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
|
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
45062238-81d6-42a5-b1f4-5eefcccc9f2a
|
an-effective-loss-function-for-generating-3d
|
2103.03390
| null |
https://arxiv.org/abs/2103.03390v2
|
https://arxiv.org/pdf/2103.03390v2.pdf
|
An Effective Loss Function for Generating 3D Models from Single 2D Image without Rendering
|
Differentiable rendering is a very successful technique that applies to a Single-View 3D Reconstruction. Current renderers use losses based on pixels between a rendered image of some 3D reconstructed object and ground-truth images from given matched viewpoints to optimise parameters of the 3D shape. These models require a rendering step, along with visibility handling and evaluation of the shading model. The main goal of this paper is to demonstrate that we can avoid these steps and still get reconstruction results as other state-of-the-art models that are equal or even better than existing category-specific reconstruction methods. First, we use the same CNN architecture for the prediction of a point cloud shape and pose prediction like the one used by Insafutdinov & Dosovitskiy. Secondly, we propose the novel effective loss function that evaluates how well the projections of reconstructed 3D point clouds cover the ground truth object's silhouette. Then we use Poisson Surface Reconstruction to transform the reconstructed point cloud into a 3D mesh. Finally, we perform a GAN-based texture mapping on a particular 3D mesh and produce a textured 3D mesh from a single 2D image. We evaluate our method on different datasets (including ShapeNet, CUB-200-2011, and Pascal3D+) and achieve state-of-the-art results, outperforming all the other supervised and unsupervised methods and 3D representations, all in terms of performance, accuracy, and training time.
|
['Pietro Liò', 'Nikola Zubić']
|
2021-03-05
| null | null | null | null |
['single-view-3d-reconstruction']
|
['computer-vision']
|
[ 1.14191838e-01 2.15578303e-01 4.13439542e-01 -3.83847266e-01
-6.98254228e-01 -3.22025567e-01 8.24758112e-01 -1.34843156e-01
-9.86910760e-02 3.65760356e-01 -3.04121822e-01 -1.14973888e-01
1.79684356e-01 -1.17579925e+00 -1.11762559e+00 -5.00282288e-01
2.02085942e-01 1.23813069e+00 4.12544668e-01 -1.23278968e-01
2.26163208e-01 1.07858622e+00 -1.78884339e+00 1.07397646e-01
7.41310775e-01 1.25638676e+00 1.41458243e-01 3.68141979e-01
-4.81086463e-01 1.75495014e-01 -3.68058890e-01 -4.78843123e-01
5.39531231e-01 -4.61728610e-02 -5.67023516e-01 3.46140981e-01
7.21928656e-01 -3.16590816e-01 1.11891240e-01 8.00618589e-01
4.61817890e-01 -7.00463578e-02 8.70548964e-01 -7.88449168e-01
-3.37832808e-01 -3.73121649e-01 -5.91730595e-01 -5.57202041e-01
4.83762085e-01 4.15349081e-02 5.04785955e-01 -1.17808032e+00
8.64634991e-01 1.35576868e+00 9.13195729e-01 4.66692150e-01
-1.44723129e+00 -5.64476669e-01 -1.39030544e-02 -2.11535186e-01
-1.27774739e+00 -6.87609911e-02 9.40969169e-01 -6.00227475e-01
1.07279861e+00 3.28782767e-01 9.15227711e-01 8.42843473e-01
1.73232034e-01 4.21336085e-01 1.24808967e+00 -4.65841293e-01
2.41433337e-01 7.76766986e-02 -3.65469605e-01 7.61316180e-01
-1.81561828e-01 2.65080601e-01 -2.75514752e-01 -1.99478373e-01
1.09316909e+00 3.23633961e-02 -3.64336312e-01 -8.04846942e-01
-9.80264783e-01 6.37837648e-01 6.44745886e-01 2.63850316e-02
-5.39098084e-01 1.75056130e-01 4.75295670e-02 1.16647080e-01
1.20328999e+00 5.51201105e-02 -2.41382271e-01 1.43659756e-01
-8.96638989e-01 3.23467433e-01 7.37854064e-01 8.60210061e-01
1.04674721e+00 1.85767457e-01 2.29646668e-01 6.50459051e-01
4.41650301e-01 7.64829099e-01 -1.51834711e-01 -1.08407164e+00
2.02302620e-01 7.51904964e-01 9.52710360e-02 -9.01254654e-01
-1.65026173e-01 -4.50891167e-01 -7.56400168e-01 9.40676510e-01
3.10320631e-02 5.33322990e-01 -1.12264729e+00 1.19734120e+00
6.96552932e-01 3.63618135e-01 -1.17876448e-01 9.37580585e-01
9.43434954e-01 6.43764555e-01 -4.12979335e-01 1.77937225e-01
9.28572178e-01 -6.52758598e-01 -2.11824730e-01 -4.99273371e-03
1.95606127e-01 -8.46819162e-01 1.05216348e+00 5.78338623e-01
-1.30231774e+00 -5.77083111e-01 -8.77065063e-01 -1.59019873e-01
-2.50837137e-03 -6.67208210e-02 3.68058085e-01 5.04244208e-01
-1.03202176e+00 9.13399279e-01 -7.43178308e-01 -1.17712058e-01
6.82289422e-01 6.02984652e-02 -4.13342476e-01 -1.22503169e-01
-5.04867971e-01 9.94841516e-01 4.46513854e-03 -6.10546535e-03
-1.01361716e+00 -1.07222342e+00 -7.20981359e-01 -1.75271556e-01
-6.70017228e-02 -1.08390188e+00 9.90625799e-01 -7.83949554e-01
-1.55214989e+00 1.53160048e+00 7.10070878e-02 -3.39751750e-01
8.46643209e-01 -1.01426803e-01 1.75380751e-01 5.35314605e-02
-5.20057231e-02 6.15488708e-01 8.53031933e-01 -1.90558338e+00
-3.15008193e-01 -4.90804285e-01 4.62188013e-03 4.05758321e-01
4.96450245e-01 -3.68254274e-01 -5.21016359e-01 -2.99563557e-01
4.55876291e-01 -6.79442942e-01 -2.04606608e-01 4.42773402e-01
-3.91441256e-01 -4.24050540e-02 7.45371878e-01 -7.48125315e-01
8.41197595e-02 -2.01776767e+00 3.40877205e-01 2.11000264e-01
1.00946084e-01 8.85147229e-02 1.63435582e-02 1.35670111e-01
1.58121681e-03 1.11532353e-01 -4.96670306e-01 -1.04655790e+00
1.19753554e-02 2.96072990e-01 -4.12171841e-01 7.14488447e-01
1.24210753e-01 7.22289205e-01 -4.86587614e-01 -1.45579576e-01
8.13432097e-01 1.03289700e+00 -5.36878765e-01 3.87085140e-01
-6.05815411e-01 9.04844046e-01 -2.06650034e-01 4.50114936e-01
1.09622049e+00 -1.62251666e-01 -2.94340491e-01 -1.54261500e-01
-1.77842960e-01 3.62353325e-01 -1.05385375e+00 1.91267121e+00
-8.10372829e-01 3.07868958e-01 6.25301618e-03 -7.34473467e-01
1.29990220e+00 1.27412945e-01 4.70779419e-01 -7.42538452e-01
7.05573931e-02 3.92495066e-01 -6.64311528e-01 -5.87462112e-02
2.41081744e-01 -4.08816129e-01 4.56834435e-01 2.30171666e-01
-9.39668640e-02 -1.01799858e+00 -6.95340931e-01 -2.35242650e-01
5.97596347e-01 6.90826595e-01 -1.18081294e-01 -9.09067020e-02
3.90224189e-01 1.25669893e-02 1.67303711e-01 2.91843146e-01
6.63502991e-01 1.31435394e+00 2.58693516e-01 -7.15726376e-01
-1.39683998e+00 -1.14935637e+00 -3.76723617e-01 1.49095848e-01
2.79350430e-01 -5.36150634e-02 -6.90257370e-01 -4.97474104e-01
1.04144119e-01 9.62231934e-01 -5.90144038e-01 1.95128173e-01
-6.49614811e-01 -2.40263000e-01 2.59720497e-02 1.95772737e-01
5.05206943e-01 -9.81501102e-01 -7.11368859e-01 -7.37500638e-02
7.06506222e-02 -1.09717119e+00 1.20805666e-01 -3.08151077e-02
-1.13091588e+00 -1.05658913e+00 -7.20135093e-01 -5.02281070e-01
8.03469360e-01 9.69351083e-03 1.59980249e+00 2.71130294e-01
-7.75619447e-02 4.61934865e-01 -1.33612618e-01 -4.66221660e-01
-5.32470524e-01 -3.01164150e-01 -2.90293187e-01 4.92592715e-03
-2.60311574e-01 -8.50048721e-01 -5.11019230e-01 3.87638271e-01
-8.24710548e-01 3.67231697e-01 1.49004117e-01 4.74583477e-01
1.29212058e+00 -9.84264240e-02 -1.81317911e-01 -8.59394670e-01
-1.75691452e-02 -7.48048127e-02 -9.25081491e-01 -4.30490300e-02
-4.91601020e-01 -2.31216475e-01 3.61864835e-01 -1.84310660e-01
-9.29196775e-01 1.84341982e-01 -6.18261456e-01 -1.17132282e+00
-2.61854947e-01 6.07004501e-02 -9.19179097e-02 -2.62954623e-01
6.41133428e-01 3.00936610e-01 3.25379707e-02 -8.27811003e-01
1.63902864e-01 1.65751815e-01 2.62970537e-01 -4.87947345e-01
9.48295832e-01 1.04925752e+00 3.30388606e-01 -7.47934580e-01
-7.44608581e-01 -2.38060892e-01 -6.81492627e-01 -4.59983408e-01
9.46143031e-01 -8.37114155e-01 -6.49099827e-01 5.37571669e-01
-1.58636725e+00 -5.79698086e-01 -6.04154885e-01 2.85054684e-01
-9.21324134e-01 2.61580884e-01 -2.95743078e-01 -8.35866332e-01
-3.07240725e-01 -1.23719573e+00 1.76403439e+00 -2.32024685e-01
3.55662107e-01 -9.66675222e-01 1.33416995e-01 3.94068062e-01
2.44740069e-01 6.87933981e-01 8.08149219e-01 1.39261320e-01
-1.01433957e+00 -1.31740868e-01 -1.78820729e-01 5.63416123e-01
-1.38138652e-01 -6.94896057e-02 -1.29353583e+00 -1.14078574e-01
4.12542373e-01 -1.42325133e-01 7.02277601e-01 4.58121687e-01
1.29949272e+00 -4.33289558e-02 -1.07148230e-01 1.10167003e+00
1.66221201e+00 -1.46462873e-01 9.59474385e-01 1.88236728e-01
9.03884590e-01 6.06328368e-01 5.24215519e-01 2.18077719e-01
3.57295543e-01 9.61055696e-01 1.20860910e+00 -1.76768124e-01
-3.75959933e-01 -4.51992482e-01 -8.36420283e-02 7.48379290e-01
-4.48154211e-01 -1.78020447e-01 -8.52321625e-01 2.83027202e-01
-1.52483070e+00 -4.74776924e-01 -5.42029083e-01 2.53959751e+00
3.58876228e-01 8.66952017e-02 -1.55553266e-01 6.79744333e-02
3.64859879e-01 1.18637070e-01 -4.16690171e-01 -3.55602235e-01
-1.46265984e-01 5.24844706e-01 3.32220614e-01 6.88728690e-01
-7.87827075e-01 9.46369171e-01 5.41297865e+00 8.23195279e-01
-1.28401852e+00 1.68977335e-01 6.99057877e-01 9.39618051e-02
-7.27195561e-01 -5.50046451e-02 -5.66779912e-01 1.80630416e-01
6.04811788e-01 3.89210045e-01 4.64755565e-01 7.58727849e-01
-4.12346832e-02 -7.41993040e-02 -1.18871725e+00 1.22599125e+00
2.25087985e-01 -1.56662095e+00 1.73349932e-01 2.45995417e-01
8.30493629e-01 2.57146478e-01 -2.58149020e-02 1.97167378e-02
7.45914206e-02 -1.24202204e+00 1.06988204e+00 8.77546787e-01
9.88759696e-01 -6.16054952e-01 5.61348498e-01 7.22073317e-01
-8.78587842e-01 6.42539263e-01 -4.79240507e-01 1.76452994e-01
3.05824548e-01 8.90606344e-01 -7.66972363e-01 9.31952059e-01
9.27145422e-01 5.83693266e-01 -2.76041180e-01 1.02762842e+00
-3.38693976e-01 1.96270600e-01 -5.24648070e-01 3.46953124e-01
7.24221915e-02 -5.04227102e-01 7.75565863e-01 6.56388879e-01
4.88687247e-01 -4.97631580e-02 2.20903996e-02 1.29966795e+00
-8.08658302e-02 8.04169402e-02 -7.60935962e-01 6.91580832e-01
7.15036318e-02 1.01206172e+00 -7.04663336e-01 -1.76576793e-01
-9.37836915e-02 9.18358624e-01 3.63456100e-01 1.00064181e-01
-6.50955379e-01 3.22553366e-01 4.11160618e-01 6.25199020e-01
4.42135483e-01 -1.97570190e-01 -4.80777919e-01 -9.85962629e-01
3.15476179e-01 -5.01066267e-01 -2.85801113e-01 -1.29203212e+00
-1.35290265e+00 8.54206502e-01 -4.16508783e-03 -1.46923220e+00
-1.42508596e-01 -5.39194286e-01 -3.99093449e-01 1.16361725e+00
-1.67287838e+00 -1.15584528e+00 -5.10553539e-01 3.54561210e-01
4.67567235e-01 1.52318612e-01 9.22190845e-01 2.09201857e-01
1.72339410e-01 -4.89085093e-02 -1.57168955e-01 -2.68009365e-01
2.63577670e-01 -1.21041965e+00 7.39147305e-01 3.38726699e-01
1.84274122e-01 -8.30344036e-02 5.42767465e-01 -6.70857370e-01
-1.31549764e+00 -1.09269536e+00 6.04385436e-01 -7.32210100e-01
-2.82326154e-02 -4.81412649e-01 -1.04314911e+00 5.79674721e-01
-1.03517592e-01 2.21931353e-01 -1.14518031e-01 -2.79033631e-01
-2.48272091e-01 9.23630446e-02 -1.47215879e+00 3.77458870e-01
1.23540711e+00 -3.89733046e-01 -4.28779036e-01 4.39575255e-01
7.10972548e-01 -1.14574766e+00 -8.05757046e-01 6.14385068e-01
3.67302150e-01 -1.47047126e+00 1.19689596e+00 -3.14460278e-01
6.49016440e-01 -3.30720901e-01 -3.94060373e-01 -1.28983748e+00
2.11966708e-02 -3.20323676e-01 1.32504836e-01 8.29142094e-01
4.35588136e-02 -6.36814475e-01 8.27033699e-01 3.07974726e-01
-5.28528273e-01 -1.00710082e+00 -1.16257596e+00 -7.30999708e-01
1.95963353e-01 -6.94827378e-01 8.70362937e-01 8.87972116e-01
-1.12681985e+00 6.58441707e-02 -1.40841650e-02 2.79786527e-01
9.12759900e-01 4.45007265e-01 1.16495407e+00 -1.51181555e+00
-1.57277420e-01 -3.24936986e-01 -5.77453732e-01 -1.08765352e+00
2.94569910e-01 -1.03691649e+00 -9.83989611e-02 -1.78650641e+00
-2.79961705e-01 -8.90996516e-01 3.63812089e-01 8.60313699e-02
4.77974206e-01 5.53194702e-01 1.67385012e-01 3.24774057e-01
-6.79465532e-02 7.81276524e-01 1.51517415e+00 2.00016554e-02
-1.02974683e-01 1.55323699e-01 -7.33735114e-02 9.18721855e-01
4.96075839e-01 -4.82736677e-01 -1.30426064e-01 -7.97003090e-01
3.48224908e-01 1.73769444e-01 8.86249900e-01 -9.14288819e-01
-1.90850422e-01 8.98550004e-02 4.65352118e-01 -1.08592868e+00
9.82188404e-01 -1.12955558e+00 7.61717319e-01 2.13667870e-01
9.00654942e-02 -1.02681696e-01 2.01456636e-01 5.16752362e-01
1.65577158e-01 -2.22413883e-01 9.96253073e-01 -2.67814964e-01
-1.72815546e-01 6.27213538e-01 4.16217804e-01 -1.94778040e-01
8.31090868e-01 -4.99100178e-01 4.37228307e-02 -2.27122039e-01
-6.29890263e-01 -2.48126805e-01 1.10791636e+00 1.82262406e-01
9.32807565e-01 -1.44512165e+00 -8.90180111e-01 4.82979894e-01
-1.34526296e-02 8.27770293e-01 2.55908281e-01 5.00889063e-01
-1.00942075e+00 2.26120651e-03 -1.22734815e-01 -1.29205954e+00
-9.92830634e-01 2.79636413e-01 7.28736103e-01 -1.53484344e-01
-1.24938536e+00 7.00713515e-01 4.19103324e-01 -1.09766304e+00
1.64814383e-01 -4.69914883e-01 1.80317894e-01 -4.63853478e-01
9.80266631e-02 7.40926713e-02 5.60380042e-01 -7.84470439e-01
-2.09949329e-01 1.01103079e+00 5.64039886e-01 -8.51242617e-02
1.69394386e+00 2.71054119e-01 -1.52815983e-01 5.40585935e-01
1.23567200e+00 2.07072832e-02 -1.43171418e+00 -1.52345821e-01
-5.44706225e-01 -8.48932445e-01 2.13259012e-01 -6.24358237e-01
-1.32262349e+00 1.09914553e+00 7.05494404e-01 1.30867571e-01
8.51975441e-01 1.54065430e-01 5.47867537e-01 -1.29442140e-01
7.11502612e-01 -4.42345858e-01 -1.13992937e-01 4.34794575e-01
1.34622049e+00 -1.03473818e+00 1.80525690e-01 -6.67386174e-01
-2.94905126e-01 9.18947041e-01 3.05119812e-01 -6.28355682e-01
7.50498235e-01 1.94065738e-02 -1.44393429e-01 -5.44960082e-01
-5.08267701e-01 1.27349319e-02 5.38696826e-01 8.30028892e-01
2.25040764e-02 6.67585135e-02 2.32035384e-01 -2.37854552e-02
-4.66064662e-01 -1.48724884e-01 1.74500555e-01 4.72363561e-01
-1.34553045e-01 -9.19135690e-01 -5.38582146e-01 3.66139501e-01
-5.14447466e-02 1.39295951e-01 -3.17519695e-01 7.65775025e-01
-2.68745376e-03 3.47964674e-01 4.84057695e-01 -3.22216660e-01
6.79152906e-01 -1.33743599e-01 9.20473635e-01 -7.53352344e-01
-5.64705789e-01 1.40918419e-01 -4.80375104e-02 -7.27575421e-01
-5.00588715e-01 -5.25374115e-01 -1.09292865e+00 -4.34058696e-01
-2.44635925e-01 -2.20339298e-01 1.07685506e+00 7.12928116e-01
2.86606044e-01 3.68569970e-01 8.57353032e-01 -1.61274290e+00
-2.19783142e-01 -6.19302690e-01 -5.04674911e-01 4.12924498e-01
2.01083750e-01 -8.91515255e-01 -5.50776780e-01 -1.78556994e-01]
|
[8.894289016723633, -3.368813991546631]
|
c88ec9c9-a4fc-46d6-9903-a52a5dcbda4c
|
exploration-of-whether-skylight-polarization
|
2012.09154
| null |
https://arxiv.org/abs/2012.09154v1
|
https://arxiv.org/pdf/2012.09154v1.pdf
|
Exploration of Whether Skylight Polarization Patterns Contain Three-dimensional Attitude Information
|
Our previous work has demonstrated that Rayleigh model, which is widely used in polarized skylight navigation to describe skylight polarization patterns, does not contain three-dimensional (3D) attitude information [1]. However, it is still necessary to further explore whether the skylight polarization patterns contain 3D attitude information. So, in this paper, a social spider optimization (SSO) method is proposed to estimate three Euler angles, which considers the difference of each pixel among polarization images based on template matching (TM) to make full use of the captured polarization information. In addition, to explore this problem, we not only use angle of polarization (AOP) and degree of polarization (DOP) information, but also the light intensity (LI) information. So, a sky model is established, which combines Berry model and Hosek model to fully describe AOP, DOP, and LI information in the sky, and considers the influence of four neutral points, ground albedo, atmospheric turbidity, and wavelength. The results of simulation show that the SSO algorithm can estimate 3D attitude and the established sky model contains 3D attitude information. However, when there are measurement noise or model error, the accuracy of 3D attitude estimation drops significantly. Especially in field experiment, it is very difficult to estimate 3D attitude. Finally, the results are discussed in detail.
|
['Tong Zhou', 'Hongyang Bai', 'Huaju Liang']
|
2020-11-30
| null | null | null | null |
['template-matching']
|
['computer-vision']
|
[-2.10664704e-01 -5.80379725e-01 2.46514201e-01 -2.32308939e-01
2.48710990e-01 -5.95663667e-01 1.90116778e-01 -7.57151186e-01
-9.46203433e-03 6.94948196e-01 3.61341760e-02 -4.49083239e-01
-1.80803298e-03 -8.77302825e-01 -3.22658032e-01 -1.40914547e+00
1.88242957e-01 2.37218067e-01 3.20521295e-02 -4.45853084e-01
4.00276452e-01 4.61325526e-01 -1.74439764e+00 -6.00507438e-01
1.36490655e+00 8.69548142e-01 2.29562148e-01 4.52538818e-01
-3.75754774e-01 4.58342701e-01 -3.58102292e-01 1.81647584e-01
4.03980225e-01 -4.45270151e-01 -2.08536401e-01 3.00498366e-01
2.80398130e-01 -1.28507346e-01 2.50584096e-01 1.55936193e+00
5.88903964e-01 -1.05663560e-01 4.73409325e-01 -9.44239855e-01
-1.36456490e-01 -4.08529222e-01 -6.63855970e-01 -1.75304219e-01
1.66708872e-01 3.41125488e-01 5.37214875e-01 -6.58192337e-01
5.40015638e-01 9.65866268e-01 3.65667850e-01 -1.73605159e-01
-5.50285161e-01 -5.74554443e-01 -1.07144013e-01 2.19877839e-01
-1.11596489e+00 1.60663098e-01 8.93859625e-01 -4.14616376e-01
1.95135131e-01 4.03764099e-01 1.30304408e+00 3.85608763e-01
4.28759933e-01 4.24229056e-01 2.05518723e+00 -5.60654104e-01
2.82471608e-02 3.44649255e-01 4.25262451e-01 5.94783485e-01
7.56055355e-01 2.82816142e-01 -2.70835578e-01 1.90972583e-03
3.44578505e-01 -2.31700540e-01 -6.78600192e-01 -6.86464190e-01
-9.20603454e-01 5.95269382e-01 3.10031444e-01 -5.16506582e-02
-4.41861957e-01 -7.13958025e-01 -5.82506001e-01 1.11511283e-01
4.22605246e-01 4.43201244e-01 -3.63061845e-01 -1.35092542e-01
-3.22484672e-01 1.88755676e-01 1.02542460e+00 6.16574228e-01
1.28508675e+00 3.61320645e-01 4.26026225e-01 5.85328519e-01
6.68311477e-01 1.69480217e+00 3.54720175e-01 -1.01388025e+00
1.57553870e-02 6.76340878e-01 4.46587235e-01 -1.12907600e+00
-4.72745985e-01 -7.61182427e-01 -6.33818746e-01 3.67994279e-01
1.22388415e-01 -4.24761951e-01 -9.01756942e-01 1.30682051e+00
7.77223468e-01 9.89207625e-02 5.26237786e-01 1.41225219e+00
7.77616322e-01 6.74395502e-01 -6.11867249e-01 -7.69181430e-01
1.31576848e+00 -7.35228300e-01 -8.95861149e-01 -4.18279767e-01
6.17178261e-01 -1.00402343e+00 8.08807015e-01 3.58956635e-01
-5.37563384e-01 -2.56461054e-01 -9.29490387e-01 6.60786152e-01
-3.33902597e-01 -3.92599553e-02 7.19518781e-01 7.05776095e-01
-7.62667000e-01 -8.89042839e-02 -3.66438091e-01 -5.04914641e-01
-4.01455730e-01 1.24079399e-02 -6.21622689e-02 -3.42410654e-02
-1.49281633e+00 9.30384815e-01 -2.27573086e-02 4.58672225e-01
-2.59390593e-01 -4.04390041e-03 -4.42017615e-01 -3.42139214e-01
1.79668188e-01 -7.24635780e-01 6.89767480e-01 -9.63573098e-01
-1.74600255e+00 6.27569914e-01 -4.16319758e-01 7.47883618e-02
-2.67119277e-02 -3.06217670e-02 -8.31307709e-01 5.41881137e-02
-7.56595805e-02 -2.89556687e-03 4.88718808e-01 -1.39889765e+00
-6.77186370e-01 -7.19729364e-01 3.20908763e-02 7.92083383e-01
6.34345189e-02 -1.23298675e-01 -3.84387255e-01 2.90963233e-01
8.05425227e-01 -1.35951185e+00 -1.33758545e-01 -2.66530484e-01
-1.86863571e-01 1.93171605e-01 7.35199749e-01 -5.54981053e-01
8.74506533e-01 -1.90356493e+00 -4.99044657e-02 3.58299613e-01
-2.37258777e-01 3.80061656e-01 1.80973530e-01 9.84778628e-02
3.34784508e-01 -2.82077074e-01 -3.69122237e-01 3.15903693e-01
-2.72849500e-01 4.54721123e-01 -2.76502579e-01 6.13882720e-01
-3.01302671e-01 3.28641087e-01 -6.18487716e-01 -2.79382110e-01
1.72762737e-01 2.98042357e-01 -3.16848695e-01 1.87591225e-01
2.47151051e-02 7.04213202e-01 -6.39477670e-01 8.90958905e-01
1.45702899e+00 -4.31822427e-02 7.86754563e-02 -4.11052912e-01
-7.50450492e-01 1.31753311e-02 -1.52660263e+00 9.29993510e-01
-2.40363181e-01 5.42887807e-01 3.42789292e-01 -5.15403271e-01
1.22681248e+00 -1.42834056e-02 2.00511977e-01 -8.27867687e-01
1.77691042e-01 5.49451470e-01 1.53008178e-01 -9.20544446e-01
3.92067373e-01 -1.69046804e-01 4.22539294e-01 1.11484952e-01
-5.43024659e-01 -6.60348892e-01 -1.22161716e-01 -2.77967542e-01
2.01132581e-01 2.76073456e-01 1.39599085e-01 -4.39760864e-01
7.20151186e-01 3.04269344e-01 7.40381539e-01 2.12870672e-01
-7.97942877e-02 5.21043003e-01 3.74780655e-01 -4.00902480e-01
-6.78212166e-01 -6.23189926e-01 -4.42248315e-01 1.00949191e-01
9.81447518e-01 2.11735964e-01 -3.92407775e-01 -2.19013840e-01
-1.00900881e-01 7.21973240e-01 -1.14606574e-01 1.28060251e-01
-3.33803594e-01 -1.65579200e+00 -2.17074841e-01 -4.23369855e-01
9.74615455e-01 -5.88087976e-01 -3.21262419e-01 -2.18247265e-01
-5.35025477e-01 -8.00502837e-01 2.02921644e-01 -3.73819292e-01
-1.02394736e+00 -1.30908692e+00 -4.67379600e-01 -2.24319443e-01
7.74583519e-01 9.06199515e-01 8.56428146e-01 -1.83863968e-01
1.30206943e-01 3.52630019e-01 -4.54781473e-01 -6.36603773e-01
2.95463558e-02 -5.66024482e-01 3.37640226e-01 2.88377911e-01
5.22292793e-01 -4.87710506e-01 -5.80410242e-01 6.90225065e-01
-4.16501969e-01 1.48665935e-01 6.00432634e-01 4.18114394e-01
5.87468505e-01 1.76624194e-01 -2.89297044e-01 -5.47844291e-01
1.46177799e-01 -2.96973825e-01 -1.15239608e+00 1.80947855e-01
-8.31211805e-01 -1.03173397e-01 3.40707414e-02 7.40019232e-02
-1.33273256e+00 -4.08992708e-01 8.44729096e-02 -1.05110198e-01
-1.30743414e-01 5.93371689e-01 -2.32785478e-01 -5.62985599e-01
2.46131673e-01 5.77342689e-01 3.59660864e-01 -6.16697907e-01
-6.30850270e-02 1.03502882e+00 2.07050256e-02 -7.56080970e-02
9.15682018e-01 8.12956870e-01 5.30000746e-01 -1.18869543e+00
-9.40614820e-01 -4.56052601e-01 -2.43687078e-01 -3.75375271e-01
6.64325058e-01 -1.03673756e+00 -6.70123637e-01 8.61276984e-01
-8.46833587e-01 2.49268740e-01 2.18157604e-01 1.29185593e+00
-7.52120540e-02 7.07043707e-01 1.37313709e-01 -1.17004955e+00
-1.75028250e-01 -1.11355937e+00 5.40316641e-01 5.99549770e-01
4.07868385e-01 -8.79381299e-01 2.85139024e-01 5.32777548e-01
3.66874099e-01 4.35038395e-02 4.45390850e-01 2.21119151e-01
-1.12237918e+00 -7.33346939e-02 -1.64385408e-01 3.93193960e-01
1.03243202e-01 2.71385163e-01 -6.95725918e-01 -2.02358484e-01
6.24654949e-01 9.79464278e-02 5.42399168e-01 5.28887749e-01
3.83441538e-01 -1.86929077e-01 -1.06808014e-01 1.34055686e+00
1.51886928e+00 4.74837095e-01 7.05647945e-01 9.56535101e-01
5.89363873e-01 8.10334206e-01 1.14745080e+00 2.95187891e-01
5.12969196e-01 5.42945206e-01 9.45653498e-01 -1.11898862e-01
7.75893480e-02 1.85980260e-01 3.35842192e-01 1.14923739e+00
-5.43950081e-01 -2.27797329e-01 -7.71809936e-01 1.44798443e-01
-1.45176899e+00 -7.68063843e-01 -7.99222410e-01 2.35847044e+00
3.21651965e-01 -1.71721056e-01 -5.93008637e-01 -3.13785315e-01
5.37453890e-01 3.38986188e-01 -3.73819292e-01 -1.11482292e-01
-7.83994675e-01 -5.34913778e-01 9.62476492e-01 7.36658156e-01
-6.73977852e-01 5.17236471e-01 5.77212381e+00 1.24733135e-01
-1.58457053e+00 -3.42232436e-01 6.42501563e-02 3.99193376e-01
-8.08709085e-01 5.32371283e-01 -1.07935476e+00 7.58920133e-01
2.72444606e-01 1.52700171e-01 4.23674196e-01 6.17091119e-01
5.35871506e-01 -5.51496387e-01 2.86113203e-01 8.74120116e-01
2.34349787e-01 -6.31194055e-01 -6.44828603e-02 3.09305459e-01
7.87581563e-01 1.71527371e-01 -2.76803877e-02 -9.48874131e-02
1.42486796e-01 -1.89674020e-01 2.04978392e-01 9.88449514e-01
4.48198915e-01 -3.51015180e-01 1.11370993e+00 5.24577618e-01
-6.56472266e-01 5.49965799e-02 -5.83002985e-01 -3.71013582e-01
4.07451272e-01 1.13532555e+00 -5.12954175e-01 9.60452676e-01
8.81862283e-01 6.87334180e-01 -3.32421392e-01 1.24374688e+00
-6.21352375e-01 7.20189333e-01 -5.51534355e-01 -2.79333264e-01
2.27611005e-01 -1.16449380e+00 1.00582540e+00 2.75956541e-01
1.02126908e+00 4.17348951e-01 -2.25315690e-01 4.71160620e-01
6.27971888e-01 2.47225210e-01 -5.55538237e-01 1.55415818e-01
1.99559420e-01 1.26930904e+00 -6.31802529e-02 -2.52504945e-01
-5.12377501e-01 3.01725417e-01 -5.02817631e-01 7.38129079e-01
-6.07196927e-01 -2.09635109e-01 9.14528131e-01 6.02975897e-02
9.16982442e-02 -2.94526815e-01 -8.78387988e-02 -1.40594482e+00
-1.58004388e-02 -6.95704818e-01 -9.30237770e-02 -1.50286329e+00
-7.84580648e-01 3.25153619e-01 -1.54521570e-01 -1.70184803e+00
1.71135843e-01 -8.32080185e-01 -6.65948987e-01 1.27040803e+00
-1.97756922e+00 -9.13759947e-01 -6.56914175e-01 1.03917554e-01
-2.19815671e-01 -1.45008594e-01 5.10370076e-01 -1.48490384e-01
-5.82839906e-01 -5.37254870e-01 7.75208592e-01 -5.07216811e-01
7.79198945e-01 -1.02010620e+00 -2.68252164e-01 7.99399436e-01
-2.72558242e-01 5.09260654e-01 1.13783908e+00 -6.35060906e-01
-1.65224302e+00 -6.35152221e-01 8.87939811e-01 -2.19173580e-01
3.94126534e-01 3.54067147e-01 -7.91051567e-01 2.23939329e-01
2.01095879e-01 -1.26309454e-01 4.17076409e-01 1.28925443e-01
1.43647075e-01 -6.33037746e-01 -8.51710856e-01 5.15534520e-01
5.98556817e-01 -8.94844756e-02 -6.59043372e-01 5.02825916e-01
5.90995789e-01 -4.99573529e-01 -4.49513525e-01 9.16299343e-01
5.99495530e-01 -1.31873977e+00 7.81649113e-01 7.87174888e-03
-2.74637997e-01 -9.57450509e-01 -2.49911115e-01 -1.39558911e+00
-2.87751585e-01 -2.33072475e-01 2.23608553e-01 8.79072726e-01
1.93593711e-01 -1.26620090e+00 6.53644025e-01 1.96504593e-01
-1.45403311e-01 -3.35353464e-01 -6.32125020e-01 -6.06091619e-01
-4.97813284e-01 8.26891810e-02 8.37064087e-01 9.61370647e-01
-4.77478623e-01 3.27645868e-01 -5.50494969e-01 9.66261089e-01
8.73102784e-01 9.67873991e-01 9.60725665e-01 -1.45985472e+00
-6.51351921e-03 9.30346549e-02 -3.63875553e-02 -1.02402568e+00
-1.12513818e-01 -3.20913047e-01 -1.29157633e-01 -1.65259576e+00
-1.18193343e-01 -4.33741510e-01 -2.41452247e-01 -2.94444740e-01
-2.96194404e-01 5.97619340e-02 -1.58470079e-01 6.89296663e-01
-1.80018187e-01 6.89619780e-01 1.80704939e+00 1.07016318e-01
-2.22846344e-01 2.70625472e-01 -2.74683982e-01 8.30210030e-01
7.51409829e-01 -1.26348078e-01 -2.11761609e-01 -6.54491425e-01
5.42273879e-01 1.39949188e-01 1.61882415e-01 -1.09206903e+00
1.87024042e-01 -5.81899941e-01 1.17603347e-01 -9.23789501e-01
4.27094996e-01 -9.23346698e-01 2.67496407e-01 4.66188073e-01
8.31946135e-01 -2.50327438e-01 -1.95472822e-01 3.59661907e-01
-6.02625608e-01 -2.68576771e-01 7.21605659e-01 -3.40420097e-01
-9.57363725e-01 3.17078143e-01 -2.68715501e-01 -3.49868685e-01
6.60006523e-01 -4.72221792e-01 -6.05710208e-01 -4.56835985e-01
-2.35596731e-01 4.27122772e-01 9.61444676e-01 -7.09805191e-02
2.04047799e-01 -1.14250755e+00 -4.68782514e-01 6.05389416e-01
1.20219104e-01 -1.10147513e-01 5.25951564e-01 1.18866742e+00
-1.07273877e+00 5.29760897e-01 -1.60385102e-01 -7.18961298e-01
-1.38856196e+00 2.63171107e-01 5.09152949e-01 3.55234355e-01
2.00624466e-02 6.25965238e-01 3.68306786e-01 -8.48067939e-01
-5.31184912e-01 5.58141246e-02 -7.08727241e-01 1.56755254e-01
1.56961799e-01 3.44104558e-01 -2.98003018e-01 -9.19770539e-01
-2.00847119e-01 1.58748877e+00 7.30506480e-01 -6.67639636e-03
9.60553646e-01 -6.33581519e-01 -6.33992255e-01 4.33826476e-01
8.37187529e-01 4.53334421e-01 -1.04091513e+00 -1.36285901e-01
-6.17236972e-01 -7.56524742e-01 1.96747109e-01 -6.60197139e-01
-8.09456527e-01 9.61757898e-01 6.92058325e-01 1.75136715e-01
1.31194830e+00 -6.18163645e-01 5.30564129e-01 5.72304249e-01
6.05781078e-01 -1.06985569e+00 -6.04503870e-01 8.02836716e-01
4.79960710e-01 -1.26419926e+00 2.02187479e-01 -6.46507800e-01
-6.65307164e-01 1.01352727e+00 6.65164113e-01 3.64490509e-01
8.22116792e-01 -2.94237852e-01 6.80608928e-01 -1.40074074e-01
-3.38005036e-01 -3.79017740e-01 1.73817165e-02 3.11699182e-01
6.29417747e-02 1.10594384e-01 -6.20444357e-01 5.37553988e-02
-1.86429784e-01 -5.07493243e-02 7.91361451e-01 6.86421335e-01
-1.10949123e+00 -1.00985324e+00 -9.42292452e-01 1.25657633e-01
3.08188517e-02 -4.64509949e-02 7.19990432e-02 5.28345287e-01
2.94532180e-01 7.80676603e-01 1.22892231e-01 -2.27126122e-01
2.53901303e-01 -1.77102059e-01 5.23564033e-02 -6.68509826e-02
4.88955081e-01 9.16461647e-02 1.93506539e-01 -3.02154988e-01
-9.97067332e-01 -4.53326881e-01 -8.39001119e-01 -2.32985228e-01
-6.85917854e-01 7.50116348e-01 9.40284967e-01 6.40489280e-01
2.51898497e-01 -8.85463208e-02 1.08345723e+00 -4.22376722e-01
-2.58495539e-01 -7.96275973e-01 -1.16257083e+00 -3.53833195e-03
2.30648771e-01 -8.92200351e-01 -1.10665071e+00 -5.05147994e-01]
|
[9.965824127197266, -2.7836861610412598]
|
9731b9cd-91cc-4458-95e1-31ea1c0c43bc
|
learning-to-detect-specular-highlights-from
| null | null |
https://dl.acm.org/doi/abs/10.1145/3394171.3413586
|
https://dl.acm.org/doi/abs/10.1145/3394171.3413586
|
Learning to Detect Specular Highlights from Real-world Images
|
Specular highlight detection is a challenging problem, and has many applications such as shiny object detection and light source estimation. Although various highlight detection methods have been proposed, they fail to disambiguate bright material surfaces from highlights, and cannot handle non-white-balanced images. Moreover, at present, there is still no benchmark dataset for highlight detection. In this paper, we present a large-scale real-world highlight dataset containing a rich variety of material categories, with diverse highlight shapes and appearances, in which each image is with an annotated ground-truth mask. Based on the dataset, we develop a deep learning-based specular highlight detection network (SHDNet) leveraging multi-scale context contrasted features to accurately detect specular highlights of varying scales. In addition, we design a binary cross-entropy (BCE) loss and an intersection-over-union edge (IoUE) loss for our network. Compared with existing highlight detection methods, our method can accurately detect highlights of different sizes, while effectively excluding the non- highlight regions, such as bright materials, non-specular as well as colored lighting, and even light sources.
|
['and Chunaxia Xiao', 'Lei Zhu', 'QiFeng Lin', 'Qing Zhang', 'Gang Fu']
|
2020-10-10
| null | null | null | null |
['highlight-detection']
|
['computer-vision']
|
[ 4.86837059e-01 -6.32480502e-01 2.85011947e-01 -1.11037098e-01
-5.01498461e-01 -6.91548645e-01 1.72083810e-01 -6.62868246e-02
1.34998098e-01 6.08361542e-01 2.65863054e-02 3.30101810e-02
3.46497446e-01 -5.24246454e-01 -4.72892135e-01 -8.61508191e-01
-2.75782328e-02 -3.92970532e-01 7.34707534e-01 -2.98806354e-02
3.86464864e-01 4.37332571e-01 -1.50234842e+00 4.83332038e-01
9.74409461e-01 1.31472373e+00 1.76181972e-01 5.21042109e-01
-7.68126771e-02 8.70644808e-01 -7.09889114e-01 -3.09891284e-01
3.99524808e-01 -3.29318732e-01 1.74725309e-01 -1.98417291e-01
1.12277389e+00 -5.68808317e-01 -2.38981277e-01 1.09742510e+00
7.90440798e-01 6.53952584e-02 6.49383128e-01 -1.40780318e+00
-6.09550774e-01 1.00713171e-01 -1.05446720e+00 1.15526371e-01
1.30898625e-01 2.19827875e-01 8.99957776e-01 -1.07050431e+00
5.46089649e-01 9.33577955e-01 8.43509972e-01 9.40622464e-02
-7.78250873e-01 -1.03845751e+00 9.85879228e-02 9.10811797e-02
-1.09138811e+00 -2.66087860e-01 1.35717952e+00 -2.37648636e-01
1.70100570e-01 5.67943513e-01 7.30882227e-01 8.45002770e-01
1.13644503e-01 1.07187712e+00 1.66336000e+00 -3.29592854e-01
-1.80115886e-02 1.36594743e-01 -2.55937606e-01 8.82460892e-01
5.06333351e-01 2.00436115e-01 -8.05395246e-01 -1.39237553e-01
5.75762570e-01 5.33785261e-02 -6.90282464e-01 -2.55170882e-01
-1.19828355e+00 3.44317436e-01 4.72616822e-01 -2.60701597e-01
-1.22521415e-01 9.47584882e-02 4.34598178e-01 -7.68052265e-02
6.76793516e-01 4.39071059e-01 -3.35328758e-01 2.73932725e-01
-1.07945693e+00 1.54732406e-01 6.29763842e-01 8.73945355e-01
6.31844401e-01 1.87664106e-01 -4.75898832e-01 9.46639955e-01
-7.54296780e-02 9.68208611e-01 -1.99735746e-01 -5.39979517e-01
2.48405397e-01 5.26593864e-01 4.25825000e-01 -1.34344602e+00
-6.33851171e-01 -6.66449785e-01 -7.17275620e-01 4.26396132e-01
1.92463204e-01 -3.03508520e-01 -5.93736351e-01 1.41564989e+00
3.76569957e-01 3.39257538e-01 -3.48832220e-01 1.36724246e+00
1.13538706e+00 6.48471236e-01 -3.22946161e-01 -1.50602058e-01
1.23878551e+00 -1.14848852e+00 -7.00791121e-01 -5.46161085e-03
-1.83680192e-01 -1.30877066e+00 1.20801580e+00 6.10300183e-01
-1.20141256e+00 -4.65377212e-01 -1.05414200e+00 -2.28696287e-01
-3.21041554e-01 3.47072631e-01 8.52857590e-01 5.18686473e-01
-8.18863094e-01 6.04142062e-02 -2.27776900e-01 1.09332643e-01
4.15536255e-01 -2.80349016e-01 1.80886194e-01 -1.68509886e-01
-8.88293505e-01 4.25237328e-01 7.10172430e-02 2.89909214e-01
-6.92439616e-01 -1.06222963e+00 -6.63097501e-01 7.27020577e-02
6.98488832e-01 -4.93896812e-01 8.39584410e-01 -1.24524510e+00
-1.39357722e+00 8.78032744e-01 -1.64080024e-01 -3.03162746e-02
8.25054705e-01 -3.23615938e-01 -7.25762248e-01 1.75814062e-01
-1.49547070e-01 1.73056111e-01 9.21693623e-01 -1.82614720e+00
-7.37827480e-01 1.69280812e-01 1.26192406e-01 1.39235049e-01
-2.65039295e-01 4.91141260e-01 -5.87385893e-01 -7.81820714e-01
-8.33743364e-02 -6.13199830e-01 3.70489061e-01 5.43367982e-01
-1.16027093e+00 3.48365694e-01 9.60435092e-01 -5.44460773e-01
1.21673739e+00 -2.16275811e+00 -6.15870118e-01 1.38606161e-01
4.12100822e-01 2.96346486e-01 -1.58093348e-01 3.19720834e-01
2.69253522e-01 -2.45242015e-01 -3.19591284e-01 -1.78800434e-01
6.38356283e-02 -6.66078508e-01 -5.04585505e-01 4.91735369e-01
4.51320738e-01 5.73573828e-01 -1.00598037e+00 -5.72677493e-01
1.73856124e-01 7.28778601e-01 4.74780686e-02 1.03561237e-01
-1.46648750e-01 -8.99205655e-02 -2.75978446e-01 1.36228108e+00
1.21314251e+00 -1.15452677e-01 -3.36485296e-01 -3.95198435e-01
-4.95257258e-01 -3.48730356e-01 -1.21490526e+00 9.37989235e-01
-4.78948772e-01 1.30159247e+00 1.84006587e-01 -8.77644941e-02
1.22441709e+00 -3.04331034e-01 3.71458590e-01 -8.19412947e-01
7.05287829e-02 3.96980852e-01 -2.47655392e-01 -5.47938406e-01
7.23591506e-01 1.10944591e-01 4.92890388e-01 2.54651815e-01
-8.11788797e-01 6.28911331e-02 1.01100989e-02 -5.50362431e-02
9.89231944e-01 1.33979050e-02 2.51496099e-02 -1.39220193e-01
3.90936255e-01 -3.66265148e-01 6.99205399e-01 6.16571248e-01
-3.02555263e-01 1.07832849e+00 4.49832976e-01 -3.76570910e-01
-6.70204341e-01 -1.22737694e+00 -2.00913846e-01 1.14236033e+00
6.64186358e-01 4.92122397e-02 -4.37956095e-01 -4.06668395e-01
3.44400227e-01 4.02486712e-01 -3.79835337e-01 -1.24061638e-02
-6.33558273e-01 -5.18906236e-01 2.19802171e-01 2.18315095e-01
8.57965231e-01 -1.20413160e+00 -9.64239776e-01 -1.07764460e-01
8.40083137e-02 -1.10905468e+00 -8.78250182e-01 -4.73712161e-02
-2.30779171e-01 -1.46352410e+00 -9.44784343e-01 -8.30313146e-01
6.21838927e-01 9.09510672e-01 1.35750675e+00 4.73597795e-02
-8.87700140e-01 2.32106283e-01 -1.34649649e-01 -9.80979979e-01
2.00117957e-02 -5.00105739e-01 -5.88626862e-01 3.14841688e-01
1.51195914e-01 1.33233881e-02 -1.24983943e+00 5.01337171e-01
-8.27656090e-01 3.64427090e-01 6.85683727e-01 7.84670472e-01
6.69441700e-01 2.98100680e-01 4.20907110e-01 -9.45500195e-01
5.62962830e-01 -1.95968509e-01 -8.00908744e-01 3.40949923e-01
-2.72417665e-01 -6.68386459e-01 6.25999987e-01 -6.16274595e-01
-1.36644876e+00 -3.20623577e-01 6.18248165e-01 -5.30840576e-01
9.91971120e-02 2.37885728e-01 -1.51318654e-01 -3.99842173e-01
5.40772855e-01 2.03047037e-01 -4.23004866e-01 -1.95492655e-01
1.08558595e-01 4.44833398e-01 4.82765049e-01 -2.81560600e-01
8.25198829e-01 9.77602184e-01 1.99586660e-01 -7.70484865e-01
-1.16952479e+00 -6.49907708e-01 -1.56305447e-01 -5.68252385e-01
2.96962351e-01 -9.09180462e-01 -7.78078020e-01 8.39559138e-01
-1.09895205e+00 -5.52693427e-01 1.25256374e-01 1.46647260e-01
-1.00809872e-01 2.47112617e-01 -4.23681527e-01 -1.24607909e+00
-6.23151004e-01 -7.94066191e-01 1.32131720e+00 5.99339128e-01
3.25222164e-01 -7.38387287e-01 -2.51747489e-01 -6.14249036e-02
7.01524138e-01 9.52163637e-01 7.76579738e-01 1.57560095e-01
-9.42815423e-01 -2.29753762e-01 -1.05408108e+00 4.18928474e-01
1.27581164e-01 5.16991258e-01 -1.15946472e+00 -1.06503926e-01
-3.92968357e-01 -1.00439005e-01 1.30090463e+00 5.84657609e-01
1.23526907e+00 1.83068942e-02 -3.05001974e-01 8.88277769e-01
1.63704693e+00 1.76569492e-01 4.16143924e-01 4.07451570e-01
8.41004491e-01 6.84702218e-01 7.95393944e-01 5.42365551e-01
1.17375612e-01 4.67442274e-01 5.59826374e-01 -8.51637185e-01
-5.33797026e-01 6.43697530e-02 2.86743045e-01 2.28961766e-01
1.26362801e-01 -5.03356755e-01 -6.07774198e-01 3.49985659e-01
-1.32323825e+00 -9.89327967e-01 -4.17560637e-01 2.18850589e+00
8.34013164e-01 1.70643866e-01 1.67779654e-01 -2.23186210e-01
9.65800822e-01 5.48062742e-01 -1.02835763e+00 4.00472200e-03
-8.23421896e-01 3.35371569e-02 6.37483299e-01 1.45723462e-01
-1.36036611e+00 7.27234066e-01 6.04000950e+00 6.07967257e-01
-1.59845471e+00 -1.23519585e-01 7.46270299e-01 -3.58232647e-01
-5.21595359e-01 -2.83804983e-01 -4.74547029e-01 9.05226529e-01
-3.33023220e-02 2.26929262e-01 2.99549907e-01 6.96862221e-01
2.78478503e-01 -3.38066727e-01 -5.05202651e-01 1.23049188e+00
5.11664987e-01 -1.13338411e+00 -3.34143341e-01 -4.28046435e-01
1.17481649e+00 9.61183831e-02 4.60192025e-01 -7.59264901e-02
-1.20330565e-01 -6.23837411e-01 8.91015589e-01 7.46671617e-01
9.10064816e-01 -7.26371944e-01 3.73034507e-01 -3.08254778e-01
-1.28818071e+00 -1.70624465e-01 -4.41134721e-01 3.91534150e-01
2.52650101e-02 1.05642962e+00 -3.71471763e-01 2.45844603e-01
5.64797342e-01 5.78295529e-01 -3.56878906e-01 1.86724067e+00
-2.43855193e-01 5.05589426e-01 -3.13345641e-01 -4.54252213e-01
1.86802857e-02 -2.49267191e-01 6.04371965e-01 1.65750563e+00
2.33835533e-01 -2.94826120e-01 3.16460103e-01 1.19352627e+00
-2.41149962e-01 1.38454288e-01 -1.96562022e-01 3.09033364e-01
5.24545550e-01 1.61008799e+00 -8.71239543e-01 -1.05930597e-01
-5.58865786e-01 9.48884785e-01 -3.50923613e-02 7.69329786e-01
-9.59576726e-01 -7.76438951e-01 4.82926995e-01 1.90313548e-01
-8.21429789e-02 1.36567578e-02 -3.85766089e-01 -1.00931966e+00
4.62997496e-01 -5.95874250e-01 6.43036515e-02 -1.33134747e+00
-1.33734679e+00 3.55598330e-01 -3.70643377e-01 -1.45751667e+00
8.87397468e-01 -8.31258953e-01 -1.25575435e+00 9.33620155e-01
-2.43980145e+00 -1.31341410e+00 -1.02494109e+00 2.08583787e-01
2.85115063e-01 3.32208015e-02 6.70755804e-02 3.53683025e-01
-4.69731599e-01 4.32304621e-01 4.86600399e-01 4.64988872e-02
1.14212966e+00 -1.27539062e+00 7.42930248e-02 8.43166411e-01
-2.04637289e-01 4.55921561e-01 7.40922272e-01 -6.28196359e-01
-1.49455535e+00 -1.18480909e+00 1.83846988e-02 -6.70348629e-02
6.21639907e-01 -4.54923183e-01 -8.28467250e-01 -1.13273919e-01
1.33253872e-01 2.71147937e-01 4.62723196e-01 -4.20828722e-02
-5.96378922e-01 -3.86505842e-01 -8.76768112e-01 6.98123932e-01
9.83824134e-01 -4.25515205e-01 3.18679482e-01 6.08627617e-01
4.60607171e-01 -6.50554240e-01 -2.49023110e-01 5.51337600e-01
7.05292404e-01 -1.38790882e+00 1.04432619e+00 -6.58061653e-02
7.35271692e-01 -3.83357167e-01 2.10213169e-01 -1.04284489e+00
-1.30797494e-02 -8.99369538e-01 -2.06644684e-01 1.36265993e+00
1.76301897e-01 -6.49069309e-01 7.66251445e-01 8.83103386e-02
-4.48046237e-01 -1.02129412e+00 -4.25702453e-01 -8.02347064e-01
-5.86915314e-01 -1.68567210e-01 7.47294009e-01 9.57989156e-01
-6.38493657e-01 -2.84530640e-01 -5.32757521e-01 3.02252173e-01
8.82686734e-01 8.83678198e-01 9.18378174e-01 -1.29076600e+00
8.68784916e-03 -8.64139557e-01 -4.50623780e-02 -6.54903293e-01
-7.67638534e-02 -5.36201417e-01 2.65740305e-01 -1.61636674e+00
4.87175554e-01 -6.16625786e-01 -7.09984362e-01 2.31714934e-01
-5.68910420e-01 2.35041022e-01 -1.16754482e-02 1.88620210e-01
-8.00197065e-01 4.39425349e-01 1.54376638e+00 -1.81624979e-01
-2.16344312e-01 -1.67618677e-01 -6.28602266e-01 9.50353503e-01
7.68679798e-01 2.75936611e-02 -2.75803417e-01 -2.49473408e-01
5.46390831e-01 -3.94761384e-01 6.75805628e-01 -9.89909768e-01
8.30097124e-02 -4.15521234e-01 6.64126813e-01 -1.00884569e+00
4.05822873e-01 -6.67094350e-01 -3.11231077e-01 1.57352880e-01
-2.85102367e-01 -3.43300432e-01 2.29260787e-01 5.66708505e-01
-5.08573726e-02 7.42683038e-02 8.54462564e-01 9.36741680e-02
-8.27161610e-01 4.06802982e-01 3.59254360e-01 3.08439910e-01
1.00663352e+00 -3.14440966e-01 -1.19493580e+00 -1.97214991e-01
4.11848783e-01 1.46655664e-01 6.82263136e-01 2.54216284e-01
7.37916887e-01 -1.08969986e+00 -7.09184945e-01 6.78409487e-02
4.90329891e-01 2.07234938e-02 6.01597607e-01 8.40623856e-01
-7.73510218e-01 -8.96103755e-02 -1.23480201e-01 -1.68425232e-01
-1.32555234e+00 4.34954822e-01 3.20198208e-01 2.91404963e-01
-7.33320057e-01 9.37049687e-01 7.10557282e-01 1.57034427e-01
2.97755748e-01 -4.32435542e-01 9.72824395e-02 -2.47917604e-02
7.33207881e-01 6.20092213e-01 -2.67877668e-01 -2.73516446e-01
9.86960903e-02 7.09297657e-01 7.99552873e-02 5.43742180e-01
9.67934430e-01 -3.25480513e-02 -1.53042406e-01 5.61762989e-01
1.08223736e+00 6.17536187e-01 -1.61015904e+00 -4.10874426e-01
-4.38868970e-01 -8.41993213e-01 1.91330284e-01 -1.09530067e+00
-1.30551898e+00 8.34866762e-01 5.90171874e-01 3.44640970e-01
1.35282910e+00 -3.74015719e-01 9.99233484e-01 1.22549802e-01
1.25511482e-01 -1.34338868e+00 4.34886426e-01 1.65735230e-01
9.74547625e-01 -1.55276549e+00 2.11059928e-01 -8.54531467e-01
-7.15769887e-01 1.25459850e+00 7.02189207e-01 -2.35216562e-02
3.75578582e-01 3.96836400e-01 2.25942492e-01 -3.03919673e-01
-3.53540391e-01 -3.99199516e-01 5.89736104e-01 5.41111708e-01
2.37843737e-01 1.86428308e-01 1.48567393e-01 3.06131393e-01
3.46720293e-02 -3.12778592e-01 5.59673071e-01 5.53645015e-01
-6.90391302e-01 -1.06170803e-01 -4.88563508e-01 6.63700759e-01
-4.20066595e-01 -4.30008709e-01 -6.65859640e-01 8.80679309e-01
1.33625999e-01 7.10157454e-01 -1.34677906e-02 1.58986133e-02
1.48113698e-01 -5.44096351e-01 1.28137544e-01 -2.12011114e-01
-3.67392033e-01 3.24295431e-01 2.99422462e-02 -3.60065252e-01
-2.92773187e-01 -6.01614952e-01 -9.82133567e-01 -7.99812526e-02
-6.71014547e-01 -3.45548898e-01 7.84209073e-01 1.41403988e-01
2.34664753e-01 6.85262561e-01 9.29881632e-01 -7.54735231e-01
-5.97583540e-02 -6.27705276e-01 -9.33277786e-01 5.99652708e-01
5.27863443e-01 -8.98932338e-01 -5.51419854e-01 -2.34003127e-01]
|
[10.800372123718262, -2.7258996963500977]
|
fe236193-6149-46ad-9b80-65eecd2c2ab2
|
iterative-gradient-encoding-network-with
|
2103.15903
| null |
https://arxiv.org/abs/2103.15903v1
|
https://arxiv.org/pdf/2103.15903v1.pdf
|
Iterative Gradient Encoding Network with Feature Co-Occurrence Loss for Single Image Reflection Removal
|
Removing undesired reflections from a photo taken in front of glass is of great importance for enhancing visual computing systems' efficiency. Previous learning-based approaches have produced visually plausible results for some reflections type, however, failed to generalize against other reflection types. There is a dearth of literature for efficient methods concerning single image reflection removal, which can generalize well in large-scale reflection types. In this study, we proposed an iterative gradient encoding network for single image reflection removal. Next, to further supervise the network in learning the correlation between the transmission layer features, we proposed a feature co-occurrence loss. Extensive experiments on the public benchmark dataset of SIR$^2$ demonstrated that our method can remove reflection favorably against the existing state-of-the-art method on all imaging settings, including diverse backgrounds. Moreover, as the reflection strength increases, our method can still remove reflection even where other state of the art methods failed.
|
['Prabir Kumar Biswas', 'Sutanu Bera']
|
2021-03-29
| null | null | null | null |
['reflection-removal']
|
['computer-vision']
|
[ 8.30469489e-01 -1.36101454e-01 5.64826131e-01 -2.62803733e-01
-7.26877928e-01 -1.43371165e-01 2.46916100e-01 -5.57330668e-01
-1.84834331e-01 5.90612352e-01 3.73858400e-02 -3.53077352e-01
-1.24626420e-01 -6.74864948e-01 -8.86988461e-01 -1.07278836e+00
-6.93506077e-02 -5.21917462e-01 2.71789998e-01 -2.80336738e-01
4.78522927e-01 5.99470317e-01 -1.60460448e+00 6.46330297e-01
6.87054873e-01 9.40650105e-01 3.19901593e-02 6.19906127e-01
3.57724249e-01 9.69497025e-01 -6.80317879e-01 -4.27880794e-01
7.14608073e-01 -4.54123884e-01 -3.67274910e-01 3.61746773e-02
1.19354963e+00 -3.83349717e-01 -4.68132943e-01 8.61571133e-01
7.99646199e-01 1.26856431e-01 6.39927506e-01 -7.55738139e-01
-7.97321022e-01 3.87721717e-01 -9.05194283e-01 2.26454347e-01
2.97506332e-01 2.12238595e-01 7.73421645e-01 -8.97894263e-01
3.75684619e-01 1.12285161e+00 1.07993639e+00 4.41667110e-01
-9.41386580e-01 -7.72136211e-01 2.59917766e-01 1.72660723e-01
-1.27200150e+00 -5.58124721e-01 8.73213410e-01 -6.26873970e-02
1.19116700e+00 4.47711140e-01 5.02342463e-01 8.60371530e-01
2.74991184e-01 5.63814998e-01 1.33578098e+00 -6.99490428e-01
-3.16958994e-01 6.30560294e-02 -2.54523754e-02 9.51027095e-01
4.79446530e-01 2.47975498e-01 -6.99035168e-01 2.02783093e-01
5.58459878e-01 -1.43242493e-01 -4.99036580e-01 9.49930251e-02
-6.25846207e-01 3.85682136e-01 4.73163366e-01 -6.93684369e-02
-7.94382617e-02 2.36062840e-01 -3.65611650e-02 4.05306965e-01
8.71082544e-01 4.18692410e-01 -2.72176623e-01 5.10490954e-01
-7.55424976e-01 -9.59987417e-02 6.37373507e-01 5.45330107e-01
5.72055578e-01 3.40987265e-01 -3.52852434e-01 1.11047721e+00
3.59349608e-01 7.85759568e-01 -3.78943652e-01 -8.93390238e-01
4.19548690e-01 3.59595448e-01 3.99096385e-02 -9.73646283e-01
-6.06045127e-01 -6.39564753e-01 -1.06948149e+00 7.07572043e-01
2.38348067e-01 -1.50519326e-01 -1.15530097e+00 1.34215426e+00
6.36339262e-02 4.42059129e-01 1.29714534e-01 9.73345160e-01
1.24223948e+00 3.60014737e-01 -4.70541447e-01 -1.85285047e-01
1.07148480e+00 -9.17839527e-01 -3.83081973e-01 -2.47177303e-01
6.88391924e-02 -1.16715121e+00 8.75764489e-01 9.39834833e-01
-1.23309064e+00 -5.82176566e-01 -1.19890237e+00 -3.98655310e-02
-2.02781078e-03 3.81057084e-01 8.69080722e-01 1.18364012e+00
-1.19701254e+00 4.51482624e-01 -5.14405906e-01 -3.52726020e-02
7.06411183e-01 3.20795685e-01 -4.16733772e-02 -6.06780410e-01
-8.21480870e-01 7.84918427e-01 -4.00726467e-01 6.19282842e-01
-8.29583049e-01 -8.55242550e-01 -6.19374812e-01 -2.07458124e-01
4.41488415e-01 -7.91910112e-01 6.70608461e-01 -8.17914844e-01
-1.64070976e+00 8.67859840e-01 -1.49328962e-01 -4.51257855e-01
5.18371284e-01 -5.38775444e-01 -5.20022511e-01 1.36225283e-01
-5.04724801e-01 2.15275377e-01 1.33312452e+00 -1.77292371e+00
-5.54207027e-01 5.60986884e-02 2.40317404e-01 3.86601031e-01
-2.48163745e-01 1.26444586e-02 -6.53783679e-01 -2.74734914e-01
-3.58106731e-03 -8.53847086e-01 -2.07569182e-01 -8.27326328e-02
-6.24127507e-01 1.75017625e-01 5.48392832e-01 -4.28063303e-01
9.14309978e-01 -2.30708647e+00 -3.81249785e-01 2.16341391e-01
3.35884899e-01 6.45676255e-01 -2.91950375e-01 2.57884324e-01
-1.83863133e-01 -1.34223521e-01 -3.09711933e-01 -2.90203810e-01
-2.03926295e-01 -4.40936148e-01 -3.05784345e-01 8.47546279e-01
2.49393851e-01 4.27400291e-01 -5.68847418e-01 -1.59148902e-01
3.27951759e-01 1.08339071e+00 -6.69327080e-01 -3.55209783e-02
1.88753083e-01 2.55022854e-01 -1.57309011e-01 6.38146281e-01
1.09364855e+00 -1.91954598e-01 -6.58908412e-02 -6.24694109e-01
-2.73189455e-01 -1.39942244e-01 -9.00728345e-01 1.18284476e+00
-6.53587043e-01 9.21684921e-01 -1.37830988e-01 -7.94233441e-01
9.46312129e-01 -7.75117055e-02 4.83033806e-01 -1.07190108e+00
-1.97154954e-01 3.41583192e-02 2.10728809e-01 -6.85535610e-01
3.69029462e-01 2.27136053e-02 4.35318470e-01 1.92030966e-01
-5.45116484e-01 -1.19485430e-01 -4.60769311e-02 -1.01561032e-01
1.39167261e+00 2.97734767e-01 -1.56617507e-01 2.15107918e-01
6.69476628e-01 -4.01119113e-01 2.90464789e-01 1.11030793e+00
1.13593653e-01 1.13104081e+00 -9.39681679e-02 -5.47685564e-01
-7.04785883e-01 -1.20932817e+00 -1.69018820e-01 9.56917107e-01
3.29861969e-01 -1.92900583e-01 -5.49230754e-01 -3.42665076e-01
-2.48355150e-01 5.23588598e-01 -2.65589505e-01 -3.28017861e-01
-8.13606262e-01 -1.43689966e+00 4.82042640e-01 2.43461370e-01
8.71334255e-01 -1.09582984e+00 -7.00505376e-01 -1.24991782e-01
-1.98936790e-01 -1.37679935e+00 -9.75753367e-03 -7.25870356e-02
-5.68059683e-01 -1.24414432e+00 -5.95389366e-01 -6.72210097e-01
8.57529998e-01 8.63769829e-01 1.51102567e+00 4.71862227e-01
-1.08491683e+00 6.06687963e-01 -2.51617491e-01 -5.88265419e-01
-3.09086125e-02 -4.26642269e-01 -4.75814462e-01 -2.35717837e-02
2.28029042e-01 -3.87637854e-01 -1.21818304e+00 1.73257902e-01
-8.51833284e-01 -3.95860001e-02 8.32597136e-01 6.38157845e-01
3.58175159e-01 4.76331294e-01 1.34923279e-01 -1.08311033e+00
4.60195661e-01 4.26669680e-02 -4.27401841e-01 2.45448411e-01
-4.90046203e-01 -2.08211496e-01 5.66057086e-01 -3.02225556e-02
-1.72668910e+00 -1.72183305e-01 -7.94233456e-02 5.73767051e-02
-5.43793216e-02 -1.17986966e-02 1.41673028e-01 -6.12299919e-01
7.88839161e-01 1.34127468e-01 -4.50737327e-01 -2.02690646e-01
2.99832702e-01 3.06854665e-01 1.70593038e-01 -2.84514338e-01
8.03283453e-01 1.05067980e+00 4.34847742e-01 -1.36610246e+00
-1.15269840e+00 -5.34562767e-01 -2.48684227e-01 -3.06112975e-01
7.27708995e-01 -9.72898424e-01 -9.90904808e-01 7.81179845e-01
-7.24779844e-01 -4.45270121e-01 3.02323382e-02 3.45014542e-01
-3.84275228e-01 4.95553970e-01 -5.45264304e-01 -1.08140910e+00
-4.34900314e-01 -9.72161472e-01 9.58701253e-01 3.08526635e-01
4.44135666e-01 -7.63860345e-01 1.70494124e-01 5.91082633e-01
6.81952119e-01 -1.60018116e-01 7.84093142e-01 2.58399636e-01
-1.01364172e+00 6.27811849e-02 -6.81056976e-01 6.46721363e-01
8.14466402e-02 -2.15258426e-03 -1.36226571e+00 -3.83865237e-01
-1.82755385e-03 -2.08355457e-01 1.61063564e+00 6.52547359e-01
1.21087801e+00 3.24695967e-02 -2.29452908e-01 1.09752822e+00
1.55257881e+00 3.07427924e-02 1.35264671e+00 3.87808651e-01
8.40306520e-01 5.52213967e-01 4.98715013e-01 4.04685855e-01
7.51565471e-02 4.87888664e-01 7.02846766e-01 -9.00351346e-01
-7.17692912e-01 4.85444665e-01 5.09522855e-01 3.09395581e-01
-6.25918269e-01 -6.32821262e-01 -5.19641817e-01 1.69166118e-01
-1.44053698e+00 -1.08614051e+00 -3.36304039e-01 2.21243787e+00
5.25136590e-01 1.23539083e-01 -2.92553037e-01 6.37368783e-02
2.55615860e-01 3.61519456e-01 -4.25512105e-01 -2.86475360e-01
-3.98528725e-01 7.00845778e-01 8.72006357e-01 3.81374031e-01
-1.12809229e+00 8.62897515e-01 6.61718178e+00 4.99021828e-01
-1.17743981e+00 -2.47213885e-01 6.48525238e-01 -3.34698945e-01
-2.91398138e-01 -2.73165524e-01 -9.93969858e-01 -1.15769304e-01
5.97929776e-01 6.45195961e-01 3.90362203e-01 1.33848429e-01
2.84188747e-01 -2.71076202e-01 -7.27661610e-01 1.06550324e+00
6.35727406e-01 -1.00771999e+00 -1.46033570e-01 -3.84486109e-01
7.84308195e-01 1.44629657e-01 5.14192224e-01 4.46364172e-02
2.85363495e-01 -1.32119334e+00 -2.99433693e-02 9.59732354e-01
7.52754509e-01 -6.42139614e-01 5.50313175e-01 -2.71742374e-01
-9.65632975e-01 2.49694493e-02 -4.75140631e-01 2.83317357e-01
-2.40156837e-02 7.63061821e-01 -7.77275264e-01 6.88767672e-01
1.13433957e+00 7.85151362e-01 -6.59834981e-01 1.24487221e+00
-3.48087162e-01 7.07057714e-01 -3.28067988e-01 4.02187467e-01
-6.98301271e-02 -3.62383902e-01 5.82747281e-01 1.36914062e+00
1.91422909e-01 5.49157374e-02 -4.20108773e-02 3.86043429e-01
-8.54079798e-02 -2.00705737e-01 -6.65619016e-01 4.80401635e-01
-8.97338539e-02 1.29909575e+00 -7.34727740e-01 5.98911569e-02
-7.50753224e-01 1.12991869e+00 1.09618746e-01 8.37875366e-01
-8.01857650e-01 -3.41364503e-01 5.41153133e-01 2.29485616e-01
5.18055797e-01 -1.34280056e-01 -3.32891680e-02 -9.80168104e-01
3.03929865e-01 -9.42531884e-01 2.68902272e-01 -9.68385279e-01
-1.44805717e+00 5.87465167e-01 -2.39055812e-01 -1.26134670e+00
4.30306137e-01 -9.42246974e-01 -5.19403100e-01 6.63296163e-01
-2.22515845e+00 -1.21842945e+00 -5.86914122e-01 7.37716675e-01
4.25483793e-01 -1.40621336e-02 5.37459373e-01 5.03971279e-01
-4.12455022e-01 6.31625533e-01 1.68701351e-01 3.14919353e-02
1.04509866e+00 -9.18104231e-01 1.36078149e-01 1.08233202e+00
2.32478544e-01 5.93544185e-01 7.56927252e-01 -3.92828017e-01
-1.59047973e+00 -9.28620756e-01 2.05756038e-01 -1.02299690e-01
2.48120293e-01 -1.96683943e-01 -7.94322729e-01 5.15492976e-01
4.50059354e-01 2.06733719e-01 7.76417732e-01 1.88965335e-01
-7.02291310e-01 -4.53958839e-01 -9.86726344e-01 8.33486021e-01
1.42772532e+00 -2.84380704e-01 -2.54460350e-02 5.37914634e-01
2.30854705e-01 -5.09144783e-01 -4.49837297e-01 5.35486996e-01
7.03178108e-01 -1.54414201e+00 1.83144104e+00 -1.80916160e-01
5.86377323e-01 -2.21708015e-01 -1.36575595e-01 -1.20099497e+00
-2.35636279e-01 -7.14612603e-01 1.79451317e-01 8.43159020e-01
4.32260722e-01 -6.74891233e-01 8.05844426e-01 1.91681176e-01
-3.68217617e-01 -5.52626252e-01 -4.61181879e-01 -5.77886164e-01
-2.41433427e-01 -5.13934076e-01 -2.18132604e-02 3.59796822e-01
-7.82925725e-01 5.69326505e-02 -7.55492091e-01 5.27045310e-01
1.12605298e+00 3.09165925e-01 8.27060759e-01 -9.28785443e-01
-4.51356858e-01 -4.60426241e-01 -2.29000136e-01 -9.16854918e-01
9.51550752e-02 -6.19367361e-01 2.95460790e-01 -1.71227932e+00
3.15150440e-01 -4.84075755e-01 -4.42511350e-01 2.46293858e-01
-4.46875215e-01 9.30199146e-01 8.69676545e-02 -9.28550884e-02
-7.71762252e-01 3.41711968e-01 1.40813029e+00 -1.41093299e-01
-2.29640678e-01 4.58332337e-02 -8.52363646e-01 1.07711291e+00
6.83352232e-01 -3.66098940e-01 -5.57243884e-01 -5.38194358e-01
6.60734296e-01 -4.13742632e-01 6.06243849e-01 -1.14144659e+00
2.11227443e-02 3.82158943e-02 6.07510686e-01 -5.24556696e-01
6.10274792e-01 -7.30702341e-01 -7.17570782e-02 2.12921813e-01
-1.93184733e-01 -5.73494613e-01 2.85320282e-01 6.54650688e-01
7.25220293e-02 8.82340968e-03 8.80426586e-01 -1.17148340e-01
-6.05968297e-01 1.79719150e-01 -1.98709697e-01 1.96610555e-01
5.66361308e-01 -2.64576167e-01 -8.15716743e-01 -2.28897944e-01
-3.17139208e-01 -2.39518270e-01 1.66817784e-01 1.93299532e-01
1.06101096e+00 -6.29546583e-01 -9.85149622e-01 2.09776938e-01
-3.56055722e-02 -2.12139815e-01 5.72089434e-01 1.00441825e+00
-6.15436554e-01 1.76051240e-02 -9.75476801e-02 -5.29646277e-01
-1.82467341e+00 2.40805238e-01 4.99419272e-01 -2.21288696e-01
-1.15888751e+00 1.11801767e+00 5.83440959e-01 9.63810384e-02
2.42490724e-01 -2.26927385e-01 -2.68814892e-01 -2.59279460e-01
6.07757151e-01 3.99847716e-01 3.53034556e-01 -2.71454453e-01
-3.56613964e-01 1.04671896e+00 -2.75625497e-01 2.02034205e-01
1.66817749e+00 -2.68764257e-01 5.07546738e-02 1.50217161e-01
1.03420353e+00 3.61958921e-01 -1.31406701e+00 -2.58052081e-01
-6.43739760e-01 -7.91137278e-01 1.51421815e-01 -8.48614156e-01
-1.56705463e+00 8.22101653e-01 7.95525253e-01 1.42690077e-01
1.58259499e+00 -4.05269742e-01 5.86209655e-01 6.56458616e-01
1.68104753e-01 -8.12962770e-01 3.50893110e-01 4.39654887e-01
8.12509596e-01 -1.48469603e+00 5.77937841e-01 -8.50733876e-01
-5.61020076e-01 1.19482291e+00 4.76832688e-01 -3.26382995e-01
6.70795202e-01 5.18690467e-01 4.35406476e-01 -4.87298548e-01
-4.84462857e-01 -2.61573434e-01 5.51026702e-01 9.65618372e-01
7.57263899e-01 -3.39342535e-01 6.75058439e-02 -1.58884630e-01
-1.44978002e-01 -2.47025371e-01 6.21878922e-01 8.61502647e-01
-3.52409363e-01 -7.09279239e-01 -3.93850923e-01 3.09614748e-01
-7.36313581e-01 -6.41134739e-01 -3.55661064e-01 7.92987227e-01
6.03524074e-02 1.16785264e+00 -2.48282641e-01 9.19570960e-03
4.33594316e-01 -5.34798384e-01 9.89463985e-01 -4.46710050e-01
-8.10895205e-01 1.14455529e-01 2.62753844e-01 -8.05823684e-01
-9.08195496e-01 -4.07542497e-01 -9.66266453e-01 -5.84635977e-03
-2.24320874e-01 -6.10339165e-01 5.15410542e-01 6.83966279e-01
1.54556066e-01 9.29299057e-01 6.15369618e-01 -8.06307912e-01
-2.35867172e-01 -6.42321706e-01 -5.15620053e-01 6.50941551e-01
5.32848775e-01 -4.36477810e-01 -6.92655981e-01 1.33840933e-01]
|
[10.566685676574707, -2.7855827808380127]
|
a0690f56-9fb5-455e-a7a0-c4d23346252c
|
multi-view-vector-valued-manifold
|
1904.03921
| null |
http://arxiv.org/abs/1904.03921v1
|
http://arxiv.org/pdf/1904.03921v1.pdf
|
Multi-view Vector-valued Manifold Regularization for Multi-label Image Classification
|
In computer vision, image datasets used for classification are naturally
associated with multiple labels and comprised of multiple views, because each
image may contain several objects (e.g. pedestrian, bicycle and tree) and is
properly characterized by multiple visual features (e.g. color, texture and
shape). Currently available tools ignore either the label relationship or the
view complementary. Motivated by the success of the vector-valued function that
constructs matrix-valued kernels to explore the multi-label structure in the
output space, we introduce multi-view vector-valued manifold regularization
(MV$\mathbf{^3}$MR) to integrate multiple features. MV$\mathbf{^3}$MR exploits
the complementary property of different features and discovers the intrinsic
local geometry of the compact support shared by different features under the
theme of manifold regularization. We conducted extensive experiments on two
challenging, but popular datasets, PASCAL VOC' 07 (VOC) and MIR Flickr (MIR),
and validated the effectiveness of the proposed MV$\mathbf{^3}$MR for image
classification.
|
['DaCheng Tao', 'Yong Luo', 'Hong Liu', 'Chao Xu', 'Chang Xu', 'Yonggang Wen']
|
2019-04-08
| null | null | null | null |
['multi-label-image-classification']
|
['computer-vision']
|
[ 1.99836064e-02 -2.78742969e-01 -1.25348955e-01 -5.74513733e-01
-5.62334180e-01 -6.48675680e-01 4.27942395e-01 7.89804310e-02
-1.03489205e-01 4.38236564e-01 -2.58418173e-02 -9.82244387e-02
-4.66081321e-01 -4.91190404e-01 -7.63307810e-01 -8.46397996e-01
-1.16957344e-01 -2.61309624e-01 4.05493053e-03 -3.97366732e-02
3.08642715e-01 3.02634627e-01 -1.95712507e+00 3.45206976e-01
6.73485875e-01 1.15609956e+00 7.94444159e-02 4.28517342e-01
1.22189894e-01 8.31621289e-01 -1.59351885e-01 -4.68234986e-01
1.65362537e-01 -3.16577196e-01 -6.05389476e-01 5.89020193e-01
8.90893459e-01 8.97415504e-02 -1.45613641e-01 1.36262012e+00
1.74436104e-02 1.82373747e-01 1.04667807e+00 -1.55203283e+00
-7.59833634e-01 -2.24059328e-01 -8.79885852e-01 1.92940652e-01
2.50049233e-01 -2.16251522e-01 1.10862184e+00 -1.34078419e+00
6.87289357e-01 1.04803336e+00 3.42692286e-01 1.34828061e-01
-1.34072292e+00 -2.40936428e-01 2.75226891e-01 2.67097920e-01
-1.53246117e+00 -2.30963275e-01 8.88937116e-01 -8.76723528e-01
3.75779361e-01 6.24422669e-01 3.87695253e-01 8.27666938e-01
2.46673241e-01 5.90791404e-01 1.45662940e+00 -3.13073903e-01
2.84587070e-02 5.58364868e-01 3.62043709e-01 9.58186388e-01
-6.33496717e-02 -7.44299544e-03 -3.57812196e-01 -2.44174480e-01
7.27819085e-01 1.86968818e-01 -3.49175304e-01 -8.39828968e-01
-1.19371545e+00 9.45211887e-01 2.45831743e-01 1.13189537e-02
7.21268281e-02 -1.65392488e-01 3.42566818e-01 1.24696426e-01
3.92083615e-01 -1.06535099e-01 -2.16955602e-01 2.57244676e-01
-3.91008943e-01 -1.73665255e-01 3.44533056e-01 1.05749106e+00
1.06327391e+00 -6.34303689e-03 4.86584641e-02 1.07215858e+00
4.16431904e-01 6.00050390e-01 1.36914432e-01 -1.10778832e+00
4.54379737e-01 7.66108990e-01 1.35145942e-02 -1.44408643e+00
-4.26049978e-01 -2.11231634e-01 -1.04040170e+00 3.05427164e-01
4.20885533e-01 1.69571072e-01 -4.72610444e-01 1.66394246e+00
5.80410838e-01 -3.12548317e-03 -2.92946734e-02 9.68298018e-01
1.02328122e+00 3.15484852e-01 -9.46216807e-02 -7.94577301e-02
1.38383961e+00 -9.85095561e-01 -4.08729464e-01 9.52694565e-02
7.18230247e-01 -7.35469460e-01 1.13721919e+00 3.02061528e-01
-8.06532979e-01 -7.14887798e-01 -1.04350650e+00 1.37716025e-01
-3.82491022e-01 3.52193773e-01 4.85200644e-01 6.32026076e-01
-7.44314134e-01 4.63456243e-01 -3.27072769e-01 -2.43092880e-01
2.46519342e-01 1.29322901e-01 -8.75850916e-01 -4.37739342e-01
-6.45397663e-01 5.47378898e-01 4.37575132e-02 1.27916023e-01
-6.39106691e-01 -3.93452227e-01 -9.63500142e-01 -3.37725967e-01
3.69877398e-01 -3.69122088e-01 2.55803615e-01 -1.06598401e+00
-1.02007902e+00 1.13677633e+00 8.61222595e-02 2.83612847e-01
3.80920559e-01 3.74662206e-02 -6.51080787e-01 3.76886040e-01
1.84391215e-01 4.04536873e-01 1.13288426e+00 -1.52676284e+00
-5.20916343e-01 -8.13937664e-01 3.02956671e-01 1.65240586e-01
-3.58149797e-01 -3.35091986e-02 -1.47477329e-01 -5.62015116e-01
3.26648772e-01 -1.01551855e+00 7.72198811e-02 -8.25899765e-02
-4.49972719e-01 -1.09391570e-01 8.46521258e-01 -5.82014680e-01
9.39873934e-01 -2.53983045e+00 4.58745748e-01 2.41597384e-01
2.01190278e-01 -8.52104798e-02 -1.80394258e-02 2.64490783e-01
-2.51761347e-01 4.87490222e-02 -1.92105323e-01 -4.84496504e-02
-1.40381664e-01 7.68013969e-02 1.74532712e-01 1.00200188e+00
4.58455570e-02 6.19846106e-01 -7.58579314e-01 -6.28933430e-01
4.77160364e-01 5.41310191e-01 -2.17721790e-01 8.87134224e-02
9.43112001e-02 6.65671527e-01 -5.53158164e-01 7.51944542e-01
7.48095214e-01 -5.47728360e-01 -9.91986319e-02 -5.29957891e-01
-1.05656311e-01 -5.43218315e-01 -1.50296640e+00 1.59959722e+00
-9.68067944e-02 2.94861645e-01 3.50949615e-01 -1.17274690e+00
8.44397485e-01 1.45109981e-01 6.75836205e-01 -3.85886848e-01
1.06110185e-01 3.38725001e-02 -2.91878849e-01 -6.40400589e-01
2.54299819e-01 1.06236778e-01 -1.37436977e-02 4.39161748e-01
1.22998357e-01 2.01228037e-01 -1.61011964e-02 1.40112251e-01
8.18046927e-01 3.39814164e-02 2.32810393e-01 -6.62354648e-01
9.99502182e-01 -2.80866206e-01 4.97806609e-01 2.77587533e-01
-2.67101705e-01 6.84302568e-01 3.94917935e-01 -5.05372882e-01
-8.09602082e-01 -1.13090158e+00 -5.38770020e-01 1.22994411e+00
5.09761155e-01 -3.08214515e-01 -5.24269342e-01 -8.97209942e-01
1.54519314e-02 2.70382851e-01 -6.46336496e-01 -2.33239427e-01
-2.12639287e-01 -6.60309196e-01 2.95294344e-01 4.66215253e-01
3.20345491e-01 -7.09504783e-01 -5.45403361e-01 -1.47662640e-01
-2.46956319e-01 -1.12112677e+00 -6.52685821e-01 7.65150413e-02
-6.28510416e-01 -1.26481354e+00 -4.75305259e-01 -7.01779544e-01
8.11296701e-01 7.13077724e-01 1.02027655e+00 9.03575346e-02
-6.02376878e-01 1.00826800e+00 -4.39077258e-01 -4.16386537e-02
-1.01494335e-01 -4.78337407e-01 7.09907711e-02 7.92360306e-01
1.84150621e-01 -5.07517755e-01 -6.14243448e-01 6.26701891e-01
-1.03763187e+00 5.14669856e-03 2.14520842e-01 1.03182685e+00
8.22934747e-01 -9.44782719e-02 5.29186070e-01 -7.31074035e-01
1.23283766e-01 -3.39080453e-01 -4.41441178e-01 4.53742594e-01
-3.34499866e-01 -1.36761948e-01 5.50736308e-01 -4.34046894e-01
-7.46241391e-01 6.68939874e-02 3.67662609e-01 -7.23419309e-01
-4.19190019e-01 4.26895767e-01 -3.56486022e-01 -2.87427068e-01
4.73388940e-01 2.92846203e-01 2.93720956e-03 -3.58070523e-01
5.29746890e-01 4.88325715e-01 2.28693202e-01 -4.48915452e-01
5.58154941e-01 7.68475711e-01 2.07515419e-01 -1.06955910e+00
-8.08116376e-01 -5.99924207e-01 -8.57676923e-01 -5.44653416e-01
1.13348043e+00 -1.01962531e+00 -5.73010564e-01 2.84959286e-01
-8.24779928e-01 3.41013283e-01 -1.26579747e-01 6.69530988e-01
-6.58588409e-01 5.38148105e-01 -4.00182396e-01 -7.23444939e-01
3.92893590e-02 -1.20092547e+00 1.00701571e+00 1.05099037e-01
1.65739477e-01 -1.12379479e+00 -2.15452492e-01 7.14078605e-01
7.32016042e-02 5.08230269e-01 1.20167553e+00 -3.54531050e-01
-4.66464758e-01 -1.34242401e-01 -5.46481073e-01 7.15877771e-01
1.90474033e-01 5.74772507e-02 -1.02503800e+00 -4.89281714e-01
9.07336995e-02 -4.58486646e-01 8.01887870e-01 3.16762239e-01
1.26566422e+00 -2.10773982e-02 -1.18034311e-01 5.80901623e-01
1.55624461e+00 4.32159659e-03 3.92176360e-01 4.67484519e-02
1.01852953e+00 9.19415891e-01 6.17297113e-01 4.01135743e-01
4.51093614e-01 6.71188891e-01 6.24993205e-01 -1.05965734e-01
2.81969517e-01 9.83996689e-02 3.70088845e-01 9.06743348e-01
-2.04964265e-01 9.17971358e-02 -5.21107435e-01 2.91788429e-01
-1.76966238e+00 -7.85486758e-01 -3.57940406e-01 2.32538199e+00
1.05041847e-01 -2.19183922e-01 8.17950070e-02 9.24504846e-02
8.37635040e-01 3.15941036e-01 -5.03750920e-01 -2.87776858e-01
-3.35614026e-01 -1.96769595e-01 3.92745495e-01 1.80513069e-01
-1.34694374e+00 5.15711188e-01 4.79282522e+00 8.77314627e-01
-1.01044476e+00 4.23936956e-02 7.12823927e-01 1.97021723e-01
-3.01848203e-01 -5.95110729e-02 -5.38490653e-01 2.96956360e-01
4.17574584e-01 1.88483089e-01 3.39263052e-01 8.19503009e-01
-1.88101619e-01 -1.32032320e-01 -1.17562783e+00 1.18325984e+00
4.48705822e-01 -1.10902691e+00 -1.40015408e-01 2.97121763e-01
6.45574212e-01 -1.54783770e-01 4.07232374e-01 3.14955250e-04
-7.18240365e-02 -1.09045577e+00 6.86491013e-01 4.13736880e-01
7.96918213e-01 -6.31826818e-01 3.16784620e-01 4.11671847e-01
-1.45930111e+00 -1.34453014e-01 -3.80951136e-01 3.03375870e-01
-8.56463090e-02 4.20755714e-01 -2.47183964e-02 9.49225903e-01
7.83467174e-01 1.13340044e+00 -8.92037511e-01 4.86811012e-01
3.48787010e-01 1.74307134e-02 1.53523237e-01 3.79062772e-01
5.63941300e-02 -8.23787570e-01 5.94717860e-01 8.76837611e-01
1.54077277e-01 -6.41359612e-02 5.46131670e-01 8.20436895e-01
2.88289577e-01 5.21382093e-01 -8.35400701e-01 1.53133012e-02
-6.44877926e-02 1.51030421e+00 -7.92942405e-01 -9.38085432e-04
-8.05079281e-01 1.02594054e+00 4.71346200e-01 4.25155938e-01
-8.29046607e-01 -4.89592478e-02 5.72949946e-01 9.15202349e-02
3.95159841e-01 -2.92334616e-01 2.52369195e-02 -1.49005759e+00
2.48468250e-01 -7.37914562e-01 3.82577270e-01 -5.64568162e-01
-1.49353802e+00 5.81646025e-01 -6.39765412e-02 -1.64585280e+00
1.75349027e-01 -6.71586514e-01 -1.98064417e-01 7.41648376e-01
-1.24637043e+00 -1.35739338e+00 -3.28097224e-01 9.65237558e-01
2.33096793e-01 -2.83190757e-01 9.48472083e-01 4.04003620e-01
-5.42298138e-01 4.53021079e-01 3.99317533e-01 1.03013195e-01
5.88390291e-01 -1.14452863e+00 -4.80573446e-01 3.22667569e-01
3.65474761e-01 4.02890384e-01 2.04528809e-01 -2.22771794e-01
-1.69107652e+00 -1.19748044e+00 3.09967071e-01 -5.87902665e-01
4.56942767e-01 -3.00676525e-01 -7.40421295e-01 5.53683758e-01
-5.79626867e-05 6.24041736e-01 9.68590617e-01 -1.35857716e-01
-7.02387273e-01 -2.87952479e-02 -1.18824470e+00 2.85464168e-01
9.00810719e-01 -5.83555698e-01 -1.49654791e-01 2.62290597e-01
4.56061214e-01 5.02086915e-02 -1.22680950e+00 6.60353720e-01
5.97987533e-01 -1.37740135e+00 1.18501246e+00 -7.59078503e-01
2.69159526e-01 -3.18421870e-01 -9.01242912e-01 -9.99181390e-01
-4.04161483e-01 1.96411788e-01 8.49362090e-02 1.26638830e+00
1.76895440e-01 -4.87695336e-01 4.70944285e-01 2.90586710e-01
-2.76368912e-02 -8.87317777e-01 -1.05862534e+00 -7.33191073e-01
4.21133153e-02 -1.58471704e-01 1.39563844e-01 1.07145095e+00
-3.24276924e-01 3.46247047e-01 -4.77681786e-01 1.92450240e-01
9.22367334e-01 3.92111659e-01 7.21994281e-01 -1.38784778e+00
-3.15567225e-01 -9.37178656e-02 -7.43554771e-01 -6.15333676e-01
1.92552552e-01 -1.01023114e+00 -3.40161622e-01 -1.04633510e+00
4.93922114e-01 -6.99074745e-01 -4.93703216e-01 -8.38179588e-02
-7.23742917e-02 2.60595411e-01 2.60838956e-01 1.71935752e-01
-7.78132796e-01 6.32759035e-01 1.33348072e+00 -2.41221368e-01
9.13155824e-02 -1.38743177e-01 -5.55255949e-01 1.04210639e+00
4.09933478e-01 -1.58533469e-01 -4.19910461e-01 -2.70701081e-01
2.70595163e-01 2.80652225e-01 6.52044773e-01 -7.49683797e-01
-8.03737417e-02 -3.45244318e-01 4.13937122e-01 -4.69473302e-01
6.21760428e-01 -9.84077752e-01 1.38628811e-01 5.68996044e-03
-1.54217452e-01 8.74169096e-02 -1.89261332e-01 1.05943656e+00
-3.32002252e-01 -3.64095345e-02 9.71657634e-01 -2.16248751e-01
-6.42369866e-01 3.13109308e-01 -4.40704040e-02 2.07405940e-01
1.35923135e+00 -3.48387152e-01 -2.35160694e-01 -1.66900709e-01
-7.23672926e-01 1.52299359e-01 4.82155025e-01 5.55027485e-01
7.19802499e-01 -1.49701476e+00 -6.34794474e-01 5.35936475e-01
6.44584537e-01 -2.19015524e-01 5.51588595e-01 1.07079971e+00
-1.17407300e-01 3.30533721e-02 -3.27754915e-01 -1.00297654e+00
-1.45403004e+00 6.92359746e-01 3.56214851e-01 -6.79549873e-02
-6.32523119e-01 8.16182733e-01 5.80148578e-01 -5.44200957e-01
1.74898624e-01 -3.30191152e-03 -5.03471792e-01 2.00964391e-01
4.31078196e-01 6.28310978e-01 -8.67240950e-02 -1.28348505e+00
-3.29104036e-01 8.75658453e-01 5.57350740e-02 2.30071396e-02
1.18056786e+00 -3.57619941e-01 -3.42573553e-01 8.04201663e-01
1.60379183e+00 -1.06438458e-01 -1.14267719e+00 -2.95517325e-01
-2.01685857e-02 -5.68779469e-01 -2.95406163e-01 -2.83481210e-01
-1.07658422e+00 9.06886458e-01 7.89478481e-01 1.67207196e-01
8.35941136e-01 7.97244161e-02 2.77338207e-01 -7.87057653e-02
6.26109838e-01 -1.08625329e+00 3.81749541e-01 2.81170219e-01
7.99844742e-01 -1.52630115e+00 -4.57481891e-02 -7.56407380e-01
-7.62499571e-01 9.86710370e-01 5.96176505e-01 -2.23453417e-01
1.08816564e+00 -3.21967334e-01 8.76775663e-03 -3.69091928e-01
-3.26259017e-01 6.57771528e-03 5.59562624e-01 3.91133249e-01
2.86311418e-01 2.11562067e-01 -3.03484172e-01 5.69890618e-01
1.80902854e-01 -5.37274301e-01 4.07478243e-01 1.03059506e+00
-2.45646447e-01 -8.96971941e-01 -4.40184146e-01 5.93743980e-01
-4.64970917e-01 2.58591861e-01 -1.72286585e-01 6.49746001e-01
4.26905096e-01 1.10206628e+00 -1.80193663e-01 -5.59481025e-01
1.29998580e-01 2.53267102e-02 5.51707685e-01 -4.55675811e-01
-3.32977861e-01 1.34775668e-01 -2.17184141e-01 -5.14579356e-01
-7.80070126e-01 -7.17009246e-01 -9.52958345e-01 8.42214376e-02
-2.83865720e-01 -3.22468905e-03 5.57456613e-01 8.57601345e-01
2.34893933e-01 2.21877217e-01 9.17898476e-01 -6.44467652e-01
-5.09345651e-01 -6.83362067e-01 -1.16039205e+00 9.34034348e-01
2.79097199e-01 -1.03576386e+00 -4.63725656e-01 2.80161768e-01]
|
[8.659677505493164, 4.394859790802002]
|
b9f55108-f784-496d-b5cd-4894e57ed906
|
prema-predictive-maintenance-of-solenoid
|
2211.12326
| null |
https://arxiv.org/abs/2211.12326v1
|
https://arxiv.org/pdf/2211.12326v1.pdf
|
PreMa: Predictive Maintenance of Solenoid Valve in Real-Time at Embedded Edge-Level
|
In industrial process automation, sensors (pressure, temperature, etc.), controllers, and actuators (solenoid valves, electro-mechanical relays, circuit breakers, motors, etc.) make sure that production lines are working under the pre-defined conditions. When these systems malfunction or sometimes completely fail, alerts have to be generated in real-time to make sure not only production quality is not compromised but also safety of humans and equipment is assured. In this work, we describe the construction of a smart and real-time edge-based electronic product called PreMa, which is basically a sensor for monitoring the health of a Solenoid Valve (SV). PreMa is compact, low power, easy to install, and cost effective. It has data fidelity and measurement accuracy comparable to signals captured using high end equipment. The smart solenoid sensor runs TinyML, a compact version of TensorFlow (a.k.a. TFLite) machine learning framework. While fault detection inferencing is in-situ, model training uses mobile phones to accomplish the `on-device' training. Our product evaluation shows that the sensor is able to differentiate between the distinct types of faults. These faults include: (a) Spool stuck (b) Spring failure and (c) Under voltage. Furthermore, the product provides maintenance personnel, the remaining useful life (RUL) of the SV. The RUL provides assistance to decide valve replacement or otherwise. We perform an extensive evaluation on optimizing metrics related to performance of the entire system (i.e. embedded platform and the neural network model). The proposed implementation is such that, given any electro-mechanical actuator with similar transient response to that of the SV, the system is capable of condition monitoring, hence presenting a first of its kind generic infrastructure.
|
['T. V. Prabhakar', 'Vishwanath Shastry', 'Harsha Yelchuri', 'Prajwal BN']
|
2022-11-21
| null | null | null | null |
['fault-detection']
|
['miscellaneous']
|
[ 1.45655885e-01 1.15128227e-01 5.30808382e-02 -1.98053140e-02
2.31933683e-01 -5.65458119e-01 3.76318634e-01 2.61692643e-01
3.37739915e-01 3.17431062e-01 -8.71439517e-01 -7.08949745e-01
-5.54374874e-01 -8.26220393e-01 -7.93483078e-01 -7.09645748e-01
-1.09745413e-01 3.64916503e-01 3.36765110e-01 -1.22857414e-01
2.02447310e-01 7.65934706e-01 -1.68412745e+00 6.93430081e-02
4.30778176e-01 1.67429554e+00 3.40052038e-01 7.38278747e-01
4.84419376e-01 8.71749997e-01 -8.16589773e-01 3.66116703e-01
1.54800624e-01 8.88409242e-02 -6.80453658e-01 2.47484699e-01
-2.76464492e-01 -4.20501173e-01 -1.00314744e-01 9.24128652e-01
1.84081793e-01 -2.62676567e-01 5.34956872e-01 -1.47871673e+00
-7.74534121e-02 4.04542893e-01 3.20880264e-01 -1.95337176e-01
3.58449608e-01 5.56367934e-01 5.16118050e-01 -5.92993975e-01
3.94563943e-01 7.64111161e-01 6.97773397e-01 -5.82971983e-02
-9.37705398e-01 -1.07697181e-01 -1.81737557e-01 1.89104974e-01
-1.01776683e+00 -2.44083643e-01 7.20788777e-01 -6.77580416e-01
1.25905240e+00 5.37696540e-01 3.69870454e-01 1.04185534e+00
8.97260010e-01 3.14780235e-01 8.72029603e-01 -2.80938268e-01
7.41645813e-01 2.39958167e-01 -2.59263478e-02 5.59731841e-01
3.50138038e-01 4.55390587e-02 1.36240199e-01 -8.95183533e-02
9.54310060e-01 2.78730780e-01 -2.45924324e-01 -2.52628684e-01
-1.13071847e+00 6.41056746e-02 -9.39836875e-02 6.76789701e-01
-6.97688818e-01 5.14324866e-02 6.02845371e-01 6.26677692e-01
-2.39120230e-01 7.38784850e-01 -8.69913042e-01 -4.26107615e-01
-3.54546577e-01 -9.92177427e-02 1.26889622e+00 9.74876463e-01
2.33954445e-01 3.81485701e-01 1.25031322e-01 4.33183700e-01
3.27884018e-01 2.50336230e-01 5.06246686e-01 -7.46491909e-01
2.15195715e-01 7.52184570e-01 5.07925987e-01 -7.66593337e-01
-7.07951844e-01 -3.29468660e-02 -8.85546327e-01 6.81820512e-01
6.00493513e-02 -3.74892086e-01 -6.15946114e-01 8.60323489e-01
1.27105832e-01 5.18108085e-02 6.17473386e-02 7.04999208e-01
1.89380109e-01 6.37835443e-01 -2.52226174e-01 -4.61156577e-01
1.35207665e+00 -3.82659316e-01 -1.28320217e+00 9.04465765e-02
5.32541692e-01 -7.19239116e-01 8.24649513e-01 9.69033122e-01
-9.50090051e-01 -8.90736997e-01 -1.73127115e+00 6.37259305e-01
-6.75359666e-01 3.62458706e-01 3.90863627e-01 5.40625691e-01
-7.49621749e-01 1.45135665e+00 -1.11895561e+00 -9.77403596e-02
-1.64277449e-01 4.65782136e-01 -2.96534419e-01 2.59088010e-01
-1.32971096e+00 1.26758063e+00 3.29016030e-01 4.15201813e-01
-1.07115233e+00 -7.05347121e-01 -6.89340174e-01 -1.76055700e-01
1.95282072e-01 -2.02264324e-01 1.35748017e+00 -2.79500037e-01
-1.90544474e+00 2.73015141e-01 6.33022189e-01 -4.10647988e-01
7.02676535e-01 -2.55410105e-01 -9.88563716e-01 1.44281834e-01
-4.31968331e-01 -3.22045207e-01 9.69458818e-01 -9.22491670e-01
-6.88993454e-01 -3.02539915e-01 3.00178844e-02 -6.15416050e-01
-1.10702790e-01 -1.01495013e-01 2.95178115e-01 -1.72087491e-01
-1.97223336e-01 -7.98543572e-01 1.17132608e-02 -4.46562134e-02
-8.50624084e-01 -3.14315975e-01 1.25651491e+00 -8.24074805e-01
9.54518378e-01 -2.07614946e+00 -3.34425837e-01 4.31658208e-01
-2.29411229e-01 5.18664122e-01 6.66641772e-01 6.99644387e-01
-4.63871151e-01 -2.84999430e-01 -1.43600702e-01 1.32946238e-01
1.24124154e-01 2.86656350e-01 -1.39072929e-02 7.75216579e-01
5.70337117e-01 2.40227863e-01 -5.59768319e-01 1.71304673e-01
9.09139156e-01 3.30752045e-01 2.45252967e-01 3.41122687e-01
-1.91067025e-01 2.72606492e-01 -5.10096192e-01 8.96285295e-01
2.88989604e-01 4.69483845e-02 1.26681447e-01 -5.34265876e-01
-3.37932885e-01 2.37252474e-01 -1.61893213e+00 1.08933520e+00
-8.38058591e-01 3.78384858e-01 5.25534034e-01 -1.11562479e+00
1.30172026e+00 9.71203506e-01 8.39577079e-01 -4.11202580e-01
6.04262412e-01 4.37864393e-01 -1.54589713e-01 -1.17597079e+00
-1.52610987e-02 2.88005352e-01 1.61314085e-01 6.27414957e-02
4.73179109e-03 -3.11126590e-01 7.52887279e-02 -3.95563304e-01
1.52614474e+00 2.76155561e-01 1.69762611e-01 -4.46795762e-01
5.89073002e-01 -2.14415595e-01 3.93057525e-01 1.24997959e-01
-1.90723035e-02 6.09798469e-02 5.08880615e-01 -2.53729522e-01
-1.16602409e+00 -6.79136992e-01 -2.27920830e-01 3.03100407e-01
2.12385446e-01 -1.29175678e-01 -6.41438484e-01 -1.83194399e-01
2.75367260e-01 8.38828266e-01 -1.17957376e-01 -5.51699221e-01
-4.86347914e-01 -2.04939738e-01 1.92358047e-01 7.48219371e-01
1.40995786e-01 -9.76371169e-01 -9.45122063e-01 6.55995250e-01
5.32668769e-01 -1.13909125e+00 3.21858466e-01 6.46385372e-01
-1.04015779e+00 -1.50071800e+00 1.73189670e-01 -7.17393279e-01
4.76534039e-01 -3.99178505e-01 7.29750276e-01 9.59030986e-02
-5.60260475e-01 3.18869382e-01 -1.52701005e-01 -4.14947093e-01
-8.28894913e-01 -5.36332846e-01 5.44078052e-01 1.46822762e-02
-1.04943059e-01 -6.27382159e-01 -2.90046304e-01 6.20100975e-01
-7.93832600e-01 -4.29390937e-01 5.02150297e-01 3.99394006e-01
4.38317418e-01 8.83121610e-01 5.71762204e-01 -5.53951442e-01
6.28786087e-01 -5.03091812e-01 -9.09318924e-01 1.03507839e-01
-9.41237032e-01 -2.17585161e-01 1.13518727e+00 -4.50874805e-01
-7.08265901e-01 3.76744896e-01 -1.84327260e-01 -6.98565364e-01
-4.82964605e-01 3.22058171e-01 -5.76497316e-01 1.72202334e-01
3.27844501e-01 -3.32746387e-01 3.56605709e-01 -7.98371315e-01
-1.82939634e-01 1.09693992e+00 7.76584029e-01 -2.29064360e-01
4.74259555e-01 8.03370029e-02 1.97450116e-01 -8.50581467e-01
2.76627213e-01 -1.26758918e-01 -3.99123311e-01 -6.06082559e-01
4.90181148e-01 -5.46339452e-01 -1.50192940e+00 6.13656700e-01
-1.02183306e+00 -2.78522164e-01 -3.69233340e-01 4.63935792e-01
-4.95818496e-01 8.77725780e-02 -8.96937847e-01 -1.04781401e+00
-3.34911168e-01 -1.28763115e+00 9.03634250e-01 -9.63650830e-03
-2.81208187e-01 -9.24306095e-01 -5.26362181e-01 -7.86933079e-02
4.16481137e-01 7.27047861e-01 7.36337900e-01 -6.07354879e-01
-2.57765293e-01 -8.75212193e-01 5.25395513e-01 1.06278729e+00
7.17093110e-01 4.16327834e-01 -1.04129481e+00 -3.88197303e-01
6.48719609e-01 1.62191570e-01 -5.25100231e-02 1.60434201e-01
1.05498672e+00 -3.38494807e-01 -4.21642601e-01 -7.73283094e-02
1.38560712e+00 7.11375177e-01 5.79835832e-01 3.14888656e-02
5.03485084e-01 4.48079586e-01 1.00559342e+00 5.52334130e-01
-4.02221203e-01 5.57027519e-01 8.80126119e-01 8.75073150e-02
3.90881091e-01 1.45466983e-01 8.57973397e-01 6.39098167e-01
-1.23044245e-01 -9.72023532e-02 -5.24935365e-01 2.44519040e-01
-1.59560394e+00 -5.33469439e-01 -6.53428614e-01 2.22308159e+00
6.63792789e-01 5.22926033e-01 -1.92624450e-01 1.07508194e+00
8.66284430e-01 -5.28654397e-01 -6.30222023e-01 -7.10798800e-01
3.53591889e-01 -4.13703658e-02 6.56464100e-01 2.40488797e-01
-1.17353463e+00 -5.55294342e-02 5.34967613e+00 2.98467159e-01
-1.27312100e+00 -1.00726038e-01 1.16309449e-01 2.36655787e-01
4.16553140e-01 -2.67545044e-01 -5.40087581e-01 7.84208834e-01
1.33860624e+00 3.05058032e-01 2.99705803e-01 1.16218269e+00
4.90456581e-01 4.68484424e-02 -1.68698931e+00 7.66914725e-01
-4.59710538e-01 -1.13698936e+00 -8.53803158e-01 1.99385677e-02
2.01007515e-01 -2.75859356e-01 -4.27408487e-01 -3.13159898e-02
-1.02311902e-01 -4.99477237e-01 9.18792248e-01 6.19731247e-01
7.27210879e-01 -6.24268949e-01 1.05201185e+00 2.67873406e-01
-9.79284883e-01 -5.24194539e-01 -3.93424556e-03 -3.08089107e-01
4.06048149e-01 1.04529452e+00 -9.74308312e-01 7.64889121e-01
6.62373245e-01 3.47952008e-01 -1.83252722e-01 5.77457428e-01
-1.78237796e-01 4.54367012e-01 -5.14216065e-01 -7.11810812e-02
-8.72554854e-02 -3.68216448e-02 5.47492862e-01 6.67922378e-01
3.77884120e-01 -4.02594477e-01 1.81889728e-01 8.33011508e-01
5.53928733e-01 -3.85806382e-01 -5.86323142e-01 -8.81417766e-02
4.60343719e-01 1.36752391e+00 -4.82403964e-01 -3.65535438e-01
-1.70357972e-01 7.02237308e-01 -6.92742705e-01 -1.62153225e-02
-8.16537917e-01 -7.10008383e-01 7.38856733e-01 2.84160256e-01
3.11603129e-01 -1.02124721e-01 -2.76788414e-01 -2.73121446e-01
4.61660475e-01 -4.24391240e-01 -4.91054207e-02 -8.42463791e-01
-1.30837488e+00 2.16737330e-01 -4.74863976e-01 -1.58486748e+00
-5.15815854e-01 -1.26979578e+00 -5.91658533e-01 4.65736359e-01
-1.04930007e+00 -6.06812954e-01 -2.30924010e-01 6.25558376e-01
4.27657783e-01 -9.13957041e-03 8.52279186e-01 5.03598988e-01
-6.83145165e-01 -7.33130500e-02 1.99320957e-01 -1.77405283e-01
2.73560405e-01 -1.35014606e+00 9.61368456e-02 6.86682105e-01
-6.64270878e-01 2.34439924e-01 1.02930307e+00 -8.27731609e-01
-2.13802648e+00 -1.10608327e+00 4.59787756e-01 -1.00458279e-01
9.39575195e-01 -2.30498925e-01 -8.79425824e-01 5.43742478e-01
2.28479225e-02 2.13362634e-01 6.82073757e-02 -5.10320425e-01
4.46781635e-01 -3.76192570e-01 -1.50927866e+00 1.30858213e-01
4.08958077e-01 -4.81386423e-01 -5.69478691e-01 7.69933283e-01
5.16184926e-01 -4.91195649e-01 -1.67118490e+00 5.64138770e-01
2.72076219e-01 -9.15217102e-01 7.46101499e-01 -3.07654947e-01
2.07402512e-01 -4.86803412e-01 1.74650565e-01 -1.09568799e+00
-1.77814111e-01 -9.27630603e-01 -5.87132394e-01 1.33932531e+00
1.97093025e-01 -8.84045124e-01 1.79047614e-01 6.15296245e-01
-8.33332002e-01 -7.81417072e-01 -8.79601836e-01 -1.03753817e+00
-6.74416304e-01 -5.17502308e-01 6.71255708e-01 9.88900244e-01
5.21974206e-01 2.03496724e-01 2.28457525e-02 7.48487651e-01
-1.09858721e-01 -1.98110119e-01 2.84883916e-01 -1.49892545e+00
-2.61009336e-01 -1.35307431e-01 -6.67814732e-01 -1.52867004e-01
-4.15820032e-01 -3.34460109e-01 2.01986477e-01 -1.58118463e+00
-1.02444160e+00 -4.20165330e-01 -4.39132661e-01 7.40929663e-01
4.85716850e-01 -2.68128544e-01 -1.82928771e-01 -1.58817992e-02
2.62523443e-02 1.09597608e-01 9.78641570e-01 -2.44226545e-01
-1.57319993e-01 4.62494642e-01 1.54031008e-01 5.87088525e-01
1.01119792e+00 1.69642102e-02 -1.34548783e-01 2.24977955e-01
2.22652420e-01 3.23767334e-01 6.72508895e-01 -1.38289368e+00
3.01497102e-01 2.11971402e-01 3.24501723e-01 -3.02294910e-01
1.59172967e-01 -1.71022141e+00 6.23410165e-01 1.01760542e+00
2.43221819e-01 1.74554080e-01 -3.98275629e-02 2.25349531e-01
-6.96543753e-02 -3.97213638e-01 5.48317015e-01 2.71084040e-01
-4.89752442e-01 -2.00742543e-01 -7.64246047e-01 -9.53996301e-01
1.26722491e+00 -2.78882772e-01 -3.68792146e-01 1.27705470e-01
-6.14598691e-01 1.12759739e-01 2.68648475e-01 6.12247288e-01
2.67980754e-01 -1.05118918e+00 -1.06064372e-01 5.66071987e-01
-6.92115650e-02 1.03999317e-01 2.02156514e-01 9.62180734e-01
-5.08373082e-01 3.22195143e-01 -5.10374531e-02 -6.42776012e-01
-9.49149549e-01 8.43013883e-01 5.86203575e-01 1.20779209e-01
-8.24975133e-01 1.24413945e-01 -8.72832775e-01 -2.30354398e-01
2.62904882e-01 -1.03529108e+00 -1.20119572e-01 -1.11546770e-01
5.58507681e-01 8.40816259e-01 9.65878069e-01 -1.93221755e-02
-3.13138306e-01 1.99848011e-01 5.42027056e-01 4.77841347e-01
1.45430303e+00 2.25504518e-01 -1.89221263e-01 7.51248777e-01
9.12377417e-01 -4.05133754e-01 -1.37297928e+00 4.40405548e-01
-3.33869383e-02 6.91718608e-02 1.85664073e-01 -9.94965255e-01
-1.20423031e+00 4.70848829e-01 8.73015106e-01 1.19272900e+00
1.00480342e+00 -3.74200284e-01 6.11063778e-01 4.80710000e-01
5.09084463e-01 -1.62809277e+00 -2.46830747e-01 1.17897093e-01
9.70263779e-01 -6.12806082e-01 2.47412324e-02 -5.63971400e-01
-3.76290768e-01 1.32977760e+00 2.26147115e-01 -1.69068620e-01
8.60768199e-01 9.76909399e-01 -2.69606203e-01 -3.32592815e-01
-5.66206694e-01 5.66185534e-01 -2.87396193e-01 7.36283898e-01
-1.22305982e-01 2.43811190e-01 2.33887225e-01 7.60089993e-01
-3.39463986e-02 2.42215306e-01 4.19194460e-01 1.32947099e+00
-4.12479460e-01 -7.67138839e-01 -7.20437884e-01 5.19892335e-01
-4.16678011e-01 8.02048385e-01 3.34495027e-03 8.94840658e-01
3.48563105e-01 1.25684512e+00 2.47447073e-01 -5.78933597e-01
1.07693958e+00 1.94871292e-01 2.08156243e-01 -3.72233875e-02
-6.81294680e-01 -8.15314204e-02 4.43433195e-01 -1.03087711e+00
-1.84917171e-02 -5.72664678e-01 -1.44891357e+00 -1.04380414e-01
-5.56138933e-01 -2.32588798e-01 1.40471184e+00 1.05210102e+00
3.93526345e-01 1.33952463e+00 1.03422570e+00 -8.26355338e-01
-9.52227771e-01 -1.23282194e+00 -1.13777804e+00 2.29045585e-01
3.00934255e-01 -8.32622409e-01 -5.30349374e-01 2.28788733e-01]
|
[6.77626371383667, 2.3470122814178467]
|
bff8bf06-0f38-419c-81f4-a98a75887d31
|
basic-level-categorization-facilitates-visual
|
1511.04103
| null |
http://arxiv.org/abs/1511.04103v3
|
http://arxiv.org/pdf/1511.04103v3.pdf
|
Basic Level Categorization Facilitates Visual Object Recognition
|
Recent advances in deep learning have led to significant progress in the
computer vision field, especially for visual object recognition tasks. The
features useful for object classification are learned by feed-forward deep
convolutional neural networks (CNNs) automatically, and they are shown to be
able to predict and decode neural representations in the ventral visual pathway
of humans and monkeys. However, despite the huge amount of work on optimizing
CNNs, there has not been much research focused on linking CNNs with guiding
principles from the human visual cortex. In this work, we propose a network
optimization strategy inspired by both of the developmental trajectory of
children's visual object recognition capabilities, and Bar (2003), who
hypothesized that basic level information is carried in the fast magnocellular
pathway through the prefrontal cortex (PFC) and then projected back to inferior
temporal cortex (IT), where subordinate level categorization is achieved. We
instantiate this idea by training a deep CNN to perform basic level object
categorization first, and then train it on subordinate level categorization. We
apply this idea to training AlexNet (Krizhevsky et al., 2012) on the ILSVRC
2012 dataset and show that the top-5 accuracy increases from 80.13% to 82.14%,
demonstrating the effectiveness of the method. We also show that subsequent
transfer learning on smaller datasets gives superior results.
|
['Garrison W. Cottrell', 'Panqu Wang']
|
2015-11-12
| null | null | null | null |
['object-categorization']
|
['computer-vision']
|
[ 1.10769756e-01 1.41556323e-01 -1.44346535e-01 -4.94565964e-01
3.50366712e-01 -3.91072571e-01 5.29517114e-01 2.83435374e-01
-7.92111874e-01 1.61735058e-01 2.53383338e-01 -1.56515881e-01
-2.66948313e-01 -6.92350686e-01 -8.06394815e-01 -3.05122107e-01
-2.85026401e-01 5.82299046e-02 1.36255220e-01 -7.23990723e-02
3.30944508e-01 8.19014788e-01 -1.91063333e+00 7.88299978e-01
6.37276292e-01 1.22552872e+00 3.67911071e-01 7.20205426e-01
2.98193027e-03 9.01640236e-01 -3.06074589e-01 -2.81295508e-01
2.50006020e-01 -2.75867134e-01 -1.20916998e+00 -1.47742718e-01
5.93186796e-01 -3.99622232e-01 -3.02939713e-01 9.56246555e-01
1.84736192e-01 1.17592297e-01 7.11189091e-01 -1.01856279e+00
-1.02142024e+00 5.67400932e-01 -2.65822083e-01 4.48438078e-01
-1.36931807e-01 1.18202634e-01 1.23546576e+00 -9.28614080e-01
4.92206872e-01 1.13818395e+00 4.66918766e-01 8.19562197e-01
-1.35981750e+00 -5.30543506e-01 4.11452234e-01 6.39590621e-01
-1.12236249e+00 -3.88908356e-01 3.12786996e-01 -7.57855535e-01
1.32819116e+00 -1.44453689e-01 1.14414191e+00 6.91174209e-01
-1.51541131e-03 8.56549561e-01 1.10609257e+00 -3.98673743e-01
1.73628375e-01 5.08196838e-02 1.76182881e-01 7.25624740e-01
5.84576540e-02 3.55441183e-01 -4.83740449e-01 4.93478864e-01
7.00916946e-01 6.58595636e-02 -1.94276832e-02 -2.50933498e-01
-1.00792253e+00 8.23113501e-01 1.22698629e+00 5.34754336e-01
-4.99570727e-01 2.71079957e-01 3.09229791e-01 4.97402340e-01
1.27656937e-01 5.47154188e-01 -6.34230733e-01 1.90464914e-01
-8.82538259e-01 3.61885503e-02 4.98487294e-01 7.17479646e-01
6.00580871e-01 1.37137309e-01 -2.22697601e-01 1.10639453e+00
6.99104667e-01 -6.19502887e-02 5.02816617e-01 -1.12284458e+00
2.25537881e-01 6.51429474e-01 -5.57741106e-01 -8.68810058e-01
-6.18082523e-01 -5.33588827e-01 -8.22593331e-01 5.33312678e-01
4.31642085e-01 2.03679979e-01 -1.06569195e+00 1.86909401e+00
-9.48835015e-02 -3.05891752e-01 -5.83824925e-02 9.80524838e-01
9.16391313e-01 4.38387364e-01 3.43458056e-01 1.76344752e-01
1.18406439e+00 -9.16329443e-01 -1.26303807e-01 -3.53858799e-01
5.21728039e-01 -4.65684950e-01 6.72690392e-01 5.45153797e-01
-1.08808696e+00 -8.11571956e-01 -1.11430383e+00 -2.31510997e-01
-5.36521792e-01 1.14675015e-01 8.80951583e-01 5.30173302e-01
-1.29802787e+00 7.83654273e-01 -7.60346472e-01 -4.63831693e-01
1.33045459e+00 7.12897956e-01 -4.43817943e-01 -8.31424966e-02
-8.03888977e-01 1.10207152e+00 6.13420665e-01 5.35662733e-02
-1.23635709e+00 -8.63562226e-01 -6.11188650e-01 3.95355850e-01
-2.69625008e-01 -6.69209421e-01 1.04985249e+00 -1.41369224e+00
-1.38726509e+00 1.11616886e+00 3.66187431e-02 -6.10303044e-01
-1.37602240e-01 -2.10296698e-02 9.67317596e-02 3.23463976e-01
-1.93432972e-01 1.41670394e+00 8.03945601e-01 -9.38022196e-01
-7.81220496e-01 -5.22451282e-01 2.30490506e-01 -1.04140498e-01
-5.45427978e-01 2.45607048e-01 -3.41384001e-02 -5.08370161e-01
2.82176405e-01 -6.03828490e-01 -1.98341876e-01 3.22655380e-01
-5.84039427e-02 -5.44409454e-01 1.94525883e-01 -6.34727240e-01
8.11818123e-01 -1.98536241e+00 2.73100615e-01 7.88213089e-02
4.20067102e-01 4.63060260e-01 -4.60535586e-01 4.06890325e-02
-5.99788904e-01 1.45418391e-01 -2.33704790e-01 -9.31756347e-02
-1.52995288e-01 1.79732874e-01 -2.50347793e-01 5.54748833e-01
4.79118258e-01 1.06991208e+00 -7.15277612e-01 -9.15610120e-02
1.46207631e-01 3.92107576e-01 -1.01578188e+00 1.70150131e-01
-5.88814840e-02 3.29572231e-01 1.10919073e-01 5.89581311e-01
4.01515216e-01 -7.35385269e-02 -1.71265770e-02 -5.77137582e-02
-3.34682852e-01 4.53485012e-01 -6.47172511e-01 1.46814764e+00
-3.89368474e-01 1.09088445e+00 -2.59998571e-02 -1.67995489e+00
6.98171675e-01 2.83094734e-01 2.89909512e-01 -1.02421308e+00
4.18492973e-01 1.00311443e-01 8.09923053e-01 -4.07340109e-01
-2.17033550e-01 -1.69141516e-01 4.40156341e-01 3.18029284e-01
8.12660575e-01 2.55171843e-02 2.83729941e-01 -4.99164090e-02
7.62813926e-01 9.34416875e-02 2.57443309e-01 -3.18652064e-01
5.30541301e-01 -9.48109925e-02 2.05250114e-01 6.20034218e-01
-2.59043843e-01 6.88224196e-01 4.66810912e-01 -6.10270143e-01
-1.02560139e+00 -8.97342503e-01 -2.75217712e-01 1.44542444e+00
-5.18553972e-01 -2.19621584e-01 -7.60275662e-01 -3.09105217e-01
-8.23538080e-02 5.18797338e-01 -9.52791989e-01 -3.29975486e-01
-5.34889162e-01 -4.14256543e-01 4.93321717e-01 7.69065380e-01
4.24245328e-01 -1.45838523e+00 -9.02470529e-01 2.20844179e-01
5.22658348e-01 -9.60648656e-01 2.51277953e-01 2.53331393e-01
-1.01008797e+00 -1.08500957e+00 -7.59549916e-01 -9.84920800e-01
7.24566460e-01 2.50124186e-02 8.90658677e-01 4.45409149e-01
-6.62990212e-01 4.37383682e-01 -1.11099370e-01 -6.04863584e-01
-2.24310353e-01 1.17063522e-01 4.57005464e-02 -1.57455638e-01
2.89448917e-01 -5.76011419e-01 -5.57844698e-01 -5.10246456e-02
-7.29790628e-01 1.04341835e-01 5.21533966e-01 8.14840972e-01
3.21276516e-01 -2.87810415e-01 4.48361635e-01 -3.90738726e-01
2.02395737e-01 -4.98823225e-01 -8.20822418e-01 4.17921366e-03
-5.88480890e-01 -2.33446900e-02 7.21349657e-01 -2.99406499e-01
-6.69884741e-01 7.52494112e-02 -6.47440851e-01 -2.43760705e-01
-5.24740994e-01 4.08427268e-01 2.75384277e-01 -3.55899781e-01
5.03975153e-01 1.94370925e-01 -1.66833803e-01 -3.58067483e-01
3.47909391e-01 3.52332950e-01 3.09472352e-01 -3.77411455e-01
5.67262352e-01 4.79578853e-01 1.07040524e-03 -9.28588808e-01
-1.13267219e+00 -2.12478653e-01 -1.01257837e+00 -3.04712892e-01
1.24646783e+00 -6.36975229e-01 -1.17221928e+00 3.73081893e-01
-1.41338754e+00 -4.83818084e-01 -2.05296040e-01 6.43130004e-01
-6.47747517e-01 -2.38056704e-01 -5.90989351e-01 -4.34053957e-01
-1.04183733e-01 -1.08316672e+00 3.88459533e-01 3.84590298e-01
-1.67194977e-01 -9.41035330e-01 -5.06777093e-02 9.07312259e-02
4.71950799e-01 -1.80251062e-01 1.19138265e+00 -7.17610419e-01
-5.67504346e-01 1.26817033e-01 -5.41616380e-01 6.45531595e-01
-2.91836023e-01 6.01310879e-02 -1.24528074e+00 -1.44501060e-01
-1.50901616e-01 -5.33475757e-01 1.38347447e+00 4.11176324e-01
1.64128470e+00 -1.25942186e-01 1.66659299e-02 9.80837643e-01
1.48498583e+00 2.00809002e-01 6.00544930e-01 3.87774318e-01
6.25706911e-01 1.11544180e+00 -2.20090542e-02 -3.52747813e-02
4.42461580e-01 4.21134502e-01 7.54096150e-01 -2.78522726e-02
-3.73606801e-01 -8.59204680e-02 8.66806433e-02 6.71573818e-01
-4.04899240e-01 2.14319125e-01 -1.01734436e+00 8.54925334e-01
-1.47469699e+00 -1.00222635e+00 3.15600894e-02 1.84262455e+00
8.01032960e-01 -4.81090415e-03 9.24360240e-04 1.85541198e-01
3.43994498e-01 -1.27570599e-01 -3.08882892e-01 -8.89357805e-01
-2.35589165e-02 7.40659654e-01 2.22826928e-01 2.40442812e-01
-9.97254193e-01 1.14009261e+00 6.68138647e+00 3.30016553e-01
-1.16901660e+00 9.88640860e-02 6.45654678e-01 -2.38073114e-02
1.33423686e-01 -2.76761055e-01 -7.62300134e-01 2.46280842e-02
8.97173166e-01 1.44933954e-01 7.42525697e-01 7.62214482e-01
-1.98121741e-01 7.52026215e-02 -1.42744768e+00 8.38618517e-01
-4.93483394e-02 -1.37401283e+00 -7.22794887e-03 7.80379213e-03
7.52587318e-01 2.89511621e-01 2.23613128e-01 5.49664378e-01
1.16743572e-01 -1.56529498e+00 7.40756691e-01 4.35808808e-01
3.00161391e-01 -5.82846344e-01 3.72105777e-01 4.46832746e-01
-9.54862356e-01 -4.11230773e-01 -8.03128183e-01 -4.01035607e-01
-4.90771979e-01 1.39399201e-01 -7.42707372e-01 -2.95177221e-01
1.10340822e+00 8.21857214e-01 -6.62333429e-01 1.28606284e+00
-5.65686405e-01 7.14001298e-01 -1.35186791e-01 -2.69438237e-01
4.48731959e-01 2.09606916e-01 1.86906874e-01 1.26791739e+00
-1.19727708e-01 3.07329714e-01 -3.72351229e-01 1.26154947e+00
-2.39540994e-01 2.21873179e-01 -5.51149726e-01 -9.79311541e-02
1.37239927e-02 1.36787760e+00 -7.83056855e-01 -2.32617572e-01
-4.62211788e-01 5.89882731e-01 8.52994740e-01 3.20605308e-01
-4.39097792e-01 -2.39525154e-01 7.05396831e-01 -2.09031656e-01
7.66288817e-01 -2.90498346e-01 -6.63968265e-01 -8.59207869e-01
-4.26646173e-01 -7.92230487e-01 2.04641342e-01 -5.88056922e-01
-1.00250351e+00 4.71287340e-01 -1.66405961e-01 -6.46765769e-01
2.42725350e-02 -1.20957720e+00 -5.81797779e-01 9.65489805e-01
-1.39292562e+00 -1.00889158e+00 -6.95234463e-02 5.97742319e-01
5.59538305e-01 -2.85560250e-01 7.85579503e-01 3.01012486e-01
-3.48659009e-01 6.07789040e-01 -3.03579718e-01 5.50273478e-01
3.49478833e-02 -1.14331532e+00 1.93479031e-01 7.12087989e-01
4.10964489e-01 8.43638718e-01 1.87054127e-01 -7.94759765e-02
-1.11075997e+00 -9.70169663e-01 7.44828224e-01 -2.26644665e-01
5.71877897e-01 -5.83261192e-01 -1.02766585e+00 6.55648291e-01
4.16072994e-01 1.75118029e-01 6.01677120e-01 1.79662049e-01
-5.80892086e-01 -2.44736820e-01 -9.71153796e-01 5.13978481e-01
1.02116370e+00 -4.43098009e-01 -9.47934985e-01 1.84118152e-01
5.48325419e-01 -2.43228599e-02 -9.10109043e-01 3.45889181e-01
8.06236506e-01 -1.15911758e+00 1.05245948e+00 -1.18659461e+00
8.16367924e-01 -3.48192416e-02 -2.71466315e-01 -1.37598503e+00
-6.35962367e-01 3.06343436e-01 -6.51730597e-02 9.00785804e-01
3.97349507e-01 -6.12473607e-01 5.51017702e-01 2.26140663e-01
-4.27382618e-01 -1.05502200e+00 -8.24678242e-01 -4.58806843e-01
3.99315506e-01 -7.19508350e-01 2.31242195e-01 6.84175193e-01
-3.35270435e-01 3.08896333e-01 1.94671422e-01 6.40819594e-02
5.02492070e-01 1.12376504e-01 1.70496523e-01 -1.36349511e+00
-2.07071438e-01 -9.56863701e-01 -5.97882807e-01 -8.87665868e-01
2.76765317e-01 -1.35834253e+00 1.09031394e-01 -1.77354765e+00
1.23412691e-01 -2.65702069e-01 -6.17932022e-01 8.69287312e-01
2.32799917e-01 4.45461273e-01 5.59622943e-01 -5.46918400e-02
-3.69380474e-01 4.10039574e-01 1.18705750e+00 -1.94362655e-01
9.37420353e-02 5.83181605e-02 -8.05287302e-01 1.08677125e+00
8.01099122e-01 -5.88999391e-01 -9.95537564e-02 -8.00791383e-01
1.77895516e-01 -4.00187135e-01 5.90335965e-01 -1.10648298e+00
1.98747262e-01 2.45042294e-02 1.01788628e+00 -3.34991544e-01
9.02031437e-02 -7.98109710e-01 -6.54244483e-01 9.71182525e-01
-6.60535634e-01 -2.33979955e-01 3.96634758e-01 -7.06112981e-02
-1.75557733e-01 -4.96122003e-01 1.29680824e+00 -1.88035950e-01
-9.96942520e-01 5.59908688e-01 -5.55452645e-01 -7.84641653e-02
8.45786452e-01 -1.57127261e-01 -2.65650600e-01 4.85791862e-02
-1.00483131e+00 1.32010415e-01 -3.14628445e-02 5.22735775e-01
8.94035578e-01 -1.22825861e+00 -6.89002335e-01 5.10036230e-01
-7.42866024e-02 -3.80124748e-01 -4.21578921e-02 8.80189300e-01
-5.81643462e-01 6.74996018e-01 -6.45251989e-01 -8.18742692e-01
-1.03752089e+00 6.12295508e-01 5.13174236e-01 3.20804000e-01
-2.32917830e-01 1.44644523e+00 3.86226147e-01 -4.36642587e-01
3.29716653e-01 -4.08705652e-01 -7.97652900e-01 2.69660115e-01
6.77476168e-01 3.39569837e-01 -1.45189995e-02 -6.37925267e-01
-4.68327075e-01 6.88286602e-01 -4.82571358e-03 1.23843811e-01
1.80222178e+00 2.75384068e-01 -5.13521969e-01 2.70790279e-01
1.49249160e+00 -4.83294040e-01 -9.53870237e-01 -8.45469311e-02
1.26179799e-01 -2.40891829e-01 1.93851087e-02 -6.85880780e-01
-1.32638001e+00 1.53655040e+00 8.41719389e-01 3.00957769e-01
1.28493166e+00 1.80847451e-01 2.40442514e-01 5.17921746e-01
1.97211534e-01 -9.83438611e-01 1.74918935e-01 8.31739366e-01
1.13223016e+00 -1.20261848e+00 -6.96216300e-02 -3.73067074e-02
-3.14576536e-01 1.33327830e+00 8.20948601e-01 -5.37952840e-01
7.52136469e-01 -2.19960824e-01 -3.80929768e-01 -1.83702618e-01
-1.05253756e+00 -4.81273025e-01 7.69885778e-01 6.97557747e-01
6.13016069e-01 -1.06349848e-01 2.30337516e-03 5.74449062e-01
-1.80308178e-01 -1.11602090e-01 2.28927717e-01 6.08491123e-01
-8.19018841e-01 -7.19688714e-01 -9.52494293e-02 6.94217205e-01
-5.08364320e-01 -3.64493847e-01 -2.77270406e-01 6.90775037e-01
7.12752998e-01 6.18242025e-01 2.63565660e-01 -1.45579398e-01
3.70011657e-01 -7.38837421e-02 9.33061898e-01 -1.07746077e+00
-8.31392944e-01 -3.29520345e-01 -3.46234888e-01 -6.32107735e-01
-4.03218746e-01 -5.32690763e-01 -1.30396163e+00 5.77336363e-02
9.42546949e-02 -2.70119488e-01 8.23650301e-01 1.09317863e+00
-7.05848783e-02 6.89388514e-01 4.99278039e-01 -9.34594929e-01
-2.61048317e-01 -9.22295094e-01 -2.87742645e-01 5.23808561e-02
4.45528120e-01 -6.86588049e-01 -1.97831646e-01 6.65319934e-02]
|
[9.622639656066895, 2.3792426586151123]
|
8fc5fb13-419a-45f2-b6e7-a311d722a855
|
what-constitutes-good-contrastive-learning-in
|
2306.12086
| null |
https://arxiv.org/abs/2306.12086v1
|
https://arxiv.org/pdf/2306.12086v1.pdf
|
What Constitutes Good Contrastive Learning in Time-Series Forecasting?
|
In recent years, the introduction of self-supervised contrastive learning (SSCL) has demonstrated remarkable improvements in representation learning across various domains, including natural language processing and computer vision. By leveraging the inherent benefits of self-supervision, SSCL enables the pre-training of representation models using vast amounts of unlabeled data. Despite these advances, there remains a significant gap in understanding the impact of different SSCL strategies on time series forecasting performance, as well as the specific benefits that SSCL can bring. This paper aims to address these gaps by conducting a comprehensive analysis of the effectiveness of various training variables, including different SSCL algorithms, learning strategies, model architectures, and their interplay. Additionally, to gain deeper insights into the improvements brought about by SSCL in the context of time-series forecasting, a qualitative analysis of the empirical receptive field is performed. Through our experiments, we demonstrate that the end-to-end training of a Transformer model using the Mean Squared Error (MSE) loss and SSCL emerges as the most effective approach in time series forecasting. Notably, the incorporation of the contrastive objective enables the model to prioritize more pertinent information for forecasting, such as scale and periodic relationships. These findings contribute to a better understanding of the benefits of SSCL in time series forecasting and provide valuable insights for future research in this area.
|
['Tristan Sylvain', 'Lili Meng', 'Qi Yan', 'Chiyu Zhang']
|
2023-06-21
| null | null | null | null |
['contrastive-learning', 'contrastive-learning']
|
['computer-vision', 'methodology']
|
[ 4.58440810e-01 -3.57678473e-01 -4.01722729e-01 -5.21206558e-01
-5.55235565e-01 -5.55362940e-01 7.39759684e-01 3.35019916e-01
-1.21925071e-01 4.05971706e-01 1.97909668e-01 -5.10286033e-01
-3.84997696e-01 -5.68597496e-01 -5.40586352e-01 -8.32952023e-01
-6.32005990e-01 -3.86835188e-02 -2.72247612e-01 -4.43110466e-01
3.50958377e-01 7.52190173e-01 -1.76842463e+00 1.09497122e-01
8.22620451e-01 1.39681363e+00 2.14895710e-01 1.65093660e-01
-1.31382778e-01 1.00026560e+00 -4.12331343e-01 -4.67191748e-02
2.34049484e-01 -2.66376585e-01 -3.79735202e-01 -3.66863539e-03
2.59627193e-01 -1.38233334e-01 -2.15557605e-01 6.51960313e-01
3.74912024e-01 3.76684576e-01 5.20920992e-01 -1.12507033e+00
-6.55504048e-01 5.99286497e-01 -3.25332940e-01 6.47080600e-01
8.63331258e-02 1.56564593e-01 1.08660638e+00 -7.32207954e-01
2.69610167e-01 1.10786295e+00 8.59211147e-01 1.59466192e-01
-1.19863224e+00 -6.80159926e-01 5.35004497e-01 3.20182025e-01
-9.47168291e-01 -4.34522301e-01 1.27341616e+00 -5.15053868e-01
1.03377914e+00 1.88676417e-01 3.65890205e-01 9.43005443e-01
2.55540818e-01 7.34449208e-01 1.30829918e+00 -7.29583681e-01
3.53141308e-01 1.89075068e-01 1.72191516e-01 3.41740400e-01
-1.58299819e-01 5.37636936e-01 -5.88673174e-01 4.31318581e-02
5.96655607e-01 2.23592177e-01 -2.22496223e-02 -2.32440069e-01
-1.00668883e+00 8.55495632e-01 5.30048251e-01 4.49962914e-01
-4.35038686e-01 -1.41039535e-01 6.16941214e-01 5.51622272e-01
9.58412468e-01 6.30701959e-01 -5.49478292e-01 3.49795893e-02
-1.06382179e+00 -1.49174884e-01 4.41133767e-01 4.73269343e-01
5.48015416e-01 5.59275568e-01 5.35573199e-05 7.62376666e-01
1.18398249e-01 4.73493487e-01 6.49136662e-01 -7.21435487e-01
3.88633102e-01 5.38639963e-01 -8.18358064e-02 -1.21270847e+00
-3.93604279e-01 -7.88573921e-01 -8.57912183e-01 1.18721247e-01
4.40303050e-02 -4.07733507e-02 -6.70827866e-01 1.82001340e+00
8.38474184e-02 2.26647735e-01 1.97951645e-02 7.32774854e-01
4.96979773e-01 7.93307722e-01 2.65034497e-01 -5.47312021e-01
9.20832992e-01 -9.30594981e-01 -7.38295436e-01 -2.23307997e-01
6.20308578e-01 -6.29529774e-01 9.25779164e-01 1.15805171e-01
-8.03513169e-01 -7.90449083e-01 -9.23257172e-01 2.10532486e-01
-5.06791770e-01 2.30230257e-01 8.25400054e-01 2.29347020e-01
-8.47566009e-01 8.14816296e-01 -1.00570095e+00 -3.40657741e-01
2.45841578e-01 1.93733111e-01 9.57852080e-02 1.03273623e-01
-1.37641537e+00 1.04661083e+00 1.13769785e-01 1.75713629e-01
-4.26193237e-01 -9.12950277e-01 -7.99413860e-01 1.35824353e-01
1.28082901e-01 -1.92095354e-01 1.22201657e+00 -1.26897836e+00
-1.32872140e+00 4.62667197e-01 -1.46858066e-01 -9.45323408e-01
3.03371549e-01 -8.59153345e-02 -6.40690565e-01 7.37993121e-02
-9.78732333e-02 4.16340202e-01 8.95732582e-01 -8.55647266e-01
-5.69076836e-01 -3.66315871e-01 -1.99280947e-01 6.00930341e-02
-5.67415774e-01 2.51211505e-02 1.94909528e-01 -1.11990166e+00
-5.72293214e-02 -7.68201947e-01 -3.16711575e-01 -9.51157883e-02
1.97649181e-01 -3.89365852e-01 7.40246177e-01 -6.71657205e-01
1.54603922e+00 -2.34677625e+00 -3.76667059e-03 2.28884071e-01
-4.35059555e-02 2.39855006e-01 -2.52299041e-01 8.87519717e-01
-1.40646040e-01 -2.50894010e-01 -1.88323945e-01 -3.30324769e-01
-1.31650239e-01 1.31464675e-01 -1.11982536e+00 3.25988293e-01
4.59171414e-01 7.17633545e-01 -9.02599156e-01 -1.10419504e-01
4.63015884e-01 6.25162542e-01 -7.10175186e-02 1.29661113e-01
-2.65431374e-01 6.73041344e-01 -4.28369552e-01 5.34487665e-01
4.32590008e-01 -5.37948489e-01 6.96674511e-02 -1.30582303e-01
-4.63765204e-01 4.38200831e-01 -6.07853770e-01 1.20055521e+00
-6.09185874e-01 9.80721056e-01 -3.31491351e-01 -1.30038917e+00
1.22442734e+00 1.62969291e-01 7.10010529e-01 -1.06822693e+00
-6.39557987e-02 2.32434720e-01 -1.64737105e-01 -3.10892105e-01
2.44839519e-01 -1.13785855e-01 2.86415607e-01 3.83508980e-01
-1.33476764e-01 3.35938632e-02 1.34876266e-01 -2.08319008e-01
7.26578057e-01 1.49904251e-01 3.54028076e-01 -1.97047889e-01
4.72493917e-01 1.56772882e-01 4.29135114e-01 6.02877975e-01
-1.27756998e-01 2.33171806e-01 5.57449125e-02 -6.75962925e-01
-9.04613078e-01 -6.92673147e-01 -3.89473051e-01 1.37852907e+00
-9.72866863e-02 -1.45702824e-01 -5.75640723e-02 -3.67996067e-01
2.52488643e-01 8.72879148e-01 -6.29060209e-01 -2.79981017e-01
-6.33124113e-01 -6.29959702e-01 7.33104721e-02 8.25407803e-01
3.35859150e-01 -1.25808048e+00 -8.53017032e-01 2.91367590e-01
1.34879649e-01 -1.02103078e+00 -2.09710807e-01 4.20293987e-01
-1.43909383e+00 -6.83218360e-01 -5.53396046e-01 -8.08952272e-01
5.46999037e-01 5.26042759e-01 1.01375294e+00 -1.14725433e-01
-7.60525167e-02 5.16333699e-01 -4.29757625e-01 -4.69652146e-01
-3.55228484e-01 2.28475586e-01 -4.38930988e-02 -1.91766885e-03
2.91520178e-01 -8.05318952e-01 -5.64039052e-01 2.62410998e-01
-7.65204310e-01 -6.87907189e-02 4.41890746e-01 8.42822790e-01
4.95750219e-01 1.62530586e-01 1.02073371e+00 -7.84280837e-01
6.94130778e-01 -7.18169570e-01 -7.26451278e-01 3.76630098e-01
-1.12119746e+00 -3.29655292e-03 9.53894377e-01 -4.87177849e-01
-1.03757203e+00 -2.49235928e-01 8.98818225e-02 -6.11086607e-01
5.59182912e-02 1.03054690e+00 6.60424352e-01 -3.59344222e-02
6.73221171e-01 4.40296143e-01 2.43119046e-01 -4.72751200e-01
1.88939199e-01 4.15544420e-01 3.32244992e-01 -3.26760888e-01
5.80011964e-01 4.57493335e-01 7.67320255e-03 -8.55213463e-01
-1.28877664e+00 -5.68899691e-01 -4.43348944e-01 -3.00077766e-01
3.10127795e-01 -9.68798220e-01 -3.16954434e-01 3.26216578e-01
-6.51590407e-01 -3.63970667e-01 -4.45051998e-01 5.95002115e-01
-4.17490542e-01 1.31987363e-01 -4.37395424e-01 -1.13798034e+00
-4.00855660e-01 -8.31090510e-01 9.32043731e-01 1.84834048e-01
-2.00894058e-01 -1.47723484e+00 4.02552821e-02 1.36532292e-01
8.09638977e-01 2.38079488e-01 1.02176750e+00 -7.12747514e-01
-2.62090176e-01 -2.40348667e-01 5.44044673e-02 4.39299583e-01
3.60947549e-01 -9.61384848e-02 -1.17619228e+00 -5.69625080e-01
1.15354814e-01 -3.01949322e-01 9.63224530e-01 5.31449497e-01
1.13775909e+00 -2.24713653e-01 -1.59124553e-01 5.38207293e-01
1.24898577e+00 4.02743667e-01 2.29597896e-01 5.76480985e-01
4.40461338e-01 8.45243990e-01 7.49940097e-01 6.11286402e-01
3.83743733e-01 4.47916776e-01 3.44926178e-01 -1.50077254e-01
-2.64737252e-02 -2.15692714e-01 3.37934494e-01 1.16634166e+00
-3.82818133e-02 1.25092873e-02 -9.75755453e-01 3.85951221e-01
-1.73418939e+00 -1.08852935e+00 4.33525831e-01 2.23039055e+00
6.71399653e-01 1.25438944e-01 1.13221161e-01 3.30953300e-01
5.97232759e-01 5.16380906e-01 -1.03738713e+00 -2.32737377e-01
-3.41153890e-01 3.85037400e-02 3.80009830e-01 1.55117184e-01
-1.38301575e+00 6.91624045e-01 7.03444815e+00 6.32612586e-01
-1.88297844e+00 -3.20033044e-01 7.80755520e-01 1.14733711e-01
-3.94547075e-01 -1.03520177e-01 -5.77581942e-01 5.42182803e-01
1.11550498e+00 -2.67948031e-01 6.78178549e-01 8.56749475e-01
4.44524080e-01 2.34699950e-01 -1.02858007e+00 7.85227239e-01
9.83350426e-02 -1.52804816e+00 -7.97313154e-02 -1.66123331e-01
7.77590930e-01 2.52371311e-01 4.05576646e-01 5.47061503e-01
-1.21778697e-01 -8.66194487e-01 7.24175036e-01 5.06466806e-01
5.61176896e-01 -5.22687376e-01 6.29566073e-01 3.15857798e-01
-1.25912511e+00 -5.00672638e-01 -1.00644857e-01 -4.00089592e-01
-7.04426169e-02 5.03830492e-01 -7.27152050e-01 5.43846309e-01
5.18152833e-01 1.14151335e+00 -4.56187546e-01 7.71662176e-01
1.06751226e-01 1.09389865e+00 -1.01238318e-01 -1.23872105e-02
2.50065655e-01 -3.27839077e-01 4.46849495e-01 1.11109853e+00
2.59461075e-01 -1.27285212e-01 2.23147273e-01 5.29022932e-01
2.90899575e-01 1.76410377e-01 -5.44632256e-01 -3.93770188e-01
6.79829240e-01 8.68998289e-01 -6.49994612e-01 -2.25693539e-01
-4.70618188e-01 2.85762846e-01 3.79298985e-01 5.14270425e-01
-5.20884514e-01 -4.60519902e-02 3.22397172e-01 8.97013322e-02
5.34002602e-01 -3.41567695e-01 -4.05405879e-01 -9.55517709e-01
1.70265883e-01 -8.70514512e-01 4.75925773e-01 -5.89489520e-01
-1.61234641e+00 7.24682391e-01 5.48034050e-02 -1.59793591e+00
-4.79082972e-01 -5.05245388e-01 -6.28141522e-01 6.10716283e-01
-2.01189804e+00 -1.01996934e+00 -1.28998339e-01 1.58299312e-01
7.92889118e-01 -4.74735439e-01 7.85455406e-01 1.91327319e-01
-5.70441365e-01 4.11075532e-01 5.55720568e-01 -3.08238775e-01
5.59381783e-01 -9.77059603e-01 3.87405097e-01 6.07233346e-01
1.52730137e-01 5.97667396e-01 6.79054797e-01 -3.32179666e-01
-1.35475588e+00 -1.14880943e+00 8.17575157e-01 -6.59435019e-02
8.48811150e-01 5.80150709e-02 -1.01637959e+00 5.01128852e-01
-1.01561723e-02 4.98100817e-02 7.58953571e-01 2.62342304e-01
-4.42752123e-01 -4.96025681e-01 -8.07891190e-01 4.54684496e-01
7.25225687e-01 -7.32016087e-01 -5.10818779e-01 2.41534546e-01
5.80858052e-01 -9.23497230e-02 -8.56735289e-01 5.51064372e-01
5.75808465e-01 -7.99176991e-01 9.55597937e-01 -5.33417225e-01
5.36489606e-01 3.10236607e-02 -1.92859799e-01 -1.32984531e+00
-4.55192208e-01 -5.50810695e-01 -3.49036247e-01 1.19838059e+00
2.18893930e-01 -9.16925371e-01 3.47026855e-01 3.99573594e-01
-1.91039383e-01 -1.00903666e+00 -6.82536244e-01 -8.71472955e-01
8.91614705e-02 -4.21095908e-01 5.12659550e-01 1.16672599e+00
-1.37540370e-01 2.13962570e-01 -4.37408686e-01 -1.44027490e-02
4.54162478e-01 5.96699655e-01 3.10097188e-01 -1.38572240e+00
-2.80002668e-03 -4.78398204e-01 -3.05353820e-01 -9.95971441e-01
3.57709408e-01 -8.68405044e-01 -2.54977435e-01 -1.24118137e+00
-2.54118532e-01 -6.15104437e-01 -8.19607198e-01 2.93397129e-01
-8.15094709e-02 -2.93957200e-02 3.08857977e-01 6.43848836e-01
-4.20862138e-01 7.32375383e-01 1.09659386e+00 -8.89265686e-02
-2.31317133e-01 3.52863282e-01 -5.54296076e-01 5.93586147e-01
9.57941651e-01 -2.02371404e-01 -6.85926378e-01 -4.35014397e-01
5.44437133e-02 2.15541661e-01 1.45570576e-01 -7.25998342e-01
2.18587771e-01 -2.15324834e-01 3.11506957e-01 -6.82178795e-01
1.92822486e-01 -8.91979575e-01 -2.25639090e-01 3.38401586e-01
-7.08236158e-01 3.11456203e-01 4.62760448e-01 5.86183012e-01
-5.26463389e-01 6.67831227e-02 5.15811741e-01 1.30749092e-01
-9.15682733e-01 2.09664311e-02 -2.92351037e-01 -8.89302865e-02
8.69419813e-01 -2.73017585e-01 -1.34648383e-01 -4.79828060e-01
-4.27905232e-01 2.92398602e-01 1.04621038e-01 5.18661320e-01
3.10823798e-01 -1.19955504e+00 -5.36107838e-01 4.89357889e-01
1.08308941e-01 -4.00849164e-01 3.37794244e-01 8.52864087e-01
-4.48732153e-02 5.88206172e-01 -3.08854461e-01 -6.56764507e-01
-8.77031922e-01 6.31529093e-01 1.71565175e-01 -3.56077760e-01
-6.00623667e-01 5.09673834e-01 1.33238554e-01 -2.42656946e-01
5.86534560e-01 -4.31001753e-01 -3.84789228e-01 2.49860793e-01
4.08977568e-01 4.87310827e-01 2.27597907e-01 -4.87997115e-01
-3.71026367e-01 5.69453418e-01 -4.66545373e-02 6.63044527e-02
1.71847343e+00 -1.76859155e-01 1.52960092e-01 8.55656028e-01
1.05533850e+00 -1.73144832e-01 -1.58607805e+00 -4.39304560e-01
1.95088342e-01 -2.34256193e-01 2.66336948e-01 -8.31228256e-01
-1.03282738e+00 8.45536411e-01 6.68170929e-01 6.10473752e-01
1.48284435e+00 -2.65234917e-01 5.57604611e-01 3.05768996e-01
2.78033257e-01 -9.36226070e-01 6.88708853e-03 5.65589428e-01
1.01809382e+00 -1.48697400e+00 8.09386447e-02 -1.14930868e-01
-6.60566151e-01 1.23115766e+00 2.55755574e-01 -2.69728899e-01
7.85711706e-01 1.07516579e-01 2.59239644e-01 -5.57556339e-02
-1.05666292e+00 -1.46284595e-01 6.32164359e-01 3.21909845e-01
6.24420285e-01 1.49455306e-03 -9.43918750e-02 2.93595880e-01
-2.57191479e-01 -4.80831601e-02 -1.51384741e-01 1.06238520e+00
-1.02999903e-01 -9.19624507e-01 -1.29595861e-01 5.68648756e-01
-2.30739459e-01 1.03676561e-02 -6.21504039e-02 5.74468136e-01
-3.70225638e-01 9.44162786e-01 2.63310373e-01 -2.68063217e-01
2.65518546e-01 2.73898374e-02 2.29691744e-01 -3.56110185e-01
-7.53031433e-01 8.96463320e-02 -1.81870669e-01 -2.35985875e-01
-6.66460454e-01 -7.52160072e-01 -9.88150775e-01 -2.64962818e-02
-3.78962934e-01 5.98113276e-02 6.33807838e-01 1.02090740e+00
5.65938175e-01 6.68867171e-01 1.22909188e+00 -8.24256897e-01
-9.96010721e-01 -8.94106686e-01 -3.50523144e-01 4.07343060e-01
6.80536091e-01 -7.15602040e-01 -6.90349102e-01 1.80035591e-01]
|
[7.05460786819458, 3.0108392238616943]
|
6a08f2ea-73e9-4193-92e0-61f195c562a5
|
a-hassle-free-machine-learning-method-for
|
1808.04694
| null |
http://arxiv.org/abs/1808.04694v1
|
http://arxiv.org/pdf/1808.04694v1.pdf
|
A Hassle-Free Machine Learning Method for Cohort Selection of Clinical Trials
|
Traditional text classification techniques in clinical domain have heavily
relied on the manually extracted textual cues. This paper proposes a generally
supervised machine learning method that is equally hassle-free and does not use
clinical knowledge. The employed methods were simple to implement, fast to run
and yet effective. This paper proposes a novel named entity recognition (NER)
based an ensemble system capable of learning the keyword features in the
document. Instead of merely considering the whole sentence/paragraph for
analysis, the NER based keyword features can stress the important clinic
relevant phases more. In addition, to capture the semantic information in the
documents, the FastText features originating from the document level FastText
classification results are exploited.
|
['Liu Man']
|
2018-08-10
| null | null | null | null |
['clinical-knowledge']
|
['miscellaneous']
|
[ 2.90930510e-01 2.50511199e-01 -3.71882200e-01 -4.03049618e-01
-7.40104198e-01 -2.98095912e-01 5.79094172e-01 1.13452077e+00
-9.23621714e-01 9.53082383e-01 4.83143091e-01 -4.31612223e-01
-8.47592652e-01 -7.03213394e-01 2.39716247e-01 -6.41171157e-01
1.77815467e-01 3.05044740e-01 -3.44833732e-03 -3.84811401e-01
6.85846865e-01 4.75639671e-01 -1.14017284e+00 6.33514166e-01
8.42050076e-01 5.73993802e-01 2.54608184e-01 7.78175235e-01
-6.87806308e-01 1.18954778e+00 -6.32351458e-01 -1.90679446e-01
-5.01674652e-01 -5.13408184e-01 -9.32243466e-01 -5.37116006e-02
-1.34844586e-01 2.41400123e-01 -2.01079115e-01 7.31902540e-01
6.06231689e-01 -6.56489730e-02 8.80262434e-01 -4.71589804e-01
-1.84157193e-01 7.02395618e-01 -2.50834346e-01 4.87662822e-01
5.45860112e-01 -5.61500072e-01 6.45942032e-01 -7.26910412e-01
8.64748180e-01 6.40186965e-01 6.15990400e-01 3.51424992e-01
-5.53798914e-01 -4.52096105e-01 -1.66500062e-01 2.82332242e-01
-1.13008642e+00 -1.73582971e-01 7.76235938e-01 -5.86195290e-01
1.05957079e+00 3.44941765e-01 3.56493831e-01 8.84202778e-01
9.50557351e-01 3.93729270e-01 1.17971551e+00 -8.95836532e-01
7.74103627e-02 6.03573918e-01 7.41746306e-01 8.08602989e-01
2.58364618e-01 -1.57159582e-01 -2.09852576e-01 -4.09996063e-01
5.37494645e-02 4.18512613e-01 -2.08054423e-01 2.75038958e-01
-9.23542857e-01 8.01241815e-01 1.69502329e-02 1.05570030e+00
-3.99915189e-01 -6.26442730e-01 9.90408063e-01 1.49964973e-01
2.52628237e-01 5.25322616e-01 -7.12194264e-01 -2.43065357e-01
-1.14579844e+00 -3.72986466e-01 1.08863175e+00 7.36173153e-01
1.72455817e-01 -3.70812267e-01 -3.77329826e-01 4.73928452e-01
1.78012028e-01 1.65169969e-01 1.09077120e+00 -1.59607261e-01
4.44542646e-01 1.00748789e+00 -2.61856318e-01 -1.13036501e+00
-7.83306897e-01 -5.09384811e-01 -7.70115793e-01 -4.00773436e-01
-9.66943726e-02 -2.93112129e-01 -1.10235929e+00 9.00472939e-01
1.55128762e-01 -3.64451796e-01 4.18645620e-01 2.26346746e-01
1.27416897e+00 2.52688706e-01 6.56274140e-01 -4.39879626e-01
1.64956522e+00 -7.08319843e-01 -1.23917544e+00 2.69712240e-01
1.09183180e+00 -1.24977052e+00 4.22828346e-01 3.89571369e-01
-5.98438680e-01 -2.66678035e-01 -9.03792799e-01 2.79908068e-03
-1.05672407e+00 4.99469757e-01 6.24609351e-01 6.16069198e-01
-7.05431759e-01 7.14172840e-01 -5.03052235e-01 -7.42132187e-01
1.43573552e-01 5.34624279e-01 -5.50890148e-01 1.14407174e-01
-1.08003175e+00 1.09111547e+00 8.22227001e-01 -1.48682380e-02
-1.23019204e-01 -3.02336514e-01 -8.21843565e-01 6.82462677e-02
1.49585262e-01 -6.04256570e-01 9.08221364e-01 -8.63152802e-01
-1.33910823e+00 7.79138148e-01 -8.48312527e-02 -2.25117087e-01
2.73247540e-01 1.07754312e-01 -8.39143455e-01 6.08485758e-01
-3.26581560e-02 1.92873925e-02 4.81283426e-01 -6.00641966e-01
-6.34438276e-01 -5.51008523e-01 -5.35608172e-01 2.17680037e-01
-6.21244013e-01 1.72763050e-01 -1.84907187e-02 -7.37463415e-01
-5.42464491e-04 -5.21957874e-01 -2.95433730e-01 -5.33881009e-01
-5.08743167e-01 -4.35245782e-01 6.12251759e-01 -6.73389554e-01
1.50940931e+00 -2.20136523e+00 -4.10601616e-01 3.62315416e-01
4.18177247e-01 2.07415894e-01 3.47779691e-01 1.03606296e+00
-1.26871631e-01 2.58212566e-01 1.42511964e-01 3.06935132e-01
-3.68495762e-01 2.02235639e-01 2.05390573e-01 2.49533996e-01
-8.91665965e-02 5.73149502e-01 -8.74615967e-01 -1.22803926e+00
2.62196928e-01 3.66721153e-01 -6.47234917e-02 -4.33804654e-02
3.96643400e-01 9.85235348e-02 -9.19311047e-01 5.28955340e-01
1.16907284e-01 -1.87283158e-01 3.90103608e-01 -2.36796543e-01
-1.42073944e-01 1.54006600e-01 -7.43962705e-01 1.43431711e+00
-4.76213455e-01 4.27254617e-01 -3.50544482e-01 -1.09056962e+00
7.94316947e-01 7.98628926e-01 6.85580075e-01 -3.58878374e-01
4.27320421e-01 3.41321111e-01 -1.18428566e-01 -1.01214969e+00
2.56878287e-01 -1.57452360e-01 -7.79779628e-02 -7.15294480e-02
2.45746270e-01 2.90214777e-01 8.38228390e-02 3.25000167e-01
1.21925211e+00 -2.26967469e-01 1.15839279e+00 -3.69592339e-01
7.35969186e-01 3.54252547e-01 2.04901308e-01 5.94907343e-01
-1.60068512e-01 2.98470408e-02 1.23816572e-01 -2.58285135e-01
-7.09827781e-01 -4.77517396e-01 -6.16682708e-01 7.45729685e-01
-1.43954024e-01 -6.95671976e-01 -6.36046886e-01 -1.05070448e+00
-2.74392992e-01 5.08327127e-01 -8.45804989e-01 -3.20020206e-02
-3.29189301e-01 -5.44210553e-01 6.42546773e-01 4.94727015e-01
2.54364401e-01 -1.11177111e+00 -6.03994191e-01 5.29574275e-01
6.10188246e-02 -9.48869586e-01 -3.00843477e-01 6.53218687e-01
-1.29109097e+00 -1.21899033e+00 -6.57967865e-01 -9.12149608e-01
7.74937332e-01 -2.26334274e-01 7.59123206e-01 -2.31830701e-02
-6.85636997e-01 3.85637879e-01 -6.13405585e-01 -6.74921274e-01
-3.25978398e-01 3.17886680e-01 -4.04540062e-01 -3.44292372e-01
9.28009987e-01 -6.97914958e-02 -4.33989972e-01 -2.64496684e-01
-8.87417078e-01 -2.90874153e-01 9.25232589e-01 1.11397505e+00
4.43288624e-01 2.41590664e-01 6.31167471e-01 -1.45193028e+00
7.83636212e-01 -7.03687668e-01 4.23342317e-01 4.78267789e-01
-1.00669098e+00 2.68284589e-01 6.39861524e-01 -2.04123527e-01
-1.11315715e+00 6.10475466e-02 -3.64699692e-01 3.89884502e-01
-7.26960003e-01 8.85895252e-01 3.06250662e-01 -2.32205670e-02
8.07666659e-01 4.76641029e-01 -3.69117916e-01 -5.66701889e-01
-1.34996980e-01 1.16863227e+00 -4.76083340e-04 -6.69952482e-02
2.28060350e-01 1.58437788e-01 1.33529067e-01 -9.69078064e-01
-8.41330826e-01 -1.07604289e+00 -6.38404369e-01 -1.01664580e-01
1.04799116e+00 -4.83166963e-01 -4.80106801e-01 -6.13298416e-02
-8.86092901e-01 5.93965054e-01 -6.78496957e-02 7.98060000e-01
-3.05237383e-01 6.17303371e-01 -7.41280615e-01 -8.54232490e-01
-6.93617582e-01 -6.57257676e-01 8.98272812e-01 3.79196852e-01
-3.87041360e-01 -1.22613716e+00 1.46959201e-01 2.57258326e-01
2.29296654e-01 2.69289285e-01 1.25114799e+00 -1.41349411e+00
3.26026738e-01 -6.67813063e-01 -5.24764843e-02 -8.36746246e-02
4.79741156e-01 -1.90689415e-03 -6.37644947e-01 9.02241692e-02
2.31633216e-01 6.14075512e-02 7.87571430e-01 1.79062873e-01
1.01623154e+00 -4.89007384e-01 -7.43442416e-01 1.37297153e-01
1.64067566e+00 5.50610423e-01 3.35183531e-01 5.66693068e-01
3.74292761e-01 6.29736960e-01 5.48372746e-01 4.98546213e-01
1.77302808e-01 1.86024800e-01 -2.20104843e-01 -1.17340505e-01
1.68807283e-01 -5.23258671e-02 -4.44695726e-02 1.05435586e+00
5.97750256e-03 -1.94486737e-01 -9.57073450e-01 3.31079572e-01
-1.61891818e+00 -8.94171894e-01 -1.66406274e-01 1.62920356e+00
1.05311537e+00 2.66090184e-01 -3.27506423e-01 4.87139761e-01
4.89250422e-01 -3.71010453e-01 -2.86104847e-02 -8.12633991e-01
3.35720628e-02 5.19370914e-01 5.96783876e-01 8.83591026e-02
-1.10833263e+00 5.23746550e-01 5.86099863e+00 8.44254553e-01
-9.23866093e-01 6.70295283e-02 2.72151411e-01 2.25059167e-01
8.72810185e-02 -3.88444752e-01 -9.45078909e-01 3.92725140e-01
1.31140041e+00 -2.00253531e-01 -5.36508083e-01 8.12057853e-01
4.13967043e-01 -3.80479574e-01 -8.53736401e-01 7.80811965e-01
1.90783396e-01 -1.22981834e+00 1.86786950e-01 6.04087748e-02
3.33982795e-01 -3.18762928e-01 -1.84426844e-01 2.00589314e-01
-2.69831091e-01 -9.51207221e-01 -9.72849876e-03 9.44831848e-01
4.96912718e-01 -6.75282657e-01 1.48517597e+00 3.77900958e-01
-9.41992462e-01 -1.11522958e-01 -1.72662467e-01 3.35022539e-01
-2.64074326e-01 4.72001344e-01 -1.18175673e+00 1.14283490e+00
4.98847246e-01 5.79293668e-01 -7.01825440e-01 1.24080396e+00
1.15920462e-01 5.80564380e-01 -1.31760046e-01 -4.08308744e-01
5.58369040e-01 2.87250191e-01 2.84662634e-01 1.84224105e+00
2.93600649e-01 3.94864351e-01 3.54185998e-01 -1.95730060e-01
1.77793756e-01 1.20377648e+00 -7.46470153e-01 -4.68519926e-01
-3.99767384e-02 1.30277169e+00 -1.15496838e+00 -4.82104242e-01
-3.72592866e-01 9.13084149e-01 -2.17808649e-01 5.13354354e-02
-2.15907186e-01 -1.08174741e+00 -2.44492307e-01 7.95985907e-02
2.52318650e-01 1.07192755e-01 -3.23902518e-01 -1.02648163e+00
-2.84982979e-01 -6.25118732e-01 9.34926689e-01 -2.53796130e-01
-1.35104489e+00 1.09620428e+00 -2.16701120e-01 -1.28988898e+00
-2.39457697e-01 -8.05512965e-01 -3.86099517e-01 6.45034611e-01
-1.49438322e+00 -1.02142179e+00 -9.45653990e-02 8.44985366e-01
6.48550689e-01 -3.49247694e-01 1.35783541e+00 3.49523038e-01
-4.56292868e-01 5.82368493e-01 3.89489353e-01 4.76363897e-01
1.02646649e+00 -1.45424199e+00 -9.08156991e-01 2.09159851e-01
-3.81766520e-02 9.38511491e-01 5.09005427e-01 -7.78930306e-01
-9.46784735e-01 -5.84750652e-01 1.76472986e+00 -3.46130222e-01
6.23506546e-01 3.11987489e-01 -6.11621380e-01 1.54067516e-01
5.32199740e-01 -4.68112290e-01 1.40232468e+00 3.71644050e-02
-5.81364445e-02 8.07038695e-02 -1.23273420e+00 1.63840637e-01
3.31104785e-01 -4.82074022e-01 -1.16387546e+00 4.54783052e-01
2.53837138e-01 -5.98572083e-02 -1.11104798e+00 2.42591515e-01
3.62551630e-01 -2.30370402e-01 6.10054731e-01 -1.13187480e+00
4.06119078e-01 -3.08517013e-02 7.46457651e-02 -9.29460168e-01
-1.61484718e-01 -8.76352787e-02 6.21508695e-02 1.11131215e+00
7.60633826e-01 -5.41522801e-01 5.37956119e-01 2.96829611e-01
-1.29449576e-01 -9.41525638e-01 -7.18598604e-01 -4.05442506e-01
-7.97504932e-02 -5.96315181e-03 -1.61238071e-02 1.22553933e+00
7.52721250e-01 4.28994298e-01 2.87738889e-02 -1.56529710e-01
3.14550579e-01 7.69278333e-02 -2.62845486e-01 -1.33820450e+00
-4.88445675e-03 -3.71062607e-01 -6.95896327e-01 -9.24734492e-03
-1.04697168e-01 -1.01219082e+00 -1.10517427e-01 -1.84829354e+00
4.60073292e-01 -1.29410177e-01 -7.21242011e-01 5.16657770e-01
-1.75218865e-01 -1.11366272e-01 -2.01534972e-01 3.28868419e-01
-5.46848178e-01 3.59564647e-02 1.05808711e+00 -1.15669608e-01
-1.08405352e-01 8.87256116e-02 -6.14670873e-01 7.56788135e-01
8.54365945e-01 -9.82557833e-01 -3.18353444e-01 2.38136575e-01
1.12870932e-01 1.98418379e-01 -1.44905344e-01 -7.99478173e-01
6.28760576e-01 -1.75471172e-01 6.79801583e-01 -7.09781170e-01
-2.38413811e-01 -1.07442081e+00 -2.32848197e-01 6.53214395e-01
-6.26949787e-01 1.10323697e-01 1.04966901e-01 6.89079642e-01
-5.24979651e-01 -8.49054635e-01 4.52773988e-01 -4.17475134e-01
-5.61165273e-01 -8.93496200e-02 -7.62541592e-01 -2.09906891e-01
1.12056065e+00 -3.41920674e-01 -4.43312898e-02 1.11870974e-01
-1.17872119e+00 -1.29753068e-01 -5.87498173e-02 1.15096599e-01
5.69000602e-01 -8.21807206e-01 -5.14916003e-01 -3.46733272e-01
3.25616747e-01 -7.72776783e-01 4.02705848e-01 9.70348120e-01
-5.33617496e-01 1.08807635e+00 -6.14531040e-02 -7.63991401e-02
-1.58532465e+00 8.09621394e-01 1.71800002e-01 -7.52993166e-01
-7.93793976e-01 4.56327409e-01 -2.72408217e-01 -1.28634021e-01
2.63402700e-01 -6.27476647e-02 -1.18581867e+00 5.79364955e-01
5.76871097e-01 9.64015797e-02 4.67502862e-01 -3.50527823e-01
-5.52290440e-01 5.58528543e-01 -3.39791566e-01 3.98739800e-02
1.26962137e+00 -8.14357921e-02 -8.89117718e-02 4.19594526e-01
1.25806332e+00 1.47101343e-01 1.12451613e-01 -2.35871151e-01
7.74831533e-01 1.87944114e-01 4.56111193e-01 -1.33945239e+00
-5.77759266e-01 8.50889444e-01 8.42980444e-01 1.81886807e-01
1.24347806e+00 -2.73842812e-01 5.62091172e-01 7.33210444e-01
1.08949997e-01 -1.34638393e+00 -3.60790491e-01 2.78412044e-01
3.80372763e-01 -1.20338416e+00 1.19373992e-01 -3.47201616e-01
-6.76419735e-01 1.83090234e+00 2.04793736e-01 2.30213448e-01
1.12788057e+00 4.28177506e-01 1.59990147e-01 -4.99662846e-01
-5.98804355e-01 -1.57881543e-01 5.14769077e-01 3.07868838e-01
8.33573282e-01 -1.52923375e-01 -1.11072028e+00 9.00053918e-01
1.17588341e-01 3.73072296e-01 4.80663121e-01 1.30918097e+00
-5.45265198e-01 -1.34803140e+00 -9.82699245e-02 8.23866189e-01
-1.20125234e+00 -4.13771451e-01 -3.77176732e-01 7.36380160e-01
1.51080579e-01 1.02144921e+00 -5.35206616e-01 -2.15421945e-01
3.72001976e-01 5.28138340e-01 2.06809402e-01 -8.84220302e-01
-1.07200491e+00 1.53001294e-01 3.18531364e-01 -1.65510789e-01
-6.34597540e-01 -5.06584764e-01 -1.45372140e+00 7.26130679e-02
-5.35119176e-01 8.68678689e-01 7.97765076e-01 1.19848454e+00
3.83237213e-01 7.75484443e-01 5.73578417e-01 1.54644713e-01
-5.01466930e-01 -1.14513373e+00 -4.32442605e-01 8.54432732e-02
1.58261165e-01 -4.06961203e-01 -2.07939416e-01 1.45164341e-01]
|
[8.462166786193848, 8.709704399108887]
|
0b91cea9-35c5-4053-b180-a5b47eda4624
|
hierarchical-cross-modal-transformer-for-rgb
|
2302.08052
| null |
https://arxiv.org/abs/2302.08052v1
|
https://arxiv.org/pdf/2302.08052v1.pdf
|
Hierarchical Cross-modal Transformer for RGB-D Salient Object Detection
|
Most of existing RGB-D salient object detection (SOD) methods follow the CNN-based paradigm, which is unable to model long-range dependencies across space and modalities due to the natural locality of CNNs. Here we propose the Hierarchical Cross-modal Transformer (HCT), a new multi-modal transformer, to tackle this problem. Unlike previous multi-modal transformers that directly connecting all patches from two modalities, we explore the cross-modal complementarity hierarchically to respect the modality gap and spatial discrepancy in unaligned regions. Specifically, we propose to use intra-modal self-attention to explore complementary global contexts, and measure spatial-aligned inter-modal attention locally to capture cross-modal correlations. In addition, we present a Feature Pyramid module for Transformer (FPT) to boost informative cross-scale integration as well as a consistency-complementarity module to disentangle the multi-modal integration path and improve the fusion adaptivity. Comprehensive experiments on a large variety of public datasets verify the efficacy of our designs and the consistent improvement over state-of-the-art models.
|
['Feihong Shen', 'Hao Chen']
|
2023-02-16
| null | null | null | null |
['rgb-d-salient-object-detection']
|
['computer-vision']
|
[ 5.81636466e-02 -1.37564585e-01 -2.24181861e-01 -2.91660964e-01
-9.99949396e-01 -5.50356150e-01 7.33560681e-01 -1.53190196e-01
-1.24762706e-01 2.96191931e-01 6.75582409e-01 1.02715842e-01
-1.34162232e-01 -6.16219938e-01 -8.05260897e-01 -6.80225611e-01
2.07902133e-01 -2.09284306e-01 8.47971797e-01 -3.26788127e-01
3.12427077e-02 3.54615986e-01 -1.67248547e+00 7.81832516e-01
8.53475392e-01 1.32039523e+00 3.00117224e-01 8.14405754e-02
-2.85183848e-03 7.62380183e-01 1.70820341e-01 -2.84983814e-01
2.18954176e-01 -3.01654279e-01 -8.23681533e-01 4.71822843e-02
7.58135855e-01 -2.24787071e-01 -4.37970340e-01 1.18277955e+00
5.47054648e-01 -1.44571126e-01 3.07858884e-01 -1.32645357e+00
-1.01847780e+00 3.66253465e-01 -9.51638758e-01 3.02567512e-01
3.15923929e-01 3.37679386e-01 1.27073860e+00 -1.13266611e+00
5.03617644e-01 1.42453218e+00 7.92132080e-01 2.48688161e-01
-1.19847047e+00 -5.22219479e-01 5.21438837e-01 3.30884039e-01
-1.25484157e+00 -2.01983511e-01 1.13426435e+00 -2.91270316e-01
9.09011960e-01 1.52787447e-01 8.05687904e-01 1.03137469e+00
6.41870201e-02 1.14299822e+00 1.32825041e+00 -3.55424136e-01
-1.47402525e-01 -7.53604174e-02 -2.21310958e-01 7.18578041e-01
-1.71317622e-01 1.48613796e-01 -9.70469773e-01 2.00548507e-02
8.13504100e-01 2.37879127e-01 -2.68672019e-01 -8.42028201e-01
-1.58763516e+00 6.48742557e-01 1.09349525e+00 4.14212435e-01
-3.22196484e-01 -8.32400024e-02 3.55363131e-01 -2.09887072e-01
4.04421419e-01 1.69818744e-01 -5.01190901e-01 3.33295554e-01
-7.43825734e-01 9.51385405e-03 2.15595439e-02 9.52705443e-01
9.66996193e-01 -2.37022430e-01 -5.28643310e-01 7.46777833e-01
4.67206627e-01 3.79451603e-01 3.97372574e-01 -6.46014273e-01
4.69255865e-01 1.10023189e+00 -3.15203339e-01 -9.52171564e-01
-5.86839616e-01 -4.83606815e-01 -9.63689804e-01 8.97645131e-02
2.00520545e-01 4.08189207e-01 -9.31611955e-01 1.99838078e+00
6.21983945e-01 1.83783382e-01 -2.05822080e-01 1.19843960e+00
1.03016794e+00 2.35071823e-01 3.72774959e-01 1.45353436e-01
1.55503762e+00 -1.21387780e+00 -4.26705539e-01 -4.05895203e-01
3.06389987e-01 -7.17190683e-01 1.34091532e+00 -2.61588186e-01
-1.09437907e+00 -6.96871579e-01 -8.77787471e-01 -5.65149724e-01
-6.24498010e-01 5.38832974e-03 7.42463827e-01 1.66820034e-01
-1.00368917e+00 8.50791931e-02 -7.81970084e-01 -3.73008698e-01
5.12115240e-01 6.91300780e-02 -5.66117883e-01 -1.76327735e-01
-1.39498115e+00 8.70818794e-01 4.36977923e-01 2.62926519e-01
-7.52081633e-01 -9.87409115e-01 -1.04909873e+00 4.36945669e-02
1.99994743e-01 -8.28792334e-01 8.40366900e-01 -1.02880728e+00
-1.06434000e+00 1.02016306e+00 -2.30280742e-01 7.43356943e-02
3.06386650e-01 -1.05782464e-01 -3.49201798e-01 2.10249484e-01
3.12372178e-01 1.01995683e+00 5.92101753e-01 -1.35415149e+00
-7.84059167e-01 -5.62430918e-01 2.67427266e-01 4.70006883e-01
-5.29991984e-01 -2.06897110e-01 -6.95562541e-01 -5.98059833e-01
4.53827560e-01 -6.12279117e-01 9.26410109e-02 2.24682391e-01
-6.51181400e-01 -1.23325907e-01 9.40706670e-01 -4.14788663e-01
9.96084630e-01 -2.28702927e+00 3.95381451e-01 -9.84921306e-02
3.16581696e-01 -5.76237589e-02 -3.80063474e-01 1.98968053e-01
-5.61908148e-02 -1.89995304e-01 -2.24674925e-01 -4.89774138e-01
1.33088395e-01 9.94106680e-02 -2.20531434e-01 5.20835340e-01
6.34913087e-01 1.39519286e+00 -8.80964935e-01 -7.09565461e-01
4.36777145e-01 7.41810977e-01 -4.58566308e-01 1.15238942e-01
-4.73306291e-02 4.47112054e-01 -5.24970055e-01 1.26798964e+00
8.45014513e-01 -5.16349435e-01 -1.29608199e-01 -8.72941911e-01
-1.38799995e-01 2.89428174e-01 -8.96207631e-01 2.17415643e+00
-3.21533442e-01 4.05842423e-01 -6.12346418e-02 -8.52724135e-01
5.81924319e-01 -1.12639084e-01 4.81372148e-01 -1.23484421e+00
2.10602377e-02 1.28673211e-01 -3.87382030e-01 -3.86836469e-01
5.72882116e-01 -1.51630923e-01 -4.39498127e-02 1.26581982e-01
2.78405935e-01 3.33801150e-01 -1.26429781e-01 2.23104373e-01
7.53775239e-01 2.77567208e-01 1.84946671e-01 -2.95990437e-01
5.60230494e-01 -1.53063357e-01 5.22027433e-01 5.80421984e-01
-4.44964290e-01 1.04078650e+00 3.79388452e-01 -3.72056931e-01
-6.92350209e-01 -1.28668594e+00 -2.05305517e-01 1.32715988e+00
7.98117340e-01 -1.20463409e-01 -3.29096407e-01 -8.00047100e-01
4.95289974e-02 2.02873141e-01 -1.00487137e+00 -3.09608817e-01
-2.63381124e-01 -7.25988805e-01 3.50464672e-01 7.46120095e-01
9.04691994e-01 -9.38688219e-01 -6.80000186e-01 -1.08596995e-01
-5.87022603e-01 -1.20119429e+00 -5.69768727e-01 2.07051903e-01
-6.49374962e-01 -1.05393040e+00 -7.25692332e-01 -8.67537618e-01
3.43666017e-01 5.15669048e-01 1.25455260e+00 -1.46893770e-01
-1.43500552e-01 3.86065632e-01 -3.81111920e-01 -2.21494902e-02
3.58697712e-01 3.02550703e-01 -3.30248237e-01 1.08435661e-01
4.01536405e-01 -6.46494448e-01 -8.60162139e-01 3.30546647e-01
-1.05998814e+00 4.09984559e-01 8.07258844e-01 1.02126849e+00
8.50889981e-01 -2.88949281e-01 3.59309316e-01 -3.62816125e-01
7.50465840e-02 -4.47613150e-01 -2.49347851e-01 6.59413338e-01
-2.78784305e-01 1.35681964e-02 1.08791649e-01 -5.09218395e-01
-1.13587606e+00 1.13866687e-01 -1.52685726e-02 -6.21315897e-01
-2.01693043e-01 3.65065902e-01 -5.33957064e-01 -2.35974148e-01
2.60238498e-01 4.54010367e-01 -2.74346858e-01 -3.92146379e-01
5.98854840e-01 1.22659288e-01 5.88211596e-01 -4.62701678e-01
6.74691319e-01 8.62118661e-01 -1.51266530e-01 -3.47563505e-01
-1.17605162e+00 -4.47905868e-01 -8.57828915e-01 -1.90814808e-01
1.02411604e+00 -1.25874674e+00 -5.07775545e-01 5.52928567e-01
-1.07362866e+00 -1.89432338e-01 -3.65277261e-01 2.49029845e-01
-3.85983169e-01 1.77117407e-01 -5.11656284e-01 -4.57230657e-01
-1.81390747e-01 -1.18951046e+00 1.83320355e+00 3.71332943e-01
3.64859134e-01 -1.04867744e+00 8.00687522e-02 3.17801297e-01
4.98632967e-01 2.81033307e-01 7.79774547e-01 -2.65586972e-01
-8.26620340e-01 1.63403690e-01 -8.35928321e-01 1.22094508e-02
9.68827084e-02 -3.58717263e-01 -1.17179275e+00 -2.21984625e-01
-4.23941165e-01 -4.14838016e-01 1.12447369e+00 4.03151512e-01
1.00572658e+00 5.32287881e-02 -3.52451742e-01 7.41093218e-01
1.43233466e+00 -4.84101802e-01 7.19589651e-01 5.79657555e-01
1.07093751e+00 5.74119747e-01 5.33326447e-01 1.92848414e-01
8.65854919e-01 7.13713288e-01 7.28798032e-01 -7.13616669e-01
-2.96402901e-01 -5.29868603e-01 2.97029346e-01 5.29507339e-01
1.35352448e-01 4.82532866e-02 -6.84971392e-01 7.17533767e-01
-1.88448322e+00 -9.10439849e-01 -2.06037927e-02 1.72792721e+00
7.54331827e-01 -6.86176866e-02 2.34806567e-01 -2.21325412e-01
5.35979509e-01 3.55697066e-01 -5.79044700e-01 3.01594257e-01
-6.51119769e-01 -2.04620153e-01 4.12030220e-01 3.11895728e-01
-1.30819333e+00 8.44513416e-01 5.95738888e+00 1.04050803e+00
-1.22229588e+00 3.83195400e-01 5.93584120e-01 -1.12788334e-01
-6.43880606e-01 -2.93980772e-03 -6.88277245e-01 2.98974514e-01
4.03535776e-02 4.23901349e-01 6.51833788e-02 6.92418933e-01
-2.03061268e-01 -1.58190399e-01 -8.61284852e-01 8.76276851e-01
1.87675402e-01 -1.25460422e+00 4.31812182e-02 -3.09652630e-02
9.48567867e-01 2.50448704e-01 2.45348901e-01 2.59620130e-01
6.26100749e-02 -6.60208762e-01 1.12916660e+00 5.59894323e-01
6.63601816e-01 -5.58972418e-01 5.71129262e-01 3.60441916e-02
-1.78030801e+00 -1.73593253e-01 -1.96031854e-01 2.28322685e-01
2.55617023e-01 5.75037062e-01 -7.73361698e-03 8.55799973e-01
1.17460930e+00 1.06199563e+00 -1.02002382e+00 8.13067913e-01
-7.26263747e-02 -3.60513963e-02 -3.20596606e-01 3.58689964e-01
3.47649068e-01 1.95767939e-01 4.26218480e-01 1.18779206e+00
1.64281577e-01 -1.74719676e-01 7.72101581e-02 1.08367550e+00
1.88294783e-01 -1.54199615e-01 -3.65836173e-01 3.02937210e-01
4.03545141e-01 1.25797153e+00 -5.81407249e-01 -2.20603451e-01
-8.47589731e-01 1.08989072e+00 4.91598874e-01 5.19053459e-01
-1.01324975e+00 -2.03443971e-02 6.70585573e-01 1.05304435e-01
5.60930610e-01 -4.71716188e-02 -4.75415021e-01 -1.34713817e+00
2.13527843e-01 -6.03664577e-01 6.42589390e-01 -1.02268434e+00
-1.69163740e+00 6.05630457e-01 -2.51290128e-02 -1.52133632e+00
3.99825931e-01 -4.55826014e-01 -3.15797657e-01 8.64688158e-01
-2.14771819e+00 -2.16004705e+00 -4.26904500e-01 1.12968314e+00
2.99877495e-01 2.19014600e-01 4.36113775e-01 4.41435009e-01
-3.76941681e-01 7.26405978e-01 -2.06716865e-01 -6.32424504e-02
6.39936268e-01 -1.01855886e+00 -1.27165049e-01 1.03898764e+00
3.72931175e-02 6.01751924e-01 2.25510836e-01 -3.70762467e-01
-1.43736255e+00 -1.19972003e+00 7.07782209e-01 -5.00621855e-01
7.23737895e-01 -2.77102560e-01 -1.05190623e+00 5.96266866e-01
2.90337592e-01 5.69055915e-01 4.51381564e-01 2.29501992e-01
-9.31262791e-01 -2.05822304e-01 -9.23111856e-01 4.13461417e-01
1.27842665e+00 -1.00493300e+00 -5.54862916e-01 4.70423279e-03
8.78642082e-01 -5.01026571e-01 -7.89010167e-01 7.66231894e-01
6.51362777e-01 -1.45894551e+00 1.25396001e+00 -3.36972296e-01
6.14866614e-01 -5.31765461e-01 -4.70693827e-01 -8.53110135e-01
-5.81924736e-01 1.80408277e-03 -8.30473453e-02 1.50752664e+00
2.27377206e-01 -4.98237729e-01 4.04224426e-01 2.92767644e-01
-3.26108664e-01 -8.59673023e-01 -1.17561746e+00 -5.06704271e-01
-3.74578275e-02 -4.08955634e-01 6.62910700e-01 9.97995019e-01
1.21050932e-01 2.42079213e-01 -3.56184095e-01 4.63038325e-01
6.02015138e-01 5.74464023e-01 4.11117971e-01 -8.68411779e-01
-2.12976575e-01 -6.89607024e-01 -4.57253367e-01 -1.19095421e+00
-4.25767474e-04 -8.06470513e-01 1.13734111e-01 -1.50049138e+00
7.04920650e-01 -3.20092797e-01 -7.71145046e-01 7.36726701e-01
-4.11402911e-01 6.85492575e-01 1.39816403e-01 2.69927174e-01
-1.11154842e+00 1.02461898e+00 1.49095643e+00 -1.82755798e-01
-1.18775122e-01 -6.31485105e-01 -8.76487434e-01 6.19206369e-01
3.10905188e-01 -1.94379821e-01 -4.17435497e-01 -6.04864478e-01
8.15795958e-02 -2.30782524e-01 9.34701502e-01 -9.13882315e-01
3.06343675e-01 -2.34100029e-01 5.85052013e-01 -8.54426503e-01
3.27830136e-01 -9.15976405e-01 -1.84878334e-02 -4.95059378e-02
-3.94741058e-01 -1.36225363e-02 3.03393155e-01 6.95399582e-01
-5.12523770e-01 7.02906489e-01 9.47265923e-01 4.08876836e-02
-1.07713270e+00 5.08635402e-01 1.95619449e-01 1.09971985e-01
9.29421365e-01 -2.21850544e-01 -6.18866384e-01 3.66878994e-02
-4.50688541e-01 2.72988617e-01 6.50606096e-01 6.79607272e-01
6.43653214e-01 -1.74072540e+00 -3.25257599e-01 3.96143913e-01
6.54391527e-01 2.29313355e-02 8.24079394e-01 1.38493085e+00
1.01058364e-01 2.70688832e-01 -4.69098717e-01 -9.55353558e-01
-9.57615018e-01 5.99911630e-01 5.54270446e-01 -3.31167102e-01
-5.10232151e-01 1.00980330e+00 7.88405240e-01 -4.55419779e-01
8.88322517e-02 -2.88611561e-01 -6.87256902e-02 4.29751426e-02
4.78664428e-01 -1.74333066e-01 8.30370188e-02 -1.02689052e+00
-7.06112504e-01 8.55050445e-01 1.44339487e-01 1.91066116e-02
1.19327426e+00 -5.79940677e-01 -2.37244636e-01 4.26835358e-01
1.28279328e+00 -3.04278880e-01 -1.59411252e+00 -7.16907978e-01
-3.56708288e-01 -6.38034225e-01 1.54617026e-01 -7.62220323e-01
-1.35268331e+00 9.66423750e-01 8.21421564e-01 2.11059079e-01
1.61318350e+00 4.78150249e-01 6.48858249e-01 -2.35232264e-01
1.73903778e-01 -7.66869903e-01 3.54153484e-01 4.36394870e-01
9.24579859e-01 -1.49968171e+00 -4.83245850e-02 -3.79690737e-01
-7.80864775e-01 7.89855838e-01 8.86779130e-01 1.94479600e-01
7.53539622e-01 3.38737331e-02 -1.22190177e-01 -4.07954752e-01
-5.50546408e-01 -6.86557531e-01 7.21721172e-01 7.29479194e-01
3.25712651e-01 -1.48829311e-01 1.56982183e-01 5.44917583e-01
3.53351533e-01 -2.05840170e-01 -2.64135629e-01 9.66933668e-01
-2.33770698e-01 -7.20618725e-01 -1.89176932e-01 -3.00914864e-04
-1.26385639e-04 -2.29927570e-01 -3.14120382e-01 8.59505177e-01
5.16839981e-01 7.11581171e-01 4.34475467e-02 -5.43300152e-01
4.46218073e-01 -1.61102504e-01 4.17236984e-01 -1.29232496e-01
-4.91583377e-01 2.67619699e-01 -3.73771667e-01 -7.97247112e-01
-1.01877773e+00 -6.73117042e-01 -9.59828496e-01 -1.26041338e-01
-2.92325675e-01 -5.46924889e-01 2.54973263e-01 9.78218138e-01
6.55294001e-01 6.04531527e-01 4.95931864e-01 -1.19036222e+00
-7.04533905e-02 -8.50578368e-01 -5.95116973e-01 5.42613447e-01
6.29004359e-01 -1.06989765e+00 -2.71931231e-01 -5.87129146e-02]
|
[9.743783950805664, -0.7476024627685547]
|
5ce2ebd1-2a10-483e-af11-144cb8a1ef52
|
coco-gan-conditional-coordinate-generative
| null | null |
https://openreview.net/forum?id=r14Aas09Y7
|
https://openreview.net/pdf?id=r14Aas09Y7
|
COCO-GAN: Conditional Coordinate Generative Adversarial Network
|
Recent advancements on Generative Adversarial Network (GAN) have inspired a wide range of works that generate synthetic images. However, the current processes have to generate an entire image at once, and therefore resolutions are limited by memory or computational constraints. In this work, we propose COnditional COordinate GAN (COCO-GAN), which generates a specific patch of an image conditioned on a spatial position rather than the entire image at a time. The generated patches are later combined together to form a globally coherent full-image. With this process, we show that the generated image can achieve competitive quality to state-of-the-arts and the generated patches are locally smooth between consecutive neighbors. One direct implication of the COCO-GAN is that it can be applied onto any coordinate systems including the cylindrical systems which makes it feasible for generating panorama images. The fact that the patch generation process is independent to each other inspires a wide range of new applications: firstly, "Patch-Inspired Image Generation" enables us to generate the entire image based on a single patch. Secondly, "Partial-Scene Generation" allows us to generate images within a customized target region. Finally, thanks to COCO-GAN's patch generation and massive parallelism, which enables combining patches for generating a full-image with higher resolution than state-of-the-arts.
|
['Hwann-Tzong Chen', 'Da-Cheng Juan', 'Yu-Sheng Chen', 'Chia-Che Chang', 'Wei Wei', 'Chieh Hubert Lin']
|
2019-05-01
| null | null | null |
iclr-2019-5
|
['scene-generation']
|
['computer-vision']
|
[ 5.37289917e-01 3.04904044e-01 2.54317164e-01 2.74042934e-02
-7.60771930e-01 -6.08668864e-01 6.94712043e-01 -5.32569587e-01
1.92555636e-01 9.32544291e-01 -8.39654654e-02 1.25111071e-02
1.85430825e-01 -1.40389037e+00 -8.92888963e-01 -1.01930666e+00
4.58393127e-01 7.20629320e-02 3.30904365e-01 -3.96711558e-01
-1.78192146e-02 6.40452921e-01 -1.51755714e+00 2.87104249e-01
9.10084963e-01 7.57598639e-01 4.39383686e-01 6.79380476e-01
-1.41034961e-01 5.13094842e-01 -6.55927241e-01 -4.31248099e-01
4.90188420e-01 -1.04373443e+00 -3.59249979e-01 2.20993862e-01
5.36258996e-01 -2.54567623e-01 -3.22184235e-01 9.30296540e-01
3.94940585e-01 -1.37257144e-01 6.36100590e-01 -1.09947538e+00
-7.32767582e-01 1.83688998e-01 -6.74533367e-01 -5.35715163e-01
2.82500774e-01 1.65957242e-01 6.79831266e-01 -5.83903968e-01
7.89917469e-01 1.02106249e+00 5.72449505e-01 7.81396329e-01
-1.42628813e+00 -5.66482604e-01 -1.37070417e-02 -2.59511828e-01
-1.27581954e+00 -1.33232791e-02 1.01982021e+00 -2.62325585e-01
3.51584613e-01 5.19408464e-01 7.53933847e-01 1.01924253e+00
4.48856145e-01 4.31349188e-01 1.33755028e+00 -6.31976664e-01
2.44839177e-01 -5.08572049e-02 -7.46508300e-01 6.28457248e-01
1.69199239e-02 3.02211016e-01 -2.99129874e-01 1.22928604e-01
1.43229961e+00 1.04234494e-01 -4.67782289e-01 -3.35977763e-01
-1.33920252e+00 8.33343744e-01 6.17972195e-01 3.43134880e-01
-5.44644594e-01 2.63117254e-01 -1.55923799e-01 8.30409229e-02
2.37541541e-01 6.25224650e-01 1.17173016e-01 2.59187847e-01
-1.13769746e+00 4.48497534e-01 4.65137243e-01 8.85161698e-01
1.07613838e+00 3.16025913e-01 -2.42516488e-01 7.49304295e-01
-8.36048275e-02 6.72040284e-01 3.66950184e-01 -9.52177286e-01
2.26263821e-01 3.78731489e-01 -3.84224835e-03 -9.85006988e-01
3.02422047e-02 -4.14915115e-01 -1.39239192e+00 8.13734174e-01
2.95891076e-01 -2.87579596e-01 -1.14490461e+00 1.75243080e+00
4.22662169e-01 -2.89326478e-02 2.76172813e-02 7.08829224e-01
6.29193187e-01 1.05187142e+00 -3.66181761e-01 -1.74291562e-02
1.15404928e+00 -1.06871319e+00 -4.29027051e-01 -8.09407160e-02
-3.06710731e-02 -8.76180410e-01 9.30013001e-01 4.30174083e-01
-1.28441226e+00 -8.16202760e-01 -1.10807610e+00 6.80589229e-02
-2.46311054e-01 -1.81175575e-01 5.22598803e-01 4.71854031e-01
-1.22387993e+00 2.99614102e-01 -4.29149628e-01 4.55493666e-02
3.16991359e-01 7.59740397e-02 -3.72185439e-01 -7.57079497e-02
-9.76559937e-01 5.01973152e-01 2.11058587e-01 -1.94683671e-01
-9.28634346e-01 -7.11901248e-01 -7.71689534e-01 4.03601527e-02
8.11299384e-02 -1.10167980e+00 8.54921579e-01 -1.20149517e+00
-1.73265314e+00 6.34712696e-01 -1.97006688e-01 -3.43260705e-01
7.75456727e-01 2.59352446e-01 -2.66214103e-01 2.63581485e-01
2.05832794e-01 1.04062188e+00 1.12888205e+00 -1.55232811e+00
-5.74347615e-01 -3.62007581e-02 3.52057777e-02 1.98313519e-01
9.94940177e-02 -4.58929390e-01 -5.14415741e-01 -8.73360693e-01
1.21890865e-01 -9.94557977e-01 -5.00776470e-01 -8.05234909e-02
-5.74152589e-01 3.49990338e-01 9.03215230e-01 -4.37561482e-01
8.07542801e-01 -2.09864974e+00 3.09359491e-01 1.50126591e-01
1.19845994e-01 1.33421689e-01 -1.91878825e-01 5.56880176e-01
-1.25291362e-01 1.67335957e-01 -4.72148329e-01 -3.14185828e-01
-1.96858466e-01 -8.14824089e-05 -6.10267460e-01 1.29624441e-01
2.41515607e-01 1.16845298e+00 -6.92575336e-01 -4.34588939e-01
4.88273144e-01 8.63112867e-01 -5.97394049e-01 2.84877867e-01
-3.07287663e-01 9.35947895e-01 -3.97549212e-01 3.15431029e-01
8.62104237e-01 -1.21354587e-01 -4.55046035e-02 -1.18669711e-01
-1.46392047e-01 -3.52401704e-01 -1.02353883e+00 1.69468915e+00
-6.31484807e-01 4.79918033e-01 -3.29771116e-02 -7.05892444e-01
1.15224421e+00 3.14201295e-01 3.61863583e-01 -7.21839905e-01
-2.27814630e-01 1.41987994e-01 -2.39350498e-01 6.18430078e-02
4.21171248e-01 -3.62540543e-01 -1.32363722e-01 3.61885041e-01
-1.92737848e-01 -7.23637760e-01 1.86271444e-01 1.12872347e-01
8.28153610e-01 3.13724041e-01 1.05704181e-01 -2.82706600e-02
5.76709390e-01 -1.71820056e-02 4.62588251e-01 7.07071543e-01
4.73315656e-01 1.39826429e+00 4.54977840e-01 -5.00394464e-01
-1.46365881e+00 -1.34333229e+00 3.73881683e-03 3.76953900e-01
3.36413741e-01 -1.68847889e-01 -9.99514878e-01 -5.06940782e-01
-3.80405664e-01 4.31292474e-01 -8.16194594e-01 1.89407215e-01
-8.06745052e-01 -4.76439863e-01 3.07655036e-01 3.27651381e-01
8.63637447e-01 -1.30985367e+00 -6.44502521e-01 3.38342100e-01
-7.39740357e-02 -8.75406206e-01 -5.10012269e-01 -2.40168616e-01
-7.06089854e-01 -9.19584274e-01 -1.21834195e+00 -8.17550480e-01
9.07344699e-01 2.66519397e-01 1.14088106e+00 -1.10990435e-01
-2.59076595e-01 1.14790676e-02 -2.09062725e-01 -1.18121512e-01
-7.00318515e-01 6.29214151e-03 -3.52377415e-01 2.29643956e-01
-5.93424141e-01 -9.68258202e-01 -8.91115725e-01 3.24328244e-01
-1.23846591e+00 6.82658970e-01 8.25033605e-01 9.15399730e-01
1.06847417e+00 3.13689560e-01 7.02109098e-01 -1.01739061e+00
3.41959387e-01 -1.93321094e-01 -6.89643979e-01 3.00137073e-01
-1.96578220e-01 -4.34696972e-02 1.05796790e+00 -4.46177065e-01
-1.26210535e+00 2.33360514e-01 -1.59674749e-01 -4.00548279e-01
-2.06893086e-01 7.66512454e-02 -4.14029568e-01 -1.37127280e-01
5.88208735e-01 6.76951349e-01 8.06516260e-02 -1.59288272e-01
6.23072267e-01 2.13705227e-01 8.19878697e-01 -3.60943884e-01
1.07378411e+00 7.08925009e-01 3.58538508e-01 -8.41860056e-01
-4.45892066e-01 2.63989091e-01 -4.91504163e-01 -1.42234638e-01
1.10548770e+00 -8.10981452e-01 -3.84708405e-01 6.68139160e-01
-1.16360056e+00 -5.29423892e-01 -5.73423922e-01 7.83235431e-02
-7.18389094e-01 1.86996087e-01 -4.62083668e-01 -5.10569990e-01
-4.12914217e-01 -1.18105876e+00 1.11219907e+00 5.53723156e-01
7.04216212e-02 -8.86378229e-01 4.20847349e-02 1.82883292e-01
5.96768320e-01 7.86581933e-01 6.83454037e-01 3.86301368e-01
-1.03229952e+00 -2.32296899e-01 -1.15692064e-01 3.17795575e-01
3.37775916e-01 7.08396509e-02 -6.94090962e-01 -2.57438511e-01
5.21182120e-02 7.95745403e-02 5.65817893e-01 4.72045332e-01
1.12115240e+00 -3.23260248e-01 -3.16341996e-01 8.52503359e-01
1.66821396e+00 3.24641705e-01 1.16680884e+00 1.60135701e-01
8.68125379e-01 2.59312481e-01 3.09621602e-01 7.56122395e-02
1.16758727e-01 8.60193074e-01 4.01013523e-01 -5.51729918e-01
-5.31407714e-01 -5.94222486e-01 9.31062996e-02 4.65795904e-01
-1.44034952e-01 -3.05556387e-01 -4.68812466e-01 4.40142810e-01
-1.51576638e+00 -1.11448979e+00 -1.03601426e-01 2.22901630e+00
7.13832080e-01 -2.13316694e-01 -1.12751946e-01 4.68143001e-02
8.73844683e-01 4.09117758e-01 -4.43675935e-01 -3.26817244e-01
-3.27285379e-01 7.09665418e-01 4.16369796e-01 4.73991126e-01
-8.27188611e-01 8.95800114e-01 6.44556999e+00 1.11517155e+00
-1.58080804e+00 1.97021943e-02 9.62176561e-01 8.08594152e-02
-6.31948888e-01 -6.54963451e-03 -6.15231574e-01 5.25207520e-01
3.66836041e-01 -2.66950667e-01 3.65932226e-01 7.94682264e-01
-4.29228134e-02 -1.39486700e-01 -6.10305071e-01 8.12294841e-01
5.91753572e-02 -1.62429881e+00 4.77065921e-01 2.53677607e-01
1.46779549e+00 -4.10599023e-01 2.57336944e-01 -9.33068469e-02
2.74547994e-01 -1.23824859e+00 6.77453995e-01 5.61153114e-01
1.24314070e+00 -1.01459241e+00 3.15445215e-01 4.11675721e-01
-1.06651497e+00 3.14264119e-01 -4.71258521e-01 2.32070416e-01
5.28097510e-01 8.12156320e-01 -5.17794251e-01 8.63315582e-01
3.93205792e-01 2.78958172e-01 -3.41673404e-01 7.69353211e-01
-5.22376120e-01 2.80364662e-01 -2.06264675e-01 3.48934114e-01
2.14712337e-01 -5.14241576e-01 4.62497264e-01 8.41855228e-01
7.45846689e-01 -2.36259624e-02 -9.63879153e-02 1.20997465e+00
1.16477497e-02 -2.77013760e-02 -8.37707639e-01 3.81453127e-01
4.12525743e-01 1.27385950e+00 -6.87702715e-01 -4.21004623e-01
-2.60473758e-01 1.20643842e+00 6.18355125e-02 3.74729216e-01
-9.61846292e-01 -5.38263738e-01 3.31030577e-01 3.61198753e-01
6.56596541e-01 -2.08253503e-01 -4.48004246e-01 -1.07991600e+00
-8.89732130e-03 -9.78807271e-01 -1.28839925e-01 -8.66683781e-01
-1.00194049e+00 1.08324349e+00 -2.59712100e-01 -1.47249854e+00
-4.75116789e-01 -1.43683195e-01 -8.90287936e-01 1.15067708e+00
-1.34217608e+00 -1.42619848e+00 -6.16103351e-01 6.81049645e-01
3.83699715e-01 -7.52893239e-02 9.45172727e-01 6.25300314e-03
-7.86348432e-02 5.44014215e-01 8.39063078e-02 5.92864342e-02
5.37219763e-01 -1.06592655e+00 5.84302425e-01 1.04917765e+00
2.88002282e-01 3.20463121e-01 4.16623026e-01 -6.31321847e-01
-1.22202897e+00 -1.20627487e+00 4.57874060e-01 -2.53840834e-01
-2.86000911e-02 -1.80524573e-01 -7.10130632e-01 3.97064388e-01
4.66945142e-01 -8.87135696e-03 1.55909464e-01 -5.60115039e-01
-2.60014266e-01 -3.14168751e-01 -1.32576036e+00 8.01839054e-01
8.05699468e-01 -2.66396672e-01 -1.25221536e-01 6.85683787e-02
6.63949192e-01 -6.22737944e-01 -7.08697140e-01 4.73614752e-01
4.35410231e-01 -1.45456648e+00 1.00991702e+00 8.27814117e-02
9.26864803e-01 -7.40013123e-01 2.70328689e-02 -1.42002010e+00
-3.41660827e-01 -8.48346472e-01 2.01307401e-01 1.21987069e+00
3.58667135e-01 -7.79857218e-01 7.90829122e-01 1.97170272e-01
-1.53226539e-01 -7.52536774e-01 -8.95043850e-01 -6.33465290e-01
2.05040231e-01 -7.35919997e-02 1.05398417e+00 7.87410378e-01
-4.47653741e-01 1.92964301e-01 -5.95263958e-01 1.85811237e-01
6.55112028e-01 5.78267694e-01 1.15638196e+00 -7.29884207e-01
-6.82581127e-01 -3.55439931e-01 -4.36159700e-01 -1.10536861e+00
-3.30602884e-01 -5.38142562e-01 3.57392952e-02 -1.54918599e+00
-4.22629900e-02 -6.37819290e-01 1.84671968e-01 2.02185601e-01
-1.86737508e-01 9.35850143e-01 4.10619199e-01 2.35844657e-01
5.80796190e-02 5.52179575e-01 1.84799719e+00 -1.40772043e-02
-2.28206784e-01 -5.07719815e-02 -7.02776909e-01 3.67800295e-01
7.20248878e-01 -5.15251942e-02 -4.84013081e-01 -2.90883869e-01
1.08067289e-01 3.30368519e-01 3.97787035e-01 -1.27862501e+00
1.21377688e-02 -1.13219425e-01 6.01695955e-01 -5.20676494e-01
4.98484820e-01 -5.60357571e-01 9.56921995e-01 3.13355923e-01
9.58717167e-02 -1.15711600e-01 4.60513011e-02 3.39913458e-01
-5.01981437e-01 -6.37540817e-02 1.07326281e+00 -3.29076052e-01
-3.63924414e-01 3.96389157e-01 -8.76953676e-02 -1.68381885e-01
1.29089403e+00 -3.40768218e-01 -2.17126220e-01 -6.71285033e-01
-5.18544853e-01 -3.77620459e-01 9.74594533e-01 1.92091897e-01
4.55908537e-01 -1.59675193e+00 -8.10292482e-01 4.88249034e-01
-2.45083973e-01 6.24834538e-01 4.96358395e-01 3.68160009e-01
-8.79504919e-01 2.24689528e-01 -3.63593787e-01 -7.22244203e-01
-9.84156549e-01 6.69618249e-01 2.55181789e-01 -3.40070784e-01
-8.29015553e-01 6.25122488e-01 9.44019496e-01 -1.77072063e-01
-4.29106444e-01 5.22797965e-02 2.22520307e-01 -3.53820026e-01
5.33068836e-01 -8.96990672e-02 -2.55535603e-01 -5.90559661e-01
1.75267205e-01 9.89382148e-01 1.88538596e-01 -3.66438389e-01
1.29507864e+00 1.26091167e-01 -9.63418856e-02 -2.91402023e-02
9.49407279e-01 4.95265603e-01 -1.56900167e+00 6.54053390e-02
-8.78151059e-01 -6.81410432e-01 -2.27733478e-01 -6.35928869e-01
-1.54659379e+00 7.02338398e-01 2.62218982e-01 3.30239654e-01
1.43597949e+00 -1.89024419e-01 8.26062322e-01 -3.51499319e-01
5.50827742e-01 -6.10398531e-01 1.63064137e-01 1.75703317e-01
1.05799782e+00 -7.51491666e-01 -8.03995654e-02 -6.14050806e-01
-6.42615318e-01 1.04961371e+00 4.38655078e-01 -4.15249079e-01
2.68165261e-01 2.90966749e-01 1.34440094e-01 3.31935436e-02
-4.05756831e-01 6.52228445e-02 3.22864830e-01 8.01181614e-01
3.28190863e-01 1.85609192e-01 -2.94461966e-01 1.17662251e-01
-5.56784570e-01 -6.41654730e-02 5.55753291e-01 5.04977942e-01
-6.43587392e-03 -1.48716736e+00 -5.66017747e-01 1.01928659e-01
-2.86505967e-01 -9.45192799e-02 -1.81498498e-01 8.38469684e-01
3.42764527e-01 6.76988542e-01 3.63177210e-02 -1.64180994e-01
7.71808997e-02 -2.32507408e-01 6.06144071e-01 -4.51927364e-01
-2.70420581e-01 5.86398691e-02 -3.67863119e-01 -4.75930810e-01
-1.85666651e-01 -3.89659882e-01 -9.83628869e-01 -2.96322674e-01
-9.99204144e-02 1.70247883e-01 6.33793890e-01 4.59070593e-01
5.90648055e-01 6.18090332e-01 9.09049749e-01 -1.08868909e+00
-3.76752056e-02 -7.26430178e-01 -4.78932858e-01 4.27782953e-01
1.19163416e-01 -1.47825748e-01 -3.26022357e-01 1.81950048e-01]
|
[11.648832321166992, -0.6373625993728638]
|
a9a81707-f3b6-466f-92f2-ea26abb6e857
|
multi-agent-deep-reinforcement-learning-for-10
|
2209.02633
| null |
https://arxiv.org/abs/2209.02633v3
|
https://arxiv.org/pdf/2209.02633v3.pdf
|
Energy Management of Multi-mode Hybrid Electric Vehicles based on Hand-shaking Multi-agent Learning
|
The future transportation system will be a multi-agent network where connected AI agents can work together to address the grand challenges in our age, e.g., mitigation of real-world driving energy consumption. Distinguished from the existing research on vehicle energy management, which decoupled multiple inputs and multiple outputs (MIMO) control into single-output(MISO) control, this paper studied a multi-agent deep reinforcement learning (MADRL) framework to deal with multiple control outputs simultaneously. A new hand-shaking strategy is proposed for the DRL agents by introducing an independence ratio, and a parametric study is conducted to obtain the best setting for the MADRL framework. The study suggested that the MADRL with an independence ratio of 0.2 is the best, and more than 2.4% of energy can be saved over the conventional DRL framework.
|
['Quan Zhou', 'Zhi Li', 'Min Hua']
|
2022-09-06
| null | null | null | null |
['total-energy', 'energy-management']
|
['miscellaneous', 'time-series']
|
[-2.80750602e-01 1.76081672e-01 -5.65949321e-01 8.28672349e-02
-1.46766394e-01 -3.33849192e-01 4.68114406e-01 -1.72527030e-01
-4.78381515e-01 1.07141876e+00 -2.25509152e-01 -3.52187365e-01
-5.33562362e-01 -1.10349739e+00 -5.38318276e-01 -1.19243610e+00
-4.86279372e-03 1.85207009e-01 9.43303257e-02 -6.23226464e-01
7.07066953e-02 4.89322603e-01 -1.29495966e+00 -6.28068149e-01
1.09801984e+00 7.97677815e-01 3.27497900e-01 5.36104918e-01
2.69885540e-01 8.71853590e-01 -5.58146775e-01 -1.05649028e-02
2.06093296e-01 -3.43633443e-01 -3.55535537e-01 -1.81234583e-01
-4.82557148e-01 -5.52221835e-01 -4.58276004e-01 8.28726232e-01
5.90971947e-01 6.02026224e-01 5.41542828e-01 -2.35078502e+00
-5.79319179e-01 5.51500082e-01 -6.67191625e-01 -7.65394792e-02
-3.18003178e-01 4.28674877e-01 6.68561816e-01 -1.32735193e-01
2.50280648e-01 1.33266342e+00 4.49815020e-02 8.09412181e-01
-9.37779307e-01 -8.35994422e-01 3.42785031e-01 6.39871955e-01
-1.23064566e+00 -1.47964105e-01 8.58484089e-01 2.33318564e-02
1.23507154e+00 1.86285123e-01 7.94621944e-01 8.14707637e-01
5.97457647e-01 6.80274308e-01 8.34036648e-01 -4.12209183e-01
4.17072058e-01 -5.16908206e-02 -1.08365692e-01 4.06420290e-01
7.98972607e-01 2.32554123e-01 8.85247141e-02 3.18037212e-01
4.22222197e-01 -1.45191476e-01 5.28007746e-01 -2.55717933e-01
-9.97173309e-01 1.06252408e+00 5.00456870e-01 8.89199600e-02
-5.05386591e-01 6.88708425e-01 1.35880247e-01 1.43773526e-01
2.06083842e-02 6.89469099e-01 -2.54968345e-01 1.25818580e-01
-2.46728554e-01 2.71967202e-01 5.67817450e-01 8.22452486e-01
7.18688905e-01 7.56787062e-01 3.67814340e-02 5.41079819e-01
4.53203350e-01 7.08360970e-01 1.80774316e-01 -1.58066237e+00
3.83082479e-01 7.75310338e-01 3.04370105e-01 -7.74322927e-01
-7.62102425e-01 2.08329130e-02 -1.01209068e+00 6.39227331e-01
1.29520312e-01 -1.07542050e+00 -6.81901753e-01 1.88671112e+00
2.40970150e-01 6.34541512e-02 5.22609055e-01 8.21190894e-01
2.82174081e-01 1.09853792e+00 1.14206292e-01 -3.15911442e-01
1.12793744e+00 -1.26726139e+00 -1.06317115e+00 -3.90233189e-01
4.04792905e-01 -1.18597195e-01 3.74515027e-01 9.77781564e-02
-1.01875007e+00 -2.42994159e-01 -1.36912286e+00 2.22150236e-01
-7.65031755e-01 2.02627517e-02 5.39386749e-01 6.29407465e-01
-1.13151813e+00 -1.17889624e-02 -8.11341941e-01 -8.41951445e-02
-1.69114470e-01 7.27094114e-01 1.28348842e-01 7.90898874e-02
-1.62239444e+00 1.37509692e+00 5.86029828e-01 2.22456083e-01
-1.04700983e+00 -3.79300386e-01 -6.60712242e-01 -1.16379037e-02
9.50264633e-01 -5.55527687e-01 1.14472795e+00 -6.82468951e-01
-1.80561626e+00 -9.37945843e-02 1.64514601e-01 -4.31517184e-01
3.45987350e-01 1.21954329e-01 -4.57750827e-01 1.29065886e-01
-2.79962629e-01 8.73781502e-01 3.97933334e-01 -1.27872753e+00
-8.16085458e-01 8.35282207e-02 6.85466588e-01 4.07869250e-01
-2.86194503e-01 -3.37036669e-01 2.26585120e-02 -1.66693121e-01
-6.40090168e-01 -1.24161494e+00 -5.60870349e-01 -4.39524680e-01
-1.35136485e-01 -7.43481934e-01 9.78505015e-01 -2.80471891e-01
1.32507980e+00 -1.58596087e+00 5.68221867e-01 1.61371555e-03
1.94089428e-01 2.90095091e-01 -2.50591785e-01 6.54371023e-01
3.90823722e-01 2.20867485e-01 1.42736241e-01 3.76879051e-02
3.00811827e-01 4.51850504e-01 2.48120621e-01 1.48690224e-01
4.64160681e-01 8.12135756e-01 -9.64404345e-01 -2.76453018e-01
3.24031442e-01 2.00706214e-01 -2.89459914e-01 3.01959097e-01
-3.06588918e-01 7.86603987e-02 -7.76141226e-01 5.22976816e-01
6.36272967e-01 1.72021255e-01 2.81123638e-01 -3.24942693e-02
-3.90102535e-01 -5.81827641e-01 -1.30410433e+00 8.66101027e-01
-4.88504469e-01 6.54949605e-01 4.90879267e-01 -9.68087375e-01
8.95473897e-01 3.22373003e-01 9.62344885e-01 -1.20656168e+00
2.97988802e-01 1.66773349e-01 3.86322469e-01 -6.47473097e-01
5.54378211e-01 2.16279596e-01 -2.97652453e-01 3.54222655e-01
-2.59722233e-01 5.26891686e-02 4.53169346e-01 -4.65746075e-02
1.05944097e+00 -1.49824783e-01 8.38426501e-02 -2.81508297e-01
4.58644032e-01 1.27996281e-01 9.21926022e-01 4.27513599e-01
-5.86872339e-01 -4.77627933e-01 5.79116881e-01 -1.78414211e-01
-1.13414657e+00 -7.04176307e-01 4.99011934e-01 8.29740644e-01
8.30511868e-01 2.88523316e-01 -8.07602704e-01 -3.00811976e-01
7.50928298e-02 1.06293702e+00 -2.71785855e-01 -6.06977344e-01
-8.06885302e-01 -9.73309696e-01 5.58218241e-01 5.16601682e-01
9.32781756e-01 -9.23391163e-01 -9.79466856e-01 2.89087653e-01
-2.06485819e-02 -1.11096346e+00 -3.44782025e-01 2.93141156e-01
-2.01734513e-01 -1.05131221e+00 -3.84140134e-01 -6.64416552e-01
5.61032832e-01 3.54698837e-01 5.05516112e-01 3.68405282e-02
5.46536706e-02 1.83753043e-01 -1.40434369e-01 -5.60159683e-01
-3.95907134e-01 2.30870783e-01 3.00301611e-01 -4.10575187e-03
2.79990047e-01 -7.22763687e-02 -7.09321618e-01 4.60654289e-01
-8.10072184e-01 3.06397557e-01 6.80308819e-01 6.61938310e-01
2.47031033e-01 5.45563757e-01 1.36678112e+00 4.33146097e-02
7.36607492e-01 -8.03222299e-01 -8.11391354e-01 3.84528935e-01
-9.95035291e-01 3.11087251e-01 7.86096394e-01 -4.91475552e-01
-1.07708228e+00 8.12397059e-03 6.52636066e-02 -2.02972308e-01
5.40379770e-02 6.35499880e-02 -4.97293383e-01 -1.00173809e-01
-2.27574155e-01 3.58537547e-02 2.48251393e-01 2.24407747e-01
2.83934325e-01 6.05603755e-01 9.01653916e-02 -5.23142219e-01
5.68460464e-01 -1.42296016e-01 5.48774123e-01 -4.46792960e-01
1.09577060e-01 2.75429636e-01 -2.77378947e-01 -6.82783067e-01
1.11619902e+00 -7.93101013e-01 -1.38286734e+00 6.96981907e-01
-9.22129035e-01 -6.65009975e-01 1.03384830e-01 5.86877584e-01
-6.66082561e-01 -1.99188247e-01 -2.94911176e-01 -1.22262990e+00
-1.41259044e-01 -1.46215892e+00 3.60429734e-01 7.29711056e-01
3.34346563e-01 -9.26600397e-01 8.78634751e-02 2.55727053e-01
6.47943079e-01 5.02081811e-01 1.04288900e+00 -2.64533013e-01
-6.50317430e-01 3.81257050e-02 1.02436736e-01 2.40047082e-01
6.25647977e-02 1.59794360e-01 -4.47285175e-01 -5.42399406e-01
-3.20935071e-01 -2.84882396e-01 3.95590276e-01 3.91589433e-01
1.03802454e+00 -3.57932866e-01 -4.98986125e-01 1.19714811e-01
1.74511421e+00 1.16263568e+00 5.91227472e-01 6.14471674e-01
5.58618188e-01 3.89787376e-01 6.65436208e-01 2.05490753e-01
1.06714654e+00 6.37841940e-01 9.92591381e-01 -2.67616689e-01
2.38308348e-02 2.18326133e-02 5.03825128e-01 6.47081435e-01
-1.75050840e-01 -7.90810227e-01 -6.96795464e-01 4.82069671e-01
-2.18546867e+00 -9.00248587e-01 3.18011791e-02 1.77415526e+00
4.42656308e-01 1.62346438e-01 2.69024760e-01 1.12938374e-01
7.97726989e-01 1.65780429e-02 -1.04644585e+00 -8.59564722e-01
3.49894762e-02 -4.57649887e-01 1.01149881e+00 4.89596099e-01
-7.80162275e-01 4.73685652e-01 6.73245621e+00 8.24108481e-01
-1.01543796e+00 -8.04532096e-02 5.52104414e-01 -2.56240010e-01
-1.90271482e-01 -2.46289343e-01 -7.79444277e-01 7.04136372e-01
1.46763766e+00 -6.40492916e-01 7.52775311e-01 4.71473813e-01
8.22495937e-01 -4.12065685e-01 -7.38056004e-01 5.08294880e-01
-3.76497865e-01 -1.04524279e+00 -3.09300363e-01 3.39390039e-01
8.81720364e-01 -8.33976567e-02 -3.78678471e-01 5.03448546e-01
7.51575470e-01 -9.00281310e-01 4.55762088e-01 6.34527564e-01
2.77527422e-01 -1.25640929e+00 7.20440626e-01 5.72553039e-01
-1.38282788e+00 -7.32499063e-01 -4.66365144e-02 -2.44995266e-01
2.61474222e-01 -6.12389185e-02 -4.32559848e-01 6.29588723e-01
2.99263179e-01 4.14570957e-01 -2.79654354e-01 8.08877885e-01
6.53534234e-02 2.67198235e-01 -3.28870505e-01 -8.04284155e-01
4.85118806e-01 -3.87861639e-01 4.85619605e-01 8.88998628e-01
3.05444956e-01 1.28557146e-01 4.12562311e-01 5.54115832e-01
-8.95847604e-02 -5.08241713e-01 -6.40678525e-01 5.06538153e-02
7.89873242e-01 1.37062430e+00 -6.58403337e-01 -2.32830554e-01
-2.87245244e-01 3.68605763e-01 8.67252424e-02 4.64131147e-01
-1.25480747e+00 -8.41212690e-01 9.28007066e-01 -2.84164518e-01
1.22765332e-01 -5.00605643e-01 -1.89900115e-01 -6.65503502e-01
-2.41319507e-01 -3.69935483e-01 -2.22918969e-02 -7.67548561e-01
-9.43012953e-01 1.62600473e-01 3.14906567e-01 -1.01562917e+00
-3.90759289e-01 -6.01128936e-01 -7.96545208e-01 4.83066767e-01
-1.95027316e+00 -9.34952378e-01 -2.88132042e-01 3.21385533e-01
6.23300791e-01 -4.38384652e-01 6.02792680e-01 5.35031021e-01
-1.45205879e+00 5.33829093e-01 5.33674359e-01 -2.36059621e-01
1.75250620e-01 -1.28269172e+00 -1.80883482e-01 8.15590858e-01
-9.89377677e-01 -1.94031671e-01 6.16339445e-01 -3.87240678e-01
-2.29385567e+00 -1.17996264e+00 2.57003874e-01 1.41055077e-01
6.29466355e-01 -1.18154418e-02 -3.23555261e-01 3.30355585e-01
9.46352124e-01 -3.44107300e-01 1.16153561e-01 -7.82234073e-01
4.64320838e-01 -5.56506753e-01 -1.34936082e+00 8.15063238e-01
6.56251967e-01 2.21994683e-01 -7.45835230e-02 -7.60392621e-02
8.34065199e-01 -1.81576282e-01 -8.54069471e-01 3.52668971e-01
4.34744418e-01 -8.81502256e-02 7.60990262e-01 -5.45833945e-01
2.45440632e-01 -5.79322219e-01 -1.84937894e-01 -1.88838780e+00
-3.61176819e-01 -4.60269898e-01 -4.90495592e-01 1.27265131e+00
2.74364710e-01 -6.47914767e-01 2.80229807e-01 6.97135448e-01
-3.19143742e-01 -7.65379429e-01 -1.21707582e+00 -8.87431383e-01
2.76454628e-01 2.07243990e-02 6.56713009e-01 5.48420131e-01
-1.55819848e-01 3.03594977e-01 -5.12925327e-01 3.72857839e-01
5.72565377e-01 -2.91659474e-01 4.54098940e-01 -8.54550183e-01
1.88303947e-01 -6.91214800e-01 -1.02887318e-01 -4.33350950e-01
2.92950898e-01 -5.47891080e-01 1.83367655e-01 -2.03043389e+00
-1.38832942e-01 -2.80909210e-01 -5.75843513e-01 6.11842036e-01
3.73101048e-02 -3.98415215e-02 3.96189988e-01 -3.64721209e-01
-5.40632665e-01 7.83380151e-01 1.28835154e+00 -4.65502739e-01
-2.13935539e-01 -7.14096352e-02 -7.68023908e-01 2.61551350e-01
1.11427188e+00 -3.36154163e-01 -7.85854936e-01 -5.11907995e-01
1.24059588e-01 3.01926225e-01 2.20787898e-01 -9.40742195e-01
4.40357715e-01 -1.11782789e+00 9.32491124e-02 -4.20408726e-01
3.35218430e-01 -1.19854915e+00 3.64035875e-01 8.17688406e-01
-2.02087849e-01 3.38829070e-01 1.75558999e-01 4.95574504e-01
1.09643200e-02 -1.45885527e-01 7.60164320e-01 1.15362331e-01
-9.71476793e-01 -2.72182897e-02 -8.84887695e-01 -3.40713739e-01
1.90720820e+00 -6.81249984e-03 -8.86734784e-01 -1.65165111e-01
-3.69990379e-01 1.30970454e+00 -1.35824665e-01 6.50289953e-01
3.39579165e-01 -1.53119040e+00 -5.91894865e-01 -1.82472661e-01
-4.51322764e-01 -3.48727971e-01 1.97436079e-01 6.55498683e-01
-2.87044287e-01 4.65926290e-01 -6.16106510e-01 -7.90243447e-02
-8.23655963e-01 4.80298042e-01 8.53687942e-01 -2.40218177e-01
-1.88056096e-01 -6.00786470e-02 -5.15259981e-01 -1.93117887e-01
1.57769710e-01 -1.60070539e-01 -2.60855347e-01 3.07615757e-01
1.16854228e-01 1.15187061e+00 -1.75973505e-01 -2.95032382e-01
-4.68776554e-01 5.64605772e-01 4.38441306e-01 -8.02478194e-03
1.41990685e+00 -4.31638569e-01 1.61807746e-01 2.72409856e-01
8.61467481e-01 -5.70823967e-01 -1.43569863e+00 3.19587409e-01
-3.33690315e-01 1.79022804e-01 4.07996029e-01 -9.80483413e-01
-1.38364410e+00 4.87975180e-01 7.11935282e-01 7.12334692e-01
1.34731102e+00 -5.36076427e-01 7.62269855e-01 6.45437539e-01
3.15271288e-01 -1.64733565e+00 1.42315358e-01 4.54721063e-01
8.00885916e-01 -1.21836627e+00 -3.92096788e-02 2.27059796e-01
-7.79268563e-01 1.19479215e+00 1.33034790e+00 -2.21045360e-01
5.57453275e-01 4.39701825e-01 -4.25752610e-01 1.49194479e-01
-1.05710065e+00 -2.61699975e-01 -2.22724289e-01 3.89986813e-01
-2.69284248e-01 3.87538791e-01 -5.39262176e-01 4.45385367e-01
3.51837069e-01 9.16465074e-02 9.74856019e-01 1.00885403e+00
-5.35854161e-01 -9.53035235e-01 -2.40697488e-01 3.13838273e-01
1.98695622e-02 8.10415149e-01 -2.21277922e-02 9.79475081e-01
2.74780065e-01 1.45909560e+00 3.01449656e-01 -5.20719767e-01
5.35611987e-01 -1.29660368e-01 8.07079002e-02 6.21469729e-02
-3.19089651e-01 -1.04993790e-01 1.48718566e-01 -2.64146864e-01
-5.19326746e-01 -2.47504607e-01 -1.73191953e+00 -6.62363946e-01
-2.83773154e-01 3.98166738e-02 8.40930760e-01 1.01384532e+00
4.73240197e-01 1.10362577e+00 1.24792290e+00 -8.03659022e-01
-3.79770190e-01 -6.77890003e-01 -3.60247254e-01 -3.60243440e-01
6.16364658e-01 -9.30247962e-01 -3.22083622e-01 -4.75728542e-01]
|
[5.462203502655029, 2.180488109588623]
|
c70fc396-07cd-4249-b5dc-c1f9f536bd6c
|
distinguishing-cause-from-effect-on
|
2303.08572
| null |
https://arxiv.org/abs/2303.08572v1
|
https://arxiv.org/pdf/2303.08572v1.pdf
|
Distinguishing Cause from Effect on Categorical Data: The Uniform Channel Model
|
Distinguishing cause from effect using observations of a pair of random variables is a core problem in causal discovery. Most approaches proposed for this task, namely additive noise models (ANM), are only adequate for quantitative data. We propose a criterion to address the cause-effect problem with categorical variables (living in sets with no meaningful order), inspired by seeing a conditional probability mass function (pmf) as a discrete memoryless channel. We select as the most likely causal direction the one in which the conditional pmf is closer to a uniform channel (UC). The rationale is that, in a UC, as in an ANM, the conditional entropy (of the effect given the cause) is independent of the cause distribution, in agreement with the principle of independence of cause and mechanism. Our approach, which we call the uniform channel model (UCM), thus extends the ANM rationale to categorical variables. To assess how close a conditional pmf (estimated from data) is to a UC, we use statistical testing, supported by a closed-form estimate of a UC channel. On the theoretical front, we prove identifiability of the UCM and show its equivalence with a structural causal model with a low-cardinality exogenous variable. Finally, the proposed method compares favorably with recent state-of-the-art alternatives in experiments on synthetic, benchmark, and real data.
|
['Catarina A. Oliveira', 'Mário A. T. Figueiredo']
|
2023-03-14
| null | null | null | null |
['causal-discovery']
|
['knowledge-base']
|
[ 6.21137023e-01 2.04253614e-01 -2.07806900e-01 -5.33613935e-02
-6.61182582e-01 -5.69146931e-01 9.27809596e-01 4.90252495e-01
-2.71559417e-01 1.21989048e+00 3.46502513e-01 -6.15201771e-01
-7.37927735e-01 -1.12899649e+00 -1.22434044e+00 -8.72235417e-01
-4.06327993e-01 2.19290391e-01 1.72865577e-02 1.93133220e-01
2.06403777e-01 2.79538073e-02 -1.44634795e+00 1.54386712e-02
5.57726502e-01 8.02685440e-01 1.69532359e-01 5.38545370e-01
1.04192942e-01 7.54379749e-01 -5.90101957e-01 -1.97152331e-01
3.24485898e-02 -7.62082100e-01 -9.84810710e-01 -4.78894353e-01
-1.76015094e-01 1.00297816e-01 -1.69131994e-01 1.12229538e+00
2.20293939e-01 -2.93042123e-01 1.02501953e+00 -1.22235858e+00
-4.73681778e-01 1.06149399e+00 -4.10149872e-01 -1.56054702e-02
5.69227040e-01 -2.83261746e-01 1.20254254e+00 -3.81290913e-01
6.54045463e-01 1.43477631e+00 5.44023037e-01 2.03240871e-01
-1.83664215e+00 -6.49506032e-01 -1.06099388e-02 -3.98055539e-02
-1.65302455e+00 -9.68771726e-02 2.79450387e-01 -4.73598331e-01
1.68354213e-01 5.22358716e-01 3.44720095e-01 1.28482497e+00
4.81654227e-01 3.25876027e-01 1.33873522e+00 -5.75268924e-01
8.58860373e-01 -1.07527040e-01 2.76935119e-02 2.94924825e-01
5.83743632e-01 4.56001669e-01 -6.68040633e-01 -6.17498934e-01
7.54033089e-01 -3.99957806e-01 -4.48375404e-01 -1.78068563e-01
-1.06789958e+00 9.15749311e-01 9.48957205e-02 5.53130567e-01
-5.56809545e-01 7.49110818e-01 2.13262308e-02 4.07011390e-01
7.84310251e-02 3.01614553e-01 -4.97249097e-01 -1.57890812e-01
-5.87482810e-01 3.02498817e-01 9.93244946e-01 6.79579079e-01
4.34164256e-01 -3.98246169e-01 -2.64924854e-01 6.62440807e-02
6.48613334e-01 6.84693158e-01 -3.35610546e-02 -5.08339226e-01
1.45618647e-01 3.02327991e-01 1.28610298e-01 -8.89476299e-01
-3.18611860e-01 -5.12621105e-01 -1.01764977e+00 -3.10265385e-02
8.25682163e-01 -1.86412558e-01 -5.83955586e-01 2.26540518e+00
-1.02867922e-02 5.25103867e-01 1.26316369e-01 7.12036908e-01
5.27432621e-01 5.35953522e-01 2.73755461e-01 -6.95288897e-01
1.18519711e+00 4.83860582e-01 -8.46406341e-01 3.21187556e-01
4.38898057e-01 -4.65207666e-01 5.50408661e-01 4.40520793e-01
-5.78292489e-01 -2.74000376e-01 -8.27172339e-01 6.49078667e-01
-1.58168867e-01 -4.30924028e-01 5.67640543e-01 9.04144645e-01
-8.11280429e-01 5.18705547e-01 -4.59979057e-01 -3.12375158e-01
1.35073394e-01 2.93574601e-01 -2.97725797e-01 1.29800707e-01
-1.67536604e+00 4.52646703e-01 3.16753745e-01 -1.88520417e-01
-1.39901888e+00 -6.47160828e-01 -4.15983915e-01 1.58416688e-01
3.70045513e-01 -6.16267383e-01 8.11321259e-01 -6.59944057e-01
-9.98710155e-01 2.40394652e-01 -1.64669007e-01 -7.44135618e-01
5.48681438e-01 2.61237714e-02 -3.83131683e-01 -1.55555651e-01
1.58986002e-01 1.73569441e-01 8.84364486e-01 -1.74348378e+00
-3.27329636e-01 -3.75359237e-01 -9.18358639e-02 -3.35731566e-01
7.97287077e-02 -2.43378833e-01 8.28613564e-02 -3.97830427e-01
1.78894907e-01 -7.73768961e-01 -3.29175591e-01 -4.11805511e-01
-9.95001733e-01 -3.44978720e-02 3.69742423e-01 -1.91061631e-01
1.63716865e+00 -2.06607032e+00 4.02231216e-02 8.22901309e-01
3.38122457e-01 -4.75772440e-01 1.35561943e-01 6.02057338e-01
-3.36817861e-01 4.51851070e-01 -4.43998724e-01 1.22362643e-01
-4.90575582e-02 2.34022379e-01 -3.04848820e-01 8.14262450e-01
-8.00231006e-03 5.43605089e-01 -9.48438883e-01 -3.25400025e-01
-4.00052294e-02 2.97881991e-01 -3.29382360e-01 1.23521574e-01
-2.40861461e-01 6.90060675e-01 -4.77271914e-01 2.89346993e-01
6.71131611e-01 -3.42774391e-01 4.56974685e-01 1.65398777e-01
-3.07314813e-01 1.45493656e-01 -1.62996435e+00 1.14338946e+00
-3.26130241e-01 4.51835722e-01 -2.10651323e-01 -1.01646566e+00
1.08157694e+00 6.38137281e-01 3.40652764e-01 -4.77403611e-01
2.79079258e-01 3.14888597e-01 3.92462820e-01 -3.07166278e-01
-1.52334701e-02 -2.34253868e-01 -2.39989460e-01 2.61604309e-01
2.06066072e-01 3.30260545e-01 -4.13878746e-02 2.08530620e-01
1.57233834e+00 -3.94096911e-01 6.65360153e-01 -6.12377286e-01
1.52772665e-01 -5.67313015e-01 4.00138855e-01 1.50428319e+00
5.36459088e-02 4.44692850e-01 1.11121464e+00 2.48266950e-01
-8.09816599e-01 -1.38998282e+00 -5.93929172e-01 3.78583550e-01
3.02175492e-01 -5.19795716e-01 -6.09199941e-01 -3.39623213e-01
2.62981266e-01 8.45038831e-01 -1.00182891e+00 -8.33369866e-02
5.07031046e-02 -1.09870160e+00 7.33416677e-01 8.17672536e-03
3.09849828e-01 -5.54321170e-01 -5.00987768e-01 3.06659847e-01
-1.17976412e-01 -6.56476557e-01 3.57270539e-01 7.83318579e-01
-6.49155557e-01 -1.21376991e+00 -3.22359085e-01 -7.82052949e-02
2.32686430e-01 -4.18068856e-01 1.05305159e+00 -2.00375721e-01
-7.12360430e-04 2.92483002e-01 -3.63023072e-01 -4.78948683e-01
-6.81780457e-01 -4.91257101e-01 1.39388934e-01 2.77919412e-01
2.53491879e-01 -7.29224980e-01 -3.55649114e-01 1.41398877e-01
-9.85657632e-01 -4.05930817e-01 7.13816047e-01 7.25774586e-01
5.66963732e-01 6.48244858e-01 6.01058245e-01 -8.51417720e-01
5.66150844e-01 -9.83467400e-01 -5.85145354e-01 -1.28551712e-02
-7.76312053e-01 3.22715640e-01 6.64933026e-01 -2.39700347e-01
-9.17225897e-01 9.13039520e-02 7.52531216e-02 2.87569240e-02
-5.08611143e-01 7.64150798e-01 -5.13022125e-01 2.87273943e-01
7.85939217e-01 1.01694517e-01 -2.51302540e-01 -4.45801377e-01
4.83701408e-01 5.17390430e-01 3.67960513e-01 -6.44424021e-01
6.64103568e-01 6.73464000e-01 6.79070532e-01 -7.06045568e-01
-4.15304959e-01 -2.68176347e-01 -5.54622173e-01 -1.41264796e-01
9.43904936e-01 -5.81705213e-01 -1.06583190e+00 1.30647793e-01
-1.36277330e+00 3.09119914e-02 -3.08337748e-01 7.71071076e-01
-6.40964925e-01 2.61378214e-02 -3.09350878e-01 -1.37097514e+00
4.23721582e-01 -8.44012737e-01 8.04136038e-01 -1.62316144e-01
-4.11415040e-01 -1.18458486e+00 1.03159942e-01 -2.87240416e-01
8.83602202e-02 4.55563724e-01 1.21624124e+00 -6.76814675e-01
-4.91851032e-01 -2.52529830e-01 -1.16647087e-01 -1.91733196e-01
-6.01022653e-02 -1.25160581e-02 -9.89434898e-01 2.01755166e-01
4.31502871e-02 2.26995260e-01 1.02043009e+00 9.38016176e-01
9.83727753e-01 -3.11527252e-01 -6.19708300e-01 5.47851808e-02
1.73794460e+00 3.20897192e-01 7.54016161e-01 2.47195572e-01
2.76519775e-01 4.87385482e-01 3.50406736e-01 8.32607150e-01
-1.00106239e-01 5.17786384e-01 7.86000371e-01 -2.00608894e-02
1.77635297e-01 -6.34213686e-01 1.71555996e-01 1.71552405e-01
-1.13689288e-01 -6.38263226e-01 -7.07220197e-01 4.44408029e-01
-1.88218653e+00 -1.18065238e+00 -9.52207148e-01 2.60182071e+00
8.43882680e-01 3.00415576e-01 -1.29289152e-02 3.91040027e-01
8.48507345e-01 -1.56141266e-01 -6.85347244e-02 -1.50619507e-01
-4.42658961e-01 2.23533884e-01 8.66162658e-01 6.70130014e-01
-1.03216124e+00 1.29243478e-01 6.61343098e+00 1.03206432e+00
-6.17429137e-01 2.24272400e-01 6.64249778e-01 4.67228174e-01
-6.60288453e-01 4.48598355e-01 -6.06810927e-01 7.68741429e-01
1.23555601e+00 -1.55852228e-01 1.62213147e-01 2.99098790e-01
5.90411484e-01 -4.87166256e-01 -1.39054835e+00 6.83400989e-01
-5.68311930e-01 -1.10211861e+00 -2.78194621e-02 4.75516677e-01
6.33191526e-01 -4.55905735e-01 1.17809489e-01 -2.30648994e-01
5.42526603e-01 -1.37671363e+00 7.33321905e-01 7.81143546e-01
7.57032156e-01 -5.77454805e-01 1.00179088e+00 2.66061902e-01
-1.01669908e+00 -9.42636430e-02 -1.90287039e-01 -4.34480131e-01
1.59240186e-01 1.26169813e+00 -5.65706313e-01 6.57777905e-01
4.63567108e-01 4.25238341e-01 -2.26654395e-01 9.67594266e-01
-2.74555773e-01 1.20611870e+00 -3.61461163e-01 -1.64302081e-01
-1.00625224e-01 2.62771570e-03 8.89933765e-01 1.27612221e+00
3.51287901e-01 -1.74737275e-02 -1.43994018e-01 1.30395401e+00
2.51235757e-02 4.68043871e-02 -7.94225931e-01 9.66018960e-02
8.48566413e-01 6.17878079e-01 -7.83059716e-01 -8.65864977e-02
-3.00133169e-01 4.79493946e-01 -3.39086473e-01 3.83711934e-01
-7.99277067e-01 1.04452074e-01 3.70400757e-01 3.43213324e-03
2.02436075e-02 2.19013408e-01 -6.98237360e-01 -5.69702506e-01
-2.29963049e-01 -5.72600961e-01 3.16981137e-01 -3.74747723e-01
-1.25391531e+00 1.57459661e-01 1.65137768e-01 -1.38896763e+00
-1.08836539e-01 -2.54396886e-01 -4.37305510e-01 1.01014948e+00
-9.27897692e-01 -5.86834788e-01 6.19697161e-02 5.64244866e-01
-2.16447026e-01 1.52129754e-01 1.00392640e+00 2.57277846e-01
-1.93793297e-01 2.51508802e-01 1.96741104e-01 -1.21307321e-01
4.08956409e-01 -1.57492793e+00 -1.25777543e-01 7.32458949e-01
3.76351885e-02 7.88642526e-01 1.30824530e+00 -9.77665067e-01
-1.41528106e+00 -1.00035584e+00 8.52772295e-01 -5.34399927e-01
9.28231120e-01 -4.90355223e-01 -7.13722944e-01 3.46751213e-01
-2.37510866e-03 -1.74076021e-01 5.82213283e-01 4.78859156e-01
-2.45888859e-01 1.86004966e-01 -9.77980793e-01 3.47783625e-01
9.61846650e-01 -2.55491346e-01 -2.80231833e-01 1.44970506e-01
6.17460430e-01 2.14155361e-01 -9.53074932e-01 3.29444319e-01
4.80291396e-01 -1.06358349e+00 6.20444357e-01 -4.18616802e-01
6.38614297e-01 -4.89555806e-01 -5.87288082e-01 -1.30028856e+00
-3.83317202e-01 -6.23511910e-01 3.54913324e-02 1.32054341e+00
6.19238198e-01 -5.23341477e-01 3.10098827e-01 2.51043111e-01
4.53673989e-01 -4.17065889e-01 -1.28965688e+00 -9.88606095e-01
3.50910313e-02 -9.31294203e-01 4.45613474e-01 8.96627426e-01
4.44135658e-04 4.27597255e-01 -4.84588504e-01 4.11977768e-01
8.35161924e-01 -1.16945982e-01 4.13986146e-01 -1.74465954e+00
-5.77794790e-01 -3.27668935e-01 -7.44102955e-01 -7.48921633e-01
-3.18139158e-02 -7.54225075e-01 1.24076016e-01 -1.18066251e+00
4.49016720e-01 -5.30304074e-01 -3.67638975e-01 9.01372582e-02
-4.18490730e-03 -5.24993129e-02 -5.95890991e-02 1.48843959e-01
-2.16210887e-01 3.91391098e-01 8.52841794e-01 -1.49210927e-03
-2.38247663e-01 8.93749818e-02 -7.24475086e-01 7.05363095e-01
4.20035392e-01 -8.52736890e-01 -1.84960678e-01 3.82564425e-01
5.18800378e-01 4.47272539e-01 7.90153682e-01 -7.32829034e-01
3.21306475e-02 -1.49573848e-01 2.35857293e-01 -2.07787052e-01
2.68181413e-02 -9.31201100e-01 5.75721562e-01 6.74105287e-01
-6.23955071e-01 -3.79628092e-01 -8.51777196e-02 1.07401669e+00
-2.11897850e-01 -2.63440579e-01 4.92233157e-01 1.90192405e-02
-4.27748799e-01 -2.17453852e-01 -4.69247729e-01 -1.87689170e-01
8.52095068e-01 1.87118649e-01 -4.16495115e-01 -7.12171555e-01
-7.70178497e-01 -1.98914483e-01 -1.26625597e-02 5.36771864e-02
4.58281428e-01 -1.15345895e+00 -8.14296007e-01 -1.20799549e-01
6.31560013e-02 -5.21621466e-01 -1.17859608e-02 1.04609191e+00
1.83854237e-01 6.15980864e-01 4.28813249e-01 -6.24623954e-01
-9.61092591e-01 6.31333411e-01 1.87529445e-01 -1.46358356e-01
-9.65257138e-02 7.15024054e-01 5.10559440e-01 -4.19092998e-02
6.32031187e-02 -2.70189077e-01 -1.96563587e-01 5.24344146e-02
4.51970756e-01 4.59143907e-01 -1.09253623e-01 -3.64925802e-01
-3.92396003e-01 1.32603437e-01 6.42748296e-01 -4.23068196e-01
1.03983247e+00 -1.61407202e-01 -3.72074276e-01 9.00508761e-01
1.15990198e+00 2.94286370e-01 -8.66664886e-01 8.40984061e-02
1.13566615e-01 -3.95442337e-01 5.09466827e-02 -9.45913970e-01
-4.65180874e-01 4.22600180e-01 8.46789956e-01 9.36348677e-01
9.82013047e-01 3.84091407e-01 -1.77387953e-01 -4.95438948e-02
4.72083271e-01 -6.72385335e-01 -4.33443606e-01 1.97562501e-01
7.72256851e-01 -9.27234888e-01 -2.62151599e-01 -6.62708163e-01
-1.26292914e-01 9.87849832e-01 -8.71318877e-02 -2.53794808e-02
1.01874065e+00 5.47420204e-01 -3.67456108e-01 -2.04271361e-01
-8.51071596e-01 -4.33482945e-01 3.46123753e-03 5.93196213e-01
5.87172508e-01 5.53872645e-01 -8.64344954e-01 7.67715394e-01
-1.08560108e-01 6.98266597e-03 7.89806902e-01 4.88254040e-01
-2.68117279e-01 -9.28299189e-01 -7.75753021e-01 5.62009394e-01
-5.18015683e-01 -3.63533646e-01 -4.66497242e-01 8.64908755e-01
3.38818699e-01 1.47727013e+00 9.06258747e-02 -4.82089758e-01
1.73184514e-01 -1.33791804e-01 1.48133531e-01 -3.09081852e-01
-1.06768981e-01 2.78820455e-01 6.50926158e-02 -4.53055322e-01
-4.88472581e-01 -8.38042200e-01 -1.16381061e+00 -3.26157540e-01
-4.75387841e-01 2.71091372e-01 5.91289282e-01 1.17172933e+00
3.83476615e-02 6.37580514e-01 5.00826299e-01 5.04930876e-02
-5.14904618e-01 -9.59896266e-01 -1.05171037e+00 3.41226757e-01
3.87281299e-01 -8.69431794e-01 -6.18069232e-01 1.81730986e-02]
|
[7.806193828582764, 5.283635139465332]
|
1fd05aff-9ff6-4530-9fd7-ead079e867b9
|
tightly-coupled-learning-strategy-for-weakly
|
2202.06470
| null |
https://arxiv.org/abs/2202.06470v1
|
https://arxiv.org/pdf/2202.06470v1.pdf
|
Tightly Coupled Learning Strategy for Weakly Supervised Hierarchical Place Recognition
|
Visual place recognition (VPR) is a key issue for robotics and autonomous systems. For the trade-off between time and performance, most of methods use the coarse-to-fine hierarchical architecture, which consists of retrieving top-N candidates using global features, and re-ranking top-N with local features. However, since the two types of features are usually processed independently, re-ranking may harm global retrieval, termed re-ranking confusion. Moreover, re-ranking is limited by global retrieval. In this paper, we propose a tightly coupled learning (TCL) strategy to train triplet models. Different from original triplet learning (OTL) strategy, it combines global and local descriptors for joint optimization. In addition, a bidirectional search dynamic time warping (BS-DTW) algorithm is also proposed to mine locally spatial information tailored to VPR in re-ranking. The experimental results on public benchmarks show that the models using TCL outperform the models using OTL, and TCL can be used as a general strategy to improve performance for weakly supervised ranking tasks. Further, our lightweight unified model is better than several state-of-the-art methods and has over an order of magnitude of computational efficiency to meet the real-time requirements of robots.
|
['N. Zheng', 'W. Zuo', 'R. Wang', 'Y. Shen']
|
2022-02-14
| null | null | null | null |
['visual-place-recognition']
|
['computer-vision']
|
[-2.44127765e-01 -6.67758584e-01 -5.13718784e-01 -1.99719593e-01
-1.22433221e+00 -3.63877952e-01 6.37022078e-01 1.80648804e-01
-4.28548783e-01 4.36362088e-01 1.66943714e-01 1.45048246e-01
-5.94702125e-01 -6.88211679e-01 -6.43958807e-01 -7.81015456e-01
-2.32289821e-01 5.50677836e-01 6.75236225e-01 -4.77135986e-01
6.23468399e-01 6.37552142e-01 -1.90767705e+00 3.44492942e-01
8.00906539e-01 1.01054370e+00 7.18428791e-01 1.73972156e-02
9.18160826e-02 7.42453575e-01 -2.44393408e-01 1.38207018e-01
4.03136790e-01 8.51092860e-02 -8.12852681e-01 -5.51952422e-01
2.28541091e-01 -2.61045545e-01 -4.43412334e-01 9.25372303e-01
6.40532672e-01 4.86254662e-01 5.49718499e-01 -1.34990835e+00
-8.24138761e-01 2.60420263e-01 -6.38623893e-01 -1.20745897e-01
4.38608617e-01 -1.11164644e-01 1.45300770e+00 -1.20708930e+00
5.19157112e-01 1.42332327e+00 4.01280731e-01 2.41614506e-01
-8.14083695e-01 -6.02090418e-01 3.47190112e-01 7.40947962e-01
-1.63790333e+00 -2.48767659e-01 7.20324457e-01 -3.27929944e-01
1.01828039e+00 1.91228569e-01 5.47482729e-01 7.47610986e-01
2.46624812e-01 9.76016104e-01 9.45718348e-01 -6.27699792e-02
8.49187747e-02 -3.40940297e-01 3.61132473e-02 6.99423254e-01
1.21926561e-01 1.69220254e-01 -1.00548387e+00 -4.12405521e-01
7.33722389e-01 3.82802308e-01 8.84683952e-02 -7.68507540e-01
-1.71468151e+00 6.72268987e-01 8.80865753e-01 1.87826172e-01
-2.64931649e-01 3.06257635e-01 5.40215373e-01 2.01627597e-01
4.82638329e-02 4.53426480e-01 -3.64289194e-01 8.62160772e-02
-5.75883865e-01 1.58590302e-01 2.90403366e-01 1.07135201e+00
1.18203676e+00 -4.24421251e-01 -3.79005998e-01 1.32866728e+00
7.15568841e-01 4.31899428e-01 8.12122524e-01 -8.64367545e-01
5.22455752e-01 6.29960716e-01 1.48774773e-01 -1.30550337e+00
-3.34357440e-01 -4.83378991e-02 -7.03034937e-01 -2.67004091e-02
-1.96600854e-01 6.30892396e-01 -1.13998282e+00 1.37789190e+00
4.54383790e-02 3.28964218e-02 -2.01051347e-02 1.07454693e+00
8.97981226e-01 9.42667246e-01 -1.16183609e-01 -2.48863138e-02
1.05743349e+00 -1.48846364e+00 -3.95568550e-01 -3.90322417e-01
5.55913925e-01 -7.69876003e-01 8.82211983e-01 1.02119178e-01
-6.46910548e-01 -4.61660415e-01 -1.14745939e+00 -3.44856292e-01
-5.66329777e-01 2.35252231e-01 6.72606349e-01 -5.17239347e-02
-1.11131740e+00 6.67251050e-01 -6.93016708e-01 -4.92413670e-01
4.63530421e-02 5.33738196e-01 -5.01227796e-01 -2.50737548e-01
-1.13828838e+00 1.03068733e+00 4.43745226e-01 3.01680207e-01
-1.15909076e+00 -2.20238015e-01 -7.21990764e-01 -1.83811292e-01
1.81633770e-01 -3.10679704e-01 1.00164402e+00 -1.72520801e-01
-1.61277330e+00 6.31194890e-01 -2.99663782e-01 -1.08437255e-01
2.96877980e-01 -2.64860272e-01 -1.46951899e-01 1.25243897e-02
5.77522337e-01 6.96251154e-01 6.04361951e-01 -1.29249978e+00
-9.47842658e-01 -4.64497417e-01 6.29134849e-02 6.13884091e-01
-3.55325907e-01 -1.20581634e-01 -8.61681283e-01 -3.62930447e-01
8.07608962e-01 -9.54952538e-01 -4.12310928e-01 -2.19702497e-01
-1.15384152e-02 -6.13225341e-01 8.13448608e-01 -3.70893538e-01
1.01359546e+00 -2.18549395e+00 2.58068651e-01 4.40916359e-01
9.77329463e-02 -8.13641995e-02 -4.57307249e-01 5.87034404e-01
3.21939737e-01 -8.26032013e-02 6.54722750e-02 -2.37185046e-01
1.70972943e-01 3.30059201e-01 -3.23565423e-01 5.95373034e-01
-1.19389154e-01 8.96336913e-01 -1.35115719e+00 -7.03766108e-01
3.69319856e-01 1.53263897e-01 -3.38778257e-01 -2.83340836e-04
1.68755203e-01 2.78662115e-01 -7.13987410e-01 1.03199804e+00
3.48064333e-01 -2.86248356e-01 -1.64428037e-02 -2.28989482e-01
-3.53716582e-01 3.22942406e-01 -1.00200307e+00 1.88269758e+00
-3.55853975e-01 3.78595233e-01 -3.94488990e-01 -8.60280454e-01
1.26729167e+00 -1.19164199e-01 6.56254053e-01 -1.15306997e+00
-2.17559069e-01 7.36656785e-01 -4.81758147e-01 -2.23748982e-01
7.65574634e-01 4.39693511e-01 -3.16707790e-01 -3.43344584e-02
-4.92862985e-02 1.72650650e-01 1.50837973e-01 -9.75353122e-02
1.02991402e+00 3.79019797e-01 2.66174197e-01 -1.57202795e-01
4.87367868e-01 1.36450142e-01 8.50089073e-01 8.61945987e-01
-3.34980279e-01 6.13571882e-01 -3.33886445e-02 -4.91297901e-01
-8.28669190e-01 -8.86958778e-01 -1.84001215e-02 1.46870768e+00
1.01081252e+00 -6.15312040e-01 3.27869691e-02 -4.99506176e-01
7.97405094e-02 8.73648077e-02 -3.77724946e-01 -2.49019578e-01
-6.28657699e-01 -6.78368092e-01 4.41198140e-01 5.13392568e-01
6.08054101e-01 -1.11674869e+00 -6.07470095e-01 2.05418214e-01
-5.38856566e-01 -8.38385582e-01 -3.64726543e-01 2.85949528e-01
-8.56097519e-01 -8.61500323e-01 -8.34022582e-01 -1.23569214e+00
4.80364919e-01 9.56493318e-01 8.78972411e-01 1.44850105e-01
6.99764192e-02 2.24814191e-01 -8.65621507e-01 1.27540395e-01
3.26289028e-01 1.48543760e-01 2.01660395e-01 -1.02128819e-01
2.22743303e-01 -4.44021076e-01 -5.98709762e-01 7.23442197e-01
-5.76324344e-01 -1.54961705e-01 1.00604177e+00 1.07923734e+00
1.01687479e+00 -2.90653467e-01 4.06554937e-01 -1.35582119e-01
4.71239775e-01 -4.17013019e-01 -7.05574691e-01 5.90645909e-01
-6.62490070e-01 3.01993102e-01 3.75615209e-01 -4.95873243e-01
-6.18843794e-01 1.98368549e-01 2.85208195e-01 -6.73520565e-01
2.03888804e-01 7.45284557e-01 6.55653700e-03 -4.39723462e-01
3.76565963e-01 4.08967495e-01 -2.92405784e-01 -4.86075640e-01
2.85321355e-01 6.60191596e-01 2.48796210e-01 -5.89183390e-01
8.39165330e-01 4.64840561e-01 5.66722415e-02 -5.97640634e-01
-6.73206627e-01 -1.02483702e+00 -5.31575084e-01 -1.87438786e-01
5.62282443e-01 -1.08520651e+00 -5.78053415e-01 4.36354816e-01
-1.05253208e+00 -1.68485537e-01 1.48366734e-01 5.95308721e-01
-6.66198671e-01 4.66125071e-01 -3.26775551e-01 -6.34414852e-01
-1.35866329e-01 -1.43065679e+00 1.40870917e+00 7.86144361e-02
3.99606168e-01 -5.26578069e-01 2.27557883e-01 2.68275857e-01
2.95571834e-01 -5.68263605e-02 5.73270380e-01 -6.39536798e-01
-1.07292414e+00 -2.24400327e-01 -4.28260714e-01 -9.99963135e-02
4.06409986e-02 -2.12485775e-01 -7.08759964e-01 -4.52299356e-01
-4.55048054e-01 -5.36365151e-01 9.60840821e-01 1.17559321e-01
9.83186662e-01 -5.72520234e-02 -6.05134785e-01 4.90098774e-01
1.45725214e+00 1.68017283e-01 6.45302773e-01 9.65182602e-01
9.82120633e-01 4.49579358e-01 1.35043299e+00 4.54812974e-01
7.89749026e-01 9.43459392e-01 5.96871972e-01 1.37188956e-01
1.93911150e-01 -3.59328032e-01 5.56388795e-01 9.77274656e-01
-2.10980192e-01 2.84464985e-01 -1.06977248e+00 7.96217740e-01
-2.54204893e+00 -9.20441568e-01 9.38854292e-02 2.48261428e+00
5.56152463e-01 -1.96321890e-01 -1.55490592e-01 -2.36588836e-01
8.42337906e-01 4.40224200e-01 -4.55488443e-01 1.22880004e-01
-1.95342712e-02 -4.21353817e-01 8.48806441e-01 3.73604476e-01
-1.15775001e+00 1.03914690e+00 5.92812109e+00 1.20747912e+00
-1.12150812e+00 1.90739140e-01 1.50502294e-01 3.69645655e-02
-8.48596692e-02 1.22375496e-01 -9.42214608e-01 2.41095796e-01
3.38110328e-01 -3.77450921e-02 6.56261683e-01 1.03627861e+00
-1.32447734e-01 7.90683106e-02 -1.11940515e+00 1.43183446e+00
1.31595135e-01 -1.04579902e+00 2.61803061e-01 1.85685605e-01
7.57110953e-01 6.08560503e-01 5.70307709e-02 6.38036489e-01
3.06124568e-01 -9.11589205e-01 9.26047921e-01 4.31161225e-01
4.88119185e-01 -6.87289715e-01 7.13419437e-01 2.67413229e-01
-1.86805856e+00 -4.97154742e-01 -8.34893942e-01 2.83643484e-01
-1.23833202e-01 2.78676391e-01 -4.23695058e-01 9.13543344e-01
1.26327884e+00 1.07242787e+00 -7.37276554e-01 1.48654306e+00
-5.69738774e-03 -3.17799538e-01 -5.24158001e-01 -2.75428027e-01
4.66910273e-01 -1.85899585e-01 5.88563800e-01 8.38697135e-01
4.27495360e-01 -1.42378315e-01 5.89634836e-01 2.64006138e-01
1.29626378e-01 2.42025644e-01 -7.24955440e-01 3.07557970e-01
7.58386791e-01 1.27684760e+00 -5.12790859e-01 -1.40615940e-01
-2.99438387e-01 8.58132184e-01 7.50329435e-01 2.51898557e-01
-7.97491610e-01 -4.77256417e-01 3.52160215e-01 -4.42524761e-01
2.92616040e-01 -4.20397252e-01 2.45228969e-02 -1.02214837e+00
3.40873033e-01 -5.39890885e-01 3.66609395e-01 -6.46811724e-01
-1.40116394e+00 5.04152477e-01 3.73210311e-02 -2.01560259e+00
2.15966068e-02 -5.95830083e-01 -1.67952865e-01 5.44096351e-01
-2.04888463e+00 -1.37499809e+00 -3.65940928e-01 6.62378669e-01
5.76281548e-01 -1.09367967e-01 5.98673820e-01 3.96049619e-01
-2.37078935e-01 4.08027738e-01 4.69766289e-01 -6.28534779e-02
9.28251743e-01 -1.02929640e+00 -5.75573072e-02 5.86617887e-01
4.10557762e-02 6.92994595e-01 2.38313541e-01 -5.71210921e-01
-1.63035989e+00 -1.21039033e+00 9.50640857e-01 -3.63693833e-01
6.05450809e-01 1.12245686e-03 -5.32682240e-01 3.80147249e-01
-1.42831624e-01 1.97863683e-01 4.72197652e-01 2.08084047e-01
-5.65998375e-01 -4.32555616e-01 -7.32014000e-01 6.96740508e-01
1.24880922e+00 -6.53100014e-01 -7.48159230e-01 4.93744522e-01
8.06153595e-01 -4.22713459e-01 -8.52703035e-01 8.49841714e-01
7.47623324e-01 -8.46273661e-01 1.20019674e+00 -7.32916519e-02
1.38029411e-01 -8.41054022e-01 -6.64315641e-01 -1.07249439e+00
-6.96719885e-01 -3.43942702e-01 4.30620089e-02 1.02552247e+00
4.57343310e-01 -6.55133426e-01 3.49009454e-01 1.05911538e-01
-3.17964315e-01 -7.21995890e-01 -9.70531762e-01 -1.05238855e+00
-3.18296552e-01 9.35400464e-03 5.71053505e-01 7.58231461e-01
6.24754429e-02 2.93226302e-01 -4.99233156e-01 4.27481145e-01
6.23321295e-01 5.08768260e-01 5.72234750e-01 -1.45184016e+00
9.42729786e-02 -4.20432568e-01 -5.28484285e-01 -1.37968361e+00
6.22321554e-02 -9.18961823e-01 6.20973587e-01 -1.90931296e+00
3.50032419e-01 -7.03464270e-01 -8.80924106e-01 7.15647876e-01
8.63850787e-02 2.85396248e-01 3.25531483e-01 9.89625812e-01
-1.38911688e+00 9.72335517e-01 1.14682949e+00 -3.37253779e-01
-3.85133147e-01 -3.50123137e-01 -3.46802801e-01 5.51587641e-01
6.11129165e-01 -5.03659844e-01 -2.52187133e-01 -6.30882025e-01
3.23002309e-01 -8.85026827e-02 3.18492442e-01 -1.10730505e+00
7.30774045e-01 -4.06654477e-01 1.57991990e-01 -1.06627440e+00
6.07298851e-01 -9.40539718e-01 -1.69363409e-01 2.60999113e-01
-1.88630491e-01 2.65327066e-01 -2.37779960e-01 7.96091378e-01
-5.22754848e-01 -9.60043669e-02 5.18166363e-01 -1.14136316e-01
-1.28164876e+00 5.48480451e-01 -1.48323461e-01 -4.18513507e-01
9.87761378e-01 -2.57396638e-01 -4.27768677e-01 2.84368433e-02
-3.16057503e-01 7.70486236e-01 5.46943963e-01 8.27372372e-01
9.84894395e-01 -1.76327014e+00 -4.72110033e-01 -1.88512579e-01
7.39408731e-01 2.90044010e-01 -4.26321896e-03 8.44301641e-01
-5.77218115e-01 5.97406507e-01 -2.41518036e-01 -8.11372697e-01
-1.01145589e+00 4.43190515e-01 4.71317880e-02 -4.27943677e-01
-4.84888345e-01 6.67593718e-01 1.13816924e-01 -8.21077228e-01
5.19144237e-01 9.76520255e-02 -5.60351372e-01 1.81417823e-01
2.80262947e-01 4.68840718e-01 1.71080440e-01 -8.28203142e-01
-7.62203991e-01 9.76864219e-01 -1.16100900e-01 -1.04127452e-01
1.34408557e+00 -3.54007870e-01 -5.45080781e-01 6.30354881e-01
1.19178069e+00 -2.30185926e-01 -9.37674582e-01 -3.65864277e-01
2.60352612e-01 -4.95455384e-01 -1.34699047e-02 -3.92694294e-01
-7.74610341e-01 5.19471526e-01 5.78674912e-01 -2.73860320e-02
8.95608187e-01 8.61012489e-02 5.20549119e-01 1.11709309e+00
1.11932147e+00 -1.30308330e+00 4.15225446e-01 1.00755274e+00
1.04681730e+00 -1.29857743e+00 -5.49561903e-02 -1.94290534e-01
-6.30898058e-01 8.94902706e-01 8.57021034e-01 -3.80424589e-01
4.60208476e-01 -5.07848263e-01 -8.04629028e-02 -1.89469814e-01
-7.33570516e-01 -3.73491883e-01 4.42990839e-01 4.65445668e-01
-5.51195554e-02 2.37463415e-03 -2.58470088e-01 1.88630432e-01
7.31661022e-02 -2.95255095e-01 -6.84993863e-02 1.16904306e+00
-7.37537920e-01 -1.01511419e+00 -3.50132436e-01 3.87537628e-01
8.84148180e-02 -7.97296911e-02 -2.02282727e-01 4.53884542e-01
-1.68760810e-02 9.74562883e-01 -3.44763607e-01 -7.72866070e-01
3.25348318e-01 -3.35015774e-01 2.00795457e-01 -5.16548395e-01
-3.25232118e-01 2.73568798e-02 -1.16257422e-01 -1.03803825e+00
-4.69562650e-01 -5.68007886e-01 -1.24574804e+00 6.58861771e-02
-4.97212052e-01 1.33525476e-01 5.21118641e-01 9.65823114e-01
4.89374131e-01 3.22838336e-01 9.35099661e-01 -1.33913696e+00
-6.38374388e-01 -7.48254895e-01 -4.61081117e-01 1.50363907e-01
3.78260881e-01 -1.09732950e+00 -3.16005498e-01 -4.01482254e-01]
|
[7.742311954498291, -1.9380050897598267]
|
3cc47838-4e70-468c-9572-8229ecc4bbea
|
a-dual-encoding-system-for-dialect
| null | null |
https://aclanthology.org/2020.vardial-1.20
|
https://aclanthology.org/2020.vardial-1.20.pdf
|
A dual-encoding system for dialect classification
|
In this paper we present the architecture, processing pipeline and results of the ensemble model developed for Romanian Dialect Identification task. The ensemble model consists of two TF-IDF encoders and a deep learning model aimed together at classifying input samples based on the writing patterns which are specific to each of the two dialects. Although the model performs well on the training set, its performance degrades heavily on the evaluation set. The drop in performance is due to the design decision which makes the model put too much weight on presence/lack of textual marks when determining the sample label.
|
['Dan Cristea', 'Petru Rebeja']
| null | null | null | null |
vardial-coling-2020-12
|
['dialect-identification']
|
['natural-language-processing']
|
[ 4.90419157e-02 -1.77979633e-01 -1.28542800e-02 -8.16760600e-01
-4.14623320e-01 -7.37054229e-01 8.09089780e-01 -1.13119997e-01
-4.70099419e-01 2.75249749e-01 5.68052053e-01 -3.33126664e-01
-2.55485445e-01 -7.05869794e-01 -5.07374592e-02 -3.73355627e-01
2.89872169e-01 9.34196293e-01 -1.65853754e-01 -3.23476970e-01
3.17501605e-01 4.90131050e-01 -1.68494678e+00 6.63131118e-01
7.80269504e-01 7.37880826e-01 6.35181963e-02 6.96802676e-01
-2.36913547e-01 8.57427001e-01 -7.86190212e-01 -6.20766461e-01
-2.45738029e-02 -3.00658226e-01 -1.00388300e+00 -2.51903329e-02
9.29268301e-01 -5.51793158e-01 -5.41547239e-01 6.84103191e-01
6.21686578e-01 -1.18389674e-01 8.99325550e-01 -7.76525676e-01
-6.53004467e-01 7.91025877e-01 -5.77767119e-02 2.53113031e-01
3.00951988e-01 -7.35813007e-02 1.22980964e+00 -8.75807166e-01
7.33231366e-01 1.01676667e+00 9.39382911e-01 5.93139470e-01
-1.23832929e+00 -4.00995314e-01 -7.40666389e-02 1.99087188e-01
-1.27567530e+00 -9.81201828e-01 3.83112997e-01 -5.69633067e-01
1.12814116e+00 3.10911119e-01 2.77644336e-01 9.86459792e-01
-2.07450479e-01 8.48930299e-01 1.32846797e+00 -5.90069413e-01
-9.43557993e-02 6.17476344e-01 5.71744025e-01 3.70999426e-01
-2.25370705e-01 1.00528188e-01 -4.14616555e-01 1.23056896e-01
2.45541304e-01 -4.07339692e-01 -1.54609680e-01 4.12600756e-01
-6.30439579e-01 5.62491775e-01 3.72545235e-02 7.63057530e-01
-2.40544334e-01 -3.58754337e-01 5.42936265e-01 5.89338243e-01
5.56194305e-01 3.30767304e-01 -5.95489264e-01 -3.25196594e-01
-1.41945279e+00 2.02069804e-01 8.15912127e-01 6.02914572e-01
2.47280642e-01 -3.64888869e-02 -6.79166198e-01 1.53576708e+00
2.99524516e-01 5.16952336e-01 4.93498504e-01 -5.55508971e-01
4.07260180e-01 6.26290798e-01 -1.73453718e-01 -5.69729984e-01
-3.79930735e-01 -4.64884639e-01 -1.78866521e-01 3.42362970e-01
7.90066898e-01 -2.71883190e-01 -1.17132080e+00 1.58465779e+00
-4.11987513e-01 -4.59965140e-01 -3.92275810e-01 7.05284536e-01
8.97869110e-01 4.36976135e-01 -7.32959956e-02 4.39702600e-01
1.24253166e+00 -8.02675843e-01 -7.09930003e-01 -1.80783153e-01
5.58680475e-01 -9.12203968e-01 1.00109696e+00 4.73518401e-01
-1.05583966e+00 -6.90366149e-01 -1.10661018e+00 -2.81447768e-01
-5.48490763e-01 5.96895397e-01 2.43431985e-01 1.16847491e+00
-1.26571405e+00 6.90852284e-01 -9.92160738e-02 -5.38381517e-01
2.80667126e-01 5.63000798e-01 -5.15522540e-01 -1.44264102e-01
-1.18498051e+00 1.23583364e+00 8.23101774e-03 5.10969236e-02
-3.94844115e-01 -4.97510463e-01 -6.49599552e-01 1.56394362e-01
-6.38686538e-01 -1.82470337e-01 1.28014004e+00 -1.24702215e+00
-1.56042683e+00 1.35303771e+00 -2.24346384e-01 -1.22070231e-01
7.40494907e-01 -1.56031605e-02 -9.31956589e-01 -2.71066636e-01
1.46254539e-01 4.82789904e-01 6.22779846e-01 -9.77198660e-01
-9.63811338e-01 -4.09393579e-01 -1.21463627e-01 9.83835235e-02
-4.20917064e-01 2.89291441e-01 -4.14616019e-01 -5.36044300e-01
-3.24042529e-01 -7.95766413e-01 3.26344162e-01 -6.86399996e-01
-2.27611274e-01 -4.59073424e-01 7.20960259e-01 -1.14325368e+00
1.59084105e+00 -2.12420297e+00 -1.97168618e-01 3.23913783e-01
-4.06438895e-02 4.22615588e-01 -7.40060806e-02 7.71792054e-01
-1.01267569e-01 8.26534107e-02 2.21634001e-01 -4.00114954e-01
1.80415019e-01 4.97199669e-02 -2.61999905e-01 3.19863230e-01
1.69315442e-01 4.26261514e-01 -6.93492830e-01 -2.19044507e-01
2.74680734e-01 4.24695581e-01 -1.75344139e-01 2.63394743e-01
1.72862098e-01 8.24896395e-02 2.56370664e-01 5.30239284e-01
8.83338034e-01 1.44999042e-01 2.44342685e-01 -1.95199810e-02
-3.94460768e-01 1.02345824e+00 -9.16965365e-01 1.23261786e+00
-6.50426149e-01 8.93795669e-01 1.61972731e-01 -4.54955012e-01
8.97989035e-01 4.09439683e-01 -1.50045529e-02 -6.45335555e-01
1.27590954e-01 5.06373167e-01 3.26391608e-01 -6.26203954e-01
6.19728982e-01 -1.49556458e-01 5.17612090e-03 7.97262788e-01
3.93156320e-01 3.91120136e-01 5.86707056e-01 -2.16511533e-01
8.84965360e-01 7.15616196e-02 -3.77500474e-01 -5.09060681e-01
8.12483668e-01 -1.76272258e-01 3.42539579e-01 7.33600259e-01
-9.35140327e-02 8.67938876e-01 4.10275728e-01 -4.04846638e-01
-1.04875362e+00 -8.17634642e-01 -4.98362839e-01 1.45321918e+00
-4.43124115e-01 -6.43336713e-01 -8.79379809e-01 -8.13300192e-01
6.94586486e-02 1.01651645e+00 -8.12241375e-01 -1.06892981e-01
-5.71922064e-01 -5.66699982e-01 9.54482675e-01 5.37961423e-01
3.46876115e-01 -1.11095929e+00 -1.18364289e-01 1.53641894e-01
-1.60187528e-01 -6.54404342e-01 -4.95321810e-01 2.78428942e-01
-3.49898040e-01 -6.71972990e-01 -6.02222502e-01 -9.05181646e-01
3.63630354e-01 -1.26748055e-01 1.17282891e+00 6.15905486e-02
1.50673762e-01 1.41425759e-01 -3.06499749e-01 -5.05909085e-01
-4.89383101e-01 5.76705098e-01 -1.87569961e-01 2.10306585e-01
9.61472273e-01 -1.70440778e-01 -3.53452802e-01 2.70845860e-01
-4.51962918e-01 -1.75743535e-01 3.81562352e-01 7.60240376e-01
-2.77795523e-01 1.47139346e-02 3.34956914e-01 -1.14504850e+00
7.75756598e-01 -3.62560868e-01 -1.05748735e-01 4.20619011e-01
-7.39478350e-01 1.23499207e-01 5.06231308e-01 -1.34305343e-01
-1.04842544e+00 2.34285109e-02 -8.37577164e-01 1.87401474e-01
-2.89750755e-01 4.04973865e-01 -2.37986043e-01 2.89463252e-01
5.50236821e-01 9.60149691e-02 -1.14415012e-01 -8.67070019e-01
-4.75004800e-02 1.58323061e+00 3.76140207e-01 -4.04582262e-01
1.68206424e-01 5.09852618e-02 -7.46054590e-01 -8.47374737e-01
-5.01551449e-01 -4.17392284e-01 -8.85238171e-01 -4.68151748e-01
6.18436337e-01 -6.50159419e-01 -3.89565587e-01 8.80782545e-01
-9.49729800e-01 -2.61663705e-01 1.69874765e-02 2.18167484e-01
-2.28172034e-01 -7.70170912e-02 -7.60517180e-01 -8.07500660e-01
-4.35726464e-01 -1.13874745e+00 8.81676853e-01 4.76768129e-02
-8.96458089e-01 -1.12896597e+00 1.79272622e-01 3.67928296e-01
5.53565741e-01 -2.97942132e-01 1.04657626e+00 -9.44089413e-01
2.02318728e-01 -4.26582515e-01 -1.13849640e-01 4.16825086e-01
1.86490789e-01 2.80053556e-01 -1.53521729e+00 -8.24666545e-02
-1.74048170e-01 -2.05946624e-01 9.23710644e-01 1.98066607e-01
8.78302217e-01 1.25419497e-01 -9.19795707e-02 4.50978518e-01
1.13823187e+00 1.72924399e-01 6.08524859e-01 3.30846667e-01
5.12898624e-01 9.28029835e-01 -3.66604999e-02 9.92766917e-02
4.64559495e-01 7.60805190e-01 -1.44593015e-01 1.40023425e-01
-4.58024442e-01 -3.59629542e-01 6.83963060e-01 6.13227785e-01
6.67192638e-02 -2.70250082e-01 -1.04012442e+00 5.74289799e-01
-1.55652857e+00 -1.01829743e+00 -4.50776428e-01 2.24343395e+00
6.53202891e-01 1.09315507e-01 4.35524702e-01 4.57178116e-01
7.89893866e-01 7.71069750e-02 -1.86073154e-01 -1.09270954e+00
-2.63642251e-01 4.06507939e-01 5.02095044e-01 7.00281501e-01
-1.01361120e+00 8.22940409e-01 7.72394848e+00 7.43869901e-01
-1.46952975e+00 -1.64468244e-01 5.95216036e-01 -3.15794125e-02
-1.55536130e-01 -4.34046924e-01 -1.04785681e+00 8.13072264e-01
1.29711545e+00 5.54162227e-02 4.62817580e-01 3.85204047e-01
1.06458321e-01 -7.53406659e-02 -1.11548913e+00 7.54554749e-01
1.62469015e-01 -9.98396099e-01 -3.72847840e-02 1.26787543e-01
2.12205991e-01 2.33329833e-01 2.45252877e-01 3.04460943e-01
3.42963129e-01 -1.30041003e+00 1.00348580e+00 3.00826162e-01
9.47116494e-01 -7.20611393e-01 8.20997715e-01 2.02089816e-01
-5.36759496e-01 -1.25341237e-01 -3.27799916e-02 -1.95340678e-01
5.02470788e-03 3.05577457e-01 -9.84232306e-01 2.31972978e-01
6.35565519e-01 6.70866430e-01 -7.53682375e-01 7.99207807e-01
-3.57493162e-01 9.93761539e-01 -2.68900365e-01 3.61228958e-02
2.34832793e-01 -2.13129073e-01 1.01527907e-01 1.77119577e+00
1.63759187e-01 -3.49611461e-01 -2.24273980e-01 5.78523576e-01
2.42455248e-02 9.41991284e-02 -3.95144224e-01 -8.49384218e-02
4.82308596e-01 1.44672430e+00 -5.56543946e-01 -3.61317545e-01
-4.75461572e-01 1.09346461e+00 5.52568376e-01 1.25277966e-01
-4.25561249e-01 -6.35120273e-01 7.10426688e-01 2.39783153e-01
3.10265362e-01 1.13890141e-01 -8.62762690e-01 -8.48456025e-01
-1.81964278e-01 -8.84876013e-01 6.01006866e-01 -6.06659591e-01
-1.36955178e+00 8.37928236e-01 -5.49869180e-01 -8.87497783e-01
-4.19733763e-01 -9.90731537e-01 -6.13372147e-01 1.54072726e+00
-9.51998115e-01 -1.26383185e+00 8.35452676e-02 2.43799016e-01
4.38984156e-01 -3.67692679e-01 9.64069784e-01 6.61303878e-01
-7.98266709e-01 8.99857402e-01 3.20026457e-01 5.90853155e-01
1.18327224e+00 -1.73999834e+00 4.93349463e-01 5.82555175e-01
1.59218729e-01 7.55788803e-01 7.00532556e-01 -3.87518853e-01
-5.92504442e-01 -7.67178953e-01 1.85064518e+00 -6.80058181e-01
5.05843580e-01 -6.08397663e-01 -6.33733749e-01 5.10254383e-01
6.57351911e-01 -7.20151544e-01 1.04432321e+00 7.17328548e-01
-5.35848618e-01 -9.19667631e-02 -1.33946824e+00 3.12214494e-01
7.41359234e-01 -8.22774827e-01 -7.34521031e-01 7.89607614e-02
-2.68535107e-01 -2.00104058e-01 -9.45881724e-01 6.57188296e-02
9.39188480e-01 -1.25042880e+00 4.48683918e-01 -4.27581191e-01
6.09939694e-01 -1.30346671e-01 3.71925570e-02 -1.60004270e+00
-8.15242529e-01 5.82405413e-03 3.85460258e-01 1.65946317e+00
7.61031866e-01 -4.01384085e-01 6.48185849e-01 7.28780389e-01
-1.28196580e-02 -2.62573540e-01 -6.46839142e-01 -5.00094831e-01
2.01413065e-01 -3.94389182e-01 6.85663283e-01 1.10102856e+00
2.75207758e-01 9.39944863e-01 -2.69768327e-01 -3.33575070e-01
4.89811115e-02 -1.45838661e-02 2.26603448e-01 -1.25093079e+00
-2.62945682e-01 -8.91860962e-01 -1.40148789e-01 -6.28259599e-01
2.08608925e-01 -1.04827070e+00 -1.14409581e-01 -1.37624574e+00
-9.45507139e-02 -5.06157398e-01 -2.42877066e-01 4.58513647e-01
-1.14754431e-01 3.48297089e-01 1.64754003e-01 2.09727153e-01
-1.92779198e-01 -1.52255341e-01 3.74771446e-01 -2.37716451e-01
-3.77990082e-02 1.92465678e-01 -6.48972631e-01 2.09727973e-01
9.02213156e-01 -6.87958598e-02 -2.45054677e-01 -8.81245255e-01
2.25460321e-01 -4.50788826e-01 -1.74846470e-01 -9.41108823e-01
3.57584804e-02 4.80263680e-01 7.47304916e-01 -5.13242722e-01
-4.35571000e-03 -7.19874918e-01 6.49105012e-02 3.46833915e-01
-7.40282118e-01 2.23736539e-01 1.68228090e-01 -3.88085127e-01
-2.42149040e-01 -5.64752460e-01 8.71049881e-01 1.22277357e-01
-5.92187166e-01 8.03784057e-02 -8.96459997e-01 -1.47364780e-01
2.84522235e-01 -3.43767881e-01 -2.05357552e-01 -3.87976468e-01
-7.95976162e-01 1.19039930e-01 5.81438780e-01 7.41811156e-01
-1.21862903e-01 -1.28385615e+00 -1.05143857e+00 6.89444423e-01
3.26262474e-01 -6.91635370e-01 2.22712010e-02 5.99883914e-01
-5.96869349e-01 5.51011443e-01 -5.44010222e-01 -2.56261438e-01
-1.48410344e+00 1.06040567e-01 4.66681927e-01 -3.35071057e-01
-1.80721492e-01 8.28784525e-01 -2.86173940e-01 -7.38166213e-01
7.99710080e-02 1.40176877e-01 -7.25694418e-01 4.96166021e-01
8.53038013e-01 7.92186618e-01 5.94050169e-01 -1.13085437e+00
-4.12718266e-01 2.95868307e-01 -6.06760681e-01 -4.84824449e-01
1.30728781e+00 -4.12076302e-02 -1.77499995e-01 6.68680608e-01
1.28890836e+00 2.23408446e-01 -7.85686731e-01 -1.85016971e-02
3.96961868e-01 -2.82699049e-01 2.12797552e-01 -1.33674395e+00
-7.78308392e-01 1.03596258e+00 7.76463628e-01 3.15902352e-01
8.28219831e-01 -3.44801158e-01 6.15399480e-01 -1.98074192e-01
-1.57286674e-01 -1.66343296e+00 -8.96890163e-01 1.09578729e+00
5.49534380e-01 -9.00355518e-01 -3.56571078e-01 -9.23702791e-02
-5.81348777e-01 1.37491870e+00 5.22009790e-01 -1.29267171e-01
6.41263902e-01 4.11246181e-01 5.12057900e-01 -8.33224729e-02
-5.78591466e-01 -3.22750211e-01 4.44705099e-01 7.47679889e-01
1.06033385e+00 6.31418377e-02 -5.61081350e-01 5.48822761e-01
-5.23943782e-01 1.06120193e-02 2.67585635e-01 4.89695519e-01
-2.45253593e-01 -1.42331851e+00 -2.27012739e-01 8.38399827e-01
-6.81211412e-01 -1.13443621e-01 -9.97395456e-01 5.53330123e-01
4.87235576e-01 1.19801986e+00 4.08697069e-01 -8.41825128e-01
5.75992465e-01 4.34437722e-01 5.76550305e-01 -4.88611221e-01
-1.37870872e+00 -2.26888657e-01 6.92360818e-01 5.93828410e-02
-5.22663668e-02 -1.04229021e+00 -8.34254444e-01 -5.67698956e-01
-6.98805079e-02 9.82414559e-02 5.14514565e-01 1.04853237e+00
1.52488872e-01 6.89415708e-02 6.82279229e-01 -4.24962729e-01
-5.18417835e-01 -1.39784002e+00 -7.37372935e-01 6.17277384e-01
3.22523385e-01 -2.92518616e-01 -1.18789695e-01 -2.59138718e-02]
|
[10.235244750976562, 10.667710304260254]
|
f16697ff-751a-4c87-8b63-c8f20bcbe317
|
active-learning-in-physics-from-101-to
|
2307.03899
| null |
https://arxiv.org/abs/2307.03899v1
|
https://arxiv.org/pdf/2307.03899v1.pdf
|
Active Learning in Physics: From 101, to Progress, and Perspective
|
Active Learning (AL) is a family of machine learning (ML) algorithms that predates the current era of artificial intelligence. Unlike traditional approaches that require labeled samples for training, AL iteratively selects unlabeled samples to be annotated by an expert. This protocol aims to prioritize the most informative samples, leading to improved model performance compared to training with all labeled samples. In recent years, AL has gained increasing attention, particularly in the field of physics. This paper presents a comprehensive and accessible introduction to the theory of AL reviewing the latest advancements across various domains. Additionally, we explore the potential integration of AL with quantum ML, envisioning a synergistic fusion of these two fields rather than viewing AL as a mere extension of classical ML into the quantum realm.
|
['Xi Chen', 'Yolanda Vives-Gilabert', 'José D. Martín-Guerrero', 'Yongcheng Ding']
|
2023-07-08
| null | null | null | null |
['active-learning', 'active-learning']
|
['methodology', 'natural-language-processing']
|
[ 6.38883710e-01 5.23245335e-01 -7.13979006e-01 -4.75834161e-01
-1.03800499e+00 -4.76362318e-01 7.29264259e-01 4.83646125e-01
-5.89942098e-01 9.64432359e-01 -9.25357938e-02 -2.63159901e-01
-2.98926346e-02 -9.70173538e-01 -3.82438689e-01 -8.91491473e-01
-5.04052937e-02 6.18068695e-01 5.08068800e-02 5.48708029e-02
4.04106438e-01 6.42481744e-01 -1.39243782e+00 5.66128977e-02
1.10333717e+00 8.11010897e-01 -1.91482767e-01 4.25640911e-01
-3.55301052e-01 8.56703877e-01 -3.79690439e-01 -4.40798193e-01
1.88878626e-01 -6.29489601e-01 -8.99610698e-01 -3.84867080e-02
1.77796856e-01 5.10648079e-02 -2.44207337e-01 9.85122859e-01
3.72198761e-01 3.86998087e-01 4.56443518e-01 -1.02898359e+00
-5.48418164e-01 6.14934742e-01 2.02938239e-03 7.78332874e-02
4.44331259e-01 3.13810498e-01 1.25621033e+00 -9.81071770e-01
6.67853415e-01 1.00916040e+00 2.82722831e-01 6.08330369e-01
-1.39343393e+00 -4.68725681e-01 1.04272924e-02 7.41384268e-01
-1.16561365e+00 -5.63031495e-01 8.36984336e-01 -3.63344640e-01
7.68651903e-01 6.31315485e-02 8.86817932e-01 9.15664494e-01
2.25453041e-02 1.10774302e+00 1.27291727e+00 -1.18587935e+00
8.98824692e-01 2.93106079e-01 3.76858681e-01 5.39021492e-01
2.47482017e-01 4.42424208e-01 -9.91900444e-01 -2.75100738e-01
2.91953593e-01 -2.36954302e-01 1.09223507e-01 -5.85669458e-01
-1.06278074e+00 9.54226494e-01 2.13377133e-01 5.77057265e-02
-3.54016483e-01 -1.22195549e-01 7.14604333e-02 2.68047243e-01
4.00188446e-01 7.94402778e-01 -3.57819200e-01 3.70237380e-02
-9.25053716e-01 1.01793848e-01 7.00083733e-01 6.35005236e-01
1.04805112e+00 -8.84490460e-02 -3.53892073e-02 3.35793078e-01
7.48466969e-01 3.77318084e-01 -5.17301299e-02 -1.09689569e+00
4.24809158e-02 8.25000226e-01 2.09595308e-01 -3.85929346e-02
-1.95248708e-01 -4.79280740e-01 -3.12890738e-01 3.12688172e-01
2.53969818e-01 1.14962440e-02 -6.71499550e-01 1.41902447e+00
4.37113285e-01 9.99210775e-02 2.53269762e-01 5.33550084e-01
8.21493149e-01 5.27871370e-01 2.65193164e-01 -7.21364439e-01
5.94858110e-01 -8.42387736e-01 -7.63014734e-01 -2.70675361e-01
4.93192941e-01 -5.29338658e-01 7.88664222e-01 7.63480663e-01
-9.64755476e-01 -5.01376688e-01 -1.24054468e+00 1.80381932e-04
-4.30999666e-01 -2.97762245e-01 9.85739350e-01 8.87344241e-01
-8.64482641e-01 7.00318933e-01 -9.98051167e-01 -3.09891522e-01
5.53243458e-01 6.54260337e-01 -1.02585904e-01 2.13759802e-02
-1.20857167e+00 1.04650819e+00 6.77125514e-01 8.71230755e-03
-6.15383685e-01 -4.65750277e-01 -7.58260012e-01 -3.82884711e-01
5.29547811e-01 -4.12372649e-01 1.37164557e+00 -8.65071535e-01
-1.78115952e+00 7.39281297e-01 -5.28430939e-01 -6.14196479e-01
3.21303368e-01 -3.37029397e-01 -5.39576292e-01 2.82676548e-01
-1.22315519e-01 7.03116834e-01 4.59560454e-01 -1.17773068e+00
-4.53325152e-01 -2.32471168e-01 1.85481936e-01 2.89425164e-01
-1.54353544e-01 -1.78637892e-01 3.85003388e-02 6.41791821e-02
3.95561039e-01 -8.47680926e-01 -4.17588055e-01 -2.47366086e-01
-2.63241112e-01 -6.07535422e-01 6.72966182e-01 1.59636304e-01
1.12193215e+00 -2.02241874e+00 1.52795345e-01 1.39844611e-01
3.95426780e-01 1.81644335e-01 3.10148954e-01 8.47092867e-01
5.92363104e-02 -1.35768935e-01 -2.95978278e-01 -1.90868765e-01
1.18409604e-01 1.16656415e-01 -2.82904506e-01 6.56078577e-01
2.18059465e-01 1.09755969e+00 -1.31473255e+00 -5.93397737e-01
4.30957228e-01 -5.63269556e-02 -3.41045052e-01 1.41007960e-01
-5.47814310e-01 1.06358850e+00 -4.29447532e-01 8.31074238e-01
4.08590138e-01 -4.80532289e-01 3.49473804e-01 3.79267000e-02
-2.66631573e-01 5.27040482e-01 -9.88461912e-01 1.62359512e+00
-6.56271772e-03 5.69546163e-01 -2.72623330e-01 -1.29515851e+00
9.02659714e-01 4.08664584e-01 7.37702966e-01 -7.93266237e-01
8.61672834e-02 4.35087383e-01 2.51342684e-01 -5.11627793e-01
3.69272023e-01 -2.96048850e-01 1.03291146e-01 6.18426740e-01
1.97857365e-01 -1.44996792e-01 2.94734687e-01 2.60239214e-01
8.67132187e-01 4.34744567e-01 7.95409977e-01 -5.84691800e-02
6.56754851e-01 1.87836781e-01 4.34377939e-01 1.00492108e+00
-6.50076151e-01 -2.86235400e-02 -9.05703232e-02 -3.50245416e-01
-7.72718906e-01 -1.30161214e+00 -5.42113245e-01 9.38432217e-01
3.35908502e-01 -5.11791766e-01 -3.98966283e-01 -7.57355094e-01
-3.57453585e-01 9.58680630e-01 -3.28343719e-01 -2.31771037e-01
-4.98154759e-01 -1.06075883e+00 9.47755799e-02 5.23941666e-02
6.01765871e-01 -1.21581864e+00 -5.57967842e-01 3.58806811e-02
1.15572177e-01 -8.81128967e-01 5.70417702e-01 5.82874238e-01
-1.02973080e+00 -1.11649418e+00 1.16204023e-02 -5.25317252e-01
5.64419389e-01 -2.40769863e-01 1.01798284e+00 -2.84570694e-01
-5.32879718e-02 4.86929566e-01 -6.36991858e-01 -6.16408467e-01
-7.28067815e-01 9.14948508e-02 4.34163548e-02 -1.16730042e-01
7.14123428e-01 -1.87797040e-01 -4.93442446e-01 -3.52801651e-01
-6.24900401e-01 3.89366969e-02 6.94877803e-01 6.65234387e-01
4.76686716e-01 -1.19961470e-01 6.31875217e-01 -1.22046185e+00
1.36879817e-01 -4.96187747e-01 -3.35350811e-01 5.08657038e-01
-8.94701123e-01 -5.78175560e-02 4.72113550e-01 -1.85213402e-01
-1.19898605e+00 1.58654347e-01 -2.11535126e-01 1.49361223e-01
-3.54070216e-01 6.29720688e-01 -1.17311731e-01 -2.27485627e-01
8.77573013e-01 4.35414836e-02 -3.04342300e-01 -1.40360460e-01
4.54314768e-01 8.16466153e-01 -2.34852638e-03 -5.56320727e-01
7.83443272e-01 3.47647697e-01 2.89192021e-01 -7.63576031e-01
-1.23437095e+00 -3.77601773e-01 -1.10891747e+00 -3.98951739e-01
5.89816868e-01 -4.92181093e-01 -4.93435770e-01 1.46392226e-01
-7.02602923e-01 -2.95874238e-01 -8.32234085e-01 9.44846392e-01
-4.71380830e-01 3.76444608e-01 -4.05143291e-01 -1.22511435e+00
-1.68836549e-01 -1.22206700e+00 6.68124259e-01 5.04409254e-01
-4.07001346e-01 -1.20314848e+00 3.58504832e-01 5.92139661e-01
6.48889840e-02 -8.33649654e-04 9.34578657e-01 -1.06563151e+00
-7.88178623e-01 -3.65228415e-01 3.61003369e-01 1.92065343e-01
1.73635632e-02 -3.00418828e-02 -1.43952477e+00 -2.97654688e-01
9.19832215e-02 -7.13377535e-01 6.63661599e-01 9.95444655e-02
7.70132303e-01 1.06770098e-01 -3.61627996e-01 1.93171158e-01
1.34729207e+00 6.17716670e-01 4.35002238e-01 2.81529993e-01
2.72790998e-01 4.69079703e-01 6.99256897e-01 1.05107158e-01
1.19126357e-01 3.32570881e-01 2.30861664e-01 -7.95223713e-02
1.18827961e-01 -1.76732279e-02 3.25248867e-01 9.74741817e-01
4.34091948e-02 3.03391144e-02 -1.01583457e+00 -8.60736445e-02
-1.72406495e+00 -1.09880209e+00 -1.50733903e-01 2.46800661e+00
9.27289248e-01 3.85030538e-01 -1.11672260e-01 3.55019480e-01
5.45339644e-01 -8.93643722e-02 -8.59889746e-01 -2.48217717e-01
-1.46795124e-01 7.83168256e-01 1.80868387e-01 5.75302303e-01
-1.27576375e+00 1.02197957e+00 7.84191227e+00 8.42161000e-01
-1.10057449e+00 2.52094686e-01 5.44525564e-01 8.64618942e-02
-2.44688675e-01 5.72200775e-01 -7.55187094e-01 1.88958839e-01
1.15069318e+00 -1.46480843e-01 2.17322543e-01 7.02606022e-01
2.25109503e-01 -4.46996361e-01 -1.30246806e+00 5.89741111e-01
-5.87737523e-02 -1.35420847e+00 -4.25863639e-02 7.37178698e-02
9.45173144e-01 2.31998548e-01 -1.52160004e-01 5.07914543e-01
2.55202144e-01 -1.00625086e+00 5.75239599e-01 6.50035679e-01
5.14258802e-01 -3.21312994e-01 6.31659627e-01 5.53215265e-01
-6.56267881e-01 -1.57681927e-01 -9.19008031e-02 -4.04117495e-01
1.48097649e-01 5.72714210e-01 -9.26528513e-01 7.10276008e-01
2.36350968e-01 7.88659871e-01 -5.86402595e-01 1.27006185e+00
-3.20399821e-01 9.70094502e-01 -1.39765665e-01 -9.33054090e-02
7.97590911e-02 -6.39042139e-01 5.19227266e-01 8.60795259e-01
-1.56683862e-01 1.25065535e-01 4.98139709e-01 9.41865325e-01
1.06227130e-01 1.86444506e-01 -4.42548007e-01 -4.25104052e-01
8.26735735e-01 1.14973783e+00 -9.45531726e-01 -5.59837103e-01
-5.86447060e-01 6.19195819e-01 2.46996790e-01 2.83193856e-01
-5.56886137e-01 -1.30262747e-02 -2.88769364e-01 -1.85105950e-01
-1.90260604e-01 -2.54404336e-01 -4.65484023e-01 -1.05907500e+00
-5.12792587e-01 -5.39041460e-01 2.49530688e-01 -3.48517567e-01
-1.30169368e+00 2.10789125e-03 8.55103955e-02 -1.19675696e+00
-9.11237821e-02 -5.27518570e-01 -4.62084353e-01 7.18967378e-01
-1.31895471e+00 -8.19991350e-01 -1.13700703e-01 1.99083149e-01
3.54035199e-01 -1.80752471e-01 1.12635541e+00 1.69441029e-01
-5.76909363e-01 2.22877324e-01 3.68668556e-01 5.40099293e-03
5.81331372e-01 -1.44594038e+00 1.82863191e-01 7.35097885e-01
4.01458412e-01 9.44105864e-01 7.41818070e-01 -5.78828514e-01
-1.37922847e+00 -7.41735578e-01 8.79205644e-01 -5.17007768e-01
5.66308498e-01 -4.08273250e-01 -5.95780730e-01 7.44534492e-01
1.95231751e-01 -7.57474676e-02 1.24378729e+00 2.88336575e-01
-1.09637402e-01 -2.95830667e-02 -1.14743221e+00 4.38829243e-01
7.17910409e-01 -7.47546256e-01 -5.68787158e-01 4.36541051e-01
3.83825004e-01 -1.08261131e-01 -8.05315197e-01 6.63202345e-01
3.56373638e-01 -1.07542372e+00 7.98268914e-01 -4.58281875e-01
-1.89278439e-01 -1.92217007e-01 -7.72997662e-02 -8.88006568e-01
-2.49833122e-01 -6.94049954e-01 -5.64989090e-01 8.40197980e-01
4.04410601e-01 -6.34837925e-01 1.18533087e+00 5.89051664e-01
-3.06939840e-01 -8.79559457e-01 -8.75548720e-01 -4.39261675e-01
2.93301463e-01 -5.93548536e-01 2.12478563e-02 9.06591773e-01
3.55760008e-01 5.39233387e-01 1.49659947e-01 -2.13937476e-01
7.48718977e-01 1.02779984e-01 4.30856615e-01 -1.50859404e+00
-3.10732335e-01 -3.35360557e-01 -4.01398242e-01 -9.82653618e-01
1.29144266e-01 -1.09854960e+00 8.59138966e-02 -1.57382345e+00
2.64523268e-01 -6.22481406e-01 -3.95437241e-01 2.99060911e-01
-1.29396513e-01 5.18322408e-01 -5.73424399e-02 4.61654305e-01
-9.36070144e-01 4.59570408e-01 1.06752133e+00 -9.74169374e-03
-3.29793125e-01 4.14304696e-02 -4.76994127e-01 6.01743162e-01
8.27783585e-01 -2.25692466e-01 -5.70942461e-01 2.68298030e-01
5.28352380e-01 -1.68473274e-01 2.45938972e-01 -1.04618609e+00
2.69472599e-01 -4.23666388e-01 5.56980193e-01 -5.67535281e-01
5.11195600e-01 -5.00797927e-01 -1.13845244e-02 4.45489705e-01
-4.90656346e-01 -5.96642971e-01 -2.46669471e-01 5.05225182e-01
-1.19520545e-01 -5.96080542e-01 9.54061091e-01 -4.66888070e-01
-9.75957930e-01 1.75586045e-01 -3.92127037e-01 -5.21844998e-02
1.28113902e+00 -3.35223705e-01 -4.07713413e-01 8.18741769e-02
-1.14289677e+00 -5.90406358e-02 4.43874210e-01 -2.42718309e-01
2.75435448e-01 -8.10997009e-01 -2.56103307e-01 3.99604179e-02
1.81000695e-01 -9.30652022e-03 1.57235339e-02 1.15179300e+00
-4.78466809e-01 6.67194009e-01 1.88528627e-01 -6.28270090e-01
-8.97976279e-01 4.53162193e-01 2.53862560e-01 -3.40681784e-02
-6.28071904e-01 7.32289374e-01 -2.28269801e-01 -6.75145566e-01
2.72377193e-01 3.69415551e-01 -3.28352243e-01 3.52891423e-02
1.99345395e-01 5.56574523e-01 1.72213599e-01 -3.66139263e-01
-3.06530774e-01 3.34578156e-02 -1.07477568e-02 -2.84114391e-01
1.11053443e+00 -1.57106966e-01 -1.60236314e-01 1.20281029e+00
7.47606099e-01 1.53059483e-01 -9.74594116e-01 -4.86031502e-01
3.99927586e-01 3.67220975e-02 1.15371905e-01 -9.74849641e-01
-3.39863688e-01 8.43567193e-01 6.92225754e-01 3.78092498e-01
7.58903444e-01 2.23893151e-01 4.43771154e-01 6.68414176e-01
6.34786487e-01 -1.23388088e+00 2.59994328e-01 3.44263673e-01
4.80838977e-02 -1.49117315e+00 4.43357378e-01 -3.55054229e-01
-4.09527332e-01 1.28598917e+00 4.40106511e-01 -1.64539739e-02
7.14432180e-01 -3.44441459e-02 -6.62327465e-03 -2.81933665e-01
-7.34912932e-01 -2.62801826e-01 3.41143101e-01 3.78904015e-01
7.99689651e-01 -9.02089942e-03 -4.25195426e-01 -1.09028053e-02
-5.56712709e-02 7.25092143e-02 3.91924083e-01 1.34124422e+00
-8.14897239e-01 -1.67034483e+00 -2.72626847e-01 5.29130638e-01
-1.13065653e-01 -5.26503809e-02 -5.65426767e-01 6.57553315e-01
6.17555797e-01 1.10412049e+00 4.28304402e-03 -1.98954403e-01
-2.92424262e-01 6.09282136e-01 9.18142080e-01 -1.01590168e+00
-3.52251470e-01 -7.02493116e-02 -8.10288489e-02 -2.36896157e-01
-9.94418561e-01 -9.40260828e-01 -1.44008195e+00 -1.28101930e-01
-5.30211210e-01 6.61430538e-01 4.30256039e-01 1.41688085e+00
6.84412643e-02 1.38681993e-01 6.60178006e-01 -6.38921678e-01
-5.23659945e-01 -8.79957795e-01 -4.34297800e-01 1.10533215e-01
2.34568074e-01 -7.68893600e-01 -3.44467968e-01 -6.29464537e-02]
|
[9.533804893493652, 4.181944370269775]
|
30b57193-6ede-43eb-8bd7-de7d00d7bec6
|
unsupervised-moving-object-detection-via
|
1901.03360
| null |
http://arxiv.org/abs/1901.03360v2
|
http://arxiv.org/pdf/1901.03360v2.pdf
|
Unsupervised Moving Object Detection via Contextual Information Separation
|
We propose an adversarial contextual model for detecting moving objects in
images. A deep neural network is trained to predict the optical flow in a
region using information from everywhere else but that region (context), while
another network attempts to make such context as uninformative as possible. The
result is a model where hypotheses naturally compete with no need for explicit
regularization or hyper-parameter tuning. Although our method requires no
supervision whatsoever, it outperforms several methods that are pre-trained on
large annotated datasets. Our model can be thought of as a generalization of
classical variational generative region-based segmentation, but in a way that
avoids explicit regularization or solution of partial differential equations at
run-time.
|
['Stefano Soatto', 'Yanchao Yang', 'Davide Scaramuzza', 'Antonio Loquercio']
|
2019-01-10
|
unsupervised-moving-object-detection-via-1
|
http://openaccess.thecvf.com/content_CVPR_2019/html/Yang_Unsupervised_Moving_Object_Detection_via_Contextual_Information_Separation_CVPR_2019_paper.html
|
http://openaccess.thecvf.com/content_CVPR_2019/papers/Yang_Unsupervised_Moving_Object_Detection_via_Contextual_Information_Separation_CVPR_2019_paper.pdf
|
cvpr-2019-6
|
['moving-object-detection', 'unsupervised-video-object-segmentation']
|
['computer-vision', 'computer-vision']
|
[ 2.48322964e-01 5.02404928e-01 7.57059967e-03 -3.28336030e-01
-4.60193098e-01 -7.52607703e-01 6.28740907e-01 -2.82978386e-01
-6.74778581e-01 6.64379776e-01 -1.16336204e-01 -3.36434275e-01
2.76391178e-01 -9.12707210e-01 -9.09181356e-01 -9.21928227e-01
2.62454227e-02 4.99855310e-01 6.30340815e-01 -7.88189396e-02
2.76506275e-01 6.27008379e-01 -9.92572725e-01 -5.69719821e-02
6.18769586e-01 7.12512851e-01 1.05048075e-01 1.02114344e+00
1.81229673e-02 1.21609414e+00 -6.50948644e-01 -2.67779052e-01
4.96667504e-01 -7.46907890e-01 -1.14895523e+00 4.90275472e-01
7.01953053e-01 -2.62781650e-01 -5.01459479e-01 1.07831323e+00
6.48037270e-02 5.01228750e-01 7.92026162e-01 -1.01666510e+00
-7.53011048e-01 1.54871762e-01 -4.08612609e-01 4.43104833e-01
-1.87047254e-02 4.07467067e-01 7.29761899e-01 -6.01953328e-01
9.73436296e-01 1.09806216e+00 5.39064348e-01 7.44505167e-01
-1.58765566e+00 -3.73231503e-03 4.36172336e-01 -9.93593186e-02
-1.12883043e+00 -2.28921518e-01 7.77675092e-01 -5.84867954e-01
7.47550845e-01 1.38925523e-01 7.62501836e-01 1.15051150e+00
3.14333975e-01 7.74477720e-01 7.39586830e-01 -3.26285869e-01
4.83163953e-01 8.26715678e-02 -2.64511049e-01 8.80575716e-01
4.04608659e-02 7.78996050e-02 1.02662146e-01 -1.40332609e-01
1.24711645e+00 -1.51997386e-02 -4.09316838e-01 -7.06935108e-01
-1.16402841e+00 1.03009248e+00 6.86248541e-01 1.76855624e-01
-4.11703855e-01 5.09468257e-01 2.20945552e-01 1.23477370e-01
4.36153769e-01 5.82827985e-01 -3.92075241e-01 3.21554095e-01
-1.23150480e+00 3.66063356e-01 6.87098622e-01 7.73310483e-01
9.96424079e-01 3.50997657e-01 -7.64730945e-02 2.65418947e-01
3.39997947e-01 2.39831850e-01 4.73642647e-01 -1.53756952e+00
1.71287525e-02 2.92170584e-01 3.01141411e-01 -8.33033979e-01
-1.61889032e-01 -1.60211131e-01 -6.23422742e-01 6.22907817e-01
6.89578712e-01 -4.05666202e-01 -1.32181895e+00 1.71497881e+00
3.77252370e-01 6.45395398e-01 4.74953651e-03 1.10453963e+00
3.58640611e-01 8.51540744e-01 7.51660718e-03 1.50368111e-02
7.56416798e-01 -1.07568908e+00 -3.98903728e-01 -5.35268247e-01
3.88978422e-01 -5.60944498e-01 8.46049547e-01 3.78263056e-01
-1.11022007e+00 -5.08679271e-01 -8.72488856e-01 -1.69386387e-01
-3.08485627e-01 -4.92913783e-01 5.34482419e-01 5.91615081e-01
-1.34917903e+00 7.60216594e-01 -1.04676259e+00 -8.94896612e-02
4.42521840e-01 2.70666361e-01 -2.81469434e-01 2.49098912e-01
-8.16177905e-01 9.76121962e-01 3.46172065e-01 2.59819090e-01
-1.19845188e+00 -5.48759043e-01 -9.67782080e-01 -2.52084583e-01
4.94012803e-01 -7.82762825e-01 1.14388525e+00 -1.59603179e+00
-1.51178002e+00 9.39046085e-01 -2.19378203e-01 -6.85606182e-01
8.91064525e-01 -1.82304576e-01 2.43411604e-02 3.88626069e-01
1.70033090e-02 1.16900265e+00 1.02532077e+00 -1.53241897e+00
-3.82079065e-01 3.41954194e-02 3.78636241e-01 -4.05375659e-02
5.37496030e-01 -1.81654826e-01 -6.31540120e-01 -8.38319719e-01
-3.84053625e-02 -1.15064442e+00 -8.93496037e-01 1.42827749e-01
-5.97036898e-01 -7.38940993e-03 1.01546896e+00 -5.13697684e-01
5.54471433e-01 -1.72454572e+00 3.84169102e-01 1.39250472e-01
2.05242753e-01 4.08951789e-01 -1.23974860e-01 -1.38433101e-02
1.71599463e-02 3.60256761e-01 -7.73861706e-01 -3.00345033e-01
-2.70471483e-01 5.35545290e-01 -3.75506550e-01 8.25975955e-01
5.16397059e-01 1.08980834e+00 -1.15164340e+00 -6.33020401e-01
6.04159594e-01 5.68490684e-01 -7.42240608e-01 1.68252498e-01
-6.17402136e-01 7.95174360e-01 -4.85191286e-01 1.63336635e-01
5.64930737e-01 -3.45665663e-01 2.78910212e-02 3.53986382e-01
-2.43426692e-02 1.39529929e-02 -1.09571743e+00 1.75300503e+00
-3.53912860e-01 1.00782549e+00 1.95030436e-01 -1.50360918e+00
5.48575640e-01 2.39519730e-01 5.72352469e-01 -2.30716020e-01
2.15924174e-01 -1.45570815e-01 -9.10395663e-03 -5.12301207e-01
3.86038303e-01 -2.32045680e-01 2.54738867e-01 2.05121487e-01
-1.69501919e-02 -3.42348099e-01 1.39244467e-01 3.52359205e-01
9.95168865e-01 5.40502369e-01 9.83736217e-02 -2.46622920e-01
4.49749857e-01 2.67676055e-01 8.87429655e-01 9.86828268e-01
-4.44583625e-01 8.34472835e-01 6.09096169e-01 -6.52902186e-01
-1.17595756e+00 -1.00446463e+00 -8.53358880e-02 6.94927037e-01
4.52013373e-01 -8.42937175e-03 -7.84339070e-01 -9.57076490e-01
-1.79036587e-01 6.34061635e-01 -1.03766358e+00 1.78561389e-01
-8.04153800e-01 -3.77775043e-01 2.66191155e-01 5.29277861e-01
4.57199007e-01 -1.42401278e+00 -8.63424838e-01 2.37254575e-01
-3.79034765e-02 -1.17621505e+00 -5.46226323e-01 1.53757736e-01
-9.51247931e-01 -1.23336005e+00 -7.34695911e-01 -6.88210785e-01
7.65401602e-01 -3.59530939e-04 1.45229089e+00 2.06273556e-01
-3.48415345e-01 4.40819621e-01 -1.02866597e-01 -2.99727827e-01
-5.31602502e-01 -3.03589463e-01 -3.69079322e-01 5.37974164e-02
9.52739716e-02 -3.40401977e-01 -8.74569356e-01 2.88708955e-01
-1.04009879e+00 -1.35854244e-01 3.43284935e-01 7.55922139e-01
6.85999393e-01 -2.31717825e-01 1.95175916e-01 -1.23032343e+00
1.09625958e-01 -3.46416354e-01 -8.64275694e-01 1.43259078e-01
-3.28626275e-01 1.97755381e-01 6.33823633e-01 -5.19175768e-01
-1.03261304e+00 3.90972525e-01 -1.82355121e-02 -8.48324835e-01
-4.42500681e-01 2.30254438e-02 1.90508455e-01 -1.81613892e-01
7.52837241e-01 1.30319163e-01 -1.54171392e-01 -1.68487981e-01
5.86192966e-01 -1.14809193e-01 7.56993055e-01 -3.63015085e-01
9.07624841e-01 9.25678194e-01 2.78837860e-01 -8.37322354e-01
-7.96551526e-01 -3.80195320e-01 -9.68517959e-01 -2.92088121e-01
1.38455510e+00 -6.18620157e-01 -2.95813143e-01 9.41202566e-02
-1.27367175e+00 -7.80437112e-01 -5.77778101e-01 4.47109312e-01
-7.06501186e-01 4.38618720e-01 -6.10831439e-01 -7.98753977e-01
2.25371018e-01 -1.23104894e+00 8.07641864e-01 2.48997286e-01
-1.11939780e-01 -1.45068491e+00 -5.45480773e-02 1.00589707e-01
2.64640421e-01 7.48045504e-01 2.82752484e-01 -4.46782827e-01
-1.00878572e+00 -9.59070027e-02 -7.41783306e-02 4.97957349e-01
-1.29484847e-01 3.36744040e-01 -9.65901315e-01 -1.31549165e-01
7.81076550e-02 -2.94954747e-01 9.78024244e-01 8.84640336e-01
1.22231483e+00 -4.23014581e-01 -2.95172900e-01 7.29456484e-01
1.53165329e+00 1.83344394e-01 7.12588072e-01 1.78892314e-01
9.77750361e-01 5.57124734e-01 2.97741234e-01 -1.32771358e-01
7.17676431e-02 4.24496323e-01 7.98764646e-01 -2.72146881e-01
6.52715610e-03 -8.89244005e-02 1.18941188e-01 -1.07135467e-01
-2.29258806e-01 -3.74621332e-01 -7.09975004e-01 6.65232480e-01
-1.91702759e+00 -1.15993369e+00 -2.41243243e-01 2.14971137e+00
5.59566915e-01 2.07631260e-01 2.08049029e-01 -2.90466517e-01
6.15435183e-01 3.05692315e-01 -6.94583237e-01 -5.29382110e-01
2.04096592e-04 1.78849846e-01 7.59430349e-01 9.81722414e-01
-1.38887036e+00 1.25498807e+00 7.25383234e+00 3.41914177e-01
-1.14220965e+00 2.19706610e-01 8.44470263e-01 -3.87508273e-02
-2.55160928e-01 2.48231605e-01 -2.84664899e-01 2.63596296e-01
6.41133845e-01 3.02660972e-01 3.61782551e-01 8.03786039e-01
2.77589470e-01 -3.03377241e-01 -1.17018616e+00 5.18235028e-01
2.30685761e-03 -1.55964124e+00 -1.58385590e-01 2.04501316e-01
8.70060205e-01 1.85663536e-01 -1.08979791e-01 1.25728818e-02
6.41593099e-01 -9.50360656e-01 8.87199759e-01 6.61557317e-01
1.05067536e-01 -5.80308855e-01 4.03039008e-01 5.47477543e-01
-7.47893810e-01 2.34545037e-01 -4.46033776e-01 9.79588479e-02
3.13814938e-01 3.76428127e-01 -5.81096292e-01 1.27258077e-01
4.45940405e-01 4.91359025e-01 -3.82270485e-01 1.03442228e+00
-4.75244194e-01 7.00889051e-01 -1.90116003e-01 2.61599422e-01
6.87180579e-01 -5.25997579e-01 8.30359399e-01 1.40231943e+00
-6.08245395e-02 5.56323752e-02 4.80943263e-01 1.03368688e+00
1.53707461e-02 -1.26859590e-01 -8.36647272e-01 4.68726188e-01
8.79673637e-04 1.06036913e+00 -1.05503678e+00 -5.08413911e-01
-3.66721123e-01 1.37579000e+00 2.93228418e-01 7.99389720e-01
-8.95626068e-01 -6.06820919e-02 4.93395686e-01 1.29282594e-01
5.25559247e-01 -2.58271456e-01 -5.17956689e-02 -1.31370890e+00
-2.72875071e-01 -1.09828070e-01 2.53297776e-01 -8.19874823e-01
-1.01130462e+00 4.00882632e-01 -4.46792990e-02 -9.98142183e-01
-6.27053797e-01 -7.32618630e-01 -9.24649000e-01 8.65152955e-01
-1.37414098e+00 -9.03112173e-01 9.97782797e-02 6.03266895e-01
6.26233697e-01 2.62995988e-01 4.68192071e-01 -1.68565184e-01
-4.71964121e-01 9.69687924e-02 -1.63402125e-01 4.53702748e-01
4.42293197e-01 -1.66319025e+00 2.90937841e-01 1.34318900e+00
4.89619911e-01 5.76863050e-01 9.71321464e-01 -4.76147205e-01
-8.05210710e-01 -1.34306586e+00 7.15418100e-01 -6.89198375e-01
4.63500202e-01 -3.24229211e-01 -1.10487366e+00 1.06899524e+00
5.75645566e-01 6.95124805e-01 1.57113820e-01 -2.18265280e-01
-1.32278383e-01 2.62152791e-01 -1.18829262e+00 6.74815118e-01
8.20818543e-01 -3.23753476e-01 -5.68064868e-01 5.62158585e-01
7.43030190e-01 -5.63980758e-01 -4.64178771e-01 -9.54770390e-03
5.37802204e-02 -1.16636622e+00 1.29880583e+00 -8.46317291e-01
4.51460838e-01 -3.52623850e-01 1.08631231e-01 -1.21235859e+00
-2.83830106e-01 -9.16012585e-01 -2.36524299e-01 7.19665110e-01
3.89473230e-01 -4.67228085e-01 1.04521418e+00 7.35369623e-01
-1.54645637e-01 -6.27943397e-01 -8.21481466e-01 -7.59733021e-01
3.43811095e-01 -5.00733793e-01 -2.54635196e-02 1.02358246e+00
-6.25642300e-01 4.57732007e-02 -3.70986551e-01 3.41889888e-01
7.03963578e-01 8.02894980e-02 7.13551998e-01 -9.93780434e-01
-2.93182343e-01 -5.35778463e-01 -6.25335693e-01 -1.09966946e+00
3.07469010e-01 -6.49325848e-01 2.33521909e-01 -1.51016307e+00
-2.28347018e-01 -2.64329284e-01 -1.57081991e-01 4.55589980e-01
-2.02463880e-01 4.78277236e-01 1.61553383e-01 1.67011872e-01
-5.24926364e-01 2.20021397e-01 1.54386163e+00 -1.72302961e-01
-3.98234427e-01 1.36033937e-01 -4.26719964e-01 1.06749809e+00
7.42286265e-01 -5.06087601e-01 -5.63667834e-01 -3.06843817e-01
-6.03168830e-02 8.64431262e-02 9.24896657e-01 -7.97422349e-01
1.36091441e-01 -4.94459361e-01 6.21514022e-01 -3.27909291e-01
3.12808067e-01 -6.68682277e-01 -1.40372425e-01 3.82706553e-01
-3.75105590e-01 -4.58853953e-02 -2.34807730e-02 8.67761672e-01
-1.12498857e-01 -4.75212693e-01 1.11734760e+00 -5.54556012e-01
-8.80349398e-01 4.12671059e-01 -8.03462803e-01 3.01613718e-01
1.16116345e+00 -3.36123347e-01 -1.75437517e-02 -5.35526872e-01
-1.10350573e+00 3.16859484e-02 7.45886743e-01 1.70046285e-01
3.86441499e-01 -9.25830185e-01 -4.77103323e-01 -5.72500974e-02
-3.71493429e-01 5.01506567e-01 6.18519485e-02 5.60265064e-01
-9.33342576e-01 2.25344196e-01 4.00885902e-02 -6.97332859e-01
-7.23663807e-01 9.64407623e-01 8.55477393e-01 -1.78516120e-01
-1.10491550e+00 1.02030516e+00 4.98598605e-01 -2.36753121e-01
1.94945335e-02 -3.91992480e-01 1.82978511e-02 -3.86586785e-01
2.05599278e-01 4.78905905e-03 -2.85177261e-01 -7.39050388e-01
-2.22883180e-01 4.36612904e-01 1.78180471e-01 -4.16531265e-01
1.12493682e+00 -1.52892485e-01 5.71266375e-03 3.67696404e-01
1.12917197e+00 5.98439276e-02 -1.93310106e+00 -8.97333771e-02
-1.31680712e-01 -5.17424524e-01 2.75044173e-01 -5.81039846e-01
-1.39075530e+00 8.37012529e-01 2.89043069e-01 4.22315687e-01
9.75943387e-01 -3.62050720e-02 5.15911639e-01 4.63432431e-01
-6.99313655e-02 -1.05473459e+00 1.53085396e-01 4.54662502e-01
7.44041920e-01 -1.33595538e+00 -1.15116298e-01 -2.24146500e-01
-8.57111514e-01 9.48310912e-01 6.44046366e-01 -8.34947050e-01
8.17365885e-01 1.59022510e-01 2.93608963e-01 -1.37620419e-01
-5.30471385e-01 -2.42575586e-01 4.13275033e-01 7.66947389e-01
2.74499148e-01 -2.48903126e-01 9.44326520e-02 -3.02511990e-01
6.95362538e-02 -1.52529791e-01 5.81218243e-01 9.86637235e-01
-3.08387637e-01 -1.00770843e+00 -3.89416635e-01 1.51922777e-01
-6.06017172e-01 1.07124843e-01 -4.05942202e-01 1.06723237e+00
3.20066065e-01 8.64259899e-01 3.61887991e-01 7.61134624e-02
-1.11782640e-01 7.68862069e-02 6.07116163e-01 -6.16487265e-01
-5.10007441e-01 3.26504350e-01 -3.36710811e-01 -8.12814593e-01
-7.31372952e-01 -7.93431163e-01 -1.31621683e+00 -6.18595295e-02
-1.24090455e-01 1.19070984e-01 4.00186181e-01 1.07947528e+00
-1.08747175e-02 5.49463630e-01 5.21524012e-01 -1.25968289e+00
1.38834506e-01 -7.35446751e-01 -4.86761451e-01 4.70173776e-01
7.69921005e-01 -3.70850593e-01 -2.53771275e-01 4.88096058e-01]
|
[9.029215812683105, -0.34647640585899353]
|
a1b9213b-0489-418a-b62a-9b154165a377
|
reason-from-context-with-self-supervised
|
2211.12817
| null |
https://arxiv.org/abs/2211.12817v2
|
https://arxiv.org/pdf/2211.12817v2.pdf
|
Reason from Context with Self-supervised Learning
|
Self-supervised learning (SSL) learns to capture discriminative visual features useful for knowledge transfers. To better accommodate the object-centric nature of current downstream tasks such as object recognition and detection, various methods have been proposed to suppress contextual biases or disentangle objects from contexts. Nevertheless, these methods may prove inadequate in situations where object identity needs to be reasoned from associated context, such as recognizing or inferring tiny or obscured objects. As an initial effort in the SSL literature, we investigate whether and how contextual associations can be enhanced for visual reasoning within SSL regimes, by (a) proposing a new Self-supervised method with external memories for Context Reasoning (SeCo), and (b) introducing two new downstream tasks, lift-the-flap and object priming, addressing the problems of "what" and "where" in context reasoning. In both tasks, SeCo outperformed all state-of-the-art (SOTA) SSL methods by a significant margin. Our network analysis revealed that the proposed external memory in SeCo learns to store prior contextual knowledge, facilitating target identity inference in the lift-the-flap task. Moreover, we conducted psychophysics experiments and introduced a Human benchmark in Object Priming dataset (HOP). Our results demonstrate that SeCo exhibits human-like behaviors.
|
['Mengmi Zhang', 'Zenglin Shi', 'Gabriel Kreiman', 'Ankur Sikarwar', 'Xiao Liu']
|
2022-11-23
| null | null | null | null |
['visual-reasoning', 'visual-reasoning']
|
['computer-vision', 'reasoning']
|
[ 4.18133080e-01 -4.70938273e-02 -2.00657770e-01 -3.34166884e-01
6.23519626e-03 -6.19437218e-01 8.98941755e-01 2.96616286e-01
-5.79045713e-01 6.81545675e-01 9.19652954e-02 -4.24727887e-01
-1.97721839e-01 -6.64919615e-01 -8.14192533e-01 -9.18892920e-01
-8.29263553e-02 5.66837080e-02 2.75075108e-01 -2.21958935e-01
5.82284510e-01 7.86781788e-01 -1.71095943e+00 5.63129067e-01
6.28083646e-01 1.10939395e+00 4.04917836e-01 1.36860758e-01
-2.66646028e-01 7.99503028e-01 -4.36518639e-01 -2.10016772e-01
4.28465754e-01 -3.42447966e-01 -7.03162193e-01 -2.43776441e-01
7.14688957e-01 7.61101674e-03 -2.80258119e-01 1.04816425e+00
5.00257313e-01 2.56718695e-01 5.26969552e-01 -1.27778447e+00
-1.36985564e+00 5.15807331e-01 -4.21248198e-01 6.62364602e-01
3.32582980e-01 4.03468132e-01 8.43525887e-01 -1.29296744e+00
5.17121017e-01 1.38809121e+00 3.98482352e-01 6.69022918e-01
-1.63406038e+00 -7.77190685e-01 6.75289989e-01 4.35665011e-01
-1.14460242e+00 -3.39550167e-01 9.05429363e-01 -3.90647650e-01
9.45726395e-01 2.18998194e-01 7.23621368e-01 1.47662902e+00
8.15405622e-02 8.07004213e-01 1.93793702e+00 -5.00499606e-01
5.33136010e-01 2.85073727e-01 1.11299329e-01 4.30334151e-01
5.19358158e-01 5.50403237e-01 -9.66771722e-01 3.21663879e-02
7.93121457e-01 -2.37643689e-01 -1.04049839e-01 -4.00125057e-01
-1.42649651e+00 5.15883684e-01 7.98699319e-01 1.87077612e-01
-1.85844705e-01 -8.55122432e-02 1.83189481e-01 2.78341949e-01
8.32400247e-02 5.91551602e-01 -1.21906750e-01 5.61430991e-01
-5.24573326e-01 1.01243153e-01 3.72901112e-01 1.00711346e+00
7.26320863e-01 2.90058434e-01 -5.02596021e-01 4.87319142e-01
1.14093177e-01 4.61892337e-01 4.91169900e-01 -7.55064666e-01
3.44440013e-01 6.83364213e-01 5.01703657e-02 -9.71463084e-01
-3.93919766e-01 -4.86819148e-01 -7.02139854e-01 4.26734924e-01
4.48765337e-01 2.97488391e-01 -9.13141966e-01 2.12873340e+00
3.98541331e-01 1.35789588e-01 9.24529880e-02 1.25909650e+00
8.72047007e-01 2.26960286e-01 7.23855793e-01 -2.86552668e-01
1.57615423e+00 -8.55762839e-01 -5.11372507e-01 -7.56749213e-01
3.08066308e-01 -6.57557607e-01 1.31529140e+00 1.58688948e-01
-8.68246257e-01 -9.39891458e-01 -9.76134539e-01 -9.60685387e-02
-7.72595942e-01 -2.29844265e-02 8.78926456e-01 5.33370435e-01
-1.00916076e+00 3.91025335e-01 -2.14837521e-01 -3.01446974e-01
8.19642782e-01 1.58539295e-01 -4.86186385e-01 7.11045414e-02
-1.28525710e+00 1.03874433e+00 5.52527666e-01 3.59021395e-01
-1.18042135e+00 -6.01261973e-01 -8.28485250e-01 1.46789983e-01
6.56148374e-01 -6.49916768e-01 6.54556990e-01 -1.19073689e+00
-1.03757977e+00 1.23156250e+00 -1.30854368e-01 -2.61137217e-01
1.21870264e-01 -2.76112825e-01 -5.29427946e-01 3.40863347e-01
2.17144713e-01 9.96745825e-01 1.10982418e+00 -1.58460939e+00
-2.05594406e-01 -5.02191186e-01 2.62944102e-01 2.26025194e-01
-1.35895371e-01 -2.53237128e-01 1.66054174e-01 -9.05610800e-01
2.85722554e-01 -9.20804501e-01 1.04842454e-01 2.76814878e-01
-1.43104792e-01 -2.53949195e-01 8.82891238e-01 -2.64636636e-01
8.63463104e-01 -2.20143890e+00 -5.82512394e-02 3.90119515e-02
1.95145220e-01 5.45963943e-01 -3.60730171e-01 3.59185159e-01
-3.66891742e-01 -1.32112071e-01 -2.86135599e-02 -7.12339133e-02
-2.87041683e-02 2.94176430e-01 -6.31014466e-01 3.42095882e-01
5.38560033e-01 1.24191725e+00 -1.15741515e+00 -6.92107022e-01
-7.53658265e-03 2.38165915e-01 -2.95075208e-01 1.74765632e-01
-3.00478429e-01 6.54583693e-01 -1.09787494e-01 8.92721236e-01
6.03086531e-01 -2.82565713e-01 3.17501903e-01 -4.21737373e-01
-1.64668500e-01 1.86334595e-01 -1.07828021e+00 1.52044225e+00
-2.13547960e-01 7.07218230e-01 -6.23993389e-02 -1.18305683e+00
9.14727628e-01 2.81404536e-02 -1.29109502e-01 -1.11791277e+00
5.68244718e-02 -1.30282015e-01 1.78547949e-01 -6.57681644e-01
1.97816834e-01 -2.76907146e-01 1.80149511e-01 2.38139212e-01
1.24232464e-01 3.68312627e-01 -1.84313278e-03 3.05344135e-01
6.80709183e-01 4.27598178e-01 5.01547098e-01 -4.93394434e-01
5.69153547e-01 -9.66963172e-02 4.98287350e-01 9.32292581e-01
-6.19389117e-01 1.64354444e-01 4.56352890e-01 -4.22054857e-01
-6.45797014e-01 -1.43362916e+00 -7.50973597e-02 1.37345719e+00
5.09540021e-01 1.01027071e-01 -4.87507224e-01 -7.25179136e-01
2.52621412e-01 8.42211425e-01 -9.61411178e-01 -4.54286426e-01
-5.29943228e-01 -5.32659829e-01 4.33647692e-01 9.18124080e-01
6.41712070e-01 -1.51337743e+00 -7.29775548e-01 -6.85296282e-02
1.88730523e-01 -1.23993182e+00 -3.81325543e-01 5.16438663e-01
-8.74646366e-01 -1.12744832e+00 -1.16581552e-01 -8.89835298e-01
8.72968256e-01 5.45531988e-01 1.05663991e+00 5.50973788e-02
-4.32371050e-01 5.91152012e-01 -4.67046462e-02 -1.92073613e-01
9.13964510e-02 -3.97344589e-01 7.94536844e-02 2.16884091e-01
5.36039770e-01 -7.86419332e-01 -8.50633025e-01 5.41701674e-01
-1.02811623e+00 8.19880888e-02 9.14219737e-01 9.58999097e-01
5.51361978e-01 -4.93979007e-01 1.06722605e+00 -6.93824053e-01
5.51190436e-01 -4.28050995e-01 -6.09440506e-01 4.53253597e-01
-7.61556625e-01 2.56994516e-01 5.43923855e-01 -7.83835888e-01
-1.40585947e+00 -1.37516083e-02 4.83550519e-01 -3.98658484e-01
-3.66582602e-01 2.16158584e-01 -3.37481171e-01 -3.11456174e-01
8.56497526e-01 2.78866887e-01 -2.48204038e-01 -1.00092068e-01
4.95890886e-01 1.97767198e-01 6.05894744e-01 -9.96041298e-01
7.23008037e-01 7.26509809e-01 3.36107820e-01 -5.01050949e-01
-1.26665509e+00 -1.89845622e-01 -7.14322329e-01 -3.31858188e-01
9.25311625e-01 -1.01964664e+00 -1.19337463e+00 1.85271695e-01
-8.62769783e-01 -4.66011673e-01 -3.58606637e-01 3.33459288e-01
-4.25993443e-01 1.61681011e-01 -3.69604886e-01 -8.41209769e-01
-5.45310900e-02 -8.76562595e-01 7.56940663e-01 4.26295608e-01
-1.43025234e-01 -9.28257644e-01 -2.16591075e-01 4.07830983e-01
5.42607665e-01 1.23442233e-01 1.17850435e+00 -6.55812144e-01
-7.36339450e-01 3.50070387e-01 -6.23616576e-01 2.12636977e-01
1.03942500e-02 -5.91876626e-01 -1.39282405e+00 -2.21172154e-01
1.52158439e-01 -4.27544922e-01 9.87644613e-01 -9.61874202e-02
1.17310822e+00 -1.32209018e-01 -3.07362348e-01 4.18020010e-01
1.31939042e+00 2.11536199e-01 5.91922164e-01 3.05628300e-01
4.43945944e-01 7.55724072e-01 5.86793303e-01 -1.37661456e-03
2.16222852e-01 3.89939159e-01 4.89493847e-01 1.13013990e-01
-4.69622970e-01 -4.66215730e-01 1.97412819e-01 3.46156955e-01
-5.11988662e-02 1.46243215e-01 -8.59422803e-01 6.09672427e-01
-1.84422302e+00 -1.02171361e+00 4.23279554e-02 1.99068928e+00
8.72423112e-01 4.17372614e-01 -2.51758069e-01 -1.36971205e-01
7.49494135e-01 2.19627783e-01 -8.74227285e-01 -2.00120136e-01
-5.04525542e-01 1.68552667e-01 9.11089182e-02 2.21162125e-01
-9.55062866e-01 1.11072731e+00 5.87915945e+00 5.66508234e-01
-1.23035145e+00 2.00084463e-01 4.85936552e-01 -6.34215102e-02
-1.93852112e-01 3.74519616e-01 -7.98560917e-01 2.72932827e-01
4.75340694e-01 1.38398081e-01 4.65848476e-01 9.67544615e-01
-1.33106217e-01 -5.37568986e-01 -1.47587442e+00 9.55223083e-01
2.48687029e-01 -1.19041169e+00 2.07389742e-01 -3.03667784e-01
7.14658260e-01 -4.12430316e-01 2.42837965e-01 5.84980607e-01
-1.97120816e-01 -1.01170039e+00 8.21378171e-01 4.79067832e-01
6.20259166e-01 -9.43656638e-02 9.51899812e-02 1.66856319e-01
-1.16322017e+00 -3.65512848e-01 -3.42593819e-01 -3.45935941e-01
-2.32054323e-01 4.14181322e-01 -5.13273656e-01 2.69969612e-01
6.79519355e-01 3.22415084e-01 -9.26692367e-01 7.97042131e-01
-6.32833421e-01 4.35838401e-01 1.80028766e-01 -2.90639214e-02
1.30276933e-01 6.58261552e-02 5.37132919e-01 1.20726275e+00
-2.73351043e-01 3.12482327e-01 1.57826185e-01 1.18817186e+00
-2.40022559e-02 -9.90865827e-02 -5.67897856e-01 1.13961048e-01
5.99537730e-01 1.16417944e+00 -9.40572739e-01 -4.98129845e-01
-3.15548599e-01 7.40880072e-01 6.95496082e-01 6.63573682e-01
-6.99270844e-01 -1.18746981e-01 4.49117154e-01 5.72814047e-02
3.44872385e-01 -3.23860586e-01 -5.57861686e-01 -1.05081165e+00
1.44662812e-01 -6.40538156e-01 3.80803794e-01 -1.09055603e+00
-1.48715270e+00 3.97447526e-01 2.91666746e-01 -1.06264222e+00
4.31713462e-01 -1.01171720e+00 -5.09834707e-01 7.60765970e-01
-1.75850499e+00 -1.05708480e+00 -5.02476037e-01 7.40334690e-01
3.70859504e-01 -1.93865076e-01 6.40465558e-01 1.02447771e-01
-3.11368704e-01 3.96565795e-01 -4.20485675e-01 2.33243350e-02
9.57591832e-01 -9.96394575e-01 -1.91338137e-02 7.62786865e-01
1.72696367e-01 1.06975186e+00 4.97144789e-01 -7.62180269e-01
-1.53205681e+00 -8.72613311e-01 5.92253566e-01 -4.86790299e-01
7.15054214e-01 -6.38690710e-01 -1.08076668e+00 6.73094809e-01
1.11688592e-01 5.85697293e-01 6.55860364e-01 1.51557457e-02
-8.18122208e-01 -2.73979604e-01 -1.02317810e+00 8.74554813e-01
1.60514271e+00 -8.42308283e-01 -1.09889305e+00 7.51946270e-02
5.89396238e-01 -1.95872650e-01 -4.29918885e-01 3.34740549e-01
5.58414876e-01 -1.09181178e+00 1.20357442e+00 -9.16841745e-01
4.65246201e-01 -4.29191858e-01 -2.20385283e-01 -1.05520320e+00
-5.32066345e-01 -2.92802274e-01 -2.50127435e-01 1.23344147e+00
4.43185419e-02 -7.94025421e-01 5.22006929e-01 4.63721067e-01
-3.63380685e-02 -6.84984744e-01 -1.00149262e+00 -1.04219234e+00
-1.10628448e-01 -2.27546260e-01 3.84715796e-01 1.09810042e+00
-1.47012189e-01 3.90481949e-01 -6.05907887e-02 3.40218991e-01
7.03010023e-01 4.52043444e-01 2.85110980e-01 -1.05081856e+00
-4.73018661e-02 -3.74722213e-01 -3.22984278e-01 -8.55683148e-01
4.44013327e-01 -1.19064093e+00 -1.23315476e-01 -1.13304758e+00
1.93627834e-01 -3.70298862e-01 -6.64529979e-01 6.65270627e-01
-4.92429405e-01 1.91244870e-01 3.85012746e-01 1.41154572e-01
-6.58902943e-01 5.04501760e-01 1.37060153e+00 -9.84337777e-02
6.14043474e-02 -5.08150220e-01 -8.94470811e-01 6.74248576e-01
7.18737185e-01 -3.97484720e-01 -4.84288603e-01 -2.72654802e-01
3.10830623e-01 -1.63931400e-01 1.07714248e+00 -8.43891501e-01
2.76585609e-01 -3.87543857e-01 8.39503348e-01 -4.43093747e-01
2.68487036e-01 -8.65053773e-01 -2.66459137e-01 4.35445845e-01
-5.01567006e-01 -2.65913340e-03 5.28653264e-01 7.18896747e-01
-1.03448629e-01 -1.91102669e-01 7.19925106e-01 -2.01398209e-01
-1.17755008e+00 -2.16290653e-01 -2.07271755e-01 2.04731166e-01
1.13146770e+00 -2.17935920e-01 -8.60103309e-01 1.24935277e-01
-9.40361738e-01 1.11174107e-01 6.61693886e-02 4.67662811e-01
7.33315706e-01 -1.50723922e+00 -2.84426391e-01 3.08535278e-01
3.30792934e-01 -2.72089332e-01 3.69391531e-01 9.14962113e-01
8.15331638e-02 5.16096711e-01 -4.60882068e-01 -6.73846841e-01
-1.02042758e+00 1.32014763e+00 3.99557129e-02 1.36600137e-01
-3.46431702e-01 7.56862819e-01 7.24597812e-01 -7.80591071e-02
3.06687772e-01 -1.42596424e-01 -7.52804801e-02 1.73253745e-01
4.20176983e-01 1.56914279e-01 -1.26679346e-01 -3.54099303e-01
-4.40189004e-01 3.69657099e-01 1.70552973e-02 2.05124423e-01
1.09283304e+00 -1.27852008e-01 -1.08679667e-01 4.71604466e-01
6.84204042e-01 -1.17061883e-01 -1.45066512e+00 -4.91855800e-01
2.43134007e-01 -6.20915353e-01 -2.74837613e-01 -1.03306723e+00
-7.85580754e-01 9.92294729e-01 7.18923032e-01 1.01723455e-01
1.26448357e+00 8.63964036e-02 2.80080717e-02 5.45855165e-01
4.01750952e-01 -1.18054748e+00 5.37106276e-01 2.77252257e-01
1.14036000e+00 -1.54338920e+00 9.63366330e-02 -3.38105768e-01
-6.79050088e-01 9.30721223e-01 1.14970994e+00 -1.05238050e-01
4.28142548e-01 5.75709119e-02 -6.95950463e-02 -2.62095332e-01
-8.84318590e-01 -3.04929793e-01 4.70378995e-01 6.54392779e-01
1.44910291e-01 -6.54821247e-02 -1.20413937e-01 4.68752563e-01
1.68253720e-01 -3.12280506e-01 2.76986416e-02 1.16695058e+00
-3.95408422e-01 -8.10757220e-01 -7.20705271e-01 7.73896053e-02
9.99238342e-02 -3.81573230e-01 -4.24161315e-01 9.01561379e-01
4.36828554e-01 6.32118106e-01 -1.02133125e-01 4.67895493e-02
2.81327724e-01 4.03287500e-01 7.18161821e-01 -5.65531731e-01
-6.11140549e-01 -1.88411072e-01 -1.00064747e-01 -5.87001681e-01
-6.18543506e-01 -5.16866624e-01 -1.11357343e+00 2.64193356e-01
-2.15145037e-01 -3.96572530e-01 4.54206139e-01 9.54270363e-01
3.74739289e-01 5.10505617e-01 3.50196600e-01 -8.76631975e-01
-3.96033466e-01 -6.50379002e-01 -3.62700582e-01 7.77529478e-01
1.65617719e-01 -1.12940776e+00 -4.27647442e-01 2.45188717e-02]
|
[10.05467700958252, 1.7358895540237427]
|
24fef516-6be1-4b16-b0e4-f61cebcbaecc
|
inno-at-semeval-2020-task-11-leveraging-pure-1
| null | null |
https://aclanthology.org/2020.semeval-1.193
|
https://aclanthology.org/2020.semeval-1.193.pdf
|
Inno at SemEval-2020 Task 11: Leveraging Pure Transfomer for Multi-Class Propaganda Detection
|
The paper presents the solution of team {''}Inno{''} to a SEMEVAL 2020 task 11 {''}Detection of propaganda techniques in news articles{''}. The goal of the second subtask is to classify textual segments that correspond to one of the 18 given propaganda techniques in news articles dataset. We tested a pure Transformer-based model with an optimized learning scheme on the ability to distinguish propaganda techniques between each other. Our model showed 0:6 and 0:58 overall F1 score on validation set and test set accordingly and non-zero F1 score on each class on both sets.
|
['Vladimir Ivanov', 'Dmitry Grigorev']
|
2020-12-01
| null | null | null |
semeval-2020
|
['propaganda-detection']
|
['natural-language-processing']
|
[ 1.26045376e-01 1.86858207e-01 -3.83389801e-01 -1.96789905e-01
-8.15474927e-01 -7.00102985e-01 1.45798254e+00 2.56710708e-01
-5.35411298e-01 5.74553192e-01 6.08707488e-01 -7.30751634e-01
-4.25314873e-01 -5.96083224e-01 -5.52178860e-01 -3.40744436e-01
-5.60775287e-02 6.22368634e-01 3.13016862e-01 -6.37098312e-01
1.00839162e+00 -7.19658658e-02 -1.11518550e+00 1.01964045e+00
7.99904704e-01 9.04421270e-01 -1.41900808e-01 9.53654766e-01
-4.91746748e-03 2.00187659e+00 -1.38078773e+00 -4.96322513e-01
2.01553538e-01 -5.71006954e-01 -1.25099218e+00 -2.68562436e-01
9.69250560e-01 3.34389396e-02 -3.62221420e-01 1.13757825e+00
2.03494832e-01 -6.92474991e-02 1.07900155e+00 -8.57561767e-01
-5.21249056e-01 1.24837673e+00 -5.40123105e-01 9.00937855e-01
6.65387511e-01 -2.25540221e-01 7.74930000e-01 -5.60545325e-01
1.06709063e+00 1.14213848e+00 7.74866343e-01 3.98547590e-01
-1.06998432e+00 -6.53477728e-01 -2.30962947e-01 3.94218147e-01
-7.97632873e-01 -2.88426697e-01 6.75581157e-01 -1.15484989e+00
9.90527749e-01 3.82917285e-01 5.86055756e-01 1.69177449e+00
8.25051427e-01 8.61028671e-01 1.55707645e+00 -3.48163217e-01
5.33543294e-03 3.30365926e-01 5.81539512e-01 8.37673008e-01
-8.01089033e-02 1.28382206e-01 -7.28488445e-01 -1.75671071e-01
2.44015366e-01 -6.43659294e-01 1.44683316e-01 6.59366488e-01
-1.21897721e+00 1.21210754e+00 2.21372321e-01 4.74899769e-01
-2.98375398e-01 -8.66452828e-02 9.01434243e-01 6.67675018e-01
8.39527726e-01 7.22310305e-01 -1.36047974e-01 -2.74396956e-01
-1.11026716e+00 7.56507576e-01 8.09016883e-01 6.05554998e-01
-6.87628239e-02 2.85588980e-01 -5.67781270e-01 6.08548641e-01
-2.38808036e-01 7.25530267e-01 3.10505986e-01 -5.87254047e-01
7.91700065e-01 3.66079599e-01 1.63273960e-02 -1.26755929e+00
-6.33295953e-01 -8.32235694e-01 -5.00294626e-01 -1.06504284e-01
4.98047799e-01 -2.99470931e-01 -1.11028814e+00 1.36354983e+00
-2.45634336e-02 -4.30588692e-01 -1.28439531e-01 5.23316026e-01
1.14711452e+00 8.27723861e-01 -5.35164811e-02 -4.28091824e-01
1.38083827e+00 -1.01922452e+00 -8.30870926e-01 -2.63368815e-01
9.54994202e-01 -1.08167827e+00 7.09495604e-01 6.92558229e-01
-6.50239766e-01 -3.34639877e-01 -8.58506441e-01 2.27873966e-01
-2.04066988e-02 2.18286291e-01 4.46424842e-01 5.68046153e-01
-2.76547730e-01 3.72139394e-01 -1.47704139e-01 -3.10939133e-01
2.42211714e-01 -3.03101629e-01 -1.36900187e-01 3.61978173e-01
-1.34756458e+00 1.42599905e+00 4.67344671e-01 -4.04646307e-01
-1.54518139e+00 -7.53466249e-01 -3.51159811e-01 -4.01635528e-01
4.76542532e-01 -3.10213007e-02 1.30614758e+00 -1.05698502e+00
-8.13492298e-01 1.16630650e+00 6.73217535e-01 -8.34030628e-01
7.01690316e-01 -5.49909413e-01 -8.72176886e-01 8.71575177e-02
6.04647398e-01 -1.70188129e-01 9.85467911e-01 -8.25534344e-01
-9.46126163e-01 -1.37115300e-01 3.00028990e-03 -2.34495834e-01
-1.05986372e-01 8.81924927e-01 7.42170751e-01 -9.81451511e-01
-2.23538995e-01 -8.32342923e-01 5.37929893e-01 -8.84577155e-01
-7.80479610e-01 -5.52562058e-01 8.09257686e-01 -1.14650762e+00
1.57168388e+00 -1.58245373e+00 1.54444322e-01 7.42092952e-02
4.06839967e-01 3.07273000e-01 8.67362097e-02 6.15531623e-01
2.33678240e-02 1.92397878e-01 4.66330379e-01 5.07437348e-01
-2.28997678e-01 -3.31436306e-01 -6.03345513e-01 8.06715190e-01
-2.07600355e-01 1.81830794e-01 -1.00617886e+00 -5.68546534e-01
-1.11742206e-01 -1.18009664e-01 -1.90381095e-01 -3.26797999e-02
-2.60571927e-01 -2.34997366e-02 -6.13524377e-01 3.40955973e-01
1.31834790e-01 1.08397126e-01 1.20876975e-01 4.11308184e-02
-4.38311487e-01 7.51465201e-01 -3.29807997e-01 1.06735814e+00
-3.72362062e-02 1.03390241e+00 -1.18732311e-01 -9.86317098e-01
8.82784426e-01 1.87581882e-01 1.79283932e-01 -8.41768086e-01
6.90338194e-01 1.48580536e-01 3.85944486e-01 -6.64272785e-01
4.15514141e-01 -3.31153452e-01 -6.61237001e-01 5.40493846e-01
8.25691670e-02 8.65079910e-02 4.03385341e-01 5.55144131e-01
1.48177910e+00 -1.60533905e-01 2.28071421e-01 -8.37445974e-01
3.82985562e-01 6.81102037e-01 2.87512749e-01 1.16410613e+00
-2.48698235e-01 2.86769625e-02 6.06782258e-01 -8.73866081e-01
-9.86786783e-01 -8.09891045e-01 -1.07060567e-01 1.62097692e+00
-3.88870239e-01 -6.03279114e-01 -4.87129778e-01 -1.29411650e+00
-4.70825762e-01 1.34266877e+00 -1.23607373e+00 1.66849233e-02
-7.22031534e-01 -5.39826989e-01 1.10590494e+00 1.74214363e-01
5.94772339e-01 -7.78114974e-01 -9.45770562e-01 1.01181760e-01
-7.01839149e-01 -6.93869710e-01 -1.20129846e-01 3.26446474e-01
-3.20176214e-01 -1.29022253e+00 -3.40728343e-01 -7.75110662e-01
-1.00445980e-02 7.21869022e-02 1.10372305e+00 -1.50422066e-01
-1.25803366e-01 -1.03950083e-01 -8.00480545e-01 -7.32863605e-01
-9.43524182e-01 1.18011706e-01 9.35022682e-02 -4.22824770e-01
2.53260314e-01 6.79241866e-02 1.42818466e-02 2.23259285e-01
-2.70921916e-01 2.00958773e-01 2.23054230e-01 7.48029411e-01
-2.47255087e-01 2.18287647e-01 3.90746593e-01 -9.86484051e-01
9.60225701e-01 -6.17837310e-01 -2.01992795e-01 1.65330157e-01
-4.56472695e-01 -2.12166458e-01 6.35420859e-01 -5.19737065e-01
-1.01093864e+00 -4.83277559e-01 -3.98769639e-02 6.76665679e-02
2.23143131e-01 6.69445038e-01 6.77997112e-01 5.20361245e-01
1.64613330e+00 1.83152243e-01 -3.24664176e-01 -5.83913565e-01
-1.08272322e-01 7.91764855e-01 6.46880507e-01 -4.83363450e-01
6.55242205e-01 1.31377563e-01 -4.91323471e-01 -6.35908186e-01
-1.53762937e+00 -4.50008065e-01 -2.54606485e-01 -5.94926953e-01
7.34962285e-01 -6.33539200e-01 -4.18815613e-01 5.78797638e-01
-1.34058177e+00 -1.10336758e-01 1.27065346e-01 4.40150797e-01
-3.86788130e-01 7.97288418e-02 -6.44136131e-01 -8.65590692e-01
-7.37921834e-01 -5.80557168e-01 7.30569482e-01 -1.70644432e-01
-5.78134954e-01 -8.17912281e-01 2.84280926e-01 6.95165277e-01
1.35248512e-01 4.35872942e-01 9.70932901e-01 -1.49029493e+00
6.66106462e-01 -4.77732986e-01 -1.14317695e-02 2.83866614e-01
-1.74288332e-01 1.58981979e-01 -5.78867197e-01 -1.47337034e-01
3.40478152e-01 -6.49559796e-01 1.01413262e+00 2.55173534e-01
4.14012313e-01 -9.15020883e-01 -3.67988706e-01 -3.11235696e-01
9.52665031e-01 2.17890561e-01 5.19695759e-01 9.04666007e-01
4.65884477e-01 6.52778685e-01 7.87343383e-01 4.28297102e-01
-5.19644432e-02 7.29409814e-01 2.02473238e-01 6.01080954e-01
-2.39617810e-01 -4.40761298e-01 7.96099544e-01 5.14666677e-01
-4.21602398e-01 -3.76678675e-01 -1.30020571e+00 5.71638465e-01
-1.61510444e+00 -1.85880232e+00 -7.54671812e-01 1.68301189e+00
7.54429042e-01 4.92309511e-01 4.43509132e-01 4.36249912e-01
8.12294245e-01 4.00178492e-01 3.69779587e-01 -8.51756632e-01
-1.19657800e-01 9.21554118e-02 2.57542521e-01 4.85394925e-01
-1.61640918e+00 9.83695149e-01 7.32546616e+00 1.43281579e+00
-1.03773689e+00 3.78272533e-01 2.00781748e-01 -3.10370266e-01
-2.02817656e-03 -2.14219213e-01 -8.51340294e-01 6.15416288e-01
1.25746095e+00 -2.62648046e-01 1.96145363e-02 8.32158625e-01
1.11312635e-01 -1.48264870e-01 -7.01329231e-01 3.22766185e-01
5.31896234e-01 -1.66188252e+00 1.92853481e-01 5.03659770e-02
8.18939328e-01 1.45879477e-01 -1.52955875e-01 7.44396031e-01
6.63247585e-01 -1.13223183e+00 1.40933800e+00 6.10824376e-02
4.36999917e-01 -6.17800117e-01 7.88865745e-01 9.82650876e-01
-4.27599728e-01 -2.81824559e-01 2.65499521e-02 -3.41108859e-01
1.72869205e-01 5.01760781e-01 -1.03008163e+00 3.63106042e-01
6.45810127e-01 5.21067441e-01 -7.37635314e-01 5.78372955e-01
-6.73825920e-01 1.26794457e+00 -1.25827074e-01 -2.60369271e-01
6.17267311e-01 4.50742006e-01 1.03041315e+00 1.53107631e+00
-6.91351481e-03 -6.84789941e-02 3.80058348e-01 3.20920616e-01
2.78734237e-01 1.68384343e-01 -5.56428015e-01 -2.31662929e-01
4.10997212e-01 8.00066352e-01 -7.30335414e-01 -5.20172358e-01
2.38755330e-01 3.59423190e-01 1.26061872e-01 -2.07761630e-01
-1.27219427e+00 -3.45392793e-01 -1.78391680e-01 6.02671683e-01
-6.91337734e-02 7.56272972e-02 -1.83681518e-01 -8.00196707e-01
-4.85637009e-01 -1.20703053e+00 7.96878457e-01 -4.93341476e-01
-1.23696578e+00 8.72595668e-01 2.68525928e-01 -9.02221322e-01
-3.09525728e-01 -5.50917387e-01 -5.37059724e-01 3.24925780e-01
-4.30602193e-01 -1.36016631e+00 1.25718098e-02 3.81469876e-01
9.56281185e-01 -8.07752311e-01 4.95516032e-01 8.24309215e-02
-2.46024340e-01 3.42183441e-01 -6.47061989e-02 5.51961005e-01
7.16476440e-01 -9.30289567e-01 4.06851694e-02 1.00038314e+00
2.79964190e-02 4.45617437e-01 1.48680425e+00 -9.85142291e-01
-9.71721649e-01 -1.01715624e+00 1.23632860e+00 -8.85831654e-01
1.20481789e+00 -1.75592024e-02 -4.16394681e-01 7.02382267e-01
3.80322546e-01 -9.42360520e-01 6.30922496e-01 5.12389481e-01
-9.73061144e-01 5.26960194e-01 -1.01615632e+00 1.51198044e-01
8.98003578e-01 -4.51173365e-01 -1.46908081e+00 9.88736153e-01
3.22806910e-02 -3.76765758e-01 -6.87750578e-01 3.33246052e-01
6.97974920e-01 -8.31542313e-01 6.69432104e-01 -1.21231842e+00
1.47420692e+00 1.20215841e-01 -2.37407729e-01 -1.11369836e+00
-5.11928678e-01 -3.66465867e-01 2.56547686e-02 9.63957131e-01
5.07017612e-01 -1.38103366e-01 3.00810099e-01 -5.00104845e-01
-3.07420850e-01 -3.43653291e-01 -1.11991954e+00 -1.08446324e+00
2.85592318e-01 -5.03627360e-01 -4.24901009e-01 1.32604945e+00
2.95728654e-01 9.77270126e-01 -8.89874160e-01 -2.75830805e-01
5.49387336e-01 2.31768694e-02 6.52181387e-01 -1.16581714e+00
-3.51991534e-01 -6.63194597e-01 -3.53111982e-01 -6.26885593e-01
5.04792146e-02 -1.04566824e+00 1.38595998e-01 -1.61934781e+00
6.41580462e-01 -1.33782059e-01 5.98270074e-02 4.97678757e-01
2.81774104e-01 2.73600876e-01 2.75226869e-02 3.92476797e-01
-4.48335499e-01 -2.65990887e-02 8.35763931e-01 -6.12443149e-01
1.42559528e-01 -7.92823806e-02 -6.54744208e-01 8.64149153e-01
6.88911080e-01 -9.38169301e-01 -3.25320482e-01 -5.77553749e-01
5.43418050e-01 6.63390569e-03 4.32541013e-01 -7.81483054e-01
-6.66879043e-02 -5.00692308e-01 1.64859891e-02 -6.71322346e-01
-1.09318763e-01 1.15349300e-01 1.29532427e-01 9.42566931e-01
-5.89120865e-01 -3.11937612e-02 5.33694774e-02 5.56304574e-01
5.07250838e-02 -4.49200332e-01 8.33017170e-01 -1.59242600e-01
-4.42876726e-01 -5.34379661e-01 -1.08979642e+00 3.78448933e-01
9.77395177e-01 5.31008653e-02 -1.17800558e+00 -1.43697545e-01
-7.29744017e-01 -2.02809889e-02 -2.57701483e-02 6.45172477e-01
1.35662243e-01 -1.01688099e+00 -1.41753137e+00 -5.12734175e-01
9.60012525e-02 -8.68906915e-01 1.42924115e-01 1.05933726e+00
-7.91755497e-01 3.14582050e-01 -4.96561795e-01 -3.14410329e-01
-1.72592926e+00 3.51809293e-01 3.12932879e-01 -5.94965875e-01
-5.96132994e-01 1.19356000e+00 -1.89853966e-01 -1.51164886e-02
-8.75766277e-02 3.05932581e-01 -5.35623908e-01 4.70138967e-01
9.26548719e-01 7.72969961e-01 -7.74709955e-02 -8.81425261e-01
-6.19282722e-01 -5.33147343e-02 -6.62405312e-01 -8.23391005e-02
1.23814404e+00 5.91687083e-01 -1.38544232e-01 4.72594887e-01
8.34696949e-01 2.26727918e-01 -7.23574683e-02 -2.67955847e-02
1.40748441e-01 -3.93617272e-01 3.10692817e-01 -1.43545806e+00
-5.25439620e-01 5.30977726e-01 4.39236999e-01 6.07170820e-01
4.23689425e-01 2.21051529e-01 3.77939999e-01 3.62004161e-01
3.40153843e-01 -1.64655542e+00 2.58149385e-01 9.05106366e-01
1.36836302e+00 -6.86838686e-01 3.76969606e-01 -4.26596254e-01
-5.82017362e-01 8.82697225e-01 2.77040094e-01 -3.79934669e-01
2.47936904e-01 8.60921517e-02 4.04056744e-04 -8.86190772e-01
-9.27711308e-01 1.76236853e-01 4.50534403e-01 4.56903696e-01
5.54211974e-01 2.25009724e-01 -1.12641716e+00 4.79322731e-01
-5.51040530e-01 -3.21376234e-01 5.70862710e-01 9.97899592e-01
-8.76965821e-01 -2.06763819e-01 -5.15604675e-01 8.15478802e-01
-8.61876309e-01 -8.69837329e-02 -1.13500428e+00 9.30402875e-01
2.42680565e-01 1.26060212e+00 -2.05187142e-01 -1.07420659e+00
2.80340374e-01 -2.77376398e-02 6.27441764e-01 -6.90864742e-01
-1.41749251e+00 1.63376048e-01 1.13964570e+00 -1.30301535e-01
-4.55123097e-01 -6.76371515e-01 -7.05613673e-01 -8.82955015e-01
-4.21077460e-01 4.57914442e-01 2.93492019e-01 1.14009202e+00
-3.52919698e-01 5.73479831e-01 5.07652700e-01 -3.58114958e-01
-1.01700938e+00 -1.49618769e+00 -3.53348076e-01 5.89016020e-01
-1.72346339e-01 -7.58324623e-01 -4.62433994e-01 6.89845532e-02]
|
[8.465950012207031, 10.675714492797852]
|
3a663ebb-b0ee-4eb6-8b7c-12f57e7d7d56
|
flexivit-one-model-for-all-patch-sizes
|
2212.08013
| null |
https://arxiv.org/abs/2212.08013v2
|
https://arxiv.org/pdf/2212.08013v2.pdf
|
FlexiViT: One Model for All Patch Sizes
|
Vision Transformers convert images to sequences by slicing them into patches. The size of these patches controls a speed/accuracy tradeoff, with smaller patches leading to higher accuracy at greater computational cost, but changing the patch size typically requires retraining the model. In this paper, we demonstrate that simply randomizing the patch size at training time leads to a single set of weights that performs well across a wide range of patch sizes, making it possible to tailor the model to different compute budgets at deployment time. We extensively evaluate the resulting model, which we call FlexiViT, on a wide range of tasks, including classification, image-text retrieval, open-world detection, panoptic segmentation, and semantic segmentation, concluding that it usually matches, and sometimes outperforms, standard ViT models trained at a single patch size in an otherwise identical setup. Hence, FlexiViT training is a simple drop-in improvement for ViT that makes it easy to add compute-adaptive capabilities to most models relying on a ViT backbone architecture. Code and pre-trained models are available at https://github.com/google-research/big_vision
|
['Filip Pavetic', 'Ibrahim Alabdulmohsin', 'Michael Tschannen', 'Matthias Minderer', 'Xiaohua Zhai', 'Simon Kornblith', 'Mathilde Caron', 'Alexander Kolesnikov', 'Pavel Izmailov', 'Lucas Beyer']
|
2022-12-15
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Beyer_FlexiViT_One_Model_for_All_Patch_Sizes_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Beyer_FlexiViT_One_Model_for_All_Patch_Sizes_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['panoptic-segmentation']
|
['computer-vision']
|
[ 3.13736171e-01 -3.81498188e-01 -2.22435519e-01 -3.07808220e-01
-8.06677997e-01 -1.08383763e+00 6.74090505e-01 -5.87548912e-02
-4.56409514e-01 1.24161460e-01 -2.01445863e-01 -4.88416255e-01
1.59979016e-01 -6.72543764e-01 -7.71735251e-01 -6.38968766e-01
2.80355364e-01 6.20549619e-01 5.49750865e-01 -1.51142441e-02
3.13703567e-01 2.83424079e-01 -1.62914777e+00 2.68407106e-01
6.25282764e-01 1.11847985e+00 4.76506054e-01 1.01958120e+00
-6.08089231e-02 4.56386894e-01 -5.24431109e-01 -5.61934531e-01
6.93926334e-01 -1.99789360e-01 -8.51493716e-01 3.39914709e-01
8.78171921e-01 -2.87608713e-01 -8.35086703e-02 8.38269293e-01
3.94469142e-01 4.73844372e-02 4.47744459e-01 -1.06790304e+00
-6.76463187e-01 4.30844098e-01 -6.78017318e-01 3.91990423e-01
9.67134163e-03 4.24640507e-01 1.19521618e+00 -5.81221044e-01
6.23539805e-01 9.96983647e-01 1.03089535e+00 2.47564092e-01
-1.45357943e+00 -3.86944711e-01 2.74592400e-01 2.24484149e-02
-1.25152719e+00 -4.51510698e-01 1.79509029e-01 -4.44109529e-01
1.07207465e+00 2.70022422e-01 6.53689742e-01 8.86768341e-01
7.24799465e-03 7.40634680e-01 1.12518013e+00 -3.05824399e-01
1.82236150e-01 -1.29304649e-02 9.85490009e-02 5.19463599e-01
1.38006568e-01 -6.16543815e-02 -1.34068012e-01 -1.09585389e-01
7.57324934e-01 9.68034640e-02 -2.54797995e-01 -4.29686069e-01
-1.05596149e+00 8.24526012e-01 4.73587424e-01 8.33812803e-02
-1.58068031e-01 4.64433700e-01 4.03518051e-01 4.79184359e-01
5.52444577e-01 4.24906582e-01 -6.14089429e-01 2.70490851e-02
-1.12856507e+00 2.61531621e-01 7.30357766e-01 7.37811863e-01
1.05901635e+00 -1.23683354e-02 -2.30416715e-01 1.02855670e+00
-5.23325317e-02 6.18799627e-01 4.22813773e-01 -1.38230729e+00
1.66999653e-01 3.54399383e-01 2.82954285e-03 -7.27763593e-01
-2.06046194e-01 -4.98577058e-01 -5.45925736e-01 1.78165346e-01
4.81285602e-01 -1.80056289e-01 -1.52927339e+00 1.56854403e+00
5.57749867e-01 3.42203617e-01 -3.18637371e-01 8.93853605e-01
6.16373658e-01 9.60984945e-01 -1.07508279e-01 2.44753361e-01
1.51653731e+00 -1.18098235e+00 1.35878874e-02 -6.30994499e-01
2.90740997e-01 -1.07244539e+00 1.20767224e+00 4.56582546e-01
-1.02128410e+00 -5.23881316e-01 -9.00036037e-01 -1.76204443e-01
-3.49491179e-01 -1.09842449e-01 6.85972452e-01 5.57651997e-01
-1.41972947e+00 7.10518718e-01 -9.07480657e-01 -6.31036997e-01
2.94572353e-01 2.71896511e-01 -1.42706245e-01 -1.62324965e-01
-6.61726654e-01 7.20738888e-01 1.24917440e-01 -6.07942156e-02
-1.01779902e+00 -7.95415640e-01 -6.04002357e-01 1.03721254e-01
3.94242972e-01 -8.32342803e-01 1.69429910e+00 -1.25680578e+00
-1.29630411e+00 9.88368928e-01 -8.49950165e-02 -6.50612891e-01
4.38790053e-01 -5.54991476e-02 7.18548596e-02 1.85243636e-01
5.98290712e-02 1.08076859e+00 1.52669418e+00 -9.36226666e-01
-6.52955413e-01 -1.83253065e-01 2.12912485e-01 1.89877793e-01
-1.58942878e-01 -7.40735531e-02 -9.29243743e-01 -5.69479823e-01
-2.12663069e-01 -1.08552623e+00 -3.91593575e-01 4.11602408e-02
-5.57283200e-02 -4.69436310e-02 6.22038007e-01 -3.47272724e-01
7.69820273e-01 -2.07656670e+00 3.51353958e-02 1.91422794e-02
2.49848142e-01 4.47199523e-01 -4.74910676e-01 5.30542254e-01
1.06779851e-01 6.07409887e-02 -4.93893564e-01 -3.27732801e-01
-1.26338273e-01 4.66733932e-01 -2.53744692e-01 4.94787484e-01
2.24190727e-01 9.03421521e-01 -6.10678077e-01 -3.77825081e-01
3.29251051e-01 2.97912866e-01 -6.22984648e-01 2.52602668e-03
-5.01148403e-01 1.39203370e-01 -3.80218893e-01 4.88328815e-01
8.65447044e-01 -6.20837629e-01 -2.33086701e-02 1.31251633e-01
-1.86238065e-01 2.56705731e-01 -1.20841050e+00 1.54335666e+00
-4.85652536e-01 7.75370777e-01 3.16861361e-01 -1.15060997e+00
6.28602147e-01 2.38740537e-02 2.32878268e-01 -6.78010046e-01
4.19101641e-02 -1.07183531e-02 -1.77858680e-01 -2.75540322e-01
6.24999583e-01 6.93678483e-02 1.43380716e-01 5.69546103e-01
1.50003865e-01 -3.82732302e-01 3.34405363e-01 1.74864694e-01
1.11133754e+00 1.04820788e-01 4.83164266e-02 -1.46202356e-01
4.28168625e-02 4.05992091e-01 3.85038525e-01 1.11855507e+00
-1.47454008e-01 9.20480847e-01 4.13107663e-01 -5.89093029e-01
-1.30367672e+00 -9.04379964e-01 -3.11361313e-01 1.44739044e+00
1.27496302e-01 -2.85073817e-01 -8.54220331e-01 -4.07293856e-01
1.69730514e-01 2.10358530e-01 -6.71836495e-01 2.38356441e-01
-4.54080939e-01 -7.30267644e-01 5.94166934e-01 4.27835256e-01
4.36482251e-01 -1.02766144e+00 -7.17594326e-01 -1.34000937e-02
7.76336342e-02 -1.05505180e+00 -6.43269122e-01 2.90567398e-01
-7.49965072e-01 -8.80876303e-01 -8.78392637e-01 -8.31509411e-01
3.06494713e-01 7.54596531e-01 1.45045948e+00 3.63380700e-01
-4.71782625e-01 4.98037070e-01 -4.62314069e-01 -2.37569585e-01
-2.72922337e-01 4.67499167e-01 -4.33027387e-01 -1.39810100e-01
-6.75257519e-02 -4.09795910e-01 -9.03457165e-01 3.29011679e-01
-1.13728011e+00 -4.24962528e-02 4.51397747e-01 8.62870038e-01
7.29200721e-01 -3.96830976e-01 4.50963043e-02 -1.01456392e+00
3.05649728e-01 -4.15425926e-01 -9.07907605e-01 2.26984128e-01
-7.45881736e-01 1.91947930e-02 5.96505165e-01 -5.28363347e-01
-7.19186187e-01 -2.93954695e-03 -1.51623785e-01 -5.22749722e-01
-1.36643469e-01 3.01829755e-01 5.06431699e-01 -1.59473166e-01
7.56177962e-01 2.02137694e-01 1.03796199e-01 -5.24406672e-01
7.05835640e-01 5.82334876e-01 4.97012854e-01 -4.28349465e-01
7.14018047e-01 4.55933630e-01 -3.69172186e-01 -6.46335721e-01
-9.40789044e-01 -6.61863208e-01 -2.60030150e-01 1.06306732e-01
7.02022791e-01 -1.01182950e+00 -2.77064860e-01 4.93491679e-01
-8.97971213e-01 -9.04653251e-01 -2.08298966e-01 -4.58582640e-02
-3.34290236e-01 3.44967574e-01 -6.06854856e-01 -1.78611889e-01
-5.56409299e-01 -1.11348844e+00 1.25983846e+00 2.06320345e-01
-2.34305416e-03 -9.25440431e-01 2.84333467e-01 3.44029844e-01
5.70744336e-01 -1.17959350e-01 4.85335439e-01 -2.71619409e-01
-6.69713199e-01 -5.86131364e-02 -4.53921080e-01 4.57248002e-01
-1.79159939e-01 3.28207344e-01 -1.10300207e+00 -4.14908588e-01
-3.10390800e-01 -3.63215595e-01 1.32478857e+00 6.68567300e-01
1.18578959e+00 -2.56261677e-01 -2.27653518e-01 1.04421425e+00
1.51284981e+00 -2.53306150e-01 6.58240318e-01 4.48356986e-01
6.29518092e-01 2.66594052e-01 3.75079870e-01 3.20262611e-01
3.59768271e-01 5.74731767e-01 4.68985349e-01 -2.70314842e-01
-2.36908421e-01 3.13948318e-02 2.03548804e-01 3.74619007e-01
1.57243460e-01 -3.24695289e-01 -1.03125763e+00 7.65623093e-01
-1.84749603e+00 -8.83228123e-01 3.23067792e-03 2.18640447e+00
7.63984740e-01 -5.35610691e-03 2.18563050e-01 -4.16072518e-01
6.82086766e-01 4.57028180e-01 -7.28421330e-01 -5.86332381e-01
1.13643110e-01 5.00223279e-01 9.09328461e-01 6.36051655e-01
-1.17888939e+00 1.28105545e+00 7.06670284e+00 9.84934032e-01
-1.56133795e+00 1.85423523e-01 7.22486079e-01 -2.21518859e-01
-1.99948043e-01 2.42376864e-01 -6.35888457e-01 3.63842100e-01
9.98088360e-01 5.32334521e-02 7.44065523e-01 9.89848614e-01
-7.56597221e-02 -1.51188329e-01 -8.21061730e-01 8.65187883e-01
-1.04401052e-01 -1.49648070e+00 8.05440620e-02 -7.15161934e-02
7.51242280e-01 8.43505740e-01 2.23479092e-01 1.67826027e-01
4.69769269e-01 -9.66782987e-01 7.78068423e-01 -1.76006462e-02
7.79693127e-01 -3.50641787e-01 4.90817994e-01 1.71699136e-01
-1.06834257e+00 -4.99407947e-02 -6.61597311e-01 -5.62886000e-02
-1.09940268e-01 4.26175177e-01 -7.66460359e-01 1.23118490e-01
1.04302025e+00 5.52945912e-01 -8.08576822e-01 1.25670028e+00
-1.26314899e-02 8.52612376e-01 -5.47370136e-01 1.98473707e-01
5.09542584e-01 -2.26791903e-01 2.97881186e-01 1.46612632e+00
2.52343088e-01 -1.36532411e-01 2.63778001e-01 6.16826892e-01
-1.81211770e-01 -1.37326762e-01 -5.84911168e-01 1.49170801e-01
5.83143830e-01 1.48823893e+00 -9.26654935e-01 -4.96466219e-01
-4.65593815e-01 9.50206399e-01 3.25498044e-01 3.50103229e-01
-8.62386644e-01 -2.43415207e-01 7.28163660e-01 1.20393746e-01
1.05295897e+00 -2.14953214e-01 -3.21623862e-01 -1.12466955e+00
-1.09350294e-01 -9.99196470e-01 4.43582654e-01 -8.31175447e-01
-1.14489698e+00 5.87310076e-01 -3.49557027e-02 -9.60530102e-01
-1.88993469e-01 -5.81610858e-01 -6.17761672e-01 6.31351054e-01
-1.60485339e+00 -1.01092184e+00 -4.52187628e-01 4.94834214e-01
6.37149751e-01 3.06615502e-01 5.36841035e-01 5.99580482e-02
-6.08778238e-01 5.69364965e-01 3.14874500e-01 -4.30842005e-02
8.11380327e-01 -1.26459730e+00 8.51163983e-01 8.92628789e-01
4.83532250e-01 4.06750143e-01 7.59226322e-01 -1.20256297e-01
-1.35470331e+00 -1.07202721e+00 4.53569502e-01 -5.01388311e-01
7.68284678e-01 -3.01354080e-01 -8.99497151e-01 8.24322641e-01
5.40538728e-01 1.26794592e-01 3.75694036e-01 2.10522398e-01
-6.07752383e-01 -3.11635911e-01 -8.87688518e-01 5.23412943e-01
8.54000926e-01 -3.22507262e-01 -2.32254401e-01 3.87690395e-01
6.88208580e-01 -3.71692538e-01 -7.26606488e-01 5.51878512e-02
5.82730830e-01 -9.82220888e-01 1.05184531e+00 -3.29497039e-01
3.99134666e-01 -3.29986036e-01 -4.65021506e-02 -1.29685080e+00
-4.98889834e-01 -6.54129386e-01 2.54622519e-01 9.36200023e-01
5.92885554e-01 -7.76245236e-01 6.93267226e-01 4.04048860e-01
-1.54130250e-01 -5.77098846e-01 -7.22535908e-01 -7.18014121e-01
2.07796171e-01 -4.74094450e-01 6.15102112e-01 8.13781321e-01
-5.75776756e-01 3.87862980e-01 -2.03089550e-01 2.20496833e-01
4.28146929e-01 4.93673444e-01 9.28535402e-01 -1.07402921e+00
-6.43019259e-01 -6.72932446e-01 -2.04213187e-01 -1.41894937e+00
-2.32549876e-01 -7.49107182e-01 4.47744653e-02 -1.53949344e+00
2.41442442e-01 -6.39926136e-01 -1.17931731e-01 8.51139843e-01
-3.19131583e-01 7.60199547e-01 3.69024277e-01 4.53983158e-01
-6.63213193e-01 1.45630285e-01 1.12680328e+00 -2.34508559e-01
-1.10513143e-01 -7.96128288e-02 -8.44572783e-01 5.64485431e-01
9.32363868e-01 -4.65651214e-01 -3.91472906e-01 -1.02105963e+00
1.98318496e-01 -1.80863336e-01 4.47328061e-01 -8.65265310e-01
-1.93568692e-02 -2.47432530e-01 2.27346197e-01 -2.01581791e-01
3.39979053e-01 -4.38579232e-01 7.53385499e-02 2.07783878e-01
-2.10753992e-01 1.81980178e-01 3.05905730e-01 4.04936224e-01
6.26309440e-02 -3.91432315e-01 9.28516984e-01 -2.79675841e-01
-7.79847682e-01 3.72309208e-01 -4.10677910e-01 1.90662429e-01
7.67200172e-01 -3.09878200e-01 -4.61806923e-01 -3.51838022e-01
-3.68720651e-01 2.52688915e-01 7.87388444e-01 3.32764417e-01
1.81503683e-01 -7.57648468e-01 -7.04996705e-01 9.62781981e-02
5.07764518e-02 6.96803406e-02 1.08575881e-01 8.05990279e-01
-8.66388559e-01 1.68151394e-01 -1.04065437e-03 -1.03793967e+00
-1.21090019e+00 5.61704695e-01 3.64142209e-01 -1.93363845e-01
-7.94573665e-01 1.07874703e+00 3.45676124e-01 -4.48096126e-01
8.41758251e-02 -4.37001646e-01 2.28448138e-01 -1.62933879e-02
4.82072860e-01 3.69758010e-02 2.02228293e-01 -2.55490631e-01
-2.24930227e-01 7.29415536e-01 -2.43263066e-01 -7.79310837e-02
1.27094519e+00 -1.17032953e-01 -4.01351042e-02 1.96848974e-01
8.99029672e-01 -2.92463601e-01 -1.66381967e+00 -3.04692030e-01
-3.38627726e-01 -5.12012601e-01 2.42887482e-01 -6.37251616e-01
-1.29118705e+00 8.18458676e-01 5.62516332e-01 4.83034998e-01
1.11421263e+00 1.33439749e-01 8.51018906e-01 4.93111432e-01
2.52041399e-01 -8.75972152e-01 -1.49728656e-01 5.86293161e-01
5.27498424e-01 -1.24900246e+00 1.09966703e-01 -1.22727752e-01
-8.76763761e-01 8.06735039e-01 3.32568616e-01 -5.57232261e-01
5.42942941e-01 3.54345143e-01 2.86589116e-01 -7.44608939e-02
-9.58667576e-01 -4.55897987e-01 8.46038684e-02 6.37067437e-01
2.34451696e-01 1.01714738e-01 1.38156128e-03 -1.74123034e-01
-2.24410504e-01 -1.79305062e-01 4.33825493e-01 7.83622265e-01
-6.27367556e-01 -1.18270350e+00 -4.81142491e-01 5.61511040e-01
-4.40542400e-01 -3.93013954e-01 -3.05306166e-01 6.61631763e-01
9.37790200e-02 8.78314972e-01 3.68055195e-01 -1.26533657e-01
9.66332257e-02 -1.72401387e-02 5.89418769e-01 -5.73276103e-01
-8.21972966e-01 7.35836774e-02 3.08322068e-02 -5.88355660e-01
-2.91958958e-01 -6.18499637e-01 -8.67648602e-01 -5.50819874e-01
-2.90570170e-01 -8.81554410e-02 6.00750923e-01 7.92240083e-01
6.60748184e-01 9.37171653e-02 5.02855241e-01 -1.06850493e+00
-6.28434956e-01 -8.75927329e-01 -3.96995932e-01 2.81785071e-01
5.10821700e-01 -3.26693028e-01 -2.84756601e-01 2.05167443e-01]
|
[9.561599731445312, 1.3129940032958984]
|
71ace4a8-ba8a-4cac-8412-ba69e988cd1e
|
efficientderain-learning-pixel-wise-dilation
|
2009.09238
| null |
https://arxiv.org/abs/2009.09238v1
|
https://arxiv.org/pdf/2009.09238v1.pdf
|
EfficientDeRain: Learning Pixel-wise Dilation Filtering for High-Efficiency Single-Image Deraining
|
Single-image deraining is rather challenging due to the unknown rain model. Existing methods often make specific assumptions of the rain model, which can hardly cover many diverse circumstances in the real world, making them have to employ complex optimization or progressive refinement. This, however, significantly affects these methods' efficiency and effectiveness for many efficiency-critical applications. To fill this gap, in this paper, we regard the single-image deraining as a general image-enhancing problem and originally propose a model-free deraining method, i.e., EfficientDeRain, which is able to process a rainy image within 10~ms (i.e., around 6~ms on average), over 80 times faster than the state-of-the-art method (i.e., RCDNet), while achieving similar de-rain effects. We first propose the novel pixel-wise dilation filtering. In particular, a rainy image is filtered with the pixel-wise kernels estimated from a kernel prediction network, by which suitable multi-scale kernels for each pixel can be efficiently predicted. Then, to eliminate the gap between synthetic and real data, we further propose an effective data augmentation method (i.e., RainMix) that helps to train network for real rainy image handling.We perform comprehensive evaluation on both synthetic and real-world rainy datasets to demonstrate the effectiveness and efficiency of our method. We release the model and code in https://github.com/tsingqguo/efficientderain.git.
|
['Xiaofei Xie', 'Felix Juefei-Xu', 'Lei Ma', 'Jingyang Sun', 'Yang Liu', 'Qing Guo', 'Wei Feng']
|
2020-09-19
| null | null | null | null |
['single-image-deraining']
|
['computer-vision']
|
[ 1.51772827e-01 -4.17434365e-01 4.02326912e-01 -5.25269449e-01
-4.98169631e-01 -2.74582624e-01 1.09558754e-01 -2.69721419e-01
-5.85706770e-01 7.76813805e-01 -2.72035003e-01 -4.74564701e-01
2.68421710e-01 -8.50115180e-01 -7.66345024e-01 -1.02139723e+00
1.34348944e-01 5.05150445e-02 2.07653135e-01 -1.67028546e-01
-6.41177967e-03 5.85944533e-01 -1.54479444e+00 -1.47239372e-01
1.59561610e+00 5.96301615e-01 6.75445318e-01 7.88664460e-01
-8.52698162e-02 6.57905459e-01 -5.40099621e-01 -2.88667884e-02
2.15311885e-01 -4.89719391e-01 -2.78679520e-01 1.65084437e-01
6.31368101e-01 -6.71402276e-01 -4.65804249e-01 1.10756016e+00
6.09903514e-01 9.34009776e-02 2.46056736e-01 -8.40337992e-01
-3.81881803e-01 1.49391487e-01 -6.71428144e-01 4.64386851e-01
-2.69977808e-01 3.18091333e-01 4.06695724e-01 -1.14220119e+00
2.46778756e-01 1.15400243e+00 4.84920412e-01 2.88109004e-01
-9.99610484e-01 -8.72293293e-01 2.52033591e-01 8.00175890e-02
-1.36931932e+00 -4.10673887e-01 4.90148604e-01 -5.65465316e-02
3.79667372e-01 5.04038274e-01 7.17673242e-01 5.36358714e-01
1.16805114e-01 8.05130363e-01 1.56727529e+00 -1.78800002e-01
1.46124065e-01 -8.87187794e-02 2.31932461e-01 5.62773168e-01
5.04027903e-01 2.04473406e-01 -4.01905291e-02 1.97974935e-01
8.50536823e-01 3.91465038e-01 -7.31362820e-01 7.13986456e-02
-1.03354371e+00 6.73349977e-01 6.20566666e-01 7.79570192e-02
-4.05800790e-01 -1.67567864e-01 -1.18088879e-01 3.90015095e-01
9.10361290e-01 1.02563590e-01 -4.62295055e-01 2.03102961e-01
-1.27782297e+00 6.29212856e-01 6.14282250e-01 5.99544466e-01
1.09983325e+00 2.75787950e-01 -1.89972162e-01 8.10306370e-01
2.41253242e-01 1.16004455e+00 1.12384586e-02 -6.74445510e-01
5.04533947e-01 2.17299163e-01 3.85248899e-01 -7.46053100e-01
-4.75411057e-01 -4.53232527e-01 -1.22713661e+00 5.18189728e-01
3.27286094e-01 -3.26331258e-01 -1.36259079e+00 1.28909504e+00
4.88851905e-01 6.41168058e-01 2.08786547e-01 1.17610931e+00
8.30293000e-01 1.03453147e+00 -1.14458755e-01 -6.36262596e-01
1.24593794e+00 -1.05053473e+00 -7.93072045e-01 -4.55424219e-01
2.74509519e-01 -8.98825169e-01 1.23879850e+00 2.79165983e-01
-8.88250113e-01 -5.66457093e-01 -1.00510812e+00 2.39318654e-01
-1.05756901e-01 1.82270452e-01 5.75562596e-01 4.05466735e-01
-7.16577649e-01 5.64791441e-01 -9.12617505e-01 -3.55145782e-01
4.05860782e-01 6.12992533e-02 -2.36193705e-02 -5.33866644e-01
-1.13615406e+00 8.22944224e-01 3.80532742e-01 5.73815942e-01
-1.02583253e+00 -8.69367898e-01 -5.95828176e-01 -4.44864109e-02
3.48489553e-01 -5.66509962e-01 1.06845546e+00 -1.01391792e+00
-1.32398784e+00 5.20976186e-01 -3.23739082e-01 -3.76933426e-01
3.44291717e-01 -7.72248328e-01 -5.52367806e-01 1.29126027e-01
-2.35156134e-01 3.63868505e-01 1.28186607e+00 -1.53343976e+00
-8.17376256e-01 -1.59087256e-01 1.41161516e-01 3.99300814e-01
-1.11823745e-01 -1.97218005e-02 -7.22160757e-01 -8.70883226e-01
1.27802759e-01 -9.07699645e-01 -4.73742217e-01 -6.51649237e-02
-2.66029090e-01 4.05822605e-01 9.93783832e-01 -8.82755339e-01
1.40090084e+00 -2.23060775e+00 -1.77307159e-01 -6.27801102e-03
4.29135978e-01 8.48760784e-01 -1.80878431e-01 5.42692654e-02
-4.66922894e-02 -1.22386850e-01 -7.98261762e-01 -3.31435591e-01
-4.46621001e-01 4.37630922e-01 -3.46794546e-01 5.73274791e-01
1.65591076e-01 6.44903541e-01 -5.98202467e-01 -3.24692398e-01
3.28840077e-01 6.66459799e-01 -1.72756955e-01 7.15979517e-01
-1.83053628e-01 5.30975938e-01 -3.91447663e-01 7.67380774e-01
1.39485955e+00 -8.48155841e-02 -2.33071461e-01 -3.54294270e-01
-3.38464886e-01 -3.18357199e-01 -1.26277888e+00 1.12313473e+00
-6.54437602e-01 3.92198205e-01 2.38976508e-01 -6.88301325e-01
9.16075706e-01 7.99964368e-02 -5.60133532e-02 -6.26211822e-01
-3.56590003e-03 2.75955200e-01 -1.19524069e-01 -5.81104636e-01
3.68194342e-01 -1.56029850e-01 5.21598339e-01 3.02317739e-01
-4.34253246e-01 -1.46162227e-01 8.20814818e-02 1.78397790e-01
8.19416821e-01 7.79285282e-02 2.26522852e-02 -1.03971064e-01
3.23962957e-01 -1.18138500e-01 6.62145019e-01 7.86985159e-01
-9.49771255e-02 9.11304653e-01 -6.01996668e-03 -5.16817749e-01
-1.00205612e+00 -1.01222670e+00 -1.76412836e-01 7.95633554e-01
4.78683472e-01 -9.82252508e-03 -8.55241954e-01 -5.31503677e-01
-4.89327237e-02 6.03882909e-01 -5.02526224e-01 1.64967373e-01
-7.19255805e-01 -1.57452142e+00 3.34585637e-01 2.04507604e-01
8.65074456e-01 -1.24840713e+00 -5.47098219e-01 2.26551384e-01
-1.90892667e-01 -1.18639576e+00 -3.06121379e-01 3.11692655e-02
-1.03595352e+00 -9.13456559e-01 -7.24718034e-01 -4.81286228e-01
7.47874439e-01 8.12653065e-01 1.23374093e+00 4.48234856e-01
-2.63676703e-01 -1.88543379e-01 -6.29068136e-01 -4.16541159e-01
3.38606274e-04 -1.88625947e-01 -1.87448695e-01 6.09226339e-02
1.32478550e-01 -7.13924587e-01 -9.29359436e-01 6.75087124e-02
-1.27190292e+00 2.11600944e-01 9.66740847e-01 8.08372915e-01
7.42490530e-01 1.03816852e-01 2.74522454e-01 -1.06081533e+00
5.12268841e-01 -3.44266564e-01 -8.16327512e-01 2.04239294e-01
-6.53691411e-01 -1.21222064e-01 5.78793287e-01 -2.72182584e-01
-1.24011576e+00 9.76470411e-02 -1.44719839e-01 -5.03694713e-01
-2.19482273e-01 4.83829349e-01 -2.32406855e-01 -2.88572788e-01
4.68183070e-01 4.41806436e-01 -4.66641933e-02 -5.48099458e-01
4.52664733e-01 5.11834741e-01 5.91299593e-01 -2.69658625e-01
1.38069308e+00 6.07817113e-01 -3.24472040e-01 -8.97695243e-01
-1.07625473e+00 -3.44077080e-01 -3.38873863e-01 -1.45229489e-01
7.33537912e-01 -1.29318750e+00 -3.84840429e-01 8.31048608e-01
-1.02496052e+00 -5.36823809e-01 6.82773069e-02 6.03842199e-01
-4.38190177e-02 4.62042779e-01 -7.18551278e-01 -8.61742973e-01
-7.74156749e-01 -8.86186004e-01 7.78216243e-01 5.61604202e-01
5.31385660e-01 -6.75589859e-01 5.24022318e-02 1.95283085e-01
7.25815833e-01 1.85054839e-01 4.47394758e-01 5.32357395e-03
-8.54386866e-01 7.80770555e-02 -7.04995930e-01 5.95984519e-01
1.01051815e-01 8.10776874e-02 -9.68313634e-01 -5.87070286e-01
1.65198579e-01 -1.28680721e-01 1.14645743e+00 4.40443367e-01
9.97258902e-01 -1.55571893e-01 -8.93498138e-02 9.87654328e-01
1.73064399e+00 -4.05844562e-02 9.87351298e-01 4.11146253e-01
8.15581799e-01 3.21117103e-01 1.08143711e+00 5.49528182e-01
3.09664667e-01 3.72977227e-01 6.76900268e-01 -7.51550436e-01
-8.06351163e-05 3.04070294e-01 3.52528334e-01 8.45484018e-01
-3.68345052e-01 -3.51836354e-01 -5.73645294e-01 4.76148248e-01
-1.76502812e+00 -7.89979279e-01 -2.73043126e-01 2.15194511e+00
8.39097440e-01 4.25316468e-02 -3.17668021e-01 -2.34112382e-01
5.33927739e-01 4.92790878e-01 -6.45319045e-01 6.45274669e-02
-2.71186084e-01 5.59338689e-01 7.01702893e-01 7.52761126e-01
-1.23106217e+00 1.15868890e+00 5.23317194e+00 8.71023595e-01
-1.20672727e+00 8.43333080e-02 5.44886827e-01 -5.69494776e-02
-1.19403623e-01 8.67623687e-02 -8.92876387e-01 6.63708091e-01
7.59886563e-01 5.06974049e-02 5.25811732e-01 4.82054979e-01
8.06127787e-01 -3.16717654e-01 -2.73323268e-01 8.83681357e-01
6.65394366e-02 -9.84900653e-01 -5.20775802e-02 -2.51666278e-01
5.75099826e-01 1.61833718e-01 -1.76194340e-01 3.20753366e-01
2.38645807e-01 -1.07120430e+00 2.55175054e-01 7.32032359e-01
7.21440554e-01 -7.15834677e-01 9.54556942e-01 2.97029465e-01
-1.11429417e+00 2.62017936e-01 -6.38331532e-01 1.30835604e-02
8.13005343e-02 1.29051363e+00 -3.34886760e-01 7.35858798e-01
1.03149819e+00 6.28811657e-01 -5.61424732e-01 1.16933465e+00
-3.96653175e-01 1.03159213e+00 -5.00257492e-01 5.00834346e-01
8.31077322e-02 -4.99071181e-01 3.33829761e-01 1.32599998e+00
4.78832126e-01 4.35815185e-01 1.83010876e-01 5.30299902e-01
5.22968695e-02 -1.48036322e-02 -2.91992009e-01 3.04057777e-01
4.59682137e-01 1.50596905e+00 -5.09405553e-01 -3.76110077e-01
-5.01725197e-01 1.23715782e+00 1.13801062e-01 7.50750124e-01
-1.14822912e+00 -6.44191921e-01 9.14738417e-01 -2.20989361e-02
4.76495832e-01 -2.64905483e-01 -6.25037104e-02 -1.36331630e+00
1.48747459e-01 -1.00721657e+00 -6.33296445e-02 -8.07902753e-01
-1.13179052e+00 8.71550262e-01 -2.63431668e-01 -1.38972819e+00
4.09932315e-01 -3.42333853e-01 -8.47121775e-01 1.09653163e+00
-2.21058130e+00 -1.10853195e+00 -1.04308546e+00 6.00334823e-01
5.37210822e-01 3.16369653e-01 4.41983819e-01 5.81942618e-01
-7.48120427e-01 3.62248093e-01 4.27900821e-01 -1.79808527e-01
1.08169198e+00 -1.05242717e+00 4.72867280e-01 1.29713690e+00
-3.31746526e-02 3.27847511e-01 9.60186899e-01 -4.71264750e-01
-1.22204220e+00 -1.58308160e+00 6.73549056e-01 -1.04657114e-01
4.01314706e-01 -2.77439982e-01 -1.47363508e+00 4.67810035e-01
4.01429161e-02 4.49545115e-01 1.36497259e-01 -2.58918166e-01
-1.53525472e-01 -5.11425018e-01 -9.55742002e-01 6.02754891e-01
7.91516483e-01 -1.02075376e-01 -3.10350567e-01 4.46406215e-01
9.41422224e-01 -6.15543664e-01 -5.80377817e-01 7.04372466e-01
3.00445497e-01 -1.23279095e+00 8.63232374e-01 -1.18067458e-01
3.90531570e-01 -7.92138040e-01 -6.32752106e-02 -1.36698198e+00
-2.97389835e-01 -2.94459641e-01 -3.14169943e-01 1.11181390e+00
2.85804808e-01 -6.76646352e-01 5.98749578e-01 3.12773213e-02
-1.23658903e-01 -8.37261617e-01 -5.04220605e-01 -5.88894129e-01
-1.73693389e-01 -2.32360557e-01 4.92060184e-01 9.60861623e-01
-9.30052280e-01 9.03244019e-02 -7.95902312e-01 7.98284173e-01
7.28478491e-01 5.92729509e-01 8.93625379e-01 -9.99137223e-01
-2.13455394e-01 1.20385662e-01 8.29407424e-02 -1.25554430e+00
-1.37121379e-01 -2.97394276e-01 4.74304736e-01 -1.54519880e+00
1.92036375e-01 -4.88863200e-01 -2.77069390e-01 5.06082833e-01
-8.52143884e-01 3.47785026e-01 3.14966381e-01 5.07759571e-01
-4.63948429e-01 7.85952210e-01 1.51009512e+00 7.73489401e-02
-3.54013562e-01 1.49604127e-01 -4.92981017e-01 7.78768361e-01
1.13870382e+00 -5.21759629e-01 -2.52374411e-01 -5.57954431e-01
-1.58246487e-01 -1.41953975e-02 3.90706778e-01 -1.26345062e+00
-8.52978528e-02 -3.59862268e-01 2.76597947e-01 -5.33189118e-01
2.17128471e-01 -6.62969649e-01 6.05644239e-03 3.94705683e-01
2.65369177e-01 1.98666193e-02 2.74518937e-01 5.11556804e-01
-4.93078083e-01 8.28334317e-03 1.10124445e+00 -2.26950529e-03
-8.43622327e-01 6.10489428e-01 -2.50450402e-01 -1.75308898e-01
7.84259439e-01 -3.22613344e-02 -5.51961124e-01 -3.44909579e-01
-5.72832406e-01 3.46207768e-01 4.68443960e-01 6.68962076e-02
8.20116878e-01 -8.40646565e-01 -1.12502098e+00 2.57077515e-01
-2.90861093e-02 2.35044599e-01 7.11186767e-01 1.02046835e+00
-8.70104074e-01 -1.77235365e-01 -1.98856592e-02 -3.34288299e-01
-1.43071973e+00 3.61089826e-01 3.30390185e-01 -3.83588135e-01
-9.08153176e-01 6.42678559e-01 4.92497355e-01 -2.98120975e-01
-3.72393839e-02 -3.58428031e-01 -1.56778768e-01 -2.46290579e-01
7.48026073e-01 1.83216818e-02 4.84711304e-02 -2.45425850e-01
-6.98784068e-02 5.22791564e-01 -1.85071155e-01 2.42480636e-01
1.34252083e+00 -1.89499989e-01 -3.30443889e-01 2.22649917e-01
6.80005610e-01 1.07824527e-01 -1.54801214e+00 -3.16482097e-01
-5.37475467e-01 -6.50901318e-01 1.79046407e-01 -7.40157127e-01
-1.46659088e+00 8.80408049e-01 1.13364482e+00 -6.42063096e-02
1.58303154e+00 -3.93765092e-01 1.04074717e+00 4.51638877e-01
2.31159274e-02 -6.15650117e-01 -2.84353316e-01 6.46926820e-01
7.59978175e-01 -1.39116907e+00 3.07129472e-01 -5.65657616e-01
-5.74648619e-01 9.41076458e-01 7.54216909e-01 -2.36177251e-01
7.59892821e-01 3.81152540e-01 5.53463697e-01 -7.38251731e-02
-4.01161551e-01 -4.62437958e-01 -1.94994405e-01 3.04561347e-01
2.62905270e-01 4.66124900e-02 -3.56956273e-01 2.46317551e-01
2.32245550e-01 1.13984697e-01 4.93640304e-01 9.29917991e-01
-9.39482749e-01 -8.36035967e-01 -7.98039436e-01 5.50769627e-01
-3.26511145e-01 -5.90781510e-01 3.00103545e-01 7.16431558e-01
2.05846101e-01 9.30169225e-01 -5.71012124e-02 -2.01330215e-01
4.59022105e-01 -4.50505406e-01 1.51925847e-01 -3.84238303e-01
-3.57746243e-01 1.06354386e-01 -2.76427448e-01 -5.26672065e-01
-7.86451280e-01 -3.09602320e-01 -1.07956553e+00 -4.33704793e-01
-3.42746705e-01 1.61093011e-01 3.56052190e-01 8.56857717e-01
2.86719769e-01 5.51400840e-01 7.45387912e-01 -9.91776228e-01
-2.96950191e-01 -1.03779960e+00 -9.58429754e-01 2.77834684e-01
4.93191272e-01 -3.70429367e-01 -6.03481948e-01 2.07428068e-01]
|
[10.907706260681152, -3.2555227279663086]
|
d628b3b2-2bc5-4657-9f9a-db22879edb0f
|
a-comparative-study-on-speaker-attributed
|
2203.16834
| null |
https://arxiv.org/abs/2203.16834v3
|
https://arxiv.org/pdf/2203.16834v3.pdf
|
A Comparative Study on Speaker-attributed Automatic Speech Recognition in Multi-party Meetings
|
In this paper, we conduct a comparative study on speaker-attributed automatic speech recognition (SA-ASR) in the multi-party meeting scenario, a topic with increasing attention in meeting rich transcription. Specifically, three approaches are evaluated in this study. The first approach, FD-SOT, consists of a frame-level diarization model to identify speakers and a multi-talker ASR to recognize utterances. The speaker-attributed transcriptions are obtained by aligning the diarization results and recognized hypotheses. However, such an alignment strategy may suffer from erroneous timestamps due to the modular independence, severely hindering the model performance. Therefore, we propose the second approach, WD-SOT, to address alignment errors by introducing a word-level diarization model, which can get rid of such timestamp alignment dependency. To further mitigate the alignment issues, we propose the third approach, TS-ASR, which trains a target-speaker separation module and an ASR module jointly. By comparing various strategies for each SA-ASR approach, experimental results on a real meeting scenario corpus, AliMeeting, reveal that the WD-SOT approach achieves 10.7% relative reduction on averaged speaker-dependent character error rate (SD-CER), compared with the FD-SOT approach. In addition, the TS-ASR approach also outperforms the FD-SOT approach and brings 16.5% relative average SD-CER reduction.
|
['Lei Xie', 'Yuxiao Lin', 'Shiliang Zhang', 'Zhihao Du', 'Fan Yu']
|
2022-03-31
| null | null | null | null |
['speaker-separation']
|
['speech']
|
[ 3.53105634e-01 -1.11192852e-01 2.56983310e-01 -5.53246856e-01
-1.48967421e+00 -2.86599308e-01 5.71603477e-01 7.89578184e-02
-2.83292621e-01 3.97497803e-01 3.20138484e-01 -2.85902560e-01
1.26953155e-01 6.41444996e-02 -3.25637519e-01 -8.17727506e-01
5.50920181e-02 4.99033362e-01 2.07939580e-01 -2.25082815e-01
2.98609883e-01 9.13072079e-02 -1.38406801e+00 4.33843672e-01
1.03163815e+00 6.77805007e-01 3.94550532e-01 8.15522015e-01
-3.13552350e-01 5.43532908e-01 -1.12127328e+00 -2.09001645e-01
-2.55731158e-02 -5.54796040e-01 -6.25436842e-01 2.33558908e-01
7.78281689e-02 5.93316928e-02 -1.73224077e-01 8.17103088e-01
8.16501200e-01 1.76304400e-01 4.31381524e-01 -1.15921915e+00
-6.66681156e-02 9.89092290e-01 -7.47912109e-01 4.73965198e-01
5.64657569e-01 -3.16551000e-01 9.04632747e-01 -1.13822544e+00
5.25620654e-02 1.48873997e+00 3.53254467e-01 6.49395645e-01
-1.01387548e+00 -7.09940434e-01 2.86266327e-01 1.21611476e-01
-1.87408090e+00 -1.01818097e+00 6.70046449e-01 -3.27328891e-01
9.90411878e-01 7.30653882e-01 7.85023123e-02 8.73550057e-01
-9.05057341e-02 8.55907142e-01 9.19890761e-01 -6.47884786e-01
2.20114335e-01 1.20365120e-01 4.48702425e-01 9.72758830e-02
-9.95361060e-02 -2.78503090e-01 -1.05171311e+00 -2.05507576e-01
1.63261384e-01 -2.57943720e-01 -4.13490593e-01 5.44476509e-01
-1.25685012e+00 3.64161462e-01 -3.46424788e-01 6.65073276e-01
-4.43772852e-01 -3.66533965e-01 4.01535541e-01 2.23885179e-01
6.14004910e-01 -5.45837656e-02 -1.76929101e-01 -1.88848883e-01
-1.41490912e+00 5.66884689e-02 6.45164430e-01 1.15934801e+00
3.48679870e-01 3.77634853e-01 -4.14022148e-01 1.07411277e+00
4.89360183e-01 5.88232934e-01 8.09686482e-01 -4.01381493e-01
7.81108558e-01 1.82055101e-01 8.94230828e-02 -6.57035947e-01
-1.16392985e-01 -3.96634310e-01 -7.45105624e-01 -5.62228680e-01
1.71235397e-01 -1.21143676e-01 -1.00442064e+00 1.56052768e+00
2.10700482e-01 2.31960386e-01 4.74192262e-01 8.54817450e-01
7.89050579e-01 1.14210463e+00 -2.14792237e-01 -7.82019377e-01
1.57249403e+00 -1.09337842e+00 -1.24433506e+00 -2.98280507e-01
4.82523352e-01 -1.22172070e+00 7.46299326e-01 4.14255679e-01
-1.21100104e+00 -5.59088767e-01 -1.08691466e+00 4.90307510e-01
1.49530023e-01 3.96309674e-01 -2.58020431e-01 7.15716779e-01
-1.03313732e+00 -6.50571510e-02 -7.84259021e-01 -4.38946784e-01
-3.48841101e-01 2.45218083e-01 7.32786432e-02 1.46872237e-01
-1.27734590e+00 6.71114743e-01 2.80359775e-01 3.64390641e-01
-8.04441988e-01 -3.02964181e-01 -6.05658889e-01 1.17923126e-01
2.87753522e-01 4.80366014e-02 1.73421383e+00 -8.90114248e-01
-1.79863739e+00 5.84552050e-01 -9.75969315e-01 -5.11223316e-01
2.02015579e-01 -1.59291923e-01 -1.01342881e+00 -1.66572221e-02
-2.10605338e-02 1.36052147e-01 7.68663228e-01 -1.35773230e+00
-7.75237918e-01 -3.40416640e-01 -4.93865460e-01 4.28198338e-01
-3.48825008e-01 5.48361480e-01 -4.98727977e-01 -7.44888663e-01
4.88896310e-01 -9.56295907e-01 1.01168975e-02 -1.13135791e+00
-6.02526844e-01 -3.78594071e-01 6.36939585e-01 -9.89437521e-01
1.94764173e+00 -2.45890737e+00 3.39954421e-02 1.99369818e-01
-7.95085058e-02 4.21182990e-01 6.49055047e-03 5.87108552e-01
-1.74708083e-01 -6.17022775e-02 -2.81750083e-01 -7.60001719e-01
-1.38233081e-01 8.48288685e-02 -3.18406433e-01 2.54712194e-01
1.11166328e-01 2.40458012e-01 -5.32331347e-01 -6.21643424e-01
-6.99010193e-02 2.93987006e-01 -3.90284993e-02 6.01018846e-01
1.91843480e-01 3.29685450e-01 -1.99989676e-01 5.82316875e-01
7.59753942e-01 2.28638753e-01 4.58466709e-01 8.26197676e-03
-4.79194880e-01 7.54848301e-01 -1.38314104e+00 1.38072443e+00
-3.65537584e-01 5.41544378e-01 3.01391274e-01 -9.68046546e-01
1.26929414e+00 1.01057112e+00 2.20877230e-01 -5.58065832e-01
1.33481592e-01 5.24311960e-01 2.85251796e-01 -2.80493528e-01
7.52532005e-01 -1.26860231e-01 -1.11689523e-01 2.36357406e-01
-1.57995939e-01 2.44632319e-01 1.25810578e-01 4.32692081e-01
8.18862736e-01 -4.25230265e-01 2.27118596e-01 -2.21050560e-01
1.01538157e+00 -2.67229140e-01 8.38652849e-01 4.77695942e-01
-3.21174413e-01 7.87528515e-01 -5.73490672e-02 3.79111357e-02
-6.40800595e-01 -8.76248658e-01 5.99157549e-02 9.94948566e-01
8.13610554e-02 -7.19134092e-01 -9.37755108e-01 -3.50854129e-01
-4.03905272e-01 9.48573649e-01 -1.21467290e-02 -5.05358959e-03
-1.03209877e+00 -7.20454454e-01 7.76154697e-01 3.49577963e-01
2.70121187e-01 -7.11193204e-01 -5.79931866e-03 4.40078884e-01
-8.55977178e-01 -1.22484517e+00 -1.09366226e+00 1.88673630e-01
-5.09864509e-01 -3.18070650e-01 -8.68489921e-01 -8.13149869e-01
4.86501515e-01 7.23068058e-01 7.14219868e-01 -5.28423041e-02
2.44081885e-01 2.09282070e-01 -5.65409005e-01 -3.67169529e-01
-8.86960506e-01 9.30566564e-02 4.35141981e-01 6.23183966e-01
6.36388600e-01 -4.09537196e-01 -2.66731530e-01 6.24184430e-01
-6.09563351e-01 -1.13662124e-01 5.11965513e-01 6.28652513e-01
3.35701525e-01 -6.69150129e-02 8.52167547e-01 -3.99776280e-01
5.76153576e-01 -2.65852839e-01 -3.67976815e-01 4.48297024e-01
-5.12862325e-01 -6.70975670e-02 4.17578697e-01 -5.69180191e-01
-1.27686059e+00 -4.50082682e-02 -3.14883292e-01 -3.03679585e-01
-7.91363120e-02 5.78818440e-01 -5.37975252e-01 5.27307451e-01
3.02546173e-01 6.81061208e-01 -1.86919138e-01 -5.09999037e-01
-1.71514571e-01 1.54028165e+00 5.49427748e-01 -3.04538488e-01
7.36665189e-01 -1.38161018e-01 -8.53576541e-01 -1.22332585e+00
-6.98098600e-01 -9.30308521e-01 -3.67471159e-01 -2.72806764e-01
7.47991621e-01 -1.21380675e+00 -5.18726647e-01 5.81631243e-01
-1.51891506e+00 1.36907756e-01 2.84280181e-01 8.30182731e-01
-2.33998656e-01 3.94080967e-01 -5.94642162e-01 -1.52665484e+00
-5.36799312e-01 -1.18322432e+00 9.99285758e-01 2.13948563e-01
-4.53280389e-01 -4.75290388e-01 -5.91991618e-02 6.93215966e-01
1.81989267e-01 -4.41834480e-01 4.07338500e-01 -1.14589858e+00
-3.66902500e-01 2.37401128e-02 1.56300545e-01 3.68519306e-01
3.50050539e-01 -6.64523765e-02 -1.20970237e+00 -4.79299456e-01
2.98413396e-01 3.41510117e-01 4.61005360e-01 4.06065643e-01
6.85536206e-01 -3.66564155e-01 -3.28187644e-01 1.92330554e-02
9.16269600e-01 7.72929251e-01 6.36068344e-01 1.77274391e-01
5.55286288e-01 5.63935459e-01 9.57849026e-01 4.93520081e-01
5.79538405e-01 9.83926237e-01 -1.92206666e-01 1.01122752e-01
-8.65816250e-02 2.16166731e-02 9.14720416e-01 1.85343945e+00
5.54310568e-02 -7.16298997e-01 -9.52453434e-01 5.43381691e-01
-1.85683143e+00 -8.30252767e-01 -4.48057741e-01 2.44157791e+00
9.02337253e-01 2.24085897e-01 2.87527114e-01 4.63706732e-01
1.28933668e+00 1.89112648e-01 -1.43281013e-01 -5.15536666e-01
-2.34378412e-01 -2.14536816e-01 1.84052989e-01 6.04697466e-01
-6.75398588e-01 7.61401057e-01 5.60811615e+00 9.32944179e-01
-1.08592308e+00 2.32108265e-01 2.07485035e-01 2.49862242e-02
-9.42767859e-02 -5.56045212e-02 -1.17237139e+00 5.95163703e-01
1.52324855e+00 -4.96693492e-01 2.47352645e-01 4.02027339e-01
5.79043031e-01 -9.13136005e-02 -1.20575500e+00 1.19153893e+00
4.78655905e-01 -8.12498569e-01 2.78556440e-02 4.39417956e-04
2.12557927e-01 -3.60334486e-01 -3.01047981e-01 2.04889923e-01
7.87132904e-02 -6.15044713e-01 1.06777203e+00 1.50144413e-01
5.45451760e-01 -8.08280766e-01 8.13015521e-01 4.75044370e-01
-1.58334649e+00 1.08687736e-01 3.02692484e-02 1.92173660e-01
5.66467524e-01 5.90486765e-01 -1.10146940e+00 1.08551335e+00
5.71774125e-01 2.33573675e-01 -1.68468311e-01 8.62218440e-01
2.58826241e-02 1.07266152e+00 -1.61756113e-01 1.59994021e-01
-3.50892283e-02 2.38860101e-02 8.81897390e-01 1.57529342e+00
5.63386381e-01 1.67650491e-01 2.98562825e-01 2.59189427e-01
1.41760260e-01 1.79235563e-01 -2.08222065e-02 5.95053062e-02
1.01776659e+00 1.11589289e+00 -6.21603012e-01 -6.21209800e-01
-1.05992891e-01 1.04719794e+00 -4.80633900e-02 3.37311953e-01
-1.10584593e+00 -4.85766441e-01 5.63336730e-01 -4.41545919e-02
3.33031178e-01 -3.67885977e-01 1.65940318e-02 -1.02205098e+00
2.11866274e-01 -1.21029758e+00 2.74043500e-01 -4.82832313e-01
-9.40198362e-01 1.01502061e+00 -8.48711580e-02 -1.42244709e+00
-1.25923157e-01 1.84293419e-01 -6.80833101e-01 8.89100552e-01
-1.50108826e+00 -6.68685973e-01 -7.09598316e-05 3.68283689e-01
1.21150446e+00 -1.91245258e-01 6.72510624e-01 7.24097729e-01
-1.08086491e+00 1.02216673e+00 5.11544086e-02 -3.75604331e-02
1.05661559e+00 -9.65009868e-01 3.23480695e-01 1.03898823e+00
1.06527336e-01 6.73800945e-01 8.50432396e-01 -4.48902100e-01
-1.30810452e+00 -1.05644631e+00 1.32531738e+00 -1.45754129e-01
3.16834837e-01 -5.78456998e-01 -1.19722843e+00 6.45430624e-01
4.03101325e-01 -5.95848680e-01 8.44657063e-01 1.85668953e-02
-2.01326758e-01 -3.75087738e-01 -7.74103463e-01 5.06641328e-01
6.40648365e-01 -5.35240471e-01 -9.71923649e-01 6.45564198e-02
1.01408839e+00 -3.46485883e-01 -7.33875751e-01 9.36081931e-02
2.56989300e-01 -7.25826561e-01 5.92451572e-01 1.55341357e-01
-2.70231307e-01 -6.47991359e-01 -4.49463964e-01 -1.26201415e+00
-1.16263151e-01 -1.07145727e+00 3.03579979e-02 2.06346726e+00
5.51997542e-01 -5.60179710e-01 1.69861168e-01 3.83662224e-01
-5.73451877e-01 -1.00096762e-01 -1.13324654e+00 -1.14821172e+00
-5.00719786e-01 -3.17309558e-01 6.24930382e-01 8.02780330e-01
8.06542486e-02 5.93454421e-01 -5.58619916e-01 6.43097103e-01
4.97079372e-01 -1.03787184e-01 6.66637182e-01 -9.15429592e-01
-1.55580401e-01 -1.61866114e-01 6.81399405e-02 -1.26210725e+00
2.21921913e-02 -6.54630065e-01 6.97879195e-01 -1.30616438e+00
7.94984028e-03 -2.15621859e-01 -3.75196189e-01 1.33003280e-01
-3.29340905e-01 -2.56496370e-01 2.48403415e-01 3.80894810e-01
-5.48356950e-01 5.37962139e-01 5.65332055e-01 -1.52013019e-01
-5.97608566e-01 1.72402635e-01 -3.75190914e-01 5.45533478e-01
8.35352063e-01 -6.97052300e-01 -1.82342276e-01 -3.56703520e-01
-5.99701047e-01 4.38575566e-01 -2.42875382e-01 -1.00044107e+00
4.97209966e-01 -9.74595770e-02 -1.56801790e-01 -1.13398504e+00
4.37640995e-01 -4.86630499e-01 2.25279346e-01 4.07684922e-01
-2.64554620e-01 1.63970619e-01 1.33793980e-01 4.22831953e-01
-5.81921518e-01 -1.01665199e-01 6.24995887e-01 2.07890511e-01
-3.39848310e-01 -1.76691219e-01 -9.39694881e-01 -1.01947904e-01
8.13284159e-01 -4.34269339e-01 -8.66568312e-02 -2.90216297e-01
-4.32736307e-01 2.58423269e-01 -1.91499919e-01 4.77933943e-01
6.61991000e-01 -1.22551084e+00 -1.02415705e+00 1.82088882e-01
1.51002482e-01 -1.01026669e-01 5.01846135e-01 1.04906750e+00
9.28235650e-02 5.24772644e-01 5.47003269e-01 -7.81773865e-01
-1.81254756e+00 3.27527165e-01 6.21783435e-02 -2.54509836e-01
-3.07776660e-01 7.24299252e-01 2.81902462e-01 -1.56409875e-01
4.97429579e-01 -2.09427252e-01 -1.84372053e-01 1.00510739e-01
6.49697065e-01 3.93278182e-01 3.95107508e-01 -1.00132477e+00
-8.08491409e-01 2.98167676e-01 -3.34422469e-01 -5.53327978e-01
9.29257333e-01 -7.50633478e-01 -3.15165892e-02 8.22975934e-01
8.84333193e-01 3.22011769e-01 -6.70785129e-01 -3.50397587e-01
4.39509183e-01 -2.19700798e-01 -6.31171763e-02 -5.60649931e-01
-5.78832924e-01 7.59755254e-01 3.84249657e-01 4.35596019e-01
1.11556709e+00 -6.27834424e-02 1.02255559e+00 1.30573809e-01
2.18992040e-01 -1.07727861e+00 9.88031402e-02 6.65938079e-01
9.51753259e-01 -9.60492432e-01 -4.16582108e-01 -4.89509106e-01
-8.42936695e-01 9.16435003e-01 7.02617884e-01 5.20040512e-01
4.25350249e-01 2.78813392e-01 3.93758953e-01 2.55695879e-01
-8.27969611e-01 -8.29531997e-02 2.60307789e-01 1.89093083e-01
7.51018524e-01 2.24769622e-01 -4.71079826e-01 8.43967259e-01
-3.19718093e-01 -3.66596639e-01 5.64845502e-01 9.43004966e-01
-5.62820256e-01 -1.26037896e+00 -7.80422449e-01 -1.81214571e-01
-3.90558183e-01 -2.08033696e-01 -2.86436707e-01 2.25017846e-01
-3.77473205e-01 1.63126791e+00 7.20948055e-02 -6.76959872e-01
5.09595573e-01 2.97915757e-01 3.92460003e-02 -6.57177746e-01
-7.71906614e-01 8.33967745e-01 3.81568849e-01 -5.82455136e-02
-5.10686398e-01 -7.92211294e-01 -1.35773790e+00 -2.40301624e-01
-7.94957697e-01 7.23655999e-01 7.41849720e-01 9.47796404e-01
4.08259124e-01 8.46006513e-01 1.12569869e+00 -6.45722151e-01
-6.93837404e-01 -1.24061048e+00 -6.17316067e-01 -1.95649210e-02
4.57914710e-01 -2.85556018e-01 -5.75095057e-01 1.26030028e-01]
|
[14.625097274780273, 6.2413153648376465]
|
0a232953-3f44-48eb-9c92-32cb00e8b74d
|
neural-transformation-fields-for-arbitrary
| null | null |
http://openaccess.thecvf.com//content/CVPR2023/html/Fu_Neural_Transformation_Fields_for_Arbitrary-Styled_Font_Generation_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Fu_Neural_Transformation_Fields_for_Arbitrary-Styled_Font_Generation_CVPR_2023_paper.pdf
|
Neural Transformation Fields for Arbitrary-Styled Font Generation
|
Few-shot font generation (FFG), aiming at generating font images with a few samples, is an emerging topic in recent years due to the academic and commercial values. Typically, the FFG approaches follow the style-content disentanglement paradigm, which transfers the target font styles to characters by combining the content representations of source characters and the style codes of reference samples. Most existing methods attempt to increase font generation ability via exploring powerful style representations, which may be a sub-optimal solution for the FFG task due to the lack of modeling spatial transformation in transferring font styles. In this paper, we model font generation as a continuous transformation process from the source character image to the target font image via the creation and dissipation of font pixels, and embed the corresponding transformations into a neural transformation field. With the estimated transformation path, the neural transformation field generates a set of intermediate transformation results via the sampling process, and a font rendering formula is developed to accumulate them into the target font image. Extensive experiments show that our method achieves state-of-the-art performance on few-shot font generation task, which demonstrates the effectiveness of our proposed model. Our implementation is available at: https://github.com/fubinfb/NTF.
|
['Yu Qiao', 'Jianjun Wang', 'Junjun He', 'Bin Fu']
|
2023-01-01
| null | null | null |
cvpr-2023-1
|
['disentanglement']
|
['methodology']
|
[ 3.81216317e-01 -2.25024119e-01 9.64269936e-02 -2.38771915e-01
-5.19760549e-01 -5.98676920e-01 7.81100512e-01 -4.53543037e-01
2.41337195e-01 5.45391977e-01 3.40619802e-01 -6.79707304e-02
3.38389009e-01 -1.00750709e+00 -7.58861363e-01 -7.11976469e-01
8.61795843e-01 9.68278423e-02 4.85832877e-02 -4.72530901e-01
3.71882230e-01 2.12931290e-01 -1.39826107e+00 3.15865636e-01
1.28745258e+00 9.67853487e-01 3.67999494e-01 8.29711616e-01
-6.94424152e-01 7.26835847e-01 -8.48374963e-01 -5.72487712e-01
3.30403566e-01 -7.28995502e-01 -3.10670942e-01 2.72190452e-01
4.46735889e-01 -6.25600755e-01 -2.53976136e-01 1.20959997e+00
6.01480067e-01 5.36135025e-02 8.16307127e-01 -1.03666580e+00
-1.60260224e+00 4.08910990e-01 -6.87378049e-01 -2.67401427e-01
2.44124264e-01 4.48758185e-01 6.60183549e-01 -1.15141726e+00
6.19064867e-01 1.35840845e+00 2.18943581e-01 8.00162971e-01
-1.20173979e+00 -7.27581322e-01 2.67458290e-01 -1.42136410e-01
-9.83263910e-01 -2.19017625e-01 1.08570802e+00 -5.72570682e-01
3.48445863e-01 3.15920889e-01 8.24550092e-01 1.18372655e+00
2.88917087e-02 8.09216917e-01 1.21878541e+00 -5.19686699e-01
2.18315586e-01 2.41637826e-01 -4.11081277e-02 4.27006692e-01
-8.38241130e-02 1.60335883e-01 -5.30696392e-01 1.50825948e-01
1.18415737e+00 -1.43977746e-01 -3.21434587e-01 -1.85581923e-01
-1.06078744e+00 6.29747272e-01 4.02613789e-01 4.71192747e-02
-1.45749807e-01 -2.80922651e-02 3.16419184e-01 2.59940565e-01
6.41642749e-01 3.41210932e-01 9.25990045e-02 -2.38086984e-01
-7.98082530e-01 3.07043761e-01 4.90873754e-01 1.35273361e+00
5.50690770e-01 2.81875581e-01 -6.86515510e-01 1.02243698e+00
1.08892240e-01 6.06388330e-01 4.19276595e-01 -5.91737866e-01
6.47162318e-01 5.81355572e-01 2.03442022e-01 -7.22948730e-01
4.02379960e-01 -3.11990064e-02 -9.63353992e-01 4.60977048e-01
3.92466903e-01 -2.24341944e-01 -7.43549943e-01 1.60962403e+00
2.59948194e-01 7.14763105e-02 -1.07942574e-01 8.59920025e-01
6.73744500e-01 1.04523945e+00 -1.03819877e-01 -7.83600882e-02
1.25496876e+00 -1.09586966e+00 -8.72923553e-01 -1.42496184e-01
2.63429105e-01 -1.13261890e+00 1.85702443e+00 3.50924790e-01
-1.33954287e+00 -1.01828694e+00 -1.04636371e+00 -2.98369706e-01
-4.51280892e-01 4.33668315e-01 4.08433199e-01 6.67706788e-01
-7.79043674e-01 4.91336465e-01 -2.69569188e-01 -2.08343759e-01
5.38480759e-01 -1.55909449e-01 3.29948813e-01 3.34075987e-02
-1.07506788e+00 5.99822998e-01 1.37114078e-01 9.99531597e-02
-4.66653466e-01 -7.79446185e-01 -6.26575828e-01 7.77158216e-02
1.61179349e-01 -5.55072248e-01 1.10256791e+00 -1.14328110e+00
-1.88989162e+00 4.81419832e-01 -2.18985379e-01 1.02476224e-01
8.03484023e-01 -2.16954589e-01 -4.27819133e-01 -2.78937846e-01
-4.09770720e-02 8.75054955e-01 1.12969255e+00 -1.26474750e+00
-4.37519908e-01 3.46233919e-02 -8.62724334e-03 3.96993548e-01
-6.78929090e-01 -4.88281921e-02 -4.24904794e-01 -1.12129128e+00
-3.24806601e-01 -6.44736826e-01 1.86479896e-01 4.60258216e-01
-4.43834126e-01 -7.74144307e-02 9.37382221e-01 -8.37663651e-01
1.23861516e+00 -2.31367421e+00 3.95384192e-01 -1.13323644e-01
1.29436433e-01 2.12996244e-01 -2.86439091e-01 3.55879009e-01
-9.26790386e-02 1.73463728e-02 -2.25894347e-01 -2.69614637e-01
-9.11336392e-03 -1.48424909e-01 -8.07703912e-01 -1.44620061e-01
2.10033208e-01 1.09625006e+00 -7.58357882e-01 -3.24968368e-01
3.93488944e-01 4.53066826e-01 -2.25463852e-01 3.32105190e-01
-4.88375962e-01 4.40183759e-01 -3.10464978e-01 5.10103822e-01
9.15087640e-01 -1.08508125e-01 -1.37932166e-01 -1.37997940e-01
-1.42292544e-01 -2.03800663e-01 -9.33255255e-01 1.82944238e+00
-5.12731016e-01 7.31818616e-01 -3.81740957e-01 -1.75023422e-01
1.31618822e+00 1.27179816e-01 -4.75458242e-03 -5.88024914e-01
3.35889786e-01 1.29745245e-01 -1.37139931e-01 -2.95113206e-01
7.19284475e-01 4.33853157e-02 3.65062132e-02 6.00176156e-01
-2.28569075e-01 -5.01032412e-01 4.18076932e-01 1.40893042e-01
3.37362558e-01 5.81308722e-01 2.30433550e-02 -1.54325411e-01
5.61673045e-01 -2.83382177e-01 1.76315963e-01 3.73927981e-01
9.55710635e-02 7.68690526e-01 5.96550941e-01 -3.60196322e-01
-1.54709888e+00 -1.29606450e+00 1.60384148e-01 7.11101115e-01
9.91186202e-02 -2.54368544e-01 -1.13938200e+00 -3.26142639e-01
-2.86896527e-01 9.97068226e-01 -5.97062945e-01 -2.78836042e-01
-6.75833464e-01 -4.20534790e-01 2.05351800e-01 6.02618515e-01
7.32637882e-01 -1.13857019e+00 -2.92776316e-01 3.13474424e-02
-1.10390104e-01 -7.85538435e-01 -1.03604066e+00 -3.70264053e-01
-8.13160837e-01 -6.45735204e-01 -1.39362681e+00 -9.38810110e-01
1.01359677e+00 2.66164362e-01 7.53093719e-01 -1.84056923e-01
-2.44814381e-01 -5.93411662e-02 -2.63767362e-01 -3.91272366e-01
-7.02351868e-01 -6.47126734e-02 -3.22598457e-01 1.35740504e-01
2.78192442e-02 -4.06404227e-01 -6.17674589e-01 1.37210831e-01
-9.57267225e-01 7.11640358e-01 4.21695322e-01 8.22703123e-01
6.20105147e-01 -3.06030631e-01 3.07024717e-01 -7.89587617e-01
1.20362377e+00 -7.61362836e-02 -7.43979871e-01 3.91340911e-01
-4.34414983e-01 8.20368528e-02 1.08627450e+00 -8.07156801e-01
-1.39688098e+00 -1.12725012e-01 2.05546543e-01 -6.64686501e-01
-4.66824844e-02 -1.15912810e-01 -4.50758159e-01 1.59096926e-01
6.17527843e-01 5.66125035e-01 8.18758532e-02 -4.75430518e-01
8.51158619e-01 7.41803348e-01 4.30204272e-01 -8.57213318e-01
1.05731034e+00 3.59777749e-01 -2.49261558e-01 -5.12850881e-01
-5.74024439e-01 3.73135090e-01 -6.25138938e-01 -5.01491666e-01
9.30775225e-01 -6.95801735e-01 -3.54158282e-01 7.98348069e-01
-1.29424214e+00 -4.27984416e-01 -7.54971623e-01 -6.54500350e-02
-6.71684623e-01 3.35702628e-01 -5.36858797e-01 -6.50315583e-01
-4.08181280e-01 -1.25631833e+00 1.19922268e+00 5.05222678e-01
-1.51531011e-01 -7.80177355e-01 -4.51498888e-02 -6.72633275e-02
4.04995143e-01 2.39779696e-01 1.33616006e+00 2.65283793e-01
-7.46566832e-01 -1.18314391e-02 -4.24934804e-01 4.37254965e-01
5.81630826e-01 1.29622608e-01 -9.98307645e-01 -1.24798194e-01
-1.16322994e-01 -2.20212132e-01 5.70534408e-01 3.19927603e-01
1.34844410e+00 -2.07235873e-01 2.03580841e-01 8.36244881e-01
1.33910537e+00 4.79695708e-01 7.64843881e-01 6.98371604e-02
9.71085131e-01 4.40647393e-01 4.71699387e-01 5.11385798e-01
1.17252499e-01 4.74821329e-01 3.65556963e-02 -1.43619105e-01
-6.11899137e-01 -5.97165942e-01 2.18531504e-01 9.51999426e-01
-9.76574048e-02 -2.52231479e-01 -4.15364891e-01 9.02049616e-02
-1.66238439e+00 -9.59763408e-01 1.43864835e-02 2.28155303e+00
9.74500060e-01 2.06363991e-01 7.89108053e-02 -1.14653394e-01
9.21522200e-01 4.33706254e-01 -9.32523966e-01 -7.05240369e-01
-1.80359855e-01 -2.45393422e-02 7.98842832e-02 2.57657766e-01
-5.99035740e-01 1.16769266e+00 5.84893465e+00 9.12795961e-01
-1.12431657e+00 -1.83363914e-01 9.83984232e-01 -1.86747670e-01
-4.03543651e-01 1.21292872e-02 -7.59086430e-01 8.59485388e-01
5.40495515e-01 -4.83931273e-01 9.59848046e-01 6.95540190e-01
3.87244821e-01 2.48588711e-01 -9.61989462e-01 1.07646537e+00
2.31286764e-01 -1.28188741e+00 4.96644169e-01 -2.04281379e-02
9.12586391e-01 -9.37300742e-01 4.45363551e-01 1.38847813e-01
2.25735262e-01 -8.21370065e-01 1.02612042e+00 8.70767891e-01
1.42748141e+00 -5.73596120e-01 -5.48723638e-02 1.87594488e-01
-1.42960346e+00 -3.20187164e-03 -6.20756149e-01 -8.15931931e-02
2.68128991e-01 3.38660270e-01 -4.23821121e-01 4.51168418e-01
3.13898176e-01 7.61183202e-01 -7.55438626e-01 7.20013618e-01
-3.00222248e-01 3.07416052e-01 1.79769784e-01 -2.58752823e-01
-3.33139896e-02 -6.48972929e-01 3.76542509e-01 6.30728900e-01
6.76731229e-01 -1.95077002e-01 -1.92244411e-01 1.57803440e+00
-1.99533045e-01 2.11626425e-01 -4.39428687e-01 -2.40732357e-01
4.69853312e-01 1.19316328e+00 -6.69356346e-01 -6.80765867e-01
-4.60689813e-01 1.29578733e+00 2.27079675e-01 5.30041933e-01
-1.04368281e+00 -7.16482282e-01 5.80239356e-01 -5.24905547e-02
1.11218408e-01 -1.10164285e-02 -7.77263165e-01 -1.24375641e+00
3.85614783e-01 -1.01858342e+00 -1.37956470e-01 -1.33995962e+00
-1.22366774e+00 6.49756074e-01 -8.11113939e-02 -1.79519641e+00
-5.22132851e-02 -6.41303480e-01 -1.12863922e+00 1.36507475e+00
-1.07627463e+00 -1.22718835e+00 -7.28488266e-01 3.50084186e-01
1.13391006e+00 -3.51448238e-01 5.89629292e-01 -3.68620194e-02
-5.61158419e-01 6.40215397e-01 1.88146949e-01 2.78404672e-02
7.88241982e-01 -1.23120117e+00 1.06084824e+00 9.33043718e-01
-2.30690718e-01 5.25833070e-01 5.27392924e-01 -8.17582548e-01
-1.18367946e+00 -1.05180454e+00 5.03681362e-01 -2.25841805e-01
4.81010467e-01 -7.46952534e-01 -8.65708292e-01 3.00183415e-01
4.72611845e-01 -3.54697078e-01 3.09195966e-01 -4.33189332e-01
-5.04914463e-01 -6.20913468e-02 -7.99789906e-01 1.20252538e+00
1.08112121e+00 -3.78237247e-01 -3.89901459e-01 2.81221092e-01
9.32227015e-01 -5.38339913e-01 -6.83853686e-01 -2.73727000e-01
5.75945318e-01 -8.86036098e-01 7.79995561e-01 -2.30881318e-01
9.30016160e-01 -3.18834215e-01 1.62536561e-01 -1.54482031e+00
-3.55334699e-01 -6.59378469e-01 -1.17097683e-01 1.57035482e+00
2.04019204e-01 -5.06292224e-01 6.46328092e-01 6.24628544e-01
-2.42638160e-02 -7.96412468e-01 -3.63905311e-01 -7.47050762e-01
4.54762280e-01 3.31469253e-02 1.26326132e+00 6.56216025e-01
-3.95707339e-01 1.67961091e-01 -6.10809088e-01 -3.81326884e-01
5.62421679e-01 3.15858424e-01 8.08571219e-01 -9.59053218e-01
-4.38014179e-01 -7.05520689e-01 2.25529730e-01 -1.11941063e+00
-3.18530910e-02 -7.86085546e-01 -1.18212305e-01 -1.71374333e+00
1.30272090e-01 -5.32275252e-02 1.98969141e-01 1.03375822e-01
-4.78567928e-01 -1.00500748e-01 5.81795633e-01 3.69140059e-01
9.26632583e-02 8.51281941e-01 2.05846453e+00 -2.52990246e-01
-3.26777101e-01 -1.88784018e-01 -8.58839273e-01 3.47763330e-01
8.72113049e-01 8.81151557e-02 -7.09588528e-01 -5.63408732e-01
1.04804568e-01 -1.15548261e-02 1.13905542e-01 -8.63983750e-01
-1.41589165e-01 -1.26351357e-01 7.64831722e-01 -4.70303386e-01
4.60517317e-01 -4.84232485e-01 -4.06172462e-02 3.56327981e-01
-5.19196749e-01 1.82143331e-01 7.96495602e-02 3.58909398e-01
-8.01463984e-03 -2.24322513e-01 8.75477254e-01 -1.06761605e-01
-5.11909902e-01 3.76409173e-01 -9.39613432e-02 -7.57345324e-03
8.86949480e-01 -6.03238225e-01 -6.26342535e-01 -2.54634976e-01
-3.46370250e-01 -9.00597349e-02 6.43175185e-01 7.59201705e-01
7.30219066e-01 -1.65307593e+00 -6.24216020e-01 4.62864190e-01
9.98032242e-02 -2.02798024e-02 3.88870180e-01 2.97629148e-01
-5.03223062e-01 1.26638487e-02 -4.30570066e-01 -2.68685192e-01
-1.07447743e+00 7.63823152e-01 1.78699434e-01 3.86875905e-02
-7.50484645e-01 7.27401078e-01 5.80431879e-01 -2.35717267e-01
3.64119709e-02 -3.83235008e-01 -4.47049327e-02 -1.17838383e-01
7.05175996e-01 4.37554866e-01 -3.91092211e-01 -3.59171361e-01
4.23267901e-01 9.10473168e-01 -2.79360592e-01 -2.39688873e-01
8.48220170e-01 -3.69685963e-02 5.66690341e-02 5.14782250e-01
1.01020062e+00 -1.10067062e-01 -1.68569458e+00 -5.97006381e-02
-5.11220515e-01 -8.25621367e-01 -3.62377465e-01 -7.98848212e-01
-9.58358943e-01 1.06256592e+00 7.20873535e-01 -8.49191546e-02
1.32660127e+00 -3.28259349e-01 9.73464370e-01 -6.62653148e-02
1.65781900e-01 -1.17070246e+00 2.72983670e-01 3.40384871e-01
1.03547394e+00 -8.97643983e-01 -2.51444608e-01 -6.02152288e-01
-8.31905484e-01 1.30910742e+00 9.29622233e-01 -2.96406299e-01
2.04853788e-01 2.98393220e-01 2.01438695e-01 2.46693283e-01
-5.84375799e-01 2.20174536e-01 1.53547049e-01 4.98596966e-01
4.70438868e-01 9.48898643e-02 -2.96005160e-01 4.76943851e-01
-4.39661413e-01 7.46392757e-02 5.83296955e-01 7.04407573e-01
-2.72652060e-01 -1.35258567e+00 -4.69528139e-01 5.05696177e-01
4.10393089e-01 -3.10857683e-01 -5.37908912e-01 3.64727557e-01
1.29731610e-01 6.14831448e-01 1.28914773e-01 -3.58443022e-01
4.50527966e-01 1.53645799e-01 5.16166985e-01 -4.40683961e-01
-5.27288675e-01 3.12958747e-01 -3.90018344e-01 -3.41933370e-01
1.86895132e-01 -6.23111367e-01 -8.15962017e-01 -3.24170530e-01
-1.71060205e-01 -1.65339753e-01 5.10800004e-01 4.81152743e-01
3.94371331e-01 9.51397479e-01 8.67557764e-01 -8.41955543e-01
-7.03584611e-01 -9.32707310e-01 -5.47324061e-01 5.24159610e-01
-1.55722335e-01 -3.92659098e-01 -1.16418608e-01 3.26338351e-01]
|
[11.635649681091309, -0.38205671310424805]
|
3d25569c-c2bb-4a99-bb84-7f85e3355a37
|
learning-to-recommend-frame-for-interactive
|
2103.10391
| null |
https://arxiv.org/abs/2103.10391v2
|
https://arxiv.org/pdf/2103.10391v2.pdf
|
Learning to Recommend Frame for Interactive Video Object Segmentation in the Wild
|
This paper proposes a framework for the interactive video object segmentation (VOS) in the wild where users can choose some frames for annotations iteratively. Then, based on the user annotations, a segmentation algorithm refines the masks. The previous interactive VOS paradigm selects the frame with some worst evaluation metric, and the ground truth is required for calculating the evaluation metric, which is impractical in the testing phase. In contrast, in this paper, we advocate that the frame with the worst evaluation metric may not be exactly the most valuable frame that leads to the most performance improvement across the video. Thus, we formulate the frame selection problem in the interactive VOS as a Markov Decision Process, where an agent is learned to recommend the frame under a deep reinforcement learning framework. The learned agent can automatically determine the most valuable frame, making the interactive setting more practical in the wild. Experimental results on the public datasets show the effectiveness of our learned agent without any changes to the underlying VOS algorithms. Our data, code, and models are available at https://github.com/svip-lab/IVOS-W.
|
['Shenghua Gao', 'Hanling Zhang', 'Shenhan Qian', 'Weixin Luo', 'Jia Zheng', 'Zhaoyuan Yin']
|
2021-03-18
| null |
http://openaccess.thecvf.com//content/CVPR2021/html/Yin_Learning_To_Recommend_Frame_for_Interactive_Video_Object_Segmentation_in_CVPR_2021_paper.html
|
http://openaccess.thecvf.com//content/CVPR2021/papers/Yin_Learning_To_Recommend_Frame_for_Interactive_Video_Object_Segmentation_in_CVPR_2021_paper.pdf
|
cvpr-2021-1
|
['interactive-video-object-segmentation']
|
['computer-vision']
|
[ 4.99680005e-02 1.19955927e-01 -3.56500655e-01 -3.99376929e-01
-7.98370779e-01 -6.25979185e-01 1.81453675e-02 -1.44099295e-01
-5.86274922e-01 5.30117750e-01 -1.13686204e-01 -1.97506830e-01
1.13241836e-01 -6.32489085e-01 -8.00682008e-01 -9.03615713e-01
6.36637062e-02 6.20103955e-01 6.25935555e-01 1.01145640e-01
1.57237619e-01 8.59814435e-02 -1.41845644e+00 1.72347695e-01
6.76436722e-01 1.20818853e+00 5.48452735e-01 8.31251204e-01
2.49127045e-01 8.35097671e-01 -5.93467772e-01 -3.37997496e-01
5.70292294e-01 -5.87753892e-01 -1.05147469e+00 4.65501845e-01
4.53193873e-01 -8.07196140e-01 -1.27191961e-01 1.02631378e+00
2.96107948e-01 4.17851299e-01 3.15043211e-01 -1.50337791e+00
-2.55836695e-02 8.81768405e-01 -5.07203579e-01 1.88882500e-01
3.92749757e-01 4.90650743e-01 1.30435014e+00 -6.54394925e-01
9.07608449e-01 9.94012058e-01 1.97525427e-01 6.74277067e-01
-9.12632763e-01 -4.14346099e-01 4.22851831e-01 4.23978657e-01
-1.34764278e+00 -3.66789281e-01 6.81608438e-01 -4.45546716e-01
4.06105846e-01 1.70091301e-01 9.61366057e-01 8.19636881e-01
-2.89346665e-01 1.35723472e+00 5.69307089e-01 -2.95711964e-01
4.39223200e-01 -6.51485845e-02 -4.16966528e-02 7.74995208e-01
-1.58609778e-01 -1.22879297e-01 -5.61662376e-01 3.11369225e-02
6.39833868e-01 7.95656815e-03 -3.22486073e-01 -4.07482684e-01
-1.03236806e+00 7.62419641e-01 1.97726086e-01 7.63235688e-02
-4.69951689e-01 3.85540962e-01 2.96009064e-01 2.58531898e-01
2.31708735e-01 2.31613487e-01 -5.47064781e-01 -3.33269596e-01
-1.22991705e+00 5.29736400e-01 7.44414985e-01 8.07416260e-01
7.98923314e-01 -2.63719976e-01 -3.06764185e-01 6.05237842e-01
4.68112379e-01 2.66716421e-01 -1.22417249e-01 -1.72428977e+00
2.92802662e-01 3.05185705e-01 4.33922976e-01 -6.05976284e-01
-7.92379826e-02 -7.41512775e-02 -2.16739014e-01 3.74334872e-01
6.90315306e-01 -4.80835110e-01 -7.63710678e-01 1.60421991e+00
6.34015262e-01 5.61311483e-01 -9.05753523e-02 1.32929373e+00
5.33004582e-01 6.08921707e-01 1.07921258e-01 -2.54933238e-01
1.23761928e+00 -1.14641547e+00 -4.81952846e-01 -7.89311007e-02
4.42683995e-01 -6.23930514e-01 1.07116950e+00 4.02937829e-01
-1.18834913e+00 -4.31013525e-01 -8.98740053e-01 2.47054964e-01
2.62869716e-01 -1.85268205e-02 3.88182670e-01 3.33158195e-01
-1.20299888e+00 6.23361170e-01 -9.75872159e-01 -1.91534206e-01
4.97464836e-01 2.88331389e-01 1.87044531e-01 8.97493437e-02
-1.09902847e+00 3.25690210e-01 4.46059823e-01 3.22811380e-02
-1.33656025e+00 -5.10481536e-01 -5.00988662e-01 -2.29496118e-02
1.11828995e+00 -6.11956716e-01 1.68183291e+00 -1.59649289e+00
-1.51253176e+00 7.11496294e-01 -1.54126883e-01 -7.33994424e-01
9.59948003e-01 -3.00200313e-01 1.47574782e-01 4.93998230e-01
2.62480587e-01 1.01986516e+00 9.64580834e-01 -1.31788242e+00
-1.17468345e+00 5.67376353e-02 7.77822256e-01 3.38603497e-01
1.01028979e-01 1.45352520e-02 -1.13453257e+00 -2.93315142e-01
-1.80586323e-01 -1.20305371e+00 -2.64431149e-01 -9.62382555e-02
-4.76548195e-01 -4.45619702e-01 6.89071298e-01 -5.99316418e-01
1.15702999e+00 -2.22865391e+00 1.10095575e-01 3.03123951e-01
1.75307244e-01 9.85027850e-02 -3.60689163e-02 3.43462825e-02
4.25416142e-01 2.17993006e-01 -1.88596711e-01 -3.15152764e-01
-1.92643657e-01 2.04819903e-01 -3.05376574e-02 3.85855436e-01
-1.34448791e-02 6.46596611e-01 -1.02906859e+00 -7.76653886e-01
1.09370813e-01 1.26142621e-01 -9.25008595e-01 6.17433548e-01
-7.31503248e-01 7.55894363e-01 -7.19092906e-01 5.41591763e-01
3.63024682e-01 -4.18406129e-01 2.34086603e-01 -4.45296615e-02
-2.04220768e-02 7.64229447e-02 -1.47269726e+00 1.62919140e+00
-9.02575925e-02 5.49606681e-01 1.03464544e-01 -8.20954382e-01
4.25355226e-01 2.04647824e-01 7.84597754e-01 -4.47376072e-01
9.56941098e-02 -5.19924350e-02 -1.56193599e-03 -7.53214478e-01
5.40587306e-01 5.30229032e-01 2.63810605e-01 4.33763623e-01
-8.12367816e-03 2.20693171e-01 6.29001737e-01 3.91335636e-01
1.09043860e+00 5.72786927e-01 2.90779304e-03 -6.93709254e-02
3.31836551e-01 5.42306202e-03 9.41401064e-01 9.76362765e-01
-4.27707672e-01 6.36030793e-01 7.14728594e-01 -5.03744423e-01
-8.57608557e-01 -7.74938166e-01 3.14018935e-01 1.29221559e+00
7.00848758e-01 -3.80652308e-01 -1.20322502e+00 -8.83382320e-01
-3.61601233e-01 5.52931368e-01 -4.29514170e-01 3.46996665e-01
-6.00839734e-01 -1.38936818e-01 2.64515519e-01 2.10523605e-01
5.86748362e-01 -1.39547622e+00 -1.02517092e+00 1.38794437e-01
-5.73264182e-01 -1.18623340e+00 -8.52185786e-01 -1.97321624e-01
-6.10345602e-01 -1.40107799e+00 -5.59440434e-01 -6.70328259e-01
5.99968135e-01 2.88111418e-01 1.05485523e+00 5.56613028e-01
1.83364332e-01 4.41355616e-01 -5.96775472e-01 -6.74049556e-02
-4.58333641e-01 1.62333131e-01 -4.08554554e-01 2.33754277e-01
-5.21567948e-02 -1.71803966e-01 -1.02105749e+00 4.11008298e-01
-7.89390564e-01 3.87530088e-01 1.65423274e-01 5.98204613e-01
9.60328758e-01 8.17229040e-03 2.86305457e-01 -1.13433838e+00
8.10552984e-02 -4.12186980e-01 -8.44848335e-01 1.99959442e-01
-4.02464777e-01 -7.88305998e-02 4.02371824e-01 -3.27287465e-01
-9.13132787e-01 2.63667911e-01 -9.69396979e-02 -5.84487379e-01
-2.79819965e-01 2.61862278e-01 -3.10223550e-01 2.03010023e-01
3.54839504e-01 -1.19921707e-01 -3.43475133e-01 -2.20329449e-01
1.70105949e-01 3.41552675e-01 2.26059705e-01 -6.27524197e-01
5.80897033e-01 3.63620788e-01 -5.21895349e-01 -5.61684728e-01
-9.81394291e-01 -5.07729292e-01 -5.59613168e-01 -6.72854125e-01
1.07346368e+00 -8.12761307e-01 -8.75275254e-01 1.41455755e-01
-1.12539232e+00 -9.33401346e-01 -3.42081130e-01 3.19831043e-01
-8.73512328e-01 2.23874554e-01 -4.67437595e-01 -8.89589846e-01
-9.44008902e-02 -1.64994180e+00 1.01970315e+00 5.71186364e-01
-2.54564106e-01 -5.90649068e-01 -2.33766541e-01 5.80177903e-01
-1.21686876e-01 2.41450116e-01 4.20732588e-01 -5.18629074e-01
-1.11666358e+00 2.36636147e-01 1.15538172e-01 2.02988952e-01
-2.07289785e-01 3.07905346e-01 -8.39466453e-01 -9.88243371e-02
-3.86383235e-01 -2.70311803e-01 7.51719236e-01 6.39184475e-01
1.50868750e+00 -2.84153253e-01 -8.62655267e-02 7.68978298e-01
1.23633802e+00 5.26504517e-01 4.84158635e-01 4.47763979e-01
7.57317603e-01 2.76717365e-01 1.14410162e+00 5.00298083e-01
4.61149126e-01 8.14833283e-01 6.68385625e-01 2.51386482e-02
-3.15051749e-02 -1.28844246e-01 6.81847274e-01 2.53970891e-01
-2.93730259e-01 -5.31426013e-01 -6.87422335e-01 4.40125287e-01
-2.19405699e+00 -1.06540287e+00 3.58388036e-01 2.10985851e+00
8.39065254e-01 2.39548311e-01 4.45646435e-01 -1.27243668e-01
8.57715130e-01 3.14041406e-01 -8.15232038e-01 -1.34742558e-01
1.61279947e-01 -2.54727095e-01 4.49700564e-01 6.28948152e-01
-1.22311163e+00 1.15877998e+00 5.01401234e+00 7.49202967e-01
-9.58364189e-01 1.67092726e-01 1.06819046e+00 -2.18673497e-01
-1.32708147e-01 2.21080273e-01 -8.10712576e-01 5.59341371e-01
7.29184031e-01 -1.43589722e-02 8.23174834e-01 7.90716529e-01
7.31243074e-01 -4.06599671e-01 -1.28477216e+00 8.47677648e-01
-1.37462810e-01 -1.37986612e+00 -1.35984167e-01 -1.17447622e-01
7.40216911e-01 2.00103939e-01 -2.19475433e-01 6.47849813e-02
4.21337157e-01 -5.01780570e-01 1.19160199e+00 4.43734407e-01
5.78904033e-01 -6.90468907e-01 5.16810894e-01 2.61155009e-01
-9.76476729e-01 2.36477330e-02 -9.02095139e-02 2.60945112e-01
2.50194132e-01 3.59841257e-01 -1.01514804e+00 1.83691621e-01
9.19153810e-01 5.64667404e-01 -3.56679440e-01 1.13585126e+00
-2.79684275e-01 8.68551850e-01 -1.43564463e-01 1.08485527e-01
2.76473314e-01 -3.95920902e-01 6.71432316e-01 1.15176904e+00
1.70639247e-01 1.27351686e-01 7.70367861e-01 6.30455494e-01
-3.20705682e-01 6.40961528e-02 -1.69145823e-01 7.86879957e-02
5.35486460e-01 1.27966952e+00 -1.25848496e+00 -4.70127076e-01
-3.84901345e-01 1.01835787e+00 1.36240736e-01 6.61502659e-01
-1.14890027e+00 7.34879225e-02 5.26737213e-01 2.75798261e-01
4.83628601e-01 -5.01240976e-03 5.28047122e-02 -1.04818964e+00
3.01721669e-03 -9.78472531e-01 5.85701883e-01 -6.18455470e-01
-7.53460705e-01 4.95662183e-01 -8.97157285e-03 -1.29397929e+00
-3.98902357e-01 -1.45776391e-01 -5.88824511e-01 3.34899545e-01
-1.35122049e+00 -7.34876573e-01 -3.81470710e-01 5.93645155e-01
9.99230981e-01 9.57230628e-02 1.94145247e-01 2.91301370e-01
-5.09841561e-01 3.98558885e-01 -3.79325360e-01 3.31148535e-01
5.75238228e-01 -1.29106021e+00 1.41771987e-01 8.92167747e-01
3.07998657e-01 -5.32411644e-03 9.19746280e-01 -5.10639846e-01
-1.30322647e+00 -1.06732166e+00 2.68422902e-01 1.87939070e-02
4.47131574e-01 -6.49197176e-02 -6.74700201e-01 6.46592200e-01
3.89544398e-01 7.40366876e-02 4.84377891e-01 -2.35721961e-01
1.11714199e-01 -2.62937583e-02 -9.98857617e-01 8.26177180e-01
1.20135176e+00 -1.59663484e-01 -2.95687485e-02 5.01785219e-01
8.92953634e-01 -6.60902202e-01 -6.57285571e-01 4.02381383e-02
5.11609375e-01 -9.59020138e-01 6.77242339e-01 -3.94797295e-01
3.90501320e-01 -7.19795167e-01 -7.05583915e-02 -1.02400887e+00
3.43263522e-02 -8.59484911e-01 -1.47414610e-01 1.19194746e+00
5.64018071e-01 -5.36321029e-02 9.14805055e-01 7.04887927e-01
4.78194244e-02 -7.64390945e-01 -7.29174733e-01 -3.63818884e-01
-5.75209081e-01 -5.43131649e-01 4.85371143e-01 6.39656663e-01
-5.31859577e-01 -4.97181853e-03 -3.64217311e-01 3.78864318e-01
6.13343477e-01 4.71853018e-01 9.50003982e-01 -9.56215858e-01
-7.64587939e-01 -3.06500942e-01 -2.32557893e-01 -1.21543860e+00
2.17780963e-01 -6.85349703e-01 4.02269602e-01 -1.44626653e+00
2.51731783e-01 -5.35279453e-01 -3.32846314e-01 4.61303115e-01
-2.59202600e-01 1.79643214e-01 6.99614584e-01 2.46656880e-01
-1.44153178e+00 2.18623832e-01 1.41205621e+00 -2.57107139e-01
-5.56549013e-01 2.63298899e-01 -5.86259544e-01 9.39768851e-01
1.02040279e+00 -5.55664003e-01 -4.55318928e-01 -4.78619397e-01
1.86409160e-01 3.62023711e-01 2.47850269e-01 -8.55783641e-01
2.95630723e-01 -6.35277152e-01 2.34266505e-01 -2.54167944e-01
1.02106772e-01 -7.71638334e-01 1.27182633e-01 4.34578478e-01
-6.11093044e-01 -2.14515969e-01 -2.70167083e-01 3.90251577e-01
1.85297746e-02 -5.23281813e-01 7.52493739e-01 -4.01917100e-02
-1.05322433e+00 7.76410401e-01 -1.42321497e-01 2.61692911e-01
1.14067137e+00 -1.82832167e-01 1.66105479e-01 -5.97565711e-01
-9.86526549e-01 6.20907784e-01 5.01259029e-01 3.38301867e-01
5.24396360e-01 -9.53941524e-01 -6.41283095e-01 -1.74578205e-01
-6.55970648e-02 2.79463559e-01 2.16964990e-01 7.13884354e-01
-5.86172700e-01 -3.75287622e-01 7.73452073e-02 -7.69973755e-01
-1.35608530e+00 4.01540101e-01 3.72029185e-01 -1.46375552e-01
-7.14315772e-01 6.38053954e-01 3.17336887e-01 2.39160150e-01
4.80673462e-01 -1.69505492e-01 -2.89473265e-01 7.37330988e-02
6.46476626e-01 3.11337829e-01 -1.47137552e-01 -5.98041773e-01
-1.92956373e-01 1.70730695e-01 -3.63755487e-02 -3.58205914e-01
1.34057689e+00 -4.45811689e-01 2.41246998e-01 5.11474490e-01
9.03368115e-01 -1.02110855e-01 -1.97402728e+00 -1.79255903e-01
1.42122246e-02 -6.02739334e-01 -1.63076837e-02 -6.03961647e-01
-1.56780672e+00 3.60633373e-01 5.39819002e-01 3.43878627e-01
1.23110640e+00 -1.33682946e-02 7.49634981e-01 3.19940001e-01
4.50199902e-01 -1.45101464e+00 -3.01708002e-02 2.83795714e-01
6.82186067e-01 -1.27635109e+00 -2.13897228e-01 -3.16046566e-01
-1.05397451e+00 1.02218425e+00 8.55241597e-01 -2.39972740e-01
5.54339230e-01 3.21049482e-01 1.56807616e-01 -6.63063303e-02
-7.69426048e-01 -3.41696173e-01 -1.48138598e-01 3.10679108e-01
2.41326660e-01 1.61304712e-01 -2.53687382e-01 2.38490954e-01
-4.53965291e-02 1.43882751e-01 6.37111425e-01 6.90733671e-01
-5.39049208e-01 -1.01614976e+00 -1.68701530e-01 4.24054772e-01
-5.92244267e-01 2.09669754e-01 -1.01439267e-01 4.85903591e-01
2.96165645e-01 9.63974655e-01 9.46494788e-02 -2.51894057e-01
2.08318800e-01 -2.68072784e-02 3.18979353e-01 -4.44865763e-01
-5.18451571e-01 3.81627053e-01 6.75687864e-02 -1.06097960e+00
-6.96061671e-01 -7.97751009e-01 -1.76171076e+00 -2.79541433e-01
-2.11904988e-01 2.21329778e-01 2.91132569e-01 1.03757167e+00
2.85484046e-01 5.74827135e-01 7.66954362e-01 -1.01335859e+00
5.46556525e-02 -4.82249558e-01 -3.30432177e-01 5.54576278e-01
2.94274181e-01 -5.74107468e-01 -4.34994921e-02 5.64884305e-01]
|
[9.195234298706055, -0.12504230439662933]
|
ded3f26d-fe50-41ce-81ce-482ef2eaef96
|
few-shot-bayesian-imitation-learning-with
|
1904.06317
| null |
https://arxiv.org/abs/1904.06317v2
|
https://arxiv.org/pdf/1904.06317v2.pdf
|
Few-Shot Bayesian Imitation Learning with Logical Program Policies
|
Humans can learn many novel tasks from a very small number (1--5) of demonstrations, in stark contrast to the data requirements of nearly tabula rasa deep learning methods. We propose an expressive class of policies, a strong but general prior, and a learning algorithm that, together, can learn interesting policies from very few examples. We represent policies as logical combinations of programs drawn from a domain-specific language (DSL), define a prior over policies with a probabilistic grammar, and derive an approximate Bayesian inference algorithm to learn policies from demonstrations. In experiments, we study five strategy games played on a 2D grid with one shared DSL. After a few demonstrations of each game, the inferred policies generalize to new game instances that differ substantially from the demonstrations. Our policy learning is 20--1,000x more data efficient than convolutional and fully convolutional policy learning and many orders of magnitude more computationally efficient than vanilla program induction. We argue that the proposed method is an apt choice for tasks that have scarce training data and feature significant, structured variation between task instances.
|
['Josh Tenenbaum', 'Leslie Pack Kaelbling', 'Kelsey R. Allen', 'Alex K. Lew', 'Tom Silver']
|
2019-04-12
| null | null | null | null |
['program-induction']
|
['computer-code']
|
[ 8.86778533e-02 1.77232444e-01 -4.51260477e-01 -4.82613444e-01
-6.43760264e-01 -8.41742754e-01 9.09856021e-01 1.99405774e-02
-7.59990394e-01 1.25272834e+00 -2.82495748e-02 -6.59195900e-01
-1.32265136e-01 -6.40207589e-01 -1.26814449e+00 -4.83053476e-01
-3.09880793e-01 8.96443486e-01 3.61847550e-01 -1.66979432e-02
3.79916847e-01 4.28787768e-01 -1.63444912e+00 1.17825605e-01
7.97933102e-01 6.78437352e-01 4.50531393e-01 9.74734604e-01
-1.13246933e-01 1.20233893e+00 -6.47737265e-01 6.35152012e-02
2.01714709e-01 -1.46484241e-01 -6.68188632e-01 -2.80601114e-01
4.21893090e-01 -7.50992656e-01 -5.79011500e-01 1.19313049e+00
5.26958331e-02 5.02518296e-01 7.41822898e-01 -1.33184242e+00
-3.97363186e-01 7.44999945e-01 -2.74229705e-01 2.69167013e-02
4.72725391e-01 7.40993798e-01 8.04292977e-01 -1.46679848e-01
5.77984631e-01 1.51120067e+00 3.00253659e-01 8.18529963e-01
-1.47651112e+00 -7.13795841e-01 3.97691876e-01 -4.09516785e-03
-7.75274456e-01 1.41212121e-02 6.00890279e-01 -7.40810812e-01
1.16169310e+00 -1.19408928e-01 7.42167771e-01 1.68989003e+00
3.90835628e-02 1.08376515e+00 1.43037331e+00 -2.60272980e-01
7.27457106e-01 9.52584893e-02 9.89491940e-02 1.09682453e+00
3.18442553e-01 8.03838134e-01 -4.69937235e-01 -4.67046767e-01
1.02453279e+00 1.06187306e-01 1.60947582e-03 -8.91624391e-01
-1.13053715e+00 9.22802210e-01 -7.71611035e-02 -1.85267851e-01
-2.98546016e-01 7.58968711e-01 5.74348748e-01 5.62976420e-01
-2.39849880e-01 7.49448121e-01 -7.99610555e-01 -6.93470180e-01
-5.70304990e-01 8.66017997e-01 1.09459269e+00 1.20010138e+00
8.79260421e-01 2.52067506e-01 -1.02526911e-01 2.15097234e-01
1.39794275e-01 6.14499390e-01 3.85080963e-01 -1.32736290e+00
2.21823037e-01 1.40285507e-01 5.20574331e-01 -3.70858639e-01
-1.61441460e-01 -4.09720838e-03 -2.61500627e-01 7.69170702e-01
6.47647500e-01 -5.63650846e-01 -8.20142448e-01 2.01667833e+00
4.04104292e-02 1.55160362e-02 -5.34673408e-02 4.32169706e-01
2.47286811e-01 4.86221582e-01 1.45971790e-01 -1.61146522e-02
9.79245305e-01 -7.90235817e-01 -3.29378009e-01 -3.31715018e-01
4.90449518e-01 -1.25928387e-01 1.53467619e+00 5.73922455e-01
-9.67335761e-01 -7.09282994e-01 -8.39271605e-01 1.89690679e-01
-3.30493629e-01 1.20292641e-02 1.03988063e+00 4.47517931e-01
-8.60699117e-01 8.16786647e-01 -1.07739449e+00 -2.32404396e-01
6.12217009e-01 4.50914204e-01 -5.32959774e-02 2.89247781e-02
-7.14864969e-01 9.85346258e-01 8.35403562e-01 -6.52392507e-01
-1.93593943e+00 -6.62535131e-01 -1.07990205e+00 4.75094169e-02
7.86872685e-01 -5.31533957e-01 1.69939196e+00 -5.72048843e-01
-2.00532103e+00 6.46885216e-01 8.36849362e-02 -6.64367557e-01
4.01086211e-01 -3.18573296e-01 1.76930606e-01 4.96891737e-02
6.45559058e-02 7.66491294e-01 8.74003530e-01 -1.03236330e+00
-8.95398915e-01 -7.91190714e-02 7.41955638e-01 1.16947770e-01
1.06377870e-01 -1.24101587e-01 8.95338133e-02 -2.78295487e-01
-7.16867924e-01 -1.01322675e+00 -3.99529576e-01 -6.77476600e-02
-2.04925716e-01 -6.83547676e-01 6.27116323e-01 -3.24924648e-01
6.09419167e-01 -2.03073907e+00 2.90586114e-01 5.85616678e-02
1.43597901e-01 5.32016195e-02 -1.24266170e-01 2.12504491e-02
3.44894707e-01 3.48120718e-03 -6.20323047e-02 7.36675411e-02
6.24122024e-01 6.73775136e-01 -5.76717973e-01 2.64790684e-01
-1.00566428e-02 8.54162514e-01 -1.34785008e+00 -2.69608200e-01
4.32872623e-01 -2.17737883e-01 -8.72776628e-01 5.59468567e-01
-1.15144992e+00 6.00235403e-01 -8.40220273e-01 3.42385203e-01
1.38208568e-01 -1.39872923e-01 3.27821940e-01 5.68138659e-01
-5.03988862e-02 5.49902380e-01 -9.06864345e-01 2.00108552e+00
-4.15196598e-01 6.04373574e-01 -1.81801185e-01 -1.34707093e+00
9.62429821e-01 6.99566603e-02 3.59769672e-01 -2.47827068e-01
1.00376047e-02 -2.38812380e-02 1.59373462e-01 -6.97834253e-01
2.61859030e-01 -1.32177725e-01 -4.89432871e-01 4.63634968e-01
5.57501912e-01 -6.62894428e-01 3.88766795e-01 5.73964939e-02
1.37586856e+00 8.45799327e-01 5.66265166e-01 -2.56623268e-01
1.60162989e-02 2.34301284e-01 6.16285324e-01 1.58324659e+00
-3.85942370e-01 -1.26657784e-01 1.03513050e+00 -8.28174770e-01
-1.09733593e+00 -1.17005718e+00 3.82147133e-01 1.35929894e+00
-1.60062462e-01 -4.29289401e-01 -4.27020043e-01 -1.02058554e+00
2.23694965e-01 1.05100691e+00 -6.65695310e-01 1.78163014e-02
-6.75242305e-01 -1.03985265e-01 6.09628558e-01 5.90624094e-01
5.84152102e-01 -1.47230029e+00 -1.07884049e+00 1.11933269e-01
3.75545442e-01 -9.82852399e-01 -6.82043061e-02 6.79540575e-01
-7.47810185e-01 -1.30961311e+00 -4.16367114e-01 -8.29270661e-01
4.04196084e-01 -3.28933418e-01 1.24231839e+00 -3.42636108e-01
-3.49548385e-02 7.36015141e-01 -5.61319664e-03 -7.02314317e-01
-6.08670592e-01 -1.02865763e-01 3.60387683e-01 -9.62205291e-01
6.09094441e-01 -8.53606701e-01 -3.07591949e-02 -2.73181647e-01
-3.81039739e-01 4.58132811e-02 6.59010112e-01 1.02868271e+00
3.47923040e-01 2.27287650e-01 1.56821102e-01 -9.50006187e-01
7.85556376e-01 -3.26306373e-01 -1.27085483e+00 6.86531365e-02
-3.10522705e-01 6.19393289e-01 8.62715304e-01 -7.84314811e-01
-1.28843653e+00 3.02993119e-01 2.43753731e-01 -5.45429468e-01
-8.98993552e-01 2.89561301e-01 1.07086338e-01 1.82942435e-01
9.77503598e-01 3.38983953e-01 3.39918956e-02 -3.47348064e-01
6.71033025e-01 -2.09564622e-02 5.69285214e-01 -1.60624599e+00
8.01624894e-01 1.45677537e-01 -1.70764253e-01 -6.18671715e-01
-5.77116013e-01 7.89557397e-02 -5.30748129e-01 9.72314626e-02
8.91503870e-01 -7.30966628e-01 -1.10624611e+00 1.72307670e-01
-9.74357009e-01 -1.23582745e+00 -6.37347460e-01 6.54102921e-01
-1.22955954e+00 8.13093409e-02 -5.73923409e-01 -8.48535895e-01
4.07551914e-01 -1.35680807e+00 7.94929922e-01 4.32132363e-01
-3.32612514e-01 -8.49423945e-01 3.21476847e-01 -1.15614630e-01
5.03801517e-02 3.19910377e-01 1.30322015e+00 -7.74192452e-01
-6.89608932e-01 2.45732769e-01 1.67506143e-01 3.30501258e-01
2.27865446e-02 -1.37278363e-01 -6.12771511e-01 -3.73179406e-01
-1.42304435e-01 -1.02189100e+00 6.75523698e-01 4.99883831e-01
1.54550254e+00 -3.84964824e-01 -4.66499716e-01 4.13382679e-01
1.12208629e+00 5.83521485e-01 2.64926761e-01 3.40106487e-01
3.50174487e-01 2.73681402e-01 5.85869849e-01 5.19692361e-01
1.72931537e-01 2.52384484e-01 4.51675832e-01 5.78529894e-01
2.60901749e-01 -7.36712635e-01 8.14235091e-01 6.77785799e-02
-8.17855895e-02 3.05473238e-01 -7.39350677e-01 5.49478054e-01
-1.91579545e+00 -1.04429376e+00 6.42952323e-01 1.98517132e+00
1.05704844e+00 5.55516005e-01 3.81915182e-01 -5.92539608e-01
3.70557070e-01 9.15659815e-02 -1.00442410e+00 -4.88602757e-01
3.79648417e-01 7.58216381e-01 2.90384024e-01 5.17205238e-01
-1.08408296e+00 1.12195337e+00 6.98063374e+00 7.52919912e-01
-6.79540396e-01 -2.29192406e-01 2.08174109e-01 -2.36292407e-01
-1.55750632e-01 1.54212087e-01 -7.97620058e-01 4.48099196e-01
7.91914165e-01 -1.67275131e-01 8.52545321e-01 1.46744967e+00
-1.50194392e-01 -1.43295988e-01 -1.58609438e+00 8.09493005e-01
-3.47359002e-01 -1.38165128e+00 -2.07311139e-01 1.03055492e-01
8.97228479e-01 1.86332121e-01 2.21759081e-01 1.15209758e+00
1.61118340e+00 -1.07049727e+00 6.42829537e-01 2.86925733e-01
5.55749178e-01 -6.20012701e-01 1.86581075e-01 8.65675092e-01
-6.61298573e-01 -4.19214457e-01 -5.59553862e-01 -5.24975657e-01
-5.00763774e-01 -1.36801362e-01 -8.93600762e-01 -1.34142786e-01
8.53096068e-01 5.39871216e-01 -1.49281129e-01 6.37844503e-01
-8.62566292e-01 5.38913548e-01 -3.56883198e-01 -7.02474117e-01
5.60401559e-01 -6.40522838e-02 5.19151866e-01 1.08571792e+00
2.06418470e-01 1.79991618e-01 5.29718757e-01 1.32665002e+00
5.70337176e-02 -6.50732040e-01 -1.23142433e+00 -3.15509856e-01
5.63147247e-01 7.81762838e-01 -2.40571380e-01 -6.63159609e-01
-4.30677384e-01 5.75403273e-01 6.54809713e-01 5.40599048e-01
-8.76101911e-01 -2.33185828e-01 9.06387627e-01 -4.27998066e-01
6.33760333e-01 -4.29191649e-01 9.00855437e-02 -1.18235981e+00
-1.85179070e-01 -1.38103473e+00 2.58858532e-01 -6.24438047e-01
-1.17509556e+00 1.99769270e-02 4.54046994e-01 -8.63115609e-01
-8.03107738e-01 -1.01628554e+00 -6.35749519e-01 5.72083533e-01
-1.38349521e+00 -6.99448228e-01 -4.32025753e-02 5.81883073e-01
6.39632702e-01 -6.58559799e-01 1.08256292e+00 -4.81445462e-01
-1.63282514e-01 3.06720048e-01 1.21324964e-01 1.53439492e-01
3.57487708e-01 -1.73273861e+00 2.41362274e-01 4.86562222e-01
2.10923135e-01 7.80001104e-01 8.94663274e-01 -4.82794762e-01
-1.41905093e+00 -8.97934377e-01 -4.80593070e-02 -5.66592455e-01
7.16103494e-01 -3.95206064e-01 -6.92420542e-01 1.24879491e+00
3.83917242e-01 -9.69761312e-02 4.27340388e-01 4.42757368e-01
-4.21712041e-01 2.80896813e-01 -1.01125681e+00 8.72571528e-01
1.02029383e+00 -6.18567526e-01 -1.21426654e+00 3.24134588e-01
7.21336663e-01 -6.67346001e-01 -5.76038241e-01 1.77658483e-01
5.63422501e-01 -8.97435367e-01 8.59005690e-01 -1.25370169e+00
5.02318025e-01 -2.57795304e-01 -2.31184021e-01 -1.53870010e+00
-2.90574312e-01 -9.25182402e-01 -5.50220728e-01 5.48971534e-01
1.11821115e-01 -2.95993298e-01 8.97869885e-01 4.61827725e-01
-4.78183962e-02 -5.52165151e-01 -5.05631328e-01 -1.05423272e+00
4.87579793e-01 -4.05566514e-01 4.91981238e-01 6.66396022e-01
2.63285041e-01 2.68617570e-01 -3.39502126e-01 1.19957946e-01
1.01530671e+00 2.61640251e-01 1.33197105e+00 -1.02060437e+00
-8.78858268e-01 -6.01536095e-01 -1.64961237e-02 -1.69444966e+00
8.16028774e-01 -6.87106073e-01 3.94914150e-01 -9.78308022e-01
2.13197112e-01 -4.08797383e-01 -2.86252439e-01 5.58399320e-01
9.48595107e-02 -8.42702389e-01 -1.59911737e-02 -2.77796537e-01
-7.62922287e-01 6.80779755e-01 1.23260939e+00 -4.13571596e-01
-2.18620032e-01 7.78583810e-02 -6.72822118e-01 1.30922055e+00
9.46838260e-01 -5.49188256e-01 -7.38926351e-01 -3.99566233e-01
1.09658264e-01 2.49737561e-01 4.63716328e-01 -1.05585337e+00
2.07425028e-01 -7.66280174e-01 3.88981164e-01 -4.07444388e-01
2.72552997e-01 -4.83625948e-01 -4.76835310e-01 5.85609376e-01
-7.69834101e-01 -2.76980430e-01 4.14772689e-01 8.36389363e-01
2.18144774e-01 -3.73746902e-01 5.59470117e-01 -6.93315685e-01
-1.19268525e+00 2.37454429e-01 -7.57353544e-01 1.08322933e-01
1.08891690e+00 1.01744756e-01 -1.95452526e-01 -4.17157084e-01
-8.33637834e-01 3.76096070e-01 3.90644580e-01 1.79748893e-01
6.11591458e-01 -1.10053217e+00 -3.70260149e-01 1.73885211e-01
-7.32914777e-03 2.45518297e-01 -8.72930810e-02 1.91198260e-01
-3.81861985e-01 4.02583331e-01 -6.45898163e-01 -5.60469508e-01
-8.78956497e-01 7.49460995e-01 1.70931265e-01 -5.46225905e-01
-6.90982878e-01 8.88156652e-01 2.15123475e-01 -8.79704833e-01
5.74719727e-01 -7.77230620e-01 2.38399189e-02 -7.06904650e-01
3.92902344e-01 -3.21769081e-02 -8.56478870e-01 4.26006287e-01
1.83899906e-02 -2.66339555e-02 -2.01169491e-01 -2.18095124e-01
1.39754188e+00 5.03363848e-01 1.73198014e-01 5.93812287e-01
5.96105635e-01 -4.07948345e-01 -2.12268853e+00 -3.53094131e-01
1.27852589e-01 -4.42117065e-01 -4.38970268e-01 -8.28466356e-01
-3.40163291e-01 1.05688727e+00 2.19799846e-01 6.77177906e-02
5.88176608e-01 1.71820283e-01 2.34798312e-01 1.07678998e+00
8.17564607e-01 -1.13742149e+00 6.35199070e-01 8.86380255e-01
6.67480588e-01 -1.33773816e+00 -8.32430720e-02 4.25290436e-01
-4.96331513e-01 1.02256632e+00 1.04039705e+00 -5.71685016e-01
2.51661956e-01 5.41654050e-01 -6.46794498e-01 -1.74936697e-01
-1.03996110e+00 -9.19326916e-02 -1.63445577e-01 1.21126604e+00
-1.10702738e-01 2.42332891e-01 3.92980605e-01 6.23944640e-01
-2.64870316e-01 2.18978673e-01 4.31620240e-01 1.20463490e+00
-7.96952069e-01 -9.53142762e-01 -4.05039601e-02 5.37914991e-01
-1.16774954e-01 1.60806805e-01 -1.25559587e-02 1.09443915e+00
9.29578170e-02 5.24816930e-01 9.10551026e-02 -2.85143703e-01
1.43242385e-02 2.63609201e-01 1.12625587e+00 -9.26615655e-01
1.50483102e-02 -2.62600452e-01 4.39015124e-03 -7.48250306e-01
-3.09707165e-01 -7.91597664e-01 -1.20944571e+00 -4.31047231e-01
2.73282588e-01 4.57837470e-02 2.90026695e-01 1.05013072e+00
-3.49671096e-02 6.62758052e-01 1.74629763e-01 -6.99945271e-01
-1.19094968e+00 -8.91222954e-01 -6.13872766e-01 2.83648163e-01
4.01513547e-01 -1.03587008e+00 -2.11673498e-01 7.21488521e-02]
|
[4.141976833343506, 1.5532859563827515]
|
d0219c4c-fb31-499d-974f-ccb2f40a309e
|
structure-alignment
|
2307.02170
| null |
https://arxiv.org/abs/2307.02170v2
|
https://arxiv.org/pdf/2307.02170v2.pdf
|
Structure Alignment
|
While many good textbooks are available on Protein Structure, Molecular Simulations, Thermodynamics and Bioinformatics methods in general, there is no good introductory level book for the field of Structural Bioinformatics. This book aims to give an introduction into Structural Bioinformatics, which is where the previous topics meet to explore three dimensional protein structures through computational analysis. We provide an overview of existing computational techniques, to validate, simulate, predict and analyse protein structures. More importantly, it will aim to provide practical knowledge about how and when to use such techniques. We will consider proteins from three major vantage points: Protein structure quantification, Protein structure prediction, and Protein simulation & dynamics. The Protein DataBank (PDB) contains a wealth of structural information. In order to investigate the similarity between different proteins in this database, one can compare the primary sequence through pairwise alignment and calculate the sequence identity (or similarity) over the two sequences. This strategy will work particularly well if the proteins you want to compare are close homologs. However, in this chapter we will explain that a structural comparison through structural alignment will give you much more valuable information, that allows you to investigate similarities between proteins that cannot be discovered by comparing the sequences alone.
|
['Sanne Abeln', 'K. Anton Feenstra', 'Robbin Bouwmeester', 'Isabel Houtkamp', 'Dea Gogishvili', 'Jose Gavaldá-Garciá', 'Olga Ivanova']
|
2023-07-05
| null | null | null | null |
['protein-structure-prediction']
|
['miscellaneous']
|
[ 3.45271170e-01 -1.13632925e-01 -1.36302188e-01 -2.63135374e-01
-2.85244405e-01 -5.23010492e-01 6.71913996e-02 4.98019129e-01
-2.81369239e-01 1.22770357e+00 -2.96828926e-01 -7.10291207e-01
-1.32240459e-01 -3.07857305e-01 -4.61693794e-01 -1.26117849e+00
-5.25597855e-02 5.17355204e-01 4.37318951e-01 -6.49237156e-01
3.30236763e-01 7.60934174e-01 -1.56096315e+00 1.65457726e-01
9.74373460e-01 2.76836544e-01 8.04288387e-01 5.89472950e-01
-2.39482343e-01 5.68377137e-01 -4.26718295e-01 -4.32318836e-01
-1.99878216e-01 -8.61988842e-01 -1.10077858e+00 -1.19079068e-01
-5.90063334e-02 2.71446228e-01 4.57594335e-01 6.61489129e-01
5.08888483e-01 -1.91837221e-01 5.97705960e-01 -5.60992718e-01
-3.38045925e-01 -8.33898932e-02 -2.27417767e-01 2.48984680e-01
8.29627872e-01 2.89396763e-01 7.73960650e-01 -6.07289314e-01
1.02334285e+00 9.56982911e-01 6.55009985e-01 5.96531332e-01
-1.48664749e+00 -2.55517606e-02 -3.32986295e-01 3.89222294e-01
-1.02280676e+00 -1.29766449e-01 3.94553989e-01 -5.32289565e-01
1.47094619e+00 3.40033621e-01 8.00314486e-01 7.62433410e-01
6.55644536e-01 1.90389276e-01 1.10480249e+00 -5.23280203e-01
1.18778452e-01 4.07262072e-02 3.53976756e-01 5.92235088e-01
1.95994735e-01 1.18217207e-01 -4.58049178e-01 -3.74478310e-01
1.79603115e-01 -8.40493888e-02 -3.27758014e-01 -9.51862693e-01
-1.01139116e+00 6.81885779e-01 3.33378613e-02 5.70089340e-01
-4.28405195e-01 -4.62637216e-01 4.16997164e-01 2.16323912e-01
-1.24801137e-01 4.20057714e-01 -8.29270005e-01 -3.10341746e-01
-6.11049414e-01 5.20685315e-01 9.92275178e-01 3.87512296e-01
6.48978472e-01 -3.63131225e-01 7.87492752e-01 7.82562375e-01
3.83139104e-01 1.07574992e-01 5.70211172e-01 -6.68421686e-01
-1.84512109e-01 2.90060818e-01 2.39352211e-01 -6.38156176e-01
-3.51968616e-01 4.02847469e-01 -5.36335051e-01 4.96248752e-01
7.47073829e-01 2.47417465e-01 -5.79456747e-01 1.40390229e+00
4.38246816e-01 -4.20159698e-01 2.89817929e-01 7.30128288e-01
9.04655039e-01 4.94398713e-01 3.84046793e-01 -8.31351399e-01
1.53677845e+00 -5.99781692e-01 -5.53732395e-01 3.16283852e-01
8.11955988e-01 -1.05139303e+00 5.32619178e-01 5.01137078e-01
-1.22208381e+00 -4.09777820e-01 -1.19620943e+00 -2.08222002e-01
-5.71932793e-01 -1.72168136e-01 5.23523569e-01 5.38517356e-01
-6.84392273e-01 1.02416313e+00 -1.13615036e+00 -8.83543432e-01
-2.80481815e-01 4.55296427e-01 -5.93915462e-01 3.29744607e-01
-1.26224983e+00 1.61888754e+00 5.97004056e-01 -4.36059743e-01
-1.81681201e-01 -5.36258638e-01 -5.19602776e-01 -2.55106896e-01
7.16207596e-03 -6.97364330e-01 1.03280735e+00 -6.08765543e-01
-1.28398991e+00 1.32262361e+00 -7.39733100e-01 -4.36096191e-01
-3.76259461e-02 3.67861807e-01 -3.46460640e-02 2.08865717e-01
-2.58076400e-01 3.03411871e-01 -4.52635288e-02 -9.76079345e-01
-4.67516094e-01 -4.40019965e-01 -1.48554102e-01 3.04960072e-01
6.77055836e-01 4.95935857e-01 1.93002135e-01 -4.09964412e-01
3.25258791e-01 -6.32827282e-01 -5.17019928e-01 -1.50598899e-01
2.67482549e-03 -3.89103442e-01 4.80100989e-01 -6.39081597e-01
7.32225776e-01 -1.86492169e+00 4.27401453e-01 1.84271157e-01
3.00712824e-01 3.75331700e-01 3.07149053e-01 1.15685678e+00
-6.78972661e-01 3.78131270e-02 -3.72802615e-01 5.86386442e-01
-2.75058329e-01 3.26635510e-01 5.85247017e-02 5.65617681e-01
-1.38783991e-01 7.69271791e-01 -6.30875587e-01 -2.72686452e-01
3.84350479e-01 6.55832708e-01 -2.13938300e-02 4.18100283e-02
-1.33991865e-02 4.59093511e-01 -3.36141199e-01 4.58843112e-01
7.83320427e-01 -2.57114172e-01 8.04205775e-01 -1.45979390e-01
-3.50578666e-01 6.90853059e-01 -6.61900461e-01 1.19107604e+00
4.21581805e-01 1.49802446e-01 -2.64384467e-02 -1.29640472e+00
1.11519217e+00 4.20319110e-01 8.16551924e-01 -4.55513984e-01
-2.42815301e-01 2.07289681e-01 5.14181614e-01 -4.69916433e-01
-1.46932648e-02 -4.84583348e-01 5.35331428e-01 3.76459181e-01
-2.16068998e-01 5.58480062e-02 5.24901807e-01 1.68111697e-01
8.50529730e-01 4.53240931e-01 8.27144861e-01 -5.00455618e-01
1.09571707e+00 4.18015718e-01 4.64661360e-01 4.26965132e-02
-2.39243492e-01 1.62551135e-01 6.45356536e-01 -7.68171728e-01
-1.52347457e+00 -9.29156601e-01 -4.65828180e-01 9.92436528e-01
1.72414482e-01 -7.68134892e-01 -1.07488263e+00 1.18105680e-01
-1.49143100e-01 6.95977062e-02 -1.73037857e-01 9.19172633e-03
-5.52001059e-01 -1.33021605e+00 -3.29011790e-02 7.00234622e-02
-1.16789574e-02 -1.08758008e+00 -7.97279477e-01 4.38776135e-01
6.52455613e-02 -4.85764384e-01 1.85683474e-01 7.02378154e-01
-9.76937473e-01 -1.51782548e+00 -5.56421936e-01 -9.35063541e-01
3.72807473e-01 2.54996121e-01 1.04971504e+00 3.63757223e-01
-5.57979941e-01 -1.19039476e-01 -2.80216962e-01 -4.21472549e-01
-6.30345225e-01 -1.48580864e-01 1.96701005e-01 -7.83776581e-01
1.01600921e+00 -9.10154879e-01 -4.17561591e-01 5.01054168e-01
-6.80237055e-01 1.25075877e-01 4.61956382e-01 8.74207675e-01
9.22228456e-01 -9.56533775e-02 5.72717264e-02 -9.16899085e-01
6.96630418e-01 -7.75882229e-02 -4.80257303e-01 5.49947858e-01
-7.62477040e-01 2.86276013e-01 3.77956271e-01 -1.68847777e-02
-8.54842663e-01 4.39364910e-01 -6.80593371e-01 5.39845347e-01
-4.48153079e-01 4.60417211e-01 -4.27467674e-01 -3.56878877e-01
6.75398946e-01 6.79504931e-01 5.74813366e-01 -7.85165131e-01
-1.18716128e-01 3.73862356e-01 3.29090029e-01 -5.53959310e-01
3.65119010e-01 1.74666137e-01 1.24572016e-01 -9.51503336e-01
-1.79341570e-01 -6.62552238e-01 -1.03738070e+00 1.74254403e-01
7.72859752e-01 -2.27081940e-01 -1.31738210e+00 2.46916935e-01
-9.27516401e-01 2.44507995e-02 2.52005905e-01 5.37216902e-01
-1.01140070e+00 1.10064054e+00 -5.24947405e-01 -5.29051423e-01
-2.69663483e-01 -1.61486793e+00 8.50399017e-01 -5.80134504e-02
-3.50692898e-01 -1.01775825e+00 3.99172276e-01 4.18477178e-01
1.19766966e-01 4.22850668e-01 1.26223278e+00 -6.13846421e-01
-2.61663646e-01 2.62832016e-01 3.42616081e-01 -5.72808497e-02
2.29854360e-01 4.26480174e-01 -3.65535587e-01 -2.24403650e-01
1.25231579e-01 -2.01397121e-01 6.46754205e-01 2.29706407e-01
8.40882659e-01 -9.41860378e-02 -6.28405988e-01 3.77056658e-01
1.44335520e+00 7.20596611e-01 8.80650520e-01 6.62250519e-01
1.90296248e-01 1.03754926e+00 9.75080788e-01 1.66010503e-02
-3.08101714e-01 8.55694771e-01 1.78420126e-01 6.94287941e-02
2.78712392e-01 1.16633765e-01 2.45869920e-01 5.77801466e-01
-7.54427373e-01 9.71337706e-02 -1.05370283e+00 -6.95262402e-02
-1.66515601e+00 -1.19624114e+00 -5.01951337e-01 2.19576788e+00
1.30048466e+00 -7.73367360e-02 5.02781212e-01 2.06135556e-01
5.44344664e-01 -3.43354940e-01 -6.55692458e-01 -8.35427701e-01
-3.11439663e-01 3.90771955e-01 4.49651361e-01 5.35614312e-01
-7.54284322e-01 6.41955078e-01 7.88781071e+00 8.38625491e-01
-1.14765561e+00 -3.35161418e-01 2.65842944e-01 1.49205074e-01
3.93102057e-02 3.15891564e-01 -7.63537288e-01 4.57341552e-01
1.38425851e+00 -3.91857833e-01 4.63949703e-02 8.71610224e-01
5.09798408e-01 -3.83338869e-01 -9.16657925e-01 7.99319565e-01
-4.54708725e-01 -1.50742102e+00 -2.18889609e-01 4.23823148e-01
1.55663148e-01 -2.16339201e-01 -2.91147113e-01 -3.28686327e-01
1.91961646e-01 -1.14538348e+00 -4.89774868e-02 6.20429397e-01
1.42191827e-01 -8.31055701e-01 7.98057616e-01 5.82216680e-01
-9.63381767e-01 6.61652505e-01 -8.99691701e-01 -1.44298241e-01
3.70683819e-01 5.25401175e-01 -6.40663266e-01 4.95072067e-01
7.16990292e-01 6.29688144e-01 -2.54005104e-01 1.00885403e+00
-4.98773605e-02 8.36176276e-02 -2.00276807e-01 -3.97494197e-01
-3.97769026e-02 -7.73008585e-01 2.16827244e-01 9.07959700e-01
-2.44232893e-01 5.09305060e-01 -1.94955319e-02 5.29526651e-01
5.57356238e-01 5.59482038e-01 -2.13641703e-01 -8.44561905e-02
-8.29225704e-02 9.78715003e-01 -4.82374221e-01 -3.09886694e-01
-6.34257376e-01 8.84929299e-01 1.30006090e-01 1.36486530e-01
-5.10280192e-01 -4.63273436e-01 1.14533579e+00 3.33564341e-01
3.13048512e-01 -2.29338273e-01 2.92423159e-01 -7.31971562e-01
-2.43777305e-01 -1.09014022e+00 2.67727654e-02 -8.44937861e-01
-1.21695483e+00 2.11866319e-01 -7.45888380e-03 -5.14830112e-01
-2.65488744e-01 -1.08857667e+00 -2.45340958e-01 1.24628651e+00
-9.95761752e-01 -6.12142324e-01 2.57114679e-01 7.17885494e-02
2.19117612e-01 7.90578201e-02 9.40127611e-01 -1.12069555e-01
-3.90601128e-01 3.40267241e-01 8.72219145e-01 -3.65437984e-01
7.33365238e-01 -1.40652716e+00 3.78333658e-01 1.34890854e-01
-4.72507387e-01 1.24549174e+00 1.25727403e+00 -7.29997694e-01
-1.24811232e+00 -3.89434308e-01 1.11165702e+00 -4.87575740e-01
5.41379869e-01 -9.58522335e-02 -1.29300988e+00 3.28238368e-01
2.97716588e-01 -6.90593541e-01 1.25821137e+00 -1.69740006e-01
1.39146876e-02 4.59087402e-01 -1.26760709e+00 2.73998290e-01
7.80155540e-01 -2.47808978e-01 -8.43374074e-01 5.69635987e-01
2.72895783e-01 -3.97558838e-01 -1.45394731e+00 3.84754419e-01
8.79669309e-01 -1.37398195e+00 1.20779264e+00 -1.02587223e+00
1.63077176e-01 -4.83777970e-01 1.98954880e-01 -5.94034135e-01
-3.88435274e-01 -5.31907022e-01 1.97008863e-01 6.10188007e-01
4.76213425e-01 -7.34848201e-01 1.03142524e+00 7.33516812e-01
8.46600831e-02 -1.07095850e+00 -6.22863948e-01 -8.63050520e-01
3.45840722e-01 3.47567976e-01 4.89461660e-01 6.73767567e-01
7.24926293e-01 4.04186785e-01 -1.36977941e-01 -4.58548754e-01
4.60235864e-01 3.49625617e-01 6.71302497e-01 -1.24757135e+00
-2.14422464e-01 -2.50502974e-01 -7.97880650e-01 -1.05193269e+00
-1.77634910e-01 -5.77627957e-01 -3.75538558e-01 -1.41257846e+00
6.02463722e-01 1.77039027e-01 8.70383158e-02 2.49383241e-01
8.17859732e-03 -2.75016371e-02 -2.93833226e-01 3.95882070e-01
-2.21711069e-01 9.16645005e-02 1.11062765e+00 2.99087793e-01
5.99613274e-03 -1.04976520e-01 -5.50646007e-01 5.16373098e-01
1.00136173e+00 -3.42106938e-01 -1.98036835e-01 6.02272332e-01
2.31442153e-01 -1.04551807e-01 -9.89532918e-02 -5.60405493e-01
-8.82026106e-02 -5.46355605e-01 2.99535573e-01 -1.01723707e+00
3.08525741e-01 -8.14287722e-01 6.20461822e-01 1.01850533e+00
-7.45029077e-02 2.15164512e-01 -1.54382619e-03 3.66984189e-01
-1.12605602e-01 -6.88375592e-01 1.16725492e+00 -4.82351273e-01
-6.55930161e-01 6.64628530e-03 -1.02502108e+00 -5.94164848e-01
1.37257409e+00 -8.48022401e-01 -1.07988268e-01 -2.84479111e-02
-1.33492649e+00 -8.02383348e-02 1.15341973e+00 -2.37661213e-01
3.54249418e-01 -8.49042594e-01 -2.04177886e-01 -7.08311573e-02
2.14144588e-01 -6.00425124e-01 1.61552995e-01 1.06129658e+00
-1.26326978e+00 9.95535672e-01 -5.69791436e-01 -5.51605403e-01
-2.17197132e+00 1.18829763e+00 6.54712021e-01 -1.57716036e-01
-2.00090691e-01 5.26391387e-01 2.06246048e-01 -4.01537091e-01
-9.41945314e-02 6.67635649e-02 -4.41959828e-01 -3.17055047e-01
6.13230467e-01 1.10258684e-01 1.93397164e-01 -9.48284686e-01
-4.51122224e-01 7.83807576e-01 -3.15992206e-01 4.87686098e-01
1.40376651e+00 -4.13731068e-01 -6.19131207e-01 1.64696619e-01
1.06312573e+00 -2.58985102e-01 -6.50583088e-01 1.58221610e-02
2.22726762e-01 -2.38184810e-01 -7.42711604e-01 -6.55170143e-01
-3.48958164e-01 8.46575439e-01 7.33778834e-01 1.22124486e-01
9.26647484e-01 1.96541071e-01 7.24296212e-01 6.39921129e-01
2.56617665e-01 -8.63218606e-01 -4.59479153e-01 4.30854261e-01
4.97579485e-01 -9.19727325e-01 2.46359751e-01 -7.09551096e-01
-3.48490268e-01 1.30950654e+00 3.48309577e-01 2.43438587e-01
3.90095115e-01 2.86456019e-01 -2.56194100e-02 -4.01787162e-01
-9.32880163e-01 -2.74402350e-01 -8.01897347e-02 9.11315024e-01
1.10496140e+00 -2.81686753e-01 -1.26471281e+00 9.69638526e-02
-4.63295400e-01 -3.18731368e-01 3.65154147e-01 1.19541109e+00
-1.04910648e+00 -2.15531445e+00 -4.87858772e-01 9.85243767e-02
-7.56539702e-01 -1.66798383e-01 -1.19255722e+00 6.83167100e-01
-1.11893989e-01 5.65918207e-01 -4.53539670e-01 6.94013759e-02
8.33304152e-02 4.11074191e-01 7.89365232e-01 -3.39159966e-01
-3.83878499e-01 1.05800644e-01 1.26054376e-01 -4.53286409e-01
-9.51271176e-01 -7.52839923e-01 -1.53176439e+00 -8.29815269e-01
-3.22668076e-01 9.22323048e-01 7.97528505e-01 8.31210732e-01
1.62012219e-01 1.01947777e-01 5.52818701e-02 -4.85077292e-01
-1.65024966e-01 -6.55674517e-01 -7.99901724e-01 3.49798352e-01
3.86425667e-02 -6.77888930e-01 8.25801457e-04 4.47869450e-01]
|
[4.741698265075684, 5.2831854820251465]
|
9f4220a3-ab5d-49f2-ab73-90feecaaf775
|
a-paraboost-stereoscopic-image-quality
|
1603.09469
| null |
http://arxiv.org/abs/1603.09469v1
|
http://arxiv.org/pdf/1603.09469v1.pdf
|
A ParaBoost Stereoscopic Image Quality Assessment (PBSIQA) System
|
The problem of stereoscopic image quality assessment, which finds
applications in 3D visual content delivery such as 3DTV, is investigated in
this work. Specifically, we propose a new ParaBoost (parallel-boosting)
stereoscopic image quality assessment (PBSIQA) system. The system consists of
two stages. In the first stage, various distortions are classified into a few
types, and individual quality scorers targeting at a specific distortion type
are developed. These scorers offer complementary performance in face of a
database consisting of heterogeneous distortion types. In the second stage,
scores from multiple quality scorers are fused to achieve the best overall
performance, where the fuser is designed based on the parallel boosting idea
borrowed from machine learning. Extensive experimental results are conducted to
compare the performance of the proposed PBSIQA system with those of existing
stereo image quality assessment (SIQA) metrics. The developed quality metric
can serve as an objective function to optimize the performance of a 3D content
delivery system.
|
['C. -C. Jay Kuo', 'Rui Song', 'Hyunsuk Ko']
|
2016-03-31
| null | null | null | null |
['stereoscopic-image-quality-assessment']
|
['computer-vision']
|
[ 6.36782348e-02 -6.10833108e-01 -6.42125495e-03 -3.75290811e-01
-1.21970248e+00 -2.45689586e-01 5.80881476e-01 5.47651425e-02
-1.91250116e-01 5.35792291e-01 4.49603230e-01 -9.12202150e-02
-2.89307445e-01 -6.21733963e-01 -3.36287290e-01 -8.94041657e-01
2.67331004e-01 3.23459208e-01 4.34683055e-01 -3.96685660e-01
5.59916198e-01 4.71246243e-01 -1.64235568e+00 4.63611573e-01
1.25526845e+00 1.53733957e+00 4.27674919e-01 6.65029645e-01
2.85424720e-02 7.20659077e-01 -5.99089384e-01 -3.72779250e-01
4.51306134e-01 -3.54930311e-01 -7.28863776e-01 8.14085826e-02
4.32392091e-01 -5.57675719e-01 3.75339948e-02 1.14414668e+00
8.41996670e-01 -5.03992103e-03 7.10113347e-01 -1.09935296e+00
-3.06712277e-02 -9.09050405e-02 -3.86205971e-01 5.17361045e-01
5.80868542e-01 2.56985456e-01 1.02670228e+00 -9.04011011e-01
3.00385267e-01 1.35433507e+00 2.68693209e-01 2.08358273e-01
-1.07271898e+00 -3.98702502e-01 -1.54145718e-01 8.09058249e-01
-1.04369700e+00 -3.81832927e-01 9.22493041e-01 -6.77680731e-01
5.23135960e-01 1.35528088e-01 8.11299562e-01 4.51913148e-01
3.66873324e-01 8.15889060e-01 1.43437517e+00 -2.75783420e-01
4.26977903e-01 1.16130717e-01 -8.78070071e-02 5.77706039e-01
-2.47379988e-01 1.49126291e-01 -4.92269456e-01 -2.19242454e-01
3.53368223e-01 -2.39683360e-01 -2.83931732e-01 -6.19723558e-01
-8.01111341e-01 6.48284018e-01 7.25515544e-01 2.85905510e-01
-5.02452075e-01 -3.94038349e-01 4.41385061e-01 3.96913469e-01
7.93113530e-01 2.23771371e-02 -2.83528343e-02 1.55594677e-03
-9.57434952e-01 3.82318318e-01 2.79963791e-01 4.60749000e-01
7.10614800e-01 -9.80400965e-02 -4.89853323e-01 1.16901946e+00
4.10726935e-01 4.88910556e-01 4.40677345e-01 -9.44575846e-01
6.28632903e-01 4.98496354e-01 1.88269645e-01 -1.03719294e+00
-1.42901883e-01 -2.98659027e-01 -6.14743054e-01 6.85572088e-01
1.56454235e-01 1.24120206e-01 -7.41521299e-01 1.18139613e+00
6.02824450e-01 -2.68854350e-01 -6.68785870e-02 1.22267282e+00
9.11167920e-01 8.65893960e-01 -1.19219571e-01 -2.17084020e-01
1.07287097e+00 -9.28738713e-01 -5.41569948e-01 1.63261220e-01
8.93231556e-02 -8.92440736e-01 7.67418146e-01 5.74973285e-01
-1.47976184e+00 -9.15761411e-01 -1.12877703e+00 9.47756246e-02
1.85760222e-02 -2.06561923e-01 1.57456979e-01 7.43565321e-01
-1.45054579e+00 3.32734942e-01 -2.69010067e-01 -1.43372506e-01
2.43876562e-01 1.91456601e-01 -9.22777727e-02 -2.69309253e-01
-1.06522799e+00 1.02130270e+00 8.80510211e-02 -1.55012965e-01
-1.18805432e+00 -6.98572874e-01 -5.44026732e-01 -4.97779474e-02
8.39115679e-02 -7.99664497e-01 1.19691956e+00 -1.12010539e+00
-1.56715894e+00 9.53571558e-01 -3.06604058e-01 -1.53272733e-01
5.97227216e-01 4.90482934e-02 -3.56252015e-01 4.34495121e-01
1.52177185e-01 4.01144803e-01 9.72087622e-01 -1.45915151e+00
-1.16197073e+00 -8.19693208e-01 2.76812404e-01 7.94872344e-01
-1.11444905e-01 2.82861948e-01 -6.39032722e-01 -4.16677505e-01
9.61217433e-02 -3.78227472e-01 1.49641037e-01 2.82364059e-02
-1.36374459e-02 -2.84844279e-01 4.33210313e-01 -9.38607693e-01
1.13656402e+00 -1.92581928e+00 4.90251362e-01 4.25314039e-01
1.06248252e-01 4.37474340e-01 -3.00277192e-02 1.64877012e-01
3.58180143e-02 -3.67897511e-01 -8.56312513e-02 -2.77671039e-01
-3.05489182e-01 -2.48768225e-01 3.37180495e-01 2.77298421e-01
1.50644928e-01 3.49778920e-01 -7.40457952e-01 -7.73536980e-01
4.51966316e-01 1.54375389e-01 -4.77673531e-01 6.63812757e-01
1.16195560e-01 5.53044140e-01 -4.48749274e-01 8.03530335e-01
9.54056263e-01 2.81813112e-03 -2.05583140e-01 -4.32965577e-01
-1.80422679e-01 1.49177060e-01 -1.15813565e+00 1.65321815e+00
-5.66690624e-01 2.73774385e-01 1.10491306e-01 -1.05047655e+00
1.02881944e+00 3.15749645e-01 8.03523362e-01 -1.14802790e+00
1.22560740e-01 2.71371037e-01 -6.76348060e-02 -6.80831730e-01
3.19077432e-01 -9.64703858e-02 2.94045478e-01 1.52608782e-01
1.95380792e-01 -3.96727473e-01 1.33370727e-01 -9.11637843e-02
8.54927897e-01 1.29605541e-02 3.27951372e-01 -1.03549205e-01
1.13872313e+00 -8.01101252e-02 4.85748470e-01 5.19866586e-01
-6.56396389e-01 6.61522508e-01 2.14147702e-01 -4.61925656e-01
-1.38517308e+00 -1.17575824e+00 -1.44211367e-01 8.69231045e-01
5.97711861e-01 4.48738150e-02 -7.29260504e-01 -4.97343093e-01
-1.63093522e-01 1.64483756e-01 -1.36228383e-01 7.15564191e-02
-3.17730337e-01 -5.89338124e-01 -6.87737539e-02 3.81476097e-02
7.76894391e-01 -8.99729192e-01 -3.53957444e-01 2.54984647e-01
-4.08756524e-01 -8.04857135e-01 -2.64802486e-01 -3.04069817e-01
-9.08906996e-01 -1.03982997e+00 -1.28030694e+00 -6.86507404e-01
1.71115503e-01 5.81660450e-01 1.11200452e+00 -1.15508795e-01
1.22889511e-01 3.89989942e-01 -5.26150465e-01 -1.61762923e-01
-6.12847984e-01 -3.05409282e-01 -1.42800927e-01 4.47662264e-01
-4.99672405e-02 -4.86658037e-01 -1.07900357e+00 4.77386951e-01
-8.62160861e-01 -8.35586488e-02 4.78155881e-01 7.40117610e-01
4.93878365e-01 1.39827728e-01 2.70898670e-01 -2.55954683e-01
5.32413542e-01 -4.67617989e-01 -7.70933151e-01 2.76423156e-01
-7.13927388e-01 -2.49468222e-01 5.11113942e-01 -1.91134326e-02
-1.52529466e+00 -4.98033494e-01 -5.27615786e-01 -2.91111350e-01
-2.94692107e-02 2.51779646e-01 -3.77227694e-01 -4.46326971e-01
6.00516617e-01 2.21815825e-01 1.33145228e-01 -5.00098348e-01
-9.09820050e-02 8.76878023e-01 2.88867146e-01 -3.27434868e-01
6.13762081e-01 4.85337824e-01 -7.98233524e-02 -4.77054328e-01
-5.14822483e-01 -8.40501547e-01 -2.75716633e-01 -8.28243017e-01
7.97597170e-01 -1.05973744e+00 -5.23418486e-01 7.75129557e-01
-1.07758534e+00 1.89256042e-01 3.55425477e-02 5.09095371e-01
-7.99534857e-01 4.76780504e-01 -4.13194031e-01 -8.66976976e-01
-5.18600643e-01 -1.55109394e+00 1.18504775e+00 3.88389826e-01
3.36773187e-01 -7.36360312e-01 2.71452636e-01 9.58846569e-01
2.85498351e-01 -5.24413995e-02 9.84324455e-01 -1.44017801e-01
-7.18714952e-01 -5.93847670e-02 -3.88968885e-01 8.05081606e-01
-1.41699240e-03 -3.30068886e-01 -9.49075699e-01 -2.86395371e-01
2.33253777e-01 -4.20105249e-01 4.86349225e-01 6.12575352e-01
1.07693195e+00 -1.22547925e-01 1.94857359e-01 6.10994399e-01
1.67939854e+00 6.68565035e-01 6.35969639e-01 5.52262247e-01
3.80623996e-01 6.87557697e-01 9.86514151e-01 4.38304573e-01
3.83565992e-01 1.07768488e+00 6.18188620e-01 -2.58628458e-01
-1.97715417e-01 1.11381471e-01 3.43941748e-01 7.68335462e-01
-2.73290694e-01 -1.38808265e-01 -7.46928394e-01 4.38774794e-01
-1.65071177e+00 -9.28394735e-01 1.68665797e-01 2.13358879e+00
5.60760140e-01 1.77166358e-01 4.08309191e-01 4.85041887e-01
8.38754177e-01 3.18144530e-01 -4.47532505e-01 -5.34322679e-01
-1.07179821e-01 -1.19547946e-02 1.63998485e-01 3.19401443e-01
-9.12989378e-01 4.03356045e-01 5.77634001e+00 1.27506065e+00
-9.87696528e-01 2.68389702e-01 7.44335830e-01 2.12110683e-01
-3.04015636e-01 -7.78571889e-02 -3.46187472e-01 7.07808793e-01
5.33814788e-01 -1.27457500e-01 4.95347530e-01 6.34511352e-01
6.40251696e-01 -4.18100178e-01 -7.20768094e-01 1.24748874e+00
1.49008945e-01 -9.95855927e-01 1.99421048e-01 -9.26054865e-02
9.09061849e-01 6.82703778e-02 2.44390190e-01 -2.22804263e-01
3.49862427e-02 -5.21011949e-01 9.13444161e-01 6.34433687e-01
5.00647008e-01 -8.25309932e-01 9.04127598e-01 2.76595503e-01
-9.51324284e-01 -3.16993147e-01 -1.89214155e-01 2.04384536e-01
3.31541240e-01 6.19256675e-01 -4.47431028e-01 1.02401578e+00
1.08359945e+00 5.50209820e-01 -4.48430717e-01 1.58678281e+00
1.16209991e-01 2.71291256e-01 8.06649104e-02 3.63110303e-04
1.79916456e-01 -2.75988370e-01 8.70660186e-01 7.45193660e-01
4.34254110e-01 9.05564651e-02 -1.76793486e-01 4.69404727e-01
1.87759340e-01 4.34374213e-01 -4.69037890e-01 7.99538612e-01
1.72757730e-01 1.02269435e+00 -8.13310519e-02 -4.60020095e-01
-5.24780869e-01 8.52221072e-01 4.14593378e-04 1.95571408e-01
-5.86400032e-01 -4.25287373e-02 5.72654963e-01 2.94630472e-02
2.16760501e-01 2.78759301e-01 -1.95283547e-01 -1.08465409e+00
1.96862847e-01 -1.07405686e+00 4.13497180e-01 -1.18238163e+00
-1.28893554e+00 5.61663985e-01 1.09284073e-01 -1.70249617e+00
-1.17080584e-01 -3.38340938e-01 -6.57857358e-01 9.44151342e-01
-1.82452548e+00 -8.74654472e-01 -5.60747981e-01 9.26966012e-01
7.25736916e-01 -3.22983533e-01 2.85369188e-01 5.05039215e-01
-2.07331955e-01 3.79086792e-01 4.75768000e-01 -4.42181110e-01
6.03847623e-01 -1.17646813e+00 -1.17752463e-01 7.37898290e-01
-3.57186347e-01 -2.52308846e-01 7.34836698e-01 -4.20028448e-01
-8.69748473e-01 -7.58428156e-01 6.69551909e-01 3.73816900e-02
2.01539546e-01 2.41330296e-01 -5.84744453e-01 -3.19586962e-01
1.40035614e-01 -1.39544725e-01 4.25888449e-01 -1.56925216e-01
-1.88936666e-01 -7.99104869e-01 -1.44314742e+00 2.50252217e-01
7.84273207e-01 -4.99105245e-01 -4.61243063e-01 1.82928652e-01
3.29461664e-01 -3.15183729e-01 -1.08346522e+00 6.84391439e-01
6.81053996e-01 -1.53551340e+00 1.10572910e+00 -9.06028077e-02
5.82557917e-01 -3.58531296e-01 -2.18946382e-01 -1.53589749e+00
-4.17274088e-01 -2.10833400e-01 1.89114675e-01 1.11522102e+00
9.68118012e-02 -2.89932668e-01 5.03980279e-01 1.91163436e-01
-2.13038489e-01 -5.88790476e-01 -1.12247229e+00 -7.12080956e-01
-1.82308987e-01 -3.00720930e-01 6.91068470e-01 3.92822087e-01
-1.81731537e-01 1.33913457e-01 -4.09307361e-01 5.53188212e-02
9.67589736e-01 3.93907279e-02 6.94641173e-01 -8.65169168e-01
-3.98796976e-01 -5.22332311e-01 -7.49127686e-01 -8.82259190e-01
-4.04974520e-01 -5.99462152e-01 1.85165424e-02 -1.48726225e+00
4.09147888e-01 -3.91287625e-01 -5.63502371e-01 -3.16669703e-01
-4.82921869e-01 2.32219547e-01 3.08172196e-01 3.81586432e-01
-5.80231845e-01 8.27922106e-01 1.42573225e+00 -5.03797233e-01
-4.86860611e-02 2.83188462e-01 -3.48523468e-01 5.42866170e-01
6.24339461e-01 -1.69015229e-01 -3.23275506e-01 -5.11516511e-01
-9.28602666e-02 6.34630024e-01 3.24482232e-01 -1.25096750e+00
9.58655775e-02 1.01927463e-02 3.14528197e-01 -6.67951167e-01
4.27639186e-01 -9.44866776e-01 7.82182440e-03 3.85209143e-01
-1.04299471e-01 9.43897739e-02 -2.41834357e-01 3.49035949e-01
-7.04095840e-01 -1.35963857e-01 1.20614815e+00 4.52014133e-02
-7.97945201e-01 2.97649145e-01 -1.16522498e-01 -2.00877532e-01
1.00667667e+00 -3.83739769e-01 -1.13918453e-01 -5.06822705e-01
-4.76001769e-01 2.25023285e-01 3.30460429e-01 2.87114859e-01
7.95429170e-01 -1.52990508e+00 -7.92151570e-01 -1.01117238e-01
4.57115173e-01 -3.33499193e-01 6.54173613e-01 6.02189839e-01
-6.69175684e-01 2.63949901e-01 -6.21103227e-01 -9.62407351e-01
-1.39408386e+00 4.60963041e-01 4.59293038e-01 -4.40013796e-01
-1.79702744e-01 5.48052847e-01 2.92823315e-01 -4.47642803e-03
2.65969843e-01 8.68036970e-02 -7.05445707e-01 -1.18246777e-02
4.92467433e-01 6.85566664e-01 5.43962777e-01 -8.06225181e-01
-2.10975811e-01 8.97646785e-01 2.07266629e-01 -2.65112132e-01
1.24248707e+00 -5.04448116e-01 6.62741363e-02 1.29474983e-01
1.23866153e+00 -3.02030563e-01 -1.19981647e+00 -4.09847349e-01
-3.80123734e-01 -8.69156182e-01 2.90316790e-01 -8.32845986e-01
-1.07342434e+00 7.88447738e-01 1.31667185e+00 1.77835301e-01
1.74392033e+00 -1.40156507e-01 7.42430329e-01 -2.75594324e-01
5.28256238e-01 -1.06422186e+00 2.33045503e-01 1.29100204e-01
8.16857755e-01 -1.50582027e+00 -1.76125258e-01 -8.78156647e-02
-6.72335982e-01 7.72859097e-01 5.44059873e-01 -3.80214602e-02
4.98298556e-01 -3.46086949e-01 7.62298629e-02 -2.33475268e-02
-4.96390730e-01 -3.95607024e-01 6.27116382e-01 7.66778469e-01
-3.20634469e-02 -1.65041268e-01 -6.47655666e-01 1.39869535e-02
1.52212769e-01 -2.18814947e-02 8.83133411e-02 6.37369275e-01
-5.92550516e-01 -9.90649402e-01 -7.18817234e-01 2.84232050e-01
-4.68395561e-01 3.08655985e-02 1.40688390e-01 2.10279599e-01
4.79146361e-01 1.30560148e+00 -7.13198110e-02 -5.39115429e-01
5.47780156e-01 -3.68839443e-01 4.38318342e-01 -1.15272485e-01
-7.03861177e-01 1.57772064e-01 -5.87369800e-02 -8.69875133e-01
-7.67222166e-01 -6.11955106e-01 -3.70790690e-01 -3.26421976e-01
-5.57685271e-02 1.13922216e-01 9.98529971e-01 8.13259840e-01
-7.22277537e-02 2.44707391e-01 1.35607278e+00 -1.00631678e+00
-3.57248962e-01 -6.84759319e-01 -5.08973598e-01 5.77324510e-01
5.64371347e-01 -6.53926194e-01 -2.77997643e-01 -2.60316599e-02]
|
[11.844151496887207, -1.9748085737228394]
|
8225f791-2d40-46fe-aca9-9bd55e088c93
|
a-frugal-approach-to-music-source-separation
| null | null |
https://hal.archives-ouvertes.fr/hal-02986241/
|
https://hal.archives-ouvertes.fr/hal-02986241/document
|
A frugal approach to music source separation
|
During the past years, deep learning brought a big step in performance of music source separation algorithms. A lot has been done on the architecture optimisation, but training data remains an important bias for model comparison. In this work, we choose to work with the frugal and well-known original TasNet neural network and to focus on simple methods to exploit a relatively important dataset. Our results on the MUSDB test set outperform all previous state of the art approaches with extra data on the following source categories:vocals, accompaniment, drums, bass and in average. We believe that our results on how to shape a training set can apply to any type of architecture.
|
['Nathan Souviraà-Labastie', 'Emery Pierson Lancaster']
|
2020-10-02
| null | null | null | null |
['music-source-separation']
|
['music']
|
[-9.94953327e-03 -2.77547747e-01 -9.44148675e-02 -9.53358039e-02
-5.89739978e-01 -6.52234912e-01 5.24623215e-01 -9.19374749e-02
-4.68848139e-01 5.51407158e-01 3.96352500e-01 -2.01260932e-02
-5.02900541e-01 -3.14302266e-01 -2.16448784e-01 -6.97087765e-01
-2.92486399e-01 6.90647423e-01 1.73760742e-01 -7.70880222e-01
1.27304137e-01 3.76492798e-01 -1.59996772e+00 4.42636758e-01
2.13349983e-01 1.02947307e+00 -1.79627091e-02 6.64455891e-01
-9.61544961e-02 8.59990060e-01 -1.00328684e+00 -3.72397870e-01
3.01132590e-01 -8.83281887e-01 -9.66771185e-01 -5.14362216e-01
2.95582652e-01 2.15157092e-01 -4.13334183e-02 8.01968098e-01
1.17427289e+00 2.69794554e-01 5.49702823e-01 -1.06276429e+00
-1.88557371e-01 1.55957568e+00 -2.42423654e-01 5.73084414e-01
8.51285458e-02 1.08616808e-02 1.27469432e+00 -5.45240700e-01
3.80199701e-01 1.17233026e+00 1.04184747e+00 5.03280938e-01
-1.22233868e+00 -7.28779614e-01 -2.90686548e-01 4.02511418e-01
-1.29207242e+00 -8.00794184e-01 1.08455861e+00 -2.63137162e-01
7.36322045e-01 4.91779655e-01 6.08843625e-01 1.34439337e+00
-4.22007531e-01 8.82644415e-01 1.15303969e+00 -6.77191138e-01
2.18977481e-01 8.43572468e-02 1.00165889e-01 1.10650904e-01
-1.40370697e-01 8.43977332e-02 -7.52992928e-01 6.58992976e-02
6.48969173e-01 -7.86449492e-01 -2.74166733e-01 6.20044814e-03
-1.21856582e+00 6.97174847e-01 3.23446661e-01 9.93511200e-01
-1.83308393e-01 2.23045021e-01 6.97671115e-01 7.71740317e-01
1.41447127e-01 9.01885509e-01 -6.25882328e-01 -6.26079023e-01
-1.29920685e+00 2.56027490e-01 1.07447028e+00 5.15360117e-01
-3.70306633e-02 6.72041714e-01 -2.54929036e-01 1.27951229e+00
2.29322109e-02 6.57854751e-02 9.95406866e-01 -1.02383065e+00
2.32065231e-01 1.42038912e-01 -3.54879379e-01 -6.59668326e-01
-5.39472997e-01 -8.97651613e-01 -8.56496811e-01 3.13543141e-01
7.69921720e-01 -1.86679646e-01 -8.30697298e-01 1.52179849e+00
-6.74388930e-03 1.72668800e-01 -5.48218712e-02 1.03207099e+00
1.11778307e+00 2.94708014e-01 -4.88614708e-01 9.46914479e-02
1.07535696e+00 -9.77151752e-01 -5.91462076e-01 4.35237661e-02
2.34025791e-01 -1.17981005e+00 1.07634079e+00 9.37633276e-01
-1.33263373e+00 -9.30348992e-01 -1.09304905e+00 2.01498926e-01
-2.72407085e-01 2.18689203e-01 6.27041876e-01 8.88179362e-01
-8.66829216e-01 1.23964179e+00 -5.09911001e-01 -3.61651361e-01
1.73243299e-01 5.68167508e-01 -5.85716218e-02 6.69117808e-01
-1.13244736e+00 6.64594352e-01 7.56784618e-01 1.97745636e-02
-9.22099411e-01 -5.46967685e-01 -1.98416606e-01 8.27637762e-02
3.42342943e-01 -5.22080421e-01 1.48447263e+00 -1.31696129e+00
-1.89661407e+00 8.68945539e-01 4.57740128e-01 -9.66382623e-01
3.73028368e-01 -2.48227477e-01 -5.04741192e-01 -2.69050777e-01
-4.17286277e-01 5.78975976e-01 7.99492538e-01 -1.08445835e+00
-5.44035852e-01 7.85235092e-02 7.78204203e-02 -7.16109723e-02
-1.55314669e-01 4.29700404e-01 7.92652294e-02 -1.11426139e+00
-1.04395531e-01 -1.04180932e+00 1.92670792e-01 -7.68207312e-01
-3.90965939e-01 -2.91016966e-01 3.06231946e-01 -4.38665926e-01
1.27684081e+00 -2.12553883e+00 5.38352847e-01 -1.52447611e-01
-2.05041409e-01 3.87545347e-01 -2.05820978e-01 5.35314083e-01
-3.32200915e-01 1.17539600e-01 -1.14628032e-01 -3.80374193e-01
3.12850446e-01 -5.69824502e-03 -5.19086421e-01 2.88553923e-01
-6.35457635e-02 5.86353779e-01 -6.69128895e-01 -4.23815101e-01
-1.32425688e-02 4.82573777e-01 -4.73379463e-01 6.34448081e-02
-1.00193404e-01 5.17450273e-01 -1.43954018e-02 5.19884586e-01
2.37991452e-01 2.12036923e-01 -1.21968567e-01 -9.06236321e-02
-1.28153324e-01 8.95768225e-01 -1.34027731e+00 1.96994531e+00
-2.75334597e-01 9.10564244e-01 2.24136829e-01 -1.05211890e+00
1.01455021e+00 6.22583091e-01 5.95075965e-01 -4.18255925e-01
7.26742923e-01 6.03717923e-01 1.03584671e+00 -1.23587452e-01
3.39974135e-01 -4.37208802e-01 1.11400299e-01 4.12629455e-01
5.82464993e-01 -1.88644290e-01 3.90558690e-01 -2.70508319e-01
7.55766392e-01 2.29718506e-01 7.46891499e-02 -4.23600465e-01
4.30147290e-01 -4.93860617e-02 3.96087646e-01 6.81293011e-01
-2.13381425e-01 9.44962442e-01 3.96295220e-01 -4.35867876e-01
-9.29714203e-01 -8.19085181e-01 -2.39145949e-01 1.35399699e+00
-5.45895278e-01 -5.24014413e-01 -7.83380389e-01 -2.79696763e-01
-3.23229611e-01 6.64650202e-01 -4.45782870e-01 -6.25072513e-03
-8.74050319e-01 -7.10476875e-01 1.13172472e+00 5.10258198e-01
3.10592532e-01 -1.52149725e+00 -4.72786456e-01 3.20512950e-01
6.56509846e-02 -6.70304239e-01 -8.65644738e-02 8.01956058e-01
-9.75716293e-01 -7.52958179e-01 -8.72942626e-01 -7.35721946e-01
-5.67043364e-01 -1.50578752e-01 1.45933962e+00 -2.05136299e-01
-1.28630579e-01 3.09867263e-02 -6.24934673e-01 -9.38864052e-01
-6.21102154e-01 5.71676195e-01 1.21447533e-01 1.73967592e-02
3.46995264e-01 -1.01358569e+00 -3.38203669e-01 2.98598856e-01
-6.39572263e-01 -3.16993833e-01 3.67093563e-01 7.33510613e-01
1.93668172e-01 2.03802988e-01 7.04942703e-01 -6.43998206e-01
7.41101563e-01 -2.92348236e-01 -1.48323119e-01 -3.08233857e-01
-3.55280638e-01 3.81807946e-02 6.53697968e-01 -5.96128702e-01
-5.38059354e-01 -1.06244668e-01 -5.43772936e-01 -5.49193323e-01
-4.26462442e-01 4.36360568e-01 7.77305737e-02 1.51225477e-01
9.53818023e-01 -6.60455925e-03 -1.41546413e-01 -1.04571760e+00
1.87992528e-01 7.13748395e-01 4.42826182e-01 -6.49452209e-01
6.86530173e-01 2.80605823e-01 -9.30515975e-02 -9.98824060e-01
-7.28048742e-01 -3.96953046e-01 -9.35195267e-01 -3.63665856e-02
6.15048707e-01 -6.29346907e-01 -5.99036038e-01 3.22587520e-01
-9.48573291e-01 -4.85529840e-01 -7.46664226e-01 5.80506802e-01
-5.05942106e-01 1.94781146e-03 -5.83710730e-01 -8.54535520e-01
-3.72440219e-01 -8.83719206e-01 7.22149134e-01 -1.36331674e-02
-6.33406162e-01 -9.55235183e-01 5.49720228e-01 1.66568920e-01
7.68468976e-01 1.32119939e-01 5.44332147e-01 -9.29268897e-01
-1.39497146e-01 4.70520742e-02 3.09026390e-01 7.48703003e-01
2.36565042e-02 -3.89868096e-02 -1.55707550e+00 -2.00428635e-01
2.81366318e-01 -5.27345181e-01 1.16732275e+00 4.06617075e-01
1.05146658e+00 -2.75051426e-02 1.22393526e-01 7.38716602e-01
8.24776411e-01 2.15141654e-01 5.28801620e-01 5.40593386e-01
5.86985409e-01 5.49002111e-01 2.91188002e-01 1.68049470e-01
-2.42487013e-01 7.81142294e-01 4.51556593e-01 -3.08084134e-02
-4.86310840e-01 9.95437130e-02 3.35697651e-01 1.25949907e+00
-6.52144015e-01 -2.64147133e-01 -7.67655432e-01 5.57460308e-01
-1.52889919e+00 -1.09508204e+00 -4.08701785e-02 2.14653730e+00
9.01737690e-01 3.56243074e-01 7.53414690e-01 9.60825503e-01
4.17255193e-01 3.40299726e-01 -8.23034793e-02 -4.67374355e-01
-3.59598041e-01 7.26102173e-01 1.25477165e-01 6.51567355e-02
-1.20436013e+00 8.04772675e-01 7.02537537e+00 1.29933631e+00
-1.46579897e+00 1.42008528e-01 2.21124709e-01 -5.60194671e-01
8.41306001e-02 -1.27579883e-01 -6.38209105e-01 3.39842588e-01
1.31312704e+00 1.27256751e-01 6.26539528e-01 7.01852858e-01
1.57538112e-02 3.95143330e-01 -1.27973652e+00 1.17297542e+00
1.72248006e-01 -1.09504962e+00 -1.47259519e-01 -2.37789989e-01
5.67993641e-01 4.07674223e-01 -4.77980785e-02 5.62209189e-01
7.27002844e-02 -1.33812487e+00 9.99321640e-01 2.86552191e-01
3.68066907e-01 -7.95892537e-01 7.74639904e-01 2.07663342e-01
-9.36235547e-01 -1.34344190e-01 -4.79226172e-01 -2.55593419e-01
8.17170739e-02 2.51291752e-01 -8.81144464e-01 5.79025149e-01
8.92532587e-01 6.85952783e-01 -5.92707157e-01 1.37465250e+00
2.20560916e-02 1.28469110e+00 -5.28198183e-01 -1.00730240e-01
3.23077917e-01 -1.71322599e-01 9.20363545e-01 1.47603953e+00
2.52762556e-01 -5.09717166e-01 -1.06995597e-01 7.50408173e-01
1.23586198e-02 2.65514165e-01 -3.40605289e-01 -1.55059233e-01
4.34558354e-02 1.36682892e+00 -9.79561985e-01 -9.00540054e-02
3.18504013e-02 6.23467624e-01 2.14093607e-02 6.16440922e-02
-7.80535161e-01 -4.42130864e-01 4.88478720e-01 7.52848461e-02
3.65655214e-01 2.89493240e-02 -2.80416012e-01 -7.90926456e-01
-3.50663424e-01 -1.28270698e+00 5.06256998e-01 -6.63818836e-01
-1.14600825e+00 1.07801580e+00 -8.98123980e-02 -1.33506954e+00
-3.58306438e-01 -8.53782117e-01 -7.94442534e-01 7.41326392e-01
-1.12024617e+00 -8.02482367e-01 6.42801747e-02 4.93073523e-01
6.48216605e-01 -6.85559750e-01 8.18107069e-01 6.73414350e-01
-3.44285131e-01 5.66400707e-01 1.42893121e-01 4.38574553e-01
9.36437726e-01 -1.42611432e+00 3.01300317e-01 4.55654860e-01
1.16833150e+00 4.42619145e-01 9.32410836e-01 1.48231119e-01
-1.01092219e+00 -5.18060446e-01 4.56346244e-01 -5.70231736e-01
7.90602744e-01 -2.46404037e-01 -7.63352215e-01 4.47194934e-01
7.81752288e-01 -4.43073034e-01 7.34038591e-01 6.60716772e-01
-2.32789084e-01 -3.53283316e-01 -5.60187399e-01 4.95996654e-01
9.55271423e-01 -3.35581511e-01 -9.22851443e-01 -7.15263337e-02
5.58619797e-01 -4.19567674e-01 -6.66030407e-01 4.48967814e-01
7.43157387e-01 -1.32477295e+00 1.09291255e+00 -7.89558887e-01
3.85949522e-01 -3.99902277e-02 -2.16880720e-02 -1.67282295e+00
-4.34242487e-01 -9.26810920e-01 1.49869829e-01 1.57528532e+00
4.31913137e-01 -2.18143597e-01 7.15357006e-01 -3.70392680e-01
-4.59414482e-01 -4.20649201e-01 -1.13241887e+00 -9.90470409e-01
3.18234831e-01 -7.52331257e-01 6.14627063e-01 9.78515148e-01
-2.07351238e-01 8.49777460e-01 -5.28546810e-01 -6.35419726e-01
2.85709888e-01 1.81951940e-01 9.03178096e-01 -1.58913624e+00
-8.06752741e-01 -1.10508490e+00 -3.61083418e-01 -5.75038791e-01
-7.66749978e-02 -1.16528535e+00 -2.51083910e-01 -1.05911469e+00
-2.08789602e-01 -4.53414261e-01 -9.46847677e-01 3.89941156e-01
3.03682953e-01 6.40047669e-01 4.86358762e-01 1.18385568e-01
-4.02430952e-01 3.72445613e-01 1.11867368e+00 -2.39954948e-01
-3.93512845e-01 2.90800780e-01 -6.66426003e-01 9.73700583e-01
1.11942041e+00 -6.59240842e-01 -2.71326959e-01 -3.29309225e-01
3.55905741e-01 -2.72951812e-01 3.52496117e-01 -1.45187378e+00
-4.39670794e-02 2.48171300e-01 7.78774470e-02 -5.26674449e-01
6.52681947e-01 -6.75334573e-01 2.48125523e-01 3.42193693e-01
-5.27430594e-01 -3.23751211e-01 4.18862998e-01 -2.94158068e-02
-5.21617293e-01 -6.71347022e-01 7.48236239e-01 -1.80420935e-01
-3.50293487e-01 -1.59926921e-01 -1.43787533e-01 3.76807094e-01
2.77989954e-01 -1.50777087e-01 1.11641176e-01 -3.82075667e-01
-9.43063378e-01 -3.59673500e-01 -2.17479486e-02 7.03637004e-01
-7.96582699e-02 -1.40466309e+00 -1.04067826e+00 1.11605022e-02
-1.58053875e-01 -4.10313845e-01 -3.29619385e-02 1.03135121e+00
-3.59539539e-01 4.70149934e-01 -5.90419888e-01 -5.73597074e-01
-1.34869432e+00 4.26760048e-01 4.50268686e-01 -8.28049481e-02
-4.80831772e-01 1.06091857e+00 -2.62929022e-01 -2.72240698e-01
6.02807999e-01 -3.36108387e-01 -4.74091470e-01 5.12762189e-01
3.25692385e-01 6.14712536e-01 8.97400007e-02 -7.16294110e-01
-2.68803865e-01 4.72847670e-01 2.15634137e-01 -3.21914345e-01
1.59244108e+00 3.08810383e-01 7.74214882e-03 1.19099593e+00
9.52554345e-01 3.50197643e-01 -6.63582087e-01 -4.22807829e-03
8.57834220e-02 -2.27626398e-01 1.86157137e-01 -9.39666271e-01
-1.15340662e+00 1.12612867e+00 6.29405618e-01 7.07845569e-01
1.10895002e+00 -1.30197601e-02 4.92308676e-01 3.52179587e-01
2.79871017e-01 -1.19157994e+00 -1.23206347e-01 5.95919788e-01
1.10523951e+00 -1.11224568e+00 -2.07234189e-01 1.58413455e-01
-6.50017202e-01 1.22539341e+00 1.81711540e-01 -4.81115848e-01
6.04950488e-01 4.16973948e-01 2.49056205e-01 -8.69243741e-02
-5.20775974e-01 -4.46299314e-01 6.31405115e-01 4.35983777e-01
1.02035880e+00 -1.80188999e-01 -2.70405710e-01 7.79709399e-01
-8.77701104e-01 4.00895551e-02 3.78083110e-01 4.95664418e-01
-4.24272835e-01 -1.39860511e+00 -6.08403206e-01 2.63893932e-01
-1.03462827e+00 -2.51039773e-01 -8.43097210e-01 1.07636738e+00
2.72976398e-01 8.93172204e-01 -2.14169607e-01 -4.55217183e-01
2.86652952e-01 4.09692883e-01 6.89991057e-01 -5.09300530e-01
-1.23383117e+00 5.35627007e-01 3.28473300e-01 -2.41044655e-01
-6.69775724e-01 -5.69476545e-01 -9.36012089e-01 -2.28096634e-01
-3.40475082e-01 3.86591971e-01 4.93232787e-01 7.42017508e-01
6.86074793e-02 1.06067348e+00 3.11986834e-01 -1.17524433e+00
-6.04668617e-01 -1.59764397e+00 -7.30553091e-01 3.89286786e-01
2.94421196e-01 -5.33287942e-01 -4.87129211e-01 8.68194252e-02]
|
[15.789474487304688, 5.3004984855651855]
|
d44cb183-647b-4afe-9608-5c052115c6c6
|
temporally-anchored-relation-extraction
| null | null |
https://aclanthology.org/P12-1012
|
https://aclanthology.org/P12-1012.pdf
|
Temporally Anchored Relation Extraction
| null |
["{\\'A}lvaro Rodrigo", 'Bernardo Cabaleiro', 'Anselmo Pe{\\~n}as', 'Guillermo Garrido']
|
2012-07-01
| null | null | null |
acl-2012-7
|
['temporal-information-extraction']
|
['natural-language-processing']
|
[-8.63703638e-02 1.71006292e-01 -6.22772932e-01 -4.08054382e-01
-8.41685571e-03 -9.08429027e-01 6.55310392e-01 -6.53472245e-01
-2.85945535e-01 1.06888819e+00 -4.63127941e-02 -1.01159286e+00
-3.91567826e-01 -9.63214397e-01 -4.95059669e-01 -6.31337762e-01
-9.79754329e-01 7.25764990e-01 3.30370307e-01 -6.93831444e-01
7.03166842e-01 7.88774848e-01 -1.68942046e+00 7.18545914e-01
7.04417467e-01 8.52217197e-01 2.49141872e-01 1.14950800e+00
-1.95044339e-01 1.55633950e+00 -7.48382092e-01 -5.46825826e-01
3.13719302e-01 -1.23176083e-01 -7.22945035e-01 -1.01074085e-01
9.28529128e-02 -8.59008506e-02 -2.09758401e-01 9.22211111e-01
5.37373662e-01 4.49454933e-02 1.08379531e+00 -1.42548037e+00
-5.91619551e-01 6.10313773e-01 -4.01565880e-02 1.21627934e-01
1.03678203e+00 -5.39447069e-01 1.19919395e+00 -1.13026452e+00
7.20913768e-01 1.26888943e+00 8.66221786e-01 5.44149756e-01
-1.22286928e+00 -1.94712028e-01 -3.26822817e-01 -9.51717794e-02
-1.46558487e+00 -3.25250506e-01 4.25783843e-02 -2.08119690e-01
1.66093647e+00 1.26596653e+00 1.20609856e+00 1.01401424e+00
1.26658809e+00 8.34431887e-01 1.04267764e+00 -5.13792276e-01
3.35295945e-01 3.66983831e-01 1.54683650e-01 6.33519173e-01
8.40953708e-01 5.26628852e-01 -7.06372619e-01 -9.13127720e-01
9.33553874e-01 -2.94925272e-01 1.71355158e-01 -5.05680561e-01
-9.05919552e-01 6.91228509e-01 1.78732842e-01 3.83959889e-01
-1.39880210e-01 9.89067405e-02 1.26390755e-01 5.30987144e-01
-2.58292928e-02 6.47037446e-01 -9.11868811e-01 -1.33165747e-01
-8.71728659e-01 5.10332465e-01 1.25398111e+00 1.52653182e+00
1.24482810e-01 2.94908643e-01 -9.34252143e-02 3.17179203e-01
8.92314315e-01 1.01808000e+00 4.28362608e-01 -1.36146402e+00
-6.87414408e-02 1.72361732e-01 5.01781464e-01 -8.52631688e-01
-6.33224547e-01 -9.64177120e-03 -8.93263519e-01 4.49267089e-01
3.49161088e-01 4.57367361e-01 -8.02827001e-01 5.07305264e-01
4.33481112e-02 -2.34125629e-01 4.53833073e-01 5.55570945e-02
4.99930978e-01 3.76208365e-01 -1.34477139e-01 -5.73289394e-01
1.06082785e+00 -1.36716676e+00 -1.35299087e+00 2.33215362e-01
9.05734658e-01 -1.07320261e+00 4.35900748e-01 5.33875942e-01
-1.55548143e+00 -1.37560293e-01 -1.08699942e+00 1.78573877e-01
-7.27255583e-01 -3.14239264e-01 8.57801437e-01 1.43120694e+00
-1.60129595e+00 9.73287821e-01 -4.91727620e-01 6.59165755e-02
1.20568443e-02 8.24621081e-01 -2.64718989e-03 4.62812334e-01
-1.33193445e+00 1.08501506e+00 2.22979754e-01 -1.21242590e-01
-1.65216476e-01 -2.13068098e-01 -8.23704481e-01 -5.63443303e-01
-4.78693932e-01 -5.29636025e-01 1.44139910e+00 -2.59346128e-01
-1.65295815e+00 9.71794367e-01 -1.42069459e-01 -1.97814897e-01
6.14786744e-01 -1.28011424e-02 -8.31891418e-01 2.42498964e-01
-1.89849049e-01 5.76383233e-01 9.28263724e-01 -1.35132408e+00
-7.59897232e-01 -1.67359829e-01 -1.23336017e-01 2.66287565e-01
-1.25510961e-01 1.89734384e-01 2.11616129e-01 -1.12999000e-01
3.27147305e-01 -7.20919967e-01 -2.53068686e-01 -5.32041907e-01
-1.46512717e-01 -7.10518599e-01 7.70373225e-01 -4.51523662e-01
1.83705616e+00 -1.67618537e+00 -2.06720144e-01 3.98590982e-01
3.57815564e-01 -1.24705513e-03 2.40583986e-01 1.08380008e+00
-2.76906848e-01 7.32199550e-01 4.11965609e-01 -1.10722095e-01
2.24991128e-01 4.85861301e-01 -4.16602850e-01 3.05609167e-01
-1.29282743e-01 1.15307164e+00 -1.16605783e+00 -5.23096442e-01
4.11106765e-01 9.42391157e-02 -4.58732933e-01 4.79237735e-01
2.99364805e-01 1.47170946e-01 -3.56553018e-01 1.39399457e+00
1.15709066e+00 -1.31984919e-01 1.45911396e-01 5.30878425e-01
-3.79135728e-01 3.55090618e-01 -6.96863770e-01 1.05554795e+00
7.08333924e-02 5.00173986e-01 1.02364108e-01 -7.94621468e-01
3.33247900e-01 8.61540735e-01 4.37155962e-01 -9.67555881e-01
-2.26398129e-02 6.92409754e-01 1.12803578e-01 -6.07703328e-01
7.58228302e-01 3.94563079e-02 -3.64872098e-01 6.40070081e-01
-2.37588286e-01 -6.59476995e-01 9.64643434e-02 3.08100313e-01
6.22585893e-01 -5.18246442e-02 5.62923312e-01 -1.05613089e+00
7.86340594e-01 -1.82965681e-01 -1.81299388e-01 1.03415680e+00
-3.09923887e-01 3.19085121e-01 2.99841821e-01 -6.65102363e-01
-6.45341039e-01 -1.12307119e+00 -4.89381433e-01 1.30636716e+00
3.24267983e-01 -4.39044595e-01 -9.54439282e-01 -2.49762803e-01
1.77620783e-01 6.89606130e-01 -5.90509653e-01 3.84124845e-01
-5.03739953e-01 -8.32535863e-01 7.39044368e-01 3.45434904e-01
-5.07752821e-02 -1.33414865e+00 -6.58416986e-01 1.25490099e-01
-2.20292807e-01 -6.63697243e-01 -6.23428151e-02 4.48765576e-01
-1.35989368e+00 -5.18594682e-01 -6.66252747e-02 -8.20914626e-01
5.87345481e-01 2.46782884e-01 1.27047324e+00 5.39230824e-01
-2.31483161e-01 4.26904231e-01 -1.21292919e-01 -4.95818377e-01
-4.59671497e-01 -8.00336525e-02 5.28869390e-01 -5.87835789e-01
5.19427478e-01 -2.50617653e-01 -7.29350567e-01 5.37953973e-01
-6.88540697e-01 1.62748516e-01 1.79803044e-01 1.04410267e+00
1.35816500e-01 -9.34035778e-02 1.22507080e-01 -6.38007045e-01
8.72274399e-01 -1.69219792e-01 -3.78732830e-01 5.77745810e-02
-6.77108407e-01 -3.74140263e-01 3.21430594e-01 -3.25342178e-01
-1.01981449e+00 -4.87835288e-01 -9.82677937e-02 2.45538145e-01
1.11353043e-02 -1.46784872e-01 6.47139177e-02 -5.24923325e-01
8.02199244e-01 9.25758183e-02 1.99174434e-02 -6.80815242e-03
3.01039815e-01 7.09525108e-01 -6.82967342e-03 -6.68678164e-01
8.44880998e-01 4.91470337e-01 7.98524171e-02 -9.57177758e-01
-1.52186140e-01 -2.60129690e-01 -9.51962709e-01 -6.54426932e-01
6.56643391e-01 -6.78531289e-01 -9.10833478e-01 3.91110867e-01
-9.38691139e-01 -3.38627815e-01 -3.91645581e-01 4.25431967e-01
-1.01278400e+00 2.75717527e-02 -3.90154392e-01 -1.27895141e+00
-5.10977268e-01 -1.02017939e+00 9.43384409e-01 5.30070923e-02
-5.10597289e-01 -1.26927447e+00 5.87685481e-02 2.71537274e-01
1.81734428e-01 -1.73075795e-01 6.90226793e-01 -2.38256708e-01
-4.24233019e-01 -1.53791070e-01 2.34436691e-02 -1.39755070e-01
1.70832314e-02 4.95917559e-01 -9.81751978e-01 -5.31145096e-01
6.65065646e-02 -1.92070693e-01 -1.08835101e-01 6.52520418e-01
5.91872573e-01 -2.29931593e-01 -8.56000841e-01 5.40386558e-01
1.38545322e+00 3.85070026e-01 5.32770038e-01 7.28214979e-01
1.41836226e-01 5.53460240e-01 9.17806149e-01 4.63203549e-01
1.30579369e-02 3.28798652e-01 2.40537539e-01 1.49327129e-01
1.11720070e-01 -1.54819340e-01 3.77893507e-01 1.16112018e+00
-8.18235934e-01 -2.69281328e-01 -5.07867396e-01 4.42987174e-01
-1.72482407e+00 -1.40330648e+00 -4.32368398e-01 6.90478683e-01
6.25676990e-01 1.56016424e-01 -1.48347050e-01 3.35214496e-01
4.99015123e-01 -2.03574806e-01 -1.19133167e-01 -1.06291151e+00
-1.43546045e-01 3.15233678e-01 7.37729073e-01 1.00061214e+00
-7.20721722e-01 1.03317809e+00 1.29781246e+01 1.02230716e+00
2.21112028e-01 1.03134915e-01 5.16071796e-01 3.48020852e-01
-4.36954498e-01 -4.56139445e-02 -1.04416132e+00 2.72933897e-02
1.38140702e+00 -4.30666685e-01 6.85999811e-01 5.44219851e-01
3.44648361e-01 -4.23268199e-01 -1.26188684e+00 5.26221812e-01
9.73738134e-02 -1.40886843e+00 -2.83300440e-04 6.85225725e-01
7.73699820e-01 -5.08050561e-01 6.22419357e-01 3.24184299e-01
6.09259963e-01 -1.14389277e+00 8.60300779e-01 2.53660440e-01
1.03040910e+00 -6.05088234e-01 5.67372203e-01 1.68872893e-01
-1.14389896e+00 -2.20873043e-01 -8.77727985e-01 -1.00755692e+00
3.93533185e-02 -1.81779593e-01 -4.29956943e-01 3.48861217e-01
9.58353162e-01 2.99398601e-01 -3.93658698e-01 9.95779395e-01
-4.78694476e-02 1.04875881e-02 -2.71853864e-01 -4.48467314e-01
4.83122796e-01 -3.54241252e-01 4.66730654e-01 1.00164843e+00
2.48499006e-01 3.51035744e-01 -9.84472036e-02 4.01770771e-01
5.45058846e-01 3.29446048e-02 -1.19659424e+00 -1.78908288e-01
2.83276141e-01 9.16795909e-01 -4.83487815e-01 -4.22520459e-01
-2.00212970e-01 8.62069130e-01 -3.55488248e-02 5.01107454e-01
-6.11489356e-01 -4.35615242e-01 9.72222984e-01 -1.27327025e-01
-1.14700586e-01 -3.48497719e-01 -6.23769283e-01 -7.30352640e-01
-5.89872956e-01 -4.54965204e-01 5.93606755e-02 -5.53365827e-01
-1.39813089e+00 5.79277515e-01 -2.27688253e-02 -1.40553558e+00
-6.99901402e-01 -1.27676582e+00 -4.76714373e-01 4.92853165e-01
-1.11898029e+00 -1.10984349e+00 2.50124663e-01 4.52870727e-01
1.64141744e-01 -5.34416080e-01 1.39563632e+00 3.57715860e-02
1.00637585e-01 9.24474537e-01 6.69434488e-01 -7.37814724e-01
5.56605101e-01 -1.27867436e+00 5.68737745e-01 -1.37897313e-01
-4.31265175e-01 9.05828118e-01 6.28349900e-01 -5.39804697e-01
-1.41196322e+00 -3.66917729e-01 1.08350635e+00 -9.83769417e-01
6.55218959e-01 -3.86345625e-01 4.23767231e-02 7.88592756e-01
7.15902448e-01 -6.18741751e-01 8.21781039e-01 -1.83753878e-01
1.80774391e-01 5.75296998e-01 -1.39248300e+00 6.12354755e-01
1.66275799e+00 -4.63594139e-01 -6.25784039e-01 7.60327101e-01
8.13696027e-01 -6.94087505e-01 -1.30082703e+00 3.34633321e-01
8.65424156e-01 -8.75409484e-01 1.61978090e+00 -1.32660246e+00
-4.43697497e-02 2.81152606e-01 -2.61993498e-01 -9.32519078e-01
-5.96193194e-01 -1.23518765e+00 -5.33532679e-01 -5.83747849e-02
5.96577883e-01 -1.13057327e+00 3.42365682e-01 8.74560475e-01
-2.82833427e-01 -6.42737269e-01 -1.06996536e+00 -1.32016802e+00
-3.35779637e-02 -1.45572275e-01 4.90409225e-01 7.63798356e-01
6.83744550e-01 1.09839931e-01 -6.36873543e-02 -1.00294888e-01
5.33176839e-01 9.55312885e-03 4.41501856e-01 -1.34294486e+00
3.86843324e-01 -5.75816095e-01 -3.07655483e-01 -9.45992947e-01
-8.85957032e-02 -8.12076271e-01 -6.53862000e-01 -1.28511906e+00
-8.32044985e-03 -1.91056758e-01 -1.11109078e-01 -1.65725678e-01
3.67937148e-01 2.13746816e-01 1.20859891e-02 1.03788137e-01
-3.71160030e-01 6.18435517e-02 1.29639816e+00 6.91750320e-05
-1.62315920e-01 4.85058486e-01 -4.81304944e-01 7.84440815e-01
8.58408585e-02 -2.99253196e-01 -6.78878546e-01 6.11881316e-02
6.69384480e-01 4.61409837e-02 3.24159935e-02 -7.42885649e-01
5.37211418e-01 -3.75702560e-01 4.78586555e-01 -1.32223868e+00
1.30741090e-01 -9.61415648e-01 6.95283338e-02 9.40189242e-01
2.74610907e-01 1.20424610e-02 8.66204947e-02 5.39500564e-02
-1.44871444e-01 -5.70943117e-01 9.21121240e-01 -4.07591403e-01
-4.92852688e-01 -5.20386267e-03 -1.03226590e+00 8.97834301e-02
9.92593169e-01 -7.84614205e-01 -3.59281451e-01 -4.20183957e-01
-8.29068601e-01 -1.95836127e-02 6.50830388e-01 3.03609259e-02
7.30431557e-01 -1.51530886e+00 -2.30721906e-01 7.18729138e-01
-3.22939813e-01 -3.74400020e-01 -1.70157343e-01 6.58265352e-01
-1.32361674e+00 1.02442718e+00 -5.41665435e-01 -4.55340147e-01
-1.14228773e+00 4.78126436e-01 4.28307921e-01 -2.41845414e-01
-2.15481281e-01 1.11954463e+00 2.71224789e-02 -8.23025763e-01
1.85185194e-01 -9.21545625e-02 -7.54407048e-01 3.55081353e-03
6.88606799e-01 1.05194807e+00 -2.91290224e-01 -6.04341030e-01
-4.56784427e-01 6.65885091e-01 2.32151806e-01 -2.87484169e-01
9.17833567e-01 -2.19243199e-01 -9.89108324e-01 4.28274393e-01
8.38715494e-01 -1.36269778e-01 -3.69319022e-02 4.16855574e-01
1.25943512e-01 -8.02164078e-01 -4.32406247e-01 -3.55811834e-01
-1.85641110e-01 5.54822803e-01 5.34874737e-01 8.99602413e-01
8.57008278e-01 -3.02566767e-01 8.18335712e-01 9.66778398e-01
5.72402716e-01 -1.68019545e+00 -2.13140488e-01 6.89524531e-01
9.12339568e-01 -9.28350806e-01 5.44190466e-01 -7.27165341e-01
-4.14997995e-01 1.32979155e+00 4.68304873e-01 -1.55325383e-01
1.27306652e+00 5.74917436e-01 1.14069022e-02 -3.70670199e-01
-9.44949508e-01 1.12705544e-01 3.75366658e-01 1.11147714e+00
5.06513238e-01 5.07374525e-01 -9.66500878e-01 3.21953118e-01
-7.28706717e-01 -2.34555230e-01 4.90474731e-01 1.41972518e+00
-6.43810987e-01 -1.20391917e+00 -7.23931909e-01 4.61561680e-01
-5.49773455e-01 -1.16372630e-01 -5.28106689e-01 8.46754074e-01
-5.76629937e-02 1.49448860e+00 -1.97535474e-03 -5.03491640e-01
4.11356747e-01 1.41089618e-01 7.21762300e-01 -1.23501487e-01
-9.04846430e-01 4.15413082e-01 3.84890139e-01 -1.23056793e+00
-8.58632207e-01 -1.05834293e+00 -1.40667629e+00 -1.19437599e+00
-5.12782812e-01 1.89310342e-01 3.83317530e-01 3.90289724e-01
-2.06836104e-01 2.85260603e-02 9.80917513e-01 -1.07949340e+00
-3.90341938e-01 -9.39418614e-01 -1.00026262e+00 -7.84516707e-02
2.89751232e-01 -8.00943017e-01 -7.83523321e-01 2.83909619e-01]
|
[-7.331573963165283, 3.6997148990631104]
|
8bb3c6d4-8930-4bce-8352-2f54572f6166
|
review-of-deep-learning-based-malware
|
2307.01494
| null |
https://arxiv.org/abs/2307.01494v1
|
https://arxiv.org/pdf/2307.01494v1.pdf
|
Review of Deep Learning-based Malware Detection for Android and Windows System
|
Differentiating malware is important to determine their behaviors and level of threat; as well as to devise defensive strategy against them. In response, various anti-malware systems have been developed to distinguish between different malwares. However, most of the recent malware families are Artificial Intelligence (AI) enable and can deceive traditional anti-malware systems using different obfuscation techniques. Therefore, only AI-enabled anti-malware system is robust against these techniques and can detect different features in the malware files that aid in malicious activities. In this study we review two AI-enabled techniques for detecting malware in Windows and Android operating system, respectively. Both the techniques achieved perfect accuracy in detecting various malware families.
|
['Seokjoo Shin', 'Nazmul Islam']
|
2023-07-04
| null | null | null | null |
['malware-detection']
|
['miscellaneous']
|
[ 2.24177599e-01 -4.92980421e-01 -3.94230217e-01 1.22341290e-01
1.58725902e-01 -1.14349306e+00 8.38227391e-01 1.25762984e-01
-7.56557006e-03 5.92982948e-01 -4.50809598e-01 -8.19493890e-01
8.17131177e-02 -7.11539328e-01 -8.22255164e-02 -7.99349129e-01
-4.43359971e-01 1.70238227e-01 2.67469198e-01 -3.34379345e-01
9.13600922e-01 8.07431996e-01 -1.46574891e+00 3.75709653e-01
9.39978004e-01 8.65285158e-01 -1.43096760e-01 1.29427671e+00
-2.86923110e-01 8.50077510e-01 -1.07087004e+00 -1.84339687e-01
3.64544719e-01 -2.77334720e-01 -7.44739652e-01 -3.90271991e-01
-2.82358774e-03 -3.06036919e-01 -1.48997664e-01 1.58749771e+00
-1.36057228e-01 -5.79742014e-01 9.81971860e-01 -1.83471334e+00
-7.44126856e-01 4.95828778e-01 -7.36902654e-01 7.92867959e-01
6.86931610e-01 4.26783264e-01 2.97577351e-01 -2.44379774e-01
-2.01486740e-02 1.28481364e+00 5.78005135e-01 1.12094986e+00
-6.47948921e-01 -1.13372147e+00 -5.06130099e-01 2.21603632e-01
-1.23020911e+00 -4.41657782e-01 7.68601775e-01 -7.49132097e-01
9.51813817e-01 6.58819795e-01 4.84811336e-01 1.32457495e+00
9.13529396e-01 3.94624174e-01 1.37083042e+00 -2.01740846e-01
2.90209800e-01 4.48557556e-01 5.18821836e-01 8.61648858e-01
7.72369683e-01 2.30633184e-01 7.61457160e-02 -9.09262776e-01
8.27505961e-02 3.31278116e-01 -2.38107979e-01 3.17804277e-01
-6.83949649e-01 1.19587135e+00 1.67564824e-01 5.34551024e-01
-1.73732907e-01 -3.50682706e-01 9.82403100e-01 1.94281474e-01
5.13207428e-02 7.52449214e-01 -5.25260627e-01 -1.85708538e-01
-2.70719111e-01 -1.77635655e-01 8.02973509e-01 3.92402500e-01
2.53065228e-01 5.85665643e-01 3.56296986e-01 1.51426017e-01
4.94067341e-01 5.81615269e-01 1.13430774e+00 -6.34457290e-01
-4.14056703e-02 7.65686095e-01 -3.60749483e-01 -1.46944475e+00
3.05680204e-02 2.62982070e-01 -6.46941841e-01 4.72209275e-01
-5.10850064e-02 -7.04291016e-02 -1.09690666e+00 9.94316518e-01
3.35836530e-01 3.30434144e-01 4.23639305e-02 -1.01066642e-01
6.41925573e-01 7.01621115e-01 1.47437453e-02 -3.69978517e-01
1.60996866e+00 -5.30172229e-01 -6.32098138e-01 -2.70768553e-01
6.92995191e-01 -5.48339367e-01 8.37014198e-01 5.13834298e-01
-4.39018905e-01 -2.50145286e-01 -1.27973473e+00 7.71518052e-01
-9.94525790e-01 -5.21476626e-01 5.00679255e-01 1.60972250e+00
-9.60653484e-01 2.52242744e-01 -7.70160258e-01 -2.95328081e-01
5.31646967e-01 7.25901783e-01 -2.63925821e-01 5.44242620e-01
-8.73810053e-01 7.39481032e-01 7.05252707e-01 -6.15009606e-01
-1.17842782e+00 -1.93238467e-01 -5.47256887e-01 -1.98910370e-01
2.62708336e-01 -8.43397826e-02 1.11426222e+00 -1.30328536e+00
-1.26631057e+00 8.01535428e-01 2.44584113e-01 -7.26163864e-01
-6.02520071e-02 1.91055283e-01 -8.72217000e-01 1.94031879e-01
-7.06408871e-03 1.85364798e-01 1.63310790e+00 -1.19799995e+00
-6.35111034e-01 -6.20703042e-01 -7.88423941e-02 -5.03292441e-01
-6.91027880e-01 5.07686436e-01 5.76375902e-01 -5.37901759e-01
-5.94081461e-01 -1.03279150e+00 2.96724111e-01 -5.39359331e-01
-6.16366565e-01 -4.95732436e-03 2.10736585e+00 -6.69373751e-01
1.37077630e+00 -2.00199366e+00 -3.82882267e-01 2.70008415e-01
6.23246193e-01 1.32299149e+00 1.07620686e-01 1.33342519e-01
1.87953580e-02 8.19660187e-01 -1.78438619e-01 5.16839385e-01
-6.76008821e-01 2.28560135e-01 -4.97967750e-01 5.30686200e-01
1.73856542e-01 5.50932586e-01 -6.95644200e-01 -5.33842027e-01
2.27592826e-01 3.54974031e-01 -3.06290388e-01 1.81388065e-01
-3.20192836e-02 5.01010597e-01 -9.24201310e-01 1.34014332e+00
6.64033592e-01 2.49937490e-01 9.24817622e-02 3.03428203e-01
2.32472062e-01 1.91321552e-01 -3.67725194e-01 2.13992875e-02
-1.53517991e-01 7.49247432e-01 3.75190854e-01 -8.42697561e-01
6.82848990e-01 3.08844149e-01 1.71362549e-01 -1.39457192e-02
8.48606408e-01 3.39388937e-01 7.37741172e-01 -6.56245291e-01
4.28711951e-01 5.71171880e-01 1.67266130e-01 4.61727023e-01
-3.82819921e-01 5.53160198e-02 4.47898060e-02 5.34684435e-02
1.60574734e+00 -7.19246209e-01 1.02927971e+00 -2.94277102e-01
1.34105444e+00 8.99656713e-02 2.67635196e-01 6.74571872e-01
-9.78281081e-01 -3.50528449e-01 2.18947396e-01 -3.10404539e-01
-6.15075827e-01 -8.08591306e-01 3.37979533e-02 9.51035440e-01
-5.84760718e-02 -3.69599968e-01 -1.12304783e+00 -1.51327860e+00
-1.82761364e-02 5.83229840e-01 -5.79963565e-01 -5.20046771e-01
-6.39663935e-01 -6.07154012e-01 1.22941816e+00 1.96144819e-01
8.25351000e-01 -1.14150298e+00 -7.07890451e-01 -1.57136530e-01
3.12715113e-01 -8.49065185e-01 -1.63844332e-01 1.89294055e-01
-8.70282412e-01 -1.47096384e+00 4.62408066e-02 -7.08180785e-01
6.19194925e-01 6.25647426e-01 7.18161106e-01 6.51232243e-01
-3.37577164e-01 2.90278018e-01 -4.90027905e-01 -7.78144777e-01
-1.28120232e+00 -1.37882888e-01 3.99415910e-01 -1.62395254e-01
7.88271904e-01 -2.83569872e-01 -5.50095811e-02 2.12301388e-01
-9.16308403e-01 -8.94657671e-01 4.38892901e-01 4.38337386e-01
-1.65362597e-01 8.15125465e-01 2.34887853e-01 -9.49442267e-01
1.18837917e+00 -1.02362561e+00 -3.38705897e-01 -8.59377831e-02
-6.35826647e-01 -1.29228026e-01 1.36631596e+00 -1.04202187e+00
-6.04210973e-01 -4.70568500e-02 -1.77062884e-01 -2.24598467e-01
-4.33510929e-01 -1.03225283e-01 -2.61933088e-01 -7.42510438e-01
8.36842000e-01 4.38785672e-01 2.44442806e-01 -1.89842477e-01
-1.52043909e-01 1.34056199e+00 3.83213252e-01 -2.15479821e-01
1.14509630e+00 3.05635393e-01 6.51393980e-02 -1.14671373e+00
-1.14943698e-01 -3.50357741e-01 -2.65704006e-01 -8.28818902e-02
7.87136197e-01 -1.15062088e-01 -8.81415129e-01 7.93569267e-01
-1.13523328e+00 3.64421368e-01 5.62123895e-01 -1.16436794e-01
6.92342361e-03 6.71959639e-01 -7.66285181e-01 -1.07853746e+00
-7.16352463e-01 -1.41976023e+00 5.73822379e-01 3.31278682e-01
-3.49783868e-01 -9.10997033e-01 1.91251844e-01 1.03333272e-01
5.55131614e-01 3.41766894e-01 9.89954948e-01 -1.35152495e+00
1.73402913e-02 -4.67797637e-01 -1.17349327e-01 3.77483010e-01
7.68546164e-01 8.28697503e-01 -9.34095085e-01 -2.17328042e-01
6.68514967e-01 1.12794220e-01 3.88483942e-01 1.33468166e-01
1.20479274e+00 -1.05778158e+00 -8.12225878e-01 4.70839143e-01
1.07959020e+00 1.26925969e+00 4.99380708e-01 4.15630043e-01
8.48760962e-01 3.81952465e-01 5.86526871e-01 7.12079406e-02
-2.32471824e-01 2.13737100e-01 7.77886748e-01 6.80348396e-01
6.72741473e-01 2.21817374e-01 9.30167377e-01 6.78716183e-01
-1.13922447e-01 -1.84326202e-01 -1.04355693e+00 -1.19534127e-01
-1.27531636e+00 -1.11389184e+00 -2.76684761e-01 2.00007820e+00
5.55659294e-01 5.19891977e-01 5.90043306e-01 5.70679426e-01
8.45607102e-01 1.75863162e-01 -5.30709445e-01 -1.29417694e+00
2.11959288e-01 2.01705709e-01 6.67935610e-01 2.66523689e-01
-1.41581511e+00 8.04843187e-01 7.03344250e+00 1.04366267e+00
-1.48815727e+00 2.27681577e-01 4.80379254e-01 4.62002575e-01
4.06502217e-01 -1.48987740e-01 -6.24428391e-01 9.96246099e-01
1.40120125e+00 -2.89690822e-01 5.05941689e-01 1.40237653e+00
-2.18359232e-01 2.72652414e-03 -6.22155428e-01 7.45735168e-01
2.53631026e-01 -9.84487474e-01 7.17591271e-02 5.00908852e-01
3.14524829e-01 -2.89697498e-01 2.57190317e-01 2.84581095e-01
3.31083953e-01 -1.08846450e+00 -6.68321997e-02 -2.92235225e-01
3.18558693e-01 -1.02461052e+00 7.88238466e-01 4.41819549e-01
-1.29690504e+00 -7.12030649e-01 -1.49480984e-01 -2.80602723e-01
-4.40191984e-01 -4.55752797e-02 -1.10433877e+00 -1.64118752e-01
7.20035195e-01 3.49950433e-01 -9.39824343e-01 5.00406623e-01
3.20577770e-02 7.99237251e-01 3.97132337e-02 -4.29926634e-01
1.66000262e-01 2.29688194e-02 9.61592436e-01 9.80009437e-01
1.94119066e-01 -3.36647108e-02 1.91057354e-01 4.55643922e-01
3.74211788e-01 -2.00163513e-01 -1.26189315e+00 -9.29514408e-01
7.79791713e-01 1.27249277e+00 -9.50306833e-01 -5.25164366e-01
-1.54234007e-01 8.82848799e-01 -2.15504393e-01 -2.56007969e-01
-9.34173167e-01 -6.85188055e-01 1.17655385e+00 2.31161013e-01
-1.98644131e-01 -3.81806254e-01 -2.47398213e-01 -8.03618491e-01
-6.30128205e-01 -1.75773942e+00 4.82083350e-01 -2.23320261e-01
-1.17537856e+00 9.53967929e-01 -3.11577693e-02 -1.16505790e+00
-4.38280642e-01 -9.57048059e-01 -1.02043271e+00 2.13638902e-01
-5.70153892e-01 -1.03981841e+00 -5.65417111e-02 6.26485646e-01
6.05934262e-01 -8.82354975e-01 7.59738266e-01 -4.83455472e-02
-8.14251065e-01 4.65363204e-01 9.63902450e-04 2.72446603e-01
1.27747282e-01 -8.62112403e-01 3.01092863e-01 9.23907042e-01
-2.13559464e-01 1.09152639e+00 7.81201959e-01 -1.08616877e+00
-1.79551423e+00 -8.51075768e-01 1.59612253e-01 -6.14139497e-01
8.28633070e-01 -1.01078674e-01 -1.01318800e+00 8.06764960e-01
2.37720981e-01 -2.20953584e-01 5.90678334e-01 -6.27498806e-01
-6.26576126e-01 2.91850835e-01 -1.84968460e+00 6.39893115e-01
5.60671806e-01 -5.16451120e-01 -3.80013824e-01 1.76566720e-01
6.95973277e-01 6.77722916e-02 -2.06695661e-01 3.80247980e-01
3.28544348e-01 -1.18151903e+00 1.17476344e+00 -7.00018883e-01
6.04464114e-02 -3.66253227e-01 5.64612187e-02 -7.68085778e-01
-2.37113684e-01 -7.65833616e-01 -8.52411985e-01 1.22669959e+00
-2.28945106e-01 -1.22878885e+00 6.17701769e-01 1.61679491e-01
2.73383170e-01 -6.54014826e-01 -5.98100543e-01 -9.01388228e-01
-2.56530195e-01 -9.83755141e-02 6.09996378e-01 1.20768774e+00
-1.28549978e-01 1.80443689e-01 -2.79829890e-01 1.44733801e-01
5.37714303e-01 -6.94396317e-01 7.47125983e-01 -1.22471094e+00
-2.92116314e-01 -8.70297015e-01 -9.29821134e-01 -2.07859650e-02
2.50254810e-01 -4.64866310e-01 -4.33521241e-01 -2.21982703e-01
1.62721694e-01 3.17112170e-02 3.57457176e-02 4.66011375e-01
-1.33204356e-01 4.53794003e-01 -2.31012590e-02 4.17505383e-01
4.55717333e-02 -1.54573441e-01 2.82756686e-01 -3.23475271e-01
-3.32306176e-01 2.28492051e-01 -7.48216450e-01 1.30326390e+00
1.31275392e+00 -7.11775124e-01 -4.43602234e-01 3.25955987e-01
-2.60270596e-01 -5.97965658e-01 1.68371469e-01 -9.64394510e-01
-3.40449721e-01 -6.30840719e-01 1.98327020e-01 -4.75328922e-01
9.52411518e-02 -8.35353315e-01 -1.39313772e-01 1.26452351e+00
3.49900126e-01 5.42176604e-01 2.47638151e-01 5.39774656e-01
1.36060759e-01 -5.53241253e-01 1.14035273e+00 -2.84680724e-01
-7.41329849e-01 9.55270529e-02 -1.12740302e+00 -2.30502367e-01
1.69112718e+00 -6.57663465e-01 -7.01930046e-01 -1.62660524e-01
-1.19776127e-03 -3.67058605e-01 6.90699100e-01 3.75224024e-01
5.40375590e-01 -9.15495753e-01 -2.13912830e-01 3.57784599e-01
-1.13048032e-01 -9.99348342e-01 -3.68993282e-01 6.35342002e-01
-9.82888222e-01 5.17761886e-01 -9.01453853e-01 -4.88771170e-01
-2.36561799e+00 1.26363325e+00 1.01046570e-01 -1.49476737e-01
7.98593685e-02 5.95136821e-01 -3.67048196e-02 -8.07993039e-02
1.37792856e-01 2.56755441e-01 -6.75572395e-01 -4.16790873e-01
9.83215332e-01 7.48063684e-01 -3.16055179e-01 -1.11381710e+00
-8.39773238e-01 2.94901967e-01 -4.08103794e-01 4.09335554e-01
5.35697341e-01 1.03661209e-01 -4.62338924e-01 -1.05132035e-03
1.26143754e+00 4.63023216e-01 -1.28440320e-01 4.61839080e-01
3.43693972e-01 -9.02634323e-01 -2.92064697e-01 -5.44934750e-01
-8.13420117e-01 7.58161366e-01 7.38873839e-01 1.23769963e+00
1.37335074e+00 -5.10032237e-01 8.78609955e-01 2.56475210e-01
3.09017450e-01 -5.75177848e-01 2.43099213e-01 5.67502677e-01
4.31637913e-01 -1.41690505e+00 6.89293593e-02 -6.18310809e-01
-4.72707689e-01 1.16520786e+00 1.00417995e+00 -2.21842006e-01
6.39218450e-01 6.90314293e-01 1.22754192e-02 -2.79928714e-01
-3.51150393e-01 2.91724294e-01 3.55979592e-01 1.32647443e+00
3.27671915e-01 2.20897466e-01 -4.94969398e-01 1.55843750e-01
-1.16250090e-01 -6.21996284e-01 7.46923983e-01 1.41834271e+00
-7.73199677e-01 -1.18210125e+00 -1.00005198e+00 7.11600959e-01
-9.54012513e-01 3.36739756e-02 -1.59019005e+00 7.34981537e-01
1.29369333e-01 1.18088198e+00 -2.28771865e-01 -1.29107320e+00
-5.51347375e-01 -6.09132238e-02 9.96624082e-02 -5.58027565e-01
-1.05524671e+00 -4.91893977e-01 6.46109059e-02 -2.27206603e-01
3.90701257e-02 -1.18431665e-01 -1.16819215e+00 -9.96236324e-01
-4.64174867e-01 1.23324297e-01 7.34579682e-01 7.91583776e-01
5.71445048e-01 1.34218827e-01 8.44816029e-01 -8.70958745e-01
-5.14026046e-01 -9.08440053e-01 -2.05099002e-01 1.46026194e-01
5.10394990e-01 -7.55287945e-01 -6.45736277e-01 -3.30640078e-01]
|
[14.412065505981445, 9.676471710205078]
|
9ea5b67f-fafd-4338-9dc0-cb031c76d7b0
|
cdistnet-perceiving-multi-domain-character
|
2111.11011
| null |
https://arxiv.org/abs/2111.11011v3
|
https://arxiv.org/pdf/2111.11011v3.pdf
|
CDistNet: Perceiving Multi-Domain Character Distance for Robust Text Recognition
|
The Transformer-based encoder-decoder framework is becoming popular in scene text recognition, largely because it naturally integrates recognition clues from both visual and semantic domains. However, recent studies show that the two kinds of clues are not always well registered and therefore, feature and character might be misaligned in the difficult text (e.g., with rare shapes). As a result, constraints such as character position are introduced to alleviate this problem. Despite certain success, a content-free positional embedding hardly associates stably with meaningful local image regions. In this paper, we propose a novel module called Multi-Domain Character Distance Perception (MDCDP) to establish a visual and semantic related positional encoding. MDCDP uses positional embedding to query both visual and semantic features following the attention mechanism. The two kinds of constrained features are then fused to produce a reinforced feature, generating a content-aware embedding that well perceives spacing variations and semantic affinities among characters, i.e., multi-domain character distance. We develop a novel network named CDistNet that stacks multiple MDCDPs to guide a gradually precise distance modeling. Thus, the feature-character alignment is well built even various recognition difficulties presented. We create two series of augmented datasets with increasing recognition difficulties and apply CDistNet to both them and six public benchmarks. The experiments demonstrate that CDistNet outperforms recent popular methods by large margins in challenging recognition scenarios. It also achieves state-of-the-art accuracy on standard benchmarks. In addition, the visualization shows that CDistNet achieves proper information utilization in both visual and semantic domains. Our code is given in https://github.com/simplify23/CDistNet.
|
['Yu-Gang Jiang', 'Hongtao Xie', 'Shancheng Fang', 'Zhineng Chen', 'Tianlun Zheng']
|
2021-11-22
| null | null | null | null |
['scene-text-recognition']
|
['computer-vision']
|
[ 1.56764269e-01 -6.67764008e-01 -1.77228957e-01 -5.03344059e-01
-6.25121355e-01 -5.50003767e-01 7.06008255e-01 3.52991112e-02
-3.75196159e-01 2.85020709e-01 4.49918061e-01 9.90907997e-02
-3.98338512e-02 -7.32438505e-01 -5.66604972e-01 -6.92937553e-01
6.78908110e-01 3.80349070e-01 3.74866575e-01 -2.24653140e-01
6.51988566e-01 3.73770088e-01 -1.46489942e+00 4.76966023e-01
1.01567221e+00 8.40754271e-01 5.86672008e-01 4.40886736e-01
-4.76114213e-01 4.27403867e-01 -5.61231792e-01 -4.56642658e-01
9.97650474e-02 -3.65017563e-01 -6.26725078e-01 8.31701979e-02
5.24952710e-01 -3.56752723e-01 -5.93964934e-01 1.16205156e+00
8.19371223e-01 8.94423723e-02 5.95501482e-01 -1.03548181e+00
-1.36039233e+00 6.25236154e-01 -7.06328273e-01 1.97177544e-01
5.81661820e-01 1.36886075e-01 1.19116712e+00 -1.35121036e+00
3.55799437e-01 1.22361052e+00 3.69372785e-01 2.94837922e-01
-8.49448919e-01 -4.85246450e-01 3.27106297e-01 5.94221711e-01
-1.63940144e+00 -1.49211377e-01 9.02332306e-01 -2.73715675e-01
1.01743758e+00 3.27833235e-01 5.58918953e-01 1.27121389e+00
-4.60783876e-02 9.65925992e-01 8.22519541e-01 -2.80040920e-01
-2.18292952e-01 1.28137752e-01 2.20996082e-01 5.61985910e-01
1.75316304e-01 -2.04503968e-01 -6.73154771e-01 3.23447883e-01
7.22898841e-01 1.79218546e-01 -3.69777828e-01 -2.18790606e-01
-1.56666791e+00 4.45556819e-01 5.51189780e-01 5.47610819e-01
8.15255269e-02 -1.18732296e-01 4.09261674e-01 1.24024868e-01
1.82604283e-01 2.54347414e-01 -1.58404604e-01 -2.93676704e-01
-5.89798629e-01 -9.16087814e-03 3.44594479e-01 1.02846587e+00
7.01297998e-01 2.95884032e-02 -3.59296024e-01 1.31437743e+00
2.80070066e-01 6.14039242e-01 6.96036756e-01 -3.02904129e-01
6.61812007e-01 9.32992995e-01 -2.36323893e-01 -1.50309217e+00
-2.53036261e-01 -3.76130372e-01 -1.00868344e+00 -1.76141277e-01
2.72254109e-01 4.51961219e-01 -7.15324104e-01 1.58834994e+00
1.06572911e-01 1.53683975e-01 -3.24522220e-02 1.20247006e+00
1.12637889e+00 7.60459542e-01 -2.64735430e-01 2.76071191e-01
1.43765306e+00 -1.17218328e+00 -6.06267869e-01 -2.83973843e-01
5.99904776e-01 -1.05989373e+00 1.38703144e+00 3.04697871e-01
-8.33719194e-01 -8.57513309e-01 -1.24622154e+00 -3.50883782e-01
-5.66322863e-01 3.95534217e-01 2.54460633e-01 3.00001979e-01
-7.71637082e-01 4.59146678e-01 -4.58918720e-01 -4.08259451e-01
2.72347510e-01 1.11469999e-01 -3.66656035e-01 -3.88911366e-01
-1.13700366e+00 8.70599747e-01 4.34796810e-01 2.52809674e-01
-6.04790628e-01 -3.80751789e-01 -8.13463569e-01 -1.00344583e-01
1.82895228e-01 -4.33597237e-01 7.79236138e-01 -8.48600507e-01
-1.24728441e+00 9.55915630e-01 -1.41656354e-01 2.67388254e-01
5.67883372e-01 -2.11307868e-01 -7.13097036e-01 -1.05670979e-02
1.62084758e-01 6.43558860e-01 7.74380505e-01 -1.05991614e+00
-4.77274746e-01 -2.76531994e-01 -9.83347297e-02 7.38762558e-01
-6.20795131e-01 4.56101596e-02 -9.21791315e-01 -8.39394689e-01
3.34501654e-01 -5.41590273e-01 3.55469644e-01 2.77436733e-01
-6.97113574e-01 -3.05515379e-01 9.03901517e-01 -5.18897295e-01
1.15097666e+00 -2.25222445e+00 3.60399991e-01 5.68854585e-02
2.37088025e-01 2.23692194e-01 -3.26970309e-01 4.01413321e-01
-5.89007102e-02 6.85861558e-02 -2.28386402e-01 -2.71744549e-01
6.18947148e-02 2.23259896e-01 -1.90964147e-01 4.50200379e-01
2.05652460e-01 1.02769506e+00 -8.25654984e-01 -5.27966201e-01
3.18941027e-01 4.83634651e-01 -3.73989910e-01 2.09320769e-01
-4.97213341e-02 3.00890744e-01 -4.90941733e-01 8.26491475e-01
9.20424461e-01 -4.52930480e-01 1.21875841e-03 -4.51931000e-01
-1.16461478e-01 9.97659341e-02 -1.24266267e+00 1.93744957e+00
-1.22520983e-01 6.02850080e-01 -4.53581512e-01 -1.07160497e+00
1.32394886e+00 -1.73558831e-01 2.15973571e-01 -9.06222999e-01
1.91063181e-01 2.66581714e-01 -6.04237653e-02 -5.50587952e-01
6.71981335e-01 2.53683686e-01 -2.25555524e-02 2.94761747e-01
-8.99014995e-02 2.22701374e-02 -7.10569173e-02 1.92349181e-01
7.82544792e-01 6.55526072e-02 2.55630732e-01 -2.07317293e-01
8.06894183e-01 -1.56400666e-01 5.26150882e-01 4.44158465e-01
-2.07202137e-01 9.42713022e-01 3.66511971e-01 -3.48813713e-01
-1.14639735e+00 -9.83371258e-01 -2.95709074e-01 1.07183623e+00
8.43292058e-01 -5.09972155e-01 -4.08941776e-01 -4.92757946e-01
4.43854220e-02 3.28033864e-01 -5.06228447e-01 -2.77194381e-01
-4.47833508e-01 -7.29042947e-01 6.80724204e-01 6.92679703e-01
8.50471854e-01 -1.01178133e+00 -1.46620423e-02 6.63158763e-03
-1.68594629e-01 -1.15303957e+00 -8.44111264e-01 -7.83900470e-02
-4.69079792e-01 -1.00600457e+00 -9.39248383e-01 -1.23902965e+00
6.70255303e-01 5.70982814e-01 8.67446780e-01 2.98933953e-01
-2.91882306e-01 9.96624753e-02 -6.20835304e-01 1.36686057e-01
4.97182198e-02 7.49169907e-04 -4.60075326e-02 7.81333596e-02
6.80752397e-01 -4.43759203e-01 -6.29931867e-01 6.47409439e-01
-7.95924008e-01 3.10093105e-01 6.42333686e-01 1.04694247e+00
6.78669035e-01 -2.61787474e-01 3.00200135e-01 -4.32279885e-01
7.44022608e-01 -3.17229569e-01 -3.43282729e-01 4.79442120e-01
-4.40936506e-01 1.21576916e-02 8.23247552e-01 -6.16950870e-01
-7.64140129e-01 -1.46797657e-01 -2.03690469e-01 -2.80171871e-01
-1.65058494e-01 5.04374027e-01 -6.32133245e-01 -9.70347673e-02
4.83591914e-01 8.52485657e-01 -1.77717075e-01 -6.12345934e-01
2.93654412e-01 8.44296873e-01 6.18691921e-01 -7.23722994e-01
7.54702866e-01 1.84008420e-01 -4.21548963e-01 -7.98399866e-01
-6.52236521e-01 -2.91693240e-01 -6.96141779e-01 1.25730522e-02
7.53844321e-01 -8.26872408e-01 -5.01231253e-01 8.67692411e-01
-1.23998487e+00 2.69210842e-02 -1.57496450e-03 4.63366479e-01
-2.60108113e-01 7.86039174e-01 -6.01662755e-01 -1.86843038e-01
-1.52408540e-01 -1.29439032e+00 1.12317741e+00 3.80106539e-01
1.18160143e-01 -9.02918279e-01 -2.86916308e-02 2.32857227e-01
2.48168290e-01 -1.29108071e-01 8.35494041e-01 -6.63064778e-01
-6.81721270e-01 4.75366600e-02 -8.03620636e-01 3.02647650e-01
2.62494475e-01 2.28597149e-01 -8.48649442e-01 -2.42268652e-01
-3.20577770e-01 -3.32075983e-01 8.68787289e-01 2.34739501e-02
1.44976842e+00 -5.50193004e-02 -2.07587674e-01 9.73865509e-01
1.32915747e+00 1.18891038e-01 7.73344576e-01 4.67940331e-01
1.12668478e+00 4.32677507e-01 6.10820830e-01 5.64021111e-01
5.42018592e-01 8.84648561e-01 3.28074694e-01 -4.20153514e-02
-2.27350593e-01 -4.19789702e-01 2.77724981e-01 1.21689212e+00
1.58375040e-01 -4.15987968e-01 -8.98432910e-01 4.29916799e-01
-1.86969447e+00 -8.44656408e-01 -1.10364206e-01 1.89632976e+00
9.26441908e-01 5.07564433e-02 -1.20823547e-01 9.81582031e-02
9.98143911e-01 3.90772671e-01 -7.60335863e-01 -3.07589442e-01
-8.28901887e-01 -1.81046978e-01 2.28120759e-01 3.97413999e-01
-8.86197865e-01 1.01238084e+00 5.15627766e+00 1.13522005e+00
-1.16095686e+00 -1.09537788e-01 5.14319658e-01 2.21338958e-01
-5.98409235e-01 -1.28344223e-01 -9.33548629e-01 6.98253930e-01
1.38497606e-01 -3.01252872e-01 3.76443177e-01 8.04564178e-01
-1.43180296e-01 2.98391581e-01 -1.07243741e+00 1.54055095e+00
4.68798369e-01 -1.35475671e+00 3.99007231e-01 -6.45349547e-02
5.25677562e-01 -1.28931805e-01 2.17246160e-01 2.04438530e-02
-9.71415043e-02 -1.24378061e+00 7.69795597e-01 5.31706095e-01
1.20567513e+00 -6.67747676e-01 6.45596921e-01 2.01123476e-01
-1.52436936e+00 1.14629172e-01 -7.17217863e-01 1.59852445e-01
1.74912121e-02 5.01325548e-01 -4.20156896e-01 6.27867639e-01
6.20569766e-01 1.40385365e+00 -8.37480903e-01 1.03241038e+00
-2.80882806e-01 9.73300412e-02 -2.30092555e-01 -1.78338602e-01
4.06110555e-01 -2.26108342e-01 4.14948046e-01 1.30452263e+00
4.23769772e-01 -6.60076812e-02 1.17144741e-01 9.04650569e-01
-8.44300687e-02 3.92479360e-01 -5.80813169e-01 1.10075898e-01
7.04895496e-01 1.17347860e+00 -5.40786326e-01 -2.88274318e-01
-5.29443979e-01 1.40625131e+00 4.38566238e-01 2.00795501e-01
-1.04420245e+00 -4.82365310e-01 7.41088033e-01 -2.92854249e-01
3.50567997e-01 -1.40173361e-01 -4.18265611e-01 -1.42813218e+00
2.99336284e-01 -1.04025996e+00 2.95972526e-01 -8.79803896e-01
-1.51971602e+00 6.68482244e-01 -3.18544686e-01 -1.49451458e+00
3.47044051e-01 -8.97696495e-01 -6.53545678e-01 9.03323889e-01
-1.52394545e+00 -1.13907444e+00 -7.43248165e-01 7.77977705e-01
6.94336355e-01 -3.27554822e-01 6.85441077e-01 4.93749797e-01
-8.30919743e-01 1.06966496e+00 2.89858520e-01 3.19714487e-01
9.91026103e-01 -1.10457110e+00 4.21085566e-01 8.23616803e-01
2.52257049e-01 4.52069193e-01 2.74074972e-01 -5.57114959e-01
-1.50548387e+00 -9.45533931e-01 7.68180251e-01 -4.41811532e-01
5.28033495e-01 -4.72836047e-01 -1.18082261e+00 3.19997132e-01
9.60793570e-02 5.45848384e-02 4.53128129e-01 -2.01864958e-01
-7.46398211e-01 -1.34082645e-01 -8.08690846e-01 7.12534726e-01
1.40028667e+00 -7.62614250e-01 -6.38776243e-01 2.12163389e-01
7.76214719e-01 -4.76724714e-01 -6.90765679e-01 2.65106291e-01
5.03345013e-01 -1.01700950e+00 1.00118363e+00 -2.49633342e-01
5.71526587e-01 -5.98973691e-01 -5.59740961e-01 -1.13455844e+00
-5.52493453e-01 -2.34014541e-01 1.02590576e-01 1.34788311e+00
3.24916661e-01 -6.23460948e-01 5.97038090e-01 -6.64871745e-03
-2.65541732e-01 -8.09212685e-01 -8.36774468e-01 -8.13926816e-01
1.38263583e-01 -3.82453531e-01 6.79104388e-01 1.20012605e+00
-1.89276109e-03 4.19703871e-01 -5.03955841e-01 5.52340634e-02
3.43351692e-01 2.60647893e-01 5.20184159e-01 -9.34260130e-01
-6.67155534e-02 -7.19724536e-01 -7.25196540e-01 -1.76647878e+00
-2.72270525e-04 -1.08354092e+00 -1.39381543e-01 -1.43269825e+00
4.87705857e-01 -4.66001451e-01 -3.76097143e-01 4.14293319e-01
-2.22455338e-01 2.61865854e-01 1.37942120e-01 3.60959411e-01
-6.77341402e-01 9.22717154e-01 1.58905447e+00 -3.99216652e-01
1.24737740e-01 -6.29981101e-01 -8.14828038e-01 5.69875121e-01
9.36296284e-01 -7.98368901e-02 -3.18726361e-01 -8.26995134e-01
2.52141118e-01 -4.36230868e-01 3.62524837e-01 -9.78564143e-01
4.76488143e-01 -1.20060004e-01 6.80257201e-01 -9.30650294e-01
3.81128132e-01 -6.83279276e-01 -1.71436056e-01 1.77630290e-01
-3.23935181e-01 3.29968452e-01 -2.60423589e-03 5.07287681e-01
-4.61718380e-01 -7.22527280e-02 6.92972243e-01 1.16770454e-01
-1.19182622e+00 3.09944600e-01 1.25122905e-01 2.54042447e-01
9.29950416e-01 -6.45183682e-01 -7.67828465e-01 -2.23373190e-01
-3.43647361e-01 2.61837691e-01 7.11016536e-01 7.10725188e-01
1.10245812e+00 -1.64727497e+00 -8.18678081e-01 3.99287790e-01
5.56214631e-01 -8.21176320e-02 4.28426981e-01 7.47090876e-01
-5.51839769e-01 2.98304170e-01 -3.41713399e-01 -8.36614728e-01
-1.17791402e+00 5.18596232e-01 3.87211293e-01 7.50251189e-02
-7.76735306e-01 9.78460848e-01 3.26262534e-01 -5.71530640e-01
3.27349126e-01 -2.72085935e-01 -2.52797872e-01 6.40120804e-02
5.99275768e-01 1.51830330e-01 -1.47287101e-01 -8.95023644e-01
-6.29561186e-01 1.22771895e+00 -2.76285589e-01 2.75613517e-01
9.98903513e-01 -2.30546013e-01 -1.45650491e-01 3.38395536e-01
1.39456999e+00 -4.66406718e-03 -1.21770418e+00 -6.19617462e-01
-1.01346828e-01 -8.47823083e-01 -3.78177732e-01 -7.11772144e-01
-1.06083560e+00 9.99714851e-01 5.39351225e-01 -1.33346736e-01
1.13269365e+00 -3.14679854e-02 7.82369196e-01 3.53353620e-01
-4.66598384e-03 -1.11222351e+00 5.02486587e-01 7.52645254e-01
1.00265133e+00 -1.38821852e+00 -4.08145748e-02 -3.98043901e-01
-9.34201598e-01 1.19474971e+00 9.87411201e-01 -2.48624906e-02
4.73697066e-01 1.65451735e-01 9.25965533e-02 -5.86858578e-02
-6.32659554e-01 -1.70030147e-01 5.75901926e-01 5.45457006e-01
4.80909228e-01 -1.99806765e-02 -1.91361994e-01 6.16087496e-01
-2.38951981e-01 -5.80402017e-01 2.78545260e-01 7.33699560e-01
-5.39575458e-01 -1.17032778e+00 -2.88417250e-01 3.00762594e-01
-1.06803983e-01 -3.61142099e-01 -5.73880076e-01 6.13390505e-01
2.41623417e-01 5.80296636e-01 2.20147863e-01 -7.50913024e-01
3.00365806e-01 -3.01086515e-01 3.15780133e-01 -4.97982115e-01
-2.53921390e-01 -3.81702147e-02 -2.11919561e-01 -5.88681102e-01
-1.62456915e-01 -3.74284416e-01 -1.24327767e+00 -4.92464095e-01
-2.87200123e-01 -1.50630400e-01 4.29977953e-01 7.05560565e-01
3.63067746e-01 5.20521760e-01 7.38884985e-01 -4.80022430e-01
-4.28852558e-01 -8.66107285e-01 -5.80405831e-01 5.32514870e-01
6.82110190e-02 -6.28194034e-01 -3.02440196e-01 -2.40651011e-01]
|
[11.69140625, 2.0540895462036133]
|
66488eb0-a16f-4d12-9aea-c98c2f890ae9
|
daa-a-delta-age-adain-operation-for-age
|
2303.07929
| null |
https://arxiv.org/abs/2303.07929v1
|
https://arxiv.org/pdf/2303.07929v1.pdf
|
DAA: A Delta Age AdaIN operation for age estimation via binary code transformer
|
Naked eye recognition of age is usually based on comparison with the age of others. However, this idea is ignored by computer tasks because it is difficult to obtain representative contrast images of each age. Inspired by the transfer learning, we designed the Delta Age AdaIN (DAA) operation to obtain the feature difference with each age, which obtains the style map of each age through the learned values representing the mean and standard deviation. We let the input of transfer learning as the binary code of age natural number to obtain continuous age feature information. The learned two groups of values in Binary code mapping are corresponding to the mean and standard deviation of the comparison ages. In summary, our method consists of four parts: FaceEncoder, DAA operation, Binary code mapping, and AgeDecoder modules. After getting the delta age via AgeDecoder, we take the average value of all comparison ages and delta ages as the predicted age. Compared with state-of-the-art methods, our method achieves better performance with fewer parameters on multiple facial age datasets.
|
['Zongjie Jiang', 'Bing Wang', 'Bin Xiao', 'Ju Tao', 'Ye Li', 'Xingpeng Zhang', 'Ping Chen']
|
2023-03-14
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Chen_DAA_A_Delta_Age_AdaIN_Operation_for_Age_Estimation_via_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Chen_DAA_A_Delta_Age_AdaIN_Operation_for_Age_Estimation_via_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['age-estimation', 'age-estimation']
|
['computer-vision', 'miscellaneous']
|
[-9.01319385e-02 -1.86764210e-01 1.19934678e-01 -9.15876687e-01
-1.75217658e-01 -2.31277823e-01 4.66797680e-01 -2.60053873e-01
-6.01560116e-01 6.87147915e-01 -6.66011199e-02 2.05221429e-01
2.59791434e-01 -9.29539680e-01 -4.92692262e-01 -9.63357508e-01
6.21284172e-03 5.61890192e-02 -4.50325459e-02 1.40437841e-01
3.23645294e-01 2.65615880e-01 -1.91858840e+00 9.13807005e-02
1.22015178e+00 1.68233275e+00 -1.69601753e-01 4.23227012e-01
-3.31283271e-01 5.26929975e-01 -8.89237285e-01 -5.12999654e-01
1.03131510e-01 -6.81008697e-01 -2.89767206e-01 -1.20137863e-01
8.54617000e-01 -7.73452342e-01 -2.78844267e-01 1.38127351e+00
5.32045901e-01 -8.00115392e-02 1.15886557e+00 -1.61424172e+00
-1.23609865e+00 3.15038681e-01 -8.69492888e-01 -1.63723320e-01
3.14961851e-01 -1.31720051e-01 2.90344626e-01 -9.44684029e-01
1.30010977e-01 1.51059675e+00 6.20024025e-01 8.85199487e-01
-5.64404368e-01 -1.30120099e+00 5.48061403e-03 4.46967721e-01
-1.36357522e+00 -1.72958583e-01 6.20166004e-01 -4.65277970e-01
2.25053251e-01 -1.11946121e-01 8.62240076e-01 9.13521051e-01
3.47501427e-01 5.52893281e-01 1.50558126e+00 -2.56007105e-01
1.04025967e-01 1.21366404e-01 -2.32712299e-01 8.81806850e-01
-1.11315921e-02 1.11200280e-01 -4.65490162e-01 1.46575004e-01
7.58895040e-01 8.80553052e-02 -1.14033736e-01 6.69079721e-02
-8.42545867e-01 6.14763319e-01 4.23031867e-01 2.31376570e-02
-1.36614978e-01 9.30898562e-02 1.34284720e-01 4.89888906e-01
5.09556174e-01 -3.27562809e-01 -5.50448716e-01 -8.25469121e-02
-7.87668169e-01 4.34686020e-02 4.61382031e-01 7.61803508e-01
1.16640842e+00 6.62719831e-03 3.09366342e-02 8.86049509e-01
2.45135188e-01 7.21781969e-01 5.80339551e-01 -9.57243562e-01
-2.48357952e-02 6.79622173e-01 -2.83508837e-01 -7.52873123e-01
1.15962774e-02 1.09684475e-01 -9.14473712e-01 7.39380836e-01
7.87537038e-01 -2.45122150e-01 -1.09600747e+00 1.96129560e+00
3.41809303e-01 2.67683953e-01 -2.40284819e-02 5.92258096e-01
7.64994919e-01 7.74782479e-01 1.30109578e-01 -5.24049997e-01
1.46317458e+00 -6.65825903e-01 -6.31779015e-01 -3.46994214e-02
2.24223450e-01 -7.81795800e-01 8.95835459e-01 5.94259202e-01
-9.31962490e-01 -7.85303891e-01 -1.00838256e+00 3.30971889e-02
-4.89804983e-01 5.32806516e-01 6.34278953e-01 5.68273842e-01
-1.14035499e+00 7.57222712e-01 -3.51694226e-01 -3.94561559e-01
5.56101024e-01 4.75281656e-01 -5.21108866e-01 2.61738747e-01
-1.09526050e+00 7.48383880e-01 2.82491535e-01 8.53173211e-02
-6.90401196e-01 -9.06442225e-01 -9.58844900e-01 -1.67342439e-01
-2.28103429e-01 -4.25538331e-01 1.05999804e+00 -1.53475595e+00
-1.49393451e+00 1.05752873e+00 -1.10788667e-03 -7.74680078e-02
4.46435601e-01 -2.19104767e-01 -6.48439944e-01 7.05099180e-02
-2.28554055e-01 1.10735810e+00 1.38645768e+00 -8.65898728e-01
-7.44150698e-01 -5.41869760e-01 -3.53916973e-01 7.97662660e-02
-8.94487917e-01 2.17508852e-01 -9.16602835e-02 -5.05787969e-01
-1.32587403e-01 -5.62926352e-01 5.24648547e-01 8.02536726e-01
4.65701342e-01 -4.62252855e-01 9.83844697e-01 -1.07345641e+00
1.34741616e+00 -2.34259200e+00 6.75256737e-03 -4.29326892e-02
2.55057573e-01 -6.82213306e-02 -2.09524184e-02 -3.48630100e-01
-2.51455098e-01 -3.61968607e-01 -1.41088277e-01 9.70738009e-02
-2.50076026e-01 -8.91073495e-02 6.76790029e-02 3.14784080e-01
3.35020393e-01 3.49651128e-01 -8.71220291e-01 -1.00626671e+00
-1.52422667e-01 5.86080134e-01 -3.30138296e-01 6.72840476e-01
8.11765715e-02 3.01644057e-01 -1.86092988e-01 5.73251784e-01
1.10252738e+00 2.70195484e-01 -2.53041595e-01 -3.57903898e-01
-8.17630515e-02 -4.45457309e-01 -9.38995063e-01 1.25900459e+00
-3.14740598e-01 3.17117751e-01 -1.35120839e-01 -5.50913751e-01
1.61233020e+00 8.06220546e-02 3.31282049e-01 -4.99615818e-01
5.37464142e-01 3.01609099e-01 1.12044714e-01 -4.64535952e-01
1.39756873e-01 -1.29921511e-01 8.25212151e-02 5.54540813e-01
2.47063994e-01 -1.30738810e-01 -1.37434661e-01 -1.85667396e-01
5.28326750e-01 2.53358483e-01 1.99184179e-01 -3.22185993e-01
9.87237394e-01 -8.73205960e-01 7.23088562e-01 -1.03382498e-01
-4.11073059e-01 5.03060818e-01 6.85365796e-01 -5.88230014e-01
-1.18165433e+00 -1.35095417e+00 -1.28007248e-01 1.01079094e+00
-4.37812023e-02 -4.92993668e-02 -1.27620816e+00 -8.40345562e-01
8.07011202e-02 -1.48969551e-03 -1.10968447e+00 -4.86758888e-01
-5.39833486e-01 -4.49709684e-01 2.56170899e-01 7.45368898e-01
1.05137360e+00 -1.24779010e+00 -2.27602348e-01 -3.56707454e-01
3.61138791e-01 -6.01893008e-01 -9.23096180e-01 -4.01996613e-01
-6.75226986e-01 -9.76338685e-01 -1.24096060e+00 -1.23537719e+00
1.02069151e+00 -3.42506886e-01 6.89897358e-01 1.14822634e-01
-2.57826656e-01 1.56007871e-01 -4.29683328e-02 -5.60389876e-01
-3.20638835e-01 -1.69690341e-01 3.53354216e-01 2.66199619e-01
6.52436435e-01 -9.41990733e-01 -1.05158842e+00 8.35229978e-02
-4.10097748e-01 -1.36576504e-01 6.09794497e-01 5.40745497e-01
1.82372496e-01 -1.43836156e-01 6.39028013e-01 -3.68659854e-01
5.88684201e-01 -1.56233802e-01 -5.63779593e-01 4.10273850e-01
-8.59369040e-01 1.98318452e-01 6.49943888e-01 -8.34272981e-01
-1.15107775e+00 -1.39396816e-01 1.43698499e-01 -3.63199085e-01
-2.74029821e-02 -7.68716112e-02 -3.83907020e-01 -1.65884227e-01
4.30159360e-01 2.48288572e-01 6.76611423e-01 -3.76613975e-01
1.36744261e-01 1.14741385e+00 8.57789218e-01 -7.42965341e-01
6.82385981e-01 2.29239315e-01 -1.09629259e-01 -4.71280903e-01
-6.05997980e-01 4.15268093e-01 -5.70028782e-01 -6.67659283e-01
9.81668591e-01 -8.71964812e-01 -9.76520479e-01 1.49259031e+00
-1.08658957e+00 -1.91592231e-01 9.07142907e-02 4.32015479e-01
-4.79433298e-01 4.67843682e-01 -7.89165497e-01 -7.60674000e-01
-6.85985208e-01 -8.34747374e-01 9.21067774e-01 9.22490239e-01
6.21343739e-02 -6.63997412e-01 -2.45817393e-01 -2.50015825e-01
2.35770419e-01 1.36966988e-01 1.06589222e+00 -7.78159453e-03
2.64258347e-02 3.78180146e-02 -4.09855098e-01 8.71113122e-01
3.54419678e-01 5.55240273e-01 -1.06888640e+00 -2.26699769e-01
-1.77697644e-01 -4.58315283e-01 7.44543016e-01 3.44433397e-01
1.69648254e+00 -2.61480778e-01 1.41896531e-01 6.86430573e-01
1.10854983e+00 5.21086454e-01 8.03533673e-01 5.44823147e-02
5.43303967e-01 6.59027219e-01 6.96443796e-01 6.03944182e-01
2.94572890e-01 1.92663655e-01 1.15732051e-01 7.99071118e-02
-7.39342421e-02 -2.62176692e-01 5.47948301e-01 8.46976876e-01
-3.37320477e-01 5.27763426e-01 -4.60436940e-01 2.64741540e-01
-1.35519361e+00 -8.36553276e-01 4.14411783e-01 2.33572268e+00
1.30265415e+00 1.01336747e-01 3.82081926e-01 5.03869951e-02
1.18078041e+00 4.98119704e-02 -6.96814358e-01 -6.59111798e-01
3.97816338e-02 3.97401124e-01 1.57305643e-01 9.90453959e-02
-9.23878372e-01 6.74772561e-01 6.34782028e+00 8.41258764e-01
-1.19296467e+00 -2.79340029e-01 8.67209792e-01 2.09757105e-01
-5.82657568e-02 -3.61975640e-01 -7.77002990e-01 8.68844330e-01
7.33877003e-01 -4.14561838e-01 8.03289652e-01 1.04559243e+00
-3.10091645e-01 -9.64120869e-03 -1.22595286e+00 1.32609391e+00
3.19121808e-01 -5.01657426e-01 1.07061818e-01 -8.73381719e-02
7.18088388e-01 -9.26862895e-01 6.15727961e-01 3.81190330e-01
1.32043567e-02 -1.00566435e+00 7.55582213e-01 8.70855391e-01
1.42578351e+00 -8.93468797e-01 6.03683233e-01 -1.66460827e-01
-1.30223501e+00 -2.89612383e-01 -5.66575587e-01 -2.33202294e-01
-4.02319849e-01 4.48432505e-01 -1.81485116e-01 6.11817762e-02
8.51135910e-01 8.22924495e-01 -8.15577507e-01 8.39289546e-01
-2.27213591e-01 7.10916594e-02 2.64029689e-02 -7.03284889e-02
-2.38357857e-01 -5.76007485e-01 -3.79532576e-01 6.72829092e-01
9.52718198e-01 -1.37636110e-01 -3.29298913e-01 5.00195503e-01
-3.34166825e-01 7.97531009e-02 -2.58887202e-01 1.06224239e-01
8.72243702e-01 1.32319582e+00 -2.92595625e-01 -4.41479981e-01
-4.78245944e-01 1.07044137e+00 3.46774548e-01 5.38809970e-02
-8.84045005e-01 -6.93792105e-01 8.18086445e-01 -1.11460984e-01
1.62784427e-01 1.97925344e-02 -1.40864670e-01 -7.47884512e-01
-2.50293538e-02 -8.58897805e-01 5.10763645e-01 -1.04746377e+00
-1.50906491e+00 5.40828645e-01 9.67488531e-03 -1.24102581e+00
-8.16956088e-02 -9.13127482e-01 -9.55115974e-01 1.02993190e+00
-1.06417775e+00 -1.06091738e+00 -8.68854284e-01 5.67797959e-01
1.99671492e-01 -4.87338334e-01 7.55595744e-01 3.14402670e-01
-4.34475839e-01 1.00052345e+00 -5.98030500e-02 3.60114068e-01
1.28300965e+00 -1.31706572e+00 1.64862648e-01 3.80455732e-01
-6.85350955e-01 4.27000016e-01 5.12130320e-01 -4.97204155e-01
-6.87163293e-01 -8.15479636e-01 7.25373328e-01 -6.40023872e-02
5.97151279e-01 -1.96246400e-01 -9.45671558e-01 4.46325928e-01
2.37875238e-01 1.52701691e-01 2.96102405e-01 -1.90838754e-01
-6.23268127e-01 -7.01817214e-01 -1.33322513e+00 6.05510235e-01
1.05538416e+00 -4.57521260e-01 -5.65698802e-01 -2.89046556e-01
6.09916806e-01 -1.23029061e-01 -1.14575982e+00 3.70759577e-01
1.03145885e+00 -9.43084836e-01 6.63769007e-01 -1.48637146e-01
7.69192815e-01 -3.05109024e-01 2.94013053e-01 -1.25392401e+00
-2.01040938e-01 -1.26765534e-01 -1.48118973e-01 1.64898932e+00
5.96134253e-02 -7.26045549e-01 8.73980761e-01 4.53074366e-01
2.06885576e-01 -9.70753074e-01 -7.48316228e-01 -7.11020410e-01
3.63651395e-01 3.62924457e-01 1.19748199e+00 7.21726120e-01
-3.40285391e-01 -3.10643706e-02 -3.53885025e-01 -1.07500613e-01
7.39995718e-01 -1.26477614e-01 4.66710657e-01 -1.49220908e+00
2.47336868e-02 -4.25605267e-01 -6.61579370e-01 -4.71324176e-01
3.76016855e-01 -4.51121122e-01 -1.29366592e-01 -1.07606387e+00
3.86074930e-01 -2.92746872e-01 -4.85477567e-01 3.50811690e-01
-3.25028688e-01 2.05910310e-01 -1.42190620e-01 -2.26327449e-01
-1.32459402e-01 7.26388991e-01 1.48003602e+00 -2.63774246e-01
3.33735868e-02 -1.65894687e-01 -5.80402672e-01 7.95125663e-01
8.53188992e-01 -1.03563510e-01 -3.51382703e-01 -2.60537058e-01
7.64597952e-02 -2.45037228e-01 -5.01957648e-02 -1.30837238e+00
1.13172933e-01 -2.33526200e-01 1.02560329e+00 -4.21434581e-01
1.53292418e-01 -6.44025028e-01 -4.89798300e-02 5.71928501e-01
-6.32656738e-02 1.70954362e-01 -1.75559282e-01 1.02529958e-01
-3.80187556e-02 -1.57087356e-01 1.03438401e+00 4.22189869e-02
-8.85188639e-01 7.15030193e-01 5.83878765e-03 -2.29458898e-01
1.11711061e+00 -4.04879451e-01 -5.18463731e-01 -4.04845089e-01
-5.10882676e-01 2.17470258e-01 6.15732133e-01 4.17609513e-01
7.69209683e-01 -1.76105976e+00 -6.93628907e-01 5.70711792e-01
1.52020007e-01 -1.84382528e-01 3.78551632e-01 4.94061381e-01
-5.63152730e-01 -5.18012881e-01 -1.03480744e+00 -3.02810520e-01
-1.55539727e+00 7.34052539e-01 5.41507423e-01 3.71247560e-01
-3.91531028e-02 9.12445426e-01 3.10459882e-01 -2.83221245e-01
2.73693442e-01 7.03563681e-04 -4.83082920e-01 2.93406099e-01
1.02363682e+00 3.92068624e-01 -3.30423474e-01 -3.76729757e-01
-2.46431544e-01 1.20125031e+00 -2.68777132e-01 -2.63138711e-02
1.13203466e+00 1.50432870e-01 -5.65042078e-01 3.39962006e-01
1.41434491e+00 -2.59737730e-01 -1.66729307e+00 4.62565795e-02
-4.61748004e-01 -4.65393990e-01 -4.18278128e-01 -6.66019142e-01
-1.31217408e+00 9.21951354e-01 1.10868573e+00 -2.98096269e-01
1.64601755e+00 -1.07993975e-01 8.09444070e-01 -1.76845174e-02
3.73938113e-01 -1.25872660e+00 2.63180852e-01 2.40991190e-01
8.03855479e-01 -1.04068005e+00 3.00174616e-02 -1.88781545e-01
-4.37351346e-01 1.29768717e+00 1.15680826e+00 -1.44539759e-01
8.15529525e-01 2.17675120e-01 3.03212702e-01 3.07286233e-01
-4.00258899e-01 1.37982890e-01 1.55182123e-01 8.88552904e-01
4.32832360e-01 -3.88808772e-02 -7.05684721e-01 6.12133086e-01
-5.10447145e-01 1.60671353e-01 1.49114370e-01 6.26280248e-01
-6.63534045e-01 -1.23481894e+00 -4.43934530e-01 6.58736587e-01
-3.34349006e-01 9.81450379e-02 -1.60796270e-01 4.37232554e-01
5.64339459e-01 4.79005069e-01 6.26972139e-01 -7.69728899e-01
-9.73244756e-02 2.03798741e-01 9.68770742e-01 -2.36633062e-01
-2.38364667e-01 -4.77329552e-01 -4.50004607e-01 -2.78582722e-01
-1.92443565e-01 -6.54680729e-01 -1.10967457e+00 -6.89335644e-01
4.20155935e-02 3.94260213e-02 5.26533663e-01 6.32211208e-01
-7.93541819e-02 1.58517092e-01 1.00039339e+00 -5.16740918e-01
-4.96789575e-01 -1.16827559e+00 -8.95166218e-01 5.95147073e-01
6.67308792e-02 -6.77282512e-01 -4.86051679e-01 2.47584686e-01]
|
[13.50946044921875, 0.7438858151435852]
|
ef37a7bd-d520-4a0e-b167-5c6ebef10fa2
|
3d-aware-adversarial-makeup-generation-for
|
2306.14640
| null |
https://arxiv.org/abs/2306.14640v1
|
https://arxiv.org/pdf/2306.14640v1.pdf
|
3D-Aware Adversarial Makeup Generation for Facial Privacy Protection
|
The privacy and security of face data on social media are facing unprecedented challenges as it is vulnerable to unauthorized access and identification. A common practice for solving this problem is to modify the original data so that it could be protected from being recognized by malicious face recognition (FR) systems. However, such ``adversarial examples'' obtained by existing methods usually suffer from low transferability and poor image quality, which severely limits the application of these methods in real-world scenarios. In this paper, we propose a 3D-Aware Adversarial Makeup Generation GAN (3DAM-GAN). which aims to improve the quality and transferability of synthetic makeup for identity information concealing. Specifically, a UV-based generator consisting of a novel Makeup Adjustment Module (MAM) and Makeup Transfer Module (MTM) is designed to render realistic and robust makeup with the aid of symmetric characteristics of human faces. Moreover, a makeup attack mechanism with an ensemble training strategy is proposed to boost the transferability of black-box models. Extensive experiment results on several benchmark datasets demonstrate that 3DAM-GAN could effectively protect faces against various FR models, including both publicly available state-of-the-art models and commercial face verification APIs, such as Face++, Baidu and Aliyun.
|
['Jing Dong', 'Yunfan Liu', 'Bo Peng', 'Ziwen He', 'Yue Jiang', 'Yueming Lyu']
|
2023-06-26
| null | null | null | null |
['face-recognition', 'face-verification']
|
['computer-vision', 'computer-vision']
|
[ 3.15935791e-01 8.49135071e-02 2.36325219e-01 -3.55887830e-01
-5.45421302e-01 -8.08507085e-01 5.85927308e-01 -6.99898124e-01
1.72185779e-01 6.22538388e-01 -2.48250321e-01 -2.70187557e-01
3.72671604e-01 -1.01066375e+00 -8.70621383e-01 -9.35189843e-01
2.08381400e-01 -1.90105274e-01 -3.24950069e-01 -2.65712261e-01
-1.18875496e-01 6.63987398e-01 -1.48234546e+00 2.94345111e-01
1.01270974e+00 1.16564524e+00 -3.03163469e-01 2.15838209e-01
-1.62671995e-03 4.08826947e-01 -8.75130892e-01 -1.07579672e+00
7.25200295e-01 -6.16166711e-01 -1.94504470e-01 -1.53200269e-01
4.76669490e-01 -4.97434855e-01 -4.80298162e-01 1.18331134e+00
7.38399982e-01 -3.96669745e-01 5.19015372e-01 -1.75735605e+00
-1.13486266e+00 2.12105468e-01 -6.64021671e-01 -3.04481715e-01
3.57309252e-01 5.17892122e-01 1.70495555e-01 -7.44126260e-01
2.56883860e-01 1.40638018e+00 7.05734432e-01 1.16813421e+00
-1.01929784e+00 -1.36285925e+00 -2.14473620e-01 -1.32865943e-02
-1.50747132e+00 -8.12399268e-01 7.66432643e-01 -3.59286994e-01
1.74449056e-01 4.43159044e-01 2.67419338e-01 1.32158470e+00
7.48828426e-02 3.91931236e-01 1.29831457e+00 -1.15903340e-01
-4.51584794e-02 3.53994161e-01 -5.33496797e-01 6.49693012e-01
2.90251166e-01 4.08260673e-01 -3.54988962e-01 -3.49742413e-01
8.57168853e-01 -5.26790284e-02 -4.87748235e-01 -9.03304070e-02
-6.26453340e-01 5.73607922e-01 5.00088394e-01 -1.07038140e-01
-1.21340752e-01 -1.37265414e-01 1.80297017e-01 2.24731684e-01
3.59018177e-01 1.70613468e-01 -1.31567925e-01 4.89954948e-01
-5.40797830e-01 2.24975213e-01 3.73971075e-01 9.75298882e-01
6.31371796e-01 3.36152703e-01 -2.68091768e-01 8.25315833e-01
4.89053130e-01 9.25621212e-01 3.75984609e-01 -5.27424276e-01
4.84582305e-01 5.68307579e-01 -1.00676648e-01 -1.22047389e+00
3.25733602e-01 -5.71401455e-02 -1.10323012e+00 4.46334153e-01
1.87586501e-01 -3.48521948e-01 -9.32310402e-01 1.88279700e+00
6.50511563e-01 3.75921935e-01 2.41713747e-01 5.87320089e-01
9.65240240e-01 7.28996336e-01 1.48484558e-01 -8.66513997e-02
1.28432560e+00 -6.00918412e-01 -6.94437265e-01 -4.63733822e-02
1.02987371e-01 -8.40292215e-01 8.84956717e-01 8.30679461e-02
-9.35700595e-01 -8.51057887e-01 -1.21041739e+00 1.59939840e-01
-2.51036525e-01 3.14572342e-02 4.19657946e-01 1.45288777e+00
-9.18522120e-01 2.56069124e-01 -3.12221140e-01 8.17627460e-02
1.03782928e+00 5.63199341e-01 -8.47666800e-01 -9.82551426e-02
-1.19311059e+00 3.99292350e-01 6.95871934e-02 3.48627925e-01
-1.11535251e+00 -8.85712922e-01 -9.45082545e-01 -1.58944517e-01
-4.58397865e-02 -4.97486115e-01 8.38324487e-01 -1.29400706e+00
-1.72115576e+00 9.16154683e-01 9.29218903e-02 -2.00696275e-01
6.53556585e-01 6.18236400e-02 -8.29087317e-01 4.33092378e-03
-2.47212172e-01 4.62681830e-01 1.46888983e+00 -1.45206809e+00
-1.13838807e-01 -5.32697678e-01 -3.36413048e-02 -2.19111070e-01
-5.16446829e-01 3.99980426e-01 -5.22538908e-02 -1.01255345e+00
-4.67508882e-01 -8.77729118e-01 1.61881655e-01 3.50647211e-01
-5.49586833e-01 3.29476207e-01 1.24188328e+00 -1.06600273e+00
8.99600089e-01 -2.28703737e+00 -5.77838160e-02 3.44217241e-01
-8.04449096e-02 9.61944401e-01 -3.06028634e-01 2.26998881e-01
-2.72958964e-01 4.71539974e-01 -4.43891972e-01 -2.63240665e-01
-8.26317072e-02 -5.94623294e-03 -5.06625473e-01 5.52661300e-01
4.82409060e-01 8.71428072e-01 -4.57938433e-01 -2.55818069e-01
1.07963286e-01 9.72289085e-01 -5.63435614e-01 5.87420642e-01
-1.09336406e-01 7.71102130e-01 -3.80136728e-01 8.38512957e-01
1.46714067e+00 2.59006262e-01 7.03740269e-02 1.64165851e-02
4.69568521e-01 -3.54393452e-01 -1.06157100e+00 9.97523308e-01
-3.75097752e-01 1.73463136e-01 2.93770134e-01 -3.44738454e-01
1.07397234e+00 4.01440173e-01 -1.50525467e-02 -5.39384604e-01
3.62716079e-01 6.36711121e-02 -4.09015194e-02 -3.04536313e-01
-3.71484719e-02 -6.96275532e-02 1.60167828e-01 2.44128823e-01
-1.21478245e-01 6.06043637e-02 -4.66218382e-01 -1.13892876e-01
7.62259305e-01 1.13398265e-02 -1.24621522e-02 -1.12439595e-01
1.06178105e+00 -8.31048191e-01 7.87343442e-01 3.19413580e-02
-9.94578749e-02 7.98297107e-01 3.97854596e-01 -1.90552369e-01
-8.51450384e-01 -9.58403707e-01 -1.70048654e-01 4.88047093e-01
1.02395684e-01 -2.54670829e-01 -1.26000834e+00 -1.12030852e+00
2.38116592e-01 3.81606996e-01 -7.07434714e-01 -4.51850444e-01
-5.90841234e-01 -6.89634562e-01 1.04288208e+00 3.75168085e-01
1.02837420e+00 -1.03427494e+00 1.41625479e-01 -2.10060090e-01
3.47448997e-02 -1.02911329e+00 -6.83009148e-01 -8.56299222e-01
-2.85172790e-01 -1.32331133e+00 -7.04440534e-01 -7.09293962e-01
1.06039786e+00 1.80471420e-01 4.84282881e-01 5.13658762e-01
-1.91012561e-01 2.35540401e-02 -2.38169417e-01 -6.22034907e-01
-9.05527413e-01 -4.19160277e-01 2.35505298e-01 8.02944064e-01
1.49895651e-02 -4.90995914e-01 -6.75934792e-01 6.02878273e-01
-1.11759555e+00 -8.51784274e-02 3.82702231e-01 7.48350084e-01
2.78241903e-01 1.20585039e-01 8.35795283e-01 -9.40249860e-01
3.92894924e-01 -4.06935245e-01 -7.16313839e-01 4.79514986e-01
-4.76737440e-01 -3.10331523e-01 9.02606785e-01 -5.25445998e-01
-1.33870101e+00 -1.37364626e-01 -5.42578101e-01 -4.37522352e-01
-6.27633333e-02 -2.42728487e-01 -1.11323667e+00 -7.15073705e-01
4.51748043e-01 4.23559219e-01 3.89268488e-01 -2.97792763e-01
2.21088767e-01 8.94162536e-01 4.77967650e-01 -5.58421910e-01
1.49712908e+00 3.91186565e-01 6.55755354e-03 -5.91062844e-01
-2.13148788e-01 3.70513171e-01 -4.13280912e-02 -2.84924120e-01
6.56145573e-01 -1.04121864e+00 -8.86610389e-01 1.38134050e+00
-1.03127778e+00 -3.97982076e-02 6.88945353e-02 -1.17481075e-01
-2.15510324e-01 4.17791963e-01 -5.68742275e-01 -7.38976598e-01
-6.48067236e-01 -1.25940096e+00 1.04576969e+00 5.33568382e-01
3.35202515e-01 -5.61338127e-01 -2.35031724e-01 6.81607008e-01
6.15751863e-01 6.59880757e-01 6.77140415e-01 -3.23059261e-01
-5.17756343e-01 -3.30358654e-01 -1.39917254e-01 1.00072634e+00
4.44368899e-01 1.10790558e-01 -1.36681271e+00 -5.46336591e-01
2.34213814e-01 -2.00980932e-01 4.97160584e-01 -2.03124896e-01
1.38387549e+00 -8.32528710e-01 -2.17140287e-01 1.06199610e+00
1.25478816e+00 2.46183798e-01 1.16192341e+00 -1.53230548e-01
8.57622802e-01 6.41382515e-01 3.31005305e-01 2.96531349e-01
8.29143971e-02 7.07131028e-01 6.69622302e-01 4.18102741e-02
-1.44718707e-01 -5.68344235e-01 6.01917624e-01 3.50434780e-01
4.87607867e-02 -3.41612667e-01 -5.16701758e-01 4.58195619e-02
-1.30955517e+00 -9.71843183e-01 1.45072967e-01 2.25102520e+00
8.90147686e-01 -2.88878411e-01 -9.55437496e-02 7.45983943e-02
1.04433751e+00 1.24792628e-01 -5.33002377e-01 -1.75537840e-01
-3.00047100e-01 3.45212132e-01 3.14275056e-01 1.26388118e-01
-1.05466723e+00 7.91762769e-01 5.17237329e+00 9.40293431e-01
-1.19529927e+00 2.25817174e-01 9.99088883e-01 1.84709609e-01
-3.51383954e-01 -3.14728975e-01 -8.95324647e-01 9.79403079e-01
6.44086659e-01 -5.63645326e-02 5.48437178e-01 8.35710406e-01
-1.71494097e-01 5.65270185e-01 -8.02756429e-01 1.00044143e+00
3.44494820e-01 -1.22440624e+00 2.43357733e-01 2.17273995e-01
8.03369462e-01 -7.52254725e-01 4.88493353e-01 1.71005026e-01
4.21689600e-02 -1.24750102e+00 6.35037482e-01 2.96786159e-01
1.32981753e+00 -9.62794423e-01 6.09463513e-01 7.06126168e-02
-1.07852066e+00 -1.20211713e-01 -2.70186961e-01 3.96064520e-01
-2.65437718e-02 3.76179457e-01 -5.84057271e-01 7.23170817e-01
5.48038721e-01 3.29594582e-01 -6.77493870e-01 5.92722356e-01
-5.73370039e-01 4.39889759e-01 -2.43410811e-01 3.48719954e-01
-3.09388489e-01 -7.18708560e-02 4.27656054e-01 6.57462239e-01
5.37482083e-01 1.27211630e-01 -4.02918488e-01 9.19905543e-01
-6.99662089e-01 7.65584782e-02 -8.12013209e-01 1.14662840e-03
6.35335505e-01 1.27237904e+00 -5.28701879e-02 1.88100412e-01
-1.22234128e-01 9.62978005e-01 1.15111984e-01 1.58425480e-01
-1.07883191e+00 -1.72855422e-01 1.16958523e+00 2.08691761e-01
1.69562802e-01 1.13413230e-01 3.91135067e-02 -1.14114857e+00
3.24440777e-01 -1.37080896e+00 2.25602120e-01 -5.00307620e-01
-1.50240719e+00 8.32376778e-01 -4.43177909e-01 -1.25757062e+00
-7.01579824e-03 -5.48204064e-01 -7.73869216e-01 1.10189569e+00
-1.53169692e+00 -1.70024562e+00 -4.15502608e-01 8.57759058e-01
9.58137214e-02 -6.69815838e-01 8.37763369e-01 4.25378174e-01
-7.92481780e-01 1.30992162e+00 -1.10948958e-01 4.62919503e-01
6.42284334e-01 -5.54416001e-01 6.91834211e-01 9.78708267e-01
-1.64342880e-01 6.31940961e-01 2.23415717e-01 -6.58276916e-01
-1.69798231e+00 -1.43844116e+00 2.59416878e-01 -3.79665196e-01
-7.74232531e-03 -6.34583175e-01 -9.70572650e-01 5.96721351e-01
1.72249824e-01 2.97310263e-01 8.25148463e-01 -6.96987569e-01
-7.24128723e-01 -4.33998406e-01 -1.99712324e+00 4.66152161e-01
9.77460980e-01 -6.72418416e-01 -9.17234421e-02 2.70002306e-01
6.24894381e-01 -2.96432316e-01 -7.85849631e-01 6.24153137e-01
5.96337020e-01 -9.56667125e-01 1.11555457e+00 -5.57434916e-01
3.26083928e-01 -4.53984141e-01 -1.17179155e-01 -1.14011300e+00
1.31462276e-01 -1.02810979e+00 -2.25213066e-01 2.09362197e+00
-3.94516885e-02 -1.00334704e+00 7.22100735e-01 5.82936704e-01
2.61897981e-01 -3.85229886e-01 -9.52121377e-01 -8.29231083e-01
8.99612680e-02 -1.90191772e-02 1.38314199e+00 1.03786409e+00
-6.01512909e-01 -2.69845665e-01 -6.36847079e-01 5.12374222e-01
7.91054845e-01 -4.22636479e-01 1.07806909e+00 -7.96123803e-01
-7.82425031e-02 -1.55690014e-01 -5.54394484e-01 -5.00219285e-01
5.57981074e-01 -7.82682002e-01 -2.44306415e-01 -6.80258393e-01
-5.19474521e-02 -7.07312524e-01 -6.11340329e-02 5.05001605e-01
-3.03944349e-01 6.06157541e-01 2.95334607e-01 -1.82227287e-02
4.12949651e-01 7.94827878e-01 1.30864108e+00 -2.23612309e-01
2.42034525e-01 8.37377980e-02 -8.88825178e-01 5.00819027e-01
8.82466972e-01 -4.45375115e-01 -4.69741136e-01 -3.51417392e-01
-1.66841656e-01 -1.13557197e-01 6.14497483e-01 -1.11777306e+00
-6.16273321e-02 -9.15604830e-02 4.81857479e-01 1.16433464e-02
3.05002153e-01 -9.72903728e-01 5.79165041e-01 2.87534654e-01
3.00467163e-02 1.88400084e-03 3.94194037e-01 3.42922956e-01
-1.68750435e-01 8.59404430e-02 1.14367294e+00 1.13916650e-01
-4.07909542e-01 7.73246765e-01 3.35157037e-01 -1.15695067e-01
1.44169247e+00 -1.28195614e-01 -6.08280659e-01 -2.04791337e-01
-6.85681701e-02 8.02497473e-03 7.59627700e-01 5.91707766e-01
7.79503405e-01 -1.65476990e+00 -1.06817079e+00 9.74681079e-01
1.27802104e-01 -1.87687188e-01 6.62104845e-01 9.07611474e-02
-5.89745104e-01 -1.19423620e-01 -4.99589592e-01 -1.50425673e-01
-1.42970479e+00 7.88880050e-01 5.25509357e-01 2.38293871e-01
-2.98170686e-01 9.59565580e-01 4.51956779e-01 -6.17686450e-01
3.05348709e-02 4.33516681e-01 -2.45484477e-03 -1.82622090e-01
9.63063419e-01 1.81276485e-01 2.65745446e-02 -1.04877007e+00
-3.38077098e-01 5.99081755e-01 8.44128430e-02 3.26429248e-01
1.06266356e+00 5.62480539e-02 -3.66543263e-01 -5.65807164e-01
1.25127399e+00 2.71979362e-01 -1.38562727e+00 1.15315020e-01
-6.86354160e-01 -9.42225933e-01 -3.32151145e-01 -6.20254338e-01
-1.60662246e+00 7.36806750e-01 7.19263136e-01 1.31959334e-01
1.25067878e+00 -3.84239048e-01 1.10156643e+00 -1.72976047e-01
4.86101538e-01 -4.31034505e-01 -1.60373628e-01 -8.15594420e-02
1.10156739e+00 -1.18467224e+00 -2.96941757e-01 -7.76623309e-01
-4.92431760e-01 7.91711271e-01 8.69732559e-01 5.86085245e-02
7.52506614e-01 2.99660474e-01 1.91279560e-01 3.27933609e-01
-2.36276969e-01 5.57814479e-01 2.91421145e-01 9.09593761e-01
-1.44622490e-01 -6.50082901e-02 9.03925151e-02 8.14012647e-01
-2.50579506e-01 -9.56493244e-02 2.55020082e-01 8.51515830e-01
2.99459696e-01 -1.44950128e+00 -6.50615871e-01 1.06410779e-01
-7.07707405e-01 6.43849522e-02 -3.41861129e-01 5.49354374e-01
5.10883033e-01 9.40643489e-01 -2.09140211e-01 -7.23596215e-01
1.51738793e-01 -1.01073422e-02 4.99965549e-01 -3.40008110e-01
-9.12378907e-01 -5.58109581e-01 -1.82784840e-01 -3.53661060e-01
-2.89809823e-01 -4.98829067e-01 -8.00906599e-01 -7.86911666e-01
-2.55008191e-01 -6.29866496e-02 6.00482464e-01 6.20935738e-01
6.09940648e-01 2.44179323e-01 1.21492362e+00 -6.08890593e-01
-5.79717040e-01 -7.05597401e-01 -5.03056169e-01 5.63506603e-01
3.12256277e-01 -5.27844787e-01 -2.69369334e-01 2.00383708e-01]
|
[12.783671379089355, 0.8111989498138428]
|
f087f966-5ac5-4588-a2aa-472369c80c5d
|
mx2m-masked-cross-modality-modeling-in-domain
|
2307.04231
| null |
https://arxiv.org/abs/2307.04231v1
|
https://arxiv.org/pdf/2307.04231v1.pdf
|
Mx2M: Masked Cross-Modality Modeling in Domain Adaptation for 3D Semantic Segmentation
|
Existing methods of cross-modal domain adaptation for 3D semantic segmentation predict results only via 2D-3D complementarity that is obtained by cross-modal feature matching. However, as lacking supervision in the target domain, the complementarity is not always reliable. The results are not ideal when the domain gap is large. To solve the problem of lacking supervision, we introduce masked modeling into this task and propose a method Mx2M, which utilizes masked cross-modality modeling to reduce the large domain gap. Our Mx2M contains two components. One is the core solution, cross-modal removal and prediction (xMRP), which makes the Mx2M adapt to various scenarios and provides cross-modal self-supervision. The other is a new way of cross-modal feature matching, the dynamic cross-modal filter (DxMF) that ensures the whole method dynamically uses more suitable 2D-3D complementarity. Evaluation of the Mx2M on three DA scenarios, including Day/Night, USA/Singapore, and A2D2/SemanticKITTI, brings large improvements over previous methods on many metrics.
|
['Wenhui Li', 'Shenghao Zhang', 'Yuanyuan Guan', 'Yonggen Ling', 'Zunran Wang', 'Boxiang Zhang']
|
2023-07-09
| null | null | null | null |
['semantic-segmentation', '3d-semantic-segmentation', 'domain-adaptation']
|
['computer-vision', 'computer-vision', 'methodology']
|
[ 1.70970466e-02 -4.59370902e-03 -3.96513999e-01 -4.39845979e-01
-9.74690318e-01 -5.38879931e-01 6.12748742e-01 -3.98753852e-01
-3.79692502e-02 3.92315179e-01 1.08776733e-01 6.82875961e-02
-2.93598115e-01 -8.15458715e-01 -4.67350036e-01 -7.66568661e-01
2.96823800e-01 7.91812837e-01 7.62969971e-01 -4.95236218e-01
1.12332188e-01 4.01239663e-01 -1.86772525e+00 5.63708842e-01
1.04819667e+00 1.11450839e+00 3.63760293e-01 1.71997651e-01
-6.19418204e-01 1.33721605e-01 -3.70504230e-01 -1.09871149e-01
6.21209681e-01 -1.87052757e-01 -1.06235850e+00 3.58541697e-01
3.66561890e-01 3.45042422e-02 1.08805887e-01 9.72113311e-01
7.34224975e-01 1.14912562e-01 6.90042675e-01 -1.33364844e+00
-2.55275995e-01 2.32031584e-01 -8.88127983e-01 -1.23606309e-01
3.88188779e-01 3.55618284e-03 5.48529863e-01 -8.13682139e-01
9.04180110e-01 1.46354806e+00 7.80659139e-01 7.68100739e-01
-1.15141368e+00 -6.36264801e-01 1.25074521e-01 1.57271817e-01
-1.28953135e+00 -1.14404060e-01 1.12775993e+00 -6.01693690e-01
7.35985756e-01 2.81825542e-01 4.13971841e-01 8.91935766e-01
-1.57476053e-01 8.95721138e-01 1.47510850e+00 -6.13060296e-01
-6.77890256e-02 4.15215552e-01 9.44970027e-02 3.16134721e-01
-5.44805646e-01 3.12493533e-01 -7.56683171e-01 -8.24790373e-02
5.16208887e-01 -4.09717768e-01 -1.65260077e-01 -6.59916520e-01
-1.08336604e+00 7.05482483e-01 1.41067579e-01 3.08119148e-01
9.51727014e-03 -8.02607536e-01 4.53096122e-01 4.17012155e-01
6.50740445e-01 7.59303123e-02 -7.08349347e-01 -1.06754899e-02
-9.81027961e-01 2.70730644e-01 3.13528329e-01 9.52240825e-01
8.88329923e-01 -2.71142066e-01 1.16919279e-01 1.33203232e+00
4.20863241e-01 7.29035676e-01 6.48249328e-01 -7.84848511e-01
5.85454524e-01 9.13213313e-01 -1.98966488e-01 -7.31677532e-01
-6.21331871e-01 -2.93821663e-01 -5.66719711e-01 3.76152456e-01
3.71134967e-01 7.34386072e-02 -1.26671875e+00 1.74130797e+00
9.19370592e-01 2.81407796e-02 2.43009403e-01 1.16191566e+00
1.20692599e+00 3.16267282e-01 3.91866043e-02 1.88390091e-01
1.20751417e+00 -9.81290400e-01 -5.73979080e-01 -3.21233690e-01
7.34322071e-01 -1.11668909e+00 1.01512241e+00 1.56751156e-01
-7.38981187e-01 -8.54293346e-01 -9.95462894e-01 9.26080793e-02
-5.92577398e-01 -9.88696050e-03 4.74277824e-01 5.76216161e-01
-7.13677526e-01 4.56555367e-01 -4.73079294e-01 -6.01772606e-01
1.25852570e-01 3.70078385e-01 -6.41810715e-01 -1.35209393e-02
-1.51794398e+00 1.00086141e+00 6.48036063e-01 -2.00802416e-01
-4.11519051e-01 -7.69383788e-01 -8.23298991e-01 -5.71505725e-01
1.89178512e-01 -5.73037446e-01 8.67636979e-01 -1.16994846e+00
-1.48813629e+00 1.41151464e+00 1.12759909e-02 -1.11591421e-01
6.79546475e-01 3.18663893e-03 -8.12646747e-01 1.37652412e-01
4.60791737e-01 9.51057613e-01 8.40022743e-01 -1.62203872e+00
-7.85355151e-01 -6.38005614e-01 -1.44216329e-01 4.67901081e-01
-1.18355528e-01 -2.58991867e-01 -7.40227342e-01 -6.93956733e-01
5.62588632e-01 -1.00324094e+00 -1.42185330e-01 -3.49092811e-01
-3.10440630e-01 -1.11181177e-01 1.13256729e+00 -7.38658249e-01
9.67473030e-01 -2.29287839e+00 1.46655798e-01 3.88919592e-01
-1.44125819e-01 2.06010684e-01 -6.83520436e-02 2.45130405e-01
-3.51409942e-01 -3.24393332e-01 -4.61867183e-01 -3.78469944e-01
-1.05301328e-01 3.03416729e-01 -1.22818500e-02 3.46646130e-01
1.67263493e-01 6.11075699e-01 -4.83449250e-01 -8.79371464e-01
5.51596344e-01 1.93188801e-01 -2.91802645e-01 4.49030884e-02
-2.78517365e-01 7.75653720e-01 -5.08004248e-01 8.58561516e-01
1.35948217e+00 -1.37320191e-01 -5.57145476e-03 -4.37502384e-01
-1.36422127e-01 1.05902813e-01 -1.56740975e+00 2.13853383e+00
-2.97774702e-01 2.38930836e-01 6.61856607e-02 -1.15905058e+00
1.13954413e+00 2.42929369e-01 9.15368855e-01 -1.00866580e+00
1.60008296e-01 5.06015480e-01 -4.07299459e-01 -6.01223528e-01
4.00328547e-01 -3.81145269e-01 -4.12946083e-02 7.93165490e-02
1.40948117e-01 -3.14947307e-01 4.77270130e-03 1.98549572e-02
4.71710742e-01 4.88544196e-01 -1.39441550e-01 -4.26680893e-01
7.01852381e-01 4.99701917e-01 8.28947067e-01 3.65783483e-01
-3.41470778e-01 1.06061947e+00 1.32946640e-01 -1.36637613e-01
-7.25381374e-01 -1.08229327e+00 -3.64954859e-01 7.63768077e-01
7.21376002e-01 -3.93317640e-02 -6.89923823e-01 -1.02953029e+00
1.20355442e-01 4.24648613e-01 -6.09201014e-01 -2.72864521e-01
-4.00260866e-01 -8.50727737e-01 4.14448798e-01 4.63173300e-01
8.57477725e-01 -7.24977076e-01 -1.59263447e-01 -3.32297496e-02
-4.78832960e-01 -1.16977596e+00 -3.40115964e-01 2.38565445e-01
-1.01555920e+00 -1.03404617e+00 -8.36014450e-01 -9.10278618e-01
3.13974798e-01 2.09372923e-01 1.21326387e+00 -3.15875798e-01
6.64273128e-02 2.15511322e-01 -5.30969739e-01 -1.25867233e-01
-3.94040108e-01 1.38939008e-01 -6.17398880e-02 5.58884405e-02
6.48578823e-01 -5.55838048e-01 -3.72050285e-01 7.28744209e-01
-7.66904414e-01 2.32780471e-01 3.95187169e-01 8.18069577e-01
8.58769596e-01 5.61025254e-02 5.60983300e-01 -1.08464980e+00
-1.07173629e-01 -4.01560754e-01 -4.49362069e-01 2.76440799e-01
-6.96843982e-01 -1.39606982e-01 2.10654929e-01 -4.97467488e-01
-1.41755676e+00 2.97012150e-01 -4.31461006e-01 -5.34884572e-01
-4.32302505e-01 3.04323137e-01 -6.13813281e-01 -1.01983309e-01
7.26265430e-01 1.33534193e-01 1.83921844e-01 -9.77002680e-01
2.27416590e-01 8.52281153e-01 5.34836471e-01 -5.42228818e-01
8.66187453e-01 7.18415678e-01 -2.05118656e-02 -6.24008179e-01
-8.49170387e-01 -8.08892667e-01 -8.92481804e-01 -1.87372640e-01
9.30481315e-01 -1.07933080e+00 -2.17475265e-01 7.25155413e-01
-8.39788854e-01 -1.35204643e-01 -3.14130723e-01 5.17948329e-01
-5.53840101e-01 4.62302685e-01 -1.50481224e-01 -4.14922327e-01
-9.81781036e-02 -1.25037169e+00 1.40521514e+00 1.80991605e-01
-8.80819280e-03 -1.00763202e+00 1.27239287e-01 8.97335649e-01
1.28680602e-01 3.77394348e-01 9.63674068e-01 -5.73872566e-01
-2.16917440e-01 9.21819359e-02 -2.58579671e-01 4.62057054e-01
2.27327734e-01 -2.75990456e-01 -1.09639776e+00 -1.68154519e-02
-1.33502021e-01 -6.37572706e-02 6.65425420e-01 3.70155126e-01
7.23875940e-01 5.50674379e-01 -4.65576380e-01 6.38414919e-01
1.29205513e+00 1.07034311e-01 6.14665627e-01 7.55770564e-01
7.47144580e-01 9.05922890e-01 1.35682356e+00 1.54974669e-01
6.39709651e-01 1.05813372e+00 3.55036587e-01 -4.68145788e-01
-5.61343849e-01 -1.03882670e-01 9.20037478e-02 7.03862309e-01
1.86110467e-01 1.77129418e-01 -1.00781238e+00 7.47557759e-01
-1.96851671e+00 -6.35981441e-01 -3.55033487e-01 2.05971861e+00
7.44674861e-01 1.24627955e-01 3.85708481e-01 1.93417504e-01
8.07219863e-01 -6.31438717e-02 -5.42549670e-01 -1.01693034e-01
-5.92904508e-01 3.87754850e-02 5.63106000e-01 4.34368938e-01
-1.24724114e+00 9.62951958e-01 6.01410341e+00 1.39575315e+00
-1.22871244e+00 3.21824789e-01 3.74972343e-01 1.85356930e-01
-4.24792856e-01 3.57613489e-02 -8.28833699e-01 4.86779988e-01
3.06308895e-01 4.91814464e-01 -4.14938964e-02 8.85539353e-01
1.09036844e-02 -2.83253819e-01 -8.06944251e-01 1.18270493e+00
5.88644808e-03 -1.17027688e+00 -7.89595991e-02 -1.62566975e-01
8.79037499e-01 1.60185434e-02 -3.98927405e-02 1.73901394e-01
1.37134437e-02 -6.83913112e-01 7.81623721e-01 4.24077243e-01
8.17099988e-01 -7.62474895e-01 6.65830910e-01 4.13223594e-01
-1.25008142e+00 1.69168398e-01 -3.68389599e-02 6.01363182e-01
3.30659300e-01 7.57917523e-01 -4.22610611e-01 1.09760118e+00
1.06839335e+00 7.63278425e-01 -4.58977640e-01 5.93434632e-01
1.94510743e-01 1.94007158e-01 -3.90432090e-01 7.73123384e-01
9.27408487e-02 -1.92175716e-01 7.87228286e-01 1.08602524e+00
1.19858518e-01 -3.14791381e-01 3.10302466e-01 5.46711862e-01
4.54925120e-01 1.19217113e-01 -4.23942864e-01 4.70758229e-01
2.65060127e-01 9.45291340e-01 -6.41857803e-01 -1.70943737e-01
-5.27121961e-01 1.01871741e+00 -1.10608526e-01 1.36438325e-01
-8.82142484e-01 -8.77513736e-02 7.64063716e-01 1.73983946e-01
1.87405512e-01 8.84110294e-03 -6.10800624e-01 -1.08047950e+00
1.82537660e-01 -9.57850516e-01 7.07472801e-01 -6.80141151e-01
-1.47273529e+00 5.79522133e-01 3.16030979e-01 -1.62965024e+00
-4.90881465e-02 -5.77955425e-01 -3.87079380e-02 9.09651637e-01
-1.63415337e+00 -1.72431374e+00 -2.56225526e-01 8.71660173e-01
5.37845969e-01 -2.61540234e-01 6.42684639e-01 9.21843648e-01
-2.85605013e-01 6.75408661e-01 -2.38292050e-02 -2.74510056e-01
1.00044215e+00 -1.18964803e+00 7.74979442e-02 6.25159085e-01
-1.72126934e-01 8.65564495e-02 7.03085959e-01 -7.06275344e-01
-8.82681251e-01 -9.86051023e-01 7.44231641e-01 -5.45805395e-01
2.27861002e-01 -2.06720427e-01 -9.50046659e-01 2.97194749e-01
-2.42440179e-01 -1.45107850e-01 4.56575572e-01 2.52408348e-02
-2.73428857e-01 -2.01393843e-01 -1.58336675e+00 2.67641366e-01
1.20675087e+00 -4.89158154e-01 -6.36993349e-01 3.74699980e-01
7.52863348e-01 -6.90728486e-01 -1.16126382e+00 8.45420718e-01
4.15631920e-01 -1.02065349e+00 1.21047568e+00 -2.46376172e-01
7.02451244e-02 -5.86085260e-01 -5.43694079e-01 -1.16875732e+00
1.01095550e-01 -2.39164233e-01 2.89189816e-01 1.56064034e+00
4.69820768e-01 -6.45854414e-01 9.48796391e-01 4.24659461e-01
-2.72600323e-01 -3.93262386e-01 -1.16069889e+00 -1.02081728e+00
2.31621176e-01 -7.27748871e-01 1.00138605e+00 1.29088271e+00
-3.58256340e-01 9.54139903e-02 -4.43606526e-01 3.41143042e-01
5.41817725e-01 4.77876812e-01 8.08424115e-01 -1.23665619e+00
-1.70365140e-01 -2.60089219e-01 -4.59418148e-01 -1.22997701e+00
2.00644568e-01 -7.98488498e-01 -8.62196162e-02 -1.26141274e+00
1.13648295e-01 -7.90231347e-01 -6.50669867e-03 4.08398956e-01
1.51857004e-01 3.48219424e-01 2.00816952e-02 3.52377415e-01
-2.60137618e-01 5.93502343e-01 1.45282006e+00 -1.47397012e-01
-3.81658375e-01 1.46928892e-01 -3.89570653e-01 7.11790919e-01
6.26281857e-01 -3.73004287e-01 -4.37441796e-01 -3.83478791e-01
-2.80906498e-01 -8.76546726e-02 2.04148203e-01 -9.98353958e-01
-3.09473742e-02 -2.39717692e-01 1.14794664e-01 -1.05704916e+00
4.86107528e-01 -1.17825830e+00 3.89638901e-01 3.36449631e-02
1.65256292e-01 -1.66162491e-01 3.08876723e-01 2.11181656e-01
-6.57681286e-01 -7.65022561e-02 1.10747623e+00 -3.31794880e-02
-1.20401478e+00 2.08761901e-01 2.20925845e-02 9.51425657e-02
1.06984651e+00 -6.88582063e-01 -1.95451856e-01 -2.77933739e-02
-1.02677417e+00 4.74212438e-01 6.35215640e-01 7.02045858e-01
2.78315783e-01 -1.64164758e+00 -4.35646951e-01 3.29961926e-01
4.07458842e-01 2.25916937e-01 7.43533731e-01 1.02487969e+00
-2.23940030e-01 1.55132025e-01 -2.40182474e-01 -1.13141227e+00
-1.38793373e+00 2.92748421e-01 5.68110228e-01 -2.75861949e-01
-5.87581992e-01 8.69102478e-01 2.65503436e-01 -1.15691578e+00
-7.54871219e-02 1.63504928e-01 -3.47042978e-01 2.39601135e-01
5.49085811e-02 2.13680819e-01 3.64544868e-01 -1.23475420e+00
-6.38966441e-01 1.15062141e+00 1.46774665e-01 -1.24418154e-01
1.21321034e+00 -5.59896767e-01 -7.70578161e-02 3.78004134e-01
1.14934158e+00 -7.08138198e-02 -1.18275011e+00 -3.82545948e-01
4.00117934e-02 -5.86481631e-01 -4.93397787e-02 -9.73176301e-01
-1.43582225e+00 6.85911119e-01 1.14397538e+00 8.06076005e-02
1.54453528e+00 2.22663671e-01 1.05362570e+00 -4.16768014e-01
2.83406019e-01 -1.49567986e+00 -1.17630638e-01 5.44509053e-01
5.47095954e-01 -1.53685653e+00 -2.30109110e-01 -8.17527235e-01
-9.24388468e-01 7.84758270e-01 9.24983025e-01 2.45685190e-01
1.07502306e+00 2.65834369e-02 6.45583153e-01 -3.67847770e-01
-2.26455897e-01 -5.21811843e-01 6.08002305e-01 9.85052228e-01
1.48251846e-01 7.56866336e-02 -2.49479771e-01 4.99062836e-01
-1.52090684e-01 -2.57812142e-01 -2.22841278e-01 8.50224197e-01
-2.27296278e-01 -1.50209916e+00 -6.94978833e-01 1.31165579e-01
-1.49815798e-01 2.84391046e-01 -2.21763983e-01 1.02298582e+00
6.98860288e-01 9.70884442e-01 -3.99740003e-02 -7.21766353e-01
6.82599843e-01 1.15342535e-01 3.76042604e-01 -3.65919352e-01
-4.79347229e-01 1.66657254e-01 1.61378860e-01 -6.67058170e-01
-8.63517106e-01 -7.21120834e-01 -1.36259675e+00 -3.05332035e-01
-6.06385112e-01 -1.40042081e-01 7.94667184e-01 1.09677577e+00
4.88061845e-01 2.89862782e-01 7.35321701e-01 -6.21676326e-01
-1.43605739e-01 -8.17007422e-01 -6.53295457e-01 8.90944779e-01
1.72777936e-01 -1.22753847e+00 -2.80674875e-01 -2.95064878e-02]
|
[9.679058074951172, 1.360481858253479]
|
c7da2153-7e56-4d59-b182-7ccd0b51dc9e
|
siammot-siamese-multi-object-tracking
|
2105.11595
| null |
https://arxiv.org/abs/2105.11595v1
|
https://arxiv.org/pdf/2105.11595v1.pdf
|
SiamMOT: Siamese Multi-Object Tracking
|
In this paper, we focus on improving online multi-object tracking (MOT). In particular, we introduce a region-based Siamese Multi-Object Tracking network, which we name SiamMOT. SiamMOT includes a motion model that estimates the instance's movement between two frames such that detected instances are associated. To explore how the motion modelling affects its tracking capability, we present two variants of Siamese tracker, one that implicitly models motion and one that models it explicitly. We carry out extensive quantitative experiments on three different MOT datasets: MOT17, TAO-person and Caltech Roadside Pedestrians, showing the importance of motion modelling for MOT and the ability of SiamMOT to substantially outperform the state-of-the-art. Finally, SiamMOT also outperforms the winners of ACM MM'20 HiEve Grand Challenge on HiEve dataset. Moreover, SiamMOT is efficient, and it runs at 17 FPS for 720P videos on a single modern GPU. Codes are available in \url{https://github.com/amazon-research/siam-mot}.
|
['Joseph Tighe', 'Davide Modolo', 'Xinyu Li', 'Andrew Berneshawi', 'Bing Shuai']
|
2021-05-25
| null |
http://openaccess.thecvf.com//content/CVPR2021/html/Shuai_SiamMOT_Siamese_Multi-Object_Tracking_CVPR_2021_paper.html
|
http://openaccess.thecvf.com//content/CVPR2021/papers/Shuai_SiamMOT_Siamese_Multi-Object_Tracking_CVPR_2021_paper.pdf
|
cvpr-2021-1
|
['online-multi-object-tracking']
|
['computer-vision']
|
[-4.75825340e-01 -4.64816242e-01 -2.89622605e-01 4.53946888e-02
-5.73767781e-01 -4.79568958e-01 4.66460973e-01 -9.95678529e-02
-6.28937542e-01 5.67881227e-01 -1.31252438e-01 8.75815451e-02
1.70920193e-01 -4.48528081e-01 -9.35160398e-01 -4.71184462e-01
-3.97908896e-01 7.05261886e-01 1.09678352e+00 -1.12800915e-02
-3.91771682e-02 3.57690126e-01 -1.54784787e+00 8.06989800e-03
4.72014785e-01 6.82137012e-01 1.31511226e-01 1.11638176e+00
3.32306713e-01 8.82442176e-01 -4.91870344e-01 -4.59509373e-01
3.61753464e-01 4.75117564e-02 -4.97758985e-01 -2.75102109e-01
1.16289139e+00 -2.41903499e-01 -6.98674858e-01 9.13048685e-01
3.33112478e-01 1.91712528e-01 3.97971213e-01 -1.88234341e+00
-3.83723170e-01 5.21272361e-01 -8.86371076e-01 9.30652380e-01
7.99014941e-02 5.44055164e-01 7.97239482e-01 -7.82077849e-01
8.67542088e-01 1.42192745e+00 8.60885143e-01 5.62000632e-01
-9.70852137e-01 -8.09439063e-01 2.89546162e-01 4.93514478e-01
-1.48981988e+00 -5.10938644e-01 1.87139556e-01 -5.99566877e-01
7.18274772e-01 2.35544026e-01 7.67577589e-01 9.93188441e-01
2.91330963e-01 1.11062574e+00 7.64019370e-01 8.65276530e-02
-1.45277426e-01 -2.03187913e-01 3.89461160e-01 8.97123694e-01
6.16340578e-01 4.27807152e-01 -6.25162601e-01 -1.18368074e-01
6.36236250e-01 -8.32541361e-02 -5.57228774e-02 -5.17507970e-01
-1.54392481e+00 5.97276688e-01 4.44577515e-01 7.89902508e-02
-2.33288810e-01 8.07351172e-01 4.13054168e-01 -1.13827124e-01
3.44395936e-01 -2.78163463e-01 -1.33064806e-01 -3.80572379e-01
-9.48048174e-01 5.66027105e-01 6.48032725e-01 1.26386189e+00
4.23939407e-01 -5.69466949e-02 -4.44731653e-01 3.51525873e-01
4.50098425e-01 8.03595901e-01 -4.13670950e-02 -1.32466841e+00
4.03876483e-01 1.41840592e-01 3.20471287e-01 -8.00554454e-01
-3.00909013e-01 -2.29006916e-01 -3.43446881e-01 8.13181177e-02
7.35077322e-01 -3.24892610e-01 -6.48810506e-01 1.62234509e+00
8.04791808e-01 9.18050289e-01 -4.52188492e-01 1.07378721e+00
8.46542835e-01 4.85494494e-01 2.09401265e-01 2.27021694e-01
1.42369723e+00 -1.60843468e+00 -4.88525122e-01 -2.52772897e-01
6.97791338e-01 -5.58138371e-01 3.86310279e-01 1.23563491e-01
-1.12677610e+00 -6.64086044e-01 -7.49197543e-01 1.39655381e-01
-3.58276367e-01 8.53909478e-02 5.75890899e-01 5.92849255e-01
-1.07725406e+00 6.08850896e-01 -1.31614625e+00 -6.66397989e-01
6.71361923e-01 3.82950455e-01 1.09379860e-02 -4.71535400e-02
-7.89275706e-01 9.15380657e-01 2.76169777e-01 -2.42271185e-01
-1.09166467e+00 -8.10414433e-01 -5.74171245e-01 -4.38556820e-02
5.18051147e-01 -7.64995754e-01 1.28782928e+00 -4.38791871e-01
-9.55364227e-01 7.30828643e-01 -3.79893273e-01 -8.02515686e-01
8.99757504e-01 -5.27932525e-01 -6.00811064e-01 1.91775486e-01
2.99759924e-01 1.08539009e+00 6.73282802e-01 -1.03234172e+00
-1.13257170e+00 -1.22314617e-01 -9.34360698e-02 -8.45150277e-02
-1.80349305e-01 3.29532474e-01 -1.02428091e+00 -3.52503091e-01
-5.43409050e-01 -1.29524732e+00 -1.33030087e-01 4.74711061e-01
-2.05011159e-01 -5.78693748e-01 1.27831578e+00 -1.90322682e-01
1.09434783e+00 -2.07306457e+00 -1.37208223e-01 -1.92352027e-01
4.78644818e-01 4.41549987e-01 -2.69281089e-01 8.11581165e-02
4.70042229e-01 -2.39201203e-01 4.30887759e-01 -5.59472919e-01
1.45091191e-01 1.20929882e-01 1.53959811e-01 7.92825520e-01
-6.48034364e-02 1.15714872e+00 -9.99052346e-01 -8.60444784e-01
3.71039450e-01 6.52890682e-01 -5.14345407e-01 -2.63726413e-01
-2.59528697e-01 4.92490530e-01 -3.55114400e-01 7.57172585e-01
8.16194654e-01 -6.43409908e-01 -1.47121549e-01 -5.27846999e-02
-4.09350425e-01 -1.32534146e-01 -1.24305296e+00 1.49436510e+00
1.18259907e-01 9.39143836e-01 3.99552919e-02 -3.37729782e-01
3.22052389e-01 7.73143023e-02 7.34879136e-01 -4.91016835e-01
1.28382131e-01 7.05891848e-02 1.63455983e-03 -9.12066996e-02
8.54195237e-01 5.73963404e-01 3.05055857e-01 1.99155018e-01
-7.78369457e-02 6.19035423e-01 7.42841482e-01 4.99709219e-01
1.11225700e+00 4.27711517e-01 -4.63136621e-02 -4.31343913e-01
3.29823703e-01 3.57197464e-01 7.76664674e-01 1.10462892e+00
-8.56531084e-01 2.52080023e-01 4.26123030e-02 -5.00513315e-01
-9.19177592e-01 -1.18227065e+00 -3.14618312e-02 1.26119816e+00
6.57942593e-01 -5.07533193e-01 -5.14114857e-01 -6.14482522e-01
3.77951264e-01 2.67089635e-01 -5.16782641e-01 2.78133810e-01
-1.01330388e+00 -6.22425854e-01 6.29509032e-01 7.36398995e-01
5.48770666e-01 -8.05955350e-01 -9.91373360e-01 1.67934895e-01
-2.53538519e-01 -1.38583016e+00 -1.02716947e+00 -3.32551718e-01
-6.83484316e-01 -1.20774639e+00 -8.85905027e-01 -4.59146589e-01
1.94002882e-01 7.75320947e-01 1.30038655e+00 3.29349816e-01
-3.93948346e-01 5.01403451e-01 -2.69045889e-01 -4.04499471e-01
-6.87046051e-02 1.84018433e-01 1.41048178e-01 -1.78066939e-01
5.71869135e-01 -1.46578029e-01 -7.61852264e-01 7.25162864e-01
-2.06106499e-01 -7.88745135e-02 3.38021725e-01 2.95180827e-01
5.48744678e-01 -3.21621567e-01 5.68138435e-02 -4.95371133e-01
-2.76425630e-01 -6.82461143e-01 -1.02989328e+00 1.66578040e-01
-1.94863632e-01 -2.39693716e-01 9.29601192e-02 -6.98819637e-01
-6.08108044e-01 1.13424346e-01 1.44147441e-01 -8.45211804e-01
1.04930522e-02 -3.27296227e-01 2.69589633e-01 -4.14980233e-01
3.18552196e-01 4.74604629e-02 -1.45042956e-01 -3.40056509e-01
3.33465666e-01 1.31544501e-01 8.05946708e-01 -6.08773589e-01
9.76884305e-01 8.66618872e-01 1.55101558e-02 -7.59875238e-01
-6.94967210e-01 -8.42625618e-01 -4.99060005e-01 -6.69837892e-01
1.07450342e+00 -1.12867546e+00 -1.25592220e+00 4.17457283e-01
-1.06256342e+00 -5.91112137e-01 -1.18781455e-01 5.46375394e-01
-3.85109127e-01 1.90999225e-01 -6.88803136e-01 -7.43250191e-01
-2.27219358e-01 -1.05652058e+00 1.20364523e+00 4.54288453e-01
-7.97114819e-02 -1.04649222e+00 2.45385975e-01 4.35790807e-01
3.80864978e-01 2.88136750e-01 -2.11581200e-01 -3.41218740e-01
-1.31744576e+00 -1.10423351e-02 -4.23328608e-01 -4.38154995e-01
-3.77322465e-01 2.74714261e-01 -7.25202739e-01 -7.35076785e-01
-7.87197888e-01 -8.12092423e-02 1.17763150e+00 6.86073124e-01
8.66235435e-01 -1.12993084e-02 -1.03749120e+00 6.97653234e-01
1.45133281e+00 3.12380232e-02 3.68759334e-01 5.19292772e-01
8.93822014e-01 1.39897401e-02 8.61871600e-01 2.36630574e-01
8.39790463e-01 1.08381903e+00 6.09266281e-01 1.13191262e-01
-5.24124861e-01 -1.53487280e-01 4.40078974e-01 5.81567883e-01
-2.72000372e-01 -5.33423126e-01 -8.33527684e-01 7.51976430e-01
-2.16914296e+00 -1.26330924e+00 -6.72792137e-01 2.06265306e+00
9.64752734e-02 3.78271967e-01 8.77656639e-01 -4.70533609e-01
9.07579839e-01 3.02281737e-01 -6.08902752e-01 2.62204766e-01
-7.23974630e-02 -1.50712177e-01 9.65237260e-01 4.07521456e-01
-1.46433496e+00 1.04953039e+00 5.80333042e+00 8.72829318e-01
-5.99299669e-01 6.28278017e-01 1.41390543e-02 -4.84574437e-01
4.78287280e-01 4.85631526e-02 -1.49027359e+00 7.39589095e-01
1.22311604e+00 -2.31756881e-01 9.87161845e-02 7.88758278e-01
1.46764502e-01 -3.28219444e-01 -9.64860916e-01 8.07928920e-01
-9.78888273e-02 -1.55845857e+00 -3.32692504e-01 3.29514742e-01
6.28063679e-01 6.83419526e-01 3.76476310e-02 4.27025616e-01
7.15233743e-01 -5.73066413e-01 1.04619825e+00 4.39289689e-01
3.14668775e-01 -5.32794178e-01 4.25108403e-01 1.67448416e-01
-2.00436640e+00 1.01076819e-01 -2.84169197e-01 2.59432673e-01
4.68092471e-01 1.02027692e-01 -3.96590382e-01 4.97564167e-01
1.06956542e+00 9.83066022e-01 -8.67230952e-01 1.68992567e+00
2.34391570e-01 5.21756828e-01 -5.99261940e-01 -1.77796245e-01
2.79965341e-01 2.22928643e-01 9.93552744e-01 1.44211864e+00
1.59444720e-01 -1.42991975e-01 5.63927770e-01 7.15311766e-01
6.72710640e-03 -4.72859085e-01 -3.57185990e-01 2.08418533e-01
7.04148531e-01 1.30500412e+00 -9.21635389e-01 -7.38243043e-01
-5.55533946e-01 8.41263950e-01 1.95412904e-01 1.31764680e-01
-1.49987745e+00 1.06494173e-01 1.00578392e+00 1.50846928e-01
8.08439672e-01 -3.82916123e-01 2.41120264e-01 -1.18329203e+00
-1.65625662e-01 -4.30360675e-01 5.23403525e-01 -6.63870156e-01
-9.60340142e-01 3.66611928e-01 7.40610138e-02 -1.28188002e+00
2.08021641e-01 -4.36025620e-01 -6.63108349e-01 3.43964785e-01
-1.57530522e+00 -1.16574371e+00 -5.54746032e-01 6.26354635e-01
5.88692605e-01 1.23899899e-01 8.02863687e-02 7.61001229e-01
-7.97144115e-01 6.30078197e-01 9.85724851e-02 3.62530857e-01
7.50744283e-01 -1.14438331e+00 8.79073620e-01 1.02201426e+00
2.00767949e-01 3.95683318e-01 8.10498416e-01 -7.71767557e-01
-1.51703906e+00 -1.50607181e+00 6.18577898e-01 -9.57497478e-01
8.53849769e-01 -2.74909467e-01 -6.81682825e-01 1.09930813e+00
2.29253188e-01 6.73378170e-01 1.74583405e-01 -1.94284067e-01
-2.09966362e-01 1.97971255e-01 -9.51341867e-01 6.22703314e-01
1.41998696e+00 9.17396396e-02 -1.47733003e-01 5.40914118e-01
7.33566642e-01 -7.97757983e-01 -9.73550141e-01 4.56668250e-02
7.62886226e-01 -8.66223812e-01 1.25121212e+00 -4.19879019e-01
-1.05012655e-01 -5.79004109e-01 -8.52332488e-02 -8.04395914e-01
-5.25856495e-01 -5.78320086e-01 -7.91521370e-01 1.02395046e+00
6.24289699e-02 -5.98512769e-01 1.01017606e+00 2.01094687e-01
3.91059369e-02 -5.09307086e-01 -9.77801979e-01 -1.43542588e+00
2.96155065e-02 -1.21755272e-01 4.44136411e-01 6.67851925e-01
-5.81307352e-01 -1.29128173e-01 -5.21402001e-01 4.71906662e-01
1.36683559e+00 -4.16310653e-02 1.19301689e+00 -1.16841388e+00
-2.29845881e-01 -3.63334626e-01 -5.99195302e-01 -1.34813976e+00
-4.03300710e-02 -7.62515247e-01 -1.80979505e-01 -1.14815331e+00
4.89570856e-01 -3.64857495e-01 -2.20097259e-01 3.62551868e-01
-2.71576285e-01 5.34606755e-01 7.72735178e-01 4.53177392e-01
-1.52130401e+00 3.39125842e-01 1.17972016e+00 -2.52292454e-01
-2.06207167e-02 -5.45977242e-02 -1.66982919e-01 5.19143343e-01
6.41466498e-01 -8.18831861e-01 2.08984643e-01 -5.70440590e-01
-4.08686608e-01 3.38874012e-02 9.55908358e-01 -1.50911295e+00
6.56564176e-01 -2.25017536e-02 2.30688006e-01 -1.07786572e+00
5.93815029e-01 -6.06653810e-01 3.41983914e-01 8.07806671e-01
2.40653604e-02 4.04583663e-01 4.61569965e-01 6.48374915e-01
1.79450944e-01 2.17881650e-01 9.77807879e-01 5.04982062e-02
-1.11456215e+00 7.88969636e-01 -2.53663361e-01 4.07372385e-01
1.35735261e+00 -2.94149011e-01 -8.18755567e-01 8.25855508e-02
-5.72935402e-01 7.92503953e-01 4.98837799e-01 6.84674144e-01
3.44035506e-01 -1.54830444e+00 -8.04702699e-01 -3.23758781e-01
1.57887802e-01 -4.33399647e-01 3.54709387e-01 1.41599727e+00
-4.34836090e-01 5.89999020e-01 -3.30841877e-02 -1.09116626e+00
-1.67158318e+00 5.57596982e-01 3.56725961e-01 -1.58375576e-01
-1.02606189e+00 7.53283083e-01 5.08381315e-02 -1.98630989e-02
2.68661529e-01 -1.66148961e-01 1.26282543e-01 -3.40483814e-01
6.53861225e-01 8.94352436e-01 -5.40003896e-01 -9.01018679e-01
-8.09493840e-01 5.44646204e-01 -1.76308975e-01 -1.24048977e-03
1.00889456e+00 -1.34325549e-01 3.11067849e-01 2.69922525e-01
8.61150205e-01 -2.25311741e-01 -1.63437140e+00 -2.05894426e-01
1.35071382e-01 -7.55806863e-01 -1.23050600e-01 -4.11887169e-01
-1.32969677e+00 4.88457352e-01 8.91110241e-01 -5.05636111e-02
4.72204804e-01 1.28623396e-01 1.09961653e+00 1.84352636e-01
6.51085198e-01 -8.55231524e-01 1.35989804e-02 4.67465937e-01
2.57615358e-01 -1.25489378e+00 1.56025775e-03 -3.51084977e-01
-3.96847576e-01 7.31519997e-01 9.31370080e-01 -2.95687884e-01
5.66957951e-01 3.86311948e-01 -1.76263273e-01 -3.56387049e-01
-8.72018635e-01 -5.02724171e-01 2.40183055e-01 5.67514062e-01
1.00743279e-01 -2.80715562e-02 3.83053645e-02 -3.25569451e-01
4.63891178e-02 3.24575990e-01 3.58158410e-01 9.97383595e-01
-5.03685236e-01 -8.09860706e-01 -6.30175054e-01 2.73294508e-01
-3.31593901e-01 2.24436969e-01 -3.86841856e-02 1.13370657e+00
1.80454433e-01 9.31426525e-01 2.12990478e-01 -1.23454973e-01
2.39784271e-01 -3.72579724e-01 5.88810921e-01 -2.08598241e-01
-6.25130117e-01 -1.67604424e-02 1.18521951e-01 -9.89411175e-01
-6.64075017e-01 -1.10001993e+00 -1.13600576e+00 -8.87614667e-01
-2.76795030e-01 -3.09714745e-03 3.83819699e-01 6.94461405e-01
4.20301288e-01 6.01685107e-01 1.41736763e-02 -1.11366856e+00
-1.24118239e-01 -6.66051984e-01 -3.66216540e-01 2.85835534e-01
4.71191227e-01 -1.11442041e+00 -1.28516361e-01 -1.14956729e-01]
|
[6.329741954803467, -2.0412309169769287]
|
3343f38d-8304-40f6-8a65-522b1cbefaea
|
teravr-empowers-precise-reconstruction-of
| null | null |
https://doi.org/10.1038/s41467-019-11443-y
|
https://www.nature.com/articles/s41467-019-11443-y.pdf
|
TeraVR empowers precise reconstruction of complete 3-D neuronal morphology in the whole brain
|
Neuron morphology is recognized as a key determinant of cell type, yet the quantitative profiling of a mammalian neuron’s complete three-dimensional (3-D) morphology remains arduous when the neuron has complex arborization and long projection. Whole-brain reconstruction of neuron morphology is even more challenging as it involves processing tens of teravoxels of imaging data. Validating such reconstructions is extremely laborious. We develop TeraVR, an open-source virtual reality annotation system, to address these challenges. TeraVR integrates immersive and collaborative 3-D visualization, interaction, and hierarchical streaming of teravoxel-scale images. Using TeraVR, we have produced precise 3-D full morphology of long-projecting neurons in whole mouse brains and developed a collaborative workflow for highly accurate neuronal reconstruction.
|
['Li-Juan Liu', 'Michael Hawrylycz', 'Zhi Zhou', 'Zongcai Ruan', 'Yike Guo', 'Hongkui Zeng', 'Zhongze Gu', 'Yun Wang', 'Yimin Wang', 'Wei Xie', 'Qingming Luo', 'Lingsheng Kong', 'Renjie Chai', 'Qi Li', 'Yaoyao Li', 'Xiangfeng Luo', 'Ning Zhong', 'Hanchuan Peng']
|
2019-08-02
| null | null | null |
nature-communicationsvolume-10-article-number
|
['electron-microscopy-image-segmentation']
|
['computer-vision']
|
[-2.87969291e-01 -4.51516002e-01 7.90879607e-01 -3.67298275e-01
-5.70058048e-01 -9.24259961e-01 3.33004445e-01 3.41160685e-01
-8.47541094e-01 4.64505345e-01 -9.86327156e-02 -3.86993200e-01
1.67728350e-01 -4.49945271e-01 -6.68648064e-01 -6.15782678e-01
3.66154476e-03 6.39031053e-01 5.52345455e-01 -1.43885612e-01
4.08357590e-01 9.73042130e-01 -1.44399130e+00 1.94014102e-01
3.58312160e-01 6.62442386e-01 6.64214432e-01 9.69405293e-01
-1.96646690e-01 -2.26417221e-02 -4.41930801e-01 -4.30334732e-03
1.28454432e-01 2.48974003e-02 -4.30320323e-01 -3.07401419e-01
4.98206288e-01 -2.19974920e-01 9.86513793e-02 7.76092410e-01
6.41991317e-01 -4.51813251e-01 4.76814270e-01 -8.42196047e-01
-1.74813420e-01 -3.81328142e-03 -2.27931201e-01 6.62868917e-01
2.75853574e-01 -6.43425912e-04 4.79527056e-01 -9.41468298e-01
1.18960822e+00 9.21918988e-01 5.65103889e-01 6.43539310e-01
-1.58891356e+00 -6.54953301e-01 -4.00756076e-02 -3.04378659e-01
-1.36345816e+00 -4.92807388e-01 -8.77221301e-02 -9.84993398e-01
1.09276950e+00 2.96913888e-02 1.34576643e+00 6.60339415e-01
8.53728116e-01 7.59321377e-02 1.22414696e+00 1.51999727e-01
4.38083917e-01 -4.78117764e-01 2.40106314e-01 7.15415776e-01
1.83854476e-01 -4.93907064e-01 -3.26185763e-01 -6.94836453e-02
1.38761139e+00 4.34983611e-01 -3.17217298e-02 -1.10811085e-01
-1.49213231e+00 1.02557115e-01 -1.21488206e-01 1.29086688e-01
-1.04022734e-01 2.93245614e-01 2.79783100e-01 -1.37752462e-02
6.57165423e-02 3.33863944e-01 -3.73099357e-01 -3.94982189e-01
-7.83856809e-01 3.43226999e-01 2.25651518e-01 8.21868062e-01
4.85183448e-01 2.85223518e-02 5.90035439e-01 6.71131551e-01
2.79599577e-01 1.85497582e-01 4.71254557e-01 -1.39537370e+00
-1.89774320e-01 8.30697477e-01 -2.48762771e-01 -4.81156886e-01
-8.49453390e-01 -1.79664806e-01 -6.38873219e-01 1.10667336e+00
8.53096247e-01 -1.02977343e-01 -7.23458946e-01 1.26834297e+00
5.49789310e-01 -3.38945270e-01 -2.89605260e-01 8.92670631e-01
9.33657885e-01 3.84343773e-01 -1.31780595e-01 2.19277203e-01
1.24743140e+00 -1.11726910e-01 -3.51594985e-01 1.27346084e-01
6.02858961e-01 -3.12559575e-01 1.06958127e+00 4.56185192e-01
-1.05560112e+00 9.46783274e-02 -9.22542810e-01 -3.95879745e-01
-4.21474427e-01 -2.21200183e-01 4.54880893e-01 1.71994165e-01
-1.26911211e+00 4.17280883e-01 -1.16604936e+00 -4.46251094e-01
1.08029175e+00 3.29788387e-01 -1.22162235e+00 1.09049790e-01
2.47693006e-02 8.64822268e-01 -1.43446118e-01 -2.99952090e-01
-8.29951346e-01 -1.05102110e+00 -3.19715232e-01 -4.40730304e-02
-2.52021641e-01 -7.50371635e-01 1.23586011e+00 1.52528957e-01
-1.26952267e+00 1.30982149e+00 -1.15713328e-01 3.09759118e-02
3.03121567e-01 2.20879659e-01 4.07160036e-02 2.03683719e-01
-1.75172284e-01 6.41793251e-01 -2.47985255e-02 -1.11672640e+00
-5.28754234e-01 -1.22100937e+00 -3.37340266e-01 2.85222698e-02
2.36309215e-01 2.85939723e-01 -2.51854509e-01 -1.58854768e-01
5.93494475e-01 -2.87543446e-01 -4.52166408e-01 7.71594286e-01
-6.77672494e-03 4.11284924e-01 8.42930138e-01 -5.20902097e-01
4.96114820e-01 -2.12410593e+00 5.67123853e-02 -1.61395177e-01
8.07385802e-01 -6.38931543e-02 2.36739591e-01 2.67535448e-01
2.09510237e-01 3.38918865e-01 1.55036906e-02 -4.59813684e-01
-2.77225763e-01 -9.87471044e-02 -7.07350224e-02 7.49779999e-01
-5.24474740e-01 6.07770562e-01 -6.03377342e-01 -5.56032598e-01
1.74852014e-01 8.63653898e-01 -5.08352101e-01 -1.28749833e-01
1.73540100e-01 7.90312409e-01 -2.32123956e-01 7.55112886e-01
4.55312431e-01 -2.05700338e-01 1.84442222e-01 1.56611681e-01
-7.30103374e-01 -1.26460567e-01 -5.77440500e-01 1.80809891e+00
-2.71803051e-01 1.08163869e+00 5.34796178e-01 -3.20584774e-01
8.51186275e-01 6.90264180e-02 3.85385901e-01 -6.40786767e-01
3.01778257e-01 1.70166552e-01 -1.52021706e-01 -2.09401846e-01
6.36793301e-02 -3.56583744e-01 5.46600148e-02 3.96781027e-01
2.67061889e-01 -3.86260152e-01 1.91580534e-01 2.40196794e-01
1.45983291e+00 1.88409984e-01 2.30257183e-01 -4.08818007e-01
-3.10776740e-01 1.68295890e-01 6.68601871e-01 7.93004334e-02
-2.93756157e-01 1.06713378e+00 7.16214001e-01 -4.15950745e-01
-1.43765056e+00 -1.32118571e+00 -5.83821476e-01 7.63647079e-01
-1.14943326e-01 -2.00018778e-01 -7.73715258e-01 4.04634029e-02
-5.11297844e-02 2.37696037e-01 -5.91010094e-01 6.61385119e-01
-3.89339596e-01 -6.05031967e-01 6.75936997e-01 1.87190577e-01
9.33049247e-02 -8.21501076e-01 -1.31825161e+00 1.80891052e-01
3.65631104e-01 -1.10505688e+00 -4.62075733e-02 2.88419992e-01
-8.36066604e-01 -1.06934452e+00 -7.08236992e-01 -7.21629798e-01
9.60740924e-01 3.37806582e-01 6.59129202e-01 -3.90955806e-02
-8.59891355e-01 -1.25805642e-02 -6.03644848e-02 -3.18786025e-01
5.23421504e-02 -5.27136505e-01 1.15548961e-01 -7.59239435e-01
-2.05769762e-01 -1.32789958e+00 -5.83669364e-01 3.77968192e-01
-7.34429479e-01 3.83882314e-01 1.51943877e-01 1.86744183e-01
1.58411586e+00 -2.45099798e-01 3.70456845e-01 -1.04093909e+00
4.29606944e-01 -3.15119445e-01 -1.14755058e+00 -5.44408895e-02
-1.10571273e-01 -5.09472251e-01 9.57045376e-01 -2.77834743e-01
-6.44149601e-01 2.47486159e-01 -5.40353596e-01 -5.79578206e-02
-5.17104268e-01 2.37645760e-01 -2.81600598e-02 -1.61842868e-01
5.40298581e-01 3.28952640e-01 3.45351607e-01 -4.12779957e-01
2.86031086e-02 3.61640334e-01 9.24773633e-01 -1.76747888e-01
2.41545364e-01 9.45517421e-01 3.18016969e-02 -1.00891292e+00
-3.14836740e-01 -3.04452568e-01 -9.86307085e-01 -6.06349826e-01
1.00836754e+00 -6.44940853e-01 -1.03214037e+00 5.87856412e-01
-8.88593614e-01 -1.05279148e+00 -3.00659072e-02 4.37552243e-01
-5.69494963e-01 -7.12299952e-03 -5.61550677e-01 -4.57183510e-01
-3.78535151e-01 -1.21592617e+00 9.07464266e-01 2.52506316e-01
-3.57129157e-01 -5.44637322e-01 6.79814696e-01 3.67981568e-02
1.80818766e-01 5.18239617e-01 1.00991035e+00 -1.48561999e-01
-5.02848566e-01 -3.33956510e-01 -1.93437636e-01 -5.29843390e-01
-3.00003827e-01 5.65627694e-01 -7.87728608e-01 1.98312491e-01
-4.38368201e-01 -3.47469449e-01 3.79069448e-01 4.43820238e-01
1.06414282e+00 1.61117956e-01 -3.57231081e-01 1.00610447e+00
1.44806278e+00 1.64311036e-01 5.80841839e-01 3.60293478e-01
4.21898305e-01 5.76402843e-01 1.87459648e-01 3.97920191e-01
4.57106501e-01 5.52284718e-01 5.80754638e-01 5.40156737e-02
-9.48899537e-02 2.24685267e-01 5.96531704e-02 7.68422067e-01
-1.89060390e-01 1.60207674e-01 -1.02680957e+00 1.97719082e-01
-1.27915514e+00 -1.06242728e+00 -5.37766397e-01 2.05093026e+00
5.70277393e-01 7.07740709e-03 3.03159893e-01 -8.24003592e-02
3.72412920e-01 -4.45200175e-01 -8.66927743e-01 -2.58513778e-01
-4.11187857e-01 -1.56164065e-01 3.21655929e-01 2.67604560e-01
-4.06460911e-01 7.42286384e-01 7.89443541e+00 2.73655295e-01
-1.36353350e+00 1.91401973e-01 4.83454138e-01 -4.70466346e-01
-3.62277001e-01 -1.68021634e-01 -9.78060603e-01 3.72808784e-01
6.11074746e-01 -2.26105154e-02 6.14528656e-01 6.36563361e-01
3.64500910e-01 -4.59541380e-01 -9.68264699e-01 1.21010697e+00
-4.01350945e-01 -1.85984123e+00 -2.83821255e-01 5.98261833e-01
2.26473153e-01 8.30994785e-01 -1.76618934e-01 -1.69851363e-01
4.89174604e-01 -1.12324500e+00 9.82127666e-01 7.78373480e-01
1.21698415e+00 -6.26222491e-01 2.71616280e-01 6.69701934e-01
-9.56413567e-01 2.05734119e-01 -6.57315791e-01 1.83503027e-03
4.82215643e-01 6.48433566e-01 -6.40538454e-01 -3.64889383e-01
9.84717906e-01 2.96612054e-01 -5.76986253e-01 1.32234001e+00
4.61571217e-01 8.79116431e-02 -4.70239609e-01 8.78272504e-02
-5.67294419e-01 -4.28723574e-01 4.43715930e-01 1.17001069e+00
5.37058890e-01 4.38848108e-01 -4.27566379e-01 1.00201738e+00
7.97775239e-02 6.66711032e-02 -7.28523135e-01 -4.12730575e-01
7.22133696e-01 1.83722126e+00 -1.39015114e+00 4.86165918e-02
-1.37699127e-01 5.69595158e-01 8.32555592e-01 1.12930555e-02
-3.43208462e-01 -2.53333122e-01 1.03857327e+00 2.68561095e-01
3.78821380e-02 -8.63731444e-01 -7.08900571e-01 -9.28151309e-01
-2.58904994e-01 -1.46300822e-01 -3.73142064e-02 -1.06494701e+00
-7.93525398e-01 6.56264484e-01 -5.84439635e-01 -8.26342463e-01
2.30881840e-01 -8.34023714e-01 -7.05328584e-01 4.66278017e-01
-6.42315388e-01 -8.89569998e-01 -2.97433168e-01 3.67254138e-01
1.58226535e-01 -4.99352664e-02 1.18182111e+00 2.38253072e-01
-8.85445595e-01 7.47790793e-03 2.30113819e-01 -2.24559620e-01
1.89427242e-01 -1.35444224e+00 3.89833659e-01 4.32121098e-01
4.08818349e-02 6.78129017e-01 7.35697150e-01 -5.78380883e-01
-1.58632743e+00 -1.04249275e+00 3.38309258e-01 -6.09255731e-01
5.52697897e-01 -6.95222855e-01 -7.72935033e-01 9.41224456e-01
-3.14336151e-01 4.38960344e-01 1.04058945e+00 -2.21713230e-01
-4.50467616e-01 1.24417275e-01 -1.33795297e+00 9.06756938e-01
1.10171604e+00 -3.90211135e-01 -1.81958646e-01 -2.77778450e-02
2.90473223e-01 -2.85615563e-01 -1.23632348e+00 -2.00141162e-01
1.20857191e+00 -1.16756105e+00 6.98088765e-01 9.46704149e-02
2.33554229e-01 -5.91045380e-01 -2.02435493e-01 -1.21167672e+00
-1.74007982e-01 -2.46359050e-01 3.81180048e-01 9.07078266e-01
4.89922941e-01 -6.23336017e-01 8.45639825e-01 4.62463886e-01
-6.80000901e-01 -1.00465906e+00 -1.02449751e+00 -4.08050954e-01
2.14325935e-01 -3.68405163e-01 3.93467337e-01 5.59769750e-01
5.30898094e-01 7.27377608e-02 6.09413743e-01 -4.73405048e-02
6.12885475e-01 -1.10676862e-01 9.69542027e-01 -1.43231678e+00
1.19383410e-01 -7.02072024e-01 -8.52077007e-01 -6.49993539e-01
-2.24593386e-01 -1.00205183e+00 -6.55131564e-02 -1.84290600e+00
4.42284018e-01 -2.65080452e-01 1.96957588e-01 3.33869100e-01
5.54414809e-01 5.59693336e-01 -1.66378543e-01 2.71299481e-01
-8.08222950e-01 1.52125314e-01 1.37140846e+00 5.14965296e-01
8.62821415e-02 -4.43398029e-01 -4.76809442e-01 9.74808753e-01
7.35126734e-01 -3.48415643e-01 -5.77493049e-02 -7.05450773e-01
4.64823753e-01 1.63579479e-01 4.22302097e-01 -1.01149166e+00
4.42854255e-01 -9.63560678e-03 6.91694558e-01 -1.06637812e+00
6.04059756e-01 -5.63525140e-01 5.32298744e-01 1.64369270e-01
-9.85870734e-02 3.74696404e-01 1.38206154e-01 2.84528434e-01
2.86875397e-01 2.64854848e-01 9.86527443e-01 -3.92313212e-01
-3.84041309e-01 3.77574861e-01 -1.19454062e+00 -6.78755715e-02
9.40479100e-01 -7.08701134e-01 -7.48711526e-01 8.39891136e-02
-9.01051879e-01 1.10901464e-02 1.33737481e+00 -4.86086220e-01
1.01425397e+00 -1.04103875e+00 -3.40954870e-01 4.34300071e-03
3.18733091e-03 3.85737091e-01 5.40444493e-01 8.64441454e-01
-1.31519365e+00 1.33573219e-01 -9.03748095e-01 -6.38909817e-01
-1.54959309e+00 5.29399812e-02 2.59506792e-01 3.23785663e-01
-1.10908878e+00 8.23701143e-01 3.80974382e-01 -7.29521215e-01
-1.18510194e-01 -3.08910847e-01 -4.99388993e-01 -1.35244250e-01
8.48734379e-01 4.37991470e-01 1.23406671e-01 -7.46237099e-01
-2.24161446e-01 7.04846919e-01 9.10709277e-02 -4.54306841e-01
1.69349003e+00 -4.80029911e-01 -3.44145030e-01 7.97190666e-01
9.46517289e-01 2.91195028e-02 -1.50461972e+00 5.34034848e-01
-4.08155203e-01 -2.95650035e-01 5.98492064e-02 -6.60169125e-01
-9.22557235e-01 1.17386782e+00 3.63340378e-01 -1.57882780e-01
5.89622915e-01 3.05453427e-02 7.63646126e-01 4.98787969e-01
9.46786225e-01 -7.67174184e-01 -2.39270732e-01 7.48807490e-01
8.18090796e-01 -6.78975821e-01 -2.19211672e-02 -3.05165619e-01
-1.62704792e-02 1.18439794e+00 6.27344310e-01 -2.33063921e-01
7.15158999e-01 8.97984684e-01 2.53061086e-01 -5.92670023e-01
-1.05290067e+00 3.02231550e-01 -3.80833417e-01 6.10083580e-01
4.53645200e-01 -1.13512330e-01 6.48093298e-02 6.88506722e-01
-5.27851522e-01 2.22616438e-02 7.09815919e-01 9.44332004e-01
-7.34526455e-01 -6.36998057e-01 -1.86956912e-01 5.18087685e-01
-6.70736194e-01 2.60868132e-01 -3.13454717e-01 4.76365060e-01
6.49048612e-02 2.57532567e-01 3.91648293e-01 -2.84327656e-01
2.15499699e-02 -1.72197774e-01 8.00441265e-01 -8.10523570e-01
-2.58754671e-01 2.60643274e-01 -1.40119791e-01 -6.51903510e-01
2.92516232e-01 -8.24973941e-01 -2.16002011e+00 -5.09862721e-01
2.02936560e-01 -3.79458904e-01 1.35868418e+00 7.87690341e-01
6.11516833e-01 4.44517463e-01 -2.94666756e-02 -1.20846188e+00
3.85788679e-01 -5.94073653e-01 -9.59226429e-01 -1.94142848e-01
1.83102354e-01 -4.06545252e-01 -5.61418235e-02 3.04256499e-01]
|
[14.152273178100586, -3.119887113571167]
|
5adb05c3-3d0b-4420-bbaa-e12723aa582b
|
nlip-noise-robust-language-image-pre-training
|
2212.07086
| null |
https://arxiv.org/abs/2212.07086v2
|
https://arxiv.org/pdf/2212.07086v2.pdf
|
NLIP: Noise-robust Language-Image Pre-training
|
Large-scale cross-modal pre-training paradigms have recently shown ubiquitous success on a wide range of downstream tasks, e.g., zero-shot classification, retrieval and image captioning. However, their successes highly rely on the scale and quality of web-crawled data that naturally contain incomplete and noisy information (e.g., wrong or irrelevant content). Existing works either design manual rules to clean data or generate pseudo-targets as auxiliary signals for reducing noise impact, which do not explicitly tackle both the incorrect and incomplete challenges simultaneously. In this paper, to automatically mitigate the impact of noise by solely mining over existing data, we propose a principled Noise-robust Language-Image Pre-training framework (NLIP) to stabilize pre-training via two schemes: noise-harmonization and noise-completion. First, in noise-harmonization scheme, NLIP estimates the noise probability of each pair according to the memorization effect of cross-modal transformers, then adopts noise-adaptive regularization to harmonize the cross-modal alignments with varying degrees. Second, in noise-completion scheme, to enrich the missing object information of text, NLIP injects a concept-conditioned cross-modal decoder to obtain semantic-consistent synthetic captions to complete noisy ones, which uses the retrieved visual concepts (i.e., objects' names) for the corresponding image to guide captioning generation. By collaboratively optimizing noise-harmonization and noise-completion schemes, our NLIP can alleviate the common noise effects during image-text pre-training in a more efficient way. Extensive experiments show the significant performance improvements of our NLIP using only 26M data over existing pre-trained models (e.g., CLIP, FILIP and BLIP) on 12 zero-shot classification datasets, MSCOCO image captioning and zero-shot image-text retrieval tasks.
|
['Xiaodan Liang', 'Chunjing Xu', 'Xiwen Liang', 'Hang Xu', 'Jianhua Han', 'Yanxin Long', 'Runhui Huang']
|
2022-12-14
| null | null | null | null |
['memorization']
|
['natural-language-processing']
|
[ 5.91482580e-01 -2.85598993e-01 5.77478809e-03 -2.08494619e-01
-1.36970842e+00 -4.55372840e-01 6.30186319e-01 -9.12600011e-02
-4.05861229e-01 5.42302370e-01 4.17503208e-01 -1.37935653e-01
3.60654965e-02 -6.19511366e-01 -1.00265360e+00 -6.76924884e-01
7.24722981e-01 3.24144244e-01 1.25113025e-01 -3.12813133e-01
8.39397237e-02 -3.53652269e-01 -1.88642490e+00 5.74121714e-01
9.74670172e-01 9.21707094e-01 5.67973495e-01 4.97683138e-01
-5.93273461e-01 5.56738615e-01 -5.74635506e-01 -5.50846696e-01
2.99546242e-01 -6.75252140e-01 -4.53100950e-01 1.93342835e-01
2.08467230e-01 -1.58465296e-01 -3.87159526e-01 1.32095575e+00
8.38353336e-01 1.21676542e-01 4.88432288e-01 -1.30296099e+00
-9.11353528e-01 6.96808338e-01 -7.40088105e-01 -5.43635711e-02
2.59917021e-01 5.32025874e-01 9.34926152e-01 -1.17699969e+00
7.67100692e-01 1.29950857e+00 4.65168864e-01 8.04542899e-01
-1.35014403e+00 -9.61864233e-01 -1.07304379e-01 2.74198323e-01
-1.61041224e+00 -5.91356039e-01 7.89458990e-01 -2.92005569e-01
3.72384489e-01 3.25504422e-01 2.55525291e-01 1.57778978e+00
-3.54658186e-01 8.34270716e-01 1.02423704e+00 -5.55913389e-01
1.77898303e-01 2.34546527e-01 -4.63675745e-02 2.73028642e-01
5.93198389e-02 -1.95428580e-01 -4.92620468e-01 1.20319275e-03
4.09766197e-01 -2.10592613e-01 -2.91322291e-01 -2.50183791e-01
-1.27707350e+00 6.28603876e-01 2.04656556e-01 3.94924015e-01
-1.85241655e-01 -1.24842137e-01 5.24617374e-01 1.61818281e-01
2.83575386e-01 3.25021625e-01 -3.95180494e-01 -3.51710953e-02
-9.63921845e-01 1.70546949e-01 3.23947579e-01 1.18546951e+00
9.56389904e-01 2.52343528e-03 -8.48385930e-01 1.34320354e+00
8.43950659e-02 8.61942470e-01 7.44750679e-01 -6.85214400e-01
8.57289910e-01 3.99535000e-01 1.72974281e-02 -9.44052994e-01
8.52945074e-03 -3.98102850e-01 -1.03567529e+00 -2.83981234e-01
3.40089761e-02 -1.11475572e-01 -1.36272538e+00 2.14452863e+00
1.92852065e-01 4.70902205e-01 2.25004062e-01 9.79935944e-01
1.00940681e+00 8.55034769e-01 3.64773095e-01 -3.07512581e-01
1.50889981e+00 -1.01730502e+00 -8.87123466e-01 -2.27469042e-01
4.33464795e-01 -9.04226720e-01 1.53768086e+00 1.27042264e-01
-7.28415132e-01 -6.43450558e-01 -8.73606324e-01 -9.14257616e-02
-3.67268920e-01 9.82681662e-02 4.90328297e-02 4.05379444e-01
-6.23296440e-01 1.15994729e-01 -1.77284241e-01 -1.66673154e-01
5.18601179e-01 -1.25084788e-01 -3.55620623e-01 -4.33444202e-01
-1.48572993e+00 5.66636384e-01 5.85089624e-01 -5.46529591e-02
-9.44576561e-01 -7.91351378e-01 -9.54493880e-01 1.13344252e-01
7.49666214e-01 -7.21629381e-01 9.93785262e-01 -1.01085436e+00
-1.04701281e+00 8.58793437e-01 -9.82758105e-02 -2.96902657e-01
4.44272578e-01 -9.10793915e-02 -4.67028230e-01 -1.10530639e-02
4.63018537e-01 9.56204176e-01 1.08344686e+00 -1.69045937e+00
-4.37557995e-01 -1.69680446e-01 -3.18608582e-01 4.56474900e-01
-4.65259880e-01 -1.29140764e-01 -1.12225246e+00 -9.35436010e-01
9.26782098e-03 -6.72919929e-01 -1.33983135e-01 -1.20504066e-01
-4.69139785e-01 -2.02861498e-03 8.18596005e-01 -6.91367090e-01
1.27937615e+00 -2.23329520e+00 5.20604476e-02 -5.29810414e-03
-6.55437037e-02 4.88980532e-01 -7.32838094e-01 2.68596947e-01
-2.68831160e-02 1.87154606e-01 -4.99286830e-01 -6.37101173e-01
-9.88737792e-02 4.86187190e-01 -4.53881860e-01 -8.50933269e-02
3.22281271e-01 1.10180712e+00 -1.02899015e+00 -8.91254365e-01
3.51516962e-01 4.23872143e-01 -3.33696425e-01 2.22410411e-01
-3.92577320e-01 3.53436053e-01 -4.24304426e-01 5.55147827e-01
6.66483760e-01 -2.77009189e-01 -1.58520535e-01 -5.09501219e-01
2.75457174e-01 -2.78006434e-01 -1.34174955e+00 2.04671383e+00
-5.31864703e-01 3.37284029e-01 1.78282931e-02 -9.75091040e-01
7.42080152e-01 2.74388462e-01 3.70470941e-01 -1.01450133e+00
2.72003472e-01 1.03130378e-01 -4.92533445e-01 -8.17789912e-01
4.16667640e-01 -1.52408257e-01 -2.42927790e-01 2.60717452e-01
3.24132234e-01 -7.21581355e-02 3.11562091e-01 5.02575397e-01
9.51947272e-01 2.33763456e-02 -2.17528213e-02 1.41208634e-01
6.76177800e-01 -4.81721982e-02 5.38562059e-01 1.06089747e+00
-1.42485321e-01 1.22278810e+00 2.64100522e-01 1.62615493e-01
-1.25976694e+00 -9.86098051e-01 -5.45742363e-02 1.10521388e+00
5.11901557e-01 -3.78499866e-01 -8.81860852e-01 -5.61253309e-01
-3.72834116e-01 8.15238059e-01 -5.51766038e-01 -3.62130046e-01
-2.47612596e-01 -7.29609966e-01 6.84197068e-01 1.99599221e-01
4.89144057e-01 -1.26665080e+00 2.10075546e-02 1.05873071e-01
-7.07818389e-01 -1.30057335e+00 -7.23405659e-01 7.49260606e-03
-2.00902894e-01 -8.61989856e-01 -8.54217231e-01 -8.71682286e-01
7.04290271e-01 6.95293665e-01 1.05576015e+00 8.27387273e-02
-4.12498683e-01 3.53113145e-01 -6.23080194e-01 -1.67959243e-01
-3.99615139e-01 -2.32036129e-01 -1.89163566e-01 3.54277551e-01
3.28401893e-01 -4.63223279e-01 -5.55213153e-01 3.07480067e-01
-1.44576359e+00 3.32586825e-01 7.64198720e-01 1.24946237e+00
7.28740454e-01 2.47868858e-02 4.77054775e-01 -6.78348720e-01
8.46067607e-01 -6.45823836e-01 -2.17954144e-01 5.47769666e-01
-4.84828770e-01 1.31823272e-01 6.60091758e-01 -7.77917802e-01
-1.20850706e+00 7.34443963e-02 -7.65996873e-02 -9.30032790e-01
-1.14691712e-01 3.44818592e-01 -3.90083909e-01 2.67629564e-01
7.55168080e-01 7.26040483e-01 -1.31549373e-01 -3.49914551e-01
7.55679309e-01 8.37271869e-01 8.15392613e-01 -6.69987440e-01
1.00219023e+00 2.56569535e-01 -4.37424839e-01 -6.49344504e-01
-9.50835168e-01 -6.68839693e-01 -3.22723836e-02 -9.70276818e-02
8.94990265e-01 -1.18881249e+00 -2.47917455e-02 5.74888527e-01
-1.24275947e+00 1.04223318e-01 -3.30601990e-01 1.21726267e-01
-3.73620123e-01 5.17774820e-01 -3.50434840e-01 -7.01177776e-01
-5.19467413e-01 -1.25799549e+00 1.37428010e+00 2.36367270e-01
1.45260513e-01 -3.18491668e-01 -8.26959014e-02 6.07764781e-01
2.95657486e-01 -1.66137338e-01 8.44534695e-01 -5.88715017e-01
-4.71769601e-01 -7.65015557e-02 -6.09212279e-01 5.43517888e-01
-6.63702711e-02 -3.94003272e-01 -9.86516833e-01 -7.28826672e-02
-2.29815617e-01 -7.08949685e-01 8.56270671e-01 1.84523702e-01
1.26102853e+00 -2.96320796e-01 -1.57408655e-01 4.38747078e-01
1.50255048e+00 -9.11288243e-03 9.23816800e-01 1.29631057e-01
7.04685926e-01 5.72867095e-01 6.52722418e-01 4.27794844e-01
2.07290813e-01 6.35964036e-01 2.58863330e-01 -6.37040734e-02
-4.84336257e-01 -6.47843957e-01 4.64920215e-02 7.84026384e-01
4.63218719e-01 -4.55304205e-01 -5.79963505e-01 7.09649980e-01
-1.86635363e+00 -9.63553846e-01 4.78553288e-02 1.99205077e+00
1.18123102e+00 1.01820521e-01 -3.27530026e-01 -8.91232491e-02
1.09387505e+00 1.81990623e-01 -5.14183939e-01 4.20498461e-01
-5.38637578e-01 1.25666231e-01 2.79078752e-01 2.65613109e-01
-9.73503292e-01 1.04112709e+00 4.47537994e+00 1.52235329e+00
-7.82878876e-01 3.39793891e-01 6.98653698e-01 -1.08557560e-01
-6.77843034e-01 3.30078656e-05 -6.14819288e-01 8.76756847e-01
5.45055926e-01 -1.10193439e-01 5.10825455e-01 7.54973710e-01
1.93556026e-01 -9.86362901e-03 -5.49819231e-01 1.38899863e+00
3.91339362e-01 -1.27165008e+00 5.04178941e-01 -3.69437218e-01
8.55776668e-01 -8.75980780e-02 9.29105654e-02 6.17696226e-01
1.47525147e-01 -7.12366045e-01 8.03710639e-01 4.99348015e-01
1.09882820e+00 -3.97445321e-01 7.02613056e-01 4.53798145e-01
-9.85184252e-01 -1.01983234e-01 -3.09866905e-01 4.66543913e-01
2.97384202e-01 6.26329958e-01 -3.36813956e-01 5.49918175e-01
6.65740311e-01 2.77797043e-01 -5.93571424e-01 8.31143022e-01
-1.85144335e-01 4.84712809e-01 -3.03501636e-01 1.52642578e-01
1.61625788e-01 -1.64992034e-01 6.31369174e-01 1.16857982e+00
3.09583545e-01 2.61888266e-01 -3.59529629e-02 9.24666643e-01
-3.46336931e-01 2.73283660e-01 -4.03495312e-01 -9.14543644e-02
6.32707357e-01 1.33322823e+00 -5.08010387e-01 -4.70077038e-01
-4.21255380e-01 1.08324146e+00 8.80035460e-02 4.67252165e-01
-9.79124248e-01 -5.48346519e-01 5.39707363e-01 -2.34312434e-02
2.42038533e-01 1.61871120e-01 -2.60490686e-01 -1.41253471e+00
3.34306687e-01 -1.06006742e+00 2.32440248e-01 -1.17892265e+00
-1.47600698e+00 5.88147223e-01 -1.40698522e-01 -1.32118332e+00
7.38160983e-02 -1.87250242e-01 -5.18141031e-01 7.14962959e-01
-1.43327868e+00 -1.33747399e+00 -5.41837990e-01 7.39663482e-01
8.90993237e-01 -6.80804104e-02 5.31212628e-01 6.48232341e-01
-5.49069762e-01 7.30103016e-01 1.27079487e-01 1.40029609e-01
1.01115739e+00 -7.62118340e-01 1.83418468e-01 9.51485038e-01
3.38334918e-01 4.43324149e-01 7.39473343e-01 -7.64139056e-01
-1.29212546e+00 -1.25384951e+00 5.40635943e-01 -3.26103926e-01
6.25400662e-01 -5.93664944e-01 -1.03826046e+00 2.52931923e-01
8.40039179e-02 5.30144647e-02 2.98371792e-01 -3.70288402e-01
-4.85210925e-01 -1.56015769e-01 -8.75727892e-01 7.86918104e-01
1.16907120e+00 -6.34317100e-01 -5.89173973e-01 3.46927971e-01
1.20641589e+00 -2.95768380e-01 -2.62848705e-01 4.09686983e-01
2.64377952e-01 -7.17234194e-01 9.98697162e-01 -6.30679190e-01
6.88382685e-01 -5.04907668e-01 -2.82206118e-01 -1.05960536e+00
-1.56877682e-01 -5.77747822e-01 1.63830176e-01 1.64093029e+00
4.14549738e-01 -2.11758241e-02 5.95542848e-01 3.37181926e-01
-1.27857730e-01 -4.25214678e-01 -9.36381340e-01 -5.20304084e-01
-3.30527455e-01 -6.43971622e-01 4.71624285e-01 9.11741972e-01
-4.36926425e-01 6.23672545e-01 -8.66198242e-01 -3.48886177e-02
6.41701162e-01 -1.60697475e-01 8.82690907e-01 -8.13510776e-01
-2.40980238e-01 -2.34878853e-01 -9.79576409e-02 -9.85198796e-01
1.56673208e-01 -7.45252252e-01 5.29716134e-01 -1.44012952e+00
4.89279538e-01 -2.66693711e-01 -4.12700623e-01 4.83114034e-01
-5.89528203e-01 3.68653208e-01 2.72809267e-01 2.57203728e-01
-1.01451755e+00 7.98848093e-01 1.30184805e+00 -3.93114984e-01
-1.43635258e-01 -4.02208894e-01 -8.01637590e-01 2.92847067e-01
4.93674219e-01 -5.56391239e-01 -5.66166043e-01 -4.78097141e-01
1.33574054e-01 8.55353624e-02 4.44926709e-01 -1.03972006e+00
2.65661240e-01 -1.23602450e-01 2.41090089e-01 -5.18697500e-01
3.35278749e-01 -7.95118749e-01 9.88285020e-02 2.60614082e-02
-4.57866848e-01 -3.24828476e-01 8.31059441e-02 9.48802352e-01
-3.53002906e-01 -4.11164373e-01 7.73547590e-01 -2.47605935e-01
-8.64801884e-01 3.83279860e-01 -3.09627447e-02 4.09529388e-01
9.38822985e-01 -7.63987973e-02 -4.15784299e-01 -4.38515395e-01
-5.29096127e-01 4.48436052e-01 3.39355916e-01 8.89838755e-01
6.59754336e-01 -1.40384793e+00 -7.98766315e-01 3.30975711e-01
5.60002685e-01 6.56943619e-02 8.05756867e-01 5.09457469e-01
5.65222418e-03 7.87359402e-02 1.21687306e-02 -6.22941613e-01
-1.18547702e+00 7.46885300e-01 -2.23903009e-03 -2.66142309e-01
-4.57738787e-01 7.58775473e-01 3.24169248e-01 -2.69330800e-01
2.92241573e-01 2.27837414e-01 5.18646128e-02 2.83967014e-02
6.40074074e-01 -8.77428278e-02 3.75691466e-02 -6.60948098e-01
-2.69113332e-02 6.52614653e-01 -7.34927729e-02 -2.50484675e-01
1.12271690e+00 -5.00814617e-01 -3.02408226e-02 2.03328222e-01
1.28503227e+00 -2.37427756e-01 -1.11407876e+00 -6.03221595e-01
-1.31225228e-01 -3.77468169e-01 -1.21965459e-04 -7.59045184e-01
-9.57709968e-01 7.55026519e-01 5.93451858e-01 -1.95627183e-01
1.28226626e+00 3.78924720e-02 1.05263162e+00 2.72804767e-01
1.09266810e-01 -1.29777515e+00 3.70737076e-01 3.65105927e-01
7.26424634e-01 -1.55213714e+00 -4.03205752e-01 -2.97535121e-01
-8.72076213e-01 6.22902393e-01 8.20207179e-01 3.42484355e-01
3.75120282e-01 1.18396118e-01 1.48766816e-01 7.48237222e-02
-6.65594280e-01 -4.92056519e-01 2.58016288e-01 6.12043917e-01
-2.55612522e-01 -2.43027017e-01 -2.62605727e-01 8.84196997e-01
1.49199337e-01 -8.44158456e-02 2.42007613e-01 6.54064953e-01
-4.85292196e-01 -9.15051997e-01 -4.91731375e-01 4.34029222e-01
-1.59999624e-01 -5.07604361e-01 -1.67282999e-01 4.42514300e-01
3.98951799e-01 1.04471648e+00 -9.74054560e-02 -5.10155559e-01
4.87764567e-01 2.30851039e-01 3.24082933e-02 -5.20066082e-01
-1.94026321e-01 2.22498909e-01 -1.75984383e-01 -2.93045670e-01
-3.18318009e-01 -2.22281784e-01 -8.90273988e-01 1.03200980e-01
-4.53448057e-01 1.63983256e-01 6.32729352e-01 9.42159295e-01
5.56761920e-01 5.24340332e-01 5.77389061e-01 -5.75509131e-01
-7.15834618e-01 -1.08331859e+00 -3.52387309e-01 1.13117611e+00
2.15218775e-02 -6.36424422e-01 -4.60673988e-01 2.74985522e-01]
|
[11.082064628601074, 0.8100055456161499]
|
6e5daadf-5e91-4a44-a75c-ce284bf1a505
|
83-imagenet-accuracy-in-one-hour
|
2011.00071
| null |
https://arxiv.org/abs/2011.00071v2
|
https://arxiv.org/pdf/2011.00071v2.pdf
|
Training EfficientNets at Supercomputer Scale: 83% ImageNet Top-1 Accuracy in One Hour
|
EfficientNets are a family of state-of-the-art image classification models based on efficiently scaled convolutional neural networks. Currently, EfficientNets can take on the order of days to train; for example, training an EfficientNet-B0 model takes 23 hours on a Cloud TPU v2-8 node. In this paper, we explore techniques to scale up the training of EfficientNets on TPU-v3 Pods with 2048 cores, motivated by speedups that can be achieved when training at such scales. We discuss optimizations required to scale training to a batch size of 65536 on 1024 TPU-v3 cores, such as selecting large batch optimizers and learning rate schedules as well as utilizing distributed evaluation and batch normalization techniques. Additionally, we present timing and performance benchmarks for EfficientNet models trained on the ImageNet dataset in order to analyze the behavior of EfficientNets at scale. With our optimizations, we are able to train EfficientNet on ImageNet to an accuracy of 83% in 1 hour and 4 minutes.
|
['Sameer Kumar', 'Yang You', 'Quoc Le', 'Mingxing Tan', 'James Demmel', 'Hieu Pham', 'Arissa Wongpanich']
|
2020-10-30
| null | null | null | null |
['2048']
|
['playing-games']
|
[-3.76407802e-01 -4.92247902e-02 -1.84933722e-01 -7.23112226e-01
-1.40979856e-01 -4.38838482e-01 6.06578924e-02 8.75493810e-02
-1.00688148e+00 4.06683087e-01 -4.32916522e-01 -6.47206783e-01
-1.69364020e-01 -8.61026645e-01 -9.40810919e-01 -2.64124185e-01
-2.55207777e-01 5.54851353e-01 2.79166341e-01 3.20966281e-02
-1.29750282e-01 7.54340887e-01 -1.65069222e+00 2.42706612e-01
3.33398014e-01 1.36879897e+00 1.67729005e-01 1.20620322e+00
-6.45921826e-02 1.03043592e+00 -9.20555472e-01 -3.07763815e-01
6.73960686e-01 3.15283924e-01 -8.15464199e-01 -2.86963314e-01
8.96323204e-01 -7.57000506e-01 -4.46654558e-01 8.02524149e-01
4.49121237e-01 6.50297198e-03 -1.11101687e-01 -1.40899062e+00
-3.46919931e-02 8.63839507e-01 1.27574116e-01 5.52355886e-01
-7.33612001e-01 3.61467928e-01 8.11665833e-01 -3.69212061e-01
4.50524211e-01 8.96524191e-01 7.99918354e-01 5.72190881e-01
-9.28775549e-01 -9.68633592e-01 -1.02099515e-01 3.52187455e-01
-1.25069344e+00 -6.75863624e-01 5.54405898e-02 -1.01316832e-01
1.77287745e+00 1.36106014e-01 8.55100632e-01 6.37208700e-01
1.63285628e-01 5.51678896e-01 6.53122842e-01 -2.00739726e-02
3.19000244e-01 -1.50700137e-01 4.68356431e-01 8.08306694e-01
4.40713495e-01 1.22144364e-01 -3.98570389e-01 9.56740230e-02
6.58728480e-01 -1.36767894e-01 2.60587543e-01 3.18606853e-01
-1.00852370e+00 7.30829060e-01 7.88294792e-01 1.90419182e-01
-3.86138380e-01 8.37851882e-01 9.72731531e-01 4.89502668e-01
5.79882443e-01 7.51828969e-01 -9.86910343e-01 -3.64156663e-01
-1.14580727e+00 2.29105651e-01 9.07108247e-01 1.23355389e+00
1.14356530e+00 4.09328252e-01 -4.18281704e-02 6.04109645e-01
-2.45893542e-02 5.12104094e-01 3.83394331e-01 -1.20888340e+00
3.31003100e-01 5.87396562e-01 -4.62573051e-01 -5.04472926e-02
-5.06755054e-01 -6.07713938e-01 -7.43408561e-01 3.44586104e-01
2.22651780e-01 -5.18131852e-01 -1.04348230e+00 1.32993305e+00
1.01255160e-02 3.96661848e-01 5.36348894e-02 6.03290498e-01
8.00857902e-01 6.36599660e-01 1.60294518e-01 2.91297972e-01
1.32103086e+00 -1.49552679e+00 -1.25172958e-01 -2.88963616e-01
1.16596305e+00 -7.17031419e-01 9.41492200e-01 2.09705606e-02
-1.05130410e+00 -6.60221934e-01 -1.34241462e+00 -1.12677105e-01
-3.91746640e-01 3.36309075e-01 8.76206756e-01 8.20389688e-01
-1.45939529e+00 1.15061712e+00 -1.45581245e+00 -3.38035226e-01
6.13557696e-01 8.66475403e-01 -3.54514122e-01 6.14238679e-02
-5.82378805e-01 7.74219394e-01 7.63520718e-01 -1.48508906e-01
-1.14054632e+00 -1.31667423e+00 -6.18555427e-01 4.18388575e-01
-3.47806662e-02 -6.18048966e-01 1.45242691e+00 -9.22544777e-01
-1.45575440e+00 5.99736512e-01 1.03285372e-01 -1.11207449e+00
1.99154779e-01 -1.13185316e-01 -1.49120733e-01 1.42511889e-01
-1.77751094e-01 9.32737827e-01 4.47439581e-01 -2.59726435e-01
-6.32065833e-01 -1.18740372e-01 2.65353113e-01 -5.18155582e-02
-8.26966763e-01 3.13589871e-01 -3.94078076e-01 -2.26666443e-02
-2.73720741e-01 -9.55568373e-01 -4.10468102e-01 -6.52503818e-02
-8.86152461e-02 -1.15775928e-01 1.06834686e+00 -3.15032661e-01
6.71597481e-01 -1.91839933e+00 -6.27589643e-01 2.17887700e-01
3.82983476e-01 6.27945244e-01 -4.49215889e-01 -2.66093880e-01
3.20493616e-02 1.55838594e-01 3.17345560e-01 -5.42040050e-01
6.91399872e-02 4.52808887e-01 -6.15449548e-02 3.28350395e-01
6.33406118e-02 9.13925290e-01 -4.74210620e-01 -2.86896944e-01
3.18130344e-01 4.75732177e-01 -8.61883640e-01 3.14090759e-01
-1.91774115e-01 -5.69660105e-02 -1.98494941e-01 5.38812101e-01
7.05031574e-01 -5.74598134e-01 1.33633882e-01 -3.50253910e-01
-3.19425017e-01 5.31570673e-01 -7.17317879e-01 1.44476628e+00
-7.67483950e-01 1.07368147e+00 1.06812216e-01 -8.02181721e-01
7.11766243e-01 2.01236635e-01 5.68915069e-01 -6.55286133e-01
3.73686761e-01 3.13824654e-01 1.99310437e-01 4.10363786e-02
5.39003849e-01 3.88871253e-01 4.19943810e-01 4.31753248e-01
6.45906746e-01 1.47611484e-01 6.71015620e-01 -1.41241252e-01
1.61377025e+00 -3.61714900e-01 -2.14798987e-01 -5.95135570e-01
6.84841052e-02 1.71630383e-01 3.62443507e-01 8.48536313e-01
-1.70737058e-01 4.58074473e-02 4.48905140e-01 -1.04737794e+00
-1.62837338e+00 -5.54067254e-01 -3.06020021e-01 1.35056245e+00
-4.28375542e-01 -7.49891758e-01 -9.57359850e-01 -3.81092459e-01
-2.90867656e-01 4.12081152e-01 -2.54183650e-01 1.41296417e-01
-7.37911284e-01 -9.90214169e-01 1.05148721e+00 8.53055954e-01
1.11006105e+00 -9.36976254e-01 -9.42446470e-01 2.90904433e-01
6.07295692e-01 -1.43427646e+00 1.39238359e-02 6.69015408e-01
-1.17447615e+00 -1.14640641e+00 -3.13968599e-01 -6.76621914e-01
7.60356009e-01 2.01467127e-02 1.43812287e+00 5.54667532e-01
-3.74155611e-01 9.31553822e-03 -1.31730303e-01 -3.46682221e-01
-2.82468945e-01 8.10848475e-01 1.20192662e-01 -5.78298032e-01
3.55772704e-01 -6.85391366e-01 -4.97010946e-01 1.77386358e-01
-7.62244463e-01 2.18508497e-01 6.13172174e-01 7.09611237e-01
5.10018468e-01 1.16648130e-01 -2.87591964e-02 -8.53281319e-01
4.74599674e-02 -2.84001172e-01 -1.27974272e+00 1.96847677e-01
-8.77377033e-01 1.94248021e-01 9.83691990e-01 -3.90784353e-01
-5.92860281e-01 5.20382561e-02 -6.16869688e-01 -8.69273365e-01
-1.35251228e-02 1.97035179e-01 3.30623418e-01 -6.18950963e-01
8.42986763e-01 -1.24688260e-01 -5.53056076e-02 -3.39746386e-01
8.10940340e-02 4.97346103e-01 1.59929529e-01 -7.76867628e-01
5.86877108e-01 1.58039808e-01 2.41970923e-02 -9.57968116e-01
-7.38927126e-01 -4.45906639e-01 -2.33221725e-01 -6.74314611e-03
6.64900959e-01 -1.17827141e+00 -1.23148513e+00 4.92543459e-01
-1.10711133e+00 -1.11614931e+00 -2.32561275e-01 5.02028644e-01
-1.35476172e-01 -4.98388886e-01 -1.02034056e+00 -1.89154103e-01
-8.75404596e-01 -1.35994172e+00 8.49202812e-01 4.60887641e-01
1.46625653e-01 -8.88988674e-01 -1.36214837e-01 2.10170180e-01
1.07119334e+00 -3.29253197e-01 6.36375546e-01 -8.82021368e-01
-9.10892189e-01 -2.13563278e-01 -5.20131886e-01 6.85608566e-01
-5.66309690e-01 7.58143216e-02 -1.03859186e+00 -5.89167416e-01
-3.09362113e-01 -5.85750639e-01 7.70828962e-01 4.28482354e-01
1.55661118e+00 -4.87753004e-01 -1.18996784e-01 1.44427514e+00
1.75291526e+00 -3.75818424e-02 5.67682981e-01 5.30873716e-01
7.23903537e-01 -1.52792066e-01 1.84038430e-01 2.89138496e-01
-2.58130133e-02 4.34544206e-01 6.53016567e-01 -1.50905950e-02
-1.18061826e-01 7.11421892e-02 -3.26781794e-02 8.40193033e-01
-3.30121934e-01 5.93273295e-03 -1.05280912e+00 3.43214273e-01
-1.43110907e+00 -6.41752958e-01 7.66055807e-02 1.97135282e+00
4.94070441e-01 2.65057236e-01 -2.59839207e-01 -2.12928295e-01
4.24914509e-01 1.77862912e-01 -6.00145400e-01 -4.65987772e-01
1.13099918e-01 8.65358174e-01 1.36153388e+00 2.89324939e-01
-9.72171009e-01 1.15404379e+00 6.98105097e+00 6.87670112e-01
-1.53880310e+00 4.51111615e-01 8.71390224e-01 -5.70679307e-01
3.31057698e-01 7.87442625e-02 -1.35976708e+00 2.72846550e-01
1.60517204e+00 -4.74872030e-02 6.79892898e-01 1.67356300e+00
-1.97634965e-01 2.87499964e-01 -9.80948150e-01 9.68377709e-01
-2.20116794e-01 -1.81293714e+00 -2.25851804e-01 1.98145822e-01
8.91946435e-01 9.97584462e-01 -2.57538885e-01 7.24138975e-01
4.57120866e-01 -1.19576669e+00 5.49586952e-01 -2.52265245e-01
8.12127471e-01 -9.21988368e-01 1.08650291e+00 2.16471747e-01
-1.19209850e+00 -9.10974145e-02 -1.07260621e+00 -1.78874910e-01
-3.53018284e-01 5.83631754e-01 -9.63860452e-01 -4.73242067e-02
1.17638946e+00 5.66541553e-01 -9.16285574e-01 1.01636088e+00
-2.46226713e-02 8.23432565e-01 -7.64065266e-01 -3.36486220e-01
3.96511018e-01 1.80107296e-01 4.97238338e-02 9.53312874e-01
4.82496291e-01 -1.35447636e-01 -1.70529708e-01 6.54255152e-01
-6.19437218e-01 2.02741120e-02 -2.61417121e-01 -5.97697683e-02
5.30236006e-01 1.72430182e+00 -7.45178223e-01 -5.81816852e-01
-1.99778825e-01 5.27380288e-01 7.38970578e-01 1.35210073e-02
-9.90918279e-01 -4.41804737e-01 1.08190870e+00 1.11821517e-02
2.27325991e-01 -3.40738833e-01 -3.40819836e-01 -1.06706178e+00
-7.77313262e-02 -7.75348783e-01 -4.96733774e-05 -6.59556508e-01
-6.66005313e-01 9.02967453e-01 -1.36451632e-01 -6.95563495e-01
7.61424825e-02 -1.22370934e+00 -7.85507619e-01 5.31172633e-01
-1.36936808e+00 -1.11672974e+00 -6.01094127e-01 4.41763222e-01
3.38488400e-01 -4.92828220e-01 8.89382124e-01 7.14030623e-01
-6.11439586e-01 8.35377812e-01 9.42239389e-02 3.33971798e-01
2.34431475e-01 -9.71783459e-01 8.81015360e-01 9.09062147e-01
9.88883302e-02 6.18168294e-01 5.05354822e-01 -2.45591119e-01
-1.48229051e+00 -1.38618743e+00 6.70269728e-01 3.55004296e-02
7.25142539e-01 -3.27248216e-01 -6.37119889e-01 1.12920296e+00
1.71159431e-01 6.40883029e-01 4.72448945e-01 4.31301981e-01
-3.92540842e-01 -6.61966085e-01 -9.18608725e-01 5.20304322e-01
1.01078486e+00 -2.41651341e-01 2.34737471e-01 7.24606752e-01
7.27848113e-01 -8.04417789e-01 -1.16380787e+00 3.71421754e-01
3.77403706e-01 -8.40701997e-01 7.72996664e-01 -5.93977332e-01
2.55922228e-01 -4.31204168e-03 -9.17572603e-02 -1.05734742e+00
-4.56554517e-02 -5.55084825e-01 -3.05255298e-02 6.80503309e-01
3.62518966e-01 -8.28956008e-01 1.24189615e+00 5.45069754e-01
-4.54651088e-01 -8.23624969e-01 -9.12364542e-01 -7.95021296e-01
-1.18747294e-01 -7.05697000e-01 9.68778074e-01 4.96187925e-01
-5.66764534e-01 6.53665513e-02 -8.26697946e-02 1.33517250e-01
6.25439823e-01 -3.77860069e-01 8.91414106e-01 -9.21166360e-01
-4.71160382e-01 -4.73278493e-01 -7.54632235e-01 -1.00698233e+00
3.34416717e-01 -8.37050498e-01 -8.97416323e-02 -9.26904023e-01
3.75270285e-02 -7.60157168e-01 -2.86998332e-01 9.80785966e-01
3.16048294e-01 4.76001948e-01 3.25620532e-01 5.62868938e-02
-8.01122546e-01 1.43560261e-01 8.13912511e-01 -1.44717351e-01
1.28158987e-01 -1.66043505e-01 -2.59776354e-01 6.21618211e-01
9.62762713e-01 -5.71624458e-01 -4.25920546e-01 -9.00708377e-01
2.08708778e-01 -3.40292335e-01 3.53755683e-01 -1.67044711e+00
4.69690591e-01 2.12257683e-01 5.12163877e-01 -3.71225208e-01
4.06769425e-01 -6.53629541e-01 1.28807947e-01 6.17367864e-01
-2.40456760e-01 4.76467341e-01 5.85979044e-01 -2.08969951e-01
-1.06381297e-01 -3.88900369e-01 9.86446202e-01 -1.14237331e-01
-7.43303597e-01 7.37503469e-01 -2.51962721e-01 2.44710618e-03
8.97820413e-01 2.60388076e-01 -9.15893614e-01 1.66202530e-01
-2.91396558e-01 2.53552288e-01 4.10125285e-01 1.46058932e-01
3.35233778e-01 -1.07951188e+00 -5.28393269e-01 3.46305639e-01
-8.13205615e-02 4.51303646e-02 3.37390512e-01 5.92906117e-01
-1.58377254e+00 6.83152676e-01 -5.38532674e-01 -6.38014734e-01
-1.49040568e+00 1.26028985e-01 7.95177639e-01 -5.92558146e-01
-5.11701465e-01 8.35496008e-01 -3.06754857e-01 -6.22651100e-01
1.74366981e-01 -4.32070225e-01 2.53117293e-01 -6.56685829e-01
6.45622790e-01 4.32102084e-01 6.78969026e-01 -5.15851602e-02
-3.69703025e-01 1.02882765e-01 -7.19237402e-02 1.09940723e-01
1.41676462e+00 7.21914887e-01 -2.76300281e-01 -1.79171249e-01
1.54324293e+00 -7.13598549e-01 -1.22933936e+00 7.68987909e-02
-2.63527095e-01 -4.29742821e-02 3.88096124e-01 -3.11463505e-01
-1.50876343e+00 7.08788872e-01 8.62684429e-01 -1.75585821e-01
1.06685293e+00 -2.62862504e-01 8.74123633e-01 1.11460006e+00
3.28254342e-01 -1.16027009e+00 -2.08116919e-01 9.89570022e-01
1.38672888e-01 -1.12763786e+00 1.03135675e-01 4.74779494e-02
1.40277088e-01 1.29281414e+00 1.02366912e+00 -4.48239148e-01
8.11132789e-01 7.72403121e-01 9.20800716e-02 -2.17903987e-01
-1.12749124e+00 3.76465246e-02 -1.29721418e-01 1.37037456e-01
3.47578615e-01 2.04312935e-01 1.13870203e-01 5.85953109e-02
-6.67884052e-01 -7.94515684e-02 3.13188463e-01 9.14455235e-01
-3.73908848e-01 -9.92651641e-01 -3.14138038e-03 8.70214820e-01
-6.89848244e-01 -4.22342062e-01 4.53107208e-01 8.18391562e-01
1.72990710e-01 6.15275919e-01 6.30846798e-01 -6.22938335e-01
9.59979594e-02 -4.07297052e-02 4.61375386e-01 -5.93881309e-01
-1.12627304e+00 -5.38996041e-01 2.24079713e-01 -7.59134829e-01
-9.97097194e-02 1.45609519e-02 -1.14909708e+00 -1.01897717e+00
-1.60083920e-01 2.83806212e-03 1.45198596e+00 1.02265668e+00
4.15154845e-01 7.51221240e-01 1.71284825e-01 -1.08069789e+00
-5.72329879e-01 -1.13398027e+00 -4.80671704e-01 -6.40946403e-02
-6.17457181e-02 -1.96182523e-02 -3.75178635e-01 -1.93696812e-01]
|
[8.506765365600586, 3.0427117347717285]
|
1ec96127-1856-4f8e-8c30-98a8a4cd8f3a
|
a-realistic-study-of-auto-regressive-language
|
2108.11857
| null |
https://arxiv.org/abs/2108.11857v2
|
https://arxiv.org/pdf/2108.11857v2.pdf
|
Probing Pre-trained Auto-regressive Language Models for Named Entity Typing and Recognition
|
Despite impressive results of language models for named entity recognition (NER), their generalization to varied textual genres, a growing entity type set, and new entities remains a challenge. Collecting thousands of annotations in each new case for training or fine-tuning is expensive and time-consuming. In contrast, humans can easily identify named entities given some simple instructions. Inspired by this, we challenge the reliance on large datasets and study pre-trained language models for NER in a meta-learning setup. First, we test named entity typing (NET) in a zero-shot transfer scenario. Then, we perform NER by giving few examples at inference. We propose a method to select seen and rare / unseen names when having access only to the pre-trained model and report results on these groups. The results show: auto-regressive language models as meta-learners can perform NET and NER fairly well especially for regular or seen names; name irregularity when often present for a certain entity type can become an effective exploitable cue; names with words foreign to the model have the most negative impact on results; the model seems to rely more on name than context cues in few-shot NER.
|
['Romain Hennequin', 'Elena V. Epure']
|
2021-08-26
| null |
https://aclanthology.org/2022.lrec-1.151
|
https://aclanthology.org/2022.lrec-1.151.pdf
|
lrec-2022-6
|
['few-shot-ner']
|
['natural-language-processing']
|
[-2.61259884e-01 -1.13342516e-02 -2.12886587e-01 -2.70800799e-01
-7.07164943e-01 -6.82824790e-01 7.54138649e-01 2.59856820e-01
-1.29220140e+00 9.85841036e-01 3.45787793e-01 -1.57727405e-01
1.28978297e-01 -9.27546442e-01 -6.88840747e-01 -3.20400894e-01
-1.45403564e-01 7.01647282e-01 2.13418946e-01 -3.79879892e-01
1.39009595e-01 5.09623051e-01 -1.41388381e+00 2.45413736e-01
9.88689125e-01 2.05028117e-01 2.62116522e-01 4.94380414e-01
-6.47296309e-01 8.05251122e-01 -9.18711901e-01 -7.74740160e-01
-2.55716704e-02 -1.66794404e-01 -9.91207898e-01 -5.27372599e-01
4.44175333e-01 1.99036419e-01 -1.20158494e-01 8.94493937e-01
9.24858093e-01 6.86105430e-01 9.31854367e-01 -7.09705830e-01
-1.01325274e+00 1.16421616e+00 -2.18445957e-01 5.13797939e-01
3.04894209e-01 1.40918747e-01 8.97883296e-01 -1.11566389e+00
1.01978230e+00 9.92133856e-01 9.71601307e-01 1.14513004e+00
-1.28083432e+00 -6.92136765e-01 1.99167132e-02 1.38302058e-01
-1.50382447e+00 -5.87588608e-01 2.74055302e-01 -3.33803266e-01
1.15260327e+00 1.41644388e-01 -4.15082015e-02 1.72334325e+00
-2.87849873e-01 5.49199343e-01 9.16289926e-01 -5.28447747e-01
1.24087505e-01 6.03136837e-01 2.64304429e-01 3.10226142e-01
2.25111082e-01 1.07412882e-01 -5.09105682e-01 -1.33604020e-01
3.99492294e-01 -1.41807199e-01 -1.47671610e-01 1.12669684e-01
-1.22941041e+00 6.83725059e-01 4.23098594e-01 8.15846145e-01
-4.82806474e-01 -2.12936267e-01 5.44044614e-01 4.11473155e-01
5.75925291e-01 1.08938050e+00 -8.82655799e-01 -2.24275142e-01
-9.12212551e-01 -1.03719331e-01 1.24644995e+00 9.49791312e-01
6.94910228e-01 1.11267706e-02 -6.20853841e-01 1.22376275e+00
-3.52720559e-01 4.52815592e-01 7.92705715e-01 -1.96663529e-01
4.21103239e-01 4.02256370e-01 1.61492258e-01 -5.94036341e-01
-5.92212796e-01 -5.33388436e-01 -8.44680607e-01 -5.50208129e-02
5.42718828e-01 -4.48056698e-01 -1.01046145e+00 1.89596772e+00
2.76394904e-01 3.46834511e-01 6.54860809e-02 4.26451415e-01
9.79134321e-01 5.98041773e-01 7.78000653e-01 -1.66817680e-01
1.46295989e+00 -8.93181682e-01 -5.77985823e-01 -3.31644684e-01
9.06859457e-01 -5.54961622e-01 1.13700235e+00 -6.25554472e-02
-7.31294930e-01 -6.57610476e-01 -5.16697705e-01 -1.43491372e-01
-1.23625147e+00 5.79781272e-03 3.27793092e-01 4.82622802e-01
-8.49087059e-01 8.06004465e-01 -4.14030045e-01 -6.44591630e-01
1.83327511e-01 3.25260967e-01 -3.98292899e-01 -1.03460096e-01
-1.55894566e+00 1.35493028e+00 7.91492999e-01 -2.33532846e-01
-6.66319489e-01 -1.11686313e+00 -8.76487017e-01 1.28413767e-01
4.10246700e-01 -7.85300255e-01 1.11082768e+00 -8.88740838e-01
-1.16377020e+00 1.10374618e+00 -1.79300189e-01 -3.69843632e-01
4.51965898e-01 -3.09806705e-01 -8.17482948e-01 -3.80616695e-01
3.43139648e-01 6.30251229e-01 5.46418130e-01 -1.25167584e+00
-5.44715166e-01 2.31982786e-02 9.49199870e-02 -3.49798123e-03
-5.30252099e-01 3.21829647e-01 5.18761612e-02 -8.14133167e-01
-5.80443680e-01 -6.16550982e-01 -2.03723073e-01 -7.31428742e-01
-5.17733335e-01 -7.06213474e-01 1.64338633e-01 -5.80680668e-01
1.61475396e+00 -2.03719950e+00 2.04657149e-02 -6.75418675e-02
9.75856856e-02 5.22210896e-01 -2.71385580e-01 5.27411461e-01
-1.88382030e-01 5.89941502e-01 2.06958735e-03 -2.01694876e-01
7.73255713e-03 -7.59378299e-02 -1.34947419e-01 -1.01150259e-01
3.95672828e-01 8.32030475e-01 -1.08861589e+00 -6.05677843e-01
-8.96781012e-02 5.16958773e-01 -4.30948794e-01 4.65027153e-01
6.74616247e-02 2.69553751e-01 -2.80695140e-01 4.70626056e-01
2.78731734e-01 -2.77603716e-01 -2.00207755e-01 -1.03654370e-01
-2.74572164e-01 3.95509571e-01 -1.17840290e+00 1.39766324e+00
-9.34085429e-01 2.68123358e-01 -3.51959556e-01 -5.97866654e-01
8.92939568e-01 4.33380961e-01 -2.14616910e-01 -5.24912477e-01
3.81087586e-02 3.51936758e-01 -1.22271463e-01 -6.48106873e-01
5.37350476e-01 -4.02273297e-01 -2.98252195e-01 3.31443906e-01
6.60419762e-01 4.87802714e-01 5.20439446e-01 1.35984514e-02
1.20367026e+00 6.35879487e-02 6.68491602e-01 -7.25257248e-02
3.37016374e-01 -2.98478029e-04 5.37760675e-01 1.22519195e+00
-1.40132621e-01 3.16474915e-01 -1.32220075e-01 -4.68048632e-01
-9.61508870e-01 -9.66444612e-01 -2.47696325e-01 2.10161042e+00
-1.11967869e-01 -2.19525382e-01 -5.27810514e-01 -8.90795290e-01
-6.30266443e-02 1.31031811e+00 -7.69375265e-01 -1.76600188e-01
-6.75931513e-01 -7.10453093e-01 7.18958616e-01 6.17229998e-01
2.23327667e-01 -1.65584040e+00 -1.16177902e-01 3.97252560e-01
4.49110344e-02 -8.13167036e-01 -6.16551399e-01 5.68985939e-01
-5.26840389e-01 -7.04034567e-01 -1.06935215e+00 -8.97828639e-01
5.98749757e-01 -2.30606258e-01 1.62387621e+00 8.37260336e-02
-2.69620031e-01 3.54214907e-01 -4.62440431e-01 -5.18673420e-01
-5.58739901e-01 6.21616244e-01 2.66770780e-01 -1.77288562e-01
7.93085098e-01 -6.09654069e-01 -7.84576014e-02 2.67957568e-01
-6.20203137e-01 -3.60391527e-01 7.47864783e-01 1.12347829e+00
1.01441808e-01 -3.29638183e-01 7.30028868e-01 -1.59793293e+00
4.30679262e-01 -9.49358940e-01 -9.65638533e-02 5.06635904e-01
-4.99180704e-01 4.20289457e-01 7.98898816e-01 -9.27101851e-01
-1.36790860e+00 -9.89110693e-02 -1.74282044e-01 -2.13843450e-01
-5.98037422e-01 5.03939033e-01 -1.51775196e-01 2.50032395e-01
1.23561263e+00 4.19985801e-02 -8.40592027e-01 -8.43043208e-01
5.04117906e-01 5.24408519e-01 4.46781427e-01 -7.58363485e-01
1.01070356e+00 -8.10799003e-02 -6.94954753e-01 -9.00386930e-01
-1.29051018e+00 -6.85927570e-01 -9.67784226e-01 1.41582228e-02
9.37067211e-01 -8.20677817e-01 -3.74698222e-01 1.96555257e-01
-1.32870746e+00 -2.86276251e-01 -3.88731629e-01 4.93182600e-01
-1.00634051e-02 -1.74609974e-01 -6.30387127e-01 -7.25744665e-01
-3.00854236e-01 -5.33989549e-01 7.12974608e-01 5.55720568e-01
-4.22932655e-01 -1.50523007e+00 3.52560490e-01 -7.31687993e-02
7.73700833e-01 -2.20513041e-03 8.76871705e-01 -1.61969328e+00
-9.64485928e-02 -1.46733031e-01 -6.59884978e-03 4.06248234e-02
-7.64368698e-02 -2.18735933e-01 -1.30615234e+00 -2.68687513e-02
-3.68400276e-01 -5.20223200e-01 9.14699256e-01 -1.44339889e-01
8.35125387e-01 -3.59966338e-01 -4.79801923e-01 4.88003254e-01
1.29368412e+00 -6.07885979e-02 3.36826295e-01 6.37133956e-01
8.47127974e-01 5.42852044e-01 2.31298357e-01 2.34165147e-01
2.00907320e-01 4.67597783e-01 -3.37486446e-01 -1.37014315e-01
-1.52654603e-01 -4.93511647e-01 3.89327347e-01 9.22874629e-01
-3.41287106e-01 -3.26023132e-01 -1.15257251e+00 6.60099328e-01
-1.29444873e+00 -1.16657853e+00 2.47745425e-01 2.25947404e+00
1.41309333e+00 1.30452886e-01 -2.64927745e-02 -5.56832492e-01
1.16955185e+00 -2.91556995e-02 -4.33402359e-01 -1.86487958e-01
-2.40281418e-01 5.19358993e-01 4.83675659e-01 2.86973000e-01
-1.28241158e+00 1.19923568e+00 6.21201277e+00 1.03160250e+00
-1.03317630e+00 3.02385092e-01 4.23479557e-01 2.83797711e-01
-1.90643966e-01 -2.01180428e-01 -1.29222560e+00 3.65892112e-01
1.34259856e+00 -3.83019507e-01 2.86642641e-01 9.90606129e-01
-2.71281958e-01 2.22953603e-01 -1.27700734e+00 9.07597542e-01
1.34621695e-01 -9.92636263e-01 2.12744802e-01 -1.58809379e-01
7.87442625e-01 1.89781353e-01 -2.95744389e-01 1.10786569e+00
6.61707401e-01 -1.04114950e+00 4.61457610e-01 7.47109175e-01
7.67585993e-01 -6.87534869e-01 9.35061991e-01 5.23774683e-01
-9.46067095e-01 -1.15340777e-01 -6.37637377e-01 1.73539951e-01
1.00526504e-01 3.84687036e-01 -9.68923092e-01 3.16203654e-01
5.59254408e-01 3.51055175e-01 -8.62791955e-01 1.23584139e+00
-2.97980785e-01 8.10105979e-01 -2.37876341e-01 -2.41049588e-01
-8.68091807e-02 3.89738858e-01 5.54880977e-01 1.76362371e+00
3.07248801e-01 5.93880601e-02 2.25836381e-01 8.73312294e-01
-5.11293828e-01 4.49768633e-01 -7.76008010e-01 -2.57563531e-01
8.34047735e-01 1.32209003e+00 -6.42988205e-01 -5.19055247e-01
-5.08715868e-01 1.18827009e+00 9.11348522e-01 4.58343059e-01
-5.73604941e-01 -7.34063685e-01 3.20526034e-01 -1.29648941e-02
3.69830281e-01 2.26465926e-01 -2.46373519e-01 -1.47059453e+00
-3.71353269e-01 -6.83402359e-01 7.54749358e-01 -5.75022638e-01
-1.90481830e+00 7.74924994e-01 -7.86599070e-02 -1.10139239e+00
-2.90043116e-01 -6.05006218e-01 -7.42932558e-01 8.32768559e-01
-1.44380999e+00 -1.04633224e+00 6.74151175e-04 4.40329194e-01
5.70008397e-01 -3.09464335e-01 1.27056682e+00 5.45469165e-01
-4.96782839e-01 9.70108211e-01 -5.63956518e-03 7.18675494e-01
1.22212064e+00 -1.50230384e+00 4.47710812e-01 6.92227006e-01
4.55611527e-01 9.82013106e-01 7.47575879e-01 -6.57224894e-01
-8.98511112e-01 -1.04374516e+00 1.55210435e+00 -9.74064648e-01
7.17001379e-01 -3.45782250e-01 -1.23391402e+00 6.93646073e-01
1.97321847e-01 1.08856738e-01 8.98033261e-01 7.26443648e-01
-8.16028118e-01 2.48202875e-01 -9.79305744e-01 5.43391645e-01
1.15814030e+00 -6.12764776e-01 -1.15230703e+00 3.02286625e-01
8.30535531e-01 -2.26258799e-01 -9.50654864e-01 8.38675648e-02
1.15716793e-01 -6.21358573e-01 8.68262589e-01 -1.24092722e+00
2.81752348e-01 3.10335755e-02 3.16251487e-01 -1.67999125e+00
-4.51357991e-01 -4.36738282e-01 1.98863983e-01 1.89267862e+00
1.04334641e+00 -3.60098988e-01 2.32202366e-01 6.85474813e-01
9.90854762e-03 -1.26053333e-01 -7.11213171e-01 -9.80045676e-01
1.29910409e-01 -2.23077908e-01 3.37536991e-01 1.69175708e+00
5.85386530e-03 8.06632936e-01 -3.74296516e-01 1.05785511e-01
3.06714982e-01 -4.75511223e-01 5.61413527e-01 -1.44075799e+00
-2.02881783e-01 -3.13504010e-01 -1.88522398e-01 -4.85291481e-01
4.36251670e-01 -1.16910803e+00 2.05809280e-01 -1.19604647e+00
3.34818482e-01 -5.57899177e-01 -6.18489563e-01 8.02149177e-01
-6.35751128e-01 4.45288559e-03 1.80764392e-01 9.50414166e-02
-7.69424736e-01 2.01684490e-01 7.34530985e-01 -8.16876739e-02
-2.67210633e-01 4.43295240e-02 -5.29474139e-01 6.26460373e-01
3.98595899e-01 -8.49970877e-01 1.55661255e-01 -3.51750225e-01
3.99498880e-01 -2.02720612e-01 2.40299162e-02 -8.10470819e-01
5.29752076e-01 -5.86638832e-03 4.42274213e-01 8.97249803e-02
-1.44541219e-01 -6.49675667e-01 -1.30503833e-01 1.08667269e-01
-6.06424272e-01 2.82788239e-02 2.27787673e-01 3.73359293e-01
-1.05753347e-01 -6.95891619e-01 7.49465406e-01 -4.73159730e-01
-1.06498313e+00 1.81383133e-01 -2.00853676e-01 8.46327901e-01
6.25359297e-01 8.03113654e-02 -2.27815866e-01 6.25763508e-03
-1.11745679e+00 -2.83227526e-02 2.72508949e-01 6.63304567e-01
3.68201770e-02 -1.20724463e+00 -8.83766651e-01 -8.01955163e-02
4.16672409e-01 -3.27591628e-01 2.92047054e-01 6.14345491e-01
4.78130057e-02 1.88056882e-02 -1.39383525e-01 -1.62400693e-01
-8.87862384e-01 7.88910806e-01 3.12437266e-01 -4.30489808e-01
-3.81447732e-01 1.18739676e+00 1.01370901e-01 -9.75333512e-01
3.20654660e-01 7.01783895e-02 -6.38055384e-01 5.60327888e-01
8.43902230e-01 4.19642091e-01 2.53274813e-02 -5.54439187e-01
-1.77298725e-01 3.37283611e-01 -3.18118304e-01 7.12360740e-02
1.44521272e+00 4.18457985e-02 4.83692400e-02 9.94513094e-01
1.10690868e+00 3.16505879e-01 -5.78182459e-01 -3.89295787e-01
7.34222174e-01 -3.91202904e-02 -2.91237772e-01 -1.05905509e+00
-5.73821247e-01 7.56905496e-01 4.53237712e-01 1.62927970e-01
5.38945019e-01 1.51969269e-01 5.38241923e-01 8.66005003e-01
3.55850518e-01 -1.09965658e+00 -3.35482001e-01 9.73445594e-01
4.76482689e-01 -1.43544102e+00 -4.19599324e-01 -1.16275229e-01
-7.24627793e-01 1.09308422e+00 7.97615230e-01 9.65407416e-02
5.90172768e-01 1.77369237e-01 2.67603219e-01 8.88905488e-03
-6.91298544e-01 -4.50928748e-01 5.14582157e-01 6.27623856e-01
8.72129798e-01 -1.11667076e-02 -3.04409772e-01 7.94673562e-01
-2.65033871e-01 -1.50870308e-01 3.70918930e-01 6.73090637e-01
-4.84627515e-01 -9.75565553e-01 -2.25403950e-01 4.15463507e-01
-6.98757946e-01 -6.07108593e-01 -3.20728630e-01 8.61990154e-01
2.10494027e-01 5.42280793e-01 -3.22202258e-02 -1.80497006e-01
6.40924394e-01 7.89982021e-01 8.40815678e-02 -1.18734074e+00
-1.09903610e+00 -4.94751036e-01 4.32637453e-01 -1.86888948e-01
-2.75466025e-01 -3.73916328e-01 -1.06941247e+00 -9.03153867e-02
-5.02048433e-01 4.45650131e-01 4.35252905e-01 1.12222004e+00
3.41268986e-01 3.36411059e-01 3.98923397e-01 -7.00888216e-01
-8.38086724e-01 -1.31907463e+00 -6.16537631e-01 8.01579535e-01
3.28267030e-02 -7.51224816e-01 -6.90663636e-01 2.13465437e-01]
|
[9.674999237060547, 9.3273286819458]
|
8eb324db-d02e-47f7-b3f4-421ff96e8656
|
g1020-a-benchmark-retinal-fundus-image
|
2006.09158
| null |
https://arxiv.org/abs/2006.09158v1
|
https://arxiv.org/pdf/2006.09158v1.pdf
|
G1020: A Benchmark Retinal Fundus Image Dataset for Computer-Aided Glaucoma Detection
|
Scarcity of large publicly available retinal fundus image datasets for automated glaucoma detection has been the bottleneck for successful application of artificial intelligence towards practical Computer-Aided Diagnosis (CAD). A few small datasets that are available for research community usually suffer from impractical image capturing conditions and stringent inclusion criteria. These shortcomings in already limited choice of existing datasets make it challenging to mature a CAD system so that it can perform in real-world environment. In this paper we present a large publicly available retinal fundus image dataset for glaucoma classification called G1020. The dataset is curated by conforming to standard practices in routine ophthalmology and it is expected to serve as standard benchmark dataset for glaucoma detection. This database consists of 1020 high resolution colour fundus images and provides ground truth annotations for glaucoma diagnosis, optic disc and optic cup segmentation, vertical cup-to-disc ratio, size of neuroretinal rim in inferior, superior, nasal and temporal quadrants, and bounding box location for optic disc. We also report baseline results by conducting extensive experiments for automated glaucoma diagnosis and segmentation of optic disc and optic cup.
|
['Sheraz Ahmed', 'Wolfgang Neumeier', 'Muhammad Naseer Bajwa', 'Muhammad Imran Malik', 'Gur Amrit Pal Singh', 'Andreas Dengel']
|
2020-05-28
| null | null | null | null |
['optic-cup-segmentation']
|
['medical']
|
[ 2.73248196e-01 9.95984674e-02 2.65610933e-01 -2.38261923e-01
-5.07588446e-01 -3.69830400e-01 1.61121234e-01 -2.09008873e-01
-4.93183553e-01 7.18881249e-01 2.24304408e-01 -5.53604662e-01
-3.32623720e-01 -4.06785607e-01 9.53930765e-02 -6.91545784e-01
-2.78387249e-01 5.14747977e-01 1.94252402e-01 2.11737886e-01
6.13650799e-01 3.84929657e-01 -1.98316824e+00 3.90201211e-01
9.56717551e-01 1.05549979e+00 2.46369854e-01 9.96911347e-01
5.59716344e-01 1.97770312e-01 -3.42719972e-01 -9.33392495e-02
8.19525540e-01 -5.47735274e-01 -1.03478289e+00 6.91050947e-01
9.77050483e-01 -4.73810554e-01 1.98377762e-02 9.94942784e-01
1.04735887e+00 -5.31155467e-01 7.15448737e-01 -5.37347078e-01
-5.11162460e-01 -3.55312020e-01 -7.63012052e-01 6.90840542e-01
-7.31005073e-02 5.28362513e-01 4.89073753e-01 -2.68911123e-01
7.06857681e-01 7.65281975e-01 4.89681810e-01 4.49139148e-01
-7.73725390e-01 -1.17757566e-01 -5.07243812e-01 -9.05293897e-02
-1.10146999e+00 -4.80151147e-01 -6.93117231e-02 -8.21136832e-01
9.34077680e-01 4.50333476e-01 1.13366604e+00 8.58668936e-04
1.58423632e-01 3.11979860e-01 1.82028937e+00 -6.36760712e-01
-2.29271315e-02 -1.73466966e-01 1.61148280e-01 6.14664435e-01
7.13358164e-01 3.84283572e-01 2.08653554e-01 -1.96949750e-01
9.99560237e-01 -5.60577691e-01 -4.01662856e-01 7.25202858e-02
-9.70681012e-01 5.89103639e-01 5.06479025e-01 -1.31308272e-01
-1.97912499e-01 -2.14572236e-01 2.90612131e-01 4.65050429e-01
3.33148062e-01 6.37079716e-01 -3.99439782e-01 1.88617185e-02
-6.92100823e-01 4.44215894e-01 1.39944479e-01 7.17821896e-01
1.48274496e-01 -5.56460559e-01 -2.85450190e-01 8.43803406e-01
5.44103324e-01 1.51621684e-01 7.34202623e-01 -9.39502716e-01
1.71893895e-01 1.05411160e+00 2.01026499e-01 -3.10282826e-01
-4.41656142e-01 -3.81772310e-01 -6.63811862e-01 5.59406579e-01
7.17034876e-01 -4.23801273e-01 -1.52175832e+00 3.15160245e-01
3.58706206e-01 1.85110018e-01 -2.62546897e-01 1.25652444e+00
8.08808148e-01 -3.40226293e-01 -2.21217439e-01 -3.24070245e-01
1.51323557e+00 -5.41312873e-01 -2.71192729e-01 -3.43894899e-01
1.11448395e+00 -1.15781808e+00 8.05408537e-01 5.00715494e-01
-1.04688549e+00 -3.00805092e-01 -5.35850644e-01 -3.19866210e-01
-4.78153564e-02 8.12934458e-01 1.00521195e+00 5.75815141e-01
-1.32277536e+00 1.50260746e-01 -5.80215216e-01 -4.62340325e-01
1.00537515e+00 7.12186813e-01 -6.05421960e-01 -1.40091538e-01
-3.23441178e-01 8.06497514e-01 3.73472005e-01 2.85256833e-01
-6.30544573e-02 -3.71870190e-01 -4.99315411e-01 -8.91331613e-01
-1.84999540e-01 -1.10691917e+00 1.23408651e+00 -5.14133573e-01
-1.00121725e+00 1.71561539e+00 -1.13957226e-01 -8.46396625e-01
5.43917775e-01 -7.14856982e-02 -2.96873510e-01 2.98297197e-01
-4.98261899e-02 7.02600718e-01 7.81793356e-01 -7.99280405e-01
-1.30995202e+00 -9.04575646e-01 -2.17407092e-01 1.43843725e-01
4.40140367e-01 4.39334750e-01 -4.64965105e-01 -2.64555216e-01
3.20072412e-01 -1.05075920e+00 -4.16512758e-01 1.78453907e-01
-4.88427192e-01 -2.15102538e-01 6.80708587e-01 -6.98600233e-01
1.28937924e+00 -1.78235233e+00 -4.73995477e-01 -1.48996145e-01
6.06241465e-01 8.73058975e-01 2.35614106e-01 -5.72376288e-02
-1.67450249e-01 2.09770277e-01 1.11493148e-01 -2.24001899e-01
-6.82086527e-01 -5.65744229e-02 2.59885758e-01 9.02702034e-01
2.93776095e-01 7.76745856e-01 -4.73806113e-01 -6.08466804e-01
2.89917529e-01 3.08673084e-01 -6.35143280e-01 -3.95124778e-03
-2.29432750e-02 5.50546885e-01 -3.41188639e-01 1.02056098e+00
6.62781835e-01 -2.99635142e-01 -3.89809459e-01 -8.49446729e-02
-3.32533509e-01 1.86656043e-02 -9.59257483e-01 1.14883590e+00
1.08596809e-01 8.45731795e-01 -2.92268485e-01 -4.86400843e-01
6.55802667e-01 1.87430277e-01 1.74299613e-01 -4.26670253e-01
5.17983675e-01 4.71195012e-01 7.61123121e-01 -8.81433129e-01
5.66844605e-02 -2.11622212e-02 6.94851160e-01 1.88996449e-01
-2.43236259e-01 -2.13831604e-01 4.51573521e-01 -3.92479092e-01
7.91897655e-01 -2.04353631e-01 6.29967630e-01 -2.47676182e-03
4.32245821e-01 2.33742580e-01 2.00692430e-01 3.80730420e-01
-6.75585151e-01 1.18278110e+00 6.40550137e-01 -7.99933851e-01
-1.30772340e+00 -5.15566289e-01 -9.34930027e-01 -1.29306316e-01
-2.24930540e-01 -1.21917568e-01 -5.72492838e-01 -3.21345180e-01
6.72242939e-02 -4.78858560e-01 -6.68944061e-01 3.85622859e-01
-2.41307393e-01 -1.20565796e+00 1.96203724e-01 1.13378219e-01
7.53053010e-01 -7.75381446e-01 -8.04298460e-01 -1.65362284e-01
3.56318265e-01 -8.18641126e-01 -9.48909596e-02 -7.92425811e-01
-1.27555776e+00 -1.76520562e+00 -1.02010882e+00 -1.17440009e+00
1.01608336e+00 1.22155152e-01 9.38044608e-01 3.80917490e-01
-1.37309408e+00 -2.77320921e-01 -1.77868873e-01 -9.03803647e-01
-4.73685898e-02 -4.71587211e-01 -1.80720672e-01 5.52764460e-02
6.63537979e-01 -1.16282851e-01 -1.40668786e+00 2.65292138e-01
-3.48625332e-01 -1.17262661e-01 9.77979422e-01 6.44410014e-01
7.35646665e-01 6.26442805e-02 2.41143644e-01 -8.56236875e-01
6.64675951e-01 -1.16829738e-01 -9.36781645e-01 -5.82968518e-02
-6.43967927e-01 -6.16444826e-01 -3.29812855e-01 1.77203074e-01
-4.31632966e-01 -1.33172667e-03 4.12898868e-01 -3.82927172e-02
-5.11504412e-01 4.58613425e-01 4.84002292e-01 -5.38386285e-01
9.08913732e-01 -1.37982816e-01 4.80436116e-01 -7.51493394e-01
-2.67191604e-02 1.55780745e+00 5.95502853e-01 1.76485345e-01
1.99250042e-01 7.28752434e-01 2.99774408e-01 -8.43789279e-01
-8.48481894e-01 -1.02150440e+00 -7.43768334e-01 3.00860442e-02
7.49242425e-01 -8.79977465e-01 -6.24596894e-01 1.01405251e+00
-6.57775879e-01 3.46375955e-03 -9.22669098e-02 6.89302385e-01
-2.44295940e-01 2.78993040e-01 -2.54038215e-01 -7.29230464e-01
-5.80471992e-01 -1.22423470e+00 1.28139496e+00 3.32191318e-01
1.63748652e-01 -6.98834717e-01 -9.52645391e-03 9.58814740e-01
3.49465132e-01 3.64437789e-01 7.74095893e-01 -3.78754400e-02
-5.33890188e-01 -6.30707681e-01 -7.74670303e-01 6.62564337e-01
2.56645828e-01 3.22159648e-01 -9.52452183e-01 -3.87693465e-01
-2.31753707e-01 -1.16185732e-01 9.88878608e-01 1.04630208e+00
9.66494679e-01 -1.05416648e-01 -3.53154123e-01 5.83855391e-01
1.62580252e+00 3.81750539e-02 1.22745240e+00 4.96441096e-01
3.30446810e-01 5.69770515e-01 7.48230338e-01 3.13182503e-01
1.89027056e-01 4.13049936e-01 3.63494128e-01 -2.22979411e-01
-6.59003794e-01 6.05911911e-01 -5.84594369e-01 3.44637126e-01
-7.51261890e-01 1.49660677e-01 -1.36492717e+00 1.05696189e+00
-1.24018240e+00 -4.27181751e-01 -6.55686319e-01 2.64059234e+00
9.47403193e-01 -1.27181634e-01 3.59250456e-01 1.23652317e-01
6.45122647e-01 -7.31962860e-01 -3.24844301e-01 1.54231921e-01
-1.37541324e-01 3.68295044e-01 6.22699320e-01 1.61588207e-01
-1.43769252e+00 8.25659513e-01 6.68363428e+00 2.48697668e-01
-1.06290042e+00 -1.45070776e-01 6.96474552e-01 -4.52086836e-01
7.56938159e-01 1.08326785e-01 -9.87417459e-01 4.04834896e-01
7.53199816e-01 2.08806232e-01 -5.91403507e-02 1.94273219e-01
4.21085864e-01 -5.95642209e-01 -6.33885980e-01 1.08179903e+00
-2.67751276e-01 -1.50294709e+00 -3.03365234e-02 6.37602627e-01
8.55555952e-01 4.61974531e-01 1.73834100e-01 -6.30481303e-01
-2.75835663e-01 -1.30376160e+00 -5.04338801e-01 6.88023746e-01
1.28490877e+00 -2.70614952e-01 1.19089806e+00 -3.24791312e-01
-3.89339894e-01 -1.61914349e-01 -4.74602938e-01 -2.40655616e-01
-2.44876325e-01 2.49977231e-01 -1.32209146e+00 4.93911235e-03
7.27006674e-01 9.34655905e-01 -1.01062524e+00 2.44028783e+00
2.85832912e-01 5.65034211e-01 -8.16428363e-02 4.12253499e-01
1.98000431e-01 -5.68612099e-01 6.01104438e-01 4.21707720e-01
2.45548218e-01 1.50209531e-01 -2.49920666e-01 3.58134449e-01
2.55062997e-01 3.33939880e-01 -5.92304885e-01 -1.90959379e-01
2.75826842e-01 9.78229225e-01 -5.57053804e-01 3.31772044e-02
-6.13692105e-01 1.99156404e-01 -2.72090554e-01 2.89452791e-01
-1.40454061e-02 -2.49876767e-01 8.14706326e-01 7.47014344e-01
-2.97797658e-02 9.68605056e-02 -3.98825645e-01 -7.66421914e-01
1.66931413e-02 -9.40933585e-01 3.84044647e-01 -8.61532509e-01
-9.67923820e-01 6.41680121e-01 -4.33629215e-01 -1.70261681e+00
2.09316894e-01 -1.07054734e+00 -4.80856150e-01 1.25976276e+00
-1.60718369e+00 -1.41921520e+00 -4.70344484e-01 4.01808679e-01
1.26199663e-01 -7.81130493e-01 7.50598490e-01 2.51993626e-01
-7.88044870e-01 2.14781836e-01 -1.22248614e-02 2.33975947e-01
8.26925337e-01 -1.33840549e+00 2.60880023e-01 7.17958212e-01
-3.60094488e-01 7.65180826e-01 4.29810971e-01 -7.96939611e-01
-9.82712507e-01 -1.23429465e+00 7.02644706e-01 -5.14535248e-01
6.05609775e-01 5.82899988e-01 -5.31375766e-01 5.58824420e-01
-9.27738994e-02 4.79206324e-01 6.94894075e-01 -2.91568483e-03
3.16320777e-01 7.91124254e-02 -1.09028292e+00 4.80590820e-01
7.14429080e-01 -1.75720960e-01 -4.06645894e-01 8.94738257e-01
1.59750372e-01 -8.22917402e-01 -9.68220413e-01 4.42618191e-01
3.56663018e-01 -1.09836733e+00 6.41989529e-01 -1.08402026e+00
2.83063143e-01 -4.97768044e-01 3.98790300e-01 -6.74681425e-01
1.82109371e-01 -9.60851908e-01 2.30809599e-01 5.74797392e-01
3.30275089e-01 -9.77608860e-01 8.58129740e-01 4.30379093e-01
-3.69478613e-01 -9.98497903e-01 -8.83882046e-01 -1.74791723e-01
2.71606706e-02 -3.44434120e-02 2.11369544e-01 3.83152634e-01
-5.57527006e-01 -2.35859856e-01 -3.89009379e-02 3.41235489e-01
6.39726341e-01 3.10649097e-01 8.47350657e-01 -1.64858878e+00
4.88036349e-02 -2.91073442e-01 -1.46224999e+00 -3.86924744e-01
-8.39469075e-01 -5.18732011e-01 -6.27999604e-01 -1.85398114e+00
9.70232636e-02 -5.95049798e-01 3.24808173e-02 2.99383879e-01
1.53834734e-03 9.85751450e-01 -4.89134729e-01 5.46246648e-01
-4.11801487e-02 -4.30524468e-01 2.06066728e+00 2.62990355e-01
-4.31065112e-01 4.45016116e-01 -7.70093262e-01 8.17041874e-01
7.73248613e-01 -7.10660517e-02 -4.98842776e-01 -3.36664259e-01
1.30661502e-01 -2.40757853e-01 7.79274821e-01 -1.00974774e+00
1.30121797e-01 2.19866544e-01 2.61733949e-01 -5.08546352e-01
9.95242596e-02 -2.04759002e-01 -2.16173351e-01 2.96896160e-01
1.42253160e-01 -6.33485079e-01 7.94568062e-02 5.03305376e-01
-4.27518696e-01 7.29452027e-03 1.06254888e+00 -4.06114608e-01
-5.27853966e-01 6.11112654e-01 -2.53265984e-02 2.41354555e-01
1.29190040e+00 -8.32002997e-01 -8.47524285e-01 3.56191427e-01
-9.65584695e-01 8.35986510e-02 6.08549535e-01 1.82636812e-01
5.55382848e-01 -6.80373847e-01 -1.24583578e+00 4.25346792e-01
4.62820262e-01 3.25129151e-01 8.14560279e-02 1.51136863e+00
-1.25139511e+00 7.75645435e-01 -3.73673558e-01 -7.20962822e-01
-1.84672391e+00 -8.31924751e-02 9.60667133e-01 3.60574692e-01
-9.47376013e-01 1.11015975e+00 -2.70563334e-01 4.29252476e-01
9.91178080e-02 -8.32395971e-01 -5.22720039e-01 -7.88109750e-02
8.89713943e-01 2.25147024e-01 4.03054982e-01 -7.21406937e-01
1.24034323e-01 8.73534262e-01 -5.53171754e-01 5.60750484e-01
1.15591300e+00 -3.69833618e-01 -5.90031087e-01 -3.99769843e-02
8.09761524e-01 -3.39374810e-01 -8.28909397e-01 -1.11401767e-01
-1.86581239e-01 -9.55712795e-01 7.08944023e-01 -1.04924691e+00
-9.80853260e-01 8.20219815e-01 1.33467948e+00 1.69554219e-01
1.25456405e+00 -3.20863053e-02 6.69028699e-01 4.85780500e-02
3.72294873e-01 -8.06815684e-01 -6.11598253e-01 1.42161893e-02
8.87768626e-01 -1.69759309e+00 2.65566766e-01 -6.96640849e-01
-6.36375725e-01 1.03017676e+00 5.49892008e-01 -3.16815972e-01
6.67894721e-01 -2.98246652e-01 6.66934252e-01 -6.39768541e-01
-4.88570094e-01 -7.70522535e-01 1.01919246e+00 8.37189138e-01
5.97565114e-01 2.79975608e-02 -6.42874658e-01 -3.40616740e-02
-3.37522864e-01 4.16637361e-01 7.88030148e-01 8.04800451e-01
-7.43494391e-01 -1.02117336e+00 -7.53216222e-02 1.38123178e+00
-8.04761648e-01 -3.02511722e-01 -4.85552788e-01 8.99580896e-01
5.62755942e-01 8.53794515e-01 1.73789784e-01 3.11357230e-01
3.45356129e-02 -3.57122272e-01 6.33508325e-01 -1.00250173e+00
-2.41875291e-01 4.28252846e-01 3.79735678e-01 -3.93271655e-01
-5.43689787e-01 -6.53903842e-01 -8.67891371e-01 2.63344049e-01
-4.63013202e-01 -2.25360900e-01 3.86375785e-01 8.77034903e-01
6.49898946e-01 2.61221547e-02 9.97826457e-02 -3.42122704e-01
-1.95889100e-01 -1.37177670e+00 -1.06286466e+00 8.91885683e-02
6.75437212e-01 -5.62831998e-01 -3.10883492e-01 3.65770131e-01]
|
[15.816308975219727, -4.006500244140625]
|
9d84bc73-92f0-4ac7-bee2-93d26dd9f6d0
|
relightable-neural-human-assets-from-multi
|
2212.07648
| null |
https://arxiv.org/abs/2212.07648v3
|
https://arxiv.org/pdf/2212.07648v3.pdf
|
Relightable Neural Human Assets from Multi-view Gradient Illuminations
|
Human modeling and relighting are two fundamental problems in computer vision and graphics, where high-quality datasets can largely facilitate related research. However, most existing human datasets only provide multi-view human images captured under the same illumination. Although valuable for modeling tasks, they are not readily used in relighting problems. To promote research in both fields, in this paper, we present UltraStage, a new 3D human dataset that contains more than 2,000 high-quality human assets captured under both multi-view and multi-illumination settings. Specifically, for each example, we provide 32 surrounding views illuminated with one white light and two gradient illuminations. In addition to regular multi-view images, gradient illuminations help recover detailed surface normal and spatially-varying material maps, enabling various relighting applications. Inspired by recent advances in neural representation, we further interpret each example into a neural human asset which allows novel view synthesis under arbitrary lighting conditions. We show our neural human assets can achieve extremely high capture performance and are capable of representing fine details such as facial wrinkles and cloth folds. We also validate UltraStage in single image relighting tasks, training neural networks with virtual relighted data from neural assets and demonstrating realistic rendering improvements over prior arts. UltraStage will be publicly available to the community to stimulate significant future developments in various human modeling and rendering tasks. The dataset is available at https://miaoing.github.io/RNHA.
|
['Jingyi Yu', 'Lan Xu', 'Wenzheng Chen', 'Kuixiang Shao', 'Qixuan Zhang', 'Teng Xu', 'Di wu', 'Kai He', 'Taotao Zhou']
|
2022-12-15
| null |
http://openaccess.thecvf.com//content/CVPR2023/html/Zhou_Relightable_Neural_Human_Assets_From_Multi-View_Gradient_Illuminations_CVPR_2023_paper.html
|
http://openaccess.thecvf.com//content/CVPR2023/papers/Zhou_Relightable_Neural_Human_Assets_From_Multi-View_Gradient_Illuminations_CVPR_2023_paper.pdf
|
cvpr-2023-1
|
['image-relighting']
|
['computer-vision']
|
[ 2.25166470e-01 -1.08394511e-01 1.91984624e-01 -2.33664095e-01
-2.61227190e-01 -3.72969389e-01 5.11004984e-01 -6.45135283e-01
2.30130762e-01 5.98696947e-01 2.24161863e-01 1.34524807e-01
4.60072398e-01 -9.54617083e-01 -8.76204491e-01 -5.42632639e-01
3.82787049e-01 2.66087502e-01 -1.86769024e-01 -4.37929809e-01
-2.02293068e-01 7.46705770e-01 -1.71014416e+00 3.88919652e-01
5.30192971e-01 9.62582409e-01 1.74286738e-02 6.85526907e-01
3.03556114e-01 4.94754434e-01 -3.79881591e-01 -6.61867559e-01
3.72454464e-01 -2.44698673e-01 -3.46936405e-01 1.81396022e-01
1.08316195e+00 -7.79190123e-01 -5.65118074e-01 5.68528056e-01
8.10860693e-01 2.04013124e-01 4.71246004e-01 -1.17855346e+00
-1.07529759e+00 -5.93023635e-02 -8.25126350e-01 -1.97932512e-01
6.35226905e-01 2.04514354e-01 7.47821033e-01 -1.13790393e+00
8.18480730e-01 1.50418222e+00 7.28204310e-01 8.45724940e-01
-1.21499336e+00 -7.62821257e-01 -1.28634959e-01 -7.59195536e-02
-9.95979548e-01 -6.04247451e-01 1.04746985e+00 -2.18340084e-01
7.32035875e-01 4.57095385e-01 1.13595974e+00 1.44917083e+00
3.44789207e-01 7.45836318e-01 1.19631290e+00 -7.49959499e-02
-6.31715879e-02 -1.14243381e-01 -3.75465244e-01 1.04505622e+00
6.88632578e-02 4.49462175e-01 -7.48650670e-01 -9.05245394e-02
1.56587040e+00 2.66970068e-01 -5.95032036e-01 -4.11936522e-01
-1.30208349e+00 6.11488342e-01 4.88797814e-01 -1.12034090e-01
-2.18263179e-01 3.96077693e-01 8.20844769e-02 -3.55156474e-02
7.21882939e-01 2.42227465e-01 -9.44544189e-03 2.01293081e-01
-6.52708709e-01 3.61972153e-01 4.44993556e-01 1.01004016e+00
7.67220795e-01 5.20387113e-01 1.11168064e-01 1.04608560e+00
4.61569391e-02 7.90746152e-01 -1.79611631e-02 -1.50601983e+00
1.99197859e-01 3.39456826e-01 1.15628473e-01 -1.29617715e+00
-4.26188886e-01 -2.54779726e-01 -1.25630617e+00 4.46369022e-01
2.31449395e-01 -3.11687812e-02 -8.53922367e-01 1.43340373e+00
5.05250275e-01 1.97560892e-01 -3.51404667e-01 1.04913008e+00
1.29871285e+00 7.07375944e-01 -3.46629024e-01 1.27100408e-01
1.35861886e+00 -1.08848846e+00 -7.49330103e-01 -5.37684187e-02
7.77456388e-02 -8.85805070e-01 1.32643664e+00 8.31153095e-01
-1.49468815e+00 -6.43967211e-01 -8.13256800e-01 -6.67276859e-01
-1.44777641e-01 -7.11274520e-02 8.11886609e-01 3.83560300e-01
-1.19075155e+00 6.39157414e-01 -6.70352936e-01 -2.87885010e-01
6.22277737e-01 -1.08363606e-01 -4.85622555e-01 -3.93574119e-01
-1.02473354e+00 5.05124688e-01 -4.66776252e-01 2.46044606e-01
-8.97079706e-01 -9.57818806e-01 -9.87946451e-01 -2.95855045e-01
2.74131615e-02 -1.18691897e+00 9.45211053e-01 -7.07273424e-01
-1.47101367e+00 1.17676616e+00 -1.44470677e-01 -1.77991558e-02
7.81405389e-01 -2.81190962e-01 -2.85548538e-01 3.86049300e-01
-2.23495230e-01 7.54871607e-01 9.73798752e-01 -1.85300553e+00
-6.21739775e-02 -4.95438248e-01 2.91056260e-02 1.90373421e-01
-2.44174510e-01 -1.10729091e-01 -5.93149185e-01 -1.00090170e+00
-6.37299940e-02 -8.28740656e-01 5.37575670e-02 6.72645271e-01
-5.76260149e-01 3.26700419e-01 9.52168941e-01 -8.89345288e-01
7.12548077e-01 -1.89325702e+00 2.60101259e-01 -2.04331636e-01
5.07772326e-01 1.02060758e-01 -1.26551762e-01 2.73791194e-01
-1.12138867e-01 8.36125612e-02 -1.10960871e-01 -8.36985469e-01
-6.51604012e-02 2.30320290e-01 -3.70575964e-01 4.82067615e-01
-1.33524925e-01 1.17893219e+00 -7.08046913e-01 -4.19340879e-01
5.35912454e-01 1.18784523e+00 -5.44400334e-01 3.20171893e-01
-3.86881530e-02 7.81361341e-01 -2.21636847e-01 9.79395509e-01
8.95959139e-01 -2.75399417e-01 -3.05999190e-01 -6.39152586e-01
1.39973924e-01 -4.73389447e-01 -7.84921527e-01 1.81482279e+00
-6.83488905e-01 7.77033269e-01 2.29879647e-01 -4.90074843e-01
8.39094877e-01 1.46777868e-01 6.01298034e-01 -7.74534881e-01
1.55181602e-01 -2.11409613e-01 -7.21520782e-01 -4.07670557e-01
6.81892812e-01 -4.17012513e-01 2.59859025e-01 3.92695546e-01
-3.03939044e-01 -5.41478395e-01 -2.91200250e-01 2.26404652e-01
5.86438954e-01 3.54237229e-01 -4.73124720e-02 8.75659063e-02
-9.41884741e-02 -4.02931094e-01 3.93743336e-01 2.63989031e-01
1.24748489e-02 1.34413481e+00 1.14672609e-01 -9.13928807e-01
-1.32647347e+00 -1.35040426e+00 -2.65591204e-01 8.99276674e-01
2.52387524e-01 -1.80580273e-01 -6.57651782e-01 -1.46136045e-01
1.01723008e-01 4.70959514e-01 -9.46174562e-01 5.74430563e-02
-7.06804454e-01 -4.61525232e-01 5.08462131e-01 5.24093986e-01
7.10295022e-01 -1.04506135e+00 -6.38621449e-01 -1.47501141e-01
-3.29579681e-01 -1.12519681e+00 -5.90295792e-01 -4.19647485e-01
-7.27083564e-01 -9.86515284e-01 -1.28972471e+00 -5.53736269e-01
6.46120429e-01 5.85925937e-01 1.60112524e+00 4.43469346e-01
-6.67935312e-01 6.29540145e-01 -2.71842666e-02 -1.81643307e-01
-2.61569589e-01 -5.20531595e-01 4.92828190e-02 -1.35964125e-01
-4.93754715e-01 -8.47243309e-01 -8.99073482e-01 4.80774432e-01
-9.91841316e-01 7.76460707e-01 7.72884339e-02 7.40457118e-01
6.79390907e-01 -2.98237532e-01 1.18333340e-01 -8.62975657e-01
1.86315000e-01 -2.50113904e-01 -3.20137650e-01 1.63091525e-01
-3.35798785e-03 -4.60250586e-01 6.12427294e-01 -4.00250435e-01
-1.22108364e+00 -3.13295215e-01 -2.22275138e-01 -9.48295951e-01
-1.93183497e-01 -1.15673512e-01 -2.16529891e-01 -2.87445694e-01
5.80989182e-01 1.35898903e-01 -7.14195371e-02 -4.89098191e-01
3.55198413e-01 2.00262487e-01 5.46039701e-01 -7.74133682e-01
7.60464430e-01 8.49887192e-01 7.76890144e-02 -1.00689411e+00
-7.58990109e-01 2.46895552e-01 -4.53352094e-01 -5.71910262e-01
8.31562579e-01 -1.03845167e+00 -8.99111450e-01 7.97807276e-01
-1.02963150e+00 -7.43167520e-01 -2.73412496e-01 -1.63347960e-01
-7.35241830e-01 3.57560009e-01 -8.39244545e-01 -5.65432131e-01
-4.48400080e-01 -1.18245053e+00 1.57129943e+00 2.70537466e-01
2.37498209e-02 -1.11513066e+00 -1.15820840e-01 8.29147160e-01
2.58408457e-01 8.74370337e-01 7.97502220e-01 5.50556123e-01
-5.38742065e-01 1.54726580e-01 -2.87119269e-01 2.80477822e-01
2.63058096e-01 2.16606781e-01 -1.18300498e+00 -3.63467008e-01
-2.39614636e-01 -5.89027524e-01 6.78053141e-01 4.13429290e-01
1.52017200e+00 -1.45326972e-01 -2.45459467e-01 1.16723609e+00
1.31278455e+00 -6.48711473e-02 8.32488835e-01 -4.29201908e-02
1.35598648e+00 6.63973212e-01 1.78095326e-01 6.26080692e-01
4.70739931e-01 8.59279871e-01 5.11858761e-01 -6.45552337e-01
-6.30197465e-01 -4.47722048e-01 3.05793975e-02 7.01786399e-01
-5.56062639e-01 -3.26075763e-01 -8.10556233e-01 2.95895845e-01
-1.34587240e+00 -9.29207563e-01 -1.33495912e-01 2.04868698e+00
7.69554198e-01 -3.18709314e-01 6.97565079e-02 -1.73546582e-01
5.18347800e-01 3.20354283e-01 -8.23021412e-01 -2.67393172e-01
-3.98951590e-01 2.66413212e-01 7.09331483e-02 4.70932782e-01
-6.90915346e-01 8.50367785e-01 6.32561588e+00 6.66130900e-01
-1.08857727e+00 -2.15362143e-02 9.92203712e-01 -3.91981810e-01
-8.18848968e-01 -4.46575046e-01 -5.57709515e-01 2.55050480e-01
3.81234884e-01 1.18588932e-01 6.29946411e-01 6.18290007e-01
2.53823847e-01 -2.76364218e-02 -9.57414687e-01 1.37821364e+00
3.15765440e-01 -1.39131641e+00 5.13128877e-01 2.50605136e-01
1.01704025e+00 -3.00397903e-01 5.76312423e-01 -6.17679618e-02
2.81660706e-01 -1.25789678e+00 6.19771361e-01 7.93934464e-01
1.32409990e+00 -7.78158247e-01 1.02061324e-01 8.85316730e-02
-1.19250596e+00 2.88021326e-01 -4.71476465e-01 1.99956179e-01
4.03236419e-01 5.90851486e-01 -1.60289511e-01 5.77723920e-01
8.97805631e-01 1.10955548e+00 -5.04112005e-01 5.59132755e-01
5.78118376e-02 1.73943877e-01 -1.37629107e-01 4.09801304e-01
-2.02405512e-01 -2.93522716e-01 3.77324432e-01 8.69247079e-01
4.17548865e-01 3.58468086e-01 2.87297252e-03 9.73539710e-01
-3.05214792e-01 -1.49603695e-01 -9.21900630e-01 3.24379772e-01
1.51435181e-01 1.23911381e+00 -4.83342409e-01 -3.36422086e-01
-2.68720657e-01 1.28753388e+00 3.16532850e-01 6.36960268e-01
-9.89264250e-01 -1.03003988e-02 7.37871647e-01 4.44178820e-01
-1.36199161e-01 -3.56464386e-01 -3.73006701e-01 -1.51211655e+00
-2.22958878e-01 -8.90444577e-01 -1.41917333e-01 -1.46995354e+00
-1.23878622e+00 6.80639684e-01 1.07061289e-01 -1.23842227e+00
3.38711366e-02 -6.37222767e-01 -3.70018542e-01 7.93081820e-01
-1.46186686e+00 -1.67412829e+00 -8.99505198e-01 8.09743226e-01
6.61015332e-01 1.24013744e-01 7.79160500e-01 3.18158209e-01
-6.21365547e-01 5.61796665e-01 -8.51467624e-02 -3.92315723e-02
9.25689995e-01 -1.00524449e+00 6.77969337e-01 4.39738601e-01
9.66070443e-02 5.51110029e-01 6.81361020e-01 -6.83107913e-01
-1.60959053e+00 -1.08484292e+00 1.37488037e-01 -4.72833425e-01
1.66839406e-01 -4.66776252e-01 -9.85812187e-01 8.06633711e-01
3.68975341e-01 3.15543532e-01 5.25656581e-01 -2.85269469e-01
-4.10544157e-01 4.61276956e-02 -1.22468853e+00 8.31230462e-01
1.51620674e+00 -3.30589831e-01 -1.64459199e-02 2.80546516e-01
6.27251029e-01 -8.05073023e-01 -1.03015399e+00 3.95938694e-01
9.97355819e-01 -1.46127760e+00 1.34548938e+00 -3.84672523e-01
9.18324769e-01 5.51629514e-02 -2.54553020e-01 -1.41247594e+00
-2.51337886e-01 -6.28480732e-01 -3.51609975e-01 9.41418886e-01
-2.04137415e-01 -6.22233868e-01 8.45459282e-01 6.99386179e-01
-1.38883173e-01 -1.08712304e+00 -5.92821181e-01 -4.28548545e-01
1.68024093e-01 -1.65607870e-01 6.52273655e-01 9.63426173e-01
-7.65926421e-01 -7.54228234e-02 -9.46920872e-01 -1.22826874e-01
1.04572308e+00 2.66591251e-01 1.09776175e+00 -1.08093512e+00
-2.20893368e-01 -1.58628076e-01 5.56389219e-04 -1.06843066e+00
1.92856595e-01 -6.28741503e-01 -2.94180721e-01 -1.59572613e+00
1.32478625e-01 -4.13549989e-01 3.47623795e-01 5.24401605e-01
1.05180657e-02 9.98813093e-01 2.52888501e-01 8.02133456e-02
-1.89716090e-02 8.17756772e-01 2.09057403e+00 -1.25576500e-02
1.24118894e-01 -2.73207605e-01 -5.98269165e-01 9.35624838e-01
6.92226231e-01 1.87424287e-01 -3.90106916e-01 -8.71512234e-01
3.02985042e-01 2.05007032e-01 8.41385067e-01 -9.37355697e-01
-1.93204522e-01 -2.46226430e-01 9.34529364e-01 -5.23596108e-01
1.00046599e+00 -6.97809041e-01 7.01349854e-01 2.05068782e-01
-6.99691623e-02 1.61627814e-01 2.66371250e-01 4.50734824e-01
1.41216710e-01 4.39681739e-01 1.04563570e+00 -4.37073678e-01
-4.76557851e-01 6.62999153e-01 1.56223282e-01 1.73389129e-02
7.92854726e-01 -4.65931058e-01 -4.95909989e-01 -6.37971342e-01
-7.85381615e-01 -7.83687755e-02 1.08132112e+00 4.01352823e-01
1.15839660e+00 -1.61733329e+00 -7.13222802e-01 4.33564931e-01
-5.18231187e-03 1.12834945e-01 7.70237446e-01 3.23059648e-01
-7.91166484e-01 -1.81772158e-01 -5.16079187e-01 -5.44320822e-01
-1.33173680e+00 3.91154170e-01 3.68998200e-01 1.15086697e-01
-1.03218460e+00 7.73754895e-01 7.48880982e-01 -3.47961396e-01
6.14164099e-02 -1.32901639e-01 4.52407487e-02 -2.17392728e-01
6.71596944e-01 4.43158507e-01 -1.39541849e-01 -9.06474948e-01
1.31618559e-01 1.08488619e+00 2.11657777e-01 1.82974711e-02
1.42187142e+00 -1.90663487e-01 8.76640528e-02 5.00293672e-01
1.30435419e+00 3.60521255e-03 -1.59608114e+00 -6.75088912e-02
-1.16007340e+00 -8.81323755e-01 -4.24653012e-03 -6.37600601e-01
-1.59582090e+00 1.18409407e+00 5.39754272e-01 -6.81157857e-02
1.28061938e+00 -1.25640839e-01 1.31511688e+00 4.59558964e-02
6.89996004e-01 -5.74224353e-01 5.33197522e-01 1.36547878e-01
1.40905678e+00 -1.12015867e+00 1.32792935e-01 -5.93724787e-01
-6.30901098e-01 9.95102108e-01 7.46540785e-01 -2.09183618e-01
4.34125602e-01 4.08799350e-01 2.39361152e-02 -3.28361303e-01
-5.93377471e-01 2.70942748e-01 3.70843053e-01 8.19222510e-01
4.17116791e-01 1.10853638e-03 3.76777083e-01 1.66937768e-01
-4.56647694e-01 -1.09854147e-01 5.58357775e-01 4.97478873e-01
-2.78275721e-02 -7.28282630e-01 -5.59552431e-01 3.90306383e-01
-2.17121437e-01 -2.36211363e-02 -2.15021342e-01 7.27133334e-01
7.87040591e-02 6.35304928e-01 5.30435778e-02 -3.41564596e-01
5.21120310e-01 -5.00427306e-01 8.58688176e-01 -4.50559586e-01
-4.67048883e-01 1.14562199e-01 -2.11473601e-03 -8.16209435e-01
-4.06638473e-01 -3.70393008e-01 -9.21718538e-01 -8.03170204e-01
1.13093860e-01 -5.48659384e-01 3.75212133e-01 3.64360034e-01
2.56664902e-01 5.26531398e-01 5.41925430e-01 -1.62107229e+00
1.60747468e-01 -5.49535751e-01 -8.04314196e-01 5.10581851e-01
4.76554841e-01 -9.32032943e-01 -2.87609816e-01 4.16879803e-01]
|
[12.34485912322998, -0.6283774375915527]
|
0b421155-75e9-4f19-b970-5f084dfdc49f
|
ssh-a-self-supervised-framework-for-image
|
2108.06805
| null |
https://arxiv.org/abs/2108.06805v2
|
https://arxiv.org/pdf/2108.06805v2.pdf
|
SSH: A Self-Supervised Framework for Image Harmonization
|
Image harmonization aims to improve the quality of image compositing by matching the "appearance" (\eg, color tone, brightness and contrast) between foreground and background images. However, collecting large-scale annotated datasets for this task requires complex professional retouching. Instead, we propose a novel Self-Supervised Harmonization framework (SSH) that can be trained using just "free" natural images without being edited. We reformulate the image harmonization problem from a representation fusion perspective, which separately processes the foreground and background examples, to address the background occlusion issue. This framework design allows for a dual data augmentation method, where diverse [foreground, background, pseudo GT] triplets can be generated by cropping an image with perturbations using 3D color lookup tables (LUTs). In addition, we build a real-world harmonization dataset as carefully created by expert users, for evaluation and benchmarking purposes. Our results show that the proposed self-supervised method outperforms previous state-of-the-art methods in terms of reference metrics, visual quality, and subject user study. Code and dataset are available at \url{https://github.com/VITA-Group/SSHarmonization}.
|
['Zhangyang Wang', 'Sarah Kong', 'Sohrab Amirghodsi', 'Simon Chen', 'Kalyan Sunkavalli', 'Zhe Lin', 'Yilin Wang', 'Jianming Zhang', 'He Zhang', 'Yifan Jiang']
|
2021-08-15
| null |
http://openaccess.thecvf.com//content/ICCV2021/html/Jiang_SSH_A_Self-Supervised_Framework_for_Image_Harmonization_ICCV_2021_paper.html
|
http://openaccess.thecvf.com//content/ICCV2021/papers/Jiang_SSH_A_Self-Supervised_Framework_for_Image_Harmonization_ICCV_2021_paper.pdf
|
iccv-2021-1
|
['image-harmonization']
|
['computer-vision']
|
[ 3.77088338e-01 -1.55270398e-01 8.54207352e-02 -2.74887234e-01
-9.23062503e-01 -5.80780327e-01 4.89703089e-01 -3.69699299e-02
-4.29149009e-02 5.89979529e-01 -3.04686222e-02 -2.51392107e-02
3.58951896e-01 -6.08429432e-01 -7.29305625e-01 -8.08807611e-01
5.08672237e-01 6.60425127e-02 8.31299126e-02 -3.25303912e-01
1.19039163e-01 2.63155580e-01 -1.76736319e+00 2.80712306e-01
1.25057864e+00 9.19192135e-01 8.07652026e-02 4.69412267e-01
-7.80113563e-02 5.55803061e-01 -7.37488747e-01 -7.08421230e-01
6.83850884e-01 -8.26368988e-01 -4.33538646e-01 3.81270021e-01
8.85737062e-01 -1.44467000e-02 -1.20349087e-01 1.32992423e+00
8.94097328e-01 1.95360199e-01 2.49108866e-01 -1.43928134e+00
-9.31362808e-01 2.26275265e-01 -9.00496423e-01 -1.70149416e-01
1.96042225e-01 5.13344109e-01 8.52429986e-01 -7.64994442e-01
6.44164562e-01 9.79572952e-01 5.14907300e-01 4.05810952e-01
-1.66800869e+00 -1.02661288e+00 -1.22018993e-01 2.54703641e-01
-1.43750668e+00 -5.85257530e-01 1.15539014e+00 -3.57371837e-01
2.08873972e-01 6.60446823e-01 7.62608886e-01 1.00439215e+00
-5.93617521e-02 7.18616426e-01 1.66039300e+00 -5.30798852e-01
6.73928708e-02 3.33204508e-01 -3.03812593e-01 4.51412857e-01
1.03846170e-01 -4.09100465e-02 -5.38068175e-01 9.01763365e-02
6.69581175e-01 -4.71977234e-01 -4.42173958e-01 -5.95995069e-01
-1.20632565e+00 4.78296250e-01 2.92776138e-01 2.17105865e-01
-4.14834395e-02 -2.28310935e-02 2.71643430e-01 1.70328036e-01
4.94042486e-01 6.13016009e-01 -1.50527641e-01 1.35610417e-01
-9.39399302e-01 6.24143146e-02 3.81434113e-01 9.89098966e-01
8.88531983e-01 2.45466650e-01 -5.37364602e-01 1.13450360e+00
-3.80079187e-02 5.99606454e-01 4.04431522e-01 -1.02685809e+00
1.90935463e-01 6.32320404e-01 2.24397749e-01 -1.08857250e+00
-2.11157039e-01 -5.47736704e-01 -9.80468810e-01 5.13121068e-01
2.30455905e-01 3.15938815e-02 -9.00594175e-01 1.83275187e+00
6.41698480e-01 2.14482993e-01 -3.48223820e-02 9.35739279e-01
1.01900864e+00 6.35948300e-01 5.03393728e-03 -3.98241967e-01
1.25921559e+00 -1.16800594e+00 -1.04114747e+00 -1.19586084e-02
-6.92554340e-02 -1.37093806e+00 1.56553555e+00 5.30237556e-01
-1.32104790e+00 -1.04202163e+00 -1.10511613e+00 -1.69762850e-01
-3.17458898e-01 3.31588805e-01 2.99045622e-01 7.35072970e-01
-1.15326822e+00 3.49786639e-01 -3.74411672e-01 -1.65010959e-01
2.55235136e-01 8.40041414e-02 -1.62565112e-01 2.52248764e-01
-1.05323732e+00 7.90314138e-01 3.07379186e-01 5.34287505e-02
-5.86459219e-01 -7.54648209e-01 -8.91196132e-01 -3.07541519e-01
3.22044402e-01 -6.61242843e-01 9.70838606e-01 -1.34349394e+00
-1.74536324e+00 1.30998087e+00 2.49107778e-02 -8.48294701e-03
6.62588179e-01 -1.42576784e-01 -5.62554002e-01 -3.25752236e-02
6.97515905e-02 6.62537456e-01 1.20259070e+00 -1.94173717e+00
-5.29536366e-01 -1.78864226e-01 -2.38061741e-01 3.47348243e-01
-2.81668901e-01 -6.44027963e-02 -1.07872379e+00 -1.09007418e+00
-9.26125273e-02 -9.22572494e-01 -9.20111593e-03 -3.54412198e-02
-7.16634095e-01 4.56617236e-01 8.85635674e-01 -9.38554883e-01
1.20301569e+00 -2.34491849e+00 9.25143063e-02 2.97154427e-01
1.03590205e-01 4.07939672e-01 -2.86413312e-01 -2.04998422e-02
-3.38658571e-01 -8.10630247e-02 -4.08697158e-01 -6.32037342e-01
1.19405061e-01 -6.01993278e-02 -1.51334673e-01 4.61549312e-01
5.61840720e-02 9.12191749e-01 -7.24615276e-01 -7.88435996e-01
5.30896425e-01 5.09814799e-01 -3.44313532e-01 5.03182888e-01
-4.33693752e-02 8.40597391e-01 9.37206149e-02 7.04230666e-01
9.82078612e-01 -1.58225402e-01 7.36135915e-02 -7.50337601e-01
-1.38793528e-01 -3.83884043e-01 -1.45635450e+00 1.88442564e+00
-4.89662558e-01 6.53793991e-01 -1.51638836e-02 -5.85006297e-01
1.12076807e+00 1.10434815e-02 5.74265957e-01 -1.17129695e+00
2.79048562e-01 2.70643942e-02 -1.87027156e-01 -2.23227933e-01
5.05058646e-01 2.15880290e-01 -4.33758534e-02 3.61387312e-01
-1.72776982e-01 -4.91926700e-01 3.38107347e-01 -7.63279647e-02
4.80346382e-01 3.84326786e-01 1.55215561e-01 -1.15336537e-01
6.06814265e-01 9.49252322e-02 6.12829149e-01 4.78265464e-01
-2.20196337e-01 1.06564438e+00 3.49153638e-01 -1.82172239e-01
-1.33342731e+00 -1.20286083e+00 -1.92199662e-01 9.91012156e-01
6.41383827e-01 -3.78483236e-01 -1.06096303e+00 -4.26333845e-02
-2.16349781e-01 5.58088362e-01 -6.43106639e-01 -6.32916242e-02
-4.85862643e-01 -8.43656719e-01 3.60434830e-01 1.27528533e-02
9.63137686e-01 -1.06332636e+00 -5.46471536e-01 -4.75935973e-02
-4.55261827e-01 -1.10303569e+00 -6.77203834e-01 -2.11293802e-01
-4.18865174e-01 -1.02817500e+00 -7.38668799e-01 -7.48060942e-01
6.33303463e-01 1.75232142e-01 1.26362932e+00 1.29198685e-01
-6.69070542e-01 3.23368937e-01 -3.92687857e-01 -1.51855752e-01
-5.43867230e-01 -2.90562719e-01 -1.56323805e-01 3.70335847e-01
-2.08981454e-01 -6.00064337e-01 -1.09719694e+00 5.48584819e-01
-1.16252780e+00 6.21872962e-01 5.42022228e-01 7.64431655e-01
1.10587621e+00 1.44253388e-01 7.51753300e-02 -8.56614470e-01
5.69668293e-01 1.25194192e-01 -7.43827701e-01 5.88849485e-01
-5.62074006e-01 -3.26063275e-01 5.27493775e-01 -4.58043098e-01
-1.25997353e+00 6.45636320e-02 2.79292136e-01 -6.06337130e-01
-9.53604653e-02 -1.13050848e-01 -5.46446800e-01 -2.81719953e-01
6.93971038e-01 2.50643760e-01 3.21348459e-02 -3.58750731e-01
7.66996205e-01 4.50570047e-01 1.09410524e+00 -5.69344580e-01
1.11715460e+00 4.06702727e-01 -2.08049819e-01 -4.78541434e-01
-6.55517042e-01 -2.15622351e-01 -5.79139650e-01 -4.12779421e-01
9.88750458e-01 -9.56532419e-01 -4.27911043e-01 6.28013909e-01
-9.41044271e-01 -6.63967013e-01 -5.03366351e-01 4.05311808e-02
-5.91512263e-01 3.80821019e-01 -3.38752270e-01 -5.84075451e-01
-4.58324164e-01 -1.14196098e+00 1.07275486e+00 4.34000582e-01
-7.36446455e-02 -6.06309175e-01 1.39017433e-01 6.13365114e-01
5.14126301e-01 7.73758352e-01 7.25427151e-01 -3.86292674e-02
-4.80499268e-01 8.78381357e-02 -3.75135452e-01 5.53509176e-01
4.51551408e-01 3.09067279e-01 -1.09562778e+00 -2.76404947e-01
-1.88628197e-01 -1.49734437e-01 4.41791624e-01 2.84401655e-01
1.28309524e+00 -2.19385311e-01 1.78370580e-01 8.88132513e-01
1.42721367e+00 1.79288238e-01 8.92070949e-01 4.94785935e-01
8.61251295e-01 5.97003043e-01 6.69244170e-01 5.34769058e-01
2.57694721e-01 9.57992613e-01 2.13528529e-01 -1.03146434e+00
-8.01036954e-01 -9.62724462e-02 8.40863585e-02 5.18125892e-01
3.17163207e-02 -2.11210832e-01 -7.18636930e-01 2.83431888e-01
-1.80210912e+00 -7.55310416e-01 -4.82866205e-02 2.28169417e+00
1.12321138e+00 -1.52866513e-01 1.41630754e-01 1.05152458e-01
1.05513096e+00 7.66298771e-02 -6.58242285e-01 -1.16857745e-01
-5.92570245e-01 3.73110056e-01 3.98545533e-01 4.19225812e-01
-1.19500458e+00 1.02364993e+00 5.64880610e+00 9.97272670e-01
-1.26425111e+00 1.95816711e-01 1.11626446e+00 -4.43939827e-02
-3.13936353e-01 -1.33188143e-01 -2.44402379e-01 4.80331123e-01
2.60751933e-01 -1.42989561e-01 6.01732075e-01 5.22487104e-01
4.26936030e-01 -1.81123510e-01 -7.11514652e-01 1.41283727e+00
2.56491840e-01 -1.17952526e+00 1.16694845e-01 -3.27805489e-01
1.06836915e+00 -5.59131563e-01 5.83765388e-01 3.40493768e-02
1.76137045e-01 -8.06644678e-01 7.75560617e-01 4.66009438e-01
1.08188653e+00 -5.03070474e-01 3.79021972e-01 -3.94824475e-01
-1.20884919e+00 2.69223183e-01 9.67900604e-02 5.87434471e-01
1.11798681e-01 4.67480391e-01 -1.96018308e-01 6.73017144e-01
1.03926492e+00 3.94080400e-01 -1.12495339e+00 1.00067377e+00
-2.67116427e-01 1.86717331e-01 9.10038315e-03 5.87646246e-01
-3.35058421e-01 -5.41953802e-01 3.87991369e-01 1.14082229e+00
1.46789640e-01 7.24276453e-02 1.27407879e-01 1.02127230e+00
-2.41067708e-02 5.82114816e-01 -2.31325328e-01 2.91378587e-01
3.89849454e-01 1.48784089e+00 -8.01028967e-01 -3.30373794e-01
-1.63098752e-01 1.28912091e+00 -7.34712109e-02 5.00819683e-01
-1.14974189e+00 -3.94701302e-01 5.07886887e-01 5.24293184e-02
6.45168275e-02 5.42556979e-02 -5.11192739e-01 -1.16582978e+00
9.20079350e-02 -1.28768742e+00 2.81583160e-01 -1.11479115e+00
-1.18247306e+00 7.02925503e-01 -5.47215343e-03 -1.52894950e+00
1.72484919e-01 -2.26376086e-01 -6.36952877e-01 7.17069983e-01
-1.46505964e+00 -1.40423048e+00 -9.59465683e-01 9.07396138e-01
3.34337980e-01 -6.50041103e-02 5.99780977e-01 5.37070334e-01
-8.57808292e-01 8.91159594e-01 2.22360156e-02 -6.04666322e-02
1.22881806e+00 -1.20824921e+00 2.91993499e-01 9.42306519e-01
1.34909526e-01 2.69929886e-01 8.95512521e-01 -4.53567415e-01
-1.28403020e+00 -1.11674285e+00 1.47237539e-01 -2.56834835e-01
3.71183157e-01 -3.46070647e-01 -8.11243773e-01 1.61568895e-01
5.35373330e-01 1.05069585e-01 7.81163394e-01 -3.73738647e-01
-3.47158790e-01 -4.61704165e-01 -1.18691373e+00 8.89217496e-01
9.24903989e-01 -3.29200834e-01 -1.09636663e-02 2.57399112e-01
6.99708581e-01 -6.72648847e-01 -9.21091676e-01 5.13122976e-01
4.34475511e-01 -1.26883519e+00 1.04098272e+00 -5.28886579e-02
3.24617445e-01 -7.99996316e-01 -1.82297543e-01 -1.17482531e+00
-9.64507759e-02 -1.12710464e+00 2.02212885e-01 1.48497319e+00
3.79447103e-01 -3.64734143e-01 5.73861480e-01 6.28919482e-01
-3.80412638e-02 -3.79401207e-01 -6.33922338e-01 -5.45037925e-01
-3.36441368e-01 -1.32638156e-01 7.17252851e-01 1.11647940e+00
-3.80727261e-01 1.54649287e-01 -7.01359749e-01 1.00251868e-01
8.34615946e-01 4.45527792e-01 1.40140808e+00 -7.37338126e-01
-2.98870713e-01 -5.33483624e-01 -2.46250197e-01 -4.10340518e-01
-9.30994824e-02 -5.47378719e-01 -3.19745317e-02 -1.32844365e+00
2.57001013e-01 -4.05201644e-01 -2.57533729e-01 5.68480909e-01
-3.24656487e-01 9.55873668e-01 4.31676835e-01 2.63059944e-01
-6.43724501e-01 7.15886056e-01 1.61345792e+00 -2.34844774e-01
-4.13710326e-01 -4.30257261e-01 -7.76791871e-01 5.20170212e-01
9.25902128e-01 -2.26988252e-02 -4.70697790e-01 -4.25787866e-01
-1.17649890e-01 -1.30423173e-01 4.04387027e-01 -1.20581961e+00
2.78903060e-02 -2.05061764e-01 5.70265770e-01 -4.44909245e-01
3.31049591e-01 -7.10364342e-01 7.07912803e-01 2.50540912e-01
-2.74353713e-01 2.66392231e-01 4.18778777e-01 1.73893780e-01
-2.80939370e-01 2.14274645e-01 1.19605696e+00 5.38792051e-02
-6.74576759e-01 2.37351716e-01 2.76448935e-01 1.01700410e-01
1.17875016e+00 -2.51567811e-01 -3.44423294e-01 -4.73709673e-01
-6.34841800e-01 1.77974720e-02 8.76341164e-01 3.69726598e-01
4.82311606e-01 -1.46942604e+00 -6.99186563e-01 1.97855994e-01
2.91422606e-01 -1.57970995e-01 5.70068777e-01 7.05076516e-01
-6.28371954e-01 -3.98972839e-01 -4.97516602e-01 -5.75012922e-01
-1.32939935e+00 5.60880005e-01 3.30718398e-01 4.36919555e-03
-5.47887206e-01 5.04770935e-01 3.02463502e-01 -1.92717403e-01
2.92170227e-01 -2.83217486e-02 -7.41962567e-02 8.90707672e-02
3.10598701e-01 1.74163729e-01 -2.24426873e-02 -8.01311731e-01
-1.31186629e-02 8.08770597e-01 2.20115423e-01 -1.57554507e-01
9.96062517e-01 -3.78677249e-01 -2.27644369e-01 2.09687531e-01
1.09058666e+00 7.78812841e-02 -1.15401936e+00 -2.32595816e-01
-4.57121640e-01 -8.79416347e-01 -1.30879685e-01 -8.64311755e-01
-1.41107810e+00 4.82290089e-01 1.18354952e+00 4.04548086e-02
1.68502235e+00 -2.25103348e-01 7.00483680e-01 -1.17270462e-01
-6.42626584e-02 -1.17307711e+00 4.34411973e-01 -1.43095866e-01
1.09955394e+00 -1.35097742e+00 1.00635462e-01 -5.20771682e-01
-9.10831273e-01 7.01843739e-01 8.65883172e-01 3.87940928e-02
2.82134086e-01 1.59345210e-01 5.85254073e-01 -8.47557411e-02
-1.81196913e-01 -2.94408828e-01 4.77217078e-01 6.26539350e-01
2.98065871e-01 -2.99554504e-02 -2.38853976e-01 2.09716976e-01
-4.45830971e-01 -3.88779551e-01 2.19293773e-01 6.78166270e-01
-2.45135091e-02 -1.26954567e+00 -6.47711337e-01 -1.25124655e-03
-1.75736278e-01 -8.40141475e-02 -3.93792808e-01 8.18494081e-01
4.43745881e-01 9.12257850e-01 3.64190228e-02 -4.20145482e-01
5.36431968e-01 -2.92056948e-01 5.49409270e-01 -2.53980517e-01
-5.20242691e-01 4.13309246e-01 -1.35596886e-01 -6.24994695e-01
-6.61190271e-01 -3.79707307e-01 -7.45465338e-01 -2.76770592e-01
-1.18496984e-01 -4.60517406e-02 6.75950646e-01 3.21877301e-01
2.69312471e-01 5.90283990e-01 9.34792936e-01 -9.68671799e-01
1.77497603e-02 -6.70919538e-01 -5.06658137e-01 9.63004351e-01
1.57126367e-01 -6.28880501e-01 -2.45391846e-01 5.58462620e-01]
|
[11.262516975402832, -1.1942129135131836]
|
a4d0e9ae-8293-4664-a820-c93cd01de780
|
pali-x-on-scaling-up-a-multilingual-vision
|
2305.18565
| null |
https://arxiv.org/abs/2305.18565v1
|
https://arxiv.org/pdf/2305.18565v1.pdf
|
PaLI-X: On Scaling up a Multilingual Vision and Language Model
|
We present the training recipe and results of scaling up PaLI-X, a multilingual vision and language model, both in terms of size of the components and the breadth of its training task mixture. Our model achieves new levels of performance on a wide-range of varied and complex tasks, including multiple image-based captioning and question-answering tasks, image-based document understanding and few-shot (in-context) learning, as well as object detection, video question answering, and video captioning. PaLI-X advances the state-of-the-art on most vision-and-language benchmarks considered (25+ of them). Finally, we observe emerging capabilities, such as complex counting and multilingual object detection, tasks that are not explicitly in the training mix.
|
['Radu Soricut', 'Neil Houlsby', 'Xiaohua Zhai', 'Anelia Angelova', 'Mojtaba Seyedhosseini', 'Alexander Kolesnikov', 'Keran Rong', 'Yuanzhong Xu', 'Anurag Arnab', 'Daniel Keysers', 'Yang Li', 'Andreas Peter Steiner', 'Kenton Lee', 'Julien Amelot', 'Lucas Beyer', 'Ibrahim Alabdulmohsin', 'Gang Li', 'Austin Waters', 'Filip Pavetic', 'Matthias Minderer', 'AJ Piergiovanni', 'Marvin Ritter', 'Paulina Pietrzyk', 'Ceslee Montgomery', 'Bo Pang', 'Mandar Joshi', 'Hexiang Hu', 'Arsha Nagrani', 'Michael Tschannen', 'Mario Lucic', 'Daniel Salz', 'Mostafa Dehghani', 'Siamak Shakeri', 'Yi Tay', 'Xiao Wang', 'Sebastian Goodman', 'Carlos Riquelme Ruiz', 'Jialin Wu', 'Soravit Changpinyo', 'Basil Mustafa', 'Piotr Padlewski', 'Josip Djolonga', 'Xi Chen']
|
2023-05-29
| null | null | null | null |
['video-captioning', 'video-question-answering']
|
['computer-vision', 'computer-vision']
|
[ 2.43888453e-01 -4.12787884e-01 -2.35381886e-01 -2.17595860e-01
-1.34690177e+00 -6.88897789e-01 1.05675983e+00 1.71878040e-01
-7.45948553e-01 4.77419853e-01 3.36187214e-01 -3.75652373e-01
4.00505483e-01 -3.41629088e-01 -1.26076138e+00 -3.21764857e-01
1.50837943e-01 7.31782734e-01 4.04294640e-01 2.80340631e-02
1.44587249e-01 3.00078895e-02 -1.50602889e+00 7.66524673e-01
3.97020638e-01 7.69732714e-01 4.04229462e-01 1.11410570e+00
-2.65219629e-01 1.28808391e+00 -3.56988788e-01 -6.33064866e-01
-1.60353333e-01 3.99899818e-02 -7.09273636e-01 1.29485324e-01
1.57087493e+00 -6.10393405e-01 -6.55693948e-01 6.27439559e-01
2.68493593e-01 -1.18014023e-01 6.90295517e-01 -1.34459066e+00
-1.06093693e+00 2.78406441e-01 -7.22991824e-01 6.03005886e-01
3.92772079e-01 8.36665034e-01 8.75740170e-01 -1.18117464e+00
8.21452022e-01 1.59624159e+00 7.00630546e-01 5.02996981e-01
-1.05964720e+00 -6.72579169e-01 2.86423653e-01 4.45449024e-01
-1.01875556e+00 -7.55781531e-01 -1.44486859e-01 -7.09451675e-01
1.50380719e+00 -2.32541025e-01 3.79087955e-01 1.41203320e+00
8.28994811e-02 1.36635077e+00 1.09910774e+00 -3.87689084e-01
-1.15102783e-01 -8.36577863e-02 3.46509278e-01 1.07404852e+00
1.15023687e-01 3.39960605e-02 -6.04734361e-01 1.10835368e-02
5.30271292e-01 -3.24492335e-01 -2.60081124e-02 -1.46515027e-01
-1.54349124e+00 7.84888029e-01 -1.25040272e-02 1.06227338e-01
-2.47743726e-02 7.39764631e-01 6.56584263e-01 -1.00906223e-01
3.91045660e-01 2.58457810e-01 -5.64329267e-01 -4.89123538e-02
-9.90366995e-01 3.63612860e-01 6.53877974e-01 1.15642512e+00
6.19212210e-01 3.09526473e-02 -7.67237306e-01 7.64972925e-01
2.36763850e-01 1.00650239e+00 1.56174570e-01 -1.27266884e+00
8.76897871e-01 2.53276855e-01 -1.44921141e-02 -4.61147338e-01
-3.83410364e-01 -2.47009039e-01 -2.84090370e-01 5.95248975e-02
8.85716140e-01 1.04110939e-02 -1.17539132e+00 1.78217340e+00
1.78195819e-01 4.13758546e-01 6.75374418e-02 5.91916382e-01
1.07378817e+00 9.27004397e-01 6.54149175e-01 1.33366704e-01
1.68473375e+00 -1.55707109e+00 -4.30355936e-01 -7.69567907e-01
4.20744926e-01 -8.92871857e-01 1.22468340e+00 2.14875668e-01
-1.34139132e+00 -7.55931437e-01 -7.39619315e-01 -5.73800206e-01
-4.82859045e-01 1.72464013e-01 6.17808878e-01 5.07037044e-01
-1.28707528e+00 -1.20116599e-01 -4.77972567e-01 -6.04683876e-01
5.60622573e-01 -1.12152383e-01 -3.70127082e-01 -8.26277852e-01
-6.83908403e-01 1.06586242e+00 3.50442320e-01 -5.96137404e-01
-1.42587125e+00 -1.06582117e+00 -9.86914694e-01 -1.07742943e-01
5.42271793e-01 -8.69868159e-01 1.51724708e+00 -6.34529829e-01
-8.12208414e-01 1.31364548e+00 -1.00643039e-02 -5.32676578e-01
3.96075279e-01 -1.60268053e-01 -3.02316546e-01 5.23605227e-01
3.41439068e-01 1.51713407e+00 9.07802999e-01 -1.12601781e+00
-1.02434385e+00 -4.36163634e-01 1.37793601e-01 3.41418117e-01
1.40455980e-02 3.41795832e-01 -1.14985466e+00 -3.15558344e-01
-6.70642138e-01 -8.81826758e-01 1.81005940e-01 3.92420709e-01
-2.17791665e-02 -3.90759319e-01 9.59136605e-01 -1.02005315e+00
4.40283567e-01 -2.03555655e+00 -1.92299560e-01 -7.69094110e-01
2.49723569e-02 3.21267635e-01 -8.16123903e-01 1.98026389e-01
8.12220201e-02 -4.36497666e-02 -4.66929702e-03 -7.36190259e-01
-5.68691790e-02 4.35540378e-01 -1.47665054e-01 4.28290993e-01
2.90321290e-01 1.32834518e+00 -1.10365736e+00 -8.73311162e-01
4.80055273e-01 1.69536084e-01 -4.65663701e-01 8.08404759e-02
-8.62165213e-01 2.82055348e-01 4.94020544e-02 8.23556542e-01
3.61674011e-01 -4.56728697e-01 -2.72259116e-01 -3.02766979e-01
-1.17633067e-01 -2.29887143e-01 -7.00424373e-01 2.04745126e+00
-6.06572092e-01 1.07080638e+00 2.95486301e-01 -7.96365917e-01
6.50661588e-02 2.99025238e-01 3.51391166e-01 -9.45175052e-01
-1.88556045e-01 4.09815274e-02 -2.62796700e-01 -8.92063797e-01
5.81164896e-01 4.34563667e-01 8.84924233e-02 2.97069937e-01
6.57549858e-01 -1.29086003e-01 6.76759303e-01 5.71381569e-01
9.71335173e-01 2.15715274e-01 2.67813385e-01 -1.62495241e-01
4.26740676e-01 2.58308530e-01 -3.06266308e-01 1.32385910e+00
-3.45746011e-01 7.21878409e-01 8.35621953e-02 -2.18031406e-01
-1.37384796e+00 -1.16929948e+00 -6.46826029e-02 1.70833170e+00
-1.63379803e-01 -1.80801481e-01 -6.21396124e-01 -6.98973000e-01
9.15049165e-02 8.61401439e-01 -5.67424119e-01 2.60547936e-01
-5.12997210e-01 -5.74815452e-01 6.63218498e-01 6.14721119e-01
3.60478401e-01 -1.24590945e+00 -3.96722108e-01 1.14347376e-01
-5.09913921e-01 -1.86138582e+00 -8.37691426e-01 -9.85955074e-02
-4.88808483e-01 -1.17431271e+00 -7.31408715e-01 -8.56830180e-01
-1.20739704e-02 6.20964348e-01 1.88009751e+00 -1.10810824e-01
-7.59611726e-01 1.30802608e+00 -1.02736540e-01 -5.18804908e-01
-5.64831018e-01 -6.38401732e-02 -4.48291004e-01 -3.25398177e-01
2.89437413e-01 5.58551066e-02 -5.58867633e-01 -3.23084672e-03
-8.73641789e-01 1.31327733e-01 7.24321902e-01 5.12237132e-01
5.12392282e-01 -8.48890424e-01 4.30474162e-01 -3.92901659e-01
2.66683757e-01 -5.32435954e-01 -5.23345828e-01 7.36718416e-01
-1.70822635e-01 -1.97203845e-01 6.02460280e-02 -5.88701844e-01
-9.11814749e-01 -6.47107437e-02 -1.96221501e-01 -4.06924427e-01
-3.36992472e-01 2.74734907e-02 1.43557563e-01 -1.25398889e-01
6.25433087e-01 5.87545931e-01 -2.28704084e-02 -1.75965637e-01
9.72579002e-01 4.48146969e-01 1.01992273e+00 -6.90229475e-01
2.51215428e-01 5.43725312e-01 -3.67780775e-01 -1.01389253e+00
-1.23709822e+00 -8.63444686e-01 -2.49641463e-01 -3.70608240e-01
1.47948360e+00 -1.34189558e+00 -6.87707245e-01 6.04103923e-01
-1.45853519e+00 -4.77853596e-01 -2.20346600e-01 2.26070821e-01
-7.13847995e-01 3.30312312e-01 -7.90994465e-01 -5.83856285e-01
-4.89650458e-01 -1.31520689e+00 1.54621124e+00 9.17981640e-02
3.29246581e-01 -7.41145015e-01 8.26923773e-02 9.37572479e-01
3.28761846e-01 -9.81610790e-02 9.93566155e-01 -5.60502410e-01
-9.67205584e-01 1.34822190e-01 -8.65836501e-01 3.44203144e-01
-6.12109125e-01 -8.18895772e-02 -9.05277014e-01 -3.40666503e-01
-3.19140464e-01 -9.34962332e-01 1.22198486e+00 6.58715725e-01
1.04249728e+00 7.03319022e-03 -2.85125613e-01 4.61923152e-01
1.49632692e+00 -2.54244488e-02 6.30647838e-01 2.12166697e-01
7.84078300e-01 4.54620510e-01 2.71226853e-01 7.26349503e-02
7.03253865e-01 7.34142780e-01 5.97006619e-01 -1.47942621e-02
-7.02456892e-01 -1.85656637e-01 5.09396613e-01 4.39011663e-01
3.37389112e-01 -3.78874034e-01 -1.11196339e+00 8.65630746e-01
-1.82051921e+00 -1.10553694e+00 -1.08062066e-01 1.77323723e+00
7.74450183e-01 -1.47346079e-01 2.36101374e-01 -5.07175684e-01
5.25362611e-01 2.94860005e-01 -5.82388222e-01 -5.42969108e-02
-1.22023843e-01 -2.91827302e-02 5.67399025e-01 4.42218423e-01
-1.25377035e+00 1.20684457e+00 7.42108870e+00 1.10149157e+00
-8.03446591e-01 6.37344480e-01 7.11804509e-01 -1.52942538e-01
1.65572271e-01 -4.31333214e-01 -1.16921711e+00 1.75038338e-01
9.19944882e-01 4.04345840e-01 4.64028239e-01 7.92555928e-01
-2.59832293e-01 -4.41474259e-01 -1.26730216e+00 1.20600235e+00
6.55781627e-01 -1.68891561e+00 1.42752618e-01 -3.21995169e-01
6.69446230e-01 8.99348915e-01 9.29834023e-02 8.04737031e-01
2.87441522e-01 -1.05029869e+00 1.05722737e+00 4.38178957e-01
9.90653694e-01 -3.03747594e-01 2.49863580e-01 2.53159285e-01
-9.86864328e-01 -1.63641766e-01 1.53598944e-02 2.52526015e-01
1.18532404e-01 5.06358817e-02 -5.83577335e-01 1.37979284e-01
8.22840571e-01 5.74990213e-01 -9.64292943e-01 9.99957323e-01
4.15278226e-02 6.88849449e-01 -3.34528744e-01 1.09730974e-01
5.76842427e-01 -2.85482109e-02 3.58521104e-01 1.76816380e+00
-3.72846867e-03 -2.74783466e-02 6.00738347e-01 6.64196372e-01
-1.48826763e-01 5.90901859e-02 -4.87959772e-01 7.02913702e-02
1.91976532e-01 1.21401322e+00 -5.43970406e-01 -7.12708354e-01
-1.05493498e+00 7.50304639e-01 4.20787930e-01 4.58681554e-01
-9.88382161e-01 3.43731612e-01 4.34150547e-01 6.06584735e-02
4.70345467e-01 -3.75028819e-01 5.15507609e-02 -1.27502990e+00
-2.03145832e-01 -1.22840273e+00 4.34462667e-01 -1.25193703e+00
-1.32303488e+00 2.03657284e-01 2.23708570e-01 -4.37294096e-01
-2.04620153e-01 -9.85547662e-01 -3.37561011e-01 3.78130913e-01
-1.79409790e+00 -1.69988430e+00 -5.02475381e-01 8.72188449e-01
1.13080716e+00 -3.12358707e-01 5.77533007e-01 6.15415514e-01
-3.38310331e-01 3.15745890e-01 2.43563175e-01 1.13201946e-01
8.74417365e-01 -1.05475008e+00 4.67502743e-01 6.76643610e-01
6.18658781e-01 -6.57823831e-02 4.67096567e-01 -3.41204822e-01
-1.43410623e+00 -1.21858513e+00 6.44330561e-01 -9.01133358e-01
9.46363509e-01 -4.16371077e-01 -5.25603890e-01 8.64539862e-01
5.76832414e-01 5.94277196e-02 2.54246771e-01 3.19708437e-02
-6.70082092e-01 1.02348082e-01 -7.59632051e-01 5.35310447e-01
1.01380825e+00 -8.62211823e-01 -4.93674576e-01 9.74236965e-01
8.51888657e-01 -3.12552154e-01 -4.50117975e-01 9.68628004e-02
5.91179907e-01 -8.72042596e-01 1.60610044e+00 -5.40074646e-01
8.08732927e-01 1.40853375e-01 -4.87095356e-01 -6.99045837e-01
-2.07777053e-01 -1.01518713e-01 -1.91704974e-01 1.11101043e+00
3.36535662e-01 -1.89920533e-02 5.79164684e-01 3.53970043e-02
-2.20270470e-01 -5.65761447e-01 -7.79539526e-01 -6.56458914e-01
1.42862841e-01 -7.23737121e-01 -8.66787583e-02 5.95476210e-01
-6.35893822e-01 6.14626110e-01 -3.48370224e-01 -1.29797414e-01
8.38397443e-01 -1.63939536e-01 8.42822671e-01 -7.57092953e-01
-5.24715483e-01 -4.22763973e-01 -8.27421322e-02 -1.08956754e+00
1.48711205e-01 -9.32173491e-01 4.68130931e-02 -1.84134805e+00
6.08591020e-01 4.02751006e-02 8.78056958e-02 4.14719850e-01
-3.08126390e-01 4.76440847e-01 6.15009189e-01 3.03194493e-01
-1.34329057e+00 1.62226677e-01 1.21867621e+00 -5.57642639e-01
4.03405637e-01 -3.18843126e-01 -3.89793843e-01 7.30168223e-01
2.98644602e-01 -3.87645572e-01 -3.64511311e-01 -1.03473103e+00
9.94087942e-03 4.50111240e-01 6.76547587e-01 -1.22361279e+00
1.83847889e-01 -4.80467174e-03 3.45817655e-01 -9.36735451e-01
5.61430454e-01 -4.41879451e-01 -3.71250033e-01 4.39862162e-01
-4.70998317e-01 1.46911815e-01 4.90964562e-01 7.03201592e-01
8.18917826e-02 -3.60791564e-01 8.72469127e-01 -3.60826463e-01
-1.17889214e+00 5.10893404e-01 -4.07680869e-01 6.92897201e-01
9.88546848e-01 -1.37340529e-02 -9.09288824e-01 -3.42493981e-01
-6.40987098e-01 4.88332093e-01 2.06851006e-01 7.93625414e-01
3.38627577e-01 -9.66778338e-01 -1.19532144e+00 -3.23690325e-01
6.31564379e-01 -3.86179626e-01 3.77410889e-01 8.33798528e-01
-5.15152812e-01 8.07354152e-01 -2.81316519e-01 -1.01154900e+00
-1.22719634e+00 8.31621885e-01 2.01475859e-01 -6.33318484e-01
-5.57500660e-01 8.86642754e-01 4.22705799e-01 -3.15102845e-01
5.10901988e-01 -3.01277339e-01 -1.18744843e-01 8.91159549e-02
7.65307188e-01 2.08954677e-01 -2.06790894e-01 -4.21387374e-01
-3.03399593e-01 5.17484069e-01 -2.21994266e-01 -2.74013966e-01
9.02107120e-01 -2.46743605e-01 6.51784465e-02 2.74966925e-01
1.21003771e+00 -5.56702197e-01 -1.47000813e+00 -4.42847937e-01
-6.51720986e-02 -1.04743533e-01 6.38281554e-03 -1.18766057e+00
-7.29231834e-01 1.12023079e+00 7.43594050e-01 -3.54728311e-01
6.37895465e-01 4.07536000e-01 6.68335497e-01 6.31362379e-01
2.62914538e-01 -1.06764269e+00 5.34470558e-01 7.22572923e-01
7.67991304e-01 -1.75413013e+00 -1.03666142e-01 -6.88009411e-02
-5.71160316e-01 7.25483119e-01 7.86437690e-01 3.70396785e-02
1.83556065e-01 3.61960143e-01 -6.06066175e-02 -1.45938292e-01
-9.25535798e-01 -4.32435393e-01 3.79043132e-01 8.97737086e-01
2.77590841e-01 -1.61342412e-01 1.48264885e-01 9.81923565e-02
4.58168507e-01 -2.09766105e-02 2.56588668e-01 6.50322258e-01
-4.71090347e-01 -4.83625084e-01 -3.85503232e-01 4.56819773e-01
-4.33826298e-01 -4.86332327e-01 1.28820166e-01 1.01620030e+00
2.43245646e-01 9.26841140e-01 1.50468588e-01 3.46081704e-01
4.45781201e-02 1.96476221e-01 9.37783241e-01 -5.98910570e-01
-5.60636282e-01 -1.46650791e-01 2.58603334e-01 -6.99770987e-01
-4.31511641e-01 -6.81947529e-01 -8.07718217e-01 -4.03391719e-02
-1.46093071e-01 -4.40581232e-01 9.60642457e-01 1.22870731e+00
1.31953284e-01 5.25798023e-01 -3.45945090e-01 -1.00473654e+00
-5.89636028e-01 -9.13709044e-01 -7.15630502e-02 5.72874188e-01
3.39702338e-01 -3.28155577e-01 6.74470812e-02 3.63521129e-01]
|
[10.855392456054688, 1.5712324380874634]
|
35a465dd-4f4b-48cc-9a13-fff32d0ec9d4
|
evaluation-of-deep-learning-based-pose
|
1602.09065
| null |
http://arxiv.org/abs/1602.09065v3
|
http://arxiv.org/pdf/1602.09065v3.pdf
|
Evaluation of Deep Learning based Pose Estimation for Sign Language Recognition
|
Human body pose estimation and hand detection are two important tasks for
systems that perform computer vision-based sign language recognition(SLR).
However, both tasks are challenging, especially when the input is color videos,
with no depth information. Many algorithms have been proposed in the literature
for these tasks, and some of the most successful recent algorithms are based on
deep learning. In this paper, we introduce a dataset for human pose estimation
for SLR domain. We evaluate the performance of two deep learning based pose
estimation methods, by performing user-independent experiments on our dataset.
We also perform transfer learning, and we obtain results that demonstrate that
transfer learning can improve pose estimation accuracy. The dataset and results
from these methods can create a useful baseline for future works.
|
['Amir Ghaderi', 'Vassilis Athitsos', 'Srujana Gattupalli']
|
2016-02-29
| null | null | null | null |
['hand-detection']
|
['computer-vision']
|
[-2.89994568e-01 -5.41314185e-01 -3.76491934e-01 -3.09306234e-01
-7.16355681e-01 -3.76363367e-01 2.92637050e-01 -8.66229236e-01
-8.35280895e-01 5.53534329e-01 4.79190230e-01 1.78365111e-02
2.91100860e-01 -1.95924118e-01 -5.26584923e-01 -4.07354414e-01
-1.37224317e-01 7.53283799e-01 6.87025964e-01 -3.45277667e-01
1.08562090e-01 6.01347446e-01 -1.59932709e+00 1.87395096e-01
3.81462008e-01 7.52465010e-01 1.65653471e-02 7.74108112e-01
4.57023680e-01 6.14080608e-01 -6.83893502e-01 -3.98881435e-01
5.22496164e-01 -2.85737783e-01 -6.77062809e-01 -2.72808112e-02
7.88910329e-01 -7.57887065e-01 -7.62531459e-01 6.63274527e-01
1.22010410e+00 -3.44131887e-02 8.27169001e-01 -1.21655869e+00
-4.24881876e-02 1.55897215e-01 -7.30787396e-01 -8.07116404e-02
7.16631293e-01 5.86558461e-01 7.20345438e-01 -9.54312861e-01
7.88444996e-01 1.38892591e+00 6.93974018e-01 8.94437671e-01
-5.95547855e-01 -9.10613596e-01 6.03578277e-02 6.41609967e-01
-1.38646472e+00 -3.45461279e-01 8.17025900e-01 -5.91127038e-01
6.91083431e-01 -1.80344433e-02 1.04325461e+00 1.30076838e+00
-2.32752681e-01 1.51742268e+00 1.20289493e+00 -5.67458630e-01
-6.69359714e-02 -3.91079158e-01 -4.92877029e-02 6.48036242e-01
4.82071877e-01 1.41809508e-01 -7.82698214e-01 2.24526256e-01
9.68919456e-01 -3.43326718e-01 -4.68227595e-01 -7.27865100e-01
-1.08618546e+00 5.07932186e-01 4.23837155e-01 4.72844578e-02
-5.31150401e-02 2.33531177e-01 4.72241461e-01 1.54147461e-01
2.32362393e-02 2.83180177e-03 -4.20613319e-01 -2.80174643e-01
-7.55117059e-01 5.90971947e-01 7.98724651e-01 9.17183936e-01
6.53064623e-02 -1.93248227e-01 -3.05319279e-01 6.57348692e-01
8.34826648e-01 6.45064831e-01 5.69875419e-01 -6.24387264e-01
6.61605120e-01 2.48342469e-01 1.58691883e-01 -7.09358990e-01
-5.55780113e-01 -3.63560729e-02 -3.30176920e-01 5.74915946e-01
1.22974741e+00 -3.15159202e-01 -1.40128589e+00 1.34540892e+00
9.68702212e-02 -2.26267964e-01 -2.54611105e-01 1.53138316e+00
1.00282943e+00 5.43143377e-02 8.38416293e-02 2.34419793e-01
1.24438250e+00 -1.12991989e+00 -6.74156010e-01 -2.19758481e-01
4.52030629e-01 -8.41810346e-01 1.17209589e+00 6.80525303e-01
-8.75116229e-01 -2.24531725e-01 -7.91428864e-01 -2.23255277e-01
-6.55915439e-02 6.53213799e-01 6.37148917e-01 8.15492511e-01
-8.09712172e-01 2.65950590e-01 -9.24930096e-01 -7.65538156e-01
4.46425945e-01 3.80337507e-01 -3.30566049e-01 7.02881953e-03
-9.27550137e-01 1.07135224e+00 1.12277985e-01 2.52235204e-01
-6.98722363e-01 1.96442362e-02 -7.11839080e-01 -5.45416355e-01
3.16725254e-01 -4.70070988e-01 1.59225643e+00 -5.63985288e-01
-1.73614061e+00 1.36120641e+00 -2.67302543e-02 -3.45133781e-01
1.18449306e+00 -6.18655562e-01 -2.66177095e-02 8.50210413e-02
-3.98253173e-01 6.62768662e-01 9.31023359e-01 -1.06435907e+00
-5.03527701e-01 -7.03769803e-01 -1.67447597e-01 1.76480502e-01
-1.59283593e-01 2.87887543e-01 -7.13467360e-01 -6.42207086e-01
2.57684141e-01 -1.11176646e+00 -2.69370526e-02 3.82071376e-01
-2.46607155e-01 -3.25956255e-01 7.05493510e-01 -1.14078009e+00
9.00132835e-01 -1.70266020e+00 3.90824944e-01 1.64828181e-01
9.57738087e-02 7.11437881e-01 1.28706042e-02 2.78306574e-01
2.98638105e-01 -2.64456034e-01 5.40905744e-02 -2.19812140e-01
-8.02533235e-03 -7.11177588e-02 -1.04463868e-01 6.64875507e-01
-2.12218836e-01 1.02434123e+00 -6.71019316e-01 -6.85778260e-01
3.27840090e-01 4.22479719e-01 -4.04644102e-01 2.77959257e-01
-1.82827041e-01 8.54887426e-01 -1.02597773e-01 8.51320982e-01
3.62150490e-01 1.26634121e-01 1.97928637e-01 -5.33310354e-01
1.23796761e-01 9.39882770e-02 -1.19848096e+00 1.69528401e+00
-7.36009702e-02 9.88686085e-01 -1.00099361e-02 -6.95006192e-01
6.23012245e-01 2.80876517e-01 4.62004006e-01 -5.23064792e-01
5.46869457e-01 3.49033743e-01 3.49192977e-01 -6.53853536e-01
2.34161735e-01 3.27890925e-02 3.40205908e-01 3.54697078e-01
4.03145142e-02 -7.80210122e-02 1.81678146e-01 -3.04412037e-01
8.27087283e-01 5.59971631e-01 3.06967288e-01 2.38789588e-01
4.09051955e-01 -1.13514222e-01 2.45353952e-01 5.72264612e-01
-7.22464323e-01 9.83099401e-01 1.37051672e-01 -5.19864559e-01
-9.84315157e-01 -8.52564991e-01 8.05300921e-02 9.96141315e-01
5.38167246e-02 -2.41119415e-01 -6.40816152e-01 -7.19586551e-01
1.92618862e-01 -7.89348483e-02 -4.37630713e-01 2.11944282e-01
-9.39967334e-01 -2.76383013e-01 7.40427673e-01 1.19010139e+00
5.42042077e-01 -1.27420425e+00 -9.04830158e-01 -2.46812493e-01
-2.00735852e-01 -1.27053964e+00 -5.09924650e-01 -4.60366338e-01
-7.09876180e-01 -1.33442652e+00 -1.45853388e+00 -1.09505761e+00
4.66545045e-01 -1.04837120e-01 7.91303098e-01 -6.74541220e-02
-5.53015530e-01 6.65763319e-01 -4.48031276e-01 -4.75763917e-01
-4.36371714e-02 5.06367572e-02 1.80490300e-01 -2.45975494e-01
4.73375112e-01 -2.85138875e-01 -7.21169949e-01 7.04852164e-01
-2.26268962e-01 -1.52251534e-02 7.24107623e-01 7.67156780e-01
3.07278395e-01 -9.01082098e-01 1.50293618e-01 -2.07667366e-01
4.91793036e-01 4.20018762e-01 -6.49388075e-01 2.05955654e-01
-7.94326328e-03 9.66997817e-02 -1.84314903e-02 -6.28848910e-01
-9.55388010e-01 4.85032320e-01 -3.66576046e-01 -3.28223944e-01
-2.45073870e-01 2.61639029e-01 -2.89775610e-01 -4.04620796e-01
6.87095821e-01 1.50609419e-01 2.87609816e-01 -6.97863758e-01
2.79294223e-01 9.10756946e-01 6.01111889e-01 -4.60979611e-01
6.49770737e-01 5.16219139e-01 -3.96610983e-02 -8.54529560e-01
-5.73074162e-01 -6.38813496e-01 -9.00724888e-01 -5.48592389e-01
7.78899372e-01 -9.64974642e-01 -1.05630839e+00 1.03636312e+00
-9.91869986e-01 -6.12479746e-01 8.49155411e-02 8.80558431e-01
-8.49702537e-01 5.12478530e-01 -5.15142143e-01 -1.00228620e+00
-1.72931746e-01 -1.12740326e+00 1.23098052e+00 1.11928441e-01
-3.84964168e-01 -5.43427646e-01 2.97354937e-01 4.26510513e-01
1.93232179e-01 2.19936222e-01 1.11024119e-01 -3.62432867e-01
-5.47663212e-01 -5.26446819e-01 -1.74359038e-01 2.07136750e-01
-2.97238007e-02 -2.06310630e-01 -8.92346919e-01 -6.04547381e-01
-7.95071185e-01 -9.08936143e-01 1.08165717e+00 5.72813153e-01
1.14857364e+00 1.26794353e-01 -3.27494085e-01 4.62131441e-01
7.99726307e-01 -2.13208925e-02 6.64107680e-01 3.28148395e-01
8.38499248e-01 5.32702923e-01 6.39703691e-01 5.18582523e-01
4.97361034e-01 1.16546786e+00 2.02382263e-02 -9.92350206e-02
-5.96372366e-01 -3.06395382e-01 3.74021769e-01 5.45949996e-01
-9.80212033e-01 1.01601563e-01 -1.23898327e+00 3.11827004e-01
-1.99222565e+00 -8.36232603e-01 1.46405622e-01 2.24119163e+00
9.35690343e-01 -1.56663701e-01 8.76095235e-01 4.91260618e-01
4.37348872e-01 -1.49436265e-01 -3.97111624e-01 2.10908964e-01
1.53655648e-01 4.10495639e-01 4.73613381e-01 2.97164053e-01
-1.40441036e+00 1.26955092e+00 6.91507435e+00 2.20892310e-01
-1.31436920e+00 -1.11159548e-01 -1.09223917e-01 -2.49760434e-01
5.95014513e-01 -5.43666959e-01 -8.95518839e-01 1.33178890e-01
1.28076121e-01 2.71050721e-01 9.28164795e-02 9.43401039e-01
6.84734881e-02 -1.16815038e-01 -1.36066151e+00 1.46530533e+00
3.22143912e-01 -5.29104650e-01 -1.25149235e-01 -1.03748888e-01
4.33241904e-01 9.64674819e-03 -3.33767161e-02 4.30580705e-01
1.76990807e-01 -1.03950977e+00 6.61420643e-01 3.99719685e-01
9.89773393e-01 -4.95815217e-01 7.35748768e-01 3.21174473e-01
-1.29679871e+00 -3.94309498e-02 -8.27915296e-02 -3.91015053e-01
2.21045524e-01 -1.96075544e-01 -7.41538763e-01 5.64762838e-02
8.76311541e-01 8.89053345e-01 -7.43458271e-01 1.63275051e+00
-7.30850101e-01 6.36749029e-01 -3.74271274e-01 -4.07734692e-01
-3.59973758e-01 1.61259547e-01 7.12067783e-01 1.21255207e+00
-4.32514064e-02 -1.15102991e-01 3.16854835e-01 4.35445398e-01
2.11785100e-02 2.52325892e-01 -4.74007010e-01 1.06899083e-01
9.32004582e-03 8.48340690e-01 -6.37218595e-01 -1.96304098e-01
-3.19392115e-01 1.10932398e+00 1.18788540e-01 3.54490936e-01
-6.56264126e-01 -3.20171714e-01 6.94420576e-01 2.71035731e-01
2.61354297e-01 -6.50610864e-01 -2.86366314e-01 -1.47896779e+00
3.85298043e-01 -1.00762737e+00 3.72785449e-01 -6.16617203e-01
-1.07760394e+00 5.47398776e-02 -7.07211392e-03 -1.49478877e+00
-5.87300360e-01 -1.20669210e+00 -2.68994629e-01 6.34208977e-01
-1.32296336e+00 -1.34941781e+00 -6.36287391e-01 7.24899113e-01
4.92307603e-01 -4.00626272e-01 6.24031782e-01 4.31900620e-01
-1.78242981e-01 1.05574572e+00 -3.64342391e-01 7.57619560e-01
1.00057125e+00 -1.16933525e+00 2.78029233e-01 7.10910082e-01
3.12817484e-01 4.41682965e-01 5.46016276e-01 -6.43053234e-01
-1.46258867e+00 -5.29833019e-01 4.48854059e-01 -7.15087712e-01
1.87270433e-01 -3.53056043e-01 -4.05803680e-01 7.59542882e-01
-1.13429859e-01 1.15697183e-01 3.73074174e-01 1.22267820e-01
-3.55358303e-01 8.63904953e-02 -1.05099344e+00 6.53883576e-01
1.50586021e+00 -2.25703388e-01 -7.90766060e-01 2.82589883e-01
1.01938751e-02 -9.24379826e-01 -4.50496107e-01 6.22269094e-01
1.43623698e+00 -6.56675220e-01 9.92074966e-01 -8.21870267e-01
2.47849166e-01 -3.49452674e-01 2.14877594e-02 -1.02185118e+00
-2.11528572e-03 -1.60867065e-01 -3.06933105e-01 6.68758571e-01
1.19342916e-01 -3.79897416e-01 1.10713923e+00 4.72591400e-01
3.59642118e-01 -4.07816082e-01 -8.76971602e-01 -1.02442765e+00
5.86182959e-02 -5.10605872e-01 3.99029441e-02 2.96273679e-01
-1.76904649e-02 2.17297107e-01 -6.04491174e-01 -9.84222144e-02
7.84412086e-01 5.28411120e-02 1.39470315e+00 -1.26754177e+00
-3.03616822e-01 -5.54893672e-01 -8.62457454e-01 -1.16969001e+00
8.18028289e-04 -4.86011058e-01 2.14858606e-01 -1.67419636e+00
1.66030049e-01 2.08951354e-01 2.16534182e-01 7.10337818e-01
2.58459803e-02 5.54495573e-01 4.48542386e-01 2.96741694e-01
-6.31892025e-01 4.74368393e-01 1.44469893e+00 -1.39931813e-01
-2.15823904e-01 3.43747377e-01 1.48967966e-01 1.00815725e+00
7.70429432e-01 4.05472815e-02 9.26905274e-02 -2.85975069e-01
-2.17598334e-01 -2.38766864e-01 5.20991504e-01 -1.28758836e+00
2.61362135e-01 1.66602254e-01 6.11854553e-01 -8.93558204e-01
4.05232131e-01 -6.82244301e-01 -5.15352666e-01 9.15728509e-01
-1.07346684e-01 -4.56564039e-01 1.74002945e-01 1.22001700e-01
-2.10534573e-01 1.34422988e-01 9.34833348e-01 -1.20651819e-01
-1.18618298e+00 3.95121843e-01 -3.58001977e-01 3.66847187e-01
8.71406972e-01 -3.40074688e-01 2.91511297e-01 -7.15038121e-01
-8.28829825e-01 1.73294663e-01 1.97615668e-01 5.64966440e-01
8.20407927e-01 -1.36688304e+00 -8.05776477e-01 1.96334302e-01
2.78659314e-01 -6.12902828e-02 -4.32346135e-01 7.07605958e-01
-7.55570233e-01 5.55192828e-01 -5.26283741e-01 -8.35543096e-01
-1.73499858e+00 8.53578746e-02 3.74063581e-01 9.49345231e-02
-6.89267635e-01 8.18029821e-01 -2.72460282e-01 -5.51308572e-01
9.29938078e-01 -3.74546081e-01 -3.18554521e-01 -7.06666261e-02
6.21917844e-01 4.66986954e-01 -1.45389035e-01 -6.40453160e-01
-5.65540016e-01 1.00944948e+00 4.36983667e-02 -3.74977201e-01
1.20946860e+00 3.37964058e-01 3.09316427e-01 1.98644981e-01
7.99091280e-01 -1.75827876e-01 -1.28931177e+00 -2.95086622e-01
1.18555408e-02 -6.61225379e-01 -2.47186109e-01 -1.07468998e+00
-1.07158923e+00 1.23585331e+00 9.35725749e-01 -7.91973174e-01
1.01904941e+00 7.56506696e-02 6.63747013e-01 5.17809808e-01
7.99077213e-01 -1.29974127e+00 3.89167339e-01 6.92011654e-01
1.31946027e+00 -1.58800161e+00 -8.30425620e-02 -3.21642905e-01
-5.15483677e-01 1.17870295e+00 9.94958580e-01 -1.41008168e-01
5.51824093e-01 2.88330376e-01 4.30318892e-01 1.36787295e-01
5.04514165e-02 -7.67995358e-01 7.57784367e-01 9.10787880e-01
6.45672739e-01 5.13956062e-02 -8.06748927e-01 7.05034018e-01
-2.05587104e-01 4.12533522e-01 2.93476209e-02 1.02159977e+00
-2.06956312e-01 -1.25701368e+00 -6.74048603e-01 2.16327623e-01
-8.47437829e-02 4.46083575e-01 -9.05507922e-01 1.05871928e+00
-1.32430673e-01 3.40451241e-01 -4.48398381e-01 -5.35293400e-01
5.85474730e-01 1.82783172e-01 1.27723932e+00 -3.77515435e-01
-1.57313004e-01 -9.18852631e-03 -6.01547882e-02 -6.09214902e-01
-3.18819255e-01 -7.32379019e-01 -1.10818303e+00 4.14503887e-02
-2.52115339e-01 -6.49535656e-01 4.85905498e-01 9.47819948e-01
-2.06548497e-01 2.44112372e-01 1.61586907e-02 -1.44227850e+00
-9.01422918e-01 -1.30869198e+00 -5.19878864e-01 6.65301859e-01
1.10369958e-01 -1.08087182e+00 4.73799407e-02 -9.86762997e-03]
|
[9.131863594055176, -6.44445276260376]
|
48eb0e8b-816e-4f8d-80a3-2b80f6b51769
|
realistic-simulation-of-users-for-it-systems
|
2111.11785
| null |
https://arxiv.org/abs/2111.11785v1
|
https://arxiv.org/pdf/2111.11785v1.pdf
|
Realistic simulation of users for IT systems in cyber ranges
|
Generating user activity is a key capability for both evaluating security monitoring tools as well as improving the credibility of attacker analysis platforms (e.g., honeynets). In this paper, to generate this activity, we instrument each machine by means of an external agent. This agent combines both deterministic and deep learning based methods to adapt to different environment (e.g., multiple OS, software versions, etc.), while maintaining high performances. We also propose conditional text generation models to facilitate the creation of conversations and documents to accelerate the definition of coherent, system-wide, life scenarios.
|
['Adrien Bécue', 'Éric Totel', 'Benjamin Costé', 'Alexandre Dey']
|
2021-11-23
| null | null | null | null |
['conditional-text-generation']
|
['natural-language-processing']
|
[-1.58180907e-01 -3.21061462e-01 -1.91901252e-01 -1.54442161e-01
-3.59938651e-01 -9.06476915e-01 9.68692422e-01 2.24220723e-01
-2.49534220e-01 5.84091544e-01 4.31697369e-02 -7.82159567e-01
2.29186609e-01 -1.07076049e+00 -2.60941774e-01 -3.78505677e-01
-1.86803937e-02 3.77863288e-01 3.66597682e-01 -5.60871325e-02
3.95923465e-01 6.11805081e-01 -1.38330483e+00 2.22036526e-01
5.12943387e-01 5.04182518e-01 -1.27360374e-01 9.31912243e-01
-2.45690554e-01 8.07241082e-01 -1.37949479e+00 -2.99489975e-01
-3.09595793e-01 -1.61387831e-01 -6.33685946e-01 -2.23297939e-01
-4.74632651e-01 -3.54479522e-01 3.11574519e-01 1.06405592e+00
3.95813882e-01 -1.48483694e-01 4.35788751e-01 -1.44574869e+00
-5.04988916e-02 1.25029552e+00 -5.56378722e-01 7.58206397e-02
4.73822504e-01 4.22398508e-01 4.77027059e-01 -1.97180025e-02
3.21954161e-01 9.59889531e-01 1.55126259e-01 5.43540716e-01
-9.83969271e-01 -7.65911460e-01 -1.39106259e-01 2.22871080e-02
-8.84018362e-01 -3.52099925e-01 1.02580452e+00 -4.60296512e-01
6.22927845e-01 2.73521692e-01 1.89877093e-01 1.91270578e+00
3.87000203e-01 4.65381354e-01 9.86759424e-01 -7.20721126e-01
6.49020553e-01 4.65402871e-01 2.64679015e-01 4.19236481e-01
4.65574145e-01 7.75184017e-03 -2.73632586e-01 -7.26530313e-01
2.95922965e-01 -2.38789231e-01 2.48826697e-01 2.11472824e-01
-8.23701203e-01 7.82647073e-01 -4.11641628e-01 7.03645051e-01
-4.28328454e-01 1.67847499e-01 7.11142242e-01 9.02875587e-02
1.86984494e-01 6.79971159e-01 -3.61614674e-01 -7.19473541e-01
-7.41405845e-01 -1.39888041e-02 1.23684275e+00 4.46284115e-01
4.48406130e-01 2.29429111e-01 -6.69189543e-02 1.04519159e-01
4.56584275e-01 3.49989325e-01 7.82499433e-01 -7.03454733e-01
2.69908130e-01 6.85957372e-01 2.89904952e-01 -8.38296592e-01
-3.84617358e-01 -5.77988207e-01 -5.66302240e-01 2.06376120e-01
7.47719482e-02 -6.35195136e-01 -4.50780213e-01 1.55669701e+00
3.61601800e-01 4.13348973e-01 -1.34671122e-01 7.99599960e-02
1.32293150e-01 5.61062634e-01 2.93356240e-01 -2.21467242e-01
1.45953476e+00 -5.15644729e-01 -6.21899664e-01 2.08912455e-02
7.01389670e-01 -4.59657282e-01 1.21624207e+00 7.08470941e-01
-5.86086392e-01 -4.49751168e-01 -1.18251908e+00 8.61406684e-01
-6.29968882e-01 -1.73276421e-02 5.11771858e-01 1.43293643e+00
-6.34234846e-01 4.04947460e-01 -1.08556950e+00 -1.08203694e-01
-1.08351104e-01 3.59650217e-02 1.81570083e-01 6.57337844e-01
-1.44702828e+00 6.97024226e-01 5.33004642e-01 -3.52973968e-01
-1.22230887e+00 -2.48475790e-01 -4.15091962e-01 1.27172261e-01
3.49828243e-01 -3.35402876e-01 1.44862247e+00 -5.62929571e-01
-1.86362255e+00 3.89280140e-01 3.10819685e-01 -7.14667380e-01
4.68781471e-01 -2.78624684e-01 -6.01321936e-01 -2.09141314e-01
-3.06701243e-01 -2.03924999e-01 1.09073830e+00 -1.23777950e+00
-4.46946263e-01 -2.51449257e-01 3.95770252e-01 -3.24661821e-01
-7.10368931e-01 5.41346312e-01 1.17352687e-01 -4.24849987e-02
-1.01961446e+00 -5.27961135e-01 -2.32427359e-01 -1.16543555e+00
-6.67548835e-01 -1.76856533e-01 1.02687216e+00 -6.57772422e-01
1.40386856e+00 -1.86738384e+00 -3.18825543e-01 5.06315351e-01
2.83543319e-01 5.30844808e-01 1.75330415e-01 5.80138147e-01
1.73940971e-01 7.30480075e-01 2.63219863e-01 -5.00035107e-01
2.27823257e-01 -7.49517158e-02 -3.44722927e-01 1.54835999e-01
1.15698420e-01 4.01684582e-01 -7.95192659e-01 -4.49657828e-01
4.47475553e-01 3.37993175e-01 -1.40138403e-01 3.89026642e-01
-6.46596670e-01 5.83006740e-01 -6.33214116e-01 4.16845411e-01
1.87501594e-01 -2.56585777e-01 4.36012924e-01 3.17796320e-01
-1.90175220e-01 2.64710546e-01 -1.24603045e+00 1.09888721e+00
-1.11594617e+00 3.06532979e-01 -1.57848403e-01 -5.78028858e-01
1.02386940e+00 3.50777745e-01 2.68505692e-01 -2.93829799e-01
7.64811933e-01 -1.71129540e-01 -2.05194689e-02 -3.49881262e-01
3.77541095e-01 4.08426940e-01 -2.04339758e-01 1.37259233e+00
-1.53868645e-01 3.27046335e-01 2.81749666e-01 1.17339738e-01
1.25947225e+00 -2.98189312e-01 6.22194171e-01 3.64146650e-01
7.55913019e-01 -4.73586500e-01 2.51276523e-01 9.44072008e-01
-1.89649895e-01 -1.78457901e-01 8.77634585e-01 -2.70916551e-01
-8.41401100e-01 -7.68353462e-01 1.97031498e-01 9.85761464e-01
-2.17060983e-01 -6.57072902e-01 -1.17064631e+00 -9.41618979e-01
-5.68815649e-01 1.13325274e+00 -5.00597894e-01 -4.50710356e-01
-4.51434076e-01 -8.24156821e-01 1.08056903e+00 2.64018029e-01
4.76941407e-01 -1.21597123e+00 -1.08470976e+00 5.29121220e-01
-1.76467642e-01 -1.12594187e+00 -3.54383327e-02 6.75276071e-02
-4.90547031e-01 -1.04688787e+00 2.95849293e-01 2.13553011e-03
1.25795648e-01 -2.07459167e-01 1.16453576e+00 2.42931709e-01
-1.77709103e-01 2.91766256e-01 -7.81180978e-01 -6.22204602e-01
-1.22386038e+00 4.56320912e-01 8.89055580e-02 2.52502739e-01
4.22952920e-01 -5.48406482e-01 -1.33102238e-01 1.74109697e-01
-1.13328350e+00 -3.75071049e-01 6.11239076e-01 5.38175046e-01
-1.27843007e-01 7.05536723e-01 4.56202328e-01 -9.64748561e-01
1.44482577e+00 -6.75670385e-01 -9.22221005e-01 4.04908299e-01
-8.76259089e-01 3.84766683e-02 8.64588618e-01 -8.11096728e-01
-9.94539380e-01 -3.45877171e-01 -1.67808115e-01 4.59307879e-02
-5.89656830e-01 4.98134881e-01 -4.41935480e-01 1.90021068e-01
9.82302070e-01 2.22317547e-01 -1.61849678e-01 -3.04443359e-01
4.99626666e-01 9.51632321e-01 2.22069710e-01 -7.95998752e-01
1.02148211e+00 8.02085623e-02 -3.34524125e-01 -6.99045718e-01
-2.06315756e-01 -8.15070868e-02 -3.10770243e-01 -3.29404145e-01
5.40051520e-01 -2.15229332e-01 -9.42671001e-01 6.80113792e-01
-1.30167902e+00 -3.73454273e-01 1.70174628e-01 1.00095719e-01
-1.61756307e-01 4.98219818e-01 -7.22466826e-01 -1.15382266e+00
-8.26850593e-01 -1.25273180e+00 7.33351767e-01 4.90244478e-01
-5.51460326e-01 -1.10568464e+00 5.71467280e-01 1.53996930e-01
5.38851142e-01 3.98811072e-01 7.83664346e-01 -1.11083090e+00
-3.48276585e-01 -6.70936048e-01 1.40339404e-01 4.87367809e-01
6.00548573e-02 8.04987371e-01 -8.98183942e-01 -2.01757446e-01
2.60591447e-01 -2.84314975e-02 -1.59929290e-01 -7.48937950e-02
1.09625781e+00 -4.77976292e-01 -1.91932544e-01 2.12286010e-01
1.06695688e+00 6.13948941e-01 5.89976728e-01 4.05249387e-01
2.05173478e-01 6.34607315e-01 3.93482178e-01 9.36407745e-01
7.74672255e-02 6.56630874e-01 4.41114336e-01 1.34422794e-01
4.96541500e-01 -2.77355641e-01 6.57737851e-01 2.64492899e-01
2.33647481e-01 -4.19798523e-01 -1.02157140e+00 2.97121261e-03
-1.32899404e+00 -9.55851376e-01 1.09901644e-01 2.09023356e+00
6.95748925e-01 6.81075633e-01 2.83472717e-01 3.57310236e-01
7.00283587e-01 2.48899192e-01 -2.91996747e-01 -5.77460885e-01
3.78430516e-01 4.63113666e-01 2.74162918e-01 4.86106962e-01
-7.75721848e-01 8.35713089e-01 5.74787092e+00 6.62179828e-01
-1.22681403e+00 4.22885299e-01 3.19678754e-01 3.01912636e-01
-1.54640839e-01 1.88564375e-01 -8.10608268e-01 8.74657273e-01
1.60226738e+00 -3.98946077e-01 4.19142812e-01 1.05171525e+00
4.32521224e-01 7.55403042e-02 -7.20727742e-01 5.32262444e-01
8.17815438e-02 -1.33808970e+00 -1.92707121e-01 1.44452840e-01
2.89765805e-01 -1.14906304e-01 -1.66768581e-01 3.37755620e-01
8.50101292e-01 -7.74732530e-01 5.90193808e-01 9.74301323e-02
4.80643809e-01 -8.04782808e-01 1.03425944e+00 5.95628202e-01
-7.85285950e-01 -6.46417513e-02 3.22283477e-01 5.23719378e-02
2.33580053e-01 7.78827131e-01 -1.27539802e+00 4.87530679e-01
3.77081037e-01 -8.62939507e-02 -8.72410059e-01 7.36117065e-01
-5.76968193e-01 9.84679818e-01 -1.88683793e-01 -3.86928886e-01
-6.57674447e-02 1.66535199e-01 4.27483767e-01 1.16858459e+00
1.16035119e-01 -5.46716571e-01 1.90980166e-01 8.33527744e-01
4.97428365e-02 1.65426563e-02 -6.12203836e-01 -3.64968985e-01
8.66298139e-01 1.50504351e+00 -6.48204923e-01 -2.49357954e-01
-5.25050936e-03 5.32705188e-01 -8.78550857e-02 1.95608720e-01
-9.61138010e-01 -5.84018946e-01 4.62609410e-01 9.25629959e-02
-5.53266108e-01 -3.98881912e-01 -2.89816350e-01 -1.02125180e+00
-2.03452960e-01 -1.16245902e+00 2.52565503e-01 -3.64374459e-01
-9.31778729e-01 8.37664664e-01 8.34116861e-02 -8.88314307e-01
-7.50908434e-01 -4.46184337e-01 -1.35175073e+00 6.19266748e-01
-1.03740466e+00 -1.16097212e+00 -3.14491391e-01 3.93827111e-01
2.96345744e-02 -4.60640460e-01 7.94236600e-01 1.71679497e-01
-9.36930537e-01 6.04181528e-01 -2.40724027e-01 2.95594156e-01
5.76806664e-01 -1.38605034e+00 5.93323052e-01 1.13473105e+00
2.92654365e-01 8.74129653e-01 8.68814409e-01 -7.51938164e-01
-1.03179085e+00 -8.85526657e-01 4.03215915e-01 -7.10630417e-01
9.75630403e-01 -5.56821227e-01 -8.12928557e-01 5.17813742e-01
1.90281957e-01 -7.54507184e-01 9.17050302e-01 2.44869769e-01
-3.90952021e-01 -1.31031454e-01 -1.11469376e+00 6.89644694e-01
2.45708480e-01 -5.53515851e-01 -1.82137951e-01 1.70408815e-01
7.97324538e-01 -1.25081003e-01 -6.78378642e-01 1.61676891e-02
3.76292557e-01 -1.10552657e+00 5.39542973e-01 -7.92840898e-01
1.48735044e-03 -3.38278949e-01 1.16872475e-01 -1.14004862e+00
7.37146288e-02 -1.06046450e+00 -6.97074533e-01 1.65290225e+00
4.29294258e-01 -7.54651010e-01 8.56150389e-01 3.64422679e-01
4.34192985e-01 -2.38065526e-01 -5.71820438e-01 -6.77934766e-01
-2.66762674e-01 -6.67841613e-01 1.25864732e+00 8.40007782e-01
6.83186948e-02 4.39423233e-01 -4.22599196e-01 3.19107980e-01
6.37137949e-01 -3.47013354e-01 1.12819338e+00 -1.23858976e+00
-6.56480432e-01 -7.23337114e-01 -2.32419387e-01 -1.24224834e-01
3.77239108e-01 -2.67708719e-01 -3.22921574e-01 -8.06984484e-01
-4.86589484e-02 -3.01995486e-01 -3.10370833e-01 2.44055465e-01
3.64737250e-02 -3.01156342e-01 4.87170042e-03 -1.76922269e-02
-5.35582840e-01 1.53275222e-01 2.72708088e-01 6.62103146e-02
-4.94758606e-01 4.15634215e-01 -6.89613700e-01 6.97433710e-01
1.42045736e+00 -5.06588757e-01 -3.61518204e-01 2.34821126e-01
1.97922781e-01 -4.89266496e-03 1.33389801e-01 -1.20749998e+00
1.60694681e-02 -4.12044704e-01 1.52798498e-03 -2.61152208e-01
-2.07444176e-01 -5.47562301e-01 2.34357208e-01 6.92915678e-01
-3.44436586e-01 2.51523435e-01 -1.46911964e-01 4.39274430e-01
-4.12215553e-02 -6.02290332e-01 6.85368955e-01 -1.16745129e-01
-3.03497672e-01 1.07850619e-01 -6.46623492e-01 -1.74276516e-01
1.10687244e+00 1.51458293e-01 -8.28369915e-01 -4.33748156e-01
-1.68160379e-01 9.31246728e-02 4.40650821e-01 5.39597809e-01
4.03357655e-01 -7.79574454e-01 -4.20235902e-01 4.33960631e-02
1.86589733e-01 -4.88745183e-01 -3.20615284e-02 3.31736654e-01
-5.19005060e-01 6.32516593e-02 -2.32577905e-01 -1.37152627e-01
-1.25931275e+00 5.10857880e-01 3.15011203e-01 -9.52376008e-01
-2.02135906e-01 1.17817760e-01 -5.03652692e-01 -2.08838597e-01
3.52622598e-01 1.99696481e-01 -5.89250565e-01 -2.90463101e-02
9.84610617e-01 3.97839934e-01 1.18717410e-01 1.43504113e-01
-2.59398967e-01 -2.66345382e-01 -1.34325460e-01 -4.85465109e-01
9.00027871e-01 1.88220337e-01 -1.71339750e-01 4.75289971e-01
5.66313803e-01 4.03034210e-01 -6.83765769e-01 -9.10260063e-03
5.54957330e-01 -1.39676154e-01 -1.02355853e-01 -9.59185719e-01
-6.72557950e-01 8.77439618e-01 4.59706306e-01 6.79363489e-01
9.00884628e-01 -1.53616145e-01 6.81509018e-01 2.11531937e-01
6.18791997e-01 -9.83841956e-01 4.64847684e-01 2.37912118e-01
3.69729638e-01 -9.06875789e-01 -1.22529663e-01 2.16142282e-01
-4.84575212e-01 1.23615265e+00 8.16848993e-01 4.03471500e-01
5.54168284e-01 7.05117941e-01 1.50099665e-01 4.28139884e-03
-1.03517663e+00 2.05719262e-01 -3.49079669e-01 8.46121848e-01
2.80757129e-01 5.55178285e-01 -2.20198497e-01 7.91051686e-01
-9.03759897e-02 -9.29515213e-02 1.08999312e+00 9.61033881e-01
-2.96575278e-01 -1.63990378e+00 -4.71226186e-01 3.57302666e-01
-6.07513070e-01 2.61151105e-01 -6.81454837e-01 4.99420434e-01
1.07799895e-01 1.26189244e+00 -3.89679730e-01 -9.26413476e-01
1.16649978e-01 2.13107556e-01 -7.37622231e-02 -6.74236059e-01
-9.45157111e-01 -3.48987520e-01 3.53537440e-01 -2.28777617e-01
9.76320058e-02 -5.07467449e-01 -8.34832370e-01 -5.29216707e-01
-3.60568047e-01 2.98769653e-01 1.03514373e+00 8.38262856e-01
2.19261989e-01 5.84176183e-01 1.10009944e+00 -3.40878308e-01
-9.34342980e-01 -1.26203990e+00 -3.86282712e-01 2.53810346e-01
-2.67781705e-01 -6.16952181e-01 -5.55301487e-01 -5.13459295e-02]
|
[5.474369525909424, 7.3453569412231445]
|
53a07257-1040-4f36-8a06-d3e2dd59039b
|
the-mixing-method-low-rank-coordinate-descent
|
1706.00476
| null |
http://arxiv.org/abs/1706.00476v3
|
http://arxiv.org/pdf/1706.00476v3.pdf
|
The Mixing method: low-rank coordinate descent for semidefinite programming with diagonal constraints
|
In this paper, we propose a low-rank coordinate descent approach to
structured semidefinite programming with diagonal constraints. The approach,
which we call the Mixing method, is extremely simple to implement, has no free
parameters, and typically attains an order of magnitude or better improvement
in optimization performance over the current state of the art. We show that the
method is strictly decreasing, converges to a critical point, and further that
for sufficient rank all non-optimal critical points are unstable. Moreover, we
prove that with a step size, the Mixing method converges to the global optimum
of the semidefinite program almost surely in a locally linear rate under random
initialization. This is the first low-rank semidefinite programming method that
has been shown to achieve a global optimum on the spherical manifold without
assumption. We apply our algorithm to two related domains: solving the maximum
cut semidefinite relaxation, and solving a maximum satisfiability relaxation
(we also briefly consider additional applications such as learning word
embeddings). In all settings, we demonstrate substantial improvement over the
existing state of the art along various dimensions, and in total, this work
expands the scope and scale of problems that can be solved using semidefinite
programming methods.
|
['Wei-Cheng Chang', 'Po-Wei Wang', 'J. Zico Kolter']
|
2017-06-01
| null | null | null | null |
['learning-word-embeddings']
|
['methodology']
|
[ 6.81096164e-04 2.37561926e-01 -4.87855226e-01 -1.48912311e-01
-1.17188466e+00 -8.99794579e-01 1.59406275e-01 8.53606611e-02
-3.06713909e-01 5.58265507e-01 2.82698154e-01 -4.82253581e-01
-5.27479470e-01 -2.94800907e-01 -7.47872293e-01 -8.79511952e-01
-2.48779118e-01 1.05773270e+00 -3.22270155e-01 -3.17535937e-01
1.36264414e-01 1.18811302e-01 -1.00044346e+00 -3.72999579e-01
1.03577721e+00 6.76910520e-01 -1.53881982e-01 4.91994828e-01
2.04884812e-01 1.22400217e-01 1.52149111e-01 -3.72054309e-01
5.88916540e-01 9.50200669e-03 -8.21805716e-01 3.74342114e-01
8.12204361e-01 -1.94940150e-01 -2.00902656e-01 1.27283549e+00
4.50957328e-01 1.26362681e-01 5.15945911e-01 -1.25715446e+00
-9.85055447e-01 4.34284419e-01 -8.72964621e-01 -8.11065435e-02
2.67391920e-01 -4.40213680e-01 1.70439923e+00 -1.41347623e+00
6.76426291e-01 1.18253112e+00 6.01375878e-01 3.56816739e-01
-1.45647109e+00 -2.20654190e-01 3.94088358e-01 -2.69287586e-01
-1.31362510e+00 -3.58824492e-01 6.80997610e-01 -5.31477213e-01
7.18058944e-01 3.37850720e-01 4.69120502e-01 3.43694240e-01
-3.94127607e-01 1.00986695e+00 1.03795862e+00 -3.77604365e-01
1.68526262e-01 1.74747229e-01 4.60583866e-01 1.10943615e+00
6.10944748e-01 -2.71925747e-01 -3.88349533e-01 -5.34833968e-01
4.83976454e-01 -4.05395180e-01 -2.42223978e-01 -9.76261854e-01
-1.33761644e+00 1.23472309e+00 -5.07639535e-02 3.92138921e-02
9.49469302e-03 1.39414012e-01 2.25287288e-01 2.49092624e-01
7.92023480e-01 3.81431639e-01 -2.89387763e-01 -2.22792223e-01
-9.77501929e-01 1.20941348e-01 1.22062874e+00 9.67004895e-01
5.43369651e-01 -5.61213642e-02 2.41124079e-01 7.07600415e-01
4.95298713e-01 9.05518830e-01 -2.22185835e-01 -8.63011003e-01
9.20762956e-01 3.21739763e-01 2.75800020e-01 -1.36893690e+00
-5.73975623e-01 -4.62556690e-01 -7.44208038e-01 -1.28078267e-01
4.01801348e-01 -3.92927319e-01 -2.69346923e-01 1.67126858e+00
3.31482649e-01 -5.48261218e-02 -8.61610994e-02 1.21857822e+00
3.50601941e-01 8.73708069e-01 -5.97626567e-01 -4.42064464e-01
1.05079269e+00 -1.16330731e+00 -7.95314610e-01 -4.21926528e-01
9.24355686e-01 -7.19947696e-01 1.08652282e+00 4.63438511e-01
-1.22182906e+00 2.56285042e-01 -1.18131816e+00 -1.19302019e-01
1.01064416e-02 4.07858253e-01 1.17665482e+00 7.79628813e-01
-1.36250687e+00 2.21960828e-01 -8.33504260e-01 -1.51220411e-01
-1.37841776e-01 6.75886452e-01 -5.60888588e-01 -9.30359289e-02
-9.99547482e-01 5.44337451e-01 5.57364821e-02 2.31022313e-01
-5.24363518e-01 -9.15287852e-01 -1.08655500e+00 -2.15752199e-01
4.40658510e-01 -5.37378073e-01 9.21843529e-01 -6.82129383e-01
-1.31850207e+00 1.14437377e+00 -4.70295519e-01 -2.17570662e-01
4.64962035e-01 -1.39679715e-01 5.59536032e-02 1.09776080e-01
1.61616743e-01 1.94317847e-01 5.78320622e-01 -1.17661703e+00
-7.98514113e-02 -4.59238321e-01 3.03614497e-01 4.48276490e-01
-8.03249300e-01 1.81633815e-01 -8.79455924e-01 -5.19185483e-01
3.16978484e-01 -1.44348347e+00 -5.34150600e-01 1.68021247e-01
-4.34748590e-01 -3.05167049e-01 3.54515672e-01 -4.05192673e-01
1.67557573e+00 -1.97370470e+00 6.51281178e-01 5.30981302e-01
4.22306091e-01 -7.82328378e-03 -1.73748001e-01 5.91960907e-01
-3.16943973e-01 1.88603967e-01 -5.03765106e-01 -6.16540909e-01
4.12081301e-01 1.83674768e-01 -2.03492880e-01 1.20491111e+00
-2.66014725e-01 6.54655397e-01 -9.77405071e-01 -2.38925606e-01
-9.83672291e-02 1.87313348e-01 -1.03472245e+00 -3.50859284e-01
7.00198188e-02 -1.32712781e-01 -4.00009543e-01 4.01033580e-01
9.08368170e-01 -5.53789198e-01 4.14485455e-01 -2.51749009e-01
-3.43365163e-01 6.83029369e-02 -1.79533434e+00 1.60891128e+00
-4.33419108e-01 7.38990605e-01 4.95636314e-01 -1.37479150e+00
6.07542038e-01 1.63796514e-01 7.92825103e-01 -2.00357378e-01
-1.41776174e-01 3.83253127e-01 -3.57589960e-01 -1.89930588e-01
4.78676617e-01 -5.52014261e-02 -1.57581583e-01 8.93523514e-01
-1.93022281e-01 -1.46075701e-02 5.30127883e-01 4.76907432e-01
7.68812895e-01 -6.49688065e-01 -1.29003301e-01 -7.37521887e-01
5.52313030e-01 -6.74687326e-02 3.74875784e-01 5.90033174e-01
7.88740963e-02 7.72560656e-01 9.42830563e-01 -2.15858757e-01
-1.17274201e+00 -1.18071985e+00 -4.68195975e-01 1.17066884e+00
1.34090498e-01 -5.31691313e-01 -7.15873301e-01 -4.31023479e-01
8.10281709e-02 2.24618152e-01 -5.85540593e-01 4.11828607e-02
-4.79864836e-01 -1.25793421e+00 2.67825395e-01 3.40399921e-01
3.21334660e-01 4.56712693e-02 4.76574481e-01 -2.48755720e-02
-4.47043888e-02 -1.07592678e+00 -1.07245111e+00 1.06207788e-01
-1.02919805e+00 -1.08509552e+00 -9.15020704e-01 -1.13013160e+00
1.23285937e+00 4.73450989e-01 1.11040926e+00 -2.59057760e-01
-1.29268274e-01 7.37222135e-01 6.69363812e-02 2.64512360e-01
2.56487846e-01 1.53694218e-02 5.35755217e-01 1.15582064e-01
2.79129334e-02 -1.84878632e-01 -5.32051623e-01 3.11278969e-01
-9.79697287e-01 -3.21193725e-01 1.80993423e-01 9.08465803e-01
8.89866471e-01 1.33548692e-01 -7.30639473e-02 -1.05696714e+00
1.10498893e+00 -5.87762654e-01 -9.72544074e-01 2.45130494e-01
-8.00280869e-01 5.12684882e-01 6.06650531e-01 -7.63926283e-02
-6.16556227e-01 2.88318217e-01 2.80700922e-01 -2.77632266e-01
8.28219891e-01 9.27353263e-01 -6.47294894e-02 -3.96974295e-01
5.79602659e-01 1.47087261e-01 5.56359477e-02 -3.64017725e-01
7.06559837e-01 5.24465263e-01 7.06488043e-02 -9.30265844e-01
1.23730719e+00 9.11530197e-01 2.51401484e-01 -8.61071050e-01
-1.13721812e+00 -8.72674406e-01 -7.29440987e-01 7.44426027e-02
5.10429084e-01 -9.97690856e-01 -7.00409591e-01 1.33509606e-01
-1.01855540e+00 -1.65068313e-01 1.23631485e-01 4.55222130e-01
-5.94508529e-01 6.84504628e-01 -4.21214879e-01 -8.11948478e-01
-2.08238393e-01 -9.91432488e-01 1.13487351e+00 -2.37744749e-01
1.25808463e-01 -1.65437472e+00 4.76951748e-01 4.61737871e-01
1.00035317e-01 8.10308754e-02 6.96809769e-01 -3.77555430e-01
-3.28022391e-01 -2.85655171e-01 -2.41610840e-01 4.43831086e-01
-1.60409719e-01 2.96300035e-02 -1.61319911e-01 -5.03791392e-01
-9.49274674e-02 -2.40592718e-01 7.99156308e-01 3.85877192e-01
6.17369950e-01 -6.47697926e-01 -1.39822781e-01 8.27522576e-01
1.50383794e+00 -4.01538491e-01 2.27653414e-01 2.81186998e-01
7.38860011e-01 4.40266877e-01 8.21997881e-01 6.19316339e-01
6.72099054e-01 6.73769236e-01 3.42859209e-01 -2.60695755e-01
4.15662676e-01 -3.43138911e-02 4.64927614e-01 1.16454077e+00
1.83818284e-02 1.92715891e-03 -1.01565135e+00 7.45425045e-01
-2.21266723e+00 -8.52735102e-01 -3.96902651e-01 2.32114291e+00
9.26763237e-01 -1.74124286e-01 2.26950601e-01 -7.93160275e-02
7.26959467e-01 3.85606349e-01 -6.19106174e-01 -6.92552090e-01
-3.69066656e-01 1.27523810e-01 1.12114346e+00 9.36361015e-01
-1.45281041e+00 1.00255799e+00 7.32731104e+00 6.26301944e-01
-7.95494795e-01 7.27504343e-02 4.45697486e-01 -2.09363580e-01
-7.07736790e-01 -1.16314374e-01 -8.87097836e-01 1.38248324e-01
4.06887233e-01 -5.07857442e-01 7.33914137e-01 1.06277072e+00
4.97028381e-01 1.81542546e-01 -1.12294114e+00 1.30046833e+00
4.43679750e-01 -1.24091220e+00 -2.56426632e-01 2.91572392e-01
1.36144650e+00 8.43850747e-02 6.78685546e-01 -2.61662722e-01
4.98296380e-01 -9.35807049e-01 3.78941983e-01 -5.56677347e-03
7.64847934e-01 -9.41864371e-01 3.84277016e-01 2.52604876e-02
-1.13875198e+00 1.52816519e-01 -5.07777333e-01 6.05654940e-02
3.48988295e-01 1.06640124e+00 -2.99732715e-01 6.75380826e-02
2.58453310e-01 8.68345141e-01 -2.52653301e-01 1.16168237e+00
-2.86773682e-01 6.39603436e-01 -8.55391920e-01 -1.44753054e-01
6.25614762e-01 -8.79228771e-01 7.67409325e-01 1.17472970e+00
8.74862596e-02 -2.94284895e-02 3.00515056e-01 6.09076560e-01
-1.11647882e-01 4.02515650e-01 -6.30108476e-01 -3.32913339e-01
1.01113409e-01 1.07491660e+00 -3.69106114e-01 1.24459304e-01
-4.74774688e-01 1.08412218e+00 5.08470237e-01 4.51648533e-01
-8.76383185e-01 -4.52573746e-01 9.81134415e-01 -1.31240790e-03
3.02775681e-01 -8.70013177e-01 -4.31321561e-01 -1.51696718e+00
4.99003470e-01 -7.55340219e-01 4.25731152e-01 -2.06528112e-01
-1.20896041e+00 1.79613531e-01 -1.22168377e-01 -1.05503130e+00
-9.67503488e-02 -9.21259880e-01 -3.84391457e-01 5.73981702e-01
-1.24224138e+00 -6.60056055e-01 3.34914535e-01 6.88657582e-01
9.51658338e-02 9.51881185e-02 5.71010232e-01 6.04986429e-01
-7.13301897e-01 7.83953488e-01 7.85439372e-01 1.34971023e-01
5.68618953e-01 -1.64757025e+00 2.24936157e-04 9.20788467e-01
1.56289846e-01 8.58162224e-01 7.39953041e-01 -3.21789026e-01
-2.25175023e+00 -7.85950959e-01 1.02984989e+00 -2.49313921e-01
1.51755238e+00 -4.16352063e-01 -2.00668693e-01 7.72544444e-01
-1.04369238e-01 1.38010934e-01 1.00064766e+00 4.72673446e-01
-3.95516664e-01 5.30231409e-02 -8.59718621e-01 7.56483436e-01
1.07571924e+00 -6.05943739e-01 -2.75641054e-01 1.12546170e+00
4.52252865e-01 -4.94415402e-01 -8.09463799e-01 2.03952286e-03
4.01579052e-01 -4.36313093e-01 1.18107748e+00 -8.18821669e-01
3.15098107e-01 -3.69604796e-01 -2.17500031e-01 -1.26855266e+00
-3.83092612e-01 -1.11499000e+00 -2.71515220e-01 8.64698470e-01
8.64917278e-01 -6.21353567e-01 9.47162330e-01 9.20356393e-01
-5.60341179e-02 -9.64668930e-01 -1.03689051e+00 -9.87473607e-01
2.42455438e-01 -4.80155289e-01 -2.19358876e-01 9.96980131e-01
3.65380079e-01 3.10211509e-01 -6.56228662e-01 4.06088233e-01
9.92496729e-01 2.79566765e-01 5.02314866e-01 -9.43239570e-01
-3.22035491e-01 -2.83553034e-01 -3.25223565e-01 -1.84116864e+00
4.88559604e-01 -1.34120619e+00 -1.94440663e-01 -1.57220495e+00
3.93657416e-01 -8.05087447e-01 -1.30193131e-02 2.18282849e-01
4.02505808e-02 3.66005361e-01 5.93242608e-02 1.18894540e-01
-1.05814087e+00 4.61195648e-01 1.30677927e+00 -2.95978487e-01
-2.55916297e-01 1.31055992e-02 -1.09063888e+00 6.64804578e-01
5.72410405e-01 -3.54022771e-01 -4.15847391e-01 -7.85495937e-01
1.04771924e+00 -2.07870215e-01 -1.23239905e-01 -4.43226993e-01
4.78477567e-01 -1.01410381e-01 -4.98645335e-01 -3.18424374e-01
2.10926116e-01 -6.83946609e-01 -2.86270857e-01 1.18715107e-01
-3.45211655e-01 1.28541842e-01 -4.13926169e-02 6.26134217e-01
-1.79465622e-01 -4.25186545e-01 6.86065018e-01 3.45867991e-01
-2.81061947e-01 7.16796100e-01 -2.95957983e-01 6.42530084e-01
9.81650472e-01 -2.54963096e-02 -5.08932546e-02 -5.90316534e-01
-7.76283026e-01 6.09164417e-01 3.28754008e-01 2.86852986e-01
4.86937433e-01 -1.70877063e+00 -6.25451624e-01 -1.84329316e-01
4.70031723e-02 -1.45314857e-01 -1.33730933e-01 1.27436411e+00
-6.91078782e-01 8.78366590e-01 7.45558679e-01 -4.84273076e-01
-1.18873274e+00 5.01358807e-01 1.54558033e-01 -4.38942790e-01
-1.78598717e-01 9.52377498e-01 1.51404828e-01 -6.31769657e-01
3.15123320e-01 -7.54449964e-02 8.09070915e-02 2.98819005e-01
3.62502545e-01 5.96108258e-01 -6.05977066e-02 -8.07873309e-01
-4.99016285e-01 9.86516178e-01 3.69119570e-02 -2.45216027e-01
1.53775585e+00 -1.65978849e-01 -6.74862862e-01 2.02180311e-01
1.89453685e+00 3.39410841e-01 -9.72043455e-01 -1.93739086e-01
-2.41443738e-01 -3.98809731e-01 3.71426910e-01 -7.79390782e-02
-1.32879257e+00 6.24137938e-01 2.01548353e-01 2.74525076e-01
7.12725580e-01 2.59208620e-01 7.43251979e-01 7.92202711e-01
2.79788941e-01 -1.31961226e+00 7.74446800e-02 8.43655825e-01
7.99868703e-01 -1.30652416e+00 1.49108872e-01 -7.49497890e-01
-5.95185697e-01 1.26414776e+00 1.01675875e-01 -5.24153471e-01
1.04454625e+00 3.37596178e-01 -3.73934448e-01 -2.46588618e-01
-5.69444954e-01 -1.65093213e-01 5.54944098e-01 4.28016990e-01
3.36839437e-01 1.33491546e-01 -6.49532259e-01 3.87578428e-01
-3.43802571e-01 -5.61003447e-01 4.49870676e-01 4.18819934e-01
-4.30643409e-01 -1.14817441e+00 -3.67294878e-01 3.15411627e-01
-3.46582502e-01 -5.16255736e-01 -4.51202840e-01 4.36972082e-01
-5.75927079e-01 9.75141525e-01 -1.52462849e-03 -4.94583547e-02
1.98389754e-01 -4.82716173e-01 5.19648492e-01 -6.10312402e-01
4.00172807e-02 -1.65057167e-01 1.78354636e-01 -5.77394009e-01
-2.83493847e-01 -8.31305981e-01 -1.14212048e+00 -5.32505095e-01
-4.36548650e-01 2.59339988e-01 6.59447312e-01 6.62258267e-01
1.78377539e-01 -2.08949581e-01 9.20366764e-01 -5.20507932e-01
-7.18590379e-01 -1.83676124e-01 -6.27723455e-01 6.37132525e-02
3.50921363e-01 -4.69240665e-01 -9.09520268e-01 -2.26453826e-01]
|
[7.046410083770752, 4.560855388641357]
|
5225ee29-bad1-4541-a930-4a5ad27195bd
|
predicting-target-language-ccg-supertags
|
1702.01147
| null |
http://arxiv.org/abs/1702.01147v2
|
http://arxiv.org/pdf/1702.01147v2.pdf
|
Predicting Target Language CCG Supertags Improves Neural Machine Translation
|
Neural machine translation (NMT) models are able to partially learn syntactic
information from sequential lexical information. Still, some complex syntactic
phenomena such as prepositional phrase attachment are poorly modeled. This work
aims to answer two questions: 1) Does explicitly modeling target language
syntax help NMT? 2) Is tight integration of words and syntax better than
multitask training? We introduce syntactic information in the form of CCG
supertags in the decoder, by interleaving the target supertags with the word
sequence. Our results on WMT data show that explicitly modeling target-syntax
improves machine translation quality for German->English, a high-resource pair,
and for Romanian->English, a low-resource pair and also several syntactic
phenomena including prepositional phrase attachment. Furthermore, a tight
coupling of words and syntax improves translation quality more than multitask
training. By combining target-syntax with adding source-side dependency labels
in the embedding layer, we obtain a total improvement of 0.9 BLEU for
German->English and 1.2 BLEU for Romanian->English.
|
['Marcin Junczys-Dowmunt', 'Rico Sennrich', 'Alexandra Birch', 'Maria Nadejde', 'Tomasz Dwojak', 'Siva Reddy', 'Philipp Koehn']
|
2017-02-03
|
predicting-target-language-ccg-supertags-1
|
https://aclanthology.org/W17-4707
|
https://aclanthology.org/W17-4707.pdf
|
ws-2017-9
|
['prepositional-phrase-attachment']
|
['natural-language-processing']
|
[-2.02945635e-01 2.21245992e-03 -6.51929021e-01 -5.41157484e-01
-1.07150209e+00 -6.58986568e-01 4.72631514e-01 1.55555353e-01
-7.78748989e-01 9.93856370e-01 5.35018623e-01 -7.67358184e-01
2.52251416e-01 -5.69718063e-01 -9.12827313e-01 -4.54296768e-01
3.84401344e-02 8.86332333e-01 -9.61755216e-02 -5.30565739e-01
-1.48928121e-01 -1.10622443e-01 -5.38903654e-01 6.86231017e-01
1.03604746e+00 4.71395522e-01 5.18096328e-01 7.41882175e-02
-5.24826229e-01 2.38413557e-01 -2.96902746e-01 -7.52183259e-01
2.52258599e-01 -4.06870663e-01 -8.44486952e-01 -4.09362882e-01
2.13227913e-01 1.48310900e-01 2.68635631e-01 8.87943447e-01
4.27138537e-01 -2.53708869e-01 5.35760880e-01 -5.32129049e-01
-7.90938258e-01 1.31900930e+00 -5.59844911e-01 2.59931117e-01
1.08775064e-01 -1.70283727e-02 1.63702822e+00 -1.21302199e+00
7.97255397e-01 1.39888418e+00 6.55267954e-01 6.37457788e-01
-1.28910649e+00 -6.71122909e-01 8.99474844e-02 1.41364345e-02
-1.10844576e+00 -3.09790194e-01 5.79459071e-01 -3.20129186e-01
1.57945728e+00 -2.32276972e-02 2.17279688e-01 1.32830644e+00
6.19159997e-01 8.39122951e-01 1.24442995e+00 -5.84799707e-01
-3.48902285e-01 5.93077466e-02 3.05151850e-01 5.69587827e-01
1.42783150e-01 1.35135978e-01 -6.09826863e-01 1.38077587e-01
4.43321437e-01 -4.45050597e-01 1.44320741e-01 2.66497284e-01
-1.65095043e+00 9.30717945e-01 2.30091512e-01 7.63684928e-01
-3.29744279e-01 2.07631826e-01 5.32314003e-01 6.98563159e-01
6.43306673e-01 5.23842394e-01 -1.12481248e+00 -2.00063035e-01
-6.97718263e-01 -1.57963142e-01 7.55541682e-01 1.09114206e+00
1.00034833e+00 5.74967936e-02 -1.27268672e-01 1.14115059e+00
1.86966181e-01 8.96952510e-01 6.64885521e-01 -3.93670827e-01
1.18286467e+00 4.17592376e-01 2.33726460e-04 -2.12623775e-01
-4.87648487e-01 -5.03464878e-01 -5.10167122e-01 -5.40099978e-01
6.41081154e-01 -6.04965210e-01 -8.28710079e-01 2.13732433e+00
-7.18419328e-02 -5.00010133e-01 3.16831410e-01 7.75992393e-01
6.84327424e-01 9.42512691e-01 2.29468539e-01 -3.56170565e-01
1.75803876e+00 -1.12064385e+00 -9.22982156e-01 -8.11794400e-01
1.19508934e+00 -1.23892951e+00 1.40861905e+00 -1.42794520e-01
-1.27006269e+00 -5.01786351e-01 -5.24344087e-01 -3.11272293e-01
-4.23108429e-01 3.17060977e-01 5.79177380e-01 4.77577001e-01
-9.75946128e-01 5.06156325e-01 -7.01850772e-01 -4.14075315e-01
-1.35450363e-01 5.75731516e-01 -4.85928297e-01 -3.09138268e-01
-1.70744586e+00 1.36226702e+00 3.92835885e-01 -1.85897008e-01
-3.01341951e-01 -6.49600744e-01 -9.48128998e-01 1.81346890e-02
1.42180145e-01 -7.38349736e-01 1.19309103e+00 -9.95730937e-01
-1.47999692e+00 8.80736768e-01 -5.10412931e-01 -3.01177084e-01
6.04643598e-02 -2.77274668e-01 -2.45734960e-01 -4.11439240e-01
2.33738199e-01 5.56958556e-01 3.64837140e-01 -8.17479730e-01
-6.62311852e-01 -5.03868043e-01 -2.43840620e-01 3.72817039e-01
-4.29011226e-01 6.21187806e-01 -2.24020079e-01 -7.44780302e-01
-1.88820176e-02 -1.08896828e+00 -3.48099470e-02 -7.08295166e-01
-1.57522634e-01 -5.01143456e-01 2.97548831e-01 -9.20629621e-01
1.26934612e+00 -1.99892402e+00 4.28431392e-01 -1.24898218e-01
-3.14143956e-01 1.59206167e-01 -4.35547918e-01 5.93712747e-01
2.31832433e-02 3.73710454e-01 3.18209343e-02 -4.67093229e-01
6.88740984e-02 7.25375473e-01 1.48608871e-02 9.89242829e-03
4.47628856e-01 1.43101072e+00 -8.44425738e-01 -2.91071773e-01
-1.62821591e-01 2.01445967e-01 -3.16742480e-01 -1.85343221e-01
-3.74449492e-01 3.21895987e-01 -3.78928602e-01 5.64888418e-01
3.37461054e-01 -4.20304649e-02 4.08311963e-01 -2.47197580e-02
-2.32333422e-01 1.23722911e+00 -4.57806349e-01 1.71917260e+00
-8.81045699e-01 2.72984773e-01 6.71117604e-02 -7.66053915e-01
8.12497675e-01 4.65580434e-01 1.59697801e-01 -9.15642381e-01
2.15323031e-01 8.18584800e-01 7.50286639e-01 -3.80831361e-01
2.60944754e-01 -6.13281548e-01 -3.67973745e-01 4.45105970e-01
2.71754026e-01 4.67033423e-02 1.45340592e-01 -2.40656123e-01
1.06600952e+00 7.66121745e-02 1.60052896e-01 -7.11347759e-01
1.24382980e-01 1.01187177e-01 8.53088379e-01 3.17323357e-01
2.59610742e-01 3.57601702e-01 4.15123254e-01 -4.99964684e-01
-1.16448081e+00 -9.10235167e-01 -1.18806206e-01 1.47992957e+00
-2.03865021e-01 -6.04297400e-01 -6.32542431e-01 -8.13876331e-01
-9.39665288e-02 9.28725362e-01 -5.08944452e-01 5.19766510e-02
-1.29705524e+00 -1.15497398e+00 4.99363124e-01 5.33726752e-01
3.05583656e-01 -1.05121660e+00 1.06760778e-01 5.26563168e-01
-7.73350954e-01 -1.34955001e+00 -7.86789536e-01 6.49151266e-01
-1.12770283e+00 -3.04521173e-01 -6.62002027e-01 -1.10829425e+00
2.45573729e-01 -4.66086343e-02 1.35990429e+00 -1.15270905e-01
4.36882973e-01 -4.60137039e-01 -4.84058410e-01 -2.82763124e-01
-5.74850023e-01 5.79804361e-01 8.27101320e-02 -3.52618486e-01
6.48506880e-01 -4.10706103e-01 -1.08056702e-02 3.79252404e-01
-4.44390774e-01 2.19092324e-01 8.24621618e-01 1.00066137e+00
3.45850796e-01 -6.35380387e-01 3.60992461e-01 -1.03249490e+00
4.21737880e-01 -3.27445298e-01 -2.59153664e-01 3.36851448e-01
-5.04242301e-01 4.95713115e-01 6.81112349e-01 -6.18306756e-01
-9.24128413e-01 -1.38541803e-01 -4.21314478e-01 2.06690401e-01
1.91865250e-01 8.24563265e-01 -3.51163059e-01 4.61434305e-01
6.50378108e-01 3.95242162e-02 -2.24009126e-01 -7.58721828e-01
2.07416058e-01 4.71883446e-01 -8.41611475e-02 -1.02447832e+00
7.21552074e-01 -2.08948538e-01 -2.70192087e-01 -5.93884170e-01
-7.02219486e-01 -2.64840811e-01 -7.08436131e-01 4.19503987e-01
1.04308033e+00 -9.50361907e-01 -8.92193317e-02 2.73444474e-01
-1.56207716e+00 -6.63472354e-01 -5.82349626e-03 8.86337876e-01
-3.85371238e-01 1.69239417e-01 -1.11652696e+00 -1.22622773e-01
-4.60817844e-01 -1.35936475e+00 1.09635508e+00 -4.46652442e-01
-3.88957590e-01 -1.26758492e+00 -1.01384670e-01 4.38050270e-01
4.45146143e-01 -3.86484355e-01 1.57018507e+00 -9.29079771e-01
-3.28851372e-01 1.17178813e-01 -2.41543442e-01 3.24097395e-01
2.53362417e-01 -4.46184188e-01 -3.87466013e-01 -1.88986510e-01
-1.12606034e-01 -3.42287906e-02 8.04801822e-01 3.90348822e-01
1.62628833e-02 -4.55304235e-01 -1.78043395e-01 4.25888687e-01
1.28908384e+00 1.83308106e-02 4.85117763e-01 3.21736932e-01
7.00384438e-01 6.82487130e-01 5.43543696e-01 -1.62786737e-01
6.89818501e-01 8.50758195e-01 3.50003317e-02 -5.77443428e-02
-3.42230320e-01 -1.73668534e-01 9.83151615e-01 1.54429126e+00
-9.97206569e-02 -4.12225455e-01 -1.20045233e+00 5.88315010e-01
-1.71312845e+00 -3.91529322e-01 -5.33498585e-01 1.93295097e+00
1.32659173e+00 1.96495578e-01 -1.53534144e-01 -4.47671026e-01
5.15255392e-01 -4.90355380e-02 9.18148607e-02 -6.78567708e-01
-3.28701705e-01 4.26734090e-01 6.60759747e-01 8.90380740e-01
-7.41130650e-01 1.75050533e+00 5.54493952e+00 8.69429052e-01
-1.12334597e+00 5.39413989e-01 1.47447228e-01 1.06475383e-01
-6.24392211e-01 2.76440412e-01 -1.22753572e+00 4.70683843e-01
1.20329678e+00 9.84717757e-02 3.57898861e-01 3.60769451e-01
1.88131824e-01 1.87141150e-01 -1.32166779e+00 4.88078564e-01
-1.33248135e-01 -1.20668924e+00 2.21623957e-01 3.43143702e-01
7.04695702e-01 5.37360191e-01 -1.50288194e-01 5.71789086e-01
5.12253046e-01 -7.87316561e-01 7.39968657e-01 -1.26204759e-01
7.28497684e-01 -4.99054432e-01 1.01666367e+00 5.76070368e-01
-8.98857713e-01 1.89133972e-01 -5.15069664e-01 -1.28162175e-01
2.85391778e-01 6.11452997e-01 -7.48244286e-01 6.91240609e-01
3.48178387e-01 5.99440455e-01 -4.14428234e-01 5.28432012e-01
-7.34103501e-01 9.27558303e-01 -3.66479158e-01 -2.26484850e-01
5.75413167e-01 -2.31673777e-01 5.89964509e-01 1.65040565e+00
4.70003247e-01 -1.81927696e-01 1.30958363e-01 3.91185492e-01
-2.35647082e-01 6.36650562e-01 -5.49568892e-01 -1.03154451e-01
2.13030532e-01 9.21478212e-01 -4.72502083e-01 -3.41010898e-01
-6.58805370e-01 9.52732742e-01 5.32913327e-01 3.15260112e-01
-7.29555964e-01 -6.20902292e-02 8.41286778e-01 1.85828447e-01
2.45880365e-01 -6.65306926e-01 -5.52583873e-01 -1.41916215e+00
1.72811851e-01 -9.30006266e-01 2.03936473e-01 -5.35773039e-01
-1.24273741e+00 7.85853744e-01 -2.05521315e-01 -8.50640118e-01
-3.80645692e-01 -7.67246902e-01 -3.96871597e-01 1.21119952e+00
-1.68830991e+00 -1.45354068e+00 6.11116230e-01 2.84501970e-01
7.70342708e-01 -1.56528696e-01 1.13421381e+00 5.01920819e-01
-5.74489057e-01 8.37356210e-01 1.81035087e-01 4.37473238e-01
1.05156648e+00 -1.21186101e+00 7.08852947e-01 7.71239698e-01
3.99828672e-01 8.42485309e-01 6.19902730e-01 -6.23455405e-01
-1.39332402e+00 -1.00719607e+00 2.05534673e+00 -6.41774833e-01
8.74465108e-01 -7.09458888e-01 -9.08362091e-01 8.35504353e-01
6.30024254e-01 -3.19954723e-01 7.33640909e-01 6.53847754e-01
-5.28863192e-01 7.23006427e-02 -6.99965596e-01 6.85287833e-01
9.74096477e-01 -4.85627115e-01 -9.54514742e-01 5.06941259e-01
1.02254498e+00 -2.43321121e-01 -7.58402765e-01 4.78871495e-01
3.05593073e-01 -1.94008499e-01 6.31679952e-01 -8.03023517e-01
6.13429427e-01 4.26154956e-02 -4.16139275e-01 -1.67864025e+00
-3.62388581e-01 -6.38280809e-01 3.40747982e-01 1.07524180e+00
1.27861297e+00 -5.82384109e-01 4.68444109e-01 1.96003333e-01
-5.00479400e-01 -6.14982665e-01 -1.15748751e+00 -1.04462421e+00
8.32847595e-01 -3.27050209e-01 2.35856026e-01 1.02551508e+00
2.56808728e-01 1.11171865e+00 -3.65323991e-01 -6.36437014e-02
2.49658942e-01 -1.53259952e-02 2.95931220e-01 -1.20092309e+00
-3.65151286e-01 -2.77864158e-01 1.14329316e-01 -1.12032270e+00
4.50271666e-01 -1.33641875e+00 5.01397997e-02 -1.67429566e+00
3.37049700e-02 -3.78507465e-01 -1.75536394e-01 9.18659508e-01
-2.43885085e-01 2.75587231e-01 4.56442460e-02 1.96006358e-01
-1.05290264e-01 3.91436905e-01 1.24519348e+00 -2.09022295e-02
-4.55674008e-02 -1.26201347e-01 -4.98742491e-01 5.46705186e-01
8.54985356e-01 -8.49886239e-01 9.70394537e-02 -1.23092210e+00
5.82762599e-01 2.54926205e-01 -3.04734915e-01 -4.05345738e-01
-2.75926199e-02 -2.37930626e-01 1.31258080e-02 -1.91000104e-01
3.67838889e-01 -6.63967371e-01 -2.33033836e-01 3.69097054e-01
-1.59950390e-01 4.92952764e-01 3.86833191e-01 -1.81724361e-04
-1.98255435e-01 -2.76590407e-01 6.01227701e-01 -3.48971128e-01
-3.12995493e-01 2.25573890e-02 -6.93346024e-01 2.70032972e-01
4.74520802e-01 5.91621734e-02 -2.30481640e-01 1.59877706e-02
-7.03203619e-01 2.74561882e-01 2.17150316e-01 7.48349190e-01
1.18043296e-01 -1.35330462e+00 -1.04124010e+00 2.19373614e-01
4.91939336e-02 -4.55815166e-01 -3.24649304e-01 1.04902768e+00
-1.33689180e-01 5.61083794e-01 -9.18876752e-02 -3.14826429e-01
-1.16859722e+00 2.32492685e-01 5.33429496e-02 -5.02155304e-01
-2.54486114e-01 9.54010367e-01 1.46508053e-01 -8.57712090e-01
-8.76343325e-02 -7.34222829e-01 1.86776534e-01 2.26459295e-01
2.08808435e-03 -5.55719733e-02 3.97908568e-01 -7.76359916e-01
-4.40782011e-01 5.79807937e-01 -3.54114711e-01 -2.64987797e-01
1.28534102e+00 -9.27110761e-02 -4.45444465e-01 4.59223360e-01
1.31854820e+00 3.47341806e-01 -5.19451916e-01 -5.54006398e-01
5.81395030e-01 2.00866107e-02 -1.73768893e-01 -1.11010730e+00
-5.71974754e-01 1.19222999e+00 1.14716273e-02 -3.65285099e-01
6.23120427e-01 1.31036520e-01 1.07594204e+00 6.12364471e-01
6.34929955e-01 -1.05751669e+00 -1.52694628e-01 1.28485847e+00
6.03964031e-01 -1.19203341e+00 -6.25946522e-01 -4.62746501e-01
-6.40712976e-01 9.88076329e-01 3.58343512e-01 1.00691125e-01
5.30124366e-01 3.54759276e-01 3.13343227e-01 1.06735922e-01
-8.83918822e-01 -1.96369380e-01 2.58861482e-01 2.67851632e-02
9.70584750e-01 2.15245292e-01 -9.17671978e-01 6.85575068e-01
-3.80017579e-01 -3.82608354e-01 2.31684163e-01 5.57879627e-01
-4.40970093e-01 -1.86579704e+00 -3.75716612e-02 2.71337330e-01
-8.72944474e-01 -8.56860459e-01 -4.69931453e-01 8.68426621e-01
1.24782972e-01 7.96264768e-01 -1.19898897e-02 -4.56176549e-01
2.65786320e-01 6.05270684e-01 4.50867325e-01 -1.18390954e+00
-9.79199409e-01 5.03264904e-01 4.78045136e-01 -2.35882491e-01
-1.30622849e-01 -7.29286373e-01 -1.27173102e+00 -2.18543693e-01
-4.24116850e-01 5.48262000e-01 8.58043492e-01 1.11342072e+00
2.37106562e-01 2.16328189e-01 3.74882430e-01 -3.73496801e-01
-6.11153662e-01 -1.33108222e+00 -1.82564646e-01 2.69925874e-02
1.76377684e-01 -3.85940939e-01 -2.94188768e-01 -1.29278004e-01]
|
[11.49057674407959, 10.240463256835938]
|
9c43a316-79e9-427d-a2b4-eb4dbf6de84a
|
inpars-toolkit-a-unified-and-reproducible
|
2307.04601
| null |
https://arxiv.org/abs/2307.04601v1
|
https://arxiv.org/pdf/2307.04601v1.pdf
|
InPars Toolkit: A Unified and Reproducible Synthetic Data Generation Pipeline for Neural Information Retrieval
|
Recent work has explored Large Language Models (LLMs) to overcome the lack of training data for Information Retrieval (IR) tasks. The generalization abilities of these models have enabled the creation of synthetic in-domain data by providing instructions and a few examples on a prompt. InPars and Promptagator have pioneered this approach and both methods have demonstrated the potential of using LLMs as synthetic data generators for IR tasks. This makes them an attractive solution for IR tasks that suffer from a lack of annotated data. However, the reproducibility of these methods was limited, because InPars' training scripts are based on TPUs -- which are not widely accessible -- and because the code for Promptagator was not released and its proprietary LLM is not publicly accessible. To fully realize the potential of these methods and make their impact more widespread in the research community, the resources need to be accessible and easy to reproduce by researchers and practitioners. Our main contribution is a unified toolkit for end-to-end reproducible synthetic data generation research, which includes generation, filtering, training and evaluation. Additionally, we provide an interface to IR libraries widely used by the community and support for GPU. Our toolkit not only reproduces the InPars method and partially reproduces Promptagator, but also provides a plug-and-play functionality allowing the use of different LLMs, exploring filtering methods and finetuning various reranker models on the generated data. We also made available all the synthetic data generated in this work for the 18 different datasets in the BEIR benchmark which took more than 2,000 GPU hours to be generated as well as the reranker models finetuned on the synthetic data. Code and data are available at https://github.com/zetaalphavector/InPars
|
['Rodrigo Nogueira', 'Jakub Zavrel', 'Roberto Lotufo', 'Vitor Jeronymo', 'Luiz Bonifacio', 'Hugo Abonizio']
|
2023-07-10
| null | null | null | null |
['synthetic-data-generation', 'retrieval', 'synthetic-data-generation', 'information-retrieval']
|
['medical', 'methodology', 'miscellaneous', 'natural-language-processing']
|
[-7.03626648e-02 -1.01089939e-01 -1.62246808e-01 -3.16681713e-01
-1.45972896e+00 -8.15343678e-01 8.78704071e-01 6.85577467e-02
-4.60876912e-01 8.02040756e-01 3.19319040e-01 -2.60679156e-01
-1.26086641e-02 -7.08619416e-01 -5.73175550e-01 -4.25997466e-01
1.38343066e-01 9.89445627e-01 6.47980496e-02 -6.26170099e-01
2.61808753e-01 2.97545612e-01 -1.90197754e+00 8.45359981e-01
7.80419767e-01 6.36488318e-01 4.28139716e-01 8.76099825e-01
-1.78409413e-01 3.85627061e-01 -8.36991429e-01 -8.93830508e-02
4.47250575e-01 -2.29333952e-01 -7.80594170e-01 -5.46834886e-01
4.82424647e-01 -2.41451606e-01 -1.79703727e-01 6.12016380e-01
9.90800261e-01 1.47847563e-01 5.22657931e-01 -1.12246096e+00
-6.34312809e-01 6.09973907e-01 -3.23367506e-01 -1.36247173e-01
5.59233487e-01 1.47511944e-01 9.92757857e-01 -1.03218329e+00
9.35488820e-01 1.21931350e+00 3.53128761e-01 6.45621121e-01
-1.20360553e+00 -5.67855358e-01 -5.62343419e-01 -1.41179204e-01
-1.37086654e+00 -4.77288932e-01 1.94313258e-01 -2.30496675e-01
9.77819324e-01 6.55476928e-01 2.03649417e-01 1.42807150e+00
-2.98196316e-01 9.00101662e-01 1.12078547e+00 -7.82232761e-01
8.10016915e-02 3.79766345e-01 2.32637540e-01 4.42115843e-01
4.90647629e-02 1.91035941e-01 -3.58127594e-01 -5.68903089e-01
5.39534390e-01 -4.63096231e-01 -1.34911165e-01 6.26840889e-02
-1.52991390e+00 8.76708090e-01 1.40165016e-01 4.25815940e-01
-2.02639610e-01 -9.95658487e-02 5.47613502e-01 6.11278534e-01
6.21583223e-01 6.66021049e-01 -6.77607596e-01 -3.05085629e-01
-9.35181677e-01 6.75530195e-01 9.08499599e-01 1.03934193e+00
5.50854981e-01 -1.77005827e-01 -6.56239390e-01 1.16488099e+00
1.18776793e-02 5.55306435e-01 7.45755136e-01 -1.02818882e+00
4.35520291e-01 5.16352236e-01 4.10261184e-01 -5.75312912e-01
-4.54245508e-01 -2.24486217e-01 -7.29523301e-01 1.96250856e-01
5.67299068e-01 -1.35873169e-01 -8.31990600e-01 1.45947671e+00
2.91777611e-01 -3.82195204e-01 1.79650560e-01 8.83037746e-01
1.16167605e+00 7.54031658e-01 3.86108533e-02 2.43399828e-03
1.52619064e+00 -1.02968144e+00 -2.98105776e-01 1.84347406e-01
1.15731740e+00 -1.26863933e+00 1.64103138e+00 4.92416054e-01
-1.10584116e+00 -6.26827240e-01 -8.31695795e-01 -2.99722522e-01
-6.59316957e-01 2.94819802e-01 7.18783081e-01 5.14511824e-01
-1.30955923e+00 4.88653511e-01 -5.41059136e-01 -4.63153362e-01
7.63906687e-02 2.63253689e-01 -3.08191866e-01 -2.18914464e-01
-1.56570995e+00 8.15886021e-01 3.83737981e-01 -3.47088128e-01
-5.32885015e-01 -8.17627430e-01 -5.90919614e-01 -1.85298756e-01
1.11293204e-01 -7.92795360e-01 1.54493713e+00 -7.96819508e-01
-1.48160315e+00 7.76596785e-01 1.42322600e-01 -2.79396415e-01
7.94361174e-01 -1.03030585e-01 -2.94707417e-01 -3.72401029e-02
-4.82982099e-02 8.70453656e-01 3.84766459e-01 -9.70589221e-01
-1.62541658e-01 -8.88998881e-02 -6.41055554e-02 3.55375409e-01
-1.65780589e-01 5.34420013e-01 -5.37559211e-01 -7.97352254e-01
-6.18426859e-01 -9.66223657e-01 -1.91905051e-01 -3.92535061e-01
-3.78264010e-01 -3.75434428e-01 4.70517755e-01 -5.30272007e-01
1.21585882e+00 -1.86085701e+00 -2.66058385e-01 9.20594856e-02
-2.69792736e-01 6.70691788e-01 -6.85172558e-01 9.46608424e-01
-1.60089165e-01 3.61162961e-01 1.84994236e-01 -1.89428329e-01
1.42145157e-01 -1.55184135e-01 -5.22747278e-01 5.61159104e-02
-6.86238483e-02 7.99648821e-01 -9.66828227e-01 -3.24239492e-01
1.30079269e-01 6.12824559e-01 -4.34906393e-01 3.68586868e-01
-7.37829089e-01 3.98076475e-01 -3.82325560e-01 3.94324154e-01
6.76927209e-01 -2.78968990e-01 -1.78319588e-02 4.95017730e-02
-3.25664610e-01 5.32947421e-01 -1.32563055e+00 1.76110041e+00
-5.89882553e-01 4.03782427e-01 -2.90852636e-01 -4.26881611e-01
8.25333118e-01 5.20078003e-01 3.34539950e-01 -7.93233454e-01
-9.85671356e-02 3.70812625e-01 -3.78045976e-01 -1.87802151e-01
9.38668847e-01 2.05134943e-01 -1.29597351e-01 8.91234934e-01
-1.37754411e-01 -4.18183714e-01 9.18900013e-01 4.85392958e-01
1.10966098e+00 2.59437472e-01 -5.36207221e-02 -1.67489976e-01
3.28808755e-01 4.20994461e-01 -6.31128922e-02 1.00126851e+00
6.20736480e-01 9.04413164e-01 2.46255249e-01 -4.52213943e-01
-1.29000628e+00 -8.62028599e-01 -4.33212996e-01 1.47203183e+00
-4.70674872e-01 -7.83864617e-01 -6.63928151e-01 -4.40855682e-01
-3.22552651e-01 9.17104125e-01 -2.51887083e-01 1.75289333e-01
-2.67573595e-01 -1.06784821e+00 8.71289134e-01 3.51550654e-02
2.52093643e-01 -1.36790681e+00 -1.81132630e-01 7.95136988e-02
-4.83372062e-01 -8.38061213e-01 -4.23931718e-01 -2.52905995e-01
-5.56217730e-01 -9.28494215e-01 -8.44284415e-01 -5.65085471e-01
4.86588746e-01 1.88248724e-01 1.47134423e+00 2.86107004e-01
-6.29627645e-01 2.28475913e-01 -5.70671022e-01 -5.44898450e-01
-9.40668583e-01 3.13046098e-01 4.24190722e-02 -6.42436564e-01
1.82887837e-01 -7.24818707e-02 -6.40746713e-01 3.49852443e-01
-1.27923834e+00 2.90525138e-01 4.35235173e-01 8.63807678e-01
5.02890944e-01 -4.45555627e-01 5.76398790e-01 -1.03647637e+00
1.10846913e+00 -4.75803822e-01 -7.48976707e-01 3.77218425e-01
-7.12262571e-01 2.69660056e-01 6.37520075e-01 -2.67749161e-01
-1.11427975e+00 -2.93315768e-01 -3.72366995e-01 1.00936793e-01
-8.40388760e-02 5.30823171e-01 2.77305156e-01 3.41846257e-01
1.27399385e+00 -3.98769863e-02 -1.86452046e-02 -7.58897603e-01
6.29226029e-01 1.04117393e+00 -1.71121154e-02 -8.41336548e-01
4.72151250e-01 6.78780824e-02 -3.47234070e-01 -8.05961549e-01
-7.49659300e-01 -5.15079379e-01 -7.11092502e-02 2.12301344e-01
2.99686909e-01 -1.05581772e+00 -4.49934453e-01 3.85482073e-01
-1.10010982e+00 -5.80843985e-01 -3.09022933e-01 3.55823934e-01
-5.71274042e-01 3.58036995e-01 -8.08641434e-01 -5.03968418e-01
-8.33026767e-01 -1.21972728e+00 1.26977956e+00 1.13036118e-01
-4.72179502e-01 -7.66118050e-01 3.93894047e-01 4.13101703e-01
6.71877742e-01 -6.47896230e-02 8.55409563e-01 -8.53488564e-01
-2.91504860e-01 -3.49435538e-01 -3.30946624e-01 2.60312438e-01
-1.52999610e-01 3.09460849e-01 -9.03865933e-01 -3.76774520e-01
-5.20035982e-01 -7.86937833e-01 6.51862264e-01 2.20574252e-02
1.16240597e+00 -4.62055683e-01 -2.52987266e-01 3.32317173e-01
1.16078627e+00 -2.92511374e-01 7.80877590e-01 4.57127988e-01
2.57222772e-01 6.42671287e-01 8.57395232e-01 4.34638590e-01
2.56354809e-01 1.04817402e+00 -9.60184485e-02 -1.72430292e-01
-2.23115057e-01 -1.28518879e-01 4.66737330e-01 8.08370531e-01
-1.62065014e-01 -4.35383320e-01 -8.32084119e-01 2.18747079e-01
-1.76063120e+00 -9.48166668e-01 -4.90209222e-01 2.69378495e+00
1.43031538e+00 -2.32845739e-01 1.33308433e-02 -2.98940301e-01
3.59499842e-01 -1.07302248e-01 -8.17438811e-02 -5.11280298e-01
-2.84744233e-01 5.89089990e-01 3.64822149e-01 6.25329018e-01
-9.11433220e-01 1.02973413e+00 6.31030893e+00 1.10171163e+00
-9.27134752e-01 2.12143380e-02 7.14115620e-01 -2.50367343e-01
-3.96021962e-01 2.96153873e-02 -1.03049672e+00 4.53651369e-01
1.37229741e+00 -2.67408222e-01 5.87692618e-01 8.58373582e-01
3.99412572e-01 -1.03947394e-01 -9.27323520e-01 7.32120633e-01
-2.76793450e-01 -1.42146969e+00 2.00798184e-01 -2.33827770e-01
7.32105494e-01 6.58577442e-01 -4.55743968e-02 4.99651313e-01
7.28012323e-01 -9.50709343e-01 3.62485439e-01 3.53468567e-01
1.20296097e+00 -4.50533271e-01 7.28037715e-01 3.99658084e-01
-5.92649877e-01 3.82373452e-01 -5.81637740e-01 1.33322015e-01
-1.17387012e-01 7.24675596e-01 -1.16717684e+00 6.97594821e-01
4.86970335e-01 1.45319775e-01 -7.95680702e-01 8.90227854e-01
-1.27760768e-01 4.41673338e-01 -4.36562330e-01 -9.62582752e-02
-5.85571378e-02 -5.08185178e-02 4.01750863e-01 1.45132065e+00
3.58439654e-01 -2.33539283e-01 1.40936285e-01 6.66716158e-01
-9.28879231e-02 4.57785070e-01 -4.62359875e-01 -2.06657141e-01
4.38797295e-01 1.51132607e+00 -3.05408448e-01 -3.85637224e-01
-2.32363731e-01 8.09994638e-01 3.54278117e-01 3.53589088e-01
-5.83524227e-01 -5.75386822e-01 5.26880980e-01 2.34227717e-01
-1.75473690e-01 -1.07117392e-01 -8.47324207e-02 -1.22536719e+00
-1.02421992e-01 -1.39690614e+00 6.99173212e-01 -1.06588757e+00
-1.26910329e+00 9.87398684e-01 3.35114181e-01 -1.18860495e+00
-6.89962566e-01 -5.55774152e-01 -1.43625304e-01 1.20745981e+00
-1.23378825e+00 -1.05426931e+00 -4.33014572e-01 3.92586499e-01
5.06746292e-01 -1.27357990e-01 1.27099788e+00 3.32347184e-01
-2.23976806e-01 6.66223109e-01 4.48910117e-01 -1.16540380e-01
1.31698537e+00 -9.70696449e-01 6.56090498e-01 2.50861436e-01
4.04915184e-01 9.82740164e-01 8.14469039e-01 -5.04789412e-01
-1.30522251e+00 -9.66223240e-01 1.05868292e+00 -7.24249363e-01
6.35250568e-01 -5.18235207e-01 -7.72095919e-01 3.46619725e-01
2.74230570e-01 -2.59656459e-01 6.06805861e-01 7.20885471e-02
-4.26572025e-01 3.42661701e-02 -9.90342557e-01 6.74194336e-01
7.88005948e-01 -2.06430435e-01 -3.97804111e-01 9.68172669e-01
6.93175375e-01 -6.51373208e-01 -8.63806427e-01 3.30145061e-01
5.02476633e-01 -8.08477223e-01 9.61972594e-01 -4.86783206e-01
4.19298738e-01 -2.35677257e-01 9.54306424e-02 -1.33749616e+00
2.75622755e-02 -6.65912092e-01 1.25954777e-01 1.22428262e+00
7.36152351e-01 -6.86040998e-01 4.31817919e-01 5.91054380e-01
9.64658055e-03 -5.96988261e-01 -4.59464759e-01 -6.03673041e-01
4.59721200e-02 -6.19569540e-01 4.44616854e-01 7.84632683e-01
-5.63401617e-02 4.30501848e-01 -1.94429159e-01 -3.39337945e-01
3.28322351e-01 3.78930150e-03 1.02844071e+00 -1.00439465e+00
-5.23752153e-01 -2.33416617e-01 2.49595478e-01 -7.96489000e-01
-6.58111051e-02 -1.19713271e+00 -1.27099857e-01 -1.58876383e+00
2.20815867e-01 -9.28816617e-01 -7.66790807e-02 7.81308055e-01
-1.02934994e-01 5.63446701e-01 7.57879987e-02 4.79622424e-01
-4.27284926e-01 2.42176816e-01 1.09267437e+00 3.04528587e-02
-3.78304511e-01 -6.17699511e-02 -7.96115160e-01 2.50145435e-01
9.14184153e-01 -5.37168026e-01 -3.66099894e-01 -3.90860111e-01
4.57191855e-01 -9.04533267e-02 1.87604204e-01 -7.58995891e-01
-1.01604797e-02 1.21941365e-01 3.35701287e-01 -4.81969118e-01
2.49314979e-01 -2.50973701e-01 2.24467754e-01 1.50825396e-01
-4.96018946e-01 2.32882395e-01 3.91591489e-01 7.96982050e-02
-1.09238341e-01 -2.57344604e-01 5.31593442e-01 -3.34426969e-01
-4.21737373e-01 1.63796961e-01 -1.83580920e-01 2.95509338e-01
6.03358686e-01 2.72594422e-01 -7.33051598e-01 -4.81640995e-01
-3.54852170e-01 1.20390929e-01 6.78149581e-01 7.54476488e-01
2.04887539e-01 -1.14915359e+00 -1.10000443e+00 2.87227839e-01
3.68113399e-01 -3.00851334e-02 2.03927532e-01 5.90081215e-01
-7.78839290e-01 6.57543123e-01 3.10483798e-02 -2.84411311e-01
-1.32000721e+00 6.03821158e-01 -1.51015356e-01 -7.07507730e-01
-3.04825246e-01 5.89169025e-01 1.76081453e-02 -8.07166636e-01
2.52945423e-02 -2.35785265e-02 -4.54444662e-02 -1.04689129e-01
9.82375205e-01 2.43797526e-01 4.44330096e-01 -2.96266586e-01
-6.36018068e-02 4.59872521e-02 -2.92903990e-01 -3.88448894e-01
1.33386946e+00 3.34579885e-01 -2.47172102e-01 2.75615007e-01
1.10317945e+00 2.01587617e-01 -6.81151986e-01 -1.08099550e-01
4.90422128e-03 -3.28806370e-01 -1.81613788e-01 -1.25473428e+00
-4.93965954e-01 5.60107648e-01 4.77227360e-01 1.52431071e-01
9.26879466e-01 -6.11666217e-03 5.85228384e-01 4.56460983e-01
5.47234774e-01 -1.10514581e+00 -2.42759343e-02 5.82598805e-01
1.22242403e+00 -1.17649150e+00 5.75618185e-02 -2.55910367e-01
-7.24296331e-01 1.07343721e+00 3.90946239e-01 1.93151191e-01
1.65977716e-01 2.67950118e-01 2.95830786e-01 1.26657877e-02
-1.07365394e+00 -5.55329621e-02 3.74111235e-01 4.63835657e-01
1.08599758e+00 -2.13407148e-02 -7.75110722e-01 4.00441200e-01
-3.80548924e-01 2.55033106e-01 4.68603969e-01 7.80803025e-01
5.40317595e-02 -1.87500489e+00 -4.06459481e-01 4.98782426e-01
-6.33439243e-01 -6.31483078e-01 -3.85475338e-01 7.77021170e-01
-4.02640015e-01 9.14374769e-01 -2.26858377e-01 -1.98292807e-01
2.78792113e-01 7.34216645e-02 4.21036422e-01 -8.84649336e-01
-7.99254894e-01 -2.20841423e-01 5.41225016e-01 -4.33711678e-01
1.83202028e-02 -4.66164947e-01 -1.07824469e+00 -3.16094756e-01
-1.00080185e-01 4.31211919e-01 9.60884333e-01 4.32347715e-01
7.72421956e-01 1.55875847e-01 3.59165221e-01 -8.84831488e-01
-8.15856755e-01 -1.34580719e+00 -2.67000049e-01 4.75417346e-01
-2.40056053e-01 -3.49649340e-01 -3.06767732e-01 -1.23741038e-01]
|
[11.288629531860352, 8.736373901367188]
|
f37d92b0-90dd-41d9-a82e-32309a12ec42
|
cluster-based-sampling-in-hindsight
|
2208.14741
| null |
https://arxiv.org/abs/2208.14741v1
|
https://arxiv.org/pdf/2208.14741v1.pdf
|
Cluster-based Sampling in Hindsight Experience Replay for Robot Control
|
In multi-goal reinforcement learning in an environment, agents learn policies to achieve multiple goals by using experiences gained from interactions with the environment. With a sparse binary reward, training agents is particularly challenging, due to a lack of successful experiences. To solve this problem, hindsight experience replay (HER) generates successful experiences from unsuccessful experiences. However, generating successful experiences without consideration of the property of achieved goals is less efficient. In this paper, a novel cluster-based sampling strategy exploiting the property of achieved goals is proposed. The proposed sampling strategy groups episodes with different achieved goals and samples experiences in the manner of HER. For the grouping, K-means clustering algorithm is used. The centroids of the clusters are obtained from the distribution of failed goals defined as the original goals not achieved. The proposed method is validated by experiments with three robotic control tasks of the OpenAI Gym. The results of experiments demonstrate that the proposed method significantly reduces the number of epochs required for convergence in two of the three tasks and marginally increases the success rates in the remaining one. It is also shown that the proposed method can be combined with other sampling strategies for HER.
|
['Dongsoo Har', 'TaeYoung Kim']
|
2022-08-31
| null | null | null | null |
['multi-goal-reinforcement-learning']
|
['methodology']
|
[ 5.02736233e-02 2.48132318e-01 -4.41909884e-04 -7.59055540e-02
-5.68479121e-01 -2.92009562e-01 4.88271803e-01 3.31770658e-01
-7.90991783e-01 1.33461118e+00 -1.48687214e-01 3.79109085e-01
-7.71491766e-01 -5.63434899e-01 -4.36599791e-01 -1.17909992e+00
-1.98245600e-01 5.31654179e-01 2.60856207e-02 -1.86049685e-01
6.28242433e-01 2.39781618e-01 -2.11714101e+00 -1.66966155e-01
1.05465984e+00 6.67287886e-01 8.55727911e-01 3.95860225e-01
1.33576810e-01 9.87072408e-01 -9.12052989e-01 4.79835749e-01
3.98271233e-01 -8.97127628e-01 -5.51917911e-01 4.00807947e-01
-4.88842368e-01 -3.03281307e-01 1.07848674e-01 1.05497098e+00
5.18647432e-01 9.34649110e-01 5.14156818e-01 -1.44055390e+00
-1.69466734e-01 6.63309634e-01 -4.03973669e-01 -2.13024870e-01
2.97137678e-01 1.00217927e-02 5.98250389e-01 -3.42167258e-01
5.54668546e-01 1.00487089e+00 1.46473795e-01 8.11590910e-01
-9.56837535e-01 -2.16429740e-01 1.64783537e-01 3.23387504e-01
-1.21767998e+00 -2.22785920e-01 7.58827865e-01 -7.42401481e-02
8.72199476e-01 -1.41991125e-02 9.61257517e-01 7.79643178e-01
-4.02312875e-02 7.77004302e-01 1.40362787e+00 -5.55832267e-01
1.08452821e+00 2.87879974e-01 -2.50655860e-01 4.39250857e-01
2.90155947e-01 3.43688041e-01 -3.38914514e-01 -1.67511180e-01
5.64756691e-01 1.10984981e-01 -1.58877850e-01 -6.57945871e-01
-9.13750648e-01 9.41426277e-01 4.00759608e-01 6.06205046e-01
-8.62789452e-01 4.42527374e-03 5.02318442e-02 4.12456870e-01
-4.75430340e-02 7.63609827e-01 -9.86743420e-02 -3.62362236e-01
-7.31270194e-01 5.30110657e-01 7.52711833e-01 9.17666316e-01
8.56366873e-01 3.16463709e-01 7.35917240e-02 7.59269893e-01
-8.38159919e-02 2.42401734e-01 9.02658343e-01 -1.10204279e+00
8.26170668e-02 6.29191279e-01 5.85049808e-01 -6.99458539e-01
-6.20833278e-01 -2.53212273e-01 -2.97906965e-01 8.61914158e-01
5.85391641e-01 -5.49543381e-01 -8.58140230e-01 1.78796816e+00
4.34461534e-01 -3.96038890e-02 5.97993493e-01 9.11988616e-01
2.82594234e-01 3.97584468e-01 -1.25100821e-01 -6.62940860e-01
5.60410857e-01 -8.59166265e-01 -7.76030838e-01 -3.12647492e-01
4.53515887e-01 -3.03686321e-01 7.31532633e-01 5.64126074e-01
-1.13959193e+00 -5.31217635e-01 -9.56415474e-01 6.67841792e-01
-1.31437436e-01 2.96571046e-01 6.92841768e-01 2.91401088e-01
-9.12723184e-01 9.92374122e-01 -8.76429558e-01 -5.68443775e-01
2.74804860e-01 3.43306452e-01 -3.93211916e-02 8.60780999e-02
-5.36426961e-01 8.97451162e-01 7.74574935e-01 -1.48188606e-01
-1.11368024e+00 -2.70382799e-02 -7.69595981e-01 1.01646818e-01
6.41474426e-01 -2.97999650e-01 1.20542228e+00 -1.11725891e+00
-1.66646898e+00 2.67384440e-01 2.35554978e-01 -4.43662286e-01
5.59732318e-01 -2.64627375e-02 2.16502443e-01 3.34971607e-01
3.18972498e-01 6.38503850e-01 7.94610083e-01 -1.40024984e+00
-9.50112343e-01 -3.94473433e-01 1.89556405e-02 8.08989465e-01
-2.52698809e-01 -5.14662683e-01 2.15978712e-01 -1.05022721e-01
1.63540155e-01 -9.83451605e-01 -6.65650368e-01 -5.79371154e-01
5.76832816e-02 -4.47864771e-01 2.57936865e-01 -6.63968623e-02
6.73624635e-01 -2.28998923e+00 4.70505089e-01 4.14343029e-01
-3.28134708e-02 -8.29910263e-02 -5.81300929e-02 6.57363653e-01
3.52503777e-01 -4.79607612e-01 1.98235698e-02 -4.42341894e-01
2.42698900e-02 4.54143018e-01 7.12148473e-02 3.68094414e-01
-1.89077467e-01 1.98633477e-01 -1.30462301e+00 -3.15078527e-01
3.86114180e-01 3.39727663e-02 -4.96979356e-01 4.52353239e-01
-2.46263251e-01 5.89716554e-01 -7.27692962e-01 2.53773272e-01
2.75869489e-01 -6.88196532e-03 6.10043883e-01 6.10055268e-01
-2.29334041e-01 -3.92094301e-03 -1.41485465e+00 1.81385970e+00
-2.28038520e-01 -6.41575754e-02 -6.21017814e-02 -1.18567789e+00
1.12044466e+00 3.97394985e-01 5.83856404e-01 -5.91160357e-01
3.42210382e-01 3.99438709e-01 2.38364875e-01 -5.57482541e-01
7.63208926e-01 -9.14840773e-02 -6.15247488e-02 3.51363152e-01
3.69950056e-01 -3.76909465e-01 5.90809941e-01 4.23766673e-02
1.01551104e+00 4.46885914e-01 7.31927216e-01 -1.07209109e-01
1.14290379e-01 3.05049002e-01 7.55071402e-01 1.29417682e+00
-3.57361555e-01 3.05141628e-01 2.43644387e-01 -1.69896305e-01
-9.09195483e-01 -9.19420719e-01 3.88838440e-01 6.99529588e-01
4.75471228e-01 -5.30882664e-02 -6.03121579e-01 -3.82454455e-01
-1.65161148e-01 9.36514258e-01 -6.23857200e-01 -1.90723225e-01
-3.29350650e-01 -6.34558618e-01 1.52395116e-02 1.18100800e-01
5.49330831e-01 -1.73937535e+00 -1.38790870e+00 5.13946593e-01
-2.15662882e-01 -6.59198284e-01 3.69813472e-01 4.71609265e-01
-8.32162499e-01 -1.10492182e+00 -7.59645939e-01 -9.70246136e-01
7.41346538e-01 3.53492171e-01 5.80213726e-01 1.85963571e-01
-2.19171464e-01 5.49783885e-01 -8.49443197e-01 -4.07977015e-01
-3.90585661e-01 -2.74630964e-01 4.16741937e-01 -2.36733094e-01
4.29416299e-01 -7.09471107e-01 -3.77653360e-01 5.10768555e-02
-5.31098127e-01 -5.12270182e-02 5.34774363e-01 1.08521557e+00
4.94843334e-01 5.83403707e-01 1.11429393e+00 -3.32736582e-01
9.21586812e-01 -4.45706457e-01 -5.26441216e-01 1.61509737e-02
-5.62619388e-01 1.09973447e-02 6.51367843e-01 -7.51858175e-01
-1.10697806e+00 -2.18456518e-02 2.67990977e-01 -3.99363846e-01
-4.32280689e-01 4.77163881e-01 2.71087259e-01 1.15932703e-01
7.70298719e-01 4.19146657e-01 2.66092509e-01 -9.61001888e-02
-7.72218267e-03 4.05909091e-01 1.29856318e-01 -5.65670252e-01
2.81638324e-01 1.90457210e-01 -5.76052777e-02 -7.43051171e-01
-5.20421505e-01 -3.81358981e-01 1.39871892e-03 -6.19522214e-01
5.95698059e-01 -7.17672467e-01 -9.82766986e-01 4.61827546e-01
-5.87854445e-01 -5.58466375e-01 -7.60307908e-01 9.55300987e-01
-1.27683663e+00 1.21702150e-01 -2.92456627e-01 -1.47230542e+00
6.41644374e-02 -1.09421384e+00 3.46075296e-01 8.81131649e-01
-1.40088871e-01 -7.13493824e-01 2.20798161e-02 -1.03284441e-01
2.39593834e-01 3.86258096e-01 3.55805576e-01 -7.69554257e-01
-2.70539194e-01 1.98316742e-02 4.85954016e-01 1.25987113e-01
2.41724804e-01 -5.45349061e-01 -5.98225892e-01 -5.91437221e-01
4.04889375e-01 -8.73364210e-01 4.24865544e-01 6.30121410e-01
6.35170341e-01 -9.24824998e-02 -3.00335824e-01 -4.49903160e-02
1.57668674e+00 8.06282401e-01 5.64192593e-01 7.63543487e-01
6.06150329e-02 6.93590105e-01 1.26124430e+00 1.09327888e+00
9.04939100e-02 3.39525044e-01 7.29069293e-01 3.50981832e-01
3.00290585e-01 -1.00247510e-01 4.03467804e-01 2.47730255e-01
-3.51434588e-01 -1.07474230e-01 -3.04616451e-01 8.81306589e-01
-2.27037573e+00 -1.32281101e+00 3.47542167e-01 2.25633025e+00
5.70568562e-01 1.09397903e-01 2.67132998e-01 3.23860496e-01
7.00248897e-01 -2.42417321e-01 -7.73586392e-01 -3.27358186e-01
1.20148525e-01 1.80474356e-01 9.85080898e-02 4.18054670e-01
-6.81012809e-01 8.14965427e-01 5.37714624e+00 8.36471081e-01
-5.32034218e-01 -1.54854521e-01 8.87197554e-02 -4.11869287e-01
2.25413576e-01 -1.45942234e-02 -5.83560705e-01 2.75447220e-01
3.68325710e-01 -3.47254634e-01 9.92177784e-01 1.11256242e+00
3.94057572e-01 -7.99389005e-01 -8.38063776e-01 8.29849839e-01
6.96169734e-02 -7.93584526e-01 -5.28344154e-01 -9.50657204e-02
8.68205369e-01 -2.16854066e-01 -1.11346617e-01 5.78443289e-01
7.78312445e-01 -6.09224081e-01 8.51122677e-01 5.35041392e-01
1.33081302e-01 -1.00565863e+00 6.58366978e-01 8.23376536e-01
-7.19663680e-01 -6.60205603e-01 -5.36025405e-01 -6.65368140e-01
1.94970239e-02 1.52372554e-01 -1.16105080e+00 7.99039423e-01
7.00067937e-01 4.59316492e-01 -2.03388840e-01 1.17181146e+00
-2.59963274e-01 2.41770789e-01 -2.91181087e-01 -6.99949443e-01
4.78407472e-01 -5.21656394e-01 7.51497805e-01 3.59327346e-01
8.04420888e-01 4.82775606e-02 4.41661417e-01 8.64602685e-01
6.54705703e-01 1.21863440e-01 -8.53663862e-01 -1.88437197e-02
6.46579266e-01 1.25729406e+00 -8.44386339e-01 -4.01379198e-01
5.83937056e-02 7.97941685e-01 5.68213344e-01 3.41823369e-01
-5.53973973e-01 -4.53326762e-01 1.42510489e-01 -3.58870119e-01
3.32750380e-01 -1.34943619e-01 -4.15290780e-02 -7.04143643e-01
-9.49002281e-02 -8.29491377e-01 3.38867545e-01 -8.58054936e-01
-9.66580451e-01 6.69425249e-01 1.09306835e-01 -1.49344373e+00
-7.56831288e-01 -5.05178422e-02 -4.27515179e-01 6.40776932e-01
-1.16153729e+00 -4.77623016e-01 -4.28430498e-01 6.43766999e-01
7.20393062e-01 -5.19461036e-01 8.56253982e-01 -2.03865901e-01
-2.06520244e-01 1.74257547e-01 3.47621888e-01 -4.98789042e-01
3.23529631e-01 -1.46148825e+00 -6.85723543e-01 5.99061489e-01
-2.39933223e-01 4.43350345e-01 1.03797221e+00 -7.17814147e-01
-7.06955969e-01 -6.59503996e-01 3.94850343e-01 5.32922745e-01
3.16284716e-01 2.33941391e-01 -4.96392190e-01 4.37773317e-01
3.16691667e-01 -5.54423392e-01 5.52513897e-01 -3.47635932e-02
6.12535655e-01 2.39831179e-01 -1.50395751e+00 8.54121208e-01
7.50215530e-01 2.99344242e-01 -8.10586214e-01 1.80374175e-01
2.42109045e-01 -4.09200817e-01 -5.32858551e-01 6.22466989e-02
3.16342682e-01 -1.13422763e+00 5.48776269e-01 -3.87652844e-01
3.31903666e-01 -3.89238447e-01 -5.14530391e-02 -1.99650717e+00
-2.26016700e-01 -5.64721465e-01 2.56398976e-01 1.01216483e+00
-4.53655757e-02 -5.72905958e-01 7.49103308e-01 1.86770335e-01
-2.35451579e-01 -4.47036088e-01 -9.62201774e-01 -9.70730186e-01
-1.77458659e-01 2.05607221e-01 3.61122638e-01 7.54091620e-01
5.13720632e-01 3.58059816e-02 -5.10062993e-01 1.65409630e-03
9.07054484e-01 2.48335138e-01 7.88982928e-01 -1.08447361e+00
-5.37467241e-01 -6.90730810e-02 -2.41121039e-01 -7.27006018e-01
-9.07327849e-05 -5.81976473e-01 5.47224879e-01 -1.64989614e+00
7.77048245e-02 -8.42582107e-01 -1.68200374e-01 3.60819221e-01
-1.29214063e-01 -2.58336455e-01 4.71679091e-01 8.41597244e-02
-8.31791162e-01 8.48806024e-01 1.29550683e+00 2.82458484e-01
-8.27040255e-01 1.72495440e-01 -4.58014190e-01 8.47312093e-01
1.28466797e+00 -5.82797468e-01 -6.66510284e-01 9.14982110e-02
9.39080119e-02 5.27934074e-01 7.21609294e-02 -1.52136838e+00
1.60652593e-01 -4.26217973e-01 3.01689684e-01 -5.01074016e-01
6.26212597e-01 -8.15717161e-01 2.19063282e-01 6.36701047e-01
-4.65623736e-01 -1.00221366e-01 -6.69132918e-03 7.68132746e-01
-8.35202113e-02 -9.15593088e-01 5.93113005e-01 -6.59138441e-01
-6.93627775e-01 -3.28990698e-01 -7.76498497e-01 -1.51117876e-01
1.51549923e+00 -3.33031833e-01 -2.39209700e-02 -5.46544492e-01
-1.19346237e+00 5.33527851e-01 3.61825377e-01 1.42317370e-01
7.23262489e-01 -1.29314232e+00 -4.10513371e-01 -7.25435698e-03
-6.17080778e-02 4.82400283e-02 3.63080919e-01 7.00808644e-01
-1.36769265e-02 -9.28473771e-02 -7.84233570e-01 -3.90801281e-01
-1.23067319e+00 4.91319090e-01 2.33244300e-01 -3.76025349e-01
-6.00847721e-01 3.84023100e-01 -2.00938910e-01 -4.18358713e-01
2.79399186e-01 -4.90065776e-02 -7.67695308e-01 4.87454273e-02
5.73318601e-01 5.62562883e-01 -2.49264434e-01 -1.07183591e-01
8.23809505e-02 8.37259367e-02 -5.34618869e-02 -4.02882189e-01
1.62877357e+00 -2.35292181e-01 2.55511522e-01 6.05971336e-01
3.83562565e-01 -3.60853314e-01 -1.58860755e+00 -2.26951558e-02
-2.09504142e-01 -5.93407393e-01 -2.89846897e-01 -7.72440493e-01
-6.58097744e-01 3.19243550e-01 5.72310209e-01 4.54784364e-01
1.19255447e+00 -3.14438492e-01 4.85609472e-02 5.29340506e-01
1.03377008e+00 -1.60094619e+00 5.10387301e-01 3.36705416e-01
7.38539457e-01 -1.08294094e+00 -2.49323651e-01 1.26816466e-01
-9.90763366e-01 9.15658236e-01 7.83945084e-01 -4.10960019e-01
7.49123842e-02 -1.29217401e-01 -1.06780685e-01 -5.67962788e-02
-6.87396288e-01 -6.58442795e-01 -6.50508165e-01 8.87973368e-01
-2.99796879e-01 1.86753675e-01 -8.91928613e-01 5.10095835e-01
-1.32162184e-01 2.04806149e-01 1.07773685e+00 1.43436730e+00
-8.32527637e-01 -9.14370418e-01 -5.61721563e-01 4.43463564e-01
-3.01802129e-01 3.07629704e-01 3.74327786e-02 7.55286992e-01
2.65153572e-02 1.47993016e+00 4.93639568e-03 -1.65427148e-01
1.82555854e-01 -1.42994793e-02 8.02879155e-01 -6.07240617e-01
-6.12261295e-01 1.74329445e-01 9.61022079e-02 -4.66224194e-01
-6.75518870e-01 -9.68533874e-01 -1.61966395e+00 -5.65085700e-03
-3.56527507e-01 5.49674511e-01 5.86893916e-01 8.58284831e-01
1.90855563e-01 6.51587188e-01 7.58849144e-01 -1.06388235e+00
-1.00431836e+00 -1.09926558e+00 -1.03006244e+00 5.34755051e-01
4.67794463e-02 -1.23955584e+00 -6.53927624e-01 -3.03021759e-01]
|
[3.973743200302124, 1.763352870941162]
|
4a679d1c-b5bb-472a-b289-ae8d1d4c428c
|
end-to-end-video-text-spotting-with
|
2203.10539
| null |
https://arxiv.org/abs/2203.10539v3
|
https://arxiv.org/pdf/2203.10539v3.pdf
|
End-to-End Video Text Spotting with Transformer
|
Recent video text spotting methods usually require the three-staged pipeline, i.e., detecting text in individual images, recognizing localized text, tracking text streams with post-processing to generate final results. These methods typically follow the tracking-by-match paradigm and develop sophisticated pipelines. In this paper, rooted in Transformer sequence modeling, we propose a simple, but effective end-to-end video text DEtection, Tracking, and Recognition framework (TransDETR). TransDETR mainly includes two advantages: 1) Different from the explicit match paradigm in the adjacent frame, TransDETR tracks and recognizes each text implicitly by the different query termed text query over long-range temporal sequence (more than 7 frames). 2) TransDETR is the first end-to-end trainable video text spotting framework, which simultaneously addresses the three sub-tasks (e.g., text detection, tracking, recognition). Extensive experiments in four video text datasets (i.e.,ICDAR2013 Video, ICDAR2015 Video, Minetto, and YouTube Video Text) are conducted to demonstrate that TransDETR achieves state-of-the-art performance with up to around 8.0% improvements on video text spotting tasks. The code of TransDETR can be found at https://github.com/weijiawu/TransDETR.
|
['Ping Luo', 'Yuanqiang Cai', 'Hong Zhou', 'Chunhua Shen', 'Ying Fu', 'Debing Zhang', 'Weijia Wu']
|
2022-03-20
| null | null | null | null |
['text-spotting']
|
['computer-vision']
|
[ 2.75828481e-01 -1.00414646e+00 -2.11748034e-01 -1.13909096e-01
-7.66151786e-01 -5.30651629e-01 6.89740300e-01 -2.37745151e-01
-4.39783782e-01 6.83164150e-02 1.79748937e-01 -2.96558380e-01
4.15411621e-01 -3.69760513e-01 -7.11441457e-01 -4.06808913e-01
5.46768129e-01 2.99013317e-01 5.85726738e-01 2.28406429e-01
4.29989666e-01 3.68658081e-02 -1.39317453e+00 5.19343674e-01
6.42017603e-01 9.24058497e-01 4.75582451e-01 1.10522771e+00
-4.30562466e-01 1.00074351e+00 -3.69910896e-01 -3.75710726e-01
5.70566654e-02 -3.46286595e-01 -5.00010729e-01 3.67888749e-01
6.61090136e-01 -8.42298508e-01 -9.55269098e-01 8.60579252e-01
5.41985214e-01 -5.23847528e-02 3.80040318e-01 -1.15055466e+00
-6.06297493e-01 6.62870109e-01 -8.97182763e-01 2.13619441e-01
7.13248253e-01 4.02988285e-01 1.00552857e+00 -1.48487234e+00
5.37450075e-01 1.08054292e+00 6.69889748e-01 6.05995059e-01
-7.66858518e-01 -8.51283610e-01 2.71106184e-01 4.19872850e-02
-1.60383296e+00 -6.60991013e-01 2.69633412e-01 -6.99504972e-01
8.08586538e-01 3.10051560e-01 5.10490417e-01 1.15493703e+00
8.50633085e-02 1.51823497e+00 4.55497861e-01 -2.28744134e-01
-1.83831379e-01 -4.01994914e-01 1.11193351e-01 7.33166933e-01
-2.62764189e-02 -2.56988168e-01 -1.04117644e+00 6.80394173e-02
6.38113856e-01 4.73224759e-01 -3.84281516e-01 1.21432491e-01
-1.57395637e+00 3.80522668e-01 -2.31895760e-01 2.88105071e-01
-1.24951512e-01 4.37175721e-01 7.98823535e-01 3.34064960e-01
5.25715411e-01 -4.13760364e-01 -1.83989391e-01 -4.29809779e-01
-1.56144845e+00 5.59823960e-02 4.33614790e-01 1.26355505e+00
5.09877622e-01 9.04644504e-02 -4.96446729e-01 7.69357860e-01
5.32982767e-01 9.36133504e-01 7.76818097e-01 -1.51459843e-01
9.50316727e-01 5.04461169e-01 -7.33522624e-02 -8.31553161e-01
4.83604521e-02 2.26750448e-01 -6.57748282e-01 -3.55041832e-01
1.86433896e-01 -2.75872618e-01 -1.19389701e+00 8.24188113e-01
2.14621469e-01 5.20675659e-01 -2.38972977e-01 9.04954493e-01
1.10928810e+00 1.02470720e+00 -1.33049414e-01 -9.56625491e-02
1.34682310e+00 -1.24224496e+00 -8.78020287e-01 -3.21761429e-01
8.46861601e-01 -1.14819336e+00 1.11210442e+00 2.65859246e-01
-9.68649864e-01 -4.19556201e-01 -7.09935069e-01 -2.87887305e-01
-1.89467430e-01 6.00126147e-01 -6.84745982e-02 3.73338968e-01
-1.02016115e+00 1.65788010e-01 -9.58277583e-01 -6.40783727e-01
3.20238143e-01 2.78744906e-01 -1.17861919e-01 -1.67427361e-01
-8.88468623e-01 5.68659566e-02 2.86231935e-01 1.45494729e-01
-1.00888681e+00 -5.15337646e-01 -6.70942545e-01 -4.99828793e-02
6.53009295e-01 -5.52093863e-01 1.36345732e+00 -1.07549262e+00
-1.41073823e+00 9.19950068e-01 -5.72212458e-01 -3.25334460e-01
9.00174439e-01 -5.79910994e-01 -3.52836818e-01 2.53878385e-01
1.72748804e-01 5.09159505e-01 1.33787453e+00 -5.50954044e-01
-9.08990860e-01 -2.14299396e-01 -6.82467043e-01 2.70368159e-01
-5.53916693e-01 5.40457666e-01 -1.23687446e+00 -1.08639586e+00
-3.42826210e-02 -8.37083280e-01 4.58159387e-01 2.41415977e-01
-5.12024522e-01 -3.94640923e-01 1.45729005e+00 -7.90663600e-01
1.63025904e+00 -2.28661847e+00 -9.96565521e-02 -2.54380167e-01
4.07960474e-01 4.21775073e-01 -8.72858018e-02 7.02868581e-01
1.20872088e-01 1.16668276e-01 1.00546636e-01 -6.83109581e-01
7.91558474e-02 -3.92356336e-01 -6.16596222e-01 6.13440037e-01
-1.49241954e-01 1.08382440e+00 -6.27314866e-01 -8.00912440e-01
5.21085322e-01 3.64219993e-01 -1.10976748e-01 4.53336090e-02
-3.56996328e-01 -9.69362631e-02 -6.94889367e-01 1.01795006e+00
4.63053286e-01 -4.83613342e-01 -1.08219087e-01 -1.24178976e-01
-2.52862990e-01 2.42336635e-02 -1.03898478e+00 1.63118756e+00
5.80714941e-02 1.27459967e+00 3.94499041e-02 -4.77011055e-01
6.63492858e-01 4.78554845e-01 5.79295576e-01 -5.69912136e-01
3.05246741e-01 1.73850343e-01 -7.65958607e-01 -7.29453385e-01
8.64015877e-01 4.73879606e-01 5.89182302e-02 6.19170070e-01
-9.93942842e-02 4.50851291e-01 3.48144650e-01 3.89316440e-01
1.19243300e+00 2.99287319e-01 -5.98805733e-02 1.67366430e-01
6.13919973e-01 1.09520569e-01 3.23369861e-01 7.96586871e-01
-3.52344602e-01 7.79424965e-01 9.54867527e-02 -3.60561937e-01
-9.28141594e-01 -6.39902830e-01 1.06886521e-01 1.38386393e+00
3.55822444e-01 -7.64214814e-01 -6.18185103e-01 -7.21207798e-01
-6.39190245e-03 1.89082697e-01 -4.25633162e-01 2.63974547e-01
-7.56450772e-01 -3.59517992e-01 9.22006190e-01 5.87828398e-01
6.33781135e-01 -9.10808563e-01 -4.33352262e-01 9.75992996e-03
-4.95842189e-01 -1.42123067e+00 -1.39088202e+00 -3.33255082e-01
-6.69641018e-01 -9.18708563e-01 -1.14096701e+00 -8.73286664e-01
3.62372041e-01 9.15492177e-01 7.57727325e-01 3.44698280e-01
-3.01959932e-01 5.93794584e-01 -7.00486839e-01 7.55437315e-02
-6.02915809e-02 -4.12963182e-02 -2.67336279e-01 2.99635828e-01
6.48085892e-01 1.85087472e-01 -4.93625641e-01 5.84321439e-01
-1.01333892e+00 4.90749300e-01 4.99669611e-01 7.57777452e-01
5.88884234e-01 -1.44248068e-01 7.93741047e-02 -5.95241308e-01
4.21288699e-01 -3.09024125e-01 -6.75753713e-01 4.94178236e-01
-3.75199527e-01 -4.85637188e-01 5.49624979e-01 -6.72346473e-01
-8.67689610e-01 3.57186705e-01 3.82423564e-03 -9.69985664e-01
-1.28592566e-01 4.81507093e-01 7.47725219e-02 8.43911469e-02
2.42945492e-01 9.48023677e-01 -2.82815337e-01 -3.11171174e-01
-2.19109561e-03 1.10024238e+00 4.30572480e-01 -2.03961641e-01
8.81706655e-01 5.09340703e-01 -6.72513485e-01 -9.67467189e-01
-4.53204900e-01 -9.74081039e-01 -5.50789356e-01 -4.33766186e-01
1.01330996e+00 -1.30168426e+00 -7.87168086e-01 9.98931050e-01
-9.77106929e-01 -6.04654849e-01 3.39594156e-01 4.97199595e-01
-4.53993171e-01 8.22879672e-01 -8.07168305e-01 -7.71719515e-01
-7.95438051e-01 -1.04991901e+00 1.62875235e+00 9.10009295e-02
1.13050751e-01 -7.87678480e-01 -2.08643436e-01 4.98583376e-01
1.90540507e-01 -3.12888980e-01 8.85242224e-02 -5.80575585e-01
-8.22391331e-01 -2.31196195e-01 -4.61391568e-01 -2.16697410e-01
-5.15907519e-02 4.26569730e-01 -7.48398006e-01 -5.11575580e-01
-2.89599091e-01 -2.28714526e-01 1.14093316e+00 3.32040161e-01
9.21122015e-01 -1.73071444e-01 -4.67206359e-01 6.61685407e-01
1.33247137e+00 3.04403424e-01 6.01216137e-01 3.19614768e-01
1.07307613e+00 1.32449284e-01 9.32593584e-01 5.53078592e-01
3.33228499e-01 6.91181779e-01 2.82012895e-02 9.42591727e-02
-2.57097393e-01 -3.26686978e-01 9.60507095e-01 8.16167653e-01
5.96406460e-01 -7.90862441e-01 -1.07186949e+00 4.45211470e-01
-2.29050565e+00 -1.08964455e+00 -3.37663352e-01 1.92335081e+00
5.00996172e-01 2.84522977e-02 3.56231958e-01 2.46921629e-02
1.25987589e+00 3.91662568e-01 -7.32127726e-01 2.89690763e-01
-1.20501399e-01 -4.17673856e-01 5.51834226e-01 2.46901304e-01
-1.27444184e+00 1.44123709e+00 5.27313471e+00 1.06279075e+00
-1.39308667e+00 -4.41653877e-02 3.29256654e-01 -2.79830873e-01
2.62893587e-01 -1.66281641e-01 -1.09606636e+00 8.37981224e-01
6.69877291e-01 -3.06124419e-01 4.32789981e-01 5.30796885e-01
5.08279562e-01 -6.18187785e-02 -9.82656062e-01 1.43503296e+00
3.76843333e-01 -1.45626116e+00 2.23746747e-01 -2.51763463e-01
3.86894017e-01 3.21383566e-01 1.41612059e-02 1.48138076e-01
1.17525421e-01 -7.73140192e-01 1.04194164e+00 3.13305438e-01
1.13577712e+00 -3.51055175e-01 3.44403684e-01 2.54989505e-01
-1.78244996e+00 1.07600734e-01 -1.00224085e-01 5.36028683e-01
2.20877841e-01 4.66643810e-01 -6.75796807e-01 5.09009719e-01
9.27097738e-01 1.49886274e+00 -5.98708391e-01 1.05377388e+00
4.55685779e-02 8.10745358e-01 -2.24875852e-01 -2.29630157e-01
2.82244295e-01 7.06486637e-03 5.46381712e-01 1.59453809e+00
4.03094828e-01 -8.02369416e-02 3.17360699e-01 5.70525289e-01
-3.41268599e-01 1.88360184e-01 -2.91382909e-01 -3.66783619e-01
5.19766271e-01 1.19883549e+00 -1.00205112e+00 -5.91560304e-01
-6.39033437e-01 1.29259622e+00 -2.97141284e-01 4.06315118e-01
-1.09993994e+00 -6.51257217e-01 2.52908200e-01 9.42101479e-02
7.43518293e-01 -1.70261264e-01 -6.26758859e-02 -1.58090794e+00
3.21316987e-01 -9.77569282e-01 3.21709573e-01 -1.09819281e+00
-9.94879007e-01 3.49567771e-01 -4.87005979e-01 -1.45198023e+00
5.07180542e-02 -5.77227592e-01 -6.77338481e-01 6.50819480e-01
-1.20023251e+00 -1.35845065e+00 -6.28295124e-01 1.00548995e+00
1.23769128e+00 -5.58280274e-02 1.26876011e-01 5.20017743e-01
-1.05236280e+00 7.69599140e-01 3.59528899e-01 6.92736387e-01
1.06443071e+00 -8.55814278e-01 8.38446856e-01 1.20816422e+00
1.32408619e-01 2.25308597e-01 3.80222261e-01 -1.04352403e+00
-2.03370237e+00 -1.35829675e+00 6.98754489e-01 -4.46319044e-01
9.36112940e-01 -6.03108644e-01 -8.71294022e-01 8.80769372e-01
2.90439814e-01 -4.63056117e-02 3.06449503e-01 -5.07540703e-01
-3.21276873e-01 1.36144668e-01 -6.46895826e-01 7.45774567e-01
1.05844438e+00 -6.61337614e-01 -2.72715628e-01 4.18995172e-01
6.79203153e-01 -5.78281522e-01 -4.72144246e-01 -4.57428098e-02
6.86355174e-01 -7.55279303e-01 6.49976552e-01 -6.29158020e-02
4.24612671e-01 -5.56572258e-01 -1.28549129e-01 -4.20573354e-01
1.16540462e-01 -9.67093587e-01 -3.66785377e-01 1.36560953e+00
1.25052020e-01 -4.07685161e-01 9.50960815e-01 2.36044511e-01
-7.84466639e-02 -5.67875624e-01 -7.71815956e-01 -6.14207983e-01
-3.51062030e-01 -5.48563480e-01 2.57788092e-01 9.75754201e-01
-2.96086799e-02 3.77907634e-01 -7.85693288e-01 -6.25171745e-03
3.98356915e-01 1.67811364e-01 9.45958972e-01 -8.55706632e-01
-1.09982789e-01 -4.43374932e-01 -2.97646999e-01 -1.92543793e+00
6.30967543e-02 -7.45100319e-01 3.42849255e-01 -1.54524803e+00
3.82240862e-01 -5.31026311e-02 2.12216705e-01 4.53070849e-01
-2.75524557e-01 1.15558401e-01 4.43943679e-01 7.83960402e-01
-1.17254114e+00 4.93308067e-01 1.00693321e+00 -2.40389526e-01
-1.59828570e-02 -1.49437830e-01 -2.02353612e-01 5.44456482e-01
5.90067744e-01 -4.88499999e-01 -4.45066653e-02 -9.00273860e-01
1.18914619e-01 3.17907453e-01 3.20739210e-01 -8.29762220e-01
8.46272290e-01 -8.65138173e-02 4.57833081e-01 -1.29028392e+00
1.65207043e-01 -8.47183585e-01 -1.53301194e-01 3.95384580e-01
-3.27658027e-01 3.66722584e-01 1.81353301e-01 6.20911181e-01
-1.39056504e-01 -2.37456262e-01 4.75026876e-01 3.41842473e-01
-9.50054407e-01 5.52517354e-01 -7.92598069e-01 8.08983669e-02
1.05114686e+00 -5.20813167e-01 -5.68714499e-01 -3.05943221e-01
-2.59831876e-01 5.56946099e-01 5.89217722e-01 6.82430327e-01
8.49933863e-01 -1.01866424e+00 -8.32854509e-01 3.49400192e-02
2.04327255e-01 -1.01468354e-01 3.56618702e-01 1.02583110e+00
-6.04671657e-01 6.26221895e-01 3.70780200e-01 -1.07510817e+00
-1.77965188e+00 5.33556640e-01 2.60129869e-01 -8.89260843e-02
-9.38726425e-01 6.20151281e-01 2.46365920e-01 2.87129749e-02
5.26287079e-01 -1.96589768e-01 1.79433763e-01 -5.06672077e-02
8.59017670e-01 3.82127166e-01 -6.85808361e-02 -7.15113819e-01
-3.92430067e-01 9.10209358e-01 -4.40838903e-01 -5.59527753e-03
8.48822296e-01 -5.08328676e-01 1.99878737e-01 3.45934510e-01
1.02677107e+00 1.03469081e-02 -1.35821176e+00 -4.68705654e-01
4.60720733e-02 -7.02583194e-01 1.17916927e-01 -5.50200999e-01
-1.13327074e+00 8.60771418e-01 5.30918360e-01 -6.27236208e-04
1.23533034e+00 -2.13697329e-01 1.16936624e+00 4.47722107e-01
-3.53055000e-02 -1.06661820e+00 2.65008688e-01 6.61874533e-01
5.37086368e-01 -1.23572659e+00 2.10597739e-02 -2.35844374e-01
-5.55359006e-01 1.20359135e+00 5.90526819e-01 1.50727272e-01
4.70310897e-01 5.69381177e-01 1.30823985e-01 -5.91528928e-03
-1.10107827e+00 -2.31416494e-01 2.49268115e-01 4.39545343e-04
5.97584844e-01 -2.98032880e-01 3.92301977e-02 7.59427622e-02
3.85518134e-01 2.75398403e-01 5.04244268e-01 1.06678629e+00
-5.57252645e-01 -6.95309877e-01 -5.72538853e-01 5.50098300e-01
-6.59286261e-01 -4.14100528e-01 -5.97701490e-01 6.81396961e-01
-4.56295967e-01 9.55574155e-01 1.13171540e-01 -6.29438102e-01
6.69149607e-02 -5.90968877e-02 -6.21543862e-02 -3.69286954e-01
-6.73825681e-01 6.79355502e-01 -2.76088089e-01 -4.33528543e-01
-2.88991004e-01 -7.31599629e-01 -1.32739520e+00 -7.15041876e-01
-6.06914818e-01 -8.26863721e-02 5.57097316e-01 7.71621585e-01
5.12095153e-01 4.07073498e-01 6.07141316e-01 -7.22876549e-01
-1.36683524e-01 -8.85580957e-01 -2.53837705e-01 2.07926273e-01
4.52941090e-01 -1.94625005e-01 -2.58720458e-01 6.12697601e-01]
|
[11.961345672607422, 2.17992901802063]
|
6da3b8c5-c0d7-4239-a48f-8ceef50388d4
|
hd-maps-fine-grained-road-segmentation-by
| null | null |
http://openaccess.thecvf.com/content_cvpr_2016/html/Mattyus_HD_Maps_Fine-Grained_CVPR_2016_paper.html
|
http://openaccess.thecvf.com/content_cvpr_2016/papers/Mattyus_HD_Maps_Fine-Grained_CVPR_2016_paper.pdf
|
HD Maps: Fine-Grained Road Segmentation by Parsing Ground and Aerial Images
|
In this paper we present an approach to enhance existing maps with fine grained segmentation categories such as parking spots and sidewalk, as well as the number and location of road lanes. Towards this goal, we propose an efficient approach that is able to estimate these fine grained categories by doing joint inference over both, monocular aerial imagery, as well as ground images taken from a stereo camera pair mounted on top of a car. Important to this is reasoning about the alignment between the two types of imagery, as even when the measurements are taken with sophisticated GPS+IMU systems, this alignment is not sufficiently accurate. We demonstrate the effectiveness of our approach on a new dataset which enhances KITTI [8] with aerial images taken with a camera mounted on an airplane and flying around the city of Karlsruhe, Germany.
|
['Sanja Fidler', 'Raquel Urtasun', 'Shenlong Wang', 'Gellert Mattyus']
|
2016-06-01
| null | null | null |
cvpr-2016-6
|
['road-segementation']
|
['computer-vision']
|
[ 1.87540248e-01 6.15300909e-02 2.09242180e-01 -4.32516366e-01
-6.44349039e-01 -8.01333368e-01 8.70440423e-01 -9.29126740e-02
-5.41594982e-01 7.58378327e-01 -5.72848320e-02 -5.11070132e-01
-1.69942066e-01 -1.32979977e+00 -7.53226578e-01 -3.88045251e-01
1.78349158e-03 7.20664322e-01 4.43119228e-01 -2.85524547e-01
2.26307467e-01 7.06154704e-01 -1.63983607e+00 -2.80000865e-01
8.44213963e-01 8.32344234e-01 4.98288333e-01 8.90847445e-01
1.81014866e-01 4.61460292e-01 -8.21233764e-02 -4.47223306e-01
5.51431358e-01 2.70532489e-01 -9.77889299e-01 7.42681563e-01
8.87264609e-01 -6.99876249e-01 -2.43814424e-01 1.24070072e+00
-1.61139011e-01 3.66231389e-02 5.35721183e-01 -1.09194934e+00
4.43662331e-02 3.68085235e-01 -4.53525215e-01 -3.50858979e-02
2.71228105e-02 1.82187527e-01 1.02281034e+00 -3.95440608e-01
4.64127809e-01 9.36875463e-01 7.39931107e-01 -4.40991223e-01
-1.16050243e+00 -1.52526677e-01 2.44482771e-01 1.49026364e-01
-1.70053899e+00 -1.12494834e-01 3.02786887e-01 -6.42183959e-01
4.51362103e-01 1.12595715e-01 4.59590822e-01 1.95050284e-01
-2.40160987e-01 4.20142591e-01 1.33805180e+00 -3.13663661e-01
2.17404366e-01 1.59800500e-01 1.07713893e-01 6.04865074e-01
3.59851032e-01 1.00413688e-01 1.45205289e-01 -4.91130911e-02
9.62846637e-01 1.81272328e-01 -8.00150558e-02 -5.01138210e-01
-1.36505544e+00 7.04204321e-01 5.80533564e-01 1.15844429e-01
-6.09437048e-01 1.17863856e-01 -9.10909772e-02 -5.00533283e-02
3.34787011e-01 1.74355477e-01 -3.25151891e-01 1.50402695e-01
-1.23341715e+00 2.43448779e-01 4.76767719e-01 8.22563827e-01
1.25783098e+00 -1.94254130e-01 4.99163032e-01 3.66613925e-01
1.86417446e-01 7.37635851e-01 -9.25146267e-02 -1.26904261e+00
4.91014928e-01 6.18823528e-01 5.91539919e-01 -1.02666223e+00
-3.33134472e-01 -2.48460814e-01 -5.50869107e-01 4.06807303e-01
4.68032032e-01 -3.32606614e-01 -9.19692874e-01 1.37633300e+00
3.43710631e-01 3.88502300e-01 1.42853204e-02 9.74529982e-01
6.72939643e-02 3.66445869e-01 -9.60881077e-03 5.06043375e-01
1.39852893e+00 -6.43032372e-01 -2.68272161e-01 -6.29262626e-01
6.55839264e-01 -6.08886003e-01 5.48305392e-01 1.34459242e-01
-6.52191699e-01 -7.95258105e-01 -9.41746652e-01 1.31270140e-01
-7.40708470e-01 5.38053334e-01 4.90408093e-01 5.79130352e-01
-1.25182271e+00 6.70032144e-01 -7.64272034e-01 -4.57388371e-01
2.05864049e-02 2.40398362e-01 -7.05333829e-01 -2.83778250e-01
-1.20982313e+00 9.79357481e-01 4.37095970e-01 1.61803856e-01
-4.98335063e-01 -1.68674752e-01 -1.01119149e+00 -7.20761195e-02
2.56779730e-01 -3.61115605e-01 9.69194949e-01 -6.33158088e-01
-9.23347771e-01 9.26342905e-01 1.31492063e-01 -9.38284576e-01
3.94740522e-01 -1.88703477e-01 -2.11525485e-01 1.22393884e-01
4.13320541e-01 9.37860012e-01 5.10791481e-01 -1.06321180e+00
-1.33114719e+00 -6.52932048e-01 6.43323183e-01 3.21839988e-01
2.95089424e-01 -3.77893031e-01 -2.53314972e-01 4.95585671e-04
1.56641543e-01 -1.15714061e+00 -4.76947933e-01 -2.69051313e-01
-4.49005574e-01 3.70973527e-01 8.93961310e-01 -9.36350346e-01
8.76806140e-01 -2.10940337e+00 4.10715006e-02 3.55765641e-01
-1.57388717e-01 5.89288361e-02 1.79743856e-01 1.88839838e-01
1.47021890e-01 4.88938875e-02 -5.21107554e-01 1.23905540e-02
5.78008369e-02 5.25001168e-01 -3.41542214e-01 4.63478684e-01
4.60141785e-02 5.61536312e-01 -7.91911781e-01 -2.87053794e-01
9.02338862e-01 1.73472106e-01 -1.30359098e-01 -1.34938270e-01
1.90968718e-02 2.30876833e-01 -3.57008576e-01 7.05662072e-01
9.24470782e-01 1.60336241e-01 1.82790354e-01 -8.44188556e-02
-7.22663045e-01 1.94037989e-01 -1.57805598e+00 1.32751107e+00
-2.06591740e-01 6.99150741e-01 3.66638333e-01 -9.20913517e-01
8.80772531e-01 1.34163368e-02 7.97562748e-02 -4.96315449e-01
-5.94062440e-04 1.02526300e-01 -3.70899409e-01 1.55901210e-02
1.05686593e+00 -2.24758968e-01 -2.40710899e-01 8.78968984e-02
-2.12412104e-01 -7.74553597e-01 4.04762983e-01 1.20120987e-01
7.81970441e-01 2.45222315e-01 3.56453151e-01 -3.99691254e-01
4.53284949e-01 5.64355016e-01 1.52656302e-01 5.37793934e-01
-1.43249303e-01 7.19901621e-01 2.71171123e-01 -4.65111583e-01
-1.22345865e+00 -1.06211102e+00 -4.53735173e-01 4.04320776e-01
2.72881925e-01 -2.43549898e-01 -9.28628564e-01 -4.67373908e-01
1.43633723e-01 7.42055058e-01 -6.38576210e-01 3.14382046e-01
-9.40533131e-02 -6.23696208e-01 4.91732925e-01 4.15368259e-01
1.04197466e+00 -1.92168310e-01 -7.17936695e-01 1.69101685e-01
-3.53065908e-01 -1.39633775e+00 1.03076003e-01 1.83471516e-01
-8.52010190e-01 -1.20298028e+00 -4.48403180e-01 -2.54058570e-01
5.59960783e-01 6.06307864e-01 9.93504047e-01 -2.64650077e-01
-6.09525144e-02 4.85771805e-01 -1.51034236e-01 -7.48557821e-02
6.49475306e-02 -4.33315262e-02 -1.34978741e-01 3.50973718e-02
1.73934504e-01 -4.84594941e-01 -3.42152715e-01 6.09788597e-01
-7.86501050e-01 1.58646703e-01 6.40047431e-01 4.74151373e-01
7.19592690e-01 7.50918210e-01 -4.22848761e-01 -7.06367970e-01
-2.20029145e-01 -5.85301340e-01 -1.06520486e+00 -6.50413483e-02
-1.56384930e-01 -1.64531395e-01 1.73136458e-01 4.64344651e-01
-9.10460532e-01 5.93370140e-01 -1.64940193e-01 1.74059905e-02
-8.67114365e-01 4.23062056e-01 -2.16812998e-01 -1.58081368e-01
4.71199661e-01 2.74851639e-03 -2.02658877e-01 -3.28067631e-01
5.49632311e-01 9.92115140e-01 1.01436710e+00 -2.64322162e-01
1.02196968e+00 9.04669285e-01 1.76975846e-01 -9.76005375e-01
-6.86560988e-01 -7.38289535e-01 -1.32326198e+00 -1.74263254e-01
8.31511378e-01 -1.14640176e+00 -6.64654896e-02 4.97042835e-01
-6.64305747e-01 -4.84029144e-01 -4.64555472e-02 4.16630328e-01
-7.46851861e-01 4.26728010e-01 -2.35834643e-01 -6.42189324e-01
1.86450928e-01 -1.12042749e+00 1.36133194e+00 2.27901980e-01
-4.60335240e-03 -1.10190868e+00 -8.97769034e-02 5.69685280e-01
7.95806199e-02 4.05822128e-01 4.72627819e-01 2.69237310e-02
-9.19565201e-01 -5.41701674e-01 -4.58747268e-01 4.38242495e-01
3.19182247e-01 2.11962432e-01 -8.11563194e-01 1.51327088e-01
-3.57100040e-01 1.70864537e-01 7.66212940e-01 4.79532778e-01
5.63633800e-01 1.32615089e-01 -4.05537874e-01 3.25549275e-01
1.60347676e+00 -1.00526653e-01 8.73542428e-01 6.26366019e-01
5.87223947e-01 6.03861153e-01 1.11793113e+00 2.33620033e-01
8.30430448e-01 9.36479688e-01 8.63282323e-01 -1.20807350e-01
7.47605264e-02 -1.00624382e-01 3.48190293e-02 -7.78917670e-02
-2.36496404e-01 7.09570721e-02 -1.05406225e+00 9.37944949e-01
-1.84562135e+00 -8.55317891e-01 -5.24009824e-01 2.50417399e+00
1.71382591e-01 6.43384606e-02 -1.39338209e-03 1.85316056e-01
8.26696694e-01 1.10721160e-02 -9.84302908e-02 3.24913580e-03
-1.09186195e-01 -2.36798421e-01 1.37551260e+00 9.04598415e-01
-1.55325663e+00 9.53490853e-01 5.93210077e+00 4.20609206e-01
-9.50577796e-01 -8.12497735e-02 4.55574900e-01 5.17364681e-01
-1.33508099e-02 5.23861051e-01 -8.72015834e-01 5.03210187e-01
1.01876688e+00 2.99686104e-01 3.13924849e-01 1.07775700e+00
4.15791944e-02 -8.10445905e-01 -4.51306343e-01 5.46556115e-01
-2.81593502e-01 -1.24430966e+00 -4.34967041e-01 4.75862145e-01
6.81234539e-01 1.87913701e-01 -1.52322799e-01 -2.51192483e-03
6.61945283e-01 -7.41417170e-01 7.22456098e-01 3.94823223e-01
5.00697494e-01 -7.50804424e-01 9.99273300e-01 5.76919973e-01
-1.19963813e+00 1.18469410e-01 -3.93804312e-01 -5.23806810e-01
3.19101393e-01 6.58693075e-01 -1.05424333e+00 6.97611749e-01
6.22927547e-01 6.50774658e-01 -6.33068204e-01 9.02657092e-01
-5.41735947e-01 4.14250821e-01 -6.60519004e-01 6.34402692e-01
8.05670261e-01 -5.54832458e-01 2.91208506e-01 8.57420802e-01
4.96766806e-01 1.42726123e-01 3.39158207e-01 3.94419998e-01
4.62120593e-01 -5.22559464e-01 -9.93460357e-01 1.75490931e-01
1.84551761e-01 1.51428425e+00 -8.48942161e-01 -4.49075252e-01
-3.51965666e-01 7.68842578e-01 4.52147052e-02 9.25772563e-02
-6.36504650e-01 -2.18138441e-01 8.13436151e-01 4.25588608e-01
3.57297957e-01 -7.55002022e-01 -2.00790048e-01 -8.68871212e-01
-1.99158445e-01 -2.34011739e-01 1.06167033e-01 -1.24032819e+00
-5.55131197e-01 4.14204180e-01 2.24402606e-01 -1.22566402e+00
-4.60605353e-01 -7.76374638e-01 -2.95804739e-01 9.49306846e-01
-1.54792964e+00 -1.16861093e+00 -4.47756767e-01 3.81619126e-01
1.43050939e-01 3.47849548e-01 7.34772980e-01 1.69590876e-01
-2.95650899e-01 -3.90558779e-01 1.32566184e-01 1.50995597e-01
4.43268269e-01 -1.61401260e+00 8.41791630e-01 1.23736179e+00
1.65715769e-01 1.70410037e-01 8.01009953e-01 -7.03385532e-01
-9.13631976e-01 -1.34371054e+00 7.74295986e-01 -5.94579697e-01
7.14419782e-01 8.55925307e-02 -5.73882997e-01 1.06288815e+00
-5.14767468e-02 -1.63269714e-01 2.20908329e-01 4.79933098e-02
1.31421879e-01 -4.65092845e-02 -1.15756404e+00 4.07284737e-01
6.78982317e-01 -6.04228675e-01 -6.19463861e-01 5.37022293e-01
1.27540469e-01 -4.94878858e-01 -7.62565434e-01 4.06378508e-01
1.75502047e-01 -1.16949058e+00 1.03552413e+00 1.06876627e-01
1.37206048e-01 -6.89407706e-01 -7.39528656e-01 -1.54245412e+00
-3.50654811e-01 1.03123225e-01 7.02262580e-01 1.04555237e+00
1.44155443e-01 -6.94410145e-01 8.38870525e-01 6.22751355e-01
-1.73179612e-01 -7.17205927e-02 -8.31599712e-01 -6.79943144e-01
-1.91878930e-01 -5.92358470e-01 6.58660591e-01 4.85395014e-01
-3.98007631e-01 1.60335764e-01 -3.86039108e-01 9.14338887e-01
7.36094236e-01 1.94908187e-01 1.01280618e+00 -1.48793125e+00
1.07506275e-01 -4.57074791e-02 -1.02255452e+00 -7.33919680e-01
5.61242178e-02 -3.39204103e-01 -2.27108821e-02 -1.76283395e+00
-2.55059421e-01 -4.93496746e-01 3.53367448e-01 2.97870219e-01
1.22401781e-01 5.46909153e-01 -6.54261261e-02 2.11847708e-01
-3.96562308e-01 1.00682475e-01 6.32203877e-01 -7.19471183e-03
2.81752080e-01 2.40434423e-01 -3.86066884e-01 1.16739130e+00
7.02025175e-01 -1.41126901e-01 -1.95525214e-01 -4.83122885e-01
6.30562305e-02 1.20463908e-01 8.36915135e-01 -1.25147820e+00
1.63979188e-01 -1.46368101e-01 3.13153118e-01 -1.08625841e+00
4.79227334e-01 -1.15629756e+00 4.33754593e-01 3.32214236e-01
2.86167353e-01 -6.01355806e-02 2.34834418e-01 4.21948105e-01
-5.25489688e-01 -2.62182742e-01 7.65271127e-01 -4.24439430e-01
-1.42286754e+00 1.50949493e-01 -4.71309334e-01 -5.65154254e-01
1.13468814e+00 -2.96419472e-01 -2.66824365e-01 -3.82335663e-01
-6.53382480e-01 4.13074344e-01 1.01216388e+00 2.97983494e-02
1.64095283e-01 -1.22523785e+00 -5.48433423e-01 3.25256526e-01
7.62665868e-02 9.63963345e-02 3.98599207e-01 7.41157353e-01
-8.47764254e-01 7.17323482e-01 -4.79923517e-01 -6.99639559e-01
-1.18380880e+00 3.57430190e-01 3.31124634e-01 -1.60526916e-01
-3.15181285e-01 3.31743091e-01 1.00945927e-01 -8.47857177e-01
-4.06673789e-01 -3.71399790e-01 -2.37079710e-01 7.21348748e-02
4.64101523e-01 3.39349478e-01 2.29291141e-01 -9.91701663e-01
-2.98389703e-01 8.28687489e-01 3.70968372e-01 -3.48660827e-01
9.19978797e-01 -7.09714711e-01 -2.11816039e-02 1.88811436e-01
6.03351116e-01 2.24658474e-02 -1.47609484e+00 -1.84579611e-01
-5.11058383e-02 -7.06381857e-01 4.72594202e-01 -5.06706893e-01
-6.87404633e-01 9.81296360e-01 6.65145040e-01 3.72071683e-01
8.82330775e-01 -2.08397955e-01 5.05899549e-01 4.20020014e-01
8.58292460e-01 -1.24026895e+00 -8.27982605e-01 5.58482468e-01
3.39183360e-01 -1.37548470e+00 6.47807270e-02 -5.67119479e-01
-6.18097782e-01 1.10628366e+00 5.47937937e-02 -3.27430695e-01
4.75887924e-01 1.75953597e-01 -7.46935681e-02 -7.28187412e-02
-1.76727787e-01 -1.01807010e+00 3.72960567e-02 7.18109906e-01
-1.60020307e-01 6.67908967e-01 2.90367633e-01 -9.28805098e-02
-3.23766083e-01 3.22603658e-02 7.57718205e-01 9.46045637e-01
-8.54456842e-01 -7.93042421e-01 -7.66637683e-01 3.04777324e-01
-1.13545351e-01 -6.60070553e-02 -2.56330967e-01 1.14530432e+00
3.09590131e-01 8.74908149e-01 3.91351253e-01 -3.54506999e-01
4.37896281e-01 -1.46628499e-01 1.88702315e-01 -5.51699340e-01
1.43669665e-01 -1.88708231e-01 4.78592634e-01 -3.39034855e-01
-3.86617333e-01 -9.45285618e-01 -9.74388659e-01 -2.77774543e-01
-1.26782626e-01 1.84053242e-01 1.01397383e+00 1.22973847e+00
9.24555138e-02 1.43906191e-01 7.05017149e-01 -1.18238175e+00
-9.76032987e-02 -8.23279679e-01 -1.16601837e+00 -4.92255986e-02
2.65051425e-01 -8.10438275e-01 -3.58095169e-01 8.53097215e-02]
|
[8.414305686950684, -2.0275676250457764]
|
5c9a1332-e0ea-47df-9281-2178519c07ec
|
maximal-clique-based-non-autoregressive-open
| null | null |
https://aclanthology.org/2021.emnlp-main.764
|
https://aclanthology.org/2021.emnlp-main.764.pdf
|
Maximal Clique Based Non-Autoregressive Open Information Extraction
|
Open Information Extraction (OpenIE) aims to discover textual facts from a given sentence. In essence, the facts contained in plain text are unordered. However, the popular OpenIE systems usually output facts sequentially in the way of predicting the next fact conditioned on the previous decoded ones, which enforce an unnecessary order on the facts and involve the error accumulation between autoregressive steps. To break this bottleneck, we propose MacroIE, a novel non-autoregressive framework for OpenIE. MacroIE firstly constructs a fact graph based on the table filling scheme, in which each node denotes a fact element, and an edge links two nodes that belong to the same fact. Then OpenIE can be reformulated as a non-parametric process of finding maximal cliques from the graph. It directly outputs the final set of facts in one go, thus getting rid of the burden of predicting fact order, as well as the error propagation between facts. Experiments conducted on two benchmark datasets show that our proposed model significantly outperforms current state-of-the-art methods, beats the previous systems by as much as 5.7 absolute gain in F1 score.
|
['Bin Wang', 'Limin Sun', 'Hongsong Zhu', 'Tingwen Liu', 'Yucheng Wang', 'Bowen Yu']
| null | null | null | null |
emnlp-2021-11
|
['open-information-extraction']
|
['natural-language-processing']
|
[-3.69823864e-03 7.69692183e-01 -8.35943520e-02 -2.12211996e-01
-9.31976974e-01 -4.72889960e-01 2.47274756e-01 5.01876354e-01
-1.20304422e-02 9.95210052e-01 4.15679157e-01 -2.20828518e-01
-2.78790414e-01 -1.15885973e+00 -9.50301707e-01 -5.72595954e-01
-1.38176516e-01 7.12019026e-01 1.55935794e-01 1.02444887e-02
8.34637582e-02 -2.80986220e-01 -1.06112039e+00 5.28269410e-01
1.13931274e+00 1.03226829e+00 9.81438085e-02 5.42129993e-01
-4.83513415e-01 1.47896028e+00 -4.30537164e-01 -1.12020075e+00
3.01349200e-02 -3.33302826e-01 -9.94245768e-01 -2.34999862e-02
9.80583578e-02 -1.19317807e-01 -5.65866411e-01 1.10344458e+00
1.31925093e-02 9.60743055e-03 4.97038096e-01 -9.45252120e-01
-7.91881382e-01 1.18551481e+00 -4.51644361e-01 3.04200083e-01
6.03763461e-01 -4.09920126e-01 1.51525748e+00 -9.06637669e-01
5.99972367e-01 8.85880709e-01 3.92916858e-01 1.32225439e-01
-8.79547179e-01 -5.45984983e-01 5.97729266e-01 2.96836734e-01
-1.38009250e+00 -2.94745952e-01 6.45872533e-01 -1.91994891e-01
9.84208226e-01 4.15572256e-01 5.45345247e-01 7.69122720e-01
5.19383073e-01 9.70599949e-01 7.19949245e-01 -1.72826737e-01
1.03832766e-01 1.44374995e-02 3.42289835e-01 8.54807138e-01
4.23366785e-01 -4.47630972e-01 -8.59064758e-01 -1.84275612e-01
2.74498791e-01 -2.05434695e-01 -2.39518702e-01 -2.40191836e-02
-1.05694723e+00 6.45694256e-01 3.47391099e-01 2.40294971e-02
-6.89055979e-01 -3.70368749e-01 2.64210671e-01 2.69579381e-01
6.15997076e-01 1.94106013e-01 -7.15272486e-01 -4.42995653e-02
-8.38069201e-01 2.97021508e-01 1.09788632e+00 9.98403609e-01
6.93878055e-01 -4.10467237e-01 -2.96225786e-01 6.57495975e-01
2.09411517e-01 2.61400729e-01 1.42935574e-01 -4.82496381e-01
9.55916822e-01 9.79180336e-01 -1.18939336e-02 -1.44090235e+00
-4.57625180e-01 -7.09884226e-01 -8.83377194e-01 -6.45159423e-01
2.15278268e-01 -3.76400799e-01 -6.48604453e-01 1.32786512e+00
4.61637259e-01 3.41075689e-01 3.06754619e-01 5.71474195e-01
1.18751609e+00 9.84513283e-01 -1.58959776e-01 -5.31311989e-01
1.42339504e+00 -8.51179540e-01 -9.57719445e-01 -3.80203426e-01
3.42173427e-01 -6.32034063e-01 6.26435280e-01 8.27825069e-01
-1.15276837e+00 -9.56180468e-02 -9.55521107e-01 -4.16175365e-01
-3.63497823e-01 2.04674572e-01 9.08910513e-01 3.69701505e-01
-6.89722002e-01 3.83940041e-01 -5.54396629e-01 3.80582273e-01
3.82080406e-01 3.28154653e-01 -2.23440602e-01 -1.42289251e-01
-1.29555285e+00 4.36423928e-01 6.94652498e-01 4.09374774e-01
-3.40609998e-01 -7.24383652e-01 -9.66693997e-01 5.69557965e-01
1.50484800e+00 -6.95399106e-01 1.02768362e+00 -3.23047519e-01
-1.22692609e+00 6.37439907e-01 -5.14048517e-01 -6.44975245e-01
4.81022112e-02 -4.99919355e-01 -7.34664500e-01 -3.78849581e-02
3.59012932e-01 1.63888127e-01 4.32144046e-01 -1.18565357e+00
-8.72404695e-01 -3.34308803e-01 2.61019409e-01 8.50988850e-02
-6.42925277e-02 -1.52997807e-01 -9.41632926e-01 -4.40479279e-01
5.21582723e-01 -5.49388409e-01 -1.47635639e-01 -8.57897580e-01
-1.03875518e+00 -3.23656887e-01 3.48839343e-01 -9.79950726e-01
1.84756732e+00 -1.78559911e+00 1.58097118e-01 2.71550208e-01
6.01358771e-01 4.17049602e-02 3.83133322e-01 4.17956948e-01
6.36348948e-02 8.22334811e-02 -2.95986980e-01 -6.32589310e-03
-9.92875323e-02 2.33208537e-01 -5.99464417e-01 -4.21524383e-02
1.64638489e-01 7.91907489e-01 -1.04360604e+00 -7.50923097e-01
-2.62661278e-01 3.52187939e-02 -7.27856696e-01 -2.66491622e-02
-7.44868100e-01 1.91734716e-01 -6.41602695e-01 3.55904698e-01
6.93924069e-01 -6.20419979e-01 4.68975365e-01 3.32759060e-02
1.00697152e-01 8.78077447e-01 -1.38775790e+00 1.53478456e+00
-3.06838244e-01 2.17176735e-01 -3.88839513e-01 -8.33519220e-01
9.13604081e-01 4.33545619e-01 5.24270713e-01 -3.02576065e-01
2.66355693e-01 1.97379608e-02 -2.22647578e-01 -6.64942026e-01
8.40246797e-01 2.16972232e-01 -2.37819970e-01 1.94658309e-01
2.88869202e-01 2.30201691e-01 5.96639574e-01 7.25826085e-01
1.08724451e+00 -1.06600840e-02 4.79379147e-01 1.09101929e-01
6.45266056e-01 -3.44006112e-04 8.79085541e-01 7.94875860e-01
5.31955123e-01 2.32145756e-01 1.14274299e+00 -4.53442127e-01
-5.73472857e-01 -9.01019037e-01 1.33928984e-01 4.66944247e-01
2.45447326e-02 -1.06289542e+00 -6.84248626e-01 -1.00614500e+00
-1.62001789e-01 7.61097431e-01 -6.04942441e-01 2.77357727e-01
-3.31601530e-01 -6.70409620e-01 1.73855633e-01 1.26169518e-01
5.91437697e-01 -8.11755240e-01 -2.07974121e-01 3.85746211e-01
-9.11253870e-01 -1.41423810e+00 -2.56756186e-01 1.42196819e-01
-5.82742095e-01 -1.18934309e+00 8.79544951e-03 -6.55032158e-01
6.18957520e-01 -6.25326261e-02 1.32671618e+00 1.46318629e-01
2.72496104e-01 -3.83744091e-01 -4.71022010e-01 -3.53936523e-01
-8.64280537e-02 2.14290753e-01 -2.24485964e-01 3.68574053e-01
5.30120015e-01 -5.29457211e-01 -3.01917315e-01 -8.04606751e-02
-6.99506938e-01 3.67740095e-01 6.84139431e-01 8.55080068e-01
8.82277012e-01 6.90891922e-01 4.94264036e-01 -1.51505625e+00
5.02685905e-01 -8.41353714e-01 -5.27762651e-01 4.86103177e-01
-5.27631044e-01 2.69161791e-01 8.33992779e-01 1.14045836e-01
-1.30083716e+00 -1.08527742e-01 -2.28955001e-01 1.54805640e-02
2.17987031e-01 1.17887700e+00 -3.95305753e-01 5.96091986e-01
3.24163467e-01 1.98858619e-01 -8.40570867e-01 -3.71902734e-01
3.70907634e-01 6.63370192e-01 9.00288641e-01 -3.03924054e-01
5.72669208e-01 2.44640097e-01 -3.90833467e-02 -3.84630740e-01
-1.57216227e+00 -4.55110103e-01 -6.09113991e-01 -8.13679248e-02
5.72921336e-01 -9.86885071e-01 -9.89859402e-01 2.45466866e-02
-1.11440909e+00 1.86224520e-01 -2.10261598e-01 2.93766588e-01
-1.66147918e-01 2.41047308e-01 -7.35752821e-01 -8.59993935e-01
-3.03051919e-01 -6.65392578e-01 8.28872025e-01 2.49450594e-01
-1.12437949e-01 -7.12079108e-01 -1.03108011e-01 4.25436378e-01
-4.83551115e-01 1.57381490e-01 1.18063223e+00 -8.58034432e-01
-9.01485622e-01 -3.12030792e-01 -4.87710200e-02 -1.34070486e-01
4.50976705e-03 -1.30096197e-01 -8.12251627e-01 2.69314766e-01
1.39523223e-01 -6.36762604e-02 9.14641559e-01 2.38353223e-01
1.09869444e+00 -6.84407294e-01 -4.03273135e-01 3.12166303e-01
1.48787141e+00 2.42449641e-01 7.62227118e-01 5.96016571e-02
7.04992890e-01 5.03585994e-01 6.34351969e-01 5.00342786e-01
7.28021562e-01 3.72160852e-01 1.92025781e-01 3.08612078e-01
2.55558401e-01 -7.57405937e-01 1.46323070e-01 1.11795843e+00
1.78066701e-01 -6.13530278e-01 -7.38659263e-01 4.96961921e-01
-1.86190176e+00 -9.53225613e-01 -4.83078122e-01 2.11926198e+00
1.05468416e+00 6.97052956e-01 -2.15269580e-01 4.49542075e-01
4.05543745e-01 1.19942307e-01 -3.09233159e-01 -3.04116964e-01
-1.82566062e-01 -1.06824962e-02 2.42136106e-01 6.97525561e-01
-9.23055172e-01 1.04770303e+00 5.60340643e+00 8.97884309e-01
-5.85800409e-01 3.77101675e-02 7.06180632e-01 -1.35035679e-01
-4.75041628e-01 2.83164650e-01 -9.18450177e-01 4.86771703e-01
8.18796396e-01 -2.12784946e-01 3.18708330e-01 6.07303679e-01
-9.74610671e-02 -6.46821022e-01 -8.81057322e-01 6.87761486e-01
1.73858389e-01 -1.47799993e+00 6.87509030e-02 -7.53855482e-02
8.54894936e-01 -3.28256428e-01 -3.89528275e-01 4.88238245e-01
5.62359869e-01 -7.17134058e-01 6.11543298e-01 6.69303834e-01
4.34352130e-01 -9.87278342e-01 8.30901623e-01 6.41833305e-01
-1.19046593e+00 -1.25893965e-01 -3.98286730e-01 -2.45529979e-01
1.70915172e-01 1.17614591e+00 -7.81673908e-01 1.22324276e+00
6.17634177e-01 6.54613435e-01 -3.74261558e-01 8.37984502e-01
-6.61739826e-01 8.23055208e-01 -1.59795761e-01 -6.88487664e-02
4.93839532e-02 -2.89656281e-01 4.81832534e-01 1.01612663e+00
2.86466610e-02 5.49910784e-01 2.53697604e-01 7.13387191e-01
-2.96367377e-01 2.17668727e-01 -3.76455873e-01 1.33818001e-01
3.66115212e-01 1.05396092e+00 -8.24351132e-01 -5.34295619e-01
-7.22698927e-01 8.63804460e-01 6.60995424e-01 1.70823649e-01
-6.86725140e-01 -4.26126242e-01 6.98515475e-02 -2.99063087e-01
5.10558188e-01 9.70671400e-02 -5.59809566e-01 -1.38694549e+00
4.90101397e-01 -8.33279252e-01 7.84279823e-01 -8.97553623e-01
-9.17328656e-01 7.91394174e-01 -2.88514584e-01 -9.44211066e-01
-3.53982300e-01 -2.32436225e-01 -3.02858979e-01 7.65691102e-01
-1.35330570e+00 -6.18926942e-01 6.84406832e-02 5.47924101e-01
5.78914821e-01 1.64034009e-01 7.69914925e-01 3.07477653e-01
-8.26835394e-01 4.37480330e-01 2.86357217e-02 2.67165750e-01
2.24558398e-01 -1.26006436e+00 6.05891049e-01 1.27823091e+00
6.82448745e-01 5.32598197e-01 7.14908063e-01 -1.16604149e+00
-1.21493638e+00 -8.07421625e-01 1.63974845e+00 -4.31007504e-01
6.32222474e-01 -4.87038314e-01 -9.10617292e-01 1.00152004e+00
3.85493636e-02 -2.02923656e-01 5.32966614e-01 6.58179462e-01
-1.53175876e-01 -1.34737596e-01 -7.90143669e-01 6.13787472e-01
8.62356365e-01 -3.15936387e-01 -6.71048462e-01 3.53337795e-01
9.12300706e-01 -8.20642531e-01 -7.45603323e-01 1.69228360e-01
3.81466478e-01 -9.45821822e-01 6.41059339e-01 -6.25947893e-01
1.03664386e+00 -1.54165700e-01 1.60350561e-01 -1.26563942e+00
-3.55448544e-01 -8.95017922e-01 -7.04813123e-01 1.36225843e+00
1.01075435e+00 -2.76446104e-01 8.27777267e-01 5.30521631e-01
-4.66743521e-02 -1.22107708e+00 -7.69617319e-01 -4.92266178e-01
-4.18959171e-01 -5.94028234e-01 8.03298295e-01 7.05134332e-01
5.41856050e-01 7.54328251e-01 -6.88122392e-01 5.94565749e-01
3.00846130e-01 3.61174077e-01 6.23804271e-01 -1.24944723e+00
-4.02068049e-01 1.02871694e-01 -3.72094102e-02 -1.45438671e+00
6.39466150e-03 -9.04167295e-01 8.36190954e-02 -1.81263709e+00
4.77750391e-01 -4.95720029e-01 -7.20482618e-02 4.11234796e-01
-6.95037305e-01 -3.29946011e-01 -1.95323210e-02 3.56108844e-02
-8.06438982e-01 4.10157114e-01 1.15078533e+00 -1.57478489e-02
-2.44279668e-01 7.42540210e-02 -1.07470906e+00 8.18291247e-01
5.78861833e-01 -7.83338487e-01 -4.79899973e-01 -3.64683181e-01
9.38372850e-01 6.79588854e-01 2.56218892e-02 -9.17263210e-01
6.32046103e-01 -8.90361238e-03 2.83344388e-01 -1.01262736e+00
1.94162235e-01 -8.25891197e-01 9.85453203e-02 5.96924860e-04
-2.07251742e-01 -1.69932082e-01 1.28038943e-01 8.23079705e-01
-3.08690667e-01 -2.29908124e-01 -1.18821278e-01 1.53261991e-02
-3.54811877e-01 3.58523488e-01 -1.26739919e-01 3.87090713e-01
6.02763772e-01 1.45662010e-01 -3.16996336e-01 -2.98822016e-01
-8.50357473e-01 5.15079558e-01 -4.51602668e-01 1.88793480e-01
5.96004128e-01 -1.02331245e+00 -7.45959520e-01 1.28459543e-01
-5.13026081e-02 4.42366689e-01 4.51914281e-01 7.40038574e-01
-2.41426248e-02 7.19275892e-01 4.26074624e-01 -1.50007248e-01
-1.01195562e+00 7.96842098e-01 -1.07934028e-01 -1.15160251e+00
-8.24957609e-01 9.27445173e-01 1.21563576e-01 -1.64788410e-01
1.57772720e-01 -4.49031949e-01 -3.55464041e-01 -8.11563153e-03
5.38842142e-01 1.70619443e-01 3.87140870e-01 -2.99602389e-01
-6.36979491e-02 8.12820420e-02 -5.13083518e-01 8.59031975e-02
1.19976246e+00 -2.70319581e-01 -2.27095172e-01 5.76715350e-01
6.83486879e-01 4.63201165e-01 -1.05756676e+00 -5.71023226e-01
1.23090975e-01 -4.91458267e-01 4.33837101e-02 -1.07394195e+00
-1.04517877e+00 2.48190969e-01 -4.17767435e-01 5.77740967e-01
1.24982345e+00 1.58983484e-01 1.05089498e+00 4.17479813e-01
4.59280312e-01 -9.24603403e-01 -4.12451178e-01 6.71179235e-01
7.69982994e-01 -9.82192814e-01 1.91859305e-01 -1.16771686e+00
-8.22234631e-01 8.46684992e-01 3.11534226e-01 -7.47914091e-02
5.93608677e-01 4.03134406e-01 -3.34009618e-01 -5.65611064e-01
-1.12321985e+00 -1.05173230e-01 2.80598819e-01 6.98584318e-02
2.34537646e-01 9.13345963e-02 -4.39233869e-01 1.29986680e+00
-5.77546835e-01 1.60573289e-01 5.40517032e-01 5.39325237e-01
-3.26899201e-01 -8.83908570e-01 -2.68854946e-01 8.75236392e-01
-7.27042496e-01 -4.76581842e-01 -4.31335717e-01 4.88749444e-01
2.24403217e-01 1.26938629e+00 -1.35386273e-01 -5.28723180e-01
3.81704569e-01 9.32523459e-02 1.96090728e-01 -5.94665408e-01
-4.95466858e-01 -4.71994504e-02 4.21484292e-01 -4.07936305e-01
-8.50128531e-02 -7.70356655e-01 -1.41174138e+00 -4.22255278e-01
-5.79558134e-01 5.00639796e-01 1.99951306e-01 1.20566559e+00
3.78844082e-01 8.42154682e-01 6.16134405e-01 6.77553043e-02
-7.05433637e-02 -8.70016396e-01 -7.11087823e-01 -7.99397230e-02
2.61666566e-01 -6.21386170e-01 -1.51714101e-01 1.42479658e-01]
|
[9.450746536254883, 8.561880111694336]
|
a57bb8a2-cdbe-47a7-8680-9db71a152b52
|
privacy-preserving-structure-from-motion
| null | null |
https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/1273_ECCV_2020_paper.php
|
https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123460324.pdf
|
Privacy Preserving Structure-from-Motion
|
Over the last years, visual localization and mapping solutions have been adopted by an increasing number of mixed reality and robotics systems. The recent trend towards cloud-based localization and mapping systems has raised significant privacy concerns. These are mainly grounded by the fact that these services require users to upload visual data to their servers, which can reveal potentially confidential information, even if only derived image features are uploaded. Recent research addresses some of these concerns for the task of image-based localization by concealing the geometry of the query images and database maps. The core idea of the approach is to lift 2D/3D feature points to random lines, while still providing sufficient constraints for camera pose estimation. In this paper, we further build upon this idea and propose solutions to the different core algorithms of an incremental Structure-from-Motion pipeline based on random line features. With this work, we make another fundamental step towards enabling privacy preserving cloud-based mapping solutions. Various experiments on challenging real-world datasets demonstrate the practicality of our approach achieving comparable results to standard Structure-from-Motion systems.
|
['Johannes L. Schönberger', 'Viktor Larsson', 'Marc Pollefeys', 'Pablo Speciale', 'Marcel Geppert']
| null | null | null | null |
eccv-2020-8
|
['image-based-localization']
|
['computer-vision']
|
[-1.47192135e-01 -2.21274067e-02 4.80725104e-03 -5.85031509e-01
-7.22691357e-01 -9.97403145e-01 6.90345824e-01 6.44575655e-02
-6.23266339e-01 7.62874126e-01 -6.90502077e-02 -1.29593536e-01
2.64111906e-02 -6.85266554e-01 -8.21380258e-01 -4.58507389e-01
-1.29213324e-02 2.83660412e-01 4.07568038e-01 2.62167808e-02
5.23545802e-01 1.03442240e+00 -1.39480412e+00 -1.14704534e-01
4.94512767e-01 9.74328399e-01 1.09174348e-01 4.84747410e-01
-7.66188465e-03 4.88565952e-01 -2.30664104e-01 -6.69309795e-01
5.45094311e-01 2.43079960e-01 -5.90577126e-01 -2.41007134e-02
6.04135990e-01 -6.21344447e-01 -5.22425532e-01 1.12772501e+00
4.20518011e-01 -1.09548084e-01 2.39161011e-02 -1.71561432e+00
-4.52053845e-01 -2.37904549e-01 -4.90011871e-01 -3.03264350e-01
7.64965892e-01 -2.12474242e-01 6.13873780e-01 -8.80955219e-01
8.95017743e-01 8.28515172e-01 8.59624863e-01 1.79674223e-01
-9.83946383e-01 -5.73145390e-01 -4.37445082e-02 2.12459639e-01
-2.11928940e+00 -6.49519801e-01 4.93097186e-01 -3.30335706e-01
5.71532845e-01 6.35228992e-01 4.22197253e-01 5.93124270e-01
3.26715820e-02 4.13314402e-01 9.37441528e-01 -3.45137715e-01
2.33246297e-01 7.07000673e-01 -2.39294350e-01 5.40269077e-01
6.13614202e-01 3.73233929e-02 -6.07126713e-01 -5.78696728e-01
6.40299082e-01 2.36981645e-01 -5.26919782e-01 -1.38620996e+00
-1.02940738e+00 5.97561359e-01 5.48313498e-01 -3.55540812e-02
-1.39542408e-02 2.19719216e-01 8.87468606e-02 1.41123891e-01
3.19713086e-01 1.87100712e-02 -5.15999049e-02 9.42879077e-03
-9.08311248e-01 2.70061433e-01 7.92390227e-01 1.53262627e+00
1.16356218e+00 -6.78874791e-01 5.03658116e-01 1.95037469e-01
4.70061868e-01 5.91470838e-01 1.75638180e-02 -1.00634646e+00
4.87422019e-01 3.28805327e-01 6.31170809e-01 -1.63679457e+00
-3.42242606e-02 -3.07156052e-02 -6.14804924e-01 2.70129561e-01
1.82245612e-01 1.35473117e-01 -4.17201310e-01 1.29071355e+00
5.05494297e-01 5.69239333e-02 -1.09245051e-02 1.01191926e+00
2.80424923e-01 3.42921317e-01 -4.66963619e-01 2.26860523e-01
1.02917564e+00 -6.27464056e-01 -6.65161252e-01 -1.12216935e-01
5.06571233e-01 -6.74248934e-01 3.86139959e-01 1.74744383e-01
-7.70046711e-01 -1.73046872e-01 -1.22646821e+00 -2.99572945e-01
-5.60383201e-01 3.52334268e-02 4.65637982e-01 8.69215012e-01
-1.37105882e+00 2.97170699e-01 -8.45010698e-01 -6.37164652e-01
4.56204206e-01 6.11719966e-01 -9.45274591e-01 -5.34863770e-01
-7.65797913e-01 8.92006099e-01 2.25543395e-01 3.01024765e-01
-3.49703997e-01 -3.64733756e-01 -8.64003778e-01 -2.49695778e-01
3.97094458e-01 -5.62595010e-01 7.38544464e-01 -3.22688580e-01
-9.77607131e-01 9.73403990e-01 -4.54570711e-01 -5.09097576e-01
1.05516672e+00 -4.35717523e-01 -1.38062388e-01 3.29699785e-01
2.85973400e-01 6.68161809e-01 6.80983424e-01 -1.48298001e+00
-6.90157115e-01 -5.98070383e-01 -6.59530535e-02 1.60563394e-01
-2.47254744e-01 -5.93111105e-02 -7.43912101e-01 1.79751273e-02
5.06799996e-01 -1.22106874e+00 -3.04857582e-01 7.85040438e-01
-3.69155347e-01 3.40373427e-01 1.12725914e+00 -5.02400100e-01
5.92923224e-01 -2.38251781e+00 -3.78186494e-01 3.86879832e-01
1.10022992e-01 2.45900434e-02 2.51370043e-01 6.56333148e-01
3.91441315e-01 8.93313363e-02 3.28799821e-02 -7.06619143e-01
1.51906470e-02 2.75442749e-01 -6.53564930e-01 1.07954240e+00
-4.62241396e-02 8.48004043e-01 -8.13771725e-01 -5.38070142e-01
4.85705316e-01 5.91587484e-01 -3.50187123e-01 3.47846858e-02
2.08248541e-01 4.75937486e-01 -2.98014313e-01 6.07568204e-01
1.42247188e+00 5.04550971e-02 1.53481737e-01 -4.54241000e-02
-3.82976443e-01 -1.94110703e-02 -1.09215808e+00 1.95550013e+00
-8.11692849e-02 7.71956444e-01 2.96400785e-01 -4.81518686e-01
1.02203619e+00 1.24710239e-01 6.19410038e-01 -3.08660537e-01
-7.28813782e-02 3.11275363e-01 -7.65007973e-01 -1.92856044e-01
1.11572826e+00 4.77871120e-01 -9.45485085e-02 1.66211277e-01
-4.98946995e-01 -3.19438905e-01 -6.55508935e-01 2.79288739e-01
9.90312159e-01 4.39665020e-01 1.42457306e-01 1.45933062e-01
5.18040419e-01 3.40935528e-01 4.32998419e-01 6.85692608e-01
-2.64138967e-01 9.77976978e-01 -8.14671442e-02 -4.26641226e-01
-1.08963776e+00 -9.58089054e-01 -5.53313531e-02 2.82067657e-01
6.89312160e-01 -4.40140247e-01 -3.90901059e-01 -4.99477953e-01
3.57801110e-01 3.18332791e-01 -3.49969029e-01 2.06654221e-01
-4.94595677e-01 -1.15179516e-01 5.24233818e-01 1.92867547e-01
6.67938650e-01 -2.57038534e-01 -8.57303202e-01 -1.01000361e-01
-2.64842600e-01 -1.43383563e+00 -2.63605833e-01 -2.53329724e-01
-5.22533059e-01 -7.81572461e-01 -5.80231249e-01 -5.14448464e-01
1.04377377e+00 9.59074557e-01 4.88902777e-01 1.48445934e-01
-3.01905066e-01 6.29849851e-01 -2.31939465e-01 -2.30660826e-01
1.06604293e-01 -1.53977945e-01 5.44347167e-02 2.00805590e-01
3.92820120e-01 -5.77396750e-01 -5.74087024e-01 5.02496362e-01
-7.18455970e-01 -1.60824195e-01 3.77330631e-01 2.65671164e-01
6.96612775e-01 -2.21289799e-01 -1.90098122e-01 -5.65304518e-01
8.68314952e-02 -6.16169572e-01 -9.98795331e-01 1.30224928e-01
-6.81947351e-01 -3.53520632e-01 2.29720265e-01 -1.09732710e-01
-5.47624350e-01 6.88023508e-01 4.18873936e-01 -7.11321473e-01
-5.62994704e-02 2.09779695e-01 -4.05982733e-01 -8.00644040e-01
1.03747606e-01 4.41607624e-01 1.82123050e-01 -3.02629679e-01
6.25530064e-01 8.70867908e-01 9.01820481e-01 -5.79421297e-02
1.23784804e+00 1.27663851e+00 1.85951576e-01 -7.80413687e-01
-2.03006268e-01 -8.54237378e-01 -9.14354146e-01 -8.87232199e-02
5.56389928e-01 -1.15949643e+00 -8.87884974e-01 3.11268866e-01
-1.35138595e+00 4.31321919e-01 6.70204982e-02 3.84351492e-01
-5.93776643e-01 8.64865780e-01 -1.45269204e-02 -9.68931019e-01
7.60554597e-02 -1.01572263e+00 1.03231943e+00 4.73758131e-02
-3.58766243e-02 -6.56378925e-01 3.49753462e-02 2.78318077e-01
5.26839972e-01 4.99560416e-01 1.13964327e-01 -3.64820719e-01
-1.35150170e+00 -7.76998699e-01 -4.06036645e-01 -7.59164095e-02
-1.24632306e-02 -3.69217575e-01 -1.02352965e+00 -5.87897241e-01
5.56350835e-02 1.51989283e-02 2.47241184e-01 -4.02879380e-02
7.06996679e-01 -1.87169135e-01 -6.20241344e-01 1.01315486e+00
1.76549733e+00 -2.17675164e-01 8.49899650e-01 6.51195228e-01
7.07703888e-01 5.71844876e-01 1.01964355e+00 4.94121939e-01
8.28546941e-01 9.19731081e-01 8.62455189e-01 9.81238037e-02
3.53323311e-01 -5.63298941e-01 1.14312030e-01 3.89090836e-01
3.22468936e-01 -1.14333245e-03 -6.72670364e-01 6.33698761e-01
-2.08277535e+00 -7.72098422e-01 -3.02475333e-01 2.53425097e+00
1.83497354e-01 -3.92267317e-01 -2.48222664e-01 -3.36463124e-01
7.66223311e-01 1.20359004e-01 -4.57014501e-01 3.85461771e-03
-1.40283078e-01 -4.78446841e-01 1.24170566e+00 4.27645892e-01
-1.05432689e+00 9.07287896e-01 5.78406143e+00 2.28206307e-01
-1.17933857e+00 9.49560702e-02 1.28447935e-01 -6.26363009e-02
-3.30542833e-01 5.54813385e-01 -7.57946372e-01 3.21647912e-01
6.02416277e-01 -4.02156383e-01 3.08957666e-01 1.10298419e+00
6.16253726e-02 -5.02818406e-01 -9.98837292e-01 1.25481570e+00
2.70922333e-01 -1.42900360e+00 -3.22194487e-01 6.51337862e-01
4.75877464e-01 3.37600291e-01 5.62228747e-02 -3.79390061e-01
2.56324321e-01 -7.76184797e-01 1.01932573e+00 3.69642496e-01
9.08851266e-01 -7.09250689e-01 7.05074489e-01 4.84212518e-01
-1.12596679e+00 2.01645061e-01 -8.87848258e-01 -2.73570622e-04
2.77397722e-01 3.81585240e-01 -1.16041827e+00 1.03862393e+00
8.70037913e-01 4.28493679e-01 -6.46277487e-01 1.34986210e+00
9.03069079e-02 -1.60493508e-01 -6.86276317e-01 4.43879664e-02
-5.37689850e-02 -2.30626717e-01 3.95688415e-01 8.29074681e-01
5.60017347e-01 -1.93060994e-01 4.51218300e-02 7.58811891e-01
9.28618461e-02 7.31054414e-03 -1.15647244e+00 3.24909449e-01
9.85808194e-01 1.19854093e+00 -6.88222826e-01 1.01760879e-01
-4.75575656e-01 1.35566187e+00 2.49802545e-01 4.03809249e-02
-7.13333130e-01 -4.00828719e-01 7.35852718e-01 2.39027813e-01
5.40522933e-01 -8.60977888e-01 -1.40201584e-01 -1.15877438e+00
4.95044112e-01 -3.29223961e-01 -7.93428645e-02 -7.88477659e-01
-7.48847425e-01 4.77950096e-01 -1.02344565e-01 -1.53252423e+00
-3.04608762e-01 -1.55078575e-01 -7.41706416e-02 8.67784977e-01
-1.77066398e+00 -1.35843194e+00 -6.32897079e-01 7.72052646e-01
-2.68483192e-01 1.71096355e-01 9.29468453e-01 2.49606594e-01
9.83797535e-02 4.75883096e-01 4.15138811e-01 2.30730891e-01
8.48164856e-01 -7.24365294e-01 7.90048242e-01 1.12680626e+00
3.15362334e-01 9.31596041e-01 4.70508605e-01 -7.22059011e-01
-2.00218034e+00 -1.01279569e+00 9.74964142e-01 -8.64141881e-01
3.68385345e-01 -8.14377725e-01 -6.01755798e-01 8.72895539e-01
-1.01967901e-01 4.60580200e-01 4.87266779e-01 -5.24368823e-01
-3.83293897e-01 -4.42420840e-02 -1.51306999e+00 4.01188403e-01
9.52885687e-01 -7.64647663e-01 -2.25561589e-01 2.45635331e-01
6.18347287e-01 -5.49612403e-01 -6.50238335e-01 3.57650034e-02
5.33664167e-01 -1.22424889e+00 1.06396019e+00 1.14803746e-01
-2.74468899e-01 -8.60969305e-01 -5.67411005e-01 -7.42750108e-01
6.61453828e-02 -9.41589475e-01 2.50608444e-01 1.37720084e+00
-1.58380747e-01 -9.25786674e-01 1.18451774e+00 1.13381565e+00
2.85825104e-01 -9.06383619e-02 -1.31687057e+00 -7.75838137e-01
-4.87688959e-01 -4.26117778e-01 8.44914675e-01 1.05072033e+00
-2.31229752e-01 -4.56677705e-01 -6.17302477e-01 8.67902756e-01
7.71629691e-01 4.85897139e-02 1.37109184e+00 -9.52077508e-01
2.87683696e-01 4.00573581e-01 -1.13490987e+00 -1.09495270e+00
2.31993804e-03 -6.80862308e-01 3.77229415e-02 -1.35409701e+00
-3.26186158e-02 -7.84590542e-01 8.30025524e-02 3.01120311e-01
3.57513279e-01 4.19676661e-01 3.47755283e-01 6.07220054e-01
-6.37569189e-01 2.86893249e-01 4.42284077e-01 1.98964998e-01
2.62262762e-01 -3.33916396e-02 -4.72014934e-01 4.91560847e-01
4.87271875e-01 -5.88076949e-01 -4.25308615e-01 -6.58006251e-01
1.73485696e-01 -1.06536731e-01 7.48281002e-01 -1.10469973e+00
8.10241818e-01 -1.34732813e-01 2.30643824e-01 -1.00480378e+00
7.63348401e-01 -1.57630479e+00 6.42669559e-01 3.79914701e-01
1.67612657e-01 1.42119899e-01 4.57220199e-03 8.57424498e-01
-1.49304509e-01 -9.60854292e-02 3.29052866e-01 -7.85507541e-03
-9.36813235e-01 3.45045269e-01 1.42059103e-01 -5.42583823e-01
1.46083748e+00 -5.42592943e-01 -3.25232089e-01 -7.79759884e-01
-2.08244875e-01 1.98887199e-01 1.40555131e+00 4.23578292e-01
8.72813523e-01 -1.21272469e+00 -4.93312120e-01 3.27287376e-01
3.59401166e-01 1.58857122e-01 -6.22878745e-02 7.21546352e-01
-9.85453010e-01 6.09666169e-01 -1.47218615e-01 -8.11812699e-01
-1.42502069e+00 7.99249351e-01 -6.99227825e-02 4.68836397e-01
-7.03037322e-01 6.42976940e-01 -1.80819437e-01 -4.51785982e-01
2.80360848e-01 4.88366298e-02 5.06067991e-01 -3.28364819e-01
6.41181827e-01 3.84760201e-01 4.03818637e-02 -1.02983236e+00
-7.91894019e-01 6.12045228e-01 -1.57110497e-01 -3.67673039e-01
1.28499019e+00 -8.61507893e-01 -3.36525500e-01 6.50382694e-03
1.35432076e+00 5.58402538e-01 -1.27156937e+00 -4.13402431e-02
2.70422995e-01 -1.13044608e+00 -1.50962129e-01 -2.87165433e-01
-8.29233944e-01 6.12184286e-01 5.83572686e-01 -6.03861958e-02
6.45949006e-01 -1.68918625e-01 6.48504734e-01 3.22614372e-01
1.42562699e+00 -7.50141978e-01 -6.56861603e-01 -5.93619468e-03
5.44526219e-01 -1.27034283e+00 3.80311519e-01 -7.51059890e-01
-4.55631912e-01 1.01068008e+00 2.12210432e-01 -7.67761096e-02
3.98794055e-01 3.17887217e-01 2.15659603e-01 9.29934382e-02
-1.60706893e-01 1.46399513e-01 -2.07721591e-01 9.87286866e-01
-2.73815393e-01 -1.68540612e-01 8.74245092e-02 5.30529320e-02
-1.61624685e-01 1.91932753e-01 6.20645165e-01 1.40575862e+00
-2.80651271e-01 -1.25050938e+00 -7.05114067e-01 -2.69844204e-01
-2.04858720e-01 2.32551470e-01 -6.01365745e-01 7.55082965e-01
-1.12644650e-01 1.00762701e+00 -2.55129822e-02 -5.18105209e-01
7.43303150e-02 -3.20579708e-01 1.96860686e-01 -2.27634594e-01
-2.06761748e-01 -2.34663978e-01 -8.30103233e-02 -9.92091954e-01
-2.57684559e-01 -8.17011654e-01 -1.07001257e+00 -5.56819499e-01
-4.22664076e-01 2.31728449e-01 1.25844955e+00 3.93202007e-01
9.13945735e-01 -4.48324531e-01 8.02875280e-01 -9.37688410e-01
-5.68014145e-01 -2.56464005e-01 -3.93673658e-01 1.19719282e-01
4.82106864e-01 -3.96445572e-01 -3.32534015e-01 -7.88091719e-02]
|
[7.591151237487793, -2.1216816902160645]
|
3579445a-c3a8-4fd4-af81-332fc6d4bbf4
|
isa-net-improved-spatial-attention-network
|
2211.02256
| null |
https://arxiv.org/abs/2211.02256v1
|
https://arxiv.org/pdf/2211.02256v1.pdf
|
ISA-Net: Improved spatial attention network for PET-CT tumor segmentation
|
Achieving accurate and automated tumor segmentation plays an important role in both clinical practice and radiomics research. Segmentation in medicine is now often performed manually by experts, which is a laborious, expensive and error-prone task. Manual annotation relies heavily on the experience and knowledge of these experts. In addition, there is much intra- and interobserver variation. Therefore, it is of great significance to develop a method that can automatically segment tumor target regions. In this paper, we propose a deep learning segmentation method based on multimodal positron emission tomography-computed tomography (PET-CT), which combines the high sensitivity of PET and the precise anatomical information of CT. We design an improved spatial attention network(ISA-Net) to increase the accuracy of PET or CT in detecting tumors, which uses multi-scale convolution operation to extract feature information and can highlight the tumor region location information and suppress the non-tumor region location information. In addition, our network uses dual-channel inputs in the coding stage and fuses them in the decoding stage, which can take advantage of the differences and complementarities between PET and CT. We validated the proposed ISA-Net method on two clinical datasets, a soft tissue sarcoma(STS) and a head and neck tumor(HECKTOR) dataset, and compared with other attention methods for tumor segmentation. The DSC score of 0.8378 on STS dataset and 0.8076 on HECKTOR dataset show that ISA-Net method achieves better segmentation performance and has better generalization. Conclusions: The method proposed in this paper is based on multi-modal medical image tumor segmentation, which can effectively utilize the difference and complementarity of different modes. The method can also be applied to other multi-modal data or single-modal data by proper adjustment.
|
['Zhanli Hua', 'Xiaohua Zhuc', 'Tianye Niu', 'Dong Liang', 'Haining Wangg', 'Fan Yang', 'Lu Zhang', 'Na Zhang', 'HaiYan Wang', 'Hao Shen', 'Zixiang Chen', 'Guoshuai Wang', 'Sijuan Zou', 'Zhengyong Huang']
|
2022-11-04
| null | null | null | null |
['tumor-segmentation']
|
['computer-vision']
|
[ 1.65350273e-01 -1.25879869e-01 -4.28739041e-01 -2.64049798e-01
-7.77347744e-01 -3.27646852e-01 9.00551379e-02 -1.25560053e-02
-7.24742293e-01 5.51509500e-01 1.09587699e-01 -4.37927753e-01
8.84172022e-02 -6.64379358e-01 -1.26551718e-01 -1.04651988e+00
3.92683297e-01 5.05335271e-01 6.00658357e-01 -8.37630555e-02
-2.06969127e-01 5.08161485e-01 -6.70423687e-01 3.81874055e-01
9.58801806e-01 1.03764760e+00 8.30644965e-01 2.71057725e-01
-2.97984809e-01 5.28564334e-01 -2.02881575e-01 6.42120913e-02
-1.78166106e-02 -4.67572182e-01 -6.90009058e-01 -8.65882933e-02
-3.72474670e-01 -2.81424493e-01 -3.34912866e-01 1.11150956e+00
5.59039354e-01 -2.04202563e-01 5.65684557e-01 -9.10326600e-01
-3.28690022e-01 6.06123865e-01 -8.99243593e-01 4.43279952e-01
-3.62374991e-01 2.72560716e-01 4.91496861e-01 -5.47568083e-01
2.00932890e-01 5.42440772e-01 7.29554236e-01 4.28340346e-01
-7.32910872e-01 -8.59898210e-01 -1.05343396e-02 3.33686322e-02
-1.39651597e+00 5.22282571e-02 5.44272959e-01 -4.36769992e-01
4.05096829e-01 3.12435269e-01 9.04289603e-01 7.74314523e-01
5.34191370e-01 8.06067646e-01 9.27024424e-01 -1.67641297e-01
2.51704332e-04 1.07354715e-01 -6.10426664e-02 7.65556693e-01
9.27175581e-02 5.49147092e-03 2.80877143e-01 1.99325547e-01
1.12019110e+00 5.86630881e-01 -6.70740187e-01 -5.91114871e-02
-1.33040547e+00 8.16927493e-01 1.01674259e+00 7.92557478e-01
-4.26150292e-01 9.70321968e-02 5.18024802e-01 -4.63164598e-01
3.11500758e-01 1.23789854e-01 -1.74812078e-01 5.24462722e-02
-9.99321997e-01 -5.07966101e-01 2.07637399e-01 4.08472925e-01
1.68840587e-01 -2.13063583e-01 -3.95329207e-01 7.98819482e-01
3.96160007e-01 4.06979114e-01 1.21858323e+00 -2.30512232e-01
1.63637176e-01 6.35071337e-01 -3.10297966e-01 -6.35160983e-01
-8.77888620e-01 -7.40557611e-01 -1.21094143e+00 -1.19910285e-01
2.88300484e-01 -6.69859201e-02 -1.39825141e+00 1.63821423e+00
2.73134977e-01 1.69457376e-01 -1.98463634e-01 1.32895005e+00
1.02640831e+00 3.56679082e-01 5.13135612e-01 -3.09936613e-01
1.56740034e+00 -1.02906907e+00 -6.80565059e-01 -2.77382620e-02
9.08547044e-01 -7.05485404e-01 9.23640311e-01 -1.15415744e-01
-7.18907416e-01 -1.90434799e-01 -8.03155720e-01 6.78604543e-02
-2.18017533e-01 4.76580888e-01 8.84511590e-01 5.74876070e-01
-6.90292835e-01 1.54205576e-01 -1.16494501e+00 -4.11575913e-01
8.61581445e-01 7.07921088e-01 -2.77083188e-01 -9.45683196e-02
-1.14991987e+00 9.24088240e-01 6.96815073e-01 1.75533950e-01
-7.87725627e-01 -7.26897001e-01 -4.91032958e-01 4.31464277e-02
4.72041219e-01 -6.18262589e-01 1.33891308e+00 -1.20270956e+00
-1.55609047e+00 5.27823210e-01 3.93069535e-03 -7.37954602e-02
4.38419789e-01 4.28451449e-01 -4.24120545e-01 4.06554252e-01
1.82760909e-01 6.99865162e-01 2.88506299e-01 -8.37770879e-01
-5.79824567e-01 -5.46823621e-01 -4.36239064e-01 3.80672276e-01
-2.17021361e-01 -1.84246927e-01 -6.60562158e-01 -6.43067598e-01
4.27218467e-01 -9.14511085e-01 -3.78260970e-01 9.56484377e-02
-4.14636523e-01 9.01626274e-02 9.51686144e-01 -8.52318704e-01
1.08049726e+00 -2.08560348e+00 5.91286309e-02 3.74387622e-01
3.94421935e-01 4.41328615e-01 1.47663385e-01 -2.54404068e-01
-1.60442173e-01 1.60533965e-01 -3.70552331e-01 1.40673295e-01
-4.67588931e-01 1.53903186e-01 2.17552960e-01 5.44011176e-01
-1.99081719e-01 1.27724588e+00 -7.34904706e-01 -8.87497127e-01
3.89005184e-01 3.37176412e-01 -1.77763239e-01 -8.15256760e-02
-1.69066433e-02 8.26678097e-01 -7.17287540e-01 8.46229374e-01
5.96481323e-01 -5.58228493e-01 2.19350997e-02 -6.15676582e-01
-5.78963645e-02 -3.22667032e-01 -5.02909422e-01 1.73506999e+00
-4.51493144e-01 4.16722745e-01 8.75283182e-02 -8.16571653e-01
4.50206071e-01 5.10654628e-01 9.31629360e-01 -9.06112313e-01
6.57489419e-01 3.30742538e-01 4.13703233e-01 -6.57931387e-01
-2.19792444e-02 -6.14960253e-01 1.76599771e-01 3.33322525e-01
-2.45205224e-01 -1.07266158e-01 -1.98324621e-01 4.96250652e-02
8.86475801e-01 -3.09944600e-01 3.36521536e-01 -1.77866727e-01
5.32174826e-01 -1.02714328e-02 6.02201104e-01 2.89932013e-01
-3.70120138e-01 7.09382951e-01 2.89212435e-01 -2.24708557e-01
-7.48413444e-01 -7.78087080e-01 -3.90944928e-01 5.59036493e-01
3.73668879e-01 7.33838603e-02 -5.84948063e-01 -8.84128809e-01
-2.11447895e-01 4.05313581e-01 -8.04618120e-01 -2.80155689e-01
-4.63651955e-01 -1.18842471e+00 5.20762920e-01 8.87566984e-01
8.87564242e-01 -8.71166229e-01 -6.10552788e-01 3.24529052e-01
-3.94764543e-01 -1.11533237e+00 -5.66468656e-01 3.45637798e-01
-9.17648494e-01 -1.07881546e+00 -1.19028628e+00 -7.10931718e-01
7.48434782e-01 2.42527574e-01 6.23935699e-01 1.36651978e-01
-3.08539629e-01 -3.70112248e-02 -2.32767165e-01 -3.70996296e-01
-1.67864710e-01 4.01567966e-01 -5.04314721e-01 -4.75329757e-02
2.65166670e-01 -2.80235261e-01 -5.97811759e-01 5.19029200e-01
-1.06500864e+00 5.93964815e-01 1.24003434e+00 1.05669999e+00
7.49565482e-01 1.05347708e-01 3.46380830e-01 -7.83460855e-01
3.65351766e-01 -6.03246510e-01 -4.18373555e-01 2.27749407e-01
-1.23667389e-01 -2.54713058e-01 5.81295788e-01 -3.96237761e-01
-1.00928211e+00 3.12278926e-01 -3.38204414e-01 -3.65242094e-01
-2.37052161e-02 7.11267114e-01 -1.85028970e-01 -3.29812884e-01
3.97060901e-01 2.46369407e-01 1.44513085e-01 -9.61720198e-02
-1.33752123e-01 6.68217599e-01 4.01007444e-01 -1.46712452e-01
3.06722879e-01 6.40842199e-01 1.71698004e-01 -5.75149417e-01
-7.32961059e-01 -4.08924550e-01 -5.10607302e-01 -1.44248813e-01
1.22624087e+00 -7.29811847e-01 -6.97845757e-01 5.01231372e-01
-9.18519020e-01 -4.22276944e-01 -8.20508972e-02 8.61539602e-01
-1.36954531e-01 2.38492534e-01 -7.54290342e-01 -1.73471764e-01
-3.74823749e-01 -1.58207810e+00 1.00217295e+00 5.84337771e-01
1.68307230e-01 -1.02375090e+00 -1.93402633e-01 1.99907348e-01
5.88360786e-01 1.80419266e-01 8.14947844e-01 -6.32557869e-01
-5.04558206e-01 -2.59096414e-01 -6.61918223e-01 -5.97773353e-03
3.46417814e-01 -9.46739838e-02 -8.00674260e-01 -8.33180249e-02
-1.15617082e-01 -1.19743511e-01 9.57373202e-01 8.83034945e-01
1.60044301e+00 2.73293704e-01 -8.42085063e-01 8.00668895e-01
1.43458056e+00 5.76363802e-01 5.72705209e-01 1.63872048e-01
1.06349909e+00 1.69562519e-01 3.72389823e-01 9.97446701e-02
2.51083732e-01 6.69996500e-01 5.97767055e-01 -6.74745023e-01
-8.36003497e-02 -4.24834676e-02 -2.10345849e-01 5.71978331e-01
-7.73612261e-02 -2.96873838e-01 -1.12104368e+00 5.15043378e-01
-1.73680961e+00 -7.14372098e-01 -1.73432246e-01 1.82541275e+00
7.22852349e-01 -1.01209566e-01 -1.23923473e-01 -1.95987802e-02
8.49894345e-01 -2.82130063e-01 -6.27248526e-01 7.63918832e-02
1.05351038e-01 2.45036691e-01 7.69041002e-01 1.57802075e-01
-1.16540778e+00 6.29141212e-01 5.52205610e+00 1.16981137e+00
-1.60222995e+00 5.09099841e-01 8.72436583e-01 9.26266015e-02
-4.82019410e-02 -4.06655908e-01 -3.45801264e-01 7.40251899e-01
5.79275548e-01 1.97522596e-01 -1.09893635e-01 6.26295328e-01
1.85822859e-01 -4.25429851e-01 -7.81312883e-01 1.03502858e+00
-1.57110840e-01 -1.23162818e+00 -2.22344577e-01 2.75454283e-01
5.19789159e-01 1.26418442e-01 1.17155552e-01 1.82437941e-01
1.38575733e-01 -1.25408125e+00 2.47734293e-01 3.81679058e-01
8.42579067e-01 -6.26157820e-01 1.36001635e+00 3.53680015e-01
-1.24711692e+00 -4.66183177e-04 -1.08017184e-01 5.48130989e-01
1.94016203e-01 6.04511082e-01 -9.77966845e-01 6.13255560e-01
4.40471858e-01 6.34638846e-01 -4.47849959e-01 1.33617687e+00
-1.11866064e-01 5.00582099e-01 -5.02539933e-01 -3.49281691e-02
3.81392837e-01 1.76441018e-02 7.85512924e-02 1.10703897e+00
4.27476615e-01 5.50305367e-01 3.36596012e-01 8.53086412e-01
7.04442784e-02 1.31629303e-01 -6.56836778e-02 -4.63760942e-02
2.10355476e-01 1.55577695e+00 -1.32440245e+00 -3.77575904e-01
-3.49350482e-01 8.76986086e-01 -7.04430044e-02 2.21731037e-01
-1.26918495e+00 -2.07627729e-01 -6.13394007e-02 1.04708999e-01
2.05880031e-01 1.03025250e-01 -5.85753202e-01 -1.02822888e+00
-4.22305614e-01 -4.87316340e-01 4.40939337e-01 -6.97697103e-01
-9.30001199e-01 6.77633107e-01 -1.00224406e-01 -1.09860146e+00
3.38100523e-01 -4.83414441e-01 -8.09313953e-01 8.62114489e-01
-1.47570872e+00 -1.25583076e+00 -7.62457252e-01 6.33776128e-01
2.55685717e-01 1.32982925e-01 5.81846178e-01 4.74942714e-01
-9.03636336e-01 5.88815391e-01 -4.75861691e-02 2.73186028e-01
4.69666749e-01 -9.82276380e-01 -4.75551397e-01 5.52327335e-01
-4.75142300e-01 1.84555858e-01 6.38212934e-02 -5.54704905e-01
-8.14043403e-01 -1.22789562e+00 2.09186211e-01 8.96071717e-02
3.77816319e-01 1.60396472e-01 -8.65351737e-01 6.96235716e-01
4.75782901e-02 2.81418920e-01 7.56584704e-01 -5.57864964e-01
3.27312112e-01 -1.08114071e-01 -1.36729455e+00 4.63974774e-01
4.43753928e-01 -2.29802936e-01 -2.14759827e-01 4.67413932e-01
6.32415056e-01 -7.87180185e-01 -7.84477472e-01 6.49764597e-01
4.14864987e-01 -7.65880048e-01 7.65562236e-01 1.88871492e-02
3.69093895e-01 -4.04776663e-01 6.20302930e-02 -1.28710127e+00
-4.18457985e-01 2.59885520e-01 6.01539254e-01 7.50624239e-01
2.94542432e-01 -8.13473105e-01 8.31999362e-01 5.34119368e-01
-6.30642176e-01 -1.08028567e+00 -8.81288886e-01 -3.39793414e-01
1.23363711e-01 -2.22101599e-01 5.47950804e-01 9.13060129e-01
-1.35155246e-01 5.47317602e-02 1.53470621e-01 1.04197457e-01
1.76788226e-01 -1.28389373e-01 1.25513747e-01 -9.50823367e-01
-2.35831007e-01 -7.20133781e-01 -3.64558727e-01 -8.95791590e-01
-1.32825315e-01 -1.08600247e+00 -1.00422077e-01 -1.61829007e+00
4.84612733e-01 -6.10323131e-01 -5.23905933e-01 7.09991992e-01
-9.86543149e-02 3.56516957e-01 -6.29982501e-02 3.45743597e-01
-3.85577202e-01 6.30188882e-01 1.82584727e+00 -1.49355993e-01
-1.54320553e-01 -8.01630989e-02 -5.30957997e-01 8.44525754e-01
7.99449027e-01 -3.39852542e-01 -3.78129452e-01 -3.46348405e-01
-3.91366601e-01 3.54379863e-01 4.78698611e-01 -1.01089299e+00
3.31802934e-01 -2.30277315e-01 8.19235206e-01 -6.05423212e-01
2.37164587e-01 -1.12807190e+00 1.74880639e-01 7.59388983e-01
5.00939153e-02 -1.08872220e-01 4.11370724e-01 3.26835394e-01
-3.84539127e-01 -7.80692250e-02 1.04432237e+00 -4.71844226e-01
-5.63422799e-01 6.83482349e-01 -3.74519795e-01 -2.09670410e-01
1.46533597e+00 -2.46553794e-01 -2.16403574e-01 -7.04381168e-02
-7.27778852e-01 2.84165561e-01 2.35289350e-01 -1.00889660e-01
5.91376245e-01 -1.39397001e+00 -3.78517985e-01 1.93562761e-01
-3.88413295e-02 1.72898591e-01 5.78542650e-01 1.68012309e+00
-7.38530576e-01 5.26045263e-01 -2.73135871e-01 -8.85382414e-01
-1.01133084e+00 2.50648707e-01 8.12214434e-01 -4.68886018e-01
-5.19944251e-01 8.67537081e-01 6.23953640e-01 -1.89365253e-01
-7.01594874e-02 -6.54503763e-01 -2.70877063e-01 -2.08289847e-01
3.25558782e-01 -9.75632221e-02 1.47760883e-01 -6.99386060e-01
-4.34454173e-01 6.92840695e-01 -3.26827407e-01 -1.14915529e-02
9.60219383e-01 7.46896565e-02 -1.30859002e-01 6.16871752e-02
1.18092906e+00 -3.84650111e-01 -9.52249885e-01 -1.27906144e-01
-3.72135520e-01 -3.32415760e-01 6.07481778e-01 -1.11681128e+00
-1.64869952e+00 9.70975697e-01 9.39828157e-01 -1.29321992e-01
1.37946498e+00 4.37977761e-02 1.06282496e+00 -2.66379863e-01
8.13152716e-02 -7.21161067e-01 7.88689926e-02 2.25010529e-01
5.64841092e-01 -1.45225656e+00 -1.32400841e-01 -6.52710378e-01
-7.83024848e-01 1.16784275e+00 8.80054295e-01 2.29302183e-01
6.54444337e-01 2.19962344e-01 1.32885709e-01 -2.79065967e-01
-1.96200326e-01 -2.02776209e-01 1.67687476e-01 3.40711683e-01
4.21623617e-01 2.80909151e-01 -3.19210529e-01 9.17950869e-01
1.14899822e-01 2.27782667e-01 2.44624943e-01 8.91488671e-01
-4.92780179e-01 -8.74874711e-01 -3.78483802e-01 5.95338583e-01
-4.74659622e-01 -1.65494189e-01 -1.61894172e-01 1.09117019e+00
3.32776934e-01 5.47688484e-01 -3.96976769e-02 -3.18166167e-01
7.35910088e-02 -2.54014105e-01 3.37873548e-01 -5.12934148e-01
-6.10494137e-01 3.84363741e-01 -3.94748718e-01 -2.35285521e-01
-3.26752394e-01 -3.76880467e-01 -1.70084608e+00 -6.71237782e-02
-5.91484964e-01 7.76791498e-02 7.20720828e-01 1.02759743e+00
7.91812912e-02 1.07156944e+00 3.84103656e-01 -6.28688335e-01
-1.61130071e-01 -9.41431761e-01 -6.09370291e-01 1.11918852e-01
1.59679428e-01 -7.95631349e-01 -6.97525963e-02 -3.50833118e-01]
|
[14.576802253723145, -2.4167816638946533]
|
c0fdb65f-ea59-4e03-9ea6-6ebce9d542fb
|
an-empirical-survey-of-unsupervised-text
|
2012.03468
| null |
https://arxiv.org/abs/2012.03468v1
|
https://arxiv.org/pdf/2012.03468v1.pdf
|
An Empirical Survey of Unsupervised Text Representation Methods on Twitter Data
|
The field of NLP has seen unprecedented achievements in recent years. Most notably, with the advent of large-scale pre-trained Transformer-based language models, such as BERT, there has been a noticeable improvement in text representation. It is, however, unclear whether these improvements translate to noisy user-generated text, such as tweets. In this paper, we present an experimental survey of a wide range of well-known text representation techniques for the task of text clustering on noisy Twitter data. Our results indicate that the more advanced models do not necessarily work best on tweets and that more exploration in this area is needed.
|
['Soroush Vosoughi', 'Ruibo Liu', 'Weicheng Ma', 'Jason Wei', 'Chongyang Gao', 'Lili Wang']
|
2020-12-07
| null |
https://aclanthology.org/2020.wnut-1.27
|
https://aclanthology.org/2020.wnut-1.27.pdf
|
emnlp-wnut-2020-11
|
['text-clustering']
|
['natural-language-processing']
|
[ 9.73567292e-02 5.09732068e-02 8.94529149e-02 -4.53433096e-01
-1.03331530e+00 -6.40901864e-01 9.46902692e-01 6.64186180e-01
-4.51857090e-01 5.40803075e-01 8.93570662e-01 -5.02633035e-01
-6.33898079e-02 -5.73629260e-01 -2.36741692e-01 -4.40297991e-01
-6.61778729e-03 8.50316107e-01 9.69530195e-02 -5.59976876e-01
1.53630301e-01 1.84809014e-01 -1.45544517e+00 3.32003266e-01
4.31243211e-01 6.05415702e-01 8.87491461e-03 5.56389153e-01
-6.77057385e-01 7.73595512e-01 -7.79148161e-01 -4.54519004e-01
-2.58867800e-01 -3.21537942e-01 -1.04928195e+00 -7.91287497e-02
1.34898067e-01 2.84612536e-01 -2.64220268e-01 8.38291109e-01
5.64819098e-01 2.85004497e-01 5.18324316e-01 -9.63231325e-01
-2.40824118e-01 1.24288309e+00 -4.40113872e-01 4.24049824e-01
3.12198430e-01 -3.67065787e-01 1.28708005e+00 -6.73929870e-01
5.95908165e-01 1.52290297e+00 7.51494527e-01 1.80938900e-01
-1.21224117e+00 -5.98477185e-01 1.83983669e-01 -4.88889925e-02
-1.42031252e+00 -2.98543215e-01 7.42646992e-01 -1.58765420e-01
1.11953676e+00 4.51216310e-01 2.39177763e-01 1.05723631e+00
3.63903423e-03 9.12027895e-01 1.04860914e+00 -4.82774287e-01
1.41107403e-02 4.48002934e-01 1.87950686e-01 2.72523165e-01
3.49915564e-01 -3.52660775e-01 -4.19663757e-01 -4.90129501e-01
-1.63006242e-02 -2.24915400e-01 -1.24807827e-01 9.07357037e-02
-1.04550838e+00 1.18343222e+00 1.27065361e-01 1.02930498e+00
-1.28722295e-01 2.83668600e-02 8.07688057e-01 2.09734261e-01
1.20195913e+00 5.71342647e-01 -4.32075381e-01 -6.61912739e-01
-1.31740141e+00 1.20802961e-01 1.03608227e+00 7.40056992e-01
6.66971743e-01 3.59589458e-02 1.97625667e-01 8.49880636e-01
1.23494588e-01 2.10824683e-01 7.44969070e-01 -6.35565937e-01
8.09838474e-01 5.17075360e-01 -7.04619214e-02 -1.25102890e+00
-5.91699779e-01 -4.47508723e-01 -7.47295201e-01 -3.82235914e-01
1.77348822e-01 -1.76790595e-01 -7.37403810e-01 1.35995400e+00
3.62195559e-02 -2.79534720e-02 -1.64267458e-02 4.30401295e-01
6.58138931e-01 9.17529106e-01 1.56661078e-01 -1.77551925e-01
9.91500974e-01 -5.31137586e-01 -7.79817641e-01 -3.57514441e-01
7.38869786e-01 -1.10537565e+00 8.60391736e-01 3.14330161e-01
-9.29651499e-01 -1.82817295e-01 -5.63640058e-01 -3.79161686e-02
-8.54457736e-01 -2.75234759e-01 7.83273101e-01 9.36930299e-01
-1.07062197e+00 4.94981170e-01 -8.17552626e-01 -9.41286147e-01
2.91352779e-01 3.48275572e-01 -1.84764832e-01 -9.38285422e-03
-1.27922082e+00 8.90605330e-01 3.95063043e-01 3.26696113e-02
-3.08141798e-01 -4.11225587e-01 -6.57751262e-01 7.80283287e-02
1.78562731e-01 -2.56432980e-01 1.28662622e+00 -7.64436543e-01
-1.25998020e+00 8.49116266e-01 -4.98971373e-01 -6.32343411e-01
3.79963338e-01 -1.36916950e-01 -4.38878447e-01 -1.79902613e-01
-1.21336374e-02 3.85136962e-01 4.44253176e-01 -1.40795267e+00
-7.51901150e-01 -3.09817553e-01 -2.28248939e-01 3.22825909e-01
-7.70023286e-01 5.99765241e-01 -1.94292709e-01 -6.71115816e-01
-1.46117672e-01 -8.42780292e-01 -4.18749362e-01 -6.40439212e-01
-3.42141598e-01 -6.41989350e-01 8.23906600e-01 -4.01669204e-01
1.29254901e+00 -2.16531515e+00 -2.04425931e-01 3.88883978e-01
1.58574611e-01 1.73230380e-01 1.75850336e-02 9.74391043e-01
-6.40859595e-04 5.00399768e-01 -6.73297495e-02 -6.61276460e-01
1.59888312e-01 4.12726879e-01 -8.00059915e-01 4.78881538e-01
-9.22428742e-02 6.37992322e-01 -1.03514147e+00 -6.87272787e-01
3.52087706e-01 5.87002039e-01 -1.74286887e-01 -2.49384597e-01
-4.01176602e-01 1.33399561e-01 -6.00057781e-01 1.87019825e-01
2.91884840e-01 -2.48611197e-01 2.19849646e-01 2.84545749e-01
-2.73218006e-01 8.45610857e-01 -1.00337100e+00 1.48747563e+00
-3.32870483e-01 1.03802633e+00 1.31109685e-01 -1.10189724e+00
7.28287697e-01 5.07215381e-01 7.07131445e-01 -4.14568931e-01
4.27745223e-01 -4.23124693e-02 -6.59851059e-02 -2.12249637e-01
1.15525949e+00 -2.35523522e-01 -1.77414075e-01 7.48487890e-01
-1.39508009e-01 -3.29161078e-01 4.33033407e-01 4.48431790e-01
1.05985105e+00 -4.19084460e-01 1.43809840e-01 -2.75944203e-01
2.86735117e-01 4.56291676e-01 -3.07183992e-02 8.36901724e-01
-5.67526817e-02 6.86640620e-01 1.14762969e-01 -1.07261471e-01
-8.44288468e-01 -5.24989545e-01 -2.09568962e-01 1.33269262e+00
-2.88141370e-01 -1.00400758e+00 -5.25151074e-01 -5.23691893e-01
-1.33020788e-01 6.93795264e-01 -4.07269835e-01 1.31280735e-01
-7.10220873e-01 -1.16597319e+00 8.34767103e-01 4.09867734e-01
2.80796569e-02 -7.78231561e-01 -2.42210329e-02 4.58131254e-01
-4.33329850e-01 -1.09678078e+00 -2.45137170e-01 3.72364283e-01
-9.65151668e-01 -5.97665191e-01 -4.41980511e-01 -8.47065091e-01
4.47571903e-01 4.85709280e-01 1.20175099e+00 -1.39802128e-01
-3.92810069e-02 4.76764977e-01 -7.10808337e-01 -6.24188006e-01
-5.74015200e-01 4.51858461e-01 -1.48893505e-01 -1.92378521e-01
6.84402466e-01 -4.07899827e-01 -6.60989359e-02 2.05407180e-02
-1.04312575e+00 -4.77097750e-01 3.24100882e-01 5.75734735e-01
2.14963257e-01 7.11696506e-01 3.30228329e-01 -1.08792102e+00
1.04628468e+00 -7.63410866e-01 -8.50738958e-02 -1.37156881e-02
-5.83464086e-01 -3.22829746e-02 8.18605959e-01 -3.52906078e-01
-9.95870471e-01 -1.71288684e-01 -4.69726861e-01 1.13744952e-01
-1.83568850e-01 1.04488277e+00 3.43953043e-01 1.79469243e-01
8.62131596e-01 3.01405907e-01 -3.51217568e-01 -5.61516523e-01
4.18248355e-01 1.06741595e+00 1.07876256e-01 -4.07978207e-01
7.86601067e-01 6.39650404e-01 -4.73237365e-01 -1.25381982e+00
-8.59032154e-01 -9.38733995e-01 -2.07604066e-01 8.75985995e-02
4.07369256e-01 -8.70741606e-01 -2.63506591e-01 -6.79128394e-02
-9.68064845e-01 -2.65595794e-01 -3.51573288e-01 3.89020503e-01
-2.09934115e-01 5.02156675e-01 -6.17452383e-01 -9.16245043e-01
-4.32155401e-01 -7.55539179e-01 1.15859795e+00 -1.38909146e-01
-7.92556882e-01 -1.38210785e+00 2.74735481e-01 3.22125494e-01
6.37407243e-01 -1.17265247e-01 8.24952543e-01 -1.02805614e+00
-4.34136242e-02 -4.64209229e-01 2.42038779e-02 1.67470545e-01
1.41127601e-01 1.24851994e-01 -9.87150431e-01 -2.86779881e-01
-3.90752964e-03 -3.38163763e-01 7.79494882e-01 2.55865782e-01
1.08887696e+00 -2.34310210e-01 -6.33417428e-01 8.23344886e-02
1.17850101e+00 -8.34574476e-02 4.60990608e-01 2.80863106e-01
4.91159946e-01 4.84651804e-01 4.25477237e-01 2.89465159e-01
6.55746937e-01 4.91251767e-01 1.37378201e-01 -9.66260061e-02
9.76071693e-03 -2.56410331e-01 2.54389524e-01 1.15537667e+00
2.29753315e-01 -6.55641377e-01 -1.18033648e+00 7.20151842e-01
-1.68045926e+00 -1.09307134e+00 -2.70611703e-01 1.71596491e+00
9.59690034e-01 1.91893131e-01 1.97458342e-01 5.57468057e-01
7.11338699e-01 3.51709217e-01 3.10523715e-02 -3.58789057e-01
-2.17248812e-01 2.16393650e-01 3.96867543e-01 4.65661615e-01
-1.10740840e+00 9.34377372e-01 7.50569439e+00 8.73920619e-01
-1.21663892e+00 1.97742343e-01 5.63775957e-01 -3.55022436e-04
-4.97586966e-01 -2.69853864e-02 -7.75006175e-01 5.39814770e-01
1.24058998e+00 -5.90569854e-01 1.86586440e-01 7.15468585e-01
4.31676328e-01 -5.64509369e-02 -7.25109935e-01 8.51812959e-01
2.81553477e-01 -1.08642089e+00 3.65136750e-02 -2.74842940e-02
6.56419337e-01 5.61859488e-01 -1.41757485e-02 4.00895089e-01
7.07646370e-01 -1.00195527e+00 5.22542417e-01 -6.18153773e-02
2.67533809e-01 -7.35942602e-01 8.58374536e-01 4.39711362e-01
-1.02392125e+00 1.24989502e-01 -4.96464759e-01 -9.86804515e-02
2.52577603e-01 8.66405129e-01 -1.01377702e+00 5.34577906e-01
7.24923670e-01 8.02083731e-01 -6.09078646e-01 9.64093506e-01
1.35942623e-01 1.22392905e+00 -7.72594988e-01 -3.87295187e-01
4.62607861e-01 9.63445827e-02 4.17839855e-01 1.73301935e+00
1.26415759e-01 -4.36835550e-02 2.85463810e-01 3.98053169e-01
-1.95340946e-01 4.13558632e-01 -7.22237408e-01 -4.48748857e-01
3.57942641e-01 1.28878582e+00 -8.95939708e-01 -4.51744944e-01
-4.45449114e-01 5.14509797e-01 1.96861565e-01 1.80612057e-01
-5.81134021e-01 -2.26351082e-01 4.89447236e-01 1.35290757e-01
1.26674801e-01 -3.54609847e-01 -4.33909535e-01 -9.87710595e-01
-2.82421082e-01 -7.40650654e-01 5.70465088e-01 -4.82277781e-01
-1.40537548e+00 6.25668824e-01 4.07178886e-02 -7.98120439e-01
-3.99961829e-01 -2.99191177e-01 -5.89866221e-01 6.17831230e-01
-1.33095706e+00 -8.11070263e-01 7.63428360e-02 6.09385610e-01
5.19491076e-01 -1.55531219e-03 9.36958253e-01 4.34108287e-01
-2.59196967e-01 3.77524942e-01 5.43609738e-01 1.90859720e-01
6.55632496e-01 -1.23908877e+00 4.55749601e-01 6.50195837e-01
5.04383743e-01 5.62266350e-01 1.16005254e+00 -5.43415666e-01
-1.23644316e+00 -1.17100656e+00 1.40473711e+00 -7.47592032e-01
9.96955574e-01 -4.93423820e-01 -7.91386127e-01 9.33932006e-01
3.88766199e-01 -4.51611340e-01 7.95426846e-01 4.28368717e-01
-2.38049895e-01 -1.53485583e-02 -9.02277708e-01 4.43518728e-01
7.01320291e-01 -6.04985952e-01 -8.03261101e-01 7.06970692e-01
5.18742919e-01 -2.42050603e-01 -6.89653754e-01 -3.62988450e-02
-1.85848922e-02 -5.39283276e-01 7.32276440e-01 -3.72998655e-01
-1.29705081e-02 -6.27817586e-02 3.15568014e-03 -1.36624312e+00
-2.32552856e-01 -7.25314319e-01 2.88676441e-01 1.47368181e+00
5.52254200e-01 -6.23147845e-01 9.69376445e-01 3.87944490e-01
-4.11248468e-02 -2.87867099e-01 -9.17442977e-01 -6.57339394e-01
4.39712763e-01 -7.81193495e-01 3.35738719e-01 1.12243974e+00
6.21469021e-01 6.53622210e-01 -2.23340213e-01 -2.76117235e-01
2.66295642e-01 9.35967267e-03 5.30140817e-01 -1.44326508e+00
1.23579942e-01 -5.37416399e-01 -3.62845570e-01 -9.21497762e-01
3.68499458e-01 -1.02221584e+00 2.54291832e-01 -1.76405752e+00
1.88055694e-01 -4.32879001e-01 -5.42081073e-02 4.05939579e-01
-1.00756422e-01 3.31956416e-01 -2.22163200e-02 1.23928219e-01
-6.95017099e-01 3.71387362e-01 6.13478661e-01 -3.01634938e-01
-1.54110566e-01 8.18634555e-02 -1.00271738e+00 5.71520865e-01
1.20919049e+00 -8.76428485e-01 -4.29448634e-01 -4.52470630e-01
5.23866594e-01 -4.56424713e-01 -2.06316859e-01 -7.92187333e-01
3.62231880e-01 1.58889666e-01 1.80301424e-02 -6.62444234e-01
4.07637686e-01 -8.49557757e-01 -1.12511031e-01 8.48368183e-02
-4.56155956e-01 2.28407621e-01 3.43606621e-01 4.85253632e-01
-5.11656582e-01 -1.41442984e-01 4.79965210e-01 -1.75610304e-01
-2.05030635e-01 -1.06467614e-02 -1.01658964e+00 2.91413426e-01
6.03969336e-01 1.73647925e-02 -2.44605497e-01 -7.37483382e-01
-4.32704031e-01 8.75339955e-02 2.15834349e-01 5.91976464e-01
3.43201548e-01 -8.46098781e-01 -7.80142307e-01 -3.69929731e-01
-3.03709036e-04 -1.59041077e-01 -2.26541445e-01 6.71010792e-01
-2.73822844e-02 9.02417481e-01 5.08352399e-01 -3.95981878e-01
-1.29938114e+00 3.24036568e-01 1.27597436e-01 -4.53649253e-01
-6.53796017e-01 7.08425522e-01 -3.33507597e-01 -4.58257943e-01
3.46062064e-01 -3.65891695e-01 -2.94506103e-01 2.85310030e-01
6.08454823e-01 3.36951166e-01 1.71405718e-01 -8.14337730e-01
-3.14015865e-01 2.39604995e-01 -1.70278773e-01 -2.66826779e-01
1.50563240e+00 -3.15208733e-01 -2.89759219e-01 8.03824246e-01
1.18159175e+00 4.91153523e-02 -3.77317816e-01 -2.87559092e-01
4.28135693e-01 -1.08562708e-01 2.23232776e-01 -5.35970688e-01
-7.97238469e-01 7.93918073e-01 1.98456943e-01 7.80175805e-01
8.37430775e-01 2.05443814e-01 7.61197448e-01 5.71324110e-01
2.84982145e-01 -1.29013944e+00 -2.28957087e-01 9.92794991e-01
5.49488604e-01 -1.16166055e+00 3.44179906e-02 -1.61755517e-01
-4.99440193e-01 9.11936820e-01 -9.15811025e-03 1.46531224e-01
8.49901378e-01 4.05315489e-01 9.78667960e-02 -3.51731509e-01
-8.39090645e-01 -4.86469060e-01 -7.33635873e-02 4.78587359e-01
8.91510487e-01 1.81283185e-03 -2.78589666e-01 1.76449507e-01
-7.04834163e-01 -1.80367500e-01 6.05335832e-01 9.88335669e-01
-6.30785286e-01 -1.13674760e+00 -5.36636889e-01 6.34453416e-01
-7.78675795e-01 -2.48879641e-01 -6.75148904e-01 8.23166132e-01
-2.20162988e-01 1.42148483e+00 1.23774305e-01 -2.51877725e-01
1.82691708e-01 3.14852685e-01 1.21699236e-01 -9.23475027e-01
-8.66783440e-01 4.61411849e-02 4.71113056e-01 -2.42383540e-01
-6.29659891e-01 -8.96959722e-01 -1.26930809e+00 -7.23997831e-01
-4.40528870e-01 6.87683940e-01 8.74663591e-01 1.01608896e+00
1.33797392e-01 2.60328114e-01 3.60523552e-01 -8.31106305e-01
-3.82421702e-01 -1.15879607e+00 -5.24053276e-01 2.50573307e-01
1.61807373e-01 -1.23279706e-01 -5.99112928e-01 2.27125734e-02]
|
[10.482566833496094, 8.354775428771973]
|
db72afc6-4505-4807-9308-cb4121a5d438
|
on-merging-feature-engineering-and-deep
|
2207.06096
| null |
https://arxiv.org/abs/2207.06096v2
|
https://arxiv.org/pdf/2207.06096v2.pdf
|
On Merging Feature Engineering and Deep Learning for Diagnosis, Risk-Prediction and Age Estimation Based on the 12-Lead ECG
|
Objective: Machine learning techniques have been used extensively for 12-lead electrocardiogram (ECG) analysis. For physiological time series, deep learning (DL) superiority to feature engineering (FE) approaches based on domain knowledge is still an open question. Moreover, it remains unclear whether combining DL with FE may improve performance. Methods: We considered three tasks intending to address these research gaps: cardiac arrhythmia diagnosis (multiclass-multilabel classification), atrial fibrillation risk prediction (binary classification), and age estimation (regression). We used an overall dataset of 2.3M 12-lead ECG recordings to train the following models for each task: i) a random forest taking the FE as input was trained as a classical machine learning approach; ii) an end-to-end DL model; and iii) a merged model of FE+DL. Results: FE yielded comparable results to DL while necessitating significantly less data for the two classification tasks and it was outperformed by DL for the regression task. For all tasks, merging FE with DL did not improve performance over DL alone. Conclusion: We found that for traditional 12-lead ECG based diagnosis tasks DL did not yield a meaningful improvement over FE, while it improved significantly the nontraditional regression task. We also found that combining FE with DL did not improve over DL alone which suggests that the FE were redundant with the features learned by DL. Significance: Our findings provides important recommendations on what machine learning strategy and data regime to chose with respect to the task at hand for the development of new machine learning models based on the 12-lead ECG.
|
['Joachim A. Behar', 'Antonio Luiz P. Ribeiro', 'Antônio H. Ribeiro', 'Jesse Read', 'Eran Zvuloni']
|
2022-07-13
| null | null | null | null |
['age-estimation', 'age-estimation']
|
['computer-vision', 'miscellaneous']
|
[ 2.35397622e-01 2.83973217e-02 -2.71875202e-03 -4.09157962e-01
-9.72714365e-01 -4.27274436e-01 2.21905008e-01 4.98495996e-01
-5.41166008e-01 9.03259099e-01 -1.20426200e-01 -7.74633050e-01
-4.67358261e-01 -4.29480731e-01 -3.01057011e-01 -6.74685955e-01
-2.83536911e-01 7.51523197e-01 -3.18657368e-01 3.14463109e-01
2.72277761e-02 6.18480384e-01 -1.23250210e+00 3.91222328e-01
7.51096189e-01 1.14613569e+00 -2.35671565e-01 7.50158012e-01
3.39311391e-01 1.05170751e+00 -7.71324456e-01 -7.38166869e-02
1.51788697e-01 -6.45786345e-01 -8.30793798e-01 -2.07520574e-01
2.60055274e-01 -1.38099074e-01 3.45413536e-01 2.16623366e-01
1.16273105e+00 -2.76752502e-01 7.23622382e-01 -1.02258658e+00
2.77713053e-02 5.16601920e-01 -3.53497684e-01 4.01038587e-01
1.97828233e-01 -3.00100278e-02 7.25176334e-01 -7.35480547e-01
4.12028611e-01 7.66850054e-01 1.23870277e+00 3.05667192e-01
-1.36001480e+00 -5.97996712e-01 -2.32499555e-01 1.32449612e-01
-1.27620518e+00 -2.64381200e-01 6.76935315e-01 -7.65822828e-01
1.07589674e+00 2.95697987e-01 8.43770921e-01 8.07895243e-01
4.50550795e-01 2.73315430e-01 1.53197467e+00 -5.53907752e-01
8.33961472e-04 3.21078539e-01 3.87309432e-01 7.35556841e-01
3.43110025e-01 3.88533443e-01 -2.55507648e-01 -7.04526484e-01
4.94137287e-01 -2.25359052e-01 -1.35159090e-01 -4.99133542e-02
-1.25185692e+00 7.77550936e-01 2.13851165e-02 4.90060151e-01
-5.82838535e-01 -6.40215501e-02 6.79346979e-01 6.98227048e-01
4.68761981e-01 8.75823498e-01 -1.02624130e+00 -2.62816519e-01
-1.20178318e+00 3.03438067e-01 9.15366888e-01 1.02061741e-01
5.82785547e-01 1.88518852e-01 -2.18691677e-01 8.31356943e-01
2.74273083e-02 3.91572028e-01 4.22823995e-01 -9.01348770e-01
7.15961158e-02 4.66456234e-01 -7.73126632e-02 -8.83701205e-01
-1.06024361e+00 -9.80576634e-01 -8.75626147e-01 2.34712020e-01
6.98201299e-01 -7.45734572e-01 -7.39524305e-01 1.62161529e+00
-2.02052251e-01 1.02027558e-01 -8.90913531e-02 7.55491376e-01
7.66881883e-01 1.07917123e-01 2.52261311e-01 -4.76868778e-01
1.21311986e+00 -2.95038104e-01 -5.94741285e-01 -7.09081441e-02
1.19385624e+00 -6.41941547e-01 6.94736898e-01 8.76371324e-01
-8.43442857e-01 -6.14767790e-01 -9.73203123e-01 2.84746319e-01
-4.99672145e-02 4.66298461e-01 6.00936294e-01 9.89249468e-01
-9.78940189e-01 9.00823593e-01 -8.28428745e-01 -3.12840670e-01
4.17558908e-01 5.62698305e-01 -3.59324843e-01 2.28406385e-01
-1.33217049e+00 1.29084110e+00 1.06635109e-01 2.98837340e-03
-5.30300021e-01 -7.55575299e-01 -7.21262872e-01 -2.85001963e-01
5.00146151e-02 -9.26172376e-01 8.03820372e-01 -1.08413196e+00
-1.09534609e+00 1.07139099e+00 1.77385416e-02 -7.29490399e-01
4.77569401e-01 -3.33813488e-01 -3.51147741e-01 5.70071675e-02
1.17683984e-01 3.11902374e-01 7.80883729e-01 -8.84597182e-01
-5.82510710e-01 -5.40296018e-01 -1.73371419e-01 1.70548670e-02
4.64953408e-02 4.81415913e-02 4.03816402e-01 -7.49456882e-01
-7.41210878e-02 -9.92723107e-01 -2.25659072e-01 -3.10015619e-01
-2.92093754e-02 -2.61417836e-01 5.12758553e-01 -9.98344719e-01
1.52505434e+00 -2.02428627e+00 8.03434923e-02 2.38611042e-01
6.26742542e-01 4.00781691e-01 1.87134579e-01 3.03272247e-01
-4.76484746e-01 1.78969547e-01 -2.64015138e-01 -2.29400378e-02
-4.19167876e-01 3.47574763e-02 1.05796754e-01 6.02742970e-01
1.28082246e-01 8.92962694e-01 -5.55127561e-01 -4.42373544e-01
3.08080763e-01 3.46315175e-01 -3.39659303e-01 4.74000014e-02
3.09761345e-01 6.39442444e-01 2.03198772e-02 4.34191555e-01
2.99633801e-01 -1.77144185e-01 2.47836336e-01 -2.81088531e-01
1.22998364e-01 4.30893898e-02 -8.70614588e-01 1.47450888e+00
-4.06728685e-01 5.20451546e-01 -3.73631835e-01 -1.21052134e+00
1.27635145e+00 6.52007818e-01 7.85837591e-01 -3.44567150e-01
9.98535380e-02 4.70009685e-01 5.38971364e-01 -6.07948244e-01
-5.86512029e-01 -6.13569319e-01 -7.80773982e-02 5.36404252e-01
6.25955015e-02 -2.32355925e-03 -2.36084312e-01 -1.79536700e-01
1.18151045e+00 1.22981891e-01 4.37275261e-01 -4.67117935e-01
4.63703007e-01 -1.30360976e-01 7.74079204e-01 9.67739761e-01
-2.23165140e-01 6.73660159e-01 6.34305060e-01 -9.19977367e-01
-5.55320024e-01 -7.93123305e-01 -3.83262634e-01 5.50252378e-01
-5.22090375e-01 -4.93822515e-01 -5.86477518e-01 -9.28109705e-01
-4.92864363e-02 4.69278872e-01 -6.15932405e-01 -2.35427901e-01
-5.33091486e-01 -1.33127832e+00 9.21420395e-01 6.58134222e-01
1.28196493e-01 -1.02293622e+00 -1.16664982e+00 3.84469062e-01
-1.54227480e-01 -7.43028522e-01 1.13043942e-01 7.17503130e-01
-1.20584726e+00 -1.22456348e+00 -7.74544597e-01 -5.64579785e-01
2.51563072e-01 -6.45090222e-01 1.23503995e+00 8.42811167e-02
-4.85156000e-01 3.11513871e-01 -3.16071361e-01 -9.22339916e-01
-3.60781223e-01 2.20821172e-01 1.03464380e-01 -2.77376566e-02
5.13460219e-01 -7.50073493e-01 -6.18637681e-01 -5.64892730e-03
-4.64631259e-01 -1.41897678e-01 7.80396760e-01 9.54253972e-01
2.88928390e-01 -2.03316569e-01 1.15204167e+00 -1.19468224e+00
7.03025460e-01 -2.41570562e-01 1.04733787e-01 1.67635158e-02
-1.11044908e+00 -1.69537559e-01 4.21091378e-01 -1.99962482e-01
-4.87556636e-01 1.76027879e-01 -2.64369845e-01 -2.99528450e-01
-3.39142829e-01 8.78983140e-01 1.05775304e-01 3.33184451e-02
9.36606407e-01 -2.39159748e-01 3.04566711e-01 -4.65737373e-01
-2.70446956e-01 7.50392377e-01 1.10211462e-01 -6.51972950e-01
3.19864720e-01 1.01415999e-01 3.13094676e-01 -5.47335386e-01
-7.71855116e-01 -2.49118358e-01 -9.15205240e-01 -2.23446582e-02
8.90736639e-01 -8.78895819e-01 -2.48288348e-01 4.33996618e-01
-8.84597719e-01 -3.20595056e-01 -3.63739789e-01 7.68909097e-01
-5.71993172e-01 3.48581225e-01 -4.50903773e-01 -8.43910277e-01
-6.97815955e-01 -9.39001739e-01 7.93340266e-01 -3.27395707e-01
-8.43785346e-01 -1.15945613e+00 1.31335124e-01 2.98000455e-01
3.60969484e-01 7.87725031e-01 1.31964660e+00 -9.45388556e-01
3.53576720e-01 -3.80932450e-01 3.65683436e-02 6.46185577e-01
3.63077432e-01 -2.81645209e-01 -1.10538018e+00 -3.65320593e-01
2.52021283e-01 -2.34800056e-01 7.07585335e-01 6.44285560e-01
1.11282086e+00 1.32305533e-01 -1.01283625e-01 5.24613738e-01
1.19632649e+00 4.48255986e-01 5.44334471e-01 2.62027979e-01
5.38455129e-01 4.44770455e-01 5.31378925e-01 4.10498321e-01
2.54625827e-01 5.44861436e-01 -1.42888442e-01 -6.82932794e-01
8.59210920e-03 3.53564680e-01 1.95038274e-01 4.69377458e-01
-4.93280947e-01 3.25917691e-01 -1.27213478e+00 4.13069695e-01
-1.78924119e+00 -6.71864092e-01 -3.73889446e-01 2.39469743e+00
5.22722304e-01 2.76837140e-01 4.70330536e-01 7.02948034e-01
3.43105793e-01 -2.20922381e-01 -4.10385758e-01 -7.37015665e-01
-1.86685652e-01 5.73832393e-01 1.57150626e-01 1.64331436e-01
-1.20162594e+00 1.89178810e-01 6.31655550e+00 3.65489841e-01
-1.43929076e+00 1.74782231e-01 9.09488082e-01 1.09922014e-01
1.68524489e-01 6.44408464e-02 -3.95896703e-01 2.53765941e-01
1.29885340e+00 8.12286884e-02 4.70299125e-02 4.48675185e-01
3.83672118e-01 -6.43339306e-02 -1.23177874e+00 1.11840665e+00
9.85636115e-02 -1.08378565e+00 -2.88785309e-01 -3.14905830e-02
2.68639266e-01 -7.17941150e-02 -3.00187349e-01 4.38828737e-01
-4.82931256e-01 -1.08551800e+00 3.47950876e-01 6.61032200e-01
1.04449689e+00 -6.17936075e-01 1.21829426e+00 4.51648653e-01
-7.27406800e-01 -1.81343406e-01 1.74852118e-01 -3.53876323e-01
2.09938213e-02 8.89585137e-01 -9.28248525e-01 8.20515394e-01
7.54212022e-01 9.19233739e-01 -7.71849513e-01 9.23786104e-01
1.29876971e-01 1.06291771e+00 -1.34599894e-01 3.69685471e-01
-5.02093323e-02 1.97191015e-01 6.38044596e-01 1.20885551e+00
1.30309239e-01 -1.77735373e-01 2.29108438e-01 5.00939608e-01
5.20529509e-01 2.89465249e-01 -7.01065600e-01 1.01669677e-01
1.24463901e-01 1.11616743e+00 -7.36166000e-01 -4.28427011e-01
-5.26680410e-01 5.95924914e-01 7.40983849e-03 1.06423207e-01
-6.32881403e-01 -5.90150058e-01 7.56575912e-02 6.65332913e-01
-1.68326095e-01 3.45884681e-01 -7.59025335e-01 -9.58114207e-01
-1.14212260e-01 -1.09076381e+00 7.16352582e-01 -4.83043909e-01
-1.40528023e+00 8.19912255e-01 -4.51930687e-02 -1.16496420e+00
-4.71334964e-01 -5.33676207e-01 -3.11756402e-01 1.23315787e+00
-1.19904399e+00 -9.53626752e-01 -3.01360134e-02 3.29278827e-01
1.69620812e-01 -2.16789365e-01 1.32923555e+00 5.72435975e-01
-3.20914328e-01 5.20155013e-01 -3.41028005e-01 1.79914549e-01
1.02898026e+00 -1.52097821e+00 -2.26620063e-01 4.92420048e-01
4.50792834e-02 4.72550660e-01 3.07082087e-01 -6.75248504e-01
-7.06114471e-01 -1.08007503e+00 1.27491891e+00 -7.22905993e-01
1.00620851e-01 6.21505454e-02 -7.81462014e-01 4.63763475e-01
-1.31924972e-01 3.23459245e-02 1.11762798e+00 3.54396164e-01
1.28146991e-01 -2.18383402e-01 -1.12162936e+00 -1.60500128e-02
5.57411730e-01 -3.96930307e-01 -6.46779120e-01 4.83079404e-02
3.13580707e-02 -1.93923056e-01 -1.31617510e+00 9.12680149e-01
8.43038321e-01 -8.12000036e-01 8.73176336e-01 -5.88104725e-01
1.88709527e-01 -1.84492365e-01 1.02902293e-01 -1.16406322e+00
-4.98590976e-01 -4.70781416e-01 -1.71339408e-01 8.23957980e-01
6.69832766e-01 -8.09911311e-01 5.39403081e-01 5.71546972e-01
-7.39902258e-02 -1.20179236e+00 -6.96314335e-01 -5.28407931e-01
3.13671887e-01 -4.46489125e-01 2.12430313e-01 1.05766761e+00
-2.52020627e-01 7.31467545e-01 -4.63332236e-01 -1.32987946e-01
3.94945353e-01 1.82851434e-01 3.55308801e-01 -1.85759819e+00
-5.32454312e-01 -2.38384098e-01 -5.41416943e-01 -1.19963855e-01
6.13919534e-02 -1.19447017e+00 -4.82932210e-01 -1.54343164e+00
1.30937755e-01 -7.60300517e-01 -7.32890904e-01 7.21287131e-01
-4.53359544e-01 3.09421927e-01 4.20671386e-05 9.28284749e-02
1.48991933e-02 -1.98381245e-01 7.49731541e-01 1.81410700e-01
-3.80511373e-01 5.05097687e-01 -9.03343618e-01 6.75490320e-01
9.27953959e-01 -6.68355763e-01 -4.41123694e-01 -1.64026186e-01
1.17769405e-01 4.67010975e-01 4.97337371e-01 -1.04257929e+00
-2.23255575e-01 4.11781847e-01 7.90872991e-01 -2.61903524e-01
9.47754551e-03 -6.61073387e-01 3.19652319e-01 7.47863889e-01
-2.22063079e-01 3.17001462e-01 3.05349052e-01 2.06769571e-01
-5.67672960e-02 -1.42148733e-01 6.72497928e-01 -6.63481876e-02
-1.84571758e-01 -1.22092897e-03 -7.18117118e-01 8.32630172e-02
8.52173984e-01 -2.75685906e-01 2.81124949e-01 -2.76725352e-01
-1.45873952e+00 -1.23069689e-01 -7.53482655e-02 2.84146875e-01
4.47757035e-01 -8.31113160e-01 -1.07494259e+00 1.69214651e-01
-5.14716208e-02 -2.41776749e-01 1.06187470e-01 1.43477356e+00
-6.29183173e-01 4.33865130e-01 -1.51409552e-01 -7.75844753e-01
-1.43948388e+00 4.30467755e-01 7.37963974e-01 -5.84194601e-01
-6.80863023e-01 6.31934524e-01 -1.78913131e-01 -4.52653736e-01
3.07672381e-01 -2.78479040e-01 -3.67671698e-01 3.70656133e-01
2.40887940e-01 5.78817904e-01 5.87683380e-01 -2.47228846e-01
-5.62375963e-01 6.23661518e-01 1.79120779e-01 -1.27512449e-02
1.34791982e+00 2.23909412e-02 -1.46044537e-01 8.84948730e-01
8.99892151e-01 -2.69701004e-01 -5.09193480e-01 3.77412885e-02
1.98998541e-01 -1.01939321e-03 2.52597064e-01 -1.42048883e+00
-1.00045383e+00 1.14031625e+00 1.01737034e+00 -6.78989813e-02
1.39294922e+00 -3.43055606e-01 2.97145486e-01 2.47266874e-01
4.71929967e-01 -7.80462921e-01 -3.77136648e-01 3.06933839e-03
7.32899249e-01 -1.14218283e+00 1.82300031e-01 -1.19818307e-01
-8.05829227e-01 1.21519089e+00 3.05442393e-01 -2.94338483e-02
1.12396133e+00 3.24441910e-01 4.25846100e-01 -3.02412748e-01
-7.69476414e-01 -5.43848090e-02 3.73910278e-01 6.14743710e-01
7.47168303e-01 1.48326457e-01 -7.78850734e-01 1.11050403e+00
4.10387591e-02 4.80251431e-01 1.79339617e-01 9.21121299e-01
1.33880572e-02 -1.43456435e+00 -1.95859909e-01 1.21390355e+00
-9.04923975e-01 -1.51283652e-01 -4.22277242e-01 5.98095715e-01
7.51535475e-01 9.80320454e-01 -3.34908366e-01 -4.44047451e-01
2.83268362e-01 4.84743774e-01 5.67641079e-01 -8.16576242e-01
-1.08248568e+00 1.98404953e-01 3.62293452e-01 -1.58575624e-01
-4.33510840e-01 -9.04359698e-01 -8.93599391e-01 2.45498896e-01
-4.24442768e-01 3.59670341e-01 2.76562661e-01 1.00208414e+00
3.41562241e-01 7.42893279e-01 5.53459942e-01 -4.52541709e-01
-4.43061769e-01 -1.14301014e+00 -6.30991638e-01 1.19476020e-01
4.33155894e-01 -5.31554580e-01 -4.72524285e-01 1.64750710e-01]
|
[14.307344436645508, 3.3087663650512695]
|
52a97910-5db0-45dc-a192-4a443ba09158
|
temporal-scale-estimation-for-oversampled
|
2109.10937
| null |
https://arxiv.org/abs/2109.10937v1
|
https://arxiv.org/pdf/2109.10937v1.pdf
|
Temporal Scale Estimation for Oversampled Network Cascades: Theory, Algorithms, and Experiment
|
Spreading processes on graphs arise in a host of application domains, from the study of online social networks to viral marketing to epidemiology. Various discrete-time probabilistic models for spreading processes have been proposed. These are used for downstream statistical estimation and prediction problems, often involving messages or other information that is transmitted along with infections caused by the process. It is thus important to design models of cascade observation that take into account phenomena that lead to uncertainty about the process state at any given time. We highlight one such phenomenon -- temporal distortion -- caused by a misalignment between the rate at which observations of a cascade process are made and the rate at which the process itself operates, and argue that failure to correct for it results in degradation of performance on downstream statistical tasks. To address these issues, we formulate the clock estimation problem in terms of a natural distortion measure. We give a clock estimation algorithm, which we call FastClock, that runs in linear time in the size of its input and is provably statistically accurate for a broad range of model parameters when cascades are generated from the independent cascade process with known parameters and when the underlying graph is Erd\H{o}s-R\'enyi. We further give empirical results on the performance of our algorithm in comparison to the state of the art estimator, a likelihood proxy maximization-based estimator implemented via dynamic programming. We find that, in a broad parameter regime, our algorithm substantially outperforms the dynamic programming algorithm in terms of both running time and accuracy.
|
['Petko Bogdanov', 'Carolyn Kaminski', 'Abram Magner']
|
2021-09-22
| null | null | null | null |
['epidemiology']
|
['medical']
|
[ 4.51707602e-01 1.16335824e-01 -5.13191335e-02 -5.39996177e-02
-4.13474858e-01 -6.57189667e-01 1.02597618e+00 3.62735301e-01
-3.73051554e-01 4.36593533e-01 4.06909585e-02 -3.89371693e-01
-2.16853857e-01 -9.05840278e-01 -6.48904920e-01 -6.75037801e-01
-5.40900946e-01 1.04784358e+00 3.52769703e-01 8.65122229e-02
1.13689728e-01 2.41949648e-01 -7.39648402e-01 -2.88237095e-01
3.91723037e-01 6.14904940e-01 -2.40566298e-01 1.21962440e+00
2.46662199e-02 4.97221619e-01 -6.22346282e-01 -6.14864230e-01
1.95042118e-01 -6.36696041e-01 -5.06324947e-01 2.11867943e-01
-5.14867425e-01 -8.50131661e-02 -3.47175688e-01 8.79890084e-01
2.70393282e-01 -4.07258809e-01 1.10285306e+00 -1.44304287e+00
-1.32235095e-01 6.83079243e-01 -9.26598430e-01 6.39781952e-01
1.29089788e-01 3.97144593e-02 1.01325202e+00 -5.49587488e-01
6.89821959e-01 1.45286560e+00 7.92017758e-01 5.24646044e-02
-1.69585264e+00 -5.76839328e-01 7.74074495e-02 -4.19695616e-01
-1.34563470e+00 -4.07614470e-01 4.50899303e-01 -7.14191854e-01
4.70476478e-01 -1.94704086e-02 6.08412921e-01 1.27360725e+00
5.48225522e-01 4.95261759e-01 1.03747702e+00 -2.05802128e-01
4.82944131e-01 -1.83829032e-02 1.28536727e-02 6.23151422e-01
5.69946289e-01 2.28712767e-01 -8.09961796e-01 -7.98559666e-01
9.04653132e-01 -2.53081381e-01 -3.07176150e-02 -2.71105655e-02
-1.03406477e+00 9.96794939e-01 -1.91747442e-01 2.66207922e-02
-2.89374411e-01 4.74271655e-01 3.93156223e-02 5.56877792e-01
1.15247130e+00 1.83702502e-02 -3.36175174e-01 -1.95523530e-01
-9.88691866e-01 2.72282094e-01 1.47801626e+00 9.81331527e-01
4.34277803e-01 -3.91497493e-01 -6.41659051e-02 3.56371105e-01
6.54090226e-01 6.34845316e-01 -1.25615343e-01 -5.81645191e-01
3.60282481e-01 2.29663640e-01 2.38172010e-01 -9.95093882e-01
-2.51226753e-01 -5.00637233e-01 -1.06193650e+00 -2.25072414e-01
8.40453804e-01 -5.20726204e-01 -6.61111832e-01 1.75601065e+00
3.80393475e-01 4.72382098e-01 -2.77512252e-01 4.47793156e-01
-1.96708813e-01 9.57010567e-01 3.41555178e-02 -7.66088188e-01
1.30819452e+00 -3.58533561e-01 -6.26151919e-01 -2.21675381e-01
3.03640157e-01 -7.76044250e-01 2.67042279e-01 3.30922097e-01
-1.03389215e+00 1.37333319e-01 -7.36944735e-01 5.24737716e-01
1.30503774e-01 -4.76117313e-01 4.53259349e-01 8.38880777e-01
-1.31661677e+00 6.60140812e-01 -9.35238242e-01 -5.86322784e-01
2.90182054e-01 1.89194545e-01 1.54781014e-01 1.27813846e-01
-1.00025511e+00 4.70625252e-01 -1.30843729e-01 1.36846960e-01
-1.21006298e+00 -6.16101265e-01 -2.11322606e-01 -7.71984830e-02
4.09986943e-01 -8.32574129e-01 1.40681541e+00 -9.51409519e-01
-1.32455075e+00 4.97972816e-01 -5.32422662e-01 -6.49581075e-01
1.02103209e+00 -4.11780141e-02 -2.30692640e-01 7.35996068e-02
7.24912733e-02 -5.75765930e-02 1.37377894e+00 -1.01855850e+00
-4.39331859e-01 -4.52582031e-01 -3.07560086e-01 6.16258122e-02
1.87213585e-01 2.81226188e-01 -4.29206014e-01 -5.70811927e-01
7.69654810e-02 -1.18588960e+00 -5.39426684e-01 7.26195285e-03
-5.32239556e-01 -1.03677921e-01 7.04886079e-01 -4.15940791e-01
9.76413667e-01 -1.61875713e+00 9.97996777e-02 5.48926055e-01
4.47961450e-01 -2.81403303e-01 1.27279371e-01 9.67895508e-01
2.92445123e-01 5.68046212e-01 -3.24505538e-01 -6.55042112e-01
-2.48988107e-01 5.78962043e-02 -3.36883336e-01 9.86139834e-01
2.47459948e-01 5.69300234e-01 -8.91562045e-01 -3.59256625e-01
-3.63815963e-01 3.43785375e-01 -2.28634253e-01 2.14667439e-01
-3.58862609e-01 4.20180887e-01 -4.42699045e-01 1.86046734e-01
3.88322890e-01 -7.38426864e-01 4.49822247e-01 5.46148658e-01
1.38816863e-01 -3.55875902e-02 -1.34384358e+00 7.48066962e-01
-3.14408988e-01 6.89434528e-01 1.60986558e-01 -5.25116146e-01
5.58199763e-01 4.99522150e-01 4.45345551e-01 3.47198218e-01
2.50309944e-01 -3.94453779e-02 4.34172563e-02 -1.68927908e-02
1.76466420e-01 -3.22707981e-01 4.97968830e-02 1.01928306e+00
-1.65027469e-01 -5.81378862e-02 1.15884505e-01 5.19627571e-01
1.64238036e+00 -4.13905501e-01 3.96684527e-01 -2.57623106e-01
-6.47038547e-03 -1.00529917e-01 3.41243744e-01 1.10204673e+00
-1.53094605e-01 1.74147367e-01 1.15483034e+00 6.61154091e-02
-1.28974271e+00 -1.17922938e+00 5.74285053e-02 9.17525470e-01
8.81906040e-03 -4.53934819e-01 -6.55910313e-01 -2.20879868e-01
4.84481566e-02 3.35714072e-01 -6.45902753e-01 6.62503112e-03
-2.10174128e-01 -1.28982234e+00 5.23456097e-01 1.52278781e-01
1.24901742e-01 -4.62626517e-01 -5.54754324e-02 4.51122910e-01
7.36892819e-02 -1.10393453e+00 -5.57073057e-01 -6.94816932e-02
-1.05355096e+00 -1.02687490e+00 -5.62161446e-01 -2.96761751e-01
6.62950397e-01 1.29703745e-01 9.37454164e-01 -2.04635873e-01
-1.35535493e-01 4.68910962e-01 1.54383481e-03 -5.72393954e-01
-8.88776302e-01 -6.07240349e-02 6.85741156e-02 3.58465672e-01
1.91191062e-01 -6.00160778e-01 -5.40358424e-01 3.65464866e-01
-8.93835962e-01 -1.72186196e-01 5.33373713e-01 3.88168603e-01
1.31201386e-01 3.88262004e-01 2.96397239e-01 -1.15454137e+00
1.04549730e+00 -8.88446569e-01 -8.88240159e-01 1.71237569e-02
-6.34966552e-01 3.46220076e-01 3.58278036e-01 -4.87999290e-01
-8.62179160e-01 -1.16631947e-01 4.96240050e-01 -7.52220824e-02
1.96810365e-01 5.04885316e-01 3.31180125e-01 2.06932619e-01
5.86758852e-01 2.25097448e-01 3.27217698e-01 -7.38704577e-02
3.04998159e-01 5.09919822e-01 -3.16905081e-02 -3.79546344e-01
9.81513619e-01 7.41655588e-01 4.01128471e-01 -1.22099447e+00
-5.60307503e-01 -7.12031543e-01 -2.11477414e-01 -3.67707819e-01
5.34889877e-01 -9.65755641e-01 -8.78757238e-01 8.24363828e-01
-1.40101111e+00 -3.33297282e-01 7.03771338e-02 3.01297665e-01
-5.55763960e-01 2.31524363e-01 -9.88753080e-01 -1.41770351e+00
-9.11138803e-02 -6.85487032e-01 9.92450893e-01 -1.31945431e-01
-2.46528044e-01 -1.36368406e+00 3.38502079e-01 -9.27419960e-02
4.39857900e-01 1.16046324e-01 7.92499840e-01 -8.33768606e-01
-7.98869014e-01 -3.57554346e-01 -1.90505788e-01 6.29167333e-02
-5.63120656e-02 3.54517192e-01 -7.08578825e-01 -2.10842580e-01
1.45257905e-01 4.34054971e-01 6.66288972e-01 6.74534798e-01
6.19079053e-01 -4.60686922e-01 -6.98109508e-01 -1.25643602e-02
1.41269600e+00 -2.24394888e-01 1.59668073e-01 -3.28834116e-01
2.69691527e-01 7.42283344e-01 1.11024670e-01 6.28563106e-01
1.49901956e-01 2.61686027e-01 1.86944172e-01 2.12164089e-01
4.65826899e-01 -4.32751209e-01 4.58087713e-01 8.14677119e-01
-2.11186171e-01 -7.21425831e-01 -9.09725547e-01 5.46067655e-01
-1.85888267e+00 -1.03402901e+00 -5.05262017e-01 2.43596506e+00
8.17546844e-01 5.21813452e-01 4.75348890e-01 -6.89475238e-02
1.08387291e+00 2.64224887e-01 -3.47792208e-01 -1.49646997e-01
1.56001195e-01 3.94567698e-02 1.03079796e+00 8.98988008e-01
-8.55493069e-01 5.32877207e-01 7.06595278e+00 5.01835942e-01
-6.29611313e-01 3.89975756e-01 7.81093299e-01 8.15197900e-02
-1.44384682e-01 3.33915114e-01 -8.22493255e-01 5.81501365e-01
1.46798873e+00 -6.19602025e-01 4.47205186e-01 4.33800578e-01
6.74798191e-01 -1.73900902e-01 -1.28235209e+00 6.18008912e-01
-2.73243517e-01 -1.07873499e+00 -2.73666799e-01 5.20054221e-01
7.37033367e-01 8.28748718e-02 -1.28230438e-01 -2.70736605e-01
9.07793701e-01 -8.27329218e-01 5.00221848e-01 5.82144499e-01
3.25032830e-01 -6.28354490e-01 4.38818455e-01 6.56149387e-01
-1.07703829e+00 2.24262938e-01 -7.70958811e-02 -1.90495789e-01
5.86272657e-01 1.30214202e+00 -1.29158497e+00 1.18401535e-01
8.64924863e-02 3.88126045e-01 -3.08823353e-03 1.00384879e+00
-2.80255556e-01 1.11927855e+00 -7.06085801e-01 -2.92165756e-01
3.99226759e-04 -4.06010062e-01 9.77572024e-01 1.15777099e+00
3.58116627e-01 -3.89913358e-02 -3.83629836e-02 7.05506563e-01
-8.02656338e-02 -1.46368384e-01 -7.72429526e-01 -4.16429937e-01
5.24582863e-01 1.00124133e+00 -1.03615534e+00 -1.41898200e-01
-7.80209005e-02 8.75484824e-01 5.42067960e-02 5.05266726e-01
-8.49115372e-01 1.74629852e-01 6.53539717e-01 5.25432110e-01
3.50687534e-01 -5.41751623e-01 -3.63299586e-02 -9.18309391e-01
-3.78505066e-02 -3.79429936e-01 1.89446300e-01 -3.12924415e-01
-1.62623751e+00 3.10787767e-01 8.55937153e-02 -7.12075293e-01
-4.63923246e-01 -3.52024585e-01 -6.12903655e-01 8.42750967e-01
-1.11328602e+00 -6.23507380e-01 2.17782125e-01 1.56444818e-01
4.21659529e-01 1.71631619e-01 3.14644128e-01 -1.78619817e-01
-4.58700180e-01 -1.44962743e-01 3.58553171e-01 -1.11556791e-01
4.57258105e-01 -1.28978848e+00 8.58523130e-01 9.43748593e-01
1.69204682e-01 3.93593729e-01 1.24562287e+00 -9.93963420e-01
-1.38040864e+00 -1.15317464e+00 1.04529035e+00 -7.65165031e-01
1.18182552e+00 -6.15897357e-01 -5.35287678e-01 7.00569272e-01
3.06039974e-02 -1.40533209e-01 5.31597674e-01 2.13881135e-01
-1.23399414e-01 5.59055284e-02 -8.31549764e-01 6.63157821e-01
1.10814989e+00 -5.05426228e-01 -7.03700408e-02 7.03755736e-01
5.27703822e-01 1.04065955e-01 -6.91001356e-01 -9.27787125e-02
4.34230208e-01 -7.26987541e-01 4.75331008e-01 -6.37432396e-01
4.67540801e-01 -3.53486091e-02 2.47686088e-01 -1.39979267e+00
-2.12206781e-01 -1.26710308e+00 -2.52000898e-01 9.88230348e-01
5.75694442e-01 -8.91040087e-01 6.72423065e-01 3.23642761e-01
7.75893331e-01 -5.31994224e-01 -9.74927962e-01 -7.79651940e-01
-1.19561940e-01 -6.11148953e-01 6.78180307e-02 5.54784954e-01
-1.97231293e-01 7.16451168e-01 -4.13900673e-01 5.26550055e-01
1.08142662e+00 -2.46515110e-01 7.72051811e-01 -1.52252948e+00
-4.94039804e-01 -3.28169525e-01 -2.77125776e-01 -1.15191281e+00
-1.28072053e-01 -3.65130782e-01 2.53170490e-01 -1.19001627e+00
3.16112787e-01 -5.02808630e-01 3.25060487e-01 -3.53228688e-01
-1.19927011e-01 -2.10827231e-01 -2.30891734e-01 5.05878925e-01
-3.34801316e-01 2.32292205e-01 9.32290912e-01 2.25319490e-01
-1.24573708e-01 5.42608619e-01 -5.30658007e-01 7.29165614e-01
6.33637309e-01 -9.66207385e-01 -4.51800466e-01 -5.20874299e-02
6.61884189e-01 4.73456919e-01 4.33664262e-01 -6.18690789e-01
4.89010274e-01 -6.45828247e-02 3.19007412e-02 -2.31755883e-01
2.93880075e-01 -5.94801009e-01 5.22483826e-01 6.28010273e-01
-4.96861666e-01 2.57698774e-01 -3.65728676e-01 1.63199663e+00
2.46140748e-01 -1.31874144e-01 5.18811882e-01 -5.01720533e-02
7.43941367e-02 5.25402367e-01 -8.64224970e-01 2.56814092e-01
1.05643153e+00 2.41918057e-01 -4.89532262e-01 -1.04866195e+00
-6.75179064e-01 1.27743691e-01 3.98101449e-01 1.51451305e-01
1.41422629e-01 -7.48250127e-01 -9.68072236e-01 -1.34248644e-01
-1.99935287e-01 -2.51513839e-01 -1.54724419e-01 1.07738125e+00
-4.69181448e-01 1.09999739e-01 6.24728680e-01 -6.88387692e-01
-1.07321298e+00 4.25599813e-01 1.95880875e-01 -5.36234856e-01
-2.06589669e-01 7.12649047e-01 9.10020247e-02 1.64782882e-01
-1.00516386e-01 -2.63002008e-01 3.60102594e-01 7.81487897e-02
4.75302100e-01 4.98818219e-01 -2.52782404e-01 -3.48907679e-01
-2.25488201e-01 1.59878954e-01 8.79819226e-03 -6.53994799e-01
9.76310670e-01 -2.42159873e-01 -2.81745613e-01 8.52829218e-01
9.02678907e-01 1.37173444e-01 -1.32994664e+00 -4.62491155e-01
1.39629930e-01 -2.65438199e-01 -3.96564566e-02 -3.10546488e-01
-6.52484894e-01 5.27306974e-01 2.05836579e-01 9.88450468e-01
5.83317161e-01 4.10887569e-01 2.99489260e-01 -5.72758261e-03
3.52557123e-01 -8.43472779e-01 -1.49822123e-02 2.69635439e-01
5.75378239e-01 -9.46744978e-01 1.11726448e-01 -8.60796273e-01
-3.80626142e-01 9.12074804e-01 -2.33310685e-01 -2.78114200e-01
1.10775137e+00 4.22389090e-01 -3.82401526e-01 -2.40623266e-01
-1.25857162e+00 8.35268348e-02 -8.98705497e-02 3.81426275e-01
1.60027325e-01 1.92218110e-01 -3.14363301e-01 4.32748757e-02
4.97515127e-02 -1.06000111e-01 7.65062690e-01 8.24899375e-01
-4.62435484e-01 -1.05548489e+00 -3.21379155e-01 7.12806046e-01
-6.35765791e-01 -1.41541496e-01 -5.50167561e-01 6.65734828e-01
-4.49416161e-01 1.22208619e+00 1.78667322e-01 8.08402672e-02
-1.50625393e-01 9.13885981e-03 2.83653021e-01 -5.81374764e-01
-2.26344347e-01 2.09914863e-01 4.11146343e-01 -2.39504486e-01
-4.61040258e-01 -9.67713416e-01 -4.49171513e-01 -7.25525558e-01
-5.68708479e-01 1.98359281e-01 6.96955979e-01 1.18708289e+00
2.03986675e-01 1.73647717e-01 8.15460443e-01 -4.97958004e-01
-9.61452723e-01 -9.92532194e-01 -8.26869011e-01 -2.25448400e-01
3.27929288e-01 -3.52720380e-01 -8.53793859e-01 2.35887900e-01]
|
[6.769892692565918, 5.033936500549316]
|
0d41ff85-2f52-4ff6-aafa-a35518d08c97
|
data-efficient-autoregressive-document
|
2211.09388
| null |
https://arxiv.org/abs/2211.09388v1
|
https://arxiv.org/pdf/2211.09388v1.pdf
|
Data-Efficient Autoregressive Document Retrieval for Fact Verification
|
Document retrieval is a core component of many knowledge-intensive natural language processing task formulations such as fact verification and question answering. Sources of textual knowledge, such as Wikipedia articles, condition the generation of answers from the models. Recent advances in retrieval use sequence-to-sequence models to incrementally predict the title of the appropriate Wikipedia page given a query. However, this method requires supervision in the form of human annotation to label which Wikipedia pages contain appropriate context. This paper introduces a distant-supervision method that does not require any annotation to train autoregressive retrievers that attain competitive R-Precision and Recall in a zero-shot setting. Furthermore we show that with task-specific supervised fine-tuning, autoregressive retrieval performance for two Wikipedia-based fact verification tasks can approach or even exceed full supervision using less than $1/4$ of the annotated data indicating possible directions for data-efficient autoregressive retrieval.
|
['James Thorne']
|
2022-11-17
| null | null | null | null |
['fact-verification']
|
['natural-language-processing']
|
[ 1.71410337e-01 1.47710219e-01 -4.24840331e-01 -4.23456252e-01
-1.51906097e+00 -7.02831209e-01 8.23808134e-01 4.78823394e-01
-6.43110096e-01 6.60468459e-01 4.18537945e-01 -3.35029304e-01
-1.78524435e-01 -6.72381401e-01 -8.52908790e-01 -2.11586788e-01
3.20626944e-02 7.48159230e-01 4.93303746e-01 -4.30352420e-01
6.25745475e-01 6.27391487e-02 -1.37238932e+00 4.47199970e-01
9.20152605e-01 8.58770013e-01 1.58962518e-01 1.03309107e+00
-4.26337808e-01 1.01309741e+00 -4.41310793e-01 -6.11914992e-01
2.64481843e-01 -1.76599413e-01 -9.11117613e-01 -1.92966461e-01
6.13194823e-01 -2.93642491e-01 -5.31080246e-01 9.13428903e-01
3.57473910e-01 3.23283017e-01 8.56484532e-01 -5.72008967e-01
-8.45310271e-01 5.63642919e-01 -4.54152733e-01 5.32807291e-01
6.15895092e-01 -1.72690004e-01 1.25831926e+00 -9.06424582e-01
8.48734438e-01 8.37466419e-01 3.39843333e-01 3.38525295e-01
-1.08594584e+00 -1.07119635e-01 4.23898734e-02 1.79369867e-01
-1.27974653e+00 -5.00730038e-01 5.98846316e-01 -4.57695842e-01
1.34943783e+00 2.08743930e-01 -1.74311618e-03 7.62034893e-01
-8.25282782e-02 8.55355918e-01 5.22664905e-01 -9.44503605e-01
3.72183397e-02 3.12723756e-01 8.45868766e-01 7.83414900e-01
2.57617265e-01 -5.34449279e-01 -7.49659419e-01 -3.27720463e-01
2.89578706e-01 -3.18614066e-01 -2.18944713e-01 -4.23395932e-01
-1.02358866e+00 8.61235380e-01 1.77944094e-01 2.30735317e-01
-6.13643050e-01 8.81789178e-02 6.21946216e-01 4.22821730e-01
4.52291042e-01 9.15985286e-01 -6.70309901e-01 -2.16005534e-01
-1.01566398e+00 1.75406277e-01 1.12468994e+00 1.14186811e+00
8.22545290e-01 -1.33035943e-01 -1.50237098e-01 8.48890781e-01
1.99138850e-01 6.13094449e-01 7.56045759e-01 -9.01955128e-01
7.02949107e-01 5.64569652e-01 4.33111995e-01 -8.87373984e-01
6.19184561e-02 -2.08672181e-01 -2.79934525e-01 -5.80902278e-01
4.46268022e-01 -1.52634438e-02 -9.32356954e-01 1.30499554e+00
2.63139755e-01 -2.57128149e-01 1.60588756e-01 6.86604440e-01
3.86954665e-01 6.49127483e-01 2.29851067e-01 -2.94082016e-01
1.53515863e+00 -1.05360126e+00 -7.38034785e-01 -3.76425743e-01
8.52651417e-01 -7.60025680e-01 1.09575343e+00 4.10614252e-01
-1.08906400e+00 -2.91949123e-01 -1.01423466e+00 -5.42900443e-01
-5.46249568e-01 2.62275279e-01 6.69701755e-01 4.71754819e-01
-7.11571157e-01 2.08895385e-01 -5.27175546e-01 -2.27526858e-01
-6.85376376e-02 5.45153618e-02 -2.38444969e-01 -2.66855836e-01
-1.46380568e+00 8.65683615e-01 3.36590350e-01 -1.21334819e-02
-7.19928801e-01 -7.04346716e-01 -7.51168907e-01 2.62974232e-01
7.66937256e-01 -7.77446628e-01 1.55396676e+00 -9.45575058e-01
-1.18674898e+00 8.42906058e-01 -5.13961732e-01 -8.85889411e-01
1.80347130e-01 -6.95134103e-01 -3.13751608e-01 6.22231066e-01
1.88544393e-01 1.94356605e-01 9.19515073e-01 -7.19264090e-01
-4.15306509e-01 -4.91058677e-01 1.32914335e-01 1.37808785e-01
-4.35441643e-01 2.39371285e-01 -8.06537688e-01 -3.70054901e-01
-1.37215131e-03 -8.96373391e-01 -1.46588981e-01 -3.32839519e-01
-1.80147946e-01 -5.43614388e-01 3.32131892e-01 -9.36181545e-01
1.49541330e+00 -1.80206466e+00 7.12732552e-03 3.91675204e-01
-7.39694759e-02 3.47850084e-01 -2.83060610e-01 6.20964646e-01
2.27966875e-01 2.22200707e-01 1.66884318e-01 3.02081257e-02
9.41398740e-02 -2.13407934e-01 -8.58040512e-01 1.70121640e-01
1.25389785e-01 9.38489318e-01 -9.03036952e-01 -6.03934228e-01
-3.90806168e-01 3.10082167e-01 -4.51639742e-01 1.34917811e-01
-7.85009027e-01 -4.02615219e-01 -7.79678822e-01 3.36167067e-01
-1.19483195e-01 -5.74658632e-01 1.88872948e-01 5.42569831e-02
2.02448726e-01 6.81846976e-01 -9.78484929e-01 1.74375415e+00
-5.37319541e-01 7.07753181e-01 -1.97629660e-01 -9.34233785e-01
9.03569639e-01 4.06544775e-01 2.40420043e-01 -7.62662530e-01
-2.77560145e-01 2.56248027e-01 -3.22806150e-01 -8.17489266e-01
1.00612032e+00 4.14095074e-02 -7.57614076e-02 5.52536488e-01
-5.58909848e-02 -1.01251658e-02 7.06721604e-01 8.41072440e-01
1.24164522e+00 8.14770684e-02 2.41100609e-01 1.23981247e-02
6.45746827e-01 4.06243622e-01 2.78750062e-01 1.18642497e+00
2.00892329e-01 2.76046067e-01 4.38300967e-01 -2.43314654e-01
-1.22437596e+00 -7.11087167e-01 6.25521913e-02 1.47579634e+00
-3.10914963e-01 -4.54592645e-01 -5.76986134e-01 -6.49067223e-01
6.62268028e-02 1.00544047e+00 -4.27536011e-01 -7.28210956e-02
-5.37831128e-01 -4.16386992e-01 5.82973599e-01 4.55110312e-01
1.31024301e-01 -9.59162891e-01 -4.34571236e-01 2.47015253e-01
-3.95797729e-01 -1.03632927e+00 -6.06247723e-01 7.54762366e-02
-7.48018026e-01 -1.17997658e+00 -7.43200123e-01 -6.97336555e-01
5.42942345e-01 3.94921482e-01 1.23103774e+00 2.43528679e-01
-2.35143051e-01 9.22014773e-01 -3.16890419e-01 -3.17047477e-01
-2.50953764e-01 4.38264400e-01 -6.99687749e-02 -3.88058960e-01
9.20561850e-01 -7.56353512e-02 -3.79430950e-01 2.10369639e-02
-9.09275115e-01 -5.06725788e-01 5.86745858e-01 8.39217961e-01
4.27919626e-01 -2.00920895e-01 8.29108953e-01 -1.15509927e+00
8.95033777e-01 -5.00112355e-01 -7.22248554e-01 8.10642779e-01
-8.41969073e-01 5.08476317e-01 2.59106040e-01 -4.30543244e-01
-1.37762249e+00 5.65241836e-02 4.65026110e-01 -2.94209450e-01
3.51358086e-01 1.03049004e+00 3.74489665e-01 4.25062329e-01
1.20743620e+00 4.89264131e-01 -1.90428659e-01 -4.50791508e-01
5.12171149e-01 7.23894060e-01 4.32506174e-01 -6.02186501e-01
7.67094553e-01 7.84045383e-02 -3.10148716e-01 -9.16110337e-01
-1.23601246e+00 -1.09984350e+00 -6.82033300e-01 2.47504818e-03
6.36374950e-01 -1.01914024e+00 -5.79372346e-01 -1.64046273e-01
-1.17470741e+00 -1.19008452e-01 -3.46910395e-02 4.30233806e-01
-6.32117748e-01 4.87804383e-01 -7.80110180e-01 -1.06783211e+00
-6.71274543e-01 -5.76621354e-01 1.09257460e+00 3.63355428e-02
-2.56313145e-01 -8.02318871e-01 2.73154318e-01 8.92381370e-01
3.61597419e-01 -6.06267512e-01 1.13030565e+00 -1.20817375e+00
-1.01041484e+00 -8.20231259e-01 -1.20099306e-01 1.71797350e-01
-3.33096772e-01 -3.24047469e-02 -7.51915455e-01 7.80884400e-02
-2.53050029e-01 -5.66214323e-01 9.46376860e-01 7.71841779e-02
6.22292042e-01 -5.23946404e-01 -2.24994153e-01 -1.38564095e-01
1.32913828e+00 4.13692407e-02 7.22985744e-01 3.92922699e-01
3.21072936e-01 8.57876301e-01 7.75800943e-01 5.81181385e-02
1.34589300e-01 5.04526138e-01 -3.63209277e-01 6.08216405e-01
1.98465101e-02 -6.30052328e-01 2.33081341e-01 8.14890444e-01
1.32602111e-01 -1.52900383e-01 -1.08512580e+00 7.69408524e-01
-1.70915723e+00 -1.22664273e+00 -1.32045373e-01 2.40014458e+00
1.18823195e+00 1.89928621e-01 -1.27605706e-01 -1.28538087e-01
4.07206953e-01 9.69450548e-03 -4.78552848e-01 -1.57875732e-01
8.80290642e-02 -1.19897224e-01 4.21413839e-01 7.23959804e-01
-9.81023073e-01 1.10663950e+00 6.57247829e+00 6.61017478e-01
-5.60698688e-01 -9.44588408e-02 2.09181651e-01 -7.73505643e-02
-4.47638720e-01 1.31446257e-01 -9.80366468e-01 2.19621047e-01
1.37012208e+00 -6.07082367e-01 4.00714576e-01 1.07853532e+00
1.68633625e-01 -4.09890234e-01 -9.77855742e-01 6.14115119e-01
5.71958542e-01 -1.26114404e+00 1.65520161e-01 -8.02322701e-02
7.20018387e-01 3.38269733e-02 -1.66211501e-01 6.41092122e-01
5.03905773e-01 -5.07473707e-01 2.33026057e-01 8.65502834e-01
3.48384649e-01 -4.08868343e-01 5.45824468e-01 5.39593995e-01
-8.12644184e-01 -1.06323160e-01 -6.05544448e-01 2.99278677e-01
1.69367164e-01 6.05143726e-01 -1.26436996e+00 3.66683006e-01
3.98208380e-01 4.35206532e-01 -6.69069529e-01 8.96951199e-01
-4.34514523e-01 5.43152750e-01 -1.90980047e-01 -4.03229088e-01
1.07388191e-01 2.00835746e-02 4.87143725e-01 1.27213073e+00
1.12273015e-01 2.91789383e-01 5.48055843e-02 2.39489958e-01
-3.83637905e-01 4.44525838e-01 -6.85192406e-01 -3.14271688e-01
4.41186756e-01 8.38760853e-01 -3.51627439e-01 -8.51789594e-01
-3.26190531e-01 8.30082595e-01 5.06053686e-01 5.67871690e-01
-3.92798036e-01 -6.63979530e-01 1.35947660e-01 1.80473551e-01
4.75850284e-01 -2.92104840e-01 2.91146077e-02 -1.28942084e+00
3.18353981e-01 -9.19560671e-01 6.60526633e-01 -1.19231820e+00
-1.12112391e+00 3.40776503e-01 -8.67749676e-02 -8.63053024e-01
-8.01181853e-01 -3.98130983e-01 -5.73784113e-02 8.15312803e-01
-1.77016497e+00 -7.89896846e-01 3.94624472e-02 3.97557408e-01
6.80514038e-01 -1.66961685e-01 7.91068137e-01 3.00550967e-01
-3.45207423e-01 3.00833642e-01 2.96329588e-01 4.20580178e-01
1.08461368e+00 -1.34103215e+00 -2.40289187e-03 9.28079188e-01
5.06836593e-01 1.22398663e+00 9.14143980e-01 -8.37301075e-01
-1.65773666e+00 -6.91112578e-01 1.50495660e+00 -7.24488914e-01
1.09090221e+00 -5.41285425e-02 -1.18649220e+00 8.38809073e-01
2.05881372e-01 -3.73362333e-01 7.43673325e-01 6.52498484e-01
-9.95522976e-01 -2.40874201e-01 -6.78098559e-01 5.10666549e-01
6.39059842e-01 -1.14620817e+00 -1.10687077e+00 8.07955384e-01
8.03192198e-01 -1.62936121e-01 -7.64537752e-01 -1.74702749e-01
5.61057627e-01 -2.32005879e-01 8.67673457e-01 -1.10839200e+00
4.92334962e-01 -1.79053649e-01 3.45206559e-02 -8.71610880e-01
-3.30451760e-03 -6.01299047e-01 -4.01691943e-01 9.75980997e-01
8.04698229e-01 -2.91184902e-01 7.62536943e-01 1.32367599e+00
1.67312279e-01 -1.93790346e-01 -5.81064224e-01 -6.59206688e-01
-9.07983556e-02 -2.45751604e-01 6.15599006e-03 7.35539436e-01
5.01070321e-01 8.41151953e-01 -1.74067125e-01 1.11467212e-01
3.43642741e-01 1.34875029e-01 8.52337420e-01 -1.25445926e+00
-2.01067165e-01 -4.87018190e-02 -2.23988339e-01 -1.45241904e+00
3.04464072e-01 -8.96253645e-01 2.72117913e-01 -1.52413702e+00
2.97279716e-01 -1.07422210e-01 -1.38335556e-01 3.46462458e-01
-2.45358959e-01 -2.15417385e-01 -9.03378278e-02 4.43688959e-01
-1.13806903e+00 3.15263927e-01 5.48462272e-01 -3.25349748e-01
-1.83116078e-01 -1.44669816e-01 -6.09825134e-01 3.78377259e-01
5.57980359e-01 -5.18437624e-01 -6.11855924e-01 -4.89105403e-01
7.55524695e-01 3.01101804e-01 6.40701177e-03 -4.29971814e-01
8.09300423e-01 4.89970818e-02 7.22178072e-02 -4.54605937e-01
2.38890216e-01 -7.23164678e-01 -2.78285891e-01 7.61347413e-02
-1.01518536e+00 5.80095947e-02 -2.32037995e-02 1.06980145e+00
-2.45541021e-01 -8.69701862e-01 1.74109682e-01 -2.93040603e-01
-6.94908738e-01 -3.86686623e-02 -5.93210220e-01 3.22601080e-01
6.64565802e-01 2.69842207e-01 -5.42505145e-01 -5.50048769e-01
-4.52795357e-01 2.76399374e-01 6.20554350e-02 3.60683054e-01
3.84158969e-01 -7.99877644e-01 -6.58154011e-01 -3.68146420e-01
2.50180662e-01 -2.52643347e-01 1.46213621e-01 5.01148522e-01
-1.88204378e-01 1.24385154e+00 4.72802848e-01 -2.00906128e-01
-1.25562382e+00 9.32813227e-01 -3.51147130e-02 -7.44434714e-01
-3.12888384e-01 5.33565402e-01 -2.70339578e-01 -1.41802922e-01
2.02021629e-01 2.32297808e-01 -1.94064796e-01 1.78541377e-01
9.00508940e-01 1.06997795e-01 2.20888540e-01 -1.30893290e-01
4.28310595e-03 1.78383887e-01 -6.69357657e-01 -3.83624285e-01
1.20796144e+00 -1.33867040e-01 -1.75169334e-02 4.67027932e-01
1.11736369e+00 6.96538463e-02 -8.25782716e-01 -6.78866267e-01
8.05952549e-01 -3.36610347e-01 1.26634970e-01 -8.76428485e-01
-5.55747688e-01 7.62273848e-01 -2.78339945e-02 2.79912382e-01
7.70062804e-01 -4.47594151e-02 5.48930824e-01 1.40481114e+00
3.77839357e-01 -1.43575430e+00 2.56027848e-01 5.63070178e-01
9.08876181e-01 -1.28775489e+00 1.97757155e-01 -2.96706259e-01
-6.60640836e-01 9.62690413e-01 4.47858125e-01 1.09494161e-02
5.66598058e-01 -2.06508189e-01 -5.45873540e-03 -3.39025170e-01
-1.16230071e+00 -1.62791073e-01 4.40787762e-01 2.83188403e-01
6.61902368e-01 -4.25545901e-01 -4.37702268e-01 2.75437027e-01
8.79601315e-02 8.87257382e-02 5.22564650e-01 9.54382360e-01
-7.29175687e-01 -8.84289086e-01 -3.51242349e-02 7.28044093e-01
-5.72484076e-01 -5.48873186e-01 -2.28550941e-01 4.93593812e-01
-1.09808600e+00 9.51298714e-01 5.58911934e-02 3.12074751e-01
2.33939469e-01 8.15749645e-01 1.47182092e-01 -8.24317098e-01
-4.90043670e-01 3.10239941e-02 5.27415037e-01 -2.25924835e-01
-4.16388631e-01 -6.35703206e-01 -1.06929505e+00 8.77591968e-02
-5.93354940e-01 7.09046602e-01 7.54392326e-01 8.23188603e-01
6.06760740e-01 1.34202242e-01 2.57085502e-01 -1.41460478e-01
-9.81025934e-01 -1.14224517e+00 -2.55193681e-01 2.95111895e-01
1.62443638e-01 -2.98388898e-01 -3.74707699e-01 4.59745735e-01]
|
[11.4216947555542, 7.877123832702637]
|
dac7ec4a-ae68-4460-a0c5-021458a4e84d
|
learning-to-automate-cryo-electron-microscopy
|
2112.01534
| null |
https://arxiv.org/abs/2112.01534v2
|
https://arxiv.org/pdf/2112.01534v2.pdf
|
Learning to automate cryo-electron microscopy data collection with Ptolemy
|
Over the past decade, cryogenic electron microscopy (cryo-EM) has emerged as a primary method for determining near-native, near-atomic resolution 3D structures of biological macromolecules. In order to meet increasing demand for cryo-EM, automated methods to improve throughput and efficiency while lowering costs are needed. Currently, all high-magnification cryo-EM data collection softwares require human input and manual tuning of parameters. Expert operators must navigate low- and medium-magnification images to find good high-magnification collection locations. Automating this is non-trivial: the images suffer from low signal-to-noise ratio and are affected by a range of experimental parameters that can differ for each collection session. Here, we use various computer vision algorithms, including mixture models, convolutional neural networks, and U-Nets to develop the first pipeline to automate low- and medium-magnification targeting. Learned models in this pipeline are trained on a large internal dataset of images from real world cryo-EM data collection sessions, labeled with locations that were selected by operators. Using these models, we show that we can effectively detect and classify regions of interest in low- and medium-magnification images, and can generalize to unseen sessions, as well as to images captured using different microscopes from external facilities. We expect our open-source pipeline, Ptolemy, will be both immediately useful as a tool for automation of cryo-EM data collection, and serve as a foundation for future advanced methods for efficient and automated cryo-EM microscopy.
|
['Tristan Bepler', 'Anchi Cheng', 'Alex J. Noble', 'Paul T. Kim']
|
2021-12-01
| null | null | null | null |
['cryogenic-electron-microscopy-cryo-em']
|
['computer-vision']
|
[ 1.00096002e-01 -5.41684747e-01 6.38826489e-01 -5.89699388e-01
-9.83487487e-01 -7.64662743e-01 2.19467863e-01 3.93807292e-01
-1.01182902e+00 6.72286093e-01 -5.66686511e-01 -4.87416416e-01
2.42822409e-01 -3.34214896e-01 -8.56838226e-01 -7.12316096e-01
-1.37448236e-01 9.72764909e-01 4.17782307e-01 1.47264421e-01
5.55561781e-01 1.05311120e+00 -1.26622570e+00 4.12510991e-01
3.91055912e-01 4.62726235e-01 1.20405304e+00 9.93333042e-01
2.64053158e-02 2.99798936e-01 -6.12857163e-01 6.38495386e-02
-1.52086010e-02 -1.64925933e-01 -1.11192000e+00 -1.52385443e-01
2.86153615e-01 -2.63322532e-01 3.36287737e-01 1.00826502e+00
6.21705055e-01 -6.01041690e-02 4.74747151e-01 -5.39493263e-01
-5.95158100e-01 -1.40598655e-01 -5.12014031e-01 7.48122513e-01
1.60284996e-01 5.78176856e-01 4.37760442e-01 -8.25286329e-01
1.13750696e+00 1.11700594e+00 4.50965852e-01 6.83993101e-01
-1.96816611e+00 -3.08629781e-01 -1.04345851e-01 2.96507686e-01
-7.59665489e-01 -4.24769163e-01 3.68540078e-01 -5.69997191e-01
1.23265755e+00 -8.49199146e-02 4.59703088e-01 8.11410189e-01
7.82115579e-01 1.70729071e-01 1.27246046e+00 -4.33874011e-01
3.37350756e-01 2.56536435e-02 -1.04110420e-01 5.95153928e-01
-1.69989213e-01 -4.23659265e-01 -9.68754664e-02 -3.06490988e-01
8.57615590e-01 3.21028709e-01 -3.65360111e-01 -4.70006168e-01
-1.39356744e+00 4.32691872e-01 3.52799118e-01 4.74837542e-01
-3.25807005e-01 -3.08458090e-01 2.79256403e-01 1.99787140e-01
2.70823807e-01 9.43959713e-01 -6.60034418e-01 -1.15055725e-01
-7.66790748e-01 1.73045292e-01 4.33236331e-01 7.37164497e-01
9.59331870e-01 -6.57776117e-01 6.04625463e-01 7.71938264e-01
-1.54927578e-02 1.74549073e-01 4.31110322e-01 -9.78186846e-01
-6.83027282e-02 4.89400595e-01 2.43915245e-01 -4.89542574e-01
-9.18091416e-01 2.14942753e-01 -7.52546906e-01 5.50909400e-01
6.05185091e-01 -1.08415429e-02 -1.17058825e+00 1.48638034e+00
3.86469603e-01 -3.79359096e-01 -4.00329977e-01 1.01678813e+00
3.92387867e-01 5.24710655e-01 3.76165137e-02 -1.84210286e-01
1.37824583e+00 -4.32303607e-01 -4.18922395e-01 -1.21965818e-01
6.42047644e-01 -9.87855077e-01 1.13965642e+00 2.32729509e-01
-1.15348935e+00 -4.10273701e-01 -1.21149039e+00 -3.28971982e-01
-7.49975443e-01 6.18737526e-02 4.68625575e-01 6.19903430e-02
-1.13872349e+00 1.02027452e+00 -1.45261395e+00 -6.37564361e-01
4.99860615e-01 5.76132476e-01 -7.85425425e-01 9.96118188e-02
-4.31171656e-01 1.10315812e+00 3.03972691e-01 6.77010864e-02
-1.01448023e+00 -8.92274141e-01 -5.22393584e-01 1.79617554e-02
-4.45031673e-02 -5.37934840e-01 1.19062710e+00 -3.20742309e-01
-1.14788830e+00 1.35662150e+00 -2.75113463e-01 -1.62625253e-01
6.18511327e-02 1.71037555e-01 -6.27694726e-02 5.64302921e-01
1.89531714e-01 7.60293245e-01 3.94050807e-01 -1.21256161e+00
-4.99526173e-01 -7.48915911e-01 -2.23559871e-01 -2.84988344e-01
1.75005972e-01 2.84029901e-01 -1.24965817e-01 2.71680176e-01
2.66946256e-01 -7.06582367e-01 -3.31068903e-01 4.00411002e-02
-1.08571634e-01 2.12660149e-01 1.24536252e+00 -6.93809152e-01
3.23811293e-01 -1.77904093e+00 2.32300699e-01 -1.38474613e-01
4.31576401e-01 2.87994087e-01 1.44109726e-01 2.20383242e-01
-1.14621624e-01 2.66617164e-03 1.03147678e-01 -4.73545760e-01
-4.43033397e-01 -5.71374111e-02 1.30973503e-01 7.65978575e-01
3.77429664e-01 7.08045959e-01 -8.73589694e-01 -5.42190492e-01
6.19144499e-01 5.83413720e-01 -2.59110570e-01 5.16919255e-01
-2.29727879e-01 8.38516176e-01 1.25383049e-01 7.45688200e-01
8.18242311e-01 -7.42394745e-01 5.85483313e-01 -3.65181983e-01
-3.20387155e-01 2.94535644e-02 -7.83534825e-01 1.65273833e+00
-2.74595529e-01 5.16246557e-01 7.81572461e-01 -7.49806166e-01
6.69065058e-01 -3.63224931e-02 3.24824005e-01 -5.29542983e-01
1.60608232e-01 1.33911490e-01 -2.59697139e-02 -5.22757232e-01
3.32505584e-01 -4.22248423e-01 4.57805485e-01 6.13943040e-01
4.47150767e-01 -1.29347622e-01 3.71443212e-01 1.54492006e-01
1.06737900e+00 2.02073678e-01 1.26600549e-01 -5.63128829e-01
3.26103300e-01 2.57699579e-01 4.09784675e-01 5.36544442e-01
-4.04501587e-01 8.79556835e-01 3.02993894e-01 -1.05288565e+00
-1.83472443e+00 -9.69597399e-01 -5.31623125e-01 1.02739179e+00
-3.08146589e-02 -3.64599861e-02 -8.30080152e-01 -3.27164710e-01
-2.25736454e-01 -1.96832269e-01 -4.71360892e-01 1.48202613e-01
-6.62044764e-01 -9.93470252e-01 3.64334956e-02 3.76027703e-01
1.16691384e-02 -1.22479665e+00 -9.32432652e-01 4.25517350e-01
5.62315285e-02 -1.10297680e+00 -1.33900225e-01 8.78920853e-01
-8.74432385e-01 -1.17788327e+00 -7.52129614e-01 -9.18908596e-01
1.03707218e+00 4.64371532e-01 1.21606147e+00 2.26827726e-01
-8.24591935e-01 2.60368977e-02 -2.56533902e-02 -2.16802403e-01
-2.96142370e-01 2.55521387e-01 3.40044200e-01 -4.49048162e-01
7.05655277e-01 -8.44085991e-01 -7.70027816e-01 5.01683950e-01
-9.26044285e-01 -1.62397642e-02 3.84457111e-01 9.38993931e-01
1.15274513e+00 -3.47026885e-01 2.87363619e-01 -8.83181095e-01
4.33625132e-01 -9.75318998e-02 -9.05655861e-01 1.81805298e-01
-3.22603434e-01 -1.75187103e-02 1.18211710e+00 -3.48910987e-01
-7.53771663e-01 5.71806617e-02 -2.63797879e-01 -4.86557096e-01
-6.17062092e-01 4.33266014e-02 -1.81664959e-01 -4.42634046e-01
7.56875873e-01 1.42355740e-01 2.40548655e-01 -5.53645790e-01
-9.46959332e-02 6.38610840e-01 6.59750640e-01 -7.40618289e-01
2.93724388e-01 9.27534163e-01 6.86013401e-02 -8.83511901e-01
-5.06013811e-01 -6.26731098e-01 -1.02906573e+00 7.03399256e-02
1.06543255e+00 -6.01751089e-01 -1.02450275e+00 3.83223355e-01
-1.08232248e+00 -3.39164972e-01 3.84794086e-01 3.27704310e-01
-6.14727795e-01 4.72914308e-01 -1.01682234e+00 -5.12029052e-01
-4.19575840e-01 -1.43104303e+00 1.25737381e+00 3.43370795e-01
-1.41870067e-01 -8.83335292e-01 2.42844168e-02 2.70850122e-01
4.25980926e-01 -7.44311586e-02 1.12211895e+00 -2.32501447e-01
-7.33845472e-01 -7.58209825e-02 -1.56393826e-01 4.34700735e-02
2.24390745e-01 2.53523767e-01 -1.04914248e+00 -6.39464557e-01
7.51079470e-02 -5.95694244e-01 7.37370908e-01 6.07371986e-01
1.22012734e+00 3.27689826e-01 -5.07503629e-01 9.68047321e-01
1.48080850e+00 3.63969386e-01 7.16791749e-01 7.66133308e-01
5.28991938e-01 6.10245585e-01 3.43744427e-01 1.20308638e-01
-6.98185042e-02 4.12881494e-01 3.64355028e-01 -3.45608175e-01
6.20016634e-01 2.53742129e-01 -1.34641096e-01 3.27880472e-01
-2.93017685e-01 2.06041321e-01 -7.73652375e-01 5.15189648e-01
-1.57597721e+00 -1.08256471e+00 1.81838766e-01 2.05805397e+00
8.36233199e-01 7.74375424e-02 3.45744252e-01 -2.41658166e-01
6.93580925e-01 -2.99726218e-01 -6.98627234e-01 -2.09771782e-01
-1.39018118e-01 1.99875072e-01 3.32585871e-01 3.66598904e-01
-1.14042497e+00 7.55348980e-01 7.18723917e+00 1.68844074e-01
-1.31454349e+00 -1.22761771e-01 8.05742085e-01 -2.50061810e-01
1.88745648e-01 -7.51926797e-03 -9.69199240e-01 6.11499965e-01
1.05090010e+00 1.65017083e-01 5.59380352e-01 9.67425764e-01
3.25465560e-01 -2.71073729e-01 -1.30684352e+00 8.67874503e-01
-3.14960718e-01 -1.70650458e+00 -1.81753069e-01 2.37594098e-01
2.12530315e-01 4.19230998e-01 -2.73099542e-01 -2.25871876e-01
2.78705597e-01 -8.94242883e-01 1.17197908e-01 2.75713384e-01
7.71226943e-01 -8.06133449e-01 8.31133962e-01 4.13074166e-01
-8.67934942e-01 3.58059704e-01 -1.07108891e+00 1.96576893e-01
5.01067400e-01 6.15136385e-01 -9.17706311e-01 2.57827323e-02
1.09621465e+00 1.93576038e-01 -3.67948174e-01 6.42421663e-01
2.86852568e-01 -8.47999007e-02 -3.88833910e-01 -3.00841872e-02
-4.49834242e-02 -2.82907158e-01 1.81554481e-01 1.42775023e+00
-1.03889294e-02 -1.68308690e-01 5.07217236e-02 1.15603125e+00
-1.14557460e-01 -2.34267294e-01 -4.42863733e-01 -1.67176172e-01
3.42152774e-01 1.83855462e+00 -1.16099787e+00 -1.56194925e-01
-2.06995308e-01 8.32984924e-01 8.67503464e-01 2.22472623e-01
-3.77321303e-01 -5.29442847e-01 8.64694059e-01 3.16253752e-01
3.21366668e-01 -5.50103784e-01 -7.53757656e-02 -1.25931370e+00
-6.96086660e-02 -8.04167271e-01 1.67233169e-01 -9.40970600e-01
-1.29512119e+00 6.16001308e-01 -2.62822300e-01 -7.08649457e-01
-7.29773939e-02 -1.11399722e+00 -5.66632211e-01 1.11737406e+00
-1.25387287e+00 -9.00919318e-01 -1.46581158e-01 2.16748968e-01
3.23249459e-01 -2.34078337e-03 1.04807544e+00 2.71277308e-01
-3.57293963e-01 3.06003578e-02 3.15036058e-01 -1.09590039e-01
8.74166310e-01 -1.54814529e+00 4.63362247e-01 6.68564975e-01
-1.90420255e-01 1.26775098e+00 7.11236119e-01 -4.38900292e-01
-1.49720860e+00 -7.01081693e-01 6.08046710e-01 -6.92421615e-01
4.83182698e-01 -6.71115637e-01 -1.16391599e+00 8.75706017e-01
1.25953242e-01 4.88245972e-02 7.79564083e-01 1.52898088e-01
-4.64543439e-02 2.83612669e-01 -1.62144721e+00 4.56885904e-01
6.15867257e-01 -6.62127674e-01 -4.26908106e-01 5.76387763e-01
3.62192065e-01 -3.56453300e-01 -1.11586213e+00 -3.63473035e-02
5.21868408e-01 -1.15628481e+00 8.69615197e-01 -7.36232042e-01
4.76654351e-01 -5.36374927e-01 1.34468183e-01 -1.23071134e+00
-7.08102643e-01 -4.97909307e-01 2.56748080e-01 6.62769377e-01
4.02136266e-01 -4.92309928e-01 9.39986050e-01 6.97434008e-01
-1.76528126e-01 -7.84651041e-01 -7.62008190e-01 -4.99693424e-01
1.05478130e-01 2.12064043e-01 4.59749013e-01 7.91292667e-01
2.72191674e-01 4.84667122e-01 -7.34849228e-03 1.30778283e-01
8.52432489e-01 3.38316947e-01 8.53793979e-01 -1.14176869e+00
-2.72516310e-01 -1.04966171e-01 -6.61174357e-01 -1.00722241e+00
-5.78556806e-02 -5.61207950e-01 9.82215255e-02 -1.40526617e+00
7.17334509e-01 3.31115648e-02 -8.90870988e-02 2.78037399e-01
-2.73728427e-02 3.24906230e-01 -1.24857441e-01 4.71827775e-01
-6.71818078e-01 -2.05877586e-03 1.21488631e+00 1.04332492e-01
-1.39536115e-03 -5.20896554e-01 -3.86202753e-01 6.33086681e-01
6.25390947e-01 -4.55342978e-01 7.09025189e-02 -3.80200058e-01
7.05249980e-02 -2.84316361e-01 2.84060329e-01 -1.14863241e+00
3.79239112e-01 -1.05406016e-01 9.68989849e-01 -9.53602254e-01
3.49417895e-01 -7.65734851e-01 4.99823019e-02 5.64958155e-02
2.87241321e-02 3.48691165e-01 8.43216851e-02 2.02017933e-01
9.43158865e-02 -1.71173722e-01 1.33913243e+00 -8.16348374e-01
-3.36449236e-01 1.88986346e-01 -6.03921831e-01 -3.28256279e-01
8.94964099e-01 -2.37214386e-01 -6.74490929e-01 -1.74266286e-02
-9.22929227e-01 3.60649526e-01 1.32928085e+00 -1.91293538e-01
6.01277709e-01 -7.35359967e-01 -1.80943951e-01 4.61009949e-01
6.64721131e-02 3.04640025e-01 5.41999519e-01 8.00910652e-01
-9.53976095e-01 2.92040944e-01 -7.84502387e-01 -9.22753811e-01
-1.32661903e+00 6.93570435e-01 5.01996040e-01 -3.07173938e-01
-7.67005920e-01 6.69342160e-01 1.35756657e-01 -7.46733010e-01
-1.37997031e-01 -2.55606741e-01 1.26640666e-02 -5.05371571e-01
1.07241511e+00 8.51704106e-02 2.60503054e-01 -3.15065414e-01
-3.10757846e-01 4.08942193e-01 -7.10043430e-01 1.07610933e-01
1.84640610e+00 -1.97927624e-01 -3.20496976e-01 4.61058408e-01
1.16606283e+00 -6.61624014e-01 -1.32922244e+00 2.21706554e-01
-1.63867369e-01 -3.00530702e-01 -1.58816233e-01 -4.82820511e-01
-6.78059936e-01 9.83382046e-01 6.16542041e-01 9.93209928e-02
9.15322602e-01 1.37860760e-01 5.42695284e-01 7.30902970e-01
7.78102815e-01 -1.16607344e+00 -2.46435046e-01 5.37531614e-01
3.06021273e-01 -1.43046844e+00 1.00322761e-01 7.61672482e-02
-1.86273232e-02 1.39275980e+00 7.29976356e-01 -5.16955145e-02
2.89206296e-01 7.64519572e-01 3.81550372e-01 -5.91859937e-01
-1.00641501e+00 2.76495725e-01 -5.39521873e-01 8.96237552e-01
6.00950241e-01 -2.08357751e-01 1.24209918e-01 3.55603918e-02
1.11230597e-01 7.05845580e-02 6.70165122e-01 1.39901078e+00
-7.22633064e-01 -1.12580287e+00 -4.79591280e-01 6.13187969e-01
-7.52693772e-01 2.22525731e-01 -4.16431218e-01 4.69476223e-01
-1.88476130e-01 6.14342630e-01 1.68488577e-01 -3.14031184e-01
1.45173907e-01 1.89844459e-01 6.65619373e-01 -6.31554246e-01
-4.81299967e-01 1.98193163e-01 -3.79724592e-01 -5.20474613e-01
-5.07774055e-01 -3.68064970e-01 -1.46709669e+00 -6.11935318e-01
-2.73402393e-01 2.46411130e-01 7.85059392e-01 8.10229719e-01
7.80800462e-01 3.10178012e-01 2.40082443e-01 -1.60660219e+00
-2.92627007e-01 -1.25376713e+00 -7.71267116e-01 5.42035937e-01
2.70433575e-01 -4.11328614e-01 -6.15058064e-01 3.69348496e-01]
|
[13.464141845703125, -3.0876641273498535]
|
ac5430dd-5227-4233-866b-cbba4649da9e
|
procst-boosting-semantic-segmentation-using
|
2204.11891
| null |
https://arxiv.org/abs/2204.11891v2
|
https://arxiv.org/pdf/2204.11891v2.pdf
|
ProCST: Boosting Semantic Segmentation Using Progressive Cyclic Style-Transfer
|
Using synthetic data for training neural networks that achieve good performance on real-world data is an important task as it can reduce the need for costly data annotation. Yet, synthetic and real world data have a domain gap. Reducing this gap, also known as domain adaptation, has been widely studied in recent years. Closing the domain gap between the source (synthetic) and target (real) data by directly performing the adaptation between the two is challenging. In this work, we propose a novel two-stage framework for improving domain adaptation techniques on image data. In the first stage, we progressively train a multi-scale neural network to perform image translation from the source domain to the target domain. We denote the new transformed data as "Source in Target" (SiT). Then, we insert the generated SiT data as the input to any standard UDA approach. This new data has a reduced domain gap from the desired target domain, which facilitates the applied UDA approach to close the gap further. We emphasize the effectiveness of our method via a comparison to other leading UDA and image-to-image translation techniques when used as SiT generators. Moreover, we demonstrate the improvement of our framework with three state-of-the-art UDA methods for semantic segmentation, HRDA, DAFormer and ProDA, on two UDA tasks, GTA5 to Cityscapes and Synthia to Cityscapes.
|
['Raja Giryes', 'Shady Abu-Hussein', 'Shahaf Ettedgui']
|
2022-04-25
| null | null | null | null |
['synthetic-to-real-translation']
|
['computer-vision']
|
[ 6.23113275e-01 2.55147964e-01 -8.80831331e-02 -4.20441359e-01
-1.12068117e+00 -6.89774215e-01 7.26453662e-01 -2.33097941e-01
-5.51720262e-01 8.45858693e-01 -2.37702489e-01 -3.47156942e-01
3.02304000e-01 -9.27585006e-01 -1.08320963e+00 -5.42830229e-01
8.02173436e-01 8.04712534e-01 4.94399607e-01 -3.26007843e-01
-2.35252708e-01 1.92070037e-01 -1.15568447e+00 3.76421452e-01
1.11282027e+00 9.60427165e-01 4.59723026e-01 3.62560660e-01
-2.09563598e-01 2.87385434e-01 -8.74394059e-01 -4.89742279e-01
3.93913507e-01 -8.28876019e-01 -1.09700656e+00 2.04775468e-01
2.51469612e-01 -1.64356664e-01 1.79620430e-01 1.05642211e+00
6.27446294e-01 1.57476217e-01 6.68504417e-01 -1.20070255e+00
-9.25144672e-01 4.87594724e-01 -5.23763299e-01 -2.04434603e-01
5.55506609e-02 3.36842015e-02 6.47551119e-01 -7.65805721e-01
1.08485603e+00 1.14316642e+00 4.94790673e-01 9.51547563e-01
-1.28054976e+00 -5.76419532e-01 -3.53971459e-02 1.42791405e-01
-1.18966842e+00 -2.09705085e-01 7.57682920e-01 -4.41797346e-01
8.25008094e-01 -5.91665320e-02 3.81808817e-01 1.32075250e+00
-5.30414283e-01 9.11234140e-01 1.17550075e+00 -8.19122493e-01
3.30443859e-01 3.28657925e-01 -2.94760704e-01 3.25360835e-01
-3.05193383e-02 -1.07549146e-01 -1.78220436e-01 3.48131716e-01
8.59284401e-01 -5.55325449e-01 -1.97957888e-01 -4.25612181e-01
-1.19341981e+00 8.41315687e-01 3.55669975e-01 2.80551225e-01
-2.11229250e-01 -2.18098357e-01 6.40283108e-01 3.76333356e-01
5.72397768e-01 4.75418091e-01 -6.46397948e-01 6.65251818e-03
-7.81541526e-01 2.60144830e-01 5.42565525e-01 1.06727660e+00
6.10707760e-01 1.42084509e-01 -1.37621224e-01 1.35939085e+00
-2.89654057e-03 5.60095847e-01 8.60039115e-01 -7.93860734e-01
6.72230363e-01 5.96339047e-01 1.78591609e-01 -4.27075326e-01
-9.91888940e-02 -1.93494365e-01 -8.88003469e-01 1.39441758e-01
6.98354423e-01 -1.71906546e-01 -1.35598040e+00 1.85329723e+00
5.01441300e-01 1.11555994e-01 6.11811638e-01 7.51137614e-01
8.01018596e-01 7.83414185e-01 9.95972827e-02 1.04687385e-01
1.24106586e+00 -1.18716943e+00 -4.40450579e-01 -5.48627198e-01
6.73809171e-01 -7.03713655e-01 1.56621385e+00 1.61472186e-01
-9.39146042e-01 -8.94289672e-01 -1.04552317e+00 -1.71341687e-01
-6.59851372e-01 5.13330698e-01 1.29411677e-02 4.75048870e-01
-1.03296602e+00 3.75843346e-01 -6.14610136e-01 -6.23353660e-01
4.36955482e-01 1.83309212e-01 -3.17670554e-01 -5.97809888e-02
-1.45834434e+00 8.88963282e-01 9.27820325e-01 -3.04800570e-01
-6.54790878e-01 -7.70813525e-01 -1.00261092e+00 -2.80468851e-01
5.37145257e-01 -6.96960449e-01 1.64470363e+00 -1.35046959e+00
-1.75357616e+00 1.25327480e+00 9.44330171e-02 -7.48153687e-01
7.74805784e-01 -9.24550220e-02 -4.68630344e-01 2.38707033e-03
4.11410064e-01 1.14492798e+00 9.93494570e-01 -1.27583349e+00
-7.39898741e-01 -2.19700962e-01 -1.01041973e-01 2.51415998e-01
-8.62327516e-02 -1.85054436e-01 -6.44509733e-01 -8.90120208e-01
-8.44625384e-02 -1.01270843e+00 -1.39450461e-01 -2.29981840e-01
-3.20929497e-01 -2.39891991e-01 8.99638593e-01 -7.17993915e-01
8.06959212e-01 -2.14055824e+00 2.56035507e-01 5.42229861e-02
-3.45168859e-01 7.70749509e-01 -3.72933865e-01 4.03658561e-02
-3.33261341e-01 -1.95724727e-03 -6.68193042e-01 -3.18506867e-01
-7.47574493e-02 3.88609082e-01 -3.21798891e-01 -3.17301303e-02
3.97187084e-01 9.33294177e-01 -8.34356725e-01 -4.26992595e-01
1.23775244e-01 2.15490326e-01 -4.29378361e-01 3.63815784e-01
-6.30082488e-01 7.41666377e-01 -1.59508616e-01 1.50376379e-01
7.24462450e-01 -9.21076685e-02 7.24212453e-02 -1.07700244e-01
2.98819207e-02 1.91100240e-01 -1.04318130e+00 1.91907990e+00
-7.51652658e-01 5.75653911e-01 -2.60124385e-01 -1.17041743e+00
1.18322062e+00 3.40446472e-01 2.28409678e-01 -1.09519017e+00
5.80772869e-02 4.97134566e-01 -6.48321062e-02 -3.84539068e-01
5.35177767e-01 -3.39015722e-01 -3.90333951e-01 3.31921250e-01
3.10005754e-01 -4.84483629e-01 4.98437703e-01 -1.21888600e-01
4.16765064e-01 5.68013906e-01 3.20060372e-01 5.55098429e-02
6.21778786e-01 4.77281332e-01 6.05540514e-01 4.80179965e-01
-1.43492624e-01 9.63477135e-01 5.71376503e-01 -2.16527075e-01
-1.50048542e+00 -9.35229123e-01 1.51036173e-01 8.93482149e-01
4.77434248e-02 1.52671993e-01 -1.42440426e+00 -1.00034344e+00
-3.33930701e-01 9.45535243e-01 -6.23574317e-01 -2.20043600e-01
-6.07195497e-01 -6.52855456e-01 7.67765641e-01 7.12674797e-01
1.12072837e+00 -1.31897938e+00 -4.72323149e-01 2.15593353e-01
-4.97445226e-01 -1.54644287e+00 -5.81056535e-01 9.09807980e-02
-6.51263952e-01 -7.22621679e-01 -1.20392990e+00 -1.03317297e+00
6.11555755e-01 4.34041619e-02 1.22060525e+00 -4.95957106e-01
8.17425177e-02 -5.94628565e-02 -5.23971975e-01 -4.66802448e-01
-1.13220859e+00 3.65702093e-01 -2.59306461e-01 1.50593352e-02
2.82302350e-01 -2.56589353e-01 -3.74440044e-01 5.24602234e-01
-1.28751099e+00 4.79767233e-01 5.16482472e-01 9.43967462e-01
8.82986069e-01 -1.16052195e-01 8.52558613e-01 -1.15072155e+00
6.67683482e-01 -2.54527658e-01 -7.74721503e-01 2.57941931e-01
-3.86938334e-01 1.55806080e-01 8.83084595e-01 -5.60039282e-01
-1.44305670e+00 3.63688082e-01 -4.99595582e-01 -3.46460491e-01
-4.81633186e-01 4.03887570e-01 -5.22618830e-01 3.12367141e-01
1.08221030e+00 2.75345802e-01 -1.09264493e-01 -4.15069103e-01
6.28511667e-01 8.55504215e-01 8.38207066e-01 -5.80699086e-01
6.99078500e-01 3.19942594e-01 -4.02876198e-01 -6.39404297e-01
-8.23384464e-01 -3.56667727e-01 -9.77565289e-01 1.92613855e-01
9.96094167e-01 -8.21759403e-01 1.21703133e-01 7.43279874e-01
-1.41297829e+00 -7.82720566e-01 -6.63118243e-01 1.60434559e-01
-8.37194741e-01 1.18267640e-01 -3.27742606e-01 -1.77520350e-01
-3.88890922e-01 -1.24908507e+00 1.15610754e+00 2.58174360e-01
-1.33669838e-01 -1.03650534e+00 1.24319039e-01 3.47349018e-01
1.05545849e-01 3.44575524e-01 9.33061182e-01 -7.71250606e-01
-1.40177712e-01 -1.84713453e-02 -4.63551432e-01 8.53395998e-01
1.42024770e-01 -2.98749924e-01 -7.43913949e-01 -3.65019776e-03
-1.56915709e-01 -4.24769521e-01 5.81627846e-01 3.44378740e-01
8.95542502e-01 1.15524782e-02 -1.62463993e-01 5.83417952e-01
1.42744827e+00 4.75197911e-01 7.47645080e-01 6.20680928e-01
6.40991986e-01 4.65948611e-01 9.47913289e-01 3.46161462e-02
3.21800113e-01 9.57797647e-01 -2.77450308e-02 -5.88411510e-01
-5.75607479e-01 -4.63469237e-01 2.98146218e-01 5.79063416e-01
2.08137780e-01 -3.25037301e-01 -1.03124630e+00 8.90009344e-01
-1.86780155e+00 -5.29703140e-01 -5.45448661e-02 2.12835336e+00
1.11818278e+00 2.15377197e-01 3.90864253e-01 -4.90403995e-02
7.98888981e-01 -2.43232250e-01 -7.40236938e-01 -5.90456128e-01
-2.02433184e-01 3.17153782e-01 5.11846662e-01 2.62552053e-01
-1.16216528e+00 1.45508373e+00 5.20732546e+00 1.17580223e+00
-1.18371153e+00 3.55064273e-01 7.37006187e-01 4.93489116e-01
9.13567748e-03 -2.78090477e-01 -7.76422501e-01 3.80729526e-01
1.09853017e+00 -1.26499653e-01 2.09648326e-01 1.10932899e+00
2.82952726e-01 2.25377716e-02 -1.09183180e+00 8.67834628e-01
8.96926299e-02 -1.26537526e+00 2.51902252e-01 -2.93292850e-01
1.01315331e+00 -1.10936634e-01 2.94564702e-02 3.64243954e-01
5.66603780e-01 -7.25181878e-01 8.40968788e-01 -1.33877814e-01
1.21100998e+00 -6.62813187e-01 7.38578498e-01 3.19827080e-01
-9.42573249e-01 3.12263727e-01 -3.45372707e-01 3.93193781e-01
1.98308408e-01 3.25101137e-01 -1.31274855e+00 6.94331408e-01
5.38466513e-01 4.87762183e-01 -5.10178983e-01 6.82657242e-01
-4.91365463e-01 5.51573873e-01 -1.93211332e-01 4.08022434e-01
4.10550088e-01 -3.41328353e-01 2.41163060e-01 1.11997461e+00
3.97226840e-01 -3.15056264e-01 -6.12893440e-02 1.02434683e+00
-3.70603532e-01 1.74337506e-01 -7.95817435e-01 1.08236365e-01
2.11720154e-01 8.61558199e-01 -9.30659652e-01 -6.74671113e-01
-3.83011729e-01 1.46717954e+00 5.59143201e-02 3.61392856e-01
-1.05448830e+00 -4.70528603e-01 3.31622809e-01 1.93443429e-03
2.86006242e-01 2.53303465e-03 -3.72970879e-01 -9.93355989e-01
1.05752654e-01 -1.10654008e+00 4.52097923e-01 -8.45725715e-01
-9.89907682e-01 9.83592927e-01 1.49542153e-01 -1.43867362e+00
-5.61367929e-01 -5.06192148e-01 -2.68716902e-01 1.01846266e+00
-1.51810813e+00 -1.48883343e+00 -2.79955357e-01 6.30958259e-01
1.07587957e+00 -1.67832449e-01 7.18101442e-01 4.57831264e-01
-4.55200881e-01 7.24660754e-01 1.87856600e-01 3.36361468e-01
8.65616977e-01 -1.32453299e+00 1.04896843e+00 9.55574274e-01
8.57714564e-02 -5.71181253e-02 5.51070094e-01 -6.53386354e-01
-4.36926156e-01 -1.43901694e+00 6.26766503e-01 -2.14676008e-01
4.28300172e-01 -3.95406097e-01 -1.07279539e+00 6.79604828e-01
2.11491734e-01 -1.47148594e-01 2.32757136e-01 -3.80044371e-01
-2.85895824e-01 -4.61584292e-02 -1.30159795e+00 7.07194984e-01
8.97766352e-01 -2.76170969e-01 -5.61051071e-01 3.62441212e-01
1.05918539e+00 -7.00198293e-01 -7.22941875e-01 3.18125188e-01
2.20735088e-01 -7.55827367e-01 1.00855577e+00 -4.58912551e-01
6.38741434e-01 -3.16827357e-01 3.64125296e-02 -1.79195678e+00
2.01523125e-01 -5.03907740e-01 5.02049804e-01 1.41514504e+00
6.69972837e-01 -6.10019386e-01 7.97404110e-01 3.53867620e-01
-2.68066764e-01 -2.94382811e-01 -8.28537226e-01 -1.04700482e+00
4.76569116e-01 -3.36185873e-01 6.95225656e-01 9.77408409e-01
-5.06182492e-01 4.99772191e-01 -1.99012563e-01 -9.64530483e-02
2.44285047e-01 9.10222530e-02 9.48401213e-01 -9.12647545e-01
-2.60436893e-01 -2.44896963e-01 -9.96302161e-03 -1.10224211e+00
8.22890773e-02 -8.65708232e-01 2.45519638e-01 -1.64137793e+00
-1.35366783e-01 -5.06601453e-01 1.65287212e-01 5.56529701e-01
-1.10270150e-01 3.73102307e-01 2.09678918e-01 1.45657703e-01
-1.65882692e-01 4.62129176e-01 1.56856310e+00 -1.64767608e-01
-4.97409672e-01 6.55101463e-02 -4.25154716e-01 6.59089327e-01
9.25918281e-01 -4.29045975e-01 -6.25347972e-01 -5.77823102e-01
-2.09016666e-01 -6.84998855e-02 1.21167034e-01 -9.79268253e-01
-2.56967992e-01 -9.01450142e-02 3.75525877e-02 -3.18895817e-01
2.16325194e-01 -6.69524372e-01 3.93299684e-02 2.25937232e-01
-3.03212702e-01 -1.58254966e-01 4.33802724e-01 1.85678318e-01
-4.79843616e-01 -4.09411639e-01 1.22442222e+00 -2.16547385e-01
-1.04328716e+00 -8.85583460e-02 1.57439023e-01 2.89808422e-01
1.15881991e+00 -3.63519073e-01 -1.34579375e-01 -2.64643282e-01
-7.78256297e-01 2.80887261e-03 4.67996508e-01 5.21231890e-01
3.98166388e-01 -1.25690281e+00 -7.86764205e-01 2.25420475e-01
2.49882147e-01 5.16511977e-01 2.60315150e-01 5.31820297e-01
-5.82417130e-01 3.85214686e-01 -4.39244926e-01 -6.22664034e-01
-1.19950008e+00 5.01717091e-01 4.72800016e-01 -4.72414196e-01
-4.77472484e-01 7.42547750e-01 3.62792701e-01 -7.25996971e-01
-1.10506169e-01 -4.31789130e-01 -7.61477649e-02 4.61971089e-02
1.94329143e-01 -2.16265488e-02 3.09876204e-01 -6.70795143e-01
-1.25661364e-03 4.85307276e-01 -1.11278042e-01 -3.48815113e-01
1.22716844e+00 -1.23776183e-01 2.12689251e-01 1.80913195e-01
1.16236317e+00 -4.71236438e-01 -1.40106797e+00 -3.53047788e-01
4.33470011e-02 -2.39118010e-01 -3.95425707e-01 -1.10408163e+00
-9.99943078e-01 1.02483058e+00 7.26004422e-01 1.42001301e-01
1.34714997e+00 2.14439034e-01 1.15173447e+00 8.10354128e-02
8.21311548e-02 -1.37333632e+00 1.32298782e-01 4.44957078e-01
9.10858154e-01 -1.34067833e+00 -5.54983139e-01 -5.75586200e-01
-1.13161969e+00 8.09957445e-01 8.06449294e-01 1.18075004e-02
1.07191093e-02 3.87366116e-02 2.74417400e-01 3.08142930e-01
-2.28040725e-01 -4.62760061e-01 1.83953673e-01 1.04317212e+00
1.76151097e-01 -4.28022034e-02 -2.61142522e-01 7.07597733e-01
-2.52048641e-01 3.10163319e-01 5.04209399e-01 6.27056956e-01
-6.53824210e-02 -1.56242692e+00 -4.43539530e-01 -1.26613239e-02
-3.33565891e-01 1.14174709e-02 -3.75153840e-01 1.10851336e+00
3.75817686e-01 6.19793773e-01 3.35918218e-02 -1.13192573e-01
7.58460343e-01 2.09445655e-01 3.06621224e-01 -7.31193960e-01
-2.40804195e-01 8.35445896e-02 1.12964734e-01 -8.16940367e-02
-4.47811812e-01 -6.16274536e-01 -1.33821714e+00 2.37869903e-01
-3.80133539e-02 -1.53799327e-02 8.79991055e-01 1.03699768e+00
4.17842954e-01 6.58285022e-01 2.84628630e-01 -5.38466036e-01
-3.38382125e-01 -1.02651489e+00 -1.03076354e-01 7.66338170e-01
-1.57065019e-02 -5.69227278e-01 2.33155727e-01 6.62034512e-01]
|
[9.795074462890625, 1.4056127071380615]
|
39bcb9e4-da00-45a2-99c1-b73d632be13f
|
improving-multimodal-named-entity-recognition
| null | null |
https://aclanthology.org/2020.acl-main.306
|
https://aclanthology.org/2020.acl-main.306.pdf
|
Improving Multimodal Named Entity Recognition via Entity Span Detection with Unified Multimodal Transformer
|
In this paper, we study Multimodal Named Entity Recognition (MNER) for social media posts. Existing approaches for MNER mainly suffer from two drawbacks: (1) despite generating word-aware visual representations, their word representations are insensitive to the visual context; (2) most of them ignore the bias brought by the visual context. To tackle the first issue, we propose a multimodal interaction module to obtain both image-aware word representations and word-aware visual representations. To alleviate the visual bias, we further propose to leverage purely text-based entity span detection as an auxiliary module, and design a Unified Multimodal Transformer to guide the final predictions with the entity span predictions. Experiments show that our unified approach achieves the new state-of-the-art performance on two benchmark datasets.
|
['Rui Xia', 'Jing Jiang', 'Jianfei Yu', 'Li Yang']
|
2020-07-01
| null | null | null |
acl-2020-6
|
['multi-modal-named-entity-recognition']
|
['natural-language-processing']
|
[ 6.82794675e-02 -1.06501184e-01 -2.49303490e-01 -2.34220341e-01
-9.55666304e-01 -4.92570609e-01 5.19396067e-01 4.16696519e-02
-4.38137263e-01 4.02268618e-01 4.54981565e-01 -3.45726311e-01
3.78908575e-01 -5.54048657e-01 -6.48997009e-01 -4.22759175e-01
3.94113153e-01 -3.30538899e-02 5.09781539e-01 -2.27543727e-01
3.10596764e-01 4.50841337e-01 -1.30673873e+00 5.49311161e-01
9.83348846e-01 1.12168002e+00 1.97236910e-01 5.09958267e-01
-6.14884853e-01 8.41722846e-01 -3.97043943e-01 -6.02283180e-01
-1.51485249e-01 -3.82181257e-01 -6.56915486e-01 3.68221164e-01
3.55677873e-01 -1.78403705e-01 -4.82595742e-01 9.00485575e-01
5.81874490e-01 2.42653992e-02 7.81855047e-01 -1.31076109e+00
-9.46903884e-01 5.00619173e-01 -9.43910241e-01 -1.52935088e-01
4.88663316e-01 1.96050465e-01 1.05656672e+00 -1.35969651e+00
7.89712548e-01 1.29566264e+00 4.63603795e-01 4.30569977e-01
-1.18326187e+00 -5.19637108e-01 5.39078236e-01 3.11234683e-01
-1.63560104e+00 -5.75475156e-01 9.56579685e-01 -3.92694116e-01
6.34362400e-01 1.01088710e-01 3.87228698e-01 1.15032375e+00
-1.03991523e-01 1.16217387e+00 9.53828275e-01 -4.01645750e-01
-2.74185315e-02 1.75828055e-01 -9.36415419e-03 8.68458152e-01
-9.71927717e-02 -3.39115322e-01 -4.48500037e-01 -1.39250368e-01
4.98054117e-01 1.15893587e-01 -3.22621256e-01 -6.65579319e-01
-1.32994747e+00 7.11407781e-01 5.07690728e-01 2.28000179e-01
-2.63894647e-01 1.58774406e-01 5.02108634e-01 -1.03931375e-01
4.28270131e-01 1.11906670e-01 -1.63427994e-01 1.57377079e-01
-8.16944182e-01 -7.00691119e-02 5.11312783e-01 9.95059431e-01
8.61328900e-01 -8.77479017e-02 -6.03395998e-01 9.13928092e-01
5.17428041e-01 5.63156009e-01 5.12388468e-01 -2.42729336e-01
7.56514192e-01 9.14505303e-01 1.73021093e-01 -1.20116794e+00
-4.25990105e-01 -8.46677050e-02 -6.33084059e-01 -1.39814511e-01
2.37553209e-01 -2.61689294e-02 -1.05057943e+00 1.52909911e+00
2.50551254e-01 -1.05182700e-01 2.82280818e-02 8.61473799e-01
1.23736262e+00 6.89739883e-01 5.44570506e-01 -8.09364319e-02
1.64299154e+00 -1.23840523e+00 -7.84845829e-01 -3.63216430e-01
6.18587136e-01 -8.56143832e-01 1.12785256e+00 -1.84378564e-01
-9.18703020e-01 -3.83079439e-01 -1.01196265e+00 -1.93524703e-01
-6.15166128e-01 5.24664521e-01 2.67294437e-01 4.81669962e-01
-7.71720350e-01 -3.10017895e-02 -6.30488455e-01 -5.08638680e-01
3.54222059e-01 -3.76760885e-02 -4.58095908e-01 -7.40624517e-02
-1.00860536e+00 7.89317846e-01 4.00679767e-01 1.52647778e-01
-6.19867027e-01 -3.57696354e-01 -1.12910283e+00 1.19909614e-01
4.05317098e-01 -6.92216933e-01 1.02966714e+00 -1.14502740e+00
-1.18335509e+00 8.49170268e-01 -5.41044652e-01 6.99468479e-02
5.82938135e-01 -1.44769877e-01 -5.87496579e-01 2.88917720e-01
-1.73561066e-01 7.44262815e-01 7.24194050e-01 -1.77513850e+00
-5.28134942e-01 -3.14630747e-01 -3.43340598e-02 2.32069656e-01
-7.76493132e-01 8.04869384e-02 -1.10275292e+00 -9.32154179e-01
4.49552387e-02 -7.53589928e-01 -2.36516848e-01 1.22218870e-01
-7.05912828e-01 -1.33586988e-01 9.63060737e-01 -7.27027178e-01
1.49346900e+00 -2.06958842e+00 1.70380145e-01 7.46159852e-02
3.44124764e-01 3.18519145e-01 -5.98212481e-01 6.79473996e-01
-6.88658208e-02 2.05176964e-01 -1.86731182e-02 -7.02767313e-01
7.30400085e-02 -9.85394865e-02 -4.85746533e-01 2.88507670e-01
3.33279848e-01 1.22110045e+00 -7.87158668e-01 -1.02712917e+00
2.10900322e-01 6.41006529e-01 -2.67274827e-01 2.79663682e-01
-1.46564335e-01 2.61270583e-01 -5.32960236e-01 7.60745466e-01
8.45788658e-01 -3.35066587e-01 3.46548468e-01 -6.76101506e-01
-1.51551053e-01 -1.70803696e-01 -1.11085117e+00 1.56662273e+00
-4.79593933e-01 4.69589323e-01 -1.12266853e-01 -7.13781118e-01
7.90119946e-01 2.46568426e-01 2.33119071e-01 -6.27565145e-01
1.10369898e-01 -8.81665200e-02 -5.27221799e-01 -7.08722055e-01
7.78152049e-01 5.81145436e-02 -1.31775826e-01 2.76969373e-01
4.33507264e-02 4.94599074e-01 1.02750082e-02 3.79256845e-01
7.90615320e-01 2.73587197e-01 3.55446279e-01 2.60130674e-01
5.47708988e-01 -1.17069259e-01 5.81727803e-01 4.44870591e-01
-3.60195190e-01 8.32348287e-01 6.58332407e-01 -3.09144109e-01
-8.99680734e-01 -1.06787086e+00 1.82377979e-01 1.40248621e+00
5.56551576e-01 -5.49176335e-01 -5.28705299e-01 -1.01804829e+00
-3.00422460e-01 4.90782559e-01 -6.45588875e-01 -1.99079756e-02
-4.86128777e-01 -5.63978314e-01 5.71980894e-01 8.79479885e-01
3.75646979e-01 -9.43141699e-01 -2.73768395e-01 7.15443268e-02
-5.60259700e-01 -1.49882901e+00 -8.74468744e-01 -3.87982130e-01
-7.10648715e-01 -8.72979224e-01 -1.09221756e+00 -9.55232263e-01
7.73392200e-01 5.52246332e-01 9.41242695e-01 7.28975162e-02
-1.04821004e-01 5.94896019e-01 -5.80566585e-01 -1.53461769e-01
-5.19942492e-02 1.67130873e-01 -3.63688409e-01 3.32491010e-01
2.35510558e-01 -1.75764173e-01 -8.42642009e-01 3.92904013e-01
-1.02358174e+00 2.41928682e-01 7.49914467e-01 8.25019419e-01
6.11089110e-01 -5.79838574e-01 6.26664281e-01 -7.36860871e-01
4.63996559e-01 -3.64163429e-01 -3.19260269e-01 8.13800693e-01
-4.39088464e-01 -4.12311964e-02 3.94989222e-01 -7.48830020e-01
-1.16950512e+00 3.93252820e-01 -7.21194148e-02 -4.86216307e-01
-1.15572385e-01 4.77194130e-01 -4.99909818e-01 -3.84604596e-02
1.99849367e-01 3.19546878e-01 -3.82680088e-01 -6.18444622e-01
7.57059038e-01 6.94746852e-01 4.18324471e-01 -3.53618115e-01
8.27784717e-01 5.09479463e-01 -3.39419067e-01 -6.98013425e-01
-7.19973981e-01 -5.21957397e-01 -5.03555954e-01 -3.57525676e-01
1.17952704e+00 -1.09063447e+00 -6.19480669e-01 2.55451351e-01
-1.41228843e+00 1.19613595e-01 1.31154329e-01 2.39563808e-01
-3.30889374e-01 8.01074445e-01 -5.00072896e-01 -9.22144592e-01
-5.39389133e-01 -1.25082743e+00 1.29280496e+00 3.35845500e-01
4.51993085e-02 -7.88954139e-01 -5.53904846e-03 3.30334067e-01
3.00912887e-01 1.87023968e-01 9.06093717e-01 -6.29934311e-01
-5.05013943e-01 -2.56008595e-01 -8.26495707e-01 3.48384678e-02
-1.06593043e-01 5.30643873e-02 -1.03152788e+00 -4.15574946e-02
-6.42394841e-01 -3.03869039e-01 1.07753158e+00 -1.46491438e-01
1.17050266e+00 -1.69637650e-01 -6.07975245e-01 4.69265401e-01
1.41649842e+00 -1.30286261e-01 8.87538671e-01 2.46445790e-01
1.12625325e+00 5.30639470e-01 6.19884253e-01 5.11281908e-01
7.87656009e-01 7.15143025e-01 4.85625237e-01 -4.09553915e-01
-2.29775205e-01 -6.48585856e-01 4.45924997e-01 7.90432453e-01
9.06882063e-02 -5.30381978e-01 -8.83666813e-01 7.53360510e-01
-2.11848974e+00 -7.79938400e-01 -1.24473497e-02 1.83906269e+00
5.16809583e-01 -2.15634555e-01 2.18259320e-01 -3.33726615e-01
1.03244293e+00 4.95173156e-01 -3.42835397e-01 -1.10336885e-01
-2.88481265e-01 -3.43524188e-01 4.59144980e-01 7.42417276e-02
-1.27824521e+00 1.12502646e+00 6.24119520e+00 1.05556417e+00
-1.09353149e+00 2.00378031e-01 4.34087634e-01 1.90863445e-01
-5.06268084e-01 -1.90270618e-01 -7.17081010e-01 3.31694484e-01
3.77483815e-01 5.61201088e-02 4.28641513e-02 8.76812756e-01
1.33639565e-02 1.75151825e-01 -8.12559485e-01 1.14624619e+00
3.68456900e-01 -1.12051022e+00 4.24963295e-01 -6.16131499e-02
5.42878568e-01 -2.67590314e-01 3.78878266e-02 2.77616173e-01
-6.95737824e-02 -9.02521193e-01 8.77179980e-01 7.01128840e-01
1.03706503e+00 -8.44227791e-01 5.86177349e-01 -4.60521281e-02
-1.66931343e+00 6.12713583e-03 -2.87298709e-01 4.87095326e-01
4.75178957e-01 3.45046192e-01 -5.57980359e-01 6.27784491e-01
2.83806592e-01 4.41004723e-01 -8.56074452e-01 1.09023583e+00
-1.89435259e-01 3.37793320e-01 1.19778082e-01 -5.83425798e-02
1.28943384e-01 1.70359179e-01 4.78456914e-01 1.52578604e+00
2.76286870e-01 -2.07430437e-01 2.80255973e-01 6.52999222e-01
-2.46774852e-01 5.86122155e-01 -4.99911219e-01 -3.35748017e-01
4.51988369e-01 1.45081067e+00 -7.49418557e-01 -2.05705270e-01
-7.52347827e-01 1.37841320e+00 6.24568522e-01 6.16834700e-01
-8.87104213e-01 -5.28363466e-01 5.00023007e-01 -1.53784990e-01
7.35002756e-01 -2.99578488e-01 -2.50075907e-01 -1.56239343e+00
9.33360234e-02 -7.12873816e-01 4.64397013e-01 -9.43488419e-01
-1.45827758e+00 6.87698960e-01 -2.40608066e-01 -1.41204858e+00
1.90315425e-01 -6.84914291e-01 -6.70147657e-01 5.51239133e-01
-1.59835923e+00 -1.69902730e+00 -3.07615727e-01 5.86477578e-01
4.59948689e-01 1.32495277e-02 7.55860209e-01 4.62383002e-01
-8.07322204e-01 8.94513845e-01 -1.72213107e-01 4.04653996e-01
9.93577302e-01 -9.29785073e-01 1.92440465e-01 7.60717809e-01
1.68077722e-01 5.86859047e-01 4.55695182e-01 -6.17301524e-01
-1.63537872e+00 -1.18105900e+00 9.81111765e-01 -3.94723415e-01
6.91013515e-01 -3.66551161e-01 -7.81262219e-01 6.75832868e-01
3.50260615e-01 3.14730480e-02 8.49324882e-01 1.42534459e-02
-7.56077588e-01 6.14319108e-02 -8.59341085e-01 9.22081232e-01
9.47663903e-01 -6.00243330e-01 -5.97088099e-01 4.23870869e-02
8.93987596e-01 -2.99672008e-01 -7.79893994e-01 3.82785290e-01
7.26010680e-01 -6.63943529e-01 1.05116272e+00 -6.31366074e-01
5.54061115e-01 -5.05965650e-01 -2.33815670e-01 -1.06119180e+00
-1.28039092e-01 -1.83918595e-01 -2.20356062e-01 1.80640852e+00
5.39640903e-01 -3.76480609e-01 5.25026143e-01 4.63436991e-01
1.01761580e-01 -7.91462779e-01 -8.07819843e-01 -4.77631748e-01
-2.43494317e-01 -4.80921030e-01 6.09287798e-01 1.00743377e+00
1.00746088e-01 5.06863415e-01 -6.81158423e-01 2.17247665e-01
2.41216481e-01 2.90239871e-01 7.32123792e-01 -6.80627584e-01
5.87641858e-02 -3.48841429e-01 -4.43826288e-01 -1.26951623e+00
1.35809034e-01 -6.17383778e-01 1.50858492e-01 -1.83457899e+00
6.50873423e-01 -1.18652575e-01 -3.83633107e-01 5.94123304e-01
-4.42260653e-01 5.24293065e-01 5.41995585e-01 9.45593789e-02
-1.06604362e+00 7.98803210e-01 1.19269967e+00 -9.66594815e-02
-2.58114692e-02 -5.97667933e-01 -7.75759935e-01 7.30724096e-01
4.85971540e-01 -1.79132685e-01 -9.63672698e-02 -5.60670793e-01
2.81282514e-01 -5.00844605e-02 4.12967414e-01 -5.52494764e-01
1.27123892e-01 -2.31582984e-01 5.64880371e-01 -8.93745542e-01
2.80249029e-01 -7.75901914e-01 -3.03671658e-01 1.07095487e-01
-3.25104356e-01 1.48356378e-01 1.03381574e-01 7.84895420e-01
-2.78391480e-01 2.02805530e-02 6.40409887e-01 1.66749477e-01
-1.08088589e+00 2.82841474e-01 -4.52577055e-01 -4.74871360e-02
9.82456565e-01 9.43556651e-02 -6.61680162e-01 -4.67154056e-01
-5.34338593e-01 3.99843782e-01 6.22174561e-01 6.23714447e-01
7.53801942e-01 -1.62100852e+00 -4.51756477e-01 -2.27898657e-01
7.73674607e-01 -4.75427240e-01 5.61887324e-01 8.79812717e-01
-2.72322744e-01 1.62714645e-01 3.50576043e-02 -4.67859834e-01
-1.26626182e+00 8.97363544e-01 7.62333125e-02 -3.06754321e-01
-4.15680319e-01 5.77799320e-01 6.64767325e-01 -3.44924867e-01
2.14938894e-01 6.27815798e-02 -5.32306433e-01 4.12294328e-01
6.47460520e-01 2.31925458e-01 -6.49549067e-02 -1.06899726e+00
-4.74070787e-01 7.66490579e-01 -9.94918868e-03 -2.86245886e-02
9.76855814e-01 -4.63821024e-01 2.09663630e-01 4.18363869e-01
1.44112635e+00 1.21099584e-01 -9.68617022e-01 -4.06445056e-01
-6.17445819e-02 -3.54440808e-01 -9.78441089e-02 -6.80089951e-01
-1.24767506e+00 1.02648282e+00 7.21458972e-01 -9.60362554e-02
1.28928781e+00 1.17196195e-01 1.14880824e+00 2.20542490e-01
-6.00041300e-02 -1.00391340e+00 1.60770804e-01 4.85983580e-01
9.17670548e-01 -1.24320972e+00 -2.36423686e-02 -5.90605736e-01
-1.12519133e+00 1.18030989e+00 6.73752248e-01 3.06931555e-01
2.78357774e-01 -1.22516401e-01 3.30981851e-01 -1.42382488e-01
-5.59706271e-01 -6.51836812e-01 8.00998747e-01 5.14729023e-01
5.86262882e-01 -1.21770546e-01 -5.16266048e-01 9.15032268e-01
5.33689737e-01 -1.83970243e-01 9.53152925e-02 8.60252321e-01
-3.94900590e-01 -1.04637158e+00 -3.76390994e-01 1.48308156e-02
-3.75013620e-01 -2.01246545e-01 -4.97707695e-01 6.00138187e-01
-2.02694256e-02 9.48492348e-01 -2.93597262e-02 -5.17376423e-01
5.23383439e-01 2.59516031e-01 2.27753505e-01 -3.20481479e-01
-3.43116671e-01 2.60670155e-01 2.04012945e-01 -6.70082748e-01
-3.46296877e-01 -2.64457792e-01 -1.14143479e+00 6.62128925e-02
-3.39307696e-01 -2.31194362e-01 6.99985683e-01 8.77329648e-01
6.13050818e-01 4.49789554e-01 5.21047831e-01 -1.04230022e+00
-2.10868329e-01 -8.51164877e-01 -3.59443605e-01 6.26409709e-01
1.78791389e-01 -5.61920822e-01 -9.95897055e-02 1.29677206e-01]
|
[10.789027214050293, 1.5515800714492798]
|
46ec63d0-3049-47d7-a763-3b1af6f0cc48
|
neural-collaborative-ranking
|
1808.04957
| null |
http://arxiv.org/abs/1808.04957v1
|
http://arxiv.org/pdf/1808.04957v1.pdf
|
Neural Collaborative Ranking
|
Recommender systems are aimed at generating a personalized ranked list of
items that an end user might be interested in. With the unprecedented success
of deep learning in computer vision and speech recognition, recently it has
been a hot topic to bridge the gap between recommender systems and deep neural
network. And deep learning methods have been shown to achieve state-of-the-art
on many recommendation tasks. For example, a recent model, NeuMF, first
projects users and items into some shared low-dimensional latent feature space,
and then employs neural nets to model the interaction between the user and item
latent features to obtain state-of-the-art performance on the recommendation
tasks. NeuMF assumes that the non-interacted items are inherent negative and
uses negative sampling to relax this assumption. In this paper, we examine an
alternative approach which does not assume that the non-interacted items are
necessarily negative, just that they are less preferred than interacted items.
Specifically, we develop a new classification strategy based on the widely used
pairwise ranking assumption. We combine our classification strategy with the
recently proposed neural collaborative filtering framework, and propose a
general collaborative ranking framework called Neural Network based
Collaborative Ranking (NCR). We resort to a neural network architecture to
model a user's pairwise preference between items, with the belief that neural
network will effectively capture the latent structure of latent factors. The
experimental results on two real-world datasets show the superior performance
of our models in comparison with several state-of-the-art approaches.
|
['Xu Congfu', 'Cao Yi', 'Yang Xin', 'Song Bo']
|
2018-08-15
| null | null | null | null |
['collaborative-ranking']
|
['graphs']
|
[-2.44724508e-02 -3.91544342e-01 -2.73544461e-01 -6.97892964e-01
-3.88942510e-01 -3.12016934e-01 6.12056911e-01 -4.42689627e-01
-2.33189359e-01 4.36770231e-01 7.55373418e-01 -1.88543290e-01
-6.64947033e-01 -9.20064330e-01 -5.36899745e-01 -6.14535332e-01
-8.26315954e-02 4.94131744e-01 -3.15619886e-01 -5.23179352e-01
2.34411538e-01 1.89975560e-01 -1.70687103e+00 7.65965641e-01
9.76132810e-01 1.10153699e+00 1.76425487e-01 1.78949237e-01
-9.89343449e-02 6.95947886e-01 -3.69329095e-01 -4.87192422e-01
4.11502481e-01 -1.27786294e-01 -4.16520894e-01 -3.17666858e-01
3.05190653e-01 -4.95666504e-01 -5.21250486e-01 8.82925570e-01
5.04209280e-01 6.76748097e-01 8.90698612e-01 -9.47132111e-01
-1.43135166e+00 1.02701938e+00 -2.14031875e-01 -6.38466850e-02
1.50895432e-01 -4.73386943e-01 1.46382284e+00 -1.04568601e+00
1.52717680e-01 1.02845144e+00 6.15137517e-01 5.75070679e-01
-8.65486085e-01 -7.81783342e-01 6.89535081e-01 2.63303101e-01
-1.10694396e+00 -7.11379573e-02 8.05077970e-01 -4.16905016e-01
6.81128144e-01 1.90428853e-01 5.09461641e-01 1.31458521e+00
2.14023888e-01 9.23127651e-01 8.43720913e-01 -2.03607976e-01
1.14280572e-02 8.92478600e-02 3.83157641e-01 3.44470739e-01
2.23897342e-02 3.37861806e-01 -3.62706214e-01 -2.83691078e-01
7.47693598e-01 8.11704099e-01 -2.03841716e-01 -3.55002910e-01
-1.04070723e+00 1.16559625e+00 6.67190552e-01 2.81851709e-01
-5.63547015e-01 -2.02707693e-01 5.03273569e-02 2.83954442e-01
7.00888634e-01 4.13617879e-01 -5.42253971e-01 3.79277050e-01
-7.26063609e-01 2.33251929e-01 5.54433048e-01 4.66049135e-01
2.95178562e-01 -9.56540089e-03 -3.86219144e-01 1.28255272e+00
7.35672235e-01 1.30257145e-01 8.39290619e-01 -6.75908089e-01
2.10751861e-01 3.86974990e-01 2.89002150e-01 -1.05258095e+00
-1.94557324e-01 -9.49558914e-01 -1.15710020e+00 -5.28894849e-02
8.29065144e-02 -3.04461926e-01 -8.11185777e-01 1.77426994e+00
-1.58188149e-01 2.67370433e-01 2.06340000e-01 9.89860892e-01
8.84114861e-01 8.40389609e-01 -2.19230011e-01 -2.12922364e-01
7.90034056e-01 -1.11036623e+00 -5.72688878e-01 4.62489277e-02
2.09909290e-01 -6.28517151e-01 9.26448345e-01 6.29899442e-01
-8.61480296e-01 -8.66650641e-01 -9.42277253e-01 1.95822939e-01
-3.43635976e-01 3.98899049e-01 1.08814287e+00 5.14131904e-01
-1.00522530e+00 9.10061717e-01 -3.50496083e-01 9.93598625e-03
2.09619254e-01 7.55832434e-01 -1.10205248e-01 -2.97018159e-02
-1.55522883e+00 5.59560895e-01 6.16712272e-02 5.10185957e-01
-6.17932498e-01 -4.11132038e-01 -3.35706711e-01 3.88619423e-01
2.77362317e-01 -8.68835568e-01 1.22382879e+00 -1.22131371e+00
-1.87087774e+00 -3.90008613e-02 9.72336233e-02 -3.07839870e-01
2.39714608e-01 -5.80982149e-01 -6.61479354e-01 -6.28498614e-01
-2.47162074e-01 4.08384591e-01 5.70110559e-01 -1.08261526e+00
-7.85097778e-01 -2.43811488e-01 4.46306169e-01 4.20645773e-01
-7.54355788e-01 -5.55031374e-03 -3.32139015e-01 -7.55506039e-01
9.81434286e-02 -9.46887195e-01 -4.20816779e-01 -6.22827411e-01
-4.06542681e-02 -7.08356798e-01 3.38021129e-01 -1.93739384e-01
1.13925636e+00 -1.94465911e+00 1.13471515e-01 3.98486793e-01
1.62107468e-01 3.87398779e-01 -3.95807236e-01 3.60691845e-01
-1.94737896e-01 -1.53132290e-01 3.72902781e-01 -3.19613248e-01
1.23273410e-01 1.04324676e-01 -6.26952350e-01 2.05826074e-01
-3.17507684e-01 8.19860578e-01 -9.05919313e-01 2.69753754e-01
1.04202114e-01 8.29560101e-01 -8.83779049e-01 4.05076146e-01
-2.01832280e-01 2.10883364e-01 -4.07396615e-01 6.05566874e-02
5.75484574e-01 -3.09183598e-01 3.47482234e-01 -2.71868229e-01
-1.13844182e-02 6.33950770e-01 -1.22766757e+00 1.35978186e+00
-4.92083907e-01 2.73504466e-01 -4.87980306e-01 -9.15273309e-01
1.11166167e+00 3.48778158e-01 5.03799140e-01 -6.57881618e-01
7.15099424e-02 3.10938619e-02 2.47454181e-01 -2.05674827e-01
6.82161450e-01 -1.28981881e-02 8.18326697e-02 4.79232013e-01
-1.41987605e-02 8.03066254e-01 -1.24783173e-01 2.46041819e-01
7.74371266e-01 1.18696101e-01 -3.42542008e-02 -1.13636106e-01
4.31853682e-01 -6.28390193e-01 6.68915510e-01 9.88591194e-01
3.39328527e-01 4.58929271e-01 5.18080406e-02 -5.36096096e-01
-6.38247788e-01 -1.00366473e+00 2.07352966e-01 1.60188556e+00
1.29926484e-02 -3.22305769e-01 -2.12336734e-01 -8.51602733e-01
-2.02432454e-01 6.25210345e-01 -7.71114588e-01 -2.51866817e-01
-2.83230275e-01 -8.42723072e-01 8.56745336e-03 6.42235398e-01
2.05981970e-01 -1.27706027e+00 2.64445990e-01 3.09810460e-01
-1.46954745e-01 -3.11790586e-01 -5.43865085e-01 3.12323362e-01
-7.11790979e-01 -6.02770627e-01 -9.19875920e-01 -7.37730324e-01
5.79404593e-01 6.12248361e-01 1.12797546e+00 -1.79519914e-02
6.35299027e-01 7.32799023e-02 -6.92696333e-01 -6.46791980e-02
4.79195081e-02 4.33489904e-02 5.46532989e-01 4.49026763e-01
3.80830139e-01 -5.80315232e-01 -8.87159050e-01 5.08721471e-01
-8.34235132e-01 -8.77902284e-02 7.98723638e-01 9.78404403e-01
4.81634140e-01 2.87975103e-01 9.89614248e-01 -1.24334204e+00
9.80581641e-01 -9.00399923e-01 -2.31832758e-01 2.51470715e-01
-7.26415038e-01 3.29149030e-02 1.06874406e+00 -7.66056001e-01
-1.16924798e+00 -1.24752469e-01 -4.44859117e-01 -4.56268400e-01
-1.04775518e-01 8.50676954e-01 -2.92892516e-01 3.93930554e-01
3.83541822e-01 1.02182627e-01 -5.84595621e-01 -7.20243752e-01
5.40707767e-01 9.22835052e-01 3.25582594e-01 -4.10576910e-01
4.79904115e-01 2.69487709e-01 -4.70622689e-01 -2.16629818e-01
-1.27800512e+00 -5.16405463e-01 -3.80411983e-01 -6.78573549e-02
5.11679530e-01 -8.40271175e-01 -6.52803123e-01 2.25698486e-01
-9.55787182e-01 -8.83509144e-02 -6.75288588e-02 8.77754748e-01
-3.19782943e-01 2.35693380e-01 -5.71500361e-01 -8.53118718e-01
-6.02263868e-01 -1.01103473e+00 4.14854765e-01 2.56348968e-01
-2.33726893e-02 -8.78018260e-01 2.19022930e-01 2.86579460e-01
6.29288554e-01 -5.86088955e-01 8.53361011e-01 -1.10328948e+00
-2.89422274e-01 -2.52487689e-01 -2.68114179e-01 5.74630022e-01
2.05786243e-01 -1.36701658e-01 -7.44497478e-01 -2.23028377e-01
-1.82280302e-01 6.46823421e-02 1.12596679e+00 5.78613043e-01
1.29125428e+00 -2.99475312e-01 -1.20762467e-01 5.28772593e-01
1.14603543e+00 5.44080496e-01 6.54415071e-01 -1.85125209e-02
7.96867251e-01 5.64134479e-01 5.21440268e-01 3.08407545e-01
4.71540779e-01 7.11178541e-01 3.67829561e-01 -1.25236690e-01
3.94943744e-01 -2.81062454e-01 5.69468200e-01 1.13488364e+00
-3.53971958e-01 -6.84562325e-01 -2.58870155e-01 2.02996373e-01
-2.25726080e+00 -1.18038416e+00 -3.12949605e-02 2.36215281e+00
4.15398002e-01 1.22106634e-01 1.20828457e-01 -2.46353865e-01
6.71617210e-01 2.50393245e-02 -5.36784828e-01 -2.70948380e-01
8.49907473e-02 9.49363485e-02 1.81993753e-01 3.44855279e-01
-1.20595920e+00 8.04096818e-01 5.85324335e+00 7.56471992e-01
-1.06042874e+00 -1.00105042e-02 6.45458937e-01 -1.98439211e-01
-4.94425088e-01 -4.01638985e-01 -9.66498733e-01 5.23164690e-01
9.31365371e-01 4.14137952e-02 5.98082602e-01 8.13209534e-01
1.77978992e-01 5.51567197e-01 -1.11985576e+00 8.29229474e-01
2.74171144e-01 -1.22494256e+00 3.40104014e-01 3.17741483e-01
1.04744327e+00 2.08423123e-01 6.32101536e-01 7.13472486e-01
8.07588696e-01 -9.06395316e-01 3.95528764e-01 1.00758147e+00
3.44781607e-01 -9.11313653e-01 1.01662207e+00 4.11899030e-01
-9.56880629e-01 -4.44241434e-01 -8.16168249e-01 -3.12029779e-01
-4.73413393e-02 5.69230378e-01 -3.40288997e-01 5.62439322e-01
7.12134063e-01 9.80948687e-01 -2.37092435e-01 1.20989656e+00
-1.14316247e-01 6.55612528e-01 9.86010302e-03 -1.60984918e-01
2.26367816e-01 -5.41285157e-01 1.74298212e-01 8.97236168e-01
6.00463748e-01 6.60964176e-02 2.18500599e-01 5.08032739e-01
-3.77617687e-01 3.82643342e-01 -5.41453540e-01 6.19681552e-02
1.21416956e-01 1.36145794e+00 -4.70849067e-01 -4.19905335e-01
-5.40688932e-01 8.66430163e-01 4.66373056e-01 3.59321862e-01
-7.51880169e-01 -1.31930217e-01 6.61181152e-01 -1.53023809e-01
5.77025950e-01 3.25171351e-02 3.41323428e-02 -1.34904397e+00
-2.88214803e-01 -8.25838506e-01 3.39954406e-01 -6.73586309e-01
-1.87527478e+00 7.57875383e-01 -3.01311642e-01 -1.45121789e+00
-2.60351688e-01 -5.63221574e-01 -7.11058676e-01 8.59200597e-01
-1.28148925e+00 -9.55830932e-01 8.37772340e-02 6.88922703e-01
5.69435060e-01 -5.36894083e-01 8.83839071e-01 5.62431514e-01
-3.65523160e-01 7.20774174e-01 7.74185240e-01 5.30864447e-02
7.10695326e-01 -1.12398124e+00 2.05966577e-01 4.69455838e-01
5.39290965e-01 1.21350777e+00 2.48585269e-01 -5.78014672e-01
-1.07253218e+00 -1.14094305e+00 8.57216716e-01 -4.08948630e-01
4.19360697e-01 -3.41789216e-01 -7.91781664e-01 5.95603585e-01
1.09005384e-01 -2.86183745e-01 1.20736885e+00 7.78702199e-01
-5.38440406e-01 -1.83877826e-01 -6.71887279e-01 6.42181516e-01
1.05167294e+00 -3.57150644e-01 -7.18189061e-01 3.11620951e-01
6.58975661e-01 1.78379655e-01 -6.11564696e-01 5.00462234e-01
1.10456133e+00 -9.72389162e-01 1.08686697e+00 -8.63509238e-01
5.36166608e-01 -2.61405021e-01 -4.22106326e-01 -1.65761292e+00
-1.07649231e+00 -2.62900978e-01 -2.12110102e-01 1.04490399e+00
4.94222432e-01 -5.02532482e-01 9.24298227e-01 3.97359192e-01
-3.14092636e-01 -1.01836705e+00 -3.29837382e-01 -3.53015661e-01
1.33751309e-03 -2.78924733e-01 6.72772944e-01 9.60487306e-01
-1.60883129e-01 7.25694954e-01 -9.74464118e-01 1.46188468e-01
2.16252923e-01 3.05951744e-01 4.52633142e-01 -1.71215141e+00
-7.20357656e-01 -3.96369308e-01 1.17326409e-01 -1.48127854e+00
2.54149169e-01 -9.10863340e-01 -3.00502591e-02 -1.81617808e+00
2.54792184e-01 -4.48313117e-01 -1.27088642e+00 3.91069591e-01
-5.71539775e-02 4.74397540e-01 4.19826955e-02 3.41775924e-01
-9.35161591e-01 6.78568065e-01 1.08644664e+00 -1.40209690e-01
-4.94381189e-01 6.05191767e-01 -1.24984896e+00 8.48549902e-01
6.37785733e-01 -3.71532738e-01 -5.60611784e-01 -5.00694335e-01
7.09753454e-01 2.23639179e-02 -1.50847375e-01 -7.79684901e-01
2.25386396e-01 -1.26990631e-01 6.91047788e-01 -7.92919755e-01
3.86400759e-01 -9.04097259e-01 1.36064693e-01 1.17094353e-01
-7.80304790e-01 -6.17002994e-02 -4.74673152e-01 6.75981522e-01
-7.74806440e-02 -1.29381254e-01 3.67172658e-01 -4.99084638e-03
-4.64811504e-01 5.90940535e-01 -4.41402465e-01 -7.28604198e-01
4.16532546e-01 8.17931443e-02 -2.40087911e-01 -7.22903490e-01
-9.28729713e-01 -3.16513740e-02 -1.06973782e-01 8.47406685e-01
7.59399295e-01 -1.47016227e+00 -6.65352225e-01 3.36777896e-01
-9.75306332e-03 -5.85178256e-01 4.91085410e-01 3.46642584e-01
1.80333465e-01 4.87828493e-01 -9.49190408e-02 -8.20038915e-02
-1.02987349e+00 5.89331567e-01 2.24895105e-01 -5.76440454e-01
-2.91598350e-01 8.74521613e-01 5.22431970e-01 -7.16150284e-01
3.70211601e-01 -4.82491665e-02 -9.47355211e-01 1.51727512e-01
6.06033385e-01 2.37942189e-01 2.79154181e-02 -6.02613807e-01
-4.50377958e-03 2.73197025e-01 -6.69989645e-01 2.06562519e-01
1.60524094e+00 -4.38112356e-02 5.79819568e-02 4.27826375e-01
1.00298226e+00 1.77847743e-01 -9.79187012e-01 -4.76246715e-01
-2.89222568e-01 -3.66151094e-01 3.27278018e-01 -1.07750094e+00
-1.23791444e+00 7.17933774e-01 8.29765618e-01 3.93423170e-01
9.80587780e-01 -2.82354385e-01 7.49761343e-01 7.06546366e-01
2.87133366e-01 -9.32338774e-01 1.04717668e-02 8.46754193e-01
6.87301695e-01 -1.14513767e+00 -1.55459478e-01 5.15923612e-02
-5.90061367e-01 8.62065852e-01 6.11673653e-01 -5.33910751e-01
8.93892944e-01 -2.96941757e-01 5.34857474e-02 1.33998021e-01
-9.19183969e-01 -1.78838462e-01 7.64502525e-01 2.66649902e-01
9.15048242e-01 3.55696261e-01 -3.90837252e-01 1.24203336e+00
-9.41180736e-02 1.59923941e-01 1.56661168e-01 2.02440575e-01
-4.14873779e-01 -1.37471581e+00 -5.08702844e-02 9.83523667e-01
-6.09645605e-01 -3.69027644e-01 -2.46344164e-01 5.13851792e-02
3.28070730e-01 1.11014414e+00 7.30416924e-02 -8.46194804e-01
2.33908385e-01 -1.05194315e-01 1.17376320e-01 -8.11044037e-01
-7.70145595e-01 2.04802185e-01 -1.80014327e-01 -3.02081406e-01
-4.01599795e-01 -3.75324965e-01 -8.11458290e-01 -4.69041653e-02
-6.72041118e-01 5.43819368e-01 5.76010466e-01 9.65251088e-01
5.58293581e-01 6.43778265e-01 9.33339119e-01 -1.04966092e+00
-6.08125210e-01 -1.04111767e+00 -6.73822701e-01 3.85464132e-01
6.26305565e-02 -7.78622746e-01 -2.44906783e-01 -3.90144229e-01]
|
[10.159652709960938, 5.629669189453125]
|
5d169344-c9f3-4d89-b6f3-7be8dab11908
|
deepflash-turning-a-flash-selfie-into-a
|
1901.04252
| null |
https://arxiv.org/abs/1901.04252v2
|
https://arxiv.org/pdf/1901.04252v2.pdf
|
DeepFlash: Turning a Flash Selfie into a Studio Portrait
|
We present a method for turning a flash selfie taken with a smartphone into a photograph as if it was taken in a studio setting with uniform lighting. Our method uses a convolutional neural network trained on a set of pairs of photographs acquired in an ad-hoc acquisition campaign. Each pair consists of one photograph of a subject's face taken with the camera flash enabled and another one of the same subject in the same pose illuminated using a photographic studio-lighting setup. We show how our method can amend defects introduced by a close-up camera flash, such as specular highlights, shadows, skin shine, and flattened images.
|
['Ugo Erra', 'Fabio Ganovelli', 'Francesco Banterle', 'Roberto Scopigno', 'Paolo Cignoni', 'Nicola Capece']
|
2019-01-14
| null | null | null | null |
['3d-depth-estimation']
|
['computer-vision']
|
[ 7.93108046e-01 1.05902977e-01 6.54172719e-01 -4.99516577e-01
-3.23206604e-01 -5.18589914e-01 2.51061082e-01 -6.25546038e-01
-1.61723673e-01 6.74925685e-01 6.48875535e-02 -3.00646693e-01
4.92121667e-01 -5.75934350e-01 -1.07993519e+00 -5.17812669e-01
4.62021410e-01 -3.89704943e-01 3.35894823e-02 1.24955691e-01
-2.83549670e-02 6.16235077e-01 -1.62891340e+00 1.81394443e-01
7.42305443e-02 1.11015606e+00 1.34816423e-01 8.04991484e-01
4.57210600e-01 7.38258600e-01 -9.87771392e-01 -9.46927130e-01
7.74908304e-01 -3.47071618e-01 -3.81646395e-01 6.78807855e-01
1.58372188e+00 -1.14185584e+00 -6.27761185e-01 8.77802968e-01
2.90614009e-01 -8.26311857e-03 -1.15883827e-01 -1.05218923e+00
-7.47227490e-01 -3.05593431e-01 -5.95775425e-01 5.54241054e-02
5.84553540e-01 3.88566107e-01 1.81533068e-01 -6.42740011e-01
5.44958055e-01 1.07719278e+00 1.01357782e+00 4.30936724e-01
-1.06386352e+00 -6.45812750e-01 -4.12924588e-01 -1.52334711e-02
-1.09112346e+00 -8.05599868e-01 9.19257939e-01 2.72346996e-02
6.44925773e-01 1.53772563e-01 8.52874041e-01 1.26712263e+00
4.79295641e-01 7.74852633e-02 1.37870109e+00 -4.93019104e-01
1.59130648e-01 3.71118575e-01 -7.23888800e-02 4.87210751e-01
2.47221842e-01 3.22162032e-01 -4.35286820e-01 -3.39944400e-02
7.38321602e-01 2.54599035e-01 -6.69328928e-01 3.13686766e-02
-5.56412220e-01 2.01040119e-01 2.08574995e-01 1.70124516e-01
-7.57377297e-02 1.23339728e-01 -2.29931459e-01 3.48308504e-01
5.30274034e-01 2.28883326e-01 -4.37928736e-01 1.49717122e-01
-1.05752492e+00 -4.10205752e-01 4.84121948e-01 6.27913117e-01
6.42423570e-01 1.11840516e-01 3.63309413e-01 4.35786545e-01
4.66236919e-02 5.56075871e-01 1.23698540e-01 -1.19972956e+00
7.42771849e-02 1.70525879e-01 4.32575643e-01 -8.38992000e-01
-2.68878996e-01 3.14392075e-02 -2.44072825e-01 8.27055752e-01
4.73364443e-01 -4.93303984e-01 -9.56180036e-01 1.37318552e+00
2.94639349e-01 4.34864640e-01 -3.93639058e-02 1.06643760e+00
8.50678325e-01 3.75648707e-01 -6.39041603e-01 -3.35334599e-01
1.18769801e+00 -7.94082701e-01 -8.47203672e-01 -3.58565032e-01
-1.11377187e-01 -1.18437016e+00 1.15244079e+00 8.10208380e-01
-1.34567583e+00 -8.27974916e-01 -1.29229450e+00 -3.59500766e-01
-2.06033841e-01 4.91277725e-01 5.10318354e-02 1.21941483e+00
-1.35767162e+00 7.23598659e-01 -2.33960107e-01 -6.12811625e-01
2.86049426e-01 2.40520108e-02 -5.59742987e-01 -4.32729155e-01
-8.26971173e-01 7.77215958e-01 -5.01099288e-01 3.62290710e-01
-7.91518033e-01 -6.95501566e-01 -5.80331385e-01 2.88254321e-02
3.35456133e-01 -2.88667768e-01 1.22972667e+00 -1.69873261e+00
-1.87121713e+00 1.27454460e+00 -1.30886868e-01 -1.17846251e-01
4.48288858e-01 -5.63930273e-01 -9.11832452e-01 4.49400246e-01
-1.96424261e-01 8.66187289e-02 1.29576814e+00 -1.22920835e+00
2.00954769e-02 -5.26524603e-01 3.01342547e-01 6.88922033e-02
-2.01367229e-01 4.87565398e-01 -5.02317667e-01 -1.71921268e-01
-4.26588982e-01 -6.82427227e-01 3.79521966e-01 4.15976375e-01
-5.54435253e-01 7.70584643e-01 1.29705620e+00 -9.82504725e-01
6.36320651e-01 -2.14854169e+00 -6.76549613e-01 -2.35736519e-01
-2.31271535e-02 7.08432317e-01 -2.16238424e-02 4.90199447e-01
-4.05920178e-01 -2.48885497e-01 -7.70172924e-02 -7.49892354e-01
-5.81798971e-01 -6.87709525e-02 -3.55825126e-01 6.55238152e-01
-8.44724774e-02 4.89202946e-01 -7.40592718e-01 -7.80753940e-02
3.59153241e-01 7.05785096e-01 2.45084777e-01 6.00694716e-01
2.28878185e-01 3.36575508e-01 3.76772016e-01 7.74431527e-01
1.25213909e+00 1.52945155e-02 2.41612032e-01 -3.75801712e-01
-7.29811043e-02 -2.49730915e-01 -9.46786880e-01 1.35930037e+00
-6.46887422e-01 1.35956800e+00 3.29379618e-01 -1.32750750e-01
8.06833923e-01 7.86680505e-02 -1.75873354e-01 -6.91023707e-01
2.51650542e-01 -7.58723170e-02 -5.27511477e-01 -8.95913482e-01
5.12915611e-01 -2.71836609e-01 5.84341586e-01 6.56396270e-01
-1.94149449e-01 -1.43715814e-01 -2.83607781e-01 8.32696483e-02
1.17509675e+00 4.89806756e-02 -6.13536611e-02 2.03887820e-01
1.46458030e-01 -6.83946908e-01 5.45528717e-02 3.89398307e-01
-7.34677687e-02 1.28566515e+00 1.02992609e-01 -8.52870941e-01
-1.06561422e+00 -1.08261943e+00 -1.67562157e-01 5.99233985e-01
8.37583840e-02 -2.12973118e-01 -9.22484398e-01 -4.15237337e-01
-2.16755643e-01 4.77021694e-01 -9.18068767e-01 -7.44786188e-02
-5.01195014e-01 -1.90822527e-01 3.78645897e-01 3.01891923e-01
9.86155808e-01 -8.85275483e-01 -1.03395259e+00 -4.81131375e-01
5.79551421e-02 -1.21319616e+00 -6.82645023e-01 -2.25125343e-01
-3.37049305e-01 -1.39501321e+00 -9.01846886e-01 -4.17664379e-01
6.63168073e-01 8.47346544e-01 1.34120774e+00 1.33329555e-01
-6.10312104e-01 6.25474691e-01 -6.34977669e-02 -3.69377822e-01
-2.85844892e-01 -7.35306382e-01 2.85273716e-02 7.08172619e-01
1.70944184e-01 -5.95706582e-01 -9.85695601e-01 2.55647689e-01
-1.00957870e+00 -2.39680663e-01 8.32828879e-02 4.57382858e-01
-1.00712836e-01 1.89397931e-01 -3.37985277e-01 -5.62240899e-01
4.48130608e-01 -9.86948013e-02 -7.10889101e-01 1.41545981e-01
-2.40463510e-01 -8.66978943e-01 5.90877235e-01 -3.36645603e-01
-1.46354938e+00 -4.58500497e-02 3.87020648e-01 -9.45075870e-01
-5.55416644e-01 -4.21847373e-01 -3.73752654e-01 -3.73627365e-01
7.16774106e-01 -8.18750411e-02 -2.89540365e-02 -4.38663602e-01
1.22178778e-01 8.70749354e-01 9.38399673e-01 2.12529987e-01
7.46644139e-01 9.76053119e-01 -4.64063175e-02 -1.16882885e+00
-8.80200565e-01 -1.28805459e-01 -5.62627733e-01 -5.04301190e-01
7.40646720e-01 -7.51633525e-01 -7.04672992e-01 1.04690409e+00
-1.16968167e+00 -6.35665059e-01 -2.09935144e-01 7.43217617e-02
-1.05342917e-01 3.57394993e-01 -1.96486130e-01 -7.05685675e-01
-4.95639890e-02 -6.87483072e-01 1.29541409e+00 8.36215436e-01
2.28212416e-01 -9.08158720e-01 1.97967485e-01 6.62136614e-01
3.41227591e-01 1.08973183e-01 2.30838269e-01 2.69666672e-01
-4.79444176e-01 -3.38804454e-01 -1.55424610e-01 8.55870664e-01
5.54351866e-01 3.87888521e-01 -1.71582639e+00 -3.52414519e-01
3.08300167e-01 -3.41397971e-01 4.59788531e-01 4.58448619e-01
1.07274413e+00 -1.73235565e-01 -2.43017361e-01 9.56007540e-01
1.54006016e+00 3.95686209e-01 1.15492833e+00 1.05969444e-01
5.37578762e-01 4.49825972e-01 1.82166532e-01 -1.80395581e-02
-2.84238964e-01 9.20584679e-01 5.09047985e-01 -7.46210635e-01
-5.97006679e-01 -8.46706405e-02 4.65222359e-01 -4.50756311e-01
8.36280063e-02 -4.87165481e-01 -5.85901916e-01 2.90371895e-01
-9.54847157e-01 -8.84496450e-01 -2.48525858e-01 2.32208347e+00
4.49706376e-01 -3.99900883e-01 3.87469530e-02 -7.32828537e-03
9.54410434e-01 4.97204177e-02 -5.58793902e-01 -4.64156717e-01
-3.10276985e-01 1.76801518e-01 5.49677074e-01 5.80591500e-01
-7.74285436e-01 5.89827359e-01 7.40903330e+00 -1.38347715e-01
-1.48165333e+00 -8.01660269e-02 8.75713289e-01 -4.14373606e-01
6.29162416e-02 -2.70550996e-01 -2.75233746e-01 7.98447132e-01
9.37865615e-01 3.44827086e-01 4.98306334e-01 5.43564558e-01
2.53995270e-01 -7.13761926e-01 -9.72416282e-01 1.07677507e+00
7.20040917e-01 -1.10171199e+00 -6.16180301e-01 1.47182494e-01
7.39269316e-01 -2.23085836e-01 5.13918698e-01 -6.00023448e-01
7.14673251e-02 -1.11304355e+00 4.75121528e-01 7.04866946e-01
1.14514101e+00 -3.93139005e-01 4.40139711e-01 -3.01530182e-01
-5.35253584e-01 1.63262393e-02 -2.02191412e-01 -1.89871803e-01
2.31765732e-01 5.38502216e-01 -1.02892947e+00 1.98426232e-01
1.02467084e+00 5.33251464e-01 -6.93339288e-01 9.20808256e-01
-2.15129346e-01 5.80246508e-01 -2.01797456e-01 6.20683372e-01
-1.83670774e-01 -4.56494838e-01 5.94811261e-01 8.33103478e-01
2.76016295e-01 1.12193674e-01 -5.62747896e-01 7.20161796e-01
-1.14612900e-01 -5.61257958e-01 -9.72645223e-01 4.77172643e-01
9.47526395e-02 1.48381817e+00 -5.81785917e-01 -1.91809729e-01
-6.35743618e-01 1.47437739e+00 -3.13201517e-01 7.11875260e-01
-7.06388891e-01 -3.82118732e-01 4.97189224e-01 5.08555770e-01
2.73575991e-01 1.31414309e-01 5.47044054e-02 -8.84852350e-01
4.18084025e-01 -4.86860782e-01 -1.53445721e-01 -2.06109738e+00
-1.04446673e+00 8.19134653e-01 -3.71393003e-02 -1.24449801e+00
2.67088026e-01 -6.59163117e-01 -1.14971316e+00 8.79907429e-01
-1.31349266e+00 -1.01074052e+00 -8.72610033e-01 7.54505813e-01
4.71417457e-01 -1.12103157e-01 6.19214833e-01 2.98067838e-01
-5.42885363e-01 4.87591892e-01 2.94790596e-01 -1.00035891e-01
9.78436649e-01 -9.22598481e-01 4.52944905e-01 1.12400305e+00
-1.79574236e-01 4.04523432e-01 7.54054189e-01 -4.70346838e-01
-1.01079583e+00 -8.95892441e-01 5.55388689e-01 -4.31273729e-01
2.34689280e-01 -3.89953315e-01 -7.54071355e-01 8.95014584e-01
7.16049314e-01 3.00693691e-01 4.97596622e-01 -2.86039978e-01
-1.94772631e-01 -5.02256930e-01 -1.25011754e+00 4.92534995e-01
7.34689713e-01 -4.96667296e-01 -4.77231264e-01 5.19285798e-01
4.94929969e-01 -5.98158240e-01 -2.84308970e-01 -3.52283686e-01
9.79790807e-01 -1.48660052e+00 8.62325788e-01 -3.51742327e-01
5.78020096e-01 -1.42393798e-01 -2.85746139e-02 -1.36894858e+00
3.06392699e-01 -1.21845651e+00 3.80397737e-01 1.13785267e+00
-2.41055503e-01 -7.45897710e-01 4.47066158e-01 6.16227150e-01
-5.24257831e-02 -4.66653317e-01 -7.66622066e-01 -6.85865462e-01
-4.71638680e-01 3.79439443e-02 5.91941774e-01 5.43536484e-01
-5.52145183e-01 1.96187183e-01 -7.35511661e-01 2.41528779e-01
5.86325884e-01 -4.13603801e-03 1.12805855e+00 -1.14066720e+00
-7.65454546e-02 4.34632152e-01 -1.87501222e-01 -5.06361306e-01
-1.33600384e-01 -6.12865528e-03 -2.17311189e-01 -1.08425295e+00
7.32521936e-02 1.79969266e-01 2.25116640e-01 1.26154423e-01
6.35935888e-02 6.84174538e-01 3.81569594e-01 1.11089349e-01
-2.87472606e-01 8.53449777e-02 9.15779710e-01 2.36107409e-01
-9.24556404e-02 4.21381406e-02 -6.05006874e-01 1.17687190e+00
5.95082462e-01 -1.86528936e-01 -4.82786417e-01 -5.37948132e-01
3.18790853e-01 1.84576243e-01 8.16850781e-01 -1.20878017e+00
-6.16821349e-02 1.30490333e-01 9.63642538e-01 -1.90148965e-01
8.59341085e-01 -1.05333698e+00 5.44107497e-01 5.11188135e-02
-7.32845813e-02 -8.96602031e-03 2.95409441e-01 5.81305981e-01
1.91570923e-01 -9.37066004e-02 9.22793210e-01 -3.89066249e-01
-1.45005569e-01 -1.20715261e-01 -1.33180141e-01 -3.79370481e-01
9.92914736e-01 -7.42752552e-01 -6.77007139e-01 -7.64979780e-01
-4.71470624e-01 -5.61873972e-01 9.82215822e-01 1.14055105e-01
7.81472921e-01 -1.02854121e+00 -2.41031066e-01 6.62671030e-01
-3.76584738e-01 -3.83630186e-01 4.43504363e-01 6.69109404e-01
-6.63308680e-01 1.23449564e-01 -4.71361428e-01 -3.81126285e-01
-1.85207379e+00 3.69314551e-01 8.35558593e-01 2.50250489e-01
-8.54836345e-01 6.73728824e-01 4.51325685e-01 2.64690042e-01
1.83383390e-01 -2.29371995e-01 2.29161888e-01 -3.84646118e-01
9.24684584e-01 5.35198867e-01 5.00090361e-01 -5.36191642e-01
-1.56552628e-01 7.49209106e-01 1.82247266e-01 -9.28260833e-02
1.12209046e+00 -5.62979758e-01 2.13201344e-01 4.63404089e-01
1.35472929e+00 4.24385339e-01 -1.59107959e+00 -8.53779390e-02
-1.29299688e+00 -1.13282609e+00 2.09115043e-01 -1.09485662e+00
-1.39109075e+00 8.52894366e-01 8.85929883e-01 2.22354084e-01
1.47349477e+00 -8.09078366e-02 8.55224371e-01 2.03378916e-01
2.02169538e-01 -9.99049485e-01 4.74632621e-01 -1.89645767e-01
9.30310011e-01 -1.22677827e+00 4.00204733e-02 -1.71382725e-01
-5.58130860e-01 1.47333002e+00 4.19700146e-01 -2.43497789e-02
4.31120157e-01 2.71122366e-01 3.91063392e-01 -3.90621841e-01
-3.45786393e-01 2.21252870e-02 6.65421709e-02 8.83041739e-01
-9.84200016e-02 -3.41564536e-01 6.30388260e-01 -3.20550986e-02
-1.95101231e-01 3.14521611e-01 1.12337375e+00 8.16785932e-01
-2.75016930e-02 -4.42246050e-01 -8.71648729e-01 1.37534365e-01
-4.79457438e-01 -2.79370155e-02 -7.68185973e-01 9.07362580e-01
4.57532316e-01 1.07371032e+00 4.26774114e-01 -1.18131176e-01
2.17055157e-01 4.40381560e-03 7.15051889e-01 -3.13861132e-01
-3.94625664e-01 -1.35052338e-01 1.22939125e-01 -6.85118914e-01
-4.94271219e-01 -7.15464413e-01 -2.67372608e-01 -9.68781710e-02
-1.22200802e-01 -7.13368893e-01 5.86124361e-01 7.32905030e-01
1.69810787e-01 3.04092944e-01 9.21902776e-01 -8.67381394e-01
1.31803125e-01 -8.59837592e-01 -9.39969897e-01 4.01530534e-01
9.44397688e-01 -5.52876532e-01 -8.61282945e-01 5.48917830e-01]
|
[10.328438758850098, -2.6937763690948486]
|
b52bfe2e-48ec-4257-bc42-37df44448b08
|
uncertainty-guided-depth-fusion-for-spike
|
2208.12653
| null |
https://arxiv.org/abs/2208.12653v2
|
https://arxiv.org/pdf/2208.12653v2.pdf
|
Uncertainty Guided Depth Fusion for Spike Camera
|
Depth estimation is essential for various important real-world applications such as autonomous driving. However, it suffers from severe performance degradation in high-velocity scenario since traditional cameras can only capture blurred images. To deal with this problem, the spike camera is designed to capture the pixel-wise luminance intensity at high frame rate. However, depth estimation with spike camera remains very challenging using traditional monocular or stereo depth estimation algorithms, which are based on the photometric consistency. In this paper, we propose a novel Uncertainty-Guided Depth Fusion (UGDF) framework to fuse the predictions of monocular and stereo depth estimation networks for spike camera. Our framework is motivated by the fact that stereo spike depth estimation achieves better results at close range while monocular spike depth estimation obtains better results at long range. Therefore, we introduce a dual-task depth estimation architecture with a joint training strategy and estimate the distributed uncertainty to fuse the monocular and stereo results. In order to demonstrate the advantage of spike depth estimation over traditional camera depth estimation, we contribute a spike-depth dataset named CitySpike20K, which contains 20K paired samples, for spike depth estimation. UGDF achieves state-of-the-art results on CitySpike20K, surpassing all monocular or stereo spike depth estimation baselines. We conduct extensive experiments to evaluate the effectiveness and generalization of our method on CitySpike20K. To the best of our knowledge, our framework is the first dual-task fusion framework for spike camera depth estimation. Code and dataset will be released.
|
['Shanghang Zhang', 'Tiejun Huang', 'Li Du', 'Lei Ma', 'Ming Lu', 'Jiyuan Zhang', 'Xiaobao Wei', 'Jiaming Liu', 'Jianing Li']
|
2022-08-26
| null | null | null | null |
['stereo-depth-estimation']
|
['computer-vision']
|
[-7.86092356e-02 -4.86496657e-01 2.08119705e-01 -5.30871153e-01
-1.05285072e+00 -3.94882202e-01 4.72036839e-01 -5.51278949e-01
-6.06092811e-01 9.92192984e-01 1.43773183e-01 1.97379902e-01
2.01871768e-01 -4.93574440e-01 -8.76908302e-01 -1.02903521e+00
4.49428648e-01 2.20060453e-01 3.83293539e-01 2.96143115e-01
4.70936209e-01 9.13057774e-02 -1.51858270e+00 4.21953827e-01
7.94313371e-01 1.35574067e+00 6.49178326e-01 6.10763788e-01
1.07281364e-03 7.08934188e-01 -2.01592162e-01 -2.15975389e-01
2.51528949e-01 -2.44430885e-01 -2.38731265e-01 -1.61827236e-01
5.28776288e-01 -1.00101280e+00 -8.11256409e-01 1.15156341e+00
6.18892252e-01 -4.03036445e-01 4.82920617e-01 -1.51594663e+00
-3.52286220e-01 3.67062539e-01 -6.17977440e-01 4.42295700e-01
2.33689085e-01 4.84577239e-01 5.81740856e-01 -9.66860056e-01
3.99502724e-01 8.79094362e-01 4.17939126e-01 6.72040582e-01
-1.03571713e+00 -9.43329513e-01 2.76430696e-01 2.64903396e-01
-1.26781380e+00 -5.76583445e-01 5.54028511e-01 -5.29472530e-01
1.00020957e+00 -4.11758959e-01 6.59494519e-01 1.35633385e+00
3.86619568e-01 1.16506040e+00 1.40970933e+00 4.50377673e-01
4.96025771e-01 -3.79653215e-01 6.19731247e-02 1.36106864e-01
2.58791745e-01 4.04021204e-01 -9.09903109e-01 2.08242118e-01
1.06480658e+00 -8.53620917e-02 -6.24774754e-01 -2.68456668e-01
-1.40583444e+00 3.35707664e-01 3.39185804e-01 -4.33448404e-02
-7.68282786e-02 5.93679488e-01 2.29404032e-01 6.88766241e-02
3.77494752e-01 -3.64362970e-02 -3.67591828e-01 -7.20745862e-01
-9.76441205e-01 4.30708408e-01 3.10723007e-01 1.06091440e+00
8.12853396e-01 4.21565138e-02 -1.44478992e-01 6.67016625e-01
3.53655756e-01 7.39310324e-01 4.97433305e-01 -1.40916371e+00
6.52387679e-01 2.55617827e-01 9.97667983e-02 -3.38743865e-01
-2.99356163e-01 -3.17862034e-01 -8.38670015e-01 4.30814147e-01
6.40756190e-01 -1.01971306e-01 -7.35653222e-01 1.81924164e+00
-2.06590116e-01 8.54939103e-01 2.04160929e-01 1.22267067e+00
7.15190291e-01 4.08448249e-01 -4.25177217e-01 -8.77031386e-02
1.08022785e+00 -6.19235158e-01 -5.48579097e-01 -7.28210032e-01
4.13595438e-01 -3.63113612e-01 6.64511383e-01 6.07655883e-01
-1.16068089e+00 -4.09686357e-01 -1.09180534e+00 -2.13819966e-01
2.93270886e-01 8.88148546e-02 5.47400832e-01 3.74501169e-01
-1.12716007e+00 3.40051919e-01 -1.07241356e+00 5.51140010e-02
6.92012608e-01 2.03404590e-01 -4.04680878e-01 -2.81553149e-01
-9.39601243e-01 5.29211581e-01 1.36624411e-01 3.11687291e-02
-1.15255213e+00 -7.45682836e-01 -9.01648223e-01 -2.31406376e-01
-6.72947019e-02 -9.07972872e-01 1.38646340e+00 -4.39416438e-01
-1.37351906e+00 9.78526473e-01 -3.89585644e-01 -7.41539776e-01
6.15141273e-01 -3.28368276e-01 1.66908845e-01 1.84469372e-01
2.26910278e-01 1.06176293e+00 5.82017601e-01 -1.07489300e+00
-8.86874497e-01 -6.87449574e-01 3.95653173e-02 2.47118115e-01
1.58483356e-01 -4.04461443e-01 -6.99353516e-01 -4.41800445e-01
5.16429484e-01 -7.11770773e-01 1.00292973e-01 1.22146264e-01
-1.54184802e-02 1.96636245e-01 5.07064700e-01 -2.94873089e-01
7.69064903e-01 -2.21934938e+00 1.93171576e-01 -6.16908371e-01
4.45436031e-01 -1.12711355e-01 1.15505733e-01 -5.27206026e-02
2.43213311e-01 -4.80040878e-01 -4.82556134e-01 -7.73712456e-01
-6.42967522e-02 1.97675243e-01 -5.15628695e-01 5.42005777e-01
9.06087831e-02 9.77941871e-01 -8.34407628e-01 -2.72403389e-01
4.11565840e-01 4.51993883e-01 -5.66581905e-01 1.02176785e-01
-1.50606986e-02 7.41919696e-01 -2.41416201e-01 7.06678808e-01
1.12495780e+00 -9.28340072e-04 -4.36442047e-01 -2.13414997e-01
-2.52617896e-01 8.22418481e-02 -9.39646125e-01 2.15531635e+00
-3.30998331e-01 9.23773527e-01 -9.12455563e-03 -5.70711255e-01
8.68600965e-01 6.16251566e-02 4.04670119e-01 -9.67425883e-01
3.33404392e-01 4.20387268e-01 -2.47554585e-01 -2.08138436e-01
3.46429825e-01 -9.66337770e-02 -6.54346272e-02 2.36239702e-01
-1.00031912e-01 -4.81635839e-01 -1.02024175e-01 4.10570353e-01
1.27024055e+00 3.92188430e-01 -1.23111971e-01 7.47974440e-02
2.05374271e-01 -4.57638800e-01 7.86427200e-01 5.45860469e-01
-6.98848128e-01 1.24686801e+00 6.16177380e-01 -5.47055937e-02
-8.81555676e-01 -1.39762354e+00 -4.15398628e-01 5.83135672e-02
6.38749242e-01 -1.09919354e-01 -8.41749132e-01 -2.55074531e-01
7.69787356e-02 4.44206297e-01 -3.83713752e-01 -1.80971518e-01
-2.12487936e-01 -6.81412756e-01 7.13222802e-01 8.99481356e-01
1.07107449e+00 -7.07400680e-01 -6.93226874e-01 2.43041679e-01
-3.91866744e-01 -1.70070469e+00 -3.56511623e-01 1.70977756e-01
-9.08317089e-01 -8.70051086e-01 -9.19087112e-01 -4.98980999e-01
1.73445210e-01 5.86739242e-01 9.22901630e-01 -4.90702122e-01
-1.78356040e-02 2.87109315e-01 -6.74093515e-02 -1.76999271e-01
2.93514639e-01 -2.75848716e-01 5.55065461e-02 -5.30439280e-02
7.49358594e-01 -7.44154155e-01 -9.48332787e-01 3.57955217e-01
-7.65540659e-01 3.34146827e-01 2.99032182e-01 6.75394595e-01
6.83421314e-01 -8.73007998e-02 5.18679082e-01 -3.14363718e-01
2.26245925e-01 -3.58908236e-01 -9.49142635e-01 -4.35750484e-01
-2.71014512e-01 8.24188888e-02 3.32503378e-01 4.53325957e-02
-1.04329777e+00 1.76222563e-01 -1.53343976e-01 -5.62523961e-01
-6.69842288e-02 9.79304910e-02 -1.72461256e-01 -9.90819782e-02
4.41953331e-01 3.91531408e-01 -2.56006926e-01 -2.26946801e-01
-1.19052105e-01 7.43859053e-01 9.52636302e-01 -3.63729089e-01
5.04328430e-01 1.01436460e+00 -1.29182011e-01 -4.07693386e-01
-6.78805232e-01 -5.96726179e-01 -3.70884627e-01 -3.76234531e-01
9.48039591e-01 -1.68219185e+00 -8.83954287e-01 1.41966021e+00
-1.19367445e+00 -5.23624122e-01 2.45401576e-01 8.68037760e-01
-9.29447472e-01 5.33182621e-01 -7.70021021e-01 -8.43077838e-01
2.28419285e-02 -1.46837127e+00 1.60053217e+00 1.45365700e-01
2.94439226e-01 -6.00268126e-01 1.21097885e-01 4.79382694e-01
8.53445381e-02 2.25481782e-02 1.68481007e-01 3.28477263e-01
-1.12058890e+00 8.36767405e-02 -7.71307349e-01 3.27834457e-01
-1.22341424e-01 -3.61663908e-01 -1.52390361e+00 -2.14232817e-01
2.58403331e-01 -5.28678417e-01 1.30430937e+00 8.06695521e-01
1.24571800e+00 7.31044769e-01 -1.84383541e-01 1.04928041e+00
1.58920991e+00 7.49431178e-02 1.11654639e+00 4.65888828e-01
8.01995695e-01 3.64655524e-01 6.66869342e-01 7.41452157e-01
7.95897484e-01 7.03437567e-01 7.02980995e-01 4.44549650e-01
-1.64603814e-01 -1.64528847e-01 4.80160594e-01 4.49920267e-01
2.45416403e-01 -1.35394454e-01 -8.25277209e-01 6.09769940e-01
-1.88129735e+00 -8.96315336e-01 -2.90909946e-01 2.31145930e+00
5.97049296e-01 4.74523604e-01 -1.39331251e-01 3.31715941e-02
4.93829966e-01 6.28875196e-02 -9.93511498e-01 1.84993044e-01
-5.29046357e-01 -2.01212034e-01 6.81565046e-01 4.38660443e-01
-8.71211112e-01 8.50533426e-01 5.56488752e+00 7.79753506e-01
-9.65650439e-01 2.81996857e-02 5.67222357e-01 -5.80567122e-01
-3.93132955e-01 1.90728214e-02 -1.17371666e+00 9.41136360e-01
6.50195420e-01 3.47975045e-02 6.14930689e-01 5.50498545e-01
3.53172779e-01 -6.86291397e-01 -1.15887427e+00 1.85349846e+00
4.61293049e-02 -1.21223044e+00 -4.61441129e-01 1.70992360e-01
8.51902664e-01 5.93761325e-01 1.21623330e-01 1.78959653e-01
2.26007566e-01 -7.91267812e-01 8.36401820e-01 6.14131808e-01
8.15099120e-01 -5.52493811e-01 7.31467783e-01 5.01286745e-01
-1.21682990e+00 -1.10879630e-01 -5.73313653e-01 -5.22700012e-01
5.67296624e-01 9.89121258e-01 -1.04004100e-01 2.86668152e-01
9.03397202e-01 1.26449955e+00 -3.07862252e-01 1.35202479e+00
-2.39210516e-01 1.06720224e-01 -2.63733298e-01 3.84381682e-01
3.05003643e-01 -1.84796914e-01 3.13079447e-01 7.06763029e-01
6.11790776e-01 1.05334967e-01 -3.14504594e-01 1.05522346e+00
-3.90715152e-02 -5.90372741e-01 -7.35317230e-01 2.00926855e-01
5.13625920e-01 9.77191150e-01 -4.64495093e-01 -2.96757340e-01
-6.40603185e-01 1.14832854e+00 3.60286653e-01 3.59916210e-01
-1.06424546e+00 -2.92858705e-02 1.09969425e+00 -1.18216183e-02
1.70113981e-01 -4.34516609e-01 -7.19720304e-01 -1.70458198e+00
2.50850648e-01 -3.89021933e-01 -1.28329813e-01 -1.39618266e+00
-1.21443093e+00 5.29345751e-01 -2.74030119e-02 -1.38019347e+00
-8.35465789e-02 -8.04778099e-01 -3.96616846e-01 8.25371563e-01
-1.87810969e+00 -7.41600215e-01 -9.19161499e-01 5.14691174e-01
6.03567243e-01 3.46765146e-02 6.49463087e-02 3.38699758e-01
-5.01867652e-01 2.76388913e-01 3.11628282e-01 -2.02268034e-01
9.52470243e-01 -1.26170003e+00 7.25884140e-01 8.61419857e-01
-2.10666165e-01 7.63386562e-02 4.68831390e-01 -5.49744129e-01
-1.22714674e+00 -9.22405422e-01 5.22567868e-01 -6.64053380e-01
5.82233429e-01 -4.47264165e-01 -8.13230038e-01 5.86805403e-01
-2.74267234e-02 9.11535695e-02 8.59420151e-02 -5.28622627e-01
-2.18688920e-01 -2.65257210e-01 -8.55352044e-01 5.34369886e-01
1.27972639e+00 -8.32839787e-01 -3.78406703e-01 1.98966227e-02
5.01319408e-01 -5.76052666e-01 -4.74743277e-01 4.63528216e-01
6.26193702e-01 -1.56358814e+00 8.69707525e-01 3.79445553e-01
5.70413709e-01 -5.02282560e-01 -4.85570461e-01 -1.24167573e+00
-7.58184567e-02 -1.27094314e-01 3.39230187e-02 1.07975984e+00
-1.29649322e-02 -8.33953857e-01 1.03660893e+00 5.15941501e-01
-5.70526123e-01 -5.93162417e-01 -1.19834733e+00 -8.63739431e-01
8.76886025e-02 -8.94242108e-01 3.70121628e-01 3.10789108e-01
-1.15046993e-01 5.52467182e-02 -3.52763116e-01 3.04883957e-01
1.01913261e+00 -4.20894623e-02 8.08693826e-01 -9.50914741e-01
-3.72381538e-01 -3.42856526e-01 -8.14087570e-01 -1.69875157e+00
2.85806268e-01 -5.26065648e-01 3.46182078e-01 -1.59540498e+00
4.50510561e-01 -1.28428981e-01 -8.98977146e-02 -7.21758083e-02
-9.14445668e-02 5.22594213e-01 -3.11289635e-02 2.82111496e-01
-5.86226344e-01 7.24221051e-01 1.12631631e+00 3.42000425e-02
1.59201309e-01 -2.84688652e-01 -4.98903185e-01 7.49477267e-01
6.47161722e-01 -1.86684474e-01 -7.03366399e-01 -9.92199421e-01
2.63285995e-01 2.99612284e-01 6.95097029e-01 -1.57509291e+00
4.93859857e-01 6.50136545e-02 3.52892697e-01 -8.44669461e-01
7.68495679e-01 -5.80261767e-01 1.36406617e-02 2.28537828e-01
6.96264133e-02 -1.62493616e-01 1.93210140e-01 7.01761127e-01
-4.48638827e-01 -3.06697227e-02 9.14344609e-01 -8.19196925e-02
-1.10715997e+00 6.24160290e-01 -3.57875228e-01 2.70619452e-01
9.97526109e-01 -5.95549941e-01 -6.44759655e-01 -4.83086377e-01
-2.46997803e-01 3.57107908e-01 8.35729182e-01 2.37584084e-01
1.00262046e+00 -1.40479302e+00 -6.32756770e-01 3.23083609e-01
3.86234909e-01 2.37401962e-01 5.91389656e-01 1.04063809e+00
-5.84583521e-01 4.31449115e-01 -3.49794179e-01 -1.02261543e+00
-8.01047266e-01 3.48637372e-01 3.94786239e-01 2.61216193e-01
-6.46982372e-01 1.14230192e+00 1.02445769e+00 6.80842996e-02
1.79423094e-01 -7.37733722e-01 1.40408829e-01 -2.00983435e-01
7.43456364e-01 2.02181667e-01 6.51783943e-02 -4.70138788e-01
-4.25479054e-01 8.76433730e-01 5.07725626e-02 -3.73570025e-01
1.00942278e+00 -6.50659680e-01 2.54205376e-01 5.99796116e-01
1.11987913e+00 -5.07572770e-01 -2.29436445e+00 -1.36517420e-01
-5.30069888e-01 -8.56062889e-01 4.74888384e-01 -4.85931695e-01
-1.18692493e+00 1.23011255e+00 6.69176936e-01 -6.14239037e-01
1.24539196e+00 -6.30301237e-02 1.17135501e+00 1.80586860e-01
8.38942885e-01 -9.23021972e-01 2.21358523e-01 5.95108867e-01
5.74029624e-01 -1.58233833e+00 -4.23126519e-01 -2.01068640e-01
-8.29960525e-01 8.05192828e-01 9.28455830e-01 -2.23720670e-01
3.46781075e-01 5.73896527e-01 1.56789683e-02 -5.38764261e-02
-9.73128080e-01 -2.31856123e-01 -2.87893206e-01 5.60852051e-01
5.89049272e-02 -1.77120849e-01 2.34608650e-01 5.76463819e-01
-9.27949604e-03 3.67331177e-01 6.17501199e-01 7.38601267e-01
-6.02403522e-01 -7.01068342e-01 -2.62699693e-01 3.44468385e-01
-1.61865115e-01 -3.19785982e-01 -9.39135700e-02 3.14871043e-01
6.40919507e-02 8.03732455e-01 2.58857220e-01 -3.69527757e-01
1.29214451e-01 -2.57930696e-01 7.93029308e-01 -3.19819778e-01
6.33787140e-02 1.15232430e-01 -2.43199274e-01 -7.63545036e-01
-2.75834501e-01 -8.53787363e-01 -1.34096611e+00 -4.70673978e-01
-3.95130180e-02 -4.03880626e-01 6.16840839e-01 1.01042390e+00
2.16338396e-01 4.61026818e-01 3.87152195e-01 -9.70494986e-01
-7.71724507e-02 -6.93888187e-01 -9.02779043e-01 2.11270064e-01
4.27424490e-01 -9.28444862e-01 -6.48906231e-01 -1.50957182e-01]
|
[8.892863273620605, -2.3481199741363525]
|
dda57cc9-a326-4531-9482-c46b5614dc18
|
rule-based-vs-neural-net-approaches-to
| null | null |
https://aclanthology.org/W18-3803
|
https://aclanthology.org/W18-3803.pdf
|
Rule-based vs. Neural Net Approaches to Semantic Textual Similarity
|
This paper presents a neural net approach to determine Semantic Textual Similarity (STS) using attention-based bidirectional Long Short-Term Memory Networks (Bi-LSTM). To this date, most of the traditional STS systems were rule-based that built on top of excessive use of linguistic features and resources. In this paper, we present an end-to-end attention-based Bi-LSTM neural network system that solely takes word-level features, without expensive feature engineering work or the usage of external resources. By comparing its performance with traditional rule-based systems against SemEval-2012 benchmark, we make an assessment on the limitations and strengths of neural net systems to rule-based systems on Semantic Textual Similarity.
|
['Linrui Zhang', 'Dan Moldovan']
|
2018-08-01
| null | null | null |
coling-2018-8
|
['sentence-pair-modeling']
|
['natural-language-processing']
|
[ 1.30540967e-01 -7.73230754e-03 -9.95792672e-02 -4.84377533e-01
-3.26826990e-01 -1.57630026e-01 6.81244075e-01 4.12095517e-01
-6.95168078e-01 4.49497104e-01 3.26306939e-01 -7.24236906e-01
-3.39650571e-01 -9.55066144e-01 -4.88414109e-01 2.26815224e-01
2.42813900e-01 7.06764936e-01 1.95027992e-01 -9.52831089e-01
6.08234823e-01 1.48625150e-01 -1.50790107e+00 7.93878555e-01
6.92386150e-01 1.52632093e+00 -5.13014644e-02 5.73346198e-01
-7.07651973e-01 1.38468707e+00 -4.51410323e-01 -5.02822161e-01
-1.11777782e-01 -3.28478366e-01 -1.38845766e+00 -8.71045530e-01
4.85762596e-01 4.32739668e-02 -4.13101822e-01 7.08536386e-01
5.51902115e-01 5.56471944e-01 4.64673072e-01 -1.13908231e+00
-1.38725150e+00 9.91093099e-01 1.78356990e-01 3.70325983e-01
5.70952415e-01 -2.37306640e-01 1.10480618e+00 -1.21976471e+00
5.15668035e-01 1.29706347e+00 1.35581028e+00 2.97912598e-01
-6.91787124e-01 -4.13348287e-01 1.46831805e-02 9.69197154e-01
-1.19101059e+00 -4.75484490e-01 6.46529555e-01 -3.29658777e-01
2.26993775e+00 2.03037888e-01 4.89081025e-01 1.21623564e+00
3.16492766e-01 5.49748838e-01 7.18362510e-01 -8.30611885e-01
3.81768569e-02 -5.56431748e-02 4.75694269e-01 6.77136242e-01
-4.94104847e-02 -2.73274779e-02 -5.66465616e-01 -1.70160934e-01
1.35104969e-01 -6.79487586e-02 2.45033890e-01 1.88092366e-01
-8.78184378e-01 9.33876216e-01 6.73246384e-01 1.09386528e+00
-4.23546761e-01 1.07580284e-02 1.02170408e+00 5.81489503e-01
5.50050080e-01 6.10827446e-01 -6.13539577e-01 -5.22130653e-02
-9.90244389e-01 -1.98542967e-01 8.22975397e-01 6.35228992e-01
4.77473915e-01 4.34618741e-01 -5.47577739e-01 1.08958960e+00
8.75535384e-02 1.62977189e-01 1.24352837e+00 -5.41501880e-01
2.71189302e-01 7.23624945e-01 -4.64232236e-01 -1.27029705e+00
-2.26532325e-01 -3.32484275e-01 -6.44216835e-01 -2.56064802e-01
-1.52833894e-01 -5.54699190e-02 -8.35734427e-01 1.30135691e+00
-3.77622157e-01 -1.37522548e-01 2.02546582e-01 5.30691922e-01
1.27899170e+00 4.99991208e-01 1.12930663e-01 1.83375239e-01
1.04395139e+00 -1.10963237e+00 -7.96725035e-01 -3.14342082e-01
8.95159066e-01 -7.73388803e-01 1.08232629e+00 3.38640809e-02
-1.14498246e+00 -7.28413999e-01 -1.06756735e+00 -3.52524608e-01
-1.30145156e+00 -2.97953457e-01 4.32027072e-01 5.85204959e-01
-1.41121554e+00 1.07370853e+00 -2.16923714e-01 -1.06037283e+00
2.38536417e-01 3.87321264e-01 -9.40579697e-02 3.53259981e-01
-1.69688404e+00 1.73193944e+00 6.56095386e-01 2.26372182e-01
-2.35099316e-01 -5.79174101e-01 -8.55529487e-01 2.66782612e-01
2.50162631e-01 -7.64361024e-01 1.47962880e+00 -1.24134517e+00
-1.65245545e+00 5.12585223e-01 -1.07085437e-01 -9.06559169e-01
4.56088223e-02 -1.36695474e-01 -7.88354516e-01 -1.46435156e-01
9.66550782e-02 6.06571794e-01 2.26727769e-01 -7.05511034e-01
-3.69045168e-01 -7.70539939e-02 -1.13125756e-01 1.29797952e-02
-9.39070940e-01 4.15350974e-01 1.07298970e-01 -6.72600925e-01
-4.35632259e-01 -5.94701827e-01 8.08184817e-02 -3.09808284e-01
-1.06442362e-01 -6.27788246e-01 9.33576584e-01 -6.94031060e-01
1.31196105e+00 -1.48237300e+00 -4.16317374e-01 1.64092466e-01
-1.55347288e-01 6.64691925e-01 -4.10164922e-01 6.93301916e-01
-2.07091287e-01 3.74566585e-01 -2.91375462e-02 6.30998313e-02
1.11871064e-01 1.06743194e-01 -4.02474701e-01 -2.34428570e-01
9.15821567e-02 1.38837624e+00 -9.76659954e-01 -6.78080440e-01
3.88051897e-01 4.35996234e-01 -1.67323753e-01 -2.54649529e-03
-5.28573133e-02 -4.68090087e-01 -1.37664586e-01 4.51426446e-01
-5.31264096e-02 2.16098744e-02 2.55486310e-01 -3.43109697e-01
-8.95977840e-02 6.37070656e-01 -3.58278781e-01 1.80052912e+00
-1.02826893e+00 9.29673374e-01 -6.86610699e-01 -1.10772896e+00
1.10781217e+00 5.76036096e-01 3.78583342e-01 -1.29225242e+00
4.83245581e-01 5.39614320e-01 4.20625284e-02 -5.51932931e-01
7.46567070e-01 -1.43801153e-01 2.64712088e-02 5.54112256e-01
2.59529620e-01 5.03182895e-02 1.58574313e-01 6.08285982e-03
9.12789583e-01 1.34511054e-01 3.78538787e-01 -2.68872947e-01
7.06952333e-01 7.29094967e-02 8.41972232e-02 7.26800859e-01
5.11642992e-02 3.09564829e-01 -1.39238477e-01 -7.68590868e-01
-1.03675961e+00 -5.33656776e-01 1.46702856e-01 1.47013342e+00
-5.54263294e-01 -4.56733018e-01 -5.78971744e-01 -7.07655132e-01
-8.31847824e-03 1.15019047e+00 -8.00363064e-01 -4.81168538e-01
-3.63749057e-01 -1.74354166e-01 1.07201183e+00 8.82408381e-01
4.78227913e-01 -1.57382703e+00 -6.04165137e-01 4.95566219e-01
-1.97697788e-01 -9.74658430e-01 -2.96904117e-01 2.97804892e-01
-7.47962117e-01 -6.42190397e-01 -3.02391469e-01 -7.01798856e-01
7.94626586e-03 -2.08851025e-02 1.32286358e+00 -1.76953897e-02
-8.91645476e-02 1.01600394e-01 -4.27011013e-01 -5.22257864e-01
-3.08913022e-01 3.25538635e-01 3.38310450e-02 -2.48504266e-01
8.76555204e-01 -4.68232155e-01 -5.59494868e-02 -2.47107241e-02
-5.74714184e-01 -3.34208876e-01 4.87068117e-01 8.21797967e-01
6.16745800e-02 -2.28477404e-01 8.18149149e-01 -5.59872746e-01
1.25425053e+00 -6.56042874e-01 -1.89627036e-02 6.22213721e-01
-1.00111163e+00 1.81630284e-01 8.67710292e-01 -2.10281909e-01
-1.06369138e+00 -3.65779817e-01 -2.32018918e-01 -5.44892311e-01
1.13527626e-01 1.01034963e+00 4.38396335e-01 -3.20078790e-01
8.20372880e-01 2.18562841e-01 -1.70379773e-01 -4.99526143e-01
3.65994006e-01 1.15628290e+00 4.52217668e-01 -4.87572402e-01
-1.12968408e-01 -3.62644307e-02 -4.53442067e-01 -5.37318289e-01
-1.05812860e+00 -2.58081466e-01 -7.43296981e-01 -1.47052556e-01
6.07749224e-01 -3.24032426e-01 -6.76731825e-01 2.60250539e-01
-1.25087023e+00 -3.87691528e-01 -4.20078605e-01 2.23380759e-01
-4.01168704e-01 1.61924228e-01 -7.96322823e-01 -4.79427904e-01
-1.08280706e+00 -5.69291532e-01 4.63104248e-01 -1.44499853e-01
-8.22683692e-01 -1.54367018e+00 8.23190659e-02 3.16124111e-01
1.19784677e+00 -9.83342826e-02 9.46440578e-01 -1.28651500e+00
5.50671995e-01 -3.63443226e-01 -2.88291454e-01 4.15238261e-01
4.79066782e-02 -1.70865636e-02 -1.10159254e+00 9.55154151e-02
-3.69265936e-02 -5.62564492e-01 8.93596649e-01 3.75557274e-01
1.23466623e+00 -3.84567410e-01 -1.13521233e-01 1.85180724e-01
1.34525836e+00 3.60435456e-01 4.03539032e-01 7.88677752e-01
7.69929707e-01 5.26691794e-01 3.03193569e-01 1.94873974e-01
4.53351140e-01 6.12664998e-01 5.68731837e-02 4.06735204e-02
-3.07569712e-01 -1.59334093e-01 3.02549541e-01 1.21237767e+00
3.34646031e-02 -3.36356968e-01 -1.28270817e+00 7.68565953e-01
-1.96015072e+00 -9.65391397e-01 -1.20539419e-01 1.77776134e+00
7.30085075e-01 2.45661557e-01 -1.63851246e-01 2.97085524e-01
7.87217021e-01 -2.04033162e-02 -3.24328750e-01 -1.41856790e+00
-1.59083024e-01 5.14052749e-01 3.44846815e-01 3.14321518e-01
-7.31389880e-01 1.07536709e+00 6.80039740e+00 1.07979131e+00
-1.33357692e+00 2.76564598e-01 3.28974485e-01 -1.44784316e-01
-2.66985655e-01 -1.71340302e-01 -4.56416219e-01 2.73539603e-01
1.59308934e+00 -4.04967457e-01 2.88319886e-01 6.78522348e-01
-1.54758841e-01 2.72964418e-01 -1.15046871e+00 6.18192613e-01
3.99904639e-01 -1.46325219e+00 4.37497705e-01 -5.22298753e-01
4.77645248e-01 4.60007459e-01 3.91431265e-02 5.43578565e-01
4.88400728e-01 -1.21879375e+00 7.50412107e-01 5.75061977e-01
7.23622501e-01 -7.31467426e-01 1.28709149e+00 -1.64693743e-01
-1.19520497e+00 -1.54611960e-01 -3.39454502e-01 -4.06721026e-01
-1.17831185e-01 3.35818231e-01 -8.55798841e-01 7.89656937e-01
7.81545937e-01 8.91611874e-01 -6.88835502e-01 6.14311695e-01
-1.04731955e-02 4.84656394e-01 -1.75991967e-01 -2.94821501e-01
7.04216659e-01 4.52735841e-01 9.22860131e-02 1.79324031e+00
4.40467298e-01 -2.64133066e-01 1.58414375e-02 7.50738740e-01
-1.22177660e-01 3.34354162e-01 -8.50521386e-01 -9.87236351e-02
5.74649990e-01 1.13749564e+00 -4.39506263e-01 -6.28246129e-01
-3.56052071e-01 8.33636701e-01 4.55656499e-01 1.27870753e-01
-6.27606332e-01 -7.99221873e-01 1.40876040e-01 -2.36945480e-01
3.68415505e-01 1.93957910e-01 -7.27394879e-01 -7.38598645e-01
-1.34976566e-01 -7.43521690e-01 7.99838662e-01 -1.09622025e+00
-1.71048963e+00 1.04666984e+00 -2.63306260e-01 -8.53659809e-01
-4.43486661e-01 -4.20882732e-01 -7.50183463e-01 9.52390373e-01
-1.38099349e+00 -1.32703495e+00 -4.89737019e-02 6.71024323e-01
6.53398514e-01 -5.42063951e-01 1.01682174e+00 3.60860378e-01
-4.04191256e-01 8.62247944e-01 1.72380015e-01 1.43357053e-01
5.85189104e-01 -8.04974079e-01 7.61578619e-01 4.65955049e-01
1.65135264e-01 7.48270571e-01 4.01498556e-01 -6.46976709e-01
-9.53003585e-01 -1.17440665e+00 1.74944735e+00 -3.69621843e-01
9.65391517e-01 1.19582750e-02 -1.10323572e+00 7.51380324e-01
8.87931228e-01 -1.68902930e-02 6.82132185e-01 2.74527937e-01
-4.72939372e-01 -1.47310749e-01 -1.03261065e+00 4.59752887e-01
1.04127192e+00 -9.62945461e-01 -1.14300978e+00 8.33898336e-02
6.97408140e-01 1.00550577e-01 -1.10272336e+00 6.33238792e-01
7.04690397e-01 -7.43356705e-01 9.12967563e-01 -8.78344119e-01
6.61477745e-01 1.79881051e-01 -2.55427271e-01 -1.46974492e+00
-5.46824992e-01 -3.69381756e-01 1.15473360e-01 1.19419074e+00
6.05824709e-01 -5.78686535e-01 2.58960217e-01 4.57144231e-01
-4.24472272e-01 -7.69997418e-01 -9.00362432e-01 -1.05173147e+00
3.12065631e-01 -5.95996857e-01 6.41302764e-01 1.39571857e+00
5.23510873e-01 5.46471417e-01 -5.09992950e-02 -6.85722053e-01
-3.34575512e-02 -5.10956720e-02 -7.71361887e-02 -1.19019961e+00
3.31802756e-01 -9.41296697e-01 -2.67988920e-01 -1.21285565e-01
5.98803818e-01 -1.24157596e+00 -5.35016470e-02 -1.84672940e+00
6.17701095e-03 -7.97867179e-02 -8.86805654e-01 9.70541477e-01
3.29188146e-02 2.21793070e-01 6.07837662e-02 -8.84313881e-02
-7.00114131e-01 6.29707456e-01 3.69175553e-01 -3.01513225e-01
4.64120731e-02 -7.29540586e-01 -5.49236417e-01 7.45869696e-01
1.15754735e+00 -5.60933769e-01 -2.49959484e-01 -7.71818221e-01
3.80142927e-01 -2.71973580e-01 1.13730937e-01 -1.10751331e+00
6.96722329e-01 -2.36441358e-03 1.80654868e-01 -4.65232164e-01
2.07702398e-01 -7.08433926e-01 -4.73077632e-02 7.08442807e-01
-8.83500099e-01 2.96129674e-01 4.21006829e-01 1.53374281e-02
-4.71462369e-01 -2.38301501e-01 4.23806548e-01 -3.35336447e-01
-8.89268339e-01 -8.55664760e-02 -7.77331769e-01 -4.53307703e-02
5.57495475e-01 -3.86932313e-01 -3.70778948e-01 -2.71414608e-01
-3.89589459e-01 9.95961502e-02 -3.35819498e-02 9.71960843e-01
7.35060096e-01 -1.64850950e+00 -7.43813813e-01 -1.10543899e-01
4.54487860e-01 -7.45740592e-01 2.93040574e-02 6.20193362e-01
-3.42940271e-01 1.16530514e+00 -7.06807733e-01 -3.60238291e-02
-1.14607942e+00 7.06492484e-01 5.58057904e-01 -2.07870781e-01
-2.85097331e-01 8.43065262e-01 -5.14935017e-01 -7.27153420e-01
4.08096723e-02 -2.49322608e-01 -6.70115948e-01 1.71291098e-01
1.33218974e-01 5.62374115e-01 6.55753076e-01 -5.18782020e-01
-4.82897699e-01 3.40919256e-01 -7.25225918e-03 -1.09201051e-01
1.37225831e+00 -8.98151100e-03 -2.62822181e-01 7.62977302e-01
1.24692893e+00 -7.46279776e-01 -9.98173654e-02 -5.34927428e-01
5.04529953e-01 -1.15169831e-01 5.24253845e-01 -1.07039881e+00
-1.15654445e+00 1.11579764e+00 4.63261932e-01 1.68214098e-01
8.19674671e-01 -5.66068530e-01 1.32059205e+00 9.76573884e-01
-1.08668074e-01 -1.58141947e+00 5.03395312e-02 1.22752154e+00
1.06065798e+00 -1.04601407e+00 -3.17182779e-01 2.07890093e-01
-5.94724059e-01 1.47250533e+00 7.57287383e-01 -1.83299487e-03
5.93132138e-01 1.66492596e-01 1.16765369e-02 -2.64683276e-01
-1.23479474e+00 -2.26887986e-01 5.69595039e-01 4.36188519e-01
7.84474254e-01 -2.85222143e-01 -3.17427427e-01 5.97780466e-01
-2.72861302e-01 4.30815816e-01 1.82394132e-01 1.02767181e+00
-4.48016107e-01 -8.81143332e-01 -6.00934140e-02 6.32746100e-01
-6.32059634e-01 -8.02174687e-01 -8.86193752e-01 3.83283615e-01
1.05534062e-01 1.10842788e+00 2.23354027e-01 -7.14755476e-01
4.34875727e-01 4.87079859e-01 2.22415000e-01 -4.47784990e-01
-1.44415057e+00 -7.03946054e-01 6.27782762e-01 -4.86508489e-01
-3.62128496e-01 -3.09638739e-01 -1.05947387e+00 -5.64887762e-01
-2.97574073e-01 1.55730441e-01 6.73932433e-01 1.15635943e+00
6.46189392e-01 6.94172263e-01 2.36739114e-01 -5.04612088e-01
-3.66347939e-01 -1.39995265e+00 1.04986072e-01 2.85870254e-01
-7.04814270e-02 -3.38147670e-01 1.08202416e-02 -1.31253287e-01]
|
[10.759201049804688, 9.557857513427734]
|
110f952d-c222-4f15-951e-60e0252338a6
|
ove6d-object-viewpoint-encoding-for-depth
|
2203.01072
| null |
https://arxiv.org/abs/2203.01072v3
|
https://arxiv.org/pdf/2203.01072v3.pdf
|
OVE6D: Object Viewpoint Encoding for Depth-based 6D Object Pose Estimation
|
This paper proposes a universal framework, called OVE6D, for model-based 6D object pose estimation from a single depth image and a target object mask. Our model is trained using purely synthetic data rendered from ShapeNet, and, unlike most of the existing methods, it generalizes well on new real-world objects without any fine-tuning. We achieve this by decomposing the 6D pose into viewpoint, in-plane rotation around the camera optical axis and translation, and introducing novel lightweight modules for estimating each component in a cascaded manner. The resulting network contains less than 4M parameters while demonstrating excellent performance on the challenging T-LESS and Occluded LINEMOD datasets without any dataset-specific training. We show that OVE6D outperforms some contemporary deep learning-based pose estimation methods specifically trained for individual objects or datasets with real-world training data. The implementation and the pre-trained model will be made publicly available.
|
['Esa Rahtu', 'Janne Heikkilä', 'Dingding Cai']
|
2022-03-02
| null |
http://openaccess.thecvf.com//content/CVPR2022/html/Cai_OVE6D_Object_Viewpoint_Encoding_for_Depth-Based_6D_Object_Pose_Estimation_CVPR_2022_paper.html
|
http://openaccess.thecvf.com//content/CVPR2022/papers/Cai_OVE6D_Object_Viewpoint_Encoding_for_Depth-Based_6D_Object_Pose_Estimation_CVPR_2022_paper.pdf
|
cvpr-2022-1
|
['6d-pose-estimation']
|
['computer-vision']
|
[ 2.45050471e-02 1.79991528e-01 -1.36339888e-01 -5.71073592e-01
-6.60303354e-01 -7.19885170e-01 6.27207935e-01 -3.49557221e-01
-3.38866770e-01 1.93729475e-01 -1.09538294e-01 -2.03545336e-02
1.30412683e-01 -3.32398027e-01 -1.00905442e+00 -2.21415639e-01
-2.02641543e-02 1.20235837e+00 5.34231842e-01 2.18637332e-01
1.76238969e-01 1.04753435e+00 -1.19613397e+00 -6.53273519e-03
3.28680247e-01 1.31597900e+00 2.34519914e-01 5.74409604e-01
1.59173042e-01 3.42782587e-01 -6.03219807e-01 -4.82376903e-01
7.14742422e-01 2.72112936e-01 -7.18307257e-01 6.71983063e-01
1.15626574e+00 -8.53858173e-01 -4.23719674e-01 5.47701538e-01
6.88800395e-01 -1.65883750e-01 4.80826527e-01 -1.10506260e+00
-2.97031999e-01 2.36651376e-02 -6.01149559e-01 -2.08853677e-01
2.96359926e-01 2.17703074e-01 5.34079790e-01 -1.27978230e+00
8.67633581e-01 1.56553376e+00 8.63102973e-01 6.74207747e-01
-1.35002756e+00 -4.71557528e-01 3.01606536e-01 -2.59450972e-01
-1.26132274e+00 -3.51194054e-01 7.91664422e-01 -5.84543347e-01
1.18550050e+00 -1.35991409e-01 7.21497416e-01 1.29192269e+00
-1.81560908e-02 8.94242227e-01 7.50686705e-01 -2.52030998e-01
9.57613662e-02 -1.65197894e-01 -3.94676894e-01 6.76306188e-01
3.88588399e-01 -1.17916614e-01 -5.93881190e-01 -1.21188894e-01
1.29618359e+00 -2.09366947e-01 8.71792138e-02 -1.50518930e+00
-1.31854963e+00 4.35630023e-01 4.60777521e-01 -4.05086905e-01
-8.34084675e-02 5.32647014e-01 2.12396756e-01 -3.16264890e-02
6.73397303e-01 3.36101204e-01 -9.07814503e-01 2.89606489e-02
-6.03865385e-01 5.38385749e-01 6.16853356e-01 1.46788657e+00
6.00783587e-01 3.24473083e-02 1.65323079e-01 5.83518684e-01
5.89594066e-01 5.44453859e-01 1.49038285e-01 -1.21127415e+00
5.72164297e-01 5.72895169e-01 4.44086343e-01 -7.84805000e-01
-5.95099807e-01 -7.10820019e-01 -1.62344828e-01 3.23216259e-01
4.20968860e-01 -5.74774966e-02 -1.15053213e+00 1.55249548e+00
7.79814780e-01 2.02950373e-01 -3.20606589e-01 9.13995266e-01
9.25571799e-01 1.83736429e-01 -4.50223625e-01 2.64766216e-01
1.13037610e+00 -1.24962127e+00 -2.12667704e-01 -6.07560515e-01
3.41968834e-01 -9.65764165e-01 7.46786535e-01 3.79184425e-01
-1.20293939e+00 -5.68589866e-01 -1.09366190e+00 -4.31384087e-01
-1.48552790e-01 3.04855078e-01 8.39802682e-01 5.06718159e-01
-1.08655632e+00 3.75370532e-01 -9.86218631e-01 -3.55394959e-01
6.07020974e-01 6.17896318e-01 -5.79265058e-01 -2.75437310e-02
-3.29312891e-01 1.07082772e+00 3.66691113e-01 2.46262744e-01
-1.28631628e+00 -7.74510443e-01 -1.04579639e+00 -5.02412558e-01
5.51431715e-01 -1.08494234e+00 1.63092792e+00 -3.41267020e-01
-1.71226299e+00 1.24037361e+00 1.39133230e-01 -5.13106883e-01
9.82423663e-01 -7.73899257e-01 2.57855088e-01 1.41701087e-01
-1.48972556e-01 1.09691596e+00 1.17749679e+00 -1.42731118e+00
-1.81507394e-01 -6.95186675e-01 3.10772479e-01 4.29346591e-01
-1.17457714e-02 -1.06307894e-01 -9.13290739e-01 -6.40328526e-01
5.42501271e-01 -1.14886403e+00 -2.59823769e-01 6.69007301e-01
-4.12992358e-01 -5.81643842e-02 1.11923814e+00 -1.48013741e-01
1.91104874e-01 -1.85144269e+00 1.94539100e-01 -1.56776533e-01
2.72485733e-01 3.31123263e-01 -4.24904853e-01 2.34483063e-01
-1.30454928e-03 -3.88909787e-01 -6.66280761e-02 -9.08637881e-01
5.93644306e-02 2.17338115e-01 -1.42313346e-01 7.48457491e-01
2.72447795e-01 1.05953252e+00 -7.62325168e-01 -2.94029355e-01
5.84817708e-01 5.84453166e-01 -6.68138981e-01 3.12098265e-01
-6.33712590e-01 6.11433387e-01 -3.08406800e-01 6.46574318e-01
1.14144027e+00 -2.32281849e-01 1.72253465e-03 -3.32735509e-01
1.60434574e-01 3.04951012e-01 -1.22941625e+00 2.42164254e+00
-4.39543098e-01 5.32373786e-01 2.20753718e-02 -6.16795838e-01
1.08734512e+00 7.88448229e-02 6.23177350e-01 -1.70913577e-01
2.87911087e-01 1.29439086e-01 -2.28931248e-01 -4.04097944e-01
2.07446590e-01 2.08560765e-01 7.42376074e-02 3.29139262e-01
4.69758719e-01 -8.40913832e-01 -1.07415080e-01 -2.49264836e-02
7.99516022e-01 7.42164135e-01 1.17012918e-01 4.13694233e-02
2.27050066e-01 -1.18451655e-01 3.33344579e-01 4.71003890e-01
-1.66265532e-01 1.06924832e+00 2.20489904e-01 -7.22108901e-01
-1.06721306e+00 -1.23959649e+00 -3.77175927e-01 7.76915193e-01
3.91159266e-01 -3.15606952e-01 -5.60182691e-01 -6.54696584e-01
4.71617848e-01 4.14822489e-01 -5.13755441e-01 1.03636563e-01
-6.74356580e-01 -4.05926555e-01 2.44254366e-01 8.15463483e-01
6.10106051e-01 -7.25481808e-01 -7.01752007e-01 1.08261496e-01
1.01963736e-01 -1.64292717e+00 -4.38272208e-01 2.56729662e-01
-1.18232489e+00 -1.01111734e+00 -6.91544056e-01 -7.06306517e-01
6.45586312e-01 3.21726143e-01 1.24689579e+00 -3.44006777e-01
-2.80461073e-01 5.85706830e-01 7.64466524e-02 -6.41959071e-01
7.18697980e-02 1.82356119e-01 1.47399515e-01 -2.00655639e-01
2.40297332e-01 -6.09512269e-01 -7.84427404e-01 4.08701032e-01
-4.60659415e-01 1.82415724e-01 5.74730992e-01 4.91957068e-01
7.15267420e-01 -5.18842757e-01 3.79115939e-02 -5.30672431e-01
-9.39728767e-02 7.11572841e-02 -9.01165485e-01 -2.62564719e-01
-9.75542516e-02 -1.65536359e-01 6.28815871e-03 -5.57459712e-01
-1.00100267e+00 5.73065281e-01 -1.09011561e-01 -7.35100091e-01
-2.72501379e-01 -1.59914553e-01 -3.22318792e-01 -4.96500522e-01
6.34611547e-01 -2.32842341e-01 -8.23230967e-02 -9.52225983e-01
3.38999540e-01 1.71923667e-01 6.00148976e-01 -5.79082251e-01
1.19887853e+00 8.32221448e-01 1.60610467e-01 -5.06350338e-01
-1.07768488e+00 -4.38233733e-01 -1.29492354e+00 -1.27033427e-01
7.55714476e-01 -1.11765492e+00 -7.96618760e-01 7.57177174e-01
-1.38350391e+00 -4.93743092e-01 -2.84102648e-01 6.01518869e-01
-9.54039574e-01 1.01385228e-01 -4.15896326e-01 -6.13696933e-01
3.32271727e-03 -1.07477903e+00 1.82879698e+00 -9.99673232e-02
-5.53165339e-02 -7.42233932e-01 4.52190712e-02 5.73387742e-01
9.23509002e-02 4.59807456e-01 5.11630595e-01 -4.41618085e-01
-9.82338309e-01 -4.45306331e-01 -2.13556096e-01 2.92864054e-01
-1.05549870e-02 -4.42672856e-02 -1.03194427e+00 -3.86305630e-01
9.30116400e-02 -6.97261214e-01 6.99998319e-01 4.37413901e-01
1.31151068e+00 -8.31502676e-03 -3.46951038e-01 1.01975644e+00
1.27356768e+00 -1.37627438e-01 2.07075924e-01 3.39581251e-01
9.91704464e-01 4.51256365e-01 5.82833111e-01 3.81013304e-01
4.22164261e-01 9.59404469e-01 1.05943680e+00 -9.54426229e-02
-2.62562603e-01 -4.50747073e-01 5.69470860e-02 4.39435124e-01
-1.29505768e-01 -1.83811203e-01 -9.41926360e-01 3.75910789e-01
-1.66902041e+00 -4.59328353e-01 9.45582893e-03 2.10694933e+00
4.86520320e-01 3.89236122e-01 9.14586335e-02 -2.48944581e-01
3.53744596e-01 3.22922647e-01 -1.00328982e+00 -3.66626382e-02
-2.44166404e-02 1.57539845e-01 5.21321356e-01 4.97594863e-01
-1.29859722e+00 1.18883908e+00 7.04471445e+00 5.03488541e-01
-1.02165771e+00 4.83810566e-02 2.62101799e-01 -3.31836194e-01
-1.62225544e-01 -1.76879354e-02 -8.94216359e-01 -2.65037745e-01
1.87294245e-01 3.34187061e-01 1.01883128e-01 1.20538914e+00
-4.81380224e-02 -8.81143939e-03 -1.47909856e+00 1.15193343e+00
2.52959132e-01 -1.17406464e+00 -2.74934154e-02 2.61818934e-02
7.70782650e-01 3.90900075e-01 2.43403822e-01 2.52171513e-02
1.86298802e-01 -8.98750424e-01 1.08198953e+00 9.53153744e-02
7.21679449e-01 -4.26031440e-01 3.20752263e-01 5.18250108e-01
-1.01036811e+00 1.16712958e-01 -5.30807495e-01 -1.02062814e-01
1.04227170e-01 3.75920594e-01 -1.10311139e+00 3.56715083e-01
7.38875091e-01 7.64679253e-01 -7.72579074e-01 1.10332131e+00
-2.07499459e-01 4.25830185e-02 -5.37739992e-01 3.54116470e-01
1.06704749e-01 8.14163089e-02 7.74902344e-01 8.78593385e-01
9.27513316e-02 -2.83347636e-01 1.00630693e-01 6.49018943e-01
-1.21419735e-01 -2.87691206e-01 -6.15217507e-01 3.58973205e-01
2.07100451e-01 1.23378754e+00 -7.48195887e-01 -1.09947063e-01
-2.57910550e-01 1.03438830e+00 3.18435967e-01 3.23121339e-01
-8.23915124e-01 7.71297142e-02 7.43308187e-01 3.00898761e-01
6.19756937e-01 -6.68422937e-01 -3.01294535e-01 -1.28896487e+00
3.14084291e-01 -6.12284839e-01 -5.59524298e-02 -1.03356981e+00
-1.16361594e+00 5.69949865e-01 3.16716820e-01 -1.36316574e+00
-1.72998235e-01 -1.08384085e+00 -4.10498977e-02 5.74407756e-01
-1.23623359e+00 -1.43537116e+00 -4.94226128e-01 3.47321182e-01
7.35024869e-01 -3.26384977e-02 6.75965726e-01 -4.41987701e-02
-1.83149159e-01 5.29067814e-01 -2.91355520e-01 5.07313684e-02
7.39062309e-01 -1.17329156e+00 8.81296635e-01 4.10881400e-01
2.42925659e-01 4.09438282e-01 6.57516301e-01 -4.57798958e-01
-1.55187309e+00 -1.07777786e+00 4.98614222e-01 -1.11064351e+00
3.92429531e-01 -1.04827392e+00 -5.22826493e-01 1.09428000e+00
2.64040922e-04 4.43103373e-01 1.28356397e-01 1.05857747e-02
-5.51910877e-01 -1.85581669e-01 -1.16355622e+00 5.85658669e-01
1.76921010e+00 -3.69706511e-01 -6.48077309e-01 6.33496284e-01
9.20612335e-01 -1.32392204e+00 -7.11419106e-01 8.84285688e-01
7.13012874e-01 -9.81240749e-01 1.35792267e+00 -5.93257248e-01
2.25794345e-01 -4.88874048e-01 -9.52806249e-02 -9.20422018e-01
-1.63710788e-01 -5.40809572e-01 -5.35227001e-01 6.53948724e-01
1.39765561e-01 -4.35048997e-01 1.22561264e+00 3.20162982e-01
-3.42261702e-01 -9.51505840e-01 -9.01554346e-01 -8.77967894e-01
-4.68763635e-02 -6.32852137e-01 5.70260942e-01 5.10826886e-01
-7.95149505e-01 3.63144487e-01 -3.82777661e-01 3.48459989e-01
8.60445082e-01 2.80022413e-01 1.55438101e+00 -1.14062178e+00
-2.61635214e-01 -1.77449226e-01 -7.95389771e-01 -1.79090905e+00
1.14717327e-01 -5.99809110e-01 -3.28508765e-02 -1.49819207e+00
9.59979892e-02 -2.20914543e-01 3.37203056e-01 4.68883872e-01
2.42059439e-01 5.95837355e-01 7.50545189e-02 8.48832950e-02
-5.29856443e-01 7.73305058e-01 1.75391519e+00 -7.67107904e-02
-8.57766941e-02 2.23275706e-01 -2.61561632e-01 1.14918458e+00
3.73551488e-01 -4.28511441e-01 -5.92816651e-01 -8.74804795e-01
5.44224158e-02 -2.13116854e-01 5.96036196e-01 -1.06493640e+00
2.40624044e-03 -1.48336723e-01 8.96171868e-01 -1.28451109e+00
9.42996204e-01 -9.87689853e-01 1.15602709e-01 3.88668031e-01
-9.50894505e-02 5.91162518e-02 3.88896793e-01 4.91061896e-01
1.13508366e-01 -1.35696772e-02 6.91615045e-01 -1.47642508e-01
-6.32040501e-01 7.67581403e-01 1.30764708e-01 -6.56800196e-02
1.11134839e+00 -5.15509784e-01 -2.23444805e-01 -2.95975626e-01
-8.21076393e-01 1.53629050e-01 7.62509942e-01 6.21275127e-01
6.56000376e-01 -1.40664887e+00 -6.12974048e-01 2.46462613e-01
3.65153760e-01 7.72435784e-01 9.44737345e-02 5.71454287e-01
-9.32814002e-01 4.23505485e-01 -9.65511203e-02 -1.22373331e+00
-8.97731245e-01 6.03113413e-01 6.51895106e-01 5.12158200e-02
-6.95037365e-01 1.22625303e+00 5.05158246e-01 -9.09170687e-01
5.46597481e-01 -4.49887186e-01 2.94131756e-01 -2.75477171e-01
1.43636361e-01 9.05351490e-02 2.29916573e-01 -7.26953864e-01
-4.97421890e-01 9.47556138e-01 -6.93297535e-02 -1.89926267e-01
1.48303676e+00 -4.93662618e-02 1.86043277e-01 4.35838997e-01
1.19133878e+00 -3.27276103e-02 -1.91903126e+00 -3.56474370e-01
-2.56152362e-01 -7.80342281e-01 -2.46260613e-01 -5.90424418e-01
-1.04173350e+00 8.47654760e-01 4.94503319e-01 -4.87977475e-01
8.28302562e-01 2.47126386e-01 5.38521528e-01 6.33954287e-01
7.21181691e-01 -8.37503552e-01 6.94245994e-01 6.04975224e-01
1.25544226e+00 -1.23870099e+00 2.35326231e-01 -6.93996489e-01
-2.24339247e-01 1.05666113e+00 1.06659257e+00 -4.68831122e-01
6.79779053e-01 4.20184016e-01 2.26223260e-01 -2.24544182e-01
-6.44026339e-01 1.83529910e-02 5.45083046e-01 8.40099633e-01
7.88483024e-02 -2.88313955e-01 2.39023417e-01 -2.96811108e-02
-3.09018105e-01 -2.85465956e-01 1.10498197e-01 9.82511222e-01
-1.56257704e-01 -1.02782321e+00 -4.10898924e-01 1.48162380e-01
-9.55460295e-02 3.48560870e-01 -6.56924367e-01 9.88593698e-01
3.34608287e-01 2.59917647e-01 3.23267490e-01 -2.10357383e-01
4.98336941e-01 -1.56991795e-01 1.23741758e+00 -9.75038469e-01
-2.00801745e-01 2.04079404e-01 9.14440956e-03 -8.34609985e-01
-5.50992250e-01 -6.43656194e-01 -8.77620280e-01 -1.09709539e-01
-3.89999509e-01 -7.41178632e-01 7.70080745e-01 8.53133440e-01
3.91920865e-01 1.78797752e-01 3.94941211e-01 -1.62989783e+00
-5.07652342e-01 -9.54432845e-01 -1.41776875e-01 1.74786568e-01
5.62638581e-01 -1.06329870e+00 -1.42690256e-01 -7.14609101e-02]
|
[7.619314193725586, -2.6631405353546143]
|
0b27344e-5854-4bdb-9cc1-9e8529f6da59
|
hierarchical-multi-instance-multi-label
|
2305.19419
| null |
https://arxiv.org/abs/2305.19419v1
|
https://arxiv.org/pdf/2305.19419v1.pdf
|
Hierarchical Multi-Instance Multi-Label Learning for Detecting Propaganda Techniques
|
Since the introduction of the SemEval 2020 Task 11 (Martino et al., 2020a), several approaches have been proposed in the literature for classifying propaganda based on the rhetorical techniques used to influence readers. These methods, however, classify one span at a time, ignoring dependencies from the labels of other spans within the same context. In this paper, we approach propaganda technique classification as a Multi-Instance Multi-Label (MIML) learning problem (Zhou et al., 2012) and propose a simple RoBERTa-based model (Zhuang et al., 2021) for classifying all spans in an article simultaneously. Further, we note that, due to the annotation process where annotators classified the spans by following a decision tree, there is an inherent hierarchical relationship among the different techniques, which existing approaches ignore. We incorporate these hierarchical label dependencies by adding an auxiliary classifier for each node in the decision tree to the training objective and ensembling the predictions from the original and auxiliary classifiers at test time. Overall, our model leads to an absolute improvement of 2.47% micro-F1 over the model from the shared task winning team in a cross-validation setup and is the best performing non-ensemble model on the shared task leaderboard.
|
['Bhuwan Dhingra', 'Anni Chen']
|
2023-05-30
| null | null | null | null |
['multi-label-learning']
|
['methodology']
|
[ 1.24330632e-01 2.58868545e-01 -4.86915320e-01 -1.37304798e-01
-8.65452647e-01 -7.48467982e-01 1.09001172e+00 3.96318108e-01
-3.53882194e-01 8.46081138e-01 5.21896183e-01 -5.41377187e-01
-3.48197259e-02 -5.12146115e-01 -6.68859243e-01 -6.08004212e-01
2.49129832e-01 3.07426065e-01 1.08069576e-01 -1.88632831e-01
5.37201285e-01 -3.97493057e-02 -1.23730493e+00 6.90685391e-01
9.15671766e-01 8.79047871e-01 -4.05524857e-02 4.87774044e-01
-2.87975490e-01 1.60795975e+00 -1.03827846e+00 -5.12294948e-01
-2.18478322e-01 -5.75496018e-01 -1.12643445e+00 -2.05444917e-01
6.73525870e-01 1.91846728e-01 1.02312468e-01 5.99092066e-01
3.25848937e-01 -1.80538282e-01 9.48512852e-01 -9.12819564e-01
-4.32014138e-01 1.14872825e+00 -6.12987518e-01 1.17566861e-01
2.14988887e-01 -6.52232707e-01 1.15278947e+00 -6.49106622e-01
8.39876831e-01 1.07799363e+00 8.41444314e-01 3.65093112e-01
-1.34724057e+00 -7.04290092e-01 1.26011685e-01 5.95381260e-01
-9.80620861e-01 -3.17168385e-01 1.06665349e+00 -8.95310581e-01
6.71786010e-01 2.19003275e-01 3.62990797e-01 1.60060084e+00
3.71756405e-01 6.94793224e-01 1.84863341e+00 -7.47515142e-01
7.10315481e-02 1.74502552e-01 4.96398270e-01 4.58166778e-01
-1.31186411e-01 -2.19059289e-01 -6.09298408e-01 -2.15978846e-01
-1.39674589e-01 -6.29397035e-01 1.92579348e-02 1.01385005e-01
-9.99926448e-01 1.09343112e+00 3.76297891e-01 6.26682222e-01
-1.57378510e-01 4.65383939e-02 7.12665200e-01 2.34626681e-01
9.09229040e-01 5.50860405e-01 -4.37393248e-01 -9.28642973e-02
-1.08235860e+00 2.74037302e-01 1.04099452e+00 3.18625301e-01
4.09106702e-01 -2.56641239e-01 -6.05980873e-01 8.43697071e-01
1.27719725e-02 1.63522989e-01 3.97729576e-01 -7.82903492e-01
4.94960427e-01 4.16491121e-01 -8.24686140e-02 -1.18160546e+00
-6.41961694e-01 -9.47314620e-01 -5.46701431e-01 2.72473856e-03
5.33499062e-01 -1.75176591e-01 -6.31621838e-01 1.78323591e+00
2.14348108e-01 -1.56120867e-01 -1.12290740e-01 4.40769613e-01
8.40507686e-01 4.77397144e-01 3.96004885e-01 -4.97187436e-01
1.37571156e+00 -1.34196961e+00 -9.47223246e-01 -1.02256112e-01
1.17467558e+00 -1.03620875e+00 7.13907123e-01 5.92733741e-01
-4.35616821e-01 -4.90842074e-01 -1.14163077e+00 -1.42346352e-01
-3.67981493e-01 1.44440219e-01 2.24325910e-01 4.60382044e-01
-4.52654898e-01 4.57951665e-01 -3.14389139e-01 -7.09789023e-02
3.33912909e-01 -3.18438441e-01 6.14142325e-03 3.25396627e-01
-1.45785856e+00 1.47733617e+00 3.63166600e-01 -2.48307422e-01
-8.11485648e-01 -6.37453318e-01 -3.86786252e-01 -3.66384000e-01
6.48460686e-01 -4.97077480e-02 1.18330300e+00 -1.03358746e+00
-1.19171989e+00 1.14848697e+00 9.41759795e-02 -5.18399715e-01
5.44953465e-01 -4.27456230e-01 -3.24233770e-01 -2.07569942e-01
3.43475848e-01 2.51975745e-01 7.34988630e-01 -1.34219861e+00
-6.19710684e-01 -1.58792496e-01 2.96937317e-01 6.55460134e-02
-3.16581398e-01 3.37807089e-01 2.79613465e-01 -8.07920933e-01
-3.23635966e-01 -1.05493748e+00 2.67880142e-01 -5.67487657e-01
-6.31823003e-01 -8.34502101e-01 8.41615975e-01 -1.00202334e+00
1.64924932e+00 -1.93517768e+00 3.20271850e-01 -1.91838890e-01
4.08777297e-01 1.67195648e-01 2.29478106e-02 3.89083624e-01
2.68699899e-02 3.46571118e-01 -3.90778556e-02 -3.26589376e-01
-1.70359835e-01 3.67678180e-02 -3.75426948e-01 4.21520859e-01
-2.15667784e-01 7.17871726e-01 -7.68587887e-01 -7.74104774e-01
2.42305989e-03 2.99740881e-01 -1.61842391e-01 -1.20358422e-01
-3.39651585e-01 6.31359220e-01 -3.26733440e-01 3.34978908e-01
2.33628049e-01 -2.92112559e-01 4.24939662e-01 -2.79176265e-01
-4.47905093e-01 7.84775496e-01 -5.40660620e-01 1.49448991e+00
-6.92843139e-01 8.61348867e-01 -6.46562725e-02 -1.22177291e+00
8.32175493e-01 3.90107900e-01 4.12388742e-01 -6.07714653e-01
3.04441571e-01 3.82828504e-01 3.03386837e-01 -2.69945443e-01
3.08389187e-01 -2.45372400e-01 -5.06812632e-01 5.72094083e-01
-5.38395159e-03 9.66588110e-02 3.97430301e-01 2.08202705e-01
1.18815899e+00 2.41808742e-01 6.16703451e-01 -4.97213185e-01
6.18650138e-01 2.06390008e-01 3.91126543e-01 8.69028449e-01
-1.44148782e-01 2.16461509e-01 6.63446844e-01 -6.33401871e-01
-8.45108569e-01 -6.62171602e-01 -6.31451607e-01 1.48464966e+00
-3.12801778e-01 -7.75049627e-01 -6.16709232e-01 -1.06779647e+00
-1.61170259e-01 1.17282236e+00 -1.10543334e+00 3.21864814e-01
-7.05569029e-01 -5.70631206e-01 8.12974811e-01 2.10164070e-01
3.47360253e-01 -1.08667576e+00 -8.97916257e-01 5.08457839e-01
-5.82968593e-01 -1.10146117e+00 9.10297967e-03 6.76888108e-01
-4.09457773e-01 -1.27779436e+00 -3.35761815e-01 -5.25394738e-01
-1.25319383e-03 -2.23617405e-01 1.05452859e+00 4.07978930e-02
-1.60960015e-02 -8.55333358e-02 -6.87260568e-01 -5.79596043e-01
-7.87179947e-01 4.21470433e-01 -3.59058261e-01 -2.11220965e-01
8.12724084e-02 -2.20828772e-01 7.08583966e-02 9.65803564e-02
-3.94466966e-01 5.11661708e-01 3.07242602e-01 9.32803690e-01
2.31935740e-01 -2.30210155e-01 7.54506230e-01 -1.17223847e+00
5.69723725e-01 -5.17475307e-01 -1.50167108e-01 3.78922045e-01
-7.97754228e-01 -1.34285614e-02 6.87355638e-01 -4.39729482e-01
-9.62255299e-01 -4.20312047e-01 1.26110418e-02 1.24770336e-01
1.52063206e-01 8.68570507e-01 2.33858258e-01 2.28363603e-01
9.54081237e-01 -2.50932038e-01 -7.97360986e-02 -6.09227896e-01
3.69857728e-01 8.67690444e-01 1.61503479e-01 -6.99149013e-01
5.74720979e-01 -8.20311755e-02 1.03398181e-01 -3.36924851e-01
-2.05200410e+00 -3.09331954e-01 -7.80440867e-01 -5.34049511e-01
8.58354151e-01 -7.39561558e-01 -3.19488645e-01 4.45346624e-01
-1.48661494e+00 -3.67364556e-01 4.69534378e-03 1.74063087e-01
-2.94040024e-01 7.18711838e-02 -5.43684602e-01 -7.77424574e-01
-3.04810762e-01 -8.57131660e-01 5.74613929e-01 -9.95053202e-02
-6.02759302e-01 -1.16304326e+00 1.99374139e-01 7.04802990e-01
3.53953809e-01 6.05705559e-01 1.28353035e+00 -1.01016319e+00
2.72923321e-01 -2.95849144e-02 2.29850244e-02 3.90257061e-01
1.90542445e-01 -1.36941656e-01 -9.66788411e-01 8.69827271e-02
2.06021443e-01 -7.25041091e-01 1.04473734e+00 1.72239915e-01
1.01684844e+00 -6.23938322e-01 -3.69458705e-01 -1.03109263e-01
1.09635401e+00 4.22121882e-02 4.12691563e-01 8.75324428e-01
7.22374022e-01 8.71159613e-01 5.07447243e-01 1.88856006e-01
3.90899539e-01 8.91836524e-01 1.97689131e-01 2.39770487e-01
-3.17185193e-01 -1.49986401e-01 4.47564453e-01 7.82603681e-01
-1.41987309e-01 -2.45043457e-01 -1.05042005e+00 3.48970920e-01
-1.76274502e+00 -1.11422789e+00 -6.59089267e-01 1.97208643e+00
1.17015874e+00 2.72456348e-01 3.97150703e-02 3.10987890e-01
6.42995715e-01 5.62250555e-01 -8.38177372e-03 -7.26058602e-01
-3.69335473e-01 1.83774784e-01 2.12889537e-01 5.48826337e-01
-1.46024680e+00 8.09738815e-01 5.99539375e+00 1.20219958e+00
-1.00830710e+00 7.19245851e-01 6.73660100e-01 1.10211382e-02
-1.57288145e-02 1.45907789e-01 -8.09085906e-01 5.80569327e-01
1.13171220e+00 -8.99960697e-02 2.26508349e-01 5.87892056e-01
-9.30714831e-02 -1.87325582e-01 -8.21633816e-01 3.62799704e-01
5.57847202e-01 -1.19321465e+00 -2.08368078e-01 1.51037350e-01
9.00963247e-01 -3.54182795e-02 -1.70382753e-01 5.61840296e-01
1.58303693e-01 -1.08110964e+00 1.09288561e+00 2.76527137e-01
3.70985240e-01 -3.75747621e-01 8.08784842e-01 6.77151978e-01
-6.33728027e-01 -2.36799851e-01 1.54669389e-01 -3.73965412e-01
2.22550035e-01 7.55650401e-01 -6.13930345e-01 6.45421684e-01
4.21770126e-01 6.15129530e-01 -8.48718584e-01 4.95274365e-01
-6.76189423e-01 1.17473352e+00 4.22195531e-02 -1.94276094e-01
3.34624171e-01 1.75679699e-01 5.42955935e-01 1.23540044e+00
4.32440871e-03 -2.80987024e-01 5.05164742e-01 4.53711033e-01
-1.44222528e-01 5.06075561e-01 -2.72160023e-01 2.90155597e-02
3.46811950e-01 1.42011511e+00 -6.53891444e-01 -3.73473406e-01
-2.40859598e-01 4.58435774e-01 5.91695607e-01 -3.25507261e-02
-1.07700729e+00 5.24709746e-02 -2.63288260e-01 2.56808132e-01
3.99898477e-02 -2.68155392e-02 -6.02584064e-01 -5.62370062e-01
-9.78754088e-02 -7.83517122e-01 4.93724376e-01 -3.88666987e-01
-1.24405670e+00 4.61220950e-01 7.11528063e-02 -1.01601338e+00
-5.15590906e-02 -6.68533623e-01 -4.43906724e-01 6.22400880e-01
-1.25806665e+00 -1.52553666e+00 -3.56004387e-02 -6.72839805e-02
6.28096282e-01 -1.06855869e-01 8.98798883e-01 8.33341405e-02
-2.92576313e-01 3.03408355e-01 -1.07220514e-02 2.17857048e-01
1.05365539e+00 -1.22212136e+00 -3.37209702e-01 4.53879327e-01
1.76283494e-01 4.35238391e-01 8.78210962e-01 -6.94485843e-01
-5.25448263e-01 -8.89931202e-01 1.51664853e+00 -6.77762747e-01
1.12885988e+00 -1.88833728e-01 -8.54406893e-01 5.23110747e-01
5.74504614e-01 -4.57831293e-01 9.95028794e-01 7.98563600e-01
-7.82653987e-01 2.57925779e-01 -6.63440943e-01 1.11316085e-01
7.92635560e-01 -4.30595785e-01 -8.82840812e-01 6.14731610e-01
3.52468312e-01 -4.01726663e-01 -8.94169033e-01 5.36076844e-01
6.44708931e-01 -7.29455769e-01 4.36638862e-01 -6.05188310e-01
1.22854865e+00 -4.03088555e-02 -1.75580800e-01 -1.24424314e+00
-3.46853435e-01 -6.12713285e-02 -3.16181749e-01 1.37335753e+00
6.83590770e-01 -3.28888863e-01 2.30012044e-01 -2.71974821e-02
-2.12615490e-01 -8.74208808e-01 -1.24305332e+00 -8.20421159e-01
5.18104434e-01 -3.77380371e-01 -2.18808651e-01 1.23426664e+00
4.88213412e-02 1.01903152e+00 -7.29706526e-01 -5.02287865e-01
5.68385065e-01 1.91449642e-01 4.86276388e-01 -1.61143184e+00
-2.48143822e-01 -5.95291495e-01 3.57939787e-02 -6.06456637e-01
6.15322649e-01 -1.42079592e+00 1.30677685e-01 -1.63561082e+00
4.91710722e-01 -5.26754439e-01 -4.40897077e-01 6.78516150e-01
-1.20648019e-01 1.23568147e-01 3.57599318e-01 5.36517739e-01
-6.48016989e-01 2.67997265e-01 1.07236230e+00 -3.79155576e-01
6.57432228e-02 -1.92806900e-01 -7.65613914e-01 8.27380538e-01
7.86453068e-01 -8.27051938e-01 -9.09585133e-03 -3.24424744e-01
2.40576163e-01 -1.61952928e-01 5.30610919e-01 -9.14777935e-01
1.41478613e-01 -1.64532259e-01 2.26864055e-01 -5.89317620e-01
1.75594121e-01 -4.07398283e-01 1.48304388e-01 6.28341019e-01
-7.29673445e-01 -2.20874891e-01 -2.87232865e-02 4.23234433e-01
8.21672827e-02 -6.74512804e-01 5.90863883e-01 8.06751177e-02
-2.40880981e-01 -5.50867379e-01 -4.60725665e-01 2.75324076e-01
9.79618430e-01 2.09159404e-01 -9.27305222e-01 -3.95561345e-02
-7.68293679e-01 -1.17582403e-01 1.20108604e-01 4.66451585e-01
-2.58498996e-01 -1.09871638e+00 -9.92573619e-01 -6.52479708e-01
4.53668088e-02 -4.01223361e-01 7.30130970e-02 1.26503217e+00
-2.45345131e-01 4.24586803e-01 -4.26499918e-02 -3.69993001e-01
-1.32065690e+00 3.87843788e-01 3.27259605e-03 -8.94059181e-01
-3.36326092e-01 5.79850972e-01 -9.73056480e-02 -2.95986563e-01
3.77005376e-02 3.06593031e-01 -6.03466988e-01 6.84001267e-01
3.35288316e-01 5.93150795e-01 6.90521672e-02 -9.10233438e-01
-3.15295070e-01 3.39842260e-01 -2.36265063e-01 -1.48498833e-01
1.23035669e+00 2.56632060e-01 -2.44314775e-01 1.14013779e+00
1.04740131e+00 3.75152469e-01 -4.85603273e-01 -2.71056682e-01
4.63513136e-01 6.61822855e-02 1.78311199e-01 -1.48090756e+00
-4.90241855e-01 7.05384970e-01 2.30854675e-01 4.28121150e-01
6.96862400e-01 1.48303702e-01 2.10458711e-01 1.56909868e-01
2.33409077e-01 -1.48807132e+00 1.68887198e-01 8.02355647e-01
1.28462517e+00 -9.18348968e-01 2.29774281e-01 -4.42169994e-01
-6.13565207e-01 1.09401274e+00 4.60857272e-01 1.16074137e-01
4.30664599e-01 2.51981705e-01 1.88289449e-01 -1.85196653e-01
-8.63340259e-01 5.18663190e-02 4.25249219e-01 1.88969463e-01
9.38515484e-01 2.99107134e-01 -1.28763437e+00 7.21740484e-01
-1.94862768e-01 -2.47502208e-01 4.28823590e-01 7.68680096e-01
-5.57856500e-01 -1.13881111e+00 -4.10363734e-01 7.07274377e-01
-7.10746586e-01 -2.38053516e-01 -6.43823385e-01 7.98671663e-01
4.07531440e-01 1.21759629e+00 -1.98535204e-01 -5.18698037e-01
3.26889046e-02 5.64070880e-01 4.90728498e-01 -6.23272657e-01
-1.04683626e+00 1.62828326e-01 7.59755969e-01 8.93328413e-02
-8.82002294e-01 -6.64196253e-01 -9.49334919e-01 -5.77711761e-02
-4.67548043e-01 4.38271612e-01 6.75980747e-01 1.33142495e+00
-1.62822790e-02 7.97765434e-01 6.76807404e-01 -5.79280853e-01
-7.35741198e-01 -1.57612455e+00 -2.33879864e-01 3.97833586e-01
-1.60671040e-01 -9.71798956e-01 -6.43639743e-01 8.52959529e-02]
|
[8.505366325378418, 10.670791625976562]
|
d4375ac3-d072-4dc0-b1c6-968c7926d294
|
its-about-time-turn-entry-timing-for-situated
| null | null |
https://aclanthology.org/2020.sigdial-1.12
|
https://aclanthology.org/2020.sigdial-1.12.pdf
|
It’s About Time: Turn-Entry Timing For Situated Human-Robot Dialogue
|
Turn-entry timing is an important requirement for conversation, and one that spoken dialogue systems largely fail at. In this paper, we introduce a computational framework based on work from Psycholinguistics, which is aimed at achieving proper turn-taking timing for situated agents. The approach involves incremental processing and lexical prediction of the turn in progress, which allows a situated dialogue system to start its turn and initiate actions earlier than would otherwise be possible. We evaluate the framework by integrating it within a cognitive robotic architecture and testing performance on a corpus of task-oriented human-robot directives. We demonstrate that: 1) the system is superior to a non-incremental system in terms of faster responses, reduced gap between turns, and the ability to perform actions early, 2) the system can time its turn to come in immediately at a transition point or earlier to produce several types of overlap, and 3) the system is robust to various forms of disfluency in the input. Overall, this domain-independent framework can be integrated into various dialogue systems to improve responsiveness, and is a step toward more natural, human-like turn-taking behavior.
|
['Matthias Scheutz', 'Antonio Roque', 'Ravenna Thielstrom', 'Felix Gervits']
| null | null | null | null |
sigdial-acl-2020-7
|
['spoken-dialogue-systems']
|
['speech']
|
[ 8.74703899e-02 7.05775201e-01 2.65012562e-01 -6.30011976e-01
-6.30498171e-01 -9.07159984e-01 9.67643499e-01 1.44179061e-01
-4.14914817e-01 7.21124709e-01 6.55409455e-01 -6.26431942e-01
3.33407633e-02 -4.67138737e-01 -9.27257389e-02 -2.36985922e-01
9.29185599e-02 7.87116587e-01 2.39102542e-01 -9.16012287e-01
5.72626173e-01 5.90301335e-01 -1.30127597e+00 4.00908709e-01
5.87822616e-01 2.56153733e-01 4.92377311e-01 8.31329703e-01
-5.95580414e-02 1.39608228e+00 -6.82544768e-01 2.13017464e-01
1.89792532e-02 -7.03899860e-01 -1.65541637e+00 1.54349640e-01
-2.56482542e-01 -4.92916137e-01 -1.49225667e-01 6.31962419e-01
3.57202351e-01 3.60601127e-01 5.48055470e-01 -9.65427756e-01
-3.09325516e-01 9.28615928e-01 3.97717535e-01 -5.97415529e-02
1.06926966e+00 5.05948126e-01 8.14325690e-01 -5.24569511e-01
6.87768877e-01 1.63374376e+00 5.13830245e-01 7.64507711e-01
-1.19215155e+00 8.74419063e-02 5.15867174e-02 -1.15338534e-01
-1.00319207e+00 -8.96975875e-01 5.08455336e-01 -3.89706671e-01
1.51525831e+00 2.96948642e-01 5.09046316e-01 8.39086950e-01
1.70119062e-01 6.82889640e-01 1.15599787e+00 -8.70136023e-01
2.34555557e-01 2.79132038e-01 6.04422428e-02 3.91656667e-01
-6.93324327e-01 1.56365797e-01 -5.07979512e-01 -1.06383124e-02
6.21643066e-01 -5.95203221e-01 -1.81749165e-01 1.12859428e-01
-1.41087604e+00 6.63713157e-01 2.12660700e-01 7.94921279e-01
-5.84297359e-01 -3.62365931e-01 5.39829969e-01 7.17914999e-01
3.01818680e-02 8.99419129e-01 -2.59707093e-01 -8.00085783e-01
-2.09995806e-01 5.83354175e-01 1.57709038e+00 1.05177939e+00
4.20567304e-01 -2.09406301e-01 -2.85255253e-01 1.04284918e+00
2.10937202e-01 1.16389930e-01 3.07061613e-01 -1.43522072e+00
4.82709080e-01 7.83230186e-01 5.99661171e-01 -5.76193213e-01
-9.57457960e-01 7.02948987e-01 4.05345075e-02 3.72383922e-01
7.71168709e-01 -3.21059734e-01 -4.23655003e-01 1.74453390e+00
4.37019616e-01 -9.25427496e-01 6.16940677e-01 8.86204243e-01
6.25815928e-01 8.29183280e-01 1.07408822e-01 -4.36747164e-01
1.37913191e+00 -8.01367223e-01 -7.88564563e-01 -3.74177396e-01
1.14167738e+00 -9.61357832e-01 1.26148212e+00 3.53177816e-01
-1.04576802e+00 -3.85294467e-01 -8.85495007e-01 -2.11088449e-01
-1.54295582e-02 -1.74968943e-01 6.10192120e-01 4.27066267e-01
-1.14507735e+00 3.61380398e-01 -6.83949530e-01 -8.04368913e-01
-7.38991976e-01 2.32859582e-01 -3.25017631e-01 2.13057771e-01
-1.18716300e+00 1.41076791e+00 4.39562947e-01 1.90941662e-01
-4.35979933e-01 2.21371111e-02 -9.11812067e-01 -2.88190663e-01
3.46497416e-01 -1.70734271e-01 2.07291937e+00 -1.00474155e+00
-2.31734538e+00 8.03374171e-01 -1.40583292e-02 -3.44184041e-01
4.28956956e-01 -8.67258683e-02 -8.63870159e-02 1.12308212e-01
9.01040360e-02 8.08830738e-01 -3.14875692e-02 -8.78524363e-01
-7.70725608e-01 -2.89734453e-01 5.29653490e-01 8.69720340e-01
3.26743066e-01 4.52697486e-01 1.47247761e-01 -9.48849246e-02
1.15632214e-01 -1.23258674e+00 -8.92645493e-02 -6.76292837e-01
-2.97320664e-01 -6.36683464e-01 4.02668118e-01 -5.54949164e-01
1.00206590e+00 -1.95951080e+00 1.89753383e-01 -4.04864289e-02
-2.59340852e-02 1.72463402e-01 -6.77766511e-03 1.20248997e+00
7.70311877e-02 -1.81950152e-01 2.27703340e-02 5.74615039e-02
1.43751591e-01 1.44498542e-01 -4.53356393e-02 1.48431584e-01
1.56136900e-02 5.77460468e-01 -8.79799843e-01 -2.95684695e-01
3.03497106e-01 4.20472175e-02 -5.14343143e-01 5.93400002e-01
-3.90523553e-01 6.62733734e-01 -3.92381310e-01 3.08444768e-01
-2.36476492e-02 3.79831403e-01 5.10435402e-01 5.46610773e-01
-6.90089047e-01 1.05224204e+00 -7.74272501e-01 1.37502265e+00
-6.29087627e-01 6.47188127e-01 3.13734919e-01 -4.56068635e-01
9.83139217e-01 6.66572809e-01 3.21279727e-02 -7.49391079e-01
8.84999931e-02 5.03313206e-02 5.93278646e-01 -7.90080249e-01
7.75501966e-01 -9.37007070e-02 -3.73517632e-01 8.87585282e-01
-3.50614756e-01 -6.27025485e-01 2.47421861e-01 2.42981315e-01
1.13868070e+00 1.76198870e-01 5.52513182e-01 -2.97585398e-01
5.27649283e-01 3.48321676e-01 2.04890981e-01 7.08519101e-01
-3.96555811e-01 9.43402350e-02 7.33991563e-01 -5.14042675e-01
-1.03640962e+00 -5.03667414e-01 2.96069235e-01 1.60997915e+00
-6.88282624e-02 -1.31667912e-01 -1.09755504e+00 -2.46490076e-01
-6.65380299e-01 1.44347823e+00 -2.83809930e-01 1.98777020e-02
-8.88668776e-01 -3.52833979e-02 6.00646019e-01 3.34502101e-01
3.12711090e-01 -1.85159731e+00 -1.13514256e+00 5.78092754e-01
-6.50807440e-01 -8.93923521e-01 -5.30348718e-01 2.88453609e-01
-4.76330280e-01 -8.64587486e-01 -1.29698947e-01 -8.90434027e-01
3.98579001e-01 2.54804641e-01 8.91372561e-01 1.67666271e-01
2.61253834e-01 5.67743421e-01 -7.67000973e-01 -3.84800881e-01
-1.21257770e+00 -1.01510389e-02 7.28149116e-02 -5.22321999e-01
3.80890518e-01 -3.56443763e-01 -2.06991136e-01 5.42770922e-01
-4.19358969e-01 4.35538322e-01 3.42240781e-01 7.68421769e-01
-3.26920748e-01 -4.04052526e-01 6.34111643e-01 -7.27064729e-01
1.42271471e+00 -1.61516428e-01 -2.54122257e-01 2.28674158e-01
-4.67028916e-01 1.76399946e-01 6.00426793e-01 -2.48173133e-01
-1.41150963e+00 1.83952466e-01 -1.92210555e-01 5.30004382e-01
-4.46110010e-01 4.00818139e-01 -9.67620015e-02 1.93462074e-01
9.94404852e-01 2.47738987e-01 5.67593336e-01 -1.20163247e-01
3.84489506e-01 1.10539401e+00 3.50111842e-01 -8.47134888e-01
-5.37259690e-02 -2.90819257e-01 -6.30362451e-01 -1.04076016e+00
-1.79619163e-01 -4.04675066e-01 -7.92915106e-01 -5.63705504e-01
5.53089440e-01 -6.22767210e-01 -9.83457625e-01 4.29305822e-01
-1.38826323e+00 -1.02143502e+00 8.55956152e-02 3.78006309e-01
-1.01861572e+00 2.85168827e-01 -7.97480047e-01 -1.12698376e+00
-1.49641037e-01 -1.12342334e+00 6.50844336e-01 3.28410894e-01
-8.74130905e-01 -7.65636384e-01 1.31833211e-01 1.67137608e-01
2.99615860e-01 -2.40916073e-01 8.64625812e-01 -1.13967228e+00
-2.27006525e-02 -7.80731067e-02 1.75722122e-01 1.03096917e-01
1.49011046e-01 7.70581067e-02 -4.61554557e-01 8.30715969e-02
9.66102183e-02 -7.10194647e-01 -1.00648612e-01 1.34165771e-03
-9.43455920e-02 -5.61105311e-01 -1.56090960e-01 -2.49664545e-01
4.92183894e-01 8.94391954e-01 6.91406369e-01 3.61346632e-01
1.13134399e-01 1.29077983e+00 1.03728521e+00 3.08424234e-01
9.51575041e-01 8.78034115e-01 5.75545710e-03 2.92769372e-01
1.91421106e-01 -1.94682568e-01 5.69196880e-01 6.77294731e-01
-1.81784749e-01 -2.18926370e-01 -1.13228703e+00 5.25799513e-01
-2.02808928e+00 -1.09064257e+00 -1.25115618e-01 1.89806831e+00
9.62567568e-01 1.66733861e-01 5.80123484e-01 1.17442785e-02
7.26654589e-01 -7.25168064e-02 -1.55847713e-01 -1.25012541e+00
5.12545884e-01 -3.26890886e-01 9.65967309e-03 9.46524084e-01
-6.70730591e-01 1.49205518e+00 6.83873558e+00 -8.20466056e-02
-1.06731784e+00 -1.54160649e-01 3.83918554e-01 2.30396658e-01
7.78193548e-02 -2.59256456e-03 -5.61794758e-01 6.82230741e-02
1.17658162e+00 -7.24914894e-02 7.79334247e-01 7.07034588e-01
6.50822401e-01 -5.13575435e-01 -1.56771302e+00 3.78478467e-01
-1.14315465e-01 -8.28078330e-01 -4.83507514e-01 -1.29754186e-01
-1.51919186e-01 -3.50911319e-01 -4.74853277e-01 5.60293078e-01
6.24884248e-01 -8.92997563e-01 9.01357234e-01 3.85887474e-01
2.20529005e-01 -7.86846280e-01 5.31887770e-01 9.59238946e-01
-6.36501253e-01 -1.30620718e-01 -2.79131606e-02 -7.01898277e-01
2.02631116e-01 -4.12649393e-01 -1.68492424e+00 1.56005561e-01
3.33102375e-01 -2.56486982e-01 -2.59202439e-02 5.22820592e-01
-4.57669497e-01 4.10407722e-01 -4.30983245e-01 -7.37196624e-01
4.09697175e-01 -1.93046972e-01 6.61468446e-01 1.17235899e+00
-1.17312998e-01 6.93392277e-01 4.51297641e-01 4.00013208e-01
5.53312361e-01 1.82779908e-01 -9.06529248e-01 2.02559736e-02
9.05557752e-01 9.45393205e-01 -8.29624116e-01 -1.50964856e-01
-9.90969986e-02 8.54639709e-01 4.48343784e-01 1.28033370e-01
-5.10529816e-01 -5.30300319e-01 3.38440597e-01 -2.19799489e-01
-1.48063943e-01 -3.83096188e-01 -9.24384296e-02 -4.66511041e-01
-7.16471598e-02 -1.19840252e+00 -6.13904148e-02 -8.38656783e-01
-6.37764156e-01 9.20256257e-01 6.78177550e-02 -9.50331926e-01
-1.06610584e+00 -3.83187771e-01 -5.99029422e-01 8.90374839e-01
-5.95507145e-01 -9.74226177e-01 2.66594011e-02 2.05332667e-01
8.08658719e-01 1.87870637e-02 1.19505405e+00 -1.41311541e-01
-2.63809472e-01 2.55408138e-01 -6.26185298e-01 8.56320485e-02
8.29362273e-01 -1.17077279e+00 4.26337719e-01 4.83411431e-01
-5.60238138e-02 7.40463376e-01 1.09215939e+00 -5.01189113e-01
-1.16605985e+00 -3.23922604e-01 1.14432478e+00 -4.52924848e-01
4.72364932e-01 -2.01298967e-01 -7.41078556e-01 8.87232423e-01
5.33072770e-01 -8.50138485e-01 4.54584926e-01 3.28211933e-01
-3.95228304e-02 2.93917060e-01 -1.12363958e+00 8.15181255e-01
8.10662925e-01 -5.01292527e-01 -1.35412526e+00 4.13348109e-01
6.65331185e-01 -7.00663567e-01 -5.20852804e-01 -1.32109344e-01
5.74635565e-01 -1.24285674e+00 4.34227467e-01 -5.10535777e-01
3.05285662e-01 -1.82112008e-01 4.15930077e-02 -1.45147467e+00
-3.44121307e-01 -1.11769712e+00 5.84500611e-01 1.29333067e+00
6.65865779e-01 -6.31610692e-01 6.35630637e-02 1.18570828e+00
-5.07444859e-01 -3.60534728e-01 -7.75206923e-01 -3.54746431e-01
-4.55551073e-02 -3.92795771e-01 3.69121671e-01 5.98602831e-01
1.09865582e+00 8.14294338e-01 -3.30696583e-01 -4.02851701e-02
-2.29051501e-01 -7.94138014e-02 1.03049171e+00 -9.09048915e-01
3.67120057e-02 -4.83507246e-01 5.09540737e-02 -1.37283862e+00
1.01015277e-01 -4.17409122e-01 8.35580409e-01 -1.43742216e+00
-4.90021080e-01 -4.57570463e-01 3.83181185e-01 7.61936426e-01
1.69629365e-01 -4.09820348e-01 3.54707360e-01 2.62841195e-01
-4.66302574e-01 3.42068106e-01 1.06278276e+00 2.87096471e-01
-1.00700533e+00 2.47753337e-01 -7.22437739e-01 8.86203945e-01
7.82443285e-01 -1.77094638e-01 -4.85717982e-01 -3.54357362e-02
-7.62628391e-02 8.17133725e-01 -1.15709640e-01 -8.21958721e-01
6.56638503e-01 -4.40797031e-01 -2.77375609e-01 -3.02758276e-01
2.94135451e-01 -4.75114584e-01 -1.05223604e-01 3.47056091e-01
-7.04617202e-01 1.48086444e-01 1.86232179e-01 1.35450810e-01
2.81686932e-02 -3.20299178e-01 6.06461763e-01 -3.86798322e-01
-7.86052704e-01 -7.36025691e-01 -1.25813329e+00 -1.56232804e-01
1.25673795e+00 -3.92032683e-01 -1.60185322e-01 -6.63052619e-01
-6.83645427e-01 4.41180587e-01 4.09870178e-01 5.07199883e-01
3.19668651e-01 -7.92940080e-01 -5.23650110e-01 -8.58851448e-02
1.19625948e-01 -1.93778530e-01 -1.31626114e-01 7.67671466e-01
-9.11713362e-01 7.79743552e-01 -4.27582324e-01 -5.19881427e-01
-1.35128975e+00 2.75954396e-01 4.35024470e-01 2.46193539e-02
-8.46829236e-01 7.24104524e-01 -1.41735226e-02 -8.29453528e-01
3.09480041e-01 -3.23942155e-01 -4.78348434e-01 -9.87132452e-03
7.36598253e-01 9.75914150e-02 -6.10138364e-02 -7.29272783e-01
-3.32646012e-01 -1.03741027e-01 -2.32203707e-01 -7.85053313e-01
1.02579212e+00 -4.51689869e-01 -2.03104526e-01 9.04264808e-01
3.52820277e-01 -2.82868966e-02 -1.20051062e+00 -4.60362360e-02
3.18904787e-01 -1.89679325e-01 -3.53602499e-01 -1.06919253e+00
5.92836253e-02 3.77737820e-01 7.15543842e-03 9.50032175e-01
5.94485521e-01 4.44407947e-02 3.64718527e-01 9.63260651e-01
6.95445955e-01 -1.44920528e+00 -3.86859477e-02 1.12241042e+00
1.12869501e+00 -9.51238811e-01 -2.07237139e-01 -2.09548965e-01
-1.21163988e+00 1.29349661e+00 7.80344903e-01 2.15314955e-01
6.65963516e-02 1.52519628e-01 5.37360787e-01 -2.26796970e-01
-1.20330191e+00 -9.45952162e-02 -4.21076655e-01 6.47427678e-01
8.31984282e-01 1.76481530e-01 -6.37285888e-01 6.15451336e-02
-6.09846532e-01 -2.47108266e-01 7.78277636e-01 1.19785941e+00
-7.04059541e-01 -1.04749930e+00 -4.87168223e-01 -6.98499158e-02
-1.30645484e-01 1.81269094e-01 -1.08853376e+00 9.58991885e-01
-4.73065495e-01 1.55492842e+00 -3.66373500e-03 -2.75997847e-01
6.24531031e-01 4.83471215e-01 4.90354836e-01 -9.10972834e-01
-9.22029018e-01 1.35759218e-02 8.51274312e-01 -6.08781517e-01
-2.89144427e-01 -8.79790008e-01 -1.57844043e+00 -3.72998297e-01
-2.27059111e-01 4.03009534e-01 5.26458323e-01 1.21755195e+00
1.51929989e-01 2.56224990e-01 7.28980303e-01 -1.07753909e+00
-6.85559809e-01 -1.40109658e+00 -1.73192948e-01 6.99053258e-02
1.57198578e-01 -3.97637427e-01 -2.20492125e-01 -7.78508782e-02]
|
[12.903738021850586, 7.953948020935059]
|
f21d37b1-9888-4685-acb4-2916c73ee777
|
hierarchical-dense-correlation-distillation-1
|
2306.15278
| null |
https://arxiv.org/abs/2306.15278v1
|
https://arxiv.org/pdf/2306.15278v1.pdf
|
Hierarchical Dense Correlation Distillation for Few-Shot Segmentation-Extended Abstract
|
Few-shot semantic segmentation (FSS) aims to form class-agnostic models segmenting unseen classes with only a handful of annotations. Previous methods limited to the semantic feature and prototype representation suffer from coarse segmentation granularity and train-set overfitting. In this work, we design Hierarchically Decoupled Matching Network (HDMNet) mining pixel-level support correlation based on the transformer architecture. The self-attention modules are used to assist in establishing hierarchical dense features, as a means to accomplish the cascade matching between query and support features. Moreover, we propose a matching module to reduce train-set overfitting and introduce correlation distillation leveraging semantic correspondence from coarse resolution to boost fine-grained segmentation. Our method performs decently in experiments. We achieve 50.0% mIoU on COCO dataset one-shot setting and 56.0% on five-shot segmentation, respectively. The code will be available on the project website. We hope our work can benefit broader industrial applications where novel classes with limited annotations are required to be decently identified.
|
['Jiaya Jia', 'Jingyong Su', 'Shu Liu', 'Chengyao Wang', 'Xiaoyang Wu', 'Zhuotao Tian', 'Bohao Peng']
|
2023-06-27
| null | null | null | null |
['few-shot-image-segmentation', 'semantic-correspondence']
|
['computer-vision', 'computer-vision']
|
[ 3.85127008e-01 2.42037684e-01 -3.64867300e-01 -8.08536351e-01
-1.00465453e+00 -5.12246609e-01 1.30519152e-01 -9.61361974e-02
-2.46606112e-01 3.08141649e-01 -2.29936764e-01 1.72905594e-01
3.79117355e-02 -8.07609081e-01 -7.16155469e-01 -4.05432999e-01
3.29595506e-01 4.32835639e-01 8.61036360e-01 -2.28457704e-01
1.93191394e-01 1.81630790e-01 -1.71580517e+00 5.25137365e-01
1.04837298e+00 1.31867635e+00 2.87634879e-01 4.45087820e-01
-1.95780292e-01 6.17107630e-01 -4.18071747e-01 -4.38531935e-01
2.94133872e-01 -2.19677940e-01 -1.01275444e+00 2.30842486e-01
5.77332377e-01 -2.77713209e-01 -2.08368465e-01 1.14301574e+00
4.10265088e-01 1.91225588e-01 4.98126060e-01 -1.25784779e+00
-7.44579077e-01 7.61718392e-01 -7.92570829e-01 2.81129032e-01
-3.63218412e-02 2.58701265e-01 1.18679702e+00 -9.72781420e-01
4.24852937e-01 1.15310168e+00 6.90605938e-01 7.75553405e-01
-1.14643264e+00 -7.47575521e-01 3.29956234e-01 4.29856390e-01
-1.41093552e+00 -3.88784438e-01 8.13768387e-01 -3.14132482e-01
7.73401260e-01 1.36057690e-01 5.69094479e-01 9.97735143e-01
-3.73501241e-01 1.20738971e+00 7.70114660e-01 -1.31360531e-01
2.51009732e-01 2.33376831e-01 7.79769123e-01 8.32619965e-01
1.10673569e-01 -2.00557187e-01 -5.56617677e-01 2.28532493e-01
6.98607683e-01 2.16354936e-01 1.12168148e-01 -4.25354630e-01
-7.50450015e-01 1.00103676e+00 8.64826739e-01 2.02476814e-01
8.55202302e-02 2.32257713e-02 4.48388487e-01 1.78478763e-01
2.76858330e-01 3.96173537e-01 -6.29666626e-01 5.51777072e-02
-1.14235139e+00 -7.96639472e-02 6.41436219e-01 1.47658634e+00
8.53544295e-01 -8.56860541e-03 -4.90434527e-01 1.11714756e+00
-4.66107987e-02 1.70595378e-01 5.99185348e-01 -1.04684651e+00
1.62654877e-01 9.06599283e-01 -2.88509488e-01 -6.33089542e-01
-2.74942487e-01 -7.05617547e-01 -6.41116381e-01 -2.39446104e-01
2.04401299e-01 7.50327334e-02 -1.40426934e+00 1.57987034e+00
3.81948978e-01 3.82372826e-01 -1.07059970e-01 1.05596423e+00
9.45787489e-01 4.05286044e-01 1.30140796e-01 1.30832210e-01
1.51989579e+00 -1.35910845e+00 -3.25319976e-01 -3.55675340e-01
5.97770035e-01 -6.19137943e-01 1.43306160e+00 1.00259639e-01
-1.05225098e+00 -8.18274677e-01 -1.17300904e+00 -2.95809567e-01
-4.45710808e-01 6.21950515e-02 7.41772711e-01 5.50631046e-01
-6.38274193e-01 7.81523407e-01 -9.65869427e-01 -2.87827581e-01
1.22433722e+00 4.93460566e-01 1.88859776e-01 -1.84436023e-01
-1.08290040e+00 3.19069147e-01 5.22091806e-01 -4.46704291e-02
-8.93142760e-01 -8.56995106e-01 -9.06848967e-01 2.13077664e-01
5.46142459e-01 -6.80067897e-01 1.33298433e+00 -1.03153002e+00
-1.23980653e+00 9.95494485e-01 7.44900033e-02 -5.40496767e-01
3.74591708e-01 -1.39930740e-01 -1.25972480e-01 2.59123743e-01
5.25112987e-01 1.04086459e+00 7.40358472e-01 -1.02426326e+00
-8.07520092e-01 -5.76916695e-01 -4.29733880e-02 7.16313422e-02
-4.73445684e-01 -2.11494863e-01 -6.67349517e-01 -6.47919357e-01
3.18527013e-01 -6.94545627e-01 -3.60673934e-01 2.25885585e-02
-5.55165052e-01 -3.36352140e-01 1.02238584e+00 -2.51942158e-01
9.91553783e-01 -2.14570880e+00 -1.48829848e-01 -1.25478074e-01
1.76369429e-01 3.36232603e-01 -1.91396341e-01 -1.48782849e-01
2.31322467e-01 4.70467471e-02 -5.15108466e-01 -1.54644683e-01
1.39873713e-01 2.86980927e-01 -2.16698602e-01 2.41622016e-01
6.24411643e-01 1.13117826e+00 -8.55059266e-01 -7.54957974e-01
2.35823859e-02 1.93064854e-01 -6.04831040e-01 2.94382006e-01
-2.49359697e-01 4.65098284e-02 -5.96826971e-01 1.09673488e+00
5.74911475e-01 -5.46609998e-01 -3.82300735e-01 -4.40446079e-01
2.02660277e-01 -1.60560042e-01 -9.24959660e-01 2.17745805e+00
-2.61903316e-01 1.49124071e-01 -1.83063820e-01 -1.17459857e+00
9.43147719e-01 -6.38795346e-02 2.88139254e-01 -6.83005929e-01
4.16118890e-01 1.96492180e-01 -7.87678286e-02 -3.54137450e-01
2.48894811e-01 -4.52312194e-02 -3.20218980e-01 3.81265692e-02
5.09837985e-01 -8.14404935e-02 3.41574550e-01 3.16003501e-01
1.06487441e+00 1.46773979e-01 9.80657190e-02 -3.88830066e-01
3.23511451e-01 2.77852863e-01 9.72266614e-01 7.90721953e-01
-4.42646354e-01 8.51075053e-01 2.15992883e-01 -3.67153972e-01
-9.20055449e-01 -1.11925817e+00 -3.17007780e-01 1.35048544e+00
6.38737679e-01 -1.29922420e-01 -1.00594389e+00 -8.19836915e-01
-2.10723147e-01 5.26307106e-01 -5.97562730e-01 -4.69641387e-01
-3.40065002e-01 -8.08657050e-01 4.67196435e-01 1.09810710e+00
7.63032138e-01 -8.26573789e-01 -5.72345316e-01 1.69563755e-01
3.36865634e-02 -1.17690969e+00 -5.76651037e-01 3.52853686e-01
-8.70933771e-01 -1.10976791e+00 -7.88375854e-01 -1.15890884e+00
5.22941351e-01 2.61843771e-01 9.73861814e-01 -1.01616099e-01
-7.60776460e-01 4.53632176e-02 -3.82467359e-01 -1.89693213e-01
1.02019772e-01 3.55341911e-01 -2.94161737e-01 1.61887258e-02
7.79386461e-01 -6.85909152e-01 -7.78075039e-01 5.20446002e-01
-6.62155211e-01 1.16024174e-01 5.08503616e-01 9.28160369e-01
8.12772572e-01 -3.04918170e-01 7.77492404e-01 -1.27497482e+00
1.02016423e-02 -4.69447702e-01 -5.89063048e-01 2.49697164e-01
-5.93409657e-01 -1.61258385e-01 6.40799284e-01 -4.62296784e-01
-1.07322872e+00 4.66059834e-01 -3.18783745e-02 -7.87577629e-01
-3.88767481e-01 -1.51129261e-01 -2.89570719e-01 -5.61212488e-02
6.53303802e-01 9.84007716e-02 -5.82696199e-01 -5.71718276e-01
5.42317927e-01 6.10510170e-01 7.35780060e-01 -6.94148779e-01
6.85433507e-01 4.56423223e-01 -3.21933895e-01 -5.30598581e-01
-1.43194771e+00 -8.89332354e-01 -8.94738972e-01 1.26297340e-01
9.61803317e-01 -1.15538239e+00 -4.08250779e-01 3.11091125e-01
-9.15510535e-01 -1.24097519e-01 -6.05889022e-01 3.42747057e-03
-6.97226524e-01 5.60698844e-02 -8.84272218e-01 -4.82634723e-01
-4.86816436e-01 -9.93167043e-01 1.33032537e+00 6.16573811e-01
-6.56237602e-02 -4.88508582e-01 -1.83215812e-01 7.56158233e-01
1.74002513e-01 1.17893860e-01 6.34365380e-01 -9.51174319e-01
-7.07112491e-01 -2.61412978e-01 -5.70871055e-01 3.01597536e-01
-1.37458473e-01 -1.83854312e-01 -1.44293761e+00 -1.90274924e-01
-9.70247164e-02 -7.00661182e-01 1.24674988e+00 3.21604431e-01
1.44646466e+00 -7.63933435e-02 -3.15158039e-01 7.46691346e-01
1.35382891e+00 1.07890464e-01 3.16675812e-01 -1.04157133e-02
1.04268062e+00 4.77653772e-01 8.59128356e-01 3.11984390e-01
2.12488383e-01 4.59571064e-01 2.09861264e-01 -2.39856407e-01
-3.84766817e-01 -3.30091804e-01 -1.31132051e-01 7.61653960e-01
2.63942331e-01 -2.39364803e-02 -7.68644094e-01 7.52564430e-01
-1.94740832e+00 -7.20594823e-01 8.82289838e-03 1.74619579e+00
8.56847405e-01 4.44845915e-01 1.72382340e-01 8.01381096e-02
9.96618450e-01 -6.09728880e-02 -9.39267516e-01 -1.05651259e-01
3.11556570e-02 2.71953732e-01 4.13558960e-01 1.92293391e-01
-1.29914486e+00 1.43835020e+00 4.85048723e+00 1.20755672e+00
-6.85446501e-01 4.01716530e-01 9.39528644e-01 -1.40254036e-01
-8.30729231e-02 -4.12181765e-03 -1.01638317e+00 4.27098662e-01
5.87693632e-01 2.44260002e-02 -4.53694910e-02 1.14537942e+00
-1.91248685e-01 1.96748227e-01 -1.09991384e+00 9.56850946e-01
-1.04696833e-01 -1.35010469e+00 3.31178829e-02 -3.55375648e-01
8.89683008e-01 1.07808501e-01 -5.61538432e-03 5.10583699e-01
2.85158277e-01 -7.70274222e-01 5.97721756e-01 3.02745193e-01
8.08884919e-01 -8.06284010e-01 6.25487864e-01 2.66395599e-01
-1.33740711e+00 -1.25680611e-01 -6.08413160e-01 1.37537718e-01
1.25216141e-01 5.32336652e-01 -6.09750748e-01 2.73289323e-01
8.29201102e-01 8.92327547e-01 -5.48443496e-01 1.02017760e+00
-3.37553732e-02 7.69286156e-01 -3.09846371e-01 1.03326410e-01
4.13799167e-01 1.62574276e-01 2.07907364e-01 1.17279232e+00
-1.27680555e-01 2.71283180e-01 4.55762327e-01 1.10096788e+00
-2.24082708e-01 -4.75516021e-02 -1.58351824e-01 1.67961940e-02
5.23181140e-01 1.42067456e+00 -1.19639170e+00 -4.40633655e-01
-4.26960528e-01 1.29291606e+00 3.88594896e-01 2.04886913e-01
-9.41492856e-01 -6.71985745e-01 4.92390960e-01 1.22718655e-01
5.27453005e-01 3.22070390e-01 -5.32215059e-01 -1.27690589e+00
-2.07507282e-01 -4.70652491e-01 5.80545247e-01 -4.46862310e-01
-1.56239808e+00 5.23955464e-01 -2.41545200e-01 -1.03998089e+00
1.74906597e-01 -4.21452314e-01 -7.12583721e-01 4.36868370e-01
-1.38406157e+00 -1.49866927e+00 -4.28061754e-01 5.50801158e-01
1.12731433e+00 -9.70206633e-02 6.67914987e-01 5.00303209e-01
-8.56473088e-01 8.52388799e-01 -3.24122965e-01 2.76064545e-01
4.43888932e-01 -1.17622554e+00 4.76279467e-01 7.71562994e-01
2.60720760e-01 3.05934578e-01 3.99106085e-01 -4.82051253e-01
-8.17118645e-01 -1.44001424e+00 3.84711653e-01 -3.24974179e-01
6.16174698e-01 -4.80341733e-01 -1.08498073e+00 5.61994910e-01
-2.53299087e-01 5.14930427e-01 7.00712264e-01 5.83733581e-02
-5.60339272e-01 -1.58335015e-01 -1.27147233e+00 3.07735890e-01
1.33645165e+00 -4.58876938e-01 -6.35782063e-01 2.61603177e-01
1.21996450e+00 -2.55337924e-01 -8.81094277e-01 5.25305808e-01
3.36321115e-01 -7.15896308e-01 9.08626080e-01 -7.42575645e-01
4.37563449e-01 -2.47184649e-01 -2.77117282e-01 -6.71643555e-01
-5.30480087e-01 -2.96760082e-01 2.13016104e-02 1.39047933e+00
5.25020957e-01 -1.13184117e-01 1.18986046e+00 5.27153373e-01
-4.35639262e-01 -8.89048517e-01 -7.91954458e-01 -8.43548834e-01
-6.25620643e-03 -3.14406604e-01 5.58847904e-01 9.66733575e-01
-1.91525042e-01 8.30130041e-01 -1.07527763e-01 2.49508068e-01
8.15097213e-01 4.75864261e-01 4.61244494e-01 -1.25210571e+00
-4.60741401e-01 -3.51224631e-01 -5.55711031e-01 -1.11921835e+00
1.47866383e-01 -9.35092986e-01 1.12296864e-01 -1.14100349e+00
5.41067004e-01 -5.80072284e-01 -5.42449474e-01 6.22410893e-01
-1.52868703e-01 6.57252908e-01 1.19639136e-01 7.21223652e-03
-1.10733974e+00 6.53416812e-01 1.29251552e+00 -4.01926786e-01
-1.75826892e-01 1.17480047e-01 -6.58816576e-01 9.09327924e-01
9.84990060e-01 -5.74993610e-01 -6.58525884e-01 -3.23425055e-01
-3.48986566e-01 -2.44418830e-01 3.63778204e-01 -1.14597225e+00
3.75452459e-01 -7.21049961e-03 4.59140539e-01 -6.68916464e-01
4.40838397e-01 -6.72951877e-01 -3.61362100e-01 3.93012941e-01
-4.47607696e-01 -6.05856895e-01 1.85355365e-01 7.26434946e-01
-1.91088542e-01 -3.74341190e-01 1.08800876e+00 -2.92041123e-01
-1.17919326e+00 6.53571665e-01 2.13412493e-01 4.74977821e-01
1.15641809e+00 -4.61868644e-01 -1.90528989e-01 2.05136627e-01
-9.82279420e-01 6.10741317e-01 3.32948655e-01 3.69633794e-01
5.88057995e-01 -1.22931147e+00 -3.61274749e-01 2.11984023e-01
3.67080569e-01 4.06185687e-01 4.65530217e-01 4.96046573e-01
-1.96948409e-01 2.31671214e-01 -2.03106865e-01 -7.42711902e-01
-1.11207855e+00 5.83298266e-01 2.47935578e-01 8.86751786e-02
-6.28727198e-01 1.45766759e+00 4.58965093e-01 -3.86031747e-01
3.53518575e-01 -1.45122960e-01 -2.01441776e-02 7.61751086e-02
4.73777592e-01 4.85300571e-01 -1.26513124e-01 -4.71008241e-01
-3.74518931e-01 7.58719325e-01 -5.01046300e-01 4.46485966e-01
1.26712286e+00 -9.95691270e-02 3.12765896e-01 4.21098888e-01
1.41354561e+00 -6.07724369e-01 -1.63818586e+00 -3.56576055e-01
8.31032246e-02 -2.95696706e-01 7.71623030e-02 -8.02852929e-01
-1.30568004e+00 1.10100067e+00 7.95772910e-01 -3.43353786e-02
1.12333584e+00 3.21145117e-01 1.27965689e+00 3.24877948e-01
2.22223550e-01 -1.26072931e+00 1.73101515e-01 2.95246869e-01
3.09031487e-01 -1.54052401e+00 -4.11376566e-01 -7.99485624e-01
-6.53804123e-01 8.80665958e-01 1.08265173e+00 -2.84430802e-01
5.30536056e-01 2.96246737e-01 -1.52759835e-01 -3.75696003e-01
-5.97516060e-01 -5.53652644e-01 3.47635239e-01 6.06953621e-01
1.91298038e-01 9.08707455e-02 2.64797080e-02 1.23083448e+00
-4.06873450e-02 2.20361836e-02 2.08864883e-01 8.13399136e-01
-8.98492217e-01 -7.29303837e-01 9.53403786e-02 6.03070021e-01
-3.18160355e-01 -1.95440665e-01 -3.62285763e-01 4.72352415e-01
3.58442128e-01 8.64622056e-01 2.60695070e-01 -4.17206734e-01
4.69536215e-01 -2.41314508e-02 3.58449668e-01 -9.92831826e-01
-4.75041389e-01 6.76260591e-02 -5.86515926e-02 -8.20580721e-01
-1.27457052e-01 -3.14898461e-01 -1.53186345e+00 2.42213700e-02
-5.91403306e-01 1.83165725e-02 1.39922678e-01 8.02281499e-01
3.63440335e-01 6.49683118e-01 5.31390309e-01 -5.79870701e-01
-5.40626168e-01 -7.06805468e-01 -6.76436961e-01 4.05915082e-01
2.33894680e-02 -5.36605477e-01 -1.34206146e-01 8.13288391e-02]
|
[9.600048065185547, 1.0830641984939575]
|
51e4fed9-986f-465b-9b12-63ba13d89567
|
p-meta-towards-on-device-deep-model
|
2206.12705
| null |
https://arxiv.org/abs/2206.12705v1
|
https://arxiv.org/pdf/2206.12705v1.pdf
|
p-Meta: Towards On-device Deep Model Adaptation
|
Data collected by IoT devices are often private and have a large diversity across users. Therefore, learning requires pre-training a model with available representative data samples, deploying the pre-trained model on IoT devices, and adapting the deployed model on the device with local data. Such an on-device adaption for deep learning empowered applications demands data and memory efficiency. However, existing gradient-based meta learning schemes fail to support memory-efficient adaptation. To this end, we propose p-Meta, a new meta learning method that enforces structure-wise partial parameter updates while ensuring fast generalization to unseen tasks. Evaluations on few-shot image classification and reinforcement learning tasks show that p-Meta not only improves the accuracy but also substantially reduces the peak dynamic memory by a factor of 2.5 on average compared to state-of-the-art few-shot adaptation methods.
|
['Lothar Thiele', 'Yongxin Tong', 'Zimu Zhou', 'Zhongnan Qu']
|
2022-06-25
| null | null | null | null |
['few-shot-image-classification']
|
['computer-vision']
|
[ 2.84834020e-02 -1.75437316e-01 -4.84009475e-01 -4.78428066e-01
-7.46043622e-01 -6.72556832e-02 2.65055865e-01 8.88142064e-02
-4.94219661e-01 6.32075429e-01 5.89151494e-02 1.02602072e-01
-3.45056579e-02 -8.75673711e-01 -6.75725043e-01 -6.08107507e-01
2.99672902e-01 5.44898510e-01 3.32462758e-01 5.42345271e-02
-1.08366206e-01 1.35517105e-01 -1.91678715e+00 3.01542014e-01
7.79931307e-01 1.36234033e+00 2.60397553e-01 6.27312124e-01
-2.64334977e-01 4.92480248e-01 -5.28981388e-01 -3.76326412e-01
1.49218619e-01 -3.85994643e-01 -4.53889728e-01 6.25934303e-02
1.77668169e-01 -5.13483286e-01 -2.88599432e-01 9.17964816e-01
8.73810768e-01 6.06718063e-01 4.46564376e-01 -1.13893604e+00
-7.50805438e-01 4.63510185e-01 -2.34888613e-01 5.32888412e-01
-3.64034809e-02 2.44677559e-01 5.19410968e-01 -8.48511755e-01
4.04719502e-01 8.34078431e-01 7.17826307e-01 1.13253319e+00
-9.79962826e-01 -7.73306012e-01 2.15100080e-01 4.11738306e-01
-1.12092936e+00 -6.47038281e-01 7.42499590e-01 8.77599139e-03
1.21396232e+00 -5.24653308e-02 7.78331876e-01 1.45774412e+00
8.07545930e-02 6.22241080e-01 6.38111770e-01 -1.53426498e-01
1.04191077e+00 1.75784424e-01 1.09398104e-01 5.67754805e-01
2.93994188e-01 -3.00650001e-01 -7.49504328e-01 -2.25908414e-01
1.61189511e-01 6.90218627e-01 2.30867818e-01 -3.84656489e-01
-4.48872954e-01 7.21525371e-01 2.75056899e-01 2.32844234e-01
-7.12666988e-01 2.40713179e-01 7.13367999e-01 1.57179505e-01
6.17499232e-01 2.11370569e-02 -7.07075715e-01 -4.95502263e-01
-8.12613428e-01 -3.09017360e-01 6.71151578e-01 9.61538911e-01
7.79943407e-01 2.81834900e-01 -1.25338033e-01 9.52840328e-01
-2.18159333e-02 5.89684844e-01 1.14742649e+00 -8.46720636e-01
4.48971510e-01 5.65596402e-01 1.03080995e-01 -3.63171607e-01
-3.32154393e-01 -3.25756192e-01 -9.09537673e-01 -1.33659542e-01
-2.45705292e-01 -3.27465713e-01 -9.72807646e-01 1.69158471e+00
6.87405169e-01 5.67018986e-01 2.09464338e-02 3.84735465e-01
5.82101107e-01 7.07358718e-01 5.57164252e-01 -4.48800296e-01
9.70061123e-01 -1.07934356e+00 -5.21100402e-01 -2.97500074e-01
4.85547751e-01 -1.06867768e-01 1.54092026e+00 2.25319251e-01
-1.03983223e+00 -7.86954045e-01 -1.27071583e+00 1.19150631e-01
-5.77120662e-01 -2.91271150e-01 4.26406950e-01 9.61946368e-01
-6.23040855e-01 7.22125113e-01 -1.14883292e+00 -5.73916733e-01
1.01385736e+00 4.66501206e-01 1.04038343e-01 -1.93197951e-01
-6.94435358e-01 4.73650724e-01 3.95550072e-01 -5.23177505e-01
-1.06516171e+00 -1.01334918e+00 -5.24982572e-01 3.09294164e-01
3.61339957e-01 -9.40483451e-01 1.47125590e+00 -8.88376474e-01
-2.01252079e+00 4.59301561e-01 -1.45316303e-01 -6.79376125e-01
4.30792004e-01 -4.17665303e-01 -8.31078172e-01 -9.74009261e-02
-2.56461501e-01 5.34507036e-01 1.41838896e+00 -8.71403873e-01
-8.24165285e-01 -4.82025325e-01 -1.37975458e-02 9.40676033e-02
-1.34549606e+00 -4.78244394e-01 -3.92931104e-01 -2.26667896e-01
-3.58581007e-01 -7.86401153e-01 -1.69394612e-01 6.86696544e-02
1.46826714e-01 -1.74734026e-01 1.20758545e+00 -3.91118824e-01
1.32192516e+00 -2.16378975e+00 -1.78802848e-01 -1.52186468e-01
9.01271105e-02 6.97741151e-01 -2.60730833e-01 1.64455086e-01
4.81624484e-01 -1.43781453e-01 -3.48844863e-02 -7.20420122e-01
1.70314983e-01 5.78841686e-01 -2.64021069e-01 1.52423337e-01
-3.11901867e-01 1.14545119e+00 -1.11139822e+00 -2.32513219e-01
4.39486951e-01 6.08316600e-01 -6.95676982e-01 1.85459375e-01
-2.86844701e-01 2.58111715e-01 -4.93770689e-01 6.50185287e-01
4.32181895e-01 -5.52614272e-01 2.49902382e-01 -1.38928190e-01
3.50088805e-01 6.76145330e-02 -8.05424392e-01 1.97897446e+00
-1.04184043e+00 1.17170610e-01 -4.22857136e-01 -9.01872933e-01
7.61958241e-01 2.07321256e-01 7.02581108e-01 -1.01258695e+00
2.08752111e-01 1.14045583e-01 -5.10696828e-01 -5.72844744e-01
3.77108276e-01 1.15360253e-01 -1.26452655e-01 5.95583856e-01
3.03808361e-01 3.25288355e-01 -6.19554259e-02 -1.89486146e-01
1.23816419e+00 -6.20110817e-02 5.41623473e-01 1.79878902e-02
1.89221606e-01 -4.24730539e-01 7.41298378e-01 9.29327309e-01
-3.95023644e-01 1.90611035e-01 -3.92566383e-01 -8.83981884e-01
-1.09957433e+00 -9.41692472e-01 1.31685659e-01 1.64986193e+00
-5.11864908e-02 -4.70249474e-01 -9.20473933e-01 -1.02538466e+00
4.11555395e-02 9.89273012e-01 -7.00472772e-01 -6.52242601e-01
-3.00059021e-01 -8.48161995e-01 2.99211033e-02 8.02149415e-01
7.77157366e-01 -1.07644963e+00 -1.32633281e+00 4.89764094e-01
2.11312696e-01 -1.01965916e+00 -4.82507348e-01 2.50373185e-01
-1.14895773e+00 -7.43299007e-01 -5.03166258e-01 -5.79162836e-01
6.13929927e-01 3.13673049e-01 1.00379741e+00 -2.67369777e-01
-2.26003498e-01 6.29257023e-01 -4.52195674e-01 -6.16884053e-01
-2.17512414e-01 4.97764498e-01 4.27369118e-01 5.66298887e-02
7.22895503e-01 -9.17282283e-01 -6.85006738e-01 1.50087759e-01
-8.52194309e-01 -2.91895300e-01 3.52718085e-01 8.95113170e-01
7.47039616e-01 1.37045205e-01 8.42956901e-01 -9.42677617e-01
4.53208417e-01 -6.45206094e-01 -4.55165774e-01 4.41159695e-01
-9.39873755e-01 -6.11159019e-02 8.84625971e-01 -1.04258966e+00
-1.21028686e+00 4.94430959e-02 -7.09386216e-03 -7.90009856e-01
1.55984620e-02 4.87673990e-02 -3.09326917e-01 9.49402899e-02
9.19677794e-01 2.81783879e-01 -5.09277403e-01 -6.31381094e-01
4.89950716e-01 9.20041740e-01 3.02366734e-01 -4.75833595e-01
5.38335860e-01 6.06859028e-01 -2.18137503e-01 -8.28134358e-01
-1.05263472e+00 -3.32648993e-01 -5.46771646e-01 2.70368438e-02
6.66841567e-01 -8.62555623e-01 -5.74398935e-01 3.14878225e-01
-6.12679780e-01 -8.89178395e-01 -1.01271808e+00 1.98760048e-01
-6.03103638e-01 -8.46245065e-02 -2.36006647e-01 -8.38261306e-01
-8.90265822e-01 -7.24600315e-01 8.65164340e-01 3.57579231e-01
-1.26530871e-01 -9.44192529e-01 1.64236009e-01 1.64021164e-01
8.08664918e-01 -1.26870856e-01 8.47035944e-01 -6.92067623e-01
-1.67723432e-01 -2.14396402e-01 9.25557613e-02 2.28880286e-01
3.91109675e-01 -5.49590409e-01 -1.26002896e+00 -5.36038160e-01
1.09057158e-01 -5.14041305e-01 6.59864724e-01 2.65123039e-01
1.60600436e+00 -5.61836123e-01 -2.85579532e-01 6.77644849e-01
1.32855761e+00 2.54327714e-01 2.09882483e-01 1.19712397e-01
5.59136331e-01 -1.24246748e-02 4.87581015e-01 1.03871119e+00
2.32510000e-01 6.41783357e-01 2.95092434e-01 4.69795138e-01
-2.25650743e-01 -4.93651152e-01 2.78329670e-01 9.46751773e-01
1.94169909e-01 -8.89648944e-02 -6.01182163e-01 5.48649430e-01
-2.22387362e+00 -9.15844500e-01 8.32408786e-01 2.28231335e+00
6.25853181e-01 2.56009728e-01 3.25685352e-01 -2.53710877e-02
5.00077128e-01 1.52783558e-01 -1.47500968e+00 -3.15972716e-01
3.30678374e-01 4.14399475e-01 3.73256475e-01 4.00561653e-03
-9.22619700e-01 8.69883060e-01 6.32801437e+00 8.69652927e-01
-1.16723037e+00 9.68336999e-01 5.89046776e-01 -7.49333143e-01
-1.33829951e-01 -3.79296839e-01 -8.75521123e-01 7.90200114e-01
1.49199927e+00 -2.31387258e-01 7.32493520e-01 1.51163423e+00
-6.02586493e-02 3.74022156e-01 -9.21301842e-01 1.42128980e+00
1.15224928e-01 -1.28588879e+00 7.42387474e-02 -1.33254258e-02
1.21337354e+00 4.07972783e-01 2.39113435e-01 7.83780575e-01
2.60973454e-01 -4.50366855e-01 4.80065495e-01 4.95022893e-01
6.94312751e-01 -9.09299552e-01 4.03802603e-01 4.16427493e-01
-1.25400984e+00 -7.42401481e-01 -8.43288362e-01 7.05664530e-02
9.42582861e-02 6.26516283e-01 -5.08852541e-01 3.92817371e-02
9.75792170e-01 7.09204495e-01 -4.75650907e-01 9.09746349e-01
3.04923169e-02 7.04371929e-01 -2.22991124e-01 -1.76822990e-01
-7.24655436e-03 2.34206229e-01 2.38589272e-01 9.93053138e-01
6.91664815e-01 7.19082728e-02 2.55769074e-01 3.88203919e-01
-5.40386438e-01 1.02456369e-01 -5.87719321e-01 -4.49561924e-02
8.11140060e-01 1.20501661e+00 -4.59016860e-01 -7.47755945e-01
-5.04690826e-01 1.35316598e+00 4.68328238e-01 3.00782353e-01
-8.56810093e-01 -1.05437458e-01 7.55583763e-01 -2.64813881e-02
7.33300269e-01 1.83248203e-02 -1.46223888e-01 -1.15791667e+00
6.14829687e-03 -6.07988656e-01 5.68236887e-01 -5.90885639e-01
-1.44484842e+00 4.11243647e-01 -1.27215520e-01 -1.20356369e+00
-3.95795584e-01 -4.10459965e-01 -5.42233825e-01 1.78777561e-01
-1.22468448e+00 -1.10699153e+00 -6.20584548e-01 6.82689607e-01
9.40693736e-01 -4.22861457e-01 1.05909777e+00 3.15343827e-01
-7.44876206e-01 8.41100216e-01 2.84413189e-01 -4.55105990e-01
3.64539623e-01 -8.46101403e-01 6.12593710e-01 5.25910020e-01
2.03153968e-01 4.90540236e-01 4.49362218e-01 -5.47743440e-01
-1.46527874e+00 -1.48437607e+00 4.40928638e-01 -4.05730933e-01
4.94404048e-01 -2.77138531e-01 -9.13254797e-01 5.84526420e-01
5.82384244e-02 6.34886324e-01 9.96159613e-01 -2.82590631e-02
-4.80629742e-01 -6.92664206e-01 -1.64653885e+00 5.36172330e-01
1.45981026e+00 -6.55421078e-01 -2.81538188e-01 3.23488593e-01
7.62178898e-01 -7.61567056e-02 -9.81459022e-01 2.39580482e-01
6.65493846e-01 -6.76342845e-01 9.70660031e-01 -9.74006593e-01
-2.60177732e-01 2.14979574e-01 -5.63207626e-01 -1.25058448e+00
-4.01374608e-01 -7.63321877e-01 -1.29170895e+00 1.24090576e+00
1.34860709e-01 -5.84877729e-01 1.05118382e+00 9.95425284e-01
-5.82616478e-02 -7.79111564e-01 -1.11237860e+00 -1.05020249e+00
-2.49258533e-01 -4.52572346e-01 9.02267456e-01 7.09368110e-01
-6.31531700e-02 3.33937109e-01 -5.36914468e-01 -2.22595945e-01
7.53139019e-01 -1.32718951e-01 8.61582100e-01 -1.30051410e+00
-6.02514565e-01 -1.46975696e-01 -2.73956656e-01 -6.74575210e-01
1.90674320e-01 -5.12796700e-01 -1.71959221e-01 -1.39079356e+00
2.48889282e-01 -2.93353617e-01 -8.45135391e-01 6.97574317e-01
-8.06886777e-02 2.01793179e-01 6.67500198e-02 2.50703871e-01
-1.10372007e+00 8.52051973e-01 6.38006687e-01 -3.31424206e-01
-6.34287715e-01 2.51075536e-01 -5.76676369e-01 6.75399542e-01
1.20493209e+00 -6.43509150e-01 -9.34640586e-01 -3.09463024e-01
5.65980896e-02 -6.64684057e-01 5.97878881e-02 -1.50549591e+00
2.96239048e-01 -1.56346411e-01 5.65753996e-01 -1.71150491e-01
7.74706542e-01 -9.49421227e-01 3.85458246e-02 6.31970167e-01
-1.48841158e-01 3.06033008e-02 2.52799630e-01 9.22208130e-01
5.25865495e-01 -3.33613783e-01 8.10489416e-01 -2.47427821e-01
-9.94393408e-01 7.52188087e-01 -4.86522168e-02 1.66981295e-01
1.08147848e+00 -5.98187707e-02 -1.94678023e-01 -1.88978449e-01
-5.53877831e-01 -2.46291831e-01 4.97048289e-01 5.26523173e-01
4.97177035e-01 -1.43491316e+00 6.67504445e-02 3.18742305e-01
4.11722839e-01 -3.12142670e-01 5.58748841e-01 4.45086658e-01
2.65414864e-01 -1.36122247e-02 -2.42328405e-01 -4.23914284e-01
-1.00514543e+00 9.78522122e-01 1.19674116e-01 -1.90373972e-01
-8.24309766e-01 6.03533387e-01 -4.51066256e-01 -3.23458672e-01
3.73172402e-01 -2.34763995e-01 1.68079734e-01 2.12143008e-02
8.85193646e-01 7.64065623e-01 2.08578005e-01 -2.53683746e-01
-2.51475781e-01 4.93133932e-01 -1.94535688e-01 3.53124030e-02
1.43787169e+00 -1.74265787e-01 6.15920246e-01 8.12233925e-01
1.30474508e+00 -4.36558634e-01 -1.74624014e+00 -5.61170161e-01
-1.83590025e-01 -3.50575864e-01 1.27340779e-01 -6.12478673e-01
-1.11322284e+00 8.48533869e-01 1.18447554e+00 9.29865241e-02
1.23317349e+00 -2.48406053e-01 1.31941843e+00 8.66193950e-01
7.34298527e-01 -1.60972512e+00 4.91277456e-01 1.42839685e-01
2.22529337e-01 -1.27095139e+00 -1.24798387e-01 3.70355725e-01
-6.58580899e-01 7.74586082e-01 7.19248056e-01 2.61117425e-02
1.01498103e+00 4.59752455e-02 -1.91607147e-01 1.20451465e-01
-8.03717792e-01 -1.44361556e-01 2.31032252e-01 9.29637969e-01
-2.18124151e-01 -8.78034811e-03 2.35529989e-01 6.17902577e-01
1.91217422e-01 5.71595073e-01 -1.40953302e-01 1.08760154e+00
-7.22990215e-01 -1.09866345e+00 2.52277076e-01 6.76151335e-01
-1.85119435e-01 2.15976030e-01 1.78891897e-01 2.94700056e-01
1.88813627e-01 9.00869966e-01 1.52226657e-01 -5.74777305e-01
3.35243106e-01 3.80030811e-01 3.50871623e-01 -6.71728849e-01
-5.01021922e-01 -4.04577792e-01 -3.33069623e-01 -7.54599214e-01
-2.57147849e-01 -4.61733907e-01 -1.12126040e+00 -3.05736035e-01
7.19559193e-02 -1.33517489e-01 8.46698701e-01 1.06382036e+00
9.78885174e-01 6.55837834e-01 5.87921202e-01 -1.14734483e+00
-9.39480960e-01 -9.13326502e-01 -3.21450382e-01 5.20863235e-01
1.66537896e-01 -6.77184045e-01 -1.25553951e-01 -2.87836026e-02]
|
[9.91006088256836, 3.234574794769287]
|
c3311eac-8bc3-4055-8f3f-48856b970bdc
|
agad-adversarial-generative-anomaly-detection
|
2304.04211
| null |
https://arxiv.org/abs/2304.04211v1
|
https://arxiv.org/pdf/2304.04211v1.pdf
|
AGAD: Adversarial Generative Anomaly Detection
|
Anomaly detection suffered from the lack of anomalies due to the diversity of abnormalities and the difficulties of obtaining large-scale anomaly data. Semi-supervised anomaly detection methods are often used to solely leverage normal data to detect abnormalities that deviated from the learnt normality distributions. Meanwhile, given the fact that limited anomaly data can be obtained with a minor cost in practice, some researches also investigated anomaly detection methods under supervised scenarios with limited anomaly data. In order to address the lack of abnormal data for robust anomaly detection, we propose Adversarial Generative Anomaly Detection (AGAD), a self-contrast-based anomaly detection paradigm that learns to detect anomalies by generating \textit{contextual adversarial information} from the massive normal examples. Essentially, our method generates pseudo-anomaly data for both supervised and semi-supervised anomaly detection scenarios. Extensive experiments are carried out on multiple benchmark datasets and real-world datasets, the results show significant improvement in both supervised and semi-supervised scenarios. Importantly, our approach is data-efficient that can boost up the detection accuracy with no more than 5% anomalous training data.
|
['Ni Zhang', 'Jian Shi']
|
2023-04-09
| null | null | null | null |
['supervised-anomaly-detection', 'semi-supervised-anomaly-detection']
|
['computer-vision', 'computer-vision']
|
[ 3.99900466e-01 -8.96335989e-02 4.21669662e-01 -3.53506565e-01
-8.48284721e-01 -4.73363072e-01 6.19056344e-01 3.32877964e-01
1.07276723e-01 4.53487515e-01 -1.15993097e-01 -2.63797373e-01
-4.70462069e-02 -8.07794988e-01 -5.80600798e-01 -7.27350354e-01
-2.40020201e-01 3.86208266e-01 1.99857429e-01 -2.25357711e-01
2.22934395e-01 4.45204198e-01 -1.69002569e+00 1.06301546e-01
1.31671166e+00 1.16871262e+00 -6.99328721e-01 5.50556183e-01
-4.03764129e-01 5.24196923e-01 -7.09339559e-01 -9.24059004e-02
6.03616953e-01 -7.00351775e-01 -3.15564781e-01 2.62311369e-01
3.65198702e-01 -3.89340192e-01 -3.25866252e-01 1.16977262e+00
3.74904126e-01 2.70943977e-02 6.76942825e-01 -1.75696409e+00
-6.24992907e-01 4.21644002e-02 -8.15878928e-01 7.36187220e-01
4.33651656e-01 3.31503540e-01 8.76399100e-01 -8.96534204e-01
8.18566084e-02 9.68280196e-01 4.55032527e-01 6.44111514e-01
-1.02206361e+00 -5.10223269e-01 3.02883804e-01 1.59317628e-02
-1.04487658e+00 -1.83665812e-01 1.12502038e+00 -2.67718911e-01
5.85145295e-01 4.58448380e-01 4.52601939e-01 1.44573832e+00
1.75607100e-01 7.25094318e-01 9.24517214e-01 -2.49212295e-01
3.85642409e-01 -3.27972978e-01 -2.68226415e-02 4.49160397e-01
5.63685119e-01 1.79681510e-01 -2.44999066e-01 -6.30827963e-01
5.55509746e-01 5.94066620e-01 -2.18143631e-02 -2.56174177e-01
-8.65308523e-01 7.68962145e-01 1.57654881e-01 2.53316700e-01
-5.26105464e-01 -4.33636576e-01 7.60332346e-01 6.43563628e-01
6.24958754e-01 3.23421299e-01 -2.28580892e-01 -1.60558268e-01
-5.15514612e-01 3.08323473e-01 4.95789140e-01 6.97222769e-01
3.68182123e-01 7.83731699e-01 -2.17077196e-01 7.67572820e-01
2.54036695e-01 6.09800935e-01 7.74364889e-01 -3.25702608e-01
4.49521035e-01 1.01251042e+00 -2.11527646e-01 -1.08701789e+00
-8.58505517e-02 -4.52921361e-01 -1.12119222e+00 2.12667197e-01
5.24591565e-01 -2.06445634e-01 -1.09090436e+00 1.57725799e+00
3.90478462e-01 5.88838577e-01 2.20190853e-01 6.10733509e-01
3.82305652e-01 3.13450187e-01 -7.17831403e-02 -1.09840721e-01
8.59699845e-01 -6.05418980e-01 -5.61037123e-01 -2.96636254e-01
7.30039060e-01 -5.89741468e-01 1.25834179e+00 3.81029516e-01
-6.96268022e-01 -2.35853955e-01 -1.00183916e+00 7.09476650e-01
-4.10891742e-01 -5.23804069e-01 5.63246965e-01 6.31335795e-01
-4.88157570e-01 4.00806785e-01 -1.04316437e+00 -1.70783758e-01
7.36090660e-01 -1.51150361e-01 -5.24636686e-01 -1.96662501e-01
-8.84893596e-01 2.45186135e-01 6.10859573e-01 4.24012728e-02
-1.00717652e+00 -6.47791564e-01 -9.96978045e-01 -1.37148172e-01
4.59313542e-01 -1.16589934e-01 8.99492264e-01 -1.15403259e+00
-7.68931925e-01 7.08713770e-01 1.00677371e-01 -6.75394595e-01
6.57459319e-01 -3.83758903e-01 -1.02403021e+00 -1.09812114e-02
2.99610812e-02 -1.92161381e-01 1.00938487e+00 -1.08467495e+00
-6.13846481e-01 -5.62494278e-01 -3.80063981e-01 -1.38532445e-01
-3.62807661e-01 -2.49223471e-01 3.74945141e-02 -1.10508180e+00
6.17026269e-01 -8.41839314e-01 -3.23068410e-01 -2.40951210e-01
-7.31029451e-01 3.83310616e-02 1.46312797e+00 -5.20514965e-01
1.04606354e+00 -2.36339307e+00 -4.82808143e-01 6.36872709e-01
2.39784062e-01 3.56198877e-01 -1.03238501e-01 2.82293499e-01
-3.09793144e-01 -6.71253577e-02 -9.24186647e-01 -2.28013322e-02
-7.15677217e-02 5.34461915e-01 -9.07469034e-01 3.58436555e-01
5.79788983e-01 7.63928950e-01 -9.99145746e-01 -2.62634218e-01
1.28020406e-01 -2.03650836e-02 -6.04659677e-01 6.29984796e-01
-1.98371649e-01 8.11150789e-01 -6.64903343e-01 1.18966341e+00
6.72634721e-01 1.65302187e-01 -3.48627180e-01 5.58715343e-01
6.01373553e-01 -3.10931206e-01 -1.25100768e+00 1.26994681e+00
-2.84728110e-02 2.96935976e-01 -3.31306607e-01 -1.32768273e+00
1.04050779e+00 2.19873652e-01 6.13890886e-01 -6.71091974e-01
-2.12814823e-01 5.44311225e-01 9.47559923e-02 -5.14813960e-01
7.58695304e-02 4.27332260e-02 -1.56515792e-01 6.46085560e-01
-9.16012004e-03 1.84078142e-01 6.66240603e-02 7.48349875e-02
1.58194613e+00 -1.17895342e-01 2.34796479e-01 1.96073622e-01
6.72436237e-01 -1.92717806e-01 8.48278522e-01 8.61884713e-01
-3.71460736e-01 8.46118510e-01 6.88041627e-01 -6.55814528e-01
-8.87320101e-01 -1.49484646e+00 -6.71698079e-02 8.13346088e-01
-1.17548242e-01 -1.49323553e-01 -5.85932791e-01 -1.30327225e+00
-5.28292125e-03 6.79625928e-01 -5.80541134e-01 -6.25814795e-01
-4.34620887e-01 -1.16645372e+00 8.55225325e-01 6.72838032e-01
5.71212828e-01 -1.27459550e+00 -3.32357019e-01 1.43600836e-01
1.16263114e-01 -9.77603257e-01 -2.22987071e-01 4.11112942e-02
-9.73245442e-01 -1.39559221e+00 -3.91516984e-01 -4.74544644e-01
1.19136167e+00 -1.02941610e-01 1.02343667e+00 2.67132968e-01
-6.12443745e-01 4.29801673e-01 -5.73417902e-01 -7.69787788e-01
-6.16982877e-01 -4.55691606e-01 3.17446768e-01 4.67345208e-01
7.03625619e-01 -8.62092853e-01 -4.14959490e-01 2.71241903e-01
-1.41559362e+00 -7.44390070e-01 6.67191446e-01 9.02937353e-01
7.91740537e-01 8.87454078e-02 1.03807878e+00 -1.14604783e+00
6.44241095e-01 -9.57290590e-01 -4.54231381e-01 -1.37776777e-01
-8.31472874e-01 -2.15790290e-02 8.39676142e-01 -3.77377093e-01
-1.00404298e+00 -2.99041748e-01 -2.44968712e-01 -6.21378005e-01
-7.66731203e-01 1.62982643e-01 -1.59913778e-01 1.59359530e-01
7.99750507e-01 6.40949428e-01 9.70982760e-02 -3.67757171e-01
-6.21581711e-02 4.40054268e-01 7.42452383e-01 -6.11319780e-01
1.31628227e+00 5.26807070e-01 1.50629923e-01 -5.55971026e-01
-8.60397398e-01 -3.81222874e-01 -5.16364455e-01 8.52739066e-02
3.71890128e-01 -6.53180540e-01 1.02682613e-01 8.71808767e-01
-4.90887284e-01 9.11212713e-02 -7.14285314e-01 1.70798644e-01
-2.80424148e-01 7.03014433e-01 -2.88348794e-01 -8.69424880e-01
-5.13309121e-01 -7.19562590e-01 8.00330520e-01 1.84625201e-02
-1.67513676e-02 -1.04908001e+00 2.43970275e-01 -8.44113082e-02
6.28961384e-01 1.04017985e+00 9.06046271e-01 -1.72298503e+00
-2.91765034e-01 -8.43344569e-01 3.69453579e-02 7.21372128e-01
4.85623926e-01 -5.52836694e-02 -9.69386816e-01 -3.63668501e-01
-2.83726752e-02 -2.54096091e-01 5.00715733e-01 -1.36430264e-01
1.65369070e+00 -4.18801427e-01 3.38406637e-02 4.61244941e-01
1.22177279e+00 1.66863993e-01 5.98227143e-01 2.40413994e-01
8.04716051e-01 3.00863504e-01 7.18705297e-01 5.99351883e-01
-1.89535573e-01 1.79269359e-01 6.66342914e-01 -3.49782519e-02
3.69059026e-01 -3.88323218e-01 3.12243819e-01 6.41613066e-01
7.76041225e-02 -4.94487844e-02 -1.06513882e+00 6.57990456e-01
-1.73453271e+00 -1.04114175e+00 -4.22490165e-02 2.37275100e+00
4.42717463e-01 6.56942010e-01 2.65195161e-01 6.63774610e-01
6.41744137e-01 1.36763975e-01 -8.76678169e-01 -3.63446862e-01
-2.99376130e-01 1.22715473e-01 -8.45483318e-02 -1.64416343e-01
-1.19925165e+00 4.43498701e-01 6.26255369e+00 6.75690532e-01
-9.02803898e-01 -1.24952666e-01 6.49695396e-01 -3.12591977e-02
-2.54498631e-01 -2.72073299e-01 -7.08043575e-02 8.33138108e-01
8.98625791e-01 -1.29030153e-01 -9.58830267e-02 1.11421168e+00
-1.19816802e-01 2.63498992e-01 -1.05424190e+00 8.61342609e-01
1.58932745e-01 -6.14222825e-01 4.27035332e-01 9.08026099e-02
7.84177721e-01 -9.25860628e-02 1.21616937e-01 5.85920334e-01
-2.31483411e-02 -8.94418776e-01 4.01137806e-02 5.60815513e-01
4.62524831e-01 -9.77684796e-01 1.03264225e+00 4.43066001e-01
-9.33455944e-01 -3.02553207e-01 -1.46826550e-01 2.09603403e-02
-1.28114462e-01 8.39332640e-01 -7.45168269e-01 6.16055429e-01
8.20881307e-01 5.40431857e-01 -7.93738246e-01 8.96789014e-01
-1.06005438e-01 1.10999513e+00 -3.82210165e-01 5.60003161e-01
3.17465454e-01 -4.38269407e-01 1.03395212e+00 8.74411166e-01
4.80856866e-01 -1.62898466e-01 2.88109332e-01 7.33379543e-01
1.06358305e-01 1.95171788e-01 -1.09119856e+00 -1.67842746e-01
3.83046865e-01 1.00831747e+00 -5.63643754e-01 -5.24355471e-02
-5.43810010e-01 9.70677495e-01 -1.83872089e-01 2.34218895e-01
-7.50223100e-01 -6.20789409e-01 5.91446280e-01 1.66921079e-01
-2.37876996e-02 1.64993078e-01 -2.39407957e-01 -1.22840559e+00
4.33157593e-01 -1.08236825e+00 8.70655239e-01 -1.64870769e-01
-1.74379528e+00 5.99805355e-01 -1.30509749e-01 -1.73379350e+00
-6.08865559e-01 -3.42824727e-01 -1.33311641e+00 4.92642671e-01
-1.17419338e+00 -9.75901425e-01 -5.07451475e-01 8.03441584e-01
5.57912171e-01 -7.59818971e-01 8.89816999e-01 2.03444406e-01
-6.89389229e-01 9.49953139e-01 1.94975063e-01 5.27808011e-01
5.93995214e-01 -1.61184216e+00 6.75399005e-01 1.45856512e+00
2.49253452e-01 2.72730976e-01 5.33885300e-01 -6.45379841e-01
-1.04456782e+00 -1.38296723e+00 1.08345658e-01 -6.45114839e-01
6.11836076e-01 -1.39910534e-01 -1.58654177e+00 7.99611568e-01
-2.50607103e-01 7.28921056e-01 8.86535406e-01 -1.16635241e-01
-4.18926150e-01 -7.94304162e-02 -1.59471035e+00 5.87046862e-01
1.04287159e+00 -2.64422834e-01 -8.02289009e-01 1.34965733e-01
5.13865411e-01 -5.15289187e-01 -5.91935039e-01 6.67033076e-01
-1.45531043e-01 -9.85593200e-01 8.26012015e-01 -9.57139671e-01
3.66553426e-01 -3.63735884e-01 -2.01104507e-01 -1.32385159e+00
2.16125399e-01 -4.42324698e-01 -8.19491446e-01 1.39761734e+00
1.62451580e-01 -1.17321146e+00 6.35010898e-01 4.49872941e-01
-2.91553289e-01 -9.62703228e-01 -1.07829177e+00 -8.45891953e-01
-1.85702994e-01 -6.48172319e-01 8.26745152e-01 1.00397503e+00
-4.00681198e-01 -2.78219968e-01 -2.04854622e-01 5.34692645e-01
8.41174304e-01 5.97237013e-02 8.33438635e-01 -1.38541067e+00
-1.34543195e-01 -1.10918783e-01 -1.03693306e+00 -2.47414321e-01
1.14942893e-01 -7.85163164e-01 -9.41470042e-02 -7.94741154e-01
-1.49341717e-01 -3.40769529e-01 -7.85016537e-01 6.16172135e-01
-5.87069988e-01 3.06060642e-01 -5.03005922e-01 9.17933360e-02
-5.43638468e-01 6.36968732e-01 7.70134509e-01 3.63354795e-02
-1.76087722e-01 3.08308512e-01 -7.23118722e-01 1.06019890e+00
8.44209433e-01 -4.81154919e-01 -6.06936097e-01 6.11843821e-03
-1.46846354e-01 -4.00516987e-01 4.31378424e-01 -1.25267076e+00
-1.50764242e-01 -9.83096808e-02 5.87023437e-01 -6.77895069e-01
-1.74225479e-01 -8.58337462e-01 -3.63856614e-01 3.58383715e-01
-2.07785532e-01 3.95390481e-01 1.76793411e-01 1.17004466e+00
-5.33879995e-01 1.87325291e-03 6.97372794e-01 1.52738079e-01
-8.17267716e-01 6.22185707e-01 -1.52739644e-01 6.44446373e-01
1.31325662e+00 -1.75088644e-01 -9.11914706e-02 -3.71722877e-01
-5.70532799e-01 2.65401483e-01 2.66445309e-01 6.30676389e-01
9.23713326e-01 -1.64239097e+00 -8.12447309e-01 8.45340610e-01
6.27049625e-01 4.14879382e-01 3.32813174e-01 7.07517982e-01
-3.55453968e-01 -3.72634791e-02 -3.29960227e-01 -8.20196509e-01
-8.86660337e-01 5.82020581e-01 2.81475514e-01 -2.05773517e-01
-9.66548681e-01 3.90851527e-01 -2.06392892e-02 -5.54345131e-01
9.42140594e-02 -7.70986173e-03 9.19143409e-02 -3.29871476e-01
6.41213179e-01 4.53228951e-01 2.21239641e-01 -4.96504933e-01
-2.23343477e-01 1.71792641e-01 -2.96830475e-01 2.58593976e-01
1.20799494e+00 1.41325504e-01 7.75607750e-02 4.98084754e-01
8.78226995e-01 4.08766977e-02 -1.10247946e+00 -4.85919625e-01
2.81258941e-01 -7.80494213e-01 -3.87377024e-01 -5.73668122e-01
-1.22834647e+00 5.91942072e-01 8.34562600e-01 5.15733302e-01
1.39262044e+00 -2.77347088e-01 9.49669003e-01 5.16463578e-01
8.70215893e-02 -1.00625277e+00 4.63453323e-01 3.03673863e-01
8.04147601e-01 -1.61094964e+00 -3.81715268e-01 -2.31371596e-02
-8.18176270e-01 8.62809122e-01 1.12210548e+00 -4.26578611e-01
5.41671991e-01 1.03294730e-01 2.63649106e-01 -2.91068047e-01
-4.24920231e-01 1.12078056e-01 4.29707646e-01 7.57468283e-01
2.34869644e-02 -1.56530157e-01 1.58012301e-01 6.89679027e-01
6.97985291e-02 -6.51976466e-01 4.53388721e-01 1.27616787e+00
-3.04468095e-01 -8.77651751e-01 -5.73730588e-01 1.06948841e+00
-8.09900820e-01 1.70579776e-01 -2.57390618e-01 6.62419558e-01
-5.87186143e-02 8.75279725e-01 2.84532517e-01 -1.87606826e-01
5.32983005e-01 5.44850647e-01 -1.93478420e-01 -5.13190091e-01
-7.08420649e-02 -1.45678714e-01 -4.28118497e-01 -7.26688027e-01
5.32660820e-02 -5.93619585e-01 -1.34065771e+00 2.65996568e-02
-4.27366495e-02 1.47988483e-01 1.30837709e-01 1.06358778e+00
4.70005184e-01 6.73925936e-01 1.03611708e+00 -2.52480268e-01
-8.02275419e-01 -9.60529208e-01 -7.51524329e-01 1.07030880e+00
5.55652142e-01 -5.64638495e-01 -8.43866706e-01 -2.83475190e-01]
|
[7.5938873291015625, 2.322298765182495]
|
f48514b8-86a5-43e3-ba82-fb5bc2d56d31
|
federated-deep-transfer-learning-for-eeg
|
2211.10976
| null |
https://arxiv.org/abs/2211.10976v3
|
https://arxiv.org/pdf/2211.10976v3.pdf
|
Federated deep transfer learning for EEG decoding using multiple BCI tasks
|
Deep learning has been successful in BCI decoding. However, it is very data-hungry and requires pooling data from multiple sources. EEG data from various sources decrease the decoding performance due to negative transfer. Recently, transfer learning for EEG decoding has been suggested as a remedy and become subject to recent BCI competitions (e.g. BEETL), but there are two complications in combining data from many subjects. First, privacy is not protected as highly personal brain data needs to be shared (and copied across increasingly tight information governance boundaries). Moreover, BCI data are collected from different sources and are often based on different BCI tasks, which has been thought to limit their reusability. Here, we demonstrate a federated deep transfer learning technique, the Multi-dataset Federated Separate-Common-Separate Network (MF-SCSN) based on our previous work of SCSN, which integrates privacy-preserving properties into deep transfer learning to utilise data sets with different tasks. This framework trains a BCI decoder using different source data sets obtained from different imagery tasks (e.g. some data sets with hands and feet, vs others with single hands and tongue, etc). Therefore, by introducing privacy-preserving transfer learning techniques, we unlock the reusability and scalability of existing BCI data sets. We evaluated our federated transfer learning method on the NeurIPS 2021 BEETL competition BCI task. The proposed architecture outperformed the baseline decoder by 3%. Moreover, compared with the baseline and other transfer learning algorithms, our method protects the privacy of the brain data from different data centres.
|
['A. Aldo Faisal', 'Xiaoxi Wei']
|
2022-11-20
| null | null | null | null |
['eeg-decoding', 'eeg-decoding']
|
['medical', 'time-series']
|
[ 2.17514366e-01 3.34385335e-02 1.93998754e-01 -4.30227309e-01
-9.12005067e-01 -6.33746862e-01 4.14503098e-01 -2.97541618e-01
-8.00806880e-01 1.52668178e+00 1.98440135e-01 -7.13613182e-02
-2.62547165e-01 -4.05027747e-01 -9.91220832e-01 -8.84308219e-01
-1.71352223e-01 2.08851308e-01 -1.33371532e-01 -9.43153873e-02
8.25135186e-02 3.05589318e-01 -1.15897930e+00 7.26391554e-01
1.12546062e+00 1.27972472e+00 9.28535089e-02 2.24038232e-02
3.35832745e-01 3.12506020e-01 -9.39428687e-01 -7.77269304e-01
4.06828254e-01 -3.71833324e-01 -8.26476514e-01 -7.91006088e-01
2.61465788e-01 -3.81122679e-01 -4.37675089e-01 1.14927459e+00
8.70438397e-01 -3.01079631e-01 6.89167857e-01 -1.88324988e+00
-7.43260324e-01 2.53699243e-01 -4.38833266e-01 -1.60083413e-01
1.60043210e-01 -8.50499570e-02 2.86205560e-01 -4.85621929e-01
4.04213190e-01 6.09909296e-01 5.43658853e-01 9.27507579e-01
-1.24260843e+00 -1.44801176e+00 -2.74638772e-01 6.22847557e-01
-1.39691317e+00 -4.56898272e-01 4.56697375e-01 -2.41821811e-01
8.79880071e-01 4.00694430e-01 7.35735655e-01 2.11131096e+00
5.51998377e-01 1.01180053e+00 1.45028055e+00 1.94133725e-02
3.08163822e-01 4.52542394e-01 2.91381348e-02 -1.67480960e-01
3.80956441e-01 1.50463626e-01 -9.99326169e-01 -2.98558891e-01
1.69617325e-01 3.00320350e-02 -8.54866982e-01 -5.43172061e-01
-1.10461688e+00 5.50477982e-01 4.10224676e-01 3.12351644e-01
-1.79034159e-01 -5.41025400e-02 7.93506861e-01 8.48651826e-01
4.86453384e-01 1.81924388e-01 -8.01320076e-01 -1.07065871e-01
-9.34973955e-01 1.32863164e-01 1.13670588e+00 1.42815578e+00
5.69261372e-01 -5.79742551e-01 -3.03872079e-01 5.05983531e-01
-1.51695117e-01 3.06794524e-01 7.76498258e-01 -3.34546655e-01
8.19197595e-01 1.86955228e-01 -6.97558969e-02 -7.53102958e-01
-2.80904651e-01 -1.64601982e-01 -1.31235337e+00 3.09154093e-01
2.81058758e-01 -6.15772665e-01 -7.31530666e-01 1.90352142e+00
-4.26373065e-01 1.37108132e-01 1.60013646e-01 8.52765203e-01
5.11432648e-01 1.60929695e-01 -9.55916122e-02 1.26377776e-01
1.09711361e+00 -5.97294450e-01 -9.18703973e-01 1.03205994e-01
5.51564157e-01 -9.27029476e-02 5.62947631e-01 9.09012794e-01
-8.03301394e-01 -3.18887569e-02 -1.19408095e+00 -5.69415428e-02
-8.99058878e-01 -5.51870428e-02 4.43515301e-01 1.08464658e+00
-1.32957828e+00 4.38352793e-01 -6.08742952e-01 -3.06382895e-01
1.33113909e+00 7.81499267e-01 -1.03823829e+00 -3.96879651e-02
-1.53689551e+00 1.10649800e+00 6.47734940e-01 -7.77237862e-02
-9.67428863e-01 -7.59553373e-01 -4.30222869e-01 2.05677390e-01
1.91687748e-01 -3.28150511e-01 5.32743335e-01 -1.01118195e+00
-1.54329371e+00 7.46788323e-01 2.96853364e-01 -8.15465748e-01
9.43561018e-01 -4.51321274e-01 -5.63303649e-01 -1.50021121e-01
-1.34540096e-01 6.62009478e-01 1.02026999e+00 -1.01858783e+00
-5.85661709e-01 -6.84732616e-01 -4.70842600e-01 -1.08167961e-01
-7.75749326e-01 -4.46042940e-02 -9.99276564e-02 -4.02220637e-01
-7.29414284e-01 -8.62095296e-01 5.02236366e-01 3.55577841e-02
-3.23288292e-01 7.31550343e-03 9.71047819e-01 -9.97372687e-01
8.60140204e-01 -2.45772481e+00 3.04921240e-01 2.94418156e-01
4.44584429e-01 5.29344678e-01 -1.18574098e-01 1.15664065e-01
-3.65569144e-01 1.19232997e-01 -3.71557891e-01 -1.69877470e-01
-1.96617562e-02 1.47644386e-01 -7.15932399e-02 7.31533289e-01
8.35111961e-02 9.95937884e-01 -6.47902668e-01 -1.10563874e-01
-1.31894007e-01 5.50253630e-01 -4.95941579e-01 2.00568408e-01
4.26942229e-01 9.11705434e-01 -6.28959537e-02 3.32447827e-01
1.06718719e+00 2.87888974e-01 -2.27747679e-01 -1.33364558e-01
2.16881782e-01 -5.19427620e-02 -6.67663395e-01 2.33580637e+00
-1.00577608e-01 6.47041917e-01 2.88793325e-01 -1.42625809e+00
8.13720703e-01 8.77595782e-01 6.65828943e-01 -8.34953427e-01
3.92816454e-01 3.29935044e-01 1.71476498e-01 -4.46757495e-01
-1.75255433e-01 1.06006935e-01 -1.58305943e-01 3.96690935e-01
5.58113396e-01 2.30964944e-01 -5.36697865e-01 7.24994903e-03
1.31111622e+00 1.77059352e-01 4.76626828e-02 -5.16525209e-01
4.75023508e-01 -4.70474899e-01 5.51580727e-01 6.20682240e-01
-6.67653739e-01 6.55342817e-01 4.62532103e-01 -1.10892013e-01
-6.33873284e-01 -8.85558248e-01 -4.24758762e-01 7.82320261e-01
-5.85390627e-02 -2.27903932e-01 -1.01310253e+00 -8.97077143e-01
1.58514246e-01 7.20136404e-01 -8.21706593e-01 -7.22236753e-01
-2.40465835e-01 -7.18721747e-01 9.63828802e-01 4.18844968e-01
7.53488183e-01 -1.15131867e+00 -8.70498300e-01 6.41592145e-02
-3.95504862e-01 -6.77649200e-01 -2.78678805e-01 8.33480179e-01
-4.88876283e-01 -8.20132673e-01 -1.16901028e+00 -5.87061703e-01
1.86837509e-01 -1.07032917e-01 7.30664670e-01 -6.28051281e-01
-2.67530471e-01 3.15966308e-01 -3.09280574e-01 -9.12441969e-01
-2.17448641e-02 3.96435171e-01 9.84821543e-02 3.24152291e-01
7.50167310e-01 -8.07236016e-01 -6.31918013e-01 4.04542983e-01
-8.50148320e-01 -1.30555527e-02 5.67387581e-01 1.10448134e+00
8.20382088e-02 -2.26810589e-01 1.08314931e+00 -6.46669328e-01
1.01467538e+00 -8.07376027e-01 -1.63194031e-01 6.05668247e-01
-4.21463400e-01 -1.13302872e-01 4.27505106e-01 -4.13757086e-01
-1.11029136e+00 -1.53843448e-01 1.74216375e-01 -3.84105533e-01
-4.21715438e-01 1.79767728e-01 -7.22646773e-01 -3.37206304e-01
5.54986477e-01 2.90327460e-01 3.75819236e-01 -4.02309120e-01
1.75400227e-01 1.21137917e+00 5.66122055e-01 -5.21272659e-01
8.57175812e-02 1.57931373e-01 -4.89047289e-01 -4.56015289e-01
6.57465756e-02 -4.12435196e-02 -7.34095097e-01 1.55350771e-02
9.63409126e-01 -1.03110719e+00 -8.00787628e-01 8.90123785e-01
-1.24979365e+00 -2.15456396e-01 -2.36853492e-03 6.31509125e-01
-7.55964875e-01 4.61338088e-02 -3.29725295e-01 -4.53348219e-01
-7.58244872e-01 -1.12306845e+00 7.13379920e-01 -2.80680388e-01
-2.33840987e-01 -5.06447077e-01 -1.64224803e-01 1.00639321e-01
6.59199417e-01 3.33129048e-01 7.89102733e-01 -1.03445816e+00
-9.75585505e-02 -3.14214379e-01 -3.03565234e-01 7.34116554e-01
3.09621096e-01 -1.05572701e+00 -1.23201108e+00 -8.48113179e-01
4.68640089e-01 -7.72845209e-01 6.57793581e-01 3.65007631e-02
1.63344085e+00 -4.24806267e-01 -5.26077986e-01 1.02670133e+00
1.25118864e+00 3.08253944e-01 1.14368129e+00 4.26641732e-01
5.64573109e-01 7.82821834e-01 -6.70674890e-02 4.43689734e-01
-3.51796225e-02 2.67676741e-01 5.59239745e-01 2.06442811e-02
2.61395872e-01 2.06609920e-01 3.67675871e-01 5.47066331e-01
-1.71854213e-01 -2.44358510e-01 -6.32314205e-01 5.38766384e-01
-1.91025245e+00 -5.37415504e-01 1.52254269e-01 2.50140214e+00
9.46412802e-01 -3.22989017e-01 3.78041640e-02 1.65821426e-02
4.81339246e-01 -4.85876173e-01 -9.82013047e-01 -2.48352334e-01
-2.70744473e-01 6.00869000e-01 8.78174961e-01 -3.02021295e-01
-9.30031598e-01 5.33953071e-01 5.18406534e+00 8.30055833e-01
-1.22386003e+00 6.59204841e-01 4.71641302e-01 -7.45029867e-01
1.36369780e-01 -6.41591549e-01 -3.35567564e-01 8.60165656e-01
1.14959586e+00 -4.70799536e-01 6.80562198e-01 4.93789107e-01
-4.73006397e-01 1.68703720e-01 -1.46172667e+00 1.49620855e+00
1.02858722e-01 -8.81747007e-01 -2.95547873e-01 4.94865596e-01
3.48634928e-01 5.12400687e-01 1.86856493e-01 3.86612773e-01
1.31285191e-01 -1.34097254e+00 7.12807298e-01 3.71230274e-01
1.18910158e+00 -9.44450259e-01 7.90723205e-01 4.26348269e-01
-3.97769392e-01 -1.76408559e-01 -3.88781160e-01 2.83485532e-01
-2.83227503e-01 1.51853696e-01 -3.24435711e-01 9.27288771e-01
1.41672921e+00 8.10603797e-01 -4.54687297e-01 1.05610275e+00
3.52997296e-02 4.32712525e-01 -3.55020463e-01 2.82348562e-02
-1.73674934e-02 -2.54976749e-01 2.33598948e-01 1.11614251e+00
5.98049819e-01 -1.48449466e-01 -6.00922346e-01 9.36608255e-01
-2.75392771e-01 8.45046267e-02 -1.01851165e+00 3.69209021e-01
1.37793913e-01 9.98930097e-01 1.42145544e-01 5.90827204e-02
-7.78926611e-01 1.54629707e+00 6.92188561e-01 5.50174296e-01
-9.20074105e-01 -4.92370516e-01 7.90040910e-01 -2.24807978e-01
1.62247896e-01 2.33694583e-01 -5.11656106e-01 -1.41997182e+00
1.30873635e-01 -1.03474796e+00 4.20492351e-01 -6.08394623e-01
-1.69621718e+00 8.29255641e-01 1.58250377e-01 -1.27102554e+00
5.99115528e-02 -8.04984808e-01 -1.50533438e-01 1.20240414e+00
-1.39110518e+00 -1.09365773e+00 -1.19681448e-01 1.51689911e+00
-2.88910210e-01 -5.46431959e-01 1.28044772e+00 6.22350216e-01
-4.52969074e-01 1.12124956e+00 5.87798476e-01 3.17104876e-01
1.27219141e+00 -8.85610878e-01 -8.78394693e-02 3.87469649e-01
-1.39492974e-01 5.29441059e-01 -9.74962115e-02 -4.69241649e-01
-1.35466051e+00 -1.15630054e+00 4.70115244e-01 -4.09409225e-01
3.49853933e-01 -9.53958511e-01 -1.12778258e+00 8.84512126e-01
8.60685587e-01 -8.66514221e-02 1.11084032e+00 -1.21870026e-01
-1.99061662e-01 -4.42624062e-01 -1.59972060e+00 4.46933448e-01
1.03107071e+00 -4.35929835e-01 -6.61472440e-01 1.98646545e-01
3.46075833e-01 -4.18703519e-02 -8.95463049e-01 1.07596777e-01
5.90627432e-01 -7.87092328e-01 7.11867988e-01 -6.98968768e-01
1.38205737e-01 2.47979894e-01 -9.94635373e-02 -1.88650441e+00
-2.43576676e-01 -8.73219848e-01 4.51282151e-02 1.14539552e+00
3.20406616e-01 -1.20241439e+00 6.39320493e-01 1.26918006e+00
1.30700871e-01 -3.29326630e-01 -1.55302131e+00 -9.61872101e-01
5.70576191e-01 -3.66036922e-01 1.01180995e+00 8.76144946e-01
5.15768588e-01 9.38539654e-02 -8.18102062e-01 -2.37849250e-01
6.83140099e-01 -4.75961655e-01 6.49771452e-01 -1.21650720e+00
8.48190486e-02 -1.69375256e-01 -6.09920740e-01 -3.97298098e-01
3.38506371e-01 -1.45315182e+00 -3.58783677e-02 -1.02516615e+00
9.33384523e-02 -3.62199992e-01 -9.12679315e-01 1.00588632e+00
2.51369864e-01 1.11699320e-01 1.07684202e-01 1.95332825e-01
-1.24012828e-01 6.51373625e-01 1.01672852e+00 -3.61279935e-01
-4.22002301e-02 -1.15689240e-01 -8.88708413e-01 2.17155367e-01
1.06928885e+00 -6.45598352e-01 -4.80295479e-01 -6.70571387e-01
-1.68202981e-01 -2.00276047e-01 5.13126075e-01 -1.22756243e+00
5.66591620e-01 6.01463199e-01 6.95775867e-01 -3.41338426e-01
1.28440231e-01 -1.46561432e+00 3.85951132e-01 3.47316384e-01
-2.59317070e-01 -5.10491192e-01 4.65979278e-01 6.62295938e-01
-1.65838093e-01 9.49311405e-02 4.72151577e-01 -1.01697594e-02
-4.29186672e-01 4.76569057e-01 -3.47792357e-01 -2.46068463e-03
1.30676329e+00 -2.92109668e-01 -1.45039544e-01 -1.69141427e-01
-7.43606627e-01 1.29666924e-01 1.20564185e-01 5.63775539e-01
6.04781628e-01 -1.24546885e+00 -9.86453652e-01 6.79031014e-01
4.32778537e-01 -2.04343811e-01 -2.54788194e-02 1.02317750e+00
7.30372444e-02 3.64016563e-01 -9.89851892e-01 -4.81872648e-01
-9.46085453e-01 7.07924843e-01 2.04875171e-01 1.19502731e-01
-8.22451353e-01 8.82495642e-01 2.77545840e-01 -4.28596407e-01
5.87294996e-01 -2.85094202e-01 2.81940289e-02 1.47924855e-01
5.68577766e-01 4.03983831e-01 6.01409376e-01 -1.92400157e-01
-6.71396434e-01 -1.24867633e-01 -3.59585673e-01 -1.80112630e-01
1.62289095e+00 9.19344649e-02 -3.32981557e-01 1.47718992e-02
1.74809480e+00 -6.41832948e-01 -1.24909997e+00 -2.01349437e-01
-1.28894985e-01 -4.77537483e-01 -5.03525920e-02 -1.29654133e+00
-1.40801847e+00 1.06422269e+00 9.32170630e-01 -2.77220368e-01
1.34905648e+00 -4.35733110e-01 8.60358596e-01 2.66051888e-01
1.24865758e+00 -1.06166625e+00 -5.02187371e-01 1.95141494e-01
1.26277530e+00 -1.24866283e+00 -3.48430783e-01 1.35459989e-01
-8.11850548e-01 9.09454644e-01 5.11839867e-01 4.52378131e-02
9.97573078e-01 5.45908988e-01 -2.34233409e-01 5.36848493e-02
-3.11289907e-01 5.50138235e-01 1.47594243e-01 1.16831362e+00
-4.71895672e-02 1.90198183e-01 -4.22659576e-01 1.27014446e+00
-1.40234694e-01 5.17187417e-01 1.83975939e-02 1.03999436e+00
4.07444119e-01 -1.17150784e+00 -2.40776673e-01 7.33702123e-01
-4.52957332e-01 -1.91080764e-01 -4.93484557e-01 6.45721197e-01
3.93784612e-01 9.60809052e-01 -2.75372922e-01 -4.93420750e-01
2.30700418e-01 3.64359885e-01 5.59615433e-01 -1.83057174e-01
-8.48690450e-01 -3.01904529e-01 -3.97420824e-01 -7.14808643e-01
-2.29250669e-01 -6.28574669e-01 -6.53439224e-01 -1.82321191e-01
-2.41805837e-01 -5.28659001e-02 7.14781106e-01 6.74895227e-01
8.92133117e-01 4.34268475e-01 2.66847610e-01 -8.76111150e-01
-5.12103975e-01 -1.03182113e+00 -9.03614640e-01 4.96966690e-01
3.11031878e-01 -5.49172878e-01 -1.63971528e-01 -6.45560175e-02]
|
[13.100032806396484, 3.449007749557495]
|
cc77fd7f-47f6-452c-9a7d-5dc3b8ca0517
|
cueing-a-pioneer-work-of-encoding-human-gaze
|
2305.15710
| null |
https://arxiv.org/abs/2305.15710v1
|
https://arxiv.org/pdf/2305.15710v1.pdf
|
CUEING: A pioneer work of encoding human gaze for autonomous driving
|
Recent analysis of incidents involving Autonomous Driving Systems (ADS) has shown that the decision-making process of ADS can be significantly different from that of human drivers. To improve the performance of ADS, it may be helpful to incorporate the human decision-making process, particularly the signals provided by the human gaze. There are many existing works to create human gaze datasets and predict the human gaze using deep learning models. However, current datasets of human gaze are noisy and include irrelevant objects that can hinder model training. Additionally, existing CNN-based models for predicting human gaze lack generalizability across different datasets and driving conditions, and many models have a centre bias in their prediction such that the gaze tends to be generated in the centre of the gaze map. To address these gaps, we propose an adaptive method for cleansing existing human gaze datasets and a robust convolutional self-attention gaze prediction model. Our quantitative metrics show that our cleansing method improves models' performance by up to 7.38% and generalizability by up to 8.24% compared to those trained on the original datasets. Furthermore, our model demonstrates an improvement of up to 12.13% in terms of generalizability compared to the state-of-the-art (SOTA) models. Notably, it achieves these gains while conserving up to 98.12% of computation resources.
|
['Xi Zheng', 'Chen Wang', 'Jianchao Lu', 'Yao Deng', 'Yiran Wang', 'Linfeng Liang']
|
2023-05-25
| null | null | null | null |
['eye-tracking']
|
['computer-vision']
|
[ 5.93755580e-02 3.26079309e-01 5.11829853e-02 -6.66366160e-01
-2.46359810e-01 -1.19521983e-01 2.68145472e-01 -2.62184709e-01
-2.84979880e-01 2.12113291e-01 1.67432815e-01 -3.01270276e-01
1.25527531e-01 -3.82710844e-01 -6.54964983e-01 -6.37393415e-01
3.81767571e-01 -9.47539657e-02 2.05857649e-01 -5.37092030e-01
4.49018359e-01 6.25140145e-02 -2.26579523e+00 1.25598893e-01
8.39079559e-01 1.19567323e+00 -2.45049149e-02 6.41862690e-01
1.26704425e-01 9.44489419e-01 -5.91310084e-01 -3.39363188e-01
2.19705790e-01 -2.35418335e-01 -5.86487055e-01 -4.41391081e-01
9.34320331e-01 -3.66191000e-01 -3.95307511e-01 9.91771162e-01
5.32165051e-01 1.73460711e-02 3.76921356e-01 -1.82477880e+00
-1.12819719e+00 2.22835466e-02 -8.78580034e-01 4.96890157e-01
2.04309657e-01 4.43911940e-01 7.44548738e-01 -6.29594564e-01
2.48634651e-01 1.13805103e+00 6.94315612e-01 1.10404003e+00
-8.72689605e-01 -1.35964072e+00 1.82647586e-01 6.74818695e-01
-1.23260415e+00 -7.22453952e-01 6.35289073e-01 -5.78301370e-01
8.27872217e-01 3.85844409e-01 3.51163119e-01 1.21706986e+00
3.23044509e-01 7.76185751e-01 1.16307282e+00 -3.42025191e-01
-9.74670127e-02 9.05012414e-02 5.39355993e-01 4.75231886e-01
2.33529925e-01 3.46962333e-01 -8.77183855e-01 4.71209794e-01
-5.48624285e-02 1.75725460e-01 -5.35196625e-02 3.96024771e-02
-9.08851862e-01 6.59984112e-01 9.69068766e-01 -2.60720760e-01
-2.97338426e-01 6.29039034e-02 6.96831420e-02 -1.26375586e-01
5.84084213e-01 5.31713307e-01 -7.86802620e-02 -2.46548399e-01
-8.04483652e-01 4.95387644e-01 4.25614357e-01 1.21698320e+00
9.61031675e-01 -2.42781743e-01 -4.25053149e-01 5.09465754e-01
5.21531940e-01 9.72307861e-01 2.82946378e-01 -8.98746490e-01
3.38402927e-01 7.93526590e-01 8.06420594e-02 -1.15555823e+00
-6.62084758e-01 -1.34463817e-01 -4.64187324e-01 5.69245934e-01
3.48143339e-01 -1.87843114e-01 -1.01925242e+00 1.82279837e+00
1.52473316e-01 1.47923693e-01 -2.92583495e-01 1.11596644e+00
1.01819432e+00 3.75652939e-01 3.54337662e-01 1.50614604e-01
1.37516749e+00 -1.10499907e+00 -1.13056552e+00 -7.27581918e-01
5.82449198e-01 -5.74785292e-01 1.17380428e+00 1.70103192e-01
-9.75348175e-01 -6.57106876e-01 -1.07922435e+00 -4.07946974e-01
-2.68982589e-01 -2.23949358e-01 3.74119282e-01 6.58322275e-01
-1.22769713e+00 1.82006806e-02 -5.20720840e-01 -3.40495050e-01
7.93899715e-01 4.66303408e-01 -1.26177862e-01 -1.15258768e-01
-9.49223638e-01 1.16178942e+00 -2.46467218e-01 1.93012133e-01
-5.42756140e-01 -9.74325955e-01 -8.37526441e-01 1.28213748e-01
1.73607126e-01 -4.90646809e-01 1.47147024e+00 -1.08594859e+00
-1.11668921e+00 7.78507948e-01 -9.04536903e-01 -5.45298040e-01
2.42253527e-01 -4.49957490e-01 -7.82127559e-01 -1.94777191e-01
2.22792208e-01 1.08186471e+00 8.12900245e-01 -1.10698295e+00
-1.21209991e+00 -4.80153531e-01 2.81004727e-01 2.09986214e-02
-3.45238000e-01 2.75938362e-01 -3.88152778e-01 -3.73215415e-02
-4.28102583e-01 -1.37416315e+00 2.55679339e-01 2.30617244e-02
-2.73616850e-01 -4.08686250e-01 1.04610813e+00 -5.49656510e-01
1.65628505e+00 -2.21941018e+00 -2.49917388e-01 -1.32909849e-01
9.93627846e-01 4.00757998e-01 2.12138444e-02 -1.77982077e-02
-1.83766291e-01 1.26703143e-01 1.80476859e-01 -7.18036771e-01
1.76576227e-01 -4.29329649e-02 -1.35775968e-01 3.68224531e-01
3.82275045e-01 1.13468385e+00 -6.57150090e-01 -2.35222265e-01
1.06951185e-01 4.20787036e-01 -5.23537219e-01 1.94948047e-01
2.83325016e-01 2.96451271e-01 -5.69478720e-02 4.83955353e-01
7.65171528e-01 -2.88722098e-01 -5.19319475e-01 2.36695390e-02
-3.13804418e-01 3.81058365e-01 -3.65776658e-01 1.19191086e+00
-1.74737453e-01 1.38953340e+00 -3.05890948e-01 -2.19926894e-01
7.43897259e-01 -4.30889614e-02 2.72209138e-01 -1.32647169e+00
5.03545880e-01 -2.16370970e-01 4.58984166e-01 -7.54465699e-01
7.19651461e-01 2.15270936e-01 9.55227539e-02 5.75317383e-01
-2.81106889e-01 6.55210614e-01 7.12715313e-02 -2.49185469e-02
1.02007604e+00 -3.14693689e-01 -2.59279728e-01 -2.84895033e-01
3.47114861e-01 -1.05024867e-01 4.89178747e-01 3.24930578e-01
-8.10562193e-01 6.49097741e-01 4.63643909e-01 -5.25296926e-01
-7.93661535e-01 -4.18478429e-01 -1.11634687e-01 1.57493317e+00
5.43332756e-01 -4.13321018e-01 -1.25216985e+00 -6.62538290e-01
-5.39940968e-02 9.67347562e-01 -1.28467941e+00 -6.65810287e-01
-4.56080168e-01 -4.84146744e-01 4.49271441e-01 5.88085115e-01
4.83695209e-01 -1.04299235e+00 -7.48948991e-01 -4.91325647e-01
-1.35690495e-01 -1.08908010e+00 -4.65182900e-01 -3.76011282e-01
-1.32599667e-01 -1.25141025e+00 -3.89730573e-01 -4.89667147e-01
6.50782347e-01 7.83018887e-01 9.80498910e-01 3.30606073e-01
6.13459907e-02 -5.94017021e-02 -2.58477032e-01 -1.39297593e+00
-8.75773132e-02 3.75461817e-01 7.41284564e-02 2.60542959e-01
1.47663403e+00 1.62738353e-01 -8.63483131e-01 4.67687905e-01
-5.03316760e-01 1.74770936e-01 3.67453068e-01 5.70130527e-01
5.36459088e-02 -3.69480759e-01 5.40450871e-01 -7.09172249e-01
6.18279874e-01 -6.68868244e-01 -4.06722724e-01 -1.37864798e-01
-1.19373059e+00 -1.44318312e-01 -1.16179295e-01 -1.71643987e-01
-1.16980910e+00 -2.52686262e-01 1.16489001e-01 -5.40242970e-01
-4.45082724e-01 -2.11491212e-02 2.24765629e-01 -1.07150100e-01
9.97187495e-01 -4.09524560e-01 3.50605845e-01 -2.09587470e-01
-4.30904888e-02 1.06681216e+00 2.71329433e-01 2.70782799e-01
6.15417123e-01 4.68194008e-01 -5.62415905e-02 -4.94984895e-01
-1.22250473e+00 -5.38009048e-01 -5.46372890e-01 -4.45975929e-01
1.09572256e+00 -9.42187905e-01 -1.22669864e+00 6.95180476e-01
-9.22893405e-01 -2.74764895e-01 1.26099020e-01 2.55140543e-01
-1.25363365e-01 -2.35321343e-01 -1.29685923e-01 -8.28337967e-01
-4.34372902e-01 -1.21745014e+00 1.23873532e+00 5.70876360e-01
-4.86815959e-01 -6.24111235e-01 -4.17474322e-02 6.75012052e-01
6.73447907e-01 -2.57925503e-02 3.32993805e-01 -4.31324840e-01
-4.69063550e-01 -2.59010851e-01 -5.57368159e-01 1.39516249e-01
5.69184162e-02 1.14774607e-01 -1.61541522e+00 -1.93488255e-01
-2.40663663e-01 1.10868089e-01 7.54229128e-01 4.13962275e-01
9.94440794e-01 -4.76185419e-02 -7.13039696e-01 5.06677806e-01
8.41764033e-01 1.92584038e-01 8.69878352e-01 5.62260032e-01
9.83224154e-01 9.50967669e-01 7.41513789e-01 9.05168951e-02
9.86772597e-01 5.57669938e-01 7.84882247e-01 -1.98715046e-01
-3.36756527e-01 -5.31367846e-02 1.32031798e-01 3.17996919e-01
-2.14965463e-01 -3.17661345e-01 -1.31288576e+00 6.35982037e-01
-1.84400296e+00 -1.04590964e+00 -7.20701337e-01 1.87054014e+00
4.87857401e-01 1.00037895e-01 3.65937233e-01 1.63464800e-01
7.39821613e-01 -9.94141102e-02 -6.89252138e-01 -4.54319239e-01
2.28586286e-01 -2.57855058e-02 6.55602336e-01 2.17317104e-01
-9.78627503e-01 7.20807731e-01 6.99564171e+00 7.02250451e-02
-1.51932514e+00 1.85610175e-01 2.66336262e-01 -7.10935593e-01
3.29840332e-02 -3.85995626e-01 -1.12003815e+00 8.67230356e-01
1.38566506e+00 -1.07723527e-01 3.21542591e-01 8.07569325e-01
2.41070822e-01 -5.68639189e-02 -9.82873976e-01 1.18123329e+00
4.74888206e-01 -9.29646909e-01 -5.93644381e-01 4.35046673e-01
6.92588091e-01 3.64961267e-01 4.44618642e-01 4.31254953e-01
-2.79924320e-03 -1.25831914e+00 1.00262928e+00 8.69055808e-01
6.99721575e-01 -7.99313009e-01 9.35355604e-01 3.01196873e-01
-5.27445555e-01 -5.57008803e-01 -9.57285836e-02 -4.68043208e-01
4.94910441e-02 7.29445890e-02 -8.81736457e-01 -6.02018014e-02
1.28437364e+00 9.43297923e-01 -1.18522012e+00 9.10079002e-01
-1.08117409e-01 6.87214553e-01 -4.47777584e-02 -1.40978098e-01
1.28811494e-01 3.38677049e-01 2.92585135e-01 8.67100060e-01
2.56244272e-01 -1.96733940e-02 -6.10769153e-01 9.63054657e-01
-1.23848291e-02 -3.80637407e-01 -5.27187228e-01 2.73320615e-01
4.42065597e-01 1.10353351e+00 1.15899906e-01 -6.24929816e-02
-5.41625857e-01 3.71512771e-01 2.88469225e-01 4.84653622e-01
-1.15951037e+00 -4.40670639e-01 1.27439785e+00 5.77937663e-01
2.32619584e-01 1.57056645e-01 -7.17050433e-01 -6.54742718e-01
2.93783881e-02 -6.75049961e-01 -6.20904304e-02 -1.24004698e+00
-1.20572746e+00 7.39290118e-01 -1.82977229e-01 -1.26524162e+00
-2.32791156e-01 -6.12001896e-01 -4.36580390e-01 1.23869598e+00
-1.87426412e+00 -1.34008992e+00 -1.02812779e+00 4.85585600e-01
3.95660669e-01 -2.83611298e-01 5.27295709e-01 4.53406364e-01
-9.02058542e-01 9.60380912e-01 -4.19358343e-01 -1.04360007e-01
1.09454930e+00 -1.13579547e+00 5.60897648e-01 8.70417833e-01
-4.81747270e-01 6.83455169e-01 8.01177680e-01 -2.43005514e-01
-1.13263166e+00 -1.07771337e+00 1.16506100e+00 -1.23401058e+00
3.94893169e-01 -3.21298391e-01 -1.03156805e+00 9.57805514e-01
6.53250575e-01 -8.68465155e-02 6.78277671e-01 4.25533146e-01
-3.43646854e-01 -3.32451582e-01 -8.57272029e-01 7.83839226e-01
1.15565646e+00 -5.29654086e-01 -6.37546003e-01 -1.66005671e-01
4.54073727e-01 -6.12046003e-01 -3.22951525e-01 2.05026284e-01
6.84527218e-01 -1.09665275e+00 5.91421902e-01 -5.41377783e-01
4.78412032e-01 -4.34200525e-01 2.79818445e-01 -1.11959493e+00
-7.51997352e-01 -5.67688882e-01 -3.59088629e-01 1.07205522e+00
3.59846383e-01 -6.95530772e-01 5.41352212e-01 1.27879655e+00
-2.86692858e-01 -4.87016231e-01 -5.20851195e-01 -6.87950850e-02
-1.37492388e-01 -5.91287076e-01 1.10818553e+00 7.29803741e-01
-3.38497944e-02 6.71658754e-01 -3.04107428e-01 2.41765097e-01
4.01580006e-01 -3.50832909e-01 1.19044948e+00 -1.41758192e+00
6.23528183e-01 -5.63009441e-01 -6.44399881e-01 -9.30137098e-01
3.04116964e-01 -3.84220541e-01 2.67986417e-01 -1.30970144e+00
-5.39042167e-02 -4.76551712e-01 -5.76542616e-01 6.13554895e-01
-4.70022827e-01 4.66603786e-01 1.86115623e-01 2.89116114e-01
-6.82203710e-01 2.23522767e-01 1.27972293e+00 -9.03851539e-02
5.16772345e-02 7.58451149e-02 -1.37749541e+00 6.01711810e-01
6.34317458e-01 -5.12635410e-01 -4.56510514e-01 -7.15639830e-01
4.75004643e-01 -7.33639717e-01 6.15272105e-01 -9.65282142e-01
6.37521744e-01 -5.23447245e-02 2.20880210e-01 -6.04631603e-01
3.61595869e-01 -7.60096312e-01 -1.85908869e-01 -6.17127828e-02
-3.61286134e-01 1.10851072e-01 4.80157554e-01 5.86014926e-01
-1.24200597e-01 5.55664115e-03 5.80059290e-01 4.68638003e-01
-8.40590835e-01 2.49130130e-01 -1.79919422e-01 -9.36717689e-02
1.22771132e+00 -3.95023942e-01 -8.73958588e-01 -3.36826265e-01
-2.21894041e-01 3.74231637e-01 5.79342544e-01 1.09174168e+00
1.32954910e-01 -1.20668793e+00 -3.82761538e-01 4.99890417e-01
5.92635334e-01 2.11301483e-02 3.85267735e-01 1.10340762e+00
-2.13256314e-01 7.16310203e-01 -4.82552618e-01 -8.94004703e-01
-1.42637074e+00 6.60716236e-01 3.12944770e-01 6.13534808e-01
-2.27727622e-01 8.60949337e-01 6.21538579e-01 4.31400947e-02
4.38860022e-02 -4.66404974e-01 -6.35282934e-01 8.47736970e-02
1.08453941e+00 6.01911247e-01 3.29618663e-01 -1.28260326e+00
-3.77083600e-01 5.18516541e-01 -1.04731731e-01 4.69413757e-01
1.16533124e+00 -5.77327907e-01 1.01238236e-01 2.50829965e-01
1.01177633e+00 -8.01931396e-02 -1.47312689e+00 -1.11176983e-01
-2.48527646e-01 -6.42555952e-01 3.62513274e-01 -7.85422385e-01
-1.01784039e+00 1.07515275e+00 9.62125540e-01 5.98093569e-01
1.21940160e+00 -5.83298951e-02 8.59853089e-01 4.49463874e-02
-4.02520262e-02 -1.03074110e+00 -3.78219187e-01 5.02873302e-01
8.19682360e-01 -1.84351587e+00 -3.98247361e-01 -2.08717316e-01
-9.78637934e-01 6.76914275e-01 9.22682524e-01 1.66900188e-03
8.29147398e-01 -4.08549421e-02 5.25870085e-01 -6.03423476e-01
-9.20060635e-01 -3.64670366e-01 8.29033732e-01 7.51474142e-01
2.18586713e-01 -4.23884317e-02 2.63141572e-01 6.30996168e-01
-5.59893012e-01 2.04171255e-01 3.23071718e-01 7.09919095e-01
-1.95800051e-01 -4.28941101e-01 -2.82991618e-01 4.55923945e-01
-4.88960415e-01 -1.75625365e-02 -3.10945928e-01 6.48661673e-01
2.78096616e-01 1.35855067e+00 4.81447250e-01 -7.49849916e-01
7.00250447e-01 1.10300466e-01 -5.53820170e-02 -4.82502192e-01
-4.60314810e-01 -3.77904892e-01 2.86407899e-02 -7.88600445e-01
-6.57348633e-01 -8.26313555e-01 -1.01095092e+00 -8.63883376e-01
-4.75524157e-01 -5.14182270e-01 6.53495073e-01 1.07891762e+00
9.37716544e-01 9.21157777e-01 4.37821001e-01 -9.67368662e-01
-8.09049383e-02 -1.19400764e+00 -2.56193012e-01 5.22547185e-01
7.80120373e-01 -1.13334167e+00 -4.05139387e-01 3.96585584e-01]
|
[14.130745887756348, 0.077862448990345]
|
883e380b-f610-44fb-9c6e-458e61bb1208
|
global-and-local-semantic-completion-learning
|
2306.07096
| null |
https://arxiv.org/abs/2306.07096v1
|
https://arxiv.org/pdf/2306.07096v1.pdf
|
Global and Local Semantic Completion Learning for Vision-Language Pre-training
|
Cross-modal alignment plays a crucial role in vision-language pre-training (VLP) models, enabling them to capture meaningful associations across different modalities. For this purpose, inspired by the success of masked language modeling (MLM) tasks in the NLP pre-training area, numerous masked modeling tasks have been proposed for VLP to further promote cross-modal interactions. The core idea of previous masked modeling tasks is to focus on reconstructing the masked tokens based on visible context for learning local-local alignment, i.e., associations between image patches and text tokens. However, most of them pay little attention to the global semantic features generated for the masked data, resulting in a limited cross-modal alignment ability of global representations to local features of the other modality. Therefore, in this paper, we propose a novel Global and Local Semantic Completion Learning (GLSCL) task to facilitate global-local alignment and local-local alignment simultaneously. Specifically, the GLSCL task complements the missing semantics of masked data and recovers global and local features by cross-modal interactions. Our GLSCL consists of masked global semantic completion (MGSC) and masked local token completion (MLTC). MGSC promotes learning more representative global features which have a great impact on the performance of downstream tasks, and MLTC can further enhance accurate comprehension on multimodal data. Moreover, we present a flexible vision encoder, enabling our model to simultaneously perform image-text and video-text multimodal tasks. Experimental results show that our proposed method obtains state-of-the-art performance on various vision-language benchmarks, such as visual question answering, image-text retrieval, and video-text retrieval.
|
['Wei Liu', 'Yujiu Yang', 'Hongfa Wang', 'Wenzhe Zhao', 'Chengfei Cai', 'Weijie Kong', 'Jie Jiang', 'Yatai Ji', 'Rong-Cheng Tu']
|
2023-06-12
| null | null | null | null |
['visual-question-answering-1', 'video-text-retrieval']
|
['computer-vision', 'computer-vision']
|
[ 3.03402394e-01 5.39064422e-05 -3.50991100e-01 -4.64185566e-01
-1.06989598e+00 -3.22946936e-01 9.77561474e-01 1.17805436e-01
-3.54463637e-01 3.25966388e-01 4.18492228e-01 -1.56333789e-01
1.68956146e-01 -3.90311927e-01 -1.03631985e+00 -6.94776595e-01
5.83769977e-01 1.91673636e-01 9.53900516e-02 -5.91343194e-02
1.05071217e-01 -1.26536041e-01 -1.60696614e+00 9.57885563e-01
8.90297234e-01 8.60012352e-01 8.09162736e-01 3.41943026e-01
-4.12415683e-01 1.00535929e+00 -1.49811059e-01 -4.84173715e-01
-1.67665526e-01 -4.40359980e-01 -7.39016473e-01 2.48242348e-01
6.70834422e-01 -3.78716111e-01 -3.35297137e-01 9.88532543e-01
2.22551137e-01 -6.38799369e-03 6.38525128e-01 -1.33020794e+00
-8.89177203e-01 5.70559800e-01 -6.88179135e-01 -2.01542258e-01
4.01685327e-01 2.63404906e-01 1.17038214e+00 -1.48911417e+00
5.00428855e-01 1.48315752e+00 1.84539214e-01 5.03545702e-01
-1.17219460e+00 -5.07386625e-01 4.60625410e-01 4.26272213e-01
-1.30535400e+00 -5.78766048e-01 6.28021419e-01 -4.54591364e-01
7.79512107e-01 1.30347908e-01 2.40145832e-01 1.24884665e+00
6.59860112e-03 1.29741251e+00 1.11277652e+00 -5.44964790e-01
-1.57026246e-01 1.40470207e-01 -9.24530253e-02 8.74882579e-01
-1.08791664e-01 -7.97944143e-02 -9.14766371e-01 2.72138447e-01
5.89845002e-01 1.63508475e-01 -3.97572815e-01 -2.45933801e-01
-1.60700393e+00 6.16241276e-01 4.85554427e-01 2.64082223e-01
-3.06900710e-01 1.47538289e-01 1.48992047e-01 2.07329337e-02
2.00996935e-01 -7.98728392e-02 -2.81772763e-01 2.64989257e-01
-9.17362630e-01 -7.37701282e-02 1.94471434e-01 9.46191132e-01
1.16337132e+00 -1.10240109e-01 -5.72606981e-01 8.20132911e-01
6.38867378e-01 6.93401217e-01 4.56117779e-01 -7.66764820e-01
8.75914454e-01 7.48170495e-01 -6.79675639e-02 -9.26129878e-01
-1.82004407e-01 -1.87894776e-01 -9.80554461e-01 -9.16979760e-02
2.23128796e-01 4.41611260e-01 -1.12912130e+00 2.01921988e+00
1.34909093e-01 9.99939218e-02 3.37979496e-01 9.12793159e-01
1.19429898e+00 7.93953419e-01 3.00568759e-01 -1.60239547e-01
1.50003839e+00 -1.33715856e+00 -6.94843590e-01 -5.14673471e-01
4.77536023e-01 -1.07622015e+00 1.36540818e+00 8.82781148e-02
-1.07733369e+00 -9.58853602e-01 -5.83084643e-01 -5.53623140e-01
-2.52201617e-01 3.43682379e-01 4.28319454e-01 1.96699444e-02
-9.35460091e-01 -4.92573343e-02 -6.88503385e-01 -3.55798006e-01
4.62146997e-01 7.44724050e-02 -4.85669762e-01 -7.23070920e-01
-1.05078983e+00 7.96971202e-01 4.11281645e-01 3.55472505e-01
-1.26511586e+00 -4.48828757e-01 -1.04111171e+00 6.76557720e-02
4.46373701e-01 -8.52935493e-01 8.74215066e-01 -1.05221355e+00
-1.03042161e+00 1.23541689e+00 -7.76925683e-01 -2.09334448e-01
2.47858658e-01 -6.11966960e-02 -2.10349381e-01 3.93923193e-01
4.02433187e-01 1.21377170e+00 1.32346666e+00 -1.62017918e+00
-5.19919395e-01 -3.49788576e-01 2.87753996e-02 4.80938941e-01
-3.17789942e-01 -8.88027996e-02 -1.00814915e+00 -7.29539931e-01
1.92734301e-01 -4.91237283e-01 4.10994254e-02 -1.45785719e-01
-3.40143144e-01 -3.07877362e-01 4.75139797e-01 -9.85505700e-01
7.59568334e-01 -2.19502783e+00 4.21809375e-01 -2.28654891e-01
3.14129591e-01 1.50979608e-01 -6.32073283e-01 5.39086223e-01
2.27600075e-02 -1.31438643e-01 -1.97875261e-01 -9.29396152e-01
1.11014940e-01 3.87441337e-01 -6.54253304e-01 8.54186490e-02
3.64383280e-01 1.47903311e+00 -6.73866808e-01 -7.69129694e-01
3.50123852e-01 3.68863851e-01 -4.10883963e-01 3.63748968e-01
-4.52410042e-01 6.93755269e-01 -3.05379719e-01 9.12363946e-01
6.33714855e-01 -4.74516481e-01 -9.84853506e-02 -4.84414548e-01
1.15766145e-01 -1.31005943e-01 -6.02840126e-01 2.13318920e+00
-5.12121856e-01 5.01015425e-01 1.07485287e-01 -1.06964016e+00
6.36525810e-01 2.42090657e-01 2.84078836e-01 -1.03288829e+00
-1.38807252e-01 6.20080493e-02 -5.47735810e-01 -7.81200469e-01
5.06707132e-01 -1.23426057e-01 2.56762784e-02 3.26656491e-01
2.34028354e-01 1.35591805e-01 -5.07656597e-02 5.76882541e-01
4.60312963e-01 2.23934487e-01 6.57538474e-02 2.43645474e-01
7.34534562e-01 -7.48648793e-02 2.42218599e-01 6.84956849e-01
-7.70682748e-03 7.68803596e-01 2.11597353e-01 3.70317474e-02
-6.66049123e-01 -1.11224079e+00 1.85065567e-01 1.26873755e+00
5.50335944e-01 -4.62708682e-01 -5.08485675e-01 -5.19451082e-01
-1.94695860e-01 5.79114497e-01 -5.02061665e-01 -2.23610327e-01
-3.94675404e-01 -4.85399485e-01 4.63858157e-01 5.31989694e-01
6.68207586e-01 -1.28011167e+00 4.00206335e-02 -2.40948930e-01
-8.66035461e-01 -1.72805607e+00 -7.40784585e-01 -1.82537735e-01
-6.96508169e-01 -1.03403378e+00 -5.75772643e-01 -1.20072854e+00
8.96324813e-01 8.36077154e-01 1.01634538e+00 9.12168920e-02
-1.59620970e-01 7.21539021e-01 -4.58569348e-01 -5.69208805e-03
-2.46674970e-01 -2.45620146e-01 -1.92469388e-01 4.90465224e-01
2.80039787e-01 -2.18512014e-01 -5.89517653e-01 3.00420821e-01
-1.21554196e+00 7.18046188e-01 1.03696680e+00 9.55758870e-01
7.59624898e-01 -3.24961990e-01 5.00658453e-01 -3.60435486e-01
2.96992123e-01 -3.12225640e-01 -2.72707105e-01 8.02570999e-01
-2.48068020e-01 1.10181998e-02 3.59415084e-01 -5.48747540e-01
-1.19994259e+00 4.17649262e-02 2.70895641e-02 -7.49974489e-01
-1.66819379e-01 7.18208790e-01 -6.15821838e-01 8.99756029e-02
2.30188757e-01 7.69795418e-01 6.47430345e-02 -4.74038571e-01
7.17217147e-01 5.72548687e-01 7.11402416e-01 -6.89863861e-01
7.18819678e-01 6.64312840e-01 -1.96844593e-01 -7.77392566e-01
-1.08679402e+00 -6.58180714e-01 -5.65306008e-01 -6.04684232e-04
1.23017919e+00 -1.38842034e+00 -6.57471478e-01 6.52838349e-01
-1.22128582e+00 -3.10910404e-01 -1.39835745e-01 3.36566597e-01
-5.78651726e-01 8.16695273e-01 -2.44724274e-01 -5.30235946e-01
-2.35847294e-01 -1.29773402e+00 1.48458683e+00 1.51636079e-01
2.73243040e-01 -9.22022402e-01 -3.94886971e-01 1.09017301e+00
2.28455558e-01 -2.29776040e-01 1.04983461e+00 -3.09072971e-01
-9.19173121e-01 1.01202816e-01 -6.96149766e-01 5.20012736e-01
-1.58492066e-02 -5.19750416e-01 -1.14134312e+00 -3.89152169e-01
8.35927483e-03 -6.71035349e-01 1.31760418e+00 3.58937174e-01
1.02517104e+00 -1.90488428e-01 -2.93786764e-01 5.91723382e-01
1.24781656e+00 -3.68747026e-01 6.72474384e-01 2.63874084e-02
1.12067509e+00 7.60866046e-01 7.61322916e-01 4.97340038e-02
8.72376919e-01 6.50902629e-01 6.15033150e-01 -3.44510168e-01
-4.58183527e-01 -6.18816316e-01 7.67563760e-01 8.27691674e-01
2.98153102e-01 -7.15187863e-02 -7.96887577e-01 6.30996406e-01
-2.21856332e+00 -7.93392241e-01 -1.21861905e-01 2.03732252e+00
9.54069078e-01 -3.19779992e-01 -3.41749430e-01 -4.54129666e-01
7.03868389e-01 3.16000134e-01 -4.31975812e-01 2.83294469e-01
-4.68455046e-01 5.03618037e-03 -6.27241582e-02 6.48546815e-01
-1.00594783e+00 1.30269933e+00 4.71168375e+00 1.12084782e+00
-8.47780108e-01 3.45874488e-01 4.92230058e-01 1.20784156e-01
-5.17219722e-01 2.18474135e-01 -8.19661379e-01 3.04756254e-01
2.32460737e-01 2.08913535e-01 4.61864829e-01 3.80574793e-01
2.24136934e-01 -2.32330948e-01 -1.10569191e+00 1.37024236e+00
3.95049155e-01 -1.27604318e+00 7.45809376e-01 -7.11873770e-02
8.40893030e-01 -8.93610492e-02 1.94446608e-01 3.80179733e-01
-1.68129832e-01 -1.21670902e+00 6.88800037e-01 6.99861705e-01
9.31343973e-01 -2.76621938e-01 5.25043011e-01 5.09904563e-01
-1.28898668e+00 -9.05495733e-02 -4.08035815e-01 3.04241866e-01
1.65064514e-01 3.98428023e-01 -6.45430088e-01 7.31302440e-01
5.42043507e-01 8.49050462e-01 -7.32143760e-01 5.76295853e-01
-3.15715939e-01 3.45986277e-01 6.46916702e-02 4.08367097e-01
2.46087596e-01 -4.28816341e-02 3.14117610e-01 1.00197887e+00
-7.78684905e-03 -1.61629766e-01 3.49179983e-01 1.10410404e+00
-1.64108247e-01 1.75333276e-01 -2.85581976e-01 -1.17488816e-01
2.26905376e-01 1.28797674e+00 -2.72858888e-01 -3.50936800e-01
-7.74698079e-01 1.11717665e+00 2.81941056e-01 7.94442415e-01
-7.15532124e-01 1.18325129e-01 4.76551294e-01 -2.00731248e-01
2.88276017e-01 -2.58189410e-01 -1.60747990e-01 -1.59728968e+00
2.54946351e-01 -9.65170443e-01 4.62446958e-01 -1.13787973e+00
-1.36472058e+00 3.34970534e-01 -2.17250939e-02 -1.16123080e+00
-2.70138383e-02 -5.21902084e-01 -3.87792349e-01 1.04989147e+00
-1.82917261e+00 -2.08549690e+00 -6.75706387e-01 1.20685697e+00
7.88717091e-01 -1.53761029e-01 5.56886613e-01 1.28216490e-01
-2.62499243e-01 4.91820574e-01 -1.54052332e-01 7.70369023e-02
9.69176173e-01 -8.52857947e-01 3.74197550e-02 9.98567581e-01
5.22100389e-01 6.41506374e-01 1.70560136e-01 -5.67859769e-01
-1.53497064e+00 -1.17944896e+00 9.01755929e-01 -5.46913207e-01
5.06519020e-01 -3.47752005e-01 -9.13781583e-01 7.21925795e-01
2.87692070e-01 -8.31162333e-02 3.60066980e-01 -1.05604775e-01
-5.49361706e-01 -1.51297063e-01 -4.89070714e-01 6.24633074e-01
8.35651994e-01 -1.08925891e+00 -6.22315228e-01 4.32417035e-01
8.78003001e-01 -1.59735113e-01 -4.05750811e-01 5.12857795e-01
3.41366708e-01 -8.41536760e-01 1.08500767e+00 -4.98869061e-01
6.68592811e-01 -4.30472881e-01 -5.39929628e-01 -7.99643278e-01
1.40850544e-01 -3.17686409e-01 -1.28053859e-01 1.38168371e+00
6.34182021e-02 -3.57306480e-01 3.83140743e-01 2.59013623e-01
-2.11949468e-01 -5.49664080e-01 -8.66969407e-01 -3.36990088e-01
-8.78136829e-02 -5.74571669e-01 1.69646338e-01 1.05057526e+00
-1.13540404e-01 4.89690721e-01 -5.86425066e-01 3.77255976e-01
7.46652961e-01 4.63081688e-01 8.89450431e-01 -7.63260782e-01
-4.24865216e-01 -3.94237578e-01 -9.57185403e-02 -1.54455602e+00
3.50296348e-01 -1.19959235e+00 1.63530096e-01 -1.75041902e+00
7.00057149e-01 1.39703555e-02 -2.98373044e-01 7.58549333e-01
-6.27082348e-01 3.91272604e-01 2.79121995e-01 5.53418756e-01
-1.00963998e+00 8.85051847e-01 1.54788065e+00 -3.38051468e-01
6.43879250e-02 -3.39894176e-01 -7.55220234e-01 5.40690660e-01
4.36539412e-01 -1.31963208e-01 -4.64296997e-01 -7.93000698e-01
1.87343165e-01 1.17169321e-01 8.74619186e-01 -5.55912733e-01
3.30752671e-01 -1.80098996e-01 3.32963377e-01 -7.93588638e-01
5.42968869e-01 -6.54030144e-01 -3.66750687e-01 7.46729076e-02
-3.48371834e-01 -2.90077478e-01 5.48537523e-02 7.65281558e-01
-7.01021791e-01 -3.62682715e-02 4.40937400e-01 -1.26361936e-01
-9.86830652e-01 2.86278307e-01 -9.59008634e-02 8.46149027e-02
7.27966309e-01 -1.15391903e-01 -5.73580205e-01 -4.36411023e-01
-8.66157711e-01 5.83202243e-01 3.91963631e-01 7.25145519e-01
1.10289180e+00 -1.29630256e+00 -7.00884521e-01 2.90150553e-01
5.20271003e-01 2.35330939e-01 6.58200145e-01 1.25816119e+00
-1.21383712e-01 6.10605896e-01 -1.54851179e-03 -1.01246786e+00
-1.40943658e+00 7.17859685e-01 1.74151614e-01 -1.92504525e-01
-5.06715059e-01 8.46938074e-01 9.66522574e-01 -3.18143427e-01
3.94166023e-01 -2.59519279e-01 -8.79555345e-02 5.71339540e-02
4.58773524e-01 -1.68529764e-01 -2.11993471e-01 -8.32570493e-01
-2.19667181e-01 8.21799576e-01 -1.41727060e-01 -1.87784433e-01
9.38811421e-01 -6.15150392e-01 -4.43975210e-01 3.88619095e-01
1.17503202e+00 -1.29401879e-02 -1.18338871e+00 -8.28915954e-01
-2.23709539e-01 -3.41680467e-01 -1.28516838e-01 -7.03262627e-01
-8.86856079e-01 1.32210481e+00 4.14563447e-01 -3.94236803e-01
1.29637754e+00 4.88124341e-01 7.43583202e-01 2.89266229e-01
2.15166673e-01 -7.26639688e-01 6.83993340e-01 4.08932328e-01
9.27439272e-01 -1.65571034e+00 -1.90931767e-01 -3.69053572e-01
-9.82049763e-01 8.74236226e-01 6.74682140e-01 4.58588690e-01
2.46599987e-01 -4.69839931e-01 6.79531395e-02 -1.55701444e-01
-7.16737211e-01 -5.99807560e-01 6.24285817e-01 5.40337503e-01
1.36184379e-01 -1.31310895e-01 1.02479354e-01 7.19725192e-01
3.46099794e-01 -2.91908264e-01 1.80617254e-02 6.30671144e-01
-4.22263026e-01 -1.08072937e+00 -3.96982163e-01 -2.14063115e-02
-9.73777920e-02 -5.39732993e-01 -4.00441170e-01 6.47900462e-01
1.75015539e-01 1.12806499e+00 -1.43516824e-01 -2.90778875e-01
-1.52713925e-01 2.39204273e-01 6.36529446e-01 -5.29598892e-01
-1.04555354e-01 2.86830038e-01 -2.03131199e-01 -5.37402391e-01
-7.30474651e-01 -4.27753955e-01 -1.35861516e+00 1.30347207e-01
-9.58533809e-02 -1.17891729e-01 5.93340158e-01 1.26689827e+00
3.53001505e-01 3.62133980e-01 4.22031075e-01 -8.31643522e-01
-2.53420621e-01 -9.79152441e-01 -2.51849681e-01 6.51003301e-01
4.08616096e-01 -4.95247841e-01 -1.27930000e-01 4.03435916e-01]
|
[10.819766998291016, 1.4585576057434082]
|
b2199e96-b2f8-4a95-9af3-f82be33ebc59
|
how-to-train-your-deep-neural-network-with
|
1612.07454
| null |
http://arxiv.org/abs/1612.07454v1
|
http://arxiv.org/pdf/1612.07454v1.pdf
|
How to Train Your Deep Neural Network with Dictionary Learning
|
Currently there are two predominant ways to train deep neural networks. The
first one uses restricted Boltzmann machine (RBM) and the second one
autoencoders. RBMs are stacked in layers to form deep belief network (DBN); the
final representation layer is attached to the target to complete the deep
neural network. Autoencoders are nested one inside the other to form stacked
autoencoders; once the stcaked autoencoder is learnt the decoder portion is
detached and the target attached to the deepest layer of the encoder to form
the deep neural network. This work proposes a new approach to train deep neural
networks using dictionary learning as the basic building block; the idea is to
use the features from the shallower layer as inputs for training the next
deeper layer. One can use any type of dictionary learning (unsupervised,
supervised, discriminative etc.) as basic units till the pre-final layer. In
the final layer one needs to use the label consistent dictionary learning
formulation for classification. We compare our proposed framework with existing
state-of-the-art deep learning techniques on benchmark problems; we are always
within the top 10 results. In actual problems of age and gender classification,
we are better than the best known techniques.
|
['Vanika Singhal', 'Shikha Singh', 'Angshul Majumdar']
|
2016-12-22
| null | null | null | null |
['age-and-gender-classification']
|
['computer-vision']
|
[-1.14540130e-01 3.65684360e-01 -1.82641238e-01 -4.09763277e-01
1.79019302e-01 -1.99362323e-01 4.93608505e-01 2.37024814e-01
-7.19267845e-01 8.56127620e-01 2.11248711e-01 -2.01622039e-01
1.60958976e-01 -1.15630877e+00 -7.20789611e-01 -1.04060435e+00
9.46027935e-02 8.10374439e-01 2.42489427e-01 -1.85043678e-01
1.01912044e-01 5.12491941e-01 -1.54833305e+00 2.33470485e-01
5.27491927e-01 1.12152863e+00 6.82397857e-02 5.08000791e-01
-4.14479434e-01 1.26343071e+00 -2.45164827e-01 -3.64224732e-01
1.18272863e-01 -1.92769155e-01 -7.75565147e-01 -3.49175930e-01
8.32339302e-02 -5.29292643e-01 -2.88816035e-01 8.87347043e-01
5.88323116e-01 3.60720605e-01 8.40050399e-01 -6.53523743e-01
-7.26123869e-01 7.01140881e-01 -2.64916688e-01 3.93578291e-01
-1.48675203e-01 -6.17947400e-01 7.73407519e-01 -1.02699840e+00
1.94102839e-01 1.07818973e+00 5.88542402e-01 7.88312197e-01
-8.96362543e-01 -7.06798613e-01 -6.21958263e-02 2.95098364e-01
-1.18793952e+00 -4.49891299e-01 6.49391115e-01 -4.47375566e-01
1.04236794e+00 -4.37901281e-02 8.28077018e-01 9.34832573e-01
4.59634542e-01 6.23685777e-01 1.26061761e+00 -6.24697506e-01
2.68552870e-01 5.01623273e-01 5.96712828e-01 1.05998492e+00
2.26749793e-01 1.17038362e-01 -1.93395928e-01 -1.95984185e-01
8.29361796e-01 3.78946930e-01 2.10487306e-01 -2.66646773e-01
-5.30812383e-01 1.22238171e+00 4.26553816e-01 6.76720679e-01
-5.86428523e-01 1.38231948e-01 3.90940338e-01 1.98156983e-01
3.82639885e-01 -9.89002138e-02 -3.83839458e-01 1.50321811e-01
-1.09011209e+00 8.45448151e-02 8.61725211e-01 4.24133331e-01
8.84905338e-01 2.27137491e-01 1.81301415e-01 1.23798370e+00
5.20287931e-01 -3.61838639e-02 9.67649639e-01 -2.90832192e-01
1.78291738e-01 6.37087047e-01 -3.96972656e-01 -7.79560447e-01
-2.76985943e-01 -6.16788924e-01 -1.07579589e+00 2.66373515e-01
1.39870539e-01 -3.68100524e-01 -1.07551491e+00 1.28080904e+00
2.54019707e-01 4.21173155e-01 2.50562429e-01 7.22616434e-01
1.09563625e+00 1.03653443e+00 1.08194679e-01 -4.11753505e-02
1.32968318e+00 -1.03856242e+00 -5.13482392e-01 -2.71400243e-01
4.59130466e-01 -4.39711422e-01 1.37851909e-01 8.76975298e-01
-1.10577738e+00 -7.90266931e-01 -1.40649605e+00 -1.89782754e-01
-8.40360343e-01 2.35858947e-01 7.37019241e-01 5.35641551e-01
-1.18156672e+00 6.95652246e-01 -7.18202055e-01 -3.43704641e-01
2.58893490e-01 7.16026366e-01 -5.09759426e-01 5.49383350e-02
-1.14178860e+00 1.23266625e+00 9.29468513e-01 1.81379482e-01
-1.39415312e+00 -7.22117200e-02 -8.76455426e-01 -8.56283978e-02
-3.18783194e-01 -5.49073040e-01 7.65325963e-01 -1.17647827e+00
-1.55947900e+00 9.34481919e-01 1.63769975e-01 -6.92191243e-01
-9.54732299e-02 -6.01919442e-02 -1.75219268e-01 -2.51422465e-01
-4.12957728e-01 7.16371119e-01 9.33678091e-01 -1.18268335e+00
-6.50431871e-01 -5.45509458e-01 2.83125192e-02 2.48189554e-01
-7.18330622e-01 -2.58301973e-01 -2.82286614e-01 -6.10448480e-01
1.95184842e-01 -8.13846111e-01 -9.69930291e-02 -7.27185667e-01
-1.45676762e-01 -5.87352455e-01 7.57058442e-01 -8.32445920e-01
1.50520051e+00 -2.15874290e+00 7.50413358e-01 3.09978127e-01
1.01125881e-01 2.53328621e-01 1.82748392e-01 2.73063630e-01
-4.27614003e-01 -4.66575891e-01 9.97309238e-02 -7.69671142e-01
-1.26942351e-01 8.62454355e-01 -1.38757050e-01 6.03220582e-01
-3.24901491e-01 4.28296089e-01 -4.64700401e-01 -6.03418171e-01
2.67471522e-01 4.53522533e-01 -4.91780549e-01 4.66274053e-01
1.79410324e-01 1.79632857e-01 -1.85799986e-01 4.17896181e-01
4.59279329e-01 1.51038364e-01 3.62442695e-02 -1.81614205e-01
-1.86415479e-01 2.90836871e-01 -1.29416728e+00 1.63070989e+00
-6.18592262e-01 4.66954172e-01 -7.32970163e-02 -1.69314003e+00
1.36310840e+00 3.94046217e-01 1.05437621e-01 -3.89829576e-01
4.17337447e-01 2.40015939e-01 -4.60932264e-03 -4.09463048e-01
4.08810109e-01 -5.22581339e-01 1.23424895e-01 2.27723103e-02
7.71022856e-01 3.82259727e-01 -2.09937364e-01 -2.08177716e-01
7.62843549e-01 -1.26735210e-01 3.92410249e-01 -2.94985294e-01
7.70001650e-01 -3.33202243e-01 3.83027166e-01 5.38735330e-01
1.63381189e-01 2.73213208e-01 1.92608297e-01 -9.21622992e-01
-1.06141973e+00 -9.13227975e-01 -2.76579022e-01 1.41680825e+00
-4.32399780e-01 -1.07025668e-01 -7.01987863e-01 -6.47092760e-01
1.25180054e-02 6.26167834e-01 -1.07403135e+00 -2.19858155e-01
-3.47342014e-01 -5.75005114e-01 5.76265097e-01 5.55206537e-01
6.30485833e-01 -1.24148333e+00 -3.57128799e-01 3.70068341e-01
4.88364846e-01 -6.45949304e-01 3.15591335e-01 1.01670527e+00
-1.24503088e+00 -5.51581860e-01 -8.05645466e-01 -1.16125369e+00
6.51249528e-01 -4.86451477e-01 1.01176667e+00 1.37658581e-01
-1.43106049e-02 -6.43376261e-02 -3.21110994e-01 -4.86724019e-01
-3.19976449e-01 1.41455099e-01 1.99639842e-01 1.11002266e-01
4.68092889e-01 -8.05214047e-01 -4.14339483e-01 -3.36094677e-01
-6.32606208e-01 -3.40837836e-02 7.30428040e-01 8.19557011e-01
5.10583341e-01 2.46058017e-01 2.31517658e-01 -9.98341024e-01
3.65999579e-01 -7.15366840e-01 -5.39866745e-01 -1.08778447e-01
-4.52971518e-01 2.65907556e-01 7.72591710e-01 -3.48615825e-01
-7.82341957e-01 1.21735327e-01 -6.19602799e-01 -4.38654453e-01
-2.29948238e-01 5.81542373e-01 8.66526067e-02 1.06625110e-01
4.52174634e-01 6.00373805e-01 -2.25730941e-01 -8.02699327e-01
1.72618389e-01 8.07260275e-01 4.39510733e-01 -5.76149344e-01
2.94873893e-01 1.55908525e-01 -3.18458229e-01 -5.73437810e-01
-7.90680170e-01 -3.59149486e-01 -7.48656869e-01 -1.20469950e-01
1.25873959e+00 -8.32217813e-01 -4.79515970e-01 5.28406203e-01
-1.16459417e+00 -1.50264949e-01 -1.30821854e-01 5.23519754e-01
-9.11619589e-02 2.96415761e-02 -8.78372967e-01 -8.83138061e-01
-5.36923885e-01 -1.16814494e+00 5.65741718e-01 2.93087780e-01
1.14118025e-01 -1.32351637e+00 4.76971179e-01 2.74091721e-01
1.76004559e-01 1.01800086e-02 1.08308542e+00 -1.15781415e+00
2.60852695e-01 -1.87459990e-01 2.02574730e-01 1.08623314e+00
-7.85334110e-02 -3.10482651e-01 -1.11556721e+00 -3.11966330e-01
2.69703567e-01 -5.37305713e-01 1.25470042e+00 4.35117006e-01
1.32387710e+00 -3.70277256e-01 -1.71488807e-01 5.62415779e-01
1.76341414e+00 4.63721395e-01 9.82528985e-01 4.60033983e-01
7.17365444e-01 3.21728408e-01 -2.90026274e-02 3.02580684e-01
4.85868782e-01 1.89309061e-01 3.72493774e-01 -1.08025178e-01
2.20034227e-01 -2.03196004e-01 5.08991897e-01 1.50261140e+00
-3.90931934e-01 4.52071726e-02 -8.11813116e-01 5.19253552e-01
-1.68592477e+00 -8.84529412e-01 2.11731076e-01 2.09069014e+00
9.15647686e-01 3.33781868e-01 8.07276592e-02 4.40543413e-01
3.08658034e-01 2.09257036e-01 -9.07091349e-02 -1.11849236e+00
2.40385547e-01 8.73666525e-01 5.48280776e-01 4.47121173e-01
-1.21873093e+00 9.01084721e-01 5.97218180e+00 9.18994486e-01
-1.17408693e+00 5.03077149e-01 5.07876396e-01 1.77471772e-01
2.95048840e-02 -9.07027945e-02 -1.17105293e+00 4.08888638e-01
1.13951218e+00 8.63404274e-01 5.53615808e-01 1.11350524e+00
-3.23415518e-01 -3.42032969e-01 -1.20428121e+00 1.19902658e+00
2.06902504e-01 -1.20367527e+00 1.27749085e-01 1.41039804e-01
6.32820368e-01 7.67488852e-02 -5.52651286e-02 7.85428405e-01
5.29961765e-01 -1.38387239e+00 4.81514305e-01 6.67068779e-01
3.16638470e-01 -1.11448014e+00 9.87852633e-01 6.31417632e-01
-9.74606276e-01 -1.19074889e-01 -9.49969828e-01 -3.02403212e-01
-4.73832309e-01 6.15801096e-01 -5.81706524e-01 3.72003824e-01
6.75213635e-01 7.86523402e-01 -2.95081645e-01 6.86046064e-01
7.17029348e-02 7.73470461e-01 -3.17683518e-01 -1.93454504e-01
5.24360120e-01 -3.69052321e-01 3.67293768e-02 1.32303476e+00
1.94081038e-01 2.31299251e-01 7.69150034e-02 4.04525846e-01
-1.95920959e-01 7.79276341e-02 -7.39815712e-01 1.04265034e-01
2.95299292e-01 1.31755519e+00 -4.63511676e-01 -6.58788621e-01
-4.84638125e-01 1.20124900e+00 6.91199780e-01 2.71658916e-02
-6.99383855e-01 -1.68085068e-01 2.57269442e-01 -1.99969456e-01
5.90454042e-01 -1.66300222e-01 -1.07499272e-01 -9.24366534e-01
-4.84165072e-01 -9.24933970e-01 6.18195295e-01 -5.46497583e-01
-1.22509170e+00 6.25598133e-01 -8.27136040e-02 -6.16101742e-01
-1.63751110e-01 -8.78512323e-01 -4.52482373e-01 9.23310697e-01
-1.35480809e+00 -1.00524759e+00 -1.04291379e-01 9.83949602e-01
5.21173835e-01 -7.50306129e-01 1.34730613e+00 6.52768373e-01
-6.78508103e-01 3.09032261e-01 3.22838843e-01 5.17815828e-01
2.55105257e-01 -1.35097563e+00 -4.82210279e-01 4.88991618e-01
3.01798850e-01 6.86881542e-01 3.85583431e-01 -4.82348353e-01
-1.27783108e+00 -6.99310899e-01 8.78338575e-01 -1.86670288e-01
4.10780817e-01 -4.94499326e-01 -7.25588858e-01 8.79746854e-01
5.64168394e-01 -2.12740991e-02 9.96147573e-01 2.51606375e-01
-1.75619692e-01 -3.32659900e-01 -1.01794314e+00 -5.75455278e-02
2.08214298e-01 -4.28635806e-01 -1.11972129e+00 -3.42896245e-02
1.60771817e-01 -4.43300009e-01 -1.10383475e+00 6.67563230e-02
7.24374533e-01 -9.38268065e-01 8.63033414e-01 -6.17412448e-01
5.82322717e-01 -1.30272582e-01 -2.65800685e-01 -1.28203559e+00
-5.05959988e-01 2.46033415e-01 -2.97875613e-01 1.17819321e+00
3.79011720e-01 -4.48968053e-01 8.02458048e-01 1.49169371e-01
-5.20763248e-02 -1.13794982e+00 -9.08390403e-01 -2.95248330e-01
2.84718513e-01 -2.66436100e-01 3.22019339e-01 1.11454070e+00
-1.88862875e-01 4.38571990e-01 -5.89261949e-01 -1.42185036e-02
4.73321199e-01 -3.41578275e-01 3.52928132e-01 -1.34067082e+00
-5.81555486e-01 -1.18350893e-01 -5.24862349e-01 -1.09143412e+00
1.35117367e-01 -1.01202190e+00 -2.46448234e-01 -1.79033589e+00
2.98662186e-01 -5.32718897e-01 -7.53276825e-01 7.21106231e-01
2.43928790e-01 1.00106560e-01 -1.81746855e-01 -1.03582172e-02
-3.01565468e-01 3.05530101e-01 9.00545120e-01 -1.62934512e-01
-1.17283044e-02 -1.59483925e-01 -5.04063606e-01 9.10827756e-01
7.59701788e-01 -6.90467536e-01 -4.05230761e-01 -4.61463422e-01
-3.38367112e-02 -1.23809792e-01 2.55792290e-01 -1.25532746e+00
4.17838395e-01 1.76039815e-01 1.01033533e+00 -7.13896990e-01
4.39553916e-01 -1.08812404e+00 1.40561640e-01 5.66731453e-01
-1.91043630e-01 1.35428265e-01 1.89512238e-01 1.88365400e-01
-1.51901245e-01 -8.81565809e-01 9.05473590e-01 -4.05957073e-01
-7.65784025e-01 2.59313136e-01 -4.96236324e-01 -3.80905211e-01
8.22029650e-01 -2.03141212e-01 2.32869983e-01 4.29385938e-02
-1.22746181e+00 -1.74821720e-01 -2.92892102e-02 3.46933007e-01
6.73016667e-01 -1.37977123e+00 -4.16196465e-01 2.29677781e-01
-4.64345783e-01 9.59659070e-02 -6.54271767e-02 5.36868811e-01
-5.97690225e-01 3.31024319e-01 -4.73612964e-01 -3.76392990e-01
-1.02147889e+00 5.80057144e-01 5.91669798e-01 -5.37826955e-01
-2.90691972e-01 1.21045244e+00 -4.55583297e-02 -3.93945783e-01
6.29949629e-01 -3.29367101e-01 -7.83122838e-01 2.24937811e-01
3.60758275e-01 1.10489048e-01 1.48724332e-01 -4.95714039e-01
-4.11832869e-01 3.10346544e-01 -2.33529031e-01 -1.19974576e-01
1.76005471e+00 3.79391581e-01 -5.96524477e-01 6.40100360e-01
1.53600407e+00 -1.91983417e-01 -4.38278645e-01 -1.95828211e-02
-2.74988383e-01 2.11957946e-01 4.76921648e-01 -5.90466976e-01
-1.27489030e+00 9.67374146e-01 9.13830221e-01 1.19653665e-01
1.07703650e+00 -1.58704489e-01 7.02358127e-01 4.14043754e-01
3.22189927e-01 -1.21670055e+00 -1.09987624e-01 7.28658676e-01
5.54116368e-01 -9.55918133e-01 3.09357583e-03 4.31449801e-01
-3.82161885e-01 1.42615068e+00 5.87310731e-01 -7.77673841e-01
1.08619523e+00 4.61918265e-01 -3.88106436e-01 -2.21257359e-01
-6.34083450e-01 5.32781817e-02 4.07288522e-01 2.77626067e-01
6.92284286e-01 -1.93012469e-02 -3.63282651e-01 1.05952036e+00
-1.90667912e-01 1.30306080e-01 -5.53390384e-02 6.54484153e-01
-7.31584311e-01 -1.30635202e+00 -3.99318248e-01 6.20657682e-01
-5.49853563e-01 -2.26509392e-01 -1.58438697e-01 5.58174014e-01
9.12619233e-01 5.46510875e-01 2.51151532e-01 -5.87650537e-01
2.05071107e-01 2.69112259e-01 8.19914520e-01 -8.49937499e-01
-8.68603945e-01 -1.35459244e-01 1.22750722e-01 -2.23485813e-01
-4.75855708e-01 -3.62587422e-01 -1.33631170e+00 -3.61413777e-01
-2.58477718e-01 4.77058262e-01 6.95292950e-01 1.11173260e+00
-2.71075666e-01 3.20108443e-01 4.02105570e-01 -7.55811155e-01
-2.81956613e-01 -1.05308950e+00 -6.49883270e-01 1.18809551e-01
3.86707515e-01 -9.24851477e-01 -1.18202269e-01 1.21510975e-01]
|
[9.207630157470703, 2.937370538711548]
|
b6f9cab4-a3ad-4ed3-9d8f-88daf89eff9b
|
recommending-complementary-products-in-e
|
1707.08113
| null |
https://arxiv.org/abs/1707.08113v1
|
https://arxiv.org/pdf/1707.08113v1.pdf
|
Recommending Complementary Products in E-Commerce Push Notifications with a Mixture Model Approach
|
Push notification is a key component for E-commerce mobile applications, which has been extensively used for user growth and engagement. The effectiveness of the push notification is generally measured by message open rate. A push message can contain a recommended product, a shopping news and etc., but often only one or two items can be shown in the push message due to the limit of display space. This paper proposes a mixture model approach for predicting push message open rate for a post-purchase complementary product recommendation task. The mixture model is trained to learn latent prediction contexts, which are determined by user and item profiles, and then make open rate predictions accordingly. The item with the highest predicted open rate is then chosen to be included in the push notification message for each user. The parameters of the mixture model are optimized using an EM algorithm. A set of experiments are conducted to evaluate the proposed method live with a popular E-Commerce mobile app. The results show that the proposed method is superior than several existing solutions by a significant margin.
|
['Qiong Zhang', 'Luo Si', 'Xiaogang Li', 'Huasha Zhao']
|
2017-07-25
| null | null | null | null |
['product-recommendation']
|
['miscellaneous']
|
[ 2.71491885e-01 -1.20866656e-01 -8.75251472e-01 -3.33326340e-01
-5.73530018e-01 -1.84376866e-01 3.65971804e-01 3.43580276e-01
-2.45820358e-01 3.05111229e-01 4.09780592e-01 -4.10368294e-01
-4.32088137e-01 -7.68062055e-01 -2.47574046e-01 -6.52778625e-01
1.33395866e-01 4.15112436e-01 3.52547407e-01 -6.12675175e-02
4.54027444e-01 -2.59980470e-01 -1.77164304e+00 7.24757791e-01
8.27147543e-01 1.07474375e+00 8.47423315e-01 8.02731216e-01
-5.70306361e-01 6.73484027e-01 -2.11233959e-01 -3.85899186e-01
-9.27229039e-03 -3.84967119e-01 -3.51821989e-01 5.32581657e-02
-2.71859884e-01 -4.52894479e-01 9.63204727e-02 7.12461591e-01
4.21025753e-01 2.91270763e-01 6.11486495e-01 -1.09848022e+00
-3.11024219e-01 8.57128501e-01 -5.38429797e-01 4.03424948e-01
4.29143190e-01 -6.71296060e-01 9.93562579e-01 -7.36274719e-01
4.49993610e-01 9.57508028e-01 4.18992996e-01 2.06364051e-01
-1.09700286e+00 -7.54682362e-01 3.29476297e-01 -2.50827130e-02
-1.19153619e+00 -1.56791583e-01 7.19989717e-01 -5.16200066e-01
5.63291728e-01 6.10123396e-01 6.36907220e-01 8.77595544e-01
4.48647618e-01 9.28263485e-01 8.04190099e-01 -2.81243742e-01
4.52587277e-01 7.28642344e-01 4.98656988e-01 -3.94754764e-03
-1.06387079e-01 -4.17295724e-01 -4.02658075e-01 -6.76840961e-01
5.94481051e-01 5.67846835e-01 9.29219872e-02 -2.07507357e-01
-8.11442971e-01 8.63397121e-01 -8.01310092e-02 1.77390486e-01
-6.87651396e-01 -4.06035691e-01 1.15600117e-01 2.09094197e-01
6.61213279e-01 -3.15735251e-01 -5.74073851e-01 -6.79282188e-01
-7.99776912e-01 3.92922282e-01 1.01095748e+00 8.82950306e-01
2.89136469e-01 -6.45922065e-01 -1.13820650e-01 1.23964334e+00
9.63208318e-01 4.28524107e-01 5.28594792e-01 -4.87415612e-01
5.92414856e-01 5.42895019e-01 4.40448374e-01 -1.03901589e+00
-2.24281609e-01 -4.29685861e-01 -3.98988008e-01 -2.57524252e-01
1.83601424e-01 -2.46844366e-01 -3.50868076e-01 1.24671900e+00
5.05173028e-01 1.91682905e-01 -1.37122989e-01 5.90446293e-01
6.12887561e-01 1.20587802e+00 3.70264679e-01 -9.49529767e-01
1.31254160e+00 -6.62464440e-01 -9.75175083e-01 7.95530528e-02
6.69045627e-01 -1.21147120e+00 9.58053648e-01 5.87513447e-01
-9.07476187e-01 -7.73472190e-01 -9.07431781e-01 5.00974894e-01
6.75593386e-04 2.11376399e-01 6.11489713e-01 9.48326290e-01
-3.14499140e-01 5.28229415e-01 -6.63925469e-01 -5.13130605e-01
5.82809336e-02 4.83482361e-01 3.56339335e-01 1.17018476e-01
-9.71307814e-01 7.45577887e-02 1.39191508e-01 -3.44557524e-01
-7.37582967e-02 -5.43311238e-01 -2.98861474e-01 1.28810510e-01
8.41868147e-02 -2.76565313e-01 1.52202117e+00 -8.24973464e-01
-1.69734943e+00 2.22859442e-01 -2.02559233e-01 -1.10894226e-01
3.59272242e-01 -3.33009869e-01 -6.83756411e-01 -4.07263666e-01
-1.78677797e-01 2.22104982e-01 9.06704903e-01 -9.31936920e-01
-1.27724147e+00 -3.01477134e-01 5.60115501e-02 5.57680249e-01
-5.56143105e-01 3.78637433e-01 -4.70099807e-01 -3.57382298e-01
3.50798577e-01 -9.77270842e-01 -1.18986657e-02 -8.51822674e-01
6.41701519e-02 -4.66020495e-01 9.31118369e-01 -8.19328606e-01
1.86120272e+00 -2.22121811e+00 -3.15180242e-01 3.90792608e-01
-1.41015336e-01 -9.06162038e-02 3.57793182e-01 5.08002937e-01
3.19226205e-01 -9.08695757e-02 4.51723367e-01 -2.06410989e-01
-1.15596438e-02 -1.39573812e-02 -3.85974139e-01 -2.09381580e-02
-7.19424844e-01 3.82343322e-01 -5.86629868e-01 -2.74578333e-01
6.39117584e-02 4.23744589e-01 -5.79682350e-01 2.32386008e-01
-1.10996678e-01 2.29507178e-01 -5.29024899e-01 2.52973914e-01
8.21229517e-01 -5.20155966e-01 4.77784932e-01 1.76740304e-01
-2.69187808e-01 4.35002893e-01 -1.59190822e+00 1.05737650e+00
-4.55884606e-01 1.89928170e-02 -1.32783070e-01 -2.74366558e-01
1.09604967e+00 3.97148132e-01 6.57170892e-01 -5.67366064e-01
3.04354399e-01 1.23106219e-01 5.88669106e-02 -5.30174017e-01
8.02374005e-01 2.41442025e-01 1.36603147e-01 6.39811337e-01
-4.63258117e-01 7.61503637e-01 2.04385236e-01 2.43622914e-01
6.04119897e-01 3.68427299e-02 1.97705105e-01 -1.63510859e-01
3.69060934e-01 -5.20224392e-01 5.96833885e-01 7.74601638e-01
-2.09544729e-02 4.20132428e-02 3.08003187e-01 -1.62802562e-01
-9.00444031e-01 -8.08670104e-01 -2.70146489e-01 1.74020803e+00
3.23134392e-01 -7.73890376e-01 -7.31793582e-01 -6.33985996e-01
-1.61908492e-01 7.17215240e-01 -3.78009796e-01 1.33653596e-01
-2.10427150e-01 -7.17393160e-01 -4.19990450e-01 3.26310694e-01
3.09332758e-01 -9.16508317e-01 -2.48827875e-01 4.78341281e-01
-3.66057038e-01 -6.90315545e-01 -5.81504762e-01 -2.28731900e-01
-1.17165148e+00 -5.67511320e-01 -7.53089905e-01 -9.33915854e-01
4.87712979e-01 5.43443620e-01 4.21872854e-01 -5.54945804e-02
4.58818138e-01 2.36759827e-01 -6.26361012e-01 -4.08709168e-01
-5.50111234e-01 3.30136538e-01 1.85516715e-01 6.74785435e-01
6.00258648e-01 -4.06767040e-01 -7.91504860e-01 8.60448480e-01
-7.20331073e-01 3.80792499e-01 3.65227371e-01 4.49634552e-01
4.60852891e-01 2.00708121e-01 8.10737908e-01 -1.23022807e+00
1.10636306e+00 -8.90608370e-01 -3.20145398e-01 -1.12676069e-01
-8.69155109e-01 -5.71774662e-01 1.51744783e-01 -9.84582186e-01
-1.39831066e+00 -9.72242728e-02 -3.74305815e-01 5.00420749e-01
-1.82603836e-01 6.60251558e-01 -2.15308145e-01 4.13299471e-01
3.56617421e-01 -5.28286546e-02 -2.97219366e-01 -9.05413926e-01
8.08205903e-02 1.43704796e+00 -1.65875122e-01 1.36804521e-01
-1.09335557e-01 -4.31826860e-02 -6.98131025e-01 -9.00388658e-01
-3.71561080e-01 -9.54550207e-01 -1.65704355e-01 -4.86570925e-01
5.46963096e-01 -6.67461693e-01 -8.21332276e-01 2.14036152e-01
-7.24336207e-01 3.70480330e-03 2.38971844e-01 9.64948595e-01
-4.48353052e-01 1.25367969e-01 -9.25556540e-01 -1.25161541e+00
-2.25090623e-01 -9.60114658e-01 6.75102234e-01 4.50604647e-01
-5.24712265e-01 -7.31869280e-01 1.71759993e-01 5.53128779e-01
3.86193961e-01 -5.59773028e-01 9.66415584e-01 -9.39282417e-01
-3.18836063e-01 -7.34561384e-01 1.19944423e-01 -1.27638653e-01
1.59495190e-01 -9.89000872e-02 -6.88655019e-01 -1.86778352e-01
1.51537523e-01 3.17662865e-01 1.52110055e-01 7.92110682e-01
8.56962681e-01 -3.81458163e-01 -7.53549635e-01 8.77287164e-02
1.07604003e+00 8.25637043e-01 8.16649616e-01 1.53436273e-01
3.10574412e-01 4.28949744e-01 1.02559352e+00 7.84375191e-01
3.44260126e-01 8.54399562e-01 1.41098559e-01 2.60678053e-01
5.02704084e-01 -6.76504970e-01 3.37968200e-01 1.33298969e+00
-1.12648755e-01 -3.66244823e-01 -4.15854216e-01 -4.07206779e-03
-2.13242912e+00 -8.58662009e-01 -4.31963384e-01 2.43742347e+00
4.61154848e-01 4.20152843e-01 5.05533040e-01 4.30836856e-01
8.17507625e-01 -2.64714599e-01 -2.22224802e-01 -2.80332357e-01
3.88751596e-01 -3.78512502e-01 3.49808276e-01 5.70335090e-01
-1.04410708e+00 4.35454339e-01 6.01752424e+00 1.02280426e+00
-7.73904264e-01 3.71079028e-01 7.95912325e-01 -1.21746361e-01
-2.94813752e-01 -2.89726872e-02 -1.18733704e+00 9.02843714e-01
1.05556178e+00 -3.96112241e-02 1.66174456e-01 9.79279339e-01
5.55947900e-01 -3.65834177e-01 -8.72686803e-01 1.24564445e+00
-5.31886332e-02 -1.07300973e+00 -2.74511337e-01 5.81052542e-01
7.83056200e-01 -3.88715863e-01 1.81431487e-01 4.79826301e-01
-2.52739251e-01 -1.65564552e-01 4.16297227e-01 5.36991775e-01
1.53767571e-01 -8.08919251e-01 7.61254728e-01 9.41689610e-01
-1.01679015e+00 -4.37975198e-01 -3.04761410e-01 -3.75205100e-01
4.24802959e-01 4.72196579e-01 -5.52317798e-01 -6.49434179e-02
6.06714725e-01 3.12776387e-01 7.29258657e-02 1.16991568e+00
3.80286932e-01 8.97662520e-01 -3.48454058e-01 -5.66535473e-01
-2.71295726e-01 -4.51422870e-01 3.35240155e-01 1.06554592e+00
7.38603771e-01 2.39242643e-01 2.04201549e-01 3.74686003e-01
-1.57889239e-02 1.02107656e+00 -2.64214039e-01 1.59275923e-02
3.03573698e-01 1.28710067e+00 -1.04113495e+00 -4.04619217e-01
-5.34551859e-01 8.10108185e-01 -2.57717699e-01 2.12342620e-01
-8.56107295e-01 -1.06062166e-01 1.19356990e-01 6.90522432e-01
3.31824064e-01 1.69824675e-01 -1.60771206e-01 -5.94229281e-01
-3.50137912e-02 -5.29579937e-01 2.76852518e-01 -5.99218726e-01
-8.27272296e-01 2.76025534e-01 -7.01583363e-03 -1.32850409e+00
-1.17914796e-01 -2.05932006e-01 -4.81914699e-01 8.35464299e-01
-6.05431974e-01 -5.05002618e-01 -1.67226627e-01 1.50383353e-01
9.09246564e-01 -2.63610721e-01 7.45350838e-01 7.16120303e-01
-4.07449871e-01 5.34827650e-01 5.05629897e-01 -5.81349075e-01
4.18153495e-01 -7.11567879e-01 2.11715922e-01 2.42654607e-01
3.76896150e-02 8.22392941e-01 8.09669197e-01 -1.11751723e+00
-1.13344789e+00 -5.96680820e-01 1.27284884e+00 -4.33901131e-01
1.63574606e-01 -4.09604043e-01 -8.06117952e-01 3.72985929e-01
3.60091850e-02 -5.27923048e-01 1.22825503e+00 4.75593477e-01
2.69935608e-01 -1.61263779e-01 -9.78543758e-01 5.51366985e-01
5.90326130e-01 -1.22177295e-01 -1.57029018e-01 2.82462686e-01
6.08877599e-01 -1.42192885e-01 -8.67121577e-01 1.66200727e-01
8.62516284e-01 -8.69532883e-01 5.79100728e-01 -4.81382638e-01
2.39470199e-01 -4.64465059e-02 -2.04395518e-01 -9.89092827e-01
-6.88204467e-01 -6.01303756e-01 -6.99692130e-01 1.48396933e+00
6.68760240e-01 -3.34335864e-01 1.05243671e+00 9.04921353e-01
4.14418250e-01 -9.17402148e-01 -3.38573784e-01 -3.65059376e-01
-8.13093662e-01 -5.00491261e-01 5.54463506e-01 5.32809317e-01
5.75457096e-01 8.14138055e-01 -7.91187823e-01 -4.74439897e-02
3.16530406e-01 -2.20241249e-02 8.20011675e-01 -1.38872802e+00
-6.87969267e-01 -2.24539444e-01 -4.67949957e-02 -1.62766278e+00
-7.50746548e-01 -5.79751611e-01 -1.15684815e-01 -1.41861248e+00
4.63358164e-01 -5.80123484e-01 -5.57486951e-01 -3.23695898e-01
-1.27163365e-01 -4.56115343e-02 1.13273738e-02 5.23270011e-01
-7.43700206e-01 1.19576357e-01 7.44921803e-01 1.73356488e-01
-1.05705380e+00 1.02174962e+00 -6.03539348e-01 7.78101027e-01
9.32867765e-01 -4.73643363e-01 -8.16079557e-01 2.77045280e-01
5.65826118e-01 3.29332471e-01 -5.08525193e-01 -5.52229047e-01
1.49832830e-01 -2.77650565e-01 3.22791785e-01 -6.99631810e-01
2.30289757e-01 -8.59065413e-01 5.10393322e-01 2.83792585e-01
-7.51438797e-01 -1.07282177e-02 -1.72225147e-01 8.20882142e-01
3.06620389e-01 -3.99937361e-01 3.18720102e-01 4.21913296e-01
-3.18804562e-01 2.63199925e-01 -5.88951170e-01 -6.90853715e-01
9.91145551e-01 -2.77274013e-01 2.03997254e-01 -7.12303460e-01
-8.94530475e-01 -1.44669876e-01 -7.07838982e-02 8.82849813e-01
4.89307582e-01 -1.53901947e+00 -3.30419064e-01 3.30276936e-01
2.76817799e-01 -8.30699146e-01 5.47654212e-01 9.52464879e-01
-1.35542780e-01 1.91624105e-01 9.22124833e-02 -6.01901650e-01
-1.74545109e+00 5.76957524e-01 -3.76929998e-01 -3.34028125e-01
-4.75582153e-01 6.16677701e-01 9.63099971e-02 -6.14267290e-02
5.36944687e-01 -2.50796825e-01 -7.02673793e-01 -1.05703138e-01
1.12622297e+00 5.99976182e-01 -7.46546090e-02 -6.26210213e-01
5.68911545e-02 1.90185189e-01 -4.35868531e-01 -2.50378788e-01
8.92379820e-01 -6.51272058e-01 3.36986214e-01 7.53571630e-01
9.44234371e-01 3.23844045e-01 -1.09117961e+00 -3.02136838e-01
-6.83042407e-02 -7.98291802e-01 1.87157288e-01 -6.01837993e-01
-5.29389083e-01 2.94159323e-01 1.05816770e+00 6.79529727e-01
8.15129876e-01 -1.12888170e-02 9.97223675e-01 -4.62454595e-02
4.29513931e-01 -1.36399329e+00 -2.19445974e-01 1.54883012e-01
3.90868515e-01 -1.13122451e+00 -3.69391263e-01 -5.90762913e-01
-7.31662989e-01 6.00600660e-01 4.08251733e-01 2.32930169e-01
1.04021704e+00 2.68345624e-01 1.10357992e-01 2.74016201e-01
-7.26424098e-01 2.90071338e-01 1.73318654e-01 3.02690476e-01
8.23751390e-01 3.65971029e-01 -1.02166557e+00 1.03682137e+00
1.24427177e-01 3.25572819e-01 1.42886445e-01 8.77965152e-01
-8.68830562e-01 -1.35360110e+00 -2.91529477e-01 9.24780726e-01
-7.20486581e-01 2.83568203e-01 1.45347580e-01 -4.81827967e-02
1.56749889e-01 1.31480706e+00 -3.78650129e-02 -8.01821530e-01
2.56856203e-01 1.61746666e-01 8.89072642e-02 -3.94722015e-01
-4.28806126e-01 7.74294078e-01 3.72491637e-03 -1.32596180e-01
-1.18303269e-01 -7.78668404e-01 -1.09542668e+00 -4.55106735e-01
-5.75236738e-01 5.80716252e-01 9.82226074e-01 6.72389150e-01
4.09460992e-01 3.59817207e-01 7.20118821e-01 -5.41255891e-01
-4.30049896e-01 -1.13787472e+00 -8.41745257e-01 5.65425277e-01
-3.11905086e-01 -5.96881688e-01 4.27102521e-02 1.07336514e-01]
|
[10.068685531616211, 5.815301418304443]
|
1f415d26-a47d-4409-a4b6-86b532f02a49
|
instance-based-model-adaptation-for-direct
|
1910.10663
| null |
https://arxiv.org/abs/1910.10663v1
|
https://arxiv.org/pdf/1910.10663v1.pdf
|
Instance-Based Model Adaptation For Direct Speech Translation
|
Despite recent technology advancements, the effectiveness of neural approaches to end-to-end speech-to-text translation is still limited by the paucity of publicly available training corpora. We tackle this limitation with a method to improve data exploitation and boost the system's performance at inference time. Our approach allows us to customize "on the fly" an existing model to each incoming translation request. At its core, it exploits an instance selection procedure to retrieve, from a given pool of data, a small set of samples similar to the input query in terms of latent properties of its audio signal. The retrieved samples are then used for an instance-specific fine-tuning of the model. We evaluate our approach in three different scenarios. In all data conditions (different languages, in/out-of-domain adaptation), our instance-based adaptation yields coherent performance gains over static models.
|
['Viet-Nhat Nguyen', 'Mattia Antonino Di Gangi', 'Marco Turchi', 'Matteo Negri']
|
2019-10-23
| null | null | null | null |
['speech-to-text-translation']
|
['natural-language-processing']
|
[ 3.61268073e-01 -8.06282908e-02 -2.98412800e-01 -5.12286901e-01
-1.52001762e+00 -7.32670426e-01 7.00190961e-01 -1.57260839e-02
-3.66896689e-01 7.24255979e-01 3.72642994e-01 -4.69026029e-01
2.33600423e-01 -3.07437271e-01 -8.01431298e-01 -4.77504551e-01
4.01391476e-01 8.14057052e-01 1.19776025e-01 -3.02766562e-01
3.22334431e-02 3.90109897e-01 -1.38529456e+00 6.32228076e-01
7.03983128e-01 7.20206261e-01 4.07273263e-01 8.37779403e-01
-1.86946213e-01 2.50551105e-01 -7.68569946e-01 -3.33182216e-01
2.04260260e-01 -5.62687635e-01 -7.18646824e-01 9.10584107e-02
2.17565715e-01 -1.57479852e-01 -6.66019917e-02 7.50828266e-01
1.00271976e+00 2.18464434e-01 3.70199800e-01 -6.39517307e-01
-1.45854563e-01 7.51064539e-01 -8.63107517e-02 4.10086870e-01
4.63581353e-01 2.03583762e-01 8.26918721e-01 -1.20581639e+00
7.69873619e-01 1.05154061e+00 4.89278406e-01 7.63627887e-01
-1.54600751e+00 -5.85894644e-01 2.32131273e-01 2.48182584e-02
-1.15913177e+00 -1.30643022e+00 6.47484601e-01 -4.67889011e-02
1.27523720e+00 3.28400433e-01 3.78986537e-01 1.42318404e+00
-1.32807925e-01 7.28994727e-01 9.22366679e-01 -6.90582216e-01
4.20185834e-01 3.58401299e-01 -3.46598238e-01 2.26853117e-01
-4.64399934e-01 9.13665295e-02 -1.09968901e+00 -3.17368180e-01
3.43238503e-01 -7.08428085e-01 -2.03749418e-01 -9.47409347e-02
-1.17997909e+00 5.48226297e-01 -1.10414699e-01 4.66055036e-01
-5.29402494e-01 -2.35281941e-02 5.15994370e-01 6.74310803e-01
7.44849503e-01 5.31202972e-01 -8.92527640e-01 -5.17597795e-01
-1.37360048e+00 3.61451715e-01 1.11265194e+00 8.75992835e-01
6.69388413e-01 1.74027070e-01 -4.28777248e-01 1.16343999e+00
5.03667705e-02 5.96693039e-01 8.22461545e-01 -6.53445184e-01
9.19020951e-01 3.88175063e-02 2.77835112e-02 -4.72804487e-01
-8.41728225e-02 -7.99272597e-01 -3.46046478e-01 -2.19313398e-01
2.85446286e-01 -4.34469938e-01 -9.01813209e-01 1.97644436e+00
4.27466720e-01 1.85551476e-02 1.42404601e-01 8.26062024e-01
2.60337204e-01 8.26380193e-01 -9.88441557e-02 -4.44613636e-01
1.05649686e+00 -8.89790893e-01 -6.55653059e-01 -4.35864151e-01
4.53026921e-01 -1.07804215e+00 1.20131195e+00 3.76856893e-01
-1.24649215e+00 -6.29878402e-01 -7.88603902e-01 1.86519772e-01
-9.91219655e-02 2.93915302e-01 1.08472325e-01 6.18293703e-01
-1.17424905e+00 4.82003033e-01 -7.44341552e-01 -4.92960423e-01
4.08692583e-02 5.92434704e-01 -1.33353472e-01 1.43464431e-01
-1.17791283e+00 8.60145807e-01 3.47015858e-01 -6.34287745e-02
-1.01123857e+00 -5.55008292e-01 -3.68593514e-01 -5.10983542e-02
4.43741113e-01 -9.22102451e-01 1.69439423e+00 -1.48439646e+00
-2.01671195e+00 6.80777073e-01 -5.84334970e-01 -5.80229521e-01
4.40210879e-01 -1.83577314e-01 -4.61372763e-01 1.13405928e-01
-1.32713899e-01 4.57218707e-01 1.26335788e+00 -1.13905442e+00
-6.53217196e-01 -1.98104650e-01 -1.31026372e-01 3.67831141e-01
-2.92000234e-01 3.11678588e-01 -7.20836461e-01 -6.65209532e-01
-8.13778304e-03 -9.30679083e-01 -1.07725494e-01 -4.80966568e-01
-3.94924700e-01 5.82372956e-02 5.60422838e-01 -7.43613958e-01
1.30860937e+00 -1.96416283e+00 3.69952738e-01 1.95788994e-01
-4.02493179e-01 3.96040946e-01 -3.18649054e-01 7.17029750e-01
2.25916468e-02 -5.87229393e-02 -1.16047308e-01 -6.93492949e-01
-1.02917954e-01 1.66922942e-01 -5.01174092e-01 1.49681166e-01
3.75342757e-01 7.17445612e-01 -8.53287935e-01 -4.64740783e-01
-3.23177464e-02 3.75099152e-01 -6.72020257e-01 6.18350744e-01
-5.40465653e-01 7.38364637e-01 -4.46199179e-01 4.34919059e-01
1.07187711e-01 1.60777196e-01 2.24416018e-01 -8.97544995e-03
2.77626403e-02 8.09554517e-01 -9.74237442e-01 1.99361992e+00
-7.77757883e-01 6.54957414e-01 9.50786322e-02 -9.55616117e-01
9.09193933e-01 7.15209782e-01 2.81893641e-01 -7.66176522e-01
-2.39478803e-04 4.94404376e-01 7.54194856e-02 -5.95816672e-01
5.03397405e-01 -1.71401069e-01 3.82540189e-02 4.74658430e-01
2.82910109e-01 -2.04131193e-03 9.60530415e-02 -5.34945317e-02
1.00866628e+00 3.58593553e-01 1.63310006e-01 -7.15392008e-02
4.25946891e-01 -8.79629478e-02 3.69893938e-01 6.80924594e-01
7.41632804e-02 7.11427569e-01 8.76637921e-02 -2.16420799e-01
-1.26097000e+00 -9.41143215e-01 8.25008601e-02 1.41634607e+00
-6.05109870e-01 -2.56627917e-01 -9.94138181e-01 -5.37021816e-01
-3.01765740e-01 7.71088004e-01 -2.26733506e-01 -1.88076153e-01
-8.65323722e-01 -6.75800443e-01 7.35359907e-01 1.78440183e-01
1.81303188e-01 -1.18347669e+00 -5.24258673e-01 4.95208532e-01
-5.01427233e-01 -1.05524063e+00 -5.46761513e-01 3.64219278e-01
-1.04894769e+00 -2.88996816e-01 -7.25352049e-01 -4.69469905e-01
4.32759553e-01 -3.96358641e-03 1.30185652e+00 -2.46330470e-01
2.48494923e-01 1.03142425e-01 -4.00339425e-01 -2.73189664e-01
-1.00420856e+00 6.61514461e-01 1.60995334e-01 2.80809551e-01
2.18209907e-01 -5.95100999e-01 -3.04868460e-01 2.13965222e-01
-7.00789869e-01 -3.25904824e-02 7.13789701e-01 9.38935339e-01
5.63845098e-01 -3.62419844e-01 8.93226862e-01 -7.39969671e-01
9.78678107e-01 -5.42836964e-01 -4.02714431e-01 1.00510158e-01
-5.52743137e-01 2.29914397e-01 8.02643955e-01 -7.69398212e-01
-1.05040383e+00 1.48175389e-01 -3.19305897e-01 -4.71643627e-01
-1.90112367e-01 8.00934017e-01 -1.79746747e-01 4.03728455e-01
8.20121706e-01 5.03661335e-01 -6.56644702e-02 -7.14883149e-01
3.34048510e-01 1.08246624e+00 4.63548213e-01 -8.17918301e-01
5.89062810e-01 -6.36783317e-02 -5.14319241e-01 -7.41277277e-01
-5.90842485e-01 -3.93944472e-01 -6.56789541e-01 -8.23628902e-02
3.84553492e-01 -8.62417877e-01 -2.48343870e-01 1.93442494e-01
-1.16547906e+00 -5.47397375e-01 -3.28778684e-01 5.04812658e-01
-7.61515796e-01 -3.80450152e-02 -2.67643452e-01 -9.81334090e-01
-4.51093435e-01 -1.10493207e+00 1.22962904e+00 -1.77736431e-01
-4.13996100e-01 -7.86864102e-01 2.35500529e-01 3.13308269e-01
6.52610302e-01 -3.21684957e-01 9.29917336e-01 -1.12587106e+00
-3.85891378e-01 -3.30391049e-01 2.92952329e-01 4.62703168e-01
3.67606394e-02 -5.51086925e-02 -1.36838675e+00 -4.12323296e-01
3.19039933e-02 -1.78595498e-01 4.67029601e-01 1.99428171e-01
8.36943030e-01 -5.04950047e-01 -2.16055378e-01 4.55677301e-01
1.00526690e+00 1.17864229e-01 3.57305169e-01 2.11789578e-01
2.64289141e-01 4.66890991e-01 4.88324046e-01 4.40486729e-01
2.01173089e-02 1.22232389e+00 9.34478045e-02 9.76007804e-02
-2.57942677e-01 -2.49952346e-01 6.78072214e-01 1.12504089e+00
1.11875139e-01 -3.47139806e-01 -7.73296058e-01 7.86835372e-01
-1.64431369e+00 -8.69382143e-01 5.42374194e-01 2.49393773e+00
1.17530358e+00 1.30175829e-01 4.38871235e-01 -4.40278091e-02
5.89577019e-01 -3.09704989e-02 -5.94313085e-01 -5.24875760e-01
1.07379980e-01 3.96051198e-01 2.55336344e-01 5.38632333e-01
-7.24981368e-01 1.10037589e+00 6.64351463e+00 9.27258193e-01
-1.37344456e+00 2.73652703e-01 4.34971124e-01 -3.86531085e-01
-2.19278604e-01 6.74552470e-02 -9.37977791e-01 5.12182593e-01
1.63997173e+00 -1.80539846e-01 8.23727310e-01 4.62219149e-01
4.81346399e-01 1.85479403e-01 -1.31430340e+00 6.27126336e-01
5.32167070e-02 -1.17963564e+00 3.89142595e-02 -6.76392391e-02
4.82613236e-01 2.95611084e-01 9.95226800e-02 3.52860957e-01
-5.51081449e-03 -7.58883536e-01 8.61000299e-01 3.79712969e-01
9.23928022e-01 -7.15533018e-01 4.71534610e-01 6.94846153e-01
-7.89831519e-01 -6.60726568e-03 -1.43785715e-01 1.54301062e-01
7.63577223e-02 4.50598508e-01 -1.44855618e+00 5.53992689e-01
3.35640520e-01 1.93819165e-01 -2.67993599e-01 6.89378798e-01
-8.40770155e-02 1.11717331e+00 -3.84295881e-01 1.10360626e-02
1.40965730e-01 1.72202930e-01 7.81666994e-01 1.45827031e+00
4.01595384e-01 -3.57363731e-01 1.74776003e-01 7.05800056e-01
-1.07791178e-01 2.78520048e-01 -5.16438484e-01 -2.71090418e-02
8.12698603e-01 8.88530195e-01 -2.75992155e-01 -4.48507309e-01
-1.12985715e-01 1.05249393e+00 4.07871604e-01 5.72005570e-01
-5.73298693e-01 -1.00249901e-01 4.81231183e-01 1.62095547e-01
3.67426902e-01 -1.74338147e-01 -6.22691885e-02 -1.19301808e+00
2.65564680e-01 -1.53523743e+00 1.80646285e-01 -5.87393165e-01
-8.73140275e-01 7.69866407e-01 2.47147679e-02 -1.10912788e+00
-9.34535205e-01 -2.17520431e-01 -3.78997624e-01 1.28785050e+00
-1.20812500e+00 -1.12033570e+00 3.67846400e-01 5.69336116e-01
9.52113986e-01 -5.47854900e-01 1.19722748e+00 4.41702783e-01
-4.32300508e-01 8.29057574e-01 2.33286768e-01 -1.03381731e-01
8.39371562e-01 -9.80039120e-01 7.20896721e-01 9.61450756e-01
4.63818371e-01 7.38596678e-01 1.03275740e+00 -5.22366941e-01
-1.52892029e+00 -8.43426228e-01 1.21602404e+00 -4.21146810e-01
5.45431018e-01 -5.94062746e-01 -9.00560856e-01 5.17051280e-01
2.73771644e-01 -1.85659334e-01 5.35635769e-01 2.57913977e-01
-3.12426209e-01 -3.21433842e-01 -9.59907591e-01 7.01941311e-01
7.62812078e-01 -8.44147921e-01 -5.18035293e-01 2.86245972e-01
7.88622141e-01 -5.87237477e-01 -7.23365068e-01 1.00729130e-01
6.47495270e-01 -7.54420757e-01 7.89285779e-01 -8.58466387e-01
1.35015905e-01 -3.10698785e-02 -4.74787712e-01 -1.59747910e+00
-3.47453379e-03 -1.23018229e+00 -1.63729832e-01 1.24352694e+00
8.33156288e-01 -5.25365531e-01 5.96347392e-01 3.45303923e-01
-3.59908789e-01 -7.68354535e-01 -1.19518268e+00 -6.87353730e-01
4.15719971e-02 -5.61887443e-01 6.77129626e-01 5.04814386e-01
-6.57441467e-02 5.93688905e-01 -4.86056119e-01 1.35543391e-01
1.30160540e-01 1.06407754e-01 1.09041297e+00 -7.83781230e-01
-8.79847527e-01 -2.74246126e-01 1.12283140e-01 -1.10937536e+00
8.69852081e-02 -8.91353607e-01 2.12319449e-01 -1.13768411e+00
-1.07701346e-01 -3.83092999e-01 -3.08376908e-01 4.19579268e-01
-1.40264273e-01 4.35547307e-02 3.15067142e-01 3.07039022e-01
-1.67859763e-01 2.80404121e-01 8.12201142e-01 -8.97007212e-02
-4.69198972e-01 3.96540254e-01 -4.27178532e-01 1.81114584e-01
8.16587150e-01 -7.15933442e-01 -4.09349978e-01 -6.19642437e-01
1.80277094e-01 3.79454970e-01 6.20012656e-02 -1.01863599e+00
4.78220396e-02 -9.10117328e-02 2.70356704e-02 -4.56380844e-01
5.38054407e-01 -7.30968475e-01 2.75030047e-01 3.31723154e-01
-6.94385350e-01 1.42192334e-01 3.62043619e-01 3.63314748e-01
-1.91127941e-01 -2.71982372e-01 6.67882442e-01 2.16315594e-02
-2.53178805e-01 5.77247888e-02 -4.13017601e-01 1.06632337e-01
3.76806021e-01 -7.43277371e-02 1.22486137e-01 -5.80607235e-01
-7.21388161e-01 -3.07655156e-01 2.30528802e-01 4.71239150e-01
2.52100289e-01 -1.10949874e+00 -9.48072553e-01 3.58516753e-01
1.76784560e-01 -2.24219918e-01 -1.86700910e-01 8.23441148e-01
9.75117832e-02 5.10705292e-01 8.55852515e-02 -6.02056026e-01
-1.28360593e+00 4.25143182e-01 4.18478847e-01 -4.36862409e-01
-3.63282949e-01 6.41799152e-01 -2.40026981e-01 -4.64740694e-01
2.63645113e-01 -1.66188732e-01 1.61002159e-01 9.93557274e-03
3.43674272e-01 1.00153297e-01 5.09781539e-01 -6.48739457e-01
-1.45681351e-01 2.55416244e-01 -1.51863143e-01 -8.86557102e-01
1.03460443e+00 -2.19571039e-01 3.65157008e-01 8.08007419e-01
1.06726682e+00 2.35947654e-01 -1.20219231e+00 -5.07662952e-01
7.52389580e-02 -4.04201865e-01 -4.14975770e-02 -1.13088119e+00
-7.67605066e-01 9.62656856e-01 4.76918876e-01 1.72785968e-02
1.25028539e+00 -2.22648755e-01 7.77491748e-01 6.07315302e-01
3.85589093e-01 -1.23854053e+00 -6.25960454e-02 5.96392035e-01
8.11156034e-01 -8.81317973e-01 -2.78187126e-01 8.85668322e-02
-6.03264511e-01 1.06632626e+00 3.02484240e-02 8.56354237e-02
2.12744981e-01 2.97676563e-01 3.83834064e-01 2.34160990e-01
-1.20430398e+00 -5.49767129e-02 3.94811183e-01 5.90223074e-01
6.07007742e-01 -9.12465341e-03 -1.36248469e-01 3.79279107e-01
-2.95763016e-01 1.41068529e-02 1.66520461e-01 7.15323329e-01
-4.31681126e-01 -1.38523138e+00 -3.08449566e-01 2.57412255e-01
-6.94157362e-01 -2.53515869e-01 -5.04431844e-01 5.29282570e-01
-2.47175157e-01 1.09218955e+00 -2.07805172e-01 -3.51505339e-01
4.63780820e-01 6.25281334e-01 3.47200096e-01 -7.54210889e-01
-1.00721931e+00 4.70110089e-01 4.33397442e-01 -4.65868056e-01
-1.73042163e-01 -8.95144224e-01 -9.74195957e-01 6.79711550e-02
-3.72280091e-01 1.32211402e-01 1.11724377e+00 1.10292029e+00
6.68292701e-01 4.79776233e-01 8.08530509e-01 -8.79151523e-01
-1.01659095e+00 -1.24431741e+00 4.67084497e-02 2.89784938e-01
3.73865962e-01 -2.13223979e-01 -1.44910783e-01 3.61532331e-01]
|
[14.519807815551758, 6.9855756759643555]
|
0c94194f-6f14-43c0-a1ee-5a7eef28bef1
|
enhancing-robustness-in-aspect-based
| null | null |
https://openreview.net/forum?id=Isf6O2t_99
|
https://openreview.net/pdf?id=Isf6O2t_99
|
Enhancing Robustness in Aspect-based Sentiment Analysis by Better Exploiting Data Augmentation
|
In this paper, we propose to leverage data augmentation to improve the robustness of aspect-based sentiment analysis models. Our method not only exploits augmented data but also makes models focus more on predictive features. We show in experiments that our method compares favorably against strong baselines on both robustness and standard datasets. In the contrary, the widely used adversarial training that only leverages the augmented data fails to improve performance due to the distribution shift caused by the augmented data.
|
['Anonymous']
|
2022-01-16
| null | null | null |
acl-arr-january-2022-1
|
['aspect-based-sentiment-analysis']
|
['natural-language-processing']
|
[ 1.18967846e-01 1.15551546e-01 -2.62445539e-01 -3.53677720e-01
-7.47755349e-01 -8.26546848e-01 1.05976558e+00 2.32956097e-01
-4.00037438e-01 3.98159027e-01 4.98673171e-01 -3.63599092e-01
2.70748287e-01 -8.04747641e-01 -5.78382611e-01 -5.72575927e-01
4.43925649e-01 2.47398049e-01 2.34890748e-02 -7.51488805e-01
3.76622468e-01 2.71237433e-01 -1.03552461e+00 2.03200638e-01
5.62753260e-01 7.36641586e-01 -7.95897126e-01 4.04190809e-01
6.73034489e-02 6.12964690e-01 -6.47461474e-01 -8.55047345e-01
4.64146316e-01 -2.76849270e-02 -3.25827122e-01 -1.66308865e-01
5.28007150e-01 -1.77343756e-01 -2.85073698e-01 9.46215987e-01
5.69034576e-01 1.56863481e-01 6.17962718e-01 -1.44044554e+00
-7.03657687e-01 6.37704015e-01 -9.61838484e-01 2.69305199e-01
2.17887878e-01 3.63418460e-01 1.20935833e+00 -6.29953921e-01
6.37752652e-01 9.75759685e-01 9.99380469e-01 5.28070569e-01
-1.31624663e+00 -7.46079624e-01 4.66422051e-01 -1.78952798e-01
-7.43133366e-01 -4.10715520e-01 1.03160727e+00 -8.61268863e-02
9.49601054e-01 2.03257382e-01 2.71693140e-01 1.53154290e+00
2.29407102e-01 8.35042775e-01 1.30952013e+00 -2.70145327e-01
2.42362186e-01 3.97974312e-01 5.68404555e-01 2.82248020e-01
4.52020615e-01 4.58322197e-01 -4.47621047e-01 -4.42466646e-01
-6.99338019e-02 -3.67304049e-02 2.39953585e-02 -2.83306271e-01
-8.66338670e-01 1.15486228e+00 2.54269898e-01 2.06737474e-01
-3.48131776e-01 3.89097608e-03 6.65330470e-01 3.24340820e-01
8.19471478e-01 9.43065584e-01 -1.11882150e+00 -2.56527990e-01
-9.00578737e-01 3.09875876e-01 6.94982171e-01 5.69947600e-01
4.38674897e-01 4.77007687e-01 -3.02916974e-01 6.35055184e-01
3.59891728e-02 7.38724053e-01 6.61037385e-01 -6.53086662e-01
4.97004420e-01 9.11833107e-01 -1.29670188e-01 -1.10155416e+00
-4.94378507e-01 -4.28491801e-01 -6.43142998e-01 1.89379007e-01
2.66886503e-01 -3.60318154e-01 -1.12662899e+00 1.79159415e+00
1.68313473e-01 7.14162588e-02 1.05418868e-01 4.33366328e-01
7.76938736e-01 2.36729473e-01 7.80391768e-02 7.61957690e-02
1.12607396e+00 -9.31516111e-01 -7.01131642e-01 -3.94509166e-01
6.61978841e-01 -7.57790804e-01 1.43257499e+00 2.44767740e-01
-9.00106490e-01 -8.67237970e-02 -1.11440861e+00 1.63688123e-01
-6.92985058e-01 -3.48877102e-01 8.30724776e-01 9.88777161e-01
-7.44872391e-01 4.03300554e-01 -7.03284740e-01 -1.16869904e-01
5.69405615e-01 3.33887011e-01 -7.88860559e-01 9.45340544e-02
-1.07256567e+00 8.07364285e-01 4.45132293e-02 -4.30957019e-01
-2.86616355e-01 -1.09022534e+00 -1.05237341e+00 3.58274914e-02
4.33254480e-01 -5.57788014e-01 1.18994820e+00 -8.00413907e-01
-1.37780011e+00 4.91320759e-01 -8.08587745e-02 -6.49906754e-01
3.92376930e-01 -6.26498997e-01 -3.80630255e-01 -1.25844210e-01
-3.42551358e-02 2.25346953e-01 9.17286992e-01 -1.26007128e+00
-2.35109568e-01 -3.52282107e-01 3.06108832e-01 -2.74207413e-01
-6.27979159e-01 2.53329296e-02 -3.37652713e-01 -1.09589446e+00
-3.98117632e-01 -1.16568124e+00 -4.90932941e-01 -7.62606680e-01
-7.33837068e-01 1.09657995e-01 1.02369869e+00 -4.32117462e-01
1.13828874e+00 -2.23881865e+00 -2.08013162e-01 3.55882525e-01
-6.63147196e-02 4.26900387e-01 -4.33410794e-01 2.55867600e-01
-2.58113056e-01 5.00097215e-01 -2.52451301e-01 -6.48822069e-01
6.86568469e-02 -2.69848993e-03 -7.04767823e-01 4.22075957e-01
1.35988578e-01 1.12655282e+00 -5.04086673e-01 1.68966744e-02
6.32988662e-02 4.28449184e-01 -9.03880835e-01 -8.46221372e-02
-1.90864220e-01 1.97978035e-01 -3.17054302e-01 6.45244479e-01
8.75848889e-01 -8.78778026e-02 -1.10532576e-02 -1.48133263e-01
4.51154560e-01 2.56769538e-01 -6.47108912e-01 1.20447981e+00
-5.77850997e-01 5.42750597e-01 -3.95760566e-01 -6.58160865e-01
7.29191065e-01 1.05794519e-01 3.35634887e-01 -8.42010975e-01
2.60418326e-01 -1.37756929e-01 -1.53048821e-02 -3.00586708e-02
6.79566026e-01 -1.83251679e-01 -3.15945148e-01 8.52621496e-01
-1.98682509e-02 -3.59213233e-01 2.82703917e-02 5.29052138e-01
1.10314703e+00 -7.61202276e-02 2.80951798e-01 -1.82446703e-01
5.25485814e-01 -7.45799020e-02 5.07855296e-01 7.56478965e-01
-6.14060685e-02 5.70628285e-01 7.88574219e-01 -4.29353833e-01
-1.04121256e+00 -7.70343840e-01 -9.39222798e-02 1.22013593e+00
-2.79193014e-01 -7.45409727e-01 -5.16335368e-01 -1.44807291e+00
2.65725851e-01 1.16913545e+00 -1.20620644e+00 -6.34434938e-01
-3.69554400e-01 -1.14620602e+00 6.48685873e-01 7.78477788e-01
3.39963555e-01 -9.29717183e-01 -2.52906587e-02 -1.88279644e-01
6.51385933e-02 -1.23671508e+00 -4.11526740e-01 1.45852774e-01
-9.40061867e-01 -1.12303567e+00 -1.19579002e-01 -3.67581435e-02
5.52932322e-01 2.83058047e-01 1.26897967e+00 1.41710013e-01
2.93738782e-01 4.98556376e-01 -5.60867488e-01 -1.00910258e+00
-3.62254679e-01 4.84154552e-01 1.20911255e-01 -1.61400288e-01
7.16575146e-01 -5.38816869e-01 -2.28019774e-01 2.96553642e-01
-1.16452909e+00 -6.63913667e-01 4.43171978e-01 9.00195360e-01
2.98003376e-01 -8.84074941e-02 7.42256999e-01 -1.58133256e+00
7.01920748e-01 -5.64645827e-01 -3.64149421e-01 -6.89050229e-03
-1.21032810e+00 1.24145836e-01 6.63640440e-01 -4.58163410e-01
-8.69468451e-01 -2.75411814e-01 -1.50607407e-01 -2.01076180e-01
-1.42140031e-01 4.87697005e-01 -9.13087949e-02 7.41565973e-03
6.81762516e-01 -5.78174591e-02 1.22101277e-01 -5.16343236e-01
5.45726538e-01 3.07752013e-01 1.94770247e-01 -3.61797571e-01
1.15850699e+00 6.83156013e-01 4.75897379e-02 -5.01475513e-01
-9.45733249e-01 -2.20341742e-01 -4.10270095e-01 4.36870337e-01
3.33637923e-01 -7.85350740e-01 -3.33292156e-01 3.73123080e-01
-7.07281470e-01 -1.89865947e-01 -5.10196626e-01 1.91467404e-01
-3.41171145e-01 2.91563153e-01 -6.88901186e-01 -5.60395002e-01
-6.03527009e-01 -9.57737684e-01 8.98003459e-01 -1.37959013e-03
-4.15329397e-01 -1.23750174e+00 4.67729360e-01 3.81422490e-01
7.37287700e-01 4.46598947e-01 8.40435982e-01 -1.44681859e+00
-1.32534504e-02 -8.50515902e-01 1.29871637e-01 5.03330529e-01
2.34938845e-01 2.55018443e-01 -1.19134343e+00 -2.89354414e-01
5.12425527e-02 -2.14313239e-01 1.12034559e+00 2.28878871e-01
8.77276957e-01 -3.58941466e-01 -8.21430888e-03 7.13326156e-01
1.12866402e+00 -2.35943571e-01 6.44841254e-01 8.24503541e-01
6.25779808e-01 6.17013216e-01 6.34807348e-01 2.66757041e-01
2.10118681e-01 6.94007695e-01 3.94158393e-01 -3.36389631e-01
1.47482067e-01 -2.11328238e-01 5.51459074e-01 4.47331935e-01
2.75230203e-02 -1.16613157e-01 -7.06039011e-01 3.88216496e-01
-1.69666731e+00 -8.28326941e-01 -1.29540116e-01 1.97505152e+00
6.67002022e-01 4.36448038e-01 1.79338574e-01 2.21164212e-01
3.05451989e-01 3.70173424e-01 -4.08982903e-01 -7.97097743e-01
-4.36442107e-01 4.65658039e-01 4.69140470e-01 4.47406620e-01
-1.22954142e+00 1.10112500e+00 7.55507469e+00 5.51998496e-01
-9.47856426e-01 8.11198652e-02 5.94996989e-01 -4.88712281e-01
-6.46010756e-01 -1.54304460e-01 -6.64534807e-01 3.48843426e-01
9.92959023e-01 -2.24905625e-01 1.10797428e-01 9.93124723e-01
-1.21448256e-01 1.32320493e-01 -9.21960413e-01 3.29480320e-01
3.17734987e-01 -1.07428336e+00 2.71410137e-01 3.38083655e-01
1.00984454e+00 -1.61296763e-02 6.80115938e-01 6.01418853e-01
6.14939928e-01 -1.01796448e+00 3.95819813e-01 1.88587025e-01
3.86231512e-01 -1.06662631e+00 1.31408072e+00 -1.29903834e-02
-5.69630504e-01 -4.34088819e-02 -1.31387949e-01 -4.81582768e-02
-1.05668351e-01 7.84180880e-01 -8.39266896e-01 4.79421198e-01
4.69542652e-01 7.10058272e-01 -1.04132497e+00 5.60179174e-01
-5.69662929e-01 1.02323651e+00 -2.10065633e-01 3.46404463e-01
1.65246919e-01 2.65704673e-02 7.00977564e-01 1.22090924e+00
-1.40328348e-01 -2.75400072e-01 -2.24892512e-01 3.38949054e-01
-2.56546944e-01 2.38416225e-01 -1.03022659e+00 1.78635828e-02
1.25522912e-01 1.27592850e+00 -4.46686357e-01 -2.75307357e-01
-6.50799572e-01 5.58889985e-01 2.31216505e-01 3.74332100e-01
-6.45230949e-01 -1.86576054e-01 7.83831000e-01 -9.23839770e-03
5.33523858e-01 5.78667037e-02 -6.86259985e-01 -1.25464404e+00
-1.15854762e-01 -1.34491193e+00 5.81104100e-01 -4.48871940e-01
-1.53966308e+00 6.60345674e-01 -2.30847135e-01 -9.36007798e-01
-5.01464427e-01 -5.60142994e-01 -8.75459492e-01 6.93956912e-01
-1.61829948e+00 -1.38148761e+00 9.19726342e-02 7.29169548e-01
2.16213837e-01 -5.58656514e-01 9.31855559e-01 -6.48168772e-02
-6.83897972e-01 1.17479122e+00 7.77598619e-02 1.88307241e-01
9.80807364e-01 -1.46209943e+00 9.30946171e-01 1.00532520e+00
3.17676395e-01 8.62410188e-01 9.67214882e-01 -5.53794801e-01
-1.18473470e+00 -8.11980247e-01 5.13300121e-01 -1.19021952e+00
9.40562904e-01 -3.10569108e-01 -9.84172761e-01 1.00564218e+00
3.16227078e-01 5.48175052e-02 1.25097466e+00 7.17903733e-01
-1.09563589e+00 -1.24272875e-01 -1.26996827e+00 4.69822347e-01
3.99725735e-01 -4.45159584e-01 -7.68593073e-01 6.30269665e-03
9.58061874e-01 -3.25567201e-02 -6.28271818e-01 6.86345220e-01
4.46154743e-01 -8.59240413e-01 9.28748786e-01 -1.38277972e+00
7.08558440e-01 5.51480837e-02 -3.15530032e-01 -1.54619968e+00
-2.00974807e-01 -3.68741095e-01 -2.61908174e-01 1.50597906e+00
7.53314734e-01 -8.75135124e-01 8.38289797e-01 8.74177814e-01
3.06911975e-01 -5.81711233e-01 -6.01605535e-01 -6.46294296e-01
4.12355095e-01 -6.57062948e-01 8.68162930e-01 1.07740796e+00
1.26572996e-02 4.47251737e-01 -3.88255149e-01 4.76302728e-02
3.52852076e-01 1.44993454e-01 1.13756621e+00 -1.06108034e+00
-2.46406391e-01 -5.15599191e-01 -4.06247556e-01 -2.64722884e-01
4.14513528e-01 -6.13678396e-01 -3.47443938e-01 -1.07346225e+00
2.53140241e-01 -3.78217816e-01 -7.56245613e-01 7.33685136e-01
-6.65961564e-01 6.91830397e-01 4.61649477e-01 -1.91734597e-01
-5.36189020e-01 5.63553214e-01 8.48012209e-01 -2.58547753e-01
-3.40077490e-01 3.29564542e-01 -1.56972587e+00 8.11431050e-01
1.09915864e+00 -5.33218563e-01 -3.02486509e-01 -1.06500089e-01
6.72364175e-01 -6.87803864e-01 5.98926954e-02 -6.03695869e-01
-1.84523359e-01 -4.88301329e-02 2.72325426e-01 -2.34787285e-01
2.43581384e-01 -8.63284826e-01 -5.55815101e-01 2.91959912e-01
-3.09246719e-01 3.68888289e-01 7.34856129e-01 7.02605903e-01
-1.75192103e-01 -7.32706487e-02 5.34802198e-01 2.24819615e-01
-1.70097187e-01 2.85277069e-01 -2.12097406e-01 2.78280467e-01
7.81846285e-01 2.94701517e-01 -6.65206850e-01 -6.01641417e-01
-2.73506373e-01 1.40012633e-02 6.96592629e-01 6.92125201e-01
2.34316975e-01 -1.20224929e+00 -6.20143890e-01 2.58379728e-01
3.22103649e-01 -6.07135117e-01 7.81183168e-02 7.23897576e-01
4.42094803e-02 2.90022075e-01 -1.85655039e-02 -2.25579560e-01
-1.22772896e+00 1.01294351e+00 6.29151985e-02 -6.05897844e-01
-4.20667350e-01 5.67940176e-01 1.26853094e-01 -8.46209288e-01
-6.12323731e-02 2.54345760e-02 -4.12721992e-01 6.30094334e-02
6.66450500e-01 2.10436195e-01 3.36280227e-01 -6.14064991e-01
-4.00400460e-01 3.93334180e-01 -5.44766545e-01 -1.69031754e-01
1.49332345e+00 2.54100442e-01 9.07059759e-02 3.76151294e-01
1.10485196e+00 8.89293075e-01 -7.87620544e-01 -2.24160343e-01
-6.59234971e-02 -6.01593137e-01 1.38320312e-01 -1.21231461e+00
-1.39547765e+00 7.19961107e-01 1.50367126e-01 2.99100667e-01
1.08047807e+00 -2.93803513e-01 6.29781783e-01 3.51023823e-01
4.12721373e-02 -9.55027938e-01 1.52530372e-01 6.15962386e-01
5.73233843e-01 -1.46947360e+00 4.34891641e-01 -2.97395855e-01
-1.12655091e+00 7.37434626e-01 6.50994062e-01 -3.55900854e-01
8.37016106e-01 4.97198761e-01 3.98197174e-01 -1.55109242e-01
-9.12152052e-01 6.57832250e-03 3.59991044e-01 7.72576928e-01
2.78744429e-01 -1.25489861e-01 -2.29880273e-01 9.55989599e-01
-6.06436074e-01 -3.98149788e-01 6.57226145e-01 8.60476851e-01
9.29148421e-02 -1.33217978e+00 -2.94074029e-01 6.66131377e-01
-9.27490294e-01 -4.31414098e-01 -5.84492862e-01 1.12830091e+00
-2.55192012e-01 1.03077865e+00 -1.38597563e-01 -7.22546041e-01
5.88397861e-01 1.87121451e-01 2.55317125e-03 -5.09182572e-01
-1.02973902e+00 -2.37901449e-01 1.33830868e-02 -6.72227085e-01
-3.56387556e-01 -7.12596655e-01 -7.62855768e-01 -6.35068715e-01
-4.27062511e-01 1.25381902e-01 7.25883961e-01 9.19328272e-01
5.36524236e-01 5.60859203e-01 8.32754850e-01 -2.79504061e-01
-6.97168410e-01 -1.22524500e+00 -4.59783405e-01 7.72546411e-01
4.49986815e-01 -6.00225747e-01 -7.09418952e-01 -3.36586893e-01]
|
[11.27287483215332, 6.902848720550537]
|
4971d73d-4d3a-44d5-a26f-687cd1abd169
|
eight-years-of-face-recognition-research
|
2208.04040
| null |
https://arxiv.org/abs/2208.04040v2
|
https://arxiv.org/pdf/2208.04040v2.pdf
|
Eight Years of Face Recognition Research: Reproducibility, Achievements and Open Issues
|
Automatic face recognition is a research area with high popularity. Many different face recognition algorithms have been proposed in the last thirty years of intensive research in the field. With the popularity of deep learning and its capability to solve a huge variety of different problems, face recognition researchers have concentrated effort on creating better models under this paradigm. From the year 2015, state-of-the-art face recognition has been rooted in deep learning models. Despite the availability of large-scale and diverse datasets for evaluating the performance of face recognition algorithms, many of the modern datasets just combine different factors that influence face recognition, such as face pose, occlusion, illumination, facial expression and image quality. When algorithms produce errors on these datasets, it is not clear which of the factors has caused this error and, hence, there is no guidance in which direction more research is required. This work is a followup from our previous works developed in 2014 and eventually published in 2016, showing the impact of various facial aspects on face recognition algorithms. By comparing the current state-of-the-art with the best systems from the past, we demonstrate that faces under strong occlusions, some types of illumination, and strong expressions are problems mastered by deep learning algorithms, whereas recognition with low-resolution images, extreme pose variations, and open-set recognition is still an open problem. To show this, we run a sequence of experiments using six different datasets and five different face recognition algorithms in an open-source and reproducible manner. We provide the source code to run all of our experiments, which is easily extensible so that utilizing your own deep network in our evaluation is just a few minutes away.
|
['Dominic Schmidli', 'Manuel Günther', 'Sébastien Marcel', 'Xinyi Zhang', 'Yu Linghu', 'Tiago de Freitas Pereira']
|
2022-08-08
| null | null | null | null |
['open-set-learning']
|
['miscellaneous']
|
[-1.51989330e-02 -3.53905648e-01 -6.43280074e-02 -8.65548372e-01
-6.06569275e-02 -1.55125588e-01 4.84357804e-01 -7.42528856e-01
-2.24727914e-01 5.43817699e-01 -1.26602247e-01 6.94093853e-02
-1.16451979e-01 -6.68415844e-01 -6.02774084e-01 -7.70331860e-01
-1.06642172e-01 3.46023142e-01 -2.78120607e-01 -4.05243367e-01
1.29119709e-01 1.04994643e+00 -2.12309384e+00 2.88397402e-01
1.80826396e-01 1.13384998e+00 -3.09230030e-01 3.52542728e-01
-1.66202545e-01 3.32111120e-01 -7.52546430e-01 -7.36535609e-01
5.17810225e-01 -2.54168957e-01 -6.09830856e-01 3.04831535e-01
8.15098405e-01 -3.49128991e-01 -3.73202354e-01 9.00992155e-01
8.89730811e-01 -2.39566848e-01 2.92851061e-01 -1.40683329e+00
-7.22413480e-01 -6.05674312e-02 -5.94729245e-01 2.17469633e-02
4.85234588e-01 1.25517011e-01 3.42598468e-01 -1.13914895e+00
4.99354124e-01 1.58260334e+00 7.04132378e-01 8.79279077e-01
-1.01751566e+00 -1.12116015e+00 -3.51348817e-02 3.47419232e-01
-1.45790565e+00 -8.79565120e-01 8.61175239e-01 -2.17851922e-01
7.47851849e-01 2.33279169e-01 4.90315229e-01 1.60170007e+00
-7.95410350e-02 4.60125118e-01 1.36804497e+00 -5.62736630e-01
-7.08144680e-02 1.53205290e-01 -3.47156301e-02 7.36974537e-01
1.74805641e-01 3.80959600e-01 -4.81027544e-01 -9.31028053e-02
7.15551198e-01 1.59696057e-01 -2.47150600e-01 -3.00441504e-01
-6.09481156e-01 8.07160318e-01 1.86537594e-01 4.60610807e-01
-8.42508152e-02 -1.31082460e-01 2.74228394e-01 6.26743197e-01
2.16988489e-01 8.75035226e-02 -4.51750159e-01 -3.22463959e-02
-9.25906837e-01 1.41599268e-01 1.15480590e+00 5.41268706e-01
8.19764078e-01 2.02003703e-01 1.68635219e-01 1.13032639e+00
3.21055204e-01 4.82008606e-01 4.77579325e-01 -7.74851799e-01
-9.36658382e-02 5.47684252e-01 -1.16233684e-01 -1.35151732e+00
-3.60802531e-01 -3.50616008e-01 -9.41807568e-01 4.84762549e-01
6.57367706e-01 -1.38353527e-01 -1.11191738e+00 1.73351908e+00
2.56775528e-01 8.67057964e-02 -2.14250356e-01 9.37899053e-01
1.01376128e+00 3.83140415e-01 -1.72436148e-01 -2.48568296e-01
1.36025906e+00 -6.98218286e-01 -8.22197318e-01 -4.15761024e-02
2.26752087e-01 -1.07688022e+00 8.01367700e-01 4.94649678e-01
-8.25543582e-01 -7.31593907e-01 -9.96184945e-01 1.72006711e-01
-5.96534431e-01 1.00171253e-01 7.51473784e-01 1.24920309e+00
-1.20424056e+00 6.15514576e-01 -5.45288980e-01 -7.41663575e-01
8.04802418e-01 6.81084394e-01 -8.72993648e-01 -3.58949751e-01
-1.02262676e+00 9.95417595e-01 -1.89258203e-01 4.95038927e-01
-8.42129171e-01 -4.30644572e-01 -4.60937411e-01 -9.75994989e-02
4.52013254e-01 -3.37098449e-01 9.00920808e-01 -1.45570481e+00
-1.56115496e+00 1.26116741e+00 -2.58833736e-01 1.10413104e-01
3.96929532e-01 -1.74444959e-01 -8.21018279e-01 -1.90889806e-01
-4.67351466e-01 5.08279622e-01 1.12015700e+00 -1.16596127e+00
-1.26658961e-01 -9.50906932e-01 -4.17177752e-02 -3.08501959e-01
-4.70349878e-01 5.59762955e-01 -4.33085054e-01 -3.77727419e-01
-1.97111905e-01 -9.73195851e-01 2.44270548e-01 2.95959443e-01
2.77682263e-02 -3.31822097e-01 1.15086162e+00 -3.73098940e-01
7.77099133e-01 -2.23698235e+00 -9.19164717e-02 4.24137004e-02
-1.11710235e-01 7.20187485e-01 -2.92336673e-01 3.00415695e-01
-3.99323881e-01 3.02387983e-01 -9.02361870e-02 -2.66415507e-01
7.16066500e-03 3.26840103e-01 -2.49498576e-01 5.63717604e-01
2.67497629e-01 7.31688797e-01 -4.34241325e-01 -1.64315373e-01
7.53100142e-02 8.82813394e-01 -3.84231478e-01 3.15621972e-01
7.15101436e-02 2.95147508e-01 -1.67861357e-01 1.00518417e+00
1.10819721e+00 1.41458467e-01 4.06406410e-02 -3.31118196e-01
8.48605633e-02 -3.78847480e-01 -1.30972779e+00 1.35417509e+00
-2.30976626e-01 7.06249595e-01 3.66023868e-01 -1.17234576e+00
1.16960180e+00 5.41786313e-01 4.28113282e-01 -7.68329024e-01
3.30767542e-01 3.04733604e-01 1.60146832e-01 -7.66420960e-01
4.81078811e-02 -1.36229619e-01 5.54447472e-01 4.47006315e-01
2.18485087e-01 1.87206239e-01 1.37247369e-01 -4.16084290e-01
8.26868474e-01 4.64949384e-02 1.12074539e-02 -1.11930527e-01
6.51835561e-01 -5.86094379e-01 4.85861808e-01 4.54432011e-01
-5.97510517e-01 6.22257829e-01 4.28230882e-01 -1.02333522e+00
-7.84493327e-01 -6.26644552e-01 -4.79626387e-01 1.04373586e+00
-3.79503965e-01 -2.13745803e-01 -7.49244690e-01 -4.36083496e-01
4.31961566e-02 -5.26554659e-02 -8.69052768e-01 9.24994200e-02
-5.49506605e-01 -1.00972188e+00 6.82746053e-01 3.21536750e-01
6.71667159e-01 -1.34514058e+00 -2.88705170e-01 -1.99981317e-01
3.00748855e-01 -1.06885684e+00 1.65484790e-02 -1.48633718e-01
-6.24471247e-01 -1.24600720e+00 -7.89148331e-01 -8.32957923e-01
7.49961436e-01 2.20819056e-01 1.21291351e+00 5.35149157e-01
-5.83060324e-01 3.76007318e-01 -2.21450254e-01 -5.60043216e-01
-2.05369443e-01 -1.75122842e-01 2.89620042e-01 3.48276019e-01
7.46344864e-01 -5.80551624e-01 -4.32344735e-01 5.60673118e-01
-8.90330493e-01 -3.74886245e-01 5.66064656e-01 9.56470191e-01
2.31653545e-02 -7.24338740e-02 6.20035529e-01 -7.56496787e-01
4.67794478e-01 -3.11337471e-01 -6.04428470e-01 2.75388867e-01
-3.99767786e-01 -1.94214121e-01 4.88057345e-01 -3.37817281e-01
-9.23704326e-01 3.66695113e-02 -4.85707015e-01 -4.28952962e-01
-6.41458988e-01 1.93718508e-01 -4.48248953e-01 -5.09315729e-01
6.27486527e-01 1.46643773e-01 3.91619027e-01 -4.92338419e-01
-1.23245446e-02 6.92953885e-01 2.06991494e-01 -5.56164503e-01
7.45408237e-01 4.97706592e-01 3.04700378e-02 -9.68974650e-01
-6.91324651e-01 1.62826180e-02 -6.44393623e-01 -2.97695965e-01
5.18884778e-01 -6.78370655e-01 -9.77361858e-01 9.46325004e-01
-1.10525179e+00 5.07035181e-02 2.11946487e-01 1.29355103e-01
-1.04988649e-01 1.91488549e-01 -3.00316602e-01 -8.20137382e-01
-1.55949026e-01 -1.37116230e+00 8.47551465e-01 4.45897579e-01
2.15721857e-02 -7.89913774e-01 -1.30934536e-01 5.19989789e-01
7.78365731e-01 2.37478763e-01 6.85313702e-01 -5.23523450e-01
-4.40489054e-01 -2.42029056e-01 -2.93442339e-01 4.65048045e-01
4.02168453e-01 3.83457690e-01 -1.45999682e+00 -5.65862298e-01
1.40506431e-01 -6.86339140e-01 6.69788957e-01 1.03839234e-01
1.45812416e+00 -1.76201880e-01 -2.30210543e-01 8.28501165e-01
1.28916800e+00 1.84550002e-01 1.06446433e+00 1.96338788e-01
5.21102250e-01 8.42675984e-01 9.67115462e-02 2.19481781e-01
2.30281707e-02 8.98130476e-01 5.35402656e-01 -2.48745605e-01
-3.64633054e-02 2.76128024e-01 4.16593999e-01 2.18658864e-01
-2.98784792e-01 -1.09062195e-01 -8.08416247e-01 1.51115239e-01
-1.26408505e+00 -1.25918412e+00 2.95874298e-01 2.18200970e+00
6.95347667e-01 -3.75157088e-01 1.29811421e-01 3.35946172e-01
6.18662179e-01 1.48960277e-01 -4.39787894e-01 -5.00716507e-01
-2.86796540e-01 5.11215031e-01 -1.65646523e-01 1.37971967e-01
-9.73185003e-01 8.67228031e-01 6.79203463e+00 6.26145303e-01
-1.78439307e+00 -2.04616666e-01 9.33174491e-01 -4.22244295e-02
1.71444163e-01 -3.51098061e-01 -8.68717611e-01 4.19469267e-01
8.13740015e-01 9.54651535e-02 5.46058416e-01 9.36505258e-01
5.47837168e-02 1.65432945e-01 -1.27809083e+00 1.40330863e+00
5.84132254e-01 -1.03630340e+00 -7.45839393e-03 1.45165965e-01
5.81630230e-01 -6.59759790e-02 3.54260832e-01 3.60714614e-01
-1.41570300e-01 -1.68883216e+00 2.50832528e-01 2.74902344e-01
6.42540693e-01 -7.13987350e-01 8.23404789e-01 -1.76202618e-02
-6.34904444e-01 -1.72697827e-01 -4.51544702e-01 -1.28198832e-01
-3.64966303e-01 5.80886185e-01 -5.20104766e-01 3.82247061e-01
8.41732979e-01 4.69907522e-01 -6.57794476e-01 7.96062410e-01
-5.21780998e-02 5.34000099e-01 -2.93844640e-01 1.34501010e-01
-9.38491449e-02 -1.73918139e-02 2.63579726e-01 1.10690546e+00
2.87443250e-01 6.59226254e-02 -3.45998891e-02 7.78756201e-01
-4.07323807e-01 1.06037855e-01 -7.29212761e-01 -1.19918756e-01
2.28238344e-01 1.46692073e+00 -5.25618255e-01 5.65621257e-02
-6.78044260e-01 6.66843891e-01 2.88636476e-01 3.23417127e-01
-6.37984693e-01 -2.72599235e-02 9.36110139e-01 1.70055121e-01
3.02010536e-01 -1.77979350e-01 -1.83912262e-01 -9.50421453e-01
1.85664788e-01 -1.36418831e+00 2.44461015e-01 -5.30362427e-01
-1.31782460e+00 7.96214581e-01 -2.41066948e-01 -7.86673009e-01
-1.32787079e-01 -9.76122200e-01 -5.10026276e-01 8.19961727e-01
-1.61739469e+00 -8.65507007e-01 -5.33160686e-01 6.67682707e-01
3.80656153e-01 -6.06320739e-01 1.22103298e+00 6.87040567e-01
-8.64329338e-01 8.79192829e-01 -1.57866359e-01 4.03307825e-01
9.67608154e-01 -4.82860237e-01 1.25822991e-01 5.51259041e-01
2.57902712e-01 7.21081674e-01 5.56590319e-01 -9.05113667e-02
-1.74489474e+00 -7.84128964e-01 7.70484209e-01 -4.48772013e-01
2.82474041e-01 -4.81050968e-01 -1.03282642e+00 4.89346415e-01
2.08595753e-01 4.41548139e-01 8.46299946e-01 2.64064223e-01
-6.17096066e-01 -6.26665294e-01 -1.31310475e+00 3.80271822e-01
1.03137982e+00 -3.40076506e-01 -2.52488434e-01 2.10254878e-01
-9.33686569e-02 -1.86829045e-01 -7.09863663e-01 6.95297658e-01
9.79118645e-01 -1.31604028e+00 8.02457988e-01 -5.75989485e-01
2.93257207e-01 -9.60970744e-02 -2.06179559e-01 -1.16717970e+00
-2.62477726e-01 -5.64238966e-01 1.10413998e-01 1.26070666e+00
1.34168223e-01 -7.05536723e-01 8.79964352e-01 5.56455433e-01
2.30010524e-01 -9.74493921e-01 -8.38462353e-01 -7.50031769e-01
1.65282679e-03 -2.16500387e-01 7.90637076e-01 9.79162753e-01
-5.25204659e-01 9.87822041e-02 -5.01123428e-01 -4.17224877e-02
4.71860856e-01 1.79106101e-01 1.08235633e+00 -1.42149830e+00
2.19690442e-01 -4.48068559e-01 -6.54395759e-01 -5.52800953e-01
4.60328430e-01 -6.64204419e-01 -3.30273688e-01 -9.38001812e-01
2.49027118e-01 -3.29514831e-01 -1.71441302e-01 8.02840829e-01
1.20458603e-01 5.91561079e-01 2.19751328e-01 1.44505724e-01
-1.19450741e-01 4.68401819e-01 1.23842919e+00 -2.05806762e-01
1.71554193e-01 -9.84196961e-02 -7.28615642e-01 6.82375610e-01
6.99613035e-01 -2.47519687e-01 -1.39664844e-01 -5.74463785e-01
-1.49555817e-01 -3.61107647e-01 3.14923376e-01 -1.15503705e+00
7.92446285e-02 -5.76761439e-02 9.01814640e-01 -8.77406672e-02
5.87827265e-01 -1.02085066e+00 2.82394260e-01 2.99785852e-01
5.29510267e-02 -8.75768661e-02 4.92349595e-01 4.23720442e-02
-3.41655016e-01 6.02480993e-02 1.07992280e+00 -1.67501539e-01
-8.09379578e-01 6.28095806e-01 -5.80918342e-02 -1.77210778e-01
1.01225245e+00 -4.97276545e-01 -2.03087181e-01 -4.02087152e-01
-6.42818391e-01 -2.92920321e-01 4.97536063e-01 8.51957202e-01
5.87152004e-01 -1.23490155e+00 -8.93437207e-01 7.42299855e-01
-4.04076874e-02 -4.67498422e-01 1.99302673e-01 5.52350163e-01
-3.19016576e-01 3.49096864e-01 -6.79970920e-01 -6.26261830e-01
-1.58393514e+00 3.71592462e-01 5.08506715e-01 9.62283015e-02
-1.85204864e-01 8.97777259e-01 -5.78239337e-02 -3.30661356e-01
4.61038023e-01 3.81143242e-01 -2.21329197e-01 5.82753718e-02
8.60475183e-01 1.62202403e-01 3.08850169e-01 -8.80571306e-01
-4.97593254e-01 7.27944076e-01 -2.13765889e-01 3.84133250e-01
1.46751082e+00 2.93033242e-01 -2.74782330e-01 9.04348642e-02
1.22726595e+00 -2.19635293e-01 -1.00021183e+00 5.26545867e-02
-2.53149152e-01 -8.02943587e-01 -9.34929475e-02 -8.76839221e-01
-1.69641995e+00 9.68207479e-01 1.01337695e+00 2.08808426e-02
1.26343989e+00 -3.69975537e-01 4.69337702e-01 3.96301180e-01
4.63783592e-01 -9.64719117e-01 1.81143999e-01 7.02115417e-01
1.22046292e+00 -1.64931440e+00 3.87180448e-02 -4.39875394e-01
-1.61322966e-01 1.36976278e+00 7.63355196e-01 1.25011340e-01
9.14800227e-01 4.58876550e-01 3.79227072e-01 -2.30941802e-01
-5.95401049e-01 3.70630845e-02 4.34540249e-02 6.87435806e-01
8.54821980e-01 -1.80264175e-01 -4.26394045e-02 3.17902058e-01
-3.04830670e-01 3.31169695e-01 1.10748552e-01 7.33207226e-01
-1.70261055e-01 -1.51763463e+00 -6.24792099e-01 2.86100894e-01
-6.90674782e-01 3.37859660e-01 -5.61067402e-01 8.74174893e-01
2.84639299e-01 9.93989527e-01 -2.49062330e-02 -3.68256032e-01
2.31497705e-01 2.84798145e-01 8.61389160e-01 -3.97476047e-01
-5.11644065e-01 -4.39528048e-01 -1.62846029e-01 -5.95298827e-01
-5.99094272e-01 -5.37572682e-01 -7.69356906e-01 -6.80136442e-01
-1.07407935e-01 -8.60406905e-02 9.30001974e-01 9.63840425e-01
4.66686100e-01 6.20955117e-02 7.52776504e-01 -1.10406494e+00
-4.67029810e-01 -1.02914560e+00 -5.15820384e-01 6.40244424e-01
2.34073684e-01 -9.36727405e-01 -4.62686419e-01 -4.14743014e-02]
|
[13.238923072814941, 0.933622419834137]
|
76d9e354-8a1f-4615-9559-c7477c424747
|
interpretable-explainability-in-facial
|
2211.04769
| null |
https://arxiv.org/abs/2211.04769v1
|
https://arxiv.org/pdf/2211.04769v1.pdf
|
Interpretable Explainability in Facial Emotion Recognition and Gamification for Data Collection
|
Training facial emotion recognition models requires large sets of data and costly annotation processes. To alleviate this problem, we developed a gamified method of acquiring annotated facial emotion data without an explicit labeling effort by humans. The game, which we named Facegame, challenges the players to imitate a displayed image of a face that portrays a particular basic emotion. Every round played by the player creates new data that consists of a set of facial features and landmarks, already annotated with the emotion label of the target facial expression. Such an approach effectively creates a robust, sustainable, and continuous machine learning training process. We evaluated Facegame with an experiment that revealed several contributions to the field of affective computing. First, the gamified data collection approach allowed us to access a rich variation of facial expressions of each basic emotion due to the natural variations in the players' facial expressions and their expressive abilities. We report improved accuracy when the collected data were used to enrich well-known in-the-wild facial emotion datasets and consecutively used for training facial emotion recognition models. Second, the natural language prescription method used by the Facegame constitutes a novel approach for interpretable explainability that can be applied to any facial emotion recognition model. Finally, we observed significant improvements in the facial emotion perception and expression skills of the players through repeated game play.
|
['Roland Klemke', 'Corrie Urlings', 'Felix Bottger', 'Deniz Iren', 'Krist Shingjergji']
|
2022-11-09
| null | null | null | null |
['facial-emotion-recognition']
|
['computer-vision']
|
[ 6.39464483e-02 6.20639563e-01 3.06453347e-01 -7.69934714e-01
-2.86198586e-01 -5.00952184e-01 4.21493679e-01 -3.22916865e-01
-2.22964272e-01 3.50768983e-01 -3.28486040e-02 4.51399654e-01
3.18266034e-01 -7.00228333e-01 -5.13900757e-01 -5.94337225e-01
-1.85462311e-01 4.73701954e-01 -5.76983452e-01 -5.86514652e-01
-2.20424563e-01 6.39883101e-01 -2.17785954e+00 6.56698167e-01
2.56409168e-01 1.38656950e+00 -2.77712673e-01 5.01877069e-01
-6.90103844e-02 9.21688795e-01 -4.95181859e-01 -8.29325020e-01
4.41175997e-01 -4.56072211e-01 -8.04197073e-01 3.21008444e-01
3.25161606e-01 -3.94069999e-01 2.05696046e-01 7.84650981e-01
3.44471842e-01 9.25453529e-02 3.33241880e-01 -1.75172150e+00
-2.39961073e-01 2.17765808e-01 -5.38655937e-01 -4.66004968e-01
6.30137384e-01 2.38459602e-01 7.29478657e-01 -8.18603039e-01
1.05622280e+00 1.06319451e+00 7.23328888e-01 1.21737421e+00
-1.12226784e+00 -9.16249812e-01 -1.41379327e-01 1.19286798e-01
-1.42166853e+00 -6.16942465e-01 8.45661342e-01 -3.95125777e-01
7.01331258e-01 3.98165792e-01 1.09507251e+00 9.90233243e-01
-5.86241603e-01 4.81018692e-01 1.32182789e+00 -5.56261718e-01
1.86612651e-01 4.25551891e-01 -1.97112262e-01 1.02458572e+00
-6.36462927e-01 3.23149025e-01 -6.64397717e-01 1.65369157e-02
4.99936432e-01 -1.76298350e-01 1.11152463e-01 -1.49961889e-01
-4.06669408e-01 7.35423028e-01 2.42453635e-01 1.35391384e-01
-6.72922075e-01 1.41129479e-01 7.17797279e-01 3.73233318e-01
6.51280046e-01 5.50698280e-01 -5.06665170e-01 -6.25456750e-01
-7.01053858e-01 1.01726063e-01 8.02440703e-01 5.35029113e-01
1.12050653e+00 2.09462596e-03 1.41557485e-01 8.43550920e-01
8.48985910e-02 2.29845539e-01 3.24113220e-01 -1.15199947e+00
-3.95319551e-01 9.01833653e-01 -7.99195394e-02 -1.36045146e+00
-5.25994718e-01 3.02283883e-01 -5.39969683e-01 5.79847276e-01
4.71543580e-01 -3.11175108e-01 -6.39754117e-01 1.94660437e+00
6.10391259e-01 1.69472575e-01 -2.69947723e-02 1.04770446e+00
9.49307978e-01 3.09582889e-01 3.54766488e-01 -1.12350062e-01
1.68118858e+00 -6.20180190e-01 -6.97971225e-01 1.02684356e-01
6.45834565e-01 -4.05513167e-01 1.27686799e+00 3.94213885e-01
-8.72086108e-01 -4.09990788e-01 -5.67645371e-01 1.36407256e-01
-6.66690707e-01 9.03364271e-02 1.21487427e+00 1.06371355e+00
-1.18112290e+00 4.85008985e-01 -3.60419750e-01 -2.78675467e-01
8.07140350e-01 6.79911196e-01 -9.75123763e-01 2.14881167e-01
-9.73993480e-01 8.15058589e-01 6.32619485e-02 2.19949141e-01
-7.86663294e-01 -7.06277132e-01 -9.59210813e-01 7.71171600e-02
3.77587587e-01 -1.79183349e-01 1.31162691e+00 -2.11027598e+00
-1.92202699e+00 1.75257981e+00 -7.73023441e-02 1.94900762e-02
3.58539760e-01 4.47536744e-02 -3.44607741e-01 2.95599490e-01
-2.28180677e-01 1.04104340e+00 8.38307977e-01 -1.25024068e+00
-1.70180440e-01 -7.08296001e-01 2.82089319e-02 2.13322826e-02
-2.18174651e-01 4.74552989e-01 -2.48981297e-01 -1.45744443e-01
-4.07365680e-01 -7.89326370e-01 -1.09922320e-01 1.67650864e-01
4.19543264e-03 -1.35566384e-01 5.14170885e-01 -5.62659800e-01
6.91361725e-01 -2.37985611e+00 -1.12707809e-01 5.62911689e-01
3.79836351e-01 2.13455558e-01 -2.17544734e-01 1.66709647e-01
-5.55286050e-01 2.33203277e-01 2.15933651e-01 -5.27262688e-01
2.09542915e-01 1.82236254e-01 -2.09553599e-01 2.29742259e-01
3.02985609e-01 1.02926540e+00 -7.60268152e-01 -4.42229182e-01
3.57314683e-02 4.58379865e-01 -6.62737966e-01 4.96308059e-01
-1.05379254e-01 3.60064358e-01 -8.15510526e-02 6.40544295e-01
5.73921382e-01 3.80598545e-01 2.15288252e-01 -9.73315910e-02
2.19920114e-01 -3.74757975e-01 -8.61249447e-01 1.43678403e+00
-4.73888159e-01 6.23330772e-01 2.58147627e-01 -7.16501236e-01
1.33011067e+00 4.73347604e-01 7.82883286e-01 -7.90119171e-01
4.87592071e-01 -4.55314070e-02 -3.11111301e-01 -7.83432126e-01
4.91195679e-01 -7.00206041e-01 -7.20464811e-02 7.21394420e-01
3.80519181e-01 -3.99744421e-01 -9.82259810e-02 -2.15479463e-01
7.58514643e-01 3.45713884e-01 2.30434075e-01 -5.47494832e-03
2.51349539e-01 -5.52368462e-02 4.13957298e-01 1.73562184e-01
-3.63570988e-01 2.68073261e-01 7.22112775e-01 -8.33361566e-01
-8.63627076e-01 -4.77308631e-01 2.88460672e-01 1.62588966e+00
-3.55736256e-01 -5.74782550e-01 -1.20276809e+00 -5.45505047e-01
-3.95380437e-01 2.38349095e-01 -1.16612732e+00 -2.35661015e-01
2.24190261e-02 -3.40587854e-01 8.84333551e-01 2.06405655e-01
3.81355613e-01 -1.48021615e+00 -8.85711253e-01 -2.40317538e-01
-2.02134907e-01 -1.08794427e+00 2.19602343e-02 -1.27026057e-02
-3.31811488e-01 -1.22487569e+00 -2.17889816e-01 -6.30783796e-01
7.67322779e-01 -3.53454173e-01 1.29484582e+00 3.77225995e-01
-6.21555030e-01 6.43167973e-01 -4.94868994e-01 -9.30603743e-01
-2.83430845e-01 -3.94373000e-01 3.11089363e-02 4.19131875e-01
7.10816622e-01 -5.61812520e-01 -2.57305235e-01 2.26595134e-01
-8.20691764e-01 1.74652562e-01 5.06326929e-02 6.88268661e-01
4.53718871e-01 -1.80907249e-01 4.70163852e-01 -8.36043358e-01
7.64153540e-01 -3.53636354e-01 -2.09611803e-01 7.61751607e-02
-1.75143495e-01 -3.79908085e-01 4.69473511e-01 -4.50945705e-01
-1.10817468e+00 5.49004018e-01 -2.92758286e-01 -3.87073040e-01
-6.59045219e-01 2.48556778e-01 -1.85454652e-01 -3.23693901e-01
7.53224552e-01 -1.01751193e-01 5.66005349e-01 -1.12652719e-01
6.80103064e-01 6.10968292e-01 5.67648351e-01 -7.74391413e-01
3.38025570e-01 4.43478733e-01 1.13263121e-02 -7.74318933e-01
-4.10343885e-01 -1.66551873e-01 -8.06523979e-01 -8.85536432e-01
7.80318081e-01 -8.21786344e-01 -1.46301484e+00 6.23364210e-01
-9.92685735e-01 -5.70926309e-01 -6.70920432e-01 3.86702865e-02
-9.24588919e-01 -2.73204669e-02 -4.05060381e-01 -1.17069995e+00
-3.24430227e-01 -7.36732125e-01 1.25095487e+00 2.42696330e-01
-7.54746616e-01 -7.27937996e-01 -4.09115059e-03 3.05404752e-01
2.22279578e-01 5.93291163e-01 6.63565576e-01 -5.70961118e-01
1.74808100e-01 -4.34848636e-01 -1.58702970e-01 2.58184075e-01
6.82109967e-03 3.11797440e-01 -1.32597995e+00 4.52877969e-01
-8.59730989e-02 -1.05461287e+00 8.65387022e-02 -1.97592616e-01
1.17207575e+00 -3.21305275e-01 1.42203093e-01 5.08010507e-01
9.58968401e-01 6.02809563e-02 9.54074860e-01 1.65417120e-01
3.76709670e-01 1.19620323e+00 6.22346520e-01 5.75822055e-01
2.32852221e-01 8.66147399e-01 4.33132410e-01 -4.89317924e-01
4.11683053e-01 -2.44272232e-01 2.49353468e-01 -6.98458776e-02
-5.76206684e-01 3.55246067e-01 -7.44538367e-01 1.59455240e-01
-1.53019345e+00 -1.17889726e+00 1.38748899e-01 1.58990753e+00
8.67960334e-01 -6.70395374e-01 4.60550725e-01 6.49532825e-02
4.04128045e-01 -2.93545276e-01 -1.97601780e-01 -9.72817063e-01
-1.10580400e-01 7.06490517e-01 -3.01446766e-01 3.22325975e-01
-8.70618284e-01 1.24477422e+00 6.71391630e+00 6.43770516e-01
-1.40615821e+00 -5.79341315e-02 1.03152478e+00 -2.13109165e-01
-3.14016417e-02 -5.21699250e-01 -1.27940878e-01 8.11105222e-02
9.19233859e-01 -2.35882267e-01 7.53312886e-01 1.15534925e+00
3.53907198e-01 -8.53114575e-02 -1.11633813e+00 1.23938835e+00
1.89997494e-01 -9.82216001e-01 -2.03582555e-01 -8.47744755e-03
2.94284463e-01 -5.16935527e-01 6.48251772e-02 4.23758715e-01
1.20712526e-01 -1.54519999e+00 6.57909691e-01 6.86487019e-01
1.19572651e+00 -7.17483878e-01 7.31830120e-01 1.03548698e-01
-7.23683536e-01 6.53823167e-02 -1.58167779e-02 -6.70506358e-01
-1.91913962e-01 -7.58096427e-02 -1.05346358e+00 8.46529454e-02
7.60049403e-01 3.05186480e-01 -4.70779538e-01 3.15128803e-01
-2.38602296e-01 3.61949533e-01 -4.44758773e-01 -1.68872133e-01
-2.42754873e-02 -2.54428923e-01 -7.69118266e-03 9.51679528e-01
1.26131058e-01 4.02448505e-01 -1.77800208e-01 8.04133832e-01
-1.01440534e-01 5.24127185e-01 -7.85858989e-01 -1.27089083e-01
3.35436151e-03 1.72348189e+00 -6.23887956e-01 -1.44793585e-01
-1.24612316e-01 1.06043255e+00 4.88160342e-01 1.52153760e-01
-6.72709167e-01 1.46487087e-01 9.07034516e-01 1.31682485e-01
-1.74353540e-01 3.04459035e-01 -1.83254674e-01 -8.00270855e-01
-9.73255113e-02 -1.13436818e+00 2.56082445e-01 -1.25254083e+00
-1.20526433e+00 1.00297594e+00 -1.82831019e-01 -8.32333326e-01
-5.47405958e-01 -7.78091371e-01 -5.48691392e-01 7.72114217e-01
-9.17335868e-01 -1.51748633e+00 -8.45816791e-01 7.99132705e-01
7.15224072e-02 -2.42308408e-01 1.58020103e+00 2.36282587e-01
-3.71875763e-01 6.01555705e-01 -7.53471136e-01 2.72682369e-01
5.17166853e-01 -8.39883983e-01 -2.22825393e-01 1.83771193e-01
3.39418113e-01 3.18569869e-01 6.79679394e-01 -1.35182396e-01
-9.99497414e-01 -6.23294473e-01 6.07773066e-01 -4.14798200e-01
5.36592424e-01 -5.39601326e-01 -6.32912517e-01 6.64282262e-01
5.19740134e-02 1.37066171e-01 1.29826915e+00 3.27057779e-01
-3.98535073e-01 -3.48590463e-02 -1.50353289e+00 6.27407849e-01
9.07413721e-01 -8.28808069e-01 -3.25919032e-01 8.60555768e-02
1.09700955e-01 -3.89257371e-01 -7.09141612e-01 9.30660367e-02
9.25967455e-01 -1.21230853e+00 5.31574249e-01 -1.20983613e+00
7.81759083e-01 2.08243489e-01 -9.89888832e-02 -1.32408178e+00
2.00576156e-01 -7.31354952e-01 2.86336809e-01 1.39811432e+00
2.59833485e-01 -3.88960868e-01 1.09490526e+00 1.46550381e+00
2.04342961e-01 -8.00690532e-01 -8.37295294e-01 -1.96019724e-01
-2.57048041e-01 -7.92203128e-01 9.79209304e-01 1.09817278e+00
5.66760540e-01 -7.06761703e-02 -3.81695151e-01 -3.77301186e-01
2.81346347e-02 -1.28125831e-01 1.28212166e+00 -1.23207474e+00
-7.61216208e-02 -5.27990282e-01 -8.81726027e-01 -3.12013507e-01
7.25110054e-01 -8.80988121e-01 1.26142964e-01 -6.22446656e-01
7.91639984e-02 -3.64005864e-01 2.36905903e-01 9.47132111e-01
6.17024116e-02 7.48331130e-01 3.73072207e-01 -2.46899858e-01
-4.99155909e-01 5.08084238e-01 1.05246711e+00 2.34752536e-01
-1.20422192e-01 -2.22197309e-01 -8.04228663e-01 9.03118670e-01
6.80862606e-01 -2.10307315e-01 -1.64945781e-01 6.98554739e-02
4.46560949e-01 -6.88739717e-02 6.40381992e-01 -5.05729079e-01
-1.79423302e-01 -7.79214427e-02 3.29213113e-01 2.75678545e-01
7.17950284e-01 -1.20099390e+00 3.17164242e-01 1.10460557e-02
-1.96307406e-01 -2.46232867e-01 6.31440222e-01 -1.45450728e-02
-4.02349383e-01 -1.03077210e-01 6.79647982e-01 -2.65972197e-01
-1.07551527e+00 1.69931442e-01 -5.08838058e-01 -3.23226392e-01
1.43682325e+00 -4.66769069e-01 3.20832692e-02 -8.14545631e-01
-1.23018467e+00 -1.23026803e-01 5.59842408e-01 2.70529270e-01
4.81182396e-01 -1.31655371e+00 -5.25024414e-01 4.55693036e-01
2.49932960e-01 -3.49692494e-01 4.53101814e-01 6.24259353e-01
-4.04054016e-01 -2.81295598e-01 -8.15861464e-01 -3.53867054e-01
-1.75983000e+00 3.06141943e-01 7.08290577e-01 -4.43501398e-02
-2.20128000e-01 9.71211731e-01 8.32491294e-02 -6.35975957e-01
-7.24870488e-02 2.77767122e-01 -3.80749732e-01 4.35089499e-01
6.71070457e-01 -6.23533211e-04 -4.71464992e-02 -1.12263322e+00
-2.19268054e-01 3.41802925e-01 5.15163958e-01 -1.91537708e-01
1.57895803e+00 3.09202522e-02 -4.45088267e-01 3.70938420e-01
1.00693607e+00 4.35121320e-02 -1.13507271e+00 2.94622451e-01
-1.04545780e-01 -5.23549139e-01 -2.49857590e-01 -9.34174836e-01
-1.23539507e+00 8.14416945e-01 6.83459163e-01 1.52861461e-01
1.58396006e+00 -1.17682647e-02 2.30542496e-01 1.63169816e-01
5.54602325e-01 -1.11859906e+00 1.14068627e-01 3.60087067e-01
9.29828525e-01 -1.26120961e+00 -4.66507554e-01 -5.95862091e-01
-1.06996822e+00 1.28401792e+00 8.14285517e-01 1.51365131e-01
5.50808311e-01 2.91701108e-01 5.09865105e-01 -7.63077199e-01
-5.88263035e-01 -3.17395687e-01 1.45905674e-01 1.04948521e+00
3.18586230e-01 2.64470041e-01 6.85548708e-02 1.25423634e+00
-7.67152727e-01 4.19668406e-01 3.13895673e-01 3.04469883e-01
4.24122438e-03 -8.62200677e-01 -2.47901738e-01 1.87606350e-01
-3.85496944e-01 1.19784646e-01 -9.89622176e-01 9.29971278e-01
4.50776339e-01 8.47951233e-01 2.45889947e-01 -6.12094581e-01
5.09163141e-01 5.17175615e-01 5.67343712e-01 -6.27548575e-01
-7.89576113e-01 -4.28580076e-01 2.81946242e-01 -9.10538793e-01
-5.34409463e-01 -4.41701531e-01 -1.44818246e+00 -3.71102482e-01
5.84700555e-02 2.80628502e-01 8.35026085e-01 9.34863210e-01
4.49676931e-01 1.35328963e-01 7.20141411e-01 -1.17145336e+00
2.09713742e-01 -8.90869617e-01 -8.70987177e-01 1.01086199e+00
-1.03922278e-01 -6.44026339e-01 -1.56679854e-01 4.52335775e-01]
|
[13.518876075744629, 1.7856042385101318]
|
d6f09529-375e-457d-85cb-920f5a729aa0
|
few-nerd-a-few-shot-named-entity-recognition
|
2105.07464
| null |
https://arxiv.org/abs/2105.07464v6
|
https://arxiv.org/pdf/2105.07464v6.pdf
|
Few-NERD: A Few-Shot Named Entity Recognition Dataset
|
Recently, considerable literature has grown up around the theme of few-shot named entity recognition (NER), but little published benchmark data specifically focused on the practical and challenging task. Current approaches collect existing supervised NER datasets and re-organize them to the few-shot setting for empirical study. These strategies conventionally aim to recognize coarse-grained entity types with few examples, while in practice, most unseen entity types are fine-grained. In this paper, we present Few-NERD, a large-scale human-annotated few-shot NER dataset with a hierarchy of 8 coarse-grained and 66 fine-grained entity types. Few-NERD consists of 188,238 sentences from Wikipedia, 4,601,160 words are included and each is annotated as context or a part of a two-level entity type. To the best of our knowledge, this is the first few-shot NER dataset and the largest human-crafted NER dataset. We construct benchmark tasks with different emphases to comprehensively assess the generalization capability of models. Extensive empirical results and analysis show that Few-NERD is challenging and the problem requires further research. We make Few-NERD public at https://ningding97.github.io/fewnerd/.
|
['Zhiyuan Liu', 'Hai-Tao Zheng', 'Pengjun Xie', 'Xu Han', 'Xiaobin Wang', 'Yulin Chen', 'Guangwei Xu', 'Ning Ding']
|
2021-05-16
| null |
https://aclanthology.org/2021.acl-long.248
|
https://aclanthology.org/2021.acl-long.248.pdf
|
acl-2021-5
|
['few-shot-ner']
|
['natural-language-processing']
|
[-5.88939011e-01 -1.68222293e-01 -2.26520166e-01 -3.47004414e-01
-7.24221528e-01 -6.66939080e-01 4.34151620e-01 3.81747037e-01
-9.97222900e-01 8.83055806e-01 5.24556756e-01 3.86121683e-02
1.60294667e-01 -9.61398840e-01 -4.07379717e-01 -3.22507948e-01
4.33868878e-02 2.80012012e-01 2.23125577e-01 -3.90934438e-01
-1.22155575e-02 1.24372602e-01 -1.45977867e+00 -1.09762773e-01
8.79880190e-01 6.33020163e-01 2.93972343e-01 4.76537704e-01
-3.73234510e-01 6.56044185e-01 -7.26019025e-01 -7.25650728e-01
1.96345851e-01 -1.10290840e-01 -1.05278659e+00 -6.05159044e-01
2.57199168e-01 -3.26686092e-02 -6.15340710e-01 1.10587978e+00
1.08904362e+00 6.85268939e-01 5.62100351e-01 -9.89645243e-01
-1.17576623e+00 8.27721834e-01 -1.00944534e-01 6.62988424e-01
3.34179550e-01 -4.12812866e-02 1.36669528e+00 -1.21370018e+00
9.61835802e-01 7.54027426e-01 1.01771700e+00 1.01340199e+00
-6.32113516e-01 -5.04630506e-01 -2.60818332e-01 5.75940311e-02
-1.56987035e+00 -3.53328258e-01 1.93756148e-01 -2.49053657e-01
1.26383793e+00 4.25326489e-02 8.00991803e-02 1.24097550e+00
-9.97931808e-02 4.94729549e-01 1.04829788e+00 -4.69834387e-01
5.09127319e-01 -2.27791071e-01 9.76156890e-01 3.75752807e-01
6.59907758e-01 -8.78764987e-02 -1.81055635e-01 -1.51185736e-01
3.02778274e-01 1.66597679e-01 -3.17093521e-01 2.01091215e-01
-1.16390681e+00 5.21340489e-01 3.20055604e-01 8.11536849e-01
-5.97816646e-01 -3.80928248e-01 5.41267574e-01 2.16811955e-01
4.53634441e-01 7.75276899e-01 -8.13286960e-01 -4.35378641e-01
-8.76733899e-01 -1.05732597e-01 1.15399945e+00 1.43169606e+00
8.51793110e-01 -1.60877481e-01 -5.38803220e-01 1.10210955e+00
-3.12115103e-01 3.30765069e-01 8.33561420e-01 -4.35756981e-01
5.34110725e-01 6.86736107e-01 2.55515516e-01 -5.63217521e-01
-6.07188046e-01 3.13351490e-02 -9.42272186e-01 -4.49515730e-01
1.39715925e-01 -7.13220179e-01 -1.13041484e+00 1.46953750e+00
4.35115904e-01 5.24145246e-01 3.57931823e-01 8.21694851e-01
1.61601949e+00 8.15627515e-01 6.21722758e-01 1.19924687e-01
1.69876897e+00 -9.62562501e-01 -7.61733592e-01 -8.55185091e-02
9.16101456e-01 -4.79658872e-01 1.06476200e+00 -3.88404161e-01
-4.84212667e-01 -3.49434882e-01 -7.77444541e-01 -2.59811640e-01
-1.06057894e+00 1.59191504e-01 5.14553428e-01 6.59596205e-01
-6.41442776e-01 6.77608550e-01 -4.58324164e-01 -7.39990234e-01
2.16245666e-01 -1.99407175e-01 -6.81516945e-01 -3.09674621e-01
-1.91646981e+00 1.02080536e+00 7.79938936e-01 -1.76196292e-01
-6.15521371e-01 -9.15512681e-01 -1.08534956e+00 2.64995337e-01
5.26442587e-01 -5.74833751e-01 1.31576586e+00 7.39210695e-02
-8.77953231e-01 9.83877420e-01 -7.12322444e-02 -4.73627508e-01
-1.66656837e-01 -2.29455784e-01 -9.64884400e-01 -2.21423760e-01
5.51245749e-01 4.22503114e-01 -5.74970199e-03 -8.14393401e-01
-6.55254781e-01 -1.74668003e-02 3.34110796e-01 -5.99754192e-02
-6.70020580e-01 3.95569324e-01 -3.09029222e-01 -8.46575618e-01
-5.60976505e-01 -6.23497367e-01 -3.33643615e-01 -6.96631491e-01
-5.31668127e-01 -7.08628833e-01 2.26102337e-01 -3.47150385e-01
1.77430677e+00 -2.07400632e+00 -3.39969546e-01 -6.14981472e-01
2.15679020e-01 5.67633569e-01 -2.90391177e-01 8.31905365e-01
-1.14661358e-01 6.21081531e-01 -1.72851413e-01 -2.73717582e-01
3.03718835e-01 1.91757560e-01 -2.41766408e-01 1.41842902e-01
-3.94265689e-02 1.11164546e+00 -1.31784832e+00 -4.44132864e-01
-8.78924578e-02 3.84713471e-01 -5.04530631e-02 3.63657206e-01
3.75505358e-01 -1.70997530e-01 -5.69342315e-01 9.00311291e-01
4.96165842e-01 -1.99865714e-01 -1.96569830e-01 -1.91384658e-01
-4.00110096e-01 2.76149005e-01 -1.11828542e+00 1.68313920e+00
-3.56601894e-01 2.94205755e-01 -4.70371336e-01 -4.57587779e-01
9.08434987e-01 5.98276556e-01 1.68244049e-01 -4.76654947e-01
3.41955066e-01 2.65749753e-01 -4.45227236e-01 -3.72612298e-01
1.09887731e+00 -3.76611769e-01 -7.64192939e-01 2.11869508e-01
6.51330292e-01 3.11454743e-01 8.20812881e-01 2.08979666e-01
1.45452023e+00 -3.00550401e-01 9.38911557e-01 -1.99983418e-01
-4.46972214e-02 1.95568904e-01 1.02026534e+00 1.11960912e+00
-5.45280159e-01 6.79172099e-01 3.42946462e-02 -3.77547801e-01
-1.07947993e+00 -8.73561740e-01 -2.60560393e-01 1.33314896e+00
2.16192484e-01 -7.51327276e-01 -8.42798948e-01 -8.94795179e-01
-3.64775002e-01 1.03597510e+00 -7.78132558e-01 1.15907431e-01
-3.67552191e-01 -8.57288063e-01 1.15868914e+00 6.44598246e-01
5.21726549e-01 -1.43195379e+00 -3.58029038e-01 2.80434072e-01
-2.86338687e-01 -1.13649833e+00 -8.56292188e-01 3.47452223e-01
-5.08952618e-01 -1.15345097e+00 -1.09788847e+00 -1.16859174e+00
3.95889491e-01 3.52340996e-01 1.34248149e+00 -1.27434686e-01
-4.14709628e-01 3.24000746e-01 -9.61710036e-01 -3.90576184e-01
9.89533216e-02 3.88249159e-01 1.68561354e-01 -5.46769202e-01
1.00560641e+00 -2.62817144e-01 -2.51624376e-01 3.76830131e-01
-7.91801929e-01 -6.02741420e-01 6.07378185e-01 1.00094843e+00
5.33338606e-01 2.30611414e-01 5.24613261e-01 -1.05527866e+00
6.34941518e-01 -1.11140192e+00 -6.56546652e-02 4.60906982e-01
-3.33757520e-01 -1.30020082e-01 7.87194729e-01 -3.98542553e-01
-1.32097018e+00 -3.55446845e-01 -2.11563423e-01 -3.08976024e-01
-7.32621729e-01 6.81420267e-01 -1.63548395e-01 4.76456672e-01
8.71758282e-01 1.07903957e-01 -1.10443485e+00 -9.50345457e-01
4.99284297e-01 9.81071532e-01 6.70687258e-01 -7.31746554e-01
5.95686793e-01 3.93864810e-02 -6.08598351e-01 -1.08900928e+00
-1.38513136e+00 -1.09215987e+00 -7.06178010e-01 1.80021629e-01
1.01903212e+00 -1.12745512e+00 1.08173944e-01 5.45048714e-01
-1.12814379e+00 -2.56002039e-01 -6.61587298e-01 4.24883366e-01
-9.52003300e-02 2.79516757e-01 -9.63373125e-01 -4.94805366e-01
-7.10214853e-01 -4.12606776e-01 9.64647174e-01 5.92431128e-01
-7.08296597e-02 -8.82582903e-01 5.88246405e-01 -5.36871962e-02
2.61680126e-01 -4.10267040e-02 3.55322957e-01 -1.31955290e+00
1.55021906e-01 -4.54800487e-01 -5.17936796e-02 6.42035082e-02
5.30006401e-02 -4.17953670e-01 -7.42034733e-01 -7.53560960e-02
-2.80451238e-01 -4.82937425e-01 8.67269337e-01 -1.15100648e-02
7.05968916e-01 -1.27445042e-01 -3.25264394e-01 2.36225709e-01
1.41655219e+00 7.44331181e-02 6.27466023e-01 5.60793400e-01
7.01379776e-01 1.81318060e-01 9.35009003e-01 5.87223411e-01
5.49282849e-01 2.36892819e-01 -1.43050805e-01 2.17004374e-01
-1.84709743e-01 -4.16798651e-01 6.16320185e-02 1.18938398e+00
-3.81255090e-01 -6.53357685e-01 -1.30278468e+00 9.63651478e-01
-1.49910736e+00 -1.36235332e+00 1.35926694e-01 1.79662108e+00
9.53791320e-01 -9.91207808e-02 -1.58764422e-01 -4.13541883e-01
1.51428831e+00 2.54133552e-01 -3.38677704e-01 4.10382822e-03
-1.98350579e-01 3.60182673e-01 4.38718200e-01 -7.21247420e-02
-1.55895531e+00 1.20719433e+00 5.57202673e+00 1.12595582e+00
-4.47859436e-01 3.60489130e-01 1.92460656e-01 2.31557757e-01
7.86541775e-02 8.94362777e-02 -1.44748271e+00 6.27222776e-01
1.08642554e+00 -5.42230844e-01 -3.93535895e-03 1.15798140e+00
-3.30012411e-01 3.49991590e-01 -7.73336411e-01 7.89986908e-01
1.13788970e-01 -1.05426943e+00 -2.24664062e-01 -1.58856332e-01
9.06042755e-01 4.27818596e-01 -7.37538815e-01 1.04695976e+00
6.05509698e-01 -8.39679837e-01 4.22648668e-01 6.25931263e-01
9.33135033e-01 -6.26093507e-01 1.18995774e+00 4.18925345e-01
-1.61339402e+00 5.37257791e-02 -8.21655512e-01 -1.37892663e-02
4.02705014e-01 6.71294808e-01 -2.49593407e-01 6.63623929e-01
1.14294457e+00 8.80742908e-01 -3.14917356e-01 1.34423351e+00
-3.54111910e-01 6.73094094e-01 -2.02532187e-01 -3.43057483e-01
3.52124989e-01 2.58900136e-01 6.33951843e-01 1.59981322e+00
5.32930970e-01 7.57852852e-01 2.01795891e-01 3.47027183e-01
-7.11230218e-01 2.84064025e-01 -5.54448605e-01 -4.28783566e-01
1.09902525e+00 1.60774899e+00 -6.84456408e-01 -6.57654643e-01
-7.10413575e-01 9.88645613e-01 7.53182590e-01 2.18080670e-01
-7.70322502e-01 -1.20632160e+00 7.05625951e-01 -3.62012774e-01
6.51236057e-01 -1.74824968e-02 1.07260615e-01 -1.79908109e+00
-2.68742740e-01 -3.80063921e-01 8.93545151e-01 -6.49757802e-01
-2.20152879e+00 8.84005129e-01 -2.76879281e-01 -1.27693212e+00
2.02650070e-01 -4.90250677e-01 -8.59152853e-01 5.48272312e-01
-1.42827153e+00 -9.10267353e-01 -3.49218309e-01 3.75324339e-01
6.57171607e-01 -5.79973832e-02 1.07413375e+00 4.98939812e-01
-9.12991345e-01 6.73042715e-01 2.20961630e-01 8.64183962e-01
1.04568732e+00 -1.40995121e+00 6.87159956e-01 9.12285447e-01
2.70483159e-02 9.88423586e-01 5.31603754e-01 -8.18905234e-01
-9.14860189e-01 -1.37566769e+00 1.56694078e+00 -6.94606781e-01
9.35519874e-01 -1.39707252e-01 -9.25378144e-01 7.62291014e-01
3.67464960e-01 4.89483654e-01 9.62208390e-01 2.23228768e-01
-4.57103640e-01 3.62675309e-01 -1.22503865e+00 4.80240464e-01
1.25959432e+00 -4.93220508e-01 -1.44790530e+00 -3.03245708e-02
8.62079144e-01 -3.20148915e-01 -1.40842140e+00 2.64144570e-01
4.08254825e-02 -5.52287817e-01 8.16421866e-01 -8.75600338e-01
3.36821973e-01 -1.66128874e-01 -1.70053452e-01 -1.69963884e+00
-4.28491652e-01 -2.60317504e-01 -3.12337041e-01 1.79556847e+00
4.42264408e-01 -4.29959476e-01 3.74321103e-01 7.62837827e-01
-4.99311596e-01 -5.67352831e-01 -8.29366446e-01 -1.18114483e+00
2.69315720e-01 -5.99244535e-02 7.75152564e-01 1.48492467e+00
2.03637898e-01 4.86681610e-01 -5.33871651e-01 1.56914875e-01
4.82948631e-01 -2.51641199e-02 3.99062485e-01 -1.19976759e+00
8.78959000e-02 -1.93289414e-01 -4.41055417e-01 -8.04384172e-01
2.54322082e-01 -9.16831255e-01 4.02717590e-01 -1.66395509e+00
5.23164988e-01 -3.71205449e-01 -6.17781579e-01 1.06187499e+00
-5.58412015e-01 1.04516290e-01 9.77411494e-02 2.45680019e-01
-1.10981524e+00 5.58783174e-01 6.92466199e-01 -6.93057105e-02
-2.45014150e-02 -4.94219184e-01 -7.10421920e-01 5.76053560e-01
8.65409195e-01 -7.32671678e-01 1.04597941e-01 -2.80177623e-01
2.58145072e-02 -2.95525461e-01 -5.98481596e-02 -9.07700300e-01
6.51518822e-01 -7.21348375e-02 4.86770682e-02 -4.11321461e-01
-9.09022391e-02 -4.22950059e-01 -1.35508537e-01 -4.54523750e-02
-2.40032330e-01 -7.89408237e-02 -3.89772616e-02 6.04012132e-01
-3.97900641e-01 -6.80571079e-01 5.44816494e-01 -3.56776923e-01
-1.81114554e+00 6.98079467e-01 -2.17438906e-01 1.03704584e+00
1.10093272e+00 7.88578987e-02 -6.02137208e-01 2.48048857e-01
-6.93747342e-01 2.24609524e-01 4.16117251e-01 6.63245618e-01
3.05271924e-01 -1.39168489e+00 -6.64619148e-01 -5.64402103e-01
6.16373777e-01 -1.28246367e-01 7.43295789e-01 4.14960355e-01
-8.43360424e-02 2.80142069e-01 -5.68897314e-02 1.85183451e-01
-9.19408202e-01 8.03969383e-01 8.37279335e-02 -3.14618319e-01
-7.92865515e-01 7.25215197e-01 -1.77245781e-01 -9.12980855e-01
-1.69207167e-03 4.82815243e-02 -5.81894100e-01 3.40939701e-01
9.55511451e-01 7.93122709e-01 -2.19837427e-02 -7.79677093e-01
-4.34445590e-01 3.14483374e-01 -4.65680547e-02 5.22815168e-01
1.60003948e+00 -2.28233188e-01 -1.07182749e-01 6.30179226e-01
1.13610005e+00 -7.79621899e-02 -7.52659082e-01 -3.02719206e-01
3.48318070e-01 -5.18415459e-02 -7.26464018e-02 -7.47314155e-01
-7.15723872e-01 5.23288190e-01 2.22356677e-01 2.33527899e-01
8.22634637e-01 2.59101123e-01 1.24914598e+00 8.27197075e-01
7.63599396e-01 -1.29890335e+00 -4.18554515e-01 1.19429147e+00
2.31386542e-01 -1.19568789e+00 -3.72760147e-01 -2.64793813e-01
-7.68818259e-01 8.38083029e-01 8.44685555e-01 -1.68328375e-01
7.35432863e-01 2.82923639e-01 2.11150330e-02 -2.69299001e-01
-6.16423666e-01 -6.93527639e-01 2.94264019e-01 3.46936971e-01
5.85848808e-01 1.07257381e-01 -5.89852870e-01 1.50065291e+00
-6.92769289e-02 1.21407270e-01 5.35439134e-01 1.16432941e+00
-7.22573221e-01 -8.54836762e-01 2.34736260e-02 5.14438093e-01
-6.19497180e-01 -6.88825428e-01 -1.00334473e-01 7.71892667e-01
1.38171479e-01 1.00805247e+00 -1.25129402e-01 -1.53502434e-01
6.83746457e-01 2.52120763e-01 -1.29368320e-01 -1.01037228e+00
-5.49745858e-01 -7.94160545e-01 3.84514481e-01 -2.43475839e-01
-3.35972488e-01 -3.27102304e-01 -1.43334544e+00 -2.91189432e-01
-4.29216266e-01 5.70636153e-01 2.19776571e-01 8.71853650e-01
6.12026215e-01 1.51743948e-01 3.72726232e-01 -7.00142145e-01
-4.64093804e-01 -1.08624387e+00 -9.39591348e-01 6.31778717e-01
-1.84035048e-01 -7.75833130e-01 -6.10687256e-01 -2.69728690e-01]
|
[9.66561222076416, 9.386706352233887]
|
00e0a591-9972-4360-b8d9-07318c86b4d6
|
modeling-feature-representations-for
|
1911.00030
| null |
https://arxiv.org/abs/1911.00030v1
|
https://arxiv.org/pdf/1911.00030v1.pdf
|
Modeling Feature Representations for Affective Speech using Generative Adversarial Networks
|
Emotion recognition is a classic field of research with a typical setup extracting features and feeding them through a classifier for prediction. On the other hand, generative models jointly capture the distributional relationship between emotions and the feature profiles. Relatively recently, Generative Adversarial Networks (GANs) have surfaced as a new class of generative models and have shown considerable success in modeling distributions in the fields of computer vision and natural language understanding. In this work, we experiment with variants of GAN architectures to generate feature vectors corresponding to an emotion in two ways: (i) A generator is trained with samples from a mixture prior. Each mixture component corresponds to an emotional class and can be sampled to generate features from the corresponding emotion. (ii) A one-hot vector corresponding to an emotion can be explicitly used to generate the features. We perform analysis on such models and also propose different metrics used to measure the performance of the GAN models in their ability to generate realistic synthetic samples. Apart from evaluation on a given dataset of interest, we perform a cross-corpus study where we study the utility of the synthetic samples as additional training data in low resource conditions.
|
['Carol Espy-Wilson', 'Rahul Gupta', 'Saurabh Sahu']
|
2019-10-31
| null | null | null | null |
['cross-corpus']
|
['computer-vision']
|
[ 4.88685459e-01 3.61749709e-01 1.59819484e-01 -5.13141215e-01
-7.59218752e-01 -5.16480684e-01 1.17229068e+00 -5.95910586e-02
-1.64286926e-01 8.10989678e-01 2.96260417e-01 3.44612986e-01
3.86364937e-01 -9.99717176e-01 -8.30849767e-01 -8.44926357e-01
2.08158717e-01 7.58202612e-01 -5.25339723e-01 -1.20639838e-01
-1.91763818e-01 3.93949270e-01 -1.70698273e+00 3.37957472e-01
8.92486632e-01 1.00492549e+00 -1.34826809e-01 7.05363929e-01
-1.65809959e-01 5.94594121e-01 -1.15639961e+00 -6.59383059e-01
1.88446179e-01 -1.02689278e+00 -4.70872402e-01 1.57872781e-01
-1.15712278e-01 2.20860034e-01 6.01836741e-02 8.91064823e-01
5.57016611e-01 7.74116954e-03 1.18349683e+00 -1.56027126e+00
-5.43609798e-01 6.40618086e-01 -8.24243724e-02 -3.63714665e-01
4.31715101e-01 2.28037775e-01 8.48495781e-01 -6.33438826e-01
7.12309599e-01 1.18586051e+00 3.36430043e-01 8.76556933e-01
-1.39168465e+00 -4.91390973e-01 -3.54869843e-01 2.68481988e-02
-1.05440021e+00 -3.62050563e-01 1.00711668e+00 -3.64656031e-01
4.43366408e-01 2.06790149e-01 7.30781913e-01 1.78342378e+00
-3.97038050e-02 9.20355260e-01 1.37289238e+00 -6.14228308e-01
5.94696581e-01 6.08363271e-01 -2.69768625e-01 1.58740520e-01
-2.15563923e-01 1.82353675e-01 -3.91166717e-01 -1.96595624e-01
3.50391686e-01 -2.35095084e-01 -3.07125539e-01 -2.37043053e-01
-7.30558693e-01 1.30560470e+00 3.29142094e-01 3.48643064e-01
-7.66817510e-01 2.75935143e-01 2.71869093e-01 3.16338271e-01
6.74507201e-01 6.24558806e-01 -4.98200394e-02 -1.68340981e-01
-9.57746923e-01 4.59527075e-01 9.92985427e-01 8.47232938e-01
8.11069071e-01 4.33018208e-01 -6.35062456e-01 9.56577659e-01
1.42906293e-01 4.39947158e-01 8.03668201e-01 -4.60183233e-01
1.45858854e-01 4.80371267e-01 1.36978542e-02 -6.92888021e-01
2.43082494e-02 -4.22197014e-01 -8.46618235e-01 2.34334558e-01
3.95562142e-01 -5.41268289e-01 -9.91940677e-01 2.11358213e+00
2.16484055e-01 2.51138568e-01 4.79533881e-01 5.30732870e-01
6.28804684e-01 9.10328031e-01 5.69204986e-02 1.29373157e-02
1.15712500e+00 -6.80386007e-01 -5.38622379e-01 -2.35009730e-01
2.90644348e-01 -6.77158177e-01 9.93780971e-01 3.55967373e-01
-9.46776927e-01 -4.53768373e-01 -8.21747720e-01 5.08770227e-01
-4.82521385e-01 2.31256947e-01 4.92383420e-01 8.94307375e-01
-1.10471308e+00 5.11149883e-01 -6.02185071e-01 -2.74243772e-01
4.23792720e-01 -1.19157610e-02 -1.55843005e-01 -1.25527620e-01
-1.24003613e+00 7.91247368e-01 2.42225341e-02 -1.42821014e-01
-1.04786527e+00 -5.12475908e-01 -8.74171138e-01 1.32292792e-01
-8.43723267e-02 -6.72078669e-01 9.50289011e-01 -1.64329493e+00
-1.83180618e+00 7.63142109e-01 -9.36987400e-02 -5.63189864e-01
6.07636333e-01 2.72472017e-02 -3.92930776e-01 -4.50958833e-02
-8.54611322e-02 8.60987484e-01 1.22820711e+00 -1.40492308e+00
-2.14639500e-01 -4.30122465e-02 -2.34484404e-01 -7.13376626e-02
-1.97328597e-01 -3.28493044e-02 -1.90942250e-02 -7.92056084e-01
-5.77215016e-01 -1.06679010e+00 -2.25938335e-01 -5.66669524e-01
-6.42501771e-01 -1.21308453e-01 5.97847462e-01 -3.09626192e-01
6.40804172e-01 -2.08632970e+00 2.98004359e-01 2.58344948e-01
-2.95427233e-01 2.02322274e-01 -1.86824724e-01 6.24552906e-01
-1.64610162e-01 2.19707057e-01 -4.20237124e-01 -6.74839377e-01
2.65325725e-01 1.96691662e-01 -5.09573758e-01 1.28991604e-01
6.55760646e-01 9.91072118e-01 -6.93836331e-01 2.32706908e-02
1.70977984e-03 7.62853324e-01 -5.46395719e-01 6.25037789e-01
-5.61702609e-01 6.38758183e-01 -3.97231996e-01 7.95150027e-02
3.72799337e-01 7.98112974e-02 5.06167039e-02 6.77099302e-02
4.56663758e-01 7.88466185e-02 -9.66372252e-01 1.32539403e+00
-7.32166767e-01 6.27531290e-01 -3.17032099e-01 -1.05063128e+00
1.12761366e+00 4.88059342e-01 1.25685140e-01 -3.04365098e-01
3.35249692e-01 1.15670115e-02 9.13540572e-02 -1.55766785e-01
4.04029906e-01 -5.03547549e-01 -4.52303171e-01 6.03354096e-01
4.08137769e-01 -4.23995048e-01 1.50654137e-01 -1.52760670e-01
9.89764571e-01 1.23109370e-01 2.30617285e-01 5.19829728e-02
3.99772912e-01 -2.42176637e-01 2.86796272e-01 5.57308853e-01
2.90062636e-01 8.64010811e-01 6.69920504e-01 -2.56701827e-01
-1.11379635e+00 -9.44811881e-01 1.17141262e-01 6.36773646e-01
-4.31368440e-01 -1.77566767e-01 -9.97839034e-01 -7.53135979e-01
-2.41743743e-01 1.25263023e+00 -9.50205564e-01 -4.73678231e-01
-1.26455098e-01 -8.49091411e-01 6.26956344e-01 3.92492473e-01
1.78473994e-01 -1.65495765e+00 -8.46731722e-01 9.83320326e-02
-1.08508572e-01 -9.31719899e-01 1.27533348e-02 4.99874264e-01
-3.89792591e-01 -6.79572344e-01 -8.46881568e-01 -4.48765457e-01
6.84468508e-01 -6.37893498e-01 1.43852496e+00 -3.80807042e-01
-1.60860971e-01 4.27385032e-01 -7.64049709e-01 -8.33188534e-01
-1.06944799e+00 9.76640806e-02 -1.64482847e-01 6.35631144e-01
3.16879094e-01 -6.65310025e-01 -3.34632218e-01 8.61154050e-02
-1.26519871e+00 -7.34801665e-02 5.87430954e-01 8.65091920e-01
5.48172712e-01 -3.07801589e-02 8.36034119e-01 -1.12284219e+00
9.17893410e-01 -8.11926365e-01 -2.01852292e-01 6.52263165e-02
-2.66839772e-01 3.05745095e-01 9.97820854e-01 -6.55371130e-01
-1.00242805e+00 1.25109494e-01 -3.08974504e-01 -7.19386816e-01
-6.47003293e-01 3.57718378e-01 -4.49106812e-01 4.18304414e-01
6.35567188e-01 4.17378277e-01 -2.62735616e-02 -2.46112481e-01
7.08192527e-01 5.57670772e-01 1.84806570e-01 -7.51802981e-01
7.82221973e-01 1.22594014e-01 -1.13316782e-01 -8.55576456e-01
-6.23455107e-01 2.54022002e-01 -2.22468838e-01 -9.81593505e-02
8.01182687e-01 -6.81613207e-01 4.82021191e-04 4.96575594e-01
-9.63255644e-01 -5.38163722e-01 -8.16034257e-01 3.17102611e-01
-8.39903951e-01 -2.09360912e-01 -2.82428741e-01 -8.94600034e-01
-2.97789544e-01 -1.11322975e+00 1.03876698e+00 4.52312410e-01
-4.80374545e-01 -9.93680477e-01 3.58660460e-01 -1.97580919e-01
5.33161104e-01 6.97359502e-01 9.03119266e-01 -1.01322639e+00
-3.94745206e-04 -4.90996361e-01 3.71842980e-01 6.99379683e-01
1.45552754e-01 -5.15374914e-02 -1.18261111e+00 -6.42664805e-02
1.12650022e-01 -5.60548425e-01 4.96955007e-01 2.67968327e-01
9.86118495e-01 -3.64442259e-01 -8.82528052e-02 5.21559536e-01
1.29665613e+00 2.58527309e-01 8.62858415e-01 -1.64519921e-01
4.08927083e-01 6.58307433e-01 2.80953199e-01 4.50375795e-01
-7.87419826e-02 7.65862226e-01 2.71788329e-01 -2.00826284e-02
-3.98195572e-02 -4.41206336e-01 3.60478640e-01 4.94538277e-01
1.16689771e-01 -6.88963354e-01 -5.19920409e-01 3.91667247e-01
-1.41126084e+00 -9.72702980e-01 2.19947144e-01 2.14834523e+00
7.87053287e-01 5.20545207e-02 2.95823961e-01 1.58657044e-01
6.37984276e-01 -2.08465457e-02 -4.33358073e-01 -5.93201756e-01
-1.84953034e-01 7.22043455e-01 -1.72512099e-01 2.78216098e-02
-8.38847876e-01 8.32553625e-01 5.88026476e+00 8.28779161e-01
-1.41229975e+00 5.05922362e-02 1.01267374e+00 -1.78058073e-01
-4.90043432e-01 -1.89949542e-01 -5.75161576e-01 7.62981892e-01
1.05964732e+00 -8.38373229e-02 4.37213093e-01 9.91158187e-01
-1.51301458e-01 5.63704744e-02 -1.31324303e+00 7.99677372e-01
2.75088310e-01 -9.77221072e-01 4.31202918e-01 2.38035664e-01
8.31517220e-01 -2.65336096e-01 2.43922770e-01 4.11911070e-01
4.34638947e-01 -1.40791130e+00 8.62824917e-01 7.71440506e-01
8.96934688e-01 -1.00900042e+00 7.78276980e-01 3.86260927e-01
-6.23174191e-01 2.96648592e-01 -2.00974897e-01 8.96284208e-02
2.05528393e-01 5.83866954e-01 -9.71296847e-01 4.67037380e-01
1.88049689e-01 2.34662235e-01 -4.41663414e-01 7.42895842e-01
-4.59189564e-01 1.02598870e+00 -3.51679087e-01 -1.59066156e-01
2.26018816e-01 -3.38242859e-01 4.69714016e-01 1.19703412e+00
5.74155807e-01 -3.88425112e-01 -1.68889031e-01 1.35764408e+00
-2.62119055e-01 1.85634851e-01 -7.37680197e-01 -3.02320421e-01
2.81261563e-01 1.46011114e+00 -6.62611544e-01 -2.87900329e-01
-1.24328382e-01 1.13469148e+00 3.83656204e-01 3.18599403e-01
-9.33847487e-01 -2.28409231e-01 6.22054219e-01 -2.19765250e-02
5.20761609e-01 3.26470166e-01 -2.35800305e-03 -1.07863951e+00
-2.11641029e-01 -1.01277840e+00 -1.17248826e-01 -7.23169029e-01
-1.57218337e+00 1.11587632e+00 -1.55085623e-01 -1.18307316e+00
-9.70931411e-01 -3.10581893e-01 -9.73214388e-01 1.15945911e+00
-1.16995072e+00 -1.04186952e+00 -3.77034605e-01 5.19474983e-01
4.51305240e-01 -4.42622542e-01 1.21112800e+00 -1.12422980e-01
-5.15855670e-01 7.19700933e-01 4.70261835e-02 1.22065112e-01
4.31705832e-01 -1.43845749e+00 3.92626047e-01 7.32293308e-01
5.65494478e-01 2.98275799e-01 9.28002477e-01 -3.19572866e-01
-1.01279426e+00 -1.19059443e+00 5.13939202e-01 -3.14402610e-01
1.18496135e-01 -6.92829788e-01 -6.79192483e-01 5.40009856e-01
4.53808904e-01 -9.64236706e-02 8.82067144e-01 -2.20096096e-01
-1.24889076e-01 1.78043574e-01 -1.33801770e+00 4.28998053e-01
5.57916701e-01 -2.87022114e-01 -2.99615830e-01 -8.45115352e-03
3.46434861e-01 -8.81318450e-02 -7.35798538e-01 1.14557452e-01
2.61881113e-01 -1.10038018e+00 5.68029463e-01 -6.38521969e-01
9.58676457e-01 6.45330362e-03 -2.05769762e-01 -2.15971661e+00
2.96166036e-02 -6.83206677e-01 1.94206554e-02 1.67666650e+00
4.58484471e-01 -4.72931445e-01 8.19109619e-01 4.63059455e-01
8.04391354e-02 -8.95030677e-01 -6.84798777e-01 -5.32082677e-01
2.44307548e-01 -4.47808027e-01 8.85408938e-01 6.67141140e-01
-3.01824927e-01 7.28425145e-01 -5.09164274e-01 -3.49122912e-01
2.08918497e-01 1.83118477e-01 1.08470857e+00 -1.04536068e+00
-6.20892823e-01 -5.49182236e-01 -4.04273838e-01 -5.69637477e-01
5.10609627e-01 -1.00086248e+00 1.11313455e-01 -1.20853364e+00
-5.81234880e-02 -5.30489802e-01 9.43473056e-02 1.97294012e-01
-8.56170505e-02 2.94373393e-01 2.78121412e-01 -1.98415071e-01
4.39061522e-02 8.98138404e-01 9.08015847e-01 -7.68981129e-03
-6.14669174e-02 3.41315150e-01 -7.34109521e-01 5.14407516e-01
7.53092468e-01 -5.00788987e-01 -6.32383347e-01 1.32892773e-01
2.15908006e-01 1.70326158e-01 2.31543273e-01 -1.04776847e+00
-4.32032555e-01 1.58308014e-01 6.15326881e-01 7.83602819e-02
5.50129235e-01 -7.12404847e-01 4.10946459e-01 1.12015024e-01
-4.26455647e-01 -1.76222622e-01 -6.60390034e-02 5.07187188e-01
-4.06470984e-01 -5.33883095e-01 7.43409753e-01 -6.25413433e-02
-2.68649518e-01 5.68361357e-02 -4.27768737e-01 3.19904625e-01
1.10846448e+00 2.87280288e-02 2.05035895e-01 -9.23892796e-01
-6.98503911e-01 -3.49034101e-01 6.24380469e-01 3.99341881e-01
3.52157295e-01 -1.38689923e+00 -8.71738732e-01 4.27227616e-01
2.42461357e-02 -1.45278946e-01 -5.54088363e-03 2.80055642e-01
-7.45328367e-02 6.45475388e-02 -2.90153593e-01 -4.23265249e-01
-8.30289721e-01 5.55298567e-01 3.63306165e-01 -5.42140186e-01
-2.35132605e-01 7.02403903e-01 3.05734962e-01 -4.30248976e-01
-1.16727047e-01 -4.66410108e-02 -3.26391160e-01 2.31631353e-01
2.59206682e-01 -8.65195543e-02 -5.54200150e-02 -6.94686055e-01
-2.37843003e-02 3.31909508e-02 3.77330482e-01 -3.23951930e-01
1.48844302e+00 4.61154282e-01 1.81806177e-01 5.56742668e-01
1.30911338e+00 1.78916633e-01 -1.20707369e+00 -1.79706663e-02
-1.71354681e-01 -1.55085072e-01 -4.24142152e-01 -9.22449529e-01
-1.22690010e+00 7.56232500e-01 4.11120385e-01 5.78206182e-01
1.19613874e+00 5.56962751e-02 4.74579275e-01 -2.91713059e-01
3.24345410e-01 -6.33657694e-01 1.26556635e-01 3.00518900e-01
8.95355761e-01 -8.77534330e-01 -5.86606562e-01 -2.59061635e-01
-1.05121088e+00 9.34607148e-01 2.83726990e-01 -5.23160100e-01
4.50070292e-01 3.78781289e-01 7.46229962e-02 -6.43816590e-02
-7.30177164e-01 -1.72694176e-01 3.67810309e-01 7.90301204e-01
4.65581566e-01 2.62387156e-01 -1.68839023e-01 8.16562593e-01
-6.72442019e-01 -5.72875142e-02 5.77304125e-01 4.60199267e-01
1.91774800e-01 -1.42302573e+00 -1.17593907e-01 6.32909298e-01
-4.27737087e-01 9.40115191e-03 -6.10938668e-01 6.76691234e-01
8.11646506e-02 8.88730764e-01 1.81274250e-01 -5.18905640e-01
1.54825836e-01 6.23450100e-01 4.22489166e-01 -6.87573910e-01
-7.22117901e-01 -8.01538080e-02 1.45954102e-01 -2.63386011e-01
-2.61761487e-01 -9.39813256e-01 -8.48592460e-01 2.32284009e-01
-7.36090392e-02 5.40855467e-01 7.70904958e-01 9.61700857e-01
4.45033044e-01 6.59495950e-01 7.78608322e-01 -9.99457181e-01
-5.54317832e-01 -1.29047716e+00 -6.23195350e-01 8.68841410e-01
-2.27142543e-01 -6.29952788e-01 -3.54532361e-01 1.30834147e-01]
|
[11.7438325881958, -0.012814158573746681]
|
903a6c02-427c-4371-84a5-34c587a8d0c1
|
c2msnet-a-novel-approach-for-single-image
|
1801.08406
| null |
http://arxiv.org/abs/1801.08406v1
|
http://arxiv.org/pdf/1801.08406v1.pdf
|
C2MSNet: A Novel approach for single image haze removal
|
Degradation of image quality due to the presence of haze is a very common
phenomenon. Existing DehazeNet [3], MSCNN [11] tackled the drawbacks of hand
crafted haze relevant features. However, these methods have the problem of
color distortion in gloomy (poor illumination) environment. In this paper, a
cardinal (red, green and blue) color fusion network for single image haze
removal is proposed. In first stage, network fusses color information present
in hazy images and generates multi-channel depth maps. The second stage
estimates the scene transmission map from generated dark channels using multi
channel multi scale convolutional neural network (McMs-CNN) to recover the
original scene. To train the proposed network, we have used two standard
datasets namely: ImageNet [5] and D-HAZY [1]. Performance evaluation of the
proposed approach has been carried out using structural similarity index
(SSIM), mean square error (MSE) and peak signal to noise ratio (PSNR).
Performance analysis shows that the proposed approach outperforms the existing
state-of-the-art methods for single image dehazing.
|
['Subrahmanyam Murala', 'Akshay Dudhane']
|
2018-01-25
| null | null | null | null |
['single-image-haze-removal']
|
['computer-vision']
|
[ 2.45927706e-01 -6.18274093e-01 7.98263669e-01 -4.09834087e-02
-3.60214978e-01 -2.04778194e-01 2.69604594e-01 -4.11754996e-02
-5.18362880e-01 8.31398964e-01 1.27117038e-01 -4.55538556e-02
-1.36772528e-01 -8.27973247e-01 -4.91705835e-01 -1.04082334e+00
6.14013821e-02 -3.68311077e-01 5.37884951e-01 -5.90097070e-01
6.84829772e-01 4.25032675e-01 -1.86211848e+00 2.66235888e-01
8.66336167e-01 1.08868718e+00 3.30535889e-01 1.11229682e+00
3.15901190e-01 1.21765792e+00 -9.37335491e-01 -5.31173125e-02
6.05698884e-01 -4.35972691e-01 -4.42446381e-01 1.25597194e-01
6.62555993e-01 -7.59787738e-01 -4.64591682e-01 1.40467799e+00
6.62382305e-01 2.79057264e-01 6.55683935e-01 -1.22037542e+00
-8.14270556e-01 -2.04918578e-01 -6.02963090e-01 8.59261811e-01
-7.36445040e-02 2.80614555e-01 2.10561827e-01 -1.01049089e+00
3.57747525e-01 1.16184700e+00 4.17981267e-01 1.40931040e-01
-6.06127322e-01 -8.62175703e-01 -5.84325552e-01 5.64750373e-01
-1.53874910e+00 -3.02893698e-01 7.98740268e-01 -2.18254760e-01
7.90543258e-01 1.86101303e-01 3.00150216e-01 2.11752728e-01
7.14574873e-01 3.77220511e-01 1.71896982e+00 -5.43998420e-01
9.68640894e-02 1.40895486e-01 -2.52716213e-01 5.18981755e-01
4.24474537e-01 3.74249101e-01 -3.10399711e-01 3.64641219e-01
6.07157767e-01 1.94339991e-01 -2.91816086e-01 2.52910197e-01
-7.44407117e-01 6.30981982e-01 5.98008692e-01 4.51308280e-01
-4.71689016e-01 2.39282519e-01 -1.91652209e-01 5.98151803e-01
2.33089432e-01 1.83492079e-01 7.28423670e-02 1.87766150e-01
-1.14155364e+00 1.38075858e-01 4.28444862e-01 7.26049840e-01
9.88883853e-01 5.65719306e-01 1.45764425e-01 5.55307209e-01
4.78920549e-01 5.92743576e-01 4.36421722e-01 -1.00197065e+00
2.42152661e-01 2.30907366e-01 2.31209323e-01 -1.25491858e+00
-4.25718650e-02 -2.86973953e-01 -1.15853977e+00 1.01324284e+00
2.47203335e-02 -1.24092616e-01 -1.43762457e+00 8.52490604e-01
1.15440123e-01 2.46395320e-01 4.00629669e-01 1.08820283e+00
9.59194720e-01 1.08500218e+00 -3.13012391e-01 -3.05531006e-02
1.03477311e+00 -9.58555222e-01 -1.04055893e+00 -1.60631090e-02
-1.28559500e-01 -1.26257086e+00 3.68684560e-01 9.26651716e-01
-1.39839458e+00 -6.84837401e-01 -1.44248593e+00 -1.00715093e-01
-8.43456209e-01 -1.09911516e-01 1.28586516e-01 7.95748651e-01
-1.42381847e+00 5.37974894e-01 -4.04541284e-01 -1.89395070e-01
3.26142728e-01 4.50398862e-01 -2.99796194e-01 -5.15924156e-01
-1.18633449e+00 1.17177141e+00 6.07974410e-01 4.15033817e-01
-1.48677838e+00 -3.40749532e-01 -5.56852520e-01 -9.07634348e-02
-1.92629527e-02 -3.61864626e-01 6.38116479e-01 -1.20386612e+00
-1.31264246e+00 4.54018265e-01 3.25793058e-01 -3.03069770e-01
1.28169388e-01 -3.58337998e-01 -7.26062775e-01 4.78469849e-01
-3.47178787e-01 4.39549297e-01 1.19206154e+00 -1.54665101e+00
-8.62292945e-01 -2.09455565e-01 6.31439537e-02 3.95632088e-01
-1.97001863e-02 1.94686845e-01 -1.04890600e-01 -6.96459591e-01
-5.06801941e-02 -3.64863515e-01 5.39665483e-03 4.47666831e-02
-2.08778977e-01 4.65312779e-01 1.18907177e+00 -9.82531607e-01
1.15120816e+00 -2.22029924e+00 -6.24870099e-02 1.40702307e-01
2.27523506e-01 6.68541789e-01 2.55080592e-02 5.61259389e-01
1.27569940e-02 -1.86863020e-01 -3.48709762e-01 -1.09443016e-01
-5.06825686e-01 6.23401739e-02 1.81405097e-01 9.65116262e-01
4.25156355e-02 2.86848605e-01 -5.63089013e-01 -4.31210309e-01
7.42551088e-01 8.92694592e-01 -2.37870887e-01 4.26223308e-01
2.62875974e-01 3.08608085e-01 1.11086003e-01 7.97176003e-01
1.20527291e+00 2.43659317e-01 -6.18520081e-01 -3.43698382e-01
-4.29848373e-01 -4.81172413e-01 -1.42393029e+00 1.24649036e+00
-4.68022108e-01 1.05552781e+00 2.28521630e-01 -7.26494253e-01
8.74595881e-01 4.53992099e-01 -3.08097731e-02 -7.69262373e-01
4.66889709e-01 3.34326804e-01 -1.46001831e-01 -8.62825036e-01
7.35768795e-01 -2.52421468e-01 7.11081684e-01 -2.29910556e-02
1.01371005e-01 -3.51319253e-01 -6.59174621e-02 -3.97499353e-02
7.97340930e-01 -2.11069420e-01 1.88415185e-01 -9.62963626e-02
8.50934803e-01 -7.15797544e-02 1.13608249e-01 5.88641167e-01
-4.96790797e-01 8.50721776e-01 -1.87988803e-01 -4.14230675e-01
-1.37813365e+00 -8.36279929e-01 2.01860949e-01 5.10348499e-01
5.02180278e-01 2.89979070e-01 -6.76806509e-01 2.41381109e-01
-3.48475307e-01 7.00173914e-01 -6.89052343e-01 -1.75202787e-01
-5.75963318e-01 -7.83693910e-01 4.72435474e-01 -2.53574818e-01
1.20637989e+00 -9.35678661e-01 -8.20166826e-01 1.22080594e-01
-9.14912950e-03 -1.09213352e+00 -5.19451536e-02 1.84810050e-02
-7.66431451e-01 -8.95128965e-01 -8.19896579e-01 -9.61074948e-01
4.89387542e-01 7.24358678e-01 7.84386754e-01 4.51478481e-01
-5.46672583e-01 1.61653459e-02 -5.59693217e-01 -5.24535477e-01
-3.39540660e-01 -5.67945123e-01 -3.83469194e-01 2.77188510e-01
2.19131261e-01 -7.88224697e-01 -1.18580282e+00 7.70947039e-02
-1.48509431e+00 -1.14783816e-01 9.52783585e-01 6.16938770e-01
9.13883820e-02 9.01825786e-01 -5.94714694e-02 -5.13702393e-01
3.54643524e-01 -5.05548358e-01 -8.57073605e-01 -7.04074726e-02
-6.97345614e-01 -2.12827981e-01 4.46783006e-01 5.71189784e-02
-1.26203358e+00 -2.54864156e-01 6.22571297e-02 -5.07695854e-01
-5.37966311e-01 3.21345896e-01 1.07828736e-01 -5.75318098e-01
5.38008153e-01 6.71816111e-01 -1.79762796e-01 -5.26300430e-01
1.78912669e-01 9.98468041e-01 9.60266650e-01 1.17433488e-01
1.28806365e+00 7.96584785e-01 1.87102243e-01 -9.35524464e-01
-3.35743517e-01 -7.23888278e-01 -5.42033076e-01 -4.47297990e-01
1.36297703e+00 -1.23611569e+00 -5.22650063e-01 9.52618182e-01
-1.16840219e+00 -2.25445721e-02 5.77267766e-01 4.03024346e-01
1.16470614e-02 5.41282535e-01 -5.90788066e-01 -1.10361695e+00
-5.27046680e-01 -1.06005704e+00 7.97172368e-01 3.76681209e-01
8.16486299e-01 -9.53630328e-01 -2.11603552e-01 4.52471405e-01
9.85278070e-01 6.82298124e-01 7.19663680e-01 7.16573298e-02
-1.02219558e+00 -1.50551409e-01 -4.89433348e-01 8.92450690e-01
2.53523380e-01 -3.93561572e-02 -1.08781588e+00 -2.12772906e-01
1.51960254e-01 7.87482336e-02 8.57071519e-01 3.47034156e-01
7.11792886e-01 -2.59818524e-01 4.57875341e-01 8.66194308e-01
2.36106038e+00 4.10890758e-01 1.19273520e+00 7.91885436e-01
6.51228368e-01 2.76594162e-01 5.03654659e-01 3.75942379e-01
1.69888604e-02 2.81755447e-01 1.01973808e+00 -3.32960933e-01
-5.75294316e-01 3.37582767e-01 3.73607904e-01 6.60166204e-01
-2.64205575e-01 -5.85781217e-01 -5.74131250e-01 6.66587889e-01
-1.30879223e+00 -8.64959002e-01 -3.43187839e-01 1.86312890e+00
5.16293347e-01 4.83919308e-02 -2.97898591e-01 4.75969702e-01
6.57083213e-01 3.92919369e-02 1.15889020e-01 -4.92008835e-01
-4.88648385e-01 3.97490084e-01 1.05751765e+00 6.49263442e-01
-1.04111600e+00 8.29833686e-01 5.40839767e+00 7.05949128e-01
-1.16647983e+00 2.96585619e-01 4.26240861e-01 6.65058568e-02
1.02523886e-01 -1.47395164e-01 -2.46783540e-01 6.91083729e-01
1.01936841e+00 2.12668329e-01 7.70161211e-01 2.64856249e-01
3.20412099e-01 -6.21643066e-01 -2.70007193e-01 1.22447348e+00
3.84736210e-01 -1.23133492e+00 2.25523557e-03 -5.26219383e-02
1.14187801e+00 1.32786512e-01 3.48581582e-01 -2.83898354e-01
1.86873242e-01 -1.16855645e+00 7.17256069e-01 6.24893904e-01
6.31296992e-01 -8.28330040e-01 1.25490820e+00 9.80597883e-02
-9.50240433e-01 -1.85396343e-01 -5.92080891e-01 3.89131457e-02
4.11034236e-03 4.17439967e-01 -6.01719141e-01 6.33973718e-01
1.00348651e+00 4.27248389e-01 -7.59711504e-01 1.38971472e+00
-5.75204007e-02 5.74013829e-01 -1.41715378e-01 3.47213954e-01
6.24498904e-01 -1.74404785e-01 3.20144445e-01 1.21798766e+00
4.93852615e-01 3.58392090e-01 -4.80599403e-01 7.17287540e-01
2.20707253e-01 -2.11738423e-01 -5.87027073e-01 2.99771428e-01
1.29855081e-01 1.12444186e+00 -6.10320151e-01 -4.75649536e-01
-5.49913287e-01 1.22190940e+00 -6.36869133e-01 4.86081183e-01
-4.93158400e-01 -8.78382325e-01 3.83178324e-01 -1.69020042e-01
6.09687924e-01 -2.94041216e-01 -4.37833071e-02 -7.51156509e-01
-4.37425643e-01 -8.84651124e-01 9.00275111e-02 -1.32319117e+00
-8.22334111e-01 6.20024681e-01 4.52241935e-02 -1.25899446e+00
4.67896551e-01 -8.22836816e-01 -6.54549778e-01 1.15375865e+00
-2.20014858e+00 -1.11559963e+00 -1.02695191e+00 9.57388818e-01
7.84421444e-01 -1.94546267e-01 2.36683458e-01 8.10457110e-01
-3.98829907e-01 2.73142487e-01 4.85741198e-01 3.70399933e-03
6.54757023e-01 -1.22202075e+00 -5.22529706e-02 1.37301254e+00
-5.65683722e-01 3.20406169e-01 1.10413110e+00 -4.93040890e-01
-1.43274462e+00 -1.15584791e+00 3.84770662e-01 -1.00670524e-01
2.59636164e-01 1.60116404e-01 -8.77902567e-01 3.31901520e-01
9.30342019e-01 1.60333246e-01 2.51310945e-01 -1.15432167e+00
-6.42262846e-02 -2.63422817e-01 -1.38047028e+00 1.51642531e-01
2.33407859e-02 -2.16691434e-01 -3.81508291e-01 1.07445128e-01
4.64449763e-01 -3.27814728e-01 -7.95132637e-01 9.22204256e-02
3.95375729e-01 -1.56694520e+00 8.92215610e-01 1.26495529e-02
4.05116528e-01 -8.24943841e-01 -4.94769633e-01 -1.20356059e+00
-2.02648610e-01 -6.84456289e-01 1.91280022e-01 7.76223004e-01
1.30123213e-01 -3.14819008e-01 5.19552290e-01 -6.15533479e-02
-4.44560766e-01 -1.41531721e-01 -6.41530097e-01 -4.82595682e-01
-4.15071659e-02 7.77904913e-02 4.64823216e-01 7.56003559e-01
-8.76062870e-01 -1.57428101e-01 -9.76633251e-01 7.64334083e-01
9.44316506e-01 -4.09634233e-01 6.26442432e-01 -9.47460651e-01
5.93673624e-02 -1.22005381e-02 -7.10929513e-01 -2.11077183e-01
-5.93623400e-01 -1.34981439e-01 2.13355884e-01 -1.61153269e+00
-1.10962726e-01 1.44736364e-01 -6.08537257e-01 -3.43633420e-03
-1.93513766e-01 7.75200307e-01 2.90891528e-01 1.71355054e-01
-3.55125576e-01 3.66662532e-01 1.39033270e+00 -1.72070667e-01
6.34899586e-02 -3.50195497e-01 -3.24893713e-01 3.98846924e-01
9.13004756e-01 -5.26685596e-01 -3.63990694e-01 -6.29060566e-01
1.92473218e-01 8.82032700e-03 5.18300474e-01 -1.58187902e+00
4.94904608e-01 -1.70996949e-01 6.55331314e-01 -7.41616428e-01
5.35809994e-01 -1.02262211e+00 3.57908458e-01 4.91412997e-01
4.09899130e-02 3.90685916e-01 1.51691481e-01 5.58782279e-01
-6.55721128e-01 -3.11982334e-01 1.14706862e+00 -4.96157587e-01
-1.03246117e+00 4.47134115e-02 -6.46191597e-01 -5.02424121e-01
8.92212808e-01 -7.13188052e-01 -5.46074629e-01 -5.25307238e-01
-1.34116933e-01 -3.55214715e-01 3.81429583e-01 8.32994133e-02
1.16904509e+00 -9.58069503e-01 -7.59679377e-01 2.58131653e-01
-2.33426001e-02 -1.03395320e-01 3.20559710e-01 9.14959550e-01
-1.50160706e+00 3.55738513e-02 -8.13458085e-01 -4.58718166e-02
-1.40541697e+00 6.16511941e-01 4.34791297e-01 4.29511309e-01
-4.75848138e-01 8.67668867e-01 -1.78394262e-02 3.83072525e-01
1.28087655e-01 -1.67566612e-01 -3.66812259e-01 -4.92672801e-01
6.24312758e-01 6.84040487e-01 1.00879781e-01 -9.25209105e-01
-8.95510316e-02 6.82968140e-01 1.28602251e-01 -1.81093261e-01
1.42509604e+00 -5.33252239e-01 -5.27041793e-01 9.53877121e-02
1.37011075e+00 -2.30238706e-01 -1.01884389e+00 -1.08018555e-01
-3.41969967e-01 -9.35385168e-01 6.92683816e-01 -8.78828585e-01
-1.35094011e+00 1.17993319e+00 1.46321607e+00 5.52502163e-02
1.70072949e+00 -7.48265147e-01 1.01644373e+00 2.07836598e-01
-9.23914760e-02 -1.00979042e+00 2.72040218e-01 1.29575580e-01
4.82793659e-01 -1.24363267e+00 2.78102487e-01 -1.27254590e-01
-4.56710488e-01 1.35394001e+00 6.54514968e-01 -3.65225881e-01
8.18984509e-01 -2.60194428e-02 3.84222716e-01 -4.30167854e-01
-4.57934588e-01 -2.52628505e-01 1.09742567e-01 6.97423577e-01
1.42411277e-01 -4.66315478e-01 1.14451937e-01 -4.27202404e-01
6.14084071e-03 -8.21577571e-03 1.15463817e+00 1.11603713e+00
-1.10990262e+00 -4.15836394e-01 -1.02371752e+00 3.41372602e-02
-8.86199653e-01 -2.89615571e-01 -8.80894512e-02 8.22444260e-01
6.60163939e-01 1.39355326e+00 -1.88778132e-01 -4.40195560e-01
4.43774574e-02 -5.50299644e-01 4.80976880e-01 -1.75178632e-01
-5.98695219e-01 1.24044329e-01 -2.61942804e-01 -2.73652434e-01
-8.30828249e-01 -8.05563480e-02 -7.25084662e-01 -4.94293213e-01
-3.13543886e-01 1.83404908e-02 1.04489172e+00 7.39265561e-01
-1.25640571e-01 6.42107546e-01 6.60977364e-01 -8.35456491e-01
7.47909248e-02 -1.12385261e+00 -9.05668616e-01 4.94774818e-01
1.10433960e+00 -5.79195976e-01 -5.29383659e-01 4.09568518e-01]
|
[10.893089294433594, -3.1173858642578125]
|
00272b50-eaa1-47f5-b8a5-9292c9020e63
|
cuda-gr-controllable-unsupervised-domain
|
2106.10852
| null |
https://arxiv.org/abs/2106.10852v4
|
https://arxiv.org/pdf/2106.10852v4.pdf
|
CUDA-GHR: Controllable Unsupervised Domain Adaptation for Gaze and Head Redirection
|
The robustness of gaze and head pose estimation models is highly dependent on the amount of labeled data. Recently, generative modeling has shown excellent results in generating photo-realistic images, which can alleviate the need for annotations. However, adopting such generative models to new domains while maintaining their ability to provide fine-grained control over different image attributes, \eg, gaze and head pose directions, has been a challenging problem. This paper proposes CUDA-GHR, an unsupervised domain adaptation framework that enables fine-grained control over gaze and head pose directions while preserving the appearance-related factors of the person. Our framework simultaneously learns to adapt to new domains and disentangle visual attributes such as appearance, gaze direction, and head orientation by utilizing a label-rich source domain and an unlabeled target domain. Extensive experiments on the benchmarking datasets show that the proposed method can outperform state-of-the-art techniques on both quantitative and qualitative evaluations. Furthermore, we demonstrate the effectiveness of generated image-label pairs in the target domain for pretraining networks for the downstream task of gaze and head pose estimation. The source code and pre-trained models are available at https://github.com/jswati31/cuda-ghr.
|
['Xin Eric Wang', 'Swati Jindal']
|
2021-06-21
| null | null | null | null |
['gaze-redirection', 'head-pose-estimation']
|
['computer-vision', 'computer-vision']
|
[-1.24408729e-01 9.21242088e-02 2.31121909e-02 -8.20728719e-01
-7.62219608e-01 -3.71913016e-01 4.01866794e-01 -6.13077402e-01
-2.31642932e-01 6.92683697e-01 2.61981398e-01 2.89325058e-01
3.63487065e-01 -2.77037919e-01 -7.47851551e-01 -8.68297517e-01
3.81190419e-01 5.07756174e-01 -4.97268736e-02 -1.79081652e-02
4.98471744e-02 2.88575977e-01 -1.85839152e+00 -8.25443193e-02
8.80791545e-01 8.31541777e-01 7.90738761e-02 4.92023975e-01
4.02846009e-01 3.87811422e-01 -5.20395160e-01 -4.99814123e-01
1.06539585e-01 -3.90759498e-01 -4.42871034e-01 5.12882650e-01
9.97045934e-01 -4.86379385e-01 -2.77329147e-01 9.02783453e-01
8.71982157e-01 2.17082039e-01 8.65935683e-01 -1.50201190e+00
-9.39285755e-01 -8.07646215e-02 -7.99987316e-01 4.50960249e-02
3.75580400e-01 5.23770630e-01 5.73970973e-01 -9.32836056e-01
6.48745716e-01 1.33262491e+00 2.52216965e-01 1.18019152e+00
-1.23759222e+00 -1.14731157e+00 3.93459976e-01 3.17373604e-01
-1.56839824e+00 -9.51810181e-01 8.39106977e-01 -4.86019582e-01
3.81593972e-01 4.78186160e-02 3.33036661e-01 1.52954304e+00
-1.57738611e-01 8.10781062e-01 1.32498872e+00 -4.83920604e-01
6.72258809e-02 2.44615555e-01 -7.41329491e-02 6.67409122e-01
1.21381968e-01 1.38851553e-01 -7.02543557e-01 1.82936087e-01
6.47703588e-01 -1.05602466e-01 -4.08488572e-01 -6.59781814e-01
-9.09901798e-01 6.86066031e-01 4.91508901e-01 -2.08909929e-01
-1.06319122e-01 -8.53480995e-02 2.99760886e-02 -2.32746154e-01
7.53862679e-01 1.09716900e-01 -3.10115695e-01 1.24130778e-01
-7.87690759e-01 3.76999676e-01 3.59213293e-01 1.31077671e+00
8.32341611e-01 1.04410194e-01 -4.77772862e-01 9.80682850e-01
6.37940109e-01 9.90921259e-01 3.96896601e-01 -1.08509517e+00
3.69354427e-01 5.43262064e-01 1.00955464e-01 -6.90743446e-01
-3.69585425e-01 -1.98102176e-01 -6.62117481e-01 3.26982766e-01
5.10301232e-01 -4.08930242e-01 -1.35415745e+00 2.22123218e+00
7.33135402e-01 -1.01678982e-01 -2.14809269e-01 1.04601467e+00
9.15571451e-01 2.85105944e-01 2.92087525e-01 -7.04284236e-02
1.35315406e+00 -9.77647066e-01 -7.35338509e-01 -6.46030247e-01
1.48578614e-01 -6.57495141e-01 1.20152473e+00 5.75409979e-02
-1.09026027e+00 -6.07529163e-01 -7.47740746e-01 -2.85194904e-01
-6.26722127e-02 4.23596650e-01 2.53678769e-01 7.31929660e-01
-1.27169693e+00 -2.56294340e-01 -7.06661582e-01 -4.95194048e-01
5.43816328e-01 3.96757752e-01 -4.20473427e-01 -2.63720840e-01
-9.26187217e-01 7.26558506e-01 4.24854420e-02 1.08922064e-01
-9.90666211e-01 -4.52210873e-01 -1.05149376e+00 -6.36969060e-02
3.52058530e-01 -8.38979423e-01 1.48424959e+00 -8.98082793e-01
-1.57188702e+00 1.10569179e+00 -4.93397176e-01 7.60386512e-02
5.26468575e-01 -1.41010195e-01 -3.27851981e-01 6.75391927e-02
1.93734303e-01 1.19605482e+00 1.17494929e+00 -1.37538779e+00
-5.18813789e-01 -7.53377378e-01 -1.58663899e-01 5.12108624e-01
-3.66446793e-01 4.43396233e-02 -8.47156942e-01 -4.77181703e-01
-3.08610380e-01 -1.20402753e+00 1.76455781e-01 8.02034512e-02
-4.75197554e-01 -4.03087795e-01 7.51399100e-01 -5.95634580e-01
1.03194082e+00 -2.14980364e+00 1.05385676e-01 -1.31809160e-01
2.97986180e-01 1.82335675e-01 -5.29110581e-02 -1.31116509e-01
-1.59479156e-01 -1.97175309e-01 6.90165386e-02 -8.48708034e-01
1.16608225e-01 -1.18481573e-02 4.49688025e-02 6.12371325e-01
1.17127605e-01 8.63398612e-01 -5.87901115e-01 -6.35255694e-01
1.61406755e-01 6.66960835e-01 -5.23391128e-01 4.99736369e-01
-2.15400785e-01 7.62297571e-01 -3.74009430e-01 5.02059102e-01
7.78996408e-01 -4.65155214e-01 1.10180110e-01 -2.60957748e-01
2.98919111e-01 -6.22159317e-02 -9.01193917e-01 1.51477385e+00
-3.15674871e-01 6.47225201e-01 -3.38560808e-03 -2.60693997e-01
8.39998722e-01 2.19539970e-01 7.39719123e-02 -9.24190342e-01
4.46438372e-01 -1.96425453e-01 -1.69167325e-01 -5.93469799e-01
3.14495772e-01 6.92219511e-02 -1.23323426e-01 5.07767141e-01
2.82619953e-01 9.23345014e-02 1.56779394e-01 8.53913501e-02
3.63646597e-01 3.46793503e-01 1.57655433e-01 -1.04668736e-03
4.12918240e-01 -4.11072940e-01 5.63040137e-01 2.25318074e-01
-4.37205791e-01 1.01144290e+00 6.46912083e-02 -7.50750452e-02
-1.05362785e+00 -9.36772943e-01 -8.14850926e-02 1.59650338e+00
6.97352365e-02 -4.75220941e-02 -1.17942381e+00 -6.92015052e-01
-1.66551322e-01 8.31236362e-01 -9.81712282e-01 -1.43329501e-01
-3.78655374e-01 -6.00514531e-01 2.94420093e-01 5.91934085e-01
5.25798023e-01 -1.02126563e+00 -3.10190409e-01 -4.58989918e-01
-4.35798258e-01 -1.13283861e+00 -8.97960722e-01 -4.42553669e-01
-4.45764482e-01 -9.79677498e-01 -1.04655766e+00 -6.87217355e-01
1.08071947e+00 2.51743168e-01 1.01930380e+00 -2.23070353e-01
-1.21428156e-02 5.64760864e-01 -1.92259699e-01 -4.98669386e-01
-2.41517022e-01 1.58303097e-01 6.93415701e-02 2.88192987e-01
5.64862311e-01 -4.30400997e-01 -9.61208999e-01 6.14765167e-01
-5.99489510e-01 3.05057108e-01 4.15686756e-01 6.56897187e-01
5.40675938e-01 -7.22607613e-01 4.48975086e-01 -1.01087272e+00
5.41853249e-01 -2.51633644e-01 -5.33482373e-01 3.21770847e-01
-7.54676580e-01 2.17587184e-02 3.79922986e-03 -4.86078590e-01
-1.54768145e+00 2.48400941e-01 5.25023378e-02 -6.16262019e-01
-5.54946303e-01 -2.28883997e-01 -6.20950222e-01 6.02585264e-02
8.57488990e-01 1.18019134e-01 9.30377170e-02 -4.65543032e-01
4.78734702e-01 7.92885065e-01 6.44207656e-01 -5.07330179e-01
7.48864770e-01 5.29000700e-01 -1.62388653e-01 -4.32767898e-01
-1.08264911e+00 -4.30661172e-01 -7.83930719e-01 -3.08783203e-01
1.06332219e+00 -1.21784735e+00 -5.43453991e-01 8.20715606e-01
-9.43766296e-01 -3.68558198e-01 7.33108073e-02 1.09764963e-01
-6.67360485e-01 1.52323693e-01 -1.17483333e-01 -6.76031709e-01
-4.39244449e-01 -1.11314368e+00 1.41967010e+00 6.05981410e-01
-8.35007578e-02 -9.45493400e-01 3.29244509e-02 7.24991798e-01
2.00813472e-01 1.39293119e-01 4.26149815e-01 -4.83611822e-01
-6.19153082e-01 -1.78238675e-01 -2.48137906e-01 2.57990807e-01
8.35257322e-02 -1.33582294e-01 -1.36506820e+00 -4.79041994e-01
-2.17711493e-01 -5.41506052e-01 4.88603950e-01 5.66659808e-01
9.77426946e-01 -2.93360204e-01 -4.62563872e-01 7.60311961e-01
9.00286496e-01 -7.74674639e-02 5.79889357e-01 2.35199794e-01
1.01512492e+00 6.93233669e-01 7.27292418e-01 3.88906896e-01
9.24083769e-01 9.33982611e-01 2.78162926e-01 -2.04980716e-01
-5.19972384e-01 -3.63204062e-01 2.61350244e-01 2.70837575e-01
-9.29399505e-02 -4.16703761e-01 -8.36193800e-01 5.11365235e-01
-1.68253136e+00 -7.51001775e-01 1.49518266e-01 2.07227969e+00
9.93767381e-01 -2.45849907e-01 4.26543653e-01 -4.07923251e-01
8.52396905e-01 -2.83178985e-02 -9.01564598e-01 1.74517967e-02
1.40640616e-01 -1.81277156e-01 3.27346146e-01 2.47748211e-01
-1.04212201e+00 9.41825151e-01 5.78445101e+00 5.29490709e-01
-1.19821942e+00 2.31707230e-01 7.48558402e-01 -2.56025374e-01
-5.78276180e-02 -5.11605620e-01 -1.24700975e+00 5.23406029e-01
7.68503368e-01 1.04255241e-03 3.73666763e-01 8.83239865e-01
2.59057462e-01 -3.96835022e-02 -1.18000090e+00 1.09608197e+00
6.80078924e-01 -7.69805193e-01 -1.19833373e-01 1.73324704e-01
8.08161557e-01 5.93101941e-02 5.28212905e-01 3.02379012e-01
3.28222424e-01 -9.58721459e-01 8.12382638e-01 5.66657543e-01
1.33515573e+00 -4.60863322e-01 3.99767101e-01 2.16831326e-01
-7.87830770e-01 -2.66875178e-02 -1.04548791e-02 3.74349236e-01
1.36398256e-01 2.01564166e-03 -1.10952246e+00 9.24452245e-02
9.71058488e-01 5.33784509e-01 -1.00746250e+00 9.29320097e-01
-5.77927232e-01 3.96738738e-01 -1.21998027e-01 3.21226686e-01
-2.85559595e-01 1.44293353e-01 4.01819140e-01 8.77571404e-01
1.40854701e-01 -2.95802532e-03 -1.60252020e-01 8.47711205e-01
-1.77913725e-01 -1.17158696e-01 -3.79871458e-01 3.74100089e-01
6.01592958e-01 1.22337282e+00 -1.61440775e-01 -1.36988565e-01
-3.99493694e-01 7.82739878e-01 4.40936536e-01 6.52732849e-01
-7.87305415e-01 7.74517953e-02 7.79792786e-01 3.97837251e-01
3.40169698e-01 1.50411285e-03 -1.82839200e-01 -1.27963257e+00
-4.63634431e-02 -1.01775563e+00 3.75508279e-01 -1.21133280e+00
-1.26656544e+00 7.48572350e-01 2.71725714e-01 -1.22784710e+00
-7.93792009e-01 -5.97781539e-01 -3.65949184e-01 1.10930705e+00
-1.49951637e+00 -1.65508604e+00 -7.26986229e-01 9.76401269e-01
6.81559324e-01 -2.17680529e-01 7.51502991e-01 2.50807375e-01
-7.88528025e-01 1.14483619e+00 -2.70779617e-02 1.17769383e-01
1.30599988e+00 -1.16124415e+00 3.32109541e-01 7.49127269e-01
-1.19387172e-01 5.05581737e-01 7.26556540e-01 -3.50700676e-01
-9.42054212e-01 -1.19506705e+00 5.70128679e-01 -8.04004490e-01
4.55163382e-02 -6.59384310e-01 -8.33536029e-01 8.88011515e-01
3.67974430e-01 1.14246704e-01 7.35112727e-01 2.73394138e-01
-4.07296747e-01 -3.01266789e-01 -1.17663836e+00 6.22802973e-01
1.05089438e+00 -3.11124831e-01 -3.82338881e-01 2.16613412e-01
4.43899989e-01 -7.51486301e-01 -4.90420282e-01 2.84096688e-01
6.14166081e-01 -1.01552927e+00 8.60505521e-01 -3.66453201e-01
1.21316515e-01 -2.50932485e-01 1.30725682e-01 -1.34619725e+00
-3.58316690e-01 -4.41487074e-01 -1.77537680e-01 1.50004816e+00
3.10010105e-01 -4.81307566e-01 6.39903128e-01 1.06140780e+00
6.60580993e-02 -5.10140240e-01 -5.76758564e-01 -3.18091810e-01
1.00358054e-02 -5.30916750e-02 7.60310829e-01 6.03000343e-01
-4.18848664e-01 7.06169724e-01 -5.08569598e-01 2.55842984e-01
7.98758507e-01 3.61214355e-02 1.22317040e+00 -1.18122768e+00
-1.07454890e-02 -7.14769214e-02 -3.09400648e-01 -9.69162822e-01
3.88708472e-01 -4.42675203e-01 1.53587192e-01 -1.32989836e+00
4.85079557e-01 -2.50672072e-01 -9.59662050e-02 6.58070147e-01
-3.89700741e-01 5.35642266e-01 1.64644733e-01 2.91949421e-01
-7.56731391e-01 6.39054298e-01 1.60182488e+00 2.09477823e-02
-8.49272609e-02 1.37320489e-01 -8.84106815e-01 6.88640177e-01
6.89735174e-01 -2.86498427e-01 -6.74211323e-01 -7.19650149e-01
-1.33582950e-01 -1.52643785e-01 3.69203568e-01 -8.33923221e-01
2.39241585e-01 -2.23345891e-01 6.94271564e-01 -4.25105363e-01
6.01202607e-01 -6.23476326e-01 -4.62185405e-02 -2.39296302e-01
-2.44678646e-01 -8.79098028e-02 3.44727337e-01 6.54578328e-01
1.72929373e-02 9.17280614e-02 9.85218704e-01 1.41712949e-01
-7.80520558e-01 5.61528623e-01 1.85870856e-01 3.03647906e-01
1.06460941e+00 -2.97728598e-01 -3.92066777e-01 -6.43894792e-01
-8.45295846e-01 3.66165131e-01 8.63491714e-01 7.41139174e-01
4.64494318e-01 -1.19155741e+00 -6.56776428e-01 3.67123991e-01
4.99348909e-01 2.01671928e-01 5.47684789e-01 6.39524937e-01
-2.51810085e-02 3.32568675e-01 -4.13339406e-01 -8.45273435e-01
-1.59459460e+00 5.60728669e-01 4.50526327e-01 1.57549515e-01
-2.49437243e-01 9.35597718e-01 9.84198272e-01 -3.28388393e-01
2.86463708e-01 3.97121944e-02 -3.91073614e-01 -8.48811939e-02
5.52796543e-01 2.23957762e-01 -9.34088752e-02 -1.05324852e+00
-2.68943965e-01 5.98790765e-01 -3.22566003e-01 6.92839269e-03
1.20637119e+00 -9.27799582e-01 3.59768331e-01 1.18347980e-01
1.00378048e+00 -1.47591606e-01 -1.76817679e+00 -2.96890736e-01
-5.65081954e-01 -6.01019740e-01 -6.74456209e-02 -1.07276475e+00
-1.19464600e+00 1.03110933e+00 9.89163220e-01 -4.52855140e-01
1.33841693e+00 2.78741598e-01 4.56564337e-01 -6.15770444e-02
1.95065185e-01 -9.40621972e-01 2.36748651e-01 2.37177312e-01
8.93420577e-01 -1.61132014e+00 -1.76818326e-01 -4.03379828e-01
-1.10815609e+00 6.01603091e-01 1.20106459e+00 2.38671869e-01
4.77738142e-01 -1.08463012e-01 4.34374779e-01 -1.51232347e-01
-5.70130467e-01 -2.61081904e-01 7.58276820e-01 1.03312826e+00
4.56774831e-01 1.00543618e-01 2.56448030e-01 3.98986220e-01
-2.71210879e-01 5.05528925e-03 2.45869324e-01 4.44697559e-01
-1.32462248e-01 -1.02249312e+00 -5.18365383e-01 1.49639204e-01
-2.94737846e-01 1.38628587e-01 -3.17152590e-01 8.05135429e-01
1.16609968e-01 8.06691349e-01 1.24524802e-01 -4.02978957e-02
3.63497883e-01 1.32379830e-01 6.89527631e-01 -8.83591413e-01
5.95268980e-02 2.18061969e-01 -2.56731436e-02 -3.32824469e-01
-5.36027789e-01 -8.01980257e-01 -8.36028814e-01 -2.27386057e-01
-2.31207073e-01 -2.26639152e-01 4.00529861e-01 8.74512553e-01
6.31148875e-01 3.43672425e-01 5.70686996e-01 -1.10848129e+00
-4.84693617e-01 -1.27767634e+00 -5.42491555e-01 6.82280838e-01
4.27230358e-01 -1.13969183e+00 -1.87894717e-01 5.36290586e-01]
|
[14.112809181213379, 0.0375446155667305]
|
bb3fddbe-d93a-4e6e-b884-ca31de3f0b72
|
revisiting-consistency-regularization-for-1
|
2204.08454
| null |
https://arxiv.org/abs/2204.08454v3
|
https://arxiv.org/pdf/2204.08454v3.pdf
|
Revisiting Consistency Regularization for Semi-supervised Change Detection in Remote Sensing Images
|
Remote-sensing (RS) Change Detection (CD) aims to detect "changes of interest" from co-registered bi-temporal images. The performance of existing deep supervised CD methods is attributed to the large amounts of annotated data used to train the networks. However, annotating large amounts of remote sensing images is labor-intensive and expensive, particularly with bi-temporal images, as it requires pixel-wise comparisons by a human expert. On the other hand, we often have access to unlimited unlabeled multi-temporal RS imagery thanks to ever-increasing earth observation programs. In this paper, we propose a simple yet effective way to leverage the information from unlabeled bi-temporal images to improve the performance of CD approaches. More specifically, we propose a semi-supervised CD model in which we formulate an unsupervised CD loss in addition to the supervised Cross-Entropy (CE) loss by constraining the output change probability map of a given unlabeled bi-temporal image pair to be consistent under the small random perturbations applied on the deep feature difference map that is obtained by subtracting their latent feature representations. Experiments conducted on two publicly available CD datasets show that the proposed semi-supervised CD method can reach closer to the performance of supervised CD even with access to as little as 10% of the annotated training data. Code available at https://github.com/wgcban/SemiCD
|
['Vishal M. Patel', 'Wele Gedara Chaminda Bandara']
|
2022-04-18
| null | null | null | null |
['semi-supervised-change-detection']
|
['computer-vision']
|
[ 3.82899314e-01 -2.58796930e-01 1.31869152e-01 -5.16295910e-01
-8.60701323e-01 -6.27969384e-01 6.64059877e-01 2.29664817e-02
-5.02893686e-01 6.79763317e-01 -1.37373675e-02 -1.35412171e-01
-5.94769679e-02 -8.51944923e-01 -6.21529460e-01 -8.96071672e-01
-1.35095194e-01 -1.61418337e-02 3.64471488e-02 -7.37168863e-02
-1.08839534e-01 6.09601200e-01 -1.47401953e+00 -1.54910475e-01
1.06216311e+00 1.12827110e+00 3.69674385e-01 3.13079298e-01
3.20509404e-01 4.30206418e-01 2.31156394e-01 -1.94980651e-02
7.29268730e-01 -4.31521297e-01 -6.79041445e-01 1.32168025e-01
3.23753983e-01 -4.54833925e-01 -4.09079790e-01 1.39401472e+00
5.86157858e-01 3.56652200e-01 3.82568836e-01 -9.61167872e-01
-4.81428772e-01 3.30745101e-01 -6.57035768e-01 5.05459249e-01
-2.01757297e-01 3.82895529e-01 1.03887367e+00 -8.64870012e-01
6.38023078e-01 7.79412270e-01 6.64952695e-01 2.04562247e-01
-1.42934203e+00 -4.68951106e-01 1.96838945e-01 1.31802991e-01
-1.58720958e+00 -4.94974524e-01 7.86058307e-01 -6.89179778e-01
6.72173679e-01 2.13350341e-01 4.90467072e-01 5.36830842e-01
-1.33025885e-01 5.07231057e-01 1.19247222e+00 -2.13053659e-01
2.56451756e-01 3.65790562e-03 -9.78049785e-02 4.56784487e-01
-1.27602192e-02 2.49737024e-01 -1.55375183e-01 3.93687189e-02
6.31727219e-01 1.64808944e-01 -4.97069538e-01 -3.92649382e-01
-1.16758513e+00 7.35746503e-01 8.12872708e-01 4.25538093e-01
-5.53807259e-01 -8.17508772e-02 1.98336929e-01 2.70864606e-01
1.03443599e+00 2.09650531e-01 -4.40071642e-01 4.69473638e-02
-1.08143961e+00 -1.79296825e-02 2.25612700e-01 4.86830056e-01
1.21289897e+00 -1.05675958e-01 1.41411498e-01 6.67271793e-01
2.44294256e-01 5.77419102e-01 5.10521948e-01 -7.35673547e-01
4.91063744e-01 4.83180642e-01 2.34392956e-01 -1.23894370e+00
-2.95632392e-01 -5.21946788e-01 -1.17892385e+00 8.35793167e-02
1.53149605e-01 -1.49198622e-01 -7.96318650e-01 1.75080001e+00
4.00642157e-01 6.63114488e-02 -2.06498638e-01 9.23218787e-01
3.26543361e-01 7.59971619e-01 -1.12501025e-01 -2.32191995e-01
8.05850983e-01 -6.26555681e-01 -5.24447441e-01 -2.52801657e-01
4.35255766e-01 -2.28175536e-01 1.06168592e+00 -1.34383008e-01
-5.78192651e-01 -4.16955322e-01 -9.78103459e-01 1.24839909e-01
-4.76917475e-01 4.04974222e-01 1.56812102e-01 1.88260540e-01
-9.17414963e-01 7.03690290e-01 -1.25316215e+00 -3.18003803e-01
5.25634110e-01 1.36028200e-01 -4.26767290e-01 -1.00089744e-01
-1.23221409e+00 6.65601134e-01 3.58559608e-01 5.40213883e-01
-8.56289625e-01 -7.22569585e-01 -8.85937274e-01 -3.51247266e-02
3.01763952e-01 -3.49202715e-02 9.76249933e-01 -1.27459657e+00
-1.22678077e+00 8.51328671e-01 -7.70799965e-02 -3.24811816e-01
8.93144429e-01 -1.97240278e-01 -3.14768642e-01 2.38296732e-01
3.43149424e-01 5.89870930e-01 5.77795863e-01 -1.18937933e+00
-4.89078999e-01 -4.92731303e-01 -7.48875290e-02 2.00993806e-01
-3.04510564e-01 -2.11756974e-01 -2.06828400e-01 -6.29897177e-01
4.46066767e-01 -1.03868639e+00 -3.86734456e-01 3.28367382e-01
-3.04895729e-01 2.49159947e-01 9.75999951e-01 -9.47047710e-01
1.00689042e+00 -2.30956101e+00 -9.46327895e-02 1.57351881e-01
1.09241754e-01 2.99724936e-01 -1.91488400e-01 3.03913921e-01
-1.30480185e-01 9.38882828e-02 -8.71786416e-01 -1.52825505e-01
-2.19043911e-01 1.18057076e-02 -3.16939861e-01 8.42980325e-01
3.47198397e-01 7.15899706e-01 -1.02206743e+00 -1.71313182e-01
3.53304774e-01 4.18656677e-01 -2.15628117e-01 7.94245526e-02
-1.52393222e-01 9.34355259e-01 -4.26266551e-01 4.42159176e-01
8.67064416e-01 -3.85306299e-01 1.77588776e-01 -1.66415587e-01
-3.24555337e-01 2.31762201e-01 -1.12560499e+00 1.48989618e+00
-3.01055163e-01 7.50884771e-01 -1.43149510e-01 -1.26545620e+00
7.80318141e-01 2.40488261e-01 6.53812349e-01 -7.18585312e-01
-2.30884373e-01 2.81725019e-01 -2.64472157e-01 -3.80726099e-01
2.44539082e-01 -2.43145719e-01 1.93511575e-01 4.10541087e-01
-3.84513706e-01 -9.77627560e-02 -4.37491685e-02 -9.48114395e-02
8.63843977e-01 2.63169646e-01 4.63323593e-01 -1.28439590e-01
4.84606534e-01 5.32456338e-02 8.86026680e-01 4.85958338e-01
-3.81001621e-01 5.13885319e-01 -7.86860567e-03 -6.20673358e-01
-9.74339664e-01 -8.91950369e-01 -3.34028661e-01 6.91280663e-01
1.46537155e-01 2.36776441e-01 -3.81091386e-01 -6.44430280e-01
-8.82092956e-03 4.19075489e-01 -5.32582939e-01 1.81378946e-02
-3.69749546e-01 -9.62017715e-01 4.19406086e-01 3.63116831e-01
1.05685019e+00 -8.30212593e-01 -6.35512948e-01 2.06863135e-01
-4.47553039e-01 -1.00237596e+00 -3.24071169e-01 1.61938563e-01
-9.40487206e-01 -7.24633634e-01 -6.01525247e-01 -4.75348294e-01
8.34508121e-01 4.17378932e-01 7.09995747e-01 -3.87276709e-01
-2.70072788e-01 3.90352383e-02 -3.25310439e-01 6.00599963e-03
-1.50662005e-01 -2.50076000e-02 6.11330010e-03 2.45707765e-01
2.64092743e-01 -7.39945889e-01 -7.21081197e-01 3.31422985e-01
-1.16994631e+00 -1.95016321e-02 4.44248229e-01 7.94432640e-01
8.52624774e-01 3.32133502e-01 5.43918490e-01 -5.80870330e-01
-8.75390843e-02 -4.75883573e-01 -8.12928617e-01 1.90602079e-01
-6.84897065e-01 -2.60058884e-02 4.69723433e-01 -2.87769675e-01
-1.18356681e+00 3.88186902e-01 1.24914847e-01 -3.98537576e-01
-1.39562950e-01 9.35674787e-01 -2.00823024e-01 9.65122040e-03
6.79350376e-01 3.35207403e-01 -1.13835461e-01 -4.46272910e-01
3.34434658e-01 7.63177276e-01 4.91105288e-01 -1.81010932e-01
9.54372585e-01 8.19713473e-01 -2.64851034e-01 -7.42375195e-01
-8.42978001e-01 -5.50199330e-01 -8.56636822e-01 -2.11267218e-01
6.70086682e-01 -1.17536855e+00 -2.21645236e-02 7.81138420e-01
-6.35514975e-01 -5.55739284e-01 -3.73669505e-01 6.81856334e-01
-3.57887805e-01 5.16331553e-01 -2.56232083e-01 -7.31516182e-01
-3.26165885e-01 -8.38347137e-01 8.48451793e-01 -8.76037497e-03
2.87291259e-01 -9.59964573e-01 2.88444877e-01 2.09789932e-01
4.67252225e-01 5.34122169e-01 5.65868318e-01 -3.20572495e-01
-6.36846602e-01 -2.82716841e-01 -3.33417773e-01 7.33128607e-01
6.14651382e-01 -2.68562734e-02 -9.96581614e-01 -4.17814940e-01
4.68633771e-02 -3.34667176e-01 1.07394266e+00 5.03521562e-01
1.07302630e+00 -3.22016478e-01 -1.84030190e-01 6.07852995e-01
1.69428158e+00 1.08638912e-01 5.38568258e-01 3.80600870e-01
6.80333674e-01 6.01828218e-01 6.68898284e-01 5.54654241e-01
3.12494636e-01 5.93015134e-01 4.07965839e-01 -1.49057880e-01
1.61167994e-01 -1.57851428e-01 3.15872669e-01 6.08842671e-01
-9.03617889e-02 -1.02408737e-01 -1.14790118e+00 7.99918592e-01
-1.85316110e+00 -1.16636944e+00 6.49918616e-02 2.43574691e+00
1.06858373e+00 -3.05638343e-01 -2.70593077e-01 1.07184999e-01
9.25647080e-01 4.29125935e-01 -9.70549762e-01 5.24136007e-01
-3.65890443e-01 -1.50030240e-01 6.47860885e-01 4.02365386e-01
-1.48123443e+00 7.51825094e-01 5.05222225e+00 5.05237222e-01
-1.56750488e+00 2.60559261e-01 6.98456705e-01 -4.29229476e-02
-2.56039619e-01 -3.02635022e-02 -4.41039324e-01 4.72631514e-01
7.02968955e-01 -1.62672117e-01 3.39086592e-01 5.97803116e-01
6.92580104e-01 -2.71788508e-01 -9.24741685e-01 8.65177631e-01
-2.34409124e-01 -1.15390754e+00 -1.81616336e-01 1.35937050e-01
1.07229829e+00 6.98399305e-01 4.60518040e-02 -9.15740877e-02
3.19349080e-01 -6.75865889e-01 6.72228098e-01 4.53256756e-01
1.00037420e+00 -3.41512591e-01 6.41430676e-01 4.35298771e-01
-1.31152320e+00 -2.01523248e-02 -3.23560387e-01 -5.24295457e-02
-5.67374639e-02 9.84229445e-01 -5.02858162e-01 6.68276250e-01
8.49443257e-01 1.10319912e+00 -4.37629938e-01 8.74136627e-01
-1.51889920e-01 5.91768622e-01 -4.90652710e-01 4.55800921e-01
2.58143395e-01 -4.74991053e-01 4.62853819e-01 8.73764396e-01
5.36164224e-01 2.46170297e-01 1.94950446e-01 9.86772120e-01
-2.05427751e-01 -4.30698171e-02 -6.85728252e-01 -1.51565462e-01
4.02757734e-01 1.22053874e+00 -5.34855366e-01 -2.01969057e-01
-3.17432910e-01 1.01896346e+00 2.06664816e-01 4.41015124e-01
-6.46796584e-01 -2.38982275e-01 6.00814581e-01 1.82168968e-02
3.84512722e-01 -2.06035718e-01 -1.22203566e-01 -1.41659963e+00
3.55153233e-01 -5.64928174e-01 3.22443724e-01 -7.15833187e-01
-1.35266399e+00 5.25248051e-01 -1.34246230e-01 -1.66251242e+00
-8.59376118e-02 -8.83181393e-02 -4.99967456e-01 9.68898058e-01
-1.94953048e+00 -9.74748611e-01 -6.37123704e-01 5.11398852e-01
2.53848195e-01 2.22126767e-01 5.77789664e-01 4.27847475e-01
-6.74854279e-01 2.63220549e-01 6.25631273e-01 2.60713279e-01
5.11251032e-01 -9.51929748e-01 3.03644508e-01 1.23820007e+00
-1.18428506e-01 1.58237800e-01 5.12842715e-01 -5.06046116e-01
-8.98703516e-01 -1.58459234e+00 8.47346902e-01 1.25083894e-01
6.86640382e-01 6.34123906e-02 -1.09140980e+00 7.96451628e-01
-2.48418510e-01 5.66633105e-01 5.20547211e-01 -4.49670404e-01
-2.92720914e-01 -4.13304299e-01 -1.27778852e+00 2.65480608e-01
8.98526013e-01 -8.48071575e-01 -2.90332556e-01 3.87565702e-01
4.59015071e-01 -1.38087690e-01 -7.52357543e-01 5.18481672e-01
2.57529050e-01 -7.90926456e-01 6.31538033e-01 -2.19794437e-01
4.85669732e-01 -6.33148313e-01 -3.90808374e-01 -1.34454656e+00
-1.61544710e-01 -2.56095797e-01 4.80479836e-01 1.06913435e+00
2.79996246e-01 -9.04427409e-01 4.82699901e-01 6.23429477e-01
7.42342994e-02 -3.54980201e-01 -1.07826889e+00 -9.08644676e-01
4.11628149e-02 -3.03595275e-01 5.04215539e-01 1.35402954e+00
-2.76027977e-01 -1.05847627e-01 -2.22466350e-01 6.93393528e-01
5.48299313e-01 2.36392930e-01 4.94688511e-01 -1.25821042e+00
-1.43764228e-01 -2.22074658e-01 -4.25251365e-01 -9.18788314e-01
1.18773833e-01 -9.14211571e-01 4.07260448e-01 -1.34877253e+00
2.77943492e-01 -7.41720140e-01 -3.58267069e-01 7.47371495e-01
-2.30522260e-01 3.23324054e-01 -9.52720642e-02 6.37601078e-01
-1.20717101e-01 9.10338700e-01 9.30828214e-01 -3.15773159e-01
-3.66186172e-01 -8.08901805e-03 -2.32345879e-01 6.25132918e-01
9.42551017e-01 -5.32543719e-01 -3.12532455e-01 -5.62067807e-01
2.17433840e-01 -9.70518515e-02 6.56151175e-01 -9.65281904e-01
9.35158953e-02 -2.48973578e-01 9.07101408e-02 -5.66044629e-01
-3.01877726e-02 -8.57712746e-01 4.17645961e-01 4.12676752e-01
-3.57007891e-01 -2.20055893e-01 1.04440339e-01 6.93808615e-01
-5.15965343e-01 -3.11532579e-02 9.70067322e-01 -9.63985920e-02
-7.95126498e-01 4.93188411e-01 -2.03627288e-01 -6.16142750e-02
9.70463514e-01 -1.88209452e-02 -2.65494227e-01 -4.19125944e-01
-6.26548469e-01 1.62545934e-01 4.88613904e-01 5.13917468e-02
3.51290554e-01 -1.17604065e+00 -7.92908609e-01 2.44637996e-01
3.50055516e-01 2.39984527e-01 4.33499485e-01 9.76675093e-01
-5.16696990e-01 2.46939152e-01 -1.62047699e-01 -6.67761207e-01
-8.85993898e-01 2.59574592e-01 6.55069053e-01 -2.07343057e-01
-7.44461536e-01 6.56148553e-01 1.10772319e-01 -7.18664050e-01
-1.54701844e-01 -4.24756169e-01 1.26553223e-01 1.58805564e-01
3.51529688e-01 3.45551193e-01 2.33047426e-01 -7.12179005e-01
-4.03510362e-01 5.10746598e-01 1.66219547e-01 -2.63215095e-01
1.66273963e+00 -4.48442400e-01 -1.00674890e-01 5.47781408e-01
1.37495697e+00 -3.04101378e-01 -1.71211123e+00 -7.84619927e-01
-8.04709718e-02 -5.51166296e-01 4.01696444e-01 -6.68458104e-01
-1.41611409e+00 8.98929954e-01 9.50044513e-01 1.49989901e-02
1.37376344e+00 -9.46485922e-02 5.32872915e-01 6.23245299e-01
2.49482408e-01 -1.06910563e+00 -2.05560967e-01 2.72171736e-01
7.70889401e-01 -1.65039945e+00 1.07176989e-01 -1.38941333e-01
-5.74426889e-01 6.90545559e-01 2.47505352e-01 2.97332313e-02
1.01526809e+00 -2.10965395e-01 1.75717816e-01 -1.94883049e-01
-3.94828916e-01 -3.15527558e-01 1.69506148e-01 3.03360254e-01
1.42765298e-01 2.16863111e-01 -1.12809740e-01 8.67318548e-03
2.37634480e-01 1.77624092e-01 4.32883382e-01 9.31723475e-01
-3.13625634e-01 -6.31180942e-01 -1.30217269e-01 4.34748322e-01
-2.83233583e-01 -2.14729965e-01 -1.00757275e-02 5.12581289e-01
4.72441167e-02 9.00196731e-01 1.58419266e-01 -2.55856276e-01
4.25883494e-02 -1.07168317e-01 -1.18148416e-01 -4.63082880e-01
-2.45467529e-01 6.46163449e-02 -2.28054062e-01 -5.60315549e-01
-9.46873605e-01 -1.04025042e+00 -1.10805964e+00 -1.27024546e-01
-2.93493718e-01 -7.62163475e-02 6.47149861e-01 9.30891216e-01
5.25014937e-01 -1.40511282e-02 1.18003368e+00 -9.43758190e-01
-6.95992053e-01 -1.04705763e+00 -7.46586263e-01 3.87458116e-01
5.50774634e-01 -4.44985539e-01 -6.85899377e-01 3.64952117e-01]
|
[9.741862297058105, -1.3024202585220337]
|
653a0181-d0ff-4600-a87c-2aefd1f0882c
|
transfer-learning-for-conflict-and-duplicate
|
2301.03709
| null |
https://arxiv.org/abs/2301.03709v1
|
https://arxiv.org/pdf/2301.03709v1.pdf
|
Transfer learning for conflict and duplicate detection in software requirement pairs
|
Consistent and holistic expression of software requirements is important for the success of software projects. In this study, we aim to enhance the efficiency of the software development processes by automatically identifying conflicting and duplicate software requirement specifications. We formulate the conflict and duplicate detection problem as a requirement pair classification task. We design a novel transformers-based architecture, SR-BERT, which incorporates Sentence-BERT and Bi-encoders for the conflict and duplicate identification task. Furthermore, we apply supervised multi-stage fine-tuning to the pre-trained transformer models. We test the performance of different transfer models using four different datasets. We find that sequentially trained and fine-tuned transformer models perform well across the datasets with SR-BERT achieving the best performance for larger datasets. We also explore the cross-domain performance of conflict detection models and adopt a rule-based filtering approach to validate the model classifications. Our analysis indicates that the sentence pair classification approach and the proposed transformer-based natural language processing strategies can contribute significantly to achieving automation in conflict and duplicate detection
|
['Devang Parikh', 'Ayse Bener', 'Mucahit Cevik', 'Savas Yildirim', 'Garima Malik']
|
2023-01-09
| null | null | null | null |
['sentence-pair-classification']
|
['natural-language-processing']
|
[ 4.30460811e-01 -8.76857117e-02 2.66075075e-01 -1.02374709e+00
-7.70560503e-01 -5.12316704e-01 3.21492463e-01 2.82502085e-01
-9.00079682e-02 1.80983722e-01 2.84441203e-01 -5.00120044e-01
-4.62066442e-01 -5.24174035e-01 -3.73101771e-01 2.93399721e-01
2.06418514e-01 4.51526910e-01 1.02932103e-01 -6.31690502e-01
7.11902380e-01 1.41718030e-01 -1.61359918e+00 9.75869775e-01
9.97383058e-01 8.66698921e-01 1.59246117e-01 4.35814083e-01
-2.41441742e-01 8.86707783e-01 -5.28459132e-01 -5.39044321e-01
3.40719968e-01 -2.57436723e-01 -8.57624710e-01 -1.29562868e-02
1.36387408e-01 7.11855665e-02 3.94157588e-01 8.28167081e-01
3.78867984e-01 -3.06097865e-01 4.04647261e-01 -1.07896090e+00
-6.94631279e-01 9.91978586e-01 -4.90810513e-01 2.98274517e-01
7.18541682e-01 -1.99537441e-01 1.10161567e+00 -9.50355589e-01
3.29824537e-01 8.39566350e-01 7.52110600e-01 2.84508169e-01
-1.13927782e+00 -6.96634293e-01 -2.47321129e-02 1.73246175e-01
-1.45898461e+00 -7.35862553e-01 7.53793716e-01 -7.24254251e-01
1.69855165e+00 1.43756092e-01 6.28320277e-02 5.04109085e-01
3.37917000e-01 2.78302103e-01 7.17304826e-01 -7.81136274e-01
5.30721396e-02 4.33779091e-01 3.41383845e-01 7.28125691e-01
1.49253428e-01 -3.21294397e-01 -6.15758181e-01 7.08436221e-02
2.04398677e-01 -3.27586561e-01 -1.00101434e-01 -1.31095245e-01
-1.06781995e+00 5.74649453e-01 -3.06494713e-01 7.80409873e-01
-3.95957470e-01 -7.83023536e-02 6.82541132e-01 1.13334739e+00
6.14453137e-01 8.97273123e-01 -9.85727012e-01 -2.51810074e-01
-1.11469030e+00 -1.38538688e-01 7.44835854e-01 1.49515176e+00
6.80315137e-01 -4.68286052e-02 -3.87444049e-01 7.41484106e-01
3.86956573e-01 -7.94837028e-02 5.04283905e-01 -5.11549652e-01
6.89047039e-01 1.18853986e+00 7.17816278e-02 -8.23091507e-01
-3.18153054e-01 -5.51441312e-01 -1.74940526e-01 -4.63284813e-02
-8.19394737e-02 -1.10044785e-01 -6.19750440e-01 1.60101879e+00
-2.02996418e-01 -3.48310977e-01 1.00944132e-01 3.33032012e-01
5.16718566e-01 2.72228032e-01 -1.79696023e-01 -2.13624120e-01
1.33474302e+00 -6.31547749e-01 -6.62550271e-01 -3.33940238e-01
1.05441689e+00 -9.16068017e-01 1.13502288e+00 3.86504680e-01
-1.14192390e+00 -6.09726608e-01 -1.27462840e+00 1.06506996e-01
-9.36124325e-02 5.50491571e-01 2.76882589e-01 7.32566595e-01
-1.17303038e+00 4.92534667e-01 -4.79195863e-01 -3.67270261e-01
1.07152686e-01 5.44739842e-01 -4.34186041e-01 -1.03918865e-01
-1.06026578e+00 1.11371851e+00 1.04529589e-01 -3.24206939e-03
-3.54479194e-01 -9.23186719e-01 -1.06184661e+00 3.22587401e-01
9.58170295e-02 -3.95892084e-01 1.39166141e+00 -1.08701921e+00
-1.13994098e+00 8.40709329e-01 -3.08932774e-02 -1.36234909e-01
-4.68397178e-02 9.24200471e-03 -6.07810020e-01 -2.83168077e-01
3.89858782e-01 6.10078573e-02 6.35478556e-01 -7.91048348e-01
-7.51424432e-01 -2.87450701e-01 1.03758238e-01 -2.02082068e-01
-5.88222146e-01 8.72074723e-01 2.08900552e-02 -3.10500741e-01
-1.21245958e-01 -4.40778583e-01 4.36191149e-02 -5.31352222e-01
1.42829582e-01 -2.52264768e-01 3.87820810e-01 -6.44554853e-01
1.58712304e+00 -2.25695539e+00 -9.06922156e-04 2.23122239e-01
6.79628849e-02 7.05054775e-02 -4.79384333e-01 6.77589595e-01
-2.41317987e-01 2.10971460e-01 -1.66412652e-01 -2.43841290e-01
1.98183462e-01 -1.07729100e-01 -9.49646086e-02 1.24311939e-01
8.72303724e-01 8.76032829e-01 -5.03298104e-01 -4.47304815e-01
2.32051462e-02 5.19957803e-02 -6.19078994e-01 2.70715803e-01
-2.26610139e-01 -4.21213023e-02 -3.50996673e-01 4.00988102e-01
4.96226341e-01 -2.95201600e-01 4.60960865e-01 -2.90750086e-01
-3.24292928e-01 7.97987044e-01 -9.17128563e-01 1.68486714e+00
-8.43749106e-01 5.78430235e-01 -6.07208312e-02 -1.02732778e+00
1.34627354e+00 5.50063789e-01 3.03423703e-01 -1.23645699e+00
1.32960439e-01 2.17093006e-01 2.77731329e-01 -8.58559668e-01
5.41833520e-01 -3.75724077e-01 -4.04704481e-01 8.13945055e-01
9.96052548e-02 -1.23209115e-02 3.47538769e-01 4.07100888e-03
1.31296301e+00 5.47181666e-02 3.56088042e-01 -4.54165429e-01
8.44976664e-01 1.98516518e-01 6.38832331e-01 3.46560299e-01
-9.50720459e-02 3.97326589e-01 5.68884552e-01 -3.34122509e-01
-7.99896121e-01 -4.25880134e-01 1.56274080e-01 1.29578555e+00
-2.39827901e-01 -6.97159052e-01 -2.90506959e-01 -7.59294152e-01
-8.05633143e-02 1.07048821e+00 -5.07556677e-01 -4.22041237e-01
-5.55696726e-01 -2.78621078e-01 5.99356651e-01 5.76985419e-01
-3.62215266e-02 -1.09049916e+00 -9.92437840e-01 4.51881707e-01
-2.70990431e-01 -1.24994576e+00 -4.65982139e-01 5.07318795e-01
-3.55911553e-01 -1.00422740e+00 7.40276128e-02 -9.91203964e-01
5.56259573e-01 9.97301489e-02 1.51926780e+00 2.38413662e-01
-1.33368485e-02 3.08022648e-01 -8.83763015e-01 -3.48103851e-01
-7.44851232e-01 7.14619234e-02 -1.39916047e-01 -2.40976155e-01
8.14305425e-01 -4.32530999e-01 7.58567601e-02 4.85939175e-01
-7.16119528e-01 1.69160813e-02 6.42687559e-01 6.13118827e-01
-5.35937883e-02 1.89906418e-01 9.97832716e-01 -7.78521895e-01
1.10852444e+00 -3.67063940e-01 -5.11567414e-01 7.51295745e-01
-1.01764572e+00 4.15454239e-01 6.65481627e-01 -6.05356917e-02
-1.14161444e+00 -3.90311442e-02 -9.58673060e-02 -6.43938035e-02
4.24062945e-02 9.98072624e-01 -7.37852305e-02 -2.50319089e-03
6.23898864e-01 1.61378965e-01 -8.16988945e-02 -3.91782165e-01
-6.05151355e-02 1.06367409e+00 1.07805356e-01 -5.95314801e-01
5.96919835e-01 -3.17941129e-01 -5.95829427e-01 -3.06565404e-01
-5.99204957e-01 -5.83499193e-01 -5.97765088e-01 -1.30759209e-01
5.99853992e-01 -8.30984414e-01 -1.95908412e-01 1.07243024e-02
-1.39309037e+00 -1.61283582e-01 -1.51684612e-01 2.37403423e-01
-3.70924860e-01 3.03850353e-01 -4.32900786e-01 -7.76401281e-01
-5.36993980e-01 -1.19350791e+00 9.55264449e-01 -1.70041114e-01
-5.32772005e-01 -6.11653030e-01 1.81877136e-01 3.88665497e-01
6.61855578e-01 -1.79341689e-01 1.35053217e+00 -8.86025608e-01
-1.07151838e-02 -1.28082216e-01 -1.71155423e-01 4.94539887e-01
4.98753071e-01 2.50460118e-01 -6.67615175e-01 -2.24749371e-01
1.64892837e-01 -3.33809704e-01 2.45103270e-01 -9.80628878e-02
5.81795573e-01 3.27202566e-02 -8.06731696e-04 -3.73607967e-03
1.32184434e+00 2.72792190e-01 4.47824836e-01 2.59967983e-01
2.67235726e-01 1.10881996e+00 9.74882722e-01 6.38484955e-01
5.37640750e-01 5.90549648e-01 1.36050880e-01 2.42801800e-01
2.01873228e-01 2.18159869e-01 4.86369371e-01 1.05330682e+00
4.41944987e-01 -6.95570633e-02 -1.25910068e+00 7.51989722e-01
-1.75057757e+00 -8.16177726e-01 -2.55067825e-01 1.87183416e+00
9.20585871e-01 3.75352859e-01 -1.73985258e-01 4.75604951e-01
6.06228352e-01 -3.63964707e-01 1.18147500e-01 -8.59026492e-01
3.53293508e-01 5.01898646e-01 1.60875905e-03 1.91142499e-01
-6.89382792e-01 7.03753412e-01 6.38639593e+00 1.09560020e-01
-1.03992772e+00 5.33143478e-03 -3.85590009e-02 1.48035750e-01
-7.14815378e-01 -7.83857927e-02 -3.69546026e-01 2.19316289e-01
1.15392399e+00 -4.87962216e-01 1.91948906e-01 6.56449199e-01
3.54318678e-01 2.41186306e-01 -1.72665393e+00 5.58183253e-01
2.31354862e-01 -1.29131401e+00 -8.72353986e-02 -4.41902161e-01
3.79674613e-01 5.20981662e-02 -2.87379831e-01 4.53598022e-01
4.89693470e-02 -8.76718044e-01 8.93154681e-01 2.46353447e-01
7.94982553e-01 -6.10618651e-01 1.00283337e+00 3.96981031e-01
-1.33467710e+00 -2.71910965e-01 -2.14974001e-01 -2.80414045e-01
-6.56415895e-02 5.04803538e-01 -1.11139739e+00 1.00213921e+00
6.25977576e-01 7.82324731e-01 -7.59719789e-01 7.25093961e-01
8.43481496e-02 1.31047264e-01 -4.62829135e-02 -1.79452766e-02
-2.01639265e-01 5.05466089e-02 2.34451424e-02 1.56214893e+00
3.20764840e-01 -2.54015207e-01 -1.07481308e-01 1.08448160e+00
1.11547999e-01 1.57299824e-02 -4.48066890e-01 -3.54751229e-01
5.56227505e-01 1.12578046e+00 -4.13538724e-01 -1.28834143e-01
-7.49817550e-01 6.02154851e-01 2.27963671e-01 -5.21578453e-02
-7.10004807e-01 -5.44677496e-01 5.16697705e-01 6.22692443e-02
6.33422792e-01 -9.68990251e-02 -5.56028187e-01 -1.02274466e+00
3.77104372e-01 -1.22744095e+00 3.00483972e-01 -6.36538625e-01
-1.18896949e+00 8.79661620e-01 -4.45940346e-02 -1.39050651e+00
-2.94707596e-01 -4.47866946e-01 -7.53084600e-01 8.11746299e-01
-1.54220283e+00 -1.15498066e+00 -2.31550261e-01 3.54546070e-01
6.03989661e-01 -2.55572855e-01 8.25778544e-01 6.57105803e-01
-4.86713618e-01 7.82103002e-01 -4.50174034e-01 -5.95392101e-02
7.81873882e-01 -1.15503466e+00 5.98155081e-01 1.16501296e+00
6.87732175e-02 9.13031101e-01 6.77556455e-01 -6.56564057e-01
-1.39653945e+00 -1.11490715e+00 1.42092669e+00 -3.71119291e-01
6.15792990e-01 -5.07056534e-01 -8.93032551e-01 7.37672687e-01
3.74125183e-01 -4.07026559e-01 9.78304803e-01 2.10441351e-01
-6.65468216e-01 -2.24662632e-01 -1.23034787e+00 -1.26605453e-02
8.07798207e-01 -8.10973287e-01 -1.06821108e+00 4.86917160e-02
8.30619574e-01 3.33107891e-03 -1.15053916e+00 4.70720232e-01
5.33838987e-01 -9.86631811e-01 3.15814793e-01 -7.59549558e-01
6.93021297e-01 -2.97046661e-01 -2.93232203e-01 -1.16682363e+00
-6.21374726e-01 -2.84047991e-01 4.54153210e-01 1.58907104e+00
8.92870724e-01 -3.83835286e-01 4.55826849e-01 9.09168899e-01
-4.57423508e-01 -5.16835809e-01 -6.73976779e-01 -4.98110712e-01
-1.81274161e-01 -5.49592078e-01 8.08570683e-01 1.03989494e+00
6.17104888e-01 5.40971875e-01 -5.95239252e-02 4.99765389e-02
1.13522299e-01 3.44694346e-01 3.98260653e-01 -1.22998869e+00
-2.97958016e-01 -3.73157412e-01 -2.41197705e-01 -5.22705436e-01
2.57591903e-01 -9.14081693e-01 2.49057114e-01 -1.62390935e+00
1.18685411e-02 -3.53898615e-01 -3.97812247e-01 7.13555634e-01
2.30137005e-01 -1.10217556e-01 -1.08783573e-01 -1.62502870e-01
-6.80200279e-01 3.38325769e-01 7.45950341e-01 -1.92989498e-01
-2.13431016e-01 3.25686708e-02 -1.18274534e+00 7.53296390e-02
8.16965401e-01 -5.73869288e-01 -7.31136262e-01 -8.61568093e-01
7.49397159e-01 3.94207239e-02 -2.08003491e-01 -7.20097899e-01
4.16575462e-01 -6.83978721e-02 -9.07314569e-02 -4.46612656e-01
-3.76234174e-01 -9.41605449e-01 8.75542834e-02 3.52466494e-01
-6.34537041e-01 6.27326012e-01 3.24250907e-01 -3.77354026e-02
-2.76205510e-01 -3.61491442e-01 5.21949232e-01 8.77990946e-03
-6.32227004e-01 -3.05664867e-01 -6.03749692e-01 7.99101666e-02
9.76174831e-01 -3.48878503e-01 -2.95873404e-01 2.73208879e-02
-2.61834294e-01 4.07988638e-01 4.00964379e-01 7.73486316e-01
8.81562710e-01 -8.03136289e-01 -9.09883797e-01 6.48973405e-01
7.12806225e-01 -2.20644191e-01 -2.06276536e-01 8.57400298e-01
-1.83663890e-01 3.56924891e-01 -5.05883336e-01 -4.07382786e-01
-1.33404160e+00 3.95219892e-01 3.89340580e-01 -3.16206038e-01
-2.19776809e-01 6.92626953e-01 -3.47705305e-01 -6.27930641e-01
5.15126027e-02 -6.12688899e-01 -3.29815894e-01 -1.20704740e-01
3.73894989e-01 4.83086780e-02 6.88181937e-01 -3.45106840e-01
-9.27609026e-01 3.77810299e-01 -4.76703763e-01 3.46031077e-02
1.49929380e+00 -1.24912225e-01 -5.71253598e-01 1.86560407e-01
9.19051886e-01 -4.89548668e-02 -3.83189678e-01 -2.23512009e-01
8.37857723e-01 -2.77764797e-01 3.20361964e-02 -1.01985490e+00
-8.56067955e-01 5.27889132e-01 1.57408252e-01 2.80677736e-01
1.36147892e+00 -1.62899107e-01 2.30505303e-01 2.93034852e-01
4.77796525e-01 -9.52651143e-01 9.57587287e-02 6.75638795e-01
8.92899990e-01 -1.16230333e+00 -2.47865781e-01 -4.68213856e-01
-6.72538221e-01 1.12148142e+00 8.85613620e-01 1.67687863e-01
4.82409447e-01 8.21954370e-01 1.86711237e-01 -3.71191591e-01
-1.29490149e+00 3.62389833e-02 2.54038036e-01 4.06867027e-01
9.58698094e-01 -2.47038364e-01 -3.95304769e-01 8.95189703e-01
6.82789180e-03 2.76864082e-01 7.42320359e-01 1.25100243e+00
-2.22513974e-01 -1.56019032e+00 1.70381308e-01 6.72242224e-01
-3.33681434e-01 -3.69925380e-01 -6.62590802e-01 4.79551136e-01
6.13072328e-02 1.39992642e+00 -1.04017623e-01 -9.98615563e-01
6.59264743e-01 2.78238297e-01 5.54069757e-01 -1.18262565e+00
-1.38439047e+00 -3.33661884e-01 5.86305439e-01 -4.33378369e-01
-5.45841813e-01 -4.80663508e-01 -1.17953324e+00 2.12795055e-03
-5.39714813e-01 4.24405456e-01 5.46912134e-01 1.19357240e+00
6.91174746e-01 8.28465760e-01 8.34874153e-01 -7.34219179e-02
-9.93278980e-01 -1.10485923e+00 -1.58980042e-01 2.03162760e-01
3.23737949e-01 -3.65164340e-01 -7.56809413e-02 1.11949138e-01]
|
[7.734042167663574, 7.873556137084961]
|
336935a1-f8ab-4e73-bd49-84b2428dbf3f
|
depth-relative-self-attention-for-monocular
|
2304.12849
| null |
https://arxiv.org/abs/2304.12849v1
|
https://arxiv.org/pdf/2304.12849v1.pdf
|
Depth-Relative Self Attention for Monocular Depth Estimation
|
Monocular depth estimation is very challenging because clues to the exact depth are incomplete in a single RGB image. To overcome the limitation, deep neural networks rely on various visual hints such as size, shade, and texture extracted from RGB information. However, we observe that if such hints are overly exploited, the network can be biased on RGB information without considering the comprehensive view. We propose a novel depth estimation model named RElative Depth Transformer (RED-T) that uses relative depth as guidance in self-attention. Specifically, the model assigns high attention weights to pixels of close depth and low attention weights to pixels of distant depth. As a result, the features of similar depth can become more likely to each other and thus less prone to misused visual hints. We show that the proposed model achieves competitive results in monocular depth estimation benchmarks and is less biased to RGB information. In addition, we propose a novel monocular depth estimation benchmark that limits the observable depth range during training in order to evaluate the robustness of the model for unseen depths.
|
['Byonghyo Shim', 'Gusang Lee', 'Jiyoung Kim', 'Kyuhong Shim']
|
2023-04-25
| null | null | null | null |
['monocular-depth-estimation']
|
['computer-vision']
|
[ 7.06992149e-02 3.13119471e-01 -1.66436970e-01 -4.96240795e-01
-3.71954799e-01 -3.98613483e-01 4.88078624e-01 2.58896612e-02
-4.13973123e-01 6.56313241e-01 1.79715797e-01 -6.59748539e-02
3.45478445e-01 -1.04356372e+00 -7.91962981e-01 -7.38635778e-01
3.65323961e-01 7.02652633e-02 5.36792576e-01 -1.37687204e-02
5.34413755e-01 5.67981184e-01 -1.70546508e+00 2.66495377e-01
6.60556316e-01 1.27697432e+00 4.17492867e-01 6.98063850e-01
-1.45016402e-01 1.00425160e+00 -6.34011567e-01 -1.63895220e-01
4.49043155e-01 -1.91046923e-01 -7.07825303e-01 -3.44869196e-02
8.60494673e-01 -1.07455623e+00 -7.05864429e-01 1.01713634e+00
4.27948296e-01 1.42190251e-02 3.90398234e-01 -1.08558023e+00
-4.15792733e-01 2.98769236e-01 -6.12835288e-01 3.56460065e-01
5.40345848e-01 3.27042282e-01 8.58866274e-01 -7.93471873e-01
5.72585166e-01 1.11349332e+00 1.71333387e-01 6.85799658e-01
-7.81832635e-01 -4.68141705e-01 6.99927807e-01 2.36930415e-01
-8.77291203e-01 -3.36824179e-01 9.32227015e-01 -1.92242011e-01
9.94568169e-01 2.92166881e-02 8.24007452e-01 1.07463944e+00
3.59011553e-02 9.79326725e-01 9.82983470e-01 -2.51473218e-01
3.98266584e-01 1.83295794e-02 -6.35224581e-02 7.23591864e-01
2.82054782e-01 3.76017511e-01 -7.98138261e-01 3.09797615e-01
9.12757814e-01 1.44599646e-01 -6.60734236e-01 -6.71099067e-01
-9.52838778e-01 6.31640851e-01 1.12044585e+00 9.53074843e-02
-1.85492396e-01 3.97917390e-01 -1.97847281e-02 2.62160855e-03
3.68690044e-01 5.60037911e-01 -5.10246694e-01 -1.00003712e-01
-5.28667569e-01 -5.70802763e-02 3.07283819e-01 8.89956534e-01
1.29016030e+00 -2.11811066e-01 2.95328666e-02 4.85746473e-01
1.72154501e-01 3.15333158e-01 4.49819148e-01 -1.23279202e+00
6.20524108e-01 8.14979076e-01 7.98720866e-02 -8.72371852e-01
-4.72627968e-01 -2.33306855e-01 -6.47707701e-01 6.35490596e-01
5.74955344e-01 9.31765605e-03 -1.04733038e+00 1.52745295e+00
2.26146117e-01 -2.97462285e-01 -6.12913147e-02 1.14939868e+00
1.06603491e+00 3.02782327e-01 -3.53478581e-01 2.60840863e-01
8.03174198e-01 -1.01154506e+00 -2.82646477e-01 -9.58782732e-01
4.94542658e-01 -4.27371025e-01 1.02375221e+00 4.40188468e-01
-1.06925941e+00 -5.81315279e-01 -1.19605899e+00 -5.69469512e-01
-3.52591187e-01 -1.06471069e-01 8.63127053e-01 4.63462293e-01
-1.19484746e+00 7.80693114e-01 -8.99673522e-01 -2.69621819e-01
4.52543914e-01 2.01416850e-01 -2.61088073e-01 -4.19199169e-01
-8.79448831e-01 5.24558663e-01 2.89333105e-01 1.93047598e-01
-8.62437367e-01 -4.60715473e-01 -1.10765433e+00 -4.01131772e-02
2.21710786e-01 -7.13874578e-01 1.02314508e+00 -9.77506340e-01
-1.52548742e+00 8.49016786e-01 -3.12554330e-01 -1.54873386e-01
7.06648350e-01 -4.69308019e-01 3.16956967e-01 4.59677577e-01
7.40090907e-02 1.24660742e+00 9.13514376e-01 -1.42933440e+00
-9.33127165e-01 -7.00529695e-01 6.34469986e-01 3.94006312e-01
-1.27698004e-01 -7.34238029e-01 -6.07810378e-01 -8.56359378e-02
6.92113936e-01 -6.18160546e-01 -3.75137827e-03 3.80191058e-01
-5.61090589e-01 1.61706567e-01 6.44313276e-01 -2.00122938e-01
8.39397132e-01 -2.05017185e+00 2.47803241e-01 -2.85155550e-02
4.72398371e-01 -3.01890254e-01 1.38756722e-01 -7.63859451e-02
1.66797802e-01 1.07951604e-01 3.45325246e-02 -4.07439083e-01
-3.90476912e-01 3.10758740e-01 -1.45508066e-01 3.96546006e-01
9.88315269e-02 7.80237854e-01 -1.14900148e+00 -2.45519757e-01
4.86686856e-01 2.02915028e-01 -6.23736084e-01 4.49226201e-01
-1.87630251e-01 4.00679678e-01 -2.48739153e-01 9.56281960e-01
8.11568856e-01 -3.90454710e-01 -5.34290746e-02 -3.53747427e-01
-9.92093608e-02 3.41989756e-01 -6.49058342e-01 1.64033151e+00
-6.30375326e-01 9.96104777e-01 -2.79280931e-01 -4.35700268e-01
8.35456371e-01 -2.18647793e-01 9.29280892e-02 -9.55223560e-01
9.72936228e-02 1.10757232e-01 -1.86979204e-01 -3.60312015e-01
7.37463713e-01 2.40787566e-01 2.59062380e-01 3.16659003e-01
-1.81824461e-01 -4.45794165e-01 -1.78420603e-01 2.26035118e-01
1.08760428e+00 3.17421496e-01 2.68676519e-01 2.88871408e-01
1.53139576e-01 -4.16741937e-01 4.80979890e-01 7.31449246e-01
-2.99631178e-01 9.65758920e-01 6.48069978e-01 -5.81107497e-01
-9.89297867e-01 -1.00187755e+00 -2.22246461e-02 8.55230153e-01
6.44079208e-01 -2.15934858e-01 -2.77468950e-01 -8.96501601e-01
2.04456281e-02 3.04168195e-01 -1.12055397e+00 -1.33285537e-01
-2.51322359e-01 -3.74206185e-01 4.85258289e-02 8.75450611e-01
6.98515892e-01 -1.07409084e+00 -1.15596187e+00 -4.71210442e-02
-8.12709779e-02 -1.26808214e+00 -1.42911434e-01 6.38103843e-01
-1.10023451e+00 -1.19434023e+00 -1.03319013e+00 -4.06284422e-01
8.67681205e-01 4.96775210e-01 1.22933185e+00 2.30259210e-01
7.33171254e-02 2.39537284e-01 -3.78241807e-01 -6.59592301e-02
9.96276736e-02 2.16333181e-01 -5.48698783e-01 -2.09911823e-01
2.84604311e-01 -4.85797912e-01 -1.26673794e+00 3.79860997e-01
-6.10048413e-01 2.16084853e-01 4.47617918e-01 8.41499567e-01
3.88848811e-01 -3.17981601e-01 5.87084740e-02 -7.79369712e-01
-4.08782884e-02 -2.66268820e-01 -6.88906729e-01 -2.02350765e-01
-4.35566515e-01 2.09239602e-01 4.31717247e-01 -2.41073504e-01
-9.49428260e-01 8.90348256e-02 -3.58427987e-02 -4.44863319e-01
-2.02130184e-01 2.70120036e-02 -2.13460326e-01 -3.90622228e-01
6.66381836e-01 1.71478122e-01 -3.85970235e-01 -2.35799477e-01
1.03131346e-01 3.21619987e-01 3.80946517e-01 -2.95182139e-01
4.07401055e-01 8.81579816e-01 3.74035388e-02 -6.49478793e-01
-1.17739272e+00 -4.19186741e-01 -5.64232230e-01 -2.35943228e-01
6.75004601e-01 -1.07162678e+00 -6.71509027e-01 7.34785259e-01
-1.09263563e+00 -7.79403627e-01 -1.56668779e-02 2.74799407e-01
-5.13987303e-01 5.57240248e-01 -7.28259563e-01 -7.68533289e-01
3.89345996e-02 -1.16372025e+00 1.21934354e+00 3.78870279e-01
-5.19087315e-02 -8.68306041e-01 -2.53017187e-01 1.68024495e-01
2.33583868e-01 2.24509791e-01 7.64407992e-01 2.19573349e-01
-1.00445807e+00 2.13910025e-02 -6.53506398e-01 2.14156717e-01
2.45420411e-01 6.57940730e-02 -1.52656960e+00 -1.66189313e-01
-1.42708987e-01 -4.93518084e-01 1.23088515e+00 3.82674962e-01
1.41308165e+00 -5.07533699e-02 -3.01861823e-01 1.05260253e+00
1.59398508e+00 1.48999095e-01 9.20469344e-01 7.25149810e-01
1.03915572e+00 6.62018418e-01 5.36608636e-01 4.68548656e-01
4.78016615e-01 5.39665639e-01 1.08458602e+00 -3.09068084e-01
-2.30531376e-02 -3.65422189e-01 7.68656358e-02 9.98302251e-02
-6.09645844e-02 -3.42069536e-01 -6.63497865e-01 6.19397402e-01
-1.54350257e+00 -5.66486478e-01 1.42326698e-01 2.26580405e+00
6.32434309e-01 4.93661970e-01 -1.12913832e-01 2.60060549e-01
2.78984725e-01 2.80899823e-01 -9.11015689e-01 -2.41534263e-01
-3.88987422e-01 -2.55616494e-02 6.89696252e-01 5.98415852e-01
-8.07300329e-01 9.54895318e-01 6.11861134e+00 2.30900288e-01
-1.25074911e+00 -2.95215011e-01 9.05362785e-01 -2.76679128e-01
-5.43137670e-01 -1.71827637e-02 -7.69994736e-01 4.98501062e-01
6.66541532e-02 4.32200193e-01 3.26126605e-01 9.69909251e-01
-1.55802935e-01 -8.83596897e-01 -1.44962847e+00 1.24448907e+00
-3.44244726e-02 -1.05356896e+00 6.59382790e-02 1.27072528e-01
8.95372748e-01 1.62793577e-01 1.46573231e-01 6.74276939e-03
1.17618032e-01 -9.83303666e-01 7.53561497e-01 3.66947204e-01
6.91918194e-01 -7.57744372e-01 9.05021846e-01 1.88766420e-01
-8.69343877e-01 -2.20930204e-01 -7.59337485e-01 -2.86479831e-01
-1.11651637e-01 6.17058516e-01 -6.94681585e-01 2.40893662e-01
8.92829120e-01 9.79664922e-01 -6.81738794e-01 1.02686417e+00
-6.45918906e-01 -1.16648659e-01 -3.28810692e-01 4.32535671e-02
2.61676848e-01 1.91612139e-01 -2.63479687e-02 6.25456572e-01
3.88059705e-01 -1.34831732e-02 -2.51996577e-01 8.44515145e-01
-2.94261780e-02 -2.39758253e-01 -7.72984922e-01 4.43828225e-01
2.75827318e-01 8.84120107e-01 -6.16834879e-01 -2.62550145e-01
-5.15715957e-01 1.35909021e+00 7.65147686e-01 4.97760147e-01
-4.31102306e-01 -9.51183140e-02 7.32311249e-01 1.85861111e-01
3.36905807e-01 2.19119638e-02 -5.69176793e-01 -1.14776027e+00
-1.66941492e-03 -3.89825583e-01 2.04059198e-01 -1.32488942e+00
-9.83503699e-01 7.64415681e-01 -3.76005113e-01 -1.15072477e+00
-1.80583313e-01 -7.29363680e-01 -3.76153886e-01 7.21887052e-01
-2.01399732e+00 -5.05773425e-01 -1.09620965e+00 4.62223738e-01
3.64455223e-01 2.81859100e-01 3.79197925e-01 9.69794989e-02
-2.91197568e-01 6.19842827e-01 -4.21655178e-01 -3.58734885e-03
6.95158899e-01 -1.46957481e+00 4.43406820e-01 6.61591291e-01
-1.71246767e-01 2.72469074e-01 5.27860641e-01 -4.42536324e-01
-1.01154649e+00 -7.55304813e-01 5.14124334e-01 -5.17062187e-01
1.84325337e-01 -2.44053200e-01 -8.49265993e-01 4.85852778e-01
-1.49786979e-01 2.22682953e-01 1.03621580e-01 2.18724646e-03
-4.23742712e-01 -1.90099284e-01 -1.08720326e+00 5.72679818e-01
1.23771966e+00 -6.70997381e-01 -1.65456995e-01 8.94792899e-02
6.57664895e-01 -6.49179459e-01 -5.04142404e-01 3.47519130e-01
5.46150506e-01 -1.83489871e+00 1.02508211e+00 6.39097765e-03
9.14997160e-01 -1.17837027e-01 -2.80306101e-01 -1.23357439e+00
1.00698425e-02 -1.78187609e-01 -2.15131909e-01 6.94341779e-01
3.77654523e-01 -5.36707997e-01 1.26751924e+00 5.82508743e-01
-9.72083881e-02 -8.85174334e-01 -8.57874036e-01 -5.31885624e-01
-9.97287780e-02 -3.66269559e-01 5.12392938e-01 6.33165300e-01
-1.38106257e-01 1.24811009e-01 -2.88851768e-01 1.82293355e-01
6.47099972e-01 2.94267178e-01 8.74571323e-01 -1.06261146e+00
-3.27567726e-01 -5.26543856e-01 -5.69035351e-01 -1.81134903e+00
-1.86108902e-01 -3.20173264e-01 1.39693171e-01 -1.52744401e+00
1.91270307e-01 -4.16658551e-01 -1.06527299e-01 4.07377750e-01
-3.92217815e-01 5.10927916e-01 7.57499561e-02 1.80476811e-02
-4.45021003e-01 6.86527669e-01 1.61108530e+00 -2.21099071e-02
-8.11999738e-02 -8.12408775e-02 -5.22311509e-01 8.44528913e-01
7.27770865e-01 -1.93848744e-01 -3.20984393e-01 -5.86686015e-01
6.28321588e-01 1.85880482e-01 5.28210342e-01 -1.05607402e+00
2.22259194e-01 -3.81959155e-02 7.83588886e-01 -8.05271983e-01
5.86363375e-01 -6.60227239e-01 -5.65594375e-01 3.53805870e-01
-6.96202442e-02 -1.40946239e-01 4.24685590e-02 5.16503096e-01
-2.09732786e-01 -1.63598552e-01 7.02052951e-01 -4.08102095e-01
-1.07118249e+00 3.50383639e-01 -1.77504532e-02 4.82004397e-02
7.28374660e-01 -6.19882226e-01 -5.42000830e-01 -6.24528050e-01
-4.95696753e-01 1.33289158e-01 9.78081763e-01 1.90580308e-01
9.71829653e-01 -1.10715330e+00 -1.22562677e-01 3.42291355e-01
4.88076597e-01 4.88583654e-01 3.51659149e-01 3.54195118e-01
-7.49507844e-01 3.44881952e-01 -3.80040139e-01 -7.86413193e-01
-9.01488185e-01 6.67459488e-01 5.92709363e-01 1.67933434e-01
-6.33635461e-01 1.12379754e+00 7.61381865e-01 -1.10803887e-01
5.16361296e-01 -7.80245423e-01 5.62840663e-02 -1.46119028e-01
4.97477859e-01 1.80800557e-01 1.11628687e-02 -1.37813106e-01
-3.35328668e-01 8.75581741e-01 -1.05846979e-01 2.82869544e-02
1.06074107e+00 -5.43279827e-01 2.12642938e-01 4.84903723e-01
1.05643237e+00 4.81673777e-02 -2.12127852e+00 -1.56530157e-01
-3.63671422e-01 -9.49244440e-01 2.77785480e-01 -7.19198287e-01
-1.41820514e+00 1.20404637e+00 4.98320878e-01 -1.41954282e-02
1.18065715e+00 4.33668606e-02 5.72928190e-01 3.34356427e-01
5.64403534e-01 -8.53585422e-01 4.83986288e-01 4.68754441e-01
7.25349247e-01 -1.69169712e+00 -9.91718625e-05 -4.20570761e-01
-4.13620204e-01 1.19311988e+00 1.15920627e+00 1.81412529e-02
3.08914632e-01 2.25018680e-01 2.64296085e-01 -1.51069939e-01
-6.08155310e-01 -3.16890955e-01 1.13386072e-01 7.57394314e-01
2.53660709e-01 -2.98559368e-01 4.59542155e-01 -1.03435591e-01
-1.76275373e-01 -2.60389447e-01 7.30849564e-01 7.95468509e-01
-5.42492568e-01 -6.90088451e-01 -1.30750462e-01 1.67029828e-01
-1.90927446e-01 -2.10433707e-01 -5.23223996e-01 7.77704120e-01
5.49542308e-02 6.97104573e-01 3.22243601e-01 -2.76308954e-01
3.46111134e-03 -4.45514500e-01 7.41337776e-01 -6.21299565e-01
-2.40668744e-01 -2.02326924e-01 -2.06003651e-01 -9.55703080e-01
-3.37681413e-01 -2.09743142e-01 -1.01838279e+00 -3.49236012e-01
-2.25581527e-01 -4.73400712e-01 5.83654284e-01 7.87909687e-01
5.75658567e-02 6.30404115e-01 5.79626203e-01 -1.14908874e+00
5.87319285e-02 -8.60947311e-01 -4.42177445e-01 2.93630511e-01
7.22460449e-01 -6.81238949e-01 -7.46991873e-01 -5.69750607e-01]
|
[8.849135398864746, -2.3470170497894287]
|
5034ca6c-d8b3-4f45-bb0c-9f151d69a14a
|
the-use-of-data-augmentation-as-a-technique
|
2205.00452
| null |
https://arxiv.org/abs/2205.00452v1
|
https://arxiv.org/pdf/2205.00452v1.pdf
|
The use of Data Augmentation as a technique for improving neural network accuracy in detecting fake news about COVID-19
|
This paper aims to present how the application of Natural Language Processing (NLP) and data augmentation techniques can improve the performance of a neural network for better detection of fake news in the Portuguese language. Fake news is one of the main controversies during the growth of the internet in the last decade. Verifying what is fact and what is false has proven to be a difficult task, while the dissemination of false news is much faster, which leads to the need for the creation of tools that, automated, assist in the process of verification of what is fact and what is false. In order to bring a solution, an experiment was developed with neural network using news, real and fake, which were never seen by artificial intelligence (AI). There was a significant performance in the news classification after the application of the mentioned techniques.
|
['Arnaldo Bispo de Jesus', 'Andre Brasil Vieira Wyzykowski', 'Mauricio S. da Cruz', 'Wilton O. Júnior']
|
2022-05-01
| null | null | null | null |
['news-classification']
|
['natural-language-processing']
|
[-5.70968576e-02 3.27745706e-01 -3.88547145e-02 -2.05945849e-01
1.68530777e-01 -4.16055590e-01 1.09375191e+00 6.66916013e-01
-5.72621465e-01 9.83201683e-01 2.84274995e-01 -5.33822060e-01
1.47010893e-01 -8.91111076e-01 -7.75250137e-01 -1.68318138e-01
2.28422001e-01 4.69318211e-01 2.40152448e-01 -5.96764147e-01
6.69927299e-01 4.57159817e-01 -1.40803802e+00 7.85345912e-01
4.71169800e-01 4.81398433e-01 -1.16766296e-01 2.77993917e-01
-6.01385713e-01 1.10773611e+00 -1.10092711e+00 -4.90835369e-01
6.37685359e-02 -4.82789397e-01 -9.88094866e-01 -2.62294054e-01
1.06210448e-01 -9.13975760e-02 -5.17073646e-02 1.34472251e+00
5.21266717e-04 -3.06853056e-01 5.01476645e-01 -1.02480721e+00
-5.87671399e-01 7.36688912e-01 -2.77221143e-01 4.69657540e-01
4.67968762e-01 -1.89942107e-01 2.73027003e-01 -4.54036951e-01
7.38224924e-01 1.24126637e+00 4.78835195e-01 2.18815416e-01
-5.87285876e-01 -5.51176548e-01 -4.51558352e-01 3.26118737e-01
-8.93660069e-01 -4.97369654e-02 7.02467799e-01 -6.25060678e-01
6.89760029e-01 8.30968991e-02 7.73148835e-01 1.12623847e+00
4.01569098e-01 3.54199678e-01 1.47780776e+00 -8.58301044e-01
1.19095169e-01 1.04056692e+00 4.92339045e-01 4.69879240e-01
7.17052460e-01 1.78643465e-01 -3.35892320e-01 -2.03001890e-02
3.55332881e-01 -2.75616437e-01 -1.93084404e-01 3.77135456e-01
-9.13689911e-01 9.69407678e-01 3.94676834e-01 1.24309587e+00
-4.46913958e-01 -4.23542619e-01 6.32901073e-01 6.64358437e-01
6.18564963e-01 8.95938516e-01 -3.14916551e-01 -4.20073360e-01
-8.71982932e-01 6.75836802e-02 1.04157376e+00 7.75015131e-02
1.15116186e-01 2.07853228e-01 4.53661680e-01 2.75307983e-01
-2.40594707e-02 3.29075664e-01 9.76998985e-01 -1.70243174e-01
3.18361729e-01 9.89973903e-01 2.89453357e-01 -1.75904155e+00
-2.24717870e-01 -6.40769482e-01 -5.82685828e-01 3.65111649e-01
7.05036759e-01 -1.12569727e-01 -6.30158663e-01 1.11259973e+00
1.99952468e-01 -2.42448561e-02 3.46962214e-01 8.72311175e-01
7.42016315e-01 9.88899946e-01 -2.19787091e-01 -2.93778002e-01
1.26198113e+00 -4.56882745e-01 -1.11247969e+00 -1.90752253e-01
5.73900938e-01 -7.68854022e-01 7.17116058e-01 5.56092680e-01
-4.18193281e-01 -3.48889828e-01 -1.21304238e+00 2.66877800e-01
-1.11660194e+00 4.30610217e-02 4.36947823e-01 9.37915266e-01
-3.87135506e-01 6.33059621e-01 -3.52493584e-01 -3.48900765e-01
1.86591938e-01 2.49521881e-02 -4.88522977e-01 1.89257815e-01
-1.40512800e+00 1.46509790e+00 9.79042768e-01 8.92400369e-02
-2.93369740e-01 -8.31386447e-03 -3.39482218e-01 -3.38067710e-02
3.07853222e-01 4.51927744e-02 7.83717930e-01 -1.60970366e+00
-1.01124167e+00 1.03065598e+00 1.51766047e-01 -9.60629761e-01
8.68685603e-01 -1.21580690e-01 -8.60357761e-01 6.23320974e-02
2.02757902e-02 1.87453553e-02 8.00003409e-01 -1.22323668e+00
-5.76107800e-01 -4.03739750e-01 -1.93113416e-01 -3.66111517e-01
-2.49007583e-01 2.57782251e-01 4.08025563e-01 -5.15095413e-01
1.04545765e-01 -4.46801782e-01 4.82165605e-01 -5.44164479e-01
-1.16509899e-01 -1.86696559e-01 1.01263106e+00 -9.97489929e-01
9.20133889e-01 -1.94657624e+00 -4.02462691e-01 2.62308747e-01
1.74606085e-01 7.42307127e-01 4.09778148e-01 3.76383752e-01
-1.58876732e-01 3.38162810e-01 6.24311157e-02 3.09897810e-01
-3.98979902e-01 3.96710247e-01 -3.73641104e-01 3.21705490e-01
2.20417276e-01 3.15690964e-01 -8.39120150e-01 -2.90051013e-01
7.25606009e-02 3.37126344e-01 2.11719219e-02 -3.17885309e-01
-1.63398653e-01 2.48845533e-01 -4.27485824e-01 1.91227660e-01
4.04444933e-01 1.69856995e-02 1.12071060e-01 -9.66005996e-02
-3.73914182e-01 3.55509877e-01 -9.62012708e-01 4.08095360e-01
9.50670801e-03 1.05796385e+00 -4.68339205e-01 -1.28146017e+00
1.24389017e+00 5.92518210e-01 5.51906712e-02 -8.78591061e-01
8.73557389e-01 5.98999560e-01 -7.33724535e-02 -8.29415083e-01
5.30438364e-01 -3.02233070e-01 2.38766223e-01 4.04545099e-01
-2.12297574e-01 1.96894016e-02 2.10017428e-01 8.66615921e-02
5.43370605e-01 -1.68932498e-01 3.72403085e-01 -1.03019536e-01
6.50683284e-01 4.64213878e-01 3.29366505e-01 6.25503182e-01
-9.59543809e-02 -1.75534964e-01 7.69830406e-01 -8.41108203e-01
-9.99583840e-01 -3.07756037e-01 8.56878087e-02 2.50584364e-01
-6.36066718e-04 2.98647374e-01 -5.55857778e-01 -6.56808972e-01
-1.16669402e-01 1.01612401e+00 -5.12892365e-01 -1.09380566e-01
-3.49917173e-01 -6.87453032e-01 5.03854632e-01 -3.99539948e-01
9.44358289e-01 -1.20889044e+00 -7.79754341e-01 3.43986809e-01
-1.69964910e-01 -1.22567618e+00 7.14983463e-01 -2.49681063e-03
-8.30227435e-01 -1.17899680e+00 -3.13991219e-01 -6.31961942e-01
4.76171076e-01 -7.94236548e-03 7.25899577e-01 4.35744524e-01
8.17528665e-02 -3.45956862e-01 -7.09967792e-01 -9.45415378e-01
-1.36762214e+00 -4.13500033e-02 6.30435068e-03 1.15371756e-02
4.92145360e-01 -2.09614411e-01 3.55440408e-01 1.45234661e-02
-1.09151673e+00 1.02698125e-01 4.45793211e-01 6.25093162e-01
-2.97544718e-01 6.35580599e-01 7.38078773e-01 -8.98324668e-01
1.04419112e+00 -4.90104914e-01 -7.42294669e-01 6.51608855e-02
-5.05764008e-01 2.22852919e-02 6.12432778e-01 -4.51085031e-01
-8.48613203e-01 -4.24927294e-01 1.75093375e-02 3.61385882e-01
-4.51823652e-01 6.92573905e-01 4.37372267e-01 -1.70842677e-01
1.13032603e+00 1.11128956e-01 1.75433919e-01 -3.25709909e-01
-1.98500633e-01 9.61093009e-01 2.41208822e-01 2.59149466e-02
7.03924894e-01 2.58862317e-01 -1.56907901e-01 -1.09964526e+00
-8.00167084e-01 -3.43246430e-01 -2.36882508e-01 -4.44143057e-01
5.86248338e-01 -4.41524446e-01 -5.14696121e-01 5.10589898e-01
-1.50948763e+00 2.28785470e-01 4.76379693e-03 6.05013371e-01
1.93526193e-01 2.48578027e-01 -4.03468549e-01 -1.08659124e+00
-2.32998416e-01 -7.32035339e-01 -1.55605435e-01 1.88574880e-01
-1.09579258e-01 -7.14139462e-01 -1.41619891e-01 6.54410839e-01
4.69977826e-01 3.83703947e-01 6.12000585e-01 -1.09536219e+00
-8.84537399e-02 -8.93891573e-01 -3.35549295e-01 5.50845981e-01
-2.79157925e-02 8.13925415e-02 -6.87169492e-01 2.91474164e-01
5.93438983e-01 -6.14327639e-02 3.36260796e-01 -2.92232096e-01
2.16803774e-01 -7.07326829e-01 -7.17656547e-03 -4.71700937e-01
1.41891885e+00 5.51789761e-01 5.75124621e-01 6.75313890e-01
-4.34776358e-02 6.91092730e-01 6.37073815e-01 1.51707202e-01
3.46432836e-03 4.62884963e-01 4.56232518e-01 1.17059164e-01
1.10616364e-01 -3.79715040e-02 2.82199681e-01 5.76367795e-01
-6.05841540e-02 -1.60818651e-01 -1.10257077e+00 4.17720199e-01
-1.44826794e+00 -1.30695724e+00 -6.70544565e-01 1.78349924e+00
6.45659268e-01 7.02591479e-01 4.83250059e-03 8.24570894e-01
7.07105815e-01 -6.10301048e-02 4.20154810e-01 -8.42749536e-01
-1.66486293e-01 -1.44568443e-01 3.62496763e-01 6.15267694e-01
-9.32593107e-01 9.08621490e-01 5.30034399e+00 4.38240051e-01
-1.68065500e+00 2.47735158e-01 4.25631583e-01 5.74157357e-01
-9.82563291e-03 -2.82098591e-01 -3.50285143e-01 7.28485763e-01
8.11013699e-01 5.95996939e-02 3.76234770e-01 6.86353564e-01
6.71296000e-01 -5.59661269e-01 -2.37180665e-01 6.22989118e-01
4.01282996e-01 -1.51126134e+00 2.33006164e-01 -2.82582849e-01
5.96085131e-01 1.28158778e-01 -4.50859606e-01 1.06181301e-01
-1.09073512e-01 -7.83275485e-01 7.24383771e-01 4.45171624e-01
-2.31223106e-01 -5.68878949e-01 1.58204484e+00 8.13185573e-01
-2.56181955e-02 -6.14226162e-02 -2.06023961e-01 -6.06714308e-01
9.14280117e-02 7.23504424e-01 -1.19025397e+00 4.23421681e-01
5.20215750e-01 7.45265931e-02 -5.77436566e-01 1.08052850e+00
-4.20800626e-01 6.14220023e-01 -3.72838438e-01 -7.88429618e-01
3.03173780e-01 -6.58526346e-02 6.30925298e-01 1.05906773e+00
-7.79617280e-02 -5.73268160e-02 -2.95902282e-01 6.93919957e-01
1.34673417e-01 4.65090662e-01 -7.47549653e-01 -5.73017240e-01
7.48888180e-02 6.56428397e-01 -8.97752583e-01 -5.26718020e-01
-1.07109480e-01 8.36228490e-01 -9.05938596e-02 2.78455298e-02
-7.16826022e-01 -1.88503817e-01 -2.77641147e-01 3.43859851e-01
-3.25331748e-01 -1.66667193e-01 -3.86932731e-01 -9.59692717e-01
-2.16203891e-02 -1.04081690e+00 7.73637593e-02 -8.75022709e-01
-9.58331525e-01 7.33767450e-01 -1.36339545e-01 -7.87086010e-01
-3.97690535e-02 -7.47938931e-01 -3.51383746e-01 6.05408967e-01
-1.12197745e+00 -7.73722708e-01 -1.13066152e-01 2.54617244e-01
2.06073210e-01 -4.66013581e-01 7.66750515e-01 3.05226684e-01
-1.41599208e-01 -1.16948694e-01 -1.41751751e-01 4.73301023e-01
2.34303162e-01 -6.70973361e-01 -7.91698471e-02 7.66653717e-01
3.24980587e-01 1.99071199e-01 1.24221396e+00 -9.26688790e-01
-6.87379837e-01 -3.25487584e-01 1.50208187e+00 -1.79047316e-01
8.22970808e-01 5.92074469e-02 -8.16266537e-01 3.01728606e-01
1.75473243e-01 -4.74841297e-01 1.65339559e-01 -2.96081752e-01
-3.43710542e-01 -3.46709578e-03 -1.49715853e+00 4.42481726e-01
-5.28049134e-02 -4.54573959e-01 -1.25149608e+00 6.16740644e-01
3.92050594e-01 1.17158694e-02 -2.27250174e-01 -9.16223302e-02
3.79926413e-01 -9.53303039e-01 4.18710262e-01 -8.78133297e-01
3.82569671e-01 -1.91342533e-01 3.57508034e-01 -1.06963480e+00
4.04543340e-01 -1.96682453e-01 5.16056240e-01 9.23881352e-01
6.55896127e-01 -1.11051393e+00 6.85371757e-01 1.31534949e-01
2.61820436e-01 -4.97668311e-02 -8.98125708e-01 -5.80710173e-01
-2.30315760e-01 -3.09807420e-01 2.07838714e-01 1.64146280e+00
1.52131721e-01 1.16058648e-01 -2.78859586e-01 2.24794254e-01
1.84306026e-01 -2.60339975e-01 6.45977676e-01 -1.44275808e+00
-9.96502489e-02 -4.31355417e-01 -8.58586550e-01 -1.53965190e-01
-2.51911491e-01 -4.35802013e-01 -3.12071860e-01 -1.39491832e+00
-2.75632799e-01 1.89288221e-02 2.20087573e-01 4.72724468e-01
2.81348377e-01 8.84417668e-02 2.06773579e-01 7.86315724e-02
2.63171881e-01 1.06898151e-01 1.11510813e+00 -5.56215681e-02
-1.15499862e-01 8.43458176e-02 -1.57986239e-01 8.84872913e-01
1.20639849e+00 -9.51982021e-01 4.97604609e-02 -1.47818685e-01
7.41230667e-01 1.31592542e-01 3.19921345e-01 -1.14895189e+00
5.36421835e-02 -3.29126045e-02 6.64923862e-02 -4.01474893e-01
1.92206115e-01 -1.25928986e+00 -2.48206942e-03 9.92037117e-01
-1.09049909e-01 2.64446110e-01 2.27835476e-01 3.10709178e-01
-5.45174003e-01 -7.17393875e-01 7.39135444e-01 -3.36313188e-01
-4.85324174e-01 -5.97280681e-01 -6.69736981e-01 -2.11396426e-01
9.35304463e-01 -2.47895494e-01 -5.93778253e-01 -5.73464096e-01
-5.63883305e-01 -2.80347377e-01 4.62516472e-02 4.67019796e-01
4.09524590e-01 -6.53024495e-01 -5.80114841e-01 -1.00931689e-01
-1.84094489e-01 -6.24554157e-01 -1.45182014e-01 8.52543771e-01
-1.26369762e+00 3.02623481e-01 -5.59570312e-01 1.19843639e-01
-1.50491536e+00 6.05872333e-01 3.47620189e-01 -1.65398479e-01
-3.46123785e-01 3.26660872e-01 -9.11770761e-01 -4.19212990e-02
1.07573971e-01 -1.78698197e-01 -7.56421983e-01 1.82590753e-01
5.58931649e-01 2.85323948e-01 7.54749328e-02 -8.34646761e-01
-5.94832711e-02 -7.13390112e-02 1.63773656e-01 -2.25730196e-01
1.35198665e+00 3.66204113e-01 -5.25924027e-01 6.31906569e-01
7.94972003e-01 2.46910438e-01 4.06675134e-03 1.65920168e-01
4.49731231e-01 -5.19920826e-01 1.69370085e-01 -1.24689436e+00
-7.02441275e-01 5.94304204e-01 4.89073128e-01 9.24754024e-01
5.19095719e-01 -3.02996963e-01 2.90286243e-01 6.01531148e-01
3.22735846e-01 -1.12655663e+00 -1.62503406e-01 6.32399738e-01
8.64989877e-01 -1.33940637e+00 -2.22594719e-02 -4.37770247e-01
-4.65048850e-01 1.42523420e+00 1.42464161e-01 -1.23317376e-01
5.08864880e-01 1.06884884e-02 2.14531168e-01 -4.81711328e-01
-1.34536654e-01 1.17597207e-01 1.24581540e-02 1.17878184e-01
3.11625004e-01 2.02429265e-01 -1.18513358e+00 2.09336966e-01
-2.42391899e-01 3.99069339e-01 9.28386033e-01 9.17472661e-01
-7.03985155e-01 -8.06731284e-01 -8.60079348e-01 3.50047320e-01
-8.28325391e-01 2.20983729e-01 -8.33945692e-01 1.07431555e+00
5.61536133e-01 1.16371691e+00 -6.15416467e-02 -2.19098836e-01
1.41005427e-01 1.32002935e-01 4.40642275e-02 -2.61010677e-01
-9.34181929e-01 -6.47621989e-01 6.08477294e-01 -1.01582050e-01
-4.65477169e-01 -1.54833212e-01 -1.16619182e+00 -4.68972832e-01
-5.89072883e-01 6.25992119e-01 1.36336875e+00 1.34765673e+00
-1.30016416e-01 2.91009188e-01 2.91864157e-01 -4.51607145e-02
-4.56873029e-01 -1.03488982e+00 -1.49499238e-01 6.10732615e-01
2.03367397e-01 -4.31145817e-01 -4.74366277e-01 -1.00953281e-01]
|
[8.19033145904541, 10.237860679626465]
|
de2c1026-2be8-4c7a-8ad5-7e9aae3cdb1c
|
neural-free-viewpoint-performance-rendering
|
2108.00362
| null |
https://arxiv.org/abs/2108.00362v2
|
https://arxiv.org/pdf/2108.00362v2.pdf
|
Neural Free-Viewpoint Performance Rendering under Complex Human-object Interactions
|
4D reconstruction of human-object interaction is critical for immersive VR/AR experience and human activity understanding. Recent advances still fail to recover fine geometry and texture results from sparse RGB inputs, especially under challenging human-object interactions scenarios. In this paper, we propose a neural human performance capture and rendering system to generate both high-quality geometry and photo-realistic texture of both human and objects under challenging interaction scenarios in arbitrary novel views, from only sparse RGB streams. To deal with complex occlusions raised by human-object interactions, we adopt a layer-wise scene decoupling strategy and perform volumetric reconstruction and neural rendering of the human and object. Specifically, for geometry reconstruction, we propose an interaction-aware human-object capture scheme that jointly considers the human reconstruction and object reconstruction with their correlations. Occlusion-aware human reconstruction and robust human-aware object tracking are proposed for consistent 4D human-object dynamic reconstruction. For neural texture rendering, we propose a layer-wise human-object rendering scheme, which combines direction-aware neural blending weight learning and spatial-temporal texture completion to provide high-resolution and photo-realistic texture results in the occluded scenarios. Extensive experiments demonstrate the effectiveness of our approach to achieve high-quality geometry and texture reconstruction in free viewpoints for challenging human-object interactions.
|
['Jingyi Yu', 'Jingya Wang', 'Lan Xu', 'Yuheng Jiang', 'Pei Lin', 'Anqi Pang', 'Yizhang Chen', 'Xin Chen', 'Guoxing Sun']
|
2021-08-01
| null | null | null | null |
['object-reconstruction']
|
['computer-vision']
|
[ 2.25951552e-01 -2.42954656e-01 3.95427406e-01 -3.49979997e-01
-6.05473936e-01 -2.18347579e-01 2.26011068e-01 -2.66709208e-01
3.60518843e-02 4.69968140e-01 1.51303550e-02 2.20058918e-01
-5.23971431e-02 -8.37342024e-01 -7.20205843e-01 -5.98800778e-01
2.30768293e-01 7.87191629e-01 3.94093335e-01 -1.22462742e-01
-3.02838951e-01 7.81583488e-01 -1.91860211e+00 4.47322041e-01
7.48681843e-01 1.24763978e+00 4.64241385e-01 9.37281907e-01
7.75187742e-03 8.91217530e-01 -2.36055985e-01 -2.23723531e-01
5.60092926e-01 -1.68969065e-01 -4.17064220e-01 6.02470994e-01
6.64142191e-01 -8.22349787e-01 -5.58152437e-01 5.31697869e-01
6.55800819e-01 3.57930213e-01 2.68387675e-01 -9.40687299e-01
-4.01011348e-01 -3.47470284e-01 -8.52590024e-01 -2.44094908e-01
6.69405937e-01 5.61780453e-01 5.01790166e-01 -1.02342606e+00
8.59349251e-01 1.43904471e+00 4.47973460e-01 6.74585402e-01
-1.27322578e+00 -5.05674720e-01 3.30165267e-01 1.31134745e-02
-1.30817819e+00 -3.18902224e-01 1.04881597e+00 -3.84317428e-01
7.23285735e-01 6.70691788e-01 1.32933950e+00 9.36107039e-01
1.75338537e-01 7.19074547e-01 1.01962793e+00 -4.20211740e-02
1.19938925e-01 6.27399907e-02 -3.17134619e-01 7.43646324e-01
1.30493358e-01 2.34124184e-01 -7.44420588e-01 -1.44754380e-01
1.69660985e+00 3.99473488e-01 -6.48318589e-01 -7.55459011e-01
-1.48333597e+00 1.40213400e-01 3.56521696e-01 -4.37973961e-02
-7.27837801e-01 3.09607118e-01 1.32750466e-01 -1.46345913e-01
6.88078284e-01 -7.16966484e-03 -4.20765340e-01 -1.91682409e-02
-5.99851966e-01 6.36664093e-01 4.97591168e-01 1.12647212e+00
4.97687936e-01 3.06159377e-01 -3.31247836e-01 6.80413604e-01
2.87552834e-01 8.59924257e-01 -1.07086584e-01 -1.22544837e+00
2.73591727e-01 6.02806151e-01 3.80411416e-01 -1.08193350e+00
-3.10770333e-01 -3.96462053e-01 -1.11839056e+00 4.70412582e-01
2.35965103e-01 3.54827821e-01 -1.06245410e+00 1.32334840e+00
1.16389251e+00 6.52489066e-01 -2.76398301e-01 1.45279932e+00
1.16112494e+00 5.78250885e-01 1.11905538e-01 -2.27121666e-01
1.54023957e+00 -7.25506306e-01 -7.74458528e-01 -5.25511466e-02
8.20105746e-02 -6.36060774e-01 1.20389068e+00 4.69627053e-01
-1.51202059e+00 -7.48160362e-01 -7.82449901e-01 -4.18062240e-01
2.85979986e-01 -1.29800618e-01 7.73117244e-01 3.52206886e-01
-5.57577848e-01 3.33846062e-01 -9.63290036e-01 -2.94924099e-02
3.60282332e-01 2.39998385e-01 -4.37683403e-01 -6.24671459e-01
-5.79090416e-01 6.35875344e-01 2.33341679e-02 5.24546266e-01
-1.02868474e+00 -1.13225472e+00 -8.41105282e-01 -2.77918607e-01
5.51495135e-01 -1.41916978e+00 1.01165533e+00 -4.90964174e-01
-1.60920680e+00 9.58108366e-01 -1.33855283e-01 2.11984292e-01
7.40285158e-01 -3.40608746e-01 -1.19881764e-01 2.75953472e-01
-2.88510323e-01 3.47387642e-01 7.90454626e-01 -1.90500343e+00
-4.18280602e-01 -6.07546329e-01 6.06039912e-02 7.56712437e-01
2.42694214e-01 -3.83555472e-01 -7.26384997e-01 -4.43344057e-01
6.57904983e-01 -6.10211790e-01 -3.61445487e-01 7.49883592e-01
-1.33460790e-01 2.56370485e-01 7.71681190e-01 -8.24463308e-01
6.99734271e-01 -1.93284249e+00 4.61535126e-01 -1.10044284e-02
7.47619510e-01 7.16171460e-03 5.59719512e-03 -8.14447403e-02
2.54588664e-01 -5.57336211e-01 -1.16422586e-01 -8.61906409e-01
-1.24715351e-01 3.69933128e-01 -2.31413975e-01 6.52741075e-01
-4.56283614e-02 1.06228840e+00 -9.62454259e-01 -4.73697037e-01
7.93338954e-01 1.17094064e+00 -7.09228098e-01 7.82713592e-01
-2.54558891e-01 1.16439903e+00 -7.02739894e-01 9.70480442e-01
1.03499138e+00 -1.54585332e-01 7.43983127e-03 -7.23984122e-01
7.24310055e-02 -2.77542830e-01 -1.34722137e+00 1.97486389e+00
-7.43830979e-01 7.62862861e-02 6.14471316e-01 -3.95777196e-01
6.85063720e-01 3.89102608e-01 6.61354184e-01 -1.04258561e+00
1.99856505e-01 1.56893268e-01 -5.58307350e-01 -6.08152628e-01
5.66928506e-01 -3.52027714e-01 2.27840006e-01 2.47696444e-01
-2.53391773e-01 -6.24145031e-01 -6.46667361e-01 3.90907526e-02
7.44105756e-01 4.94416445e-01 -9.00611840e-03 1.84432819e-01
4.20541793e-01 -2.48395592e-01 5.04975915e-01 3.43232363e-01
2.76264250e-02 1.02173209e+00 -2.37388238e-01 -8.46275568e-01
-1.20642447e+00 -1.39020693e+00 -3.54741642e-04 8.05408835e-01
7.04716444e-01 -2.70811524e-02 -3.89562458e-01 -6.15520775e-02
1.77819747e-02 4.48168486e-01 -5.53037643e-01 4.71572243e-02
-1.04565024e+00 -4.14465010e-01 -2.34944329e-01 3.70796412e-01
4.52916026e-01 -1.20695674e+00 -1.05392718e+00 3.01487923e-01
-4.00726497e-01 -1.36649835e+00 -2.13296428e-01 -1.72565728e-01
-9.64864969e-01 -1.08961296e+00 -9.27967370e-01 -4.57314491e-01
5.84790170e-01 4.84301716e-01 1.24295640e+00 2.50828534e-01
-6.91336513e-01 7.50955462e-01 -5.09281643e-02 1.27722204e-01
2.68327910e-03 -6.62709773e-01 -2.78945118e-02 2.58672148e-01
-4.98294353e-01 -8.16274643e-01 -9.87133801e-01 5.23761630e-01
-8.84534836e-01 6.31799221e-01 1.90783814e-01 6.00439787e-01
1.00527477e+00 -5.86687438e-02 -1.53851464e-01 -5.94152808e-01
7.65817091e-02 -1.46872595e-01 -4.50094312e-01 3.37139606e-01
1.72875330e-01 -5.63045979e-01 2.49021262e-01 -7.72049904e-01
-1.39222264e+00 2.27453172e-01 -4.88958992e-02 -1.13476467e+00
-2.26449832e-01 -1.67314738e-01 -3.98768634e-01 -1.77221090e-01
3.44863415e-01 3.34422290e-01 -3.59490216e-01 -5.09058177e-01
4.39278394e-01 2.28726752e-02 7.67536223e-01 -1.00554574e+00
8.57599556e-01 8.01279128e-01 -7.71589801e-02 -7.21644223e-01
-6.22896850e-01 -4.21040505e-01 -6.71390235e-01 -6.76084220e-01
1.15005350e+00 -1.16034007e+00 -1.28000951e+00 5.92029214e-01
-1.23811364e+00 -6.81818426e-01 -6.47478759e-01 5.74202240e-01
-9.10130858e-01 1.99343815e-01 -6.38743818e-01 -1.11412477e+00
-4.18645293e-01 -1.18558204e+00 1.76837766e+00 5.85373305e-02
6.05211891e-02 -6.61660373e-01 -1.82626575e-01 5.42001724e-01
2.98829794e-01 7.81925857e-01 4.11490828e-01 5.43890715e-01
-1.29606843e+00 1.46404300e-02 -3.80516768e-01 -2.11343870e-01
5.03290594e-02 -3.91428262e-01 -9.99473929e-01 -2.40666673e-01
2.13489592e-01 -3.54055136e-01 2.91579247e-01 5.60686707e-01
1.20966768e+00 -5.78201711e-02 -1.90806940e-01 1.03939724e+00
1.42077589e+00 -8.83171335e-02 5.59887230e-01 -2.32272595e-01
1.48237300e+00 7.54436314e-01 8.18618834e-01 7.16261387e-01
5.86636424e-01 1.03981698e+00 4.72125351e-01 -3.57931614e-01
-4.62497830e-01 -3.23326945e-01 -1.90197602e-01 8.71193171e-01
-5.57020426e-01 -2.87709504e-01 -6.26397252e-01 1.52180076e-01
-1.75629950e+00 -6.35213077e-01 -3.61489654e-01 2.24525952e+00
5.66220522e-01 -1.07624345e-01 2.33387407e-02 -4.76672947e-02
2.54110664e-01 -2.10309476e-02 -7.62546122e-01 2.14948654e-01
-3.20124552e-02 2.22085252e-01 1.81524485e-01 5.14098883e-01
-4.09165800e-01 8.81207705e-01 4.93697166e+00 7.96906590e-01
-9.31175947e-01 5.20294130e-01 5.30191422e-01 -4.59727943e-01
-6.26606643e-01 -3.44948292e-01 -4.15695071e-01 -3.02521382e-02
2.14844897e-01 1.97723567e-01 4.69870508e-01 5.53970158e-01
3.00681323e-01 -2.50462055e-01 -9.46796536e-01 1.44040883e+00
-6.16428293e-02 -1.13901818e+00 1.67733785e-02 1.68401182e-01
6.85719967e-01 -4.61496770e-01 -8.75791442e-03 -3.92791629e-02
1.88379019e-01 -7.73791373e-01 1.14304543e+00 1.01302230e+00
1.07365394e+00 -5.94979048e-01 1.78000584e-01 3.47314566e-01
-1.58327055e+00 1.94373786e-01 -2.45989576e-01 1.30251303e-01
7.87064672e-01 7.43438125e-01 -2.26270586e-01 7.74011731e-01
9.35162127e-01 6.43762410e-01 -1.01792425e-01 8.58996034e-01
2.12087750e-01 -1.13388024e-01 -2.69170284e-01 5.25032878e-01
-2.68708676e-01 -2.67526805e-01 5.44700444e-01 7.59600937e-01
2.20893309e-01 9.39381778e-01 5.78081131e-01 1.08025837e+00
4.45341051e-01 -6.28152341e-02 -4.05335248e-01 3.76954705e-01
1.33067489e-01 1.25417316e+00 -7.75641143e-01 -5.15927494e-01
1.23275165e-02 1.30821586e+00 3.59654367e-01 6.65109396e-01
-1.00591600e+00 3.12579453e-01 6.41981721e-01 6.88583732e-01
1.68790191e-01 -5.88718891e-01 -3.34220350e-01 -1.40051830e+00
3.32200259e-01 -5.14031529e-01 -3.28006782e-02 -1.22308147e+00
-1.10214925e+00 8.03330719e-01 1.28153086e-01 -1.36152744e+00
1.30419716e-01 -2.22806677e-01 -2.14415640e-01 9.24014091e-01
-1.28912938e+00 -1.65462995e+00 -9.04671073e-01 1.05466890e+00
7.04354405e-01 4.34210151e-01 5.21669149e-01 4.60846990e-01
-2.00842664e-01 1.81556493e-01 -6.05130136e-01 -5.88148355e-01
3.49341780e-01 -8.82770002e-01 2.02694699e-01 3.29890549e-01
-2.95111895e-01 3.92365336e-01 6.04935467e-01 -9.31493342e-01
-2.00165606e+00 -9.82802391e-01 1.38589382e-01 -5.94331741e-01
-6.95222244e-02 -6.56823993e-01 -1.04793894e+00 4.61362839e-01
-3.25298131e-01 5.30145407e-01 4.15693559e-02 -2.74982482e-01
-2.30116799e-01 -8.15510303e-02 -1.48788619e+00 7.38050342e-01
1.50974822e+00 -4.94412571e-01 -1.09724134e-01 3.70837033e-01
8.97398949e-01 -1.19720793e+00 -9.78530049e-01 6.80230379e-01
9.42881465e-01 -1.06326914e+00 1.51661336e+00 -1.17851809e-01
2.98971325e-01 -6.22680962e-01 -3.64852369e-01 -7.34539688e-01
-2.41519302e-01 -4.76785213e-01 -6.12485707e-01 7.11997569e-01
-5.57514966e-01 -3.46695900e-01 8.11334431e-01 8.17912817e-01
-2.22413689e-01 -9.28392768e-01 -9.44700241e-01 -4.23799723e-01
-5.01393616e-01 -6.67452633e-01 6.10861897e-01 8.29308629e-01
-7.03889489e-01 -1.77874461e-01 -7.79569268e-01 3.35295409e-01
1.02678668e+00 2.43100643e-01 1.13670468e+00 -1.13336039e+00
-5.21384835e-01 -1.14745185e-01 -2.89449185e-01 -1.20144999e+00
-1.39271602e-01 -3.69707286e-01 2.83443958e-01 -1.60253859e+00
1.00836404e-01 -7.10620046e-01 2.16741979e-01 -3.41796950e-02
-1.67343557e-01 6.90515399e-01 2.34127894e-01 1.76724061e-01
-5.06539464e-01 8.92243505e-01 2.02160025e+00 2.45729923e-01
-3.38022590e-01 -1.76469862e-01 -8.21377486e-02 6.89191222e-01
6.42770454e-02 -2.30347618e-01 -4.90045786e-01 -6.05722129e-01
2.37817392e-02 6.94389105e-01 9.93535101e-01 -1.04179800e+00
7.81926513e-02 -3.86764139e-01 7.66507268e-01 -8.61157000e-01
1.06607997e+00 -1.15537560e+00 9.42582011e-01 2.59557933e-01
3.68457139e-02 -1.26858354e-01 2.27112815e-01 7.53911436e-01
2.19914123e-01 6.88303769e-01 7.35726476e-01 -3.93447876e-01
-4.47407722e-01 9.35268283e-01 2.15367064e-01 -1.71599895e-01
9.28297400e-01 -4.68258083e-01 2.42030889e-01 -3.27666134e-01
-1.04936576e+00 1.62902713e-01 6.89374745e-01 3.93166542e-01
1.22255111e+00 -1.46130311e+00 -7.08803058e-01 5.35785794e-01
7.93317780e-02 5.39322853e-01 1.10704458e+00 7.99260199e-01
-7.23511815e-01 -1.17895581e-01 -2.18205258e-01 -9.27109957e-01
-1.20631039e+00 3.83534253e-01 4.55447316e-01 -2.15118334e-01
-1.25139630e+00 8.10731769e-01 8.95513654e-01 -6.44219398e-01
2.82006651e-01 -4.56687748e-01 3.17926675e-01 -5.72387457e-01
6.22879088e-01 3.36529195e-01 -1.10336401e-01 -8.15619588e-01
-3.27533819e-02 1.08308101e+00 3.58024627e-01 -1.76001623e-01
1.40821695e+00 -3.05051535e-01 1.28082827e-01 5.96512437e-01
9.87587571e-01 2.66457982e-02 -1.63753939e+00 -3.29845130e-01
-9.03285503e-01 -1.09804726e+00 1.17337583e-02 -6.65367723e-01
-1.34572852e+00 9.85786200e-01 6.71041369e-01 -4.10820246e-01
1.19202554e+00 -4.79501076e-02 1.08568299e+00 -6.35420308e-02
8.31643403e-01 -4.76733893e-01 1.28724992e-01 7.61549994e-02
1.22444069e+00 -9.59805906e-01 2.76997149e-01 -9.85514343e-01
-5.39067626e-01 8.71764123e-01 9.56767440e-01 -1.48078158e-01
5.03904700e-01 1.28453538e-01 -2.10269541e-01 -5.45348644e-01
-6.28195643e-01 -1.87896919e-02 5.79976141e-01 7.10756004e-01
1.44116864e-01 1.64086372e-01 2.99027562e-01 2.82756597e-01
1.37720238e-02 1.09592773e-01 7.60990977e-02 8.54802251e-01
-1.82596490e-01 -5.37211299e-01 -4.73026574e-01 2.05651999e-01
1.21201135e-01 1.26959100e-01 1.84284449e-01 5.87648690e-01
3.07171285e-01 4.14398462e-01 1.28260940e-01 -2.58047432e-01
7.71535754e-01 -4.09799606e-01 1.09082961e+00 -5.42767584e-01
-6.20462179e-01 4.72230256e-01 -1.23483248e-01 -1.22308385e+00
-4.13473636e-01 -3.33479613e-01 -1.21262538e+00 -5.03802419e-01
-7.49770254e-02 -4.62458760e-01 5.91212213e-01 6.16816640e-01
2.10924760e-01 7.69829571e-01 3.59303415e-01 -1.73907554e+00
1.18506826e-01 -4.56052870e-01 -8.71836126e-01 6.24311268e-01
4.20231491e-01 -8.72833073e-01 -1.71525761e-01 9.92716104e-03]
|
[7.216442584991455, -1.3651942014694214]
|
4f4f4499-9860-48a8-9594-e3063e78d99e
|
adapting-self-supervised-models-to-multi
|
2211.00482
| null |
https://arxiv.org/abs/2211.00482v1
|
https://arxiv.org/pdf/2211.00482v1.pdf
|
Adapting self-supervised models to multi-talker speech recognition using speaker embeddings
|
Self-supervised learning (SSL) methods which learn representations of data without explicit supervision have gained popularity in speech-processing tasks, particularly for single-talker applications. However, these models often have degraded performance for multi-talker scenarios -- possibly due to the domain mismatch -- which severely limits their use for such applications. In this paper, we investigate the adaptation of upstream SSL models to the multi-talker automatic speech recognition (ASR) task under two conditions. First, when segmented utterances are given, we show that adding a target speaker extraction (TSE) module based on enrollment embeddings is complementary to mixture-aware pre-training. Second, for unsegmented mixtures, we propose a novel joint speaker modeling (JSM) approach, which aggregates information from all speakers in the mixture through their embeddings. With controlled experiments on Libri2Mix, we show that using speaker embeddings provides relative WER improvements of 9.1% and 42.1% over strong baselines for the segmented and unsegmented cases, respectively. We also demonstrate the effectiveness of our models for real conversational mixtures through experiments on the AMI dataset.
|
['Sanjeev Khudanpur', 'Paola García', 'Desh Raj', 'Zili Huang']
|
2022-11-01
| null | null | null | null |
['target-speaker-extraction']
|
['audio']
|
[ 2.95419782e-01 3.77624393e-01 -1.36631787e-01 -7.95427024e-01
-1.52806938e+00 -2.97753155e-01 6.11849606e-01 -6.03783913e-02
-4.22548831e-01 4.28245097e-01 7.83359826e-01 -3.49139750e-01
2.74156451e-01 1.00386404e-01 -4.70890850e-01 -7.46279597e-01
1.31031021e-01 5.27915835e-01 -1.26636773e-01 -1.25788063e-01
-4.60024089e-01 2.58313924e-01 -1.31525457e+00 5.02439320e-01
8.31468701e-01 4.87869382e-01 2.60915279e-01 8.08368087e-01
-2.15411931e-01 7.21656621e-01 -9.69373345e-01 -2.74365395e-01
2.11396776e-02 -3.14808458e-01 -6.67452335e-01 5.08492470e-01
4.08074498e-01 -3.29907626e-01 -5.54611504e-01 7.11362660e-01
7.58737326e-01 1.33429304e-01 7.00463057e-01 -1.08791018e+00
-2.34988958e-01 1.25577021e+00 -4.86258179e-01 1.87236145e-01
9.25175250e-02 -1.11712921e-04 1.00038338e+00 -1.05398679e+00
-1.05838664e-03 1.67578876e+00 3.35263073e-01 9.05920982e-01
-1.37874222e+00 -6.14988983e-01 3.39101106e-01 9.11143050e-02
-1.38371217e+00 -1.47778034e+00 7.97468841e-01 -2.14185670e-01
1.13353074e+00 4.09266979e-01 -1.45583317e-01 1.32730532e+00
-5.35923123e-01 1.24578738e+00 7.69371092e-01 -5.79988301e-01
1.65964395e-01 5.36275208e-01 3.39952260e-01 1.30040139e-01
-1.24131739e-01 -3.30874801e-01 -7.01860666e-01 -1.40897930e-01
1.66683137e-01 -3.06385845e-01 -4.03523058e-01 1.08994380e-01
-1.13369858e+00 7.51203060e-01 -9.73854661e-02 4.62710261e-01
-3.52746010e-01 -1.56242013e-01 3.10220808e-01 2.80144304e-01
8.49451542e-01 7.96161219e-02 -3.55944723e-01 -2.47734204e-01
-1.21283197e+00 -2.95824409e-02 1.05353856e+00 9.27160323e-01
4.71680433e-01 3.81549627e-01 -1.96705207e-01 1.65680540e+00
4.97111529e-01 4.76925761e-01 8.24877322e-01 -6.77425027e-01
9.53945041e-01 1.59409657e-01 -1.68120302e-02 -1.02553621e-01
-7.72338212e-02 -4.31818902e-01 -6.83142424e-01 -3.02898198e-01
1.83790773e-01 -3.12897593e-01 -1.00372314e+00 1.89689219e+00
2.03225106e-01 1.86465412e-01 6.92421377e-01 7.07747877e-01
8.34637940e-01 9.21354473e-01 6.55560791e-02 -5.13341725e-01
1.13925517e+00 -1.22196889e+00 -1.08350492e+00 -6.15196466e-01
6.64378703e-01 -8.80492449e-01 9.16597545e-01 2.93975919e-01
-1.15223503e+00 -4.61686730e-01 -1.05963147e+00 2.47048467e-01
7.41161825e-03 3.30108374e-01 -9.20483172e-02 1.06745410e+00
-1.02805364e+00 2.37094998e-01 -8.66021276e-01 -4.42159504e-01
1.41613021e-01 4.62131530e-01 -2.04807758e-01 -2.30627432e-01
-1.01923776e+00 8.03655505e-01 5.56168221e-02 -6.67615756e-02
-9.45691884e-01 -5.14108479e-01 -9.54594970e-01 1.78564772e-01
2.22220257e-01 -1.54419437e-01 1.67412329e+00 -7.98244476e-01
-1.91873527e+00 6.28559172e-01 -7.55418181e-01 -6.31147981e-01
2.63086438e-01 -3.73965651e-01 -8.24690282e-01 6.70512468e-02
-2.65173018e-01 5.99277794e-01 9.71138775e-01 -1.36379635e+00
-4.33341950e-01 -3.03387314e-01 -2.93514609e-01 5.61027706e-01
-5.42158782e-01 2.59364814e-01 -2.52349854e-01 -6.18485272e-01
1.83311313e-01 -8.34209025e-01 -1.80957094e-01 -7.70865381e-01
-5.28371632e-01 -3.54005963e-01 1.01505983e+00 -9.47145402e-01
1.20262706e+00 -2.48928547e+00 -1.32672982e-02 -4.27290276e-02
-7.11833462e-02 6.29580259e-01 -1.97069526e-01 3.67756605e-01
-1.71104193e-01 -7.52463937e-03 -3.41936886e-01 -1.14019597e+00
1.47127360e-01 2.26883233e-01 -3.24948072e-01 2.20982790e-01
2.67386824e-01 5.46839952e-01 -3.93173635e-01 -3.43003243e-01
2.04359695e-01 5.06334186e-01 -4.23727870e-01 6.63930357e-01
-1.55686715e-03 3.25148106e-01 1.59703389e-01 4.41926390e-01
6.16147757e-01 2.25211248e-01 3.34989280e-01 -1.34325579e-01
7.72782117e-02 8.42405617e-01 -1.20145595e+00 1.57855678e+00
-8.04132819e-01 5.85639417e-01 7.45354891e-01 -1.02553082e+00
7.86356091e-01 8.24790180e-01 1.95429668e-01 -1.55813321e-01
-3.90791669e-02 2.43407890e-01 2.47844726e-01 -3.54202360e-01
3.81707042e-01 -4.04657066e-01 -2.42266990e-02 5.37553132e-01
3.65423262e-01 -7.51530677e-02 -1.30543113e-01 2.34016344e-01
1.04050314e+00 -6.17454052e-01 2.04788595e-01 -1.26752436e-01
4.89255250e-01 -4.57488686e-01 4.92834926e-01 6.98028684e-01
-4.17201728e-01 7.16827512e-01 -6.47718413e-03 2.91605085e-01
-7.00125217e-01 -1.21480048e+00 3.50045897e-02 1.27671504e+00
-3.11560690e-01 -4.13605392e-01 -1.00176394e+00 -4.60943341e-01
-2.74007320e-01 1.06151319e+00 1.54844970e-01 -1.21738382e-01
-7.82465875e-01 -9.50125635e-01 6.43257737e-01 4.76354182e-01
2.49909759e-01 -6.55481577e-01 2.01128811e-01 5.83762527e-01
-4.95724589e-01 -1.59343517e+00 -7.52370715e-01 3.14531356e-01
-7.17634857e-01 -3.36880207e-01 -8.14899385e-01 -8.46640527e-01
2.73767889e-01 4.49948311e-01 6.73182130e-01 -5.72143614e-01
-9.94789298e-04 5.84183812e-01 -2.90626943e-01 -2.78407812e-01
-1.08686554e+00 1.55933172e-01 5.61549366e-01 4.50307310e-01
5.12124538e-01 -5.72964251e-01 -2.56312907e-01 4.36418355e-01
-6.50326431e-01 -3.71055678e-02 7.08977818e-01 6.63418353e-01
-3.48238498e-02 -1.84859410e-01 1.03713441e+00 -6.49004698e-01
7.43771732e-01 -3.99658799e-01 5.58073521e-02 3.43946666e-01
-2.70873904e-01 1.13470115e-01 3.03758800e-01 -7.75836051e-01
-1.34326804e+00 4.95592877e-02 -5.49652040e-01 -3.86686325e-01
-5.87922573e-01 2.10409597e-01 -8.66399825e-01 5.88789105e-01
5.62934518e-01 3.09570253e-01 2.50393748e-02 -8.49065185e-01
5.05746484e-01 1.75098884e+00 4.90887791e-01 -4.64308113e-01
5.24642348e-01 1.01607181e-01 -9.43525136e-01 -1.57169878e+00
-7.08703160e-01 -8.65191400e-01 -3.75702620e-01 1.37738600e-01
7.30605006e-01 -1.45206809e+00 -1.76093414e-01 4.09113228e-01
-1.17119575e+00 -3.72810334e-01 -2.66350538e-01 9.16190028e-01
-5.46569347e-01 4.07172084e-01 -7.46187806e-01 -1.37398732e+00
-3.04624856e-01 -1.42162597e+00 9.28444147e-01 -8.04599747e-02
-5.03223360e-01 -7.89515138e-01 -5.63083328e-02 8.56251419e-01
5.16581833e-01 -7.45110452e-01 7.36158788e-01 -1.37331247e+00
-1.54557273e-01 -1.39479060e-02 1.65074915e-01 8.78939390e-01
4.99761820e-01 -4.63214129e-01 -1.74611914e+00 -4.55116361e-01
2.42628500e-01 -3.09583306e-01 9.00965214e-01 3.63858014e-01
8.18987012e-01 -5.59962034e-01 -3.11851144e-01 9.79522318e-02
6.83476210e-01 1.86438173e-01 2.06070468e-01 -3.61047238e-01
7.18422294e-01 7.45480537e-01 1.19263969e-01 1.97708145e-01
4.02862132e-01 7.29569435e-01 -8.48154947e-02 4.35227230e-02
-5.27565956e-01 -1.67682469e-01 9.59939718e-01 1.58731925e+00
5.04942954e-01 -5.40468335e-01 -7.39204347e-01 7.82257736e-01
-1.64230072e+00 -8.07386696e-01 1.20782711e-01 2.41666842e+00
1.02768791e+00 4.48943488e-02 3.28831255e-01 3.68520707e-01
1.01349008e+00 3.06599796e-01 -3.84489268e-01 -3.69983524e-01
-1.01803377e-01 7.45624229e-02 2.30783418e-01 8.07123303e-01
-1.10029328e+00 8.86876464e-01 6.01445055e+00 9.98563290e-01
-8.52537096e-01 5.99962294e-01 6.17250502e-01 -3.19897622e-01
-1.34581462e-01 -3.33902508e-01 -1.11816645e+00 2.36441016e-01
1.67604637e+00 -4.25958447e-02 3.68644238e-01 6.79852903e-01
3.64198536e-01 1.93345204e-01 -1.40081978e+00 1.23881447e+00
3.46183091e-01 -7.74173617e-01 -1.25502691e-01 -1.49694672e-02
4.53241229e-01 1.98651731e-01 7.19097629e-02 5.88180602e-01
3.53189617e-01 -8.16818357e-01 5.81635714e-01 -2.36278474e-01
6.25182450e-01 -6.11784101e-01 5.97352564e-01 5.87690651e-01
-9.59710777e-01 -1.53279439e-01 -1.44250080e-01 3.22174877e-01
4.24675524e-01 6.99382901e-01 -1.47434247e+00 3.59189719e-01
2.61270195e-01 1.99651062e-01 -1.14226855e-01 7.53134549e-01
-1.60873771e-01 1.26862657e+00 -4.49205458e-01 2.65683949e-01
-4.79066832e-04 1.42681241e-01 7.84171820e-01 1.66830838e+00
2.27774769e-01 -1.45002738e-01 3.90369296e-02 7.09911346e-01
-3.03252935e-01 1.56364545e-01 -3.96367610e-01 -2.62031823e-01
6.57608092e-01 9.97745395e-01 -8.83704796e-02 -5.26685417e-01
-3.85285616e-01 1.13624537e+00 2.78963953e-01 6.08071864e-01
-5.46711981e-01 -2.65018553e-01 1.01997697e+00 1.32878244e-01
2.57073849e-01 -2.99380720e-01 -5.40261008e-02 -1.21635258e+00
4.98375781e-02 -1.00048220e+00 1.04996331e-01 -2.27094725e-01
-1.35630202e+00 7.42967308e-01 9.59133133e-02 -9.94105279e-01
-6.03632569e-01 -2.86483675e-01 -6.96796000e-01 1.03403342e+00
-1.42561162e+00 -1.02811575e+00 3.03731620e-01 3.21912318e-01
1.24161708e+00 -4.29398030e-01 8.89964461e-01 4.38132137e-01
-8.52086008e-01 8.92885983e-01 2.74186581e-01 1.40870914e-01
8.52648556e-01 -1.08868420e+00 5.66816926e-01 7.80265689e-01
4.29795831e-01 5.21709442e-01 7.81321287e-01 -1.82675704e-01
-1.13297176e+00 -1.24871993e+00 1.08994174e+00 -2.59846151e-01
3.87888789e-01 -8.51641417e-01 -1.08021057e+00 7.91519463e-01
4.17031139e-01 -3.31370026e-01 1.05404055e+00 2.54062265e-01
-3.51742744e-01 -2.15722278e-01 -1.02742445e+00 5.47315300e-01
7.48639524e-01 -6.81937754e-01 -9.17062283e-01 2.19602212e-01
1.11070979e+00 -3.43835168e-02 -7.09499061e-01 1.94392949e-01
1.05768450e-01 -5.85332215e-01 8.58084381e-01 -4.42682236e-01
-2.37674534e-01 -3.66893187e-02 -6.70784771e-01 -1.74489570e+00
9.77831185e-02 -8.21020961e-01 -2.71549463e-01 1.74791169e+00
6.26271546e-01 -5.31075895e-01 4.80571628e-01 6.76082611e-01
-3.38684142e-01 -1.37509003e-01 -1.15336490e+00 -9.82058048e-01
-1.00483457e-02 -6.24328196e-01 4.96822894e-01 6.71532691e-01
2.69252926e-01 7.17841804e-01 -3.74555796e-01 5.49929500e-01
6.18646324e-01 -2.74485707e-01 7.36918151e-01 -7.98203230e-01
-4.79420334e-01 -2.41685301e-01 -1.80499151e-01 -1.45912063e+00
5.15565515e-01 -8.31210315e-01 5.54509103e-01 -1.35642457e+00
-3.75189334e-02 -2.81319469e-01 -2.63317347e-01 2.32211754e-01
-2.27027729e-01 -2.97236532e-01 1.22495808e-01 2.15807974e-01
-4.44127500e-01 7.82267869e-01 4.11071658e-01 -3.40566993e-01
-4.50102180e-01 3.17366838e-01 -6.62312806e-01 6.58369660e-01
9.59787667e-01 -4.59695131e-01 -4.74095643e-01 -4.24672544e-01
-9.53820884e-01 1.88184917e-01 -1.96383312e-01 -1.12615848e+00
4.81047332e-02 1.61020204e-01 -1.64483041e-01 -3.63908976e-01
9.55599308e-01 -5.04804432e-01 -2.68757820e-01 4.84870039e-02
-6.39660776e-01 -4.80360568e-01 2.43610919e-01 5.74604750e-01
-3.16727668e-01 -3.19165468e-01 7.87933648e-01 8.74925628e-02
-1.90448269e-01 -6.53885826e-02 -7.09631026e-01 8.44583288e-02
5.89300215e-01 2.08079666e-01 -1.14229947e-01 -7.66807973e-01
-9.94159818e-01 -8.50604661e-03 -3.60988304e-02 4.66152221e-01
5.29280484e-01 -1.13510764e+00 -9.76791203e-01 3.20186704e-01
2.14190707e-01 -1.27582559e-02 3.06606740e-01 7.43979812e-01
2.65593767e-01 4.71789092e-01 6.37222767e-01 -6.47424519e-01
-1.52799022e+00 2.82138884e-01 2.17763528e-01 -8.25907886e-02
-3.18539828e-01 9.43639576e-01 4.41574156e-01 -5.95812380e-01
6.84514284e-01 -4.32479262e-01 3.17974240e-02 -8.83360282e-02
7.23758936e-01 3.42606813e-01 3.71595383e-01 -8.42558861e-01
-4.20431495e-01 -5.39830513e-02 -4.49532449e-01 -6.90958023e-01
1.18872428e+00 -3.98896009e-01 4.51901078e-01 8.73229265e-01
1.27635539e+00 1.14827842e-01 -1.13236713e+00 -6.31127238e-01
5.41316010e-02 -5.78474291e-02 1.14166699e-01 -5.98918498e-01
-6.30228996e-01 1.28172541e+00 4.87100154e-01 1.72288164e-01
6.78312242e-01 3.21617931e-01 8.17331135e-01 4.04393822e-01
-5.85660748e-02 -1.06336927e+00 -1.12443738e-01 2.92519510e-01
8.33368182e-01 -1.31141865e+00 -5.10577738e-01 -4.41183329e-01
-9.14652824e-01 7.95643151e-01 4.33584511e-01 4.80717212e-01
6.75929487e-01 3.99385780e-01 3.92078549e-01 3.87910724e-01
-9.55256879e-01 -2.11536095e-01 1.07895762e-01 5.76759815e-01
5.48613310e-01 4.02284235e-01 2.63397962e-01 8.73269141e-01
-2.08855823e-01 -5.02265871e-01 5.81769347e-01 8.19320261e-01
-6.74212277e-01 -1.21609998e+00 -6.07833683e-01 3.31612915e-01
-3.25153828e-01 -1.76055565e-01 -2.65735149e-01 2.50442743e-01
-4.10718650e-01 1.67968404e+00 1.30574524e-01 -3.59705478e-01
3.80502075e-01 7.29371548e-01 2.65874743e-01 -1.00667882e+00
-3.76897722e-01 6.34873688e-01 4.00635183e-01 -5.37927076e-02
-3.76560569e-01 -7.51500845e-01 -9.67599571e-01 9.97017473e-02
-3.96659106e-01 3.68660122e-01 9.93505001e-01 1.01087260e+00
3.58655393e-01 6.03129148e-01 7.44118869e-01 -8.49354208e-01
-1.01923609e+00 -1.59655058e+00 -6.96650445e-01 2.68368453e-01
5.36811411e-01 -2.94612050e-01 -7.60489404e-01 1.98257953e-01]
|
[14.58682918548584, 6.37366247177124]
|
6363e596-c3f7-4355-a230-76a12db50dde
|
statistical-nlg-for-generating-the-content
| null | null |
https://aclanthology.org/W18-6561
|
https://aclanthology.org/W18-6561.pdf
|
Statistical NLG for Generating the Content and Form of Referring Expressions
|
This paper argues that a new generic approach to statistical NLG can be made to perform Referring Expression Generation (REG) successfully. The model does not only select attributes and values for referring to a target referent, but also performs Linguistic Realisation, generating an actual Noun Phrase. Our evaluations suggest that the attribute selection aspect of the algorithm exceeds classic REG algorithms, while the Noun Phrases generated are as similar to those in a previously developed corpus as were Noun Phrases produced by a new set of human speakers.
|
['Kees Van Deemter', 'Xiao Li', 'Chenghua Lin']
|
2018-11-01
| null | null | null |
ws-2018-11
|
['referring-expression-generation']
|
['computer-vision']
|
[ 3.10345948e-01 9.15496826e-01 -2.18871728e-01 -7.88282335e-01
-1.30841911e+00 -8.24763834e-01 1.13695145e+00 1.90649271e-01
-4.36726928e-01 1.36496997e+00 6.95552230e-01 -1.81618035e-01
7.94534981e-02 -1.00255966e+00 -4.28202778e-01 -4.55619663e-01
1.83989421e-01 1.09230411e+00 6.10244237e-02 -5.37227392e-01
2.28372410e-01 7.57375538e-01 -1.57186091e+00 1.76107809e-01
6.38668537e-01 4.64671373e-01 -2.28303999e-01 2.23446727e-01
-7.49397635e-01 5.47555029e-01 -1.29995227e+00 -7.30708301e-01
-4.37371358e-02 -6.44024014e-01 -9.81328309e-01 1.32344142e-01
1.28719985e-01 2.24199176e-01 4.40605313e-01 9.79812860e-01
5.45666635e-01 4.00514960e-01 6.34677231e-01 -9.72497344e-01
-7.84489036e-01 1.30673909e+00 1.09029375e-01 3.59438024e-02
1.35596120e+00 -2.34807119e-01 1.21108174e+00 -6.97148323e-01
9.15945351e-01 1.75892425e+00 2.81003773e-01 8.67864907e-01
-1.42131460e+00 -6.13806725e-01 5.10751717e-02 -4.36810225e-01
-1.65841520e+00 -5.63846231e-01 6.37512565e-01 -8.18137918e-03
1.07182407e+00 6.20443165e-01 7.89968073e-01 1.14338756e+00
-1.42904907e-03 6.31006062e-01 1.13586831e+00 -1.02398407e+00
3.26729089e-01 4.44507480e-01 -6.63751215e-02 -6.11556636e-04
2.11240143e-01 2.12956235e-01 -6.38990223e-01 -5.39414883e-01
4.80976939e-01 -1.06803632e+00 -1.88504592e-01 1.78177744e-01
-1.34338713e+00 1.05126143e+00 -1.05738387e-01 7.65615106e-01
-6.78595662e-01 2.19519287e-01 3.73778313e-01 1.65231079e-01
3.86814386e-01 9.34052706e-01 -5.41948915e-01 -2.44089410e-01
-5.92166841e-01 9.31041479e-01 1.22579908e+00 1.50677443e+00
5.49402714e-01 1.96661890e-01 -1.95971966e-01 5.62822163e-01
4.47643995e-01 4.80441004e-01 4.59330857e-01 -1.17134082e+00
3.19041103e-01 5.00476658e-01 3.36151302e-01 -7.14170158e-01
-5.13508767e-02 -1.25075327e-04 1.34883612e-01 -3.14790398e-01
4.04862195e-01 -2.34472185e-01 -5.74075878e-01 1.91757417e+00
4.20734674e-01 -7.78163791e-01 5.44210196e-01 4.24316257e-01
1.03071010e+00 8.24160516e-01 8.86518419e-01 -5.58897674e-01
1.51712358e+00 2.25293003e-02 -1.18938792e+00 -2.38687038e-01
9.85495090e-01 -7.62381732e-01 1.22303653e+00 1.42560720e-01
-1.31106126e+00 -3.02722216e-01 -7.27039993e-01 -1.59357086e-01
-4.73404467e-01 -3.78187060e-01 1.17552412e+00 9.30107057e-01
-1.06453347e+00 1.29565507e-01 -1.16846129e-01 -5.83399832e-01
4.43740655e-03 5.15095353e-01 -3.47858429e-01 4.28052455e-01
-1.43266582e+00 8.98669243e-01 9.29768682e-01 -2.89923489e-01
2.74997037e-02 -3.61617863e-01 -9.97746408e-01 -3.85742188e-01
2.16601878e-01 -7.09220409e-01 1.37152827e+00 -1.02664137e+00
-1.31514776e+00 1.43111384e+00 -5.91100752e-01 -2.98267365e-01
2.82570533e-02 -1.29352808e-01 -7.37675309e-01 -1.29149467e-01
6.15810335e-01 9.37760770e-01 3.33823711e-01 -1.75371468e+00
-8.44321072e-01 -3.03136528e-01 1.03786215e-01 3.46943915e-01
4.71657932e-01 8.56202304e-01 -5.83482487e-03 -8.73577833e-01
3.48957777e-01 -5.60889304e-01 -1.13260053e-01 -7.38873780e-01
-7.41443276e-01 -9.72130299e-01 1.85657233e-01 -8.75073969e-02
1.32516885e+00 -1.81261945e+00 -7.70200565e-02 5.78155696e-01
-1.95773914e-01 -2.04788625e-01 1.82860881e-01 5.22541583e-01
-4.57799464e-01 6.99731112e-01 -6.20515794e-02 1.72966078e-01
3.25857341e-01 6.12265766e-01 -6.74455106e-01 -5.25904708e-02
3.03166717e-01 9.84530747e-01 -8.21095407e-01 -8.93852234e-01
-1.57075703e-01 1.73387364e-01 -2.24023834e-01 4.31367010e-02
-3.24670970e-01 5.90060614e-02 -8.65711272e-01 7.83871293e-01
1.57126233e-01 4.71264780e-01 1.81369543e-01 1.03563070e-01
-1.58975273e-01 8.72258842e-01 -1.13074172e+00 1.14331472e+00
-1.75822988e-01 2.00771734e-01 -3.93268615e-01 -4.86149549e-01
1.34702647e+00 6.91482484e-01 1.37267560e-01 -3.18527043e-01
2.82300264e-01 5.44504225e-01 7.82318860e-02 -4.44347858e-01
5.61932206e-01 -9.87934887e-01 -8.04049194e-01 6.28802180e-01
-1.99592635e-01 -9.56717908e-01 3.59427541e-01 5.41003123e-02
6.39734983e-01 6.13871753e-01 9.98296618e-01 -5.02945065e-01
7.48490751e-01 5.10143876e-01 4.96636122e-01 6.59141541e-01
1.72139332e-01 1.27462149e-01 4.77961183e-01 -4.38761652e-01
-8.29408646e-01 -1.08012843e+00 -5.81737280e-01 1.19310117e+00
-2.78861105e-01 -2.82280266e-01 -8.47384453e-01 -5.90341747e-01
-2.69334048e-01 1.94449270e+00 -5.71302235e-01 3.13587725e-01
-8.78695190e-01 -6.95506394e-01 7.62996614e-01 3.16754490e-01
-6.91990405e-02 -1.80526996e+00 -5.45281529e-01 7.50062525e-01
-3.35672677e-01 -1.05386662e+00 -1.13685563e-01 3.71927202e-01
-6.19780958e-01 -5.78227997e-01 -8.44840035e-02 -8.33497465e-01
6.28830016e-01 -5.90578675e-01 1.56049371e+00 -1.15580838e-02
2.74720937e-01 1.49116993e-01 -6.03163838e-01 -9.62372422e-01
-9.78875816e-01 -1.55083481e-02 -1.49353489e-01 -4.89387423e-01
1.26723289e+00 -4.37236667e-01 3.87096584e-01 -1.65733416e-02
-9.30265665e-01 -2.65890688e-01 2.30844095e-01 4.50353414e-01
7.01777756e-01 -5.86702786e-02 7.63759971e-01 -1.33865011e+00
1.25252903e+00 -1.98267952e-01 -2.23284096e-01 8.39712247e-02
-3.06211799e-01 2.69767284e-01 3.11838835e-01 -1.92643389e-01
-1.28378189e+00 1.86676443e-01 -4.50556725e-01 6.94410861e-01
-7.85883605e-01 3.71457607e-01 -6.75948799e-01 3.04051429e-01
9.62175131e-01 1.20926037e-01 -1.67983294e-01 -2.02544197e-01
5.96292078e-01 4.87885326e-01 5.46337247e-01 -1.31297171e+00
8.88817370e-01 4.55552414e-02 -3.16081755e-02 -5.91057301e-01
-8.45266521e-01 -2.92609446e-02 -8.33923936e-01 -3.50461863e-02
6.99093938e-01 -7.03441381e-01 -4.97870356e-01 -2.85393864e-01
-1.48358822e+00 8.17962661e-02 -1.08642650e+00 4.22213674e-01
-8.47144783e-01 -1.13547184e-01 -3.10880672e-02 -9.62624252e-01
-2.38173932e-01 -6.83369994e-01 1.37859774e+00 -1.15736157e-01
-1.23781800e+00 -1.07564747e+00 -1.47876725e-01 -2.32291818e-02
1.53213188e-01 4.23165530e-01 1.36862361e+00 -1.27683055e+00
-1.30387858e-01 -3.83703262e-01 5.36760986e-02 -1.73334554e-01
2.89601505e-01 2.22764164e-01 -6.78411186e-01 3.20003629e-01
1.40174016e-01 -2.36609846e-01 -1.30688146e-01 1.75189450e-01
2.73349345e-01 -6.46800756e-01 -3.84090841e-01 1.09426878e-01
1.19345284e+00 7.99242437e-01 7.25296676e-01 5.23593009e-01
1.96054801e-01 1.33850062e+00 8.84340048e-01 -4.67685834e-02
4.54115093e-01 7.61736751e-01 -1.38990283e-01 1.39987856e-01
1.11801073e-01 -2.36447811e-01 2.17218757e-01 3.87122512e-01
-9.38736647e-02 -4.58493263e-01 -9.99408305e-01 6.04497254e-01
-1.49746192e+00 -9.98102903e-01 -3.35127562e-01 1.77985573e+00
1.25568700e+00 9.22548175e-02 1.15967914e-01 1.08066805e-01
6.03250682e-01 -2.32630735e-03 1.37388706e-01 -9.16998208e-01
-5.56582451e-01 7.26073444e-01 9.40651670e-02 5.97674072e-01
-7.11681724e-01 1.50104642e+00 8.19002438e+00 6.77904427e-01
-3.90121639e-01 -2.79627293e-01 3.88292372e-01 2.67580986e-01
-1.02518165e+00 1.73117459e-01 -1.12600064e+00 -8.24065655e-02
1.02374363e+00 -8.69353414e-01 7.89813697e-02 6.96783602e-01
2.72587508e-01 -1.28623366e-01 -1.32581186e+00 5.50168633e-01
2.08273977e-01 -9.63616073e-01 6.72543168e-01 3.78321148e-02
2.44090945e-01 -7.88845539e-01 -4.82675225e-01 3.01294565e-01
4.80432630e-01 -1.14256954e+00 9.25522089e-01 4.00211722e-01
5.22253394e-01 -1.02475810e+00 9.97567177e-01 6.88896775e-02
-8.23228002e-01 3.64736736e-01 -2.18449846e-01 2.59718895e-02
4.33041871e-01 5.52284420e-02 -9.78419363e-01 3.23458821e-01
2.62278199e-01 -8.37413669e-02 -5.61620295e-01 5.13024509e-01
-7.29064107e-01 5.70586026e-01 -3.88018548e-01 -6.67968512e-01
2.96021461e-01 -2.31444255e-01 9.72986758e-01 1.35939491e+00
4.35918838e-01 4.55240577e-01 -1.54458955e-02 1.10840786e+00
2.16493040e-01 9.26477134e-01 -1.03952467e+00 1.14360685e-02
8.33173573e-01 8.92686367e-01 -5.41665494e-01 -6.32162929e-01
-2.59353817e-02 4.66176808e-01 -1.71829224e-01 3.76195967e-01
-3.99986982e-01 -4.91606295e-01 1.69827521e-01 2.39009693e-01
-1.80813953e-01 6.52257726e-02 -2.71300435e-01 -5.26364088e-01
-8.29113871e-02 -1.14192021e+00 3.03400785e-01 -1.03725183e+00
-1.28367686e+00 8.30083489e-01 6.54147744e-01 -7.32951403e-01
-1.06293011e+00 -4.47956890e-01 -4.71749008e-01 1.23625445e+00
-6.81851208e-01 -1.10759068e+00 3.71396899e-01 3.62204462e-01
3.36847842e-01 -3.36504757e-01 1.51674151e+00 -3.59606892e-01
-1.08421771e-02 3.53562951e-01 -9.01337326e-01 -1.03963865e-02
3.21283460e-01 -1.42564845e+00 5.20153284e-01 6.79128945e-01
3.52537483e-01 1.18357611e+00 1.31742525e+00 -7.32354164e-01
-9.95870292e-01 -7.21300781e-01 1.85522747e+00 -6.99586689e-01
6.69829428e-01 -1.62995011e-01 -1.00318301e+00 9.46493149e-01
3.66990745e-01 -4.69942719e-01 9.44004953e-01 2.84643203e-01
3.21506374e-02 2.86573082e-01 -1.42811775e+00 8.38988900e-01
9.64812458e-01 -3.37119401e-01 -1.39479625e+00 5.00906587e-01
8.28439713e-01 -2.96390355e-01 -7.58409262e-01 2.31698260e-01
2.51295447e-01 -6.36449158e-01 7.71264493e-01 -9.67316151e-01
-3.43705639e-02 -3.27173054e-01 -4.67822850e-01 -1.12036848e+00
7.56258741e-02 -8.64618123e-01 3.81278992e-01 1.70385897e+00
8.55515480e-01 -7.96632349e-01 6.22215331e-01 1.29013133e+00
-5.06285205e-02 -3.27515602e-01 -1.16102707e+00 -6.76834047e-01
3.05047721e-01 -8.92493010e-01 1.21637857e+00 7.77420819e-01
1.92517906e-01 7.04325736e-01 3.03900212e-01 -2.97724515e-01
4.96557027e-01 -5.46418168e-02 6.75477862e-01 -1.28695917e+00
1.87623799e-02 -3.85467380e-01 -5.20456016e-01 -6.29638314e-01
6.36427999e-01 -8.97705674e-01 7.95887113e-02 -1.39255941e+00
-4.05389547e-01 -5.20263314e-01 3.85452092e-01 6.31411552e-01
-3.96257751e-02 -3.48913968e-02 1.07654721e-01 5.43846563e-02
5.64770848e-02 4.19207811e-01 1.15797865e+00 3.57515156e-01
-6.06287599e-01 7.50111416e-02 -1.37306046e+00 8.96053374e-01
6.86374843e-01 -8.67177248e-01 -2.73231596e-01 7.82027096e-02
5.00570416e-01 9.09387320e-02 1.52443558e-01 -4.54063982e-01
-8.28456283e-02 -4.06214386e-01 3.20314169e-01 -4.80444998e-01
1.57916129e-01 -6.82303786e-01 4.80970621e-01 -7.24713644e-03
-5.31672776e-01 3.42510313e-01 3.05141717e-01 -1.34551376e-01
-4.37207669e-01 -6.96882188e-01 3.22472274e-01 -3.16949546e-01
-7.02266335e-01 -3.77388597e-01 -7.18327284e-01 2.46077448e-01
9.81237888e-01 -6.35870755e-01 -5.06226979e-02 -4.82748419e-01
-8.51487994e-01 -2.37559512e-01 2.34372124e-01 1.24540210e-01
5.50634444e-01 -1.53601754e+00 -7.71560848e-01 -2.12725475e-02
2.96720952e-01 2.68362872e-02 -7.32928753e-01 1.14001676e-01
-3.60693663e-01 5.20095110e-01 1.09442398e-02 3.95721979e-02
-9.46789742e-01 2.95704663e-01 1.08471297e-01 2.46258423e-01
-6.55516982e-01 1.02965891e+00 -7.64006227e-02 -5.17075896e-01
-1.09003792e-02 -1.08284809e-01 -5.39681077e-01 2.81055629e-01
3.53789002e-01 -3.14886481e-01 -2.83093244e-01 -1.33947003e+00
-3.29174459e-01 3.05434793e-01 2.87915263e-02 -6.49080634e-01
1.06440926e+00 -1.49111658e-01 -4.78420377e-01 8.94164741e-01
7.01841831e-01 6.44993842e-01 -9.95768532e-02 -1.88139319e-01
5.41733146e-01 -3.53448570e-01 -4.58535969e-01 -6.97631598e-01
-2.66476363e-01 5.12905009e-02 -1.24295555e-01 5.33588052e-01
9.08824503e-01 4.53397423e-01 4.66487169e-01 5.94931781e-01
7.57808447e-01 -1.19636774e+00 -6.25702500e-01 2.65188545e-01
1.34671354e+00 -6.38314188e-01 9.18808430e-02 -8.47355545e-01
-8.48059475e-01 1.23146355e+00 2.44091943e-01 3.88961136e-02
2.62473345e-01 3.13695520e-01 4.52782243e-01 -5.74981868e-01
-8.08802307e-01 -3.44677687e-01 1.97986230e-01 9.52856243e-01
1.17053747e+00 2.59916663e-01 -8.74709308e-01 4.85746324e-01
-1.14668047e+00 -3.24335009e-01 6.03381515e-01 8.36666226e-01
-3.15117866e-01 -1.55747831e+00 -5.58245718e-01 3.66451830e-01
-7.62773335e-01 -3.51006299e-01 -9.05939817e-01 1.62619519e+00
2.05556095e-01 1.11191952e+00 5.49461730e-02 1.31380454e-01
7.92572200e-01 5.61110437e-01 3.66430253e-01 -1.16634917e+00
-8.66717935e-01 1.92102343e-01 1.07260084e+00 -4.22071695e-01
-8.87881637e-01 -1.13710153e+00 -1.66026068e+00 -1.01412078e-02
-2.45956048e-01 7.53876805e-01 4.94699091e-01 1.12973034e+00
-2.34432712e-01 2.45243475e-01 3.19512427e-01 -4.99554902e-01
-3.11613888e-01 -8.68346632e-01 -8.04914176e-01 5.12930930e-01
-4.08644259e-01 -5.32047868e-01 -6.51855648e-01 1.46942511e-01]
|
[10.484550476074219, 9.150739669799805]
|
212604d9-0d9f-49b4-9f2e-785a9c5679bb
|
multimodal-audio-textual-architecture-for-1
|
2306.06819
| null |
https://arxiv.org/abs/2306.06819v2
|
https://arxiv.org/pdf/2306.06819v2.pdf
|
Multimodal Audio-textual Architecture for Robust Spoken Language Understanding
|
Recent voice assistants are usually based on the cascade spoken language understanding (SLU) solution, which consists of an automatic speech recognition (ASR) engine and a natural language understanding (NLU) system. Because such approach relies on the ASR output, it often suffers from the so-called ASR error propagation. In this work, we investigate impacts of this ASR error propagation on state-of-the-art NLU systems based on pre-trained language models (PLM), such as BERT and RoBERTa. Moreover, a multimodal language understanding (MLU) module is proposed to mitigate SLU performance degradation caused by errors present in the ASR transcript. The MLU benefits from self-supervised features learned from both audio and text modalities, specifically Wav2Vec for speech and Bert/RoBERTa for language. Our MLU combines an encoder network to embed the audio signal and a text encoder to process text transcripts followed by a late fusion layer to fuse audio and text logits. We found that the proposed MLU showed to be robust towards poor quality ASR transcripts, while the performance of BERT and RoBERTa are severely compromised. Our model is evaluated on five tasks from three SLU datasets and robustness is tested using ASR transcripts from three ASR engines. Results show that the proposed approach effectively mitigates the ASR error propagation problem, surpassing the PLM models' performance across all datasets for the academic ASR engine.
|
['Chao Xing', 'Mehdi Rezagholizadeh', 'Anderson R. Avila']
|
2023-06-12
| null | null | null | null |
['spoken-language-understanding', 'spoken-language-understanding', 'automatic-speech-recognition']
|
['natural-language-processing', 'speech', 'speech']
|
[ 2.57178783e-01 3.08630407e-01 3.33084822e-01 -3.81613553e-01
-1.26613569e+00 -3.98969114e-01 8.16106379e-01 -2.35236064e-02
-2.92061329e-01 4.18291092e-01 6.88645005e-01 -4.40636545e-01
1.58889309e-01 -3.67765665e-01 -8.33976388e-01 -2.27916166e-01
3.64942998e-01 3.70045513e-01 -1.94625273e-01 -4.21018541e-01
-3.95483570e-03 1.85574308e-01 -1.65578413e+00 6.67428315e-01
9.35064793e-01 1.07658851e+00 2.99992651e-01 1.38398898e+00
-4.29527402e-01 1.00110447e+00 -8.85989130e-01 -1.45766705e-01
-8.44546333e-02 -2.61596322e-01 -8.04158866e-01 -3.25904065e-03
3.17220271e-01 -4.55581874e-01 -4.58182842e-01 5.67923248e-01
8.42960775e-01 1.35056376e-01 5.34804225e-01 -1.05008936e+00
-7.21974909e-01 9.48420525e-01 -6.53774068e-02 -1.03362620e-01
8.68580520e-01 2.51298904e-01 1.10117698e+00 -1.06968415e+00
4.01393443e-01 1.69750392e+00 5.49924433e-01 7.81126976e-01
-1.14536405e+00 -6.25328839e-01 -4.86446964e-03 8.56024250e-02
-1.24538159e+00 -1.08853197e+00 5.04422963e-01 -2.44641885e-01
1.52098453e+00 7.60156959e-02 -5.48314191e-02 1.47053993e+00
1.08443707e-01 1.10779536e+00 7.61405885e-01 -6.67390704e-01
1.65078506e-01 5.33599675e-01 3.74027163e-01 4.85814393e-01
-6.34818614e-01 4.75560091e-02 -8.51777494e-01 9.78536680e-02
2.81875849e-01 -2.82884181e-01 -5.28032839e-01 4.85305309e-01
-8.60438108e-01 5.70964754e-01 8.27989876e-02 5.36702871e-01
-4.27292943e-01 -2.41977349e-01 5.05468786e-01 7.41382241e-01
4.75944400e-01 2.24506289e-01 -5.28046906e-01 -4.60747987e-01
-1.00464296e+00 -3.89000446e-01 9.90041316e-01 9.98663008e-01
5.52902460e-01 2.95928121e-01 -3.94033879e-01 1.35459232e+00
7.53704011e-01 6.18412316e-01 7.55896330e-01 -5.35081983e-01
6.59471333e-01 5.70727527e-01 -1.66305616e-01 -4.14418489e-01
-1.50369033e-01 -2.32795373e-01 -6.17488742e-01 -3.70595865e-02
1.76329277e-02 -2.80410528e-01 -1.20015335e+00 1.44082892e+00
-2.06036866e-01 1.88591152e-01 6.18921101e-01 7.91033804e-01
1.26097310e+00 1.23266721e+00 -3.52023467e-02 -6.23544194e-02
1.06477511e+00 -1.21175873e+00 -1.17906189e+00 -2.24831581e-01
6.84685409e-01 -9.16608512e-01 1.19344032e+00 4.59304661e-01
-1.05638313e+00 -7.71824300e-01 -1.18455970e+00 -1.57161281e-01
-3.57918113e-01 4.16051596e-01 -1.30947337e-01 8.30611706e-01
-1.43873191e+00 2.62375414e-01 -4.27527308e-01 -6.18256152e-01
1.75132770e-02 1.59103259e-01 -5.51322281e-01 -1.39198348e-01
-1.27443826e+00 9.26687598e-01 1.48007557e-01 2.79427767e-01
-1.19698513e+00 -6.13147199e-01 -1.01514256e+00 2.34085813e-01
2.58936882e-01 -2.85495102e-01 1.52281678e+00 -8.86987507e-01
-2.08840990e+00 4.15780246e-01 -3.70294809e-01 -6.24789059e-01
3.95531297e-01 -6.40019238e-01 -6.79923654e-01 1.40568405e-01
-4.51607198e-01 2.77500689e-01 9.86310959e-01 -1.29281235e+00
-5.30024409e-01 -3.99440587e-01 -4.19113785e-01 7.86569118e-02
-2.56402999e-01 6.25605807e-02 -1.02557965e-01 -3.79989326e-01
-1.00786783e-01 -3.37707996e-01 3.02863747e-01 -6.12019658e-01
-2.84568369e-01 -3.74108672e-01 9.04786527e-01 -1.24618733e+00
1.50063992e+00 -2.09152985e+00 -4.45995145e-02 -3.02112773e-02
-2.02811688e-01 7.59714007e-01 -5.42813957e-01 7.02404082e-01
-7.10811839e-02 1.54076576e-01 -3.99982631e-02 -7.42530823e-01
7.51134977e-02 2.74375230e-01 -5.65888226e-01 2.09814329e-02
4.79998678e-01 6.75946951e-01 -6.39474452e-01 1.71616431e-02
4.93252099e-01 7.89974570e-01 -4.25815642e-01 9.19714272e-01
-2.62068361e-01 1.88105896e-01 -5.03504425e-02 5.12617648e-01
3.39318812e-01 2.98553526e-01 -1.83601797e-01 -1.40641332e-01
-1.92476869e-01 8.46228123e-01 -1.12989390e+00 1.59068274e+00
-1.00110924e+00 5.24450660e-01 5.59367895e-01 -8.99323225e-01
1.31620860e+00 1.02113700e+00 2.29752343e-02 -6.33078754e-01
3.83666933e-01 4.33296591e-01 -5.09705357e-02 -4.41285640e-01
3.59546751e-01 -1.27000064e-01 2.27182776e-01 4.23151493e-01
7.66400039e-01 -2.71929771e-01 -3.49357158e-01 4.73045737e-01
1.36910582e+00 -2.05968827e-01 7.84197524e-02 9.01045203e-02
9.87208784e-01 -4.00250673e-01 -4.77148499e-03 9.98486459e-01
-3.17569733e-01 8.47479522e-01 6.23177625e-02 1.81387007e-01
-9.13638890e-01 -1.07906306e+00 3.06701977e-02 1.09524131e+00
-5.02392292e-01 -5.72879314e-01 -9.02139306e-01 -4.51850772e-01
-2.20553637e-01 1.22255647e+00 -1.54443488e-01 -2.36560106e-01
-1.47714674e-01 -6.68239370e-02 8.90623331e-01 4.06304985e-01
2.47941926e-01 -1.06821191e+00 2.01278538e-01 5.27191103e-01
-2.72997171e-01 -1.37673736e+00 -5.28492570e-01 1.32106543e-01
-5.95938087e-01 -5.84099233e-01 -6.36541843e-01 -3.00080597e-01
7.26961121e-02 2.90423691e-01 8.66445720e-01 -2.16879040e-01
-9.68854725e-02 1.09295511e+00 -8.64227474e-01 -5.61359882e-01
-1.27341640e+00 -2.27057692e-02 3.21184486e-01 3.37347418e-01
4.07626331e-01 -4.23429281e-01 1.92537787e-03 2.87525803e-01
-8.89531136e-01 -1.73905015e-01 5.59096336e-01 9.73316133e-01
1.02918528e-01 -2.10676119e-01 9.49768245e-01 -4.10305798e-01
8.20695519e-01 -2.84244150e-01 -1.41940251e-01 4.67723668e-01
-4.97336060e-01 1.02156788e-01 5.28989792e-01 -2.77415723e-01
-1.33950627e+00 -1.57882318e-01 -6.64976180e-01 -4.69344109e-01
-4.15917635e-01 2.89375216e-01 -5.56888461e-01 3.89595777e-01
4.98537123e-01 2.72758901e-01 1.60510778e-01 -6.68045878e-01
5.92671394e-01 1.88154256e+00 5.21675408e-01 -2.74523526e-01
3.61560374e-01 -6.73056766e-02 -1.00850391e+00 -1.66829109e+00
-5.72917283e-01 -7.75339186e-01 -4.79090601e-01 -4.23383445e-01
8.41660142e-01 -1.17748511e+00 -5.24208426e-01 5.58418036e-01
-1.60788536e+00 -1.51559100e-01 8.42779758e-04 5.84220409e-01
-5.44557273e-01 2.55729556e-01 -6.71022058e-01 -1.51176691e+00
-6.52823031e-01 -1.27369714e+00 1.25720847e+00 1.38050333e-01
-4.19159949e-01 -6.95923150e-01 -2.60186821e-01 1.01271868e+00
5.52382052e-01 -6.67577326e-01 6.30986631e-01 -1.08567476e+00
-2.85175115e-01 -2.11409479e-01 -1.28638387e-01 9.33169484e-01
1.14568345e-01 6.62110280e-03 -1.73420060e+00 -3.21383923e-01
1.97010964e-01 -4.21765089e-01 8.62642884e-01 1.73316047e-01
4.78297353e-01 -4.86580431e-01 1.45299330e-01 9.87730026e-02
9.18805242e-01 3.36936414e-01 7.23473012e-01 -1.40663043e-01
5.14476061e-01 7.19566822e-01 2.35851794e-01 2.75757432e-01
1.48414567e-01 5.12263298e-01 1.97746485e-01 1.91252217e-01
-3.60130638e-01 -4.62014705e-01 1.17841470e+00 1.48959863e+00
2.85336167e-01 -5.21744251e-01 -8.40560794e-01 5.14610469e-01
-1.72445762e+00 -5.64954877e-01 1.56010482e-02 2.16420722e+00
8.22353840e-01 -9.71267298e-02 -2.43043438e-01 2.90358454e-01
4.30730313e-01 2.15597644e-01 -1.07087709e-01 -8.13719213e-01
-1.18132703e-01 1.75046474e-01 1.41279116e-01 9.67046618e-01
-6.85840487e-01 1.02251518e+00 5.62136459e+00 7.27571249e-01
-9.75788772e-01 3.32700312e-01 2.40150109e-01 6.99282587e-02
-4.10394490e-01 -5.07365108e-01 -8.14331770e-01 1.39380932e-01
1.54394472e+00 6.35930598e-02 5.51412821e-01 6.48743629e-01
5.75427771e-01 -7.36035332e-02 -1.04970992e+00 9.71540570e-01
3.31528902e-01 -8.03600609e-01 3.27460498e-01 -3.89141291e-01
2.26188138e-01 2.82902181e-01 -1.99064285e-01 5.83460093e-01
2.53425211e-01 -1.21389210e+00 5.22378325e-01 3.59196693e-01
5.95433116e-01 -5.81315219e-01 1.03916550e+00 4.39982265e-01
-9.73838687e-01 -1.20269492e-01 -8.86706710e-02 1.19850403e-02
2.48860836e-01 3.38568360e-01 -1.47640467e+00 6.29748344e-01
6.11360788e-01 4.05580938e-01 -2.55974293e-01 4.03445601e-01
-3.21404189e-01 1.00265205e+00 -4.17498767e-01 -5.85751142e-03
3.63177836e-01 -3.28222066e-02 8.62749815e-01 1.55287945e+00
3.46242696e-01 8.56658444e-02 -2.57686973e-01 8.01101029e-01
-1.65738150e-01 3.39416653e-01 -8.64911795e-01 -5.81901312e-01
4.76986945e-01 8.26325715e-01 -4.18103896e-02 -3.58234435e-01
-5.85184157e-01 1.15803611e+00 8.89195651e-02 4.85933453e-01
-3.97460729e-01 -2.83812463e-01 7.02514827e-01 -4.48144898e-02
-9.30247921e-03 -2.85586923e-01 -1.92722440e-01 -1.12673128e+00
-1.63714699e-02 -9.95370448e-01 1.34565458e-01 -1.01554823e+00
-1.12456572e+00 9.09096956e-01 -6.04495466e-01 -8.39713037e-01
-3.46836776e-01 -6.59711838e-01 -4.93182927e-01 1.08073127e+00
-1.50222600e+00 -1.10984468e+00 1.99147570e-03 4.33190346e-01
1.15184450e+00 -5.92224300e-01 9.02536571e-01 4.31887597e-01
-5.73528409e-01 6.82163119e-01 6.98266998e-02 4.78032343e-02
8.59569967e-01 -1.10461092e+00 1.87118277e-01 7.82485306e-01
2.42737234e-01 6.49079740e-01 7.01783180e-01 -5.16105592e-01
-1.55160940e+00 -9.91069198e-01 1.12965012e+00 -4.55079585e-01
7.07410574e-01 -4.38385934e-01 -1.21059203e+00 7.12400198e-01
5.35276771e-01 -6.27112865e-01 6.05930984e-01 -5.76039143e-02
-3.79490227e-01 -1.68214701e-02 -9.03808594e-01 3.65870327e-01
6.66167676e-01 -1.07447183e+00 -1.01635206e+00 -1.16660856e-02
1.10602224e+00 -2.09937822e-02 -8.80529165e-01 3.55969816e-01
4.97189581e-01 -7.89849102e-01 7.25430548e-01 -6.07598305e-01
5.16751528e-01 -1.32575706e-01 -5.69404066e-01 -1.57223856e+00
2.88502693e-01 -7.09487975e-01 -8.14995170e-02 1.80397034e+00
7.38218129e-01 -4.81166065e-01 -1.11971907e-02 5.99457502e-01
-2.85904437e-01 -8.87017772e-02 -1.11299098e+00 -5.65812051e-01
-4.23798889e-01 -9.89080310e-01 1.37268454e-01 4.94070292e-01
3.89097780e-02 9.89717722e-01 -4.18203622e-01 3.96751285e-01
2.96626985e-01 -7.10530400e-01 8.18940639e-01 -9.42751944e-01
-7.69668594e-02 -2.90474236e-01 -1.50915235e-01 -1.05712593e+00
2.89214730e-01 -8.31995428e-01 4.65550840e-01 -1.63046610e+00
-3.40104520e-01 1.68248609e-01 -1.50259659e-01 3.44255030e-01
-7.53500015e-02 -3.77807885e-01 3.35926026e-01 -9.63361859e-02
-2.90117234e-01 1.10906160e+00 8.45764875e-01 -3.14372957e-01
-5.52912712e-01 5.52151492e-03 -3.67221802e-01 5.88170171e-01
4.64846909e-01 -1.34368643e-01 -2.87416667e-01 -3.82979512e-01
-4.51866746e-01 2.98219413e-01 6.49260730e-02 -9.75997984e-01
3.77743423e-01 6.71383023e-01 -6.55736327e-02 -6.71013355e-01
4.76520598e-01 -9.10454810e-01 -3.36200774e-01 -1.01342537e-02
-4.87367958e-01 -4.59550232e-01 3.73412430e-01 3.49037349e-01
-5.50239384e-01 -1.81252196e-01 8.13095033e-01 1.29214406e-01
-5.33383310e-01 -3.06668311e-01 -7.81535029e-01 -3.96418333e-01
3.80122751e-01 1.39107913e-01 -2.58684427e-01 -9.26227033e-01
-7.16259420e-01 4.38987911e-01 -2.07625985e-01 8.66005898e-01
8.70325923e-01 -9.70287800e-01 -8.07408273e-01 4.82360750e-01
1.62138075e-01 -8.95345360e-02 4.04151767e-01 7.36527920e-01
-1.62767507e-02 6.68814361e-01 2.46482506e-01 -5.10983109e-01
-1.40857744e+00 1.36478975e-01 4.22806770e-01 4.43192944e-02
-5.10531127e-01 8.90478909e-01 -5.74622676e-03 -1.01224267e+00
8.84627342e-01 -3.56794983e-01 -3.93422782e-01 2.02357799e-01
9.36507940e-01 5.23691714e-01 2.50054002e-01 -8.14846933e-01
-1.59105584e-01 1.57545358e-01 -6.74679801e-02 -5.48231363e-01
1.21403122e+00 -6.11778378e-01 2.03718051e-01 7.92915821e-01
1.28814363e+00 -2.52435416e-01 -7.06753373e-01 -4.21379179e-01
2.29779482e-02 1.92840397e-01 3.90250355e-01 -1.09229183e+00
-5.83186626e-01 1.34140110e+00 7.01755941e-01 7.34205768e-02
9.65513170e-01 -4.43587303e-02 1.06077421e+00 5.12260735e-01
1.49718046e-01 -1.34299505e+00 8.56257230e-03 9.07291114e-01
1.29781365e+00 -1.46096444e+00 -6.55381203e-01 -1.44646719e-01
-7.61983871e-01 1.13165307e+00 4.33288455e-01 2.45823830e-01
6.60918593e-01 3.46467108e-01 5.61079085e-01 2.43385911e-01
-9.09953773e-01 -2.43741959e-01 2.34402627e-01 4.92149651e-01
5.83425879e-01 1.10091336e-01 -1.35899140e-02 1.15867507e+00
-2.50353009e-01 3.61606553e-02 5.59770167e-01 5.42571902e-01
-4.42787468e-01 -8.50030005e-01 -6.07626498e-01 2.60830432e-01
-2.91394919e-01 -1.58952817e-01 -7.05260336e-01 3.43929321e-01
-4.23650652e-01 1.68082690e+00 -1.40122473e-01 -8.15232694e-01
6.83652461e-01 6.96922600e-01 -3.93101759e-02 -6.30505621e-01
-8.53708029e-01 3.18115830e-01 3.00802350e-01 -6.04653239e-01
-2.38068089e-01 -2.59222835e-01 -1.20021641e+00 1.15121394e-01
-4.74904478e-01 1.46235988e-01 8.71996343e-01 1.12196124e+00
4.10815865e-01 9.76472080e-01 6.65647984e-01 -6.01431310e-01
-6.80623949e-01 -1.54212070e+00 -5.89738369e-01 1.73544168e-01
7.13371575e-01 -2.99992442e-01 -5.98021507e-01 -1.59707796e-02]
|
[14.196942329406738, 6.777410984039307]
|
29a85c3f-7594-41d7-841b-665817c4ba97
|
lstm-ccg-parsing
| null | null |
https://aclanthology.org/N16-1026
|
https://aclanthology.org/N16-1026.pdf
|
LSTM CCG Parsing
| null |
['Mike Lewis', 'Luke Zettlemoyer', 'Kenton Lee']
|
2016-06-01
| null | null | null |
naacl-2016-6
|
['ccg-supertagging']
|
['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]
|
[-7.263617992401123, 3.790346384048462]
|
71737789-6da0-4cd3-938b-4f98bd27b059
|
when-to-use-what-an-in-depth-comparative
|
2211.08228
| null |
https://arxiv.org/abs/2211.08228v1
|
https://arxiv.org/pdf/2211.08228v1.pdf
|
When to Use What: An In-Depth Comparative Empirical Analysis of OpenIE Systems for Downstream Applications
|
Open Information Extraction (OpenIE) has been used in the pipelines of various NLP tasks. Unfortunately, there is no clear consensus on which models to use in which tasks. Muddying things further is the lack of comparisons that take differing training sets into account. In this paper, we present an application-focused empirical survey of neural OpenIE models, training sets, and benchmarks in an effort to help users choose the most suitable OpenIE systems for their applications. We find that the different assumptions made by different models and datasets have a statistically significant effect on performance, making it important to choose the most appropriate model for one's applications. We demonstrate the applicability of our recommendations on a downstream Complex QA application.
|
['Yunyao Li', 'ChengXiang Zhai', 'Kevin Chen-Chuan Chang', 'Ishan Jindal', 'Kevin Pei']
|
2022-11-15
| null | null | null | null |
['open-information-extraction']
|
['natural-language-processing']
|
[ 7.09626228e-02 1.17683530e-01 -2.36300766e-01 -5.66173673e-01
-9.75863874e-01 -8.72705817e-01 4.40396547e-01 2.72332460e-01
-5.21093726e-01 6.99712098e-01 5.14486790e-01 -9.06842470e-01
-5.13140738e-01 -3.98106575e-01 -5.12743294e-01 -2.39324406e-01
3.01259160e-01 6.13624096e-01 2.07684740e-01 -3.04007381e-01
3.27910334e-01 3.49340498e-01 -1.02484429e+00 5.91781795e-01
9.80001092e-01 4.63263780e-01 2.90151834e-01 5.78576028e-01
-4.14774477e-01 7.96351790e-01 -5.53322792e-01 -6.06140733e-01
2.57209599e-01 -3.34159940e-01 -1.35623932e+00 -7.32559860e-01
3.25050980e-01 5.62103391e-02 7.74954036e-02 7.91649342e-01
7.32146561e-01 7.76999071e-03 9.19028640e-01 -9.14012492e-01
-8.12942207e-01 9.27185714e-01 -4.01282217e-03 5.05866349e-01
1.35290399e-01 3.14257562e-01 1.36733377e+00 -5.02817690e-01
1.02791739e+00 9.71727252e-01 6.25587642e-01 4.27607745e-01
-1.17790174e+00 -5.60587168e-01 1.28830979e-02 1.36110023e-01
-1.15403593e+00 -7.26344883e-01 2.23480910e-01 -4.59476650e-01
1.23384953e+00 7.11365193e-02 6.68020099e-02 8.83580089e-01
4.85689044e-01 4.65035707e-01 1.03590906e+00 -5.71864188e-01
9.94364098e-02 8.34751874e-02 5.90604782e-01 3.32276970e-01
5.32329261e-01 -1.05782546e-01 -3.05461526e-01 -3.14711064e-01
1.40569225e-01 -7.35231698e-01 -4.26255047e-01 -1.88590437e-02
-9.01305020e-01 8.54116380e-01 3.43408436e-01 5.51667333e-01
-1.48948565e-01 -1.77792862e-01 5.94793677e-01 4.34087515e-01
4.45133448e-01 8.31949592e-01 -9.32631314e-01 -2.11471379e-01
-7.57664621e-01 3.99095327e-01 1.04219639e+00 6.95373774e-01
4.76699501e-01 -4.74020392e-01 -2.07317844e-01 9.34383631e-01
3.59733164e-01 -9.90662202e-02 4.91731584e-01 -1.16846395e+00
7.21439242e-01 4.44842070e-01 1.32745683e-01 -6.15105569e-01
-5.20298481e-01 -3.50890487e-01 -1.14019722e-01 1.04314066e-01
9.46452618e-01 -5.45688808e-01 -6.13425076e-01 1.58019447e+00
-7.32027665e-02 -3.16644013e-01 3.43010694e-01 6.60777628e-01
9.37084854e-01 5.71449637e-01 4.56820518e-01 -1.37354177e-03
1.33181441e+00 -5.27986050e-01 -6.72950685e-01 -6.08988285e-01
1.11277902e+00 -9.86053765e-01 8.89326096e-01 3.16976190e-01
-8.73983443e-01 -2.51599073e-01 -9.59712386e-01 -5.63884974e-01
-4.91719574e-01 -1.72026560e-01 7.25763857e-01 6.05956316e-01
-8.00936282e-01 7.80807257e-01 -4.89937365e-01 -4.77534026e-01
3.06556821e-01 3.04533184e-01 -3.22444528e-01 -8.24822858e-03
-1.22218251e+00 1.39907181e+00 4.90447730e-01 1.04701415e-01
-1.97197437e-01 -6.00983024e-01 -5.55160403e-01 1.87206194e-01
3.02690089e-01 -8.07156086e-01 1.50670779e+00 -5.40007055e-01
-9.93060291e-01 6.33632481e-01 -2.12120056e-01 -5.20474255e-01
1.64397404e-01 -1.93067625e-01 -3.80992591e-01 -3.03169042e-01
2.18216941e-01 5.44456303e-01 1.69444516e-01 -7.60366976e-01
-7.90257633e-01 -4.20124024e-01 2.06109494e-01 2.40754798e-01
5.09719504e-03 6.26206160e-01 -2.85227180e-01 -2.05065846e-01
8.20421875e-02 -8.29706728e-01 -1.67047411e-01 -4.72194791e-01
-1.23078577e-01 -6.71956480e-01 3.55690598e-01 -8.00087333e-01
1.41426539e+00 -1.96531308e+00 1.18680403e-01 -2.10980177e-01
1.66477665e-01 1.65036768e-01 4.56080772e-04 5.43513775e-01
8.87030214e-02 5.23110390e-01 -4.24624160e-02 9.74753946e-02
1.31552413e-01 3.04506034e-01 -2.19181061e-01 1.48532271e-01
2.61244684e-01 7.04838514e-01 -8.73552442e-01 -6.36232972e-01
-3.56614560e-01 4.31259483e-01 -5.69861352e-01 5.56332467e-04
-4.37123150e-01 2.84727782e-01 -5.68770587e-01 3.78304064e-01
2.38973737e-01 -2.47319430e-01 3.06028247e-01 -8.15397575e-02
-3.55817467e-01 1.18893242e+00 -7.80262828e-01 1.58102643e+00
-3.87492031e-01 1.00909507e+00 -1.00744970e-01 -4.46423024e-01
7.11517751e-01 6.14803076e-01 9.14699733e-02 -3.49063039e-01
3.36590677e-01 3.95778000e-01 7.09714592e-01 -5.87286770e-01
4.27195996e-01 -1.91487774e-01 1.56355262e-01 6.96236551e-01
3.57137978e-01 1.91236347e-01 4.91629541e-01 1.39939472e-01
1.33188415e+00 2.93547064e-01 2.95441210e-01 -3.81856024e-01
-3.58940586e-02 2.87859946e-01 8.18367898e-01 7.39317298e-01
-3.66718322e-01 4.33118254e-01 5.78240573e-01 -3.30281287e-01
-8.02751184e-01 -5.58549106e-01 -4.89833921e-01 1.11288023e+00
-4.19649839e-01 -4.81371224e-01 -8.14451993e-01 -8.82121444e-01
-3.02537024e-01 9.06788349e-01 -1.57232568e-01 3.01658899e-01
-4.66253340e-01 -6.87094867e-01 8.20435822e-01 2.98658878e-01
9.18267220e-02 -1.22238326e+00 -6.89441919e-01 2.07747236e-01
-6.83837891e-01 -1.16151798e+00 -2.21814334e-01 5.70820987e-01
-1.03548169e+00 -1.01939821e+00 -3.03757727e-01 -7.28832483e-01
2.11202398e-01 2.47082021e-02 1.13299632e+00 -3.24017227e-01
2.95440555e-01 1.64392907e-02 -5.18805563e-01 -6.80469811e-01
-7.12068737e-01 5.98678648e-01 -1.12310775e-01 -5.50148487e-01
8.61455977e-01 -1.34003341e-01 -2.45224059e-01 -2.46582907e-02
-5.49815536e-01 -1.23472005e-01 6.59286380e-01 7.07248330e-01
2.60991752e-02 2.64351722e-02 6.63984418e-01 -1.22813392e+00
1.09168971e+00 -6.78028107e-01 -2.99626112e-01 5.05654812e-01
-8.94446135e-01 2.90997118e-01 2.86523759e-01 1.13899531e-02
-1.32799625e+00 -2.26666361e-01 -4.43768054e-01 2.77754068e-01
-3.92650992e-01 8.45746458e-01 -2.27053046e-01 4.28106226e-02
1.03620684e+00 -5.84621847e-01 -1.98973820e-01 -9.22734976e-01
4.31162566e-01 9.99868035e-01 3.49413604e-01 -4.47481185e-01
2.66618848e-01 -1.18478581e-01 -4.73739475e-01 -8.36048841e-01
-8.09734702e-01 -3.66847396e-01 -7.67538846e-01 2.76830435e-01
8.88992548e-01 -7.08095670e-01 -2.59087950e-01 -1.53156957e-02
-1.39187455e+00 -4.23290998e-01 9.31933802e-03 3.85475308e-01
-3.59419763e-01 3.41474712e-01 -7.50333250e-01 -4.97892916e-01
-5.47312915e-01 -1.32828617e+00 3.91300619e-01 2.64778942e-01
-7.45096803e-01 -9.61583018e-01 2.07236320e-01 7.06881344e-01
2.90371031e-01 -2.95340478e-01 1.15284419e+00 -1.00073290e+00
-3.12919676e-01 1.30577773e-01 -2.40838990e-01 2.84324139e-01
8.87123868e-02 1.00119837e-01 -1.15191042e+00 1.92889854e-01
-1.86442614e-01 -1.57913476e-01 5.84012449e-01 4.05936986e-01
5.64756989e-01 -1.69500649e-01 -4.89328623e-01 4.45274502e-01
1.25890446e+00 1.89020589e-01 5.25253654e-01 5.36908984e-01
3.35021943e-01 8.43244016e-01 5.30527055e-01 -1.91569537e-01
3.71382415e-01 4.45509285e-01 -1.86686352e-01 3.41287464e-01
1.69351697e-02 -1.97167471e-01 3.86657834e-01 5.80243587e-01
6.90322220e-02 -4.85883564e-01 -1.09290981e+00 5.92941999e-01
-1.63938808e+00 -7.93768942e-01 -1.51896596e-01 1.92661297e+00
9.32268143e-01 4.59488243e-01 -1.56579375e-01 -1.29680485e-01
3.13803554e-01 -2.96269190e-02 -1.51672021e-01 -8.69446933e-01
7.12604001e-02 2.95650035e-01 4.37940240e-01 4.44891781e-01
-8.28006327e-01 8.40298831e-01 7.11109591e+00 5.98928571e-01
-8.62995625e-01 2.69784808e-01 5.31837523e-01 1.47543684e-01
-1.52039915e-01 5.13967395e-01 -1.24401391e+00 3.28049153e-01
1.52269197e+00 -4.61986475e-02 1.07938990e-01 4.13027704e-01
3.40951413e-01 -4.47196774e-02 -1.30752861e+00 2.40458414e-01
-3.30183983e-01 -1.38851440e+00 -2.52234131e-01 1.78409535e-02
1.89520240e-01 5.88725269e-01 -4.52550203e-01 4.92916673e-01
7.84904420e-01 -9.09587979e-01 3.57631981e-01 2.39256337e-01
2.36091271e-01 -2.79664248e-01 6.81886137e-01 5.00220597e-01
-3.99472296e-01 -2.63190597e-01 -3.28833073e-01 -1.75318822e-01
3.46502006e-01 2.05183938e-01 -1.09328246e+00 3.54553282e-01
7.06648707e-01 7.67640918e-02 -5.02483189e-01 1.02914047e+00
-3.36014211e-01 9.26349699e-01 -3.73072445e-01 -1.06008515e-01
2.63081521e-01 -5.52419536e-02 3.21780801e-01 1.46729565e+00
5.28698713e-02 1.76255077e-01 -2.50600457e-01 7.37885237e-01
3.06144971e-02 3.80158961e-01 -8.17997932e-01 -3.25442255e-01
5.39605916e-01 1.03714132e+00 -5.28611898e-01 9.34925899e-02
-8.81412923e-01 6.61964059e-01 5.67282975e-01 1.58841625e-01
-3.05248439e-01 -5.18657267e-01 7.68182635e-01 9.04763564e-02
8.66956264e-02 -1.85082719e-01 -3.91700953e-01 -9.77365851e-01
-2.09800810e-01 -9.44242716e-01 8.78633201e-01 -1.00509214e+00
-1.28420079e+00 5.98168850e-01 4.31474485e-02 -7.15736866e-01
-5.20053685e-01 -7.83110559e-01 -5.81067622e-01 1.21757615e+00
-1.41551936e+00 -8.79667401e-01 2.58633643e-01 -9.19217989e-02
4.25550461e-01 5.75366430e-03 8.43276858e-01 2.48864636e-01
-5.90232015e-01 3.85118723e-01 1.75093964e-01 4.99827772e-01
9.93288577e-01 -1.10045445e+00 7.22997606e-01 8.27061594e-01
4.01368350e-01 1.02793813e+00 8.51513445e-01 -6.32496357e-01
-1.11604083e+00 -6.21569037e-01 1.65200281e+00 -1.02127218e+00
5.96718431e-01 4.73288074e-02 -9.23148215e-01 1.10919249e+00
4.72715080e-01 -3.21323901e-01 1.01095164e+00 7.45327353e-01
-1.94926560e-01 1.16943359e-01 -1.03693724e+00 4.83565748e-01
8.02637160e-01 -3.70960385e-01 -9.82751131e-01 1.26335874e-01
8.51068199e-01 -3.95903498e-01 -1.17030072e+00 4.25621688e-01
6.10349357e-01 -7.55561054e-01 5.85339606e-01 -9.86747384e-01
3.52653086e-01 -2.89318282e-02 -1.25632241e-01 -1.42863059e+00
-5.37196457e-01 -3.16855997e-01 3.08703929e-01 1.42891359e+00
1.07530499e+00 -7.46150255e-01 5.10290623e-01 9.06295657e-01
-1.92345098e-01 -8.61344278e-01 -6.97622657e-01 -3.94377202e-01
5.97732306e-01 -6.48229420e-01 4.82641816e-01 8.98763955e-01
4.03516620e-01 9.68933523e-01 1.67814702e-01 2.26216018e-01
2.23177686e-01 -1.33300237e-02 6.06963992e-01 -1.29323304e+00
-3.17372590e-01 -3.62149507e-01 -1.49776805e-02 -8.88021946e-01
5.84434383e-02 -9.92176354e-01 -1.31267041e-01 -1.63392603e+00
9.68353599e-02 -8.58306646e-01 -3.16348791e-01 5.97481370e-01
-9.21667665e-02 -4.15620580e-03 3.14680576e-01 2.73417830e-01
-2.63571411e-01 -1.12219580e-01 6.09746873e-01 1.26643793e-03
-4.41623718e-01 3.43897380e-02 -1.37731576e+00 6.93765759e-01
1.02331591e+00 -4.35955107e-01 -4.10269946e-01 -8.48929465e-01
5.46339989e-01 -1.13348261e-01 -1.29486844e-01 -8.82478416e-01
2.37979203e-01 -1.27045542e-01 1.54457986e-01 -4.03016448e-01
6.67209644e-03 -5.45172036e-01 -3.56944390e-02 1.81573927e-01
-4.99408692e-01 8.10059905e-02 3.68131727e-01 8.50443467e-02
1.19014874e-01 -9.78706002e-01 4.41299319e-01 -2.89585143e-01
-6.87827468e-01 1.07212020e-02 -5.04065335e-01 4.48935360e-01
4.30001259e-01 1.17968023e-01 -3.75472039e-01 -1.80797100e-01
-6.27653301e-01 2.90152162e-01 2.11034849e-01 5.53764164e-01
1.93023786e-01 -7.16974974e-01 -7.81345427e-01 -1.13415316e-01
1.60286739e-01 -5.72624654e-02 -1.39561996e-01 6.27096057e-01
-5.02615333e-01 9.05148625e-01 -1.65141895e-01 -1.18452638e-01
-1.23315477e+00 3.23234975e-01 3.52172315e-01 -5.17783463e-01
-5.83012879e-01 5.49963355e-01 -2.56501079e-01 -6.13532186e-01
9.06490833e-02 -3.29795182e-01 -3.67364079e-01 2.47092079e-02
5.67775369e-01 3.69757324e-01 4.57865804e-01 -5.74277759e-01
-6.24420047e-02 -4.16587815e-02 -4.53519881e-01 -2.72692829e-01
1.24373913e+00 2.84354445e-02 5.12693524e-02 5.38924038e-01
8.20827007e-01 -1.02469511e-01 -8.81045282e-01 -3.46230656e-01
2.78371364e-01 -1.91826612e-01 2.88648844e-01 -9.97903943e-01
-5.16112685e-01 9.19443667e-01 6.09592617e-01 1.17906585e-01
6.87331438e-01 -1.33738061e-02 6.74142003e-01 4.88216311e-01
4.40717667e-01 -1.15399301e+00 -8.88627648e-01 7.75842488e-01
5.87177455e-01 -1.07853627e+00 1.36159703e-01 -4.25021440e-01
-6.50946021e-01 9.14701164e-01 4.16391701e-01 3.56841922e-01
7.03890800e-01 1.64576083e-01 3.29105675e-01 -3.63965273e-01
-1.06879532e+00 -2.17539787e-01 -3.43027860e-02 6.54306173e-01
1.05467629e+00 -1.04612850e-01 -9.05003846e-01 7.95076311e-01
-2.03278303e-01 2.13686153e-01 4.54373956e-01 9.24610555e-01
-5.34847140e-01 -1.45289445e+00 -2.90476084e-01 7.21782088e-01
-7.85887182e-01 -2.34713882e-01 -5.91195643e-01 7.93531060e-01
3.31124127e-01 1.17822850e+00 -1.45981506e-01 -1.94456294e-01
2.07361773e-01 6.63177669e-01 5.73215365e-01 -9.50133860e-01
-8.07462573e-01 -3.09941798e-01 6.98133945e-01 -1.77374274e-01
-5.78902662e-02 -9.07183707e-01 -1.26503038e+00 -2.71894455e-01
-6.19005740e-01 3.58274996e-01 5.51710010e-01 1.15952623e+00
7.32687473e-01 2.88792044e-01 -5.74423969e-02 -4.19527233e-01
-7.13945568e-01 -1.30309582e+00 -1.44413769e-01 1.75899148e-01
-1.02143086e-01 -5.47293723e-01 -5.23273088e-02 -1.20047368e-01]
|
[10.478492736816406, 9.202512741088867]
|
d21b02b8-1eba-4cec-83d0-29b62595954a
|
knowledge-augmented-frame-semantic-parsing
|
2303.14375
| null |
https://arxiv.org/abs/2303.14375v1
|
https://arxiv.org/pdf/2303.14375v1.pdf
|
Knowledge-augmented Frame Semantic Parsing with Hybrid Prompt-tuning
|
Frame semantics-based approaches have been widely used in semantic parsing tasks and have become mainstream. It remains challenging to disambiguate frame representations evoked by target lexical units under different contexts. Pre-trained Language Models (PLMs) have been used in semantic parsing and significantly improve the accuracy of neural parsers. However, the PLMs-based approaches tend to favor collocated patterns presented in the training data, leading to inaccurate outcomes. The intuition here is to design a mechanism to optimally use knowledge captured in semantic frames in conjunction with PLMs to disambiguate frames. We propose a novel Knowledge-Augmented Frame Semantic Parsing Architecture (KAF-SPA) to enhance semantic representation by incorporating accurate frame knowledge into PLMs during frame semantic parsing. Specifically, a Memory-based Knowledge Extraction Module (MKEM) is devised to select accurate frame knowledge and construct the continuous templates in the high dimensional vector space. Moreover, we design a Task-oriented Knowledge Probing Module (TKPM) using hybrid prompts (in terms of continuous and discrete prompts) to incorporate the selected knowledge into the PLMs and adapt PLMs to the tasks of frame and argument identification. Experimental results on two public FrameNet datasets demonstrate that our method significantly outperforms strong baselines (by more than +3$\%$ in F1), achieving state-of-art results on the current benchmark. Ablation studies verify the effectiveness of KAF-SPA.
|
['Wei Peng', 'Jingyuan Yang', 'Yajing Sun', 'Rui Zhang']
|
2023-03-25
| null | null | null | null |
['semantic-parsing']
|
['natural-language-processing']
|
[ 4.78178382e-01 2.84845054e-01 -3.80752206e-01 -6.25905097e-01
-8.53671312e-01 -5.55240214e-01 6.92961693e-01 -4.51251827e-02
-5.77995121e-01 6.18681610e-01 5.31570971e-01 -1.91662505e-01
1.12097509e-01 -9.71857131e-01 -8.04078639e-01 -3.96808326e-01
4.77369666e-01 3.29492420e-01 5.67554832e-01 -2.84121335e-01
2.07930610e-01 -1.17812589e-01 -1.50136817e+00 7.32370794e-01
6.33676589e-01 9.27277029e-01 8.11566412e-01 2.89055198e-01
-8.25549126e-01 1.01869261e+00 -6.44705296e-01 -5.11431456e-01
-9.10527706e-02 -4.37476844e-01 -1.35968864e+00 -2.25545049e-01
-4.66768406e-02 -2.09334940e-02 2.88746506e-02 9.72699583e-01
4.38788652e-01 3.30212116e-01 7.15023279e-02 -7.11815357e-01
-5.60095429e-01 1.04535961e+00 -1.54980674e-01 4.66278553e-01
5.76412380e-01 -1.95991382e-01 1.18169844e+00 -1.08698165e+00
8.06271136e-01 1.80170059e+00 4.83801454e-01 7.18578160e-01
-1.06553495e+00 -4.56993699e-01 6.08319700e-01 5.70025682e-01
-9.03391600e-01 -5.43139398e-01 9.98329759e-01 -1.38786435e-01
1.26517558e+00 3.44961174e-02 4.35389638e-01 1.30450070e+00
-1.99784264e-01 1.17623270e+00 8.97489548e-01 -7.83278644e-01
3.43521327e-01 -3.31425369e-01 6.66670144e-01 5.74530780e-01
-8.99133086e-02 -1.30273461e-01 -9.94759321e-01 -1.29649818e-01
8.23026538e-01 -3.42037439e-01 -1.38524175e-01 9.56231356e-02
-1.12493551e+00 1.06345332e+00 2.07413599e-01 5.03428102e-01
-5.27874589e-01 -2.77759023e-02 6.18598521e-01 3.93585209e-03
3.86270761e-01 4.48742032e-01 -5.77396214e-01 -4.16241080e-01
-4.50262845e-01 4.49805319e-01 5.70029795e-01 8.48875761e-01
6.67059004e-01 -5.44854142e-02 -5.25082648e-01 1.16703963e+00
2.99590826e-01 3.27454567e-01 7.22283721e-01 -9.82914507e-01
7.86426306e-01 8.20812821e-01 -5.32922931e-02 -8.24487567e-01
-5.25349200e-01 -1.93678364e-02 -1.53656989e-01 -5.70566356e-01
3.78525674e-01 7.13107223e-03 -8.36176515e-01 2.19033790e+00
5.12497246e-01 4.71174449e-01 4.03662205e-01 9.48050857e-01
9.50465679e-01 6.55365586e-01 7.77638555e-01 -2.44172797e-01
1.77678132e+00 -8.01878035e-01 -9.38287914e-01 -6.71210110e-01
7.91070282e-01 -6.56536043e-01 1.38195121e+00 -2.03027558e-02
-1.08732164e+00 -7.05502450e-01 -7.94899046e-01 -1.87468365e-01
-1.66499183e-01 1.25335464e-02 6.68716669e-01 3.99583250e-01
-7.04924345e-01 3.94815177e-01 -1.10078204e+00 -2.95751870e-01
5.12786627e-01 1.84188709e-01 -1.84155360e-01 3.51189487e-02
-1.71794271e+00 8.26315761e-01 9.18341577e-01 6.83202967e-02
-5.19184768e-01 -6.06887043e-01 -1.13401222e+00 -2.29416955e-02
6.38781905e-01 -6.82572365e-01 1.42655933e+00 -1.03624892e+00
-1.58667862e+00 7.96778321e-01 -6.34081781e-01 -6.83663607e-01
-1.73423827e-01 -4.96752590e-01 -3.23545486e-01 3.36468518e-01
4.62306082e-01 7.60858953e-01 7.59885907e-01 -9.47598994e-01
-8.17264378e-01 -1.76307872e-01 4.93563026e-01 3.94742757e-01
-1.70124158e-01 2.44899556e-01 -3.41815561e-01 -8.60485494e-01
3.57150972e-01 -7.46395111e-01 -4.30075675e-02 -7.84054279e-01
-1.79313689e-01 -6.72223568e-01 7.59182811e-01 -6.66529536e-01
1.49698591e+00 -1.98838758e+00 2.93202966e-01 -2.31947735e-01
-1.59454927e-01 6.19995236e-01 -1.80615615e-02 3.44082490e-02
1.46141620e-02 -4.69749942e-02 -1.07968308e-01 -3.42924803e-01
-1.93372369e-02 6.47016644e-01 -5.23368120e-01 -1.98733985e-01
5.99891305e-01 1.02339911e+00 -1.14452660e+00 -5.45988441e-01
2.90877432e-01 3.62661093e-01 -6.44781888e-01 1.74809918e-01
-5.78988433e-01 3.36516023e-01 -8.39958370e-01 5.91415763e-01
2.87342191e-01 -4.27301258e-01 6.20122433e-01 -3.08646768e-01
1.72712952e-01 8.19411814e-01 -1.17231464e+00 1.97897124e+00
-4.04389143e-01 1.01146668e-01 -3.23059887e-01 -1.14160514e+00
8.15706134e-01 5.00360370e-01 7.08490834e-02 -7.85781860e-01
2.04094216e-01 2.30975635e-02 -1.40924975e-01 -3.50401759e-01
4.60727066e-01 -9.75359380e-02 -2.52410918e-01 9.46981087e-02
3.74669224e-01 3.14037561e-01 2.81440169e-01 2.49798354e-02
1.13558650e+00 4.22563970e-01 3.05278152e-01 -3.90340984e-01
7.75456667e-01 1.27465501e-01 1.07292783e+00 6.02037489e-01
-1.88003570e-01 4.13937241e-01 5.13270497e-01 -6.98641598e-01
-4.28467005e-01 -7.92730808e-01 -3.05689890e-02 1.56513715e+00
3.26790541e-01 -7.56292403e-01 -1.06099665e+00 -9.04297709e-01
-4.32813197e-01 9.14615631e-01 -4.63396072e-01 -1.29223183e-01
-1.10677898e+00 -8.73834848e-01 5.20830035e-01 9.40576077e-01
7.29288161e-01 -1.37026215e+00 -8.58082354e-01 6.46001399e-01
-6.66983664e-01 -1.47482789e+00 4.71744873e-02 2.28336170e-01
-6.53518915e-01 -1.17071998e+00 -1.27544612e-01 -8.25545847e-01
3.52653384e-01 2.16232076e-01 1.22296619e+00 5.40612964e-03
4.16567862e-01 1.04927681e-01 -7.74597287e-01 -4.17521536e-01
-2.06101045e-01 1.25488922e-01 -1.82346731e-01 3.64276804e-02
7.24889576e-01 -1.62544087e-01 -3.48985463e-01 3.31556261e-01
-8.29723537e-01 3.14358532e-01 1.63181245e-01 1.15888834e+00
7.27869451e-01 -2.85100698e-01 7.85271585e-01 -1.20629632e+00
4.40006971e-01 -3.48623276e-01 -5.40484667e-01 2.77280539e-01
-1.98135883e-01 2.47686639e-01 4.92622972e-01 -3.87739271e-01
-1.52957714e+00 -4.31427360e-03 -2.23806441e-01 -2.55465925e-01
-7.52357468e-02 5.26432514e-01 -4.72515285e-01 4.36662227e-01
5.69027126e-01 -5.15156612e-02 -3.49861801e-01 -4.89427567e-01
3.69107813e-01 1.40970692e-01 5.15022993e-01 -1.19608569e+00
2.99533397e-01 2.57307678e-01 -3.28772098e-01 -5.77251375e-01
-1.38550878e+00 -2.05083564e-01 -4.97938454e-01 6.74131736e-02
1.19539189e+00 -1.08684433e+00 -1.29691944e-01 3.66522312e-01
-1.45357895e+00 -2.10815817e-01 -1.86744526e-01 2.70698518e-01
-5.29200017e-01 1.93921909e-01 -6.05641127e-01 -5.76981783e-01
-3.00884128e-01 -1.40912843e+00 1.22824740e+00 4.42335337e-01
-4.87730056e-01 -1.07330108e+00 -2.35743254e-01 5.38728058e-01
2.20550627e-01 4.75316718e-02 1.13878119e+00 -9.22455370e-01
-4.10119593e-01 3.84062946e-01 -1.16316497e-01 1.54079959e-01
-1.74914580e-02 -5.90660453e-01 -1.23268223e+00 2.05889136e-01
6.31221831e-02 -2.25851849e-01 7.97765195e-01 4.72943008e-01
1.03769755e+00 -3.35220322e-02 -4.82129186e-01 3.60466570e-01
1.09055161e+00 3.70792478e-01 5.49729347e-01 6.66714907e-01
7.51624286e-01 4.46849018e-01 8.96716237e-01 1.41214237e-01
7.05032587e-01 7.36733794e-01 -1.24018071e-02 3.55370462e-01
-2.06231445e-01 -4.06549215e-01 4.49959010e-01 6.10347271e-01
2.25338433e-02 -6.90453500e-02 -9.32513654e-01 5.23020148e-01
-2.18733811e+00 -8.16247165e-01 1.01502001e-01 1.79087174e+00
1.08541358e+00 5.13124704e-01 -1.63205653e-01 1.65084034e-01
9.40814674e-01 2.79686868e-01 -2.21598789e-01 -2.87008017e-01
-3.62388283e-01 7.52405047e-01 6.85655028e-02 5.69247365e-01
-1.36491060e+00 1.75890291e+00 5.40382671e+00 9.55512643e-01
-9.99531686e-01 4.55706418e-01 4.56569552e-01 2.52143800e-01
-1.45831704e-01 3.28461409e-01 -1.23026192e+00 5.10945797e-01
9.85422969e-01 1.39230236e-01 1.26593947e-01 9.40124393e-01
1.04416892e-01 -8.45278278e-02 -1.02114785e+00 8.40552032e-01
-2.51769155e-01 -1.67803633e+00 1.93755597e-01 -5.82023621e-01
3.15126419e-01 -1.68180585e-01 -4.90901291e-01 6.06867850e-01
3.23302835e-01 -6.46009862e-01 8.66383135e-01 2.91240841e-01
3.68771702e-01 -4.66833115e-01 6.95798993e-01 1.62183955e-01
-1.42438424e+00 -1.44501731e-01 -4.56527293e-01 -1.83902889e-01
4.78859544e-01 4.62020487e-01 -8.38600218e-01 6.62270844e-01
5.51473558e-01 6.89306676e-01 -3.65430921e-01 3.00728381e-01
-6.60799265e-01 9.42669630e-01 -2.61566639e-01 1.19595990e-01
2.40705371e-01 2.85042673e-01 4.96482581e-01 1.08747435e+00
-9.02806073e-02 4.01093155e-01 4.38625097e-01 8.09259772e-01
9.32787359e-03 9.22154933e-02 -1.29721522e-01 -3.40444148e-02
8.91700685e-01 1.03926802e+00 -9.26362813e-01 -5.00166476e-01
-5.56261599e-01 6.03730857e-01 4.14585829e-01 2.22355440e-01
-8.10830355e-01 -2.05550995e-02 7.94149280e-01 -1.12486809e-01
3.99491489e-01 -5.16288579e-02 -1.92698359e-01 -1.21826816e+00
3.72333117e-02 -7.57899344e-01 7.38279879e-01 -6.40009105e-01
-1.06666493e+00 7.33338058e-01 3.55537146e-01 -6.86863065e-01
-4.32489127e-01 -6.23876393e-01 -3.95232409e-01 8.30155611e-01
-1.66783154e+00 -1.28572583e+00 4.55403365e-02 5.34586787e-01
1.15298891e+00 -1.04412667e-01 1.00731993e+00 1.53933361e-01
-7.37892568e-01 3.05074185e-01 -7.56589532e-01 3.52679580e-01
3.78151536e-01 -1.17027843e+00 5.87458491e-01 1.05250776e+00
2.24101573e-01 6.20348394e-01 4.18613255e-01 -7.16675103e-01
-1.36645424e+00 -1.10481799e+00 1.13182914e+00 -5.58315873e-01
4.59504664e-01 -2.18681440e-01 -1.17568707e+00 7.75513709e-01
-8.86593312e-02 9.70152840e-02 5.71338952e-01 2.42305785e-01
-4.00362849e-01 3.24726224e-01 -8.56269360e-01 4.55823779e-01
1.35625494e+00 -5.57361543e-01 -1.29633701e+00 1.65112376e-01
1.00447798e+00 -7.46153712e-01 -6.57300055e-01 5.77398419e-01
1.82074159e-01 -6.50733769e-01 1.12955439e+00 -6.21054053e-01
2.52973855e-01 -3.45105767e-01 -4.49621767e-01 -1.00765681e+00
-2.83508897e-01 -3.07907164e-01 -1.92895129e-01 1.42712343e+00
2.77666479e-01 -4.98803198e-01 5.53057492e-01 6.81372583e-01
-2.77697951e-01 -6.52747869e-01 -1.04408610e+00 -4.31056529e-01
-1.92687511e-01 -6.58404768e-01 7.44728684e-01 9.33157086e-01
-5.22938892e-02 8.01667392e-01 1.08089760e-01 3.30156177e-01
1.01738960e-01 1.38041034e-01 3.25167835e-01 -1.20987749e+00
-2.52163231e-01 -3.10374320e-01 -2.09623337e-01 -9.64514673e-01
8.56322289e-01 -1.00031006e+00 -1.17012106e-01 -1.43873990e+00
-7.07600266e-02 -5.31103134e-01 -3.83283436e-01 8.25027585e-01
-7.49779165e-01 -2.77203530e-01 3.59519541e-01 1.18433841e-01
-7.73157775e-01 3.87959093e-01 8.36990118e-01 2.30617970e-01
-3.41850609e-01 -3.23208332e-01 -8.45870614e-01 1.05549741e+00
5.36609709e-01 -3.36546242e-01 -5.09660840e-01 -8.88285697e-01
1.27369195e-01 2.29729805e-02 3.83497238e-01 -8.37369978e-01
-5.09446375e-02 -1.43251166e-01 2.61239082e-01 -4.38814163e-01
3.74105185e-01 -4.23061579e-01 -4.39303406e-02 4.83219624e-02
-3.85962158e-01 1.54409632e-01 2.50365108e-01 4.13131475e-01
-3.59779656e-01 -3.49342495e-01 5.64607680e-01 -3.75216275e-01
-1.36623812e+00 -1.45844325e-01 -5.05645648e-02 4.19545680e-01
6.67175949e-01 4.59140614e-02 -4.96952981e-01 1.94740593e-01
-7.63873100e-01 2.06685454e-01 7.24623427e-02 6.73465312e-01
4.65976596e-01 -1.33559310e+00 -2.56540895e-01 2.59198874e-01
4.99241687e-02 2.90855378e-01 1.10474542e-01 4.39682275e-01
-2.03603238e-01 3.17732215e-01 5.46095818e-02 -7.65817285e-01
-9.45706546e-01 1.61910549e-01 1.10136069e-01 -4.39185351e-01
-6.63844705e-01 1.12594998e+00 3.00569028e-01 -2.95955777e-01
5.27562015e-02 -5.37084818e-01 -4.77408677e-01 1.25235870e-01
6.94326997e-01 1.14500625e-02 1.40766576e-01 -7.59102881e-01
-4.56198782e-01 3.55949879e-01 -3.88447344e-02 -7.69804092e-03
1.33143008e+00 -6.97076321e-02 7.28458986e-02 2.12251663e-01
7.60896564e-01 -3.52692544e-01 -1.26136911e+00 -7.37931073e-01
6.11937821e-01 -3.14101726e-01 9.62499604e-02 -8.07935834e-01
-7.49641955e-01 6.91898346e-01 3.56171548e-01 -3.13946232e-02
1.07644916e+00 2.42928296e-01 1.01095331e+00 2.97342122e-01
4.45523828e-01 -1.20172369e+00 -4.52983081e-02 9.78036284e-01
4.17429209e-01 -1.10622704e+00 -5.44638872e-01 -6.52067602e-01
-7.51975298e-01 1.08760321e+00 7.37739503e-01 -7.62710394e-03
4.01359558e-01 1.85661748e-01 1.00246906e-01 -1.01316422e-01
-8.88635814e-01 -3.70400637e-01 1.33831188e-01 3.63696009e-01
5.20159662e-01 9.01598483e-02 -5.42791009e-01 1.39051914e+00
-1.66622996e-01 -1.36432678e-01 -2.26013623e-02 1.36208820e+00
-5.85279107e-01 -1.67467225e+00 -4.50271398e-01 1.36236131e-01
-5.88556230e-01 -1.65975481e-01 -1.08677402e-01 5.38505614e-01
2.13412866e-01 9.95651245e-01 -2.47878935e-02 -2.39194170e-01
4.17401254e-01 8.34194779e-01 3.67618114e-01 -9.00537014e-01
-6.43185258e-01 2.21297771e-01 4.17618692e-01 -8.56283426e-01
-8.83588672e-01 -7.65239239e-01 -1.75182366e+00 2.77023017e-01
-3.36585313e-01 7.41567016e-02 3.01551968e-01 1.47628427e+00
4.74668324e-01 6.30386889e-01 1.86841786e-01 -7.27938354e-01
-2.97254145e-01 -8.11622024e-01 1.63650841e-01 5.90526104e-01
-1.17832638e-01 -1.24174321e+00 1.75224438e-01 3.10055971e-01]
|
[10.295258522033691, 9.254343032836914]
|
c90561ad-7a35-4c48-8d1d-999a40b97921
|
fademl-understanding-the-impact-of-pre
|
1811.01444
| null |
http://arxiv.org/abs/1811.01444v1
|
http://arxiv.org/pdf/1811.01444v1.pdf
|
FAdeML: Understanding the Impact of Pre-Processing Noise Filtering on Adversarial Machine Learning
|
Deep neural networks (DNN)-based machine learning (ML) algorithms have
recently emerged as the leading ML paradigm particularly for the task of
classification due to their superior capability of learning efficiently from
large datasets. The discovery of a number of well-known attacks such as dataset
poisoning, adversarial examples, and network manipulation (through the addition
of malicious nodes) has, however, put the spotlight squarely on the lack of
security in DNN-based ML systems. In particular, malicious actors can use these
well-known attacks to cause random/targeted misclassification, or cause a
change in the prediction confidence, by only slightly but systematically
manipulating the environmental parameters, inference data, or the data
acquisition block. Most of the prior adversarial attacks have, however, not
accounted for the pre-processing noise filters commonly integrated with the
ML-inference module. Our contribution in this work is to show that this is a
major omission since these noise filters can render ineffective the majority of
the existing attacks, which rely essentially on introducing adversarial noise.
Apart from this, we also extend the state of the art by proposing a novel
pre-processing noise Filter-aware Adversarial ML attack called FAdeML. To
demonstrate the effectiveness of the proposed methodology, we generate an
adversarial attack image by exploiting the "VGGNet" DNN trained for the "German
Traffic Sign Recognition Benchmarks (GTSRB" dataset, which despite having no
visual noise, can cause a classifier to misclassify even in the presence of
pre-processing noise filters.
|
['Semeen Rehman', 'Muhammmad Abdullah Hanif', 'Faiq Khalid', 'Muhammad Shafique', 'Junaid Qadir']
|
2018-11-04
| null | null | null | null |
['traffic-sign-recognition']
|
['computer-vision']
|
[ 4.55114514e-01 8.57195035e-02 3.66047740e-01 -1.58804934e-02
-4.16601837e-01 -9.26943600e-01 9.55785453e-01 1.22185320e-01
-5.47220826e-01 7.54063606e-01 -4.96805131e-01 -6.31944299e-01
-1.68304145e-01 -9.70044792e-01 -1.05806160e+00 -8.52851331e-01
-1.18123755e-01 2.32781127e-01 4.20242220e-01 -3.10059451e-02
1.16676278e-01 1.05409884e+00 -1.44955373e+00 1.06788278e-01
4.64740485e-01 9.01737034e-01 -5.14136076e-01 6.81285322e-01
1.51891455e-01 7.85762846e-01 -1.06758749e+00 -7.57820189e-01
5.16425729e-01 -2.53453910e-01 -3.17693323e-01 -3.41595441e-01
5.54459035e-01 -4.27545726e-01 -5.24926126e-01 1.28898942e+00
5.59866607e-01 -1.98573023e-01 4.52460229e-01 -1.50657535e+00
-4.14447449e-02 6.23438060e-01 3.87880951e-03 1.48006305e-02
-1.69216320e-01 6.71876490e-01 4.85867500e-01 -3.15680444e-01
6.51286602e-01 1.26341558e+00 6.62851453e-01 7.02589214e-01
-1.26079369e+00 -8.86855423e-01 -1.88629478e-01 7.41325840e-02
-1.09610224e+00 -2.56843060e-01 8.10740948e-01 -3.59381467e-01
5.57340860e-01 3.80462527e-01 3.16586584e-01 1.67000163e+00
1.37877464e-01 3.86967421e-01 1.18842566e+00 -3.67699027e-01
5.01292050e-01 2.15907209e-02 8.82683881e-03 4.64366347e-01
6.26765251e-01 5.82363486e-01 -6.78946078e-02 -2.81376660e-01
4.20817792e-01 -1.10439770e-01 -1.50220439e-01 -3.10850769e-01
-7.51154006e-01 8.17300320e-01 6.28433883e-01 2.49056965e-01
-2.32980579e-01 5.49937904e-01 7.71245837e-01 3.57185423e-01
1.06016211e-01 4.32282984e-01 -5.11386812e-01 1.46488622e-01
-6.17379248e-01 1.96324512e-01 8.25881958e-01 4.10354257e-01
6.06801748e-01 5.22542000e-01 -7.46846646e-02 1.40400007e-01
2.49552310e-01 6.17182910e-01 2.17742398e-01 -2.86684513e-01
4.98940229e-01 6.15377903e-01 -1.55053422e-01 -1.05552530e+00
-4.16563660e-01 -5.85560799e-01 -8.80329132e-01 9.12210703e-01
7.55756736e-01 -5.12077570e-01 -1.02920020e+00 1.66841161e+00
3.14367294e-01 3.44852090e-01 2.93283969e-01 5.35884559e-01
4.38317358e-01 2.01953501e-01 3.38778585e-01 4.40095589e-02
9.35614347e-01 -3.48897785e-01 -4.45009887e-01 3.38495448e-02
3.88187081e-01 -4.35699970e-01 8.07251751e-01 5.64098895e-01
-5.85281134e-01 -3.58482242e-01 -1.25536323e+00 5.16773105e-01
-9.41526353e-01 -3.15733820e-01 3.97660017e-01 1.28916407e+00
-5.50343752e-01 8.14010203e-01 -7.78919697e-01 -1.11383058e-01
7.81196475e-01 5.64871967e-01 -3.43658388e-01 -1.06147073e-01
-1.39748764e+00 9.76718247e-01 3.72660756e-01 2.36937732e-01
-1.30259264e+00 -7.76676118e-01 -5.77015102e-01 4.23004963e-02
4.19029862e-01 -3.65319252e-01 6.54132783e-01 -1.01132214e+00
-1.32802665e+00 6.15094900e-01 5.81638575e-01 -9.82148111e-01
1.10207152e+00 -6.78113699e-02 -6.15577817e-01 1.09109551e-01
-5.51450312e-01 3.63637656e-01 1.14913511e+00 -1.32023680e+00
-1.74258068e-01 -3.93574923e-01 1.62603840e-01 -5.87026179e-01
-3.50349307e-01 7.79520199e-02 1.42722696e-01 -8.06351840e-01
-4.91511524e-01 -8.10154855e-01 -4.12824094e-01 6.78764880e-02
-6.99752688e-01 2.66211510e-01 1.19332111e+00 -4.49235439e-01
9.31389570e-01 -2.20654750e+00 -1.78324819e-01 5.61226189e-01
1.21479839e-01 9.87946868e-01 -6.26840293e-02 3.58103037e-01
-1.40147164e-01 3.38261813e-01 -4.76109177e-01 -1.53813526e-01
1.02169700e-01 4.27364349e-01 -6.12133682e-01 6.33131206e-01
3.81399900e-01 8.74015450e-01 -6.60206437e-01 8.16351622e-02
5.81044018e-01 6.86438978e-01 -3.28239173e-01 -7.42237642e-02
-4.26282674e-01 3.82961333e-01 -2.46306509e-01 3.28676432e-01
6.35202765e-01 4.29424793e-01 1.38545245e-01 -3.12360555e-01
2.58337080e-01 -6.00773767e-02 -1.13933814e+00 8.91523957e-01
-9.56207812e-02 5.83914757e-01 -1.12105154e-01 -1.06877708e+00
7.79504478e-01 2.25451350e-01 7.01729953e-02 -4.90306228e-01
3.96262914e-01 3.39902133e-01 3.22544903e-01 -3.21939617e-01
-7.37568438e-02 -2.80417278e-02 -1.10415846e-01 1.72988877e-01
5.66556305e-02 1.43584535e-01 -1.14081040e-01 -7.43277147e-02
1.50185978e+00 -1.55105710e-01 1.02556109e-01 5.19249924e-02
6.46988153e-01 -7.27138370e-02 3.53958309e-01 9.53405499e-01
-2.70019531e-01 2.81045556e-01 7.37763226e-01 -4.11507875e-01
-1.02272856e+00 -1.10127831e+00 -1.43678635e-01 4.26302969e-01
-1.74772471e-01 1.06521197e-01 -9.43303287e-01 -1.23994720e+00
1.66147232e-01 7.15295315e-01 -6.04501367e-01 -4.55844373e-01
-6.02568865e-01 -9.08266723e-01 1.45397878e+00 2.57160276e-01
7.66906261e-01 -1.21617293e+00 -6.38420761e-01 2.09521845e-01
5.50057411e-01 -1.20495033e+00 3.58560175e-01 4.75962132e-01
-6.84783936e-01 -1.48586714e+00 -2.14634717e-01 -7.11480677e-02
5.99082410e-01 -5.67311525e-01 7.81451881e-01 2.74893701e-01
-5.37431121e-01 1.80349499e-01 -3.77238899e-01 -5.34138262e-01
-1.10553515e+00 7.63369305e-03 9.39468592e-02 2.25396961e-01
2.22707704e-01 -6.92488015e-01 -2.73754090e-01 2.04542190e-01
-1.38908494e+00 -5.96303105e-01 5.16787827e-01 6.34555697e-01
1.45075783e-01 3.72129261e-01 8.22821677e-01 -1.23949409e+00
4.33027387e-01 -2.95375317e-01 -9.91133630e-01 8.62894580e-02
-4.75173652e-01 1.07745595e-01 1.19895089e+00 -6.81706965e-01
-5.57061732e-01 8.58145803e-02 -5.15550613e-01 -5.88774145e-01
-6.33451104e-01 2.21826836e-01 -8.44765186e-01 -5.87754965e-01
8.61281514e-01 8.19975808e-02 -1.04524359e-01 -3.84024024e-01
3.59324098e-01 3.02169561e-01 6.01751804e-01 -2.11263195e-01
1.44817507e+00 5.40724099e-01 6.02418840e-01 -6.19360268e-01
-3.68549734e-01 2.00636148e-01 -4.15349871e-01 -1.77573621e-01
6.46604300e-01 -3.49109679e-01 -9.47196126e-01 9.72293019e-01
-9.22173977e-01 -4.52943474e-01 -2.43495166e-01 1.53960764e-01
-5.74931167e-02 3.82230043e-01 -4.15968567e-01 -8.22837949e-01
-2.58326352e-01 -1.22136056e+00 2.58168697e-01 6.38805628e-02
1.63580522e-01 -9.76510048e-01 -3.11014205e-01 5.52723035e-02
5.84035039e-01 8.67892146e-01 1.12803817e+00 -1.24205077e+00
-5.77768445e-01 -5.27836859e-01 -6.09910041e-02 8.95258129e-01
-1.29132763e-01 1.42274424e-01 -1.36114407e+00 -2.14232802e-01
6.75805882e-02 -2.16747537e-01 7.44262218e-01 -9.91563797e-02
1.00475669e+00 -4.36647475e-01 -1.24566458e-01 5.38260937e-01
1.75821245e+00 2.49674723e-01 9.47742462e-01 3.37305188e-01
8.07210267e-01 3.06365162e-01 5.40940985e-02 2.06879675e-01
-4.21756893e-01 6.96702600e-01 1.13863969e+00 -1.40908808e-01
-6.91566616e-02 -1.87812477e-01 4.66595590e-01 -1.38100162e-01
1.48695946e-01 -5.12602150e-01 -8.03997636e-01 8.80651847e-02
-1.42109251e+00 -9.95009542e-01 -3.39315087e-01 2.25736737e+00
4.60595697e-01 5.84002554e-01 -6.73931465e-02 7.94916034e-01
5.53130984e-01 7.50197768e-02 -6.01570725e-01 -3.72558326e-01
-2.32431948e-01 5.99724948e-01 1.01178527e+00 3.39460552e-01
-1.32374990e+00 8.58013272e-01 5.12898350e+00 9.24912930e-01
-1.28302860e+00 6.70331568e-02 2.88189143e-01 1.59379885e-01
-2.92186737e-02 -2.02162355e-01 -7.58724570e-01 5.63122690e-01
9.98635888e-01 3.95926565e-01 3.94751608e-01 7.71448076e-01
1.58854842e-01 1.06109262e-01 -9.65023696e-01 6.05925798e-01
9.62627605e-02 -1.18038213e+00 2.47808680e-01 1.24599718e-01
3.82298142e-01 -6.28431737e-02 1.45085782e-01 3.66130412e-01
4.93091226e-01 -1.12525463e+00 5.71892619e-01 2.79506505e-01
4.71191317e-01 -1.05217898e+00 7.68764079e-01 3.13646138e-01
-6.04539037e-01 -2.59796381e-01 -2.10350156e-01 3.03916968e-02
6.89963177e-02 6.36177003e-01 -7.82592654e-01 6.36648238e-01
2.71444917e-01 1.44263431e-01 -6.97960436e-01 8.79477501e-01
-5.83793640e-01 9.97422576e-01 -4.12662923e-01 8.62294659e-02
3.16439837e-01 3.17674875e-01 7.16142356e-01 1.01258373e+00
-1.22451559e-01 -4.34712440e-01 -1.10831425e-01 8.47880661e-01
-3.26497555e-01 -3.00856203e-01 -7.68896163e-01 7.93514103e-02
2.65138298e-01 1.12633920e+00 -6.57433569e-01 -8.57555643e-02
-1.65348485e-01 8.38982761e-01 1.92284549e-03 2.96084136e-01
-9.54529822e-01 -2.79696465e-01 6.90647840e-01 5.68057001e-02
4.99859661e-01 -7.38713965e-02 -2.78605729e-01 -8.02053452e-01
7.49756470e-02 -1.15130699e+00 1.71293393e-01 -2.52881646e-01
-1.37001669e+00 4.42957908e-01 -2.07953885e-01 -1.09865499e+00
-4.12977785e-02 -9.17891920e-01 -4.75022644e-01 7.26386368e-01
-1.39607179e+00 -1.26657736e+00 -8.87425244e-02 7.84255683e-01
1.11494049e-01 -4.25550640e-01 7.94808209e-01 5.60764372e-01
-5.72262764e-01 8.73517573e-01 2.36728694e-02 5.23009300e-01
4.15321141e-01 -9.66180801e-01 4.41607237e-01 1.24651146e+00
8.30918923e-02 2.64461339e-01 9.00339484e-01 -6.56782031e-01
-1.20990944e+00 -1.44909847e+00 4.19671953e-01 -5.47390759e-01
8.14737737e-01 -7.55095959e-01 -9.03585970e-01 5.18241704e-01
-2.29204684e-01 3.76780957e-01 2.28754923e-01 -5.63324273e-01
-6.12805963e-01 -2.90012240e-01 -1.62002230e+00 6.96137071e-01
6.85795963e-01 -4.81056690e-01 -2.74882585e-01 1.67041957e-01
5.63497186e-01 -2.16052547e-01 -6.03480756e-01 5.81274211e-01
2.93764889e-01 -1.05636358e+00 1.19316268e+00 -7.14402616e-01
2.69016266e-01 -4.40317631e-01 -4.92169298e-02 -1.00311410e+00
4.50343460e-01 -6.60044312e-01 -1.22599728e-01 1.57564247e+00
2.24981219e-01 -9.19748664e-01 7.59504497e-01 4.37223911e-01
4.92394716e-03 -3.69796872e-01 -1.18117511e+00 -8.65246415e-01
1.02030352e-01 -7.55457044e-01 5.09274364e-01 9.39864576e-01
-7.15841532e-01 -1.88096136e-01 -3.78721565e-01 5.58892787e-01
6.64445221e-01 -6.24537468e-01 1.00913906e+00 -1.20221841e+00
-3.78550768e-01 -3.99751842e-01 -9.60419714e-01 -2.09717989e-01
1.60765439e-01 -7.38004446e-01 -7.36937523e-02 -6.71686828e-01
-4.25459385e-01 -4.93874758e-01 -2.60867000e-01 5.36631942e-01
-1.89334434e-02 4.32766259e-01 3.27007443e-01 -5.99436136e-03
-1.35132194e-01 4.63650674e-02 7.06628144e-01 -1.56050801e-01
2.10822240e-01 2.98141688e-01 -1.74221680e-01 8.21040571e-01
7.20152795e-01 -9.27443504e-01 -1.85138956e-01 -3.33396383e-02
3.67103189e-01 -4.23365682e-01 1.08225763e+00 -1.32581389e+00
1.23820148e-01 3.11464936e-01 3.07780087e-01 -2.84413844e-01
1.22240782e-02 -1.36459672e+00 3.41512293e-01 9.54935431e-01
-2.28093073e-01 -1.80151343e-01 4.53711063e-01 6.63594782e-01
-1.39867160e-02 -3.97841275e-01 1.03710961e+00 9.25908983e-02
-5.63867629e-01 1.02773257e-01 -5.80753326e-01 -3.63372490e-02
1.16104436e+00 -9.79667008e-02 -5.95276415e-01 8.81173685e-02
-6.32449985e-01 -2.67939657e-01 4.20590997e-01 1.94180131e-01
2.78061956e-01 -8.51672411e-01 -4.46714729e-01 3.73270273e-01
-1.95188403e-01 -3.10922880e-02 1.96629748e-01 4.72602814e-01
-6.70841455e-01 7.44202062e-02 -3.06573749e-01 -2.51969069e-01
-1.31109381e+00 9.31675911e-01 6.89270198e-01 -5.03219664e-01
-4.48753297e-01 5.07125318e-01 -2.95214713e-01 -4.01279747e-01
5.31970918e-01 -1.77907810e-01 1.43342838e-01 -1.09906591e-01
2.64132053e-01 4.22461390e-01 4.20895487e-01 -3.37641299e-01
-3.12378019e-01 1.88250735e-01 1.75889388e-01 1.94363534e-01
1.16916561e+00 3.93500358e-01 2.12611377e-01 7.12143183e-02
9.80049968e-01 1.47152483e-01 -1.16123009e+00 7.78990686e-02
1.30691335e-01 -1.23914883e-01 -1.74400970e-01 -1.07087374e+00
-1.26372898e+00 1.03184509e+00 9.17549968e-01 3.59035939e-01
1.05955005e+00 -4.17850822e-01 5.15583098e-01 4.55967098e-01
3.44393462e-01 -5.84637225e-01 -6.34070486e-02 3.26937914e-01
4.70194310e-01 -1.01890159e+00 -1.61409542e-01 -1.99711546e-01
-2.05807060e-01 1.09301567e+00 3.52665365e-01 -3.38554412e-01
5.62101901e-01 6.13547444e-01 2.70200253e-01 -9.42907855e-02
-3.17099363e-01 1.68592129e-02 -1.14802286e-01 7.94369936e-01
-2.93158948e-01 -1.99373186e-01 -4.90762778e-02 5.27536333e-01
1.40398756e-01 -3.45992073e-02 5.39493561e-01 7.27557242e-01
-1.57246262e-01 -1.31126738e+00 -5.57634354e-01 2.64384449e-01
-8.14584792e-01 -1.16939940e-01 -4.66563314e-01 1.07347178e+00
4.45569724e-01 7.63119042e-01 -3.14394295e-01 -4.29769754e-01
4.58069056e-01 1.67046130e-01 2.01927096e-01 -1.25853390e-01
-1.33746850e+00 -5.62716961e-01 -1.99884223e-03 -4.91728365e-01
-2.02085316e-01 -4.73245174e-01 -8.95891488e-01 -4.26537424e-01
-2.14030176e-01 -4.02263582e-01 9.19086635e-01 1.11834025e+00
8.40001404e-02 7.03810751e-01 5.61008811e-01 -7.36559212e-01
-9.66826499e-01 -8.10739279e-01 -4.27475333e-01 6.89107955e-01
3.04883987e-01 -6.06716633e-01 -8.71937573e-01 -1.07270546e-01]
|
[5.535895824432373, 7.800463676452637]
|
a8aabf9d-9b30-4b0c-b3a3-050882f87e79
|
cmuq-hybrid-sentiment-classification-by
| null | null |
https://aclanthology.org/S14-2028
|
https://aclanthology.org/S14-2028.pdf
|
CMUQ-Hybrid: Sentiment Classification By Feature Engineering and Parameter Tuning
| null |
['Kamla Al-Mannai', 'Sabih Bin Wasi', 'Behrang Mohit', 'Houda Bouamor', 'Rukhsar Neyaz', 'Hanan Alshikhabobakr']
|
2014-08-01
| null | null | null |
semeval-2014-8
|
['twitter-sentiment-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]
|
[-7.273940563201904, 3.725492238998413]
|
ea9ba6b1-3f3e-4c98-8a82-6dc3ed4d29d4
|
feature-forwarding-for-efficient-single-image
|
1904.09059
| null |
https://arxiv.org/abs/1904.09059v2
|
https://arxiv.org/pdf/1904.09059v2.pdf
|
Feature Forwarding for Efficient Single Image Dehazing
|
Haze degrades content and obscures information of images, which can negatively impact vision-based decision-making in real-time systems. In this paper, we propose an efficient fully convolutional neural network (CNN) image dehazing method designed to run on edge graphical processing units (GPUs). We utilize three variants of our architecture to explore the dependency of dehazed image quality on parameter count and model design. The first two variants presented, a small and big version, make use of a single efficient encoder-decoder convolutional feature extractor. The final variant utilizes a pair of encoder-decoders for atmospheric light and transmission map estimation. Each variant ends with an image refinement pyramid pooling network to form the final dehazed image. For the big variant of the single-encoder network, we demonstrate state-of-the-art performance on the NYU Depth dataset. For the small variant, we maintain competitive performance on the super-resolution O/I-HAZE datasets without the need for image cropping. Finally, we examine some challenges presented by the Dense-Haze dataset when leveraging CNN architectures for dehazing of dense haze imagery and examine the impact of loss function selection on image quality. Benchmarks are included to show the feasibility of introducing this approach into real-time systems.
|
['Seung Jae Lee', 'Peter Morales', 'Tzofi Klinghoffer']
|
2019-04-19
| null | null | null | null |
['image-cropping']
|
['computer-vision']
|
[ 4.47688133e-01 -2.16178358e-01 5.60193956e-01 -1.97420686e-01
-2.70635873e-01 -8.54876786e-02 2.61291236e-01 -2.72652894e-01
-6.44747317e-01 2.66927898e-01 -9.25828889e-02 -5.09041667e-01
3.05094123e-01 -9.84971941e-01 -9.94048715e-01 -9.55639243e-01
-8.30171779e-02 -4.55213308e-01 6.96652651e-01 -4.37681198e-01
2.62089014e-01 3.47008377e-01 -1.76862335e+00 4.64800566e-01
8.01352203e-01 1.34975851e+00 1.96179360e-01 1.05997002e+00
3.16421390e-01 9.99139965e-01 -5.97453177e-01 -3.14133495e-01
6.88365698e-01 -1.63672969e-01 -1.25415683e-01 1.29167959e-01
1.13819110e+00 -1.00554204e+00 -6.34200156e-01 1.02987230e+00
6.32031500e-01 -2.73136813e-02 3.49434584e-01 -1.08585942e+00
-7.55092025e-01 -4.20545340e-02 -5.82348287e-01 5.85967422e-01
-4.00887489e-01 4.34099853e-01 5.53029180e-01 -1.04380238e+00
2.72854418e-01 8.63102496e-01 7.38804519e-01 2.86456376e-01
-8.80828500e-01 -8.69321287e-01 -1.52643904e-01 4.52709466e-01
-1.49337578e+00 -6.84432149e-01 3.67925227e-01 -3.75928581e-01
1.25874734e+00 -4.33417223e-02 5.94933689e-01 2.74545997e-01
6.02514505e-01 3.53775024e-01 9.94598866e-01 -3.69426757e-01
2.91281372e-01 -9.85903740e-02 8.86523339e-04 8.72261882e-01
7.21371353e-01 4.52453285e-01 -6.36618495e-01 2.80969381e-01
7.55276382e-01 -1.59047589e-01 -4.89275038e-01 6.59360737e-02
-7.14837015e-01 7.64516294e-01 6.91404045e-01 -2.42427945e-01
-3.31363261e-01 5.35290480e-01 -1.90202538e-02 3.65477711e-01
6.51708663e-01 2.90564328e-01 -9.57529768e-02 2.57759720e-01
-1.17173445e+00 3.80550027e-01 4.15762365e-01 9.68682170e-01
9.73394692e-01 3.50328684e-01 -1.10872038e-01 4.91710097e-01
1.36656120e-01 5.66447198e-01 9.32426751e-02 -1.00184155e+00
3.12887907e-01 1.00891806e-01 1.82157308e-01 -7.87321627e-01
-2.29112953e-01 -3.39901924e-01 -8.12589824e-01 9.03674483e-01
1.55471787e-01 -2.50560820e-01 -1.42419136e+00 1.01934958e+00
2.81884581e-01 4.84219939e-01 -1.61219630e-02 1.02182376e+00
8.49699497e-01 9.18804288e-01 -4.62949812e-01 2.28241637e-01
1.23971677e+00 -1.34781516e+00 -5.94876409e-01 -3.28946769e-01
3.00375789e-01 -8.12994301e-01 5.98229408e-01 5.87230802e-01
-1.32392311e+00 -4.86159921e-01 -1.65557921e+00 -6.14122093e-01
-4.69029903e-01 -1.20364502e-01 3.34431112e-01 8.64861012e-01
-1.52225697e+00 6.09973192e-01 -9.07137394e-01 1.24879695e-01
5.12852073e-01 4.95313585e-01 -9.86418724e-02 -4.66923267e-01
-1.06981194e+00 8.16271067e-01 3.54255050e-01 2.38579020e-01
-9.51708615e-01 -8.70231986e-01 -8.25636446e-01 2.21059948e-01
-4.51020384e-03 -9.14577127e-01 1.13276637e+00 -7.90335238e-01
-1.51338041e+00 5.82856476e-01 -6.82525635e-02 -7.93790936e-01
2.42852062e-01 -3.42940003e-01 -4.11019057e-01 2.86357790e-01
-3.60438079e-01 8.23717594e-01 1.50669420e+00 -1.02014315e+00
-1.03687823e+00 -1.27034932e-01 7.73649067e-02 1.85612157e-01
-2.15521574e-01 -9.58326608e-02 -6.54940605e-01 -6.68539941e-01
-2.38966778e-01 -7.64443278e-01 -2.53949463e-01 3.53072703e-01
-1.18551426e-01 4.65523541e-01 9.86209273e-01 -6.84851825e-01
1.19927883e+00 -2.41939068e+00 -2.78851986e-01 -1.23334944e-01
5.54563701e-01 5.87734938e-01 -1.15536056e-01 7.00983480e-02
8.19713399e-02 -9.85217988e-02 -3.40521216e-01 -5.25836349e-01
-4.30418938e-01 1.09279841e-01 -1.47665828e-01 6.50488973e-01
5.27823508e-01 6.30709648e-01 -4.88027006e-01 -1.83940709e-01
4.57487315e-01 8.64981771e-01 -9.23309743e-01 4.18925732e-01
1.34112127e-03 3.32510360e-02 1.04116671e-01 5.16545296e-01
1.10686266e+00 -1.13519385e-01 -6.18743062e-01 -3.10573399e-01
-5.41276038e-01 2.12094814e-01 -8.56649160e-01 1.28394091e+00
-7.46307254e-01 1.24779928e+00 3.02820235e-01 -3.20117027e-01
5.43257594e-01 6.32024631e-02 -1.29665196e-01 -7.92089760e-01
9.68139768e-02 2.27912739e-01 -7.52476603e-02 -4.24016446e-01
9.89529490e-01 -4.49789539e-02 5.60685933e-01 4.27889228e-02
-1.41138777e-01 -4.35360372e-01 8.10723379e-03 5.33268414e-03
1.07328176e+00 -2.51056612e-01 -6.34784475e-02 -2.23448262e-01
1.38569370e-01 -3.41527909e-02 1.34786591e-01 7.83245981e-01
-2.25634545e-01 1.08086598e+00 1.68427452e-01 -6.82371199e-01
-1.40503442e+00 -1.03025711e+00 -3.17243725e-01 9.12721217e-01
2.94456303e-01 -3.00648957e-01 -7.47586548e-01 7.97223896e-02
-2.22398475e-01 6.93563282e-01 -6.39653981e-01 -2.97578216e-01
-5.81367135e-01 -9.75306451e-01 4.39777523e-01 2.51603276e-01
1.05647480e+00 -4.67278957e-01 -1.17781341e+00 8.99848044e-02
5.79435937e-02 -1.47717094e+00 -3.88974100e-01 3.03181589e-01
-6.58033669e-01 -6.96935594e-01 -6.92624807e-01 -6.91074967e-01
3.66479069e-01 8.50354493e-01 1.02084672e+00 2.14837968e-01
-3.49069357e-01 -9.82635468e-03 -4.70846772e-01 -7.65924752e-01
-2.01112419e-01 -1.01980641e-01 -3.14103037e-01 1.53575033e-01
1.10043138e-01 -5.41840315e-01 -1.27307534e+00 1.64444461e-01
-1.24784458e+00 3.19525391e-01 5.69226682e-01 7.43123770e-01
3.77464741e-01 3.93926740e-01 -2.00859189e-01 -6.71537340e-01
1.48140967e-01 -4.22010720e-01 -9.27745581e-01 -2.01655507e-01
-8.15618098e-01 8.74897242e-02 4.64509159e-01 -1.22053869e-01
-1.07665479e+00 2.54934877e-02 -1.92723945e-01 -6.64133966e-01
1.19996287e-01 1.58180997e-01 1.48314878e-01 -5.99422395e-01
7.12967932e-01 2.36036986e-01 -8.30018669e-02 -2.07341328e-01
2.90207356e-01 8.27405334e-01 9.60192680e-01 1.97691307e-03
9.75859225e-01 9.11771238e-01 3.67897227e-02 -1.24824715e+00
-6.41698360e-01 -4.82348889e-01 -2.88143456e-01 -1.04758963e-01
1.26840997e+00 -1.48959994e+00 -3.23612988e-01 8.77154171e-01
-1.25196266e+00 -6.73521101e-01 5.80917932e-02 3.61689776e-01
-1.66366398e-01 1.66708037e-01 -8.15688968e-01 -5.29109538e-01
-4.91004646e-01 -1.26244938e+00 1.26038957e+00 2.46659651e-01
6.82541966e-01 -6.06870651e-01 -2.72229344e-01 2.46890903e-01
9.21075046e-01 1.53329328e-01 6.79221272e-01 2.71147639e-01
-1.05923605e+00 1.84084833e-01 -7.11007893e-01 7.45769143e-01
-1.27517059e-01 -8.50594789e-03 -1.34538913e+00 -4.71226484e-01
1.58772230e-01 2.85493955e-02 1.41483915e+00 5.82202077e-01
1.02848029e+00 -2.13436529e-01 2.61383832e-01 1.45152211e+00
1.72272801e+00 9.85172018e-02 1.19713616e+00 5.81395388e-01
8.89018416e-01 2.66034335e-01 3.64149451e-01 3.46879661e-01
2.41277158e-01 6.65926337e-01 8.27579260e-01 -4.53764021e-01
-5.01084745e-01 2.75854856e-01 4.62053925e-01 4.68354881e-01
-2.13491127e-01 -4.89819139e-01 -6.19404376e-01 4.81721342e-01
-1.32931495e+00 -6.21967912e-01 -2.42062926e-01 2.06859922e+00
5.48180103e-01 9.10072923e-02 -3.10684711e-01 -8.42151418e-02
4.75291163e-01 4.24261242e-01 -4.36513424e-01 -5.18002570e-01
-1.11428536e-01 6.44881189e-01 1.25113213e+00 6.09875739e-01
-1.23492455e+00 1.06152534e+00 6.14182186e+00 7.17984974e-01
-1.29646325e+00 1.94868878e-01 5.87533653e-01 -4.20875311e-01
5.88009283e-02 -2.66977876e-01 -8.62214684e-01 4.64106888e-01
1.20490158e+00 1.21322952e-01 5.69942772e-01 4.44179684e-01
2.87531346e-01 -3.84346068e-01 -7.53588498e-01 1.23922479e+00
1.03338361e-01 -1.57627809e+00 4.47444320e-02 1.20372109e-01
9.46044803e-01 5.52089334e-01 4.31421816e-01 -1.32646576e-01
1.84435099e-01 -1.08196974e+00 8.96672189e-01 1.27293363e-01
1.08615327e+00 -6.91935360e-01 7.95401573e-01 -1.30315229e-01
-1.13530159e+00 -8.06507915e-02 -6.82935476e-01 -1.99623883e-01
1.63178831e-01 8.26727211e-01 -6.05056763e-01 3.17501098e-01
1.16186309e+00 3.69814634e-01 -7.27813184e-01 1.11571550e+00
-7.76183903e-02 6.19732440e-01 -4.10912722e-01 4.54976320e-01
3.59166294e-01 1.01059906e-01 3.51311803e-01 1.40064991e+00
4.60847467e-01 3.64833593e-01 -4.17990208e-01 6.88428104e-01
-2.13238467e-02 -5.07497847e-01 -5.02660632e-01 3.80418986e-01
2.38722414e-01 1.00282979e+00 -4.92813915e-01 -2.82712609e-01
-7.33690560e-01 1.23304975e+00 5.25761619e-02 3.99603903e-01
-9.04133856e-01 -6.68217182e-01 1.04824638e+00 3.50584716e-01
8.40979040e-01 -4.36219484e-01 -3.28858346e-01 -1.13905764e+00
-6.96583986e-02 -7.34127998e-01 -2.66666850e-03 -1.00909865e+00
-9.14894044e-01 6.45275474e-01 -2.06051227e-02 -9.76732790e-01
1.67508468e-01 -9.52820778e-01 -5.26240468e-01 1.02474880e+00
-2.21836448e+00 -8.70351851e-01 -9.40373719e-01 6.90086365e-01
5.43905437e-01 1.65718019e-01 3.52200717e-01 6.81007683e-01
-6.04087591e-01 5.50162792e-01 2.25451812e-01 -1.30406283e-02
6.54437065e-01 -1.03227532e+00 9.59381163e-01 1.45008230e+00
-2.03544974e-01 2.07446799e-01 7.86084771e-01 -4.41664219e-01
-1.29362893e+00 -1.49817693e+00 3.53465229e-01 -2.02565432e-01
2.93613613e-01 -3.78052682e-01 -9.50414300e-01 5.82222104e-01
3.66630644e-01 4.38052922e-01 2.81860471e-01 -6.18458748e-01
-4.43556219e-01 -3.14431638e-02 -9.20495689e-01 6.10224009e-01
7.28390694e-01 -4.23868835e-01 -1.86549097e-01 4.54970524e-02
1.09679520e+00 -7.94751763e-01 -5.44099092e-01 2.74383783e-01
5.35008729e-01 -1.58813131e+00 1.00293493e+00 -1.76013205e-02
6.79317117e-01 -5.53944051e-01 -1.43108413e-01 -1.24099362e+00
-4.47804421e-01 -6.96579218e-01 -1.87742352e-01 4.92660522e-01
2.93898374e-01 -6.24431610e-01 8.21448088e-01 4.12869930e-01
-5.43508470e-01 -6.50711298e-01 -7.34247625e-01 -7.14818180e-01
1.68023957e-03 -3.62832785e-01 6.11208856e-01 5.48080385e-01
-8.70910943e-01 3.30653563e-02 -6.22407973e-01 7.55169868e-01
7.52238691e-01 -2.24937394e-01 7.66702652e-01 -6.17937863e-01
-3.64817768e-01 -1.53562874e-01 -6.10538006e-01 -1.19440877e+00
-5.27943850e-01 -2.72889256e-01 2.90155530e-01 -1.31939733e+00
-2.12587088e-01 -5.93185686e-02 -1.53097466e-01 2.07077175e-01
-3.14493746e-01 7.89316833e-01 2.88624108e-01 1.60266623e-01
-2.82486349e-01 3.94062281e-01 1.09116066e+00 -9.11948010e-02
-2.15282276e-01 -2.97841698e-01 -3.61421466e-01 5.26462257e-01
7.66097009e-01 -4.44868445e-01 -3.74675184e-01 -1.05493867e+00
2.18809620e-01 -3.59194279e-01 6.48895919e-01 -1.35114098e+00
5.66554070e-01 2.12786525e-01 4.70663071e-01 -3.58854115e-01
4.60433036e-01 -6.58799708e-01 3.01608257e-02 4.32761908e-01
1.08782642e-01 2.62630045e-01 3.99897307e-01 5.06627202e-01
-2.86263227e-01 -1.55671043e-02 1.19379306e+00 1.02519937e-01
-9.18493867e-01 4.90481406e-01 -4.82394993e-01 -1.38919160e-01
8.03973496e-01 -4.23448503e-01 -7.27279902e-01 -3.95346552e-01
-2.63323128e-01 -1.94185111e-03 6.05664253e-01 1.68694034e-01
8.87013435e-01 -7.74411201e-01 -8.72489572e-01 4.46674973e-01
1.04748353e-01 3.81755352e-01 4.32759911e-01 5.86713612e-01
-1.44243646e+00 2.34654397e-01 -3.27638328e-01 -4.15301025e-01
-1.17765605e+00 5.17789125e-01 5.50430179e-01 7.22199753e-02
-8.21155548e-01 1.12418485e+00 6.27664328e-01 5.32626748e-01
7.22701922e-02 -5.54041803e-01 2.45314389e-01 -5.17782509e-01
9.45658743e-01 4.05964881e-01 4.57709759e-01 -3.81370097e-01
6.54700212e-03 3.17413211e-01 -1.60814464e-01 -7.83350542e-02
1.32970846e+00 -3.58819723e-01 -1.35865763e-01 -4.32329020e-03
1.23408186e+00 -1.72988907e-01 -1.75022769e+00 -1.32577121e-01
-5.19462943e-01 -7.76326597e-01 8.53801906e-01 -3.91296178e-01
-1.46362531e+00 1.14844227e+00 9.95183170e-01 -1.12707108e-01
1.57866895e+00 -5.15330017e-01 1.15734565e+00 3.51017714e-01
1.79188460e-01 -6.87001050e-01 -1.40244767e-01 4.60390151e-01
4.80694801e-01 -1.20879138e+00 1.18683241e-01 -4.67978835e-01
-3.06491435e-01 1.13237786e+00 6.53381348e-01 -1.98813111e-01
7.59504497e-01 5.36780953e-01 1.63616240e-01 -3.64748478e-01
-6.93959713e-01 -2.99851447e-01 1.94534630e-01 5.90837955e-01
8.37072060e-02 -1.67089596e-01 2.76822358e-01 -1.72048241e-01
-3.27013284e-01 -4.06252816e-02 9.85375583e-01 8.95320058e-01
-6.91746831e-01 -5.42646587e-01 -4.45540816e-01 3.07081878e-01
-5.02565801e-01 -7.04996407e-01 4.54851612e-02 6.35960281e-01
4.68696266e-01 1.06442893e+00 4.58565325e-01 -5.05183578e-01
1.78511888e-01 -4.74903047e-01 5.39058447e-01 -4.67404753e-01
-5.94273329e-01 -9.14580822e-02 -3.62508893e-02 -6.91650331e-01
-1.49730429e-01 -1.06057569e-01 -7.41605043e-01 -7.33601093e-01
-3.66558135e-01 -3.25611830e-01 7.69874096e-01 5.44350088e-01
5.05539834e-01 6.78447723e-01 4.67303753e-01 -1.22859848e+00
-1.27831548e-01 -6.84877932e-01 -7.81721652e-01 -5.15937954e-02
1.01631045e+00 -4.44192827e-01 -4.71376657e-01 1.96742669e-01]
|
[10.944223403930664, -3.0784668922424316]
|
f478ea61-6621-4613-a95c-22ab23d00925
|
disentangled-representation-learning-for-1
|
2211.00437
| null |
https://arxiv.org/abs/2211.00437v3
|
https://arxiv.org/pdf/2211.00437v3.pdf
|
Disentangled representation learning for multilingual speaker recognition
|
The goal of this paper is to learn robust speaker representation for bilingual speaking scenario. The majority of the world's population speak at least two languages; however, most speaker recognition systems fail to recognise the same speaker when speaking in different languages. Popular speaker recognition evaluation sets do not consider the bilingual scenario, making it difficult to analyse the effect of bilingual speakers on speaker recognition performance. In this paper, we publish a large-scale evaluation set named VoxCeleb1-B derived from VoxCeleb that considers bilingual scenarios. We introduce an effective disentanglement learning strategy that combines adversarial and metric learning-based methods. This approach addresses the bilingual situation by disentangling language-related information from speaker representation while ensuring stable speaker representation learning. Our language-disentangled learning method only uses language pseudo-labels without manual information.
|
['Jee-weon Jung', 'Hee Soo Heo', 'Jaesung Huh', 'Joon Son Chung', 'Youkyum Kim', 'Kihyun Nam']
|
2022-11-01
| null | null | null | null |
['speaker-recognition']
|
['speech']
|
[-2.97048688e-02 -2.16575954e-02 -2.96909481e-01 -5.40053964e-01
-1.20172715e+00 -8.19831192e-01 9.51033294e-01 -4.78495598e-01
-4.51633602e-01 8.40522945e-01 4.73462641e-01 -4.05157566e-01
1.72439918e-01 -4.27846313e-01 -4.13877964e-01 -8.30931604e-01
4.07989621e-02 5.15506864e-01 -4.33132678e-01 -4.53169286e-01
-1.12915501e-01 6.59792304e-01 -1.21362448e+00 2.00283453e-01
7.75502980e-01 2.05496266e-01 -2.45048210e-01 6.65689290e-01
2.92789168e-03 5.16381264e-01 -8.45938802e-01 -7.11456180e-01
5.71365058e-01 -5.30081332e-01 -6.31488442e-01 -1.05848491e-01
7.74859607e-01 -1.85414016e-01 -4.98593062e-01 9.34045613e-01
8.93348515e-01 -2.62273729e-01 7.81762779e-01 -1.39517474e+00
-5.17602563e-01 9.08130407e-01 -3.76802415e-01 4.53816265e-01
6.08015120e-01 1.83208045e-02 9.91922736e-01 -7.11742580e-01
2.84025252e-01 1.45569193e+00 4.41980600e-01 7.72218406e-01
-1.43768275e+00 -1.12553024e+00 1.96629360e-01 2.22212315e-01
-1.79080904e+00 -8.98053586e-01 1.04933035e+00 -5.04204214e-01
4.79139477e-01 6.32004976e-01 1.91418275e-01 1.53268754e+00
-1.91079974e-01 6.65354669e-01 1.65668869e+00 -6.25371218e-01
-1.75928250e-01 7.51437783e-01 1.44400537e-01 5.87618053e-01
7.51680285e-02 6.89103186e-01 -9.33707833e-01 -8.84328261e-02
2.76807725e-01 -4.11007702e-01 -3.30926299e-01 -3.34843963e-01
-1.57387912e+00 7.97937751e-01 -1.25716180e-02 4.25000042e-01
-3.04792244e-02 -3.38469684e-01 4.18722361e-01 8.13319147e-01
2.52781332e-01 1.10687390e-01 -1.21751785e-01 1.35248303e-01
-1.26018095e+00 1.03437133e-01 8.83631110e-01 9.52045739e-01
5.15766799e-01 4.31727409e-01 -2.06955180e-01 8.13326120e-01
5.26720166e-01 9.08864975e-01 6.51737690e-01 -3.28953117e-01
6.23645306e-01 2.01987371e-01 -2.87147194e-01 -8.23483288e-01
-4.46327254e-02 -4.56014127e-01 -7.44225085e-01 2.68577456e-01
4.56988245e-01 -6.88678175e-02 -6.51668429e-01 2.04585528e+00
8.00819993e-02 1.19089104e-01 5.06211698e-01 8.82380426e-01
1.00921357e+00 4.82452244e-01 -4.00589332e-02 -5.62133566e-02
1.35406041e+00 -8.19715083e-01 -5.36230981e-01 -4.24239561e-02
1.35517642e-01 -9.05722320e-01 9.29667592e-01 2.13821247e-01
-7.52074838e-01 -3.09091985e-01 -1.31734669e+00 1.00007616e-01
-2.96931058e-01 -5.27845463e-03 5.65494835e-01 1.45693874e+00
-1.09953690e+00 -5.06520905e-02 -5.41378260e-01 -2.28517368e-01
1.48734748e-01 5.27288020e-01 -7.72469938e-01 1.79426000e-02
-1.52614963e+00 1.25120342e+00 1.13876589e-01 1.36423307e-02
-1.48287165e+00 -3.75221968e-01 -8.84884477e-01 -1.96686104e-01
9.86974686e-03 -3.60930592e-01 1.04484165e+00 -1.10034120e+00
-1.68081832e+00 1.28766143e+00 -3.45929503e-01 -3.30593348e-01
9.14269149e-01 3.44292313e-01 -9.22944725e-01 -2.79203534e-01
-5.87006249e-02 2.73483515e-01 9.74406183e-01 -1.39733481e+00
-1.97235197e-01 -5.10505974e-01 -7.51947984e-02 3.63088429e-01
-8.54377076e-02 3.16558421e-01 6.63316026e-02 -7.30406165e-01
7.05366358e-02 -7.37097263e-01 2.79552609e-01 -3.75095338e-01
-5.65347314e-01 1.34286746e-01 4.91344213e-01 -9.77171481e-01
7.73140252e-01 -2.30073786e+00 3.48435909e-01 2.33997479e-01
4.79518734e-02 1.45976692e-01 -3.52930725e-01 4.33945209e-01
-2.85829693e-01 1.06234044e-01 -2.69877613e-01 -6.00052714e-01
2.06874579e-01 8.24465379e-02 -5.04691064e-01 8.21258724e-01
-6.62724152e-02 4.96135890e-01 -6.89250886e-01 -6.44987047e-01
7.12766275e-02 6.56651556e-01 -3.64793181e-01 1.26829281e-01
4.01174307e-01 6.85392559e-01 -1.17287308e-01 7.26787269e-01
7.68356860e-01 7.32967377e-01 1.72871798e-01 -1.65898681e-01
1.31499544e-01 5.69242239e-01 -1.18975794e+00 1.63953567e+00
-6.91203833e-01 8.96661580e-01 2.36391410e-01 -9.35383677e-01
9.98739064e-01 7.41166770e-01 2.03521419e-02 -5.99481106e-01
2.29688492e-02 3.09077084e-01 3.16812664e-01 -3.29280406e-01
6.82305619e-02 -7.94129491e-01 -2.08787888e-01 4.45846289e-01
2.18755603e-01 -7.84578621e-02 -1.32136270e-01 1.60477638e-01
5.06273270e-01 -1.58692330e-01 4.65918094e-01 -4.29583251e-01
9.54118967e-01 -5.58225691e-01 6.42330766e-01 4.84100074e-01
-6.92951262e-01 4.91637379e-01 3.17597836e-01 -4.79799882e-03
-6.92581832e-01 -1.67275119e+00 -3.07918817e-01 1.04343903e+00
-1.00399233e-01 8.24651048e-02 -6.81869149e-01 -6.69297934e-01
2.67735653e-04 9.30570364e-01 -4.06384200e-01 -3.21325541e-01
-4.67969865e-01 -7.23307788e-01 1.25996518e+00 2.83683445e-02
3.83151680e-01 -7.37115204e-01 1.89831257e-01 -2.89279133e-01
-1.95974976e-01 -8.15337598e-01 -8.04566801e-01 -1.86491638e-01
-4.73293155e-01 -9.76240993e-01 -7.97926605e-01 -9.51623917e-01
4.40924793e-01 2.68977112e-03 1.15108323e+00 -4.92896587e-01
3.35261300e-02 3.05167556e-01 -5.75365722e-02 -3.09500784e-01
-9.87611294e-01 6.62963986e-02 4.97849047e-01 3.20090026e-01
7.35322118e-01 -4.92412806e-01 -1.35671377e-01 5.65084100e-01
-4.92708325e-01 -7.41783455e-02 4.85279351e-01 7.56655514e-01
-4.92393598e-02 -2.11341634e-01 8.87275159e-01 -7.18778372e-01
5.36897779e-01 -3.70178729e-01 -3.89607430e-01 6.31905854e-01
-3.34339857e-01 2.21101463e-01 2.67236471e-01 -6.01853967e-01
-9.17703450e-01 -9.85788479e-02 -6.51137829e-02 -2.92249508e-02
-1.65179983e-01 2.33238131e-01 -7.64743388e-01 -2.33193934e-02
6.64758265e-01 6.05162024e-01 1.11199446e-01 -3.34453791e-01
4.67779458e-01 1.22136104e+00 4.84404802e-01 -5.54330647e-01
1.06183851e+00 2.78796554e-01 -5.10230720e-01 -9.98925984e-01
-2.85679281e-01 -4.12615150e-01 -6.37488782e-01 -1.51492640e-01
6.82254076e-01 -1.27245355e+00 -6.82617724e-01 4.41437393e-01
-1.01666212e+00 2.50717849e-01 -4.91275676e-02 7.72066653e-01
-4.22551334e-01 2.13632613e-01 -3.09406042e-01 -9.22027767e-01
-2.30362609e-01 -1.42176795e+00 9.67004955e-01 -3.08590502e-01
-2.15415835e-01 -9.83038306e-01 3.59526664e-01 7.85333633e-01
4.28252816e-01 9.69565809e-02 6.70752168e-01 -1.14788389e+00
-4.42195177e-01 -2.67885447e-01 1.75161317e-01 5.19477606e-01
5.09728372e-01 -4.11396176e-01 -1.37435246e+00 -6.22522354e-01
9.79993045e-02 -1.15956634e-01 7.19346285e-01 -1.46013379e-01
2.05374330e-01 -5.12394726e-01 3.27462368e-02 5.32178998e-01
1.06783557e+00 -7.43953511e-02 4.36220497e-01 -8.03500339e-02
8.86287928e-01 9.39172208e-01 -3.62811059e-01 -2.44133979e-01
3.44824672e-01 8.44924033e-01 -1.81225434e-01 1.55596077e-01
-2.86204755e-01 -3.66820753e-01 8.52864265e-01 1.23269558e+00
2.08346784e-01 -6.67316914e-02 -9.60008323e-01 4.41803604e-01
-1.12765670e+00 -1.17591655e+00 2.98584044e-01 2.51461935e+00
1.02004004e+00 -1.87801987e-01 2.63108969e-01 3.24176043e-01
8.41126204e-01 2.31013268e-01 -2.44845450e-01 -3.16068709e-01
-4.35570568e-01 -1.07140481e-01 4.61211711e-01 1.12426162e+00
-7.42588341e-01 8.07494402e-01 6.73337841e+00 6.05974972e-01
-1.33156919e+00 3.81202012e-01 3.43807906e-01 -2.61605531e-01
-6.23920381e-01 -9.92241800e-02 -7.07194090e-01 2.48989061e-01
1.00943410e+00 -6.81791544e-01 6.28313243e-01 4.56935376e-01
9.02311727e-02 5.54109812e-01 -1.54206777e+00 1.23490155e+00
7.76258409e-01 -5.90335727e-01 2.52941996e-01 8.39294344e-02
4.72556800e-01 1.43994704e-01 2.71971971e-01 4.87422884e-01
3.14844161e-01 -1.36009812e+00 9.56089973e-01 2.36075729e-01
1.01599789e+00 -6.80834949e-01 5.50174534e-01 3.50617856e-01
-7.72309244e-01 1.91657484e-01 1.78296253e-01 2.14927465e-01
6.08742647e-02 2.71543831e-01 -1.05399621e+00 4.79407579e-01
1.99902207e-01 3.31624925e-01 -5.26291728e-01 5.88601828e-01
-3.09189469e-01 7.96734869e-01 -1.38623580e-01 1.76695809e-01
-3.07810336e-01 -1.69341698e-01 1.01044500e+00 1.16015339e+00
1.46854758e-01 -3.78175527e-01 6.24704082e-03 8.13378096e-01
-1.79220721e-01 2.83609718e-01 -9.45957243e-01 -1.45902857e-01
5.74589968e-01 7.66835630e-01 -2.12690875e-01 -1.32168993e-01
-4.69978690e-01 1.01225138e+00 1.94615602e-01 5.51566005e-01
-4.98380333e-01 -4.52080667e-02 7.30137050e-01 -9.68832672e-02
-2.28541344e-01 -1.78282380e-01 -2.53754944e-01 -1.47808528e+00
-4.10164259e-02 -1.46476293e+00 1.51123270e-01 -1.58837348e-01
-1.41169000e+00 8.85217845e-01 9.25014615e-02 -1.09905303e+00
-4.97568071e-01 -4.33757275e-01 -4.43361074e-01 1.40147483e+00
-1.33059525e+00 -1.37554586e+00 1.55787081e-01 6.02340996e-01
4.19520974e-01 -9.88319755e-01 1.08773315e+00 5.02723813e-01
-6.56556368e-01 1.20649433e+00 1.16581321e-01 2.33861491e-01
8.58931065e-01 -1.23926497e+00 2.36064956e-01 9.25149798e-01
5.64962447e-01 8.91116679e-01 7.85664976e-01 -2.41898075e-01
-1.16233730e+00 -5.90686202e-01 1.31720078e+00 -7.60132194e-01
4.34143871e-01 -9.32052910e-01 -6.95554078e-01 8.64766896e-01
4.81557786e-01 -2.36847624e-01 1.15739787e+00 2.75318742e-01
-9.59187686e-01 -4.28707212e-01 -1.37185764e+00 6.88279927e-01
7.57139325e-01 -1.26035953e+00 -9.14615929e-01 1.59542114e-01
5.07125378e-01 -1.79569051e-02 -5.66380918e-01 9.92378816e-02
5.48666000e-01 -7.10927188e-01 1.05763054e+00 -6.60645425e-01
-1.35514185e-01 -3.37518811e-01 -5.61524332e-01 -1.40123212e+00
1.05649255e-01 -6.66272521e-01 1.78625599e-01 1.53773630e+00
6.41440153e-01 -9.78272319e-01 4.95365560e-01 3.51657927e-01
3.42208862e-01 -8.19545332e-03 -1.46062493e+00 -9.27418232e-01
5.19681156e-01 -1.70920894e-01 8.69265676e-01 1.30859613e+00
3.39128710e-02 5.46748638e-01 -5.57242513e-01 4.00812030e-01
8.59345853e-01 -8.02943856e-02 7.57708907e-01 -9.94814813e-01
-3.44890773e-01 -6.34597898e-01 -6.88913882e-01 -3.91146392e-01
7.17577577e-01 -1.40005589e+00 -8.64491761e-02 -9.96565044e-01
3.38984519e-01 -2.58047342e-01 -4.58017945e-01 1.62805587e-01
-8.21576416e-02 1.96626544e-01 1.03118718e-01 1.47612453e-01
7.78735355e-02 6.19734704e-01 9.19161260e-01 -6.10487700e-01
1.20033152e-01 8.89029875e-02 -8.00555050e-01 3.17504197e-01
7.38088906e-01 -4.59421992e-01 -6.01084113e-01 -3.21740031e-01
-2.41857663e-01 -3.71147459e-03 3.60148966e-01 -6.89055145e-01
-7.30600907e-03 2.05902755e-02 -2.01722924e-02 6.36054799e-02
1.51348963e-01 -8.63787830e-01 1.38558999e-01 3.80654961e-01
-5.26750803e-01 -1.08140125e-03 2.07314208e-01 3.54476929e-01
-4.06330973e-01 -1.36762321e-01 7.24883318e-01 7.31280670e-02
-2.60394335e-01 8.42716470e-02 -3.79925281e-01 7.87746012e-02
9.33810532e-01 -4.39477749e-02 -4.15251330e-02 -3.59643877e-01
-6.91849828e-01 1.45224920e-02 3.08655381e-01 7.69023061e-01
2.60614783e-01 -1.48509383e+00 -1.39441597e+00 5.68225920e-01
3.02500367e-01 -5.28867245e-01 1.03417814e-01 4.92488444e-01
-5.29789388e-01 4.66939270e-01 -1.14952065e-01 -3.42724264e-01
-1.43880260e+00 5.03477573e-01 7.26976395e-01 3.61454301e-02
-1.22428164e-01 8.79038811e-01 1.56898290e-01 -1.06860280e+00
3.89063627e-01 1.93344161e-01 -1.33529529e-01 5.81959933e-02
5.51506400e-01 3.86850566e-01 9.92375538e-02 -1.48146260e+00
-6.43607855e-01 2.25140452e-01 -2.89629132e-01 -7.71884143e-01
7.41918564e-01 -1.23911947e-01 4.11283709e-02 7.56065845e-01
1.44735360e+00 5.37407935e-01 -7.14951396e-01 -3.18499148e-01
-1.16385877e-01 -4.50304806e-01 9.17691141e-02 -9.60191965e-01
-9.32156920e-01 8.85633111e-01 9.95069802e-01 -1.58737138e-01
8.25370491e-01 -5.25440983e-02 2.51867920e-01 7.18248338e-02
4.93231833e-01 -6.32511556e-01 -3.96011174e-01 1.03870697e-01
1.19484794e+00 -1.52066886e+00 -1.60749435e-01 -3.86557132e-01
-6.91388965e-01 4.15775537e-01 3.62320989e-01 2.44992137e-01
6.31198823e-01 1.98363420e-02 7.12841928e-01 1.18611053e-01
-2.52702773e-01 1.17544577e-01 5.90612590e-01 7.86319137e-01
7.13804424e-01 5.06338298e-01 -2.07151324e-01 4.43261176e-01
-8.91883373e-01 -6.10035419e-01 2.03541741e-01 5.34481704e-01
1.93351299e-01 -1.41075742e+00 -6.32052600e-01 2.98794657e-02
-4.39975291e-01 -2.64172286e-01 -6.22554302e-01 6.33360445e-01
-4.11408357e-02 1.15351450e+00 -1.65704414e-01 -3.46315414e-01
3.84094149e-01 3.65634412e-01 5.99747777e-01 -5.22991836e-01
-4.99472231e-01 -3.48246038e-01 6.38334379e-02 -1.21738903e-01
-4.85467881e-01 -9.28746462e-01 -6.78331733e-01 -4.99160409e-01
-2.25756541e-01 2.84707248e-01 7.80710399e-01 9.50427234e-01
-1.00219525e-01 3.20681423e-01 8.01776767e-01 -5.89773536e-01
-9.27870512e-01 -1.12322044e+00 -7.21496403e-01 4.88641500e-01
8.58601570e-01 -5.12762010e-01 -6.00320280e-01 4.23766999e-03]
|
[14.309356689453125, 6.19891357421875]
|
dd6d31d0-03b3-4fb5-9ba3-8c2f1864d1a3
|
artificial-neuronal-ensembles-with-learned
|
2301.07187
| null |
https://arxiv.org/abs/2301.07187v2
|
https://arxiv.org/pdf/2301.07187v2.pdf
|
Artificial Neuronal Ensembles with Learned Context Dependent Gating
|
Biological neural networks are capable of recruiting different sets of neurons to encode different memories. However, when training artificial neural networks on a set of tasks, typically, no mechanism is employed for selectively producing anything analogous to these neuronal ensembles. Further, artificial neural networks suffer from catastrophic forgetting, where the network's performance rapidly deteriorates as tasks are learned sequentially. By contrast, sequential learning is possible for a range of biological organisms. We introduce Learned Context Dependent Gating (LXDG), a method to flexibly allocate and recall `artificial neuronal ensembles', using a particular network structure and a new set of regularization terms. Activities in the hidden layers of the network are modulated by gates, which are dynamically produced during training. The gates are outputs of networks themselves, trained with a sigmoid output activation. The regularization terms we have introduced correspond to properties exhibited by biological neuronal ensembles. The first term penalizes low gate sparsity, ensuring that only a specified fraction of the network is used. The second term ensures that previously learned gates are recalled when the network is presented with input from previously learned tasks. Finally, there is a regularization term responsible for ensuring that new tasks are encoded in gates that are as orthogonal as possible from previously used ones. We demonstrate the ability of this method to alleviate catastrophic forgetting on continual learning benchmarks. When the new regularization terms are included in the model along with Elastic Weight Consolidation (EWC) it achieves better performance on the benchmark `permuted MNIST' than with EWC alone. The benchmark `rotated MNIST' demonstrates how similar tasks recruit similar neurons to the artificial neuronal ensemble.
|
['David J. Freedman', 'Matthew J. Tilley', 'Michelle Miller']
|
2023-01-17
| null | null | null | null |
['rotated-mnist']
|
['computer-vision']
|
[ 6.98803544e-01 2.01259796e-02 2.90628016e-01 -3.66329439e-02
2.01822653e-01 -4.66260642e-01 6.12475574e-01 1.27650082e-01
-8.21836472e-01 1.19505143e+00 1.34629449e-02 1.61468722e-02
-2.43037730e-01 -1.01527274e+00 -1.09587920e+00 -1.26088476e+00
-2.07196057e-01 3.39947671e-01 4.83735472e-01 -3.29962969e-01
2.80961066e-01 6.43239021e-01 -1.90455389e+00 4.59923774e-01
5.26930332e-01 7.54054308e-01 5.86379826e-01 5.55191159e-01
-2.45777033e-02 7.12096512e-01 -6.29926145e-01 1.73330277e-01
6.41882047e-02 -5.16011894e-01 -6.73817158e-01 -3.29536162e-02
2.80492127e-01 3.17569166e-01 -3.74169737e-01 6.42458498e-01
3.07200521e-01 5.52038252e-01 7.29698539e-01 -6.88005805e-01
-8.62004638e-01 5.16483784e-01 -2.39296015e-02 5.79410315e-01
-2.27571502e-02 3.56006205e-01 6.76994979e-01 -9.94033635e-01
7.52588987e-01 8.95574450e-01 7.41056383e-01 8.64212513e-01
-1.80396223e+00 -5.76584339e-01 3.17582041e-01 -3.38330060e-01
-1.16489601e+00 -5.49677134e-01 3.75475496e-01 -2.70910591e-01
1.51463807e+00 1.30528912e-01 9.68514204e-01 1.28533161e+00
8.52158427e-01 3.02870840e-01 1.10282576e+00 -4.25436258e-01
5.39030671e-01 6.50826097e-03 1.35683492e-01 6.12500429e-01
5.90694070e-01 5.08679986e-01 -8.88996005e-01 -6.20500334e-02
8.09020817e-01 2.97943562e-01 -2.99827784e-01 -2.15991616e-01
-1.11379981e+00 5.54601371e-01 4.01614696e-01 7.39160597e-01
-5.43336332e-01 4.11952794e-01 1.17677465e-01 6.92744792e-01
2.32196927e-01 8.47029030e-01 -5.80247462e-01 4.35901433e-01
-8.37834120e-01 1.39605582e-01 5.99596560e-01 6.51496828e-01
9.87907350e-01 3.72154474e-01 -1.31595209e-01 8.71576548e-01
-6.63687885e-02 3.98793131e-01 8.35104644e-01 -7.41421521e-01
1.84589271e-02 6.24106765e-01 -2.62018412e-01 -8.34078193e-01
-3.38883907e-01 -7.09528267e-01 -1.15467393e+00 3.42607439e-01
5.44525683e-01 7.96113536e-02 -1.28408968e+00 2.26449990e+00
-3.37750673e-01 1.91092744e-01 1.52654037e-01 4.42043334e-01
4.56029773e-01 6.37397230e-01 4.13008600e-01 -3.46766084e-01
9.97000635e-01 -5.06192565e-01 -4.80072677e-01 -6.74539864e-01
4.81255561e-01 -2.78831422e-01 9.87529516e-01 3.99324417e-01
-1.24902368e+00 -6.26449704e-01 -1.25673664e+00 1.95745662e-01
-8.21072519e-01 -4.41910505e-01 5.94426394e-01 3.31388414e-01
-1.55189586e+00 8.61916482e-01 -7.65906453e-01 -3.61560494e-01
5.56803524e-01 6.44929409e-01 -5.13436794e-01 1.49542503e-02
-1.25243521e+00 1.07239544e+00 6.86488152e-01 -3.73724774e-02
-1.23115385e+00 -4.81725514e-01 -6.17448032e-01 2.81498969e-01
-6.28500432e-02 -9.77601588e-01 7.05670655e-01 -1.20562947e+00
-1.19429243e+00 9.12047863e-01 -7.21679926e-02 -5.21651387e-01
-1.40644267e-01 1.31922275e-01 -3.61589938e-01 -1.64309204e-01
-1.79538131e-01 9.14032161e-01 1.04996049e+00 -1.42256391e+00
-2.98488110e-01 -3.12855065e-01 -2.05527708e-01 6.04763580e-03
-3.94136757e-01 -4.48511779e-01 -4.67520691e-02 -1.02140951e+00
2.78075695e-01 -1.08678067e+00 -3.18763912e-01 -2.17039019e-01
3.07131354e-02 1.85320437e-01 5.79190075e-01 -1.36168018e-01
1.30886781e+00 -2.15196037e+00 6.31380677e-01 4.05370772e-01
2.00601310e-01 1.89193442e-01 -4.69975144e-01 2.96358615e-01
-3.92817944e-01 1.52755350e-01 -7.30440080e-01 -1.60183117e-01
-4.25172120e-01 7.78456151e-01 -3.10461193e-01 5.45543171e-02
4.85822827e-01 8.67697597e-01 -9.21638072e-01 1.69037029e-01
-4.06283081e-01 4.97403204e-01 -5.67503154e-01 -7.22087696e-02
-2.50742733e-01 3.16700310e-01 4.12233472e-02 3.85656804e-01
2.91109473e-01 -3.10881108e-01 3.49991232e-01 4.13399965e-01
-3.77789363e-02 2.25318477e-01 -9.07461345e-01 1.66593027e+00
-3.73924345e-01 4.02482271e-01 -2.00426519e-01 -1.02482355e+00
9.24467862e-01 3.80390584e-01 -2.62270980e-02 -7.61469603e-01
7.71903619e-02 3.82643849e-01 3.14000517e-01 -1.44188136e-01
2.60712206e-01 -3.73425990e-01 3.35582695e-03 7.25139380e-01
5.39329410e-01 -5.02343569e-03 4.58066285e-01 2.08196774e-01
1.62065399e+00 -1.95455402e-01 6.10828027e-02 -5.67749858e-01
3.78958553e-01 -1.42080441e-01 6.21973813e-01 1.19949079e+00
1.79670230e-02 3.56614351e-01 3.39820087e-01 -7.12946236e-01
-1.09700751e+00 -1.16574681e+00 -1.43477041e-02 1.53416812e+00
-2.98105061e-01 5.56842983e-02 -4.92375016e-01 -4.28136408e-01
1.19805247e-01 6.77924335e-01 -1.05687785e+00 -8.34568083e-01
-7.25989044e-01 -1.02124894e+00 5.82889378e-01 5.62106192e-01
3.17532569e-01 -1.73223591e+00 -7.73046076e-01 4.83617038e-01
2.59297192e-01 -3.11166644e-01 -2.67275333e-01 1.23195064e+00
-1.42982244e+00 -9.65432465e-01 -4.75303471e-01 -1.09299505e+00
1.16553378e+00 2.91134506e-01 1.35279357e+00 5.84161699e-01
-2.31339306e-01 1.71144977e-01 4.03282009e-02 -1.53245375e-01
-2.59382695e-01 2.67045856e-01 2.86520183e-01 -2.38037437e-01
1.01894177e-01 -1.14649522e+00 -4.43447173e-01 4.92149442e-02
-1.50010324e+00 1.19393906e-02 7.85379350e-01 1.38886213e+00
6.98206723e-01 1.53412949e-03 8.36979091e-01 -1.28125870e+00
6.80970430e-01 -5.15887499e-01 -1.80514559e-01 1.07741691e-01
-6.75959051e-01 6.94913566e-01 6.36972725e-01 -8.44508111e-01
-9.00544941e-01 6.49792850e-02 4.08073403e-02 -8.14100951e-02
-2.80432496e-02 4.95396495e-01 7.04521239e-02 -6.94194436e-02
1.01982093e+00 6.16069794e-01 -2.67200153e-02 -1.47879258e-01
4.05450612e-02 -2.46984109e-01 4.79010820e-01 -8.08959961e-01
5.55431247e-01 1.98642969e-01 1.13143355e-01 -4.21739578e-01
-6.49026155e-01 1.07515492e-01 -6.66051328e-01 -1.87743500e-01
5.37153184e-01 -5.62079966e-01 -4.31426376e-01 5.92593968e-01
-1.00197637e+00 -6.34967029e-01 -7.68002331e-01 2.35035464e-01
-5.40797710e-01 -4.96253669e-01 -6.84292674e-01 -5.39401054e-01
-8.33547637e-02 -8.61303091e-01 4.47462708e-01 2.46990234e-01
-3.36472929e-01 -1.11910713e+00 3.01133066e-01 -5.07187188e-01
7.39928365e-01 9.11653042e-02 1.35253751e+00 -6.04410112e-01
-3.67007524e-01 2.63910532e-01 1.19210772e-01 4.63609785e-01
1.50441229e-01 -2.72030652e-01 -9.75007057e-01 -6.52011633e-01
-1.47518581e-02 -3.02728415e-01 1.60911667e+00 1.77352026e-01
1.05361271e+00 -4.74067211e-01 -4.38187599e-01 4.06727582e-01
1.48749745e+00 3.81871879e-01 7.96677887e-01 3.43201190e-01
2.10748643e-01 5.35095334e-01 -2.10561812e-01 -6.89348131e-02
-3.43584865e-01 -2.16603391e-02 3.25298727e-01 2.08359808e-01
-7.38721788e-02 -1.61542185e-02 2.92593122e-01 1.06077504e+00
-4.69439775e-01 -2.05338687e-01 -9.41373944e-01 5.90650856e-01
-1.72948766e+00 -1.10003042e+00 3.36950719e-01 2.31684947e+00
1.18366575e+00 5.57489634e-01 -5.01470029e-01 3.14362288e-01
4.54923630e-01 6.57573342e-02 -6.56324267e-01 -4.64685321e-01
-6.06977999e-01 9.34765339e-01 3.82859230e-01 4.81279075e-01
-6.75271332e-01 8.68744671e-01 7.02459669e+00 5.09916306e-01
-1.06252813e+00 1.44384354e-01 6.09590352e-01 -3.40318888e-01
-4.50046510e-01 -4.58620787e-02 -7.01961994e-01 4.12275881e-01
1.22313452e+00 1.10183455e-01 8.24363649e-01 3.51390004e-01
-1.43089533e-01 -3.01394284e-01 -1.04208624e+00 5.14942646e-01
-1.09019049e-01 -1.53443658e+00 2.87242740e-01 -2.63230167e-02
9.09824252e-01 -1.06280617e-01 3.33046049e-01 5.09060204e-01
4.63845283e-01 -1.22561574e+00 4.98234570e-01 1.12921143e+00
5.83653450e-01 -8.29235017e-01 4.33034211e-01 3.55466306e-01
-7.24659383e-01 -2.88554668e-01 -5.88162184e-01 -1.70642689e-01
-3.80458772e-01 5.40585458e-01 -3.82288128e-01 -2.31299922e-01
6.37206674e-01 4.23224032e-01 -7.50874221e-01 8.00186276e-01
-2.28466362e-01 4.94440407e-01 -1.56956881e-01 1.57867402e-01
1.11644745e-01 3.30468453e-02 4.71624434e-01 1.20023918e+00
4.20152038e-01 2.84241766e-01 -2.65396506e-01 7.86082327e-01
-3.86400908e-01 -2.94083327e-01 -9.99270976e-01 -9.47387218e-02
4.62748408e-01 9.33027744e-01 -9.26462710e-01 -2.63082922e-01
1.08160004e-01 9.12084818e-01 7.51585186e-01 6.87778473e-01
-4.10595536e-01 -2.45450422e-01 4.52409059e-01 3.09915721e-01
3.22395980e-01 -3.21818531e-01 -3.68047267e-01 -7.43580222e-01
-3.63552451e-01 -6.94926679e-01 1.92226499e-01 -7.83988357e-01
-1.23282957e+00 7.91277468e-01 -3.98135453e-01 -6.80509925e-01
-3.60235982e-02 -6.48543060e-01 -4.93040085e-01 8.32760572e-01
-1.23499966e+00 -6.31711543e-01 -8.88693184e-02 6.02210462e-01
2.43051559e-01 -4.40837801e-01 1.24396741e+00 1.96551040e-01
-5.19491732e-01 3.12693477e-01 2.31496751e-01 -2.83313960e-01
5.78214049e-01 -1.09499478e+00 1.01096660e-01 6.91319644e-01
1.88713223e-01 1.13322687e+00 6.50399506e-01 -6.68659985e-01
-1.05410373e+00 -1.32063830e+00 1.12990272e+00 -3.12413186e-01
2.94674635e-01 -5.42228222e-01 -1.47809768e+00 8.06368053e-01
1.38592243e-01 2.15130523e-01 7.70704150e-01 2.45879013e-02
-5.07277310e-01 -3.22816633e-02 -9.71806884e-01 4.76722091e-01
1.29713666e+00 -4.36950684e-01 -8.13579679e-01 1.06167123e-01
6.57761633e-01 -8.82470608e-02 -5.70798993e-01 2.62939394e-01
5.59023499e-01 -8.22700441e-01 8.45737934e-01 -1.01454842e+00
3.16346735e-01 -3.38333726e-01 -3.82381044e-02 -1.58633292e+00
-7.82537341e-01 -1.83064461e-01 -3.11194628e-01 7.88342893e-01
7.19715297e-01 -8.50000679e-01 6.61225319e-01 2.23402798e-01
-2.15034246e-01 -8.10359299e-01 -8.40223789e-01 -7.59194672e-01
1.71565965e-01 -4.40463275e-02 2.76683867e-01 8.80399346e-01
-3.66984963e-01 3.75472337e-01 -7.01858774e-02 -2.63561994e-01
2.31834695e-01 -4.17110711e-01 1.05504245e-01 -1.44714558e+00
-3.60201806e-01 -5.68226516e-01 -6.51446953e-02 -5.87883711e-01
6.12044595e-02 -1.12105620e+00 1.39494121e-01 -1.09909797e+00
1.95031047e-01 -6.16011202e-01 -1.01673138e+00 9.01985705e-01
-6.74910471e-02 3.45341325e-01 8.66066851e-03 3.80025834e-01
-2.93073028e-01 3.57465297e-01 1.10717738e+00 -1.81627363e-01
-4.43202138e-01 -2.49034911e-01 -7.62281120e-01 5.23814976e-01
8.44503343e-01 -9.83185947e-01 -4.31263983e-01 -4.58594888e-01
6.23462856e-01 -3.36475790e-01 3.04415435e-01 -1.21654248e+00
4.67064321e-01 -9.57643688e-02 8.55605781e-01 -7.14266747e-02
1.78922310e-01 -6.02173209e-01 6.17111325e-01 8.30419898e-01
-4.48799938e-01 3.90872180e-01 5.71829140e-01 7.05837071e-01
1.31776463e-03 -3.09617937e-01 9.48804021e-01 -4.24269587e-01
-4.86893654e-01 1.28132880e-01 -1.08731246e+00 -4.02507409e-02
7.55039513e-01 -3.98685038e-01 -2.35369563e-01 1.70467898e-01
-1.27862108e+00 -1.04641244e-01 4.14505005e-01 2.47576058e-01
8.74146163e-01 -1.36545610e+00 -4.60680664e-01 5.14260173e-01
-3.08239087e-02 -8.95642191e-02 6.03661500e-02 6.42401218e-01
-3.07713300e-01 3.96994501e-01 -7.02063322e-01 -2.46499032e-01
-7.69821942e-01 5.52372694e-01 3.12384844e-01 -5.59870362e-01
-2.04938278e-01 1.08390725e+00 2.71559298e-01 -3.60725611e-01
8.65169913e-02 -3.21935624e-01 -2.07050353e-01 1.21841207e-01
3.77643883e-01 -1.29933078e-02 3.29948783e-01 -2.80415177e-01
-2.28013217e-01 7.90075213e-02 -4.45872337e-01 -4.69420403e-02
1.70654035e+00 2.00173795e-01 -3.58439296e-01 7.37764657e-01
6.54032052e-01 -2.80268639e-01 -1.23917520e+00 -2.91391164e-01
9.67404991e-02 -5.94637319e-02 -1.39767826e-01 -8.42817962e-01
-1.16079879e+00 6.65183306e-01 5.22802293e-01 9.31321383e-02
1.29978096e+00 -4.39100564e-01 4.73405033e-01 5.93078494e-01
5.59233546e-01 -1.11743283e+00 5.40696144e-01 1.11895168e+00
9.53326941e-01 -6.33502364e-01 -1.53648704e-01 2.17958421e-01
-1.30125508e-01 1.05265248e+00 5.90786576e-01 -5.07669866e-01
6.15968347e-01 4.35672641e-01 -4.39167291e-01 -2.30340898e-01
-1.27532959e+00 -8.51222947e-02 6.96817040e-02 5.71011364e-01
3.93062085e-01 -2.05802187e-01 -3.04037333e-01 5.41241229e-01
4.23379950e-02 -5.37941754e-02 3.97619873e-01 1.30312026e+00
-8.90508831e-01 -1.03386605e+00 -1.25765726e-01 6.95220411e-01
-2.44752392e-01 -4.42877620e-01 -2.57425368e-01 7.20972717e-01
4.95359957e-01 4.43306118e-01 1.55520827e-01 -2.85892665e-01
2.58335054e-01 5.74462533e-01 7.43844807e-01 -1.09459758e+00
-9.90616500e-01 -2.98007756e-01 -3.62852871e-01 -2.76276052e-01
-2.63228148e-01 -5.60458243e-01 -1.30575907e+00 -2.87497908e-01
-1.49617106e-01 5.46734743e-02 2.07647353e-01 7.16667116e-01
4.16607589e-01 9.45094705e-01 1.51338190e-01 -8.23218703e-01
-2.07208812e-01 -7.96182215e-01 -7.14943945e-01 4.50530916e-01
2.73006648e-01 -8.33423018e-01 -4.43751544e-01 3.09671909e-01]
|
[8.730279922485352, 3.211463451385498]
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.