -1.12851627e-01 -7.24795282e-01 -3.25494528e-01
1.97260961e-01 1.84881583e-01 -4.87531304e-01 -3.98216337e-01
-1.25554287e+00 8.23992968e-01 2.81860560e-01 4.19312328e-01
-2.01753929e-01 -2.18634441e-01 4.34178740e-01 2.83328861e-01
2.92480886e-01 6.21045791e-02 -3.84884655e-01 -1.70664594e-01
-4.08775419e-01 3.66430543e-02 5.18063486e-01 5.30326962e-01
1.01028943e+00 -1.12410828e-01 2.16666749e-03 1.03973174e+00
5.87397158e-01 5.51017523e-01 5.84390879e-01 -6.27200603e-01
7.10304677e-01 1.04309189e+00 3.12466659e-02 -1.21362627e+00
-4.77364480e-01 -2.96753999e-02 -1.14539921e+00 -1.53619364e-01
2.62832433e-01 -3.32501113e-01 -4.27705795e-01 2.19196987e+00
2.37374321e-01 -2.36456856e-01 2.45460913e-01 1.08222401e+00
9.64991391e-01 7.00734913e-01 -7.38994963e-03 -2.69185752e-01
1.14946008e+00 -5.68018258e-01 -7.79736578e-01 -1.72968403e-01
5.47683299e-01 -7.28448033e-01 1.02342045e+00 3.62147927e-01
-8.77509356e-01 -4.80629027e-01 -1.03336322e+00 1.21203870e-01
-4.73332733e-01 2.48500377e-01 6.39669120e-01 3.96763921e-01
-6.60963774e-01 2.33231574e-01 -7.66131163e-01 -1.03913367e-01
6.56653615e-03 4.27686125e-01 -7.30235040e-01 -6.98586255e-02
-1.33281481e+00 7.77137578e-01 5.09222269e-01 4.93788272e-01
-5.76049685e-01 -3.70046794e-01 -1.09424269e+00 -5.47393076e-02
2.76015639e-01 -5.75737894e-01 7.14767039e-01 -1.15057194e+00
-1.36458242e+00 4.14482951e-01 -2.88912833e-01 1.99936554e-01
2.01888785e-01 -1.84744224e-01 -3.86924207e-01 -7.23335370e-02
-2.27904797e-01 6.38800561e-01 9.31492567e-01 -1.49218714e+00
-2.26235479e-01 -5.65760970e-01 -1.12851664e-01 3.79759520e-01
-7.35775054e-01 -2.55679309e-01 -5.67930460e-01 -5.34041941e-01
2.81132162e-01 -8.76934588e-01 -1.07978500e-01 -4.40028399e-01
-4.04077560e-01 -3.31287831e-01 7.33017445e-01 -4.63305295e-01
1.58168876e+00 -2.46792793e+00 9.11052287e-01 2.48404399e-01
1.03436954e-01 -7.82154277e-02 -4.38730896e-01 6.67342067e-01
-2.09196657e-01 1.14766158e-01 -3.54454696e-01 -6.70521796e-01
2.06482664e-01 2.58708626e-01 -2.65672892e-01 3.86780083e-01
4.29248065e-01 7.68005788e-01 -7.20723748e-01 -4.26142812e-01
2.33945921e-01 8.32741082e-01 -2.91673094e-01 2.24680319e-01
9.53719243e-02 5.83523452e-01 -4.59808141e-01 8.26501727e-01
8.18531573e-01 -1.13322236e-01 3.63626212e-01 -5.78800321e-01
9.88962129e-03 -9.98000726e-02 -1.41632366e+00 1.82232273e+00
-3.09244663e-01 2.58275211e-01 1.38740808e-01 -1.17776060e+00
8.11271131e-01 3.36985111e-01 6.21420205e-01 -5.89694619e-01
3.47185642e-01 1.15428701e-01 -1.69759259e-01 -4.85451132e-01
3.83023649e-01 -1.80391505e-01 -3.57056171e-01 2.89234370e-01
1.69581518e-01 2.02135414e-01 1.72839999e-01 2.16371313e-01
3.62203717e-01 1.81744527e-02 8.49686712e-02 -4.97393534e-02
7.50569582e-01 -6.00910723e-01 7.12671757e-01 2.75967330e-01
9.37227905e-03 5.64609408e-01 7.47225106e-01 -9.04838517e-02
-5.25422513e-01 -1.08115625e+00 -1.19757913e-01 1.12256515e+00
4.84144926e-01 -5.45810819e-01 -3.72853577e-01 -7.16482282e-01
-9.45043378e-03 3.88298333e-01 -6.87805772e-01 -3.93537939e-01
-1.57262251e-01 -1.08033073e+00 3.12849075e-01 2.62598306e-01
9.07541960e-02 -6.54907942e-01 -1.75741017e-01 -1.27009809e-01
-6.94158256e-01 -6.84956908e-01 -4.86564845e-01 2.41697133e-01
-7.11689055e-01 -1.04728615e+00 -6.39890790e-01 -5.08554220e-01
7.62922883e-01 2.85507560e-01 8.15816879e-01 -1.03429981e-01
1.15029281e-02 5.72977066e-01 -4.46655661e-01 -1.29827455e-01
-5.69473729e-02 4.28760946e-02 1.97033025e-02 5.45420945e-01
2.63923287e-01 -3.11781466e-01 -5.17613113e-01 3.35568428e-01
-1.26833928e+00 1.39231324e-01 5.86536825e-01 1.20100546e+00
4.93877858e-01 -5.56869954e-02 5.14231384e-01 -2.52497286e-01
6.49562716e-01 -6.75951660e-01 -3.63700420e-01 3.33709061e-01
-3.10144454e-01 1.96634173e-01 3.67560595e-01 -6.65641308e-01
-1.04463601e+00 1.23728469e-01 -6.46964312e-02 -5.27116656e-01
6.22395277e-02 8.14981401e-01 -5.83789051e-01 1.46699905e-01
1.08784579e-01 2.96139568e-01 2.35331029e-01 -4.60979223e-01
3.66676956e-01 5.71689427e-01 -3.68269943e-02 -6.47013068e-01
3.60916376e-01 3.69482368e-01 -8.57058838e-02 -9.21500742e-01
-5.47566772e-01 -4.63974953e-01 -5.22338748e-01 -3.20725352e-01
6.65991366e-01 -8.95277262e-01 -9.13649201e-01 4.98872101e-01
-1.11624229e+00 1.54035911e-01 9.14684385e-02 6.00940943e-01
-2.67867535e-01 4.90985632e-01 -4.79623675e-01 -1.10089993e+00
-1.28567412e-01 -1.41506886e+00 1.24267375e+00 3.89037549e-01
-1.19714461e-01 -1.10118532e+00 1.67600736e-01 3.53201449e-01
1.39545090e-02 2.38233343e-01 1.01058841e+00 -5.08604646e-01
-1.34777784e-01 -2.72810221e-01 -1.14605308e-01 6.22751296e-01
7.57660195e-02 6.39413223e-02 -9.88204956e-01 -3.68874103e-01
-1.86198413e-01 -5.57150483e-01 1.16324568e+00 1.89129502e-01
6.97488189e-01 -1.38597041e-01 -2.50794709e-01 4.03948605e-01
1.27730620e+00 -7.10412441e-03 5.01623571e-01 3.01382579e-02
6.87234581e-01 8.45537245e-01 5.31850040e-01 6.39023542e-01
5.20349443e-01 6.23689413e-01 5.62745214e-01 -9.70279798e-02
5.83225965e-01 -1.20409831e-01 6.67916954e-01 1.11512840e+00
7.90242255e-02 -2.24855244e-01 -6.22493446e-01 3.50313216e-01
-2.23502183e+00 -7.57626951e-01 7.60480240e-02 2.38422036e+00
7.87369847e-01 -3.17626745e-01 1.93740889e-01 9.22329649e-02
5.56159019e-01 2.94977456e-01 -4.03799117e-01 -2.94214904e-01
-4.23408866e-01 -3.04662913e-01 -8.73636156e-02 5.92272580e-01
-1.24351525e+00 5.93622804e-01 4.73325872e+00 8.91745508e-01
-1.07796085e+00 8.62408429e-02 4.59201753e-01 -1.48335651e-01
-7.04938293e-01 -6.73888102e-02 -7.35086322e-01 6.23002589e-01
6.29144251e-01 1.03584930e-01 3.70241612e-01 4.00008768e-01
8.01579356e-02 -1.31400660e-01 -9.50913787e-01 1.19598305e+00
4.63566110e-02 -7.99165547e-01 2.95141190e-01 6.00116178e-02
4.94188488e-01 -2.03068987e-01 3.05926055e-01 3.87343675e-01
-2.40752771e-01 -8.77711713e-01 5.65649986e-01 9.11760986e-01
4.07329649e-01 -9.78505135e-01 9.91476059e-01 4.43187058e-01
-1.23991132e+00 -1.41048402e-01 -2.41509542e-01 2.79396683e-01
9.72350016e-02 5.74009359e-01 7.29169324e-02 9.88052249e-01
5.13557553e-01 7.64515281e-01 -4.99963135e-01 6.55314863e-01
8.33001658e-02 4.59888428e-02 -2.53165156e-01 -2.44106241e-02
1.30815729e-01 -4.93114918e-01 6.04518652e-01 1.16131353e+00
3.39717448e-01 -1.66084096e-01 2.50783443e-01 6.41499579e-01
6.59116283e-02 3.63285363e-01 -6.83208346e-01 -3.81839424e-01
1.02149345e-01 1.34546030e+00 -3.00206363e-01 -3.02799456e-02
-4.51223403e-01 8.66249204e-01 3.45984519e-01 5.11319995e-01
-8.72985959e-01 -2.81153083e-01 7.65924811e-01 -5.22349179e-01
9.20369402e-02 -1.14124767e-01 -2.45643318e-01 -1.55053091e+00
9.61850658e-02 -8.45572650e-01 5.80517471e-01 -4.06751156e-01
-1.51389956e+00 6.01085246e-01 4.97738719e-02 -1.40410411e+00
-1.74031287e-01 -6.81622028e-01 -4.33445632e-01 8.77862394e-01
-1.33913934e+00 -1.27589631e+00 -1.28655106e-01 7.50338435e-01
1.77575961e-01 -6.60830513e-02 8.88788104e-01 5.25079310e-01
-9.60815549e-01 7.70345986e-01 1.85565650e-01 -1.23738267e-01
7.61915445e-01 -9.80317175e-01 -7.33829558e-01 4.80395228e-01
-2.32565403e-03 8.18361878e-01 5.23794711e-01 -3.36542726e-01
-1.81073332e+00 -6.69864118e-01 7.01259732e-01 -2.95473039e-01
7.65419126e-01 -3.82275671e-01 -9.33156669e-01 3.36240977e-01
1.84395567e-01 -4.13956285e-01 1.03131020e+00 2.12156713e-01
-5.15828729e-01 -3.92729253e-01 -7.07950652e-01 6.08862460e-01
4.56014901e-01 -6.23975277e-01 -3.51071835e-01 9.23200101e-02
5.02384663e-01 -2.72812471e-02 -9.12217557e-01 6.44867599e-01
7.64323354e-01 -9.55336154e-01 8.69884551e-01 -5.56108475e-01
3.21790278e-01 -3.10981244e-01 -5.28926015e-01 -1.16909909e+00
1.74081326e-03 -5.32347262e-01 -2.28670448e-01 1.45594251e+00
5.35726964e-01 -6.06550753e-01 2.85205930e-01 6.22795403e-01
1.02603614e-01 -9.49443877e-01 -1.10442674e+00 -3.90596002e-01
1.11093953e-01 -4.78128999e-01 2.98152655e-01 8.50174725e-01
3.39127421e-01 4.30066794e-01 -6.71516120e-01 1.84058279e-01
4.26506042e-01 2.99238175e-01 6.34039462e-01 -1.07665229e+00
-9.42395031e-02 -6.44450903e-01 -2.54331112e-01 -8.70190620e-01
3.67512077e-01 -8.89947891e-01 -1.26153395e-01 -1.10493624e+00
4.84444708e-01 -4.12674323e-02 -7.36334682e-01 5.31213701e-01
-2.82447159e-01 1.81881562e-01 3.53895068e-01 2.12324813e-01
-7.52414823e-01 1.09999275e+00 1.07180583e+00 -3.33536476e-01
-2.90299386e-01 -2.68174186e-02 -8.63550544e-01 6.36131167e-01
5.90258300e-01 -1.92081377e-01 -4.35329944e-01 -2.32678801e-01
5.14969051e-01 2.10960153e-02 2.39099577e-01 -3.88138831e-01
1.03096984e-01 -2.13638008e-01 2.77098656e-01 -5.64734995e-01
7.48799622e-01 -8.79019320e-01 4.16868776e-02 1.10112809e-01
-2.06730217e-01 6.80161342e-02 3.27933580e-01 5.78889847e-01
-6.10662997e-01 -2.10524395e-01 6.63119912e-01 1.81744054e-01
-4.90551680e-01 2.14012593e-01 -2.07247794e-01 -1.09413318e-01
7.80875385e-01 1.75784543e-01 -3.06624413e-01 -4.00797933e-01
-8.01828861e-01 4.86288488e-01 3.06162506e-01 6.03170514e-01
8.58381271e-01 -1.55965614e+00 -5.20800829e-01 3.59995514e-01
3.86643946e-01 -3.45706433e-01 6.54833734e-01 1.36693919e+00
3.21821123e-01 2.67475128e-01 9.40413773e-02 -6.47556424e-01
-1.14579499e+00 5.09453237e-01 2.54625052e-01 -2.34151393e-01
-6.79614171e-02 5.06930947e-01 4.52938527e-01 -6.53226554e-01
3.09168994e-01 -3.38754170e-02 -4.28897858e-01 6.02070868e-01
3.82776141e-01 2.14380294e-01 -1.24409728e-01 -1.00993228e+00
-5.17950296e-01 6.97139680e-01 -5.33214919e-02 -1.29785806e-01
1.04562449e+00 -3.40546936e-01 -3.65756750e-01 8.89049292e-01
1.52025151e+00 -1.12709641e-01 -1.01026869e+00 -1.54472977e-01
-1.01514392e-01 -2.33704209e-01 -5.27754845e-03 -5.26771426e-01
-1.01895618e+00 1.08314371e+00 6.50756419e-01 2.04590097e-01
1.32801127e+00 5.50472811e-02 5.67338824e-01 2.19822749e-01
6.54437467e-02 -9.81769204e-01 2.14909375e-01 4.98814583e-01
1.08227098e+00 -1.49460578e+00 -2.26796255e-01 -7.74890631e-02
-1.00849056e+00 1.07791221e+00 6.25568271e-01 2.84660250e-01
7.51415670e-01 -5.01527393e-04 3.48207504e-02 -9.44028795e-02
-7.58410692e-01 -2.22455710e-01 7.46467054e-01 1.74992159e-01
4.86786932e-01 6.99541047e-02 -3.90815049e-01 1.12125087e+00
3.86037290e-01 -4.29592669e-01 -1.41210482e-01 7.68498123e-01
-9.18611214e-02 -1.18345046e+00 -3.23606223e-01 5.06671108e-02
-1.39799982e-01 1.25412093e-02 -4.93949890e-01 7.64479280e-01
2.01746836e-01 9.66666281e-01 -1.79275841e-01 -6.56222284e-01
2.79359102e-01 2.87353069e-01 3.26916933e-01 -3.44666801e-02
-3.07621002e-01 4.23010379e-01 -1.94306478e-01 -4.78254527e-01
-5.50458372e-01 -5.79131842e-01 -1.04659605e+00 -1.25958264e-01
-4.63436365e-01 2.31862828e-01 6.22396111e-01 8.45401704e-01
4.57337022e-01 2.66922504e-01 7.86511600e-01 -1.06463599e+00
-4.27512527e-01 -9.58413184e-01 -4.49651510e-01 4.76359636e-01
4.70363677e-01 -8.75437677e-01 -5.70836663e-01 -2.49931738e-01] | [13.094594955444336, 5.0102949142456055] |
57fdd85c-19a6-4792-b940-b9c3231b2347 | robust-human-matting-via-semantic-guidance | 2210.05210 | null | https://arxiv.org/abs/2210.05210v1 | https://arxiv.org/pdf/2210.05210v1.pdf | Robust Human Matting via Semantic Guidance | Automatic human matting is highly desired for many real applications. We investigate recent human matting methods and show that common bad cases happen when semantic human segmentation fails. This indicates that semantic understanding is crucial for robust human matting. From this, we develop a fast yet accurate human matting framework, named Semantic Guided Human Matting (SGHM). It builds on a semantic human segmentation network and introduces a light-weight matting module with only marginal computational cost. Unlike previous works, our framework is data efficient, which requires a small amount of matting ground-truth to learn to estimate high quality object mattes. Our experiments show that trained with merely 200 matting images, our method can generalize well to real-world datasets, and outperform recent methods on multiple benchmarks, while remaining efficient. Considering the unbearable labeling cost of matting data and widely available segmentation data, our method becomes a practical and effective solution for the task of human matting. Source code is available at https://github.com/cxgincsu/SemanticGuidedHumanMatting. | ['Shan Liu', 'Ying Shan', 'Lei Sun', 'Bingtao Fu', 'Yu Li', 'Ye Zhu', 'Xiangguang Chen'] | 2022-10-11 | null | null | null | null | ['image-matting'] | ['computer-vision'] | [ 1.92621574e-01 2.62580723e-01 -1.46631449e-01 -3.63929302e-01
-8.00056398e-01 -3.67396414e-01 3.17678750e-01 -2.50128210e-01
-2.95618594e-01 5.02841055e-01 -1.28626764e-01 -1.16175719e-01
4.65718120e-01 -8.19098771e-01 -1.06251109e+00 -4.09580141e-01
6.73172534e-01 9.41993594e-01 3.41476500e-01 -1.38783470e-01
-4.55264263e-02 -1.40005097e-01 -1.19810939e+00 2.27346838e-01
1.29006767e+00 7.46705055e-01 3.42082441e-01 5.79306185e-01
-1.54671356e-01 6.58598840e-01 -3.87002051e-01 -7.92957962e-01
4.71966863e-01 -5.39716601e-01 -1.17550027e+00 5.85933089e-01
8.59206796e-01 -3.43378544e-01 -2.49870062e-01 1.22950613e+00
1.20050505e-01 7.36324908e-03 3.70635360e-01 -1.45152617e+00
-6.19173467e-01 7.96583891e-01 -6.18776560e-01 -2.31943056e-01
2.51952499e-01 3.95656914e-01 1.09315181e+00 -9.83403683e-01
4.38247383e-01 1.29273868e+00 7.23329425e-01 6.77945852e-01
-9.95554864e-01 -4.18933272e-01 1.89974412e-01 1.73438847e-01
-1.06674337e+00 -3.35116565e-01 5.86157084e-01 -4.00270998e-01
4.12056059e-01 3.02901775e-01 9.39291120e-01 1.02534115e+00
-8.15726072e-02 1.49259675e+00 1.25648165e+00 -2.45445669e-01
1.43021360e-01 -2.62319148e-01 2.18466371e-01 9.72161770e-01
4.65646952e-01 -4.71345931e-01 -3.72086942e-01 2.18373120e-01
8.31335783e-01 2.07789972e-01 -2.24943414e-01 -3.89755219e-01
-1.56571984e+00 6.27906978e-01 7.75413811e-01 1.23953752e-01
-3.72923672e-01 4.97223258e-01 1.86827749e-01 9.55419317e-02
7.37835109e-01 4.17407423e-01 -8.51327404e-02 -2.29523927e-01
-1.43760526e+00 3.56545001e-01 6.26149654e-01 1.16275144e+00
8.68509471e-01 2.72298716e-02 -2.50705257e-02 7.25152314e-01
1.73269242e-01 8.99112225e-01 1.83534101e-01 -1.21706855e+00
5.01500964e-01 7.73477972e-01 2.44758368e-01 -9.30377185e-01
-1.61583811e-01 -2.74703324e-01 -8.20979416e-01 -1.20170061e-02
7.72944391e-01 2.30194092e-01 -1.51280582e+00 1.42603755e+00
4.23329175e-01 2.10544541e-01 -4.46186423e-01 1.17962122e+00
6.57326400e-01 5.07484376e-01 6.82656169e-02 2.87588626e-01
1.27648556e+00 -1.74188530e+00 -7.41069794e-01 -7.64543474e-01
3.77368301e-01 -8.30114365e-01 1.45882189e+00 4.61690187e-01
-1.42875981e+00 -4.69707727e-01 -9.28246975e-01 -3.46747130e-01
-3.03751230e-01 -2.92773098e-02 8.15709591e-01 5.74413478e-01
-1.07102215e+00 6.43189371e-01 -1.10592854e+00 -4.58090872e-01
7.73758829e-01 9.56273824e-02 -1.56316221e-01 -4.75357980e-01
-8.38950753e-01 9.49876189e-01 3.30162913e-01 3.78731996e-01
-1.17321289e+00 -3.96254450e-01 -1.04810882e+00 -3.27777296e-01
6.65074050e-01 -1.05655682e+00 1.42546451e+00 -1.02814198e+00
-1.27455652e+00 1.13325083e+00 -2.14974105e-01 -4.86557692e-01
1.07808137e+00 -7.62612164e-01 1.68617010e-01 2.26517931e-01
3.29565823e-01 8.86770010e-01 9.87417102e-01 -1.42109811e+00
-1.77355230e-01 -1.50193676e-01 -8.54328834e-03 2.60102898e-01
-1.66657433e-01 -2.90410578e-01 -7.57298648e-01 -8.67610455e-01
3.69974792e-01 -9.85867620e-01 -4.23746079e-01 3.87155712e-02
-7.66534150e-01 -1.00674888e-03 5.09387314e-01 -9.96682048e-01
8.95438910e-01 -1.75834274e+00 5.77196777e-01 6.46264181e-02
6.15019381e-01 1.75605774e-01 -8.34191814e-02 1.60913080e-01
3.78083318e-01 9.45964456e-03 -8.74381244e-01 -6.44422948e-01
2.10018963e-01 2.96628088e-01 -1.53135166e-01 4.95502412e-01
-6.37771934e-02 1.62474692e+00 -1.12044871e+00 -5.89778960e-01
4.05782372e-01 1.51521727e-01 -3.41603756e-01 5.06534278e-01
-5.42433858e-01 3.15096229e-01 -3.11947018e-01 9.42180812e-01
6.73352659e-01 -6.58705473e-01 -6.91303760e-02 -1.61139831e-01
5.04585147e-01 1.49601519e-01 -8.80807996e-01 2.15151525e+00
-6.85775876e-02 3.13422084e-01 1.73350066e-01 -9.75577712e-01
5.70120931e-01 4.04169150e-02 3.49646062e-01 -4.51623589e-01
3.33072305e-01 3.50258529e-01 -4.13437784e-01 -2.93260247e-01
5.56767881e-01 -6.53982535e-02 -5.01765721e-02 7.03710079e-01
2.16680299e-02 -6.76643908e-01 2.27397457e-01 6.29156172e-01
1.04812384e+00 1.80195704e-01 -6.49135036e-04 -2.82568932e-01
-2.85868142e-02 2.60802746e-01 4.51983571e-01 7.70372629e-01
-3.61814260e-01 9.33639109e-01 7.66297057e-02 -5.15174091e-01
-1.26670146e+00 -1.23238015e+00 4.82122749e-01 1.02206254e+00
6.93304837e-01 -4.35724884e-01 -1.51505911e+00 -7.97500432e-01
-5.71883582e-02 5.27567089e-01 -6.49590254e-01 6.91490918e-02
-7.19908774e-01 -6.93012655e-01 3.11872631e-01 7.05619574e-01
7.57328689e-01 -1.18084359e+00 -4.14422423e-01 -5.70028275e-02
-6.03486538e-01 -1.27634835e+00 -7.54874229e-01 -3.07661235e-01
-9.18358326e-01 -1.14950454e+00 -1.11926425e+00 -6.91537619e-01
8.93536270e-01 5.13127089e-01 1.52082324e+00 6.37061536e-01
-2.60388702e-01 5.38922668e-01 -4.01447415e-01 -2.16108084e-01
-2.83260971e-01 3.70219857e-01 -2.20047072e-01 -2.11405545e-01
1.74667254e-01 -2.48834640e-01 -7.24085331e-01 4.88901824e-01
-8.63741338e-01 8.06231737e-01 5.81356704e-01 7.28413045e-01
4.78729814e-01 -2.53556699e-01 3.25336576e-01 -1.26827216e+00
2.12900400e-01 -3.39383602e-01 -3.05668831e-01 2.97390997e-01
-6.18251443e-01 -2.27815226e-01 2.71034002e-01 -1.82067961e-01
-8.60403299e-01 1.58552788e-02 -5.10085076e-02 -4.24576610e-01
-1.99290976e-01 8.05974081e-02 -2.03716367e-01 9.74818245e-02
5.37655592e-01 1.71308637e-01 6.90343902e-02 -5.10215223e-01
7.17222333e-01 2.41671249e-01 7.34359801e-01 -7.82348990e-01
1.19293296e+00 6.63174450e-01 -4.24294800e-01 -3.05762291e-01
-1.34975159e+00 -4.74994838e-01 -7.89449215e-01 -2.72920668e-01
1.00081587e+00 -9.66417015e-01 -2.83877850e-01 9.00644720e-01
-9.67335165e-01 -9.91390169e-01 -1.96387604e-01 -1.48056239e-01
-7.16505229e-01 5.47678053e-01 -1.15949118e+00 -5.19167483e-01
-6.04675353e-01 -1.23392582e+00 1.29418921e+00 2.86271330e-02
-2.39878342e-01 -9.08728957e-01 -3.83632541e-01 1.17249179e+00
3.12327772e-01 3.04695100e-01 3.25597644e-01 -8.41073543e-02
-9.29096341e-01 -1.08653054e-01 -4.00156468e-01 3.33911836e-01
1.37259379e-01 -3.50269794e-01 -8.28621686e-01 -3.03618729e-01
-1.97322935e-01 -5.35065830e-01 1.15002096e+00 3.10118973e-01
1.20245862e+00 -2.79943228e-01 -2.25736976e-01 6.27598286e-01
1.05326605e+00 -3.99326056e-01 7.23018587e-01 3.31161261e-01
1.44492340e+00 3.54900271e-01 8.88807535e-01 7.70415738e-02
7.00400054e-01 4.08362687e-01 4.82124180e-01 -5.52123845e-01
-3.86692792e-01 -3.65573168e-01 2.63450205e-01 1.00013244e+00
-1.13658890e-01 -3.77295524e-01 -9.80811775e-01 6.80331469e-01
-2.25655389e+00 -7.16447771e-01 -2.93480307e-01 1.84478951e+00
1.07396007e+00 3.28397721e-01 3.61622602e-01 9.44602415e-02
7.08772838e-01 1.62047237e-01 -6.25787735e-01 4.45576645e-02
-1.65883750e-02 3.58064204e-01 4.27718759e-01 5.85385740e-01
-1.17178845e+00 1.52204418e+00 6.16429329e+00 8.72774541e-01
-6.42710268e-01 4.04619932e-01 8.72145593e-01 8.18197355e-02
-3.82577807e-01 -1.25546142e-01 -3.77445608e-01 5.99233210e-01
4.77795422e-01 8.62761773e-03 6.48422301e-01 8.75703812e-01
6.73009083e-02 -2.13146076e-01 -9.69531655e-01 1.03622723e+00
1.46379396e-01 -1.14930546e+00 1.32129997e-01 -2.00207636e-01
8.09032440e-01 -2.03679413e-01 4.86430228e-02 2.77346652e-02
5.42217433e-01 -1.25528455e+00 1.02642322e+00 3.47205937e-01
5.15552521e-01 -4.75495875e-01 6.77300930e-01 2.48417839e-01
-1.11678672e+00 4.70016301e-01 -3.61142099e-01 -8.24140161e-02
4.78687018e-01 9.84473944e-01 -7.47962475e-01 4.46962118e-01
6.03422821e-01 9.89858866e-01 -8.23188603e-01 9.34892774e-01
-6.37281299e-01 7.57770240e-01 -2.53988951e-01 3.55621189e-01
2.19057247e-01 -1.42566323e-01 2.44780421e-01 1.09783828e+00
2.25247949e-01 2.09799204e-02 5.82523882e-01 9.97442365e-01
-2.53287554e-01 -6.27385080e-02 -2.23713726e-01 -2.20976934e-01
2.81101495e-01 1.35693741e+00 -1.28697407e+00 -7.27756143e-01
-7.87208006e-02 1.62613809e+00 4.01135534e-01 3.40529919e-01
-1.10081995e+00 6.41336739e-02 3.17461342e-01 2.91323453e-01
-7.89756235e-03 -4.16645378e-01 -8.02267373e-01 -1.43048477e+00
1.25860170e-01 -1.06293011e+00 3.58253777e-01 -8.80320668e-01
-1.34341860e+00 4.98336047e-01 -1.75467372e-01 -8.96192193e-01
1.66497037e-01 -5.67595005e-01 -4.87227559e-01 4.30483222e-01
-1.15757763e+00 -1.76021838e+00 -8.15541983e-01 4.61150765e-01
9.16845560e-01 2.97888815e-01 3.45123440e-01 3.32122803e-01
-7.32190013e-01 5.73537350e-01 -6.09898627e-01 2.39243880e-01
6.62658095e-01 -1.64700472e+00 1.03216970e+00 1.20180297e+00
2.62213558e-01 5.35821021e-01 8.09009969e-01 -9.86902356e-01
-1.29395199e+00 -1.26218927e+00 5.95749617e-01 -8.16761971e-01
4.26577955e-01 -6.16385341e-01 -9.57941532e-01 9.97389138e-01
5.25408566e-01 -2.02389926e-01 1.93257824e-01 -6.68677986e-02
-3.41829568e-01 3.22389483e-01 -1.03627062e+00 7.23476410e-01
1.41481209e+00 -1.55038446e-01 -5.85619867e-01 5.92272222e-01
8.74548316e-01 -7.64404595e-01 -5.42008519e-01 3.73517722e-01
2.94849902e-01 -8.17657471e-01 9.84576583e-01 -2.69210011e-01
6.57335997e-01 -5.17295003e-01 9.70366299e-02 -1.25667024e+00
-2.02144906e-01 -5.93388915e-01 -2.97163457e-01 8.90030384e-01
3.34981680e-01 -4.04536724e-01 1.06922984e+00 8.65754426e-01
-3.83080393e-01 -7.58204699e-01 -4.67424631e-01 -8.05876672e-01
-2.44850572e-02 -3.86905313e-01 4.56789374e-01 1.06719458e+00
-2.41220772e-01 3.22382413e-02 -8.28921616e-01 -2.41913032e-02
1.04719770e+00 3.22245747e-01 1.02877271e+00 -7.24582911e-01
-4.17705357e-01 -3.00972879e-01 -2.62305319e-01 -1.23267269e+00
1.26153186e-01 -8.70179892e-01 4.61905926e-01 -1.83842230e+00
5.34407020e-01 -4.20293242e-01 1.46953724e-02 8.80566061e-01
-7.49580562e-01 7.82518923e-01 2.34977305e-01 2.31478661e-01
-1.07919943e+00 5.62261343e-01 1.52036858e+00 -3.77426535e-01
3.18648607e-01 -1.81038350e-01 -5.74660361e-01 8.23972702e-01
8.23532045e-01 -3.41209829e-01 -2.47866526e-01 -7.71099687e-01
8.96841437e-02 -4.39339757e-01 5.86857975e-01 -9.51682806e-01
1.64919719e-01 -4.41679001e-01 2.37728477e-01 -4.71516788e-01
4.40488636e-01 -5.15038371e-01 1.60892442e-01 4.83938783e-01
-8.41921791e-02 1.32744148e-01 -5.14617078e-02 3.34336251e-01
5.32374233e-02 -1.09702475e-01 9.49406981e-01 -5.13919652e-01
-8.07083309e-01 4.89539385e-01 -7.65042612e-03 3.63069504e-01
9.62878287e-01 -1.34103000e-01 -2.73882568e-01 -4.28635895e-01
-6.51058912e-01 3.64475310e-01 1.01390064e+00 3.37849349e-01
6.56411827e-01 -1.22994447e+00 -5.76503038e-01 -2.05189198e-01
-1.02092296e-01 4.98984724e-01 1.25455588e-01 7.84020185e-01
-9.14405763e-01 -1.68444902e-01 -5.77996187e-02 -7.06863403e-01
-9.29935396e-01 4.14515316e-01 2.26273626e-01 -5.37013356e-03
-8.38187218e-01 1.17490029e+00 3.28702182e-01 -2.75972933e-01
5.26955351e-02 -3.60648453e-01 6.32058799e-01 -3.38298857e-01
3.60465586e-01 4.27527219e-01 -8.30050111e-02 -5.21306634e-01
-1.83874801e-01 4.57256645e-01 -1.09606221e-01 -7.89185464e-02
1.21852601e+00 -2.08259076e-01 -2.87095666e-01 4.59950656e-01
6.15170598e-01 -3.62066776e-01 -1.29333031e+00 -2.15064153e-01
5.89363798e-02 -7.14482367e-01 -2.69742101e-01 -6.57853365e-01
-1.42423439e+00 8.69966030e-01 2.14839622e-01 3.62521969e-02
9.77655172e-01 2.91986346e-01 1.60783803e+00 8.76863226e-02
5.86987257e-01 -1.16114473e+00 4.82493401e-01 2.28573352e-01
7.36965656e-01 -1.52664089e+00 4.28321995e-02 -8.16407919e-01
-7.44006753e-01 5.61332524e-01 9.77815390e-01 -2.75824159e-01
1.33268535e-01 1.00539953e-01 3.27358186e-01 -4.73969817e-01
-2.08077893e-01 -1.96405023e-01 4.08915907e-01 2.35211924e-01
3.00034970e-01 4.18701619e-01 -8.13269317e-02 3.35579425e-01
-4.75502849e-01 -1.83951095e-01 3.49925578e-01 8.67486000e-01
-6.20015085e-01 -1.08637345e+00 -2.59607315e-01 4.70223576e-01
-2.19547778e-01 -1.56064972e-01 -5.43993771e-01 4.59859997e-01
-4.41236682e-02 8.13457251e-01 -2.38032758e-01 -2.11535290e-01
5.39558791e-02 4.09657806e-02 6.12807333e-01 -7.46200144e-01
-3.53811055e-01 -1.38014108e-02 -9.72635150e-02 -8.78767729e-01
-4.44826007e-01 -3.65397185e-01 -1.21635687e+00 -6.12903237e-01
-2.43684411e-01 -8.39883536e-02 2.32901841e-01 1.12376630e+00
1.52225584e-01 3.76298577e-01 1.62421450e-01 -9.91309524e-01
-1.01128176e-01 -9.01616693e-01 -4.11158353e-01 9.71668005e-01
7.39532337e-03 -6.21581614e-01 -1.65042043e-01 4.81246322e-01] | [10.692398071289062, -0.9482375383377075] |
7a598d0e-ef8f-4f4d-ad3a-be2abb812cf8 | motion-magnification-algorithms-for-video | 2211.16046 | null | https://arxiv.org/abs/2211.16046v1 | https://arxiv.org/pdf/2211.16046v1.pdf | Motion Magnification Algorithms for Video-Based Breathing Monitoring | In this paper, we present two video processing techniques for contact-less estimation of the Respiratory Rate (RR) of framed subjects. Due to the modest extent of movements related to respiration in both infants and adults, specific algorithms to efficiently detect breathing are needed. For this reason, motion-related variations in video signals are exploited to identify respiration of the monitored patient and simultaneously estimate the RR over time. Our estimation methods rely on two motion magnification algorithms that are exploited to enhance the subtle respiration-related movements. In particular, amplitude- and phase-based algorithms for motion magnification are considered to extract reliable motion signals. The proposed estimation systems perform both spatial decomposition of the video frames combined with proper temporal filtering to extract breathing information. After periodic (or quasi-periodic) respiratory signals are extracted and jointly analysed, we apply the Maximum Likelihood (ML) criterion to estimate the fundamental frequency, corresponding to the RR. The performance of the presented methods is first assessed by comparison with reference data. Videos framing different subjects, i.e., newborns and adults, are tested. Finally, the RR estimation accuracy of both methods is measured in terms of normalized Root Mean Squared Error (RMSE). | ['Riccardo Raheli', 'Francesco Pisani', 'Gianluigi Ferrari', 'Davide Alinovi', 'Veronica Mattioli'] | 2022-11-29 | null | null | null | null | ['motion-magnification'] | ['computer-vision'] | [ 6.98858976e-01 -8.48891214e-02 6.64671510e-02 -1.19845495e-01
-5.14056921e-01 -3.29774201e-01 1.02289945e-01 9.35571417e-02
-5.61738193e-01 6.20127380e-01 1.26325607e-01 2.15686113e-01
-3.32675219e-01 -2.60168284e-01 -5.55558838e-02 -1.05383313e+00
-3.14247102e-01 -4.45684999e-01 3.19380879e-01 3.53234828e-01
2.42093533e-01 6.27401829e-01 -1.66123819e+00 4.53424267e-03
6.80454612e-01 9.49268222e-01 3.12702805e-01 1.15521193e+00
2.55043060e-01 8.35552692e-01 -6.62761033e-01 5.25237359e-02
-9.84993018e-03 -7.69779325e-01 -3.35620731e-01 1.46676645e-01
1.90203011e-01 -5.19441307e-01 -3.68613571e-01 7.79812932e-01
6.88574910e-01 3.46684813e-01 7.15344787e-01 -5.32701612e-01
4.28962082e-01 1.82573974e-01 -6.61630988e-01 8.57569039e-01
6.27274632e-01 2.18855795e-02 3.06209296e-01 -9.09273446e-01
4.44990337e-01 7.13735044e-01 6.96444511e-01 5.31670094e-01
-9.12293553e-01 -3.34443778e-01 -5.53341031e-01 3.68724793e-01
-1.41923630e+00 -7.11444557e-01 1.13392198e+00 -4.87748355e-01
5.70092320e-01 6.12095416e-01 6.38425827e-01 7.33670354e-01
3.38939935e-01 5.31661659e-02 7.88635135e-01 -5.27902424e-01
1.01927981e-01 1.44135296e-01 -2.05661371e-01 6.10541642e-01
3.26354027e-01 -1.87776595e-01 -5.15606344e-01 -3.20026129e-02
6.76323473e-01 -1.58924699e-01 -8.71925473e-01 9.05392170e-02
-1.37045610e+00 3.83533835e-01 -1.01387888e-01 7.61132181e-01
-6.62597418e-01 1.34135425e-01 2.88333744e-01 -6.24806955e-02
2.77719557e-01 6.33984804e-02 -5.43711800e-03 -4.75725949e-01
-1.26621556e+00 -2.48927385e-01 7.24780321e-01 6.79338217e-01
3.65214854e-01 6.88650683e-02 -4.04666573e-01 6.55022740e-01
5.00974953e-01 4.62089628e-01 7.16910660e-01 -1.05580544e+00
5.10717988e-01 1.16468437e-01 2.47502133e-01 -1.16723037e+00
-6.64100349e-01 -1.40176699e-01 -9.31899965e-01 -2.90428400e-01
4.51944798e-01 -2.02834070e-01 -2.01864824e-01 1.35656893e+00
7.03631043e-01 3.95609289e-01 2.91471601e-01 9.06335890e-01
9.39459801e-01 6.37950063e-01 -1.47045597e-01 -1.26810920e+00
1.45310640e+00 -5.01798570e-01 -1.24382973e+00 5.72073422e-02
2.15709955e-01 -9.19152975e-01 3.22773606e-01 3.19072247e-01
-1.48718119e+00 -7.94462264e-01 -1.00462294e+00 4.16250199e-01
4.57096726e-01 3.53856653e-01 -1.83329538e-01 6.49691463e-01
-7.34655380e-01 9.26274419e-01 -1.11730385e+00 -2.09644660e-01
-8.25317428e-02 1.78555906e-01 -1.43344954e-01 2.23173022e-01
-9.85248446e-01 6.31103277e-01 3.01700711e-01 4.64993924e-01
-3.26223373e-01 -5.44485211e-01 -7.07049012e-01 -2.69080758e-01
6.25085160e-02 -4.19879049e-01 1.18879306e+00 -7.43437946e-01
-1.57638001e+00 5.42702854e-01 -3.70089233e-01 -1.96577504e-01
7.91780531e-01 -1.50706291e-01 -4.88077343e-01 1.11785913e+00
-4.73016202e-01 -7.30370134e-02 1.34482443e+00 -6.25959694e-01
-3.31910133e-01 -8.47631693e-02 -5.34434736e-01 3.62633914e-01
-2.12131754e-01 2.89872259e-01 -4.70887572e-01 -8.43512833e-01
6.21227264e-01 -7.00090766e-01 1.04469977e-01 1.13073625e-01
1.38449669e-01 1.94564655e-01 6.60501897e-01 -1.05246592e+00
1.66338909e+00 -2.24686480e+00 6.01834059e-02 9.58429591e-04
3.02427232e-01 2.18553513e-01 2.72787094e-01 3.93516004e-01
-1.28646521e-02 -5.20344496e-01 -4.02100384e-01 -5.68168722e-02
-8.37161064e-01 -3.79734226e-02 2.69964486e-01 1.10861838e+00
1.76181197e-01 3.02792341e-01 -8.72186780e-01 -7.48706698e-01
5.50810099e-01 8.58318627e-01 -2.98825860e-01 6.83189273e-01
3.95308375e-01 1.12528217e+00 -5.28713346e-01 3.92677784e-01
5.86015046e-01 1.28660545e-01 2.56323010e-01 -7.74883687e-01
-3.12016606e-01 2.25527465e-01 -1.25627971e+00 1.29560864e+00
-4.46705878e-01 7.32925534e-01 2.36358166e-01 -8.63503575e-01
7.78804064e-01 9.66143906e-01 1.03731537e+00 -3.34292442e-01
2.67529786e-01 3.27764660e-01 1.35291785e-01 -1.22968066e+00
1.88524544e-01 3.64184454e-02 6.93161607e-01 3.17886263e-01
-2.22901836e-01 -2.02294946e-01 3.11381370e-01 -4.35032964e-01
1.06184590e+00 2.00327545e-01 7.76328921e-01 -2.05245331e-01
9.32309747e-01 -6.06245935e-01 3.39579135e-01 3.09486151e-01
-3.85821521e-01 7.32401431e-01 1.72502846e-02 -2.06650808e-01
-6.69964492e-01 -6.80214524e-01 -2.63976753e-01 3.09949785e-01
5.04831895e-02 -1.68504983e-01 -9.63793278e-01 -2.72525549e-01
-4.29698467e-01 5.40031075e-01 -2.65271604e-01 -2.06143796e-01
-1.22603607e+00 -9.69119966e-01 3.81154507e-01 2.87050873e-01
2.89943993e-01 -1.14124513e+00 -1.61698866e+00 5.83885372e-01
-6.31442964e-01 -1.38184881e+00 -4.15061295e-01 -8.80369172e-02
-1.20031536e+00 -1.28522444e+00 -1.06224346e+00 -3.09836835e-01
5.14222980e-01 3.96823317e-01 7.95873225e-01 3.42419930e-02
-6.51027381e-01 6.17615759e-01 -4.58549410e-01 -7.87608474e-02
-6.76105797e-01 -5.93293309e-01 7.16908425e-02 2.51673311e-01
-6.66757524e-02 -7.18529642e-01 -1.06439221e+00 5.26674867e-01
-8.58703613e-01 -2.59345025e-01 3.99854451e-01 2.77357817e-01
4.16534692e-01 1.01787396e-01 2.19895050e-01 -1.37325078e-01
4.21777666e-01 -2.76130348e-01 -6.12525046e-01 1.46058258e-02
-2.43847340e-01 -1.43582568e-01 4.03835505e-01 -7.69201815e-01
-9.08182681e-01 -5.39702959e-02 -1.93808414e-02 -5.42293251e-01
-1.01591729e-01 -1.62004319e-03 2.99634129e-01 -1.16086222e-01
5.63861847e-01 3.55799586e-01 -5.30682094e-02 -2.62067437e-01
-8.01813602e-02 6.65463328e-01 7.97408760e-01 2.23407801e-02
7.72572219e-01 4.72251266e-01 3.39179546e-01 -1.42952240e+00
-4.37234551e-01 -8.58301163e-01 -7.10101247e-01 -7.54099488e-01
1.34620142e+00 -7.19546914e-01 -6.99720442e-01 6.66061223e-01
-1.30465734e+00 4.68540676e-02 -2.67395139e-01 1.29976308e+00
-6.68105960e-01 8.68051589e-01 -5.51400363e-01 -1.36561012e+00
-6.36794567e-01 -8.84643435e-01 6.68738306e-01 3.58959734e-01
-3.38316202e-01 -9.40386474e-01 1.17639251e-01 3.44835103e-01
3.91198277e-01 5.56991100e-01 3.86977047e-01 -9.81397852e-02
-2.74738550e-01 -1.74491003e-01 1.26949772e-01 2.57634670e-01
5.26641965e-01 2.35349536e-01 -1.06826401e+00 -3.77614915e-01
8.28881502e-01 3.34271163e-01 2.83742547e-01 6.47675872e-01
9.59361613e-01 -4.55031276e-01 -1.16501011e-01 5.83704531e-01
1.24781358e+00 4.48230892e-01 6.13049388e-01 -3.31605494e-01
4.92345810e-01 1.03767574e+00 8.61795425e-01 8.12987149e-01
-1.59130096e-01 7.41230488e-01 4.13669825e-01 2.90981054e-01
-1.89451799e-01 3.07106495e-01 3.95814925e-01 1.31226003e+00
-6.97311401e-01 -3.19765121e-01 -4.32001650e-01 2.81556010e-01
-1.28083539e+00 -9.80210841e-01 -4.48414981e-01 2.52823496e+00
7.59666502e-01 -2.24096045e-01 2.22547390e-02 6.27811432e-01
8.85996521e-01 2.60188073e-01 -3.03981483e-01 -1.57808989e-01
2.33195469e-01 2.29892194e-01 4.58488196e-01 3.05219173e-01
-9.80509400e-01 -1.96593508e-01 5.99450016e+00 5.19106269e-01
-1.09257197e+00 6.30280972e-02 2.59522051e-01 2.05839574e-02
6.09864295e-02 -5.34965336e-01 -5.19675434e-01 5.67499936e-01
1.17868364e+00 1.71754673e-01 3.35711390e-01 2.43123725e-01
7.49777377e-01 -4.20115173e-01 -8.25260162e-01 1.25431371e+00
7.73309469e-02 -4.83529449e-01 -5.89356422e-01 -2.00793654e-01
3.96097302e-01 -4.46173728e-01 -3.48840952e-01 -5.04719913e-01
-1.14343882e+00 -6.91105962e-01 5.94842494e-01 8.32866311e-01
1.01086402e+00 -5.46602726e-01 5.38919508e-01 4.86530215e-01
-1.53460336e+00 -7.04109445e-02 -2.86079347e-01 4.47759300e-01
2.80778438e-01 8.06901395e-01 -6.76306665e-01 6.96090817e-01
4.53641087e-01 5.63542247e-01 -1.20704122e-01 1.09643519e+00
-2.59900689e-01 6.39831424e-01 -4.73145753e-01 1.72799766e-01
-2.16548800e-01 -4.38576192e-01 1.02959526e+00 1.30174077e+00
7.52692163e-01 5.77686667e-01 -5.42883396e-01 6.67066932e-01
3.65827411e-01 2.30975479e-01 -2.38777116e-01 -1.43606544e-01
1.99962482e-01 1.37257993e+00 -8.12728047e-01 -1.62152410e-01
-5.42413294e-01 8.18239391e-01 -4.72860157e-01 2.45110422e-01
-9.18486476e-01 -3.26293945e-01 2.58218318e-01 3.95227462e-01
2.63706207e-01 -2.10317090e-01 4.12409455e-01 -9.56672251e-01
1.74047127e-01 -6.89955771e-01 2.85207212e-01 -6.70101881e-01
-4.48231757e-01 6.01922274e-01 2.23663539e-01 -1.51865375e+00
-6.42078400e-01 -1.25770345e-02 -8.13789964e-01 7.27324903e-01
-1.44115853e+00 -4.83869836e-02 -6.90764606e-01 5.10037720e-01
7.50709176e-01 3.76398712e-01 4.75471079e-01 5.68537474e-01
-4.59009022e-01 4.10256505e-01 -7.95957372e-02 -1.98322356e-01
4.35817003e-01 -8.47062945e-01 -8.40088427e-02 9.28629339e-01
-1.89920887e-01 4.07129735e-01 9.77109492e-01 -5.88499844e-01
-1.40873468e+00 -9.03468668e-01 7.63158500e-01 -2.18196865e-02
2.33632728e-01 1.43223494e-01 -1.02634156e+00 -2.75566634e-02
-1.43598795e-01 3.44092906e-01 3.17569584e-01 -1.26201856e+00
3.99414331e-01 -1.71469614e-01 -1.35940993e+00 7.50621557e-02
5.87759852e-01 -2.88349837e-01 -5.94544768e-01 1.54788541e-02
5.13327599e-01 -4.30362374e-01 -1.20908785e+00 7.36532450e-01
7.22830474e-01 -1.17583203e+00 8.98482263e-01 2.38261774e-01
-1.95179433e-02 -5.20912409e-01 2.13436931e-02 -6.80859089e-01
5.41981235e-02 -1.10889935e+00 -5.92614412e-01 1.07758808e+00
-1.46648839e-01 -5.25761962e-01 5.15399575e-01 1.85822889e-01
2.17347324e-01 -4.61306870e-01 -1.02212071e+00 -4.17099178e-01
-9.08137202e-01 -3.57066065e-01 -1.75122023e-01 6.84016347e-01
-2.14849770e-01 -7.70633072e-02 -4.31427956e-01 2.34877452e-01
7.43951023e-01 -2.01304585e-01 2.31277436e-01 -8.98114145e-01
-5.02276838e-01 4.09378186e-02 -4.42666620e-01 -7.99513698e-01
-5.09064198e-01 -4.85501364e-02 2.73951918e-01 -1.24786520e+00
-1.51864767e-01 1.68198392e-01 -3.17553371e-01 -4.03488159e-01
-3.15899670e-01 3.43172550e-01 -1.15682587e-01 2.98513472e-01
7.84833729e-03 1.48153752e-01 1.18921173e+00 3.21980655e-01
-4.29361671e-01 6.38682187e-01 2.60355651e-01 8.73034537e-01
6.57973289e-01 -4.60523963e-01 -4.91860121e-01 1.40082374e-01
1.44860502e-02 7.48471439e-01 1.75972968e-01 -1.34796429e+00
-7.54892007e-02 2.27263540e-01 3.11236888e-01 -8.09394538e-01
4.49129522e-01 -8.86082113e-01 4.89053071e-01 8.82991552e-01
-2.47230411e-01 1.92822278e-01 -3.26494761e-02 5.08504808e-01
-2.36144617e-01 -6.30626023e-01 1.18587828e+00 -4.44273204e-02
-9.61301550e-02 5.80983311e-02 -7.23657131e-01 -3.00885662e-02
8.32229316e-01 -5.02941549e-01 1.19000398e-01 -4.65936363e-01
-6.39172733e-01 -4.65372622e-01 2.67672651e-02 -1.27632827e-01
8.08384061e-01 -9.62359428e-01 -6.75715148e-01 1.96633577e-01
-1.00584827e-01 -1.33476287e-01 5.17622530e-01 1.81506836e+00
-8.95147026e-01 2.07417607e-01 1.92873895e-01 -7.27375984e-01
-1.67801523e+00 5.95094442e-01 4.54131693e-01 -4.23073098e-02
-7.09365129e-01 7.04892278e-01 -6.33045798e-03 6.24668062e-01
2.45250821e-01 -6.82537019e-01 -8.45533371e-01 2.81767637e-01
9.07636106e-01 9.59317803e-01 -8.95824097e-03 -8.03365827e-01
-4.46136534e-01 1.16945779e+00 5.95078588e-01 -4.79016379e-02
8.66282463e-01 -6.81671143e-01 -6.04064986e-02 4.98182505e-01
1.28896475e+00 3.94069612e-01 -1.25336981e+00 1.03427693e-01
-4.33786586e-02 -3.94293576e-01 -2.11226940e-01 -1.05906062e-01
-1.06808376e+00 1.01407981e+00 8.56794298e-01 3.04404914e-01
1.66126311e+00 -4.52542841e-01 5.80484211e-01 -4.15348820e-02
-3.81953157e-02 -8.01292598e-01 1.40502796e-01 -3.49546611e-01
7.11393178e-01 -8.30228865e-01 2.85841554e-01 -7.23080933e-01
-1.92217991e-01 1.52087557e+00 1.04218252e-01 1.68944150e-01
3.67052048e-01 1.15884997e-01 1.43471211e-01 2.64537483e-01
-5.57618439e-01 4.25844733e-03 5.63122094e-01 4.09049571e-01
6.72326684e-01 -4.78021413e-01 -7.26819515e-01 -1.33107677e-01
2.30699331e-01 -6.23634383e-02 6.94275081e-01 7.83529937e-01
-6.26650870e-01 -4.16692287e-01 -8.04101706e-01 1.12157300e-01
-1.10246718e+00 1.44606531e-01 2.40518913e-01 5.89717448e-01
-6.48103133e-02 1.33182132e+00 -7.97098354e-02 1.36314616e-01
3.55232328e-01 -1.03615016e-01 9.37050581e-01 -8.81815553e-02
-4.72441971e-01 5.76446295e-01 5.13728447e-02 -7.46690989e-01
-1.04793811e+00 -7.01411486e-01 -1.18041170e+00 3.43503028e-01
-3.01615715e-01 2.14183390e-01 8.72264743e-01 6.37989879e-01
-3.13583314e-01 7.94830680e-01 8.83237123e-01 -8.02227437e-01
-3.85797739e-01 -8.92788172e-01 -5.78456819e-01 3.33859116e-01
6.74413443e-01 -2.44713306e-01 -8.93823385e-01 2.00201750e-01] | [13.936112403869629, 2.9482533931732178] |
c484c209-de08-4a6f-95c5-09c25c0e181b | learning-monocular-visual-odometry-through | 1903.10543 | null | https://arxiv.org/abs/1903.10543v2 | https://arxiv.org/pdf/1903.10543v2.pdf | Learning Monocular Visual Odometry through Geometry-Aware Curriculum Learning | Inspired by the cognitive process of humans and animals, Curriculum Learning (CL) trains a model by gradually increasing the difficulty of the training data. In this paper, we study whether CL can be applied to complex geometry problems like estimating monocular Visual Odometry (VO). Unlike existing CL approaches, we present a novel CL strategy for learning the geometry of monocular VO by gradually making the learning objective more difficult during training. To this end, we propose a novel geometry-aware objective function by jointly optimizing relative and composite transformations over small windows via bounded pose regression loss. A cascade optical flow network followed by recurrent network with a differentiable windowed composition layer, termed CL-VO, is devised to learn the proposed objective. Evaluation on three real-world datasets shows superior performance of CL-VO over state-of-the-art feature-based and learning-based VO. | ['Andrew Markham', 'Sen Wang', 'Muhamad Risqi U. Saputra', 'Niki Trigoni', 'Pedro P. B. de Gusmao'] | 2019-03-25 | null | null | null | null | ['monocular-visual-odometry'] | ['robots'] | [-1.08304285e-01 -4.84516658e-02 3.06111928e-02 -3.85400474e-01
-1.48221910e-01 -2.95525938e-01 5.64909577e-01 -3.51802468e-01
-6.15332484e-01 6.73438251e-01 8.61974657e-02 -9.44797024e-02
-2.03405749e-02 -6.25073433e-01 -9.76908624e-01 -3.90325457e-01
9.22560692e-02 7.42542744e-02 6.99278489e-02 -3.46429385e-02
3.47315431e-01 4.69544888e-01 -1.49175954e+00 -3.65212083e-01
1.04886436e+00 6.62535310e-01 5.91521442e-01 6.06683671e-01
1.70241401e-01 1.12613988e+00 -4.63852584e-01 3.65031622e-02
4.66390103e-01 -4.77320671e-01 -7.16858923e-01 1.87461093e-01
9.10197675e-01 -1.03847675e-01 -6.17068529e-01 8.67757380e-01
5.23233831e-01 4.60642546e-01 5.28155684e-01 -1.02329683e+00
-7.97916591e-01 6.37894198e-02 -2.37345308e-01 2.35274538e-01
3.63136023e-01 2.71361321e-01 1.11063480e+00 -1.16341746e+00
9.14599061e-01 1.25962961e+00 5.54920495e-01 5.32151103e-01
-1.23345327e+00 -4.59001452e-01 4.25607085e-01 4.32797402e-01
-1.43985403e+00 -1.68194726e-01 9.63598967e-01 -5.88763654e-01
1.03986216e+00 -2.09646463e-01 7.04199553e-01 7.19619989e-01
3.83491069e-01 8.53619993e-01 8.78200173e-01 -2.92752773e-01
-6.00798130e-02 -4.28040251e-02 -3.76616776e-01 1.17622852e+00
2.11113930e-01 3.42788368e-01 -5.65803170e-01 3.37758034e-01
1.07111156e+00 -1.49029493e-01 -4.21444416e-01 -1.08169520e+00
-1.21923399e+00 8.39145064e-01 8.79093945e-01 -1.04282029e-01
-2.74510771e-01 4.01074857e-01 1.13176972e-01 4.59485263e-01
5.53795934e-01 7.96661913e-01 -3.21704179e-01 1.93661109e-01
-7.90107489e-01 2.55968064e-01 4.85269099e-01 1.03789210e+00
9.08094168e-01 4.35189933e-01 -4.80927713e-02 5.97255707e-01
4.01355654e-01 2.70054907e-01 4.24405098e-01 -9.26720083e-01
3.95170808e-01 6.05005324e-01 -2.69188154e-02 -8.01807344e-01
-6.12986147e-01 -7.45301306e-01 -5.78189135e-01 6.01523221e-01
3.92420173e-01 -2.51718789e-01 -9.93314564e-01 1.87409973e+00
4.09197032e-01 5.94784021e-01 5.75407855e-02 1.17688739e+00
7.16826022e-01 5.24447680e-01 -6.41075522e-02 2.11075261e-01
8.31129968e-01 -1.23300755e+00 -3.54447365e-01 -2.74874061e-01
5.19457757e-01 -6.99349821e-01 1.06302965e+00 1.70612276e-01
-1.16240060e+00 -9.12305892e-01 -1.24269962e+00 -4.41433042e-01
-7.38968849e-02 3.28970373e-01 5.45598030e-01 3.36793691e-01
-1.09950817e+00 6.10616624e-01 -8.50386024e-01 -1.82787850e-01
3.39673012e-01 3.48873734e-01 -1.29317135e-01 -2.91496459e-02
-9.82725501e-01 1.00188446e+00 1.66619346e-01 1.53228983e-01
-1.05423880e+00 -1.03548074e+00 -1.37122381e+00 -9.87449363e-02
3.74635011e-01 -1.25172532e+00 1.03025329e+00 -7.82862842e-01
-2.06998062e+00 7.82956183e-01 -2.27066413e-01 -5.35700321e-01
7.96955645e-01 -6.87069595e-01 1.79842174e-01 1.02077894e-01
-1.12198263e-01 8.60580981e-01 1.23199058e+00 -1.13673615e+00
-4.15619344e-01 -1.31013915e-01 2.28860289e-01 5.34048855e-01
1.22093827e-01 -4.58433658e-01 -6.31069466e-02 -4.83386397e-01
1.48239031e-01 -9.33175623e-01 -4.25842553e-01 3.49731833e-01
-2.19155356e-01 -2.51493156e-01 5.30798197e-01 -5.15062153e-01
9.19956744e-01 -1.81729817e+00 6.51427984e-01 1.32699693e-02
3.70367259e-01 3.18350434e-01 -2.20515072e-01 1.85001865e-02
2.31287658e-01 -4.55104053e-01 -5.87412827e-02 -5.98455966e-01
-9.28424299e-02 2.04954237e-01 -2.92048901e-01 7.15549707e-01
5.86242914e-01 1.06056178e+00 -1.12196875e+00 -2.92984307e-01
5.35409391e-01 6.88710570e-01 -9.43715692e-01 6.22968197e-01
-2.66246676e-01 8.35013151e-01 -1.88062206e-01 4.25364286e-01
4.92210746e-01 -2.33170033e-01 -7.58298710e-02 -8.53049979e-02
-3.59724224e-01 3.33039314e-01 -1.10338604e+00 2.02733707e+00
-7.67072678e-01 7.76684940e-01 -2.50967056e-01 -9.29002225e-01
1.17711258e+00 -7.88144320e-02 2.02045977e-01 -6.31625533e-01
8.97385404e-02 4.29090299e-03 -1.64396867e-01 -3.50777626e-01
6.58430874e-01 -2.78031603e-02 2.06908047e-01 9.63411406e-02
4.32616740e-01 -4.70776558e-01 -1.71967462e-01 -1.31463945e-01
6.35654211e-01 8.41113508e-01 3.83996189e-01 -2.58696914e-01
1.01085162e+00 -2.48289391e-01 6.27661526e-01 5.19578218e-01
-2.96069264e-01 5.78222334e-01 9.06652063e-02 -5.61090708e-01
-1.16647339e+00 -1.23592806e+00 7.10174441e-03 9.21035230e-01
4.18609530e-01 -1.06689982e-01 -4.92410839e-01 -3.56611103e-01
1.51254520e-01 2.60740876e-01 -5.76068282e-01 -2.01734200e-01
-8.99909258e-01 -2.38200516e-01 2.86632895e-01 5.52507162e-01
7.94377327e-01 -9.19334948e-01 -9.25838113e-01 3.82058918e-01
2.37451196e-02 -1.33079886e+00 -8.17527711e-01 -1.99511707e-01
-7.91400373e-01 -1.04498255e+00 -7.55306959e-01 -1.05734241e+00
6.98273838e-01 3.64718884e-01 1.08360910e+00 -4.10142168e-02
-4.66337323e-01 3.10957104e-01 1.78279147e-01 -2.74543434e-01
1.13292173e-01 4.14708167e-01 2.90870238e-02 -2.56573036e-02
6.60831407e-02 -6.94594324e-01 -7.56159782e-01 2.00983465e-01
-4.56564128e-01 1.28939778e-01 5.88377535e-01 9.70259428e-01
5.11119545e-01 -6.97037458e-01 3.87307525e-01 -6.05308175e-01
4.06479925e-01 -2.66260743e-01 -1.01912045e+00 1.12232000e-01
-5.35145342e-01 5.10257781e-01 7.18061745e-01 -5.45777321e-01
-1.09703720e+00 2.82186627e-01 1.59635127e-01 -7.95090437e-01
1.57655701e-01 3.94093782e-01 1.29097536e-01 -5.78107774e-01
5.89407325e-01 2.89525896e-01 -1.07143084e-02 -2.97176301e-01
5.31857550e-01 6.90002888e-02 8.26547444e-01 -4.94246185e-01
1.17140901e+00 5.96005797e-01 3.33623201e-01 -8.14626873e-01
-9.60360944e-01 -5.93131602e-01 -9.62319970e-01 -2.49037564e-01
8.61017168e-01 -1.17986751e+00 -9.89366770e-01 5.26405871e-01
-1.10774529e+00 -5.36603510e-01 -2.45832190e-01 8.50575209e-01
-8.32161665e-01 2.77366996e-01 -4.35957223e-01 -6.22342706e-01
-3.19536179e-01 -9.33505118e-01 8.28204751e-01 3.43796521e-01
4.04486433e-02 -1.20502484e+00 4.55814868e-01 5.83076328e-02
2.72552013e-01 3.90193582e-01 4.52537924e-01 1.24064967e-01
-9.65007067e-01 3.21204394e-01 -2.63202220e-01 3.21981907e-01
9.68588144e-02 -1.68386370e-01 -8.79839122e-01 -5.02671421e-01
-1.86931193e-01 -5.00473559e-01 9.89304185e-01 4.41854119e-01
9.69039321e-01 1.08399481e-01 -1.72081534e-02 1.24235427e+00
1.50232172e+00 3.95973660e-02 4.56082851e-01 4.02634442e-01
1.17792654e+00 5.09748638e-01 4.48120505e-01 3.55286509e-01
6.87503815e-01 6.20248199e-01 3.68136406e-01 -4.92771417e-02
-1.73671678e-01 -5.28359354e-01 4.69028682e-01 9.08142090e-01
-1.71440542e-01 3.40354651e-01 -6.76368117e-01 5.68649292e-01
-1.75473666e+00 -8.25105846e-01 3.53989184e-01 2.04801250e+00
7.84263074e-01 2.01851845e-01 4.13724445e-02 -1.43323272e-01
4.39136952e-01 4.43433583e-01 -7.73257256e-01 -4.06071931e-01
9.39064939e-03 3.16193074e-01 5.70298433e-01 8.92648697e-01
-1.21264088e+00 1.22194088e+00 6.23251677e+00 6.42857701e-02
-1.54805648e+00 -1.26475647e-01 3.51151749e-02 -2.23418176e-02
-1.14701442e-01 -1.02935441e-01 -8.45286548e-01 1.49428956e-02
5.01975775e-01 -2.39411101e-01 4.92336363e-01 8.65130365e-01
1.53472364e-01 1.89332306e-01 -1.15103102e+00 1.02741778e+00
2.49500319e-01 -1.36433589e+00 -5.73782576e-03 -1.87269211e-01
9.35033619e-01 2.29373798e-01 2.16143027e-01 3.25374931e-01
4.50755477e-01 -1.16429412e+00 8.78397048e-01 7.31530130e-01
7.10885704e-01 -5.96764147e-01 3.69572252e-01 2.06948534e-01
-1.52130401e+00 -4.71487418e-02 -6.02074623e-01 -3.50456715e-01
1.01357631e-01 1.52605578e-01 -7.58368909e-01 6.98029995e-01
3.21426511e-01 1.15556574e+00 -5.81057608e-01 1.44499385e+00
-5.43205321e-01 2.39343211e-01 -4.97984253e-02 -7.92801306e-02
5.27531981e-01 -2.19935462e-01 6.41016245e-01 9.00539815e-01
2.21279170e-02 -1.78161532e-01 3.40933472e-01 1.16857672e+00
-1.77012458e-01 2.22658329e-02 -6.71085417e-01 4.78400052e-01
2.88125068e-01 1.01424122e+00 -5.79312667e-02 -9.87264216e-02
-5.55460751e-01 8.17274988e-01 8.21969092e-01 4.05908585e-01
-5.77621043e-01 -4.91133392e-01 9.10404146e-01 -1.88874900e-01
5.51969767e-01 -6.22790396e-01 -1.81768715e-01 -1.38935959e+00
-1.91000253e-02 -3.09908628e-01 8.88620839e-02 -7.14824617e-01
-9.93602514e-01 5.66350579e-01 -1.87436193e-01 -1.38211870e+00
-3.26081961e-01 -6.50011241e-01 -7.39966035e-01 9.21866000e-01
-2.16548181e+00 -1.07105970e+00 -7.05501676e-01 7.58695424e-01
4.85762566e-01 -1.17525190e-01 4.07075673e-01 1.84535041e-01
-2.74451137e-01 5.40003419e-01 -1.85059562e-01 1.58335298e-01
7.45973170e-01 -1.44697809e+00 7.01802731e-01 8.52969885e-01
2.04722166e-01 4.65681851e-01 6.84863150e-01 -3.78787488e-01
-1.21065712e+00 -1.27078640e+00 8.06357622e-01 -4.52439219e-01
5.12203276e-01 -3.27170193e-01 -9.40424144e-01 8.82903874e-01
3.30206603e-02 2.57809132e-01 7.57187307e-02 -2.29736149e-01
-4.93726850e-01 -3.81659716e-02 -9.97025490e-01 6.84544384e-01
1.39227188e+00 -6.68134868e-01 -7.58093059e-01 4.27220725e-02
8.57802808e-01 -8.12073171e-01 -7.45911360e-01 3.35931659e-01
5.55113077e-01 -5.87661684e-01 1.21646214e+00 -7.08876073e-01
3.07783455e-01 -3.70246530e-01 1.64607704e-01 -1.60214376e+00
-3.47871423e-01 -9.48007941e-01 -1.67149290e-01 5.49840689e-01
5.02740778e-02 -6.59209251e-01 1.01369619e+00 8.97498503e-02
-3.91823679e-01 -7.00369895e-01 -8.42680991e-01 -8.57853115e-01
2.75583833e-01 -1.09502576e-01 3.49016517e-01 8.18994522e-01
-2.05517545e-01 3.02603692e-01 -5.03924251e-01 2.47183278e-01
9.24840927e-01 7.82827362e-02 1.08627439e+00 -1.29553139e+00
-2.81525224e-01 -2.43894562e-01 -7.42807388e-01 -1.54987156e+00
4.80052173e-01 -9.25077617e-01 3.90373431e-02 -1.23020303e+00
-2.73740649e-01 -1.49232090e-01 -7.85125718e-02 2.24119313e-02
-3.71100515e-01 4.61140424e-02 5.21819413e-01 1.33170411e-01
-4.62911904e-01 9.49812472e-01 1.85357511e+00 -2.31623277e-02
-5.39900005e-01 1.55470828e-02 -1.99433416e-01 7.10566223e-01
6.62976325e-01 -1.85798362e-01 -7.04363883e-01 -4.99453634e-01
1.13980643e-01 -1.90857232e-01 5.47506154e-01 -1.14789248e+00
4.88815099e-01 -7.57079870e-02 4.05459672e-01 -5.54512382e-01
3.10240299e-01 -3.87598127e-01 -3.11116189e-01 7.15660572e-01
-4.02159333e-01 7.34934881e-02 6.75569335e-03 4.98914629e-01
-2.07545340e-01 1.05197020e-01 8.80410016e-01 -7.36671463e-02
-1.00401556e+00 4.70864743e-01 -1.90465376e-02 4.26322252e-01
8.26897204e-01 -1.05802894e-01 -2.68595904e-01 -3.55749577e-01
-8.06654036e-01 5.13059855e-01 2.29948223e-01 4.36515927e-01
9.62860286e-01 -1.35394812e+00 -6.53672278e-01 3.91516685e-01
1.51584908e-01 2.82313973e-01 -7.76710510e-02 6.66558325e-01
-8.69426727e-01 6.46428287e-01 -4.58099812e-01 -7.09938228e-01
-1.19389522e+00 4.53101665e-01 6.84303284e-01 -2.69044399e-01
-7.96237350e-01 9.80568647e-01 3.31764936e-01 -8.33919764e-01
3.08775723e-01 -6.93595350e-01 -2.47652486e-01 -3.83715898e-01
2.61120766e-01 4.18002725e-01 -3.06896299e-01 -6.09929740e-01
-2.22246259e-01 8.52957189e-01 1.52861536e-01 -4.03436907e-02
1.38624728e+00 -3.27199936e-01 3.15369636e-01 4.30769414e-01
1.29737163e+00 -3.85302454e-01 -2.06689715e+00 -7.43783712e-01
1.33418888e-01 -5.81894517e-01 -4.79336269e-02 -2.16905370e-01
-9.97502863e-01 1.00924861e+00 5.04948258e-01 -5.57716370e-01
8.12599897e-01 -2.69817382e-01 6.72866583e-01 5.99538505e-01
3.51110548e-01 -8.80245924e-01 4.53399539e-01 7.17256129e-01
8.56443405e-01 -1.29472792e+00 -8.82570073e-03 -4.55288887e-01
-5.66259503e-01 1.09543836e+00 8.47259760e-01 -7.83729374e-01
6.95365846e-01 -2.77287394e-01 1.24147937e-01 -1.26674592e-01
-8.02824557e-01 -3.87828052e-01 6.88120663e-01 5.79960763e-01
3.03690404e-01 -2.99209982e-01 -6.09045997e-02 -1.31893724e-01
-4.75219756e-01 1.17418416e-01 5.65379381e-01 9.36658263e-01
-4.92645353e-01 -8.84878337e-01 -1.89308077e-01 -3.04355919e-02
-1.51451796e-01 -7.01314509e-02 -6.63252547e-02 7.83098936e-01
2.32025273e-02 4.16993052e-01 1.62214726e-01 -2.48677507e-01
4.87333477e-01 -9.82042030e-02 8.65201414e-01 -6.14351571e-01
-5.41137993e-01 -3.20503801e-01 -1.58244580e-01 -7.38364756e-01
-6.59375370e-01 -6.91567779e-01 -1.15865672e+00 -2.34910712e-01
-4.11650501e-02 -2.03428224e-01 4.15656090e-01 1.00615275e+00
1.76876783e-01 6.06516302e-01 7.97636211e-01 -1.10314786e+00
-7.03833818e-01 -6.84553564e-01 -1.76366955e-01 3.59728694e-01
8.37401152e-01 -1.02419603e+00 -2.85721391e-01 1.09690167e-01] | [8.350001335144043, -2.138300657272339] |
15121d37-56ef-4b19-a6ed-ee0a0790412c | kradagrad-kronecker-approximation-domination | 2305.19416 | null | https://arxiv.org/abs/2305.19416v1 | https://arxiv.org/pdf/2305.19416v1.pdf | KrADagrad: Kronecker Approximation-Domination Gradient Preconditioned Stochastic Optimization | Second order stochastic optimizers allow parameter update step size and direction to adapt to loss curvature, but have traditionally required too much memory and compute for deep learning. Recently, Shampoo [Gupta et al., 2018] introduced a Kronecker factored preconditioner to reduce these requirements: it is used for large deep models [Anil et al., 2020] and in production [Anil et al., 2022]. However, it takes inverse matrix roots of ill-conditioned matrices. This requires 64-bit precision, imposing strong hardware constraints. In this paper, we propose a novel factorization, Kronecker Approximation-Domination (KrAD). Using KrAD, we update a matrix that directly approximates the inverse empirical Fisher matrix (like full matrix AdaGrad), avoiding inversion and hence 64-bit precision. We then propose KrADagrad$^\star$, with similar computational costs to Shampoo and the same regret. Synthetic ill-conditioned experiments show improved performance over Shampoo for 32-bit precision, while for several real datasets we have comparable or better generalization. | ['Luke Walters', 'Alexander Moreno', 'Jonathan Mei'] | 2023-05-30 | null | null | null | null | ['stochastic-optimization'] | ['methodology'] | [-3.63510877e-01 -1.20228104e-01 -6.53456226e-02 -2.54583687e-01
-8.11680257e-01 -5.28272688e-01 2.87126750e-01 1.18177041e-01
-9.26169753e-01 9.22450244e-01 -9.21934471e-02 -5.35580575e-01
-2.36170739e-01 -6.23169720e-01 -1.08893085e+00 -6.69481397e-01
-2.98443466e-01 3.84884924e-01 -1.84915274e-01 -1.55643923e-02
1.19844384e-01 3.82416904e-01 -1.19618428e+00 1.59401968e-01
7.93564498e-01 1.34696984e+00 1.39615089e-01 7.14803576e-01
2.60269195e-01 7.65495956e-01 -2.47132614e-01 -8.82069647e-01
8.35981786e-01 5.58781624e-02 -6.66945755e-01 -3.83945912e-01
8.82827580e-01 -4.90299970e-01 -6.06901586e-01 1.27568519e+00
6.33271754e-01 2.82124311e-01 5.39291382e-01 -1.07435715e+00
-4.93529111e-01 8.75618637e-01 -5.44468641e-01 9.00881179e-03
-1.87914923e-01 1.15454748e-01 1.22146928e+00 -1.05001342e+00
2.52324551e-01 1.23399341e+00 1.15936828e+00 3.63112122e-01
-1.48077309e+00 -7.83066213e-01 -6.81938138e-03 2.83926755e-01
-1.50744796e+00 -4.59121257e-01 2.81186640e-01 -2.32595205e-01
1.13505650e+00 2.57601857e-01 8.14107239e-01 8.63394737e-01
3.27296436e-01 1.06186831e+00 9.55321014e-01 -2.13785097e-01
5.33062339e-01 -9.23919752e-02 1.26810148e-01 9.57073808e-01
4.72538680e-01 1.14876822e-01 -8.95723820e-01 -4.25525337e-01
5.11198819e-01 -7.07493722e-03 -3.56058806e-01 -4.84598368e-01
-8.97684932e-01 9.68129933e-01 5.28729737e-01 -1.58605427e-01
-4.08275068e-01 6.89258635e-01 5.17570794e-01 3.54817539e-01
3.83406222e-01 3.86785179e-01 -5.91293871e-01 -5.35884738e-01
-1.23815978e+00 5.74068606e-01 9.28863943e-01 6.19175255e-01
9.54622805e-01 1.78931370e-01 3.92307453e-02 8.09300005e-01
4.25318092e-01 5.66732764e-01 6.96775436e-01 -1.14775300e+00
4.86734450e-01 9.13450047e-02 -2.46928837e-02 -1.04280317e+00
-5.90455115e-01 -8.68326783e-01 -1.36929929e+00 2.13520676e-01
5.17529964e-01 -2.13574409e-01 -8.48235428e-01 1.72464049e+00
2.92000443e-01 4.72179838e-02 6.10143878e-02 9.52382028e-01
1.64027587e-01 5.86037934e-01 -3.08836460e-01 1.17679305e-01
1.10646939e+00 -1.10565352e+00 -4.36534613e-01 -5.73908389e-01
1.04963624e+00 -6.58870578e-01 1.08185422e+00 8.57884765e-01
-1.02427018e+00 -8.25474113e-02 -1.30252135e+00 -2.72429287e-01
2.62378883e-02 4.50507849e-01 1.31418061e+00 9.69632268e-01
-1.27276695e+00 1.14633167e+00 -1.17158425e+00 3.69489819e-01
5.80261528e-01 5.37068307e-01 -2.63858885e-01 2.20112801e-02
-1.28169191e+00 7.24668384e-01 4.32158887e-01 2.05570549e-01
-7.50941396e-01 -1.16698647e+00 -7.01003075e-01 1.81484684e-01
1.18547812e-01 -7.92617798e-01 1.21959591e+00 -7.28650630e-01
-1.78091156e+00 6.68878257e-01 5.90456389e-02 -1.25260925e+00
7.14880109e-01 -7.36514926e-01 1.41211420e-01 -1.91841945e-01
-3.23743403e-01 5.77177644e-01 9.43781257e-01 -5.50482452e-01
-3.71223509e-01 -6.08619511e-01 1.03879869e-01 3.35672498e-03
-5.58674693e-01 -6.18230939e-01 -3.87911320e-01 -7.94765174e-01
2.65000880e-01 -1.24424672e+00 -4.13520008e-01 1.31729588e-01
-7.80621171e-02 9.85764414e-02 1.68014482e-01 -6.75996959e-01
1.39566898e+00 -2.23728824e+00 7.78832063e-02 3.03785384e-01
4.32280242e-01 4.40800607e-01 -4.10913937e-02 1.72056884e-01
4.82248142e-02 -8.17089006e-02 -4.39958513e-01 -5.92904925e-01
3.16176683e-01 3.62226546e-01 -4.06671107e-01 7.68679500e-01
-2.96627820e-01 6.81740344e-01 -7.58406520e-01 1.88289452e-02
-1.73740655e-01 5.80271840e-01 -1.29802859e+00 -3.51843059e-01
1.63704064e-02 -2.10562244e-01 4.10187617e-02 2.75273621e-01
1.03160393e+00 -2.83834696e-01 2.58847237e-01 -4.06119972e-01
2.44873300e-01 5.33867240e-01 -1.60714149e+00 1.73923028e+00
-9.38629806e-01 6.72735155e-01 4.75822657e-01 -9.38824594e-01
6.76736414e-01 -1.61362737e-01 2.92817742e-01 -3.90188307e-01
2.56381668e-02 6.31561697e-01 -1.28572360e-01 2.79369414e-01
6.92769945e-01 -3.91696356e-02 3.12526554e-01 2.83627868e-01
1.35345548e-01 -2.07025930e-01 3.99409235e-01 4.63447779e-01
1.04224885e+00 -7.84614533e-02 -2.16047335e-02 -5.42663276e-01
4.15703446e-01 -3.37453604e-01 6.51110709e-01 9.18544233e-01
-4.44448516e-02 4.47471768e-01 5.12250006e-01 -6.48816645e-01
-8.81983817e-01 -9.64671791e-01 -2.83285260e-01 1.05801547e+00
-3.04424495e-01 -8.95798266e-01 -7.41067767e-01 -5.19335330e-01
3.83438587e-01 5.35826027e-01 -4.18201417e-01 -2.58583844e-01
-5.52004278e-01 -1.06987751e+00 7.26915896e-01 3.12410295e-01
6.31797314e-01 -2.97954500e-01 -5.17001033e-01 1.62086204e-01
1.82408974e-01 -6.68887138e-01 -6.57467365e-01 5.54532588e-01
-1.05456269e+00 -6.41662717e-01 -8.25533569e-01 -2.78971106e-01
4.96348530e-01 -8.78813267e-02 1.13403237e+00 -2.08120510e-01
-2.07474872e-01 -7.00637251e-02 1.38265463e-02 -7.36186653e-02
-5.15953712e-02 2.01840609e-01 4.86485898e-01 2.84542013e-02
-7.90139064e-02 -7.47773647e-01 -8.01888466e-01 -1.17787734e-01
-9.22549605e-01 -3.97043042e-02 6.85026050e-01 1.23474228e+00
7.29788303e-01 -2.25178316e-01 -4.88012433e-02 -8.21557164e-01
6.85108483e-01 -1.21556893e-01 -9.93280709e-01 -3.53811309e-02
-9.49131191e-01 7.68938184e-01 9.33300555e-01 -3.97672921e-01
-5.88271677e-01 1.61113873e-01 -4.07294214e-01 -6.96494639e-01
8.58892024e-01 6.29829109e-01 9.02219489e-02 -4.25115019e-01
8.95858347e-01 2.50644922e-01 -1.17066085e-01 -7.19447911e-01
3.87911856e-01 4.38342124e-01 3.34641725e-01 -7.58509338e-01
4.84378964e-01 3.80043179e-01 2.71983773e-01 -4.79013175e-01
-1.09878671e+00 2.57964432e-02 -2.27484703e-01 3.15688223e-01
1.81580082e-01 -1.15788674e+00 -9.79424417e-01 5.76616764e-01
-9.80047107e-01 -5.46828449e-01 -2.96128184e-01 8.48621666e-01
-4.81431752e-01 4.74005014e-01 -8.25193226e-01 -6.11773670e-01
-5.91346085e-01 -1.14650023e+00 7.98472285e-01 -9.38595012e-02
9.05769877e-03 -8.46484303e-01 9.59935039e-02 2.93360710e-01
3.61912191e-01 -2.54081041e-01 4.94708002e-01 -2.12132722e-01
-3.74423712e-01 -5.44056483e-02 -2.15681076e-01 8.00638497e-01
-3.73762608e-01 -3.14250380e-01 -8.00733864e-01 -6.61006629e-01
1.33462563e-01 -4.70154226e-01 1.17218971e+00 1.76080376e-01
1.40590215e+00 -7.10103989e-01 1.20447092e-01 1.35843873e+00
1.18393409e+00 -2.56887108e-01 2.64412165e-01 4.39488024e-01
6.23971939e-01 -1.22530617e-01 3.23320478e-01 7.31988549e-01
1.61383167e-01 4.96856898e-01 2.91235477e-01 3.70744705e-01
6.94160257e-03 -2.96887666e-01 6.39645517e-01 9.93169785e-01
7.59024099e-02 1.81286130e-02 -9.82593954e-01 2.43136764e-01
-1.83064401e+00 -5.72253644e-01 -2.60015160e-01 2.36556888e+00
1.25426519e+00 2.45560437e-01 -3.06343138e-01 1.02286167e-01
1.31897077e-01 1.82474852e-01 -7.37787604e-01 -6.23118281e-01
-1.59270555e-01 4.98636633e-01 1.33101594e+00 9.26282883e-01
-9.25875545e-01 7.80388892e-01 5.57083750e+00 1.38035226e+00
-1.08612514e+00 2.89406300e-01 8.07243049e-01 -4.94130939e-01
-3.88785779e-01 4.78967354e-02 -9.16148961e-01 4.55302447e-01
1.02446258e+00 -4.94615026e-02 9.76509154e-01 1.07427490e+00
-3.13768297e-01 -1.19847074e-01 -1.03079331e+00 1.54989374e+00
-1.50870681e-01 -1.30509925e+00 -2.04069942e-01 1.86040729e-01
8.15628052e-01 3.31618041e-01 2.87713379e-01 4.32234824e-01
4.47256297e-01 -1.01033580e+00 8.47677350e-01 3.49152058e-01
5.31372249e-01 -9.43303466e-01 6.29682660e-01 3.53729963e-01
-7.18999863e-01 1.08535076e-02 -6.87286913e-01 -2.12375373e-01
-4.83040586e-02 1.32977092e+00 -6.04012668e-01 2.22720608e-01
8.22533786e-01 4.67782289e-01 -5.08139312e-01 7.03789651e-01
-1.61579013e-01 7.09827304e-01 -9.73150134e-01 -1.88707441e-01
4.79412884e-01 -5.22071958e-01 4.94251668e-01 1.06097400e+00
4.43391114e-01 -1.94969922e-01 -1.46085069e-01 3.13679665e-01
-3.46774310e-01 1.18003011e-01 -3.84000828e-03 -7.77949467e-02
3.46936554e-01 1.07993388e+00 -3.66101980e-01 -4.04055327e-01
6.90178648e-02 1.14162004e+00 6.88751638e-01 1.54899463e-01
-7.65846908e-01 -3.38653088e-01 1.01687551e+00 -6.31708950e-02
4.25275266e-01 -5.00463545e-01 -4.38249022e-01 -1.48346245e+00
1.37296945e-01 -1.16161907e+00 2.10438013e-01 -1.88699827e-01
-1.13544953e+00 3.24230254e-01 -3.92446995e-01 -8.54673505e-01
-8.18795934e-02 -8.76708031e-01 2.08599091e-01 7.91339457e-01
-1.13733065e+00 -5.50041795e-01 2.38168448e-01 3.74954164e-01
3.32758352e-02 -8.26491565e-02 6.04991615e-01 6.85106575e-01
-5.50005972e-01 1.13534284e+00 8.60987842e-01 -1.98484465e-01
5.42524338e-01 -1.25373375e+00 3.81160051e-01 6.58532798e-01
4.36930865e-01 9.39953983e-01 7.34300196e-01 -3.73208344e-01
-1.96660972e+00 -1.01587260e+00 6.21533334e-01 1.56293899e-01
8.33365977e-01 -4.71934825e-01 -7.47015536e-01 6.59852028e-01
-5.62471114e-02 2.68599600e-01 5.05330205e-01 3.03562015e-01
-4.18433130e-01 -5.94145954e-01 -1.13840461e+00 7.36813605e-01
1.03862453e+00 -4.62342322e-01 1.47057265e-01 7.47028649e-01
3.98695558e-01 -8.50794673e-01 -8.98253918e-01 3.13621521e-01
6.41949236e-01 -1.21697271e+00 9.88228798e-01 -4.27066594e-01
1.99426278e-01 -9.55743641e-02 -3.01074505e-01 -1.38849545e+00
-2.02149317e-01 -1.04770601e+00 -6.55287325e-01 7.02596664e-01
4.39918697e-01 -8.58019948e-01 9.30415988e-01 5.79167366e-01
-7.31110498e-02 -1.15007341e+00 -1.06755793e+00 -9.16754425e-01
2.80047774e-01 -7.55509317e-01 4.41175789e-01 6.56203568e-01
-2.17058524e-01 1.92930460e-01 -5.70502102e-01 -7.75494725e-02
6.58386827e-01 -1.61298379e-01 8.02256286e-01 -7.13070333e-01
-8.44135106e-01 -5.01028597e-01 -4.18789923e-01 -1.52145112e+00
3.55624370e-02 -1.24913788e+00 -2.08800361e-01 -9.69428718e-01
-7.14818612e-02 -6.14825666e-01 -2.85820425e-01 5.44606924e-01
-1.53746292e-01 4.42170084e-01 3.67065281e-01 2.54677404e-02
-3.46182823e-01 8.59061658e-01 9.73020136e-01 -1.97607890e-01
-1.40724583e-02 -1.34867877e-01 -4.70993370e-01 8.47675443e-01
7.54367769e-01 -4.29747403e-01 -2.39497349e-01 -7.62755394e-01
8.78132820e-01 -2.69935548e-01 2.31944472e-01 -1.25275254e+00
1.95500389e-01 2.68443555e-01 2.71801680e-01 -4.57028061e-01
3.90908718e-01 -5.32577693e-01 1.35814309e-01 8.61879468e-01
-3.25808704e-01 2.58996427e-01 3.26120824e-01 4.17715967e-01
-6.51904419e-02 -5.02128899e-01 9.89661634e-01 3.36588174e-02
-2.07003757e-01 4.40554768e-01 -2.44405374e-01 2.84370363e-01
5.69469571e-01 3.66085708e-01 2.89679673e-02 -4.11511242e-01
-5.91580927e-01 1.64300650e-01 3.45908374e-01 -1.39716968e-01
4.83887106e-01 -1.36227310e+00 -5.33671260e-01 2.24977329e-01
-6.31640434e-01 9.12957042e-02 3.73924822e-01 9.94911730e-01
-1.06399143e+00 4.72095698e-01 1.65871933e-01 -4.59115595e-01
-9.46864963e-01 2.94706106e-01 2.20290974e-01 -6.46519482e-01
-3.93443316e-01 1.35418510e+00 -2.58913003e-02 -5.63398838e-01
3.72854859e-01 -3.56640667e-01 6.42819762e-01 1.32297069e-01
5.14704764e-01 5.64075112e-01 5.29242516e-01 -1.11402631e-01
-3.29217523e-01 2.16999963e-01 -1.58449858e-01 -3.07476044e-01
1.33455193e+00 9.16851610e-02 -2.13551015e-01 2.66623437e-01
1.63326728e+00 1.04812659e-01 -1.35744536e+00 -2.21635714e-01
-1.51646703e-01 -6.10772014e-01 5.02933562e-01 -4.68380988e-01
-1.44281971e+00 1.04662395e+00 8.31957936e-01 -1.26536861e-01
9.80975032e-01 -7.75996149e-01 1.15895402e+00 9.84951615e-01
4.83635575e-01 -1.11704993e+00 -3.65818530e-01 8.97068679e-01
7.45142817e-01 -1.00536311e+00 3.72313768e-01 1.39142960e-01
-4.46135938e-01 1.08836389e+00 3.18739027e-01 -2.37745866e-01
8.22605729e-01 3.79995972e-01 -2.62516469e-01 2.69723654e-01
-7.70247161e-01 2.58164972e-01 1.68214798e-01 9.23597142e-02
3.07944715e-01 8.15808252e-02 -6.39251053e-01 4.89872187e-01
-7.73834646e-01 2.07568277e-02 1.95628285e-01 7.05534160e-01
-1.87798813e-01 -1.10630512e+00 -3.32546473e-01 6.48740649e-01
-6.01160467e-01 -6.06846154e-01 -2.84701847e-02 2.75502264e-01
-1.01572461e-01 4.57444370e-01 -1.06808320e-01 -4.04163420e-01
-1.30819216e-01 -8.94636586e-02 6.38902664e-01 -1.67806461e-01
-7.47908175e-01 -2.98555791e-01 1.13576641e-02 -7.95533597e-01
3.07854474e-01 -5.24127901e-01 -1.10149467e+00 -6.25570714e-01
-1.64079189e-01 2.89961785e-01 8.28831077e-01 6.71535552e-01
6.13425732e-01 1.11253209e-01 4.71680909e-01 -8.28591526e-01
-1.26763797e+00 -8.95983100e-01 -6.12761915e-01 1.16897941e-01
3.91990125e-01 -5.02818584e-01 -5.99783361e-01 -2.72550911e-01] | [8.408596992492676, 3.3758227825164795] |
d7589ca7-4456-44cf-a0c3-fe46fca8fb3a | a-dual-symmetric-gauss-seidel-alternating | 1902.09135 | null | https://arxiv.org/abs/1902.09135v2 | https://arxiv.org/pdf/1902.09135v2.pdf | A Dual Symmetric Gauss-Seidel Alternating Direction Method of Multipliers for Hyperspectral Sparse Unmixing | Since sparse unmixing has emerged as a promising approach to hyperspectral unmixing, some spatial-contextual information in the hyperspectral images has been exploited to improve the performance of the unmixing recently. The total variation (TV) has been widely used to promote the spatial homogeneity as well as the smoothness between adjacent pixels. However, the computation task for hyperspectral sparse unmixing with a TV regularization term is heavy. Besides, the convergence of the primal alternating direction method of multipliers (ADMM) for the hyperspectral sparse unmixing with a TV regularization term has not been explained in details. In this paper, we design an efficient and convergent dual symmetric Gauss-Seidel ADMM (sGS-ADMM) for hyperspectral sparse unmixing with a TV regularization term. We also present the global convergence and local linear convergence rate analysis for this algorithm. As demonstrated in numerical experiments, our algorithm can obviously improve the efficiency of the unmixing compared with the state-of-the-art algorithm. More importantly, we can obtain images with higher quality. | ['Peipei Tang', 'Zheng Ma', 'Chengjing Wang', 'Longfei Ren'] | 2019-02-25 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 4.39000160e-01 -6.12489045e-01 9.41977799e-02 -9.85833164e-03
-7.76766360e-01 -2.79963791e-01 2.05332920e-01 -4.53637183e-01
2.58510914e-02 6.32683873e-01 1.22159488e-01 -1.01317629e-01
-4.27638501e-01 -5.27990162e-01 -4.23488379e-01 -1.45939314e+00
4.88058746e-01 4.51643839e-02 -5.29899001e-01 -1.09154731e-01
-1.07850462e-01 2.63926506e-01 -1.14399910e+00 -1.84661038e-02
1.45160460e+00 8.51554632e-01 4.64477569e-01 1.54256120e-01
2.24986002e-02 4.84562099e-01 -4.12023105e-02 1.86256543e-01
5.23444951e-01 -6.87101364e-01 -4.71933603e-01 8.28483462e-01
6.22785032e-01 -1.64794520e-01 -4.22320008e-01 1.62102592e+00
4.21286017e-01 2.51984537e-01 3.71494591e-01 -1.16441274e+00
-5.79347253e-01 4.02517378e-01 -1.35499620e+00 -1.14908159e-01
-3.04759413e-01 -4.28429507e-02 7.24275410e-01 -9.31060791e-01
3.08388233e-01 9.48570073e-01 4.29322511e-01 -8.39783549e-02
-1.14768457e+00 -7.17363358e-01 1.14332676e-01 1.23127908e-01
-1.65495908e+00 -2.46688083e-01 9.28349197e-01 -3.94966960e-01
2.07412690e-01 5.59726894e-01 8.05112302e-01 3.21074247e-01
-2.39743710e-01 7.99718559e-01 1.19388163e+00 -3.66592139e-01
-2.88345870e-02 -3.34227681e-01 -8.57812166e-02 6.11956716e-01
5.29973924e-01 1.56948026e-02 -1.32019386e-01 -1.82944596e-01
9.08796191e-01 4.17339563e-01 -6.45723581e-01 -2.43517369e-01
-1.34956419e+00 7.50576496e-01 5.67010820e-01 4.11895573e-01
-6.22327268e-01 -1.03500679e-01 6.50665862e-03 8.86651278e-02
7.75637805e-01 5.43753654e-02 -3.62079986e-03 4.64535445e-01
-1.32942486e+00 -5.13904952e-02 1.10114172e-01 8.14684093e-01
1.05829942e+00 4.89256710e-01 3.57446112e-02 1.10687888e+00
6.70819819e-01 9.95555401e-01 3.47469389e-01 -1.05137277e+00
5.05019367e-01 4.69276100e-01 8.60916674e-02 -1.26989830e+00
-1.92757413e-01 -6.32251382e-01 -1.51724756e+00 9.84724239e-02
4.29201089e-02 -2.25530639e-01 -7.31664300e-01 1.40665281e+00
3.91168475e-01 6.38658166e-01 8.46500471e-02 1.28579950e+00
4.40502971e-01 1.24731350e+00 -8.96839648e-02 -9.53804314e-01
9.80429053e-01 -1.10252583e+00 -1.11039460e+00 -4.27265942e-01
3.77563715e-01 -1.11206603e+00 4.15293902e-01 4.73203838e-01
-1.03639150e+00 -3.94012034e-01 -9.61086333e-01 1.55387700e-01
3.21391493e-01 2.80229479e-01 7.45090961e-01 3.78170907e-01
-6.04235291e-01 2.49506965e-01 -8.92379403e-01 -4.33434173e-02
3.56975466e-01 6.46748096e-02 -3.96625608e-01 -4.95505333e-01
-8.98104787e-01 5.50858736e-01 4.72715825e-01 5.25920510e-01
-7.21190989e-01 -5.75156212e-01 -8.13021302e-01 -9.80940536e-02
2.13992119e-01 -5.74686587e-01 6.50167406e-01 -1.38970876e+00
-1.37776411e+00 6.43813312e-01 -4.06666934e-01 2.49191776e-01
1.84334248e-01 2.32753288e-02 -5.56754172e-01 2.56186217e-01
1.55830279e-01 1.65906578e-01 1.11671579e+00 -1.36170721e+00
-4.37024891e-01 -4.56625938e-01 -5.18988848e-01 3.72658789e-01
-4.78049278e-01 -2.26494223e-01 -2.13027328e-01 -1.03899801e+00
9.34098184e-01 -1.12023127e+00 -4.82123375e-01 -1.49471730e-01
-3.40684056e-01 5.43174803e-01 9.65913951e-01 -1.14263809e+00
1.29003930e+00 -2.45212674e+00 5.27288616e-01 2.61685163e-01
1.23313062e-01 4.13762540e-01 -1.85553282e-01 3.28734905e-01
-4.99077022e-01 -2.59555459e-01 -8.53780270e-01 -1.02232561e-01
-3.32214832e-01 9.21551809e-02 -1.44488752e-01 9.95782793e-01
-3.45684737e-01 6.09411120e-01 -9.92176771e-01 -2.89321840e-01
3.32622468e-01 4.89240229e-01 -3.54242772e-01 1.56443015e-01
5.47584742e-02 7.75692463e-01 -3.80903095e-01 6.63633764e-01
1.38397145e+00 -3.42392415e-01 2.22023770e-01 -6.04024947e-01
-4.43764478e-01 -4.82932985e-01 -1.50520372e+00 1.97971261e+00
-2.14648455e-01 3.11688066e-01 7.67198265e-01 -1.22974050e+00
7.19744861e-01 4.30587769e-01 7.92833209e-01 -4.13930506e-01
-6.13850206e-02 5.15445888e-01 -2.93000102e-01 -6.08511269e-01
3.58496428e-01 -6.32172942e-01 6.80975974e-01 2.14579359e-01
-3.51597607e-01 -1.50240332e-01 1.12042561e-01 3.10088675e-02
3.59358490e-01 -6.93797097e-02 2.36236542e-01 -5.84638953e-01
7.21303582e-01 1.01704203e-01 6.10487700e-01 1.59055933e-01
1.58519059e-01 5.87530434e-01 -2.10375115e-01 -1.04589880e-01
-9.33655918e-01 -6.50186956e-01 -3.51473391e-01 5.42166889e-01
4.06968892e-01 2.09226124e-02 -4.35515940e-01 -8.44626725e-02
-3.43984604e-01 3.10327709e-01 -3.48328166e-02 5.95800579e-02
-3.24259728e-01 -1.47264373e+00 6.17755018e-02 5.83186634e-02
8.22220504e-01 -4.64952052e-01 3.80564630e-01 2.88185686e-01
-6.29093826e-01 -8.51122439e-01 -5.66733539e-01 -3.31918806e-01
-1.25435436e+00 -8.79364252e-01 -1.16977346e+00 -9.42218363e-01
9.35916364e-01 1.00657368e+00 4.55866754e-01 -1.13238700e-01
-6.47891238e-02 -8.73105451e-02 -2.31952429e-01 2.19858568e-02
-1.04858927e-01 -2.86587715e-01 -5.59894741e-02 5.42906404e-01
-1.29405379e-01 -7.11664319e-01 -6.03712678e-01 3.69451344e-01
-1.55439198e+00 4.60507929e-01 6.56195521e-01 7.84414589e-01
8.16640496e-01 6.29839957e-01 1.07612208e-01 -7.19397187e-01
1.56542972e-01 -5.61561465e-01 -7.79992998e-01 2.23794281e-01
-5.20366371e-01 -2.69876868e-01 3.16816211e-01 -3.15048009e-01
-1.22931349e+00 1.63612977e-01 1.41372094e-02 -5.93301415e-01
2.06871539e-01 9.33309853e-01 -4.56924975e-01 -4.16407466e-01
3.46810341e-01 5.73984683e-01 1.67450279e-01 -5.86302161e-01
4.78509933e-01 7.32341766e-01 3.10408473e-01 -4.22632217e-01
1.14776349e+00 8.65127623e-01 2.97903597e-01 -1.18241870e+00
-1.08640122e+00 -8.02106261e-01 -2.10528076e-01 7.22575262e-02
7.35381663e-01 -1.32149422e+00 -3.88895541e-01 9.17253137e-01
-8.84496987e-01 -2.53904492e-01 7.30502829e-02 7.49877393e-01
-2.18515545e-01 9.16551709e-01 -6.17143035e-01 -7.12255359e-01
-2.49851704e-01 -1.21578157e+00 6.01501524e-01 1.30740672e-01
4.93102461e-01 -1.11717486e+00 1.49204388e-01 5.63463688e-01
4.70471889e-01 4.18857872e-01 5.72606981e-01 2.07845703e-01
-6.13008916e-01 4.96731922e-02 -4.69213039e-01 5.43179810e-01
3.43586713e-01 -1.65199056e-01 -7.65304923e-01 -5.18484294e-01
5.05814373e-01 7.49633163e-02 8.56515050e-01 8.12754929e-01
1.02174103e+00 -5.90208650e-01 -2.04432100e-01 1.11433101e+00
1.89431596e+00 1.12628657e-02 7.05025792e-01 4.24201399e-01
1.13741672e+00 4.64888006e-01 6.37495756e-01 6.21570885e-01
-6.73802048e-02 2.70481497e-01 5.93387008e-01 -4.95794088e-01
4.04939592e-01 8.39390978e-02 1.36889040e-01 1.19819820e+00
-3.18023324e-01 -5.61608039e-02 -6.06880546e-01 4.10891742e-01
-2.05014801e+00 -1.13018215e+00 -9.12295938e-01 2.14572763e+00
6.32802844e-01 -7.93580234e-01 -3.63180995e-01 2.91450620e-01
1.05468369e+00 7.08000481e-01 -4.64662015e-01 4.06550169e-01
-4.05852228e-01 -1.34768575e-01 6.82037771e-01 8.11385453e-01
-1.12815583e+00 6.84126854e-01 5.80482197e+00 1.09960365e+00
-1.10086119e+00 3.55013698e-01 4.53388572e-01 1.64798796e-02
-4.92478043e-01 1.79129228e-01 -2.16078073e-01 4.38320756e-01
2.89454520e-01 2.50602700e-02 8.25849831e-01 4.77878481e-01
5.13125360e-01 8.31286311e-02 -2.33062401e-01 1.42946255e+00
3.54588032e-02 -1.14476669e+00 2.02753603e-01 2.94394016e-01
1.40168667e+00 -1.63534820e-01 9.64139104e-02 -2.16293231e-01
-9.52791050e-02 -8.38877261e-01 4.72436786e-01 3.55305433e-01
7.93341815e-01 -7.51171649e-01 4.63536650e-01 4.01470900e-01
-1.23696971e+00 8.47571120e-02 -7.07391858e-01 -6.02789223e-02
4.98069376e-01 1.34092069e+00 1.90340906e-01 9.55453157e-01
2.02998519e-01 1.14768052e+00 5.96830472e-02 9.78059947e-01
-3.93178254e-01 6.18024409e-01 -2.75402725e-01 6.76020086e-01
5.90584934e-01 -1.23581004e+00 7.54916608e-01 7.52232671e-01
7.05736578e-01 6.36510193e-01 3.53187591e-01 7.49907315e-01
8.22052360e-02 3.21285516e-01 -3.31542552e-01 -3.44191521e-01
1.03745587e-01 1.43640435e+00 -2.87501305e-01 -2.59832412e-01
-4.47841555e-01 9.51283395e-01 -3.24426442e-01 6.95129871e-01
-6.17761195e-01 1.12701885e-01 6.39961064e-01 3.06385774e-02
1.18040688e-01 -3.40868115e-01 -2.00769141e-01 -1.59001601e+00
-1.13976501e-01 -1.15586472e+00 3.62054080e-01 -8.39689374e-01
-1.29161823e+00 3.36018384e-01 -3.56101781e-01 -1.54692268e+00
5.99384904e-01 -4.03875679e-01 -4.97583896e-01 1.21392465e+00
-1.87122703e+00 -1.31041515e+00 -7.29744196e-01 7.20406532e-01
2.12369353e-01 -1.63561497e-02 4.78198469e-01 4.67749476e-01
-8.50176036e-01 -1.98847696e-01 6.88699961e-01 -1.30216226e-01
6.51609838e-01 -8.36814940e-01 -5.48656642e-01 1.33163106e+00
-2.00525120e-01 5.01395941e-01 6.89511240e-01 -5.86725056e-01
-1.36981356e+00 -1.24716783e+00 3.93296540e-01 3.64452690e-01
7.64128387e-01 6.03245258e-01 -1.01464331e+00 8.75485480e-01
4.69064921e-01 1.28325522e-01 9.15969074e-01 -2.60048449e-01
-1.21336222e-01 -2.94267774e-01 -9.32436466e-01 4.88542765e-01
6.59375310e-01 -2.92373925e-01 -1.62286311e-01 7.68130720e-01
4.08658415e-01 -4.40078974e-01 -9.34504211e-01 4.56835538e-01
2.35679522e-01 -8.55276644e-01 1.01665425e+00 -1.78350925e-01
5.47645330e-01 -7.96111763e-01 -3.41058403e-01 -1.49289930e+00
-8.14531565e-01 -8.52492392e-01 4.62908968e-02 1.11411774e+00
1.74509093e-01 -6.87431455e-01 4.69511509e-01 1.98534876e-01
-2.47278184e-01 -2.24790588e-01 -5.14815331e-01 -6.75825655e-01
-1.70379296e-01 -6.77604079e-02 4.06827331e-01 1.49341428e+00
1.33862505e-02 1.41021773e-01 -8.13893974e-01 8.18335354e-01
1.12164998e+00 4.30177450e-01 4.01988417e-01 -1.04343939e+00
-2.71226108e-01 -3.55070949e-01 7.87552074e-02 -1.10243285e+00
1.85802400e-01 -1.05945802e+00 -6.57217130e-02 -1.73680270e+00
4.32200104e-01 -3.77215177e-01 -3.96840751e-01 2.57842422e-01
-5.51948309e-01 2.65359759e-01 6.15095254e-03 7.17351258e-01
-1.83517769e-01 7.80794322e-01 1.68363845e+00 -5.27386367e-01
-2.47105345e-01 -2.59848982e-01 -6.85290754e-01 5.63004732e-01
6.47237897e-01 -4.95662510e-01 -3.06397885e-01 -6.80936337e-01
3.54214996e-01 2.94424742e-01 -4.20735590e-03 -8.35772812e-01
1.14062853e-01 -6.63073838e-01 1.43953204e-01 -4.03085649e-01
3.08578640e-01 -9.96721566e-01 7.03620374e-01 4.83299196e-01
5.77568524e-02 -4.08592165e-01 8.43996555e-02 5.53650618e-01
-6.39418066e-01 -3.09306502e-01 1.08790922e+00 -1.84454650e-01
-5.85852981e-01 8.07428837e-01 -1.74824044e-01 -2.17990801e-01
8.91742170e-01 9.35782939e-02 -1.57383442e-01 -2.10005149e-01
-6.49135530e-01 2.09029227e-01 5.83557606e-01 -2.36912414e-01
5.47321379e-01 -1.52802205e+00 -9.94383276e-01 2.79221565e-01
5.46256639e-02 7.20693022e-02 8.38432193e-01 1.33758628e+00
-8.86356652e-01 1.40257001e-01 -1.22926235e-01 -5.34979165e-01
-1.10873461e+00 7.51237273e-01 4.35629338e-01 -1.38243273e-01
-3.93162429e-01 6.88872159e-01 4.34013754e-01 -2.86518842e-01
-3.05598438e-01 1.45712450e-01 1.07970200e-02 5.07180914e-02
6.45975769e-01 7.23294377e-01 -1.83035403e-01 -8.50447953e-01
-8.35632235e-02 7.12498724e-01 5.04875124e-01 -1.48021638e-01
1.40499067e+00 -3.70979488e-01 -9.79778409e-01 1.74453765e-01
1.37458825e+00 2.48194441e-01 -1.20760596e+00 -2.80874163e-01
-5.65713525e-01 -7.24248886e-01 6.34051323e-01 -2.35852614e-01
-1.42313087e+00 8.76303971e-01 4.28109169e-01 2.05649510e-01
1.30143833e+00 -6.94206476e-01 9.44880247e-01 5.64115904e-02
2.12753221e-01 -8.20241392e-01 -2.57759094e-01 2.52144426e-01
7.82590449e-01 -1.33965516e+00 3.30480069e-01 -6.91192091e-01
-4.97373462e-01 9.53532040e-01 1.70212269e-01 -5.70300827e-03
6.47292674e-01 -1.06755212e-01 2.45805934e-01 -1.67206079e-01
1.41118661e-01 -2.50171632e-01 2.99131155e-01 1.77578136e-01
3.52165550e-01 1.96780622e-01 -4.44925129e-01 2.79081687e-02
4.55974847e-01 -1.04662836e-01 2.61263818e-01 6.22230887e-01
-5.54308116e-01 -1.03298485e+00 -8.75250995e-01 2.09898293e-01
-3.19491386e-01 -3.60240281e-01 2.26918235e-01 1.87857881e-01
4.19611372e-02 1.22069085e+00 -4.00644004e-01 5.38925827e-02
-1.77943408e-02 -1.82407543e-01 3.79192680e-01 -5.33058405e-01
1.35285065e-01 6.30741477e-01 -4.16794717e-01 -2.53624529e-01
-9.67199445e-01 -4.92856562e-01 -9.57718790e-01 -3.14620346e-01
-6.19439781e-01 3.94751191e-01 5.26440442e-01 9.19267952e-01
1.16959512e-01 2.67327935e-01 8.82065535e-01 -9.76375639e-01
-2.76850641e-01 -9.36090350e-01 -1.35263813e+00 3.99957746e-01
3.02025795e-01 -1.74657747e-01 -6.74404144e-01 1.82440683e-01] | [10.106460571289062, -2.0343573093414307] |
d9ce17af-226f-49bd-b1b4-21be6cc61ccc | cblue-a-chinese-biomedical-language | 2106.08087 | null | https://arxiv.org/abs/2106.08087v6 | https://arxiv.org/pdf/2106.08087v6.pdf | CBLUE: A Chinese Biomedical Language Understanding Evaluation Benchmark | Artificial Intelligence (AI), along with the recent progress in biomedical language understanding, is gradually changing medical practice. With the development of biomedical language understanding benchmarks, AI applications are widely used in the medical field. However, most benchmarks are limited to English, which makes it challenging to replicate many of the successes in English for other languages. To facilitate research in this direction, we collect real-world biomedical data and present the first Chinese Biomedical Language Understanding Evaluation (CBLUE) benchmark: a collection of natural language understanding tasks including named entity recognition, information extraction, clinical diagnosis normalization, single-sentence/sentence-pair classification, and an associated online platform for model evaluation, comparison, and analysis. To establish evaluation on these tasks, we report empirical results with the current 11 pre-trained Chinese models, and experimental results show that state-of-the-art neural models perform by far worse than the human ceiling. Our benchmark is released at \url{https://tianchi.aliyun.com/dataset/dataDetail?dataId=95414&lang=en-us}. | ['Xiaozhuan Liang', 'Xin Shang', 'Ningyu Zhang', 'Kangping Yin', 'Chuanqi Tan', 'Qingcai Chen', 'Buzhou Tang', 'Kunli Zhang', 'Hongying Zan', 'Jun Yan', 'Linfeng Li', 'Zheng Yuan', 'Hui Zong', 'Baobao Chang', 'Zhifang Sui', 'Guotong Xie', 'Yuan Ni', 'Luo Si', 'Fei Huang', 'Mosha Chen', 'Jian Xu', 'Lei LI', 'Zhen Bi'] | 2021-06-15 | null | https://aclanthology.org/2022.acl-long.544 | https://aclanthology.org/2022.acl-long.544.pdf | acl-2022-5 | ['medical-concept-normalization', 'medical-relation-extraction', 'sentence-pair-classification'] | ['medical', 'medical', 'natural-language-processing'] | [ 4.06838059e-01 3.88449281e-02 -3.21471423e-01 -5.40778458e-01
-9.80189025e-01 -5.70413284e-02 2.42719680e-01 4.28403974e-01
-8.75847936e-01 9.06354368e-01 4.70146239e-01 -4.92883623e-01
1.82092890e-01 -4.63846326e-01 -2.87548244e-01 -4.43901747e-01
1.05071932e-01 7.67073095e-01 -3.91749859e-01 -2.47090328e-02
4.01386991e-02 6.87874705e-02 -5.76218188e-01 7.09488213e-01
9.93578911e-01 6.34119093e-01 7.89793208e-02 7.46303201e-01
-1.61465973e-01 9.04190481e-01 -5.67719340e-01 -5.97213864e-01
-5.29218793e-01 -6.30835176e-01 -1.25335479e+00 -5.59952199e-01
-2.45953009e-01 -2.36815233e-02 -1.75603956e-01 1.14893603e+00
9.39444423e-01 -4.67306614e-01 5.94016850e-01 -9.38304484e-01
-9.24979687e-01 7.82097399e-01 -2.73703545e-01 3.89028102e-01
2.77410954e-01 -1.61251143e-01 6.94443166e-01 -7.47065842e-01
9.33204710e-01 9.61834133e-01 6.28144920e-01 1.14880061e+00
-7.33378291e-01 -8.81279588e-01 -1.92599207e-01 3.23231995e-01
-1.40282583e+00 -6.57164574e-01 2.81292230e-01 -1.48855433e-01
1.18185747e+00 2.15079382e-01 3.30117978e-02 1.30037022e+00
5.07191002e-01 1.12708521e+00 7.16011584e-01 -4.67628151e-01
7.79732689e-03 -5.74864307e-03 4.76290077e-01 5.28926790e-01
2.99757302e-01 -2.04570770e-01 -2.84111798e-01 -1.19344659e-01
2.76919782e-01 -8.23887661e-02 -4.49652463e-01 5.73401809e-01
-1.73243856e+00 7.09507823e-01 3.46510231e-01 7.26661682e-01
-3.60728443e-01 -1.81046709e-01 7.79374897e-01 3.91723305e-01
7.54857004e-01 5.56785464e-01 -8.44732702e-01 -3.24511975e-01
-5.26221395e-01 -8.42282996e-02 1.07419920e+00 8.81294429e-01
-4.87043671e-02 -3.42622876e-01 -1.16084144e-01 1.01320410e+00
-9.88092180e-03 4.51455593e-01 8.73299360e-01 -5.10801077e-01
6.81381226e-01 5.91151834e-01 -4.60082710e-01 -7.85137892e-01
-8.47440243e-01 -3.33739877e-01 -1.59216070e+00 -7.43327618e-01
3.29771250e-01 -5.03473341e-01 -8.59490871e-01 1.68367958e+00
1.81194320e-02 1.05022497e-01 4.97075051e-01 3.48146409e-01
1.32899499e+00 5.48566222e-01 5.19341052e-01 -2.89010465e-01
1.53107202e+00 -8.91857207e-01 -1.15169239e+00 -3.29711348e-01
1.18084824e+00 -7.78458178e-01 6.11868262e-01 3.29941809e-01
-7.91514277e-01 -3.00900102e-01 -6.77237749e-01 -2.95967907e-01
-5.68955541e-01 2.38913849e-01 5.45407176e-01 6.09085739e-01
-8.30314279e-01 3.02757949e-01 -1.15019131e+00 -5.88270187e-01
5.85493386e-01 2.48626411e-01 -6.12725914e-01 -2.19971374e-01
-1.47677410e+00 1.15304840e+00 7.35837281e-01 1.08072452e-01
-4.43180114e-01 -9.45971847e-01 -7.98700273e-01 -1.96677819e-01
7.30600357e-02 -7.15934217e-01 1.22290504e+00 -6.91937089e-01
-9.74489450e-01 1.27824998e+00 -2.49834538e-01 -5.46639144e-01
2.16316938e-01 -2.33668804e-01 -8.28475058e-01 3.12839188e-02
2.55202889e-01 7.78839588e-01 -1.92452654e-01 -6.82861626e-01
-3.73860270e-01 -4.95381415e-01 -5.30110717e-01 -8.53229985e-02
-5.40484190e-01 4.86640394e-01 -5.82062662e-01 -6.24938011e-01
-3.99715990e-01 -7.61797488e-01 -4.10997450e-01 -2.56320417e-01
-8.35024953e-01 -2.46451393e-01 9.42700654e-02 -1.01379406e+00
1.57111728e+00 -2.06429195e+00 -2.11418703e-01 -2.14582369e-01
1.29976541e-01 3.23322743e-01 -3.26829344e-01 3.93495798e-01
-4.73383695e-01 5.89649439e-01 -4.95320559e-01 -3.05782288e-01
-2.44786665e-01 9.59056243e-02 1.80636838e-01 2.18296528e-01
3.89330477e-01 1.17065620e+00 -8.71355712e-01 -8.46338689e-01
-2.13799968e-01 3.34742576e-01 -4.80950445e-01 2.06665531e-01
7.64760226e-02 5.64297795e-01 -8.67510319e-01 8.34500492e-01
4.12604213e-01 -6.23013914e-01 2.59465635e-01 -9.80681032e-02
2.80732661e-01 2.53117383e-01 -5.05508721e-01 1.82425356e+00
-1.35006368e-01 3.57546210e-01 -4.14553285e-02 -1.21509683e+00
4.72973794e-01 7.07175910e-01 6.51834548e-01 -7.03558683e-01
3.59305739e-01 1.92804292e-01 2.87381738e-01 -7.21883953e-01
7.99199790e-02 -1.20830592e-02 -2.02351794e-01 2.34507844e-01
1.96060166e-02 3.87921557e-02 3.42939407e-01 2.46875495e-01
1.21401465e+00 -3.12868297e-01 8.82490575e-01 -3.65191430e-01
6.69249117e-01 2.07502887e-01 7.31544435e-01 5.73605835e-01
-4.55857009e-01 5.74943602e-01 4.28213686e-01 -5.30764282e-01
-7.68793166e-01 -7.85790920e-01 -5.62074661e-01 8.50330591e-01
-3.04173976e-01 -5.59943676e-01 -1.10645103e+00 -5.65681100e-01
-3.54299337e-01 7.77529418e-01 -7.34769583e-01 -2.77001619e-01
-4.84486639e-01 -1.48861825e+00 1.03726482e+00 5.82701027e-01
4.96487916e-01 -1.39506328e+00 -2.25944474e-01 3.40091228e-01
-5.84351659e-01 -1.31169772e+00 -5.63632607e-01 9.64352414e-02
-9.39566135e-01 -1.08839452e+00 -7.83025920e-01 -8.76809776e-01
7.75944650e-01 -4.82134938e-01 1.37489545e+00 3.04080188e-01
-6.15491271e-01 2.52504826e-01 -3.89164537e-01 -1.14648032e+00
-8.16290498e-01 4.39478368e-01 -9.22335163e-02 -4.43401694e-01
9.17332411e-01 1.04642406e-01 -4.84005064e-01 1.46150645e-02
-1.09993660e+00 4.91347194e-01 7.35448182e-01 1.22265518e+00
4.61415172e-01 -4.06781226e-01 9.92328763e-01 -1.26173532e+00
6.94090307e-01 -6.28350258e-01 1.47486150e-01 6.68963790e-01
-4.04453635e-01 -1.07324913e-01 5.04628003e-01 -1.77837446e-01
-1.01620173e+00 -1.20957099e-01 -7.51392484e-01 3.44993949e-01
-5.65202475e-01 1.09976780e+00 -1.17252432e-01 5.64459264e-01
6.47042334e-01 2.61462450e-01 -3.70370178e-03 -6.16115987e-01
1.82074785e-01 1.17438531e+00 5.40404201e-01 -2.77014673e-01
2.26404555e-02 2.28178307e-01 -6.13317549e-01 -1.01832342e+00
-1.03882480e+00 -4.40971106e-01 -7.45968819e-01 2.68957675e-01
1.44053066e+00 -1.01840937e+00 -5.46029806e-01 4.93705750e-01
-1.34859526e+00 -1.66125715e-01 2.35815763e-01 6.40611410e-01
-1.73605874e-01 2.27653340e-01 -1.11792696e+00 -2.89465040e-01
-9.62467849e-01 -1.09964156e+00 8.51265371e-01 1.80651575e-01
-5.55365443e-01 -1.15311348e+00 2.18942389e-01 6.55801415e-01
4.09912556e-01 1.21849895e-01 1.21925640e+00 -1.34746373e+00
1.66340306e-01 -3.46804261e-01 -2.12040827e-01 3.42511773e-01
1.87979475e-01 -3.12266737e-01 -8.16318452e-01 -1.25556588e-01
-5.01203015e-02 -5.08773744e-01 8.04984868e-01 4.22995687e-01
1.50571394e+00 1.25455791e-02 -7.16049373e-01 4.29414868e-01
1.02797902e+00 4.66556847e-01 6.51385963e-01 2.37543881e-01
5.13868332e-01 3.22607964e-01 2.94903070e-01 2.93184906e-01
5.92172265e-01 1.99185908e-01 -1.65540814e-01 -5.32251596e-01
4.15370949e-02 1.40906453e-01 8.97836592e-03 1.42713952e+00
3.47071066e-02 -3.53116542e-01 -1.36156988e+00 5.76809525e-01
-1.49785113e+00 -5.93349755e-01 8.67672041e-02 1.85855818e+00
1.52200174e+00 -1.07705817e-02 -4.54112023e-01 -3.67290527e-01
4.57863122e-01 -2.51004517e-01 -4.68231529e-01 -3.75976294e-01
-1.51632428e-01 2.96457589e-01 2.62093544e-01 1.08998105e-01
-1.32682478e+00 8.41296017e-01 5.81822777e+00 9.16324973e-01
-1.01440620e+00 2.06783965e-01 1.49935842e+00 3.11945021e-01
4.51958925e-02 -6.91136718e-01 -7.97111928e-01 3.48283857e-01
1.34971249e+00 -4.51351941e-01 -1.30012080e-01 5.77819526e-01
1.96560785e-01 1.64277852e-01 -1.28529048e+00 8.85478854e-01
2.75110006e-01 -1.39780128e+00 1.08406015e-01 -2.46709332e-01
6.02232993e-01 5.73766530e-01 -2.40386665e-01 2.79187024e-01
1.15728490e-01 -1.31500113e+00 -8.18583965e-02 6.06445312e-01
9.30811644e-01 -4.07110035e-01 1.28239858e+00 3.86837691e-01
-8.22620749e-01 1.36835441e-01 -4.42616969e-01 3.74250591e-01
2.38801628e-01 6.40606761e-01 -8.36380482e-01 8.80738497e-01
8.10253441e-01 9.18363571e-01 -5.33459842e-01 8.95452738e-01
1.63348466e-02 8.37377071e-01 -1.37033626e-01 -2.10533991e-01
-5.42825460e-02 1.22333104e-02 1.58605829e-01 1.69019556e+00
8.42616484e-02 4.41610366e-01 7.98674226e-02 6.74561441e-01
-4.83522654e-01 5.41197777e-01 -4.71944928e-01 -4.68363702e-01
2.44308233e-01 1.02294707e+00 -7.22105026e-01 -7.09159017e-01
-5.53115726e-01 6.44034863e-01 3.04044306e-01 1.52194589e-01
-7.73754835e-01 -4.37531352e-01 5.71803689e-01 -5.21966338e-01
-3.59243453e-01 1.04255803e-01 -4.85839367e-01 -1.36857057e+00
-3.22532386e-01 -1.29250395e+00 7.51912594e-01 -4.86775368e-01
-1.60112321e+00 8.42264235e-01 -7.86417425e-02 -9.59814131e-01
-1.42261297e-01 -7.78896749e-01 -4.13612574e-01 6.59384012e-01
-1.33577383e+00 -1.06050837e+00 7.41735324e-02 4.29009289e-01
6.87352479e-01 -2.74821818e-01 1.48950231e+00 7.96663761e-01
-9.26972508e-01 7.63757646e-01 2.76073635e-01 8.53500366e-01
1.08424485e+00 -9.67465162e-01 5.20710289e-01 5.38062215e-01
2.35776361e-02 7.39990771e-01 2.60362864e-01 -7.70843804e-01
-1.11881649e+00 -1.01032400e+00 1.50238550e+00 -5.84424675e-01
6.73035502e-01 -3.43041480e-01 -1.00559115e+00 7.09516406e-01
4.46500927e-01 -2.42313370e-01 1.16549003e+00 2.66496949e-02
1.24202676e-01 9.55574214e-02 -8.70961726e-01 5.18637955e-01
7.73313820e-01 -3.64751875e-01 -7.37122774e-01 6.05348110e-01
7.25599170e-01 -4.19649720e-01 -1.32034624e+00 7.17104793e-01
4.63856071e-01 -2.03455612e-01 9.74459350e-01 -1.27022076e+00
7.36230910e-01 7.01977462e-02 1.90351605e-01 -1.07888246e+00
-1.90391272e-01 -1.91817731e-01 4.80051160e-01 1.06523907e+00
1.00309873e+00 -6.70462847e-01 5.65557659e-01 7.24738896e-01
-9.57162082e-02 -9.21391845e-01 -9.38487351e-01 -2.50985205e-01
6.04174316e-01 -4.68492329e-01 2.57182628e-01 1.26246810e+00
4.11493838e-01 3.50770801e-01 -1.64520994e-01 -1.87789023e-01
3.01065415e-01 -1.24605417e-01 2.89286196e-01 -8.21036458e-01
7.61432052e-02 -4.48422849e-01 -2.27403760e-01 -7.26099491e-01
1.41428679e-01 -1.19978058e+00 3.34610008e-02 -1.64543545e+00
7.50185132e-01 -1.62582606e-01 -5.46690822e-01 9.04164016e-01
-6.43116117e-01 9.64898989e-02 9.07955971e-03 -4.40303683e-02
-6.83580041e-01 3.08277130e-01 1.11586106e+00 -3.48654360e-01
8.28486681e-02 -1.39518410e-01 -8.46298516e-01 6.30594671e-01
9.32689190e-01 -5.08453250e-01 7.83180147e-02 -7.25461185e-01
-7.51669705e-02 3.01746149e-02 -1.73385590e-01 -7.33125269e-01
3.96835178e-01 1.24793194e-01 5.30907869e-01 -5.00861168e-01
-2.02486083e-01 -4.05501485e-01 3.79770808e-02 9.15262640e-01
-8.13556731e-01 3.82656395e-01 3.49601418e-01 1.68516830e-01
-4.14330333e-01 -7.86647946e-02 6.29882455e-01 -1.80117086e-01
-5.58323979e-01 3.29136610e-01 -5.54825664e-01 4.61142153e-01
6.35743916e-01 2.05926850e-01 -4.33703691e-01 -1.39788136e-01
-7.69930661e-01 5.46985745e-01 -1.56448096e-01 6.29274905e-01
5.69062114e-01 -9.79760349e-01 -1.28220046e+00 -3.14970836e-02
4.65998173e-01 -9.19120908e-02 4.76430178e-01 1.13253462e+00
-7.36354589e-01 8.33440721e-01 -5.54711558e-02 -4.29409862e-01
-1.39732373e+00 4.78922993e-01 2.79231548e-01 -6.82979763e-01
-4.03710783e-01 7.31806278e-01 4.04525459e-01 -8.07879925e-01
2.13987514e-01 -5.37345290e-01 -4.71648842e-01 -2.50016004e-01
9.08266842e-01 -3.91545258e-02 1.61709309e-01 -4.06675726e-01
-6.40273750e-01 2.16308847e-01 -3.57407302e-01 4.31915492e-01
1.42130303e+00 1.58064857e-01 -3.84833038e-01 3.17986995e-01
1.41815484e+00 -5.29507995e-01 -9.35292151e-03 -2.91661680e-01
5.36563516e-01 2.81152755e-01 -2.96133906e-01 -1.22472465e+00
-1.09175408e+00 9.17741239e-01 6.19252563e-01 -2.16170132e-01
1.24505019e+00 6.65867552e-02 8.19698811e-01 6.66116774e-01
4.75767590e-02 -1.04886973e+00 -3.45175236e-01 7.26639748e-01
7.75492847e-01 -1.57123327e+00 6.29311055e-03 -4.46154088e-01
-8.38530123e-01 9.91920948e-01 5.14271438e-01 5.38331449e-01
1.02876484e+00 5.72803736e-01 6.13415360e-01 -2.00286433e-01
-8.18844616e-01 2.08643347e-01 3.40115696e-01 3.73187453e-01
1.17574084e+00 1.53430641e-01 -6.44959629e-01 1.03493762e+00
-9.64984745e-02 4.61223304e-01 2.82435834e-01 8.43385696e-01
1.53932035e-01 -1.31093776e+00 -5.65559790e-03 8.35766733e-01
-1.27743936e+00 -5.98837137e-01 -3.84969175e-01 8.35105717e-01
-1.12703957e-01 9.76201952e-01 -1.13622598e-01 -1.94403548e-02
3.22113991e-01 2.21678063e-01 3.52482498e-02 -6.95269287e-01
-6.62843466e-01 7.24881189e-03 3.83305162e-01 -4.63872850e-01
-5.37167788e-01 -4.30062205e-01 -1.50720763e+00 -3.95794325e-02
-2.31276214e-01 2.87352771e-01 5.40731072e-01 1.07532310e+00
5.74860334e-01 7.02296913e-01 -3.51209268e-02 -2.17420571e-02
-3.65269721e-01 -1.26214921e+00 -1.51750952e-01 4.38875854e-01
-4.64014895e-02 -4.96984720e-02 7.96876773e-02 3.91022384e-01] | [8.529882431030273, 8.759053230285645] |
15ea70d2-1e47-4415-8717-5eec4eec632e | multi-modal-bifurcated-network-for-depth | 2105.00690 | null | https://arxiv.org/abs/2105.00690v2 | https://arxiv.org/pdf/2105.00690v2.pdf | Multi-modal Bifurcated Network for Depth Guided Image Relighting | Image relighting aims to recalibrate the illumination setting in an image. In this paper, we propose a deep learning-based method called multi-modal bifurcated network (MBNet) for depth guided image relighting. That is, given an image and the corresponding depth maps, a new image with the given illuminant angle and color temperature is generated by our network. This model extracts the image and the depth features by the bifurcated network in the encoder. To use the two features effectively, we adopt the dynamic dilated pyramid modules in the decoder. Moreover, to increase the variety of training data, we propose a novel data process pipeline to increase the number of the training data. Experiments conducted on the VIDIT dataset show that the proposed solution obtains the \textbf{1}$^{st}$ place in terms of SSIM and PMS in the NTIRE 2021 Depth Guide One-to-one Relighting Challenge. | ['Sy-Yen Kuo', 'Hao-Lun Luo', 'Wei-Ting Chen', 'Hao-Hsiang Yang'] | 2021-05-03 | null | null | null | null | ['image-relighting'] | ['computer-vision'] | [ 3.06129962e-01 4.99058440e-02 -1.35895107e-02 -5.03427804e-01
-5.10249496e-01 -2.60250598e-01 2.86853760e-01 -4.79883403e-01
-2.98358679e-01 5.40135384e-01 2.67005891e-01 -2.35292673e-01
2.52237171e-01 -8.85425508e-01 -8.68652821e-01 -6.78627193e-01
4.43483144e-01 2.96599772e-02 7.62838647e-02 -1.98661596e-01
5.27552724e-01 3.99606079e-01 -1.46093953e+00 3.42937112e-01
9.11818862e-01 1.27083337e+00 4.10983264e-01 5.94811559e-01
1.18155770e-01 9.92276490e-01 -3.27102631e-01 -3.71393055e-01
4.19052601e-01 -4.85816956e-01 -5.76749086e-01 1.80094391e-01
4.64179784e-01 -1.00783598e+00 -7.40377247e-01 7.53323674e-01
6.76993847e-01 -4.46435884e-02 4.81233478e-01 -1.10693502e+00
-3.05074155e-01 2.07323283e-01 -9.53014612e-01 -1.19290790e-02
9.68429893e-02 4.16582912e-01 4.41843390e-01 -7.08938301e-01
6.19291902e-01 1.19979906e+00 2.38402516e-01 3.64251167e-01
-8.51336956e-01 -6.51049852e-01 -1.41475290e-01 3.25176477e-01
-1.29312146e+00 -2.83795327e-01 6.66562915e-01 -2.64035970e-01
6.97242558e-01 3.97889167e-02 5.72101653e-01 7.41542578e-01
4.67806101e-01 5.87744594e-01 1.01294148e+00 -4.27943081e-01
1.28133118e-01 -3.16003501e-01 -3.28963131e-01 8.23084950e-01
-1.43021256e-01 4.13915426e-01 -7.47500420e-01 4.67018366e-01
1.08247674e+00 -1.43327162e-01 -3.57262433e-01 -2.16433570e-01
-1.19575846e+00 5.83764613e-01 7.28937626e-01 -7.57098645e-02
-1.63768962e-01 5.62221885e-01 4.27572757e-01 1.46110639e-01
4.26657200e-01 1.38548017e-01 -3.08759570e-01 -8.53251815e-02
-7.79014111e-01 -7.30344802e-02 3.02894205e-01 8.98323238e-01
1.25500810e+00 3.55622098e-02 -1.89227402e-01 7.27312982e-01
2.05571622e-01 3.66085202e-01 2.97160685e-01 -1.28411436e+00
7.23626137e-01 4.35525388e-01 5.59763536e-02 -9.79040563e-01
-3.50437164e-01 -1.72345668e-01 -1.05959642e+00 3.71271342e-01
-3.77197377e-02 -2.20099449e-01 -1.22278774e+00 1.28620434e+00
2.97504932e-01 3.33196670e-01 1.06922567e-01 9.90624249e-01
8.03128839e-01 8.59192967e-01 -4.00753796e-01 -5.90474643e-02
1.06236112e+00 -1.18766820e+00 -5.08076668e-01 -3.54630053e-01
6.49031162e-01 -6.50818586e-01 1.02071083e+00 6.43509090e-01
-9.67756450e-01 -5.99770606e-01 -1.33578146e+00 -5.20276189e-01
3.89489792e-02 1.23309314e-01 4.61584121e-01 4.58524674e-01
-1.11623287e+00 5.95630169e-01 -6.29906952e-01 5.90582518e-03
3.26185822e-01 3.08116823e-01 1.14766052e-02 -5.38509429e-01
-1.10449803e+00 6.86973631e-01 5.40981770e-01 1.15829214e-01
-1.15037537e+00 -5.37213683e-01 -9.11590874e-01 -1.02517307e-01
2.25201137e-02 -5.36352217e-01 1.02398872e+00 -9.47594523e-01
-1.66607678e+00 6.53520525e-01 -7.85181448e-02 -3.84402752e-01
5.55975616e-01 -1.35876685e-01 -5.05687296e-02 4.21094269e-01
-1.27488121e-01 1.27570534e+00 9.25603271e-01 -1.45561683e+00
-8.14301729e-01 -1.20541342e-01 7.32556209e-02 4.29303855e-01
-2.02765927e-01 -3.86397421e-01 -8.94756556e-01 -3.54253024e-01
3.89395863e-01 -7.89659619e-01 -1.88605234e-01 1.96234345e-01
-4.83952701e-01 3.55961710e-01 7.07082450e-01 -7.54295766e-01
9.94082987e-01 -2.34587693e+00 1.13140561e-01 1.56110331e-01
3.03090632e-01 -1.21901043e-01 -2.29887530e-01 8.07811469e-02
-2.69305408e-02 -5.23198806e-02 -1.00167096e-01 -4.52781200e-01
-2.22261801e-01 2.93405533e-01 -4.01326269e-02 5.43202341e-01
-1.39645696e-01 8.79384518e-01 -7.57904649e-01 -5.86570442e-01
5.58350384e-01 6.44443154e-01 -6.82853997e-01 1.87141463e-01
-1.82152480e-01 6.96982741e-01 -1.09566964e-01 2.69338399e-01
1.07741463e+00 -2.98502743e-02 -7.23403543e-02 -5.83909094e-01
-2.48687387e-01 7.64610395e-02 -7.88401842e-01 2.09083223e+00
-5.94716430e-01 7.63573170e-01 -3.14647883e-01 -5.39632320e-01
8.71914387e-01 8.44967440e-02 4.92383569e-01 -1.23899388e+00
3.60548764e-01 1.80970252e-01 -2.59370178e-01 -4.11085606e-01
6.42848492e-01 1.90662891e-01 -6.94945129e-03 2.92756826e-01
-1.81282043e-01 -4.54157025e-01 -1.53326998e-02 6.20443672e-02
7.44783044e-01 2.46600732e-01 -2.73012906e-01 8.77726972e-02
2.48617500e-01 -3.09920222e-01 6.05777502e-01 2.80130476e-01
-1.19073484e-02 9.30894196e-01 4.80782568e-01 -5.73856413e-01
-1.35096133e+00 -9.11441386e-01 6.29956350e-02 6.44539714e-01
4.69756454e-01 -1.35291964e-01 -8.08456719e-01 -3.43510538e-01
-5.40247381e-01 7.30365992e-01 -5.53733647e-01 -3.40925217e-01
-5.40797412e-01 -5.21131873e-01 4.13390219e-01 1.67133376e-01
1.23336172e+00 -8.73048484e-01 -8.73521328e-01 4.11364250e-02
-6.74001992e-01 -1.11417305e+00 -5.91685414e-01 2.25604907e-01
-8.78593624e-01 -6.92760408e-01 -7.13158667e-01 -8.96387041e-01
6.91505611e-01 2.93018430e-01 8.48271132e-01 -5.09006530e-02
-1.19636022e-01 -2.02073276e-01 -4.81549412e-01 -2.26404481e-02
-4.01847690e-01 9.92782265e-02 -4.18676496e-01 -6.07706681e-02
-9.12367776e-02 -4.94660854e-01 -1.17985690e+00 3.40408057e-01
-1.18052280e+00 6.92576349e-01 5.60140431e-01 6.45066381e-01
5.54283738e-01 1.91799596e-01 2.43703797e-01 -2.81985700e-01
3.87753807e-02 -4.15732898e-02 -7.80265212e-01 -6.35866448e-02
-6.35078192e-01 1.83001712e-01 4.66709048e-01 -1.42972469e-01
-1.11093020e+00 2.44038686e-01 -2.34450832e-01 -4.17040884e-01
1.85061559e-01 3.16808313e-01 -2.87916511e-01 -2.67321169e-01
6.17022693e-01 4.57583845e-01 -4.65556085e-02 -1.31605536e-01
6.10697269e-01 6.48953080e-01 7.14158356e-01 -2.49788851e-01
8.68253231e-01 6.81946754e-01 1.98225654e-03 -3.69293988e-01
-7.10487366e-01 3.57298106e-02 -5.38740218e-01 -3.84340793e-01
1.05360830e+00 -1.05519700e+00 -7.37750769e-01 8.68660748e-01
-1.20054269e+00 -5.89325368e-01 -7.89962485e-02 2.73449391e-01
-7.34808624e-01 5.23742020e-01 -6.72508717e-01 -2.42373168e-01
-5.09177625e-01 -1.37659371e+00 1.07939303e+00 4.14401710e-01
3.58571887e-01 -5.43281317e-01 -1.87225699e-01 3.86008292e-01
3.76561254e-01 2.09241614e-01 9.86698687e-01 3.18051189e-01
-1.03823674e+00 1.83541507e-01 -6.75769866e-01 5.88262916e-01
2.21190155e-01 -1.35183379e-01 -9.87149596e-01 -4.03095663e-01
-3.46388109e-02 -3.45857352e-01 8.86856139e-01 5.00170588e-01
1.45115864e+00 -1.25322357e-01 -6.82168081e-02 1.10533726e+00
1.67075741e+00 4.20246631e-01 1.36230302e+00 6.02039576e-01
9.86904681e-01 3.85550529e-01 5.79508424e-01 7.25957572e-01
5.28822422e-01 5.21390378e-01 8.21157575e-01 -5.25798202e-01
-3.28456402e-01 -3.00935209e-01 2.50924528e-01 5.14745414e-01
2.84916967e-01 -4.28172529e-01 -6.65263712e-01 3.22641045e-01
-1.35684204e+00 -5.48556805e-01 4.74825390e-02 2.13933039e+00
9.06298876e-01 1.55709594e-01 -4.06790584e-01 1.20946981e-01
6.44710839e-01 3.11938941e-01 -8.20709407e-01 -3.89015198e-01
-2.04129711e-01 1.65399313e-01 8.32490206e-01 6.44276798e-01
-6.36239171e-01 1.01929080e+00 5.45110130e+00 8.96740735e-01
-1.51785326e+00 -1.75530046e-01 1.10638595e+00 -6.39167503e-02
-3.63550305e-01 -3.24427336e-02 -5.16765237e-01 3.90093386e-01
9.86344635e-01 4.58450578e-02 6.61937952e-01 3.78954738e-01
5.57161629e-01 -5.46711683e-01 -1.00520420e+00 1.29455268e+00
2.41454884e-01 -1.40701640e+00 3.74608077e-02 1.25764925e-02
7.85726249e-01 1.79917738e-01 3.01377714e-01 -4.61299019e-03
3.62874866e-02 -9.32365954e-01 7.60256410e-01 3.47928345e-01
1.20514643e+00 -1.01915145e+00 4.41976130e-01 -1.17702000e-02
-1.04980707e+00 -7.58976638e-02 -2.62107819e-01 2.76063323e-01
2.93137938e-01 6.96948946e-01 -6.65085077e-01 7.53372490e-01
7.06376553e-01 7.55199194e-01 -5.55817366e-01 1.14450932e+00
-2.78712988e-01 3.14449549e-01 -2.85266459e-01 4.46062982e-01
8.39994475e-02 -9.94496569e-02 1.05096497e-01 6.85681224e-01
5.47625959e-01 5.34679554e-02 -3.83892536e-01 6.30794644e-01
-2.43702084e-01 -1.40997261e-01 -3.44697565e-01 4.30932641e-01
2.60515064e-01 1.12681723e+00 -5.65960884e-01 -2.23815069e-01
-2.30357036e-01 1.54517937e+00 9.32059810e-02 4.27370280e-01
-1.14480889e+00 -4.51510757e-01 3.73761088e-01 3.86631005e-02
3.39192361e-01 -2.29031369e-01 -4.33899462e-01 -9.69259202e-01
-1.30995452e-01 -8.07849467e-01 -7.20820017e-03 -1.39063311e+00
-7.00473845e-01 7.38780141e-01 -4.57727760e-02 -1.45132601e+00
-4.84551229e-02 -3.90161932e-01 -2.61472553e-01 8.75249445e-01
-1.89197624e+00 -1.07075167e+00 -8.47589910e-01 6.10926390e-01
5.73673189e-01 1.27145797e-01 2.49678180e-01 5.24020433e-01
-5.72662592e-01 5.13894200e-01 2.33182058e-01 3.83557938e-02
8.37783039e-01 -8.16283941e-01 5.48750222e-01 8.39910448e-01
-4.20959890e-01 -1.45056427e-01 6.10083938e-01 -4.57093537e-01
-1.38779056e+00 -1.32583725e+00 4.85167056e-01 1.51597679e-01
2.31132746e-01 -2.46993586e-01 -5.23093760e-01 4.34160769e-01
4.37193125e-01 -9.05456841e-02 5.31911366e-02 -6.62185967e-01
-2.17090204e-01 -4.24450696e-01 -1.19864714e+00 6.83743238e-01
7.78380871e-01 -3.91998678e-01 1.62731886e-01 3.15262645e-01
1.06885302e+00 -8.44836473e-01 -6.73116386e-01 2.53739893e-01
3.90759438e-01 -1.17058527e+00 8.51242781e-01 3.22617799e-01
1.09694481e+00 -3.53455186e-01 -3.36050808e-01 -1.26305592e+00
-7.87425637e-02 -5.70989907e-01 2.79710293e-01 9.57813084e-01
2.54013330e-01 -5.67458510e-01 1.03617454e+00 4.10559833e-01
-3.87166947e-01 -7.52291620e-01 -1.08920765e+00 -4.53390092e-01
1.09826364e-01 -3.14139396e-01 7.30204225e-01 5.37700355e-01
-3.80743921e-01 2.84282684e-01 -6.59064651e-01 2.49195918e-01
6.58396602e-01 -9.60916951e-02 7.92240739e-01 -6.41673803e-01
-1.75807253e-01 6.21581972e-02 -2.58966208e-01 -1.56025028e+00
-1.21038944e-01 -5.61690271e-01 1.92312792e-01 -1.69495082e+00
1.97135717e-01 -5.03063738e-01 -4.26566415e-02 5.44518948e-01
-1.68762766e-02 3.48778635e-01 9.84446481e-02 1.04414538e-01
-3.02431583e-01 8.26009095e-01 1.61451364e+00 -2.20033526e-01
-2.68575579e-01 -5.59373736e-01 -6.47977948e-01 3.23294550e-01
8.34695101e-01 -3.04773986e-01 -6.77694023e-01 -9.17261243e-01
2.90187985e-01 2.84763038e-01 3.98886919e-01 -1.13280106e+00
1.13847502e-01 -1.81348488e-01 4.47691381e-01 -1.01383066e+00
5.49390495e-01 -6.68779731e-01 5.79126477e-02 6.09181225e-01
-9.06985402e-02 4.07991447e-02 2.14157805e-01 2.51050830e-01
-1.05462559e-01 -9.20370221e-02 1.07728565e+00 1.65420145e-01
-8.45139325e-01 4.62616205e-01 -1.97149038e-01 -1.89239040e-01
9.28623915e-01 -5.02743959e-01 -6.27647340e-01 -5.69613934e-01
-1.84645250e-01 3.07118237e-01 6.51648402e-01 2.11822003e-01
9.68299150e-01 -1.28834128e+00 -5.95653594e-01 2.67628521e-01
1.83818012e-01 3.76933277e-01 5.78021884e-01 5.91101885e-01
-1.07844150e+00 6.32458106e-02 -3.98363322e-01 -6.72752976e-01
-8.54769826e-01 3.32426965e-01 7.97210991e-01 -1.39652207e-01
-6.56687975e-01 9.26264286e-01 3.46290678e-01 -1.99337497e-01
6.62560388e-02 -3.65121245e-01 1.43667594e-01 -2.35420123e-01
3.64741415e-01 1.04771212e-01 1.64505050e-01 -4.28191990e-01
-2.82404236e-02 5.55131972e-01 -2.53417224e-01 -4.24388915e-01
1.18067515e+00 -5.35602093e-01 -9.23550650e-02 6.82002679e-02
1.54326189e+00 -4.85776246e-01 -1.84820998e+00 -9.80282500e-02
-5.83759725e-01 -7.16509581e-01 4.38037127e-01 -8.64314198e-01
-1.57725370e+00 8.79162073e-01 1.12712157e+00 -4.70685184e-01
1.55837488e+00 -4.09392267e-01 1.13419735e+00 1.76485643e-01
3.56284797e-01 -1.04382074e+00 4.82532650e-01 4.95372623e-01
7.91831911e-01 -1.09756887e+00 2.72228234e-02 -2.15698346e-01
-5.82336128e-01 1.11150014e+00 7.78194010e-01 1.15241401e-01
3.70115906e-01 2.20853940e-01 2.32496157e-01 -1.22273073e-01
-4.73332196e-01 1.75802186e-01 -1.84886754e-01 5.99212170e-01
1.44300058e-01 -4.00847524e-01 -1.18332379e-01 -1.62508905e-01
-3.02162319e-01 2.03270838e-01 8.29621434e-01 8.14789355e-01
-4.44895923e-01 -9.39857483e-01 -1.88189045e-01 2.24192426e-01
-1.32768061e-02 -3.95347893e-01 1.43407777e-01 4.06360924e-01
1.90360263e-01 1.00416350e+00 3.88373397e-02 -6.62845552e-01
3.06934327e-01 -6.02787852e-01 5.82630932e-01 -2.79886752e-01
-3.24125260e-01 -9.88667132e-04 1.22313630e-02 -6.93065703e-01
-2.42356047e-01 -1.32217869e-01 -1.47167647e+00 -5.36117256e-01
-9.22611635e-03 -3.23560476e-01 9.66424286e-01 8.54589522e-01
4.47894573e-01 7.20297217e-01 1.11964869e+00 -1.16567957e+00
9.38089564e-03 -7.04138756e-01 -5.23653567e-01 3.23833749e-02
4.25415814e-01 -3.98862720e-01 -3.92270386e-01 1.93922192e-01] | [10.777402877807617, -2.1869451999664307] |
7bf9cf02-0e3a-411f-9686-ba3da3dba380 | locate-who-you-are-matching-geo-location-to | 2104.09119 | null | https://arxiv.org/abs/2104.09119v2 | https://arxiv.org/pdf/2104.09119v2.pdf | Locate Who You Are: Matching Geo-location to Text for User Identity Linkage | Nowadays, users are encouraged to activate across multiple online social networks simultaneously. Anchor link prediction, which aims to reveal the correspondence among different accounts of the same user across networks, has been regarded as a fundamental problem for user profiling, marketing, cybersecurity, and recommendation. Existing methods mainly address the prediction problem by utilizing profile, content, or structural features of users in symmetric ways. However, encouraged by online services, users would also post asymmetric information across networks, such as geo-locations and texts. It leads to an emerged challenge in aligning users with asymmetric information across networks. Instead of similarity evaluation applied in previous works, we formalize correlation between geo-locations and texts and propose a novel anchor link prediction framework for matching users across networks. Moreover, our model can alleviate the label scarcity problem by introducing external data. Experimental results on real-world datasets show that our approach outperforms existing methods and achieves state-of-the-art results. | ['Yangyang Li', 'Xueqi Cheng', 'HuaWei Shen', 'Hao Gao', 'Yongqing Wang', 'Jiangli Shao'] | 2021-04-19 | null | null | null | null | ['anchor-link-prediction'] | ['graphs'] | [-4.98362295e-02 -5.77812940e-02 -7.96833515e-01 -3.47165883e-01
-3.95926476e-01 -7.54551291e-01 6.38675332e-01 7.31275797e-01
-2.33351901e-01 4.18814063e-01 3.15387279e-01 -1.01326749e-01
-5.26261866e-01 -9.06018734e-01 -3.07305474e-02 1.34231905e-02
-3.61920297e-02 5.55009425e-01 5.30883014e-01 -2.13609099e-01
3.65192205e-01 2.51668930e-01 -9.97679234e-01 1.53356135e-01
9.32128727e-01 9.78054523e-01 -8.77356827e-02 -7.92501122e-02
-3.67160708e-01 2.73158491e-01 -1.71806410e-01 -9.52981532e-01
1.78321645e-01 -1.18195817e-01 -7.84594774e-01 -6.25680014e-03
2.42719218e-01 -1.12505406e-01 -4.15217131e-01 1.33057868e+00
3.11656743e-01 -1.09264970e-01 4.63436663e-01 -1.52009547e+00
-6.96339071e-01 1.00384462e+00 -1.02350581e+00 1.73396870e-01
6.15761101e-01 -6.51173234e-01 1.59785891e+00 -6.03343248e-01
5.79688132e-01 1.05614233e+00 9.27292883e-01 2.03148499e-02
-1.06912768e+00 -6.95497036e-01 3.04782420e-01 2.05726147e-01
-1.27899623e+00 -1.40875146e-01 8.09985101e-01 -4.69700277e-01
-6.83970302e-02 4.43075895e-01 4.04057115e-01 9.78233933e-01
-5.93163192e-01 6.98243439e-01 8.29358816e-01 -6.57601357e-02
-4.04972047e-01 4.18275446e-01 1.37950316e-01 5.28059423e-01
3.37750405e-01 -4.99533564e-01 -4.31192845e-01 -5.32893896e-01
2.15009406e-01 4.69147295e-01 -3.50852042e-01 -3.62071097e-01
-1.13580620e+00 6.89760029e-01 5.30457079e-01 5.50213218e-01
-4.41418439e-02 -6.81588352e-01 3.71293038e-01 1.13058873e-01
6.79833353e-01 3.47969770e-01 -4.74437982e-01 2.02552453e-01
-8.56005907e-01 7.76656391e-03 1.00219822e+00 1.12590194e+00
7.85951614e-01 -8.29542577e-01 -5.40573858e-02 1.00459576e+00
5.86356044e-01 1.16563544e-01 6.45192802e-01 -5.99825084e-01
8.04309130e-01 1.10498834e+00 1.97043687e-01 -1.86354291e+00
-5.39128721e-01 -7.26698637e-01 -8.52066278e-01 -9.08911586e-01
6.38314784e-01 -2.44026706e-01 1.51049912e-01 1.68958354e+00
5.61223567e-01 4.83651727e-01 -4.79801804e-01 5.45522749e-01
7.84604073e-01 2.64085889e-01 -3.13369185e-02 -3.74626428e-01
1.28781033e+00 -8.21086287e-01 -6.44958258e-01 -3.65598239e-02
7.77975202e-01 -9.17390525e-01 5.55819035e-01 -1.53014153e-01
-8.19245636e-01 -3.32856476e-01 -5.62058926e-01 2.74359018e-01
-5.10983765e-01 -6.48310184e-02 5.60953557e-01 6.03275418e-01
-7.11085618e-01 9.25232887e-01 -1.01223320e-01 -7.23674834e-01
3.65736246e-01 2.88763732e-01 -2.27941424e-01 4.83492091e-02
-1.42508841e+00 3.10256273e-01 2.62208819e-01 -4.72826697e-02
3.88837636e-01 -6.06986821e-01 -5.71234226e-01 4.69168723e-02
4.95543718e-01 -2.99691409e-01 1.01138306e+00 -8.14929128e-01
-9.42125976e-01 8.98792684e-01 -1.30272135e-01 -1.66794464e-01
6.76607370e-01 1.78317651e-01 -1.07272065e+00 -2.36723602e-01
4.24366891e-01 -4.49454635e-02 4.82996285e-01 -1.24496150e+00
-1.01669443e+00 -4.91237789e-01 1.57451794e-01 5.20736724e-02
-9.91882145e-01 1.16765894e-01 -7.86024988e-01 -6.90250278e-01
5.01245558e-01 -1.03099144e+00 2.19825655e-02 -2.06395373e-01
-7.65934348e-01 -5.08539021e-01 7.96588361e-01 -6.63978338e-01
1.65233159e+00 -1.92719555e+00 -2.61394642e-02 9.08060908e-01
5.63564599e-01 1.29044622e-01 7.67533332e-02 6.02446616e-01
1.14946976e-01 3.90487581e-01 2.32517540e-01 -4.21135694e-01
1.56733751e-01 -1.10324457e-01 -2.31255218e-01 5.12379944e-01
-4.88778949e-01 7.13336706e-01 -1.19705760e+00 -8.60073149e-01
-3.98825586e-01 4.88621071e-02 -3.55556160e-01 -9.02171433e-02
1.53544784e-01 4.59385574e-01 -6.97101593e-01 7.16919661e-01
6.60952747e-01 -6.75745130e-01 7.83543944e-01 -3.38110834e-01
6.85753673e-02 2.77586788e-01 -1.11450946e+00 1.37947524e+00
-1.93677649e-01 3.69941741e-01 1.05280071e-01 -9.55758631e-01
8.37520361e-01 1.81611791e-01 8.53435576e-01 -5.84004343e-01
2.02266634e-01 3.40489417e-01 -1.46575764e-01 -4.23312843e-01
5.06597757e-01 5.07057250e-01 -1.88973665e-01 6.71367764e-01
-4.74660635e-01 8.06397915e-01 3.77944589e-01 4.56533015e-01
7.63220370e-01 -2.17543080e-01 1.50997564e-01 -6.89470172e-02
7.50798821e-01 -3.25334042e-01 5.46025395e-01 5.54826260e-01
-1.81373999e-01 2.39979491e-01 5.70358634e-01 -4.49798554e-02
-9.39709783e-01 -5.70758700e-01 -3.50530520e-02 1.29267216e+00
5.36182761e-01 -6.37556255e-01 -6.67398274e-01 -1.03049159e+00
3.39549243e-01 4.60399874e-02 -5.22821665e-01 2.56679863e-01
-4.87353146e-01 -8.14182937e-01 3.73314112e-01 1.59981459e-01
4.08859581e-01 -3.70618880e-01 7.28912652e-01 3.17531705e-01
-7.14742064e-01 -1.46180129e+00 -9.27433670e-01 -8.13473523e-01
-7.04068184e-01 -1.21841621e+00 -6.43365145e-01 -7.57702947e-01
6.59073353e-01 7.14429855e-01 9.77561116e-01 5.95957875e-01
2.11946994e-01 6.71068430e-02 -4.39443469e-01 1.72782093e-01
2.65686270e-02 7.40294993e-01 2.50487000e-01 7.15811789e-01
4.26742077e-01 -9.27920222e-01 -7.64981270e-01 9.89850640e-01
-7.11854160e-01 -1.83730707e-01 2.96233058e-01 3.88009608e-01
1.33521378e-01 9.98728499e-02 8.37158620e-01 -1.52660084e+00
7.67672241e-01 -1.13605213e+00 -2.78523266e-01 3.26938808e-01
-8.62322748e-01 -3.07368547e-01 4.90692586e-01 -4.80321229e-01
-6.42732859e-01 -3.04346234e-01 2.83120889e-02 3.58324140e-01
2.72401899e-01 6.38723254e-01 -2.66337693e-01 -1.92552239e-01
2.89191127e-01 -9.11344364e-02 -1.80972829e-01 -7.59275079e-01
1.77475676e-01 9.75363076e-01 1.59441352e-01 -3.41554344e-01
1.13569391e+00 5.97607136e-01 1.88953169e-02 -2.78594822e-01
-1.35587001e+00 -1.09354889e+00 -8.03716838e-01 -1.31798953e-01
1.90965220e-01 -6.87801003e-01 -1.16145790e+00 2.16491893e-01
-8.29175115e-01 4.89254534e-01 3.69921505e-01 2.28248522e-01
1.01648800e-01 8.45258772e-01 -5.52884579e-01 -5.12206256e-01
-9.75509137e-02 -6.81243718e-01 5.16165078e-01 2.34534562e-01
-5.00506878e-01 -1.37815559e+00 2.56852936e-02 7.76665151e-01
3.86966377e-01 1.24324262e-01 8.01047683e-01 -1.19504631e+00
-4.44037646e-01 -6.69695020e-01 -8.07860553e-01 -4.31300372e-01
4.66219991e-01 -2.43295565e-01 -5.67813754e-01 -2.49414697e-01
-5.76951683e-01 1.78997859e-01 4.48596418e-01 -3.20361167e-01
1.32557559e+00 -7.40145326e-01 -8.65227044e-01 3.88280898e-01
1.08356452e+00 -3.71734232e-01 1.59639984e-01 4.21955049e-01
9.34288204e-01 1.04281259e+00 5.43896556e-01 7.86549270e-01
8.62257719e-01 1.00024915e+00 4.67028260e-01 -2.38836519e-02
3.79547715e-01 -6.45755172e-01 -7.15523586e-02 8.21699858e-01
-2.00871043e-02 -3.71721447e-01 -5.94307363e-01 5.25322855e-01
-2.12132430e+00 -1.09938633e+00 -5.06752610e-01 2.08137560e+00
7.08655417e-01 2.34415159e-01 6.64405644e-01 1.39309242e-01
1.22223985e+00 3.34647521e-02 -4.06050265e-01 4.08363700e-01
-4.21168804e-02 -5.05746186e-01 7.30412900e-01 3.94300759e-01
-8.52488875e-01 4.71389890e-01 5.26661491e+00 8.77004325e-01
-6.79992974e-01 2.86401242e-01 5.21125972e-01 1.50090352e-01
-4.39594358e-01 -2.33422406e-03 -1.10364008e+00 1.05836380e+00
6.60568476e-01 -2.79108077e-01 3.87896806e-01 5.94046652e-01
1.13417253e-01 3.74233663e-01 -9.21226323e-01 7.92773962e-01
8.04786920e-04 -1.24120235e+00 -1.54833570e-01 4.18510586e-01
8.77258956e-01 -1.67285159e-01 1.44135326e-01 1.91933624e-02
2.57272333e-01 -3.97784412e-01 3.47460777e-01 3.86569709e-01
3.95777434e-01 -6.94492042e-01 7.28784859e-01 4.56909329e-01
-1.52553487e+00 -2.36759007e-01 -3.75876203e-02 2.33529299e-01
2.83214658e-01 7.05896318e-01 -6.38727725e-01 8.34775805e-01
4.47128922e-01 1.17263877e+00 -6.12844586e-01 1.17846644e+00
-4.57965885e-04 4.71245319e-01 -2.30268747e-01 -1.12216733e-01
2.88284659e-01 -4.40475374e-01 3.20045084e-01 1.01053286e+00
5.51708579e-01 -1.69718161e-01 4.07669067e-01 3.79577845e-01
-6.30251467e-01 5.68886340e-01 -4.85170066e-01 -8.30664299e-03
8.23268890e-01 1.67883623e+00 -5.68289161e-01 -1.14401981e-01
-6.49041355e-01 8.61161649e-01 3.62724870e-01 1.62324890e-01
-6.66360974e-01 -3.82434309e-01 5.50889432e-01 7.60554314e-01
-1.89775899e-01 -8.76236483e-02 4.20482010e-02 -1.06300521e+00
5.73486760e-02 -4.65074867e-01 6.58386767e-01 -3.32117379e-01
-1.81430256e+00 3.56019557e-01 -3.47808927e-01 -1.55418384e+00
-1.01914341e-02 -3.88413757e-01 -5.30485570e-01 6.87157273e-01
-1.79865479e+00 -1.26494610e+00 -4.29425776e-01 6.80417657e-01
-2.27548797e-02 -1.84258625e-01 4.48559135e-01 8.57775927e-01
-6.90442860e-01 8.85931194e-01 5.06554365e-01 5.15721738e-01
8.83513153e-01 -1.05586529e+00 3.51698965e-01 4.25357759e-01
8.67555216e-02 7.62456834e-01 4.05801654e-01 -7.94671953e-01
-9.75636721e-01 -1.23260534e+00 1.36639655e+00 -4.34929490e-01
1.28756940e+00 -9.93720815e-02 -8.55164707e-01 6.00344419e-01
-2.36877371e-02 -1.18251123e-01 1.28598607e+00 5.10419428e-01
-4.16201890e-01 -2.79185653e-01 -1.10530055e+00 5.60008287e-01
1.42799556e+00 -5.13905048e-01 4.66537587e-02 6.98549449e-01
4.77237195e-01 -1.99774742e-01 -1.13942909e+00 1.54277205e-01
5.94933331e-01 -6.25980616e-01 1.01990640e+00 -5.94969392e-01
1.11424230e-01 -1.60096318e-01 1.35305017e-01 -1.17900872e+00
-4.80998099e-01 -8.98061931e-01 -2.40225151e-01 1.92418027e+00
6.90940201e-01 -8.74528468e-01 1.01546955e+00 6.38283134e-01
3.25235486e-01 -5.55913985e-01 -4.83659953e-01 -5.46876013e-01
-5.18251121e-01 1.03241585e-01 9.74755704e-01 1.44776857e+00
4.98581439e-01 4.12584782e-01 -5.01683712e-01 3.25185597e-01
8.13971877e-01 4.17619824e-01 5.46664357e-01 -2.08798099e+00
-1.31048545e-01 -6.01466298e-01 -2.03747258e-01 -1.22090137e+00
3.91294062e-01 -1.06680632e+00 -6.11840785e-01 -1.47371936e+00
3.25488627e-01 -1.09116256e+00 -2.84094661e-01 9.91846621e-02
4.93653417e-02 4.62891340e-01 -1.28541574e-01 7.17611969e-01
-9.15080905e-01 6.27010241e-02 1.07355380e+00 -9.56663787e-02
-6.02198206e-02 5.58606446e-01 -1.06265640e+00 8.64728749e-01
7.67949760e-01 -5.08606970e-01 -2.95774519e-01 -1.75317049e-01
9.05999303e-01 -7.32432213e-03 -3.15489620e-02 -5.81118584e-01
5.58179200e-01 -1.71727628e-01 1.55288383e-01 -6.12941623e-01
1.84216827e-01 -1.22040045e+00 -4.26491126e-02 1.46216303e-01
-3.00355762e-01 2.37310484e-01 -5.28578341e-01 1.03401220e+00
-6.84727579e-02 -3.60870600e-01 4.68903631e-01 -5.16015738e-02
-3.27785432e-01 8.97408545e-01 7.57089630e-02 1.90579757e-01
8.38471413e-01 -8.34393352e-02 -4.65546042e-01 -4.35155809e-01
-6.70990646e-01 6.63457096e-01 4.79579657e-01 5.09141743e-01
1.91420808e-01 -1.56022048e+00 -5.58550775e-01 -2.44057402e-01
4.25331712e-01 -5.89293778e-01 -8.43785424e-03 9.11766648e-01
2.10853115e-01 2.81178951e-01 7.36175627e-02 -2.94621736e-01
-1.48218191e+00 4.50506628e-01 -3.87859158e-02 -4.82202619e-01
-7.34512135e-02 5.23744643e-01 -3.86824042e-01 -5.27386904e-01
1.80201173e-01 4.74138200e-01 -8.16622198e-01 7.93783069e-01
4.15644705e-01 4.65964526e-01 -3.22271958e-02 -9.28927541e-01
-1.19363233e-01 5.63103378e-01 -3.08553994e-01 3.08163524e-01
1.25726664e+00 -7.07168639e-01 -3.24705064e-01 2.60100931e-01
1.27379251e+00 4.19243932e-01 -5.77003539e-01 -1.09173393e+00
5.45646787e-01 -6.96637034e-01 -3.53699774e-01 -1.78185582e-01
-1.22568715e+00 3.59258652e-01 1.89116243e-02 9.98855174e-01
6.61575496e-01 4.15967628e-02 1.24265492e+00 1.16510577e-01
3.06806952e-01 -1.06680155e+00 1.13015279e-01 2.96038449e-01
1.94329262e-01 -1.41512728e+00 -3.17494944e-02 -8.44133198e-01
-3.55007589e-01 7.88538516e-01 5.36381006e-01 4.24986273e-01
1.07199144e+00 -4.32578593e-01 -4.21424866e-01 -1.05997667e-01
-1.04972377e-01 -3.17619383e-01 4.96682525e-01 3.17450553e-01
5.05053520e-01 -5.05308025e-02 -3.80805284e-01 5.07947803e-01
-4.46879007e-02 -4.34549391e-01 2.92369395e-01 6.35843396e-01
-3.02146584e-01 -1.63380158e+00 2.09559537e-02 6.78216755e-01
-7.91629970e-01 -1.02273405e-01 -4.81872946e-01 4.26546186e-01
7.45759718e-03 9.67850268e-01 -7.00449198e-02 -5.06857693e-01
1.75891653e-01 -2.07297117e-01 -1.15494512e-01 -6.11511886e-01
-6.14735007e-01 -2.67998129e-01 2.38224611e-01 -1.79984614e-01
-5.73628664e-01 -8.24022651e-01 -8.37596953e-01 -9.07499552e-01
-6.22856438e-01 4.74959463e-01 5.27638912e-01 9.41355944e-01
5.50231040e-01 1.01963960e-01 1.23942173e+00 -4.31183308e-01
-3.39481145e-01 -8.46984446e-01 -7.51655459e-01 7.71349192e-01
1.41819373e-01 -6.53229773e-01 -3.12456608e-01 -2.26709679e-01] | [7.443170547485352, 6.284997940063477] |
bb0516b5-9673-4e40-8ade-b9b081de1008 | improving-long-context-document-level-machine | 2306.05183 | null | https://arxiv.org/abs/2306.05183v1 | https://arxiv.org/pdf/2306.05183v1.pdf | Improving Long Context Document-Level Machine Translation | Document-level context for neural machine translation (NMT) is crucial to improve the translation consistency and cohesion, the translation of ambiguous inputs, as well as several other linguistic phenomena. Many works have been published on the topic of document-level NMT, but most restrict the system to only local context, typically including just the one or two preceding sentences as additional information. This might be enough to resolve some ambiguous inputs, but it is probably not sufficient to capture some document-level information like the topic or style of a conversation. When increasing the context size beyond just the local context, there are two challenges: (i) the~memory usage increases exponentially (ii) the translation performance starts to degrade. We argue that the widely-used attention mechanism is responsible for both issues. Therefore, we propose a constrained attention variant that focuses the attention on the most relevant parts of the sequence, while simultaneously reducing the memory consumption. For evaluation, we utilize targeted test sets in combination with novel evaluation techniques to analyze the translations in regards to specific discourse-related phenomena. We find that our approach is a good compromise between sentence-level NMT vs attending to the full context, especially in low resource scenarios. | ['Hermann Ney', 'Christian Herold'] | 2023-06-08 | null | null | null | null | ['nmt'] | ['computer-code'] | [ 4.16345417e-01 -1.35252282e-01 -2.68408924e-01 -2.46225432e-01
-9.39349174e-01 -7.13638663e-01 7.24344075e-01 3.47599089e-01
-5.21903396e-01 9.03316438e-01 4.37725514e-01 -6.70799434e-01
1.37682080e-01 -4.65209514e-01 -6.24501586e-01 -6.48775697e-01
4.86396432e-01 5.24735689e-01 2.00402915e-01 -4.75247145e-01
4.05776232e-01 3.08881253e-01 -1.34130883e+00 4.16827828e-01
1.26000857e+00 5.47898412e-01 6.66483641e-01 4.88899142e-01
-5.19453466e-01 4.82824117e-01 -8.44347179e-01 -5.02462804e-01
-7.84086715e-03 -8.27986717e-01 -9.67258334e-01 -6.17339090e-02
3.80851060e-01 -3.16142924e-02 3.52192253e-01 1.15590739e+00
5.52061617e-01 8.41475502e-02 2.99436301e-01 -6.62296772e-01
-4.41833436e-01 8.68437827e-01 -4.43872154e-01 6.25208318e-01
1.65698290e-01 5.09985462e-02 9.43228185e-01 -8.25509071e-01
6.71538413e-01 1.25068998e+00 2.49826223e-01 4.83440131e-01
-1.05936635e+00 -1.85996950e-01 4.33306187e-01 3.38316381e-01
-1.05970716e+00 -5.50119162e-01 7.23519087e-01 -2.51726806e-01
1.21603775e+00 5.17423809e-01 3.29676896e-01 1.34747541e+00
4.09176648e-01 5.57066858e-01 1.10511470e+00 -8.00928354e-01
1.27817199e-01 1.89295307e-01 2.84519643e-01 8.72554407e-02
1.60064742e-01 -3.10976028e-01 -3.62938911e-01 -6.43982738e-02
3.12837273e-01 -4.85233098e-01 -3.83830637e-01 1.90299064e-01
-1.39084816e+00 7.74313331e-01 1.21264003e-01 9.55737472e-01
-4.77190256e-01 -1.10301830e-01 5.85893333e-01 5.63743114e-01
5.08694589e-01 8.62269044e-01 -5.15604198e-01 -5.73573828e-01
-8.96926701e-01 4.41971309e-02 6.68687761e-01 8.12010705e-01
6.79726899e-01 -1.44979209e-01 -3.67902130e-01 8.61454606e-01
-8.56179669e-02 4.04745817e-01 6.15762532e-01 -6.39884114e-01
1.02084148e+00 5.87876439e-01 7.92795271e-02 -8.31441104e-01
-2.24721178e-01 -5.56289315e-01 -6.02979660e-01 -2.43880689e-01
5.56536257e-01 -3.47895682e-01 -6.46462739e-01 2.09667420e+00
1.36955559e-01 -5.09455621e-01 -1.64914038e-02 1.12102795e+00
5.07555962e-01 7.97073185e-01 -3.25417668e-02 -6.31296039e-01
1.53590477e+00 -9.29961145e-01 -1.05712903e+00 -6.24049962e-01
6.75330937e-01 -1.20989537e+00 1.39933717e+00 -6.28956780e-02
-1.29061067e+00 -5.38554192e-01 -1.01927555e+00 -9.99120548e-02
-2.83311009e-01 1.89748898e-01 2.57418036e-01 5.08337259e-01
-9.65702534e-01 5.74198544e-01 -7.07729995e-01 -6.32512271e-01
-2.97442555e-01 3.66780430e-01 -1.27780080e-01 -2.31569465e-02
-1.51293743e+00 1.28197837e+00 1.17615134e-01 2.80937433e-01
-1.41985044e-01 -1.57577515e-01 -6.06207728e-01 2.45713621e-01
4.71523672e-01 -7.12357759e-01 1.09080517e+00 -1.31511199e+00
-1.58796251e+00 6.06177747e-01 -4.90459591e-01 -3.11860204e-01
5.78593075e-01 -2.70283192e-01 -6.01954237e-02 -5.18439151e-02
1.73659325e-02 4.44786608e-01 6.23212397e-01 -9.47985113e-01
-3.08289587e-01 -3.39304447e-01 2.00080618e-01 6.12947226e-01
-3.96506548e-01 3.73224676e-01 -5.40545940e-01 -6.31997406e-01
-7.68566057e-02 -1.04717648e+00 -3.57159562e-02 -7.05557644e-01
-2.93884933e-01 -2.27885082e-01 7.19479442e-01 -7.22832084e-01
1.48057127e+00 -1.98218405e+00 4.57673073e-01 -1.01391420e-01
-3.95124584e-01 4.36117083e-01 -2.38658711e-01 6.24520957e-01
1.69360146e-01 2.63648301e-01 -1.83830678e-01 -4.04579014e-01
-2.81693518e-01 1.83457047e-01 -3.35840434e-01 8.78925472e-02
4.68729377e-01 1.04833400e+00 -6.28464043e-01 -3.54957283e-01
-1.99899882e-01 2.75152087e-01 -4.44710582e-01 1.39510604e-02
-3.78742963e-01 6.46414876e-01 -2.82696664e-01 2.16401950e-01
3.68937701e-01 -1.25791132e-01 4.08639550e-01 1.26100168e-01
-1.58807918e-01 8.63168657e-01 -9.60967839e-01 1.64704108e+00
-4.53395784e-01 7.87684083e-01 6.40252009e-02 -8.87125611e-01
7.76522696e-01 4.84688461e-01 -1.23532703e-02 -8.99123311e-01
2.53513306e-01 3.30340266e-01 5.86157084e-01 -4.92336661e-01
6.17047191e-01 -1.51929796e-01 5.19348076e-03 6.62013590e-01
-2.99043238e-01 1.45678565e-01 3.42514485e-01 -2.53321379e-02
8.09552252e-01 9.08271447e-02 3.66802007e-01 -5.34604549e-01
5.58113933e-01 6.55628517e-02 6.39777243e-01 6.03604376e-01
-1.28539890e-01 6.58297062e-01 7.04264581e-01 -1.03579909e-01
-1.08903801e+00 -3.27365130e-01 1.66513667e-01 1.27332997e+00
1.68021828e-01 -4.92394030e-01 -1.12587011e+00 -5.68390012e-01
-5.33166289e-01 7.68043160e-01 -5.06964326e-01 -1.45817861e-01
-1.13205600e+00 -8.14077735e-01 2.66463727e-01 4.39991862e-01
2.50122696e-01 -1.09746385e+00 -8.12793553e-01 3.13698292e-01
-7.23728895e-01 -1.20763600e+00 -7.65012681e-01 3.09905320e-01
-1.06688714e+00 -6.12489700e-01 -6.37161553e-01 -6.76784039e-01
4.43793148e-01 3.25488836e-01 1.14208448e+00 2.75710255e-01
3.74297023e-01 -2.58633226e-01 -5.10489285e-01 -1.97612643e-01
-7.45133281e-01 5.01081467e-01 -7.93989375e-02 -8.79935622e-02
2.95117646e-01 -5.16698778e-01 -3.03615242e-01 3.79793912e-01
-6.56094968e-01 1.91269428e-01 7.13755846e-01 8.85963678e-01
2.52686828e-01 -2.43849710e-01 6.25503182e-01 -8.38882208e-01
9.90626395e-01 -1.72698513e-01 -2.38512188e-01 3.46602678e-01
-3.79113823e-01 1.59115553e-01 8.13823640e-01 -6.99517787e-01
-9.08599138e-01 -4.37694132e-01 -1.62029371e-01 -1.04466148e-01
-8.06500614e-02 6.63754284e-01 -3.69255245e-01 4.50216025e-01
6.44548714e-01 2.09762529e-01 -1.00771025e-01 -4.54697847e-01
7.48594180e-02 7.53260314e-01 1.32611021e-01 -8.02626491e-01
4.68308449e-01 -7.78307021e-02 -3.20094585e-01 -7.20038950e-01
-5.81538320e-01 -1.86962202e-01 -6.59918189e-01 1.83091044e-01
7.82898724e-01 -4.93032187e-01 -2.77949542e-01 -1.05090328e-01
-1.52417183e+00 -3.67930382e-01 4.04086225e-02 5.29460549e-01
-4.34978336e-01 4.11000550e-01 -7.36756742e-01 -6.19573176e-01
-4.60806429e-01 -1.65585232e+00 9.90018368e-01 -2.75473483e-02
-5.94668746e-01 -7.87037730e-01 -9.26896837e-03 3.26521248e-01
6.28328919e-01 -7.29872063e-02 1.41801977e+00 -8.71051610e-01
-4.94363099e-01 1.01786770e-01 -1.74728721e-01 1.82046503e-01
1.48730829e-01 -1.60819560e-01 -8.26249123e-01 -2.62296200e-01
2.46291891e-01 -4.16301228e-02 5.65332949e-01 1.94155276e-01
5.99844038e-01 -3.54582995e-01 -9.43570882e-02 3.64135206e-02
1.02988255e+00 4.17218655e-01 6.28115714e-01 3.23567927e-01
5.38740993e-01 7.95495272e-01 7.14087784e-01 -5.03775440e-02
2.25312486e-01 1.15316713e+00 1.62068605e-01 -2.46465057e-02
-2.56238699e-01 3.67746316e-02 6.51310980e-01 1.19579852e+00
-1.20177276e-01 -4.24615681e-01 -8.62998903e-01 5.14662564e-01
-1.89882243e+00 -8.76900852e-01 -8.30839351e-02 2.21772099e+00
1.00597513e+00 3.61329019e-01 3.60781588e-02 1.15543202e-01
8.54220510e-01 2.15242393e-02 -2.65595555e-01 -9.34187651e-01
-5.88737912e-02 -1.00858994e-01 3.97496782e-02 7.09737062e-01
-7.23838210e-01 9.81808543e-01 5.91875744e+00 7.21412122e-01
-1.44437623e+00 2.71876067e-01 5.62652588e-01 -9.57467481e-02
-4.79734927e-01 6.40034601e-02 -8.18599582e-01 6.91628098e-01
1.08595669e+00 3.07958145e-02 4.06484008e-01 5.21686494e-01
4.02048558e-01 -1.21643811e-01 -1.27861416e+00 6.58779800e-01
1.95883796e-01 -1.04046965e+00 6.13128096e-02 2.37188399e-01
6.13198578e-01 1.68499723e-02 -1.65784970e-01 3.15375835e-01
-3.92994851e-01 -7.47100413e-01 7.40503192e-01 4.65386882e-02
6.25586569e-01 -7.33785152e-01 1.01083291e+00 6.22208595e-01
-8.28164279e-01 1.91152468e-01 -2.20086351e-01 -1.59448177e-01
3.38455997e-02 4.49079067e-01 -8.76280487e-01 5.90136528e-01
3.57666135e-01 2.64930695e-01 -6.43874347e-01 6.52215540e-01
-3.07482898e-01 6.30863309e-01 -2.57387370e-01 -3.20621401e-01
3.83982897e-01 -1.47252768e-01 8.34990621e-01 1.37133145e+00
3.42687190e-01 -1.07058056e-01 4.78830412e-02 6.67275071e-01
-5.10172993e-02 4.87869591e-01 -4.85089004e-01 -9.10030603e-02
3.45258355e-01 9.64207947e-01 -8.63112032e-01 -3.19065928e-01
-3.45431894e-01 1.11343396e+00 4.82029915e-01 3.50105941e-01
-7.28581369e-01 -2.23875493e-01 6.40542865e-01 -1.36307313e-03
2.52193749e-01 -3.22588265e-01 -5.28362095e-01 -1.25676501e+00
4.75044399e-01 -1.23735213e+00 6.57879710e-02 -5.22909880e-01
-8.47021341e-01 9.08508956e-01 -1.53837427e-01 -1.13992572e+00
-3.87068033e-01 -3.81493270e-01 -6.73112810e-01 1.04812086e+00
-1.53505445e+00 -8.30655158e-01 1.49261191e-01 2.00426266e-01
1.00595176e+00 1.55729264e-01 8.03161323e-01 3.70039552e-01
-6.15808487e-01 7.07557917e-01 -6.54544234e-02 -8.53474885e-02
8.89821053e-01 -1.10481954e+00 4.20442194e-01 1.17998099e+00
2.49149337e-01 1.04084623e+00 9.09525990e-01 -5.52660406e-01
-1.37815714e+00 -5.84355295e-01 1.61666417e+00 -5.31154752e-01
3.99331003e-01 -5.59679091e-01 -1.15845919e+00 3.69058132e-01
5.44471085e-01 -4.82538313e-01 5.32189608e-01 4.05233979e-01
-6.19621016e-02 6.44902438e-02 -7.38623440e-01 1.01597333e+00
8.65545034e-01 -5.03605545e-01 -8.76009226e-01 1.60009310e-01
9.19400990e-01 -4.53586519e-01 -6.30336285e-01 3.65453005e-01
2.79863924e-01 -7.23208249e-01 4.50108767e-01 -4.78553414e-01
6.14650369e-01 -3.37407261e-01 -1.58610046e-01 -1.36624432e+00
-2.63341635e-01 -7.14633763e-01 3.19500528e-02 1.33758128e+00
7.38748729e-01 -5.37646830e-01 3.80766124e-01 4.43601996e-01
-2.13706613e-01 -8.54617953e-01 -1.07721329e+00 -6.30975962e-01
2.97559321e-01 -3.04727763e-01 6.27056181e-01 9.76969600e-01
2.06561357e-01 9.17262435e-01 -3.30453455e-01 -3.86382155e-02
-2.43599474e-01 4.64704156e-01 5.88760853e-01 -9.29383755e-01
-3.67864609e-01 -7.37787724e-01 5.48610166e-02 -9.78007078e-01
2.48584785e-02 -7.89287150e-01 2.12175295e-01 -1.47805297e+00
2.40334287e-01 -2.57091403e-01 -1.35777429e-01 2.39949092e-01
-5.80783308e-01 6.47689104e-02 3.87488335e-01 3.69115740e-01
-3.42339277e-01 3.90667051e-01 1.33038509e+00 2.20029801e-02
-3.66276979e-01 -2.79571507e-02 -7.66380191e-01 3.12685966e-01
7.50848293e-01 -4.04763192e-01 -3.26912016e-01 -8.16520989e-01
2.32149556e-01 5.67587689e-02 -2.11293101e-01 -5.68100274e-01
1.87606782e-01 -3.81142199e-01 1.22009143e-02 -4.52551335e-01
1.87067747e-01 -8.26543152e-01 -1.07074484e-01 3.63377422e-01
-5.20557106e-01 4.61491376e-01 2.71550834e-01 2.62771040e-01
-2.77085811e-01 -3.48353118e-01 6.40284300e-01 -1.97541073e-01
-2.72625804e-01 -2.47300625e-01 -5.71386337e-01 5.73988035e-02
6.52403653e-01 -1.87170163e-01 -4.32209492e-01 -4.14143294e-01
-5.32712162e-01 8.55777636e-02 6.20200455e-01 7.51009405e-01
8.26825500e-02 -1.18400788e+00 -7.64123976e-01 1.27255738e-01
1.02315277e-01 -2.73816109e-01 -9.85560566e-02 1.16226566e+00
-1.95875227e-01 6.55629873e-01 -9.53726545e-02 -5.83493292e-01
-1.37174869e+00 5.73652148e-01 1.74368232e-01 -4.73872721e-01
-4.97123271e-01 5.22583008e-01 5.18042520e-02 -1.34770378e-01
1.65195167e-01 -5.29850245e-01 -2.74422169e-01 2.71780282e-01
6.21271551e-01 3.88678871e-02 5.06755710e-01 -6.50386691e-01
-4.46565747e-01 5.63715339e-01 -1.99728623e-01 -2.91039854e-01
9.20385361e-01 -4.13251460e-01 -2.30047002e-01 5.83507180e-01
9.03322220e-01 2.00208724e-01 -8.23864043e-01 -2.28548229e-01
2.70731598e-01 -2.72679418e-01 -1.85002908e-01 -9.19946611e-01
-5.07445335e-01 1.14125037e+00 1.95142105e-01 2.75391698e-01
1.13164783e+00 -1.46829814e-01 7.67780840e-01 3.52012873e-01
3.00840020e-01 -1.18898666e+00 -5.89354485e-02 8.31706822e-01
9.79414642e-01 -1.20913494e+00 -1.49424449e-01 -3.46647769e-01
-7.06799567e-01 1.11380744e+00 5.68914950e-01 2.41514653e-01
-5.28712384e-02 1.95708871e-01 2.40386039e-01 2.45821089e-01
-1.01861525e+00 -1.15697265e-01 3.19298953e-01 1.49418160e-01
9.04289842e-01 -1.22950472e-01 -8.11843097e-01 4.81405586e-01
-2.49048874e-01 -4.42340016e-01 4.20902371e-01 7.88583696e-01
-4.39283013e-01 -1.45760739e+00 -4.84922409e-01 8.94389823e-02
-7.04294145e-01 -2.72646487e-01 -6.81141973e-01 6.36642039e-01
9.20613557e-02 1.11919332e+00 1.22773737e-01 -1.49858311e-01
3.77548546e-01 3.66138577e-01 3.88419390e-01 -7.40059495e-01
-9.82152164e-01 3.58491868e-01 3.06352735e-01 -4.01537806e-01
-2.84761339e-01 -7.67131984e-01 -9.42208886e-01 -3.12513322e-01
-5.00274539e-01 2.39648387e-01 7.69684076e-01 1.25162983e+00
4.10340577e-01 5.20980954e-01 4.44020748e-01 -6.50521994e-01
-4.56033111e-01 -1.27811944e+00 2.40427226e-01 3.14782709e-01
3.19063365e-01 -4.19822216e-01 -2.37170368e-01 -2.11386293e-01] | [11.5418119430542, 10.061893463134766] |
ed133690-186c-4624-9305-e4fe2d9fbe51 | video-compression-with-arbitrary-rescaling | 2306.04202 | null | https://arxiv.org/abs/2306.04202v1 | https://arxiv.org/pdf/2306.04202v1.pdf | Video Compression with Arbitrary Rescaling Network | Most video platforms provide video streaming services with different qualities, and the quality of the services is usually adjusted by the resolution of the videos. So high-resolution videos need to be downsampled for compression. In order to solve the problem of video coding at different resolutions, we propose a rate-guided arbitrary rescaling network (RARN) for video resizing before encoding. To help the RARN be compatible with standard codecs and generate compression-friendly results, an iteratively optimized transformer-based virtual codec (TVC) is introduced to simulate the key components of video encoding and perform bitrate estimation. By iteratively training the TVC and the RARN, we achieved 5%-29% BD-Rate reduction anchored by linear interpolation under different encoding configurations and resolutions, exceeding the previous methods on most test videos. Furthermore, the lightweight RARN structure can process FHD (1080p) content at real-time speed (91 FPS) and obtain a considerable rate reduction. | ['Li Zhang', 'Junlin Li', 'Hao Jiang', 'Shijie Zhao', 'Mengxi Guo'] | 2023-06-07 | null | null | null | null | ['video-compression'] | ['computer-vision'] | [ 3.30107868e-01 -3.16444993e-01 -3.99314106e-01 -1.29589081e-01
-3.63316685e-01 -1.79318264e-01 9.96676311e-02 -3.70990157e-01
-1.41942337e-01 3.80152941e-01 4.03503507e-01 -3.73944998e-01
1.48852691e-01 -8.72161508e-01 -5.44444680e-01 -5.76543510e-01
-3.43754917e-01 -2.28846610e-01 5.80228746e-01 -2.70096928e-01
3.02376986e-01 3.96248430e-01 -1.68888915e+00 7.05268919e-01
7.69951522e-01 1.32910299e+00 6.83108389e-01 7.96374619e-01
-6.39847592e-02 9.66700792e-01 -5.06060839e-01 -3.18118751e-01
4.76872593e-01 -5.07938623e-01 -5.10522902e-01 1.59741297e-01
-1.21767148e-01 -9.28515792e-01 -6.46695256e-01 1.00626349e+00
4.30322766e-01 -1.07895076e-01 3.64549726e-01 -7.90919244e-01
-3.11256409e-01 7.72145271e-01 -6.52801991e-01 3.91230911e-01
5.15573502e-01 -3.22787017e-02 4.88136560e-01 -5.36054969e-01
3.96299571e-01 1.23006082e+00 3.44695270e-01 6.03709400e-01
-1.02193391e+00 -7.81786382e-01 -2.69366980e-01 4.20348287e-01
-1.57866240e+00 -7.26557314e-01 5.45308650e-01 -3.27169057e-03
6.65515065e-01 3.46229941e-01 7.62176692e-01 7.66103208e-01
1.24227583e-01 2.23139778e-01 3.34532768e-01 -1.82758078e-01
1.12826891e-01 -2.54065424e-01 -7.75930583e-01 1.90045938e-01
1.15367305e-02 -1.61326230e-01 -3.52319449e-01 2.10030481e-01
1.46970510e+00 -2.04786912e-01 -8.48234713e-01 3.60663235e-02
-1.26087272e+00 4.20616537e-01 1.44181490e-01 1.44507542e-01
-3.18252325e-01 3.15385222e-01 8.10580909e-01 3.58341634e-01
2.88334340e-01 3.19467601e-03 -2.53096342e-01 -4.70356226e-01
-1.08384776e+00 -4.71307747e-02 4.97213423e-01 1.42679381e+00
2.61456698e-01 4.03347701e-01 -2.19110504e-01 1.14216554e+00
2.44481087e-01 4.57034379e-01 5.99023521e-01 -1.55192709e+00
9.19907928e-01 -1.66160595e-02 2.71256901e-02 -1.03601575e+00
1.06452480e-01 -1.10548243e-01 -1.22383666e+00 -7.34129548e-02
5.43242693e-02 2.17098445e-01 -6.19516015e-01 1.20716977e+00
-7.01968670e-02 2.82628357e-01 6.74451217e-02 1.04666185e+00
5.59093058e-01 1.27179587e+00 -1.52263165e-01 -9.44318295e-01
1.33580625e+00 -7.88218617e-01 -9.46495175e-01 3.54581743e-01
4.39628571e-01 -8.92866731e-01 9.42503035e-01 5.06322503e-01
-1.61867309e+00 -9.17112350e-01 -1.27115095e+00 -2.64643133e-02
6.00406408e-01 -1.87677950e-01 4.86077145e-02 5.14685452e-01
-1.09596419e+00 6.19304061e-01 -5.30123770e-01 1.80231795e-01
2.56275952e-01 2.32009292e-01 -2.92924251e-02 -2.29112208e-01
-1.27521217e+00 2.91782349e-01 5.53949118e-01 -1.03722155e-01
-1.03713274e+00 -7.21546948e-01 -5.95121384e-01 3.10389638e-01
2.12002471e-01 -3.31150651e-01 1.08376014e+00 -1.09672844e+00
-1.88303578e+00 3.13661456e-01 -1.18882814e-02 -5.22613108e-01
6.72954917e-01 -5.47998026e-02 -7.19582736e-01 7.71992922e-01
-3.26450676e-01 5.07440209e-01 1.14865518e+00 -9.52349186e-01
-6.99220300e-01 3.08998883e-01 -2.52320953e-02 2.53701955e-01
-4.02804285e-01 4.07378316e-01 -8.18184793e-01 -8.07367086e-01
2.50510156e-01 -4.00199383e-01 1.99077390e-02 1.53146088e-01
1.73574820e-01 3.68709713e-01 9.80960548e-01 -8.35618079e-01
1.48508883e+00 -2.46959162e+00 1.35579839e-01 -1.00329421e-01
1.80492282e-01 5.67400336e-01 -1.73563927e-01 3.96878459e-02
-6.71814382e-02 2.58277237e-01 8.26845765e-02 1.08966783e-01
-5.10534942e-01 1.21138223e-01 -3.17341298e-01 3.01987201e-01
-1.49885833e-01 3.41016084e-01 -7.26446092e-01 -6.71667039e-01
2.07013965e-01 7.95863986e-01 -9.43260610e-01 5.24082661e-01
3.75999138e-02 2.82309234e-01 -2.31505454e-01 5.51497340e-01
9.95451987e-01 -2.64696658e-01 3.93798023e-01 -7.13396847e-01
-2.47437745e-01 1.54991642e-01 -1.12680924e+00 1.41543400e+00
-5.26433647e-01 7.08064377e-01 2.75100112e-01 -8.03254068e-01
1.01863897e+00 6.38470650e-01 5.65567732e-01 -1.00907993e+00
2.25462034e-01 2.40228891e-01 -1.90777615e-01 -6.29649520e-01
7.14493155e-01 2.56769136e-02 4.67312455e-01 -4.92880084e-02
-2.38206550e-01 -2.74040718e-02 2.00077236e-01 -5.74148726e-03
8.87123942e-01 7.42942793e-03 1.19956605e-01 -2.33806700e-01
8.20774913e-01 -6.46858692e-01 7.70466387e-01 2.24047944e-01
-1.51755288e-01 9.31248665e-01 4.64775741e-01 -4.07564044e-01
-1.74603128e+00 -1.01961684e+00 -3.30691785e-01 8.15168440e-01
5.00672936e-01 -5.94509006e-01 -7.64425099e-01 9.51166376e-02
-6.41266346e-01 5.27042985e-01 2.19794735e-01 -9.97272059e-02
-8.00567865e-01 -2.89887756e-01 4.40742731e-01 2.29412273e-01
1.04575658e+00 -6.77544355e-01 -6.14137948e-01 5.75888157e-01
-5.47250390e-01 -1.46356773e+00 -4.90824938e-01 -4.43625510e-01
-1.04690647e+00 -7.59365976e-01 -9.50799227e-01 -8.42971742e-01
3.39457214e-01 5.97741008e-01 9.03260529e-01 2.25453615e-01
1.69045642e-01 -1.96172431e-01 -8.37467670e-01 4.62155968e-01
-9.11241829e-01 -1.37181625e-01 6.17753603e-02 7.83813465e-03
-3.57895821e-01 -8.47189665e-01 -8.20167065e-01 6.07464492e-01
-1.19735813e+00 6.10995352e-01 1.48366928e-01 3.41295034e-01
5.41177750e-01 4.30432200e-01 4.01018828e-01 -1.75126106e-01
2.82356620e-01 -2.99134254e-01 -7.39045262e-01 -3.33607383e-02
-3.21282059e-01 -1.75114080e-01 1.23385072e+00 -5.46714306e-01
-1.13585258e+00 -5.38336158e-01 -4.92764652e-01 -6.81142986e-01
2.93705165e-01 1.80159621e-02 -2.56992012e-01 -8.02749321e-02
3.10801923e-01 3.91665518e-01 4.62023392e-02 -3.48501384e-01
1.71390753e-02 8.76817882e-01 5.92577398e-01 -3.08338165e-01
6.26674294e-01 3.06091726e-01 -8.87973905e-02 -7.67434716e-01
-2.13263080e-01 1.73491552e-01 -1.03240542e-01 -4.79077280e-01
7.94172108e-01 -1.31391025e+00 -6.95048451e-01 4.52861160e-01
-1.15303957e+00 -3.82104844e-01 -6.16955683e-02 7.22987175e-01
-6.89479291e-01 6.53212070e-01 -1.20292640e+00 -4.67061043e-01
-3.30458760e-01 -1.49522209e+00 7.06624568e-01 2.64266998e-01
4.16456759e-01 -4.26856190e-01 -4.58570659e-01 1.46030948e-01
8.56668472e-01 -1.37661934e-01 9.74649847e-01 2.08951771e-01
-7.59187996e-01 2.95075834e-01 -6.20939732e-01 4.94125217e-01
1.61804616e-01 3.25251907e-01 -5.42230368e-01 -4.73967433e-01
1.81457326e-01 1.29353153e-02 4.58079427e-01 5.14675677e-01
1.62964976e+00 -5.36093295e-01 6.81929067e-02 1.27971506e+00
1.73043561e+00 5.24654150e-01 1.39098108e+00 4.45583493e-01
4.69505042e-01 8.77977684e-02 5.73931992e-01 8.96234870e-01
2.12110654e-02 7.88036644e-01 5.80974460e-01 1.73773497e-01
-7.69147798e-02 -2.60358095e-01 5.23586571e-01 1.26431310e+00
-6.62520289e-01 -5.68472147e-01 -3.99250627e-01 -2.57952530e-02
-1.29452717e+00 -1.25804722e+00 -8.78686979e-02 2.26752639e+00
9.00875986e-01 2.54923731e-01 7.00267404e-02 3.34658831e-01
1.07014179e+00 3.94563824e-01 -2.99253434e-01 -4.91267294e-01
1.67776663e-02 -1.48329437e-01 6.67332232e-01 2.23313943e-01
-5.20302773e-01 4.95155483e-01 6.92006159e+00 1.18751156e+00
-1.16714418e+00 -1.03852823e-01 7.48306036e-01 -2.10732967e-01
-3.25989813e-01 -1.35062486e-01 -5.89686811e-01 9.52040732e-01
1.36159992e+00 -2.64461339e-01 7.54733682e-01 8.40545833e-01
7.40676701e-01 1.88278317e-01 -6.27900064e-01 1.26064217e+00
-2.03036591e-01 -1.50988376e+00 3.36134285e-01 7.22822454e-03
4.97740179e-01 -2.41780847e-01 -1.04735486e-01 8.30203518e-02
-3.14027160e-01 -5.91824889e-01 8.45723748e-01 1.49862707e-01
1.55333829e+00 -8.93365622e-01 5.95707476e-01 3.04988734e-02
-1.61230421e+00 -2.68827975e-01 -8.02384853e-01 1.30187556e-01
5.09997666e-01 5.23877025e-01 -1.68169364e-01 4.31012779e-01
9.28089976e-01 7.03192115e-01 -1.04303837e-01 9.11490440e-01
1.10462882e-01 6.65152788e-01 -4.69264835e-02 4.57972974e-01
-2.12684080e-01 -3.53202134e-01 4.16165441e-01 1.14008892e+00
8.21844339e-01 3.42864752e-01 -3.06047589e-01 3.97437483e-01
-1.90050885e-01 1.06210448e-01 -1.57691836e-02 2.39215642e-01
6.60879135e-01 8.20408106e-01 -5.38667142e-01 -4.96707231e-01
-4.73220676e-01 1.03082907e+00 -3.00569385e-01 3.92040998e-01
-1.46885395e+00 -5.03998041e-01 6.99707091e-01 4.62620407e-01
4.91842538e-01 -2.54290313e-01 3.65036994e-01 -1.31377602e+00
-3.69089469e-02 -1.06886125e+00 -9.94354114e-02 -9.58949387e-01
-6.15535915e-01 6.97208345e-01 -1.13118710e-02 -1.88701475e+00
-3.19201909e-02 -2.37386882e-01 -1.62381545e-01 5.67885518e-01
-1.68910682e+00 -2.10551232e-01 -5.80547631e-01 6.55167460e-01
8.87649238e-01 -1.16351083e-01 3.04729253e-01 7.93982565e-01
-4.33902800e-01 5.69819331e-01 2.51863688e-01 -7.77978823e-02
4.81570750e-01 -3.67733210e-01 1.53081208e-01 8.60805511e-01
-7.07364500e-01 1.19981520e-01 8.18066597e-01 -3.36755425e-01
-1.61076736e+00 -1.04488766e+00 2.93092877e-01 7.26226211e-01
4.26958472e-01 -1.52622223e-01 -1.19240475e+00 2.01614186e-01
2.26531729e-01 1.63306162e-01 4.31796700e-01 -8.62031162e-01
-3.41732919e-01 -5.29234529e-01 -1.16857481e+00 6.62934661e-01
1.15839219e+00 -3.61505479e-01 -3.10608037e-02 2.38883607e-02
1.32260978e+00 -5.63964128e-01 -1.11120379e+00 5.29215455e-01
4.31606144e-01 -1.39703619e+00 1.02714241e+00 5.60233518e-02
1.02117741e+00 -4.06565845e-01 -5.00530362e-01 -7.68755794e-01
-5.64875126e-01 -7.02706814e-01 -3.00173134e-01 1.27983022e+00
-6.27991483e-02 -3.00207049e-01 5.48199892e-01 2.99037755e-01
-1.04575105e-01 -4.58825648e-01 -9.21426296e-01 -7.76479721e-01
-5.15496969e-01 -5.49376011e-01 7.84561694e-01 5.70603907e-01
-5.88394236e-03 -1.03154205e-01 -7.49965012e-01 2.32763052e-01
5.93441308e-01 -2.81598866e-01 4.83290166e-01 -6.21182919e-01
-3.78973424e-01 -2.96943009e-01 -5.80131710e-01 -1.63849914e+00
-3.21314514e-01 -4.98551339e-01 -1.65027335e-01 -1.15189171e+00
1.32920593e-01 -3.77761155e-01 -3.83023880e-02 -2.11010158e-01
1.82048321e-01 3.07225943e-01 2.82346338e-01 5.24843752e-01
-4.28820670e-01 5.69735050e-01 1.51030421e+00 1.87697902e-01
-1.52828604e-01 -1.96409017e-01 -2.57248908e-01 4.62719798e-01
6.98081255e-01 -1.33750468e-01 -6.27800465e-01 -6.48181856e-01
2.70216435e-01 8.47496450e-01 -3.15320422e-03 -1.22474337e+00
-7.32689574e-02 -3.05940509e-01 2.92308331e-01 -4.31415588e-01
2.14205012e-01 -9.91889358e-01 5.30467272e-01 5.11943042e-01
-3.87953520e-01 1.50574669e-01 -2.76145693e-02 2.74036735e-01
-3.50800723e-01 -1.21414177e-01 1.14400208e+00 7.03084096e-02
-6.33290291e-01 4.71860379e-01 -7.27815151e-01 7.00262329e-03
8.91614735e-01 -4.12374526e-01 -2.60549158e-01 -5.59925139e-01
-1.78255364e-01 -1.55846402e-01 7.88289964e-01 1.53011099e-01
9.90279675e-01 -1.46039462e+00 -7.42111564e-01 4.70950633e-01
-3.19029182e-01 -2.83873621e-02 6.29948854e-01 4.55238432e-01
-1.29841709e+00 7.46517032e-02 -4.01635706e-01 -5.90443671e-01
-1.10850084e+00 7.22176433e-01 3.16917062e-01 -3.12469956e-02
-8.81946325e-01 4.72273409e-01 -4.15351950e-02 6.04853332e-01
7.34555870e-02 -1.03579976e-01 -2.98953384e-01 -2.67362624e-01
9.84682560e-01 5.51641107e-01 -1.71412677e-01 -5.23980021e-01
1.10028811e-01 4.82276559e-01 5.53929396e-02 7.92383924e-02
1.18390501e+00 -5.31839013e-01 2.38656178e-02 -1.19113341e-01
1.44944727e+00 -1.56043069e-02 -1.58549213e+00 9.81495827e-02
-6.95094287e-01 -1.07059956e+00 8.94434378e-02 1.52297616e-01
-1.54375863e+00 5.07340610e-01 5.61589479e-01 4.63061303e-01
1.65457726e+00 -3.69147420e-01 1.08349884e+00 -2.45325401e-01
7.13597894e-01 -1.08279061e+00 -2.98038516e-02 2.35764235e-01
6.62012815e-01 -5.84593832e-01 1.25766575e-01 -5.68749189e-01
-5.29159963e-01 1.48230588e+00 3.72926116e-01 -1.87997390e-02
2.49437258e-01 4.64026213e-01 -1.77852422e-01 5.07619619e-01
-8.00555110e-01 5.00001431e-01 -2.01441601e-01 4.45794046e-01
4.36217159e-01 -2.50309408e-01 -4.52360213e-01 -7.84018263e-02
-1.00748232e-02 2.26006314e-01 9.53663290e-01 4.74386007e-01
-6.08826756e-01 -8.18637252e-01 -3.26726109e-01 4.09157604e-01
-6.48763597e-01 -2.59646654e-01 9.49285150e-01 4.10542279e-01
-8.78019929e-02 9.77916777e-01 4.06998575e-01 -6.09489739e-01
1.26329497e-01 -5.32922864e-01 2.68707722e-01 2.07496528e-02
-1.06058545e-01 4.31398720e-01 -3.78443040e-02 -8.62799227e-01
-5.48821151e-01 -2.86214799e-01 -1.28382885e+00 -8.12516809e-01
-1.19854756e-01 1.14101335e-01 4.35557604e-01 5.61537564e-01
3.41222435e-01 6.98516667e-01 1.17655420e+00 -8.87810767e-01
-2.01153353e-01 -4.62382823e-01 -7.81490743e-01 3.49245220e-01
2.18859360e-01 -2.19571486e-01 -5.46808243e-01 2.97834694e-01] | [11.160856246948242, -1.7746981382369995] |
251570c5-0cda-4156-ae5d-076b264ae3e9 | segment-graph-based-image-filtering-fast | null | null | http://openaccess.thecvf.com/content_iccv_2015/html/Zhang_Segment_Graph_Based_ICCV_2015_paper.html | http://openaccess.thecvf.com/content_iccv_2015/papers/Zhang_Segment_Graph_Based_ICCV_2015_paper.pdf | Segment Graph Based Image Filtering: Fast Structure-Preserving Smoothing | In this paper, we design a new edge-aware structure, named segment graph, to represent the image and we further develop a novel double weighted average image filter (SGF) based on the segment graph. In our SGF, we use the tree distance on the segment graph to define the internal weight function of the filtering kernel, which enables the filter to smooth out high-contrast details and textures while preserving major image structures very well. While for the external weight function, we introduce a user specified smoothing window to balance the smoothing effects from each node of the segment graph. Moreover, we also set a threshold to adjust the edge-preserving performance. These advantages make the SGF more flexible in various applications and overcome the "halo" and "leak" problems appearing in most of the state-of-the-art approaches. Finally and importantly, we develop a linear algorithm for the implementation of our SGF, which has an O(N) time complexity for both gray-scale and high dimensional images, regardless of the kernel size and the intensity range. Typically, as one of the fastest edge-preserving filters, our CPU implementation achieves 0.15s per megapixel when performing filtering for 3-channel color images. The strength of the proposed filter is demonstrated by various applications, including stereo matching, optical flow, joint depth map upsampling, edge-preserving smoothing, edges detection, image abstraction and texture editing. | ['Xiaopeng Zhang', 'Feihu Zhang', 'Longquan Dai', 'Shiming Xiang'] | 2015-12-01 | null | null | null | iccv-2015-12 | ['stereo-matching'] | ['computer-vision'] | [ 2.73870349e-01 -3.01372945e-01 1.03347120e-03 -1.80178508e-01
-7.92100579e-02 -2.63993889e-01 2.02546254e-01 1.53811991e-01
-5.86659253e-01 4.74289477e-01 -1.01142293e-02 -2.15705425e-01
4.57186922e-02 -1.15794337e+00 -4.41879600e-01 -7.56045699e-01
8.08002800e-02 -4.87605810e-01 9.73970830e-01 -2.85142884e-02
5.07328093e-01 7.52921581e-01 -1.66729164e+00 4.15389333e-03
1.20092928e+00 1.17967796e+00 2.51380384e-01 6.02540135e-01
-3.29168677e-01 3.32221568e-01 -3.22813213e-01 -3.28072250e-01
4.73915577e-01 -3.47961485e-01 -3.85232389e-01 3.75675857e-01
6.90184593e-01 -3.57727855e-01 -4.06450778e-01 1.37042153e+00
5.05650103e-01 2.50124902e-01 3.50348324e-01 -1.12470460e+00
-2.86700904e-01 7.08178198e-03 -9.02190089e-01 1.72303930e-01
9.67087746e-02 1.20672867e-01 4.84147191e-01 -7.44185448e-01
6.98484063e-01 1.14934576e+00 5.25969803e-01 2.52300531e-01
-1.37363183e+00 -5.51763892e-01 1.67535335e-01 1.89703181e-01
-1.32257128e+00 -3.15013587e-01 8.52265358e-01 -3.72342706e-01
4.43210632e-01 3.83423746e-01 8.32330227e-01 2.21866980e-01
4.43697572e-01 4.88532424e-01 1.20846498e+00 -4.59691286e-01
1.20925769e-01 -1.26630589e-01 1.57280743e-01 9.44525361e-01
2.74419934e-01 1.15716361e-01 -3.27377796e-01 -1.79343835e-01
1.45536947e+00 8.54313150e-02 -6.79266691e-01 -5.56776583e-01
-1.07761705e+00 5.73226213e-01 4.11951333e-01 8.04702118e-02
-2.17953488e-01 1.77693948e-01 3.47702295e-01 1.10817686e-01
4.59090769e-01 -1.76798463e-01 -6.25294223e-02 1.95449263e-01
-8.65954220e-01 -3.23273465e-02 5.86004019e-01 7.30618179e-01
1.07611036e+00 -1.16231769e-01 -3.16733807e-01 7.51487434e-01
1.09668754e-01 3.38985860e-01 2.48474598e-01 -1.20369899e+00
1.21410914e-01 6.86520875e-01 4.82773632e-02 -1.17825067e+00
-3.30029935e-01 -2.69296497e-01 -1.16892397e+00 8.19509149e-01
4.40150946e-01 7.29094818e-02 -9.70194697e-01 1.46196806e+00
5.54280877e-01 4.59014684e-01 -3.23615819e-01 9.08398628e-01
6.54372036e-01 5.90996563e-01 -1.79299518e-01 -3.68661672e-01
1.49148178e+00 -9.45713937e-01 -7.51792967e-01 1.80196278e-02
2.49885798e-01 -8.81774664e-01 1.22465563e+00 3.27900320e-01
-1.23243892e+00 -5.94548464e-01 -9.92691457e-01 -2.92582214e-01
-2.54767895e-01 -4.16327678e-02 5.75541139e-01 6.50751173e-01
-9.12599802e-01 6.53835952e-01 -7.99118519e-01 -1.61834911e-01
1.25124946e-01 2.10477665e-01 -2.99503833e-01 -9.49077904e-02
-9.12234008e-01 5.46988785e-01 1.45639613e-01 2.59291828e-02
-5.91987781e-02 -8.13595951e-01 -8.60554099e-01 2.09189311e-01
2.39173546e-01 -7.67735839e-01 5.17709315e-01 -9.52094555e-01
-1.68885112e+00 1.00236881e+00 -4.24543440e-01 -2.53301769e-01
5.58339357e-01 5.86402081e-02 -3.27838391e-01 2.90553778e-01
-1.76991180e-01 4.10629511e-01 1.03365362e+00 -1.03797638e+00
-7.82696545e-01 -3.17454040e-01 7.93946385e-02 1.53418571e-01
-3.25589746e-01 -8.77117738e-02 -9.07679975e-01 -8.98629129e-01
2.60767430e-01 -5.73877692e-01 -4.58471775e-01 5.78563213e-01
-1.37981787e-01 1.79824382e-01 9.95885134e-01 -4.61677551e-01
1.67933583e+00 -2.54403234e+00 -1.03627138e-01 4.04717386e-01
3.65968436e-01 3.21204394e-01 -7.00106323e-02 6.15947023e-02
1.31572142e-01 -2.92167991e-01 -4.93927479e-01 -1.62997350e-01
-3.57930899e-01 1.47191375e-01 9.50626191e-03 6.27684414e-01
-1.12175085e-01 5.29694438e-01 -7.32022583e-01 -5.55558503e-01
5.46235263e-01 7.03761280e-01 -6.97453320e-01 9.62175727e-02
3.65077525e-01 3.19878250e-01 -2.90437877e-01 3.09785753e-01
1.28796625e+00 -1.31025687e-01 -9.59104076e-02 -5.75270295e-01
-6.13258958e-01 -1.07705884e-01 -1.80954111e+00 1.52956808e+00
-4.46152478e-01 4.81432736e-01 4.82292801e-01 -7.12305129e-01
9.90220368e-01 -9.77870002e-02 4.75063086e-01 -6.87784255e-01
4.85874526e-02 1.77686468e-01 -2.64812678e-01 -2.23213002e-01
5.27785718e-01 4.90087457e-02 4.77513254e-01 2.82282114e-01
-4.01369750e-01 -6.29795641e-02 3.52180839e-01 3.02925110e-02
8.77929389e-01 -1.02386676e-01 3.10529172e-01 -6.31196916e-01
1.00847900e+00 -3.84794533e-01 7.77801871e-01 5.49179614e-01
-3.33221376e-01 8.69585991e-01 4.53124136e-01 -4.48401928e-01
-6.38792396e-01 -1.07440436e+00 -4.80947763e-01 7.46526837e-01
6.93229139e-01 -3.74282926e-01 -9.06578839e-01 -3.35361093e-01
2.73761954e-02 1.27616316e-01 -3.80711049e-01 -3.65262404e-02
-7.07689345e-01 -6.23509467e-01 7.31388405e-02 3.15157861e-01
9.86793697e-01 -6.05032623e-01 -8.04134905e-01 3.42900634e-01
3.34422141e-02 -1.09041202e+00 -1.04807866e+00 -1.74445882e-01
-1.01741886e+00 -1.08792365e+00 -7.62286901e-01 -8.25794518e-01
9.00183856e-01 5.30994952e-01 8.68373573e-01 2.64396608e-01
-3.78789365e-01 1.72297522e-01 -1.14651084e-01 6.97791502e-02
2.61876993e-02 -3.65607262e-01 -3.65269840e-01 3.93686533e-01
-7.61765465e-02 -6.42980635e-01 -1.12592411e+00 5.08941889e-01
-1.14716089e+00 3.17720413e-01 2.87896693e-01 8.12114239e-01
8.33163500e-01 2.08643034e-01 -9.91313756e-02 -8.01501274e-01
5.98964751e-01 2.56799281e-01 -1.03575540e+00 1.76233739e-01
-5.59161425e-01 4.09628674e-02 7.86769629e-01 -5.18640876e-01
-1.06300128e+00 1.60841435e-01 2.75261747e-03 -3.25159937e-01
2.04826564e-01 3.72233614e-02 -6.73493445e-02 -5.87893903e-01
2.95686871e-01 1.42520979e-01 2.71549314e-01 -4.52938318e-01
3.57312024e-01 4.24364984e-01 7.58531690e-01 -3.92158389e-01
6.39611721e-01 9.88219380e-01 4.03108597e-01 -8.95790875e-01
-4.87039685e-01 -4.50684011e-01 -6.17506742e-01 -1.27437115e-01
8.18439424e-01 -6.56451881e-01 -9.44433868e-01 8.54378104e-01
-1.11486173e+00 -3.20373654e-01 -2.61684775e-01 3.95940959e-01
-3.83782744e-01 6.55699313e-01 -7.24390864e-01 -6.34513378e-01
-4.13587123e-01 -1.26751149e+00 8.35101724e-01 5.15933692e-01
2.71677822e-01 -1.11925662e+00 -1.44734159e-01 -2.54115701e-01
6.59286261e-01 4.11105335e-01 8.58817697e-01 4.73100781e-01
-8.56770039e-01 1.03399195e-01 -6.56131566e-01 5.14468551e-01
1.83881700e-01 1.71345472e-01 -7.31104732e-01 -3.78978968e-01
-5.59141003e-02 3.76768440e-01 9.72190619e-01 7.43326843e-01
1.24528408e+00 6.95741251e-02 -2.05118373e-01 1.21413112e+00
1.67569232e+00 6.84970766e-02 1.03230822e+00 2.72368491e-01
6.11208320e-01 4.96095449e-01 5.09034276e-01 4.55152810e-01
1.24752522e-01 7.58350492e-01 2.37053856e-01 -6.69560909e-01
-4.59540367e-01 -9.24374629e-03 2.08371118e-01 5.63227296e-01
-2.45986804e-01 1.24324828e-01 -3.41893882e-01 1.56660706e-01
-1.66551948e+00 -7.69876003e-01 -4.27121013e-01 2.61765409e+00
7.03867853e-01 7.44772106e-02 -9.06698182e-02 2.09530607e-01
9.34056222e-01 3.40683520e-01 -4.33846593e-01 -4.77655590e-01
-1.68724254e-01 5.14633715e-01 8.80933881e-01 8.58797967e-01
-1.14203715e+00 8.22903335e-01 6.28911161e+00 9.36784029e-01
-1.28487587e+00 -1.68628603e-01 6.09083831e-01 2.71613538e-01
-2.20906809e-01 7.09516183e-02 -7.57406771e-01 5.59691310e-01
2.87947536e-01 -3.18406165e-01 3.65082204e-01 5.51949382e-01
5.00384867e-01 -4.08712685e-01 -5.10065377e-01 1.16205490e+00
-1.14976630e-01 -1.28725255e+00 -6.50994033e-02 6.27776459e-02
6.15799248e-01 -3.76524419e-01 -7.11947903e-02 -3.96356314e-01
-1.53824210e-01 -6.97423637e-01 4.47405130e-01 3.53149056e-01
8.63645971e-01 -8.39217186e-01 3.05865109e-01 1.25963092e-02
-1.58933687e+00 1.04560435e-01 -5.60260773e-01 8.61732736e-02
3.32269877e-01 1.05737591e+00 1.58253327e-01 3.74044627e-01
6.50417626e-01 6.19439423e-01 -3.58189076e-01 1.35868287e+00
-1.71695538e-02 -5.69655187e-03 -4.27671462e-01 3.81603062e-01
1.11470364e-01 -7.96004832e-01 5.13932884e-01 1.25511467e+00
3.43596607e-01 2.82555044e-01 1.86457247e-01 7.66961694e-01
9.02659670e-02 2.41617262e-01 -2.11668611e-01 5.48610151e-01
3.55794042e-01 1.27606630e+00 -9.38308597e-01 -4.23366904e-01
-7.03950584e-01 1.29908311e+00 3.50662023e-02 5.00224829e-01
-7.38228202e-01 -9.86301422e-01 9.64371920e-01 3.88337582e-01
3.48573714e-01 -3.20103317e-01 -4.36047643e-01 -1.26125848e+00
1.46335408e-01 -5.53991795e-01 3.48698944e-01 -5.06901383e-01
-9.67273653e-01 4.63573843e-01 -2.18767017e-01 -1.34681749e+00
3.87402654e-01 -7.05111444e-01 -7.45100379e-01 9.40557599e-01
-1.91822052e+00 -5.75062692e-01 -6.61667526e-01 9.64417338e-01
2.20794976e-01 4.05386060e-01 4.30309892e-01 6.28062129e-01
-5.33171237e-01 4.23916906e-01 1.00169361e-01 9.89071205e-02
8.38821530e-01 -9.50413525e-01 3.51352304e-01 1.16125381e+00
-2.69828230e-01 6.20431244e-01 4.39125687e-01 -4.49061036e-01
-1.26421249e+00 -8.57556343e-01 5.74717641e-01 2.78119713e-01
3.35665345e-01 -2.67015904e-01 -1.15487611e+00 3.19957614e-01
5.63667864e-02 5.28697848e-01 1.91851869e-01 -5.14558792e-01
-2.15761378e-01 -4.09544677e-01 -1.33951747e+00 6.63975775e-01
1.08765042e+00 -2.49530256e-01 -9.00414959e-02 -1.01355329e-01
4.55144703e-01 -6.14833355e-01 -9.20479536e-01 4.14487869e-01
6.81823313e-01 -1.53569555e+00 1.18429887e+00 6.46418855e-02
5.73937260e-02 -7.33268738e-01 2.12299973e-01 -1.06165898e+00
-6.34752810e-01 -8.26868355e-01 2.59557944e-02 1.08949554e+00
-2.90357489e-02 -9.08633769e-01 7.66326010e-01 5.35320818e-01
-8.59051794e-02 -6.11383915e-01 -9.79402542e-01 -5.62767625e-01
-3.79905969e-01 -1.03051536e-01 4.52236623e-01 6.93428934e-01
-2.58070588e-01 -1.73517600e-01 -1.80683941e-01 2.30099812e-01
1.00740039e+00 3.51843953e-01 7.26401687e-01 -1.15502465e+00
-2.70993203e-01 -6.51242852e-01 -5.94311297e-01 -1.48590982e+00
-2.23765984e-01 -5.09791255e-01 -3.59749883e-01 -1.46166873e+00
1.18786274e-02 -3.66879940e-01 -1.79358050e-01 9.52248126e-02
-3.66351515e-01 4.32494521e-01 1.93806395e-01 -9.27812234e-02
-2.59442162e-02 3.04420203e-01 1.68214858e+00 2.28265777e-01
-4.66438144e-01 -3.45162414e-02 -3.69750112e-01 8.46283138e-01
4.73521531e-01 -5.97441681e-02 -3.02675009e-01 -4.22454506e-01
-2.62388855e-01 -1.69307247e-01 2.65806943e-01 -8.69655192e-01
2.89478123e-01 -2.86906064e-01 2.41219372e-01 -4.21686053e-01
2.25133747e-01 -9.38754022e-01 1.14230521e-01 6.78576589e-01
1.14428714e-01 -3.15014310e-02 1.23746112e-01 3.84361088e-01
-3.86897832e-01 -2.05067284e-02 1.51049888e+00 -2.73067113e-02
-8.90531659e-01 5.68619728e-01 -2.17998445e-01 -7.53734261e-02
1.08924878e+00 -5.70312560e-01 -3.07881325e-01 -2.56574363e-01
-4.02660578e-01 1.27233118e-01 7.86757290e-01 -6.68550050e-03
8.15381706e-01 -1.36431885e+00 -4.58283007e-01 7.39918828e-01
-9.24234614e-02 -1.18970081e-01 4.70660448e-01 1.01956797e+00
-9.15695071e-01 -1.06042914e-01 -4.17235672e-01 -5.07075489e-01
-1.35248673e+00 4.80884045e-01 3.53608221e-01 -2.22262576e-01
-1.03977633e+00 6.03496075e-01 5.41370511e-01 2.80343860e-01
3.03114027e-01 -4.50463146e-01 2.27472167e-02 -2.07544267e-01
8.39940429e-01 6.27524853e-01 -5.35827018e-02 -4.72118199e-01
-1.75420627e-01 1.14764619e+00 2.24848360e-01 6.71395473e-03
9.36241329e-01 -4.13754880e-01 -3.91874194e-01 -8.72557424e-03
1.25683117e+00 3.57730836e-01 -1.49088287e+00 -1.80113167e-01
-3.83308113e-01 -9.40365851e-01 4.52397168e-01 -1.59833357e-01
-1.42426455e+00 9.04214263e-01 6.87498927e-01 3.21256042e-01
1.56466198e+00 -6.06639743e-01 1.05559707e+00 -3.08832288e-01
2.05690235e-01 -1.05652344e+00 -2.84991652e-01 3.15629065e-01
5.73267579e-01 -8.33464503e-01 9.26919580e-02 -8.78673196e-01
-1.62635833e-01 1.27980947e+00 4.72268701e-01 -3.08322519e-01
8.46181810e-01 6.66225135e-01 7.55536184e-02 2.34568194e-01
-2.06147701e-01 -4.19726074e-01 2.87345409e-01 5.77544689e-01
3.75553429e-01 -1.47252724e-01 -8.95839989e-01 -1.06749125e-01
2.77168542e-01 4.65649106e-02 5.03665209e-01 7.12455273e-01
-7.74474561e-01 -1.22452748e+00 -4.21463966e-01 2.22899944e-01
-3.48431647e-01 -1.63517758e-01 2.27390960e-01 6.17833316e-01
1.09241135e-01 6.22558057e-01 3.80206555e-01 1.51664410e-02
6.51564777e-01 -2.84496814e-01 4.87956643e-01 -1.86299458e-01
-4.37555313e-01 3.17003727e-01 -2.70120680e-01 -9.58569586e-01
-5.26425064e-01 -1.76934689e-01 -1.31377411e+00 -4.37738627e-01
-1.79847464e-01 -1.54203787e-01 3.85708600e-01 2.45314062e-01
4.98763978e-01 2.76956052e-01 6.91746294e-01 -7.49094486e-01
-3.42665277e-02 -4.63949889e-01 -9.39086914e-01 6.27334356e-01
3.14564794e-01 -6.49822354e-01 -4.86922801e-01 3.65296975e-02] | [11.058063507080078, -2.5060036182403564] |
1838eb76-1927-4f02-891f-9fd99c3eb417 | interpretable-and-accurate-fine-grained | 2005.10411 | null | https://arxiv.org/abs/2005.10411v1 | https://arxiv.org/pdf/2005.10411v1.pdf | Interpretable and Accurate Fine-grained Recognition via Region Grouping | We present an interpretable deep model for fine-grained visual recognition. At the core of our method lies the integration of region-based part discovery and attribution within a deep neural network. Our model is trained using image-level object labels, and provides an interpretation of its results via the segmentation of object parts and the identification of their contributions towards classification. To facilitate the learning of object parts without direct supervision, we explore a simple prior of the occurrence of object parts. We demonstrate that this prior, when combined with our region-based part discovery and attribution, leads to an interpretable model that remains highly accurate. Our model is evaluated on major fine-grained recognition datasets, including CUB-200, CelebA and iNaturalist. Our results compare favorably to state-of-the-art methods on classification tasks, and our method outperforms previous approaches on the localization of object parts. | ['Zixuan Huang', 'Yin Li'] | 2020-05-21 | interpretable-and-accurate-fine-grained-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Huang_Interpretable_and_Accurate_Fine-grained_Recognition_via_Region_Grouping_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Huang_Interpretable_and_Accurate_Fine-grained_Recognition_via_Region_Grouping_CVPR_2020_paper.pdf | cvpr-2020-6 | ['fine-grained-visual-recognition'] | ['computer-vision'] | [ 2.00664610e-01 3.16199988e-01 -3.69687051e-01 -6.61646485e-01
-5.31340420e-01 -8.36078525e-01 8.98493350e-01 -6.48783939e-03
-3.42514552e-02 4.59391326e-01 1.94064513e-01 -3.46373841e-02
5.54439947e-02 -6.29413545e-01 -1.26093340e+00 -4.20506239e-01
1.22198522e-01 9.10199404e-01 3.73788148e-01 2.38790959e-01
3.94275248e-01 1.00405848e+00 -1.60922444e+00 7.52254248e-01
3.12899411e-01 1.60018671e+00 -1.43989667e-01 4.26109821e-01
-2.69081295e-01 9.84552562e-01 -6.43904805e-01 -2.40808949e-01
3.31064284e-01 1.61743209e-01 -1.27193964e+00 7.03633010e-01
8.41103196e-01 -5.26405752e-01 -1.17619485e-01 8.97209525e-01
-1.54847234e-01 -2.24973768e-01 1.14571857e+00 -1.39231455e+00
-8.19027483e-01 5.35493791e-01 -5.48993528e-01 1.34154126e-01
-2.31135171e-02 1.37108669e-01 1.09722006e+00 -9.64536548e-01
4.95269895e-01 1.48710012e+00 8.15931022e-01 3.09032291e-01
-1.43186128e+00 -4.49050486e-01 5.36728144e-01 -7.32387155e-02
-1.68414259e+00 -6.36400938e-01 4.51413155e-01 -7.50544310e-01
8.03443551e-01 3.53520066e-02 4.80584413e-01 7.72036195e-01
-1.53088376e-01 1.01999938e+00 1.13294375e+00 -3.74229819e-01
3.47054720e-01 -1.50926374e-02 3.92957091e-01 9.42550480e-01
3.37924868e-01 -6.93462044e-02 -3.06746751e-01 -1.26427323e-01
1.00609803e+00 2.40834102e-01 5.14460988e-02 -6.25029981e-01
-1.18442345e+00 5.24243414e-01 9.48795676e-01 1.40206218e-01
-4.25912261e-01 5.96032381e-01 1.28770605e-01 -3.45770389e-01
6.06164157e-01 2.31097445e-01 -6.13602281e-01 3.33607435e-01
-1.03789687e+00 1.09743848e-01 7.59777129e-01 8.47791493e-01
9.86896813e-01 -2.22267866e-01 -4.81812745e-01 7.69811273e-01
7.22796261e-01 1.80449978e-01 1.54441334e-02 -1.22526252e+00
-3.82951871e-02 1.06491411e+00 1.34121001e-01 -4.88217950e-01
-1.24371640e-01 -5.93920887e-01 -5.15607059e-01 7.62840092e-01
5.00294447e-01 5.06862581e-01 -1.54791164e+00 1.34812880e+00
3.53223681e-01 -1.31191522e-01 -5.08931875e-01 8.56115818e-01
7.67138124e-01 7.11358245e-03 2.63836890e-01 6.43277824e-01
1.58928275e+00 -1.06658697e+00 -2.51693904e-01 -2.87925810e-01
7.52798021e-02 -4.36498463e-01 8.22327852e-01 2.03040287e-01
-7.44943798e-01 -6.52876616e-01 -8.99316669e-01 -2.63162434e-01
-6.87746465e-01 3.49449813e-01 8.02323043e-01 4.68496054e-01
-1.02426517e+00 7.77983248e-01 -7.81921864e-01 -2.41075829e-01
1.38053560e+00 4.96584743e-01 -6.56676054e-01 1.81501377e-02
-2.16823071e-01 8.33626986e-01 5.08086681e-01 4.18121256e-02
-1.42325127e+00 -5.57260931e-01 -8.60935152e-01 1.69340193e-01
6.34830222e-02 -8.05086911e-01 1.32933366e+00 -1.31075108e+00
-9.21252847e-01 1.30798173e+00 -2.62255877e-01 -3.86959821e-01
3.42260748e-01 -9.06313807e-02 1.90652370e-01 2.79934943e-01
2.85502791e-01 1.26656926e+00 1.09844315e+00 -1.87295258e+00
-6.80501044e-01 -4.56913680e-01 3.40020120e-01 -2.35619694e-02
3.65296975e-02 1.91254765e-02 -5.18437922e-01 -5.95130265e-01
1.58530399e-01 -6.27115190e-01 -3.63106467e-02 7.14120746e-01
-6.51512086e-01 -5.30874610e-01 7.33991325e-01 -4.27651197e-01
4.72418427e-01 -2.13183808e+00 -2.62762427e-01 -7.57045252e-03
6.24645054e-01 -8.47882852e-02 -1.61089778e-01 -1.08898014e-01
-7.12012500e-02 3.67456764e-01 -1.60022616e-01 -6.32224500e-01
3.38080496e-01 3.13602120e-01 -4.52130258e-01 7.48796999e-01
6.11189842e-01 1.24776232e+00 -5.62605321e-01 -6.15689099e-01
3.58264953e-01 2.87298560e-01 -2.59015292e-01 2.04271197e-01
-4.89950776e-01 -4.35789376e-02 -4.51577485e-01 1.34043074e+00
6.46412134e-01 -7.86434650e-01 -1.73396423e-01 -5.94691515e-01
5.85112274e-02 3.09531987e-01 -8.50513160e-01 1.30453205e+00
1.04153588e-01 5.22284806e-01 1.85863033e-01 -8.43911469e-01
5.65717161e-01 2.74243243e-02 1.18230589e-01 -3.93045187e-01
2.60228306e-01 -2.17643715e-02 -1.01453707e-01 -7.41945952e-02
2.90535510e-01 -4.47685532e-02 5.96084185e-02 7.31124520e-01
3.93027335e-01 -3.83333601e-02 -4.62233536e-02 1.71773151e-01
9.55863535e-01 5.07182121e-01 5.58292508e-01 -3.73312473e-01
-2.28101155e-03 1.44172445e-01 2.39695936e-01 9.33467805e-01
-4.48997945e-01 8.50093603e-01 2.96739042e-01 -6.44993603e-01
-1.09931040e+00 -1.12175143e+00 -3.45580220e-01 1.40589368e+00
1.87662944e-01 8.75497144e-03 -7.71634042e-01 -1.21095574e+00
5.04414797e-01 2.95708656e-01 -1.35736239e+00 1.91510543e-01
-2.09401354e-01 -2.69735664e-01 6.85083508e-01 9.44820702e-01
4.63157892e-01 -1.26235497e+00 -2.74297774e-01 -1.94435164e-01
4.34582718e-02 -9.55419123e-01 -2.62433410e-01 6.02343440e-01
-8.29656899e-01 -1.22504747e+00 -4.35799062e-01 -7.44016886e-01
1.10735273e+00 3.17650676e-01 1.47690189e+00 3.36917222e-01
-5.08809209e-01 4.34130728e-01 -1.58137128e-01 -4.99164343e-01
-1.92965031e-01 -1.52969107e-01 -2.43308604e-01 -1.71832517e-02
3.35149169e-01 -2.49937311e-01 -5.36859334e-01 5.31566918e-01
-7.91092098e-01 -5.14903925e-02 8.92829001e-01 7.19554365e-01
8.41222763e-01 -2.11954564e-01 1.11669935e-01 -8.09883177e-01
-4.08454891e-03 -3.12018901e-01 -4.82441753e-01 5.01973748e-01
-3.65170985e-01 2.49645069e-01 1.56007871e-01 -3.19707513e-01
-9.38970983e-01 3.33697766e-01 1.35726780e-01 -7.91125178e-01
-9.14378464e-01 -2.09064186e-01 -2.84806252e-01 -4.92545336e-01
6.72954261e-01 9.80877429e-02 -1.53047472e-01 -7.90044427e-01
6.30763769e-01 7.30445921e-01 6.19663596e-01 -6.72836959e-01
6.19626462e-01 9.26957846e-01 -9.11234468e-02 -4.86100346e-01
-1.19293320e+00 -5.70064843e-01 -1.01349115e+00 -7.17826840e-03
7.59706378e-01 -1.03391719e+00 -8.32106531e-01 4.93166655e-01
-1.28563571e+00 -4.37915415e-01 -4.80762035e-01 -1.15086116e-01
-5.54058015e-01 2.01215222e-01 -6.45470262e-01 -5.98193347e-01
-2.78322816e-01 -9.47179496e-01 1.72439110e+00 1.21674716e-01
-2.80773550e-01 -8.58645737e-01 -3.70617718e-01 6.64544106e-01
1.23439789e-01 1.08726896e-01 6.97953939e-01 -7.84330606e-01
-1.07835066e+00 -1.09871134e-01 -8.61150563e-01 2.25010216e-01
9.28544924e-02 3.40584889e-02 -1.33705711e+00 5.62100904e-03
-4.81001794e-01 -7.32651412e-01 1.09572053e+00 3.88470024e-01
1.70813167e+00 -4.60557371e-01 -7.37753928e-01 5.88321984e-01
1.19687247e+00 -2.91053772e-01 4.53983605e-01 3.52100670e-01
8.80807579e-01 5.86521506e-01 4.86781031e-01 3.30764711e-01
2.85867453e-01 3.96675617e-01 9.42893803e-01 -3.54544550e-01
-5.97604930e-01 -1.72653839e-01 -1.23575822e-01 -3.08650017e-01
-1.63665831e-01 -3.01073879e-01 -6.79018676e-01 7.04576373e-01
-1.73471653e+00 -9.26659465e-01 1.17415406e-01 1.67513227e+00
6.67324960e-01 7.36135766e-02 1.08982712e-01 -7.84055367e-02
6.25329852e-01 -3.32273985e-03 -7.64999747e-01 -1.16702318e-01
-1.34875933e-02 1.54184461e-01 5.69924295e-01 2.82472819e-01
-1.54484880e+00 1.07362378e+00 7.67626762e+00 8.73675585e-01
-7.64685214e-01 3.67027856e-02 1.04334772e+00 2.37529814e-01
-1.52843282e-01 -1.51572138e-01 -9.82671857e-01 1.03150167e-01
3.18212837e-01 4.76686358e-01 4.38845158e-01 1.04750299e+00
-2.17437863e-01 -2.75291447e-02 -1.50415242e+00 6.82352066e-01
2.01010108e-01 -1.52532053e+00 2.26222023e-01 2.05178559e-01
7.01537549e-01 2.29847044e-01 -1.98773727e-01 4.19068113e-02
6.00361407e-01 -1.43452740e+00 1.33195281e+00 5.60551226e-01
7.31612682e-01 -2.92704940e-01 5.47483504e-01 2.10189030e-01
-1.06732500e+00 -4.07472588e-02 -4.00499195e-01 -7.26823509e-02
-2.77924716e-01 4.18696314e-01 -1.16212785e+00 1.37178764e-01
8.06187749e-01 7.81895459e-01 -7.79514670e-01 9.94958997e-01
-4.19770926e-01 5.83355308e-01 -2.98752934e-01 2.26095721e-01
2.96982881e-02 4.00214642e-01 1.10553026e-01 1.22575939e+00
-2.82983065e-01 3.23564447e-02 3.60108554e-01 1.42301559e+00
-4.49831814e-01 -5.58266759e-01 -3.15036863e-01 -1.75534219e-01
2.72601485e-01 1.45464051e+00 -1.04128981e+00 -6.22400284e-01
-7.33542740e-02 1.00451982e+00 8.52698922e-01 2.32827738e-01
-6.17337584e-01 -1.10072732e-01 8.37266684e-01 1.63142383e-01
8.36001992e-01 1.04737185e-01 -7.86690235e-01 -9.20984626e-01
-9.64090750e-02 -7.02758014e-01 3.28280956e-01 -9.57181156e-01
-1.57402956e+00 5.94561458e-01 -1.00240037e-01 -9.56003129e-01
-2.35531516e-02 -1.12194264e+00 -4.27909672e-01 8.22472394e-01
-1.38515997e+00 -1.73954618e+00 -3.63844395e-01 4.41923112e-01
3.98464739e-01 1.73747510e-01 8.50587130e-01 3.42690386e-02
-2.56961584e-01 5.10062158e-01 -3.54855329e-01 4.97136801e-01
3.48859757e-01 -1.42480636e+00 5.41736126e-01 7.05269456e-01
3.61177057e-01 8.23461831e-01 5.32963157e-01 -6.50729239e-01
-8.98061216e-01 -1.38588285e+00 4.73175138e-01 -9.37737286e-01
5.30622780e-01 -5.71974933e-01 -6.61847651e-01 8.94192696e-01
7.95015600e-03 7.22087324e-01 4.43002552e-01 2.86240578e-01
-8.48608017e-01 8.08923617e-02 -1.24223316e+00 6.92953467e-02
1.17641878e+00 -6.70789421e-01 -8.92102242e-01 3.98353755e-01
5.87306559e-01 -1.81350961e-01 -7.46467590e-01 4.39051330e-01
8.36491108e-01 -8.64126861e-01 1.21359098e+00 -8.14288735e-01
4.10787553e-01 -7.48829603e-01 -3.11258942e-01 -8.67971659e-01
-8.51290703e-01 3.57979566e-01 -3.07979435e-01 1.14532518e+00
2.08793297e-01 -2.88862139e-01 8.78533304e-01 6.19268000e-01
-2.78605521e-01 -6.63930595e-01 -7.40911484e-01 -6.47699237e-01
-2.37059146e-01 -2.75266886e-01 4.40626025e-01 5.81597686e-01
-5.06045401e-01 9.86310095e-02 1.80328175e-01 2.85287887e-01
9.80022192e-01 6.75600886e-01 7.24904478e-01 -1.34449887e+00
-3.44653130e-01 -7.09561050e-01 -6.44067645e-01 -1.06519365e+00
3.14794779e-01 -7.04123795e-01 5.10952771e-01 -1.73736072e+00
7.21457362e-01 -5.15956819e-01 -3.31036031e-01 1.30641627e+00
-5.77934878e-03 1.06668997e+00 5.24566993e-02 4.28390115e-01
-1.10405505e+00 3.73940736e-01 1.04088211e+00 -4.81846958e-01
4.42485929e-01 -1.85604498e-01 -8.18873823e-01 9.59180057e-01
4.08864588e-01 -4.90832478e-01 3.08036476e-01 -3.43516409e-01
-3.47943515e-01 -7.33165681e-01 9.08139884e-01 -6.98923230e-01
-1.24262802e-01 -4.47801501e-02 1.06327939e+00 -7.33702123e-01
4.40889239e-01 -7.49134004e-01 1.18095815e-01 3.12901497e-01
-2.41374612e-01 -6.69699967e-01 2.42132172e-01 5.44741631e-01
-1.19095877e-01 -4.59937416e-02 8.19430828e-01 -3.96284580e-01
-9.34790373e-01 4.17433679e-01 -1.36072472e-01 -1.06184632e-01
8.97779942e-01 -4.10575747e-01 -6.43398404e-01 1.09822780e-01
-6.90878332e-01 -6.37025386e-02 7.98322380e-01 3.10982555e-01
4.23557311e-01 -1.30822742e+00 -4.37610418e-01 2.17538223e-01
5.62376857e-01 1.24414312e-02 -5.62217645e-02 3.91307533e-01
-4.02722031e-01 4.94214565e-01 -3.47254187e-01 -8.39788437e-01
-1.27721310e+00 6.10826612e-01 5.41168571e-01 3.83766331e-02
-3.23877931e-01 9.72056031e-01 8.20782483e-01 -4.28616732e-01
3.59353632e-01 -5.06345332e-01 -5.94216846e-02 -2.81634271e-01
4.75279331e-01 9.32725668e-02 -3.43499221e-02 -7.59097576e-01
-7.52120614e-01 4.39257264e-01 -1.59781501e-01 1.97812572e-01
1.24106824e+00 -6.35630712e-02 -3.45849335e-01 3.86178941e-01
1.05385315e+00 -2.49891371e-01 -1.70119667e+00 -1.82477057e-01
-1.69526905e-01 -4.89794463e-01 2.63831541e-02 -1.31035626e+00
-8.10455263e-01 7.48678327e-01 4.76923019e-01 5.78553118e-02
7.61774778e-01 7.18108118e-01 7.19287898e-04 3.72850060e-01
4.83545542e-01 -5.33873737e-01 -2.02723835e-02 5.24260223e-01
8.91799510e-01 -1.55174053e+00 2.36869663e-01 -4.51677799e-01
-1.93235740e-01 1.07860720e+00 7.03874350e-01 -2.06469551e-01
5.69060504e-01 3.83055538e-01 3.92948091e-02 -3.57374787e-01
-5.77956498e-01 -4.47681040e-01 8.65151882e-01 5.28819501e-01
3.58014703e-01 3.09710085e-01 3.91458362e-01 5.25870800e-01
2.30115905e-01 -4.58094291e-03 -8.52344632e-02 8.38022947e-01
-6.56419218e-01 -7.77398050e-01 -6.37434065e-01 5.87960660e-01
-5.81914723e-01 -2.20883097e-02 -8.28578591e-01 6.73789859e-01
5.26527226e-01 6.99160516e-01 4.33297545e-01 -2.09491905e-02
-9.04107168e-02 -2.07376629e-02 7.50279844e-01 -7.80006766e-01
-5.32521904e-01 -6.00603856e-02 7.97838345e-02 -8.22329998e-01
-5.25161028e-01 -4.46361750e-01 -1.25446081e+00 -8.23777094e-02
-3.65931243e-01 -1.85662284e-01 7.00624585e-01 1.33541870e+00
4.57134873e-01 5.90012729e-01 2.57600725e-01 -1.34891117e+00
-2.62503415e-01 -1.01201642e+00 -6.24156237e-01 6.05724633e-01
5.05946577e-01 -8.15535903e-01 -2.99309671e-01 2.55983382e-01] | [9.537657737731934, 1.495361328125] |
d6ae7b91-32e8-4a17-8d51-3866940f8df5 | ps8-net-a-deep-convolutional-neural-network | 2009.10380 | null | https://arxiv.org/abs/2009.10380v1 | https://arxiv.org/pdf/2009.10380v1.pdf | PS8-Net: A Deep Convolutional Neural Network to Predict the Eight-State Protein Secondary Structure | Protein secondary structure is crucial to creating an information bridge between the primary and tertiary (3D) structures. Precise prediction of eight-state protein secondary structure (PSS) has significantly utilized in the structural and functional analysis of proteins in bioinformatics. Deep learning techniques have been recently applied in this research area and raised the eight-state (Q8) protein secondary structure prediction accuracy remarkably. Nevertheless, from a theoretical standpoint, there are still lots of rooms for improvement, specifically in the eight-state PSS prediction. In this study, we have presented a new deep convolutional neural network (DCNN), namely PS8-Net, to enhance the accuracy of eight-class PSS prediction. The input of this architecture is a carefully constructed feature matrix from the proteins sequence features and profile features. We introduce a new PS8 module in the network, which is applied with skip connection to extracting the long-term inter-dependencies from higher layers, obtaining local contexts in earlier layers, and achieving global information during secondary structure prediction. Our proposed PS8-Net achieves 76.89%, 71.94%, 76.86%, and 75.26% Q8 accuracy respectively on benchmark CullPdb6133, CB513, CASP10, and CASP11 datasets. This architecture enables the efficient processing of local and global interdependencies between amino acids to make an accurate prediction of each class. To the best of our knowledge, PS8-Net experiment results demonstrate that it outperforms all the state-of-the-art methods on the aforementioned benchmark datasets. | ['Won-Sook Lee', 'Md. Aminur Rab Ratul', 'M. Hamed Mozaffari', 'Maryam Tavakol Elahi'] | 2020-09-22 | null | null | null | null | ['protein-secondary-structure-prediction'] | ['medical'] | [ 3.10482711e-01 -2.71927238e-01 5.27839847e-02 -4.83643055e-01
-3.32101464e-01 -3.18842560e-01 1.48461595e-01 4.66643304e-01
-3.25655043e-01 1.02497494e+00 4.45189280e-03 -6.58413708e-01
-4.79235016e-02 -4.41977948e-01 -8.65905166e-01 -1.10357034e+00
-1.44399419e-01 5.84313385e-02 3.36791307e-01 -3.39522958e-01
3.69182587e-01 6.97642803e-01 -1.23378015e+00 6.73741043e-01
1.13016319e+00 1.03235114e+00 4.83818531e-01 4.61405665e-01
-3.88917416e-01 7.88632631e-01 -4.97922540e-01 -2.77084887e-01
-1.31473109e-01 -3.89000297e-01 -8.72805119e-01 -3.75584304e-01
-8.38830620e-02 -2.09396146e-03 5.14171273e-02 7.22579539e-01
6.92938209e-01 -2.33140066e-01 6.02707446e-01 -8.03228080e-01
-7.09700406e-01 3.15122038e-01 -6.66146159e-01 8.83691683e-02
1.86942950e-01 3.72899592e-01 1.05073500e+00 -6.82243347e-01
6.80894434e-01 8.79225552e-01 6.76710308e-01 3.11695606e-01
-1.19884884e+00 -6.61076963e-01 -1.15024656e-01 7.10922360e-01
-9.45005834e-01 1.52479604e-01 6.07519388e-01 -2.91709274e-01
1.53820407e+00 1.30718946e-02 6.30189061e-01 1.00889647e+00
8.68826807e-01 7.79381633e-01 1.14680564e+00 3.81525652e-03
3.97328734e-02 -5.37870765e-01 4.12041873e-01 4.69241232e-01
1.46231696e-03 -7.72127286e-02 -5.61078429e-01 -3.91494006e-01
3.25689077e-01 2.02141687e-01 -2.95781612e-01 -4.53724205e-01
-1.18854284e+00 5.25796652e-01 7.34025180e-01 2.12615311e-01
-5.63070297e-01 -4.22195226e-01 5.54974020e-01 2.92050034e-01
7.06695691e-02 2.64646411e-01 -1.24446666e+00 -1.45285428e-01
-3.42907101e-01 1.93879902e-01 7.23052561e-01 4.94883388e-01
8.03876996e-01 -3.56919080e-01 2.31975988e-01 6.03043497e-01
3.20531845e-01 2.32091382e-01 6.53576136e-01 -2.82116413e-01
1.62453264e-01 8.81863713e-01 1.84320450e-01 -7.91062355e-01
-7.82894969e-01 -5.32058835e-01 -1.01049721e+00 -6.39165342e-02
1.58675969e-01 -4.84837592e-02 -8.93384933e-01 1.62346160e+00
4.12491381e-01 4.52577434e-02 3.23430896e-01 9.25333023e-01
8.88100386e-01 9.56215203e-01 4.05860275e-01 -4.72955585e-01
1.18635857e+00 -6.63696468e-01 -4.89042878e-01 2.85669118e-01
6.61716282e-01 -9.73573327e-01 8.06774199e-01 4.91021752e-01
-5.62875569e-01 -6.26255631e-01 -1.24944389e+00 -2.40589380e-01
-2.43816197e-01 4.38555658e-01 6.52279973e-01 -4.52425964e-02
-6.70725465e-01 9.18778002e-01 -1.00322616e+00 -2.55131394e-01
1.68822542e-01 7.51068115e-01 -5.15110016e-01 4.24509972e-01
-1.50427079e+00 8.78603339e-01 7.00792730e-01 1.20123394e-01
-6.08403027e-01 -8.24155927e-01 -3.53960365e-01 2.86932439e-01
-4.53852005e-02 -3.55333894e-01 1.12373447e+00 -1.00036800e+00
-1.48741555e+00 5.39478183e-01 -2.65476733e-01 -6.47646606e-01
-8.23506266e-02 -2.47223705e-01 -4.18123305e-01 1.50806487e-01
-1.18470296e-01 5.35793006e-01 2.25013077e-01 -8.97761583e-01
-5.76997638e-01 -5.58061659e-01 -1.55084193e-01 2.53015637e-01
1.13114864e-01 7.62514919e-02 3.04102212e-01 -4.34649438e-01
1.54628202e-01 -8.70057940e-01 -3.77643704e-01 -1.77034140e-01
-2.83213943e-01 -4.34112936e-01 6.11749887e-01 -7.96709538e-01
9.56885815e-01 -1.93518090e+00 1.22404978e-01 3.27742770e-02
2.31697306e-01 8.57144952e-01 -1.27928212e-01 9.15389180e-01
-5.98728299e-01 -3.19366753e-01 -2.68499970e-01 3.52675855e-01
-2.44701222e-01 2.09364500e-02 -2.50332147e-01 4.66826707e-01
5.44904351e-01 8.11385155e-01 -5.39123654e-01 -1.45038471e-01
2.43821487e-01 7.43973017e-01 -3.11251432e-01 2.60985941e-01
-3.11347812e-01 4.97777998e-01 -4.96288329e-01 5.01011074e-01
1.04973257e+00 -5.62568009e-01 6.39313161e-01 -3.91866714e-01
-1.56151474e-01 5.55065751e-01 -6.85383976e-01 1.72384965e+00
-3.12233251e-02 1.62224546e-01 -2.38966867e-01 -1.06835628e+00
1.18546295e+00 1.03179343e-01 8.28865886e-01 -5.79443932e-01
3.46568674e-02 2.17901587e-01 7.08890855e-01 -3.27821702e-01
-3.35680842e-02 -8.09417889e-02 3.46814126e-01 1.43738061e-01
1.66817650e-01 7.59815276e-01 3.97764035e-02 1.63034886e-01
1.13917542e+00 3.55029821e-01 5.12818635e-01 -4.74152356e-01
1.00482488e+00 -1.00414827e-01 9.43313658e-01 6.29351009e-03
-2.20365793e-01 2.33601213e-01 7.22552717e-01 -8.73581052e-01
-1.03816020e+00 -8.95183623e-01 -2.98834115e-01 1.01616549e+00
1.25834182e-01 -5.89468777e-01 -7.36421525e-01 -6.58407629e-01
-4.17323746e-02 1.41321912e-01 -3.56322706e-01 -2.68028468e-01
-6.90821588e-01 -1.14584398e+00 2.51110256e-01 5.38302898e-01
4.49553698e-01 -1.42649519e+00 -3.70707840e-01 6.47729993e-01
-2.21798271e-02 -6.71297133e-01 -6.74039945e-02 8.40653181e-01
-9.29540515e-01 -1.42842877e+00 -4.82169539e-01 -1.00759208e+00
2.91714579e-01 1.02203503e-01 8.63195598e-01 -5.07087670e-02
-1.63942292e-01 -7.71417916e-01 -4.86430079e-01 -3.85011494e-01
-2.80235976e-01 2.60018170e-01 6.37847334e-02 -5.35186157e-02
9.99267757e-01 -8.69207919e-01 -9.60522294e-01 3.25703710e-01
-7.59453773e-01 1.47367716e-01 1.10432816e+00 1.12782621e+00
7.31065333e-01 -2.60937572e-01 9.20088708e-01 -7.28533804e-01
3.89337540e-01 -4.59132314e-01 -4.68935013e-01 2.25562409e-01
-5.84893644e-01 2.68079460e-01 1.09657347e+00 -6.84802905e-02
-1.23163664e+00 5.26775360e-01 -8.37089300e-01 1.72843903e-01
-3.61457139e-01 5.71781814e-01 -4.38131839e-01 -1.02422856e-01
5.57317913e-01 6.93190873e-01 1.81129560e-01 -7.84321189e-01
-1.89017653e-02 8.50406289e-01 2.09215462e-01 -3.92108142e-01
1.98363408e-01 1.09816723e-01 2.29983553e-01 -6.89050138e-01
-6.80528820e-01 -6.11830533e-01 -7.91600466e-01 4.88597125e-01
7.44546771e-01 -6.78883493e-01 -1.25285351e+00 8.39917362e-01
-1.24632812e+00 -1.06848720e-02 3.75769764e-01 3.34275335e-01
-5.59188724e-01 9.49910939e-01 -8.14185202e-01 -2.77645677e-01
-9.46274638e-01 -1.45181143e+00 1.00269103e+00 1.41439527e-01
1.06535390e-01 -4.46787208e-01 1.89625993e-01 2.03086361e-01
6.29571527e-02 8.84395167e-02 1.21126437e+00 -9.76713598e-01
-3.39057654e-01 4.04784590e-01 -4.11704391e-01 2.84602404e-01
2.63936192e-01 3.26210521e-02 -6.39257371e-01 -3.20686191e-01
-1.16323330e-01 -3.65600646e-01 9.50875640e-01 3.59529793e-01
1.15700173e+00 -7.85891861e-02 -2.13979393e-01 6.16568208e-01
1.45798147e+00 7.36629844e-01 6.60825729e-01 4.99736488e-01
6.05034232e-01 4.13796246e-01 9.16667461e-01 4.62227076e-01
4.75106314e-02 3.80515099e-01 5.90513647e-01 -6.81667477e-02
2.99723744e-01 2.56319083e-02 4.10599679e-01 8.78659546e-01
1.66252181e-02 -1.57666370e-01 -9.14484739e-01 2.24502102e-01
-1.89811409e+00 -6.55351877e-01 -4.24043328e-01 1.67266476e+00
1.24034441e+00 8.19864273e-02 -1.17271461e-01 -6.60338774e-02
5.86415112e-01 8.59055743e-02 -1.09166420e+00 -5.13330400e-01
-2.04258576e-01 4.61392134e-01 2.45280445e-01 7.74094835e-02
-1.28160727e+00 7.91598380e-01 5.30758047e+00 8.54431331e-01
-1.23382354e+00 -4.51067060e-01 6.19529426e-01 1.87686324e-01
1.92744195e-01 -3.67543884e-02 -9.44864035e-01 7.32028067e-01
1.13624012e+00 2.60291338e-01 4.22945693e-02 8.75510275e-01
2.24879548e-01 2.37167016e-01 -6.86071157e-01 6.92587316e-01
-4.25749660e-01 -1.46913576e+00 1.51796207e-01 -4.85088304e-03
5.32878101e-01 3.42045397e-01 -1.56043962e-01 3.78475315e-03
1.16660431e-01 -8.71805549e-01 -7.64858797e-02 3.05822104e-01
5.82829893e-01 -1.30206954e+00 1.13901567e+00 4.14389461e-01
-1.15008390e+00 2.47683469e-02 -6.51580215e-01 9.12228897e-02
2.21364666e-02 6.81264699e-01 -7.63767302e-01 6.97585821e-01
7.96409070e-01 8.39931190e-01 -4.22714084e-01 6.89053237e-01
-2.20628843e-01 5.65124810e-01 1.58654768e-02 -2.75324851e-01
3.25050920e-01 -3.37617308e-01 1.63708583e-01 1.09067106e+00
-2.13121369e-01 2.58750677e-01 4.26628068e-02 4.15877521e-01
1.57482028e-01 4.34060335e-01 -1.51608400e-02 -1.59383178e-01
-8.15817807e-03 1.24132502e+00 -4.07731086e-01 -2.14108407e-01
-3.82120758e-01 6.98544383e-01 5.32678723e-01 1.05725840e-01
-8.31641316e-01 -7.97575653e-01 1.26083493e+00 -1.84499741e-01
4.52907503e-01 -2.97774494e-01 4.74006981e-02 -1.15709591e+00
-1.52282372e-01 -1.27605724e+00 1.99374899e-01 -6.61624968e-01
-1.46250582e+00 6.16637766e-01 -6.84484720e-01 -8.88485014e-01
1.46024242e-01 -1.06241953e+00 -3.52832556e-01 1.01824534e+00
-1.58639145e+00 -1.23022759e+00 3.76551598e-02 1.96353093e-01
4.25037831e-01 -2.58301854e-01 7.92124510e-01 1.38412371e-01
-5.24035931e-01 4.06189173e-01 7.86684573e-01 -9.90446955e-02
8.77212942e-01 -1.35939479e+00 7.03896582e-01 5.61704576e-01
-3.38247895e-01 1.09986663e+00 5.75051665e-01 -8.35419416e-01
-1.42109740e+00 -1.04825330e+00 9.34951246e-01 -2.11085542e-03
7.09617853e-01 -6.24978915e-02 -1.21719217e+00 3.38837028e-01
2.13200554e-01 -3.27343315e-01 1.04226041e+00 -4.91128340e-02
1.05451494e-02 -7.06809536e-02 -9.00511742e-01 3.90633106e-01
1.04776514e+00 -3.19766551e-01 -5.42364478e-01 1.94465518e-01
9.89733160e-01 -3.06165874e-01 -1.18959188e+00 7.54290938e-01
6.82156265e-01 -1.11135328e+00 9.45099950e-01 -1.04001355e+00
5.85791647e-01 -5.23731947e-01 2.46745907e-02 -1.06594205e+00
-5.67225754e-01 -3.78754735e-01 -2.38060758e-01 7.45236814e-01
3.32846731e-01 -6.08403563e-01 9.87332582e-01 2.73534525e-02
-5.27396798e-01 -1.41547263e+00 -5.68009615e-01 -2.71568507e-01
1.93727419e-01 2.37820148e-01 8.66905391e-01 7.36403763e-01
-1.59527943e-01 5.61292827e-01 -1.90341845e-01 3.07066576e-03
1.90563083e-01 6.17483377e-01 5.59778631e-01 -1.29154110e+00
-2.42744967e-01 -2.51199938e-02 -4.84120220e-01 -1.27256107e+00
2.00513631e-01 -7.04441130e-01 -1.87088802e-01 -1.31236100e+00
4.12591666e-01 -6.56376407e-02 -9.65155005e-01 5.10499537e-01
-3.52995843e-01 -8.49222615e-02 -3.25965375e-01 1.23723879e-01
-5.31460941e-01 7.22630918e-01 1.40870583e+00 -4.15835157e-02
-1.46023661e-01 -2.31852289e-02 -5.18594027e-01 4.48684901e-01
1.16451299e+00 -1.36897862e-01 -3.11192364e-01 5.67091554e-02
2.76792914e-01 -1.71406776e-01 8.60011578e-02 -8.19507658e-01
-7.39645511e-02 -2.42281631e-01 7.97853649e-01 -1.21950316e+00
1.97277233e-01 -7.31044829e-01 1.65940687e-01 9.18810785e-01
-3.48230869e-01 8.38228911e-02 1.68232068e-01 5.88394880e-01
-1.87765062e-01 7.52086118e-02 8.80380929e-01 -2.22021490e-02
-7.67467260e-01 4.04896885e-01 -4.07420516e-01 -4.68731076e-01
9.90252435e-01 3.22605558e-02 -3.80756229e-01 3.59385759e-01
-8.65442097e-01 9.58498493e-02 4.83955353e-01 1.61242366e-01
6.14932179e-01 -9.73284483e-01 -3.49708289e-01 3.27382326e-01
7.38261566e-02 -1.29841864e-01 5.32900333e-01 7.90879726e-01
-1.02077723e+00 8.36063683e-01 -7.46066332e-01 -6.81590140e-01
-1.40119874e+00 7.31658340e-01 2.83453584e-01 -3.60910505e-01
-3.87918413e-01 5.74956000e-01 4.38424766e-01 -5.57565212e-01
2.01281090e-03 -4.24254298e-01 -5.13221383e-01 -1.61864713e-01
2.99737424e-01 1.44938961e-01 3.70279908e-01 -7.69939840e-01
-5.25840044e-01 4.36907053e-01 -5.81841588e-01 8.57274115e-01
1.57614911e+00 1.12296559e-01 -4.62151200e-01 -4.06758115e-02
1.35980010e+00 -4.82297242e-01 -1.33007169e+00 -1.80184782e-01
1.30689830e-01 4.05930765e-02 -4.38566148e-01 -1.24155235e+00
-6.47612810e-01 1.04942274e+00 7.64982939e-01 -3.05531174e-01
1.14265716e+00 -3.16283137e-01 1.34583807e+00 7.30771542e-01
3.20301056e-01 -7.40849435e-01 -2.57216901e-01 6.35908902e-01
5.39158165e-01 -1.31361866e+00 -7.18328506e-02 -1.82707667e-01
-6.06886685e-01 1.19027126e+00 8.09639275e-01 -1.25586525e-01
4.20603424e-01 4.31640334e-02 1.25990480e-01 -1.71490416e-01
-1.02923882e+00 8.44756961e-02 2.25092351e-01 5.95061719e-01
7.73079216e-01 1.41416006e-02 -9.09580708e-01 8.19869637e-01
-2.41610706e-02 -9.52167735e-02 1.56073216e-02 1.01355684e+00
-6.54013395e-01 -1.48733211e+00 1.63018256e-01 4.22239453e-01
-8.46743286e-01 -2.65332222e-01 -8.70645404e-01 4.25103009e-01
1.80614457e-01 6.19211912e-01 -3.21816057e-01 -4.31120455e-01
-1.46783954e-02 8.02141652e-02 2.08060622e-01 -2.61138469e-01
-8.07855606e-01 -9.53506865e-03 1.82223152e-02 -6.07074678e-01
-4.04798955e-01 -4.96842653e-01 -1.66027975e+00 -2.21971765e-01
-4.13399369e-01 4.98922020e-01 6.64669096e-01 7.95076787e-01
1.01644099e+00 5.54741144e-01 5.92630684e-01 -4.88549441e-01
-7.68156469e-01 -1.03697491e+00 -4.45127934e-01 2.89269686e-01
3.09096575e-01 -4.85677212e-01 1.37221605e-01 -3.80952246e-02] | [4.699323654174805, 5.60764741897583] |
ed105c48-38c3-4c75-8682-25c94e11ea10 | artificial-perceptual-learning-image | 2106.07559 | null | https://arxiv.org/abs/2106.07559v1 | https://arxiv.org/pdf/2106.07559v1.pdf | Artificial Perceptual Learning: Image Categorization with Weak Supervision | Machine learning has achieved much success on supervised learning tasks with large sets of well-annotated training samples. However, in many practical situations, such strong and high-quality supervision provided by training data is unavailable due to the expensive and labor-intensive labeling process. Automatically identifying and recognizing object categories in a large volume of unlabeled images with weak supervision remains an important, yet unsolved challenge in computer vision. In this paper, we propose a novel machine learning framework, artificial perceptual learning (APL), to tackle the problem of weakly supervised image categorization. The proposed APL framework is constructed using state-of-the-art machine learning algorithms as building blocks to mimic the cognitive development process known as infant categorization. We develop and illustrate the proposed framework by implementing a wide-field fine-grain ecological survey of tree species over an 8,000-hectare area of the El Yunque rainforest in Puerto Rico. It is based on unlabeled high-resolution aerial images of the tree canopy. Misplaced ground-based labels were available for less than 1% of these images, which serve as the only weak supervision for this learning framework. We validate the proposed framework using a small set of images with high quality human annotations and show that the proposed framework attains human-level cognitive economy. | ['Tian Zheng', 'Douglas C. Morton', 'Helen Jin', 'María Uriarte', 'Chengliang Tang'] | 2021-06-02 | null | null | null | null | ['image-categorization'] | ['computer-vision'] | [ 6.57922506e-01 2.37308383e-01 -2.96000749e-01 -4.94337231e-01
-3.56075078e-01 -6.88365757e-01 4.02512938e-01 4.42294776e-01
-6.36606276e-01 7.12164581e-01 -3.79543841e-01 -3.71570468e-01
-2.66879618e-01 -7.24455416e-01 -8.29869568e-01 -4.85855103e-01
-1.12478301e-01 3.45816463e-01 2.25985005e-01 5.40203303e-02
7.02816471e-02 2.83653021e-01 -2.08875299e+00 3.17057908e-01
1.01667762e+00 8.42087090e-01 8.69719505e-01 5.04170716e-01
1.00147590e-01 8.17399859e-01 -3.32684249e-01 -3.47243786e-01
3.12820882e-01 3.28589766e-03 -8.73361111e-01 4.82684046e-01
8.09137940e-01 -3.67988437e-01 2.67928421e-01 1.27926779e+00
1.86563700e-01 -2.18276709e-01 7.08796084e-01 -1.10031271e+00
-7.40883470e-01 5.22253871e-01 -6.36754632e-01 1.42795041e-01
-2.87467957e-01 -2.06503049e-01 1.22678232e+00 -8.60210538e-01
4.10072148e-01 1.07342637e+00 6.09264851e-01 3.28363985e-01
-1.34194434e+00 -6.32303119e-01 3.06423485e-01 3.00207973e-01
-1.38925123e+00 -2.93676287e-01 6.93992615e-01 -7.33097136e-01
4.32556957e-01 -2.73607243e-02 5.05347133e-01 6.33231282e-01
-2.82564998e-01 6.81025922e-01 1.44830108e+00 -6.58730626e-01
4.32958215e-01 3.13477665e-01 3.30846936e-01 9.63712096e-01
3.69443953e-01 2.27830246e-01 -2.96426147e-01 1.64884664e-02
7.37512112e-01 9.58522931e-02 3.84305976e-02 -4.84263569e-01
-8.45907927e-01 7.66820252e-01 7.87117243e-01 2.91981697e-01
-3.95719260e-01 -3.01661551e-01 6.38834015e-02 3.49703848e-01
7.42552578e-01 2.24377289e-01 -5.41283011e-01 4.28295016e-01
-9.39224839e-01 -1.69741675e-01 5.49584448e-01 6.88803196e-01
8.68089616e-01 -9.73655954e-02 3.60712439e-01 1.23470724e+00
1.91707462e-01 7.69748032e-01 3.26027900e-01 -9.02744591e-01
1.20826513e-01 7.28949189e-01 2.34944776e-01 -9.91201758e-01
-3.46119076e-01 -5.56923628e-01 -7.94811785e-01 5.29889405e-01
5.55082440e-01 1.89334843e-02 -1.00775290e+00 1.87507915e+00
4.53933299e-01 -1.03852764e-01 3.90462577e-02 8.55741441e-01
7.19875872e-01 3.50998998e-01 3.08707774e-01 -2.66892195e-01
1.22937500e+00 -1.02064955e+00 -2.60116071e-01 -4.81401503e-01
2.41523355e-01 -4.00758773e-01 1.18212593e+00 4.57421869e-01
-6.04495347e-01 -8.21441174e-01 -9.90425944e-01 3.48865867e-01
-4.54073429e-01 6.76797032e-01 7.86236584e-01 5.13653457e-01
-9.02535081e-01 4.12092268e-01 -7.05949187e-01 -6.24950588e-01
6.59716785e-01 3.26993525e-01 -6.25163972e-01 -3.37558836e-01
-6.60102725e-01 6.35865211e-01 4.97920632e-01 1.61127150e-01
-1.20562160e+00 -3.13323200e-01 -7.31862724e-01 1.46637067e-01
5.49832702e-01 5.67363314e-02 1.29581201e+00 -1.48687541e+00
-1.02473092e+00 1.44951260e+00 6.93179965e-02 -2.87043959e-01
1.44986406e-01 -3.09406191e-01 -1.46176562e-01 3.46547663e-01
5.01525283e-01 8.11201692e-01 1.04117346e+00 -1.50497317e+00
-8.45431268e-01 -8.09316099e-01 1.79830104e-01 7.35990703e-02
-5.46481788e-01 -1.07891738e-01 1.97542623e-01 -6.56828165e-01
2.64490545e-01 -9.45848107e-01 -1.33587465e-01 1.48212790e-01
1.90766081e-02 -1.40455320e-01 7.29894757e-01 -6.19003892e-01
7.22050905e-01 -2.15389442e+00 -3.33283357e-02 -1.75638735e-01
1.74761623e-01 4.26970869e-01 -2.57096678e-01 8.51460397e-02
-4.74884212e-02 -1.33753702e-01 -5.82439482e-01 8.73497576e-02
-3.09598893e-01 5.08891642e-01 -2.09236965e-01 3.12038243e-01
2.15796456e-01 3.90008837e-01 -1.08535063e+00 -4.95822191e-01
2.26550326e-01 -5.63912690e-02 -3.32605690e-01 5.15985668e-01
-2.82617390e-01 4.23954844e-01 -2.95835227e-01 9.74341512e-01
5.68450689e-01 -2.27409005e-01 2.92333990e-01 1.10166699e-01
-2.26318270e-01 -4.49062943e-01 -1.00648785e+00 1.61035478e+00
-3.30539733e-01 3.47489774e-01 4.39114571e-01 -1.55965316e+00
8.97545218e-01 1.35822237e-01 1.99520588e-01 -3.80895823e-01
1.10093635e-02 2.14825764e-01 2.02047795e-01 -6.04150772e-01
7.32484087e-02 -2.36930430e-01 7.29413182e-02 3.81333798e-01
4.36444074e-01 -3.64957780e-01 2.70632297e-01 -1.61128610e-01
1.02636647e+00 1.81665614e-01 6.15292490e-01 -4.89461035e-01
6.44820213e-01 3.93691510e-01 6.90324724e-01 8.28049421e-01
-4.41689014e-01 2.26117730e-01 -6.73148930e-02 -5.18109977e-01
-1.09015429e+00 -1.07995999e+00 -4.58714545e-01 1.67242432e+00
1.76584267e-03 1.16334669e-01 -9.13807034e-01 -8.03529680e-01
-1.20096043e-01 2.47270226e-01 -6.56589568e-01 1.20112039e-01
-1.31717578e-01 -7.87271678e-01 3.81907612e-01 4.82401341e-01
8.60461116e-01 -1.37310159e+00 -7.40530729e-01 1.57368913e-01
4.94648553e-02 -1.08600271e+00 3.33503038e-01 4.31032956e-01
-8.12515140e-01 -1.32021415e+00 -7.05925465e-01 -1.10578382e+00
9.20550287e-01 6.03549004e-01 1.10431194e+00 1.51623845e-01
-4.39309269e-01 3.21197122e-01 -5.91599107e-01 -6.46273911e-01
-3.12958539e-01 -2.21057348e-02 1.92515701e-01 3.64970602e-03
4.55799520e-01 -6.69160485e-01 -2.05114454e-01 3.99568498e-01
-7.39158750e-01 4.28692065e-02 8.88085186e-01 1.21408832e+00
5.77859640e-01 4.00486499e-01 8.52135837e-01 -9.08612967e-01
8.94686207e-02 -5.30869246e-01 -8.63670111e-01 4.37926650e-01
-3.68897438e-01 -1.13508686e-01 7.80987978e-01 -4.95169699e-01
-1.18359423e+00 4.10578012e-01 8.78333598e-02 -2.78657898e-02
-8.46080959e-01 4.47254837e-01 -1.51370779e-01 -1.46080628e-01
7.38488019e-01 -8.47967491e-02 -3.21360141e-01 -6.37120426e-01
3.48835766e-01 1.10863483e+00 1.02873206e+00 -8.58701885e-01
8.46311569e-01 4.54979509e-01 -1.21361978e-01 -1.13484275e+00
-1.53882563e+00 -5.62830448e-01 -1.18363118e+00 -2.91566998e-01
7.98016846e-01 -1.00171185e+00 -3.87291729e-01 5.48005462e-01
-7.19777882e-01 -5.14450133e-01 -2.02949643e-01 5.44753194e-01
-5.16317248e-01 1.46830216e-01 -3.35818797e-01 -9.92577434e-01
-3.00741762e-01 -7.16052949e-01 1.09162593e+00 2.59847611e-01
2.72966981e-01 -6.89543843e-01 -3.14423628e-02 7.15540051e-01
4.60074209e-02 8.60409066e-02 1.11196768e+00 -5.45737505e-01
-2.39260897e-01 6.48273602e-02 -4.82255995e-01 7.53569007e-01
2.76586533e-01 -2.77642328e-02 -1.19744051e+00 -3.57175946e-01
7.39238188e-02 -9.12331104e-01 8.08890939e-01 2.21791953e-01
1.12070072e+00 7.01521486e-02 -7.02923313e-02 2.48407811e-01
1.51046777e+00 2.14751109e-01 -1.01160534e-01 8.91355798e-02
6.53609931e-01 1.01498020e+00 8.75461400e-01 2.98572242e-01
2.53539562e-01 2.81855553e-01 5.28985381e-01 -3.20964932e-01
2.05239886e-03 -2.20572293e-01 4.10870574e-02 7.18441010e-01
-2.51829892e-01 6.94086328e-02 -1.03499234e+00 7.01299667e-01
-1.83350801e+00 -8.82548094e-01 8.53396058e-02 2.24872351e+00
7.86897719e-01 1.71156958e-01 -9.83798355e-02 5.47480762e-01
8.71030807e-01 -1.19504720e-01 -5.47312081e-01 -6.86842203e-03
-6.45364672e-02 3.30483824e-01 3.10293943e-01 3.10067236e-01
-1.46341479e+00 1.19960797e+00 5.89378023e+00 6.44560814e-01
-1.01717925e+00 9.26845670e-02 4.42343384e-01 4.97841060e-01
4.00142968e-01 -2.10555866e-01 -4.45914328e-01 2.23573834e-01
6.52892351e-01 4.03683513e-01 5.67726612e-01 1.14302766e+00
1.16017468e-01 -3.25432897e-01 -1.00908458e+00 7.75394976e-01
1.12003118e-01 -7.58790851e-01 -3.03271085e-01 -1.53500855e-01
7.05173016e-01 2.36748555e-03 -1.89231426e-01 3.01892430e-01
5.67817152e-01 -8.91922951e-01 6.85469866e-01 8.82630050e-03
9.35808599e-01 -4.56586301e-01 7.17276216e-01 8.05726409e-01
-1.18873644e+00 -5.95728338e-01 -6.67560518e-01 -4.69211400e-01
-4.88476455e-01 5.75825691e-01 -6.54010832e-01 1.37186900e-01
1.11789668e+00 7.01581776e-01 -9.17954445e-01 7.80104518e-01
-5.53807378e-01 1.01338112e+00 -2.99749285e-01 2.73083866e-01
3.45605046e-01 -1.33025020e-01 4.87432741e-02 7.64951169e-01
-1.25429081e-03 4.88511294e-01 5.22964597e-01 5.71434259e-01
-2.20351160e-01 2.11476803e-01 -8.49630952e-01 -9.32299569e-02
4.10661489e-01 1.30839264e+00 -9.11973059e-01 -3.33877981e-01
-5.27331293e-01 6.23749971e-01 7.49672771e-01 9.24744159e-02
-2.86174119e-01 -2.41580270e-02 5.69762364e-02 -1.03843778e-01
1.66911960e-01 -7.51560330e-02 -1.79632992e-01 -1.16013551e+00
-1.53327331e-01 -6.86821580e-01 3.95964295e-01 -7.83944964e-01
-1.45615196e+00 4.97038633e-01 1.12299919e-01 -9.33012486e-01
-2.89582312e-02 -6.84403658e-01 -3.52994472e-01 4.75048959e-01
-1.50089633e+00 -1.52854335e+00 -7.75357187e-01 5.08039474e-01
7.71397293e-01 -4.65624362e-01 1.02988636e+00 2.42865339e-01
-4.23633456e-01 1.01551868e-01 1.29783332e-01 7.92577267e-02
5.72602391e-01 -1.40016448e+00 -2.91601807e-01 8.34791303e-01
4.05376911e-01 1.36512563e-01 4.22901154e-01 -5.05238295e-01
-8.45540464e-01 -1.14533532e+00 6.42480195e-01 -9.58124325e-02
6.37913704e-01 -3.98287117e-01 -9.67139721e-01 5.76382697e-01
-1.14335462e-05 2.49761760e-01 7.87885666e-01 1.15169480e-01
-3.04405689e-01 -2.37632975e-01 -1.36992800e+00 1.07125394e-01
1.08266079e+00 -4.90022928e-01 -9.90325451e-01 5.91361582e-01
3.75279635e-01 2.27521226e-01 -7.64532208e-01 6.22616827e-01
4.45940495e-01 -8.17088187e-01 8.52168977e-01 -5.21037459e-01
5.08567333e-01 -3.91840488e-01 -4.93733615e-01 -1.17553854e+00
-3.96454692e-01 1.88849345e-01 3.66409123e-01 1.31488717e+00
4.64295037e-02 -1.85840935e-01 8.63961816e-01 2.08524495e-01
-5.51401861e-02 -3.76730353e-01 -5.58421075e-01 -6.95409775e-01
-3.55955260e-03 -3.35595936e-01 1.28661692e-01 9.86170471e-01
-3.14450681e-01 5.83899498e-01 -2.35699803e-01 4.60351378e-01
1.07458746e+00 4.73076403e-01 6.58759475e-01 -1.72967064e+00
-2.67953753e-01 8.64001364e-02 -5.17842352e-01 -5.61237156e-01
5.38150907e-01 -7.96898782e-01 3.10187340e-01 -1.19743288e+00
4.60897833e-01 -6.96718276e-01 -2.37204343e-01 8.13741803e-01
-1.94798782e-01 4.86461639e-01 5.58067188e-02 3.96133244e-01
-3.71582091e-01 4.11405116e-01 8.71190786e-01 -4.62523729e-01
1.10738479e-01 1.49163201e-01 -4.76553500e-01 1.27150321e+00
7.80127823e-01 -5.83845079e-01 -6.07497036e-01 -5.80408514e-01
-5.02613127e-01 -9.94638205e-02 3.89211148e-01 -1.05221653e+00
-7.42928609e-02 -4.60807890e-01 3.49004209e-01 -5.08221328e-01
1.82172745e-01 -1.18628848e+00 -1.46367729e-01 4.81609046e-01
-3.50058973e-01 -3.90733033e-01 -6.71232268e-02 4.29877788e-01
-3.36632729e-01 -5.10850430e-01 1.04566240e+00 -4.16071385e-01
-1.12250412e+00 1.68077677e-01 -1.66189834e-01 -3.13689299e-02
1.17857695e+00 1.01395771e-01 -2.80059397e-01 3.03055365e-02
-8.35416555e-01 5.49508892e-02 5.18716216e-01 2.52321184e-01
4.09997016e-01 -9.61338818e-01 -7.60165155e-01 3.48954022e-01
4.11023647e-01 4.32500243e-02 -2.93448195e-02 1.35421172e-01
-4.86471921e-01 2.83081919e-01 -7.15255320e-01 -8.08334351e-01
-1.44585395e+00 7.31799781e-01 -5.02750650e-02 -2.83803549e-02
-3.71133894e-01 6.59898400e-01 4.64325964e-01 -6.77089751e-01
2.50319391e-01 -1.31013393e-01 -5.10020852e-01 -2.57008281e-02
3.95826548e-01 1.20064832e-01 -6.03481941e-02 -8.41681957e-01
-1.16521902e-01 5.75096786e-01 2.72382945e-01 8.41959342e-02
1.55395842e+00 -1.00021817e-01 -1.19001888e-01 4.07264292e-01
6.71770096e-01 -3.63013953e-01 -1.15709507e+00 -3.86613756e-01
2.71263003e-01 -4.78302538e-01 7.53081366e-02 -8.52642596e-01
-9.80901420e-01 1.12691975e+00 9.47451770e-01 1.71026364e-01
1.32020175e+00 7.33098090e-02 1.15372144e-01 9.13837731e-01
7.57889152e-01 -1.14958620e+00 -4.32105409e-03 1.93514839e-01
6.99537873e-01 -1.85300219e+00 -2.60828026e-02 -5.18510044e-01
-4.03237283e-01 8.25453877e-01 8.76951158e-01 -6.01319484e-02
6.31247163e-01 -5.26320748e-02 1.59110144e-01 -9.80261862e-02
-6.13306999e-01 -5.44041216e-01 1.02380179e-01 8.65659952e-01
2.49591246e-01 3.48359197e-01 -3.14036787e-01 5.66814482e-01
-1.51234623e-02 6.59644902e-02 3.30001712e-01 1.05520988e+00
-8.43044698e-01 -9.41929817e-01 -4.87135082e-01 5.60734570e-01
-4.96031433e-01 -8.32362697e-02 -3.28802764e-01 6.02479219e-01
6.14754379e-01 1.21315706e+00 -1.09560490e-01 -2.66588479e-01
8.91779140e-02 -1.77443489e-01 4.67161685e-01 -9.24626529e-01
-3.49522054e-01 -8.91053006e-02 -3.03072054e-02 -1.33598417e-01
-9.12312686e-01 -4.18625295e-01 -8.79204512e-01 1.20229006e-01
-2.75604784e-01 2.55072713e-01 6.87793255e-01 9.09941435e-01
-2.46028766e-01 8.31760466e-02 8.40163231e-01 -9.48740065e-01
-5.04185319e-01 -1.08887255e+00 -8.73810530e-01 5.07601798e-01
1.47848189e-01 -9.27997231e-01 -2.48206571e-01 5.29993713e-01] | [9.660857200622559, 2.4263370037078857] |
96104e20-e0e9-4767-9f51-6897246a5ea7 | autosamp-autoencoding-mri-sampling-via | 2306.02888 | null | https://arxiv.org/abs/2306.02888v2 | https://arxiv.org/pdf/2306.02888v2.pdf | AutoSamp: Autoencoding MRI Sampling via Variational Information Maximization | Accelerated MRI protocols routinely involve a predefined sampling pattern that undersamples the k-space. Finding an optimal pattern can enhance the reconstruction quality, however this optimization is a challenging task. To address this challenge, we introduce a novel deep learning framework, AutoSamp, based on variational information maximization that enables joint optimization of sampling pattern and reconstruction of MRI scans. We represent the encoder as a non-uniform Fast Fourier Transform that allows continuous optimization of k-space sample locations on a non-Cartesian plane, and the decoder as a deep reconstruction network. Experiments on public MRI datasets show improved reconstruction quality of the proposed AutoSamp method over the prevailing variable density and variable density Poisson disc sampling. We demonstrate that our data-driven sampling optimization method achieves 4.4dB, 2.0dB, 0.75dB, 0.7dB PSNR improvements over reconstruction with Poisson Disc masks for acceleration factors of R = 5, 10, 15, 25, respectively. Furthermore, we analyze the characteristics of the learned sampling patterns with respect to changes in acceleration factor, measurement noise, underlying anatomy, and coil sensitivities. We show that all these factors contribute to the optimization result by affecting the sampling density, k-space coverage and point spread functions of the learned sampling patterns. | ['John M. Pauly', 'Shreyas S. Vasanawala', 'Morteza Mardani', 'Cagan Alkan'] | 2023-06-05 | null | null | null | null | ['anatomy'] | ['miscellaneous'] | [ 1.89826444e-01 1.47740180e-02 -1.90338552e-01 -4.08270597e-01
-1.11815965e+00 -3.73727977e-01 1.75251275e-01 -5.55478968e-02
-6.06114209e-01 7.86159575e-01 5.30216634e-01 -1.74392149e-01
-4.13509846e-01 -5.33442438e-01 -1.02421308e+00 -1.12349713e+00
-2.62154460e-01 3.31218779e-01 1.07860856e-01 3.26830357e-01
-1.69646256e-02 4.22462553e-01 -6.62524521e-01 1.89670339e-01
7.16353714e-01 1.12641644e+00 4.30817306e-01 4.48971182e-01
1.90853268e-01 6.65057123e-01 -4.31829602e-01 -1.85609847e-01
2.91023940e-01 -1.12547755e-01 -6.72742307e-01 -5.56064732e-02
-1.23374805e-01 -5.87042570e-01 -5.91664970e-01 9.86272335e-01
8.17974627e-01 3.78647670e-02 8.40854704e-01 -6.35142684e-01
-3.35395753e-01 8.85049582e-01 -7.68336475e-01 4.19675678e-01
-3.90987277e-01 2.79024452e-01 4.36306059e-01 -5.37779808e-01
4.26033765e-01 7.09451556e-01 8.04218709e-01 3.93767506e-01
-1.44544697e+00 -6.17108583e-01 -4.34190422e-01 1.52724981e-01
-1.66164196e+00 -1.97055727e-01 7.59660661e-01 -5.18809378e-01
4.15738523e-01 2.32360467e-01 4.94218022e-01 8.26252520e-01
7.57027328e-01 5.39679110e-01 1.16324317e+00 -7.96257257e-02
4.37240601e-01 -7.08284974e-02 -1.22473232e-01 3.52458090e-01
6.39988706e-02 3.17310989e-02 -2.61004895e-01 -2.85822332e-01
1.34235144e+00 -9.28624496e-02 -6.22500539e-01 -4.78674948e-01
-1.16094148e+00 7.70315528e-01 6.82487845e-01 1.33769467e-01
-7.69934833e-01 3.27286929e-01 4.62497532e-01 -2.30603024e-01
1.97576970e-01 5.50512075e-01 -2.03431517e-01 -4.27282304e-02
-1.03944468e+00 1.09763712e-01 4.00411695e-01 8.19471002e-01
2.57606208e-01 1.22091681e-01 -5.95874727e-01 9.61464524e-01
2.57159561e-01 5.75552225e-01 3.68918389e-01 -1.02114737e+00
1.77899361e-01 -9.56560895e-02 5.37255332e-02 -9.18551266e-01
-5.44839561e-01 -8.67615461e-01 -1.13138461e+00 -2.73590982e-01
2.58844763e-01 -3.72583717e-01 -7.77148843e-01 1.73004401e+00
5.01529276e-01 1.69361144e-01 -2.54043758e-01 1.29871047e+00
5.63527882e-01 5.21833360e-01 -8.19994211e-02 -5.50006390e-01
1.33778918e+00 -5.81744373e-01 -9.77719724e-01 9.47799981e-02
1.86531514e-01 -6.85677052e-01 9.15769935e-01 4.23514277e-01
-1.32172990e+00 -1.59752071e-01 -9.64230359e-01 3.37487966e-01
7.01270282e-01 1.25159040e-01 3.87752116e-01 5.64031959e-01
-8.76607358e-01 6.29710734e-01 -9.99515951e-01 4.87513155e-01
8.14986110e-01 4.41351503e-01 5.21632582e-02 -2.35320166e-01
-1.01384223e+00 6.42808855e-01 2.99239643e-02 -1.41177058e-01
-1.12057924e+00 -1.27343810e+00 -4.73900676e-01 -1.25380382e-01
3.28250229e-01 -4.65219170e-01 1.10206044e+00 -4.37855661e-01
-1.34333825e+00 4.42894965e-01 7.51693174e-02 -7.53867507e-01
5.83362818e-01 -1.61567748e-01 -1.24237783e-01 4.23985749e-01
9.71797481e-02 3.93035889e-01 8.13910484e-01 -1.00924981e+00
1.29577756e-01 -5.60816050e-01 -2.75658935e-01 2.37546220e-01
6.15569279e-02 -2.40305558e-01 -3.99926990e-01 -5.70608497e-01
1.44146934e-01 -8.08134735e-01 -6.20288372e-01 1.50903523e-01
-5.23091972e-01 4.08513159e-01 1.49707615e-01 -8.99134517e-01
1.03476262e+00 -2.26058865e+00 2.09640816e-01 2.79698312e-01
4.72247481e-01 -3.12667549e-01 2.44479850e-01 -1.89101025e-01
1.54394448e-01 -1.71479136e-01 -4.20748144e-01 -3.60189602e-02
-4.14475769e-01 1.33040935e-01 1.04338564e-01 1.01984775e+00
-1.69365346e-01 6.81458890e-01 -8.65485668e-01 -4.12484348e-01
3.14080268e-01 8.34421694e-01 -8.81419659e-01 2.03752488e-01
2.34626159e-01 8.10977519e-01 -4.08989042e-01 3.59122843e-01
1.04631126e+00 -4.09136802e-01 3.25300962e-01 -8.18183184e-01
-3.78577970e-02 4.28111851e-02 -1.03238010e+00 1.82489884e+00
-5.13406634e-01 3.29505235e-01 3.15399170e-01 -1.03656805e+00
6.68133438e-01 1.72634184e-01 9.38901603e-01 -7.76547194e-01
2.61606663e-01 1.38279587e-01 1.69804543e-01 -3.95535022e-01
1.99062914e-01 -4.84601378e-01 1.73545539e-01 3.06441456e-01
-2.67253388e-02 -4.52912711e-02 -4.28972661e-01 1.03209138e-01
1.06673610e+00 -4.64853317e-01 -7.73833543e-02 -7.11622059e-01
1.19263984e-01 -3.17623019e-01 5.18062174e-01 9.88769293e-01
-2.48752147e-01 8.99635553e-01 5.30545890e-01 -3.77523094e-01
-1.44811428e+00 -1.33731782e+00 -8.17014396e-01 3.37589949e-01
2.10738525e-01 -3.45958844e-02 -7.71825254e-01 -3.59246522e-01
-2.52877742e-01 6.76023960e-01 -5.46954572e-01 -2.27686137e-01
-7.17037618e-01 -1.13293636e+00 4.49549943e-01 2.70502985e-01
5.25322318e-01 -4.80338335e-01 -7.76321232e-01 3.80477905e-01
-3.88820171e-01 -1.19014883e+00 -7.02202082e-01 1.95053875e-01
-8.72003376e-01 -7.52580702e-01 -9.25031900e-01 -3.27924222e-01
6.56757593e-01 -1.10686757e-01 1.05549860e+00 -3.52444291e-01
-4.93366301e-01 2.79804409e-01 -1.68807551e-01 -1.19161606e-01
-3.05570871e-01 -5.62970974e-02 -1.45664951e-02 -6.28169030e-02
-2.20734179e-01 -7.72750378e-01 -1.26302350e+00 3.52369428e-01
-8.95509958e-01 8.58110264e-02 7.21631944e-01 9.18054342e-01
9.50319469e-01 1.17537923e-01 2.83921957e-01 -5.23757160e-01
4.51701343e-01 -7.19408453e-01 -4.41210657e-01 1.82991935e-04
-3.42802972e-01 4.84592855e-01 4.09719259e-01 -4.91165072e-01
-7.36437023e-01 1.63485818e-02 -2.73699909e-01 -6.76649868e-01
2.31300592e-01 3.29473674e-01 1.00320593e-01 -1.93331078e-01
8.08713555e-01 6.45939291e-01 3.93050343e-01 -2.90271312e-01
2.42216036e-01 4.57894593e-01 3.54009867e-01 -6.12755120e-01
2.97131270e-01 8.43725026e-01 6.21121377e-02 -7.83258200e-01
-5.68200409e-01 -2.06851900e-01 -1.68965906e-01 -3.88134778e-01
9.40520704e-01 -7.83820570e-01 -5.62160075e-01 3.51466775e-01
-8.71160626e-01 -2.00517565e-01 -4.00514454e-01 7.35113084e-01
-4.85089242e-01 3.83459896e-01 -6.89224780e-01 -6.91107929e-01
-7.61102855e-01 -1.87625635e+00 1.01368737e+00 -1.04073711e-01
7.36995712e-02 -6.47652566e-01 -2.11344972e-01 2.89684474e-01
7.97637820e-01 3.27490509e-01 9.27720308e-01 -7.35399574e-02
-7.82189488e-01 1.83058053e-01 -1.76969543e-01 4.03207839e-01
8.47060382e-02 -7.20788419e-01 -4.60992545e-01 -3.69103044e-01
5.48923910e-01 -6.58464581e-02 4.47602898e-01 1.26158476e+00
1.68121850e+00 -4.58648980e-01 -1.34731874e-01 1.12267399e+00
1.49334717e+00 2.52090305e-01 7.54813433e-01 7.21293362e-03
5.77248216e-01 1.18863717e-01 2.36860290e-01 8.36529911e-01
1.66935146e-01 8.59211683e-01 4.70634013e-01 8.62784684e-02
-1.36043400e-01 2.32410431e-02 -2.53495097e-01 7.45498240e-01
1.15026571e-01 2.83601195e-01 -9.11273539e-01 4.95112240e-01
-1.22360420e+00 -5.36161304e-01 7.96772614e-02 2.29202604e+00
1.12555826e+00 -2.14488104e-01 1.09728619e-01 -1.32626995e-01
5.78171492e-01 -2.28339825e-02 -9.36807454e-01 1.82455659e-01
6.63800910e-02 1.35469675e-01 1.13429296e+00 5.14342904e-01
-8.31918776e-01 2.51737177e-01 6.18568850e+00 1.08323157e+00
-1.35011196e+00 5.40218174e-01 1.07092798e+00 -6.30986691e-01
-4.77780908e-01 -4.91029561e-01 -5.39863050e-01 8.33983302e-01
8.23929608e-01 8.21176171e-02 7.45083153e-01 7.46401131e-01
3.20778340e-01 1.38153769e-02 -6.36322975e-01 1.23108268e+00
-2.38615781e-01 -1.61675155e+00 -1.63316786e-01 2.88850397e-01
5.48538804e-01 2.71994323e-01 2.47196779e-01 4.19967808e-03
-4.19880077e-03 -1.20524621e+00 6.64147437e-01 4.09452975e-01
1.09779930e+00 -9.60376859e-01 5.21917045e-01 2.16330603e-01
-5.58697164e-01 1.19691372e-01 -3.45610589e-01 4.89030898e-01
4.21964288e-01 1.16647124e+00 -9.70812738e-01 3.15283895e-01
7.12138116e-01 1.14667572e-01 1.19489275e-01 1.02410674e+00
2.76527882e-01 6.64683104e-01 -5.24485469e-01 8.02755952e-02
1.25263512e-01 -6.42593205e-02 7.04983413e-01 1.03873575e+00
2.99035013e-01 2.60739714e-01 -1.03820994e-01 1.07839203e+00
2.92131435e-02 -2.21585054e-02 -1.75989375e-01 4.06255960e-01
6.26388311e-01 1.04532778e+00 -5.97425818e-01 1.36182606e-01
-5.86610436e-02 6.95298553e-01 1.64247513e-01 2.97738701e-01
-1.16973937e+00 3.94604169e-02 6.02143049e-01 5.91047823e-01
2.02679425e-01 -3.45800847e-01 -4.29594547e-01 -9.54312086e-01
1.07375368e-01 -8.97505879e-01 1.42942676e-02 -3.54277581e-01
-1.06411958e+00 5.60557127e-01 3.00446600e-01 -8.45576704e-01
2.89080627e-02 -2.23447770e-01 -1.46089688e-01 9.16080177e-01
-1.09685314e+00 -4.50805932e-01 -1.18960932e-01 3.93234551e-01
3.31727356e-01 -6.03932235e-03 2.69099891e-01 6.22679532e-01
-2.08411410e-01 7.19126880e-01 3.60134989e-01 6.61899149e-02
1.44288138e-01 -9.32037413e-01 1.92832388e-02 4.89758521e-01
-2.74786443e-01 4.80212688e-01 7.99252808e-01 -5.13934076e-01
-1.66449797e+00 -1.08571637e+00 -9.56441015e-02 -4.66636270e-02
4.41336453e-01 -3.44535291e-01 -8.31656992e-01 1.77296072e-01
-6.95980564e-02 2.41765440e-01 6.74848318e-01 -2.50860035e-01
1.76215187e-01 -2.43024707e-01 -1.64744163e+00 4.77905512e-01
8.52133632e-01 -2.16562599e-01 1.10229149e-01 5.80736339e-01
8.48048270e-01 -8.80697727e-01 -1.32310450e+00 5.41290283e-01
4.97627079e-01 -7.73211598e-01 1.28915095e+00 -1.46977305e-01
5.97306430e-01 2.47112494e-02 -4.76617455e-01 -1.31827128e+00
-4.71159130e-01 -4.42343622e-01 -3.28187868e-02 5.28753519e-01
2.62320817e-01 -3.52618068e-01 8.63305748e-01 5.13844907e-01
-2.86183476e-01 -1.29009759e+00 -1.53784192e+00 -5.00076890e-01
3.11070323e-01 -5.53090155e-01 5.87769210e-01 7.12483168e-01
-3.90696645e-01 -5.50685972e-02 -5.53652823e-01 3.77957195e-01
1.16799057e+00 -2.91610122e-01 2.18156293e-01 -3.66663963e-01
-7.72749305e-01 -2.33984753e-01 -2.96692729e-01 -1.32431507e+00
-3.65652531e-01 -7.00664222e-01 -1.17060937e-01 -1.35061193e+00
3.77263933e-01 -7.36872971e-01 -3.12766492e-01 -6.67316765e-02
2.60898676e-02 3.18915062e-02 -1.18107498e-01 2.57110476e-01
-2.43489400e-01 6.33498549e-01 1.68152618e+00 -2.36350112e-03
-4.42312658e-02 -1.41503796e-01 -5.95363259e-01 3.77228826e-01
7.03380704e-01 -4.51040238e-01 -5.04467189e-01 -7.30807126e-01
1.61302760e-02 4.69680607e-01 2.70300061e-01 -1.06601679e+00
-6.95846742e-03 -4.51512858e-02 5.35427451e-01 -4.52379882e-01
2.63784349e-01 -8.05985093e-01 3.65883857e-01 4.48667735e-01
-4.06139255e-01 -2.31530935e-01 2.68133372e-01 4.49606568e-01
2.07098454e-01 -1.69845819e-01 1.26287258e+00 -1.46916986e-01
6.32581413e-02 4.21397060e-01 -2.39380911e-01 9.88479033e-02
7.93128729e-01 1.07527889e-01 2.52596885e-01 -2.65077323e-01
-8.32016230e-01 -1.45967063e-02 -4.47110757e-02 5.13666794e-02
6.89067721e-01 -1.43486285e+00 -8.55615914e-01 2.06053481e-01
-4.41143811e-01 2.45510682e-01 8.16704452e-01 1.27095401e+00
-7.21558511e-01 1.96146369e-01 -4.91623655e-02 -1.09214318e+00
-7.84930766e-01 1.70690924e-01 7.06868470e-01 -4.34592068e-01
-7.71574736e-01 9.29157734e-01 2.58937061e-01 -3.38296026e-01
7.95383453e-02 -3.86517167e-01 6.51880875e-02 -4.42708969e-01
5.53272605e-01 1.69704467e-01 4.65110280e-02 -3.09224069e-01
-2.85395056e-01 2.68235445e-01 -2.36607477e-01 -2.16565326e-01
1.34143507e+00 -1.36259660e-01 2.24069223e-01 1.53794885e-01
1.35377872e+00 -2.03756332e-01 -1.54344761e+00 -3.75952423e-01
-6.12520993e-01 -5.01689732e-01 6.16396368e-01 -8.01912785e-01
-1.41879678e+00 5.63919067e-01 1.00822794e+00 -1.21782333e-01
1.09156966e+00 1.25345230e-01 8.69388878e-01 -3.88019443e-01
5.71915686e-01 -7.09926784e-01 -4.17952612e-02 1.13908015e-01
9.63383615e-01 -9.26770270e-01 9.20677483e-02 -2.12168634e-01
-7.23124266e-01 6.91435158e-01 3.40940893e-01 -2.01010391e-01
6.79267585e-01 6.84555769e-01 -2.44504169e-01 -3.35614502e-01
-3.34527999e-01 6.51017725e-01 2.20479473e-01 4.51914161e-01
3.65087003e-01 3.21328729e-01 -3.52095336e-01 7.10521638e-01
-1.83678254e-01 1.85210798e-02 3.21200430e-01 6.36155248e-01
-1.84183866e-01 -6.02914453e-01 -4.47733134e-01 8.55176508e-01
-8.00819874e-01 -1.51962981e-01 5.75854003e-01 4.44517493e-01
-1.51025832e-01 3.78920525e-01 8.96450505e-02 -1.89135969e-01
2.52219528e-01 -4.52668220e-01 7.78454959e-01 -1.81650221e-01
-3.75877827e-01 3.39238703e-01 -2.38321692e-01 -6.04492545e-01
1.48370704e-02 -7.43470907e-01 -1.22781956e+00 -1.12634934e-01
-1.58504203e-01 2.50629485e-01 9.19781625e-01 8.28817308e-01
4.51465368e-01 8.14001620e-01 6.49594843e-01 -6.28310442e-01
-9.21415567e-01 -7.66419232e-01 -7.08074033e-01 2.20318690e-01
3.84475470e-01 -6.44217849e-01 -2.80339211e-01 -4.93278027e-01] | [13.54008674621582, -2.3761818408966064] |
9b698372-e145-4f64-864d-5c9966c666d7 | thompson-sampling-efficiently-learns-to | 2206.09977 | null | https://arxiv.org/abs/2206.09977v1 | https://arxiv.org/pdf/2206.09977v1.pdf | Thompson Sampling Efficiently Learns to Control Diffusion Processes | Diffusion processes that evolve according to linear stochastic differential equations are an important family of continuous-time dynamic decision-making models. Optimal policies are well-studied for them, under full certainty about the drift matrices. However, little is known about data-driven control of diffusion processes with uncertain drift matrices as conventional discrete-time analysis techniques are not applicable. In addition, while the task can be viewed as a reinforcement learning problem involving exploration and exploitation trade-off, ensuring system stability is a fundamental component of designing optimal policies. We establish that the popular Thompson sampling algorithm learns optimal actions fast, incurring only a square-root of time regret, and also stabilizes the system in a short time period. To the best of our knowledge, this is the first such result for Thompson sampling in a diffusion process control problem. We validate our theoretical results through empirical simulations with real parameter matrices from two settings of airplane and blood glucose control. Moreover, we observe that Thompson sampling significantly improves (worst-case) regret, compared to the state-of-the-art algorithms, suggesting Thompson sampling explores in a more guarded fashion. Our theoretical analysis involves characterization of a certain optimality manifold that ties the local geometry of the drift parameters to the optimal control of the diffusion process. We expect this technique to be of broader interest. | ['Mohsen Bayati', 'Mohamad Sadegh Shirani Faradonbeh', 'Mohamad Kazem Shirani Faradonbeh'] | 2022-06-20 | null | null | null | null | ['thompson-sampling'] | ['methodology'] | [ 2.82564819e-01 2.55028814e-01 -3.49581957e-01 1.28039777e-01
-4.15030152e-01 -7.62934506e-01 2.84080684e-01 3.27509165e-01
-4.39861298e-01 9.81357217e-01 -2.36671150e-01 -5.24990082e-01
-6.58419371e-01 -4.82797384e-01 -7.90278792e-01 -1.20683599e+00
-3.80284369e-01 5.32135010e-01 -1.10042743e-01 -2.83621669e-01
1.97863132e-01 2.44760975e-01 -9.03123319e-01 -4.03214931e-01
9.44934607e-01 1.19468439e+00 -1.15984552e-01 8.54759574e-01
3.29583764e-01 5.75474560e-01 -4.32019979e-01 -2.99491465e-01
6.70407891e-01 -5.55742323e-01 -3.39878768e-01 2.66635507e-01
-1.59825295e-01 -2.87020534e-01 -1.35026157e-01 1.26095057e+00
6.75913513e-01 1.26889870e-01 5.80915570e-01 -1.11929595e+00
-4.78708506e-01 7.05474436e-01 -7.41510153e-01 3.72376919e-01
-1.90465882e-01 4.89276022e-01 7.04760075e-01 -7.98904151e-02
5.12937486e-01 1.13335407e+00 4.82168138e-01 7.10073113e-01
-1.34921014e+00 -4.35084373e-01 4.93320167e-01 -2.56139427e-01
-1.02018011e+00 -1.62441254e-01 4.26525056e-01 -4.72342372e-01
3.41736287e-01 2.38910466e-01 9.09461915e-01 1.10139275e+00
7.64486074e-01 9.19037402e-01 1.18069518e+00 -6.01925887e-02
9.06607211e-01 1.36309281e-01 -3.12265664e-01 6.11309767e-01
7.35006034e-01 2.38584489e-01 -2.36817017e-01 -3.22576106e-01
1.00697982e+00 4.03629951e-02 -4.84591186e-01 -9.34479058e-01
-1.09742141e+00 8.16022098e-01 -5.91259077e-02 -9.86495707e-03
-7.66258240e-01 3.68654341e-01 1.12873688e-01 8.10954034e-01
4.46372062e-01 4.49867278e-01 -3.84277135e-01 -4.17507082e-01
-5.31048179e-01 5.56352079e-01 1.28717959e+00 1.05798542e+00
4.87464406e-02 1.20234385e-01 -3.47320288e-01 9.10557657e-02
6.20144717e-02 7.72719860e-01 2.19793200e-01 -1.12574756e+00
4.31442171e-01 5.94430156e-02 7.81850994e-01 -6.56083405e-01
-1.98663965e-01 -8.18274856e-01 -8.85518253e-01 5.14272392e-01
9.79448974e-01 -9.26611006e-01 -3.84114712e-01 1.81883991e+00
4.62897152e-01 -1.12446263e-01 1.08645797e-01 1.04523635e+00
-8.37985098e-01 4.02407765e-01 -3.75927567e-01 -9.49654341e-01
9.67435420e-01 -4.56112891e-01 -9.26478267e-01 -2.32144189e-03
3.00310701e-01 -3.44215691e-01 6.67677999e-01 6.64218128e-01
-1.30332041e+00 1.96204588e-01 -1.02416730e+00 7.16143548e-01
2.09795341e-01 -2.80498207e-01 4.65198517e-01 8.12176645e-01
-8.63734663e-01 9.26102161e-01 -1.03284061e+00 -4.67298180e-01
4.06718165e-01 1.98370263e-01 4.62915480e-01 2.72886127e-01
-9.08927619e-01 6.28760338e-01 -1.15896627e-01 1.86827898e-01
-1.30327213e+00 -8.78372669e-01 -1.65028974e-01 -1.88361838e-01
1.09469306e+00 -9.54494953e-01 1.61027658e+00 -9.94282246e-01
-1.88998628e+00 2.14452535e-01 -1.91497579e-02 -9.28298712e-01
1.44142568e+00 -2.10410148e-01 1.03343755e-01 6.96610734e-02
-1.02880582e-01 -1.93860263e-01 1.21567464e+00 -8.79190207e-01
-5.95961273e-01 -5.93446136e-01 1.66199267e-01 3.24246347e-01
-2.37315476e-01 -5.04225492e-01 1.48451716e-01 -6.82833374e-01
-1.28977060e-01 -1.12862694e+00 -6.71527028e-01 3.61761630e-01
-2.00075403e-01 -8.66887067e-03 4.21269178e-01 -1.20547175e-01
1.18240130e+00 -1.75871241e+00 3.22645128e-01 2.56184023e-02
1.05245337e-01 -1.30953625e-01 3.89844418e-01 6.35526359e-01
3.56908143e-01 2.27645293e-01 -4.68979985e-01 -1.23838581e-01
1.25669047e-01 3.63133512e-02 -5.48885465e-01 9.55899477e-01
-9.13135782e-02 7.01277375e-01 -1.09046078e+00 2.32214853e-02
-1.94663137e-01 3.25422995e-02 -3.41701686e-01 -1.36456266e-01
-4.11347777e-01 4.56317097e-01 -9.59283113e-01 5.14566243e-01
3.33186179e-01 -2.45762393e-01 3.71856332e-01 3.80413264e-01
-1.80380285e-01 -6.35488093e-01 -1.39354777e+00 1.32418036e+00
-3.52994800e-01 4.04654890e-01 4.85921562e-01 -9.75864649e-01
5.82179308e-01 2.29117796e-01 6.92148983e-01 -3.01488400e-01
3.01072299e-01 1.84318572e-01 4.06964570e-02 -5.31856716e-01
2.26813838e-01 -4.83441085e-01 1.34585142e-01 7.93821037e-01
-4.72349614e-01 -1.06839858e-01 1.02664433e-01 -7.80480728e-03
1.21654165e+00 -9.32510570e-02 2.67247081e-01 -8.11035872e-01
1.73174247e-01 -8.74051899e-02 6.45387888e-01 9.79402959e-01
-5.09484053e-01 1.38959542e-01 9.11789298e-01 -8.68400335e-02
-1.09319115e+00 -8.68439734e-01 -7.85479099e-02 4.40610617e-01
3.13857406e-01 -6.25096783e-02 -8.69265676e-01 -4.07145679e-01
5.14206171e-01 7.56476402e-01 -1.03886354e+00 -1.28342986e-01
-2.87260115e-01 -9.59734082e-01 4.45578508e-02 2.34523460e-01
3.37933213e-01 -4.21251833e-01 -1.02487898e+00 5.18690765e-01
5.10310173e-01 -7.16994643e-01 -5.46126068e-01 1.08194739e-01
-1.10077786e+00 -1.27305746e+00 -1.07413805e+00 -9.42332670e-02
7.07933247e-01 4.69018705e-02 6.33400083e-01 -4.64453667e-01
-1.28154591e-01 7.45140791e-01 1.16533183e-01 -7.79976845e-01
-3.50613624e-01 -1.02269463e-01 3.97079825e-01 4.10263985e-01
-7.42967203e-02 -3.56841743e-01 -9.86606240e-01 4.74526674e-01
-7.86219180e-01 -2.88018525e-01 4.65619057e-01 7.56234705e-01
7.65703022e-01 4.70389456e-01 5.43202937e-01 -6.82532907e-01
1.11302757e+00 -4.52701867e-01 -1.15891445e+00 1.96093604e-01
-9.26935256e-01 3.98308545e-01 6.58629358e-01 -8.09377372e-01
-9.44344103e-01 -4.29237746e-02 6.46164477e-01 -5.18658698e-01
2.83886313e-01 1.00609303e-01 1.47868440e-01 5.22840954e-02
4.90689665e-01 2.90837198e-01 3.78042370e-01 -2.05416292e-01
2.44234949e-01 3.51802528e-01 4.94094677e-02 -8.76913667e-01
5.81131041e-01 9.26522195e-01 4.67489183e-01 -8.08587611e-01
-6.92491591e-01 -1.21687107e-01 -2.59879053e-01 -3.87274712e-01
5.96246004e-01 -6.03622556e-01 -1.24906528e+00 4.50430214e-01
-6.09466910e-01 -5.97195506e-01 -8.01831067e-01 3.05410028e-01
-1.16783929e+00 7.02930391e-02 -4.53779817e-01 -1.71743202e+00
-1.13439500e-01 -9.08887029e-01 7.85463154e-01 1.93574399e-01
2.70696077e-02 -1.13987684e+00 1.06208779e-01 -1.96170762e-01
5.96100807e-01 6.18210673e-01 4.22078252e-01 -1.99100286e-01
-6.79684043e-01 -5.87813556e-02 5.35161376e-01 1.16191886e-01
5.88967577e-02 -2.65876651e-01 -4.79352713e-01 -6.47552848e-01
7.62935281e-01 2.58729924e-02 6.79578364e-01 8.55601251e-01
8.60852003e-01 -5.52095830e-01 -4.05891806e-01 3.65450025e-01
1.47590685e+00 5.03710985e-01 3.99649069e-02 4.45024848e-01
2.69021809e-01 6.56017423e-01 7.72737324e-01 1.03464603e+00
6.12346129e-03 2.94716358e-01 5.44074059e-01 4.13147211e-01
7.14364767e-01 -1.88859865e-01 4.55538660e-01 2.10143849e-01
-1.29242884e-02 -2.28591368e-01 -6.81015491e-01 5.27328253e-01
-2.07507420e+00 -8.60531986e-01 8.33069161e-02 2.73103714e+00
8.87479186e-01 2.32500970e-01 4.05048341e-01 -1.76897459e-02
6.66705370e-01 -4.53251414e-02 -1.39317477e+00 -2.16746107e-01
-1.85025454e-01 -3.48934770e-01 1.22358215e+00 4.55044627e-01
-8.67555678e-01 4.12153184e-01 6.07763004e+00 5.55141449e-01
-1.01422834e+00 6.04213513e-02 8.89968216e-01 -6.70225561e-01
1.16306506e-02 -7.44665861e-02 -7.54287481e-01 5.16207635e-01
8.68193090e-01 -8.93149018e-01 6.41190290e-01 8.86972606e-01
6.42977178e-01 -2.04798952e-01 -1.07090735e+00 8.92462313e-01
-3.87323737e-01 -9.94501293e-01 -5.28677166e-01 4.06331122e-01
9.92201388e-01 -3.79233748e-01 3.41374606e-01 -7.63101727e-02
5.55259585e-01 -6.58010781e-01 7.98663735e-01 7.87628651e-01
3.15797746e-01 -1.00968242e+00 4.90187436e-01 5.38767159e-01
-7.27098346e-01 -5.16240478e-01 -4.48123634e-01 -1.52284771e-01
4.79256600e-01 9.19981658e-01 -4.12597775e-01 2.20915854e-01
4.54419523e-01 6.79637074e-01 3.38035151e-02 1.08929932e+00
-5.49177825e-02 6.23397291e-01 -4.72700685e-01 -5.56591392e-01
1.64064959e-01 -5.14043689e-01 1.01643538e+00 6.35791898e-01
3.63899112e-01 5.23484536e-02 2.23684357e-03 7.96567917e-01
2.35341966e-01 2.96013746e-02 -6.40628099e-01 -4.10923421e-01
1.73870221e-01 8.30036163e-01 -8.60672414e-01 -1.60488456e-01
1.03159644e-01 9.81955349e-01 -7.14472160e-02 4.52908993e-01
-6.90765977e-01 -1.71504468e-01 1.00783122e+00 1.75766572e-01
5.11958838e-01 -4.00826007e-01 -3.79108578e-01 -1.07184553e+00
2.31510386e-01 -6.76030695e-01 3.05832624e-01 -2.88242362e-02
-1.41696143e+00 1.23525865e-01 -1.27300546e-01 -1.13042510e+00
-3.43298316e-01 -3.43398869e-01 -1.01513267e-01 6.03079200e-01
-1.20212507e+00 -6.40822202e-02 3.09144855e-01 3.71700495e-01
5.92732966e-01 2.93839872e-02 2.69978493e-01 -3.13481331e-01
-7.77444661e-01 1.57049805e-01 9.27654803e-01 -5.06907344e-01
4.35123116e-01 -1.55944180e+00 2.11361334e-01 7.85366893e-01
-5.13551950e-01 5.21831691e-01 1.19727969e+00 -8.54371786e-01
-2.19885778e+00 -8.97119224e-01 -1.66197382e-02 -3.00422788e-01
1.22398925e+00 -3.52598488e-01 -7.17767954e-01 3.54452431e-01
1.59789547e-01 -2.64757484e-01 9.63673294e-02 -2.71304518e-01
5.36784649e-01 -2.27909341e-01 -1.16051435e+00 8.10003400e-01
1.15318370e+00 1.83558892e-02 -1.89472288e-02 4.70802873e-01
5.35037756e-01 -5.06007195e-01 -9.12096858e-01 6.90681562e-02
5.65609455e-01 -6.72333717e-01 5.55410445e-01 -6.59275591e-01
5.28931990e-03 -1.75031587e-01 -6.83905110e-02 -1.52347803e+00
2.86522806e-02 -1.58404946e+00 -5.42699218e-01 9.49406385e-01
2.93045908e-01 -8.26122582e-01 8.10441852e-01 5.99506319e-01
5.07022798e-01 -1.04087806e+00 -9.35714245e-01 -1.34063101e+00
4.41659242e-01 -2.42823198e-01 2.86538690e-01 6.97565496e-01
-1.15879830e-02 -8.73054564e-03 -4.24808532e-01 1.72083706e-01
1.12567079e+00 3.40328477e-02 4.56676304e-01 -9.67441320e-01
-6.05607212e-01 -6.54138982e-01 6.53843954e-02 -1.18871737e+00
-1.48064718e-01 -3.18562388e-01 1.03424303e-01 -1.23654735e+00
-1.49990022e-01 -4.92607921e-01 -1.66392043e-01 -2.58059889e-01
-1.43992633e-01 -6.25645578e-01 1.35374516e-01 1.66098222e-01
-5.28267264e-01 7.72162735e-01 1.50924444e+00 -7.26379156e-02
-3.97003084e-01 6.40822709e-01 -8.12806308e-01 3.53514582e-01
9.90565658e-01 -4.75053340e-01 -9.17384386e-01 -1.83226734e-01
2.75318831e-01 4.86706823e-01 -3.40794213e-02 -8.10989738e-01
3.29247206e-01 -5.86491585e-01 -6.49323016e-02 -2.03293428e-01
5.26586175e-02 -7.17950344e-01 1.83575109e-01 9.37182724e-01
-5.51894605e-01 2.01864541e-01 -4.75159883e-02 1.36952722e+00
4.39463079e-01 -1.65401265e-01 9.00430739e-01 -1.60953537e-01
-2.38437355e-02 5.37029982e-01 -6.04627192e-01 3.29607189e-01
1.39746940e+00 -6.67348178e-03 -1.26537099e-01 -7.54381597e-01
-6.72650158e-01 5.41370094e-01 5.14104187e-01 1.62155703e-01
1.71634361e-01 -8.81033599e-01 -6.21148229e-01 -2.21977998e-02
-4.33859438e-01 -8.34519938e-02 2.56112546e-01 1.11625147e+00
-1.44397616e-01 2.75310963e-01 1.21476434e-01 -6.54664576e-01
-8.10464025e-01 8.96537602e-01 6.20865822e-01 -2.07469225e-01
-6.34991407e-01 4.80534703e-01 1.10921860e-01 4.16964144e-01
3.11990201e-01 -5.72439373e-01 3.95469636e-01 1.16909437e-01
7.25753188e-01 6.89415932e-01 -1.81343257e-01 2.03062460e-01
-1.60008408e-02 5.19466579e-01 1.67938862e-02 -5.13411820e-01
1.16193032e+00 -4.46570188e-01 2.30627641e-01 7.08125234e-01
6.07987463e-01 -2.21884117e-01 -1.92558730e+00 -1.63203686e-01
5.91823459e-02 -3.40146899e-01 -2.82831769e-02 -6.44645154e-01
-1.06078613e+00 6.74443007e-01 7.57729411e-01 6.77099645e-01
9.96080935e-01 -4.05745029e-01 3.54829669e-01 4.25376624e-01
5.42056024e-01 -1.28711998e+00 -7.27214515e-02 2.24805057e-01
8.92619491e-01 -9.84164834e-01 -1.25353605e-01 -7.55880252e-02
-8.28704298e-01 9.63254631e-01 2.85361081e-01 -2.90381402e-01
7.14280128e-01 5.14216363e-01 -1.34620667e-01 2.07951352e-01
-1.16333127e+00 -5.47907036e-03 -3.81573051e-01 4.40022945e-01
-4.12305258e-02 2.20391586e-01 -5.79451025e-01 3.78091872e-01
1.04375228e-01 1.97529331e-01 6.90074801e-01 1.24555254e+00
-4.49051261e-01 -8.62677753e-01 -4.40583318e-01 3.12064171e-01
-6.55256093e-01 2.27607206e-01 -2.26365522e-01 6.92318976e-01
-6.05065703e-01 9.09228325e-01 -1.60940170e-01 3.26917738e-01
4.75732356e-01 -1.77922830e-01 5.91314256e-01 -2.68750545e-02
-3.00779670e-01 2.09415153e-01 -2.09011599e-01 -6.09550953e-01
-6.56412989e-02 -1.15070784e+00 -8.92004907e-01 -4.30500597e-01
-4.79284562e-02 1.12882249e-01 6.45188034e-01 8.03330719e-01
3.69341135e-01 5.05648196e-01 6.83164716e-01 -3.69120538e-01
-1.58982825e+00 -6.59970105e-01 -1.02058780e+00 3.91501226e-02
5.62003314e-01 -6.25252128e-01 -5.06706238e-01 -2.65070826e-01] | [4.510873794555664, 2.7094788551330566] |
081df2fe-138b-4ede-91ec-e676922a5602 | explaining-groups-of-points-in-low | 2003.01640 | null | https://arxiv.org/abs/2003.01640v3 | https://arxiv.org/pdf/2003.01640v3.pdf | Explaining Groups of Points in Low-Dimensional Representations | A common workflow in data exploration is to learn a low-dimensional representation of the data, identify groups of points in that representation, and examine the differences between the groups to determine what they represent. We treat this workflow as an interpretable machine learning problem by leveraging the model that learned the low-dimensional representation to help identify the key differences between the groups. To solve this problem, we introduce a new type of explanation, a Global Counterfactual Explanation (GCE), and our algorithm, Transitive Global Translations (TGT), for computing GCEs. TGT identifies the differences between each pair of groups using compressed sensing but constrains those pairwise differences to be consistent among all of the groups. Empirically, we demonstrate that TGT is able to identify explanations that accurately explain the model while being relatively sparse, and that these explanations match real patterns in the data. | ['Sriram Sankararaman', 'Jonathan Terhorst', 'Ameet Talwalkar', 'Gregory Plumb'] | 2020-03-03 | null | https://proceedings.icml.cc/static/paper_files/icml/2020/2264-Paper.pdf | https://proceedings.icml.cc/static/paper_files/icml/2020/2264-Paper.pdf | icml-2020-1 | ['counterfactual-explanation'] | ['miscellaneous'] | [ 4.32578534e-01 6.80505991e-01 -4.92141038e-01 -3.53625238e-01
-5.61366618e-01 -6.43418252e-01 7.51362741e-01 4.04893279e-01
4.09028888e-01 4.14049119e-01 9.09426928e-01 -5.65211117e-01
-6.88982069e-01 -5.96963704e-01 -7.38004744e-01 -5.06234705e-01
-6.13653898e-01 7.54165590e-01 -4.81026143e-01 2.53634810e-01
5.78737319e-01 3.81359458e-01 -1.57618821e+00 7.42724657e-01
7.09347427e-01 7.02681005e-01 1.83167934e-01 3.29334170e-01
-2.28414878e-01 4.59692985e-01 -2.36827284e-01 1.40622362e-01
6.39427781e-01 -7.25507438e-01 -1.04730642e+00 2.10234895e-01
1.61820248e-01 -2.24994849e-02 2.50758410e-01 7.80675173e-01
-2.07955405e-01 3.21153738e-02 7.05589354e-01 -1.48644340e+00
-4.56100732e-01 8.94147158e-01 -5.33229113e-01 -8.18852633e-02
6.48523092e-01 -5.60389645e-02 1.32393146e+00 -8.52236390e-01
7.84779131e-01 1.44811511e+00 5.47728479e-01 3.80382895e-01
-1.69129479e+00 -5.79820395e-01 3.53516221e-01 1.00443624e-01
-1.15147436e+00 -3.85763377e-01 8.03286135e-01 -5.18623590e-01
5.65218508e-01 8.87844622e-01 8.03629696e-01 5.40138602e-01
1.72465160e-01 4.31251258e-01 1.09586942e+00 -5.15086234e-01
7.54306436e-01 -2.24957377e-01 5.36911823e-02 6.61887586e-01
4.59982604e-01 2.13561468e-02 -1.08105206e+00 -6.19733214e-01
4.47641641e-01 3.11700791e-01 -2.39193663e-01 -7.83296347e-01
-1.40937924e+00 9.90051448e-01 3.09292525e-01 5.07235527e-03
-5.99133492e-01 1.20379709e-01 -5.57466298e-02 1.62392467e-01
3.20345849e-01 9.34999228e-01 -6.35341942e-01 8.98254216e-02
-7.52010107e-01 3.18505019e-01 8.12737286e-01 5.35153031e-01
1.16841936e+00 -4.87774938e-01 3.23004335e-01 1.19897105e-01
3.33344340e-01 -8.47373903e-02 5.34796357e-01 -1.19398582e+00
4.41119075e-01 8.79030764e-01 -3.73302470e-03 -1.29536986e+00
-3.56936157e-01 -1.65834188e-01 -6.21441901e-01 6.34866953e-02
2.20343068e-01 -1.14106955e-02 -7.20722318e-01 1.95153534e+00
2.97521323e-01 2.33559459e-01 8.15638229e-02 8.54711711e-01
2.00748220e-01 3.83926272e-01 -2.36370131e-01 -4.53832179e-01
8.43116939e-01 -3.94402504e-01 -5.16398430e-01 -5.91767251e-01
9.61162567e-01 -3.47724020e-01 9.99450862e-01 6.91068470e-02
-8.87604892e-01 -2.54773080e-01 -9.97223258e-01 2.14224294e-01
1.02393562e-02 -1.72970831e-01 9.11752760e-01 1.20234422e-01
-8.92617047e-01 7.49192655e-01 -7.69311666e-01 -3.48938584e-01
2.86350369e-01 3.72883350e-01 -5.05315065e-01 -1.13680407e-01
-7.72014260e-01 5.58471501e-01 3.39048237e-01 -1.26571894e-01
-8.10588300e-01 -9.10075784e-01 -8.29985261e-01 2.03697905e-01
4.94609952e-01 -6.52577519e-01 7.17736602e-01 -7.81164706e-01
-5.31538546e-01 7.06515729e-01 -7.48152018e-01 -5.25869370e-01
3.36702853e-01 7.32085854e-02 -8.39662552e-02 -2.53353477e-01
3.79117608e-01 5.60946405e-01 5.42434454e-01 -1.64669263e+00
-4.88697141e-01 -6.59848392e-01 -4.50905442e-01 1.00520611e-01
5.49228825e-02 -3.46501321e-01 -3.52716856e-02 -5.15660286e-01
1.11936343e+00 -1.20133924e+00 -5.01230776e-01 9.30824056e-02
-9.44917321e-01 5.44579998e-02 8.88780355e-01 -6.35427475e-01
1.21213210e+00 -2.07766342e+00 2.75271147e-01 7.95947015e-01
6.76856339e-01 -6.65519893e-01 -6.16614660e-03 6.32578611e-01
-4.45833236e-01 4.62928683e-01 -5.23030162e-01 -5.02179682e-01
-1.71836615e-01 6.42076850e-01 -6.36147559e-01 3.91869903e-01
-1.80640947e-02 6.12825334e-01 -8.90005469e-01 -6.02618232e-02
-5.98613247e-02 -2.03107029e-01 -7.23586798e-01 2.82288283e-01
-4.88527656e-01 6.38246596e-01 -3.90961468e-01 2.04517975e-01
6.58592165e-01 -3.96521568e-01 8.47222686e-01 -9.02119353e-02
-2.08625183e-01 5.38844526e-01 -1.52225149e+00 1.48101759e+00
8.04498866e-02 6.71807706e-01 -1.94328606e-01 -1.09890258e+00
1.06585681e+00 -3.66593637e-02 8.24557543e-01 -1.83488742e-01
-3.72607350e-01 2.74706930e-01 5.87057434e-02 -6.44455194e-01
1.97430998e-01 -2.15252653e-01 1.41176889e-02 1.26455152e+00
-4.98270988e-01 2.68995911e-01 -1.77369982e-01 3.12193274e-01
1.15500140e+00 -3.29743028e-01 5.87523103e-01 -3.76454264e-01
-1.36384457e-01 1.19790010e-01 6.78840578e-01 9.57162321e-01
4.51030225e-01 8.21269929e-01 6.74965560e-01 -1.12045491e+00
-9.08401430e-01 -8.87557805e-01 3.35409343e-01 5.03069937e-01
3.12433690e-01 -9.27550733e-01 -3.08790892e-01 -7.06738889e-01
2.67715961e-01 1.00149047e+00 -1.21374464e+00 -1.97858721e-01
-3.02590758e-01 -5.45085549e-01 -7.87775442e-02 3.87006730e-01
8.85634050e-02 -6.13602281e-01 -1.03442776e+00 -1.60500392e-01
-3.13308835e-01 -6.10846281e-01 -2.06927001e-01 4.80440408e-01
-1.00552607e+00 -1.45191240e+00 2.53288001e-01 -1.46324724e-01
1.03341603e+00 3.49210471e-01 1.11666906e+00 2.05868110e-01
-1.02929942e-01 1.54125700e-02 -1.90361202e-01 -4.67292100e-01
-3.23021710e-01 -4.01641667e-01 2.23989427e-01 7.77524412e-02
4.41767097e-01 -6.96080625e-01 -4.00169015e-01 3.80652934e-01
-7.71760762e-01 5.66950858e-01 4.12558436e-01 4.03878659e-01
8.92599642e-01 1.15133807e-01 1.95681393e-01 -9.95289981e-01
6.66437387e-01 -8.82436037e-01 -8.81698802e-02 3.79476577e-01
-9.51206744e-01 7.31400251e-01 4.54147607e-01 -2.21735820e-01
-6.13275290e-01 1.81890398e-01 5.80354869e-01 -5.16733885e-01
-8.70406181e-02 1.03491712e+00 -1.24340147e-01 5.20141542e-01
6.32600009e-01 1.64283976e-01 2.33585641e-01 -8.54850233e-01
4.86327618e-01 3.23495656e-01 4.95536298e-01 -5.23141801e-01
8.14136326e-01 8.54624510e-01 2.19476476e-01 -3.62284929e-01
-7.86459148e-01 -1.06706932e-01 -8.86311471e-01 1.01437829e-01
5.65158069e-01 -6.25513792e-01 -4.55745429e-01 -6.35666907e-01
-9.42707241e-01 -1.24545701e-01 -6.24372303e-01 5.61245680e-01
-6.52353644e-01 1.12629801e-01 3.78576547e-01 -7.99624205e-01
1.34883657e-01 -9.89320695e-01 9.35633302e-01 -1.73463330e-01
-1.01396370e+00 -9.28588867e-01 2.52332717e-01 1.40345395e-01
-2.86855479e-03 6.96651340e-01 1.40415561e+00 -1.03818202e+00
-8.35086882e-01 -8.86594579e-02 1.21090315e-01 -4.95340079e-01
5.45114040e-01 -3.74122262e-02 -5.27100980e-01 -2.81746417e-01
-2.78494488e-02 1.75614968e-01 6.71803415e-01 5.77640831e-01
1.45808518e+00 -8.96918476e-01 -7.39207327e-01 6.17285192e-01
1.06574273e+00 8.65353867e-02 1.57076105e-01 3.20370108e-01
5.24083614e-01 8.11179876e-01 5.26526153e-01 6.53996468e-01
4.92952526e-01 7.11619318e-01 7.09510326e-01 -8.86329114e-02
1.59859002e-01 -6.47509992e-01 3.50874439e-02 3.19760293e-01
1.43822685e-01 8.59574825e-02 -1.06413317e+00 5.37131727e-01
-2.21698141e+00 -9.14877355e-01 -3.88376378e-02 2.27609730e+00
3.33203614e-01 -1.15671106e-01 5.28321713e-02 1.96452007e-01
4.65014577e-01 1.16992258e-01 -8.34330201e-01 -5.72412431e-01
6.24641078e-03 -2.55215287e-01 2.26108685e-01 4.83898848e-01
-5.56943834e-01 2.69251943e-01 7.25079346e+00 -5.48032075e-02
-8.35449815e-01 -4.84693527e-01 7.35138297e-01 -9.08186808e-02
-1.07297945e+00 7.20231056e-01 -1.52393520e-01 3.51307184e-01
6.53983057e-01 -4.08628285e-01 6.33839428e-01 5.67284346e-01
6.03940904e-01 2.33413223e-02 -1.77381277e+00 6.16044462e-01
-9.02881846e-03 -1.84517860e+00 4.54361230e-01 4.61391956e-01
7.71257579e-01 -2.62714595e-01 1.10378020e-01 -5.00073373e-01
4.51378733e-01 -1.30945039e+00 7.10112512e-01 5.38770735e-01
5.12114108e-01 -5.42349994e-01 3.05308610e-01 5.80291748e-01
-1.01282120e+00 -4.20021564e-01 -2.96238869e-01 -5.04046202e-01
-1.19907096e-01 5.35726726e-01 -1.15974164e+00 4.90637183e-01
5.92991710e-01 8.78713965e-01 -4.83532399e-01 6.31160200e-01
-9.11890790e-02 4.24259454e-01 -3.20646405e-01 5.47794640e-01
-1.39858099e-02 -2.85548002e-01 7.26149559e-01 4.47343528e-01
5.95417917e-01 6.34031445e-02 2.82034755e-01 1.20100522e+00
5.02258120e-03 -3.91547710e-01 -7.77632952e-01 -8.81058723e-03
8.78954947e-01 6.16106093e-01 -5.51421225e-01 -2.49776974e-01
-8.06223154e-02 5.99885166e-01 2.26623967e-01 1.80058375e-01
-1.75720036e-01 4.38148707e-01 1.09718812e+00 2.54113406e-01
-2.10930221e-02 -2.20409855e-01 -8.14043105e-01 -1.28535092e+00
1.54095739e-02 -1.14179599e+00 7.83818662e-01 -8.93674672e-01
-9.44197416e-01 1.86052933e-01 2.41416171e-01 -1.31353199e+00
-6.60535634e-01 -5.17131202e-02 -8.41188073e-01 8.95267665e-01
-8.05335462e-01 -7.64183760e-01 -2.64994025e-01 3.11073512e-01
5.11902869e-01 -7.95896444e-03 9.69174325e-01 -6.38761520e-01
-4.15887713e-01 2.01010898e-01 2.34288156e-01 -3.08938235e-01
2.79809088e-01 -1.24175990e+00 5.54806650e-01 8.62424850e-01
5.20017803e-01 1.16591752e+00 1.09643316e+00 -9.43380892e-01
-1.35631514e+00 -8.87268245e-01 1.05925953e+00 -4.34566617e-01
4.31026727e-01 -2.90073276e-01 -9.76732016e-01 1.09579170e+00
-9.98462141e-02 -1.15487553e-01 9.32873666e-01 5.02507210e-01
-4.66578990e-01 2.60008067e-01 -1.09631455e+00 6.65454924e-01
1.11491930e+00 -4.16139871e-01 -6.22438669e-01 3.67406845e-01
5.52331865e-01 -1.58955783e-01 -6.69415772e-01 2.03589961e-01
6.54533923e-01 -1.25024724e+00 9.61237848e-01 -1.01435769e+00
7.13554323e-01 -4.82776910e-01 -3.88856232e-01 -1.63062179e+00
-2.10321099e-01 -6.78491175e-01 -1.96389765e-01 8.14796448e-01
4.98511434e-01 -5.68412125e-01 1.07110453e+00 1.19736695e+00
2.38065764e-01 -8.97870064e-01 -8.54892910e-01 -4.83028769e-01
-3.19074839e-01 -2.86782235e-01 1.29667270e+00 1.35667837e+00
2.67579526e-01 1.06219336e-01 -5.40140510e-01 2.53511637e-01
8.17235053e-01 9.33289766e-01 9.26845014e-01 -1.56448126e+00
-2.24173412e-01 4.33994085e-02 -2.74920642e-01 -7.69689023e-01
6.75446168e-02 -9.81117249e-01 -2.41371721e-01 -1.66513097e+00
3.73586923e-01 -6.64336145e-01 4.62529063e-02 6.65881693e-01
-1.38113454e-01 -4.04153258e-01 3.42820227e-01 7.97159016e-01
-3.78630936e-01 2.53685981e-01 7.45620489e-01 6.45184442e-02
-4.20192242e-01 -2.00210467e-01 -1.32350421e+00 7.18539953e-01
6.61050260e-01 -7.43693113e-01 -4.34634805e-01 -5.26438117e-01
3.39193225e-01 -8.16742107e-02 2.71139562e-01 -7.12814033e-01
2.77368933e-01 -5.70044577e-01 4.96108145e-01 -5.42756140e-01
7.66781271e-02 -8.49507332e-01 8.54498506e-01 6.15781605e-01
-1.01299596e+00 2.60113418e-01 -1.38505116e-01 7.14106977e-01
-3.15150768e-02 1.17447093e-01 2.90480494e-01 -1.84836611e-01
-5.58398426e-01 5.12311086e-02 -1.35928959e-01 -1.25307605e-01
8.44743550e-01 -4.35420543e-01 -2.41722479e-01 -6.97388589e-01
-8.72368753e-01 3.14386338e-01 7.21638262e-01 3.16911876e-01
7.25918591e-01 -1.28076065e+00 -5.97068846e-01 5.96908987e-01
2.82596022e-01 1.15611203e-01 -1.63168553e-02 5.02434611e-01
8.56481045e-02 3.04631859e-01 -1.29952043e-01 -5.33820629e-01
-1.15173542e+00 4.56696868e-01 2.67732978e-01 -1.46984890e-01
-8.16670597e-01 5.00128806e-01 3.20728779e-01 -4.68718827e-01
-2.02800274e-01 -4.43175822e-01 -1.10989332e-01 1.21219018e-02
4.80864882e-01 2.84422696e-01 -1.72021732e-01 -4.44433212e-01
-4.78811443e-01 4.67957795e-01 2.66468436e-01 3.90728749e-02
1.66192663e+00 -3.29725742e-01 -2.41841808e-01 4.98149097e-01
1.16456711e+00 3.98605242e-02 -1.36978984e+00 -1.81933269e-01
3.61517191e-01 -9.58792508e-01 -2.85482049e-01 -6.40212953e-01
-7.41347671e-01 4.66128170e-01 1.72777370e-01 5.99667192e-01
9.36661124e-01 4.67796773e-01 1.89655915e-01 3.37072462e-01
9.22063887e-02 -5.92460632e-01 -5.70566282e-02 3.49532217e-02
1.19312203e+00 -1.00082636e+00 2.60293454e-01 -3.56422126e-01
-5.91038644e-01 1.02139890e+00 2.10987419e-01 1.33983567e-01
5.44376135e-01 -1.59552763e-03 -1.91188324e-02 -6.85862780e-01
-1.11400902e+00 1.54788598e-01 2.86799550e-01 5.55677414e-01
5.72641008e-02 2.26891100e-01 -6.49790391e-02 4.28695679e-01
-5.90636849e-01 -2.35437736e-01 4.99282241e-01 7.90137708e-01
-2.49340683e-01 -9.80872393e-01 -5.93660891e-01 9.73534942e-01
9.41189080e-02 1.05066992e-01 -8.91000926e-01 6.45438969e-01
1.05506249e-01 8.33082914e-01 3.58431160e-01 -7.34066546e-01
2.16793880e-01 9.04053524e-02 -6.77884594e-02 -6.52748644e-01
-1.34107858e-01 -1.20541379e-01 2.42424365e-02 -9.07465160e-01
-3.73966902e-01 -9.49157953e-01 -1.33644569e+00 -2.14293480e-01
5.76086044e-02 4.53026175e-01 6.84059858e-01 1.29590988e+00
7.46545434e-01 1.22988813e-01 7.84011662e-01 -5.80745459e-01
-4.34109330e-01 -4.19655234e-01 -5.02502680e-01 8.28490317e-01
6.20297849e-01 -7.22091198e-01 -6.15129054e-01 7.78056458e-02] | [8.781569480895996, 5.678574562072754] |
1f105f2f-8b4c-4fe5-9a3a-bf8c81dec0c8 | unsupervised-acoustic-unit-discovery-by | 2104.00994 | null | https://arxiv.org/abs/2104.00994v2 | https://arxiv.org/pdf/2104.00994v2.pdf | Unsupervised Acoustic Unit Discovery by Leveraging a Language-Independent Subword Discriminative Feature Representation | This paper tackles automatically discovering phone-like acoustic units (AUD) from unlabeled speech data. Past studies usually proposed single-step approaches. We propose a two-stage approach: the first stage learns a subword-discriminative feature representation and the second stage applies clustering to the learned representation and obtains phone-like clusters as the discovered acoustic units. In the first stage, a recently proposed method in the task of unsupervised subword modeling is improved by replacing a monolingual out-of-domain (OOD) ASR system with a multilingual one to create a subword-discriminative representation that is more language-independent. In the second stage, segment-level k-means is adopted, and two methods to represent the variable-length speech segments as fixed-dimension feature vectors are compared. Experiments on a very low-resource Mboshi language corpus show that our approach outperforms state-of-the-art AUD in both normalized mutual information (NMI) and F-score. The multilingual ASR improved upon the monolingual ASR in providing OOD phone labels and in estimating the phone boundaries. A comparison of our systems with and without knowing the ground-truth phone boundaries showed a 16% NMI performance gap, suggesting that the current approach can significantly benefit from improved phone boundary estimation. | ['Odette Scharenborg', 'Laureano Moro-Velázquez', 'Piotr Żelasko', 'Siyuan Feng'] | 2021-04-02 | null | null | null | null | ['acoustic-unit-discovery'] | ['speech'] | [ 2.37666324e-01 5.53770177e-02 -1.91576123e-01 -5.13007402e-01
-1.68308771e+00 -5.69776237e-01 3.88228774e-01 -1.02067478e-01
-5.59543252e-01 5.81780851e-01 3.77776623e-01 -5.08799255e-01
2.44777009e-01 -2.65813202e-01 -5.60111880e-01 -7.97239184e-01
1.51130527e-01 7.52832055e-01 2.21411124e-01 2.12238468e-02
1.33768292e-02 7.92413056e-02 -1.49571240e+00 1.97385475e-01
9.38605428e-01 7.49019504e-01 8.73482227e-01 7.26114929e-01
-3.79543751e-01 3.56676400e-01 -6.83175445e-01 3.28323878e-02
2.30793983e-01 -6.41000092e-01 -6.89605355e-01 2.56794930e-01
2.11779192e-01 2.65330225e-01 -6.79548383e-02 8.81839752e-01
5.81035674e-01 2.81897753e-01 7.90937960e-01 -5.51630855e-01
-4.69764352e-01 1.17770731e+00 -2.47757137e-01 9.72180516e-02
2.24593505e-01 -3.64049047e-01 1.20646918e+00 -1.23592639e+00
3.34486157e-01 1.20345497e+00 5.40321350e-01 5.98790288e-01
-1.24267364e+00 -6.60809696e-01 1.37816861e-01 2.17174426e-01
-1.80861640e+00 -8.94370377e-01 7.28604019e-01 -5.12715101e-01
1.28727686e+00 2.71470219e-01 1.94608644e-01 8.58124852e-01
-2.69118339e-01 7.25953817e-01 1.10538793e+00 -1.03579295e+00
3.16596508e-01 3.62052798e-01 2.90920109e-01 3.69447589e-01
-1.54080749e-01 -5.48012778e-02 -4.38598931e-01 -8.41651782e-02
3.02321255e-01 -5.80128074e-01 -3.66881453e-02 2.57618129e-02
-1.19764304e+00 9.03948426e-01 -2.17732832e-01 9.57355261e-01
-4.10137266e-01 -3.41815531e-01 4.23999548e-01 2.80269682e-01
7.28180587e-01 2.74973840e-01 -4.64966923e-01 -2.95360267e-01
-1.18529260e+00 -3.62138778e-01 6.95476532e-01 8.18557858e-01
9.71005797e-01 2.93260276e-01 2.33864337e-01 1.43155420e+00
4.63170350e-01 5.04547894e-01 9.72426832e-01 -6.27081513e-01
5.14298558e-01 2.22761288e-01 -2.72906244e-01 -2.76439458e-01
-8.70427787e-02 -2.29198143e-01 -4.54588503e-01 -3.38309646e-01
2.50599891e-01 -1.04323618e-01 -1.11078441e+00 1.63624537e+00
1.75002724e-01 3.22569937e-01 4.01700377e-01 6.08316302e-01
5.02187133e-01 1.08727241e+00 -1.54905558e-01 -5.98469138e-01
1.20845163e+00 -1.01733494e+00 -8.38305533e-01 -3.51769000e-01
8.05329978e-01 -1.04368138e+00 1.05546558e+00 3.98500979e-01
-9.27834392e-01 -8.31099749e-01 -9.31448579e-01 1.87150508e-01
-3.03014964e-01 4.45792496e-01 3.75056677e-02 9.86317515e-01
-1.03904355e+00 1.94216277e-02 -6.73490167e-01 -2.99087048e-01
-2.80278176e-01 2.94837028e-01 -4.05225337e-01 -1.03710935e-01
-1.39576030e+00 7.07742453e-01 6.00726485e-01 -2.35944018e-02
-9.49121118e-01 -4.46427256e-01 -9.54140365e-01 -2.73354560e-01
2.23469764e-01 1.21435396e-01 1.15610278e+00 -8.11624169e-01
-1.83835030e+00 8.47977459e-01 -6.61283314e-01 -5.11636972e-01
-5.61937317e-02 -7.52678066e-02 -6.60849273e-01 -3.89363840e-02
-1.92479547e-02 4.79122698e-01 8.31586003e-01 -1.40409803e+00
-9.45379376e-01 -3.39016676e-01 -6.02753639e-01 4.48752493e-01
-3.84089291e-01 -3.77402008e-02 -6.85121417e-01 -5.84079504e-01
2.09053308e-01 -9.96935666e-01 -3.68142836e-02 -1.20732033e+00
-3.98770630e-01 -7.03798950e-01 6.42496586e-01 -1.06157267e+00
1.73384202e+00 -2.23620749e+00 1.95665866e-01 3.83237243e-01
-2.90904939e-01 3.70944202e-01 -8.22189003e-02 3.08343261e-01
5.73637150e-02 -7.95892477e-02 -4.08051968e-01 -8.45930874e-01
2.52350699e-02 4.61847425e-01 -1.45352021e-01 2.73634076e-01
5.89990653e-02 4.92265373e-01 -6.96808577e-01 -4.77338016e-01
2.96726435e-01 5.56769669e-01 -1.90248102e-01 3.16328973e-01
7.17776790e-02 5.65065205e-01 1.88574389e-01 4.78668392e-01
4.68509257e-01 6.37307346e-01 4.94235665e-01 -1.54257454e-02
-3.73394012e-01 7.84485698e-01 -1.19625020e+00 1.78115451e+00
-8.54934931e-01 5.84579587e-01 1.04175784e-01 -1.32998538e+00
1.30114007e+00 6.21972322e-01 2.26902127e-01 -4.36779350e-01
-7.01735392e-02 7.79431403e-01 1.69040099e-01 -2.99855411e-01
5.41742206e-01 -1.11279033e-01 -3.35696429e-01 1.97232425e-01
3.16486865e-01 -2.20013648e-01 -2.41840500e-02 -1.34712100e-01
7.61784673e-01 -1.29331619e-01 3.27509284e-01 -3.08162719e-01
7.64958024e-01 -1.23189971e-01 6.23558819e-01 6.30697489e-01
-1.33983523e-01 7.37960458e-01 -1.10240690e-01 7.88802058e-02
-9.39276755e-01 -1.12557733e+00 -2.44238213e-01 1.25008202e+00
-3.39921862e-01 -2.83953667e-01 -1.08807659e+00 -6.50600195e-01
-2.13847041e-01 9.50411737e-01 -3.09813648e-01 2.49099299e-01
-8.67759049e-01 -4.34811652e-01 8.59274030e-01 2.91058362e-01
-7.18178079e-02 -1.02835178e+00 2.88572490e-01 5.40680587e-01
-4.31791812e-01 -1.20058739e+00 -7.49943018e-01 6.05248630e-01
-4.84999806e-01 -4.24593687e-01 -8.96489382e-01 -1.41971672e+00
4.08382326e-01 9.96022746e-02 8.25436056e-01 -4.84117270e-01
8.70208591e-02 1.34794548e-01 -6.18048429e-01 -1.60155773e-01
-9.25152242e-01 3.62956911e-01 5.28562009e-01 1.81236744e-01
8.31475616e-01 -3.12559247e-01 -3.06186099e-02 3.27496409e-01
-6.00147069e-01 -5.97512543e-01 6.73999727e-01 7.54550934e-01
9.56350863e-01 -3.26413964e-03 9.78872955e-01 -7.77977884e-01
5.22360027e-01 -4.79386121e-01 -3.86682212e-01 2.48562902e-01
-5.99621236e-01 -6.67730421e-02 6.93138301e-01 -6.32853150e-01
-1.11272049e+00 4.66449976e-01 -6.10713959e-01 -3.89615268e-01
-4.71880883e-01 5.12251973e-01 -5.38760900e-01 4.38699424e-01
4.67838913e-01 5.16282856e-01 -1.66164100e-01 -8.49518418e-01
4.19732422e-01 1.51620698e+00 6.30466163e-01 -4.41572011e-01
5.79028189e-01 -8.24027061e-02 -8.84040296e-01 -1.47321117e+00
-6.01339996e-01 -1.03385890e+00 -8.91385972e-01 -2.77990159e-02
9.66017246e-01 -1.06885135e+00 -1.14119817e-02 5.14731407e-01
-1.04334378e+00 -2.89932251e-01 -4.10094857e-01 9.87892091e-01
-4.53531265e-01 4.65961188e-01 -5.14755845e-01 -1.03966665e+00
-1.62055641e-01 -1.21216011e+00 1.17515624e+00 -1.20749444e-01
-2.88465679e-01 -1.01885116e+00 5.15537977e-01 5.27921617e-01
2.03336909e-01 -5.76835394e-01 8.41911197e-01 -1.14804983e+00
3.38406041e-02 -1.09918945e-01 2.89146543e-01 9.02467191e-01
3.68517816e-01 -3.20069462e-01 -1.26610315e+00 -2.87484258e-01
4.72366400e-02 -1.17229491e-01 7.31118381e-01 5.40266752e-01
6.91119730e-01 -4.30339091e-02 -1.93894014e-01 2.59184986e-01
1.03115022e+00 5.81304312e-01 4.24767584e-01 -1.07855856e-01
7.67424941e-01 7.14137435e-01 6.82642937e-01 5.53102419e-02
4.98247802e-01 7.52778947e-01 -1.90742180e-01 -2.37153806e-02
-2.95337826e-01 -2.60298193e-01 8.58301759e-01 2.07631946e+00
3.26739371e-01 -2.04856157e-01 -9.79143798e-01 1.00412107e+00
-1.58636403e+00 -6.20236456e-01 1.36290863e-01 2.37417626e+00
9.21684921e-01 2.18079522e-01 3.99583340e-01 4.24053580e-01
1.04798687e+00 -1.74183846e-02 -2.91624069e-01 -7.09250212e-01
-2.26455286e-01 1.51449278e-01 2.59847045e-01 9.20190632e-01
-9.07731235e-01 1.38261843e+00 6.02899981e+00 1.41112995e+00
-9.83116210e-01 4.22877610e-01 4.56047446e-01 3.30695838e-01
-3.21270794e-01 -9.11618620e-02 -1.11717999e+00 4.43741053e-01
1.50452399e+00 2.04740793e-01 4.22866493e-01 7.17836380e-01
2.07266212e-01 -8.05280730e-02 -9.73320842e-01 9.57001507e-01
3.09452713e-01 -9.37847078e-01 -9.70713645e-02 1.73274592e-01
7.84461915e-01 3.69179130e-01 -1.52671650e-01 4.99968320e-01
7.89669529e-02 -8.34552050e-01 6.85678482e-01 2.52509058e-01
1.05031681e+00 -1.02599049e+00 7.61396825e-01 5.05252302e-01
-1.58742666e+00 2.90304631e-01 -5.01797497e-01 1.68128237e-01
2.61992872e-01 6.42543674e-01 -1.29051518e+00 5.33367157e-01
4.39764202e-01 4.13077414e-01 -1.37487635e-01 7.13375986e-01
2.45493948e-02 1.41429985e+00 -4.46937650e-01 1.53690437e-02
3.83020818e-01 -4.35344309e-01 5.97857654e-01 1.74844015e+00
5.11018038e-01 -3.40317398e-01 2.96885759e-01 3.26607347e-01
1.12073429e-01 6.22874618e-01 -5.79029918e-01 -1.21818788e-01
8.28097999e-01 9.24741328e-01 -6.76683486e-01 -4.10154700e-01
-4.37669933e-01 1.01011276e+00 2.74677724e-01 2.37634033e-01
-4.11803037e-01 -5.65374255e-01 7.51761019e-01 -2.07543373e-01
5.87307870e-01 -5.18669903e-01 1.47545226e-02 -8.84574473e-01
-1.24393024e-01 -8.36823463e-01 1.30689144e-01 -9.52059329e-02
-1.30481482e+00 1.00904191e+00 -2.25410238e-01 -1.17399001e+00
-7.19348848e-01 -3.75313282e-01 -3.39924186e-01 1.07818174e+00
-1.45934975e+00 -1.08090353e+00 4.00375873e-01 4.75907087e-01
1.12972343e+00 -5.28117776e-01 1.14651144e+00 4.83019769e-01
-5.32995105e-01 8.65235627e-01 5.46466947e-01 2.50494748e-01
7.12678969e-01 -1.31264639e+00 5.20338237e-01 8.55385244e-01
7.35545039e-01 4.03116852e-01 3.91483516e-01 -5.47445059e-01
-1.06698585e+00 -1.04051828e+00 1.38013232e+00 -3.13855290e-01
6.98886335e-01 -8.04198086e-01 -1.04205585e+00 4.93634582e-01
1.28651142e-01 -2.49523044e-01 1.00719655e+00 1.55256942e-01
-1.52283236e-01 -7.72326440e-02 -7.51896024e-01 2.44319215e-01
8.05554271e-01 -1.01131952e+00 -8.89141023e-01 1.07496962e-01
9.97607410e-01 7.81394914e-02 -9.00232613e-01 1.08673066e-01
4.40964818e-01 -5.89198351e-01 6.80502355e-01 -1.28717497e-01
-5.21091938e-01 -3.93251926e-01 -6.99315012e-01 -1.59772849e+00
-8.14749002e-02 -7.02969134e-01 1.61088824e-01 1.67557824e+00
7.97521472e-01 -5.48811793e-01 4.58060473e-01 -3.05528343e-01
-5.89679599e-01 -3.91460270e-01 -1.29004979e+00 -1.17581797e+00
1.61590979e-01 -9.24809694e-01 3.26756805e-01 8.00379574e-01
9.72015038e-02 4.99987304e-01 -3.59002113e-01 3.01238388e-01
4.23608124e-01 -2.23799214e-01 3.96259308e-01 -1.01100814e+00
-2.63644159e-01 -1.40781760e-01 -3.88113171e-01 -1.27307510e+00
6.52966440e-01 -1.09985232e+00 6.76574230e-01 -1.14017355e+00
-2.60792911e-01 -6.16447628e-01 -5.02551138e-01 2.53770381e-01
-1.09263748e-01 3.31906199e-01 4.37465757e-02 2.76667267e-01
-4.15988654e-01 2.96432495e-01 4.48513687e-01 -2.10382044e-02
-6.23791277e-01 1.81901649e-01 -2.15776026e-01 6.62616789e-01
7.92947173e-01 -5.96485257e-01 -3.00268352e-01 -3.06729376e-01
-5.00399530e-01 -8.67447257e-02 -6.04846656e-01 -1.05273438e+00
6.24288023e-02 2.96596825e-01 -1.63195863e-01 -7.73449063e-01
4.46868569e-01 -6.61597311e-01 -3.85044701e-02 2.05279723e-01
-2.51645327e-01 -1.85892418e-01 4.19712141e-02 3.20564121e-01
-6.02303863e-01 -4.85178143e-01 6.54239237e-01 7.92032927e-02
-7.40001738e-01 -2.03173846e-01 -9.53914464e-01 1.00159869e-01
8.72489512e-01 -2.93290168e-01 2.90836245e-01 -3.11224401e-01
-8.13531518e-01 -2.61037856e-01 1.87771544e-01 6.26873195e-01
4.73261148e-01 -1.17300105e+00 -9.39045906e-01 4.67964053e-01
2.14924544e-01 -2.49986455e-01 6.50357231e-02 6.65273964e-01
-7.12623671e-02 6.46748781e-01 4.22913194e-01 -8.86434317e-01
-1.21257782e+00 2.71372199e-01 5.53418361e-02 -2.86086798e-01
-2.13473305e-01 9.97275293e-01 2.76965410e-01 -8.95431876e-01
5.33821166e-01 -1.20955825e-01 -2.77764946e-01 4.10850286e-01
3.38860303e-01 3.17790329e-01 2.97731817e-01 -1.31694674e+00
-5.78395367e-01 7.46999621e-01 -1.09108917e-01 -5.27833462e-01
1.07246494e+00 -7.60466933e-01 2.25770205e-01 1.17323053e+00
1.35768485e+00 4.89178270e-01 -8.79275680e-01 -4.52742189e-01
3.70056272e-01 -1.12231486e-02 1.08162962e-01 -5.81591070e-01
-5.97142637e-01 9.79270756e-01 6.84280097e-01 3.07387888e-01
8.85465324e-01 3.14197749e-01 9.77652252e-01 2.89557576e-01
2.94922173e-01 -1.51673114e+00 -2.72268921e-01 9.08313513e-01
5.09169459e-01 -1.27543640e+00 -7.24998057e-01 -3.64926755e-01
-8.69713068e-01 9.18363333e-01 2.57432491e-01 -4.32853471e-04
7.45721579e-01 3.68913859e-01 5.51414132e-01 2.85290301e-01
-4.60407406e-01 -5.81036150e-01 5.54599881e-01 7.19389439e-01
7.03420579e-01 4.51446593e-01 -3.49420458e-01 6.07103884e-01
-3.91945004e-01 -6.99222684e-01 1.37671471e-01 4.99607116e-01
-6.82085097e-01 -1.41931689e+00 -6.20595813e-01 1.62659779e-01
-3.87833863e-01 -3.20525259e-01 -3.26519847e-01 5.60352266e-01
2.54360229e-01 1.34721398e+00 2.75478095e-01 -7.39231348e-01
1.65713325e-01 6.10774219e-01 8.35512429e-02 -9.28343654e-01
-4.35012400e-01 7.47836709e-01 1.75097495e-01 -2.57324815e-01
-4.58865225e-01 -8.99161279e-01 -1.29179573e+00 4.66800243e-01
-7.26194859e-01 6.77730620e-01 1.06380689e+00 1.05899322e+00
9.44735557e-02 4.21596974e-01 9.84141350e-01 -9.91460443e-01
-3.29680264e-01 -1.22473609e+00 -6.74721539e-01 1.46693602e-01
2.63505936e-01 -3.36844802e-01 -6.66218340e-01 1.32159486e-01] | [14.435632705688477, 6.678610324859619] |
e4a50e39-ab0d-4ae0-8734-64fa8483a9a9 | graph-representation-for-order-aware-visual | null | null | http://openaccess.thecvf.com//content/CVPR2023/html/Qiu_Graph_Representation_for_Order-Aware_Visual_Transformation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Qiu_Graph_Representation_for_Order-Aware_Visual_Transformation_CVPR_2023_paper.pdf | Graph Representation for Order-Aware Visual Transformation | This paper proposes a new visual reasoning formulation that aims at discovering changes between image pairs and their temporal orders. Recognizing scene dynamics and their chronological orders is a fundamental aspect of human cognition. The aforementioned abilities make it possible to follow step-by-step instructions, reason about and analyze events, recognize abnormal dynamics, and restore scenes to their previous states. However, it remains unclear how well current AI systems perform in these capabilities. Although a series of studies have focused on identifying and describing changes from image pairs, they mainly consider those changes that occur synchronously, thus neglecting potential orders within those changes. To address the above issue, we first propose a visual transformation graph structure for conveying order-aware changes. Then, we benchmarked previous methods on our newly generated dataset and identified the issues of existing methods for change order recognition. Finally, we show a significant improvement in order-aware change recognition by introducing a new model that explicitly associates different changes and then identifies changes and their orders in a graph representation. | ['Hirokatsu Kataoka', 'Kenji Iwata', 'Fumiya Matsuzawa', 'Yanjun Sun', 'Yue Qiu'] | 2023-01-01 | null | null | null | cvpr-2023-1 | ['visual-reasoning', 'visual-reasoning'] | ['computer-vision', 'reasoning'] | [ 4.19926494e-01 -3.18779826e-01 -5.72610721e-02 -3.17087471e-01
2.76241213e-01 -8.91713560e-01 1.05973828e+00 6.56729758e-01
-6.09879605e-02 3.90674263e-01 3.39774400e-01 -3.46299678e-01
-2.72690415e-01 -7.22750664e-01 -5.19091964e-01 -1.96989849e-01
-2.97148913e-01 2.73627281e-01 5.61142564e-01 -3.62094641e-01
6.47805095e-01 9.53460515e-01 -2.10730195e+00 4.35169816e-01
6.62661731e-01 6.46337509e-01 -1.55154690e-02 9.27257717e-01
-3.20226461e-01 1.39948988e+00 -6.33882582e-01 6.30258918e-02
-9.06622037e-02 -8.39216113e-01 -1.05061245e+00 5.17914951e-01
6.38419747e-01 -2.83839732e-01 -5.07524610e-01 9.75680053e-01
-8.21380243e-02 2.73006618e-01 6.21737480e-01 -1.48647130e+00
-6.97032571e-01 3.84502947e-01 -3.03063154e-01 8.88948262e-01
1.02469969e+00 1.35929301e-01 8.04309070e-01 -3.44012648e-01
1.09102452e+00 1.13472211e+00 2.72232801e-01 2.51997441e-01
-1.17736197e+00 -2.44262651e-01 6.59452260e-01 1.02164555e+00
-9.97882962e-01 -3.88528287e-01 1.13826489e+00 -7.07168281e-01
1.44098449e+00 7.07498968e-01 1.04278219e+00 8.74607325e-01
2.81537414e-01 7.60899842e-01 1.21380568e+00 -5.69488585e-01
2.40302190e-01 -3.68367851e-01 4.09111410e-01 6.83364272e-01
1.17605209e-01 -1.57823920e-01 -6.53104126e-01 2.95517504e-01
8.00330162e-01 5.76739833e-02 -4.24349129e-01 -5.78849971e-01
-1.52789879e+00 1.45049751e-01 3.69339645e-01 8.15883458e-01
-3.50524217e-01 1.31512294e-02 4.09183621e-01 5.08603811e-01
2.05734476e-01 5.38820684e-01 8.60081892e-03 -3.09700400e-01
-6.69201374e-01 -1.71811152e-02 5.91604233e-01 7.17454791e-01
7.98636556e-01 -7.88415298e-02 -3.20218593e-01 1.82915628e-01
-2.35441342e-01 4.85395901e-02 4.45187300e-01 -8.21159005e-01
2.13066712e-01 1.13970900e+00 1.36000454e-01 -1.55302143e+00
-5.98846138e-01 -2.39907742e-01 -7.25363553e-01 1.22583501e-01
3.86724502e-01 8.06107521e-01 -8.89801919e-01 1.64040589e+00
2.42635563e-01 3.03762645e-01 -2.51291096e-01 6.68795824e-01
6.17514491e-01 5.81414938e-01 -1.22283511e-01 -5.71051240e-01
1.46781003e+00 -6.77421927e-01 -9.83782470e-01 -1.16128214e-01
3.54135275e-01 -5.38430750e-01 1.30778909e+00 2.27921993e-01
-1.06719458e+00 -6.61949098e-01 -9.63384509e-01 -9.07400921e-02
-5.83910882e-01 -2.19001397e-01 6.41708553e-01 2.53345191e-01
-1.30871713e+00 6.55354440e-01 -9.41243112e-01 -7.55617023e-01
-9.43608806e-02 -2.71165352e-02 -4.05830979e-01 2.46979818e-01
-9.42040443e-01 9.65448081e-01 3.71368021e-01 -1.06024131e-01
-5.31258345e-01 -4.79162604e-01 -8.57070386e-01 2.19941195e-02
4.96513277e-01 -7.04471529e-01 1.12387133e+00 -8.68126750e-01
-1.24615836e+00 1.07951939e+00 -5.23615420e-01 -3.57869655e-01
4.95331854e-01 2.03311414e-01 -7.23404825e-01 3.05315554e-01
-1.59628883e-01 3.23360473e-01 7.33133614e-01 -1.36961591e+00
-7.11351454e-01 -2.74015605e-01 4.20254797e-01 3.76602411e-01
-2.32617646e-01 -1.79818068e-02 -5.60466111e-01 -3.63705486e-01
3.80483508e-01 -6.62814140e-01 1.86816184e-03 -1.05846971e-01
-2.46837243e-01 -1.14781164e-01 7.21692383e-01 -6.36773527e-01
1.67621124e+00 -2.04611492e+00 1.68253183e-01 2.77972966e-01
4.66343641e-01 -3.37638967e-02 4.79041040e-02 7.00009406e-01
-4.16821986e-01 3.30861025e-02 -2.49457344e-01 6.91136997e-03
6.90628141e-02 3.15499365e-01 -4.53534633e-01 2.87599891e-01
-8.36791620e-02 8.44603539e-01 -1.15488732e+00 -3.81515652e-01
4.11837518e-01 9.60386693e-02 -3.91676635e-01 1.26635984e-01
-4.76935841e-02 4.76450354e-01 4.29708362e-02 4.97023314e-01
3.39364558e-01 -2.79407531e-01 3.83439541e-01 -5.04955828e-01
-3.72860909e-01 2.25462541e-01 -1.28440356e+00 1.16239285e+00
-6.16194531e-02 9.16126609e-01 -6.77773833e-01 -1.01454508e+00
7.45087445e-01 5.12366816e-02 5.17632604e-01 -1.18427742e+00
-1.48927927e-01 -3.12777460e-01 1.71602704e-02 -8.07286203e-01
7.39827931e-01 3.78806025e-01 4.19977494e-02 5.03613770e-01
-4.18023765e-01 -2.25452989e-01 8.35820019e-01 3.18509459e-01
1.41304576e+00 1.14739738e-01 1.00572932e+00 4.28766608e-02
7.21702278e-01 2.29546234e-01 3.50334257e-01 7.10738003e-01
-4.51395780e-01 4.09976512e-01 7.66126812e-01 -9.17680800e-01
-7.82277524e-01 -1.23542523e+00 4.10578430e-01 8.86257052e-01
4.11384404e-01 -6.51524603e-01 -4.43951368e-01 -6.06111825e-01
-1.54122472e-01 9.13173735e-01 -8.46952140e-01 -2.29135916e-01
-7.05433309e-01 -3.68324876e-01 2.09460229e-01 4.29149091e-01
5.12913048e-01 -1.19876540e+00 -1.08032215e+00 1.15311310e-01
-5.05290031e-01 -1.02040541e+00 -3.57050389e-01 -1.02763288e-01
-7.33228505e-01 -1.44957721e+00 3.65481861e-02 -6.62385702e-01
9.64497328e-01 4.78011996e-01 1.33692360e+00 1.08759075e-01
-4.31886137e-01 9.25812960e-01 -4.06460136e-01 -1.84236884e-01
-5.63882232e-01 -5.05291164e-01 -6.11076355e-02 1.93147808e-02
1.08662821e-01 -7.24025309e-01 -4.55696911e-01 2.52572447e-01
-1.06659973e+00 3.68716866e-01 2.32846409e-01 2.92778254e-01
5.17028570e-01 3.30815583e-01 -7.47243613e-02 -8.69808257e-01
7.80937135e-01 -3.95561792e-02 -4.49498832e-01 6.94563448e-01
-7.15360343e-01 1.65013418e-01 5.16801834e-01 -4.65669453e-01
-1.17333364e+00 5.30671589e-02 3.73906225e-01 -2.04537421e-01
-4.17680264e-01 4.36547935e-01 3.15545201e-02 1.55577853e-01
5.80673754e-01 5.73553026e-01 -3.25740337e-01 -1.09950006e-02
5.13783753e-01 -5.57484059e-03 9.91533160e-01 -3.13018590e-01
7.52182305e-01 7.34626472e-01 3.62474769e-02 -6.26846850e-01
-3.53203267e-01 -6.22306168e-01 -1.03716421e+00 -7.87402391e-01
6.36936843e-01 -4.02907312e-01 -8.01482379e-01 5.39210677e-01
-1.15630472e+00 -1.87105343e-01 -5.41039646e-01 8.62497389e-02
-8.10241640e-01 7.61772513e-01 -5.61245084e-01 -5.67314327e-01
1.53720170e-01 -7.66006947e-01 7.19740689e-01 1.06882721e-01
-6.37165189e-01 -1.16135669e+00 3.34878802e-01 -2.40531587e-03
2.65259594e-01 4.05873209e-01 1.27135324e+00 -1.91349611e-01
-6.33902907e-01 1.89135820e-01 -1.09772831e-01 -3.81703526e-01
6.79249167e-01 3.16197693e-01 -4.52366561e-01 -1.02430992e-01
1.12483658e-01 4.78196323e-01 6.12521350e-01 1.43366709e-01
1.06153142e+00 -4.42773283e-01 -3.55829626e-01 3.39143425e-01
1.26398349e+00 7.74781704e-01 8.08975279e-01 6.43724263e-01
6.20730400e-01 5.65038085e-01 3.91506314e-01 4.71808523e-01
6.75368130e-01 7.13939071e-01 4.52739596e-01 -1.91735458e-02
-3.55243355e-01 -2.66182244e-01 1.83091715e-01 8.49991977e-01
-3.01463515e-01 -2.99424022e-01 -1.07492197e+00 6.12512648e-01
-1.98472536e+00 -1.37214983e+00 -3.17545027e-01 1.91360414e+00
6.89721107e-01 1.81095466e-01 7.39434436e-02 4.78575200e-01
7.64617443e-01 3.85675251e-01 -5.20669758e-01 -3.86555016e-01
-9.47878584e-02 -1.35520026e-01 3.84778753e-02 3.54891330e-01
-9.99426842e-01 8.97153139e-01 7.27746201e+00 1.03796653e-01
-1.02864969e+00 -3.13961804e-01 2.77063310e-01 1.81129038e-01
-4.62643802e-01 1.54148430e-01 -1.89125568e-01 1.47182688e-01
4.90142435e-01 -3.66908550e-01 6.54384196e-01 2.94473678e-01
2.83596307e-01 -3.69147807e-01 -1.37414610e+00 1.31314313e+00
5.61772108e-01 -1.29691339e+00 4.68469560e-01 -3.77859741e-01
5.84465206e-01 -7.57263839e-01 -1.47044986e-01 -7.64752030e-02
2.49828979e-01 -6.89302146e-01 9.44347739e-01 1.07345450e+00
2.97468960e-01 -4.01627749e-01 2.04031408e-01 1.65166423e-01
-1.51157403e+00 -1.67516200e-03 -5.13415672e-02 -6.91720963e-01
7.31886178e-02 4.17966247e-01 -8.72174561e-01 7.80148923e-01
6.84674203e-01 1.02160716e+00 -1.13349450e+00 9.43822742e-01
-4.11449134e-01 3.49406183e-01 1.10614538e-01 4.15733606e-02
-1.95322305e-01 -2.56588429e-01 6.69270635e-01 1.04657280e+00
3.44127476e-01 1.05859116e-01 3.79946455e-03 6.53635919e-01
3.42415839e-01 -1.56109288e-01 -8.11322987e-01 -1.33362576e-01
3.43441814e-01 9.92052734e-01 -1.43177748e+00 -6.25401556e-01
-3.87171477e-01 1.30117798e+00 2.71004647e-01 4.02961642e-01
-7.49473333e-01 -2.29841664e-01 6.62344992e-01 1.32499129e-01
6.36942461e-02 -5.89424789e-01 -5.34975752e-02 -1.29015148e+00
2.33813137e-01 -9.31087673e-01 6.68476462e-01 -1.03883350e+00
-9.94956315e-01 5.39674282e-01 3.87796313e-01 -1.41542506e+00
-4.69370127e-01 -3.12397957e-01 -5.81374347e-01 1.54876232e-01
-1.31960869e+00 -8.75028431e-01 -7.01148152e-01 8.59931111e-01
5.27239561e-01 3.64498287e-01 5.54918051e-01 9.51524898e-02
-4.68471825e-01 1.01404451e-01 -3.02249163e-01 -8.39488506e-02
5.15134096e-01 -1.39088738e+00 5.98772109e-01 1.43322599e+00
5.08974254e-01 6.08438909e-01 9.93251562e-01 -7.26317763e-01
-1.27818644e+00 -6.85929716e-01 1.04519928e+00 -4.48789626e-01
8.27148259e-01 -2.84803420e-01 -1.17572176e+00 9.24495339e-01
3.37524563e-01 -3.05510193e-01 1.15001820e-01 -7.76040778e-02
-5.48962295e-01 -1.74078390e-01 -7.40881801e-01 1.08515215e+00
1.74987066e+00 -6.86308205e-01 -9.20805514e-01 1.84986621e-01
4.32275444e-01 -5.10245919e-01 -4.41542774e-01 2.70198077e-01
2.63078451e-01 -1.33235407e+00 9.55481052e-01 -5.57871819e-01
1.01548962e-01 -7.45593667e-01 1.52727827e-01 -1.44325769e+00
-7.30084658e-01 -6.07415557e-01 -1.57308772e-01 1.17376828e+00
-2.29628831e-01 -7.47915268e-01 3.87256145e-01 4.54172790e-01
-1.32331610e-01 1.09321028e-02 -6.19588017e-01 -9.00173962e-01
-7.84890056e-01 -4.46233511e-01 7.94080019e-01 1.20544755e+00
3.03965867e-01 7.42551908e-02 -1.75926924e-01 1.71709269e-01
1.65559530e-01 4.60337639e-01 6.93933606e-01 -1.16300082e+00
-1.29729025e-02 -8.46859992e-01 -9.32069302e-01 -6.65400147e-01
1.43764794e-01 -7.97351956e-01 -6.05390482e-02 -1.98739803e+00
2.91858196e-01 2.64186442e-01 -2.86905408e-01 6.88336790e-01
-1.71403050e-01 6.09457232e-02 2.88912266e-01 3.56171936e-01
-7.48569429e-01 1.46967053e-01 1.28384304e+00 -3.05302531e-01
-2.92640507e-01 -3.87131214e-01 -4.22314078e-01 6.56227052e-01
6.05841815e-01 -4.42696735e-02 -7.24860191e-01 -2.50050813e-01
5.26565731e-01 -5.37862070e-02 4.93897200e-01 -1.28080559e+00
5.52222788e-01 -6.12904489e-01 2.59045035e-01 -7.31433511e-01
-3.23182903e-02 -8.92845333e-01 7.53882587e-01 6.80026054e-01
-2.52955705e-01 6.37512982e-01 2.48890534e-01 5.04339755e-01
-4.29186732e-01 1.20454282e-01 4.09605354e-01 -1.00983664e-01
-1.55433595e+00 -9.63843688e-02 -8.84442747e-01 -2.95712918e-01
1.44833982e+00 -5.16224444e-01 -6.90125287e-01 -5.75445533e-01
-7.57761717e-01 -1.15255741e-02 7.18282938e-01 6.15944207e-01
6.88060403e-01 -1.35746479e+00 -1.75766066e-01 2.69754082e-01
4.54564303e-01 -5.74185908e-01 4.99984741e-01 7.24034846e-01
-7.78208911e-01 3.70494425e-01 -7.12567627e-01 -5.98654926e-01
-1.60056722e+00 8.91863823e-01 2.39187881e-01 -3.54689032e-01
-6.88166618e-01 2.24285990e-01 1.83133468e-01 -9.59518328e-02
-3.28085721e-02 -8.99843812e-01 -6.20179117e-01 3.14098030e-01
5.52346408e-01 4.49605972e-01 1.88240051e-01 -6.76203668e-01
-6.34692252e-01 7.71264493e-01 1.97751984e-01 1.21534709e-02
1.04941082e+00 -4.13757175e-01 -4.63518053e-01 8.34767997e-01
7.05454886e-01 -3.61062847e-02 -1.13853645e+00 -1.81486040e-01
2.81353027e-01 -5.15958667e-01 -5.19452870e-01 -6.86816812e-01
-6.22729897e-01 5.63838363e-01 4.21691567e-01 5.39835572e-01
1.65767360e+00 6.26612827e-02 3.41183037e-01 4.35274780e-01
3.02208394e-01 -8.22501242e-01 4.08098191e-01 7.96106935e-01
1.00314260e+00 -7.68654704e-01 4.67951931e-02 -5.88832855e-01
-5.69618702e-01 1.26808071e+00 6.11107230e-01 2.20540762e-01
2.48843178e-01 2.21847855e-02 2.64729038e-02 -3.78605902e-01
-7.55358398e-01 -6.15137696e-01 5.52313805e-01 7.08421588e-01
1.53316081e-01 2.03595236e-02 -2.54423022e-01 -2.48442560e-01
-2.13178024e-01 -1.93011895e-01 7.96036303e-01 1.26187563e+00
-3.40899855e-01 -9.56300139e-01 -4.70869154e-01 2.53903747e-01
2.09995270e-01 1.70303404e-01 -6.89221919e-01 9.59338903e-01
-3.32254283e-02 9.37085271e-01 5.04330993e-01 -4.46969330e-01
6.86433017e-01 8.60372335e-02 7.39741266e-01 -4.43293720e-01
-4.38188314e-01 -4.30649549e-01 -1.11069893e-02 -8.74949396e-01
-7.15099573e-01 -8.63577485e-01 -1.18493068e+00 -2.20586896e-01
5.33037364e-01 -2.57032782e-01 3.27090353e-01 9.57494020e-01
3.83438438e-01 9.53953505e-01 4.67108846e-01 -4.90981609e-01
1.81657180e-01 -4.79680479e-01 -4.15995747e-01 1.05654585e+00
3.09293658e-01 -6.40973985e-01 -2.67679811e-01 6.22992992e-01] | [8.777017593383789, 0.8789583444595337] |
865a6267-f091-4c5c-acf3-ef5fa92101f7 | temporal-knowledge-graph-completion-using-box | 2109.08970 | null | https://arxiv.org/abs/2109.08970v1 | https://arxiv.org/pdf/2109.08970v1.pdf | Temporal Knowledge Graph Completion using Box Embeddings | Knowledge graph completion is the task of inferring missing facts based on existing data in a knowledge graph. Temporal knowledge graph completion (TKGC) is an extension of this task to temporal knowledge graphs, where each fact is additionally associated with a time stamp. Current approaches for TKGC primarily build on existing embedding models which are developed for (static) knowledge graph completion, and extend these models to incorporate time, where the idea is to learn latent representations for entities, relations, and timestamps and then use the learned representations to predict missing facts at various time steps. In this paper, we propose BoxTE, a box embedding model for TKGC, building on the static knowledge graph embedding model BoxE. We show that BoxTE is fully expressive, and possesses strong inductive capacity in the temporal setting. We then empirically evaluate our model and show that it achieves state-of-the-art results on several TKGC benchmarks. | ['İsmail İlkan Ceylan', 'Ralph Abboud', 'Johannes Messner'] | 2021-09-18 | null | null | null | null | ['temporal-knowledge-graph-completion'] | ['knowledge-base'] | [-3.53868723e-01 5.89005470e-01 -8.62183213e-01 -9.09702256e-02
-3.49029243e-01 -6.09151781e-01 6.86516881e-01 5.25240421e-01
-2.29833320e-01 7.87328780e-01 6.44099832e-01 -2.83853620e-01
-4.86711264e-01 -1.20010722e+00 -8.10955465e-01 -2.47692838e-01
-7.77896047e-01 7.39591181e-01 3.64971906e-01 4.04339358e-02
-2.57357150e-01 2.00459972e-01 -1.06502223e+00 8.66385847e-02
3.43858123e-01 6.08813643e-01 -6.29470646e-01 7.42912948e-01
-2.17240244e-01 1.38962138e+00 -2.34890938e-01 -8.50583255e-01
-9.83996987e-02 2.56488174e-01 -1.16116667e+00 -2.90497333e-01
6.37395214e-03 -5.89407444e-01 -1.36913693e+00 4.53048199e-01
-2.22020932e-02 4.78189856e-01 3.93579185e-01 -1.76972914e+00
-9.98013377e-01 1.03695226e+00 -1.30823717e-01 2.61207134e-01
6.66076481e-01 -2.82359391e-01 1.35255301e+00 -7.53273606e-01
1.03123593e+00 1.02291071e+00 8.38464677e-01 2.99647510e-01
-1.18582511e+00 -2.61085719e-01 5.04463971e-01 7.61086166e-01
-1.23995769e+00 -2.52306372e-01 6.00511491e-01 -4.26896811e-01
1.55503583e+00 -1.31345063e-01 7.60686040e-01 1.03083265e+00
6.45934492e-02 8.36274683e-01 5.69891870e-01 -4.16163892e-01
1.19499877e-01 -2.30522528e-01 5.45763791e-01 1.07684529e+00
4.78810281e-01 1.64708525e-01 -1.14006770e+00 -3.27088863e-01
5.45757651e-01 1.60799891e-01 -1.67571634e-01 -6.54321432e-01
-1.36420655e+00 7.60753036e-01 3.10132831e-01 9.44625735e-02
-2.49592945e-01 7.56041050e-01 6.30724311e-01 4.67339963e-01
6.44734919e-01 1.33533016e-01 -7.35383391e-01 -2.40368590e-01
-6.74182832e-01 2.38208175e-01 1.37478209e+00 1.21869791e+00
8.24091852e-01 -9.67683718e-02 -3.41225535e-01 1.60780802e-01
1.28007427e-01 8.61124769e-02 3.51580828e-01 -9.47724044e-01
3.76732737e-01 8.66840541e-01 3.02231550e-01 -9.47528362e-01
-2.88970858e-01 -1.39787480e-01 -1.76081076e-01 -3.09959710e-01
1.90033957e-01 6.91727176e-02 -1.05323422e+00 1.70436323e+00
3.72365534e-01 9.84129548e-01 4.04399663e-01 3.46653104e-01
8.08481872e-01 6.44639611e-01 3.93230915e-02 -1.69376701e-01
1.18859935e+00 -1.01802349e+00 -8.78761351e-01 -1.11101627e-01
8.32832992e-01 4.81316224e-02 5.82522035e-01 1.49617210e-01
-7.68234909e-01 3.77874151e-02 -9.54012513e-01 -4.71746475e-01
-1.07811368e+00 -1.30187139e-01 1.27051282e+00 3.26996028e-01
-1.19988120e+00 7.79888511e-01 -1.32492483e+00 -2.91429639e-01
2.14658782e-01 2.63646960e-01 -6.35293603e-01 -4.88936126e-01
-1.57829285e+00 8.05467367e-01 1.03185344e+00 3.99639234e-02
-9.69621599e-01 -8.32078159e-01 -1.34989440e+00 2.56079435e-01
1.05296016e+00 -9.34981823e-01 1.20418131e+00 -4.45093736e-02
-1.14844990e+00 5.77727675e-01 -3.69208127e-01 -8.66014719e-01
7.42800385e-02 -3.40877295e-01 -1.07240772e+00 2.27978807e-02
2.61005789e-01 1.52199969e-01 6.69432759e-01 -1.03751469e+00
-4.21258360e-01 -3.16028297e-01 4.73499954e-01 -2.27005020e-01
-3.22558731e-01 -2.70483524e-01 -9.41053152e-01 -4.76595581e-01
-1.05177499e-01 -7.23179758e-01 1.83546028e-04 -8.00883025e-02
-4.27729636e-01 -4.77110237e-01 1.03064001e+00 -9.45688426e-01
1.70731175e+00 -1.88502502e+00 2.23014623e-01 3.24110061e-01
5.28696597e-01 -5.80435386e-03 1.35037586e-01 9.17833388e-01
-1.03371970e-01 -2.93355342e-02 -1.12303227e-01 -4.66869771e-01
4.75415975e-01 8.17112446e-01 -7.31226444e-01 2.39084154e-01
-2.14696415e-02 1.46550632e+00 -1.36808562e+00 -5.26610732e-01
8.26647282e-02 3.68383914e-01 -3.89360964e-01 -1.13781258e-01
-7.54183233e-01 -4.63136554e-01 -5.33375561e-01 7.63014436e-01
1.84369206e-01 -6.14797533e-01 7.29130626e-01 -2.23621935e-01
2.61696190e-01 2.05561176e-01 -1.16463947e+00 1.90914524e+00
-2.72798955e-01 4.10338283e-01 -5.30466437e-01 -6.43176615e-01
4.35621172e-01 6.70082986e-01 6.36007130e-01 -1.57420069e-01
-2.52811015e-01 -1.41447499e-01 -6.87415004e-01 -4.85099554e-01
8.40817630e-01 1.49424747e-01 -2.73424119e-01 6.96150482e-01
5.44867694e-01 2.79999018e-01 3.95757437e-01 9.43103611e-01
1.47063696e+00 4.85945791e-01 3.53268057e-01 3.75226557e-01
1.02972388e-02 6.13141172e-02 6.46538258e-01 4.92503196e-01
-3.99654731e-02 -3.37557971e-01 8.21914673e-01 -6.33884132e-01
-6.00703776e-01 -1.33699763e+00 6.12823129e-01 9.28501964e-01
-9.70958844e-02 -1.27353942e+00 2.57302113e-02 -1.13362980e+00
6.68436468e-01 8.59927475e-01 -8.97148073e-01 -3.62131149e-01
-3.68372381e-01 -1.67408854e-01 6.11033142e-01 1.00736618e+00
4.58208732e-02 -7.20918417e-01 -2.12732852e-01 5.63854694e-01
-1.36257812e-01 -1.29970241e+00 -2.77072072e-01 -2.11953800e-02
-8.28785419e-01 -1.51132977e+00 1.09839089e-01 -6.26230657e-01
4.14841563e-01 -6.18197210e-03 1.28343248e+00 -5.44685833e-02
-1.42168164e-01 1.14570785e+00 -4.84337479e-01 -3.18494767e-01
1.49532467e-01 -2.74921861e-02 1.84602365e-01 -1.25369444e-01
4.24134880e-01 -7.92655468e-01 -3.59963119e-01 -1.87300846e-01
-1.17870677e+00 -1.50140032e-01 6.42270893e-02 7.38751948e-01
7.70520091e-01 4.39218938e-01 5.10277212e-01 -1.26731122e+00
4.26926613e-01 -7.53602386e-01 -5.48183501e-01 6.79993749e-01
-8.46683562e-01 4.02507454e-01 3.80638778e-01 -2.30326906e-01
-1.02984893e+00 -1.49003968e-01 4.14419264e-01 -7.77918458e-01
5.10108292e-01 1.19376123e+00 3.64249386e-02 5.07077239e-02
2.91163534e-01 2.08761334e-01 -3.96914661e-01 -4.00499761e-01
1.21139514e+00 -7.48871863e-02 9.32188451e-01 -9.12394166e-01
1.00717294e+00 8.73843729e-01 9.35692166e-04 -2.00124905e-01
-1.03089643e+00 -5.71637869e-01 -6.27093613e-01 2.29156855e-02
4.62907791e-01 -1.12197065e+00 -9.05799210e-01 -9.02224407e-02
-9.23364282e-01 -6.10506117e-01 -6.56533420e-01 5.96425354e-01
-6.12688243e-01 5.48495889e-01 -7.53041148e-01 -6.34068966e-01
-1.86161846e-01 -2.54877508e-01 9.21655655e-01 -1.43885419e-01
-1.98876068e-01 -1.59429038e+00 3.45287859e-01 9.44188163e-02
1.03713617e-01 5.42988300e-01 1.19718897e+00 -5.90709269e-01
-9.22163904e-01 -3.47177476e-01 -6.95704147e-02 -1.77548796e-01
1.18504047e-01 -3.03917937e-02 -6.71133161e-01 -3.30259174e-01
-6.84791386e-01 -4.30989236e-01 1.14509809e+00 -4.64584772e-03
1.05493307e+00 -5.45903146e-01 -8.41191351e-01 6.17552638e-01
1.68182588e+00 -1.82241380e-01 5.61062694e-01 1.96195349e-01
7.38254964e-01 1.15907915e-01 5.31783044e-01 5.35384297e-01
9.99584198e-01 5.54563582e-01 2.64673054e-01 4.93880779e-01
-1.48754969e-01 -8.83310258e-01 2.99089789e-01 8.07964265e-01
-2.60632604e-01 -1.53263152e-01 -8.79697621e-01 1.28169501e+00
-2.30069041e+00 -1.11159611e+00 1.66435882e-01 1.86583197e+00
1.01703787e+00 8.32994431e-02 -2.24569570e-02 3.72825079e-02
3.56234580e-01 4.85292733e-01 -6.35380685e-01 -3.04348171e-01
-2.27667615e-02 3.96920741e-01 7.40119100e-01 6.60771132e-01
-9.70578134e-01 1.30671537e+00 6.88187885e+00 4.40280497e-01
-4.60763961e-01 3.52622598e-01 -2.91221201e-01 -9.63468328e-02
-6.63037360e-01 5.93685508e-01 -6.67433143e-01 5.46731390e-02
8.59126031e-01 -7.64442265e-01 6.49071217e-01 7.88563609e-01
-4.93555069e-01 9.96491164e-02 -1.41021562e+00 7.14737296e-01
9.21028629e-02 -1.59475017e+00 -2.43954919e-02 -1.58499256e-01
9.76538956e-01 -1.10923313e-01 -3.56272161e-01 7.89806008e-01
1.13162374e+00 -8.62537205e-01 3.38041753e-01 7.69002497e-01
9.75342929e-01 -6.21148705e-01 3.69359523e-01 -1.00480221e-01
-1.55402493e+00 -1.78853989e-01 -1.61732450e-01 -5.78682721e-02
4.42802250e-01 7.01197922e-01 -1.05269599e+00 1.39262807e+00
5.34456313e-01 1.15504777e+00 -4.74284142e-01 8.46012950e-01
-7.44890153e-01 7.65436113e-01 -1.86261639e-01 4.65091914e-01
-6.15375265e-02 2.07845569e-01 4.50276405e-01 9.96541202e-01
1.61714047e-01 -6.73325062e-02 2.04281896e-01 6.73526525e-01
-4.88770515e-01 -5.15048802e-01 -8.60887289e-01 -5.33847928e-01
7.14477360e-01 8.48273873e-01 -3.94855112e-01 -6.09115601e-01
-6.93177104e-01 1.05684841e+00 6.47751033e-01 7.48315394e-01
-7.86409199e-01 -4.14114922e-01 6.68857396e-01 -2.76439965e-01
6.51315033e-01 -5.58897674e-01 2.15531990e-01 -1.46362233e+00
2.23750651e-01 -2.96136469e-01 1.23058796e+00 -8.52700412e-01
-1.22877109e+00 -1.35052469e-04 2.82080114e-01 -5.38576961e-01
-3.67143065e-01 -4.71063077e-01 -5.46556711e-01 4.93790179e-01
-1.75179923e+00 -1.50383890e+00 -1.23991117e-01 1.07260394e+00
-1.06017135e-01 1.74185872e-01 1.07283199e+00 3.63883115e-02
-3.52798790e-01 4.12035763e-01 1.47513121e-01 3.27865154e-01
5.16346633e-01 -1.47770381e+00 7.24765122e-01 1.00379717e+00
4.34237510e-01 7.36546159e-01 4.36057210e-01 -1.10938585e+00
-1.87192309e+00 -1.45582545e+00 1.31266522e+00 -9.05684173e-01
1.23562801e+00 -3.49831462e-01 -8.34627330e-01 1.94734192e+00
-1.42889708e-01 3.44541728e-01 6.46821141e-01 7.44684994e-01
-1.03379202e+00 -8.79766420e-02 -7.22249925e-01 5.75136602e-01
1.33634281e+00 -9.35450196e-01 -1.10636604e+00 4.76685613e-01
1.44904625e+00 -4.99762654e-01 -1.48478305e+00 1.86253265e-01
4.80617702e-01 -3.15843582e-01 1.08936143e+00 -1.21355128e+00
1.60737662e-03 -4.47530627e-01 -7.38036772e-03 -1.22720349e+00
-3.16850156e-01 -8.36834013e-01 -1.45603049e+00 1.06978643e+00
4.36939955e-01 -7.24523902e-01 1.01757133e+00 7.19886541e-01
-2.97712814e-02 -6.37422323e-01 -9.64344025e-01 -1.16863215e+00
-5.11116266e-01 -5.72563648e-01 9.53269601e-01 1.32050169e+00
5.58240473e-01 5.20385355e-02 -5.53768337e-01 4.56917346e-01
5.41444719e-01 5.03628790e-01 6.13704264e-01 -1.29547203e+00
-4.03765440e-01 2.15752810e-01 -5.15526593e-01 -6.00881577e-01
5.10743499e-01 -1.06719196e+00 -4.60709602e-01 -2.15647531e+00
8.44509006e-02 -6.97603673e-02 -4.59922224e-01 1.08539462e+00
-1.10806592e-01 -4.09207672e-01 -2.40777507e-01 -8.27171206e-02
-9.94323313e-01 7.17715561e-01 7.68597484e-01 -2.44606271e-01
-1.92538142e-01 -3.17199230e-01 -6.10807419e-01 2.60101438e-01
4.98764485e-01 -3.85907203e-01 -8.13603938e-01 -5.41283429e-01
8.02544653e-01 2.21259236e-01 5.66969991e-01 -5.50127983e-01
7.52810299e-01 -4.70068976e-02 5.01166619e-02 -4.83126074e-01
5.62086999e-01 -8.32386494e-01 3.13136429e-01 1.51444241e-01
7.94827640e-02 -9.81744472e-03 3.15860927e-01 1.33388424e+00
-4.12102193e-01 3.25194150e-01 -3.61839712e-01 1.05119951e-01
-1.38991666e+00 8.11176956e-01 9.60439369e-02 4.42492813e-02
1.02256513e+00 -5.11516295e-02 -7.38250375e-01 -4.49329466e-01
-1.17005229e+00 7.50153899e-01 2.31928617e-01 3.64726067e-01
8.26798320e-01 -1.64925086e+00 -2.52355874e-01 -3.12654257e-01
6.20583892e-01 4.69407514e-02 2.83017993e-01 7.19134450e-01
-4.53363135e-02 4.61019725e-01 3.26724797e-01 2.11032301e-01
-8.57471943e-01 1.12905252e+00 -3.69621925e-02 -6.84398651e-01
-8.48387539e-01 6.42482936e-01 -3.76049012e-01 -4.11581427e-01
3.68604153e-01 -6.02789521e-01 6.04262501e-02 -4.25019376e-02
2.59842694e-01 5.38445950e-01 -2.46004178e-03 -8.57578814e-02
-4.81146187e-01 2.03131765e-01 -1.88672334e-01 -8.44842046e-02
1.42821217e+00 1.30752176e-01 -2.39947990e-01 5.07429123e-01
1.00511849e+00 -2.69913021e-03 -1.04265392e+00 -9.18927252e-01
3.15184593e-01 -4.19536352e-01 -3.53009328e-02 -7.84025133e-01
-7.92335093e-01 2.42796361e-01 -2.85041839e-01 -4.16057818e-02
8.88714671e-01 2.72226572e-01 9.46255565e-01 6.97573125e-01
5.41578293e-01 -1.17240298e+00 5.82664693e-03 6.15694046e-01
5.92164159e-01 -8.14027369e-01 2.64559746e-01 -4.94239241e-01
-5.27329206e-01 1.00971043e+00 3.36212009e-01 7.66897723e-02
9.49440360e-01 1.90745611e-02 -5.21274090e-01 -7.43024826e-01
-1.24328768e+00 -4.57893282e-01 -1.07009141e-02 7.68421173e-01
-1.60426646e-01 3.31407785e-01 8.65965337e-02 8.96277964e-01
-1.62993111e-02 6.60990775e-01 4.68246460e-01 1.28798699e+00
-7.47544169e-02 -1.32488430e+00 9.52778384e-02 4.74637598e-01
-1.31337702e-01 -5.34320176e-02 -3.83790523e-01 1.04108799e+00
-2.58372314e-02 8.46569538e-01 -2.27684170e-01 -5.66971004e-01
3.23133975e-01 5.49495935e-01 5.92558265e-01 -7.70347118e-01
5.02961352e-02 -6.93632305e-01 5.63636184e-01 -8.83586228e-01
-2.86826521e-01 -4.03990924e-01 -1.28509688e+00 -5.44202685e-01
-9.07611623e-02 3.03402454e-01 2.49724299e-01 7.26683795e-01
5.61039567e-01 6.34964705e-01 4.32050377e-02 -1.55910850e-01
-8.38748589e-02 -4.36248004e-01 -9.52437043e-01 3.57421488e-01
2.26308152e-01 -8.47735286e-01 7.33310264e-03 2.33990788e-01] | [8.568046569824219, 7.9064860343933105] |
aa0429fe-0b08-4261-918f-e7854e4c89ad | a-report-on-the-2020-vua-and-toefl-metaphor | null | null | https://aclanthology.org/2020.figlang-1.3 | https://aclanthology.org/2020.figlang-1.3.pdf | A Report on the 2020 VUA and TOEFL Metaphor Detection Shared Task | In this paper, we report on the shared task on metaphor identification on VU Amsterdam Metaphor Corpus and on a subset of the TOEFL Native Language Identification Corpus. The shared task was conducted as apart of the ACL 2020 Workshop on Processing Figurative Language. | ['Egon Stemle', 'Chee Wee (Ben) Leong', 'Xianyang Chen', 'Beata Beigman Klebanov', 'Rutuja Ubale', 'Chris Hamill'] | 2020-07-01 | null | null | null | ws-2020-7 | ['native-language-identification'] | ['natural-language-processing'] | [-6.75350521e-03 1.88359588e-01 -2.14983061e-01 -2.42632017e-01
-3.14176172e-01 -9.55731034e-01 1.19430208e+00 2.49408409e-01
-8.25018227e-01 6.79066658e-01 8.52562249e-01 -7.13273227e-01
-1.83184016e-02 -2.09965557e-01 7.94460922e-02 2.03109980e-01
-4.88915294e-03 6.44516468e-01 -2.39418432e-01 -7.00592399e-01
6.55124545e-01 1.80032089e-01 -6.82147026e-01 7.99573421e-01
4.68909293e-01 3.50016832e-01 4.13818121e-01 7.25628316e-01
-3.69770736e-01 8.20121050e-01 -7.78970838e-01 -3.14715713e-01
-7.32854977e-02 -1.20297849e-01 -1.32739723e+00 -6.38656676e-01
2.98282474e-01 5.21657988e-02 -3.04640681e-02 7.61309981e-01
4.01204526e-01 3.32772344e-01 9.73891735e-01 -8.67074013e-01
-6.70721352e-01 1.52134204e+00 1.27954865e-02 7.96099365e-01
9.79802430e-01 -4.73517478e-01 1.38671303e+00 -1.20605934e+00
1.20969200e+00 2.01953936e+00 5.16250134e-01 7.34854639e-01
-1.49057615e+00 -4.98604685e-01 2.92136371e-01 3.94397736e-01
-1.55081880e+00 -7.06425190e-01 1.17979300e+00 -5.06915927e-01
1.67104781e+00 4.52260941e-01 5.05574882e-01 1.64327872e+00
-1.82945326e-01 9.57196295e-01 1.09689391e+00 -1.11040533e+00
-8.43910575e-02 3.19473207e-01 4.03195292e-01 2.55942732e-01
-4.41620886e-01 1.57925487e-01 -7.04620838e-01 -5.23315370e-01
8.26538324e-01 -1.12718868e+00 -3.42476815e-01 -3.42727788e-02
-1.45389295e+00 1.23249900e+00 2.84930825e-01 1.12913013e+00
7.44796097e-02 1.97514459e-01 1.11245656e+00 5.39681077e-01
4.61725861e-01 9.65173900e-01 -1.21035703e-01 -2.07083762e-01
-4.79067951e-01 1.55347854e-01 8.68694663e-01 7.39870250e-01
-9.99228805e-02 -3.92597169e-02 3.85201991e-01 1.34371519e+00
3.13722730e-01 9.12007242e-02 9.52494264e-01 -5.11356831e-01
4.28413540e-01 2.41329744e-01 7.39408955e-02 -1.09307802e+00
-4.04368520e-01 1.28101990e-01 -1.20112464e-01 -5.54526687e-01
1.27272829e-01 1.14108073e-02 -6.14229977e-01 1.93325174e+00
-2.32240647e-01 -3.68222982e-01 2.69109368e-01 9.39468384e-01
1.21431613e+00 3.39712381e-01 4.73604411e-01 -2.11386889e-01
1.56305504e+00 -5.60470402e-01 -6.91285014e-01 -3.35199684e-01
1.14812934e+00 -1.16107881e+00 1.38767254e+00 2.95153916e-01
-9.35954690e-01 -7.56612897e-01 -1.21925843e+00 -5.78017950e-01
-9.35317993e-01 1.21260814e-01 7.53450632e-01 7.09227085e-01
-1.18701041e+00 -1.33415908e-01 -3.13003570e-01 -1.02372527e+00
-2.77887374e-01 -1.70705810e-01 -2.59912103e-01 4.27217185e-01
-1.87629330e+00 1.37425768e+00 8.66729856e-01 -1.62524715e-01
-1.36213899e-01 -5.72518706e-01 -9.28761899e-01 -6.09308779e-01
3.17209959e-02 -4.35384512e-01 1.15602756e+00 -1.07526541e+00
-1.12389159e+00 1.81099200e+00 5.40891774e-02 -5.85157812e-01
3.98373336e-01 -5.92234433e-01 -9.65189159e-01 -1.53446987e-01
4.53104854e-01 1.03352511e+00 3.54290247e-01 -1.14736855e+00
-2.53674805e-01 -5.19085005e-02 3.08192044e-01 9.52658504e-02
-4.59627628e-01 4.84350681e-01 3.91600549e-01 -1.17794836e+00
-1.18374787e-01 -7.41377354e-01 4.15140659e-01 -6.45338714e-01
-3.09714854e-01 -7.24728107e-01 7.50906765e-01 -8.05525243e-01
1.43927431e+00 -2.15227604e+00 5.94798625e-01 8.85826498e-02
1.01803161e-01 -2.33793974e-01 -8.29704031e-02 8.85522962e-01
-2.01558173e-01 5.23227394e-01 -1.32740021e-01 -2.71928132e-01
3.46907645e-01 2.96145111e-01 -9.36053872e-01 1.63197935e-01
-1.64654195e-01 1.11831260e+00 -1.10950804e+00 -8.01393092e-01
4.69256282e-01 9.55907553e-02 -2.55256109e-02 -3.40674639e-01
-1.72575742e-01 2.56589875e-02 -1.21574976e-01 2.72010982e-01
8.01145434e-02 4.91550863e-01 5.49307227e-01 -1.46149516e-01
-4.04815972e-01 8.86731684e-01 -3.44918519e-01 2.26656461e+00
-8.13348889e-01 9.12287295e-01 1.42896116e-01 -9.01808321e-01
7.17772305e-01 7.00063586e-01 -2.07850337e-01 -4.98734564e-01
3.62578928e-01 5.23298740e-01 4.25447345e-01 2.08143443e-02
8.79022121e-01 -6.01232290e-01 -7.00675309e-01 6.06088936e-01
-1.30543306e-01 -5.22108674e-01 1.75035834e-01 4.64301914e-01
5.89994550e-01 5.76059408e-02 9.80904877e-01 -1.19539583e+00
9.81553257e-01 3.74972373e-01 -8.66651386e-02 3.39031279e-01
-1.36855230e-01 -9.23971366e-03 1.02058373e-01 -4.51100886e-01
-7.47887492e-01 -1.30324054e+00 -4.65602577e-01 1.03478837e+00
-2.64899097e-02 -1.07024491e+00 -4.64688212e-01 -5.03550172e-01
-9.47404206e-02 1.65752077e+00 -6.16802573e-01 6.09314442e-02
-1.18203676e+00 7.57756382e-02 1.21144164e+00 5.34690440e-01
2.42958739e-01 -1.39023960e+00 -8.65226030e-01 2.39949033e-01
-3.53297621e-01 -1.15143335e+00 -3.62143189e-01 -2.31582820e-01
-3.22218478e-01 -9.49293077e-01 -1.37163311e-01 -1.18224311e+00
-6.50325865e-02 1.89737193e-02 1.40127170e+00 1.81008160e-01
-3.39093238e-01 3.69095832e-01 -2.82803237e-01 -6.67126477e-01
-7.39807725e-01 1.43916771e-01 4.52235863e-02 -1.07141864e+00
8.29495132e-01 -2.85649657e-01 2.80902982e-01 -6.32102787e-02
-5.49640596e-01 1.83885336e-01 -2.95499593e-01 1.03490210e+00
-1.23316445e-03 -6.28942013e-01 5.42285025e-01 -8.29520941e-01
1.47431850e+00 -3.18459481e-01 -1.08118065e-01 2.17373475e-01
-1.55169899e-02 -3.77166331e-01 6.67667806e-01 -3.35916817e-01
-1.09358788e+00 -3.92093271e-01 -8.42726231e-03 -2.40632370e-01
4.67388183e-02 5.97927153e-01 -6.22006245e-02 2.59170532e-01
8.05925608e-01 -1.18382588e-01 -2.70472556e-01 -4.17385310e-01
8.93298447e-01 8.01363528e-01 8.10489893e-01 -9.92698252e-01
3.58074427e-01 6.73833536e-03 -2.30804771e-01 -1.34897709e+00
-2.14965761e-01 -4.57817614e-01 -9.28838909e-01 -9.98534635e-02
7.87176728e-01 -6.52763963e-01 -1.43330127e-01 -1.42577663e-01
-1.58734727e+00 -2.53612101e-01 -1.41745627e-01 6.39536008e-02
-5.77359557e-01 3.96248013e-01 -4.70205545e-01 -8.08157802e-01
-4.27184850e-01 -5.47513247e-01 9.25810039e-01 -5.41583598e-01
-1.55070698e+00 -1.86019886e+00 3.79794657e-01 1.04600176e-01
2.94426918e-01 1.89018279e-01 1.37652194e+00 -7.44386971e-01
3.81426245e-01 2.34259322e-01 -1.52109846e-01 -9.34260711e-03
1.37418956e-01 -4.04928327e-01 -6.44355297e-01 -1.11739211e-01
-3.21686976e-02 -6.79509580e-01 8.69509280e-01 3.75275239e-02
2.84588039e-01 8.86014700e-02 -5.56024969e-01 1.06416099e-01
1.06656098e+00 2.28007168e-01 1.19353384e-01 4.91703838e-01
4.60595071e-01 6.92301929e-01 5.64320683e-01 -6.42072856e-02
4.76276934e-01 6.90632463e-01 -1.91201746e-01 2.52006292e-01
-3.06159973e-01 -2.77248174e-01 3.85444999e-01 9.77221012e-01
3.01571161e-01 -1.19585112e-01 -1.59900737e+00 8.24137568e-01
-1.84338748e+00 -5.85055888e-01 -3.06226350e-02 1.60960472e+00
7.25782275e-01 1.34336710e-01 2.32220590e-01 -2.05133334e-01
6.12329900e-01 3.83413702e-01 1.42349049e-01 -1.12591624e+00
-5.45524538e-01 5.07241368e-01 -1.56182498e-01 1.16800332e+00
-1.10386097e+00 1.93247437e+00 7.42613268e+00 8.96541953e-01
-8.08241785e-01 3.18287730e-01 1.19638123e-01 -2.14562401e-01
-5.63007116e-01 -4.19368520e-02 -3.19886684e-01 -7.01740459e-02
1.02364051e+00 -6.31238341e-01 4.05311197e-01 6.10912323e-01
-2.32926205e-01 -6.82034343e-02 -1.37433553e+00 1.05171728e+00
2.93998063e-01 -1.20653343e+00 1.96561143e-01 -1.94041759e-01
-6.74834996e-02 -1.68720931e-01 -2.37841874e-01 2.75853634e-01
3.31519805e-02 -1.22983897e+00 8.33526194e-01 1.37004539e-01
8.37852776e-01 -7.65544891e-01 5.73238134e-01 1.41518146e-01
-1.11065483e+00 1.95949033e-01 -2.99483091e-01 -4.88796890e-01
5.30506015e-01 -2.11151004e-01 -9.25350606e-01 2.74721503e-01
6.85832351e-02 7.39711046e-01 -5.44790566e-01 -1.06329523e-01
-4.94042069e-01 5.23750663e-01 -3.19264859e-01 -6.50096536e-01
1.58354342e-01 3.43030095e-02 1.42178690e+00 1.75300765e+00
-2.01464072e-01 2.74792135e-01 4.67249274e-01 9.35610950e-01
1.53982088e-01 5.90177774e-01 -7.74506986e-01 -7.34811902e-01
8.98393512e-01 1.13807893e+00 -8.62035573e-01 -5.32936573e-01
5.67998812e-02 1.05719125e+00 4.69498008e-01 -3.61046474e-03
-3.82375002e-01 -1.25632226e-01 6.92706048e-01 -2.44430035e-01
-3.31349671e-01 -5.03996015e-01 -6.07133448e-01 -1.03075624e+00
-3.84881735e-01 -4.36458886e-01 4.18882698e-01 -8.39853048e-01
-1.78079212e+00 7.47026980e-01 5.87365270e-01 -5.55698752e-01
-7.07036853e-01 -9.37626243e-01 -6.80842340e-01 1.11003590e+00
-6.12593472e-01 -1.45757794e+00 4.31809038e-01 5.16728520e-01
8.18145275e-01 -5.73274910e-01 1.57062209e+00 -2.57176429e-01
-8.73909369e-02 6.42971337e-01 -6.03093684e-01 7.43583217e-03
7.92750716e-01 -1.40905249e+00 8.80947471e-01 6.20599806e-01
4.40966636e-01 1.32480884e+00 1.01541150e+00 -6.31541729e-01
-9.47892010e-01 -2.33327121e-01 1.30153501e+00 -4.78760272e-01
1.58395648e+00 -5.42070627e-01 -6.41071677e-01 6.99239075e-01
9.05489802e-01 -7.89221346e-01 8.45730245e-01 5.29229522e-01
-4.87860769e-01 9.89981890e-01 -1.14184105e+00 9.18907046e-01
1.72934818e+00 -1.04239357e+00 -1.64187229e+00 5.93559444e-01
8.22235703e-01 -3.69878232e-01 -7.08322942e-01 1.14908323e-01
4.67356473e-01 -3.15553367e-01 1.13636661e+00 -8.84971499e-01
5.57897925e-01 6.95390999e-02 -6.56771541e-01 -1.58418739e+00
-3.09441596e-01 -6.19519770e-01 2.46448457e-01 1.25206017e+00
2.02050582e-01 -7.34453380e-01 5.34408689e-02 1.73123777e-01
-4.02925313e-02 -1.44789353e-01 -1.58476186e+00 -6.52305961e-01
5.47354996e-01 -6.73899174e-01 4.70801890e-02 1.21406853e+00
9.73226845e-01 1.11300886e+00 2.54194319e-01 -8.11366796e-01
4.64535832e-01 6.39012232e-02 4.57443088e-01 -1.04009724e+00
-3.54088873e-01 -9.37018335e-01 -6.29306197e-01 -5.99214256e-01
1.06068838e+00 -1.30223548e+00 -1.55462503e-01 -1.26000547e+00
-1.86379537e-01 1.19353332e-01 -2.14715466e-01 3.53076130e-01
2.21977353e-01 1.84249178e-01 4.74438965e-01 1.35760024e-01
-1.46726575e-02 3.39344889e-01 4.96518910e-01 -1.69931322e-01
-1.29798517e-01 -9.88332629e-01 -7.31361985e-01 8.31157982e-01
8.04968178e-01 -4.89819460e-02 -7.09721267e-01 -4.88854736e-01
1.61063150e-02 -2.86183059e-01 4.00205940e-01 -2.59505779e-01
-1.19677924e-01 -1.04307383e-01 1.27156317e-01 -6.21169806e-01
6.76620066e-01 -2.73145974e-01 -1.80453494e-01 4.92994100e-01
-3.70555520e-01 5.43894231e-01 7.96878159e-01 -1.52492493e-01
-1.96347296e-01 -1.23754859e-01 3.73032302e-01 -4.04352814e-01
-1.15471721e+00 -8.74084771e-01 -8.34571004e-01 4.24171150e-01
8.16112876e-01 -3.54250997e-01 -4.82327789e-01 -1.17661439e-01
-6.07609451e-01 3.89444381e-01 5.81709802e-01 1.12222588e+00
8.12101781e-01 -1.46334612e+00 -6.96996987e-01 -9.81245488e-02
2.86635101e-01 -9.54256892e-01 -3.14150184e-01 4.96587276e-01
-3.31534117e-01 1.03653932e+00 -4.19950247e-01 -1.21285379e-01
-1.54433167e+00 5.26484489e-01 2.06515476e-01 1.29977375e-01
-7.69818783e-01 1.22262645e+00 1.91749543e-01 -4.66421068e-01
-3.67511094e-01 1.69326603e-01 -4.09767270e-01 6.37661815e-02
4.01134104e-01 2.58882582e-01 -6.08883202e-01 -1.35389245e+00
-9.25698757e-01 1.53399304e-01 -2.35122591e-01 -1.14480889e+00
7.79522955e-01 -2.60799259e-01 -7.83197284e-01 1.20993543e+00
1.21450162e+00 3.96715730e-01 3.21857512e-01 7.47847408e-02
6.88235939e-01 -2.46417746e-01 1.06579438e-01 -1.06261981e+00
2.28818342e-01 6.24928176e-01 -7.16899149e-03 3.28726709e-01
6.97649419e-01 2.86269765e-02 9.03757930e-01 3.19747537e-01
2.71940261e-01 -1.42152321e+00 -4.89345402e-01 1.05866575e+00
1.70330620e+00 -7.84776032e-01 -1.02023423e-01 -6.57003522e-01
-6.04579568e-01 1.45788658e+00 5.12030840e-01 -3.91184926e-01
5.46865404e-01 -2.91929394e-02 3.45321864e-01 -4.85362142e-01
-7.59842992e-01 1.95580035e-01 4.98842418e-01 5.08258522e-01
1.23786390e+00 6.79192543e-01 -1.37794089e+00 7.06490636e-01
-1.09919107e+00 -4.62014914e-01 3.43156338e-01 7.65001714e-01
-1.74921483e-01 -1.26782787e+00 -2.25527048e-01 -2.75092795e-02
1.36091337e-02 -7.02298701e-01 -1.73738241e+00 1.37486184e+00
-4.70514521e-02 1.03199959e+00 2.23805949e-01 -4.34106946e-01
1.44012615e-01 4.95983541e-01 9.31856096e-01 -7.53528476e-01
-6.56380773e-01 -6.95292652e-02 1.14010489e+00 -4.53956217e-01
-4.74600494e-01 -9.10261571e-01 -1.56848550e+00 -1.66130811e-01
2.86700502e-02 3.97642821e-01 3.80217671e-01 1.20138800e+00
-2.25521058e-01 2.49639183e-01 -1.90677971e-01 -8.86081219e-01
-7.34490529e-02 -1.54892576e+00 -4.46728915e-01 1.30613819e-01
-1.78194016e-01 -9.10093069e-01 -4.94246960e-01 -3.24773252e-01] | [10.749269485473633, 9.20854377746582] |
5a7b25b9-a4ff-4d77-9405-9f2df61fd6a2 | transfer-learning-with-semi-supervised | 2306.16760 | null | https://arxiv.org/abs/2306.16760v1 | https://arxiv.org/pdf/2306.16760v1.pdf | Transfer Learning with Semi-Supervised Dataset Annotation for Birdcall Classification | We present working notes on transfer learning with semi-supervised dataset annotation for the BirdCLEF 2023 competition, focused on identifying African bird species in recorded soundscapes. Our approach utilizes existing off-the-shelf models, BirdNET and MixIT, to address representation and labeling challenges in the competition. We explore the embedding space learned by BirdNET and propose a process to derive an annotated dataset for supervised learning. Our experiments involve various models and feature engineering approaches to maximize performance on the competition leaderboard. The results demonstrate the effectiveness of our approach in classifying bird species and highlight the potential of transfer learning and semi-supervised dataset annotation in similar tasks. | ['Chris Hayduk', 'Murilo Gustineli', 'Nathan Zhong', 'Anthony Miyaguchi'] | 2023-06-29 | null | null | null | null | ['feature-engineering', 'transfer-learning'] | ['methodology', 'miscellaneous'] | [-7.44344369e-02 -2.16968626e-01 -2.40624949e-01 -7.01584876e-01
-6.98830605e-01 -1.05023122e+00 5.18056214e-01 8.27273075e-03
-9.50373352e-01 7.54691839e-01 5.46234488e-01 2.13392209e-02
-6.75152941e-03 -4.03802454e-01 -6.59855962e-01 2.90286075e-02
-5.13764024e-01 4.22006220e-01 4.92299460e-02 -2.74600118e-01
-5.22664301e-02 4.71803173e-02 -1.55321038e+00 3.22678924e-01
4.52027708e-01 6.51531339e-01 -1.15936361e-02 9.59953666e-01
3.77660021e-02 1.20038068e+00 -8.55671406e-01 -5.24679422e-01
6.18980885e-01 -3.70852917e-01 -1.10181320e+00 -5.97411633e-01
1.23326385e+00 -3.09448719e-01 -2.30445012e-01 9.18383598e-01
6.31473243e-01 2.47008532e-01 6.02425098e-01 -1.48331261e+00
-7.15077817e-01 1.01283681e+00 -1.91040322e-01 7.19658256e-01
2.11115137e-01 8.79283249e-02 1.47301507e+00 -6.90122306e-01
6.12947166e-01 7.17921495e-01 1.41641653e+00 8.51738095e-01
-1.29080856e+00 -1.25553703e+00 -1.39256909e-01 1.53657094e-01
-1.45601928e+00 -4.89701539e-01 5.62928855e-01 -8.59933913e-01
1.13339829e+00 3.14795971e-01 7.90185273e-01 7.44488120e-01
-4.15528774e-01 7.66996086e-01 1.07758534e+00 -2.75281042e-01
-6.23303838e-02 4.81403917e-01 2.98170179e-01 7.75009632e-01
1.81269795e-02 6.55417979e-01 -9.68764365e-01 -3.09951842e-01
4.43237215e-01 -3.29490721e-01 -1.05690427e-01 -3.37830871e-01
-1.28116405e+00 9.41173494e-01 8.61723959e-01 1.09910548e-01
-1.21155962e-01 3.98472309e-01 8.64758134e-01 3.45893443e-01
8.02290678e-01 1.00160289e+00 -8.41396868e-01 -7.60660172e-02
-1.32762229e+00 2.40628093e-01 1.08704090e+00 1.05344808e+00
7.75034726e-01 6.51974380e-02 7.65818283e-02 1.04702187e+00
2.74155974e-01 2.46721327e-01 3.49752009e-01 -8.30285490e-01
1.23328209e-01 2.64158726e-01 4.03157584e-02 -4.87630099e-01
-4.67771947e-01 -5.19360423e-01 -1.49830103e-01 -8.15435424e-02
4.82685640e-02 -4.64536190e-01 -6.14836514e-01 1.81305981e+00
2.53845513e-01 5.93455791e-01 -1.62071809e-01 8.28762650e-01
1.37379992e+00 5.26333690e-01 7.54062891e-01 5.56870103e-01
1.13130617e+00 -1.41676748e+00 -1.74411163e-01 -1.18772045e-01
6.71082556e-01 -4.36163396e-01 8.62360060e-01 1.71547398e-01
-4.76694554e-01 -8.06091011e-01 -1.02266693e+00 -8.75033159e-03
-6.95034504e-01 3.83285820e-01 7.93189585e-01 6.90506577e-01
-1.06906128e+00 3.63042891e-01 -5.56474388e-01 -4.99916315e-01
5.49172461e-01 4.40605849e-01 -5.79153597e-01 4.66931432e-01
-1.15907478e+00 8.70746732e-01 5.45916677e-01 -1.72981411e-01
-1.45626521e+00 -1.30160940e+00 -8.02007675e-01 -8.55943933e-02
-5.78903221e-02 -3.48077059e-01 1.55121875e+00 -1.09067905e+00
-9.64743614e-01 1.27988350e+00 7.52156258e-01 -9.23040748e-01
2.87056342e-03 -3.44267607e-01 -6.31064653e-01 8.65553245e-02
3.62010568e-01 1.39088690e+00 4.62251067e-01 -7.90777087e-01
-1.06843150e+00 1.52409986e-01 2.07713589e-01 2.46520564e-01
-6.08217895e-01 3.83959770e-01 3.30410600e-01 -7.01695740e-01
-9.81357098e-01 -9.01605904e-01 -2.43280292e-01 4.62496132e-02
5.50743863e-02 -3.29964042e-01 3.27595234e-01 -5.98262668e-01
1.11181951e+00 -2.22047186e+00 -2.53209919e-01 -1.26472712e-01
2.47289017e-01 3.12550455e-01 -4.87034202e-01 4.68291879e-01
2.08488777e-02 1.63699716e-01 -1.15302995e-01 -6.46679029e-02
3.79411094e-02 1.57614097e-01 -4.29621667e-01 4.56906050e-01
8.25923979e-02 6.65090621e-01 -1.38338518e+00 -6.28414035e-01
2.84161001e-01 2.92026281e-01 -9.21691000e-01 5.50741851e-01
-1.07596457e-01 2.17697978e-01 -3.25535499e-02 4.00112033e-01
2.50754833e-01 3.59237790e-01 -1.64178148e-01 -2.21948236e-01
-3.69476676e-01 5.61833858e-01 -7.76637018e-01 2.05765247e+00
-5.74266434e-01 8.42000723e-01 2.81622648e-01 -7.21830308e-01
9.75477278e-01 9.27819833e-02 4.40820396e-01 -1.03192680e-01
-1.88255310e-01 1.82826743e-01 2.14923779e-03 -2.03845769e-01
5.78600228e-01 -2.13628396e-01 -3.88123602e-01 5.68017423e-01
1.09219825e+00 -2.74154365e-01 3.41437489e-01 1.81342125e-01
9.45392787e-01 5.58016896e-01 3.47678035e-01 -7.68359900e-01
3.09178948e-01 5.20762980e-01 3.23724598e-01 7.52306223e-01
-6.13345981e-01 1.47915989e-01 -2.47372225e-01 -9.06281829e-01
-7.42974341e-01 -8.00844491e-01 -2.48700917e-01 2.12670255e+00
-2.72253782e-01 -7.71501124e-01 -6.81879699e-01 -1.19115055e+00
-3.94641096e-03 6.45525634e-01 -9.72612083e-01 -1.81909651e-01
-2.02653095e-01 -4.89368260e-01 1.25406694e+00 7.19752729e-01
3.14555377e-01 -1.40849352e+00 -1.02227557e+00 1.25415921e-01
1.06491707e-01 -5.59510112e-01 -8.41365933e-01 6.97635949e-01
-2.06134155e-01 -1.16941690e+00 -2.96278566e-01 -1.23606145e+00
3.45936567e-01 6.58508390e-02 1.43413913e+00 -9.72106829e-02
-6.45729005e-01 7.99633384e-01 -4.03672665e-01 -6.05354905e-01
-2.81919837e-01 3.26779574e-01 2.30721578e-01 -4.47746903e-01
9.18479681e-01 -3.26513797e-01 -5.27775764e-01 2.86788881e-01
-6.01271212e-01 -2.05545649e-01 2.40980580e-01 9.33730543e-01
3.71903270e-01 -3.16847414e-01 7.65165329e-01 -9.98873711e-01
5.84597588e-01 -5.78783989e-01 -5.80553591e-01 2.16244087e-01
-2.62534231e-01 -1.28040507e-01 6.38892174e-01 -3.73239279e-01
-8.26618373e-01 5.65287590e-01 -2.43042052e-01 -2.40346670e-01
-4.14742917e-01 3.63646477e-01 2.32718989e-01 -4.35270429e-01
1.06942892e+00 -7.83087239e-02 -3.24707031e-01 -7.29803979e-01
9.10494268e-01 9.00904715e-01 8.60325575e-01 -5.80256402e-01
5.90278804e-01 3.73410173e-02 -7.61809707e-01 -7.55488515e-01
-1.20525086e+00 -8.92618775e-01 -9.73382890e-01 -1.36970103e-01
8.75060081e-01 -1.45368016e+00 -3.05552870e-01 -1.69704244e-01
-7.89800763e-01 -4.98681277e-01 -9.58482623e-01 7.44437099e-01
-4.86839533e-01 -2.37483084e-01 -3.83397192e-01 -5.17479897e-01
-7.46408582e-01 -4.75933105e-01 9.20905709e-01 8.59404355e-02
-5.02433002e-01 -1.06593430e+00 1.08018911e+00 1.79692522e-01
4.61960435e-01 -1.37477934e-01 2.15782523e-01 -1.52846551e+00
8.34157541e-02 -3.87770683e-02 -4.42740209e-02 5.97102165e-01
-1.13669515e-01 -1.12995893e-01 -1.01912129e+00 -2.20531866e-01
-6.84621513e-01 -9.19995308e-01 9.55131352e-01 7.50577003e-02
8.84019434e-01 -3.18301290e-01 -1.47838071e-01 8.90337527e-01
1.22282243e+00 -3.66677403e-01 -3.13918591e-01 2.73335934e-01
5.97817600e-01 7.61088908e-01 5.72537601e-01 2.93248028e-01
4.12522763e-01 3.68645877e-01 1.25512958e-01 -1.16951920e-01
-4.88074571e-01 -9.38592792e-01 2.84450829e-01 8.26837242e-01
3.82139444e-01 1.50287777e-01 -1.09396458e+00 9.81558502e-01
-1.48142517e+00 -1.08234918e+00 1.77960753e-01 1.95931149e+00
1.16642869e+00 -3.36711854e-01 5.86624742e-01 -5.11902213e-01
5.68011880e-01 -4.81376052e-03 -2.57576406e-01 -4.90700811e-01
1.54874086e-01 6.11488700e-01 7.21800148e-01 3.57236326e-01
-1.61942935e+00 1.26363361e+00 8.35972214e+00 7.12945998e-01
-7.37840533e-01 4.37505662e-01 3.22035439e-02 -3.02022636e-01
8.95664319e-02 -1.20685436e-01 -9.23904002e-01 2.04507798e-01
1.23125148e+00 -2.30864897e-01 5.83399415e-01 1.06266844e+00
-5.55548668e-01 4.03922617e-01 -1.28927505e+00 8.23383689e-01
4.55489129e-01 -1.13332939e+00 -1.26722336e-01 -2.40594387e-01
7.71506846e-01 8.30475807e-01 -2.99054414e-01 8.95494461e-01
9.20572460e-01 -9.54687715e-01 7.96649337e-01 2.05781937e-01
1.01512039e+00 -6.35365486e-01 5.88048995e-01 2.45921746e-01
-1.23288405e+00 -7.79867321e-02 -5.20693779e-01 -2.36297429e-01
-1.69694483e-01 -1.87646806e-01 -1.00490665e+00 -9.54218488e-03
1.08399665e+00 1.01373017e+00 -1.00705504e+00 1.56445396e+00
-4.42599326e-01 1.25481141e+00 -5.84244013e-01 -2.34928071e-01
3.30713958e-01 1.55287266e-01 3.88969928e-01 1.77644849e+00
-1.73215061e-01 -2.58853853e-01 5.57121277e-01 7.81511247e-01
-4.18369740e-01 3.25745106e-01 -7.07497001e-01 -2.08210573e-01
5.90833247e-01 1.58584118e+00 -4.39887971e-01 -3.47741157e-01
-4.34724480e-01 6.21516407e-01 5.17404974e-01 -1.10810608e-01
-8.37660730e-01 -4.27466273e-01 5.53663850e-01 -1.05239518e-01
3.85699540e-01 1.63817570e-01 1.57143563e-01 -1.03598905e+00
-7.11887360e-01 -8.21753860e-01 7.94605494e-01 -4.38177705e-01
-1.76166701e+00 7.82532930e-01 2.36799031e-01 -1.29005551e+00
-3.77917625e-02 -5.50922215e-01 -3.97113234e-01 4.75271553e-01
-1.32949090e+00 -1.70414460e+00 -6.47528544e-02 4.99852031e-01
4.61005419e-01 -5.07721543e-01 1.14168572e+00 4.86118376e-01
-5.77939972e-02 6.71817183e-01 3.27946525e-03 3.81238192e-01
1.00235236e+00 -1.71027648e+00 5.72293282e-01 5.47922969e-01
1.14605248e+00 5.87360322e-01 3.91973346e-01 -4.86404479e-01
-9.22957242e-01 -1.27822387e+00 6.09049559e-01 -9.21886265e-01
9.80881453e-01 -7.09435463e-01 -3.27101111e-01 8.16912293e-01
4.97107297e-01 2.62362570e-01 1.67832518e+00 4.39924449e-01
-7.61843622e-01 -9.02947634e-02 -1.05673587e+00 -1.22348599e-01
1.04699993e+00 -9.74468946e-01 -1.05593896e+00 4.04009253e-01
5.79854012e-01 -9.93403792e-02 -8.48738790e-01 1.76577196e-01
7.37357974e-01 -2.58770347e-01 8.19991469e-01 -1.14316106e+00
2.44148895e-01 -4.77316618e-01 -7.41674155e-02 -1.75412869e+00
-4.76703405e-01 -5.77651858e-01 2.22838253e-01 1.40408397e+00
4.53277498e-01 -3.68727669e-02 6.57442570e-01 1.49879098e-01
-2.03760549e-01 8.28637034e-02 -8.51868927e-01 -6.53971255e-01
1.58207506e-01 -2.88044333e-01 3.99787664e-01 1.05335188e+00
3.64251761e-03 7.45205104e-01 -6.30756259e-01 -1.39583915e-01
5.53723872e-01 5.59559502e-02 1.04618347e+00 -1.44669187e+00
-3.40528071e-01 -1.82015762e-01 -3.94705683e-01 -9.20910776e-01
4.86090332e-01 -1.42384303e+00 4.61261630e-01 -1.17637658e+00
2.30777860e-01 -5.41996479e-01 -6.83230519e-01 9.59595680e-01
3.67300138e-02 8.50382268e-01 1.52675167e-01 1.85862303e-01
-1.10710335e+00 2.36940145e-01 3.37042570e-01 -2.01503605e-01
-6.86635226e-02 -2.50346512e-01 -8.21373641e-01 6.69675291e-01
5.65751493e-01 -8.75401676e-01 -2.25453690e-01 -6.84657454e-01
7.17996359e-02 -8.00904572e-01 4.87485707e-01 -1.12602544e+00
3.38410676e-01 3.07923585e-01 2.52257794e-01 -4.53076273e-01
5.56591316e-04 -9.85073924e-01 1.24617238e-02 3.30063969e-01
-1.01269805e+00 -2.61056095e-01 5.23236811e-01 4.67392713e-01
-2.91411281e-01 -4.27913576e-01 5.06447732e-01 -1.17360495e-01
-1.19042599e+00 2.88978696e-01 -2.09182411e-01 4.45594668e-01
8.98748219e-01 4.43355918e-01 -2.89205402e-01 -9.23886448e-02
-7.08789468e-01 5.00521779e-01 1.83359101e-01 6.98307872e-01
2.08036125e-01 -1.17712235e+00 -1.16554832e+00 2.77991086e-01
5.20597577e-01 -6.50392890e-01 -1.07557900e-01 3.59516382e-01
-8.00515354e-01 3.68687421e-01 -5.72307885e-01 -2.97566921e-01
-1.24738705e+00 5.01320243e-01 4.31312591e-01 -6.11671880e-02
-1.28321528e-01 1.41238976e+00 -1.51748657e-02 -1.37061727e+00
3.07260960e-01 8.88178125e-02 -2.60981709e-01 7.99254104e-02
6.92841411e-01 3.26184571e-01 -1.29294321e-01 -8.10977101e-01
-5.25064170e-01 4.89508025e-02 -7.19594583e-02 -5.35561405e-02
1.63971364e+00 3.92869592e-01 7.78981969e-02 4.08440143e-01
1.21223795e+00 2.63184160e-01 -1.06458139e+00 -1.44154146e-01
1.42468989e-01 -2.10099339e-01 2.72137195e-01 -1.32491684e+00
-6.63790166e-01 8.91903520e-01 7.95944571e-01 1.20900925e-02
7.14975476e-01 -5.57474187e-03 6.38294637e-01 5.19229889e-01
4.50254023e-01 -1.03849292e+00 -4.66052771e-01 3.28775913e-01
6.26458824e-01 -1.02737844e+00 4.79098745e-02 1.31050199e-01
-6.94380701e-01 9.52262640e-01 9.89511907e-01 -5.10061443e-01
8.56729209e-01 4.96692240e-01 2.53419101e-01 -3.42135012e-01
-7.53105044e-01 -4.96866047e-01 5.37514687e-01 8.04557085e-01
7.98906088e-01 2.77364284e-01 -1.82492629e-01 7.53642440e-01
-6.21743560e-01 6.74068853e-02 4.62443978e-02 7.91218400e-01
-2.86409527e-01 -9.63532090e-01 1.25767276e-01 5.92169225e-01
-4.83681709e-01 -5.17471313e-01 -1.04768860e+00 7.08098412e-01
6.89682305e-01 7.68600941e-01 -3.95559110e-02 -5.16492963e-01
3.61428052e-01 1.78947486e-02 3.88046116e-01 -1.18211246e+00
-1.44847929e+00 -2.25390628e-01 4.54311639e-01 -3.65685135e-01
-8.80465567e-01 -4.09861565e-01 -7.53118813e-01 2.95892328e-01
-6.65712595e-01 7.52835035e-01 6.18826687e-01 2.56422400e-01
2.11227447e-01 9.04656723e-02 4.70340252e-01 -6.79276109e-01
-5.85839450e-01 -1.10356879e+00 -5.76762617e-01 3.78881156e-01
2.93440759e-01 -4.86154616e-01 -3.55906516e-01 2.65056252e-01] | [15.096685409545898, 5.18523645401001] |
ab531bb3-9c9c-4b08-8769-ebe777690b37 | mask-dngan-multi-stage-raw-video-denoising | 2103.02861 | null | https://arxiv.org/abs/2103.02861v3 | https://arxiv.org/pdf/2103.02861v3.pdf | Multi-Stage Raw Video Denoising with Adversarial Loss and Gradient Mask | In this paper, we propose a learning-based approach for denoising raw videos captured under low lighting conditions. We propose to do this by first explicitly aligning the neighboring frames to the current frame using a convolutional neural network (CNN). We then fuse the registered frames using another CNN to obtain the final denoised frame. To avoid directly aligning the temporally distant frames, we perform the two processes of alignment and fusion in multiple stages. Specifically, at each stage, we perform the denoising process on three consecutive input frames to generate the intermediate denoised frames which are then passed as the input to the next stage. By performing the process in multiple stages, we can effectively utilize the information of neighboring frames without directly aligning the temporally distant frames. We train our multi-stage system using an adversarial loss with a conditional discriminator. Specifically, we condition the discriminator on a soft gradient mask to prevent introducing high-frequency artifacts in smooth regions. We show that our system is able to produce temporally coherent videos with realistic details. Furthermore, we demonstrate through extensive experiments that our approach outperforms state-of-the-art image and video denoising methods both numerically and visually. | ['Nima Khademi Kalantari', 'Libing Zeng', 'Avinash Paliwal'] | 2021-03-04 | null | null | null | null | ['video-denoising'] | ['computer-vision'] | [ 4.39854980e-01 -2.05371559e-01 3.97053003e-01 -2.06539661e-01
-7.77629733e-01 -4.88323241e-01 4.13140297e-01 -1.91421211e-01
-6.26805067e-01 6.67149842e-01 1.08755231e-01 6.29541576e-02
2.58311599e-01 -6.65487647e-01 -9.36683178e-01 -8.90568316e-01
-1.77539643e-02 -3.49074304e-01 2.98878849e-01 -5.29957004e-02
2.53469013e-02 5.81610084e-01 -1.29610562e+00 2.34687448e-01
6.65385902e-01 1.01675987e+00 1.19775601e-01 7.39927649e-01
3.47510844e-01 8.42044473e-01 -5.74625850e-01 -2.98769593e-01
5.83574653e-01 -7.44234622e-01 -5.47402322e-01 3.51579696e-01
5.53577363e-01 -7.51503885e-01 -4.39778745e-01 1.15324068e+00
2.78927892e-01 4.15839821e-01 1.69302568e-01 -7.47828186e-01
-2.37194136e-01 6.84416592e-02 -7.55458593e-01 1.31753698e-01
2.79338032e-01 2.99400419e-01 5.43163717e-01 -8.24556589e-01
6.53642833e-01 1.11370063e+00 6.81221545e-01 5.11052608e-01
-1.34636140e+00 -4.86251116e-01 1.28621906e-01 1.29529461e-01
-1.13064480e+00 -6.63415849e-01 1.01532769e+00 -2.78958589e-01
3.75367135e-01 1.31642371e-01 6.11365914e-01 8.69046628e-01
2.86309510e-01 2.99746186e-01 8.23254824e-01 -3.72590601e-01
2.16807231e-01 -6.52854860e-01 -3.51917952e-01 6.19472265e-01
-2.04418108e-01 2.35443249e-01 -3.77003193e-01 2.20074579e-02
9.93595064e-01 1.19847529e-01 -5.75992584e-01 -3.15948054e-02
-1.22419751e+00 5.83845079e-01 5.35398304e-01 3.09983045e-01
-8.09114695e-01 4.70894992e-01 2.19500631e-01 2.41902709e-01
6.56491995e-01 5.35623394e-02 4.13521789e-02 2.45486468e-01
-1.35588753e+00 7.06092194e-02 3.88978660e-01 3.49436462e-01
8.09842765e-01 6.94958195e-02 -1.50713861e-01 6.21614158e-01
2.36114830e-01 1.31452918e-01 1.79259092e-01 -1.67774785e+00
3.72620344e-01 -1.49537846e-01 2.65635908e-01 -9.53539491e-01
5.55685796e-02 -1.14866003e-01 -1.18381405e+00 6.29615426e-01
5.04119396e-01 -3.14490318e-01 -9.47022200e-01 1.73893857e+00
2.98863202e-01 6.71799481e-01 1.12485185e-01 1.15291798e+00
3.26798648e-01 8.70092452e-01 6.36653453e-02 -5.23374438e-01
1.03504646e+00 -9.42037225e-01 -9.30907726e-01 -1.09112471e-01
1.52188018e-02 -1.05579925e+00 5.34615755e-01 5.38204551e-01
-1.57981181e+00 -8.11284363e-01 -1.06150818e+00 -2.06792399e-01
1.77899882e-01 1.92027297e-02 4.81131859e-02 8.79527777e-02
-1.11728024e+00 9.38198328e-01 -1.18925571e+00 8.47282484e-02
3.31594527e-01 1.82032153e-01 -5.19262195e-01 -1.95936486e-01
-1.13789999e+00 7.25338042e-01 1.06053233e-01 5.23411214e-01
-1.11750305e+00 -5.11057734e-01 -1.08521557e+00 1.41719636e-02
1.99252844e-01 -8.87520492e-01 9.67646062e-01 -1.37434614e+00
-1.63113940e+00 5.64970970e-01 -4.33864325e-01 -4.86076444e-01
7.26517856e-01 -3.67358148e-01 -2.35788465e-01 5.07452369e-01
1.46479916e-03 6.28181517e-01 1.25476074e+00 -1.44120836e+00
-4.76517498e-01 7.84396231e-02 1.12537399e-01 6.41186684e-02
-2.42840387e-02 -1.07387369e-02 -6.58306241e-01 -1.02519941e+00
2.39844993e-01 -7.96739578e-01 -2.55977690e-01 3.91640007e-01
-2.28883818e-01 4.67521667e-01 9.48710978e-01 -1.01370394e+00
1.03192246e+00 -2.47386336e+00 4.69382823e-01 1.30914271e-01
2.61753738e-01 3.24626446e-01 -2.41349876e-01 1.07873604e-01
-2.17071593e-01 -1.77040324e-01 -4.76356536e-01 -8.37631345e-01
-5.31549156e-01 3.35519940e-01 -3.19919914e-01 5.50208688e-01
3.42660248e-01 6.37776673e-01 -9.66857135e-01 -2.40731791e-01
6.13231122e-01 9.35057402e-01 -6.62282109e-01 3.84798884e-01
1.71123464e-02 8.75093997e-01 -1.36569470e-01 2.17751786e-01
8.37237954e-01 1.47392407e-01 9.41352621e-02 -6.06676042e-01
-1.56493917e-01 1.67248566e-02 -1.19992435e+00 1.65823138e+00
-5.78899086e-01 8.41040254e-01 5.72106183e-01 -1.13583720e+00
6.30483568e-01 4.13101524e-01 5.90390623e-01 -3.95352095e-01
1.38329878e-01 1.54238358e-01 -2.59996414e-01 -5.02933204e-01
3.19693059e-01 -3.48319054e-01 1.41061768e-01 9.31035802e-02
4.87381220e-02 -1.40468165e-01 1.25779688e-01 7.57496133e-02
1.09301555e+00 2.65448451e-01 1.61547288e-02 1.68905228e-01
7.88565636e-01 -4.23052847e-01 6.37891233e-01 3.86676162e-01
-1.95978060e-01 1.12330198e+00 3.47883433e-01 -5.76807261e-01
-1.14332795e+00 -1.20028341e+00 2.92356551e-01 5.19228637e-01
2.75175124e-01 -2.30843991e-01 -9.23757374e-01 -4.71681833e-01
-3.78587633e-01 3.34690899e-01 -5.69620490e-01 -2.05051795e-01
-9.55865920e-01 -2.12763295e-01 1.67773873e-01 4.24519926e-01
7.29479909e-01 -9.14628685e-01 -5.80011606e-01 4.00071234e-01
-4.79221910e-01 -1.16095579e+00 -7.07863033e-01 1.08416051e-01
-7.79834270e-01 -9.31526303e-01 -8.47503364e-01 -8.85162830e-01
8.31333756e-01 3.12667429e-01 9.38750446e-01 2.35310674e-01
1.94254122e-03 -6.33254424e-02 -1.92232758e-01 2.80243516e-01
-5.08760154e-01 -5.08138716e-01 -1.57572359e-01 4.19603914e-01
-3.43972594e-01 -8.06620717e-01 -8.97489130e-01 2.03538567e-01
-1.37015593e+00 1.34033471e-01 3.11491698e-01 9.15244102e-01
7.36767650e-01 3.61116439e-01 1.44179374e-01 -3.52058172e-01
3.06330651e-01 -2.01539025e-01 -7.26711571e-01 -7.22785220e-02
2.96141148e-01 -3.14055197e-02 9.03741419e-01 -4.19502139e-01
-1.03249109e+00 3.46707672e-01 -3.61131251e-01 -8.05090308e-01
-1.04768999e-01 2.71769166e-01 -1.66201040e-01 -1.15462609e-01
3.05139542e-01 1.43060625e-01 5.90217188e-02 -3.90482605e-01
3.04149121e-01 2.49064386e-01 1.09761572e+00 -2.82577246e-01
9.63518500e-01 7.64904022e-01 1.34092253e-02 -6.99288428e-01
-8.40704739e-01 -1.70877263e-01 -6.03474498e-01 -4.68362629e-01
1.29350686e+00 -9.97860909e-01 -5.53256035e-01 7.03958988e-01
-1.41838741e+00 -4.13941652e-01 -2.34439641e-01 5.44039547e-01
-5.58596671e-01 6.19165957e-01 -8.66230309e-01 -3.75194281e-01
-1.75994679e-01 -1.32495356e+00 1.05837953e+00 2.24422857e-01
-1.07220344e-01 -9.98394191e-01 4.08978686e-02 1.88952073e-01
3.58750790e-01 5.05562663e-01 4.12617683e-01 1.00360528e-01
-6.90035582e-01 5.45149855e-02 3.99476029e-02 8.04658294e-01
2.89016277e-01 1.09140731e-01 -7.57632971e-01 -3.41300398e-01
3.78514498e-01 -5.18819969e-03 1.16587126e+00 6.20737791e-01
1.10321474e+00 -1.76084682e-01 1.70153733e-02 1.02050352e+00
1.43672585e+00 1.15011223e-01 9.73574996e-01 3.11505526e-01
6.90697610e-01 4.46979672e-01 4.56624568e-01 1.70404658e-01
7.35794082e-02 5.96634626e-01 5.98404050e-01 -3.80088091e-01
-2.81823277e-01 -1.79980565e-02 4.74933177e-01 5.70517302e-01
-2.29496971e-01 -3.35548460e-01 -3.70050579e-01 5.52868068e-01
-1.80077684e+00 -1.07017720e+00 4.37007062e-02 2.12990808e+00
8.01846206e-01 5.65685369e-02 -2.31483817e-01 1.80519208e-01
8.60382617e-01 3.38455081e-01 -2.22596779e-01 -1.90098569e-01
-8.46600011e-02 4.39471573e-01 3.49566877e-01 8.56660247e-01
-1.29111588e+00 7.45778084e-01 5.89250660e+00 5.16797900e-01
-1.37041223e+00 5.10293879e-02 7.47649252e-01 -2.33253449e-01
-4.66688760e-02 3.57085504e-02 -1.23129986e-01 5.81706345e-01
6.40391052e-01 1.94330484e-01 6.70202792e-01 2.27480397e-01
5.85394681e-01 -3.51734310e-01 -9.09045160e-01 7.49144316e-01
-1.21901043e-01 -1.27672660e+00 -9.90490541e-02 -1.37225062e-01
9.83214021e-01 -2.75399923e-01 -1.29780769e-01 -3.32548171e-01
-1.63361570e-03 -8.71325850e-01 9.48785841e-01 7.60159194e-01
5.37613869e-01 -8.50471020e-01 7.09652781e-01 1.51909336e-01
-1.15315318e+00 2.16367647e-01 -2.08846286e-01 2.88725141e-02
6.87226236e-01 8.63977432e-01 1.35399535e-01 6.63474202e-01
7.69496083e-01 7.77405918e-01 3.84263881e-02 9.41307604e-01
-4.45251405e-01 4.08322573e-01 -2.84561217e-01 9.31737781e-01
2.33975485e-01 -5.12598455e-01 5.95030308e-01 9.72541928e-01
5.85367799e-01 3.72390091e-01 8.45191628e-02 6.57951295e-01
-1.28255352e-01 -5.25703669e-01 -4.03965324e-01 4.81797636e-01
1.00510918e-01 1.24150264e+00 -5.30602634e-01 -5.95286131e-01
-4.80737835e-01 1.49271417e+00 -1.62316114e-02 6.16546571e-01
-8.62527847e-01 -3.01966697e-01 7.74753809e-01 2.09961124e-02
6.04345977e-01 -4.30235296e-01 -8.66715610e-02 -1.32047927e+00
2.69604445e-01 -7.77436256e-01 1.48788065e-01 -9.67580140e-01
-1.02051198e+00 7.12219656e-01 -2.83018649e-01 -1.42952061e+00
-3.13555896e-01 -9.95924845e-02 -8.60611260e-01 1.05277157e+00
-1.63327038e+00 -8.87894511e-01 -4.78663027e-01 8.08314204e-01
4.34780896e-01 3.50125521e-01 4.21989739e-01 4.57635105e-01
-4.20045465e-01 1.45994216e-01 1.23286188e-01 2.31299445e-01
9.38167334e-01 -9.77140367e-01 2.53073603e-01 1.35437930e+00
-2.19377711e-01 4.78552639e-01 7.85470009e-01 -5.48861086e-01
-8.92418742e-01 -1.18900228e+00 7.03469634e-01 2.46659756e-01
4.25938785e-01 -1.15415469e-01 -1.21615064e+00 6.66504681e-01
5.23790300e-01 4.20184195e-01 1.57179490e-01 -8.86405587e-01
-1.11205749e-01 -2.60360420e-01 -1.14425802e+00 4.38557833e-01
6.68116033e-01 -4.97922570e-01 -5.80377400e-01 -6.20212959e-05
4.89128917e-01 -4.74524409e-01 -8.36025536e-01 4.13936526e-01
4.48596627e-01 -1.21878147e+00 1.12176406e+00 -6.35918150e-06
7.35896826e-01 -7.52496362e-01 2.02847980e-02 -1.48364282e+00
-1.35378167e-01 -1.03186476e+00 -3.91055085e-02 1.15825260e+00
-6.41171038e-02 -4.81994838e-01 4.87016648e-01 2.76898772e-01
-1.74899265e-01 -4.03066754e-01 -9.75559592e-01 -5.68652928e-01
-1.19285114e-01 -1.99676231e-01 9.16138589e-02 7.55350590e-01
-3.94941151e-01 -1.12106301e-01 -6.45283222e-01 4.07589912e-01
8.55984867e-01 -6.03578500e-02 5.16381621e-01 -6.12553895e-01
-2.88302243e-01 -1.21035993e-01 -2.71953344e-01 -1.20940089e+00
9.73369107e-02 -1.77014530e-01 3.71778935e-01 -1.34122884e+00
-5.54539710e-02 6.15591481e-02 -1.96770787e-01 3.58504057e-01
-3.67220819e-01 7.67639279e-01 2.53519148e-01 2.44957820e-01
-2.88743854e-01 4.94644165e-01 1.29376304e+00 -4.98204213e-03
-2.59705987e-02 -2.08561480e-01 -2.58841097e-01 7.87334204e-01
5.65401256e-01 -3.77549201e-01 -2.77518295e-02 -5.81360757e-01
-2.95057416e-01 2.83714861e-01 8.08662772e-01 -1.16858637e+00
1.76243424e-01 3.11016086e-02 7.42526829e-01 -3.55569065e-01
5.12461662e-01 -9.57296550e-01 4.91968185e-01 5.63735187e-01
-2.10699201e-01 -8.39715824e-02 2.38702193e-01 3.97442281e-01
-5.99084854e-01 -7.96447545e-02 1.24625230e+00 1.11781314e-01
-3.32086384e-01 1.62225932e-01 -5.59532106e-01 -3.09678435e-01
1.07501495e+00 -1.18172884e-01 3.02151460e-02 -6.67586029e-01
-1.01063228e+00 1.62060782e-02 7.94339418e-01 1.01603530e-01
5.90819001e-01 -1.29995859e+00 -5.68118572e-01 4.81583148e-01
-5.74226022e-01 9.48974267e-02 3.85405004e-01 8.54054153e-01
-8.86189103e-01 -2.39300236e-01 -2.86586583e-01 -5.87226510e-01
-1.26831555e+00 6.35473669e-01 5.30478776e-01 -1.44960865e-01
-6.41818821e-01 5.55442393e-01 3.14339310e-01 2.21125543e-01
1.83331817e-01 -3.64406675e-01 5.14639989e-02 -1.12528995e-01
5.63000500e-01 2.45253712e-01 -2.20140424e-02 -7.33086765e-01
-2.17036679e-01 8.20461690e-01 1.87555879e-01 -3.59602392e-01
1.62039745e+00 -2.99169213e-01 -2.20835373e-01 -7.59860426e-02
1.54191744e+00 2.05200225e-01 -1.97118616e+00 -1.04072895e-02
-5.09392381e-01 -6.07462704e-01 2.48951837e-01 -3.66378337e-01
-1.62191832e+00 6.94180310e-01 5.42951822e-01 2.35210627e-01
1.71738732e+00 -2.21740901e-01 1.10769486e+00 -2.07920030e-01
-1.33024737e-01 -7.41836727e-01 1.48352474e-01 3.61863226e-01
7.48355925e-01 -9.26463425e-01 -4.58543152e-02 -4.44614947e-01
-1.86653465e-01 1.28195953e+00 2.19967484e-01 -4.20206934e-01
5.42050362e-01 3.83524090e-01 4.00812358e-01 3.76728177e-02
-5.56290805e-01 -9.97154862e-02 2.74395615e-01 3.62200141e-01
3.70878190e-01 -4.62211967e-01 -2.84621984e-01 -1.96322799e-02
1.63369492e-01 1.46267503e-01 5.27290106e-01 8.76220703e-01
-2.87266642e-01 -1.20622146e+00 -6.73871875e-01 -1.53839365e-01
-7.03104734e-01 -7.39460066e-02 -3.73663418e-02 5.47992349e-01
2.14047372e-01 1.04313290e+00 2.48089954e-01 -1.78119451e-01
3.30051839e-01 -2.36933932e-01 5.26850283e-01 -2.15656504e-01
-5.43338239e-01 5.52003801e-01 -1.42272487e-01 -8.93704057e-01
-7.38019526e-01 -5.83053887e-01 -1.09009230e+00 -3.14322352e-01
4.61990479e-03 -7.35690668e-02 4.11322415e-01 8.76666546e-01
1.46408990e-01 7.45340586e-01 8.57553244e-01 -1.53089523e+00
-1.47345260e-01 -6.06763661e-01 -1.38478965e-01 6.62265003e-01
7.96515048e-01 -2.84975588e-01 -6.25884116e-01 5.87364018e-01] | [11.296283721923828, -2.0782501697540283] |
f8373282-6728-4cc1-93ec-eca56ad23046 | mucs-lt-edi-eacl2021-cohope-hope-speech | null | null | https://aclanthology.org/2021.ltedi-1.27 | https://aclanthology.org/2021.ltedi-1.27.pdf | MUCS@LT-EDI-EACL2021:CoHope-Hope Speech Detection for Equality, Diversity, and Inclusion in Code-Mixed Texts | This paper describes the models submitted by the team MUCS for “Hope Speech Detection for Equality, Diversity, and Inclusion-EACL 2021” shared task that aims at classifying a comment / post in English and code-mixed texts in two language pairs, namely, Tamil-English (Ta-En) and Malayalam-English (Ma-En) into one of the three predefined categories, namely, “Hope_speech”, “Non_hope_speech”, and “other_languages”. Three models namely, CoHope-ML, CoHope-NN, and CoHope-TL based on Ensemble of classifiers, Keras Neural Network (NN) and BiLSTM with Conv1d model respectively are proposed for the shared task. CoHope-ML, CoHope-NN models are trained on a feature set comprised of char sequences extracted from sentences combined with words for Ma-En and Ta-En code-mixed texts and a combination of word and char ngrams along with syntactic word ngrams for English text. CoHope-TL model consists of three major parts: training tokenizer, BERT Language Model (LM) training and then using pre-trained BERT LM as weights in BiLSTM-Conv1d model. Out of three proposed models, CoHope-ML model (best among our models) obtained 1st, 2nd, and 3rd ranks with weighted F1-scores of 0.85, 0.92, and 0.59 for Ma-En, English and Ta-En texts respectively. | ['H L Shashirekha', 'Aparna B K', 'Fazlourrahman Balouchzahi'] | null | null | null | null | eacl-ltedi-2021-4 | ['hope-speech-detection'] | ['natural-language-processing'] | [-3.77611041e-01 2.88643707e-02 -1.44494876e-01 -1.07618198e-01
-1.12193060e+00 -3.09214413e-01 6.96445704e-01 1.87351435e-01
-7.33344972e-01 8.68250668e-01 7.08339512e-01 -8.08840156e-01
3.52696627e-01 -2.76701540e-01 -3.43722254e-01 -3.35579693e-01
3.96808416e-01 2.81624585e-01 8.70038643e-02 -3.21265876e-01
2.85812259e-01 2.79717962e-03 -1.20870447e+00 5.16660213e-01
9.63725269e-01 5.84357679e-01 4.59367692e-01 1.13730443e+00
-3.35735500e-01 1.19181013e+00 -2.16420963e-01 -7.45514333e-01
-2.30018392e-01 -4.69743073e-01 -8.84297371e-01 -5.11295497e-01
2.27436095e-01 -1.34036630e-01 -1.89672425e-01 1.16942537e+00
6.54703498e-01 9.48947966e-02 7.46781647e-01 -8.72546732e-01
-7.54731059e-01 1.28403163e+00 -5.23383141e-01 2.20519334e-01
3.13372493e-01 1.47992847e-02 1.09320343e+00 -1.26831329e+00
5.88688612e-01 1.42338979e+00 8.59629452e-01 7.18660712e-01
-7.45458782e-01 -7.89042652e-01 -4.80067343e-01 2.21209437e-01
-1.28831863e+00 -6.02994502e-01 2.83371419e-01 -6.22847915e-01
1.66262257e+00 3.24139714e-01 3.15009266e-01 1.13397539e+00
5.81946194e-01 9.22920346e-01 1.16617703e+00 -6.45253003e-01
-1.96998626e-01 6.47470593e-01 3.18750501e-01 5.69838047e-01
-3.65011334e-01 -1.49765164e-01 -5.11880457e-01 -3.18261497e-02
-5.54566048e-02 -2.65256256e-01 -5.69591522e-02 6.34797692e-01
-1.22742379e+00 9.24449921e-01 -9.37792212e-02 7.82900214e-01
-4.23191041e-01 -1.70461968e-01 8.60330641e-01 3.66057187e-01
4.37949538e-01 -9.66437012e-02 -5.09697855e-01 -5.97401977e-01
-1.29255915e+00 2.25251503e-02 7.06786454e-01 1.14648044e+00
3.92865658e-01 2.55549401e-01 -2.92546153e-01 1.15989518e+00
5.12472332e-01 5.97915709e-01 1.07774687e+00 -1.67362630e-01
7.84963012e-01 3.87983203e-01 -2.19983235e-01 -6.07103288e-01
-3.75142783e-01 -3.25738877e-01 -9.49366450e-01 -2.27838740e-01
-2.19216365e-02 -4.31642771e-01 -9.45611477e-01 1.61550498e+00
-3.97688061e-01 -1.31594107e-01 6.74957931e-01 1.86316550e-01
1.46543467e+00 1.01770830e+00 2.23823056e-01 -1.74334124e-01
1.20359230e+00 -1.46070313e+00 -6.74508035e-01 -9.43818130e-03
8.46117079e-01 -1.33870554e+00 9.04162347e-01 1.88436732e-02
-1.43465829e+00 -7.07510412e-01 -7.99277365e-01 -2.08642900e-01
-5.28361559e-01 5.99574268e-01 -1.77041575e-01 7.08144367e-01
-1.36361480e+00 1.69407398e-01 -4.02545035e-01 -6.17176414e-01
-1.56727299e-01 2.24195421e-01 -3.97392482e-01 3.21822256e-01
-1.23287106e+00 8.97488534e-01 4.66352224e-01 -7.60021210e-02
-1.07008040e+00 -2.27577403e-01 -8.61760259e-01 4.86327745e-02
-3.28744322e-01 -1.02352895e-01 1.22183096e+00 -1.10379672e+00
-1.38462603e+00 1.08343005e+00 -4.30233896e-01 -5.65330923e-01
5.15453756e-01 -3.43700089e-02 -8.57016206e-01 -3.65372837e-01
1.05436750e-01 4.06912237e-01 4.14897710e-01 -8.83167803e-01
-9.41917539e-01 -4.88498211e-02 -2.36033961e-01 2.98704416e-01
-2.98582613e-01 8.28639328e-01 8.01509395e-02 -3.53218943e-01
-3.26158792e-01 -7.21522033e-01 2.16219202e-01 -8.94419074e-01
-7.60400593e-01 -5.00006795e-01 8.23789060e-01 -1.27925563e+00
1.58582103e+00 -2.23526907e+00 -1.51305171e-02 -1.48050785e-01
1.56940967e-02 5.98463356e-01 -1.51291460e-01 7.69342422e-01
-3.87341082e-02 3.10169786e-01 1.71415016e-01 -7.62663245e-01
-6.35966212e-02 1.52777746e-01 -1.86345380e-04 1.69396803e-01
7.90629387e-02 7.60234296e-01 -8.76103222e-01 -4.99133497e-01
1.24064587e-01 3.60596210e-01 -2.76059002e-01 3.55638206e-01
2.36625090e-01 2.20482022e-01 2.65718214e-02 4.25184786e-01
5.80011129e-01 3.06215376e-01 -3.70749757e-02 6.09513894e-02
-8.12974870e-01 6.04451835e-01 -1.00571203e+00 9.98366416e-01
-7.03622401e-01 7.24064946e-01 1.69405475e-01 -7.22842097e-01
1.01419699e+00 7.57383347e-01 -2.08206281e-01 -4.79852378e-01
4.39897954e-01 7.86234677e-01 4.15970087e-01 -6.57299280e-01
1.03566480e+00 -1.94842994e-01 2.53587719e-02 8.00045431e-02
4.78333592e-01 1.04866877e-01 1.56458303e-01 1.00589827e-01
7.14758515e-01 -3.44478875e-01 4.29516196e-01 -4.45363998e-01
1.19523537e+00 -2.87119180e-01 3.80421132e-01 4.32941556e-01
-3.93185645e-01 4.67256814e-01 4.25115705e-01 1.07820123e-01
-1.33894277e+00 -7.75456429e-01 -1.91648275e-01 1.31782556e+00
-3.80608618e-01 -2.72702515e-01 -6.04098618e-01 -4.54800516e-01
-3.70707065e-01 1.15835726e+00 -3.99450481e-01 -6.14434062e-03
-5.43916881e-01 -3.28002870e-01 1.03399432e+00 3.89558882e-01
6.40072823e-01 -1.36419928e+00 -4.61849533e-02 2.16537893e-01
-2.86412418e-01 -9.89767492e-01 -8.10167074e-01 2.63457924e-01
-2.80674249e-01 -5.66474676e-01 -9.02154744e-01 -1.27461219e+00
1.61601529e-01 -2.72941053e-01 8.52912366e-01 -2.50614695e-02
9.36274901e-02 -6.97125569e-02 -6.18033707e-01 -3.86038572e-01
-1.00017428e+00 1.88504964e-01 -4.27202173e-02 2.77965087e-02
4.84835386e-01 -1.10671684e-01 -6.44986192e-03 -1.22686714e-01
-2.28886187e-01 2.47279853e-01 5.87674260e-01 1.19323790e+00
1.65085331e-01 -5.08730471e-01 5.83193183e-01 -6.80678904e-01
8.42132568e-01 -7.63774037e-01 5.50247217e-03 2.17140928e-01
-4.62518036e-01 -2.30608508e-01 8.76464427e-01 -4.39844519e-01
-1.00016654e+00 -3.79246801e-01 -9.51872885e-01 -1.91146046e-01
-1.59608461e-02 8.22374105e-01 -4.19606678e-02 3.83628637e-01
2.94566125e-01 6.50702894e-01 -3.16876501e-01 -4.12421584e-01
1.09787948e-01 1.62063253e+00 5.92401326e-01 -2.57594615e-01
3.21010441e-01 -2.66236275e-01 -8.42584312e-01 -9.87957776e-01
-5.42403579e-01 -7.75342405e-01 -6.32751524e-01 -1.13733254e-01
1.40422404e+00 -1.04958963e+00 -2.95127481e-01 9.46004689e-01
-1.40811527e+00 -1.46596134e-01 4.06554900e-02 7.30421126e-01
-2.26870298e-01 3.31414133e-01 -9.64570045e-01 -1.18967068e+00
-8.88226867e-01 -1.42177796e+00 7.14747131e-01 3.40862095e-01
-1.23799339e-01 -1.21126509e+00 1.81455677e-03 4.39031810e-01
5.36677480e-01 -1.19912386e-01 1.05202782e+00 -1.34164667e+00
2.70205796e-01 -9.06340703e-02 -3.23140830e-01 1.02090323e+00
-2.11232960e-01 2.61123508e-01 -8.17662120e-01 -2.39585876e-01
-1.88962489e-01 -5.73014975e-01 6.82422280e-01 3.03792268e-01
5.56541324e-01 -3.94609988e-01 1.32704079e-01 3.35505664e-01
1.35238397e+00 3.47123414e-01 6.02039695e-01 1.19197018e-01
5.54943979e-01 6.47231787e-02 3.47157091e-01 3.40398848e-01
7.64042437e-01 2.87024111e-01 1.92832321e-01 4.58297469e-02
-2.35180408e-01 -3.66088569e-01 9.21975911e-01 1.92923737e+00
-7.47187883e-02 -4.66095954e-01 -1.18789828e+00 9.20856833e-01
-1.55475748e+00 -1.09843528e+00 -7.80715048e-01 2.21929765e+00
8.56255710e-01 2.57553846e-01 2.50836372e-01 9.70532447e-02
7.69191861e-01 1.64241225e-01 1.86357871e-02 -1.21172464e+00
-4.05246764e-01 2.00184062e-01 5.57761192e-01 7.42601216e-01
-1.04237545e+00 1.00765860e+00 4.80657434e+00 1.06306624e+00
-1.26380599e+00 4.47947800e-01 4.72415328e-01 5.80393076e-01
-2.85058081e-01 2.88897660e-02 -1.30662370e+00 7.38086104e-01
1.64311087e+00 -3.78295213e-01 1.82856470e-01 7.77250826e-01
3.22077990e-01 3.66667733e-02 -7.32857168e-01 8.63185406e-01
3.74911845e-01 -1.17036200e+00 2.36623688e-03 -1.88456103e-01
8.57721388e-01 6.79326117e-01 3.66282947e-02 8.32215726e-01
5.47768056e-01 -1.12478483e+00 1.08318377e+00 4.05279487e-01
9.97665286e-01 -7.41962433e-01 1.44165409e+00 8.05303395e-01
-1.21575427e+00 -2.94532683e-02 -3.45829666e-01 6.80030137e-02
1.71282634e-01 6.58137053e-02 -6.44279420e-01 7.52273798e-01
5.00113964e-01 7.96560705e-01 -2.57103860e-01 7.89415002e-01
9.20118485e-03 9.96103466e-01 5.13477139e-02 -4.77484345e-01
7.64378190e-01 1.55101851e-01 5.60863733e-01 1.84121454e+00
5.64139843e-01 -3.22551787e-01 1.93128467e-01 5.85187554e-01
-1.99955136e-01 6.44351363e-01 -1.86616391e-01 -1.28832087e-01
5.74614048e-01 1.16448760e+00 -1.21200621e-01 -5.96916914e-01
-6.14044189e-01 8.17260563e-01 2.44058549e-01 3.75454389e-02
-7.40076065e-01 -7.52444983e-01 6.33423030e-02 -4.15655300e-02
1.81888804e-01 -1.50174841e-01 9.99647379e-02 -1.06786680e+00
-3.88111591e-01 -1.02277589e+00 1.94465473e-01 -6.68742061e-01
-1.23267508e+00 1.05667722e+00 -1.98520645e-01 -1.05032015e+00
-2.63761342e-01 -6.58111870e-01 -8.65591347e-01 1.40893126e+00
-1.54434764e+00 -1.47783744e+00 -4.62378711e-02 7.03053057e-01
1.09736001e+00 -8.32515061e-01 7.83731759e-01 5.56496203e-01
-5.97094595e-01 9.51410174e-01 6.13782883e-01 4.94550318e-01
5.95617294e-01 -1.31023335e+00 1.05159298e-01 7.37559140e-01
-1.52915329e-01 4.13524657e-01 4.86081362e-01 -4.91791934e-01
-8.62951100e-01 -1.16034687e+00 1.88270211e+00 -8.62222910e-02
8.20533872e-01 -3.63132030e-01 -6.22561038e-01 7.66697347e-01
6.82508945e-01 -4.03541178e-01 5.64385772e-01 -2.39223212e-01
-1.85730547e-01 3.87825102e-01 -1.18402219e+00 3.35136116e-01
4.20413703e-01 -7.82919049e-01 -7.22677827e-01 3.90304267e-01
6.51472270e-01 -5.85009217e-01 -9.58286762e-01 -1.03168972e-01
4.09380227e-01 -1.07136035e+00 4.16324675e-01 -5.28914332e-01
7.22173810e-01 1.34319648e-01 -5.77324808e-01 -1.08245134e+00
-1.96898300e-02 -4.45959479e-01 2.78157413e-01 1.55669701e+00
8.45707178e-01 -6.22617960e-01 8.96581560e-02 -1.47207543e-01
-6.25774205e-01 -5.82105756e-01 -1.22858429e+00 -7.49363303e-01
6.25683725e-01 -5.23227155e-01 1.94833353e-01 9.95084524e-01
1.04847006e-01 6.03765309e-01 -5.72469652e-01 -1.98224455e-01
1.09789878e-01 -6.78750575e-01 5.56904495e-01 -8.14736962e-01
-1.18107647e-01 -5.88116348e-01 -3.07426929e-01 -9.01338816e-01
3.90238345e-01 -1.21652853e+00 3.46194729e-02 -1.60855567e+00
4.21988517e-01 -2.10354552e-01 -1.61843881e-01 4.81857598e-01
-6.77230256e-03 7.46107027e-02 3.17394286e-01 2.42062077e-01
-2.35196471e-01 4.12967086e-01 6.60290778e-01 -3.22948322e-02
-8.41881931e-02 1.16105147e-01 -1.24225862e-01 6.56718910e-01
6.09124124e-01 -3.31145465e-01 3.51197347e-02 -2.53050208e-01
1.03321843e-01 2.15995759e-01 3.36665399e-02 -1.09674597e+00
1.23261005e-01 6.12960979e-02 -9.85561162e-02 -8.90421748e-01
1.53715804e-01 -3.56147766e-01 -2.50217497e-01 6.99665844e-01
-5.98286629e-01 3.93697113e-01 8.48270506e-02 -1.01189487e-01
-5.29218137e-01 -8.03465188e-01 1.00928557e+00 -1.79829806e-01
-4.99587864e-01 -1.58152699e-01 -7.26667941e-01 2.61559129e-01
8.71816337e-01 -1.00382723e-01 -3.45079690e-01 -5.00520706e-01
-7.73011446e-01 2.67503738e-01 -5.29859215e-02 4.63820934e-01
5.60956359e-01 -1.14782286e+00 -1.30837715e+00 1.91758037e-01
2.91175358e-02 -5.88940442e-01 2.97769070e-01 1.17455435e+00
-5.25345147e-01 5.11765301e-01 -6.16957061e-02 -2.42635354e-01
-1.61280334e+00 1.20630547e-01 3.99659276e-01 -6.82927668e-01
-1.05751418e-01 1.07155919e+00 -3.60184550e-01 -8.44547927e-01
1.84553131e-01 -3.15907419e-01 -5.23616910e-01 8.98910919e-04
3.34035426e-01 4.71877664e-01 -8.13984685e-03 -1.45852053e+00
-3.86684358e-01 2.79613137e-01 -2.38506436e-01 -2.16390863e-01
1.12271619e+00 -1.52798891e-01 -3.50312293e-01 7.79044807e-01
1.33636141e+00 4.50664550e-01 -1.55524284e-01 -8.19767118e-02
1.29441321e-01 1.31709814e-01 -9.18473676e-03 -7.78372049e-01
-6.17545843e-01 1.18873048e+00 6.21515155e-01 1.31457403e-01
7.23718166e-01 1.11746807e-02 9.84758914e-01 -8.20507631e-02
-1.70869470e-01 -1.18594635e+00 -2.00620398e-01 1.18497920e+00
8.51141453e-01 -1.10989165e+00 -5.92316031e-01 3.33879411e-01
-8.67381513e-01 1.33363545e+00 5.14726698e-01 -2.79829632e-02
7.82798052e-01 1.50164425e-01 1.58330023e-01 2.76857436e-01
-8.98947537e-01 -1.68185860e-01 3.77065569e-01 4.33220744e-01
1.06090724e+00 2.14016378e-01 -7.62867093e-01 6.29473746e-01
-5.68248570e-01 -1.69804975e-01 7.99660921e-01 4.02026057e-01
-6.48679793e-01 -8.78491044e-01 -3.00131626e-02 6.63742483e-01
-8.76176178e-01 -6.04737759e-01 8.88326380e-04 8.46311629e-01
5.02576172e-01 1.07357049e+00 4.35517505e-02 -8.38691950e-01
8.76890793e-02 3.45310599e-01 -3.40254724e-01 -6.26568079e-01
-1.18136239e+00 1.73440993e-01 5.60414910e-01 2.50021040e-01
-2.96503723e-01 -8.04012239e-01 -1.29403532e+00 -4.17655170e-01
-4.73091036e-01 3.01861376e-01 7.77600586e-01 1.07923388e+00
-7.10847154e-02 3.32046032e-01 4.69163060e-01 -4.37605262e-01
-7.72661686e-01 -1.47532678e+00 -5.52829146e-01 2.20987812e-01
4.44600672e-01 -8.32974911e-04 -5.56039333e-01 -2.22679228e-02] | [9.393171310424805, 10.678457260131836] |
bf447dbb-53b6-40c6-85e6-b8bc78456363 | cnn-based-methods-for-object-recognition-with | 2305.12417 | null | https://arxiv.org/abs/2305.12417v1 | https://arxiv.org/pdf/2305.12417v1.pdf | CNN-based Methods for Object Recognition with High-Resolution Tactile Sensors | Novel high-resolution pressure-sensor arrays allow treating pressure readings as standard images. Computer vision algorithms and methods such as Convolutional Neural Networks (CNN) can be used to identify contact objects. In this paper, a high-resolution tactile sensor has been attached to a robotic end-effector to identify contacted objects. Two CNN-based approaches have been employed to classify pressure images. These methods include a transfer learning approach using a pre-trained CNN on an RGB-images dataset and a custom-made CNN (TactNet) trained from scratch with tactile information. The transfer learning approach can be carried out by retraining the classification layers of the network or replacing these layers with an SVM. Overall, 11 configurations based on these methods have been tested: 8 transfer learning-based, and 3 TactNet-based. Moreover, a study of the performance of the methods and a comparative discussion with the current state-of-the-art on tactile object recognition is presented. | ['Jesús M. Gómez-de-Gabriel', 'Alfonso J. García-Cerezo', 'Juan M. Gandarias'] | 2023-05-21 | null | null | null | null | ['object-recognition'] | ['computer-vision'] | [ 6.19804919e-01 1.44249082e-01 1.60242915e-01 -3.09290439e-01
-2.82251984e-01 -2.65313517e-02 4.14290965e-01 -7.41087645e-02
-8.46592963e-01 6.22091770e-01 -4.42015111e-01 5.62656410e-02
-1.58915758e-01 -9.33171034e-01 -1.01972485e+00 -6.02280736e-01
1.46262810e-01 5.87820172e-01 7.27640867e-01 -6.10202737e-02
6.75952911e-01 6.16573155e-01 -1.87200117e+00 5.10906577e-01
5.47224045e-01 1.79475939e+00 4.53284949e-01 5.25860965e-01
-2.21906424e-01 2.02813193e-01 -3.44750673e-01 -1.01629114e-02
4.39184964e-01 -9.68915001e-02 -6.33168221e-01 -2.64332891e-01
2.88019180e-01 -3.95064920e-01 -2.09281594e-01 4.69308943e-01
4.94299233e-01 -3.01147252e-01 9.37966049e-01 -9.00147974e-01
-4.95643973e-01 3.11993450e-01 -4.91525466e-03 -4.54103231e-01
3.44340086e-01 -2.02351674e-01 1.74779415e-01 -9.21259165e-01
4.35876071e-01 1.09049606e+00 9.20495927e-01 4.80197936e-01
-9.31643844e-01 -2.48578951e-01 -6.78370237e-01 3.86409581e-01
-8.76027107e-01 1.57796636e-01 1.00905955e+00 -4.51661140e-01
1.12617850e+00 3.08719367e-01 8.35903049e-01 1.44757891e+00
6.40315533e-01 4.23281699e-01 1.63363063e+00 -7.35248685e-01
6.43927574e-01 2.88677514e-01 -3.08619946e-01 3.90987307e-01
3.27643842e-01 3.17386270e-01 -5.62549055e-01 -6.21528663e-02
1.36312258e+00 -2.00174734e-01 8.05216357e-02 -6.55719936e-01
-7.92328179e-01 6.07271254e-01 9.34957564e-01 3.89155746e-01
-7.00463891e-01 1.36017352e-01 5.61883926e-01 2.37869844e-01
3.08163673e-01 7.04933822e-01 -6.13728702e-01 -8.43229219e-02
-5.73346555e-01 -5.55308871e-02 1.20057106e+00 5.74528515e-01
9.24793720e-01 -1.89833969e-01 -1.16511829e-01 9.29759324e-01
2.84496218e-01 4.23918724e-01 4.78089392e-01 -6.27136767e-01
1.79002181e-01 6.54812276e-01 1.54896706e-01 -9.31712627e-01
-2.66714394e-01 4.30828869e-01 -5.25857985e-01 8.87257397e-01
5.02609968e-01 -2.93555498e-01 -1.24567199e+00 7.69945025e-01
4.86409618e-03 -1.35271236e-01 -1.74137708e-02 1.03264832e+00
6.51704550e-01 3.00116211e-01 -8.19842219e-02 4.81635034e-01
7.58608401e-01 -9.50975001e-01 -2.85272926e-01 -2.45761737e-01
3.23956348e-02 -6.65672719e-01 8.21937501e-01 5.89078367e-01
-6.80918634e-01 -8.24174523e-01 -1.36092341e+00 2.02678695e-01
-9.65588391e-01 5.54157376e-01 4.05197144e-01 3.54825228e-01
-7.34929144e-01 1.31745601e+00 -1.06109285e+00 -7.32570887e-01
3.68951887e-01 6.06405795e-01 -4.33796346e-01 1.37241781e-01
-9.60083961e-01 1.42081487e+00 6.37856662e-01 5.18005669e-01
-1.29266351e-01 -2.02187940e-01 -5.41941583e-01 -1.40600160e-01
-4.91292663e-02 -3.18420321e-01 1.08878911e+00 -1.03750849e+00
-2.40541625e+00 8.93146396e-01 5.43482661e-01 -4.00327384e-01
8.52135003e-01 -7.05162585e-01 5.89571148e-02 4.20153230e-01
-2.92905807e-01 5.54289162e-01 9.10121322e-01 -1.55874431e+00
-3.08856487e-01 -2.61396527e-01 -3.63821447e-01 -8.55917484e-02
-3.07924986e-01 -1.66228101e-01 -9.64195579e-02 -1.18265361e-01
1.89586699e-01 -8.24430645e-01 1.40712500e-01 4.34446454e-01
-5.00122786e-01 -7.72097930e-02 1.06376255e+00 -6.20007157e-01
3.31619143e-01 -1.80482495e+00 2.11363494e-01 5.92693090e-01
-3.01850855e-01 7.66154647e-01 3.93768400e-03 6.64231360e-01
2.53296316e-01 -3.82390976e-01 -1.06934689e-01 -2.10788995e-01
-8.02209005e-02 5.62073290e-01 1.87756568e-02 3.37857418e-02
5.10545492e-01 9.77452040e-01 -6.22717023e-01 -2.70578146e-01
9.01452780e-01 4.92294222e-01 4.15661186e-02 5.09793937e-01
-1.57858297e-01 4.44360554e-01 -4.46400493e-01 6.95635438e-01
7.13159800e-01 2.87040472e-01 7.53979245e-03 -4.90259290e-01
-3.71380657e-01 -1.92266077e-01 -7.24360228e-01 1.63815761e+00
-6.14090860e-01 3.35310906e-01 2.02752113e-01 -1.10810864e+00
1.66259634e+00 8.60736445e-02 2.74829835e-01 -7.37107098e-01
6.60403430e-01 8.04427028e-01 -7.58742020e-02 -8.30511451e-01
1.19066894e-01 1.72613606e-01 1.88484490e-01 -5.72317429e-02
2.70499796e-01 -2.07267523e-01 -5.39262652e-01 -9.50169981e-01
9.94764566e-01 6.63026333e-01 -2.76848733e-01 -1.34426221e-01
4.81555045e-01 9.41022262e-02 -4.19298820e-02 8.06073308e-01
-3.53149995e-02 8.20420325e-01 1.45979017e-01 -6.87412500e-01
-1.32450044e+00 -1.10165417e+00 -1.53313726e-01 7.31344223e-01
1.63378283e-01 4.51775342e-01 -7.97293007e-01 -1.80480793e-01
6.11580551e-01 -1.00837521e-01 -1.11824155e+00 -8.83187056e-02
-5.86721182e-01 5.13658673e-02 2.80576736e-01 1.09216022e+00
9.28636909e-01 -1.62961113e+00 -1.34699643e+00 3.22272718e-01
5.06141841e-01 -1.03491247e+00 7.25633681e-01 7.46975720e-01
-8.54856491e-01 -1.10201347e+00 -8.00864518e-01 -9.31117594e-01
5.82731724e-01 -5.38791299e-01 3.77273709e-01 -3.09388220e-01
-6.34009361e-01 4.40718472e-01 -7.94814289e-01 -8.80715609e-01
-2.72182673e-01 2.77236670e-01 -4.05625291e-02 -5.89173324e-02
5.13000488e-01 -4.75909978e-01 -5.96452355e-01 2.47826681e-01
-5.86331069e-01 -1.31565258e-01 1.22204685e+00 9.65759575e-01
6.15690112e-01 -6.02257907e-01 4.46138352e-01 -5.16372859e-01
6.64096713e-01 1.42548874e-01 -2.10085586e-01 2.91998595e-01
-2.86924064e-01 -1.18726820e-01 6.11743867e-01 -6.90409720e-01
-8.47095370e-01 4.33453768e-01 -2.27054462e-01 -8.00065160e-01
-4.16224539e-01 5.52612185e-01 2.58644849e-01 -6.72092199e-01
8.16538632e-01 -1.00507058e-01 7.37843752e-01 -6.28016710e-01
-1.40925750e-01 1.09842610e+00 7.48721838e-01 -4.78159428e-01
6.07414067e-01 1.58874020e-01 3.00000366e-02 -7.63925731e-01
-4.77658302e-01 -2.30975151e-01 -1.12259924e+00 -5.30009151e-01
1.06743872e+00 -1.88710332e-01 -5.08274853e-01 1.24078500e+00
-1.00788999e+00 -8.30060780e-01 -2.17242703e-01 4.37685877e-01
-7.33445525e-01 1.43939510e-01 -8.15312326e-01 -7.13327885e-01
-5.94368696e-01 -7.23848224e-01 1.07132757e+00 5.51501870e-01
-1.17914081e-01 -8.41575801e-01 3.15749943e-02 5.07346019e-02
1.07872915e+00 6.38735831e-01 6.32993698e-01 -5.93344450e-01
3.74921225e-02 -7.19921589e-01 -3.60112816e-01 6.71932101e-01
3.66358906e-01 3.74455862e-02 -1.20005560e+00 -4.99689393e-02
-1.93299174e-01 -8.14651906e-01 8.45330238e-01 2.13832051e-01
1.21353483e+00 5.04321158e-02 -3.86786550e-01 -8.52504596e-02
1.59645402e+00 2.80870527e-01 1.00965261e+00 5.61005116e-01
3.94133747e-01 5.12783706e-01 6.68758988e-01 1.84006065e-01
-3.50966752e-01 5.85839808e-01 5.12268960e-01 -3.46438795e-01
1.25581950e-01 -2.98244685e-01 -1.48976922e-01 4.53535497e-01
-6.80904686e-01 2.75273889e-01 -9.38776731e-01 6.31371588e-02
-1.90501392e+00 -4.50354934e-01 2.59994209e-01 2.09417057e+00
6.20552123e-01 3.89712840e-01 -1.95881531e-01 5.61621964e-01
6.13340795e-01 -5.38019359e-01 -7.00370967e-01 -8.38535666e-01
2.08187401e-01 9.62436140e-01 5.41104198e-01 1.15842924e-01
-1.43502009e+00 7.60844946e-01 6.17125607e+00 2.44131878e-01
-1.98526919e+00 -3.06921959e-01 7.06702098e-02 5.73978126e-01
5.79386532e-01 -5.64236820e-01 -3.41844596e-02 2.99250394e-01
5.92817545e-01 4.06855434e-01 3.30965161e-01 1.05050719e+00
-4.48944747e-01 -6.85129464e-01 -1.18232405e+00 9.10759687e-01
1.52908683e-01 -1.00912094e+00 -2.04657838e-01 -1.81848690e-01
1.95400178e-01 9.39730555e-02 -1.29270092e-01 2.52756298e-01
-4.91173685e-01 -8.55010331e-01 3.14348459e-01 1.03450608e+00
9.32015896e-01 -2.02844396e-01 1.30731547e+00 -1.41696215e-01
-8.82014036e-01 -1.05951026e-01 -7.38186896e-01 -4.94032025e-01
-2.57887691e-01 1.97430044e-01 -7.33867466e-01 7.49730527e-01
1.25999308e+00 4.02375728e-01 -4.43753570e-01 1.16181433e+00
-2.69775186e-02 2.02975035e-01 -4.92283672e-01 -5.95469475e-01
1.84789881e-01 -2.35450760e-01 5.79812704e-03 1.15271413e+00
4.44522679e-01 -4.38355237e-01 -1.47312768e-02 1.06141639e+00
5.05854964e-01 -2.41930666e-03 -7.42490232e-01 1.91664755e-01
3.48528534e-01 1.40856552e+00 -7.47382045e-01 -1.16693497e-01
-1.33689180e-01 1.18325317e+00 2.97403187e-01 1.54419959e-01
-4.23061162e-01 -8.08518171e-01 2.48445794e-01 6.97153658e-02
7.85613239e-01 -2.01752588e-01 -4.95521575e-01 -5.47629893e-01
1.83774427e-01 -1.04588235e-03 -4.93304163e-01 -1.15743434e+00
-1.66279292e+00 4.42869365e-01 -6.63138777e-02 -1.12226045e+00
-9.78748873e-02 -1.66595769e+00 -7.98753321e-01 9.17036772e-01
-1.23857498e+00 -1.02638662e+00 -7.71237731e-01 2.36958623e-01
1.41894236e-01 -8.47934335e-02 1.22104764e+00 -1.76467195e-01
-1.45080447e-01 1.76572815e-01 1.26710191e-01 4.51801002e-01
7.52676785e-01 -1.24271667e+00 -1.41017646e-01 -2.85304245e-02
-7.14833796e-01 1.59694061e-01 3.74615878e-01 -5.17483771e-01
-1.56081569e+00 -9.88318086e-01 4.29025084e-01 -6.99299946e-02
5.46435475e-01 -4.78370875e-01 -1.13058794e+00 3.63933712e-01
1.34179905e-01 9.42337215e-02 2.68505931e-01 -1.91389993e-01
-2.26847336e-01 -1.61254272e-01 -1.37668049e+00 1.29602343e-01
4.76358384e-01 -5.81387460e-01 -1.01012087e+00 -7.90890157e-02
-1.86888371e-02 -8.32373977e-01 -1.31084919e+00 8.70675564e-01
1.04786098e+00 -1.04496443e+00 7.20624268e-01 -9.68416259e-02
1.00334489e+00 1.16950385e-01 1.11582868e-01 -1.27734208e+00
-8.91792625e-02 3.11813146e-01 -2.37980746e-02 6.82549655e-01
2.59356529e-01 -1.01146924e+00 1.04280198e+00 6.74450815e-01
-2.90375084e-01 -1.18001425e+00 -1.10367763e+00 -8.95974994e-01
2.68161297e-01 1.29308328e-01 -4.68595065e-02 4.41676676e-01
2.66003728e-01 1.03191398e-01 1.24734621e-02 -2.07730964e-01
2.53493488e-01 -1.41095251e-01 5.16161501e-01 -1.72021699e+00
1.01500019e-01 -3.11888933e-01 -8.81520748e-01 -4.55242127e-01
-1.80584565e-02 -2.98470050e-01 4.21865851e-01 -1.74306142e+00
-1.99133620e-01 -5.14575064e-01 -4.11121398e-01 9.85204518e-01
3.52554798e-01 1.22874841e-01 -4.19492275e-02 8.62821713e-02
-7.35741705e-02 4.36310083e-01 1.37298298e+00 -8.37115943e-02
-1.99299231e-01 3.31610441e-02 1.40858486e-01 2.87832707e-01
1.20483780e+00 1.83152556e-01 1.16641238e-01 -2.23834917e-01
-1.74622580e-01 -1.69886306e-01 7.53815472e-01 -1.72944951e+00
3.84182334e-01 4.73937780e-01 8.32527757e-01 -3.96348000e-01
5.11393130e-01 -1.07624483e+00 -7.53445132e-03 9.20175433e-01
-1.50873199e-01 -5.54808795e-01 4.17751670e-01 2.65224367e-01
-3.87398541e-01 -2.06530020e-01 6.21665001e-01 -1.71009824e-01
-6.26242936e-01 -2.52372026e-01 -4.54720646e-01 -7.67905533e-01
1.11268008e+00 -6.67957485e-01 -3.25070173e-01 1.09172158e-01
-8.80463183e-01 -2.52864480e-01 3.82394224e-01 4.56704170e-01
8.82477105e-01 -1.30826688e+00 -1.91389173e-01 2.77304739e-01
8.47018585e-02 -1.57578718e-02 5.40830716e-02 6.69720590e-01
-8.90038848e-01 4.82170105e-01 -1.20364916e+00 -8.34359407e-01
-8.58358204e-01 3.37856382e-01 6.39766455e-01 2.01902419e-01
-3.53012085e-01 5.16466558e-01 -8.65661204e-01 -7.96663046e-01
3.85532051e-01 -5.90061545e-01 -2.80355513e-01 -4.52719182e-01
-1.75146013e-01 3.24638009e-01 3.79313946e-01 -4.26328406e-02
-2.04100132e-01 9.76618707e-01 3.02054137e-01 1.16582952e-01
1.32032752e+00 5.97127974e-01 -1.11000843e-01 7.16437221e-01
1.21458149e+00 -8.69551182e-01 -1.45250618e+00 -7.74036273e-02
1.05302539e-02 -1.09266594e-01 -1.96660534e-01 -1.17221546e+00
-7.52976894e-01 9.54058230e-01 8.50623190e-01 2.62479693e-01
9.90953743e-01 -2.15017036e-01 4.07416880e-01 6.79844558e-01
9.17988181e-01 -1.60810888e+00 3.83576065e-01 6.11001968e-01
1.44725513e+00 -1.31058526e+00 -3.50672007e-01 -8.55120540e-01
-2.25361675e-01 1.69055223e+00 9.57163870e-01 -7.41446912e-01
8.11265707e-01 3.30829650e-01 2.92850435e-01 5.19087240e-02
-8.97779986e-02 -5.78869730e-02 1.74307972e-01 8.77288878e-01
8.41762424e-02 4.87889461e-02 -4.01407242e-01 3.14087152e-01
-4.69620004e-02 6.71948552e-01 1.24814380e-02 1.64332712e+00
-5.72765946e-01 -1.06051970e+00 -3.68236363e-01 7.65181303e-01
4.40783426e-02 4.79904652e-01 -8.17717552e-01 8.90760422e-01
2.64700115e-01 5.56878805e-01 3.18442285e-01 -9.47473764e-01
6.07210934e-01 2.07848474e-01 1.01596951e+00 -4.05058712e-01
-7.73653328e-01 -3.51465613e-01 4.57232539e-03 -6.83308601e-01
-5.89670599e-01 -4.35414076e-01 -8.21097016e-01 2.47475490e-01
-2.38927484e-01 -2.76569068e-01 1.15440583e+00 8.60688448e-01
2.10462019e-01 3.91930968e-01 4.65402216e-01 -1.87003279e+00
-6.00298643e-01 -1.49558127e+00 -6.89737976e-01 2.03411043e-01
-2.53418326e-01 -9.62400913e-01 -2.26968423e-01 -2.12762132e-01] | [5.85770845413208, -0.7982520461082458] |
222b2a15-8646-4d56-a956-67c50201c5f4 | sketchxai-a-first-look-at-explainability-for | 2304.11744 | null | https://arxiv.org/abs/2304.11744v1 | https://arxiv.org/pdf/2304.11744v1.pdf | SketchXAI: A First Look at Explainability for Human Sketches | This paper, for the very first time, introduces human sketches to the landscape of XAI (Explainable Artificial Intelligence). We argue that sketch as a ``human-centred'' data form, represents a natural interface to study explainability. We focus on cultivating sketch-specific explainability designs. This starts by identifying strokes as a unique building block that offers a degree of flexibility in object construction and manipulation impossible in photos. Following this, we design a simple explainability-friendly sketch encoder that accommodates the intrinsic properties of strokes: shape, location, and order. We then move on to define the first ever XAI task for sketch, that of stroke location inversion SLI. Just as we have heat maps for photos, and correlation matrices for text, SLI offers an explainability angle to sketch in terms of asking a network how well it can recover stroke locations of an unseen sketch. We offer qualitative results for readers to interpret as snapshots of the SLI process in the paper, and as GIFs on the project page. A minor but interesting note is that thanks to its sketch-specific design, our sketch encoder also yields the best sketch recognition accuracy to date while having the smallest number of parameters. The code is available at \url{https://sketchxai.github.io}. | ['Yi-Zhe Song', 'Tao Xiang', 'Kaiyue Pang', 'Ke Li', 'Yulia Gryaditskaya', 'Zhiyu Qu'] | 2023-04-23 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Qu_SketchXAI_A_First_Look_at_Explainability_for_Human_Sketches_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Qu_SketchXAI_A_First_Look_at_Explainability_for_Human_Sketches_CVPR_2023_paper.pdf | cvpr-2023-1 | ['sketch-recognition'] | ['computer-vision'] | [ 1.35531902e-01 3.52562994e-01 -3.76115710e-01 -2.61514843e-01
-1.02310635e-01 -1.08144355e+00 9.92205203e-01 -4.87061560e-01
3.11639011e-01 3.57326448e-01 5.90988278e-01 -4.52284843e-01
-5.72858512e-01 -6.35026276e-01 -5.30570805e-01 -5.01883328e-02
-1.71194300e-02 4.12675291e-01 -4.30592984e-01 -9.04797241e-02
6.52908862e-01 8.33156884e-01 -1.32006168e+00 4.38305438e-01
3.30625564e-01 7.11495459e-01 1.29868373e-01 8.95587027e-01
-4.23045337e-01 5.78383625e-01 -5.40012598e-01 -7.28613377e-01
1.70935899e-01 -4.45794851e-01 -7.37804174e-01 -9.39782485e-02
8.37535501e-01 -5.24185538e-01 -7.65932798e-01 4.94091749e-01
1.62649184e-01 -2.00216234e-01 8.35383117e-01 -1.65845668e+00
-1.22250175e+00 7.21504092e-01 -5.71757853e-02 -1.79792002e-01
4.90386009e-01 2.81740129e-01 1.30069661e+00 -1.27800119e+00
1.03970742e+00 1.14193058e+00 8.01182032e-01 7.10802436e-01
-1.19662249e+00 -5.76575100e-01 -7.13093728e-02 -9.34007838e-02
-1.26215160e+00 -4.14180189e-01 1.04254544e+00 -3.79960686e-01
5.54741025e-01 8.10594082e-01 7.89799094e-01 1.35534513e+00
-2.75704384e-01 1.30385864e+00 8.67001712e-01 -3.28950316e-01
1.22515425e-01 -1.64083634e-02 3.53523083e-02 8.23020816e-01
1.09022059e-01 1.03581414e-01 -6.72573566e-01 1.74190830e-02
1.47815812e+00 3.85764748e-01 -1.73241019e-01 -5.70577145e-01
-1.50723016e+00 5.62559247e-01 5.62665939e-01 3.76023531e-01
-1.69596747e-02 4.49555308e-01 9.26887803e-03 3.64443719e-01
1.53374761e-01 9.21574771e-01 -1.65022120e-01 -4.89971280e-01
-8.83747935e-01 3.03267479e-01 1.09545827e+00 1.32491994e+00
5.84953725e-01 -8.54301751e-02 -4.54068072e-02 6.28524899e-01
-1.74369849e-02 5.44838607e-01 -1.51305392e-01 -1.19462597e+00
3.15541327e-01 6.32860780e-01 -5.23758233e-02 -1.23166645e+00
-1.37210861e-01 -2.25421682e-01 -1.00284100e+00 6.84777975e-01
6.85422719e-01 2.29775786e-01 -6.05081022e-01 1.51976728e+00
-4.05230850e-01 -7.80380592e-02 -3.99511397e-01 1.03775585e+00
7.98521280e-01 4.72504556e-01 -1.06013983e-01 6.30925536e-01
1.40371037e+00 -7.49075413e-01 -5.42269886e-01 -6.88125193e-03
1.15468070e-01 -7.01586664e-01 1.57779741e+00 4.44574177e-01
-1.28507364e+00 -5.97380161e-01 -1.24758565e+00 -3.38780642e-01
-6.86566055e-01 4.73823160e-01 1.08982193e+00 5.57603061e-01
-1.01313329e+00 9.65202808e-01 -4.87721056e-01 -6.76363230e-01
7.65252411e-01 8.62921774e-03 -5.43712676e-01 5.26861884e-02
-7.86443830e-01 7.82723665e-01 -4.08644021e-01 3.92717309e-02
-3.11291188e-01 -8.06686640e-01 -5.49979210e-01 2.98993945e-01
3.52187932e-01 -6.96106911e-01 1.00165308e+00 -9.88980234e-01
-1.26763403e+00 6.99274659e-01 -1.04835041e-01 7.34771639e-02
6.84306204e-01 -1.22909240e-01 -4.37537283e-02 1.26294866e-01
-9.96100307e-02 8.95534515e-01 8.23122501e-01 -1.33360040e+00
-7.81769156e-02 -1.96638614e-01 2.35067844e-01 -6.56636879e-02
-1.66097879e-01 -2.92867482e-01 -4.48524207e-01 -9.87518668e-01
1.48731187e-01 -1.01361036e+00 2.60772347e-01 9.79818225e-01
-4.08313096e-01 -2.85065413e-01 8.45131516e-01 -5.28163671e-01
1.28053570e+00 -2.10656381e+00 1.96845651e-01 4.44678813e-01
6.95087612e-01 7.42276162e-02 -2.46154785e-01 9.25094724e-01
-8.43578652e-02 7.49931574e-01 -1.96717143e-01 -3.72174978e-01
5.90608120e-01 2.22714454e-01 -5.47757447e-01 2.13402852e-01
2.26218864e-01 1.35985541e+00 -8.83282959e-01 -2.04896808e-01
2.54569262e-01 5.56106389e-01 -3.87230992e-01 1.14269100e-01
-2.44356636e-02 5.82349524e-02 -3.55386496e-01 7.66098499e-01
5.23766458e-01 -3.17059338e-01 1.80602372e-01 -3.06557149e-01
-2.06059620e-01 6.97394684e-02 -1.16235316e+00 1.98611736e+00
-2.90984601e-01 1.27487469e+00 -8.16603228e-02 -3.80334139e-01
9.97861862e-01 1.71986416e-01 9.01127383e-02 -3.95426512e-01
-3.62015784e-01 1.28193825e-01 -2.13557631e-01 -2.88796037e-01
5.06536484e-01 1.80213839e-01 -1.27934096e-02 1.03247356e+00
-2.76100487e-01 -4.76275772e-01 -1.59903497e-01 5.03572464e-01
1.06519496e+00 3.04771870e-01 2.36271724e-01 -3.78816932e-01
3.74888293e-02 -2.95680732e-01 -2.74452269e-01 1.13269019e+00
1.33542851e-01 1.06409407e+00 6.08508289e-01 -9.94645059e-01
-1.59797597e+00 -1.40026760e+00 -3.37297507e-02 8.51042569e-01
1.03373406e-02 -5.25855303e-01 -6.08940125e-01 -4.12856430e-01
3.53966385e-01 6.01503134e-01 -9.38323438e-01 3.48029315e-01
-5.23453593e-01 4.94020015e-01 7.05053627e-01 7.10673392e-01
1.99890092e-01 -1.48765230e+00 -5.46800256e-01 -3.90653789e-01
1.02646768e-01 -3.72860134e-01 -6.31394207e-01 -1.44722223e-01
-5.83765984e-01 -8.56679916e-01 -7.60549426e-01 -6.04036868e-01
7.53080606e-01 3.27739805e-01 1.37397265e+00 8.08293641e-01
-5.66126108e-01 7.50206113e-01 -1.03354134e-01 -3.64881754e-01
-1.81789786e-01 3.41374338e-01 -2.72155643e-01 -4.33606386e-01
-4.06959392e-02 -8.58947873e-01 -8.18258762e-01 3.40994388e-01
-1.02333653e+00 6.29686892e-01 6.18490160e-01 6.54537678e-01
1.05860293e-01 -4.73665237e-01 2.65989572e-01 -9.09960508e-01
8.97860825e-01 -2.97901928e-01 3.37205417e-02 5.02838433e-01
-5.67274868e-01 1.97682559e-01 4.66071576e-01 -3.61818552e-01
-6.83735669e-01 -7.92317688e-02 2.22349733e-01 -2.42889404e-01
-2.51934946e-01 5.27043156e-02 -8.83315429e-02 1.20968878e-01
7.05599785e-01 6.89977258e-02 2.80212611e-01 -5.07969737e-01
9.55613434e-01 4.67225015e-01 7.04537451e-01 -8.71521652e-01
1.03469491e+00 4.60607976e-01 -6.97885230e-02 -8.86959553e-01
-2.88431495e-01 8.93597454e-02 -9.02906954e-01 -3.76753062e-01
5.02978563e-01 -2.71556318e-01 -1.26946938e+00 2.83019617e-02
-1.31500494e+00 -4.93934631e-01 -4.70336735e-01 -3.61503065e-02
-8.69111836e-01 1.39482483e-01 -5.25738120e-01 -7.41675794e-01
-1.54928505e-01 -7.71049023e-01 1.10516453e+00 7.29269013e-02
-9.87671912e-01 -9.69360828e-01 -1.87076211e-01 -4.41611782e-02
7.09872723e-01 3.27685863e-01 1.10397816e+00 -3.52180600e-01
-1.02450156e+00 -1.39159352e-01 -6.25541329e-01 -2.31073543e-01
1.29921719e-01 2.44550988e-01 -9.52933908e-01 1.35762477e-02
-5.22750854e-01 -2.14942709e-01 5.01739383e-01 1.43578816e-02
1.37609756e+00 -6.30149126e-01 -3.25674772e-01 6.92376614e-01
1.29868162e+00 -1.06721304e-01 8.13384235e-01 2.19651252e-01
8.62080634e-01 4.66025412e-01 3.27917859e-02 4.64916140e-01
2.46366888e-01 5.61832070e-01 1.94065437e-01 -3.51038218e-01
-4.59624052e-01 -8.64981294e-01 -2.66704708e-01 4.76632386e-01
-3.28999251e-01 -3.78651112e-01 -9.46221948e-01 2.09182292e-01
-1.86861503e+00 -1.25688684e+00 -3.60222347e-03 1.95353878e+00
4.75763083e-01 -7.17633218e-02 2.37122893e-01 2.31886551e-01
3.99209738e-01 2.38033235e-01 -3.61997992e-01 -5.41791141e-01
-1.85350329e-01 3.13348055e-01 1.77901939e-01 7.15281546e-01
-6.53919697e-01 8.62270892e-01 6.71434546e+00 5.00960410e-01
-7.34020770e-01 -5.67231834e-01 4.81657147e-01 -8.45569745e-02
-9.16006446e-01 1.54167667e-01 -7.21438229e-02 4.63277578e-01
2.72072405e-01 -5.88532984e-02 1.17525911e+00 6.91827238e-01
5.02914637e-02 3.07980418e-01 -1.59790826e+00 1.06959593e+00
1.76183898e-02 -1.89854562e+00 3.54014516e-01 4.59093451e-02
2.01501086e-01 -5.97207069e-01 3.08282614e-01 -1.34253263e-01
7.94168115e-02 -1.43417931e+00 1.00468218e+00 1.04298449e+00
1.22875559e+00 -5.60024559e-01 1.56671666e-02 5.43986782e-02
-1.14182329e+00 4.79869135e-02 -2.19003320e-01 -4.20560211e-01
-2.13100910e-01 -5.14309891e-02 -6.79065526e-01 8.09772238e-02
3.42283785e-01 8.67928028e-01 -7.11356580e-01 6.77762091e-01
-3.18544000e-01 2.77144432e-01 -1.34907633e-01 -4.22896922e-01
2.70408709e-02 -2.66281247e-01 4.72598255e-01 1.43887234e+00
4.05328423e-01 4.13196951e-01 -3.44739020e-01 1.44760919e+00
-1.04283780e-01 -4.56870168e-01 -1.17747116e+00 -1.93985149e-01
9.37401831e-01 1.17896438e+00 -8.04144382e-01 -2.44557157e-01
-1.30692706e-01 1.37015402e+00 4.06616867e-01 4.62228924e-01
-5.47817945e-01 -6.47529244e-01 7.55525053e-01 2.23265395e-01
1.47304267e-01 -5.39340973e-01 -9.73219693e-01 -8.90866220e-01
5.46087809e-02 -9.41997170e-01 -1.40059873e-01 -1.41891491e+00
-1.36430025e+00 4.11840886e-01 -1.76250070e-01 -1.06429577e+00
-1.86467692e-01 -8.60368729e-01 -8.78741801e-01 7.96730876e-01
-7.96177626e-01 -1.51509559e+00 -5.25839031e-01 2.08147064e-01
6.17613733e-01 -8.65022615e-02 9.68083799e-01 -2.30225753e-02
-2.56937712e-01 7.71664560e-01 -1.06943339e-01 2.22182348e-01
5.04207075e-01 -1.49708021e+00 1.24973571e+00 2.78645962e-01
7.05239236e-01 1.11221600e+00 6.08248353e-01 -5.69570839e-01
-1.83908892e+00 -2.43342519e-01 1.01834667e+00 -1.18444920e+00
6.54931247e-01 -7.80327857e-01 -8.28505456e-01 7.38089383e-01
2.52152622e-01 -3.80167991e-01 4.20038342e-01 2.96708852e-01
-5.01339316e-01 2.75454134e-01 -8.80952954e-01 1.02531457e+00
1.76243055e+00 -1.00745070e+00 -4.75803196e-01 1.15146682e-01
3.03662896e-01 -2.47591302e-01 -8.62431943e-01 -2.85707325e-01
1.42739797e+00 -1.04944098e+00 1.25622511e+00 -6.82083607e-01
8.60195994e-01 -1.77524462e-01 8.48030597e-02 -1.03555954e+00
-7.20253587e-01 -9.51274753e-01 3.70109491e-02 1.27241361e+00
4.99745309e-01 -3.99064004e-01 1.11913645e+00 1.11276793e+00
1.92748517e-01 -8.68175447e-01 -4.41020280e-01 -6.62089825e-01
-3.15276459e-02 -4.92206901e-01 9.36528563e-01 9.47498858e-01
4.45311040e-01 7.85848722e-02 -4.57679093e-01 -2.89389849e-01
5.72061419e-01 2.49044642e-01 1.17400944e+00 -1.41094410e+00
-1.08199202e-01 -9.05317783e-01 -2.72428691e-01 -1.26949811e+00
-9.45339054e-02 -1.02449405e+00 -3.34792316e-01 -1.53533459e+00
3.35453451e-01 -6.97234392e-01 1.95431963e-01 5.79121411e-01
5.69209270e-02 3.25024933e-01 7.45112896e-01 5.90588689e-01
-3.32403809e-01 1.60404578e-01 1.37502825e+00 -8.67350698e-02
3.75916883e-02 -2.48669460e-01 -9.14172828e-01 4.86471236e-01
7.26437926e-01 -1.06836259e-01 -5.12367725e-01 -7.16041028e-01
2.95271993e-01 2.32509926e-01 6.65297031e-01 -7.62247920e-01
3.20814073e-01 -1.37601480e-01 7.60713458e-01 -4.49555993e-01
4.86929625e-01 -8.88951182e-01 4.98888910e-01 2.44935423e-01
-7.52813339e-01 4.49492127e-01 4.67809364e-02 3.31663311e-01
2.89081693e-01 -9.32315663e-02 1.45300940e-01 -2.82521427e-01
-5.20493090e-01 3.27175856e-01 -2.70661265e-01 -1.36334226e-01
3.55597824e-01 -6.60865426e-01 -5.03036380e-01 -8.57681870e-01
-5.72763979e-01 -1.84972152e-01 9.33485806e-01 6.01077557e-01
7.87353396e-01 -1.68697584e+00 -4.40283597e-01 3.77796203e-01
-2.52675451e-02 -4.60461289e-01 -2.39557377e-03 4.73326333e-02
-5.46201050e-01 3.19702417e-01 -5.37330210e-01 -2.62057245e-01
-9.31252897e-01 4.91165221e-01 7.50286058e-02 4.10401285e-01
-9.48014200e-01 5.84789813e-01 1.78648412e-01 -3.95657867e-01
3.55245352e-01 -4.75511789e-01 3.58166873e-01 -2.58638263e-01
5.05903006e-01 5.59154034e-01 -5.73854566e-01 -7.31146485e-02
-2.23029524e-01 6.38818860e-01 3.81923705e-01 -2.88368285e-01
1.33327413e+00 -4.44015861e-02 7.32765496e-02 4.10451174e-01
1.14186430e+00 5.61919622e-02 -1.45867383e+00 1.43091559e-01
4.05388810e-02 -6.96824312e-01 -4.89075840e-01 -1.13783669e+00
-7.03076422e-01 1.03092146e+00 1.53555781e-01 4.66954350e-01
5.02156079e-01 4.54249159e-02 5.42370260e-01 3.76535803e-01
9.53232795e-02 -5.39493442e-01 2.14849025e-01 1.96130663e-01
1.46834922e+00 -9.79892850e-01 2.21389890e-01 -1.41726702e-01
-6.28474712e-01 1.48209989e+00 2.83431828e-01 -1.96444079e-01
5.09819686e-01 3.78045529e-01 -5.87172136e-02 -5.40478766e-01
-4.54801500e-01 2.22507030e-01 5.53293586e-01 6.16149724e-01
4.59483713e-01 3.06176573e-01 1.69213161e-01 4.02123451e-01
-5.93957901e-01 4.76631112e-02 2.54029542e-01 5.81950128e-01
-2.41094142e-01 -1.06664956e+00 -1.29250333e-01 4.18776184e-01
2.97027200e-01 -4.27013300e-02 -1.06495285e+00 1.13179946e+00
-1.40959591e-01 4.76234317e-01 2.10460141e-01 -3.45916510e-01
4.18825209e-01 9.90688801e-02 6.45904839e-01 -6.20577335e-02
-3.92263502e-01 -7.31479943e-01 4.60555628e-02 -6.65430725e-01
1.50441080e-01 -3.86098623e-01 -8.70903552e-01 -9.26805139e-01
2.62655228e-01 -1.86307177e-01 7.89472461e-01 5.17917693e-01
5.27744055e-01 5.98346069e-02 3.94161940e-01 -1.25919092e+00
-1.93808049e-01 -5.97329855e-01 -5.24937391e-01 6.58274710e-01
3.66854161e-01 -3.78805339e-01 -3.80222708e-01 -6.86729327e-02] | [11.704503059387207, 0.369618684053421] |
822760e6-fa58-48b4-8e32-3bb2be2b4e8f | fine-mixing-mitigating-backdoors-in-fine | 2210.09545 | null | https://arxiv.org/abs/2210.09545v1 | https://arxiv.org/pdf/2210.09545v1.pdf | Fine-mixing: Mitigating Backdoors in Fine-tuned Language Models | Deep Neural Networks (DNNs) are known to be vulnerable to backdoor attacks. In Natural Language Processing (NLP), DNNs are often backdoored during the fine-tuning process of a large-scale Pre-trained Language Model (PLM) with poisoned samples. Although the clean weights of PLMs are readily available, existing methods have ignored this information in defending NLP models against backdoor attacks. In this work, we take the first step to exploit the pre-trained (unfine-tuned) weights to mitigate backdoors in fine-tuned language models. Specifically, we leverage the clean pre-trained weights via two complementary techniques: (1) a two-step Fine-mixing technique, which first mixes the backdoored weights (fine-tuned on poisoned data) with the pre-trained weights, then fine-tunes the mixed weights on a small subset of clean data; (2) an Embedding Purification (E-PUR) technique, which mitigates potential backdoors existing in the word embeddings. We compare Fine-mixing with typical backdoor mitigation methods on three single-sentence sentiment classification tasks and two sentence-pair classification tasks and show that it outperforms the baselines by a considerable margin in all scenarios. We also show that our E-PUR method can benefit existing mitigation methods. Our work establishes a simple but strong baseline defense for secure fine-tuned NLP models against backdoor attacks. | ['Xu sun', 'Chenguang Wang', 'Xingjun Ma', 'Lingjuan Lyu', 'Zhiyuan Zhang'] | 2022-10-18 | null | null | null | null | ['sentence-pair-classification'] | ['natural-language-processing'] | [ 7.29490370e-02 -7.59707540e-02 -3.58026326e-02 -1.55828670e-01
-1.08227015e+00 -1.36268926e+00 7.20864832e-01 3.76392573e-01
-7.23622262e-01 4.86165822e-01 4.56568718e-01 -8.34947705e-01
2.78976798e-01 -8.82978201e-01 -1.05777335e+00 -8.15557420e-01
1.05391666e-01 -1.37332782e-01 2.18723565e-01 -4.43252981e-01
2.84569949e-01 6.67177439e-01 -7.54570603e-01 4.05390590e-01
5.03020108e-01 4.75766718e-01 -2.98810959e-01 7.75863409e-01
-2.08081961e-01 4.14789528e-01 -7.71434724e-01 -8.13883424e-01
6.74555063e-01 1.59416512e-01 -5.68393528e-01 -7.43455589e-01
5.04781306e-01 -4.79413152e-01 -5.16436815e-01 1.43578410e+00
4.65133518e-01 -4.39872071e-02 3.23696733e-01 -1.20678985e+00
-7.79909492e-01 9.44458187e-01 -6.65176570e-01 3.26251328e-01
-1.40879437e-01 3.29972804e-01 9.19813037e-01 -9.34450507e-01
2.40828544e-01 1.39453793e+00 7.12837219e-01 9.78583097e-01
-1.37803888e+00 -1.18565559e+00 -5.19758947e-02 -3.99433464e-01
-1.21638763e+00 -5.19278646e-01 6.18457794e-01 -3.88984561e-01
1.14523351e+00 1.13305084e-01 -5.33736981e-02 1.57581472e+00
6.42108798e-01 5.79954386e-01 8.20051014e-01 -4.35216457e-01
3.92291814e-01 4.64000434e-01 6.34267211e-01 4.75274861e-01
5.42964280e-01 3.36694628e-01 -6.15417361e-01 -1.05395401e+00
5.90532348e-02 2.69457363e-02 -2.39194885e-01 -8.93208012e-02
-6.40575528e-01 1.30661964e+00 2.03897089e-01 3.16622108e-01
-9.00478512e-02 4.70438339e-02 6.90933764e-01 3.17740738e-01
4.23344821e-01 7.94470310e-01 -8.44139397e-01 2.23967582e-01
-8.55177641e-01 2.28270397e-01 9.89817262e-01 5.21510541e-01
6.50254488e-01 -4.11805511e-02 -2.96338081e-01 6.11052513e-01
3.47425312e-01 4.60279346e-01 5.42433321e-01 -2.27138788e-01
9.61916327e-01 3.67002375e-02 -1.27853259e-01 -8.98795724e-01
3.62186357e-02 -2.12782264e-01 -6.96940601e-01 9.70626697e-02
2.78730601e-01 -6.89728796e-01 -9.10598934e-01 2.13650489e+00
3.05333555e-01 7.13237077e-02 2.97568977e-01 3.49209964e-01
2.89162636e-01 9.11609709e-01 2.06877455e-01 2.26899996e-01
1.45538163e+00 -9.98413920e-01 -4.34862763e-01 -3.74379486e-01
8.37953031e-01 -5.62437654e-01 1.20496559e+00 3.97978365e-01
-7.35570252e-01 -1.66216493e-01 -1.56473124e+00 -1.07977100e-01
-7.75712967e-01 -4.88206357e-01 4.37134206e-01 1.35277498e+00
-8.86781633e-01 7.50270844e-01 -7.99112558e-01 6.98135570e-02
6.75319612e-01 4.75536376e-01 -6.43513322e-01 -5.16061559e-02
-1.73816812e+00 7.10942864e-01 2.47290760e-01 -4.28984240e-02
-1.15119052e+00 -1.12449849e+00 -8.31798851e-01 1.58716902e-01
1.16733462e-01 -5.45544267e-01 1.03367817e+00 -1.51622847e-01
-1.41136563e+00 6.19325817e-01 1.20728210e-01 -9.23536003e-01
2.64216632e-01 -5.94216585e-01 -3.15195858e-01 1.54256895e-01
-7.53072575e-02 4.23239708e-01 1.31441033e+00 -1.06534851e+00
-2.65431404e-01 -2.29658067e-01 2.57921904e-01 -2.85044849e-01
-1.13891232e+00 2.34006196e-01 1.28088325e-01 -7.59879768e-01
-6.85613275e-01 -6.22500360e-01 -3.13255966e-01 -3.37108135e-01
-8.91187429e-01 7.51417801e-02 8.85376573e-01 -3.45265210e-01
1.31735051e+00 -2.32671094e+00 -2.64274687e-01 2.19984740e-01
3.33171934e-01 1.13343656e+00 -6.29857242e-01 7.25210130e-01
-2.95076311e-01 6.35418892e-01 -1.90600023e-01 -7.92090237e-01
2.43931755e-01 2.32467726e-01 -1.24135995e+00 4.35170531e-01
2.67947733e-01 8.20179641e-01 -8.68262827e-01 1.88179165e-02
9.44692045e-02 6.01499498e-01 -7.24650919e-01 6.04127301e-03
-1.18151337e-01 -2.83590168e-01 -2.89389044e-01 2.95521200e-01
1.00922477e+00 3.39909554e-01 -1.98835522e-01 -1.76174909e-01
2.68715382e-01 5.79255283e-01 -9.24325287e-01 1.13279879e+00
-3.58240455e-01 3.83125424e-01 1.10892832e-01 -4.26864982e-01
7.04524279e-01 2.44273916e-01 -1.90812603e-01 -8.46989527e-02
3.47831190e-01 9.70015302e-02 -1.03643179e-01 -1.64545551e-01
4.56812829e-01 -4.04860258e-01 -3.49943101e-01 7.69671559e-01
2.92814314e-01 -2.25370880e-02 -6.31786659e-02 7.15206444e-01
1.56854844e+00 -6.30974591e-01 5.37018031e-02 -5.41109145e-01
6.19261086e-01 -3.90734464e-01 4.92484391e-01 1.07062531e+00
-4.45579827e-01 3.20108086e-01 6.52357936e-01 -2.46971577e-01
-9.31918681e-01 -1.18188167e+00 -7.08064064e-02 1.01946867e+00
-3.73959810e-01 -7.84428716e-01 -1.10646999e+00 -1.04949534e+00
2.79996932e-01 8.33508968e-01 -7.35240340e-01 -7.71552980e-01
-3.42273623e-01 -7.85254240e-01 1.49308085e+00 5.72851658e-01
2.75515079e-01 -8.77769768e-01 1.70636754e-02 8.69417191e-02
3.11331242e-01 -8.92860651e-01 -9.82616961e-01 7.34321296e-01
-5.91137946e-01 -7.79458880e-01 -2.60877907e-01 -4.68947679e-01
6.03554070e-01 3.07933539e-01 6.65303051e-01 5.13105541e-02
-1.14028826e-01 6.17833866e-04 -1.51760399e-01 -5.44442773e-01
-6.45007670e-01 2.20165864e-01 6.54135585e-01 1.09363049e-01
8.90576661e-01 -5.50000727e-01 -7.77885988e-02 -1.57530442e-01
-1.35064769e+00 -7.29286611e-01 3.86492610e-01 8.83593142e-01
2.68043846e-01 3.11781168e-01 4.85808402e-01 -1.19350064e+00
1.16835523e+00 -4.29926842e-01 -6.04687989e-01 1.86337397e-01
-1.25268519e-01 4.16098297e-01 1.01792431e+00 -9.15628493e-01
-5.00466049e-01 -1.75173849e-01 -5.98867297e-01 -7.44889796e-01
-3.73754501e-02 1.63524225e-01 -5.43290555e-01 -3.00415337e-01
9.22817647e-01 7.99852423e-03 -3.99911344e-01 -5.47321379e-01
6.98503733e-01 5.49687505e-01 4.90994513e-01 -7.09511220e-01
1.67325723e+00 3.27866673e-01 -2.41886854e-01 -8.52961302e-01
-1.01314104e+00 -1.63609117e-01 -1.82038188e-01 7.24086523e-01
7.68454790e-01 -7.05304146e-01 -4.16974634e-01 7.20707297e-01
-1.25725961e+00 -3.39687139e-01 -2.30769277e-01 2.31602088e-01
2.41703257e-01 7.06882477e-01 -1.13615024e+00 -7.46087313e-01
-8.05533886e-01 -1.05068815e+00 7.83918262e-01 1.37437820e-01
-1.80188596e-01 -1.06581616e+00 3.26079726e-01 2.00030521e-01
3.83477718e-01 3.50092947e-02 1.19394541e+00 -1.27916276e+00
-1.71226472e-01 -4.13359314e-01 1.13688946e-01 9.92800117e-01
7.00502023e-02 7.96004012e-03 -1.31497455e+00 -5.35229027e-01
6.33906662e-01 -4.21218187e-01 1.11243415e+00 -3.31971236e-02
8.76469731e-01 -6.02963805e-01 -2.42094159e-01 7.33337343e-01
1.29136515e+00 -2.39521191e-01 5.69402277e-01 3.78769606e-01
7.34547317e-01 3.13448220e-01 5.66274412e-02 1.44051522e-01
-2.16354311e-01 -2.89830519e-03 3.19224983e-01 2.02963263e-01
4.68989849e-01 -7.95161784e-01 7.68926799e-01 7.21837938e-01
9.66877341e-01 -3.18706810e-01 -6.31319284e-01 5.02458870e-01
-1.39215279e+00 -7.61490762e-01 2.06634447e-01 2.05062604e+00
1.15985405e+00 7.00844288e-01 -2.61992335e-01 2.47871533e-01
6.36753380e-01 4.56626654e-01 -4.37836677e-01 -9.51168776e-01
-1.75293714e-01 6.67631686e-01 8.94228160e-01 8.60962629e-01
-1.41108739e+00 1.18687832e+00 5.73267508e+00 1.15785944e+00
-9.93363082e-01 3.24918598e-01 3.22771907e-01 -3.21876913e-01
-4.97449547e-01 -3.67097296e-02 -1.38676739e+00 3.49477053e-01
1.14467263e+00 -1.37374908e-01 4.08738166e-01 8.87824595e-01
-2.59244353e-01 4.64952171e-01 -1.16910541e+00 5.96581459e-01
3.61129045e-02 -1.48951328e+00 2.59679407e-01 3.51477295e-01
5.87590992e-01 2.57108748e-01 2.51473933e-01 4.65194762e-01
6.74161077e-01 -8.80068183e-01 4.62499678e-01 -1.11639120e-01
4.99999583e-01 -1.16045737e+00 5.51509023e-01 5.78165829e-01
-8.34334373e-01 -2.67209351e-01 -6.67919636e-01 1.73967600e-01
1.48523286e-01 9.83193994e-01 -6.00006342e-01 1.58914626e-01
5.76160192e-01 1.90024842e-02 -4.10628915e-01 3.90724897e-01
-5.97230136e-01 9.03384864e-01 -5.39972544e-01 -3.44195552e-02
5.78260183e-01 1.21290766e-01 6.89984143e-01 1.36916983e+00
-2.34814525e-01 -1.60653353e-01 -3.34793389e-01 1.00648367e+00
-5.16901195e-01 -2.54242003e-01 -8.53247702e-01 -5.02728760e-01
6.83322668e-01 1.45082140e+00 -3.05129290e-01 -1.77347913e-01
-2.31539190e-01 8.00483584e-01 4.85189199e-01 2.41526484e-01
-5.99334896e-01 -8.66780519e-01 1.26588440e+00 -2.31442109e-01
3.96724910e-01 -2.76794702e-01 -2.33115628e-01 -1.24061346e+00
-7.38478601e-02 -1.06425393e+00 2.51852572e-01 -3.23113054e-01
-1.85922539e+00 7.20199168e-01 -2.75831133e-01 -5.97101629e-01
1.28011853e-01 -7.40656734e-01 -8.20329487e-01 1.18749189e+00
-1.48957646e+00 -9.08809006e-01 7.36533582e-01 8.38359952e-01
1.67645499e-01 -2.98790932e-01 9.75996614e-01 1.97495341e-01
-8.28324854e-01 1.25309587e+00 1.87601477e-01 6.58204377e-01
7.37806380e-01 -1.06805539e+00 9.75864232e-01 1.44770575e+00
3.91683519e-01 1.53733885e+00 5.86464763e-01 -7.07740724e-01
-1.41772687e+00 -1.20795965e+00 1.18818831e+00 -7.70225406e-01
1.07623374e+00 -1.18555450e+00 -1.13356328e+00 6.82402432e-01
1.67793795e-01 -1.13954525e-02 1.06510627e+00 -8.89937058e-02
-1.12122452e+00 -5.32091074e-02 -1.38289273e+00 7.28502154e-01
3.72971088e-01 -1.04416358e+00 -9.86063182e-01 4.40711588e-01
1.32681465e+00 -1.19204529e-01 -4.50605601e-01 -4.05127555e-02
3.59149426e-01 -4.36182261e-01 1.10903037e+00 -1.06336617e+00
2.67944753e-01 -7.22370669e-02 -3.85296255e-01 -1.32268584e+00
-8.35525170e-02 -1.13528800e+00 -2.65856981e-01 1.60852325e+00
4.08764958e-01 -1.05732071e+00 7.27680206e-01 7.46738315e-01
1.31879836e-01 -5.70781827e-01 -7.90250719e-01 -8.73067856e-01
8.09004426e-01 -4.79274780e-01 7.10670888e-01 9.68319833e-01
-7.31451660e-02 3.40040296e-01 -4.12889093e-01 6.17760837e-01
8.00302088e-01 -6.78091824e-01 7.29503155e-01 -7.48501122e-01
-4.40750629e-01 -1.50197297e-01 1.09562045e-02 -7.81092823e-01
4.04300332e-01 -8.35846484e-01 -5.26251756e-02 -6.82787359e-01
2.89260019e-02 2.00427021e-03 -6.06777966e-01 7.44765043e-01
-3.51734489e-01 1.86810717e-01 2.46201754e-01 -1.24835931e-01
2.71559078e-02 4.31253165e-01 4.86309826e-01 -1.53648615e-01
-1.64605394e-01 -1.28731400e-01 -1.34295595e+00 7.06886590e-01
8.59254360e-01 -1.15360641e+00 -4.60520178e-01 -4.53080505e-01
2.41243958e-01 -7.58495033e-01 2.44651094e-01 -7.99264669e-01
3.09695423e-01 1.04869463e-01 -8.80846828e-02 -4.23109293e-01
1.35380000e-01 -5.84942102e-01 -5.54421723e-01 5.77614486e-01
-3.69900972e-01 -7.71453455e-02 5.98951995e-01 6.02446675e-01
1.34822637e-01 -4.92203236e-01 8.77756536e-01 -1.97729492e-03
-4.95278500e-02 3.56836587e-01 -3.87273073e-01 2.68768221e-01
6.73380315e-01 2.42166817e-01 -7.06753075e-01 9.28676948e-02
-2.51935154e-01 1.82210594e-01 1.94816560e-01 5.40743172e-01
4.54418331e-01 -1.13757479e+00 -4.11682606e-01 6.40306354e-01
-1.42287701e-01 -2.96476454e-01 2.31850207e-01 2.79890209e-01
-2.73369133e-01 4.62417126e-01 1.46014929e-01 2.64755506e-02
-1.20919061e+00 1.13639045e+00 2.73424774e-01 -5.92208385e-01
-4.55915093e-01 1.39427221e+00 4.44315612e-01 -6.64052963e-01
3.26981962e-01 -3.28194469e-01 2.37372547e-01 -7.73816332e-02
9.96913493e-01 -5.72797991e-02 2.22288072e-01 -3.55958134e-01
-5.96793890e-01 3.12884361e-01 -6.06464684e-01 -2.49968112e-01
1.30582476e+00 2.56645113e-01 -1.12962641e-01 1.27650917e-01
1.65973723e+00 5.82272470e-01 -9.83877540e-01 -4.51451957e-01
-1.53192788e-01 -2.85549164e-01 8.77383575e-02 -5.79992652e-01
-8.63431275e-01 1.26617551e+00 6.27764985e-02 1.75095692e-01
8.16109478e-01 -3.61468047e-01 1.39190006e+00 6.81698442e-01
1.78630546e-01 -8.34588647e-01 8.00044239e-02 6.72443867e-01
4.26028073e-01 -5.88255823e-01 -3.30393165e-01 -1.47317842e-01
-4.67963219e-01 8.62022161e-01 3.55412692e-01 -5.08111358e-01
9.51265275e-01 8.18726480e-01 2.49784753e-01 1.35146961e-01
-8.68720412e-01 6.25268936e-01 -8.36206302e-02 5.62792778e-01
-2.90927529e-01 -2.10858881e-01 7.78440712e-03 1.18691885e+00
-3.90281796e-01 -3.30711663e-01 4.82808620e-01 1.02245188e+00
-4.15084630e-01 -1.51826608e+00 -5.29984474e-01 3.17243874e-01
-8.31987381e-01 -6.51771247e-01 -5.19628286e-01 3.85620117e-01
3.34936641e-02 1.07580614e+00 -4.73847896e-01 -8.18130553e-01
2.77398676e-01 3.16848457e-01 3.76438699e-03 -8.57223272e-01
-1.20208406e+00 -3.35870922e-01 -1.77249327e-01 -5.24918795e-01
3.60091180e-01 -3.62107784e-01 -9.05452907e-01 -6.35805607e-01
-5.44252574e-01 2.48679206e-01 5.94694614e-01 9.99403536e-01
4.59388644e-01 1.98794812e-01 7.42711067e-01 -8.00879419e-01
-1.45665014e+00 -7.23482251e-01 -7.24137962e-01 2.41690680e-01
7.07831025e-01 -2.85681456e-01 -1.01024997e+00 -3.59830379e-01] | [6.105504989624023, 7.7920355796813965] |
f55f1113-c922-4705-a5cd-b761c2922a93 | continual-learning-for-anomaly-detection-in | 2004.07941 | null | https://arxiv.org/abs/2004.07941v1 | https://arxiv.org/pdf/2004.07941v1.pdf | Continual Learning for Anomaly Detection in Surveillance Videos | Anomaly detection in surveillance videos has been recently gaining attention. A challenging aspect of high-dimensional applications such as video surveillance is continual learning. While current state-of-the-art deep learning approaches perform well on existing public datasets, they fail to work in a continual learning framework due to computational and storage issues. Furthermore, online decision making is an important but mostly neglected factor in this domain. Motivated by these research gaps, we propose an online anomaly detection method for surveillance videos using transfer learning and continual learning, which in turn significantly reduces the training complexity and provides a mechanism for continually learning from recent data without suffering from catastrophic forgetting. Our proposed algorithm leverages the feature extraction power of neural network-based models for transfer learning, and the continual learning capability of statistical detection methods. | ['Yasin Yilmaz', 'Keval Doshi'] | 2020-04-15 | null | null | null | null | ['anomaly-detection-in-surveillance-videos', 'anomaly-detection-in-surveillance-videos'] | ['computer-vision', 'methodology'] | [ 9.57111493e-02 -2.47334376e-01 -2.06239909e-01 -1.75713792e-01
-4.96944219e-01 -2.36376122e-01 4.00813043e-01 1.01382129e-01
-3.78687143e-01 5.07745326e-01 -1.92558050e-01 -5.02239347e-01
-1.09324411e-01 -6.06943905e-01 -9.82720792e-01 -6.36333704e-01
-5.14096379e-01 2.01471169e-02 5.62652290e-01 9.92516428e-02
-8.51703286e-02 3.82233858e-01 -1.69566917e+00 3.15407604e-01
8.93052101e-01 1.37147117e+00 -5.92358828e-01 6.74387515e-01
9.20621827e-02 1.07383013e+00 -3.66346836e-01 -5.09689033e-01
4.35561180e-01 -2.56050289e-01 -3.17931384e-01 2.80251473e-01
7.71120787e-01 -1.07224786e+00 -7.08158970e-01 1.03124321e+00
1.85985774e-01 -2.32506637e-02 6.01160824e-01 -1.54520488e+00
-5.05726993e-01 -1.97260782e-01 -6.50132954e-01 7.29275703e-01
1.51226416e-01 3.96554470e-01 6.12863004e-01 -9.66094971e-01
1.27423733e-01 8.87864470e-01 1.00395966e+00 6.85423434e-01
-7.69656897e-01 -6.91577911e-01 4.39241260e-01 5.28867841e-01
-9.65615392e-01 -4.56776172e-01 7.31059194e-01 -6.36033833e-01
9.16429579e-01 -1.51553959e-01 1.11870503e+00 1.32393599e+00
5.64439476e-01 1.17200041e+00 4.96970326e-01 -1.53596580e-01
3.71838480e-01 -2.02865288e-01 -6.09702580e-02 1.06690168e+00
5.43144345e-01 3.19908410e-01 -5.44466138e-01 -4.41835284e-01
5.38753331e-01 8.97832930e-01 6.96131438e-02 -5.05314231e-01
-5.79932928e-01 8.26362729e-01 8.80332291e-02 3.45107429e-02
-6.81501746e-01 7.68438652e-02 7.82822669e-01 8.11316907e-01
6.80156469e-01 4.09230739e-02 -6.40485466e-01 -3.05342555e-01
-9.16881561e-01 6.23551346e-02 6.06668055e-01 7.54183888e-01
5.05509257e-01 4.20641184e-01 -3.87548916e-02 2.51651347e-01
1.01494171e-01 4.90872860e-01 4.66255933e-01 -7.34388351e-01
4.39311750e-02 8.03711772e-01 -5.19487970e-02 -1.11071086e+00
-3.45277220e-01 -1.95833743e-01 -9.95475709e-01 2.87646174e-01
3.50007325e-01 -4.11590993e-01 -8.45188618e-01 1.47908354e+00
4.78014171e-01 7.06736803e-01 -1.73347313e-02 4.42879140e-01
2.74657339e-01 6.27173364e-01 1.94478258e-01 -3.72854948e-01
8.81415069e-01 -8.90048981e-01 -6.75264180e-01 -2.68945515e-01
6.47534490e-01 -2.86151826e-01 6.36170268e-01 7.05113709e-01
-6.61596477e-01 -3.67619574e-01 -9.52187121e-01 4.36985284e-01
-2.21732855e-01 -7.72437900e-02 8.97543192e-01 5.48146725e-01
-9.38105702e-01 3.12173456e-01 -1.18728364e+00 -5.12198687e-01
9.14166749e-01 2.05489665e-01 -3.18355978e-01 -3.69217217e-01
-7.88866699e-01 6.49203241e-01 2.24364340e-01 -2.91289035e-02
-1.17691100e+00 -7.07979500e-01 -7.55336761e-01 -1.25334248e-01
8.38581145e-01 -5.58687031e-01 1.23976660e+00 -1.24737513e+00
-1.18902206e+00 5.17555952e-01 -4.63792384e-02 -9.51021373e-01
6.78443789e-01 -8.40649247e-01 -7.25412667e-01 2.10604668e-01
-1.62882477e-01 3.81177783e-01 1.48192251e+00 -6.38267994e-01
-1.02077532e+00 -4.70973432e-01 -1.17006667e-01 -1.74426064e-01
-9.23461735e-01 -1.70627221e-01 -1.12757988e-01 -7.69769013e-01
-1.03447407e-01 -7.22724974e-01 -8.74219611e-02 6.67637348e-01
2.05614790e-01 -3.12609553e-01 1.59146369e+00 -7.68379509e-01
1.19094718e+00 -2.30301666e+00 -2.01613337e-01 -2.01633405e-02
3.13797742e-01 7.41078556e-01 -2.17697490e-02 1.27200767e-01
1.66648701e-01 -4.04317379e-01 -1.60425141e-01 -1.02186449e-01
-5.18433750e-01 2.17746660e-01 -3.93332005e-01 6.08909011e-01
6.37352407e-01 8.05958271e-01 -1.03517783e+00 -3.49924445e-01
1.98221490e-01 1.48975864e-01 -8.43526721e-01 4.90890890e-01
-1.66929677e-01 2.73754865e-01 -4.58101362e-01 1.08438230e+00
4.19480115e-01 -2.23511547e-01 -1.92166418e-01 1.95613265e-01
-3.25070918e-02 -3.45775157e-01 -6.21981323e-01 1.51134431e+00
2.17352062e-01 5.91704309e-01 -5.74197248e-02 -1.45051968e+00
6.41218603e-01 5.15026510e-01 8.88528764e-01 -6.81022406e-01
-6.44053221e-02 1.93894312e-01 -7.08490089e-02 -8.96925092e-01
1.06334187e-01 2.50427544e-01 1.75745726e-01 2.88875788e-01
2.12961942e-01 5.07609844e-01 2.83944216e-02 8.74231607e-02
1.47419441e+00 2.32921448e-02 4.46086586e-01 7.00141415e-02
3.15106899e-01 1.01928025e-01 9.08405781e-01 8.75153422e-01
-7.31795609e-01 7.54692256e-02 4.27629262e-01 -1.27306294e+00
-1.07747030e+00 -8.36303890e-01 1.23599984e-01 1.31272542e+00
-2.91427433e-01 -5.37924618e-02 -6.26498699e-01 -1.08301651e+00
2.31018350e-01 2.60403216e-01 -7.84879863e-01 -6.83588743e-01
-6.42572701e-01 -8.70234251e-01 4.63651389e-01 6.07846379e-01
5.42908549e-01 -9.94530797e-01 -9.36593175e-01 3.20034713e-01
2.57483602e-01 -9.84515846e-01 -1.79755688e-01 -1.94871366e-01
-1.09606647e+00 -1.51098442e+00 -5.53912520e-01 -5.41527629e-01
6.32390082e-01 6.70698166e-01 1.01670170e+00 1.29257455e-01
-6.80376887e-01 1.05160153e+00 -4.24831212e-01 -8.53896618e-01
-3.60831171e-01 -1.05597951e-01 5.00664771e-01 3.26761514e-01
8.68884325e-01 -4.38396096e-01 -6.91492379e-01 5.79668358e-02
-1.07123923e+00 -5.19219458e-01 8.50951970e-01 1.13770294e+00
3.04938704e-01 -1.55658629e-02 9.26530838e-01 -6.21496141e-01
3.50611329e-01 -9.03227806e-01 -9.57252264e-01 2.61854649e-01
-5.51813424e-01 -3.25315654e-01 5.97296059e-01 -6.18199766e-01
-7.80957997e-01 4.74133641e-02 1.70175597e-01 -8.47649455e-01
-1.90381691e-01 1.16752326e-01 2.47731894e-01 -1.09988481e-01
6.39556825e-01 3.89250726e-01 4.20444846e-01 -1.14038385e-01
-1.18076749e-01 3.46419781e-01 3.77320886e-01 -1.39408603e-01
7.93992043e-01 6.22572362e-01 -5.69355395e-03 -9.55283105e-01
-1.08740366e+00 -4.13571328e-01 -5.72090030e-01 -4.06332135e-01
7.67062664e-01 -1.09068799e+00 -5.92514038e-01 9.08704460e-01
-1.04952610e+00 -3.99755426e-02 -3.90910029e-01 4.55889970e-01
-2.55800217e-01 5.08136153e-01 -4.46994245e-01 -1.06192994e+00
-5.57048619e-01 -6.44811749e-01 7.57131636e-01 7.47107714e-02
1.73160106e-01 -9.45025921e-01 6.52200682e-03 3.76250595e-02
6.74943984e-01 4.36872691e-01 6.05826914e-01 -7.10525930e-01
-7.10754156e-01 -7.50582337e-01 -1.17285281e-01 5.98932624e-01
2.83110946e-01 6.74210722e-03 -9.33764160e-01 -7.65439510e-01
3.24120149e-02 -5.31771243e-01 1.03041255e+00 4.31262612e-01
1.36622739e+00 -5.96490204e-01 -3.59070003e-01 5.46916783e-01
1.02412856e+00 3.66542250e-01 3.72713506e-01 3.02815169e-01
6.09038949e-01 2.78692305e-01 6.75565064e-01 7.78270125e-01
1.86626673e-01 6.96877316e-02 6.13696516e-01 -1.70704693e-01
2.39383101e-01 -1.71784684e-01 5.66132903e-01 4.63913262e-01
4.85038245e-03 -1.07919322e-02 -9.77410853e-01 5.26352704e-01
-2.17673349e+00 -1.24491227e+00 4.03120607e-01 2.16078973e+00
4.25791353e-01 2.16312766e-01 2.48270974e-01 6.05393760e-02
4.13755089e-01 -1.56662434e-01 -1.06745005e+00 -5.17563289e-03
-3.69069539e-02 -3.70269477e-01 3.16714734e-01 -1.96781918e-01
-1.76186836e+00 8.00021410e-01 6.89407730e+00 5.53608656e-01
-1.15416992e+00 8.04215074e-02 7.03922331e-01 -2.40556464e-01
4.74652022e-01 -4.69918579e-01 -6.71682954e-01 5.53511322e-01
9.10775900e-01 2.17379909e-02 1.88124806e-01 1.29782581e+00
-2.05745846e-02 -2.83617768e-02 -1.19900393e+00 1.07290280e+00
3.31212491e-01 -1.21499145e+00 3.61386240e-01 -7.60909319e-02
7.58524001e-01 1.56189397e-01 2.67254293e-01 4.70847696e-01
-2.51340270e-02 -7.16535687e-01 2.36535117e-01 5.53191781e-01
5.30673385e-01 -7.33802736e-01 9.18067873e-01 4.14716721e-01
-8.53883564e-01 -8.07685614e-01 -4.33868974e-01 -1.29345194e-01
-6.61259145e-02 6.92326307e-01 -5.58521748e-01 1.61546499e-01
1.18311048e+00 9.42339778e-01 -6.23209238e-01 1.37055480e+00
2.41914541e-01 9.91990566e-01 -1.97552934e-01 2.76019543e-01
3.73624206e-01 -4.91650961e-02 7.00529814e-01 1.07227898e+00
5.32943308e-01 -1.09122768e-01 4.95077491e-01 2.05287725e-01
-2.06558276e-02 -1.49182886e-01 -1.24510252e+00 -1.13569416e-01
2.85316557e-01 8.02832961e-01 -1.65361434e-01 -4.20049340e-01
-8.93074930e-01 9.33080137e-01 2.44846091e-01 1.79218352e-01
-7.39736855e-01 3.34216170e-02 7.97240198e-01 1.59411311e-01
4.46497768e-01 -2.71763951e-01 2.12468669e-01 -1.30310380e+00
3.61657709e-01 -1.03234410e+00 7.97389984e-01 1.62689641e-01
-1.74310470e+00 3.75696510e-01 -1.61093161e-01 -1.57881689e+00
-4.14067239e-01 -7.73719311e-01 -7.88026392e-01 -1.31270081e-01
-1.69642758e+00 -1.08831167e+00 -3.57295305e-01 9.30223167e-01
8.77740443e-01 -8.31160247e-01 9.01255906e-01 4.98830855e-01
-6.94381356e-01 7.70510077e-01 3.08010846e-01 2.67862111e-01
7.05377638e-01 -8.27067673e-01 3.62381607e-01 1.11611998e+00
-1.21848539e-01 2.29604825e-01 3.50875765e-01 -6.20156050e-01
-1.74874818e+00 -1.39766002e+00 1.14246249e-01 -4.23504084e-01
8.26753616e-01 -2.44227484e-01 -1.25446033e+00 7.35698462e-01
-1.58947200e-01 5.29869497e-01 8.53666425e-01 3.04234046e-02
-5.02543688e-01 -3.95531029e-01 -1.19092727e+00 3.63611847e-01
1.07363474e+00 -2.82543391e-01 -4.90392357e-01 4.30737644e-01
5.54162085e-01 -1.12485990e-01 -2.64161676e-01 5.66529214e-01
5.34740031e-01 -9.40762937e-01 7.66044021e-01 -9.98196244e-01
2.24960923e-01 2.50136927e-02 1.87088251e-01 -9.70418215e-01
-2.66463161e-01 -7.32158899e-01 -1.00292110e+00 7.21548021e-01
1.17263272e-01 -7.89656997e-01 7.71849334e-01 6.52585089e-01
-1.11281238e-01 -8.11319649e-01 -1.23013723e+00 -1.03854692e+00
-2.36589178e-01 -3.95320475e-01 1.10995911e-01 8.41860294e-01
-4.35978472e-01 -6.94628060e-02 -1.00051439e+00 2.22227722e-01
9.10384476e-01 -2.64590591e-01 8.41193259e-01 -1.53777575e+00
-1.91396177e-01 -1.75543547e-01 -7.22034991e-01 -9.26908493e-01
-8.38801190e-02 -2.61714309e-01 -1.96557090e-01 -9.19616699e-01
1.66064501e-01 -1.96592942e-01 -6.00442827e-01 5.14916122e-01
-2.10551649e-01 1.07455947e-01 -3.07079107e-01 2.43842319e-01
-1.19371939e+00 7.55146682e-01 6.01897180e-01 -7.64475539e-02
-2.27031261e-01 1.36362463e-01 -3.18176895e-01 1.02096760e+00
6.68973327e-01 -4.15892512e-01 -2.93881446e-01 -5.28544247e-01
-1.61448531e-02 -1.90640688e-01 4.84591186e-01 -1.34876394e+00
4.43065673e-01 -5.93675934e-02 6.57665670e-01 -4.52253461e-01
2.04653218e-01 -1.12980568e+00 -5.63510299e-01 7.80344784e-01
9.34144855e-02 3.90301466e-01 4.85276490e-01 1.29730523e+00
-1.59714356e-01 2.78872043e-01 6.36232316e-01 -4.70678788e-03
-9.02834117e-01 8.75084162e-01 -6.61235452e-01 5.16015710e-03
1.47595561e+00 -2.27150202e-01 -4.35355008e-02 -3.68938893e-01
-3.97919267e-01 3.64188284e-01 1.51229799e-01 7.25886285e-01
8.82588565e-01 -1.12263417e+00 -7.49966145e-01 4.91199195e-01
2.38964155e-01 -1.06551602e-01 3.97227734e-01 8.37065160e-01
-2.53719151e-01 1.95560247e-01 -3.55136603e-01 -7.94097602e-01
-1.20895314e+00 9.38277185e-01 2.01389402e-01 -5.98599054e-02
-9.27875757e-01 5.90421379e-01 4.73878048e-02 7.89056718e-02
5.74603081e-01 1.09126521e-02 1.72232110e-02 -6.20363615e-02
8.64572465e-01 4.68287587e-01 3.24809365e-03 -2.90261973e-02
-2.55545855e-01 1.93356693e-01 -5.56214631e-01 4.63892162e-01
1.50429106e+00 2.66984165e-01 1.25367135e-01 3.45157146e-01
8.06764781e-01 -6.09137416e-01 -1.79382718e+00 -4.53415960e-01
2.41051793e-01 -5.52223444e-01 1.17922816e-02 -3.43676507e-01
-1.04945445e+00 8.73133898e-01 9.25083518e-01 2.70752102e-01
1.29683316e+00 -3.03148210e-01 1.09313858e+00 1.00330627e+00
3.83155316e-01 -1.23338151e+00 6.75192058e-01 5.68791449e-01
5.85815132e-01 -2.00396991e+00 -1.33214846e-01 7.21871257e-02
-4.11759645e-01 1.18780494e+00 8.54640722e-01 -1.97936326e-01
8.27692628e-01 2.28822887e-01 -9.19023827e-02 -2.03571916e-01
-8.87117267e-01 1.75168574e-01 3.17220747e-01 6.97578073e-01
7.48997852e-02 -3.39319021e-01 2.32528076e-01 2.02087924e-01
6.13376200e-01 2.29092479e-01 1.41862333e-01 1.26807129e+00
-5.94796479e-01 -6.14956379e-01 -1.59868106e-01 8.04410398e-01
-6.31656528e-01 6.48007765e-02 -1.90615654e-02 4.88663912e-01
7.27991238e-02 8.04139435e-01 2.61455238e-01 -2.68657267e-01
1.97962254e-01 2.51326770e-01 9.91985500e-02 -3.02308917e-01
-1.14784479e-01 -2.85266012e-01 -3.97185147e-01 -7.62751162e-01
-2.40737677e-01 -7.64676213e-01 -6.49836421e-01 -2.46982828e-01
-1.36456236e-01 -1.77326351e-01 2.87333518e-01 1.02526546e+00
6.99886322e-01 2.96299517e-01 6.56766415e-01 -5.31615198e-01
-7.95629263e-01 -7.89168715e-01 -2.07337439e-01 3.72640342e-01
8.79356265e-01 -7.28905737e-01 -2.82927752e-01 8.64884630e-02] | [7.862951755523682, 1.6220593452453613] |
467c83ce-ac9c-42c8-9e16-b9f34e8c533e | 190910363 | 1909.10363 | null | https://arxiv.org/abs/1909.10363v2 | https://arxiv.org/pdf/1909.10363v2.pdf | Shadow Transfer: Single Image Relighting For Urban Road Scenes | Illumination effects in images, specifically cast shadows and shading, have been shown to decrease the performance of deep neural networks on a large number of vision-based detection, recognition and segmentation tasks in urban driving scenes. A key factor that contributes to this performance gap is the lack of `time-of-day' diversity within real, labeled datasets. There have been impressive advances in the realm of image to image translation in transferring previously unseen visual effects into a dataset, specifically in day to night translation. However, it is not easy to constrain what visual effects, let alone illumination effects, are transferred from one dataset to another during the training process. To address this problem, we propose deep learning framework, called Shadow Transfer, that can relight complex outdoor scenes by transferring realistic shadow, shading, and other lighting effects onto a single image. The novelty of the proposed framework is that it is both self-supervised, and is designed to operate on sensor and label information that is easily available in autonomous vehicle datasets. We show the effectiveness of this method on both synthetic and real datasets, and we provide experiments that demonstrate that the proposed method produces images of higher visual quality than state of the art image to image translation methods. | ['Matthew Johnson-Roberson', 'Ram Vasudevan', 'Alexandra Carlson'] | 2019-09-23 | null | null | null | null | ['image-relighting'] | ['computer-vision'] | [ 7.25961208e-01 -1.58524126e-01 1.12910785e-01 -7.11511552e-01
-3.12309712e-01 -6.04820371e-01 7.62990832e-01 -4.49371755e-01
-4.05785531e-01 8.38310361e-01 -2.52367705e-01 -3.54156584e-01
3.26240480e-01 -7.88652003e-01 -1.08490443e+00 -7.14197993e-01
3.65449816e-01 2.54575551e-01 5.01170516e-01 -4.05422837e-01
9.67649072e-02 5.00091076e-01 -1.86427188e+00 4.58575301e-02
9.15006876e-01 8.71689260e-01 4.18744057e-01 5.23951650e-01
-1.17398605e-01 8.07244480e-01 -5.31950533e-01 -2.37706155e-01
3.49590927e-01 -4.77716297e-01 -3.03496122e-01 3.54403794e-01
7.81192720e-01 -3.72284770e-01 -3.27725351e-01 8.50674570e-01
4.31978434e-01 8.01448226e-02 7.26574004e-01 -1.55852079e+00
-6.12737179e-01 -2.74655409e-02 -6.20559394e-01 1.38411939e-01
-1.08792275e-01 3.45448017e-01 3.42376620e-01 -7.48213947e-01
5.83564341e-01 1.26229310e+00 5.29475570e-01 4.38633174e-01
-1.03679168e+00 -7.98227549e-01 -1.35668650e-01 1.43530980e-01
-1.00017369e+00 -4.40382898e-01 9.38837230e-01 -5.30304074e-01
5.35177112e-01 2.23823249e-01 5.85351765e-01 1.09996390e+00
3.32183987e-01 7.56097019e-01 1.70238471e+00 -3.33996683e-01
2.85633713e-01 4.48216975e-01 -1.53653741e-01 6.06474936e-01
1.12196334e-01 6.11249268e-01 -4.67628658e-01 4.63457078e-01
5.18721402e-01 -1.48630276e-01 -2.08902553e-01 -4.33119714e-01
-8.79097700e-01 7.75190890e-01 7.22809911e-01 -1.32623702e-01
1.03078224e-02 3.86050045e-01 2.09407032e-01 1.56444147e-01
4.07556742e-01 7.57928789e-02 -2.56676495e-01 1.77888453e-01
-8.27595830e-01 1.66146979e-01 5.25206685e-01 7.90079951e-01
1.11446464e+00 4.12798762e-01 -1.87806055e-01 7.20465720e-01
1.98202670e-01 1.19700861e+00 2.39329740e-01 -8.50869358e-01
2.73604214e-01 3.00193131e-01 2.57081926e-01 -1.04687035e+00
-3.02406192e-01 -3.70069474e-01 -6.47708595e-01 5.91307342e-01
1.91711798e-01 -2.38246545e-01 -1.29594827e+00 1.69349027e+00
4.22706842e-01 2.71721542e-01 2.13448793e-01 1.10381055e+00
1.08981097e+00 5.40322959e-01 2.97605060e-02 2.94226930e-02
9.56502438e-01 -1.26527381e+00 -7.47642756e-01 -8.66592050e-01
1.42913103e-01 -9.33277607e-01 1.08645260e+00 2.82858778e-02
-6.17063880e-01 -8.52658033e-01 -1.17936850e+00 -9.78581384e-02
-6.70258999e-01 2.39403564e-02 7.65106320e-01 8.94827366e-01
-9.08046126e-01 1.25884488e-01 -4.03261274e-01 -4.84703064e-01
4.85457063e-01 1.69779092e-01 -1.54533789e-01 -3.01247716e-01
-1.25847387e+00 1.07801759e+00 2.98187464e-01 1.95272684e-01
-1.26335192e+00 -5.90661466e-01 -7.20873356e-01 -3.31442654e-01
2.72295535e-01 -4.40727293e-01 9.20809627e-01 -1.36139631e+00
-1.48071313e+00 9.71581638e-01 -2.38333166e-01 -5.19994438e-01
6.95352912e-01 -5.81335165e-02 -3.42610002e-01 -2.37836987e-01
7.47259706e-02 1.08669770e+00 1.02135742e+00 -1.96623695e+00
-6.37376845e-01 -2.00244144e-01 6.72343969e-02 3.76504421e-01
5.59778586e-02 -2.23152310e-01 -5.48690081e-01 -1.79285005e-01
-3.33655298e-01 -1.20806181e+00 -1.39246702e-01 9.23216157e-03
-2.18726680e-01 2.44454652e-01 1.18357730e+00 -5.05441487e-01
2.57488608e-01 -2.04864979e+00 -2.29984060e-01 -1.01003475e-01
-1.83972538e-01 4.36977535e-01 -1.65552318e-01 2.78889954e-01
1.47498235e-01 -1.63458526e-01 -5.25186360e-01 -3.52227122e-01
-8.60953704e-02 7.45701969e-01 -4.79017138e-01 5.93030691e-01
1.50506958e-01 1.11503649e+00 -8.87142241e-01 -4.96753305e-01
8.47714007e-01 6.67854249e-01 -9.58877802e-02 1.88256636e-01
-3.38336378e-01 1.01858604e+00 -1.60849050e-01 4.94740784e-01
1.04269207e+00 4.07762110e-01 -3.44805598e-01 -6.83074594e-02
-2.33370259e-01 -1.75283849e-01 -8.26929271e-01 1.49567235e+00
-5.60938418e-01 1.26213169e+00 -1.70564756e-01 -8.64550948e-01
9.50971544e-01 -6.33486062e-02 1.66221723e-01 -1.10772526e+00
3.82516146e-01 1.36062309e-01 -1.27094597e-01 -6.32680535e-01
4.61680055e-01 -1.12268977e-01 -1.71350148e-02 1.79788038e-01
-3.15490663e-01 -6.84321523e-01 2.21940756e-01 -2.56710202e-01
6.20165884e-01 2.75105208e-01 -2.41209000e-01 -8.49661976e-03
5.02003431e-01 1.76549479e-01 5.58610737e-01 7.71352351e-01
-2.39934176e-01 7.61768639e-01 2.77624913e-02 -3.98003697e-01
-1.18113172e+00 -1.03409207e+00 -3.22392613e-01 1.03120065e+00
5.33518612e-01 6.01869285e-01 -8.38335156e-01 -3.92226517e-01
9.40108821e-02 8.42864752e-01 -7.00127900e-01 -8.54827836e-02
-3.51448148e-01 -8.43877375e-01 5.69419920e-01 2.77276784e-01
1.11631405e+00 -1.19908047e+00 -1.00512481e+00 2.02910090e-03
-1.86286017e-01 -1.58587432e+00 -1.84432983e-01 2.53274858e-01
-4.65149015e-01 -9.57827151e-01 -3.94794524e-01 -8.12675834e-01
7.10398734e-01 6.88125789e-01 1.07594550e+00 -1.14107259e-01
-5.29663026e-01 2.32887670e-01 -2.96645254e-01 -7.20527291e-01
-5.58344364e-01 -2.72206962e-01 -2.29569197e-01 3.94784391e-01
7.97086284e-02 -3.28966618e-01 -7.66062856e-01 4.75962490e-01
-1.23873019e+00 3.13849390e-01 8.09616089e-01 6.87250853e-01
3.60311419e-01 1.47711426e-01 2.05329672e-01 -1.05382037e+00
2.48104781e-01 -2.00879052e-01 -6.58455729e-01 1.35054663e-01
-6.13041878e-01 -5.64029887e-02 5.44054866e-01 -1.33892491e-01
-1.35523212e+00 2.77309299e-01 -2.78945900e-02 -2.69254476e-01
-5.10078847e-01 4.33902889e-02 -1.24445967e-01 -5.40872872e-01
6.54403448e-01 3.40602934e-01 -2.00313970e-01 1.05853220e-02
6.87581182e-01 7.49236822e-01 8.21349382e-01 -2.99215347e-01
1.09039259e+00 9.08176780e-01 1.52906984e-01 -8.60352635e-01
-8.57645392e-01 -3.75053823e-01 -7.54445493e-01 -5.48360586e-01
9.78915870e-01 -9.65783238e-01 -2.85885662e-01 6.61974430e-01
-1.05357540e+00 -6.23576164e-01 2.38031764e-02 1.65104285e-01
-4.65066552e-01 6.25623316e-02 -5.20793051e-02 -8.24250996e-01
-7.55905062e-02 -1.23885584e+00 1.14801097e+00 5.70671439e-01
3.46522033e-01 -9.28937674e-01 -1.56404391e-01 7.21742094e-01
4.56737578e-01 6.63366079e-01 6.46359026e-01 2.14869589e-01
-7.67836034e-01 -5.95480762e-02 -5.84695160e-01 6.06662571e-01
3.05754811e-01 -5.25031798e-02 -1.33670557e+00 -2.61238992e-01
-8.48325193e-02 -3.30828905e-01 1.11888850e+00 2.69331455e-01
7.60846913e-01 1.26787901e-01 -3.55251491e-01 7.33534694e-01
1.60028124e+00 3.25044364e-01 8.58326018e-01 4.45723861e-01
9.70684111e-01 5.87409198e-01 8.02932918e-01 2.24120282e-02
3.71809989e-01 7.35003829e-01 7.06612706e-01 -9.03597593e-01
-7.45169699e-01 -5.83384931e-02 3.51230472e-01 1.79283097e-01
1.19070083e-01 -3.59843761e-01 -7.99429476e-01 6.55291200e-01
-1.70971107e+00 -7.22751319e-01 -3.73656243e-01 2.14691710e+00
6.52671576e-01 1.44280225e-01 -3.56360316e-01 2.86013149e-02
5.49043775e-01 1.83451071e-01 -7.20121920e-01 -5.46467543e-01
-4.17523116e-01 -8.31023135e-05 1.15025151e+00 6.28105760e-01
-1.16702342e+00 1.37812972e+00 5.75198221e+00 5.20080388e-01
-1.60565650e+00 2.66408086e-01 6.95355296e-01 3.25161576e-01
-2.80574203e-01 -7.26301298e-02 -5.46214759e-01 4.16051269e-01
7.90913165e-01 2.09956467e-01 6.16205692e-01 6.68553591e-01
5.86394966e-01 -3.87883306e-01 -9.38759983e-01 8.61854672e-01
5.26116133e-01 -9.15754378e-01 -9.63997766e-02 -1.21901251e-01
1.23614728e+00 2.64274538e-01 4.33598220e-01 1.63320631e-01
3.38935316e-01 -1.02737069e+00 7.08415568e-01 3.21750760e-01
6.56205952e-01 -7.72051930e-01 6.56743765e-01 1.99250370e-01
-8.79609406e-01 -2.38162652e-02 -4.27813709e-01 7.55821727e-03
1.11497626e-01 4.92968380e-01 -1.10782003e+00 4.06303436e-01
7.25837946e-01 4.95930403e-01 -8.40159774e-01 9.41929162e-01
-4.48745787e-01 5.04464209e-01 -2.17615470e-01 1.90750659e-01
3.58599186e-01 -2.87598282e-01 2.34416708e-01 1.18023932e+00
3.35590504e-02 -3.12190562e-01 2.22798586e-01 8.54993463e-01
9.83360261e-02 -1.84740409e-01 -1.01739728e+00 4.20902938e-01
2.24079415e-01 1.32139838e+00 -8.72913003e-01 -1.02791131e-01
-4.14440215e-01 1.07549953e+00 -1.29259065e-01 6.40010536e-01
-1.24374366e+00 -1.00117102e-01 4.14908499e-01 1.34821847e-01
3.76320392e-01 -2.74365634e-01 -4.80729222e-01 -6.35543048e-01
-1.13749258e-01 -6.31129920e-01 -2.79241741e-01 -1.10669518e+00
-8.52979898e-01 5.79553008e-01 -3.83979641e-02 -1.09518337e+00
-3.45116220e-02 -5.42024374e-01 -4.83839273e-01 6.61646962e-01
-2.14256763e+00 -1.57988644e+00 -8.35206926e-01 6.18894339e-01
7.37996578e-01 -7.90135469e-03 4.55276996e-01 4.17831361e-01
-3.80583078e-01 3.80871683e-01 3.45480800e-01 1.37444334e-02
8.25854182e-01 -1.10206103e+00 4.16085273e-01 1.05315638e+00
9.60996374e-02 -7.41177723e-02 9.93764341e-01 -3.84703755e-01
-1.42253363e+00 -1.58020711e+00 3.41589510e-01 -4.45969641e-01
1.65443167e-01 -4.55363840e-01 -5.39972484e-01 4.97412741e-01
4.91086721e-01 1.26074746e-01 2.04680800e-01 -4.31742907e-01
-1.19910866e-01 -5.27314901e-01 -1.17209971e+00 6.15701020e-01
8.29851389e-01 -3.08574378e-01 -3.57770622e-01 4.95623648e-01
5.32270074e-01 -6.35017753e-01 -1.62457258e-01 6.08686209e-01
3.79581332e-01 -1.19026828e+00 1.03164458e+00 4.17701229e-02
3.94105762e-01 -6.01230502e-01 -1.59837097e-01 -1.36622715e+00
1.94879994e-01 -3.88952821e-01 5.16101778e-01 1.21514487e+00
3.29134554e-01 -6.68948054e-01 7.46719241e-01 4.53781605e-01
-3.48779231e-01 -2.79539376e-01 -8.49352956e-01 -6.32075071e-01
5.68413176e-02 -4.37887758e-01 4.80284333e-01 6.76182806e-01
-1.03663015e+00 3.45081627e-01 -6.12030268e-01 3.80176604e-01
7.43536055e-01 3.21499616e-01 1.27866697e+00 -1.03692198e+00
-1.18961707e-01 -1.11297801e-01 -3.75874370e-01 -8.40110183e-01
1.47084728e-01 -6.10513508e-01 5.88641047e-01 -1.65907311e+00
1.27707869e-01 -5.66511571e-01 -8.99253041e-02 4.00796652e-01
-1.15715601e-01 7.98647881e-01 1.64398074e-01 1.34555697e-01
-3.54073495e-01 6.66641831e-01 1.58153319e+00 -3.16671640e-01
-2.13696100e-02 -2.44154796e-01 -2.89434731e-01 6.44068122e-01
7.67893195e-01 -4.59287912e-01 -6.53080702e-01 -7.25621819e-01
-1.77472472e-01 -2.48106629e-01 5.07955372e-01 -1.27526939e+00
3.72538855e-03 -3.47278774e-01 3.84745181e-01 -5.85112095e-01
5.08037448e-01 -7.85817206e-01 1.58769310e-01 3.96232784e-01
-5.53426370e-02 -3.34341764e-01 3.53789479e-01 5.42136133e-01
-1.17939278e-01 7.05802962e-02 1.01221752e+00 -1.20077334e-01
-1.12397742e+00 9.20620561e-02 -1.55028239e-01 -1.02681471e-02
1.18605387e+00 -3.47704709e-01 -3.97367448e-01 -3.54633719e-01
-4.50935506e-04 4.93762381e-02 5.87661386e-01 7.65351593e-01
5.65833032e-01 -1.15918338e+00 -6.53489411e-01 2.72733718e-01
2.30072051e-01 1.22778766e-01 1.58758342e-01 6.32372439e-01
-7.78455496e-01 4.37160611e-01 -3.86196792e-01 -9.04210806e-01
-1.28679645e+00 4.17821079e-01 3.84655863e-01 2.79532224e-01
-5.67741454e-01 6.87213838e-01 5.57865620e-01 -5.19378781e-01
5.78640290e-02 -3.28548014e-01 -4.72501665e-02 -3.24064285e-01
1.41762123e-01 -5.73730916e-02 1.60225689e-01 -9.54955101e-01
-2.10575387e-01 7.90006399e-01 1.92979246e-01 -9.11977887e-02
1.03482151e+00 -3.95453393e-01 5.63944280e-02 5.39839804e-01
1.07573521e+00 -3.37523699e-01 -1.65940487e+00 -3.98301519e-02
-5.51797748e-01 -7.56466210e-01 3.14958304e-01 -9.89491224e-01
-1.31065762e+00 9.03071463e-01 1.19468987e+00 -3.60586606e-02
1.12928927e+00 -2.95403063e-01 1.04919624e+00 2.67677367e-01
3.10178846e-01 -1.14837730e+00 2.54280530e-02 4.48949218e-01
7.49495685e-01 -1.78896642e+00 -1.66032374e-01 -4.53464270e-01
-8.26877594e-01 6.95506334e-01 6.90993309e-01 -1.15846217e-01
2.84846514e-01 2.20361009e-01 5.96221328e-01 -8.08998663e-03
-3.40025336e-01 -4.93739218e-01 1.77668750e-01 8.69370818e-01
9.40432847e-02 1.01905204e-01 -6.89046904e-02 -3.09759736e-01
-2.22626790e-01 5.70656210e-02 6.46552026e-01 7.76644468e-01
-4.32968825e-01 -9.12017524e-01 -5.89755058e-01 -2.19852552e-02
-5.90813272e-02 5.55165708e-02 -4.24309254e-01 8.36960196e-01
7.50005901e-01 1.25810242e+00 -8.17689672e-02 -1.84342802e-01
2.49253511e-01 -2.61302173e-01 5.12688637e-01 -3.65613967e-01
-1.51306182e-01 -2.09650815e-01 -4.91485409e-02 -2.82618433e-01
-5.24896443e-01 -5.32287002e-01 -1.20334542e+00 -1.92084387e-01
-1.31167203e-01 -2.34090090e-01 1.16871643e+00 1.11515617e+00
9.73610114e-03 8.07459354e-01 9.04712856e-01 -1.03002083e+00
1.06149800e-01 -8.12973797e-01 -4.44092304e-01 6.74605489e-01
5.52561939e-01 -9.86176789e-01 -3.65407050e-01 2.65591592e-01] | [8.852160453796387, -1.7563341856002808] |
9717152d-3e55-4c60-8afa-549acae716a3 | a-dataset-of-inertial-measurement-units-for | 2307.02480 | null | https://arxiv.org/abs/2307.02480v1 | https://arxiv.org/pdf/2307.02480v1.pdf | A Dataset of Inertial Measurement Units for Handwritten English Alphabets | This paper presents an end-to-end methodology for collecting datasets to recognize handwritten English alphabets by utilizing Inertial Measurement Units (IMUs) and leveraging the diversity present in the Indian writing style. The IMUs are utilized to capture the dynamic movement patterns associated with handwriting, enabling more accurate recognition of alphabets. The Indian context introduces various challenges due to the heterogeneity in writing styles across different regions and languages. By leveraging this diversity, the collected dataset and the collection system aim to achieve higher recognition accuracy. Some preliminary experimental results demonstrate the effectiveness of the dataset in accurately recognizing handwritten English alphabet in the Indian context. This research can be extended and contributes to the field of pattern recognition and offers valuable insights for developing improved systems for handwriting recognition, particularly in diverse linguistic and cultural contexts. | ['Rahul Mishra', 'Hari Prabhat Gupta'] | 2023-07-05 | null | null | null | null | ['handwriting-recognition'] | ['computer-vision'] | [ 1.98101044e-01 -7.45025337e-01 -5.52534401e-01 -3.32820237e-01
1.20420298e-02 -8.49755466e-01 4.82872754e-01 -6.27233028e-01
-2.56842017e-01 4.04126942e-01 2.90369630e-01 -1.39130905e-01
-4.16141778e-01 -4.45688963e-01 -3.73788327e-01 -4.94081676e-01
3.37951750e-01 2.68041402e-01 -5.12920730e-02 -1.46471947e-01
8.77396941e-01 1.17720997e+00 -1.44230223e+00 6.78518116e-02
5.76181591e-01 5.88638604e-01 6.31631315e-01 1.19798768e+00
-1.13332786e-01 9.27567363e-01 -8.94654632e-01 6.93558622e-03
2.92545408e-01 -3.66445720e-01 -3.75793964e-01 5.76941133e-01
6.19084418e-01 -4.97555584e-01 -5.09894192e-01 6.61204219e-01
7.05794930e-01 2.41318032e-01 5.67226171e-01 -3.24335158e-01
-1.00279260e+00 1.51982948e-01 -4.23629940e-01 1.86626315e-01
7.00028896e-01 -1.70147583e-01 4.59360003e-01 -7.41594970e-01
1.01429772e+00 8.25974286e-01 5.94415367e-01 4.15349126e-01
-4.57347453e-01 -3.77650917e-01 1.71008110e-02 1.18465506e-01
-1.22622943e+00 -3.95484030e-01 8.73011291e-01 -3.68569613e-01
9.61246252e-01 5.65067410e-01 5.03703058e-01 1.29854143e+00
2.48964548e-01 1.00639403e+00 1.25030458e+00 -6.64918780e-01
3.46947387e-02 3.87789309e-02 3.35379720e-01 4.09231484e-01
1.42718792e-01 -1.66780036e-02 -9.24261749e-01 3.19974184e-01
1.09164143e+00 -1.36764848e-03 8.39513764e-02 -6.52937591e-02
-1.24408042e+00 1.98099583e-01 4.23616245e-02 5.83425939e-01
-4.35266227e-01 -2.67154306e-01 3.67349058e-01 2.18189374e-01
-9.58838910e-02 5.18676639e-01 -1.35587573e-01 -1.12800848e+00
-1.10364580e+00 1.12917602e-01 8.19767773e-01 1.27742028e+00
-6.52390346e-02 4.27729487e-01 2.47485787e-01 1.15331483e+00
3.41728836e-01 1.06070626e+00 8.51368487e-01 -6.26981974e-01
7.05613792e-01 5.22307158e-01 1.91095144e-01 -1.06632316e+00
-3.99605691e-01 -1.82122216e-01 -3.49710852e-01 -1.52024943e-02
3.30207556e-01 -1.36434913e-01 -9.78983164e-01 7.70835519e-01
-1.10220701e-01 -1.13911420e-01 1.43359497e-01 1.22389913e+00
3.47759962e-01 2.45433524e-01 -3.93049419e-01 1.24293029e-01
1.16303504e+00 -6.38032973e-01 -9.09772515e-01 -2.35222921e-01
1.53087288e-01 -1.19113708e+00 1.14443660e+00 4.66243327e-01
-6.26283407e-01 -8.27080309e-01 -1.38154078e+00 -4.22189608e-02
-2.20924750e-01 8.02419901e-01 5.23405194e-01 9.70662236e-01
-5.96983433e-01 2.98789054e-01 -1.21179068e+00 -6.98269963e-01
1.59936219e-01 2.35204011e-01 -2.36805931e-01 1.85883105e-01
-4.76021558e-01 8.52574646e-01 1.54466942e-01 3.24698210e-01
-1.18419014e-01 -3.17951478e-02 -3.38929176e-01 -7.46848762e-01
-1.37414575e-01 2.50709921e-01 9.08547938e-01 -7.16855764e-01
-2.05552530e+00 6.60786867e-01 -2.90620804e-01 -1.49695307e-01
9.56150413e-01 -3.40856314e-01 -8.56818199e-01 7.00931996e-02
-1.04132995e-01 -7.57186934e-02 7.71992326e-01 -7.70875514e-01
-6.38171554e-01 -6.01534069e-01 -6.63262188e-01 2.32226729e-01
-4.84962374e-01 2.62130737e-01 -7.64611840e-01 -1.01169670e+00
4.30754811e-01 -1.22100389e+00 3.96074086e-01 -4.37954277e-01
-2.64366418e-01 2.54487842e-01 1.36226058e+00 -1.06750166e+00
1.08446288e+00 -2.07825661e+00 8.95762295e-02 3.58755469e-01
-2.52700865e-01 6.19237065e-01 1.80887535e-01 4.75719959e-01
6.68702126e-01 -3.20590854e-01 2.84846932e-01 -1.01888277e-01
-1.54948831e-01 5.91858923e-01 -5.89736998e-01 3.84717554e-01
-2.09463090e-02 6.79600954e-01 -7.81527102e-01 -4.55629826e-02
4.91952360e-01 3.68462443e-01 2.42905483e-01 8.61471053e-03
1.99083775e-01 5.30027092e-01 -7.70545065e-01 1.19887698e+00
5.45169413e-01 3.52704376e-01 2.24153042e-01 5.86733297e-02
-4.29651707e-01 -6.33183271e-02 -1.46109593e+00 1.27911937e+00
-6.97234273e-01 1.22539008e+00 -1.87237933e-01 -4.58244562e-01
1.51307964e+00 7.49222338e-02 3.11921120e-01 -9.02629972e-01
1.54499158e-01 4.95934427e-01 -2.92367693e-02 -8.31288457e-01
1.21435297e+00 4.40538853e-01 1.20822839e-01 4.59998518e-01
-2.68846512e-01 -7.86705613e-02 2.90291220e-01 -2.25758538e-01
5.48910260e-01 4.80033398e-01 -3.17874067e-02 -1.68547541e-01
3.39846939e-01 -1.33746415e-01 2.85057783e-01 9.48824584e-01
-2.20960855e-01 7.05460429e-01 -1.45362213e-01 -6.35531843e-01
-1.17830122e+00 -8.71245384e-01 -4.53747272e-01 6.71351790e-01
1.76066428e-01 3.21066752e-02 -2.18093187e-01 -4.54167500e-02
2.29435772e-01 2.04091057e-01 -4.10514206e-01 3.90025467e-01
-7.55852282e-01 -4.85773593e-01 9.32144165e-01 1.16183090e+00
9.45302129e-01 -8.72592151e-01 -1.00010157e+00 4.21512455e-01
1.67614266e-01 -1.45208633e+00 -1.94060192e-01 -1.23110786e-01
-1.03738999e+00 -9.15761471e-01 -8.81007612e-01 -6.37890458e-01
5.88292062e-01 7.70024508e-02 4.47254688e-01 -3.46485227e-01
-4.29497033e-01 5.58559537e-01 -5.70067704e-01 -8.57911646e-01
-3.31199169e-01 2.97089577e-01 5.02215743e-01 7.18051791e-02
5.01036406e-01 -1.22020110e-01 -1.75065681e-01 5.71400106e-01
-4.55088586e-01 -1.98038548e-01 4.71202046e-01 4.78908956e-01
4.07027960e-01 -3.59415352e-01 1.81268886e-01 -6.18405640e-01
9.52213407e-01 -4.45711128e-02 -6.31851614e-01 3.73726159e-01
-2.34509051e-01 -7.24578947e-02 7.58729517e-01 -6.05873287e-01
-1.15345144e+00 -3.59431319e-02 1.35953337e-01 8.38718284e-03
-2.47412100e-01 3.69575381e-01 2.27184951e-01 -4.48963672e-01
6.30746841e-01 5.37096620e-01 -1.54323662e-02 -6.57020450e-01
-1.85848717e-02 1.61854362e+00 8.88529301e-01 -6.80263937e-01
1.08297758e-01 5.61064661e-01 6.36279285e-02 -1.64975786e+00
-7.53525048e-02 -4.68659848e-01 -1.09340107e+00 -5.95499694e-01
4.89224225e-01 -5.49452901e-01 -6.03363037e-01 1.07466877e+00
-6.74054205e-01 -6.96393624e-02 5.41633442e-02 8.90686214e-01
-3.59205693e-01 4.42280382e-01 -6.42230690e-01 -1.23623979e+00
-8.32278132e-02 -1.08626997e+00 1.38382506e+00 6.49984181e-01
-4.80722874e-01 -9.91840780e-01 1.44805368e-02 4.26344305e-01
4.81834322e-01 1.52964741e-01 1.69802353e-01 -2.26794153e-01
-3.58462155e-01 -7.16934800e-01 8.87153298e-02 1.72358319e-01
5.26946664e-01 5.96268356e-01 -8.70002210e-01 -1.04581341e-01
-2.67381310e-01 -1.13677152e-01 3.22776794e-01 1.67532021e-03
7.20454216e-01 1.93630219e-01 -2.55362336e-02 5.77847064e-01
1.14757740e+00 7.34530330e-01 5.83163857e-01 6.35058582e-01
8.70016992e-01 1.24660678e-01 7.71862447e-01 6.34163082e-01
7.90857598e-02 8.47566843e-01 -4.43829477e-01 2.65839785e-01
-2.35595986e-01 -7.88582787e-02 3.79051358e-01 1.24037588e+00
-8.04066777e-01 -1.44295776e-02 -1.16147280e+00 3.90824854e-01
-1.56966555e+00 -8.64140928e-01 -4.49980982e-02 2.03685498e+00
6.17327869e-01 -2.82414138e-01 2.46180117e-01 3.47285569e-01
4.58787620e-01 1.13864250e-01 -5.33963621e-01 -9.28953826e-01
-4.56374079e-01 -2.27170922e-02 9.81141627e-01 4.94296700e-01
-1.03435445e+00 1.04584241e+00 7.39631653e+00 1.88967973e-01
-1.99252748e+00 -6.30514503e-01 -7.83220157e-02 7.89427012e-02
5.68399310e-01 -3.75178128e-01 -1.18940794e+00 5.61670065e-01
7.34958887e-01 4.14030887e-02 5.97106814e-01 6.02700710e-01
3.26465935e-01 -1.74302086e-01 -6.48830175e-01 9.67020273e-01
3.58331740e-01 -1.19053745e+00 -1.03142984e-01 8.04226622e-02
8.91535819e-01 2.30918318e-01 1.21931233e-01 -3.74754995e-01
3.22074331e-02 -6.08363330e-01 8.77900302e-01 1.08667874e+00
6.44910634e-01 -3.99037212e-01 6.23121560e-01 3.54343146e-01
-1.04003298e+00 -9.03175026e-02 -4.86492813e-01 -4.66387004e-01
-5.40952906e-02 1.51238099e-01 -1.04173052e+00 5.61130166e-01
3.00512701e-01 9.02449310e-01 -6.13689899e-01 6.48999214e-01
-6.77429438e-02 6.88200891e-01 -4.14694518e-01 -6.57653213e-01
1.41078219e-01 -5.28816879e-01 6.23618782e-01 1.44809222e+00
4.35404658e-01 -2.23772854e-01 1.12549320e-01 2.83249736e-01
1.61718652e-01 1.74372233e-02 -6.06462002e-01 -5.63503146e-01
5.31107426e-01 7.01011539e-01 -8.78116965e-01 -2.49359161e-01
-1.63628876e-01 1.12948084e+00 -5.09418435e-02 3.93336922e-01
-5.10060787e-01 -6.26293421e-01 3.37643623e-01 -2.81619132e-01
2.48840094e-01 -1.13400233e+00 -9.88329828e-01 -1.11863971e+00
4.85559165e-01 -8.29947770e-01 -1.49105161e-01 -6.44015849e-01
-7.88075686e-01 4.69540775e-01 -2.67831564e-01 -1.40259564e+00
-4.19186115e-01 -1.07018530e+00 -2.46009961e-01 9.86360312e-01
-8.66258621e-01 -1.43851984e+00 -4.68719393e-01 2.15321526e-01
7.71415234e-01 -8.71007264e-01 8.39551091e-01 1.92202091e-01
-6.83735788e-01 6.39263868e-01 7.60824978e-01 2.85695016e-01
5.73764443e-01 -1.13716185e+00 3.30762446e-01 1.14750278e+00
4.63103592e-01 9.70703006e-01 4.87757176e-01 -8.42516661e-01
-2.04115438e+00 -7.33986676e-01 7.30300546e-01 -8.02946389e-01
5.77508807e-01 -3.08860511e-01 -5.06506860e-01 6.04328275e-01
-1.07088521e-01 -2.72704601e-01 6.36650741e-01 9.34655964e-02
-1.57877609e-01 -5.67058399e-02 -8.63196015e-01 6.64571583e-01
9.83098924e-01 -9.89289105e-01 -5.70902765e-01 1.11130819e-01
-5.48349738e-01 -9.58332300e-01 -1.00553107e+00 -6.13453351e-02
1.39943624e+00 -4.61180359e-01 7.36441553e-01 -5.99082589e-01
4.09392744e-01 -3.24152708e-01 -3.49047244e-01 -1.04858589e+00
-1.05843544e-01 -5.25254309e-01 -3.52768719e-01 1.31200612e+00
1.05079733e-01 -6.85386121e-01 7.55226016e-01 7.74224162e-01
3.08573782e-01 -4.36396658e-01 -6.34111285e-01 -1.08012509e+00
-2.99600095e-01 -5.26613593e-01 5.08050144e-01 7.88963318e-01
-1.17958253e-02 -3.06568146e-01 -7.95145035e-01 1.22878700e-01
3.54889870e-01 2.56877482e-01 9.83264565e-01 -7.93836832e-01
-1.98850438e-01 -3.20634931e-01 -1.09255958e+00 -9.57736552e-01
-1.43970862e-01 -5.72134018e-01 -3.03221434e-01 -1.11286044e+00
-3.62813890e-01 -4.49715137e-01 1.64097756e-01 6.52552471e-02
2.80374419e-02 4.71334845e-01 5.75378537e-01 4.70822841e-01
-1.49475068e-01 -6.73946887e-02 1.22162282e+00 -1.32226467e-01
-2.87272215e-01 1.55986696e-01 -2.53987581e-01 3.96488518e-01
8.31296086e-01 1.86280102e-01 -2.36519754e-01 -8.35780978e-01
-1.41700715e-01 -3.11555684e-01 -1.37726218e-01 -1.08652258e+00
4.47571129e-01 -1.18669018e-01 7.33738959e-01 -5.02849162e-01
5.84392734e-02 -7.37659693e-01 -4.95189205e-02 3.75888735e-01
-6.27020374e-02 2.52857447e-01 3.35891455e-01 2.23846018e-01
-3.30350161e-01 -9.89051014e-02 5.89822173e-01 2.73964167e-01
-1.30191660e+00 -3.72321874e-01 -6.29313409e-01 -5.70849836e-01
9.89572227e-01 -8.47761154e-01 -8.41056183e-02 -8.78828242e-02
-5.30801773e-01 -2.60935098e-01 5.10472119e-01 9.61054742e-01
4.50863719e-01 -1.18323421e+00 -4.93979186e-01 7.54974782e-01
2.26699710e-02 -6.10797286e-01 -1.62385385e-02 6.01077974e-01
-1.32747936e+00 7.71232486e-01 -8.16116154e-01 -6.96207881e-01
-1.49806583e+00 -2.05034584e-01 2.83382922e-01 2.73637950e-01
-6.39384687e-01 5.80338955e-01 -1.05870366e+00 -4.89916384e-01
2.87522793e-01 -4.38467979e-01 -1.14011742e-01 -1.28585368e-01
6.52777374e-01 7.56633222e-01 2.87009299e-01 -8.84666622e-01
-4.02225822e-01 1.00129688e+00 -1.05537064e-01 -2.45747134e-01
1.03786337e+00 -1.00125156e-01 2.77376443e-01 7.65603423e-01
7.45928049e-01 5.89786112e-01 -1.13506496e+00 4.74899486e-02
1.90361723e-01 -1.05158246e+00 -1.34812891e-01 -8.89400423e-01
-7.28820145e-01 6.57640159e-01 8.53005469e-01 -2.17529774e-01
9.90950406e-01 -6.60442889e-01 4.18118179e-01 8.22157502e-01
7.14645147e-01 -1.85505354e+00 -4.21353102e-01 8.75134528e-01
9.55550194e-01 -7.90151238e-01 1.33702949e-01 -7.28992000e-02
-8.79902422e-01 1.63433707e+00 2.87935197e-01 -1.57873020e-01
2.39637777e-01 7.22317278e-01 7.12743580e-01 1.55183822e-01
1.02039240e-01 2.25168556e-01 2.26968557e-01 7.50705838e-01
8.80217552e-01 4.64877903e-01 -6.06274426e-01 2.13527232e-01
-3.93524170e-01 3.55107754e-01 5.77555299e-01 1.59218192e+00
-5.05352728e-02 -1.32883942e+00 -7.54297614e-01 6.27160370e-01
-2.12499976e-01 5.53414524e-01 -9.14155066e-01 6.40439928e-01
-2.01647744e-01 6.96032465e-01 8.06654096e-02 -6.87295496e-01
6.19793832e-01 6.89874068e-02 8.03828835e-01 -2.70981994e-03
-3.69351685e-01 -1.69155568e-01 5.00249527e-02 -3.25591087e-01
-3.47015947e-01 -1.00429225e+00 -1.05014396e+00 -3.40570003e-01
-2.70788252e-01 -3.82421941e-01 1.15946639e+00 8.40496600e-01
5.72560370e-01 5.16223192e-01 7.40769446e-01 -8.97664011e-01
-5.89170754e-01 -1.14150739e+00 -8.99318337e-01 3.88961524e-01
9.50040966e-02 -5.19296169e-01 4.14674163e-01 1.95646450e-01] | [11.857401847839355, 2.5644547939300537] |
475b5549-f7fb-44a2-b15c-4334ac53f7ad | ttpp-temporal-transformer-with-progressive | 2003.03530 | null | https://arxiv.org/abs/2003.03530v1 | https://arxiv.org/pdf/2003.03530v1.pdf | TTPP: Temporal Transformer with Progressive Prediction for Efficient Action Anticipation | Video action anticipation aims to predict future action categories from observed frames. Current state-of-the-art approaches mainly resort to recurrent neural networks to encode history information into hidden states, and predict future actions from the hidden representations. It is well known that the recurrent pipeline is inefficient in capturing long-term information which may limit its performance in predication task. To address this problem, this paper proposes a simple yet efficient Temporal Transformer with Progressive Prediction (TTPP) framework, which repurposes a Transformer-style architecture to aggregate observed features, and then leverages a light-weight network to progressively predict future features and actions. Specifically, predicted features along with predicted probabilities are accumulated into the inputs of subsequent prediction. We evaluate our approach on three action datasets, namely TVSeries, THUMOS-14, and TV-Human-Interaction. Additionally we also conduct a comprehensive study for several popular aggregation and prediction strategies. Extensive results show that TTPP not only outperforms the state-of-the-art methods but also more efficient. | ['Xiaojiang Peng', 'Yu Qiao', 'Wen Wang', 'Jian Cheng', 'Yanzhou Su'] | 2020-03-07 | null | null | null | null | ['action-anticipation'] | ['computer-vision'] | [ 5.09791791e-01 -2.69546099e-02 -5.50603151e-01 -4.35555667e-01
-5.69464922e-01 7.88245052e-02 7.59669960e-01 -2.82222867e-01
-1.43656343e-01 5.86021721e-01 7.91807055e-01 1.95610411e-02
1.73923910e-01 -6.22497141e-01 -4.54883516e-01 -4.25695479e-01
-4.50872421e-01 1.61367450e-02 6.12991214e-01 -3.42348628e-02
1.63345039e-01 1.33244410e-01 -1.82498252e+00 9.20027375e-01
3.88720393e-01 1.29683244e+00 1.66063886e-02 6.97824776e-01
6.48078471e-02 1.95295668e+00 -1.62368476e-01 -3.98207158e-01
1.25655442e-01 -5.78841150e-01 -8.64069700e-01 7.63008669e-02
-1.53985709e-01 -6.70838594e-01 -7.01785386e-01 3.29863518e-01
9.24193859e-02 2.88991272e-01 1.70247555e-01 -1.34177983e+00
-3.80097121e-01 7.09337950e-01 -2.15154752e-01 4.62620467e-01
6.25280976e-01 6.02533758e-01 1.26305723e+00 -5.95480978e-01
6.62019014e-01 1.16914809e+00 7.95465469e-01 5.58363736e-01
-8.42749834e-01 -4.73813742e-01 6.21423423e-01 8.14789414e-01
-9.16502893e-01 -6.45414770e-01 7.17872024e-01 -3.73264581e-01
1.42172503e+00 1.91142321e-01 1.00922954e+00 1.35345197e+00
4.92067605e-01 1.49832618e+00 6.64659739e-01 -6.68372810e-02
8.81382674e-02 -4.64344442e-01 8.31340402e-02 5.98459303e-01
-6.35671318e-01 2.10065693e-01 -9.72843289e-01 1.39575154e-01
4.70579773e-01 4.86977160e-01 -3.37667987e-02 9.21763629e-02
-1.44221115e+00 4.21780854e-01 1.87845871e-01 1.82353050e-01
-1.07450688e+00 3.10217112e-01 5.87368250e-01 1.47115022e-01
6.87008500e-01 2.12227330e-02 -4.40471292e-01 -8.79322171e-01
-7.92703748e-01 3.15596700e-01 5.55383801e-01 6.79990649e-01
5.93841136e-01 5.67496195e-02 -6.41439795e-01 3.66377771e-01
2.72051871e-01 -1.12057500e-01 7.79565871e-01 -1.26356888e+00
5.85454047e-01 7.77584076e-01 2.65721530e-01 -8.73883665e-01
-3.25426608e-01 -1.06090419e-01 -6.33406758e-01 -1.48266301e-01
8.44515022e-03 -1.27793804e-01 -8.62139463e-01 1.40010202e+00
1.61335528e-01 8.67905974e-01 2.66493022e-01 7.24519730e-01
5.77533722e-01 9.25191164e-01 3.89663041e-01 -4.19808090e-01
8.22167456e-01 -1.33259273e+00 -7.56833613e-01 -2.23391429e-01
6.15703464e-01 -3.66113394e-01 8.46735895e-01 2.83902019e-01
-1.10887384e+00 -7.80035555e-01 -6.84069097e-01 3.29012261e-03
5.27831241e-02 2.31499568e-01 7.04817235e-01 -6.72880039e-02
-8.66537631e-01 1.09814787e+00 -1.42643988e+00 -3.26907367e-01
5.29441059e-01 1.63307577e-01 -1.25120223e-01 3.11418235e-01
-1.22034740e+00 7.70858467e-01 5.05974352e-01 2.04609767e-01
-1.08924091e+00 -5.71809173e-01 -8.06596518e-01 2.07380667e-01
4.67652023e-01 -4.43509459e-01 1.71197844e+00 -1.02243638e+00
-2.07546711e+00 1.59413338e-01 -5.01189768e-01 -1.03980207e+00
5.25279045e-01 -6.92067802e-01 -5.90758443e-01 7.38334805e-02
5.43495417e-02 6.56181633e-01 6.91601813e-01 -3.75796616e-01
-1.16078055e+00 -1.24734202e-02 1.63511023e-01 3.98322612e-01
-3.35648686e-01 2.76579678e-01 -3.46548110e-01 -4.55282629e-01
5.93820438e-02 -9.47960496e-01 -4.97407436e-01 -3.06155384e-01
-3.23102444e-01 -5.56474268e-01 7.95777559e-01 -5.50497770e-01
1.67011964e+00 -2.11493468e+00 6.02983385e-02 -3.30664694e-01
1.43687837e-02 1.55467734e-01 -1.15932956e-01 5.96215010e-01
-1.14241257e-01 -3.30645144e-01 2.51872480e-01 -5.30499339e-01
2.47287545e-02 3.31094235e-01 -6.43210948e-01 2.15208352e-01
2.34195262e-01 9.97066915e-01 -9.88774359e-01 -5.61492562e-01
4.64146823e-01 3.72885704e-01 -4.91560876e-01 5.17861903e-01
-6.18554652e-01 4.85089242e-01 -6.40262604e-01 5.55143714e-01
-2.26622477e-01 -5.44823349e-01 3.33249748e-01 -1.26369625e-01
-5.34012496e-01 7.39156842e-01 -6.18610084e-01 1.58159196e+00
-1.40533894e-01 6.00338697e-01 -1.08738792e+00 -8.82253587e-01
6.93925619e-01 5.98794639e-01 9.78395581e-01 -6.74152374e-01
-8.47427994e-02 -2.14284718e-01 -3.57616633e-01 -8.50783408e-01
8.19843590e-01 7.90817216e-02 -1.53308790e-02 3.70574802e-01
-6.37324080e-02 7.47238576e-01 2.27185592e-01 -5.62707186e-02
1.44858301e+00 9.09934163e-01 3.39449376e-01 5.64969063e-01
4.82693523e-01 2.79677100e-02 1.11716247e+00 4.71093327e-01
-4.42815840e-01 4.20675218e-01 7.06996322e-01 -9.35833633e-01
-7.39446759e-01 -7.88589895e-01 6.54106021e-01 1.46849358e+00
-6.10541366e-02 -7.47857392e-01 -2.25791916e-01 -8.20288122e-01
-4.24856782e-01 8.29953551e-01 -7.76226938e-01 -1.16948277e-01
-7.48998523e-01 -4.81573790e-01 2.52375096e-01 9.11202073e-01
7.32138097e-01 -1.53348613e+00 -1.14695036e+00 5.03678679e-01
-5.43930709e-01 -1.15744889e+00 -1.19205251e-01 -2.50172794e-01
-9.25130308e-01 -9.24669325e-01 -5.33699989e-01 -2.60842025e-01
3.91879715e-02 1.11202225e-01 1.04366159e+00 -2.30421245e-01
3.36720407e-01 1.77300185e-01 -7.79247284e-01 -1.27703816e-01
-2.37492070e-01 8.67391899e-02 3.74351032e-02 4.44484711e-01
5.95833361e-01 -7.13795304e-01 -7.22648680e-01 2.37558112e-01
-3.79127353e-01 5.95879078e-01 5.92108905e-01 5.18559039e-01
7.03226626e-01 5.84700145e-02 2.36502558e-01 -6.05380118e-01
1.35726303e-01 -6.64428413e-01 -2.06501812e-01 4.36344594e-01
-3.53876859e-01 -2.02724170e-02 5.62018752e-01 -4.13880140e-01
-1.40525174e+00 3.99385095e-01 -1.30150869e-01 -6.60352826e-01
-5.34627289e-02 6.77749753e-01 2.13889971e-01 6.50329590e-01
3.41361821e-01 5.05851507e-01 -3.20649981e-01 -3.77361327e-01
7.96833932e-02 3.36830139e-01 5.47327280e-01 7.79140508e-03
2.69610107e-01 5.29860437e-01 -1.35218769e-01 -2.74274230e-01
-1.13607955e+00 -3.09536606e-01 -6.10562384e-01 -5.68436623e-01
9.74166989e-01 -1.04041743e+00 -8.56930792e-01 6.23251379e-01
-1.07183444e+00 -6.49787188e-01 -3.59639525e-01 6.31805539e-01
-9.64648187e-01 2.56864667e-01 -8.53250861e-01 -9.33337152e-01
-3.67341310e-01 -8.59341264e-01 8.57904553e-01 1.74414530e-01
-4.67644840e-01 -5.76265097e-01 2.21926376e-01 2.68703729e-01
2.17282683e-01 3.00902069e-01 3.63865346e-01 -4.40613598e-01
-7.13612020e-01 -2.36883029e-01 8.16014558e-02 4.63756919e-02
4.00113128e-03 1.70836940e-01 -7.69160807e-01 9.81862918e-02
-3.76247704e-01 -3.09311390e-01 9.28482354e-01 3.83024633e-01
1.27412021e+00 -4.01806563e-01 -3.07735592e-01 3.02686721e-01
9.90737021e-01 3.75468433e-01 1.01252878e+00 4.79483753e-01
5.21702826e-01 4.48527873e-01 1.03278756e+00 9.66103196e-01
7.88305700e-01 7.26466119e-01 4.30728853e-01 3.02508801e-01
1.11488335e-01 -5.78920126e-01 8.84342730e-01 7.97453582e-01
-5.93701065e-01 -3.26761305e-01 -8.58856440e-01 5.29959977e-01
-2.40183282e+00 -1.52630508e+00 1.84124142e-01 1.69377542e+00
5.73173463e-01 2.81580478e-01 3.03807288e-01 1.46544844e-01
4.66364533e-01 6.10996723e-01 -6.38662934e-01 -3.84229086e-02
2.61091650e-01 -1.34087026e-01 5.52202910e-02 6.47006184e-02
-1.42430294e+00 1.13055253e+00 6.29898787e+00 4.48160619e-01
-1.01024520e+00 8.08608532e-02 6.90847337e-01 -3.49490464e-01
1.61200449e-01 1.34110183e-01 -7.07922220e-01 6.31858587e-01
1.29335392e+00 -2.20685735e-01 5.39362878e-02 9.02501166e-01
5.28166771e-01 -3.92718315e-02 -1.13214433e+00 8.24619293e-01
-1.49744630e-01 -1.54256678e+00 8.66114944e-02 -2.49314338e-01
5.31046212e-01 2.86973149e-01 1.06837749e-01 7.03812659e-01
5.16427398e-01 -7.18144357e-01 7.15381920e-01 1.08762217e+00
3.76697540e-01 -5.59834480e-01 5.93315542e-01 2.97553152e-01
-1.48901427e+00 -5.48312485e-01 -8.30968916e-02 -5.67270339e-01
6.90134466e-01 1.43144757e-01 -6.13380075e-01 4.36056495e-01
9.39701796e-01 1.72497535e+00 -4.35932130e-01 7.79701233e-01
-2.57671177e-01 7.89293528e-01 -1.90637552e-03 1.38085717e-02
4.84044373e-01 2.05054119e-01 2.57480919e-01 9.09037650e-01
4.36946094e-01 4.34513777e-01 4.32372302e-01 3.23067784e-01
2.11353853e-01 -1.97555438e-01 -4.12563533e-01 -2.59206086e-01
2.14780658e-01 8.67768884e-01 -3.85960281e-01 -7.72204399e-01
-7.36599684e-01 9.81284201e-01 2.71636844e-01 3.14916760e-01
-1.18949938e+00 1.44572526e-01 6.84779406e-01 1.06486445e-02
5.20007372e-01 -8.83440673e-02 1.53749168e-01 -1.41448212e+00
-5.75080281e-03 -6.88048303e-01 6.35849655e-01 -1.00803864e+00
-8.86059105e-01 7.98713624e-01 4.19322848e-02 -1.85926330e+00
-8.14090610e-01 -4.14374880e-02 -8.49623561e-01 2.48968229e-01
-1.41636598e+00 -1.22606897e+00 -2.35653013e-01 5.99131286e-01
1.09102046e+00 -2.27264717e-01 7.89476275e-01 7.45691136e-02
-9.37626600e-01 1.27179876e-01 -3.46078902e-01 2.16249302e-02
1.54726028e-01 -7.47847915e-01 6.44938529e-01 9.38063979e-01
3.80523242e-02 1.38399214e-01 7.72824824e-01 -7.58913398e-01
-1.16645896e+00 -1.34914482e+00 1.12955296e+00 -2.81415045e-01
9.28174019e-01 2.83825457e-01 -7.56215513e-01 1.21331382e+00
1.87042058e-01 1.64631113e-01 6.63179517e-01 2.81560384e-02
3.77435014e-02 -5.98826781e-02 -4.99260902e-01 7.71937132e-01
1.21985888e+00 -4.00151402e-01 -6.24724030e-01 2.34342426e-01
8.03953171e-01 -3.09271812e-01 -9.05138731e-01 5.40645003e-01
9.45191324e-01 -1.33828437e+00 6.63440764e-01 -7.93996453e-01
8.30141485e-01 -6.65917248e-02 -1.15775183e-01 -8.59071016e-01
-5.16199946e-01 -8.78569245e-01 -7.94403672e-01 9.79829490e-01
3.81817222e-01 -2.30733246e-01 1.19570041e+00 7.81787455e-01
-2.12029621e-01 -1.08511996e+00 -8.07392120e-01 -5.86715460e-01
-7.17955053e-01 -6.91468835e-01 6.95426226e-01 4.99823123e-01
2.06684083e-01 1.23230293e-01 -9.01345909e-01 -1.13604777e-01
1.88073397e-01 3.09164494e-01 7.69905508e-01 -8.66361320e-01
-3.83842260e-01 -4.85289358e-02 -6.50717676e-01 -1.36129379e+00
2.53344506e-01 -2.30093464e-01 1.62317783e-01 -1.31879163e+00
1.63660392e-01 1.20602641e-02 -7.33615816e-01 8.28946948e-01
-4.63984013e-02 1.26359075e-01 3.73985738e-01 4.63389188e-01
-1.36066806e+00 9.50077713e-01 8.65171134e-01 8.44646692e-02
-3.38238299e-01 3.81969303e-01 -2.14416794e-02 9.39517617e-01
9.00847375e-01 -3.47324014e-01 -5.33886254e-01 -3.72493535e-01
1.99352592e-01 4.25152302e-01 4.71358597e-01 -1.46285129e+00
1.95988014e-01 -4.76369202e-01 3.72694165e-01 -9.85204339e-01
6.07212782e-01 -6.15389585e-01 2.71011680e-01 4.36932206e-01
-7.09815681e-01 2.71738708e-01 -2.58206427e-01 6.62702680e-01
-3.37669671e-01 3.05124730e-01 2.49770150e-01 -2.45571584e-01
-1.33128560e+00 5.38866460e-01 -6.98736787e-01 -4.41912323e-01
1.20620799e+00 -3.33315104e-01 -1.16388984e-01 -5.27003586e-01
-9.34618235e-01 2.69824773e-01 4.45269421e-02 5.59458375e-01
7.67755091e-01 -1.48474777e+00 -5.50907135e-01 4.01354581e-02
-8.57261382e-03 -4.65273857e-01 5.05850911e-01 1.15812123e+00
-2.46583924e-01 4.12222803e-01 -3.05774480e-01 -5.35437346e-01
-1.38351524e+00 5.30701697e-01 1.41288564e-01 -6.22090697e-01
-1.01514232e+00 6.96112990e-01 -1.79605305e-01 2.17120618e-01
3.26308101e-01 -3.86840940e-01 -6.34127975e-01 -3.29169864e-03
7.36251712e-01 5.54604173e-01 -6.06213570e-01 -7.54572213e-01
-2.18425497e-01 1.60759374e-01 -1.62519529e-01 -5.25596701e-02
1.56717765e+00 -2.99840868e-01 1.96562916e-01 7.49271274e-01
9.25157011e-01 -8.59204352e-01 -1.87392688e+00 -4.45382953e-01
1.64721742e-01 -5.12994647e-01 -2.22731039e-01 -4.31065917e-01
-1.07964182e+00 6.72281682e-01 4.37368453e-01 2.69602507e-01
1.34626269e+00 -1.92161590e-01 1.24703455e+00 4.72507477e-01
4.13832158e-01 -1.16707122e+00 2.83466697e-01 6.50745511e-01
7.52516091e-01 -1.09282959e+00 1.25984132e-01 -1.11498125e-01
-1.06409156e+00 9.97815251e-01 8.47698092e-01 -1.32097006e-01
4.66553271e-01 -1.75166935e-01 -7.22003356e-02 -6.73535466e-02
-1.56475282e+00 -3.09513748e-01 1.04227751e-01 1.04455031e-01
4.92816716e-01 -1.29984453e-01 2.67870203e-02 4.43341315e-01
-6.94768620e-04 6.54617071e-01 3.08247298e-01 1.11078572e+00
-3.43112618e-01 -8.18118155e-01 1.18297905e-01 5.51730633e-01
-5.22152066e-01 1.09934434e-01 -3.03688995e-03 3.80656242e-01
8.62280428e-02 8.21069300e-01 2.67794639e-01 -8.98821175e-01
1.79848835e-01 1.55724585e-01 7.35802352e-02 -3.73349875e-01
-4.94291514e-01 -1.27728596e-01 2.40760624e-01 -1.34043455e+00
-9.42070603e-01 -9.46322680e-01 -1.18732119e+00 -2.03458861e-01
8.79603475e-02 -9.66352671e-02 -4.42720689e-02 1.01691484e+00
6.25057101e-01 6.92223966e-01 6.88376844e-01 -1.11629796e+00
-3.49599332e-01 -1.05055571e+00 -8.18351209e-02 3.66962582e-01
2.37398595e-01 -7.12277055e-01 3.20641585e-02 3.40852678e-01] | [8.193334579467773, 0.5084204077720642] |
ba6d3455-dfbd-4f4f-aa46-bef4901ba88b | guided-table-structure-recognition-through | 2104.10538 | null | https://arxiv.org/abs/2104.10538v1 | https://arxiv.org/pdf/2104.10538v1.pdf | Guided Table Structure Recognition through Anchor Optimization | This paper presents the novel approach towards table structure recognition by leveraging the guided anchors. The concept differs from current state-of-the-art approaches for table structure recognition that naively apply object detection methods. In contrast to prior techniques, first, we estimate the viable anchors for table structure recognition. Subsequently, these anchors are exploited to locate the rows and columns in tabular images. Furthermore, the paper introduces a simple and effective method that improves the results by using tabular layouts in realistic scenarios. The proposed method is exhaustively evaluated on the two publicly available datasets of table structure recognition i.e ICDAR-2013 and TabStructDB. We accomplished state-of-the-art results on the ICDAR-2013 dataset with an average F-Measure of 95.05$\%$ (94.6$\%$ for rows and 96.32$\%$ for columns) and surpassed the baseline results on the TabStructDB dataset with an average F-Measure of 94.17$\%$ (94.08$\%$ for rows and 95.06$\%$ for columns). | ['Muhammad Zeshan Afzal', 'Muhammad Noman Afzal', 'Marcus Liwicki', 'Didier Stricker', 'Khurram Azeem Hashmi'] | 2021-04-21 | null | null | null | null | ['table-recognition'] | ['computer-vision'] | [ 1.78940520e-01 1.84603687e-02 -3.17399174e-01 -3.02569926e-01
-1.21309352e+00 -7.53899813e-01 3.14176530e-01 7.32416987e-01
-1.87455434e-02 6.40220642e-01 5.93720423e-03 -3.70587528e-01
-2.09832579e-01 -9.69080150e-01 -8.84627342e-01 -4.07136917e-01
-4.13522929e-01 7.07262099e-01 1.18049182e-01 -1.56409994e-01
6.84694827e-01 4.53247964e-01 -1.74221754e+00 9.07469213e-01
5.50105095e-01 1.66996682e+00 -1.59293786e-01 3.48579526e-01
-6.35032773e-01 8.68095338e-01 -8.53533983e-01 -5.90504646e-01
2.90238798e-01 -1.78905681e-01 -7.76681721e-01 2.57599503e-01
8.94515753e-01 -5.78228161e-02 -1.55326620e-01 8.22524786e-01
1.97415501e-02 -2.03266010e-01 6.42898023e-01 -9.14195836e-01
-3.50832015e-01 7.23952830e-01 -7.95849919e-01 8.98714289e-02
5.97100079e-01 -5.60949802e-01 1.32976007e+00 -1.12126756e+00
8.32422316e-01 1.08729839e+00 6.03828967e-01 1.83629226e-02
-1.23164558e+00 -6.97290778e-01 2.97758192e-01 2.15930670e-01
-1.76209462e+00 -6.68299198e-01 5.63971519e-01 -4.07345057e-01
1.05798602e+00 4.18195188e-01 1.16994552e-01 5.35316944e-01
1.68918148e-01 9.34065700e-01 9.33214247e-01 -5.56440234e-01
9.86615866e-02 1.75802425e-01 4.10980523e-01 9.50406194e-01
7.60211706e-01 -5.37725627e-01 -7.37355590e-01 -1.78411514e-01
3.99220526e-01 -1.80627808e-01 1.52312487e-01 -5.69799602e-01
-1.21030176e+00 4.75363582e-01 2.82161415e-01 1.56946108e-01
-1.92390218e-01 -3.98267359e-01 5.74838877e-01 -5.56551069e-02
-9.84151587e-02 3.34837407e-01 -3.53545815e-01 -9.15293582e-03
-1.03792703e+00 2.09773958e-01 8.46841872e-01 1.56452215e+00
6.04048014e-01 3.06556597e-02 -4.46629822e-01 9.19551015e-01
2.53916830e-01 5.17693281e-01 -2.94704348e-01 -6.12197876e-01
1.28116751e+00 1.15500498e+00 1.62185833e-01 -1.23641479e+00
-2.45914623e-01 -4.88300204e-01 -1.02458024e+00 -1.51330113e-01
3.68631124e-01 6.57822847e-01 -1.04959440e+00 9.34951365e-01
2.61258800e-02 -8.64399433e-01 2.73108166e-02 2.72250414e-01
8.33250642e-01 7.26645827e-01 -3.77014995e-01 -2.26166993e-01
1.66255045e+00 -8.92855942e-01 -1.00615478e+00 -2.57512838e-01
6.48266435e-01 -9.91265059e-01 1.10366642e+00 6.83669865e-01
-1.17320800e+00 -6.64551795e-01 -1.41187012e+00 1.51182547e-01
-5.95199168e-01 4.16664183e-01 5.29252708e-01 1.02463496e+00
-6.78347588e-01 6.56824112e-02 -4.93384033e-01 -2.07489744e-01
4.72467393e-01 4.49796438e-01 -3.62884045e-01 -1.93318903e-01
-5.61743796e-01 3.90995443e-01 5.18807292e-01 -1.54233184e-02
-5.98400414e-01 -4.77442950e-01 -8.78701746e-01 1.34909719e-01
9.57315505e-01 -3.83731313e-02 8.68471205e-01 -6.43329397e-02
-8.42622638e-01 1.07120538e+00 -3.36702228e-01 -5.09113073e-01
1.41382277e-01 -9.87648964e-02 -5.92679143e-01 9.60238874e-02
2.52414674e-01 3.86963636e-01 2.93047488e-01 -1.52444124e+00
-9.23406184e-01 -5.38059533e-01 -1.73676610e-01 -1.17847109e-02
-1.85384437e-01 -1.71432793e-01 -9.42841887e-01 -7.30531812e-01
6.96101725e-01 -5.77730238e-01 1.89855874e-01 -1.69304341e-01
-8.73658538e-01 1.45952791e-01 5.95560372e-01 -6.91821098e-01
1.82300162e+00 -1.92990851e+00 -3.68048161e-01 5.74181080e-01
1.66131750e-01 1.15022980e-01 3.01798940e-01 7.03310609e-01
-2.69293180e-03 2.73866802e-01 -2.95834482e-01 -2.39152133e-01
1.58994645e-01 3.54188420e-02 -5.07110417e-01 2.66339511e-01
-1.68543875e-01 4.99783993e-01 -2.56039321e-01 -6.53269708e-01
3.96637619e-02 7.46749490e-02 -5.19824922e-01 3.20784524e-02
-1.32835340e-02 -4.43687320e-01 -2.95080334e-01 1.39878488e+00
9.38078642e-01 -1.90020069e-01 7.09152400e-01 -3.55462283e-01
-2.57087976e-01 5.45114696e-01 -1.67022145e+00 1.39451933e+00
8.41079131e-02 2.05909982e-01 1.50237724e-01 -1.01393986e+00
1.42851615e+00 -7.16604292e-02 5.01532137e-01 -1.02358651e+00
-1.23657428e-01 3.02293748e-01 -1.30257845e-01 5.12922443e-02
6.33935034e-01 4.68463570e-01 -5.49275279e-01 1.49386823e-01
-2.13024467e-01 7.88215622e-02 7.95629084e-01 2.90247232e-01
1.18761897e+00 1.20790847e-01 5.07610559e-01 -6.21079862e-01
8.42445612e-01 2.85082877e-01 4.46226150e-01 9.40469980e-01
4.15293686e-02 6.86359465e-01 5.91704667e-01 -5.98753452e-01
-9.55322266e-01 -1.30999994e+00 -2.94585854e-01 8.22783530e-01
2.58591585e-02 -1.01837099e+00 -8.40315044e-01 -7.92850316e-01
1.68566555e-01 5.57998598e-01 -7.14854360e-01 3.27456325e-01
-6.13408864e-01 -6.37602210e-01 4.98296052e-01 8.48363578e-01
8.45784247e-01 -8.84074867e-01 -2.88466573e-01 2.71292865e-01
-2.90387511e-01 -1.29576433e+00 -2.75291651e-01 5.19722879e-01
-9.00038719e-01 -1.12743843e+00 -3.50126587e-02 -7.14113712e-01
7.25616574e-01 1.18536837e-01 1.52074015e+00 1.68766558e-01
-3.63117427e-01 5.15884459e-02 -2.84149051e-01 -3.64294201e-01
-1.80154637e-01 2.49835432e-01 -1.66049153e-01 -1.55475974e-01
3.25094134e-01 4.33284380e-02 -4.29555982e-01 7.66801476e-01
-6.83683634e-01 -8.47249106e-02 7.05170155e-01 8.30990493e-01
1.23171711e+00 3.07322234e-01 1.81792215e-01 -1.25540519e+00
1.16685428e-01 -7.91211650e-02 -1.00736785e+00 5.94949186e-01
-1.03527677e+00 1.89908952e-01 6.74510241e-01 3.15845937e-01
-7.81874061e-01 2.51064569e-01 -1.77942589e-01 1.06231961e-02
-1.18593715e-01 5.38583994e-01 -5.33762574e-01 5.01130223e-01
4.14967597e-01 2.54875034e-01 -4.66960102e-01 -7.82258272e-01
-1.35524735e-01 6.61213875e-01 6.27152383e-01 -6.47087276e-01
8.27453732e-01 6.18729711e-01 1.75897628e-01 -4.02722120e-01
-7.09547162e-01 -5.18350720e-01 -8.46860349e-01 1.06259294e-01
5.13688803e-01 -7.68016756e-01 -8.93138289e-01 2.24666953e-01
-5.40843070e-01 1.51724681e-01 6.77348226e-02 -1.30398303e-01
-4.12911326e-01 3.07874262e-01 -2.62298614e-01 -8.96424353e-01
-3.95791560e-01 -1.12315941e+00 1.06986880e+00 -2.28453785e-01
-3.21139961e-01 -5.19945383e-01 -4.01016265e-01 7.00788617e-01
3.97790819e-02 2.62703598e-01 1.29933250e+00 -5.88861525e-01
-9.61136758e-01 -4.11727458e-01 -5.59328139e-01 -1.11520551e-01
2.29341537e-01 -3.52228470e-02 -7.24221826e-01 -1.56221077e-01
-3.81682992e-01 -3.68744209e-02 7.14702666e-01 2.42664255e-02
1.40399158e+00 -4.42703903e-01 -6.79808021e-01 4.45186824e-01
1.49018466e+00 9.78231013e-01 1.04380381e+00 8.21000695e-01
5.59951782e-01 5.29818118e-01 1.07370746e+00 6.43199742e-01
4.47706670e-01 8.88624132e-01 4.83715028e-01 1.17658444e-01
3.22437175e-02 -3.82144660e-01 -2.77450541e-04 4.97347653e-01
3.61463130e-01 -6.17417395e-01 -1.40835106e+00 5.81051767e-01
-1.68606925e+00 -7.18498647e-01 -2.12193444e-01 2.28981614e+00
5.52363098e-01 8.10832679e-01 2.00396016e-01 7.90418446e-01
6.12079084e-01 1.68721393e-01 -2.21277341e-01 -5.32229722e-01
-3.47071439e-01 2.61986285e-01 6.48380101e-01 7.61420280e-02
-1.37474155e+00 6.04185939e-01 6.16320992e+00 9.18289840e-01
-4.01135653e-01 -6.36808753e-01 9.99873519e-01 9.75358784e-02
4.64895330e-02 -2.00833634e-01 -1.39554918e+00 2.34925255e-01
9.53766286e-01 1.57741934e-01 1.39585599e-01 9.18927610e-01
-3.92498791e-01 -4.51634496e-01 -1.21047711e+00 1.24656022e+00
3.00935000e-01 -1.69567573e+00 1.58506215e-01 1.36730030e-01
4.88902122e-01 -6.85281098e-01 2.80383915e-01 1.82578698e-01
-5.65374717e-02 -1.12663293e+00 9.76415992e-01 2.34519452e-01
9.55676496e-01 -9.68022108e-01 7.50563264e-01 -1.70236418e-03
-1.62376535e+00 -1.54462367e-01 -2.99745530e-01 1.90221637e-01
-3.44523877e-01 3.71959627e-01 -8.59268367e-01 7.47563481e-01
1.23928177e+00 4.43593562e-01 -8.27421904e-01 7.38977194e-01
6.33547455e-02 5.36426723e-01 -3.80763024e-01 6.74152374e-03
4.40432169e-02 -1.02845654e-01 1.71690673e-01 1.35631466e+00
2.18447194e-01 -2.14594468e-01 1.58614218e-01 4.91543114e-01
-3.36684108e-01 3.59205276e-01 -5.16044438e-01 -1.74319874e-02
5.26395380e-01 8.61552358e-01 -1.17570806e+00 -4.98969942e-01
-2.93273985e-01 4.91297007e-01 7.78877288e-02 -5.98195828e-02
-6.81591630e-01 -6.11061871e-01 2.75122166e-01 4.36461031e-01
8.37094963e-01 -1.24064669e-01 -9.12783325e-01 -8.75082254e-01
5.50476134e-01 -1.24111295e+00 9.78314817e-01 -4.34844553e-01
-7.94281960e-01 8.33548963e-01 2.40305215e-01 -1.37602282e+00
-1.84987068e-01 -8.16298664e-01 1.66380167e-01 6.35963500e-01
-1.02312982e+00 -7.72194803e-01 -3.14618975e-01 2.47137249e-01
3.77154022e-01 -4.60503846e-01 1.03712249e+00 3.34451795e-01
-6.70640349e-01 1.04049659e+00 4.68735576e-01 4.32109177e-01
6.40731037e-01 -1.36552536e+00 5.71182430e-01 9.51474667e-01
3.65779132e-01 8.35342884e-01 4.70231324e-01 -5.34598827e-01
-1.91372478e+00 -7.37844527e-01 9.23120558e-01 -6.16382778e-01
3.52981061e-01 -1.00696754e+00 -8.46465528e-01 5.35903931e-01
1.13140635e-01 -1.48778856e-01 5.90824366e-01 3.14041823e-01
-7.24830449e-01 -7.71794319e-01 -1.46348429e+00 5.93246460e-01
1.19344473e+00 -2.80580431e-01 -3.70577365e-01 -3.68626863e-02
-3.00411321e-02 -5.60789645e-01 -1.06823909e+00 5.81504643e-01
8.46197248e-01 -1.03313661e+00 1.11194181e+00 5.93063645e-02
1.88482255e-01 -5.71923077e-01 -7.40003824e-01 -2.74661601e-01
-2.10669696e-01 -2.69347280e-01 -3.11144650e-01 1.36880314e+00
7.50401855e-01 -3.02328408e-01 1.21987200e+00 3.78192693e-01
-2.93975502e-01 -7.63886273e-01 -8.43505740e-01 -9.36156034e-01
-2.89053082e-01 -2.01450557e-01 5.76181471e-01 5.28398931e-01
-1.17579237e-01 2.53682554e-01 -5.44191636e-02 2.98158498e-03
7.75985539e-01 5.42110980e-01 9.73382533e-01 -9.64045405e-01
4.05432060e-02 -3.35134178e-01 -3.86226773e-01 -1.06456995e+00
-4.57895964e-01 -6.95860386e-01 -8.55788514e-02 -1.75588787e+00
3.16288888e-01 -3.42374772e-01 -5.38656831e-01 5.08246779e-01
1.55421302e-01 3.89870942e-01 1.89499125e-01 2.33899549e-01
-7.78897345e-01 -8.80658254e-02 7.39119828e-01 -5.35168469e-01
8.09492022e-02 -3.10782999e-01 -9.11442995e-01 3.64579141e-01
8.22136939e-01 -6.00031793e-01 -3.23548079e-01 -3.98940593e-02
1.67841151e-01 5.24015687e-02 -2.66552866e-01 -1.15961397e+00
1.85079306e-01 1.29984310e-02 7.49418736e-01 -1.74586940e+00
2.68164873e-01 -8.81014049e-01 1.06615797e-01 4.71724182e-01
-3.84282917e-02 5.63919485e-01 6.02237999e-01 3.68421316e-01
-4.57895994e-01 -9.10296384e-03 5.96157551e-01 1.07522560e-02
-6.80771649e-01 -1.07508756e-01 -3.52706134e-01 -4.18737717e-02
9.95475709e-01 -7.18700230e-01 -4.91011798e-01 -4.46737632e-02
-3.68431926e-01 4.85784598e-02 3.74965221e-01 2.78075755e-01
7.56988347e-01 -1.19642913e+00 -4.16949302e-01 4.67655957e-01
6.79228842e-01 -8.43081698e-02 1.77764595e-02 4.43425834e-01
-7.82862544e-01 1.00691116e+00 -2.04893410e-01 -6.73074841e-01
-1.45243967e+00 8.44907463e-01 -2.71699969e-02 -7.30627477e-01
-2.47786671e-01 7.09923029e-01 1.20207131e-01 -1.00359708e-01
5.78361809e-01 -4.48747188e-01 -9.03408006e-02 3.00486714e-01
4.63881135e-01 3.73023659e-01 7.84783423e-01 -5.06859601e-01
-8.23000789e-01 4.97243524e-01 -5.70108891e-01 7.41092712e-02
1.07601726e+00 -1.51073486e-02 -6.95401952e-02 3.55678886e-01
8.99146914e-01 2.47697875e-01 -7.44490683e-01 -8.31996948e-02
5.55346787e-01 -6.83335125e-01 -4.28441077e-01 -1.24965942e+00
-7.09794283e-01 5.84774613e-01 6.25756502e-01 3.05756897e-01
1.27948511e+00 -1.85437769e-01 4.25780416e-01 6.80340052e-01
6.34765506e-01 -1.14571679e+00 -2.78632604e-02 4.66283113e-01
8.31630230e-01 -1.13128829e+00 4.33198452e-01 -8.38521838e-01
-3.85568976e-01 1.06009793e+00 6.31350636e-01 1.64511606e-01
4.52279985e-01 7.51661241e-01 9.26713049e-02 -2.85610318e-01
-8.84936869e-01 3.13601196e-02 5.75958908e-01 4.62604731e-01
6.63416564e-01 -3.79012115e-02 -1.93489328e-01 5.70110917e-01
-2.25138351e-01 -4.83398467e-01 2.79807687e-01 1.39990020e+00
-3.65932345e-01 -1.42573237e+00 -8.61967266e-01 6.72718763e-01
-7.53455997e-01 -1.87747627e-01 -6.53189361e-01 1.17270756e+00
-1.77501708e-01 1.00967395e+00 2.04985812e-01 -3.99317920e-01
6.36325598e-01 1.17966771e-01 4.21275437e-01 -5.52697122e-01
-7.04359710e-01 2.56434321e-01 3.74487340e-01 -5.15082836e-01
-8.24238807e-02 -7.38313138e-01 -1.21568739e+00 -3.95886451e-01
1.18569486e-01 2.27031380e-01 5.67906022e-01 2.87519157e-01
4.65456277e-01 4.73014325e-01 3.39733958e-01 6.55549020e-02
-1.37814403e-01 -6.77779198e-01 -5.16401470e-01 2.43353412e-01
-2.46337447e-02 -8.24559093e-01 9.94717926e-02 2.30331913e-01] | [11.696083068847656, 3.036895990371704] |
979fd3de-0988-4413-b2d8-04db8ec53506 | online-3d-bin-packing-reinforcement-learning | 2208.07123 | null | https://arxiv.org/abs/2208.07123v1 | https://arxiv.org/pdf/2208.07123v1.pdf | Online 3D Bin Packing Reinforcement Learning Solution with Buffer | The 3D Bin Packing Problem (3D-BPP) is one of the most demanded yet challenging problems in industry, where an agent must pack variable size items delivered in sequence into a finite bin with the aim to maximize the space utilization. It represents a strongly NP-Hard optimization problem such that no solution has been offered to date with high performance in space utilization. In this paper, we present a new reinforcement learning (RL) framework for a 3D-BPP solution for improving performance. First, a buffer is introduced to allow multi-item action selection. By increasing the degree of freedom in action selection, a more complex policy that results in better packing performance can be derived. Second, we propose an agnostic data augmentation strategy that exploits both bin item symmetries for improving sample efficiency. Third, we implement a model-based RL method adapted from the popular algorithm AlphaGo, which has shown superhuman performance in zero-sum games. Our adaptation is capable of working in single-player and score based environments. In spite of the fact that AlphaGo versions are known to be computationally heavy, we manage to train the proposed framework with a single thread and GPU, while obtaining a solution that outperforms the state-of-the-art results in space utilization. | ['Sukhan Lee', 'Aaron Valero Puche'] | 2022-08-15 | null | null | null | null | ['3d-bin-packing'] | ['miscellaneous'] | [-4.74448800e-02 6.52216822e-02 -2.87691534e-01 1.25180930e-01
-5.76104701e-01 -3.17514271e-01 2.16395892e-02 3.65999162e-01
-6.45597100e-01 9.25359666e-01 -3.11511517e-01 -3.73382747e-01
-6.35156870e-01 -1.01742709e+00 -8.17761123e-01 -8.15404892e-01
-3.78439814e-01 9.89875436e-01 3.55281740e-01 -4.45628554e-01
2.77148098e-01 5.63915312e-01 -1.66762745e+00 6.13879487e-02
8.46166313e-01 1.19623280e+00 6.80186689e-01 6.79405808e-01
5.27408626e-03 4.71616089e-01 -7.09787488e-01 -4.39388782e-01
7.31092632e-01 -5.02964973e-01 -8.29379022e-01 3.54333192e-01
-1.20612860e-01 -2.17940301e-01 4.57728021e-02 7.10810304e-01
7.10744858e-01 3.43157083e-01 3.22309405e-01 -1.25739872e+00
-2.42367536e-01 5.81714153e-01 -6.98794484e-01 2.39728212e-01
3.57713699e-01 3.13737541e-01 1.06207705e+00 4.29611653e-02
3.98450971e-01 1.03860128e+00 1.65138364e-01 4.25474614e-01
-1.23675585e+00 -2.75863677e-01 8.16132128e-02 4.39870834e-01
-9.62321341e-01 1.73646510e-01 3.88176471e-01 -6.16537668e-02
1.20238805e+00 5.53692818e-01 1.27832985e+00 8.10777247e-01
1.85223773e-01 1.00833094e+00 1.11795247e+00 -5.71389198e-01
7.05215335e-01 -1.31866068e-01 -1.54895708e-01 4.88677919e-01
4.55829442e-01 2.78521270e-01 -4.15536970e-01 -5.14626428e-02
6.55109823e-01 -4.96216007e-02 5.61257452e-02 -8.77873421e-01
-6.97284758e-01 1.10108149e+00 4.29646730e-01 4.38715816e-02
-5.45185864e-01 1.02267735e-01 4.24805522e-01 2.13698119e-01
3.34756851e-01 8.43110681e-01 -3.56023222e-01 -4.82999802e-01
-8.22785616e-01 9.25389946e-01 1.11166298e+00 9.90599036e-01
5.53810179e-01 -6.19288348e-02 -4.84933019e-01 6.65463150e-01
2.44387649e-02 3.08041573e-01 4.86069560e-01 -7.87220776e-01
6.17438912e-01 5.14767468e-01 3.34258556e-01 -7.60916293e-01
-6.22364283e-01 -5.73839366e-01 -5.69819510e-01 4.16965336e-01
3.73539895e-01 9.52392537e-03 -8.21763217e-01 1.34657824e+00
4.40708786e-01 6.21465407e-02 -2.06683632e-02 1.07074594e+00
2.11739853e-01 6.80648148e-01 -1.92262545e-01 -3.92461658e-01
1.27959549e+00 -1.33048165e+00 -5.80929816e-01 -2.08303899e-01
4.59795326e-01 -4.51233387e-01 7.78053224e-01 7.53981471e-01
-1.42401028e+00 -3.20846736e-01 -1.22054458e+00 4.27784294e-01
-4.37468648e-01 -4.72733617e-01 8.31497848e-01 8.99539709e-01
-6.27682447e-01 8.59534264e-01 -9.73369896e-01 -1.98943064e-01
4.24268007e-01 7.51605034e-01 -1.52614132e-01 -1.43188864e-01
-8.18385422e-01 9.56663013e-01 8.56392086e-01 -1.41063839e-01
-6.53796017e-01 -3.94708127e-01 -6.98545635e-01 2.04394430e-01
1.06301808e+00 -6.12628043e-01 1.22615159e+00 -3.95764947e-01
-1.91649985e+00 4.72527117e-01 3.64057630e-01 -1.03579271e+00
6.53752387e-01 -1.86823979e-01 8.68821070e-02 -2.61617862e-02
-1.69906586e-01 3.95965517e-01 7.32629061e-01 -1.03258550e+00
-9.43521023e-01 -3.94663990e-01 3.46279860e-01 3.65047246e-01
-2.04069898e-01 -2.10724443e-01 -4.75104123e-01 -3.90291810e-01
-2.87999243e-01 -9.30400312e-01 -6.97380245e-01 -6.51289284e-01
-2.34253973e-01 -1.80141374e-01 1.04802556e-01 -4.74181384e-01
1.30921078e+00 -1.89370620e+00 5.28386116e-01 2.45878324e-01
-7.48559684e-02 5.19186199e-01 -4.91420440e-02 5.25939703e-01
3.19839150e-01 -1.37664720e-01 -5.89264743e-02 -4.45513815e-01
4.87654865e-01 5.98430634e-01 1.12931304e-01 3.51633400e-01
1.08283177e-01 7.69163370e-01 -8.68845522e-01 -3.12184840e-01
3.75235707e-01 -3.19519579e-01 -1.13518703e+00 4.65909630e-01
-4.81065303e-01 7.01866625e-03 -3.27885747e-01 5.06149054e-01
8.68643999e-01 -6.52940646e-02 2.05033720e-01 3.28075081e-01
-8.65165144e-02 1.52832372e-02 -1.57283032e+00 1.63753450e+00
-2.72925496e-01 -3.11541289e-01 8.74338225e-02 -1.41768384e+00
8.71232688e-01 -1.57126993e-01 7.06652045e-01 -9.72394407e-01
3.22223723e-01 2.16537476e-01 1.85653508e-01 -5.13916731e-01
8.36527526e-01 -7.45139644e-02 -1.92214206e-01 2.21061364e-01
-7.19710141e-02 -4.09861296e-01 9.46241379e-01 -3.84288222e-01
1.12175870e+00 1.69100508e-01 5.40261805e-01 -1.00387499e-01
3.42386693e-01 4.03987654e-02 4.81113344e-01 9.55156028e-01
-2.26382241e-02 2.72817492e-01 6.35033011e-01 -3.72275710e-01
-1.11704111e+00 -7.90466070e-01 1.48280382e-01 1.32448137e+00
4.27465320e-01 -2.24233210e-01 -7.41496086e-01 -3.48886609e-01
3.11527312e-01 5.74069619e-01 -4.22160417e-01 2.11991463e-02
-5.81970811e-01 -8.66918325e-01 -9.34295058e-02 3.18260074e-01
3.29196602e-01 -1.21502686e+00 -1.14209986e+00 6.31340384e-01
2.27821678e-01 -1.08826423e+00 -1.45635083e-01 7.60548413e-01
-6.99072361e-01 -9.14142549e-01 -6.10758245e-01 -6.07806146e-01
2.49610066e-01 1.25126213e-01 1.07757020e+00 -1.74671218e-01
-4.48693156e-01 -5.86554687e-03 -8.75133216e-01 -6.19269967e-01
-2.28551015e-01 3.20651293e-01 -7.40580410e-02 -3.28962207e-01
1.20360442e-01 -3.95417809e-01 -5.44359565e-01 3.09792370e-01
-9.23934996e-01 1.80334836e-01 8.40113759e-01 9.40498769e-01
6.17279649e-01 2.68810928e-01 4.07276660e-01 -7.41524994e-01
7.94875681e-01 -3.57397884e-01 -9.31969166e-01 -8.36659446e-02
-3.96793514e-01 1.74592108e-01 8.76506209e-01 -4.05842274e-01
-4.42277431e-01 -8.19351971e-02 -2.85947919e-01 -2.55973220e-01
-1.05288364e-01 2.72984684e-01 -4.29167114e-02 5.69162145e-02
3.54211241e-01 1.89990923e-01 2.68091351e-01 -4.33809876e-01
1.38760716e-01 4.45405424e-01 2.01766565e-01 -7.54126370e-01
4.01698589e-01 -1.33971777e-02 3.12461913e-01 -4.04197246e-01
-9.14895713e-01 -6.30102396e-01 -3.20784897e-01 8.45877081e-02
7.51507342e-01 -4.82061118e-01 -1.40399194e+00 2.81088084e-01
-7.20168889e-01 -5.55615008e-01 -5.99547923e-01 2.75532216e-01
-1.23597729e+00 2.68284768e-01 -4.68680412e-01 -9.22097564e-01
-2.57335067e-01 -1.15966964e+00 8.69042754e-01 3.50126922e-01
2.10265800e-01 -4.39758956e-01 9.68291610e-02 5.15370667e-01
3.92125160e-01 3.18297893e-01 7.18827367e-01 -8.07446063e-01
-5.55532396e-01 -1.41890034e-01 6.88254982e-02 2.74302304e-01
-9.81702581e-02 -5.63656151e-01 -2.07163319e-01 -7.25909889e-01
-9.38510969e-02 -4.11168903e-01 5.20725608e-01 3.59362185e-01
1.35885775e+00 -3.78668815e-01 -6.25878796e-02 5.30520320e-01
1.51805210e+00 3.76305014e-01 3.54716510e-01 8.40325236e-01
1.48975477e-01 2.16977969e-01 1.11353815e+00 9.84062552e-01
1.87487692e-01 1.00951850e+00 1.07896101e+00 -9.12231132e-02
3.70634317e-01 2.24988032e-02 6.20590188e-02 3.11638623e-01
-2.80893475e-01 -7.14797497e-01 -5.12852430e-01 3.52879286e-01
-2.33914995e+00 -1.00315571e+00 1.97598308e-01 2.33729434e+00
6.10359907e-01 6.37345552e-01 7.69025266e-01 3.88428807e-01
3.98493558e-01 4.36428841e-03 -5.44835150e-01 -9.67193723e-01
1.63278744e-01 6.26246572e-01 9.01897013e-01 2.45558888e-01
-1.04580772e+00 7.75928378e-01 5.64043760e+00 1.11876869e+00
-6.83946848e-01 -1.81763917e-01 3.74621212e-01 -4.36569512e-01
3.64859134e-01 -4.57826495e-01 -9.68332171e-01 7.67295420e-01
8.45838606e-01 -2.59815484e-01 8.22097957e-01 1.02016318e+00
1.91672087e-01 -4.45192218e-01 -1.04208529e+00 9.53659654e-01
1.41349286e-02 -1.24076772e+00 -2.67486274e-01 4.94843215e-01
6.06713533e-01 -4.13939416e-01 1.13410868e-01 4.88775313e-01
4.04928178e-01 -9.80124235e-01 6.65628850e-01 2.44716927e-01
1.03352658e-01 -1.33485126e+00 1.17543626e+00 7.32942939e-01
-7.56738663e-01 -5.04662991e-01 -6.76094353e-01 -2.90604532e-01
3.28818589e-01 3.80662620e-01 -9.12807047e-01 8.13520551e-01
6.66882277e-01 4.53840047e-02 -2.61732519e-01 1.66852117e+00
4.79249433e-02 1.99729368e-01 -4.04311478e-01 -5.50747871e-01
5.97901106e-01 -3.77601594e-01 5.13561845e-01 1.02167606e+00
3.66389811e-01 1.60726830e-01 6.32202744e-01 5.82795203e-01
2.55617589e-01 2.37743258e-01 -3.30843687e-01 1.23289302e-01
2.37402946e-01 1.14679706e+00 -7.68373489e-01 5.90203963e-02
-7.27487206e-02 9.56609905e-01 6.07000411e-01 -2.07479343e-01
-1.02914965e+00 -1.21307969e-01 6.65054858e-01 1.43174678e-01
9.94145155e-01 -3.41442347e-01 -2.02533789e-03 -5.17141581e-01
-9.12047774e-02 -9.35781538e-01 4.39934105e-01 -2.61482149e-01
-1.25363696e+00 3.91332299e-01 1.13582052e-02 -1.18263638e+00
-4.17367876e-01 -9.01259542e-01 -1.63209334e-01 5.83986878e-01
-1.50114262e+00 -6.57849967e-01 -2.24739052e-02 4.26043689e-01
6.16027951e-01 -1.84231430e-01 7.21766412e-01 1.65505737e-01
-5.59993744e-01 6.28060699e-01 2.19158068e-01 -5.15712023e-01
2.28034571e-01 -1.58709526e+00 1.30372092e-01 5.66130459e-01
-1.88217074e-01 -2.32502203e-02 1.01216853e+00 -3.13288569e-01
-1.75946331e+00 -7.32482791e-01 1.31163016e-01 2.72665881e-02
5.73504031e-01 -4.23855513e-01 -4.98389035e-01 1.10500246e-01
6.36691004e-02 -4.54451926e-02 6.78513825e-01 8.55118930e-02
4.77726936e-01 -1.79764003e-01 -1.21263051e+00 5.07972538e-01
1.02898884e+00 4.71844435e-01 -4.65904206e-01 3.26324373e-01
7.67078221e-01 -9.47784662e-01 -7.76234329e-01 2.84841359e-01
1.45318210e-01 -1.23005986e+00 9.06631947e-01 -7.51172483e-01
3.97828281e-01 -6.91717565e-02 -1.38237715e-01 -1.45341861e+00
-5.88152468e-01 -7.05598056e-01 -5.00869930e-01 8.71891022e-01
1.72492892e-01 -3.64655912e-01 1.29552400e+00 4.94998038e-01
-2.41279483e-01 -1.26578200e+00 -1.04868805e+00 -1.20534873e+00
-1.01059675e-01 -2.60452390e-01 7.57760584e-01 1.26795232e-01
1.51667327e-01 1.27513871e-01 -6.56907558e-01 -2.08449569e-02
4.09538537e-01 5.30650079e-01 9.71572101e-01 -8.70839655e-01
-8.80322874e-01 -6.00765288e-01 -4.55619574e-01 -1.16796243e+00
-1.65775195e-01 -6.59728467e-01 1.27270743e-01 -1.51238430e+00
1.73548624e-01 -5.44792473e-01 -2.87145197e-01 3.92697334e-01
-5.66026345e-02 8.31234604e-02 5.99932373e-01 -3.60199660e-01
-9.54018831e-01 5.07677317e-01 1.46621001e+00 7.43077621e-02
-4.40094352e-01 4.14711535e-01 -5.42738438e-01 2.55422860e-01
8.51987481e-01 -3.46085906e-01 -1.11529544e-01 -2.69602891e-02
2.48777270e-01 3.05525303e-01 -7.14125857e-02 -9.78785098e-01
7.29847550e-02 -3.54443669e-01 1.71834379e-01 -6.87144339e-01
4.43049580e-01 -1.05828559e+00 -3.29832919e-02 7.30161786e-01
8.17580894e-03 1.95353895e-01 3.23381573e-01 5.77536225e-01
3.36464979e-02 -6.66533232e-01 6.55415058e-01 -2.38846257e-01
-5.62044621e-01 3.99887025e-01 -3.51475567e-01 -5.35805710e-02
1.53011227e+00 -4.19519871e-01 -1.27766222e-01 6.60512298e-02
-5.99268019e-01 5.88610470e-01 2.38617286e-01 2.50645518e-01
2.15616181e-01 -1.15910256e+00 -5.82585692e-01 1.61059916e-01
-1.55865371e-01 1.03011951e-01 3.25540513e-01 6.02946222e-01
-7.84008920e-01 5.00759900e-01 -6.76708698e-01 -5.00030577e-01
-1.17115867e+00 1.13863504e+00 8.50923359e-03 -1.14878166e+00
-5.38457334e-01 6.81311607e-01 -4.06944633e-01 -2.21892640e-01
3.40468854e-01 -2.18627527e-01 -1.61495447e-01 1.20892778e-01
3.57767433e-01 7.62189448e-01 3.45458508e-01 3.93729955e-02
-1.28302470e-01 1.65480196e-01 -9.91988555e-02 1.56930089e-01
1.82210779e+00 1.84385926e-01 2.06433490e-01 -1.61881521e-01
6.67118847e-01 -2.56764919e-01 -1.21961892e+00 6.52114600e-02
-1.11461215e-01 -7.27873802e-01 -2.05707371e-01 -7.59577274e-01
-8.59741271e-01 4.35096949e-01 5.49160719e-01 8.32517624e-01
1.23273456e+00 -1.78460926e-01 8.60363126e-01 2.51294225e-01
8.07539344e-01 -1.25774360e+00 1.92461044e-01 6.08795047e-01
6.30647898e-01 -1.06792808e+00 1.02097996e-01 -2.73470372e-01
-6.52835846e-01 9.95206833e-01 7.43014574e-01 -4.11853731e-01
1.21456258e-01 4.86576945e-01 -6.61578715e-01 9.51314997e-03
-7.28334904e-01 -6.55082941e-01 -9.92166158e-03 5.50556600e-01
-2.11020231e-01 2.69614547e-01 -6.82543695e-01 6.32278979e-01
-5.55750251e-01 -2.17213377e-01 4.17961389e-01 1.10741246e+00
-6.63883090e-01 -1.37735903e+00 -4.30306226e-01 5.80902338e-01
-4.04656976e-01 2.90919781e-01 1.84953377e-01 8.58131945e-01
1.67454287e-01 8.25704038e-01 1.08070180e-01 -1.39169574e-01
6.00566685e-01 -2.86507398e-01 8.77476692e-01 -6.45599663e-01
-6.49922669e-01 -4.94702952e-03 4.50248979e-02 -7.24770725e-01
-2.57960975e-01 -4.75887746e-01 -1.00512218e+00 -4.98670131e-01
-4.00367588e-01 2.22246021e-01 6.13810420e-01 9.95028496e-01
2.03102529e-01 7.65238345e-01 6.37596726e-01 -1.09278584e+00
-8.77322793e-01 -5.85040689e-01 -1.01487505e+00 3.88183147e-01
4.98748310e-02 -9.33607042e-01 -1.01545462e-02 -6.29336774e-01] | [4.928703308105469, 2.6695752143859863] |
ae1bf041-cf54-4cd2-a5ae-215f809d08e0 | efficient-neural-net-approaches-in-metal | 2208.04150 | null | https://arxiv.org/abs/2208.04150v1 | https://arxiv.org/pdf/2208.04150v1.pdf | Efficient Neural Net Approaches in Metal Casting Defect Detection | One of the most pressing challenges prevalent in the steel manufacturing industry is the identification of surface defects. Early identification of casting defects can help boost performance, including streamlining production processes. Though, deep learning models have helped bridge this gap and automate most of these processes, there is a dire need to come up with lightweight models that can be deployed easily with faster inference times. This research proposes a lightweight architecture that is efficient in terms of accuracy and inference time compared with sophisticated pre-trained CNN architectures like MobileNet, Inception, and ResNet, including vision transformers. Methodologies to minimize computational requirements such as depth-wise separable convolution and global average pooling (GAP) layer, including techniques that improve architectural efficiencies and augmentations, have been experimented. Our results indicate that a custom model of 590K parameters with depth-wise separable convolutions outperformed pretrained architectures such as Resnet and Vision transformers in terms of accuracy (81.87%) and comfortably outdid architectures such as Resnet, Inception, and Vision transformers in terms of faster inference times (12 ms). Blurpool fared outperformed other techniques, with an accuracy of 83.98%. Augmentations had a paradoxical effect on the model performance. No direct correlation between depth-wise and 3x3 convolutions on inference time, they, however, they played a direct role in improving model efficiency by enabling the networks to go deeper and by decreasing the number of trainable parameters. Our work sheds light on the fact that custom networks with efficient architectures and faster inference times can be built without the need of relying on pre-trained architectures. | ['Sabeesh Ethiraj', 'Bharath Kumar Bolla', 'Rohit Lal'] | 2022-08-08 | null | null | null | null | ['defect-detection'] | ['computer-vision'] | [-2.96284765e-01 1.93498611e-01 5.05910397e-01 -2.61938334e-01
-3.91407646e-02 -1.91560969e-01 3.53109807e-01 -1.52767375e-01
-5.39583385e-01 1.42251328e-01 -1.51078105e-01 -6.99057996e-01
-2.66528666e-01 -1.03392339e+00 -5.80844939e-01 -5.92281640e-01
1.95419267e-01 2.52595603e-01 3.54460537e-01 -3.77214551e-01
5.04373729e-01 6.97986305e-01 -1.54731965e+00 4.36051458e-01
4.87847865e-01 1.16452169e+00 5.75550914e-01 6.42764986e-01
-6.96931556e-02 9.29965258e-01 -7.76636600e-01 -4.33518797e-01
1.92085549e-01 2.37103373e-01 -7.83988357e-01 -2.55498916e-01
2.59855419e-01 -6.21447980e-01 -5.33750176e-01 5.58821380e-01
7.18887985e-01 -3.45309712e-02 2.21904576e-01 -8.50281656e-01
-8.55439782e-01 6.24047816e-01 -3.88475895e-01 3.09723794e-01
-1.85109407e-01 4.73021805e-01 5.91013014e-01 -1.07764840e+00
2.05119088e-01 1.09232223e+00 1.10266173e+00 3.81521583e-01
-9.23550069e-01 -6.05419874e-01 -1.29327223e-01 2.67824411e-01
-1.19817221e+00 -4.39910293e-01 3.82095277e-01 -3.10406655e-01
1.49916220e+00 1.80471495e-01 5.53781509e-01 7.30617583e-01
3.29254627e-01 3.38169038e-01 7.69054174e-01 -4.78756100e-01
2.10783988e-01 9.52633321e-02 1.09812170e-01 9.43750679e-01
3.46741557e-01 -2.16380894e-01 -3.48040283e-01 3.49846810e-01
1.17964149e+00 2.38087475e-01 -3.52835134e-02 2.39697516e-01
-9.14933503e-01 7.55163729e-01 4.58705306e-01 5.32222271e-01
-6.04258239e-01 5.32100499e-01 5.46874285e-01 3.49509746e-01
5.42255044e-01 9.69055355e-01 -6.93971217e-01 -2.30207339e-01
-8.36175978e-01 -8.76276195e-02 6.58967376e-01 6.78505301e-01
7.74066746e-01 3.84390891e-01 -8.74330327e-02 1.11994612e+00
4.68769036e-02 2.81576216e-01 4.71399337e-01 -1.13593578e+00
3.18202615e-01 8.66178691e-01 -2.63229311e-01 -9.62757826e-01
-4.51311260e-01 -4.16325718e-01 -7.05909669e-01 6.17046714e-01
4.71636325e-01 -2.17904240e-01 -1.42927277e+00 1.07547188e+00
-1.06495969e-01 -2.24403664e-02 -1.98739082e-01 7.95213699e-01
5.72299004e-01 5.75377405e-01 -1.51793018e-01 5.89968085e-01
1.48415208e+00 -1.07875180e+00 -5.19606352e-01 -3.16276580e-01
5.67804098e-01 -1.08231878e+00 1.23272049e+00 3.87439042e-01
-1.32942855e+00 -7.49344170e-01 -1.30535352e+00 -2.97010392e-01
-5.53315699e-01 2.95782149e-01 1.08890581e+00 7.36539543e-01
-1.34566557e+00 1.03202879e+00 -1.07304060e+00 -3.13634545e-01
5.73517561e-01 7.27409840e-01 -2.78860569e-01 -2.89224923e-01
-6.79682851e-01 1.26246119e+00 2.95760393e-01 6.32370234e-01
-8.48679006e-01 -9.42157030e-01 -6.43775165e-01 2.80051470e-01
2.91361630e-01 -7.62276232e-01 1.11525548e+00 -6.69483066e-01
-1.60826945e+00 4.69657272e-01 1.53597564e-01 -5.43739796e-01
2.85993010e-01 -4.76395011e-01 -1.77325845e-01 1.72359362e-01
-1.65149391e-01 5.85433781e-01 6.41717374e-01 -6.71700418e-01
-6.65303707e-01 -6.06498420e-02 4.60245997e-01 -2.25274161e-01
-6.38193369e-01 5.34023494e-02 -3.25792760e-01 -3.62618357e-01
2.40694702e-01 -8.51300001e-01 -3.42393696e-01 -9.48760286e-02
-1.37403533e-01 -2.12351114e-01 1.01535058e+00 -9.52138364e-01
9.30484653e-01 -1.96501863e+00 -5.75697720e-01 7.56345615e-02
3.32737625e-01 7.30119407e-01 3.26912403e-02 3.57949197e-01
-1.87948465e-01 1.46073535e-01 1.36305973e-01 -2.77569711e-01
-2.40762666e-01 3.32707852e-01 1.48419932e-01 2.63491541e-01
5.33433795e-01 8.57489347e-01 -5.46336293e-01 -5.01867719e-02
6.34937942e-01 6.18416011e-01 -7.10942328e-01 4.69491854e-02
7.90666938e-02 9.64051951e-03 -1.18897595e-02 6.88751936e-01
6.36218309e-01 -2.99183428e-01 -2.40942001e-01 -2.93168753e-01
-2.00791821e-01 3.15319568e-01 -9.56269026e-01 1.63897872e+00
-9.38520372e-01 8.87847245e-01 1.71367183e-01 -1.09248173e+00
9.34612095e-01 4.58887637e-01 3.38049203e-01 -8.19940448e-01
2.95542032e-01 2.61099696e-01 1.84308648e-01 -6.45442903e-01
6.68782055e-01 1.24558821e-01 4.61551160e-01 1.56614512e-01
6.88308030e-02 -8.53792429e-02 2.33217716e-01 -1.36671901e-01
1.45455003e+00 -2.52148975e-02 -5.24606407e-01 -3.11519772e-01
2.71764457e-01 -2.85320491e-01 2.46849060e-01 6.00383520e-01
4.43511963e-01 7.86734104e-01 2.62470514e-01 -8.03426921e-01
-1.21470964e+00 -8.30476224e-01 1.59121946e-01 7.64840364e-01
-1.90424293e-01 -3.98711056e-01 -6.51143134e-01 -2.92264462e-01
-1.05403885e-01 4.70485598e-01 -3.77471328e-01 -1.85933292e-01
-7.33873725e-01 -6.70066655e-01 6.32525682e-01 9.53120708e-01
8.81016910e-01 -9.01550829e-01 -7.78836846e-01 3.20990086e-01
2.64286697e-01 -1.21008873e+00 1.24280520e-01 3.32144141e-01
-1.12899005e+00 -1.03272700e+00 -6.92529738e-01 -7.20665753e-01
8.94518137e-01 1.75539359e-01 9.76160526e-01 4.97622341e-01
-5.14776289e-01 5.32922149e-02 -1.05405904e-01 -5.86589932e-01
-2.90702701e-01 2.57922292e-01 -2.61173874e-01 -5.69033980e-01
3.02722037e-01 -5.65804422e-01 -7.95041859e-01 3.12622428e-01
-8.05152953e-01 2.48680711e-01 1.02564883e+00 8.48752201e-01
6.23377785e-02 2.50198960e-01 4.45988417e-01 -8.32420290e-01
6.18179798e-01 -2.30243489e-01 -3.82666469e-01 7.99963623e-03
-1.10116553e+00 4.14894074e-02 5.95440209e-01 -1.61530599e-01
-1.15375543e+00 -2.60285258e-01 -5.82190692e-01 -4.44115639e-01
6.89819502e-03 4.92108971e-01 4.33824569e-01 -2.38062173e-01
5.54534018e-01 -3.70189071e-01 3.31605136e-01 -6.24156296e-01
1.05729729e-01 6.70903683e-01 4.21059817e-01 -5.10596395e-01
5.48108459e-01 3.81517351e-01 -1.11479819e-01 -9.86036420e-01
-3.36285293e-01 -2.51152247e-01 -4.05200124e-01 -9.88722593e-02
8.67846787e-01 -9.62720275e-01 -8.07204604e-01 9.06919122e-01
-1.16562414e+00 -3.55848521e-01 -3.43787611e-01 3.24016392e-01
-7.89914727e-02 1.22224696e-01 -1.17821884e+00 -6.26717865e-01
-5.48033834e-01 -1.12429273e+00 6.29558682e-01 3.39935243e-01
-2.66314626e-01 -8.77001345e-01 -7.12152719e-01 5.99053800e-01
1.09144080e+00 -8.18982720e-02 9.86732841e-01 -4.65495080e-01
-6.55177772e-01 -4.40444112e-01 -4.44063306e-01 1.02784467e+00
1.73277542e-01 9.17158127e-02 -1.14099932e+00 -1.48404092e-01
4.56431583e-02 4.55609933e-02 7.16831982e-01 4.38311249e-01
1.45411527e+00 -1.12572148e-01 -8.70490521e-02 4.46859270e-01
1.41346228e+00 3.72341812e-01 1.07352614e+00 4.41329807e-01
7.27223396e-01 3.16074282e-01 1.47566989e-01 6.70580715e-02
2.65478879e-01 2.11074397e-01 5.53915679e-01 -5.21928847e-01
-5.33179581e-01 8.76439456e-03 7.26663619e-02 9.38463926e-01
-6.13116205e-01 8.28491077e-02 -1.11631322e+00 6.00482941e-01
-1.49877191e+00 -5.92104554e-01 -1.51416868e-01 1.81615698e+00
4.90867376e-01 6.12304986e-01 -4.34347570e-01 4.13858682e-01
5.52322030e-01 -2.10945696e-01 -2.42469490e-01 -7.31154621e-01
2.33133584e-01 8.84395182e-01 8.51224899e-01 1.38874322e-01
-5.91594100e-01 7.98602045e-01 6.50812340e+00 7.86354780e-01
-1.40620601e+00 1.58383608e-01 6.96717441e-01 -9.35882404e-02
1.21284746e-01 -1.33140445e-01 -6.11769497e-01 4.93005842e-01
1.25509632e+00 4.70607132e-01 3.74850512e-01 1.09804523e+00
6.70093969e-02 -2.53821254e-01 -1.09974241e+00 7.86209047e-01
-2.14135543e-01 -1.79182875e+00 -2.95092702e-01 4.45025042e-02
5.54102719e-01 1.27273500e-01 -5.45482934e-02 3.97849679e-01
3.66265297e-01 -1.08448493e+00 6.17658854e-01 3.47601533e-01
5.43431997e-01 -7.36692131e-01 1.26944768e+00 -6.49022833e-02
-8.96251500e-01 -4.13267493e-01 -5.31714261e-01 -6.96728110e-01
2.65467409e-02 8.37586761e-01 -1.16569555e+00 4.22989309e-01
1.14645123e+00 2.83934325e-01 -5.68059683e-01 9.15222526e-01
-1.63037434e-01 6.24374270e-01 -3.12419504e-01 1.53790298e-03
2.64253914e-01 1.95279240e-03 -2.14699641e-01 1.03762317e+00
3.63553584e-01 -1.86904490e-01 -3.74805540e-01 8.61266732e-01
7.93550014e-02 -5.81328630e-01 -3.73055160e-01 -8.84073451e-02
5.57040036e-01 1.26944387e+00 -1.00613654e+00 -1.58040166e-01
-5.39162636e-01 7.14814126e-01 -1.08888382e-02 1.82366624e-01
-9.43780601e-01 -6.01955533e-01 7.41924644e-01 4.22817916e-01
4.92630869e-01 -2.99906045e-01 -8.12462866e-01 -4.58781928e-01
-4.87705842e-02 -6.68210208e-01 -5.79899922e-02 -8.63709688e-01
-9.40043330e-01 4.34930444e-01 -4.07399118e-01 -5.18327951e-01
9.97149944e-02 -9.58289981e-01 -7.01469660e-01 9.33522642e-01
-1.34346366e+00 -1.21286976e+00 -4.42296207e-01 2.83934653e-01
7.57446110e-01 -7.96385333e-02 7.05969155e-01 7.47297049e-01
-7.31053412e-01 5.33304036e-01 -2.02294663e-01 2.30606794e-01
2.95732379e-01 -1.05781531e+00 6.02253795e-01 9.41270888e-01
-2.11735383e-01 9.09307420e-01 5.05603254e-01 -2.38763303e-01
-1.43966174e+00 -7.97427416e-01 4.83543336e-01 -3.34534407e-01
4.50203657e-01 -2.57247418e-01 -7.08174586e-01 6.23168647e-01
2.00260356e-01 -1.16939031e-01 1.38093993e-01 3.06478500e-01
7.99241848e-03 -2.86991537e-01 -1.07867229e+00 5.04603922e-01
9.11738038e-01 -6.67140365e-01 -2.76710749e-01 1.57656357e-01
6.50372148e-01 -5.43072224e-01 -9.73161042e-01 4.24520761e-01
5.44189155e-01 -9.89076376e-01 1.13480425e+00 -2.11494535e-01
8.66687417e-01 5.76174520e-02 1.11132398e-01 -1.08746588e+00
-4.72082824e-01 -2.45593920e-01 -1.10976666e-01 1.25633109e+00
5.85294068e-01 -1.01730740e+00 1.05898201e+00 1.01082170e+00
-7.65731394e-01 -1.20563447e+00 -6.48678124e-01 -7.20559359e-01
-7.14094639e-02 -4.09907013e-01 6.41059697e-01 5.92116416e-01
-4.65848982e-01 1.98169976e-01 8.93388838e-02 1.80224642e-01
-1.33387938e-01 -4.84755337e-01 6.03560925e-01 -1.11043227e+00
-6.66787773e-02 -4.43824381e-01 -5.58641732e-01 -7.80389071e-01
-3.47951859e-01 -4.88242865e-01 -6.70091063e-02 -1.79129481e+00
-3.18683326e-01 -7.29362607e-01 -2.80962795e-01 7.37227142e-01
1.13677196e-01 2.40232319e-01 1.44102639e-02 -1.01128623e-01
2.63777245e-02 -2.06597950e-02 1.07935047e+00 -1.01696484e-01
1.84888795e-01 -1.76556826e-01 -8.20420623e-01 8.87082934e-01
8.44992042e-01 -1.65206879e-01 -5.17729640e-01 -1.03162313e+00
2.48206243e-01 -3.29944134e-01 4.07085389e-01 -1.41839492e+00
4.74589765e-01 4.16283965e-01 5.94821393e-01 -4.25229460e-01
5.43673754e-01 -8.16087484e-01 2.93526262e-01 7.06236780e-01
2.19990492e-01 3.55980575e-01 4.30169463e-01 -6.92629963e-02
-8.92763436e-02 -5.14612198e-01 6.36818945e-01 -3.18144321e-01
-9.37184811e-01 -3.45424786e-02 -3.48426938e-01 -4.23236012e-01
9.16533589e-01 -7.59017885e-01 -5.15635252e-01 7.38426298e-02
-4.68395889e-01 1.45996222e-03 3.61262321e-01 4.28918093e-01
6.57702804e-01 -8.29965413e-01 -3.98784280e-01 3.75152320e-01
-5.03988087e-01 3.25696617e-01 4.25522089e-01 8.22920322e-01
-1.21053445e+00 3.91284138e-01 -3.58252764e-01 -5.32103837e-01
-1.00924814e+00 1.89665437e-01 2.90213704e-01 -2.07420379e-01
-7.41189063e-01 1.29078472e+00 -1.83486685e-01 -2.26818562e-01
2.21885011e-01 -6.07589662e-01 1.66069061e-01 -1.51938617e-01
4.47340488e-01 9.13489878e-01 7.96876669e-01 1.80904984e-01
-1.89028457e-01 2.93827116e-01 -4.54081208e-01 3.78581613e-01
1.75763106e+00 1.26224101e-01 -2.77074605e-01 9.48151425e-02
1.00489140e+00 -2.72888631e-01 -1.20144296e+00 1.60189882e-01
9.65241566e-02 -4.78612393e-01 3.34134877e-01 -9.70557988e-01
-1.67236543e+00 7.74242520e-01 8.32288742e-01 2.57016361e-01
1.07003760e+00 -1.79311529e-01 1.02190638e+00 3.82684499e-01
3.51322711e-01 -1.35341704e+00 3.02964270e-01 5.71805835e-01
6.20343029e-01 -1.02367914e+00 -1.32394344e-01 -4.45511371e-01
-8.99990425e-02 1.30633688e+00 8.12377751e-01 -3.55412483e-01
6.25476956e-01 7.47071266e-01 -3.72883864e-02 -7.52215028e-01
-5.80671191e-01 1.17869750e-01 -2.55736649e-01 5.66055000e-01
6.22840583e-01 -2.36001506e-01 -8.02225322e-02 4.26916122e-01
-3.66774589e-01 2.26335928e-01 4.61630136e-01 1.08812916e+00
-3.36672634e-01 -7.55343676e-01 -2.63209403e-01 7.12474167e-01
-8.04279208e-01 -1.80930048e-01 2.40058005e-01 9.15732086e-01
3.69266063e-01 1.08118415e+00 2.51515865e-01 -6.30139947e-01
5.56590438e-01 -6.29221350e-02 3.72803748e-01 -6.71793044e-01
-8.83710623e-01 -5.76468825e-01 3.64212692e-01 -6.11995280e-01
3.48072574e-02 -1.66508570e-01 -1.19987237e+00 -8.80631983e-01
-7.87703753e-01 -1.87692732e-01 1.07563221e+00 9.47427630e-01
4.96360242e-01 1.20880949e+00 2.82721758e-01 -9.53010261e-01
-5.14179409e-01 -1.13564503e+00 -3.34268570e-01 -4.36494015e-02
-1.54287238e-02 -4.50639248e-01 -1.47945672e-01 -1.36974556e-02] | [7.830989360809326, 2.045464515686035] |
55c55c8d-9ce9-45b1-822f-08e447a29730 | generative-aspect-based-sentiment-analysis | 2211.07743 | null | https://arxiv.org/abs/2211.07743v1 | https://arxiv.org/pdf/2211.07743v1.pdf | Generative Aspect-Based Sentiment Analysis with Contrastive Learning and Expressive Structure | Generative models have demonstrated impressive results on Aspect-based Sentiment Analysis (ABSA) tasks, particularly for the emerging task of extracting Aspect-Category-Opinion-Sentiment (ACOS) quadruples. However, these models struggle with implicit sentiment expressions, which are commonly observed in opinionated content such as online reviews. In this work, we introduce GEN-SCL-NAT, which consists of two techniques for improved structured generation for ACOS quadruple extraction. First, we propose GEN-SCL, a supervised contrastive learning objective that aids quadruple prediction by encouraging the model to produce input representations that are discriminable across key input attributes, such as sentiment polarity and the existence of implicit opinions and aspects. Second, we introduce GEN-NAT, a new structured generation format that better adapts autoregressive encoder-decoder models to extract quadruples in a generative fashion. Experimental results show that GEN-SCL-NAT achieves top performance across three ACOS datasets, averaging 1.48% F1 improvement, with a maximum 1.73% increase on the LAPTOP-L1 dataset. Additionally, we see significant gains on implicit aspect and opinion splits that have been shown as challenging for existing ACOS approaches. | ['Lu Wang', 'Joseph J. Peper'] | 2022-11-14 | null | null | null | null | ['aspect-based-sentiment-analysis'] | ['natural-language-processing'] | [ 4.41530824e-01 4.74091440e-01 -2.42391571e-01 -7.69624949e-01
-1.29172182e+00 -7.03817546e-01 5.64946055e-01 1.06452471e-02
1.27838969e-01 6.14236534e-01 5.00995815e-01 -3.76451969e-01
4.31067228e-01 -8.39607835e-01 -6.80658519e-01 -4.78403687e-01
2.58288413e-01 5.43615222e-01 -5.91933370e-01 -6.18564427e-01
1.12035982e-01 -2.64531195e-01 -1.46458757e+00 8.40121090e-01
9.86178756e-01 1.24336731e+00 -5.25141656e-01 7.17643023e-01
-2.31733009e-01 8.55610967e-01 -9.10899580e-01 -1.20758796e+00
1.12064660e-01 -6.33295476e-01 -5.87325037e-01 5.24428561e-02
5.45646191e-01 -1.98361138e-03 3.70217085e-01 6.35024250e-01
4.99963403e-01 -3.62842053e-01 9.87659514e-01 -1.30908275e+00
-8.69051933e-01 9.72273529e-01 -6.24134123e-01 -2.63817221e-01
2.74007887e-01 2.09777117e-01 1.74455976e+00 -1.15006435e+00
4.70482439e-01 1.42804825e+00 8.10821414e-01 4.85762477e-01
-1.22439301e+00 -6.25881195e-01 4.85754281e-01 -2.38430560e-01
-8.06550682e-01 -4.27185744e-01 7.18479097e-01 -2.12604716e-01
1.20002723e+00 2.70456523e-01 7.35843718e-01 1.32074893e+00
4.71632540e-01 1.41196382e+00 1.24660707e+00 -1.95697188e-01
2.82779306e-01 3.50268066e-01 2.68747330e-01 3.13378811e-01
4.26176637e-01 -2.84073234e-01 -9.83831167e-01 -5.62576670e-03
-2.44943388e-02 -3.59891295e-01 7.58840069e-02 1.30149983e-02
-8.87360632e-01 1.13443291e+00 2.96364248e-01 -1.74783245e-01
-4.22557622e-01 -1.31861731e-01 3.50486159e-01 3.91435891e-01
1.05045903e+00 7.16956854e-01 -9.61879373e-01 -2.74484098e-01
-8.39464903e-01 3.85140598e-01 1.10466194e+00 1.16258311e+00
7.44198680e-01 4.97516751e-01 -3.33594084e-01 8.22501242e-01
2.30833024e-01 9.82826173e-01 5.27165055e-01 -4.26518202e-01
6.14375293e-01 9.85020936e-01 -2.94207811e-01 -8.92520607e-01
-3.40069920e-01 -8.21199894e-01 -9.48395133e-01 -7.65107125e-02
-1.58883899e-01 -5.13965070e-01 -1.26547647e+00 1.58632469e+00
1.34947702e-01 -3.31769854e-01 4.55916792e-01 3.23506385e-01
9.76125121e-01 5.66946149e-01 6.19448088e-02 3.44873369e-02
1.48445010e+00 -1.17887783e+00 -6.44271076e-01 -8.12420428e-01
6.98461831e-01 -7.72567630e-01 1.10566056e+00 4.21598017e-01
-1.03552532e+00 -3.76787961e-01 -1.12578070e+00 -2.52835061e-02
-6.81514025e-01 1.96698636e-01 1.04134846e+00 8.24272156e-01
-1.04060662e+00 1.29866973e-01 -5.58420539e-01 3.19724470e-01
5.87334812e-01 3.63464862e-01 -1.57401577e-01 1.09885111e-01
-9.95023608e-01 3.80258441e-01 1.71886161e-02 6.28579631e-02
-4.22650695e-01 -9.96982515e-01 -1.31920135e+00 7.10074306e-02
3.65649372e-01 -1.11872458e+00 1.41108823e+00 -1.27722609e+00
-1.43442643e+00 6.59378529e-01 -6.43393099e-01 -4.94560272e-01
2.70529855e-02 -4.89042312e-01 -5.87222934e-01 -2.87341684e-01
4.79474366e-01 7.87606180e-01 1.00057459e+00 -1.30300975e+00
-5.37176490e-01 -4.23846304e-01 1.52835444e-01 3.76143754e-01
-4.33183402e-01 -3.19473386e-01 -1.66515812e-01 -8.49362493e-01
-1.37407601e-01 -1.05068457e+00 -4.58934665e-01 -8.22502673e-01
-7.07666516e-01 -2.27011666e-01 5.35504222e-01 -4.20281589e-01
1.17317557e+00 -1.97145438e+00 -4.73009534e-02 9.82890055e-02
6.34130687e-02 1.87596560e-01 -4.02797401e-01 2.58326441e-01
-1.10019170e-01 3.61655354e-01 -4.29684371e-01 -8.28489363e-01
2.53209174e-01 -2.61143800e-02 -6.97395325e-01 -3.76005650e-01
8.66938472e-01 1.37294781e+00 -6.49916291e-01 -1.76385999e-01
-4.75820690e-01 4.47841525e-01 -9.35018957e-01 2.34455854e-01
-5.18120050e-01 4.94325347e-03 -3.22792739e-01 9.71303344e-01
6.02289140e-01 -4.50991213e-01 7.31981769e-02 -1.68646306e-01
2.70981222e-01 8.64836574e-01 -8.26988339e-01 1.21278477e+00
-7.46353745e-01 4.99215931e-01 -2.40978509e-01 -4.00117159e-01
1.03755927e+00 1.79961786e-01 1.94990128e-01 -7.11141109e-01
2.22385317e-01 2.89532393e-01 -1.27972692e-01 2.08959043e-01
1.05160224e+00 -1.97440013e-01 -5.71259260e-01 5.29448628e-01
2.99427211e-01 -8.40413868e-01 6.18320346e-01 4.27353501e-01
7.68601000e-01 -1.41466614e-02 3.20580900e-01 -1.01956517e-01
4.32424068e-01 1.26409844e-01 4.89901900e-01 6.73903048e-01
3.26359212e-01 9.58954871e-01 9.52553689e-01 -3.03290874e-01
-7.60484040e-01 -7.81658709e-01 1.97328497e-02 1.04737759e+00
-2.57240653e-01 -1.01439810e+00 -5.01748085e-01 -1.10482514e+00
-1.27293006e-01 9.43455100e-01 -8.78704906e-01 -2.24311948e-01
-2.84149617e-01 -1.23204863e+00 1.09096043e-01 7.84385622e-01
2.06415892e-01 -1.11974871e+00 1.18178181e-01 1.19252339e-01
-1.63501292e-01 -1.21674180e+00 -4.79885787e-01 4.32353079e-01
-7.32000291e-01 -7.73437500e-01 -4.45130825e-01 -6.23524904e-01
7.51811266e-01 2.05059140e-03 1.85893595e+00 -3.30153018e-01
9.82117653e-02 2.14483157e-01 -5.69830716e-01 -9.17061806e-01
-3.14724207e-01 6.84056163e-01 -3.45296144e-01 2.19296157e-01
7.29436517e-01 -5.39753914e-01 -6.22017920e-01 2.79484806e-03
-8.61002922e-01 2.45612159e-01 9.87993598e-01 9.27779615e-01
8.03538442e-01 -3.71316642e-01 7.92280614e-01 -1.64727736e+00
8.20681751e-01 -6.52067006e-01 -1.26196042e-01 -7.83409029e-02
-1.07825863e+00 1.33954093e-01 7.62603939e-01 -1.15236029e-01
-9.51751351e-01 -2.44298533e-01 -4.41146672e-01 1.50130510e-01
1.63247027e-02 7.91187525e-01 -1.66213408e-01 5.77424109e-01
4.67332393e-01 3.54137361e-01 -3.13264996e-01 -1.43030301e-01
5.76952517e-01 8.12789679e-01 2.97223151e-01 -3.77747506e-01
6.84088409e-01 2.14727297e-01 -2.40116432e-01 -5.88717937e-01
-1.55245829e+00 -4.06241536e-01 -3.19943070e-01 2.10006446e-01
5.95876753e-01 -1.48954606e+00 -2.61937439e-01 6.19283199e-01
-9.36600149e-01 -1.91868767e-01 -5.24911344e-01 -1.81773752e-02
-3.28571290e-01 -1.67175725e-01 -5.22722125e-01 -7.30350196e-01
-1.14094520e+00 -1.07599556e+00 1.64265454e+00 4.43537205e-01
-6.25611246e-01 -9.39420581e-01 1.96702108e-01 6.50572181e-01
5.95871866e-01 1.68309033e-01 8.73729527e-01 -8.97531986e-01
-4.58420724e-01 -1.91601232e-01 1.25874683e-01 6.66335046e-01
1.93943694e-01 -1.04061969e-01 -1.06018412e+00 -2.11233720e-01
-2.03054801e-01 -4.79788363e-01 9.34623778e-01 1.20653763e-01
8.35732520e-01 -6.68686807e-01 1.28950894e-01 6.74254596e-01
1.09082627e+00 1.39643341e-01 6.00194812e-01 1.97000384e-01
7.60702968e-01 3.21148485e-01 6.67575657e-01 2.61318088e-01
8.10751140e-01 2.69674450e-01 2.68892705e-01 -3.06550443e-01
-2.55994052e-01 -3.31896186e-01 8.11357856e-01 1.39000702e+00
2.63764501e-01 -4.00676608e-01 -4.70676512e-01 8.87599409e-01
-1.50697613e+00 -5.87982535e-01 -2.47993723e-01 1.62845898e+00
1.16241705e+00 3.74188453e-01 4.10125591e-03 -3.26051414e-02
8.93130600e-02 5.89850962e-01 -6.78769708e-01 -9.56140757e-01
-5.24144828e-01 7.80888557e-01 6.64802715e-02 3.68801057e-01
-1.18475413e+00 9.79584217e-01 5.97164917e+00 7.23409712e-01
-1.05616438e+00 -1.37146741e-01 1.16423059e+00 -9.55097601e-02
-1.01500010e+00 -3.28548476e-02 -1.11851406e+00 4.18878078e-01
1.00674081e+00 -2.69820035e-01 -1.17767230e-01 1.24060726e+00
-2.43303239e-01 1.26160949e-01 -9.48229730e-01 6.53826356e-01
5.64951777e-01 -1.06093049e+00 6.43416345e-01 -1.40435752e-02
1.45991635e+00 -4.82276604e-02 5.14157176e-01 7.91183114e-01
4.19869334e-01 -1.10697615e+00 5.49506903e-01 7.12096617e-02
7.66924083e-01 -9.96605396e-01 1.10772705e+00 -2.34156325e-01
-1.16140091e+00 1.36592269e-01 -3.37875858e-02 8.76653790e-02
2.62714952e-01 9.33830738e-01 -9.95101094e-01 7.29087830e-01
4.59625393e-01 1.07682908e+00 -8.29170465e-01 4.58995253e-01
-6.29014671e-01 8.73511493e-01 -2.45767329e-02 -2.48913094e-01
3.60616952e-01 -2.17476606e-01 5.38187623e-01 1.27414417e+00
2.73375899e-01 -2.72649109e-01 -2.43195459e-01 6.90316379e-01
-4.34774846e-01 2.47095227e-01 -6.26264513e-01 -4.55977678e-01
-4.80877049e-02 1.46580458e+00 -3.79974544e-01 -5.13450682e-01
-3.11727554e-01 7.84454703e-01 7.58926496e-02 2.72033334e-01
-6.27388835e-01 -3.66508424e-01 9.07315552e-01 -2.27820836e-02
6.89110816e-01 7.71278664e-02 -6.47759497e-01 -1.39972854e+00
1.80737078e-01 -1.57097006e+00 3.01137209e-01 -8.21095347e-01
-1.33135581e+00 1.10114551e+00 -3.05820286e-01 -1.13802648e+00
-8.74768674e-01 -7.85272300e-01 -5.55018306e-01 7.67394006e-01
-1.58490300e+00 -1.68827617e+00 -6.28196402e-03 2.21549436e-01
9.94887114e-01 -3.03184569e-01 9.22272801e-01 2.18084976e-02
-3.73382509e-01 8.79981875e-01 -3.19389045e-01 5.91588654e-02
7.17361808e-01 -1.72847605e+00 1.01523089e+00 9.10487831e-01
4.24115926e-01 6.94892347e-01 5.15120625e-01 -6.04271948e-01
-1.41926587e+00 -1.20813286e+00 1.15959120e+00 -7.70701051e-01
6.73107982e-01 -6.49004579e-01 -4.75691527e-01 9.02589738e-01
5.27515650e-01 -4.83590722e-01 1.05128682e+00 7.36295938e-01
-6.19063973e-01 -1.78573176e-01 -5.86003065e-01 8.16956520e-01
6.85906231e-01 -5.85645556e-01 -4.71030146e-01 1.26298040e-01
1.04223561e+00 -4.54653352e-01 -6.75406635e-01 8.60849202e-01
5.64044654e-01 -8.97122324e-01 6.69077277e-01 -7.00042307e-01
1.04900169e+00 -9.07309353e-02 -1.20862596e-01 -1.75971532e+00
1.96573101e-02 -7.22373843e-01 -3.88541311e-01 1.41098166e+00
1.18709397e+00 -6.98407590e-01 8.49995613e-01 4.61147368e-01
-1.68606892e-01 -1.32901657e+00 -3.30815852e-01 -2.44421631e-01
8.23794901e-02 -5.31512797e-01 8.79563332e-01 5.91644883e-01
-2.89838433e-01 1.17824817e+00 -2.76255190e-01 -1.87272206e-01
2.38511577e-01 6.54931426e-01 1.06789362e+00 -7.90753424e-01
-4.35301602e-01 -4.62415278e-01 -1.99155554e-01 -1.22107780e+00
2.06934344e-02 -8.09178710e-01 -1.29951969e-01 -1.59207451e+00
1.91217646e-01 -2.32190549e-01 -1.18239097e-01 5.86483896e-01
-6.46155179e-01 4.60303813e-01 1.61244109e-01 -3.75189841e-01
-7.06228912e-01 9.31667507e-01 1.21404779e+00 -5.01907647e-01
-7.09939003e-02 3.81016046e-01 -1.69691610e+00 6.11575663e-01
7.63781190e-01 -4.47250098e-01 -5.49648583e-01 -2.52855599e-01
8.75864208e-01 -4.80682611e-01 -2.61140168e-01 -7.84803629e-01
-1.25574425e-01 1.98147148e-01 2.72103697e-01 -9.36420143e-01
5.39316952e-01 -3.11229885e-01 -2.65927732e-01 -3.30045037e-02
-2.54503101e-01 3.74018341e-01 3.13954800e-01 4.63899136e-01
-6.70857489e-01 1.74428761e-01 1.06337562e-01 2.56504938e-02
-3.91052455e-01 2.29061469e-01 -1.31308734e-01 5.47380149e-01
4.89474326e-01 2.08473802e-01 -5.44360697e-01 -7.49326766e-01
-3.31801921e-01 1.90475076e-01 1.15271583e-01 7.57954180e-01
4.84668791e-01 -1.03158522e+00 -9.63119149e-01 4.67003286e-01
2.75835097e-01 2.57266760e-01 2.17116281e-01 6.36195779e-01
2.32756808e-02 4.64558423e-01 2.89588928e-01 -4.85849798e-01
-1.12468600e+00 8.10755566e-02 -1.82111692e-02 -7.84621060e-01
-7.72650391e-02 9.88884807e-01 4.69187617e-01 -6.91018045e-01
-4.22661960e-01 -4.36569870e-01 -3.50053072e-01 3.60530645e-01
4.49092835e-01 -2.63124943e-01 4.33505565e-01 -6.39214396e-01
-1.78971902e-01 4.29396957e-01 -5.06003261e-01 4.98648640e-03
1.46516430e+00 9.45745222e-03 -1.42524824e-01 4.46223199e-01
1.24768078e+00 5.08339703e-01 -1.02344370e+00 2.32130904e-02
-2.25866035e-01 -9.18728188e-02 -2.39930138e-01 -1.23414958e+00
-1.14009881e+00 6.13121390e-01 -6.47493303e-02 2.95423895e-01
1.20534825e+00 1.53669380e-02 1.02494264e+00 1.92787424e-01
1.04729064e-01 -7.94859946e-01 1.19722635e-01 8.97885203e-01
8.72785747e-01 -1.39252937e+00 3.30462828e-02 -4.27788973e-01
-1.19527245e+00 6.82345808e-01 6.79782927e-01 -2.93851481e-03
4.80313808e-01 5.16048133e-01 4.34554040e-01 -3.83781791e-01
-1.29002631e+00 -3.39627743e-01 6.26084626e-01 3.75725299e-01
7.08317399e-01 2.56351769e-01 -8.93347189e-02 1.16221881e+00
-1.10145998e+00 -5.25902152e-01 3.90127927e-01 7.07422078e-01
1.58608496e-01 -1.23851287e+00 1.81886390e-01 8.62906635e-01
-7.12971032e-01 -7.53791034e-01 -7.59042084e-01 5.73377013e-01
-2.14218110e-01 1.08268940e+00 5.97412400e-02 -4.66776937e-01
4.78337020e-01 2.01614186e-01 -4.92553413e-02 -8.00241053e-01
-9.40762162e-01 -7.24829063e-02 5.41247368e-01 -4.33544338e-01
-1.88843742e-01 -5.51884115e-01 -7.84918725e-01 7.59043619e-02
-2.27889240e-01 3.78591686e-01 7.55490303e-01 9.51765478e-01
8.45490277e-01 8.43958139e-01 8.31546843e-01 -5.33383727e-01
-3.58886153e-01 -9.73537803e-01 -3.15732449e-01 4.02562112e-01
3.58580768e-01 -1.74399957e-01 -5.43933332e-01 1.15631476e-01] | [11.485977172851562, 6.717613220214844] |
17fb37bf-482a-42c7-a7aa-a51ec1a5e038 | corpus-driven-lexical-analysis-norms-and | null | null | https://aclanthology.org/W13-3807 | https://aclanthology.org/W13-3807.pdf | Corpus-driven Lexical Analysis: Norms and Exploitations in Word Use | null | ['Patrick Hanks'] | 2013-11-01 | null | null | null | ws-2013-11 | ['lexical-analysis'] | ['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] | [-7.403311252593994, 3.7471799850463867] |
63f1375b-6d71-44ff-823e-64df86dfdaa3 | a-unified-compression-framework-for-efficient | 2304.00471 | null | https://arxiv.org/abs/2304.00471v2 | https://arxiv.org/pdf/2304.00471v2.pdf | A Unified Compression Framework for Efficient Speech-Driven Talking-Face Generation | Virtual humans have gained considerable attention in numerous industries, e.g., entertainment and e-commerce. As a core technology, synthesizing photorealistic face frames from target speech and facial identity has been actively studied with generative adversarial networks. Despite remarkable results of modern talking-face generation models, they often entail high computational burdens, which limit their efficient deployment. This study aims to develop a lightweight model for speech-driven talking-face synthesis. We build a compact generator by removing the residual blocks and reducing the channel width from Wav2Lip, a popular talking-face generator. We also present a knowledge distillation scheme to stably yet effectively train the small-capacity generator without adversarial learning. We reduce the number of parameters and MACs by 28$\times$ while retaining the performance of the original model. Moreover, to alleviate a severe performance drop when converting the whole generator to INT8 precision, we adopt a selective quantization method that uses FP16 for the quantization-sensitive layers and INT8 for the other layers. Using this mixed precision, we achieve up to a 19$\times$ speedup on edge GPUs without noticeably compromising the generation quality. | ['Hyoung-Kyu Song', 'Sungsu Lim', 'Hyungshin Kim', 'Shinkook Choi', 'Hancheol Park', 'Daeun Seo', 'Jaemin Kang', 'Bo-Kyeong Kim'] | 2023-04-02 | null | null | null | null | ['talking-face-generation', 'face-generation'] | ['computer-vision', 'computer-vision'] | [ 3.65426987e-01 3.35064858e-01 1.68472111e-01 -1.82587162e-01
-8.83311033e-01 -4.49348062e-01 5.97050011e-01 -7.08714604e-01
-1.54494688e-01 7.61411965e-01 -4.47245352e-02 -4.38525915e-01
5.08718610e-01 -1.02654636e+00 -6.77051246e-01 -8.71098638e-01
3.76586735e-01 -6.98646083e-02 -6.25230223e-02 -2.38991871e-01
-1.99837208e-01 4.17625517e-01 -1.54510486e+00 3.59307319e-01
6.90182745e-01 1.21926355e+00 -8.78348276e-02 6.63783252e-01
9.57282186e-02 8.23307395e-01 -8.69098306e-01 -1.05667436e+00
5.33153892e-01 -7.49011755e-01 -4.77122515e-01 -1.47429397e-02
4.02299523e-01 -5.65546930e-01 -8.21373701e-01 1.11384416e+00
9.54659224e-01 -1.72365755e-01 2.89649576e-01 -1.46757138e+00
-3.77361208e-01 5.15831172e-01 -5.10140896e-01 -2.28814527e-01
8.43524337e-02 5.81375122e-01 5.42417049e-01 -7.99335659e-01
4.99842346e-01 1.47398698e+00 4.75519985e-01 8.51157308e-01
-1.00100231e+00 -1.39727831e+00 -3.01951051e-01 1.36776805e-01
-1.80689812e+00 -1.08671594e+00 6.40534043e-01 -9.67916548e-02
6.18305504e-01 3.43585610e-01 4.39585835e-01 1.13523388e+00
-5.44202551e-02 3.23075771e-01 8.63394678e-01 -1.23899974e-01
4.75328229e-02 1.67562179e-02 -8.46008062e-01 8.98745120e-01
7.88793713e-02 1.43524230e-01 -4.88201380e-01 -1.42482325e-01
1.01265621e+00 -3.69120747e-01 -2.85728216e-01 2.01287284e-01
-9.02164757e-01 8.88859987e-01 3.24192643e-01 -1.50710180e-01
-2.33661205e-01 5.98986447e-01 3.12385708e-01 1.67726666e-01
2.38998994e-01 3.80974375e-02 1.03695221e-01 -1.19930193e-01
-1.04520810e+00 1.50979951e-01 6.87840939e-01 1.26225460e+00
7.37182260e-01 5.55367291e-01 -3.27905059e-01 6.73774540e-01
2.21621126e-01 8.72014701e-01 1.82435408e-01 -1.05563128e+00
5.40255725e-01 2.52652973e-01 -2.55215824e-01 -9.89499569e-01
1.56391114e-02 -2.72103935e-01 -1.37447631e+00 8.18749294e-02
2.23419622e-01 -6.46994650e-01 -8.17797840e-01 1.71200788e+00
4.54210639e-01 4.46879327e-01 4.69057501e-04 8.38815808e-01
8.23973477e-01 9.31215465e-01 -3.12843397e-02 -2.52009153e-01
1.41988838e+00 -8.86982501e-01 -5.27658641e-01 -1.53143942e-01
3.97105813e-01 -1.02872694e+00 7.41977632e-01 2.76891142e-01
-1.26717496e+00 -7.00582743e-01 -1.20993185e+00 -1.93498090e-01
1.37663797e-01 2.67784178e-01 6.83431149e-01 1.14856315e+00
-1.12961829e+00 4.65058804e-01 -5.12063026e-01 3.81779999e-01
6.16506577e-01 6.60496116e-01 -2.51573682e-01 4.38876897e-02
-1.27065694e+00 2.98013121e-01 2.00144440e-01 5.00336252e-02
-8.17116499e-01 -7.64885843e-01 -9.18747067e-01 2.40079924e-01
3.23851526e-01 -7.93790042e-01 1.11631227e+00 -9.71579850e-01
-2.15735245e+00 5.28146267e-01 -6.88192397e-02 -5.23823798e-01
5.41790962e-01 2.48114944e-01 -6.26854181e-01 1.87105700e-01
-3.17021400e-01 8.96584630e-01 1.36141920e+00 -9.43199277e-01
-4.95628297e-01 -1.85306668e-01 2.40989570e-02 8.36630464e-02
-4.73148346e-01 -1.56908885e-01 -6.55886590e-01 -9.20773745e-01
-2.37678394e-01 -1.23977256e+00 -1.34918883e-01 8.21600035e-02
-3.47596169e-01 1.94440767e-01 7.91514575e-01 -5.60525060e-01
1.14179862e+00 -2.34696126e+00 -2.13622659e-01 2.71749675e-01
2.66030073e-01 6.38280034e-01 -1.14660434e-01 4.01283130e-02
1.23774245e-01 3.78367268e-02 -1.46497712e-01 -4.37214494e-01
2.76549291e-02 2.72432230e-02 -4.97925580e-01 3.94335181e-01
2.66133219e-01 8.99754524e-01 -6.57476842e-01 -4.95660305e-01
1.90473348e-01 9.81953561e-01 -8.68747592e-01 1.62615553e-01
-1.22781150e-01 3.93298358e-01 -1.62690222e-01 2.58325130e-01
8.56377840e-01 -1.10827021e-01 2.52181441e-01 -1.51012480e-01
2.67022729e-01 2.77024567e-01 -1.11698163e+00 1.62253213e+00
-8.45373333e-01 6.14176393e-01 3.93815398e-01 -4.50157613e-01
9.89156783e-01 4.39582825e-01 1.14360608e-01 -7.44531631e-01
4.89863604e-01 2.04065442e-01 5.03879040e-02 3.50827500e-02
4.91730005e-01 -3.03794980e-01 -1.92673281e-01 1.97419882e-01
-7.88952261e-02 -4.43168312e-01 -7.53166005e-02 7.18019605e-02
8.65219057e-01 -1.87312409e-01 -5.12209646e-02 -1.87039319e-02
7.83013523e-01 -4.81511533e-01 5.90146840e-01 1.61440104e-01
1.35092884e-01 7.42956698e-01 5.72432160e-01 -8.29292834e-02
-9.13994312e-01 -9.31576669e-01 2.59454161e-01 9.06888545e-01
7.20262751e-02 -5.21091163e-01 -1.10830641e+00 -3.46804142e-01
-3.30791056e-01 6.08614087e-01 -1.94366455e-01 -5.19467890e-01
-6.81499302e-01 -7.45284736e-01 1.19443989e+00 3.92081171e-01
9.91845727e-01 -9.39418614e-01 -3.03417861e-01 6.99377358e-02
-1.01816297e-01 -1.24942148e+00 -8.28915238e-01 -5.00768363e-01
-3.43953639e-01 -6.07980251e-01 -7.92933047e-01 -9.20248568e-01
4.91869569e-01 1.09577879e-01 8.01121175e-01 8.16404372e-02
-2.42818117e-01 -2.37503931e-01 3.96879353e-02 -2.51188904e-01
-7.75144160e-01 1.14793241e-01 1.80086326e-02 2.22801685e-01
-1.80564448e-02 -4.39469129e-01 -8.21948051e-01 3.07575315e-01
-8.15043986e-01 2.42162794e-01 5.37850142e-01 8.84953499e-01
4.33769941e-01 9.30753872e-02 6.30340636e-01 -7.87280440e-01
3.90761167e-01 -1.54171750e-01 -8.75097990e-01 -8.94193724e-02
-2.90858179e-01 4.83597256e-02 9.48391855e-01 -4.06404287e-01
-1.06934249e+00 1.07050076e-01 -5.48161209e-01 -6.09298587e-01
3.87240916e-01 -2.87692010e-01 -5.20125389e-01 -3.10238361e-01
5.63800931e-01 2.64699191e-01 1.16538621e-01 1.66906089e-01
5.03062665e-01 8.43641281e-01 6.74710751e-01 -3.09245378e-01
9.11821485e-01 3.24261129e-01 7.95802921e-02 -1.01860881e+00
-2.35070795e-01 1.33152515e-01 1.62552804e-01 1.03838090e-02
6.31700397e-01 -1.32152855e+00 -1.20492923e+00 5.80939651e-01
-1.27309513e+00 -3.19406271e-01 -1.84335485e-01 2.79958576e-01
-5.50173581e-01 1.93880096e-01 -6.45234466e-01 -5.88295102e-01
-7.56169617e-01 -1.38413227e+00 1.20862842e+00 2.56718576e-01
-1.91426079e-04 -3.83538485e-01 -4.12115395e-01 4.18507487e-01
6.29557371e-01 1.41961649e-01 5.44314325e-01 1.74515024e-01
-8.11735630e-01 -1.14919455e-03 -3.72435868e-01 4.55333233e-01
5.05514480e-02 -1.79633588e-01 -1.17724586e+00 -3.84252489e-01
6.62934734e-03 -1.69013053e-01 6.94578946e-01 -4.65092957e-02
1.39922214e+00 -5.47184229e-01 -1.93720624e-01 1.05505896e+00
1.12061214e+00 3.84718239e-01 9.92729783e-01 -3.85932833e-01
8.11168015e-01 2.67528117e-01 1.55963838e-01 6.84274435e-01
8.69898722e-02 6.85270667e-01 2.87827432e-01 -1.18493922e-01
-5.55203795e-01 -3.30397367e-01 5.64031899e-01 7.07295477e-01
3.20099853e-02 -4.52153504e-01 -5.21108031e-01 1.76419780e-01
-1.19526303e+00 -9.71857667e-01 2.75582671e-01 2.25690222e+00
1.04338229e+00 6.82337955e-02 -9.23307091e-02 2.18512371e-01
8.32520127e-01 3.53091747e-01 -4.62230265e-01 -4.19456989e-01
8.86329077e-03 8.26077878e-01 6.18416786e-01 5.48625469e-01
-8.73689055e-01 1.18512857e+00 5.61785936e+00 1.39237392e+00
-1.38355374e+00 1.11416966e-01 9.11495090e-01 -3.57554466e-01
-2.20978454e-01 -2.52021044e-01 -8.90672386e-01 6.14019215e-01
1.12608325e+00 -2.50696063e-01 6.09217405e-01 7.76104629e-01
-1.51343644e-02 3.18701863e-01 -8.17348123e-01 1.46733618e+00
1.34143233e-01 -1.47571671e+00 1.54689744e-01 1.59402773e-01
7.42996097e-01 -4.85039443e-01 3.57821047e-01 2.36215264e-01
1.99014425e-01 -1.27291024e+00 6.83844566e-01 -9.19058472e-02
1.59142911e+00 -1.18544209e+00 4.87966001e-01 -6.75345585e-03
-1.15961921e+00 1.23838842e-01 -4.35495287e-01 3.68390605e-02
2.06807047e-01 4.74983573e-01 -8.94194186e-01 2.09956214e-01
3.20330322e-01 4.25476879e-02 -2.69098133e-01 4.24518198e-01
-2.89899468e-01 4.87149864e-01 -3.01621377e-01 7.27092996e-02
3.87460589e-02 -1.04822047e-01 2.68604517e-01 9.49474931e-01
6.81736767e-01 2.98230320e-01 -2.31116012e-01 6.62613571e-01
-6.56734407e-01 -1.07147381e-01 -5.44327676e-01 3.51312831e-02
5.82517803e-01 1.22366977e+00 -4.71648067e-01 -4.28989679e-01
-2.23015457e-01 1.32460570e+00 -2.12183461e-01 -1.39251715e-02
-1.39168859e+00 -5.83131015e-01 1.01545119e+00 3.07347387e-01
2.59766489e-01 -1.80136606e-01 -2.16270387e-01 -9.72949803e-01
4.80382331e-02 -1.14297533e+00 -1.50757834e-01 -4.49262112e-01
-6.69388175e-01 9.01912153e-01 -4.04546529e-01 -1.11880052e+00
-4.58790779e-01 -2.57763922e-01 -3.43896210e-01 9.53293085e-01
-1.25570548e+00 -9.91743207e-01 -4.23613846e-01 8.30408454e-01
2.48824105e-01 -2.90086001e-01 7.03072190e-01 6.06370449e-01
-6.72107875e-01 1.30908251e+00 -1.45292878e-01 2.19796017e-01
7.29627490e-01 -5.72603405e-01 8.52188349e-01 8.69789422e-01
-8.12876597e-03 5.59348226e-01 3.82183343e-01 -3.92029822e-01
-1.60048640e+00 -1.36435294e+00 6.41411781e-01 2.08589032e-01
3.16721678e-01 -7.11400032e-01 -5.16164601e-01 3.56104225e-01
3.48216027e-01 5.17527461e-02 4.95554894e-01 -6.63900197e-01
-3.14566791e-01 -3.65857750e-01 -1.25675464e+00 1.00017023e+00
1.21649027e+00 -5.73421180e-01 1.51637226e-01 2.83983052e-01
7.48467386e-01 -6.72511518e-01 -7.21226037e-01 2.20039368e-01
4.55108255e-01 -9.90797818e-01 1.06348968e+00 1.94117904e-01
4.82719004e-01 -2.93205917e-01 5.48099056e-02 -1.10767019e+00
-1.81896597e-01 -1.43170750e+00 -1.73175871e-01 1.30439699e+00
2.50059992e-01 -6.73597813e-01 1.09544098e+00 3.58099580e-01
1.25536835e-02 -4.85338986e-01 -1.00387180e+00 -6.29453838e-01
-2.83641927e-02 -3.21960330e-01 1.02226579e+00 4.59878683e-01
-2.00567275e-01 5.32841623e-01 -6.17619514e-01 1.84508383e-01
4.52062011e-01 -8.04471076e-02 1.06851828e+00 -6.48429573e-01
-3.94034326e-01 -3.77188802e-01 -4.78570670e-01 -1.21039689e+00
2.02968046e-01 -8.24610114e-01 7.73667842e-02 -9.11678851e-01
-3.33956152e-01 -6.19285107e-01 3.06891501e-01 2.96266645e-01
-7.75861293e-02 8.24464440e-01 3.30831319e-01 -2.18585968e-01
-5.69780953e-02 7.33445168e-01 1.43642378e+00 -1.51998281e-01
-5.20136692e-02 -1.15844481e-01 -8.29458654e-01 5.78883469e-01
8.56693745e-01 -1.91557080e-01 -7.35329390e-01 -4.68168795e-01
4.18525189e-02 3.64915103e-01 4.35402691e-01 -1.27517390e+00
1.52991161e-01 5.53902909e-02 3.41589719e-01 -8.12001154e-03
7.78724015e-01 -5.71957767e-01 4.41677272e-01 5.68244874e-01
4.03460041e-02 -7.18659088e-02 2.93611646e-01 2.74765015e-01
-1.31937444e-01 2.08151370e-01 1.09697092e+00 2.05060497e-01
-4.28801775e-01 6.34441257e-01 -9.53615978e-02 5.73814586e-02
1.20680940e+00 2.41227597e-02 -3.95206958e-01 -6.19794548e-01
-4.03297275e-01 -1.07787110e-01 5.43151736e-01 4.14658427e-01
6.08981729e-01 -1.40310311e+00 -8.01972091e-01 7.09199965e-01
-3.41105372e-01 1.75100282e-01 5.54507554e-01 3.70030880e-01
-8.73966992e-01 4.14356083e-01 -9.42403674e-02 -2.49868408e-01
-1.21462405e+00 5.50911903e-01 1.66689157e-01 3.87872942e-02
-4.67978150e-01 1.06609619e+00 4.45890725e-01 1.21243596e-01
-6.81804772e-03 7.71496743e-02 2.23111421e-01 -1.08181976e-01
7.19103575e-01 5.59557140e-01 1.22987650e-01 -7.98093379e-01
-2.20244110e-01 3.42639893e-01 2.03566611e-01 -1.80174857e-01
8.33095431e-01 3.94139625e-02 1.63835846e-02 -4.56325144e-01
1.28139853e+00 8.32440257e-02 -1.41468167e+00 1.43118784e-01
-7.72005200e-01 -4.61919636e-01 9.56046358e-02 -2.48937875e-01
-1.72967124e+00 1.02485037e+00 5.04312932e-01 -2.27623701e-01
1.30500102e+00 -3.22206438e-01 1.35177028e+00 6.77548498e-02
6.04288220e-01 -8.32467198e-01 -1.39899701e-01 2.25151822e-01
6.09297395e-01 -7.96717167e-01 -9.15615633e-02 -9.27753508e-01
-5.48365235e-01 8.07244837e-01 5.27614474e-01 -4.86734025e-02
4.78181690e-01 5.81475019e-01 -1.17512688e-01 2.61922747e-01
-5.50070703e-01 4.22849618e-02 2.61684507e-02 5.96929193e-01
2.17364043e-01 1.27126321e-01 -2.95705318e-01 4.86144453e-01
-7.38592446e-01 7.30934441e-02 4.79000777e-01 4.23999697e-01
1.73950922e-02 -1.13315260e+00 -3.70393485e-01 2.53724217e-01
-6.42152011e-01 -3.97754639e-01 3.45848612e-02 4.96204823e-01
2.96796024e-01 1.10165668e+00 2.37320974e-01 -6.25766277e-01
8.30525979e-02 -1.41337708e-01 4.88968223e-01 -3.90826613e-01
-6.74853861e-01 4.80981730e-02 -5.00719361e-02 -5.14188826e-01
2.41117384e-02 -1.88203320e-01 -1.09704661e+00 -9.94363427e-01
-1.62884340e-01 -8.10309276e-02 4.55051124e-01 5.19140601e-01
7.69614100e-01 5.10060608e-01 7.81795204e-01 -7.33712614e-01
-4.61075693e-01 -6.20880485e-01 -3.88473123e-01 5.02622798e-02
3.58526744e-02 -4.03508395e-01 -1.04546584e-01 1.52619496e-01] | [12.146272659301758, -0.3871088922023773] |
e56a06bb-619a-4fd7-8c3f-bb8ba838586e | phyaat-physiology-of-auditory-attention-to | 2005.11577 | null | https://arxiv.org/abs/2005.11577v1 | https://arxiv.org/pdf/2005.11577v1.pdf | PhyAAt: Physiology of Auditory Attention to Speech Dataset | Auditory attention to natural speech is a complex brain process. Its quantification from physiological signals can be valuable to improving and widening the range of applications of current brain-computer-interface systems, however it remains a challenging task. In this article, we present a dataset of physiological signals collected from an experiment on auditory attention to natural speech. In this experiment, auditory stimuli consisting of reproductions of English sentences in different auditory conditions were presented to 25 non-native participants, who were asked to transcribe the sentences. During the experiment, 14 channel electroencephalogram, galvanic skin response, and photoplethysmogram signals were collected from each participant. Based on the number of correctly transcribed words, an attention score was obtained for each auditory stimulus presented to subjects. A strong correlation ($p<<0.0001$) between the attention score and the auditory conditions was found. We also formulate four different predictive tasks involving the collected dataset and develop a feature extraction framework. The results for each predictive task are obtained using a Support Vector Machine with spectral features, and are better than chance level. The dataset has been made publicly available for further research, along with a python library - phyaat to facilitate the preprocessing, modeling, and reproduction of the results presented in this paper. The dataset and other resources are shared on webpage - https://phyaat.github.io. | ['Jesús Requena Carrión', 'Nikesh Bajaj', 'Francesco Bellotti'] | 2020-05-23 | null | null | null | null | ['lwr-classification', 'semanticity-prediction', 'attention-score-prediction', 'noise-level-prediction'] | ['time-series', 'time-series', 'time-series', 'time-series'] | [ 1.87854990e-01 -2.40713239e-01 4.23441559e-01 -2.87659705e-01
-4.56242770e-01 -1.84891358e-01 1.34614855e-01 7.52741564e-03
-5.05899012e-01 8.28103244e-01 2.57821977e-01 -5.82190305e-02
-7.00694025e-02 -1.67971864e-01 -2.32936993e-01 -7.07722306e-01
-4.72790860e-02 -2.00026453e-01 -1.73047781e-02 7.86397383e-02
6.08766496e-01 2.58400977e-01 -1.76791477e+00 1.67561904e-01
7.24851191e-01 1.46100831e+00 6.52791381e-01 5.92344880e-01
2.65841484e-01 1.74371168e-01 -8.24765861e-01 -1.36904240e-01
-3.90910991e-02 -6.03449464e-01 -5.72896659e-01 -3.40528309e-01
-2.03998804e-01 2.46579871e-01 -2.00853005e-01 1.17797041e+00
1.02397990e+00 2.41311908e-01 5.07758439e-01 -1.27143693e+00
-6.46671295e-01 4.43235993e-01 -2.45576635e-01 8.09253693e-01
7.07753479e-01 3.68753463e-01 7.86150694e-01 -9.67980325e-01
-6.20315373e-02 7.74630308e-01 2.37837493e-01 6.65381253e-01
-9.54111099e-01 -1.02677119e+00 -2.86140710e-01 7.66006887e-01
-1.52431476e+00 -7.44558036e-01 9.03060079e-01 -5.62402606e-01
1.19624877e+00 4.81517702e-01 9.42073405e-01 1.29330504e+00
5.00420988e-01 2.77641714e-01 1.36379576e+00 -2.76227772e-01
5.19384027e-01 4.10655528e-01 3.08630645e-01 5.18947430e-02
-2.66925339e-02 3.35259326e-02 -1.03606129e+00 -1.56087056e-01
5.07314682e-01 -3.42880428e-01 -6.58620417e-01 3.71489495e-01
-1.04036427e+00 5.80392957e-01 2.22334355e-01 6.23683333e-01
-8.21512461e-01 -2.37642899e-01 2.99256027e-01 2.50912189e-01
2.90212542e-01 4.20282871e-01 -3.81071746e-01 -5.52088082e-01
-5.26413560e-01 -2.24675775e-01 6.65746033e-01 5.43079793e-01
3.54793996e-01 2.00733751e-01 -1.46560237e-01 1.07518542e+00
1.57759070e-01 6.92085147e-01 1.05660129e+00 -5.44713259e-01
-1.90738235e-02 1.82477698e-01 -1.03196204e-01 -9.92163897e-01
-6.26753986e-01 -2.01369822e-01 -6.60548449e-01 -2.29666367e-01
-7.39835668e-04 -1.54408306e-01 -4.03702378e-01 1.68061531e+00
-1.00514740e-01 1.79750994e-01 -7.03838468e-02 1.09761810e+00
8.69462013e-01 5.09905636e-01 2.25643814e-01 -7.09651887e-01
1.83309114e+00 -2.42940947e-01 -1.14266348e+00 -4.19638485e-01
-1.90875530e-01 -7.35862315e-01 1.34690976e+00 6.34941161e-01
-1.01319039e+00 -6.60074711e-01 -9.17709649e-01 3.48647535e-01
-2.63197869e-01 -1.39320306e-02 4.77500916e-01 7.96698928e-01
-1.06542730e+00 1.66870579e-01 -6.76453173e-01 -3.98602128e-01
4.12292331e-01 3.56629848e-01 -3.85945231e-01 3.85070413e-01
-1.47852838e+00 1.06634188e+00 1.21630065e-01 2.22372070e-01
-5.72687745e-01 -4.89957333e-01 -6.08891368e-01 1.52696595e-01
-3.58570814e-02 -3.52046400e-01 1.11484993e+00 -7.69102216e-01
-1.55405235e+00 7.61183739e-01 -4.19161528e-01 -1.91489562e-01
-2.71725535e-01 -1.17195725e-01 -6.70065165e-01 2.15239078e-01
3.04922229e-03 3.83592039e-01 8.53636205e-01 -5.51616609e-01
-8.39353353e-02 -4.91236269e-01 -6.52781188e-01 6.05390221e-02
-4.42258149e-01 5.71564138e-01 2.64860950e-02 -5.24370313e-01
-6.58029616e-02 -6.74612820e-01 1.95137560e-01 -5.05057335e-01
-3.26327860e-01 -3.23486537e-01 9.03671235e-02 -8.61391783e-01
1.15497780e+00 -2.48081470e+00 6.49638250e-02 2.09611684e-01
6.49490161e-03 1.17306106e-01 -7.20514059e-02 7.40353838e-02
-3.32446724e-01 1.19624265e-01 -3.05277020e-01 4.66627292e-02
1.02566466e-01 -3.88975352e-01 1.89784530e-03 4.54355389e-01
1.06147848e-01 6.59274459e-01 -5.28325319e-01 -1.38451293e-01
9.48355645e-02 6.05140984e-01 -2.39681810e-01 3.56530041e-01
3.97754371e-01 5.01425207e-01 -3.15747894e-02 4.59242910e-01
4.99329537e-01 2.43627563e-01 -3.83818448e-01 -1.15242518e-01
-1.96539372e-01 4.00764883e-01 -8.04699838e-01 1.38424897e+00
-3.81738544e-01 8.92534494e-01 -2.51904652e-02 -8.38483810e-01
1.10443819e+00 5.91687918e-01 2.07319677e-01 -9.98507142e-01
6.66116357e-01 1.32065624e-01 6.37149990e-01 -8.44656229e-01
-1.09652253e-02 -2.08691746e-01 6.30476400e-02 4.86357123e-01
7.15037510e-02 -2.88702935e-01 8.69360715e-02 1.80089730e-04
8.96769881e-01 -6.04566276e-01 6.35753989e-01 -3.77316505e-01
6.41497195e-01 -5.63157201e-01 2.74004191e-01 5.04800320e-01
-5.30120134e-01 2.40167201e-01 4.01058644e-01 3.11171114e-02
-5.05387187e-01 -1.06562841e+00 -4.86634910e-01 1.03505909e+00
-1.73954085e-01 -1.08908646e-01 -9.51758862e-01 2.82202423e-01
-3.32342416e-01 9.87310112e-01 -4.91295785e-01 -4.98411179e-01
1.79676652e-01 -7.52415538e-01 3.64398539e-01 4.06022370e-01
3.24722141e-01 -1.81424451e+00 -8.45425308e-01 2.99082957e-02
-5.26042223e-01 -1.21664512e+00 -4.22655702e-01 3.65281641e-01
-3.49631548e-01 -7.85112321e-01 -4.67344195e-01 -6.26383781e-01
1.64131403e-01 2.27874108e-02 5.80507874e-01 -3.14585418e-01
-7.39376843e-01 5.36582947e-01 -3.05919796e-01 -1.00558519e+00
1.11086421e-01 -1.93995550e-01 5.27668178e-01 1.59911245e-01
9.15410876e-01 -8.88256013e-01 -6.02344811e-01 1.42524451e-01
-5.28024793e-01 -3.03996921e-01 5.60913384e-01 4.23235506e-01
4.42370951e-01 -1.45573869e-01 9.77390170e-01 1.13555841e-01
1.28161848e+00 -5.14529943e-01 -2.17324883e-01 -2.35250697e-01
-2.02165425e-01 -3.62863541e-01 4.73827124e-01 -6.31383061e-01
-6.76308692e-01 -1.96585998e-01 -1.65659949e-01 -2.38207370e-01
-5.59857309e-01 4.87376750e-01 -2.46648043e-01 7.23304302e-02
7.73300946e-01 4.03384984e-01 -6.41058162e-02 -1.47482961e-01
-2.05779448e-01 1.23892272e+00 6.87651634e-01 -1.44781858e-01
1.48845568e-01 -1.13399692e-01 -4.78287995e-01 -1.26073611e+00
-5.17029583e-01 -3.94319654e-01 -4.73584920e-01 -5.10878742e-01
8.99954200e-01 -6.82211339e-01 -8.92686009e-01 6.64777398e-01
-1.03596509e+00 -1.24427550e-01 2.86916971e-01 1.06383216e+00
-5.49961150e-01 1.09102778e-01 -2.11770922e-01 -1.21888673e+00
-7.79354155e-01 -1.01770794e+00 4.55723435e-01 3.91766697e-01
-6.27064645e-01 -4.35615599e-01 -5.12474589e-02 3.98639113e-01
6.04449749e-01 -2.82270908e-01 5.54452896e-01 -8.37723076e-01
3.52783382e-01 -2.43037224e-01 4.33547236e-02 4.17870164e-01
2.14703709e-01 -2.93842554e-01 -1.49848557e+00 3.16598117e-02
7.94101894e-01 -3.12288076e-01 4.06800568e-01 7.20206380e-01
1.33394682e+00 -3.50321122e-02 8.99940655e-02 1.60121694e-01
7.87293971e-01 5.75178385e-01 7.83326387e-01 2.08311137e-02
6.01718500e-02 6.40329361e-01 2.70429775e-02 5.89272439e-01
-3.74858943e-03 4.04406399e-01 2.37817287e-01 3.26412112e-01
1.35771334e-01 3.69734466e-01 5.60786843e-01 1.08381712e+00
-4.93126139e-02 -9.90415812e-02 -1.01177216e+00 4.81483728e-01
-1.15081286e+00 -1.11111057e+00 -2.56385475e-01 2.35288000e+00
7.38413513e-01 5.11984192e-02 7.73452073e-02 6.42700493e-01
7.52936006e-01 -3.34858358e-01 -6.42973900e-01 -6.62190080e-01
-4.36554439e-02 6.96609616e-01 -2.16995895e-01 3.14301252e-01
-6.59438312e-01 5.77066243e-01 5.74517155e+00 4.02922332e-01
-1.52795494e+00 7.49907121e-02 2.95574337e-01 -3.54570419e-01
1.78492352e-01 -6.25402629e-01 -2.70049572e-01 6.90961123e-01
1.46532166e+00 -7.25874484e-01 9.04070258e-01 5.23705363e-01
6.77912712e-01 -5.50175726e-01 -9.18421865e-01 1.47605598e+00
3.17927897e-01 -4.47735786e-01 -4.11959082e-01 -2.03468397e-01
-1.50036231e-01 6.02755323e-02 2.23571703e-01 8.39284509e-02
-5.14638841e-01 -1.14804018e+00 4.93995190e-01 7.96710491e-01
8.37499082e-01 -5.78934252e-01 7.44144082e-01 4.10541713e-01
-9.34888780e-01 -2.28662699e-01 -3.83905858e-01 -5.44014156e-01
4.69757477e-03 6.09287679e-01 -6.97678685e-01 8.95886123e-02
9.62049901e-01 3.16251278e-01 -4.57140505e-01 1.21445441e+00
-3.79457086e-01 7.89318860e-01 -2.21053720e-01 -5.60835898e-01
-4.35070127e-01 1.22467786e-01 4.66234684e-01 1.01125193e+00
6.30726099e-01 4.93443370e-01 -4.98532563e-01 1.02931046e+00
2.14666590e-01 3.95072877e-01 -5.01924992e-01 -1.00011714e-01
7.56733000e-01 1.40397787e+00 -4.84875888e-01 1.24462331e-02
-2.96952099e-01 7.33446419e-01 2.26034853e-03 3.01482022e-01
-8.24973583e-01 -8.19511592e-01 4.84624773e-01 -1.78647920e-01
-2.12962642e-01 4.17189486e-02 -4.83911783e-01 -8.79952073e-01
1.97589070e-01 -7.17944503e-01 8.88772383e-02 -1.29489207e+00
-1.30450320e+00 8.59579146e-01 2.01777220e-02 -8.28980386e-01
-7.86871389e-02 -5.10990739e-01 -6.95693254e-01 1.46776533e+00
-1.08572292e+00 -6.42927364e-02 -4.72780496e-01 9.06547666e-01
5.87035894e-01 -4.68278900e-02 1.06716740e+00 2.19423473e-01
-7.50190914e-01 4.98804450e-01 -5.51044226e-01 -1.34756237e-01
7.54093111e-01 -7.44422734e-01 -6.27711490e-02 7.06078529e-01
-4.56187464e-02 7.13801980e-01 6.50117576e-01 -2.88225114e-01
-1.13832188e+00 -6.30892158e-01 1.04303861e+00 -5.01458049e-02
7.14455485e-01 -4.83043104e-01 -9.75513637e-01 2.36522242e-01
5.42509735e-01 -2.27844179e-01 1.14674890e+00 -2.26565644e-01
1.73392162e-01 -3.59831862e-02 -1.13384295e+00 3.89740676e-01
6.99563563e-01 -7.74235427e-01 -9.30472314e-01 8.00748467e-02
3.30974132e-01 -1.45659652e-02 -8.12376738e-01 -6.28319196e-03
5.37934363e-01 -8.29136431e-01 5.99902511e-01 -3.28836381e-01
3.52319889e-02 2.65513267e-02 -1.15459956e-01 -1.67782426e+00
-4.39711154e-01 -5.85318446e-01 7.89993331e-02 8.94004881e-01
3.65797848e-01 -8.98723304e-01 -5.88968694e-02 6.38392866e-01
-1.68809429e-01 -3.89546365e-01 -1.15699875e+00 -6.52083099e-01
-9.23189893e-02 -7.94176221e-01 2.99033076e-01 5.54952979e-01
8.67994189e-01 5.51182985e-01 -1.23482168e-01 2.55227059e-01
3.44407558e-01 -4.02782559e-01 2.09232092e-01 -1.35228717e+00
1.77338779e-01 -5.50561368e-01 -5.34511924e-01 -4.77819443e-01
2.46788085e-01 -9.48356986e-01 2.49508366e-01 -1.36022115e+00
3.51750523e-01 3.22773546e-01 -5.88720322e-01 5.87570071e-01
-2.93326020e-01 1.60578698e-01 -4.73825224e-02 -1.03849284e-01
3.25662345e-02 6.72083974e-01 8.18142176e-01 -6.01659156e-03
-2.81277984e-01 2.35929880e-02 -9.53888595e-01 4.57477719e-01
1.38735902e+00 -5.05965471e-01 -4.51339930e-01 -1.01720013e-01
-2.95418739e-01 2.69116879e-01 3.81287068e-01 -1.37564671e+00
2.91126072e-01 1.08328305e-01 4.48192805e-01 -2.23841056e-01
6.64163709e-01 -4.94959742e-01 3.28417756e-02 4.33064997e-01
-3.22390139e-01 1.03463896e-01 5.82529068e-01 1.63335457e-01
-2.85456777e-02 -3.22974712e-01 8.05526555e-01 8.29381570e-02
-3.56303930e-01 3.09087001e-02 -7.47115850e-01 -3.79041247e-02
9.24557865e-01 -2.73830444e-01 -2.37243816e-01 -4.86260265e-01
-9.50208068e-01 -1.73353747e-01 -2.28618994e-01 4.06474710e-01
8.61815870e-01 -1.15168965e+00 -6.82821095e-01 5.27283490e-01
-4.37541530e-02 -7.03327179e-01 5.55161953e-01 1.27863562e+00
1.86847359e-01 5.41574717e-01 -7.26015329e-01 -4.10863578e-01
-1.36548221e+00 5.19958377e-01 2.73385704e-01 6.96418166e-01
-2.26770088e-01 8.30334961e-01 6.66616186e-02 2.77664453e-01
3.40328604e-01 -2.93546289e-01 -6.39472127e-01 2.13798910e-01
9.71496046e-01 2.54742205e-01 3.02181989e-01 -6.95858359e-01
-5.58559239e-01 1.96307763e-01 2.54801482e-01 -2.94429809e-01
1.38109183e+00 -1.82286829e-01 -2.08544940e-01 8.59019518e-01
8.46147180e-01 3.98618216e-03 -4.89172310e-01 6.49777502e-02
-3.01912069e-01 -3.40349495e-01 1.30798206e-01 -9.54481602e-01
-1.01513028e+00 1.16182709e+00 9.01343465e-01 4.16610092e-01
1.36075222e+00 -8.80183373e-03 3.22583795e-01 2.52782941e-01
2.90391326e-01 -1.01441157e+00 1.02612726e-01 3.93426776e-01
1.21896279e+00 -9.07254517e-01 -2.47582018e-01 -2.09397376e-01
-8.20646226e-01 9.60210264e-01 6.48009360e-01 1.72100291e-01
7.41264462e-01 3.58334422e-01 3.19067836e-02 -7.83634856e-02
-8.22014511e-01 -1.28929883e-01 2.06988394e-01 8.37293804e-01
6.31465077e-01 1.63059935e-01 -5.34269571e-01 1.52485323e+00
-7.34790325e-01 -3.34174782e-02 5.17846584e-01 5.04699588e-01
-6.17478549e-01 -3.92185479e-01 -6.98026419e-01 6.65237665e-01
-4.37231779e-01 -4.25810158e-01 -5.10221958e-01 3.49092066e-01
-1.76457927e-01 1.36747575e+00 1.35949150e-01 -5.41477978e-01
5.41393220e-01 5.19447684e-01 2.07909435e-01 -3.95434052e-01
-4.40127820e-01 8.72306451e-02 -1.97120100e-01 -3.95942926e-01
-2.74532437e-01 -9.04292166e-01 -1.28707147e+00 -5.15099689e-02
-1.80181816e-01 3.53527963e-01 8.00857008e-01 6.40224874e-01
4.26607490e-01 6.54993534e-01 5.63217402e-01 -9.84982848e-01
-1.94121748e-01 -1.61282563e+00 -6.78408861e-01 1.37375116e-01
-2.15827934e-02 -8.52014899e-01 -8.48363161e-01 8.00465941e-02] | [13.218949317932129, 3.382521390914917] |
c2cbf74e-293c-45ce-b2c8-299a6fa4473c | magvlt-masked-generative-vision-and-language | 2303.12208 | null | https://arxiv.org/abs/2303.12208v1 | https://arxiv.org/pdf/2303.12208v1.pdf | MAGVLT: Masked Generative Vision-and-Language Transformer | While generative modeling on multimodal image-text data has been actively developed with large-scale paired datasets, there have been limited attempts to generate both image and text data by a single model rather than a generation of one fixed modality conditioned on the other modality. In this paper, we explore a unified generative vision-and-language (VL) model that can produce both images and text sequences. Especially, we propose a generative VL transformer based on the non-autoregressive mask prediction, named MAGVLT, and compare it with an autoregressive generative VL transformer (ARGVLT). In comparison to ARGVLT, the proposed MAGVLT enables bidirectional context encoding, fast decoding by parallel token predictions in an iterative refinement, and extended editing capabilities such as image and text infilling. For rigorous training of our MAGVLT with image-text pairs from scratch, we combine the image-to-text, text-to-image, and joint image-and-text mask prediction tasks. Moreover, we devise two additional tasks based on the step-unrolled mask prediction and the selective prediction on the mixture of two image-text pairs. Experimental results on various downstream generation tasks of VL benchmarks show that our MAGVLT outperforms ARGVLT by a large margin even with significant inference speedup. Particularly, MAGVLT achieves competitive results on both zero-shot image-to-text and text-to-image generation tasks from MS-COCO by one moderate-sized model (fewer than 500M parameters) even without the use of monomodal data and networks. | ['Jongmin Kim', 'Donghoon Lee', 'DaeJin Jo', 'Sungwoong Kim'] | 2023-03-21 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Kim_MAGVLT_Masked_Generative_Vision-and-Language_Transformer_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Kim_MAGVLT_Masked_Generative_Vision-and-Language_Transformer_CVPR_2023_paper.pdf | cvpr-2023-1 | ['text-infilling'] | ['natural-language-processing'] | [ 7.02348173e-01 2.29206994e-01 2.83367813e-01 -4.89813685e-01
-1.19329476e+00 -3.34983468e-01 1.15614903e+00 -4.99002606e-01
-3.47649992e-01 7.49663174e-01 2.00014740e-01 -3.37571919e-01
4.51623231e-01 -7.14074910e-01 -1.17524898e+00 -9.02888238e-01
7.54591048e-01 8.30549181e-01 4.83895056e-02 7.89474547e-02
-7.79088885e-02 -1.05084859e-01 -1.45714414e+00 6.88593149e-01
7.56886482e-01 9.06663179e-01 9.66047525e-01 9.04598892e-01
-2.84263015e-01 9.10759270e-01 -1.73442125e-01 -6.56416178e-01
2.60865301e-01 -7.64830947e-01 -4.67114359e-01 4.42817807e-01
6.95716977e-01 -4.84684020e-01 -3.26004475e-01 6.82629526e-01
7.25489259e-01 -2.04851590e-02 8.25605154e-01 -1.43443012e+00
-9.00224447e-01 9.25431132e-01 -8.18685412e-01 -3.87039334e-01
2.59943426e-01 5.05670428e-01 7.60755658e-01 -1.45345330e+00
9.30807292e-01 1.39381754e+00 5.23407340e-01 5.74955881e-01
-1.41048980e+00 -6.46514773e-01 1.78541765e-01 -5.89375533e-02
-1.38455999e+00 -6.51159763e-01 3.35317671e-01 -4.06069309e-01
1.12781191e+00 2.41911598e-02 4.42196906e-01 1.50685751e+00
1.91093162e-01 1.03509057e+00 1.10401356e+00 -3.99817854e-01
-1.77593455e-01 4.62304503e-02 -5.35740614e-01 8.94866347e-01
-1.27876878e-01 9.68399085e-03 -6.20857239e-01 2.39277735e-01
8.09138715e-01 -1.56976923e-01 -9.84566361e-02 -2.72281710e-02
-1.40792906e+00 7.85555780e-01 -3.58523354e-02 -1.59948971e-02
-3.23232293e-01 4.32090402e-01 2.23356664e-01 1.37394249e-01
4.38523650e-01 -1.91500083e-01 1.67582917e-03 6.59703165e-02
-1.31263709e+00 2.53454566e-01 5.06673157e-01 1.36890614e+00
8.93519163e-01 3.96985412e-01 -7.62269318e-01 8.38002861e-01
3.70305061e-01 9.43771958e-01 4.39831764e-01 -6.80172861e-01
8.61259460e-01 1.42763570e-01 -2.48895809e-01 -4.65571076e-01
-1.75761338e-02 -1.97319120e-01 -1.33086586e+00 -7.15123489e-02
6.44044355e-02 -4.23974216e-01 -1.49123585e+00 1.71880472e+00
1.64858580e-01 2.43293837e-01 9.11176726e-02 7.46871054e-01
1.10446334e+00 1.00288057e+00 6.98362440e-02 -2.52325147e-01
1.14181602e+00 -1.27860403e+00 -7.00340629e-01 -4.30667907e-01
3.83876681e-01 -1.07375193e+00 9.46503758e-01 2.47360706e-01
-1.44722295e+00 -7.89283991e-01 -5.17302096e-01 -4.34629321e-01
-1.48387253e-01 4.58942294e-01 5.22358477e-01 3.44282717e-01
-1.36830294e+00 -2.38206144e-02 -6.20692730e-01 -2.73106009e-01
2.88656563e-01 7.76417777e-02 -4.46383417e-01 -3.53821665e-01
-9.10797596e-01 5.68791449e-01 4.25851136e-01 2.85291791e-01
-1.45296478e+00 -5.40857375e-01 -1.21869206e+00 -6.99324384e-02
2.96054482e-01 -1.21795011e+00 1.04940522e+00 -8.47102880e-01
-1.63569999e+00 9.16049898e-01 -5.81280529e-01 -5.54611266e-01
8.92527759e-01 5.16095106e-03 -5.87092265e-02 -2.79120132e-02
2.75573164e-01 1.41993213e+00 1.36683416e+00 -1.47739196e+00
-7.02533245e-01 6.36643395e-02 -4.96048123e-01 4.59358394e-01
1.44470915e-01 -1.48948744e-01 -1.06505919e+00 -8.09028625e-01
-6.49711490e-02 -8.94066870e-01 -1.15965053e-01 -2.32050136e-01
-8.14227879e-01 -1.11333780e-01 6.80375695e-01 -7.54723549e-01
8.96612048e-01 -1.95765388e+00 3.29304397e-01 -1.09601066e-01
1.63505390e-01 6.73939884e-02 -6.12866938e-01 5.79451263e-01
-8.98447260e-03 -3.05404961e-02 -3.12304974e-01 -1.25891519e+00
2.14902952e-01 3.11128974e-01 -6.58863008e-01 1.21185631e-01
2.83316642e-01 1.34780645e+00 -5.80161810e-01 -8.58072877e-01
3.13014299e-01 4.79464859e-01 -6.24163330e-01 3.82985443e-01
-5.02514958e-01 3.66280943e-01 -5.20132519e-02 5.48246384e-01
6.98444188e-01 -4.55366969e-01 3.16792458e-01 -3.98639143e-01
3.65231819e-02 -7.04424158e-02 -8.92743409e-01 1.84434998e+00
-5.39447546e-01 7.49270141e-01 -7.99161121e-02 -7.05550551e-01
6.70315981e-01 3.57067317e-01 1.14699237e-01 -7.49213099e-01
8.67953226e-02 4.87196967e-02 -3.51863205e-01 -3.92607331e-01
8.15941632e-01 -2.26580217e-01 -4.04707491e-02 5.77526391e-01
4.24797475e-01 -3.11990738e-01 4.61631209e-01 5.59549868e-01
6.18120074e-01 5.40472090e-01 -2.54260689e-01 3.97717535e-01
1.76144674e-01 -2.46199429e-01 1.77572191e-01 1.11664057e+00
4.68991816e-01 9.72627580e-01 3.09531510e-01 1.46099821e-01
-1.22214019e+00 -1.32609046e+00 6.59577027e-02 1.18372810e+00
8.17241520e-02 -4.26253796e-01 -7.61982203e-01 -5.29099405e-01
-2.47951135e-01 9.32204485e-01 -5.24412096e-01 1.73104241e-01
-5.24792254e-01 -9.04093742e-01 7.56008744e-01 4.24105525e-01
6.79694235e-01 -1.30594051e+00 5.29329553e-02 8.48612487e-02
-5.18816471e-01 -1.51729834e+00 -8.22506487e-01 3.54834683e-02
-4.84942257e-01 -5.06972432e-01 -8.81475091e-01 -1.12394273e+00
7.66447902e-01 1.82730824e-01 1.11530507e+00 -2.67388046e-01
-2.73335397e-01 4.79830861e-01 -2.32592881e-01 -1.64469481e-01
-6.91876888e-01 -1.17376842e-01 -3.68630081e-01 2.97073126e-01
-3.39035213e-01 -4.79246795e-01 -4.77982342e-01 2.01811090e-01
-1.13573587e+00 1.07258701e+00 1.00079942e+00 1.08362401e+00
9.35093045e-01 -2.77936429e-01 3.61964822e-01 -9.04498875e-01
5.40026367e-01 -5.05643606e-01 -4.19176847e-01 5.30891836e-01
-6.47426665e-01 1.27825543e-01 3.40454668e-01 -6.32850945e-01
-1.56395793e+00 2.06126451e-01 -2.48519897e-01 -7.25554764e-01
-5.77313043e-02 4.45300192e-01 -2.25339100e-01 2.68216640e-01
2.19539270e-01 9.48786318e-01 -8.21165666e-02 -2.21406996e-01
9.09300864e-01 5.02175629e-01 7.48756588e-01 -5.80729604e-01
7.83972621e-01 3.67194980e-01 -1.67233169e-01 -6.58943176e-01
-5.58188915e-01 4.17093299e-02 -2.49399543e-01 -1.72936171e-01
1.15977681e+00 -1.24944603e+00 -4.31465119e-01 7.91409671e-01
-1.26984370e+00 -7.32027411e-01 -1.83736056e-01 2.08772287e-01
-7.93378234e-01 4.85352904e-01 -8.32315624e-01 -6.61376953e-01
-6.59986019e-01 -1.38602054e+00 1.64043784e+00 -3.49924527e-02
1.09852791e-01 -9.34236944e-01 -1.79894716e-01 5.32109141e-01
2.78844595e-01 -1.15493894e-01 5.81174612e-01 -2.59180307e-01
-9.93321180e-01 4.73567061e-02 -4.78311926e-01 2.97574371e-01
-2.11143658e-01 -1.20664388e-01 -8.57118070e-01 -3.17779511e-01
-2.47969911e-01 -5.86576223e-01 1.16552413e+00 5.55028379e-01
1.15054476e+00 -1.79774612e-01 -3.35884005e-01 9.13088262e-01
1.39815545e+00 1.64518803e-02 9.31254685e-01 -2.05067709e-01
1.07134676e+00 2.49584138e-01 4.33492661e-01 4.54439312e-01
6.20918274e-01 5.76294124e-01 2.90865332e-01 -3.71793509e-01
-6.77492261e-01 -6.91558361e-01 6.69044673e-01 7.73049533e-01
9.94181111e-02 -8.92556727e-01 -5.66071689e-01 5.48733711e-01
-2.00305200e+00 -9.11128163e-01 -2.92395562e-01 1.91454959e+00
9.40376759e-01 -1.38677582e-01 -3.52061778e-01 -5.99877656e-01
7.59934187e-01 1.42388090e-01 -4.06021446e-01 -2.66750883e-02
-4.63085175e-01 1.93966866e-01 3.54854912e-01 6.82992816e-01
-9.11425114e-01 1.38454628e+00 6.06081915e+00 1.33192837e+00
-9.97488916e-01 3.16570044e-01 8.59797239e-01 -1.60072818e-01
-5.04229426e-01 -4.92625013e-02 -1.19939983e+00 5.65259695e-01
6.60433531e-01 1.14638388e-01 4.90009308e-01 4.57220852e-01
2.29037404e-01 -2.52489597e-01 -1.08959019e+00 1.27448535e+00
3.18460464e-01 -1.50986981e+00 7.04737127e-01 1.09727830e-02
1.08259714e+00 9.61347744e-02 2.59056211e-01 3.98943841e-01
5.09197056e-01 -1.11080360e+00 1.05286610e+00 7.22496808e-01
1.29230857e+00 -2.80400723e-01 4.47678775e-01 3.74862611e-01
-1.20052242e+00 2.72386879e-01 -2.27269962e-01 4.80597764e-01
6.29650772e-01 5.52006960e-01 -9.64669406e-01 6.68960452e-01
4.38162327e-01 7.24429131e-01 -4.71609026e-01 5.43899655e-01
-3.40470374e-01 5.75580657e-01 -1.77932858e-01 3.46217722e-01
2.26345003e-01 -2.30083093e-01 4.28519160e-01 1.40182304e+00
6.49597943e-01 -2.63551235e-01 2.03307092e-01 1.20406604e+00
-1.91691011e-01 -8.33085105e-02 -4.95586455e-01 -1.79898009e-01
2.90319681e-01 1.44277430e+00 -5.74624002e-01 -7.03245163e-01
-2.57273883e-01 1.44051135e+00 1.57702431e-01 5.59469581e-01
-1.12246847e+00 7.71643072e-02 1.35230795e-01 2.48027090e-02
5.07724822e-01 -6.82372674e-02 -1.71016783e-01 -1.25588679e+00
-9.21887681e-02 -7.98716962e-01 1.37648001e-01 -1.38059556e+00
-1.29264927e+00 7.34967947e-01 8.62510055e-02 -9.75006104e-01
-6.76663816e-01 -2.24485546e-01 -4.68495369e-01 1.11339211e+00
-1.39199853e+00 -1.84348011e+00 -4.25804168e-01 9.39746082e-01
8.14751089e-01 -2.25854099e-01 4.80733395e-01 2.86238521e-01
-4.78522658e-01 6.77097380e-01 -2.66797878e-02 6.37956336e-02
8.85106027e-01 -1.07882452e+00 6.10798001e-01 9.35828388e-01
2.53945529e-01 3.37797433e-01 5.24745882e-01 -9.14860785e-01
-1.42755914e+00 -1.46481812e+00 8.63769352e-01 -3.66929442e-01
4.57360685e-01 -6.07573092e-01 -4.87592101e-01 1.08684492e+00
7.65993655e-01 -2.52530992e-01 1.86341897e-01 -4.54717696e-01
-1.14883065e-01 1.78938564e-02 -6.60082996e-01 8.02182972e-01
1.20239818e+00 -4.73975718e-01 -1.66489303e-01 4.86280948e-01
8.48380804e-01 -6.96698844e-01 -6.14310384e-01 2.39532128e-01
4.02814209e-01 -9.01338518e-01 1.02795887e+00 -1.85290262e-01
8.36933672e-01 -3.19611043e-01 -2.49568522e-01 -1.01192617e+00
-1.21359453e-01 -7.78356612e-01 1.73012167e-01 1.56517088e+00
4.47768450e-01 -5.00261486e-01 5.55801809e-01 2.46240363e-01
-4.13480550e-01 -5.74596226e-01 -6.85339212e-01 -4.48764563e-01
-1.24915339e-01 -6.82969749e-01 3.68916065e-01 6.13732457e-01
-7.02828944e-01 6.56464994e-01 -9.98472571e-01 -1.45660877e-01
7.61458933e-01 3.17435622e-01 1.07574964e+00 -3.86648208e-01
-6.44221067e-01 -2.84845203e-01 1.55125663e-01 -1.41725612e+00
8.53845999e-02 -1.12543499e+00 4.64471519e-01 -1.76404881e+00
5.74624717e-01 -1.98059931e-01 3.01314682e-01 7.26957083e-01
-3.18509877e-01 4.73117203e-01 3.20002466e-01 1.07316509e-01
-6.80911779e-01 8.50578785e-01 1.51271033e+00 -2.30741724e-01
9.63254049e-02 -4.08659965e-01 -4.47322428e-01 4.02295053e-01
2.84134597e-01 -2.77568102e-01 -6.80418313e-01 -5.92781901e-01
2.57658482e-01 4.08110231e-01 6.37784243e-01 -6.50152385e-01
2.50015944e-01 -5.23564741e-02 5.23353577e-01 -8.73680413e-01
5.32678366e-01 -3.35973471e-01 6.00907385e-01 -6.62414879e-02
-3.19245011e-01 -6.34780079e-02 4.38130610e-02 5.75818658e-01
-1.88635796e-01 -2.44549643e-02 4.51385111e-01 -1.20503232e-01
-8.00848782e-01 4.64104354e-01 -3.77011806e-01 -6.20666752e-03
8.26713681e-01 -2.07418352e-01 -5.25669813e-01 -5.91938198e-01
-8.07354093e-01 2.28744581e-01 2.96474665e-01 5.49082875e-01
9.40648794e-01 -1.37115157e+00 -9.70966041e-01 3.83155406e-01
1.92133635e-02 1.71022579e-01 6.76439404e-01 1.03825057e+00
-1.86843023e-01 4.24713701e-01 1.63271144e-01 -8.44809353e-01
-1.37143040e+00 4.38809663e-01 2.19665721e-01 -5.25966585e-01
-6.46661103e-01 8.70362759e-01 1.03202868e+00 -3.90039533e-01
3.10262181e-02 -1.27952024e-01 3.24874699e-01 -1.12826385e-01
4.04095292e-01 -1.35496572e-01 -2.09779322e-01 -6.14260733e-01
2.33773172e-01 3.93011481e-01 -5.39504699e-02 -4.67873454e-01
1.05907917e+00 -4.06980693e-01 -1.69342965e-01 2.39966497e-01
8.73266399e-01 -2.20648736e-01 -1.44347203e+00 -2.64202744e-01
-7.24997461e-01 -2.17845321e-01 -5.89126572e-02 -8.87550890e-01
-1.16456020e+00 9.33858275e-01 4.03286844e-01 -2.21872330e-01
1.26132524e+00 9.12405699e-02 9.41223502e-01 1.48486793e-01
2.19907954e-01 -8.12952459e-01 2.76657283e-01 5.27707338e-01
1.12645185e+00 -1.22573280e+00 -2.39018336e-01 -3.65284264e-01
-1.03816080e+00 6.64798021e-01 6.28224730e-01 2.32433304e-01
2.36602589e-01 3.56352657e-01 -1.44313350e-01 9.65074152e-02
-1.19540286e+00 -3.35951030e-01 4.54706430e-01 6.19427621e-01
3.62174422e-01 -3.96565907e-03 6.59331381e-02 3.62947583e-01
-9.57524627e-02 7.95281157e-02 2.45179132e-01 5.87177932e-01
-1.99563399e-01 -1.21003008e+00 -2.97760457e-01 5.44043660e-01
-1.51920170e-01 -7.75197208e-01 -1.12530634e-01 6.70173764e-01
2.27620021e-01 7.73038268e-01 2.48548254e-01 -3.21152151e-01
-1.93502620e-01 5.19666709e-02 7.29160130e-01 -5.43597877e-01
-4.37791973e-01 6.53699815e-01 8.33163783e-02 -4.09663826e-01
-3.11505586e-01 -6.47965848e-01 -1.14791644e+00 -3.17143589e-01
-1.02667861e-01 -2.90531397e-01 6.29825413e-01 1.11387038e+00
5.32676578e-01 5.76474607e-01 1.81304470e-01 -1.13205409e+00
-8.67964551e-02 -1.23905313e+00 -4.69613433e-01 3.49161059e-01
-1.47519475e-02 -2.37029836e-01 -2.00042874e-02 6.19363427e-01] | [11.210923194885254, 0.48760393261909485] |
66c3c5b9-14f7-4881-843b-fe799767228c | safe-continuous-control-with-constrained | 2104.06922 | null | https://arxiv.org/abs/2104.06922v1 | https://arxiv.org/pdf/2104.06922v1.pdf | Safe Continuous Control with Constrained Model-Based Policy Optimization | The applicability of reinforcement learning (RL) algorithms in real-world domains often requires adherence to safety constraints, a need difficult to address given the asymptotic nature of the classic RL optimization objective. In contrast to the traditional RL objective, safe exploration considers the maximization of expected returns under safety constraints expressed in expected cost returns. We introduce a model-based safe exploration algorithm for constrained high-dimensional control to address the often prohibitively high sample complexity of model-free safe exploration algorithms. Further, we provide theoretical and empirical analyses regarding the implications of model-usage on constrained policy optimization problems and introduce a practical algorithm that accelerates policy search with model-generated data. The need for accurate estimates of a policy's constraint satisfaction is in conflict with accumulating model-errors. We address this issue by quantifying model-uncertainty as the expected Kullback-Leibler divergence between predictions of an ensemble of probabilistic dynamics models and constrain this error-measure, resulting in an adaptive resampling scheme and dynamically limited rollout horizons. We evaluate this approach on several simulated constrained robot locomotion tasks with high-dimensional action- and state-spaces. Our empirical studies find that our algorithm reaches model-free performances with a 10-20 fold reduction of training samples while maintaining approximate constraint satisfaction levels of model-free methods. | ['J. Marius Zöllner', 'Karam Daaboul', 'Moritz A. Zanger'] | 2021-04-14 | null | null | null | null | ['safe-exploration'] | ['robots'] | [ 2.80403912e-01 3.07895243e-01 -4.68848139e-01 9.26060900e-02
-1.03774202e+00 -4.79941100e-01 5.66091955e-01 1.07838124e-01
-9.63989258e-01 1.55539286e+00 -1.66371614e-01 -4.65472609e-01
-6.31853282e-01 -4.90859687e-01 -6.70762122e-01 -6.18679523e-01
-7.77167320e-01 5.84114134e-01 8.89530703e-02 -9.03986022e-02
5.20411432e-01 3.44105870e-01 -1.66110921e+00 -5.47887564e-01
9.46158588e-01 1.14249861e+00 3.31650853e-01 7.78640628e-01
5.11874855e-01 4.85399604e-01 -5.15769064e-01 3.16939712e-01
2.87516385e-01 -4.30620015e-01 -5.42110205e-01 -8.60708132e-02
-3.27949345e-01 -4.63350028e-01 1.80549785e-01 1.06800604e+00
4.02655661e-01 7.41658509e-01 6.03473246e-01 -1.30033243e+00
9.43033099e-02 3.37266088e-01 -1.60063654e-01 1.25497565e-01
2.53256202e-01 6.09074712e-01 6.29662454e-01 -2.23745584e-01
4.96446699e-01 1.27700555e+00 5.12914896e-01 4.84577984e-01
-1.40163434e+00 -3.17888319e-01 2.21393541e-01 -1.37767091e-01
-1.16093516e+00 -1.76115632e-01 1.87403992e-01 -4.38882917e-01
1.01997530e+00 1.56942736e-02 8.97911847e-01 1.03858817e+00
7.01796472e-01 3.70751977e-01 1.22835827e+00 -3.80041659e-01
1.09908903e+00 -1.38437256e-01 -4.63793606e-01 5.70703626e-01
6.14879429e-01 9.95101213e-01 -3.08594018e-01 -2.92942971e-01
7.30901480e-01 -4.91002142e-01 -5.53760342e-02 -8.33776176e-01
-9.71220255e-01 1.03286564e+00 -3.27698588e-02 -3.01533371e-01
-5.52438676e-01 4.54789698e-01 5.00706315e-01 3.42789173e-01
1.53564081e-01 9.24495757e-01 -6.14405036e-01 -6.19425833e-01
-7.02864230e-01 8.90587211e-01 9.43038642e-01 8.94853830e-01
2.24675700e-01 4.85270351e-01 -3.48052196e-02 4.27914053e-01
3.63371015e-01 3.84784520e-01 4.34053630e-01 -1.41849625e+00
3.93530458e-01 8.57921690e-02 8.36645365e-01 -5.39824665e-01
-3.24882895e-01 -5.19144952e-01 -2.24693090e-01 9.32611048e-01
4.96482849e-01 -4.79749471e-01 -6.12324297e-01 1.93232727e+00
5.12344778e-01 -1.53295219e-01 2.05951780e-01 7.87457645e-01
-7.27909446e-01 4.90369111e-01 1.80719048e-01 -8.67909133e-01
7.87768602e-01 -5.07952094e-01 -8.53058934e-01 -2.67893493e-01
6.21883512e-01 -8.46171230e-02 1.29351044e+00 6.71083808e-01
-1.23753130e+00 -3.07385158e-02 -1.34707785e+00 6.45490348e-01
-8.35445449e-02 -3.94351989e-01 4.33573902e-01 5.53170323e-01
-5.76959133e-01 1.12685144e+00 -1.05948079e+00 -2.46404663e-01
3.06203008e-01 3.68082345e-01 1.08992688e-01 3.73174816e-01
-1.00246346e+00 1.51055956e+00 7.29674935e-01 -9.46664065e-02
-1.10522079e+00 -4.71534640e-01 -7.38560140e-01 -1.32713482e-01
9.02457833e-01 -2.52445936e-01 1.67357516e+00 -5.23896694e-01
-1.96153402e+00 -9.02269408e-02 4.43468630e-01 -8.49832475e-01
8.23501945e-01 -3.26596975e-01 -1.27840359e-02 -4.59015779e-02
-1.11408085e-02 3.18637580e-01 8.64118516e-01 -1.17211318e+00
-6.68599129e-01 -1.16929309e-02 -5.42282835e-02 5.09341002e-01
-2.29550507e-02 -3.45796764e-01 3.59862208e-01 -4.44555283e-01
-1.70340538e-01 -1.16830552e+00 -6.63009346e-01 1.43968746e-01
6.49330486e-03 9.51670036e-02 4.79191333e-01 -2.60836393e-01
1.32347393e+00 -1.61317694e+00 2.40676314e-01 2.64840931e-01
-5.12108684e-01 8.11909288e-02 5.84846027e-02 5.31752646e-01
3.96743715e-01 1.66027173e-02 -6.44138575e-01 -1.38171231e-02
3.81695777e-01 5.42661250e-01 -6.07791841e-01 6.11525536e-01
1.18860034e-02 4.72906917e-01 -1.17508781e+00 -4.38211799e-01
3.48049998e-01 -2.65594590e-02 -7.24368811e-01 2.32338578e-01
-7.65178978e-01 3.33465606e-01 -7.19747424e-01 4.42505836e-01
1.18641235e-01 2.17279598e-01 3.86794090e-01 4.92720723e-01
-3.91001105e-01 7.99862891e-02 -1.28662968e+00 1.42291594e+00
-4.33068186e-01 1.76705062e-01 2.67228186e-01 -9.70209479e-01
8.32390666e-01 -2.60871034e-02 3.79336834e-01 -5.67816138e-01
2.65445083e-01 3.96633774e-01 -9.76761729e-02 -2.57672608e-01
6.01796269e-01 -4.07520801e-01 -2.73477048e-01 3.89730155e-01
-1.50357217e-01 -7.61660039e-01 1.38796106e-01 -4.35660690e-01
1.01627004e+00 8.39891970e-01 6.08915448e-01 -7.57741392e-01
1.60255164e-01 2.31247813e-01 6.67828918e-01 9.07391846e-01
-4.55647498e-01 -7.95617104e-02 6.71300292e-01 -2.98925638e-01
-8.98500681e-01 -8.63596022e-01 -3.47653568e-01 7.93788195e-01
1.08167954e-01 -1.69564486e-01 -6.36235535e-01 -4.62263763e-01
3.82737249e-01 1.37530947e+00 -7.26943195e-01 -3.21071714e-01
-5.17871499e-01 -6.79527879e-01 2.38071099e-01 2.89218515e-01
5.78495786e-02 -9.76600766e-01 -1.71431363e+00 4.73444968e-01
3.49824548e-01 -5.96859574e-01 -1.18829809e-01 5.79609454e-01
-9.73750949e-01 -1.08353698e+00 -3.28904718e-01 -1.53266668e-01
4.24037009e-01 -6.82668805e-01 7.82805681e-01 -3.41600358e-01
-3.39475662e-01 4.19438809e-01 -9.80801135e-02 -5.49465895e-01
-5.29493034e-01 -3.14003795e-01 5.51582396e-01 -7.15994954e-01
-1.81446895e-01 -3.71378869e-01 -3.82985204e-01 2.90967375e-01
-5.55196762e-01 -2.30131522e-01 2.56578207e-01 1.11010218e+00
7.58570433e-01 1.41742840e-01 7.19439089e-01 -1.72363698e-01
1.10476005e+00 -4.40865904e-01 -1.39326656e+00 1.78405255e-01
-9.94249046e-01 5.25583327e-01 5.30288875e-01 -8.76554251e-01
-9.28266823e-01 -3.15757319e-02 4.18561578e-01 -4.22389001e-01
2.60610342e-01 5.31720817e-01 2.56950468e-01 3.14234802e-03
6.91516161e-01 7.25880638e-02 4.89541501e-01 -1.67083532e-01
1.90206081e-01 2.07008213e-01 2.89233088e-01 -1.19673526e+00
5.05606651e-01 3.74175832e-02 4.71916765e-01 -6.48317873e-01
-5.44273853e-01 3.23178589e-01 -2.08958551e-01 -3.93019080e-01
5.00864208e-01 -4.28185850e-01 -1.02781212e+00 -2.54738647e-02
-6.07185662e-01 -9.38318133e-01 -8.52221310e-01 6.90490305e-01
-1.60148942e+00 1.92527160e-01 -1.72346219e-01 -1.48101580e+00
-2.94838473e-02 -1.05234551e+00 5.96364141e-01 1.13166481e-01
-4.74871367e-01 -8.77087772e-01 4.38945085e-01 -4.30479288e-01
3.87065679e-01 7.56489515e-01 8.14237118e-01 -4.04786259e-01
-2.54979819e-01 -9.38393399e-02 3.81970257e-01 1.29374787e-01
-1.48762807e-01 -2.88195252e-01 -5.08908749e-01 -7.03241706e-01
1.88221380e-01 -7.84044802e-01 3.88660282e-01 5.59331357e-01
9.83325779e-01 -7.21939504e-01 -2.13845521e-01 2.52905548e-01
1.47091615e+00 5.53216636e-01 7.86702111e-02 6.53409183e-01
-1.82322338e-01 7.49606133e-01 1.41714633e+00 1.14751148e+00
-9.66085419e-02 6.74070597e-01 6.99005902e-01 8.14452767e-01
6.07507765e-01 -4.83967781e-01 3.79115850e-01 3.69367562e-02
2.22422466e-01 -1.34226039e-01 -7.61683106e-01 4.43588555e-01
-2.03544116e+00 -8.49006295e-01 5.79382300e-01 2.81268096e+00
1.01183033e+00 4.92316127e-01 2.63168633e-01 -3.51489820e-02
4.12352145e-01 -1.41512170e-01 -1.10555398e+00 -7.40300834e-01
4.17869478e-01 -4.76953387e-02 7.84919143e-01 8.96296382e-01
-8.47698510e-01 6.10655427e-01 6.73742342e+00 7.87482679e-01
-6.22704864e-01 -2.75176823e-01 2.39254713e-01 -4.36401427e-01
-2.03634445e-02 1.13359913e-01 -6.34602845e-01 7.25396514e-01
1.30986285e+00 -6.87286139e-01 7.32513964e-01 1.16470933e+00
5.01462817e-01 -7.34589994e-01 -1.16353452e+00 4.27774459e-01
-6.02488816e-01 -1.10122287e+00 -5.23168445e-01 3.74210685e-01
6.13882244e-01 -1.72892421e-01 7.00250641e-02 6.21684849e-01
6.24884725e-01 -9.49551702e-01 1.08067226e+00 6.67679369e-01
6.97354078e-01 -1.14637661e+00 3.82800370e-01 6.55808508e-01
-8.27480078e-01 -6.57224655e-01 -2.50357151e-01 -3.84010911e-01
3.13662827e-01 1.57698601e-01 -7.34235764e-01 8.22064877e-02
3.41236144e-01 2.62941360e-01 1.70871332e-01 9.97931242e-01
6.18966036e-02 3.69581401e-01 -7.78368652e-01 -6.39921010e-01
5.59214175e-01 -3.48888934e-01 7.68971741e-01 8.61771047e-01
4.39463228e-01 -5.18422015e-03 2.72295505e-01 9.57854748e-01
8.44486117e-01 -3.50811243e-01 -7.62761712e-01 -1.82076752e-01
7.26556778e-01 5.87755859e-01 -7.00498700e-01 -4.65372056e-02
3.51630121e-01 3.69187176e-01 2.28907987e-01 1.97193429e-01
-8.07694197e-01 -4.10080612e-01 6.53901517e-01 -1.84789613e-01
2.38137260e-01 -5.62708020e-01 -2.39519954e-01 -6.52670741e-01
-5.31351827e-02 -8.80833030e-01 3.13241482e-01 -2.68517345e-01
-8.58045101e-01 2.15494409e-01 6.41267538e-01 -1.30269587e+00
-9.70693171e-01 -5.37352204e-01 -3.63491207e-01 7.28529871e-01
-1.22884786e+00 -3.11866432e-01 4.20191377e-01 1.83423281e-01
6.60605192e-01 -6.27546310e-02 8.94176483e-01 -5.05680263e-01
-4.84284043e-01 2.15855435e-01 4.26081389e-01 -8.14172149e-01
2.22204447e-01 -1.24082053e+00 1.50264397e-01 5.68430126e-01
-8.57569456e-01 4.05616462e-01 1.31280923e+00 -9.60511327e-01
-1.54687226e+00 -8.26622427e-01 2.76083142e-01 -3.83806646e-01
9.20851052e-01 -7.80603439e-02 -7.02874780e-01 4.30675298e-01
-2.67086834e-01 -9.89280120e-02 8.36778060e-02 -1.41576275e-01
2.74723500e-01 2.67906755e-01 -1.45562017e+00 7.90253520e-01
8.29146147e-01 -5.91302998e-02 -5.99545479e-01 1.76582858e-01
6.13898933e-01 -3.95876795e-01 -9.21391964e-01 6.06359720e-01
7.42327988e-01 -6.65619731e-01 6.85553491e-01 -6.68707013e-01
-5.09737507e-02 -1.03553362e-01 -3.67829084e-01 -1.41494524e+00
2.53788587e-02 -9.93080676e-01 -4.51687425e-01 5.15618682e-01
8.10855925e-02 -6.67167664e-01 6.74849868e-01 8.33129883e-01
-4.57472168e-02 -1.12826645e+00 -1.45078492e+00 -1.29774106e+00
4.23005223e-01 -4.54268873e-01 1.99716941e-01 5.34002721e-01
5.08739889e-01 -3.48806888e-01 -3.86856288e-01 -1.29709877e-02
1.05450869e+00 -9.92704183e-02 4.45328116e-01 -8.41133416e-01
-5.11194289e-01 -4.88288820e-01 1.30226389e-01 -6.18377864e-01
3.35188717e-01 -1.43998340e-01 6.13313675e-01 -1.19570601e+00
-4.21416938e-01 -5.90985060e-01 -6.31869361e-02 2.80672371e-01
1.63126513e-01 -6.40426278e-01 1.79312766e-01 -7.36786500e-02
-6.05158806e-01 1.08610380e+00 1.04791427e+00 3.65658104e-01
-5.19801259e-01 1.02833092e-01 -7.41403624e-02 7.67343700e-01
9.10972178e-01 -4.56604272e-01 -9.37630236e-01 1.46667302e-01
2.11394459e-01 6.01791203e-01 1.09901458e-01 -1.16877651e+00
-4.85172868e-02 -9.50111270e-01 3.24416347e-02 -3.86000842e-01
3.16478699e-01 -7.46241331e-01 6.80992231e-02 1.03238988e+00
-7.88090885e-01 8.90392959e-02 4.68612343e-01 1.01811945e+00
3.73311311e-01 -3.39673191e-01 9.88149762e-01 -1.91254050e-01
-5.29131711e-01 -5.13421446e-02 -8.87537539e-01 2.48155415e-01
1.28557777e+00 -2.37802178e-01 1.28835544e-01 -2.44306192e-01
-6.28984749e-01 5.93560576e-01 5.15594244e-01 2.79676408e-01
5.90130627e-01 -8.83176446e-01 -2.30006069e-01 9.97581184e-02
-1.60405964e-01 -1.58926189e-01 -1.95908681e-01 4.93187875e-01
-2.96874225e-01 4.06663656e-01 -3.70526165e-01 -2.21212417e-01
-6.85540676e-01 6.07558608e-01 6.06750548e-01 -4.12696987e-01
-4.53053653e-01 4.29442972e-01 -4.33407515e-01 -3.16156298e-01
4.86776322e-01 -4.83197808e-01 2.40247726e-01 -7.99181759e-02
2.58837461e-01 7.21740484e-01 -3.47823799e-01 9.26058665e-02
-2.76977777e-01 2.50778735e-01 2.19558984e-01 -7.64382243e-01
1.02728510e+00 -4.33389992e-02 4.94794607e-01 5.81855774e-01
6.35878503e-01 -7.15305328e-01 -2.10128069e+00 2.33345106e-01
3.79339963e-01 -4.53068644e-01 1.69894502e-01 -1.00820744e+00
-2.40832046e-02 4.95400399e-01 6.93807781e-01 6.53207526e-02
6.22392535e-01 -6.94326103e-01 1.52824253e-01 7.27188170e-01
9.43548858e-01 -1.67908311e+00 -1.42480098e-02 5.25513887e-01
9.89271045e-01 -9.42053914e-01 3.69198740e-01 4.24733192e-01
-7.26586998e-01 9.02223825e-01 7.15076506e-01 -2.98182994e-01
3.77385050e-01 3.77779871e-01 -4.41204637e-01 3.25902998e-01
-1.10458803e+00 5.10607474e-02 -2.72564501e-01 6.15076661e-01
-3.31083357e-01 1.02232262e-01 -5.84398627e-01 3.76275033e-01
-6.73984289e-02 2.58201748e-01 4.40572649e-01 1.48861158e+00
-9.14762020e-01 -8.44068110e-01 -3.15080315e-01 3.01097393e-01
-1.14804223e-01 3.85505646e-01 2.53536701e-01 1.03168321e+00
-4.08929557e-01 8.84351313e-01 4.38735709e-02 -2.89822333e-02
1.67317346e-01 8.04091617e-02 5.62032878e-01 -3.84112865e-01
-1.45651937e-01 2.92096417e-02 3.98329616e-01 -9.66083050e-01
4.51509394e-02 -9.03638542e-01 -1.38363230e+00 -3.32469642e-02
-3.11076492e-01 3.51634234e-01 7.98639417e-01 9.18136537e-01
1.96941033e-01 2.45904833e-01 5.71325302e-01 -1.06045258e+00
-1.81181741e+00 -6.33455753e-01 -6.00119889e-01 -1.87907085e-01
6.02603257e-01 -1.21925533e+00 -6.28334522e-01 -4.99292046e-01] | [4.400390625, 2.225665330886841] |
228cfd0a-f11a-483c-88a0-4511d8831400 | adv-bot-realistic-adversarial-botnet-attacks | 2303.06664 | null | https://arxiv.org/abs/2303.06664v1 | https://arxiv.org/pdf/2303.06664v1.pdf | Adv-Bot: Realistic Adversarial Botnet Attacks against Network Intrusion Detection Systems | Due to the numerous advantages of machine learning (ML) algorithms, many applications now incorporate them. However, many studies in the field of image classification have shown that MLs can be fooled by a variety of adversarial attacks. These attacks take advantage of ML algorithms' inherent vulnerability. This raises many questions in the cybersecurity field, where a growing number of researchers are recently investigating the feasibility of such attacks against machine learning-based security systems, such as intrusion detection systems. The majority of this research demonstrates that it is possible to fool a model using features extracted from a raw data source, but it does not take into account the real implementation of such attacks, i.e., the reverse transformation from theory to practice. The real implementation of these adversarial attacks would be influenced by various constraints that would make their execution more difficult. As a result, the purpose of this study was to investigate the actual feasibility of adversarial attacks, specifically evasion attacks, against network-based intrusion detection systems (NIDS), demonstrating that it is entirely possible to fool these ML-based IDSs using our proposed adversarial algorithm while assuming as many constraints as possible in a black-box setting. In addition, since it is critical to design defense mechanisms to protect ML-based IDSs against such attacks, a defensive scheme is presented. Realistic botnet traffic traces are used to assess this work. Our goal is to create adversarial botnet traffic that can avoid detection while still performing all of its intended malicious functionality. | ['Wim Mees', 'Jean-Michel Dricot', 'Thibault Debatty', 'Tayeb Kenaza', 'Benjamin Cochez', 'Islam Debicha'] | 2023-03-12 | null | null | null | null | ['network-intrusion-detection'] | ['miscellaneous'] | [ 3.17841560e-01 3.44136246e-02 3.83895822e-02 -7.31737763e-02
2.27077037e-01 -1.11991870e+00 7.87622511e-01 -1.94058672e-01
-4.77469653e-01 4.15336937e-01 -8.13337326e-01 -1.05101573e+00
2.49819476e-02 -1.06965160e+00 -4.95778531e-01 -5.19821584e-01
-4.35821533e-01 1.07710779e-01 5.00186682e-01 -3.04209262e-01
4.57172751e-01 1.26726770e+00 -1.20258653e+00 1.61093667e-01
4.21280742e-01 6.54525995e-01 -5.23843765e-01 1.13027978e+00
-1.18198562e-02 8.96520019e-01 -1.13793516e+00 -5.02878547e-01
8.26826036e-01 -2.67699957e-01 -7.10895121e-01 -2.08190814e-01
3.54159266e-01 -4.47048396e-01 -4.29971665e-01 1.20208597e+00
1.56245321e-01 -2.49735206e-01 4.06771183e-01 -2.12127972e+00
-2.44160756e-01 3.83675665e-01 -2.44647130e-01 1.48387209e-01
1.20952912e-02 7.64226377e-01 4.81255859e-01 -2.87457369e-03
4.10166234e-01 1.39547431e+00 3.07137966e-01 8.20564866e-01
-1.23256981e+00 -1.13545644e+00 -1.13265894e-01 -1.13146596e-01
-1.04265749e+00 -2.33642176e-01 7.21604109e-01 -3.88111204e-01
8.02522063e-01 4.61399823e-01 1.89057410e-01 1.28029561e+00
5.63579023e-01 1.15790755e-01 1.48065782e+00 -5.93297064e-01
3.86055082e-01 7.83928573e-01 1.32246345e-01 5.71229935e-01
7.90695965e-01 5.95367551e-01 4.25139397e-01 -6.22127950e-01
6.75017834e-01 -1.50293574e-01 2.88549829e-02 -1.75793678e-01
-5.39738119e-01 1.13353837e+00 3.38437527e-01 3.80759895e-01
-2.50203341e-01 5.21516427e-02 6.01199865e-01 6.54008746e-01
-1.23242937e-01 7.49470472e-01 -5.34299850e-01 2.41570160e-01
-4.48720872e-01 3.04409303e-02 1.15980792e+00 4.15713817e-01
5.56166351e-01 4.56452459e-01 6.54872179e-01 1.31508157e-01
4.70301330e-01 7.10281253e-01 1.63092077e-01 -6.82654500e-01
6.16255924e-02 5.00237525e-01 -2.56038100e-01 -1.39957261e+00
-1.66066483e-01 -9.88716930e-02 -5.02671897e-01 1.10737729e+00
6.55723810e-01 -3.99580657e-01 -7.66687155e-01 1.72220874e+00
3.93828213e-01 3.14824104e-01 1.96502179e-01 4.93991792e-01
6.15070872e-02 5.17977536e-01 3.12082261e-01 -1.69861659e-01
1.12278128e+00 -2.65933931e-01 -2.37854570e-01 -8.69025588e-02
5.36801934e-01 -7.61552870e-01 8.85791659e-01 3.55081797e-01
-4.15160924e-01 -3.63047361e-01 -1.16817153e+00 9.21546459e-01
-9.07392204e-01 -6.76816881e-01 7.94209599e-01 1.72632432e+00
-8.14372778e-01 4.81767595e-01 -7.93045521e-01 -5.60095310e-01
3.22454780e-01 7.21842885e-01 -2.79022723e-01 3.17485213e-01
-1.29045033e+00 9.47121680e-01 3.59469086e-01 -3.62457603e-01
-9.70507145e-01 -4.04569685e-01 -4.77285594e-01 -1.91087380e-01
3.79827440e-01 -2.87113249e-01 8.32319260e-01 -1.26767981e+00
-1.24352789e+00 7.25709319e-01 5.09918630e-01 -9.14409518e-01
5.51735222e-01 2.08521456e-01 -8.03204119e-01 2.10949868e-01
-2.11793408e-01 2.89694279e-01 1.26498222e+00 -1.31091666e+00
-4.59603339e-01 -1.64801583e-01 7.20003009e-01 -5.47454357e-01
-4.70362961e-01 4.77122426e-01 5.63807428e-01 -4.36572790e-01
-5.25022924e-01 -1.23133612e+00 -3.69446844e-01 -1.22991838e-01
-5.74348509e-01 2.44893759e-01 1.57127643e+00 -9.27362293e-02
9.56633151e-01 -1.99319696e+00 -5.51004410e-01 6.66316867e-01
2.24482670e-01 1.05944037e+00 -1.21883467e-01 6.48880601e-01
-1.84627637e-01 8.88790488e-01 -1.78942177e-02 3.86149645e-01
-7.43637681e-02 3.42062801e-01 -7.78118849e-01 5.77160835e-01
4.72947448e-01 5.16431153e-01 -5.42513549e-01 -3.36426735e-01
4.46635842e-01 3.70687604e-01 -6.41588688e-01 3.25108260e-01
-1.32815763e-01 4.61384654e-01 -6.63839638e-01 5.83208680e-01
6.42747819e-01 1.78804994e-01 2.98405796e-01 -1.38152633e-02
4.53704596e-02 -1.11975908e-01 -1.13432050e+00 2.37943724e-01
-4.22996014e-01 6.27864122e-01 1.73960134e-01 -1.07172894e+00
6.65874064e-01 2.29464546e-01 4.13193703e-01 -3.09110165e-01
5.03612041e-01 1.30615488e-01 7.14150667e-01 -3.62366855e-01
-7.93246645e-03 -1.62208360e-02 7.24737942e-02 8.19797635e-01
-2.68783778e-01 9.89886597e-02 -2.14605797e-02 3.21829706e-01
1.50040960e+00 -3.35148245e-01 3.60417515e-01 -1.00456856e-01
9.98387516e-01 1.32676646e-01 3.82974386e-01 1.11448228e+00
-5.48131287e-01 -2.74648875e-01 5.28820157e-01 -5.29321730e-01
-9.90833700e-01 -1.09349298e+00 -2.11452376e-02 6.40969634e-01
-1.51437782e-02 -1.91152766e-01 -9.85799432e-01 -1.23407626e+00
-5.80375828e-02 6.21618807e-01 -4.05070841e-01 -4.43983436e-01
-7.19331264e-01 -7.16490805e-01 1.40714574e+00 -9.84876007e-02
6.35726154e-01 -1.07734823e+00 -8.77046168e-01 2.89997846e-01
4.92305100e-01 -1.33322167e+00 1.27458826e-01 -2.61462703e-02
-6.12936676e-01 -1.70912313e+00 3.36178869e-01 -2.74219990e-01
7.68869579e-01 3.85326594e-01 6.52261436e-01 6.45521164e-01
-6.35655463e-01 5.27639031e-01 -5.54332674e-01 -7.01859236e-01
-1.36907780e+00 -7.18264282e-02 3.63530219e-01 5.56566454e-02
5.39793611e-01 -6.29350364e-01 -2.00541034e-01 5.19897878e-01
-1.37846792e+00 -6.35163963e-01 6.61307573e-01 4.28224593e-01
-1.86296761e-01 7.60412037e-01 8.25542748e-01 -1.22417605e+00
7.30523348e-01 -5.00601172e-01 -8.61339211e-01 1.83867991e-01
-6.34857476e-01 -1.60321489e-01 1.37026238e+00 -9.79562461e-01
-6.73630595e-01 -2.84746051e-01 -1.90243214e-01 -2.76953042e-01
-7.08493888e-01 -1.15217485e-01 -4.40886348e-01 -8.02053452e-01
1.03772652e+00 2.39173211e-02 3.35470229e-01 5.05823307e-02
1.56809226e-01 7.65577912e-01 -4.17941362e-02 -6.23847783e-01
1.64112926e+00 6.10415459e-01 4.36381608e-01 -1.23374188e+00
-2.79508561e-01 8.09808969e-02 -2.89087772e-01 -4.00210738e-01
5.35416842e-01 -2.88813449e-02 -1.06388342e+00 6.36329353e-01
-1.00432980e+00 -2.07205221e-01 2.63624996e-01 1.84900582e-01
-2.50387162e-01 5.89541674e-01 -4.87721175e-01 -9.75992799e-01
-2.06064701e-01 -1.34435976e+00 -7.37310993e-03 2.05921859e-01
-1.08590722e-01 -1.10246730e+00 -4.15127911e-02 3.50067347e-01
7.38934040e-01 6.02135420e-01 9.35458004e-01 -1.35364628e+00
-6.21505916e-01 -4.73868221e-01 -7.88646787e-02 7.10723281e-01
3.01448375e-01 5.82755148e-01 -9.21684086e-01 -5.76012373e-01
4.11481917e-01 -6.32563978e-02 8.86271745e-02 -1.56561226e-01
6.19145274e-01 -8.16574037e-01 -7.21019208e-02 4.09668565e-01
1.68538165e+00 6.43063128e-01 7.64504790e-01 5.80425024e-01
4.80132073e-01 7.01220870e-01 5.57233036e-01 2.05099508e-02
-3.63695115e-01 4.70004976e-01 8.41277063e-01 8.02288577e-02
3.17529231e-01 -1.61785819e-02 6.00898027e-01 -9.36860070e-02
2.06677377e-01 -1.58008918e-01 -8.91657472e-01 -1.95668206e-01
-1.29228783e+00 -1.16872549e+00 -2.28290007e-01 2.27273369e+00
3.27840388e-01 6.83624446e-01 3.05525988e-01 4.30680722e-01
8.03125143e-01 7.50789791e-02 -4.45735127e-01 -1.04466414e+00
2.39454821e-01 4.98410523e-01 8.78376245e-01 4.95707661e-01
-1.28105307e+00 1.08255780e+00 5.60899305e+00 6.56299710e-01
-1.49914646e+00 -7.21861944e-02 3.40541542e-01 4.85377520e-01
3.53449285e-01 5.07293344e-01 -6.08762383e-01 5.03720760e-01
1.32053411e+00 -1.71609774e-01 6.57247066e-01 9.86484289e-01
3.94325197e-01 1.49474204e-01 -6.44151151e-01 7.66050890e-02
-7.63034299e-02 -9.81405377e-01 9.43767652e-02 4.74439740e-01
5.76341413e-02 -2.63227105e-01 9.09665227e-02 4.25041705e-01
5.63347936e-01 -1.16272271e+00 1.59463346e-01 -2.35113591e-01
1.71201259e-01 -9.55350876e-01 5.96752226e-01 5.55468202e-01
-8.20929706e-01 -2.17369631e-01 -2.84116238e-01 -2.77055562e-01
-1.85344636e-01 2.30016336e-01 -1.10736382e+00 2.55533516e-01
2.14464888e-01 -1.98939860e-01 -7.44120240e-01 6.21591866e-01
-2.35810190e-01 1.02935004e+00 -4.00352865e-01 1.04492065e-02
4.37764466e-01 8.36926028e-02 6.53910458e-01 1.06400859e+00
-3.03910315e-01 -3.70957553e-02 2.74155110e-01 8.30011725e-01
2.76770979e-01 -4.93666492e-02 -1.26702321e+00 -3.37181985e-01
5.99588335e-01 1.29464018e+00 -9.33616161e-01 4.05508876e-02
-2.53756672e-01 5.86078525e-01 -1.27419427e-01 1.45966768e-01
-1.10597801e+00 -5.11851728e-01 1.01110423e+00 2.48322099e-01
-2.84237295e-01 -4.62686509e-01 -1.15558662e-01 -8.87615740e-01
-4.55292046e-01 -1.47872841e+00 3.48459423e-01 -5.76629676e-03
-1.33544075e+00 6.60682023e-01 -4.23162356e-02 -9.96051550e-01
-2.15466976e-01 -6.30997300e-01 -8.97272885e-01 6.06094182e-01
-1.24056566e+00 -1.28086567e+00 2.53564358e-01 6.84084594e-01
2.26850018e-01 -5.98105371e-01 7.36366451e-01 1.67995334e-01
-6.09007180e-01 7.37141967e-01 -4.18736845e-01 5.25755525e-01
4.20287877e-01 -7.06252933e-01 4.10023481e-01 1.36165941e+00
2.07849547e-01 1.00060618e+00 8.69886935e-01 -7.64478087e-01
-1.33054972e+00 -1.09276056e+00 1.96293637e-01 -4.78872359e-01
9.04271781e-01 -3.29153419e-01 -8.99211764e-01 6.92684293e-01
-8.63666087e-03 -5.53719848e-02 5.61793089e-01 -4.93215591e-01
-7.10136533e-01 9.14425030e-02 -1.84442818e+00 8.75227511e-01
4.78333622e-01 -5.64311802e-01 -2.97878057e-01 2.91021407e-01
6.49320602e-01 3.75897139e-01 -5.34261048e-01 3.26742232e-01
4.73593086e-01 -1.15777016e+00 1.16671240e+00 -1.00136065e+00
-1.22941181e-01 -4.35201705e-01 -1.10699952e-01 -8.25582623e-01
1.87459171e-01 -6.23048127e-01 -6.81508705e-02 1.31879592e+00
8.02220404e-02 -1.34169340e+00 7.03184307e-01 6.33255720e-01
5.75978935e-01 -2.77996659e-01 -7.60130465e-01 -1.18673444e+00
2.81800628e-01 -5.42105913e-01 3.40518862e-01 1.19960678e+00
-2.60513365e-01 -2.78810393e-02 -3.09505254e-01 8.18114758e-01
9.81807768e-01 -5.04394472e-01 1.27554607e+00 -1.07740271e+00
-4.05124515e-01 -3.91253233e-01 -8.96975636e-01 -5.30329607e-02
3.83667052e-01 -5.73861897e-01 -5.75464606e-01 -5.22383332e-01
-4.27301675e-01 -7.53879488e-01 -1.67843506e-01 4.65647519e-01
2.26346493e-01 2.79800415e-01 5.59771776e-01 2.88130552e-01
1.46502912e-01 -3.45044345e-01 6.17752492e-01 -4.13885601e-02
-4.04509716e-02 2.32642487e-01 -4.33090419e-01 9.80317771e-01
1.31128395e+00 -8.46257210e-01 -6.63039625e-01 3.86847198e-01
-2.55665332e-01 -3.01506996e-01 5.86546957e-01 -1.07130957e+00
1.76825106e-01 -7.09111094e-01 -1.92976996e-01 2.60992557e-01
4.13050205e-02 -1.24047375e+00 9.02099982e-02 1.03495491e+00
-7.98861124e-03 3.02646775e-03 1.65286034e-01 3.48839998e-01
1.30757987e-01 -3.51273119e-01 1.21970856e+00 -2.11824641e-01
-5.03249943e-01 1.49533734e-01 -7.66747415e-01 2.28466485e-02
1.60108876e+00 -3.50234360e-01 -6.31454289e-01 -2.59773791e-01
-2.79649526e-01 -1.80610031e-01 8.39771986e-01 4.42012936e-01
4.21197623e-01 -6.47732913e-01 -3.76292199e-01 4.41717952e-01
-3.34459633e-01 -7.90555179e-01 -1.75853893e-01 4.27531838e-01
-9.46704447e-01 3.28888029e-01 -6.25855625e-01 -1.64267153e-01
-1.58138442e+00 9.69729841e-01 5.28447807e-01 -3.48493427e-01
-3.74022037e-01 2.57041991e-01 4.64835279e-02 -4.21328843e-01
1.33055866e-01 4.45651650e-01 -5.41126132e-02 -4.71065879e-01
5.61011553e-01 2.25890368e-01 -2.99879164e-01 -4.64255393e-01
-5.53472519e-01 1.25711918e-01 -2.11320475e-01 -2.94090342e-02
8.27781796e-01 2.93262541e-01 -2.21400648e-01 -3.34653586e-01
9.22695994e-01 2.15688631e-01 -5.65990448e-01 1.47166222e-01
1.80284381e-01 -8.68898690e-01 -3.28797638e-01 -7.53326833e-01
-1.07938159e+00 7.80461967e-01 5.56285560e-01 9.78059232e-01
1.01126242e+00 -7.02392399e-01 5.75696766e-01 6.32854700e-01
7.34923542e-01 -7.10666418e-01 1.37726411e-01 3.10338616e-01
1.99574456e-01 -1.24112022e+00 -2.09548309e-01 -6.72479689e-01
-3.11830610e-01 1.24001002e+00 8.02052200e-01 -5.36078811e-01
7.13765681e-01 5.71104586e-01 3.04011345e-01 -1.31475329e-01
-5.48709929e-01 1.57943085e-01 -3.85902107e-01 1.07255995e+00
-1.69049993e-01 6.48699775e-02 -3.50376278e-01 -1.05750687e-01
1.20136283e-01 -3.03641468e-01 1.14092171e+00 1.23633051e+00
-5.12860656e-01 -1.67631674e+00 -7.67483592e-01 2.54889816e-01
-7.93029308e-01 2.12940589e-01 -8.47271383e-01 1.16490030e+00
8.49432871e-02 1.41766727e+00 -5.34150779e-01 -7.50385642e-01
1.30504131e-01 -3.41422483e-02 -1.99943478e-03 -5.28418779e-01
-1.08434367e+00 -6.27137184e-01 1.53728366e-01 -4.72475827e-01
-2.62655884e-01 -2.45633185e-01 -1.02875960e+00 -9.53255415e-01
-3.10781419e-01 1.42179266e-01 8.52787375e-01 8.44515443e-01
1.76888496e-01 3.40310812e-01 1.14830160e+00 -3.62286210e-01
-9.22797561e-01 -3.83801162e-01 -3.90836686e-01 3.81463051e-01
2.25059073e-02 -6.17187977e-01 -8.94453585e-01 -2.33805642e-01] | [5.50906229019165, 7.526325225830078] |
77aa180c-6328-452a-8c84-74be21525c38 | supervised-fine-tuning-evaluation-for-long | 2211.07696 | null | https://arxiv.org/abs/2211.07696v1 | https://arxiv.org/pdf/2211.07696v1.pdf | Supervised Fine-tuning Evaluation for Long-term Visual Place Recognition | In this paper, we present a comprehensive study on the utility of deep convolutional neural networks with two state-of-the-art pooling layers which are placed after convolutional layers and fine-tuned in an end-to-end manner for visual place recognition task in challenging conditions, including seasonal and illumination variations. We compared extensively the performance of deep learned global features with three different loss functions, e.g. triplet, contrastive and ArcFace, for learning the parameters of the architectures in terms of fraction of the correct matches during deployment. To verify effectiveness of our results, we utilized two real world datasets in place recognition, both indoor and outdoor. Our investigation demonstrates that fine tuning architectures with ArcFace loss in an end-to-end manner outperforms other two losses by approximately 1~4% in outdoor and 1~2% in indoor datasets, given certain thresholds, for the visual place recognition tasks. | ['Esa Rahtu', 'Farid Alijani'] | 2022-11-14 | null | null | null | null | ['visual-place-recognition'] | ['computer-vision'] | [-1.86176389e-01 -3.87550682e-01 2.43378058e-01 -7.97808290e-01
-4.18875962e-01 -8.44224811e-01 5.82315028e-01 1.25089079e-01
-7.86549330e-01 8.79071772e-01 -3.98488902e-02 -5.55842277e-03
-3.56239229e-01 -8.11729014e-01 -1.27988780e+00 -5.59999049e-01
-5.14620483e-01 -1.06181763e-01 1.88242570e-01 -2.11611718e-01
3.02044481e-01 8.14231634e-01 -1.65035319e+00 5.87803870e-02
6.67501330e-01 1.30682468e+00 -1.44222910e-02 7.26638138e-01
2.29903594e-01 5.82541943e-01 -5.72635829e-01 -3.86833519e-01
7.28947878e-01 3.59345645e-01 -5.53023696e-01 -1.67104095e-01
7.27618039e-01 -2.54936337e-01 -5.52718759e-01 6.75865531e-01
7.74165809e-01 3.63247216e-01 4.65762973e-01 -1.27160478e+00
-7.39913523e-01 2.73775822e-03 -3.02689791e-01 4.90184247e-01
1.48479968e-01 3.15218836e-01 7.07128525e-01 -9.74623144e-01
3.55334699e-01 1.12848175e+00 1.08272779e+00 -5.86615549e-03
-8.62726927e-01 -7.23059416e-01 3.48025948e-01 1.34722888e-01
-1.48430252e+00 -4.91613150e-01 5.46370268e-01 -4.69838738e-01
1.37583578e+00 3.88710201e-02 1.23134255e-01 1.04807603e+00
2.89705455e-01 4.12748307e-01 1.04023099e+00 -3.77176225e-01
8.02545473e-02 1.60326123e-01 -2.04012170e-01 8.47168446e-01
2.05999687e-01 2.61403590e-01 -6.66021228e-01 5.58527187e-02
8.21818769e-01 9.29966792e-02 -4.05591242e-02 -2.73928255e-01
-1.13976240e+00 5.32370865e-01 1.23564672e+00 1.57233849e-01
-4.43630427e-01 5.80196738e-01 3.88207078e-01 2.85845578e-01
4.02361453e-01 2.93695003e-01 -4.56906825e-01 6.87663779e-02
-8.55403364e-01 3.98868561e-01 4.36799526e-01 1.06096911e+00
1.06441367e+00 -2.34606832e-01 -5.24757206e-01 9.61423457e-01
4.75152731e-01 4.86212522e-01 4.14034873e-01 -4.40911680e-01
6.09767556e-01 3.43489498e-01 3.72308910e-01 -1.24834228e+00
-4.14564788e-01 -8.21004510e-01 -6.27252102e-01 2.64421433e-01
1.26603127e-01 -4.10716802e-01 -1.35866261e+00 1.78947163e+00
1.43687934e-01 3.55879575e-01 5.74561907e-03 9.14840221e-01
7.38018334e-01 5.46401799e-01 1.74656853e-01 7.53565609e-01
1.18398166e+00 -1.42697346e+00 -2.82742560e-01 -5.37593782e-01
1.60055488e-01 -8.73913467e-01 1.01567471e+00 -1.47839040e-01
-6.93135321e-01 -7.89503515e-01 -1.45249760e+00 -1.84039623e-01
-8.29812348e-01 5.17342985e-01 5.45539498e-01 5.99800587e-01
-1.29554462e+00 7.45111763e-01 -8.21139157e-01 -7.38847852e-01
6.30134344e-01 5.24627388e-01 -6.56512678e-01 -1.24674134e-01
-8.96092832e-01 8.06417167e-01 4.74836119e-02 4.77511466e-01
-1.30816936e+00 -7.58836329e-01 -8.07059467e-01 3.08061421e-01
-2.30698600e-01 -6.81695282e-01 1.10104918e+00 -8.66971374e-01
-1.12227416e+00 9.86618400e-01 -1.21851407e-01 -6.52704537e-01
7.61642814e-01 -4.69757259e-01 -3.86573493e-01 -2.10023150e-01
2.78069586e-01 7.12797582e-01 4.25444752e-01 -1.13809752e+00
-7.46989667e-01 -3.94342065e-01 4.25149560e-01 1.30306989e-01
8.48420803e-03 -1.29704684e-01 -2.97138035e-01 -4.48553562e-01
-1.09387629e-01 -8.36137056e-01 -3.14835489e-01 4.39552635e-01
-4.37433213e-01 5.78844398e-02 6.43111825e-01 -5.22280097e-01
5.98776400e-01 -2.40683055e+00 -5.03697097e-01 3.25054049e-01
-2.01989502e-01 5.80895096e-02 -1.04569435e-01 5.75187683e-01
1.49787813e-01 2.32657596e-01 -3.88551652e-01 -7.33198702e-01
2.05074608e-01 2.49135405e-01 -7.31182173e-02 6.95669651e-01
3.92373651e-01 6.16209030e-01 -7.67198741e-01 9.12344828e-02
3.79429370e-01 7.24890172e-01 -3.19317281e-01 4.33745176e-01
2.88392544e-01 1.61338463e-01 -3.38183433e-01 7.57135391e-01
9.44404840e-01 9.72062945e-02 -3.87819856e-01 -3.17589752e-02
-4.67520654e-01 1.57968342e-01 -9.93146598e-01 1.87452257e+00
-7.27864623e-01 8.73790681e-01 -1.20666690e-01 -5.47211766e-01
1.09176958e+00 -9.96171609e-02 4.31948155e-02 -1.03918862e+00
-3.24165151e-02 2.80117065e-01 -2.84517735e-01 -2.90832132e-01
5.58907568e-01 1.00961499e-01 -1.60627022e-01 -3.37173849e-01
4.97963637e-01 6.37916803e-01 4.89414595e-02 -4.37499553e-01
1.02132094e+00 1.52671695e-01 -6.50005862e-02 -4.57187623e-01
4.48409736e-01 -1.32345200e-01 3.74818653e-01 9.51197505e-01
-4.99291629e-01 1.03441179e+00 2.78271317e-01 -7.12963760e-01
-1.11640322e+00 -1.04924917e+00 -1.61552235e-01 1.09654236e+00
2.25569651e-01 3.87049355e-02 -5.45949042e-01 -5.00646830e-01
2.99511999e-01 2.50864565e-01 -9.17337298e-01 -1.91851899e-01
-3.64842892e-01 -7.43517220e-01 1.00666904e+00 8.56931508e-01
1.18591261e+00 -1.01573360e+00 -9.17111933e-01 1.18484914e-01
1.80471405e-01 -1.19646978e+00 -1.32778361e-01 5.46617091e-01
-2.91993916e-01 -1.00191832e+00 -7.94582248e-01 -8.00010085e-01
5.65960288e-01 2.02600539e-01 1.04453719e+00 -1.40645251e-01
-2.77023733e-01 1.27245069e-01 -2.55723029e-01 -5.10556757e-01
7.11345315e-01 3.48855495e-01 -2.37464115e-01 2.74085611e-01
1.13404639e-01 -8.76828432e-01 -1.10382009e+00 2.61389226e-01
-6.12339497e-01 -6.60394192e-01 5.89411139e-01 7.23267972e-01
4.32141662e-01 -2.24487424e-01 -7.81941488e-02 -4.91115153e-01
4.33742911e-01 -6.04996502e-01 -7.08461225e-01 2.65325457e-01
-3.79177511e-01 1.38581216e-01 5.64340472e-01 1.19674273e-01
-7.71844327e-01 1.27347931e-01 -2.45950148e-01 -4.74234790e-01
-3.93110842e-01 3.49001914e-01 7.63107985e-02 -6.56036973e-01
9.36693966e-01 1.38609439e-01 -5.23505986e-01 -4.21894103e-01
1.21126488e-01 3.43763828e-01 6.69328213e-01 -5.04998863e-01
8.93161476e-01 5.27821720e-01 -3.46901652e-04 -6.39920056e-01
-6.28129959e-01 -6.37060821e-01 -6.26770079e-01 -4.07712534e-02
8.46335888e-01 -1.18105900e+00 -5.20050228e-01 7.09163785e-01
-8.96981061e-01 -2.33273700e-01 5.45681529e-02 1.62706479e-01
-3.07486385e-01 -2.22073227e-01 -1.31543100e-01 -6.79762602e-01
-3.90767872e-01 -9.68284190e-01 1.13315153e+00 7.44502127e-01
5.08264840e-01 -8.98822129e-01 1.34720117e-01 -8.32002163e-02
9.71065879e-01 7.74273992e-01 4.94003624e-01 -4.38512862e-01
-7.57750213e-01 -3.08924943e-01 -3.99937719e-01 2.51571536e-01
-1.06269881e-01 5.08271484e-03 -1.55327928e+00 -4.27610666e-01
-7.73296654e-01 -3.68472934e-01 9.17881131e-01 2.69145340e-01
1.12685061e+00 -2.21397519e-01 -2.89254159e-01 1.16191196e+00
2.02197695e+00 2.06906125e-01 1.04775095e+00 8.69413495e-01
6.41651452e-01 2.57104039e-01 4.69948977e-01 6.15669787e-01
3.96701604e-01 5.49192131e-01 7.89367855e-01 -3.46947998e-01
1.10370524e-01 -3.53885531e-01 -5.62506989e-02 -3.00929010e-01
-8.66323113e-02 -3.11640888e-01 -9.19768929e-01 1.01647174e+00
-1.91696799e+00 -7.45801389e-01 3.74986500e-01 2.22730184e+00
1.89761862e-01 1.47983387e-01 7.15294704e-02 -1.03857934e-01
5.58710098e-01 4.92635369e-01 -4.77563322e-01 -3.57975841e-01
-2.22772241e-01 3.23698670e-01 1.08256435e+00 2.86489099e-01
-1.55055594e+00 9.29065049e-01 5.99954176e+00 2.53684908e-01
-1.64663744e+00 -2.52007134e-02 7.12501287e-01 -3.32039371e-02
2.15394810e-01 -2.62767989e-02 -6.31824791e-01 3.55296284e-01
1.03732061e+00 1.20646015e-01 4.09536183e-01 1.07587981e+00
-8.90924688e-03 3.08061112e-03 -1.00243509e+00 1.17665589e+00
-1.64144978e-01 -1.34933925e+00 -1.97050050e-01 -1.29749492e-01
8.60919654e-01 6.19099438e-01 3.38994950e-01 4.08673465e-01
1.41978130e-01 -1.36886728e+00 1.15926301e+00 4.79296684e-01
5.00841260e-01 -9.77617979e-01 9.87849295e-01 -1.02745473e-01
-1.26324642e+00 -2.09724873e-01 -5.46243131e-01 -1.27309412e-01
-1.58116400e-01 2.56197363e-01 -7.92060256e-01 6.80722713e-01
1.18298292e+00 6.12083077e-01 -9.22921121e-01 1.78976083e+00
-3.18707228e-02 3.50733757e-01 -5.73278725e-01 1.48047749e-02
1.00735235e+00 4.50102597e-01 6.53991327e-02 1.48622680e+00
4.57206041e-01 -3.90095919e-01 9.63050202e-02 9.78600740e-01
-1.47636130e-01 -1.44747406e-01 -6.81748688e-01 2.88833499e-01
6.39845669e-01 1.25626671e+00 -3.44007343e-01 1.45597205e-01
4.04747501e-02 9.98570859e-01 5.22950947e-01 4.95829642e-01
-1.01279747e+00 -8.87400389e-01 1.03463447e+00 4.80498634e-02
7.25663900e-01 -2.76293337e-01 -2.92541385e-01 -8.39924335e-01
3.83736819e-01 -7.36941546e-02 -2.29864810e-02 -7.70420015e-01
-1.13763475e+00 8.28990042e-01 -1.06890105e-01 -9.50860977e-01
-3.97039112e-03 -9.28102612e-01 -8.72185826e-01 1.15449655e+00
-2.08901763e+00 -1.25751567e+00 -8.90505135e-01 8.64200711e-01
2.06107423e-01 -1.36506429e-03 6.77663386e-01 7.15417385e-01
-5.30131102e-01 1.08041418e+00 3.49964678e-01 4.52916771e-01
5.58896601e-01 -1.18522346e+00 6.26828730e-01 1.03226423e+00
-8.86177421e-02 6.68000877e-01 5.97600877e-01 9.52896252e-02
-9.66217399e-01 -1.37931752e+00 7.89934099e-01 -1.01327896e-01
1.04504362e-01 -6.16274416e-01 -3.26452762e-01 7.42642105e-01
4.42281574e-01 6.74398899e-01 3.98280472e-01 8.96152556e-02
-4.46973205e-01 -5.17497897e-01 -1.56034851e+00 4.21554059e-01
1.15615237e+00 -6.04389966e-01 -5.71736805e-02 2.79465705e-01
5.97433209e-01 -5.93207836e-01 -7.78041184e-01 5.08388996e-01
5.68790495e-01 -1.19612110e+00 1.06789792e+00 -5.69045365e-01
3.98523897e-01 -5.80062687e-01 -5.85999608e-01 -1.30414093e+00
-7.38923907e-01 -2.44989201e-01 6.78474247e-01 1.24263716e+00
6.97647452e-01 -6.59731865e-01 6.13588870e-01 1.79759562e-01
-5.00173628e-01 -7.85189092e-01 -1.10890830e+00 -6.86352909e-01
-9.72573385e-02 -9.81130078e-02 6.22979999e-01 7.15603828e-01
-9.11845803e-01 -9.95852500e-02 -3.85371268e-01 6.12280726e-01
4.26304460e-01 -2.01005921e-01 6.99764729e-01 -9.12878692e-01
6.79112747e-02 -2.32902110e-01 -1.00571001e+00 -6.50555015e-01
-4.35546897e-02 -5.41706204e-01 1.36970967e-01 -1.65716970e+00
-1.39221415e-01 -4.82952297e-01 -7.89292157e-01 6.75389469e-01
2.13022277e-01 4.13324118e-01 1.38803899e-01 7.31450599e-03
-5.44770122e-01 5.60842991e-01 6.13621414e-01 -2.97409356e-01
1.05831273e-01 -1.89541131e-01 -4.31488127e-01 2.66169578e-01
6.66343749e-01 -3.61625373e-01 -3.87838602e-01 -8.99708331e-01
-1.13370307e-01 -7.58444130e-01 7.26957381e-01 -1.62747395e+00
1.67990923e-01 9.67899486e-02 8.08723867e-01 -4.58246112e-01
2.94991225e-01 -9.32778299e-01 -1.42983030e-04 2.62291253e-01
-2.29722887e-01 4.32053298e-01 6.72608137e-01 4.94068593e-01
-4.53247577e-01 2.76440591e-01 7.91559219e-01 -1.50862053e-01
-1.13733494e+00 3.49860400e-01 1.85790300e-01 -2.28873268e-01
1.11913908e+00 -4.08610851e-01 -5.97943425e-01 -1.03948750e-01
-5.25497079e-01 3.02027047e-01 2.82608479e-01 5.03459036e-01
5.56789398e-01 -1.56509864e+00 -5.15855908e-01 3.19486946e-01
4.19713020e-01 3.28739397e-02 3.65136445e-01 6.38901770e-01
-1.24201107e+00 5.91630936e-01 -8.12050462e-01 -6.99328899e-01
-8.38914156e-01 2.57838249e-01 9.18539703e-01 -1.15183800e-01
-1.34055898e-01 1.21998167e+00 -2.56053377e-02 -7.71945298e-01
5.86577713e-01 -7.35470474e-01 1.09785460e-01 -4.34854358e-01
3.29487234e-01 1.61717042e-01 4.01553065e-01 -4.72264469e-01
-8.79544258e-01 5.73974013e-01 -1.80427115e-02 8.13177899e-02
1.45250487e+00 7.09695220e-02 2.08503306e-01 2.39225402e-01
1.40199518e+00 -2.63646096e-01 -1.75323641e+00 1.15248598e-01
-2.13238657e-01 -6.71903491e-01 -2.23556254e-02 -1.16569400e+00
-1.05047762e+00 7.11379230e-01 1.25913906e+00 -2.23756343e-01
1.10107505e+00 -4.47932363e-01 4.81520474e-01 4.93666500e-01
5.35286188e-01 -7.27339327e-01 -3.16897452e-01 6.89749062e-01
9.57359791e-01 -1.27448344e+00 -3.99713188e-01 -9.87083837e-02
-3.17163736e-01 8.09318602e-01 7.08630145e-01 -8.30442727e-01
7.81759679e-01 5.54210134e-02 1.63987607e-01 -1.25152960e-01
-4.89151180e-01 -1.90953314e-01 3.43283445e-01 6.68915153e-01
4.50542480e-01 -1.68597028e-02 1.57329395e-01 1.28720552e-01
-3.15894067e-01 -2.05571473e-01 3.28292698e-02 1.12807083e+00
-3.14006418e-01 -5.75828016e-01 -2.79931694e-01 -6.25000000e-02
-4.61007386e-01 -3.33347879e-02 -4.23227847e-01 9.58094656e-01
5.75903416e-01 7.40093708e-01 1.11724332e-01 -5.78668296e-01
8.95246744e-01 -2.07960352e-01 4.36649501e-01 -1.53617918e-01
-9.81768429e-01 -5.07188678e-01 1.17750145e-01 -7.84450769e-01
-3.36803257e-01 -3.69044363e-01 -9.36839700e-01 -3.20737869e-01
1.37947127e-01 -2.00394303e-01 9.44089949e-01 7.78394938e-01
6.95621610e-01 6.54239595e-01 7.13215828e-01 -1.09376609e+00
-5.29289246e-01 -1.12348926e+00 -4.02621806e-01 3.14343184e-01
8.42984796e-01 -8.71563077e-01 -1.90535665e-01 -3.23402077e-01] | [7.776882648468018, -1.845916509628296] |
aab4ac39-583c-4e7b-97db-0b7b8f618c74 | challenges-and-opportunities-for-computer | 2004.06180 | null | https://arxiv.org/abs/2004.06180v1 | https://arxiv.org/pdf/2004.06180v1.pdf | Challenges and Opportunities for Computer Vision in Real-life Soccer Analytics | In this paper, we explore some of the applications of computer vision to sports analytics. Sport analytics deals with understanding and discovering patterns from a corpus of sports data. Analysing such data provides important performance metrics for the players, for instance in soccer matches, that could be useful for estimating their fitness and strengths. Team level statistics can also be estimated from such analysis. This paper mainly focuses on some the challenges and opportunities presented by sport video analysis in computer vision. Specifically, we use our multi-camera setup as a framework to discuss some of the real-life challenges for machine learning algorithms. | ['Neha Bhargava', 'Fabio Cuzzolin'] | 2020-04-13 | null | null | null | null | ['sports-analytics'] | ['computer-vision'] | [-1.45942941e-02 -2.31486574e-01 -2.14857221e-01 -1.81732461e-01
-4.58627045e-01 -5.28402686e-01 3.88281971e-01 6.21145606e-01
-7.04151452e-01 2.61379898e-01 2.52205908e-01 3.85988802e-01
-3.72755736e-01 -5.46304286e-01 -5.32192409e-01 -6.98608637e-01
-2.63447583e-01 4.32358801e-01 4.57445174e-01 -6.72629237e-01
4.72813606e-01 5.04738629e-01 -2.31893277e+00 6.12105966e-01
-1.74405977e-01 8.70901108e-01 3.38446110e-01 1.01273799e+00
1.75753027e-01 8.72110367e-01 -5.69698393e-01 -8.30945075e-01
3.11389029e-01 -2.22394168e-01 -4.90563035e-01 1.38655573e-01
4.56475854e-01 7.81184286e-02 -3.12367171e-01 8.44848633e-01
3.27977687e-01 1.15350835e-01 2.13334382e-01 -1.14790809e+00
5.01699150e-01 5.21914959e-01 -4.42925006e-01 1.02478015e+00
7.42393315e-01 3.91990006e-01 8.02799821e-01 -5.58108509e-01
6.44870281e-01 9.24520791e-01 6.96251988e-01 9.37569290e-02
-7.75903404e-01 -3.00443381e-01 -2.77446330e-01 9.80942726e-01
-1.06069160e+00 -3.87027144e-01 8.35324943e-01 -6.55453861e-01
3.29230547e-01 5.38787007e-01 1.34245205e+00 9.24521565e-01
1.55337960e-01 1.00415790e+00 1.20508921e+00 -6.14416122e-01
3.65196504e-02 -1.73404723e-01 6.05907023e-01 2.97954142e-01
2.64094234e-01 2.19208553e-01 -1.13840330e+00 2.83193409e-01
6.19778931e-01 -1.14273280e-01 1.57069713e-01 -1.53729945e-01
-1.30614901e+00 9.01502490e-01 -2.65484620e-02 1.66362897e-01
-4.12538558e-01 1.35957956e-01 8.83241713e-01 2.28434488e-01
1.84493199e-01 4.42560911e-01 3.50653455e-02 -9.65825200e-01
-8.57050717e-01 6.34166062e-01 4.67269033e-01 4.32125568e-01
5.39547741e-01 -1.09442428e-01 -3.06176743e-03 6.68481767e-01
-8.53099823e-02 4.14585732e-02 3.27279538e-01 -1.17408359e+00
3.26563269e-01 3.74320209e-01 -2.98988253e-01 -1.18568361e+00
-5.79473913e-01 -3.63266170e-01 -2.13794738e-01 2.90918529e-01
7.77680576e-01 7.08973184e-02 -2.56815791e-01 8.05327058e-01
4.26536620e-01 3.41676772e-01 -9.29824114e-02 1.33242106e+00
1.19014788e+00 1.81668386e-01 -2.52557606e-01 -1.84679180e-01
1.59401906e+00 -8.06577027e-01 -6.24677479e-01 -4.11087304e-01
3.81322145e-01 -8.25538337e-01 8.52206290e-01 9.36565518e-01
-1.06153810e+00 -8.07132959e-01 -6.65985346e-01 2.11832091e-01
-1.42602712e-01 2.47540642e-02 7.13645160e-01 6.32882893e-01
-4.15996820e-01 5.91945589e-01 -8.71041238e-01 -4.70597208e-01
-4.20736673e-04 6.63411468e-02 -4.17156637e-01 -8.99178631e-05
-9.45384920e-01 8.20786893e-01 5.96466184e-01 2.99097121e-01
-6.94308639e-01 -5.16474128e-01 -9.41958010e-01 -2.34100625e-01
6.58196628e-01 -4.33927123e-03 1.16163027e+00 -7.37392604e-01
-1.07194269e+00 1.47537541e+00 2.79095262e-01 -6.02898479e-01
4.07869041e-01 -3.95978481e-01 -2.42623881e-01 2.42064625e-01
2.65093505e-01 4.38537002e-02 3.82636070e-01 -6.96473837e-01
-1.04175758e+00 -5.41283727e-01 1.36531055e-01 1.57464266e-01
2.28114203e-02 5.02377748e-01 -7.48783767e-01 -4.82021779e-01
-1.46874815e-01 -8.79252970e-01 -3.06424052e-01 -6.05497956e-01
-7.14305714e-02 -1.52574077e-01 3.55863243e-01 -6.10435724e-01
1.17863476e+00 -2.03504062e+00 3.69451046e-01 1.78568065e-01
2.55505949e-01 2.07341477e-01 3.36062849e-01 7.24682391e-01
9.90656987e-02 -6.22386217e-01 5.41933537e-01 2.49035582e-01
-3.79704893e-01 3.62578005e-01 1.05520263e-01 5.07565916e-01
-2.06320196e-01 6.66048408e-01 -8.80512297e-01 -7.03868091e-01
5.74340045e-01 -1.07252754e-01 -2.50626743e-01 -3.51528451e-03
2.58080870e-01 5.32824695e-01 -2.92249650e-01 8.25972319e-01
5.75229339e-02 5.83706915e-01 1.01394095e-01 -1.65972233e-01
-5.58886707e-01 -2.91958570e-01 -1.49775994e+00 1.70643592e+00
3.67955603e-02 1.12512314e+00 1.21941216e-01 -1.78469527e+00
9.86607552e-01 -1.58458903e-01 7.29605973e-01 -6.34374797e-01
1.52632892e-01 -1.09009609e-01 2.04197794e-01 -1.08901596e+00
7.39798963e-01 -2.17346996e-01 -3.90676796e-01 1.41528726e-01
8.52562711e-02 1.58012837e-01 8.58666837e-01 -1.52909860e-01
1.00317025e+00 1.25550568e-01 2.63113379e-01 -8.38222057e-02
4.88600791e-01 6.34573936e-01 6.10449553e-01 8.48496795e-01
-3.37606788e-01 6.10286593e-01 5.63650310e-01 -1.03452682e+00
-1.09445369e+00 -8.38553011e-01 1.07062861e-01 1.15319479e+00
1.62639946e-01 -6.84455812e-01 -6.80921674e-01 6.63894368e-03
-2.33841352e-02 -1.02140293e-01 -6.14281833e-01 -1.14842564e-01
-7.10382819e-01 -5.78035951e-01 4.71117795e-01 6.52707219e-01
6.92735463e-02 -1.17136621e+00 -1.11875820e+00 3.42669874e-01
-3.62736881e-01 -1.38533521e+00 4.04885039e-02 -1.41591907e-01
-7.78326392e-01 -1.51226497e+00 -4.57963049e-01 -4.34553444e-01
-2.31521651e-01 4.46120262e-01 1.41608858e+00 7.94428587e-02
-8.79415452e-01 8.07916939e-01 -7.82654941e-01 -9.43280697e-01
-1.97748616e-01 -1.48653299e-01 2.67575204e-01 2.79829502e-01
8.70408654e-01 -2.30493948e-01 -4.70355004e-01 4.44280475e-01
-5.09253502e-01 -1.61055714e-01 3.83842468e-01 4.97611344e-01
8.77784252e-01 -6.01851344e-02 -5.27175926e-02 -3.92614216e-01
6.67168260e-01 -4.29886431e-01 -5.00777721e-01 3.77466418e-02
-8.61334205e-02 -4.73885536e-01 7.54855946e-02 -4.31238890e-01
-4.46006656e-01 1.94871947e-01 6.38419017e-02 -4.86309439e-01
-4.29317862e-01 6.68454170e-01 3.36583406e-01 1.22158127e-02
8.81584764e-01 1.53701320e-01 3.34880143e-01 -6.55849516e-01
-9.29502100e-02 6.20331705e-01 9.03731763e-01 -6.52048826e-01
3.27577412e-01 6.94921136e-01 4.37594175e-01 -1.38959861e+00
-1.08290219e+00 -1.17826915e+00 -9.39328849e-01 -1.35663486e+00
9.13462460e-01 -8.94539177e-01 -1.34082389e+00 3.65890890e-01
-4.85045165e-01 -1.91934966e-03 -6.50568902e-01 9.15001750e-01
-8.18455696e-01 4.06182289e-01 -5.38405418e-01 -9.64103222e-01
9.41371247e-02 -9.02596533e-01 1.04540920e+00 5.25349855e-01
-2.89569318e-01 -7.98891068e-01 4.44911778e-01 1.27156067e+00
-2.50960648e-01 7.69898474e-01 -1.95667908e-01 -5.15799999e-01
-1.03377797e-01 -4.81863528e-01 3.98748636e-01 2.69462913e-01
-4.58858967e-01 -4.33097482e-02 -7.71508694e-01 1.86617188e-02
-2.84563661e-01 -3.45680743e-01 8.72635365e-01 7.33316720e-01
1.11995447e+00 5.62312067e-01 -6.77428097e-02 4.87125933e-01
1.07110929e+00 -2.79862761e-01 5.61274886e-01 9.66159225e-01
7.08047688e-01 1.16231608e+00 1.35670567e+00 6.21373117e-01
2.67062247e-01 1.02555442e+00 5.39320052e-01 5.56123070e-02
-1.95202976e-02 1.60535857e-01 3.53194535e-01 6.57681584e-01
-1.13723207e+00 5.54576695e-01 -1.12535572e+00 4.67169762e-01
-2.14382219e+00 -1.55006015e+00 -1.08810580e+00 2.01446462e+00
5.76848574e-02 2.63872385e-01 1.04539573e+00 6.63119018e-01
8.29935133e-01 -2.23666597e-02 -1.58532619e-01 -8.15610409e-01
-1.10508241e-01 3.42982560e-01 6.27322912e-01 -1.11359291e-01
-1.25987196e+00 6.51952446e-01 6.82139540e+00 8.62739265e-01
-7.00764000e-01 1.14088394e-01 1.30251646e-01 -6.30201578e-01
6.47093534e-01 -5.10346629e-02 -7.61615634e-01 3.36007506e-01
9.29562986e-01 -2.77236030e-02 5.08179218e-02 8.92948389e-01
5.59669137e-01 -6.01079583e-01 -8.30604792e-01 1.27123713e+00
2.75163740e-01 -1.54135251e+00 -5.81094444e-01 4.04110879e-01
1.54439047e-01 -9.66653880e-03 -3.89232546e-01 1.77678630e-01
-2.42372975e-01 -7.08506107e-01 9.93331373e-01 8.16897154e-01
3.31689149e-01 -1.08899271e+00 8.85034561e-01 5.11871815e-01
-1.11316741e+00 -4.41119134e-01 -5.11911750e-01 -9.43629980e-01
1.63489103e-01 5.67559898e-01 -2.81080782e-01 5.22541583e-01
1.34979177e+00 9.49161947e-01 -4.70734090e-01 1.41417265e+00
1.08506866e-01 5.79593301e-01 -4.03291643e-01 -1.16529271e-01
2.88490891e-01 -5.23362756e-01 7.91785240e-01 1.29035997e+00
2.12251246e-02 1.03297420e-01 3.98535162e-01 7.99925998e-02
6.54762506e-01 2.22167328e-01 -7.49365866e-01 1.06119983e-01
-1.77841917e-01 1.48931777e+00 -9.56721842e-01 -1.65632606e-01
-5.98263144e-01 3.76223624e-01 1.36256769e-01 -2.69526899e-01
-4.56370890e-01 -1.09393433e-01 1.05785334e+00 6.11511946e-01
1.83637038e-01 -3.63832116e-01 -1.61165670e-01 -1.24134314e+00
5.49375415e-02 -9.40396488e-01 7.96869457e-01 -6.14319861e-01
-8.99256885e-01 -7.83695951e-02 3.13407481e-01 -1.59400749e+00
-2.43859470e-01 -7.40408301e-01 -8.49311173e-01 3.06508809e-01
-1.00570929e+00 -1.00144279e+00 -4.70868856e-01 5.45080483e-01
1.02570462e+00 -5.29921830e-01 2.69133300e-01 1.27582908e-01
-5.89463294e-01 4.72302362e-02 -7.46825114e-02 3.34687114e-01
4.32728171e-01 -1.22188747e+00 2.54550986e-02 8.19341838e-01
6.02397263e-01 -2.73384392e-01 1.22296965e+00 -4.40372944e-01
-1.64425468e+00 -2.66627401e-01 4.07524765e-01 -7.78907299e-01
7.86151826e-01 -1.13061123e-01 -4.85653430e-01 4.25162375e-01
-2.04748780e-01 -1.29053637e-01 1.12001383e+00 5.09614825e-01
3.94403279e-01 -3.43537211e-01 -4.91263956e-01 2.51005918e-01
8.88392091e-01 -2.75466591e-02 -8.29538226e-01 2.61236668e-01
-3.44570190e-01 -6.80286646e-01 -1.02841520e+00 9.25965700e-03
9.18367028e-01 -1.39719141e+00 1.21647263e+00 -8.76465738e-01
6.26132369e-01 -2.68517062e-02 1.02603927e-01 -9.47437763e-01
-2.12533653e-01 -2.64167130e-01 3.67367305e-02 9.14814055e-01
-1.39107674e-01 2.39919752e-01 1.19408357e+00 3.53615582e-01
-2.35508442e-01 -3.17185551e-01 -1.03360760e+00 -8.36898029e-01
-3.37601483e-01 -1.09505212e+00 4.80396263e-02 7.93366373e-01
1.68837667e-01 -3.80828045e-02 -6.34005904e-01 -1.44156650e-01
1.11226606e+00 1.68642476e-01 1.30280483e+00 -1.61579335e+00
-3.85543048e-01 -4.47264403e-01 -1.40169597e+00 -2.85544574e-01
-2.45208964e-01 -5.21107316e-01 -3.33414048e-01 -1.17284870e+00
2.83182353e-01 1.81988060e-01 -9.39301699e-02 -1.16759196e-01
9.09429938e-02 7.64871299e-01 5.37900627e-01 2.81610429e-01
-6.38317347e-01 -3.15574378e-01 1.09506357e+00 1.10108137e-01
4.78289276e-02 3.47500563e-01 -3.75847429e-01 9.20222700e-01
7.99884140e-01 -3.98105532e-01 1.49735019e-01 -1.43158332e-01
5.66251874e-01 1.30327106e-01 5.19849241e-01 -1.25438416e+00
2.43276641e-01 -4.76295710e-01 1.18932731e-01 -4.91158545e-01
5.44284225e-01 -4.32663918e-01 1.48313537e-01 6.60912752e-01
-1.36146739e-01 8.08466896e-02 -8.59644040e-02 7.33546972e-01
-7.80844331e-01 -5.10617375e-01 3.09021324e-01 -4.96559680e-01
-1.25744987e+00 -1.60897225e-01 -6.80618644e-01 4.76343520e-02
1.48303604e+00 -8.90876591e-01 6.48246780e-02 -2.96029478e-01
-1.24278736e+00 5.26175320e-01 1.67568207e-01 5.11511147e-01
6.24003828e-01 -1.33575821e+00 -1.06376135e+00 -1.33041153e-02
4.55226868e-01 -4.00813401e-01 4.64967608e-01 1.08717203e+00
-1.09604418e+00 1.39141381e-01 -7.97750115e-01 -9.62770164e-01
-1.94123137e+00 5.76881707e-01 1.99525300e-02 -1.61153644e-01
-8.11835289e-01 4.32703495e-01 -4.32478756e-01 2.31966361e-01
2.59636007e-02 1.15968414e-01 -1.02097845e+00 5.85298300e-01
7.56907225e-01 8.61779213e-01 1.30309328e-01 -9.35116470e-01
-4.19301391e-01 8.51995587e-01 3.02366942e-01 -1.24333046e-01
1.64330697e+00 -9.07298997e-02 1.91130117e-01 9.92072225e-01
3.73150319e-01 -2.12678343e-01 -1.06031001e+00 -7.97645301e-02
2.87326485e-01 -7.28395224e-01 1.65358454e-01 -3.43374461e-02
-1.15766633e+00 1.01874352e+00 6.30148172e-01 5.37089169e-01
1.09855974e+00 3.79332274e-01 5.96360207e-01 1.96977198e-01
4.22289610e-01 -1.79839778e+00 1.11752197e-01 1.51293993e-01
6.79208517e-01 -1.51790571e+00 5.69890402e-02 -2.83167124e-01
-9.98572230e-01 1.69214487e+00 2.97861308e-01 -3.84627134e-01
5.44285119e-01 2.27303579e-01 2.30799526e-01 -7.20251203e-01
-4.75301027e-01 -8.52459550e-01 3.38142484e-01 8.18553388e-01
1.77538410e-01 1.58216044e-01 -7.64269173e-01 5.12877882e-01
-4.53308046e-01 -3.89648713e-02 1.04454076e+00 1.08881342e+00
-6.88305378e-01 -1.11725080e+00 -9.52053130e-01 4.85083550e-01
-7.33576298e-01 5.36934733e-01 -3.07532281e-01 8.03923249e-01
3.35963756e-01 9.69225347e-01 2.51542956e-01 -6.28685772e-01
7.99108148e-01 -1.39399275e-01 5.87535679e-01 -5.48140585e-01
-9.06245530e-01 3.74131203e-02 4.61825997e-01 -7.65011311e-01
-9.52910662e-01 -1.25012159e+00 -4.89479899e-01 -3.50203425e-01
-1.89573709e-02 1.04718402e-01 9.74567175e-01 1.05380452e+00
-2.60192215e-01 3.86396289e-01 4.61814791e-01 -1.11805022e+00
1.73825622e-01 -6.87734246e-01 -1.01124883e+00 5.64548135e-01
-7.82715529e-02 -9.19050932e-01 2.58677173e-02 3.62858236e-01] | [7.405196189880371, 0.22444649040699005] |
789a04f6-759d-4fbc-9516-1e04ce024a9d | machine-learning-for-synthetic-data | 2302.04062 | null | https://arxiv.org/abs/2302.04062v4 | https://arxiv.org/pdf/2302.04062v4.pdf | Machine Learning for Synthetic Data Generation: A Review | Machine learning heavily relies on data, but real-world applications often encounter various data-related issues. These include data of poor quality, insufficient data points leading to under-fitting of machine learning models, and difficulties in data access due to concerns surrounding privacy, safety, and regulations. In light of these challenges, the concept of synthetic data generation emerges as a promising alternative that allows for data sharing and utilization in ways that real-world data cannot facilitate. This paper presents a comprehensive systematic review of existing studies that employ machine learning models for the purpose of generating synthetic data. The review encompasses various perspectives, starting with the applications of synthetic data generation, spanning computer vision, speech, natural language processing, healthcare, and business domains. Additionally, it explores different machine learning methods, with particular emphasis on neural network architectures and deep generative models. The paper also addresses the crucial aspects of privacy and fairness concerns related to synthetic data generation. Furthermore, this study identifies the challenges and opportunities prevalent in this emerging field, shedding light on the potential avenues for future research. By delving into the intricacies of synthetic data generation, this paper aims to contribute to the advancement of knowledge and inspire further exploration in synthetic data generation. | ['Wenqi Wei', 'Huazheng Wang', 'Minjie Shen', 'Yingzhou Lu'] | 2023-02-08 | null | null | null | null | ['synthetic-data-generation', 'synthetic-data-generation'] | ['medical', 'miscellaneous'] | [ 3.96234870e-01 4.72610533e-01 -4.75936949e-01 -4.68049645e-01
-7.90663719e-01 -3.97264391e-01 5.98800838e-01 4.78959709e-01
-4.97884542e-01 9.44528818e-01 2.54969478e-01 -3.09002817e-01
-2.04778202e-02 -1.02547622e+00 -4.38073605e-01 -5.58927834e-01
1.99189126e-01 2.96342194e-01 -8.28148365e-01 -9.58538651e-02
2.35468641e-01 4.02795047e-01 -1.63601923e+00 4.12672818e-01
1.03482282e+00 8.96439552e-01 -3.58129501e-01 8.34424868e-02
-1.67856395e-01 5.51241279e-01 -1.05588436e+00 -9.52909231e-01
4.55388784e-01 -2.94195414e-01 -5.74545741e-01 2.00760558e-01
1.37261629e-01 -5.20543456e-01 -1.13382146e-01 7.68426180e-01
6.21318460e-01 -1.82584487e-02 4.78417993e-01 -1.83335519e+00
-1.11491454e+00 3.20801020e-01 -1.27439529e-01 -1.88208610e-01
2.89742380e-01 2.72773147e-01 7.98007011e-01 -8.29866111e-01
5.27745724e-01 1.01748896e+00 4.99364972e-01 9.51807439e-01
-1.26027358e+00 -9.06263411e-01 -6.65280744e-02 -5.30515723e-02
-1.32654977e+00 -5.83363950e-01 6.67460144e-01 -4.19636160e-01
4.10327524e-01 3.86867434e-01 6.97672963e-01 1.44405043e+00
3.19272317e-02 9.02531505e-01 1.22000086e+00 -3.62022877e-01
3.07724983e-01 7.14888096e-01 -2.66619343e-02 1.37515008e-01
6.72568917e-01 3.37687343e-01 -6.70287848e-01 -4.28120315e-01
3.19463283e-01 1.04865342e-01 -9.07705575e-02 -5.21993399e-01
-1.02537715e+00 1.28012204e+00 1.24203011e-01 3.69562656e-02
-3.97873312e-01 -5.37645757e-01 4.42080826e-01 1.89956293e-01
5.41278660e-01 5.10481894e-01 -3.02218348e-01 1.26645386e-01
-1.03157282e+00 4.57842141e-01 8.05061400e-01 1.22603440e+00
4.76493984e-01 2.81499684e-01 -1.33832723e-01 7.78722644e-01
4.00691211e-01 5.41669071e-01 4.13445473e-01 -7.69906700e-01
5.13074279e-01 4.70111758e-01 8.31854194e-02 -1.12727857e+00
-1.69004634e-01 -9.25008878e-02 -1.09271693e+00 1.34514838e-01
2.40837246e-01 -3.00876945e-01 -6.56622112e-01 1.61502981e+00
3.39627385e-01 -4.67091024e-01 3.57170135e-01 6.52968824e-01
1.14659500e+00 4.12751794e-01 2.55542904e-01 -3.06687832e-01
1.26475883e+00 -4.62122619e-01 -1.00177646e+00 3.14082876e-02
3.03366214e-01 -5.50519168e-01 8.82804394e-01 2.98229992e-01
-1.07974899e+00 -3.67811590e-01 -7.63801277e-01 -3.19866359e-01
-7.86374927e-01 -1.51083782e-01 7.25884020e-01 1.16834712e+00
-8.23598325e-01 1.84423223e-01 -5.70266068e-01 -3.88236165e-01
1.03038931e+00 1.61979735e-01 -4.11848366e-01 4.80791507e-03
-1.35654259e+00 7.08712697e-01 3.17523330e-01 2.09259596e-02
-5.41872025e-01 -8.43534231e-01 -1.01775265e+00 -2.17725784e-01
3.93997043e-01 -8.19139600e-01 9.95645940e-01 -6.46402478e-01
-1.03075218e+00 9.35757995e-01 -3.86296096e-03 -7.25931525e-01
6.72099233e-01 -5.07612415e-02 -6.24810874e-01 -3.31011951e-01
2.01030359e-01 6.53365910e-01 6.14611983e-01 -1.23385835e+00
-6.60864353e-01 -5.28507173e-01 -3.30188245e-01 2.51660086e-02
-2.87828833e-01 -7.97455609e-02 -1.72462352e-02 -7.85684228e-01
-5.08743703e-01 -7.56704032e-01 -3.89002889e-01 7.07874596e-02
-5.80894768e-01 3.40337493e-02 7.97891438e-01 -3.91908497e-01
1.36801887e+00 -2.16712570e+00 -4.70182419e-01 1.37676373e-01
3.09886307e-01 5.68830550e-01 1.65151302e-02 7.75843680e-01
-7.05360770e-02 5.60301483e-01 -2.16536790e-01 -2.62123585e-01
-3.03136706e-02 1.28230914e-01 -4.81349468e-01 2.46659428e-01
3.65451694e-01 1.01926410e+00 -7.36952066e-01 -1.48147434e-01
3.16819221e-01 4.86786962e-01 -5.49858153e-01 1.77274778e-01
-5.55497687e-03 4.56032395e-01 -4.00504768e-01 7.64715850e-01
8.37610066e-01 -4.36074547e-02 -3.85128733e-05 1.11960303e-02
-1.18879020e-01 2.56922871e-01 -8.37612867e-01 1.17757070e+00
-2.19657958e-01 4.40716863e-01 -1.55448718e-02 -8.69905353e-01
1.03208983e+00 3.08392078e-01 6.03070676e-01 -7.32633293e-01
4.87620980e-02 3.02813202e-01 -2.28312880e-01 -6.83315933e-01
7.22246945e-01 -1.96770310e-01 -1.08422168e-01 5.29390931e-01
-4.12363529e-01 -2.80623704e-01 -4.90481965e-02 9.19745639e-02
6.06830060e-01 -2.52907515e-01 5.20890653e-01 2.22912341e-01
2.04454958e-01 3.52727771e-02 4.24800932e-01 5.99262714e-01
-2.37456679e-01 5.45064151e-01 4.00514066e-01 -5.97818196e-01
-1.06427169e+00 -9.66097474e-01 -1.25302553e-01 6.88475311e-01
-3.63925606e-01 -3.97868603e-01 -7.88508296e-01 -8.16615641e-01
3.39513481e-01 1.11152947e+00 -7.13244617e-01 -3.70345324e-01
1.79390423e-02 -1.08887231e+00 6.02780938e-01 1.92187846e-01
3.63964498e-01 -1.13974929e+00 -7.54401743e-01 2.51734024e-03
-4.20920163e-01 -9.82058585e-01 -1.98568195e-01 -2.77845055e-01
-8.77001226e-01 -1.13745618e+00 -4.26003844e-01 -2.58543134e-01
8.27256441e-01 3.29736024e-01 1.01859605e+00 -1.35454312e-01
-3.88855487e-01 2.90414989e-01 -3.94401759e-01 -1.08719206e+00
-6.66526556e-01 1.10677652e-01 -5.52185886e-02 -1.10791080e-01
9.57030773e-01 -7.40490183e-02 -5.07357419e-01 7.06031621e-02
-1.29589939e+00 -7.03799427e-02 4.64422613e-01 1.10839939e+00
5.03163815e-01 4.77497578e-02 1.00690246e+00 -1.36734796e+00
1.14563584e+00 -8.82743776e-01 -3.77849102e-01 1.12536028e-01
-1.22589839e+00 -3.08060586e-01 5.03622770e-01 5.61652565e-03
-1.00366604e+00 -2.73877889e-01 -1.04439266e-01 -1.92872360e-01
-2.79439747e-01 3.55406791e-01 -2.65638888e-01 2.83028573e-01
8.26327085e-01 1.94482341e-01 6.72068417e-01 -3.73473674e-01
5.57225168e-01 1.02231598e+00 1.10920124e-01 -3.66375864e-01
5.07842243e-01 5.64812779e-01 -2.93794394e-01 -7.90931463e-01
-6.68975830e-01 4.51626116e-03 -3.97638500e-01 7.53354132e-02
6.90548420e-01 -6.71888113e-01 -5.91760874e-01 5.65367043e-01
-7.91375160e-01 1.22015022e-01 -7.71245480e-01 3.39823812e-01
-5.31634569e-01 1.99671879e-01 -1.68734103e-01 -1.01255953e+00
-5.77366948e-01 -1.08519936e+00 7.88260758e-01 1.80175751e-01
-5.06786644e-01 -8.32381010e-01 -2.78047025e-01 9.04472411e-01
4.91787106e-01 6.11216784e-01 8.50369334e-01 -8.75783741e-01
-5.44501662e-01 -5.22215843e-01 -6.76470995e-02 4.51540232e-01
5.04682243e-01 1.71858445e-01 -1.05897760e+00 -3.14002275e-01
9.23455507e-02 -4.90712672e-01 3.48869592e-01 3.09821695e-01
1.38759947e+00 -7.67832994e-01 -1.12240724e-01 4.41844165e-01
1.14490414e+00 4.26614493e-01 5.44333220e-01 1.50923505e-01
3.79080564e-01 1.18190575e+00 8.29965889e-01 9.03477609e-01
7.31029093e-01 2.79180735e-01 4.04426992e-01 -3.72780025e-01
1.81382626e-01 -4.45792317e-01 -1.56090662e-01 2.60354340e-01
1.77439719e-01 -3.46982807e-01 -7.21503079e-01 5.99290431e-01
-1.58828950e+00 -1.10220778e+00 -2.15269342e-01 2.59727001e+00
7.48048127e-01 -3.43092114e-01 2.71828711e-01 6.25338033e-02
5.57810366e-01 1.12551041e-01 -6.88785374e-01 -7.37661541e-01
-1.77906677e-01 1.67706043e-01 5.35665214e-01 5.98438038e-03
-9.82168913e-01 6.24065042e-01 7.40986109e+00 4.71799105e-01
-1.05659497e+00 -1.43558651e-01 1.02976048e+00 -3.17500770e-01
-6.29642010e-01 -3.79657120e-01 -5.94835758e-01 6.80887520e-01
9.79448557e-01 -6.04954004e-01 3.59674633e-01 8.36054564e-01
4.63961989e-01 1.78234521e-02 -1.10835624e+00 8.85527968e-01
1.98478073e-01 -1.48693800e+00 3.93445015e-01 4.95632976e-01
6.33248985e-01 1.09663066e-02 6.06105268e-01 7.93955475e-02
2.11851090e-01 -1.31051171e+00 5.53254306e-01 4.72375065e-01
9.65468168e-01 -8.38451028e-01 7.90705264e-01 3.61509860e-01
-3.40028316e-01 -6.21056296e-02 -1.52002424e-01 7.20359311e-02
1.28244296e-01 6.26876950e-01 -1.04731596e+00 6.98986590e-01
6.17295802e-01 4.42032993e-01 -3.04703414e-01 7.41767943e-01
2.16371343e-01 2.01339379e-01 2.81551369e-02 2.46413015e-02
-3.75417583e-02 -2.98447788e-01 2.25986063e-01 9.77434456e-01
4.07053798e-01 6.01242557e-02 -1.85146406e-01 1.04621029e+00
-1.76213697e-01 2.43234158e-01 -1.10607433e+00 -5.17273188e-01
7.34408796e-01 9.28333282e-01 -9.37143564e-02 -6.27541766e-02
-7.92375684e-01 4.95946765e-01 9.69107524e-02 3.37119132e-01
-4.01693404e-01 -1.14443578e-01 7.69301116e-01 3.47345501e-01
-3.77619117e-01 6.51052743e-02 -7.27727532e-01 -9.69040096e-01
-2.05518436e-02 -1.40031314e+00 7.13550031e-01 -4.68985617e-01
-1.47363925e+00 5.14457643e-01 1.29913598e-01 -1.35810804e+00
-5.71548283e-01 -2.93243527e-01 -1.27743527e-01 1.05210829e+00
-1.26270545e+00 -1.08926928e+00 -2.25084469e-01 3.87665212e-01
3.41814816e-01 -4.96870995e-01 1.19524527e+00 2.84034550e-01
-5.46158254e-01 8.56603920e-01 2.09325314e-01 9.30058807e-02
7.41163135e-01 -7.81535029e-01 6.22484684e-01 5.44160187e-01
-6.22461215e-02 9.10666406e-01 4.16567832e-01 -6.29211068e-01
-1.14430296e+00 -1.27738881e+00 1.05830693e+00 -6.06594622e-01
1.93023711e-01 -4.44458306e-01 -7.93074250e-01 6.17512882e-01
2.40322694e-01 -2.73204774e-01 1.64013374e+00 -8.47891793e-02
-1.70190796e-01 -2.13643014e-01 -1.82753253e+00 6.71095133e-01
5.54017901e-01 -4.04286534e-01 -1.88955277e-01 1.09467492e-01
4.39662457e-01 -3.01848024e-01 -9.58995581e-01 2.12137580e-01
6.55811429e-01 -1.05202520e+00 8.15068364e-01 -9.22501445e-01
4.85537112e-01 2.04195350e-01 1.33720029e-03 -1.27893293e+00
-1.34494886e-01 -7.46118844e-01 -5.46906367e-02 1.32528794e+00
5.31827807e-01 -8.28652322e-01 1.04407024e+00 1.54570353e+00
3.00362527e-01 -8.45011294e-01 -7.33556151e-01 -4.38221544e-01
1.16967946e-01 -4.68108535e-01 1.11799145e+00 1.23064303e+00
-9.97330248e-02 2.09136717e-02 -6.89916253e-01 -3.68267328e-01
7.71690249e-01 -1.37463704e-01 1.13931537e+00 -1.13150787e+00
3.98743600e-01 -2.86926448e-01 -2.42870003e-01 -4.93148118e-01
-7.45774657e-02 -8.03237319e-01 -5.39215505e-01 -1.50632894e+00
5.41167445e-02 -5.29079616e-01 -1.29155712e-02 3.41183901e-01
-1.17189549e-02 2.45253637e-01 3.20096493e-01 -5.15646264e-02
3.00690401e-02 5.64138353e-01 1.33131182e+00 1.01780985e-03
-1.41913623e-01 3.88662905e-01 -1.54340625e+00 3.87901276e-01
1.01568365e+00 -3.06335509e-01 -7.69708991e-01 -3.27326208e-01
9.74064693e-02 -7.50865936e-02 2.44042799e-01 -2.87589312e-01
-2.06449237e-02 -5.10828674e-01 3.58937472e-01 -6.58652902e-01
2.77543277e-01 -1.12465918e+00 4.50233310e-01 3.38856488e-01
-4.89574820e-01 4.96008247e-02 1.60596684e-01 5.01588106e-01
-2.98553288e-01 -1.62530676e-01 7.23104775e-01 -2.69993871e-01
-3.39236230e-01 5.47738075e-01 -4.08132136e-01 1.31194621e-01
1.40602922e+00 -3.38017285e-01 -3.57478827e-01 -6.02509379e-01
-5.79010069e-01 3.81168664e-01 4.35105771e-01 8.04646373e-01
7.68301427e-01 -1.45242000e+00 -8.00962865e-01 7.32965946e-01
4.92568165e-01 1.94533169e-01 3.12848240e-01 4.23303396e-01
-2.23292246e-01 4.29908246e-01 -2.78491646e-01 -9.86609012e-02
-1.21992803e+00 8.70487332e-01 -2.31069829e-02 1.27615556e-01
-2.97871798e-01 5.23091495e-01 1.29085451e-01 -5.02649724e-01
2.73522824e-01 -9.21446458e-02 -8.91440883e-02 2.44649932e-01
5.45975387e-01 6.02019072e-01 5.46432659e-02 -7.16022253e-01
-2.53848612e-01 -2.41865680e-01 -2.98312753e-01 3.95258442e-02
1.01685154e+00 -3.00352931e-01 -1.08167551e-01 3.33492607e-01
1.01349652e+00 3.47054540e-03 -8.16375494e-01 -2.16198027e-01
-2.13851243e-01 -8.53015661e-01 -3.94598126e-01 -9.39170480e-01
-1.09337485e+00 8.14304709e-01 3.78306955e-01 4.52452302e-01
1.30408537e+00 -2.56326944e-01 7.27670372e-01 -5.52748628e-02
4.50962275e-01 -1.35076749e+00 -2.03068227e-01 5.80025278e-02
1.01088476e+00 -1.57963300e+00 -5.87842651e-02 -3.39413464e-01
-1.05217767e+00 8.48828971e-01 5.05356669e-01 4.45141077e-01
7.30701327e-01 1.84744731e-01 2.91086644e-01 -5.50283454e-02
-6.32340670e-01 1.90174252e-01 1.09054372e-01 1.13784850e+00
4.23297256e-01 2.21309125e-01 -6.30242288e-01 5.93612254e-01
-5.56193709e-01 2.51798898e-01 6.29812062e-01 8.63333464e-01
1.23402588e-01 -1.57197618e+00 -5.07042408e-01 1.03879774e+00
-6.79633260e-01 -8.22719485e-02 -6.52481437e-01 7.60481715e-01
2.55930841e-01 1.28655684e+00 1.34222537e-01 -2.89209336e-01
3.47746700e-01 -6.68928958e-03 -1.82847366e-01 -6.45425618e-01
-6.03222013e-01 -1.66201502e-01 1.07616372e-01 -3.90783519e-01
-3.35272491e-01 -8.85366142e-01 -8.46231163e-01 -6.81095123e-01
-8.45980719e-02 2.18173712e-01 8.07914793e-01 4.62000161e-01
7.04826236e-01 2.31325865e-01 6.31542504e-01 -1.85206346e-02
-6.57784164e-01 -6.37198806e-01 -4.44136441e-01 6.25572741e-01
4.07917678e-01 -3.11123729e-01 1.79027636e-02 5.03016775e-03] | [6.333499908447266, 6.723089694976807] |
4fc129b6-c36c-4f82-8b5e-fd6001193439 | physics-informed-machine-learning-for-sensor | 2006.13380 | null | http://arxiv.org/abs/2006.13380v1 | http://arxiv.org/pdf/2006.13380v1.pdf | Physics-informed machine learning for sensor fault detection with flight test data | We develop data-driven algorithms to fully automate sensor fault detection in
systems governed by underlying physics. The proposed machine learning method
uses a time series of typical behavior to approximate the evolution of
measurements of interest by a linear time-invariant system. Given additional
data from related sensors, a Kalman observer is used to maintain a separate
real-time estimate of the measurement of interest. Sustained deviation between
the measurements and the estimate is used to detect anomalous behavior. A
decision tree, informed by integrating other sensor measurement values, is used
to determine the amount of deviation required to identify a sensor fault. We
validate the method by applying it to three test systems exhibiting various
types of sensor faults: commercial flight test data, an unsteady aerodynamics
model with dynamic stall, and a model for longitudinal flight dynamics forced
by atmospheric turbulence. In the latter two cases we test fault detection for
several prototypical failure modes. The combination of a learned dynamical
model with the automated decision tree accurately detects sensor faults in each
case. | [] | 2020-06-23 | null | null | null | null | ['physics-informed-machine-learning'] | ['graphs'] | [ 2.95699984e-01 -1.21191852e-01 1.84751838e-01 -9.10649374e-02
-4.42015320e-01 -4.25471365e-01 4.17125046e-01 7.05724776e-01
2.22750992e-01 7.63699949e-01 -5.43481946e-01 -4.00790781e-01
-6.24917626e-01 -4.61884379e-01 -5.69685400e-01 -8.99312735e-01
-4.03629005e-01 4.08323765e-01 5.07852554e-01 -3.36678892e-01
2.96997070e-01 8.01038921e-01 -1.65236807e+00 1.31510213e-01
5.90282500e-01 1.37612808e+00 -1.14707947e-01 1.04054737e+00
3.24580908e-01 5.05825102e-01 -8.84358883e-01 8.94697785e-01
2.97832578e-01 -4.00994509e-01 -3.82936776e-01 5.67802608e-01
1.16131380e-01 -1.05357289e-01 -1.40030384e-01 7.34598815e-01
2.50459518e-02 6.54422343e-02 6.53793275e-01 -1.26152933e+00
4.10952747e-01 4.82014529e-02 3.05321254e-02 2.79128700e-01
5.70206761e-01 3.56401712e-01 4.02143031e-01 -4.88285959e-01
1.88627481e-01 7.74271727e-01 7.11767375e-01 1.28895089e-01
-1.23406112e+00 -8.18278193e-02 2.33781408e-03 7.09133446e-02
-1.14960968e+00 -2.04481289e-01 8.44500005e-01 -7.28630245e-01
9.27157640e-01 1.89111516e-01 7.14324057e-01 5.77945590e-01
1.14354205e+00 -2.05339387e-01 1.25903833e+00 -2.33360440e-01
8.49447668e-01 -5.28074382e-03 2.85329193e-01 7.82940924e-01
5.46145558e-01 7.03311443e-01 -3.17516536e-01 -6.16103768e-01
4.15400714e-01 1.93756506e-01 -4.59529489e-01 -2.75554717e-01
-7.06729650e-01 1.88036531e-01 5.67806773e-02 2.40919277e-01
-6.27986908e-01 4.21981909e-04 3.12862337e-01 8.81253660e-01
4.29999799e-01 6.42135561e-01 -9.67017174e-01 -8.67834967e-03
-8.99486303e-01 2.41591021e-01 1.25345755e+00 3.29212099e-01
7.17069387e-01 6.20920122e-01 3.67963076e-01 1.82169363e-01
1.20799258e-01 5.93011677e-01 6.22049153e-01 -7.19672263e-01
-2.91477442e-01 7.59021759e-01 4.84435052e-01 -6.61215544e-01
-4.96021450e-01 -4.85237151e-01 -5.94627440e-01 6.33421004e-01
2.34431535e-01 -3.66804898e-01 -7.59823322e-01 1.16004348e+00
4.58210886e-01 6.21446371e-01 1.52211502e-01 7.47333586e-01
-2.96021014e-01 5.98896146e-01 -6.73547566e-01 -8.02303791e-01
7.31070995e-01 -1.58350468e-01 -6.49901330e-01 3.60472240e-02
6.67697728e-01 -4.28686887e-01 3.99250001e-01 8.14659297e-01
-6.83485866e-01 -8.47052813e-01 -1.41478276e+00 1.18683088e+00
-1.93049982e-01 -1.70184404e-01 -3.61418456e-01 1.80834711e-01
-8.23490560e-01 1.26527870e+00 -1.11834455e+00 -3.24590653e-01
-5.45256734e-01 1.57367676e-01 -1.45933360e-01 2.25412145e-01
-1.04840863e+00 8.41678917e-01 3.28401059e-01 1.27981827e-01
-1.37828493e+00 -5.53798556e-01 -5.78510463e-01 -2.58666396e-01
4.57727462e-01 -4.65531379e-01 1.16776299e+00 -5.53500354e-01
-1.43729651e+00 1.72865704e-01 -1.41650811e-01 -5.87725222e-01
4.04833734e-01 -4.40527022e-01 -9.17623878e-01 -9.03244410e-03
-2.32088700e-01 -7.05436885e-01 1.46223092e+00 -1.15748525e+00
-4.69860256e-01 -1.58880591e-01 -3.18849921e-01 -4.55716550e-01
-2.44662538e-02 -5.93817770e-01 5.81963599e-01 -1.07840702e-01
3.94338548e-01 -1.05183387e+00 -1.85671777e-01 -3.08521688e-01
-5.39999843e-01 2.52447650e-02 1.17779088e+00 -3.40902597e-01
1.38630128e+00 -2.04712462e+00 1.76787153e-01 4.73599494e-01
-2.08357181e-02 4.32540737e-02 3.29559416e-01 1.02286863e+00
-3.02134454e-01 -4.21264201e-01 -5.88679790e-01 1.73730955e-01
-4.52454686e-01 4.34043556e-01 -6.17177248e-01 7.38780141e-01
4.56990629e-01 -1.23394564e-01 -7.42901087e-01 1.14810295e-01
4.08489257e-01 -1.12066314e-01 -1.43923938e-01 7.21057355e-01
-4.11875784e-01 6.09446764e-01 -4.10803646e-01 6.05934918e-01
1.39535233e-01 2.61055306e-02 -6.55905232e-02 -8.12039003e-02
-3.16576213e-01 -1.09648973e-01 -1.48061705e+00 9.72252488e-01
-3.94282401e-01 2.92632908e-01 2.10633084e-01 -1.14549232e+00
1.28742635e+00 4.95195925e-01 7.81281650e-01 1.08004436e-02
2.48477429e-01 3.80766988e-01 9.29571316e-02 -8.47489715e-01
5.25985379e-03 -1.87253952e-02 -6.92645088e-02 2.89580405e-01
-1.69399843e-01 -6.30506814e-01 -6.97493106e-02 -2.60898143e-01
1.91420996e+00 1.95506215e-02 3.90135974e-01 -5.13466895e-01
7.34956980e-01 1.64670497e-01 7.35371888e-01 5.70799410e-01
-1.21804707e-01 1.14673272e-01 4.38320100e-01 -7.00305104e-01
-7.68807352e-01 -9.36865687e-01 -2.08550334e-01 2.81707227e-01
1.48472041e-01 -3.61348987e-01 -4.47631717e-01 -4.11749452e-01
5.29814661e-01 5.78839004e-01 -5.10332584e-01 -8.44269097e-01
-5.07817626e-01 -5.00431716e-01 6.25528172e-02 4.18483876e-02
3.71678583e-02 -5.99038124e-01 -1.04893148e+00 6.10522985e-01
6.06014371e-01 -7.82261968e-01 3.23789001e-01 6.22915745e-01
-1.00608039e+00 -1.72048140e+00 3.14855725e-01 -1.56223908e-01
6.76325560e-01 -2.49804214e-01 8.68378341e-01 3.97007585e-01
-4.87155288e-01 5.61870039e-01 -2.08294883e-01 -3.17474663e-01
-1.11545753e+00 -6.96436346e-01 6.93840384e-01 1.77784473e-01
-3.35581362e-01 -4.29044396e-01 -1.50116861e-01 3.96056384e-01
-1.00956440e+00 -5.54627717e-01 2.47479185e-01 9.16449308e-01
5.30120790e-01 8.18726063e-01 5.28823793e-01 -5.39043128e-01
6.27938271e-01 -7.26751626e-01 -9.05024111e-01 1.43591166e-01
-1.02185261e+00 2.27998465e-01 9.32636142e-01 -6.15633130e-01
-6.00407064e-01 2.03683287e-01 3.31561714e-01 -7.25418568e-01
-5.45837104e-01 6.80416405e-01 2.61558145e-01 1.63152024e-01
7.69990087e-01 1.83520347e-01 2.92521715e-01 -4.40449208e-01
-2.34792471e-01 4.52287585e-01 5.89469194e-01 -5.65493762e-01
9.97718692e-01 1.12912133e-01 5.44875026e-01 -1.12742543e+00
-6.18290484e-01 -3.90219003e-01 -7.26708889e-01 -5.96369445e-01
1.67478129e-01 -5.54580212e-01 -5.97498775e-01 7.70331621e-01
-8.09614480e-01 -3.24512303e-01 -5.31603217e-01 5.04204452e-01
-6.34572268e-01 -2.85422504e-02 -7.04262197e-01 -1.25067008e+00
-1.73591860e-02 -8.60272110e-01 1.11757052e+00 4.49845716e-02
-3.10898691e-01 -1.20179331e+00 5.79559445e-01 -4.40388322e-01
4.22923535e-01 9.05042291e-01 7.67317593e-01 -8.12554955e-01
-1.61538392e-01 -9.78178084e-01 6.69535041e-01 6.15730762e-01
5.24086058e-01 4.16805476e-01 -7.83301234e-01 -8.34770679e-01
7.51718521e-01 -2.02598292e-02 4.53797489e-01 1.54197127e-01
5.62180638e-01 -1.02821045e-01 -5.24893463e-01 8.07300657e-02
1.53972244e+00 3.40303898e-01 -1.85930192e-01 -1.94816273e-02
3.13549399e-01 4.31474477e-01 9.58108008e-01 7.42550433e-01
-4.48246062e-01 4.20345545e-01 6.33467138e-01 2.25387946e-01
5.16335785e-01 2.02507451e-01 7.79433548e-01 7.50723779e-01
3.50539476e-01 -7.94164091e-02 -9.99105453e-01 2.81080097e-01
-1.61467564e+00 -4.85297889e-01 -1.82356060e-01 2.55591965e+00
4.17233825e-01 6.55030608e-01 -5.26503958e-02 7.94517100e-01
6.03786111e-01 -2.48544946e-01 -9.42824185e-01 -3.62148464e-01
2.66245723e-01 1.96849406e-02 3.86235625e-01 4.93921727e-01
-8.55761647e-01 1.79083198e-01 6.17676163e+00 -1.21957570e-01
-1.35856640e+00 -3.70549142e-01 6.21044226e-02 1.89280689e-01
3.82387727e-01 1.82537019e-01 -5.65379560e-01 4.42691416e-01
1.69899416e+00 -2.03260317e-01 1.97473899e-01 6.65046096e-01
4.33596283e-01 -2.58059293e-01 -1.29612851e+00 2.69113481e-01
-1.65894732e-01 -9.27770674e-01 -5.42664766e-01 -1.02524469e-02
4.22748923e-01 -1.67435303e-01 -2.90537536e-01 3.85060348e-02
1.38749138e-01 -4.03187394e-01 4.62127179e-01 1.02247655e+00
4.18216258e-01 -3.84208173e-01 9.36933994e-01 5.50261974e-01
-1.19350457e+00 -6.60320699e-01 5.52892685e-04 -5.69548965e-01
2.53185570e-01 1.22041368e+00 -9.35247779e-01 6.24777019e-01
3.27865422e-01 8.61533403e-01 -4.87213045e-01 9.54444230e-01
7.09320279e-03 9.27189767e-01 -5.03705084e-01 2.50770092e-01
-2.19633982e-01 -3.64832059e-02 1.11475062e+00 6.35747612e-01
5.84003806e-01 -1.04387484e-01 7.00654149e-01 4.78023291e-01
8.52666140e-01 -4.55670267e-01 -9.96304810e-01 -5.86638460e-03
6.00977242e-01 1.02268040e+00 -5.89721441e-01 -4.31091845e-01
3.82874869e-02 5.63562810e-01 -1.68150306e-01 2.43386105e-01
-6.15373850e-01 -1.06038041e-01 8.31577659e-01 4.35493618e-01
2.74788216e-02 -4.51083958e-01 1.41115665e-01 -8.82155478e-01
-1.74047023e-01 -5.31811893e-01 2.49897629e-01 -6.39036894e-01
-1.15988421e+00 6.29607856e-01 2.05752775e-01 -1.75224853e+00
-9.78979170e-01 -6.79481745e-01 -9.64550018e-01 7.03170180e-01
-1.04748893e+00 -2.26049364e-01 -3.62859249e-01 3.65336865e-01
4.11885232e-01 -1.60931423e-01 8.64288926e-01 -2.79818326e-01
-8.01274300e-01 -3.00347954e-01 1.35147631e-01 -4.64265496e-01
6.99552178e-01 -1.29425895e+00 1.18887946e-01 9.20935154e-01
-3.22233737e-01 1.78496212e-01 1.42888618e+00 -1.11351192e+00
-1.87162602e+00 -1.20969796e+00 2.80849010e-01 -4.19661909e-01
1.07084632e+00 1.36770770e-01 -1.42083442e+00 2.81729788e-01
-1.23255603e-01 4.80209351e-01 8.01301375e-02 -5.72211206e-01
1.13618933e-01 -3.97123128e-01 -1.19143236e+00 -1.08524226e-01
4.16763395e-01 -3.45565349e-01 -7.67137527e-01 3.38193297e-01
5.35824955e-01 -4.85979825e-01 -1.16898561e+00 7.97225714e-01
1.15655795e-01 -9.17113364e-01 3.18236858e-01 -5.90531707e-01
2.03306433e-02 -5.58591187e-01 -1.27708867e-01 -1.69574380e+00
-2.55706519e-01 -7.91249871e-01 -4.76071864e-01 8.15764844e-01
3.97992358e-02 -8.94874930e-01 3.84943128e-01 3.59241307e-01
-4.47046757e-01 -8.63174915e-01 -1.00552750e+00 -9.90246654e-01
-3.46639961e-01 -3.18221897e-01 2.17494115e-01 6.18300378e-01
8.32706466e-02 1.97306201e-02 -3.17104235e-02 8.51716757e-01
6.22523844e-01 1.65682778e-01 5.32928884e-01 -1.51304162e+00
-4.72678810e-01 1.74617961e-01 -7.20561028e-01 -1.77639261e-01
1.59205332e-01 -2.58941501e-01 4.51784611e-01 -8.31536531e-01
-8.04977357e-01 -1.04343593e-02 -5.71852446e-01 1.16919525e-01
7.99974613e-03 -4.38570559e-01 -3.89635682e-01 2.05846727e-01
-3.17565016e-02 2.58411735e-01 3.74345779e-01 2.14882255e-01
-2.76144028e-01 4.92207646e-01 2.24526927e-01 7.69621909e-01
8.77007484e-01 -3.73139977e-01 -4.11572129e-01 3.30078036e-01
-3.78542878e-02 8.14772904e-01 4.65641081e-01 -1.83094418e+00
2.33778641e-01 -2.76205271e-01 3.37180525e-01 -4.55257118e-01
1.65757611e-01 -1.20551300e+00 5.42211950e-01 1.12431407e+00
-1.95939824e-01 6.51797295e-01 3.58896047e-01 9.62898970e-01
-4.25540447e-01 1.68086980e-02 8.35492671e-01 3.22856367e-01
-5.94038725e-01 5.10192513e-02 -7.56822050e-01 -4.05487299e-01
1.20106471e+00 -3.86056006e-02 -1.06415309e-01 -2.28172779e-01
-8.93488169e-01 7.15008080e-02 5.38364410e-01 2.03184694e-01
8.19979072e-01 -8.35914731e-01 -5.78146279e-01 8.73508155e-01
5.33356443e-02 -3.61596495e-01 -4.93706577e-02 8.59143138e-01
-9.37466398e-02 2.50149429e-01 -1.87657718e-02 -9.85532820e-01
-1.10030675e+00 6.23566926e-01 7.42186427e-01 -5.57656027e-02
-3.95114005e-01 2.57970631e-01 -5.69940805e-01 2.45475601e-02
-2.32879356e-01 -7.29422927e-01 2.60000765e-01 -3.24371636e-01
4.02471274e-01 4.53547835e-01 5.73931158e-01 -2.78339893e-01
-2.83407211e-01 5.75434983e-01 4.31545466e-01 -6.00326527e-03
9.52467740e-01 -6.15535155e-02 -1.05605692e-01 1.30982542e+00
8.46827626e-01 -2.93734878e-01 -1.61364067e+00 -1.72763675e-01
2.77781844e-01 -2.50630472e-02 2.30936576e-02 -7.13715374e-01
-7.29425251e-01 3.43142390e-01 7.97527730e-01 1.00921798e+00
1.34934151e+00 -2.62698114e-01 4.41265702e-01 4.90825951e-01
6.16136551e-01 -7.98120320e-01 -9.97443423e-02 5.81342638e-01
8.06673408e-01 -8.51518095e-01 6.89477995e-02 -1.60062417e-01
-9.20645371e-02 1.54591393e+00 6.45483196e-01 -7.37822831e-01
1.19876814e+00 7.27270603e-01 -1.68194287e-02 -2.13757724e-01
-1.43182921e+00 1.18784130e-01 1.44886121e-01 2.37196743e-01
-7.20583126e-02 -1.47157624e-01 1.01059280e-01 -1.24339424e-01
6.41472405e-03 -3.53441909e-02 5.78836918e-01 1.34337270e+00
-8.31509948e-01 -7.47493088e-01 -8.00626338e-01 6.67323470e-01
4.17022891e-02 5.86368680e-01 -3.30706477e-01 7.64159858e-01
-1.55725665e-02 1.15661800e+00 2.20917940e-01 -7.96201646e-01
6.36988401e-01 4.92537469e-01 1.00187518e-01 -5.45576274e-01
-3.86713892e-01 4.55935188e-02 -1.19705102e-04 -8.16876709e-01
-6.56925663e-02 -7.01996863e-01 -1.28650236e+00 4.54858281e-02
-2.62881845e-01 4.37588215e-01 6.95270598e-01 1.26127946e+00
3.79231453e-01 8.89088809e-01 1.09258544e+00 -7.13353336e-01
-9.21032190e-01 -9.23400581e-01 -8.33279490e-01 3.08255315e-01
9.11713839e-01 -8.75772655e-01 -9.21236098e-01 1.69664368e-01] | [6.588922023773193, 2.490638494491577] |
fbdcbf09-0146-47ad-8b37-97f9544bb0a5 | logical-story-representations-via-framenet | null | null | https://openreview.net/forum?id=76qMNL5EUjq | https://openreview.net/pdf?id=76qMNL5EUjq | Logical Story Representations via FrameNet + Semantic Parsing | We present a means of obtaining rich semantic representations of stories by combining neural FrameNet identification, a formal logic-based semantic parser, and a hierarchical event schema representation. The final schematic representation of the story abstracts constants to variables, preserving their types and relationships to other individuals in the story. All identified FrameNet frames are incorporated as temporally bound ``episodes'' and related to one another in time. The semantic role information from the frames is also incorporated into the final schema's type constraints. We describe this system as well as its possible applications to question answering and open-domain event schema learning. | ['Anonymous'] | 2022-01-16 | null | null | null | acl-arr-january-2022-1 | ['formal-logic'] | ['reasoning'] | [ 8.58870894e-02 6.89273775e-01 -5.48259020e-01 -6.23236716e-01
-2.70413250e-01 -7.19416380e-01 9.64008093e-01 6.70822978e-01
-3.64851594e-01 1.15166152e+00 7.87383676e-01 1.88595816e-01
-3.44542593e-01 -1.29123354e+00 -7.15076149e-01 -2.93044388e-01
-1.83315098e-01 6.54490829e-01 8.29097867e-01 -4.12553847e-01
-7.56819397e-02 7.28032887e-02 -1.71529782e+00 8.05920660e-01
-4.92795035e-02 8.77804577e-01 1.63063869e-01 4.75943536e-01
-7.15803742e-01 1.72788155e+00 -7.86438048e-01 -4.67229277e-01
-4.87115502e-01 -7.10602701e-01 -1.37310624e+00 -1.36195170e-02
-2.29906067e-01 -9.44399685e-02 -5.64612031e-01 5.73134303e-01
-1.14302166e-01 7.99110055e-01 1.58727840e-01 -1.36637151e+00
-5.51621974e-01 1.12949836e+00 2.92475581e-01 5.85625529e-01
1.05624223e+00 -5.58433414e-01 1.39579093e+00 -4.22582269e-01
1.41679204e+00 1.53430223e+00 5.11606932e-01 6.13317549e-01
-1.09307933e+00 -5.95643856e-02 3.78188401e-01 7.06352115e-01
-1.02619016e+00 -4.73303020e-01 1.00476611e+00 -3.62353563e-01
1.55155647e+00 1.81443959e-01 1.02823901e+00 1.18766260e+00
4.55895811e-02 7.47851133e-01 3.70610893e-01 -7.18219459e-01
6.83584750e-01 -6.67088628e-02 5.45331597e-01 7.59258986e-01
-1.17567852e-01 -9.51626226e-02 -9.91625726e-01 -2.80408591e-01
9.59951580e-01 -2.92998552e-01 2.96112955e-01 -2.27998704e-01
-1.21455586e+00 9.26314771e-01 9.19693485e-02 5.27585268e-01
-3.30592871e-01 6.40866876e-01 7.15297103e-01 9.26986262e-02
3.60898346e-01 3.29699606e-01 -3.85683328e-01 -4.92200069e-02
-4.07196522e-01 6.85468435e-01 8.79060924e-01 1.18590713e+00
5.29271483e-01 7.45435618e-03 -9.54723805e-02 6.09177768e-01
3.06231648e-01 -2.28181198e-01 4.15241778e-01 -1.33600807e+00
1.04560949e-01 7.58084416e-01 2.38187477e-01 -9.35225070e-01
-5.74407339e-01 2.13118136e-01 1.13522418e-01 -5.50462484e-01
1.90059558e-01 1.83372393e-01 -3.86343688e-01 2.10514402e+00
3.93638432e-01 5.30450523e-01 4.18631822e-01 4.96502101e-01
1.07767177e+00 8.45932245e-01 7.17866838e-01 -5.76739490e-01
1.86556351e+00 -3.46874356e-01 -1.13270414e+00 -1.98655546e-01
5.52616537e-01 -6.33353814e-02 4.75391090e-01 -1.93672463e-01
-1.35261810e+00 -1.38198778e-01 -8.16312611e-01 -3.19246352e-01
-5.66479206e-01 -5.89191139e-01 1.05427551e+00 4.53038514e-02
-8.66119742e-01 3.73245537e-01 -9.72300708e-01 -6.98099673e-01
3.12042773e-01 1.43761411e-01 -2.55382210e-01 3.86100531e-01
-1.84651029e+00 9.28656280e-01 1.14010966e+00 -4.52931762e-01
-8.85066569e-01 -5.23131490e-01 -1.45050395e+00 6.27460554e-02
6.54561400e-01 -8.28527391e-01 1.65495420e+00 -8.36501241e-01
-1.17911291e+00 1.15464008e+00 -5.41451633e-01 -8.61664772e-01
-2.49499902e-01 9.66177881e-02 -9.50412869e-01 6.98609769e-01
2.34104544e-01 6.04213238e-01 4.24515575e-01 -9.95792627e-01
-8.69623005e-01 -3.06877941e-01 8.44262660e-01 2.97570527e-01
1.29843295e-01 7.86919773e-01 -2.19387710e-01 -7.80347347e-01
3.31215799e-01 -3.79956305e-01 -1.08895242e-01 -1.79658249e-01
4.53813449e-02 -6.65432036e-01 5.77354789e-01 -3.87587935e-01
1.40636528e+00 -1.83962309e+00 2.98157424e-01 -1.51680246e-01
-5.18478788e-02 -5.62860250e-01 2.66636014e-01 4.89907205e-01
-3.69850188e-01 5.04788663e-03 6.55647293e-02 1.40277520e-01
2.56713569e-01 9.29088175e-01 -6.77176118e-01 2.02639788e-01
1.30919188e-01 9.03447628e-01 -1.15438759e+00 -8.70966434e-01
3.73232663e-01 4.79148589e-02 -5.01145959e-01 9.02736038e-02
-1.03124464e+00 -1.00654729e-01 -7.96494424e-01 3.97611558e-01
-2.94280767e-01 -6.00948930e-02 5.94748974e-01 -1.01836532e-01
-9.66414437e-02 9.59986091e-01 -1.28498292e+00 1.90737438e+00
-8.17957819e-02 5.69295645e-01 -4.21272695e-01 -1.07809675e+00
5.84888875e-01 1.01611054e+00 4.89162743e-01 -4.85712886e-01
1.53331190e-01 -4.02807653e-01 -6.57128930e-01 -6.81218684e-01
6.31095886e-01 -6.62576497e-01 -7.05849946e-01 6.02180481e-01
5.11802554e-01 -1.07145719e-01 7.29900599e-01 4.75911915e-01
8.38350654e-01 6.36058331e-01 8.32122207e-01 -2.37743318e-01
5.03772020e-01 3.92307848e-01 7.80901551e-01 5.18391371e-01
-9.14142579e-02 4.62099761e-02 7.79124498e-01 -7.66116142e-01
-9.39794838e-01 -1.17370629e+00 -1.62467226e-01 1.34324968e+00
3.61480534e-01 -9.69136596e-01 -6.47312522e-01 -3.28333527e-01
-4.17360216e-01 1.14815736e+00 -7.79145241e-01 2.54965685e-02
-8.65618348e-01 -3.41299146e-01 7.56249666e-01 9.39190865e-01
1.03986956e-01 -1.45928931e+00 -1.11255240e+00 6.87417090e-01
-6.07998490e-01 -1.12226319e+00 2.59133607e-01 3.10282499e-01
-6.17175221e-01 -1.18167162e+00 4.06530738e-01 -9.37416971e-01
2.71971762e-01 -4.57027406e-01 1.33921289e+00 -1.34751901e-01
1.44382879e-01 5.17428815e-01 -4.26460445e-01 -4.73946035e-01
-3.68537933e-01 -4.34811950e-01 -1.07294358e-01 -2.26302400e-01
4.30833668e-01 -4.81920600e-01 -2.51626465e-02 1.61858290e-01
-1.04051912e+00 4.96638827e-02 -7.76326895e-01 4.92828876e-01
6.73725426e-01 4.64713156e-01 4.81808901e-01 -9.92183506e-01
2.43388429e-01 -7.51868665e-01 -5.15978396e-01 4.78491932e-01
2.25328818e-01 8.43833834e-02 3.04968238e-01 -3.38706255e-01
-1.75548542e+00 -1.85627729e-01 2.90332902e-02 5.74834459e-02
-4.19951379e-01 7.20748782e-01 -4.45458591e-01 9.25610900e-01
8.68515134e-01 -1.69913203e-01 -4.85832900e-01 -9.50288624e-02
5.41583419e-01 -2.31432155e-01 9.34418499e-01 -1.07990205e+00
3.86923701e-01 7.05716968e-01 -6.54707104e-02 -4.99474138e-01
-1.24484539e+00 -2.24920675e-01 -6.54697239e-01 -3.46591949e-01
1.20313561e+00 -8.55149925e-01 -5.61393023e-01 1.02270141e-01
-1.48414028e+00 -5.66421337e-02 -9.73516762e-01 3.02509606e-01
-1.03756690e+00 -2.97500864e-02 -7.81860709e-01 -6.19953394e-01
3.73351753e-01 -5.18299222e-01 7.53374815e-01 3.39416534e-01
-9.98644292e-01 -1.39080441e+00 3.37338895e-02 1.17573753e-01
-2.88734853e-01 5.08695781e-01 1.24709785e+00 -8.63274097e-01
-1.80642381e-01 6.69400766e-02 2.95030117e-01 -4.98451471e-01
-1.01888947e-01 -7.97763392e-02 -6.30709648e-01 3.62491101e-01
3.29654574e-01 -2.96240926e-01 3.24778020e-01 4.62140501e-01
7.28835821e-01 -4.05216157e-01 -5.31943679e-01 1.73440129e-02
1.48007095e+00 8.45039427e-01 6.15187705e-01 3.79767805e-01
2.08197013e-01 7.57804155e-01 3.44486624e-01 5.36502957e-01
6.25238180e-01 5.47226906e-01 -3.85589376e-02 5.21984994e-01
-9.21858773e-02 -5.76030433e-01 2.76225448e-01 1.67659566e-01
-8.36569369e-02 -4.14576977e-01 -8.26990187e-01 7.53933012e-01
-2.17234778e+00 -1.82270193e+00 -2.14181185e-01 1.45133948e+00
1.01462734e+00 3.12502831e-01 3.14954817e-02 1.45910814e-01
1.06330764e+00 6.06945753e-01 -2.97280610e-01 -4.13456172e-01
-5.60692251e-01 2.59874910e-01 -6.32583722e-02 6.10792816e-01
-1.06553066e+00 1.28602445e+00 7.63264418e+00 5.03752887e-01
-1.55723199e-01 4.04951721e-01 1.12425901e-01 -1.28970563e-01
-7.06385016e-01 5.67326128e-01 -5.86176276e-01 8.14359412e-02
1.09468329e+00 -5.71535587e-01 4.67565060e-01 6.36217058e-01
1.22058064e-01 -3.86566788e-01 -1.48261178e+00 5.81871808e-01
1.93379104e-01 -2.00817275e+00 2.15128347e-01 -6.72549427e-01
3.47930104e-01 -6.75204694e-01 -9.38504100e-01 1.52570188e-01
6.18682861e-01 -5.70605159e-01 1.54887569e+00 5.40348589e-01
5.41563928e-01 -7.01157987e-01 2.44705915e-01 -3.70351993e-03
-1.50701761e+00 -2.24187553e-01 -1.73092559e-01 -5.30400097e-01
6.32099330e-01 1.59815297e-01 -3.39707583e-01 7.65900135e-01
5.06509185e-01 8.71063232e-01 2.70039178e-02 3.70180607e-01
-7.34278440e-01 5.33257425e-01 -2.74281561e-01 8.66230503e-02
-8.52609240e-03 7.05381334e-02 7.25962400e-01 9.57436442e-01
-1.50452062e-01 9.14493084e-01 1.38054863e-01 1.01400137e+00
3.18712555e-02 -4.32459772e-01 -5.79664409e-01 -8.67658406e-02
7.22005367e-01 5.00576258e-01 -1.06159127e+00 -7.58674502e-01
-4.61527944e-01 5.20919323e-01 -8.30898285e-02 3.08486760e-01
-1.14675212e+00 -2.20549777e-01 6.67082250e-01 -6.84511289e-03
2.30084993e-02 2.07217671e-02 -6.45490661e-02 -1.18972683e+00
-2.32989863e-01 -3.13338459e-01 1.22183049e+00 -1.36813378e+00
-9.00045753e-01 4.83600616e-01 7.31692433e-01 -5.02958894e-01
-5.42487681e-01 -2.19562590e-01 -5.06871641e-01 2.89321154e-01
-8.49107683e-01 -1.06205809e+00 1.74523458e-01 1.03683043e+00
7.26274848e-01 1.16331138e-01 1.26092899e+00 -1.68118197e-02
-3.77963841e-01 -1.00358382e-01 -7.21354604e-01 2.59697706e-01
3.43158506e-02 -9.96127725e-01 2.25730896e-01 9.42636549e-01
1.73033565e-01 5.29430509e-01 9.20018256e-01 -9.17517483e-01
-1.08085454e+00 -7.98717201e-01 1.48943746e+00 -7.34232903e-01
1.02849960e+00 -4.26208317e-01 -9.65940833e-01 1.47614574e+00
2.40634859e-01 -2.06325784e-01 7.52364635e-01 1.61045685e-01
-4.91159648e-01 2.37488702e-01 -1.07673311e+00 5.22765279e-01
1.42979062e+00 -9.02672052e-01 -1.90182996e+00 3.75138402e-01
1.19116688e+00 -5.53159058e-01 -8.03546131e-01 -4.18315409e-03
2.08611637e-01 -5.98109424e-01 1.06300879e+00 -1.22246778e+00
4.99660194e-01 -3.82234484e-01 -3.91613454e-01 -6.20250642e-01
-3.24258059e-01 -5.25359750e-01 -3.01606864e-01 1.33150041e+00
1.83380872e-01 -2.32324094e-01 5.53306699e-01 8.70332658e-01
-2.62629330e-01 -3.00256386e-02 -1.07725501e+00 -4.85559434e-01
-2.70022929e-01 -8.93755555e-01 8.08083832e-01 1.20068502e+00
7.21723258e-01 4.28922623e-01 1.08610876e-01 7.35874772e-02
2.96857685e-01 2.12029740e-01 -3.40009034e-01 -1.37572539e+00
-5.55896200e-02 -2.21126109e-01 -4.54124063e-01 -3.75478566e-01
7.96060681e-01 -9.99638319e-01 -6.49569482e-02 -1.76244545e+00
5.22202291e-02 -4.16137688e-02 -2.33569250e-01 8.49667728e-01
3.82281274e-01 -2.04706937e-01 -1.62663177e-01 -7.46032521e-02
-8.16527188e-01 2.70695865e-01 5.67720234e-01 1.27348676e-02
-2.46671990e-01 -5.74679494e-01 -5.23240566e-01 1.08619833e+00
6.60650611e-01 -6.47871077e-01 -8.56485963e-01 -4.59377229e-01
7.24888325e-01 7.77538717e-01 5.45548797e-01 -6.97711229e-01
3.32203835e-01 -5.80529809e-01 2.19133094e-01 -3.46870989e-01
4.65278327e-01 -6.78511441e-01 7.79811680e-01 4.57488485e-02
-8.03411126e-01 7.61050954e-02 2.82759935e-01 3.89478683e-01
-4.36199546e-01 -3.58862728e-01 7.13182390e-01 -4.42866772e-01
-1.25771689e+00 -3.03459913e-01 -7.74824381e-01 2.39248618e-01
1.29200125e+00 -2.77030885e-01 -3.37105185e-01 -2.04389796e-01
-1.34783947e+00 1.05859771e-01 1.98141724e-01 5.26930392e-01
5.91799736e-01 -1.38324165e+00 -1.69303492e-01 -9.87118557e-02
-5.58302738e-02 2.07642652e-02 2.10539564e-01 1.26404077e-01
-3.85822952e-01 4.66727436e-01 -4.06417161e-01 -9.76695418e-02
-8.76967251e-01 6.87648356e-01 3.95144790e-01 -1.76135823e-02
-8.31534445e-01 8.14090192e-01 1.98795763e-03 1.31041571e-01
2.20969707e-01 -4.67891991e-01 -4.81180012e-01 5.04273355e-01
6.32543623e-01 1.42743498e-01 -1.90960750e-01 -7.90305078e-01
-6.30402446e-01 -1.07343737e-02 3.27916801e-01 -5.22384942e-01
1.42547798e+00 -3.19145769e-01 -5.86722136e-01 8.33075106e-01
6.46134257e-01 -4.70406651e-01 -9.15210187e-01 -3.76107514e-01
5.83279371e-01 -2.06184328e-01 -3.25678378e-01 -4.97908801e-01
-4.68758881e-01 4.88345139e-03 -2.36607730e-01 5.26410103e-01
8.78752589e-01 8.31897438e-01 5.90770662e-01 2.90599316e-01
5.85705757e-01 -1.18600035e+00 1.39130689e-02 7.10719109e-01
8.43414843e-01 -3.08497548e-01 -7.99280126e-03 -5.34826458e-01
-5.59257746e-01 9.76213932e-01 3.91536891e-01 -1.67336971e-01
4.08971608e-01 3.71494919e-01 -3.77111346e-01 -5.54188669e-01
-1.15091550e+00 -1.23909645e-01 -2.18454421e-01 4.88465279e-01
2.64506221e-01 -4.88925762e-02 -4.43653911e-01 1.08342302e+00
-2.65459538e-01 1.16870932e-01 6.24037802e-01 1.33290064e+00
-3.81648839e-01 -1.05284464e+00 -4.28105026e-01 1.23401545e-02
-4.39372569e-01 1.63127378e-01 -2.88181931e-01 8.87414277e-01
4.75910991e-01 9.32441592e-01 7.19879389e-01 2.02644214e-01
4.66676921e-01 7.17331111e-01 7.73626089e-01 -9.88393188e-01
-5.41954994e-01 -6.46129018e-03 7.72285700e-01 -6.91906691e-01
-1.01079535e+00 -9.67135012e-01 -2.20622897e+00 -7.35939220e-02
6.37183860e-02 2.78843552e-01 8.93519074e-02 1.24404657e+00
-8.39184597e-02 6.06502116e-01 2.82498337e-02 -2.53595382e-01
3.73309851e-01 -2.45063782e-01 -4.47017193e-01 6.96901679e-01
3.41524221e-02 -8.52055132e-01 -4.24741246e-02 6.51665986e-01] | [10.291892051696777, 9.145818710327148] |
fb675ea2-d7b1-455f-bcd8-46946d2dbcbf | tapir-tracking-any-point-with-per-frame | 2306.08637 | null | https://arxiv.org/abs/2306.08637v1 | https://arxiv.org/pdf/2306.08637v1.pdf | TAPIR: Tracking Any Point with per-frame Initialization and temporal Refinement | We present a novel model for Tracking Any Point (TAP) that effectively tracks any queried point on any physical surface throughout a video sequence. Our approach employs two stages: (1) a matching stage, which independently locates a suitable candidate point match for the query point on every other frame, and (2) a refinement stage, which updates both the trajectory and query features based on local correlations. The resulting model surpasses all baseline methods by a significant margin on the TAP-Vid benchmark, as demonstrated by an approximate 20% absolute average Jaccard (AJ) improvement on DAVIS. Our model facilitates fast inference on long and high-resolution video sequences. On a modern GPU, our implementation has the capacity to track points faster than real-time. Visualizations, source code, and pretrained models can be found on our project webpage. | ['Andrew Zisserman', 'Joao Carreira', 'Yusuf Aytar', 'Ankush Gupta', 'Dilara Gokay', 'Mel Vecerik', 'Yi Yang', 'Carl Doersch'] | 2023-06-14 | null | null | null | null | ['visual-tracking', 'motion-estimation'] | ['computer-vision', 'computer-vision'] | [ 1.55891255e-01 -5.70344806e-01 -3.97721797e-01 -4.98744175e-02
-1.15277314e+00 -6.87167883e-01 5.89958549e-01 5.21489322e-01
-5.09784639e-01 2.69341081e-01 1.10968478e-01 3.41029488e-03
-8.62719398e-03 -7.42403090e-01 -7.47965455e-01 -4.60208982e-01
-4.97745693e-01 5.07158220e-01 1.03240502e+00 1.23886749e-01
5.79878271e-01 7.55719900e-01 -1.54705954e+00 6.63866550e-02
2.61159152e-01 1.31894386e+00 2.05531716e-01 9.75933492e-01
1.41160429e-01 6.05191827e-01 -2.94621766e-01 -2.74038225e-01
2.38480419e-01 4.51955758e-02 -6.89303935e-01 -3.06568742e-01
9.15574193e-01 -4.21421796e-01 -4.53377306e-01 8.35632920e-01
4.69612002e-01 3.48798394e-01 3.55696708e-01 -1.02605546e+00
-1.36981100e-01 -9.82763171e-02 -8.27799499e-01 5.41246831e-01
6.45747244e-01 3.70543487e-02 1.12235510e+00 -1.13470030e+00
7.85906792e-01 8.75724077e-01 1.01976132e+00 -1.18031077e-01
-1.05480087e+00 -6.04387522e-01 7.21913278e-02 3.67887467e-01
-1.61915958e+00 -4.11993235e-01 3.94419104e-01 -4.41730410e-01
9.88669634e-01 2.09634498e-01 7.93202579e-01 5.11688292e-01
1.30476773e-01 6.10947371e-01 4.09076810e-01 -1.36922553e-01
9.86944363e-02 -4.87527460e-01 -1.17192492e-01 8.75703394e-01
-4.69576269e-02 2.31405810e-01 -8.11321318e-01 -5.23223162e-01
1.06068850e+00 -9.51956958e-02 -2.39286929e-01 -6.22824788e-01
-1.49773037e+00 4.74889874e-01 3.87535125e-01 -6.17917404e-02
-6.15776122e-01 4.04721737e-01 3.20137292e-01 1.77578911e-01
2.42217198e-01 4.47797656e-01 -2.71388024e-01 -6.99337780e-01
-1.11950147e+00 3.27290237e-01 6.28348231e-01 1.07592583e+00
7.44527221e-01 -4.87153888e-01 -1.46234572e-01 5.26238620e-01
-1.89620815e-02 6.15225315e-01 -9.81487334e-02 -1.34595919e+00
2.97441691e-01 4.18522358e-01 2.93121308e-01 -1.28062725e+00
-2.30550319e-01 -2.44126126e-01 -4.17685837e-01 2.21957073e-01
2.09512487e-01 6.14836812e-02 -4.77503926e-01 1.47200859e+00
6.58119559e-01 7.12309182e-01 -4.68345165e-01 9.74719107e-01
5.41018605e-01 6.76018834e-01 -1.36202872e-01 -9.07644629e-02
1.26055229e+00 -9.75071192e-01 -4.07177150e-01 1.71124175e-01
4.00279284e-01 -9.54049289e-01 7.61134148e-01 2.69586205e-01
-1.34623492e+00 -5.78687072e-01 -1.01609480e+00 -1.10260651e-01
-1.62154660e-01 -5.69368303e-02 4.87217933e-01 1.83172151e-01
-1.24924350e+00 7.49794245e-01 -1.19313860e+00 -3.75808626e-01
4.51493502e-01 4.59171653e-01 -3.65570456e-01 1.50289740e-02
-8.04148674e-01 5.02347410e-01 -4.49559353e-02 -2.15367824e-01
-7.09979951e-01 -9.62814510e-01 -7.54146934e-01 2.12717578e-02
5.03819108e-01 -6.36237979e-01 1.13687909e+00 -5.02471566e-01
-1.22056580e+00 8.72098863e-01 -6.02570713e-01 -3.80370349e-01
6.89663827e-01 -6.59461200e-01 -2.22402483e-01 5.14231503e-01
1.50655493e-01 6.45430028e-01 8.05633605e-01 -9.97069776e-01
-9.86304641e-01 -1.91877127e-01 5.47058918e-02 2.83309847e-01
7.45240077e-02 1.63780630e-01 -1.29828000e+00 -7.04215944e-01
3.19859385e-02 -9.71066296e-01 -8.99289846e-02 4.85812515e-01
-7.86655676e-03 -1.34830892e-01 8.29945028e-01 -5.48685551e-01
1.42326236e+00 -2.12345028e+00 -4.96834591e-02 4.31565702e-01
2.25103199e-01 3.13289106e-01 -1.66427270e-02 4.92260367e-01
9.18830857e-02 -3.02129745e-01 1.02781318e-01 -1.76703244e-01
-2.85897642e-01 -1.28025651e-01 -2.78706878e-01 6.36048079e-01
-1.69141740e-02 7.89294958e-01 -1.19722593e+00 -5.80703974e-01
2.10513353e-01 4.80567038e-01 -6.94747269e-01 1.33893967e-01
-1.43751572e-03 2.39048779e-01 -3.21777731e-01 4.64474022e-01
4.61068094e-01 -6.14161849e-01 -1.15329111e-02 -2.86362529e-01
-2.90927082e-01 2.72506118e-01 -1.36143136e+00 2.02126455e+00
2.59135105e-02 7.92275012e-01 -1.31355137e-01 -4.93507206e-01
7.64088452e-01 1.77847326e-01 9.39066350e-01 -4.30623740e-01
-3.00375819e-01 -1.32334931e-03 -4.80338246e-01 -2.46130079e-01
8.54230940e-01 5.35453916e-01 3.10980737e-01 4.75659937e-01
-4.27917451e-01 2.67406940e-01 1.64412633e-01 2.17714950e-01
1.50978374e+00 4.40154314e-01 6.35365024e-02 -1.15669474e-01
4.50068712e-01 2.36649662e-01 4.69505012e-01 7.70150721e-01
-1.98406860e-01 8.26973379e-01 2.91563869e-01 -5.40279031e-01
-1.00767076e+00 -1.24551475e+00 -7.24367201e-02 1.06404996e+00
5.97664297e-01 -6.72522008e-01 -4.73417312e-01 -2.67927259e-01
2.15691462e-01 9.12330970e-02 -4.63441014e-01 2.23434046e-01
-8.19163382e-01 -1.09765500e-01 2.97135770e-01 7.63601243e-01
4.04907048e-01 -7.23819911e-01 -9.59964216e-01 7.80081451e-02
-1.76227659e-01 -1.01324797e+00 -1.08826184e+00 -2.27908999e-01
-9.12856877e-01 -1.29988730e+00 -6.30914986e-01 -7.20121443e-01
4.37599182e-01 5.24681866e-01 1.22475493e+00 3.65983307e-01
-2.76519209e-01 5.12216747e-01 -2.24047840e-01 4.95756790e-02
1.14398479e-01 1.20823182e-01 5.99135505e-03 -2.24295557e-01
4.91696656e-01 -3.45131427e-01 -9.40047503e-01 5.91262996e-01
-4.80954111e-01 -4.22446690e-02 2.17524394e-01 4.59651500e-01
1.08552802e+00 -1.36170805e-01 7.12377280e-02 -3.21606427e-01
1.53733313e-01 -3.85060996e-01 -9.79640603e-01 2.29083151e-01
-4.19777364e-01 -1.69107348e-01 1.27000809e-01 -3.03421825e-01
-5.53756654e-01 2.18153358e-01 -1.49274200e-01 -8.76480818e-01
7.35950936e-03 3.64417225e-01 3.30356836e-01 -2.15718120e-01
3.48255515e-01 1.57099217e-01 -3.31450738e-02 -5.21726489e-01
1.99454814e-01 2.05973133e-01 9.34169531e-01 -5.98717093e-01
9.49543417e-01 7.22640693e-01 1.48902252e-01 -8.53328884e-01
-5.13458014e-01 -9.14088309e-01 -9.37983871e-01 -2.58062929e-01
8.66263270e-01 -1.02819431e+00 -1.08665586e+00 2.95451850e-01
-9.64414001e-01 -3.48523527e-01 -6.64623305e-02 6.41397357e-01
-6.50178313e-01 3.81573379e-01 -7.00385809e-01 -4.33382392e-01
-2.12232858e-01 -8.70397329e-01 1.33856535e+00 1.84407011e-01
-3.61707956e-01 -9.29563999e-01 4.11222816e-01 6.15288280e-02
2.06647113e-01 3.22026849e-01 5.08469701e-01 -3.26440275e-01
-1.13065112e+00 -4.07374710e-01 -3.55367541e-01 -3.32719237e-01
6.07905835e-02 3.69142562e-01 -5.46027958e-01 -5.99848092e-01
-3.83370608e-01 7.59788901e-02 7.52117693e-01 4.88378912e-01
1.03174508e+00 3.42984349e-02 -5.94625473e-01 8.95222485e-01
1.42446995e+00 1.05839178e-01 4.86258268e-01 5.08729279e-01
6.22404397e-01 9.62261930e-02 9.62807417e-01 4.62854177e-01
3.98292512e-01 9.99805629e-01 3.77499610e-01 -1.04667850e-01
-1.46820530e-01 -3.42203319e-01 1.14786856e-01 6.92050338e-01
-3.74943614e-01 2.20985357e-02 -9.17456329e-01 6.16050959e-01
-2.08064365e+00 -1.31250024e+00 -2.22631052e-01 2.52674890e+00
4.30330306e-01 1.79265261e-01 4.08315778e-01 -1.62127480e-01
7.05399156e-01 3.05563986e-01 -6.28607571e-01 8.79627303e-04
7.90205300e-02 2.43022114e-01 7.67734468e-01 5.14855146e-01
-1.29681766e+00 7.76886523e-01 7.11601210e+00 6.10418856e-01
-1.02619636e+00 -4.53799330e-02 3.38661939e-01 -4.49783891e-01
1.21279716e-01 -5.75258993e-02 -7.20128715e-01 3.91094863e-01
6.47467077e-01 -1.32381350e-01 3.88010591e-01 8.37783277e-01
2.96104997e-01 -2.12357804e-01 -1.07019818e+00 1.16306674e+00
4.51123901e-02 -1.77495766e+00 -1.28474444e-01 2.23186731e-01
6.35933816e-01 4.99254346e-01 -1.38189495e-01 -1.11185625e-01
1.97260872e-01 -7.09563613e-01 6.06302619e-01 6.98477805e-01
9.28192556e-01 -7.38844514e-01 4.24506217e-01 -2.87553985e-02
-1.68489754e+00 2.58621007e-01 -2.16319352e-01 1.09225743e-01
1.36472210e-01 1.10603556e-01 -4.24581170e-01 4.43384379e-01
9.00449991e-01 9.32883084e-01 -4.74148691e-01 1.48612416e+00
1.53215691e-01 4.23019677e-01 -5.38735569e-01 2.28993088e-01
2.61215001e-01 -8.50610211e-02 6.15001202e-01 1.25638175e+00
4.40970004e-01 1.40275389e-01 3.46455723e-01 3.59181851e-01
8.53532776e-02 3.62198986e-02 -3.31447899e-01 3.45851690e-01
7.78516352e-01 1.10574603e+00 -8.04457128e-01 -3.97485942e-01
-5.52117109e-01 1.00970852e+00 2.29825214e-01 2.69209355e-01
-9.91934299e-01 -5.86125672e-01 9.69028711e-01 2.49234736e-01
7.91979074e-01 -4.92495596e-01 -6.13573827e-02 -1.03537977e+00
1.68399006e-01 -7.04393506e-01 5.33104360e-01 -8.34511936e-01
-9.11357105e-01 4.66430932e-01 -1.16365008e-01 -1.41018665e+00
-2.66006768e-01 -3.61252815e-01 -5.66516042e-01 8.32204223e-01
-1.39137626e+00 -7.15676129e-01 -6.01077080e-01 7.14531302e-01
3.71975332e-01 1.50534809e-01 6.94871604e-01 4.84882176e-01
-3.55774462e-01 4.21402603e-01 1.45035103e-01 3.09673429e-01
7.05088317e-01 -1.09944987e+00 9.05356407e-01 7.96722770e-01
4.00006413e-01 6.41307235e-01 5.54175258e-01 -7.39028573e-01
-1.57775342e+00 -8.65728319e-01 8.05996895e-01 -8.25623274e-01
7.29344904e-01 -2.03510579e-02 -1.06593895e+00 7.67589569e-01
3.22340801e-03 1.31220415e-01 5.13693988e-01 2.16505006e-01
-3.94984633e-01 -1.03916056e-01 -7.85230696e-01 5.72243094e-01
1.27398062e+00 -5.68444610e-01 -3.97054762e-01 2.92517096e-01
4.00815755e-01 -7.73523927e-01 -1.22476530e+00 2.94345707e-01
8.97907317e-01 -1.09746945e+00 1.23716950e+00 -3.26776952e-01
2.18180194e-01 -4.93111014e-01 -1.41991332e-01 -6.45781934e-01
-4.89434481e-01 -1.00078082e+00 -5.66373229e-01 6.37340009e-01
1.02332637e-01 -2.23137036e-01 9.69384134e-01 4.88934755e-01
1.12260208e-01 -1.13054919e+00 -8.29575300e-01 -7.84161687e-01
-5.09787738e-01 -5.98085761e-01 6.82966828e-01 6.93774760e-01
-5.82718067e-02 5.42127043e-02 -1.39874309e-01 4.82556403e-01
6.66853905e-01 3.59373480e-01 1.05606091e+00 -1.03593671e+00
-2.39756346e-01 -5.98564923e-01 -6.78542912e-01 -1.79274464e+00
-2.97420681e-01 -4.54487771e-01 -6.61587268e-02 -1.39152777e+00
9.23900530e-02 -6.61997259e-01 -4.34577048e-01 2.49998584e-01
-3.10927063e-01 4.48894709e-01 9.96107087e-02 6.30169153e-01
-9.56182778e-01 1.87892333e-01 9.07890677e-01 3.46336246e-01
-1.48299620e-01 6.77887425e-02 -2.17875570e-01 8.12774599e-01
3.77701163e-01 -4.38595086e-01 -1.57384336e-01 -5.33320069e-01
1.22964859e-01 2.52423674e-01 3.88784617e-01 -1.31691051e+00
4.09830481e-01 -9.55532417e-02 4.73345548e-01 -9.68523443e-01
7.07823217e-01 -7.53600895e-01 2.73630798e-01 3.99392068e-01
-3.41950446e-01 6.60652876e-01 1.98079437e-01 7.09931016e-01
-1.27911165e-01 1.13969982e-01 6.16122007e-01 1.82557240e-01
-8.16383183e-01 7.99627900e-01 -5.25827371e-02 1.04924403e-01
1.07675529e+00 -4.37985480e-01 -2.67525375e-01 -5.45181274e-01
-4.51423913e-01 3.17443699e-01 1.00447679e+00 5.90901077e-01
5.77805161e-01 -1.35412061e+00 -5.26968956e-01 9.41159204e-02
7.65651688e-02 -1.85034737e-01 3.94901959e-03 1.06182039e+00
-8.58079612e-01 3.55413169e-01 5.82148321e-02 -8.89538407e-01
-1.48224926e+00 5.83028018e-01 2.53387004e-01 -4.08063233e-02
-9.42553163e-01 8.88205945e-01 1.38173895e-02 2.64560789e-01
4.52162892e-01 -8.79423544e-02 1.25505343e-01 -1.47666872e-01
8.24146986e-01 6.09785497e-01 -3.12512293e-02 -9.70377088e-01
-6.57692552e-01 1.01465678e+00 1.29705062e-02 -5.21475151e-02
1.13142824e+00 -1.50459662e-01 1.29709512e-01 4.58861589e-01
1.29295480e+00 2.27051094e-01 -1.66391432e+00 -3.64170462e-01
1.71375990e-01 -1.03997898e+00 -3.24579217e-02 -3.53536963e-01
-9.19301748e-01 4.31096226e-01 5.95521331e-01 4.63525876e-02
9.98184204e-01 -3.31917182e-02 1.04310560e+00 3.48874062e-01
5.57599068e-01 -8.53153467e-01 -3.05742323e-02 4.09498960e-01
5.89416265e-01 -1.07238138e+00 2.09498957e-01 -3.02699476e-01
-4.99389529e-01 1.10093415e+00 4.00386721e-01 -3.58725011e-01
7.61566222e-01 3.76808345e-01 9.92256850e-02 -3.37275714e-01
-1.00308371e+00 -8.74152631e-02 3.77278894e-01 4.82955843e-01
3.44712377e-01 -3.36356282e-01 9.64135826e-02 -2.67610520e-01
-2.33532079e-02 1.31248087e-01 8.44463557e-02 9.79394257e-01
-5.28061867e-01 -8.37200165e-01 -3.09748620e-01 4.29113954e-01
-4.66788083e-01 -1.72739681e-02 1.57075822e-01 5.63621640e-01
-1.52696118e-01 7.58728266e-01 6.53690636e-01 -4.22698617e-01
3.87505680e-01 -2.85853982e-01 3.64327550e-01 -3.32830042e-01
-5.57141542e-01 1.37284741e-01 -6.17633313e-02 -1.22745442e+00
-4.66315031e-01 -9.36249018e-01 -1.27347291e+00 -6.86137080e-01
-1.55641660e-01 1.62061319e-01 4.61117178e-01 4.84995455e-01
7.94609845e-01 1.72205910e-01 4.22670335e-01 -1.06407559e+00
-1.38506353e-01 -4.59734529e-01 -2.41218567e-01 4.11875695e-01
4.02598619e-01 -7.03573048e-01 -1.38819749e-02 1.68882847e-01] | [8.062812805175781, -1.3626630306243896] |
09c34f1c-ef95-4e51-adb2-7f23c4055101 | safe-exploration-of-state-and-action-spaces | 1402.0560 | null | http://arxiv.org/abs/1402.0560v1 | http://arxiv.org/pdf/1402.0560v1.pdf | Safe Exploration of State and Action Spaces in Reinforcement Learning | In this paper, we consider the important problem of safe exploration in
reinforcement learning. While reinforcement learning is well-suited to domains
with complex transition dynamics and high-dimensional state-action spaces, an
additional challenge is posed by the need for safe and efficient exploration.
Traditional exploration techniques are not particularly useful for solving
dangerous tasks, where the trial and error process may lead to the selection of
actions whose execution in some states may result in damage to the learning
system (or any other system). Consequently, when an agent begins an interaction
with a dangerous and high-dimensional state-action space, an important question
arises; namely, that of how to avoid (or at least minimize) damage caused by
the exploration of the state-action space. We introduce the PI-SRL algorithm
which safely improves suboptimal albeit robust behaviors for continuous state
and action control tasks and which efficiently learns from the experience
gained from the environment. We evaluate the proposed method in four complex
tasks: automatic car parking, pole-balancing, helicopter hovering, and business
management. | ['Javier Garcia', 'Fernando Fernandez'] | 2014-02-04 | null | null | null | null | ['safe-exploration'] | ['robots'] | [ 2.62517810e-01 1.88255653e-01 -1.20843984e-01 1.85211912e-01
-2.92352498e-01 -3.24258566e-01 5.12962520e-01 4.19949442e-01
-6.45525217e-01 1.19495356e+00 -4.13730323e-01 -4.65293348e-01
-5.33456624e-01 -6.13121152e-01 -5.31777978e-01 -8.26307535e-01
-5.70871770e-01 4.56493020e-01 1.87997282e-01 -3.87671471e-01
4.14327383e-01 5.32328904e-01 -1.68591094e+00 -5.73552310e-01
1.04596615e+00 8.57499123e-01 3.32567066e-01 4.12697524e-01
2.42570639e-01 7.12977231e-01 -5.37657142e-01 3.60180765e-01
2.82387465e-01 -5.40434957e-01 -8.35076630e-01 2.01516375e-01
-3.95750940e-01 -1.44193873e-01 2.00820088e-01 1.20010018e+00
1.85535088e-01 7.47989655e-01 3.52438062e-01 -1.44457817e+00
3.01270187e-01 4.65745896e-01 -4.13173139e-01 1.10255614e-01
2.03075394e-01 6.56007886e-01 4.51956838e-01 -2.58114219e-01
3.25213611e-01 1.22115254e+00 -1.96526968e-03 5.18714488e-01
-1.23856068e+00 -4.21820611e-01 5.29643178e-01 2.70316035e-01
-1.17466068e+00 -1.88518405e-01 4.86617327e-01 -3.88019592e-01
8.97218406e-01 -9.13161878e-03 8.16155910e-01 8.37581635e-01
5.77153921e-01 6.57932520e-01 1.19188333e+00 -2.32570797e-01
9.00256693e-01 4.58951034e-02 -2.52214283e-01 5.49782813e-01
1.84742153e-01 7.07070172e-01 -2.95989126e-01 -2.01989010e-01
3.42736632e-01 -1.03811778e-01 -1.36929363e-01 -8.48958611e-01
-1.03925979e+00 6.17600739e-01 2.29306549e-01 5.81157096e-02
-7.28250563e-01 8.46830457e-02 3.80521715e-01 6.42777860e-01
-3.01529050e-01 8.97578180e-01 -1.69610709e-01 -6.97126329e-01
-5.09860873e-01 5.26305020e-01 7.92814553e-01 4.05142546e-01
6.44189060e-01 4.27813411e-01 1.75741166e-01 2.65913099e-01
2.98664756e-02 1.93198800e-01 2.69945145e-01 -7.53335357e-01
5.15747130e-01 5.99132001e-01 7.32906520e-01 -6.42250180e-01
-4.05407071e-01 -3.70626360e-01 -4.87616956e-01 9.74866807e-01
4.27859604e-01 -4.51362848e-01 -7.70984828e-01 1.67692554e+00
7.90441632e-01 -7.45710060e-02 2.23452330e-01 8.68971884e-01
-3.05310190e-01 4.99784142e-01 1.87181488e-01 -5.51918745e-01
8.36829364e-01 -5.67543805e-01 -9.12790596e-01 -5.91325045e-01
3.96197796e-01 -2.24332094e-01 1.10284185e+00 4.24713701e-01
-1.28359580e+00 -3.51776093e-01 -1.25213337e+00 7.11991012e-01
-3.30244899e-01 -3.63774657e-01 3.46925497e-01 1.12677835e-01
-5.29653788e-01 8.79640937e-01 -9.71705854e-01 -3.15051556e-01
5.12955599e-02 5.57021141e-01 -1.03432499e-01 2.88470149e-01
-1.31777549e+00 1.37688053e+00 6.94881558e-01 3.87596011e-01
-1.33903694e+00 -1.39512807e-01 -6.55649602e-01 1.34568617e-01
1.22694790e+00 -1.93020388e-01 1.35185301e+00 -5.95975816e-01
-1.73228276e+00 2.20500352e-03 3.55571687e-01 -5.29930055e-01
6.90187871e-01 -3.45461041e-01 -1.88417763e-01 -2.37192899e-01
-1.29072711e-01 1.79137141e-01 1.03447807e+00 -1.05515051e+00
-7.97521651e-01 -4.68473375e-01 2.57485569e-01 7.45736361e-01
-1.83746099e-01 -4.77465838e-01 3.41051728e-01 -1.59647048e-01
-3.44775245e-02 -1.17514741e+00 -6.15414023e-01 -3.11893493e-01
-2.48317212e-01 -1.52251005e-01 7.91760981e-01 -1.35641351e-01
1.19769442e+00 -2.07359743e+00 4.22436565e-01 3.82266045e-01
-2.83603191e-01 3.35211992e-01 1.00973733e-01 7.49054611e-01
8.20986107e-02 -1.99323729e-01 -2.12628782e-01 1.18564516e-01
-4.83116284e-02 4.28125024e-01 -3.62954855e-01 3.32881570e-01
6.57015219e-02 4.71758574e-01 -1.00675786e+00 -3.11834246e-01
2.33301386e-01 -4.37811539e-02 -3.27414930e-01 4.10365611e-01
-3.04482847e-01 6.59906030e-01 -8.94791067e-01 4.29212451e-01
1.56052142e-01 1.90512687e-01 2.35419944e-01 5.03601134e-01
-4.31446731e-01 2.14743376e-01 -1.54007781e+00 1.07624221e+00
-4.87630278e-01 7.85066038e-02 3.31012607e-01 -1.10801983e+00
8.62982869e-01 1.89419061e-01 5.23449838e-01 -8.14282000e-01
2.25641578e-01 1.63586169e-01 2.12604612e-01 -4.93965298e-01
5.38661659e-01 -1.81531593e-01 -1.93843931e-01 6.68384910e-01
-3.21111172e-01 -4.16765779e-01 2.64996618e-01 -2.17632100e-01
9.87325132e-01 2.61610270e-01 5.64102232e-01 -3.98520738e-01
6.18542433e-01 2.80443192e-01 6.11353338e-01 7.26247430e-01
-4.79194492e-01 -5.48153281e-01 8.55355799e-01 -5.12188315e-01
-6.05061591e-01 -8.17883551e-01 1.82634711e-01 8.21736991e-01
4.50012863e-01 2.23688385e-03 -6.03734732e-01 -6.37035906e-01
2.43688505e-02 9.98596489e-01 -5.66372395e-01 -6.20823741e-01
-6.54107034e-01 -3.30551207e-01 -1.14036575e-01 1.80921838e-01
4.96837467e-01 -1.41069043e+00 -1.59256434e+00 5.89577258e-01
1.06033191e-01 -6.06953323e-01 -2.69944966e-01 6.33986831e-01
-1.06758738e+00 -1.10015130e+00 -2.14396909e-01 -4.53431517e-01
6.79178655e-01 -1.68000042e-01 6.90611064e-01 -1.32044420e-01
-2.49812514e-01 2.43532643e-01 -7.35708773e-02 -3.28066707e-01
-5.53075254e-01 -4.34549376e-02 3.42814833e-01 -2.19892994e-01
1.07398272e-01 -3.30229491e-01 -3.89802694e-01 5.46079814e-01
-7.53012836e-01 -7.46430531e-02 3.60663235e-01 1.06719565e+00
5.14457583e-01 7.26415694e-01 7.77208507e-01 -4.83264148e-01
1.00124109e+00 -2.80660987e-01 -1.02629673e+00 3.93339545e-01
-8.19007695e-01 3.72402430e-01 7.67774820e-01 -6.46832585e-01
-1.07822609e+00 2.97104925e-01 2.57490784e-01 -5.27211688e-02
-1.37504265e-01 4.80575800e-01 -1.81484833e-01 1.01991743e-01
5.97734749e-01 2.94417650e-01 3.62527192e-01 -1.49668410e-01
1.29975885e-01 3.34811747e-01 2.57300526e-01 -7.80380309e-01
8.11415136e-01 3.99147309e-02 3.78426343e-01 -6.51606262e-01
-2.12617919e-01 -1.46535441e-01 -5.02445817e-01 -3.55508536e-01
5.22314131e-01 -4.60614473e-01 -1.05425370e+00 3.09756368e-01
-4.73153412e-01 -5.46737075e-01 -5.58694363e-01 1.84882134e-01
-1.03038478e+00 1.79540783e-01 3.06119286e-02 -1.31822193e+00
1.20915845e-02 -1.27402091e+00 4.87600058e-01 4.55695421e-01
-3.18714917e-01 -5.57182491e-01 9.30897221e-02 -9.05351490e-02
3.13594431e-01 4.34143573e-01 9.98613834e-01 -2.99863487e-01
-5.49057245e-01 -1.09339096e-02 5.11826634e-01 -7.41942525e-02
1.93659604e-01 -3.63111526e-01 -4.12483692e-01 -6.38629377e-01
2.01195091e-01 -6.66398048e-01 2.86947995e-01 1.14452630e-01
8.30521524e-01 -5.55321157e-01 -2.16411427e-01 1.24056416e-03
1.26850069e+00 9.13078845e-01 3.46614957e-01 6.67900920e-01
4.79647554e-02 9.39205706e-01 1.42181540e+00 8.29960048e-01
7.80267045e-02 6.64175451e-01 8.27107370e-01 2.26706475e-01
7.15682209e-01 -4.35300201e-01 4.37672555e-01 -7.33840391e-02
1.96764320e-01 -5.04519567e-02 -8.60812366e-01 6.16175950e-01
-2.14197278e+00 -7.80230045e-01 4.96955186e-01 2.58550096e+00
8.34882259e-01 3.98783773e-01 2.78912038e-01 2.40887105e-01
4.84117359e-01 -7.23575428e-02 -1.18494058e+00 -6.47526622e-01
4.67536390e-01 -2.28515401e-01 2.18269780e-01 5.80152154e-01
-8.06473434e-01 7.25187600e-01 5.78448105e+00 5.41722059e-01
-1.03194058e+00 -4.82921302e-01 4.45765197e-01 -1.49017066e-01
1.81672558e-01 7.08991215e-02 -6.04278326e-01 4.61062819e-01
8.65801334e-01 -2.58473009e-01 8.17763686e-01 1.04576743e+00
4.65645015e-01 -7.69733727e-01 -1.16917217e+00 5.23633480e-01
-5.92455447e-01 -6.91942096e-01 -6.45859540e-01 -2.14729588e-02
5.47860146e-01 -5.23888171e-01 1.90484077e-01 5.20726740e-01
3.96870047e-01 -8.67942035e-01 5.76330245e-01 1.85308412e-01
3.82434785e-01 -1.28991234e+00 3.74590963e-01 9.10481215e-01
-8.85872304e-01 -6.89122200e-01 -4.62304801e-02 -3.10071290e-01
2.52669811e-01 7.00051486e-02 -9.68442023e-01 2.71774679e-02
4.56340849e-01 1.34244293e-01 -1.46580204e-01 1.05562711e+00
-3.86938006e-01 5.46229482e-02 -3.03238869e-01 -6.22402072e-01
4.99522895e-01 -3.75383705e-01 6.92012668e-01 3.65387350e-01
2.40239277e-01 4.28034440e-02 3.94542158e-01 8.71464014e-01
7.65699863e-01 -2.23304018e-01 -9.47255075e-01 -1.68011785e-01
3.52404475e-01 9.60355341e-01 -8.07172596e-01 -1.36313304e-01
2.58084476e-01 5.14354110e-01 2.68631876e-01 4.50867921e-01
-7.31660664e-01 -4.62168396e-01 8.22800338e-01 8.37135464e-02
1.32382229e-01 -4.53867108e-01 -1.46925598e-01 -5.96858025e-01
-8.76706392e-02 -1.04586971e+00 3.53998303e-01 -2.41575167e-01
-5.96319377e-01 4.67549026e-01 1.69149891e-01 -1.26716399e+00
-7.58422494e-01 -1.09772786e-01 -5.72508633e-01 6.61978066e-01
-1.30954635e+00 -2.10877255e-01 1.15733340e-01 6.39415383e-01
5.99792838e-01 -1.83626503e-01 5.67671120e-01 -1.53385818e-01
-6.21188998e-01 3.16970982e-02 1.96329772e-01 -6.76104665e-01
1.67773321e-01 -1.25274527e+00 9.14338455e-02 7.61006057e-01
-4.65063453e-01 2.99452662e-01 1.16216433e+00 -8.06892693e-01
-1.64444160e+00 -7.58513510e-01 5.05781293e-01 6.58698454e-02
7.11991966e-01 -2.03886822e-01 -8.53075206e-01 2.61695832e-01
-5.05160391e-02 -4.08001721e-01 -1.25104278e-01 -2.03491678e-03
5.35707414e-01 -1.03478402e-01 -1.26817822e+00 9.69317675e-01
7.55372822e-01 -2.41223842e-01 -5.52124798e-01 4.25610431e-02
2.70031631e-01 -3.88537705e-01 -3.54706258e-01 2.85216838e-01
2.80343682e-01 -7.14644134e-01 6.74043357e-01 -8.12435031e-01
-2.82655191e-02 -4.55646932e-01 3.23132694e-01 -1.68180823e+00
-1.14242010e-01 -9.25148189e-01 -2.87277371e-01 5.74780047e-01
1.55338556e-01 -4.88924950e-01 8.29487443e-01 7.34641552e-01
1.59586862e-01 -9.80038166e-01 -1.30105054e+00 -9.27855492e-01
-4.77529392e-02 4.20291424e-02 5.52851081e-01 4.32986319e-01
4.85496223e-01 9.39558968e-02 -5.33768713e-01 1.00627013e-01
6.88598216e-01 3.47554862e-01 6.77366138e-01 -8.56082022e-01
-2.68353403e-01 -2.50464380e-01 4.72498238e-02 -8.21316481e-01
1.00446165e-01 -8.19110963e-03 7.57097661e-01 -1.26633620e+00
-3.99750918e-01 -5.42800963e-01 -3.26165646e-01 4.62506592e-01
-7.54108727e-02 -7.50021696e-01 2.49413013e-01 3.38180363e-02
-6.72711134e-01 9.41064060e-01 1.43801820e+00 6.59045652e-02
-6.88632488e-01 4.55024838e-01 -2.35919356e-01 5.28950691e-01
9.19800222e-01 -3.02914143e-01 -8.84612918e-01 8.54535475e-02
2.41656512e-01 7.45958865e-01 7.00229853e-02 -1.08131313e+00
1.87980965e-01 -8.89886260e-01 -1.76507503e-01 -4.78016764e-01
4.86900866e-01 -1.15964627e+00 2.59960800e-01 1.21595752e+00
-4.41084951e-01 3.26893955e-01 1.84662089e-01 7.92564929e-01
-1.79257870e-01 -3.21468830e-01 9.52350080e-01 -1.83772042e-01
-8.07930291e-01 6.53121993e-02 -9.79252994e-01 1.56327039e-02
1.52093887e+00 -2.96284676e-01 -1.35476455e-01 -3.83777678e-01
-8.87713730e-01 8.28346729e-01 2.38874167e-01 5.70083916e-01
5.59389412e-01 -9.79246140e-01 -4.46484536e-02 4.13560480e-01
-6.85540661e-02 -6.66408092e-02 2.72045165e-01 6.67812049e-01
-4.31801341e-02 2.44750947e-01 -6.77507520e-01 -3.38562965e-01
-1.05579066e+00 6.76203966e-01 4.35754240e-01 -4.26103324e-01
-5.32285631e-01 4.00271654e-01 -4.88038320e-04 -1.77976280e-01
5.64733624e-01 -3.93147290e-01 -3.53814304e-01 8.09233338e-02
3.05237561e-01 5.92146575e-01 -7.71241412e-02 -1.17143460e-01
-2.66571730e-01 1.61917225e-01 9.61972699e-02 -2.44394526e-01
1.16146445e+00 -1.70982018e-01 3.54182810e-01 4.83915210e-01
4.73085195e-01 -8.39493752e-01 -1.73269033e+00 -5.87325245e-02
2.44626477e-01 -4.94291127e-01 7.00534135e-02 -8.25380445e-01
-4.45947587e-01 7.85605907e-01 6.82382584e-01 4.38354939e-01
1.08122718e+00 -5.67359030e-01 4.46288198e-01 7.41349399e-01
8.26699615e-01 -1.72715771e+00 3.30874056e-01 6.87841356e-01
8.22137058e-01 -1.07635069e+00 -7.46829659e-02 2.58654326e-01
-1.01683307e+00 9.20408726e-01 1.06848645e+00 9.98315290e-02
5.44881344e-01 1.83759749e-01 -1.35025978e-01 -8.22542980e-02
-9.90945518e-01 -1.18322067e-01 -1.50759652e-01 5.59301317e-01
-3.25932771e-01 8.22895020e-02 -2.92131662e-01 -1.09751210e-01
3.48429084e-02 -9.43907276e-02 5.98343849e-01 1.33199251e+00
-9.12913382e-01 -1.03084469e+00 -5.16000628e-01 2.55987763e-01
-4.20272956e-03 6.40387952e-01 -1.71515986e-01 9.01109695e-01
4.59882729e-02 7.67494082e-01 -4.55826186e-02 3.05134561e-02
4.43689972e-01 1.06756590e-01 2.87705451e-01 -6.10812306e-01
-4.17106777e-01 -2.29812730e-02 1.64276809e-01 -9.11813319e-01
1.39144016e-03 -8.94616008e-01 -1.44806695e+00 2.43982207e-02
-2.42008701e-01 3.86375785e-01 6.91519201e-01 8.99858296e-01
1.31901145e-01 5.12266815e-01 7.78477609e-01 -4.57620472e-01
-1.37635827e+00 -5.82602084e-01 -7.61182189e-01 6.79985136e-02
7.05170572e-01 -1.15854871e+00 -2.51637876e-01 -5.90657175e-01] | [4.537845134735107, 2.008737802505493] |
ab406c8d-1418-4111-b22c-d24f693c9a20 | improved-tree-search-for-automatic-program | 2303.07166 | null | https://arxiv.org/abs/2303.07166v1 | https://arxiv.org/pdf/2303.07166v1.pdf | Improved Tree Search for Automatic Program Synthesis | In the task of automatic program synthesis, one obtains pairs of matching inputs and outputs and generates a computer program, in a particular domain-specific language (DSL), which given each sample input returns the matching output. A key element is being able to perform an efficient search in the space of valid programs. Here, we suggest a variant of MCTS that leads to state of the art results on two vastly different DSLs. The exploration method we propose includes multiple contributions: a modified visit count, a preprocessing procedure for the training dataset, and encoding the part of the program that was already executed. | ['Lior Wolf', 'Aran Carmon'] | 2023-03-13 | null | null | null | null | ['program-synthesis'] | ['computer-code'] | [ 5.71008801e-01 1.70284927e-01 -7.21410215e-01 -2.58525938e-01
-9.17219281e-01 -8.42899203e-01 5.59643388e-01 2.71255821e-01
-9.31869820e-02 7.25321412e-01 -1.20092764e-01 -8.91011298e-01
2.54133970e-01 -1.34572101e+00 -9.70867991e-01 -1.00720078e-01
-2.09359098e-02 2.72650093e-01 5.89190483e-01 3.90308239e-02
5.79934835e-01 3.94652873e-01 -2.04267979e+00 4.94197309e-01
1.05346429e+00 8.24367583e-01 2.56259054e-01 1.01659310e+00
-4.87325311e-01 7.09839821e-01 -5.82800329e-01 -5.31472445e-01
9.21045095e-02 -6.66511893e-01 -1.14812815e+00 -6.12864085e-02
7.50181600e-02 1.55431563e-02 1.03845239e-01 1.10640693e+00
-7.14463815e-02 -1.97770260e-02 2.99310535e-01 -1.43785679e+00
-1.07468456e-01 1.05324113e+00 -1.01617552e-01 -7.19488040e-02
6.74240828e-01 6.69768274e-01 1.18591332e+00 -4.88443255e-01
8.09234738e-01 1.02388942e+00 1.54377416e-01 7.74596632e-01
-1.77015650e+00 -2.47422233e-01 -3.43514681e-01 -2.54482687e-01
-1.12861466e+00 -4.42510456e-01 4.80069041e-01 -4.82916206e-01
1.50830996e+00 5.11197865e-01 8.92454267e-01 7.91923761e-01
2.29514852e-01 9.13756609e-01 1.15531278e+00 -7.50507772e-01
6.21171057e-01 3.93108279e-01 9.31018293e-02 8.33688259e-01
1.67136326e-01 4.06262279e-01 -3.80642653e-01 -7.54562259e-01
3.62003058e-01 -4.91647273e-01 7.88354576e-02 -7.89459348e-01
-1.13197494e+00 9.01828051e-01 -1.05517037e-01 3.84415865e-01
5.25861345e-02 5.59672058e-01 6.03239775e-01 5.81596792e-01
-1.87967539e-01 9.15625632e-01 -5.11192858e-01 -4.55283701e-01
-1.20898855e+00 5.57867765e-01 1.18053210e+00 1.12866867e+00
8.99843931e-01 8.01123977e-02 -3.50661784e-01 -7.08527118e-03
8.55497420e-02 3.11746240e-01 2.01119483e-01 -8.42352688e-01
6.41333461e-01 1.07312930e+00 -8.02196935e-02 -3.35861772e-01
1.52664572e-01 -2.09296644e-01 -9.21165422e-02 3.87537718e-01
1.89987242e-01 -2.08691761e-01 -3.45476896e-01 1.83167470e+00
1.97039261e-01 -9.14318860e-02 3.30478936e-01 1.64356753e-01
4.46389228e-01 6.80304825e-01 -1.19608708e-01 -2.39774808e-01
9.55458701e-01 -1.06624770e+00 -2.65221864e-01 -5.10975122e-01
1.04537189e+00 -3.65749687e-01 1.36116159e+00 4.50993985e-01
-1.45073891e+00 -4.16138351e-01 -1.37141585e+00 1.63196221e-01
-3.23508978e-01 1.67947114e-01 8.21358681e-01 8.18455517e-01
-1.34460878e+00 7.00216413e-01 -5.95847785e-01 6.67672008e-02
2.40378916e-01 6.02395177e-01 7.62257352e-03 -1.07199803e-01
-7.27063954e-01 6.99159443e-01 7.90497541e-01 -4.60267842e-01
-1.01670659e+00 -6.01424575e-01 -1.17550540e+00 7.15076849e-02
6.25397980e-01 -7.17551112e-01 1.65062261e+00 -1.29270685e+00
-1.49465370e+00 1.06127250e+00 -4.40963447e-01 -5.72687268e-01
4.96672779e-01 5.38460612e-01 -2.86244452e-01 -2.84685016e-01
4.61272486e-02 4.99274015e-01 7.34175026e-01 -1.04140115e+00
-7.92764127e-01 -1.51154086e-01 2.64647633e-01 -4.15562451e-01
8.63853171e-02 2.43908867e-01 -3.32652211e-01 -3.54966849e-01
-5.04806519e-01 -8.95178795e-01 -4.96189713e-01 -2.35644057e-01
-4.03713942e-01 -2.98206389e-01 2.13445678e-01 -3.87400746e-01
1.66441905e+00 -2.28871107e+00 5.63686490e-01 5.47717810e-01
-8.36283434e-03 -4.97853346e-02 -9.13973376e-02 5.16823053e-01
1.37019783e-01 2.75246173e-01 -6.78285778e-01 -7.93632716e-02
4.69079673e-01 5.32916412e-02 -4.44595337e-01 1.08633801e-01
5.11312783e-01 1.22077703e+00 -1.03166819e+00 -6.65738285e-01
-1.16575785e-01 -4.86326426e-01 -9.30487573e-01 5.00794709e-01
-1.14775944e+00 -3.11850440e-02 -4.76348579e-01 5.23245811e-01
3.85828912e-01 -2.03441083e-01 3.68916035e-01 3.50526571e-01
-4.35561031e-01 5.42048931e-01 -1.20997977e+00 1.98931015e+00
-6.10208392e-01 6.06143773e-01 -2.91118294e-01 -6.46875381e-01
1.01002097e+00 2.22782537e-01 -4.35839929e-02 -4.78114486e-01
-1.56647004e-02 7.21095622e-01 -5.88791780e-02 -5.50571620e-01
6.15042090e-01 1.08010270e-01 -6.26560867e-01 8.77238214e-01
-2.21813340e-02 -6.74778402e-01 5.84462702e-01 1.28881872e-01
1.33208680e+00 4.02220875e-01 7.06797481e-01 -4.19622838e-01
7.92684972e-01 5.81885993e-01 2.77170867e-01 7.67799973e-01
3.42955917e-01 7.00286776e-02 1.19871855e+00 -3.98529768e-01
-1.21034682e+00 -6.38244033e-01 2.26371810e-01 9.05426264e-01
-2.64984310e-01 -6.51938200e-01 -9.29058313e-01 -8.80511463e-01
-9.98160318e-02 1.23821592e+00 -5.72340310e-01 -3.05836141e-01
-1.00009668e+00 -1.46935508e-01 7.61659443e-01 4.30432826e-01
1.36172816e-01 -1.40219700e+00 -1.40528882e+00 2.80254453e-01
2.49142468e-01 -7.29585350e-01 -4.67290998e-01 6.01221621e-01
-1.18277240e+00 -1.02389359e+00 -6.59586042e-02 -1.04431379e+00
6.87750161e-01 -4.55295444e-01 1.45139885e+00 4.66647446e-01
-1.95997477e-01 -5.07556386e-02 -3.56355719e-02 -2.59151459e-02
-1.15181339e+00 2.87482917e-01 -6.68622315e-01 -4.61781979e-01
2.46332526e-01 -3.06031048e-01 1.63195968e-01 -1.93668336e-01
-1.09329545e+00 2.35167265e-01 5.11541426e-01 8.09551001e-01
5.54135323e-01 1.86102197e-01 1.48824215e-01 -1.21145570e+00
4.96839583e-01 -4.02071744e-01 -1.11861861e+00 4.96310323e-01
-6.81388319e-01 8.31845164e-01 1.12016082e+00 -3.24216902e-01
-7.43511975e-01 4.31315303e-01 -1.35924786e-01 1.38296500e-01
-9.51225907e-02 5.21084726e-01 -3.46835822e-01 -3.30852345e-02
8.86795402e-01 6.00538492e-01 -1.03505142e-01 -1.50449514e-01
3.76612514e-01 2.70308942e-01 5.71045637e-01 -1.08147931e+00
7.61442780e-01 -1.78663760e-01 1.38209820e-01 -7.26166219e-02
-1.77864954e-02 1.59025609e-01 -5.80228984e-01 2.40428478e-01
3.17158222e-01 -1.98796153e-01 -4.49421495e-01 1.66430131e-01
-1.16891897e+00 -9.37150896e-01 -7.47233927e-01 -2.28401408e-01
-8.48524988e-01 1.15453400e-01 -3.09310891e-02 -7.73857474e-01
-2.52457470e-01 -1.67856383e+00 9.28913653e-01 4.31251228e-02
-5.84216714e-01 -6.55406117e-01 3.85335267e-01 -4.82471049e-01
2.39717066e-01 4.29438293e-01 1.66995120e+00 -7.88757443e-01
-9.48168039e-01 -1.02111429e-01 2.01739475e-01 2.46369839e-01
-8.32320452e-02 4.72868353e-01 -6.02533340e-01 -9.01154205e-02
-1.36929587e-01 -2.97228068e-01 2.28236496e-01 9.55125615e-02
1.18927300e+00 -5.00906110e-01 -5.80984473e-01 5.86080790e-01
1.63597298e+00 3.41385335e-01 4.70966935e-01 3.52763951e-01
-3.70857283e-03 4.47438449e-01 6.24783874e-01 3.65038991e-01
6.46051243e-02 6.25479102e-01 2.47601300e-01 2.40023106e-01
1.46967828e-01 -4.86809194e-01 5.28108180e-01 2.77805358e-01
5.11528730e-01 1.60966609e-02 -1.02529013e+00 8.66475165e-01
-1.64758158e+00 -6.73894763e-01 1.32381350e-01 2.34158349e+00
1.19712317e+00 3.44365507e-01 4.84322101e-01 3.69806677e-01
3.29391599e-01 3.70603241e-02 -5.59146941e-01 -1.02268898e+00
1.59828246e-01 1.01188135e+00 4.96310294e-01 5.69484890e-01
-6.27033889e-01 8.17162991e-01 7.12384176e+00 7.28311181e-01
-9.84141350e-01 -1.65523812e-01 3.43539357e-01 6.45902902e-02
-9.35391247e-01 3.46730679e-01 -6.18639410e-01 2.91006684e-01
1.39767039e+00 -8.75244856e-01 8.10816944e-01 1.21372211e+00
-2.92185485e-01 -2.75515467e-01 -1.76322532e+00 3.13426465e-01
-2.79185355e-01 -1.53317082e+00 -9.83501151e-02 -2.16656163e-01
8.39479625e-01 -4.73756313e-01 5.21299653e-02 4.27270830e-01
5.21201789e-01 -1.03709710e+00 1.12548733e+00 2.06261978e-01
8.94105971e-01 -1.07438564e+00 2.83968002e-01 4.87842739e-01
-1.22654021e+00 -1.59716263e-01 1.32668316e-01 9.57049131e-02
-3.90732318e-01 2.69464791e-01 -9.01149690e-01 3.36333781e-01
2.13749930e-01 4.57468361e-01 -6.97103858e-01 1.01200128e+00
-5.21172881e-01 3.95362973e-01 4.73173754e-03 -5.72370291e-01
3.07832152e-01 1.14033431e-01 4.89540070e-01 1.37453651e+00
4.67480719e-01 -8.24056491e-02 6.30706027e-02 1.59636462e+00
1.20311538e-02 3.52425203e-02 -8.92651260e-01 -4.07280475e-01
5.32407820e-01 7.97195196e-01 -5.82002342e-01 -6.88477695e-01
-4.66732204e-01 6.99122012e-01 2.87487060e-01 9.21220407e-02
-6.97285175e-01 -5.14605522e-01 3.30588490e-01 -1.25348970e-01
2.78471857e-01 -2.65841365e-01 -6.03662610e-01 -8.95603538e-01
3.88561338e-01 -1.48721433e+00 3.68944854e-01 -2.63347954e-01
-2.66128063e-01 7.05737531e-01 8.77685100e-02 -8.23567212e-01
-8.03653359e-01 -3.42919320e-01 -7.68922150e-01 1.07504761e+00
-1.30080748e+00 -6.63865626e-01 6.21810369e-02 3.05755734e-01
5.63419640e-01 -1.06832303e-01 8.28993022e-01 2.25671250e-02
-4.39545423e-01 7.34033406e-01 -2.87271142e-01 -3.83498609e-01
-7.98650682e-02 -1.38836813e+00 1.08785915e+00 1.20401359e+00
-2.33740985e-01 7.39799559e-01 8.05669129e-01 -5.48447132e-01
-2.24608755e+00 -9.87154782e-01 9.96633410e-01 -3.11114609e-01
7.06364512e-01 -3.41540456e-01 -6.85865581e-01 8.42937052e-01
7.94127700e-04 -1.88178733e-01 2.67578125e-01 -5.08749068e-01
-1.55072302e-01 1.72826082e-01 -1.24996924e+00 6.79376245e-01
9.52867210e-01 -6.43834829e-01 -5.07504284e-01 6.90643955e-03
9.65048254e-01 -8.06474507e-01 -6.26454771e-01 7.69363195e-02
3.73320967e-01 -8.52604091e-01 6.85686707e-01 -6.85301542e-01
8.15270782e-01 -3.57635349e-01 -2.45374441e-01 -1.14559269e+00
2.45957702e-01 -9.72217321e-01 -5.60620010e-01 1.04750597e+00
7.23573446e-01 -6.10335350e-01 8.22913587e-01 9.43573117e-01
-1.03097722e-01 -9.10507560e-01 -7.37654030e-01 -7.76296496e-01
7.60802850e-02 -6.11676395e-01 1.22048593e+00 2.50055224e-01
3.20917457e-01 5.25316298e-02 -4.42302562e-02 -1.16507374e-01
4.21330661e-01 8.63270879e-01 8.60025883e-01 -9.11933184e-01
-7.81886160e-01 -6.58432603e-01 1.68108493e-01 -9.06353593e-01
4.19368684e-01 -1.19158041e+00 8.96410421e-02 -1.04060686e+00
1.12150893e-01 -5.23493946e-01 2.99190730e-01 4.75371420e-01
9.77712646e-02 -5.36427200e-01 -8.93098190e-02 -1.23216689e-01
-2.58193880e-01 6.88357837e-03 8.81124020e-01 -2.85430342e-01
-5.09970367e-01 3.10007513e-01 -5.84201455e-01 2.58163452e-01
5.55173695e-01 -8.34052205e-01 -2.16824710e-01 -2.34135360e-01
7.09412873e-01 5.69990873e-01 3.17138046e-01 -8.59151781e-01
2.34438419e-01 -6.49895549e-01 -3.75264734e-01 -4.08712924e-01
-2.25567624e-01 -7.15999424e-01 6.04278445e-01 1.01706040e+00
-8.44517052e-01 3.12071383e-01 4.06442612e-01 7.52799958e-02
-3.00285637e-01 -9.58778441e-01 7.93837130e-01 -3.66578579e-01
-8.79742742e-01 3.31462808e-02 -3.67565900e-01 2.39255294e-01
1.08558011e+00 -2.28312626e-01 -4.00950126e-02 3.50825965e-01
-2.43240312e-01 5.40502854e-02 1.04034007e+00 9.38198417e-02
3.91651869e-01 -1.16138327e+00 -3.66293758e-01 5.69955707e-01
2.85555094e-01 -1.61964163e-01 -4.46225375e-01 3.37753087e-01
-5.18903255e-01 4.71947819e-01 -1.57150730e-01 -1.58615083e-01
-1.13153803e+00 8.13449502e-01 2.50661343e-01 -4.84553605e-01
-3.93407613e-01 6.09230578e-01 -3.26214731e-01 -5.21452963e-01
1.49963154e-02 -8.76246870e-01 2.34912649e-01 -4.20689791e-01
4.91303176e-01 2.31494650e-01 8.09726045e-02 -5.57678677e-02
-4.87043500e-01 2.74152100e-01 4.73966390e-01 -3.40575665e-01
1.25169003e+00 6.38322294e-01 -4.90629941e-01 3.78558934e-01
1.17323732e+00 5.42831495e-02 -7.67840981e-01 -3.58180583e-01
4.22962278e-01 -5.78313172e-01 -1.61621317e-01 -6.13240898e-01
-7.86670446e-01 7.17808306e-01 7.00256452e-02 3.23510379e-01
1.19238532e+00 -6.06651008e-02 6.64550900e-01 4.85404879e-01
8.12061071e-01 -7.58622825e-01 -2.26207264e-02 4.39945757e-01
4.14313406e-01 -7.26608694e-01 -1.89964250e-01 -1.93277031e-01
-4.54467475e-01 1.32698965e+00 5.43764710e-01 -2.14736968e-01
1.48213357e-02 7.77974665e-01 -8.15188110e-01 -1.11157000e-01
-1.06705332e+00 2.87710540e-02 3.21704388e-01 5.26068687e-01
2.38223195e-01 2.83082053e-02 -2.84824252e-01 6.29173934e-01
-3.64344865e-01 3.13836336e-01 5.60263038e-01 1.51550031e+00
-4.81085628e-01 -1.64044595e+00 -1.62587866e-01 6.18720651e-01
-3.57258171e-01 -1.79709896e-01 -4.58826810e-01 7.35201597e-01
-1.00957051e-01 5.70432723e-01 -2.25308582e-01 -2.95335352e-01
4.64106232e-01 2.12982088e-01 6.32370472e-01 -1.05219281e+00
-1.04462898e+00 -5.59474170e-01 3.29840809e-01 -6.96937382e-01
-6.09931350e-02 -7.02244341e-01 -1.38155508e+00 -2.59306312e-01
-1.64542750e-01 2.99774617e-01 6.26270711e-01 7.81638145e-01
1.92279786e-01 5.30762315e-01 7.11569071e-01 -4.79680061e-01
-6.32956147e-01 -1.27658129e-01 -2.31994033e-01 -5.37992269e-02
4.61061597e-01 -1.61148235e-01 -1.75178677e-01 -9.99074243e-03] | [8.202512741088867, 7.350993633270264] |
5ba4befc-c283-42be-837f-51846c4cd35c | openceres-when-open-information-extraction | null | null | https://aclanthology.org/N19-1309 | https://aclanthology.org/N19-1309.pdf | OpenCeres: When Open Information Extraction Meets the Semi-Structured Web | Open Information Extraction (OpenIE), the problem of harvesting triples from natural language text whose predicate relations are not aligned to any pre-defined ontology, has been a popular subject of research for the last decade. However, this research has largely ignored the vast quantity of facts available in semi-structured webpages. In this paper, we define the problem of OpenIE from semi-structured websites to extract such facts, and present an approach for solving it. We also introduce a labeled evaluation dataset to motivate research in this area. Given a semi-structured website and a set of seed facts for some relations existing on its pages, we employ a semi-supervised label propagation technique to automatically create training data for the relations present on the site. We then use this training data to learn a classifier for relation extraction. Experimental results of this method on our new benchmark dataset obtained a precision of over 70{\%}. A larger scale extraction experiment on 31 websites in the movie vertical resulted in the extraction of over 2 million triples. | ['Prashant Shiralkar', 'Xin Luna Dong', 'Colin Lockard'] | 2019-06-01 | null | null | null | naacl-2019-6 | ['open-information-extraction'] | ['natural-language-processing'] | [ 1.66579977e-01 9.63868678e-01 -6.65437579e-01 -3.27813596e-01
-8.17912579e-01 -7.70172954e-01 4.48933333e-01 5.31778872e-01
-1.76689476e-01 1.28013992e+00 1.99709475e-01 -2.25164428e-01
-1.17961712e-01 -1.16268587e+00 -8.10369074e-01 3.31127696e-04
-1.85669497e-01 7.26942956e-01 5.71931541e-01 -2.35244125e-01
-7.46939778e-02 -3.36687565e-02 -1.58122945e+00 6.41041160e-01
1.01762342e+00 1.11366010e+00 -2.74052352e-01 1.37281090e-01
-6.81442678e-01 1.29917622e+00 -3.44835967e-01 -1.13009870e+00
3.08636039e-01 -1.50602251e-01 -1.57884014e+00 2.43041664e-02
3.77645314e-01 2.72263139e-01 -7.91550204e-02 1.16275287e+00
1.08935907e-01 -3.93174589e-01 4.66970444e-01 -1.34571874e+00
-4.48990464e-01 1.27468765e+00 -4.29804325e-01 1.83748603e-01
7.11870790e-01 -6.84148788e-01 1.66408408e+00 -6.52149856e-01
1.19857442e+00 6.29153550e-01 5.56299090e-01 1.63192153e-01
-8.93126309e-01 -5.11182189e-01 -2.50820994e-01 1.71053723e-01
-1.44453406e+00 -2.96622783e-01 4.79533315e-01 -3.74977082e-01
9.21799839e-01 3.67718309e-01 6.07319415e-01 5.58106124e-01
2.57715601e-02 7.21008718e-01 1.24453366e+00 -9.25832093e-01
8.61908775e-04 6.82886779e-01 6.51577592e-01 9.16803360e-01
9.44953203e-01 -3.30582708e-01 -5.57923317e-01 -2.08228201e-01
3.22061598e-01 -5.27114332e-01 -1.94008246e-01 -3.57210249e-01
-8.40287507e-01 6.80575907e-01 3.16579044e-01 4.88972902e-01
-2.24710554e-01 -6.59873903e-01 4.22780246e-01 2.74544448e-01
6.85173750e-01 5.95637679e-01 -8.92005026e-01 2.35195890e-01
-6.77438200e-01 3.21501136e-01 1.73330581e+00 1.28285909e+00
1.02123654e+00 -9.27657425e-01 1.03470497e-01 7.02336729e-01
1.34308770e-01 2.01280713e-01 1.36438504e-01 -7.42996156e-01
9.77322698e-01 1.28202903e+00 3.39933485e-01 -1.09331322e+00
-4.51412618e-01 -1.03206471e-01 -2.49900714e-01 -3.82905155e-01
4.62436557e-01 -3.84792507e-01 -4.89070505e-01 1.20200765e+00
6.88950121e-01 -3.45574975e-01 2.88873017e-01 6.36340022e-01
1.29813969e+00 3.73909354e-01 -5.40044568e-02 -5.30756950e-01
1.63812923e+00 -8.29309165e-01 -9.85912025e-01 -3.35542075e-02
7.66220450e-01 -6.05847061e-01 6.06955588e-01 2.73675054e-01
-7.56078243e-01 1.08317263e-01 -1.17588854e+00 -3.34662110e-01
-9.35215890e-01 9.40459874e-03 9.44045782e-01 5.44943154e-01
-4.91286725e-01 4.69341397e-01 -4.42613065e-01 -6.04364634e-01
3.86511624e-01 3.66269320e-01 -5.18385828e-01 6.13179691e-02
-1.53187037e+00 9.00562108e-01 8.19446385e-01 -3.77293974e-01
-7.92506244e-03 -4.37252253e-01 -9.04784858e-01 1.02500089e-01
1.38746262e+00 -6.54404223e-01 1.15019131e+00 -6.48464620e-01
-7.55094469e-01 1.30017507e+00 -6.98907450e-02 -6.71246231e-01
1.27778919e-02 -3.36251676e-01 -8.24559689e-01 1.12836383e-01
5.99539936e-01 1.24575607e-01 1.48239732e-01 -1.48989034e+00
-1.13074481e+00 -4.99413431e-01 4.64498907e-01 -5.21914810e-02
-4.57778484e-01 3.33383203e-01 -4.71167445e-01 -3.06806087e-01
2.26702541e-01 -8.76145065e-01 -7.29967952e-02 -7.06751108e-01
-8.45194340e-01 -6.13284826e-01 3.95368278e-01 -7.19908834e-01
1.71056831e+00 -1.36236095e+00 -1.75296649e-01 3.33075821e-01
6.20011091e-01 1.38860241e-01 4.87722695e-01 5.35156071e-01
-9.76055562e-02 5.16931534e-01 -8.58674347e-02 3.24713230e-01
-5.10201231e-03 3.20835143e-01 -3.93337905e-01 -2.43467659e-01
-3.25154588e-02 8.18720102e-01 -1.02540112e+00 -1.20918596e+00
-6.00822031e-01 -1.84893772e-01 -3.73139620e-01 3.22939083e-02
-6.92954659e-01 -2.56752878e-01 -6.25674486e-01 7.63685524e-01
3.75369221e-01 -4.45163488e-01 6.00327551e-01 -4.80628937e-01
-1.18605658e-01 5.95917284e-01 -1.23570609e+00 1.31671381e+00
-1.08452469e-01 2.43299469e-01 -4.03086275e-01 -5.79299808e-01
7.39000559e-01 4.79721844e-01 1.05917084e+00 -3.86801451e-01
3.00668031e-01 3.17679584e-01 -2.55941570e-01 -1.06448352e+00
6.34275556e-01 3.12506966e-02 -2.37102836e-01 4.65619475e-01
4.16277081e-01 1.99123755e-01 1.10191488e+00 5.22914946e-01
1.31764436e+00 2.24609345e-01 5.51544964e-01 -1.16493627e-01
5.35343528e-01 6.23865962e-01 7.48039603e-01 2.64003783e-01
1.43438995e-01 1.81139577e-02 7.54690886e-01 -6.89885974e-01
-9.50591981e-01 -5.85316002e-01 -3.34439486e-01 5.93299448e-01
1.36783039e-02 -1.27245951e+00 -8.30338359e-01 -1.28602016e+00
-5.37547879e-02 2.85684258e-01 -5.21615505e-01 4.16794419e-01
-3.22904438e-01 -6.53719962e-01 2.99635112e-01 3.88118904e-03
6.34702682e-01 -9.24528718e-01 -3.00124317e-01 1.62268057e-01
-9.80248570e-01 -1.79533613e+00 1.61970630e-01 2.26190999e-01
-4.91565287e-01 -1.59467494e+00 1.33099347e-01 -9.23431098e-01
7.37242937e-01 -1.99489847e-01 1.59869123e+00 3.27782370e-02
-5.43304197e-02 -1.42533436e-01 -7.05959976e-01 -5.55403531e-01
-2.58178890e-01 5.13535500e-01 -8.02014694e-02 -2.07465053e-01
1.03768003e+00 -2.32846498e-01 1.29531935e-01 2.22009182e-01
-7.34735608e-01 1.01313584e-01 4.36681330e-01 4.56596047e-01
6.32528603e-01 6.45942152e-01 4.86924171e-01 -1.73034990e+00
5.94991624e-01 -5.39498568e-01 -5.66420615e-01 6.92332983e-01
-9.10603344e-01 3.43818635e-01 2.91462302e-01 1.14218913e-01
-1.20061457e+00 1.19093210e-01 3.55003886e-02 4.81412798e-01
6.98232576e-02 1.11242998e+00 -4.12967712e-01 1.22254588e-01
8.25367808e-01 -4.69965607e-01 -4.57397729e-01 -4.06463027e-01
4.28124875e-01 9.33818638e-01 3.76480848e-01 -8.11469495e-01
8.34299326e-01 3.32574159e-01 -5.88327199e-02 -4.05707121e-01
-1.77296317e+00 -5.19857526e-01 -9.38566744e-01 2.05673594e-02
6.65924728e-01 -9.22542870e-01 -4.58251923e-01 -1.86295584e-01
-8.47795248e-01 3.59222069e-02 -4.02900517e-01 1.62673667e-01
-2.44314373e-01 1.57332599e-01 -4.84071285e-01 -5.65137208e-01
-4.22968984e-01 -4.64479148e-01 6.37791753e-01 2.39855677e-01
-2.00558722e-01 -8.98851871e-01 3.97610754e-01 8.00947547e-01
-2.22576320e-01 3.91867161e-01 9.39213276e-01 -9.84318018e-01
-6.77596927e-01 -3.51028085e-01 -3.69138092e-01 1.00034133e-01
2.45468393e-01 -6.91641048e-02 -6.49911642e-01 3.58892977e-01
-2.92889208e-01 -5.18295586e-01 4.67708379e-01 -3.48074466e-01
6.54627502e-01 -7.11181879e-01 -6.50751770e-01 1.47929892e-01
1.63601875e+00 -2.13026330e-01 5.48683107e-01 5.79808533e-01
6.93316162e-01 8.44236016e-01 9.34663773e-01 3.29193681e-01
8.71413529e-01 5.91558635e-01 7.37432241e-02 2.26536676e-01
3.40983942e-02 -5.59696317e-01 -7.62358084e-02 7.79972076e-01
-4.59732831e-01 -2.58256476e-02 -9.64819431e-01 5.50870836e-01
-1.74568534e+00 -1.00051200e+00 -4.74646002e-01 1.96216142e+00
1.66135573e+00 4.15870190e-01 1.51402935e-01 3.52610797e-01
6.16558194e-01 -2.29773074e-01 5.25213741e-02 -5.69652431e-02
-1.73142985e-01 2.55755305e-01 6.99282050e-01 3.06535274e-01
-1.26612031e+00 1.16431010e+00 5.63837099e+00 5.85481703e-01
-5.79981029e-01 3.56737711e-02 2.48150110e-01 3.06537956e-01
-3.36209029e-01 5.83113074e-01 -1.31620455e+00 1.99429244e-01
1.01181138e+00 -4.65674877e-01 2.20775828e-01 7.63729095e-01
-2.34749049e-01 -4.00448352e-01 -1.05800009e+00 5.53582728e-01
1.82740360e-01 -1.39381802e+00 -9.33009535e-02 2.07890198e-01
9.93992329e-01 -7.50811622e-02 -7.25056231e-01 4.60477084e-01
8.42355311e-01 -5.56498528e-01 4.57964003e-01 3.35434139e-01
6.45399630e-01 -4.51995850e-01 9.53125417e-01 5.24569035e-01
-1.18408084e+00 -5.67833818e-02 -2.64071763e-01 -1.26637816e-01
9.37647671e-02 1.04641080e+00 -9.12803888e-01 9.70735490e-01
8.47548306e-01 5.75291932e-01 -8.59872997e-01 9.05414879e-01
-5.59971631e-01 5.86725354e-01 -4.66473341e-01 -2.30371773e-01
-2.01752514e-01 -1.54851228e-01 1.91914588e-01 1.03091943e+00
-1.62478715e-01 3.10968667e-01 3.17691743e-01 4.32073772e-01
-5.69346726e-01 5.35609782e-01 -6.78706944e-01 -2.24862188e-01
3.42378914e-01 1.58211672e+00 -7.84081936e-01 -4.97024745e-01
-7.47958958e-01 4.52176094e-01 8.24377358e-01 1.16413705e-01
-6.73430264e-01 -6.63222313e-01 -1.13989726e-01 3.46925586e-01
2.09088638e-01 4.71323356e-02 -3.57191175e-01 -1.49751079e+00
3.60998929e-01 -7.70496011e-01 8.95627201e-01 -6.91963434e-01
-1.18181050e+00 7.98246086e-01 1.65812284e-01 -1.04631257e+00
-3.12080085e-01 -5.07332385e-01 1.41328797e-01 5.12035728e-01
-1.49891484e+00 -1.15208769e+00 -2.52305508e-01 4.11642760e-01
3.46519761e-02 9.43847895e-02 8.95800769e-01 6.99133456e-01
-5.84410429e-01 2.04030797e-01 -4.12879169e-01 7.62886226e-01
7.53627002e-01 -1.51293468e+00 1.07492730e-01 9.28653896e-01
5.79768360e-01 7.28732049e-01 7.59854913e-01 -1.10433042e+00
-1.09141219e+00 -8.95116985e-01 1.87120485e+00 -8.88032258e-01
9.91935849e-01 -3.36307287e-01 -6.59853339e-01 1.08217824e+00
5.16234040e-01 2.27067843e-01 9.77632225e-01 5.39658129e-01
-4.57484454e-01 -1.48159891e-01 -1.19866049e+00 3.73615146e-01
1.23444021e+00 -3.16651583e-01 -9.21478570e-01 6.65895164e-01
8.61351728e-01 -5.28721213e-01 -1.24515462e+00 5.16482294e-01
3.86731684e-01 -5.87844729e-01 5.77793300e-01 -8.88116121e-01
6.77795112e-01 -2.82418996e-01 -6.63654730e-02 -1.00199437e+00
7.15824515e-02 -5.49238861e-01 -5.03225327e-01 1.71292424e+00
9.82710779e-01 -3.31148773e-01 8.02971900e-01 9.39222336e-01
2.55629361e-01 -8.94783914e-01 -4.42161888e-01 -5.72057664e-01
-2.37657368e-01 -1.84175014e-01 4.48409706e-01 1.01537764e+00
7.93076515e-01 8.69508088e-01 -1.00450270e-01 2.01413766e-01
6.72008991e-01 5.73694587e-01 7.26931810e-01 -1.54171801e+00
9.31505784e-02 3.24218065e-01 -1.98869169e-01 -6.14978492e-01
1.29142091e-01 -1.11384034e+00 -2.26104483e-01 -1.91162372e+00
4.56277221e-01 -8.79287362e-01 9.86544490e-02 8.64136338e-01
-1.61108464e-01 1.21252641e-01 -2.10888088e-01 2.21662730e-01
-8.96504700e-01 -3.00870766e-03 1.06048453e+00 6.98828995e-02
-5.46530373e-02 -5.71446158e-02 -1.04015017e+00 7.66008437e-01
5.23043931e-01 -7.53133714e-01 -3.36176485e-01 -5.14465524e-03
9.38441455e-01 -9.89656672e-02 -1.99548513e-01 -9.41571176e-01
3.27368110e-01 -1.28034100e-01 -1.01572841e-01 -5.31799972e-01
-1.99777275e-01 -1.21206689e+00 -9.81715322e-03 -6.33932129e-02
-2.70241916e-01 -2.97629893e-01 -2.31559962e-01 2.33032629e-01
-3.37779552e-01 -5.23294032e-01 8.21276307e-02 -3.14519614e-01
-6.31567597e-01 2.69055158e-01 2.05429077e-01 6.53830767e-01
1.04464900e+00 2.87279844e-01 -5.96307337e-01 1.02022983e-01
-8.52435887e-01 3.69274944e-01 2.36893028e-01 3.77467841e-01
1.05387293e-01 -1.24259317e+00 -5.23015440e-01 -1.61475256e-01
3.94867122e-01 -5.52271865e-03 -4.59038913e-01 5.12763739e-01
-4.40002710e-01 5.37554145e-01 3.93942781e-02 -3.37843597e-02
-1.23330247e+00 8.82338107e-01 -4.81411666e-02 -8.86798501e-01
-4.93653089e-01 3.57918859e-01 -7.27949202e-01 -4.23800975e-01
1.22291073e-01 -4.14575398e-01 -6.63680673e-01 2.94606239e-01
4.03289229e-01 -4.86305170e-02 3.82880211e-01 -6.35030568e-01
-1.56944439e-01 1.41616061e-01 -1.05816364e-01 -3.95067036e-02
1.46766841e+00 -2.20041335e-01 -5.21533310e-01 3.98564816e-01
8.67776632e-01 5.60486794e-01 -5.26943326e-01 -7.17171133e-01
7.82533646e-01 -3.89681816e-01 -2.10104868e-01 -9.51747119e-01
-9.81599152e-01 -5.30341417e-02 -1.95839465e-01 8.97090256e-01
7.17112064e-01 4.32564408e-01 9.38313067e-01 5.97774684e-01
9.67110038e-01 -1.17475379e+00 -3.82636338e-01 6.25861943e-01
5.39664805e-01 -1.43712616e+00 3.75508934e-01 -1.39762843e+00
-5.35702109e-01 9.03060079e-01 6.47471547e-01 2.44771361e-01
8.79537404e-01 5.71581244e-01 3.51680852e-02 -7.04277158e-01
-7.58988440e-01 -7.21660376e-01 1.81177184e-01 3.67123991e-01
5.90827525e-01 -6.47349730e-02 -8.16181540e-01 9.96472716e-01
-4.91262823e-01 4.29572791e-01 5.47585189e-01 1.19242644e+00
-6.46376610e-01 -1.40680373e+00 -1.09618902e-01 8.34011972e-01
-7.83527374e-01 -2.65024602e-01 -6.36232555e-01 7.10748136e-01
4.78259444e-01 1.08511937e+00 -3.85243297e-01 -4.26602483e-01
4.23537523e-01 2.74400055e-01 5.68709135e-01 -9.31618392e-01
-5.27123034e-01 -3.50590795e-01 9.95180666e-01 -3.18534821e-01
-8.61712754e-01 -6.25804961e-01 -1.49357343e+00 -4.73649092e-02
-6.62211120e-01 6.85955405e-01 4.37654912e-01 1.13882124e+00
1.12237975e-01 2.46865898e-01 4.56885099e-01 8.10600221e-02
-2.41985336e-01 -8.62335503e-01 -6.25910640e-01 5.98304689e-01
-2.40171313e-01 -8.09501886e-01 -7.89170638e-02 5.52761316e-01] | [9.353689193725586, 8.574577331542969] |
edbfce72-de9b-4fdc-a8cb-d2bbde79e031 | towards-a-general-deep-feature-extractor-for | 2201.07781 | null | https://arxiv.org/abs/2201.07781v1 | https://arxiv.org/pdf/2201.07781v1.pdf | Towards a General Deep Feature Extractor for Facial Expression Recognition | The human face conveys a significant amount of information. Through facial expressions, the face is able to communicate numerous sentiments without the need for verbalisation. Visual emotion recognition has been extensively studied. Recently several end-to-end trained deep neural networks have been proposed for this task. However, such models often lack generalisation ability across datasets. In this paper, we propose the Deep Facial Expression Vector ExtractoR (DeepFEVER), a new deep learning-based approach that learns a visual feature extractor general enough to be applied to any other facial emotion recognition task or dataset. DeepFEVER outperforms state-of-the-art results on the AffectNet and Google Facial Expression Comparison datasets. DeepFEVER's extracted features also generalise extremely well to other datasets -- even those unseen during training -- namely, the Real-World Affective Faces (RAF) dataset. | ['Alice Othmani', 'Liam Schoneveld'] | 2022-01-19 | null | null | null | null | ['facial-emotion-recognition'] | ['computer-vision'] | [-7.51465410e-02 -1.36723043e-02 9.15329009e-02 -9.46076095e-01
-6.30570352e-02 -3.52171361e-01 8.79377723e-01 -3.57325584e-01
-4.31211650e-01 5.27980387e-01 7.24816844e-02 2.97281265e-01
3.01538467e-01 -4.18092519e-01 -3.85368675e-01 -7.87889123e-01
-1.49651691e-01 1.08073935e-01 -6.39716446e-01 -5.79491317e-01
-1.02120295e-01 9.19399679e-01 -1.66842091e+00 4.34593409e-01
-1.11601420e-01 1.68118799e+00 -5.17315269e-01 5.66954792e-01
-3.32053751e-01 1.16449058e+00 -6.78358436e-01 -8.31229627e-01
5.87939173e-02 -3.71131092e-01 -7.87208796e-01 1.16797499e-01
6.00685060e-01 -2.26323396e-01 -1.06611647e-01 9.08643246e-01
7.11238325e-01 1.37140499e-02 8.25565875e-01 -1.68323064e+00
-5.85917890e-01 -2.54264355e-01 -8.69502187e-01 -8.50300714e-02
3.08313400e-01 -5.55550195e-02 6.29461527e-01 -1.25493252e+00
8.94939780e-01 1.49337935e+00 6.03660405e-01 9.88828301e-01
-8.01259279e-01 -9.52522457e-01 3.51724550e-02 2.02214181e-01
-1.09099030e+00 -7.49377370e-01 1.05235684e+00 -4.27525997e-01
9.91233647e-01 1.39722228e-01 7.13345230e-01 1.60120273e+00
-9.62738693e-03 1.12016737e+00 1.32478988e+00 -2.93303609e-01
4.71104309e-02 3.07268411e-01 -3.47626656e-01 8.88683379e-01
-3.94459069e-01 1.01737931e-01 -7.46495068e-01 -5.73100895e-02
4.02107775e-01 5.37082739e-03 -2.28314504e-01 -2.34844744e-01
-4.74115998e-01 8.47262979e-01 6.31673157e-01 3.02484572e-01
-7.22920716e-01 4.43828672e-01 9.86063898e-01 6.49339914e-01
7.58399546e-01 2.23036215e-01 -5.37094653e-01 -3.62417012e-01
-9.15601254e-01 3.66449878e-02 6.47247970e-01 3.12604994e-01
9.27437365e-01 4.69923407e-01 -1.15456812e-01 1.06773853e+00
3.30088526e-01 3.27852368e-01 4.39627588e-01 -5.73074341e-01
-2.03633502e-01 6.43004596e-01 -1.87632352e-01 -1.30881083e+00
-5.72306752e-01 -5.33469096e-02 -1.10286272e+00 7.99013495e-01
-3.97563241e-02 -5.03421783e-01 -9.27470505e-01 1.91511774e+00
3.16417217e-01 -4.78126109e-02 1.25761986e-01 1.15629101e+00
1.26890695e+00 6.28010094e-01 4.83053923e-01 2.08460521e-02
1.32101643e+00 -7.74700940e-01 -8.60985160e-01 -2.92865962e-01
4.15183246e-01 -5.51428437e-01 8.90627980e-01 6.25501394e-01
-7.68128991e-01 -5.35860181e-01 -8.58298421e-01 -5.39065041e-02
-6.92744970e-01 3.42236698e-01 1.01877904e+00 7.52141118e-01
-1.32152367e+00 1.79972410e-01 -2.94722587e-01 -3.99239779e-01
1.02508485e+00 6.09122574e-01 -1.22828948e+00 1.56483546e-01
-9.67893302e-01 9.76352632e-01 -2.85748634e-02 3.80470395e-01
-8.70888114e-01 -3.68608773e-01 -9.77172852e-01 5.76207153e-02
4.00854163e-02 -4.42962050e-01 1.15545785e+00 -2.39649439e+00
-1.75808918e+00 1.38395131e+00 -2.20674053e-01 -5.63995987e-02
2.23624125e-01 -3.12117159e-01 -6.60843730e-01 2.80176699e-01
-3.95493984e-01 9.27035928e-01 1.33375788e+00 -1.30835700e+00
5.07224947e-02 -7.00429559e-01 -2.67494291e-01 -2.16812879e-01
-6.26630247e-01 5.14753819e-01 -1.43657118e-01 -5.48167050e-01
-7.49687135e-01 -7.25891709e-01 2.54962966e-03 5.09881914e-01
-1.54762506e-01 -5.47838509e-01 1.33264935e+00 -3.22440237e-01
8.11695516e-01 -2.45190668e+00 1.31138489e-01 3.12732875e-01
3.49190176e-01 5.80942035e-01 -4.52888817e-01 2.82760888e-01
-5.68141699e-01 9.96541884e-03 -7.37989023e-02 -5.83190382e-01
2.52393067e-01 2.12790251e-01 -8.92272033e-03 5.89838386e-01
7.31692493e-01 1.18752718e+00 -6.16555274e-01 -2.97558755e-01
1.74745023e-01 9.07155037e-01 -1.99210986e-01 4.60841805e-01
-2.10467316e-02 1.66166842e-01 -2.38792628e-01 8.29377651e-01
7.52251446e-01 -8.18198174e-02 -2.84419581e-03 -5.65165617e-02
2.70496488e-01 -6.50723755e-01 -5.81409395e-01 1.54241502e+00
-7.83786774e-01 1.22337139e+00 4.46561337e-01 -1.08599329e+00
1.40534341e+00 3.93906027e-01 3.44494820e-01 -7.62374520e-01
6.43669069e-01 -7.83339813e-02 -2.09409460e-01 -7.92085886e-01
2.42796302e-01 -7.81753719e-01 -1.43970102e-01 2.87542284e-01
6.15825593e-01 1.31431013e-01 -2.45250151e-01 1.60496701e-02
6.86511219e-01 6.19496554e-02 4.26627785e-01 -1.58991605e-01
6.40797317e-01 -5.97334981e-01 2.33841911e-01 5.09433076e-03
-3.27984959e-01 2.45404154e-01 6.87773108e-01 -7.64907062e-01
-4.76964146e-01 -5.32803953e-01 5.41263632e-02 1.49990654e+00
-3.15124094e-01 -4.59686697e-01 -6.83798492e-01 -9.03562844e-01
1.09572619e-01 1.80137724e-01 -1.31232560e+00 -2.05825448e-01
1.01099871e-01 -4.55651909e-01 7.72936165e-01 6.36354566e-01
5.45845211e-01 -1.52491724e+00 -7.68441498e-01 7.98418932e-03
2.29279011e-01 -1.13877344e+00 2.80605704e-01 1.11310787e-01
-1.07981935e-01 -8.74221563e-01 -8.06387663e-01 -6.04770482e-01
5.40446579e-01 -4.25296366e-01 1.33957672e+00 1.23426013e-01
-5.15232742e-01 7.31827140e-01 -3.02752763e-01 -9.53443170e-01
-1.29437938e-01 -3.56301785e-01 -8.96878764e-02 8.32582414e-01
8.44510198e-01 -3.80190998e-01 -5.26620209e-01 -1.26458835e-02
-8.61414254e-01 -4.70938861e-01 5.77036858e-01 7.71472156e-01
2.28064552e-01 -5.12956858e-01 7.55985498e-01 -6.17403150e-01
6.66487098e-01 -4.66565549e-01 -5.03854007e-02 5.42151406e-02
-1.01583293e-02 -1.66402474e-01 5.39603710e-01 -2.83009648e-01
-1.14382923e+00 1.01096600e-01 -5.34669578e-01 -7.14745224e-01
-6.53285623e-01 3.17488402e-01 -3.12170953e-01 -3.99367779e-01
5.57010531e-01 -1.73038214e-01 3.39363784e-01 -1.82332382e-01
3.70453477e-01 7.52283931e-01 5.19847929e-01 -3.30000401e-01
2.81043112e-01 6.25330925e-01 2.48098969e-01 -1.12412989e+00
-6.32129669e-01 -4.26618308e-01 -4.78191048e-01 -6.37168705e-01
1.04001629e+00 -1.02826977e+00 -1.02688468e+00 6.68103337e-01
-1.18036771e+00 -3.37770998e-01 -2.81736050e-02 3.29551809e-02
-6.42506599e-01 9.81638059e-02 -4.31187212e-01 -9.98818278e-01
-5.33888757e-01 -8.75601292e-01 1.49427843e+00 2.64942646e-01
-5.07566929e-01 -1.08984184e+00 1.49683535e-01 -1.69829920e-01
5.59756041e-01 8.28686833e-01 4.85078812e-01 -3.95282149e-01
3.01241606e-01 -3.09887290e-01 -4.28948373e-01 6.77160323e-01
-1.10230722e-01 2.93637663e-01 -1.50863326e+00 -1.41172454e-01
-2.22075447e-01 -1.15760517e+00 1.06803524e+00 3.61256823e-02
1.55614018e+00 -2.68373609e-01 -1.27493858e-01 7.15776324e-01
1.30669558e+00 -1.58735558e-01 9.25519645e-01 2.01169714e-01
4.74777907e-01 8.69368970e-01 3.48796219e-01 6.39975667e-01
-2.16719247e-02 7.31679082e-01 8.41510594e-01 -7.40754843e-01
2.33621985e-01 2.24363670e-01 5.47155380e-01 4.37028073e-02
-3.09371769e-01 -2.56615907e-01 -7.02534080e-01 3.75842661e-01
-1.62356639e+00 -1.05982578e+00 1.51103884e-01 1.39818299e+00
5.94383597e-01 -5.01230776e-01 1.82112679e-01 6.28114194e-02
2.03817025e-01 4.63940084e-01 -2.79708236e-01 -1.25695765e+00
-2.94288367e-01 8.42358410e-01 -2.29194999e-01 -3.76638700e-03
-1.22128963e+00 1.05060947e+00 5.50784779e+00 6.78576827e-01
-1.88700902e+00 3.65202203e-02 9.32220221e-01 -1.83997452e-01
1.56972796e-01 -8.17228377e-01 -3.14909101e-01 1.45038739e-02
1.07218170e+00 2.18113914e-01 1.11645430e-01 1.18198550e+00
-1.89356267e-01 1.12867936e-01 -9.55185473e-01 1.62163663e+00
4.90076095e-01 -9.98236239e-01 1.62330200e-03 -2.17787936e-01
4.82867926e-01 -1.65018186e-01 3.55627030e-01 6.04102433e-01
-1.22045666e-01 -1.74016893e+00 3.83544415e-01 5.48197329e-01
1.19769716e+00 -1.02626693e+00 8.51265788e-01 -2.23309457e-01
-7.88191199e-01 -1.31666943e-01 -2.71208584e-01 -1.35975137e-01
-2.90560015e-02 3.73113155e-01 -5.96031427e-01 2.60423571e-01
8.92416120e-01 8.35584164e-01 -6.51761353e-01 3.05169642e-01
-1.93570629e-01 4.59672123e-01 -1.36804417e-01 -2.09324628e-01
6.32850766e-01 1.09858468e-01 1.38246581e-01 1.72639596e+00
1.47842914e-01 -8.87950584e-02 -3.42743933e-01 5.21308184e-01
-4.31933522e-01 3.11069965e-01 -9.24148202e-01 -2.52919674e-01
-4.22405571e-01 1.88008523e+00 -2.29000300e-01 -2.33424038e-01
-5.19248366e-01 1.39119077e+00 5.72894871e-01 5.38321555e-01
-5.98258197e-01 -2.97455460e-01 1.18123376e+00 -1.76213205e-01
3.29248458e-01 2.41947129e-01 3.71390074e-01 -7.84648061e-01
-2.10214019e-01 -1.00124490e+00 5.56624830e-02 -1.03887987e+00
-1.50948858e+00 1.00692439e+00 -3.48504037e-01 -6.40101016e-01
-6.10616744e-01 -1.35843730e+00 -6.80285990e-01 6.66334212e-01
-1.69105339e+00 -1.31347489e+00 -6.41082525e-01 1.09883368e+00
1.40577823e-01 -4.99810100e-01 1.21691227e+00 8.87592286e-02
-3.72220725e-01 6.35106325e-01 -4.38566297e-01 5.34897387e-01
9.03906703e-01 -1.08035171e+00 -1.15414225e-01 1.29784375e-01
3.07243258e-01 2.16002777e-01 4.60609347e-01 2.10363343e-01
-1.28941870e+00 -1.05785644e+00 6.96873963e-01 -3.26307714e-01
3.86001676e-01 -6.15852475e-01 -7.77949214e-01 5.86165845e-01
7.20121861e-01 5.77506304e-01 1.16110742e+00 2.31505379e-01
-6.79935575e-01 -2.48175129e-01 -1.20696294e+00 2.76950449e-01
7.53243148e-01 -8.94570589e-01 -2.00623676e-01 -6.52198419e-02
-1.65583134e-01 1.57943834e-02 -6.16847932e-01 5.58803260e-01
8.32699418e-01 -1.29334807e+00 7.28206873e-01 -1.11514723e+00
7.43161201e-01 1.13351829e-01 -1.34795442e-01 -1.55284488e+00
-4.95741516e-02 -6.72090292e-01 -1.30078301e-01 1.05847359e+00
1.07136704e-01 -4.12470698e-01 7.13972986e-01 3.62847805e-01
3.14727157e-01 -8.99844527e-01 -8.62563312e-01 -4.00499314e-01
-1.37509733e-01 -3.69896024e-01 5.92964828e-01 1.33344626e+00
-6.25335872e-02 5.92799723e-01 -5.73180735e-01 -3.95159751e-01
5.75124100e-02 4.47471946e-04 1.14515841e+00 -1.35547936e+00
1.08121917e-01 -6.07021868e-01 -1.10755837e+00 -4.05572772e-01
9.97011244e-01 -7.89648712e-01 -2.52087355e-01 -1.14865506e+00
9.10453349e-02 1.55373931e-01 -3.49655628e-01 8.65995884e-01
7.85718188e-02 6.62975252e-01 2.88907379e-01 -6.43556237e-01
-5.12388706e-01 9.44340527e-01 1.09269285e+00 -1.28609493e-01
2.92589694e-01 -3.98939252e-01 -6.22719228e-01 8.51158142e-01
8.35388422e-01 -9.46176946e-02 -2.25915208e-01 -1.36662731e-02
3.32838029e-01 -1.93805456e-01 7.41182804e-01 -6.30805790e-01
-2.32481465e-01 1.47511631e-01 1.04741955e+00 -2.88051609e-02
6.69410765e-01 -1.06477606e+00 -1.26668662e-01 -1.41695768e-01
-1.39105260e-01 2.32168287e-01 7.66162872e-01 2.71979392e-01
-6.38612092e-01 6.72046095e-02 8.26775134e-01 -2.66982298e-02
-1.24013257e+00 5.21071434e-01 -4.28248435e-01 -1.95273310e-01
1.11749136e+00 -1.30081639e-01 -1.75637990e-01 -7.85325766e-01
-8.91525328e-01 -1.65037587e-01 2.17683792e-01 7.13129580e-01
7.60097086e-01 -1.45336235e+00 -8.64498377e-01 3.42527509e-01
4.49538857e-01 -5.44062316e-01 2.81879574e-01 7.06585526e-01
-2.08877861e-01 -9.49977860e-02 -7.59898961e-01 -6.11670792e-01
-1.61410248e+00 5.34108222e-01 6.34788394e-01 3.16416249e-02
-1.68565273e-01 1.22621155e+00 4.41091031e-01 -4.17940259e-01
-4.89141792e-02 2.44140133e-01 -3.27380747e-01 3.94603550e-01
8.43507648e-01 -2.18444437e-01 3.35798599e-02 -1.14672518e+00
-5.20350993e-01 4.89963412e-01 1.39189035e-01 1.18930914e-01
1.47330284e+00 1.76067561e-01 -2.86583364e-01 1.97806820e-01
1.81701624e+00 -1.79751605e-01 -1.05689454e+00 1.85198769e-01
-5.07026874e-02 -3.82736295e-01 2.02689990e-01 -9.33720708e-01
-1.59652400e+00 1.40664947e+00 7.93968737e-01 -1.07413188e-01
1.56524050e+00 2.41511725e-02 3.26694518e-01 4.08337474e-01
3.27516906e-02 -9.83327806e-01 4.35661048e-01 3.52829307e-01
1.39891374e+00 -1.45638609e+00 -3.25482517e-01 3.25093418e-02
-1.07338989e+00 1.40026581e+00 7.53003597e-01 -1.42920837e-02
7.63977230e-01 4.68987107e-01 5.08685112e-01 -7.08486736e-01
-8.35351229e-01 -3.15564126e-01 3.45708728e-01 7.27620244e-01
6.51062846e-01 3.21026845e-03 2.90793806e-01 6.79286420e-01
-1.09811850e-01 3.01514953e-01 4.77671623e-02 6.53005242e-01
-5.00476621e-02 -8.28625977e-01 -6.41897619e-02 2.26792127e-01
-7.66844630e-01 2.07730189e-01 -1.10188890e+00 1.00420213e+00
2.49825269e-02 7.16145992e-01 2.23209068e-01 -2.35993996e-01
3.27378243e-01 4.72849369e-01 6.25653267e-01 -2.18588844e-01
-8.70995104e-01 -2.34595448e-01 2.33037144e-01 -9.45056140e-01
-9.66624856e-01 -2.57657707e-01 -1.14896286e+00 -3.26448768e-01
1.89882323e-01 -1.50669307e-01 7.23142743e-01 6.84716463e-01
3.80605966e-01 2.61704266e-01 6.29534960e-01 -1.14769065e+00
3.85993905e-02 -1.02865434e+00 -7.21908569e-01 7.69863725e-01
5.93117058e-01 -8.21849048e-01 -1.58415079e-01 -7.77136460e-02] | [13.572559356689453, 1.8235673904418945] |
1f844c08-94bd-461e-99a4-cdb578857b75 | multi-scale-multi-band-densenets-for-audio | 1706.09588 | null | http://arxiv.org/abs/1706.09588v1 | http://arxiv.org/pdf/1706.09588v1.pdf | Multi-scale Multi-band DenseNets for Audio Source Separation | This paper deals with the problem of audio source separation. To handle the
complex and ill-posed nature of the problems of audio source separation, the
current state-of-the-art approaches employ deep neural networks to obtain
instrumental spectra from a mixture. In this study, we propose a novel network
architecture that extends the recently developed densely connected
convolutional network (DenseNet), which has shown excellent results on image
classification tasks. To deal with the specific problem of audio source
separation, an up-sampling layer, block skip connection and band-dedicated
dense blocks are incorporated on top of DenseNet. The proposed approach takes
advantage of long contextual information and outperforms state-of-the-art
results on SiSEC 2016 competition by a large margin in terms of
signal-to-distortion ratio. Moreover, the proposed architecture requires
significantly fewer parameters and considerably less training time compared
with other methods. | ['Naoya Takahashi', 'Yuki Mitsufuji'] | 2017-06-29 | null | null | null | null | ['audio-source-separation', 'music-source-separation'] | ['audio', 'music'] | [ 2.18941793e-01 -3.99567604e-01 1.65570781e-01 -2.30784491e-02
-1.22639787e+00 -2.02333286e-01 2.72059679e-01 -1.30389139e-01
-3.94820690e-01 7.27566719e-01 2.07928032e-01 5.31391539e-02
-3.16475123e-01 -2.92638391e-01 -5.31216323e-01 -9.03931201e-01
-1.41888291e-01 -1.52180925e-01 1.40327409e-01 -1.73722759e-01
-2.39906400e-01 4.87377971e-01 -1.59230065e+00 4.62243944e-01
7.17771828e-01 1.46214998e+00 4.44271974e-02 7.00026095e-01
2.36411870e-01 7.27408051e-01 -5.88805616e-01 -1.07870072e-01
1.54726818e-01 -6.34390354e-01 -5.03366053e-01 -1.02758318e-01
4.26672220e-01 -3.02449733e-01 -4.39329863e-01 1.13720453e+00
1.14046788e+00 2.12808967e-01 5.74281752e-01 -1.01386333e+00
-9.17465389e-02 8.22900534e-01 -5.25933027e-01 4.96874958e-01
1.34656191e-01 -9.83714312e-02 9.24424827e-01 -1.08294761e+00
-5.76285720e-02 9.91348088e-01 9.12985563e-01 1.86417207e-01
-1.33527124e+00 -8.17573667e-01 -5.90406209e-02 5.52482367e-01
-1.46277821e+00 -8.45200419e-01 1.12653363e+00 -2.29628786e-01
7.35522449e-01 1.37607798e-01 3.70424360e-01 1.11653590e+00
-3.10006082e-01 7.21561313e-01 9.21656013e-01 -5.29820740e-01
2.68513650e-01 -4.81474437e-02 -1.41167641e-01 2.37817988e-01
-1.90080493e-03 2.95598190e-02 -6.75443888e-01 -1.07511416e-01
5.27413428e-01 -4.02598143e-01 -6.62004352e-01 -6.55080900e-02
-9.11851525e-01 7.18307257e-01 5.99989831e-01 6.44069195e-01
-4.02236849e-01 2.97194242e-01 5.14091671e-01 2.25161970e-01
7.33225822e-01 2.78576106e-01 -2.89512962e-01 -4.32145112e-04
-1.45909643e+00 2.18604088e-01 6.53611004e-01 6.45852923e-01
2.11586013e-01 8.02050769e-01 6.30140305e-04 1.14897549e+00
1.19393416e-01 2.29073644e-01 5.45534790e-01 -7.50860870e-01
4.47463572e-01 -8.13881401e-03 -1.86630189e-01 -7.37866700e-01
-4.93898332e-01 -1.10331309e+00 -1.20379961e+00 2.47754022e-01
2.91706681e-01 -4.13494140e-01 -8.73847306e-01 1.61703265e+00
2.52155364e-01 4.91609216e-01 1.66848928e-01 1.14726400e+00
1.11660695e+00 8.35394561e-01 -1.94209516e-01 -1.30276725e-01
1.09266770e+00 -1.02830029e+00 -6.99511766e-01 -8.24759677e-02
-4.44862619e-02 -9.11349714e-01 6.40400708e-01 7.45965302e-01
-1.32233477e+00 -9.48984802e-01 -1.25914645e+00 1.67420119e-01
-1.28539979e-01 4.75683004e-01 1.65294692e-01 6.79673970e-01
-1.05553031e+00 7.58245409e-01 -5.54023266e-01 8.11878666e-02
4.85923737e-01 3.91908884e-01 -2.30346814e-01 7.34935254e-02
-1.32596278e+00 4.72122699e-01 2.96592593e-01 2.35311106e-01
-1.28103876e+00 -8.98991644e-01 -9.37600076e-01 5.51439762e-01
4.66452897e-01 -4.75786150e-01 1.34214389e+00 -1.18074846e+00
-1.80257201e+00 2.71721095e-01 2.64578968e-01 -7.89052486e-01
2.92896777e-01 -6.34039760e-01 -6.89373553e-01 6.14597321e-01
-1.79001108e-01 4.75280166e-01 1.31143391e+00 -9.05534208e-01
-6.08738363e-01 1.00707188e-01 4.57462147e-02 2.34204251e-02
-3.21308762e-01 2.66855091e-01 -1.03640273e-01 -1.08832562e+00
-1.04606971e-01 -6.99637949e-01 -7.30676651e-02 -1.94567487e-01
-2.71278441e-01 -3.47454473e-02 5.62206447e-01 -7.96750963e-01
1.06985056e+00 -2.61908937e+00 2.88537920e-01 -6.91346675e-02
1.50685087e-01 5.92647493e-01 -2.31363192e-01 3.31928223e-01
-5.70178688e-01 -2.59895295e-01 -4.26080912e-01 -5.08969903e-01
1.01640504e-02 -5.62071025e-01 -4.03032124e-01 6.23449564e-01
3.60072523e-01 2.86026865e-01 -6.48829341e-01 -2.45784044e-01
9.61630419e-02 7.64185488e-01 -6.32524192e-01 3.07240009e-01
1.76780909e-01 4.08522576e-01 1.35853633e-01 3.52885872e-01
8.28814566e-01 9.45299193e-02 -8.01995248e-02 -2.01291203e-01
-1.62265182e-01 6.19701207e-01 -1.41269493e+00 2.02412415e+00
-6.60997033e-01 8.66080344e-01 6.61851525e-01 -1.21632290e+00
6.99157834e-01 1.02311170e+00 4.90178466e-01 -3.32593918e-01
3.32124561e-01 5.53120613e-01 3.11046779e-01 -3.59785169e-01
9.76994187e-02 -3.23154181e-01 2.30578229e-01 1.02773353e-01
5.62451482e-01 -1.04966506e-01 6.28234968e-02 -1.61705613e-01
9.37987745e-01 -7.67188296e-02 1.23750612e-01 -3.17836165e-01
8.74083102e-01 -5.43205917e-01 5.17217100e-01 3.43081236e-01
-9.23049003e-02 8.60898554e-01 2.80692846e-01 -5.10927662e-02
-7.94229627e-01 -9.75704312e-01 -1.20457880e-01 9.80959594e-01
-1.80236250e-01 -2.66512603e-01 -9.73979890e-01 -1.86290830e-01
-2.00515643e-01 3.81568938e-01 -1.97304457e-01 -1.28004268e-01
-6.55225575e-01 -6.17841125e-01 9.60190654e-01 5.38129628e-01
6.97503448e-01 -1.02236283e+00 -4.78797823e-01 4.25946087e-01
-3.48362327e-01 -1.25660229e+00 -2.04634845e-01 5.17312109e-01
-5.65608501e-01 -7.30592728e-01 -1.08989775e+00 -7.94246674e-01
9.24821123e-02 1.29271924e-01 8.53602707e-01 -5.30684948e-01
-3.16400647e-01 3.66704799e-02 -3.43509704e-01 -5.72433054e-01
-2.47661486e-01 3.40629548e-01 7.18952864e-02 5.42551517e-01
-5.69194332e-02 -9.49299634e-01 -6.57609522e-01 1.31024331e-01
-1.02553511e+00 -3.03413033e-01 5.52781165e-01 7.85549164e-01
2.27941483e-01 5.71574152e-01 1.14727080e+00 -1.89231873e-01
6.70311511e-01 -4.79526788e-01 -5.55902123e-01 -4.01771069e-01
-2.56095201e-01 -1.71866298e-01 8.87192369e-01 -4.90827262e-01
-1.15813851e+00 6.71632886e-02 -5.19433379e-01 -6.45462573e-01
-4.06379998e-01 4.69894290e-01 -2.63202786e-01 1.17479743e-04
5.53396940e-01 8.84472579e-02 -1.76956475e-01 -9.68245268e-01
6.76335692e-02 8.80429566e-01 5.71234405e-01 -1.62237257e-01
5.37291527e-01 4.49340314e-01 -1.55676771e-02 -9.22680676e-01
-8.19577336e-01 -6.52408361e-01 -4.52360421e-01 -5.87989688e-02
8.78042102e-01 -1.27011871e+00 -3.65187138e-01 7.40311205e-01
-1.20527005e+00 -2.16602698e-01 -3.37753415e-01 8.81423414e-01
-4.87149954e-01 1.35435477e-01 -6.46307707e-01 -8.12723279e-01
-4.49265718e-01 -1.19322491e+00 8.29454303e-01 5.60928062e-02
-1.42439827e-01 -6.03710473e-01 6.40078858e-02 9.50283110e-02
7.01365709e-01 1.54942855e-01 6.48391783e-01 -6.05270147e-01
-2.93605834e-01 -1.32792264e-01 -1.52893096e-01 9.23522770e-01
6.76335068e-03 -4.95358706e-01 -1.66773117e+00 -3.15420777e-01
4.23352689e-01 -3.45070183e-01 1.21483266e+00 6.15455151e-01
1.15744245e+00 -1.16648562e-01 1.69020414e-01 7.23611414e-01
1.33253372e+00 1.60063550e-01 6.21743560e-01 1.70494951e-02
5.01075327e-01 4.96366411e-01 9.64065716e-02 3.13733786e-01
-2.14940473e-01 6.67461395e-01 5.16488910e-01 -4.36253220e-01
-6.38869047e-01 1.60642296e-01 3.39874923e-01 7.92471766e-01
2.95885857e-02 -1.53363213e-01 -5.25813282e-01 8.03799987e-01
-1.63072383e+00 -9.80342388e-01 5.24354540e-03 1.88327074e+00
9.09805834e-01 2.18859270e-01 4.21348006e-01 1.10208869e+00
6.37568176e-01 3.93115342e-01 -2.26027325e-01 -2.15501994e-01
-7.40096122e-02 6.74250484e-01 2.17715815e-01 2.54415751e-01
-1.34211254e+00 3.98237973e-01 6.13851261e+00 1.38250804e+00
-1.38351548e+00 3.14839154e-01 3.08755398e-01 -4.85830873e-01
3.12261730e-01 -5.83897412e-01 -5.06935060e-01 3.42007607e-01
1.29235756e+00 2.21955359e-01 4.68500555e-01 6.99499726e-01
2.63189021e-02 -2.44172141e-02 -1.09372854e+00 1.27543306e+00
1.44355625e-01 -1.05211210e+00 -4.52136785e-01 -2.59551495e-01
5.18281341e-01 -1.18397400e-02 2.44547889e-01 1.18487567e-01
-2.52351880e-01 -1.02815306e+00 1.08025730e+00 2.50326604e-01
8.75796556e-01 -1.09204090e+00 8.29148710e-01 2.32151419e-01
-1.33717752e+00 -2.90575445e-01 -1.98978439e-01 -1.80716977e-01
1.86090365e-01 9.30201709e-01 -5.43707311e-01 7.85321832e-01
9.58190322e-01 5.46472192e-01 -1.35919735e-01 1.35510933e+00
-2.31191844e-01 1.12251258e+00 -2.81760037e-01 4.56484824e-01
4.37966079e-01 1.92834198e-01 8.12748134e-01 1.50906670e+00
3.46750528e-01 -1.89138338e-01 -8.81414637e-02 8.76105726e-01
-2.13497490e-01 -8.96219835e-02 -4.03564036e-01 1.86195940e-01
-5.04073575e-02 1.28812516e+00 -5.76745927e-01 -2.86957383e-01
-3.38465244e-01 6.34755611e-01 -4.50120792e-02 3.99695933e-01
-9.18324709e-01 -8.48412454e-01 6.74515843e-01 2.37313777e-01
7.94198573e-01 -6.37318119e-02 -1.71018526e-01 -8.28776360e-01
2.37536412e-02 -1.11466157e+00 1.74225688e-01 -7.51353443e-01
-1.11401975e+00 1.04606295e+00 4.79717255e-02 -1.48380601e+00
-2.23818049e-01 -5.37276804e-01 -5.53016901e-01 9.29095328e-01
-1.86334491e+00 -7.92548418e-01 -2.25021571e-01 7.30277538e-01
7.23424613e-01 -3.77773255e-01 8.03156078e-01 8.93484533e-01
-4.75564778e-01 4.08740401e-01 1.21539712e-01 2.28118882e-01
7.23217547e-01 -9.98940349e-01 1.34545103e-01 9.35871184e-01
2.78565735e-01 2.15037376e-01 5.82736969e-01 -9.83361676e-02
-7.93656707e-01 -1.11593246e+00 6.10488117e-01 3.88203591e-01
6.44186735e-01 -4.89050746e-01 -8.58350158e-01 5.39718121e-02
7.20237195e-01 1.80506706e-01 7.37570882e-01 -1.81602225e-01
-4.45396721e-01 -6.34072840e-01 -9.55585003e-01 1.33992776e-01
7.03493476e-01 -7.53503680e-01 -4.99231279e-01 -1.27686467e-02
6.36276186e-01 -3.33887815e-01 -4.47401047e-01 5.00374258e-01
2.58153141e-01 -1.13778448e+00 1.15666175e+00 -3.63765180e-01
4.91411030e-01 -3.43750983e-01 -1.50194705e-01 -1.57813656e+00
-4.19048846e-01 -8.59624803e-01 -1.22676760e-01 1.30366993e+00
3.40574712e-01 -2.66513228e-01 3.43551457e-01 -1.56470731e-01
-4.72500652e-01 -7.20021486e-01 -1.12283397e+00 -9.54920888e-01
-5.83904609e-02 -6.01009369e-01 5.70160449e-01 4.78068709e-01
-5.75748533e-02 4.78049755e-01 -5.95015168e-01 2.68177569e-01
6.11754954e-01 -1.83208603e-02 2.84467965e-01 -1.16863573e+00
-4.71590281e-01 -4.57430691e-01 -3.94311517e-01 -8.20217252e-01
2.58675516e-01 -7.78009176e-01 2.31934369e-01 -1.28879583e+00
-2.84812242e-01 -1.23221360e-01 -6.49796307e-01 7.27750733e-02
-7.59649090e-03 6.57897532e-01 3.52954894e-01 -1.32508397e-01
-4.21414018e-01 6.49043083e-01 6.81932032e-01 -4.16543812e-01
-1.14133827e-01 1.80991694e-01 -5.28276265e-01 9.68937755e-01
8.18772852e-01 -6.23527527e-01 -4.08749998e-01 -4.42176253e-01
-5.92078567e-02 1.38804510e-01 5.84574401e-01 -1.60052347e+00
1.35832578e-01 4.49878156e-01 3.89605105e-01 -6.06174350e-01
7.45811343e-01 -1.00650072e+00 1.78572059e-01 2.28246123e-01
-4.73631829e-01 -4.27707821e-01 5.16408384e-01 4.02349174e-01
-8.08229625e-01 -3.43987703e-01 1.14770663e+00 9.29077715e-02
-1.90686584e-01 3.95059623e-02 -5.77075005e-01 2.14475933e-02
6.78510308e-01 2.00568095e-01 1.04392186e-01 -4.62259829e-01
-8.14462721e-01 -3.87393832e-01 -3.70003521e-01 1.59886003e-01
5.91778874e-01 -1.51543117e+00 -8.82040620e-01 2.07851395e-01
-3.14815700e-01 7.69282207e-02 4.65339869e-01 1.14732599e+00
-2.73387283e-01 4.51043338e-01 -1.56508029e-01 -4.47160870e-01
-1.22415709e+00 5.38470328e-01 5.16730428e-01 -2.65028864e-01
-6.09626532e-01 1.12969959e+00 3.67394656e-01 2.06930429e-01
5.91785967e-01 -3.29447806e-01 -2.74874359e-01 3.15450579e-01
6.55994415e-01 7.52776146e-01 4.79232818e-01 -5.85811198e-01
-3.60618383e-01 3.90498102e-01 2.50588924e-01 -3.90367597e-01
1.52361023e+00 1.62544087e-01 -3.39430664e-03 4.57998663e-01
1.38530743e+00 1.93111047e-01 -1.19296467e+00 -3.84523809e-01
-2.68497020e-01 -3.14597338e-01 4.89814967e-01 -7.16634035e-01
-1.35197270e+00 1.39199317e+00 8.01348746e-01 3.95203680e-01
1.55024052e+00 -2.67960429e-01 8.26679111e-01 3.06602917e-03
-1.02008842e-01 -1.05350471e+00 2.15643495e-01 2.96650916e-01
1.19352674e+00 -1.03158832e+00 -1.30224556e-01 -2.46743128e-01
-2.92034745e-01 1.06833541e+00 2.29852542e-01 -3.53725791e-01
9.06382442e-01 5.56361675e-01 6.18749559e-02 7.40292519e-02
-4.41364348e-01 -3.78455460e-01 6.38198435e-01 4.83591884e-01
4.97750759e-01 -2.43637577e-01 2.00986080e-02 8.84429395e-01
-8.33869502e-02 -2.74274498e-02 2.06393361e-01 5.63235521e-01
-3.73431265e-01 -7.70080805e-01 -5.91859400e-01 3.58123742e-02
-1.02563000e+00 -3.68138403e-01 -2.51540005e-01 3.85523319e-01
2.86452264e-01 1.21973836e+00 -7.03262612e-02 -9.42074433e-02
3.19983691e-01 1.37083456e-01 3.15378666e-01 -4.21436369e-01
-8.44933867e-01 8.06453407e-01 1.64132118e-01 -3.38666916e-01
-6.60011530e-01 -4.93989944e-01 -9.67421770e-01 7.75463972e-03
-5.62958956e-01 1.51428029e-01 6.82622135e-01 6.96244895e-01
2.92877465e-01 1.20228565e+00 6.83707893e-01 -1.26900375e+00
-6.04980826e-01 -1.28802907e+00 -8.70614409e-01 1.27615720e-01
9.95693922e-01 -5.12649834e-01 -5.02049267e-01 1.99826434e-01] | [15.34446907043457, 5.578188419342041] |
d85e40a3-77de-45d9-84cb-62a2d4f92c71 | generalizing-to-unseen-elements-a-survey-on | 2302.01859 | null | https://arxiv.org/abs/2302.01859v1 | https://arxiv.org/pdf/2302.01859v1.pdf | Generalizing to Unseen Elements: A Survey on Knowledge Extrapolation for Knowledge Graphs | Knowledge graphs (KGs) have become effective knowledge resources in diverse applications, and knowledge graph embedding (KGE) methods have attracted increasing attention in recent years. However, it's still challenging for conventional KGE methods to handle unseen entities or relations during the model test. Much effort has been made in various fields of KGs to address this problem. In this paper, we use a set of general terminologies to unify these methods and refer to them as Knowledge Extrapolation. We comprehensively summarize these methods classified by our proposed taxonomy and describe their correlations. Next, we introduce the benchmarks and provide comparisons of these methods from aspects that are not reflected by the taxonomy. Finally, we suggest some potential directions for future research. | ['Huajun Chen', 'Jeff Z. Pan', 'Zezhong Xu', 'Yuxia Geng', 'Wen Zhang', 'Mingyang Chen'] | 2023-02-03 | null | null | null | null | ['knowledge-graph-embedding'] | ['graphs'] | [-2.35522702e-01 3.89462441e-01 -6.69159114e-01 -1.81586310e-01
5.03643043e-02 -5.56710124e-01 4.32050407e-01 1.85994238e-01
-5.81065379e-02 9.48626518e-01 1.22461602e-01 -4.07108277e-01
-5.37716329e-01 -1.13797474e+00 -4.41703737e-01 -2.94841558e-01
-2.45256066e-01 2.72674292e-01 4.30911750e-01 -2.17612401e-01
2.16263592e-01 3.89162213e-01 -1.31867814e+00 -1.35780737e-01
9.04815793e-01 6.78826630e-01 -3.07592183e-01 6.92224950e-02
-2.62932122e-01 8.70774150e-01 -6.59651577e-01 -9.74989593e-01
-4.55072187e-02 -1.12534977e-01 -1.25184333e+00 -3.00310403e-01
1.80828795e-01 -5.45270331e-02 -9.23598230e-01 1.16460454e+00
3.81966740e-01 2.89011061e-01 7.42285728e-01 -1.78717291e+00
-1.13008487e+00 8.18190992e-01 -2.76032448e-01 2.54968792e-01
5.14476538e-01 -4.10884440e-01 1.13585150e+00 -8.78829420e-01
6.81560397e-01 1.11958814e+00 7.15088487e-01 3.29285413e-01
-6.99892521e-01 -6.35894060e-01 4.27944809e-01 9.78918672e-01
-1.78290129e+00 -1.66478515e-01 7.82017708e-01 -3.00301075e-01
1.10820150e+00 2.06108302e-01 7.83615112e-01 7.13981509e-01
-6.76644295e-02 8.65749359e-01 8.47239196e-01 -6.11711621e-01
3.62792574e-02 2.97129512e-01 5.44150651e-01 7.37718523e-01
1.04438198e+00 -7.51473829e-02 -6.38084233e-01 -1.79799899e-01
6.98864996e-01 -2.94208348e-01 -4.59376454e-01 -7.86295235e-01
-1.05644786e+00 9.33282197e-01 4.08656925e-01 3.15758884e-01
-6.14353195e-02 1.29064515e-01 3.27971280e-01 2.00831801e-01
3.10251892e-01 5.06668687e-01 -5.49219370e-01 -1.03572281e-02
-4.41300422e-01 1.72491744e-01 1.06915808e+00 1.32995021e+00
5.98463178e-01 1.24049611e-01 1.92185208e-01 6.16470456e-01
2.19910905e-01 7.18177482e-03 1.14952601e-01 -4.33055609e-01
3.26935112e-01 8.36715043e-01 5.73012196e-02 -1.50545084e+00
-1.88730836e-01 -3.72932136e-01 -7.17041731e-01 -5.49829364e-01
-4.33834642e-02 1.18738033e-01 -6.77617729e-01 1.47721517e+00
4.79748696e-01 4.62005466e-01 4.08599943e-01 4.54210371e-01
1.28326833e+00 3.60900104e-01 2.08486423e-01 -1.56425446e-01
1.23499155e+00 -1.00258446e+00 -1.07893157e+00 3.80313359e-02
5.41672766e-01 -3.45234007e-01 6.16373122e-01 -1.16269872e-01
-8.03122520e-01 -2.35768244e-01 -9.65533078e-01 1.03159368e-01
-1.17211497e+00 -1.70846090e-01 1.21633720e+00 7.31968999e-01
-1.02737069e+00 5.49262166e-01 -8.79898906e-01 -5.52622020e-01
3.74160767e-01 4.54933017e-01 -3.92418355e-01 -6.18832484e-02
-1.89391673e+00 1.17112052e+00 1.01122940e+00 3.50920290e-01
-3.30785662e-01 -4.94315475e-01 -1.14109325e+00 1.11751571e-01
8.36406112e-01 -9.39385176e-01 8.42941403e-01 -9.99493245e-03
-1.13749611e+00 5.07652462e-01 -5.86784706e-02 -1.74260885e-01
-3.37574147e-02 -1.91217214e-01 -1.12689912e+00 2.95799095e-02
-1.05800688e-01 2.01912493e-01 2.22413301e-01 -1.34575188e+00
-5.99094391e-01 -1.40228227e-01 5.16374767e-01 1.77690402e-01
-7.20632195e-01 1.77491426e-01 -8.62669170e-01 -6.80821896e-01
5.81807308e-02 -7.77459443e-01 5.41819911e-03 -4.79857475e-01
-7.13877916e-01 -4.65502441e-01 8.03808808e-01 -3.99232924e-01
1.94076085e+00 -1.80377817e+00 2.72797763e-01 3.74793917e-01
4.16346073e-01 5.03550053e-01 2.12663218e-01 9.05083418e-01
-1.27951860e-01 4.89680380e-01 4.92456146e-02 2.46149525e-01
2.33704388e-01 4.22134340e-01 -2.92838812e-01 9.25541669e-02
-5.84303103e-02 1.42034602e+00 -1.09252226e+00 -6.99873328e-01
2.75024474e-02 3.64554167e-01 -5.32265566e-02 -7.20076710e-02
-1.77492835e-02 -5.43916225e-03 -7.40241289e-01 1.12048233e+00
4.79297787e-01 -5.26790977e-01 6.28904998e-01 -5.87744713e-01
2.78495580e-01 1.78805023e-01 -1.26069367e+00 1.05613172e+00
1.23549148e-01 3.77928704e-01 -5.09418845e-01 -1.08713996e+00
6.97741985e-01 2.34365359e-01 2.33012721e-01 -3.71717513e-02
-6.40319809e-02 1.67075187e-01 1.60770640e-02 -5.42096972e-01
7.95905471e-01 9.74910241e-03 -1.20417893e-01 5.37507273e-02
1.92182995e-02 2.03596503e-01 3.89909893e-01 4.83966768e-01
1.14328456e+00 -4.00210172e-02 8.33356142e-01 -3.00902966e-02
5.46426773e-01 3.07619035e-01 6.08213365e-01 6.52251065e-01
-3.87323707e-01 -1.46225289e-01 3.56190860e-01 -4.86827940e-01
-4.45881933e-01 -1.05777264e+00 -7.75662661e-02 5.03682911e-01
5.33685446e-01 -1.09261751e+00 -2.63916701e-01 -1.15851307e+00
3.13006490e-01 6.15890503e-01 -7.22385228e-01 -3.28216732e-01
-1.69294953e-01 -8.35153759e-01 7.80553937e-01 8.97020161e-01
6.05679095e-01 -1.05198359e+00 1.12969942e-01 1.38483599e-01
-1.72495633e-01 -1.14636123e+00 6.39281720e-02 -2.10430995e-01
-8.92932892e-01 -1.48062479e+00 -3.05536062e-01 -9.26016450e-01
7.97012508e-01 5.30744135e-01 1.31632686e+00 1.85985684e-01
-2.85361912e-02 7.38510430e-01 -8.94113064e-01 -3.68065149e-01
-3.20751667e-02 2.79615879e-01 3.35441530e-01 -4.17165875e-01
8.29422832e-01 -5.91563463e-01 -4.01609063e-01 4.28259373e-01
-9.50457990e-01 -7.16855004e-02 5.62589467e-01 6.26766324e-01
7.02302873e-01 6.94722772e-01 7.90901184e-01 -1.21231937e+00
7.43208408e-01 -6.06982887e-01 -5.08304834e-01 8.22538197e-01
-9.91260946e-01 -1.26141161e-01 3.37173879e-01 -1.35099709e-01
-7.50593007e-01 -4.78314698e-01 1.81916147e-01 -3.73190969e-01
2.18561798e-01 1.16335726e+00 -2.97810853e-01 -5.90777338e-01
2.47942328e-01 1.00385286e-01 -4.67964739e-01 -5.07635713e-01
5.67199707e-01 5.00194550e-01 2.75938034e-01 -5.81020355e-01
1.05439341e+00 1.17738456e-01 -2.64821723e-02 -6.76282465e-01
-1.05522466e+00 -4.32074904e-01 -5.08275926e-01 -1.05761886e-01
5.45616627e-01 -7.05846071e-01 -7.02836812e-01 2.67789334e-01
-1.04368758e+00 1.91881850e-01 -1.33612603e-01 5.97074330e-01
-1.23511292e-01 6.70873284e-01 -5.26408970e-01 -5.77568829e-01
-1.72144920e-01 -7.33101189e-01 3.18199009e-01 4.35733497e-01
-1.05231486e-01 -1.42378557e+00 -5.12975920e-03 4.54067945e-01
4.19859558e-01 2.51644820e-01 1.11137509e+00 -6.64057672e-01
-6.37165010e-01 -2.34117538e-01 -3.55815351e-01 2.76881695e-01
3.89344841e-01 -9.23174545e-02 -5.10974348e-01 -2.43789777e-01
-4.45675939e-01 -1.90929741e-01 7.55164921e-01 -6.09934777e-02
1.11671364e+00 -2.69355983e-01 -9.31340694e-01 5.10628760e-01
1.54293180e+00 3.34992319e-01 7.66584873e-01 5.23368299e-01
9.71407294e-01 4.28285301e-01 7.94423461e-01 -6.99505629e-03
8.51202309e-01 5.72993279e-01 1.88372761e-01 2.02398717e-01
2.41794530e-02 -4.89914149e-01 6.34466186e-02 1.45867085e+00
-5.02712667e-01 -3.53111118e-01 -1.01541877e+00 7.06355870e-01
-2.04317284e+00 -7.43803799e-01 -1.45683646e-01 1.87347448e+00
8.95377278e-01 -1.13787778e-01 -1.84615090e-01 1.36595234e-01
9.48043644e-01 7.41430968e-02 -4.35472637e-01 -4.96817045e-02
-2.95975566e-01 1.00979634e-01 4.21797484e-01 3.15579951e-01
-1.04808187e+00 1.39793122e+00 7.60049677e+00 7.22889125e-01
-3.91537726e-01 -1.35572955e-01 -7.55206496e-03 4.56951916e-01
-3.14528286e-01 4.09269989e-01 -9.72077847e-01 1.87671408e-01
6.81555569e-01 -6.02957726e-01 2.62127966e-01 1.01392734e+00
-4.78699028e-01 -8.43064860e-03 -9.59668934e-01 9.37856853e-01
1.08484514e-01 -1.35190296e+00 3.58750999e-01 -1.19369626e-01
8.04096460e-01 -3.91426742e-01 -2.52058595e-01 6.39460146e-01
4.18168396e-01 -1.05281210e+00 -7.32625350e-02 4.84142959e-01
7.42057264e-01 -7.57267416e-01 9.36048329e-01 -4.58006151e-02
-1.48532903e+00 2.32568249e-01 -5.28858006e-01 -7.63687864e-02
2.27341607e-01 6.18209243e-01 -6.82912290e-01 1.50792384e+00
6.95821583e-01 8.31622362e-01 -6.82785451e-01 1.15235639e+00
-6.42899871e-01 5.92645466e-01 -4.39040996e-02 -2.07179729e-02
-1.27202854e-01 -1.63850665e-01 2.92879581e-01 9.46742475e-01
1.60222933e-01 1.63992569e-01 4.35745232e-02 6.26847863e-01
-1.82954535e-01 -1.58672016e-02 -6.27970874e-01 -5.37762284e-01
9.20193851e-01 1.05824745e+00 -7.11052537e-01 -5.22983015e-01
-8.55814159e-01 6.31899834e-01 4.63427782e-01 5.35887718e-01
-8.00277650e-01 -7.66859114e-01 5.56265116e-01 -2.24298656e-01
2.60871112e-01 -2.51550615e-01 2.62436479e-01 -1.44490480e+00
-1.28588239e-02 -6.21315122e-01 7.71045506e-01 -7.84148157e-01
-1.62342012e+00 4.64171946e-01 4.09917206e-01 -8.90981078e-01
5.84512800e-02 -5.85950136e-01 -2.23935097e-01 6.60317063e-01
-1.83183599e+00 -1.05452526e+00 -3.87779593e-01 5.46492934e-01
-2.78366745e-01 -3.33917402e-02 1.08795893e+00 4.05555308e-01
-7.73512840e-01 6.95062578e-01 3.95459346e-02 3.49122792e-01
5.42291999e-01 -1.25928760e+00 4.36167061e-01 5.82866967e-01
1.57057405e-01 1.08492160e+00 5.12977719e-01 -1.01805758e+00
-1.44482553e+00 -1.22520769e+00 1.18084919e+00 -5.11852801e-01
8.62913907e-01 -4.30872999e-02 -1.16752601e+00 1.24638510e+00
2.43948083e-02 1.37807131e-01 1.04487121e+00 6.02814674e-01
-4.13636982e-01 1.23748286e-02 -8.88338268e-01 6.72408640e-01
1.40296078e+00 -3.85561258e-01 -9.95907962e-01 3.82295042e-01
7.00522661e-01 -3.23260128e-01 -1.31513250e+00 7.88383663e-01
3.96131963e-01 -6.04147911e-01 9.09682274e-01 -9.37818348e-01
-1.25339761e-01 -6.18060291e-01 6.74826875e-02 -1.38178360e+00
-5.71739435e-01 -4.16924387e-01 -8.89736772e-01 1.25441241e+00
2.78837293e-01 -9.14466679e-01 1.06387782e+00 6.11600935e-01
3.26693710e-03 -9.82210636e-01 -5.13688087e-01 -1.30565631e+00
-1.84460923e-01 -2.58883357e-01 7.74855435e-01 1.47423136e+00
4.89977181e-01 4.42888826e-01 -3.08140337e-01 3.24239314e-01
5.55611074e-01 2.79916823e-01 6.98385060e-01 -1.35363412e+00
-4.29539010e-02 -2.44100437e-01 -8.70757043e-01 -9.10850525e-01
2.81235099e-01 -1.02002215e+00 -7.22903669e-01 -2.09048748e+00
4.34794903e-01 -4.22082514e-01 -5.73747814e-01 8.95677805e-01
-3.57604682e-01 -2.09206603e-02 -1.93085387e-01 2.67450422e-01
-8.75307024e-01 6.82880700e-01 1.16525578e+00 -7.54614100e-02
-3.91837209e-02 -2.66919643e-01 -1.02186334e+00 5.49547851e-01
8.25890303e-01 -1.91916317e-01 -7.67152965e-01 -1.57842606e-01
5.42577386e-01 -2.43514583e-01 1.25331134e-01 -6.44381702e-01
4.96945679e-01 -3.78557235e-01 5.17503619e-02 -6.00924730e-01
8.18820298e-02 -8.39947999e-01 5.70333421e-01 1.11583725e-01
2.39895776e-01 4.35787551e-02 2.62662858e-01 7.60244131e-01
-6.64415300e-01 -1.98651835e-01 2.44964287e-01 1.23782083e-01
-1.26412332e+00 4.57862705e-01 5.00055552e-02 8.03479925e-02
1.18122363e+00 -3.43511313e-01 -4.66278911e-01 -2.21066788e-01
-7.78838873e-01 4.82211560e-01 3.60370040e-01 5.17505467e-01
7.00503767e-01 -1.64272344e+00 -2.95599759e-01 -3.81320894e-01
4.91351455e-01 -6.94924071e-02 2.95630187e-01 6.98901892e-01
-2.98858792e-01 7.82025278e-01 1.86838359e-02 1.94339871e-01
-1.13160598e+00 8.96789551e-01 1.15766846e-01 -3.46334964e-01
-4.21423346e-01 6.26609921e-01 4.32798080e-02 -3.00367296e-01
1.46782443e-01 3.19347193e-04 -5.72342396e-01 -6.56077191e-02
3.77498627e-01 5.27769744e-01 1.33086573e-02 -5.02886772e-01
-7.04051256e-01 4.51185584e-01 -2.71762758e-01 5.85349917e-01
1.28719974e+00 -5.45642711e-02 -3.50855082e-01 4.30217326e-01
8.41644049e-01 4.73263860e-02 -3.60542446e-01 -4.47060674e-01
2.69661397e-01 -5.14886081e-01 -1.69208765e-01 -9.46264446e-01
-1.14450598e+00 2.95870483e-01 -5.01304194e-02 4.41234589e-01
1.02807045e+00 1.68425947e-01 6.89156830e-01 5.07765710e-01
7.29092062e-01 -1.15181589e+00 -3.12348455e-01 6.11183763e-01
6.03326857e-01 -1.07250857e+00 2.87526935e-01 -1.27926660e+00
-4.76858407e-01 9.29467976e-01 6.84971154e-01 3.02580953e-01
9.83506680e-01 -1.31473497e-01 -2.88673401e-01 -5.98817647e-01
-4.86348450e-01 -3.57883155e-01 3.84117961e-01 8.77213836e-01
2.72225678e-01 1.17140681e-01 -4.83495086e-01 9.48784292e-01
-1.27008766e-01 1.22840546e-01 3.20936143e-01 1.15967441e+00
-2.13599339e-01 -1.47040081e+00 -8.73218626e-02 6.08512104e-01
-1.89769819e-01 -2.75636971e-01 -6.09396994e-01 1.32437027e+00
-4.18847688e-02 9.13471341e-01 -5.25530100e-01 -8.07137012e-01
6.05042756e-01 1.68712333e-01 5.41208625e-01 -7.62678385e-01
-1.26425996e-02 -7.02268124e-01 4.11076933e-01 -2.71547109e-01
-6.00754619e-01 -3.31790000e-01 -1.10749722e+00 -6.33096635e-01
-7.46451199e-01 4.82244760e-01 2.65376627e-01 6.39607549e-01
5.98339081e-01 4.82171744e-01 1.24615580e-01 -8.70413259e-02
-2.15674311e-01 -6.62503242e-01 -9.07675445e-01 1.45247966e-01
-5.22496581e-01 -1.29189754e+00 -4.29289967e-01 -1.37019321e-01] | [8.824671745300293, 7.8840250968933105] |
87a37de1-c187-4e97-a4f0-5e7f6c7a1b8a | a-new-spatio-temporal-loss-function-for-3d | 2210.08562 | null | https://arxiv.org/abs/2210.08562v1 | https://arxiv.org/pdf/2210.08562v1.pdf | A New Spatio-Temporal Loss Function for 3D Motion Reconstruction and Extended Temporal Metrics for Motion Evaluation | We propose a new loss function that we call Laplacian loss, based on spatio-temporal Laplacian representation of the motion as a graph. This loss function is intended to be used in training models for motion reconstruction through 3D human pose estimation from videos. It compares the differential coordinates of the joints obtained from the graph representation of the ground truth against the one of the estimation. We design a fully convolutional temporal network for motion reconstruction to achieve better temporal consistency of estimation. We use this generic model to study the impact of our proposed loss function on the benchmarks provided by Human3.6M. We also make use of various motion descriptors such as velocity, acceleration to make a thorough evaluation of the temporal consistency while comparing the results to some of the state-of-the-art solutions. | ['Thibaut Le Naour', 'Sylvie Gibet', 'Mansour Tchenegnon'] | 2022-10-16 | null | null | null | null | ['3d-human-pose-estimation'] | ['computer-vision'] | [-4.8832834e-01 -2.6032138e-01 -3.6575812e-01 -3.8297787e-02
-3.8539934e-01 -3.2114398e-01 6.0624206e-01 -3.6184177e-01
-7.6105845e-01 4.9151459e-01 3.7582350e-01 1.9078265e-01
1.8672766e-01 -4.3532246e-01 -8.6366874e-01 -4.5619324e-01
-6.5156722e-01 1.8097784e-01 3.8511699e-01 -2.7055455e-02
-4.7604896e-02 6.0901892e-01 -9.9760669e-01 8.9314699e-02
2.0784621e-01 7.7268970e-01 -1.2574165e-02 6.7208493e-01
4.2019632e-01 7.8438121e-01 -4.8599535e-01 -3.3176368e-01
3.4442094e-01 -3.7287384e-01 -7.2976410e-01 1.5301840e-01
6.3641143e-01 -4.5677608e-01 -8.9950484e-01 7.7411348e-01
5.2059895e-01 4.9197990e-01 7.2868055e-01 -1.2786394e+00
-3.0275506e-01 4.1914668e-02 -5.5302811e-01 1.8772356e-01
7.9295325e-01 2.6436076e-01 7.5708002e-01 -9.1104478e-01
1.2921921e+00 1.5049226e+00 9.7147620e-01 5.0361031e-01
-1.1440228e+00 -2.9836512e-01 -8.8882200e-02 5.6764525e-01
-1.5587013e+00 -2.9335305e-01 9.0650755e-01 -4.0245050e-01
1.0364963e+00 -6.7410865e-03 8.2583928e-01 1.3446844e+00
8.5124791e-01 9.4311905e-01 3.8556874e-01 -1.5423428e-01
-1.7139398e-02 -4.8525807e-01 -1.0389231e-01 1.1726456e+00
1.1287580e-01 3.7714813e-03 -4.6322632e-01 -1.2406608e-01
1.0757449e+00 -1.3559541e-01 -4.7077602e-01 -8.2514089e-01
-1.4611696e+00 6.5803891e-01 5.6561708e-01 2.2718272e-01
-3.8283277e-01 8.7886757e-01 4.6969566e-01 2.0458195e-02
3.4973389e-01 -1.1639805e-01 -8.1844524e-02 -2.1341130e-01
-1.0745114e+00 5.5568737e-01 7.1945745e-01 8.2870269e-01
4.4077772e-01 1.3428439e-01 -3.7956989e-01 3.4607208e-01
6.2051630e-01 3.3359081e-01 4.4318238e-01 -1.3221841e+00
5.6937999e-01 3.2912210e-01 3.8401854e-01 -1.3787836e+00
-6.1824763e-01 -2.9093665e-01 -7.9220939e-01 1.4287275e-01
5.2887446e-01 -9.8957993e-02 -9.1631353e-01 1.7439673e+00
2.4033433e-01 6.5797764e-01 -2.5419191e-01 1.2320718e+00
6.4738953e-01 6.1428362e-01 -9.0885669e-02 9.0068974e-02
9.2858982e-01 -1.2432141e+00 -8.0672163e-01 5.4887109e-03
7.0412958e-01 -5.7481027e-01 8.5008889e-01 7.6759093e-02
-1.1408610e+00 -7.3569322e-01 -1.2256260e+00 -2.2726552e-01
-1.7047949e-02 4.3751070e-01 3.1529906e-01 9.1580026e-02
-1.0244110e+00 1.0634546e+00 -1.4540501e+00 -5.4334915e-01
-1.3377476e-01 7.5974613e-02 -7.4903250e-01 7.9407260e-02
-1.2431756e+00 1.0996469e+00 1.6271544e-01 1.5463696e-01
-1.0413587e+00 -2.5509730e-01 -1.1515532e+00 -2.4501519e-01
4.7591068e-02 -1.0743390e+00 9.6193284e-01 -4.3296438e-01
-1.4260827e+00 8.9087051e-01 -1.0425672e-01 -4.7940609e-01
1.1113948e+00 -5.4541928e-01 -4.2933308e-02 3.2739764e-01
1.1828047e-02 7.5216895e-01 8.3390993e-01 -9.1525316e-01
-1.1090905e-01 -1.4709604e-01 -9.5840856e-02 -3.6895610e-02
1.2436539e-01 -4.2368454e-01 -9.2074180e-01 -1.0326464e+00
1.4533958e-01 -1.3394036e+00 -2.3320068e-01 3.3386457e-01
-3.8275871e-01 -1.3654894e-01 8.8541609e-01 -1.0178612e+00
1.2961497e+00 -1.9368807e+00 6.9591206e-01 -2.0789940e-02
-5.9496474e-02 9.9599421e-02 -1.6156933e-01 4.9911821e-01
4.6185772e-03 -1.8150173e-02 -1.0555527e-01 -8.0885392e-01
3.1616448e-03 1.5200709e-01 9.9242650e-02 9.4810301e-01
2.2047223e-01 1.0058722e+00 -9.2110962e-01 -4.8266563e-01
2.1618320e-01 7.5559711e-01 -4.6651095e-01 4.6423125e-01
3.0269234e-02 6.3505369e-01 -3.9147758e-01 2.7512655e-01
3.7509605e-01 -2.2661972e-01 8.3615288e-02 -5.4968399e-01
9.5572010e-02 5.7992317e-02 -1.2177005e+00 2.4747005e+00
-1.8586376e-01 5.3274310e-01 -2.2181264e-01 -5.6678391e-01
5.2463830e-01 3.7611225e-01 7.5139391e-01 -3.2548955e-01
1.8449348e-01 -1.6385278e-01 -2.0165090e-01 -5.4356200e-01
4.2742187e-01 2.2918759e-01 1.4828731e-01 2.1063209e-01
2.2907604e-01 7.2852023e-02 1.1943663e-01 8.3860286e-02
1.1541692e+00 8.1193197e-01 1.4115690e-01 -1.3832149e-01
6.1979949e-01 -4.6309315e-02 5.2354413e-01 2.2474228e-01
-3.9038736e-01 7.8826338e-01 4.2462912e-01 -6.4499515e-01
-1.1795950e+00 -1.3038414e+00 3.3557755e-01 4.6203914e-01
1.9733551e-01 -5.5174673e-01 -6.1319554e-01 -7.3967993e-01
1.0992619e-01 4.1586798e-01 -6.4893442e-01 -2.6500994e-01
-8.9344895e-01 -4.0759155e-01 6.0328430e-01 8.2009268e-01
6.2756914e-01 -6.7839801e-01 -8.4198231e-01 8.5125342e-02
-4.1444474e-01 -1.3927116e+00 -8.0098635e-01 -4.3792492e-01
-9.8633844e-01 -1.0501392e+00 -1.0136944e+00 -6.7424226e-01
2.7341908e-01 -1.6707433e-02 1.0618856e+00 3.5638634e-02
-1.5685359e-01 7.3441935e-01 -2.5156763e-01 2.6219076e-01
-1.9269833e-01 -7.6361872e-02 1.5017845e-01 -5.1586375e-02
-1.7623253e-01 -5.4143363e-01 -8.4928799e-01 3.8377917e-01
-8.2557821e-01 -8.4424712e-02 2.0115155e-01 7.3733371e-01
5.5038071e-01 -2.9834884e-01 3.2645367e-02 -3.2704118e-01
3.6743757e-01 -3.1253424e-01 -3.4906372e-01 1.5845473e-01
-3.2220550e-02 3.2501853e-01 2.4399292e-01 -6.3847345e-01
-7.0067185e-01 3.7659919e-01 -1.3125435e-01 -9.5007616e-01
3.1099656e-01 3.7397864e-01 2.5021976e-01 -2.4702089e-01
4.6831682e-01 -7.9459190e-02 1.1191958e-02 -5.3338945e-01
5.8410215e-01 8.8572912e-03 7.4633390e-01 -3.6790961e-01
7.1109372e-01 6.5335220e-01 3.7781301e-01 -6.9412881e-01
-4.1449606e-01 -5.5951518e-01 -9.2869693e-01 -2.7473983e-01
1.2622006e+00 -9.5019656e-01 -6.6668850e-01 5.1711017e-01
-1.5802875e+00 -4.5164591e-01 7.5990252e-02 7.7467889e-01
-9.3969631e-01 8.3346170e-01 -9.8228204e-01 -6.2454182e-01
-1.9322655e-01 -1.0760347e+00 1.2778511e+00 -2.8210613e-01
-3.9474696e-01 -1.2027092e+00 5.2242243e-01 -1.0849521e-01
5.6390919e-02 9.0019953e-01 5.8891743e-01 -1.0696399e-01
-5.6680423e-01 -3.6230838e-01 1.5628661e-01 2.4616271e-01
-7.5225860e-02 6.3478954e-02 -6.9722801e-01 -6.1888450e-01
-8.8776657e-03 -1.6913602e-01 1.0871255e+00 3.8861418e-01
7.7998161e-01 -1.0456182e-01 -4.2421082e-01 8.3989060e-01
1.3443768e+00 -2.3372966e-01 9.2894810e-01 3.0400380e-01
1.0627356e+00 5.0214213e-01 6.5333164e-01 4.9467975e-01
3.8177535e-01 1.0974053e+00 4.0717548e-01 1.6942640e-01
-3.3802733e-01 -4.8318502e-01 6.0363549e-01 8.5771316e-01
-5.0578165e-01 -2.2807798e-01 -9.0542215e-01 3.7840781e-01
-2.0379004e+00 -7.3870730e-01 1.9455567e-02 2.2948151e+00
3.7925652e-01 1.8990961e-01 2.9517335e-01 -1.1566322e-01
5.0252235e-01 5.1653755e-01 -2.0313628e-01 -6.2625416e-02
2.1959263e-01 -5.6215655e-02 5.3839517e-01 6.7095619e-01
-1.3030124e+00 9.2495590e-01 6.7890282e+00 5.6880820e-01
-1.1124483e+00 5.3555544e-02 2.2389619e-01 1.8660951e-02
9.6864983e-02 -1.2606077e-01 -4.4562331e-01 2.0971653e-01
8.0564153e-01 2.0213986e-02 1.0836971e-01 5.6438529e-01
4.6571398e-01 -3.3690520e-02 -1.4049381e+00 1.0771068e+00
3.5596341e-02 -1.1542426e+00 1.4918914e-01 -2.3674828e-01
7.1833110e-01 -6.3986175e-02 -1.1624055e-01 -1.5243237e-02
-1.5946618e-01 -9.2261678e-01 8.7409693e-01 9.0028411e-01
5.3412771e-01 -6.7944014e-01 6.5246218e-01 3.4352607e-01
-1.5574172e+00 3.1837499e-01 -2.8334987e-01 8.1502069e-03
5.0877088e-01 1.7582272e-01 -5.8987612e-01 6.9996101e-01
5.5416727e-01 9.7885269e-01 -5.9078664e-01 1.1428596e+00
-2.9375148e-01 2.1977034e-01 -1.2980305e-01 2.7644670e-01
4.4082049e-01 -9.8791428e-02 7.9655814e-01 1.3389832e+00
3.1956187e-01 -3.8045609e-01 3.6794922e-01 8.5526311e-01
3.2049317e-02 -7.6398693e-02 -8.0680615e-01 -2.2361528e-02
2.3256887e-02 9.5723975e-01 -4.7840822e-01 -1.9688007e-01
-3.1138861e-01 1.3929193e+00 4.0984839e-01 5.7229614e-01
-1.1016078e+00 -1.4289159e-01 7.3694634e-01 1.8290628e-01
2.6904240e-01 -1.0810702e+00 3.2706550e-01 -1.2808723e+00
8.7560676e-02 -3.8969058e-01 2.6398870e-01 -8.1842798e-01
-1.0962260e+00 3.9810318e-01 3.1269020e-01 -1.4083416e+00
-7.2587711e-01 -6.1331552e-01 -6.8386292e-01 7.3080063e-01
-9.5965326e-01 -9.3983459e-01 -3.4418949e-01 4.9209818e-01
3.2278186e-01 2.1417247e-01 4.8950478e-01 3.3287811e-01
-4.0373209e-01 5.9999102e-01 -3.8659438e-01 4.3778428e-01
8.3367085e-01 -1.0219849e+00 9.7971290e-01 8.1322676e-01
1.3362361e-01 3.9431834e-01 7.2409099e-01 -7.3924106e-01
-1.6062030e+00 -1.1445948e+00 7.2069961e-01 -7.7685130e-01
5.8755744e-01 -2.0716152e-01 -9.1407591e-01 9.3903577e-01
2.7702443e-02 2.4427518e-01 -7.9415821e-02 -6.4981633e-01
1.2693114e-02 2.7461109e-01 -9.6366620e-01 6.6116142e-01
1.2865485e+00 -4.6999112e-01 -4.9929386e-01 8.4472477e-02
8.0268162e-01 -6.3732767e-01 -9.1703796e-01 5.9424603e-01
7.4161804e-01 -8.5767400e-01 1.2172239e+00 -6.9178373e-01
2.2550532e-01 -3.2088187e-01 -5.7895537e-02 -1.2038760e+00
-5.0811803e-01 -6.1957562e-01 -3.8228583e-01 5.7027006e-01
-1.2033005e-02 -1.4336571e-01 1.0395205e+00 2.8162614e-01
1.6145979e-01 -8.5765368e-01 -9.8721468e-01 -1.1081772e+00
-7.0633397e-02 -3.2778686e-01 -6.8919197e-02 8.3714402e-01
-2.3945594e-01 2.3960769e-01 -7.7491730e-01 -3.4717306e-02
7.4995661e-01 -3.3369577e-01 8.9170438e-01 -6.2740082e-01
-2.4840514e-01 -1.9522350e-01 -1.1223915e+00 -1.2592865e+00
4.2561254e-01 -6.9416988e-01 -5.7491265e-02 -1.5138953e+00
4.7187819e-03 4.2510870e-01 -1.2457370e-01 8.2929462e-02
2.4788212e-02 3.1269109e-01 4.1394669e-01 1.8393658e-01
-5.4346967e-01 7.8466493e-01 1.5893110e+00 -1.6010225e-01
-2.0038585e-01 -8.1764854e-02 4.9259704e-01 1.0142171e+00
4.2821327e-01 -4.0258414e-01 -2.8635532e-01 -4.8720396e-01
-1.0095358e-01 3.9121279e-01 6.2268847e-01 -1.3468878e+00
3.1133685e-01 7.9429969e-02 5.3435040e-01 -7.2514039e-01
5.2045554e-01 -7.1984869e-01 4.4231808e-01 9.1210294e-01
-2.6794216e-01 5.4284173e-01 3.4735121e-02 7.4906778e-01
-2.2889000e-01 7.8122191e-02 7.1552283e-01 -1.6834952e-01
-9.8940581e-01 4.7484967e-01 -7.5071022e-02 8.1619576e-02
8.9284110e-01 -1.0014178e-01 1.4295572e-01 -7.1047032e-01
-1.0535140e+00 1.3931036e-01 6.9681913e-01 3.3168283e-01
8.2092351e-01 -1.8724905e+00 -6.8932903e-01 -1.7610651e-01
-2.2982717e-02 -4.0583816e-01 -7.0925034e-03 1.0409337e+00
-1.0671526e+00 5.2295583e-01 -3.0248022e-01 -7.6538467e-01
-1.1151193e+00 5.9210140e-01 5.6548965e-01 -4.8779520e-01
-7.9173601e-01 4.6754673e-01 1.1686044e-01 -1.3759629e-01
4.9411479e-01 -5.4294008e-01 3.5778906e-03 -4.4820097e-01
2.8719860e-01 8.3681864e-01 -2.1241856e-01 -8.5083669e-01
-7.0803344e-01 7.9017830e-01 6.8489379e-01 -4.0594590e-01
9.7653103e-01 -1.2243849e-01 7.7257439e-02 5.6911880e-01
1.6031971e+00 -1.0598015e-01 -1.6520841e+00 -1.1702424e-01
1.1264766e-01 -4.7847253e-01 -2.8235921e-01 -3.5287961e-01
-1.0813528e+00 9.1831446e-01 8.4144282e-01 -3.1727770e-01
6.7999327e-01 -3.0057132e-01 9.2987555e-01 4.6647781e-01
5.3517342e-01 -8.7473214e-01 4.5882425e-01 5.9065372e-01
1.1831768e+00 -1.0362097e+00 1.9031422e-01 -4.1021359e-01
-4.8645797e-01 1.1558175e+00 4.8713878e-01 -6.9550121e-01
7.3263299e-01 -6.7741685e-02 5.3202424e-02 8.9135647e-02
-5.7707387e-01 -1.6572258e-01 9.1153246e-01 4.3226236e-01
6.7286289e-01 -1.1812495e-01 -5.5301756e-01 7.6833203e-02
1.6688442e-01 3.1161575e-02 2.0002420e-01 8.8675743e-01
-1.2253823e-01 -1.0568014e+00 -3.1779018e-01 -3.0493617e-01
-4.8846036e-01 3.8741180e-01 -5.1721084e-01 1.2006303e+00
-3.8938433e-02 5.2960825e-01 -8.5997514e-02 -6.2408602e-01
5.7859504e-01 1.0417821e-01 9.0648347e-01 -3.7251326e-01
-1.0628755e-01 1.6815308e-01 2.5082535e-01 -1.1944056e+00
-6.4651340e-01 -3.7728953e-01 -1.3666966e+00 -2.8928754e-01
1.9158073e-01 -1.7524195e-01 4.6701118e-01 6.9705504e-01
2.1436499e-01 4.3799821e-01 3.2561356e-01 -1.4508126e+00
-6.0366100e-01 -9.4380677e-01 -5.8406907e-01 8.9632148e-01
3.9390308e-01 -1.0225486e+00 -2.3701747e-01 7.2330691e-02] | [7.268174171447754, -0.4094148278236389] |
d37b61ea-46b6-415a-8155-9b8ed329c9a4 | dexart-benchmarking-generalizable-dexterous | 2305.05706 | null | https://arxiv.org/abs/2305.05706v1 | https://arxiv.org/pdf/2305.05706v1.pdf | DexArt: Benchmarking Generalizable Dexterous Manipulation with Articulated Objects | To enable general-purpose robots, we will require the robot to operate daily articulated objects as humans do. Current robot manipulation has heavily relied on using a parallel gripper, which restricts the robot to a limited set of objects. On the other hand, operating with a multi-finger robot hand will allow better approximation to human behavior and enable the robot to operate on diverse articulated objects. To this end, we propose a new benchmark called DexArt, which involves Dexterous manipulation with Articulated objects in a physical simulator. In our benchmark, we define multiple complex manipulation tasks, and the robot hand will need to manipulate diverse articulated objects within each task. Our main focus is to evaluate the generalizability of the learned policy on unseen articulated objects. This is very challenging given the high degrees of freedom of both hands and objects. We use Reinforcement Learning with 3D representation learning to achieve generalization. Through extensive studies, we provide new insights into how 3D representation learning affects decision making in RL with 3D point cloud inputs. More details can be found at https://www.chenbao.tech/dexart/. | ['Xiaolong Wang', 'Yuzhe Qin', 'Helin Xu', 'Chen Bao'] | 2023-05-09 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Bao_DexArt_Benchmarking_Generalizable_Dexterous_Manipulation_With_Articulated_Objects_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Bao_DexArt_Benchmarking_Generalizable_Dexterous_Manipulation_With_Articulated_Objects_CVPR_2023_paper.pdf | cvpr-2023-1 | ['robot-manipulation'] | ['robots'] | [-3.33856881e-01 1.71710867e-02 -1.86298251e-01 3.36889215e-02
-3.93401086e-02 -7.89846957e-01 1.15924433e-01 -5.79966247e-01
-3.00789446e-01 6.41224861e-01 -3.17737490e-01 -7.54409805e-02
-3.53503555e-01 -4.58581924e-01 -8.25202525e-01 -5.99976420e-01
-2.94218421e-01 8.89168143e-01 1.63695499e-01 -5.16358376e-01
2.02971622e-01 1.03999484e+00 -1.42833877e+00 4.10159305e-02
5.65967381e-01 8.14793587e-01 7.98384070e-01 5.63013554e-01
3.06205302e-01 3.88819098e-01 -6.50381327e-01 2.55388975e-01
8.27348948e-01 1.92978650e-01 -6.39766991e-01 -5.29743265e-04
-7.21981823e-02 -7.64934242e-01 -5.58693707e-01 7.45266020e-01
4.79257017e-01 5.54541290e-01 8.24979782e-01 -1.59287024e+00
-6.77103341e-01 6.71249866e-01 -3.42258841e-01 -1.74428508e-01
3.52842152e-01 7.16810763e-01 6.90693736e-01 -5.64666212e-01
5.21277547e-01 1.70537484e+00 1.96573332e-01 7.07703352e-01
-7.83216596e-01 -6.53346479e-01 4.31554168e-01 -8.16099495e-02
-9.87344623e-01 3.49893086e-02 6.57626271e-01 -5.08662820e-01
8.56087565e-01 -8.29974934e-02 4.98664409e-01 1.50019598e+00
5.56419075e-01 8.93297315e-01 9.40654516e-01 4.26803343e-02
2.77326941e-01 -4.06356603e-01 -1.93905026e-01 4.95599598e-01
5.46934903e-01 2.31090456e-01 -6.55776188e-02 1.10346183e-01
1.26865482e+00 5.13178408e-01 -2.47655451e-01 -9.20116663e-01
-1.40986562e+00 5.03686428e-01 6.92922831e-01 7.45437741e-02
-5.72479188e-01 4.22429234e-01 2.33172655e-01 4.75068688e-01
-5.18425465e-01 9.72535551e-01 -6.90355182e-01 -1.43900633e-01
1.85106426e-01 8.05369556e-01 1.12163246e+00 1.59284413e+00
4.43998098e-01 -1.14263661e-01 -1.28185242e-01 6.00510299e-01
1.18696928e-01 7.78816700e-01 1.76843613e-01 -1.46659935e+00
7.29577899e-01 3.91729087e-01 5.91261089e-01 -7.03106225e-01
-7.13160515e-01 -6.18845858e-02 -6.04365706e-01 8.17799330e-01
3.08971763e-01 -2.39480287e-01 -9.51128542e-01 1.46997631e+00
2.45538533e-01 -6.85810387e-01 -1.26693502e-01 1.29677284e+00
1.46541283e-01 3.44194263e-01 -8.98824185e-02 2.28228971e-01
9.90823269e-01 -1.09167492e+00 -6.58476770e-01 -1.21970713e-01
2.62416691e-01 -3.05145800e-01 1.42500567e+00 6.92209542e-01
-7.56314158e-01 -6.25113606e-01 -9.61471498e-01 -7.75573868e-03
-3.64086658e-01 3.34127881e-02 8.24590027e-01 -1.65384766e-02
-4.28028822e-01 8.10986757e-01 -1.19572401e+00 -3.57638806e-01
2.74715811e-01 6.00835264e-01 -2.86186010e-01 -3.42530191e-01
-8.16249907e-01 1.46801484e+00 4.95640486e-01 1.09241553e-01
-1.27849376e+00 -1.83580071e-01 -6.47619188e-01 -2.55772978e-01
8.74838293e-01 -6.42831326e-01 1.52534664e+00 -9.26610976e-02
-1.87985575e+00 3.79319310e-01 5.47735989e-01 4.92879972e-02
7.48468757e-01 -7.25146711e-01 1.53746888e-01 5.45337014e-02
3.91800739e-02 7.06517160e-01 1.01300538e+00 -1.67982244e+00
-2.93871969e-01 -3.73201489e-01 4.50088710e-01 3.86882305e-01
1.94105029e-01 -5.64389646e-01 -2.89168447e-01 -6.14271462e-01
5.58522381e-02 -1.48383820e+00 -1.67529568e-01 3.94638538e-01
-1.94618493e-01 -4.43427145e-01 1.04693174e+00 -2.44637549e-01
3.24468464e-01 -2.02996635e+00 7.15695798e-01 -5.76207088e-03
1.29643142e-01 -3.71940285e-02 -1.95234761e-01 5.69610894e-01
4.05508101e-01 -3.89088206e-02 1.36517389e-02 4.10344824e-02
2.97148943e-01 7.22276866e-01 -4.89503622e-01 3.37336123e-01
1.46002844e-01 8.92309308e-01 -1.15124536e+00 -8.91405419e-02
2.28310466e-01 4.52030040e-02 -6.36118770e-01 4.20916051e-01
-4.89165485e-01 8.11391950e-01 -1.00654602e+00 9.70189631e-01
5.25226235e-01 -3.72924283e-02 -1.33218421e-02 9.20221806e-02
1.69719663e-02 -1.48395553e-01 -9.82287407e-01 1.91029072e+00
-5.61495721e-01 -1.73408557e-02 3.55979919e-01 -7.61000693e-01
9.53513265e-01 -2.61732768e-02 5.53209603e-01 -2.44454637e-01
3.12495679e-01 2.88061261e-01 3.58523041e-01 -8.03256392e-01
3.53366941e-01 6.61514103e-02 -3.09832186e-01 2.89271861e-01
-1.93708818e-02 -9.80259180e-01 -1.67848859e-02 -3.65972310e-01
1.21789718e+00 5.84899485e-01 1.76192403e-01 -1.74232721e-01
-1.82308301e-01 1.19128801e-01 2.67745346e-01 1.08296204e+00
-3.15643221e-01 2.88094312e-01 2.57232130e-01 -3.24678183e-01
-9.43803668e-01 -1.53934729e+00 1.83294490e-01 1.18258250e+00
5.84830940e-01 3.03671926e-01 -2.06385672e-01 -5.71522474e-01
6.56570792e-01 8.10957432e-01 -4.77874160e-01 -2.48038575e-01
-8.26144993e-01 -9.10429507e-02 1.37680128e-01 6.73624039e-01
2.88181990e-01 -1.42057562e+00 -1.20550418e+00 -1.40430219e-02
2.05640644e-01 -9.62598801e-01 -2.75889367e-01 3.97762537e-01
-8.01506281e-01 -1.22023344e+00 -7.81533241e-01 -8.06497693e-01
6.30122006e-01 4.27799076e-01 5.86186230e-01 -2.13353500e-01
-3.57703388e-01 7.48722494e-01 -7.15846121e-01 -6.03927970e-01
-2.34875157e-01 2.27533519e-01 5.58598578e-01 -7.92237520e-01
-1.31508172e-01 -5.89828968e-01 -3.89257759e-01 6.29533827e-01
-6.78030670e-01 -1.40908942e-01 1.09443343e+00 6.71312749e-01
2.99265444e-01 8.55895802e-02 3.66747618e-01 -1.34530440e-01
7.33268797e-01 -5.06832182e-01 -4.87531394e-01 1.82897616e-02
7.91625157e-02 2.04954162e-01 7.17869699e-01 -1.09825015e+00
-7.55890608e-01 1.06085807e-01 3.30415845e-01 -9.37377274e-01
-1.72370538e-01 1.58222690e-01 -1.78976670e-01 -5.13988547e-02
5.61378002e-01 -1.03532858e-01 2.48267084e-01 -5.03512383e-01
3.74689817e-01 7.02667713e-01 3.60119998e-01 -1.21932518e+00
8.68874848e-01 3.04314673e-01 2.13935882e-01 -4.99564260e-01
-4.87122804e-01 -1.68709710e-01 -8.37319136e-01 -3.41403902e-01
6.31929100e-01 -5.62980592e-01 -1.40412474e+00 3.62474918e-01
-1.12101734e+00 -9.89365578e-01 -1.99824423e-01 6.65465474e-01
-1.19864905e+00 3.36654373e-02 -5.14868855e-01 -6.83806241e-01
1.49900287e-01 -1.36145341e+00 1.17407393e+00 -2.13988335e-03
-2.69505531e-01 -1.70400977e-01 -2.60721177e-01 1.38044339e-02
3.55357170e-01 4.18097079e-01 7.61360586e-01 -4.76918936e-01
-5.71618140e-01 -8.31808969e-02 1.61781654e-01 1.67218402e-01
4.50346291e-01 -4.19299126e-01 -3.60728741e-01 -7.83315659e-01
1.44767925e-01 -7.72881031e-01 3.68331045e-01 1.08889928e-02
1.51710331e+00 -9.14834067e-02 -4.51133966e-01 3.12197685e-01
8.37595284e-01 3.85288984e-01 1.42008066e-01 2.58601516e-01
7.73437083e-01 6.62065685e-01 1.15954459e+00 5.51435709e-01
9.61551219e-02 5.40988684e-01 8.30953300e-01 6.68871522e-01
1.42388999e-01 -2.24598557e-01 3.25057507e-01 5.31153023e-01
-3.54756415e-01 -3.82599086e-01 -9.77872074e-01 2.11190417e-01
-1.85552049e+00 -5.90290070e-01 4.02893484e-01 1.92017531e+00
5.19321859e-01 1.54218435e-01 2.48407200e-02 -2.47447461e-01
6.40431166e-01 -2.37879038e-01 -1.27245677e+00 -8.33855495e-02
4.95645106e-01 -6.95600808e-02 6.18419468e-01 1.38811707e-01
-7.43560731e-01 9.30055201e-01 6.05187941e+00 3.81302476e-01
-1.10338295e+00 -3.11148107e-01 -2.48257995e-01 -3.86409074e-01
3.44214290e-01 -3.77111971e-01 -6.38003767e-01 4.69086409e-01
1.18370853e-01 9.10132304e-02 8.93116355e-01 1.19686878e+00
8.65209326e-02 -8.85101259e-02 -1.57094920e+00 9.77632046e-01
-2.89206713e-01 -3.89140397e-01 2.46166259e-01 2.97247171e-02
4.46545511e-01 1.07763015e-01 2.22886458e-01 8.18351686e-01
6.24147296e-01 -1.13755929e+00 7.87621498e-01 5.76614022e-01
6.67472959e-01 -4.13131952e-01 4.29276466e-01 7.89036155e-01
-8.56058776e-01 -6.10913575e-01 -5.37134290e-01 -3.89495850e-01
7.68308192e-02 -5.42550832e-02 -8.23406339e-01 3.68953347e-01
7.86958277e-01 3.81205767e-01 1.42333940e-01 7.41959035e-01
-2.71926075e-01 -2.61153877e-01 -4.45271224e-01 -3.77223134e-01
5.23225106e-02 -7.63837472e-02 6.08550727e-01 6.05869234e-01
3.48491728e-01 3.68371159e-01 7.12876797e-01 9.39062059e-01
-6.76237326e-03 -4.60044116e-01 -8.54598820e-01 -1.08635858e-01
4.99069303e-01 9.69413280e-01 -4.88802999e-01 5.33319041e-02
-7.29011670e-02 1.04466939e+00 6.07947111e-01 4.16648358e-01
-8.18474710e-01 -5.01688063e-01 8.33248496e-01 -1.71776384e-01
3.85185599e-01 -9.76369500e-01 2.04746109e-02 -9.95804965e-01
2.15994522e-01 -1.00040185e+00 -1.26362920e-01 -9.97002125e-01
-1.43488693e+00 9.69200134e-02 3.58347148e-01 -1.34754622e+00
-1.26973569e-01 -1.23747861e+00 -1.70555279e-01 6.88221037e-01
-1.04116821e+00 -9.02158618e-01 -5.76456010e-01 4.46490049e-01
8.17064881e-01 -1.39044419e-01 5.96335769e-01 -3.28990728e-01
-1.51029602e-01 1.85189664e-01 -9.08327624e-02 2.26026885e-02
7.88153350e-01 -1.02402186e+00 2.09435657e-01 -4.70642634e-02
-6.03139937e-01 7.99238980e-01 6.37062371e-01 -8.80691588e-01
-2.14278769e+00 -8.47116590e-01 -4.27483916e-01 -7.76314020e-01
6.12588227e-01 -4.95131433e-01 -7.40821481e-01 9.19436455e-01
-2.63476044e-01 1.36101946e-01 -1.71539839e-02 -5.88580146e-02
-6.04463741e-02 2.28083760e-01 -1.25955725e+00 8.73156846e-01
1.56666481e+00 -5.60831316e-02 -9.99506831e-01 4.13796037e-01
8.71868372e-01 -9.08988118e-01 -1.02543664e+00 7.69270897e-01
8.05194199e-01 -3.08071464e-01 9.07015622e-01 -8.63514900e-01
3.81895989e-01 -4.44810867e-01 -3.34551394e-01 -1.57278478e+00
-5.62678933e-01 -5.26415765e-01 -4.25147235e-01 3.86449009e-01
-9.17144492e-02 -6.27651393e-01 5.59924543e-01 3.02821755e-01
-2.25819692e-01 -8.31559598e-01 -6.83871031e-01 -1.36746681e+00
3.35249215e-01 -4.41381820e-02 4.97038394e-01 6.28534675e-01
1.08608223e-01 2.16674674e-02 -2.68768758e-01 5.87577522e-01
4.16769415e-01 1.72313288e-01 1.07916725e+00 -1.20365334e+00
-5.73793411e-01 -4.07133788e-01 -1.26456469e-01 -1.44922972e+00
4.63828266e-01 -7.68205285e-01 6.20201766e-01 -1.36178350e+00
-2.25220807e-02 -8.06461573e-01 -8.74880478e-02 5.21494210e-01
5.99300042e-02 -4.64577436e-01 5.46685755e-01 5.01661479e-01
-5.76175869e-01 7.72082031e-01 2.09365392e+00 -2.27512121e-01
-2.18587250e-01 3.06738198e-01 -3.55959117e-01 5.80313683e-01
1.27579284e+00 -3.02822024e-01 -3.96889776e-01 -7.37041175e-01
-1.09115198e-01 1.94228977e-01 3.66994619e-01 -1.07686961e+00
1.08534358e-01 -6.64076865e-01 4.36085939e-01 -3.71819258e-01
5.66046238e-01 -1.14465034e+00 -9.75477397e-02 9.56572950e-01
-2.76360244e-01 1.45432323e-01 2.04863831e-01 7.04654515e-01
3.46486568e-01 -2.70314425e-01 5.10605693e-01 -4.88821000e-01
-6.51411235e-01 4.18423176e-01 -3.91998142e-01 -8.05654973e-02
1.29734898e+00 4.93935756e-02 -1.82596251e-01 -1.67126790e-01
-8.02294493e-01 6.79238141e-01 7.03290880e-01 8.40233803e-01
4.97524410e-01 -1.08995426e+00 -2.89552122e-01 -7.80911446e-02
7.06424490e-02 5.07923663e-01 -4.59071584e-02 3.83660465e-01
-4.82961386e-01 1.85889900e-01 -7.10977316e-01 -5.97348511e-01
-6.97036564e-01 9.12147224e-01 2.12422982e-01 1.82161033e-01
-5.35421371e-01 4.52246845e-01 1.72689497e-01 -9.70465124e-01
4.51119810e-01 -8.88947368e-01 2.39624828e-01 -6.02451324e-01
-6.14357106e-02 5.64140260e-01 -3.78218919e-01 1.10062264e-01
-1.82712451e-01 5.17492950e-01 -9.14015919e-02 7.97949880e-02
1.29729521e+00 2.02610925e-01 1.65714413e-01 6.21076584e-01
7.70763636e-01 -2.82905787e-01 -1.81376839e+00 1.34885445e-01
-2.77133971e-01 -4.82042700e-01 -5.08736670e-01 -8.19597304e-01
-5.83285928e-01 6.92459226e-01 3.57330650e-01 -1.70073748e-01
4.83845383e-01 2.33162403e-01 4.97257739e-01 1.48960614e+00
1.13316798e+00 -1.14564538e+00 6.45387232e-01 9.76058245e-01
1.71534932e+00 -1.33570850e+00 3.00999656e-02 -2.79806554e-01
-6.59371912e-01 1.10961843e+00 1.06649053e+00 -4.71421570e-01
4.29087639e-01 3.22918177e-01 -1.70720071e-01 -1.40452787e-01
-4.13717061e-01 -3.22420802e-03 -1.98967382e-02 8.01310778e-01
-2.14150056e-01 2.41444603e-01 2.59718001e-01 4.67163563e-01
-3.46818000e-01 -3.79200578e-02 3.99045289e-01 1.46930444e+00
-7.15242803e-01 -7.37819850e-01 -4.44748014e-01 4.52892184e-01
-1.34408642e-02 6.09020591e-01 -2.68119574e-01 1.09816337e+00
-4.18370627e-02 6.90531731e-01 -9.96866450e-02 -2.57396072e-01
6.65158987e-01 -2.16183528e-01 1.01618350e+00 -8.54845881e-01
-8.80997702e-02 -5.54109812e-01 -4.64554578e-01 -7.88959503e-01
-3.90696824e-02 -6.11987114e-01 -1.56079662e+00 -1.25624135e-01
-1.42240189e-02 -2.90806979e-01 7.12645471e-01 6.59293413e-01
2.92316645e-01 7.18197048e-01 3.93890321e-01 -1.82536876e+00
-1.42324054e+00 -1.10457385e+00 -6.42201245e-01 4.42920029e-01
4.70405072e-01 -1.48359847e+00 -3.06479514e-01 -4.22728121e-01] | [4.828277587890625, 0.4440566301345825] |
f841248e-7f68-4e8c-9e68-1b3f35165a6f | personalized-popular-music-generation-using | 2105.04709 | null | https://arxiv.org/abs/2105.04709v1 | https://arxiv.org/pdf/2105.04709v1.pdf | Personalized Popular Music Generation Using Imitation and Structure | Many practices have been presented in music generation recently. While stylistic music generation using deep learning techniques has became the main stream, these models still struggle to generate music with high musicality, different levels of music structure, and controllability. In addition, more application scenarios such as music therapy require imitating more specific musical styles from a few given music examples, rather than capturing the overall genre style of a large data corpus. To address requirements that challenge current deep learning methods, we propose a statistical machine learning model that is able to capture and imitate the structure, melody, chord, and bass style from a given example seed song. An evaluation using 10 pop songs shows that our new representations and methods are able to create high-quality stylistic music that is similar to a given input song. We also discuss potential uses of our approach in music evaluation and music therapy. | ['Roger B. Dannenberg', 'Ye Wang', 'Xichu Ma', 'Shuqi Dai'] | 2021-05-10 | null | null | null | null | ['music-generation', 'music-generation'] | ['audio', 'music'] | [ 2.06885651e-01 -2.51125749e-02 1.20730504e-01 -1.24256276e-02
-5.45668364e-01 -8.67586911e-01 4.45369482e-01 -2.40928322e-01
1.73356935e-01 5.62186897e-01 4.79246587e-01 2.71349490e-01
-3.98277074e-01 -8.20482671e-01 -4.96133655e-01 -3.59443575e-01
1.56383723e-01 7.60939717e-01 -3.10320735e-01 -5.91299891e-01
2.64647692e-01 5.43763578e-01 -1.59582901e+00 6.48032486e-01
5.62361181e-01 6.19169176e-01 2.28420049e-01 8.93763244e-01
-1.14601240e-01 5.87668777e-01 -1.10404038e+00 -2.09718257e-01
3.43204707e-01 -8.89118969e-01 -6.94238305e-01 -4.67938408e-02
6.14489973e-01 -1.84017047e-01 3.57039385e-02 7.46372044e-01
7.92340219e-01 1.09361291e-01 7.89285243e-01 -8.12786400e-01
-4.56549853e-01 1.24487257e+00 -2.33581334e-01 -3.41498464e-01
4.89947379e-01 1.86366588e-01 1.27565825e+00 -4.29032058e-01
6.16179228e-01 9.64204431e-01 8.69674742e-01 8.80247116e-01
-1.63057268e+00 -9.90570068e-01 -5.07839382e-01 -1.67485163e-01
-1.23878312e+00 -1.36596084e-01 1.00466657e+00 -3.65585059e-01
4.87147391e-01 3.55755866e-01 1.23044014e+00 1.20462763e+00
-5.96598722e-02 9.49983597e-01 8.23189974e-01 -4.85892266e-01
6.82122484e-02 -1.28769562e-01 -5.30775011e-01 1.97201118e-01
-1.99520633e-01 2.83914804e-01 -8.50554407e-01 -2.79637109e-02
1.17406416e+00 -5.04193127e-01 -1.96718685e-02 -2.05909342e-01
-1.43227696e+00 7.14777350e-01 2.57712722e-01 7.60208189e-01
-3.39902550e-01 4.47043955e-01 4.35655832e-01 3.18574816e-01
-1.19945742e-01 1.46624911e+00 -3.92095476e-01 -5.21701217e-01
-1.39370203e+00 9.92028773e-01 8.54670227e-01 7.54876971e-01
3.83551806e-01 6.61888957e-01 -3.50587606e-01 9.71114516e-01
-2.14337796e-01 3.92267346e-01 8.46840024e-01 -1.19275296e+00
-9.62900519e-02 4.47143823e-01 -8.39589536e-02 -8.62571597e-01
-2.96149790e-01 -7.51165211e-01 -8.08128774e-01 5.96612215e-01
5.20406723e-01 -2.99568307e-02 -6.89320683e-01 1.79048419e+00
-2.71072388e-01 2.22140074e-01 -1.59398660e-01 6.97071075e-01
7.33701706e-01 6.00487173e-01 -2.40836322e-01 2.96547753e-03
9.22093511e-01 -6.04602039e-01 -5.74270546e-01 9.26501863e-03
1.50896227e-02 -1.08810127e+00 1.44787049e+00 8.21617961e-01
-1.44296932e+00 -9.04103339e-01 -1.05827844e+00 3.10542416e-02
1.80435941e-01 2.18281001e-01 7.41001189e-01 7.14211345e-01
-6.65496409e-01 1.18952727e+00 -3.68177682e-01 6.42702403e-03
3.08822036e-01 5.51492870e-01 -9.04314145e-02 6.50556684e-01
-9.38699305e-01 5.48390865e-01 5.55650294e-01 -2.53483653e-01
-9.17083561e-01 -1.02034640e+00 -3.34517688e-01 1.01269692e-01
-1.27297759e-01 -9.75871444e-01 1.67054355e+00 -1.26020896e+00
-1.97531247e+00 7.83822715e-01 3.99424314e-01 -3.43340904e-01
2.60333896e-01 -2.71845371e-01 -3.71599436e-01 -2.70128965e-01
-1.05969928e-01 7.77412951e-01 9.00244832e-01 -1.19752836e+00
-4.62968379e-01 1.29628524e-01 -2.87922081e-02 1.42619088e-01
-4.12491888e-01 -1.33471191e-01 -1.83332995e-01 -1.19921076e+00
-6.73061907e-02 -1.05195320e+00 -8.53222162e-02 -3.36266607e-01
-6.23130560e-01 2.01614127e-02 2.91530252e-01 -2.18065634e-01
1.43125272e+00 -2.19759417e+00 5.20153701e-01 3.40977162e-01
-1.72838658e-01 2.39722773e-01 -2.49495164e-01 4.55630898e-01
-1.59821764e-01 1.19296446e-01 3.25465202e-02 -6.53698593e-02
1.08908683e-01 1.10798173e-01 -4.98254597e-01 -3.24355036e-01
-5.78578971e-02 7.95952559e-01 -8.80232573e-01 -2.16696143e-01
-3.22669148e-02 3.34448755e-01 -1.05578065e+00 1.97543621e-01
-6.07908070e-01 6.53417885e-01 -2.45556995e-01 4.58727598e-01
8.77716485e-03 1.79708645e-01 2.21926808e-01 -4.50605154e-01
1.41021654e-01 4.46382821e-01 -1.40998328e+00 2.05306196e+00
-5.90397120e-01 5.39234221e-01 -3.53116453e-01 -4.86234337e-01
1.23399055e+00 4.56583738e-01 4.94783640e-01 -3.97200674e-01
1.67585254e-01 4.77323413e-01 4.29260522e-01 -2.18529150e-01
6.57947958e-01 -7.79755771e-01 -2.87592262e-01 8.24506104e-01
1.55173510e-01 -9.25542116e-01 3.15808266e-01 -3.40366393e-01
7.70210922e-01 4.77033466e-01 1.89968392e-01 -1.37280658e-01
1.42257214e-01 4.32879999e-02 4.22887087e-01 6.72069252e-01
4.60452378e-01 9.89090741e-01 3.27369153e-01 -6.12651229e-01
-1.33818936e+00 -1.19657469e+00 1.66281193e-01 1.10037494e+00
-4.86453801e-01 -6.53499126e-01 -7.41199195e-01 -2.35034060e-02
-6.32601157e-02 6.76878273e-01 -4.99142230e-01 -2.73635030e-01
-9.16367292e-01 -4.67874825e-01 9.43750918e-01 4.08554286e-01
1.70886256e-02 -1.77974200e+00 -5.29149652e-01 5.30890703e-01
-1.57207444e-01 -4.75878686e-01 -5.88282943e-01 -1.53265763e-02
-9.76450264e-01 -8.13229561e-01 -6.27174318e-01 -9.51879978e-01
-9.64955911e-02 -4.86172169e-01 1.46339083e+00 -2.45519564e-01
-5.87924242e-01 -5.99768646e-02 -1.45180613e-01 -6.86555564e-01
-9.54571307e-01 4.25757706e-01 2.43928343e-01 -1.86850935e-01
-8.45114067e-02 -1.12250960e+00 -4.22722518e-01 -1.33545911e-02
-1.06010044e+00 4.06530559e-01 5.76101124e-01 8.36212993e-01
6.61540389e-01 1.22520208e-01 8.03154051e-01 -6.84262872e-01
1.10414457e+00 1.82691235e-02 -8.57732520e-02 -1.54553428e-01
-3.15077931e-01 1.34981960e-01 7.00818002e-01 -1.04546118e+00
-6.39757633e-01 1.49825409e-01 -2.30105191e-01 -4.32267994e-01
-6.14282005e-02 4.98189807e-01 4.28502634e-02 4.23721492e-01
1.20330608e+00 2.79261619e-01 1.23045772e-01 -6.15362763e-01
5.60669899e-01 4.18492019e-01 8.42962742e-01 -8.92639518e-01
8.84072363e-01 6.43676966e-02 -1.92673523e-02 -5.86272001e-01
-6.95932090e-01 1.19724810e-01 -6.89476430e-01 -9.57873091e-02
4.88293052e-01 -6.15723073e-01 -8.84851515e-01 4.06840950e-01
-9.34327364e-01 -3.91882420e-01 -1.06592882e+00 3.73337299e-01
-1.18500137e+00 -1.58837900e-01 -6.40125930e-01 -6.04507148e-01
-6.19326472e-01 -9.05451298e-01 1.07093477e+00 1.50629416e-01
-1.15974784e+00 -7.01812148e-01 4.58968699e-01 4.46472540e-02
4.58406448e-01 5.49316585e-01 1.09988046e+00 -3.22635353e-01
-4.55408126e-01 -2.67420709e-01 4.93414074e-01 3.22340578e-01
5.40336370e-01 1.64286733e-01 -9.42421973e-01 -7.23955706e-02
-3.79578203e-01 -5.19312561e-01 4.38559055e-01 2.73699105e-01
1.06088221e+00 -1.75822720e-01 2.47285634e-01 7.30096161e-01
1.07006633e+00 2.76076198e-01 5.46238720e-01 3.40579957e-01
6.61102891e-01 4.30506289e-01 3.02934498e-01 5.07962763e-01
-3.69501531e-01 1.06698573e+00 3.83156240e-02 1.13650449e-01
-5.48841119e-01 -7.15864956e-01 3.74833435e-01 9.20917213e-01
-3.44419867e-01 1.71050549e-01 -7.80555487e-01 5.77108085e-01
-1.56272101e+00 -1.47298074e+00 1.95027798e-01 1.91035485e+00
1.26552689e+00 2.06941590e-01 6.90724075e-01 7.97824323e-01
3.63028526e-01 -1.54495612e-01 -4.98953015e-01 -7.01573849e-01
-1.72054484e-01 1.01598728e+00 -2.40359589e-01 7.36561269e-02
-7.90852368e-01 1.11221123e+00 7.56323862e+00 9.44470584e-01
-1.32731235e+00 -4.17032331e-01 9.51341540e-02 -5.38844407e-01
-5.10513186e-01 -3.01240653e-01 -4.58513677e-01 1.77559704e-01
6.46854103e-01 -4.49397147e-01 9.13894713e-01 6.53905094e-01
2.42355824e-01 6.39356375e-01 -1.28926384e+00 1.12809038e+00
-8.54087994e-02 -1.71895099e+00 5.73419631e-01 4.97972071e-02
8.66530776e-01 -4.11547124e-01 2.66998470e-01 4.41043884e-01
2.29183778e-01 -1.56173313e+00 1.06911159e+00 6.76500142e-01
1.01225173e+00 -8.76257241e-01 2.37552106e-01 1.99307293e-01
-1.01854050e+00 -1.59850970e-01 7.82448202e-02 -2.40027472e-01
-6.91900402e-03 1.21671431e-01 -1.06595838e+00 3.83447349e-01
4.59616870e-01 6.97609961e-01 -4.40162778e-01 1.14333749e+00
-1.47780344e-01 7.15642452e-01 -2.32575536e-01 -2.23752856e-01
-5.26798368e-02 -2.26359237e-02 6.10199749e-01 1.09312892e+00
6.64844513e-01 -3.12585413e-01 -1.66573152e-02 1.23741198e+00
6.52555078e-02 3.78849477e-01 -5.23193717e-01 -5.19772649e-01
2.64521986e-01 1.06652570e+00 -3.86085540e-01 -1.31054416e-01
3.68044406e-01 8.19697857e-01 -1.89856868e-02 6.03581592e-02
-4.75059241e-01 -1.86357141e-01 7.85062730e-01 4.25024658e-01
1.14008682e-02 -2.19803855e-01 -7.01014459e-01 -8.38640809e-01
-3.91553640e-01 -1.62853587e+00 -2.57794168e-02 -9.80546176e-01
-1.24038053e+00 6.51897967e-01 -2.56648898e-01 -1.53937232e+00
-8.02699447e-01 -4.73753422e-01 -8.57222140e-01 8.96038532e-01
-4.44659203e-01 -1.24372447e+00 -8.07459354e-02 4.63230878e-01
6.33943439e-01 -7.37085044e-01 1.37706053e+00 1.81535885e-01
1.05139278e-01 5.57775676e-01 -1.39459267e-01 9.78744403e-02
6.90046012e-01 -1.46983969e+00 4.08547014e-01 5.04851900e-02
8.95022392e-01 5.79539418e-01 9.12226617e-01 -3.61306161e-01
-1.00573993e+00 -6.55051112e-01 7.19573438e-01 -4.45123285e-01
5.21973193e-01 -1.91619977e-01 -5.46005785e-01 2.78339118e-01
1.49719343e-01 -7.37777889e-01 9.63534534e-01 4.24907804e-01
-2.07950100e-01 -3.39496344e-01 -7.54716098e-01 9.54947770e-01
1.03795528e+00 -6.02729440e-01 -7.52991557e-01 3.30347009e-02
2.63133645e-01 -4.13439482e-01 -8.55955064e-01 5.25825262e-01
1.14562035e+00 -9.11400259e-01 1.02705598e+00 -7.44758368e-01
6.41775429e-01 -4.91923451e-01 -5.16095385e-03 -1.53833199e+00
-5.90383649e-01 -9.77022767e-01 1.85772125e-02 1.15927505e+00
4.41967934e-01 4.89850372e-01 1.04233193e+00 -3.93484421e-02
-1.75359659e-02 -4.10286069e-01 -5.01455724e-01 -7.73414552e-01
3.45914602e-01 -6.46986127e-01 9.61866677e-01 8.12127352e-01
-3.82256806e-02 7.04731226e-01 -6.83574975e-01 -3.33279729e-01
3.15382242e-01 6.29172802e-01 1.22838688e+00 -1.37050533e+00
-9.46114838e-01 -9.72007453e-01 -3.93119246e-01 -4.63672012e-01
4.30810172e-03 -1.20950091e+00 -1.98015898e-01 -1.40602529e+00
4.84865159e-02 -4.58403200e-01 -4.16297525e-01 4.33653146e-01
1.64823264e-01 6.04025602e-01 5.37429035e-01 1.11348957e-01
1.86032876e-01 4.23610628e-01 1.59112251e+00 -3.79823297e-01
-5.49614251e-01 3.60709339e-01 -8.74406874e-01 7.75253654e-01
9.99298871e-01 -4.09413040e-01 -4.52823609e-01 -1.95098236e-01
5.39442480e-01 1.09687097e-01 4.70561124e-02 -1.42171097e+00
-7.29831979e-02 -1.86454237e-01 5.68241358e-01 -3.83464962e-01
5.14912724e-01 -4.08258557e-01 6.84269428e-01 4.83342826e-01
-6.41732574e-01 -4.65982631e-02 4.48117912e-01 9.59252641e-02
-3.41148764e-01 -3.40057403e-01 7.61569619e-01 -2.16262951e-01
-2.75213003e-01 8.07598885e-03 -3.73385012e-01 9.54971462e-03
4.50716496e-01 -3.29450011e-01 4.30470318e-01 -5.63709855e-01
-1.23016369e+00 -5.67694664e-01 3.65537167e-01 7.30188549e-01
4.14034039e-01 -1.74936616e+00 -8.67103636e-01 3.77737492e-01
3.47000100e-02 -2.59802103e-01 -3.79671305e-02 -6.82689399e-02
-6.54725313e-01 -4.45160680e-02 -6.61334872e-01 -3.56713116e-01
-1.22145832e+00 3.07444006e-01 4.05756414e-01 -2.73126692e-01
-6.20881081e-01 6.71913207e-01 7.44628683e-02 -5.82571447e-01
1.95239246e-01 -2.71451265e-01 -3.14283103e-01 3.46956425e-03
4.17094469e-01 1.04961976e-01 -2.50346333e-01 -3.31021875e-01
1.02849133e-01 7.50800073e-01 3.31555992e-01 -5.60445905e-01
1.31187546e+00 6.77184761e-01 1.31444573e-01 8.71738493e-01
5.08433819e-01 5.09064853e-01 -9.26538825e-01 1.14141420e-01
1.23012327e-01 -2.60902047e-01 -2.44992912e-01 -1.06409049e+00
-7.71079540e-01 7.06853926e-01 4.79377478e-01 1.35802358e-01
1.15215361e+00 -2.43184254e-01 9.16333914e-01 4.20550257e-01
4.56006616e-01 -1.06401575e+00 4.83757585e-01 5.35251439e-01
1.29301870e+00 -4.58417386e-01 -2.33563349e-01 8.91684219e-02
-5.97521782e-01 1.32271171e+00 4.68677759e-01 -4.80092674e-01
5.19879103e-01 4.21026289e-01 1.31372198e-01 4.35308442e-02
-5.76296031e-01 -1.75276861e-01 5.94510078e-01 5.69153845e-01
9.05966222e-01 1.75469756e-01 -1.15957446e-01 7.23126769e-01
-1.21805859e+00 2.83594847e-01 5.10810912e-01 3.63179177e-01
-2.87431836e-01 -1.66097939e+00 -3.16165984e-01 2.17861027e-01
-6.20704532e-01 -2.00622901e-01 -6.19261563e-01 6.45723820e-01
5.85107207e-01 5.71398616e-01 6.59281239e-02 -6.74812794e-01
5.82588911e-01 1.33302093e-01 9.12736595e-01 -9.19175148e-01
-1.04296112e+00 4.28114504e-01 4.12104987e-02 -2.06189558e-01
-2.00992092e-01 -5.32403886e-01 -1.11752617e+00 -2.92959362e-01
4.67136241e-02 2.16701567e-01 6.55643523e-01 5.00476360e-01
1.54355988e-01 8.65366817e-01 4.80069190e-01 -1.14719594e+00
-3.73106748e-01 -1.16202319e+00 -9.74498868e-01 7.59409785e-01
-2.52133012e-02 -2.98752099e-01 4.78215776e-02 3.56698513e-01] | [16.061080932617188, 5.578774452209473] |
3bdf13a8-488f-4b8b-90ed-f490a2654fc9 | spark-deficient-gabor-frames-for-inverse | 2110.09296 | null | https://arxiv.org/abs/2110.09296v1 | https://arxiv.org/pdf/2110.09296v1.pdf | Spark Deficient Gabor Frames for Inverse Problems | In this paper, we apply star-Digital Gabor Transform in analysis Compressed Sensing and speech denoising. Based on assumptions on the ambient dimension, we produce a window vector that generates a spark deficient Gabor frame with many linear dependencies among its elements. We conduct computational experiments on both synthetic and real-world signals, using as baseline three Gabor transforms generated by state-of-the-art window vectors and compare their performance to star-Gabor transform. Results show that the proposed star-Gabor transform outperforms all others in all signal cases. | ['Holger Rauhut', 'Vasiliki Kouni'] | 2021-10-13 | null | null | null | null | ['speech-denoising'] | ['speech'] | [ 3.99292648e-01 -4.85099584e-01 2.57708758e-01 1.28556192e-01
-8.34744573e-01 -5.69229484e-01 2.92209923e-01 -4.12160754e-01
-2.32564941e-01 5.51388919e-01 6.19863331e-01 -2.40795016e-01
-2.10189134e-01 -7.36918449e-01 -4.80290473e-01 -8.88945520e-01
-5.54451406e-01 -5.12618423e-01 2.84554094e-01 -2.49982432e-01
6.87739477e-02 2.61579126e-01 -1.05979228e+00 2.60634750e-01
3.42398673e-01 1.16338432e+00 4.62655760e-02 1.00249791e+00
4.10800427e-01 5.91385849e-02 -7.90481508e-01 -3.01870584e-01
5.44111073e-01 -5.40026903e-01 -8.05892125e-02 1.72126099e-01
-2.52839681e-02 3.15168761e-02 -6.24452412e-01 1.25840223e+00
9.86222029e-01 3.96842062e-01 5.12910783e-01 -4.20150101e-01
-1.00013816e+00 4.86714482e-01 -3.29905957e-01 8.49982619e-01
4.95013118e-01 -2.02779591e-01 5.61706960e-01 -1.26651132e+00
3.32926691e-01 1.15682411e+00 1.17472565e+00 2.47122556e-01
-9.96132910e-01 -5.67826390e-01 -4.54286098e-01 2.59131461e-01
-1.35447860e+00 -9.75050092e-01 1.22322905e+00 8.20099656e-03
1.30789351e+00 6.18091524e-01 6.21933818e-01 1.05006170e+00
3.19403797e-01 2.73510069e-01 1.03766680e+00 -1.03594780e+00
3.06442738e-01 -6.18932188e-01 8.38476494e-02 6.42784357e-01
2.05415130e-01 3.91811013e-01 -9.35811162e-01 -4.97144729e-01
6.70594096e-01 -2.00896710e-01 -3.73269618e-01 5.10058284e-01
-1.03211308e+00 7.40525424e-01 -3.41041744e-01 6.03557587e-01
-6.47390127e-01 2.98407942e-01 2.91395217e-01 5.48587024e-01
1.07615483e+00 -7.07925707e-02 1.71363913e-02 -1.79250479e-01
-1.11276412e+00 3.02838027e-01 6.00799084e-01 5.21095395e-01
1.84968770e-01 8.86396289e-01 -5.20717382e-01 7.87430763e-01
3.09805244e-01 1.02180898e+00 7.52307832e-01 -5.87446749e-01
4.40745503e-01 -5.90576470e-01 9.91147384e-02 -1.33553243e+00
-1.06694341e-01 -7.72566736e-01 -1.01172209e+00 -3.33736628e-01
-7.66413286e-02 -2.47717589e-01 -9.56759751e-01 1.18318391e+00
1.55319467e-01 9.21670735e-01 4.70348984e-01 6.58985615e-01
8.14171672e-01 6.85326219e-01 -4.75994110e-01 -6.59431100e-01
1.29311037e+00 -5.72171688e-01 -1.36898863e+00 -1.39907420e-01
-4.73168679e-02 -1.45475340e+00 6.88921094e-01 5.62332630e-01
-9.30634141e-01 -6.30447507e-01 -1.06400979e+00 6.42634690e-01
3.43088731e-02 3.12995106e-01 7.63154924e-01 1.35190284e+00
-1.05447614e+00 4.33472067e-01 -7.91261137e-01 -7.70429596e-02
1.42046854e-01 -1.33176014e-01 -2.67389834e-01 1.25781372e-01
-1.17248666e+00 3.94790500e-01 1.01058260e-01 3.77724558e-01
-6.68103933e-01 -4.40663457e-01 -9.21847939e-01 -4.97344434e-02
-3.36553268e-02 -3.40589195e-01 9.41480219e-01 1.05741680e-01
-1.46587670e+00 3.84912252e-01 -5.25453091e-01 -8.58112693e-01
8.03030878e-02 -1.69968158e-01 -1.00147533e+00 2.89840609e-01
-8.83266330e-02 -5.24801552e-01 1.66845453e+00 -7.24613547e-01
-1.04113743e-01 -2.80787081e-01 -4.68532413e-01 -1.54929057e-01
-4.36469227e-01 7.57301822e-02 -1.70552917e-02 -1.48037922e+00
6.83876038e-01 -5.48180223e-01 -1.14786632e-01 -6.25524402e-01
-3.71261954e-01 -2.07596030e-02 1.13864815e+00 -1.00696623e+00
1.56342268e+00 -2.48689461e+00 -2.63034493e-01 2.72969782e-01
-1.53686672e-01 4.83396202e-01 -6.77973107e-02 7.58268118e-01
-1.64818957e-01 -2.14415900e-02 9.30805802e-02 -6.80673122e-01
-1.21968932e-01 2.12332845e-01 -8.17501903e-01 9.32321072e-01
-1.40881985e-01 3.80584151e-01 -4.57019478e-01 -3.42503756e-01
2.03057215e-01 7.41991282e-01 -5.91534555e-01 -1.46320593e-02
2.97110349e-01 3.45001787e-01 -6.20206714e-01 8.71894598e-01
8.76993120e-01 4.90308523e-01 -3.16048384e-01 -6.72209084e-01
5.63249663e-02 -1.14927001e-01 -1.27401614e+00 1.70056510e+00
-4.85663503e-01 6.27900898e-01 -1.95168667e-02 -1.34956384e+00
1.39841115e+00 7.15524077e-01 3.70392412e-01 -4.56060261e-01
3.29334172e-03 2.59554893e-01 -2.39198893e-01 -8.56507480e-01
2.34889776e-01 -2.81780809e-01 1.95774049e-01 3.63038093e-01
3.76246035e-01 -3.60716045e-01 -7.40464628e-02 -2.67493099e-01
1.26217818e+00 -3.31096709e-01 1.31976053e-01 -1.52093202e-01
5.54814458e-01 -6.06267512e-01 4.81528342e-01 9.67540383e-01
-2.37948075e-01 7.75828660e-01 -2.03102008e-01 -4.68779266e-01
-9.46574926e-01 -1.00488770e+00 6.95470721e-02 9.03591812e-01
-1.24474473e-01 -4.74119365e-01 -6.70026660e-01 1.07229225e-01
-3.60447258e-01 3.91518861e-01 -4.17926431e-01 -1.30479276e-01
-7.17409194e-01 -9.09259200e-01 1.08208644e+00 2.39076331e-01
7.50267863e-01 -7.54515171e-01 -7.52380848e-01 3.74859780e-01
-1.82951987e-01 -1.36720908e+00 -7.13783383e-01 5.61315343e-02
-6.35896146e-01 -6.75218999e-01 -6.73732996e-01 -8.57231736e-01
1.32283136e-01 5.31477988e-01 7.57676184e-01 -2.28060231e-01
-2.96785772e-01 4.12216544e-01 -9.28657115e-01 -6.14472866e-01
-8.16787779e-02 -5.93601346e-01 2.90823102e-01 3.85131478e-01
-1.09362513e-01 -1.08498800e+00 -6.33025408e-01 2.95084156e-03
-9.14972126e-01 -5.56890965e-01 1.38516322e-01 1.05910659e+00
5.96440315e-01 6.30905032e-01 5.05086601e-01 -3.28759283e-01
1.28411841e+00 -6.02945350e-02 -3.99400830e-01 1.70686752e-01
-2.95596063e-01 -2.97586918e-02 4.41028237e-01 -7.03819990e-01
-8.68936360e-01 -2.06722021e-01 -4.53623503e-01 -5.81257343e-01
4.70953614e-01 7.26652622e-01 2.66098410e-01 -3.19615573e-01
8.13399851e-01 6.66282952e-01 -1.71546221e-01 -6.49908960e-01
2.75391281e-01 6.53659403e-01 8.18213761e-01 -6.75070047e-01
9.83364463e-01 5.28846622e-01 7.11641684e-02 -1.43134582e+00
-5.03727555e-01 -3.85966390e-01 5.85132558e-03 -3.50295827e-02
6.05905116e-01 -6.43471777e-01 -5.47269881e-01 4.87414956e-01
-1.42729044e+00 2.00689420e-01 -4.93998408e-01 8.84864390e-01
-4.39240038e-01 4.42969471e-01 -5.33381999e-01 -1.29715180e+00
-7.50556529e-01 -9.77973521e-01 1.17570031e+00 -4.80890088e-02
2.21119255e-01 -8.64223361e-01 2.23386407e-01 -2.08888605e-01
9.02674913e-01 2.36684859e-01 2.22608969e-01 -5.41404665e-01
5.87939695e-02 -6.02178991e-01 4.15222168e-01 4.45992023e-01
3.90223920e-01 -4.16701972e-01 -1.11216760e+00 -4.11325574e-01
6.57298207e-01 2.39668489e-01 9.67122078e-01 7.62022495e-01
1.00126660e+00 -5.93974650e-01 -8.41737390e-02 1.01640165e+00
1.18506289e+00 5.14818668e-01 5.88453889e-01 9.79441181e-02
3.47661339e-02 -8.63380432e-02 6.88099146e-01 5.63836813e-01
-4.76485491e-01 5.73066771e-01 1.74514711e-01 8.51030424e-02
-3.01282257e-01 1.36112738e-02 2.60304302e-01 1.15567636e+00
-4.17479187e-01 -2.12198183e-01 -3.59844148e-01 4.39703345e-01
-1.56882524e+00 -1.42194116e+00 2.78480053e-01 1.84835911e+00
5.72197258e-01 -6.03493266e-02 -3.72697785e-03 7.63320208e-01
5.62020004e-01 7.70342052e-01 -1.28329083e-01 -2.76719868e-01
-5.26002407e-01 1.02359056e+00 6.43050313e-01 6.29091501e-01
-1.23525023e+00 8.98762882e-01 7.77177477e+00 1.20808148e+00
-1.15773630e+00 5.07398307e-01 3.75714272e-01 4.34222609e-01
-2.95769244e-01 -3.13445389e-01 -4.22619611e-01 7.89702237e-02
6.29353046e-01 -5.23517668e-01 6.61671638e-01 4.65679586e-01
3.57628733e-01 3.19861501e-01 -2.47364908e-01 1.56145537e+00
2.11771995e-01 -1.42437947e+00 -3.61437023e-01 -2.91785955e-01
5.31875372e-01 -3.20263892e-01 4.29125577e-01 -2.24010408e-01
-1.84834793e-01 -1.16546845e+00 8.12124014e-01 5.07967889e-01
1.00929344e+00 -6.14286602e-01 7.04489946e-01 2.57577866e-01
-1.44467974e+00 1.93194300e-01 -5.78479171e-01 -1.33827120e-01
1.99387204e-02 9.14668024e-01 -7.53595769e-01 6.49464011e-01
7.16179729e-01 3.53053302e-01 -8.75917822e-02 7.73915708e-01
1.46829918e-01 1.21912658e+00 -4.73783016e-01 5.26782796e-02
1.11358296e-02 -1.89898834e-01 8.46438766e-01 1.50088561e+00
9.42620456e-01 3.96790624e-01 1.47870898e-01 3.57310623e-01
5.08877277e-01 -1.28844574e-01 -8.18958700e-01 -9.44588557e-02
6.90650523e-01 7.38267422e-01 -7.10134387e-01 -1.84647679e-01
-3.30910027e-01 8.95830095e-01 -8.25647056e-01 6.10936463e-01
-5.70474386e-01 -5.73749781e-01 5.47077954e-01 -1.37481794e-01
7.44498909e-01 -7.14298308e-01 -8.99483114e-02 -1.04257607e+00
1.87884673e-01 -9.06230688e-01 3.55134606e-02 -4.22590941e-01
-1.15188158e+00 1.08294475e+00 -5.34529723e-02 -1.36957669e+00
-3.66805702e-01 -4.21294689e-01 -7.11140335e-01 1.08694804e+00
-1.04012179e+00 -1.14264345e+00 4.27189879e-02 1.09196651e+00
5.60481250e-01 -6.79219425e-01 1.28114223e+00 2.42965624e-01
6.28819317e-02 5.45572340e-01 1.31390378e-01 4.41530868e-02
2.21971765e-01 -4.50177848e-01 8.45550835e-01 1.38390791e+00
5.46328366e-01 6.46376610e-01 1.41014469e+00 -5.73376536e-01
-1.72715247e+00 -8.36473584e-01 4.65440720e-01 2.76623696e-01
4.87608850e-01 -1.63684741e-01 -4.30972904e-01 6.25270903e-01
1.34527951e-01 7.50247538e-01 7.17239082e-01 -3.45692843e-01
-2.86257893e-01 -1.37099355e-01 -1.45187151e+00 9.59476680e-02
1.22873938e+00 -7.12827563e-01 -7.48856008e-01 5.21006823e-01
9.97352302e-01 -6.41079426e-01 -6.40812397e-01 4.58142489e-01
6.24335527e-01 -8.62432063e-01 1.67402589e+00 -4.52081800e-01
-3.52452606e-01 -3.30155581e-01 -9.85307574e-01 -1.52738678e+00
-2.33366877e-01 -1.11755586e+00 -1.80368528e-01 1.09659350e+00
-4.15934389e-03 -1.01656342e+00 7.06638992e-01 -5.09274364e-01
-2.99492031e-01 -5.92393696e-01 -1.52093649e+00 -1.16528487e+00
-4.92302239e-01 -9.58146155e-01 6.66199028e-01 7.32941091e-01
-2.21304253e-01 5.80532253e-02 -8.72784317e-01 4.30622905e-01
8.85105193e-01 -3.55347991e-02 3.62353355e-01 -7.70616353e-01
-7.57142544e-01 2.29382396e-01 -7.49069571e-01 -7.30505824e-01
2.03361511e-02 -4.71833855e-01 -1.89375132e-01 -1.18909204e+00
-6.69442713e-01 -1.39050514e-01 -1.99486464e-01 9.28475037e-02
6.30423054e-02 7.18572676e-01 8.52575600e-02 -6.64570704e-02
5.06368130e-02 6.12728894e-01 1.05704868e+00 -3.33797216e-01
-2.52587944e-01 3.73804212e-01 -4.36215311e-01 6.80283964e-01
7.22750843e-01 -4.22028929e-01 -4.68422294e-01 -3.54181081e-01
-6.98349625e-02 4.19777930e-01 2.75365055e-01 -1.31464231e+00
2.22778112e-01 7.42441192e-02 3.11413556e-01 -4.94355172e-01
8.89568686e-01 -5.43292940e-01 2.99550086e-01 5.09434402e-01
1.04136683e-01 1.93872720e-01 2.02949673e-01 5.23564696e-01
-4.55342054e-01 -1.14456177e-01 3.79134923e-01 1.97604876e-02
-1.89189449e-01 1.62515566e-01 -2.10604355e-01 -6.81229085e-02
5.05173683e-01 -2.72892624e-01 -2.83956259e-01 -6.98592126e-01
-9.68818069e-01 -9.32204366e-01 -4.15663153e-01 -1.42693520e-01
1.13551891e+00 -1.52123821e+00 -9.29439306e-01 7.41950750e-01
-4.00388539e-01 -4.56740826e-01 1.17723502e-01 5.45515001e-01
-6.02622569e-01 4.57166523e-01 9.23388302e-02 -4.48155969e-01
-1.48421609e+00 2.74044931e-01 1.41149685e-01 -1.07082531e-01
-7.82982886e-01 1.06593764e+00 -4.76292312e-01 1.81004420e-01
4.19642515e-02 -4.20811057e-01 -7.57783800e-02 -3.84704083e-01
8.46428692e-01 2.95568585e-01 3.86084795e-01 -7.11019099e-01
-4.51177478e-01 8.20851624e-01 6.44191980e-01 -6.43245697e-01
1.54818702e+00 1.65790215e-01 2.35730154e-03 1.87773019e-01
1.11293304e+00 7.52828479e-01 -5.48251808e-01 -3.92936379e-01
-4.53584373e-01 -1.00089824e+00 1.34362057e-01 -6.06588498e-02
-9.57073271e-01 5.40446877e-01 1.11825180e+00 7.19949782e-01
1.56670380e+00 -2.31102258e-01 7.12571979e-01 3.83193791e-01
4.51299876e-01 -6.07851565e-01 3.20791930e-01 3.88205051e-01
1.03421223e+00 -5.68898678e-01 8.22014809e-02 -6.68966413e-01
-4.43848483e-02 1.17198467e+00 -4.14301425e-01 -6.04304194e-01
1.09717166e+00 5.17608702e-01 1.02186061e-01 1.83198918e-02
-2.17510372e-01 -1.30347595e-01 2.43705288e-01 1.11655939e+00
3.59209687e-01 1.87949061e-01 -6.42453909e-01 6.94771111e-01
-6.66053176e-01 -1.07285656e-01 1.28231138e-01 8.14798295e-01
-4.84573692e-01 -1.09099638e+00 -1.20570838e+00 3.15106869e-01
-8.78552437e-01 -3.76629233e-01 4.03022200e-01 3.17210466e-01
-8.90559554e-02 1.37076986e+00 -1.25016093e-01 -6.98878169e-01
4.62784559e-01 -4.25619818e-02 6.41916215e-01 -1.87674165e-01
-1.22585215e-01 8.10460806e-01 1.62384361e-01 -2.10260183e-01
-5.69491267e-01 -5.16908526e-01 -7.54792571e-01 -2.07287803e-01
-4.76642996e-01 3.68462116e-01 7.84069896e-01 8.10855925e-01
6.96537271e-02 8.03381622e-01 7.05284715e-01 -1.01782703e+00
-1.05788720e+00 -1.27327049e+00 -8.61270189e-01 7.73030892e-02
7.51486719e-01 -5.39843023e-01 -4.86792594e-01 2.95061499e-01] | [15.386903762817383, 5.584723949432373] |
5dfbe222-fe8f-4a3d-b5ac-2228496bbd8b | graphsearchnet-enhancing-gnns-via-capturing | 2111.02671 | null | https://arxiv.org/abs/2111.02671v5 | https://arxiv.org/pdf/2111.02671v5.pdf | GraphSearchNet: Enhancing GNNs via Capturing Global Dependencies for Semantic Code Search | Code search aims to retrieve accurate code snippets based on a natural language query to improve software productivity and quality. With the massive amount of available programs such as (on GitHub or Stack Overflow), identifying and localizing the precise code is critical for the software developers. In addition, Deep learning has recently been widely applied to different code-related scenarios, e.g., vulnerability detection, source code summarization. However, automated deep code search is still challenging since it requires a high-level semantic mapping between code and natural language queries. Most existing deep learning-based approaches for code search rely on the sequential text i.e., feeding the program and the query as a flat sequence of tokens to learn the program semantics while the structural information is not fully considered. Furthermore, the widely adopted Graph Neural Networks (GNNs) have proved their effectiveness in learning program semantics, however, they also suffer the problem of capturing the global dependencies in the constructed graph, which limits the model learning capacity. To address these challenges, in this paper, we design a novel neural network framework, named GraphSearchNet, to enable an effective and accurate source code search by jointly learning the rich semantics of both source code and natural language queries. Specifically, we propose to construct graphs for the source code and queries with bidirectional GGNN (BiGGNN) to capture the local structural information of the source code and queries. Furthermore, we enhance BiGGNN by utilizing the multi-head attention module to supplement the global dependencies that BiGGNN missed to improve the model learning capacity. The extensive experiments on Java and Python programming language from the public benchmark CodeSearchNet confirm that GraphSearchNet outperforms current state-of-the-art works. | ['Guozhu Meng', 'Yang Liu', 'JingKai Siow', 'Lei Ma', 'Xiaofei Xie', 'Shangqing Liu'] | 2021-11-04 | null | null | null | null | ['code-summarization', 'code-search', 'code-search', 'vulnerability-detection'] | ['computer-code', 'computer-code', 'computer-vision', 'miscellaneous'] | [-2.2315912e-01 -2.3898175e-01 -5.9157133e-01 -1.8462820e-01
-7.0156968e-01 -5.5486184e-01 3.9371572e-02 4.9484310e-01
-7.5344056e-02 9.6469551e-02 2.1643403e-01 -5.5435288e-01
-1.2760081e-02 -8.4436893e-01 -8.1894964e-01 -2.1682234e-01
2.4324618e-03 -7.7395737e-02 4.9049994e-01 -1.5836746e-01
4.8876905e-01 -2.1224043e-01 -1.5149578e+00 1.9559254e-01
1.3397117e+00 8.1916690e-01 6.8811226e-01 3.5211274e-01
-8.9255571e-01 1.1403012e+00 -4.7112828e-01 -4.8004919e-01
-2.2214687e-01 -2.3883590e-01 -9.6982557e-01 -7.1062475e-01
2.6228201e-01 -2.4210910e-01 -3.4963402e-01 1.6211619e+00
3.3536902e-01 -1.9287989e-01 -8.0462188e-02 -1.1948957e+00
-8.6007947e-01 1.0548570e+00 -6.8092942e-01 2.0633127e-01
5.1439416e-01 1.4910789e-01 1.4002516e+00 -6.9252646e-01
5.0225437e-01 1.0895076e+00 6.8660861e-01 3.8748431e-01
-6.7803830e-01 -6.5715688e-01 3.1585756e-01 4.6317279e-01
-1.3141838e+00 -1.0965635e-01 9.0902913e-01 -5.7884663e-01
1.3238181e+00 1.2197993e-01 4.1581297e-01 8.7101436e-01
2.3039719e-01 7.3024619e-01 1.7049736e-01 -1.8309297e-01
-5.6977575e-03 -1.5188743e-01 5.7113379e-01 1.1201775e+00
2.1405075e-01 -2.2187871e-01 2.2853812e-02 -5.8138728e-01
3.8164207e-01 3.2783750e-01 -3.5739201e-01 -3.4460509e-01
-8.4101182e-01 8.2486248e-01 9.1382974e-01 4.6732932e-01
-1.2005470e-01 5.2598989e-01 9.2295092e-01 2.5044277e-01
1.3793790e-01 3.1081730e-01 -4.8504168e-01 -1.8465699e-01
-7.1535838e-01 1.3523589e-01 9.3622404e-01 1.2840501e+00
1.1700268e+00 1.4633082e-01 -1.7389682e-01 8.8249540e-01
6.2264270e-01 3.5742062e-01 7.0170701e-01 -2.3588099e-01
1.0245934e+00 1.4704192e+00 -4.6366772e-01 -1.4562283e+00
-1.8496300e-01 -6.0234421e-01 -6.9462842e-01 -2.4231626e-01
-6.9518410e-02 2.3780139e-01 -7.5955874e-01 1.6456113e+00
4.2713847e-02 -6.2290870e-02 -2.9354664e-02 7.5780439e-01
1.2618130e+00 6.3303214e-01 2.2394426e-01 2.3800120e-01
1.1950619e+00 -1.0930585e+00 -2.9711562e-01 -5.9612596e-01
1.0398787e+00 -5.9451938e-01 1.3664684e+00 -1.2021700e-01
-5.8279198e-01 -5.9877431e-01 -8.6394739e-01 -3.0284336e-01
-4.0400946e-01 1.8279342e-01 6.4290422e-01 3.3018607e-01
-1.0497346e+00 2.8133816e-01 -7.8270507e-01 -2.7752471e-01
4.3681267e-01 3.4203765e-01 -1.1103999e-01 -2.7199644e-01
-1.2062274e+00 2.7852327e-01 7.4031711e-01 9.7092897e-02
-1.0000261e+00 -6.0511589e-01 -1.0809292e+00 7.1498358e-01
6.7062962e-01 -3.7679523e-01 1.0860604e+00 -1.0603992e+00
-7.9478073e-01 7.2801429e-01 -3.1352494e-02 -2.0523797e-01
4.3657217e-02 -4.0931784e-02 -3.0918527e-01 -1.6144431e-01
2.4457251e-01 -4.2236775e-02 4.4368112e-01 -9.0615416e-01
-4.2054990e-01 -3.3844623e-01 5.0474614e-01 -2.5543496e-01
-5.1884401e-01 3.3490840e-01 -7.9804146e-01 -5.7235861e-01
-1.7028740e-01 -6.0255146e-01 -9.7111285e-02 -4.7871092e-01
-5.6570345e-01 -6.5308744e-01 6.5073162e-01 -1.0793784e+00
1.9126879e+00 -2.2718883e+00 2.1049412e-01 9.9172339e-02
6.5993363e-01 4.0636882e-01 -2.3863155e-01 4.8715255e-01
-2.1972915e-02 4.2599279e-01 -2.9708448e-01 1.6093309e-01
1.7736518e-01 -2.9570848e-02 -3.3909318e-01 1.3981065e-01
-1.8279249e-03 1.2950487e+00 -1.0710400e+00 -5.9376293e-01
-3.0671254e-01 4.0468365e-02 -7.8423923e-01 3.0952284e-01
-6.1276686e-01 -1.3822043e-01 -9.5262659e-01 7.3424166e-01
5.1576060e-01 -6.3895351e-01 -3.0665597e-02 6.7839369e-02
1.4145843e-02 3.7587941e-01 -6.7074519e-01 1.8833078e+00
-7.2152990e-01 3.5432151e-01 2.6996970e-02 -1.1014813e+00
1.0352591e+00 1.1789803e-01 8.3947264e-02 -9.2188865e-01
-1.5701658e-01 4.9718451e-01 -1.5059106e-01 -1.0608381e+00
1.6774473e-01 4.6657994e-01 -4.6405002e-01 5.0780749e-01
2.6884085e-02 2.7563253e-01 1.3825370e-01 4.3075484e-01
1.4123285e+00 1.7366907e-01 2.9954696e-01 -2.5415042e-01
9.3411839e-01 1.0064920e-01 5.5693835e-01 7.9560256e-01
1.9965662e-01 1.7725091e-01 7.8638089e-01 -4.8983228e-01
-6.6020423e-01 -4.3940303e-01 4.9223250e-01 1.3771433e+00
2.9748884e-01 -7.2651207e-01 -1.0021564e+00 -1.0752542e+00
-9.3148753e-02 4.0555537e-01 -3.4994981e-01 -5.8897096e-01
-8.8153267e-01 -4.9289653e-01 6.7229211e-01 5.8218074e-01
5.5043894e-01 -1.4020449e+00 -4.4065100e-01 2.5622535e-01
-2.2081405e-01 -7.8995407e-01 -8.7571514e-01 -1.1023764e-01
-5.5025643e-01 -1.4232183e+00 -3.9388484e-01 -1.1352434e+00
5.8238816e-01 1.5221170e-01 1.4462303e+00 9.9582356e-01
-2.1775836e-01 1.0565061e-01 -5.9633589e-01 1.0542723e-01
-5.2815878e-01 5.9402800e-01 -6.9952637e-01 -3.9454114e-01
5.4886234e-01 -4.9400777e-01 -3.9249077e-01 4.8446219e-02
-1.1265979e+00 -7.5409219e-02 6.6181415e-01 7.6794869e-01
2.3788710e-01 1.2192799e-02 6.0602039e-01 -9.1708118e-01
8.2599264e-01 -9.7195774e-01 -9.6816027e-01 5.4839486e-01
-6.5327430e-01 3.4120759e-01 1.1087433e+00 -1.6653797e-01
-9.1586041e-01 -2.0599960e-01 -3.9626044e-01 -3.5087729e-01
1.7308493e-01 1.3554623e+00 -2.5581369e-01 -2.3586507e-01
6.1867917e-01 5.4853088e-01 -3.2514220e-01 -7.5300699e-01
9.1211520e-02 7.1688449e-01 4.5549121e-01 -7.1749973e-01
7.3109889e-01 -1.6829222e-01 -4.3563232e-01 -2.3170309e-01
-4.7535554e-01 -5.7840073e-01 -2.4880138e-01 2.4136379e-01
7.5914478e-01 -6.0877287e-01 -7.1252626e-01 5.5524623e-01
-1.4380109e+00 -1.8004130e-01 3.4418219e-01 2.5235578e-02
-2.0791028e-01 6.3821524e-01 -4.2966062e-01 -3.8310945e-01
-6.6290230e-01 -1.5723088e+00 1.0272446e+00 3.1154728e-01
2.4784668e-01 -1.0637562e+00 -6.5711974e-03 7.1825415e-02
5.5654514e-01 1.5291793e-02 1.6602597e+00 -7.8042418e-01
-1.0096033e+00 -2.5110492e-01 -5.6763160e-01 1.2241836e-01
1.7617804e-01 -1.1398569e-01 -5.6149983e-01 -3.0946368e-01
-7.8953348e-02 -2.6407123e-01 8.0461621e-01 -4.0285274e-02
1.4145114e+00 -6.4362043e-01 -4.3302774e-01 8.3449996e-01
1.8431474e+00 1.9277722e-01 4.4908679e-01 3.5783049e-01
1.3257190e+00 4.1659310e-01 1.6725735e-01 2.7518591e-01
6.7031884e-01 4.5246267e-01 9.4654089e-01 2.6198283e-01
-6.1225113e-03 -5.5089766e-01 2.8183249e-01 1.0834714e+00
4.5501524e-01 -1.3358448e-01 -1.3638790e+00 6.9979143e-01
-2.0000505e+00 -6.5164542e-01 -7.4316673e-02 2.0521376e+00
8.8549435e-01 -9.9467829e-02 -1.9928686e-01 -3.9588273e-01
7.5274700e-01 2.7095732e-01 -7.9773772e-01 -1.2239189e-01
3.0668199e-01 -2.4131898e-02 3.6844537e-01 3.1079286e-01
-7.7306569e-01 9.7159362e-01 4.5542750e+00 1.0034882e+00
-1.1308620e+00 1.4677677e-01 1.1410730e-01 4.8147881e-01
-7.8710324e-01 2.6731825e-01 -6.9161713e-01 6.7818415e-01
7.0899606e-01 -6.1735207e-01 7.0015073e-01 1.2815547e+00
-1.9487004e-01 2.5766701e-01 -9.8166263e-01 9.5231980e-01
-8.3156191e-03 -1.2834951e+00 4.3342229e-02 -2.8620914e-01
4.9646634e-01 3.8318160e-01 -3.5876945e-01 8.0524713e-01
3.9209265e-01 -9.0535635e-01 6.2729859e-01 3.6480492e-01
7.0213276e-01 -4.8784032e-01 8.4867179e-01 4.8532504e-01
-1.6550275e+00 -5.1522124e-01 -4.6361667e-01 3.1935069e-01
-4.8207146e-01 4.0328214e-01 -5.5825305e-01 8.3415473e-01
8.3048624e-01 9.4673872e-01 -1.0812182e+00 1.2619144e+00
-1.9713180e-01 5.5434459e-01 1.9302625e-02 -3.1103930e-01
4.4290197e-01 1.0650197e-01 3.2981083e-01 1.1810613e+00
4.3589598e-01 -3.8635111e-01 4.9762326e-01 1.5705727e+00
-3.6909547e-01 3.4941769e-01 -5.1497960e-01 -3.7359411e-01
3.7054914e-01 9.9466115e-01 -4.8131940e-01 -2.4400336e-01
-9.2755371e-01 5.0984198e-01 6.5281940e-01 5.4223919e-01
-6.9785303e-01 -9.8293006e-01 4.1661024e-01 -7.3551148e-02
2.1674271e-01 2.4178000e-02 7.2782457e-02 -1.3165325e+00
6.1805129e-01 -1.0881382e+00 6.1505908e-01 -6.1433214e-01
-9.9222451e-01 7.4245787e-01 -2.4208892e-02 -9.5879495e-01
-2.4529235e-01 -2.8360394e-01 -8.7280345e-01 1.1225158e+00
-1.6146929e+00 -1.1156124e+00 -6.4627951e-01 4.1470233e-01
4.9759772e-01 -2.9750124e-01 5.1114756e-01 5.6659192e-01
-6.7877203e-01 7.5346923e-01 -5.9888568e-02 5.7171375e-01
2.1355709e-01 -1.1119283e+00 8.6332583e-01 9.6819472e-01
-8.7907165e-02 1.1633738e+00 2.1011674e-01 -8.3805150e-01
-1.7230475e+00 -1.1887891e+00 7.1947300e-01 -9.3579061e-02
8.6270571e-01 -4.1631997e-01 -1.5022231e+00 6.4311957e-01
-1.2685343e-02 1.7078115e-01 1.5589698e-01 -6.1714735e-02
-6.3593763e-01 -6.4483956e-02 -6.3407111e-01 3.2605261e-01
1.0962087e+00 -8.8108706e-01 -3.9977053e-01 2.4898835e-01
1.2657741e+00 -4.5981175e-01 -6.1909741e-01 2.0276627e-01
2.0182700e-01 -8.5986817e-01 7.8054553e-01 -5.6831849e-01
5.4082602e-01 -3.0154505e-01 3.4919240e-02 -9.9389523e-01
-9.2148252e-02 -4.2978412e-01 -2.2766578e-01 1.3781763e+00
2.0389445e-01 -7.4336886e-01 6.7230105e-01 3.9287347e-01
-3.3458295e-01 -7.7862060e-01 -5.9896559e-01 -5.2091312e-01
-6.6779241e-02 -3.4655273e-01 1.0563847e+00 8.5397315e-01
5.6932613e-02 1.8324688e-01 -5.3805966e-02 1.0484845e-01
4.6642044e-01 5.5926687e-01 5.6189376e-01 -1.3175273e+00
-3.9034772e-01 -7.9014385e-01 -4.2647499e-01 -1.0808827e+00
6.8402714e-01 -1.3542376e+00 -1.1134948e-02 -1.6256657e+00
4.7063151e-01 -4.4435063e-01 -3.4434238e-01 7.3189610e-01
-6.1526620e-01 -6.5459073e-01 -1.3514510e-01 3.8860351e-01
-7.3511225e-01 4.4168785e-01 7.5690222e-01 -6.0257918e-01
-1.2507518e-01 -3.1542577e-02 -8.8945985e-01 4.0347990e-01
5.2794963e-01 -6.9925839e-01 -5.9922808e-01 -9.5625299e-01
7.8898466e-01 2.6951683e-01 5.7499683e-01 -7.4092674e-01
5.7190335e-01 -8.1423886e-02 -4.2788154e-01 -2.9728043e-01
-6.2286502e-01 -7.3909169e-01 -3.7355829e-02 4.9760941e-01
-3.6021388e-01 2.5681713e-01 3.0988204e-01 6.7048115e-01
-5.2869171e-01 -6.8100011e-01 4.1686186e-01 -5.0751591e-01
-1.1808729e+00 6.4543605e-01 1.8768415e-01 4.8959753e-01
5.3565264e-01 9.1503285e-02 -6.7696530e-01 5.2795608e-02
-1.3007595e-01 5.9738022e-01 6.0477388e-01 8.7060171e-01
5.8410722e-01 -1.2700396e+00 -4.7667453e-01 2.3011242e-01
5.9448332e-01 9.0586230e-02 3.2942501e-01 6.1717194e-01
-6.4632517e-01 4.8627305e-01 8.9062698e-02 -3.4630632e-01
-1.1293848e+00 8.3368307e-01 4.2488781e-01 -4.2647320e-01
-5.1419687e-01 7.7677703e-01 3.9345387e-01 -6.7070836e-01
2.9209861e-01 -5.2739841e-01 -4.5529670e-01 -3.5550678e-01
4.7701812e-01 1.1993559e-01 1.3697052e-01 -4.2902714e-01
-5.5150390e-01 6.8112165e-01 -1.5168124e-01 7.9194319e-01
1.2456720e+00 4.2191956e-02 -8.3712041e-01 -2.2453269e-02
1.6135676e+00 2.1353273e-03 -7.7885199e-01 -5.0906205e-01
5.2035558e-01 -3.7327719e-01 9.4882119e-03 -5.2050668e-01
-1.3750626e+00 9.9853116e-01 2.6335937e-01 3.6853787e-01
1.0386922e+00 1.8335208e-01 9.0887851e-01 5.6957972e-01
5.4274648e-01 -6.1648059e-01 -7.4985293e-03 5.9583157e-01
6.7629141e-01 -1.1866798e+00 -3.4367254e-01 -1.5521096e-01
-6.6515677e-02 1.2289523e+00 9.4260490e-01 7.3196471e-02
4.3837512e-01 1.1391988e-01 -1.9033422e-01 -5.7031834e-01
-5.2233040e-01 -6.5756373e-02 3.3333188e-01 3.6811414e-01
4.5759141e-01 -2.9438344e-01 -4.6568010e-02 8.2565737e-01
2.1662946e-01 -3.4578230e-02 3.8088870e-01 9.0605396e-01
-4.4707596e-01 -1.0762360e+00 -2.3922757e-03 6.7887354e-01
-5.5857748e-01 -6.4630789e-01 -3.0556709e-01 5.0157338e-01
-1.3951571e-01 7.5912273e-01 -5.0541776e-01 -4.6087629e-01
3.6879110e-01 -2.9493725e-02 -2.2051311e-01 -9.1145200e-01
-8.1636935e-01 -4.7873995e-01 -4.1291741e-01 -7.4583107e-01
2.3762286e-02 -1.2629031e-01 -1.4499140e+00 -6.1047059e-02
-3.7109214e-01 4.1830313e-01 4.0815184e-01 7.6687098e-01
5.8870500e-01 6.9481426e-01 4.4764361e-01 -3.0272588e-01
-6.6226161e-01 -6.9161367e-01 -1.3624072e-01 3.5915518e-01
5.7794297e-01 -4.4117084e-01 -3.6049816e-01 -8.4534340e-02] | [7.482914924621582, 8.041744232177734] |
75fc935d-408c-4128-bd89-f31fbe418f5e | demographic-influences-on-contemporary-art | 2009.14545 | null | https://arxiv.org/abs/2009.14545v2 | https://arxiv.org/pdf/2009.14545v2.pdf | Demographic Influences on Contemporary Art with Unsupervised Style Embeddings | Computational art analysis has, through its reliance on classification tasks, prioritised historical datasets in which the artworks are already well sorted with the necessary annotations. Art produced today, on the other hand, is numerous and easily accessible, through the internet and social networks that are used by professional and amateur artists alike to display their work. Although this art, yet unsorted in terms of style and genre, is less suited for supervised analysis, the data sources come with novel information that may help frame the visual content in equally novel ways. As a first step in this direction, we present contempArt, a multi-modal dataset of exclusively contemporary artworks. contempArt is a collection of paintings and drawings, a detailed graph network based on social connections on Instagram and additional socio-demographic information; all attached to 442 artists at the beginning of their career. We evaluate three methods suited for generating unsupervised style embeddings of images and correlate them with the remaining data. We find no connections between visual style on the one hand and social proximity, gender, and nationality on the other. | ['Yuta Nakashima', 'Noa Garcia', 'Nikolai Huckle'] | 2020-09-30 | null | null | null | null | ['art-analysis'] | ['computer-vision'] | [ 1.10563971e-01 1.51114585e-02 7.96537772e-02 -2.76919544e-01
-6.76769838e-02 -9.58422899e-01 1.23764026e+00 6.25364959e-01
-3.60867292e-01 7.11940587e-01 6.34698987e-01 2.66236197e-02
-4.86449242e-01 -1.09307992e+00 2.74422183e-03 -5.34387887e-01
1.09442011e-01 7.10446179e-01 -5.62537797e-02 -3.97245616e-01
5.32157540e-01 7.20455825e-01 -1.62948263e+00 2.01283917e-01
1.75835773e-01 8.54835272e-01 -7.66411610e-03 6.31034553e-01
-5.63443422e-01 5.49116075e-01 -5.83595991e-01 -1.12752485e+00
1.07327946e-01 -3.75883996e-01 -7.52146006e-01 5.47877178e-02
6.72148824e-01 2.55354732e-01 -4.61714536e-01 8.75988185e-01
4.82684821e-01 2.72222273e-02 9.14209723e-01 -1.06301117e+00
-8.14750135e-01 7.17663705e-01 -5.07016301e-01 1.98177323e-01
4.19059932e-01 2.37378944e-02 1.21582615e+00 -6.76295280e-01
1.26379800e+00 1.41207182e+00 6.34788811e-01 1.29134193e-01
-1.34043217e+00 -2.94628173e-01 -4.12935950e-02 -8.05030465e-02
-1.27685821e+00 -2.01719135e-01 1.21333969e+00 -7.59910703e-01
2.36598536e-01 2.78509140e-01 1.13223374e+00 1.25345576e+00
-1.53247133e-01 2.15851620e-01 1.26822484e+00 -6.46772742e-01
2.12287784e-01 2.71019340e-01 -1.74352258e-01 6.52424932e-01
2.46123135e-01 -3.52947980e-01 -6.64433360e-01 -3.56118716e-02
9.59427714e-01 2.04429835e-01 6.83912039e-02 -4.21875238e-01
-1.45513451e+00 6.60932481e-01 5.94156384e-01 8.61591220e-01
-3.06396693e-01 -1.50453241e-03 3.42614114e-01 3.35843474e-01
6.87979102e-01 7.39365876e-01 1.31599933e-01 -4.13303047e-01
-9.48854327e-01 1.19990364e-01 7.59354889e-01 5.33800364e-01
8.09258580e-01 -1.13934740e-01 2.32570451e-02 1.01103616e+00
1.25035837e-01 3.29973876e-01 2.13042825e-01 -7.27760851e-01
2.83446401e-01 1.02903664e+00 -2.76982337e-01 -1.52967954e+00
-3.94145213e-02 -2.96373308e-01 -6.64320350e-01 4.65265810e-01
6.04794085e-01 1.90939501e-01 -7.60062039e-01 1.37014067e+00
1.40025616e-01 -2.99328208e-01 -3.97320539e-01 6.18985653e-01
7.08667517e-01 1.05872609e-01 1.08735345e-01 8.43831822e-02
1.56839597e+00 -4.08309937e-01 -7.26244569e-01 1.92392021e-01
1.27650350e-01 -1.13473535e+00 1.08106947e+00 4.06917542e-01
-1.08218849e+00 -4.88140136e-01 -9.52450991e-01 -1.49214104e-01
-9.60168004e-01 -1.86577260e-01 7.63538122e-01 7.18917072e-01
-1.02589393e+00 9.57503080e-01 -4.64174062e-01 -7.36757755e-01
8.06308508e-01 1.36568835e-02 -5.53175807e-01 1.64275765e-01
-5.81120312e-01 8.81987274e-01 8.88400376e-02 -3.19097340e-01
-8.83721635e-02 -5.52564681e-01 -7.14694858e-01 -3.57706040e-01
4.54150327e-02 -5.51192760e-01 5.34312665e-01 -1.05763674e+00
-1.28505766e+00 1.37052691e+00 3.51475596e-01 -2.09421813e-02
6.55349553e-01 4.95696627e-02 -8.36495638e-01 3.00862938e-01
1.09704703e-01 6.16986513e-01 8.34681809e-01 -1.29117548e+00
-4.66401517e-01 -5.79531848e-01 2.89984494e-01 -1.11528270e-01
-7.98800349e-01 -9.45222974e-02 -4.10361558e-01 -1.01654792e+00
1.14830576e-01 -8.59085441e-01 4.17323522e-02 1.33921057e-01
-4.29122627e-01 -2.53255486e-01 6.56298101e-01 -3.93850178e-01
1.33026958e+00 -2.13942242e+00 4.67442453e-01 5.70380926e-01
4.08356667e-01 -9.10804868e-02 1.92152634e-01 9.14599836e-01
-9.39883664e-03 4.84344214e-01 -1.27441987e-01 -4.96301591e-01
2.86160916e-01 1.52492404e-01 -7.84597322e-02 4.38640177e-01
7.85438269e-02 7.66087830e-01 -1.04129481e+00 -8.30231547e-01
3.41322541e-01 5.77977598e-01 -2.29308262e-01 -2.40716413e-01
3.39051448e-02 3.28933567e-01 -4.38931346e-01 6.97381198e-01
2.10261300e-01 8.96540359e-02 3.37748110e-01 -2.26548389e-01
-3.04658860e-01 -1.99178439e-02 -1.01552141e+00 1.94150615e+00
-4.55460817e-01 1.19354737e+00 -1.89715415e-01 -4.69263196e-01
1.08673215e+00 -2.20785057e-03 4.91615474e-01 -4.24040407e-01
3.03832144e-01 7.26201236e-02 -2.89555132e-01 -3.65168333e-01
7.91135728e-01 -3.97344828e-01 -2.06384674e-01 5.80343068e-01
-4.41292748e-02 -1.10263214e-01 3.88784230e-01 4.12455976e-01
1.00361478e+00 2.40773395e-01 1.54454663e-01 -3.65886271e-01
2.43675128e-01 9.29802191e-03 -6.29770681e-02 2.36186787e-01
1.06598146e-01 7.69387364e-01 3.97541463e-01 -7.38693297e-01
-1.31119382e+00 -1.24965930e+00 -3.47401559e-01 9.29702282e-01
-1.50883228e-01 -5.85675776e-01 -3.80821675e-01 -4.28622454e-01
2.19124109e-01 2.83008486e-01 -8.70394170e-01 3.30276161e-01
-2.71943957e-01 -3.93555462e-02 4.89202589e-01 2.05741853e-01
1.44067705e-01 -1.32308412e+00 -3.75110924e-01 -7.63901919e-02
2.53460079e-01 -6.82869732e-01 -8.84577259e-02 -2.64899254e-01
-6.11028850e-01 -1.23395073e+00 -9.12636936e-01 -7.12156296e-01
7.17069507e-01 -7.98687935e-02 1.43655491e+00 1.77788332e-01
-6.10537708e-01 7.77135491e-01 -3.49819362e-01 -4.40515369e-01
-2.26032212e-01 1.14830032e-01 -1.90200321e-02 3.60869139e-01
2.68758923e-01 -1.11894500e+00 -4.76530969e-01 -1.05234377e-01
-1.04379189e+00 -1.06708296e-01 5.85095882e-01 2.88802475e-01
3.76702219e-01 -3.98177244e-02 2.85402030e-01 -1.10879350e+00
4.61993217e-01 -5.55986166e-01 6.92240521e-02 1.01853512e-01
-3.51709813e-01 -1.85841948e-01 1.88677117e-01 -3.11188072e-01
-8.93230379e-01 -1.43297136e-01 7.41635486e-02 -3.85963917e-01
-3.30925852e-01 2.82129049e-01 6.29824549e-02 -2.45188177e-02
7.85216808e-01 -2.10528314e-01 4.32921872e-02 -8.56871784e-01
5.52958131e-01 6.76773608e-01 6.35445058e-01 -4.76281315e-01
1.15502083e+00 6.89587712e-01 2.21104711e-01 -1.18699849e+00
-5.72522700e-01 -3.26886654e-01 -1.06329238e+00 -6.23406231e-01
9.26271856e-01 -5.47598541e-01 -7.03088045e-01 1.30407929e-01
-8.20135534e-01 2.05393173e-02 -6.69223785e-01 4.26160038e-01
-3.57296854e-01 3.37105334e-01 -2.23646432e-01 -9.03933525e-01
2.01439857e-01 -4.52428937e-01 8.71107817e-01 -1.22831166e-01
-6.58522129e-01 -1.29606509e+00 1.11203335e-01 3.39241534e-01
1.91294670e-01 1.02285135e+00 1.04569399e+00 -3.40764821e-01
-3.53944778e-01 -3.00194412e-01 -2.72316664e-01 1.40607461e-01
3.59235138e-01 1.56995401e-01 -9.93343532e-01 -6.58127246e-04
-5.71462989e-01 -1.75320104e-01 8.41016710e-01 -1.26105860e-01
9.25470531e-01 -9.18575823e-02 -2.13735431e-01 2.86320597e-01
1.58124161e+00 -2.06076920e-01 6.83412313e-01 5.84287643e-01
1.03153682e+00 7.60114670e-01 -1.17560113e-02 5.09261668e-01
1.90029770e-01 5.54336131e-01 2.37847283e-01 9.75979790e-02
-3.80615205e-01 -4.23522919e-01 1.45902736e-02 8.14601421e-01
-8.22325826e-01 -7.73359761e-02 -9.55477715e-01 7.11169720e-01
-1.47479796e+00 -1.17274451e+00 -3.39441150e-01 2.16742802e+00
6.78764462e-01 3.45307708e-01 4.33503091e-01 5.50540149e-01
6.65259004e-01 4.74067867e-01 -4.47866656e-02 -3.21577966e-01
-3.05046529e-01 4.16184843e-01 2.77217835e-01 1.79719821e-01
-7.38046944e-01 5.93440592e-01 6.10004187e+00 6.89750373e-01
-7.56210148e-01 -1.37775287e-01 3.24051857e-01 -2.58169860e-01
-6.66827679e-01 5.82638988e-03 -2.49502689e-01 5.61977088e-01
5.89910686e-01 1.01344138e-01 5.59816360e-01 6.33849561e-01
-2.81563222e-01 2.40804926e-02 -1.25128782e+00 1.18706131e+00
2.11602136e-01 -1.70901918e+00 1.74444571e-01 2.69108802e-01
6.44987941e-01 -4.72886145e-01 1.63360815e-02 4.75790501e-02
5.96933551e-02 -1.00442290e+00 9.78554368e-01 8.37015808e-01
9.66881394e-01 -8.24877083e-01 3.99879277e-01 -2.82391459e-01
-1.03756464e+00 1.19808033e-01 -7.41362721e-02 -5.02845049e-01
9.76344198e-02 5.27677178e-01 -1.91806987e-01 6.22061133e-01
6.21720135e-01 9.29552674e-01 -9.81636226e-01 6.84206724e-01
-8.00266191e-02 2.11352497e-01 -1.20518200e-01 -2.64467269e-01
2.01022103e-01 -4.75444764e-01 5.09828091e-01 1.27098525e+00
2.82117784e-01 -2.55483478e-01 -7.50267953e-02 6.81077242e-01
-1.47989348e-01 2.60484427e-01 -9.26343799e-01 -6.33485913e-01
4.55004185e-01 1.50262618e+00 -1.27281201e+00 -1.06173530e-01
-1.76084071e-01 6.96199000e-01 3.95916313e-01 7.59284869e-02
-3.14976722e-01 -4.57062453e-01 6.05207622e-01 6.22231960e-01
-2.92563066e-03 -4.77225274e-01 -3.05374444e-01 -7.88678885e-01
-6.66458607e-02 -4.58467454e-01 3.14859033e-01 -7.35472262e-01
-1.82422006e+00 5.94766498e-01 -2.98347414e-01 -9.85341191e-01
1.00555152e-01 -6.23296201e-01 -5.24802625e-01 6.30683780e-01
-9.07926977e-01 -1.48866236e+00 -2.84842342e-01 4.68553692e-01
2.93565750e-01 -4.25799489e-01 9.69016612e-01 3.48721564e-01
-3.12997103e-01 2.08884493e-01 5.80197647e-02 2.22612083e-01
6.94805861e-01 -1.53067410e+00 2.18603328e-01 2.50648946e-01
8.17491829e-01 4.39633846e-01 6.52229726e-01 -6.94823623e-01
-1.28528357e+00 -5.97279131e-01 1.10874236e+00 -8.79114151e-01
1.14200556e+00 -4.82628703e-01 -3.21628541e-01 5.90767443e-01
4.12417531e-01 -4.00570035e-01 8.86988521e-01 6.33809686e-01
-3.15248579e-01 1.26810640e-01 -8.73673260e-01 7.79753447e-01
1.42346060e+00 -7.88057208e-01 -9.27893281e-01 3.68241996e-01
1.66465446e-01 6.34908974e-02 -1.28629029e+00 -3.07726681e-01
7.31542528e-01 -9.90013361e-01 9.99461412e-01 -4.21510607e-01
8.66175473e-01 -5.58888018e-02 -1.29761860e-01 -1.13456464e+00
-4.47009087e-01 -6.95829034e-01 4.28429037e-01 1.94981158e+00
2.51317620e-01 -3.52132410e-01 9.41650271e-01 3.67219806e-01
-1.25949562e-01 -4.53624666e-01 -7.56297529e-01 -6.21342540e-01
-1.02320768e-01 -5.09940445e-01 7.00930119e-01 1.24392366e+00
-1.52899802e-01 4.46573377e-01 -2.59627640e-01 -6.48019016e-01
4.50528562e-01 2.58021295e-01 8.17240238e-01 -1.85096169e+00
4.62751016e-02 -7.31064439e-01 -9.76942658e-01 -6.18237816e-02
-1.15632787e-01 -9.61104989e-01 -6.41160786e-01 -1.56230783e+00
3.82881388e-02 -6.26288056e-01 -2.31120571e-01 9.27961469e-02
4.44638133e-01 9.22004998e-01 4.66969162e-01 3.55324984e-01
-4.52453077e-01 1.72799319e-01 1.37122357e+00 -1.85199857e-01
-8.91582593e-02 -4.58812177e-01 -7.32117951e-01 1.04996347e+00
6.20580792e-01 -1.99907050e-01 -1.68149844e-01 -3.11583161e-01
7.19308913e-01 -5.48320115e-01 4.21079189e-01 -1.21636355e+00
-2.07517985e-02 5.77629842e-02 7.19212294e-01 -4.81442302e-01
5.76259077e-01 -1.01235986e+00 4.22404051e-01 5.50837442e-02
-3.61218780e-01 1.12642363e-01 -7.38623589e-02 8.24033439e-01
-1.68767437e-01 -1.02692403e-01 4.95633394e-01 -2.90929765e-01
-5.91944814e-01 1.82848468e-01 -1.84092432e-01 1.84117749e-01
8.06748927e-01 -5.68030417e-01 -2.63520479e-01 -2.73565352e-01
-1.09753311e+00 -4.56799448e-01 1.01272786e+00 5.19603252e-01
2.35469699e-01 -1.73532856e+00 -4.77263719e-01 6.27254024e-02
3.02486598e-01 -4.53255683e-01 1.82387695e-01 3.79101753e-01
-6.31665945e-01 -1.62763491e-01 -7.77299643e-01 -2.88348734e-01
-1.22285914e+00 5.13751805e-01 -4.13956165e-01 4.88587767e-02
-6.76473141e-01 7.94272780e-01 -2.16816038e-01 1.41519904e-02
8.92148688e-02 1.79382533e-01 -5.57306528e-01 9.19709980e-01
2.84152448e-01 5.24087429e-01 -1.42910957e-01 -1.03067398e+00
-1.21057875e-01 6.87290847e-01 4.10480797e-01 -2.17903063e-01
1.69054413e+00 -1.73922956e-01 -2.72364020e-01 1.02737963e+00
1.04781687e+00 4.61534381e-01 -8.13955605e-01 -1.02927886e-01
2.33803049e-01 -9.61304009e-01 -3.79974395e-02 -6.71609938e-01
-1.00791132e+00 8.96658063e-01 4.66000199e-01 8.51138830e-01
7.00299442e-01 4.25708555e-02 4.45963919e-01 1.96126625e-01
3.61870021e-01 -1.30263281e+00 3.53038669e-01 9.58783627e-02
1.04614174e+00 -1.00393414e+00 3.36526185e-01 -3.43725741e-01
-5.26372373e-01 1.24346066e+00 -2.49957871e-02 -1.97780952e-01
7.27477074e-01 -2.01735958e-01 -5.22128902e-02 -6.52909160e-01
-1.44631848e-01 -4.65112329e-01 3.65777135e-01 6.49242997e-01
5.06839573e-01 1.49051502e-01 -2.66290784e-01 3.03766429e-01
-6.73354149e-01 -3.41099858e-01 3.70266736e-01 9.53243911e-01
-2.06814513e-01 -1.28942847e+00 -4.44922149e-01 5.32400072e-01
-6.16097331e-01 2.09324822e-01 -9.59007919e-01 1.13655305e+00
4.67376679e-01 6.60013080e-01 4.66324836e-01 -4.22074586e-01
3.49556148e-01 1.23191081e-01 6.48994267e-01 -5.94832957e-01
-7.18654871e-01 -3.84409368e-01 4.46827024e-01 -1.45126268e-01
-7.12274849e-01 -6.63543224e-01 -6.33349657e-01 -7.81296730e-01
2.29989335e-01 -1.16077103e-02 9.20955718e-01 7.23811150e-01
1.37949601e-01 6.95676088e-01 4.25891042e-01 -1.38458467e+00
3.98916006e-01 -8.66853178e-01 -1.14065301e+00 9.35972869e-01
-1.03101514e-01 -8.21691990e-01 -2.71971405e-01 2.59068161e-01] | [11.302689552307129, 0.29054874181747437] |
dffee6cb-7a6a-4798-a53b-b4c23f70699e | enabling-classifiers-to-make-judgements | 2210.07652 | null | https://arxiv.org/abs/2210.07652v1 | https://arxiv.org/pdf/2210.07652v1.pdf | Enabling Classifiers to Make Judgements Explicitly Aligned with Human Values | Many NLP classification tasks, such as sexism/racism detection or toxicity detection, are based on human values. Yet, human values can vary under diverse cultural conditions. Therefore, we introduce a framework for value-aligned classification that performs prediction based on explicitly written human values in the command. Along with the task, we propose a practical approach that distills value-aligned knowledge from large-scale language models (LLMs) to construct value-aligned classifiers in two steps. First, we generate value-aligned training data from LLMs by prompt-based few-shot learning. Next, we fine-tune smaller classification models with the generated data for the task. Empirical results show that our VA-Models surpass multiple baselines by at least 15.56% on the F1-score, including few-shot learning with OPT-175B and existing text augmentation methods. We suggest that using classifiers with explicit human value input improves both inclusivity & explainability in AI. | ['Pascale Fung', 'Mona Diab', 'Zhaojiang Lin', 'Andrea Madotto', 'Tiezheng Yu', 'Yejin Bang'] | 2022-10-14 | null | null | null | null | ['text-augmentation'] | ['natural-language-processing'] | [ 4.19614017e-01 5.48122585e-01 -8.74447107e-01 -6.69742107e-01
-7.22603023e-01 -1.69646934e-01 8.40637028e-01 2.81255007e-01
-6.31617963e-01 1.03065634e+00 8.49139631e-01 -1.15140513e-01
6.30727485e-02 -7.34656692e-01 -5.46067417e-01 -5.98403253e-02
5.04671395e-01 8.26891720e-01 -3.22269291e-01 -6.60735786e-01
3.99756312e-01 -7.53106400e-02 -1.57524431e+00 5.07233500e-01
1.22202456e+00 7.41472542e-01 -4.03491020e-01 7.01540351e-01
-5.34092844e-01 1.18725729e+00 -7.74307430e-01 -8.72865379e-01
-8.07191208e-02 -6.34588599e-01 -9.22046959e-01 -1.78148180e-01
4.76810426e-01 -5.72968900e-01 1.29722163e-01 8.31690431e-01
4.90676969e-01 4.61191297e-01 1.05614281e+00 -1.44021273e+00
-1.11411214e+00 1.33776653e+00 -3.53187442e-01 -1.83658615e-01
4.11596060e-01 4.06782687e-01 1.13229597e+00 -8.17668021e-01
8.28885496e-01 1.35565007e+00 7.13514388e-01 1.09869921e+00
-1.38613367e+00 -8.45445573e-01 -9.19543654e-02 2.57850617e-01
-8.23330998e-01 -4.11231995e-01 6.23688221e-01 -5.70430338e-01
9.78603959e-01 3.08987588e-01 7.93352485e-01 1.58527160e+00
-3.14565837e-01 7.61824489e-01 9.75516796e-01 -6.92053437e-01
2.21142471e-01 2.87971526e-01 3.63067478e-01 5.39582372e-01
2.33177707e-01 -2.36929953e-01 -9.03725147e-01 -2.13925362e-01
4.13950890e-01 -1.89375684e-01 1.47070020e-01 -3.18541914e-01
-1.55440104e+00 1.13844860e+00 2.54200906e-01 1.15002900e-01
-9.35752839e-02 1.02018356e-01 3.81085902e-01 1.59777641e-01
8.05388212e-01 1.32248962e+00 -5.61496258e-01 -5.35606444e-01
-8.74043345e-01 4.24308747e-01 8.10619652e-01 1.09253812e+00
3.62987489e-01 1.05989397e-01 -8.08375239e-01 1.16381919e+00
-1.68318406e-01 5.93145013e-01 8.22371840e-01 -1.01406693e+00
3.46829712e-01 6.50273442e-01 2.47465253e-01 -6.00511670e-01
-3.19188178e-01 6.70953467e-02 -4.52033341e-01 -1.46014288e-01
5.43526292e-01 -3.65196049e-01 -1.19875562e+00 1.76629400e+00
3.67179334e-01 -5.34883030e-02 1.46128133e-01 5.99928021e-01
1.02856112e+00 5.69109440e-01 4.80501831e-01 3.55144292e-02
1.32611680e+00 -1.13003492e+00 -9.92730796e-01 -2.90919453e-01
1.18081653e+00 -2.91458696e-01 1.68094599e+00 2.28510737e-01
-7.99700797e-01 -4.27112341e-01 -1.03253782e+00 -3.71863097e-01
-5.82417011e-01 -1.68958277e-01 6.91522777e-01 9.32258248e-01
-4.42888707e-01 7.29979575e-01 -2.77407944e-01 -4.10185516e-01
5.64896107e-01 1.65636584e-01 -1.58627316e-01 3.70786488e-02
-1.47015440e+00 1.03503835e+00 5.16311705e-01 -7.69447684e-01
-6.21167004e-01 -9.97792304e-01 -1.31835115e+00 -1.25820816e-01
4.98701245e-01 -4.08984095e-01 1.42454791e+00 -8.35478961e-01
-1.33421886e+00 9.47978497e-01 4.51156124e-03 -5.15347183e-01
5.09412289e-01 -4.34225917e-01 -2.22498283e-01 -2.98060507e-01
3.06117773e-01 1.05700994e+00 7.11519182e-01 -9.14941072e-01
-4.39161688e-01 7.28747025e-02 4.70120050e-02 2.40152314e-01
-8.93336177e-01 2.91492015e-01 -7.95417950e-02 -8.81357431e-01
-6.48835897e-01 -7.43993044e-01 -2.70840079e-01 4.30555977e-02
-3.04345965e-01 -2.27502599e-01 2.01165974e-01 -6.04011655e-01
1.27386475e+00 -1.81905544e+00 1.20278098e-01 -3.08339652e-02
1.72348201e-01 2.56798923e-01 -3.95211101e-01 8.13325420e-02
1.22331902e-01 3.43526989e-01 -2.70154178e-01 -1.23293251e-01
-1.82915088e-02 1.56303734e-01 -3.23172837e-01 -5.56839444e-03
3.30870390e-01 1.27533543e+00 -1.15487516e+00 -7.46121347e-01
2.28723824e-01 1.93431869e-01 -8.88223350e-01 9.53333452e-02
-4.87002730e-01 2.34028623e-01 -4.13539596e-02 6.56569600e-01
2.99178183e-01 -6.76571056e-02 1.54269502e-01 2.93725222e-01
1.86015815e-01 1.02112733e-01 -8.04154277e-01 1.68298793e+00
-2.89435685e-01 5.14624000e-01 -6.68992162e-01 -6.62131369e-01
1.02663434e+00 -5.37557788e-02 2.46550411e-01 -6.42802119e-01
2.75400043e-01 -9.81031545e-03 9.11546722e-02 -4.11485285e-01
8.05944622e-01 -2.59534270e-01 -5.43777525e-01 4.72955108e-01
2.69937605e-01 -5.43808520e-01 1.73794463e-01 3.22310895e-01
9.04000878e-01 2.22958490e-01 6.38638973e-01 -1.24661356e-01
2.67956883e-01 3.07093054e-01 5.65729737e-01 7.38901079e-01
-4.54669118e-01 3.93735647e-01 7.37666786e-01 -3.55739355e-01
-1.27096844e+00 -6.69018388e-01 -2.16265827e-01 1.79437959e+00
-1.83702875e-02 -6.55457616e-01 -8.06739748e-01 -7.35195935e-01
1.87791169e-01 1.45530915e+00 -1.05296946e+00 -4.08885568e-01
-2.13598549e-01 -6.27426505e-01 5.44676125e-01 7.15036035e-01
1.53407976e-01 -1.19498122e+00 -5.40353596e-01 1.45659536e-01
-1.68070570e-01 -1.13459349e+00 -2.53409415e-01 1.19051076e-01
-5.96547365e-01 -7.58558273e-01 -7.86113262e-01 -6.49008930e-01
4.99958634e-01 -1.44353583e-01 1.12782109e+00 8.56128857e-02
-4.38809767e-02 1.22698516e-01 -7.06921220e-01 -9.09068823e-01
-5.41447580e-01 1.65692329e-01 9.63393003e-02 -2.39227980e-01
8.16498697e-01 6.40459508e-02 -6.26769895e-03 1.58201262e-01
-3.36308211e-01 5.29868305e-01 2.87490666e-01 1.17350912e+00
1.83536068e-01 -6.96166694e-01 6.54999256e-01 -1.49410558e+00
9.37932253e-01 -4.23864394e-01 2.87577249e-02 3.59113008e-01
-6.92590714e-01 1.72740057e-01 3.93410712e-01 -8.68811369e-01
-1.17137563e+00 -3.26943509e-02 7.19976202e-02 -7.81506896e-02
-1.54360846e-01 3.48840266e-01 -1.85827557e-02 2.58809388e-01
1.19544840e+00 -1.57278448e-01 2.52001472e-02 -1.62306577e-01
8.60789955e-01 1.05124736e+00 5.71469486e-01 -6.60391867e-01
7.36766100e-01 -5.50130494e-02 -4.10714954e-01 -8.75436246e-01
-1.28692150e+00 -1.99084163e-01 -6.48790956e-01 -2.66105562e-01
9.70964432e-01 -9.89634931e-01 -3.01972002e-01 2.77676076e-01
-1.06373823e+00 -5.28828442e-01 -6.24927700e-01 4.96623307e-01
-4.84849155e-01 -6.35160133e-02 -6.57307744e-01 -8.35673153e-01
-4.95352089e-01 -7.07755268e-01 9.72981870e-01 3.02602261e-01
-9.32931125e-01 -7.36528039e-01 -2.60237828e-02 9.08157527e-01
3.93335253e-01 3.43264431e-01 1.25782347e+00 -1.24810863e+00
2.68213034e-01 -1.68372095e-01 -8.17255229e-02 8.50496888e-02
-9.97379124e-02 -2.01306060e-01 -9.64219689e-01 3.25139165e-01
-4.67964172e-01 -1.13419890e+00 6.33560419e-01 2.77846098e-01
1.17809796e+00 -3.02582055e-01 -9.90946218e-02 5.07263422e-01
7.49663115e-01 7.32045174e-02 5.79365551e-01 3.64065140e-01
7.84451008e-01 8.88075233e-01 1.05841839e+00 8.39575708e-01
3.01931232e-01 3.53321940e-01 -5.17013706e-02 -2.05023284e-03
-4.60523032e-02 -5.16529322e-01 2.22218841e-01 3.76385570e-01
-1.14455223e-01 1.66062281e-01 -1.21635044e+00 6.42301798e-01
-1.93344331e+00 -1.01423109e+00 -6.32074103e-02 1.97746825e+00
1.51316595e+00 3.35393459e-01 3.03994894e-01 -7.12086931e-02
7.72675037e-01 1.38369694e-01 -7.99575031e-01 -8.75737071e-01
4.75199521e-02 1.58778802e-01 2.66698241e-01 2.92006731e-01
-1.06481731e+00 1.54979992e+00 6.77518082e+00 7.90709913e-01
-6.10992432e-01 2.10447192e-01 8.47990870e-01 -5.16275525e-01
-5.32807410e-01 -2.63683885e-01 -8.84588957e-01 3.98956150e-01
1.12846482e+00 -5.92019439e-01 3.66672337e-01 1.21272278e+00
-1.61410943e-01 2.18686491e-01 -1.08777273e+00 8.92751753e-01
5.77776670e-01 -1.27988660e+00 2.65792310e-01 -2.98499446e-02
1.05476964e+00 -3.59578639e-01 -1.48050142e-02 9.84800935e-01
7.42476106e-01 -1.22629488e+00 7.45976388e-01 2.45651543e-01
1.21838176e+00 -8.82446289e-01 6.35621011e-01 4.84158933e-01
-5.71774185e-01 -2.42799714e-01 -5.02427340e-01 -4.26515043e-01
-4.71087061e-02 1.88700452e-01 -9.39117670e-01 -2.92521745e-01
3.52884293e-01 7.17519224e-01 -8.05835903e-01 4.93491769e-01
-5.70598006e-01 5.03489137e-01 2.91824877e-01 -5.92202485e-01
1.16398074e-02 2.07445189e-01 7.60729313e-02 1.14348042e+00
-5.28639853e-02 1.97869152e-01 3.65440287e-02 1.03592551e+00
-4.82852489e-01 4.58502114e-01 -7.76734471e-01 -6.68278456e-01
6.20383978e-01 1.13583648e+00 -3.47496569e-01 -7.22597480e-01
-6.26228273e-01 8.43866050e-01 3.08797896e-01 -4.65989262e-02
-9.35984135e-01 -6.94903016e-01 6.50113225e-01 1.04406759e-01
-2.82603234e-01 1.80823386e-01 -8.29518080e-01 -1.24295521e+00
-5.50791562e-01 -1.29716122e+00 3.55484009e-01 -8.15111458e-01
-1.17506516e+00 3.65725517e-01 1.90738454e-01 -1.07692325e+00
-6.55870080e-01 -7.06370771e-01 -2.61180729e-01 6.08169377e-01
-1.00555968e+00 -1.61092281e+00 -3.04336101e-01 1.91717774e-01
9.55554485e-01 -5.02300978e-01 1.06603134e+00 -3.92569512e-01
-3.81278157e-01 8.07186723e-01 -1.16067484e-01 4.09058839e-01
1.09435987e+00 -1.29570127e+00 8.85271788e-01 5.62403321e-01
-2.34521292e-02 6.47499561e-01 7.92227805e-01 -1.11135793e+00
-8.78459334e-01 -8.05011213e-01 1.05691373e+00 -8.62092614e-01
8.20937932e-01 -6.79400921e-01 -8.42569411e-01 6.40642643e-01
1.54555857e-01 -3.16996723e-01 1.20865774e+00 6.43828154e-01
-5.94168723e-01 4.31334227e-01 -1.22269893e+00 8.93131971e-01
1.16647792e+00 -5.10881245e-01 -1.16084051e+00 3.28020245e-01
1.11189604e+00 -2.14533508e-01 -8.15018713e-01 1.78918079e-01
7.25429058e-01 -2.10036233e-01 8.44910145e-01 -1.33399725e+00
1.12653971e+00 4.67991173e-01 -1.23467676e-01 -1.49727690e+00
-2.85705447e-01 -5.05134225e-01 -2.19037354e-01 1.34656298e+00
7.10826099e-01 -1.11201838e-01 8.61764669e-01 1.43857026e+00
2.15287536e-01 -6.59406185e-01 -4.86771971e-01 -6.99392080e-01
4.22193378e-01 -5.67536175e-01 7.56379068e-01 1.50521684e+00
4.85697627e-01 6.91761255e-01 -8.63605499e-01 -6.89584017e-01
6.04051173e-01 -2.27821082e-01 1.03199446e+00 -1.52910221e+00
-1.17862195e-01 -2.62035966e-01 -9.66085419e-02 -2.24763453e-01
4.99876857e-01 -9.97204125e-01 -6.09290134e-03 -1.52471685e+00
6.41675770e-01 -5.08141145e-02 -1.76503196e-01 9.44343090e-01
-4.04931575e-01 1.79234773e-01 4.17464554e-01 -8.97157639e-02
-5.34979045e-01 5.59194744e-01 1.14617050e+00 -2.81824082e-01
-5.36275148e-01 -6.01738513e-01 -1.12147677e+00 8.21724534e-01
8.24223340e-01 -4.83402818e-01 -3.63289118e-01 -1.76639259e-01
3.36670905e-01 -2.79194713e-01 -6.95941672e-02 -9.91042554e-01
-7.82218650e-02 -8.32895517e-01 7.43411839e-01 -7.43079185e-02
4.81200397e-01 -2.68412977e-01 -5.97092986e-01 4.55488056e-01
-1.09496486e+00 -3.49425286e-01 1.07324272e-01 3.03792417e-01
2.94123799e-01 -5.27187288e-01 8.12689126e-01 -1.39647022e-01
-9.84877348e-01 -1.20709985e-01 -1.62391126e-01 4.43642259e-01
1.04823995e+00 -2.20997274e-01 -8.06582212e-01 -4.98820752e-01
-5.31685650e-01 4.00373220e-01 4.78899717e-01 8.04714262e-01
6.30470753e-01 -1.43166995e+00 -7.82988608e-01 3.21883500e-01
7.44131625e-01 -5.22336423e-01 2.19104579e-03 4.31289762e-01
-3.28680634e-01 1.58934742e-01 -6.53141797e-01 -6.56776130e-02
-1.32641208e+00 5.66018462e-01 -1.42448142e-01 7.01103434e-02
-3.89313877e-01 9.73722219e-01 -2.67388076e-01 -7.82644629e-01
2.88971066e-01 -8.04261565e-02 -5.70504129e-01 4.98175383e-01
8.78267646e-01 2.57447332e-01 -6.49444044e-01 -6.34021401e-01
-1.74541339e-01 2.81371862e-01 -2.68779248e-01 -5.76234818e-01
1.50631225e+00 3.93556356e-01 3.42004001e-01 5.80372751e-01
8.12925696e-01 -5.72548285e-02 -1.01134264e+00 -2.35995620e-01
2.28363693e-01 -6.67072058e-01 -2.76396275e-02 -1.24069715e+00
-3.11695784e-01 8.51253569e-01 2.15629622e-01 -4.60606158e-01
5.95007956e-01 -5.06250635e-02 7.56793439e-01 7.45116651e-01
2.18337998e-01 -1.80635989e+00 3.98461848e-01 6.72587454e-01
9.39783990e-01 -1.46491945e+00 2.47733686e-02 -2.87765443e-01
-1.59971309e+00 7.41104305e-01 1.12017155e+00 4.25065696e-01
1.75704464e-01 -1.33559275e-02 8.80356058e-02 2.22583115e-01
-9.56788003e-01 -3.99508566e-01 4.55596268e-01 8.82416666e-01
8.06143939e-01 1.93138659e-01 -4.84161764e-01 1.05507290e+00
-5.67105651e-01 1.85920373e-01 6.94420576e-01 9.58164573e-01
-6.63660586e-01 -9.30069864e-01 -3.36429089e-01 8.95043969e-01
-1.80333525e-01 -4.49274987e-01 -9.62367773e-01 6.20644271e-01
1.68272018e-01 8.41847658e-01 6.06716536e-02 -7.28518367e-01
1.59912467e-01 5.25950313e-01 2.02876449e-01 -1.03171957e+00
-7.05478966e-01 -3.23584974e-01 5.73556602e-01 -3.87456656e-01
5.32952771e-02 -4.98206556e-01 -1.35236192e+00 -4.13614362e-01
-1.63661152e-01 -9.37714353e-02 5.39520860e-01 7.82175303e-01
6.51074350e-02 3.51011813e-01 7.86169022e-02 -5.77889204e-01
-8.41134787e-01 -1.21611285e+00 -6.51505113e-01 8.95873308e-01
-1.90886036e-01 -6.72682524e-01 -2.30884895e-01 -4.77533452e-02] | [10.929702758789062, 8.248601913452148] |
78948d15-bc27-4285-881a-ae7200bcf7d9 | dynamic-programming-in-rank-space-scaling-1 | 2205.00484 | null | https://arxiv.org/abs/2205.00484v1 | https://arxiv.org/pdf/2205.00484v1.pdf | Dynamic Programming in Rank Space: Scaling Structured Inference with Low-Rank HMMs and PCFGs | Hidden Markov Models (HMMs) and Probabilistic Context-Free Grammars (PCFGs) are widely used structured models, both of which can be represented as factor graph grammars (FGGs), a powerful formalism capable of describing a wide range of models. Recent research found it beneficial to use large state spaces for HMMs and PCFGs. However, inference with large state spaces is computationally demanding, especially for PCFGs. To tackle this challenge, we leverage tensor rank decomposition (aka.\ CPD) to decrease inference computational complexities for a subset of FGGs subsuming HMMs and PCFGs. We apply CPD on the factors of an FGG and then construct a new FGG defined in the rank space. Inference with the new FGG produces the same result but has a lower time complexity when the rank size is smaller than the state size. We conduct experiments on HMM language modeling and unsupervised PCFG parsing, showing better performance than previous work. Our code is publicly available at \url{https://github.com/VPeterV/RankSpace-Models}. | ['Kewei Tu', 'Wei Liu', 'Songlin Yang'] | 2022-05-01 | null | https://aclanthology.org/2022.naacl-main.353 | https://aclanthology.org/2022.naacl-main.353.pdf | naacl-2022-7 | ['constituency-grammar-induction'] | ['natural-language-processing'] | [ 1.44746631e-01 3.42377514e-01 -1.69695884e-01 -2.69411087e-01
-9.07204866e-01 -6.82257950e-01 6.77759647e-01 -1.38808697e-01
5.70537113e-02 5.12135983e-01 3.89459044e-01 -9.75130916e-01
1.99371446e-02 -7.52938271e-01 -7.12355196e-01 -6.92899704e-01
-2.56228298e-01 5.81289232e-01 6.63544595e-01 1.24201074e-01
-2.08619628e-02 5.23641884e-01 -1.45951664e+00 1.20621189e-01
7.07522750e-01 2.12090760e-01 6.42899871e-01 1.06668901e+00
-4.00348872e-01 6.32111430e-01 -1.91337287e-01 -5.36724269e-01
1.72191318e-02 -4.47252512e-01 -1.04812956e+00 1.68552101e-02
-7.77918622e-02 -1.56686231e-01 -4.90038842e-01 1.01424575e+00
1.37705430e-01 3.85372519e-01 2.27830708e-01 -1.25566363e+00
-1.69231579e-01 9.94648159e-01 4.49115112e-02 4.46900018e-02
4.64705497e-01 -1.23500519e-01 1.25406265e+00 -5.72348416e-01
6.44547701e-01 1.74342084e+00 2.40249842e-01 5.61635315e-01
-1.45074654e+00 -5.43724358e-01 4.81665462e-01 1.49113774e-01
-1.19963288e+00 -3.67276162e-01 4.34847474e-01 -2.97938645e-01
1.33859837e+00 5.30748129e-01 6.06484592e-01 1.13210964e+00
1.59953028e-01 8.65883231e-01 1.09144163e+00 -6.87489688e-01
3.60441387e-01 -5.72461069e-01 5.89780986e-01 9.03426051e-01
1.93371862e-01 -1.60831120e-02 -5.66405118e-01 -5.61634958e-01
7.27641940e-01 -5.60434647e-02 1.63216054e-01 -1.94807313e-02
-1.05815923e+00 9.78226364e-01 -4.03434247e-01 1.97738856e-01
-1.24735400e-01 3.11130732e-01 2.22894743e-01 1.10518187e-01
8.70165229e-02 -7.11929873e-02 -3.34672600e-01 -5.63232422e-01
-5.29217839e-01 3.86225075e-01 1.10363388e+00 1.23449278e+00
9.41946507e-01 -2.12861329e-01 1.09690435e-01 8.01871359e-01
5.33423781e-01 5.61468601e-01 4.32461016e-02 -1.08712792e+00
5.19120634e-01 5.08414090e-01 1.42702798e-03 -5.29968739e-01
-3.67378831e-01 -1.51492264e-02 -6.40309453e-01 -3.62212896e-01
5.50743580e-01 9.17456374e-02 -1.25780261e+00 1.80679893e+00
4.65082347e-01 1.77808359e-01 -1.60506219e-02 4.92873341e-01
3.98578703e-01 1.07016909e+00 3.82386357e-01 -3.73153418e-01
1.51578951e+00 -8.40051413e-01 -7.81439722e-01 -4.91208315e-01
1.07848716e+00 -6.51427925e-01 1.03137827e+00 3.59672546e-01
-9.73706365e-01 -3.05116683e-01 -7.28171408e-01 -1.60166860e-01
-1.30880773e-01 -9.65414494e-02 9.64629233e-01 8.16540956e-01
-1.27701533e+00 5.37528336e-01 -1.87010765e+00 -4.35409337e-01
-3.09729725e-01 3.63536328e-01 -3.11411023e-01 -9.20589268e-02
-1.20569086e+00 7.04823911e-01 6.24886155e-01 1.20248847e-01
-8.76727760e-01 2.73624808e-02 -1.01919281e+00 1.65448971e-02
6.13803685e-01 -2.64693677e-01 1.43551219e+00 -7.81632140e-02
-1.52806056e+00 3.49738240e-01 -6.78108215e-01 -3.26786995e-01
3.20474207e-02 -3.66641097e-02 -4.87052917e-01 -1.35660633e-01
2.13902332e-02 2.36313924e-01 4.56817180e-01 -9.49595332e-01
-6.05918825e-01 -4.03886318e-01 1.74591154e-01 2.38025095e-02
1.83450669e-01 4.20266926e-01 -8.94521415e-01 -1.91544160e-01
6.89655423e-01 -1.45455635e+00 -5.41074932e-01 -9.80860591e-01
-8.59362781e-01 -4.62256521e-01 4.26499069e-01 -9.38846529e-01
1.77260649e+00 -2.00947714e+00 2.88058072e-01 2.43237764e-01
-2.17963271e-02 2.06968531e-01 -1.48596704e-01 9.46053565e-01
1.68120340e-01 2.66961277e-01 -3.43111277e-01 -4.83166784e-01
3.51713479e-01 8.57473314e-01 -2.34467983e-01 4.07417677e-03
2.70845406e-02 7.66588330e-01 -8.74669790e-01 -6.26253903e-01
2.27056831e-01 4.08219062e-02 -6.08942091e-01 2.92850584e-01
-4.52970713e-01 2.03957587e-01 -5.59074044e-01 5.71239889e-01
5.46989620e-01 -3.43887806e-01 1.07171404e+00 2.51691610e-01
-1.63083136e-01 7.91612744e-01 -1.38580501e+00 1.57947421e+00
-2.31585428e-01 8.14085156e-02 -8.70491713e-02 -6.73211515e-01
7.88184762e-01 2.63195693e-01 -3.98408249e-03 -6.46234080e-02
-3.80113631e-01 2.27111176e-01 1.42865956e-01 -3.15004468e-01
4.48408246e-01 -6.44868165e-02 -3.67334187e-01 3.88465136e-01
2.03590930e-01 7.32005611e-02 7.06445336e-01 5.89142144e-01
1.37054408e+00 3.22829008e-01 3.79280329e-01 -7.86349643e-03
2.35006541e-01 -2.93192476e-01 9.53082621e-01 8.99254382e-01
7.41473362e-02 1.47817343e-01 8.63947630e-01 -3.01363051e-01
-8.48554611e-01 -1.23322046e+00 1.49686202e-01 1.42159522e+00
-1.87283471e-01 -1.17014861e+00 -6.87845051e-01 -5.27715266e-01
-3.22891980e-01 8.13222289e-01 -2.29138717e-01 2.32424274e-01
-8.48206699e-01 -8.46615791e-01 6.12649858e-01 7.11918116e-01
6.73208982e-02 -9.53407764e-01 -1.98829263e-01 5.08471608e-01
-3.39323670e-01 -1.51025224e+00 -1.13796756e-01 4.16904092e-01
-9.36234474e-01 -8.08556795e-01 -9.94202569e-02 -7.36713171e-01
4.53500360e-01 -1.56115875e-01 9.35597420e-01 -1.40361348e-03
2.94441015e-01 5.44825476e-03 -4.64424103e-01 -3.12697738e-02
-9.84198570e-01 2.39771634e-01 2.24439614e-02 -3.84094000e-01
3.50960463e-01 -7.42438376e-01 1.23106549e-02 3.96263659e-01
-9.29321885e-01 6.05150223e-01 4.66119945e-01 7.42646992e-01
5.37426233e-01 -3.02036032e-02 -3.66619006e-02 -1.10619485e+00
3.04822743e-01 -2.39868462e-01 -9.46106374e-01 2.62219042e-01
-3.31999391e-01 3.77648830e-01 5.51973999e-01 -2.31446937e-01
-1.08960378e+00 2.44517699e-01 -4.54606533e-01 -1.84940979e-01
-4.35007900e-01 7.24852681e-01 -4.30650055e-01 4.81043041e-01
1.92259669e-01 3.21774095e-01 -2.67924249e-01 -7.93383658e-01
2.72945702e-01 4.01357561e-01 3.49750668e-01 -7.67827511e-01
6.45022452e-01 2.01442420e-01 2.35970184e-01 -7.81395912e-01
-4.38020676e-01 -4.02401894e-01 -5.83842933e-01 6.71724975e-02
7.82636106e-01 -6.74859107e-01 -5.60950577e-01 2.47064158e-01
-1.17522347e+00 -5.43623269e-01 4.20435131e-01 6.95676148e-01
-5.43997049e-01 7.83278942e-01 -1.03191972e+00 -1.34289610e+00
8.16815272e-02 -1.33198559e+00 1.16410542e+00 -7.80257806e-02
-2.26556018e-01 -9.33315933e-01 1.80189595e-01 3.01491261e-01
-2.21006513e-01 2.08315670e-01 1.19804549e+00 -8.28654528e-01
-6.10202134e-01 -1.01961344e-01 2.11357653e-01 2.30320886e-01
-1.50863469e-01 1.84617087e-01 -6.59122705e-01 -2.18668163e-01
-2.96537489e-01 1.96216851e-01 6.52154863e-01 1.40921921e-01
7.61745691e-01 -4.38072354e-01 -5.87151587e-01 2.19002053e-01
1.16877675e+00 5.37776768e-01 7.16063023e-01 5.35236299e-02
7.99044847e-01 3.91405553e-01 5.12016952e-01 2.34467283e-01
5.97729683e-01 6.55715346e-01 7.12743178e-02 3.92412990e-01
1.63550422e-01 -6.31852567e-01 7.75176406e-01 1.44467068e+00
-3.26175958e-01 -3.06085616e-01 -1.38857114e+00 3.29745412e-01
-2.10778522e+00 -7.07685053e-01 -4.11652476e-01 2.04659462e+00
7.25528359e-01 1.87800735e-01 1.73456714e-01 1.67983741e-01
8.27218235e-01 1.08289205e-01 -7.65429661e-02 -7.06762493e-01
7.10778534e-02 2.56734312e-01 4.45181727e-01 8.68525982e-01
-9.34571266e-01 1.30154037e+00 6.19104338e+00 7.75144994e-01
-4.52259570e-01 1.43595845e-01 1.37901157e-01 1.20857574e-01
-2.21546471e-01 8.03861916e-01 -1.19937456e+00 5.33939958e-01
1.52095401e+00 2.15774655e-01 6.70504570e-01 6.78019226e-01
1.56365499e-01 -2.56009281e-01 -1.02503550e+00 6.51438057e-01
-5.18265426e-01 -1.02347398e+00 -8.75546364e-04 3.04219544e-01
3.83911669e-01 9.05057043e-02 -5.36503792e-01 5.77132523e-01
9.06401634e-01 -6.38754189e-01 4.99132484e-01 2.53653646e-01
4.35870081e-01 -5.87890506e-01 4.58821297e-01 5.90574443e-01
-1.45000196e+00 -3.82127706e-03 -4.28604990e-01 -1.46716282e-01
5.72005510e-01 4.93063211e-01 -8.68056059e-01 7.11126745e-01
5.09565592e-01 1.26885980e-01 -3.82874787e-01 4.78721976e-01
-4.13086891e-01 1.21408999e+00 -4.87060785e-01 -1.19963828e-02
2.93658257e-01 -5.73119044e-01 6.53007925e-01 1.44354796e+00
2.40408361e-01 3.33093762e-01 5.19084513e-01 4.59662586e-01
4.04489100e-01 -8.96306857e-02 -3.91850740e-01 -5.77956498e-01
7.20635056e-01 9.76745486e-01 -8.54819715e-01 -5.58635116e-01
-4.75711644e-01 6.18683279e-01 4.40455735e-01 3.34729016e-01
-7.94571459e-01 -2.81831622e-01 6.62934422e-01 4.20315750e-02
2.81019270e-01 -7.04191208e-01 3.18088174e-01 -1.28883290e+00
-1.19630955e-02 -9.08061326e-01 6.65022969e-01 -5.18862605e-01
-7.51267791e-01 5.38737118e-01 3.94315243e-01 -7.41485536e-01
-8.36312413e-01 -5.17840683e-01 -2.55930811e-01 7.88462520e-01
-7.93321431e-01 -1.16579664e+00 2.65067399e-01 4.16748554e-01
4.83595222e-01 3.70079190e-01 1.09805286e+00 -1.38796801e-02
-8.00372243e-01 2.36821219e-01 -4.74459827e-02 1.00995623e-01
-7.82214925e-02 -1.33891368e+00 1.00069702e+00 1.31839156e+00
1.45315900e-01 1.06214106e+00 7.76980877e-01 -9.76917624e-01
-1.85305297e+00 -9.62260902e-01 1.25240600e+00 -4.65283304e-01
7.78237402e-01 -9.59428489e-01 -1.05628443e+00 1.23116529e+00
-1.17688015e-01 -3.21096271e-01 5.90960145e-01 5.74340522e-01
-3.92736107e-01 4.71359044e-01 -5.51714659e-01 8.10547471e-01
1.44238663e+00 -6.73066139e-01 -4.78991747e-01 3.51338148e-01
9.31148231e-01 -5.39321005e-01 -1.05686212e+00 2.62451410e-01
5.11716187e-01 -8.87277722e-01 6.07482076e-01 -7.03282475e-01
-9.62342769e-02 -5.15140414e-01 -4.09623414e-01 -1.02035582e+00
-4.91341442e-01 -1.02198398e+00 -5.73473752e-01 1.20301783e+00
4.80044812e-01 -8.61736655e-01 5.17254591e-01 7.57218897e-01
-3.68385732e-01 -5.65210521e-01 -8.70516479e-01 -1.17408025e+00
-1.18434615e-01 -8.53686154e-01 8.38371277e-01 5.63362658e-01
2.34685987e-01 3.49854052e-01 -3.45241696e-01 5.91448903e-01
5.81870079e-01 2.17661768e-01 8.69846523e-01 -1.11705661e+00
-5.47391951e-01 -7.15508237e-02 -3.40583652e-01 -9.42455113e-01
2.81286448e-01 -1.01809919e+00 9.73242894e-02 -1.62668383e+00
2.73738921e-01 -2.18046680e-01 3.50490510e-02 8.13482463e-01
-1.70486331e-01 -2.91453958e-01 3.73295128e-01 1.45123199e-01
-6.33166015e-01 6.38210252e-02 9.79999065e-01 3.44301373e-01
-2.10791633e-01 -2.48099864e-02 -3.10335487e-01 6.79732680e-01
8.95839453e-01 -4.78350878e-01 -2.91416943e-01 -3.02983701e-01
2.26678073e-01 5.24655700e-01 7.01851100e-02 -7.82164991e-01
1.07043371e-01 -3.40977222e-01 -1.89765707e-01 -8.04694116e-01
4.54134703e-01 -1.99954838e-01 8.88538659e-01 3.99365842e-01
-2.02996656e-01 2.55412519e-01 5.65133728e-02 6.11593306e-01
-1.61513165e-02 -1.07284315e-01 4.18783069e-01 -1.14688583e-01
-6.29619539e-01 1.11601077e-01 -7.78581321e-01 -2.72763312e-01
3.55782211e-01 -4.03517969e-02 -1.36636779e-01 -3.06084365e-01
-1.25901318e+00 -3.35463323e-02 3.50902796e-01 3.23430955e-01
3.38440895e-01 -1.11730778e+00 -3.39766264e-01 1.88421935e-01
2.56788507e-02 -3.75034064e-02 1.75886229e-01 7.72333980e-01
-3.79419625e-01 7.75972664e-01 2.97594398e-01 -5.59748173e-01
-1.28429878e+00 4.61880833e-01 -1.81467205e-01 -6.82817638e-01
-6.19611800e-01 5.52462935e-01 2.16460600e-01 -7.73915946e-01
-4.57897317e-03 -7.64354885e-01 1.80091813e-01 -3.57438624e-01
4.18075770e-01 5.64421356e-01 4.50784378e-02 -6.33028567e-01
-4.84326333e-01 6.31188080e-02 -9.13352147e-02 -4.00892466e-01
1.19186342e+00 -1.71510264e-01 -2.84855366e-01 5.26489139e-01
9.98685420e-01 -7.09350109e-02 -1.06254768e+00 -1.25815496e-01
5.44547677e-01 -2.39731967e-01 -2.67386913e-01 -4.49468911e-01
-4.55126703e-01 6.12547338e-01 1.10075690e-01 4.83710319e-01
8.14597011e-01 3.11622351e-01 7.90003002e-01 3.43836546e-01
8.06074917e-01 -9.11447346e-01 -5.95054805e-01 8.98937523e-01
4.53686059e-01 -6.00142062e-01 -3.33792597e-01 -7.85973489e-01
-6.23922408e-01 8.29045475e-01 2.16832146e-01 1.69713736e-01
4.14021462e-01 3.78536135e-01 -1.98032752e-01 -1.97883733e-02
-1.13734734e+00 -3.20263267e-01 -8.00727084e-02 3.98536593e-01
3.60486269e-01 6.96647465e-01 -4.42312717e-01 8.56811464e-01
-4.63668078e-01 -2.09281906e-01 4.54837620e-01 1.27256656e+00
-3.31096888e-01 -1.57297695e+00 -3.99892688e-01 3.95075381e-01
-4.61343229e-01 -2.13770270e-01 -1.47476196e-01 7.48557448e-01
-3.04362744e-01 1.22937930e+00 -1.30994216e-01 -5.38729608e-01
1.32474408e-01 6.13174558e-01 4.62892950e-01 -9.22097266e-01
-2.28205279e-01 4.22866672e-01 5.68913162e-01 -7.86140263e-01
-1.37858972e-01 -9.08576667e-01 -1.38682938e+00 -3.12852144e-01
-4.82917458e-01 4.55597460e-01 7.10156739e-01 9.05586541e-01
3.01339805e-01 1.74069300e-03 3.68472606e-01 -4.17735487e-01
-5.00910819e-01 -9.74828005e-01 -7.74288476e-01 1.63393646e-01
-1.10075824e-01 -6.00627422e-01 -3.30525428e-01 5.46829402e-02] | [10.362467765808105, 9.574357032775879] |
f6bef3cb-b9e7-41a1-8a2a-65fb5aac37ee | adaptive-graph-based-total-variation-for | 1610.00893 | null | http://arxiv.org/abs/1610.00893v3 | http://arxiv.org/pdf/1610.00893v3.pdf | Adaptive Graph-based Total Variation for Tomographic Reconstructions | Sparsity exploiting image reconstruction (SER) methods have been extensively
used with Total Variation (TV) regularization for tomographic reconstructions.
Local TV methods fail to preserve texture details and often create additional
artefacts due to over-smoothing. Non-Local TV (NLTV) methods have been proposed
as a solution to this but they either lack continuous updates due to
computational constraints or limit the locality to a small region. In this
paper, we propose Adaptive Graph-based TV (AGTV). The proposed method goes
beyond spatial similarity between different regions of an image being
reconstructed by establishing a connection between similar regions in the
entire image regardless of spatial distance. As compared to NLTV the proposed
method is computationally efficient and involves updating the graph prior
during every iteration making the connection between similar regions stronger.
Moreover, it promotes sparsity in the wavelet and graph gradient domains. Since
TV is a special case of graph TV the proposed method can also be seen as a
generalization of SER and TV methods. | ['Pierre Vandergheynst', 'Faisal Mahmood', 'Ulf Skoglund', 'Nauman Shahid'] | 2016-10-04 | null | null | null | null | ['tomographic-reconstructions'] | ['medical'] | [ 5.53672791e-01 7.81673193e-02 8.02049786e-03 -4.38856445e-02
-4.98872906e-01 -5.36807440e-02 4.54388112e-01 3.18507731e-01
-2.38625795e-01 7.69971371e-01 1.47383705e-01 4.83299121e-02
-4.30983275e-01 -8.48901391e-01 -5.14667571e-01 -9.27206993e-01
1.03125975e-01 1.92118108e-01 6.42249882e-01 -1.43073902e-01
2.32201874e-01 5.26243567e-01 -1.02594984e+00 1.30759925e-01
8.19723070e-01 6.74633205e-01 4.15950507e-01 2.70664006e-01
-2.73795873e-01 7.68155754e-01 -7.61172175e-02 1.07859701e-01
1.83198437e-01 -8.97852421e-01 -7.25633442e-01 5.74931860e-01
5.93228161e-01 6.33493215e-02 -1.30085722e-01 1.10630023e+00
3.69714022e-01 4.23425198e-01 3.45679849e-01 -6.65044963e-01
-3.89678419e-01 1.45139799e-01 -1.10676205e+00 3.41537237e-01
3.65207285e-01 -3.67602557e-01 6.44428015e-01 -1.10665929e+00
9.75175858e-01 9.72541273e-01 9.66439486e-01 1.27325743e-01
-1.73283052e+00 -5.54805845e-02 -3.62440124e-02 1.93224028e-01
-1.32584500e+00 -3.63104045e-01 1.12469685e+00 -3.10440212e-01
5.76588094e-01 4.81349796e-01 8.02688777e-01 4.92433220e-01
3.44040453e-01 5.21163702e-01 1.31884336e+00 -6.89212859e-01
3.11379671e-01 -9.83112305e-02 1.46953195e-01 6.97555363e-01
1.57140940e-01 -1.66057602e-01 -4.85821426e-01 -3.97698551e-01
1.14282727e+00 -8.54795203e-02 -6.23096108e-01 -6.73272252e-01
-1.06987846e+00 9.05087471e-01 4.06242818e-01 8.91426504e-01
-7.17782557e-01 1.42864734e-01 4.44759965e-01 2.83138812e-01
7.32996702e-01 1.95188317e-02 2.95229226e-01 3.96513432e-01
-1.25200415e+00 7.58767873e-02 4.57680196e-01 5.03974795e-01
9.78368163e-01 3.73217463e-01 -3.50887999e-02 8.86676431e-01
2.54506469e-01 2.05269963e-01 2.49732897e-01 -9.92596328e-01
7.03985170e-02 2.28845611e-01 -1.85069278e-01 -1.54854083e+00
-2.83365041e-01 -4.96825457e-01 -1.07165873e+00 2.55046099e-01
5.09025335e-01 3.96905869e-01 -8.10059309e-01 1.39373302e+00
5.56330204e-01 3.84781778e-01 -3.07770491e-01 8.32359731e-01
7.31687784e-01 7.16110528e-01 -6.90351650e-02 -7.78450906e-01
8.89947951e-01 -6.79303825e-01 -1.18226838e+00 -4.78036478e-02
5.06350875e-01 -8.67732584e-01 7.24995255e-01 4.99740332e-01
-1.18843675e+00 -4.51361507e-01 -7.29629457e-01 2.63976511e-02
3.95297110e-02 -3.17837089e-01 2.93102771e-01 5.79829812e-01
-1.39005435e+00 6.92726851e-01 -7.94804990e-01 -4.19893414e-01
5.55502325e-02 2.19127923e-01 -3.21439326e-01 -3.67898464e-01
-7.92155623e-01 7.49514759e-01 -4.82467972e-02 1.15073822e-01
-4.51535285e-01 -6.12361729e-01 -8.75372171e-01 -2.29479641e-01
2.78373361e-01 -4.81771499e-01 3.08477730e-01 -1.26279807e+00
-1.37296224e+00 8.13211441e-01 -5.14823735e-01 -3.21351677e-01
6.83017075e-01 2.58097857e-01 -2.07069710e-01 3.89387786e-01
8.23053792e-02 2.82577449e-03 1.26678777e+00 -1.42019248e+00
4.83488292e-02 -3.69638383e-01 -4.55641448e-01 1.19132802e-01
1.65757071e-02 -1.86004817e-01 -3.87361825e-01 -1.08852053e+00
7.59226680e-01 -7.29402602e-01 -3.69421214e-01 1.71843737e-01
-1.76015913e-01 1.43959671e-01 1.04182422e+00 -8.67666960e-01
1.16521239e+00 -2.22376990e+00 3.66215229e-01 6.32353067e-01
4.88296181e-01 -2.67660543e-02 2.99778930e-03 5.74609816e-01
-8.80778581e-02 -2.61054218e-01 -6.68254256e-01 -2.88973033e-01
-6.27374470e-01 4.83469337e-01 2.32506946e-01 1.06875789e+00
-3.37956131e-01 4.62186128e-01 -8.84460270e-01 -8.01549673e-01
5.97306311e-01 8.79038334e-01 -4.92417246e-01 -3.38440269e-01
1.65029079e-01 9.53243852e-01 -5.31412840e-01 1.60426557e-01
9.82965708e-01 -2.25688681e-01 4.29076165e-01 -2.73952216e-01
-2.64925152e-01 -2.41116241e-01 -1.25680923e+00 1.51472390e+00
-3.80827487e-01 5.56903303e-01 5.22804916e-01 -1.57337213e+00
9.84350562e-01 5.31025469e-01 1.00719285e+00 -6.54411972e-01
-6.99810833e-02 3.98640007e-01 -4.26686168e-01 -4.52242434e-01
3.62215191e-01 -5.76129794e-01 6.82396650e-01 1.50084212e-01
-2.62620896e-01 -1.12041719e-01 1.20832786e-01 1.73258275e-01
9.96622324e-01 -1.11664450e-02 5.34207940e-01 -6.66947424e-01
7.10341215e-01 -1.58455834e-01 6.16352677e-01 7.20625341e-01
-1.05614819e-01 7.94888496e-01 1.19895913e-01 -4.90855664e-01
-1.07283390e+00 -8.57388318e-01 -5.19855917e-01 3.11813354e-01
3.14392745e-01 -3.53057384e-01 -3.69099796e-01 -3.52592796e-01
-3.34035397e-01 4.71673548e-01 -6.55347705e-01 2.68735737e-01
-8.67411613e-01 -6.68692350e-01 -6.71422407e-02 -2.16901973e-02
6.70592189e-01 -7.85195172e-01 -3.81371886e-01 5.97011626e-01
-4.41138446e-01 -1.13775039e+00 -3.62288952e-01 -1.07139498e-01
-1.39408672e+00 -8.90436530e-01 -1.16882145e+00 -8.24198902e-01
7.90989995e-01 4.32494491e-01 9.25272822e-01 2.59579092e-01
-1.36589035e-01 4.92083013e-01 -4.02327478e-01 3.82669270e-01
-3.66597950e-01 -5.39451301e-01 -2.63820648e-01 5.94230950e-01
-3.83684993e-01 -7.71485925e-01 -5.22759914e-01 3.82043004e-01
-1.02575552e+00 -4.75786487e-03 1.39894933e-01 1.03507507e+00
1.14388335e+00 1.76278651e-01 2.70576686e-01 -1.13041067e+00
4.43350673e-01 -3.56274396e-01 -4.91189361e-01 2.85855472e-01
-5.54892838e-01 -9.06233564e-02 4.38655883e-01 -2.37886205e-01
-1.14870119e+00 1.34718120e-01 -3.87049988e-02 -4.71799940e-01
2.22955525e-01 7.57953346e-01 4.13766384e-01 -8.16311419e-01
5.33628941e-01 5.40999651e-01 1.53997511e-01 -5.37190020e-01
-3.34739275e-02 -1.66255012e-01 1.47216022e-01 -3.07877153e-01
4.91005957e-01 9.55520153e-01 5.48151374e-01 -1.60753322e+00
-3.85320604e-01 -6.16926193e-01 -4.79270279e-01 -5.07889330e-01
8.74422371e-01 -3.54369670e-01 -2.60790676e-01 1.56622067e-01
-8.79329801e-01 -2.99493492e-01 -4.77894336e-01 6.01596057e-01
-5.72203815e-01 1.12803519e+00 -4.70007986e-01 -8.70949328e-01
-7.39893317e-02 -1.04938209e+00 6.86704040e-01 -1.92560568e-01
-4.69225198e-02 -1.64943862e+00 1.98793709e-01 9.75376964e-02
7.19929338e-01 7.81464577e-01 6.76682591e-01 2.03276336e-01
-2.69183099e-01 -8.58293548e-02 -1.29461214e-01 2.28784129e-01
2.19849750e-01 -2.41635546e-01 -4.05060679e-01 -5.80404758e-01
6.37706816e-01 9.24709067e-02 8.16922784e-01 1.18347311e+00
8.28021705e-01 -2.40033507e-01 -2.94016600e-01 5.98011136e-01
1.88133693e+00 -1.62654221e-01 8.68569314e-01 3.90604705e-01
6.56255841e-01 6.19321048e-01 2.57248551e-01 4.50152844e-01
-1.95688382e-01 9.45130944e-01 2.64443696e-01 -5.34614861e-01
-5.20951748e-01 1.59859121e-01 8.55066180e-02 9.04578149e-01
-4.42175329e-01 -1.22238860e-01 -6.42549634e-01 6.59671247e-01
-1.85250223e+00 -9.70792294e-01 -8.70765924e-01 2.40237999e+00
5.50808132e-01 -2.08267197e-01 -4.36288267e-02 4.70914871e-01
9.15404499e-01 4.25177157e-01 -4.42151129e-02 -3.79648358e-01
-3.37254018e-01 2.88508207e-01 6.74053848e-01 9.06504214e-01
-7.35089302e-01 5.43990433e-01 6.23584509e+00 1.14676702e+00
-1.02195227e+00 4.85121965e-01 3.32876772e-01 3.09104651e-01
-4.07780886e-01 2.47936308e-01 -1.11421093e-01 1.23281181e-01
3.58391762e-01 2.12179292e-02 2.98613846e-01 3.07386070e-01
5.12531102e-01 -5.23449004e-01 -4.15527731e-01 9.94866490e-01
6.67817891e-02 -1.28002846e+00 7.85821527e-02 2.08528340e-01
8.57025087e-01 -1.86742619e-01 -4.70848829e-02 -5.30981481e-01
-2.57823408e-01 -7.95361936e-01 3.90598029e-01 4.89195317e-01
6.76767826e-01 -5.96638918e-01 6.26746655e-01 1.11785002e-01
-1.38803744e+00 4.63044286e-01 -3.73119473e-01 7.61976242e-02
5.78390598e-01 1.02062559e+00 -3.26304525e-01 8.29354823e-01
5.73317826e-01 8.64464223e-01 -1.75791398e-01 9.74419892e-01
4.23164815e-02 5.23391187e-01 -4.96350676e-01 4.61876631e-01
4.59176093e-01 -5.98473310e-01 1.07300258e+00 8.69878769e-01
3.51251870e-01 1.75875276e-01 3.74821961e-01 7.42380202e-01
3.94370258e-01 4.23076153e-01 -7.19880700e-01 3.82827878e-01
-1.08488306e-01 8.53071094e-01 -1.06049263e+00 -3.90843391e-01
-5.59093893e-01 9.94580090e-01 5.25422627e-03 5.81771076e-01
-3.47568125e-01 8.44592452e-02 -1.74697548e-01 5.86905003e-01
3.52430135e-01 -2.70562708e-01 -1.85047641e-01 -1.08993053e+00
-8.47491696e-02 -5.85697055e-01 5.02896428e-01 -4.86299068e-01
-1.14473343e+00 5.55676341e-01 4.50150631e-02 -1.27088225e+00
3.12932312e-01 -6.24038279e-02 -4.04926538e-01 6.83781743e-01
-1.51290703e+00 -1.08641553e+00 -2.57013947e-01 1.08810532e+00
6.34981573e-01 4.62208807e-01 5.61134636e-01 3.22801560e-01
-6.84906766e-02 1.94771290e-01 3.66432041e-01 -3.66286993e-01
4.19204265e-01 -9.31408226e-01 -3.27477485e-01 1.00987244e+00
-5.31855524e-02 5.25671005e-01 9.96355414e-01 -9.73087847e-01
-1.22174978e+00 -7.25033998e-01 9.74175453e-01 3.75747949e-01
3.84603202e-01 2.20644683e-01 -1.25140762e+00 5.72380602e-01
3.54385734e-01 3.04575980e-01 3.05082768e-01 -1.70403346e-01
1.32673725e-01 5.42853363e-02 -1.34688938e+00 2.35985681e-01
6.83602810e-01 -3.51912975e-01 -1.75843313e-01 3.93196821e-01
-1.15364879e-01 -2.60567158e-01 -1.05015242e+00 2.94141233e-01
1.92589626e-01 -1.28807962e+00 9.96215820e-01 3.12573582e-01
1.04144998e-01 -4.32688594e-01 -2.02832813e-03 -1.02627254e+00
-4.90154624e-01 -6.48723483e-01 3.04132611e-01 8.39684248e-01
2.63788681e-02 -8.64919126e-01 7.08272934e-01 8.01904127e-02
-1.89399585e-01 -4.14033622e-01 -1.30875504e+00 -9.14248586e-01
-3.48106354e-01 -3.79634202e-01 -1.68847501e-01 1.31733346e+00
9.27739516e-02 -3.34831625e-02 -7.44447708e-01 4.87679988e-02
1.25657761e+00 9.26008895e-02 3.08094591e-01 -1.02021849e+00
-4.48878586e-01 -3.17976236e-01 -4.78377402e-01 -8.36757481e-01
-3.10696244e-01 -9.26847875e-01 -1.27715811e-01 -1.63104558e+00
8.30596760e-02 -5.70628524e-01 -1.41240284e-01 1.77879930e-01
1.76522911e-01 5.67986846e-01 -1.67897437e-02 5.59440255e-01
-1.25745595e-01 4.20035720e-01 1.65347862e+00 6.49439618e-02
-1.76854759e-01 -8.25281069e-02 1.25958994e-01 5.61465800e-01
5.42318285e-01 -4.91105944e-01 -4.02076095e-01 -3.21977109e-01
2.10239664e-01 5.19720554e-01 3.84941816e-01 -8.01787615e-01
1.90169066e-01 1.73778199e-02 1.09472774e-01 -4.14500058e-01
4.06141222e-01 -9.19179201e-01 6.72801673e-01 6.41819358e-01
-1.55370668e-01 -1.47504747e-01 -4.19255495e-02 8.26049149e-01
-4.49560255e-01 -3.25238466e-01 1.23247993e+00 -4.24341977e-01
-6.09978974e-01 1.42972618e-01 -5.67791045e-01 -2.74361432e-01
8.00624251e-01 -6.67966962e-01 3.56351018e-01 -6.31680191e-01
-1.18946588e+00 -2.44492054e-01 6.01825118e-01 -4.12724853e-01
9.27439690e-01 -1.15669537e+00 -8.19701374e-01 1.10949390e-01
-2.72776693e-01 -2.88123816e-01 5.75585663e-01 1.65038025e+00
-9.03007567e-01 1.61562279e-01 -1.53251812e-01 -8.52966011e-01
-1.44500482e+00 5.95215499e-01 2.66661406e-01 -5.00150442e-01
-1.37774348e+00 6.19919002e-01 4.64013577e-01 1.07931964e-01
-1.42538816e-01 5.82014173e-02 -2.32919991e-01 -1.48142859e-01
3.02182585e-01 5.50788105e-01 3.76291722e-02 -9.34188306e-01
-3.31204295e-01 1.10892797e+00 2.86593854e-01 -2.10816607e-01
1.47576022e+00 -4.47343349e-01 -5.21125019e-01 3.13472360e-01
1.23619998e+00 2.54732102e-01 -9.40214336e-01 -3.69989365e-01
-7.16602132e-02 -7.22368836e-01 7.02010453e-01 -3.18553373e-02
-1.47691691e+00 6.07766926e-01 5.40818155e-01 5.06570578e-01
1.29943311e+00 -1.13824077e-01 7.39417315e-01 -3.80212605e-01
3.23499143e-01 -9.27776158e-01 -3.83704379e-02 5.08467592e-02
1.01517808e+00 -9.80090141e-01 6.15460277e-01 -7.76622832e-01
-3.61274838e-01 1.20375276e+00 -1.84845880e-01 -4.77465451e-01
7.82530963e-01 -4.02308330e-02 -2.38815442e-01 -5.49162567e-01
-7.79698268e-02 -1.45086706e-01 3.61056060e-01 5.29614508e-01
5.95134676e-01 -2.33851582e-01 -1.13200986e+00 -5.99051774e-01
5.32928050e-01 -2.74742432e-02 4.52524126e-01 1.05054653e+00
-2.77890027e-01 -9.73924756e-01 -6.75265849e-01 2.13879421e-01
-4.74981338e-01 -6.82976842e-02 1.60734840e-02 7.75590062e-01
-1.73842788e-01 1.01088536e+00 -3.02587330e-01 7.93332532e-02
1.49318308e-01 -3.88985395e-01 7.43971825e-01 -2.73566008e-01
-4.80902433e-01 7.69610405e-01 1.02651559e-01 -7.00550020e-01
-1.02341688e+00 -7.98805296e-01 -1.05416322e+00 -3.90462160e-01
-5.06343126e-01 1.89296216e-01 5.21382987e-01 8.76515627e-01
-1.22572161e-01 4.70899135e-01 5.04445672e-01 -9.38683867e-01
2.19373971e-01 -6.42131329e-01 -1.01262510e+00 4.50997382e-01
2.87664860e-01 -5.68433285e-01 -5.04332900e-01 1.41044006e-01] | [11.675834655761719, -2.4191482067108154] |
acbbbae0-6cd3-474a-8285-539ca4002c2f | online-categorical-subspace-learning-for | 1609.08235 | null | http://arxiv.org/abs/1609.08235v1 | http://arxiv.org/pdf/1609.08235v1.pdf | Online Categorical Subspace Learning for Sketching Big Data with Misses | With the scale of data growing every day, reducing the dimensionality (a.k.a.
sketching) of high-dimensional data has emerged as a task of paramount
importance. Relevant issues to address in this context include the sheer volume
of data that may consist of categorical samples, the typically streaming format
of acquisition, and the possibly missing entries. To cope with these
challenges, the present paper develops a novel categorical subspace learning
approach to unravel the latent structure for three prominent categorical
(bilinear) models, namely, Probit, Tobit, and Logit. The deterministic Probit
and Tobit models treat data as quantized values of an analog-valued process
lying in a low-dimensional subspace, while the probabilistic Logit model relies
on low dimensionality of the data log-likelihood ratios. Leveraging the low
intrinsic dimensionality of the sought models, a rank regularized
maximum-likelihood estimator is devised, which is then solved recursively via
alternating majorization-minimization to sketch high-dimensional categorical
data `on the fly.' The resultant procedure alternates between sketching the new
incomplete datum and refining the latent subspace, leading to lightweight
first-order algorithms with highly parallelizable tasks per iteration. As an
extra degree of freedom, the quantization thresholds are also learned jointly
along with the subspace to enhance the predictive power of the sought models.
Performance of the subspace iterates is analyzed for both infinite and finite
data streams, where for the former asymptotic convergence to the stationary
point set of the batch estimator is established, while for the latter sublinear
regret bounds are derived for the empirical cost. Simulated tests with both
synthetic and real-world datasets corroborate the merits of the novel schemes
for real-time movie recommendation and chess-game classification. | ['Georgios B. Giannakis', 'Yanning Shen', 'Morteza Mardani'] | 2016-09-27 | null | null | null | null | ['movie-recommendation'] | ['miscellaneous'] | [ 2.52678782e-01 -2.12182164e-01 -3.60839725e-01 -8.96898508e-02
-9.67062712e-01 -9.69195545e-01 5.60053468e-01 -8.98555666e-02
-2.77661145e-01 5.89780331e-01 3.28891158e-01 -9.58807245e-02
-7.87288070e-01 -3.31800640e-01 -5.61402440e-01 -1.00604212e+00
-1.28034741e-01 6.16258800e-01 -3.32652301e-01 1.50985435e-01
3.75150025e-01 4.67125326e-01 -1.57552707e+00 8.51719007e-02
7.06843972e-01 1.21265519e+00 -1.36137918e-01 6.78822041e-01
-1.62431881e-01 1.41519487e-01 -6.55512959e-02 -5.85365117e-01
5.78826368e-01 -4.30170819e-02 -9.91788432e-02 3.82413894e-01
2.99157590e-01 -4.49116498e-01 -4.38329220e-01 1.01784670e+00
2.89535493e-01 1.13481665e-02 9.03427601e-01 -1.27726674e+00
-1.28906950e-01 2.54732251e-01 -5.25950611e-01 -1.89286396e-01
1.99876115e-01 -1.67978719e-01 1.20325959e+00 -1.36829150e+00
3.63541186e-01 1.06447780e+00 6.17069423e-01 2.11402625e-01
-1.60220623e+00 -5.50375879e-01 1.34826377e-02 -1.61537573e-01
-1.55029786e+00 -5.39783001e-01 9.48866963e-01 -5.88821054e-01
1.27658486e-01 3.88205945e-01 4.67631668e-01 1.14415336e+00
-1.92982256e-01 8.45767856e-01 1.14282048e+00 -2.19490603e-01
6.72159791e-01 2.33785838e-01 2.01285422e-01 5.35571337e-01
4.30552751e-01 5.44125736e-02 -9.20392394e-01 -8.42038572e-01
9.16232705e-01 3.62444550e-01 -6.98873624e-02 -1.13846731e+00
-1.03555202e+00 8.48564625e-01 -3.51582557e-01 -1.09267026e-01
-5.20501733e-01 -1.71422996e-02 3.79516423e-01 2.42932588e-01
4.30917621e-01 5.21315336e-02 -2.44525418e-01 -3.99214238e-01
-1.22555900e+00 3.90611023e-01 1.15555429e+00 1.00925577e+00
4.88290131e-01 -2.83931103e-02 4.73477095e-02 8.18663657e-01
2.17262402e-01 4.63856548e-01 1.39232099e-01 -1.05470574e+00
8.05688798e-01 3.83237034e-01 4.08945739e-01 -9.31203365e-01
6.11047372e-02 -5.10007143e-01 -1.32246792e+00 -9.47338641e-02
7.99320102e-01 1.23557672e-01 -4.35724437e-01 1.59656870e+00
4.08469647e-01 7.05984309e-02 4.71571870e-02 1.04143143e+00
4.21657898e-02 6.02377176e-01 -1.81542188e-01 -6.29540622e-01
1.20010900e+00 -3.35989863e-01 -6.26077354e-01 2.80834109e-01
2.12708414e-01 -3.97083819e-01 9.00249124e-01 8.86814058e-01
-1.08065748e+00 -3.23586553e-01 -9.48211730e-01 1.02652922e-01
4.68884557e-02 3.50644410e-01 6.67723775e-01 6.78474247e-01
-7.38874018e-01 5.53235173e-01 -6.16046846e-01 3.57019529e-03
3.05215508e-01 3.66595924e-01 -1.94500819e-01 -2.48288810e-01
-9.62435782e-01 -2.97067016e-02 4.26233262e-02 2.23181561e-01
-7.20904529e-01 -7.70223677e-01 -3.63367140e-01 1.50438890e-01
4.70781058e-01 -4.88426805e-01 7.75729239e-01 -6.11992121e-01
-1.46754074e+00 4.06811684e-01 -2.65414953e-01 -2.61677235e-01
8.72295380e-01 -2.32075363e-01 2.18100753e-02 2.59324521e-01
-4.14026119e-02 8.88188854e-02 1.45327592e+00 -1.21825969e+00
-5.45723617e-01 -6.05483353e-01 -1.49444491e-01 4.30527091e-01
-5.45327246e-01 -4.39382076e-01 -5.69229186e-01 -8.53475690e-01
5.73605239e-01 -1.00248623e+00 -2.36778244e-01 2.91280001e-01
-1.46414891e-01 -1.36905104e-01 5.55010378e-01 -7.11503923e-01
1.34955084e+00 -2.47490263e+00 5.52093327e-01 4.47939724e-01
2.69192785e-01 -1.65697977e-01 2.06309035e-01 6.32973433e-01
1.27784774e-01 -2.46472523e-01 -4.30245161e-01 -5.34223914e-01
2.11993486e-01 2.85858214e-01 -8.61868441e-01 6.82712078e-01
-1.35408863e-01 5.58887601e-01 -7.25982785e-01 -2.01738343e-01
1.20317973e-02 1.98388711e-01 -6.73001349e-01 3.41299832e-01
2.03524083e-02 2.54138976e-01 -4.80933547e-01 6.17167115e-01
7.30554760e-01 -1.85707286e-01 3.83228213e-01 -1.50626719e-01
-6.47910088e-02 -1.40732273e-01 -1.88624537e+00 1.84270203e+00
-6.89154193e-02 5.87642454e-02 4.83778149e-01 -1.06638813e+00
7.89641917e-01 3.35349977e-01 6.72504067e-01 -2.72107840e-01
-3.52208525e-01 4.18715149e-01 -5.95994413e-01 -2.51778722e-01
5.95542729e-01 -2.96284139e-01 -4.00062144e-01 3.41977775e-01
-4.32583084e-03 5.43640405e-02 4.08812091e-02 2.90598810e-01
7.85885096e-01 1.22850850e-01 1.52513459e-01 -2.66404659e-01
4.21090871e-01 -1.88107312e-01 4.83565897e-01 9.52680945e-01
-5.17667346e-02 5.20360827e-01 8.52257967e-01 -3.73998940e-01
-1.27595079e+00 -1.24515057e+00 -2.35494301e-01 9.85923231e-01
1.23761006e-01 -3.30612838e-01 -5.15967011e-01 -4.57475036e-01
3.60969186e-01 4.72580880e-01 -4.15177941e-01 4.31252308e-02
-2.63159484e-01 -6.56602383e-01 3.11346054e-01 7.82035962e-02
3.10496129e-02 -2.90492326e-01 -4.03347045e-01 3.10728580e-01
-9.36845019e-02 -8.55387509e-01 -4.12178069e-01 2.98654109e-01
-1.12547231e+00 -7.75946617e-01 -8.43639374e-01 -2.68775761e-01
5.14359534e-01 2.52077311e-01 5.80274999e-01 -6.46480501e-01
-7.59602934e-02 6.90256119e-01 -1.50721237e-01 1.55423462e-01
-8.15723091e-03 -1.41419187e-01 4.78695422e-01 6.32949233e-01
4.11290005e-02 -7.03415215e-01 -6.29279912e-01 2.44795784e-01
-1.11753011e+00 8.40364099e-02 6.38804436e-01 1.14789379e+00
6.50686979e-01 7.70436926e-03 5.14991403e-01 -6.82470322e-01
6.32609010e-01 -7.56170332e-01 -7.17193365e-01 1.04844473e-01
-6.15947783e-01 2.34879896e-01 7.77978003e-01 -7.35137761e-01
-8.40976834e-01 2.95785934e-01 3.88843298e-01 -8.47451329e-01
-7.20771626e-02 4.82348979e-01 -1.82162702e-01 2.65304536e-01
3.07146311e-01 5.71734786e-01 6.40541092e-02 -7.18482614e-01
5.87144732e-01 7.27488816e-01 5.63213944e-01 -8.60513926e-01
9.20732141e-01 7.36592352e-01 3.18855226e-01 -1.04971886e+00
-6.24609888e-01 -7.86811829e-01 -7.16958106e-01 -1.24326579e-01
3.95682484e-01 -9.71920848e-01 -6.21936321e-01 2.84560591e-01
-8.93705785e-01 2.28859469e-01 -5.60378015e-01 5.30596316e-01
-8.09486568e-01 5.94205022e-01 -5.81990719e-01 -1.28146303e+00
-1.51499838e-01 -8.04295063e-01 1.15795970e+00 -2.94488221e-01
-1.25855014e-01 -6.92108154e-01 1.51551530e-01 2.80666381e-01
-9.24752131e-02 -4.37471531e-02 1.14344835e+00 -5.73432148e-01
-6.90395415e-01 -5.69660664e-01 -1.54269382e-01 1.97244629e-01
-1.94728658e-01 -4.98286098e-01 -7.75063753e-01 -6.11518204e-01
3.19052219e-01 -2.88625002e-01 5.61596096e-01 2.37885132e-01
1.10730267e+00 -6.73530221e-01 -5.63021339e-02 7.02620924e-01
1.34522879e+00 -1.46304682e-01 1.13742419e-01 -1.68487892e-01
6.21776581e-01 8.14252675e-01 6.97333217e-01 1.17662609e+00
2.19443336e-01 5.90214431e-01 1.54151231e-01 3.51530701e-01
5.21265566e-01 -5.16058087e-01 3.52871269e-01 1.15221608e+00
1.59324005e-01 7.82192498e-02 -5.96386552e-01 3.87553811e-01
-1.92149258e+00 -1.03150892e+00 1.72607265e-02 2.67163825e+00
6.74841762e-01 -7.03245103e-02 3.43482435e-01 4.84450281e-01
5.77085197e-01 1.50737226e-01 -8.26883912e-01 -6.80637956e-02
-1.08195610e-01 -2.14394912e-01 4.85794097e-01 2.25757986e-01
-1.12544274e+00 4.05032009e-01 5.28645229e+00 1.05249584e+00
-7.96292782e-01 -1.74958929e-01 5.73853254e-01 -3.06765497e-01
-4.15709436e-01 1.49102658e-01 -7.39693701e-01 6.18231058e-01
8.60348642e-01 -1.35331079e-01 7.95835018e-01 7.81612515e-01
2.14514017e-01 -1.48434667e-02 -1.16357052e+00 1.46632648e+00
-9.08031166e-02 -1.10524464e+00 1.82694182e-01 5.08481801e-01
6.00571096e-01 -4.40163910e-01 3.26059341e-01 2.38637805e-01
-8.89582112e-02 -5.39289594e-01 9.34603035e-01 8.43998075e-01
1.13662171e+00 -8.65012586e-01 2.04916582e-01 7.23266304e-01
-1.10705638e+00 -4.44445968e-01 -5.11029780e-01 -6.17531314e-03
-5.44316368e-03 6.71938837e-01 -4.03881937e-01 4.30367947e-01
4.68798995e-01 6.08299315e-01 -1.74234688e-01 8.59700382e-01
4.32333678e-01 7.36271262e-01 -6.55361116e-01 8.14978778e-02
1.07592307e-01 -8.59395921e-01 8.59647810e-01 9.95354116e-01
4.82599825e-01 4.10705209e-01 3.50409113e-02 7.27268159e-01
1.07158914e-01 7.23037124e-02 -6.20472431e-01 -1.10814683e-01
4.28767323e-01 1.18983257e+00 -5.17129064e-01 -1.78395391e-01
-3.00110757e-01 1.11420500e+00 2.13301748e-01 6.42138064e-01
-4.41186368e-01 -4.61083662e-04 5.44711053e-01 1.67908326e-01
4.81178492e-01 -6.00979507e-01 -5.61297297e-01 -1.37657225e+00
2.95123667e-01 -9.85268176e-01 5.10761738e-01 -6.26127124e-02
-1.38956904e+00 1.37289520e-03 1.11445524e-01 -1.52106154e+00
-4.11276877e-01 -3.84051770e-01 -1.32162720e-01 8.35518181e-01
-1.06614828e+00 -8.25066090e-01 6.48323372e-02 6.24258339e-01
5.51387906e-01 -3.06531757e-01 7.11410820e-01 2.13480964e-01
-4.28448826e-01 6.23363197e-01 1.03782368e+00 -3.20612043e-01
3.78062755e-01 -1.24516845e+00 6.31732270e-02 5.42473614e-01
1.84352189e-01 6.61023855e-01 6.22638464e-01 -7.02425718e-01
-2.21463823e+00 -6.86646938e-01 4.97629493e-01 -3.38027686e-01
9.08907056e-01 -8.83238494e-01 -9.32536781e-01 3.18422280e-02
-7.03736067e-01 -4.10936661e-02 6.91596150e-01 5.35650216e-02
-4.19621319e-01 -2.09933087e-01 -9.51212883e-01 5.59907913e-01
7.97223330e-01 -8.68823111e-01 -2.86196530e-01 1.26772940e-01
3.58757079e-01 -5.07979169e-02 -9.06243920e-01 1.30792886e-01
8.65636647e-01 -7.19429731e-01 1.00277889e+00 -7.11487055e-01
2.94881582e-01 -9.02379602e-02 -5.23210406e-01 -7.31189907e-01
-2.69816935e-01 -9.35568869e-01 -6.80055737e-01 1.24021482e+00
-8.31053406e-03 -2.15920016e-01 1.15883696e+00 8.90082538e-01
4.31706160e-01 -8.31447840e-01 -1.40315759e+00 -6.92526281e-01
-5.38938008e-02 -5.14759719e-01 1.65086016e-01 7.89137959e-01
5.09882346e-02 2.44104981e-01 -7.36895323e-01 1.87625989e-01
1.30901945e+00 3.62354696e-01 8.43838274e-01 -1.32020712e+00
-5.85881591e-01 -9.85184163e-02 -1.73803389e-01 -1.46896851e+00
-8.59432574e-03 -8.70728791e-01 -1.98154107e-01 -7.88783371e-01
3.13656032e-01 -5.77967465e-01 -2.58437783e-01 -2.12713525e-01
2.00894047e-02 -1.11075215e-01 2.52193004e-01 7.00776398e-01
-6.24343693e-01 8.38177621e-01 7.27385283e-01 4.01208224e-03
-3.12116355e-01 2.87987530e-01 -3.99797082e-01 5.08041382e-01
2.16155306e-01 -5.08203447e-01 -4.61309522e-01 -2.70915367e-02
3.00113589e-01 7.95816720e-01 4.29030180e-01 -6.58505917e-01
3.57064724e-01 -1.05713464e-01 3.05737466e-01 -7.77138591e-01
6.61801815e-01 -9.46699977e-01 2.21700445e-01 1.72050238e-01
-5.19925833e-01 -2.13540405e-01 -2.14135110e-01 1.24437070e+00
-2.56231725e-02 -1.73774719e-01 5.10592163e-01 2.79457629e-01
-1.81269780e-01 4.47419256e-01 -3.04249227e-01 1.39247437e-04
7.94285059e-01 -2.81521380e-01 3.53435189e-01 -5.75141907e-01
-7.26352334e-01 -4.47054803e-02 3.15508872e-01 2.54283577e-01
7.06684828e-01 -1.44850588e+00 -5.55657208e-01 3.94323379e-01
1.18868088e-03 -1.15257189e-01 3.73611540e-01 7.38615334e-01
1.37308747e-01 5.19543946e-01 1.39130160e-01 -6.93561018e-01
-7.40038931e-01 7.05212891e-01 -1.25020728e-01 -3.10277462e-01
-3.82446110e-01 4.68346626e-01 3.19412112e-01 -2.35666156e-01
4.73939866e-01 -2.01605186e-02 2.26363470e-03 4.26194787e-01
2.79063016e-01 7.67410159e-01 -2.42159098e-01 -4.89777654e-01
7.76322260e-02 2.62801528e-01 -8.75919014e-02 -4.68560815e-01
1.34182465e+00 -3.92250448e-01 -3.12533565e-02 8.47083569e-01
1.22327697e+00 -6.31094277e-02 -1.77418184e+00 -5.66272795e-01
2.87978381e-01 -6.29160523e-01 -6.01871274e-02 -2.75643855e-01
-4.23118293e-01 8.05263579e-01 4.98971999e-01 2.03834087e-01
9.02259231e-01 -2.48580709e-01 4.97940809e-01 3.15078527e-01
6.49250269e-01 -1.01874816e+00 -5.01260944e-02 3.57056171e-01
9.00688767e-01 -9.03523028e-01 -5.22003393e-04 -2.40041777e-01
-6.27615213e-01 1.08626878e+00 -1.03090301e-01 -1.21508375e-01
6.20160818e-01 -3.09385806e-02 -4.85810995e-01 7.66934976e-02
-7.63703287e-01 3.71879131e-01 4.83764559e-01 1.70020804e-01
-1.35730490e-01 1.53463751e-01 -2.84587234e-01 1.00076616e+00
6.41594231e-02 -1.06204093e-01 3.72802913e-01 5.83174348e-01
-6.93186373e-02 -8.65368605e-01 -5.81371486e-01 8.40195358e-01
-2.44765356e-01 -5.89026976e-03 -1.46754086e-01 3.09190989e-01
-3.36627930e-01 7.35870779e-01 -1.25626475e-01 -1.26945019e-01
3.48967850e-01 2.35198498e-01 1.30545750e-01 -1.65264264e-01
-2.40556314e-03 4.78991777e-01 -2.48400688e-01 -5.47247767e-01
1.94508746e-01 -1.21713352e+00 -6.66460454e-01 -1.89924389e-01
-1.26656920e-01 3.58916491e-01 7.38283098e-01 7.19955921e-01
2.78922230e-01 -3.26142699e-01 1.01391506e+00 -1.04614520e+00
-1.57049513e+00 -4.45271611e-01 -9.96836901e-01 5.00124753e-01
4.23313826e-01 -5.27583122e-01 -5.96777856e-01 -2.61823758e-02] | [7.3801703453063965, 4.349123477935791] |
06b149ed-0a9f-4c03-8efb-2d00448de775 | hand-pose-estimation-in-the-task-of | null | null | https://ieeexplore.ieee.org/document/9319235 | https://ieeexplore.ieee.org/document/9319235 | Hand Pose Estimation in the Task of Egocentric Actions | In this article we tackle the problem of hand pose estimation when the hand is interacting with various objects from egocentric viewpoint. This entails a frequent occlusion of parts of the hand by the object and also self-occlusions of the hand. We use a Voxel-to-Voxel approach to obtain hypotheses of the hand joint locations, ensemble the hypotheses and use several post-processing strategies to improve on the results. We utilize models of prior hand pose in the form of Truncated Singular Value Decomposition (SVD) and the temporal context to produce refined hand joint locations. We present an ablation study of the methods to show the influence of individual features of the post-processing. With our method we were able to constitute state-of-the-art results on the HANDS19 Challenge: Task 2 - Depth-Based 3D Hand Pose Estimation while Interacting with Objects, with precision on unseen test data of 33.09 mm. | ['Zdenĕk Krňoul', 'Jakub Kanis', 'Marek Hrúz'] | 2021-01-11 | null | null | null | ieee-access-2021-1 | ['3d-hand-pose-estimation', 'pose-estimation', 'hand-pose-estimation', '3d-hand-pose-estimation'] | ['computer-vision', 'computer-vision', 'computer-vision', 'graphs'] | [-8.76260027e-02 5.72695099e-02 4.78658170e-01 -1.02420747e-01
-6.14650071e-01 -6.60872400e-01 4.50177580e-01 -2.52773672e-01
-6.83403850e-01 5.74244320e-01 3.57496828e-01 1.56663403e-01
-1.78862527e-01 -4.44128029e-02 -4.47366804e-01 -6.61298275e-01
-3.76005881e-02 1.12990224e+00 5.07393241e-01 4.18123566e-02
3.29123259e-01 1.01575172e+00 -1.55290353e+00 3.42361033e-01
-4.19625528e-02 5.87620199e-01 2.76327223e-01 8.43822479e-01
4.40286636e-01 1.01674207e-01 -5.68930805e-01 -1.23763181e-01
3.97776514e-01 2.10407674e-01 -9.32114959e-01 2.75472611e-01
6.30979717e-01 -6.88668728e-01 -3.41684997e-01 3.80711257e-01
1.13766229e+00 2.70350546e-01 8.49064529e-01 -8.18765700e-01
2.69220889e-01 1.21294022e-01 -8.74895334e-01 2.19148010e-01
7.62129664e-01 1.72802448e-01 6.63465917e-01 -1.18191409e+00
1.24515283e+00 1.30219460e+00 5.41937172e-01 4.31926936e-01
-1.15400469e+00 -8.88016522e-02 3.73031616e-01 3.57014269e-01
-1.67850542e+00 -4.83621478e-01 5.59279740e-01 -6.82696283e-01
1.35292482e+00 3.51474255e-01 6.78913414e-01 1.14918864e+00
8.99616927e-02 7.52059281e-01 1.05003309e+00 -8.20753694e-01
9.32898074e-02 -1.42914042e-01 1.19946286e-01 4.28657830e-01
4.05315086e-02 -1.93935279e-02 -6.29470229e-01 -2.64108986e-01
1.08876526e+00 -2.58595794e-01 -1.16328031e-01 -6.06958747e-01
-1.29059482e+00 2.35100836e-01 3.33595484e-01 4.11459208e-01
-7.07874656e-01 1.52978480e-01 2.35210255e-01 -3.35500658e-01
5.19577146e-01 -1.27596017e-02 -6.51877046e-01 -1.81307673e-01
-9.15112674e-01 9.10337210e-01 7.17143536e-01 9.78604436e-01
1.45047382e-01 -6.21203244e-01 -6.30889475e-01 3.68860245e-01
4.50263202e-01 1.75733611e-01 -1.29552215e-01 -8.39855015e-01
7.18364239e-01 1.33342385e-01 3.62907737e-01 -4.67937231e-01
-8.02787781e-01 -3.46225053e-01 -1.37043357e-01 4.12336022e-01
1.02928460e+00 -1.13246754e-01 -1.08804536e+00 1.35259342e+00
7.73596704e-01 -3.48705798e-01 -7.14611948e-01 1.19033182e+00
4.92699921e-01 -1.23881131e-01 5.00717051e-02 -1.49738565e-01
1.58138371e+00 -6.18312657e-01 -6.43149137e-01 -2.84944456e-02
3.48042458e-01 -8.78763616e-01 9.23841059e-01 6.16537869e-01
-1.30768895e+00 -4.67992872e-01 -5.91160715e-01 -8.50499719e-02
-7.77797103e-02 3.01705122e-01 3.54696453e-01 4.90573645e-01
-7.90814102e-01 7.10936010e-01 -1.06096697e+00 -3.29260737e-01
3.02868307e-01 4.02216673e-01 -5.50880492e-01 1.52900800e-01
-5.63373804e-01 1.16158259e+00 1.58700719e-01 3.96963537e-01
-7.11933315e-01 -7.12310612e-01 -4.39088285e-01 -2.59984344e-01
5.05842030e-01 -9.24743474e-01 1.21270716e+00 -1.77696273e-01
-1.32325900e+00 1.07644057e+00 -4.59668726e-01 2.29682013e-01
1.21747136e+00 -5.77937305e-01 3.49266261e-01 2.38075092e-01
-1.16829164e-01 4.29798305e-01 1.05189395e+00 -1.23854446e+00
-2.53883779e-01 -1.12922549e+00 -1.90580457e-01 2.67988145e-01
2.55588025e-01 1.45224407e-01 -7.16561198e-01 -6.34081244e-01
6.05632305e-01 -1.26654315e+00 -2.01408654e-01 -2.42113397e-02
-4.92426246e-01 -4.40007210e-01 6.53112829e-01 -1.11382520e+00
9.10401046e-01 -1.88100934e+00 7.30608404e-01 4.47064072e-01
3.00718069e-01 -1.54782057e-01 8.63423164e-04 2.52109617e-01
-6.52673841e-02 -4.67854530e-01 1.05296165e-01 -8.05498600e-01
2.84674317e-02 -9.72217470e-02 8.21355358e-02 8.85851502e-01
-2.65426815e-01 8.00674021e-01 -5.12906790e-01 -6.32055819e-01
2.86032587e-01 9.33261514e-01 -6.04961097e-01 2.52673894e-01
-2.35941932e-01 8.12125802e-01 -4.73458290e-01 5.51530600e-01
6.33860767e-01 3.33607703e-01 1.13584355e-01 -4.86987442e-01
-2.25364193e-01 1.20551445e-01 -1.51377571e+00 1.99263370e+00
-2.29002118e-01 2.34505758e-01 2.92424202e-01 -1.47017434e-01
1.14192009e-01 5.89114964e-01 5.59917152e-01 -6.64997175e-02
5.53894997e-01 2.61204094e-02 -2.94954097e-03 -6.61804259e-01
6.06069565e-02 -7.29155838e-02 5.18251777e-01 4.90818769e-01
1.02245286e-01 -1.37338415e-01 -2.21699998e-02 -9.25548300e-02
8.02737594e-01 6.69840336e-01 1.78445905e-01 -1.98829964e-01
2.87703693e-01 -3.26104820e-01 -3.54401350e-01 4.85444278e-01
-6.33720607e-02 9.98753726e-01 3.78118634e-01 -2.11834162e-01
-1.04928255e+00 -1.06811094e+00 -3.10436130e-01 1.01867414e+00
-4.51117486e-01 -2.93994337e-01 -9.38170791e-01 -5.75567603e-01
2.35408694e-01 5.11061013e-01 -8.76557052e-01 4.91640270e-01
-6.81824744e-01 -3.25146675e-01 4.50724624e-02 9.20200348e-01
-1.21980526e-01 -1.20830750e+00 -8.88284087e-01 6.19741306e-02
-1.04142562e-01 -1.11225307e+00 -4.07991409e-01 -9.22581274e-03
-8.85298669e-01 -1.04562414e+00 -1.22966731e+00 -5.46618998e-01
8.02346289e-01 -9.75955352e-02 8.06891739e-01 -3.86653721e-01
-7.90487528e-01 7.34225094e-01 -2.07432836e-01 -3.33284736e-01
2.19045356e-01 1.28065750e-01 1.88135475e-01 -2.39057645e-01
-8.18990692e-02 -6.74098074e-01 -7.33742416e-01 3.73220950e-01
-2.86164373e-01 -2.92419612e-01 3.77933592e-01 4.11423117e-01
5.56332767e-01 -2.49126345e-01 -1.55702442e-01 -4.75415349e-01
5.19501328e-01 1.47528380e-01 -3.12868685e-01 1.66136965e-01
1.05322860e-01 1.28878197e-02 -2.25702539e-01 -6.70283675e-01
-1.14591980e+00 5.57378054e-01 -1.14751585e-01 -6.13524973e-01
-2.69386470e-01 8.10409859e-02 -2.31791496e-01 -1.67252511e-01
7.29558051e-01 -2.73246884e-01 -4.43586744e-02 -8.63653004e-01
4.58751440e-01 4.65396881e-01 3.22077483e-01 -7.72189081e-01
3.26745450e-01 6.72828317e-01 3.37837487e-01 -8.02054107e-01
-4.67784166e-01 -5.23737550e-01 -1.64191353e+00 -4.36031014e-01
9.10812736e-01 -4.51647311e-01 -1.03552556e+00 4.41953093e-01
-1.58897316e+00 -3.71312588e-01 -6.03081286e-02 6.36032641e-01
-8.06473553e-01 4.05206650e-01 -3.73016685e-01 -1.13349271e+00
-1.91243857e-01 -1.17598152e+00 1.51002383e+00 -4.03224379e-01
-6.62952483e-01 -6.52138174e-01 6.63801506e-02 3.69867355e-01
4.16044705e-03 2.81925082e-01 7.14706898e-01 -3.12341958e-01
-3.80902499e-01 -5.39204299e-01 -4.65266295e-02 -1.01389937e-01
-1.06290601e-01 -2.17207745e-01 -1.36712933e+00 -5.13075352e-01
-6.81627393e-02 6.72947392e-02 5.78995824e-01 8.14510167e-01
1.10484111e+00 7.01047853e-02 -4.45139021e-01 1.12910688e-01
9.77953196e-01 -3.24390590e-01 5.61216950e-01 9.70921442e-02
8.11963975e-01 1.22273755e+00 5.97945213e-01 8.47693920e-01
8.85900483e-02 1.13838732e+00 3.75492662e-01 3.84184629e-01
-2.73422033e-01 1.61179364e-01 -2.29700074e-01 -4.79413681e-02
-9.92982626e-01 9.65108350e-02 -1.05373645e+00 4.25559133e-01
-1.57077074e+00 -6.10195756e-01 -3.15830499e-01 2.28809261e+00
3.57871771e-01 1.19178645e-01 6.53130054e-01 3.50898921e-01
4.00953084e-01 -1.77880377e-01 -3.89508635e-01 1.63876757e-01
5.42640269e-01 3.96994382e-01 3.52581471e-01 8.36569428e-01
-9.24989104e-01 9.23512816e-01 6.89056015e+00 3.92542183e-01
-6.79674327e-01 2.43858755e-01 -1.88590690e-01 -4.90230113e-01
2.32229277e-01 -2.36539930e-01 -8.94055068e-01 -1.36459127e-01
1.65087208e-01 6.19567275e-01 4.77705181e-01 6.13980055e-01
1.13289624e-01 -4.71323013e-01 -1.37288678e+00 1.00975156e+00
1.79604530e-01 -4.21783119e-01 -2.50978500e-01 1.45293817e-01
3.51295501e-01 -2.59334743e-01 -1.85215529e-02 -2.41496071e-01
-5.25526464e-01 -9.11379099e-01 9.48015988e-01 7.14833438e-01
6.70043170e-01 -4.94761258e-01 4.60175216e-01 3.43180627e-01
-1.19450045e+00 9.68458429e-02 1.97327975e-02 -1.98243141e-01
3.40736181e-01 2.27391347e-01 -9.89637017e-01 2.47137800e-01
7.31336772e-01 3.84436548e-02 -5.26136935e-01 8.75903010e-01
-1.92815870e-01 -3.50045711e-02 -5.55117249e-01 1.52682364e-01
-9.54187363e-02 2.55264908e-01 9.31644440e-01 1.01331198e+00
-8.68356153e-02 2.23766387e-01 -1.14980415e-01 7.16295481e-01
6.35301948e-01 6.39353245e-02 -2.87638724e-01 4.73139048e-01
-9.30817947e-02 1.16700518e+00 -9.00382400e-01 -8.23919922e-02
-5.07956035e-02 1.30455339e+00 2.70120054e-01 5.58993936e-01
-5.35201371e-01 -2.30440974e-01 4.16141540e-01 7.08609462e-01
5.07404089e-01 -6.77027404e-01 -3.15165490e-01 -8.21237803e-01
7.16245234e-01 -5.50983369e-01 2.49102861e-01 -9.22332764e-01
-9.07622039e-01 5.06296754e-01 5.34616709e-01 -7.29288876e-01
-4.49422926e-01 -9.94519413e-01 -3.08198273e-01 1.28991616e+00
-7.00397015e-01 -1.26624119e+00 -5.49297273e-01 6.99001253e-01
4.82929051e-01 1.41999885e-01 8.19740772e-01 -4.30623144e-02
-1.08722083e-01 3.53151143e-01 -3.84870112e-01 -2.37096995e-01
6.96938574e-01 -1.08424914e+00 3.74615788e-01 3.67918313e-01
8.17170441e-02 7.44037390e-01 9.01353657e-01 -8.11258495e-01
-1.37764585e+00 -2.83211201e-01 8.43293250e-01 -1.25486755e+00
2.29011372e-01 -6.05176449e-01 -5.13612986e-01 9.10602748e-01
-2.53997803e-01 5.78139201e-02 2.63011187e-01 3.92732948e-01
-1.49940610e-01 4.67362583e-01 -1.31343436e+00 5.47676265e-01
1.38595879e+00 -4.96338964e-01 -6.70266211e-01 4.61010218e-01
4.00359966e-02 -7.74106145e-01 -6.71819627e-01 2.91843295e-01
1.24549186e+00 -8.11276734e-01 1.26852322e+00 -8.76929820e-01
-7.07781762e-02 -1.66967690e-01 -1.13961466e-01 -9.08380985e-01
-3.79833788e-01 -3.66522759e-01 -2.79907614e-01 7.70512760e-01
-2.98254602e-02 -2.36600250e-01 1.08385098e+00 7.57332981e-01
2.37919226e-01 -6.91100597e-01 -1.39593303e+00 -5.37063181e-01
1.91247575e-02 -4.90454555e-01 -2.31869575e-02 1.51228219e-01
7.37832859e-02 4.87273410e-02 -1.47815747e-02 2.14419514e-01
8.45050871e-01 -1.91326186e-01 8.00976634e-01 -1.27426970e+00
-4.03992623e-01 -3.77327293e-01 -4.82507706e-01 -8.79262865e-01
2.45727256e-01 -4.80249226e-01 7.10691884e-02 -1.51990640e+00
2.42473379e-01 3.16534787e-02 9.86116752e-02 3.02760452e-01
-3.11663933e-02 3.02006692e-01 1.05824612e-01 9.48624313e-02
-1.61561221e-01 3.72201577e-02 1.56988430e+00 1.91304535e-01
-3.82879078e-01 2.50834644e-01 1.44200660e-02 6.83802009e-01
4.52808201e-01 -3.66408885e-01 -1.39649987e-01 -2.83836603e-01
-8.03566948e-02 5.69436848e-02 7.04960465e-01 -7.58616924e-01
3.95452194e-02 1.73872992e-01 7.72306383e-01 -1.09970403e+00
6.41238511e-01 -9.61579680e-01 2.10955814e-01 5.66204906e-01
-1.23556502e-01 -7.65125006e-02 3.67655247e-01 2.23500624e-01
5.57875514e-01 -3.32510859e-01 5.37951052e-01 -2.29318932e-01
-2.33893201e-01 2.03889638e-01 -2.01571271e-01 -1.86514765e-01
1.03254604e+00 -3.03022832e-01 4.07567948e-01 -2.13007808e-01
-1.58217072e+00 -1.38103366e-01 3.10494721e-01 2.58131444e-01
4.97855902e-01 -9.75912333e-01 -8.06714237e-01 3.24162126e-01
-4.95242029e-02 4.57025990e-02 3.65361601e-01 1.26148653e+00
-5.07745564e-01 7.05515385e-01 -2.55586237e-01 -7.32003570e-01
-1.77034509e+00 5.77638268e-01 3.40207487e-01 -3.15413848e-02
-6.84005380e-01 9.63846982e-01 2.24741492e-02 -1.75031990e-01
7.67786086e-01 -3.59131843e-01 -9.07398686e-02 3.56100649e-01
5.37278235e-01 9.01765883e-01 3.76429796e-01 -6.63428783e-01
-6.41710937e-01 9.93845344e-01 4.16475348e-02 -6.51660979e-01
1.26379681e+00 4.43296582e-02 -7.36179948e-02 2.65659809e-01
9.86579239e-01 1.42814130e-01 -1.41619837e+00 -2.66076904e-02
-2.90309221e-01 -7.23204315e-01 4.19561155e-02 -9.52978730e-01
-7.68017828e-01 1.01841104e+00 8.37421000e-01 -3.99526924e-01
5.75548530e-01 4.68131095e-01 1.76602468e-01 3.19484413e-01
6.04675412e-01 -9.99147236e-01 2.19386201e-02 2.80606568e-01
1.65192544e+00 -8.37701738e-01 3.47882152e-01 -7.45862722e-01
-4.26243544e-01 1.13177609e+00 4.98058021e-01 -1.08202860e-01
7.97644973e-01 2.82045543e-01 -1.32944599e-01 -3.58609080e-01
-1.02935597e-01 -2.87507981e-01 7.22794890e-01 6.50119364e-01
5.96059024e-01 -1.33168101e-02 -4.10512000e-01 4.55973655e-01
-3.80183049e-02 -1.58961892e-01 -2.27693021e-01 1.28474140e+00
-1.68370828e-02 -1.05427098e+00 -7.14717209e-01 2.35388801e-01
-2.95480013e-01 1.79545566e-01 -5.34249842e-01 8.62408817e-01
7.91476294e-02 4.57728177e-01 2.31728349e-02 -2.22911298e-01
9.79032278e-01 4.52854484e-01 1.68656158e+00 -7.24186361e-01
-5.08985460e-01 4.83461171e-01 -8.42354372e-02 -7.14492559e-01
-2.96191096e-01 -1.20065832e+00 -1.09622777e+00 -9.28697884e-02
-3.13312232e-01 -4.29000854e-01 8.85055542e-01 1.05462945e+00
1.75608918e-01 5.05703032e-01 -1.77757512e-03 -1.98632646e+00
-6.80766225e-01 -1.27839029e+00 -7.82567382e-01 2.45933309e-01
2.06309035e-01 -1.24356031e+00 -2.11899266e-01 -1.70883611e-01] | [6.555068016052246, -0.7934607267379761] |
f9e51745-974e-4fc3-b6a8-46f6304c7cba | roberta-a-robustly-optimized-bert-pretraining | 1907.11692 | null | https://arxiv.org/abs/1907.11692v1 | https://arxiv.org/pdf/1907.11692v1.pdf | RoBERTa: A Robustly Optimized BERT Pretraining Approach | Language model pretraining has led to significant performance gains but careful comparison between different approaches is challenging. Training is computationally expensive, often done on private datasets of different sizes, and, as we will show, hyperparameter choices have significant impact on the final results. We present a replication study of BERT pretraining (Devlin et al., 2019) that carefully measures the impact of many key hyperparameters and training data size. We find that BERT was significantly undertrained, and can match or exceed the performance of every model published after it. Our best model achieves state-of-the-art results on GLUE, RACE and SQuAD. These results highlight the importance of previously overlooked design choices, and raise questions about the source of recently reported improvements. We release our models and code. | ['Myle Ott', 'Mike Lewis', 'Luke Zettlemoyer', 'Mandar Joshi', 'Danqi Chen', 'Yinhan Liu', 'Omer Levy', 'Veselin Stoyanov', 'Naman Goyal', 'Jingfei Du'] | 2019-07-26 | null | https://openreview.net/forum?id=SyxS0T4tvS | https://openreview.net/pdf?id=SyxS0T4tvS | null | ['type-prediction', 'document-image-classification', 'multi-task-language-understanding', 'lexical-simplification', 'riddle-sense', 'linguistic-acceptability'] | ['computer-code', 'computer-vision', 'methodology', 'natural-language-processing', 'natural-language-processing', 'natural-language-processing'] | [-2.58225113e-01 -1.10528871e-01 -4.61472213e-01 -3.21743190e-01
-1.03267276e+00 -8.09951127e-01 5.23106098e-01 -3.06222495e-02
-7.31344700e-01 8.07517231e-01 2.88638026e-01 -8.09655786e-01
-7.49247223e-02 -3.05749804e-01 -1.00077283e+00 -3.47730219e-01
-2.08568141e-01 6.59354508e-01 2.94545978e-01 -2.44537875e-01
1.43584803e-01 1.86086893e-01 -1.16334200e+00 4.76631254e-01
4.88182068e-01 5.02080679e-01 -1.59732565e-01 8.17437053e-01
1.70100123e-01 8.37963462e-01 -8.09618652e-01 -7.30936885e-01
2.00641915e-01 3.85373719e-02 -9.21294808e-01 -3.61759722e-01
4.76343393e-01 -2.82215714e-01 -1.97752744e-01 5.60310483e-01
7.24710107e-01 -4.81618941e-03 4.29218113e-01 -1.22276318e+00
-5.41931927e-01 1.14946365e+00 -4.82528448e-01 4.45129782e-01
-2.11324334e-01 2.95673281e-01 1.07648087e+00 -5.29640377e-01
4.03862238e-01 8.89751256e-01 9.68296826e-01 5.13758004e-01
-1.50431979e+00 -8.57042670e-01 1.97725058e-01 -1.88366443e-01
-1.50752044e+00 -7.78251529e-01 1.52918786e-01 -3.57021570e-01
1.20328796e+00 5.48986681e-02 5.10944486e-01 1.27229631e+00
1.41015649e-03 7.19272673e-01 1.04340243e+00 -6.01722121e-01
1.03539024e-02 1.83386907e-01 1.10117793e-01 5.90754151e-01
4.94827271e-01 8.29624981e-02 -4.93374288e-01 -3.99006456e-01
5.90913892e-01 -3.73759836e-01 -1.26514763e-01 -2.74897724e-01
-1.17528951e+00 7.47378230e-01 2.46413320e-01 4.42081273e-01
1.67564198e-01 5.19314110e-01 4.91454661e-01 3.24756324e-01
2.49491528e-01 6.95843816e-01 -9.80285943e-01 -6.75097823e-01
-1.03983641e+00 3.61833274e-01 1.00423849e+00 8.84284139e-01
5.55402398e-01 -1.49706930e-01 4.50935140e-02 7.41367698e-01
-3.74553204e-02 1.93549797e-01 6.07228994e-01 -1.02909589e+00
6.72296226e-01 1.80337042e-01 1.15105294e-01 -5.39173961e-01
-4.61971670e-01 -5.98368287e-01 -4.26611602e-01 -9.22034774e-03
7.19577551e-01 -6.31908774e-01 -8.20382893e-01 1.75499153e+00
-8.99800584e-02 3.88497449e-02 -8.45134780e-02 6.07556224e-01
7.80894637e-01 2.63558596e-01 2.72630811e-01 2.36466169e-01
1.02636361e+00 -1.07842755e+00 -1.75298482e-01 -4.72476006e-01
1.15960252e+00 -8.49834442e-01 1.40823746e+00 3.94580692e-01
-1.25189555e+00 -2.03203171e-01 -1.10169375e+00 -1.33016095e-01
-3.11112106e-01 1.96205422e-01 1.00167346e+00 1.06071448e+00
-1.16787434e+00 6.69048667e-01 -8.49121451e-01 -3.44319105e-01
1.78090930e-01 6.71006143e-01 -1.95196673e-01 2.98750788e-01
-1.05354333e+00 9.22813594e-01 4.62563515e-01 -1.39168888e-01
-5.35346627e-01 -7.28916049e-01 -4.00131047e-01 9.52155441e-02
5.67790687e-01 -6.39285207e-01 1.74388599e+00 -7.09411323e-01
-1.52672076e+00 7.98227727e-01 1.31048352e-01 -6.66020989e-01
4.26013470e-01 -4.41341579e-01 -3.77332985e-01 -3.78673315e-01
-3.67638677e-01 6.54740810e-01 3.54212582e-01 -9.74295795e-01
-3.84833306e-01 7.14869648e-02 1.65650040e-01 7.21602589e-02
-3.80643904e-01 2.69243658e-01 -6.90569937e-01 -6.12453103e-01
-2.13355750e-01 -9.93512332e-01 -2.01066762e-01 -5.92544496e-01
-5.39915673e-02 2.50497125e-02 1.58089846e-01 -4.43070501e-01
1.52714491e+00 -2.22475195e+00 -2.10943744e-01 1.23204783e-01
1.11729957e-01 2.62017161e-01 -2.86892921e-01 3.97056550e-01
-6.77372292e-02 7.29809940e-01 1.01628222e-01 -3.73155057e-01
7.53368810e-02 1.33923456e-01 -1.77946895e-01 3.78628552e-01
-7.18253404e-02 8.64797711e-01 -4.65218931e-01 -4.78891224e-01
-2.31429666e-01 3.04595083e-01 -1.00416923e+00 -1.15272857e-01
-2.75940746e-01 1.80480741e-02 -1.80748403e-01 3.78075570e-01
2.58365452e-01 -4.98775542e-01 2.10825160e-01 3.01344395e-02
-9.35106128e-02 7.04761505e-01 -9.70466971e-01 1.56284285e+00
-4.03418005e-01 5.31277537e-01 6.82474673e-03 -5.15493572e-01
5.30967414e-01 2.48015881e-01 8.61782208e-02 -4.23922062e-01
2.33106524e-01 1.55997843e-01 4.24383700e-01 -2.35060364e-01
7.19336629e-01 -2.41463836e-02 -6.00951836e-02 6.16744637e-01
-5.21452837e-02 -1.67434484e-01 2.98101664e-01 2.19009683e-01
1.00164747e+00 5.63833602e-02 2.23360141e-03 -3.76517117e-01
-1.67095184e-01 5.96190877e-02 4.95904356e-01 9.93284106e-01
-6.43941462e-02 7.13483691e-01 7.61266053e-01 -3.70036751e-01
-1.14466786e+00 -6.93408370e-01 -1.78690106e-01 1.55057311e+00
-2.78961271e-01 -8.04571986e-01 -7.99211204e-01 -6.52360737e-01
6.71106353e-02 7.05205679e-01 -7.07592309e-01 -1.20455146e-01
-7.79071152e-01 -8.89203072e-01 9.80856538e-01 7.64823318e-01
2.59794980e-01 -6.60183907e-01 -5.42740643e-01 6.71207756e-02
9.29813161e-02 -1.17503047e+00 -3.60664696e-01 3.58686835e-01
-1.01733041e+00 -8.15335095e-01 -4.03837293e-01 -5.78820825e-01
4.15994078e-01 -1.51972920e-02 1.54652441e+00 4.07380879e-01
-7.46775120e-02 -1.15046483e-02 -2.30670974e-01 -4.27924007e-01
-5.09201407e-01 7.54783750e-01 -5.02324924e-02 -7.02498794e-01
2.77812183e-01 -4.66607034e-01 -3.73901218e-01 3.09868097e-01
-6.24672830e-01 2.00751498e-01 5.78503132e-01 9.56979930e-01
4.03687134e-02 -7.89523721e-02 2.76541799e-01 -1.31520045e+00
6.44925416e-01 -3.15338582e-01 -6.36008859e-01 3.15014452e-01
-8.05421472e-01 3.23839515e-01 4.27460670e-01 -4.50475305e-01
-7.61582017e-01 -4.62019444e-01 -8.31133276e-02 -8.23071599e-02
4.62181568e-02 5.16368747e-01 2.72602111e-01 -9.08923298e-02
8.52313638e-01 -4.06119436e-01 -1.10720977e-01 -7.52996266e-01
3.47497672e-01 4.89275068e-01 2.53412724e-01 -1.06333756e+00
4.13807243e-01 1.26995910e-02 -3.58953565e-01 -3.17181349e-01
-6.61443055e-01 -1.20552704e-01 -4.72813189e-01 4.42335546e-01
3.22683305e-01 -1.14574695e+00 -5.99580050e-01 1.55531898e-01
-8.00376892e-01 -9.41475749e-01 -1.59100324e-01 5.17633080e-01
-3.01534623e-01 -3.12940404e-02 -8.20751786e-01 -3.42863858e-01
-3.34093750e-01 -1.33144355e+00 7.42050469e-01 5.95538989e-02
-4.51446921e-01 -1.03369200e+00 1.51758447e-01 4.48157787e-01
5.38950145e-01 -1.97902113e-01 9.81402576e-01 -7.68254042e-01
-4.15698230e-01 -4.13840637e-02 -2.30398253e-01 2.52165020e-01
-3.48245502e-01 3.54207456e-01 -8.99969101e-01 -4.68175948e-01
-4.39633667e-01 -5.72434783e-01 7.81262040e-01 3.03531498e-01
1.10118926e+00 -2.48611718e-01 -2.72721440e-01 6.92436576e-01
1.26757216e+00 -1.97178960e-01 3.78197551e-01 6.71787918e-01
4.87949550e-01 1.72553986e-01 2.24305272e-01 3.13585699e-01
3.06570411e-01 5.50465286e-01 -1.34939507e-01 -6.52684420e-02
-1.64036937e-02 -1.96472675e-01 4.64064062e-01 7.68997431e-01
-9.33886617e-02 -3.00877035e-01 -1.28238809e+00 3.93299162e-01
-1.64292884e+00 -6.10578477e-01 1.11286901e-01 2.21684909e+00
1.02008486e+00 5.81949353e-01 3.57405007e-01 -1.12106644e-01
3.30070883e-01 3.70706990e-02 -2.55586296e-01 -5.81354260e-01
-1.52174070e-01 3.78420830e-01 9.23640251e-01 5.85727096e-01
-7.79327989e-01 1.21305013e+00 8.52547932e+00 7.41553307e-01
-1.10359490e+00 2.18548372e-01 8.56455803e-01 -6.62426770e-01
-2.16090128e-01 2.55180448e-01 -9.64375317e-01 2.86500007e-01
1.35659420e+00 -3.36348861e-01 6.06616259e-01 6.96214795e-01
-1.12306155e-01 -3.03048790e-01 -1.20265639e+00 5.85388660e-01
5.44880927e-02 -1.31941748e+00 -2.73588270e-01 4.21438627e-02
7.91128039e-01 6.19511247e-01 2.48321459e-01 6.25683427e-01
8.09024036e-01 -1.25067496e+00 7.39765525e-01 7.88018107e-02
6.91272557e-01 -8.13561797e-01 6.69115663e-01 2.76031762e-01
-5.95794082e-01 -2.06453770e-01 -2.29474276e-01 -3.54832441e-01
-1.58429220e-01 3.53304386e-01 -9.18859482e-01 6.22494929e-02
7.48089731e-01 2.37917617e-01 -9.92866158e-01 9.75649714e-01
-1.23464562e-01 1.21879423e+00 -5.82780719e-01 -1.47537962e-01
2.26256996e-01 2.54206151e-01 7.31547400e-02 1.25494289e+00
-6.56351298e-02 -2.83427779e-02 5.01254294e-03 4.97161478e-01
-2.78229624e-01 1.42237738e-01 -3.29839349e-01 -3.45364571e-01
4.73694026e-01 8.93993497e-01 -6.24334514e-01 -3.58377457e-01
-4.62779760e-01 2.85232574e-01 6.53174579e-01 4.60559845e-01
-8.62129390e-01 -2.10699346e-02 5.21651924e-01 2.14573622e-01
3.31252337e-01 -3.93309921e-01 -5.49796343e-01 -1.15008104e+00
-1.46735430e-01 -1.13647103e+00 5.81557035e-01 -5.25484204e-01
-9.15418625e-01 3.77512783e-01 3.16907376e-01 -6.03827655e-01
-3.46702069e-01 -6.15630150e-01 -4.42400634e-01 7.10648417e-01
-1.01696479e+00 -8.60272765e-01 1.96913391e-01 4.77128401e-02
2.59193361e-01 -1.50908798e-01 7.14060426e-01 3.52351815e-01
-7.75004268e-01 1.14277768e+00 3.44284862e-01 1.24254465e-01
9.36077893e-01 -9.72043633e-01 8.65494549e-01 5.77084184e-01
2.43387967e-01 9.25479114e-01 6.03130817e-01 -4.24929619e-01
-1.22228599e+00 -4.66303647e-01 8.78237903e-01 -8.04281831e-01
8.29182148e-01 -5.41172862e-01 -6.69662654e-01 1.10424197e+00
3.78872484e-01 -3.37921858e-01 8.30896735e-01 8.48091125e-01
-4.21943098e-01 1.48138907e-02 -7.42715061e-01 8.07432771e-01
1.08288264e+00 -3.42305958e-01 -2.33617902e-01 2.20849827e-01
8.15814376e-01 -6.58529401e-01 -8.18715870e-01 3.79959077e-01
6.56108737e-01 -1.17238045e+00 8.61487567e-01 -8.98788571e-01
3.23472559e-01 3.29923928e-01 6.89054513e-03 -1.09272861e+00
-3.47606897e-01 -8.03949237e-01 -1.18885953e-02 1.17967010e+00
8.95937443e-01 -5.71668983e-01 9.40572262e-01 1.04896474e+00
1.77399606e-01 -8.90525997e-01 -5.82455873e-01 -7.57701755e-01
6.22015476e-01 -4.90884006e-01 9.19426441e-01 9.23815429e-01
2.83368714e-02 4.64862972e-01 -3.09153974e-01 -2.48498961e-01
1.60550714e-01 -6.81130141e-02 1.00076139e+00 -8.56438279e-01
-7.35256433e-01 -6.39369845e-01 4.92799729e-02 -9.99397635e-01
9.34554636e-02 -7.97463596e-01 -1.94386333e-01 -1.04753435e+00
2.99023151e-01 -8.32727551e-01 -2.30587989e-01 7.10305095e-01
-1.94783673e-01 2.85885572e-01 3.12538296e-01 1.33521289e-01
-5.32127857e-01 1.04655646e-01 8.95122647e-01 2.32281163e-01
-2.44342491e-01 -1.33285329e-01 -9.54348385e-01 6.77556515e-01
9.91562247e-01 -4.16308701e-01 -3.89186889e-01 -9.11219776e-01
5.97759604e-01 -3.05849195e-01 6.40875101e-02 -1.02163696e+00
2.83738911e-01 -2.13880017e-01 2.85013437e-01 -2.39830539e-01
1.33689731e-01 -5.08140564e-01 7.87994787e-02 2.63976276e-01
-5.02879977e-01 6.63435638e-01 6.62244141e-01 7.73700625e-02
1.53078154e-01 -3.97984743e-01 5.81525266e-01 -1.18568063e-01
-3.78213108e-01 -7.08570629e-02 -2.80438483e-01 6.00735366e-01
5.97050846e-01 -1.90183334e-02 -3.71821761e-01 -3.01485687e-01
-4.41743046e-01 1.91851303e-01 6.41054332e-01 2.67714411e-01
-1.88826650e-01 -8.93624485e-01 -5.45967579e-01 1.55397817e-01
-1.91358536e-01 1.40616652e-02 -8.10029060e-02 7.72273421e-01
-6.72155976e-01 3.06246996e-01 2.24842697e-01 -3.41741949e-01
-1.27684522e+00 3.44340175e-01 2.80349731e-01 -4.23241436e-01
-4.82505172e-01 9.93725955e-01 -1.84662357e-01 -4.43974018e-01
4.56537813e-01 -3.94677281e-01 3.12370807e-01 -1.68056011e-01
3.29050124e-01 2.44746774e-01 2.39469677e-01 -2.51753271e-01
-1.83748633e-01 2.55839646e-01 -4.81276125e-01 -3.74217093e-01
1.33409405e+00 3.12688828e-01 8.65351483e-02 4.66839403e-01
1.08857179e+00 1.10982195e-01 -1.12149954e+00 -3.16472113e-01
-6.76838011e-02 -3.32527548e-01 1.42504901e-01 -9.72298443e-01
-9.10050988e-01 8.53238463e-01 3.25689137e-01 -2.05490276e-01
8.20924461e-01 -4.68235500e-02 6.60511911e-01 5.85477471e-01
3.81039977e-01 -1.19938576e+00 -1.49372384e-01 6.77460194e-01
5.35397232e-01 -1.17408562e+00 3.85849565e-01 -1.36390358e-01
-5.99974155e-01 8.08490098e-01 7.61560261e-01 1.88878268e-01
5.91400921e-01 7.31332481e-01 1.13388941e-01 1.06773160e-01
-1.12388396e+00 1.77237958e-01 -5.47596812e-02 1.72442243e-01
7.71761537e-01 8.84049088e-02 -2.73528963e-01 6.32606387e-01
-7.27779806e-01 2.27733657e-01 2.77900308e-01 1.21187425e+00
-1.43220156e-01 -1.38557267e+00 -3.23385656e-01 3.96762133e-01
-6.51573241e-01 -4.85344112e-01 -3.14463168e-01 1.08910155e+00
-4.49823700e-02 5.99956989e-01 1.30356044e-01 -5.11171162e-01
2.80580461e-01 3.50530297e-01 5.96702576e-01 -5.91568649e-01
-1.12024045e+00 -3.83483358e-02 4.49655235e-01 -2.91961193e-01
-2.65428089e-02 -6.60812914e-01 -1.12182140e+00 -8.14590633e-01
-4.14534658e-01 2.88601100e-01 6.06331706e-01 7.90556133e-01
6.57217979e-01 1.92039937e-01 8.77609700e-02 -5.74396789e-01
-7.94466376e-01 -9.20809090e-01 -2.39657044e-01 2.32748196e-01
1.40099093e-01 -4.55734253e-01 -4.08740044e-01 1.39974924e-02] | [10.643017768859863, 8.235651969909668] |
087dd54a-e04d-4b87-ae36-b39350f79330 | mice-mixture-of-contrastive-experts-for-1 | 2105.01899 | null | https://arxiv.org/abs/2105.01899v1 | https://arxiv.org/pdf/2105.01899v1.pdf | MiCE: Mixture of Contrastive Experts for Unsupervised Image Clustering | We present Mixture of Contrastive Experts (MiCE), a unified probabilistic clustering framework that simultaneously exploits the discriminative representations learned by contrastive learning and the semantic structures captured by a latent mixture model. Motivated by the mixture of experts, MiCE employs a gating function to partition an unlabeled dataset into subsets according to the latent semantics and multiple experts to discriminate distinct subsets of instances assigned to them in a contrastive learning manner. To solve the nontrivial inference and learning problems caused by the latent variables, we further develop a scalable variant of the Expectation-Maximization (EM) algorithm for MiCE and provide proof of the convergence. Empirically, we evaluate the clustering performance of MiCE on four widely adopted natural image datasets. MiCE achieves significantly better results than various previous methods and a strong contrastive learning baseline. | ['Jun Zhu', 'Chongxuan Li', 'Tsung Wei Tsai'] | 2021-05-05 | mice-mixture-of-contrastive-experts-for | https://openreview.net/forum?id=gV3wdEOGy_V | https://openreview.net/pdf?id=gV3wdEOGy_V | iclr-2021-1 | ['image-clustering'] | ['computer-vision'] | [-6.41524121e-02 -1.68338284e-01 -3.18924218e-01 -6.21992528e-01
-9.17305052e-01 -6.65440798e-01 6.95411861e-01 -2.93783605e-01
-4.80779976e-01 4.35905099e-01 2.49499515e-01 1.31827772e-01
-2.67829597e-01 -4.06864166e-01 -3.82789999e-01 -1.02131140e+00
-3.99691500e-02 9.81109321e-01 2.17505395e-01 4.55202758e-01
1.77490085e-01 2.70459980e-01 -1.53850842e+00 5.40936291e-01
9.88373399e-01 5.55871487e-01 4.66516465e-01 3.39920521e-01
-1.37118533e-01 7.80491412e-01 -2.63595760e-01 -9.45241898e-02
2.05417760e-02 -4.34359014e-01 -8.39944661e-01 6.34384155e-01
2.67361104e-01 3.47782336e-02 -4.83749807e-02 1.21738541e+00
1.28304079e-01 4.44191128e-01 1.16826499e+00 -1.39276350e+00
-5.82182765e-01 6.62728369e-01 -8.86144638e-01 1.98291540e-01
-1.95025861e-01 -7.47123435e-02 1.26132345e+00 -1.12650287e+00
5.77163279e-01 1.43821788e+00 3.44643086e-01 6.28057897e-01
-1.48726749e+00 -6.77899957e-01 6.62171841e-01 7.58102015e-02
-1.47649550e+00 -3.83570611e-01 8.65741670e-01 -4.74763453e-01
5.07980704e-01 7.41712824e-02 2.98715472e-01 1.13713694e+00
-3.71814758e-01 1.18358135e+00 1.28407383e+00 -2.36161307e-01
6.31497741e-01 1.21856391e-01 2.59295315e-01 8.83096218e-01
-1.24337608e-02 -1.97747648e-01 -4.52540547e-01 -3.74774426e-01
6.59276545e-01 1.59912422e-01 -4.27654088e-02 -7.05130279e-01
-7.82102346e-01 1.08566916e+00 2.13912144e-01 1.09037042e-01
-3.56025189e-01 1.02937251e-01 -6.17117248e-02 -1.00242354e-01
6.18762016e-01 2.66538113e-01 -5.91593385e-01 5.41286647e-01
-1.35911846e+00 -3.65086980e-02 6.25361562e-01 8.29473972e-01
8.89583945e-01 -6.49384037e-02 -2.95081794e-01 7.97386527e-01
9.47892487e-01 4.68154013e-01 3.72820169e-01 -1.47015965e+00
-8.40062741e-03 4.35535073e-01 -5.95922545e-02 -5.54635763e-01
-1.38675094e-01 -8.08233619e-01 -8.05126071e-01 2.21934602e-01
2.36266822e-01 8.41139182e-02 -1.16742229e+00 2.12860680e+00
1.65185899e-01 5.46706140e-01 -9.22635421e-02 6.25078678e-01
5.27717650e-01 4.71008837e-01 4.87217277e-01 -4.05553401e-01
1.09605837e+00 -1.26665783e+00 -6.14999056e-01 -2.75431782e-01
2.68175334e-01 -4.68750596e-01 9.10582483e-01 4.59410340e-01
-1.05221319e+00 -6.35207415e-01 -8.95809472e-01 3.44494879e-01
-1.09019235e-01 1.23569220e-01 6.66616797e-01 6.48111284e-01
-1.20664477e+00 4.88876969e-01 -1.04824591e+00 1.23476498e-01
8.12058449e-01 3.70802522e-01 -2.31317729e-01 -2.25002095e-01
-6.87643588e-01 4.24300790e-01 4.50832546e-01 -1.52339101e-01
-1.29204512e+00 -3.60355586e-01 -8.81315887e-01 1.53465658e-01
2.59105682e-01 -9.12731349e-01 1.00895393e+00 -1.16747081e+00
-1.43035460e+00 1.23470056e+00 -6.03798449e-01 -1.97663188e-01
3.37126762e-01 -4.42543536e-01 -1.97204888e-01 3.87181252e-01
3.64263713e-01 1.02529502e+00 1.18045914e+00 -1.80125964e+00
-7.17236280e-01 -4.34285671e-01 -3.56085241e-01 3.00831944e-01
-1.41842619e-01 3.26120183e-02 -9.11780775e-01 -6.66361272e-01
3.22581291e-01 -9.29100394e-01 -5.80827653e-01 -3.40771645e-01
-3.02235097e-01 -3.76953274e-01 6.87361240e-01 -4.12586153e-01
9.90606368e-01 -2.35978603e+00 4.26088899e-01 3.33015203e-01
6.01077437e-01 -1.24423489e-01 -2.24692687e-01 -5.68266548e-02
1.33656813e-02 -7.73907825e-02 -6.29032850e-01 -7.57130086e-01
3.01881343e-01 4.28687215e-01 -3.30855310e-01 4.99990106e-01
1.22357950e-01 8.99840832e-01 -1.10388267e+00 -8.09458375e-01
2.48508915e-01 3.34110290e-01 -6.57090008e-01 4.38491434e-01
-2.80627459e-01 6.58022583e-01 -2.77509779e-01 5.38706958e-01
8.61429513e-01 -8.06312025e-01 7.74638355e-01 -1.20588779e-01
3.87999117e-01 9.08393264e-02 -1.22082913e+00 1.81582034e+00
7.19546154e-02 2.56891012e-01 2.40349501e-01 -1.09476399e+00
5.70874631e-01 3.43049735e-01 4.60887402e-01 -2.86878675e-01
-1.17866382e-01 1.31470680e-01 -3.67185473e-01 -2.32721731e-01
1.72447070e-01 -3.50063294e-01 -2.44275197e-01 6.36827409e-01
5.28949320e-01 6.26201043e-04 1.45002119e-02 5.46011269e-01
7.02167869e-01 3.69489670e-01 2.01201871e-01 -6.13774657e-01
2.72252709e-01 -2.08557874e-01 8.22589338e-01 9.90872502e-01
-5.10570347e-01 5.52679241e-01 2.90965643e-02 -1.38099447e-01
-6.07414126e-01 -1.83887744e+00 1.23063743e-01 1.57971549e+00
1.79412052e-01 -3.86984825e-01 -7.39147723e-01 -1.03557420e+00
-3.15863818e-01 6.75043404e-01 -6.90397501e-01 -1.78744018e-01
-2.78889745e-01 -1.11114621e+00 2.83793271e-01 7.71212399e-01
5.51935375e-01 -8.98815930e-01 -3.23397845e-01 -9.31158513e-02
-2.67894119e-01 -1.03759718e+00 -3.78348023e-01 5.12273312e-01
-1.06399834e+00 -9.45980430e-01 -3.64307046e-01 -1.05271053e+00
8.09627235e-01 2.65129030e-01 1.62263238e+00 -2.39254564e-01
1.27549902e-01 4.95186597e-01 -8.02344382e-02 9.22008529e-02
-1.37086079e-01 -1.73877940e-01 -1.98612232e-02 -2.57759318e-02
5.34861982e-01 -7.56626427e-01 -5.78921497e-01 1.87953234e-01
-1.00060844e+00 -3.64415981e-02 6.28728569e-01 7.83569634e-01
7.12722361e-01 3.49139482e-01 3.48865569e-01 -1.28781819e+00
2.05044255e-01 -8.03986013e-01 -3.51389259e-01 2.62382567e-01
-4.66195136e-01 3.26173633e-01 2.08673939e-01 -4.16144729e-01
-1.39691758e+00 2.40944594e-01 1.51718900e-01 -6.81655169e-01
-4.82558817e-01 3.34813714e-01 -4.74484831e-01 5.51521361e-01
3.50232214e-01 1.74609452e-01 -4.17761385e-01 -4.97121900e-01
8.76724958e-01 4.27407444e-01 8.21401894e-01 -8.69975924e-01
6.30218983e-01 9.48821843e-01 -3.79693180e-01 -4.00745600e-01
-1.34826970e+00 -7.16704905e-01 -8.66896868e-01 6.41577085e-03
1.07971597e+00 -1.31105781e+00 -2.37775251e-01 3.48170817e-01
-8.04115415e-01 -3.75833571e-01 -3.14398587e-01 5.41905582e-01
-5.74747026e-01 4.74038810e-01 -8.93341243e-01 -7.26112604e-01
9.53666866e-02 -1.03443670e+00 9.26573336e-01 1.59547642e-01
-3.36510390e-01 -1.32522142e+00 1.36268988e-01 4.88538176e-01
2.93459576e-02 -4.57372665e-02 1.03701842e+00 -7.28411973e-01
-4.67832416e-01 2.88069755e-01 -8.54906738e-02 3.52997065e-01
1.44449607e-01 -6.77673891e-03 -1.30342972e+00 -3.59053314e-01
2.61734158e-01 -3.95539910e-01 1.40135789e+00 6.72504842e-01
1.23292685e+00 1.52046397e-01 -6.31042957e-01 5.41171253e-01
1.41407979e+00 -6.89723119e-02 4.22829300e-01 -1.16534635e-01
7.58452833e-01 5.99345446e-01 1.93161607e-01 3.80777180e-01
4.04474735e-01 1.30042389e-01 2.36794397e-01 3.61315208e-03
6.35959879e-02 -3.14087957e-01 2.80919760e-01 1.01773834e+00
6.46315701e-03 -1.01513609e-01 -6.28028393e-01 5.56157768e-01
-1.97227073e+00 -1.00301433e+00 5.94865307e-02 1.92185473e+00
8.93512189e-01 1.41463771e-01 1.27284542e-01 -9.91093144e-02
9.64645684e-01 1.68823913e-01 -4.34331179e-01 2.67706603e-01
-3.40009212e-01 5.54942548e-01 3.78212966e-02 4.88846660e-01
-1.40905225e+00 1.11181045e+00 7.81870890e+00 1.08245254e+00
-7.25723684e-01 4.43864793e-01 7.38110840e-01 3.50439698e-02
-5.58072507e-01 9.79867503e-02 -5.41602373e-01 5.44683695e-01
7.22547472e-01 2.89202154e-01 1.78532556e-01 8.25909019e-01
-3.83437686e-02 -2.74779499e-02 -1.14978743e+00 9.40204442e-01
3.49971130e-02 -1.14487720e+00 1.61583781e-01 -2.08616443e-02
1.14713168e+00 1.69742703e-01 4.90794718e-01 4.83746052e-01
1.06874299e+00 -1.12833548e+00 5.79872847e-01 4.70803738e-01
3.58973414e-01 -6.13855600e-01 3.44526857e-01 5.38585722e-01
-1.07407069e+00 -1.21515803e-01 -2.72621155e-01 1.12607591e-02
-1.26605451e-01 6.42005682e-01 -2.83245504e-01 3.60568494e-01
7.65790641e-01 6.72910511e-01 -5.38934648e-01 7.36286759e-01
-5.92119038e-01 9.85582530e-01 -1.86029375e-01 6.15493357e-01
3.39303136e-01 -4.14150029e-01 2.98046887e-01 1.37665939e+00
-2.13822246e-01 -2.02453494e-01 5.17032266e-01 1.29455376e+00
-2.88593709e-01 -2.76131183e-01 -1.20379291e-01 6.83482289e-02
6.63056016e-01 1.36227047e+00 -1.11267507e+00 -7.95087755e-01
-3.51865143e-01 1.26916683e+00 7.45649993e-01 7.20934033e-01
-7.94296265e-01 3.55679482e-01 2.69441366e-01 -3.08671683e-01
8.01100910e-01 -2.10617661e-01 -2.46817261e-01 -1.25296080e+00
-3.04674298e-01 -7.46417940e-01 8.23990583e-01 -6.27844393e-01
-1.78226674e+00 4.74615365e-01 1.65776834e-01 -9.39947724e-01
-1.43961519e-01 -4.33643073e-01 -6.22260273e-01 7.15134859e-01
-1.43017936e+00 -1.11549425e+00 -9.81467366e-02 8.22754622e-01
6.38085127e-01 -5.46206981e-02 9.54050183e-01 4.48000319e-02
-4.77837741e-01 3.08924317e-01 2.78684795e-01 8.91849250e-02
6.69735312e-01 -1.61435044e+00 -1.21689089e-01 6.86131954e-01
5.56979060e-01 8.51134658e-01 4.80806708e-01 -4.69155997e-01
-5.15019298e-01 -1.14359570e+00 4.69491571e-01 -6.44723296e-01
3.26313674e-01 -2.45835289e-01 -9.09954846e-01 9.74345565e-01
4.26037163e-01 1.43102109e-01 1.29667687e+00 2.13347971e-01
-8.42880070e-01 3.89346302e-01 -1.12243927e+00 3.51543933e-01
9.56537366e-01 -7.71227002e-01 -1.05804157e+00 1.61154103e-02
4.41481650e-01 3.07578415e-01 -5.80908954e-01 3.72350693e-01
2.40151390e-01 -1.10696566e+00 8.73601615e-01 -7.91341305e-01
2.60240406e-01 -4.44173664e-01 -4.92894024e-01 -1.32488501e+00
-9.14478898e-01 -2.76588202e-01 -3.47687602e-01 1.14672768e+00
2.95995712e-01 -3.15802038e-01 9.62580740e-01 3.79352361e-01
-4.82617766e-02 -3.43882889e-01 -5.94543040e-01 -5.79530954e-01
2.06952095e-01 -4.27577168e-01 2.10430264e-01 1.23326182e+00
-2.92784005e-01 7.91950881e-01 -2.04202220e-01 4.09078151e-01
1.22676218e+00 4.69252169e-01 2.19008163e-01 -1.44194090e+00
-7.90858030e-01 -4.29637432e-01 1.22779123e-01 -1.31826305e+00
6.20759904e-01 -1.07014775e+00 2.60810792e-01 -1.34304142e+00
9.31299031e-01 -3.44942868e-01 -9.00342703e-01 1.58901706e-01
-7.59125352e-01 4.40596998e-01 -7.55696669e-02 5.95255017e-01
-1.49410009e+00 5.72848737e-01 7.96331346e-01 -2.30363458e-01
-7.05033317e-02 -1.55950487e-01 -8.45069528e-01 1.09932482e+00
6.09813988e-01 -6.59596086e-01 -5.54634750e-01 -4.35330451e-01
1.53111815e-02 -4.01362062e-01 2.09366143e-01 -7.36541271e-01
1.90834954e-01 -4.14203033e-02 7.03192055e-01 -6.76754951e-01
3.00990701e-01 -5.90294003e-01 -1.53413340e-02 2.32991576e-01
-3.83184075e-01 -3.71970952e-01 -1.05273165e-01 7.64158487e-01
-3.62795472e-01 -5.32622188e-02 1.09341168e+00 -3.57713938e-01
-9.20869112e-01 3.72398973e-01 -6.42280757e-01 3.80483896e-01
8.33773732e-01 6.23019934e-02 -6.05620118e-03 -1.80627227e-01
-1.26397371e+00 3.45085889e-01 4.47107881e-01 1.36364973e-03
6.08449459e-01 -1.28718281e+00 -6.47959471e-01 1.49008200e-01
8.93106312e-02 1.62232757e-01 2.76176900e-01 8.51356566e-01
1.05114225e-02 -3.06099560e-02 -1.25942886e-01 -8.27711940e-01
-9.84209120e-01 7.57969737e-01 2.64288336e-01 -5.52786171e-01
-2.69695759e-01 1.11080503e+00 7.80382395e-01 -5.94466746e-01
6.25219867e-02 2.59648085e-01 -2.39775926e-01 -2.87070554e-02
2.66696960e-01 3.21535021e-01 -4.37934250e-01 -5.18684149e-01
-4.68839288e-01 2.70606428e-01 -1.80382580e-01 -4.45025086e-01
1.41089618e+00 -2.87297845e-01 -2.10438535e-01 6.31044924e-01
1.04920912e+00 1.23595640e-01 -1.65192640e+00 -4.77353394e-01
4.11570728e-01 -1.03249826e-01 -1.63968969e-02 -5.95162690e-01
-1.03393257e+00 8.87084544e-01 4.51031983e-01 -1.87344506e-01
1.09678447e+00 5.40713966e-01 2.70161748e-01 1.30226508e-01
7.70340711e-02 -1.31992054e+00 4.45006818e-01 3.04511964e-01
1.44289806e-01 -1.23843169e+00 -1.92007840e-01 -4.72786874e-01
-6.77960277e-01 5.65471113e-01 4.42459017e-01 -4.95156020e-01
9.66221392e-01 3.24524999e-01 1.44781172e-01 -4.73510504e-01
-8.24594378e-01 -3.22372437e-01 3.38131249e-01 6.84985459e-01
3.35558504e-01 2.96165317e-01 1.37517944e-01 8.51307213e-01
3.80239904e-01 -2.63984621e-01 -2.25226119e-01 8.34994733e-01
-5.24782360e-01 -9.47108805e-01 -2.04261065e-01 2.09832117e-01
-4.65141386e-01 -2.00443104e-01 -2.30430335e-01 4.92553413e-01
3.32225561e-01 1.09711981e+00 2.85914868e-01 -1.82246789e-01
-4.07940090e-01 3.40654224e-01 7.63201594e-01 -8.48427236e-01
-1.33958727e-01 6.22905195e-01 -4.25799668e-01 -3.31792980e-01
-9.72044349e-01 -6.33827925e-01 -1.38799202e+00 2.20288873e-01
-2.02493370e-01 5.17431259e-01 1.52716562e-01 1.25003982e+00
2.08170488e-01 4.87276644e-01 6.89441144e-01 -6.30790591e-01
-5.66910148e-01 -1.03968966e+00 -9.07727361e-01 9.28748965e-01
-1.17347322e-01 -6.67785525e-01 -6.10156476e-01 3.30471903e-01] | [9.387797355651855, 2.926084280014038] |
566803ac-807d-4fb4-96fd-206e5759c1f4 | inter-brain-substrates-of-role-switching | 2303.04902 | null | https://arxiv.org/abs/2303.04902v1 | https://arxiv.org/pdf/2303.04902v1.pdf | Inter-brain substrates of role switching during mother-child interaction | Mother-child interaction is highly dynamic and reciprocal. Switching roles in these back-and-forth interactions serves as a crucial feature of reciprocal behaviors while the underlying neural entrainment is still not well-studied. Here, we designed a role-controlled cooperative task with dual EEG recording to study how differently two brains interact when mothers and children hold different roles. When children were actors and mothers were observers, mother-child inter-brain synchrony emerged within the theta oscillations and the frontal lobe, which highly correlated with children's attachment to their mothers. When their roles were reversed, this synchrony was shifted to the alpha oscillations and the central area and associated with mothers' perception of their relationship with their children. The results suggested an observer-actor neural alignment within the actor's oscillations, which was modulated by the actor-toward-observer emotional bonding. Our findings contribute to the understanding of how inter-brain synchrony is established and dynamically changed during mother-child reciprocal interaction. | ['Xiaoli Guo', 'Yunting Zhang', 'Shanbao Tong', 'Fan Jiang', 'Yue Fang', 'Wen Shi', 'Qi Zhu', 'Haiwa Wang', 'Jiayang Xu', 'Saishuang Wu', 'Yamin Li'] | 2023-03-08 | null | null | null | null | ['eeg', 'eeg'] | ['methodology', 'time-series'] | [-2.29021594e-01 2.11690724e-01 1.13471178e-02 -5.89548588e-01
5.94483197e-01 -3.46503854e-01 4.11393166e-01 3.57526183e-01
-1.19579688e-01 4.64891255e-01 1.23476714e-01 6.79866791e-01
-1.52182430e-01 -6.50475264e-01 -4.91529882e-01 -1.31290066e+00
-4.57236499e-01 7.36264363e-02 -1.00665383e-01 -2.91380525e-01
3.14873070e-01 2.04408303e-01 -1.56833553e+00 4.34716105e-01
3.94933939e-01 5.82369447e-01 1.14345029e-01 3.53976786e-01
1.81724504e-01 4.66790527e-01 -5.39703310e-01 -2.26350814e-01
-2.59973079e-01 -8.55661392e-01 -3.18552464e-01 -2.22779751e-01
-5.46663880e-01 -1.70478579e-02 2.86581367e-01 8.79086375e-01
4.40435231e-01 -2.08758250e-01 3.76131654e-01 -1.45932055e+00
-5.93853354e-01 1.03098416e+00 -9.40890133e-01 8.36616099e-01
5.86914599e-01 -3.36307049e-01 3.57326418e-01 -7.78283477e-01
4.70590830e-01 1.03601515e+00 2.39553124e-01 7.42896497e-01
-1.44144917e+00 -1.30647969e+00 2.08059028e-01 3.72194737e-01
-1.39015985e+00 -8.37881207e-01 8.87575686e-01 -4.69780833e-01
8.05414915e-01 -1.99446306e-01 1.74484921e+00 1.37031496e+00
7.56578505e-01 -5.50691262e-02 1.06568968e+00 -1.40197307e-01
-1.08741134e-01 2.88360000e-01 5.15664704e-02 8.20551161e-03
2.77656883e-01 9.10600498e-02 -1.07555687e+00 -3.62918675e-01
5.62900066e-01 -3.62041146e-01 -5.72990239e-01 1.56792879e-01
-1.20618510e+00 5.71757376e-01 4.72310074e-02 1.31636262e+00
-4.05852795e-01 1.69767410e-01 3.06774616e-01 4.48498189e-01
5.98378003e-01 1.63973555e-01 -3.01952809e-01 -5.21238506e-01
-5.55841506e-01 -1.84672982e-01 4.45373923e-01 3.77032906e-01
2.56528318e-01 -3.31052899e-01 2.26004511e-01 6.57982588e-01
2.63637334e-01 2.05433875e-01 4.42370027e-01 -5.54669499e-01
-1.48540854e-01 4.69907790e-01 -3.07843864e-01 -1.16021454e+00
-7.79356956e-01 -3.68111253e-01 -6.44252777e-01 -7.15866387e-02
2.90512502e-01 -2.34098539e-01 4.79829609e-01 2.21649790e+00
5.31151474e-01 1.98609084e-01 5.47120124e-02 7.99174607e-01
8.91859412e-01 3.99757534e-01 3.89190093e-02 -9.05600905e-01
1.49221289e+00 -2.04957545e-01 -1.18216336e+00 1.20442994e-01
2.53332406e-01 -5.00645578e-01 2.52457410e-01 2.63044946e-02
-1.41536677e+00 -3.85808080e-01 -9.41681206e-01 3.93136203e-01
-2.84366459e-02 -4.53042537e-01 5.18384278e-01 3.44062775e-01
-1.13247740e+00 4.62342322e-01 -1.01400626e+00 -6.87499523e-01
4.22852516e-01 4.11790729e-01 -7.98155010e-01 8.21093917e-01
-1.05902684e+00 9.54208672e-01 -3.04235090e-02 3.78126293e-01
-5.45291543e-01 -7.71745861e-01 -4.04815644e-01 3.52176487e-01
-3.83862257e-01 -3.02305102e-01 9.94529843e-01 -1.38226533e+00
-1.79055893e+00 1.22764528e+00 -9.99296457e-02 2.34889925e-01
-3.88627052e-01 3.20847541e-01 -7.02195406e-01 6.12273693e-01
2.95392424e-02 5.07379413e-01 4.13779795e-01 -1.14688623e+00
-5.15998676e-02 -8.39023113e-01 -4.03967470e-01 2.86509067e-01
-2.42632672e-01 5.32373130e-01 6.24120414e-01 -3.68624181e-01
6.75447762e-01 -7.22747386e-01 5.00221789e-01 -1.61515534e-01
3.36094946e-01 -4.16419566e-01 3.16856951e-01 1.94052845e-01
9.53287363e-01 -2.61429214e+00 2.42324010e-01 2.61920571e-01
6.29629970e-01 -4.83406425e-01 4.81753238e-03 8.71559680e-01
-8.58764470e-01 -1.94696784e-01 3.19912732e-01 1.12455241e-01
-2.70399779e-01 -9.06187668e-02 2.26048037e-01 1.41586494e+00
3.32818106e-02 6.21412456e-01 -9.90516722e-01 -2.58257836e-01
-5.71346879e-01 4.70174730e-01 -3.35177213e-01 7.14118004e-01
4.99717355e-01 1.08226883e+00 -1.99636862e-01 4.23565358e-01
8.11152399e-01 -6.38742968e-02 4.26826090e-01 2.00547557e-03
-4.37798053e-01 1.62969545e-01 -2.17914060e-01 1.15506005e+00
1.48647428e-01 1.11090195e+00 7.56847024e-01 -1.18108225e+00
1.02074814e+00 8.11511993e-01 4.69638616e-01 -9.46951270e-01
6.88445747e-01 1.43729430e-02 1.06356680e+00 -7.70957112e-01
-5.51866591e-01 -3.68555784e-01 3.40093315e-01 8.22591841e-01
2.29577437e-01 -2.19541401e-01 7.58381486e-02 -8.57821330e-02
7.47332752e-01 -1.99949563e-01 6.27121150e-01 -9.22379255e-01
1.92043725e-02 -8.75548363e-01 5.73094010e-01 -1.68188855e-01
-2.18498752e-01 1.06432147e-01 1.17427683e+00 -1.60713837e-01
-1.10238902e-01 -9.51114535e-01 -3.87769729e-01 1.17988801e+00
2.53016710e-01 6.54641241e-02 -8.92960608e-01 3.30078751e-01
-3.04476291e-01 6.49198353e-01 -1.25443828e+00 -6.54904723e-01
-4.14394408e-01 -6.81830406e-01 2.73879915e-01 1.29601732e-01
1.57052130e-01 -1.23046267e+00 -1.23994970e+00 2.64489323e-01
1.43136352e-01 -8.34235311e-01 -3.91845942e-01 4.15282577e-01
-7.55273759e-01 -1.17146587e+00 -2.77373821e-01 -6.99071825e-01
8.47251594e-01 3.86315398e-02 8.90077174e-01 1.26924530e-01
-1.60843488e-02 1.94872692e-01 -2.70931512e-01 -4.68413264e-01
-2.88807273e-01 -3.13764334e-01 4.10762072e-01 2.85154190e-02
3.06393832e-01 -1.70040762e+00 -6.92103863e-01 4.63057190e-01
-7.06944466e-01 1.12437949e-01 7.84534514e-02 3.15547138e-01
-3.21415365e-01 -5.34897685e-01 8.09269071e-01 -3.21074218e-01
5.38057208e-01 -9.81145561e-01 -3.71287942e-01 -6.79355264e-02
-2.46298179e-01 -4.04604167e-01 2.71861672e-01 -9.61151361e-01
-1.20655715e+00 -7.42296398e-01 2.98778743e-01 1.76291406e-01
-2.13648200e-01 2.56035119e-01 -3.24474014e-02 1.74380783e-02
3.57562363e-01 -1.32591009e-01 6.64014667e-02 2.56675988e-01
-4.55261648e-01 5.20165503e-01 4.30972904e-01 -4.59150016e-01
1.65033937e-01 2.22520292e-01 -1.84678003e-01 -5.94029367e-01
-5.50447464e-01 2.96944141e-01 -5.72912931e-01 -5.68826556e-01
1.23022926e+00 -8.03139567e-01 -1.16440606e+00 4.27025795e-01
-1.46288180e+00 -3.97176504e-01 2.09878460e-01 1.06781960e+00
-6.96745813e-02 -3.04980546e-01 -4.32634681e-01 -5.48833072e-01
-4.19532180e-01 -7.55939186e-01 4.53618616e-01 5.93548298e-01
-6.61717176e-01 -7.17169583e-01 6.62207603e-01 1.15583986e-01
3.68971378e-01 2.48714551e-01 1.13630688e+00 -8.88276458e-01
2.57396042e-01 2.42080256e-01 3.28740686e-01 -3.63997936e-01
3.30407441e-01 2.65357316e-01 -8.81995976e-01 6.47450760e-02
6.94896936e-01 -1.64370105e-01 -2.56978184e-01 3.27688456e-01
5.34534335e-01 -1.71518192e-01 -4.15940672e-01 4.65108514e-01
6.55912995e-01 9.55240488e-01 3.16201150e-01 -3.95796776e-01
-1.49098456e-01 1.19401515e+00 2.40826398e-01 4.65968758e-01
3.45421880e-01 3.28997791e-01 4.91818696e-01 3.94226126e-02
5.00871718e-01 2.08937615e-01 4.50157672e-01 1.05679667e+00
-2.20712006e-01 -7.77200013e-02 -5.69026113e-01 4.73362178e-01
-1.36137819e+00 -1.02465034e+00 -5.12675881e-01 1.80170739e+00
1.04160964e+00 -1.54015467e-01 -2.93428361e-01 -1.83377266e-01
7.51291275e-01 -2.06083447e-01 -3.71648788e-01 -5.15043616e-01
3.61344665e-02 3.88100922e-01 -4.62577164e-01 -3.63884792e-02
-3.97949219e-02 4.70586836e-01 6.49289751e+00 5.36901765e-02
-1.24214089e+00 4.55395669e-01 6.18478417e-01 -5.98207355e-01
-2.31267527e-01 -1.92462042e-01 -8.52224901e-02 4.86436963e-01
9.35619712e-01 -2.86188662e-01 5.54854155e-01 2.19574764e-01
3.86133045e-01 -8.02271843e-01 -1.57256389e+00 1.07667279e+00
6.97419494e-02 -6.80077016e-01 -1.21050298e+00 7.71723539e-02
2.02013731e-01 -4.69296157e-01 -2.93147385e-01 -2.34817192e-01
-3.70058358e-01 -1.14092350e+00 6.91368818e-01 4.85387832e-01
6.66738987e-01 -5.24137974e-01 6.17285013e-01 3.65430474e-01
-1.08915997e+00 2.55050898e-01 1.32514969e-01 -6.59551620e-01
1.23274386e-01 5.01653135e-01 -3.54477704e-01 -3.40442747e-01
9.52122688e-01 4.04487282e-01 8.26912522e-02 2.30196133e-01
-2.77946830e-01 5.25357544e-01 -6.56574070e-02 -3.04548919e-01
-5.06683886e-01 -6.45796299e-01 6.40742064e-01 5.72694957e-01
3.46143901e-01 7.85160482e-01 -8.02120268e-01 1.34762561e+00
2.30935603e-01 -1.03014894e-01 -5.83052337e-01 -2.79609084e-01
6.04412913e-01 1.49121356e+00 -1.26518071e+00 -1.99048314e-02
-7.80701190e-02 5.54883540e-01 3.67973745e-01 2.30979100e-01
-1.06509960e+00 -2.34192871e-02 7.09460318e-01 4.85544242e-02
-2.95049489e-01 -4.28058468e-02 -7.25105405e-02 -1.03060174e+00
-2.83274949e-02 -2.37462282e-01 -2.83838779e-01 -9.24522817e-01
-7.62975752e-01 6.85270131e-01 5.29017389e-01 -5.33466578e-01
1.40232325e-01 4.44455355e-01 -1.09400797e+00 4.41420168e-01
-6.17002189e-01 -7.06356883e-01 -3.72592479e-01 2.89672762e-01
1.68354977e-02 2.54909039e-01 9.00997698e-01 1.20167241e-01
-9.31983054e-01 3.90666753e-01 -6.63341165e-01 -1.41608626e-01
6.76506937e-01 -4.83802289e-01 -4.75803673e-01 2.27814555e-01
-2.26406217e-01 7.61702657e-01 5.14292717e-01 -3.17690730e-01
-1.00642371e+00 -1.75260767e-01 1.07516837e+00 1.11013934e-01
7.43752480e-01 -1.05319202e+00 -8.38115335e-01 4.49155360e-01
1.03225207e+00 1.26436815e-01 1.30960596e+00 -9.73770022e-02
1.52759179e-01 -2.98109502e-01 -1.08220780e+00 5.52919567e-01
1.31187534e+00 -3.24238986e-01 -7.23798096e-01 1.00965716e-01
4.78207707e-01 -1.31675333e-01 -9.87215638e-01 1.68744519e-01
9.05690372e-01 -1.49101818e+00 2.41748169e-01 -1.15926810e-01
2.57110685e-01 1.02168351e-01 1.36479557e-01 -1.49083114e+00
-4.26768690e-01 -4.93591666e-01 5.59401035e-01 1.56980669e+00
6.68030679e-02 -1.24250734e+00 -2.51627803e-01 3.41183931e-01
9.51899588e-02 -7.52299905e-01 -1.36827528e+00 7.67640769e-02
-1.22244328e-01 1.70182601e-01 6.34849548e-01 1.01228607e+00
9.26399589e-01 3.54527950e-01 4.35333908e-01 -2.47803275e-02
-1.59028456e-01 -1.22875549e-01 1.65877357e-01 -1.13795233e+00
-3.76879796e-02 -5.24669826e-01 -2.77475417e-01 1.21204508e-02
4.07482415e-01 -7.15837479e-01 -1.11503877e-01 -8.59652460e-01
3.89220089e-01 7.27970228e-02 1.13670252e-01 4.75824535e-01
1.79231405e-01 -1.43264413e-01 -2.39809051e-01 -2.21645460e-01
-9.99742523e-02 7.58560419e-01 1.09839058e+00 7.02196300e-01
-5.98227561e-01 -1.19648799e-01 -8.69280636e-01 7.76282907e-01
6.68754220e-01 -7.68426180e-01 -4.74441618e-01 -1.11105412e-01
5.59267879e-01 7.99017966e-01 -6.64897040e-02 -6.54151917e-01
3.00730467e-01 -1.03900336e-01 2.82935113e-01 -2.01970279e-01
2.50601500e-01 -9.09369349e-01 7.13801324e-01 5.91434062e-01
-1.36561543e-01 3.24414104e-01 2.20559519e-02 -1.63073063e-01
-2.38281637e-02 1.20280728e-01 1.17667842e+00 2.33459517e-01
5.67297518e-01 -1.75948367e-01 -1.03438139e+00 -1.60251409e-01
1.73912978e+00 -1.94375962e-01 -3.21590036e-01 -3.12655121e-01
-8.73161376e-01 6.14342093e-02 3.25140923e-01 1.18891284e-01
3.20529431e-01 -1.20822108e+00 -3.96161109e-01 4.84665543e-01
-2.63998151e-01 -1.47052258e-01 4.31832701e-01 1.86957371e+00
-1.11972123e-01 -4.96066064e-02 -4.00894940e-01 -6.21768057e-01
-1.56098199e+00 -2.43460964e-02 3.97819161e-01 2.63714939e-01
-2.78097928e-01 1.26107121e+00 6.52031600e-01 2.32395783e-01
-1.17630444e-01 -4.89002407e-01 -7.23217666e-01 8.32421541e-01
4.58345532e-01 4.22918469e-01 -2.25495368e-01 -5.57330728e-01
-6.35275900e-01 5.57405651e-01 4.65940088e-01 4.82304879e-02
1.56774187e+00 -1.37210324e-01 -1.25309646e+00 9.96258855e-01
1.04023755e+00 1.65589750e-01 -6.45568848e-01 4.38853294e-01
-4.26739722e-01 -3.19049001e-01 -5.47132075e-01 -2.96938121e-01
-1.44371653e+00 9.21655416e-01 5.45285881e-01 4.18187708e-01
1.24374187e+00 5.06922364e-01 1.86805680e-01 -1.53164551e-01
4.86713499e-01 -8.97031903e-01 2.38684669e-01 8.72658938e-03
1.32301140e+00 -6.61681056e-01 -8.12352523e-02 -2.73405731e-01
-4.05758202e-01 1.20640635e+00 9.47914779e-01 -2.40124181e-01
1.06621265e+00 7.96678662e-01 -2.32707620e-01 -4.93350923e-01
-1.57007253e+00 5.31068623e-01 1.06511548e-01 5.68182588e-01
9.42802489e-01 1.59558766e-02 -6.28228903e-01 1.11147428e+00
-5.36452472e-01 -2.79778898e-01 4.02826250e-01 6.42803073e-01
-1.22255236e-01 -8.08470130e-01 -6.60167754e-01 2.30998114e-01
-4.61214900e-01 1.13748424e-01 -3.33533198e-01 8.28691900e-01
6.56780779e-01 1.09261143e+00 6.42842650e-01 -9.58254859e-02
7.63987452e-02 2.10279316e-01 9.61805284e-01 -8.05709660e-01
-9.13886964e-01 5.62218726e-01 -3.58531438e-02 -6.83048248e-01
-9.12178993e-01 -1.17897606e+00 -1.63875175e+00 -7.25008771e-02
-3.02278578e-01 3.42149854e-01 4.61799592e-01 1.00056314e+00
3.39577943e-01 3.95217419e-01 6.08478904e-01 -1.06935060e+00
4.58918929e-01 -1.13958931e+00 -9.57786977e-01 -1.46834403e-01
3.80510926e-01 -8.62766862e-01 -5.31508088e-01 -1.36260152e-01] | [12.913161277770996, 3.3674049377441406] |
c92f4136-0e18-4def-8457-f0bdc940d4e3 | the-second-place-solution-for-cvpr-vision-23 | 2306.14116 | null | https://arxiv.org/abs/2306.14116v1 | https://arxiv.org/pdf/2306.14116v1.pdf | The Second-place Solution for CVPR VISION 23 Challenge Track 1 -- Data Effificient Defect Detection | The Vision Challenge Track 1 for Data-Effificient Defect Detection requires competitors to instance segment 14 industrial inspection datasets in a data-defificient setting. This report introduces the technical details of the team Aoi-overfifitting-Team for this challenge. Our method focuses on the key problem of segmentation quality of defect masks in scenarios with limited training samples. Based on the Hybrid Task Cascade (HTC) instance segmentation algorithm, we connect the transformer backbone (Swin-B) through composite connections inspired by CBNetv2 to enhance the baseline results. Additionally, we propose two model ensemble methods to further enhance the segmentation effect: one incorporates semantic segmentation into instance segmentation, while the other employs multi-instance segmentation fusion algorithms. Finally, using multi-scale training and test-time augmentation (TTA), we achieve an average mAP@0.50:0.95 of more than 48.49% and an average mAR@0.50:0.95 of 66.71% on the test set of the Data Effificient Defect Detection Challenge. The code is available at https://github.com/love6tao/Aoi-overfitting-team | ['Zhengtao Zhang', 'Fei Shen', 'Chengkan Lv', 'Danfeng Liu', 'Yonghao He', 'Jianwen Han', 'Hengliang Luo', 'Zhen Qu', 'Xian Tao'] | 2023-06-25 | null | null | null | null | ['instance-segmentation', 'defect-detection'] | ['computer-vision', 'computer-vision'] | [ 2.79573381e-01 3.00143510e-01 2.23374456e-01 -3.53791505e-01
-1.21476638e+00 -3.97389084e-01 6.11683726e-02 -1.69710979e-01
-7.81370178e-02 1.14042714e-01 -4.03981537e-01 -3.00794423e-01
2.54992023e-02 -6.26516998e-01 -7.89191067e-01 -5.88736355e-01
2.61132240e-01 4.65579331e-01 3.11017811e-01 -1.04425639e-01
2.38444924e-01 1.34740785e-01 -1.30821443e+00 5.32318473e-01
1.11920273e+00 1.23804605e+00 2.29045317e-01 6.89917266e-01
1.60728931e-01 4.53670889e-01 -9.63057399e-01 -5.10278404e-01
5.78010321e-01 -3.12029570e-01 -9.10785317e-01 3.82791996e-01
7.32195139e-01 -1.90178186e-01 5.89714795e-02 1.13717794e+00
4.46213782e-01 -1.60253197e-01 4.60005194e-01 -1.28612339e+00
-5.09834945e-01 6.63529634e-01 -8.05682123e-01 2.66640782e-01
-2.44702250e-01 4.85957712e-01 9.95338321e-01 -8.56741130e-01
4.48475331e-01 1.05819547e+00 8.80120337e-01 3.40962172e-01
-1.32080758e+00 -5.05404890e-01 1.92339793e-01 1.46525860e-01
-1.09084642e+00 -1.27645686e-01 7.24683046e-01 -5.10834992e-01
9.66423512e-01 1.67272240e-01 5.32349885e-01 9.11737323e-01
3.92725356e-02 1.17352819e+00 1.03348172e+00 -3.53229851e-01
1.85133461e-02 -2.07373232e-01 4.87153441e-01 8.81109416e-01
1.90203652e-01 3.42291147e-02 -2.20102042e-01 4.71273094e-01
1.04041541e+00 -1.49901211e-01 -1.51593313e-01 1.20292626e-01
-9.35382366e-01 8.15482914e-01 5.56416810e-01 3.25051457e-01
-1.91959366e-01 1.80947289e-01 5.84373355e-01 4.94589061e-01
5.37306190e-01 9.07487929e-01 -8.29764485e-01 1.47283703e-01
-1.01391339e+00 2.70960145e-02 4.70233291e-01 1.08708918e+00
5.81321418e-01 3.32597911e-01 -4.05181736e-01 1.13484812e+00
2.60296702e-01 3.55709881e-01 -4.42557689e-03 -1.32027769e+00
5.82773685e-01 5.12383044e-01 -1.80182159e-01 -4.74350482e-01
-4.45779622e-01 -7.87877858e-01 -7.17029393e-01 4.55121934e-01
5.76658905e-01 -1.91305906e-01 -1.46768403e+00 1.06763768e+00
3.84676568e-02 1.07212603e-01 -2.32325569e-01 7.76641607e-01
8.46862316e-01 1.77088514e-01 8.57870206e-02 2.45083839e-01
1.31737304e+00 -1.31181979e+00 -4.52231824e-01 -2.45567992e-01
7.59795427e-01 -8.16867352e-01 1.21579194e+00 8.53143632e-01
-1.13558412e+00 -1.03576136e+00 -1.24420822e+00 1.75670490e-01
-1.01785198e-01 4.81379896e-01 3.25348020e-01 5.93926489e-01
-9.89210725e-01 6.99086487e-01 -8.78013670e-01 -6.58267438e-02
1.01322258e+00 3.17776948e-01 -1.04575269e-01 -2.37557426e-01
-7.05580235e-01 6.16536558e-01 3.59428197e-01 2.24567696e-01
-1.34815180e+00 -1.05251503e+00 -6.65740192e-01 -3.20631772e-01
5.28059483e-01 -4.21625912e-01 1.30299437e+00 -7.92520046e-01
-1.12014246e+00 9.91384983e-01 4.73298341e-01 -6.30163848e-01
7.19812989e-01 -5.16250014e-01 -2.11026073e-01 1.98472798e-01
3.32612216e-01 6.99051082e-01 6.35360181e-01 -1.52318609e+00
-1.03262174e+00 -4.81216758e-01 8.15331265e-02 -2.49565184e-01
2.27383807e-01 -1.31437346e-01 -7.38768876e-01 -6.65071547e-01
1.87384278e-01 -6.67693198e-01 -3.17770392e-01 -2.77607679e-01
-9.63562310e-01 -1.95583060e-01 1.05824065e+00 -1.18794692e+00
8.73280585e-01 -2.24243665e+00 -2.65545577e-01 2.03452438e-01
3.57938051e-01 4.56576020e-01 -2.97441363e-01 -1.91480383e-01
-1.15386531e-01 1.54301211e-01 -6.79291308e-01 -6.51119828e-01
2.63002561e-03 4.03986946e-02 2.02346176e-01 3.74954581e-01
5.13311863e-01 9.20938194e-01 -5.80451429e-01 -4.08565164e-01
6.32945120e-01 4.26974148e-02 -5.87272882e-01 1.34969503e-01
-1.39410153e-01 5.90454042e-01 -2.75684565e-01 1.14125013e+00
7.93182135e-01 -5.22685330e-03 -2.38069952e-01 -6.54500306e-01
-6.63534403e-02 -2.77569234e-01 -1.05944407e+00 2.04170871e+00
-4.37614918e-01 4.24248397e-01 3.37985158e-01 -1.44622052e+00
1.08144045e+00 2.74804145e-01 6.18274927e-01 -8.32164586e-01
2.15220660e-01 2.84579575e-01 1.20113734e-02 -4.90097821e-01
1.38276875e-01 7.27560744e-02 -6.29471755e-03 -6.81965649e-02
5.11250257e-01 -4.42623377e-01 5.21383286e-01 -1.43146351e-01
1.15219545e+00 2.10407332e-01 -4.29365635e-01 -4.51974362e-01
1.61888734e-01 1.66300282e-01 7.78012156e-01 8.20236742e-01
-4.24927890e-01 1.15303171e+00 4.89311069e-01 -2.50220895e-01
-1.01493132e+00 -1.14540982e+00 -3.09672654e-01 6.27192616e-01
1.75618336e-01 2.24109683e-02 -1.05083132e+00 -1.18036807e+00
5.65634854e-03 8.11805367e-01 -5.42687058e-01 -1.93930641e-01
-5.29523730e-01 -8.06335807e-01 6.95517540e-01 9.27275002e-01
9.75467145e-01 -9.79415953e-01 -4.30909276e-01 2.49073461e-01
-1.49509639e-01 -1.13653219e+00 -1.42570540e-01 3.20329279e-01
-8.40949178e-01 -1.26305854e+00 -7.05288589e-01 -9.52570915e-01
4.65182483e-01 -9.52485651e-02 1.41289294e+00 3.25121701e-01
-6.32058799e-01 2.84888297e-01 -4.70495462e-01 -6.66445613e-01
-3.76884878e-01 -1.51155442e-01 -5.93682170e-01 -2.44869605e-01
1.62962139e-01 -2.32579529e-01 -6.48273647e-01 4.74480659e-01
-5.85819066e-01 7.04616532e-02 4.73376691e-01 8.86018574e-01
7.69820154e-01 1.38753042e-01 7.12845385e-01 -9.66296375e-01
2.17773974e-01 -1.72874078e-01 -6.96362019e-01 3.00550431e-01
-8.04934859e-01 -4.71931875e-01 1.90960407e-01 -1.32129684e-01
-1.13461804e+00 -4.32540849e-02 -4.63774234e-01 -7.31604040e-01
-4.27210778e-01 2.53916472e-01 -2.46642306e-01 1.73252165e-01
6.27820909e-01 -3.27529192e-01 -8.75027776e-02 -6.67896152e-01
2.26660341e-01 4.82338667e-01 7.47679293e-01 -5.90276659e-01
5.12677193e-01 4.26228017e-01 -3.15210581e-01 -6.64504588e-01
-9.22367811e-01 -5.40242374e-01 -7.15618253e-01 -3.85221243e-01
1.28607476e+00 -8.64167631e-01 -2.28319883e-01 9.46918130e-01
-1.03939867e+00 -7.06631422e-01 -7.09119856e-01 2.73426443e-01
-6.05561197e-01 2.58495659e-01 -9.42740679e-01 -5.31656504e-01
-4.92186546e-01 -1.52825487e+00 1.26608968e+00 1.77807480e-01
1.59834027e-01 -8.65342021e-01 -4.62092876e-01 1.03035867e+00
2.21988395e-01 6.13071263e-01 6.74937785e-01 -8.35784256e-01
-3.47264796e-01 -2.60392547e-01 -4.73792374e-01 1.07629645e+00
2.72198655e-02 2.26897970e-01 -1.19346261e+00 -2.54029095e-01
-1.38903677e-01 -2.06211165e-01 1.17792153e+00 8.50073159e-01
1.26527452e+00 3.12681794e-01 -1.73446849e-01 3.22012186e-01
1.37604332e+00 2.84809381e-01 7.57577837e-01 2.25720227e-01
9.24083948e-01 5.47874570e-01 9.22388911e-01 3.08560431e-01
1.04173891e-01 5.46337247e-01 7.42282331e-01 -7.36613393e-01
-6.71690524e-01 1.34021342e-01 8.79101530e-02 5.84020019e-01
-2.35780701e-01 -5.18417433e-02 -1.05736864e+00 8.95093501e-01
-1.52059972e+00 -5.60293317e-01 -6.26718163e-01 1.80846655e+00
4.14951771e-01 5.20042658e-01 1.35735869e-01 5.01786351e-01
8.39379847e-01 -2.62865335e-01 -5.91022611e-01 -1.84535593e-01
-9.42609608e-02 3.92304659e-01 5.44773757e-01 2.69726664e-01
-1.33370221e+00 8.76607001e-01 5.50477409e+00 1.01661301e+00
-6.42937124e-01 4.25769657e-01 9.82730925e-01 1.60831496e-01
4.62056436e-02 -2.02068090e-01 -5.39372683e-01 3.40962201e-01
6.81755185e-01 4.89057362e-01 6.13571294e-02 7.12656796e-01
1.06328771e-01 -1.54747292e-01 -8.32976818e-01 7.33305156e-01
-1.10544525e-01 -1.25231326e+00 -4.15346980e-01 -2.42584180e-02
1.07970810e+00 1.92901000e-01 1.24340512e-01 5.16417027e-01
4.99094576e-01 -8.39550257e-01 9.53792751e-01 3.88693035e-01
8.57781231e-01 -5.84119260e-01 1.03523481e+00 -1.46331981e-01
-1.03512180e+00 -3.73950928e-01 -2.03503355e-01 4.70609486e-01
1.81225315e-01 1.05051851e+00 -7.69034445e-01 9.91013408e-01
1.07943296e+00 7.42692113e-01 -6.12479746e-01 1.25316823e+00
-2.60917306e-01 1.09756184e+00 -1.80398226e-01 8.45373213e-01
2.31011599e-01 -1.59982026e-01 3.98840189e-01 1.25357389e+00
2.83132792e-01 -3.62336785e-01 3.85821879e-01 1.06760538e+00
-1.03297614e-01 -3.79518390e-01 -2.66594768e-01 3.05457234e-01
2.05101892e-01 1.27374136e+00 -9.79999721e-01 -2.32438177e-01
-4.83294010e-01 8.42368782e-01 -2.40296379e-01 3.40259016e-01
-1.00697136e+00 -3.50579947e-01 4.94445860e-01 -1.48590654e-01
4.34976071e-01 1.15104392e-01 -8.38795364e-01 -6.27277017e-01
-1.51288211e-01 -9.92835224e-01 4.96947676e-01 -6.65358722e-01
-1.47962737e+00 4.20535743e-01 -1.57970101e-01 -1.16485441e+00
4.45469052e-01 -8.52168620e-01 -8.32372427e-01 6.85614705e-01
-1.24342680e+00 -1.60435033e+00 -4.35282886e-01 3.21747661e-01
1.20340729e+00 -1.91294417e-01 2.72351652e-01 4.88236755e-01
-1.01054049e+00 6.62079275e-01 -1.74607798e-01 3.37684810e-01
4.00361747e-01 -1.53556097e+00 6.17515981e-01 1.13760662e+00
5.97099103e-02 -9.40808207e-02 6.41868532e-01 -6.91418290e-01
-7.73034394e-01 -1.47677326e+00 7.60632232e-02 -7.02067077e-01
3.80452037e-01 -9.18619633e-02 -8.86766970e-01 6.49078369e-01
1.80423021e-01 -1.74819771e-02 2.53067017e-01 -2.43178643e-02
-2.02043101e-01 -1.46578729e-01 -1.50631344e+00 1.22330919e-01
9.28550482e-01 -3.03836077e-01 -3.75669152e-01 4.22221452e-01
8.83062601e-01 -5.11051059e-01 -1.31973636e+00 9.06897426e-01
1.36212826e-01 -8.68078768e-01 9.64115083e-01 -2.72165358e-01
5.75815022e-01 -3.20647746e-01 -3.38881880e-01 -1.14951634e+00
-2.50354737e-01 -1.58925459e-01 1.78827330e-01 1.41404986e+00
7.00905442e-01 -3.60382348e-01 7.07509041e-01 2.98478544e-01
-9.61048841e-01 -7.93697536e-01 -9.66014624e-01 -8.60039830e-01
2.70121515e-01 -8.20164680e-01 6.11960232e-01 7.15996027e-01
-6.34274006e-01 4.64050919e-02 1.92689463e-01 3.83079618e-01
7.82460392e-01 -3.38414381e-03 4.84171450e-01 -1.28451908e+00
-1.28416076e-01 -5.30834317e-01 -2.80313879e-01 -7.51413465e-01
-2.63464689e-01 -8.70427549e-01 2.55521357e-01 -1.74386156e+00
-1.09956130e-01 -4.32153553e-01 -5.02138495e-01 6.61331177e-01
-3.16833816e-02 4.91922975e-01 3.22109431e-01 -8.25096294e-03
-6.46195590e-01 2.17721045e-01 1.51162124e+00 -5.75351655e-01
7.87495449e-02 2.35325709e-01 -6.15729332e-01 5.78694701e-01
1.11103380e+00 -2.41723880e-01 -2.69532353e-01 -6.82141721e-01
-5.25697947e-01 -1.42927065e-01 6.26773119e-01 -1.29153991e+00
-2.78296292e-01 3.58492017e-01 3.00095826e-01 -6.12664223e-01
1.11882724e-01 -6.98453963e-01 -1.29853547e-01 6.55697048e-01
-3.39817926e-02 -1.59048229e-01 3.77407461e-01 3.89111936e-01
-1.91530585e-01 -3.64157587e-01 9.46233034e-01 -2.90173262e-01
-8.90327096e-01 1.48178846e-01 -1.31041288e-01 1.05190404e-01
9.23800766e-01 -4.12270367e-01 -4.14575994e-01 2.51015604e-01
-9.25154269e-01 4.43899989e-01 2.46046677e-01 4.55317616e-01
4.29575861e-01 -9.68438148e-01 -9.75543201e-01 2.78897017e-01
2.10532784e-01 3.19321930e-01 6.81849599e-01 1.18073976e+00
-4.85757023e-01 5.74718043e-02 -1.67511508e-01 -9.91714656e-01
-1.04289985e+00 9.64896604e-02 6.43041909e-01 -5.02067983e-01
-6.40803576e-01 1.35153639e+00 2.96786547e-01 -5.98252833e-01
8.00401252e-03 -4.93449032e-01 -6.55886233e-02 -1.27185956e-01
6.60261884e-02 8.71519744e-01 5.80163419e-01 -1.99811369e-01
-1.02557965e-01 4.08276349e-01 4.01040390e-02 2.35465243e-01
1.45081460e+00 8.99811387e-02 1.28829286e-01 1.47524908e-01
9.91984189e-01 -6.09972656e-01 -1.58922708e+00 9.48892757e-02
1.10665165e-01 -2.18689665e-01 2.61528790e-01 -1.36245191e+00
-1.80046296e+00 8.93264890e-01 1.05762935e+00 4.36712325e-01
1.24897969e+00 1.03353560e-01 7.84763157e-01 4.63301083e-03
3.55239719e-01 -1.40037978e+00 2.54104048e-01 3.24564457e-01
1.17502069e+00 -1.47100925e+00 -3.28434825e-01 -6.01779580e-01
-8.06509137e-01 8.12452912e-01 9.52571392e-01 -2.24452227e-01
5.91234624e-01 4.89074469e-01 3.73493046e-01 -7.04443514e-01
-2.66386241e-01 -1.88007265e-01 1.96637332e-01 9.02795494e-01
2.08472267e-01 1.32416189e-01 6.38007279e-03 8.31755400e-01
8.88124183e-02 -2.01104164e-01 2.13949442e-01 7.32210517e-01
-2.48743594e-01 -7.29785681e-01 -3.73673916e-01 1.03569198e+00
-4.74009395e-01 1.38861984e-01 2.03058980e-02 7.91866720e-01
5.04583478e-01 1.40779018e+00 1.10497795e-01 -7.01556385e-01
9.21479464e-01 -1.65470958e-01 4.19217557e-01 -5.76852918e-01
-9.27263379e-01 3.07314098e-01 2.57277250e-01 -7.75891900e-01
-2.75946409e-01 -7.58610368e-01 -1.35339344e+00 9.35166627e-02
-8.30790102e-01 -9.22280774e-02 6.01615846e-01 8.45947623e-01
2.32903078e-01 1.22190309e+00 5.14986038e-01 -8.33374381e-01
-4.08446938e-01 -1.20609641e+00 -8.02318513e-01 4.10202533e-01
1.09863363e-01 -7.00527132e-01 -2.81561077e-01 2.73791790e-01] | [7.678824424743652, 1.8586496114730835] |
c3709261-581c-49d5-9809-5cb6cfffeb90 | select-label-and-mix-learning-discriminative | 2012.03358 | null | https://arxiv.org/abs/2012.03358v2 | https://arxiv.org/pdf/2012.03358v2.pdf | Select, Label, and Mix: Learning Discriminative Invariant Feature Representations for Partial Domain Adaptation | Partial domain adaptation which assumes that the unknown target label space is a subset of the source label space has attracted much attention in computer vision. Despite recent progress, existing methods often suffer from three key problems: negative transfer, lack of discriminability, and domain invariance in the latent space. To alleviate the above issues, we develop a novel 'Select, Label, and Mix' (SLM) framework that aims to learn discriminative invariant feature representations for partial domain adaptation. First, we present an efficient "select" module that automatically filters out the outlier source samples to avoid negative transfer while aligning distributions across both domains. Second, the "label" module iteratively trains the classifier using both the labeled source domain data and the generated pseudo-labels for the target domain to enhance the discriminability of the latent space. Finally, the "mix" module utilizes domain mixup regularization jointly with the other two modules to explore more intrinsic structures across domains leading to a domain-invariant latent space for partial domain adaptation. Extensive experiments on several benchmark datasets for partial domain adaptation demonstrate the superiority of our proposed framework over state-of-the-art methods. | ['Abir Das', 'Kate Saenko', 'Rogerio Feris', 'Rameswar Panda', 'Aadarsh Sahoo'] | 2020-12-06 | null | null | null | null | ['partial-domain-adaptation'] | ['methodology'] | [ 3.96433055e-01 -3.66048723e-01 -5.13930023e-01 -5.87226689e-01
-7.67238915e-01 -7.65397370e-01 4.51119512e-01 -2.15364903e-01
-2.11702004e-01 8.53027940e-01 2.76072137e-02 1.35160521e-01
-5.05808331e-02 -4.23872739e-01 -4.80579257e-01 -1.06470764e+00
5.40885568e-01 4.96184409e-01 2.61312842e-01 3.24980170e-02
8.92455652e-02 2.01592967e-01 -1.22550356e+00 1.77840546e-01
1.31820238e+00 7.78617561e-01 4.86702397e-02 -2.49483690e-01
-2.59381145e-01 2.13055447e-01 -3.97802591e-01 -1.36393895e-02
3.62921983e-01 -5.98908544e-01 -4.46996808e-01 3.60871941e-01
3.68709207e-01 -1.87371969e-01 -2.69256290e-02 1.36783051e+00
5.21735668e-01 2.02412918e-01 1.15534592e+00 -1.48640561e+00
-8.32783580e-01 4.32559475e-02 -8.24487150e-01 -2.02102866e-02
2.81764902e-02 7.33390525e-02 6.99458182e-01 -1.02143502e+00
6.64853752e-01 1.36341298e+00 4.11317259e-01 7.26158440e-01
-1.57484937e+00 -1.20429790e+00 4.06874746e-01 1.87092163e-02
-1.34616435e+00 -3.02232236e-01 1.14043999e+00 -7.46768594e-01
3.84425431e-01 -2.37588987e-01 7.02342242e-02 1.33252895e+00
9.01731923e-02 7.85393596e-01 1.22553968e+00 -3.84774566e-01
3.58366281e-01 7.23563671e-01 2.18116015e-01 3.26435477e-01
5.54837845e-02 4.64467630e-02 -3.49903703e-01 -4.63434130e-01
7.66100764e-01 1.31152600e-01 -1.59712374e-01 -1.27578318e+00
-1.11492574e+00 8.84765744e-01 2.33151376e-01 1.79797709e-01
-2.07737908e-01 -5.74688196e-01 5.43935001e-01 4.18508798e-01
7.34440982e-01 1.28951684e-01 -5.09320557e-01 4.41304088e-01
-6.85045063e-01 1.26156271e-01 3.84022623e-01 1.19593894e+00
8.75919521e-01 -6.30259961e-02 -2.39967242e-01 1.39461672e+00
3.66900206e-01 4.98681277e-01 7.35551953e-01 -5.43655217e-01
7.41341472e-01 8.85865927e-01 1.87768817e-01 -6.59226894e-01
-7.94010162e-02 -6.60241663e-01 -9.77220953e-01 1.32428631e-01
3.53727728e-01 -9.83232260e-02 -8.98028016e-01 2.19133258e+00
5.65580606e-01 1.89462081e-01 2.27290839e-01 7.00224996e-01
3.99679691e-01 5.57003081e-01 4.33065772e-01 -3.17103982e-01
9.56296563e-01 -8.74463379e-01 -5.25267541e-01 -4.31646228e-01
6.68103933e-01 -6.80642843e-01 1.16627455e+00 6.43294230e-02
-6.05680108e-01 -6.52919710e-01 -1.10360444e+00 1.47529989e-01
-1.89453825e-01 4.34735209e-01 2.42831096e-01 3.02064359e-01
-5.16264021e-01 2.62475669e-01 -5.18132389e-01 -4.65962797e-01
5.04347503e-01 4.59863186e-01 -5.75041831e-01 -3.01249236e-01
-1.13679850e+00 4.51208860e-01 5.84110558e-01 -4.73779291e-01
-8.29378784e-01 -6.44642889e-01 -9.09717321e-01 -1.34009302e-01
1.57123148e-01 -5.75639486e-01 8.65538359e-01 -1.33901834e+00
-1.37816572e+00 1.04282308e+00 -2.45910421e-01 1.47652492e-01
5.16361773e-01 -1.66128516e-01 -5.02913177e-01 -2.26137131e-01
5.92882991e-01 5.57256460e-01 1.05989611e+00 -1.20596123e+00
-7.11721003e-01 -6.99171543e-01 -5.60442090e-01 4.30255026e-01
-6.58610582e-01 -1.60244748e-01 -3.03557903e-01 -8.71901333e-01
4.41763043e-01 -1.06920731e+00 2.88843736e-02 -3.01706437e-02
-1.47502795e-01 -4.52949971e-01 8.91453922e-01 -4.31457847e-01
1.07171464e+00 -2.66989326e+00 3.82218748e-01 1.73079625e-01
-8.72307941e-02 2.46841267e-01 -3.30207646e-01 5.52047528e-02
-6.07307732e-01 -2.71338612e-01 -3.58523995e-01 -2.77939528e-01
-3.45968641e-02 1.77090868e-01 -4.31525290e-01 6.22402668e-01
2.88069814e-01 4.33794588e-01 -8.36842299e-01 -4.76416528e-01
9.27676409e-02 1.77550837e-01 -5.37325025e-01 3.78232270e-01
-4.08726074e-02 9.13108766e-01 -5.56961358e-01 5.52190721e-01
1.08618438e+00 -2.32884690e-01 2.41992816e-01 -8.31272975e-02
1.85299098e-01 1.37892723e-01 -1.35664880e+00 1.73968923e+00
-3.82891931e-02 -6.55859429e-03 4.81645688e-02 -1.10209167e+00
1.22634602e+00 2.57488132e-01 5.22454321e-01 -6.38521731e-01
-1.57350935e-02 4.68503684e-01 -2.44838193e-01 -3.12522352e-01
-5.13561517e-02 -5.35652757e-01 -2.08206549e-01 4.07142848e-01
4.29139704e-01 2.92296767e-01 -1.71574578e-01 -8.88602436e-02
6.61993265e-01 4.04030859e-01 4.24359053e-01 -3.16090047e-01
8.69441211e-01 -5.83282858e-02 1.08494425e+00 3.39604944e-01
-6.57264888e-01 4.52276766e-01 3.81004572e-01 -8.52487683e-02
-1.04295981e+00 -1.32682979e+00 -1.93774521e-01 1.48774314e+00
4.08230901e-01 2.12461367e-01 -6.00380659e-01 -1.16973186e+00
7.72260204e-02 6.65172696e-01 -5.71261227e-01 -6.01110995e-01
-4.86020327e-01 -7.24898398e-01 2.93960273e-01 5.75471163e-01
5.84674716e-01 -8.45491529e-01 2.39544347e-01 2.09170491e-01
-3.22712302e-01 -7.22744167e-01 -6.95747077e-01 3.70431215e-01
-1.09059227e+00 -7.80300260e-01 -9.33223963e-01 -1.15933800e+00
9.94688869e-01 3.30375761e-01 8.02493870e-01 -5.35325110e-01
2.58707047e-01 8.11623931e-02 -2.44347796e-01 -8.29553828e-02
-4.14887726e-01 2.32448593e-01 3.89075726e-01 2.20550492e-01
8.03228319e-01 -6.56804442e-01 -3.69592130e-01 6.39186919e-01
-9.48648870e-01 -2.06724018e-01 6.90385163e-01 1.09065104e+00
6.66889071e-01 1.08107693e-01 9.81797814e-01 -1.05520725e+00
5.14380991e-01 -8.33233833e-01 -4.38314915e-01 2.91098297e-01
-6.35698974e-01 2.52430826e-01 6.98036671e-01 -8.96286190e-01
-1.40237820e+00 3.41188103e-01 1.97450295e-01 -4.52375650e-01
-4.39863622e-01 2.27952152e-02 -8.18737626e-01 2.26110309e-01
7.72164166e-01 5.03949881e-01 1.03742875e-01 -6.64198220e-01
2.29721770e-01 7.31462002e-01 5.10951579e-01 -7.12439179e-01
1.07065618e+00 5.03082097e-01 -2.95713037e-01 -3.48463684e-01
-9.68680680e-01 -8.52822185e-01 -1.10535252e+00 3.52138221e-01
6.28919542e-01 -1.20140469e+00 3.47101092e-02 5.35861790e-01
-8.49004805e-01 -5.81917353e-03 -3.57903957e-01 5.79461873e-01
-5.24553239e-01 4.27353233e-01 -9.69259366e-02 -5.23535848e-01
-1.08677708e-02 -1.02450073e+00 8.73405755e-01 2.68541634e-01
-2.00922325e-01 -1.11044121e+00 2.79106617e-01 1.59569457e-01
7.13202655e-02 4.96555939e-02 1.14670670e+00 -1.03012788e+00
-1.71138346e-01 -5.28694019e-02 -3.61854225e-01 7.73591101e-01
5.53226888e-01 -5.66082120e-01 -8.72824669e-01 -6.52666926e-01
1.75001174e-01 -3.77115339e-01 8.08078170e-01 1.84639215e-01
7.04621017e-01 -2.26268783e-01 -6.58596098e-01 6.99273169e-01
1.34138536e+00 3.08265239e-01 2.32230499e-01 4.01900768e-01
6.42440736e-01 6.42515123e-01 9.98514295e-01 4.18022603e-01
1.37964606e-01 5.57611465e-01 -9.73999791e-04 -1.04192115e-01
1.04576372e-01 -4.15475518e-01 5.13444424e-01 6.43364966e-01
6.01282954e-01 5.61041795e-02 -6.53289914e-01 7.25777566e-01
-1.87767577e+00 -7.13159561e-01 3.11715424e-01 2.35766792e+00
1.00956035e+00 1.05219796e-01 1.26103804e-01 -1.52685598e-01
1.04337418e+00 -5.05077140e-03 -1.00304973e+00 1.29455915e-02
-1.03835203e-01 -1.13063172e-01 3.53342444e-01 1.94306180e-01
-1.41401196e+00 9.33404446e-01 5.45099640e+00 9.86795843e-01
-1.10756743e+00 4.19618934e-02 2.54310727e-01 2.35286027e-01
-3.03626567e-01 4.04430777e-02 -1.04191875e+00 6.19650006e-01
3.59089851e-01 -1.23647057e-01 1.10939868e-01 1.22749949e+00
-1.21963196e-01 2.60312349e-01 -1.16451716e+00 8.25856805e-01
1.73445255e-01 -5.30169547e-01 1.27660140e-01 5.60453907e-02
9.25291181e-01 -2.23948330e-01 3.11572790e-01 6.40712261e-01
3.11858624e-01 -5.68788946e-01 3.52956653e-01 2.24049777e-01
1.14643300e+00 -7.70884216e-01 5.80534995e-01 6.63036585e-01
-1.08662677e+00 -1.92050427e-01 -6.25061870e-01 1.32459432e-01
-1.62916794e-01 4.14695024e-01 -8.13711107e-01 4.96895820e-01
4.76202548e-01 1.06785083e+00 -6.84903204e-01 7.14087784e-01
-1.14434853e-01 4.67484653e-01 -1.89643860e-01 5.10777950e-01
7.25947171e-02 -3.80532205e-01 6.89388275e-01 9.53562737e-01
2.78780639e-01 -5.87442219e-02 3.87515098e-01 8.27703774e-01
1.10643469e-02 2.06242263e-01 -7.96936691e-01 1.06938280e-01
6.03258133e-01 9.18626249e-01 -4.04375464e-01 -3.52302194e-01
-3.78634542e-01 1.22945428e+00 4.41270351e-01 6.16251945e-01
-7.47877598e-01 -2.57457763e-01 7.08339632e-01 1.63406562e-02
1.50109917e-01 1.01284571e-02 -2.85392970e-01 -1.51980054e+00
4.65732776e-02 -8.99946988e-01 6.69496238e-01 -3.47304791e-01
-1.82882190e+00 3.40262741e-01 5.62963039e-02 -1.75175309e+00
-1.65629342e-01 -3.85975897e-01 -2.72385150e-01 1.12040734e+00
-1.46228337e+00 -1.20280516e+00 -2.53566712e-01 9.53349352e-01
5.61050117e-01 -5.98958075e-01 7.91610897e-01 4.61751670e-01
-5.96829951e-01 8.79308999e-01 8.10114682e-01 5.19039743e-02
1.44021463e+00 -1.09834194e+00 4.39411551e-02 6.29469156e-01
-3.43771845e-01 8.43481123e-01 3.55570018e-01 -7.70453990e-01
-5.38457870e-01 -1.37733185e+00 7.09185958e-01 -4.27632481e-01
2.60223091e-01 -4.61491913e-01 -1.32410574e+00 9.30796742e-01
-1.94599152e-01 -2.21188515e-02 9.41058576e-01 3.72881778e-02
-6.94899261e-01 -3.13844651e-01 -1.32522571e+00 2.90882945e-01
7.97420144e-01 -3.72049809e-01 -8.49827588e-01 1.94717780e-01
5.85375547e-01 -1.25122368e-01 -6.61614358e-01 5.81883252e-01
3.42440575e-01 -7.69723594e-01 1.02896893e+00 -6.07232869e-01
1.17527254e-01 -4.41329628e-01 -1.54427096e-01 -1.41599417e+00
-6.86086118e-01 -6.09609969e-02 2.76840538e-01 1.80855751e+00
1.07395537e-02 -7.77793050e-01 7.81666517e-01 4.39966410e-01
1.50540486e-01 -2.41530553e-01 -9.57287192e-01 -9.62218046e-01
4.06987935e-01 6.48697317e-02 4.94760156e-01 1.27296698e+00
-2.47272581e-01 7.25504339e-01 -4.14123893e-01 1.37256652e-01
8.73904169e-01 1.88528463e-01 8.26979578e-01 -1.53367031e+00
-5.73995477e-03 -4.33185436e-02 -1.34029299e-01 -1.25363958e+00
5.22758305e-01 -1.02381432e+00 8.04753006e-02 -1.06735790e+00
4.76442724e-01 -6.33483350e-01 -7.51520395e-01 5.24394214e-01
-2.67544627e-01 7.13508725e-02 -1.69580519e-01 6.59458578e-01
-6.51265800e-01 8.82089913e-01 1.21713912e+00 -1.28138408e-01
-4.65074807e-01 9.86116659e-03 -8.62321913e-01 7.67342985e-01
7.02640295e-01 -7.80307829e-01 -6.18069053e-01 -2.34695271e-01
-3.77294302e-01 -2.48528436e-01 2.12001845e-01 -9.65484619e-01
-9.49846506e-02 -5.33650279e-01 6.74268007e-01 -4.69613999e-01
1.52103469e-01 -1.02758574e+00 -9.53570530e-02 1.92117766e-01
-4.07310814e-01 -4.33046460e-01 2.10036620e-01 7.98740327e-01
-3.59797299e-01 -3.25370133e-02 1.49096191e+00 5.78121915e-02
-7.67377496e-01 2.46753916e-01 1.64771229e-02 2.33600885e-01
1.22078586e+00 -1.89844102e-01 -1.42389402e-01 8.14375654e-02
-6.40182018e-01 3.88916194e-01 6.92918956e-01 5.79193771e-01
3.50508332e-01 -1.77356994e+00 -6.84098423e-01 6.85581028e-01
5.77100039e-01 -2.46210806e-02 3.64839256e-01 5.54074526e-01
1.31903782e-01 3.15783232e-01 -5.66632271e-01 -8.51755619e-01
-1.15952837e+00 7.77223647e-01 1.65365577e-01 -3.41206461e-01
-4.24772024e-01 8.26034069e-01 1.01539910e+00 -9.24728394e-01
1.56338692e-01 1.55140758e-01 -2.63947874e-01 1.20045960e-01
2.40615100e-01 2.96080917e-01 -3.19028556e-01 -7.19936132e-01
-6.38998806e-01 6.78535223e-01 -3.66128385e-01 5.28276758e-03
1.20163143e+00 -5.60336351e-01 -1.08297355e-02 4.63887364e-01
1.33251345e+00 -1.76158950e-01 -1.61208487e+00 -7.49709249e-01
8.03215206e-02 -4.85672593e-01 -5.37562013e-01 -7.59640872e-01
-5.74261785e-01 1.00651276e+00 9.49771404e-01 -3.49826992e-01
1.35503697e+00 -1.05512269e-01 6.04771018e-01 -4.89751287e-02
1.23692520e-01 -1.34514916e+00 2.46842355e-01 5.66455126e-01
7.35988259e-01 -1.43605483e+00 -1.41002446e-01 -4.03334796e-01
-8.72858942e-01 7.25327134e-01 9.91284907e-01 -2.65896291e-01
5.96011341e-01 -3.02918643e-01 2.03600645e-01 2.56400019e-01
-4.83775258e-01 9.40462500e-02 4.57508504e-01 8.41315508e-01
3.25089365e-01 -7.77155310e-02 -2.42677122e-01 7.88161039e-01
2.96594441e-01 1.05297208e-01 -3.74093428e-02 6.91461384e-01
-4.66366678e-01 -1.51624584e+00 -5.01608908e-01 1.18889160e-01
-2.00931042e-01 2.12241247e-01 -4.00971860e-01 6.35122538e-01
4.27056342e-01 6.10414743e-01 -1.92394271e-01 -2.49725848e-01
4.52375054e-01 5.54785848e-01 1.83341041e-01 -9.82246816e-01
-9.26887318e-02 3.44573468e-01 -5.86934924e-01 -2.47812107e-01
-2.38211721e-01 -6.97768033e-01 -9.93004978e-01 2.71221936e-01
-2.78638631e-01 2.16759637e-01 2.07789719e-01 8.99854362e-01
4.54607427e-01 3.23345482e-01 7.34830737e-01 -3.93968344e-01
-1.00173712e+00 -1.07447791e+00 -9.17996645e-01 9.47968066e-01
4.09721673e-01 -1.11426353e+00 -3.99815828e-01 2.34261364e-01] | [10.363417625427246, 3.1160271167755127] |
64eece2b-ed16-4893-86e4-41f73c71c629 | metoomaastricht-building-a-chatbot-to-assist | 1909.02809 | null | https://arxiv.org/abs/1909.02809v1 | https://arxiv.org/pdf/1909.02809v1.pdf | #MeTooMaastricht: Building a chatbot to assist survivors of sexual harassment | Inspired by the recent social movement of #MeToo, we are building a chatbot to assist survivors of sexual harassment cases (designed for the city of Maastricht but can easily be extended). The motivation behind this work is twofold: properly assist survivors of such events by directing them to appropriate institutions that can offer them help and increase the incident documentation so as to gather more data about harassment cases which are currently under reported. We break down the problem into three data science/machine learning components: harassment type identification (treated as a classification problem), spatio-temporal information extraction (treated as Named Entity Recognition problem) and dialogue with the users (treated as a slot-filling based chatbot). We are able to achieve a success rate of more than 98% for the identification of a harassment-or-not case and around 80% for the specific type harassment identification. Locations and dates are identified with more than 90% accuracy and time occurrences prove more challenging with almost 80%. Finally, initial validation of the chatbot shows great potential for the further development and deployment of such a beneficial for the whole society tool. | ['Gerasimos Spanakis', 'William Lopez Jaramillo', 'Misha Glazunov', 'Tobias Bauer', 'Emre Devrim', 'Balaganesh Mohan'] | 2019-09-06 | null | null | null | null | ['temporal-information-extraction'] | ['natural-language-processing'] | [-3.41675192e-01 4.88149494e-01 -4.40223701e-02 -2.64249563e-01
-8.05678129e-01 -5.18230915e-01 5.59115291e-01 7.15991020e-01
-9.70895231e-01 9.29092824e-01 3.81864935e-01 -3.72859925e-01
-5.07903516e-01 -9.56102014e-01 -8.44305977e-02 -3.68722290e-01
-1.80807307e-01 8.57004285e-01 2.49152958e-01 -3.22262019e-01
5.56229472e-01 7.90834606e-01 -1.13557458e+00 1.38506785e-01
5.68202138e-01 3.79634798e-01 -2.13505216e-02 5.49676657e-01
-1.02645971e-01 9.45360780e-01 -1.25604105e+00 -6.39551640e-01
-2.28390858e-01 -2.70340472e-01 -1.51132321e+00 -4.15326655e-01
-2.74316669e-01 -3.63670945e-01 6.08398244e-02 5.39395392e-01
4.56039757e-01 2.12043494e-01 5.23527741e-01 -1.34218585e+00
1.01311125e-01 5.04090190e-01 -9.43872631e-02 2.86568195e-01
8.15518737e-01 -3.72563899e-01 4.28273559e-01 -4.12561476e-01
1.06456387e+00 1.05107105e+00 9.85255778e-01 3.17808121e-01
-8.09830487e-01 -6.13150299e-01 -6.68372571e-01 5.13610244e-01
-1.62885535e+00 -3.59583974e-01 2.83660531e-01 -3.60689193e-01
1.03268540e+00 4.75850523e-01 5.40640295e-01 1.36227238e+00
-3.71997237e-01 6.53395727e-02 8.08014929e-01 -3.45390916e-01
3.05850506e-01 4.17633563e-01 -4.34236880e-03 2.31808975e-01
1.45729423e-01 -4.80736822e-01 -4.28833276e-01 -7.33931959e-01
5.44321954e-01 2.10668951e-01 4.23344523e-01 4.70142007e-01
-7.09960818e-01 1.13335788e+00 4.19588327e-01 1.01341844e+00
-4.05539811e-01 -1.73780680e-01 8.22126269e-01 2.27082148e-01
4.10423249e-01 6.08884513e-01 -1.32854776e-02 -1.02786517e+00
-8.09865892e-01 4.06618416e-01 1.17895377e+00 6.46066606e-01
4.35732901e-01 -4.45948273e-01 3.93076986e-02 1.07631266e+00
-4.02013734e-02 1.74997926e-01 1.20498016e-01 -7.45003343e-01
5.02013266e-01 9.03087080e-01 2.99595654e-01 -1.31202960e+00
-8.61051142e-01 2.86784898e-02 -6.16737902e-01 -1.67815000e-01
8.10567558e-01 -6.21374965e-01 -2.70945370e-01 1.34803104e+00
4.54090416e-01 3.04605931e-01 -2.27079198e-01 5.76846600e-01
7.06153870e-01 4.08786833e-01 3.36618513e-01 -3.11797589e-01
1.60765386e+00 -1.28774658e-01 -1.05587423e+00 -1.87052175e-01
1.16920030e+00 -7.38565922e-01 3.03767771e-01 1.91885740e-01
-1.00972319e+00 8.92648324e-02 -5.51423013e-01 1.61823690e-01
-8.57120693e-01 -1.62227526e-01 6.37281239e-01 1.10707188e+00
-7.78796554e-01 4.66676623e-01 -6.33934021e-01 -1.35308897e+00
3.13567013e-01 2.17972592e-01 -6.28391922e-01 -7.45534375e-02
-1.28168166e+00 1.35340440e+00 1.01632020e-02 -1.45190194e-01
-1.85470730e-01 -5.20579040e-01 -8.24551702e-01 -1.37056530e-01
3.91355425e-01 -1.61506891e-01 8.51243377e-01 -3.53785396e-01
-7.19356239e-01 1.33306551e+00 8.67314450e-03 -3.99995148e-01
6.41127706e-01 8.69440138e-02 -5.83969891e-01 1.83695689e-01
6.69790745e-01 -4.87721264e-02 -2.02671792e-02 -7.39308417e-01
-7.21220076e-01 -5.78059494e-01 -1.17533483e-01 -1.47038862e-01
-5.90211153e-01 9.65596080e-01 1.96602941e-01 -4.11750495e-01
-2.57246435e-01 -8.15507948e-01 -9.39508229e-02 -4.71275389e-01
-2.44848967e-01 -2.74228692e-01 8.02382231e-01 -1.22258306e+00
1.59057510e+00 -1.94324195e+00 -3.03359002e-01 1.14497460e-01
-1.38861611e-01 4.91539299e-01 2.06194952e-01 1.39478874e+00
1.52751967e-01 2.60370135e-01 1.20054930e-01 -3.17963034e-01
-1.65465921e-01 3.56158555e-01 -5.45724817e-02 5.16483426e-01
1.47086844e-01 4.49052274e-01 -1.10098779e+00 -6.57452583e-01
9.45434049e-02 3.01809758e-01 -1.48503274e-01 1.50488839e-01
7.29613066e-01 1.39163330e-01 -4.89745378e-01 6.21559381e-01
7.41768003e-01 4.11112845e-01 4.38417882e-01 5.48945665e-01
-5.07463634e-01 3.38946790e-01 -1.16074431e+00 1.22712088e+00
-3.20785850e-01 6.20123386e-01 4.09232825e-01 -1.00015843e+00
1.25053275e+00 4.42845166e-01 5.80488563e-01 -4.15566355e-01
2.00408474e-01 2.30635881e-01 -2.69247264e-01 -1.02482235e+00
6.94447100e-01 -2.23966762e-01 -4.34058100e-01 4.43265826e-01
1.02507107e-01 2.58992881e-01 1.84276886e-02 2.65359342e-01
1.52427399e+00 -3.54206026e-01 4.46677625e-01 -1.13950536e-01
1.89067721e-01 1.86858222e-01 4.79559958e-01 7.04182327e-01
-5.96486092e-01 2.39782408e-01 8.60498607e-01 -5.12531698e-01
-9.59951937e-01 -5.49181700e-01 -3.41275483e-01 9.87616420e-01
-1.41575515e-01 -4.25571710e-01 -8.35640550e-01 -7.46255994e-01
-3.28190714e-01 8.07628334e-01 -7.80530512e-01 2.26885658e-02
-4.49952602e-01 -7.87875056e-01 1.10756576e+00 2.53520489e-01
5.37996590e-01 -1.12845314e+00 -4.71761584e-01 4.50455487e-01
-4.29370105e-01 -1.17575276e+00 3.18209887e-01 1.67467315e-02
-3.17435771e-01 -1.13598919e+00 -6.32913470e-01 -6.74479604e-01
2.50037581e-01 -1.98172808e-01 5.70121050e-01 5.53983450e-01
-6.93043053e-01 1.68699056e-01 -7.64698625e-01 -6.27865136e-01
-4.50899661e-01 2.58369356e-01 -8.52619559e-02 -8.38429481e-02
5.01785278e-01 -8.78985643e-01 -3.33541840e-01 4.97770965e-01
-7.37759829e-01 -6.36509120e-01 3.95962549e-03 2.26212770e-01
-6.35488868e-01 -4.08810556e-01 1.09337795e+00 -1.07318377e+00
5.79710424e-01 -1.12326443e+00 2.14215040e-01 -6.43205643e-02
-1.04724891e-01 -8.98225963e-01 7.88283721e-02 -2.81384408e-01
-8.57021272e-01 -2.65959680e-01 -5.51370203e-01 4.19458449e-01
-7.26822853e-01 2.58680969e-01 2.16080025e-01 1.46655098e-01
9.11149681e-01 -2.85796702e-01 2.05715880e-01 -4.83921707e-01
-3.17985207e-01 1.13213623e+00 1.55132115e-01 -2.17513904e-01
5.74279010e-01 4.45403218e-01 5.16568460e-02 -1.30493212e+00
-4.36251879e-01 -1.12818408e+00 -6.94710195e-01 -4.83337909e-01
8.94814730e-01 -4.69950408e-01 -1.07288647e+00 3.62682700e-01
-1.32479393e+00 -1.72648668e-01 1.98690355e-01 2.82478064e-01
-1.26341775e-01 1.12210900e-01 -5.02049804e-01 -1.53275597e+00
9.95886400e-02 -3.92483801e-01 7.89685249e-01 2.19260037e-01
-6.60450876e-01 -1.25099802e+00 3.79155666e-01 7.96901166e-01
5.89573860e-01 6.86097980e-01 6.23071551e-01 -1.27013075e+00
2.69282043e-01 -7.09841251e-01 -3.45612407e-01 -3.30423027e-01
-1.04629077e-01 -9.30993333e-02 -8.66005659e-01 7.22396672e-02
-3.32804024e-01 -1.75224304e-01 1.81746244e-01 -1.21633992e-01
3.70326489e-01 -4.33415860e-01 -5.88718116e-01 -2.80289531e-01
1.05409372e+00 3.73813689e-01 1.09660697e+00 7.33280957e-01
1.58680320e-01 9.65726674e-01 7.89209604e-01 7.55715847e-01
4.68378723e-01 1.04272759e+00 4.90303367e-01 -3.04212701e-02
3.10910523e-01 -1.18500941e-01 -2.48357002e-02 -1.44813105e-01
-4.98459071e-01 5.49622923e-02 -1.19860554e+00 8.38078260e-01
-1.95661378e+00 -1.31220293e+00 -6.09901845e-01 2.06536365e+00
4.26427633e-01 -1.80270538e-01 8.65030229e-01 3.84470403e-01
8.30272913e-01 -4.73265797e-02 3.63200426e-01 -8.76776457e-01
3.29673499e-01 8.85014907e-02 4.16338056e-01 3.35348755e-01
-9.36070323e-01 6.87060058e-01 5.83792019e+00 1.04102540e+00
-5.89584887e-01 2.37783760e-01 4.67187464e-01 3.00911605e-01
3.70122880e-01 1.42422184e-01 -6.59879029e-01 4.15417761e-01
1.13915777e+00 6.95835650e-02 3.07971567e-01 8.22905958e-01
5.22256374e-01 -5.02207041e-01 -3.90963465e-01 6.89507306e-01
1.67962372e-01 -1.51156282e+00 -7.84427822e-01 3.50203335e-01
9.15064737e-02 -2.82445461e-01 -6.71071351e-01 5.82796574e-01
-1.47442250e-02 -8.78276706e-01 3.07511002e-01 6.12650037e-01
4.90516156e-01 -9.05146718e-01 1.10446048e+00 6.78088665e-01
-7.22036839e-01 -2.74606079e-01 -1.96369410e-01 -5.53286076e-01
4.81161326e-01 3.53158474e-01 -1.51899588e+00 6.15797698e-01
7.98964202e-01 2.71240383e-01 -4.95935857e-01 1.39442396e+00
8.35226178e-02 7.93765724e-01 -3.60224932e-01 -4.90916312e-01
2.50614464e-01 4.38410714e-02 6.40609264e-01 1.41448832e+00
3.72279465e-01 1.69232145e-01 -1.50908157e-01 3.58447194e-01
1.95617601e-01 2.75062919e-01 -7.65896797e-01 1.09430142e-01
6.06462657e-01 1.40825844e+00 -8.51628959e-01 1.81167677e-01
2.93097664e-02 8.37779045e-01 4.20959234e-01 -3.92440781e-02
-6.07009172e-01 -7.15596080e-01 5.87676048e-01 6.88836575e-01
-1.30479544e-01 5.25139049e-02 -2.37269908e-01 -5.19946158e-01
-3.48765433e-01 -3.48701954e-01 7.41107285e-01 -4.77686822e-01
-1.02883589e+00 5.44987023e-01 1.78681418e-01 -8.93512666e-01
-5.25219619e-01 -2.39310771e-01 -9.63726342e-01 6.42647982e-01
-6.81822717e-01 -1.42654502e+00 -1.34273127e-01 3.25641125e-01
5.23058288e-02 8.52237865e-02 1.33871663e+00 5.68778098e-01
-5.56816876e-01 4.12890881e-01 -2.99170077e-01 6.04310393e-01
6.81723535e-01 -9.38648105e-01 -1.83782965e-01 3.20497155e-01
-3.87928128e-01 4.29266751e-01 6.37062311e-01 -8.22440147e-01
-8.59790325e-01 -8.86865139e-01 1.80479848e+00 -7.22128153e-01
9.87361073e-01 -6.92204595e-01 -7.54071176e-01 4.65211242e-01
2.84275040e-02 -5.59488058e-01 9.57525849e-01 5.11859655e-01
-1.19808644e-01 3.36477518e-01 -1.80546355e+00 3.16059560e-01
8.92950594e-01 -2.66953409e-01 -4.67248768e-01 7.30833292e-01
1.78971082e-01 -1.53237984e-01 -8.91811311e-01 -1.34644210e-01
3.51841271e-01 -1.03465474e+00 7.38933027e-01 -9.00541365e-01
2.42589593e-01 4.85615164e-01 3.52722436e-01 -7.82998443e-01
8.18036422e-02 -8.26619446e-01 5.70982695e-01 1.84646881e+00
3.92366469e-01 -6.39127076e-01 6.36260331e-01 1.00580072e+00
2.25877374e-01 -2.72996604e-01 -1.48921931e+00 -5.58660626e-01
-6.43960908e-02 -5.85526943e-01 3.26685578e-01 1.51683390e+00
6.72705650e-01 1.48626134e-01 -4.54642355e-01 1.99138150e-01
1.44780099e-01 -5.87033212e-01 6.88243151e-01 -1.12882543e+00
1.99058905e-01 -3.81285064e-02 -8.01614463e-01 2.17750624e-01
-1.11383311e-01 -3.66755128e-01 -3.84709179e-01 -1.52283382e+00
1.95973471e-01 -5.80229759e-01 4.33152884e-01 6.60695612e-01
2.33400479e-01 3.94043505e-01 8.33351836e-02 -1.41791970e-01
-4.71707106e-01 -1.29632980e-01 4.76257056e-01 5.25033474e-01
-2.61820555e-01 4.75663602e-01 -6.11572504e-01 6.72984898e-01
5.77640235e-01 -7.40472972e-01 3.47525328e-01 2.33180374e-01
6.45127416e-01 6.08696878e-01 6.42315686e-01 -9.62232888e-01
2.55043894e-01 -4.43357050e-01 -1.30940452e-01 -2.59691685e-01
3.72588992e-01 -7.54286170e-01 1.96810320e-01 4.39340711e-01
-1.31080791e-01 -2.53455788e-01 1.73320591e-01 2.77741909e-01
-8.60606208e-02 -8.68717253e-01 4.65298802e-01 -5.55650033e-02
-9.94578898e-02 -1.18724823e-01 -9.55935776e-01 -8.32892433e-02
1.38534045e+00 -2.10926130e-01 -5.29169261e-01 -7.56593645e-01
-9.47004259e-01 8.14610049e-02 1.81533799e-01 1.91731289e-01
1.44503772e-01 -9.84390020e-01 -7.75946677e-01 -2.37454578e-01
9.49632972e-02 -7.71623969e-01 5.91323972e-01 1.12984145e+00
-5.85632384e-01 1.20087378e-01 -2.90143430e-01 1.00038186e-01
-1.50235069e+00 2.79066414e-01 1.58324063e-01 -3.22787374e-01
-4.00224984e-01 4.24617261e-01 -8.51215303e-01 -4.89520520e-01
2.57600158e-01 3.53700936e-01 -8.50191295e-01 7.54963577e-01
8.28556120e-01 9.74728346e-01 2.95696199e-01 -8.50996733e-01
-6.98762298e-01 -5.55427521e-02 1.99203163e-01 -2.20366254e-01
1.90293992e+00 6.74447268e-02 -2.75282830e-01 1.62730291e-01
1.12000287e+00 1.24339685e-01 -2.73506880e-01 1.75623566e-01
2.97064573e-01 -5.23189783e-01 -4.90932971e-01 -8.39275837e-01
-3.11030835e-01 6.06276691e-01 1.65597022e-01 1.12237036e+00
6.53538585e-01 1.32539421e-01 6.51734054e-01 2.57619709e-01
4.21436578e-01 -1.08815265e+00 -1.90993235e-01 5.21245539e-01
9.72946167e-01 -9.35793102e-01 -1.84259966e-01 -4.59152222e-01
-7.50259459e-01 1.08574677e+00 1.95996851e-01 -1.65764078e-01
5.09593785e-01 4.63380694e-01 4.45744619e-02 -3.89144689e-01
-2.22697690e-01 -2.98180789e-01 -2.62350380e-01 1.16174364e+00
5.32743074e-02 4.14948501e-02 -9.99074757e-01 9.13112879e-01
-1.23713605e-01 -2.03902826e-01 8.28234136e-01 9.68636274e-01
-4.74429190e-01 -1.12185538e+00 -6.04727447e-01 4.08092856e-01
-8.05197597e-01 3.05581897e-01 -8.93945873e-01 1.19692814e+00
1.91675156e-01 1.20660293e+00 1.42963892e-02 -2.92287081e-01
5.69498479e-01 1.70307219e-01 -3.82506251e-02 -7.03006446e-01
-1.33317935e+00 -2.46204570e-01 9.90217566e-01 -4.17008609e-01
-3.65074039e-01 -9.35752451e-01 -1.16629732e+00 -5.27847469e-01
-1.10141933e-01 5.21527469e-01 1.07952404e+00 1.09406269e+00
1.11586899e-01 -6.54683076e-03 5.92766464e-01 -6.96177244e-01
-1.17821386e-02 -1.23905945e+00 -8.34654510e-01 4.74350750e-01
-9.20818225e-02 -6.11695766e-01 -4.34197783e-01 -4.01256472e-01] | [8.710012435913086, 10.516627311706543] |
64aa8c4a-a919-4da9-bf0c-a12875821973 | multi-hop-question-answering-via-reasoning | 1910.02610 | null | https://arxiv.org/abs/1910.02610v2 | https://arxiv.org/pdf/1910.02610v2.pdf | Multi-hop Question Answering via Reasoning Chains | Multi-hop question answering requires models to gather information from different parts of a text to answer a question. Most current approaches learn to address this task in an end-to-end way with neural networks, without maintaining an explicit representation of the reasoning process. We propose a method to extract a discrete reasoning chain over the text, which consists of a series of sentences leading to the answer. We then feed the extracted chains to a BERT-based QA model to do final answer prediction. Critically, we do not rely on gold annotated chains or "supporting facts:" at training time, we derive pseudogold reasoning chains using heuristics based on named entity recognition and coreference resolution. Nor do we rely on these annotations at test time, as our model learns to extract chains from raw text alone. We test our approach on two recently proposed large multi-hop question answering datasets: WikiHop and HotpotQA, and achieve state-of-art performance on WikiHop and strong performance on HotpotQA. Our analysis shows the properties of chains that are crucial for high performance: in particular, modeling extraction sequentially is important, as is dealing with each candidate sentence in a context-aware way. Furthermore, human evaluation shows that our extracted chains allow humans to give answers with high confidence, indicating that these are a strong intermediate abstraction for this task. | ['Shih-ting Lin', 'Jifan Chen', 'Greg Durrett'] | 2019-10-07 | null | null | null | null | ['multi-hop-question-answering'] | ['knowledge-base'] | [ 8.85324627e-02 7.13636816e-01 -1.25726432e-01 -5.78678370e-01
-1.37158859e+00 -9.16849673e-01 3.91686887e-01 4.83957708e-01
-6.31183207e-01 9.87293363e-01 4.07873183e-01 -6.66089892e-01
-1.97123691e-01 -9.25382257e-01 -9.29964185e-01 7.66876712e-02
2.56840020e-01 1.20706427e+00 8.70085299e-01 -6.26525998e-01
4.77783047e-02 7.80049488e-02 -1.20721281e+00 1.05805862e+00
1.01923084e+00 9.10258830e-01 -9.69472975e-02 9.94873524e-01
-5.20869493e-01 1.60316861e+00 -4.97933686e-01 -1.13642168e+00
-1.28563002e-01 -5.58973312e-01 -1.82731998e+00 -5.75019300e-01
3.78688842e-01 -2.05208763e-01 3.99784856e-02 7.01706469e-01
2.22633496e-01 -1.07970215e-01 5.10319829e-01 -8.75252485e-01
-2.97759324e-01 1.12743342e+00 -4.56889113e-03 1.90417722e-01
8.17646801e-01 1.88920483e-01 1.66180444e+00 -6.19229913e-01
9.44672763e-01 1.05082393e+00 7.61711538e-01 6.28102541e-01
-1.10186315e+00 -2.28350580e-01 8.25714320e-02 4.98633593e-01
-7.28690982e-01 -4.30646598e-01 5.49043775e-01 -1.22005828e-01
1.39444780e+00 3.58457088e-01 3.12928408e-01 9.11825359e-01
4.48301435e-02 1.00896740e+00 9.11992550e-01 -7.20597088e-01
1.50658384e-01 8.67702067e-02 6.72154069e-01 9.18790698e-01
-2.55743694e-02 -4.30229127e-01 -6.15106523e-01 -3.18419456e-01
-5.60216829e-02 -7.00224459e-01 -3.00130963e-01 -1.41031414e-01
-9.89789248e-01 9.60014939e-01 4.09235924e-01 2.59726673e-01
-3.82522017e-01 3.85974757e-02 3.96722704e-01 6.45004928e-01
-1.75469611e-02 8.43454540e-01 -8.13151181e-01 -7.58544207e-02
-8.08104992e-01 4.63013321e-01 1.63040006e+00 7.03957558e-01
8.45604241e-01 -8.88569534e-01 -4.37479436e-01 5.29892087e-01
1.75411791e-01 2.93499440e-01 8.24473500e-02 -1.46537471e+00
1.03881228e+00 8.76843512e-01 3.79226774e-01 -7.29146361e-01
-5.82794070e-01 -2.57082760e-01 -1.98486239e-01 -2.93753266e-01
9.54482496e-01 -2.82483757e-01 -5.20600438e-01 1.79587996e+00
4.35521364e-01 -3.24931294e-01 3.11565548e-01 6.79546535e-01
8.15971851e-01 4.49802667e-01 1.01025194e-01 -9.85127091e-02
1.66537392e+00 -1.15138912e+00 -6.58341706e-01 -5.63868344e-01
1.03294086e+00 -3.96761179e-01 1.04337859e+00 4.59439695e-01
-1.39760888e+00 -1.81933343e-01 -8.34568083e-01 -4.30630654e-01
-1.61116734e-01 -4.90933396e-02 4.72336948e-01 9.58564654e-02
-8.10936630e-01 4.97848690e-01 -4.95849162e-01 -1.79726347e-01
1.93228856e-01 3.19384605e-01 -2.79787987e-01 -2.60461926e-01
-1.58640039e+00 1.16940832e+00 5.34829378e-01 3.77384052e-02
-3.98985952e-01 -4.67963785e-01 -7.26377130e-01 2.35427991e-01
8.00826550e-01 -1.02195573e+00 1.91083038e+00 -7.05468237e-01
-1.33325291e+00 8.08624506e-01 -5.55649519e-01 -7.57242382e-01
5.08507013e-01 -3.81055057e-01 -1.98647916e-01 7.62076497e-01
2.86268502e-01 5.89826226e-01 3.75530720e-01 -1.11099422e+00
-8.40158522e-01 -2.99929261e-01 6.52059793e-01 -4.14690003e-02
-2.51817815e-02 1.24662869e-01 -4.84632701e-01 1.37787089e-01
2.65262816e-02 -7.39753187e-01 -3.14503349e-02 -3.11595857e-01
-4.62503016e-01 -7.19353735e-01 3.06680560e-01 -8.20685029e-01
1.15076661e+00 -1.48625290e+00 4.98491861e-02 1.84203789e-01
3.51333946e-01 2.24480852e-01 -2.03139916e-01 6.78588510e-01
2.55276889e-01 6.99540153e-02 -4.23870742e-01 -2.64487594e-01
2.68261939e-01 3.67432266e-01 -5.62165499e-01 -1.37849882e-01
5.70831299e-01 1.28366315e+00 -9.90416348e-01 -8.30972195e-01
-4.46541578e-01 -1.80864513e-01 -8.64594758e-01 4.02261734e-01
-1.07386672e+00 1.40673354e-01 -5.33279538e-01 2.35587448e-01
2.77353466e-01 -6.05789959e-01 4.17717487e-01 -1.06262200e-01
3.39615047e-01 9.13532019e-01 -9.09129739e-01 1.66657591e+00
-5.60192108e-01 3.31419706e-01 6.02545729e-03 -8.12368810e-01
5.21895349e-01 3.64920795e-01 -9.96127799e-02 -8.95653307e-01
-4.69253398e-02 4.65812653e-01 4.18876857e-02 -8.31793368e-01
3.54710847e-01 -1.56890795e-01 -1.92971662e-01 7.94373572e-01
2.97211528e-01 -1.32533699e-01 6.23473227e-01 5.99473000e-01
1.45938468e+00 1.51938349e-01 1.37219414e-01 3.91513668e-02
7.63658762e-01 5.65699518e-01 4.84330624e-01 9.87960339e-01
6.72623292e-02 3.89010221e-01 9.33079600e-01 -4.59436715e-01
-7.44566262e-01 -8.68977785e-01 3.45063299e-01 1.18475890e+00
-1.23817794e-01 -6.12693071e-01 -8.65802348e-01 -1.34086514e+00
-3.50970179e-01 1.14160955e+00 -5.39171040e-01 2.48301793e-02
-1.11884749e+00 -2.06286564e-01 8.38783324e-01 5.35428762e-01
4.47849184e-01 -1.31498349e+00 -7.29507983e-01 4.74882334e-01
-8.51440549e-01 -1.41008067e+00 -2.36279946e-02 4.04016644e-01
-5.85958302e-01 -1.41684520e+00 -8.74557197e-02 -7.64094532e-01
2.58806556e-01 -3.30755949e-01 1.78743732e+00 4.92565632e-01
2.35607088e-01 3.85586053e-01 -4.38967615e-01 -2.38611966e-01
-5.97871780e-01 4.83245164e-01 -5.62355459e-01 -3.01321357e-01
5.75061798e-01 -4.23715144e-01 -4.07279968e-01 1.22995026e-01
-5.66745222e-01 1.25798080e-02 6.33928657e-01 8.34792197e-01
2.71393985e-01 -2.82148749e-01 6.10009789e-01 -1.25668585e+00
8.25524092e-01 -4.73074168e-01 -3.64487976e-01 7.48685420e-01
-3.14952433e-01 5.48285067e-01 8.38745952e-01 -8.42577778e-03
-1.25768316e+00 -1.01121098e-01 -5.86695552e-01 2.38047391e-01
-3.03310573e-01 7.21651733e-01 -4.81856167e-02 5.19282699e-01
8.86741817e-01 -7.97434747e-02 -1.79604560e-01 -3.24813932e-01
6.04616642e-01 4.30784047e-01 6.07825518e-01 -1.04856431e+00
7.03578413e-01 3.31947058e-01 -1.82523400e-01 -2.74496853e-01
-1.60690284e+00 -4.19506997e-01 -6.49504840e-01 2.50987977e-01
1.02575171e+00 -4.76986587e-01 -1.18570268e+00 -1.77172020e-01
-1.65023553e+00 -5.20731032e-01 -3.43100369e-01 1.85068831e-01
-6.10810697e-01 2.32965946e-01 -9.74587202e-01 -7.64114499e-01
-5.59685528e-01 -7.78711736e-01 7.71944642e-01 1.24195538e-01
-6.40076458e-01 -9.40667808e-01 2.52319455e-01 9.61824775e-01
1.12413749e-01 -1.79972976e-01 1.33375037e+00 -1.25472867e+00
-6.29023492e-01 -1.01920016e-01 -2.74305284e-01 1.67886674e-01
-5.43049932e-01 -3.24512959e-01 -1.01696062e+00 2.38716319e-01
8.32261369e-02 -8.63286138e-01 7.71250844e-01 -2.93455899e-01
6.94277525e-01 -5.22263587e-01 -1.98879242e-01 -6.92499354e-02
1.17678273e+00 -1.39677063e-01 4.81186658e-01 3.86763036e-01
3.23187709e-01 1.11494422e+00 4.54920620e-01 -2.02593595e-01
8.48321855e-01 3.90110105e-01 3.28767210e-01 3.96347880e-01
-9.06971395e-02 -4.88873512e-01 1.03285134e-01 6.65423810e-01
2.50444829e-01 -3.06843072e-01 -1.29439557e+00 6.50174022e-01
-2.06524611e+00 -1.10027003e+00 -5.15663445e-01 1.74660933e+00
1.25869727e+00 5.13777733e-01 -7.37052932e-02 1.05101027e-01
2.19327569e-01 -1.54917344e-01 -2.88714379e-01 -3.69713485e-01
-2.00233236e-02 6.58183694e-01 -5.49227893e-02 8.64764512e-01
-7.50453234e-01 1.11691940e+00 5.91097260e+00 5.24381638e-01
-6.44275606e-01 1.82293758e-01 2.41789728e-01 9.86761004e-02
-6.23221397e-01 3.79527241e-01 -8.68094862e-01 4.20971122e-03
1.23301542e+00 3.16349000e-01 1.65214688e-01 5.16493022e-01
-3.06589425e-01 -2.84107566e-01 -1.34902501e+00 3.01515520e-01
-8.11003298e-02 -1.62845719e+00 5.11943363e-02 -4.70578551e-01
3.07386637e-01 7.26729706e-02 -5.59805751e-01 6.72977030e-01
8.15588892e-01 -1.03969669e+00 7.80079663e-01 6.52185142e-01
8.13815668e-02 -6.76440001e-01 8.91996503e-01 8.01186442e-01
-8.41939688e-01 -3.38788241e-01 -1.57681137e-01 -1.20619297e-01
4.06592757e-01 4.98749465e-01 -1.00154567e+00 8.14845920e-01
4.73624617e-01 4.56604995e-02 -7.84655094e-01 7.10987091e-01
-1.04868233e+00 9.75913286e-01 -4.09507751e-01 -3.26530635e-01
4.15264070e-01 2.44765103e-01 3.27029228e-01 1.19443369e+00
-1.08304463e-01 3.54306549e-01 6.21032305e-02 8.84562612e-01
-3.47374231e-01 -3.74777131e-02 -1.96906969e-01 4.42012809e-02
4.01855320e-01 1.08800364e+00 -4.04458851e-01 -4.96041805e-01
-4.12225455e-01 8.51022124e-01 8.98984790e-01 2.60134637e-01
-6.39607430e-01 -5.18544853e-01 -1.13896374e-02 1.86623354e-02
5.38608491e-01 1.88351553e-02 -1.45899311e-01 -1.19613945e+00
3.70877594e-01 -1.14723837e+00 9.33491707e-01 -8.92967105e-01
-1.30383563e+00 7.00587273e-01 -1.79078907e-01 -5.49216390e-01
-6.27744973e-01 -4.83702809e-01 -7.24389315e-01 8.33921134e-01
-1.74441659e+00 -1.22517014e+00 1.00254938e-02 6.10079706e-01
3.08458090e-01 3.24485540e-01 8.33842874e-01 1.10107578e-01
-1.72729358e-01 5.20167768e-01 -6.17098987e-01 5.27046204e-01
7.13490307e-01 -1.40501022e+00 2.96714067e-01 8.33690286e-01
5.37992299e-01 8.86629760e-01 7.00471580e-01 -5.28592408e-01
-1.32834184e+00 -6.75246119e-01 1.60380304e+00 -1.09841406e+00
7.81173289e-01 -1.88333109e-01 -1.21294987e+00 8.99324119e-01
5.77479422e-01 -2.17399940e-01 6.36660516e-01 5.94073117e-01
-7.01149046e-01 -2.64338423e-02 -8.60734940e-01 4.35341865e-01
9.86844301e-01 -7.77064800e-01 -1.47868311e+00 4.15511101e-01
1.07686532e+00 -4.26020861e-01 -6.87159836e-01 3.22898060e-01
2.84614325e-01 -1.04117417e+00 6.52649701e-01 -1.05397606e+00
7.46236145e-01 -2.52951562e-01 -7.51589164e-02 -9.37944710e-01
9.35074687e-02 -5.35882652e-01 -4.65329409e-01 1.03567684e+00
1.08146167e+00 -4.64270949e-01 8.00093830e-01 8.98464799e-01
-8.83164257e-02 -8.63903999e-01 -8.42767000e-01 -3.10823172e-01
1.99270755e-01 -6.35825157e-01 5.18634737e-01 7.18665183e-01
2.40030080e-01 1.08408296e+00 1.28078997e-01 3.32787424e-01
5.24942100e-01 4.88806546e-01 7.07254648e-01 -1.27798569e+00
-6.04933739e-01 -1.41250476e-01 4.79478568e-01 -1.18630052e+00
4.51234162e-01 -8.94260347e-01 1.95875317e-01 -1.97421849e+00
5.19101806e-02 -3.90429646e-01 -2.47437079e-02 6.19853079e-01
-3.10688227e-01 -2.35193744e-02 5.84317334e-02 1.73596129e-01
-1.10457253e+00 1.86804116e-01 1.02644610e+00 -1.43504605e-01
3.85327004e-02 -4.48715128e-02 -8.29848886e-01 8.87665331e-01
5.78152061e-01 -5.76987028e-01 -4.24058288e-01 -7.20256448e-01
9.73506808e-01 4.84646499e-01 3.18133146e-01 -8.02297592e-01
7.29723036e-01 -4.51796763e-02 6.43997416e-02 -4.15694684e-01
3.40928316e-01 -6.98103607e-01 -4.10776854e-01 5.12360394e-01
-6.54625952e-01 -2.73764040e-02 -7.53576383e-02 3.81771803e-01
-3.19377095e-01 -6.30648911e-01 4.00161803e-01 -3.68248403e-01
-5.33610165e-01 -1.40122652e-01 -1.68623835e-01 6.36012912e-01
6.34662390e-01 2.34445766e-01 -5.69111705e-01 -5.00184774e-01
-8.38848174e-01 6.51844144e-01 -3.32738571e-02 2.06948072e-02
3.41401726e-01 -7.63231516e-01 -7.67252505e-01 -3.73331726e-01
1.39408514e-01 1.87584043e-01 -3.11718546e-02 8.69180560e-01
-4.89480257e-01 5.42189658e-01 1.61448240e-01 -3.42286736e-01
-9.86207306e-01 4.41484392e-01 4.04952675e-01 -9.44040656e-01
-4.16360676e-01 8.93796027e-01 -3.65539402e-01 -7.59483397e-01
9.59237590e-02 -4.94505256e-01 -4.54909593e-01 3.52346569e-01
5.72848737e-01 -1.97292820e-01 4.09634173e-01 2.52397042e-02
-3.49878848e-01 2.22452030e-01 -2.19375625e-01 -4.06849086e-01
1.27831817e+00 2.22267546e-02 -3.75223577e-01 2.57462233e-01
8.24357867e-01 2.46023417e-01 -7.52485156e-01 -3.62194270e-01
5.98928750e-01 2.19842233e-02 -5.19647181e-01 -1.22230971e+00
-3.61461878e-01 9.39810276e-01 -3.86078745e-01 3.76197308e-01
6.86983585e-01 3.91675562e-01 1.17075682e+00 1.17182493e+00
4.11621481e-01 -1.03459954e+00 2.92087466e-01 9.85209823e-01
7.75606453e-01 -1.14689422e+00 -4.06761646e-01 -2.61409789e-01
-6.14402711e-01 1.05500019e+00 7.28331327e-01 1.79721624e-01
1.20314799e-01 1.57871842e-01 1.33155778e-01 -3.69627446e-01
-1.31829929e+00 -3.18672359e-01 2.14392096e-01 2.24597216e-01
3.33086491e-01 -3.71879071e-01 -2.66507417e-01 8.13646555e-01
-4.05163527e-01 -2.13899221e-02 3.45642328e-01 1.06674981e+00
-7.09363103e-01 -1.23450685e+00 -1.73868701e-01 3.84537041e-01
-4.93769318e-01 -2.13604376e-01 -7.64199734e-01 6.60390556e-01
-9.20812599e-03 1.30656004e+00 -3.51901084e-01 -8.73646736e-02
4.54677463e-01 6.77749038e-01 5.45748770e-01 -6.92514002e-01
-9.70063746e-01 -6.75805271e-01 8.80276978e-01 -6.13958120e-01
-3.04439753e-01 -2.79215425e-01 -1.57576060e+00 -1.81228340e-01
-2.97283173e-01 9.41815615e-01 2.74359822e-01 1.44133008e+00
3.57219160e-01 3.12919110e-01 7.67313391e-02 6.96033314e-02
-7.77813971e-01 -8.46480966e-01 2.45006815e-01 5.18876374e-01
2.83374578e-01 -1.37228429e-01 -1.81326807e-01 -8.50490406e-02] | [10.91812515258789, 7.9351983070373535] |
ecd6df1f-c7a8-46ef-ad07-cb4217c66121 | wildtrack-a-multi-camera-hd-dataset-for-dense | null | null | http://openaccess.thecvf.com/content_cvpr_2018/html/Chavdarova_WILDTRACK_A_Multi-Camera_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Chavdarova_WILDTRACK_A_Multi-Camera_CVPR_2018_paper.pdf | WILDTRACK: A Multi-Camera HD Dataset for Dense Unscripted Pedestrian Detection | People detection methods are highly sensitive to occlusions between pedestrians, which are extremely frequent in many situations where cameras have to be mounted at a limited height. The reduction of camera prices allows for the generalization of static multi-camera set-ups. Using joint visual information from multiple synchronized cameras gives the opportunity to improve detection performance. In this paper, we present a new large-scale and high-resolution dataset. It has been captured with seven static cameras in a public open area, and unscripted dense groups of pedestrians standing and walking. Together with the camera frames, we provide an accurate joint (extrinsic and intrinsic) calibration, as well as 7 series of 400 annotated frames for detection at a rate of 2 frames per second. This results in over 40,000 bounding boxes delimiting every person present in the area of interest, for a total of more than 300 individuals. We provide a series of benchmark results using baseline algorithms published over the recent months for multi-view detection with deep neural networks, and trajectory estimation using a non-Markovian model. | ['François Fleuret', 'Luc van Gool', 'Stéphane Bouquet', 'Pierre Baqué', 'Pascal Fua', 'Louis Lettry', 'Cijo Jose', 'Tatjana Chavdarova', 'Timur Bagautdinov', 'Andrii Maksai'] | 2018-06-01 | null | null | null | cvpr-2018-6 | ['multiview-detection'] | ['computer-vision'] | [-7.07109496e-02 -4.97757107e-01 2.58465648e-01 -2.72033632e-01
-5.10641217e-01 -6.64514422e-01 5.32508731e-01 -1.35851488e-01
-9.36500609e-01 6.93339765e-01 5.00838794e-02 4.38627511e-01
6.25023842e-01 -7.38309145e-01 -7.42063344e-01 -5.74772477e-01
-2.16460899e-01 6.06969059e-01 7.25666225e-01 9.07419398e-02
-3.37228179e-01 4.93257821e-01 -1.62002063e+00 2.88565844e-01
-1.47853891e-04 6.02478921e-01 1.37713989e-02 1.16900575e+00
6.80666208e-01 3.69391710e-01 -7.33923495e-01 -1.00773156e+00
4.84089434e-01 2.78076231e-01 -2.11483270e-01 5.22867918e-01
1.13946521e+00 -9.56482232e-01 -5.85996568e-01 8.24399233e-01
6.38401747e-01 8.45678672e-02 3.99089336e-01 -1.24683559e+00
-8.38715211e-02 -2.40206253e-02 -9.89047170e-01 5.79694152e-01
6.89126849e-01 5.06465554e-01 7.07341671e-01 -7.29508996e-01
4.51572299e-01 1.63459599e+00 1.02485549e+00 6.06524825e-01
-1.07895279e+00 -5.72386682e-01 1.65587991e-01 2.75436133e-01
-1.65093315e+00 -5.83317876e-01 2.76471823e-01 -5.01903892e-01
9.79565263e-01 1.09567322e-01 1.15324271e+00 1.59925795e+00
-1.05832368e-01 6.20586038e-01 6.72765851e-01 -3.04297686e-01
-1.32022128e-01 2.89357215e-01 1.37806728e-01 5.83980024e-01
8.63118649e-01 2.59522527e-01 -3.47432613e-01 -2.16590986e-01
7.31265187e-01 4.71364796e-01 2.98541393e-02 -4.69803393e-01
-1.20170856e+00 9.23837483e-01 1.81074306e-01 -2.10312665e-01
-1.30105436e-01 1.77422240e-01 5.61251521e-01 -3.49183202e-01
2.09096566e-01 -3.99372190e-01 -1.74574241e-01 -5.19499276e-03
-9.84161377e-01 4.91776437e-01 6.36944175e-01 1.17183542e+00
4.94789511e-01 -9.73630771e-02 -1.31077468e-01 4.46624994e-01
2.52184600e-01 1.02703178e+00 -9.02653635e-02 -1.10817719e+00
7.70172000e-01 5.99483490e-01 5.57869196e-01 -1.17367613e+00
-4.29779619e-01 -1.17893785e-01 -9.37589586e-01 1.06705971e-01
9.42827702e-01 -5.66224456e-01 -3.94969463e-01 1.47548854e+00
4.76040304e-01 1.57479897e-01 -2.28931829e-01 1.03957486e+00
5.37291110e-01 4.63231593e-01 -1.43697113e-01 -1.37601318e-02
1.84764588e+00 -9.33121622e-01 -3.06980044e-01 -5.67434192e-01
1.42400861e-01 -5.88763833e-01 4.60136026e-01 4.18042362e-01
-9.24090564e-01 -6.53577924e-01 -8.19663525e-01 5.08501641e-02
-4.29664642e-01 4.20031607e-01 3.70213777e-01 1.15904403e+00
-9.99628842e-01 1.44789830e-01 -7.81054258e-01 -8.77536237e-01
3.78448695e-01 3.36151689e-01 -5.51307976e-01 -1.04002558e-01
-9.93767202e-01 8.57566893e-01 1.74497381e-01 9.91815105e-02
-1.05944324e+00 -3.64862770e-01 -7.94904470e-01 -9.90786850e-02
3.30234975e-01 -8.03041339e-01 9.61733699e-01 -6.78270698e-01
-9.31293249e-01 1.13457239e+00 -2.15761155e-01 -8.04582536e-01
1.13952196e+00 -5.86933315e-01 -4.38471615e-01 2.68429548e-01
3.54023457e-01 7.59156704e-01 6.75567329e-01 -8.29430580e-01
-1.12713921e+00 -5.71392298e-01 3.07488814e-02 3.03865999e-01
-4.76582229e-01 5.51193357e-01 -9.07898784e-01 -1.87356606e-01
-6.73020482e-01 -1.03448880e+00 -3.16879421e-01 1.12448715e-01
-4.94300812e-01 1.87300250e-01 8.40551019e-01 -8.08865845e-01
7.93287277e-01 -1.90156424e+00 7.05570579e-02 -2.47146145e-01
2.11234331e-01 3.26334357e-01 2.82844096e-01 9.24221501e-02
1.55734047e-01 -2.31522813e-01 1.68641239e-01 -7.78930426e-01
-5.75745068e-02 7.96723142e-02 4.00073491e-02 9.09808934e-01
-1.52701557e-01 5.56558430e-01 -6.88251436e-01 -6.69138372e-01
6.35479748e-01 6.40706241e-01 -3.89484257e-01 2.10842296e-01
4.59807754e-01 2.95573771e-01 7.11028231e-03 6.99512541e-01
7.42060065e-01 -8.26189145e-02 -7.60665536e-02 -1.37443274e-01
-1.77473843e-01 -5.53804040e-01 -1.67643428e+00 1.30858099e+00
7.47010857e-02 9.65419114e-01 1.01346523e-01 -4.38222587e-01
4.25499499e-01 2.29002386e-01 4.22756404e-01 -7.88329169e-02
1.00721046e-01 -3.35226715e-01 -3.65221173e-01 -1.75023958e-01
5.57354748e-01 4.50765789e-01 -4.32765968e-02 1.02594465e-01
-1.77705530e-02 7.29344547e-01 8.00204635e-01 2.56948233e-01
9.84498203e-01 -1.46503672e-01 3.92577976e-01 1.02903411e-01
5.36911905e-01 -1.24422438e-01 4.87169057e-01 9.07366097e-01
-7.48497665e-01 7.42655694e-01 1.45038813e-01 -1.05855441e+00
-1.47132456e+00 -1.21712577e+00 2.09284779e-02 8.81827295e-01
1.34541005e-01 -3.60324621e-01 -9.87182021e-01 -2.16141567e-01
-1.11913733e-01 5.24133742e-02 -5.18079162e-01 1.77674860e-01
-8.53932202e-01 -1.07719147e+00 7.48110890e-01 7.36817658e-01
9.58103418e-01 -3.95452112e-01 -1.12326014e+00 7.35227112e-03
-3.99080455e-01 -1.71972001e+00 -6.28871322e-01 -3.25685620e-01
-3.67541283e-01 -1.36963773e+00 -1.21626663e+00 -3.65119338e-01
4.12796021e-01 7.37179160e-01 1.27008712e+00 -2.24562928e-01
-6.93373144e-01 5.83240569e-01 8.22865441e-02 -2.53384292e-01
6.53889626e-02 -2.15649590e-01 5.23853838e-01 1.98365152e-01
8.78083825e-01 -1.23165630e-01 -7.61982322e-01 5.14847219e-01
-8.31424594e-02 -5.46170697e-02 3.15578580e-01 3.35857540e-01
2.81914502e-01 -5.84408417e-02 -2.66060144e-01 -3.24350536e-01
-3.68755125e-02 -1.46183729e-01 -1.08530605e+00 1.83990940e-01
1.64283454e-01 -6.89525545e-01 3.11136723e-01 -5.43233633e-01
-8.79227757e-01 7.01883733e-01 2.54485220e-01 -4.91273880e-01
-6.44337773e-01 -7.34637260e-01 -1.46528542e-01 -7.88214523e-03
7.70010710e-01 7.83590227e-03 -1.91464439e-01 -2.54267186e-01
4.18028504e-01 4.89898235e-01 6.47004008e-01 -1.94411457e-01
7.82034278e-01 9.68609810e-01 -3.98476534e-02 -1.08018196e+00
-6.74840152e-01 -8.76266539e-01 -1.21356821e+00 -5.37610292e-01
1.16375017e+00 -1.55135751e+00 -9.83473361e-01 5.83314300e-01
-1.42689240e+00 9.32403058e-02 5.82072735e-02 4.45854723e-01
-2.34427065e-01 6.51113033e-01 -6.52357817e-01 -1.09193850e+00
-1.89728469e-01 -1.09834635e+00 1.35934794e+00 2.83792317e-01
-5.48203401e-02 -8.57266426e-01 -5.56011498e-02 4.44658995e-01
1.19916517e-02 5.99898219e-01 -4.98458892e-01 -2.96038151e-01
-6.91745818e-01 -6.29108369e-01 -4.00419980e-01 8.13098848e-02
-2.49199092e-01 1.92428306e-01 -9.99734879e-01 -7.08729327e-01
-3.62925082e-01 -2.51874700e-03 9.13730383e-01 8.43222022e-01
4.46458042e-01 -1.40025079e-01 -7.95907974e-01 5.41681886e-01
1.27243614e+00 -1.66029930e-01 3.13308686e-01 4.23079252e-01
8.79596651e-01 5.80896199e-01 2.69931644e-01 7.17192471e-01
4.29940939e-01 1.13581657e+00 3.97692800e-01 -1.36929631e-01
-3.94479260e-02 7.71121010e-02 6.14843607e-01 -7.95817673e-02
-5.46594143e-01 -2.33017251e-01 -9.19332802e-01 5.17515779e-01
-1.79719126e+00 -1.44772434e+00 -5.18050075e-01 2.50832105e+00
1.70467362e-01 1.21469505e-01 1.00503528e+00 -1.44286707e-01
1.43934870e+00 7.58822402e-03 -2.82542080e-01 4.03961509e-01
-1.36505663e-01 -6.74379289e-01 1.11753035e+00 4.09566969e-01
-1.77377856e+00 6.13781035e-01 6.81330490e+00 3.83468747e-01
-3.75950903e-01 2.26791576e-01 4.46530581e-01 -5.62142253e-01
9.59280252e-01 -3.14872861e-01 -1.67714238e+00 6.21956468e-01
1.12654912e+00 1.34245366e-01 2.20036566e-01 1.03804493e+00
1.88440725e-01 -2.90958405e-01 -1.06222522e+00 1.48477936e+00
3.40272695e-01 -1.14281642e+00 -2.55894482e-01 3.88570875e-01
5.01168668e-01 9.36365649e-02 -2.42819414e-02 -2.89276592e-03
3.70962530e-01 -6.22877836e-01 8.22569251e-01 2.38013908e-01
6.19971812e-01 -9.17584896e-01 8.66075218e-01 4.52322364e-01
-1.61995995e+00 -2.96467215e-01 -7.34015346e-01 -1.46463171e-01
5.51236331e-01 3.23770434e-01 -5.83442628e-01 1.98854119e-01
1.25451612e+00 8.73330474e-01 -8.42584252e-01 1.06654334e+00
2.17271242e-02 1.16286010e-01 -6.92805290e-01 3.13489288e-02
-1.91362761e-02 -9.85275060e-02 6.83461607e-01 1.63762164e+00
2.70570397e-01 -1.95653796e-01 3.60349000e-01 5.23885906e-01
1.55314460e-01 -3.50046486e-01 -7.97672212e-01 7.10672915e-01
2.87539601e-01 1.45667613e+00 -7.75982976e-01 -5.11676371e-01
-8.64539683e-01 1.21262646e+00 7.53819421e-02 2.37049505e-01
-1.26851344e+00 3.70456949e-02 7.62811303e-01 4.88398641e-01
3.69160771e-01 -3.99870694e-01 4.21960533e-01 -1.44392407e+00
1.65286869e-01 -5.50627053e-01 5.95974028e-01 -3.96154970e-01
-1.21069396e+00 3.41149747e-01 3.47270668e-01 -1.28923273e+00
-2.79871553e-01 -8.39973152e-01 -4.16045725e-01 6.83632910e-01
-1.08569908e+00 -1.38232422e+00 -5.68843663e-01 8.20235789e-01
6.50514126e-01 -4.81372356e-01 6.25233352e-01 8.23311150e-01
-9.29727972e-01 3.56845737e-01 3.56610212e-03 6.80100262e-01
6.83320463e-01 -1.31875789e+00 7.65937209e-01 1.39983809e+00
1.05543490e-02 3.48040700e-01 7.96611011e-01 -4.85435247e-01
-1.10775173e+00 -1.32899964e+00 5.03531456e-01 -1.19128025e+00
3.63710701e-01 -7.29373217e-01 -3.29986602e-01 9.05849218e-01
8.50135460e-02 2.90923893e-01 5.92116594e-01 -8.87644216e-02
-5.89413941e-02 -1.77251190e-01 -1.02162778e+00 4.32554722e-01
9.60821390e-01 -1.40858144e-01 -3.47133011e-01 6.57002509e-01
2.93503761e-01 -5.93586743e-01 -4.38039899e-01 -2.66568840e-01
7.40577042e-01 -1.29602110e+00 1.63160026e+00 -3.19485426e-01
-1.81055263e-01 -3.11676800e-01 -3.01926553e-01 -6.29783094e-01
-3.49325418e-01 -4.18022901e-01 -3.37974906e-01 1.19280434e+00
-9.11508277e-02 -3.06910604e-01 9.55115736e-01 7.61687636e-01
4.18547660e-01 4.36612517e-02 -1.09154022e+00 -8.64901304e-01
-3.39116395e-01 -8.67350399e-02 1.59893990e-01 2.08311722e-01
-6.47490442e-01 2.44230151e-01 -9.31755483e-01 6.00964665e-01
1.28347850e+00 -3.33414704e-01 1.24980891e+00 -1.29894042e+00
-2.66387880e-01 -1.02829590e-01 -7.49491513e-01 -1.03111434e+00
-2.05388412e-01 -9.80362296e-02 -3.15187454e-01 -1.12417173e+00
7.96916783e-01 4.60075706e-01 3.88285875e-01 -1.41240293e-02
-9.75274853e-03 5.17349839e-01 1.40120700e-01 2.44661897e-01
-9.49204862e-01 -4.58195545e-02 4.63403821e-01 -1.08897358e-01
1.21481113e-01 2.97754258e-01 -1.66088834e-01 1.17622006e+00
3.83857012e-01 -3.21643353e-01 1.12294264e-01 -4.32853341e-01
-9.37180594e-02 -2.93861181e-01 1.01260209e+00 -1.49809420e+00
3.93455118e-01 1.49397567e-01 1.11540568e+00 -8.43414605e-01
9.53256607e-01 -8.83483768e-01 3.41377705e-01 8.18615854e-01
1.11300625e-01 3.74429315e-01 4.97167930e-02 9.09319341e-01
1.04605846e-01 6.32766038e-02 1.14966130e+00 -6.55044496e-01
-9.43433344e-01 3.79961669e-01 -3.01175982e-01 -1.25831757e-02
1.47413766e+00 -4.43301916e-01 -2.88184583e-01 -2.45901212e-01
-7.70179451e-01 9.68324542e-02 5.20206928e-01 4.21286553e-01
4.88346756e-01 -1.22537899e+00 -8.00404549e-01 1.73144490e-01
-2.19395440e-02 -2.42272109e-01 3.27316791e-01 6.16447389e-01
-6.20941877e-01 5.96340120e-01 -3.59819680e-01 -1.03377426e+00
-1.77947652e+00 6.96683586e-01 5.35301328e-01 -8.45078602e-02
-7.90334702e-01 5.58231413e-01 5.26739247e-02 8.12739432e-02
2.89587021e-01 -4.71949466e-02 -2.83893526e-01 1.72281563e-01
1.10472465e+00 8.34528863e-01 -3.66091132e-01 -1.32953775e+00
-5.53671956e-01 7.28981793e-01 -7.46449968e-03 -2.24045649e-01
9.30975616e-01 -4.99247551e-01 3.08999360e-01 1.62300169e-01
9.20926988e-01 -1.57647759e-01 -1.75835359e+00 -6.37328103e-02
-3.11600894e-01 -7.95095444e-01 -2.18731478e-01 -8.47938880e-02
-9.47315574e-01 1.03970158e+00 1.15288520e+00 1.27216503e-01
4.99213278e-01 -9.43634138e-02 7.18983948e-01 4.34873939e-01
5.88343620e-01 -1.01253974e+00 7.34722838e-02 2.28359222e-01
3.49624455e-01 -1.55814433e+00 2.06308857e-01 -3.55135828e-01
-4.69751686e-01 1.06749153e+00 7.26629436e-01 -1.12505443e-01
1.10499166e-01 2.32076913e-01 -2.69186288e-01 1.39124215e-01
-3.87580991e-01 -3.87558609e-01 -1.09401487e-01 1.02499568e+00
-6.76193610e-02 1.42281711e-01 5.87595940e-01 2.05516443e-01
1.45005956e-01 -4.47508425e-01 6.31752312e-01 6.54251158e-01
-5.71179807e-01 -7.15919793e-01 -1.02595055e+00 1.44880816e-01
-4.04337108e-01 2.36565232e-01 -2.60608941e-01 9.35623884e-01
2.64338732e-01 1.09916270e+00 1.53457716e-01 -1.27911763e-02
5.01063824e-01 -2.31163338e-01 4.23871845e-01 -4.14599478e-01
-3.01720977e-01 -2.43911818e-02 4.87868935e-02 -6.32248580e-01
-6.14218712e-01 -1.19811809e+00 -5.06193697e-01 -7.67704308e-01
-1.67192608e-01 -4.71353710e-01 2.79320210e-01 6.93293273e-01
5.97875118e-02 1.06231652e-01 1.89897746e-01 -1.47473598e+00
-3.22702438e-01 -7.92011559e-01 -4.55858678e-01 4.70870346e-01
4.65753317e-01 -8.15712750e-01 -3.40261549e-01 4.78554219e-01] | [7.373427867889404, -1.0006755590438843] |
b2447fbc-382c-4647-bfd3-a29a6db9f95b | fusing-wearable-imus-with-multi-view-images | 2003.11163 | null | https://arxiv.org/abs/2003.11163v2 | https://arxiv.org/pdf/2003.11163v2.pdf | Fusing Wearable IMUs with Multi-View Images for Human Pose Estimation: A Geometric Approach | We propose to estimate 3D human pose from multi-view images and a few IMUs attached at person's limbs. It operates by firstly detecting 2D poses from the two signals, and then lifting them to the 3D space. We present a geometric approach to reinforce the visual features of each pair of joints based on the IMUs. This notably improves 2D pose estimation accuracy especially when one joint is occluded. We call this approach Orientation Regularized Network (ORN). Then we lift the multi-view 2D poses to the 3D space by an Orientation Regularized Pictorial Structure Model (ORPSM) which jointly minimizes the projection error between the 3D and 2D poses, along with the discrepancy between the 3D pose and IMU orientations. The simple two-step approach reduces the error of the state-of-the-art by a large margin on a public dataset. Our code will be released at https://github.com/CHUNYUWANG/imu-human-pose-pytorch. | ['Wen-Jun Zeng', 'Zhe Zhang', 'Wenhu Qin', 'Chunyu Wang'] | 2020-03-25 | fusing-wearable-imus-with-multi-view-images-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Zhang_Fusing_Wearable_IMUs_With_Multi-View_Images_for_Human_Pose_Estimation_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Zhang_Fusing_Wearable_IMUs_With_Multi-View_Images_for_Human_Pose_Estimation_CVPR_2020_paper.pdf | cvpr-2020-6 | ['3d-absolute-human-pose-estimation'] | ['computer-vision'] | [-2.47778609e-01 1.82972610e-01 -2.22436979e-01 -1.93006113e-01
-4.97143984e-01 -4.81764108e-01 5.47774732e-01 -6.68940425e-01
-3.68386000e-01 4.08790976e-01 6.23773634e-01 3.47732127e-01
1.67969048e-01 -4.32724744e-01 -7.32242465e-01 -1.89447135e-01
8.33074003e-03 5.95045090e-01 -1.66488394e-01 -1.18138991e-01
-7.66047984e-02 4.43878323e-01 -1.07064962e+00 -1.62873596e-01
1.81700528e-01 8.01548243e-01 -1.64297536e-01 6.00940585e-01
5.13400793e-01 1.53459506e-02 -2.61997938e-01 -4.07927752e-01
6.64399087e-01 -2.40747795e-01 -5.02504349e-01 3.96138668e-01
6.99719548e-01 -4.80413437e-01 -5.01310766e-01 9.06145453e-01
7.09760427e-01 -8.70531723e-02 4.52414095e-01 -1.21352506e+00
-3.58751655e-01 -1.02400005e-01 -1.04186893e+00 -3.28690141e-01
9.62599814e-01 -5.01408167e-02 8.30870748e-01 -1.19336331e+00
1.03281772e+00 1.52266467e+00 8.61770153e-01 6.75932407e-01
-1.04670858e+00 -3.93366009e-01 1.40701681e-01 -5.81578091e-02
-1.51463509e+00 -7.92257339e-02 1.05524266e+00 -6.21742129e-01
6.00290477e-01 1.61377952e-01 9.92845058e-01 1.33690131e+00
1.38204306e-01 8.32440674e-01 9.51269627e-01 -4.90962386e-01
-2.64101207e-01 -2.99221516e-01 1.72238238e-03 9.60398674e-01
2.56564945e-01 1.49523184e-01 -6.75320208e-01 -1.19983122e-01
1.36922693e+00 5.28741777e-01 -2.15574875e-01 -7.03660965e-01
-1.47623289e+00 5.63373923e-01 5.10572433e-01 -1.48248702e-01
-4.75703329e-01 2.47053102e-01 1.17018551e-03 -8.32525268e-02
4.79050070e-01 -8.23114067e-04 -4.64095354e-01 -6.13422412e-03
-2.51192361e-01 4.55457360e-01 3.54152799e-01 1.00614560e+00
5.64192593e-01 -3.09506059e-01 2.33965173e-01 7.76387513e-01
5.69419920e-01 6.06767178e-01 1.17587574e-01 -1.19283438e+00
6.64138913e-01 6.92032576e-01 1.96237087e-01 -1.05277085e+00
-7.61936188e-01 -3.48593205e-01 -8.31191659e-01 5.03591485e-02
6.28034651e-01 -4.45640057e-01 -8.08233202e-01 1.75168741e+00
6.87479436e-01 -2.47745246e-01 -4.51607317e-01 1.39715099e+00
5.03523588e-01 2.35986844e-01 -3.96896482e-01 3.80874842e-01
1.63590229e+00 -9.96099353e-01 -3.98976356e-01 -5.36779165e-01
1.97412297e-01 -6.28810108e-01 8.54932606e-01 2.98780143e-01
-1.18736696e+00 -7.37223625e-01 -1.03778100e+00 -1.31475806e-01
2.61063650e-02 5.06831706e-01 2.88175642e-01 4.34864193e-01
-6.78026915e-01 5.88530779e-01 -9.71761048e-01 -5.04148543e-01
-1.83877379e-01 3.18041235e-01 -8.01435888e-01 1.15098760e-01
-1.16300213e+00 9.52300966e-01 1.63764566e-01 4.10412639e-01
-3.74897152e-01 -2.98327416e-01 -1.03368032e+00 -5.76160491e-01
6.95314348e-01 -1.14560425e+00 1.05971456e+00 -5.10873973e-01
-1.52474678e+00 1.05385661e+00 -2.08486632e-01 2.88825780e-02
9.95337188e-01 -9.18343365e-01 2.92803608e-02 2.15256885e-01
2.15593707e-02 9.02696073e-01 8.74745071e-01 -1.37196589e+00
-2.61011511e-01 -9.05763924e-01 2.08756421e-02 6.53059661e-01
1.00322403e-01 -4.17186946e-01 -1.01903939e+00 -7.72039771e-01
7.64800608e-01 -1.30573487e+00 -2.42953151e-01 5.18929243e-01
-8.02460015e-01 -1.67661697e-01 3.99713844e-01 -9.24674928e-01
7.01115906e-01 -1.90220761e+00 6.79338872e-01 4.23365831e-01
3.17799777e-01 -2.59324551e-01 6.96415380e-02 2.43616968e-01
-6.71920329e-02 -2.61348814e-01 1.60090774e-01 -5.36535203e-01
-1.91950682e-03 1.30131230e-01 2.61314183e-01 9.09218013e-01
-1.00831851e-01 8.60476196e-01 -7.06377864e-01 -3.31956863e-01
2.25406900e-01 7.10883260e-01 -5.11274874e-01 4.06138092e-01
1.88430563e-01 7.40122259e-01 -3.07386607e-01 5.77630758e-01
5.42527854e-01 -1.84258193e-01 3.72254133e-01 -4.73170280e-01
6.31284192e-02 1.74655944e-01 -1.49863398e+00 2.04048514e+00
-9.33982283e-02 6.78277016e-02 -1.57440394e-01 -4.85482931e-01
7.87207246e-01 4.30377543e-01 5.16548574e-01 -3.02949876e-01
1.33526847e-01 5.80923110e-02 -3.75852495e-01 -3.09882909e-01
7.77066574e-02 -6.34571984e-02 -5.67140803e-02 2.57321149e-01
1.81558561e-02 2.71157354e-01 -1.13503478e-01 -1.23460084e-01
3.92023534e-01 7.51446486e-01 5.29817045e-01 8.87536779e-02
3.90086919e-01 -5.53740382e-01 4.96346474e-01 1.77949890e-01
-9.55514386e-02 9.46594894e-01 5.75869918e-01 -6.90795660e-01
-1.21827412e+00 -1.48941708e+00 1.86205342e-01 6.20131195e-01
2.45776419e-02 -6.06215537e-01 -8.05566669e-01 -7.43013203e-01
3.22645098e-01 1.61077790e-02 -7.25422144e-01 6.17475025e-02
-8.79504383e-01 -2.20267504e-01 3.62350076e-01 8.09557974e-01
5.24284065e-01 -4.06946391e-01 -6.56158686e-01 -2.85223663e-01
-5.45123219e-01 -1.19274759e+00 -8.83871615e-01 -2.07240283e-01
-8.17835212e-01 -1.09363234e+00 -1.35283852e+00 -5.46838224e-01
9.25062478e-01 1.68792740e-01 7.12271035e-01 -1.66154295e-01
-4.47337963e-02 5.34533978e-01 -1.56237379e-01 -1.10248759e-01
2.82378167e-01 -2.57818609e-01 6.70275509e-01 9.49921161e-02
2.22521231e-01 -6.20192230e-01 -8.23566318e-01 5.86405575e-01
-1.76515743e-01 3.60772908e-01 2.65913039e-01 6.83432341e-01
5.84651411e-01 -7.10706711e-01 -1.27613237e-02 -2.22972959e-01
1.22411959e-01 -3.70798260e-02 -4.24569994e-01 6.61814511e-02
-1.21888839e-01 -3.98132689e-02 1.33735538e-01 -3.23165387e-01
-7.14783728e-01 5.68118870e-01 -1.33289695e-01 -5.91896236e-01
-2.06790939e-01 2.17291951e-01 -4.80665080e-02 1.54990435e-01
5.51303387e-01 -8.34504291e-02 3.89506787e-01 -8.26183915e-01
4.28584784e-01 4.67928052e-01 6.65310323e-01 -6.02362990e-01
8.48907471e-01 7.16691554e-01 2.54030347e-01 -7.03236580e-01
-9.43996429e-01 -5.49627900e-01 -1.35566545e+00 -4.82287943e-01
9.58046973e-01 -1.14273095e+00 -8.02660346e-01 4.52579886e-01
-1.31352627e+00 4.71089110e-02 2.09569316e-02 7.77408183e-01
-7.83393204e-01 8.62434685e-01 -6.79154336e-01 -6.89921677e-01
-3.96834761e-01 -9.89034235e-01 1.42460394e+00 -8.88541043e-02
-6.57395482e-01 -7.03769326e-01 3.07192206e-01 3.51319909e-01
-3.92121404e-01 5.54001927e-01 4.18507129e-01 -3.34580876e-02
-1.24343909e-01 -4.62604284e-01 3.01588308e-02 1.86355203e-01
-8.14687982e-02 -3.93206000e-01 -7.40510762e-01 -3.15386325e-01
-1.99829921e-01 -2.21639901e-01 4.76717919e-01 5.54921508e-01
5.84239542e-01 -1.33275539e-01 -1.77800164e-01 6.43334508e-01
9.91788208e-01 -3.09400976e-01 4.34779674e-01 5.29578149e-01
8.95931721e-01 8.91908467e-01 3.95913541e-01 5.28149009e-01
4.54711854e-01 1.19868219e+00 3.64429802e-01 2.27824897e-02
-5.22530936e-02 -6.49073303e-01 5.31723976e-01 6.74621880e-01
-7.02820122e-01 2.46884122e-01 -8.04289520e-01 2.06814542e-01
-1.90600491e+00 -4.60881084e-01 2.01058760e-02 2.34122276e+00
4.95655149e-01 1.09607294e-01 5.97128272e-01 -8.95248130e-02
8.29723716e-01 1.68212608e-01 -4.31836754e-01 2.88746566e-01
1.55823603e-01 -2.70885944e-01 3.23840559e-01 5.18100619e-01
-1.02045965e+00 6.21887088e-01 5.54108906e+00 1.96683824e-01
-7.84936428e-01 -5.45131080e-02 1.78828880e-01 -2.60496348e-01
1.32889390e-01 -1.93321124e-01 -9.06193733e-01 1.86499462e-01
3.91779065e-01 3.63202900e-01 1.54586971e-01 8.22558284e-01
4.60413583e-02 6.12958968e-02 -1.22105694e+00 1.28845000e+00
2.29357377e-01 -9.88145828e-01 -8.15282017e-03 1.67800754e-01
4.66450334e-01 -1.53370267e-02 -1.06528118e-01 -2.39478707e-01
-2.90548980e-01 -6.24576151e-01 1.00250483e+00 6.00870371e-01
9.78434086e-01 -5.94687998e-01 5.07868290e-01 4.61584896e-01
-1.37457299e+00 2.41027579e-01 -2.57065773e-01 -2.11287990e-01
5.03639400e-01 3.32577586e-01 -2.97686517e-01 5.39749622e-01
8.34684551e-01 8.24502230e-01 -3.56136709e-01 6.99957013e-01
-6.23228252e-01 -1.26322061e-01 -5.39366305e-01 2.28559271e-01
-2.81955805e-02 -4.08999920e-01 8.44247937e-01 8.78485084e-01
2.76413530e-01 -1.56039551e-01 3.49540591e-01 6.33780956e-01
9.68920141e-02 -1.84709236e-01 -6.73301399e-01 4.50527459e-01
1.02081031e-01 1.23338401e+00 -4.17292386e-01 -2.44492114e-01
-2.73460895e-01 1.30893779e+00 2.77240604e-01 3.95437837e-01
-7.51876771e-01 1.46711813e-02 6.14020109e-01 2.29794696e-01
3.01251430e-02 -5.44861436e-01 -7.97286257e-02 -1.48447192e+00
6.50066137e-01 -7.39424169e-01 3.68278474e-01 -9.00064170e-01
-1.12985396e+00 3.56698960e-01 2.33308867e-01 -1.57665646e+00
-5.38255095e-01 -8.87213349e-01 -1.66624367e-01 8.64888191e-01
-7.89118409e-01 -1.14864850e+00 -2.47588053e-01 5.57316840e-01
3.90658349e-01 1.60920501e-01 7.40744948e-01 1.41703367e-01
-3.98224443e-01 5.42860866e-01 -3.65742266e-01 4.38982278e-01
8.97427738e-01 -1.15845525e+00 7.73133576e-01 5.55268228e-01
1.15065463e-01 8.36011171e-01 5.77404201e-01 -6.89616799e-01
-1.43956339e+00 -6.63085699e-01 8.84653211e-01 -9.07911301e-01
3.31036389e-01 -5.80511868e-01 -1.88603297e-01 1.10050988e+00
-1.25702292e-01 5.10705560e-02 3.98667097e-01 2.53387064e-01
-6.04242384e-01 1.52919039e-01 -9.83454645e-01 9.02895629e-01
1.31974399e+00 -4.13099855e-01 -8.21659148e-01 3.38216990e-01
3.67423445e-01 -8.31853807e-01 -8.94533575e-01 2.80060679e-01
1.19712567e+00 -8.00005555e-01 1.34049213e+00 -7.44007647e-01
2.78104007e-01 -4.76746976e-01 -2.79105753e-01 -1.26850581e+00
-3.97972316e-01 -6.70955896e-01 -2.94351399e-01 5.30734301e-01
1.34958565e-01 -5.09904683e-01 1.01893270e+00 2.10380331e-01
2.61075646e-01 -9.27291751e-01 -1.05291283e+00 -8.39374304e-01
-2.84473747e-01 -3.22706342e-01 6.60857558e-02 4.73811150e-01
1.55168608e-01 5.30938685e-01 -9.16973472e-01 2.57719278e-01
8.98880243e-01 2.08268631e-02 1.12060428e+00 -1.03659225e+00
-4.60946858e-01 -3.26659121e-02 -6.14227712e-01 -1.58353341e+00
-3.36389303e-01 -5.74918926e-01 -1.64828166e-01 -1.22559345e+00
3.75397027e-01 3.89737576e-01 9.24725011e-02 3.76457900e-01
-1.56231403e-01 4.69430953e-01 5.18673539e-01 3.12090874e-01
-3.91334772e-01 4.65263814e-01 1.36997867e+00 3.10134172e-01
-5.53201102e-02 1.00597166e-01 -3.48893821e-01 1.26247561e+00
6.81218266e-01 -3.79951239e-01 7.06742778e-02 -3.04962903e-01
1.02695204e-01 1.24518365e-01 8.51728737e-01 -9.63945210e-01
3.28364559e-02 7.54630491e-02 9.75835443e-01 -9.14008141e-01
7.53623724e-01 -7.13574350e-01 3.79773796e-01 7.51665831e-01
-1.95788205e-01 4.00269508e-01 -2.18494803e-01 5.84325016e-01
1.62360832e-01 1.87678277e-01 6.62101090e-01 -4.48198050e-01
-4.34865981e-01 3.94880265e-01 5.48136570e-02 -1.17572166e-01
7.63273716e-01 -4.41381127e-01 -1.04214370e-01 -6.89638913e-01
-1.06948090e+00 1.94423288e-01 5.25287628e-01 5.35986245e-01
7.61618316e-01 -1.64696467e+00 -5.63971043e-01 4.50848907e-01
7.86027759e-02 2.20507607e-02 2.48808160e-01 1.09342158e+00
-6.14394784e-01 5.53445339e-01 -4.41540092e-01 -8.99080634e-01
-1.41445196e+00 2.38395110e-01 5.28370142e-01 -2.51427591e-01
-9.42700148e-01 6.85141623e-01 2.81336039e-01 -9.48908389e-01
3.75932038e-01 7.39052221e-02 -3.30790132e-03 -1.73361316e-01
3.77422720e-01 6.68047905e-01 -3.56293023e-01 -9.61279273e-01
-4.88632053e-01 1.37616789e+00 7.79011846e-02 -4.58472043e-01
1.24691308e+00 -2.46494949e-01 5.85923605e-02 2.93151587e-01
1.47135448e+00 9.95118245e-02 -1.64089727e+00 -3.08426470e-01
-2.62473136e-01 -5.98188698e-01 -5.36959589e-01 -5.48027337e-01
-9.70571816e-01 8.13271523e-01 6.30445600e-01 -3.20068002e-01
7.70253360e-01 1.63330361e-01 6.91896439e-01 4.11545008e-01
4.80628759e-01 -1.08984208e+00 4.21529800e-01 4.83855873e-01
1.37400663e+00 -9.96178448e-01 3.92978132e-01 -6.78697050e-01
-5.41478634e-01 1.18260550e+00 6.50842845e-01 -3.93904537e-01
6.68981671e-01 -1.43596798e-01 2.98757851e-01 -2.41738498e-01
-1.59861371e-01 -8.13186020e-02 8.02141011e-01 5.70655644e-01
4.39381033e-01 2.57782936e-01 -2.94654608e-01 3.05729747e-01
-2.26839274e-01 -1.45940304e-01 1.57515425e-02 1.06343675e+00
-1.69494793e-01 -1.09211552e+00 -6.86217487e-01 -1.04186036e-01
-3.78952950e-01 3.81300539e-01 -6.31737590e-01 1.05301070e+00
9.78738144e-02 4.62634444e-01 -7.31487796e-02 -6.85351014e-01
8.40147078e-01 1.26127690e-01 7.65956700e-01 -4.39734310e-01
-9.25787911e-02 6.66625559e-01 2.01684162e-01 -8.95566344e-01
-4.90355730e-01 -7.40947366e-01 -1.07551777e+00 -2.00965628e-01
-2.73793712e-02 -3.39819252e-01 4.93330866e-01 7.88518250e-01
3.36708397e-01 -1.13466412e-01 4.33619797e-01 -1.36812246e+00
-7.39015400e-01 -8.77185762e-01 -5.76329172e-01 4.88739550e-01
3.36169392e-01 -1.03557909e+00 -9.21513513e-02 -6.90295845e-02] | [7.0166239738464355, -0.941461980342865] |
c43773c0-3b35-4e57-9269-eee37073e563 | generative-hierarchical-features-from | 2007.10379 | null | https://arxiv.org/abs/2007.10379v2 | https://arxiv.org/pdf/2007.10379v2.pdf | Generative Hierarchical Features from Synthesizing Images | Generative Adversarial Networks (GANs) have recently advanced image synthesis by learning the underlying distribution of the observed data. However, how the features learned from solving the task of image generation are applicable to other vision tasks remains seldom explored. In this work, we show that learning to synthesize images can bring remarkable hierarchical visual features that are generalizable across a wide range of applications. Specifically, we consider the pre-trained StyleGAN generator as a learned loss function and utilize its layer-wise representation to train a novel hierarchical encoder. The visual feature produced by our encoder, termed as Generative Hierarchical Feature (GH-Feat), has strong transferability to both generative and discriminative tasks, including image editing, image harmonization, image classification, face verification, landmark detection, and layout prediction. Extensive qualitative and quantitative experimental results demonstrate the appealing performance of GH-Feat. | ['Yujun Shen', 'Ceyuan Yang', 'Bolei Zhou', 'Yinghao Xu', 'Jiapeng Zhu'] | 2020-07-20 | null | http://openaccess.thecvf.com//content/CVPR2021/html/Xu_Generative_Hierarchical_Features_From_Synthesizing_Images_CVPR_2021_paper.html | http://openaccess.thecvf.com//content/CVPR2021/papers/Xu_Generative_Hierarchical_Features_From_Synthesizing_Images_CVPR_2021_paper.pdf | cvpr-2021-1 | ['image-harmonization'] | ['computer-vision'] | [ 6.29336417e-01 2.34620765e-01 4.10603248e-02 -5.55595219e-01
-7.96034873e-01 -5.61267614e-01 7.74638057e-01 -6.24753416e-01
2.96938747e-01 7.96383739e-01 2.49095961e-01 -6.58535063e-02
5.76366112e-02 -8.38842630e-01 -1.15625203e+00 -9.59497035e-01
2.77199060e-01 4.14501168e-02 -4.15036947e-01 -1.81028083e-01
6.97767958e-02 7.09416211e-01 -1.38530278e+00 2.98995614e-01
9.14756596e-01 1.05754662e+00 6.50578886e-02 5.48159301e-01
2.65394241e-01 8.33251774e-01 -6.84648693e-01 -5.85082471e-01
3.89898092e-01 -9.27152812e-01 -5.50372601e-01 4.29929674e-01
5.49975097e-01 -3.32698792e-01 -3.68986428e-01 9.99181986e-01
4.83591408e-01 -6.89973831e-02 9.21378493e-01 -1.56676543e+00
-1.24842715e+00 4.26722914e-01 -6.34044349e-01 -3.78163695e-01
3.38454634e-01 3.79488885e-01 1.02991092e+00 -9.35757518e-01
7.05724716e-01 1.36832631e+00 6.07234061e-01 7.92279899e-01
-1.50153816e+00 -8.16280246e-01 -2.16255784e-01 4.01392654e-02
-1.37412632e+00 -3.62849444e-01 9.54842031e-01 -5.25538743e-01
2.29976386e-01 2.95519829e-01 6.13025427e-01 1.29086530e+00
2.28781477e-01 8.12618732e-01 1.19158983e+00 -3.27438384e-01
1.70680769e-02 -5.47475815e-02 -8.57088804e-01 9.22610402e-01
-1.17403362e-02 2.99541026e-01 -4.63277191e-01 2.56641299e-01
9.81961846e-01 4.72640060e-02 -3.83993566e-01 -4.53465819e-01
-9.30992723e-01 1.03530955e+00 7.78968394e-01 -4.49547991e-02
-1.85671702e-01 2.77392268e-01 4.62954715e-02 2.25363895e-01
2.09804013e-01 5.96810997e-01 -8.68692324e-02 3.49114120e-01
-8.41965258e-01 2.23830208e-01 3.63510340e-01 1.20544589e+00
1.04302096e+00 6.70005083e-01 -5.69687843e-01 8.37151170e-01
1.79060310e-01 6.92485988e-01 4.21657354e-01 -1.04935253e+00
1.38343535e-02 5.56182683e-01 -3.76116514e-01 -9.38413084e-01
4.43617553e-02 -3.44332397e-01 -1.35176075e+00 3.24291497e-01
1.18070297e-01 -1.71158582e-01 -9.98859107e-01 2.03077197e+00
2.26591423e-01 -6.05170429e-02 -9.09220614e-03 5.45658410e-01
9.25938070e-01 6.52413011e-01 -7.36422911e-02 2.17511952e-02
8.70398223e-01 -8.88290763e-01 -4.13590342e-01 -1.14394121e-01
8.12095702e-02 -8.77884865e-01 9.59372342e-01 3.50386510e-03
-1.12491286e+00 -8.18218648e-01 -9.50531542e-01 -1.95090771e-01
-1.64957851e-01 2.09434286e-01 6.09199286e-01 4.00441676e-01
-1.13403058e+00 4.18465108e-01 -2.64556020e-01 -2.10652336e-01
8.13579679e-01 2.91757613e-01 -5.01635611e-01 -1.41598836e-01
-8.87918293e-01 4.20251161e-01 1.34057015e-01 3.94171514e-02
-1.29801190e+00 -7.04173267e-01 -1.11410558e+00 7.57664815e-02
-3.36786662e-03 -1.18417239e+00 9.61931944e-01 -1.12581682e+00
-1.73571074e+00 9.98770356e-01 -6.95949197e-02 -2.86047310e-01
4.36319798e-01 2.30214551e-01 -2.11856753e-01 2.58843368e-03
2.68725693e-01 1.06426907e+00 1.40751660e+00 -1.45281827e+00
-4.62478161e-01 -1.88043043e-01 -7.02822730e-02 -8.90500695e-02
-2.70457536e-01 -3.69284660e-01 -3.24748188e-01 -1.05437219e+00
-2.99541622e-01 -9.02009547e-01 -1.67428464e-01 1.07022427e-01
-7.98729539e-01 -2.88812574e-02 8.85848463e-01 -5.27250469e-01
7.30825782e-01 -2.18359876e+00 3.80884588e-01 2.55691886e-01
1.26237750e-01 1.16879880e-01 -4.08775538e-01 3.72647941e-01
-1.71354577e-01 2.02837557e-01 -3.80815715e-01 -1.83461696e-01
1.05043389e-01 1.31077036e-01 -5.96935630e-01 2.69282520e-01
5.63243806e-01 1.51647937e+00 -6.43194854e-01 -4.20740932e-01
2.93747336e-01 6.21291757e-01 -7.79591084e-01 6.25058949e-01
-2.66503602e-01 9.14646089e-01 -4.66530889e-01 6.90527618e-01
5.79829574e-01 -3.77204180e-01 7.25117698e-02 -3.60002637e-01
3.13549817e-01 -3.30544502e-01 -4.20355439e-01 1.58125782e+00
-4.85350460e-01 8.89293253e-01 -2.55141199e-01 -9.68251348e-01
1.10510743e+00 -5.81856146e-02 3.11429232e-01 -8.81277323e-01
1.81633607e-02 -1.37179911e-01 8.21449968e-04 -3.32528353e-01
2.30333641e-01 -9.97005105e-02 -2.18626171e-01 2.67385691e-01
2.74554849e-01 -4.58121687e-01 -7.96199143e-02 2.05255836e-01
8.34755361e-01 1.94536895e-01 2.72674382e-01 -1.42609805e-01
4.35398757e-01 -4.12849784e-01 4.43431377e-01 6.79212809e-01
2.00337157e-01 9.21864808e-01 5.98865569e-01 -1.99472696e-01
-1.35794592e+00 -1.16624773e+00 9.78360418e-03 9.69400704e-01
-1.05974376e-02 -2.29906514e-01 -1.06554615e+00 -7.09853053e-01
2.99797170e-02 4.41206485e-01 -9.23080981e-01 -5.13505876e-01
-4.96097952e-01 -5.13990641e-01 5.55782437e-01 5.69984615e-01
6.09779596e-01 -1.26218069e+00 -1.15451880e-01 -2.00391918e-01
1.87898397e-01 -1.16459215e+00 -8.60086560e-01 -1.04333393e-01
-4.31811333e-01 -8.29682946e-01 -7.82265246e-01 -9.67281640e-01
8.91041696e-01 -2.96469741e-02 1.18734133e+00 3.58346812e-02
-5.87086320e-01 4.07338023e-01 -2.22186163e-01 -1.72854871e-01
-5.94885528e-01 -5.69680370e-02 -2.88892657e-01 4.05153394e-01
-3.03755403e-01 -6.98249817e-01 -7.40882814e-01 1.86629489e-01
-1.03934729e+00 2.92963743e-01 8.58857751e-01 1.23085868e+00
6.27362728e-01 -5.66450469e-02 5.72899044e-01 -1.18639123e+00
4.81297165e-01 -2.24217117e-01 -4.58300591e-01 3.58638138e-01
-4.44217145e-01 2.26265594e-01 8.57319951e-01 -2.57604182e-01
-1.09205151e+00 2.18570754e-01 -2.73708880e-01 -4.83068079e-01
-1.41844854e-01 1.12680972e-01 -7.14293838e-01 -3.58787835e-01
4.40838605e-01 5.59672534e-01 3.72655541e-02 -1.73732899e-02
7.25486755e-01 1.77846387e-01 8.13922048e-01 -5.57888329e-01
1.31699848e+00 3.80998671e-01 3.62260580e-01 -8.25175464e-01
-8.66646588e-01 3.42137903e-01 -4.63327706e-01 1.72335301e-02
1.03517044e+00 -8.86542499e-01 -7.04135835e-01 5.26160955e-01
-1.12181771e+00 -5.46570957e-01 -5.57606161e-01 -1.13549948e-01
-9.18727994e-01 1.03355438e-01 -3.63309234e-01 -4.13045436e-01
-3.39637727e-01 -1.20684516e+00 1.40802789e+00 4.62255329e-01
1.83290150e-02 -1.06421041e+00 -2.61654288e-01 2.25749180e-01
3.57171655e-01 7.36414254e-01 1.17529809e+00 1.50665501e-02
-8.40990603e-01 7.14962035e-02 -2.76679218e-01 5.71075857e-01
2.74601012e-01 1.56151995e-01 -9.59684730e-01 -3.63519907e-01
-2.01378345e-01 -4.58102077e-01 8.14446747e-01 4.46627349e-01
1.81156027e+00 -5.60746670e-01 -1.50113136e-01 1.28373539e+00
1.29706109e+00 1.86735898e-01 9.54720795e-01 -1.07109554e-01
1.08213603e+00 4.66983259e-01 2.18962744e-01 3.09475392e-01
1.85787752e-01 7.58021712e-01 5.49292862e-01 -5.20892859e-01
-4.96701241e-01 -7.76171744e-01 3.63784760e-01 5.53528726e-01
-1.32273985e-02 -3.24778885e-01 -5.08384824e-01 2.50374377e-01
-1.59925365e+00 -9.30759907e-01 3.55326235e-01 1.73691106e+00
9.84661162e-01 -2.91794688e-01 -9.27851424e-02 -1.65098652e-01
7.79318988e-01 1.06741697e-01 -6.51269794e-01 -2.45310158e-01
-3.27045739e-01 5.82658947e-01 3.57948184e-01 2.49951243e-01
-9.78635132e-01 1.06443691e+00 6.53521919e+00 1.04318583e+00
-1.23519349e+00 -8.87232572e-02 1.01629269e+00 3.00205737e-01
-5.87111413e-01 -1.22084171e-01 -5.83895028e-01 5.26712894e-01
3.47546011e-01 -2.51747280e-01 5.24918318e-01 8.05703700e-01
-1.52422473e-01 5.43078601e-01 -1.30856144e+00 1.12281382e+00
3.61131281e-01 -1.52833271e+00 6.78812027e-01 1.93905100e-01
1.26201451e+00 -5.91337264e-01 8.26028764e-01 1.37816146e-01
5.43656468e-01 -1.55340648e+00 6.83544874e-01 6.06594622e-01
1.36569476e+00 -8.21532488e-01 3.13623965e-01 -1.03516884e-01
-9.94114697e-01 -9.93615482e-03 -2.53608048e-01 3.87624234e-01
6.99740797e-02 2.23096341e-01 -8.18286598e-01 4.38518047e-01
4.40846503e-01 9.07917798e-01 -7.72987127e-01 6.85141623e-01
-4.92221951e-01 4.19942260e-01 3.07234317e-01 3.99689198e-01
1.72364816e-01 -1.95337191e-01 2.92959660e-01 8.34477663e-01
5.79003334e-01 -2.37639159e-01 -4.65246886e-02 1.19261134e+00
-5.48061550e-01 -1.73750356e-01 -9.04387355e-01 -1.96148306e-01
2.62614459e-01 1.31858933e+00 -3.42925251e-01 -6.83156848e-02
-1.41024396e-01 1.19128442e+00 2.26482525e-01 4.33563322e-01
-9.08769429e-01 -4.83785510e-01 6.77514017e-01 1.13548040e-01
5.29038727e-01 1.74659029e-01 -2.72509813e-01 -1.03932118e+00
-2.31555551e-01 -1.06057656e+00 -1.07924484e-01 -9.55038428e-01
-1.46435153e+00 7.87887633e-01 -2.73640215e-01 -1.26113641e+00
-6.41010702e-01 -6.73826396e-01 -9.33179498e-01 6.93523467e-01
-1.14897907e+00 -1.74793935e+00 -6.12671971e-01 8.17657650e-01
4.93864745e-01 -6.63686395e-01 7.62342811e-01 1.69226989e-01
-5.89976013e-01 1.10806775e+00 7.75886849e-02 3.53402466e-01
6.81535840e-01 -1.31960475e+00 4.39867616e-01 8.13459754e-01
3.45238745e-01 4.84039038e-01 5.23392141e-01 -2.45372072e-01
-1.40717781e+00 -1.59580922e+00 3.01251590e-01 -3.33098590e-01
3.61574948e-01 -4.75419611e-01 -6.30740106e-01 7.65630782e-01
4.06032383e-01 1.68246865e-01 5.07658958e-01 -3.50910813e-01
-5.95967174e-01 -3.16035450e-01 -1.13045847e+00 5.80596387e-01
1.23661554e+00 -6.46313190e-01 -1.66912284e-02 2.95129657e-01
6.88048542e-01 -4.17305082e-01 -8.77125859e-01 5.16034901e-01
5.82250476e-01 -9.20433998e-01 1.14530981e+00 -6.41391158e-01
9.62754011e-01 -1.46040872e-01 -2.06295952e-01 -1.49210644e+00
-4.61360455e-01 -7.46360838e-01 1.44571275e-01 1.48250461e+00
2.22080052e-01 -5.16680777e-01 6.40995204e-01 6.71910271e-02
-2.06744492e-01 -7.62412429e-01 -4.54271019e-01 -5.60633004e-01
2.55612761e-01 2.19552383e-01 9.21896458e-01 8.97312403e-01
-6.47777736e-01 4.81361032e-01 -6.95847452e-01 -1.45954534e-01
7.51010835e-01 4.87270623e-01 1.10950458e+00 -9.47395682e-01
-4.79491889e-01 -4.40882444e-01 -5.63127160e-01 -1.05959630e+00
5.08965433e-01 -1.01795352e+00 1.15201421e-01 -1.19389296e+00
3.09235990e-01 -3.63206416e-01 9.88946036e-02 4.42502111e-01
-1.92325443e-01 8.11705768e-01 2.95844823e-01 -8.53312947e-03
-4.26515698e-01 9.00357842e-01 1.72238445e+00 -3.65293711e-01
2.80587554e-01 -1.33592173e-01 -1.01724994e+00 5.45849860e-01
6.31126404e-01 -1.46103188e-01 -5.01202285e-01 -4.45152014e-01
3.10594328e-02 -1.25131562e-01 5.82484424e-01 -8.12844872e-01
-2.26437091e-03 -3.02184284e-01 9.23498929e-01 -1.36638075e-01
3.01528364e-01 -6.61295056e-01 4.15800422e-01 3.09083641e-01
-4.31401432e-01 -1.58795625e-01 2.07184777e-02 4.80686486e-01
-5.00602782e-01 6.54099807e-02 9.94558394e-01 9.09329858e-04
-6.52369022e-01 7.23583519e-01 1.44372523e-01 2.04613611e-01
1.13716733e+00 -9.93606746e-02 -4.10899639e-01 -6.80329144e-01
-5.55489659e-01 -1.40880838e-01 6.40599132e-01 4.31362420e-01
7.74386346e-01 -1.80517721e+00 -7.40651608e-01 6.07221246e-01
2.03181282e-01 8.53057019e-03 3.21411312e-01 4.17500287e-01
-3.68611336e-01 2.22099364e-01 -5.86118162e-01 -7.41620123e-01
-9.67384398e-01 5.39851248e-01 1.82782739e-01 -6.52790144e-02
-4.53623801e-01 8.95860672e-01 1.05865622e+00 -2.38314793e-01
-5.70454486e-02 7.42387697e-02 1.70081720e-01 -2.42103666e-01
2.55007774e-01 -1.14063494e-01 -3.09592485e-01 -8.60072970e-01
4.80253063e-02 7.71987975e-01 8.46669748e-02 2.44874194e-01
1.36267650e+00 2.20047332e-05 -2.01719716e-01 -4.72759977e-02
1.46345365e+00 -1.00136682e-01 -1.65508103e+00 -1.13992319e-01
-3.76793772e-01 -5.21807849e-01 -3.64934385e-01 -5.43627501e-01
-1.56532836e+00 8.93965423e-01 4.55664098e-01 1.70532256e-01
1.34842443e+00 1.77035257e-01 6.85014665e-01 -1.03436008e-01
2.70777136e-01 -5.90490580e-01 5.38323343e-01 3.33300084e-01
1.30017507e+00 -1.18014181e+00 -2.82769561e-01 -4.84213173e-01
-6.77520871e-01 9.51491356e-01 7.33239055e-01 -2.16692626e-01
4.25430655e-01 2.27552101e-01 -2.77096421e-01 9.92816165e-02
-4.82904643e-01 -1.70683801e-01 5.59270203e-01 1.02282405e+00
3.11359435e-01 1.80982515e-01 1.61852911e-01 3.61811221e-01
-5.71760535e-01 -2.75347471e-01 2.86591440e-01 4.22113806e-01
2.10922826e-02 -1.26011670e+00 -6.15955666e-02 3.88620138e-01
-3.74704570e-01 -1.30324364e-01 -5.31771064e-01 6.64535642e-01
3.79868299e-01 3.98821414e-01 7.54852816e-02 -5.23039043e-01
2.50223745e-02 -2.20724851e-01 7.79619217e-01 -6.28221095e-01
-2.49560162e-01 -1.07959487e-01 -5.07746935e-01 -5.84161937e-01
-3.05927545e-01 -4.92664725e-01 -8.29207599e-01 -3.31362605e-01
1.03701428e-01 -1.37890175e-01 4.25898045e-01 7.06744254e-01
4.46194977e-01 7.33944714e-01 9.60067928e-01 -8.50161493e-01
-3.84786278e-01 -8.26813102e-01 -5.90720177e-01 7.59005606e-01
4.65147704e-01 -5.46896994e-01 -2.22696155e-01 4.99114603e-01] | [11.702592849731445, -0.3892733156681061] |
a400ff64-76c4-487d-adba-42ca9497a89e | improving-hospital-mortality-prediction-with | 1811.12276 | null | http://arxiv.org/abs/1811.12276v2 | http://arxiv.org/pdf/1811.12276v2.pdf | Improving Hospital Mortality Prediction with Medical Named Entities and Multimodal Learning | Clinical text provides essential information to estimate the acuity of a
patient during hospital stays in addition to structured clinical data. In this
study, we explore how clinical text can complement a clinical predictive
learning task. We leverage an internal medical natural language processing
service to perform named entity extraction and negation detection on clinical
notes and compose selected entities into a new text corpus to train document
representations. We then propose a multimodal neural network to jointly train
time series signals and unstructured clinical text representations to predict
the in-hospital mortality risk for ICU patients. Our model outperforms the
benchmark by 2% AUC. | ['Taha Kass-Hout', 'Borui Zhang', 'Mohammed Khalilia', 'Mohammad Taha Bahadori', 'Daniel Navarro', 'Ram Bhakta', 'Mengqi Jin', 'Arun Ravi', 'Parminder Bhatia', 'Aaron Colak', 'Tiberiu Doman', 'Selvan Senthivel', 'Matthieu Liger', 'Busra Celikkaya'] | 2018-11-29 | null | null | null | null | ['negation-detection'] | ['natural-language-processing'] | [ 3.25773656e-01 2.13281482e-01 -2.79093266e-01 -5.22458553e-01
-8.04884911e-01 -3.44977677e-01 -2.76794434e-02 1.04736960e+00
-6.48234546e-01 8.43762100e-01 7.93633223e-01 -6.07647121e-01
-3.59820098e-01 -7.41831005e-01 -1.20304130e-01 -2.70897686e-01
-3.59786600e-01 7.85697222e-01 -5.70166945e-01 1.91009521e-01
-2.62840390e-01 5.11372268e-01 -5.92267692e-01 9.70120311e-01
8.91190886e-01 1.02875721e+00 -4.69152868e-01 7.81165600e-01
-9.36968401e-02 1.50439656e+00 -6.38198614e-01 -1.90529704e-01
-1.05814137e-01 -3.97644371e-01 -6.42376304e-01 -2.52189159e-01
-3.37300390e-01 -3.49416405e-01 -4.91572499e-01 6.25330806e-01
7.98213124e-01 -8.05204958e-02 8.30802917e-01 -6.78227246e-01
-4.82128829e-01 8.01782548e-01 2.43195713e-01 3.73568118e-01
3.96131754e-01 9.90336388e-02 8.32326710e-01 -7.38108933e-01
5.99395514e-01 4.95724499e-01 9.53090012e-01 4.63568181e-01
-9.36075091e-01 -3.69119078e-01 -2.13867109e-02 -2.31668949e-01
-1.19924581e+00 -1.99029639e-01 4.16270167e-01 -5.14824331e-01
1.39159679e+00 1.52926715e-02 6.00890994e-01 1.26768780e+00
6.15978062e-01 5.07988214e-01 3.19211602e-01 -1.63697287e-01
2.79112309e-02 -1.42019182e-01 4.09966767e-01 8.33057165e-01
2.47730285e-01 -3.22346464e-02 -4.96364683e-01 -6.41960323e-01
4.59626377e-01 8.14376295e-01 -4.39068109e-01 4.28871840e-01
-1.62609553e+00 7.50120282e-01 3.24313134e-01 2.32991055e-01
-8.11867416e-01 1.54586043e-02 8.27435911e-01 4.95001405e-01
1.36345819e-01 7.42214024e-01 -9.84673381e-01 -1.58343881e-01
-8.03086340e-01 -2.54855663e-01 9.76566076e-01 6.68553114e-01
-1.64755911e-01 -5.09170406e-02 -4.42975253e-01 6.57890260e-01
9.51465145e-02 5.92770159e-01 1.00371623e+00 -5.23032188e-01
7.08346069e-01 9.39077616e-01 -2.67926813e-03 -5.40816665e-01
-9.78859127e-01 -3.12756628e-01 -1.15505183e+00 -7.12279201e-01
-4.04429324e-02 -7.87118614e-01 -8.46139371e-01 1.29858589e+00
-2.71580696e-01 1.90614879e-01 6.14615202e-01 3.68131697e-01
1.13609779e+00 4.87093925e-01 5.49439728e-01 -6.46158755e-01
1.37213850e+00 -5.64950824e-01 -1.02218664e+00 -1.61311775e-02
1.14329851e+00 -2.63801992e-01 1.87643111e-01 3.50818276e-01
-1.08703399e+00 6.00420833e-02 -7.17721462e-01 1.82866603e-01
-3.69356960e-01 4.69325930e-01 3.11674237e-01 1.81480162e-02
-7.76504755e-01 6.99133158e-01 -1.28009236e+00 -2.39447236e-01
5.85828900e-01 5.49743652e-01 -5.17009139e-01 1.40361592e-01
-1.27557063e+00 8.60565007e-01 6.26677990e-01 1.97992295e-01
-2.30952293e-01 -9.76566434e-01 -1.10248876e+00 2.29949370e-01
-3.68759707e-02 -1.29126716e+00 1.23020470e+00 -3.80639464e-01
-1.07869864e+00 8.93590510e-01 -2.79503137e-01 -8.38571966e-01
2.09771067e-01 -4.22129989e-01 -7.47355998e-01 4.19235349e-01
-7.14280382e-02 1.38961434e-01 2.31811807e-01 -4.60342705e-01
-4.24189568e-01 -3.76621127e-01 -7.63771117e-01 -1.28928110e-01
-2.78351545e-01 3.04837022e-02 -1.23623788e-01 -9.20684993e-01
6.98763058e-02 -6.48064792e-01 -6.84374988e-01 -4.31358933e-01
-5.94557881e-01 -1.76796779e-01 1.80191532e-01 -6.01373851e-01
1.74630344e+00 -1.95449436e+00 -1.00586355e-01 2.05415174e-01
7.34898329e-01 1.86927587e-01 6.43341094e-02 7.32251823e-01
-4.50120240e-01 2.13278428e-01 -1.89777479e-01 -2.56324232e-01
-4.87599969e-01 4.44162712e-02 -5.39458752e-01 2.28762314e-01
8.77152920e-01 1.26401579e+00 -8.78471851e-01 -6.38229907e-01
-2.85949260e-02 6.13935232e-01 -6.75669611e-01 5.84956467e-01
2.74835108e-03 5.66170871e-01 -7.29994655e-01 7.14848518e-01
-9.88569334e-02 -8.19796860e-01 4.12923396e-01 -1.16078950e-01
3.04388821e-01 4.83365864e-01 -4.43317622e-01 1.38922119e+00
-2.17508167e-01 4.18074727e-01 -3.87062907e-01 -9.28647518e-01
8.04729998e-01 8.06416273e-01 1.08939672e+00 -3.30633432e-01
4.89948779e-01 1.06414534e-01 -1.72335938e-01 -1.06612384e+00
-1.20323934e-01 -4.17607456e-01 -2.78383940e-01 2.33126760e-01
-1.82925295e-02 1.96367696e-01 -2.85969436e-01 6.47989884e-02
1.52342927e+00 -2.86456764e-01 8.65657091e-01 2.04260558e-01
3.62945050e-01 6.70809820e-02 6.52683675e-01 6.51696503e-01
-3.43581289e-01 6.30520701e-01 6.14818931e-01 -8.25217187e-01
-5.15717030e-01 -1.25660670e+00 -3.83660555e-01 7.22761154e-01
-5.57208657e-01 -6.16084516e-01 -2.57935729e-02 -7.63201773e-01
1.41682878e-01 6.64079964e-01 -7.59065986e-01 -4.98090893e-01
-7.36127019e-01 -8.68225098e-01 7.14782059e-01 1.03765595e+00
-3.76000792e-01 -1.33818388e+00 -8.99113834e-01 6.06405258e-01
-3.16467673e-01 -1.14590669e+00 -4.72999513e-01 7.76689887e-01
-1.19469011e+00 -1.31396294e+00 -4.43114698e-01 -1.06176448e+00
5.44369519e-01 -6.76845789e-01 1.20324934e+00 -1.17909022e-01
-5.39807320e-01 3.58874112e-01 -2.96986014e-01 -7.40569830e-01
-5.21370888e-01 2.28465572e-02 1.63940042e-01 -1.10958248e-01
7.72583008e-01 -3.97857785e-01 -9.05517042e-01 -2.73667455e-01
-9.16226089e-01 -2.75447637e-01 4.87452894e-01 1.13486910e+00
7.23617136e-01 -5.22118926e-01 9.86135244e-01 -1.12371159e+00
9.55862999e-01 -8.49734783e-01 2.84446925e-02 2.13045850e-01
-8.04932177e-01 1.78098336e-01 8.56580675e-01 -2.88872302e-01
-6.09688461e-01 2.44799376e-01 -5.08673862e-02 -6.56877995e-01
-2.25129724e-01 1.10379541e+00 5.50477922e-01 8.71789873e-01
5.46720445e-01 1.82579473e-01 -2.83713669e-01 -1.58771977e-01
-7.98812509e-02 8.47068965e-01 6.78529024e-01 -6.27965927e-02
1.63028255e-01 2.28059500e-01 7.95085058e-02 -4.43518192e-01
-1.20800281e+00 -8.01377535e-01 -8.04558575e-01 3.38285893e-01
1.33012748e+00 -1.10909903e+00 -1.01103437e+00 -1.09192654e-01
-1.20532012e+00 7.99795613e-02 -4.82189119e-01 9.39567089e-01
-4.86648500e-01 1.91195402e-02 -9.10898089e-01 -5.91086984e-01
-8.35777640e-01 -6.00702584e-01 1.19932461e+00 -1.27673686e-01
-5.41572452e-01 -1.23098385e+00 3.65545213e-01 5.29263616e-02
1.55810729e-01 7.67697155e-01 1.10069060e+00 -1.49764681e+00
1.43699542e-01 -6.01498306e-01 -7.07109198e-02 1.70210019e-01
4.22284901e-01 -1.05664544e-01 -6.83155775e-01 -2.07005084e-01
1.86282918e-01 -2.07890555e-01 1.12170577e+00 5.75561523e-01
1.11091757e+00 -3.78099620e-01 -5.01941025e-01 9.76645231e-01
1.30258787e+00 4.58868504e-01 2.03587547e-01 -2.73101646e-02
5.76455832e-01 3.77288431e-01 2.88985074e-01 1.01370990e+00
3.31986874e-01 -3.08736354e-01 6.01691194e-02 -6.37049135e-03
6.49674237e-01 3.17188399e-03 1.50006339e-01 1.16104460e+00
2.64298152e-02 -2.96182066e-01 -1.52334118e+00 4.63844925e-01
-1.68157065e+00 -8.69059026e-01 -7.85030890e-03 1.73622406e+00
1.10867739e+00 -1.12373149e-02 -2.60937423e-01 -1.03241265e-01
3.55883777e-01 -3.57242376e-01 -6.36135578e-01 -1.97597504e-01
-5.16160168e-02 4.32171822e-01 3.88589621e-01 8.02892745e-02
-1.32383180e+00 5.57461262e-01 6.85469055e+00 -3.93958539e-01
-1.18812656e+00 -1.52689025e-01 7.33840823e-01 -4.27336752e-01
2.17981681e-01 -6.82473838e-01 -4.48434114e-01 1.56327620e-01
1.55249596e+00 -4.15989518e-01 -1.52344376e-01 6.06839299e-01
4.21244353e-01 4.94972438e-01 -1.62867141e+00 1.22159243e+00
9.06428993e-02 -1.51709092e+00 2.14574367e-01 -1.47004798e-01
6.05081916e-01 4.23218250e-01 -1.88547179e-01 2.69980758e-01
2.90584266e-01 -1.56441629e+00 -1.21460080e-01 1.11618757e+00
1.08154655e+00 -4.17683929e-01 1.26919758e+00 2.19291925e-01
-1.07844281e+00 -3.83747995e-01 1.11900181e-01 3.08811754e-01
3.18832457e-01 2.63549685e-01 -1.34956932e+00 5.14487803e-01
4.84931082e-01 1.01480865e+00 -4.10846025e-01 1.08786452e+00
9.41804275e-02 8.27166438e-01 -7.95239806e-02 2.61598796e-01
7.23673403e-03 1.76312655e-01 2.99875438e-01 1.33854842e+00
3.50667834e-01 7.69134283e-01 3.65952015e-01 6.05277658e-01
-3.91883254e-01 4.11760211e-01 -8.84747446e-01 -5.88201463e-01
2.31256083e-01 8.35575938e-01 -3.16741616e-01 -5.68372130e-01
-5.29060483e-01 9.18919384e-01 2.42981747e-01 4.69552964e-01
-6.72295153e-01 -4.96442616e-01 2.73428828e-01 8.66673365e-02
1.35865763e-01 1.05812252e-01 -4.45867777e-01 -1.61480105e+00
-2.57626951e-01 -6.84017003e-01 9.66698527e-01 -6.60518706e-01
-1.61118376e+00 9.56342220e-01 -6.34245396e-01 -1.61584437e+00
-6.62933290e-01 -8.61248136e-01 -6.56114817e-01 8.95876884e-01
-1.57577908e+00 -7.42012501e-01 -2.48183712e-01 7.75651634e-01
1.91716135e-01 -4.11775023e-01 1.40681338e+00 1.92562222e-01
-7.66394317e-01 3.87573034e-01 1.02266163e-01 9.09013391e-01
1.01118851e+00 -1.16235721e+00 -4.59064990e-02 8.57659206e-02
-4.10514027e-01 1.09911573e+00 1.04181901e-01 -9.16358471e-01
-1.16784942e+00 -1.68420947e+00 1.28935277e+00 -7.56319642e-01
8.56765747e-01 3.47030312e-01 -1.00733387e+00 1.15154016e+00
1.53750032e-01 2.12418333e-01 1.50217068e+00 -2.48310417e-01
-2.75637716e-01 2.23508745e-01 -9.05921102e-01 3.47198069e-01
5.99229872e-01 -6.86399639e-01 -1.20649505e+00 3.77476156e-01
1.16965461e+00 -2.18478397e-01 -1.79307389e+00 6.55822635e-01
2.47828543e-01 -2.11873963e-01 8.91161203e-01 -1.64743733e+00
7.17904568e-01 9.86048654e-02 5.82486801e-02 -1.15444708e+00
-6.38183355e-02 -6.02060676e-01 -3.15949529e-01 4.74366277e-01
7.60956764e-01 -8.20259094e-01 6.22947514e-01 6.87620640e-01
-1.95554569e-02 -1.04566133e+00 -7.12131560e-01 -4.84002642e-02
2.46997729e-01 -3.03015977e-01 3.78854305e-01 1.15443063e+00
6.16966307e-01 5.20964265e-01 6.93328232e-02 3.30727011e-01
1.17624231e-01 1.70074925e-01 -1.08369894e-01 -1.65515304e+00
-1.84060670e-02 -4.40044373e-01 -4.03297365e-01 -3.63820702e-01
1.37939021e-01 -1.24666607e+00 -8.58578533e-02 -1.91435015e+00
2.39277124e-01 1.63741820e-02 -9.82254446e-01 7.02172935e-01
-4.12826836e-01 -2.37518147e-01 -1.88162252e-01 2.09525242e-01
-6.51683688e-01 5.23190260e-01 8.62097144e-01 -2.32463971e-01
-6.56681776e-01 1.62111297e-01 -5.11999369e-01 6.44850910e-01
8.13753068e-01 -7.16364324e-01 -1.25403151e-01 -2.08649606e-01
3.49199742e-01 8.82410705e-01 1.14519268e-01 -7.77725399e-01
3.27728152e-01 -1.09300658e-01 7.06778109e-01 -5.60791373e-01
-3.90514359e-02 -9.66703892e-01 -5.36020637e-01 7.57732689e-01
-8.20541799e-01 4.91357028e-01 3.89358580e-01 6.77268267e-01
-6.23249769e-01 1.77407965e-01 3.57589334e-01 2.10869825e-04
7.36781433e-02 4.96168524e-01 -7.27616191e-01 4.56367701e-01
9.30047214e-01 2.30533943e-01 -2.90594280e-01 -2.39857778e-01
-1.49141979e+00 5.19143522e-01 -2.08289236e-01 2.83263832e-01
9.83437896e-01 -1.08181679e+00 -9.11019325e-01 3.19525123e-01
3.68231475e-01 -6.79967105e-02 1.08050734e-01 1.22407722e+00
-6.96053028e-01 7.11415708e-01 1.19114332e-01 -3.94246370e-01
-1.13073111e+00 7.27950275e-01 4.93984848e-01 -5.07496893e-01
-7.99465299e-01 5.17928720e-01 -2.61633787e-02 -3.71091992e-01
4.80902642e-01 -9.81652558e-01 -5.79629898e-01 2.89930105e-01
5.39604008e-01 -3.12552035e-01 4.66858894e-02 -2.67317533e-01
-5.43066800e-01 2.38637969e-01 3.18035744e-02 2.64821500e-01
1.79529607e+00 1.79883108e-01 -2.11540937e-01 6.80668652e-01
1.40553772e+00 -3.78782116e-02 -4.03886884e-01 -3.01240057e-01
3.07745606e-01 4.74271268e-01 -1.83113530e-01 -9.50397730e-01
-8.17903340e-01 8.15635562e-01 4.20192152e-01 5.80875315e-02
1.32665515e+00 8.82500783e-03 7.57201374e-01 1.07555878e+00
-3.77511770e-01 -6.70244455e-01 1.33142741e-02 6.75954938e-01
7.37501383e-01 -1.47654009e+00 -2.54700422e-01 -1.03284009e-01
-1.17024398e+00 1.55587983e+00 3.08971763e-01 -8.17502588e-02
1.15678632e+00 4.50218529e-01 4.52429444e-01 -4.40953165e-01
-1.35543513e+00 3.44269280e-03 3.29120427e-01 3.00592601e-01
8.31636250e-01 7.09056780e-02 5.48215881e-02 1.31149745e+00
-8.68451223e-02 5.22748411e-01 5.10161161e-01 9.98866498e-01
-1.23881750e-01 -8.68841767e-01 -3.12002134e-02 1.04184866e+00
-1.00101912e+00 -7.17912734e-01 -3.99263591e-01 1.88500240e-01
-6.68907389e-02 9.60465252e-01 9.18095112e-02 -3.24953288e-01
7.25153625e-01 8.09020162e-01 -1.08787172e-01 -9.92989838e-01
-1.12506294e+00 -5.82521893e-02 1.87701941e-01 -5.79556584e-01
-1.76031962e-01 -5.99174082e-01 -1.81384146e+00 4.97104287e-01
1.46876127e-01 1.37241736e-01 2.04705611e-01 6.14989102e-01
7.98175097e-01 1.23787546e+00 3.70831639e-01 -1.43246669e-02
-6.47352040e-01 -9.09639537e-01 -2.30766594e-01 5.23350239e-01
8.86790812e-01 9.37840808e-03 -2.93698519e-01 5.15921831e-01] | [7.960072994232178, 6.527502536773682] |
1f3e2377-15fe-4e2c-a9ed-b665171174ac | a-billion-scale-foundation-model-for-remote | 2304.05215 | null | https://arxiv.org/abs/2304.05215v1 | https://arxiv.org/pdf/2304.05215v1.pdf | A Billion-scale Foundation Model for Remote Sensing Images | As the potential of foundation models in visual tasks has garnered significant attention, pretraining these models before downstream tasks has become a crucial step. The three key factors in pretraining foundation models are the pretraining method, the size of the pretraining dataset, and the number of model parameters. Recently, research in the remote sensing field has focused primarily on the pretraining method and the size of the dataset, with limited emphasis on the number of model parameters. This paper addresses this gap by examining the effect of increasing the number of model parameters on the performance of foundation models in downstream tasks such as rotated object detection and semantic segmentation. We pretrained foundation models with varying numbers of parameters, including 86M, 605.26M, 1.3B, and 2.4B, to determine whether performance in downstream tasks improved with an increase in parameters. To the best of our knowledge, this is the first billion-scale foundation model in the remote sensing field. Furthermore, we propose an effective method for scaling up and fine-tuning a vision transformer in the remote sensing field. To evaluate general performance in downstream tasks, we employed the DOTA v2.0 and DIOR-R benchmark datasets for rotated object detection, and the Potsdam and LoveDA datasets for semantic segmentation. Experimental results demonstrated that, across all benchmark datasets and downstream tasks, the performance of the foundation models and data efficiency improved as the number of parameters increased. Moreover, our models achieve the state-of-the-art performance on several datasets including DIOR-R, Postdam, and LoveDA. | ['Taekyung Lee', 'Junghoon Seo', 'Keumgang Cha'] | 2023-04-11 | null | null | null | null | ['object-detection-in-aerial-images'] | ['computer-vision'] | [ 1.34540170e-01 -4.12319303e-01 1.16209805e-01 -3.95045489e-01
-4.10315573e-01 -6.57402813e-01 3.60349089e-01 -2.11503506e-01
-7.16454327e-01 9.09937471e-02 -2.30121642e-01 -7.70311832e-01
-9.43637639e-02 -7.94676363e-01 -7.55644202e-01 -5.48496068e-01
-8.13087299e-02 2.80862778e-01 4.90908533e-01 -2.14050204e-01
2.15092048e-01 6.21330500e-01 -1.59100306e+00 2.16168866e-01
1.03497112e+00 9.58728254e-01 6.01880491e-01 7.16944635e-01
2.41758451e-01 1.66035309e-01 -4.99046922e-01 -7.93317184e-02
6.10867143e-01 -8.06647465e-02 -7.17639863e-01 6.39149174e-02
5.34805715e-01 -5.58920383e-01 -3.69887725e-02 8.25583577e-01
6.61674201e-01 2.73472577e-01 5.49421906e-01 -9.92700338e-01
-7.43083537e-01 6.20001793e-01 -8.26213896e-01 6.26724064e-01
-4.77318794e-01 2.90639102e-01 1.07185507e+00 -9.71546948e-01
2.04503283e-01 1.17896879e+00 9.58752990e-01 2.28429064e-01
-9.88121808e-01 -6.49893224e-01 4.07355487e-01 1.37000484e-02
-1.43507922e+00 -3.23145181e-01 1.49178341e-01 -6.52157247e-01
1.25927639e+00 3.05154100e-02 5.93298495e-01 4.91232991e-01
-9.19768512e-02 5.51487267e-01 1.30700314e+00 -3.82609010e-01
1.65003762e-01 3.63591090e-02 1.69458449e-01 6.76952422e-01
4.01689678e-01 1.94100067e-01 -2.13716030e-01 1.51595071e-01
9.81617570e-01 -1.98027134e-01 -1.72259927e-01 1.04690671e-01
-8.62176180e-01 9.45075393e-01 7.35311508e-01 4.38477099e-02
-9.04954523e-02 2.27735475e-01 5.86051457e-02 2.73535009e-02
6.00884080e-01 5.32141745e-01 -6.94957316e-01 2.79294640e-01
-8.64941537e-01 -8.71751830e-02 4.78120327e-01 6.65776789e-01
7.42180109e-01 -3.50017287e-02 5.11431843e-02 9.68598962e-01
4.32921380e-01 7.42418528e-01 2.82331705e-01 -9.16571319e-01
6.51562035e-01 5.51830530e-01 7.27373660e-02 -7.43202448e-01
-4.84789699e-01 -7.78209865e-01 -3.68203312e-01 2.77411431e-01
5.33399403e-01 -3.76297414e-01 -1.63976598e+00 1.41006863e+00
2.00788692e-01 9.98695288e-03 -3.94610083e-03 1.12840867e+00
1.04939604e+00 7.80728638e-01 1.41490906e-01 4.80298191e-01
1.31387770e+00 -1.12480009e+00 -1.33754564e-02 -8.08374584e-01
6.43270850e-01 -7.96239495e-01 1.34966469e+00 1.86652943e-01
-5.85951567e-01 -8.75186026e-01 -9.45022523e-01 -8.63517746e-02
-3.76997590e-01 6.59224391e-01 8.51134658e-01 5.82949162e-01
-1.06507647e+00 5.80715239e-01 -9.48925376e-01 -7.16550708e-01
5.04421830e-01 1.41437113e-01 -1.37399122e-01 -1.09778099e-01
-8.81348968e-01 9.16770279e-01 4.19358909e-01 3.52992892e-01
-9.86016035e-01 -8.63975346e-01 -5.22575736e-01 2.38848746e-01
1.62430942e-01 -5.45289874e-01 1.18921721e+00 -7.47101963e-01
-1.00757945e+00 9.44032192e-01 4.46810514e-01 -4.26515847e-01
3.93599778e-01 -4.72206384e-01 -8.73367116e-02 8.50592256e-02
9.88208279e-02 1.22557306e+00 7.62366474e-01 -1.24890125e+00
-7.71159112e-01 -4.79663014e-01 2.30014578e-01 4.93403733e-01
-1.77487403e-01 -1.65709108e-01 -5.85833371e-01 -4.37726557e-01
1.40683889e-01 -9.79956388e-01 -3.70681465e-01 1.08424395e-01
-1.23338953e-01 3.99181359e-02 9.11050081e-01 -5.84289193e-01
7.23056376e-01 -2.40070248e+00 -3.36883515e-01 2.73362994e-02
2.44203024e-04 5.43619037e-01 -4.40210253e-01 1.89310849e-01
1.11361302e-01 3.27810019e-01 -1.64319634e-01 -1.35427907e-01
-3.48363012e-01 4.28546548e-01 -2.50941575e-01 3.55543107e-01
3.58670801e-02 7.85707831e-01 -5.43440461e-01 -2.72976279e-01
2.97786981e-01 4.89344716e-01 -6.16924047e-01 8.69260505e-02
-3.53371762e-02 3.48383456e-01 -4.43505883e-01 5.76172113e-01
6.44390583e-01 -5.20330131e-01 -2.49519512e-01 -3.00133646e-01
-2.68502176e-01 8.47975090e-02 -9.34652090e-01 1.36600935e+00
-4.13352191e-01 6.25808835e-01 1.56509608e-01 -8.55958581e-01
1.02935660e+00 -1.24026030e-01 2.25192681e-01 -7.56205022e-01
7.69299269e-02 2.28408664e-01 2.13279843e-01 -5.15000224e-01
5.35970926e-01 -1.14031568e-01 2.88639069e-01 1.53104022e-01
-3.27340752e-01 -2.72213459e-01 3.19304705e-01 -7.71819949e-02
6.63831592e-01 1.02226838e-01 -1.62718087e-01 -1.41181335e-01
6.35257224e-03 4.71840084e-01 1.97668582e-01 1.00304246e+00
-1.35520905e-01 6.63591146e-01 -7.79887885e-02 -3.75360847e-01
-8.95069540e-01 -1.00528586e+00 -3.99755985e-01 1.31637359e+00
2.29632676e-01 -1.44975513e-01 -7.20446646e-01 -4.33286756e-01
2.33302370e-01 5.33860803e-01 -4.76933658e-01 -1.05675578e-01
-2.64394790e-01 -1.28107512e+00 6.88324749e-01 9.85253632e-01
1.03411806e+00 -9.82914925e-01 -9.77743268e-01 -7.34342821e-03
-1.23323075e-01 -1.32072890e+00 -2.46068314e-01 3.04977655e-01
-1.23762643e+00 -1.00694370e+00 -4.90331620e-01 -5.65028846e-01
6.43181443e-01 8.45555842e-01 9.81664896e-01 2.63143092e-01
-4.10528451e-01 1.28894597e-01 -3.56661618e-01 -4.71960247e-01
3.15897390e-02 2.81152040e-01 -4.28348094e-01 -3.41940403e-01
8.94118398e-02 -2.75121123e-01 -6.89122975e-01 4.19384629e-01
-8.45887840e-01 1.44327059e-01 7.87446976e-01 6.17973924e-01
5.26257873e-01 -1.77137274e-03 3.64812225e-01 -8.18025887e-01
2.20597476e-01 -1.27273127e-01 -7.06963360e-01 1.01977950e-02
-8.87299240e-01 -1.32656753e-01 3.35307032e-01 -1.66987881e-01
-1.10888064e+00 1.01725996e-01 -2.90658742e-01 -3.27207059e-01
-6.64946064e-02 8.52617800e-01 1.36214271e-01 -1.12014823e-02
8.40533972e-01 -1.78184599e-01 -2.31346726e-01 -7.80856311e-01
5.17244875e-01 6.99205399e-01 4.41870809e-01 -4.82995152e-01
7.40383327e-01 6.15302026e-01 -2.03875124e-01 -9.97939825e-01
-1.22228587e+00 -6.38550401e-01 -4.83469039e-01 1.13066673e-01
1.01775849e+00 -1.21861613e+00 -1.38591558e-01 8.80407333e-01
-7.68134832e-01 -8.37043524e-01 -1.93019003e-01 5.65863609e-01
-3.37423198e-02 1.77124217e-01 -5.01660764e-01 -5.02427697e-01
-4.83052194e-01 -1.17854738e+00 1.01391244e+00 2.96490878e-01
2.97771662e-01 -8.69000673e-01 -2.71854639e-01 7.14395523e-01
6.21707797e-01 -8.83727595e-02 7.96653867e-01 -2.02692613e-01
-4.08577412e-01 1.46588102e-01 -6.90183759e-01 5.56701899e-01
-7.74680972e-02 8.86465684e-02 -1.04043841e+00 -2.08477989e-01
-3.54877979e-01 -3.25871855e-01 1.41922867e+00 6.33262455e-01
1.04966319e+00 7.88429677e-02 -3.89170229e-01 1.08374667e+00
1.38959229e+00 -2.63703428e-03 5.80929697e-01 6.51126146e-01
8.56144488e-01 4.73115623e-01 6.61890686e-01 2.46265113e-01
4.62677896e-01 4.44927335e-01 5.59579730e-01 -4.52693909e-01
-1.94053039e-01 -7.77397603e-02 1.71110153e-01 2.32980266e-01
-2.67845929e-01 -1.03843272e-01 -1.21119034e+00 6.41002595e-01
-1.56215453e+00 -7.80137897e-01 -2.91362643e-01 1.91354299e+00
6.04555428e-01 6.70191348e-02 -5.36990725e-02 -1.40479044e-03
5.97455502e-01 1.73593417e-01 -6.28744483e-01 -1.56268299e-01
1.07213609e-01 6.42457381e-02 8.56474519e-01 4.02245343e-01
-1.27804458e+00 1.24699461e+00 6.16983271e+00 5.02337277e-01
-1.46478534e+00 5.47319055e-02 7.55609632e-01 1.83575079e-02
1.16365209e-01 3.72784585e-02 -1.17242217e+00 1.69336736e-01
6.97850883e-01 4.12415832e-01 2.39023134e-01 1.07905591e+00
3.97532195e-01 -3.29910517e-01 -7.61596382e-01 6.16702616e-01
-2.08301887e-01 -1.21857178e+00 -4.53177430e-02 1.18310876e-01
8.67500842e-01 7.70549536e-01 9.65748951e-02 5.07322609e-01
1.96434706e-01 -1.23424518e+00 8.45161796e-01 -4.04045219e-04
6.74000800e-01 -4.48862135e-01 7.11006641e-01 2.61213392e-01
-1.18224359e+00 -2.34790415e-01 -4.58235502e-01 -1.42720759e-01
-4.06233631e-02 6.66608810e-01 -8.19565058e-01 9.46404934e-02
1.26251352e+00 6.05616927e-01 -8.73227119e-01 1.14007962e+00
-3.24167639e-01 8.64878535e-01 -6.42249584e-01 3.61894161e-01
4.39551145e-01 -1.78754225e-01 6.59835711e-02 1.22974360e+00
3.34277421e-01 2.03159302e-01 4.44944590e-01 6.59886658e-01
-1.46068828e-02 -2.31517941e-01 -2.80937821e-01 -7.34045655e-02
6.44736588e-01 1.39169419e+00 -9.33604658e-01 -5.80402389e-02
-3.70950013e-01 5.23708642e-01 2.36223340e-01 3.96851748e-01
-1.04977310e+00 -9.77322385e-02 6.31862342e-01 3.12444299e-01
7.15740800e-01 -4.75552648e-01 -5.30593991e-01 -8.82213771e-01
-2.29726415e-02 -7.02966213e-01 4.92964417e-01 -9.87946689e-01
-1.08051026e+00 4.08482730e-01 2.03038588e-01 -6.80795372e-01
3.61657172e-01 -8.31659257e-01 -6.06769979e-01 9.29779172e-01
-1.83507466e+00 -1.23228991e+00 -8.18030238e-01 2.99170136e-01
4.19448674e-01 1.58570841e-01 3.59650284e-01 1.66420773e-01
-6.37123764e-01 4.10130113e-01 1.19647563e-01 3.12291831e-01
4.99902934e-01 -1.07068574e+00 6.43889427e-01 1.04605234e+00
5.47496267e-02 4.93277967e-01 3.97567809e-01 -4.23062533e-01
-1.22981644e+00 -1.32407403e+00 4.02179569e-01 -2.56790340e-01
5.77148557e-01 -1.44524455e-01 -9.06734884e-01 7.28145301e-01
-3.43331695e-01 2.85061479e-01 4.49859560e-01 1.09590322e-01
-4.14303690e-01 -2.87954360e-01 -9.26701903e-01 2.94943392e-01
1.20749974e+00 -3.37307692e-01 -4.20338482e-01 1.76853254e-01
5.80148816e-01 -6.47836685e-01 -8.13721061e-01 6.07013345e-01
5.53030968e-01 -9.88621056e-01 1.14338338e+00 -1.96672097e-01
5.27332604e-01 -3.41890901e-01 -1.97396368e-01 -1.31696832e+00
-4.30655539e-01 1.12852268e-01 2.16470122e-01 9.90504086e-01
6.29368782e-01 -6.31305993e-01 7.60555923e-01 2.22945407e-01
-5.65542638e-01 -8.03911924e-01 -5.35954773e-01 -7.57935822e-01
8.55509490e-02 -3.19288790e-01 3.37303102e-01 6.77697897e-01
-9.81097341e-01 5.69573283e-01 7.42110386e-02 4.85904217e-01
3.12659860e-01 3.88709038e-01 8.30320060e-01 -1.20722628e+00
-5.55543482e-01 -3.98888320e-01 2.08167285e-02 -1.27081192e+00
-3.62399608e-01 -8.17646146e-01 1.87689066e-01 -2.11456418e+00
1.03557408e-01 -8.15956950e-01 3.98087725e-02 9.47025836e-01
-5.33132434e-01 4.02758688e-01 3.22595030e-01 5.35356045e-01
-1.10662691e-01 3.87966454e-01 1.31697834e+00 4.05151024e-02
-3.70514452e-01 -1.13972232e-01 -8.97374213e-01 7.17669547e-01
1.00124252e+00 -2.98904479e-01 -6.82893634e-01 -1.12700784e+00
1.99270044e-02 -4.71192062e-01 6.02955461e-01 -9.67305601e-01
-3.29755209e-02 -2.20296845e-01 5.21306753e-01 -6.84913158e-01
2.28693336e-01 -6.10096693e-01 -4.19005454e-02 5.09811997e-01
1.70462411e-02 9.90416035e-02 5.53698063e-01 3.90156478e-01
6.37834296e-02 -1.80406049e-01 1.09283781e+00 -1.95825577e-01
-1.16127956e+00 1.74282342e-01 -8.08789879e-02 1.97330013e-01
7.36459494e-01 -4.87738788e-01 -7.05653071e-01 1.53101414e-01
-5.41443944e-01 3.27929437e-01 2.44598389e-01 4.21133310e-01
4.21606153e-01 -6.28158927e-01 -8.47141564e-01 2.61658847e-01
4.85258773e-02 4.80533063e-01 -3.10972016e-02 9.09841359e-01
-7.22814143e-01 2.53402889e-01 -2.60227710e-01 -8.73866558e-01
-1.09566355e+00 4.13517579e-02 6.32883966e-01 -2.08040178e-01
-6.77639782e-01 1.03735781e+00 3.43369484e-01 -5.05239427e-01
1.39717266e-01 -6.18224978e-01 -1.63460463e-01 7.19355866e-02
2.57255465e-01 4.22516048e-01 2.59574860e-01 -3.54210109e-01
-5.42558134e-01 6.62198246e-01 -8.52615386e-02 -5.03085786e-03
1.60979450e+00 -7.86479563e-02 1.17309555e-01 3.75920385e-02
7.86128819e-01 -4.06977206e-01 -1.57575750e+00 -3.15366983e-02
-2.78370291e-01 -3.53556246e-01 5.49331367e-01 -9.64195371e-01
-1.30503798e+00 9.78687346e-01 8.07123780e-01 -1.55535102e-01
1.06197119e+00 6.54646307e-02 5.94853997e-01 4.45352226e-01
3.07867169e-01 -1.03604376e+00 1.05127446e-01 7.95236409e-01
7.56596684e-01 -1.41193199e+00 2.01831907e-01 -4.09392744e-01
-4.80551273e-01 8.18204045e-01 9.24819052e-01 5.15391380e-02
4.92960185e-01 2.81542689e-01 3.84176672e-01 -4.05201614e-01
-3.44414741e-01 -4.52714443e-01 4.01961982e-01 5.39503694e-01
4.68162984e-01 9.22840536e-02 -1.02628261e-01 2.72216558e-01
-2.98332959e-01 -9.24679637e-02 3.62650692e-01 9.84379768e-01
-9.25533235e-01 -6.43742621e-01 -5.32562077e-01 4.67951834e-01
-3.25078368e-01 -4.32179779e-01 -3.41849625e-01 9.39505577e-01
2.34373331e-01 1.03211665e+00 9.74203050e-02 -2.48020187e-01
4.90773708e-01 -2.17590526e-01 4.54467446e-01 -9.70406532e-01
-6.79315090e-01 -6.44207969e-02 1.95858821e-01 -2.43820146e-01
-4.14246500e-01 -5.04777789e-01 -1.35874867e+00 -2.42060199e-01
-6.51222646e-01 7.53523931e-02 7.90177047e-01 9.62299645e-01
4.31596041e-01 4.10192579e-01 2.51861483e-01 -8.84508193e-01
-9.67180789e-01 -1.05541456e+00 -3.77765924e-01 6.77040890e-02
-3.42286974e-02 -5.63405097e-01 -3.62740546e-01 2.41429675e-02] | [9.345931053161621, -0.99358731508255] |
e157efc8-68e7-4951-9d20-6c09a81ee474 | method-for-specifying-location-data | 2304.11926 | null | https://arxiv.org/abs/2304.11926v1 | https://arxiv.org/pdf/2304.11926v1.pdf | Method for Specifying Location Data Requirements for Intralogistics Applications | Various applications leverage location data to increase transparency, efficiency, and safety in intralogistics. There are several properties of location data, such as the data's degrees of freedom, system latency, update rate, or accuracy. To select a suitable indoor localization system, corresponding data requirements must be derived by analyzing the considered application. To date, the dependencies of the system performance and location data requirements have not been satisfactorily described in the literature. Thus, no method exists to adequately derive location data requirements. For intralogistics, such a method is of particular relevance due to the high-cost sensitivity and heterogeneity of partially safety-relevant indoor localization applications. To fill this gap, a method for selecting and quantifying location data requirements for the application in intralogistics is presented in this work, creating substantial added value for warehouse managers and system integrators. The method is based on a spatial model that is built on the premise that location data is used to determine the presence or absence of an entity in a multidimensional interest space. The usage of the method is demonstrated in an exemplary case study for the application of 'Automated Pallet Booking'. | ['Jochen Kreutzfeldt', 'Johannes Hinckeldeyn', 'Markus Knitt', 'Jakob Schyga'] | 2023-04-24 | null | null | null | null | ['indoor-localization'] | ['computer-vision'] | [-2.82745570e-01 -4.28374112e-01 -9.64172930e-02 -4.45859164e-01
-3.83065224e-01 -7.58700132e-01 3.79220068e-01 1.13339078e+00
-3.59064370e-01 7.59784043e-01 -1.62042186e-01 -4.72286105e-01
-9.84223545e-01 -9.12638068e-01 -3.88372749e-01 -4.83791262e-01
-1.93429098e-01 3.66492093e-01 1.81791261e-01 -1.10888466e-01
4.49014992e-01 9.26313579e-01 -1.71849728e+00 -4.21210229e-02
7.26797104e-01 1.19195056e+00 4.15296644e-01 3.93759280e-01
-7.87485391e-02 2.75278956e-01 -8.41563523e-01 2.81832397e-01
2.13421062e-01 -1.69532493e-01 -3.18996549e-01 -1.61798462e-01
-3.24847579e-01 -1.85150459e-01 4.53058451e-01 5.83086610e-01
4.19949263e-01 1.03046268e-01 5.56640446e-01 -1.46099675e+00
-1.73028275e-01 2.71243811e-01 -6.88380376e-02 2.99284965e-01
5.60655534e-01 -1.97550908e-01 6.87671900e-01 -6.61196768e-01
5.18467307e-01 4.99889761e-01 4.38541144e-01 -3.97628784e-01
-1.32588875e+00 -1.83467358e-01 1.22101061e-01 -5.74905910e-02
-1.84560108e+00 -2.28822052e-01 6.57656431e-01 -5.04027724e-01
8.19831312e-01 5.27351856e-01 5.70718169e-01 4.98880982e-01
3.73339742e-01 1.49814442e-01 1.02622712e+00 -6.21779084e-01
8.49981368e-01 7.78509319e-01 6.31280914e-02 -1.60456657e-01
8.17318499e-01 -2.18503132e-01 -4.86358911e-01 -3.35464627e-02
5.95425725e-01 1.04938999e-01 -1.13186135e-03 -9.50672209e-01
-8.95962656e-01 5.42784393e-01 2.11745426e-01 7.95873463e-01
-4.67976213e-01 -2.07171589e-01 2.47484207e-01 6.61952645e-02
2.45670751e-01 4.84950989e-01 -4.52940524e-01 -3.39505732e-01
-1.01891887e+00 1.74782097e-01 1.10916936e+00 1.23295367e+00
6.66885316e-01 -3.18295181e-01 6.60979822e-02 1.01090722e-01
3.63131613e-01 3.64237100e-01 1.47629445e-02 -4.58970189e-01
4.85646933e-01 8.43172789e-01 6.88144267e-01 -1.09045112e+00
-6.51442051e-01 -2.91683257e-01 -3.43225420e-01 -7.12550711e-03
3.85544419e-01 -1.73534438e-01 -2.00847030e-01 1.36993670e+00
4.62866008e-01 -2.81778783e-01 2.14295983e-01 8.53554845e-01
3.94267380e-01 4.92955118e-01 1.05224431e-01 -2.87381351e-01
1.24475622e+00 1.21393092e-01 -9.41891432e-01 1.62966460e-01
8.13169658e-01 -7.65200496e-01 8.14164937e-01 3.10581475e-01
-9.12531495e-01 -4.19903219e-01 -1.07836187e+00 4.31409568e-01
-1.12883282e+00 3.19147885e-01 3.34910721e-01 8.61184597e-01
-7.35043228e-01 2.67892871e-02 -6.64605498e-01 -6.40198290e-01
-2.88197190e-01 4.58554208e-01 -1.32503971e-01 2.19714522e-01
-1.15273833e+00 1.21717918e+00 3.27261627e-01 2.98848778e-01
-1.33477345e-01 -6.86182559e-01 -7.57521689e-01 1.04827836e-01
2.45887637e-01 -2.54505575e-01 8.95334482e-01 -2.05442235e-01
-1.05124581e+00 1.13962866e-01 1.80739537e-01 -3.95983994e-01
6.23557508e-01 5.48108555e-02 -9.56202149e-01 -1.79514512e-01
3.31246942e-01 -3.26539725e-01 4.12053168e-02 -1.30844247e+00
-9.32744324e-01 -4.61065620e-01 1.33620992e-01 5.93878096e-04
-3.75572778e-03 -9.86755937e-02 -8.30936953e-02 -6.98468164e-02
5.77248633e-02 -5.71035862e-01 -1.26376033e-01 -4.89100039e-01
-6.92287311e-02 2.32019454e-01 8.20597589e-01 -3.91028106e-01
1.59821999e+00 -2.19883466e+00 -5.09882569e-01 7.36136615e-01
-2.68675059e-01 3.34657319e-02 4.63308871e-01 1.15296471e+00
2.10744664e-01 -1.52667627e-01 1.05471231e-01 9.43943858e-02
2.89466858e-01 1.21601231e-01 -1.98000580e-01 6.81853354e-01
-1.21605098e-01 4.13000822e-01 -7.47481108e-01 -2.85760254e-01
8.78841639e-01 7.26896524e-01 -1.41419694e-01 -5.50125465e-02
8.39929134e-02 2.45130226e-01 -6.19162858e-01 5.93211234e-01
6.77863717e-01 -4.20783311e-02 4.36673135e-01 -2.23101079e-01
-7.25170255e-01 1.95676044e-01 -1.70697916e+00 1.40263164e+00
-8.99354279e-01 2.59676725e-01 1.05680525e-01 -5.82854688e-01
1.28016067e+00 1.93757787e-01 9.17752624e-01 -6.81370795e-01
1.43331796e-01 5.51577151e-01 -1.07901834e-01 -6.07922912e-01
8.60228419e-01 5.11155009e-01 -4.88686800e-01 4.27478015e-01
-4.00764465e-01 4.12600711e-02 4.19789076e-01 -1.89064875e-01
9.07972276e-01 2.26692423e-01 4.96163547e-01 -4.51105505e-01
6.20173872e-01 1.74237788e-01 1.13435052e-01 4.65670466e-01
-8.35870951e-02 1.79466635e-01 2.79216379e-01 -2.50708014e-01
-7.14599609e-01 -7.89542496e-01 -3.23451430e-01 6.17802501e-01
5.16974986e-01 -4.98119652e-01 -7.49417543e-01 -3.99298429e-01
2.94451892e-01 1.08305573e+00 -3.81325543e-01 3.86812277e-02
-3.50100845e-01 -3.08538646e-01 -3.29169892e-02 3.15540373e-01
2.44560555e-01 -5.59146464e-01 -1.55752254e+00 4.77087200e-01
3.82503830e-02 -1.01619422e+00 1.82569157e-02 3.85936528e-01
-6.87448859e-01 -9.63694930e-01 -9.16044787e-02 -1.44770950e-01
8.34890962e-01 3.32735062e-01 8.13149691e-01 -2.53377169e-01
-2.58610576e-01 5.79044342e-01 -3.83878827e-01 -5.56324661e-01
-1.36848256e-01 3.54870379e-01 1.04562990e-01 3.44535224e-02
6.59719825e-01 -1.63345218e-01 -5.77456594e-01 6.29874110e-01
-9.65947568e-01 -5.46698451e-01 5.58827758e-01 2.58314013e-01
5.85486054e-01 4.91921514e-01 7.92521834e-01 -5.86401582e-01
8.81305575e-01 -6.97177947e-01 -1.07865286e+00 4.22675133e-01
-9.69528735e-01 -1.40786335e-01 6.47698164e-01 -1.91507060e-02
-7.80928493e-01 2.34436631e-01 3.68956655e-01 2.46824905e-01
-6.50895476e-01 7.96136379e-01 -4.06283557e-01 6.80989213e-03
4.40900862e-01 5.84817752e-02 -6.16164953e-02 -4.59548771e-01
1.15238987e-01 7.27890909e-01 2.97080338e-01 -4.43136007e-01
3.15769285e-01 2.49169216e-01 2.16839939e-01 -7.09582388e-01
-8.40825737e-02 -7.33877778e-01 -8.65390956e-01 -2.57178068e-01
4.66033876e-01 -4.27087545e-01 -1.04556739e+00 -3.30073178e-01
-9.64369535e-01 1.29874814e-02 -5.14217257e-01 6.41853392e-01
-6.35082245e-01 -3.11472993e-02 2.97223657e-01 -1.20928288e+00
2.53647417e-01 -1.18064892e+00 9.23424184e-01 -1.58203449e-02
-4.34684694e-01 -1.03668237e+00 -2.24221364e-01 -1.21699877e-01
7.68693984e-01 4.22125131e-01 6.59717143e-01 -8.08841109e-01
-7.78557479e-01 -7.98358679e-01 -4.35178280e-02 -2.09772423e-01
4.92297828e-01 -3.16063464e-02 -5.94765484e-01 -3.56020898e-01
-4.51438455e-03 6.05077267e-01 -2.53000259e-01 4.41711038e-01
7.38837481e-01 -1.42119870e-01 -4.65734541e-01 1.03322789e-01
1.89382327e+00 4.86618668e-01 2.48081103e-01 5.19136310e-01
-9.21844225e-03 9.80191767e-01 1.44655561e+00 8.49156201e-01
4.45880383e-01 1.05926466e+00 5.60307324e-01 4.41557914e-02
2.93814898e-01 -8.27239305e-02 -6.86630281e-03 1.57864138e-01
9.37322602e-02 -3.63344759e-01 -9.70811486e-01 6.30965114e-01
-1.97819531e+00 -5.52976787e-01 -2.57135063e-01 2.58790445e+00
-1.06730126e-01 3.63422781e-02 2.54378706e-01 4.42006528e-01
6.79144740e-01 -1.50638252e-01 -1.85527503e-01 -4.78494197e-01
3.38856071e-01 -3.55430007e-01 1.10357404e+00 3.94276738e-01
-7.87377119e-01 1.83106422e-01 5.79908133e+00 2.74014652e-01
-9.85244691e-01 1.23192770e-02 1.12386949e-01 8.77278969e-02
-1.30843237e-01 9.54795405e-02 -9.40132082e-01 7.76861429e-01
1.34399831e+00 -1.73570514e-01 7.60366023e-02 8.76827240e-01
7.04650760e-01 -6.65764511e-01 -1.16808820e+00 8.76327097e-01
-1.60167918e-01 -1.22980022e+00 -5.24595499e-01 4.56939071e-01
2.99866974e-01 -5.52130282e-01 9.58575159e-02 -2.73142993e-01
-2.50310063e-01 -6.45818889e-01 8.49231303e-01 6.53024316e-01
4.89126265e-01 -1.06049109e+00 9.99307513e-01 4.03648645e-01
-1.24063504e+00 -2.87961513e-01 6.21670820e-02 1.35113271e-02
4.92410243e-01 5.51030874e-01 -1.15690815e+00 8.95739079e-01
5.07939398e-01 2.10384838e-02 -3.63624871e-01 1.25268972e+00
2.14493722e-01 1.35574251e-01 -6.15162015e-01 -4.86554980e-01
1.41789570e-01 -3.32198441e-01 1.95031330e-01 1.19073200e+00
7.29202151e-01 -2.90761173e-01 -1.49754256e-01 9.29897547e-01
8.70209634e-01 4.45919275e-01 -9.39093471e-01 2.22651049e-01
8.78082454e-01 1.16668153e+00 -9.15413976e-01 2.72600114e-01
-4.63261515e-01 3.60144854e-01 -5.27687788e-01 4.53184754e-01
-7.47496963e-01 -5.30811429e-01 4.19142812e-01 6.59443617e-01
2.54573047e-01 -5.57531834e-01 -5.33366561e-01 -2.99596727e-01
1.65980384e-01 -3.10667992e-01 2.66107202e-01 -4.83199865e-01
-7.71669030e-01 3.52851629e-01 4.67191666e-01 -1.57061756e+00
-5.67420542e-01 -6.42106473e-01 6.74669892e-02 1.03316009e+00
-1.24580288e+00 -1.15248191e+00 -5.17581165e-01 3.47338378e-01
8.35093961e-04 -1.42402932e-01 9.40186620e-01 4.44796383e-01
-3.77506852e-01 1.81139424e-01 3.24912310e-01 -4.95200068e-01
5.18526077e-01 -1.29027510e+00 5.93869854e-03 8.66497874e-01
-2.41006255e-01 8.74411225e-01 9.89889383e-01 -6.79485321e-01
-1.66333044e+00 -8.02437365e-01 1.18198609e+00 -5.87361157e-01
6.38593495e-01 -5.57689726e-01 -4.37795669e-01 4.68536109e-01
-3.07467431e-01 1.25708118e-01 8.73489559e-01 -4.57639620e-02
2.32932612e-01 -6.93390906e-01 -1.49402022e+00 3.21576148e-01
3.62130791e-01 -4.62166160e-01 -1.57878116e-01 3.60097475e-02
8.89082178e-02 -3.96031946e-01 -1.25041413e+00 1.75676122e-01
4.08725947e-01 -1.06580627e+00 9.47641373e-01 3.06042060e-02
-2.71717429e-01 -6.08445227e-01 -5.11187136e-01 -9.05890584e-01
-1.91726647e-02 -3.08232188e-01 -8.61315578e-02 1.52444434e+00
4.57878649e-01 -8.75825346e-01 6.19741678e-01 1.20810044e+00
1.10819437e-01 -6.78217292e-01 -1.00352061e+00 -8.86466324e-01
-6.54721677e-01 -3.45967203e-01 1.28568304e+00 6.81589961e-01
1.72202047e-02 -1.17266685e-01 7.45017752e-02 6.00238919e-01
2.00661317e-01 2.03681082e-01 9.58406627e-01 -1.30634916e+00
2.88695633e-01 -3.33867297e-02 -6.84351981e-01 -4.88268912e-01
-4.26905900e-01 -3.92786026e-01 -1.42654479e-01 -1.86215127e+00
-6.60039425e-01 -9.97541249e-01 -4.41712737e-01 3.03954352e-02
7.66393542e-01 -2.36447215e-01 4.02475670e-02 2.12223709e-01
-4.13738906e-01 -1.42191753e-01 4.65840399e-01 2.76577532e-01
-6.82253718e-01 4.15658891e-01 -5.13919711e-01 2.49020636e-01
8.29172611e-01 -2.65927583e-01 -7.87464738e-01 -9.60534513e-02
3.57946992e-01 1.41257882e-01 3.12717974e-01 -1.05562615e+00
5.64691901e-01 -3.42582822e-01 1.98729068e-01 -9.48378444e-01
4.51983362e-01 -1.79530585e+00 8.26811373e-01 3.74411494e-01
-1.28815651e-01 3.52353275e-01 7.02130347e-02 4.49114442e-01
-2.42679819e-01 -4.05854970e-01 1.37133271e-01 3.25331986e-01
-9.28807199e-01 -1.50867358e-01 -3.73627454e-01 -7.00142324e-01
1.59121418e+00 -7.12986231e-01 -2.02864110e-01 -1.60263196e-01
-4.45587933e-01 -1.37579944e-02 5.47258198e-01 3.45311671e-01
4.24660593e-01 -1.14149475e+00 -1.36353552e-01 4.04005051e-01
3.71656865e-01 -1.55043662e-01 1.42324179e-01 9.74125803e-01
-5.73198378e-01 8.62157762e-01 -2.19963595e-01 -6.07369602e-01
-9.44025099e-01 7.92251348e-01 7.89089352e-02 9.71325487e-02
-2.02902451e-01 -1.02072701e-01 -3.34329993e-01 -8.03597867e-02
7.52343237e-02 -6.84909940e-01 -1.98181838e-01 3.20137680e-01
3.00708741e-01 6.14888787e-01 5.26146412e-01 -7.05651224e-01
-7.54325986e-01 3.82092983e-01 4.12751108e-01 -1.63234681e-01
1.25311124e+00 -6.77356303e-01 -7.37839565e-02 6.28905058e-01
7.47641623e-01 2.73531079e-01 -9.57848191e-01 2.31293976e-01
4.15511429e-01 -6.96469665e-01 -3.92755792e-02 -8.58170569e-01
-5.85800767e-01 3.05842161e-01 8.05358768e-01 8.78570616e-01
1.13500726e+00 -3.38030547e-01 8.61909986e-02 1.27803441e-02
9.18336093e-01 -1.49765146e+00 -5.82173467e-01 5.36654964e-02
7.70377219e-01 -8.61360669e-01 5.78216240e-02 -4.08978760e-01
-3.49078804e-01 1.02044570e+00 2.38292411e-01 2.75389969e-01
8.96505415e-01 6.83352113e-01 -5.86988591e-02 -1.91874072e-01
-1.55807391e-01 -1.15356378e-01 -1.58481166e-01 7.75049388e-01
4.37237710e-01 3.00006986e-01 -7.40832150e-01 8.01340759e-01
-9.87908095e-02 1.69072837e-01 4.15506691e-01 1.29658747e+00
-2.84086883e-01 -1.06040359e+00 -6.56743526e-01 1.39829993e-01
-1.83765575e-01 4.46388334e-01 -1.27802581e-01 1.23125589e+00
3.58643204e-01 1.18290818e+00 2.61942856e-02 -1.45834759e-01
8.17904413e-01 -2.90173709e-01 1.41885743e-01 -4.49806035e-01
-5.61750770e-01 -1.01881906e-01 2.01046988e-01 -4.05611783e-01
-2.44073734e-01 -8.64243746e-01 -1.24111438e+00 -2.83610165e-01
-4.86835212e-01 5.02905548e-01 1.48491144e+00 6.65963948e-01
6.37650728e-01 6.27050042e-01 5.99287868e-01 -5.55420995e-01
-4.45170909e-01 -4.42083150e-01 -9.06566620e-01 9.51847360e-02
1.27050325e-01 -8.80362868e-01 -1.26866192e-01 -3.49959135e-01] | [6.374463081359863, 1.6939343214035034] |
6233931a-78ad-4b46-a007-5f6ea1374345 | unixlong-at-semeval-2020-task-6-a-joint-model | null | null | https://aclanthology.org/2020.semeval-1.96 | https://aclanthology.org/2020.semeval-1.96.pdf | UNIXLONG at SemEval-2020 Task 6: A Joint Model for Definition Extraction | Definition Extraction is the task to automatically extract terms and their definitions from text. In recent years, it attracts wide interest from NLP researchers. This paper describes the unixlong team{'}s system for the SemEval 2020 task6: DeftEval: Extracting term-definition pairs in free text. The goal of this task is to extract definition, word level BIO tags and relations. This task is challenging due to the free style of the text, especially the definitions of the terms range across several sentences and lack explicit verb phrases. We propose a joint model to train the tasks of definition extraction and the word level BIO tagging simultaneously. We design a creative format input of BERT to capture the location information between entity and its definition. Then we adjust the result of BERT with some rules. Finally, we apply TAG{\_}ID, ROOT{\_}ID, BIO tag to predict the relation and achieve macro-averaged F1 score 1.0 which rank first on the official test set in the relation extraction subtask. | ['Jianping Shen', 'Mo Yang', 'Jiang Lianxin', 'Haiqin Yang', 'Jian Ma', 'Shuyi Xie'] | 2020-12-01 | null | null | null | semeval-2020 | ['definition-extraction'] | ['natural-language-processing'] | [ 3.63278896e-01 3.42792064e-01 -4.12186801e-01 -4.55861241e-01
-8.01995814e-01 -7.20448732e-01 7.32410431e-01 4.09468323e-01
-6.32608831e-01 1.22894001e+00 3.33580017e-01 -4.53188330e-01
-2.86034733e-01 -8.94974709e-01 -3.76678318e-01 -4.31063920e-01
-4.08412665e-02 7.24979043e-01 1.81429222e-01 -1.77952856e-01
-6.00664467e-02 1.72798395e-01 -1.08719528e+00 4.29656118e-01
9.72641468e-01 9.27623212e-01 2.54873902e-01 3.34052205e-01
-4.72020507e-01 5.49119651e-01 -8.08091700e-01 -6.22416735e-01
6.75153278e-04 -2.83742189e-01 -1.26054847e+00 -5.68593144e-01
1.73904493e-01 3.99234682e-01 -1.25819042e-01 1.04095078e+00
5.09892762e-01 9.46672261e-02 8.21883678e-01 -9.01478648e-01
-3.27173769e-01 1.32890427e+00 -5.01386046e-01 3.34507465e-01
3.46309394e-01 -2.92644829e-01 1.38829637e+00 -8.32549751e-01
9.51567292e-01 8.76172304e-01 5.91205895e-01 5.62154233e-01
-1.02717495e+00 -6.73372805e-01 4.39482033e-02 1.38197377e-01
-1.61665714e+00 -2.94768274e-01 2.10259303e-01 -3.47958297e-01
1.49849391e+00 5.19296288e-01 3.32668632e-01 9.29399073e-01
2.63622552e-01 5.77873349e-01 7.94660687e-01 -6.64158881e-01
-9.12910700e-02 -1.38095573e-01 5.16685247e-01 6.57983422e-01
3.05241078e-01 -3.20149630e-01 -6.53565288e-01 -5.69888912e-02
2.23038375e-01 -3.91899794e-01 -3.15318465e-01 5.10250747e-01
-1.28664291e+00 3.86752278e-01 2.79088527e-01 6.51904583e-01
-3.66595417e-01 1.68130755e-01 3.52060735e-01 -3.22685950e-02
5.46092808e-01 8.53639841e-01 -1.08161080e+00 -6.61911536e-03
-7.35120058e-01 3.05226058e-01 8.39062631e-01 1.44439864e+00
5.45995414e-01 -8.32197368e-01 -6.11603200e-01 1.01643384e+00
1.93168342e-01 3.59770268e-01 4.51747566e-01 -3.38615030e-01
7.42063403e-01 6.80236399e-01 1.42617766e-02 -6.28375113e-01
-6.33873880e-01 -7.20146596e-01 -7.01211274e-01 -6.94387674e-01
2.17034966e-01 -6.51112020e-01 -1.06437063e+00 1.66520512e+00
3.80130321e-01 1.87722728e-01 -7.46467933e-02 4.11855102e-01
1.51172173e+00 4.72337782e-01 4.77628022e-01 -4.08580363e-01
1.85942805e+00 -7.90719032e-01 -1.06981981e+00 -4.56268311e-01
1.03899419e+00 -7.21253991e-01 5.39461255e-01 5.26607549e-03
-6.13157988e-01 -3.20919603e-01 -9.43606436e-01 -1.12813950e-01
-8.10832858e-01 2.27195323e-01 7.35589564e-01 1.67445615e-01
-4.17350233e-01 4.81321961e-01 -5.27541995e-01 -2.02871636e-01
1.60697594e-01 4.67987895e-01 -4.54983205e-01 1.01041541e-01
-1.87538564e+00 1.20747149e+00 8.92203867e-01 1.65621683e-01
-1.41610920e-01 -8.79204810e-01 -9.02855515e-01 2.64593586e-02
4.79266107e-01 -8.69117200e-01 1.18028998e+00 -3.35482955e-01
-6.17328107e-01 1.15955436e+00 -2.30271071e-01 -5.02617419e-01
-7.79233947e-02 -3.65211517e-01 -6.83454335e-01 -3.39061946e-01
7.42642820e-01 3.26568633e-01 9.97416005e-02 -7.87154794e-01
-1.01122713e+00 -2.92758882e-01 -3.36008556e-02 7.36971945e-02
6.52731061e-02 2.07853645e-01 -7.21116245e-01 -6.09068394e-01
-6.61910400e-02 -6.14216745e-01 -2.49290347e-01 -8.13266814e-01
-7.76070774e-01 -9.53141630e-01 1.31829426e-01 -7.41056323e-01
1.83192897e+00 -1.78287184e+00 -9.76477340e-02 2.47160107e-01
4.59256500e-01 3.03182423e-01 1.54311344e-01 5.09031475e-01
-2.60791481e-01 3.94960225e-01 -1.66383758e-01 -2.55231336e-02
1.08145632e-01 1.48345575e-01 -2.68781390e-02 -3.72788869e-02
1.55505151e-01 1.04114616e+00 -1.15236568e+00 -8.09086502e-01
-1.71828613e-01 1.28199890e-01 -2.44550291e-04 2.06845403e-01
-6.07302308e-01 -1.42245427e-01 -7.43352473e-01 4.70464855e-01
4.68435973e-01 -2.46148512e-01 5.34908950e-01 -3.51045787e-01
-8.68102610e-02 1.01061749e+00 -7.53435135e-01 1.60950017e+00
-3.23166251e-01 4.22944903e-01 -2.14492589e-01 -7.22047448e-01
8.13896775e-01 4.45375800e-01 5.60622394e-01 -3.96193445e-01
9.60881934e-02 2.46159524e-01 1.04490623e-01 -6.79727852e-01
2.81665325e-01 -2.09750891e-01 -4.39251751e-01 -6.55096695e-02
8.05132389e-02 -1.13065749e-01 4.89270180e-01 2.42572948e-01
1.64365411e+00 1.95085377e-01 6.45456493e-01 -2.80831665e-01
5.35776615e-01 8.53356719e-02 1.00470328e+00 6.31620944e-01
1.53166130e-01 2.57220924e-01 5.04265368e-01 -1.75012171e-01
-3.68427932e-01 -9.46409047e-01 -4.09650415e-01 8.33244145e-01
-2.26099305e-02 -1.02481735e+00 -5.83721697e-01 -1.13617384e+00
-1.11998141e-01 8.72921824e-01 -6.95422411e-01 1.04205333e-01
-8.22319686e-01 -1.01823664e+00 7.31598616e-01 3.13234150e-01
6.98833048e-01 -1.15107095e+00 -3.86968702e-02 4.36111867e-01
-6.49355769e-01 -1.31524718e+00 -5.02809346e-01 5.56855798e-01
-8.75784010e-02 -1.01778316e+00 -2.89299816e-01 -1.14263153e+00
2.89533705e-01 -6.08024776e-01 1.47281659e+00 4.41274419e-02
-2.28822470e-01 -4.49831098e-01 -3.86724561e-01 -6.21531963e-01
4.45113257e-02 5.60043633e-01 -2.21369341e-01 -3.37854892e-01
9.63476479e-01 -3.34872603e-01 -4.26319391e-01 3.51178041e-03
-6.07218981e-01 1.16210632e-01 5.55077553e-01 5.54221690e-01
7.05978513e-01 1.36081189e-01 8.08961272e-01 -1.31432974e+00
6.81626320e-01 -4.86377776e-01 -2.14035094e-01 4.91616696e-01
-8.55244696e-01 1.85061410e-01 2.75847584e-01 -1.79025173e-01
-9.52483356e-01 1.91095054e-01 -4.61294442e-01 5.70708632e-01
-2.55781952e-02 9.34913099e-01 -6.59788013e-01 6.04296267e-01
5.62666833e-01 1.19632566e-02 -7.02791333e-01 -5.79052031e-01
3.75471801e-01 9.63651717e-01 4.95265841e-01 -5.28583050e-01
5.91820419e-01 -3.32338274e-01 -6.39898703e-02 -5.93889534e-01
-1.49839222e+00 -6.68811321e-01 -7.41224170e-01 1.67654365e-01
1.01980400e+00 -6.98635817e-01 -4.42484409e-01 1.24200612e-01
-1.43169594e+00 -2.58207589e-01 3.28907296e-02 4.76471603e-01
5.94015792e-02 1.26774952e-01 -5.74877501e-01 -4.97789592e-01
-7.19184816e-01 -4.73410100e-01 9.70145643e-01 2.68908322e-01
-6.82319045e-01 -9.55772161e-01 1.90447241e-01 3.61836106e-01
-1.00165002e-01 4.17320281e-01 1.13202310e+00 -8.83129120e-01
-1.90475538e-01 -1.61904544e-01 -3.31953347e-01 -1.14664920e-01
3.41250718e-01 -1.86902612e-01 -5.33408403e-01 4.50309575e-01
-4.86028045e-01 1.38648793e-01 1.01282918e+00 1.33501530e-01
1.00839090e+00 -4.59895521e-01 -9.71445203e-01 3.01627010e-01
1.17140532e+00 1.30649775e-01 4.60185319e-01 3.00401032e-01
6.49124563e-01 6.29227042e-01 9.11268353e-01 1.26887470e-01
5.04996002e-01 8.28220367e-01 3.93009633e-02 -4.17171009e-02
-1.51654497e-01 -2.48694286e-01 4.14389074e-02 5.10598063e-01
1.72362197e-02 -4.42788839e-01 -1.14380717e+00 8.05104077e-01
-1.79300070e+00 -8.23136568e-01 -4.75160688e-01 1.83468199e+00
1.61322737e+00 3.16453159e-01 -2.16063663e-01 -1.00058392e-01
6.78141415e-01 -2.85983384e-02 -2.00000238e-02 -1.50637344e-01
-1.17032260e-01 7.15580881e-01 5.92731774e-01 6.35479450e-01
-1.28383434e+00 1.32374179e+00 5.38170910e+00 1.29950953e+00
-5.94280303e-01 2.52839383e-02 3.79016012e-01 2.22028479e-01
-4.03580844e-01 6.03575371e-02 -1.45276868e+00 6.27974749e-01
9.73185062e-01 -2.04235598e-01 -1.21000484e-01 1.73410520e-01
-2.88288798e-02 7.68779069e-02 -1.20338464e+00 7.37712443e-01
-6.05707578e-02 -1.20078170e+00 -8.54431689e-02 1.47488236e-01
3.44434321e-01 4.11432758e-02 -5.38767934e-01 6.47729456e-01
3.62946332e-01 -1.33640432e+00 1.75052300e-01 6.37029111e-01
9.31888640e-01 -5.06971896e-01 1.09589612e+00 4.29927617e-01
-1.36469495e+00 4.51322019e-01 -9.78005379e-02 9.27579179e-02
3.08240205e-01 1.14533031e+00 -1.10426533e+00 6.81829154e-01
3.52565378e-01 5.36365151e-01 -4.58161324e-01 8.84625793e-01
-8.26593518e-01 6.90348029e-01 -4.98652488e-01 -4.92693812e-01
-4.59055193e-02 8.42153281e-02 3.83776665e-01 1.77488291e+00
-1.14566935e-02 5.03165722e-01 3.52593549e-02 7.39924729e-01
-4.78193402e-01 2.85349965e-01 -4.24816579e-01 -1.52825713e-01
7.91436255e-01 1.36755872e+00 -5.35893023e-01 -3.53023797e-01
-1.62301973e-01 7.26788104e-01 5.10926664e-01 2.31209487e-01
-8.92560363e-01 -8.81650388e-01 4.61680382e-01 -3.46499868e-02
1.21782631e-01 -1.15097895e-01 -4.65557307e-01 -1.10755014e+00
-2.22514290e-02 -5.49481571e-01 6.39027476e-01 -5.61105490e-01
-1.26291966e+00 6.78096414e-01 4.60365675e-02 -6.64075971e-01
-3.14481109e-01 -5.92926621e-01 -3.40231597e-01 9.56908226e-01
-1.10427451e+00 -1.27736950e+00 3.19218151e-02 1.10923365e-01
3.53335708e-01 9.16953757e-02 1.04257286e+00 5.44547081e-01
-6.45803094e-01 8.44114363e-01 -3.44771087e-01 8.17870140e-01
7.97606826e-01 -1.51596546e+00 4.90321428e-01 6.54144049e-01
3.10571343e-01 1.07895291e+00 8.04530323e-01 -1.03094971e+00
-7.50763774e-01 -1.21830702e+00 2.19378924e+00 -7.07583010e-01
6.82381928e-01 -4.99285311e-01 -5.55401921e-01 7.32875168e-01
2.62097627e-01 -2.14652136e-01 9.77015376e-01 6.67735219e-01
-1.90374792e-01 -1.17047571e-01 -1.01491976e+00 5.11869192e-01
1.24592924e+00 -3.38666201e-01 -9.09300447e-01 6.27046824e-01
9.81631875e-01 -3.90136302e-01 -1.19473279e+00 6.11248612e-01
5.62745512e-01 -1.28765911e-01 8.35693479e-01 -7.01196134e-01
4.20352459e-01 -4.00998712e-01 9.18457136e-02 -1.08132935e+00
-3.65838051e-01 -6.49461746e-01 -1.21624045e-01 1.80151963e+00
1.38708949e+00 -3.28416556e-01 5.18387973e-01 7.16905475e-01
-1.00951968e-02 -9.48901653e-01 -9.31498349e-01 -5.01631856e-01
5.96447401e-02 -3.00267935e-01 5.45623541e-01 9.80825245e-01
3.75366509e-01 1.23940516e+00 -3.78222317e-02 4.18900400e-02
3.13483149e-01 2.50451893e-01 3.09040099e-01 -1.12391150e+00
-3.31286728e-01 -1.60283625e-01 -1.28138766e-01 -1.14418793e+00
3.29660743e-01 -1.11991858e+00 3.02972287e-01 -1.92778051e+00
2.42599592e-01 -6.07988358e-01 -3.78172457e-01 8.59072566e-01
-6.12939298e-01 -1.93187781e-03 -2.73109406e-01 -2.20161334e-01
-6.77552760e-01 2.73018301e-01 8.78465474e-01 -2.23389283e-01
-1.52221754e-01 8.04316401e-02 -8.06648850e-01 4.43586588e-01
6.93069756e-01 -6.16673589e-01 -6.16579354e-02 -4.10881549e-01
4.11062539e-01 -2.27652758e-01 -1.70924872e-01 -4.16323662e-01
-1.62594754e-03 -2.73694783e-01 2.70703018e-01 -7.19403923e-01
8.11580271e-02 -5.48217714e-01 5.37605770e-02 1.95420608e-01
-4.63303834e-01 -9.85272005e-02 -6.28424063e-02 1.81728348e-01
-2.23553535e-02 -4.28603828e-01 3.80744278e-01 -1.82113439e-01
-3.82533252e-01 4.08420533e-01 -2.19106182e-01 2.94771016e-01
6.89601004e-01 4.08090979e-01 -3.87521952e-01 1.38528109e-01
-8.23685110e-01 6.33911431e-01 -2.73194224e-01 3.25940043e-01
2.74803579e-01 -1.19103003e+00 -9.35863018e-01 -3.14586788e-01
2.11332798e-01 3.79549712e-01 -4.01620597e-01 6.99642062e-01
-3.16651523e-01 5.56229413e-01 4.68058020e-01 -4.02045175e-02
-1.51789737e+00 3.90451312e-01 4.30904776e-01 -8.08661938e-01
-4.38036114e-01 1.34668767e+00 1.75918475e-01 -5.24847925e-01
4.07261103e-02 -3.60838443e-01 -5.39901733e-01 7.53053799e-02
6.12572014e-01 1.09124165e-02 6.64780363e-02 -3.18797052e-01
-8.15671027e-01 2.21116975e-01 -1.90966904e-01 -1.78670779e-01
1.43432069e+00 -3.35619375e-02 -5.39192080e-01 2.53622413e-01
1.17754698e+00 2.04385266e-01 -2.66159654e-01 -3.66856903e-01
7.72382796e-01 8.24235156e-02 -1.14883885e-01 -1.26134765e+00
-5.83501935e-01 3.70742440e-01 8.94309580e-02 9.84860957e-02
7.65184879e-01 3.73302698e-01 9.73827064e-01 2.86332726e-01
1.28413036e-01 -1.24759126e+00 -5.50315201e-01 8.76940668e-01
9.05847132e-01 -9.15084541e-01 1.14041969e-01 -9.83617783e-01
-9.68494490e-02 7.54290283e-01 6.34487331e-01 2.23856613e-01
7.11938977e-01 4.64139044e-01 -5.41288070e-02 -4.35089678e-01
-8.68584692e-01 -6.85869515e-01 5.23246229e-01 5.55977643e-01
1.04528117e+00 1.83554694e-01 -1.19228601e+00 1.05322039e+00
-5.04224479e-01 -1.94916919e-01 -1.72261164e-01 5.67551076e-01
-3.81019205e-01 -1.55651510e+00 2.70744532e-01 7.91819096e-01
-9.93809819e-01 -7.63492227e-01 -8.11786652e-01 5.12252331e-01
7.93600082e-01 1.16962445e+00 -3.63686115e-01 -5.05488634e-01
2.87857056e-01 2.34384820e-01 4.13724899e-01 -1.16270947e+00
-8.32781017e-01 3.20293680e-02 1.00321913e+00 -9.83788967e-02
-3.52158278e-01 -6.72446072e-01 -1.76832998e+00 1.28005683e-01
-6.20776951e-01 7.65213072e-01 3.15364927e-01 1.43149960e+00
2.64165938e-01 7.03611434e-01 2.25469425e-01 1.94977596e-01
-1.99605450e-01 -1.45459485e+00 -5.99199772e-01 2.66070306e-01
-4.46537212e-02 -4.89353448e-01 -1.32463709e-01 4.23946455e-02] | [9.49594783782959, 8.825020790100098] |
79fae7a2-ddb5-4731-9ea3-7a6ace7d26d7 | benchmarking-graphormer-on-large-scale | 2203.04810 | null | https://arxiv.org/abs/2203.04810v2 | https://arxiv.org/pdf/2203.04810v2.pdf | Benchmarking Graphormer on Large-Scale Molecular Modeling Datasets | This technical note describes the recent updates of Graphormer, including architecture design modifications, and the adaption to 3D molecular dynamics simulation. With these simple modifications, Graphormer could attain better results on large-scale molecular modeling datasets than the vanilla one, and the performance gain could be consistently obtained on 2D and 3D molecular graph modeling tasks. In addition, we show that with a global receptive field and an adaptive aggregation strategy, Graphormer is more powerful than classic message-passing-based GNNs. Empirically, Graphormer could achieve much less MAE than the originally reported results on the PCQM4M quantum chemistry dataset used in KDD Cup 2021. In the meanwhile, it greatly outperforms the competitors in the recent Open Catalyst Challenge, which is a competition track on NeurIPS 2021 workshop, and aims to model the catalyst-adsorbate reaction system with advanced AI models. All codes could be found at https://github.com/Microsoft/Graphormer. | ['Tie-Yan Liu', 'Di He', 'Chang Liu', 'Shengjie Luo', 'Jiyan He', 'Jiacheng You', 'Yifei Shen', 'Guolin Ke', 'Shuxin Zheng', 'Yu Shi'] | 2022-03-09 | null | null | null | null | ['graph-regression'] | ['graphs'] | [-1.67723030e-01 4.81787659e-02 -1.82649329e-01 1.77811552e-02
-4.16935831e-01 -5.04639685e-01 6.04109108e-01 4.94075298e-01
-5.40677547e-01 9.66404021e-01 -2.17625901e-01 -7.35144317e-01
-4.15925384e-02 -7.35405505e-01 -9.65329230e-01 -8.42747867e-01
-3.46191198e-01 8.73458147e-01 1.20214343e-01 -4.60744292e-01
4.33982670e-01 7.40799367e-01 -1.05490518e+00 3.37014914e-01
7.74041772e-01 6.48343623e-01 3.66579711e-01 7.61547804e-01
-3.83349210e-02 8.54773164e-01 -1.09943688e-01 -2.60336339e-01
1.31011069e-01 -6.25229552e-02 -9.36012506e-01 -6.99772239e-01
4.56339270e-01 3.73692125e-01 -7.25084305e-01 8.58486593e-01
8.70372117e-01 4.65561122e-01 5.75417995e-01 -9.45937514e-01
-6.09999478e-01 6.67467415e-01 -1.15730584e-01 3.08107018e-01
2.29571402e-01 4.59528804e-01 9.70904648e-01 -9.41617072e-01
8.21057439e-01 1.35248697e+00 5.62367916e-01 7.07672477e-01
-1.36396110e+00 -6.63245320e-01 4.01556373e-01 5.69043636e-01
-1.46559763e+00 -2.69016773e-01 2.60936320e-01 -1.41037613e-01
1.79680300e+00 1.31664753e-01 8.27403069e-01 1.13374221e+00
7.04086602e-01 5.80215573e-01 7.48381555e-01 -1.31950513e-01
5.82255840e-01 -4.95507330e-01 3.56699824e-01 9.00274396e-01
4.57436562e-01 9.76741761e-02 -5.29865503e-01 -7.04867601e-01
6.62020624e-01 2.24985089e-03 -1.63973361e-01 -4.25744653e-01
-1.16528451e+00 9.53303576e-01 8.03377330e-01 2.50790894e-01
-6.71000659e-01 7.14141548e-01 2.71711856e-01 2.34993592e-01
4.32869554e-01 7.80928135e-01 -3.09259534e-01 3.15187052e-02
-4.58993256e-01 6.67432964e-01 1.03946161e+00 9.42940176e-01
6.11866057e-01 3.83384377e-02 -9.36620124e-03 3.31412822e-01
4.15400833e-01 3.27252239e-01 1.26864165e-01 -8.24585617e-01
4.17443275e-01 3.99308443e-01 5.36174094e-03 -4.95093018e-01
-9.34924841e-01 -4.78085607e-01 -1.00153351e+00 1.13841474e-01
1.56393185e-01 -1.78066105e-01 -1.11963582e+00 1.49921453e+00
2.98177302e-01 1.43532306e-01 2.71339089e-01 7.15288758e-01
1.14074242e+00 8.13935339e-01 3.86459351e-01 -2.57357925e-01
9.84286427e-01 -1.30719948e+00 -6.33061767e-01 -1.04496598e-01
1.05651307e+00 -4.62574810e-01 3.97519022e-01 6.52470291e-01
-1.29517996e+00 -2.17408925e-01 -1.12913322e+00 -1.43639162e-01
-8.57424438e-01 -3.17628741e-01 1.46676064e+00 5.00336707e-01
-1.14347589e+00 1.00799060e+00 -1.03979993e+00 -5.09563684e-01
3.58723313e-01 8.87561381e-01 -4.57535207e-01 -2.42628530e-01
-1.34110057e+00 1.04582465e+00 5.61532557e-01 9.48091596e-02
-1.10457909e+00 -8.09661746e-01 -4.76240247e-01 -3.61637473e-02
2.66692996e-01 -1.07037187e+00 1.17913568e+00 -1.59573570e-01
-1.52886581e+00 3.61876041e-01 -1.47868410e-01 -9.31612015e-01
1.24557547e-01 3.50508302e-01 -1.05241627e-01 -1.43802240e-01
-4.38426763e-01 7.50991166e-01 1.67060852e-01 -6.74845874e-01
-9.33102146e-02 -4.20800418e-01 1.33385912e-01 4.20579016e-01
3.16443406e-02 -3.85402530e-01 -3.45010102e-01 -9.99559388e-02
-4.01911773e-02 -1.01985681e+00 -1.06674898e+00 -4.92705524e-01
-3.16348433e-01 -4.87908483e-01 1.25160068e-01 -1.11483306e-01
1.02831662e+00 -1.55137885e+00 6.77513063e-01 4.41554457e-01
8.63034606e-01 3.40531945e-01 -2.69761175e-01 8.34375381e-01
-2.97137171e-01 1.84203405e-02 -4.87008169e-02 -4.54252101e-02
5.26289046e-02 2.21704040e-02 1.96198478e-01 4.64842856e-01
-3.51277560e-01 1.22854877e+00 -8.36149633e-01 -3.94780114e-02
4.39248979e-02 4.43780333e-01 -6.77954793e-01 -2.85009801e-01
-6.19674802e-01 9.90460888e-02 -5.86741090e-01 5.59099615e-01
8.42646778e-01 -8.57543409e-01 4.97163296e-01 -1.27711505e-01
-9.80399027e-02 2.47516468e-01 -6.94564700e-01 2.19979477e+00
1.09239286e-02 1.40658051e-01 4.50535379e-02 -9.01879191e-01
7.26744294e-01 2.99934596e-01 6.34228289e-01 -8.97643447e-01
-6.27666116e-02 2.48947039e-01 4.27621484e-01 4.71896678e-02
3.08667660e-01 -7.52215367e-03 3.67560893e-01 1.61359936e-01
2.31055841e-01 -2.54842669e-01 4.37958688e-01 5.93843162e-01
1.44854319e+00 -5.02141304e-02 2.11067557e-01 -5.20789266e-01
4.42637533e-01 2.16297671e-01 1.61979824e-01 1.01078939e+00
-2.71447059e-02 2.30628222e-01 4.13970649e-01 -9.30913091e-01
-9.28228140e-01 -7.18794465e-01 1.92951839e-02 1.00435042e+00
7.61807114e-02 -1.22282755e+00 -7.17089951e-01 -4.10976678e-01
1.90967932e-01 6.67664766e-01 -5.65246761e-01 -2.68921047e-01
-2.54947484e-01 -1.25352657e+00 4.78067875e-01 2.12953493e-01
3.83921117e-02 -9.16671753e-01 2.23278254e-01 6.47935748e-01
3.60071659e-01 -7.04915345e-01 -2.67761588e-01 7.02287376e-01
-9.46873784e-01 -1.15047574e+00 -6.29269183e-01 -3.88670832e-01
2.70985037e-01 4.27673042e-01 1.06166470e+00 8.70093480e-02
-4.03377265e-01 -5.35733588e-02 -7.29122609e-02 -5.94303787e-01
-3.59736681e-01 6.02850854e-01 8.27196836e-02 -6.27136350e-01
3.90221983e-01 -5.80355287e-01 -8.33266675e-01 -6.75247684e-02
-7.16797888e-01 2.60459185e-01 3.92658651e-01 6.59717083e-01
6.22822464e-01 -3.66773069e-01 3.60304266e-01 -9.91378546e-01
8.50177586e-01 -5.80412269e-01 -5.89556217e-01 2.10524440e-01
-1.02971828e+00 3.74533504e-01 7.17524409e-01 -1.23104222e-01
-4.52981681e-01 1.02726214e-01 -4.24123734e-01 -3.48814964e-01
1.62463531e-01 7.53092468e-01 2.06465751e-01 -5.45597434e-01
1.00475597e+00 1.50455698e-01 3.11094001e-02 -4.31276649e-01
3.98405939e-01 1.83338910e-01 -2.05792770e-01 -8.26954067e-01
3.52362633e-01 2.08717167e-01 5.43092966e-01 -6.81609392e-01
-2.76858121e-01 -2.72901177e-01 -2.06584290e-01 9.15815830e-02
7.50347793e-01 -9.86836851e-01 -1.19158423e+00 3.33267242e-01
-1.33923042e+00 -4.97057587e-01 1.39829352e-01 6.15302622e-01
-5.24221480e-01 2.86799878e-01 -7.86702454e-01 -5.08770943e-01
-6.89751029e-01 -1.33496284e+00 7.49861240e-01 9.59261321e-03
7.02061579e-02 -1.09838212e+00 4.03816968e-01 2.79329479e-01
6.48539424e-01 9.06870961e-02 1.15371776e+00 -7.72345960e-01
-6.43323720e-01 -2.24244446e-01 2.21090112e-02 -1.81849658e-01
-2.24327266e-01 -8.35332945e-02 -7.55754173e-01 -5.07006526e-01
-6.68288589e-01 -4.26687777e-01 1.36590934e+00 5.59969902e-01
9.74776506e-01 -6.31692484e-02 -6.63873911e-01 6.94553316e-01
1.10736799e+00 3.87921304e-01 6.96792543e-01 1.70945436e-01
8.80944252e-01 -5.11703044e-02 8.18158388e-02 2.52434343e-01
3.93655092e-01 6.11466289e-01 8.95461917e-01 -1.83537230e-01
5.53139709e-02 -8.16111490e-02 2.72644937e-01 9.34382737e-01
-7.71523654e-01 -5.39704323e-01 -1.09278798e+00 -2.58118093e-01
-2.26595187e+00 -9.32593644e-01 -3.18592548e-01 1.98718858e+00
4.69541132e-01 -8.78180712e-02 1.13528468e-01 -5.07495821e-01
5.35700798e-01 1.93078965e-01 -9.88241553e-01 -6.50984526e-01
-3.12059790e-01 6.46101892e-01 7.90593863e-01 6.88267231e-01
-8.45643103e-01 1.43522751e+00 7.14306450e+00 1.25893819e+00
-9.85383809e-01 2.27956012e-01 5.51624358e-01 -3.64506483e-01
-4.18399304e-01 9.00792554e-02 -8.83936584e-01 9.86210182e-02
1.41291988e+00 -4.06722903e-01 9.13653910e-01 6.36862993e-01
8.90581086e-02 3.27285618e-01 -1.13149714e+00 9.16024566e-01
-2.88476348e-01 -2.23071456e+00 2.50856578e-01 4.24434930e-01
5.76925099e-01 8.68454993e-01 -1.38431802e-01 3.27678412e-01
4.12426114e-01 -1.54908085e+00 2.50072360e-01 6.20356143e-01
3.70423555e-01 -8.93477738e-01 5.63477337e-01 1.68802485e-01
-1.07969964e+00 1.97139263e-01 -8.33908498e-01 -2.30411783e-01
-3.27530056e-01 2.51818269e-01 -7.26249039e-01 8.27861845e-01
1.33856714e-01 9.27516282e-01 -4.03856844e-01 1.18118262e+00
2.43201137e-01 4.39261198e-01 -4.38626707e-01 -6.66723013e-01
5.96476197e-01 -3.91151279e-01 3.77961308e-01 1.01459110e+00
1.25989676e-01 1.04555368e-01 9.12268907e-02 7.70337939e-01
-3.60295624e-01 2.69559205e-01 -7.54002988e-01 -3.68986189e-01
3.14752012e-01 1.20411694e+00 -4.25598711e-01 -1.96946740e-01
-2.82176942e-01 7.01627731e-01 6.06333256e-01 6.92018211e-01
-9.37801778e-01 -2.25270763e-01 6.74898148e-01 -4.10000086e-02
2.42444217e-01 -4.64437574e-01 3.64173025e-01 -1.05195153e+00
-5.01327395e-01 -9.00177300e-01 3.35455149e-01 -8.24527264e-01
-1.22273993e+00 6.86365008e-01 -4.24261332e-01 -4.55815256e-01
2.45842829e-01 -1.14062679e+00 -4.93518859e-01 8.97854090e-01
-1.32472360e+00 -8.05133104e-01 1.41823336e-01 6.22478724e-01
4.20400947e-02 -2.72112161e-01 1.22292006e+00 3.28292251e-01
-7.41426468e-01 3.74833494e-01 5.39690375e-01 -6.88439250e-01
5.08939087e-01 -1.15693164e+00 9.89264727e-01 2.17408150e-01
6.37434050e-02 8.18606913e-01 5.96729159e-01 -5.77588439e-01
-2.20158553e+00 -1.05715430e+00 4.98712331e-01 -3.98022562e-01
8.04762781e-01 -5.72189093e-01 -7.84929216e-01 5.57524920e-01
3.20833564e-01 2.76117884e-02 6.20506704e-01 2.47337118e-01
-2.52578527e-01 1.39811754e-01 -8.48488688e-01 4.77018714e-01
1.38768315e+00 -1.97665691e-01 1.15175806e-01 1.17747629e+00
8.00879478e-01 -6.75623894e-01 -1.18355155e+00 1.95561349e-01
4.18343991e-01 -5.74627280e-01 1.13708353e+00 -1.34822071e+00
8.16481262e-02 -1.87935203e-01 -1.40635476e-01 -1.21164560e+00
-8.38636100e-01 -1.11578381e+00 -4.19695735e-01 1.46722347e-01
8.49082589e-01 -7.35305250e-01 9.34652627e-01 5.43561816e-01
-3.43298465e-01 -9.90554512e-01 -1.17968929e+00 -8.03925097e-01
6.04786336e-01 -1.79591164e-01 5.88278472e-01 6.58421636e-01
2.25819632e-01 5.74992418e-01 -3.50410938e-01 3.55040506e-02
6.35625184e-01 -7.51463175e-02 5.15967786e-01 -1.30760121e+00
-3.17202300e-01 -4.20108616e-01 -4.78405267e-01 -1.01519358e+00
1.56089529e-01 -1.61485517e+00 -6.42416894e-01 -1.71985888e+00
3.03738832e-01 -5.15041538e-02 -6.58310950e-01 5.60299397e-01
2.36933026e-02 -1.08483121e-01 1.86066791e-01 1.74070299e-01
-1.03760421e+00 6.70646787e-01 1.46696544e+00 -4.14402872e-01
-7.68714920e-02 -1.99214175e-01 -7.98224092e-01 1.59153312e-01
9.47473943e-01 -5.26771426e-01 -2.11761475e-01 -2.05472156e-01
6.61748588e-01 9.94155258e-02 2.20186383e-01 -1.02130818e+00
5.56785047e-01 -2.53288597e-01 3.04633915e-01 -4.55153763e-01
5.41652620e-01 -4.10958409e-01 3.92885715e-01 7.45429158e-01
-1.63299739e-01 1.78625807e-01 4.96595889e-01 6.80782795e-01
2.64909685e-01 -2.78231245e-03 5.02725661e-01 -4.80358779e-01
-3.00804645e-01 9.62569535e-01 -6.99658513e-01 -3.77936035e-01
8.39412570e-01 5.29276282e-02 -8.70357335e-01 -1.44097935e-02
-9.56283569e-01 3.76406044e-01 4.28749025e-01 -4.13585128e-03
4.71363634e-01 -1.21273911e+00 -4.98640269e-01 -1.55452952e-01
1.58404514e-01 -1.65367216e-01 3.60247344e-01 9.45037544e-01
-8.08274209e-01 9.98935521e-01 -2.17534415e-02 -2.79162019e-01
-9.80475366e-01 7.06137419e-01 6.57523453e-01 -5.07682860e-01
-4.06735092e-01 8.12678635e-01 1.34266376e-01 -5.38066208e-01
1.28546536e-01 -2.72312433e-01 1.40294477e-01 -3.65870655e-01
3.40179652e-01 3.17377269e-01 5.15526474e-01 5.38138710e-02
-6.24591351e-01 2.84960538e-01 -5.84976852e-01 3.84197563e-01
1.64857435e+00 5.58481395e-01 -1.53248280e-01 5.53487651e-02
9.11633551e-01 -5.36337614e-01 -9.26330924e-01 -4.31377441e-02
-1.87788874e-01 3.20515245e-01 3.74627024e-01 -7.62806296e-01
-9.54863608e-01 9.49715316e-01 5.13977945e-01 2.19888225e-01
4.53121334e-01 -8.30125734e-02 6.33557320e-01 1.25465083e+00
5.06973267e-01 -7.61293530e-01 -3.39602083e-01 8.34470928e-01
7.71379948e-01 -8.85719955e-01 2.20730439e-01 -1.49610862e-01
-3.50143641e-01 1.24983847e+00 4.98554587e-01 -7.09646121e-02
4.51240391e-01 1.81447074e-01 -5.71886837e-01 -7.31439412e-01
-1.28824914e+00 -8.86848047e-02 1.05156451e-01 5.11759460e-01
7.41151690e-01 1.30002066e-01 -2.45871201e-01 4.43935722e-01
1.73068628e-01 1.46844499e-02 3.44932556e-01 8.29707742e-01
-5.23035586e-01 -1.47332847e+00 1.26261473e-01 3.64210367e-01
-9.95530710e-02 -7.29923427e-01 -8.19752097e-01 6.61497295e-01
-2.77752817e-01 6.62626922e-01 -3.22728127e-01 -2.91082054e-01
3.64622980e-01 1.37505770e-01 9.55107152e-01 -5.04295707e-01
-7.63322353e-01 -2.15674832e-01 2.91949421e-01 -9.37912047e-01
-2.94875532e-01 -2.40027979e-01 -1.54907882e+00 -9.92770731e-01
-3.15026730e-01 7.11014986e-01 8.02111149e-01 5.77664018e-01
9.35470164e-01 5.65531969e-01 1.05071858e-01 -1.01283777e+00
-3.23081613e-01 -8.42738271e-01 -5.92697442e-01 -2.67838806e-01
3.11583579e-02 -5.94869971e-01 -1.33476853e-01 -5.89088202e-01] | [5.341142654418945, 5.724493980407715] |
711b321e-9e72-4d5d-9c93-ea77b7d20e77 | objects-as-context-for-detecting-their | 1703.09529 | null | http://arxiv.org/abs/1703.09529v3 | http://arxiv.org/pdf/1703.09529v3.pdf | Objects as context for detecting their semantic parts | We present a semantic part detection approach that effectively leverages
object information.We use the object appearance and its class as indicators of
what parts to expect. We also model the expected relative location of parts
inside the objects based on their appearance. We achieve this with a new
network module, called OffsetNet, that efficiently predicts a variable number
of part locations within a given object. Our model incorporates all these cues
to detect parts in the context of their objects. This leads to considerably
higher performance for the challenging task of part detection compared to using
part appearance alone (+5 mAP on the PASCAL-Part dataset). We also compare to
other part detection methods on both PASCAL-Part and CUB200-2011 datasets. | ['Vittorio Ferrari', 'Abel Gonzalez-Garcia', 'Davide Modolo'] | 2017-03-28 | objects-as-context-for-detecting-their-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Gonzalez-Garcia_Objects_as_Context_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Gonzalez-Garcia_Objects_as_Context_CVPR_2018_paper.pdf | cvpr-2018-6 | ['semantic-part-detection'] | ['computer-vision'] | [ 2.10441321e-01 2.73160666e-01 -1.20153219e-01 -6.06211603e-01
-4.56708699e-01 -6.88515544e-01 7.22762406e-01 2.27244701e-02
-2.33495355e-01 3.23594272e-01 9.67017487e-02 2.71103144e-01
3.00185025e-01 -7.48293877e-01 -1.20188773e+00 -2.09981367e-01
-1.48735708e-03 6.90185666e-01 8.42562079e-01 -3.76718529e-02
1.05001666e-02 7.92644739e-01 -1.91478646e+00 5.91507316e-01
2.23717526e-01 1.28997600e+00 3.36356968e-01 6.46267414e-01
-1.80763230e-01 8.02357495e-01 -4.40052897e-01 -3.86753768e-01
6.13717794e-01 -1.21348994e-02 -8.75885248e-01 4.59059834e-01
1.09319401e+00 -3.82411152e-01 -3.97479624e-01 8.50128174e-01
1.57810390e-01 -1.11454129e-01 6.24758303e-01 -1.32953012e+00
-1.57052100e-01 5.86961806e-01 -5.57216763e-01 1.32641464e-01
3.74672890e-01 2.45328397e-01 1.16686511e+00 -7.23232448e-01
8.46803844e-01 1.47225273e+00 1.01571870e+00 3.86496812e-01
-1.24445307e+00 -2.98247904e-01 3.96694362e-01 2.28173032e-01
-1.21217406e+00 -5.04878044e-01 6.94944561e-01 -5.32047451e-01
1.06205201e+00 7.10399449e-02 8.62347424e-01 6.61567986e-01
-1.81045458e-02 1.32743251e+00 6.42610371e-01 -1.84787437e-01
1.39050245e-01 1.50652751e-02 3.73866946e-01 6.08686805e-01
3.59410107e-01 -1.25814244e-01 -3.50485116e-01 1.03203282e-02
5.57925344e-01 2.86052555e-01 -1.53750658e-01 -8.04453611e-01
-9.13511515e-01 3.08505595e-01 8.01645041e-01 -3.95329893e-02
-5.71027577e-01 5.90713680e-01 2.83357263e-01 -2.16564357e-01
3.80867839e-01 1.84272960e-01 -1.11278367e+00 1.08702242e-01
-9.44844663e-01 4.92730200e-01 8.46933484e-01 1.27489197e+00
1.05749381e+00 -3.22403431e-01 -4.11306977e-01 8.36772740e-01
3.91040236e-01 3.71625751e-01 -1.42576039e-01 -1.18370903e+00
2.42027909e-01 9.85694826e-01 2.02767313e-01 -2.42188185e-01
-4.00455087e-01 -3.21796805e-01 -9.89306867e-02 5.09174347e-01
4.36181486e-01 1.89446524e-01 -1.40250015e+00 1.57543206e+00
6.55246377e-01 -1.69860184e-01 -3.99405062e-01 6.71257794e-01
8.63289237e-01 7.90161639e-02 1.04240492e-01 7.34463096e-01
1.69772220e+00 -1.18519711e+00 -4.04079348e-01 -6.58771515e-01
3.35897774e-01 -7.08673298e-01 6.22875214e-01 9.06003639e-02
-8.89455855e-01 -7.38127470e-01 -9.61334467e-01 -1.45313606e-01
-6.56632125e-01 4.85245854e-01 9.62711334e-01 3.79203439e-01
-1.10108078e+00 8.18577528e-01 -9.14422095e-01 -3.72045457e-01
1.02057409e+00 3.45256805e-01 -4.22125936e-01 -3.40934813e-01
-2.97112256e-01 7.48222172e-01 7.15786695e-01 -8.42034519e-02
-7.58686960e-01 -6.45844460e-01 -1.28179908e+00 2.05973521e-01
4.35460955e-01 -8.03460062e-01 1.44312692e+00 -1.30790818e+00
-7.98236489e-01 8.87008131e-01 -1.04778394e-01 -5.34921587e-01
5.72459877e-01 -3.25769901e-01 -3.29430178e-02 2.59667039e-01
2.12887898e-01 1.51839900e+00 6.94638610e-01 -1.36490524e+00
-7.83201098e-01 -3.58761489e-01 2.06803396e-01 1.71672374e-01
3.56275350e-01 -3.81953828e-02 -8.86052966e-01 -3.03591400e-01
2.28944495e-01 -7.77444303e-01 -6.89731464e-02 9.52358484e-01
-5.01599729e-01 -3.94243270e-01 1.10357881e+00 -4.65590388e-01
3.57706338e-01 -1.95999551e+00 -5.01249015e-01 -2.17579141e-01
1.57064870e-01 1.89198196e-01 -3.47363114e-01 2.17364684e-01
-2.03724243e-02 -2.96283036e-01 -2.76598006e-01 -4.65655357e-01
4.98010628e-02 2.39240244e-01 -1.65763637e-03 4.65130866e-01
5.43068290e-01 1.24038875e+00 -6.37409627e-01 -6.40650451e-01
3.14604551e-01 6.04968548e-01 -6.53632522e-01 1.70919858e-02
-5.48616707e-01 -3.50792289e-01 -2.70779222e-01 1.05107212e+00
1.05882049e+00 -3.43248039e-01 -7.32792541e-02 -2.45699674e-01
8.35137367e-02 4.56859767e-01 -1.13081920e+00 1.53847408e+00
2.34769024e-02 6.33069158e-01 3.41235161e-01 -4.53896463e-01
5.01005173e-01 1.02759555e-01 5.36287189e-01 -2.60829419e-01
-7.44268596e-02 -1.28114372e-01 2.21910462e-01 -1.59915313e-01
5.04480720e-01 1.49051353e-01 2.72780091e-01 1.76999897e-01
3.64568472e-01 -1.80482447e-01 3.97197634e-01 2.19974831e-01
1.42519486e+00 8.49716902e-01 2.99364597e-01 -3.08058947e-01
5.16426340e-02 2.12429836e-01 5.53919613e-01 8.99304390e-01
-3.99299264e-01 8.80599082e-01 4.05104220e-01 -6.24676287e-01
-8.48434865e-01 -1.02072906e+00 -2.76986808e-01 1.07703900e+00
2.91684598e-01 -4.14595962e-01 -6.18659019e-01 -1.17223680e+00
4.82363045e-01 5.14719546e-01 -1.06306970e+00 1.36440203e-01
-5.76052248e-01 -2.04356983e-01 2.04792753e-01 1.05295610e+00
3.57582808e-01 -1.48365068e+00 -6.81602299e-01 2.40653995e-02
5.87379709e-02 -1.21317899e+00 -5.55472970e-01 6.91781640e-01
-7.59704828e-01 -1.51793504e+00 -4.84726697e-01 -7.31069624e-01
8.19183052e-01 3.84955466e-01 1.58001244e+00 1.59002274e-01
-7.33803928e-01 6.69806004e-01 -2.76487768e-01 -6.75415993e-01
-3.00813198e-01 -2.76348233e-01 -4.39598411e-01 -3.44223171e-01
4.26896811e-01 -2.82749057e-01 -7.75403857e-01 3.88006479e-01
-7.82715261e-01 4.19673659e-02 9.50130463e-01 4.58394974e-01
7.04359412e-01 -2.07489714e-01 -1.67245269e-01 -6.67641163e-01
-3.75036478e-01 -2.76685715e-01 -6.88635111e-01 3.33874643e-01
-1.48526102e-01 5.42182289e-02 2.56083965e-01 -3.47735733e-01
-9.62557852e-01 6.68267727e-01 -5.08114919e-02 -4.35580879e-01
-5.05403519e-01 -4.43843037e-01 -4.08483893e-01 -1.32748438e-02
3.65799278e-01 2.09170461e-01 -8.68097767e-02 -5.90843856e-01
5.00392020e-01 3.10082614e-01 6.38891399e-01 -3.84820610e-01
6.83082461e-01 8.09481621e-01 -2.84613110e-02 -6.35517776e-01
-9.85518157e-01 -9.61871564e-01 -9.44242716e-01 -8.76036286e-02
7.99123645e-01 -1.17651832e+00 -6.88092172e-01 5.21942735e-01
-1.33784866e+00 -4.63389903e-01 -5.83338797e-01 2.29387030e-01
-5.25562584e-01 3.10381383e-01 -6.22550607e-01 -6.69472754e-01
-3.41379166e-01 -8.84813190e-01 1.69201028e+00 1.11657560e-01
1.33111784e-02 -6.35313034e-01 -2.71646053e-01 1.52860284e-01
1.85292616e-01 9.10501927e-02 3.70353252e-01 -7.52193451e-01
-9.29019451e-01 -3.40444058e-01 -5.21556854e-01 2.22215608e-01
2.93273404e-02 1.59392938e-01 -1.02749264e+00 -5.41895293e-02
-3.02917063e-01 -3.12149912e-01 1.15093064e+00 4.41983342e-01
1.30391693e+00 -2.12863714e-01 -7.88486123e-01 4.60021764e-01
1.48686647e+00 -1.48434997e-01 4.91003007e-01 1.70826301e-01
7.08471715e-01 6.10682070e-01 7.33569562e-01 3.83548975e-01
2.84653634e-01 7.16876566e-01 7.81130314e-01 -6.60851076e-02
-7.03437686e-01 -2.73082942e-01 2.69249231e-02 -2.62459695e-01
2.55257130e-01 -3.56355131e-01 -7.07087815e-01 6.24110997e-01
-1.66758335e+00 -1.05566156e+00 -3.66430342e-01 1.78518009e+00
5.05737603e-01 3.91525567e-01 2.04053134e-01 -4.10975546e-01
8.32695603e-01 -3.57745066e-02 -6.44359589e-01 7.55281821e-02
-1.39251262e-01 3.41756679e-02 7.95270443e-01 9.80736092e-02
-1.39116406e+00 9.78612781e-01 7.76308441e+00 6.61521673e-01
-5.56901395e-01 -1.86436355e-01 5.99829733e-01 4.97771092e-02
2.00822437e-03 6.66291192e-02 -1.20295048e+00 2.26369873e-01
3.70855659e-01 6.24530792e-01 3.39151710e-01 1.30897987e+00
-2.91627020e-01 -5.09417057e-01 -1.44505382e+00 6.30064249e-01
1.19332016e-01 -1.04467535e+00 -1.04475103e-01 3.67752090e-02
6.12776995e-01 5.95217705e-01 -4.46075231e-01 4.00752693e-01
3.89244795e-01 -7.08369195e-01 1.24082446e+00 4.84475553e-01
4.57804829e-01 -3.38281780e-01 6.18874729e-01 3.16388130e-01
-1.46525860e+00 -4.84723151e-02 -5.75128853e-01 -5.22372164e-02
9.81176458e-03 4.05113786e-01 -1.16486025e+00 1.99599057e-01
7.63611495e-01 7.46918797e-01 -9.78045762e-01 1.43780148e+00
-4.10824269e-01 4.28570807e-01 -6.65552139e-01 1.65197194e-01
-7.49511197e-02 1.99799940e-01 2.77603656e-01 1.21237695e+00
-7.80330151e-02 -2.02160522e-01 2.60465950e-01 1.16998124e+00
-2.48870403e-01 -3.52800339e-01 -4.08583343e-01 1.82871655e-01
4.72323239e-01 1.53454185e+00 -1.16436660e+00 -7.24406719e-01
-4.85040903e-01 1.00475872e+00 5.23747504e-01 -1.80741344e-02
-7.20548451e-01 -3.12573373e-01 9.29523110e-01 2.19416171e-01
1.04717767e+00 1.94087535e-01 -1.84120238e-01 -7.46116757e-01
2.07338005e-01 -4.90809679e-01 3.22493196e-01 -1.05017292e+00
-1.14604437e+00 4.08190906e-01 7.00816438e-02 -1.15610540e+00
8.96920338e-02 -1.09557402e+00 -7.54946470e-01 4.72336084e-01
-1.43542433e+00 -1.39016497e+00 -4.34906632e-01 1.61390468e-01
6.86542153e-01 4.19795781e-01 4.51304734e-01 -1.31289363e-01
-2.34214365e-01 3.31876725e-01 -3.30099463e-01 4.01819617e-01
4.92940217e-01 -1.62307072e+00 9.18335378e-01 6.85741425e-01
3.27992499e-01 3.40333819e-01 5.98366976e-01 -7.77556479e-01
-1.06538498e+00 -1.21541762e+00 4.52323347e-01 -9.56772268e-01
2.58428335e-01 -6.50502682e-01 -5.64174116e-01 8.80666792e-01
1.31175160e-01 6.95278406e-01 1.47552624e-01 5.15139662e-02
-5.66234112e-01 7.37571940e-02 -1.31884348e+00 3.09047729e-01
1.40213859e+00 -3.72771859e-01 -6.03261888e-01 3.58552814e-01
7.43861735e-01 -3.59742045e-01 -5.14081359e-01 6.87760711e-01
6.64349377e-01 -1.28803432e+00 1.13875020e+00 -4.27139103e-01
2.54666716e-01 -4.51770544e-01 -2.72518247e-01 -8.82261634e-01
-5.57989061e-01 5.25313430e-03 -2.65409738e-01 1.10101032e+00
1.87900037e-01 -2.71280110e-01 9.92287993e-01 6.00421011e-01
-3.36296856e-01 -8.04990888e-01 -5.86539865e-01 -7.62423694e-01
-4.26182240e-01 -4.08315718e-01 6.56370461e-01 2.46016771e-01
-5.84289789e-01 -1.41763324e-02 3.91997516e-01 7.24121705e-02
3.81581813e-01 3.94445747e-01 8.53680134e-01 -1.22761822e+00
-4.52738434e-01 -6.15420222e-01 -9.65071917e-01 -1.18858135e+00
7.52752498e-02 -8.79024446e-01 5.69591105e-01 -1.81467247e+00
6.12254500e-01 -1.99040920e-01 -3.73774230e-01 9.85250950e-01
-3.99700493e-01 5.21084845e-01 1.53400570e-01 7.76906237e-02
-1.05641460e+00 5.13675988e-01 1.03799200e+00 -1.57267347e-01
2.27713481e-01 2.26797059e-01 -4.76031244e-01 9.57356930e-01
4.91839856e-01 -5.80192745e-01 2.09832862e-01 -1.54220238e-01
-3.53154808e-01 -2.83319950e-01 6.31279171e-01 -1.29309070e+00
-3.23849499e-01 1.83340743e-01 1.02596056e+00 -1.26997972e+00
6.12916708e-01 -9.28101838e-01 -3.18935029e-02 4.94078130e-01
-9.37956050e-02 -1.44650266e-01 3.73577118e-01 7.18102515e-01
8.37593600e-02 -1.77921534e-01 7.36442745e-01 -3.76630276e-01
-9.85127509e-01 3.90140891e-01 -1.28446504e-01 1.56438269e-03
1.03059781e+00 -4.91586924e-01 -2.49481037e-01 -2.53806282e-02
-5.98784804e-01 2.95706689e-01 8.64833653e-01 6.30107462e-01
3.19568872e-01 -1.20588291e+00 -4.77321386e-01 1.00712821e-01
4.06352997e-01 1.47625312e-01 7.65089616e-02 6.23002648e-01
-5.29886365e-01 2.74887741e-01 -1.09832846e-01 -8.83377790e-01
-1.47431719e+00 6.80893362e-01 4.37537134e-01 7.71709457e-02
-7.28816509e-01 1.10949302e+00 4.58061278e-01 -6.73152864e-01
2.20008194e-01 -3.45456570e-01 1.60700351e-01 -2.28801548e-01
4.08867419e-01 3.09481800e-01 -3.76431495e-02 -7.16150403e-01
-7.15354323e-01 3.75521630e-01 -8.96847397e-02 2.21688345e-01
1.36475646e+00 2.92537421e-01 -1.65817350e-01 6.13497570e-02
1.18450153e+00 8.07955861e-02 -2.01867580e+00 -2.03713968e-01
5.68358861e-02 -3.64937574e-01 -2.69385368e-01 -1.23019254e+00
-9.49550867e-01 5.02228677e-01 5.45503020e-01 -8.44728723e-02
6.82366490e-01 6.11176610e-01 3.03973228e-01 5.15844107e-01
4.21016932e-01 -9.31990027e-01 -9.45815258e-03 4.55838174e-01
6.40510499e-01 -1.37156808e+00 2.76918918e-01 -6.80800676e-01
-3.44628960e-01 1.08844960e+00 1.01600528e+00 -1.45835638e-01
5.72916508e-01 3.94617081e-01 6.56958595e-02 -4.48372990e-01
-6.63852692e-01 -6.69978440e-01 6.96490943e-01 6.26725376e-01
3.69742751e-01 7.34718889e-02 1.56863257e-01 1.46789417e-01
-5.85013814e-02 -3.39730144e-01 1.71760261e-01 1.24906647e+00
-6.65571630e-01 -9.26473141e-01 -4.40435857e-01 6.27731562e-01
-4.56892252e-01 9.06420574e-02 -7.53806591e-01 9.06287670e-01
5.18530369e-01 7.33848333e-01 2.38969266e-01 1.00872070e-01
6.94391951e-02 2.42443055e-01 6.85197413e-01 -8.07381749e-01
-4.29644287e-01 -2.79422471e-04 1.35127064e-02 -9.79131043e-01
-3.62803191e-01 -7.17206120e-01 -1.27576125e+00 3.39880526e-01
-6.02003336e-01 -1.79924399e-01 8.00619245e-01 8.25594664e-01
4.51920599e-01 5.92328906e-01 1.84331715e-01 -1.27400076e+00
-3.83340150e-01 -1.11386526e+00 -6.39499962e-01 5.61544716e-01
1.73320979e-01 -5.85813522e-01 -1.41956240e-01 -2.30285406e-01] | [9.276906967163086, 0.8759414553642273] |
edc27c25-38ab-42e6-900c-0388db7ffba5 | deploying-a-robust-active-preference | 2306.04061 | null | https://arxiv.org/abs/2306.04061v1 | https://arxiv.org/pdf/2306.04061v1.pdf | Deploying a Robust Active Preference Elicitation Algorithm: Experiment Design, Interface, and Evaluation for COVID-19 Patient Prioritization | Preference elicitation leverages AI or optimization to learn stakeholder preferences in settings ranging from marketing to public policy. The online robust preference elicitation procedure of arXiv:2003.01899 has been shown in simulation to outperform various other elicitation procedures in terms of effectively learning individuals' true utilities. However, as with any simulation, the method makes a series of assumptions that cannot easily be verified to hold true beyond simulation. Thus, we propose to validate the robust method's performance in deployment, focused on the particular challenge of selecting policies for prioritizing COVID-19 patients for scarce hospital resources during the pandemic. To this end, we develop an online platform for preference elicitation where users report their preferences between alternatives over a moderate number of pairwise comparisons chosen by a particular elicitation procedure. We recruit Amazon Mechanical Turk workers ($n$ = 193) to report their preferences and demonstrate that the robust method outperforms asking random queries by 21%, the next best performing method in the simulated results of arXiv:2003.01899, in terms of recommending policies with a higher utility. | ['Phebe Vayanos', 'Simon Blessenohl', 'Patrick Vossler', 'Caroline M. Johnston'] | 2023-06-06 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [ 4.88696955e-02 2.55551815e-01 -3.70149910e-01 -4.51931953e-01
-1.13692963e+00 -1.17554355e+00 2.08345264e-01 2.77541131e-01
-7.60507941e-01 1.08922827e+00 4.93914574e-01 -7.99892604e-01
-5.80604732e-01 -5.35656631e-01 -4.80722964e-01 -3.92018557e-01
-3.52542341e-01 9.95917201e-01 -5.51462352e-01 3.37403007e-02
3.70645225e-01 2.42343217e-01 -1.08187735e+00 1.91613913e-01
8.73830736e-01 1.02866614e+00 -3.33585173e-01 4.40793604e-01
4.16073024e-01 2.45207980e-01 -2.89599180e-01 -5.86751521e-01
5.03001511e-01 -5.87514490e-02 -9.46301043e-01 -4.24091905e-01
-5.55851050e-02 -8.69711280e-01 1.42236978e-01 6.50067210e-01
1.12139797e+00 2.79264897e-01 6.89715385e-01 -1.16540885e+00
-5.83088398e-01 8.22639406e-01 -4.21831049e-02 7.45071545e-02
9.70675647e-01 3.63310099e-01 1.46627736e+00 -6.41308784e-01
6.57986581e-01 1.18411303e+00 3.50511461e-01 3.88000160e-01
-1.32348824e+00 -5.46414316e-01 1.77311137e-01 -1.94105223e-01
-1.33028603e+00 -2.08592623e-01 4.14962173e-01 -4.82350260e-01
4.96602237e-01 6.93203151e-01 5.51354349e-01 1.12136400e+00
-3.25813562e-01 6.95807874e-01 1.48309803e+00 5.57395630e-02
7.61825621e-01 5.37383139e-01 3.61003280e-02 2.68580884e-01
5.09351850e-01 2.14935005e-01 -5.59094906e-01 -1.16406953e+00
3.91083241e-01 -1.40466496e-01 -2.99140960e-01 -3.11749607e-01
-1.10655916e+00 1.05940711e+00 1.07505828e-01 -3.87640834e-01
-9.09899592e-01 -2.05088094e-01 -5.35072424e-02 2.77494311e-01
3.73079449e-01 1.02908325e+00 -6.21144891e-01 -1.68364942e-01
-7.50345469e-01 7.48651743e-01 1.19026279e+00 9.74331737e-01
3.50347936e-01 -4.88177985e-01 -6.84719980e-01 4.42190856e-01
2.21803784e-01 6.33237004e-01 1.34821907e-01 -1.43637323e+00
3.83825570e-01 4.04909104e-01 1.32809854e+00 -1.07627523e+00
-4.89271075e-01 -2.13826984e-01 -2.44425967e-01 -2.44160816e-01
4.82095927e-01 -1.13272345e+00 -1.85803488e-01 1.71965003e+00
4.33125675e-01 -3.32094938e-01 -1.74869355e-02 1.15218222e+00
3.86652350e-01 4.21795547e-01 9.12968442e-02 -3.90255570e-01
1.35766184e+00 -4.89889942e-02 -3.14992785e-01 3.08352616e-03
4.62954879e-01 -4.69495952e-01 1.09509087e+00 3.56340945e-01
-1.28516054e+00 4.02885556e-01 -3.86202306e-01 5.53360283e-01
-6.11994676e-02 -4.57148015e-01 8.25154185e-01 9.09548700e-01
-9.55878258e-01 3.19603026e-01 -1.04255155e-01 -3.36898834e-01
5.38958013e-01 5.42188346e-01 1.98643282e-01 -1.18744873e-01
-1.42035019e+00 7.61060894e-01 -2.14944154e-01 -2.38699287e-01
-6.80655003e-01 -1.05584669e+00 -3.47053498e-01 2.62136012e-01
6.43507183e-01 -1.11752367e+00 1.56817317e+00 -6.99688196e-01
-1.43012762e+00 4.28339779e-01 2.77031548e-02 -5.31568110e-01
9.19269145e-01 6.10323958e-02 -1.43265367e-01 1.30359739e-01
1.23167776e-01 5.07603288e-01 1.85921758e-01 -1.19150925e+00
-7.34008491e-01 -1.81120977e-01 5.96161783e-01 4.97282952e-01
-1.13565840e-01 2.06248954e-01 3.00731540e-01 -3.12496215e-01
-4.15463358e-01 -9.59577322e-01 -6.74992085e-01 -5.21757126e-01
-3.53957117e-01 -1.99672893e-01 -1.15650170e-01 -2.73664057e-01
1.01821065e+00 -1.60316622e+00 -4.04745787e-01 8.40796351e-01
-1.16531692e-01 -2.35337853e-01 -1.34907454e-01 7.15070128e-01
4.66191530e-01 7.09360361e-01 -1.92462072e-01 1.72757488e-02
7.18626678e-01 -1.72976866e-01 -2.76256293e-01 4.40705866e-01
-1.97722599e-01 8.91352534e-01 -9.85069394e-01 -4.94748831e-01
-2.89740860e-01 9.32799745e-03 -1.02847421e+00 1.78019047e-01
-2.21964672e-01 3.05301815e-01 -7.85762191e-01 5.86642444e-01
4.78442967e-01 -2.96123207e-01 7.39168048e-01 2.46157706e-01
-3.08647528e-02 3.83005887e-01 -1.19328940e+00 1.10606158e+00
-2.23158017e-01 -1.25583977e-01 1.66987211e-01 -6.79631829e-01
4.03124034e-01 5.03768623e-01 9.66670096e-01 -4.82070386e-01
2.86091834e-01 1.74560934e-01 2.62008719e-02 -6.53445721e-01
2.20393419e-01 -1.04036272e-01 -4.58518982e-01 1.21793664e+00
-5.64584017e-01 -1.86782733e-01 -4.72580120e-02 2.27217987e-01
9.68813598e-01 -5.47782719e-01 2.07096398e-01 -7.97391593e-01
1.44931644e-01 2.71626264e-01 4.63424057e-01 1.21655667e+00
-2.24117681e-01 3.54416937e-01 4.70643938e-01 -3.55050564e-01
-6.05289519e-01 -9.04904366e-01 1.06055737e-01 1.11064851e+00
9.27845165e-02 1.99523836e-01 -7.24031985e-01 -5.90734780e-01
4.47397947e-01 1.14894581e+00 -5.97924888e-01 1.89194024e-01
4.46093306e-02 -6.90614760e-01 3.05535644e-01 2.57413447e-01
1.20421633e-01 -1.11190784e+00 -1.11903942e+00 4.75588925e-02
-4.49675679e-01 -7.75404155e-01 -1.06089449e+00 -5.11982381e-01
-4.64677989e-01 -1.16793668e+00 -1.00269818e+00 -8.54102597e-02
6.81668282e-01 4.13743705e-02 1.11819601e+00 -1.97468504e-01
7.36501068e-02 1.17521107e+00 -1.91219494e-01 -6.01931095e-01
1.89880162e-01 -5.43022156e-03 4.76503164e-01 5.61806411e-02
5.88019371e-01 -4.06498462e-01 -1.29859066e+00 4.39904004e-01
-8.75188828e-01 -5.07494867e-01 4.22657788e-01 4.96350855e-01
2.15546668e-01 -3.20960850e-01 7.93466389e-01 -1.25924528e+00
1.44006205e+00 -9.61253524e-01 -6.98655665e-01 2.69061953e-01
-1.03904200e+00 -2.15438768e-01 4.16108847e-01 -3.82133424e-01
-9.60211754e-01 -1.16944797e-01 4.64802861e-01 2.20217034e-01
-1.37949646e-01 8.50132585e-01 4.47845683e-02 1.62094340e-01
7.96731591e-01 -3.29512149e-01 -7.94609785e-02 -3.40676516e-01
3.30459177e-01 6.97769940e-01 6.49767183e-03 -1.05297577e+00
6.21325910e-01 1.92666471e-01 -4.39003021e-01 -2.24286959e-01
-6.80632234e-01 -2.65286446e-01 4.14270721e-02 2.51437929e-02
5.62929094e-01 -6.94076478e-01 -1.67518318e+00 -3.10703635e-01
-7.59092212e-01 -5.17377257e-01 -1.46230891e-01 6.70659482e-01
-7.13082135e-01 6.47629201e-02 -2.61594415e-01 -1.32609856e+00
-2.79437631e-01 -1.04074991e+00 3.98369640e-01 2.24731669e-01
-7.11730957e-01 -8.56147528e-01 -1.46817947e-02 5.93669176e-01
7.27124155e-01 1.20530047e-01 8.99153531e-01 -1.02869773e+00
-3.21604967e-01 -3.98642033e-01 -4.31936532e-02 -3.78231525e-01
2.63534244e-02 -1.83951825e-01 -7.56840408e-01 -5.73154092e-01
-2.49815032e-01 -3.70902896e-01 1.33876041e-01 7.05020249e-01
1.20516086e+00 -1.01656461e+00 -1.86908007e-01 -1.36167854e-02
1.07835162e+00 3.90133977e-01 1.42652556e-01 1.11662336e-01
-2.24253193e-01 1.01257801e+00 7.89793193e-01 1.10426712e+00
7.15162992e-01 3.70104939e-01 1.67017549e-01 7.04013109e-02
1.13958883e+00 -2.85377860e-01 9.71314386e-02 -1.14969559e-01
-1.63741231e-01 -2.64789045e-01 -8.46465647e-01 4.70083356e-01
-1.88321340e+00 -7.51022875e-01 7.66681552e-01 2.66625428e+00
8.72587204e-01 1.32570844e-02 8.36996913e-01 -4.74761546e-01
4.42019701e-01 -3.03133219e-01 -9.13925529e-01 -7.35332310e-01
1.60670802e-01 6.75972849e-02 8.05924952e-01 5.89550436e-01
-5.28196216e-01 3.25565904e-01 7.23210430e+00 8.22581798e-02
-5.63770354e-01 -1.18861549e-01 1.12261033e+00 -5.16966999e-01
-1.08423030e+00 -1.43092975e-01 -2.63384968e-01 3.96117330e-01
1.05185080e+00 -8.56904268e-01 9.59021509e-01 5.82916200e-01
5.12350082e-01 -1.33973688e-01 -1.33569896e+00 8.31668079e-01
-2.33185887e-01 -1.27720940e+00 -8.03989097e-02 2.07242951e-01
9.64619756e-01 -3.86640221e-01 5.72631299e-01 1.35235354e-01
1.02427328e+00 -1.29246581e+00 5.38084567e-01 6.48475528e-01
6.52786255e-01 -9.34957027e-01 8.41522872e-01 4.85425711e-01
-2.91786522e-01 -4.05820549e-01 -2.19434932e-01 -2.03597784e-01
2.52840072e-01 5.68508625e-01 -1.23877180e+00 3.94699454e-01
5.13653934e-01 -1.73650876e-01 2.90077310e-02 1.06820297e+00
2.88937956e-01 7.89397240e-01 -4.19337571e-01 -3.83986443e-01
3.58328611e-01 -2.38442123e-01 2.59539515e-01 1.08356845e+00
5.65745234e-01 7.94029355e-01 1.14713222e-01 9.74280357e-01
-2.46096432e-01 2.93961346e-01 -6.65778577e-01 -3.90167445e-01
1.05517590e+00 9.99691844e-01 -2.12686241e-01 -1.55019611e-01
-1.15395794e-02 3.72490555e-01 -4.41892505e-01 8.18952262e-01
-6.16641939e-01 6.57986943e-03 7.08300292e-01 1.56209350e-01
-5.44597954e-02 4.48192209e-02 -3.30996275e-01 -5.67804396e-01
-3.04897487e-01 -1.27299225e+00 8.25860023e-01 -5.39096653e-01
-1.41042101e+00 1.24863721e-01 3.41848224e-01 -7.91752160e-01
-5.11422694e-01 -4.81627434e-01 -4.17912811e-01 1.20501137e+00
-1.18669724e+00 -3.97080034e-01 9.80232134e-02 5.23988008e-01
1.23458460e-01 1.46936372e-01 9.49690700e-01 4.78985533e-03
-3.63298118e-01 7.54728436e-01 -4.83020879e-02 -3.96027982e-01
6.79154336e-01 -1.19419324e+00 -1.93277188e-02 9.40884128e-02
-4.32274669e-01 8.66896272e-01 1.00368071e+00 -4.96029496e-01
-1.63088894e+00 -5.49595654e-01 6.88840687e-01 -4.93488878e-01
3.30823511e-01 -2.13005468e-01 -1.70991257e-01 6.64209843e-01
1.86003178e-01 -5.24289310e-01 1.14706969e+00 9.69249308e-02
4.54490855e-02 7.35587925e-02 -1.65323234e+00 9.24011588e-01
9.70140636e-01 -5.28585374e-01 -3.05002004e-01 5.48916698e-01
6.76223695e-01 -7.80600831e-02 -1.00892043e+00 2.54014343e-01
8.10953021e-01 -6.97535396e-01 1.07262158e+00 -1.12246764e+00
1.36548534e-01 8.28087479e-02 -4.09909189e-01 -1.30667460e+00
-3.52812678e-01 -1.12159240e+00 3.67564231e-01 9.19994116e-01
9.29795861e-01 -1.22129786e+00 7.40874529e-01 1.60726929e+00
6.55981958e-01 -8.49004269e-01 -6.54844820e-01 -2.88575292e-01
2.95668960e-01 -1.13370053e-01 1.06068635e+00 9.96862710e-01
2.85271853e-01 -1.68960050e-01 -3.73571903e-01 2.64563501e-01
7.32758403e-01 3.47458631e-01 5.24197698e-01 -1.17623973e+00
-2.82379508e-01 -4.98581499e-01 5.25203526e-01 -8.23692501e-01
-2.74675712e-02 -3.68301600e-01 -9.29342210e-02 -1.55374098e+00
2.66007245e-01 -6.90756619e-01 -6.11767232e-01 9.04367343e-02
-2.91080475e-02 -1.28941283e-01 -8.65369141e-02 -1.12503819e-01
-5.62614560e-01 2.34737977e-01 1.03812802e+00 -7.84618333e-02
-2.73171693e-01 4.99538809e-01 -1.48099577e+00 3.70721519e-01
1.03760743e+00 -3.19580197e-01 -6.08647585e-01 -4.02723700e-02
5.53153872e-01 4.14836437e-01 2.13407338e-01 -7.33213127e-02
3.11176032e-01 -9.62940156e-01 1.16251059e-01 -2.28512630e-01
1.14402875e-01 -8.15382302e-01 3.05021137e-01 3.84445339e-01
-8.03723335e-01 4.07649815e-01 6.33766055e-02 4.34282929e-01
5.32149196e-01 -1.45246968e-01 7.71723166e-02 -3.49187285e-01
6.31198958e-02 5.22612154e-01 -6.38279319e-01 4.24497455e-01
8.54418516e-01 6.20581806e-02 -2.41587117e-01 -1.11393440e+00
-5.89770019e-01 6.86046004e-01 4.53324556e-01 -3.03404052e-02
4.53425467e-01 -1.00554311e+00 -8.17931414e-01 -2.99581766e-01
-9.62799639e-02 -3.42166722e-01 2.19834104e-01 9.81149912e-01
-5.40736355e-02 6.81258619e-01 -3.52816880e-02 8.75381380e-02
-7.86107004e-01 6.30272567e-01 2.99676448e-01 3.71163525e-02
7.90558904e-02 5.76867640e-01 -4.61800285e-02 -6.01424396e-01
2.68447012e-01 -3.80486101e-02 -7.17883632e-02 1.70291469e-01
3.20959508e-01 6.98112607e-01 -3.56670231e-01 -9.98108238e-02
-5.16861200e-01 -5.22795543e-02 2.38479018e-01 -6.24072790e-01
1.15555322e+00 -2.73113132e-01 1.57833710e-01 1.41720727e-01
7.17093706e-01 2.96550244e-01 -9.71033096e-01 -1.45178795e-01
2.08415240e-01 -6.21473730e-01 -3.76387447e-01 -1.26066101e+00
-5.62190354e-01 1.33271515e-01 4.50760871e-01 4.80764508e-01
8.25308740e-01 -3.50437105e-01 5.19805610e-01 5.90135813e-01
5.35350204e-01 -1.19747281e+00 -3.46201926e-01 -1.38736948e-01
7.07725883e-01 -1.23415923e+00 -2.01565132e-01 1.06004644e-02
-1.11328709e+00 4.32575881e-01 8.98637623e-02 8.19060355e-02
7.31105030e-01 -2.09073424e-01 2.13470668e-01 -2.85880983e-01
-8.85380149e-01 2.40163896e-02 2.75296867e-01 4.30241823e-01
3.50002468e-01 7.51949966e-01 -6.12692237e-01 9.30405319e-01
-4.01455253e-01 4.90319692e-02 5.65696895e-01 6.48627162e-01
-1.78297117e-01 -9.41272557e-01 -2.38276690e-01 8.92361164e-01
-5.15395820e-01 -1.09707974e-01 -6.36479199e-01 4.43090081e-01
-4.19381827e-01 1.41962755e+00 -1.51052147e-01 -1.44379482e-01
5.53223908e-01 5.53375222e-02 9.20204446e-02 -3.71078461e-01
-6.88525200e-01 -3.67007732e-01 4.63603318e-01 -7.13564813e-01
-3.49519819e-01 -8.97409022e-01 -8.54677022e-01 -5.24809539e-01
9.37180296e-02 6.17597282e-01 1.73857719e-01 7.47572362e-01
6.40946507e-01 -1.94206640e-01 9.79306996e-01 -6.20687008e-01
-1.29006708e+00 -4.76540715e-01 -6.44406796e-01 5.74805439e-01
1.90412164e-01 -5.99120975e-01 -4.14475679e-01 -6.33028686e-01] | [9.149348258972168, 5.441460132598877] |
9d322119-8627-4ae0-9fb8-f176d4f8c845 | deep-tractable-probabilistic-models-for-moral | 1810.03736 | null | https://arxiv.org/abs/1810.03736v3 | https://arxiv.org/pdf/1810.03736v3.pdf | Learning Tractable Probabilistic Models for Moral Responsibility and Blame | Moral responsibility is a major concern in autonomous systems, with applications ranging from self-driving cars to kidney exchanges. Although there have been recent attempts to formalise responsibility and blame, among similar notions, the problem of learning within these formalisms has been unaddressed. From the viewpoint of such systems, the urgent questions are: (a) How can models of moral scenarios and blameworthiness be extracted and learnt automatically from data? (b) How can judgements be computed effectively and efficiently, given the split-second decision points faced by some systems? By building on constrained tractable probabilistic learning, we propose and implement a hybrid (between data-driven and rule-based methods) learning framework for inducing models of such scenarios automatically from data and reasoning tractably from them. We report on experiments that compare our system with human judgement in three illustrative domains: lung cancer staging, teamwork management, and trolley problems. | ['Lewis Hammond', 'Vaishak Belle'] | 2018-10-08 | null | null | null | null | ['moral-scenarios'] | ['miscellaneous'] | [ 3.76692206e-01 1.11672008e+00 -1.09363817e-01 -8.47679198e-01
-4.74487185e-01 -3.30596775e-01 8.27997029e-01 3.79085511e-01
-5.59175491e-01 9.28971410e-01 2.61140198e-01 -6.11686647e-01
-6.79174006e-01 -5.00624418e-01 -1.85929835e-01 -5.20931780e-01
1.05619967e-01 7.92978525e-01 1.18488364e-01 -2.69095778e-01
3.88967395e-01 3.31398904e-01 -1.48329604e+00 -5.02710007e-02
1.07758832e+00 7.82946944e-01 -3.99216026e-01 6.89273715e-01
2.67621934e-01 1.79916823e+00 -3.33071262e-01 -7.10625052e-01
6.22883961e-02 -1.42929241e-01 -1.23625326e+00 1.34528846e-01
-1.49635235e-02 -3.05581868e-01 3.12954247e-01 9.43622470e-01
5.28692119e-02 -1.38868690e-01 1.02420318e+00 -1.83762681e+00
-2.61307865e-01 8.08892787e-01 7.92658478e-02 -4.05631244e-01
3.51490319e-01 2.00695589e-01 9.65551198e-01 1.95124798e-04
3.69254529e-01 1.49329388e+00 4.55576777e-01 8.11754405e-01
-1.21625268e+00 -2.42856249e-01 4.51915078e-02 1.78070396e-01
-1.03980494e+00 -4.68976945e-01 5.14305294e-01 -8.96040976e-01
6.94517672e-01 2.48245642e-01 1.48505539e-01 1.05608082e+00
4.64816540e-01 6.38300359e-01 1.20757568e+00 -4.69765931e-01
5.58928907e-01 7.22341120e-01 1.15268305e-01 5.18480241e-01
4.05149907e-01 2.01513663e-01 -3.41824621e-01 -5.45054018e-01
4.02592003e-01 -3.41465890e-01 1.65216371e-01 -8.06165934e-01
-1.01470172e+00 1.09840369e+00 -7.45241866e-02 4.36639115e-02
-2.78453648e-01 3.18598241e-01 4.29141641e-01 2.47331649e-01
5.67810684e-02 6.17855668e-01 -4.44839031e-01 -1.59917936e-01
-4.73556846e-01 5.70930779e-01 1.37268281e+00 8.80044341e-01
4.06620741e-01 -3.76738846e-01 6.58660848e-03 2.31499001e-01
6.02917969e-01 4.72559780e-01 1.12835243e-01 -1.45534813e+00
2.13068813e-01 6.38774812e-01 6.67581141e-01 -9.81097043e-01
-4.27148879e-01 3.51074070e-01 -4.54353541e-01 5.96399724e-01
5.39758563e-01 -3.06706995e-01 -2.55180269e-01 1.73439276e+00
3.58512640e-01 -3.32296342e-01 4.67589527e-01 6.19802654e-01
2.11445034e-01 2.41749585e-02 3.41234922e-01 -3.51592302e-01
1.22715735e+00 -3.63682985e-01 -5.84509790e-01 -2.65837848e-01
1.08680463e+00 -2.94874042e-01 6.64957583e-01 5.79273105e-01
-1.08325744e+00 -3.18829715e-02 -6.96519971e-01 -7.75507092e-02
5.22744358e-02 -3.41636539e-01 7.89772868e-01 7.29747057e-01
-9.09178376e-01 5.25660872e-01 -7.55316973e-01 -4.85423088e-01
3.90543461e-01 3.02798778e-01 -3.43144506e-01 2.14299977e-01
-1.35556662e+00 1.52757108e+00 1.96503893e-01 1.27631903e-03
-1.03701174e+00 -1.39854088e-01 -9.31024969e-01 -6.45795017e-02
7.19021738e-01 -5.94451845e-01 1.50436687e+00 -1.09493601e+00
-1.41353035e+00 1.11932731e+00 2.75087833e-01 -8.35378826e-01
8.86232316e-01 -1.22805290e-01 -2.06821367e-01 -2.51180053e-01
1.94088817e-01 4.96479094e-01 6.80187404e-01 -1.14370310e+00
-6.85716271e-01 -1.64688036e-01 5.84721625e-01 2.58211672e-01
-6.26779869e-02 5.02098262e-01 7.36381412e-01 1.31136835e-01
-4.34703350e-01 -1.14519620e+00 -5.52035391e-01 2.10277259e-01
-4.22327965e-01 -6.93186939e-01 2.02147111e-01 -1.38465255e-01
1.03420389e+00 -2.02679062e+00 3.04581784e-02 2.64309555e-01
1.20760702e-01 9.47823077e-02 4.09773529e-01 1.89563364e-01
-4.21373174e-02 2.04340577e-01 -3.85619223e-01 1.68203011e-01
6.30759656e-01 5.44893265e-01 -4.08091426e-01 4.03112590e-01
3.78421128e-01 3.68420005e-01 -1.11425650e+00 -1.08351302e+00
2.29146883e-01 -8.22365656e-02 -5.69223285e-01 5.00261843e-01
-1.30125120e-01 1.79999128e-01 -7.62424231e-01 2.35469401e-01
2.27446616e-01 6.34100735e-02 5.80904543e-01 5.13887584e-01
6.00443333e-02 3.63476992e-01 -1.13750446e+00 9.62255895e-01
-3.88483524e-01 1.11150451e-01 1.58879593e-01 -7.81845808e-01
7.59961963e-01 4.11454231e-01 1.61844403e-01 -2.02143192e-01
2.75224000e-01 1.50224203e-02 5.78334145e-02 -8.00154269e-01
2.42987007e-01 -8.31585348e-01 -8.28456342e-01 8.17219555e-01
-2.58475929e-01 -5.11569321e-01 -5.61560430e-02 1.88548625e-01
1.18871212e+00 3.46280217e-01 6.89027190e-01 -4.37975764e-01
6.80641055e-01 4.14969593e-01 7.97724187e-01 6.56677306e-01
-8.08103144e-01 -5.63219748e-02 1.22129929e+00 -7.19586432e-01
-7.54799664e-01 -9.36954141e-01 -1.55103445e-01 1.08405340e+00
1.93221703e-01 -7.11455345e-02 -9.78654802e-01 -1.01006389e+00
2.30342701e-01 1.41411817e+00 -5.68281293e-01 -2.33421713e-01
-2.85004824e-01 -3.13464046e-01 5.20920336e-01 4.66397941e-01
-9.51716900e-02 -1.19929147e+00 -1.21306705e+00 1.18869036e-01
1.15779312e-02 -9.43953097e-01 -4.36773971e-02 3.54866475e-01
-5.53006887e-01 -1.35288942e+00 1.58406109e-01 -2.96041548e-01
7.00083196e-01 -2.28460446e-01 1.30772734e+00 2.07409129e-01
-9.57288891e-02 5.11025429e-01 2.80302484e-02 -7.97018886e-01
-1.05607677e+00 -3.57693374e-01 5.15659451e-01 -2.55203903e-01
5.92537642e-01 -3.40138823e-01 -1.00235991e-01 5.63305497e-01
-8.42060089e-01 5.90035357e-02 3.94770026e-01 5.48053801e-01
1.00538798e-01 1.68777168e-01 5.01298964e-01 -1.34508085e+00
8.86260331e-01 -4.96036112e-01 -6.36105835e-01 4.47791219e-01
-9.33104038e-01 3.42170447e-01 4.85911101e-01 3.54484767e-02
-1.16058958e+00 1.88684851e-01 3.63418132e-01 -2.87585333e-02
-6.55091465e-01 4.47585583e-02 -3.99220139e-01 3.29724520e-01
8.05065334e-01 -4.62550014e-01 2.27822393e-01 1.23689398e-01
5.24831235e-01 9.46825385e-01 3.73444885e-01 -1.07416010e+00
6.57362998e-01 3.54650825e-01 8.05549994e-02 -3.08231652e-01
-1.01976323e+00 -1.27432585e-01 -6.61423922e-01 -1.89944550e-01
1.04387057e+00 -7.46323705e-01 -1.13879120e+00 5.97470291e-02
-9.31891620e-01 -4.10302311e-01 -3.67531300e-01 2.34300628e-01
-1.24423754e+00 2.57040471e-01 -1.69981018e-01 -1.30120039e+00
6.38542473e-02 -1.01278877e+00 7.30572879e-01 5.04750870e-02
-9.74862099e-01 -1.04765308e+00 1.17461137e-01 7.06886232e-01
7.21939132e-02 3.47549468e-01 1.15854084e+00 -9.77889299e-01
-2.00168625e-01 -2.38124549e-01 2.94321962e-02 5.36506951e-01
6.82672765e-03 -1.76964160e-02 -9.49936986e-01 2.52773073e-02
1.67549461e-01 -8.59165549e-01 1.47776417e-02 -1.30679570e-02
1.01492846e+00 -6.15846038e-01 -3.79381686e-01 -1.30760357e-01
9.12557602e-01 -6.97641820e-02 5.25077879e-01 3.25885028e-01
3.01376760e-01 1.42999995e+00 1.04910541e+00 4.64311481e-01
9.10943747e-01 4.95251745e-01 5.49348593e-01 3.71608466e-01
6.12893105e-01 -8.46563354e-02 3.92100662e-01 3.88154164e-02
-3.45649384e-02 1.84470370e-01 -1.28085220e+00 6.41251564e-01
-2.37178087e+00 -1.11448979e+00 -1.39096826e-01 2.02751851e+00
9.09430623e-01 4.13780183e-01 1.27540037e-01 7.69003406e-02
6.25529408e-01 -1.37336999e-01 -5.67236900e-01 -1.01499379e+00
3.93783718e-01 -5.40466607e-01 3.62291843e-01 6.61517203e-01
-1.09085727e+00 5.67034423e-01 6.57072067e+00 3.59256566e-01
-4.11064297e-01 7.81563867e-04 7.87331939e-01 2.43590642e-02
-4.30150747e-01 5.02254188e-01 -6.38109386e-01 2.94192404e-01
1.19119275e+00 -3.17612857e-01 1.30586699e-01 1.25752521e+00
2.36612841e-01 -3.12760234e-01 -1.75545180e+00 5.62537432e-01
-1.21102389e-03 -6.36291265e-01 -4.23287362e-01 6.43745065e-02
4.87695366e-01 -5.06199837e-01 -1.83213308e-01 4.69882786e-01
1.18875885e+00 -1.35298073e+00 1.00220704e+00 7.27315247e-01
2.72253841e-01 -9.01988029e-01 9.53561187e-01 8.46391141e-01
-2.42968440e-01 -4.72278178e-01 -3.15191060e-01 -3.24703157e-01
2.58147046e-02 5.99964857e-01 -1.05943465e+00 3.14713657e-01
3.65548104e-01 3.65049809e-01 -2.40930006e-01 4.13302869e-01
-5.92924953e-01 2.93266296e-01 -1.01690926e-01 -1.76581711e-01
9.30569842e-02 6.58282638e-02 1.83497027e-01 9.39942837e-01
-1.10441603e-01 6.35306239e-02 5.00429235e-02 1.00230598e+00
2.86349714e-01 -4.69829172e-01 -7.10857391e-01 7.96209574e-02
2.41199225e-01 1.39670467e+00 -3.82496297e-01 -2.11878479e-01
1.79131236e-02 3.82887006e-01 2.46431872e-01 -1.37884125e-01
-8.39463472e-01 -2.54139721e-01 8.92710924e-01 2.16052711e-01
-3.05212379e-01 2.10068896e-01 -3.44935268e-01 -8.41185629e-01
-1.81641266e-01 -9.23357606e-01 5.25110543e-01 -9.13544357e-01
-1.34744489e+00 3.13035786e-01 3.02308917e-01 -1.10223091e+00
-8.59306872e-01 -7.27555513e-01 -6.31627142e-01 5.86794794e-01
-1.50029516e+00 -9.11895573e-01 1.04614578e-01 4.22626555e-01
-1.53086875e-02 1.30227897e-02 1.02113962e+00 -4.06109840e-01
-2.85603583e-01 1.16324447e-01 -3.92384529e-01 -1.44066349e-01
8.32409739e-01 -1.53382063e+00 -7.60792121e-02 4.64845777e-01
-5.31415999e-01 4.03716505e-01 9.80493784e-01 -3.79514992e-01
-1.13439798e+00 -1.16085267e+00 1.39780700e+00 -1.55979848e+00
7.82607198e-01 -2.93786377e-01 -6.70195580e-01 7.08350420e-01
-5.44360839e-03 -1.35027885e-01 7.79357433e-01 2.49929667e-01
-3.43587488e-01 -5.23628257e-02 -1.49461341e+00 4.37936127e-01
8.52714717e-01 -4.71253723e-01 -1.12546563e+00 5.38293719e-01
2.75384337e-01 -3.62547822e-02 -7.60947227e-01 3.73232514e-02
3.60323191e-01 -1.16768539e+00 6.33953094e-01 -9.13647830e-01
5.68446040e-01 -3.72689098e-01 -8.50932896e-02 -1.04713321e+00
-2.82817245e-01 -7.40008533e-01 1.66892707e-01 1.18883228e+00
2.96751767e-01 -5.03514111e-01 5.50576270e-01 1.57137597e+00
2.79190719e-01 -5.68562269e-01 -9.42097247e-01 -5.37444949e-01
3.90491694e-01 -5.07390440e-01 5.08264542e-01 8.98142576e-01
5.08139253e-01 2.60068506e-01 -3.83280963e-01 1.74082041e-01
8.95051479e-01 1.04396969e-01 6.95233464e-01 -1.65440881e+00
-7.46005103e-02 -2.14082241e-01 -4.23176199e-01 -1.71693146e-01
6.28139496e-01 -7.09569693e-01 6.09063387e-01 -1.57998800e+00
3.17263007e-01 -7.85557091e-01 -1.29352555e-01 9.06640172e-01
-4.93938774e-02 -5.70339084e-01 8.98561403e-02 2.05074921e-01
-1.00789297e+00 2.19305694e-01 7.44194865e-01 1.55774608e-01
3.66878599e-01 6.05892017e-02 -1.20238745e+00 1.31261599e+00
6.17582440e-01 -7.70882249e-01 -5.05216181e-01 -1.32126525e-01
6.36623323e-01 2.98845977e-01 5.87991178e-01 -6.46566570e-01
3.38723063e-01 -1.01771784e+00 -2.21018359e-01 1.55094191e-01
-8.12985748e-02 -1.05064964e+00 1.25779942e-01 5.15708506e-01
-7.39684343e-01 -3.81194502e-01 -4.29037511e-01 5.68767667e-01
1.43136710e-01 -5.12711763e-01 7.66938686e-01 -1.87444732e-01
-5.17646968e-01 -2.90652692e-01 -7.84378111e-01 1.49240926e-01
1.50438035e+00 -5.44593856e-02 -1.93410724e-01 -4.69200104e-01
-6.41188443e-01 6.91145957e-01 7.52154469e-01 3.36065471e-01
3.29309016e-01 -1.10016608e+00 -7.39437521e-01 -1.84379183e-02
3.81376356e-01 1.14964157e-01 -2.64751539e-02 7.67898083e-01
-2.01946720e-01 4.94026721e-01 -2.11349487e-01 -3.43139976e-01
-1.12426758e+00 5.54238856e-01 4.52762872e-01 -2.91476995e-01
-1.29846990e-01 5.91614008e-01 3.82659137e-01 -7.91339397e-01
3.75098079e-01 -1.44058868e-01 -3.51205021e-01 -1.93841606e-01
3.90150577e-01 2.39876717e-01 -2.83613592e-01 -5.94090641e-01
-7.03360081e-01 1.13011993e-01 -1.23560568e-02 9.04996693e-02
1.23209715e+00 -1.61928266e-01 -1.68253943e-01 4.03416783e-01
2.25204468e-01 -3.96246821e-01 -1.35541320e+00 8.73085409e-02
6.23335421e-01 -1.98207051e-01 -2.39615783e-01 -8.18487763e-01
-1.74513906e-01 8.65443468e-01 1.76374927e-01 4.37421530e-01
7.72052109e-01 1.01041563e-01 2.06606612e-01 6.50713623e-01
7.86904454e-01 -1.40381026e+00 -1.53767660e-01 3.74807298e-01
7.60321558e-01 -1.26747251e+00 6.09704591e-02 -1.73153758e-01
-1.23885679e+00 8.33313286e-01 6.92584157e-01 -4.73935977e-02
4.67300326e-01 3.74489754e-01 3.29489470e-01 -1.39041632e-01
-1.19023514e+00 1.52315930e-01 -3.16570342e-01 7.32578874e-01
2.25634843e-01 4.61666852e-01 -1.10320643e-01 7.75467992e-01
-1.42992049e-01 2.79317975e-01 1.03612959e+00 1.30854261e+00
-5.51411152e-01 -1.19378471e+00 -4.67497170e-01 3.80456150e-01
-3.39742184e-01 4.20302004e-01 -7.74043739e-01 5.89039207e-01
3.70077550e-01 1.27564156e+00 -1.52405426e-01 -1.49280220e-01
3.82997662e-01 1.01381429e-02 3.18775237e-01 -8.74172330e-01
-5.20222664e-01 -6.58670723e-01 6.17145360e-01 -6.80016398e-01
-7.08604455e-01 -8.44536424e-01 -1.23497260e+00 -2.83035338e-01
-1.78575665e-01 4.06675637e-01 4.28701341e-01 1.16868293e+00
-2.95656808e-02 2.33186811e-01 7.41541624e-01 -5.14745831e-01
-1.31949162e+00 -5.49032152e-01 -6.18577242e-01 5.13125598e-01
1.43268481e-01 -6.19982302e-01 -5.73397219e-01 1.96019575e-01] | [8.816980361938477, 6.131811618804932] |
a83f279a-ee30-4498-ad2b-792676dffb65 | motiontrack-end-to-end-transformer-based | 2306.17000 | null | https://arxiv.org/abs/2306.17000v1 | https://arxiv.org/pdf/2306.17000v1.pdf | MotionTrack: End-to-End Transformer-based Multi-Object Tracing with LiDAR-Camera Fusion | Multiple Object Tracking (MOT) is crucial to autonomous vehicle perception. End-to-end transformer-based algorithms, which detect and track objects simultaneously, show great potential for the MOT task. However, most existing methods focus on image-based tracking with a single object category. In this paper, we propose an end-to-end transformer-based MOT algorithm (MotionTrack) with multi-modality sensor inputs to track objects with multiple classes. Our objective is to establish a transformer baseline for the MOT in an autonomous driving environment. The proposed algorithm consists of a transformer-based data association (DA) module and a transformer-based query enhancement module to achieve MOT and Multiple Object Detection (MOD) simultaneously. The MotionTrack and its variations achieve better results (AMOTA score at 0.55) on the nuScenes dataset compared with other classical baseline models, such as the AB3DMOT, the CenterTrack, and the probabilistic 3D Kalman filter. In addition, we prove that a modified attention mechanism can be utilized for DA to accomplish the MOT, and aggregate history features to enhance the MOD performance. | ['Michael Happold', 'Lingji Chen', 'Yiluan Guo', 'Chengjie Zhang', 'Ce Zhang'] | 2023-06-29 | null | null | null | null | ['object-tracking', 'multiple-object-tracking'] | ['computer-vision', 'computer-vision'] | [ 1.66621823e-02 -4.30140525e-01 -1.69712305e-01 -3.12938899e-01
-1.00777686e+00 -4.09072727e-01 7.96195984e-01 -5.05331494e-02
-6.08993649e-01 3.35131556e-01 -2.75702745e-01 -9.47477110e-03
-1.40347049e-01 -7.14798272e-01 -9.52757359e-01 -9.69792068e-01
3.12661380e-01 6.32008851e-01 1.21709156e+00 -2.55416214e-01
1.15788601e-01 2.96201020e-01 -2.11386514e+00 8.48339722e-02
7.99454391e-01 1.19613481e+00 6.90307796e-01 7.36872554e-01
-1.12574175e-01 7.05945015e-01 -2.85349876e-01 -4.08845991e-01
1.78721905e-01 1.36990175e-01 -1.16551280e-01 -3.34596425e-01
9.18670475e-01 -3.78076702e-01 -4.38612282e-01 1.04824007e+00
6.38819993e-01 1.39323249e-01 4.99279678e-01 -1.73064911e+00
-2.60692924e-01 3.30218107e-01 -5.55638433e-01 4.34812039e-01
-2.56371498e-02 2.57069528e-01 7.68570602e-01 -1.08283079e+00
3.96572322e-01 1.53629005e+00 6.79156005e-01 4.35640663e-01
-7.98209190e-01 -1.04435873e+00 1.62524551e-01 7.51496315e-01
-1.43440270e+00 -4.88256603e-01 5.82775235e-01 -4.53057349e-01
5.65401137e-01 2.71631449e-01 6.09302819e-01 8.18081975e-01
5.06702423e-01 1.29629958e+00 7.84861684e-01 -6.49116933e-02
-1.20088533e-01 2.93989003e-01 3.47553879e-01 6.86291277e-01
3.33408833e-01 5.37020087e-01 -7.21921623e-01 2.21837983e-01
1.40631512e-01 2.82591164e-01 1.91972151e-01 -5.09391129e-01
-1.33264983e+00 6.51209772e-01 5.23381710e-01 1.47843093e-01
-4.06150758e-01 3.93653184e-01 1.18606262e-01 3.56868766e-02
1.68352768e-01 -1.26523867e-01 -1.27303228e-01 2.71199085e-02
-7.59357989e-01 4.77070093e-01 1.55544236e-01 1.27834189e+00
8.82302463e-01 -1.04367934e-01 -8.66092384e-01 3.82185429e-01
6.63072586e-01 1.32214069e+00 2.45646864e-01 -8.43915105e-01
2.92544693e-01 4.59216744e-01 3.09675068e-01 -9.27159786e-01
-3.69046897e-01 -5.39324224e-01 -4.85709131e-01 3.91527861e-01
2.93109976e-02 2.74649352e-01 -9.82138097e-01 1.81693566e+00
7.00917363e-01 5.53255677e-01 1.75816223e-01 8.53302300e-01
1.05408919e+00 6.03001177e-01 1.95672125e-01 -1.62860051e-01
1.54772484e+00 -1.10993481e+00 -1.10452497e+00 -2.29490593e-01
7.08605111e-01 -7.44727015e-01 6.73569083e-01 9.66554508e-02
-8.10964584e-01 -9.17645335e-01 -1.05658436e+00 2.89667584e-02
-4.35186327e-01 1.59294397e-01 1.77280352e-01 6.08037353e-01
-9.12449002e-01 -3.06263827e-02 -1.03944099e+00 -5.34174740e-01
4.37961191e-01 3.84088576e-01 -6.62512332e-03 -2.41433099e-01
-9.93403971e-01 1.07666183e+00 6.02857888e-01 -1.21140875e-01
-1.29462981e+00 -5.77500999e-01 -5.47706485e-01 -1.92666546e-01
5.55257738e-01 -8.32473338e-01 1.23656929e+00 -2.86144108e-01
-1.07271278e+00 6.34516358e-01 -5.84831595e-01 -7.57807195e-01
4.88373756e-01 -3.80013615e-01 -4.86370087e-01 1.71251725e-02
3.53531420e-01 9.38895524e-01 9.15727198e-01 -1.26869941e+00
-1.43438685e+00 -3.74466956e-01 3.75833036e-03 3.26854557e-01
-3.27792913e-01 -2.10779309e-02 -8.38490784e-01 -2.90676564e-01
5.15044257e-02 -1.00675225e+00 3.45964059e-02 2.55807221e-01
-1.01034202e-01 -6.27363384e-01 1.53233016e+00 -2.10057721e-01
1.02198613e+00 -2.27343607e+00 -3.46981660e-02 -3.17024857e-01
4.28543627e-01 2.87010700e-01 -5.02500422e-02 -7.17490315e-02
7.31047094e-01 -5.06786704e-01 1.39403462e-01 -6.11817956e-01
1.47998840e-01 2.17238694e-01 -1.44838676e-01 4.30928439e-01
1.72985330e-01 1.11999190e+00 -9.59895194e-01 -9.33333397e-01
5.58430076e-01 2.84422457e-01 -4.15978849e-01 8.66892338e-02
-3.62686843e-01 3.60141844e-01 -5.52252114e-01 8.06638181e-01
1.01566565e+00 -1.44875690e-01 -3.72914374e-01 -5.27717888e-01
-4.81508046e-01 -2.38853455e-01 -1.15979874e+00 1.45862186e+00
-3.09566595e-02 7.26122856e-01 1.36227205e-01 -5.33882201e-01
7.63159335e-01 9.39043760e-02 6.02360666e-01 -9.56428945e-01
1.67319506e-01 1.53308183e-01 -2.08724588e-01 -4.08285052e-01
1.08320200e+00 1.86157957e-01 -2.10364293e-02 -1.69631615e-01
-1.13202833e-01 3.06495488e-01 2.19053835e-01 4.05605823e-01
1.12652075e+00 4.59631607e-02 -2.12302759e-01 -7.92325959e-02
5.30338109e-01 4.60515529e-01 7.82636821e-01 9.56420362e-01
-5.44023097e-01 1.21528395e-01 -5.28099000e-01 -4.66116667e-02
-7.71990001e-01 -1.12704504e+00 -2.75490940e-01 1.12654781e+00
1.17524314e+00 -2.74960130e-01 -1.61679059e-01 -5.30784309e-01
2.35137016e-01 6.88571751e-01 -3.23688507e-01 -4.45761621e-01
-5.76825738e-01 -7.13273883e-01 5.31673133e-01 5.76142967e-01
8.38655949e-01 -6.49525821e-01 -8.86381924e-01 3.38380665e-01
-5.58602810e-01 -1.35780084e+00 -4.42974240e-01 6.09069057e-02
-5.09148777e-01 -9.33237135e-01 -4.44356591e-01 -5.75331032e-01
1.12656727e-01 8.95216286e-01 5.70072770e-01 -1.47065833e-01
9.90984067e-02 4.10735875e-01 -2.42922425e-01 -8.74433398e-01
-2.81809032e-01 -2.43309975e-01 3.02017629e-01 2.79205173e-01
5.28435290e-01 -8.94381329e-02 -6.15981400e-01 7.96770215e-01
-7.13330388e-01 -1.59364508e-03 9.27682877e-01 6.06414974e-01
6.99877203e-01 -5.14412709e-02 2.74909079e-01 -1.88470051e-01
-2.43093386e-01 -4.85365808e-01 -8.86823714e-01 2.11071134e-01
-7.70418704e-01 -1.03098668e-01 5.40325716e-02 -5.82256854e-01
-9.30891335e-01 2.96870977e-01 -1.25056933e-04 -9.62974429e-01
8.29155073e-02 6.95706457e-02 -2.34271899e-01 -2.92308301e-01
4.71820205e-01 4.38537538e-01 -1.64636150e-01 -3.59867066e-01
3.41449261e-01 5.98054767e-01 7.99971044e-01 -1.98637247e-01
1.14862454e+00 8.08383942e-01 1.54606715e-01 -5.33791363e-01
-7.52729058e-01 -1.02321255e+00 -3.16235989e-01 -7.20208466e-01
1.02028942e+00 -1.39991677e+00 -9.41220284e-01 5.78660309e-01
-1.14536119e+00 9.68517885e-02 2.40752883e-02 7.92919695e-01
-4.34425682e-01 1.40151694e-01 -2.12997827e-03 -9.68734443e-01
-2.13733062e-01 -1.22047341e+00 1.55385268e+00 4.48522240e-01
4.98704910e-01 -3.95163447e-01 1.60989501e-02 5.10185361e-01
4.76144999e-01 1.53436037e-02 1.43147603e-01 -4.86377776e-01
-1.38673174e+00 -8.89122114e-02 -3.56634408e-01 -2.61435121e-01
-1.89723730e-01 -2.17224330e-01 -1.14471984e+00 -5.58938384e-01
-3.14739525e-01 -9.41867009e-02 1.28571379e+00 3.88726711e-01
5.27671933e-01 7.27586076e-02 -1.15121186e+00 4.74199116e-01
1.40500832e+00 4.63950872e-01 4.15360153e-01 3.80175561e-01
7.83415496e-01 1.00404866e-01 1.28832519e+00 2.48482436e-01
7.79156029e-01 1.15794539e+00 7.93198287e-01 2.29083404e-01
-4.51851845e-01 -1.08175471e-01 6.43723965e-01 6.15126491e-01
3.80561829e-01 -6.09137893e-01 -7.37191737e-01 9.09875751e-01
-2.19183660e+00 -1.09360111e+00 -4.69084501e-01 1.98797238e+00
2.51627922e-01 4.59259063e-01 1.89508274e-01 2.98821973e-03
8.52029204e-01 1.23911843e-01 -7.62594402e-01 4.97862488e-01
-2.76188880e-01 -5.26683211e-01 7.61279762e-01 3.00206035e-01
-1.26442420e+00 9.88421440e-01 5.38174343e+00 1.25063872e+00
-8.99913728e-01 4.58220810e-01 -3.35951805e-01 -7.62085095e-02
-1.29431441e-01 -5.00606038e-02 -1.71796584e+00 4.01817530e-01
9.93405223e-01 -2.20889777e-01 -3.95712666e-02 7.58839965e-01
1.54200152e-01 -1.18757978e-01 -9.06563818e-01 1.08100760e+00
6.12920113e-02 -1.28201079e+00 3.79186012e-02 -8.50109309e-02
4.91429746e-01 4.11157548e-01 1.79052979e-01 5.57754278e-01
3.33091438e-01 -3.22946459e-01 1.21573889e+00 6.95061386e-01
5.63762009e-01 -3.57908070e-01 6.92047417e-01 5.49880326e-01
-1.85710406e+00 -2.88371325e-01 -2.17356771e-01 4.32858527e-01
3.12967569e-01 3.85556877e-01 -8.32762539e-01 8.83352637e-01
1.05198479e+00 8.35474849e-01 -6.43735349e-01 1.46669495e+00
4.05752361e-01 5.66707671e-01 -7.10689068e-01 -2.42453262e-01
2.29726925e-01 2.93456286e-01 1.11725831e+00 9.06422436e-01
4.23536807e-01 -2.08367463e-02 4.93722767e-01 7.42261887e-01
9.43809971e-02 -2.86609143e-01 -5.99143684e-01 2.07649931e-01
7.74342954e-01 1.26207566e+00 -3.91920388e-01 -5.80098033e-01
-4.18548197e-01 4.10778731e-01 -4.75431085e-02 -4.63392062e-04
-1.26462388e+00 -2.61990100e-01 6.54141486e-01 3.61742564e-02
7.87151754e-01 -1.34751812e-01 -1.19841779e-02 -7.84186661e-01
1.10466406e-01 -4.31061685e-01 5.05921900e-01 -1.02598727e+00
-7.82925129e-01 3.64906132e-01 1.16338238e-01 -1.64300334e+00
2.89980143e-01 -3.97409588e-01 -2.83130676e-01 3.71488839e-01
-1.89703369e+00 -1.41832769e+00 -7.54312456e-01 7.89111912e-01
5.56866288e-01 -1.08249106e-01 1.57017127e-01 8.84853542e-01
-4.59432364e-01 7.55849421e-01 3.62388402e-01 -2.45936662e-01
6.50018811e-01 -8.54474008e-01 -1.03246644e-02 1.05370831e+00
3.62565443e-02 6.06842898e-02 8.70757878e-01 -7.31264949e-01
-1.86854768e+00 -1.79785895e+00 5.45216262e-01 -8.21551859e-01
2.99596846e-01 -1.52044803e-01 -6.81620598e-01 6.55012429e-01
1.26606613e-01 7.50027373e-02 -1.49406403e-01 -3.24713975e-01
-5.47134690e-02 -5.32672763e-01 -1.13291717e+00 3.58669668e-01
1.14856720e+00 -1.29701674e-01 -5.08385718e-01 2.99972713e-01
1.10027122e+00 -5.43456614e-01 -6.13500297e-01 8.37583542e-01
5.84740281e-01 -5.76752007e-01 1.00349212e+00 -1.10719018e-01
-3.64112467e-01 -1.02220082e+00 -5.62043250e-01 -6.93732679e-01
-4.68199104e-01 -5.16098678e-01 -3.65257829e-01 1.25785768e+00
4.61747758e-02 -5.54199874e-01 7.11490333e-01 -9.93307158e-02
-6.01201475e-01 -1.81500763e-01 -1.27318513e+00 -1.13167119e+00
-4.96523529e-01 -4.60463613e-01 5.98775029e-01 3.65694672e-01
-6.37476444e-01 3.28407615e-01 -4.58258480e-01 9.34951246e-01
1.01949453e+00 2.74383366e-01 1.08301413e+00 -1.35037315e+00
1.33355781e-01 -2.49629632e-01 -6.53246880e-01 -1.45740116e+00
-2.96647340e-01 -8.25207651e-01 5.82060754e-01 -1.41499829e+00
3.43619972e-01 -6.10839665e-01 -7.41163790e-01 2.44798541e-01
-2.58946151e-01 2.10230321e-01 3.00060689e-01 4.03098106e-01
-1.34414732e+00 9.82474804e-01 1.13947034e+00 -6.50041461e-01
6.02414832e-02 3.95076051e-02 -2.50513971e-01 3.01674813e-01
3.78089100e-01 -8.67083967e-01 -2.05656126e-01 -4.44758624e-01
-2.11467713e-01 5.01811206e-02 7.43658066e-01 -1.52335334e+00
7.87974656e-01 -1.57861367e-01 -1.56022073e-03 -1.66843522e+00
8.67636383e-01 -9.62279975e-01 3.26321691e-01 7.39851832e-01
3.26508060e-02 1.78837880e-01 4.47213739e-01 1.06381738e+00
-1.48175567e-01 2.76159942e-01 8.55943739e-01 3.26342732e-01
-1.26931953e+00 5.79654992e-01 -3.54175776e-01 -2.30947956e-01
1.35282874e+00 -2.12671384e-01 -6.69167459e-01 3.37028541e-02
-2.33976871e-01 9.08336759e-01 6.41502291e-02 7.28411853e-01
7.00161994e-01 -1.73850608e+00 -7.61272728e-01 2.06492841e-02
6.18261874e-01 -1.22664012e-01 1.89128473e-01 1.32965553e+00
8.20628032e-02 5.57129562e-01 -1.67625904e-01 -1.37771523e+00
-1.72715783e+00 6.22230172e-01 2.20522985e-01 -5.21811545e-02
-5.64829707e-01 8.18939030e-01 2.87674367e-01 -2.21478581e-01
3.70190382e-01 -2.23821163e-01 -2.45160267e-01 -1.13970526e-01
4.68126982e-01 3.57643068e-01 -3.51477973e-02 -8.43213201e-01
-6.79729581e-01 7.45541155e-01 -9.32474956e-02 -1.05769455e-03
9.04681504e-01 -5.23988903e-01 4.62201446e-01 4.21364844e-01
7.41109431e-01 -1.70148373e-01 -1.17687738e+00 -5.68776906e-01
-2.77769148e-01 -4.57478940e-01 2.80047864e-01 -6.67736351e-01
-8.64037097e-01 7.50674367e-01 1.41766560e+00 -6.03811769e-03
9.35048103e-01 1.92888126e-01 1.05701184e+00 6.11243606e-01
6.22566879e-01 -6.82158649e-01 1.05719857e-01 5.07286429e-01
4.73591536e-01 -1.48718345e+00 -3.03748280e-01 -1.96492538e-01
-3.32827538e-01 5.48008263e-01 9.03328121e-01 4.04422700e-01
6.02712274e-01 1.81586474e-01 3.57365073e-03 -3.92640382e-01
-1.01835799e+00 -7.11055756e-01 3.37454647e-01 4.94474649e-01
-5.49740076e-01 -1.10148624e-01 -3.98089625e-02 1.33959919e-01
2.00564966e-01 3.76723148e-02 2.41193473e-02 9.95645106e-01
-9.93872225e-01 -6.41874313e-01 -6.31922424e-01 4.04707372e-01
-1.25050006e-04 1.73600182e-01 1.62110925e-01 7.77785182e-01
4.39935476e-01 1.28381944e+00 -8.25476497e-02 -9.21913266e-01
4.31377262e-01 -1.73167378e-01 3.07056040e-01 -4.03651260e-02
-2.66548663e-01 5.97948506e-02 3.77370343e-02 -6.11709177e-01
-6.72153652e-01 -8.67346644e-01 -1.24788284e+00 -3.18797201e-01
-8.83853614e-01 -8.90753139e-03 6.12371981e-01 9.65000033e-01
5.11626244e-01 7.42093921e-01 4.91788805e-01 -6.48937702e-01
-3.11849058e-01 -8.51262450e-01 -2.38342613e-01 1.96498543e-01
7.02133715e-01 -1.30738997e+00 -9.66730118e-02 -1.26615083e-02] | [6.635846138000488, -2.1578176021575928] |
1bc92cab-f067-4811-9dfa-0acfe0f5eab8 | a-cognitively-driven-weighted-entropy-model | 2112.06876 | null | https://arxiv.org/abs/2112.06876v2 | https://arxiv.org/pdf/2112.06876v2.pdf | A cognitively driven weighted-entropy model for embedding semantic categories in hyperbolic geometry | In this paper, an unsupervised and cognitively driven weighted-entropy method for embedding semantic categories in hyperbolic geometry is proposed. The model is driven by two fields of research in cognitive linguistics: the first is the statistical learning theory of language acquisition and the proposal of using high-dimensional networks to represent semantic knowledge in cognition, and the second is the domain-specific approach to semantic communication. Weighted conditional entropy of word co-occurrence is proposed as the embedding metric, and the two weighting parameters are collocation diversity and conditional probability ranking in the corresponding statistical distribution. The Boltzmann distribution is then used on the weighted-entropy metric and embedded into a hyperbolic Poincare disk model. Testing has been in particular performed in the domains of basic color and kinship words, which belong to the classes that domain-specificity focused research in cognitive semantics has most intensively investigated. Results show that this new approach can successfully model and map the semantic relationships of popularity and similarity for most of the basic color and kinship words in English and have potential to be generalized to other semantic domains and different languages. Generally, this paper contributes to both computational cognitive semantics and the research on network and geometry-driven language embedding in computational linguistics and NLP. | ['Eugene Yu Ji'] | 2021-12-13 | null | null | null | null | ['language-acquisition'] | ['natural-language-processing'] | [-1.74688518e-01 6.34777367e-01 1.68707147e-02 -3.95216078e-01
2.66724676e-01 -5.23906291e-01 7.72998273e-01 5.70626855e-01
-9.69329178e-01 3.16117227e-01 4.25866246e-01 -3.25435430e-01
-9.79420960e-01 -1.22214222e+00 3.46880890e-02 -6.64103031e-01
-3.42285782e-01 8.05772364e-01 3.55481803e-01 -4.58286047e-01
4.51758623e-01 3.54465783e-01 -1.64494145e+00 -3.21469933e-01
8.37623239e-01 5.90821147e-01 4.28751409e-01 4.30597216e-01
-4.21705693e-01 4.98646021e-01 -2.43864328e-01 -5.82874656e-01
-3.39528650e-01 -5.54682493e-01 -1.21691489e+00 -4.94586200e-01
-2.31725797e-01 4.06639159e-01 -4.17166501e-01 1.21431112e+00
4.29995835e-01 4.35013175e-01 1.23891568e+00 -1.06204712e+00
-1.19851387e+00 9.56463397e-01 2.72612751e-01 3.64180833e-01
2.72664934e-01 -8.57278407e-01 1.22646058e+00 -9.12324548e-01
7.39551306e-01 1.51814151e+00 4.29941535e-01 3.58032912e-01
-1.04220724e+00 -6.48402050e-02 -2.92468488e-01 7.11214244e-01
-1.62370467e+00 4.73060936e-01 8.95969212e-01 -7.99561918e-01
6.28197730e-01 4.29571979e-02 1.06115973e+00 7.49241829e-01
1.65992931e-01 1.94709882e-01 1.25193775e+00 -7.75947630e-01
7.17296124e-01 5.72731078e-01 3.61571670e-01 9.01551425e-01
3.67672950e-01 1.15279056e-01 -4.76046413e-01 2.39552930e-02
5.58461666e-01 -3.34768891e-01 2.82981962e-01 -6.52794898e-01
-1.19013107e+00 1.39230311e+00 5.93824506e-01 9.90715921e-01
-1.54256493e-01 -5.46341464e-02 4.08211589e-01 1.21527895e-01
8.03463519e-01 4.77508098e-01 -1.19673669e-01 5.63376546e-02
-6.98800325e-01 3.70419979e-01 8.76288235e-01 6.70127809e-01
6.99901760e-01 -2.47626439e-01 2.54482448e-01 9.67758179e-01
7.07520902e-01 5.02609015e-01 8.33848774e-01 -6.53519154e-01
-1.48372203e-01 6.25919223e-01 -5.83786905e-01 -1.52842546e+00
-7.25010753e-01 -8.24778825e-02 -2.59601563e-01 -5.23185134e-02
2.68295616e-01 2.07458243e-01 -3.02263230e-01 1.90326858e+00
1.11938797e-01 -2.07665130e-01 2.91758448e-01 8.80694211e-01
1.00869846e+00 3.38596135e-01 3.02410752e-01 5.50649092e-02
1.84864771e+00 -2.50134379e-01 -7.28291154e-01 1.33142084e-01
8.57543766e-01 -4.79796618e-01 1.04435575e+00 -1.72758490e-01
-9.11711752e-01 -3.84201199e-01 -1.09199214e+00 -2.26895645e-01
-1.41149950e+00 -5.08472025e-01 6.09511435e-01 6.99461579e-01
-1.38947117e+00 3.70631814e-01 -3.57025325e-01 -8.58283520e-01
1.64377913e-01 3.28090936e-02 -8.09651017e-02 1.32417828e-01
-1.91127658e+00 1.33630443e+00 8.99598658e-01 -4.51103032e-01
-1.36154547e-01 -3.24954957e-01 -1.08158302e+00 1.15684733e-01
-2.39474133e-01 -6.04053319e-01 4.81389821e-01 -7.01524079e-01
-1.26026380e+00 1.39477789e+00 2.43462801e-01 -4.21910882e-01
-2.44562719e-02 3.17828655e-01 -6.05015934e-01 2.96225220e-01
1.68642655e-01 8.58292282e-01 3.92233849e-01 -1.03457189e+00
-2.26173624e-01 -5.39567232e-01 1.23327240e-01 5.07483423e-01
-9.00113046e-01 -1.18683852e-01 2.30236560e-01 -8.09014320e-01
4.30600226e-01 -9.63451684e-01 2.62309730e-01 -1.65200114e-01
1.60681233e-01 -7.21662402e-01 4.09888148e-01 -6.93005741e-01
1.15993512e+00 -2.20447016e+00 5.35139620e-01 7.36855745e-01
4.19730514e-01 -1.97406992e-01 1.32390663e-01 5.19531846e-01
5.61823808e-02 8.38427395e-02 -3.00131619e-01 3.36505592e-01
4.02531892e-01 3.08077574e-01 -6.40733773e-03 3.64481866e-01
-3.40313256e-01 8.27334166e-01 -1.26325762e+00 -7.37129033e-01
1.84444875e-01 6.20112777e-01 -5.48345327e-01 -2.58220255e-01
8.10512975e-02 -2.30036050e-01 -3.62529129e-01 -1.97452586e-02
4.00009334e-01 1.06965609e-01 2.48017877e-01 -7.28575587e-02
-1.41919300e-01 2.50485301e-01 -8.23785722e-01 1.68753099e+00
-4.43170398e-01 9.70425844e-01 -6.19952917e-01 -1.22562551e+00
1.26122534e+00 2.16024052e-02 1.49826586e-01 -5.86378634e-01
3.36386383e-01 2.28265226e-01 4.01417375e-01 -5.52546144e-01
8.26856554e-01 -3.59589130e-01 -3.34135383e-01 6.28624558e-01
3.95614296e-01 -5.32481015e-01 1.97183758e-01 3.79781812e-01
6.59701467e-01 -1.21465093e-02 3.52815717e-01 -1.06021118e+00
6.63194954e-01 -2.39060774e-01 -1.27122179e-01 1.83059067e-01
-1.52639821e-01 8.75626877e-02 4.36476469e-01 -1.15831725e-01
-1.12611771e+00 -1.60179794e+00 -5.25460541e-01 1.46226811e+00
2.99800396e-01 -4.12727892e-01 -1.08539271e+00 -3.06091130e-01
-6.09493889e-02 1.22859359e+00 -7.64676332e-01 -5.65227866e-01
3.28694098e-02 -5.89164376e-01 6.03513837e-01 3.29702199e-01
3.42087209e-01 -1.27634370e+00 -5.30862987e-01 -7.02659637e-02
-7.86804631e-02 -7.52477348e-01 -3.26613970e-02 5.14073558e-02
-7.49467075e-01 -1.02321815e+00 -6.62612438e-01 -1.20917368e+00
5.34809053e-01 -2.40777045e-01 9.99030352e-01 -1.90765768e-01
-5.67011416e-01 1.01878619e+00 -6.72893286e-01 -6.34189844e-01
-3.82236391e-01 3.81663279e-03 2.76344836e-01 -2.68857569e-01
1.00108600e+00 -6.59238577e-01 -4.41978127e-01 1.32492378e-01
-1.12442577e+00 -4.76604074e-01 9.89726931e-02 5.91463804e-01
2.37883963e-02 1.20319262e-01 4.86780077e-01 -4.75112855e-01
9.72715318e-01 -7.00246930e-01 -1.00868329e-01 1.30733684e-01
-7.13997900e-01 2.84540862e-01 2.49517579e-02 -2.14419663e-01
-8.18841040e-01 -5.86452842e-01 1.59272298e-01 2.62914836e-01
-1.08457126e-01 8.58198047e-01 1.46078035e-01 4.90374938e-02
6.99925005e-01 4.74250138e-01 3.82294446e-01 -8.70012790e-02
8.00239563e-01 5.67599833e-01 4.23428826e-02 -5.11226296e-01
4.93230134e-01 3.73310149e-01 1.83197871e-01 -1.42254257e+00
-5.40950358e-01 -2.48711556e-01 -8.56021345e-01 -3.31718892e-01
1.44629240e+00 -4.81319904e-01 -6.96267068e-01 8.27524215e-02
-9.40757811e-01 1.51346281e-01 -5.55877805e-01 9.68645334e-01
-9.19099271e-01 4.26421762e-01 -4.02111471e-01 -7.11690962e-01
4.35912125e-02 -5.49304068e-01 7.20002592e-01 -2.49603204e-02
-5.08018553e-01 -2.00738883e+00 3.77080172e-01 7.28888586e-02
3.64743799e-01 -1.06582001e-01 1.53291094e+00 -9.21657324e-01
8.09915140e-02 9.48843062e-02 -1.89763978e-01 3.97612065e-01
-2.30094805e-01 -3.57491791e-01 -6.45500124e-01 2.47888844e-02
8.74402002e-02 -2.85849452e-01 6.88470840e-01 1.94371909e-01
8.92820776e-01 1.04773782e-01 -1.23554997e-01 9.29299667e-02
1.43434834e+00 1.70955315e-01 5.33429325e-01 3.49682540e-01
3.09020251e-01 1.17942929e+00 3.59754235e-01 1.96213067e-01
8.54087770e-01 5.22056341e-01 5.96798249e-02 2.42013723e-01
1.20715514e-01 -1.51135236e-01 1.83413416e-01 1.53087878e+00
-2.63650209e-01 7.37870932e-02 -1.20294034e+00 6.85358524e-01
-1.62699640e+00 -1.12139809e+00 2.00691931e-02 1.78939819e+00
6.05233669e-01 -1.19849913e-01 1.28151879e-01 3.17191243e-01
9.28216755e-01 1.58736616e-01 2.31161695e-02 -6.95091367e-01
-3.15896779e-01 3.75146776e-01 2.73683220e-01 8.60680819e-01
-8.90936732e-01 1.13795447e+00 6.12137365e+00 9.69605982e-01
-5.25220871e-01 2.53641903e-01 9.44857001e-02 4.39590156e-01
-5.16417444e-01 -8.62261876e-02 -3.05973887e-01 4.19148088e-01
9.76339817e-01 -5.01711667e-01 2.84492373e-01 6.48938775e-01
-9.56820175e-02 -1.31698012e-01 -9.24867094e-01 8.86012793e-01
6.12308204e-01 -8.92689586e-01 1.51532620e-01 2.69286513e-01
3.15706015e-01 -3.68397653e-01 1.65594950e-01 1.23898759e-01
2.28416011e-01 -9.47770655e-01 7.53681660e-01 9.07252491e-01
6.44784451e-01 -9.11557794e-01 7.64070630e-01 4.66827080e-02
-1.08006299e+00 -6.49037659e-02 -7.46715844e-01 -1.70591429e-01
-5.52297160e-02 4.47643340e-01 -6.75839901e-01 2.09045857e-01
6.18801296e-01 6.68741286e-01 -6.39048040e-01 5.31977773e-01
-2.42415100e-01 5.12305915e-01 -8.72631595e-02 -9.09397602e-01
2.83848852e-01 -5.69301903e-01 6.27196014e-01 1.24970400e+00
5.17956972e-01 1.69800803e-01 -2.17431441e-01 1.06937420e+00
4.69829172e-01 5.89204371e-01 -7.83195913e-01 -2.26970196e-01
6.51317358e-01 1.07784474e+00 -1.22967041e+00 -1.72222748e-01
-1.56731948e-01 8.22738886e-01 2.88034916e-01 1.00582436e-01
-7.32399702e-01 -7.23068535e-01 5.29725492e-01 2.18536258e-01
4.89495918e-02 -4.92418021e-01 -1.48047030e-01 -7.66331255e-01
-4.29614991e-01 2.71797121e-01 3.93003613e-01 -6.87045157e-01
-1.49162579e+00 5.22923887e-01 5.12028575e-01 -6.19824946e-01
-1.34215504e-01 -1.12169325e+00 -2.34159827e-01 8.79696071e-01
-8.76603603e-01 -8.54699612e-01 1.55649167e-02 8.13623607e-01
4.74357232e-03 -6.64036691e-01 1.31031430e+00 4.80889864e-02
3.09709370e-01 2.73150086e-01 3.48935813e-01 2.04042464e-01
2.43869886e-01 -1.68278718e+00 2.21201051e-02 -6.46704286e-02
2.77671933e-01 4.48046654e-01 6.03112221e-01 -4.61059928e-01
-6.81827903e-01 -6.17378175e-01 1.51267970e+00 -4.64536577e-01
1.04186213e+00 -6.64303184e-01 -6.71683133e-01 1.07114375e-01
6.13354482e-02 -4.76323396e-01 1.20723081e+00 1.05767176e-01
-2.96514690e-01 2.29145527e-01 -1.17944181e+00 7.08501637e-01
1.25674605e+00 -9.37049031e-01 -1.29521322e+00 6.20393097e-01
7.58234322e-01 4.37496632e-01 -9.86624539e-01 -2.67675668e-01
3.28514576e-01 -8.82534742e-01 1.06814277e+00 -4.56278741e-01
2.45484665e-01 7.26463571e-02 -4.55146611e-01 -1.53795683e+00
-6.34383023e-01 7.82989189e-02 2.73454189e-01 1.02085829e+00
2.22295836e-01 -8.18464041e-01 7.30019629e-01 1.65589780e-01
8.12593326e-02 -3.54073554e-01 -1.14412427e+00 -7.71135390e-01
5.20042300e-01 -4.11588699e-01 3.35941344e-01 1.15135562e+00
6.15583301e-01 4.37196940e-01 4.49548900e-01 -4.04248387e-01
7.36005008e-01 -4.97199953e-01 -3.02775145e-01 -1.94061375e+00
2.18802109e-01 -7.44795322e-01 -1.30234289e+00 -2.84513533e-01
7.43455648e-01 -1.65660155e+00 -1.97476134e-01 -1.44378614e+00
2.40913648e-02 -3.16433102e-01 -2.28951082e-01 -2.43865132e-01
3.42276186e-01 1.48020506e-01 9.01792794e-02 -1.05245352e-01
-2.77304471e-01 6.80096865e-01 8.66586328e-01 1.18557326e-01
-3.00171003e-02 -5.42439997e-01 -5.29356003e-01 9.67005074e-01
9.19279337e-01 -3.32116574e-01 -8.83861363e-01 1.05244450e-01
7.48798430e-01 -4.32601690e-01 6.38222575e-01 -8.96669328e-01
2.12664112e-01 1.06730610e-01 1.42083108e-01 -1.09715842e-01
3.95122558e-01 -9.77594912e-01 -3.61911446e-01 4.29847687e-01
-5.35411775e-01 2.59674251e-01 -1.99294627e-01 5.89727759e-01
-1.75472170e-01 -4.82823968e-01 7.77844489e-01 -2.56312173e-02
-7.98563242e-01 -9.21496078e-02 -7.22425818e-01 5.12631953e-01
1.26205862e+00 -5.33754408e-01 -1.03459395e-02 -3.34664404e-01
-1.27558160e+00 -2.95922518e-01 9.87270921e-02 6.37437403e-01
8.03963780e-01 -1.70805824e+00 -7.60101080e-01 7.96152055e-02
1.75121352e-01 -7.99088359e-01 1.92311168e-01 4.51274455e-01
-7.32384801e-01 6.21058285e-01 -2.76199251e-01 -4.25987542e-01
-9.98792171e-01 5.92143655e-01 1.12718992e-01 3.25184792e-01
-3.06868941e-01 8.18664372e-01 4.22476754e-02 -7.34918237e-01
4.66393195e-02 7.70627856e-02 -7.62502730e-01 5.69936931e-01
9.72635597e-02 7.78089523e-01 -3.51079762e-01 -1.28004467e+00
-4.45720822e-01 7.18602002e-01 3.78371537e-01 -5.90189755e-01
1.15743160e+00 -3.07772160e-01 -7.11291373e-01 1.02796757e+00
1.40972602e+00 -3.27639192e-01 -1.95411697e-01 -3.39809537e-01
3.25905859e-01 -9.01452303e-02 -3.74762006e-02 -6.77765012e-01
-4.65409040e-01 9.42243695e-01 8.37430239e-01 8.17739189e-01
4.87034231e-01 5.22571743e-01 1.23754613e-01 3.87486190e-01
3.21968913e-01 -1.68067086e+00 1.24897525e-01 9.06226277e-01
8.37319016e-01 -5.11264443e-01 -2.76271850e-01 -5.09012461e-01
-5.09502590e-01 1.25390625e+00 2.18257666e-01 -3.07553053e-01
1.42071152e+00 -3.19787651e-01 -2.13627383e-01 -5.84168732e-01
-6.99780658e-02 -3.79208535e-01 5.79341650e-01 8.63219023e-01
7.54983246e-01 4.58628386e-01 -9.87158835e-01 4.68709320e-01
-9.94373441e-01 -5.41024208e-01 2.91235059e-01 6.82105064e-01
-7.20934272e-01 -7.65715241e-01 -2.79338181e-01 2.35809982e-01
-1.08070811e-03 -3.01266700e-01 -5.50272763e-01 8.51675034e-01
4.42991078e-01 6.85485363e-01 5.35523951e-01 -3.57980877e-01
1.59112737e-02 5.09641230e-01 5.80557227e-01 -4.50155348e-01
-6.68947026e-02 -5.03186941e-01 -1.61103100e-01 -2.53053278e-01
-6.30133867e-01 -6.93157613e-01 -1.49270618e+00 -3.61857235e-01
-8.71727690e-02 5.96677780e-01 9.71676826e-01 9.72365201e-01
2.05361508e-02 4.16698992e-01 2.04696178e-01 -7.49088645e-01
-2.23814934e-01 -1.03856373e+00 -1.27524936e+00 4.23987687e-01
-4.19416934e-01 -1.08723962e+00 -5.52054286e-01 -1.17938980e-01] | [10.370676040649414, 8.964204788208008] |
c7bdda06-9bc8-4394-a359-c67282492d5f | enhancing-face-recognition-with-self | null | null | http://openaccess.thecvf.com//content/CVPR2022/html/He_Enhancing_Face_Recognition_With_Self-Supervised_3D_Reconstruction_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/He_Enhancing_Face_Recognition_With_Self-Supervised_3D_Reconstruction_CVPR_2022_paper.pdf | Enhancing Face Recognition With Self-Supervised 3D Reconstruction | Attributed to both the development of deep networks and abundant data, automatic face recognition (FR) has quickly reached human-level capacity in the past few years. However, the FR problem is not perfectly solved in case of uncontrolled illumination and pose. In this paper, we propose to enhance face recognition with a bypass of self-supervised 3D reconstruction, which enforces the neural backbone to focus on the identity-related depth and albedo information while neglects the identity-irrelevant pose and illumination information. Specifically, inspired by the physical model of image formation, we improve the backbone FR network by introducing a 3D face reconstruction loss with two auxiliary networks. The first one estimates the pose and illumination from the input face image while the second one decodes the canonical depth and albedo from the intermediate feature of the FR backbone network. The whole network is trained in end-to-end manner with both classic face identification loss and the loss of 3D face reconstruction with the physical parameters. In this way, the self-supervised reconstruction acts as a regularization that enables the recognition network to understand faces in 3D view, and the learnt features are forced to encode more information of canonical facial depth and albedo, which is more intrinsic and beneficial to face recognition. Extensive experimental results on various face recognition benchmarks show that, without any cost of extra annotations and computations, our method outperforms state-of-the-art ones. Moreover, the learnt representations can also well generalize to other face-related downstream tasks such as the facial attribute recognition with limited labeled data. | ['Xilin Chen', 'Shiguang Shan', 'Jie Zhang', 'Mingjie He'] | 2022-01-01 | null | null | null | cvpr-2022-1 | ['3d-face-reconstruction', 'face-identification', 'face-reconstruction'] | ['computer-vision', 'computer-vision', 'computer-vision'] | [ 1.08414799e-01 1.67578638e-01 5.05187025e-04 -7.91352272e-01
-2.38046452e-01 -2.99963981e-01 5.24526656e-01 -5.33141673e-01
-2.27478072e-01 3.97915244e-01 -1.49722457e-01 1.43581539e-01
-6.60682917e-02 -7.55533457e-01 -8.73096466e-01 -1.11868727e+00
2.57178426e-01 2.81713128e-01 -5.39357185e-01 -4.59573828e-02
6.72197668e-03 9.98332202e-01 -1.85176814e+00 -1.59306869e-01
5.22354662e-01 1.70415771e+00 -8.68269429e-02 2.87773609e-02
-7.52842724e-02 7.13984489e-01 -8.40559751e-02 -3.46241891e-01
5.17399311e-01 -2.45218337e-01 -3.58784884e-01 4.10558611e-01
7.36811042e-01 -4.63616848e-01 -6.21354640e-01 1.00000691e+00
6.50740623e-01 -1.30572990e-02 3.78565073e-01 -1.13995695e+00
-6.32593334e-01 -3.23534086e-02 -5.28541982e-01 -3.36304724e-01
2.83285439e-01 2.94792354e-01 5.98754883e-01 -1.14289999e+00
3.68420929e-01 1.52731776e+00 6.36062980e-01 7.78490543e-01
-1.14475965e+00 -8.44001710e-01 2.90995389e-01 1.11912973e-02
-1.67143512e+00 -8.80591869e-01 1.20959222e+00 -2.44196549e-01
4.23934162e-01 2.22930461e-02 3.21143031e-01 1.12297571e+00
-3.50063801e-01 5.51676929e-01 1.05348134e+00 -2.99744129e-01
-3.96633521e-02 1.03878714e-01 -1.96259707e-01 1.00633800e+00
-1.01347454e-01 2.74838358e-01 -5.89740574e-01 1.51701733e-01
9.32000995e-01 3.39381218e-01 -5.06766021e-01 -5.63010573e-01
-7.64907002e-01 5.83222508e-01 5.45137763e-01 -1.19397147e-02
-2.55069703e-01 -6.31873831e-02 5.89683950e-02 3.45201254e-01
7.67361045e-01 -9.33798254e-02 -4.09376651e-01 3.43436360e-01
-7.16544509e-01 -7.93592706e-02 7.14135528e-01 7.04782128e-01
1.22733486e+00 2.06894502e-01 -3.56645440e-03 8.08941662e-01
6.54334486e-01 9.14638162e-01 1.49634123e-01 -8.17338169e-01
2.77852565e-01 7.40973532e-01 -2.16231719e-02 -1.02158725e+00
-3.85117114e-01 -6.01267815e-01 -1.08272362e+00 3.26559216e-01
4.67175484e-01 8.47987682e-02 -8.75973046e-01 2.02040625e+00
5.91052234e-01 1.29910305e-01 -1.18199075e-02 1.11223257e+00
8.40224683e-01 2.44521856e-01 -3.30136806e-01 -3.39247972e-01
1.18982244e+00 -7.05310225e-01 -6.22317553e-01 -1.39750928e-01
2.48299852e-01 -7.43817687e-01 8.31533790e-01 2.19266281e-01
-8.84417832e-01 -7.87379503e-01 -8.62539351e-01 -1.15122505e-01
-1.00805566e-01 3.49729687e-01 5.58648109e-01 7.34775782e-01
-9.56802785e-01 6.18655443e-01 -6.69584394e-01 -2.46200174e-01
7.29460359e-01 6.05874836e-01 -6.96627200e-01 -3.61347198e-01
-9.54190910e-01 6.19785786e-01 1.27258636e-02 6.59910262e-01
-1.10164976e+00 -6.75368071e-01 -9.10930336e-01 3.09930779e-02
4.93472934e-01 -5.53961456e-01 6.36408627e-01 -1.22934592e+00
-1.94002235e+00 1.06784534e+00 -2.33049229e-01 2.48053238e-01
5.94179213e-01 -1.50962189e-01 -2.63256460e-01 2.04326645e-01
-1.40946046e-01 5.33036828e-01 1.29522491e+00 -1.32195699e+00
-9.16009843e-02 -1.00910974e+00 7.81126618e-02 1.28625572e-01
-5.64660430e-01 -1.71107605e-01 -5.79689085e-01 -3.72120112e-01
2.82779515e-01 -7.69408405e-01 2.63023555e-01 5.65791130e-01
-2.38607109e-01 -8.51127729e-02 8.30502272e-01 -5.66646516e-01
4.82958168e-01 -2.39256406e+00 2.91572154e-01 1.26673669e-01
1.28889292e-01 1.43246382e-01 -3.78854841e-01 -5.30973896e-02
-3.56148034e-01 -1.66011587e-01 -1.60893261e-01 -7.31702447e-01
-7.00286180e-02 1.62352115e-01 -2.59611338e-01 8.27126563e-01
5.10064244e-01 7.43568957e-01 -7.22409606e-01 -2.22947672e-01
2.58640498e-01 9.96953785e-01 -5.68157136e-01 5.50969720e-01
-1.90364629e-01 9.19671357e-01 -4.41427618e-01 6.95961535e-01
1.19804358e+00 -8.94897655e-02 1.39498800e-01 -4.74256277e-01
-4.66701724e-02 7.71697983e-02 -1.13864422e+00 1.86958003e+00
-5.70021927e-01 1.53477475e-01 4.57109809e-01 -1.13703883e+00
1.14509821e+00 2.94504404e-01 5.20226955e-01 -7.33051717e-01
2.31560677e-01 1.35930449e-01 -3.62695277e-01 -4.89323318e-01
-3.95034194e-01 -1.39559895e-01 4.07504797e-01 3.52080226e-01
2.23689541e-01 1.16040692e-01 -4.11749035e-01 -2.72411048e-01
6.14459157e-01 4.46188062e-01 -1.38051420e-01 -2.01391235e-01
1.00139952e+00 -7.51366675e-01 6.51796103e-01 1.66131034e-01
1.81042969e-01 5.55789471e-01 4.26622182e-01 -6.99431300e-01
-7.05227315e-01 -9.65499341e-01 -2.48051763e-01 7.82964051e-01
5.86640127e-02 -4.50529233e-02 -7.93568194e-01 -8.36530030e-01
1.36190668e-01 6.47292882e-02 -7.49491155e-01 -2.80117542e-01
-5.86826384e-01 -6.22518003e-01 4.66010839e-01 3.16060543e-01
8.04783762e-01 -7.36550927e-01 -1.00840583e-01 -2.14106962e-01
-5.79110533e-03 -1.27812195e+00 -5.49704194e-01 -9.08851475e-02
-6.13465250e-01 -1.02711296e+00 -6.69357598e-01 -6.49745405e-01
1.00781369e+00 3.77209753e-01 6.92974448e-01 2.76464224e-01
-3.55806261e-01 2.99209058e-01 -1.72129795e-01 -1.04305997e-01
3.53614497e-03 -1.38115168e-01 2.78866142e-01 9.30031359e-01
1.78357080e-01 -7.84117877e-01 -7.74791896e-01 4.54297990e-01
-7.58363843e-01 -1.68527365e-01 5.66597879e-01 9.44563508e-01
5.10264814e-01 -1.47569761e-01 4.25848007e-01 -6.11816108e-01
-2.32442990e-01 -2.26380751e-01 -7.10854471e-01 1.67301700e-01
-5.01820087e-01 1.53761923e-01 7.77596235e-01 -3.89177173e-01
-1.22205019e+00 4.53042030e-01 -3.18494767e-01 -7.51850307e-01
-2.63169438e-01 6.22388572e-02 -9.13289607e-01 -5.74511886e-01
3.12004179e-01 3.80238563e-01 4.89416301e-01 -6.95947349e-01
1.99329540e-01 4.79847848e-01 3.82742077e-01 -5.47318637e-01
1.14786744e+00 8.36101055e-01 4.00250912e-01 -9.75194335e-01
-1.18665612e+00 -9.60754305e-02 -7.50484407e-01 -2.41413832e-01
6.12809062e-01 -1.05892801e+00 -1.22631764e+00 8.39337349e-01
-1.17043734e+00 -5.40795587e-02 -1.55877963e-01 3.25876415e-01
-3.58204007e-01 4.73414242e-01 -4.26042646e-01 -9.03855503e-01
-2.61127979e-01 -1.18794668e+00 1.33696711e+00 2.35144168e-01
6.41246259e-01 -7.10509241e-01 -3.85924995e-01 4.39089954e-01
2.88375705e-01 1.56601101e-01 8.21211934e-01 -1.16313949e-01
-8.01856875e-01 -1.61568522e-01 -5.21900296e-01 6.94938660e-01
3.03287268e-01 -3.73187840e-01 -1.58255649e+00 -4.80395228e-01
4.15939003e-01 -5.08561492e-01 9.68606353e-01 6.56409794e-03
1.33971643e+00 -3.59174162e-01 2.80302744e-02 1.22105324e+00
1.28243911e+00 -2.13579401e-01 5.71763694e-01 -2.60667205e-01
1.03033006e+00 1.00209796e+00 2.99728990e-01 4.57034439e-01
3.63789439e-01 7.05314875e-01 8.43672693e-01 -2.18075410e-01
-1.43277943e-01 -3.22639525e-01 4.85201240e-01 5.15802562e-01
-2.14612305e-01 1.26028489e-02 -4.90941405e-01 -1.22854628e-01
-1.68974686e+00 -7.13964045e-01 3.68698984e-01 2.40643954e+00
6.45436585e-01 -3.65096658e-01 -1.76492333e-01 6.23508506e-02
5.75382948e-01 2.75722772e-01 -8.89535010e-01 2.81333089e-01
-4.15236413e-01 1.90083697e-01 1.85646564e-01 4.52048808e-01
-9.56474781e-01 9.47201788e-01 4.73123455e+00 7.17304647e-01
-1.49769187e+00 5.38417771e-02 8.53479028e-01 -2.01358031e-02
-1.59467816e-01 -9.14367065e-02 -9.77147579e-01 1.94413587e-01
4.85588819e-01 4.70395952e-01 7.14918256e-01 7.05241323e-01
1.10242084e-01 3.57445538e-01 -1.42694461e+00 1.24746728e+00
2.86038607e-01 -9.42775011e-01 1.46608755e-01 3.05684298e-01
3.33937526e-01 -2.59189427e-01 1.42546609e-01 1.16654485e-01
-4.10273671e-01 -1.23003387e+00 7.19747901e-01 7.36463845e-01
1.16304016e+00 -6.60211444e-01 5.92782021e-01 2.59526134e-01
-1.16300404e+00 -1.45939901e-01 -4.52619463e-01 -8.62819403e-02
-1.71305880e-01 6.94051564e-01 -3.03725332e-01 6.50365055e-01
5.73570430e-01 8.82960320e-01 -3.94864947e-01 4.26280081e-01
-3.36544812e-01 1.75128073e-01 -4.00893271e-01 4.47291940e-01
-4.44524623e-02 -3.80861104e-01 4.00674522e-01 5.95841885e-01
2.79989541e-01 4.67195250e-02 1.79434419e-01 1.00599432e+00
-4.16550606e-01 3.75074893e-02 -6.16659105e-01 1.60391435e-01
2.97091782e-01 1.46687829e+00 -3.11205655e-01 2.22789437e-01
-2.80516565e-01 1.09034610e+00 4.82000440e-01 4.29604292e-01
-6.43975258e-01 2.42039878e-02 8.62618864e-01 2.60921270e-01
2.86643654e-01 -8.37089568e-02 1.21821351e-01 -1.25724149e+00
3.67378473e-01 -6.88312709e-01 -5.56490384e-02 -4.25928563e-01
-1.36078727e+00 6.84769094e-01 -4.21930611e-01 -9.52544272e-01
-2.17055604e-02 -9.59496081e-01 -4.70777571e-01 8.84464204e-01
-2.10855699e+00 -1.41501701e+00 -6.18484437e-01 8.67352664e-01
1.06213547e-01 -2.01506391e-01 8.56991708e-01 6.00498676e-01
-8.54779720e-01 9.27168190e-01 -1.41370133e-01 3.37124109e-01
8.61633718e-01 -8.17282856e-01 2.48741824e-02 4.45598602e-01
-1.34032533e-01 5.30628264e-01 1.79136455e-01 -2.05372810e-01
-2.03211141e+00 -1.13948274e+00 6.06987715e-01 -2.16819763e-01
3.15587461e-01 -7.29585409e-01 -7.79174328e-01 4.08966839e-01
-3.05733055e-01 5.89115202e-01 3.99379283e-01 -4.78886627e-02
-8.44313562e-01 -9.26515222e-01 -1.23193109e+00 2.03934208e-01
1.38220119e+00 -8.19828689e-01 4.19344231e-02 5.17212570e-01
6.05134249e-01 -1.69865787e-01 -8.22350681e-01 4.97437298e-01
6.41278684e-01 -9.39920664e-01 1.12249255e+00 -3.63385409e-01
1.15834482e-01 -2.64021695e-01 -3.34818423e-01 -1.00205147e+00
-1.47704542e-01 -6.66419268e-01 2.47809775e-02 1.52718782e+00
1.45120788e-02 -7.58894980e-01 8.17865670e-01 3.82423222e-01
-1.37577055e-03 -7.27435708e-01 -1.14678776e+00 -7.27475226e-01
-4.83089834e-02 -1.47869587e-01 7.03170478e-01 8.56161058e-01
-6.75168455e-01 3.84745389e-01 -4.86356109e-01 3.88024181e-01
1.03012574e+00 3.42668325e-01 7.24578440e-01 -1.57104683e+00
-4.45325226e-02 -2.50408560e-01 -3.38697702e-01 -1.25553429e+00
7.34591782e-01 -8.74254882e-01 -2.14684196e-02 -6.49606109e-01
1.29702583e-01 -4.76916015e-01 -1.19732112e-01 5.37778914e-01
6.84224591e-02 3.88921916e-01 1.85941696e-01 1.97505191e-01
-2.69505471e-01 1.06180584e+00 1.30038750e+00 -1.69464022e-01
1.36404172e-01 -1.22623466e-01 -5.11499405e-01 7.53225446e-01
4.33569252e-01 -1.87259659e-01 -2.59327769e-01 -6.94027722e-01
-4.53224257e-02 -4.04047519e-02 6.69976652e-01 -7.68584907e-01
1.92175493e-01 -8.04798678e-02 7.24402845e-01 -1.60274938e-01
6.18019700e-01 -1.00486326e+00 -4.23725843e-02 2.25500196e-01
-5.81698231e-02 -4.78726149e-01 -8.27095658e-02 5.27449846e-01
-1.26723185e-01 8.88781250e-02 1.13346291e+00 -8.35721418e-02
-3.80839020e-01 8.49580705e-01 3.01911384e-01 -2.28805780e-01
7.58920014e-01 -1.70369044e-01 -9.10909474e-02 -2.38436162e-01
-5.15673816e-01 7.07259551e-02 5.59696138e-01 3.40255022e-01
6.51242733e-01 -1.42668939e+00 -7.24183321e-01 7.50817180e-01
9.68760550e-02 2.61257589e-01 4.56721216e-01 7.81686902e-01
-2.48766601e-01 2.51705408e-01 -1.89580351e-01 -7.25652397e-01
-9.54144001e-01 6.37202382e-01 5.78924000e-01 1.95965171e-02
-4.92405981e-01 8.09171736e-01 5.77181160e-01 -7.09755063e-01
5.16987920e-01 2.47337312e-01 -4.12450805e-02 9.20779184e-02
6.68067276e-01 5.45612015e-02 8.24311301e-02 -9.49986815e-01
-5.53899586e-01 1.04403365e+00 5.24491332e-02 3.31556022e-01
1.49198747e+00 -2.76902884e-01 -3.72996300e-01 1.49765655e-01
1.58347976e+00 -2.49136016e-01 -1.59666860e+00 -3.92466247e-01
-3.68376613e-01 -5.05254626e-01 2.57333159e-01 -4.63549793e-01
-1.73599851e+00 1.24250150e+00 7.26819158e-01 -4.03822631e-01
1.37158906e+00 -1.26825228e-01 5.18068612e-01 4.60132331e-01
4.23718452e-01 -6.99823797e-01 1.80402279e-01 5.79605639e-01
1.01278710e+00 -1.40167546e+00 -2.19023764e-01 -5.81513524e-01
-2.62109358e-02 1.27829349e+00 6.96773648e-01 8.45968723e-03
8.78543317e-01 -3.25287059e-02 -3.40249427e-02 -8.12946558e-02
-3.62088412e-01 -1.45402357e-01 2.69539595e-01 4.68243420e-01
2.21059307e-01 -2.43324161e-01 3.46325487e-01 5.27864516e-01
1.11012682e-01 -7.18288422e-02 -1.88387603e-01 3.91935080e-01
-1.08195119e-01 -1.01251733e+00 -2.68481374e-01 1.15790032e-02
-2.27925375e-01 7.82041773e-02 -2.96850592e-01 5.70868969e-01
4.06984061e-01 7.00914502e-01 9.26628113e-02 -3.14860970e-01
2.95246094e-01 9.01833624e-02 7.15544105e-01 -5.02780795e-01
-2.48205110e-01 1.52345151e-01 -3.89132112e-01 -7.69924283e-01
-4.93494034e-01 -4.47272211e-01 -1.11535466e+00 -3.46692353e-01
-3.28465968e-01 -5.11256866e-02 6.95837140e-01 9.40809965e-01
4.63177383e-01 1.73315987e-01 1.32996190e+00 -1.25236630e+00
-6.84621751e-01 -9.07513916e-01 -6.72794521e-01 5.02182841e-01
6.77638531e-01 -8.67734849e-01 -6.07753754e-01 -1.48405448e-01] | [13.199358940124512, 0.42939144372940063] |
34aa37d8-eb10-4e8b-bd11-08e9f322c88c | eliciting-knowledge-from-large-pre-trained | 2211.01587 | null | https://arxiv.org/abs/2211.01587v2 | https://arxiv.org/pdf/2211.01587v2.pdf | Eliciting Knowledge from Large Pre-Trained Models for Unsupervised Knowledge-Grounded Conversation | Recent advances in large-scale pre-training provide large models with the potential to learn knowledge from the raw text. It is thus natural to ask whether it is possible to leverage these large models as knowledge bases for downstream tasks. In this work, we answer the aforementioned question in unsupervised knowledge-grounded conversation. We explore various methods that best elicit knowledge from large models. Our human study indicates that, though hallucinations exist, large models post the unique advantage of being able to output common sense and summarize facts that cannot be directly retrieved from the search engine. To better exploit such generated knowledge in dialogue generation, we treat the generated knowledge as a noisy knowledge source and propose the posterior-based reweighing as well as the noisy training strategy. Empirical results on two benchmarks show advantages over the state-of-the-art methods. | ['LiWei Wang', 'Michael R. Lyu', 'Jianqiao Zhao', 'Yanyang Li'] | 2022-11-03 | null | null | null | null | ['common-sense-reasoning'] | ['reasoning'] | [ 0.12401599 1.0696228 -0.31426755 -0.02625161 -1.1972302 -0.62504375
0.96481156 0.05355621 -0.22130537 1.4139161 0.92314297 -0.22910546
0.04435662 -0.810803 -0.60695755 -0.41430327 0.33736685 0.6906473
0.16510262 -0.56291443 0.31405312 -0.04192489 -1.1532507 0.7088551
1.0745888 0.64839184 0.35466322 0.4434344 -0.43314454 1.093658
-1.0210814 -0.792558 -0.16332622 -0.61174697 -1.4592326 -0.02916657
-0.07838064 -0.25011155 -0.13423628 0.8450162 0.63188183 0.3977877
0.6258082 -0.83823484 -0.81278443 1.4289435 -0.03167785 0.25228298
0.748513 0.06512157 1.1059141 -0.9538283 0.75484633 1.2858753
0.51636946 0.77619267 -1.1265544 -0.23654866 0.07657115 0.3069885
-1.238713 -0.6318549 0.7282193 -0.14142774 1.2962897 0.18774478
0.5201035 1.7053654 -0.2627113 1.1389438 1.0718517 -0.6714527
0.07421445 0.51881593 0.08282391 0.6388403 0.20817287 -0.01177182
-1.0865328 -0.5472633 0.4281577 -0.57844543 -0.6496156 0.34297743
-1.3383672 0.8770594 0.24224159 0.16331178 -0.56252337 -0.17870168
0.04548699 0.18639964 0.6008308 1.1795179 -0.6691159 -0.512206
-0.8117798 0.37788776 1.5203327 1.0377735 0.710871 -0.14050703
-0.6563063 0.8417185 0.12362264 0.3084065 0.8164978 -1.0112805
0.6815678 0.37942606 0.38629434 -0.67564285 -0.08575338 -0.4321458
-0.6862304 -0.538096 0.2743498 -0.64384854 -0.56406176 1.6518189
0.17003101 0.4407047 0.78257996 0.6112316 1.0767351 0.70452905
-0.03243607 -0.41533208 1.26577 -1.2237018 -0.8446866 -0.33390912
0.57962847 -0.703606 1.0391283 0.3790802 -1.1478351 -0.36454597
-0.7933663 0.01616847 -0.37041408 0.02730172 0.69691205 0.32964447
-0.9099481 0.57220507 -0.43597192 -0.29315582 0.16090061 -0.0661471
0.05159931 0.10800645 -1.7750018 1.155577 0.6734565 0.19718051
-0.9611311 -0.50067914 -0.7024764 0.20375258 0.84273404 -1.1075633
1.7074386 -0.71425873 -1.9178807 0.3874482 -0.40019125 -0.657546
0.24983184 -0.4569198 -0.09877162 0.09562124 0.10818003 0.50448763
0.5912306 -1.5981451 -0.52934736 0.15476897 0.3946787 0.43218884
-0.3209918 -0.24050122 -0.23294616 -0.4324393 -0.27800867 -0.8955603
-0.23106426 -0.95364505 -1.0336154 -0.47158384 0.25667948 -0.47114974
1.3639944 -1.4426534 0.1201077 0.04114756 0.16166092 0.4249777
-0.08693028 0.85059416 0.40587255 0.39918143 0.02675861 -0.26394954
-0.08351775 0.2833789 -0.83337986 -0.58336514 0.3694461 1.1889458
-1.2964389 -0.47509483 -0.27631986 0.08704937 -0.37394142 0.47296992
-0.705302 0.43481025 -0.8041107 0.29888394 0.07078622 -0.6069231
0.2290249 -0.00638771 0.26842377 0.79534334 -0.8603672 1.6809167
-0.47475535 0.40162146 -0.25138766 -0.8140734 0.6736405 0.75209445
-0.1520867 -0.10283928 -0.06532187 0.03573865 -0.06053254 -0.7278266
0.9219002 -0.4773036 0.05515395 0.5153069 0.36517558 -0.28163093
0.1467715 0.59272265 1.0724344 -0.04763905 0.37868378 0.04058556
0.35388395 0.28263402 0.3196099 1.144064 0.32533976 0.24838029
0.53249097 0.18778266 -0.46244913 -0.96940494 0.32256156 1.0219717
-0.05805907 -0.7665102 -0.693176 -0.7697557 -0.19125149 1.1768912
-0.4352723 -0.23737416 -0.3270153 -0.5935866 0.875758 0.5490009
0.42771456 -1.2460912 -0.125374 0.31023213 -0.8436511 -1.2185755
-0.05066416 0.12519361 -0.7319149 -0.8608031 -0.642146 -0.46479002
0.4100473 0.16818687 1.5197337 -0.07209051 0.34051743 0.5112003
-0.64488363 -0.5932156 -0.7012237 0.43976593 -0.03130708 -0.14376718
0.34478724 -0.53648424 -0.25199056 -0.06905037 -0.59957546 0.27493474
0.7822938 1.1327262 0.34329826 -0.08945858 1.2194839 -1.1771724
1.4647751 -0.8567296 -0.04288336 0.51138765 -0.6168016 0.42114687
0.6392518 -0.36018756 -1.678268 -0.37734345 -0.07187077 -0.10789245
-0.13557786 1.1551993 0.01833315 0.53172463 0.96923023 0.38940257
-0.38237146 -0.515828 0.98210037 0.72300434 0.32673118 -0.9853425
0.5531382 0.1598668 -0.49206972 -0.94554466 -1.6869891 -0.44405538
-0.37507018 0.11377693 0.75864637 -0.8818394 -0.33404353 -0.07578254
-1.555252 -0.3252219 -0.48427805 0.24505821 -0.5588683 0.45936805
-0.5068786 -0.89755714 -0.59320456 -0.6419896 0.9311792 0.40906122
-0.5529742 -1.0674794 0.20865695 0.72277075 0.33683988 -0.23012753
0.8806189 -1.3018771 -0.66725725 0.1716346 0.07744121 0.29056415
0.1677252 -0.3719709 -1.2353009 0.18705373 0.04198804 -0.8693911
1.0620141 -0.0282202 0.8153875 -0.79110444 -0.3245109 -0.02116828
0.9630432 -0.2779396 0.5078576 -0.02279796 0.45422837 0.7553361
0.44710335 0.4252398 0.5307765 0.41708153 -0.19283208 0.29183045
-0.08268346 -0.7610426 0.3154889 0.9475631 -0.25407442 -0.42726144
-0.91216016 0.78854966 -2.0288792 -1.3827754 0.22983615 1.7957479
1.7430615 0.15199882 -0.2281736 -0.3439763 0.49221623 0.18622084
-0.24061461 -0.19092265 -0.14499216 0.22054766 -0.1204939 0.7429743
-0.60396904 1.1683586 6.2667527 1.0792353 -0.5223797 0.13405943
0.45365992 0.04378391 -0.5487585 0.21265252 -1.1202004 0.14995879
1.2353889 -0.63408583 0.26377285 0.75472605 0.21490902 -0.27871695
-1.249989 0.63701504 0.17209442 -1.56992 0.43752304 -0.14450094
1.1633224 -0.01012186 -0.3810354 0.70590025 0.6763758 -1.1303376
0.36718714 0.8912124 0.24140067 -0.4596176 0.8999946 0.92722625
-0.60915005 0.02009681 -0.29785472 -0.16695789 0.49928576 0.5892069
-1.513387 0.7075018 0.04778564 0.32233953 -0.48686028 0.6741419
-0.6688786 0.9386257 -0.19829915 -0.31307745 0.29260266 0.20084007
0.53909045 1.2491496 0.17282993 0.37501714 0.3136085 1.1342407
-0.36597857 0.02245252 -0.7451329 -0.4318359 0.64546376 1.1356778
-0.20887747 -0.75191075 -0.4122716 0.94960284 0.55289143 0.47535628
-0.2735876 -0.22711143 0.22485432 -0.24528092 0.12151774 -0.02719089
-0.1154104 -1.404898 -0.07971876 -1.0440974 0.27422926 -0.89712375
-1.5720074 0.6965841 0.13381682 -0.6778013 -0.8706579 -0.20732349
-0.60492235 0.9171001 -1.6917597 -1.1463636 -0.07268151 0.49930578
0.81404847 -0.22382039 1.1600643 -0.44827697 -0.15701452 0.2909515
-0.04269382 0.15730688 0.7148531 -1.41057 0.06037959 0.55484784
0.49437708 1.0125719 0.6875607 -0.85842896 -1.1481348 -0.84877664
1.1840024 -1.000954 0.75113344 -0.06387746 -1.0272336 0.6850618
0.6291111 -0.59664875 0.94583094 0.5767784 -0.19895309 0.30861974
-0.7820328 0.78706 0.83986384 -0.56950057 -1.3521703 0.39361006
0.9792047 -0.33165443 -0.6638118 0.06180591 0.14536072 -0.7183433
0.8440231 -0.9609142 0.75899476 0.14257263 0.04781249 -1.8063825
0.20633364 -1.1977717 -0.56394863 1.4618505 1.0698452 -0.40464765
0.5929037 0.7890074 -0.18084587 -0.8071078 -0.71760684 -0.6447502
0.20113839 -0.3086745 0.42579317 0.9324207 0.7744183 1.1824287
-0.41244465 -0.01574186 0.24432412 0.24736375 0.7064389 -1.1630255
-0.46151194 -0.11811838 0.33598658 -1.4420599 0.5385256 -0.859663
0.28348908 -1.7198938 0.31482148 -0.17387787 -0.16194381 0.5479517
-0.5105788 -0.25588188 0.0470015 0.01686657 -0.78569645 0.8634318
1.4092813 0.03846347 -0.3656337 0.0626123 -1.257892 0.79803044
0.7802861 -0.38337252 -0.8137694 -0.3337298 0.55794805 0.3293606
0.29165125 -0.6123127 0.4479725 -0.29922828 -0.065345 -0.36743444
0.63905185 -0.26354277 -0.435961 -0.09994863 -0.8048363 -0.3828577
0.0663143 0.8349901 -0.4442114 -0.52582884 0.16797131 -0.5762929
-0.51533484 -0.23664628 -0.5708574 0.69226414 0.3184246 0.20016892
-0.64795506 -0.9276807 -0.68707573 0.23301436 -0.18034407 0.43993217
0.68462753 -1.0828182 -0.75232244 -0.21489073 0.15211357 -0.04863838
-0.06248636 0.67685634 0.33090475 0.7744953 0.392917 -0.05652754
-0.876948 0.39873362 0.08609045 -0.60910326 -0.37316927 0.98043346
-0.05331405 -0.4081812 0.11821601 -0.27123022 -0.38738886 0.11061788
0.619369 0.25053197 -0.13132018 -0.25400004 0.12366873 -0.06872118
-0.05671683 -0.5587592 1.0622674 -0.29818347 -0.02370965 0.53882563
0.5644085 0.2516693 -0.80933934 -0.6403869 0.28232482 -0.1749907
-0.08935116 -1.2639424 -0.35823753 0.80042857 -0.25322783 0.56777537
0.6193166 0.28332567 0.87079716 1.2050891 0.37288976 -1.2311118
0.3736214 1.0549772 1.1056076 -1.2688817 -0.21213348 -0.4086804
-1.2084115 0.91146183 0.74943006 0.3585984 0.46432778 0.03409794
0.14832523 -0.29791924 -1.3547281 -0.5481453 0.290468 0.5992073
0.50236446 -0.12821238 -0.16470301 1.1795363 -0.5934668 0.07101502
0.6257342 0.7409941 -0.6502158 -1.0656722 -0.07751006 0.51909995
-0.41716397 -0.5667278 -0.94122386 0.3528292 -0.21877083 1.4278483
-0.6869293 -0.20911588 0.26180342 0.71420306 0.25966173 -1.2523997
-0.6408555 -0.18936232 0.87504315 -0.33066064 -0.47719717 -0.12876275
-1.1034482 0.10483609 -0.70431554 0.67737854 0.15193836 1.281626
0.57401925 0.3218718 0.22988074 -0.65495694 -1.0570126 -1.4495236
-0.28526682 0.278888 0.17311166 -0.47653887 -0.36964402 0.22769128] | [12.158616065979004, 8.205634117126465] |
592ed164-44cb-4a67-ba4a-98add45a36f7 | deep-learning-for-energy-time-series-analysis | 2306.09129 | null | https://arxiv.org/abs/2306.09129v2 | https://arxiv.org/pdf/2306.09129v2.pdf | Deep Learning for Energy Time-Series Analysis and Forecasting | Energy time-series analysis describes the process of analyzing past energy observations and possibly external factors so as to predict the future. Different tasks are involved in the general field of energy time-series analysis and forecasting, with electric load demand forecasting, personalized energy consumption forecasting, as well as renewable energy generation forecasting being among the most common ones. Following the exceptional performance of Deep Learning (DL) in a broad area of vision tasks, DL models have successfully been utilized in time-series forecasting tasks. This paper aims to provide insight into various DL methods geared towards improving the performance in energy time-series forecasting tasks, with special emphasis in Greek Energy Market, and equip the reader with the necessary knowledge to apply these methods in practice. | ['Anastasios Tefas', 'Nikos Nikolaidis', 'Pavlos Tosidis', 'Theodoros Manousis', 'Efstratios Kakaletsis', 'Paraskevi Nousi', 'Charalampos Symeonidis', 'Maria Tzelepi'] | 2023-06-15 | null | null | null | null | ['time-series'] | ['time-series'] | [-2.23796055e-01 -5.01102269e-01 -1.65311322e-01 -1.38102859e-01
-1.51298359e-01 -5.76908290e-01 8.74754727e-01 2.53369153e-01
-8.22875947e-02 4.98072803e-01 1.43570319e-01 -3.91978890e-01
-3.56832117e-01 -9.65137243e-01 -2.70948291e-01 -9.56697047e-01
-2.31808186e-01 1.44477487e-01 -6.49277985e-01 -2.86596924e-01
1.14874668e-01 8.39139640e-01 -1.50483835e+00 7.88400229e-03
5.88444114e-01 1.08661127e+00 2.95201808e-01 5.68608820e-01
-3.28611702e-01 6.72064781e-01 -5.00892639e-01 -1.89600050e-01
-1.77092984e-01 -2.55996376e-01 -4.32953179e-01 -2.63889343e-01
-6.11644983e-01 -6.62804022e-02 -3.26764345e-01 6.35779977e-01
7.01066434e-01 5.17835617e-01 7.19047785e-01 -1.19721484e+00
-5.92046142e-01 5.13620555e-01 -9.42104962e-03 4.56643939e-01
2.68800497e-01 2.88503207e-02 8.35967362e-01 -6.00596130e-01
5.54531701e-02 6.30959570e-01 5.78566015e-01 1.95069760e-01
-7.99682856e-01 -3.83257091e-01 -2.13446423e-01 8.76031399e-01
-8.10463130e-01 -2.24574462e-01 1.55252409e+00 -4.82105523e-01
1.56832540e+00 1.61495313e-01 9.10281956e-01 1.21712577e+00
7.29389548e-01 9.09268320e-01 9.04183388e-01 -4.79818374e-01
5.28697491e-01 -1.97213665e-01 2.94352192e-02 -5.46093434e-02
-1.81009635e-01 4.44456488e-01 -2.75121778e-01 2.69318730e-01
1.89104527e-02 2.24894926e-01 1.59074917e-01 -1.78120825e-02
-9.66998816e-01 8.79750311e-01 4.04496074e-01 1.12201869e+00
-8.03996265e-01 9.77404937e-02 8.32344294e-01 4.57638979e-01
9.04933035e-01 2.13640630e-01 -6.80088818e-01 -3.27975899e-01
-9.22578812e-01 6.88206106e-02 6.75334573e-01 2.55530030e-01
2.54990160e-01 9.52055931e-01 1.74249917e-01 5.81074655e-01
-1.64379075e-01 6.45079434e-01 1.03183687e+00 -5.24026930e-01
1.74522325e-02 4.51358110e-01 2.15318128e-02 -9.42052007e-01
-8.95557642e-01 -5.50938010e-01 -1.32102549e+00 7.92749301e-02
-5.39656822e-03 -3.66091788e-01 -4.78481561e-01 1.25089943e+00
-2.27759823e-01 7.19609484e-02 9.63613093e-02 1.42959595e-01
4.88956600e-01 1.20997036e+00 3.13790262e-01 -6.93897307e-01
8.61977577e-01 -5.34079134e-01 -9.66700315e-01 8.12626928e-02
5.17890036e-01 -4.32683498e-01 3.07059824e-01 3.49940717e-01
-8.51759911e-01 -8.26408744e-01 -7.50022948e-01 1.97542429e-01
-1.14117658e+00 3.63237202e-01 5.82039595e-01 4.47952062e-01
-9.86123919e-01 8.18713069e-01 -9.17143822e-01 -3.94411594e-01
1.14492022e-01 1.91096529e-01 1.65108919e-01 5.55136800e-01
-1.26300919e+00 1.31709504e+00 6.82912350e-01 4.40395683e-01
-4.17337865e-01 -6.75275385e-01 -7.20203757e-01 2.49941602e-01
-6.29265606e-02 -5.88055909e-01 1.03988302e+00 -7.80826151e-01
-1.52756822e+00 3.75199109e-01 -2.11845323e-01 -9.00017738e-01
3.90586078e-01 -5.39617799e-02 -1.13210678e+00 -5.38452029e-01
-3.66033316e-01 -9.38771889e-02 7.42333472e-01 -4.29724723e-01
-6.31155133e-01 -3.46660495e-01 -6.27972186e-01 -1.38450041e-01
-4.77426410e-01 -3.07360321e-01 6.42726958e-01 -6.63393915e-01
-2.64281869e-01 -5.82400203e-01 -1.17195733e-01 -8.11461687e-01
7.07220808e-02 -1.12982428e+00 1.02368081e+00 -9.60086226e-01
1.35872185e+00 -1.93446398e+00 1.46268532e-01 1.56817883e-01
-2.52209038e-01 3.69292080e-01 2.22303674e-01 8.97054613e-01
-5.42240322e-01 -3.15387994e-01 2.42870748e-01 -8.26616660e-02
4.91292030e-01 3.67851675e-01 -5.74826837e-01 3.81952882e-01
5.07336259e-02 1.30979371e+00 -8.97280872e-01 3.70634735e-01
1.08786297e+00 4.77865934e-01 4.52511102e-01 7.16617554e-02
-4.45948511e-01 3.98887962e-01 -4.03264463e-01 3.05998921e-01
7.65945613e-02 -1.70765892e-01 1.39554739e-02 -3.39244843e-01
-5.07603884e-01 8.54207277e-02 -7.79000640e-01 1.13073337e+00
-8.64137828e-01 1.08869421e+00 -6.19957209e-01 -1.68653572e+00
8.71319830e-01 5.72286129e-01 1.17334318e+00 -1.07418835e+00
2.40947887e-01 4.34604794e-01 -1.20233402e-01 -6.52285635e-01
4.24408495e-01 -1.03222363e-01 1.72708675e-01 3.06696236e-01
-4.53975201e-02 5.29786870e-02 3.50977302e-01 -5.78479171e-01
4.06089932e-01 7.20371632e-03 5.41656733e-01 -1.84522316e-01
7.38737226e-01 -1.69854626e-01 8.33214894e-02 -8.63015652e-02
6.97520301e-02 -1.68173403e-01 5.08046597e-02 -1.11337757e+00
-1.25386751e+00 -6.18905246e-01 -1.43540010e-01 1.20807755e+00
-6.21490836e-01 2.51704548e-02 -1.83571979e-01 -3.08688521e-01
5.38391881e-02 1.38458550e+00 -5.29172003e-01 -6.04786202e-02
-6.58413708e-01 -9.05187309e-01 -2.16924533e-01 7.79037774e-01
3.37026179e-01 -1.45259774e+00 -6.46098614e-01 6.91003025e-01
1.08931959e-01 -7.05856681e-01 1.15803890e-01 4.83470440e-01
-6.45976424e-01 -7.48431385e-01 -9.70390320e-01 -7.28434026e-01
1.35852218e-01 -1.65988341e-01 1.36711168e+00 -4.49010044e-01
-1.98136255e-01 4.37830091e-01 -2.80067235e-01 -7.40599930e-01
-6.72387302e-01 1.97380438e-01 9.10028890e-02 -1.60745248e-01
5.90628922e-01 -7.21521735e-01 -4.48157281e-01 -1.45145297e-01
-6.22603714e-01 -1.07910819e-01 1.81538701e-01 6.59282684e-01
3.19448560e-01 8.82591367e-01 1.09293711e+00 -2.24155322e-01
8.57834101e-01 -7.64670849e-01 -7.84955919e-01 2.84887433e-01
-9.79866683e-01 -2.64232785e-01 1.01891959e+00 -2.76556462e-01
-6.76862121e-01 -1.48416951e-01 -4.76920873e-01 -3.83842528e-01
-2.56398559e-01 6.00002408e-01 1.37855187e-01 1.59540251e-01
2.08480671e-01 8.52994144e-01 -3.18335652e-01 -6.13871515e-01
2.04998240e-01 3.79643679e-01 3.51095021e-01 -2.97718477e-02
6.79936945e-01 -2.35803006e-03 3.68330538e-01 -8.71815622e-01
-3.87497097e-01 -3.90741050e-01 -5.61967134e-01 -3.15409392e-01
7.37098813e-01 -6.46909535e-01 -6.88518703e-01 6.70502126e-01
-8.95082772e-01 -2.15537816e-01 -4.32562828e-01 2.63538986e-01
-5.05483091e-01 -1.01305686e-01 -2.86801308e-01 -9.99204874e-01
-5.39905906e-01 -6.81479692e-01 6.61836803e-01 4.03364658e-01
-1.25557557e-01 -1.81371570e+00 1.85921356e-01 -2.65631884e-01
8.09907913e-01 5.39670467e-01 1.09376526e+00 -8.39023232e-01
1.13320902e-01 -4.11115021e-01 2.99181312e-01 6.08553410e-01
3.32241476e-01 -9.04024690e-02 -9.67517912e-01 -3.67153853e-01
1.50974602e-01 2.27343500e-01 5.53804159e-01 7.00826764e-01
1.35806334e+00 -4.86650225e-03 -3.53564531e-01 2.61630386e-01
1.40359879e+00 7.74350286e-01 2.50344217e-01 2.65984684e-01
4.30727750e-01 4.17212069e-01 3.75833288e-02 5.29181898e-01
3.87233585e-01 4.50707734e-01 2.59071141e-01 -1.34085640e-01
2.20941097e-01 -2.30922364e-02 3.37707341e-01 9.53960657e-01
-1.31117567e-01 -6.03720009e-01 -8.04058552e-01 7.41405010e-01
-1.60708702e+00 -1.31039584e+00 -1.55792087e-01 1.55729103e+00
9.94463041e-02 -5.15913963e-03 3.64064097e-01 6.13646388e-01
3.21716279e-01 5.18766522e-01 -7.65461743e-01 -6.30668461e-01
-2.16936350e-01 9.72804874e-02 2.78067738e-01 2.03047972e-02
-1.11052573e+00 2.17833370e-01 6.60622311e+00 7.73758054e-01
-1.31076074e+00 -6.10232800e-02 6.43779159e-01 2.98519470e-02
-3.36942792e-01 -6.66334391e-01 -3.31081510e-01 1.00350761e+00
1.35535479e+00 -7.63669074e-01 7.43647635e-01 7.33198404e-01
6.03473365e-01 1.14752680e-01 -1.27394354e+00 1.18158436e+00
-2.14817479e-01 -1.39686847e+00 -2.59691447e-01 -5.32052293e-03
9.57251430e-01 3.01738203e-01 1.17162786e-01 3.69454235e-01
1.66465547e-02 -8.84817898e-01 3.68409365e-01 7.65249372e-01
1.86974093e-01 -8.68964493e-01 1.00683784e+00 5.00045180e-01
-1.35885918e+00 -5.49484909e-01 -1.34755462e-01 -2.68635124e-01
2.83001900e-01 8.78567636e-01 -5.41571200e-01 9.75693464e-01
6.85677648e-01 1.11355436e+00 -6.06083199e-02 5.83097994e-01
1.54329240e-01 5.85585892e-01 -4.36227500e-01 -3.62684935e-01
1.53477818e-01 -3.45676064e-01 2.56608278e-01 1.08315158e+00
7.11592138e-01 -2.14142347e-04 -7.18400627e-02 5.34611166e-01
1.06491461e-01 4.62297834e-02 -6.04887605e-01 -4.95821595e-01
1.36375368e-01 9.93296802e-01 -4.87656593e-01 -2.83911437e-01
-4.71268117e-01 7.35358536e-01 -2.32282430e-01 4.71978188e-01
-5.25770545e-01 -1.62758604e-01 5.34547806e-01 -1.10176347e-01
2.82713890e-01 -4.47535485e-01 -3.32412302e-01 -8.63808334e-01
-1.79362133e-01 -2.48551875e-01 5.25355399e-01 -8.88038754e-01
-1.59263134e+00 4.27784532e-01 2.51839515e-02 -1.11499882e+00
-1.04093146e+00 -8.72420192e-01 -1.03309357e+00 1.02494311e+00
-1.78131485e+00 -1.02410281e+00 2.23590266e-02 5.73259354e-01
1.01013672e+00 -6.01810932e-01 8.32423091e-01 1.82780445e-01
-6.45757496e-01 -6.96633290e-03 9.79069114e-01 -6.63298219e-02
-2.28606075e-01 -1.36816263e+00 6.57091022e-01 6.66580737e-01
1.12997733e-01 -5.67682981e-02 8.29568744e-01 -2.32672051e-01
-1.50680065e+00 -7.72748291e-01 1.11773551e+00 -1.84849575e-01
8.52294207e-01 2.98839174e-02 -5.95869303e-01 4.90375370e-01
6.01292253e-01 -1.98526144e-01 6.33780420e-01 -2.02074394e-01
3.75416428e-01 -4.05876994e-01 -1.02954733e+00 1.89976618e-01
3.60251278e-01 -4.79288816e-01 -5.19261301e-01 5.49254537e-01
-3.20651606e-02 -1.58166081e-01 -1.04466534e+00 3.58028978e-01
4.40907925e-01 -7.55880296e-01 1.01637411e+00 -3.39133382e-01
1.68173596e-01 1.50744379e-01 -6.28992990e-02 -1.75486207e+00
-5.13983190e-01 -6.90974414e-01 -8.66726816e-01 1.16603827e+00
-1.60728410e-01 -8.92009139e-01 4.61636692e-01 3.61357659e-01
-1.63778812e-01 -6.39709532e-01 -1.11787331e+00 -6.66251063e-01
2.54423946e-01 -5.55606246e-01 9.60076749e-01 9.18958485e-01
-1.46346360e-01 6.38814494e-02 -3.30478489e-01 -6.36814833e-02
2.95200676e-01 4.52756166e-01 1.51989862e-01 -1.31222069e+00
2.01427341e-01 -9.27866042e-01 -1.60304472e-01 -3.93956959e-01
4.35802668e-01 -8.52613270e-01 -3.93286884e-01 -1.66778088e+00
-4.19845372e-01 2.50441343e-01 -7.37893820e-01 3.82499397e-01
3.30376506e-01 -1.28007054e-01 1.52923644e-01 -1.77531037e-02
5.74536249e-02 7.06498742e-01 7.07994938e-01 -2.30685145e-01
-1.04508795e-01 6.48828089e-01 1.85656324e-02 5.95478952e-01
1.06475973e+00 1.73798651e-01 -4.78178144e-01 -2.95104414e-01
3.89515489e-01 1.25565812e-01 3.17079008e-01 -1.00280929e+00
1.59642994e-01 -2.10980475e-01 9.58982408e-01 -1.07987928e+00
-1.89683903e-02 -1.30412209e+00 6.90757096e-01 6.35204554e-01
-9.64458808e-02 6.16160691e-01 2.99235553e-01 3.07841569e-01
-1.09887369e-01 7.81280082e-03 3.35105360e-01 -4.42499444e-02
-1.37476528e+00 2.04221591e-01 -5.24558783e-01 -4.91633415e-01
1.18591130e+00 -1.18211158e-01 -2.00124234e-01 -3.47298652e-01
-1.14808524e+00 3.40021938e-01 -8.55012685e-02 7.55239487e-01
9.12963524e-02 -1.43397856e+00 -5.87198317e-01 4.22277957e-01
-3.29785496e-01 -6.64541245e-01 4.90000010e-01 5.78530729e-01
-2.05709979e-01 9.11343277e-01 -1.94858283e-01 -4.40952420e-01
-6.49818540e-01 7.46018708e-01 6.98131800e-01 -3.57862025e-01
-7.08700597e-01 1.71181515e-01 -4.60256606e-01 1.37668893e-01
1.79187298e-01 -3.46833080e-01 -7.01459408e-01 5.43654978e-01
3.11879843e-01 9.81622040e-01 4.74923432e-01 -4.91350621e-01
-3.49243492e-01 6.45001471e-01 5.67471445e-01 5.41893244e-01
1.70022309e+00 -2.49300808e-01 -2.75964215e-02 8.65365863e-01
1.32804012e+00 -5.11672318e-01 -9.22879994e-01 2.60655545e-02
2.96768337e-01 1.75821856e-01 6.02406979e-01 -1.04719126e+00
-1.16855943e+00 8.83965671e-01 9.82348084e-01 1.15764606e+00
1.55286956e+00 -2.86345959e-01 1.00749540e+00 1.78511009e-01
4.32415605e-01 -1.22747409e+00 -3.64382118e-01 4.24556464e-01
7.56621122e-01 -1.21635735e+00 -1.61801696e-01 5.20830214e-01
-1.27516210e-01 1.54543519e+00 -1.49175927e-01 -4.77942079e-02
9.28315461e-01 2.13165954e-01 -2.65172005e-01 8.59133154e-02
-6.60727799e-01 -2.62659580e-01 5.74272633e-01 5.32245815e-01
4.93458331e-01 2.64657229e-01 -2.52821296e-01 4.52069521e-01
-2.05827653e-01 3.39882880e-01 -6.13835938e-02 4.88685071e-01
-8.53788331e-02 -8.77824128e-01 -1.08153634e-01 4.75344628e-01
-4.29775208e-01 2.77651131e-01 1.05188027e-01 4.82030332e-01
1.68457299e-01 9.80412960e-01 3.85885566e-01 -2.11108357e-01
3.25831205e-01 4.58074242e-01 1.56929910e-01 8.44719633e-02
-6.28330469e-01 -6.41402155e-02 -2.11496264e-01 -3.29261988e-01
-3.89468580e-01 -6.75704479e-01 -8.97469223e-01 -6.24364018e-01
-1.80837456e-02 4.43076640e-02 1.00198555e+00 1.27471459e+00
1.45368883e-02 1.02784932e+00 1.03914797e+00 -1.13598669e+00
-5.24319410e-01 -9.80791926e-01 -6.03082120e-01 2.73942173e-01
4.13464278e-01 -2.56738633e-01 -3.93657058e-01 7.15395622e-03] | [6.161076545715332, 2.77988862991333] |
27ad5a6a-d0a3-45cb-96d2-4f92263b108c | aesthetic-driven-image-enhancement-by | 1707.05251 | null | http://arxiv.org/abs/1707.05251v2 | http://arxiv.org/pdf/1707.05251v2.pdf | Aesthetic-Driven Image Enhancement by Adversarial Learning | We introduce EnhanceGAN, an adversarial learning based model that performs
automatic image enhancement. Traditional image enhancement frameworks typically
involve training models in a fully-supervised manner, which require expensive
annotations in the form of aligned image pairs. In contrast to these
approaches, our proposed EnhanceGAN only requires weak supervision (binary
labels on image aesthetic quality) and is able to learn enhancement operators
for the task of aesthetic-based image enhancement. In particular, we show the
effectiveness of a piecewise color enhancement module trained with weak
supervision, and extend the proposed EnhanceGAN framework to learning a deep
filtering-based aesthetic enhancer. The full differentiability of our image
enhancement operators enables the training of EnhanceGAN in an end-to-end
manner. We further demonstrate the capability of EnhanceGAN in learning
aesthetic-based image cropping without any groundtruth cropping pairs. Our
weakly-supervised EnhanceGAN reports competitive quantitative results on
aesthetic-based color enhancement as well as automatic image cropping, and a
user study confirms that our image enhancement results are on par with or even
preferred over professional enhancement. | ['Yubin Deng', 'Chen Change Loy', 'Xiaoou Tang'] | 2017-07-17 | null | null | null | null | ['image-cropping'] | ['computer-vision'] | [ 6.34045184e-01 2.62705266e-01 2.26357251e-01 -5.09309351e-01
-8.65365267e-01 -5.38009048e-01 4.58472133e-01 -1.58625841e-01
-4.94715899e-01 3.12601835e-01 -4.48026843e-02 -2.03775242e-01
1.49727449e-01 -8.00389349e-01 -1.06638730e+00 -5.56903780e-01
4.91621383e-02 -2.66924530e-01 -4.00268197e-01 -6.76670969e-01
-6.92818239e-02 1.41843148e-02 -1.50774682e+00 3.73600215e-01
1.18291080e+00 1.05811858e+00 -4.85860072e-02 9.59187210e-01
3.82249266e-01 7.66951859e-01 -4.29554015e-01 -8.52921486e-01
6.77535176e-01 -4.64116395e-01 -6.65297151e-01 5.13327420e-01
7.16760397e-01 -6.12512350e-01 -2.17247978e-01 1.14770138e+00
6.98216259e-01 -1.41004115e-01 3.32517534e-01 -1.45215666e+00
-1.25659311e+00 5.40237129e-01 -7.08184958e-01 -4.48059946e-01
3.35262418e-02 3.99028182e-01 1.06948864e+00 -9.43201482e-01
4.68211710e-01 9.38949168e-01 8.98299575e-01 4.74888682e-01
-1.15421867e+00 -4.83507454e-01 -1.00216016e-01 -6.32877275e-02
-1.08839428e+00 -2.82218754e-01 1.09179592e+00 4.99490313e-02
3.80921841e-01 2.35704094e-01 6.85660601e-01 7.95254707e-01
-8.22975859e-02 1.10003865e+00 1.66855943e+00 -5.37435532e-01
8.49378332e-02 2.66443342e-01 -4.75021154e-01 7.54767179e-01
-2.46176392e-01 5.10421276e-01 -5.48202217e-01 5.14453769e-01
7.58175850e-01 -5.54010458e-03 -1.38565525e-01 -3.68552238e-01
-9.54437375e-01 5.55193186e-01 9.77432668e-01 2.20423434e-02
-4.00540054e-01 4.18152362e-01 3.05187255e-01 5.75646579e-01
5.52468717e-01 6.65301323e-01 -3.92156869e-01 1.18058585e-01
-1.15784264e+00 -6.40157387e-02 3.00299317e-01 9.99805212e-01
1.02930582e+00 2.62855321e-01 -1.08165160e-01 6.80647790e-01
-8.22672173e-02 6.58442557e-01 -6.32670596e-02 -1.17181122e+00
1.91132799e-01 3.71663302e-01 1.53302133e-01 -8.25922549e-01
-1.73135549e-02 -5.91790795e-01 -8.71200562e-01 9.69948769e-01
1.26710549e-01 -3.02319884e-01 -1.00222540e+00 1.81100106e+00
3.86581346e-02 -1.75785720e-01 2.57917315e-01 1.02339005e+00
6.25442505e-01 4.29024458e-01 4.10576463e-01 1.51504263e-01
1.17641509e+00 -1.45296669e+00 -7.81902075e-01 -1.36147112e-01
3.71072441e-01 -9.35722709e-01 1.54377007e+00 5.99400878e-01
-1.55101156e+00 -7.76993394e-01 -1.17079544e+00 -3.05109531e-01
-5.18239439e-01 3.37582320e-01 8.88349295e-01 9.84665036e-01
-1.45136511e+00 7.90142000e-01 -5.80782473e-01 3.17730047e-02
5.65450370e-01 2.62816370e-01 -5.30979455e-01 -2.08403274e-01
-1.12724173e+00 8.17927122e-01 3.44696820e-01 1.53477386e-01
-9.87122953e-01 -8.99511337e-01 -1.10215676e+00 1.59516811e-01
1.59223065e-01 -5.64589024e-01 9.44303572e-01 -1.77271569e+00
-1.45791972e+00 1.13729846e+00 4.53842580e-01 -4.74557817e-01
7.33878374e-01 -3.79541069e-01 -3.89404446e-01 3.34044248e-01
-2.11887598e-01 1.20374966e+00 1.08057606e+00 -1.69924355e+00
-5.38728416e-01 8.91079679e-02 4.47820455e-01 3.38723928e-01
-8.06215346e-01 -3.65547910e-02 -5.11766613e-01 -8.54161203e-01
-4.43736494e-01 -7.37831891e-01 -2.94901341e-01 4.30651367e-01
-3.75187159e-01 4.51104522e-01 1.01977170e+00 -1.02063823e+00
8.70201588e-01 -2.11852145e+00 -4.62444536e-02 2.50493586e-01
1.88213319e-01 1.59004137e-01 -6.38458252e-01 1.57096959e-03
-2.96166420e-01 2.16483310e-01 -7.08530962e-01 -7.35267997e-01
2.37897858e-01 -1.59215853e-02 1.20817898e-02 2.61078358e-01
7.82127857e-01 1.19119000e+00 -1.06615269e+00 -5.41904211e-01
5.20753384e-01 9.70366716e-01 -7.18488753e-01 4.56200540e-01
-5.25378212e-02 3.21623832e-01 1.14256650e-01 9.05102611e-01
1.00256932e+00 -1.07551113e-01 5.71633801e-02 -5.56585431e-01
4.80776727e-02 -3.55034739e-01 -1.03629696e+00 1.87416136e+00
-9.82970774e-01 6.21099889e-01 2.63038337e-01 -7.22754657e-01
6.84096456e-01 2.20225319e-01 4.52457845e-01 -8.56108487e-01
1.89388767e-01 9.36116725e-02 -3.79341304e-01 -1.81629360e-01
8.25169623e-01 -1.78966448e-01 -2.94089504e-02 6.51729524e-01
2.62118638e-01 -4.46153313e-01 8.88774022e-02 2.57329941e-01
9.25944865e-01 5.23369431e-01 3.47876586e-02 -6.22479394e-02
3.39034468e-01 -3.22922438e-01 9.35447440e-02 4.61071312e-01
-1.43317029e-01 8.50890398e-01 2.00244382e-01 -1.30579993e-01
-1.52064073e+00 -1.01501799e+00 8.04452002e-02 1.39580524e+00
3.07125479e-01 -2.52730280e-01 -1.10514665e+00 -8.33387971e-01
-4.22988743e-01 4.73724157e-01 -9.02042806e-01 -1.39525846e-01
-3.89500916e-01 -5.99087596e-01 4.67955619e-01 5.71898699e-01
9.17547345e-01 -1.21520758e+00 -4.87119615e-01 -1.42550796e-01
-1.01251423e-01 -1.12166464e+00 -7.52386272e-01 3.29723388e-01
-6.15729094e-01 -9.68957305e-01 -9.77154732e-01 -9.69005466e-01
9.74526227e-01 1.50345489e-01 1.18460572e+00 2.58793265e-01
-2.32623383e-01 5.96457064e-01 -4.01395947e-01 -3.75057012e-01
-5.95644772e-01 -4.74607199e-02 -3.98895800e-01 -5.92140853e-02
-2.23389417e-01 -5.76187730e-01 -1.00912869e+00 2.88957268e-01
-1.27669406e+00 3.06362450e-01 9.95767772e-01 1.04826856e+00
5.24716318e-01 3.50171447e-01 3.24758857e-01 -8.34981501e-01
5.73903620e-01 1.05336554e-01 -3.42823386e-01 5.02381146e-01
-1.01973915e+00 1.06290333e-01 5.36809564e-01 -3.41593027e-01
-1.38850045e+00 2.47117475e-01 -3.23488146e-01 -3.29803616e-01
-9.59959999e-02 3.57109636e-01 -3.24952811e-01 -5.59482694e-01
6.88874781e-01 1.22542232e-01 5.32424524e-02 -1.66249260e-01
8.12083006e-01 3.01741660e-01 9.77996945e-01 -4.29016858e-01
1.17607450e+00 4.68179435e-01 -2.04992205e-01 -3.24090540e-01
-7.09978759e-01 -1.22763142e-01 -4.76346672e-01 -4.81459409e-01
7.55823910e-01 -1.11444807e+00 -5.27005017e-01 6.40842497e-01
-8.44004691e-01 -7.01057911e-01 -6.30501568e-01 -5.01590855e-02
-6.21674776e-01 4.68294382e-01 -8.14880490e-01 -8.08839977e-01
-5.97605705e-01 -9.42598104e-01 1.29338884e+00 2.01546997e-01
3.09493840e-01 -1.18988812e+00 -3.03448886e-01 4.87185866e-01
7.00188458e-01 6.74648941e-01 4.43155229e-01 2.82841623e-01
-3.68813157e-01 -2.45475471e-01 -4.98179138e-01 9.59785402e-01
7.74207637e-02 -1.76484153e-01 -1.21846950e+00 -2.77095079e-01
-3.20111990e-01 -7.59799719e-01 9.31424022e-01 2.47352704e-01
1.26363766e+00 -1.16653472e-01 3.31484288e-01 9.04521346e-01
1.70569909e+00 -3.76874983e-01 1.01920128e+00 3.71546388e-01
7.32790053e-01 5.79429328e-01 8.51577759e-01 3.58181983e-01
1.82822093e-01 3.12831104e-01 8.20209801e-01 -1.33316207e+00
-3.63962263e-01 -4.13597405e-01 4.37220991e-01 2.56384760e-01
-3.89408827e-01 -1.66093826e-01 -3.76571715e-01 5.43367386e-01
-1.37744629e+00 -7.64149189e-01 2.02426836e-01 1.84533298e+00
1.21765804e+00 -4.38760370e-02 5.39066829e-02 4.01567250e-01
7.46094048e-01 8.51521716e-02 -4.42264169e-01 -4.67874974e-01
-4.27961081e-01 6.35842085e-01 7.33395457e-01 4.96262908e-01
-1.28729331e+00 1.06689215e+00 6.54074764e+00 6.86367095e-01
-1.14450753e+00 2.22084686e-01 9.94966447e-01 2.14533612e-01
-5.78628182e-01 -1.72324479e-01 1.69756174e-01 1.69169307e-01
3.91021222e-01 2.30952963e-01 6.60024464e-01 8.67797256e-01
9.65175033e-02 1.77370161e-01 -8.41332853e-01 9.32711720e-01
8.21948349e-02 -1.01070487e+00 -2.00361341e-01 -8.64322856e-02
1.13558280e+00 -5.10273457e-01 8.11150253e-01 2.36267671e-01
4.41410124e-01 -1.10447359e+00 8.51410687e-01 1.84370056e-01
1.26278174e+00 -9.14222658e-01 7.69275546e-01 -2.72253603e-01
-1.11994267e+00 -9.04498808e-03 -9.67097431e-02 1.68937117e-01
3.11802030e-01 4.59611475e-01 -4.76006985e-01 5.82569480e-01
7.81930864e-01 7.50477612e-01 -7.27738559e-01 6.13222361e-01
-7.81363785e-01 5.33257186e-01 1.19131960e-01 6.08696342e-01
1.82777733e-01 -1.82251465e-02 1.43889844e-01 1.38406456e+00
2.83262163e-01 -5.15939407e-02 -1.34267822e-01 9.46031511e-01
-4.90451425e-01 8.72767046e-02 -1.79100066e-01 -2.27314144e-01
-2.12695494e-01 1.69634593e+00 -5.93540013e-01 -1.96713299e-01
-2.24285156e-01 1.63142180e+00 -5.18033234e-03 3.63985687e-01
-1.03217864e+00 -5.44802248e-01 2.47322828e-01 -3.71674821e-02
3.97854179e-01 9.50733647e-02 -4.61674422e-01 -9.49646831e-01
-1.54143181e-02 -1.20435107e+00 1.06086873e-01 -1.17075658e+00
-1.20097041e+00 7.28407502e-01 -5.59383094e-01 -1.39320421e+00
1.40970170e-01 -7.80481219e-01 -6.67752564e-01 6.62679136e-01
-1.92095172e+00 -1.86247826e+00 -7.49327183e-01 7.46301115e-01
1.67671174e-01 1.42694354e-01 6.94610894e-01 4.60870534e-01
-3.20760049e-02 9.12408292e-01 -2.27454290e-01 2.20066741e-01
1.12131619e+00 -1.77359438e+00 3.40818375e-01 1.08383560e+00
-1.07890978e-01 2.86221176e-01 9.38307345e-01 -2.47674048e-01
-1.13396156e+00 -1.20239174e+00 4.67220515e-01 -1.22733340e-01
2.97658414e-01 -2.76620626e-01 -5.34487307e-01 4.08930957e-01
8.91460955e-01 1.87643036e-01 5.85157394e-01 -6.51726220e-03
-5.12864649e-01 -1.53510690e-01 -1.38035452e+00 5.19858003e-01
1.16238558e+00 -6.10427976e-01 -1.05026387e-01 2.01339647e-01
8.05826783e-01 -4.01804805e-01 -1.03664017e+00 3.67777318e-01
5.13919592e-01 -1.12833452e+00 1.15553164e+00 -3.66353333e-01
1.11968505e+00 -3.30129385e-01 8.24005082e-02 -1.38507891e+00
-4.96226437e-02 -6.99173629e-01 1.79389253e-01 1.24584758e+00
4.50088292e-01 -3.39617766e-02 7.60319054e-01 5.05118549e-01
-1.82577118e-01 -5.50244212e-01 -8.25998634e-02 -6.28351808e-01
-5.21097109e-02 -5.53036451e-01 5.98566115e-01 1.03989887e+00
-1.59933522e-01 -8.47458541e-02 -9.60164607e-01 2.06431672e-01
8.28700781e-01 2.59174854e-02 7.88293362e-01 -5.25743365e-01
-6.62345111e-01 -2.87144274e-01 -2.72967309e-01 -6.78744018e-01
-8.12907238e-03 -6.83893025e-01 3.26880485e-01 -1.34837008e+00
3.30512136e-01 -5.40171087e-01 -6.36835277e-01 7.53225923e-01
-3.94718587e-01 9.46228445e-01 1.92424566e-01 -2.20547438e-01
-8.91466141e-01 5.79090297e-01 1.49522817e+00 -5.29218614e-01
-3.81181762e-02 -3.63319367e-01 -1.15356064e+00 5.56298256e-01
7.10413277e-01 -1.79131463e-01 -4.20337647e-01 -4.64267761e-01
4.27020788e-01 -3.74637961e-01 6.75480068e-01 -1.02871335e+00
-8.14529061e-02 1.17561288e-01 5.58775127e-01 -4.60814685e-02
1.54035449e-01 -9.11733747e-01 -1.99427858e-01 3.57363760e-01
-5.95902264e-01 -5.05734943e-02 2.86942512e-01 3.54733914e-01
-3.88881922e-01 7.59392381e-02 9.37502503e-01 2.35890131e-02
-9.09190357e-01 3.06970358e-01 1.49706826e-01 -8.51939172e-02
8.06445956e-01 -1.46125302e-01 -4.12829630e-02 -8.31676424e-01
-6.48319542e-01 1.75974425e-02 7.70972490e-01 2.25643083e-01
7.34883130e-01 -1.35364878e+00 -7.52662599e-01 -6.45636767e-02
2.93316185e-01 -3.15604508e-01 5.44726789e-01 6.44013882e-01
-4.53071445e-01 -4.18808371e-01 -7.13688910e-01 -3.88621688e-01
-1.31290126e+00 9.16626453e-01 4.30780470e-01 -4.54518348e-01
-3.10771346e-01 8.11461687e-01 2.08948240e-01 -4.07301039e-01
1.78729534e-01 -6.85891062e-02 1.28338695e-01 -4.59472984e-01
4.59499031e-01 1.21527359e-01 8.98604244e-02 -2.56407440e-01
5.95779493e-02 3.84985745e-01 1.12955838e-01 -3.10207963e-01
1.49187672e+00 -2.41734162e-01 9.41734165e-02 -3.55275422e-01
1.10548079e+00 7.60459378e-02 -1.74273944e+00 -1.76632389e-01
-4.84163493e-01 -4.79541749e-01 2.93213040e-01 -1.31114209e+00
-1.56696737e+00 9.03047919e-01 1.05894876e+00 -1.33070961e-01
1.99237514e+00 -3.75253558e-01 8.85896325e-01 1.38961017e-01
1.23798832e-01 -1.40408039e+00 5.34573257e-01 -1.28121078e-01
9.90681946e-01 -1.76483369e+00 -1.12313800e-01 -4.37090874e-01
-9.22595739e-01 9.03345704e-01 5.11627614e-01 -1.93136230e-01
3.32163781e-01 7.24590421e-01 3.91006529e-01 -6.55766809e-03
-2.89629728e-01 -4.69696939e-01 5.06626666e-01 7.88608730e-01
5.43051183e-01 -8.17669854e-02 -1.23007178e-01 4.29641306e-01
-1.97536826e-01 -4.27326886e-03 3.14335763e-01 8.44841421e-01
-5.19635044e-02 -1.26125765e+00 -2.09454656e-01 -2.07495034e-01
-4.33738679e-01 -4.35389876e-01 -3.14339161e-01 9.37168002e-01
4.30979073e-01 8.89530659e-01 -1.09193951e-01 -5.87181389e-01
3.30178231e-01 -3.36628288e-01 6.10872924e-01 -3.74464579e-02
-9.53064084e-01 2.71828417e-02 -1.34712771e-01 -5.49446166e-01
-6.63735449e-01 -2.15730950e-01 -9.10032451e-01 -3.39410305e-01
-3.30465734e-01 -1.76498026e-01 8.04501951e-01 7.16486335e-01
1.12687118e-01 7.28301883e-01 8.88762355e-01 -8.24477017e-01
-2.74140239e-01 -7.15540826e-01 -6.09909713e-01 9.45144415e-01
3.78903776e-01 -1.70408770e-01 -2.45224908e-01 6.63254917e-01] | [11.40692138671875, -1.5529546737670898] |
95811514-c479-4c8b-829b-8af546065f85 | learning-representations-for-incomplete-time | null | null | https://ojs.aaai.org/index.php/AAAI/article/view/17070 | https://ojs.aaai.org/index.php/AAAI/article/view/17070/16877 | Learning Representations for Incomplete Time Series Clustering | Time-series clustering is an essential unsupervised technique for data analysis, applied to many real-world fields, such as medical analysis and DNA microarray. Existing clustering methods are usually based on the assumption that the data is complete. However, time series in real-world applications often contain missing values. Traditional strategy (imputing first and then clustering) does not optimize the imputation and clustering process as a whole, which not only makes per- formance dependent on the combination of imputation and clustering methods but also fails to achieve satisfactory re- sults. How to best improve the clustering performance on incomplete time series remains a challenge. This paper pro- poses a novel unsupervised temporal representation learning model, named Clustering Representation Learning on Incom- plete time-series data (CRLI). CRLI jointly optimizes the im- putation and clustering process to impute more discrimina- tive values for clustering and make the learned representa- tions possessed good clustering property. Also, to reduce the error propagation from imputation to clustering, we introduce a discriminator to make the distribution of imputation values close to the true one and train CRLI in an alternating train- ing manner. An experiment conducted on eight real-world in- complete time-series datasets shows that CRLI outperforms existing methods. We demonstrates the effectiveness of the learned representations and the convergence of the model through visualization analysis. Moreover, we reveal that the joint training strategy can impute values close to the true ones in those important sub-sequences, and impute more discrim- inative values in those less important sub-sequences at the same time, making the imputed sequence cluster-friendly. | ['Garrison W. Cottrell', 'Sen Li', 'Chuxin Chen', 'Qianli Ma'] | 2021-05-18 | null | null | null | aaai-2021-5 | ['time-series-clustering'] | ['time-series'] | [ 1.99678123e-01 -6.51593149e-01 -1.55579448e-01 -6.07576132e-01
-7.47933269e-01 -4.74984616e-01 6.03705794e-02 9.97967571e-02
-2.40221307e-01 8.00664246e-01 1.70205474e-01 -9.14631039e-02
-6.18463874e-01 -7.88591743e-01 -7.08279312e-01 -1.23971474e+00
-1.17852159e-01 6.57294393e-01 -4.36162859e-01 2.02418536e-01
1.13555364e-01 2.98116535e-01 -1.38403046e+00 4.14536297e-01
1.17212617e+00 4.44729269e-01 4.45918739e-02 1.56232953e-01
-2.45336995e-01 4.36624199e-01 -7.04609692e-01 2.20411405e-01
-1.04075007e-03 -6.13101065e-01 -1.91237018e-01 -2.64775716e-02
-5.39057314e-01 2.76416093e-01 -2.20795050e-01 8.49978149e-01
4.77848113e-01 1.64153501e-01 7.90017188e-01 -1.45629168e+00
-5.03758132e-01 7.28548110e-01 -9.84117508e-01 -1.47779286e-01
-1.15687713e-01 -9.18358415e-02 4.08404291e-01 -3.22654068e-01
4.47766066e-01 9.98494506e-01 7.99072981e-01 2.93438464e-01
-1.57779038e+00 -9.39493716e-01 -1.03989802e-02 1.35591030e-01
-1.64236391e+00 -7.91027397e-02 8.33176911e-01 -4.67031807e-01
3.69509786e-01 4.20872122e-01 2.75270015e-01 1.05045104e+00
3.36871952e-01 5.64386368e-01 1.10557640e+00 -3.95480134e-02
3.31285655e-01 -2.71806479e-01 1.65150926e-01 -5.60246222e-03
1.57193273e-01 5.32771870e-02 -9.65787172e-02 -1.38623685e-01
6.31607950e-01 6.74571812e-01 -2.33660921e-01 -2.77357638e-01
-1.50774395e+00 5.08748531e-01 2.45275885e-01 6.43885136e-01
-4.26786870e-01 -1.95700973e-01 3.81630540e-01 2.39319399e-01
3.98408085e-01 1.55954003e-01 -5.10817170e-01 -8.17890167e-02
-1.03652370e+00 -1.46391630e-01 3.72683704e-01 6.70471191e-01
7.00038791e-01 3.58271860e-02 -2.53505290e-01 8.86123776e-01
-8.54486451e-02 2.87433684e-01 8.49276364e-01 -7.84755766e-01
2.46954322e-01 7.09304512e-01 1.32326022e-01 -1.01789320e+00
-5.14691949e-01 -6.74727142e-01 -1.62962139e+00 -1.49219796e-01
5.06426454e-01 -3.91047835e-01 -9.20338035e-01 2.02786732e+00
1.91611543e-01 7.62259603e-01 1.32265598e-01 7.85266876e-01
4.45540011e-01 9.02221680e-01 9.67528448e-02 -8.45375836e-01
8.64257097e-01 -3.04668188e-01 -8.91058326e-01 3.10220271e-01
5.77989221e-01 -6.29017234e-01 5.51796377e-01 4.15150583e-01
-4.64679718e-01 -6.59403205e-01 -7.29456961e-01 3.45007569e-01
-2.36213148e-01 1.81329161e-01 6.86431110e-01 2.36948058e-01
-4.82567400e-01 6.48752689e-01 -9.90415633e-01 -6.88922331e-02
2.76459694e-01 5.97317278e-01 -5.48474967e-01 -5.02266623e-02
-9.21642125e-01 2.84838289e-01 4.85425085e-01 3.82259458e-01
-5.66993356e-01 -8.80132198e-01 -4.79442269e-01 -3.35628949e-02
7.75014982e-02 -4.23076332e-01 5.63060820e-01 -1.09434962e+00
-1.00832021e+00 4.41635221e-01 -4.00918543e-01 -1.86825901e-01
3.31895620e-01 2.57467210e-01 -6.99591219e-01 -3.52247626e-01
1.82073519e-01 3.03853750e-01 5.43605626e-01 -1.13866198e+00
-2.29377881e-01 -6.43525720e-01 -7.45029211e-01 -1.37492016e-01
-2.76827663e-01 -3.65673274e-01 -3.17490488e-01 -9.73875821e-01
5.85771084e-01 -7.32887149e-01 -4.35521483e-01 -4.60712403e-01
-3.99228841e-01 -1.06318519e-01 8.91578972e-01 -6.12103879e-01
1.15422297e+00 -2.35074091e+00 3.31698447e-01 3.64688724e-01
-4.09835055e-02 -8.27464163e-02 -3.09191138e-01 5.44266999e-01
-6.72042131e-01 2.69641392e-02 -5.69188237e-01 -8.40461925e-02
-2.28691310e-01 4.00548756e-01 -1.78987950e-01 6.28281653e-01
-7.27643631e-03 5.54006815e-01 -8.20216596e-01 -3.10903579e-01
1.80412322e-01 5.63150167e-01 -2.96322167e-01 3.41357380e-01
-1.45859241e-01 9.81391490e-01 -3.68544430e-01 3.37250859e-01
8.49625409e-01 -1.74980879e-01 3.86287481e-01 -2.13859394e-01
-2.20955104e-01 -4.20811623e-01 -1.18328822e+00 1.62805653e+00
-1.77295133e-01 2.95810074e-01 -2.54989207e-01 -1.57560492e+00
1.28284240e+00 3.00016046e-01 1.09819949e+00 -6.18160844e-01
1.13625102e-01 -2.61383913e-02 1.31408319e-01 -5.70415199e-01
-1.44411311e-01 -1.11275434e-01 -4.57364544e-02 3.83095920e-01
-3.25642884e-01 4.06169474e-01 5.21104597e-02 -1.61371715e-02
7.95442879e-01 2.30978280e-01 -1.43643208e-02 -1.20726570e-01
4.01594043e-01 -3.48736234e-02 1.18882513e+00 4.80546921e-01
1.02879390e-01 7.43028224e-01 6.31121039e-01 -3.25493217e-01
-9.82899666e-01 -9.89693165e-01 -2.12203503e-01 6.92675054e-01
6.03698716e-02 -8.73423889e-02 -6.33457124e-01 -4.67154235e-01
-1.08553804e-01 6.70450389e-01 -7.54143417e-01 -2.62715369e-01
-3.81947637e-01 -1.38779211e+00 3.29134613e-01 4.64268625e-01
3.08304697e-01 -9.09921527e-01 -6.39567524e-02 4.21123296e-01
-6.31754875e-01 -5.04392326e-01 -3.21869403e-01 4.03217226e-01
-1.06255627e+00 -8.52913857e-01 -5.30292988e-01 -6.75362825e-01
9.42510784e-01 2.19446808e-01 6.10857189e-01 9.64022204e-02
-2.15711936e-01 -2.05227986e-01 -3.56521845e-01 -1.56716987e-01
-1.72163993e-01 -2.52499301e-02 9.04347077e-02 3.49539399e-01
6.72982693e-01 -8.60182464e-01 -4.30431724e-01 4.86885279e-01
-9.47930574e-01 5.70584685e-02 5.10435164e-01 1.03015232e+00
9.51798797e-01 6.31892145e-01 9.27846789e-01 -8.59078050e-01
2.67299056e-01 -9.18018103e-01 -4.82246339e-01 3.41459900e-01
-4.07665610e-01 1.45898208e-01 8.85565162e-01 -6.97201014e-01
-8.76363635e-01 3.91711473e-01 -1.63214132e-01 -7.58831263e-01
-1.93961099e-01 7.83437192e-01 -2.90666848e-01 5.44089079e-01
4.13548172e-01 6.47522330e-01 3.66838574e-01 -5.05634069e-01
1.34045621e-02 7.27218330e-01 6.48140550e-01 -6.25709295e-01
5.93166828e-01 4.12646800e-01 -1.51540805e-02 -5.78810155e-01
-4.09355432e-01 -5.20781338e-01 -7.47363150e-01 -1.27368271e-01
6.36996090e-01 -9.04990375e-01 -9.06059504e-01 4.48590428e-01
-7.56532788e-01 -1.16887070e-01 -2.33704634e-02 6.45020247e-01
-5.04717171e-01 4.10414279e-01 -2.89676338e-01 -7.94417024e-01
-2.59058885e-02 -9.45375204e-01 7.51049936e-01 1.09925017e-01
-1.62459731e-01 -7.89088488e-01 1.39987081e-01 9.36291441e-02
1.20590672e-01 6.97812080e-01 1.02227330e+00 -6.15052104e-01
-3.64096373e-01 -2.69338250e-01 -7.34472349e-02 -2.89666466e-02
3.87967587e-01 2.22100616e-01 -7.54439771e-01 -2.39148214e-01
9.44206864e-02 1.97544292e-01 7.88440526e-01 6.55775189e-01
1.64960468e+00 -1.69618234e-01 -4.04668808e-01 7.33330965e-01
1.36144876e+00 4.38728452e-01 9.65385675e-01 2.23424565e-02
4.56116587e-01 6.76759660e-01 6.50971591e-01 5.74365914e-01
1.65633142e-01 5.11822760e-01 2.13239178e-01 -1.68231428e-01
4.05223757e-01 -2.50496566e-01 1.86621889e-01 1.04014945e+00
-1.45057529e-01 -1.01659678e-01 -6.75971925e-01 5.73693693e-01
-2.25060821e+00 -1.40419376e+00 -4.58033442e-01 2.39026475e+00
9.61543441e-01 -3.77863079e-01 2.25457504e-01 3.78392190e-01
9.02613640e-01 -2.70571440e-01 -7.30317891e-01 5.42218424e-02
-3.13739926e-01 -7.58432150e-02 1.08627297e-01 1.04553565e-01
-9.40992951e-01 3.26450557e-01 5.68993568e+00 7.56904721e-01
-1.13631332e+00 2.18739640e-03 9.52720881e-01 2.13022262e-01
-2.80503988e-01 -1.21774681e-01 -2.59032637e-01 9.46382284e-01
1.03281832e+00 -2.76261806e-01 4.81081486e-01 4.84825581e-01
6.34779155e-01 1.91420212e-01 -9.76694882e-01 1.06175840e+00
-2.46135235e-01 -1.18911314e+00 -1.40413240e-01 9.02557746e-03
7.43396938e-01 -2.93430924e-01 -8.34121555e-02 3.88297826e-01
3.06444824e-01 -1.20161605e+00 1.51930422e-01 9.46490049e-01
5.82300007e-01 -1.03384209e+00 1.02954447e+00 6.87381804e-01
-9.88460600e-01 -1.31905138e-01 -5.22810161e-01 -8.23064148e-03
2.08161888e-03 9.47753727e-01 -5.19479215e-01 8.55475783e-01
5.91464877e-01 1.02295065e+00 -3.36656749e-01 1.19891417e+00
1.27568230e-01 5.94864547e-01 -2.18693718e-01 1.81114316e-01
-2.51844078e-01 -7.67312706e-01 5.33876866e-02 1.05015147e+00
6.48275137e-01 3.60112756e-01 2.84844398e-01 7.33991683e-01
5.07314764e-02 3.07286829e-02 -4.46680754e-01 -1.68670908e-01
6.88622713e-01 1.07975531e+00 -7.41976321e-01 -1.53917387e-01
-1.25280291e-01 7.46995509e-01 1.70109615e-01 6.60987973e-01
-9.57073271e-01 -3.14089537e-01 5.53165793e-01 -1.14924520e-01
8.55922550e-02 -3.35359365e-01 -6.02781296e-01 -1.11227036e+00
-1.85352907e-01 -9.04260576e-01 6.88559294e-01 -7.01612294e-01
-1.60145855e+00 3.85191947e-01 -1.27296522e-01 -1.57610774e+00
-2.93508470e-01 -1.41526330e-02 -5.68877399e-01 8.26833665e-01
-1.04198003e+00 -9.61880505e-01 -3.08841050e-01 7.80911624e-01
2.94242471e-01 4.72255573e-02 8.10766697e-01 3.96374851e-01
-9.71130371e-01 5.45440555e-01 7.67811716e-01 2.49270186e-01
8.10395300e-01 -7.99332261e-01 -3.67300361e-01 7.48875737e-01
-1.01729587e-01 9.73964930e-01 6.49991453e-01 -7.91940391e-01
-1.42365789e+00 -1.39598775e+00 6.61467791e-01 -1.80872828e-01
4.04180467e-01 -2.92721063e-01 -1.28719938e+00 5.40536404e-01
-1.00590825e-01 -1.92636505e-01 9.72986341e-01 2.56217659e-01
-2.81090021e-01 -5.08519709e-01 -1.10320747e+00 3.93093139e-01
6.13321006e-01 -2.23228544e-01 -4.15616184e-01 4.53638643e-01
4.99736398e-01 -1.67025834e-01 -1.12283432e+00 4.93498385e-01
4.92583454e-01 -8.66256118e-01 1.02988100e+00 -5.81367910e-01
3.80018413e-01 -8.16417217e-01 -7.60784522e-02 -1.18081689e+00
-5.03701866e-01 -2.43502632e-01 1.85344085e-01 1.47904730e+00
1.74709037e-01 -5.32484055e-01 6.32872760e-01 4.11108941e-01
-4.37237024e-02 -4.48866934e-01 -9.28834140e-01 -8.14758182e-01
6.84696212e-02 -2.39270121e-01 8.80032361e-01 1.40464497e+00
-7.60160834e-02 3.21709062e-03 -5.16012251e-01 3.97488415e-01
8.84964943e-01 4.26004052e-01 6.91375554e-01 -1.33778095e+00
-1.95182756e-01 -1.28605500e-01 -2.66098410e-01 -4.46553499e-01
2.50285536e-01 -7.81728685e-01 2.20094323e-01 -1.26360917e+00
4.03567672e-01 -6.86440468e-01 -7.11012542e-01 6.32296860e-01
-1.63522348e-01 1.19468413e-01 -3.41220915e-01 4.56980824e-01
-2.65668422e-01 4.57757115e-01 1.04378200e+00 -2.66026199e-01
-4.76891100e-01 1.59600809e-01 -7.08656132e-01 2.35454574e-01
8.31441760e-01 -6.19334340e-01 -6.09867990e-01 -2.70945549e-01
-1.93572521e-01 5.61549723e-01 1.29075900e-01 -9.09270406e-01
3.63654017e-01 -5.33078432e-01 8.91097963e-01 -8.40146720e-01
4.21725214e-02 -9.00563657e-01 1.02306175e+00 3.86353314e-01
-2.77815014e-01 9.66456067e-03 1.00775406e-01 6.78216934e-01
-3.10790330e-01 9.55247432e-02 7.20963418e-01 8.85767192e-02
-4.26351905e-01 3.83354336e-01 -2.14176744e-01 -2.91237384e-01
1.11894166e+00 -2.49348983e-01 -1.29543608e-02 -3.12182754e-01
-8.30930591e-01 4.55814183e-01 4.10150617e-01 3.46261322e-01
4.55063939e-01 -1.29023254e+00 -8.98783445e-01 5.03522515e-01
-5.47822975e-02 -1.49895862e-01 7.10317671e-01 9.69778955e-01
-6.04985058e-02 -5.06128781e-02 -2.89941669e-01 -9.02954280e-01
-1.03881395e+00 9.02472079e-01 -6.23438507e-03 -1.69239901e-02
-5.11756063e-01 2.04254732e-01 1.84238359e-01 -6.35326266e-01
1.60488814e-01 2.35897359e-02 -3.03049982e-01 -7.68467598e-03
4.22824174e-01 2.92691290e-01 -1.46709997e-02 -4.55711722e-01
-2.38974988e-01 5.84006786e-01 1.65803335e-03 1.92270383e-01
1.61955059e+00 -9.24912617e-02 -5.20850420e-01 6.09340370e-01
1.14445400e+00 -3.66758913e-01 -1.09821069e+00 1.19943753e-01
-4.40343991e-02 -4.53061253e-01 -2.38144651e-01 -7.95857787e-01
-1.05782902e+00 8.07311833e-01 6.36979878e-01 -9.26834494e-02
1.39412487e+00 -4.30616766e-01 3.50819170e-01 9.79522392e-02
4.25437987e-01 -8.01998436e-01 -1.82634577e-01 1.40215009e-01
5.63510478e-01 -8.82322371e-01 -1.91929951e-01 -2.93200135e-01
-5.28475940e-01 1.01784563e+00 5.04196763e-01 -2.16115750e-02
5.02839267e-01 3.32527846e-01 1.44779803e-02 2.05000430e-01
-6.74948633e-01 2.02337027e-01 -8.85851160e-02 8.36133718e-01
6.44966781e-01 1.87054828e-01 -4.27846491e-01 9.22630310e-01
1.11793913e-01 3.22525948e-01 4.15794164e-01 4.90209192e-01
-5.90746552e-02 -1.16981816e+00 -5.82942128e-01 5.38514376e-01
-2.34924704e-01 2.93438315e-01 -8.01470056e-02 5.42093396e-01
2.25049868e-01 1.08259821e+00 3.19969147e-01 -4.24393982e-01
3.89328822e-02 1.00797629e-02 4.06868160e-02 -1.98019072e-01
-3.45481843e-01 5.06517649e-01 -3.47989768e-01 -3.38290632e-01
-5.05655587e-01 -7.16261089e-01 -1.49779701e+00 -5.01265347e-01
-2.13052288e-01 5.26062429e-01 4.02828515e-01 9.18980896e-01
6.37934864e-01 7.97581971e-01 1.04041409e+00 -4.02884483e-01
-2.45947286e-01 -7.80120194e-01 -6.96024001e-01 6.52196825e-01
1.66627839e-01 -3.67606908e-01 -2.56330878e-01 2.05601946e-01] | [7.297471046447754, 3.576587438583374] |
bf71a71d-d56c-4903-bec5-7e6e97805341 | shikra-unleashing-multimodal-llm-s | 2306.15195 | null | https://arxiv.org/abs/2306.15195v2 | https://arxiv.org/pdf/2306.15195v2.pdf | Shikra: Unleashing Multimodal LLM's Referential Dialogue Magic | In human conversations, individuals can indicate relevant regions within a scene while addressing others. In turn, the other person can then respond by referring to specific regions if necessary. This natural referential ability in dialogue remains absent in current Multimodal Large Language Models (MLLMs). To fill this gap, this paper proposes an MLLM called Shikra, which can handle spatial coordinate inputs and outputs in natural language. Its architecture consists of a vision encoder, an alignment layer, and a LLM. It is designed to be straightforward and simple, without the need for extra vocabularies, position encoder, pre-/post-detection modules, or external plug-in models. All inputs and outputs are in natural language form. Referential dialogue is a superset of various vision-language (VL) tasks. Shikra can naturally handle location-related tasks like REC and PointQA, as well as conventional VL tasks such as Image Captioning and VQA. Experimental results showcase Shikra's promising performance. Furthermore, it enables numerous exciting applications, like providing mentioned objects' coordinates in chains of thoughts and comparing user-pointed regions similarities. Our code, model and dataset are accessed at https://github.com/shikras/shikra. | ['Rui Zhao', 'Feng Zhu', 'Richong Zhang', 'Weili Zeng', 'Zhao Zhang', 'Keqin Chen'] | 2023-06-27 | null | null | null | null | ['image-captioning'] | ['computer-vision'] | [-1.54922545e-01 2.46006578e-01 -6.15657344e-02 -5.89956164e-01
-9.79572773e-01 -7.95506477e-01 9.44774866e-01 1.70054153e-01
-4.90913659e-01 4.56481874e-01 5.91146111e-01 -2.14469939e-01
3.99233729e-01 -5.12088537e-01 -4.47949231e-01 -3.95150363e-01
5.59994757e-01 7.40317047e-01 4.23381180e-02 -4.05967861e-01
2.18827993e-01 3.12341690e-01 -9.04648840e-01 6.26832604e-01
4.32035297e-01 6.91089571e-01 3.64912450e-01 9.86119092e-01
-3.62161756e-01 1.25703847e+00 -6.02514267e-01 -7.56965816e-01
-2.40301117e-01 -5.21211565e-01 -9.71571624e-01 7.25736842e-02
8.69702816e-01 -4.30859983e-01 -3.80387068e-01 7.76883006e-01
7.42212772e-01 2.13801935e-01 4.99381751e-01 -1.27529502e+00
-1.15631366e+00 4.02478755e-01 -5.58164597e-01 -1.08636603e-01
9.45337057e-01 3.97643507e-01 1.06941235e+00 -1.17967856e+00
5.47989666e-01 1.68445146e+00 5.20878017e-01 8.88425291e-01
-8.91936541e-01 -8.49445686e-02 6.56040832e-02 -3.24167535e-02
-1.48739755e+00 -6.35764003e-01 6.58576846e-01 -3.84459287e-01
1.03983653e+00 6.42176390e-01 4.61512655e-01 1.10995352e+00
-2.42715284e-01 1.16827273e+00 6.79814339e-01 -3.72276694e-01
-1.45836845e-01 3.29269797e-01 3.13846320e-02 8.79290819e-01
-5.39101005e-01 -4.69062209e-01 -5.45772731e-01 -7.30142295e-02
1.02333808e+00 4.00780365e-02 -2.44460285e-01 -3.70513111e-01
-1.73831284e+00 7.67467976e-01 7.58360386e-01 2.02448204e-01
-1.04502790e-01 3.73566031e-01 3.78188372e-01 1.47156388e-01
8.83803368e-02 4.99783814e-01 1.72056188e-03 2.97296569e-02
-5.33640504e-01 3.06045234e-01 4.95100826e-01 1.20422554e+00
6.20111644e-01 -4.44503903e-01 -6.45087123e-01 1.02476573e+00
4.71416324e-01 7.75678754e-01 3.44855994e-01 -1.18782532e+00
8.16946805e-01 7.67063141e-01 2.08717734e-01 -1.18883789e+00
-4.51124966e-01 8.39900747e-02 -9.56636548e-01 -3.19307178e-01
3.91752571e-01 -4.89771403e-02 -4.03440505e-01 1.63601661e+00
3.37091744e-01 -2.37348974e-01 2.82913029e-01 1.21515310e+00
1.52286708e+00 9.26335394e-01 2.17001475e-02 1.24228060e-01
1.58523738e+00 -1.33419347e+00 -7.72870302e-01 -3.02107006e-01
6.06228054e-01 -7.75440693e-01 1.44706142e+00 -2.21102312e-01
-1.37846923e+00 -7.19098508e-01 -3.41906607e-01 -9.41556513e-01
-5.65203309e-01 2.73847729e-01 5.19537687e-01 -5.19986600e-02
-1.48528135e+00 -3.53299439e-01 -1.98996186e-01 -6.18990302e-01
4.07282375e-02 6.07049577e-02 -4.62987095e-01 5.83685003e-02
-1.21682012e+00 9.85997021e-01 -9.99868512e-02 3.66909355e-01
-5.87496519e-01 -2.77707815e-01 -1.28690445e+00 -2.58820683e-01
2.67894655e-01 -1.02013743e+00 1.75100303e+00 -7.53465414e-01
-1.25908744e+00 1.15982807e+00 -3.74860227e-01 -3.49595755e-01
6.98164940e-01 -1.54516444e-01 -1.78733677e-01 2.75960118e-01
2.70510763e-01 1.33554351e+00 5.32246530e-01 -1.09518766e+00
-3.25569928e-01 -3.63830686e-01 5.03275335e-01 6.47248149e-01
-1.41979931e-02 2.16479272e-01 -9.59299326e-01 -3.15832138e-01
-2.09218815e-01 -6.97575569e-01 -2.13014156e-01 3.41884047e-01
-6.69572830e-01 -5.46793699e-01 6.42215073e-01 -5.74972093e-01
9.59819198e-01 -2.08382440e+00 3.06690216e-01 -2.17263415e-01
3.11557561e-01 1.52307656e-02 -4.29802030e-01 7.53233135e-01
3.47886652e-01 -4.08316068e-02 -9.54441428e-02 -5.10119677e-01
1.61225885e-01 5.06764129e-02 -3.71672183e-01 4.57728833e-01
1.97400466e-01 1.61680591e+00 -9.81280327e-01 -7.09968328e-01
6.27834976e-01 5.62384307e-01 -2.09689066e-01 4.34464216e-01
-4.17666912e-01 5.00387847e-01 -2.07377851e-01 5.33086777e-01
2.31071144e-01 -6.10331178e-01 -1.81092307e-01 -2.60328680e-01
-2.16527805e-01 1.66906208e-01 -8.37419152e-01 1.83779347e+00
-6.71044648e-01 9.92585182e-01 9.08843875e-02 -4.43624049e-01
8.37950885e-01 3.02378744e-01 1.08209383e-02 -9.40054476e-01
-9.34836715e-02 -1.71341091e-01 -3.98236811e-01 -7.38555551e-01
7.43178546e-01 3.84281814e-01 -4.67732280e-01 3.90587002e-01
-2.19497219e-01 -3.15144897e-01 6.39875755e-02 6.78813815e-01
5.65851986e-01 -1.08400516e-01 4.14280504e-01 1.62476450e-01
7.36046910e-01 1.18251324e-01 -1.94830179e-01 7.89419234e-01
-6.24276027e-02 7.61852622e-01 3.64153892e-01 -4.97409463e-01
-9.93147671e-01 -1.00907016e+00 2.36908197e-01 1.41750038e+00
5.16207695e-01 -5.59178174e-01 -5.45517981e-01 -3.67115676e-01
-1.55784354e-01 7.53426373e-01 -5.27800560e-01 3.07322145e-01
-5.26935101e-01 2.28115618e-02 7.32121587e-01 6.39232814e-01
6.70454919e-01 -1.31692302e+00 -7.55957663e-01 -2.97184706e-01
-6.51817799e-01 -1.20321965e+00 -9.60576713e-01 -5.30795276e-01
-1.79630905e-01 -9.32826102e-01 -1.11132276e+00 -1.05715692e+00
7.60150075e-01 5.66095352e-01 1.33195388e+00 -3.10266651e-02
-1.44984916e-01 8.39547098e-01 -1.79668650e-01 -2.48678327e-01
-2.13351816e-01 -1.44192606e-01 -2.26196110e-01 -9.34539065e-02
4.16844010e-01 1.00397341e-01 -5.65678477e-01 3.35302681e-01
-5.30572772e-01 6.16319776e-01 5.30056059e-01 6.87419772e-01
5.93438625e-01 -8.68937314e-01 3.66565913e-01 -5.65735221e-01
1.07967293e+00 -1.60174638e-01 -4.24079120e-01 5.85914135e-01
1.79451093e-01 -2.33604521e-01 4.47566301e-01 -4.15803045e-01
-7.35633731e-01 1.34905316e-02 -1.50987521e-01 -2.87726074e-01
-3.63719255e-01 2.19525203e-01 -1.92766920e-01 1.79792777e-01
4.12396252e-01 4.85219896e-01 -2.76653580e-02 -2.07533732e-01
9.55729008e-01 1.01523232e+00 1.00512314e+00 -3.01388979e-01
5.52604258e-01 3.76629949e-01 -3.70036691e-01 -1.09441912e+00
-7.80188143e-01 -7.06907928e-01 -8.76982212e-01 -4.42599118e-01
1.12391353e+00 -1.13380885e+00 -1.10867763e+00 1.21313512e-01
-1.62650490e+00 -3.87894243e-01 -1.08632267e-01 1.83706358e-02
-6.11967683e-01 3.43154907e-01 -5.93216598e-01 -8.20540011e-01
-4.72671926e-01 -1.01863348e+00 1.41312146e+00 4.43468213e-01
-4.74987417e-01 -1.10984182e+00 -8.52085724e-02 7.07749546e-01
3.85013670e-01 -2.23067347e-02 3.19814295e-01 -5.43575704e-01
-5.69174170e-01 -2.06120431e-01 -5.17845988e-01 -1.69027880e-01
8.93543735e-02 -2.73307949e-01 -7.65513361e-01 -2.07791060e-01
-5.15666962e-01 -7.12086618e-01 6.36664510e-01 1.77983046e-01
9.79254127e-01 -6.99688494e-01 -1.51464507e-01 3.72208089e-01
8.16390216e-01 2.38753352e-02 3.82870167e-01 -6.95406571e-02
9.83825684e-01 8.24512482e-01 5.36373854e-01 1.67948335e-01
1.11505842e+00 9.58307624e-01 3.87976527e-01 -5.67130744e-01
-1.78292781e-01 -3.61875862e-01 4.06035095e-01 7.11834729e-01
2.02633321e-01 -4.22461838e-01 -1.00138211e+00 5.74925005e-01
-2.06142330e+00 -1.15303826e+00 -2.17317894e-01 1.67969191e+00
7.51713574e-01 -5.52511513e-01 2.15759873e-01 -6.84200168e-01
5.90036511e-01 4.32608873e-01 -5.51869035e-01 -5.16850054e-01
-2.59157658e-01 -6.97363734e-01 1.04864486e-01 7.93101490e-01
-1.07195592e+00 1.19408035e+00 5.46278095e+00 3.21004689e-01
-8.80359709e-01 4.76787016e-02 7.85021901e-01 7.12370872e-02
-4.71812963e-01 -3.30571532e-01 -7.44315922e-01 1.09983131e-01
5.84373176e-01 1.64203510e-01 4.06753749e-01 6.26168728e-01
3.94071043e-01 2.57430393e-02 -1.29074502e+00 1.55873775e+00
5.32083452e-01 -1.56869805e+00 2.80823201e-01 -2.86816418e-01
2.66380936e-01 -4.55569550e-02 5.66087477e-02 2.27883413e-01
2.84464151e-01 -1.30639839e+00 6.67725027e-01 7.93066740e-01
1.06626272e+00 -4.38165754e-01 7.22527504e-01 4.55256671e-01
-1.07412076e+00 1.99178234e-01 -3.40244681e-01 -7.49128684e-02
4.69650924e-01 -1.95520505e-01 -1.00094569e+00 2.29124188e-01
4.68528092e-01 5.70652962e-01 -6.45475686e-01 7.69230187e-01
-3.74307096e-01 -8.54226351e-02 -9.30170938e-02 -4.19545174e-01
5.25367498e-01 -9.82415378e-02 3.72096628e-01 1.53647745e+00
5.61650246e-02 1.36132002e-01 2.79795498e-01 8.08113754e-01
-1.00479849e-01 2.58599102e-01 -8.85061443e-01 2.74005868e-02
5.70396781e-01 1.26489770e+00 -2.51542270e-01 -4.26104128e-01
-7.25927830e-01 1.37085879e+00 4.17545766e-01 4.03389931e-01
-8.18802655e-01 -2.88497478e-01 4.63495344e-01 6.16977066e-02
-2.09700435e-01 -2.56812543e-01 2.59187818e-02 -1.14883292e+00
-3.22338045e-02 -8.71840775e-01 4.33653444e-01 -1.43101501e+00
-1.22654915e+00 5.79126239e-01 -6.56515509e-02 -1.02459037e+00
-5.19242585e-01 -4.77898329e-01 -4.60237890e-01 9.52493310e-01
-1.19228923e+00 -1.68995821e+00 -5.20139992e-01 9.40954745e-01
1.07125652e+00 -1.28593013e-01 8.31771910e-01 4.29992303e-02
-4.26488131e-01 6.38721585e-01 -4.17121112e-01 5.84955692e-01
9.35194969e-01 -1.33947229e+00 5.37289679e-01 5.47511637e-01
4.67429489e-01 8.27998221e-01 4.71384794e-01 -3.84886384e-01
-1.48017931e+00 -1.03694272e+00 1.27846897e+00 -8.20822418e-01
5.50474524e-01 -5.75589120e-01 -7.13082969e-01 7.93821573e-01
7.78268993e-01 -5.13332933e-02 5.53951025e-01 -1.32613733e-01
-3.09755564e-01 1.06365435e-01 -9.21020031e-01 1.00343215e+00
8.53951871e-01 -8.14699352e-01 -6.17926240e-01 6.34836495e-01
8.87163758e-01 -7.32234180e-01 -5.33912659e-01 -9.06935893e-03
3.72620195e-01 -7.63169467e-01 1.31299531e+00 -6.42564595e-01
4.60459322e-01 -3.36755723e-01 -2.27550149e-01 -9.94158387e-01
-2.31103450e-01 -7.26656854e-01 -1.09556668e-01 1.34198380e+00
3.50744933e-01 -2.69463956e-01 2.38728300e-01 7.15536594e-01
-1.11053526e-01 -5.37565410e-01 -7.58428276e-01 -1.29203215e-01
-2.36381322e-01 -2.83364773e-01 4.87876207e-01 1.00430238e+00
1.72903061e-01 9.18652952e-01 -6.30559862e-01 3.91166925e-01
2.96611965e-01 3.10172260e-01 1.23060513e+00 -7.67832696e-01
-1.09873610e-02 -4.66511697e-01 -7.97635168e-02 -1.65426683e+00
3.67304645e-02 -7.63945043e-01 4.82854135e-02 -2.20430851e+00
2.23371968e-01 -1.46866903e-01 2.58835137e-01 7.43125856e-01
5.35297692e-02 3.60874057e-01 4.26400423e-01 3.32388967e-01
-1.28907979e+00 4.42048162e-01 1.37632525e+00 -4.52948004e-01
-2.97610521e-01 -9.11085755e-02 -7.10454643e-01 8.32546055e-01
8.47092211e-01 3.33577126e-01 -5.24199247e-01 -8.60746324e-01
2.39426479e-01 4.31408435e-01 8.86989951e-01 -6.26912713e-01
6.49157763e-01 -4.29433823e-01 4.93649900e-01 -9.34188962e-01
7.18078673e-01 -6.01132810e-01 -2.28080198e-01 3.33912298e-02
-8.49778950e-01 6.76686704e-01 2.93798745e-02 2.75798351e-01
-2.63980865e-01 6.37761578e-02 5.48729181e-01 -4.64194715e-01
-1.09296072e+00 3.93852815e-02 -1.79974660e-01 7.48986891e-03
1.00641596e+00 -3.12830247e-02 -5.70120633e-01 -9.89331484e-01
-4.57701474e-01 7.84832478e-01 5.17986596e-01 6.29713714e-01
1.08800960e+00 -1.34021997e+00 -7.24321783e-01 -3.06339040e-02
4.64626580e-01 2.68934131e-01 3.78720820e-01 8.32672775e-01
-5.48686206e-01 7.87254393e-01 7.54564852e-02 -6.86779797e-01
-1.42737067e+00 6.73347533e-01 4.48493928e-01 2.10683718e-01
-8.04121315e-01 1.07293725e+00 4.91193920e-01 -7.81362116e-01
4.46017385e-01 -9.67001617e-02 -4.34199303e-01 7.11372644e-02
9.01233256e-01 -3.17919930e-03 -5.33855200e-01 -1.17850399e+00
-5.11872411e-01 6.30640090e-01 5.35592921e-02 -2.46132329e-01
8.02192450e-01 -7.21756339e-01 -3.56619179e-01 7.64073372e-01
9.48626399e-01 6.86512813e-02 -9.46169078e-01 -3.94321531e-01
-2.54456431e-01 -2.84511447e-01 -4.78774190e-01 -9.00784910e-01
-5.97801805e-01 1.16740477e+00 2.70275116e-01 1.65098682e-01
8.15378308e-01 6.27465606e-01 6.44633889e-01 6.16102338e-01
1.66024417e-01 -8.09224069e-01 4.86763000e-01 5.94141126e-01
1.40664172e+00 -1.42783141e+00 -3.88594925e-01 -6.30150437e-02
-1.25255084e+00 9.79008853e-01 1.01963067e+00 3.72623503e-01
-1.45393133e-01 -3.73519235e-03 5.77113330e-01 -3.42714041e-01
-9.39690351e-01 -2.58902311e-01 3.58804107e-01 5.80970645e-01
6.26286447e-01 5.11115603e-02 7.48085082e-02 3.14691365e-01
-3.36817443e-01 -5.35336971e-01 3.73214602e-01 4.08071935e-01
-3.77634227e-01 -6.13471091e-01 -3.91615599e-01 8.27469081e-02
1.52454888e-02 -1.93008751e-01 -9.32356000e-01 6.67910933e-01
-2.37867519e-01 1.02907801e+00 3.88133198e-01 -1.26292944e-01
3.19333076e-01 -1.28479004e-01 1.51458561e-01 -6.29312158e-01
-5.63339949e-01 1.05341552e-02 1.46983951e-01 -5.79542577e-01
-4.59326833e-01 -4.04733360e-01 -1.35698581e+00 -5.67294210e-02
8.29086304e-02 1.76120743e-01 2.93314219e-01 6.39173388e-01
2.45249897e-01 3.08420926e-01 1.90139264e-01 -6.98965013e-01
-2.38580063e-01 -1.08688664e+00 3.16829607e-02 4.48234111e-01
6.17941678e-01 -1.23519257e-01 2.07393020e-01 -4.48384024e-02] | [10.868599891662598, 1.489579439163208] |
c63387ad-7451-477d-9c8e-67e39fe1970f | referring-to-screen-texts-with-voice | 2306.07298 | null | https://arxiv.org/abs/2306.07298v1 | https://arxiv.org/pdf/2306.07298v1.pdf | Referring to Screen Texts with Voice Assistants | Voice assistants help users make phone calls, send messages, create events, navigate, and do a lot more. However, assistants have limited capacity to understand their users' context. In this work, we aim to take a step in this direction. Our work dives into a new experience for users to refer to phone numbers, addresses, email addresses, URLs, and dates on their phone screens. Our focus lies in reference understanding, which becomes particularly interesting when multiple similar texts are present on screen, similar to visual grounding. We collect a dataset and propose a lightweight general-purpose model for this novel experience. Due to the high cost of consuming pixels directly, our system is designed to rely on the extracted text from the UI. Our model is modular, thus offering flexibility, improved interpretability, and efficient runtime memory utilization. | ['Vincent Renkens', 'Hong Yu', 'Alkesh Patel', 'Hoang Long Nguyen', 'Ing-Marie Jonsson', 'Anand Dhoot', 'Shruti Bhargava'] | 2023-06-10 | null | null | null | null | ['visual-grounding', 'navigate'] | ['computer-vision', 'reasoning'] | [ 4.07730728e-01 8.97530094e-02 -1.92874730e-01 -2.92775214e-01
-3.73616070e-01 -8.02011967e-01 6.57986701e-01 3.72075826e-01
-2.92181432e-01 4.30034399e-01 6.63249910e-01 -5.91682673e-01
1.41946925e-02 -6.67705059e-01 -9.22285095e-02 6.99069649e-02
2.98647881e-01 2.48114944e-01 2.15179175e-01 -5.64059205e-02
4.71352905e-01 5.10609031e-01 -1.66356552e+00 4.64834839e-01
6.83691382e-01 6.30953670e-01 4.19003218e-01 7.52721250e-01
-5.19215286e-01 6.62576795e-01 -5.99469483e-01 -5.11994839e-01
-4.40160662e-01 -2.65357882e-01 -9.06019747e-01 3.16364840e-02
4.83523279e-01 -6.15165472e-01 -1.29680812e-01 7.53262639e-01
4.38948214e-01 2.70571738e-01 2.88827866e-01 -1.26507318e+00
-4.13153589e-01 6.26561403e-01 -3.85082781e-01 -1.07151963e-01
1.01109648e+00 -1.17377914e-01 1.16968822e+00 -7.01436460e-01
5.19063115e-01 1.12876463e+00 7.25320995e-01 5.09057224e-01
-1.04950845e+00 -8.30933228e-02 6.32049382e-01 6.10885443e-03
-1.12888229e+00 -6.62482917e-01 5.49758375e-01 -2.57435083e-01
9.71465588e-01 9.23204541e-01 4.31074053e-01 1.12055516e+00
-5.94119012e-01 1.00319397e+00 6.52884960e-01 -7.04449117e-01
2.24868342e-01 5.30003667e-01 3.87166798e-01 6.65837407e-01
2.49131583e-02 -7.05156446e-01 -5.76672018e-01 -3.04513246e-01
7.71371722e-01 4.69285697e-01 -6.37122273e-01 -2.40335301e-01
-1.08562648e+00 1.95928663e-01 2.08852217e-01 4.73143160e-01
-2.15159416e-01 -1.35797635e-01 1.62053540e-01 2.34197434e-02
-4.50035036e-02 3.61002088e-01 -3.23945940e-01 -9.33357716e-01
-5.72198868e-01 -2.95972507e-02 1.20027530e+00 1.17440343e+00
6.98870659e-01 -5.89173734e-01 3.45913954e-02 9.33525801e-01
2.75139272e-01 4.39195633e-01 2.67459989e-01 -9.44807172e-01
5.79532206e-01 8.87360275e-01 5.31997144e-01 -1.09426868e+00
-5.53828418e-01 7.44347367e-03 -4.53632593e-01 -2.03908645e-02
5.22425056e-01 -8.06288868e-02 -3.79225343e-01 1.38369560e+00
8.11862573e-02 -8.06459337e-02 -3.26275140e-01 6.44340396e-01
8.50728869e-01 3.07633579e-01 1.25995949e-01 4.48150933e-02
1.66994548e+00 -6.16448224e-01 -8.51763606e-01 -3.56287450e-01
7.88851798e-01 -8.15737426e-01 1.65621769e+00 1.28809348e-01
-8.64948452e-01 -5.04959345e-01 -6.61449552e-01 -3.31265211e-01
-6.49734080e-01 4.85940695e-01 7.67224789e-01 8.31029415e-01
-7.74014413e-01 5.95013916e-01 -8.08016360e-01 -9.52659249e-01
1.69923738e-01 1.51188031e-01 -1.05643116e-01 3.04851532e-01
-7.05558658e-01 6.12888694e-01 -9.78408977e-02 -1.25230625e-01
3.71691167e-01 -6.19100928e-01 -7.43769348e-01 3.21459919e-01
4.39197481e-01 -6.50234818e-01 1.53849280e+00 -6.29103184e-01
-1.39442503e+00 7.97208667e-01 -5.83455384e-01 1.74303073e-02
5.61126173e-01 -4.83572155e-01 -3.32165599e-01 -6.17534667e-02
-7.62966722e-02 3.42456877e-01 4.38479573e-01 -1.17710435e+00
-8.59525025e-01 -5.21288991e-01 4.97463048e-01 2.70000547e-01
-6.98019147e-01 1.33501559e-01 -8.60387444e-01 -4.05653596e-01
1.32035986e-02 -9.41989899e-01 9.63796005e-02 3.62648070e-02
-3.47918659e-01 -9.46454611e-03 9.74715829e-01 -7.87315190e-01
1.95243871e+00 -2.32015395e+00 -3.75202596e-02 2.54005313e-01
3.18903327e-01 1.73820153e-01 3.67364645e-01 7.39000022e-01
3.99747759e-01 3.14099133e-01 4.66688722e-02 -5.81265867e-01
2.70778120e-01 8.10846165e-02 -4.45532620e-01 -2.21943304e-01
-2.52683103e-01 7.86594927e-01 -9.39000249e-01 -5.59396625e-01
3.91778857e-01 6.17690921e-01 -4.40100044e-01 7.34354630e-02
-5.48774190e-02 1.40200615e-01 -5.63705146e-01 5.74546516e-01
4.21737939e-01 -3.86778772e-01 4.53634530e-01 -2.10985824e-01
-2.55833566e-01 6.04536176e-01 -1.44842815e+00 1.49676490e+00
-9.06051099e-01 8.16557288e-01 2.03193963e-01 -2.94026911e-01
5.17563343e-01 1.49182960e-01 8.12251121e-02 -4.42623079e-01
-6.71837032e-02 2.36462299e-02 -5.01994908e-01 -5.08998573e-01
7.97784328e-01 6.94531500e-01 7.94928521e-02 8.12009633e-01
-8.04428399e-01 1.19546019e-01 1.82952181e-01 3.66452992e-01
1.08434987e+00 2.89891541e-01 5.00606954e-01 2.03832418e-01
5.83707929e-01 -3.50396037e-01 -1.24932036e-01 9.18658912e-01
9.26427022e-02 5.68972409e-01 4.37081188e-01 -2.72128075e-01
-6.35732472e-01 -9.45718765e-01 1.68244749e-01 1.31009269e+00
5.70360385e-03 -9.25371051e-01 -8.14941943e-01 -6.90393746e-01
-1.21966667e-01 8.24684620e-01 -1.88225701e-01 2.82297134e-01
-6.66815281e-01 -1.34701774e-01 1.91693395e-01 8.80183339e-01
4.93240505e-01 -9.75469589e-01 -9.61410701e-01 8.94930363e-02
-4.28852350e-01 -9.97251630e-01 -6.69780076e-01 -1.91032618e-01
-1.04243231e+00 -8.52545798e-01 -6.87599301e-01 -7.05429494e-01
6.47866845e-01 5.80017984e-01 1.10902309e+00 3.23427469e-01
-2.66186625e-01 8.27416718e-01 -2.49096185e-01 -4.15952355e-01
9.74434465e-02 3.92055750e-01 -2.34293759e-01 4.27540997e-03
6.51469588e-01 -6.39201820e-01 -6.47822142e-01 5.71886897e-01
-7.99794495e-01 2.13403955e-01 1.49925873e-01 3.43906939e-01
9.51589495e-02 -2.17694119e-01 1.73585445e-01 -1.10719311e+00
7.86942005e-01 -1.60217196e-01 -2.47388065e-01 3.29721510e-01
-3.22212011e-01 6.02529682e-02 6.26549244e-01 -2.36964226e-01
-1.24084759e+00 1.42126560e-01 -7.06816837e-02 2.98187107e-01
-4.54310954e-01 2.30391636e-01 -5.71761847e-01 2.05951691e-01
5.96732140e-01 -1.61128715e-02 -3.29306185e-01 -8.98467481e-01
6.66663289e-01 1.25725770e+00 5.62737584e-01 -5.32128572e-01
5.17087758e-01 6.30403996e-01 -4.83397067e-01 -1.15241826e+00
-4.94265229e-01 -8.05827022e-01 -6.68287754e-01 -1.15098655e-01
3.94089967e-01 -4.16328490e-01 -1.21672344e+00 7.51989260e-02
-1.17925370e+00 -3.32953542e-01 -3.55090611e-02 1.55494258e-01
-3.49071980e-01 5.42789996e-01 -3.72119606e-01 -1.05012047e+00
-1.22237638e-01 -7.01023936e-01 1.13612401e+00 4.60730016e-01
-9.49414194e-01 -1.08164978e+00 -4.07004118e-01 2.84056395e-01
5.32049894e-01 -1.18291445e-01 8.19187880e-01 -4.65054572e-01
-6.73000455e-01 -3.14021260e-01 -5.43599963e-01 -2.94442415e-01
4.98484075e-01 2.27036513e-02 -1.06680584e+00 1.33869369e-02
-3.73918205e-01 1.87334687e-01 5.47567248e-01 9.12752971e-02
1.40477514e+00 -3.85141104e-01 -8.10701549e-01 4.16001260e-01
9.35166657e-01 3.42953116e-01 6.88653469e-01 4.69708174e-01
8.57825339e-01 6.55513406e-01 3.75667006e-01 6.87434077e-01
4.63952959e-01 1.01747835e+00 1.53543726e-01 -6.51716813e-02
-2.19797809e-02 -4.24462467e-01 4.40434143e-02 3.14797491e-01
-2.96726078e-01 -4.54521120e-01 -7.76618600e-01 4.25606042e-01
-1.99227667e+00 -9.44971561e-01 -2.63590723e-01 2.31419325e+00
5.56714535e-01 1.16287619e-01 9.76568386e-02 1.83691949e-01
5.63850701e-01 6.79957047e-02 -2.72660494e-01 -2.48938382e-01
3.54054391e-01 -3.84172648e-02 2.29146138e-01 6.82461143e-01
-9.26100612e-01 7.33746827e-01 6.05086899e+00 3.09128821e-01
-8.79049957e-01 -3.15489709e-01 3.67997855e-01 -1.49107322e-01
-3.74980748e-01 -4.94807065e-02 -1.00039434e+00 4.57135975e-01
6.48897290e-01 2.54801691e-01 6.35679901e-01 7.73512185e-01
2.81614810e-01 -2.79250234e-01 -1.41793787e+00 1.31374156e+00
-2.71511395e-02 -1.13872075e+00 -9.83368978e-02 6.63552135e-02
1.50580451e-01 -5.45886457e-01 6.58760220e-02 1.97311807e-02
-1.57583132e-02 -7.61479795e-01 4.38251257e-01 6.51034832e-01
6.87567830e-01 -5.15862346e-01 3.68500561e-01 1.03525721e-01
-1.18795776e+00 3.06393709e-02 1.80923343e-01 -4.28629547e-01
2.11322755e-01 2.81439990e-01 -9.38992620e-01 5.45264548e-03
7.55508959e-01 4.97780740e-01 -6.67544603e-01 8.14843059e-01
-1.08942486e-01 2.33460248e-01 -3.97269577e-01 -4.16700989e-01
-1.83838099e-01 1.11776562e-02 2.65681505e-01 1.35051024e+00
4.00723875e-01 2.06922069e-02 7.26237148e-02 5.65021217e-01
-1.66916296e-01 2.12164611e-01 -6.78046346e-01 -2.68580973e-01
6.74521744e-01 1.45807707e+00 -9.68129277e-01 -4.55318183e-01
-6.45569682e-01 1.30214131e+00 2.02833399e-01 4.43981797e-01
-6.83728337e-01 -9.41704631e-01 7.88300991e-01 4.03836101e-01
6.74672127e-02 -4.42115366e-01 -5.12193978e-01 -1.01951909e+00
2.97300398e-01 -8.77800345e-01 2.03142732e-01 -9.53200400e-01
-6.53858602e-01 3.41795325e-01 -2.23875359e-01 -1.10416317e+00
-3.67777109e-01 -6.64764881e-01 -5.55770993e-01 9.35896754e-01
-1.06957042e+00 -8.64355385e-01 -8.46449375e-01 3.89625639e-01
4.60774750e-01 2.11731657e-01 1.02838075e+00 4.13875669e-01
-5.32301128e-01 5.41700006e-01 1.66426867e-01 2.39050597e-01
8.52641761e-01 -1.37225091e+00 7.98834145e-01 4.71528739e-01
3.29406619e-01 1.38858962e+00 5.57937503e-01 -5.70472240e-01
-1.35075545e+00 -4.13265407e-01 1.32628429e+00 -9.32378888e-01
5.92798591e-01 -4.47561443e-01 -7.96954930e-01 8.02703321e-01
1.22834243e-01 -6.48532033e-01 9.90161240e-01 7.30025411e-01
-2.19946772e-01 2.46527139e-02 -7.40890384e-01 1.22219110e+00
1.44876480e+00 -8.31854880e-01 -5.74741364e-01 1.43409535e-01
5.11000991e-01 -5.06280363e-01 -5.07347107e-01 -4.12520766e-01
8.41656089e-01 -9.08089399e-01 1.07366359e+00 -4.42156643e-01
1.42145321e-01 -2.96443939e-01 1.27375886e-01 -1.02190685e+00
-1.50045892e-02 -1.10187900e+00 -1.89483523e-01 1.64125574e+00
3.83489788e-01 -6.95025504e-01 7.54245043e-01 1.21875119e+00
1.23495543e-02 -1.70203626e-01 -3.18272084e-01 -2.62058854e-01
-1.08738995e+00 -7.33978033e-01 1.01486182e+00 7.32990086e-01
5.50523043e-01 4.32848930e-01 -2.90546060e-01 3.77678126e-02
2.64376163e-01 1.45477712e-01 9.92228389e-01 -1.48027253e+00
-3.28717381e-01 -5.11020541e-01 -4.32582758e-02 -1.60604239e+00
-3.88872504e-01 -4.74177212e-01 -3.27247769e-01 -1.74159861e+00
-4.09274995e-02 -5.21995664e-01 1.12057224e-01 6.27301872e-01
-1.51600599e-01 1.05491206e-01 3.55649799e-01 1.47638321e-01
-7.88906753e-01 -1.42361075e-01 7.44053006e-01 4.43189032e-02
-7.74186254e-01 3.84325922e-01 -1.01583123e+00 1.03324676e+00
6.99623227e-01 9.64389071e-02 -3.65594357e-01 -4.58748043e-01
4.26886976e-01 -1.78576112e-01 3.34174722e-01 -9.40721512e-01
3.45730841e-01 -9.75540504e-02 3.59100252e-01 -5.21192849e-01
4.08348888e-01 -1.11199951e+00 -9.92079675e-02 1.25904575e-01
-3.92272145e-01 -3.92803475e-02 1.65197998e-01 4.76954877e-01
7.84237534e-02 -4.59364384e-01 9.46891159e-02 -3.18126887e-01
-7.08485544e-01 -2.43251622e-01 -5.63637674e-01 -1.97416529e-01
4.58557457e-01 -4.94379669e-01 -3.36907327e-01 -6.85806990e-01
-7.68885493e-01 9.13855210e-02 7.09305108e-01 5.99813342e-01
4.50014383e-01 -1.14020932e+00 1.75806135e-01 1.56908959e-01
2.73240477e-01 -2.75002033e-01 1.29096195e-01 4.52451348e-01
-4.04972792e-01 6.43980026e-01 1.18986621e-01 -3.36534351e-01
-1.70103776e+00 3.26337606e-01 -6.85225576e-02 1.33706450e-01
-6.39499664e-01 5.64318240e-01 -2.46915426e-02 -2.26479217e-01
7.15661108e-01 -6.13350511e-01 -4.53545332e-01 2.19204813e-01
1.18049979e+00 7.33329952e-01 2.67912783e-02 -2.50661045e-01
-3.28985721e-01 5.97774267e-01 -4.98830713e-02 -7.63127729e-02
1.01564467e+00 -5.65767705e-01 1.44301709e-02 5.31671524e-01
9.75187957e-01 5.59125245e-01 -1.13272071e+00 -2.66387820e-01
3.34779650e-01 -6.73398495e-01 -1.45603955e-01 -8.30979645e-01
-4.37342793e-01 8.83487701e-01 3.99570405e-01 5.88926613e-01
1.06040967e+00 4.48746532e-02 8.87965262e-01 6.92540407e-01
2.83457696e-01 -1.04902864e+00 -7.83162490e-02 4.83549058e-01
5.62806308e-01 -1.25632310e+00 -3.71320918e-02 -4.73693430e-01
-6.30097628e-01 1.31504393e+00 5.06330669e-01 8.17350507e-01
2.21348673e-01 2.82111406e-01 3.82217467e-02 -3.70196663e-02
-3.93636078e-01 -3.83278757e-01 2.41716340e-01 8.42729926e-01
8.24623942e-01 -2.40324456e-02 -6.18677139e-02 4.76420850e-01
-2.48659343e-01 5.28019946e-03 4.61054116e-01 1.08183026e+00
-5.42722166e-01 -1.25142157e+00 -3.94812852e-01 3.44381660e-01
-3.31715941e-01 -2.59413123e-01 -6.96549952e-01 4.63128000e-01
-3.57138455e-01 1.08123362e+00 1.14224114e-01 -1.70118332e-01
5.62266231e-01 3.45385134e-01 2.08729133e-01 -5.76887250e-01
-5.40482819e-01 -2.68539131e-01 3.11033934e-01 -5.49758971e-01
-2.26770520e-01 -6.31308138e-01 -1.30485570e+00 -2.55125463e-01
-2.82652378e-02 2.21528094e-02 8.04658413e-01 6.33998990e-01
6.36001289e-01 4.14218903e-01 2.58268446e-01 -7.38015592e-01
-5.74345514e-02 -7.77552366e-01 -1.57379657e-01 4.51199025e-01
2.12787896e-01 -3.53921086e-01 -1.62270859e-01 2.21637458e-01] | [11.794472694396973, 2.7496795654296875] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.