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 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
7acef441-09be-41a9-a5d3-4421387aa2d1 | why-did-the-chicken-cross-the-road-rephrasing | 2211.07516 | null | https://arxiv.org/abs/2211.07516v2 | https://arxiv.org/pdf/2211.07516v2.pdf | Why Did the Chicken Cross the Road? Rephrasing and Analyzing Ambiguous Questions in VQA | Natural language is ambiguous. Resolving ambiguous questions is key to successfully answering them. Focusing on questions about images, we create a dataset of ambiguous examples. We annotate these, grouping answers by the underlying question they address and rephrasing the question for each group to reduce ambiguity. Our analysis reveals a linguistically-aligned ontology of reasons for ambiguity in visual questions. We then develop an English question-generation model which we demonstrate via automatic and human evaluation produces less ambiguous questions. We further show that the question generation objective we use allows the model to integrate answer group information without any direct supervision. | ['Benjamin Van Durme', 'Yi Zhou', 'Jimena Guallar-Blasco', 'Elias Stengel-Eskin'] | 2022-11-14 | null | null | null | null | ['question-generation'] | ['natural-language-processing'] | [ 4.57358301e-01 9.87892807e-01 2.68890411e-01 -6.72718823e-01
-1.30572629e+00 -1.17864025e+00 7.10371494e-01 1.88584432e-01
-2.47210488e-01 6.71857953e-01 7.59814978e-01 -7.96851218e-01
-9.86643806e-02 -5.54473698e-01 -6.44843042e-01 2.89983362e-01
4.31883305e-01 5.93755066e-01 4.11161035e-01 -4.75685239e-01
3.27126473e-01 1.83062144e-02 -1.60339820e+00 9.82143164e-01
9.77127254e-01 6.26440048e-01 1.74421817e-01 8.74298751e-01
-7.64522791e-01 1.24696136e+00 -7.67469645e-01 -8.10848534e-01
5.09811416e-02 -7.19366491e-01 -1.72670019e+00 5.58922350e-01
8.95567060e-01 -4.01657999e-01 5.50883375e-02 8.24943006e-01
-1.09512888e-01 -1.08356461e-01 5.36735952e-01 -1.24027109e+00
-1.00613463e+00 4.38174874e-01 5.78198880e-02 2.53273726e-01
1.11766422e+00 6.83429763e-02 1.48876464e+00 -1.05767739e+00
1.17682469e+00 1.47503853e+00 3.39067221e-01 8.34415019e-01
-1.36889338e+00 -2.33001076e-02 3.30252945e-01 4.17108566e-01
-1.08273804e+00 -2.69404233e-01 6.39946759e-01 -4.23998058e-01
1.00962877e+00 6.72443986e-01 4.02421296e-01 8.33713353e-01
-6.09576590e-02 7.09773481e-01 1.13705468e+00 -5.52524328e-01
1.23132430e-01 1.58966318e-01 4.00411010e-01 5.88322103e-01
2.43021417e-02 -4.47722614e-01 -3.53850663e-01 -1.19999506e-01
4.78209108e-01 -6.48592532e-01 -2.63167441e-01 -1.07693233e-01
-1.24747789e+00 8.71859610e-01 5.08783817e-01 1.21957794e-01
-2.31922194e-01 9.04232934e-02 -5.03033353e-03 5.22505224e-01
-5.86938262e-02 1.10134852e+00 -3.07539284e-01 3.31917614e-01
-3.67545635e-01 6.64665818e-01 1.08618319e+00 1.04356885e+00
1.03247428e+00 -6.12553060e-01 -3.16280454e-01 6.68982387e-01
3.05616498e-01 2.78778881e-01 1.18354619e-01 -1.90810978e+00
3.58710587e-01 8.42785358e-01 5.08684695e-01 -1.22822273e+00
-2.30171680e-01 3.41383666e-02 -3.36725749e-02 1.11013092e-01
6.36421263e-01 2.00659528e-01 -7.91493654e-01 1.68453908e+00
3.63831013e-01 -6.43449545e-01 2.46408552e-01 7.86923945e-01
1.20959485e+00 6.11336827e-01 3.24420273e-01 5.46353757e-02
1.96343446e+00 -8.73841524e-01 -9.46461916e-01 -5.71618319e-01
3.52511108e-01 -9.35695589e-01 1.62116110e+00 7.19139129e-02
-1.00352263e+00 -4.04885113e-01 -6.82279289e-01 -7.71868765e-01
-4.14523989e-01 -2.94519700e-02 2.21947834e-01 1.96225435e-01
-1.27675533e+00 -9.75700244e-02 -1.28539175e-01 -5.71704805e-01
1.77452862e-01 3.49495336e-02 -1.75378397e-01 -1.30692989e-01
-1.08534479e+00 1.01128185e+00 1.11484252e-01 -2.72826314e-01
-3.56972694e-01 -3.34446937e-01 -1.14289165e+00 -2.49085918e-01
7.15188086e-01 -1.28524005e+00 1.73483276e+00 -1.07218945e+00
-8.72646451e-01 1.48314476e+00 -6.13707781e-01 -4.29646730e-01
1.60746440e-01 -9.20680463e-02 -2.68217981e-01 8.25538695e-01
8.79601061e-01 1.57439983e+00 7.46405184e-01 -1.74151373e+00
-7.23767936e-01 -3.08776826e-01 9.43359017e-01 3.69252414e-01
1.06121756e-01 2.19976202e-01 -3.19466472e-01 -8.04708302e-01
3.69978458e-01 -7.05627441e-01 -2.05884859e-01 2.60620058e-01
-3.42447788e-01 -3.96943301e-01 6.97806239e-01 -9.84744072e-01
1.09243679e+00 -1.66291058e+00 -1.01530217e-01 -7.63629153e-02
4.89106238e-01 -4.92795080e-01 -4.52783853e-01 3.78132105e-01
-3.99332605e-02 6.00591302e-01 -3.47315282e-01 -1.10870851e-02
1.92835227e-01 7.03622878e-01 -7.01011598e-01 -4.73949492e-01
7.82084465e-01 1.18224728e+00 -8.54901493e-01 -9.02229488e-01
-9.20594782e-02 6.30970299e-02 -8.28301549e-01 4.84011769e-01
-9.42940235e-01 2.58332610e-01 -3.87101889e-01 6.18965268e-01
4.00310665e-01 -6.76205456e-01 9.04156268e-03 -3.77879709e-01
1.19926132e-01 5.27309954e-01 -9.99542415e-01 1.43691456e+00
-3.58369768e-01 6.63876772e-01 1.77792907e-01 -3.82977456e-01
6.78370535e-01 2.23111451e-01 -1.69576898e-01 -5.83759487e-01
-2.45393634e-01 1.65396139e-01 -2.14782134e-01 -1.14654493e+00
6.56270206e-01 -2.51675457e-01 -3.54434878e-01 6.77350938e-01
1.44964345e-02 -7.68528163e-01 5.24661481e-01 8.01429927e-01
9.70277369e-01 2.08889976e-01 4.42803353e-01 -2.54725903e-01
7.26096332e-01 6.19129419e-01 4.79148206e-04 9.67979550e-01
-2.68518507e-01 8.02282453e-01 5.25440812e-01 -6.06650949e-01
-9.73736405e-01 -1.32878244e+00 2.14110360e-01 1.13500106e+00
2.73249388e-01 -5.75157404e-01 -8.71247172e-01 -9.14713025e-01
-2.26557404e-01 9.42322552e-01 -6.45319879e-01 5.44086516e-01
-6.75141692e-01 -1.63077842e-02 2.95683801e-01 2.20579639e-01
1.94378868e-01 -1.34550929e+00 -9.08670783e-01 2.47684633e-03
-8.40339363e-01 -1.50623298e+00 -4.20631975e-01 -2.05463633e-01
-4.37181979e-01 -1.38312566e+00 -3.28379601e-01 -1.11881363e+00
9.08999145e-01 1.62764192e-01 1.71989453e+00 5.65384030e-01
-1.84776470e-01 1.15681183e+00 -4.55723941e-01 -3.09969574e-01
-7.45760500e-01 -1.15725778e-01 -7.59715676e-01 -3.49250108e-01
4.22790587e-01 -2.25202486e-01 -5.67287624e-01 2.81653702e-01
-1.33364689e+00 1.41689822e-01 2.63021857e-01 4.11572099e-01
5.16660094e-01 -7.63157845e-01 6.20002449e-01 -7.42056727e-01
1.01510406e+00 -2.48412997e-01 -2.65324146e-01 6.53621256e-01
-9.91580337e-02 3.49921465e-01 2.11001769e-01 -6.56961650e-02
-1.18095207e+00 -1.02843838e-02 -3.46027613e-01 2.84543037e-01
-5.43647170e-01 3.00563574e-01 -1.25085309e-01 1.66415676e-01
9.35619593e-01 -1.53870404e-01 -1.82519794e-01 -1.00434512e-01
9.07107711e-01 2.84455866e-01 9.60589767e-01 -8.37045312e-01
8.43692899e-01 6.39506698e-01 -2.64327288e-01 -7.32444227e-01
-1.16835606e+00 -2.63392746e-01 -2.97772884e-01 -3.32076311e-01
1.19875753e+00 -6.48430943e-01 -5.61433792e-01 -4.43170339e-01
-1.82908332e+00 1.29507422e-01 -5.94137132e-01 -1.24138877e-01
-7.55027711e-01 5.63962102e-01 -4.17625993e-01 -7.32783973e-01
6.32814551e-03 -9.99213219e-01 1.15019679e+00 1.27078563e-01
-9.34321046e-01 -8.11133504e-01 -9.63259935e-02 9.44301903e-01
2.09155425e-01 1.53235793e-01 1.41297579e+00 -7.65148938e-01
-9.22404587e-01 4.28495139e-01 -4.79947776e-01 -3.35734263e-02
2.74109274e-01 -2.29203880e-01 -7.31678188e-01 3.98140073e-01
2.01740742e-01 -7.19706416e-01 4.98026282e-01 -7.97581598e-02
8.90337944e-01 -6.38419807e-01 -4.62136790e-02 -2.00194761e-01
1.32100010e+00 -1.33870706e-01 5.47196865e-01 4.62900817e-01
4.30943727e-01 1.30519485e+00 5.34321129e-01 -1.98326081e-01
1.12906885e+00 3.56691480e-01 4.93007153e-01 -3.38328630e-02
-3.59814078e-01 -3.42227340e-01 -1.53269589e-01 3.27916890e-01
6.02244377e-01 -1.77300423e-01 -1.14904976e+00 1.07150555e+00
-1.80933619e+00 -9.29956377e-01 -4.55798239e-01 1.41725302e+00
1.13706851e+00 -1.84494615e-01 -1.46491183e-02 -2.18924895e-01
6.62858069e-01 4.34694514e-02 -2.25524098e-01 -4.97005135e-01
-3.15607995e-01 2.57087111e-01 -3.66565526e-01 1.19407105e+00
-9.06880498e-01 1.07923996e+00 7.57246065e+00 2.62959272e-01
-2.30520234e-01 -1.97024226e-01 7.15744495e-01 4.09395993e-01
-1.23754513e+00 4.13869351e-01 -4.39376742e-01 -2.30856150e-01
2.78958648e-01 -1.93655998e-01 3.69858921e-01 3.05014521e-01
1.24427252e-01 -4.41792011e-01 -1.25442970e+00 7.65196562e-01
4.85051066e-01 -1.48436797e+00 7.89880455e-01 -3.88339072e-01
5.33497274e-01 -6.79165483e-01 -2.18835652e-01 1.50805945e-02
4.85195100e-01 -1.17826724e+00 8.07784975e-01 4.85509336e-01
3.11162233e-01 -3.94042820e-01 4.35045004e-01 2.42349684e-01
-9.48265791e-01 -6.42806962e-02 -3.71398106e-02 -3.31433594e-01
3.92139107e-01 1.77237660e-01 -8.84474695e-01 3.40023428e-01
7.41934240e-01 -8.82793497e-03 -1.16104543e+00 5.16322076e-01
-7.79313982e-01 6.54364824e-02 -2.95786649e-01 -1.97522156e-02
1.60860717e-01 -6.24899492e-02 6.16870940e-01 9.09372330e-01
1.44984871e-01 5.64577818e-01 2.60409057e-01 1.21117318e+00
-9.20576677e-02 5.12844250e-02 -7.08157957e-01 1.37824789e-01
3.27577919e-01 1.13744617e+00 -7.35246539e-01 -5.51490188e-01
-3.83990377e-01 1.07877195e+00 3.36882204e-01 4.90062982e-01
-4.38623995e-01 -4.56045181e-01 3.94300610e-01 1.18809700e-01
5.55711985e-02 -6.42748997e-02 -2.56542891e-01 -1.16205454e+00
3.33356619e-01 -1.37027025e+00 8.15756798e-01 -1.60851276e+00
-1.36392796e+00 8.04351985e-01 1.00781448e-01 -8.59397829e-01
-7.11790085e-01 -7.28565097e-01 -2.71347523e-01 7.36370325e-01
-1.63884628e+00 -1.14570427e+00 -3.56931716e-01 3.81050229e-01
7.85988629e-01 5.09525895e-01 8.31287265e-01 -1.35995924e-01
1.49676666e-01 3.28921676e-02 -1.09182394e+00 8.25900286e-02
7.11307049e-01 -1.43020523e+00 4.44740355e-01 7.54979968e-01
3.11639667e-01 8.70665014e-01 1.21698129e+00 -6.19492710e-01
-8.82345498e-01 -7.38601565e-01 1.56468320e+00 -9.57898736e-01
8.53666663e-01 -1.89681157e-01 -1.13599205e+00 8.17359149e-01
1.01753139e+00 -3.63374650e-01 8.53131056e-01 -3.08758348e-01
-6.40344977e-01 4.43712264e-01 -1.24699306e+00 9.95463252e-01
1.12829411e+00 -8.99810970e-01 -1.65048552e+00 4.21441227e-01
1.12274575e+00 -3.12055409e-01 -4.35564637e-01 3.74483258e-01
1.85663432e-01 -1.14071321e+00 8.83155346e-01 -8.48532438e-01
7.61591434e-01 -7.75581539e-01 -2.81706005e-01 -8.51264179e-01
-5.86000010e-02 -5.09075344e-01 2.64999568e-01 1.11229384e+00
7.82882750e-01 -3.08295786e-01 7.47304559e-01 9.99594152e-01
9.59756523e-02 -3.83896708e-01 -8.92446697e-01 -3.63858193e-01
1.33621367e-02 -4.77743596e-01 4.99668926e-01 7.77129889e-01
2.43781462e-01 7.20603108e-01 2.36651540e-01 2.06660226e-01
3.46171021e-01 3.48071545e-01 6.91297472e-01 -9.13520813e-01
-9.26105455e-02 -2.38669887e-01 8.28174800e-02 -1.27449489e+00
1.36831939e-01 -7.18926251e-01 2.88207293e-01 -2.22069025e+00
6.58464879e-02 2.04732671e-01 5.54382503e-01 3.34074169e-01
-5.31405509e-01 3.58090371e-01 2.71660507e-01 7.94123262e-02
-1.05228269e+00 1.94918975e-01 1.27283692e+00 -1.30375430e-01
1.40714496e-01 -6.67336702e-01 -1.32586801e+00 9.97018397e-01
6.92890465e-01 -3.25438946e-01 -4.74087268e-01 -8.41976404e-01
8.91233563e-01 -1.53550074e-01 7.56152332e-01 -6.12369299e-01
1.45816937e-01 -2.97394246e-01 1.22162104e-01 -4.45759356e-01
1.11581489e-01 -8.23496997e-01 -1.88283995e-01 2.35880271e-01
-7.55588472e-01 5.60731411e-01 1.95595428e-01 1.98348671e-01
-4.45191592e-01 -5.06553829e-01 2.84113407e-01 -4.33288187e-01
-6.94676101e-01 -3.11769038e-01 -6.77616358e-01 6.66878223e-01
7.16459811e-01 -2.81591237e-01 -5.46962678e-01 -1.01465201e+00
-1.15900397e+00 7.05636203e-01 6.15151644e-01 4.49334145e-01
8.63245487e-01 -1.45152116e+00 -8.04503620e-01 -3.32578987e-01
6.48296833e-01 -1.03365347e-01 4.02423069e-02 5.12936786e-02
-5.56871951e-01 3.13118815e-01 1.23303153e-01 -5.45867145e-01
-1.23792493e+00 4.78704572e-01 4.59587455e-01 -1.74072221e-01
-1.05539843e-01 6.95698202e-01 3.92169446e-01 -6.80569828e-01
6.69216961e-02 -5.76350987e-01 -5.96058309e-01 5.47431000e-02
4.63249803e-01 -1.44842416e-01 -3.54798824e-01 -5.77240884e-01
-3.63996714e-01 7.85053790e-01 1.36664987e-01 -6.19164228e-01
8.71584654e-01 -6.92059755e-01 -4.27127481e-01 1.80675924e-01
8.52977216e-01 2.50712156e-01 -8.01557660e-01 -4.56312075e-02
4.66155261e-01 -3.23053330e-01 -8.49489331e-01 -8.99505079e-01
-1.10405162e-01 6.27714217e-01 1.29487604e-01 6.99207008e-01
1.06072533e+00 8.35869789e-01 5.77880204e-01 8.35336149e-01
4.92079444e-02 -9.34300184e-01 5.94023347e-01 7.38393605e-01
1.57508814e+00 -1.24725819e+00 -1.99628606e-01 -9.25543606e-01
-6.17663741e-01 1.13712513e+00 8.90245080e-01 1.62807330e-01
1.44515440e-01 1.90653950e-02 7.75365710e-01 -6.02816939e-01
-8.80631864e-01 -4.46057558e-01 2.28794649e-01 8.08155596e-01
4.29843098e-01 -4.39401716e-01 -4.84045953e-01 5.66012323e-01
-5.99497318e-01 -2.39764586e-01 8.04270446e-01 9.98792529e-01
-6.54101729e-01 -1.16562235e+00 -7.40888834e-01 1.23274036e-01
-3.60160410e-01 -2.59249002e-01 -1.06345809e+00 7.27704465e-01
1.19818248e-01 1.44841886e+00 1.55786946e-01 1.39846176e-01
4.53069657e-01 2.44457066e-01 6.24248326e-01 -8.82144332e-01
-5.57817876e-01 -3.99492621e-01 6.22124672e-01 -3.64040405e-01
-5.62232375e-01 -3.82037431e-01 -1.49708056e+00 2.79845864e-01
1.47165924e-01 3.18575919e-01 2.49141216e-01 1.13578498e+00
5.49062669e-01 2.74556011e-01 2.52453834e-01 -8.89679417e-02
-2.40578547e-01 -4.34365273e-01 4.06826526e-01 9.28292036e-01
5.34764767e-01 5.71758859e-02 -5.00331759e-01 6.35100126e-01] | [10.994869232177734, 1.759125828742981] |
b4e4190b-a6ee-47f0-97b0-d142154be4f5 | vision-based-lane-detection-and-tracking | 2210.10233 | null | https://arxiv.org/abs/2210.10233v3 | https://arxiv.org/pdf/2210.10233v3.pdf | Vision-Based Robust Lane Detection and Tracking under Different Challenging Environmental Conditions | Lane marking detection is fundamental for both advanced driving assistance systems. However, detecting lane is highly challenging when the visibility of a road lane marking is low due to real-life challenging environment and adverse weather. Most of the lane detection methods suffer from four types of challenges: (i) light effects i.e., shadow, glare of light, reflection etc.; (ii) Obscured visibility of eroded, blurred, colored and cracked lane caused by natural disasters and adverse weather; (iii) lane marking occlusion by different objects from surroundings (wiper, vehicles etc.); and (iv) presence of confusing lane like lines inside the lane view e.g., guardrails, pavement marking, road divider etc. Here, we propose a robust lane detection and tracking method with three key technologies. First, we introduce a comprehensive intensity threshold range (CITR) to improve the performance of the canny operator in detecting low intensity lane edges. Second, we propose a two-step lane verification technique, the angle based geometric constraint (AGC) and length-based geometric constraint (LGC) followed by Hough Transform, to verify the characteristics of lane marking and to prevent incorrect lane detection. Finally, we propose a novel lane tracking technique, by defining a range of horizontal lane position (RHLP) along the x axis which will be updating with respect to the lane position of previous frame. It can keep track of the lane position when either left or right or both lane markings are partially and fully invisible. To evaluate the performance of the proposed method we used the DSDLDE [1] and SLD [2] dataset with 1080x1920 and 480x720 resolutions at 24 and 25 frames/sec respectively. Experimental results show that the average detection rate is 97.55%, and the average processing time is 22.33 msec/frame, which outperform the state of-the-art method. | ['Shamim Ahmad', 'Muhammad Rafiqul Islam', 'Manoranjan Paul', 'Boshir Ahmed', 'Samia Sultana'] | 2022-10-19 | null | null | null | null | ['lane-detection'] | ['computer-vision'] | [ 5.64104021e-02 -4.53554988e-01 1.81591272e-01 1.02847219e-02
-9.48394835e-02 -5.37311316e-01 4.00579870e-01 -1.34383619e-01
-4.54315573e-01 9.30173039e-01 -5.45856841e-02 -8.42462420e-01
2.86385119e-01 -6.58178508e-01 -3.47427547e-01 -7.83277392e-01
-1.13742407e-02 -3.19838941e-01 1.10639906e+00 -2.08116010e-01
7.21500218e-01 8.56026530e-01 -1.71249950e+00 -2.77571619e-01
1.06295979e+00 6.49014354e-01 4.23672915e-01 7.82446504e-01
-2.52761096e-01 6.58226013e-01 -3.54592800e-01 7.66087621e-02
3.82807195e-01 -1.27358884e-01 7.30643747e-03 4.74568516e-01
4.06179726e-01 -6.45388901e-01 -2.96089619e-01 1.16865826e+00
2.65770137e-01 1.49443626e-01 5.82032740e-01 -1.60102320e+00
9.62220654e-02 -4.87478644e-01 -1.19484639e+00 6.22180164e-01
2.71867756e-02 4.72421706e-01 1.74942650e-02 -9.07346010e-01
6.21070564e-01 1.37994921e+00 4.70705003e-01 -2.49796640e-03
-7.06193805e-01 -8.35155666e-01 -1.45575613e-01 7.10017800e-01
-1.46237099e+00 -4.57429141e-01 9.05956149e-01 -3.77528489e-01
2.90039510e-01 1.68754116e-01 1.47056282e-01 1.92467839e-01
6.80674374e-01 4.57513213e-01 1.50511527e+00 -3.99540365e-01
-2.78566718e-01 1.34463146e-01 2.76139140e-01 1.06854451e+00
5.10353625e-01 4.20556217e-01 1.48492083e-01 2.63900816e-01
5.43781519e-01 -1.14791758e-01 -4.27288681e-01 -2.91043639e-01
-1.16963279e+00 4.53615159e-01 1.54797912e-01 -8.82973000e-02
-1.99559614e-01 -2.58173913e-01 4.55063283e-01 -3.98780294e-02
-2.33517319e-01 -4.42765564e-01 2.79050507e-02 -5.28337583e-02
-5.05001545e-01 1.06974661e-01 3.95858407e-01 1.08334649e+00
9.68048036e-01 1.94868118e-01 1.79113120e-01 5.50801039e-01
4.05173361e-01 9.20232892e-01 1.23595528e-03 -6.79832041e-01
6.63998842e-01 1.89130962e-01 3.80942017e-01 -1.33531117e+00
-5.33357561e-01 -2.61861682e-01 -6.51254952e-01 9.47611690e-01
4.95327502e-01 -3.21570575e-01 -8.56652915e-01 1.07051218e+00
4.11280006e-01 5.69988430e-01 3.11551034e-01 8.80145192e-01
6.68330491e-01 1.09726286e+00 7.29165692e-03 -4.64294285e-01
1.38946843e+00 -1.02028978e+00 -1.08375955e+00 -4.60419923e-01
6.03982091e-01 -1.28583622e+00 7.80737102e-01 2.32400984e-01
-6.17117941e-01 -7.27584660e-01 -1.28219533e+00 5.61098643e-02
-5.20721197e-01 3.48733008e-01 5.04838675e-02 6.49785936e-01
-7.43128598e-01 -1.20400272e-01 -1.53612822e-01 -2.55424052e-01
-7.48657584e-02 -1.64003998e-01 -3.76549125e-01 -5.45756638e-01
-1.16734827e+00 1.05228925e+00 4.26098943e-01 4.95219171e-01
-3.58635306e-01 -3.83591741e-01 -8.71013880e-01 -3.97368699e-01
5.44455290e-01 8.08007568e-02 6.09090447e-01 -5.30772269e-01
-1.15830433e+00 8.22678089e-01 -4.53125566e-01 -2.83326298e-01
7.74242043e-01 -1.08520672e-01 -1.05553389e+00 -7.33114555e-02
1.69803575e-01 3.22654039e-01 5.92364550e-01 -1.57916522e+00
-1.32345974e+00 -1.65832996e-01 -5.09503126e-01 3.01566869e-01
8.18218529e-01 2.72763800e-02 -3.04372072e-01 -9.41793770e-02
9.25288722e-02 -9.62018132e-01 -3.89681105e-03 7.90205672e-02
-6.08468056e-01 1.55154802e-02 1.79325116e+00 -8.44513297e-01
1.41322124e+00 -2.47147298e+00 -9.85719085e-01 2.72998273e-01
-8.75241384e-02 7.82260060e-01 6.06045462e-02 2.63325185e-01
1.51520565e-01 -1.98171139e-01 -1.27626881e-01 3.39974076e-01
-3.16956073e-01 2.50963598e-01 -3.74666154e-01 8.28129947e-01
-1.00694396e-01 4.04400378e-01 -7.75696099e-01 -6.64121985e-01
7.31278360e-01 3.11418772e-01 3.45091611e-01 8.01244751e-02
7.52396822e-01 1.74816176e-01 -9.28317457e-02 6.21955991e-01
1.50560021e+00 7.76804745e-01 -5.42225778e-01 -3.75548005e-01
-1.07672453e+00 -1.95306748e-01 -1.69379878e+00 3.84496152e-01
-3.13381732e-01 1.10290694e+00 2.32887015e-01 -2.25941554e-01
1.31398547e+00 -3.16486508e-02 -2.36989871e-01 -7.88592815e-01
1.37844071e-01 4.23963994e-01 4.54356037e-02 -6.96965456e-01
6.72324836e-01 7.14293718e-02 6.70142844e-02 7.87284747e-02
-9.17532563e-01 2.82564670e-01 1.34856895e-01 -6.37948811e-02
8.14716160e-01 -3.07958215e-01 2.09410101e-01 -1.75745279e-01
1.13620901e+00 -5.32122105e-02 9.16804612e-01 5.19679844e-01
-9.39646661e-01 2.19391540e-01 4.99874175e-01 -3.11316162e-01
-1.06738353e+00 -1.07153141e+00 -1.50507256e-01 5.08103549e-01
7.66113102e-01 3.93591732e-01 -3.97906005e-01 -2.83918649e-01
1.36861533e-01 9.10530746e-01 -9.91945043e-02 -2.53449790e-02
-8.75153363e-01 -4.11628246e-01 3.67206097e-01 1.54157504e-01
1.14430344e+00 -8.72405171e-01 -8.66558671e-01 1.89702988e-01
1.75862044e-01 -1.36495328e+00 -4.75282520e-01 -3.81693482e-01
-4.89468277e-01 -1.15215528e+00 -4.52375591e-01 -9.61123228e-01
6.60224795e-01 1.11261284e+00 5.08714974e-01 2.22641498e-01
-1.51004493e-01 -1.14021078e-01 -1.00547001e-01 -4.93309468e-01
-3.78868520e-01 -5.40320814e-01 -2.49908134e-01 1.63237840e-01
1.24522477e-01 2.92699151e-02 -6.51223838e-01 7.65009522e-01
-5.68151355e-01 2.71022469e-01 8.08279216e-01 5.32811463e-01
2.86772698e-01 4.63035911e-01 4.67344493e-01 -7.04412639e-01
6.07767642e-01 -2.04366535e-01 -1.04401386e+00 -1.72433078e-01
-5.79222620e-01 -4.02045190e-01 4.97849822e-01 -2.75691729e-02
-1.38278174e+00 3.65332398e-03 1.02529630e-01 -1.92527711e-01
-6.79497182e-01 2.76000947e-02 -2.06018403e-01 -1.96655422e-01
1.11621924e-01 2.88599938e-01 3.28315943e-01 -9.42104161e-02
3.17550480e-01 8.27990294e-01 1.11737716e+00 1.97721973e-01
1.22336185e+00 5.61046660e-01 4.48510855e-01 -1.13734043e+00
-5.73508665e-02 -6.88731790e-01 -6.51638687e-01 -7.03431845e-01
9.40272391e-01 -6.67399704e-01 -6.98871613e-01 8.29932511e-01
-1.19756532e+00 -2.31502522e-02 6.95804060e-01 4.58046049e-01
-1.14316963e-01 8.51745307e-01 -3.60349357e-01 -1.10648429e+00
-8.99435282e-02 -1.03315520e+00 5.91106534e-01 6.87539577e-01
5.90641677e-01 -8.23796093e-01 -4.24818456e-01 1.34270489e-01
3.67975771e-01 6.42749369e-01 7.14643896e-01 1.21403962e-01
-9.17559087e-01 -2.84763873e-01 -7.22249627e-01 1.25532806e-01
2.17452869e-01 2.08519429e-01 -6.58361316e-01 5.06422780e-02
-4.73102063e-01 4.51808095e-01 7.49602735e-01 2.00697780e-01
3.65313083e-01 -3.32559384e-02 -5.87087035e-01 3.47166002e-01
1.55125833e+00 7.33833909e-01 1.13103127e+00 5.89820445e-01
6.76581860e-01 5.47757566e-01 1.25700057e+00 -2.27695107e-02
3.27913612e-01 5.18438935e-01 3.06324542e-01 -5.98322928e-01
-3.68508607e-01 -6.71910569e-02 4.78379041e-01 4.54398215e-01
9.03239250e-02 -2.07119405e-01 -8.40956807e-01 4.65708762e-01
-1.61802757e+00 -1.06702077e+00 -1.15734267e+00 2.36073232e+00
1.48149371e-01 6.12326622e-01 -6.37633726e-02 4.64430422e-01
1.09611905e+00 1.45432591e-01 -2.93702781e-01 -7.83255219e-01
-2.41666496e-01 -6.39362335e-01 1.27426970e+00 9.28148568e-01
-1.08159745e+00 8.75582993e-01 4.70324039e+00 7.74063230e-01
-1.34810245e+00 -2.51875907e-01 4.49913383e-01 8.78367484e-01
1.44907296e-01 -4.78347950e-02 -1.08601952e+00 8.07016373e-01
4.28310126e-01 -9.92220268e-02 8.88236985e-02 4.09473807e-01
7.64514089e-01 -8.71435106e-01 -4.62392829e-02 9.16578293e-01
5.68525754e-02 -7.16366112e-01 -4.97102857e-01 8.93562287e-02
4.37855363e-01 -2.60788023e-01 -5.62402308e-02 1.63359135e-01
-2.65004277e-01 -5.57706952e-01 6.31305397e-01 4.65121686e-01
6.18415475e-01 -1.05581808e+00 9.71544385e-01 4.31976855e-01
-1.60766816e+00 5.56368940e-02 -2.82092452e-01 7.09865689e-02
5.95130384e-01 4.16697264e-01 -1.12700784e+00 2.84891814e-01
1.91620395e-01 4.58414078e-01 -6.32198215e-01 1.33547997e+00
-3.44229966e-01 3.85604471e-01 -2.72676110e-01 3.57775688e-02
6.17032051e-01 -4.10921454e-01 8.88141692e-01 1.48435736e+00
2.47819707e-01 1.62245587e-01 2.17631921e-01 5.46197057e-01
6.14761531e-01 4.67953831e-02 -8.64032328e-01 9.72683489e-01
6.27467096e-01 1.20565736e+00 -7.98944950e-01 -1.96813509e-01
-5.77916026e-01 6.34499133e-01 -4.46349472e-01 5.84885895e-01
-9.85367417e-01 -9.77738798e-01 3.61146629e-01 2.46525332e-01
2.20870301e-01 -6.00823939e-01 -1.29877359e-01 -3.93261552e-01
-3.51200327e-02 -2.53802329e-01 9.60811675e-02 -9.76179838e-01
-5.46615660e-01 1.80879876e-01 -8.75479206e-02 -1.51152217e+00
2.76789308e-01 -5.53980231e-01 -9.23214912e-01 1.08818400e+00
-2.30505514e+00 -8.36643338e-01 -6.27183139e-01 3.74151498e-01
7.00369835e-01 -7.88308773e-03 -6.18506558e-02 5.94781816e-01
-6.79808974e-01 3.59258085e-01 1.79410830e-01 9.35802311e-02
7.20546126e-01 -7.77291715e-01 3.86000611e-02 1.28952897e+00
-9.06049132e-01 -2.93560307e-02 9.62002218e-01 -6.82825029e-01
-1.29927552e+00 -1.15985656e+00 9.55116570e-01 1.37825668e-01
4.04922366e-01 -5.68143614e-02 -1.09996247e+00 2.85877496e-01
7.13564754e-02 1.02353111e-01 -1.01153046e-01 -7.77424157e-01
2.53250778e-01 -3.83653313e-01 -1.16936922e+00 6.62860930e-01
5.49262881e-01 9.19394754e-03 -3.36681753e-01 -2.00867385e-01
5.83078116e-02 -4.96159315e-01 -4.72657606e-02 4.62832689e-01
6.04438782e-01 -1.26243961e+00 7.66304314e-01 2.36115813e-01
-2.97979742e-01 -1.22089112e+00 1.68299675e-01 -8.82386565e-01
-2.78071404e-01 -4.75839853e-01 5.49851120e-01 1.29888868e+00
2.76295781e-01 -9.52116013e-01 3.87182206e-01 3.53106916e-01
-5.63434303e-01 -5.24701715e-01 -7.80204773e-01 -6.73675835e-01
-5.06239057e-01 -6.30740300e-02 1.85415849e-01 6.56122863e-01
-4.50982630e-01 2.69541204e-01 -3.44515294e-01 6.56138897e-01
8.65588129e-01 -1.61510170e-01 1.11356604e+00 -9.52536762e-01
7.32329607e-01 -3.61617625e-01 -6.46476746e-01 -1.08413124e+00
-1.48687083e-02 -2.28859499e-01 3.09921801e-01 -1.52708638e+00
-2.48376653e-01 -5.04419029e-01 -6.25804290e-02 -5.04826196e-02
-2.66917646e-01 -5.23562506e-02 2.43948087e-01 1.44021481e-01
-3.31646204e-01 8.28910470e-02 1.36619270e+00 1.34542137e-01
-1.80992648e-01 -1.92715565e-03 6.39272705e-02 7.44269311e-01
8.62685978e-01 1.01082558e-02 -4.07799900e-01 -4.30669226e-02
-4.07759070e-01 1.71625823e-01 4.52473372e-01 -1.19216251e+00
3.95999521e-01 -3.45764995e-01 2.71270901e-01 -1.40148985e+00
-4.12879065e-02 -9.81774330e-01 -4.11318541e-02 7.52036989e-01
4.01202083e-01 4.38865840e-01 4.52026457e-01 7.04517722e-01
-1.22002251e-01 -1.81641132e-01 1.12922668e+00 1.64227560e-01
-1.66558766e+00 -2.29877055e-01 -8.29178452e-01 -3.40977281e-01
1.54875326e+00 -7.95555770e-01 -7.90874243e-01 -3.31254005e-01
4.44713905e-02 4.95812625e-01 4.79419231e-01 4.24491733e-01
8.10101390e-01 -1.29572701e+00 -7.31177747e-01 4.65386122e-01
4.43062596e-02 -3.60991210e-01 3.36340159e-01 1.16658497e+00
-1.06941354e+00 5.58211386e-01 -3.84312630e-01 -4.82399434e-01
-1.62666714e+00 6.04688406e-01 2.81624734e-01 1.20648459e-01
-7.45903790e-01 4.91772071e-02 1.30540863e-01 2.55228996e-01
1.51773114e-02 -2.69871742e-01 -5.40306330e-01 -2.58146107e-01
6.36098623e-01 9.48118746e-01 -1.87530518e-01 -1.26989746e+00
-2.65737534e-01 9.70836401e-01 1.46206319e-01 3.72339934e-02
3.70225281e-01 -7.22598732e-01 1.00233622e-01 3.81631553e-01
1.07645071e+00 3.83618861e-01 -1.34221411e+00 9.02723968e-02
-8.20136666e-02 -7.46263504e-01 2.43208840e-01 -6.41173422e-01
-6.64733589e-01 8.17779303e-01 1.03490448e+00 1.14698462e-01
8.98007035e-01 -6.25539482e-01 1.11819625e+00 -9.60721225e-02
2.60536075e-01 -1.16985142e+00 -5.93134940e-01 6.29799187e-01
5.70135713e-01 -1.10123956e+00 4.57736896e-03 -8.24629068e-01
-6.97572410e-01 1.29090774e+00 6.30892992e-01 -4.30088080e-02
4.75217491e-01 3.14062268e-01 4.25353348e-01 2.09544450e-01
-4.09320295e-01 -5.17573297e-01 1.23106584e-01 6.35418952e-01
1.12161813e-02 2.32821591e-02 -6.64101720e-01 -3.44994843e-01
1.58054277e-01 -1.49717957e-01 9.69803512e-01 1.07283247e+00
-1.25646055e+00 -4.65831637e-01 -1.01199710e+00 1.91716731e-01
-1.87824331e-02 2.84899831e-01 2.46514514e-01 1.03706121e+00
5.01334190e-01 1.16995835e+00 1.94224238e-01 -2.29919970e-01
5.91397524e-01 -7.44389072e-02 -1.72307506e-01 1.43834412e-01
4.02499557e-01 1.17806636e-01 3.56079191e-01 -2.51244873e-01
-2.84502804e-02 -7.65785336e-01 -1.56055486e+00 -4.83937055e-01
-1.66914657e-01 -3.78846936e-02 6.70736015e-01 7.03497350e-01
1.84648320e-01 2.19634131e-01 8.22510064e-01 -5.33010662e-01
1.03655703e-01 -6.17722273e-01 -7.40726709e-01 3.75629634e-01
7.23224103e-01 -9.79299009e-01 -5.05002677e-01 9.61109251e-02] | [8.006423950195312, -1.3952820301055908] |
9ab8c092-bde5-4bce-9c1e-45829dfba7cb | two-souls-in-an-adversarial-image-towards | 2109.12459 | null | https://arxiv.org/abs/2109.12459v2 | https://arxiv.org/pdf/2109.12459v2.pdf | Two Souls in an Adversarial Image: Towards Universal Adversarial Example Detection using Multi-view Inconsistency | In the evasion attacks against deep neural networks (DNN), the attacker generates adversarial instances that are visually indistinguishable from benign samples and sends them to the target DNN to trigger misclassifications. In this paper, we propose a novel multi-view adversarial image detector, namely Argos, based on a novel observation. That is, there exist two "souls" in an adversarial instance, i.e., the visually unchanged content, which corresponds to the true label, and the added invisible perturbation, which corresponds to the misclassified label. Such inconsistencies could be further amplified through an autoregressive generative approach that generates images with seed pixels selected from the original image, a selected label, and pixel distributions learned from the training data. The generated images (i.e., the "views") will deviate significantly from the original one if the label is adversarial, demonstrating inconsistencies that Argos expects to detect. To this end, Argos first amplifies the discrepancies between the visual content of an image and its misclassified label induced by the attack using a set of regeneration mechanisms and then identifies an image as adversarial if the reproduced views deviate to a preset degree. Our experimental results show that Argos significantly outperforms two representative adversarial detectors in both detection accuracy and robustness against six well-known adversarial attacks. Code is available at: https://github.com/sohaib730/Argos-Adversarial_Detection | ['Bo Luo', 'Fengjun Li', 'Chao Lan', 'Sana Awan', 'Sohaib Kiani'] | 2021-09-25 | null | null | null | null | ['adversarial-attack-detection', 'adversarial-attack-detection'] | ['computer-vision', 'knowledge-base'] | [ 6.17373765e-01 2.96021521e-01 3.53101254e-01 -1.71095468e-02
-5.38580954e-01 -1.20864749e+00 6.08609617e-01 -4.43921596e-01
-1.41389975e-02 5.77630758e-01 -3.20569724e-01 -2.11353526e-01
5.56686401e-01 -9.41371441e-01 -1.26900172e+00 -1.11003208e+00
1.47932306e-01 2.06982568e-01 1.04956694e-01 -1.26549199e-01
-1.05250902e-01 4.83858436e-01 -1.28875554e+00 4.22038972e-01
5.95324695e-01 8.30737472e-01 -2.78522700e-01 8.02480698e-01
3.63116205e-01 9.01241660e-01 -1.29162216e+00 -7.77789056e-01
7.81787395e-01 -5.17206073e-01 -2.12158889e-01 1.40695632e-01
7.44758606e-01 -6.25892818e-01 -6.90511763e-01 1.87054527e+00
4.28129792e-01 -3.74203652e-01 4.09191281e-01 -1.72217309e+00
-9.09205496e-01 5.29366255e-01 -5.18062592e-01 8.88437778e-02
1.64135501e-01 6.31124139e-01 4.33914393e-01 -6.10367775e-01
7.11436331e-01 1.39647985e+00 3.41336071e-01 1.20317590e+00
-1.24096918e+00 -1.12652862e+00 2.23319829e-01 -7.66714811e-02
-1.12435257e+00 -4.30769742e-01 9.82712686e-01 -4.17626828e-01
6.82516769e-02 4.75407660e-01 4.81230259e-01 1.88735330e+00
3.79439324e-01 4.76031393e-01 1.35862076e+00 -1.31373048e-01
3.93691331e-01 4.17214245e-01 -4.59828198e-01 6.46245480e-01
4.25038904e-01 7.53563821e-01 -4.25672457e-02 -1.82226539e-01
5.73947668e-01 1.87104777e-01 -4.67831790e-01 -2.73361623e-01
-8.05872500e-01 7.52106428e-01 5.86724877e-01 -5.86188585e-02
-3.04764390e-01 2.16965124e-01 2.38153636e-01 4.31460261e-01
1.65770173e-01 2.62571007e-01 -7.21967667e-02 4.94207352e-01
-3.32325459e-01 5.46969213e-02 4.96988028e-01 8.38359714e-01
5.59923887e-01 5.25996685e-01 -9.47472751e-02 2.87076086e-01
2.03731835e-01 9.13801551e-01 4.45765138e-01 -7.53117979e-01
3.78941476e-01 6.15985334e-01 5.13873436e-02 -1.10311687e+00
1.09857455e-01 -5.06307960e-01 -9.25602436e-01 1.13862181e+00
5.26938856e-01 -4.67973530e-01 -1.19329000e+00 2.07176161e+00
3.30546826e-01 2.49130681e-01 5.24095893e-01 9.31618869e-01
8.17220211e-01 5.06357670e-01 -1.23229787e-01 4.64947000e-02
1.18663347e+00 -6.56111658e-01 -3.96491140e-01 -4.55655724e-01
2.53614813e-01 -6.00377619e-01 7.24177599e-01 2.35329539e-01
-9.12940264e-01 -4.27500665e-01 -1.37630081e+00 7.71691680e-01
-3.20271254e-01 -4.02442887e-02 -2.04882193e-02 9.76925254e-01
-8.94451499e-01 2.76431471e-01 -4.65566248e-01 8.74854252e-02
5.32418787e-01 3.58932428e-02 -3.74828070e-01 -5.95084392e-02
-1.32421732e+00 4.64864105e-01 2.42750242e-01 6.05574762e-03
-1.74639976e+00 -4.86362100e-01 -4.99424428e-01 -1.68332428e-01
3.23582768e-01 -4.29827183e-01 7.12009609e-01 -1.84967566e+00
-1.01353371e+00 1.03041792e+00 3.23601127e-01 -5.52960336e-01
1.19117188e+00 2.11894959e-01 -6.50572002e-01 1.80563971e-01
-4.48972955e-02 5.55505633e-01 1.42753446e+00 -1.87983763e+00
-3.47526103e-01 -6.44613326e-01 2.93365389e-01 2.10840534e-03
-1.66465670e-01 -1.89128086e-01 -5.54036126e-02 -9.19494033e-01
-1.73358142e-01 -1.20693731e+00 1.70332845e-02 1.55119047e-01
-9.94064033e-01 4.51105237e-01 1.23721910e+00 -4.18798685e-01
4.77781028e-01 -2.50444221e+00 -2.40151107e-01 3.05954397e-01
5.46721041e-01 5.40570796e-01 -3.39501858e-01 1.24449797e-01
-4.22030449e-01 3.21418107e-01 -1.00660704e-01 6.65028719e-03
-2.27877170e-01 9.65187848e-02 -7.72789598e-01 7.60124743e-01
8.55000466e-02 7.23418832e-01 -1.09091914e+00 9.32368562e-02
-2.46009622e-02 4.87018853e-01 -1.12852201e-01 2.73388863e-01
-1.45226866e-01 5.49014390e-01 -2.61473835e-01 5.47050178e-01
1.04011106e+00 2.19365597e-01 -7.89124072e-02 -2.37177581e-01
3.57738882e-01 -4.57223922e-01 -9.96959031e-01 4.64866340e-01
6.56680390e-02 7.14080632e-01 -1.93706080e-02 -7.47281790e-01
9.09696460e-01 4.12993014e-01 3.97368371e-02 -4.22552139e-01
3.29801768e-01 3.91645655e-02 1.19505897e-01 -2.16576338e-01
-4.64789346e-02 -7.70715177e-02 6.78679496e-02 4.05499279e-01
-6.30733147e-02 2.16833189e-01 -1.87147245e-01 3.54389697e-01
1.35052109e+00 -2.77086258e-01 -6.37778267e-03 2.49533176e-01
4.46754098e-01 -1.67084903e-01 7.65585005e-01 1.04193592e+00
-5.60775220e-01 4.55160618e-01 7.56726384e-01 -3.64039570e-01
-1.07411861e+00 -1.58554780e+00 1.91075563e-01 6.43829167e-01
4.10152495e-01 3.93068910e-01 -1.03805852e+00 -1.15925169e+00
-2.24370323e-02 6.24136508e-01 -8.60518992e-01 -6.78407013e-01
-1.96312234e-01 -4.40355986e-01 1.06245220e+00 3.00918788e-01
7.16523528e-01 -1.16954529e+00 -4.84884143e-01 -2.53013879e-01
9.74043682e-02 -1.15741599e+00 -4.24598962e-01 -4.67920117e-02
-3.57870221e-01 -1.23435283e+00 -4.37504172e-01 -6.19071066e-01
1.31237245e+00 2.30226904e-01 6.64739966e-01 1.92955539e-01
-3.10262799e-01 1.44239202e-01 -3.14774692e-01 -7.07611620e-01
-1.27451503e+00 -8.27773869e-01 1.60502538e-01 4.19743001e-01
-7.34011233e-02 -3.49842072e-01 -6.28133118e-01 2.82969773e-01
-1.18267107e+00 -1.74710125e-01 5.65953553e-01 7.32108414e-01
4.76365268e-01 2.37141773e-01 4.96344924e-01 -1.06604779e+00
3.88315469e-01 -4.11396325e-01 -7.73578227e-01 1.77658871e-01
-3.99450660e-01 -1.98000580e-01 1.19609821e+00 -9.22822058e-01
-7.72419870e-01 2.72731073e-02 6.87252823e-03 -1.05947161e+00
-4.94928092e-01 -1.83642700e-01 -4.80296940e-01 -2.60969669e-01
9.06793416e-01 5.65370381e-01 1.14451118e-01 2.10677132e-01
3.56971830e-01 3.13177377e-01 6.97306573e-01 1.77705400e-02
1.51561201e+00 8.02623272e-01 -1.20102860e-01 -4.93298650e-01
-6.19867682e-01 2.46478960e-01 -7.29969442e-02 -6.43000484e-01
7.33074009e-01 -7.71490335e-01 -4.48541254e-01 1.08909583e+00
-1.15199339e+00 -2.49933377e-01 -2.77337968e-01 1.25884995e-01
-2.18021244e-01 2.35122785e-01 -4.63688850e-01 -8.23580742e-01
-3.21918458e-01 -1.33929062e+00 7.35243320e-01 3.39502364e-01
2.55526423e-01 -8.61489058e-01 -2.53987402e-01 2.90559053e-01
2.24678829e-01 8.50120068e-01 8.21970999e-01 -1.03910971e+00
-6.86821282e-01 -5.96312940e-01 -1.11111425e-01 7.07030714e-01
1.05107427e-01 2.63162494e-01 -1.26489055e+00 -5.31392217e-01
2.82752186e-01 -2.91869253e-01 5.80146611e-01 8.09872374e-02
8.87266457e-01 -8.33188832e-01 -3.18498582e-01 6.14051640e-01
1.43528509e+00 7.36341596e-01 7.52668381e-01 2.38652140e-01
6.76519394e-01 3.21223348e-01 1.02419510e-01 2.18553469e-01
-2.65026391e-01 3.57315451e-01 1.18765569e+00 -3.73050988e-01
-1.13097027e-01 -3.00771892e-01 8.47367346e-01 -1.08088151e-01
2.65508473e-01 -6.65340781e-01 -4.67110395e-01 1.93277329e-01
-1.48217630e+00 -1.40814519e+00 -3.43237184e-02 2.21982145e+00
6.69497550e-01 4.73958015e-01 -3.64606939e-02 1.31343761e-02
1.20376039e+00 3.76960337e-01 -1.03341258e+00 -3.45933557e-01
-2.60696918e-01 -2.11704388e-01 6.95442080e-01 3.19206029e-01
-1.15057588e+00 7.99801886e-01 4.95495749e+00 5.20652413e-01
-1.20098269e+00 5.82143292e-02 9.17098999e-01 -1.04261696e-01
-3.13830912e-01 -1.98566288e-01 -5.08024991e-01 6.56740546e-01
5.18934786e-01 -1.71449438e-01 5.86358726e-01 8.92853618e-01
1.42435119e-01 4.76838291e-01 -8.84755552e-01 5.02037764e-01
3.96808267e-01 -1.13985801e+00 3.23362440e-01 1.41248122e-01
8.77667427e-01 -1.19496256e-01 5.91788411e-01 1.24118095e-02
7.57884085e-01 -9.15724277e-01 1.00084722e+00 3.09476256e-01
7.96769917e-01 -6.60751343e-01 8.19591999e-01 2.82403082e-01
-8.04019511e-01 -1.85225621e-01 -1.72788411e-01 4.17621851e-01
-2.53322929e-01 4.27986145e-01 -8.05909157e-01 9.53722745e-02
5.35364330e-01 2.53323793e-01 -5.48541963e-01 4.92068172e-01
-5.75717866e-01 6.03540659e-01 2.22007245e-01 2.74554551e-01
1.08777314e-01 -4.52335440e-02 1.12988460e+00 6.04100466e-01
9.59504694e-02 -2.92379826e-01 -5.53933717e-02 1.19498062e+00
-3.96596313e-01 -4.33037311e-01 -1.06091988e+00 1.57252401e-01
6.38904333e-01 1.20575750e+00 -5.91409206e-01 -3.66721511e-01
-1.27158225e-01 1.20095301e+00 2.57119667e-02 6.68574512e-01
-1.24337280e+00 -1.81955770e-01 7.52043247e-01 -1.26527861e-01
1.55832544e-01 5.05343020e-01 -7.52114952e-02 -1.00652206e+00
1.51211113e-01 -1.16443443e+00 1.28410146e-01 -8.44885051e-01
-1.33069992e+00 9.13027763e-01 -3.93270582e-01 -1.52853417e+00
-1.53317079e-01 -4.93258566e-01 -1.00698447e+00 7.00284600e-01
-9.05164897e-01 -9.62571383e-01 -5.57317019e-01 6.70374930e-01
2.81193763e-01 -3.84204745e-01 6.18002892e-01 -3.15933414e-02
-5.94993711e-01 1.05118597e+00 1.44826412e-01 7.08747327e-01
5.89532793e-01 -1.14755356e+00 5.65558374e-01 1.29998004e+00
2.10469991e-01 9.79269966e-02 8.40537071e-01 -6.56428754e-01
-1.11104035e+00 -1.67008305e+00 7.55537674e-02 -3.60856414e-01
7.51355112e-01 -4.39442754e-01 -7.98714876e-01 8.19129825e-01
-1.93929579e-02 2.45242819e-01 3.05157065e-01 -1.03301632e+00
-7.28438199e-01 -1.44585416e-01 -1.57375836e+00 1.00630617e+00
9.42851782e-01 -3.82132053e-01 -1.03694029e-01 3.56165916e-01
5.56809366e-01 -4.98227626e-01 -3.62467915e-01 3.42417330e-01
4.95713562e-01 -1.23291671e+00 1.09567368e+00 -5.43222785e-01
5.29766738e-01 -4.16641742e-01 -1.08533144e-01 -1.55959153e+00
4.88729961e-03 -4.71237689e-01 5.86501360e-02 1.17405117e+00
4.17834014e-01 -9.89553452e-01 7.59266138e-01 1.52340904e-01
1.66855901e-01 -6.02773368e-01 -9.71879184e-01 -9.16425347e-01
3.76278162e-02 -8.91574398e-02 5.84887803e-01 9.31366384e-01
-6.20733380e-01 -1.41212225e-01 -4.25343156e-01 7.59446740e-01
9.03995752e-01 -1.91211164e-01 8.53217661e-01 -8.81989121e-01
-2.14284748e-01 -3.54058683e-01 -8.60192537e-01 -4.19497848e-01
1.23667449e-01 -6.99189425e-01 4.01540063e-02 -8.79561782e-01
1.43132836e-01 -1.70659795e-01 -1.73460141e-01 5.34274042e-01
-3.27849239e-01 6.32660747e-01 3.38784575e-01 1.83402538e-01
-1.00085668e-01 1.90935969e-01 1.17547631e+00 -5.21091700e-01
6.07493967e-02 1.71062008e-01 -7.48110533e-01 9.00300682e-01
9.61210728e-01 -6.83119953e-01 -3.70454699e-01 -1.51200548e-01
-9.77749228e-02 -1.44271450e-02 9.80218589e-01 -1.05049706e+00
-2.38487959e-01 -3.55241187e-02 6.01858735e-01 -1.00457080e-01
1.72821075e-01 -9.36943650e-01 3.10854733e-01 9.00670290e-01
-3.67530555e-01 1.00945249e-01 4.24843393e-02 8.59735012e-01
1.27771199e-01 -2.82909304e-01 1.20891225e+00 -1.86472595e-01
-4.84918475e-01 3.42412293e-01 -4.84860331e-01 1.12759553e-01
1.52726090e+00 -2.65530765e-01 -1.11486828e+00 -5.39848149e-01
-7.14091003e-01 -1.96178108e-01 7.48244047e-01 4.96434033e-01
8.82783175e-01 -1.32867801e+00 -8.35346997e-01 5.93995869e-01
1.86376095e-01 -3.50827932e-01 2.27357075e-01 4.02550191e-01
-3.28969628e-01 -2.93539077e-01 -2.58151650e-01 -3.69783163e-01
-1.51350331e+00 7.82580078e-01 7.22272456e-01 -8.31413791e-02
-5.54290175e-01 7.79750168e-01 6.24649167e-01 -1.42511666e-01
2.12001368e-01 2.41987258e-01 -1.26306275e-02 -3.27767015e-01
5.22988379e-01 1.20179482e-01 -2.28297830e-01 -8.67247701e-01
-2.94424742e-01 7.45489597e-02 -1.19431943e-01 6.20586015e-02
7.61377573e-01 2.01615885e-01 1.26486659e-01 6.21478669e-02
1.26946414e+00 2.45226383e-01 -1.49608755e+00 7.38813076e-04
-5.89446425e-01 -6.07075393e-01 -2.81189680e-01 -8.79330337e-01
-1.58534765e+00 5.12348533e-01 1.06550419e+00 5.25588572e-01
1.10854113e+00 1.05938084e-01 5.99565506e-01 -2.69169584e-02
1.05854519e-01 -5.57976365e-01 3.96311283e-01 2.81291269e-02
8.94226789e-01 -1.10983932e+00 -4.13301080e-01 -1.26460448e-01
-6.83643043e-01 8.67904663e-01 9.71597075e-01 -4.31496322e-01
2.37390652e-01 3.43486428e-01 5.84103167e-01 -7.29388893e-02
-5.85986733e-01 2.23394677e-01 -4.98472899e-02 9.48173702e-01
-4.47212517e-01 2.14336827e-01 4.32682246e-01 1.65922746e-01
-2.91065782e-01 -6.53249502e-01 8.84785771e-01 5.61246514e-01
-2.72903532e-01 -5.57963550e-01 -8.21994603e-01 3.26121420e-01
-5.58626235e-01 -2.75927410e-02 -8.93543720e-01 5.87894797e-01
3.81127983e-01 9.23639715e-01 1.04921393e-01 -8.00723195e-01
2.59608418e-01 -2.18152523e-01 2.35289872e-01 -3.41588259e-01
-4.20084119e-01 -2.78236479e-01 -3.36611778e-01 -5.17986476e-01
-4.29226384e-02 -4.35261905e-01 -1.09623086e+00 -2.66307533e-01
-2.38601610e-01 -2.72667348e-01 4.11715627e-01 8.51125896e-01
2.04902917e-01 5.95939934e-01 1.25083482e+00 -6.47105634e-01
-7.49033213e-01 -5.95603049e-01 -3.21565896e-01 7.99571097e-01
6.97183967e-01 -4.43808705e-01 -1.14838445e+00 1.44049197e-01] | [5.569993019104004, 7.92805290222168] |
4b97bc0d-8cfc-4c6f-9f87-6673209efc00 | studying-the-control-of-non-invasive | 1511.06004 | null | http://arxiv.org/abs/1511.06004v1 | http://arxiv.org/pdf/1511.06004v1.pdf | Studying the control of non invasive prosthetic hands over large time spans | The electromyography (EMG) signal is the electrical manifestation of a
neuromuscular activation that provides access to physiological processes which
cause the muscle to generate force and produce movement. Non invasive
prostheses use such signals detected by the electrodes placed on the user's
stump, as input to generate hand posture movements according to the intentions
of the prosthesis wearer. The aim of this pilot study is to explore the
repeatability issue, i.e. the ability to classify 17 different hand postures,
represented by EMG signal, across a time span of days by a control algorithm.
Data collection experiments lasted four days and signals were collected from
the forearm of a single subject. We find that Support Vector Machine (SVM)
classification results are high enough to guarantee a correct classification of
more than 10 postures in each moment of the considered time span. | ['Mara Graziani'] | 2015-11-18 | null | null | null | null | ['electromyography-emg'] | ['medical'] | [ 5.02145350e-01 1.70605138e-01 -3.06985646e-01 2.00947180e-01
4.69275266e-02 -4.20972437e-01 2.73506165e-01 -5.74646831e-01
-4.67640102e-01 1.02258122e+00 -9.63503495e-02 -1.10883944e-01
-2.70088285e-01 -3.77642125e-01 -4.95468318e-01 -6.17689490e-01
-1.90404341e-01 1.59381390e-01 2.99818162e-02 -7.21210837e-02
2.10335538e-01 6.28028631e-01 -1.67921185e+00 3.58749241e-01
5.80150902e-01 8.16997111e-01 5.83154678e-01 3.86099756e-01
3.74298751e-01 -7.03216121e-02 -9.96647358e-01 9.00634080e-02
1.63928568e-01 -5.48453033e-01 -4.29649264e-01 1.09761894e-01
-5.71158528e-02 -1.05004676e-01 -9.09984335e-02 7.38584876e-01
6.65105224e-01 1.42857945e-02 6.51240408e-01 -9.14798796e-01
-1.65106043e-01 3.04065138e-01 5.70133217e-02 1.12412095e-01
6.58402979e-01 4.13274914e-01 3.55384201e-01 -3.64464492e-01
1.08105469e+00 4.71751392e-01 4.90352184e-01 7.94226944e-01
-1.30357265e+00 -4.79833126e-01 -5.36285102e-01 4.70433295e-01
-1.05015910e+00 -2.45859817e-01 9.47231233e-01 -7.18934774e-01
8.55621517e-01 7.57238567e-01 1.22231019e+00 1.72912467e+00
8.59616578e-01 4.28479195e-01 1.42127573e+00 -5.38769245e-01
4.52916294e-01 2.37631783e-01 -1.28058657e-01 -2.28701711e-01
3.83785486e-01 3.02936137e-01 -4.05855238e-01 -1.08242102e-01
9.96778309e-01 -2.35162526e-01 -7.66220033e-01 9.24879462e-02
-1.08097196e+00 2.57236391e-01 1.89221203e-01 9.84201014e-01
-1.17279553e+00 -6.61547780e-02 3.70503902e-01 5.40812671e-01
-1.07081801e-01 7.51907051e-01 -6.28581941e-01 -8.44296813e-01
-7.11970389e-01 1.99798360e-01 9.71881926e-01 3.49768758e-01
-2.77071953e-01 -3.92910019e-02 -2.28490666e-01 5.58308065e-01
-3.68299372e-02 3.88909161e-01 6.55520439e-01 -7.99163640e-01
1.29016697e-01 5.75015724e-01 1.00288279e-01 -7.72728264e-01
-5.69041073e-01 -4.41342443e-01 -3.70580763e-01 6.91579044e-01
4.70553666e-01 -5.22417128e-01 -5.18226624e-01 1.51988041e+00
1.15993112e-01 -2.37570837e-01 -2.27595091e-01 1.21989870e+00
9.39639136e-02 1.15158722e-01 -6.74609914e-02 -6.52453780e-01
1.34548712e+00 -4.56782579e-02 -8.06152642e-01 -1.10315897e-01
2.67992198e-01 -5.61840355e-01 1.08729994e+00 9.67394352e-01
-8.97110701e-01 -6.14630640e-01 -1.09590685e+00 8.28304589e-01
1.60628092e-02 3.91768068e-01 4.10988510e-01 7.00696290e-01
-1.80918574e-01 1.14092302e+00 -8.23732793e-01 -2.43580639e-01
-3.25562865e-01 4.82602954e-01 -5.51929176e-01 7.28241086e-01
-1.19411731e+00 1.10573959e+00 2.35156670e-01 5.63824594e-01
9.41583291e-02 -3.90578032e-01 -1.37453899e-01 -3.95602226e-01
-1.69611908e-02 -5.72058380e-01 7.26602912e-01 -8.40279520e-01
-1.78408909e+00 9.91784155e-01 1.13980398e-01 -1.00899227e-01
7.98993587e-01 -1.78283215e-01 -4.31099117e-01 -9.90139693e-03
-2.17663661e-01 -1.00576259e-01 8.42920899e-01 -7.45732665e-01
8.99797678e-02 -7.42490888e-01 -6.14388406e-01 -1.01128809e-01
-1.11109309e-01 -7.31816143e-02 2.25830935e-02 -5.99023700e-01
7.90158808e-02 -1.21744311e+00 7.77212828e-02 -3.00230443e-01
-4.43993568e-01 -2.32356697e-01 4.03035969e-01 -1.06848645e+00
1.18802083e+00 -2.24467158e+00 7.33047426e-01 5.65420926e-01
-3.18957686e-01 2.80996680e-01 3.32172930e-01 5.05071342e-01
-2.09574550e-01 -4.06912595e-01 -8.07527900e-02 4.78754789e-01
-2.05593526e-01 3.81531417e-01 2.17287228e-01 5.63745618e-01
-2.61178374e-01 5.99888623e-01 -4.41358775e-01 3.59900817e-02
1.98036477e-01 3.24869514e-01 1.68426082e-01 3.01538974e-01
-3.87777872e-02 5.84245861e-01 -2.05798894e-01 4.39172804e-01
2.76897460e-01 3.81119817e-01 4.68340129e-01 -4.36309814e-01
6.33139908e-02 1.88356236e-01 -1.22039449e+00 1.44461358e+00
-3.59226882e-01 5.62419415e-01 9.87560749e-02 -7.80308366e-01
1.16320550e+00 8.34977090e-01 6.65948987e-01 -6.78341687e-01
4.53538418e-01 7.53762424e-01 5.19348979e-01 -9.66072738e-01
-3.17759037e-01 -1.66130185e-01 1.79404303e-01 3.99669409e-01
-2.23778099e-01 2.19098881e-01 9.51476209e-03 -7.71403134e-01
1.03132176e+00 2.43071124e-01 2.74797946e-01 -2.27872923e-01
2.68205047e-01 -1.12773493e-01 1.28634870e-01 2.59606957e-01
6.92366511e-02 1.76031783e-01 2.89299399e-01 -5.71590178e-02
-6.25916243e-01 -9.33897197e-01 -3.20657849e-01 3.99506032e-01
-2.43114337e-01 3.24967295e-01 -8.47723365e-01 -6.47782981e-02
4.13022250e-01 6.20181262e-01 -5.24762094e-01 -4.33709532e-01
-5.65210938e-01 -3.05158764e-01 2.05625877e-01 5.35361826e-01
-1.83430195e-01 -1.76321709e+00 -1.20659959e+00 6.30233288e-01
-6.29239976e-02 -7.98539579e-01 1.39132485e-01 3.04099381e-01
-1.13530469e+00 -1.11480331e+00 -8.10012639e-01 -3.74031723e-01
4.06270087e-01 -5.44183373e-01 2.84375250e-01 -1.73962608e-01
-5.85769653e-01 2.29627028e-01 -4.53939408e-01 -3.58221859e-01
-4.64185506e-01 4.20190468e-02 4.39038277e-01 -1.98503152e-01
3.79377306e-01 -9.60696280e-01 -5.20542741e-01 4.04035628e-01
-3.18735331e-01 -1.71138093e-01 8.21903288e-01 4.92179692e-01
5.05920708e-01 -2.66182452e-01 5.90719461e-01 -2.25871786e-01
1.25433648e+00 -1.80128664e-01 -6.13616873e-03 5.59590347e-02
-4.80867267e-01 -1.71657518e-01 3.43287021e-01 -8.79156351e-01
-7.18536258e-01 -6.20251633e-02 -2.46755436e-01 -1.47030160e-01
-4.33295101e-01 3.45330715e-01 -1.46150813e-01 -6.14120178e-02
1.08975840e+00 4.38208252e-01 5.94599426e-01 -6.30996406e-01
-3.60102020e-02 1.16341698e+00 9.57090974e-01 -4.60483223e-01
2.12228820e-01 -6.17655739e-02 1.67374164e-01 -9.80557919e-01
2.96905309e-01 -3.64986956e-01 -6.14170015e-01 -8.75404656e-01
5.49908817e-01 -1.67538732e-01 -9.24771428e-01 4.16906416e-01
-1.12837923e+00 -3.63574803e-01 -8.39956924e-02 9.88440275e-01
-8.48112464e-01 6.78554997e-02 -5.33790708e-01 -1.15868402e+00
-6.88325286e-01 -6.60205185e-01 8.20029557e-01 -9.14879292e-02
-1.16036880e+00 -3.92021447e-01 -1.74568817e-02 2.19722331e-01
9.45682451e-02 6.97317898e-01 4.85404611e-01 -1.47069246e-01
1.97050765e-01 -6.83829606e-01 5.44328868e-01 4.59043145e-01
3.53075087e-01 -2.33048171e-01 -7.66547143e-01 -3.40265363e-01
6.09259129e-01 1.88257992e-01 2.15330586e-01 3.88500631e-01
7.42649317e-01 -1.36496574e-01 -3.08218449e-01 -4.10329066e-02
1.18454480e+00 6.22692585e-01 1.11187923e+00 2.02703640e-01
1.85516581e-01 8.98497045e-01 6.06704235e-01 -2.52695084e-02
-8.19331527e-01 1.15245402e+00 -1.67732406e-02 2.67054737e-01
-1.16101071e-01 2.89252281e-01 3.81251663e-01 4.49785173e-01
-9.70498025e-01 3.18285644e-01 -3.81178349e-01 2.08513632e-01
-1.42424011e+00 -9.46228385e-01 -6.31874800e-01 2.30231786e+00
8.67700040e-01 3.02470595e-01 3.43110144e-01 9.12297606e-01
4.99978065e-01 -4.88222182e-01 -5.17269194e-01 -5.26911438e-01
3.24780673e-01 5.78239322e-01 4.35857624e-01 1.16820350e-01
-1.92703500e-01 -1.04494458e-02 6.38104820e+00 2.26895019e-01
-1.66565311e+00 -1.48110867e-01 -5.24193168e-01 -3.96263689e-01
8.77275020e-02 -4.55922037e-01 -2.32673362e-01 1.02143753e+00
9.25967574e-01 -1.51461080e-01 4.52136934e-01 7.83980668e-01
6.13399804e-01 -2.39236280e-01 -9.82241154e-01 8.04305911e-01
-1.49854377e-01 -9.46277142e-01 -6.82726324e-01 3.45909566e-01
3.36128980e-01 -3.10905933e-01 -2.13149011e-01 -2.11011127e-01
-1.10132515e+00 -7.28808105e-01 5.82445741e-01 9.72458541e-01
8.02470028e-01 -1.33057415e-01 6.39711440e-01 5.98961651e-01
-5.36561430e-01 -2.23810717e-01 9.28345397e-02 -4.50434208e-01
4.38183427e-01 4.79129285e-01 -6.44410908e-01 3.30271065e-01
2.19172806e-01 1.18245609e-01 4.62050028e-02 1.03415263e+00
-4.13377315e-01 5.87556839e-01 -4.45621550e-01 -5.06715715e-01
-2.56381243e-01 -1.49643600e-01 1.06009853e+00 7.07771599e-01
3.93496633e-01 -2.36993715e-01 -6.64259493e-01 1.03577614e+00
6.28590703e-01 9.88886040e-03 -5.73923409e-01 -9.57631171e-02
3.56062561e-01 8.77102852e-01 -3.82094353e-01 1.63836077e-01
1.75417125e-01 1.16999805e+00 -4.46132332e-01 3.04947853e-01
-3.14761460e-01 -5.80774844e-01 4.95116681e-01 5.08590579e-01
-8.44826177e-02 -2.30509996e-01 -6.97966039e-01 -5.04154921e-01
1.02812314e+00 -7.41551697e-01 -1.17447868e-01 -8.52522075e-01
-9.02156234e-01 1.39293373e-01 -4.72933874e-02 -1.21877587e+00
-7.94099212e-01 -1.09908104e+00 -6.65125906e-01 1.24411571e+00
-1.97308466e-01 -3.17460537e-01 -1.85860142e-01 2.45866701e-01
2.71119416e-01 6.77570179e-02 9.83582437e-01 1.77033395e-01
-2.58464724e-01 2.26342991e-01 -1.83254063e-01 -2.88380444e-01
3.93940300e-01 -1.06851852e+00 -5.34265302e-02 5.14354825e-01
-5.32144867e-02 9.26631808e-01 1.14141917e+00 -7.94944465e-01
-1.39162588e+00 -1.59722671e-01 1.01265538e+00 -2.48595774e-01
3.03153992e-01 -7.49052316e-02 -8.24978590e-01 9.49108526e-02
-2.75589913e-01 -6.09632552e-01 6.52724564e-01 -1.03686951e-01
4.19486076e-01 -5.17544448e-02 -1.15445685e+00 4.89789844e-01
1.15149474e+00 -5.67146659e-01 -1.04667258e+00 7.72820413e-02
-2.37932414e-01 -4.05289471e-01 -1.38948917e+00 3.53013128e-01
1.42880762e+00 -6.11781776e-01 7.16102064e-01 -5.46177089e-01
1.88671708e-01 4.74670064e-03 2.66530901e-01 -1.39384258e+00
3.70079167e-02 -5.96700430e-01 -2.64464289e-01 6.05429709e-01
2.82721728e-01 -8.43065083e-01 7.41825461e-01 3.72369111e-01
1.06194437e-01 -8.29609811e-01 -1.27230048e+00 -1.13526142e+00
-4.29446489e-01 -5.03226817e-01 2.91580796e-01 4.31457341e-01
8.83783460e-01 4.71294187e-02 -4.26674902e-01 -3.07148218e-01
4.87092346e-01 1.02763414e-01 6.79301322e-01 -1.39988172e+00
-6.15754485e-01 -4.76990491e-01 -8.25870097e-01 -4.13428992e-01
-1.07709996e-01 -6.68322980e-01 -3.59446108e-02 -1.57179439e+00
-4.18251365e-01 1.57706931e-01 -1.92547128e-01 2.76270270e-01
1.63594007e-01 -1.37527008e-02 4.26896550e-02 1.85957342e-01
8.58792961e-01 -1.38080716e-01 1.37607288e+00 1.77608073e-01
-6.39463663e-01 6.55762672e-01 -3.34564388e-01 1.29208043e-01
7.43584037e-01 -3.89732867e-01 -3.68461490e-01 3.03314239e-01
-8.00444856e-02 3.94308627e-01 4.49821472e-01 -1.13148928e+00
-7.71834627e-02 -4.37806770e-02 5.79612434e-01 -2.15999126e-01
2.32612565e-01 -1.10530162e+00 1.09966123e+00 1.17155993e+00
-1.11327231e-01 -5.21639645e-01 -1.03375815e-01 1.59831420e-01
2.72703208e-02 -3.09687525e-01 2.58712500e-01 1.49146557e-01
-4.74287093e-01 -4.23778117e-01 -6.06144190e-01 -7.45034158e-01
1.17216909e+00 -9.82814193e-01 2.22018510e-01 8.17781612e-02
-1.26837981e+00 -4.75328416e-01 1.92211732e-01 5.69250047e-01
2.47109979e-01 -1.26901484e+00 -3.25371951e-01 3.80986452e-01
1.84497293e-02 -9.47346330e-01 2.33694196e-01 1.23721254e+00
-2.47324765e-01 4.71221924e-01 -8.63359392e-01 -5.67121923e-01
-1.57338536e+00 1.27068684e-01 4.58113700e-01 1.94913521e-01
-9.59504068e-01 3.32361609e-01 -1.00696635e+00 2.97082007e-01
9.93265808e-02 -1.97773069e-01 -3.12028140e-01 -7.18938559e-02
3.05137604e-01 5.38156986e-01 2.87531465e-01 -2.49146745e-01
-2.78921962e-01 5.46618342e-01 7.14606106e-01 -2.16218323e-01
9.52058434e-01 4.15752143e-01 6.92858920e-02 6.95294023e-01
9.15431023e-01 -2.59011909e-02 -8.68859828e-01 5.33900142e-01
-6.43140972e-02 -5.65936327e-01 -2.59527087e-01 -1.10720730e+00
-5.06397545e-01 5.68638742e-01 1.01165473e+00 2.49509349e-01
1.21420622e+00 -9.05425772e-02 6.88948393e-01 1.91471532e-01
9.40703690e-01 -1.40574908e+00 -3.59910339e-01 -5.22006273e-01
1.47689152e+00 -3.12878549e-01 -2.78947383e-01 -6.02955341e-01
-3.94360006e-01 1.28854370e+00 2.60199457e-01 -1.63611174e-01
4.07356679e-01 2.17333138e-01 5.34017710e-03 -1.13687836e-01
-2.56013811e-01 8.68824217e-03 5.95547616e-01 8.60247254e-01
4.01414394e-01 6.51343405e-01 -1.63834167e+00 7.76210368e-01
-3.65875840e-01 8.57469320e-01 1.36239931e-01 9.48242724e-01
-3.77453297e-01 -1.20292842e+00 -4.81961399e-01 7.85074830e-01
-1.27770543e-01 8.24473441e-01 -5.46753228e-01 9.58173156e-01
2.29098052e-01 7.97527254e-01 -9.98040140e-02 -6.49035990e-01
9.97678638e-01 3.18694502e-01 1.06830299e+00 -3.20799768e-01
-5.61501682e-01 -8.29291642e-02 2.81989396e-01 -7.44065523e-01
-1.29718989e-01 -8.66802692e-01 -1.26300490e+00 1.09411657e-01
-6.22753985e-02 1.18038945e-01 1.14696217e+00 9.98049974e-01
-3.75713632e-02 4.97173876e-01 5.27515650e-01 -9.37997103e-01
-1.04555333e+00 -1.57741332e+00 -1.05849159e+00 6.73885107e-01
-2.97301739e-01 -8.82176578e-01 -4.44966912e-01 -1.14416696e-01] | [6.874439239501953, 0.22244961559772491] |
3938254d-16b3-4dc5-a670-bf211bdfae84 | polarity-loss-for-zero-shot-object-detection | 1811.08982 | null | https://arxiv.org/abs/1811.08982v3 | https://arxiv.org/pdf/1811.08982v3.pdf | Polarity Loss for Zero-shot Object Detection | Conventional object detection models require large amounts of training data. In comparison, humans can recognize previously unseen objects by merely knowing their semantic description. To mimic similar behaviour, zero-shot object detection aims to recognize and localize 'unseen' object instances by using only their semantic information. The model is first trained to learn the relationships between visual and semantic domains for seen objects, later transferring the acquired knowledge to totally unseen objects. This setting gives rise to the need for correct alignment between visual and semantic concepts, so that the unseen objects can be identified using only their semantic attributes. In this paper, we propose a novel loss function called 'Polarity loss', that promotes correct visual-semantic alignment for an improved zero-shot object detection. On one hand, it refines the noisy semantic embeddings via metric learning on a 'Semantic vocabulary' of related concepts to establish a better synergy between visual and semantic domains. On the other hand, it explicitly maximizes the gap between positive and negative predictions to achieve better discrimination between seen, unseen and background objects. Our approach is inspired by embodiment theories in cognitive science, that claim human semantic understanding to be grounded in past experiences (seen objects), related linguistic concepts (word vocabulary) and visual perception (seen/unseen object images). We conduct extensive evaluations on MS-COCO and Pascal VOC datasets, showing significant improvements over state of the art. | ['Nick Barnes', 'Shafin Rahman', 'Salman Khan'] | 2018-11-22 | null | null | null | null | ['zero-shot-object-detection'] | ['computer-vision'] | [ 2.89967448e-01 1.64227411e-01 -1.29255414e-01 -5.93841970e-01
-2.25867167e-01 -3.85847658e-01 8.74117136e-01 5.49200296e-01
-5.82348466e-01 3.13757300e-01 1.53194740e-01 1.39112353e-01
-3.25877443e-02 -8.49054098e-01 -6.84131742e-01 -6.62516236e-01
1.35173634e-01 2.47987345e-01 4.59385633e-01 -2.63344705e-01
2.42959827e-01 3.10281396e-01 -1.91008770e+00 4.77204293e-01
5.85650325e-01 1.22839868e+00 6.10792518e-01 3.54213715e-01
-3.44983369e-01 8.59828532e-01 -3.60427529e-01 -4.43762243e-01
1.49587452e-01 -3.55550975e-01 -7.55129516e-01 1.60326064e-01
5.47077537e-01 -8.75159651e-02 -1.50378510e-01 1.42280042e+00
1.87134430e-01 3.73561561e-01 6.92576289e-01 -1.42725515e+00
-1.29933119e+00 1.94976494e-01 -3.46971452e-01 4.07830268e-01
2.87467122e-01 1.86885789e-01 1.20028567e+00 -1.01393437e+00
5.27468622e-01 1.32991314e+00 1.72971904e-01 7.11790383e-01
-1.26506078e+00 -3.73234361e-01 4.39959288e-01 6.26668692e-01
-1.34818149e+00 -2.78051853e-01 8.24798405e-01 -5.70766389e-01
9.19934690e-01 1.30919769e-01 7.13517666e-01 1.20650232e+00
-1.85110882e-01 7.20012546e-01 9.04936016e-01 -4.90707010e-01
4.84767050e-01 6.10670865e-01 2.69158036e-01 3.96913797e-01
3.60857517e-01 3.59598428e-01 -4.02137041e-01 3.00499856e-01
2.89039195e-01 4.29737926e-01 -2.34266028e-01 -7.43675649e-01
-1.21895301e+00 7.24713027e-01 8.73746634e-01 6.03134215e-01
-3.49341989e-01 1.68884844e-02 1.96148440e-01 9.71822068e-02
1.99612156e-01 3.85670245e-01 -1.79221615e-01 2.97460914e-01
-5.18188596e-01 -2.11682804e-02 4.73377824e-01 9.35882986e-01
1.10400045e+00 -1.01941802e-01 -3.40037972e-01 6.32169127e-01
5.10517538e-01 5.75340271e-01 7.11975217e-01 -5.10081530e-01
2.37491027e-01 8.84364784e-01 5.29316776e-02 -1.20241225e+00
-6.27581999e-02 -4.34478968e-01 -5.14214277e-01 3.66412133e-01
1.46276757e-01 3.91580760e-01 -1.19941032e+00 1.97137511e+00
3.14357311e-01 3.71530712e-01 3.86783123e-01 1.26914799e+00
9.61544156e-01 4.51902300e-01 4.67088789e-01 1.07840993e-01
1.67985690e+00 -7.75356114e-01 -5.91663897e-01 -8.20478380e-01
4.48155969e-01 -4.28731143e-01 1.14774179e+00 -5.28100953e-02
-5.89620829e-01 -7.66891539e-01 -1.23086488e+00 -2.38161981e-01
-9.52861249e-01 2.40915548e-02 4.54514563e-01 4.46546257e-01
-6.97202086e-01 4.35681581e-01 -5.69014847e-01 -6.12841845e-01
6.96496189e-01 1.14316851e-01 -4.46555406e-01 -7.60751516e-02
-1.12527764e+00 1.07831168e+00 1.04547489e+00 -4.48068529e-02
-1.30020630e+00 -5.85721314e-01 -1.16550887e+00 1.86800957e-01
6.25671506e-01 -6.13913894e-01 9.81452942e-01 -1.45313203e+00
-8.76521647e-01 1.25773656e+00 -5.81953377e-02 -5.70055664e-01
1.87206179e-01 -1.46739051e-01 -6.02382779e-01 1.99918345e-01
1.79065377e-01 8.19717109e-01 1.03477240e+00 -1.52469361e+00
-8.11139703e-01 -5.51931083e-01 2.12846443e-01 2.53433943e-01
-4.98321533e-01 -2.66479433e-01 -9.01629180e-02 -4.76859897e-01
2.12403551e-01 -5.55928946e-01 4.08797786e-02 3.09120893e-01
-6.44740313e-02 -3.68892968e-01 1.04904234e+00 -3.11437726e-01
6.62063241e-01 -2.43261552e+00 8.34245607e-02 -5.97206503e-02
3.51912111e-01 3.50330621e-01 -4.45458412e-01 3.44646387e-02
-1.03001595e-01 -1.69025123e-01 -8.36998224e-02 -1.30172297e-01
6.37888489e-03 4.36404347e-01 -4.39834267e-01 3.47104251e-01
4.79260385e-01 1.11830556e+00 -1.26793981e+00 -3.76522332e-01
5.16688108e-01 4.04196143e-01 -3.66220117e-01 2.48656198e-01
-2.42750347e-01 2.33743057e-01 -4.42701995e-01 5.77027202e-01
4.89144862e-01 -3.09175789e-01 4.56752256e-02 -3.31669033e-01
1.90390751e-01 -6.05200902e-02 -1.08912551e+00 1.66004205e+00
-4.26115334e-01 6.33071601e-01 -3.27697039e-01 -1.33379006e+00
1.04969108e+00 1.45264223e-01 8.59254003e-02 -1.02340865e+00
3.35648119e-01 3.44922021e-02 -6.51291758e-02 -7.78893411e-01
1.89512789e-01 -5.46858251e-01 3.46484222e-02 5.48017547e-02
4.22911644e-01 1.24342702e-01 2.59182025e-02 3.25893432e-01
5.16016185e-01 -9.20885988e-03 5.33821285e-01 -6.85658678e-02
5.99047244e-01 3.17631252e-02 4.31527227e-01 6.91539526e-01
-3.95762175e-01 5.21829367e-01 1.24845199e-01 -3.61391962e-01
-9.31207120e-01 -1.49222791e+00 -4.51414324e-02 1.38935554e+00
7.53749847e-01 -1.39972329e-01 -4.85453784e-01 -6.61411703e-01
1.74252167e-01 1.05445886e+00 -9.46070611e-01 -5.32587945e-01
-1.47697195e-01 -4.69165534e-01 4.44705337e-02 8.16084385e-01
2.94642329e-01 -1.21503437e+00 -9.57709312e-01 2.66695116e-02
1.03005238e-01 -1.07401645e+00 -9.77861881e-02 2.15347230e-01
-5.96556187e-01 -1.15026212e+00 -4.66189116e-01 -9.47500706e-01
7.40335643e-01 5.07734954e-01 1.06821525e+00 1.03333713e-02
-6.56901777e-01 6.29008174e-01 -6.45786345e-01 -6.56293213e-01
-2.72159040e-01 -5.59370637e-01 1.64721757e-01 4.72085804e-01
8.77903700e-01 -3.18261296e-01 -7.11468816e-01 2.15803415e-01
-9.47211266e-01 -1.37854308e-01 6.75810039e-01 8.09361756e-01
5.75187743e-01 -2.65787333e-01 4.95500714e-01 -4.03699726e-01
1.93668619e-01 -5.11356115e-01 -3.15827876e-01 5.40519238e-01
-4.40363735e-01 1.54391199e-01 4.20074314e-01 -7.05268443e-01
-1.13197553e+00 -3.17675658e-02 3.59875858e-01 -6.88088894e-01
-4.14271384e-01 6.45381436e-02 -3.08177173e-01 6.69567287e-02
6.84506595e-01 4.05707866e-01 -7.89589360e-02 -4.57016587e-01
8.41522574e-01 5.93388855e-01 7.14706898e-01 -4.44377512e-01
7.37579405e-01 8.84734333e-01 -2.55606055e-01 -6.95351779e-01
-1.17342067e+00 -8.50637615e-01 -7.99627125e-01 -4.13493849e-02
1.17951286e+00 -8.43180239e-01 -6.14928961e-01 -6.76552877e-02
-1.11856675e+00 2.76829153e-01 -6.88195765e-01 6.59411430e-01
-5.82834899e-01 2.37872556e-01 3.36258449e-02 -7.95411766e-01
-1.33374915e-01 -7.25572586e-01 1.24449027e+00 3.33110422e-01
-1.60823077e-01 -8.74251187e-01 -1.18517362e-01 1.46875933e-01
1.95024088e-01 1.63905919e-01 1.00057578e+00 -9.98067915e-01
-5.05170286e-01 -1.36849001e-01 -6.70627952e-01 6.42553389e-01
9.90399122e-02 -4.46998268e-01 -1.14053500e+00 -4.12944794e-01
3.14949036e-01 -3.10427725e-01 9.88904953e-01 1.78191867e-02
8.45605612e-01 -2.71656364e-01 -5.55038512e-01 3.01460505e-01
1.67598915e+00 2.36238867e-01 4.74226028e-01 6.78465590e-02
7.05820143e-01 7.94677734e-01 7.28394330e-01 2.17826813e-01
1.06242716e-01 4.41742092e-01 6.04520977e-01 -4.17780876e-03
-1.63731828e-01 -5.09199381e-01 1.08553924e-01 3.75141948e-01
5.14240600e-02 -1.20290101e-01 -8.87983322e-01 7.66395211e-01
-1.75937676e+00 -9.97547507e-01 2.19284549e-01 2.18209219e+00
3.94431353e-01 1.44224137e-01 -6.83486313e-02 -3.17392312e-02
8.40982080e-01 9.96609330e-02 -6.22825801e-01 -1.28206283e-01
-1.79668471e-01 4.00434248e-02 1.85007080e-01 8.93053636e-02
-9.82147634e-01 1.02437735e+00 5.16105843e+00 6.85449302e-01
-1.08115196e+00 4.34494764e-01 2.43659049e-01 -1.82411253e-01
-4.90506887e-02 2.04253450e-01 -7.29570329e-01 3.51253539e-01
4.48537797e-01 -4.64099675e-01 1.64402395e-01 1.07753193e+00
-1.51024640e-01 3.05203218e-02 -1.58030438e+00 1.10392654e+00
4.73201036e-01 -1.07070291e+00 4.52027857e-01 -2.27538362e-01
4.91596341e-01 -1.79079443e-01 1.51692927e-01 4.77139741e-01
3.36841233e-02 -1.00126672e+00 9.36554253e-01 7.02023029e-01
4.13287103e-01 -3.82407874e-01 6.09297156e-01 3.45321238e-01
-1.18322968e+00 -3.84920955e-01 -7.45845377e-01 -1.61746398e-01
-2.03721114e-02 2.01590940e-01 -8.04536104e-01 3.34533781e-01
7.18113661e-01 9.02568698e-01 -6.91444576e-01 8.82177114e-01
-4.31365639e-01 8.11475962e-02 7.79676512e-02 -1.69135883e-01
2.91711748e-01 1.18653261e-05 5.79453349e-01 8.80159974e-01
1.27468392e-01 2.67481029e-01 2.07690924e-01 1.36615002e+00
8.99739042e-02 1.57886922e-01 -6.10612988e-01 -1.16083123e-01
3.32727611e-01 1.11216187e+00 -6.71695352e-01 -6.02449536e-01
-5.02476513e-01 1.03823793e+00 4.27965909e-01 4.28736448e-01
-8.14281106e-01 -2.97302753e-01 7.31037378e-01 7.15067796e-03
4.85949188e-01 4.35297675e-02 3.79692316e-02 -1.12678599e+00
6.95950836e-02 -3.67796302e-01 4.20840770e-01 -8.51294935e-01
-1.56398237e+00 4.16610926e-01 -6.56199828e-02 -1.34747541e+00
6.25554249e-02 -9.85391676e-01 -4.86800164e-01 6.75849438e-01
-1.54611301e+00 -1.24987423e+00 -5.31464219e-01 5.33892214e-01
5.25520980e-01 -4.29567024e-02 6.66053593e-01 1.03794649e-01
-1.95353236e-02 3.17061454e-01 -1.22835405e-01 2.15638116e-01
4.55652446e-01 -1.19723141e+00 1.48545653e-01 6.06979668e-01
5.26587367e-01 5.54630756e-01 6.85253024e-01 -5.48029602e-01
-1.17069292e+00 -1.12108779e+00 6.40792072e-01 -6.11338794e-01
6.95063829e-01 -4.89283800e-01 -1.20982814e+00 5.97762465e-01
-1.94323719e-01 2.84151375e-01 5.87061465e-01 -3.59434895e-02
-7.65840352e-01 -1.43362001e-01 -1.00735009e+00 4.20046419e-01
1.17622375e+00 -7.65754580e-01 -1.31321657e+00 1.80212855e-01
8.47860575e-01 9.18911248e-02 -5.23496389e-01 4.73655343e-01
4.80419159e-01 -8.38861406e-01 1.20433152e+00 -1.06532621e+00
1.07385725e-01 -4.66212720e-01 -4.43512619e-01 -1.15915656e+00
-2.77910233e-01 2.00661808e-01 -8.77566915e-03 1.00864589e+00
9.44503769e-02 -4.24975276e-01 4.94388759e-01 3.94145995e-01
8.46754014e-02 -4.55944687e-01 -8.86022568e-01 -1.10856068e+00
-1.34516388e-01 -4.46366608e-01 5.55464625e-01 1.01786232e+00
-4.58108336e-02 5.72031617e-01 -1.25324935e-01 2.83389360e-01
6.37383103e-01 3.07700425e-01 3.28686059e-01 -1.38711286e+00
-1.52556807e-01 -4.14505571e-01 -1.13586318e+00 -8.03158522e-01
1.60982519e-01 -1.21876168e+00 1.09203890e-01 -1.44221544e+00
4.76821780e-01 -1.34385377e-01 -7.76062310e-01 5.67195177e-01
-2.14330390e-01 3.48426104e-01 4.69378412e-01 3.33664566e-02
-9.70344007e-01 7.39096403e-01 1.14896214e+00 -4.15158421e-01
-2.75123045e-02 -3.93114924e-01 -7.64938653e-01 7.52336323e-01
4.78765368e-01 -4.99238580e-01 -4.93535459e-01 -4.77033585e-01
1.03596836e-01 -4.25485641e-01 1.10431194e+00 -1.13368595e+00
1.85839027e-01 -1.63889632e-01 4.46435899e-01 -3.01162750e-01
4.17981178e-01 -1.04213500e+00 -2.73082078e-01 5.24556875e-01
-3.68728936e-01 -4.40361857e-01 1.13875963e-01 1.02481091e+00
-2.95865238e-01 -3.54308486e-01 9.99060988e-01 -1.51341841e-01
-1.45373642e+00 1.05489552e-01 1.24408059e-01 1.75647393e-01
1.43672001e+00 -5.05790710e-01 -3.28989446e-01 2.37064864e-02
-1.05191636e+00 1.51998803e-01 3.41639787e-01 9.23601210e-01
9.73918796e-01 -1.36894763e+00 -4.67876196e-01 3.74002188e-01
7.19993055e-01 -3.40984046e-01 3.64578128e-01 5.21020591e-01
3.91405039e-02 3.34484190e-01 -3.22625965e-01 -8.35184634e-01
-1.19417453e+00 1.26315677e+00 3.58941525e-01 5.82011282e-01
-5.93989730e-01 9.06724811e-01 7.52418220e-01 -2.47860432e-01
2.06985861e-01 -3.82194668e-01 -2.18420669e-01 1.90771013e-01
7.26370215e-01 2.29300529e-01 -2.17162356e-01 -8.40799093e-01
-5.11474252e-01 6.19691730e-01 -5.77939041e-02 1.13484509e-01
1.05192447e+00 -1.85175702e-01 2.71264631e-02 5.64794779e-01
1.48928893e+00 -4.22963887e-01 -1.09888613e+00 -5.74551105e-01
2.18915462e-01 -6.31934583e-01 -7.00822398e-02 -7.76135027e-01
-7.94204175e-01 1.20584488e+00 1.03408968e+00 5.67856170e-02
9.86270547e-01 5.81852257e-01 4.60494787e-01 3.54551524e-01
3.17144930e-01 -9.82026339e-01 6.16365612e-01 2.22515836e-01
7.16766417e-01 -1.61058533e+00 -3.02079499e-01 -4.52362269e-01
-8.03936303e-01 9.22691524e-01 6.48618639e-01 -6.42563924e-02
4.82047886e-01 -2.92551547e-01 -1.60195395e-01 -4.25211161e-01
-5.03509283e-01 -7.03156471e-01 7.20540643e-01 7.54000187e-01
-1.31079212e-01 2.17167974e-01 1.33004356e-02 7.77242780e-01
6.53039813e-02 -1.58384472e-01 1.44637637e-02 8.07043552e-01
-8.39437068e-01 -6.32830620e-01 -1.91930905e-01 3.29955518e-01
2.66940761e-02 -1.10301219e-01 -4.06738877e-01 6.36457622e-01
5.53575039e-01 8.80877852e-01 4.62457776e-01 -1.93147376e-01
3.45796496e-01 -8.45752191e-03 4.58860427e-01 -9.67651665e-01
-3.18646505e-02 -4.18090880e-01 -4.28222656e-01 -4.66686130e-01
-4.64262187e-01 -2.84587026e-01 -1.31015134e+00 2.69203991e-01
-2.67791420e-01 8.08645263e-02 6.80645943e-01 1.06288040e+00
2.73521006e-01 5.57249904e-01 3.81669551e-01 -5.02288699e-01
-6.59994721e-01 -6.81722581e-01 -5.51421463e-01 1.13374364e+00
2.98379838e-01 -1.08374131e+00 -4.53252852e-01 1.13647997e-01] | [9.950551986694336, 2.015659809112549] |
29d6c760-7cba-48e6-9896-7bd83dce922d | task-agnostic-structured-pruning-of-speech | 2306.01385 | null | https://arxiv.org/abs/2306.01385v2 | https://arxiv.org/pdf/2306.01385v2.pdf | Task-Agnostic Structured Pruning of Speech Representation Models | Self-supervised pre-trained models such as Wav2vec2, Hubert, and WavLM have been shown to significantly improve many speech tasks. However, their large memory and strong computational requirements hinder their industrial applicability. Structured pruning is a hardware-friendly model compression technique but usually results in a larger loss of accuracy. In this paper, we propose a fine-grained attention head pruning method to compensate for the performance degradation. In addition, we also introduce the straight through estimator into the L0 regularization to further accelerate the pruned model. Experiments on the SUPERB benchmark show that our model can achieve comparable performance to the dense model in multiple tasks and outperforms the Wav2vec 2.0 base model on average, with 72% fewer parameters and 2 times faster inference speed. | ['Yulong Wan', 'Hongbin Suo', 'Wei-Qiang Zhang', 'Siyuan Wang', 'Haoyu Wang'] | 2023-06-02 | null | null | null | null | ['model-compression'] | ['methodology'] | [-1.09050088e-01 -1.59150381e-02 -5.85337937e-01 -4.66249406e-01
-7.85743356e-01 1.52466312e-01 3.76324326e-01 1.31055312e-02
-6.71372056e-01 6.37624860e-01 4.57383722e-01 -6.99295640e-01
2.71536082e-01 -5.69606066e-01 -5.70993304e-01 -4.79394197e-01
3.49136114e-01 5.59298337e-01 3.67918968e-01 -7.32703879e-02
-8.41076300e-02 1.21791199e-01 -1.42774844e+00 1.46773085e-01
8.32458794e-01 1.01676643e+00 7.18406558e-01 6.42797649e-01
-2.45707184e-01 7.84759641e-01 -4.01891589e-01 -5.34684122e-01
-2.50187665e-02 -4.33228686e-02 -8.88354361e-01 -6.96717501e-02
6.32498860e-01 -4.68726128e-01 -6.20396852e-01 7.08138883e-01
5.46238005e-01 5.13358057e-01 3.21345687e-01 -7.38161087e-01
-3.05956811e-01 1.02758920e+00 -6.18815482e-01 3.25253040e-01
-3.38053435e-01 -1.46270514e-01 1.28216207e+00 -1.09845161e+00
2.37309471e-01 1.24777925e+00 6.37013674e-01 6.96158171e-01
-1.08670807e+00 -7.29279876e-01 1.92374989e-01 5.53147018e-01
-1.49921811e+00 -8.28263760e-01 4.29079980e-01 1.47673205e-01
1.67529905e+00 2.88508624e-01 3.39069605e-01 1.00871050e+00
-1.85552269e-01 1.04573727e+00 6.90144956e-01 -6.20544255e-01
1.55413345e-01 1.67760104e-01 5.67969322e-01 8.81297886e-01
3.89944464e-01 -1.37245864e-01 -5.53463280e-01 -1.43813029e-01
3.44337285e-01 -1.29786491e-01 -2.72566408e-01 1.21560827e-01
-3.93009424e-01 1.04003417e+00 3.15066546e-01 2.42803708e-01
-1.01906605e-01 4.09126043e-01 6.07185960e-01 1.90844357e-01
7.68765450e-01 3.22538733e-01 -7.42563188e-01 -4.35207099e-01
-1.20790267e+00 1.29277647e-01 7.20226586e-01 1.06499755e+00
6.46004081e-01 5.38818717e-01 -2.64396697e-01 1.38616133e+00
3.00195456e-01 3.36038679e-01 8.18825603e-01 -8.26146126e-01
6.08025849e-01 1.00500666e-01 -4.21492964e-01 -6.17688656e-01
-2.12411270e-01 -8.28116119e-01 -9.24317002e-01 -3.63606364e-01
-5.58607616e-02 -6.78566890e-03 -1.10464418e+00 1.52180207e+00
6.49555475e-02 3.52199137e-01 -1.54419258e-01 6.35081232e-01
8.32634032e-01 9.13290679e-01 1.96788415e-01 -2.20990002e-01
1.06094897e+00 -1.76757550e+00 -8.67328882e-01 -6.70723379e-01
1.04967308e+00 -6.79088056e-01 1.15896749e+00 3.76506835e-01
-1.29797435e+00 -6.97794139e-01 -1.05131185e+00 -5.03575504e-01
-9.79391038e-02 1.94729418e-01 8.37346077e-01 7.42380321e-01
-9.32587922e-01 6.17146373e-01 -1.12293243e+00 -1.90545261e-01
5.68478763e-01 2.78614908e-01 -5.19608147e-02 -2.19562560e-01
-9.35892820e-01 1.04190731e+00 4.78479385e-01 -4.36306983e-01
-7.40945458e-01 -8.62911344e-01 -9.56861973e-01 6.36257350e-01
2.45549604e-01 -3.88958126e-01 1.50812924e+00 -1.90445110e-01
-1.55183458e+00 4.07447100e-01 -7.56928980e-01 -1.18703043e+00
1.16285630e-01 -4.55965042e-01 -2.31270447e-01 -1.92853048e-01
-5.47974288e-01 8.14138591e-01 6.81288779e-01 -7.87710369e-01
-7.14409709e-01 -5.70265204e-02 -2.44569570e-01 1.15926258e-01
-9.98335779e-01 -1.33968547e-01 -1.02251232e+00 -6.44742012e-01
-1.45992711e-01 -6.40413940e-01 -4.37486082e-01 -3.60399604e-01
-1.44590169e-01 -4.20684069e-01 9.36843038e-01 -8.04634631e-01
1.84485316e+00 -2.11330318e+00 6.91454932e-02 -6.33151876e-03
3.38731974e-01 7.88340271e-01 -2.71061450e-01 6.67009726e-02
4.23845887e-01 1.65473580e-01 -1.23735145e-01 -1.00398421e+00
-2.72528410e-01 6.43299580e-01 -3.49536806e-01 1.25466332e-01
-1.74132079e-01 8.74765933e-01 -6.91852868e-01 -5.95521033e-01
2.27406994e-01 7.24636197e-01 -9.61073816e-01 1.40334651e-01
-1.15190826e-01 -1.88645929e-01 -5.47067262e-02 3.13733727e-01
6.55776441e-01 -1.67928010e-01 1.11523785e-01 1.04449969e-02
1.38314903e-01 9.52743411e-01 -5.92610836e-01 1.58085716e+00
-1.00444818e+00 8.80644619e-01 1.21804893e-01 -1.00368392e+00
8.49230468e-01 2.65754759e-01 -3.65702510e-02 -6.92722023e-01
3.12113345e-01 1.33202419e-01 1.14764608e-01 -3.21930945e-02
6.30800664e-01 2.92458266e-01 5.06284654e-01 8.84852633e-02
3.85276377e-01 -2.81196922e-01 2.99210191e-01 2.18950540e-01
1.04420125e+00 -2.57800043e-01 2.42975980e-01 -1.49085760e-01
4.18123335e-01 -1.28868312e-01 5.97917080e-01 7.73010850e-01
-2.89996006e-02 5.33569038e-01 -6.41788775e-03 -2.95026720e-01
-1.06346881e+00 -6.11923516e-01 -1.96912393e-01 1.28929532e+00
-3.15210760e-01 -1.00365150e+00 -7.63048828e-01 -5.38598537e-01
-7.94490427e-02 1.17987311e+00 -1.78342685e-01 -2.31072947e-01
-7.35056102e-01 -7.03754127e-01 5.63037753e-01 7.79557526e-01
5.54683089e-01 -7.52103329e-01 -3.44972730e-01 3.23306292e-01
-2.96945184e-01 -1.36101532e+00 -4.83901232e-01 4.96403843e-01
-1.27673244e+00 -4.45530236e-01 -4.66995090e-01 -8.59498024e-01
4.85005826e-01 5.89906275e-01 1.24059808e+00 3.19766283e-01
-1.60986036e-02 -1.97764590e-01 -4.96568948e-01 -4.38312650e-01
-2.70201653e-01 5.10040224e-01 1.68290779e-01 -4.33008999e-01
4.38432813e-01 -4.05201405e-01 -2.36181557e-01 7.79458657e-02
-5.16622245e-01 1.51295483e-01 4.00986582e-01 1.30788517e+00
7.77935505e-01 8.78256410e-02 1.96100414e-01 -1.03036392e+00
5.55382669e-01 -2.42003590e-01 -5.26501417e-01 1.49411693e-01
-1.10186887e+00 1.69024974e-01 6.21075034e-01 -3.64359438e-01
-1.17370868e+00 -8.63859504e-02 -6.12939477e-01 -6.57877505e-01
1.53212145e-01 5.49953282e-01 3.10861804e-02 7.61333620e-03
4.49690402e-01 2.06466839e-01 -1.36339366e-01 -9.57568526e-01
4.31838423e-01 8.34219992e-01 3.40604842e-01 -3.04297060e-01
6.20626509e-01 8.94645154e-02 -3.97837192e-01 -1.18301439e+00
-1.01060593e+00 -6.57312810e-01 -1.89791858e-01 3.93257886e-01
5.48263371e-01 -1.05262458e+00 -4.71042171e-02 5.63450158e-02
-1.29045630e+00 -6.68059707e-01 -3.50271106e-01 6.93225563e-01
-2.54687399e-01 4.63525862e-01 -9.29385006e-01 -7.44147718e-01
-8.80746901e-01 -1.07807648e+00 6.99749231e-01 -1.46459803e-01
-2.38918185e-01 -9.02974844e-01 -6.93791658e-02 5.49601853e-01
7.46191382e-01 -8.58627021e-01 1.02002370e+00 -7.70093322e-01
-1.27090365e-01 -3.09997648e-01 -2.10202113e-01 7.59429574e-01
-1.13962136e-01 -2.12166518e-01 -1.20356286e+00 -4.25146729e-01
-4.47648428e-02 -5.01579225e-01 1.38150573e+00 5.18657446e-01
1.63657236e+00 -4.53761786e-01 -3.75509828e-01 9.36020195e-01
1.09174752e+00 5.63003458e-02 5.54838836e-01 -1.41861094e-02
7.53102183e-01 8.83819684e-02 3.77122641e-01 3.27380210e-01
9.28206369e-02 7.05285788e-01 1.99069053e-01 -7.39138797e-02
-6.98216975e-01 -3.36098433e-01 3.67385030e-01 1.80329525e+00
-1.06554158e-01 -4.34586138e-01 -6.09404743e-01 5.55003285e-01
-1.60872138e+00 -7.69307017e-01 -2.04323128e-01 2.19157267e+00
1.04522729e+00 4.21115667e-01 -3.45188200e-01 3.17536622e-01
2.97499418e-01 3.63038331e-01 -1.47658989e-01 -8.07086051e-01
-9.12486017e-02 7.41863847e-01 7.65444994e-01 8.69346023e-01
-9.53820467e-01 1.42595673e+00 6.67514801e+00 1.51515090e+00
-8.95316720e-01 6.49598897e-01 7.11613357e-01 -4.67439443e-01
-1.63023934e-01 -1.73283100e-01 -1.37254310e+00 2.31760442e-01
1.27430940e+00 -9.58991125e-02 4.52605247e-01 1.14772606e+00
1.48552343e-01 2.69971192e-01 -7.50648379e-01 1.07407832e+00
1.36113420e-01 -1.40205979e+00 1.57988966e-01 1.05230607e-01
7.01638281e-01 5.54414511e-01 7.76088005e-03 6.79643989e-01
3.96500736e-01 -9.62185681e-01 3.64518404e-01 -1.64709762e-01
7.94631600e-01 -7.97844946e-01 7.06654310e-01 3.67072731e-01
-1.07683265e+00 3.87677476e-02 -1.02526236e+00 -1.58385783e-01
2.97792196e-01 6.01527035e-01 -9.04319704e-01 1.77497417e-02
6.74336910e-01 3.96110892e-01 -4.94291544e-01 1.07041812e+00
-3.02789658e-01 1.41509211e+00 -4.45193321e-01 -1.29371844e-02
4.58970100e-01 8.76306519e-02 3.01835269e-01 1.56255078e+00
2.86256164e-01 -7.36526027e-02 -1.50615141e-01 3.59874308e-01
-3.60511273e-01 2.06809565e-01 -3.22822243e-01 2.80961152e-02
6.18570268e-01 1.07184935e+00 -1.74045563e-01 -5.99917412e-01
-7.48163462e-01 8.96177888e-01 7.54293442e-01 2.46828347e-01
-9.24162328e-01 -2.34741360e-01 8.25831413e-01 1.51843071e-01
7.12678552e-01 -3.42132568e-01 -4.13310707e-01 -1.08768344e+00
-3.58971208e-01 -7.73880303e-01 1.58377826e-01 -4.97322947e-01
-8.11755240e-01 8.74074638e-01 -2.23623067e-01 -6.28979623e-01
-2.00318158e-01 -5.82208931e-01 -5.12167156e-01 7.34877408e-01
-1.63579881e+00 -8.01585197e-01 -1.65418789e-01 2.83843100e-01
1.15473890e+00 -3.06357056e-01 1.11909831e+00 5.41147053e-01
-8.19998980e-01 1.15351439e+00 2.72303015e-01 -3.31759825e-02
4.82064575e-01 -8.76992106e-01 7.58778334e-01 7.73663402e-01
5.41037261e-01 4.67763305e-01 5.36921799e-01 -4.65062857e-01
-1.18644178e+00 -1.36786437e+00 1.27281535e+00 3.14352848e-02
4.59661722e-01 -2.88469285e-01 -1.13547456e+00 7.38932967e-01
3.68151873e-01 -1.56194940e-01 5.67872047e-01 4.77578402e-01
-6.71127200e-01 -9.55066681e-02 -8.48466575e-01 6.52174950e-01
1.11805248e+00 -2.88286686e-01 -5.39946377e-01 4.59163696e-01
1.14709377e+00 -4.17613149e-01 -5.56478560e-01 3.77660245e-01
3.06632757e-01 -5.87676704e-01 1.02167642e+00 -6.52911186e-01
3.84447426e-01 3.85057777e-01 -3.88142735e-01 -1.37108445e+00
-6.67424262e-01 -4.92405713e-01 -9.07438934e-01 1.15125608e+00
6.50944948e-01 -4.10644382e-01 9.20990944e-01 4.58327293e-01
-3.66146594e-01 -1.02640069e+00 -1.01242208e+00 -1.05577981e+00
1.03364296e-01 -8.47427666e-01 4.33977813e-01 5.60151935e-01
-1.86063927e-02 5.63730180e-01 -5.62327147e-01 -2.44411498e-01
4.95906085e-01 -3.42093527e-01 6.70144379e-01 -9.69570100e-01
-5.11636794e-01 -6.15455151e-01 -2.00582191e-01 -1.67310786e+00
2.45729476e-01 -9.79103446e-01 -1.49883274e-02 -1.48817837e+00
1.89170867e-01 -5.68686903e-01 -4.43854481e-01 6.11214280e-01
-2.12828472e-01 2.88853288e-01 8.24834928e-02 6.25566393e-02
-4.08739924e-01 6.35249257e-01 8.78719807e-01 -4.28311259e-01
-2.19700620e-01 -8.49946439e-02 -4.82478738e-01 7.88687825e-01
9.96597588e-01 -4.71759349e-01 -5.34211099e-01 -1.17611897e+00
-3.63098562e-01 -4.03562039e-01 -2.82523543e-01 -1.07134283e+00
3.26520979e-01 2.41824344e-01 5.31295203e-02 -6.33263111e-01
7.33092427e-01 -5.63248038e-01 -5.91502190e-01 4.54833031e-01
-3.65638226e-01 -1.81080833e-01 3.55511814e-01 3.87032658e-01
-3.98735523e-01 -5.42160928e-01 9.17256594e-01 1.16441764e-01
-5.68967760e-01 3.38121593e-01 -3.60601366e-01 1.22421265e-01
3.99893165e-01 1.14291765e-01 -2.77117416e-02 -4.97780651e-01
-3.26491207e-01 2.18380406e-01 -8.61581042e-02 3.52131009e-01
7.54139185e-01 -1.11141622e+00 -7.31230199e-01 3.32349092e-01
-1.27115041e-01 -5.09056821e-02 1.79609239e-01 7.54055977e-01
-3.65037918e-01 7.77977943e-01 3.97647500e-01 -3.24861288e-01
-1.55851948e+00 4.23386335e-01 5.26279546e-02 -5.39251029e-01
-7.40683675e-01 1.36287177e+00 3.11999433e-02 -1.27855852e-01
7.49426961e-01 -4.82088029e-01 -6.00860082e-02 -1.61859185e-01
7.30238080e-01 7.67347455e-01 3.92604172e-01 -5.07509649e-01
-1.44968614e-01 1.70264602e-01 -4.66593385e-01 -1.51819382e-02
1.38170469e+00 -4.67464849e-02 1.42378286e-01 6.80196136e-02
1.41537499e+00 1.15446160e-02 -9.30693924e-01 -5.89442730e-01
-7.87994638e-02 -4.45978433e-01 7.51231670e-01 -4.60331798e-01
-1.49005890e+00 1.28336215e+00 3.10710281e-01 -2.05854625e-02
9.09320891e-01 -9.18611810e-02 1.44002450e+00 6.04221106e-01
2.96855301e-01 -1.18578112e+00 -4.38994497e-01 8.54365587e-01
6.33565605e-01 -1.09605944e+00 1.68124631e-01 -4.47443515e-01
-6.10568762e-01 7.24946141e-01 7.60340393e-01 1.96706671e-02
7.57519484e-01 5.38203478e-01 -2.12998524e-01 2.08333030e-01
-1.17302275e+00 -2.67788619e-01 3.24561328e-01 4.08444047e-01
5.69039583e-01 1.40399903e-01 -5.21228790e-01 7.29852200e-01
-3.80268961e-01 -1.89304709e-01 1.12698250e-01 5.80937028e-01
-7.02280939e-01 -1.12952793e+00 8.03858265e-02 7.57426381e-01
-5.58327436e-01 -7.04973698e-01 2.04731524e-02 4.60394531e-01
-1.41670942e-01 1.17434072e+00 2.97796637e-01 -6.37528300e-01
1.92111641e-01 2.76734382e-01 4.67095286e-01 -6.07137084e-01
-6.14035189e-01 2.97708929e-01 3.22467268e-01 -5.70886910e-01
1.09412178e-01 -1.94065303e-01 -1.10845971e+00 -6.03377879e-01
-7.44142234e-01 2.60486305e-01 6.94538593e-01 8.08728993e-01
5.32738566e-01 5.73629200e-01 2.53153652e-01 -5.71236670e-01
-8.31839263e-01 -1.39874339e+00 -3.24493796e-01 1.36484176e-01
8.04122090e-02 -5.36669016e-01 -5.05852818e-01 -1.18825853e-01] | [8.867088317871094, 3.608731508255005] |
13b6362b-0c25-4079-8fda-15f38a14f83c | alibaba-translate-china-s-submission-for-wmt-1 | 2210.10049 | null | https://arxiv.org/abs/2210.10049v2 | https://arxiv.org/pdf/2210.10049v2.pdf | Alibaba-Translate China's Submission for WMT 2022 Quality Estimation Shared Task | In this paper, we present our submission to the sentence-level MQM benchmark at Quality Estimation Shared Task, named UniTE (Unified Translation Evaluation). Specifically, our systems employ the framework of UniTE, which combined three types of input formats during training with a pre-trained language model. First, we apply the pseudo-labeled data examples for the continuously pre-training phase. Notably, to reduce the gap between pre-training and fine-tuning, we use data pruning and a ranking-based score normalization strategy. For the fine-tuning phase, we use both Direct Assessment (DA) and Multidimensional Quality Metrics (MQM) data from past years' WMT competitions. Finally, we collect the source-only evaluation results, and ensemble the predictions generated by two UniTE models, whose backbones are XLM-R and InfoXLM, respectively. Results show that our models reach 1st overall ranking in the Multilingual and English-Russian settings, and 2nd overall ranking in English-German and Chinese-English settings, showing relatively strong performances in this year's quality estimation competition. | ['Jun Xie', 'Derek F. Wong', 'Xiangnan He', 'Wenqiang Lei', 'Baosong Yang', 'Dayiheng Liu', 'Yu Wan', 'Keqin Bao'] | 2022-10-18 | null | null | null | null | ['xlm-r'] | ['natural-language-processing'] | [ 9.77120176e-03 -2.99930602e-01 -2.38340795e-01 -6.97347879e-01
-1.78602970e+00 -6.88411295e-01 4.55932021e-01 2.42145315e-01
-8.02111208e-01 1.03556454e+00 5.18587649e-01 -5.38515389e-01
-4.23500165e-02 -4.70711410e-01 -6.50606096e-01 -5.52857071e-02
3.41743946e-01 7.43957818e-01 -1.59451142e-01 -7.46750474e-01
3.39226097e-01 5.66384047e-02 -9.73014474e-01 6.98627055e-01
1.44073617e+00 8.44621480e-01 -3.93730029e-02 8.78963768e-01
2.93506514e-02 3.35865736e-01 -7.69531548e-01 -1.37115777e+00
2.18593106e-01 -2.81854033e-01 -9.40436721e-01 -3.39074880e-01
4.56098884e-01 -1.71290681e-01 -4.70343344e-02 8.47865343e-01
7.74701416e-01 7.86411762e-02 5.93123198e-01 -1.00621033e+00
-8.45473170e-01 8.72556865e-01 -1.24116473e-01 1.26703620e-01
3.87337059e-01 4.71750975e-01 1.35565650e+00 -1.35915625e+00
7.16618836e-01 1.16894042e+00 4.86175209e-01 5.33101976e-01
-1.21312749e+00 -4.35480297e-01 -4.10391614e-02 2.66502827e-01
-9.32845414e-01 -7.73020804e-01 1.96543097e-01 -2.18751624e-01
1.28262448e+00 2.25151390e-01 3.04467529e-01 1.21163023e+00
4.91290510e-01 9.04111087e-01 1.31946909e+00 -4.77403402e-01
6.21646941e-02 1.97029829e-01 1.49230376e-01 5.14330626e-01
-7.64862075e-02 2.31106147e-01 -7.49191463e-01 -7.92502761e-02
2.14067891e-01 -6.08084440e-01 -2.25301251e-01 8.88887420e-02
-1.66749191e+00 6.26181662e-01 1.29005790e-01 1.77132159e-01
-3.10276270e-01 -3.37352663e-01 4.12888408e-01 1.01035404e+00
6.43612385e-01 7.63599217e-01 -9.26979661e-01 -8.01422238e-01
-1.19001377e+00 2.42357716e-01 6.88355148e-01 1.00402462e+00
4.25852656e-01 -4.97004271e-01 -6.57196581e-01 1.19966352e+00
1.46183476e-01 7.70037353e-01 5.96990943e-01 -8.63844097e-01
1.39514029e+00 4.21481639e-01 2.51400262e-01 -2.95015156e-01
-3.37590933e-01 -7.94321895e-01 -8.28766942e-01 -6.18713982e-02
4.81246889e-01 -1.09810144e-01 -8.66341650e-01 1.90248632e+00
-2.35823691e-01 -6.49857640e-01 3.59182060e-01 8.53660107e-01
7.54705429e-01 5.43040991e-01 -5.19473292e-02 -3.62978667e-01
1.16172779e+00 -1.43174434e+00 -8.00968826e-01 -9.56021249e-02
9.59613621e-01 -1.03576982e+00 1.66949224e+00 6.75259352e-01
-1.56197393e+00 -8.04338694e-01 -1.04217029e+00 -2.79772609e-01
-3.88280928e-01 4.02151823e-01 9.85945985e-02 4.80660290e-01
-1.22434795e+00 9.61443186e-01 -6.95995629e-01 -8.75036120e-02
9.02302191e-02 2.22159609e-01 -5.18624246e-01 -2.52843946e-01
-1.52588677e+00 1.25954902e+00 8.99293423e-02 6.17199615e-02
-7.54355907e-01 -6.05546534e-01 -6.55994773e-01 -2.32163087e-01
1.11962043e-01 -8.34924638e-01 1.41317809e+00 -6.03244960e-01
-1.76018715e+00 8.23663473e-01 -1.63001880e-01 -1.22138858e-01
5.84247112e-01 -5.12364805e-01 -6.83128715e-01 -4.01092052e-01
2.54531384e-01 5.14417410e-01 2.67943054e-01 -7.09192753e-01
-5.88303566e-01 -2.33859107e-01 5.11527061e-02 4.89632934e-01
-1.24988660e-01 2.81447142e-01 -7.17420518e-01 -6.51736081e-01
5.16239554e-02 -6.86804175e-01 -2.24814385e-01 -6.18902624e-01
-4.05146062e-01 -1.95467666e-01 -2.99346328e-01 -1.16010201e+00
1.74625337e+00 -1.89681542e+00 7.88031340e-01 9.78522152e-02
-2.69257277e-02 1.04008630e-01 -8.11639845e-01 4.86978054e-01
1.74910024e-01 2.54512072e-01 -2.70879269e-01 -9.12140369e-01
3.04169625e-01 -6.45631168e-04 -3.62555906e-02 -5.47597359e-04
6.18237853e-01 1.13855398e+00 -9.44890678e-01 -3.80908191e-01
-2.29996741e-01 -1.19161814e-01 -7.55201340e-01 3.19778770e-01
-1.95504501e-01 4.20751095e-01 8.89616907e-02 7.02284038e-01
5.88574052e-01 -1.22931346e-01 9.68377218e-02 -1.89595416e-01
-1.06230251e-01 1.14822686e+00 -9.69717383e-01 2.28044200e+00
-8.26762855e-01 2.61045605e-01 -1.68225572e-01 -2.25889668e-01
7.02312529e-01 1.38061672e-01 6.41073808e-02 -1.18514907e+00
-2.85374194e-01 6.76043749e-01 1.62522033e-01 -2.11621165e-01
9.11928117e-01 1.02676630e-01 -2.80187905e-01 5.71402073e-01
5.07783890e-01 -2.05359668e-01 6.62363231e-01 3.45455140e-01
1.07559681e+00 4.63392615e-01 7.75360018e-02 -1.53389320e-01
4.84039366e-01 1.21690882e-02 7.25887954e-01 5.21280348e-01
-2.34805256e-01 7.14810014e-01 3.92568797e-01 -1.43487900e-01
-1.16133535e+00 -1.22374344e+00 -1.75498724e-01 1.34872460e+00
-1.97040275e-01 -1.07092822e+00 -6.60551906e-01 -9.29857671e-01
-2.93382883e-01 6.95826769e-01 -3.45928401e-01 -1.99470118e-01
-6.25233889e-01 -9.63565588e-01 6.13741755e-01 3.60884935e-01
3.43620628e-01 -9.86861646e-01 4.14622575e-02 5.01023293e-01
-7.80431569e-01 -9.57352638e-01 -5.53984761e-01 2.24250749e-01
-9.17367876e-01 -6.25124395e-01 -7.46841550e-01 -3.88720721e-01
3.43754031e-02 -2.43277326e-01 1.91883063e+00 -1.20192461e-01
4.22559083e-01 -1.94436818e-01 -4.94329542e-01 -4.10444140e-02
-5.84876537e-01 6.04977489e-01 2.93009400e-01 -4.82632518e-01
2.86820054e-01 -3.08666587e-01 -4.09861684e-01 5.22540212e-01
-4.55382109e-01 3.12671810e-01 9.03398812e-01 1.02743423e+00
6.27164900e-01 -5.37796855e-01 6.77520156e-01 -5.24065197e-01
9.54451501e-01 -2.74307489e-01 -4.08440113e-01 6.50773764e-01
-8.85082543e-01 2.21870691e-01 5.58633804e-01 -2.28742119e-02
-6.84169471e-01 -6.54534459e-01 -4.64240551e-01 1.91149190e-01
2.15382725e-01 7.50237226e-01 -5.93824685e-01 4.34364021e-01
6.37537479e-01 6.20016679e-02 -3.66275579e-01 -8.48291874e-01
5.72115541e-01 9.23241496e-01 4.69098061e-01 -8.57885063e-01
6.94571614e-01 -4.61576343e-01 -3.84144008e-01 6.48736730e-02
-8.69631290e-01 -1.96152762e-01 -7.87900209e-01 1.69323906e-01
7.63313472e-01 -1.05195606e+00 -2.08157346e-01 3.92232209e-01
-1.22835732e+00 -3.19697082e-01 -1.41309530e-01 6.53201461e-01
-6.31685078e-01 1.91234380e-01 -9.50570524e-01 -4.08507854e-01
-6.98927343e-01 -1.56102693e+00 1.23884523e+00 -2.37021491e-01
-8.30394104e-02 -8.40920985e-01 4.73426819e-01 6.67427957e-01
6.78134322e-01 -2.23825231e-01 1.01139033e+00 -5.51416218e-01
-2.96252996e-01 1.88006133e-01 -1.76451921e-01 6.97911203e-01
-9.81104523e-02 -3.49792205e-02 -7.22195506e-01 -4.00344521e-01
-3.14625084e-01 -7.19099402e-01 7.55819798e-01 -3.05493679e-02
5.49374580e-01 -8.10328275e-02 3.67433876e-01 7.05197275e-01
1.16091084e+00 -3.02834570e-01 5.95737278e-01 6.02255583e-01
3.76054615e-01 4.62993085e-01 9.40189838e-01 1.19203165e-01
6.94689393e-01 1.07848203e+00 3.69635187e-02 7.38634914e-02
-1.97982788e-01 -2.28487015e-01 8.87487113e-01 1.65380657e+00
-2.07393944e-01 -2.48925105e-01 -7.81803966e-01 2.24103630e-01
-1.68795288e+00 -6.27483010e-01 2.61701532e-02 2.38211322e+00
1.25288081e+00 4.35922682e-01 2.07504347e-01 -7.87643641e-02
2.86719412e-01 -2.75336415e-01 -3.14764977e-01 -7.44557083e-01
-4.66012359e-01 4.71920520e-01 1.85638279e-01 7.01358318e-01
-8.74909520e-01 1.13922393e+00 6.28626347e+00 9.70844388e-01
-6.51318967e-01 3.71010423e-01 6.35112286e-01 -3.65247488e-01
-5.01940012e-01 -1.55559614e-01 -7.84501255e-01 4.25162941e-01
1.70968390e+00 -1.35101274e-01 6.32028162e-01 3.31690490e-01
2.33537599e-01 2.17657372e-01 -1.25303996e+00 8.48434627e-01
-3.08013916e-01 -8.99358690e-01 1.95886865e-01 1.68399602e-01
7.50494301e-01 5.31656742e-01 1.07608870e-01 9.52691436e-01
3.75046998e-01 -1.10828698e+00 7.64362156e-01 6.57094657e-01
1.14546204e+00 -8.48163426e-01 1.04999578e+00 4.53717113e-01
-8.25614393e-01 9.11617056e-02 -6.14897132e-01 8.62526745e-02
2.73285747e-01 8.41112316e-01 -3.57149094e-01 1.17686260e+00
4.30303872e-01 7.43559182e-01 -9.13976490e-01 8.46212864e-01
-3.56351703e-01 6.23336256e-01 -2.23659817e-02 3.11601013e-01
1.60874650e-01 -4.71652240e-01 4.45581019e-01 1.33504975e+00
3.69197160e-01 -3.33119482e-01 -1.03975303e-01 6.48278296e-01
-5.84329545e-01 4.66850221e-01 -3.18979211e-02 6.25420967e-03
3.39708239e-01 1.23326588e+00 1.92791328e-01 -5.53975940e-01
-3.42384160e-01 1.23866987e+00 7.31490076e-01 1.36396617e-01
-8.18665206e-01 -4.28787231e-01 7.17869222e-01 -1.98884755e-01
-1.76435113e-01 -1.46641329e-01 -3.97176415e-01 -1.69664073e+00
3.10061663e-01 -1.35843408e+00 1.21882506e-01 -6.06570542e-01
-1.30976367e+00 9.49787378e-01 -3.09102029e-01 -1.46854365e+00
-4.79589492e-01 -6.51422262e-01 -2.17901245e-01 1.41616070e+00
-1.73165047e+00 -8.91117930e-01 2.89490104e-01 2.27287367e-01
5.88080764e-01 -3.35250169e-01 1.03045714e+00 7.17366576e-01
-5.45892477e-01 1.19982004e+00 4.62072313e-01 1.45129472e-01
1.37192857e+00 -1.53250539e+00 8.80804121e-01 8.41290355e-01
2.76618987e-01 6.64267480e-01 3.78389776e-01 -5.82757056e-01
-1.30028582e+00 -1.06036794e+00 1.60720026e+00 -1.06112993e+00
5.41803777e-01 -4.59573388e-01 -7.88115919e-01 2.22885370e-01
2.72982568e-01 -3.31154644e-01 7.53291249e-01 6.01144195e-01
-3.90762478e-01 3.65646482e-02 -1.00311673e+00 3.94486070e-01
1.13179064e+00 -5.41918695e-01 -7.32327878e-01 3.78816068e-01
9.91182387e-01 -6.55131936e-01 -1.32863140e+00 5.29461861e-01
5.66461205e-01 -7.49071240e-01 5.71811438e-01 -1.18286335e+00
9.50627148e-01 -1.07127115e-01 -6.44852817e-01 -1.95742154e+00
-5.08751035e-01 -5.88228762e-01 -4.43550237e-02 1.24143779e+00
1.14888942e+00 -2.42658362e-01 2.56634176e-01 2.73427486e-01
-4.79067892e-01 -1.04899025e+00 -9.79461193e-01 -8.39002907e-01
5.50952137e-01 -4.46017027e-01 7.85636187e-01 7.19770610e-01
-1.46155562e-02 7.08137035e-01 -4.21567768e-01 -1.98315531e-01
5.44466436e-01 -8.16448703e-02 8.18015873e-01 -9.02772605e-01
-7.21860945e-01 -6.42043233e-01 -3.75013538e-02 -1.02510154e+00
-2.08639219e-01 -1.10885322e+00 -1.70323357e-01 -1.68385279e+00
2.84424692e-01 -3.14381927e-01 -6.55013621e-01 2.71814257e-01
-6.19672775e-01 2.70480782e-01 3.75204742e-01 3.69281471e-01
-9.56007302e-01 6.90873265e-01 1.40591538e+00 -1.28468454e-01
-2.20632061e-01 1.03179976e-01 -8.09573412e-01 1.11258209e-01
6.35938704e-01 -3.22822720e-01 -6.57747313e-02 -1.06612360e+00
6.87414467e-01 1.32014409e-01 -2.44870305e-01 -8.57666731e-01
-2.09856898e-01 -1.34436846e-01 2.71533698e-01 -5.78903496e-01
2.00549498e-01 -2.98792332e-01 -1.82049334e-01 2.16865689e-01
-6.00279689e-01 7.84267008e-01 1.08421542e-01 8.06800846e-04
-3.50020766e-01 8.52051973e-02 6.71194673e-01 1.18898712e-01
-1.45490423e-01 2.46561199e-01 1.21540584e-01 3.75528216e-01
1.35986149e-01 2.82249302e-01 -6.02454424e-01 -1.96807683e-01
-8.59925687e-01 4.06821877e-01 4.81869429e-01 6.47115648e-01
4.65594053e-01 -1.69809389e+00 -1.41055799e+00 3.80177051e-02
5.34484446e-01 -3.55554134e-01 -2.20688935e-02 1.04429460e+00
-8.77371356e-02 5.76326966e-01 -2.58840322e-01 -3.99589896e-01
-8.49540889e-01 1.57456160e-01 1.70046508e-01 -1.02748561e+00
-1.48391485e-01 6.56819701e-01 -4.75960106e-01 -1.05088460e+00
2.01279484e-02 -4.54681575e-01 -2.32769139e-02 -1.86513737e-01
6.69341385e-01 4.16006923e-01 7.73746908e-01 -6.18413389e-01
-2.99502671e-01 3.52474719e-01 -2.74304301e-01 -6.75090849e-01
1.25344408e+00 -2.13993609e-01 -1.22697122e-01 5.75672865e-01
1.17717671e+00 3.10376704e-01 -6.99500501e-01 -5.26336551e-01
2.91612595e-01 -2.97305971e-01 7.56939733e-03 -1.72925484e+00
-6.64370894e-01 1.06481802e+00 5.09892285e-01 -4.50017869e-01
1.16955256e+00 -3.37059557e-01 8.14657927e-01 5.07177413e-01
5.03261089e-01 -1.36265171e+00 -1.42336562e-01 9.52931523e-01
1.02405131e+00 -1.36484313e+00 -2.78848767e-01 4.19219099e-02
-7.34100044e-01 9.60995972e-01 5.91607749e-01 4.29537535e-01
1.31997913e-01 -6.68278337e-02 2.43319318e-01 3.04566920e-01
-1.40252244e+00 -6.89489543e-02 8.08106303e-01 3.31468433e-01
9.22134578e-01 3.21987182e-01 -6.03743434e-01 1.04719985e+00
-6.27118111e-01 1.36492420e-02 3.07445109e-01 6.21363640e-01
-3.33977550e-01 -1.65202403e+00 -1.35481507e-01 4.54242796e-01
-5.26085734e-01 -7.01401830e-01 -3.23726416e-01 4.93704438e-01
-6.94470704e-02 1.16381955e+00 -3.37887466e-01 -8.46084952e-01
8.41709316e-01 2.12165982e-01 6.60970867e-01 -6.10877931e-01
-1.01242626e+00 -1.12119183e-01 4.05242383e-01 -8.41393352e-01
-6.67644590e-02 -4.42067713e-01 -7.50010788e-01 -3.96097511e-01
-2.54620194e-01 4.69206482e-01 9.51856494e-01 9.19192314e-01
4.67614979e-01 5.09456992e-01 6.72330618e-01 -4.24960583e-01
-1.08381760e+00 -1.58009648e+00 -5.97565733e-02 5.37588716e-01
1.71126917e-01 -4.16659832e-01 -2.50081182e-01 -4.10213023e-01] | [11.644231796264648, 10.304625511169434] |
0a942087-8285-4efb-8923-c8f7d730ef89 | dataset-bias-in-the-natural-sciences-a-case | 2105.02637 | null | https://arxiv.org/abs/2105.02637v1 | https://arxiv.org/pdf/2105.02637v1.pdf | Dataset Bias in the Natural Sciences: A Case Study in Chemical Reaction Prediction and Synthesis Design | Datasets in the Natural Sciences are often curated with the goal of aiding scientific understanding and hence may not always be in a form that facilitates the application of machine learning. In this paper, we identify three trends within the fields of chemical reaction prediction and synthesis design that require a change in direction. First, the manner in which reaction datasets are split into reactants and reagents encourages testing models in an unrealistically generous manner. Second, we highlight the prevalence of mislabelled data, and suggest that the focus should be on outlier removal rather than data fitting only. Lastly, we discuss the problem of reagent prediction, in addition to reactant prediction, in order to solve the full synthesis design problem, highlighting the mismatch between what machine learning solves and what a lab chemist would need. Our critiques are also relevant to the burgeoning field of using machine learning to accelerate progress in experimental Natural Sciences, where datasets are often split in a biased way, are highly noisy, and contextual variables that are not evident from the data strongly influence the outcome of experiments. | ['Alpha A. Lee', 'Philippe Schwaller', 'Ryan-Rhys Griffiths'] | 2021-05-06 | null | null | null | null | ['chemical-reaction-prediction', 'reagent-prediction'] | ['medical', 'medical'] | [ 6.48520112e-01 -1.69896930e-01 -2.15445578e-01 -2.45358154e-01
-3.90751570e-01 -7.34407842e-01 6.49157405e-01 7.14951575e-01
-5.22802949e-01 9.19432998e-01 2.54880428e-01 -8.45206559e-01
-2.91713700e-03 -5.78839958e-01 -7.67011881e-01 -9.27548230e-01
5.50304413e-01 2.51308471e-01 -2.35470936e-01 -5.38466461e-02
8.50664258e-01 8.73347640e-01 -1.39552915e+00 1.33140936e-01
7.53628016e-01 4.01050955e-01 -7.96341896e-02 5.46488643e-01
-2.32127905e-01 9.81505513e-01 -6.40148818e-01 -2.16751069e-01
6.93265796e-02 -6.00311041e-01 -6.60463989e-01 -1.26312703e-01
8.90184045e-02 2.97708333e-01 9.37033445e-02 7.01119840e-01
5.25057793e-01 1.13291599e-01 8.67012024e-01 -8.93886924e-01
-4.35202897e-01 2.40688518e-01 -2.80125648e-01 8.31636190e-02
2.39609733e-01 7.24664092e-01 8.89645696e-01 -9.01310682e-01
6.81461811e-01 1.00232947e+00 3.72479737e-01 2.69628048e-01
-1.49235046e+00 -6.47245347e-01 3.34954023e-01 -1.70761421e-01
-1.04858315e+00 -7.25183427e-01 5.37763417e-01 -9.76743877e-01
9.70441759e-01 4.71158653e-01 6.15570724e-01 1.05524695e+00
2.95199662e-01 1.25922576e-01 9.16361094e-01 -2.82005697e-01
5.73732972e-01 6.09599762e-02 -3.25010478e-01 2.56037503e-01
5.53750575e-01 3.04926813e-01 -3.58009189e-01 -2.42275491e-01
3.97383600e-01 2.66291589e-01 -1.73750967e-01 -4.11571860e-01
-1.25605905e+00 1.08188665e+00 2.65013520e-02 3.73451293e-01
-1.59616753e-01 -3.21648508e-01 5.24542034e-01 2.05917329e-01
2.28267521e-01 1.41350722e+00 -8.35498750e-01 -1.96708217e-01
-9.79750514e-01 3.75344992e-01 8.91010761e-01 5.88304698e-01
6.06283605e-01 1.71333551e-01 4.66674387e-01 4.87587959e-01
1.49624884e-01 -1.24154776e-01 4.01630193e-01 -7.73673952e-01
9.23881531e-02 4.71983939e-01 3.11328471e-01 -7.19840288e-01
-3.64712864e-01 -2.35983089e-01 -4.84097749e-01 3.60701799e-01
7.90209055e-01 -1.82553604e-01 -7.97640741e-01 1.44409013e+00
2.77150601e-01 -5.76252639e-01 1.41079590e-01 6.78264499e-01
5.82259834e-01 5.43942571e-01 5.78298450e-01 -6.34832561e-01
7.53682613e-01 -5.26159525e-01 -4.41048771e-01 -1.81958005e-01
1.17552376e+00 -1.07153177e+00 9.76626277e-01 7.48883784e-01
-1.04292738e+00 -1.21334180e-01 -1.07573867e+00 -2.65731990e-01
-8.85311246e-01 -3.23308617e-01 7.33338058e-01 3.64115089e-01
-2.35956371e-01 1.18966174e+00 -5.15017629e-01 -2.60777086e-01
4.56459284e-01 3.06715518e-01 -4.07181710e-01 1.10702477e-01
-6.42131805e-01 1.21531653e+00 4.44312990e-01 1.42995819e-01
-4.96852160e-01 -1.05631471e+00 -6.48440659e-01 -1.45906940e-01
5.71721077e-01 -2.46893644e-01 9.90743935e-01 -6.89796209e-01
-1.15800571e+00 6.80941939e-01 -1.91365585e-01 6.29132390e-02
5.88976920e-01 1.91949591e-01 -4.15698469e-01 -4.93537217e-01
1.29445359e-01 3.12881708e-01 2.03134269e-01 -1.22605181e+00
-7.19889522e-01 -4.62645203e-01 -4.91978049e-01 -6.77960962e-02
4.00824308e-01 7.82246962e-02 2.58869261e-01 -5.88883579e-01
2.62858152e-01 -6.96722806e-01 -6.09595418e-01 -1.93804219e-01
-2.55468011e-01 -1.85781330e-01 6.60072207e-01 -2.41311133e-01
1.09004295e+00 -1.91954970e+00 1.74289986e-01 3.32215011e-01
3.79961759e-01 1.59295159e-03 2.82967627e-01 7.37675071e-01
-6.03093863e-01 5.94340324e-01 -1.27206817e-01 2.97903270e-01
-1.51510671e-01 -3.14117409e-02 -1.79261118e-01 6.51719153e-01
2.34731525e-01 5.71327090e-01 -1.16500390e+00 -5.51657006e-02
1.89829409e-01 2.42135935e-02 -3.19473743e-01 2.44772322e-02
-4.18860584e-01 6.37668669e-01 -3.33847910e-01 7.29498386e-01
2.79328763e-01 -1.60371587e-01 3.33224058e-01 -9.13644806e-02
-9.22851443e-01 6.02884769e-01 -9.93204653e-01 1.37379718e+00
-1.38699859e-01 2.88560539e-01 -3.88798773e-01 -9.09367263e-01
9.46828187e-01 3.41942728e-01 7.20895350e-01 -4.98281091e-01
1.73058659e-01 4.76484269e-01 4.46731985e-01 -5.52350879e-01
3.90149087e-01 -4.05899793e-01 3.56179357e-01 4.84493732e-01
-3.04772437e-01 -4.48916316e-01 5.08215189e-01 -7.20399246e-02
9.05006111e-01 1.37039080e-01 3.05208355e-01 -3.91191721e-01
1.72182918e-01 2.44750381e-01 8.98819566e-01 5.86296320e-01
-1.37070805e-01 4.52934086e-01 8.85747731e-01 -5.15977740e-01
-1.55297828e+00 -4.35913384e-01 -2.79015124e-01 8.83353591e-01
-4.51245099e-01 -4.72176224e-01 -2.19432220e-01 -5.29767752e-01
7.26299286e-02 7.08065152e-01 -5.48281610e-01 -2.77767509e-01
-2.88587093e-01 -1.00135231e+00 1.63243458e-01 3.08544993e-01
-4.80021417e-01 -7.74132490e-01 -2.67825514e-01 5.26655853e-01
4.13757443e-01 -6.83599889e-01 -1.24804191e-01 7.28856921e-01
-7.82897353e-01 -1.51112342e+00 -3.70837748e-01 -6.41938865e-01
6.54480994e-01 1.10926621e-01 1.01175976e+00 1.04408957e-01
-4.72509205e-01 -4.93029319e-02 -2.72471875e-01 -9.44957495e-01
-8.25232089e-01 1.02339730e-01 -1.03674151e-01 -4.36623544e-01
6.76565766e-01 -6.27204955e-01 -6.77698314e-01 1.09392911e-01
-9.45501089e-01 -2.17970759e-01 5.83531439e-01 7.88638294e-01
4.20805126e-01 5.34765758e-02 7.71486759e-01 -1.35641706e+00
4.17196751e-01 -5.80340087e-01 -5.61277151e-01 3.05803061e-01
-1.03248930e+00 1.87436417e-01 7.24511325e-01 -4.68344420e-01
-7.11535752e-01 8.81508067e-02 1.80459360e-03 4.12140973e-02
-3.59582871e-01 4.45453107e-01 -3.69672954e-01 -1.31835761e-02
9.43167210e-01 -1.51482463e-01 1.32879794e-01 -4.98264372e-01
3.50739732e-02 4.13699418e-01 -2.61950232e-02 -7.13891149e-01
4.24370885e-01 4.49910127e-02 3.00407380e-01 -8.74894440e-01
-4.03229088e-01 -3.22921902e-01 -5.20323336e-01 1.21974908e-01
5.85645914e-01 -6.15791440e-01 -8.24026585e-01 1.03571065e-01
-8.41878653e-01 -3.34999382e-01 -3.47074717e-01 6.06895447e-01
-4.56464082e-01 2.48911768e-01 -4.27766740e-01 -7.31537640e-01
3.91251817e-02 -1.54048479e+00 5.57104111e-01 -4.55396213e-02
-7.66242564e-01 -7.61431217e-01 -5.58554716e-02 3.59640747e-01
7.20596239e-02 7.26760924e-01 1.23116863e+00 -7.87898362e-01
-4.68358248e-01 -2.43905485e-01 1.13239706e-01 -5.86065650e-02
6.35325074e-01 6.39936149e-01 -9.53422248e-01 -3.64451557e-02
2.48868912e-02 -2.93693304e-01 5.07436574e-01 1.85686111e-01
1.00577462e+00 -1.74707845e-01 -1.61700994e-01 2.48685405e-01
1.36767662e+00 7.90781677e-01 3.49688649e-01 3.64393264e-01
7.75281489e-01 9.45937395e-01 4.74361211e-01 3.51842642e-01
-2.27926150e-01 9.06529129e-02 1.27005249e-01 -2.66844600e-01
5.08288085e-01 -2.92619467e-01 -8.48794058e-02 2.85888135e-01
-8.66460428e-02 -2.12489903e-01 -7.61871219e-01 3.92082214e-01
-1.78347206e+00 -9.02691185e-01 -4.08502430e-01 2.34759045e+00
1.17619240e+00 2.12074429e-01 4.80839282e-01 4.65425819e-01
3.02203476e-01 -1.46640331e-01 -5.39641559e-01 -7.81147063e-01
-1.10406823e-01 1.44031316e-01 7.25746214e-01 5.32119274e-01
-7.17782080e-01 8.39825153e-01 7.10783052e+00 4.90226597e-01
-1.39690471e+00 -5.40043831e-01 8.11187625e-01 -2.97877248e-02
-2.96244413e-01 2.73630559e-01 -4.95773643e-01 6.04133964e-01
1.04032278e+00 -2.17695296e-01 2.81085849e-01 5.38019598e-01
8.10223341e-01 -2.34176576e-01 -1.62599504e+00 5.36068320e-01
-4.61353719e-01 -1.47694957e+00 -1.57658890e-01 1.85744956e-01
5.60189605e-01 -3.24275762e-01 9.15670544e-02 -2.59282380e-01
5.55801034e-01 -1.31123185e+00 5.48643529e-01 3.44447464e-01
3.52270603e-01 -7.11833596e-01 2.81668633e-01 2.02286988e-01
-4.48183239e-01 -8.18544999e-03 -3.62120420e-01 -3.57821375e-01
-2.67323434e-01 7.04772949e-01 -1.03033817e+00 2.76898265e-01
3.96233350e-01 6.66382968e-01 -2.98304588e-01 9.75399852e-01
3.32507826e-02 4.39513415e-01 -2.07172036e-01 -2.38775074e-01
2.27311477e-01 -3.75976354e-01 1.26945391e-01 1.16902435e+00
-6.42009676e-02 5.32726645e-02 4.58616614e-02 9.03707027e-01
2.01226816e-01 3.00110579e-01 -9.06954885e-01 -4.88683373e-01
4.34515178e-01 8.10096443e-01 -7.57862747e-01 3.06558963e-02
-6.49195552e-01 2.82538980e-01 -7.53765106e-02 4.72957045e-01
-3.08985740e-01 8.65516514e-02 6.91036344e-01 2.87904203e-01
-2.56602559e-02 -1.10758349e-01 -6.61610961e-01 -8.13115656e-01
-2.04331309e-01 -1.14616382e+00 1.58882245e-01 -5.25698006e-01
-1.14178443e+00 -1.90702945e-01 -3.93572539e-01 -6.94192588e-01
1.14392206e-01 -7.34337568e-01 -4.37962294e-01 9.00466621e-01
-1.16847026e+00 -4.15441364e-01 3.04560512e-01 -7.07523003e-02
4.61064100e-01 1.22033954e-01 7.86161900e-01 3.40685576e-01
-7.25880742e-01 2.42133170e-01 5.63701928e-01 -2.52119601e-01
1.02468204e+00 -1.14959157e+00 1.86063766e-01 4.96124297e-01
-1.89970523e-01 7.14119434e-01 1.09302914e+00 -7.10382938e-01
-1.30883443e+00 -9.46874797e-01 8.86347353e-01 -7.17296839e-01
7.56983221e-01 -3.50599289e-01 -9.68518972e-01 4.92019594e-01
-3.53255242e-01 -4.56378222e-01 1.09743476e+00 2.99910635e-01
-3.15980077e-01 1.34617016e-01 -1.22248578e+00 8.41630399e-01
6.66745603e-01 -2.11908594e-01 -8.92544687e-02 5.55545330e-01
5.40400386e-01 -2.72304833e-01 -1.09186947e+00 4.24783453e-02
4.15772378e-01 -4.82376248e-01 8.58999014e-01 -1.28541481e+00
5.51577210e-01 -3.76434326e-01 4.95906062e-02 -1.12132740e+00
-5.42594433e-01 -5.55703521e-01 3.13853711e-01 9.13016677e-01
8.35759997e-01 -5.51518381e-01 1.09268546e+00 1.37039900e+00
-6.77039325e-02 -7.66268909e-01 -2.98655361e-01 -4.44367677e-01
5.59849858e-01 -8.56930669e-03 5.92870772e-01 1.22196937e+00
1.69166043e-01 5.16397715e-01 -5.52445352e-02 -2.17652261e-01
2.84968525e-01 -1.25800744e-01 7.15792477e-01 -1.30101943e+00
-2.48343442e-02 -6.73572540e-01 -3.01382124e-01 -3.91725212e-01
-2.57997900e-01 -5.03098965e-01 2.39987448e-02 -9.95743155e-01
-2.42166929e-02 -4.42128509e-01 -7.89186135e-02 2.04851463e-01
-2.42237434e-01 -2.61120528e-01 -4.70692888e-02 1.78173706e-01
-1.22217655e-01 1.00350499e-01 1.17887092e+00 -1.47063479e-01
-4.14665014e-01 -1.16931103e-01 -1.34107053e+00 6.28741562e-01
8.14119160e-01 -4.96018171e-01 -3.49940836e-01 8.34493563e-02
5.16948998e-01 -3.08717191e-01 6.15072697e-02 -5.42737722e-01
2.23424464e-01 -8.60558510e-01 6.25933826e-01 -2.16179490e-01
-7.48262256e-02 -9.61100101e-01 5.28450131e-01 3.46178949e-01
-5.51262856e-01 7.01891556e-02 1.04727544e-01 3.19971293e-01
2.12434947e-01 -4.04230297e-01 7.43478537e-01 -6.23156846e-01
-7.46902525e-02 2.89828241e-01 -6.44146979e-01 -5.88893257e-02
9.13411677e-01 -3.63934994e-01 -3.29258829e-01 1.66871189e-03
-7.02612400e-01 1.58971012e-01 8.17149520e-01 1.32332265e-01
2.07205042e-02 -7.22767532e-01 -4.47829813e-01 -2.34569106e-02
2.08886415e-01 4.17907923e-01 -1.11213438e-02 7.12832630e-01
-5.43890297e-01 4.37761486e-01 2.64637709e-01 -9.45276394e-02
-9.97076571e-01 9.96623933e-01 2.39279062e-01 9.91938338e-02
-3.87075454e-01 5.04484475e-01 1.87804028e-01 -1.22790240e-01
3.08341444e-01 -3.81815612e-01 9.18101743e-02 2.02261746e-01
3.73091996e-01 3.16269845e-01 3.74921501e-01 -2.86649764e-01
-4.22310829e-02 -1.78224109e-02 -5.07830739e-01 3.44587088e-01
1.64458966e+00 3.79709005e-01 -8.85634124e-02 7.04173386e-01
1.04385912e+00 7.49054737e-03 -1.17350543e+00 2.57107969e-02
1.57565951e-01 -3.71645361e-01 5.65072941e-03 -8.09572041e-01
-5.57180583e-01 6.66643202e-01 1.05657376e-01 2.12834865e-01
5.41269243e-01 -3.64670455e-01 3.00486356e-01 4.81765389e-01
-2.97603190e-01 -1.59894061e+00 -2.00922951e-01 1.22724697e-01
7.13743091e-01 -1.43642366e+00 4.58034515e-01 -2.39697024e-01
-3.60136002e-01 1.16756105e+00 3.90607357e-01 3.91544491e-01
3.85680348e-01 1.79205835e-01 -1.13751195e-01 -2.07064509e-01
-6.90419495e-01 2.50094067e-02 -4.79020059e-01 4.33892846e-01
8.76601458e-01 -1.99176148e-01 -6.50784791e-01 1.59089550e-01
-7.17787221e-02 1.83430575e-02 6.68807268e-01 1.51153648e+00
-5.36758006e-01 -1.46405089e+00 -5.87402344e-01 5.60098529e-01
-6.24674499e-01 -2.15283841e-01 -9.16475236e-01 6.98227346e-01
1.71088636e-01 1.06706560e+00 -2.16361210e-01 -6.40138835e-02
3.86453867e-01 3.82286608e-01 3.18151683e-01 -6.18512869e-01
-5.45063376e-01 7.50845894e-02 1.19669065e-01 -2.03552067e-01
-3.02322358e-01 -8.68370771e-01 -1.01912951e+00 -7.79748440e-01
-5.18809438e-01 5.59922934e-01 8.12996864e-01 9.97642756e-01
3.51152629e-01 5.00854075e-01 6.88702643e-01 -4.99348551e-01
-2.95074970e-01 -5.46223521e-01 -4.74320948e-01 2.19072878e-01
4.44513857e-01 -5.10358453e-01 -5.87981343e-01 6.15417100e-02] | [4.92990779876709, 5.754281044006348] |
a69c3067-f6ad-48d2-a83e-ae0a8c85d52e | comprehensive-soccer-video-understanding | 1912.04465 | null | https://arxiv.org/abs/1912.04465v4 | https://arxiv.org/pdf/1912.04465v4.pdf | SoccerDB: A Large-Scale Database for Comprehensive Video Understanding | Soccer videos can serve as a perfect research object for video understanding because soccer games are played under well-defined rules while complex and intriguing enough for researchers to study. In this paper, we propose a new soccer video database named SoccerDB, comprising 171,191 video segments from 346 high-quality soccer games. The database contains 702,096 bounding boxes, 37,709 essential event labels with time boundary and 17,115 highlight annotations for object detection, action recognition, temporal action localization, and highlight detection tasks. To our knowledge, it is the largest database for comprehensive sports video understanding on various aspects. We further survey a collection of strong baselines on SoccerDB, which have demonstrated state-of-the-art performances on independent tasks. Our evaluation suggests that we can benefit significantly when jointly considering the inner correlations among those tasks. We believe the release of SoccerDB will tremendously advance researches around comprehensive video understanding. {\itshape Our dataset and code published on https://github.com/newsdata/SoccerDB.} | ['Yudong Jiang', 'Leilei Chen', 'Canjin Wang', 'Changliang Xu', 'Kaixu Cui'] | 2019-12-10 | null | null | null | null | ['highlight-detection'] | ['computer-vision'] | [-2.35794038e-01 -4.98909473e-01 -7.52373397e-01 -6.09854646e-02
-8.48413944e-01 -6.39936924e-01 1.27172157e-01 -2.70467699e-01
-5.57788908e-01 6.38223171e-01 4.29381639e-01 2.02217221e-01
2.82251328e-01 -3.90983999e-01 -1.04095912e+00 -5.31096339e-01
-3.46421987e-01 2.09262982e-01 8.96214068e-01 -4.02290583e-01
4.48046535e-01 8.04977398e-03 -1.86690438e+00 9.41144764e-01
2.95666426e-01 1.07672465e+00 7.18440264e-02 8.79884064e-01
5.33387184e-01 1.17591035e+00 -5.37404001e-01 -5.49779356e-01
1.60796300e-01 -2.10287288e-01 -6.49939537e-01 -5.74834943e-02
6.61653817e-01 -4.98779237e-01 -9.09517884e-01 8.13297808e-01
4.48684633e-01 3.01679820e-01 1.01366438e-01 -1.44426322e+00
-3.23290825e-01 6.50458455e-01 -7.14499831e-01 1.02978492e+00
7.76307106e-01 2.76379555e-01 1.14546967e+00 -9.86120641e-01
9.54608381e-01 8.11906695e-01 5.92993319e-01 4.34142798e-01
-5.43873370e-01 -1.13074279e+00 3.22277874e-01 1.04985964e+00
-1.28839493e+00 -5.98821163e-01 4.57922041e-01 -4.93404806e-01
7.90452302e-01 2.11832121e-01 1.18080711e+00 1.44705033e+00
-3.14789056e-03 1.52863610e+00 6.19816959e-01 8.69030282e-02
-9.98216122e-02 -2.95443624e-01 3.04853708e-01 5.69849849e-01
1.37160748e-01 1.21922739e-01 -1.30955815e+00 2.69091636e-01
7.41330028e-01 -1.21788897e-01 -3.46435934e-01 -4.04527575e-01
-1.50845313e+00 6.00018203e-01 3.82847190e-02 -2.80408729e-02
-8.71143714e-02 2.28841990e-01 1.07349002e+00 1.88759491e-01
3.85362327e-01 1.14297934e-01 -3.74462545e-01 -1.04795825e+00
-8.65247309e-01 8.32886815e-01 6.57774270e-01 1.11042130e+00
7.85144791e-02 -4.85771112e-02 -1.20830715e-01 6.85492992e-01
-1.49094209e-01 3.81952435e-01 9.62822586e-02 -1.24012470e+00
7.36064136e-01 1.88086987e-01 7.94972032e-02 -1.03895593e+00
-2.97887474e-01 -2.17218950e-01 -7.52113387e-02 -2.85366952e-01
7.69015610e-01 7.21961409e-02 -5.67463160e-01 1.36485100e+00
1.76448956e-01 7.50739992e-01 -3.64880145e-01 1.32050109e+00
1.31797409e+00 5.62042058e-01 7.11314157e-02 1.70492128e-01
1.63259506e+00 -1.30420685e+00 -6.80827022e-01 -1.77011773e-01
4.11750048e-01 -6.70453191e-01 1.00860369e+00 9.73367989e-01
-1.09798801e+00 -6.65302813e-01 -1.10988593e+00 -2.09217474e-01
-1.93668708e-01 1.84840083e-01 8.67636800e-01 5.39008439e-01
-5.96942365e-01 1.39278531e-01 -1.04562593e+00 -4.15357560e-01
4.87173021e-01 -1.36647210e-01 -4.47412759e-01 -2.88412392e-01
-1.32809269e+00 6.99802935e-01 6.35087490e-01 -8.51643756e-02
-1.28543341e+00 -8.39397490e-01 -8.08777511e-01 -2.39992365e-01
1.15264630e+00 -6.87354505e-02 1.33235145e+00 -6.67807281e-01
-1.10850704e+00 1.19834816e+00 1.34193033e-01 -4.78075653e-01
5.05604684e-01 -7.14709401e-01 -6.68224037e-01 5.93547702e-01
3.78136009e-01 6.87666297e-01 4.29900318e-01 -7.49390662e-01
-1.05641675e+00 -2.78537929e-01 2.04101801e-01 2.92364806e-01
-1.39248699e-01 5.87585628e-01 -1.34859681e+00 -9.58522141e-01
-4.33144681e-02 -9.90897119e-01 1.73553362e-01 -1.58584014e-01
-2.27141023e-01 -5.67952059e-02 7.04846799e-01 -9.23767984e-01
1.42153418e+00 -2.14019012e+00 1.80303842e-01 -3.50359887e-01
1.92559406e-01 7.68927336e-02 8.10405165e-02 3.31472903e-01
-1.03024669e-01 -1.99807718e-01 3.70352000e-01 1.02639213e-01
1.58062130e-02 7.66420439e-02 -3.65408599e-01 5.59840083e-01
-1.54748887e-01 9.71113622e-01 -9.77037728e-01 -6.39405668e-01
2.60500759e-01 1.82488725e-01 -7.86689281e-01 -7.85291865e-02
-5.42673767e-02 3.97011012e-01 -5.05585670e-01 1.03694904e+00
3.77652198e-01 5.69841154e-02 7.86506608e-02 -2.02743635e-01
-9.88646373e-02 2.17698187e-01 -1.17243409e+00 2.14950442e+00
3.45029026e-01 1.08185363e+00 8.83524269e-02 -1.19859862e+00
4.43521887e-01 2.76695341e-01 8.05049360e-01 -6.23087764e-01
7.58535638e-02 -1.63615689e-01 -2.30824184e-02 -7.55079865e-01
8.53983223e-01 1.13850728e-01 -5.38389325e-01 7.61046782e-02
2.57714748e-01 1.79946154e-01 8.90992701e-01 5.48164904e-01
1.11011791e+00 4.60772097e-01 1.67543218e-01 -3.40791941e-02
1.48489848e-01 2.98998773e-01 9.03585851e-01 9.67020929e-01
-7.37848818e-01 7.69943655e-01 8.58705223e-01 -7.31330454e-01
-9.63207901e-01 -1.10383499e+00 1.62679143e-02 1.51133072e+00
4.13598120e-01 -7.22576141e-01 -6.34059012e-01 -4.74124372e-01
-8.68910775e-02 2.54224420e-01 -6.54818714e-01 -5.99617027e-02
-9.81693387e-01 -6.77257538e-01 8.03748667e-01 8.92069340e-01
5.83159029e-01 -1.12726963e+00 -3.25778544e-01 1.88838512e-01
-7.09835708e-01 -1.61795723e+00 -6.92548096e-01 -3.63074362e-01
-6.74027562e-01 -1.53831708e+00 -6.94072068e-01 -5.49833357e-01
-2.21878901e-01 6.12842441e-01 1.48436856e+00 -4.45185117e-02
-4.77564573e-01 4.12857056e-01 -8.63487542e-01 -3.27662289e-01
1.92537785e-01 -5.39099053e-03 9.67908725e-02 -2.55690157e-01
8.05859029e-01 -1.97575837e-01 -5.71673930e-01 7.74656892e-01
-6.63955152e-01 1.69579521e-01 1.75475284e-01 4.69561249e-01
5.50102711e-01 -7.74710923e-02 2.92056024e-01 -4.94339943e-01
-4.89055440e-02 -5.85225940e-01 -3.30055743e-01 8.95692781e-02
2.38817766e-01 -7.51216590e-01 1.05137704e-02 -2.94815809e-01
-9.25569952e-01 -1.27059370e-01 -4.52382229e-02 -4.42436963e-01
-2.80557096e-01 3.09933722e-01 -5.88508062e-02 3.96309905e-02
5.33525586e-01 3.17082822e-01 -3.53336364e-01 -5.31994522e-01
3.63590986e-01 2.76184946e-01 9.57379639e-01 -7.52050340e-01
3.02189022e-01 6.49723232e-01 -5.90287089e-01 -8.17103803e-01
-9.73748744e-01 -9.29117739e-01 -5.61071336e-01 -9.16511834e-01
1.04606938e+00 -1.38334596e+00 -1.10847795e+00 7.31054425e-01
-7.77364075e-01 -2.18171388e-01 -2.39782147e-02 6.85004115e-01
-6.60362780e-01 5.44360340e-01 -9.79935229e-01 -4.46732730e-01
2.15570316e-01 -1.04207337e+00 1.02124989e+00 2.91384310e-01
3.89331533e-03 -3.55301321e-01 1.35156989e-01 1.08822644e+00
-2.97921807e-01 1.90832227e-01 -1.84450418e-01 -6.12375677e-01
-8.09852600e-01 -3.55655462e-01 -1.90409437e-01 1.43484160e-01
-4.83076185e-01 1.21950217e-01 -6.41521275e-01 -2.74673998e-01
-3.91064316e-01 -6.00622773e-01 1.42131937e+00 6.53977811e-01
1.46647131e+00 3.72720301e-01 -2.86707968e-01 6.08146369e-01
8.50636184e-01 4.42673832e-01 8.58973861e-01 5.85004389e-01
7.06178725e-01 3.58311445e-01 1.16166317e+00 6.77636981e-01
6.56050563e-01 9.68353152e-01 2.64684647e-01 1.44455761e-01
-8.67290422e-02 -3.17866385e-01 7.06151545e-01 6.97388768e-01
-7.94122815e-01 -3.53276581e-01 -9.68945801e-01 6.18031442e-01
-2.03182936e+00 -1.44964576e+00 -5.38643785e-02 1.87522972e+00
5.22981524e-01 4.51572269e-01 6.49902284e-01 1.12554513e-01
8.46902013e-01 5.15197694e-01 -4.90266353e-01 2.45806798e-01
-2.79943764e-01 7.42502734e-02 5.50918519e-01 -2.51922943e-02
-1.71508884e+00 1.14747810e+00 6.31062698e+00 1.17991769e+00
-6.09682500e-01 2.65149355e-01 3.77462894e-01 -8.56057048e-01
5.33289254e-01 -2.26216435e-01 -1.06315660e+00 5.53212166e-01
6.70758724e-01 2.01711468e-02 3.21445391e-02 9.77688432e-01
3.81482929e-01 -3.73488247e-01 -9.14158165e-01 1.19198000e+00
5.22632003e-01 -1.60805666e+00 -2.84784615e-01 -2.12143585e-01
5.95989883e-01 1.08315215e-01 -1.47350812e-02 6.43433988e-01
2.09462523e-01 -7.84670293e-01 1.04939270e+00 4.82794225e-01
5.05470216e-01 -7.53082395e-01 5.67650497e-01 -6.35603294e-02
-1.48444998e+00 -1.35798737e-01 -2.00178176e-01 -2.39770889e-01
5.35996556e-01 1.39597282e-01 -1.73657700e-01 3.93816799e-01
1.31279433e+00 1.53826118e+00 -5.15873373e-01 1.35852361e+00
-1.80406123e-01 7.09594071e-01 -4.55686040e-02 1.39381737e-01
1.69108599e-01 -1.54809272e-02 7.12303817e-01 1.21271777e+00
8.07655677e-02 6.08827412e-01 5.57076156e-01 2.00677797e-01
3.82033773e-02 -5.26496693e-02 -4.16457653e-01 -2.77679414e-01
2.25525245e-01 9.81317401e-01 -1.12444246e+00 -6.40990555e-01
-7.06286311e-01 7.34014511e-01 -1.85235832e-02 2.99897820e-01
-1.50112283e+00 -1.41459450e-01 9.55424845e-01 1.73600674e-01
3.92528772e-01 -3.27396929e-01 6.91663623e-02 -1.73746586e+00
7.70680234e-02 -1.16295910e+00 8.92846525e-01 -9.03432786e-01
-1.03452992e+00 1.48205146e-01 1.27054006e-01 -1.42604280e+00
1.79457754e-01 -7.71288514e-01 -2.42127314e-01 -2.72179753e-01
-8.86929393e-01 -9.82647300e-01 -3.84244710e-01 6.62922084e-01
1.06832409e+00 -1.58583879e-01 8.95303860e-02 7.96089590e-01
-9.15657222e-01 3.98299962e-01 -1.72000304e-01 5.99714696e-01
1.02230561e+00 -8.09051812e-01 2.24706218e-01 7.85906971e-01
4.61323112e-01 -2.73963921e-02 8.32021356e-01 -6.50829852e-01
-1.35098135e+00 -5.25499165e-01 3.19327086e-01 -8.30026507e-01
9.09112632e-01 -5.34361005e-01 -6.86097026e-01 9.64612007e-01
7.64641389e-02 -1.81407616e-01 6.33769512e-01 1.05969585e-01
-2.66444117e-01 -8.77485424e-02 -3.31869215e-01 5.22193134e-01
1.40158975e+00 -3.72927129e-01 -6.06156170e-01 4.64424670e-01
5.19869387e-01 -8.85957778e-01 -9.25965667e-01 4.22328234e-01
9.44585741e-01 -1.16357362e+00 1.32157671e+00 -8.64304900e-01
7.58447230e-01 -1.69737920e-01 -1.28217757e-01 -6.41687512e-01
-5.33551760e-02 -1.58636883e-01 -1.73297673e-01 8.01808417e-01
1.80579379e-01 1.20349787e-01 1.20340788e+00 5.76286852e-01
-3.78239691e-01 -4.74547386e-01 -8.40654790e-01 -7.54041255e-01
-2.93287337e-01 -9.20646548e-01 1.01720490e-01 8.82956982e-01
3.61703575e-01 -3.14623505e-01 -9.46313083e-01 -1.45521000e-01
6.59174502e-01 1.72508024e-02 1.04739273e+00 -6.89117491e-01
-3.15624028e-01 -4.14874852e-01 -7.80095100e-01 -1.36376703e+00
-4.49254736e-02 -4.72738266e-01 -8.57367963e-02 -9.73127544e-01
7.30627656e-01 1.15463749e-01 -5.12747228e-01 4.29577202e-01
-1.21016093e-01 8.12771916e-01 2.52406418e-01 1.20891474e-01
-1.46155131e+00 3.13107818e-01 1.30338728e+00 2.64692493e-02
5.13757057e-02 -7.60768205e-02 -4.35079157e-01 8.69740903e-01
6.70906723e-01 -3.17466468e-01 -9.29145142e-02 -2.44601503e-01
9.51551870e-02 4.44583654e-01 4.46584016e-01 -1.12174499e+00
3.08358043e-01 -2.91677743e-01 2.25276336e-01 -9.73258555e-01
5.70850194e-01 -5.14257140e-02 9.95834358e-03 3.80725682e-01
-2.89753467e-01 -4.08483036e-02 3.18145901e-01 6.47391319e-01
-6.69163287e-01 2.00970888e-01 4.90378082e-01 -2.93980300e-01
-1.69378471e+00 4.65724319e-01 -6.08288467e-01 5.57764947e-01
1.31462026e+00 -4.25795406e-01 -3.79748255e-01 -5.05578756e-01
-1.10918033e+00 4.33250278e-01 3.30266386e-01 9.95243609e-01
6.11722946e-01 -1.24079430e+00 -8.94425333e-01 -1.55200258e-01
3.39864939e-01 -5.33087671e-01 7.90070593e-01 9.64604676e-01
-8.04147065e-01 5.00671029e-01 -5.82509995e-01 -5.78424871e-01
-1.49539566e+00 4.67038423e-01 -1.18171521e-01 -7.43646100e-02
-8.92986715e-01 1.08107185e+00 2.21720591e-01 -2.07640566e-02
4.21756744e-01 -1.83814302e-01 -3.73299390e-01 3.51082653e-01
8.87161314e-01 6.77600682e-01 -2.40105391e-01 -7.01966882e-01
-6.03755593e-01 4.70617145e-01 -1.65740758e-01 2.88448539e-02
1.24973738e+00 6.33320445e-03 2.57431507e-01 3.30472916e-01
6.96997106e-01 -2.43142724e-01 -1.53459108e+00 -5.58369718e-02
-3.54241207e-02 -8.51370811e-01 -1.10679865e-01 -5.91691911e-01
-1.46003211e+00 6.93188548e-01 2.83979833e-01 -2.59603739e-01
1.05828094e+00 2.11093500e-01 8.63265932e-01 2.03488588e-01
5.51817000e-01 -1.43014896e+00 4.68938202e-01 4.45413679e-01
8.32498133e-01 -1.35318482e+00 1.76216394e-01 -5.57941020e-01
-1.04680932e+00 8.05773377e-01 9.03816521e-01 -3.65964204e-01
4.53525126e-01 3.98048878e-01 -1.22420959e-01 -3.89171571e-01
-9.43503320e-01 -1.43329650e-01 2.52294540e-01 3.19385797e-01
2.32178122e-01 4.61074151e-02 -4.18696254e-01 9.81008112e-01
-2.19366267e-01 2.02436805e-01 6.50679410e-01 8.43460262e-01
-6.43907607e-01 -8.14245701e-01 -3.67521524e-01 1.81816071e-01
-7.07294464e-01 -7.72101134e-02 -1.35632649e-01 1.16384470e+00
1.93478420e-01 8.65002155e-01 1.08556546e-01 -4.67408389e-01
4.95550305e-01 -1.15048409e-01 5.54688275e-01 -3.30556631e-01
-3.83108437e-01 1.21450005e-02 5.55128276e-01 -1.08880103e+00
-5.49937725e-01 -9.28544998e-01 -9.34218287e-01 -5.46083987e-01
3.43271717e-02 3.72948870e-02 2.87157387e-01 6.02044940e-01
3.42451483e-02 3.53097707e-01 3.76902358e-03 -1.04089606e+00
2.72894681e-01 -6.88698888e-01 -8.53331149e-01 5.04111588e-01
-1.91892102e-01 -1.14130938e+00 -9.01691169e-02 3.71776432e-01] | [7.950480937957764, 0.2038504034280777] |
56f9997f-5125-429b-864c-807c407f31db | meta-learning-runge-kutta | null | null | https://openreview.net/forum?id=rkesVkHtDr | https://openreview.net/pdf?id=rkesVkHtDr | Meta-Learning Runge-Kutta | Initial value problems, i.e. differential equations with specific, initial conditions, represent a classic problem within the field of ordinary differential equations(ODEs). While the simplest types of ODEs may have closed-form solutions, most interesting cases typically rely on iterative schemes for numerical integration such as the family of Runge-Kutta methods. They are, however, sensitive to the strategy the step size is adapted during integration, which has to be chosen by the experimenter. In this paper, we show how the design of a step size controller can be cast as a learning problem, allowing deep networks to learn to exploit structure in the initial value problem at hand in an automatic way. The key ingredients for the resulting Meta-Learning Runge-Kutta (MLRK) are the development of a good performance measure and the identification of suitable input features. Traditional approaches suggest the local error estimates as input to the controller. However, by studying the characteristics of the local error function we show that including the partial derivatives of the initial value problem is favorable. Our experiments demonstrate considerable benefits over traditional approaches. In particular, MLRK is able to mitigate sudden spikes in the local error function by a faster adaptation of the step size. More importantly, the additional information in the form of partial derivatives and function values leads to a substantial improvement in performance. The source code can be found at https://www.dropbox.com/sh/rkctdfhkosywnnx/AABKadysCR8-aHW_0kb6vCtSa?dl=0 | ['Kristian Kersting', 'Patrick Schramowski', 'Nadine Behrmann'] | 2019-09-25 | null | null | null | null | ['numerical-integration'] | ['miscellaneous'] | [-3.61645132e-01 -2.00183704e-01 -1.69003487e-01 3.11153457e-02
-5.59960604e-01 -6.60445035e-01 2.73998886e-01 4.97764722e-02
-3.63759547e-01 1.00399899e+00 -3.77254635e-01 -5.42825639e-01
-3.60258460e-01 -4.81691480e-01 -7.58624256e-01 -1.05529213e+00
1.37848958e-01 1.46228522e-01 1.20600365e-01 -6.02620244e-01
3.00361603e-01 7.67097771e-01 -1.23353112e+00 -3.47234488e-01
8.69468927e-01 1.17082453e+00 -2.87677366e-02 7.95339882e-01
6.73807785e-02 4.45665479e-01 -4.93312657e-01 1.48726642e-01
4.90457475e-01 -6.22970223e-01 -6.54304802e-01 -1.84753761e-01
-1.07513100e-01 -8.14252645e-02 -2.20875274e-02 9.98082042e-01
7.94929981e-01 3.81357700e-01 6.50823474e-01 -1.21521223e+00
-4.58356068e-02 7.81417117e-02 -2.32142493e-01 2.92626888e-01
-1.30045889e-02 4.85452563e-01 6.29193246e-01 -8.50156546e-01
4.61648732e-01 8.81259561e-01 8.89303625e-01 4.04081702e-01
-1.50024855e+00 -6.01067960e-01 -1.32554427e-01 -2.53236927e-02
-1.42107975e+00 -4.34172332e-01 6.29216671e-01 -5.81294596e-01
5.24225116e-01 1.67850778e-01 5.74353635e-01 7.13622272e-01
2.14267328e-01 3.64027649e-01 1.01557875e+00 -4.03311908e-01
5.41376472e-01 3.64610106e-01 -2.74423100e-02 4.21393156e-01
1.13323547e-01 6.45047575e-02 1.26463324e-01 -1.65090784e-01
1.31081378e+00 -1.30551472e-01 -5.15700400e-01 -6.46888971e-01
-7.33225703e-01 9.59632635e-01 4.64568526e-01 4.49788183e-01
-3.58682394e-01 2.11962268e-01 3.93252194e-01 6.12752140e-01
2.64576763e-01 9.71438944e-01 -7.44583488e-01 -1.77556798e-01
-4.24918711e-01 5.86858213e-01 8.96395981e-01 6.19616210e-01
7.81800449e-01 2.06924781e-01 -9.53247994e-02 6.88906312e-01
-4.93428186e-02 3.96409869e-01 4.38288629e-01 -1.22339678e+00
5.07536791e-02 5.49031913e-01 4.47274864e-01 -8.35452318e-01
-4.21042979e-01 -4.15671557e-01 -6.03197336e-01 7.44518578e-01
8.86473715e-01 -7.39883780e-01 -5.37246346e-01 1.61710167e+00
6.56666338e-01 -1.03835516e-01 -7.34848995e-03 9.72034752e-01
2.08489578e-02 7.85489321e-01 -2.58713156e-01 -3.04941952e-01
8.92591000e-01 -6.21556640e-01 -6.27510428e-01 1.01043724e-01
8.68104935e-01 -5.50429404e-01 1.15303898e+00 3.36635023e-01
-9.74561155e-01 -4.12676930e-01 -9.04878557e-01 2.01613188e-01
-4.70068753e-01 1.37727857e-01 4.97559272e-03 2.44359463e-01
-1.06746256e+00 1.17077935e+00 -6.46283865e-01 -1.48472831e-01
2.18406972e-02 6.19876504e-01 -1.13217950e-01 5.71598351e-01
-1.39659703e+00 8.69024992e-01 4.37722385e-01 2.80961275e-01
-5.50761998e-01 -9.56915438e-01 -5.61938822e-01 1.62813552e-02
6.54552460e-01 -3.91531199e-01 1.42294049e+00 -1.24597371e+00
-1.91402268e+00 4.45466518e-01 1.20282747e-01 -3.93616587e-01
8.33315372e-01 -8.67505595e-02 1.62448287e-02 1.77772015e-01
-1.93726867e-01 1.71579301e-01 1.02814245e+00 -9.61495101e-01
-4.27543998e-01 7.78945088e-02 1.66277871e-01 -5.01203127e-02
-2.31167287e-01 -1.36998773e-01 -8.55930001e-02 -5.26050150e-01
-4.18475330e-01 -9.12460744e-01 -5.00004768e-01 3.35847795e-01
5.48294140e-03 -6.17006086e-02 6.11172378e-01 -5.32961607e-01
1.28663516e+00 -2.07695365e+00 7.75636733e-02 2.57982165e-01
-1.80472732e-02 7.49794781e-01 1.71743020e-01 6.19540989e-01
-3.65989596e-01 7.85941333e-02 -2.48890728e-01 1.57183468e-01
-6.41242415e-02 -1.53118223e-01 -2.07302541e-01 6.19135797e-01
4.52854872e-01 7.32807815e-01 -8.09531689e-01 -1.53042540e-01
2.58423328e-01 5.92587888e-01 -4.62869108e-01 3.31425101e-01
-3.79567087e-01 8.03604186e-01 -7.07294405e-01 -1.35303095e-01
3.78924102e-01 -1.82320952e-01 5.25179468e-02 6.34984206e-03
-6.63600802e-01 8.32476243e-02 -1.54576182e+00 1.28386033e+00
-5.54727733e-01 4.56884205e-01 4.82116401e-01 -1.37154198e+00
7.51706004e-01 5.73483050e-01 5.93602300e-01 -4.43110049e-01
5.80699563e-01 5.21959126e-01 2.37222258e-02 -3.91168445e-01
-4.41992246e-02 -2.43911982e-01 1.26824886e-01 1.77950919e-01
-2.64189720e-01 -3.74256223e-01 4.40306365e-01 -1.59288108e-01
7.63078690e-01 1.97206587e-01 5.58255017e-01 -6.12765729e-01
7.53681004e-01 1.84934497e-01 4.81031537e-01 2.97002256e-01
-4.53393906e-03 4.37408894e-01 9.86652851e-01 -2.57402450e-01
-1.18958580e+00 -2.56312430e-01 -5.34415007e-01 7.33230710e-01
-1.41695186e-01 -1.26500443e-01 -8.38574529e-01 -1.38744324e-01
1.98479503e-01 6.06869638e-01 -5.72127759e-01 -4.06612068e-01
-7.82851815e-01 -3.45778495e-01 2.22774044e-01 5.87246895e-01
2.75056660e-01 -9.35197592e-01 -8.13391447e-01 3.60374033e-01
5.38836956e-01 -6.75435603e-01 -4.68722314e-01 5.49052119e-01
-9.69670236e-01 -9.15511489e-01 -9.74309564e-01 -3.77131790e-01
5.81795633e-01 -2.76100010e-01 5.00202835e-01 1.81272477e-01
-2.15436056e-01 4.30098861e-01 -8.18482637e-02 -1.12049840e-01
-4.89786088e-01 1.92078039e-01 1.94859147e-01 6.42618388e-02
-3.29198450e-01 -6.05519772e-01 -4.30515558e-01 3.25442821e-01
-8.56842935e-01 -2.63034075e-01 2.68970758e-01 9.29861486e-01
3.98183554e-01 -1.10114314e-01 5.87032378e-01 -6.09370470e-01
6.86372578e-01 -4.00160730e-01 -1.25213814e+00 -1.00356810e-01
-6.56061232e-01 4.11725223e-01 1.27463746e+00 -6.61983669e-01
-6.41637206e-01 1.29529938e-01 -3.13349485e-01 -7.06952572e-01
-1.59977421e-01 4.25844729e-01 -5.50627597e-02 -4.26912606e-01
6.45718575e-01 -9.75658968e-02 4.76239800e-01 -5.84099770e-01
1.53139114e-01 3.15360457e-01 1.90752268e-01 -6.71268761e-01
6.05722606e-01 1.50885507e-01 2.14775741e-01 -8.29998732e-01
-3.60049307e-01 -3.70781243e-01 -6.59014821e-01 -8.72459188e-02
4.58143443e-01 -5.20504236e-01 -1.02488947e+00 5.55182338e-01
-8.54821026e-01 -8.89615536e-01 -6.50163412e-01 1.93767413e-01
-7.22157955e-01 6.12078570e-02 -6.07580185e-01 -8.74874830e-01
-9.33359936e-02 -1.13195276e+00 4.94850755e-01 4.33948517e-01
-1.84885889e-01 -1.28292322e+00 1.22344792e-01 -4.55983341e-01
6.46840334e-01 4.61829334e-01 7.27348089e-01 -6.61726773e-01
-1.89636186e-01 -3.07074666e-01 1.30970672e-01 5.80224514e-01
8.45352635e-02 2.24258959e-01 -7.28107333e-01 -2.47289538e-01
3.26932937e-01 -1.75622702e-01 4.60493892e-01 5.57479501e-01
6.83521807e-01 -4.88566697e-01 -2.13547453e-01 6.75929010e-01
1.48379910e+00 3.78320217e-01 2.82963187e-01 4.04114693e-01
7.06638277e-01 3.84085685e-01 5.21431983e-01 6.60965204e-01
-1.48844317e-01 7.11566985e-01 2.60478705e-01 -1.12224326e-01
3.27567160e-01 1.49562553e-01 3.42186391e-01 3.87894332e-01
-7.54594058e-02 1.36777714e-01 -9.02302921e-01 3.32982689e-01
-1.78601360e+00 -6.80476964e-01 -1.46209598e-01 2.38699603e+00
1.15031326e+00 9.99171212e-02 2.95359582e-01 4.61025864e-01
6.21200621e-01 -1.12396806e-01 -7.08345413e-01 -7.65909016e-01
2.35290527e-01 1.14743419e-01 5.47913373e-01 7.77650535e-01
-1.03350401e+00 5.14842987e-01 5.14948463e+00 7.70715415e-01
-1.71560848e+00 -2.44725466e-01 5.23437738e-01 5.85328005e-02
7.64913410e-02 -9.10910591e-02 -8.84361565e-01 6.58023119e-01
1.16155207e+00 -2.72394657e-01 5.44063330e-01 8.65453899e-01
5.85446119e-01 -2.18356445e-01 -8.41845036e-01 5.77401519e-01
-5.46279728e-01 -1.14779150e+00 -5.65959692e-01 -5.85033521e-02
7.27698088e-01 -4.82089758e-01 1.60361286e-02 2.73228973e-01
2.79285684e-02 -8.03093731e-01 7.77682304e-01 5.61814725e-01
5.90329647e-01 -8.14351857e-01 4.93025303e-01 4.54382181e-01
-1.06415391e+00 -3.09273183e-01 -2.01081574e-01 -2.67338306e-01
5.12688272e-02 5.49447596e-01 -5.96365094e-01 3.47891480e-01
4.19229686e-01 4.78067487e-01 -2.13360831e-01 1.32024431e+00
-1.54880136e-01 5.27345359e-01 -5.01930773e-01 -2.74695605e-01
3.15697908e-01 -4.20921743e-01 6.64583862e-01 9.27742481e-01
3.91338915e-01 2.03351349e-01 2.46531311e-02 9.59899426e-01
3.50199640e-01 2.91483641e-01 -4.00710374e-01 -1.39014348e-01
1.51689142e-01 1.09109735e+00 -5.24747610e-01 -1.68681294e-01
-1.92588493e-01 5.13332725e-01 2.98179388e-01 5.53814054e-01
-7.60437369e-01 -6.37331724e-01 8.17451060e-01 3.80398154e-01
7.24379003e-01 -2.70897299e-01 -1.99493021e-01 -8.72381449e-01
8.08438510e-02 -9.94563460e-01 1.94214717e-01 -4.21883911e-01
-7.89999843e-01 3.69979560e-01 1.50969595e-01 -1.24520206e+00
-7.36825526e-01 -8.51498008e-01 -6.79526567e-01 1.00222480e+00
-1.16566288e+00 -3.35430056e-01 1.71703815e-01 5.90196311e-01
2.13700742e-01 2.64529586e-01 5.55168748e-01 3.40732604e-01
-7.45091856e-01 2.78144628e-01 5.96690953e-01 -1.34743024e-02
5.80524862e-01 -1.23232114e+00 6.89092726e-02 7.32213855e-01
-4.91093695e-01 6.14831328e-01 1.12690520e+00 -2.79976398e-01
-1.35334206e+00 -5.83201051e-01 5.85799038e-01 -6.64130822e-02
1.01109636e+00 -2.15575591e-01 -1.28669035e+00 3.57005090e-01
-3.61049734e-02 3.52185369e-01 -1.97876379e-01 -3.08195859e-01
1.39344037e-01 -3.76617998e-01 -9.23759580e-01 6.03260815e-01
3.17717999e-01 -2.19595686e-01 -1.40931517e-01 -1.27402350e-01
5.14957249e-01 -5.95931113e-01 -9.25379336e-01 3.13638031e-01
4.56381232e-01 -6.35105431e-01 8.09507132e-01 -7.42859006e-01
2.86884457e-01 -3.65919799e-01 3.66461337e-01 -1.48090482e+00
-6.28498420e-02 -9.69818234e-01 -2.08027795e-01 1.14799798e+00
3.47526133e-01 -1.01837230e+00 3.68088961e-01 8.89371872e-01
1.94910448e-02 -1.17915356e+00 -7.96960950e-01 -8.94770980e-01
5.45446098e-01 -1.58339605e-01 2.13167861e-01 8.74493599e-01
4.78371233e-02 -4.43716794e-02 -1.15106039e-01 1.70164734e-01
1.19666040e-01 1.82359072e-03 6.94668114e-01 -1.13802803e+00
-4.02819216e-01 -7.26860046e-01 -9.56492126e-02 -8.43868136e-01
2.21041846e-04 -4.03034598e-01 1.68347165e-01 -1.07614803e+00
-7.09709942e-01 -5.21627665e-01 -1.93500206e-01 4.01467472e-01
-2.03512624e-01 -2.00895771e-01 1.40724689e-01 4.17634891e-03
-6.38659000e-02 3.60345781e-01 1.07005525e+00 2.78338432e-01
-6.52988195e-01 3.88064712e-01 -3.69468004e-01 6.67262733e-01
1.11842549e+00 -5.20576835e-01 -2.95402259e-01 -7.22146258e-02
2.08852202e-01 2.77580082e-01 4.50686723e-01 -1.06026077e+00
3.33753377e-01 -3.13515693e-01 6.87263533e-02 2.94174679e-04
2.44998485e-01 -7.57226646e-01 1.33464634e-01 5.91557264e-01
-3.98250699e-01 1.57392517e-01 4.26366806e-01 1.56416908e-01
-6.72807172e-02 -5.78648388e-01 1.06161451e+00 -8.39392021e-02
-5.08273840e-01 6.93816841e-02 -6.31851077e-01 1.77434787e-01
8.88630986e-01 -1.55367807e-01 1.30376324e-01 -4.63388830e-01
-5.10839641e-01 2.96543002e-01 4.58635002e-01 2.94569638e-02
1.63129363e-02 -1.09282017e+00 -3.19529295e-01 2.70197034e-01
-3.72513652e-01 1.35177812e-02 1.04720801e-01 1.27407110e+00
-4.16773736e-01 3.84298921e-01 -1.93681628e-01 -3.73635143e-01
-8.93932164e-01 5.60531557e-01 8.94347787e-01 -2.02410191e-01
-4.27589118e-01 7.06298292e-01 3.00316736e-02 -1.87160522e-01
2.38478169e-01 -6.56626582e-01 -7.88389593e-02 2.57967085e-01
3.65172148e-01 4.78488326e-01 3.66418734e-02 -3.76412272e-01
-2.21532539e-01 6.59720421e-01 3.00947607e-01 -3.71160433e-02
1.35263467e+00 -5.42285256e-02 9.98340398e-02 5.79619467e-01
1.44112885e+00 -1.25164807e-01 -1.67871523e+00 2.07500421e-02
5.45505136e-02 -1.15946956e-01 -1.06119953e-01 -3.67234468e-01
-1.21032369e+00 8.01120460e-01 5.26733756e-01 4.10156310e-01
1.10214388e+00 -4.70902801e-01 5.64252198e-01 4.16613936e-01
1.88388452e-01 -1.09523416e+00 -1.20424278e-01 6.97689354e-01
8.75840545e-01 -1.03935266e+00 1.14482231e-02 -7.86903426e-02
-4.95422542e-01 1.37902975e+00 3.29544306e-01 -5.21397054e-01
7.08871961e-01 4.76646870e-01 3.15766186e-01 1.87739983e-01
-5.64370930e-01 -1.53416306e-01 2.26908728e-01 7.37022758e-02
4.35854167e-01 -3.74887019e-01 -6.19847834e-01 4.08233166e-01
9.93661135e-02 8.34233612e-02 4.04166609e-01 9.64218497e-01
-3.68167281e-01 -1.08084607e+00 -5.96100152e-01 9.82001573e-02
-5.40638328e-01 1.45101845e-01 -1.20617405e-01 9.82829988e-01
-1.04416303e-01 4.75307435e-01 -8.48193169e-02 1.99507579e-01
5.43832958e-01 2.69709885e-01 2.07826152e-01 -3.43493789e-01
-9.01231945e-01 1.19305536e-01 -1.40997738e-01 -5.77824831e-01
-1.19240358e-01 -5.85702121e-01 -1.50463247e+00 -1.86663583e-01
-2.96490848e-01 3.85784745e-01 4.71266121e-01 9.62337673e-01
3.45077038e-01 4.12221968e-01 6.23244226e-01 -1.04067123e+00
-7.99402535e-01 -5.61250627e-01 -5.83041012e-01 1.27446190e-01
7.35070884e-01 -7.42051780e-01 -8.30685973e-01 -1.32388383e-01] | [6.532041072845459, 3.4455363750457764] |
7793df3e-f9d9-4ec9-b19b-ad1347186669 | tune-your-place-recognition-self-supervised | 2203.04446 | null | https://arxiv.org/abs/2203.04446v3 | https://arxiv.org/pdf/2203.04446v3.pdf | Self-Supervised Domain Calibration and Uncertainty Estimation for Place Recognition | Visual place recognition techniques based on deep learning, which have imposed themselves as the state-of-the-art in recent years, do not generalize well to environments visually different from the training set. Thus, to achieve top performance, it is sometimes necessary to fine-tune the networks to the target environment. To this end, we propose a self-supervised domain calibration procedure based on robust pose graph optimization from Simultaneous Localization and Mapping (SLAM) as the supervision signal without requiring GPS or manual labeling. Moreover, we leverage the procedure to improve uncertainty estimation for place recognition matches which is important in safety critical applications. We show that our approach can improve the performance of a state-of-the-art technique on a target environment dissimilar from its training set and that we can obtain uncertainty estimates. We believe that this approach will help practitioners to deploy robust place recognition solutions in real-world applications. Our code is available publicly: https://github.com/MISTLab/vpr-calibration-and-uncertainty | ['Giovanni Beltrame', 'Pierre-Yves Lajoie'] | 2022-03-08 | null | null | null | null | ['visual-place-recognition'] | ['computer-vision'] | [-2.03131825e-01 -6.72093257e-02 -1.35480672e-01 -6.41438186e-01
-1.02351749e+00 -8.33458066e-01 4.50545341e-01 8.75964612e-02
-3.91249239e-01 7.76420414e-01 -1.95426181e-01 -2.68248707e-01
-2.87173893e-02 -6.38778925e-01 -1.19641221e+00 -4.24550295e-01
-7.75839388e-02 6.45533025e-01 2.53106743e-01 3.67110893e-02
3.09574544e-01 7.17764080e-01 -1.32888484e+00 -3.40025306e-01
9.50332403e-01 8.97488892e-01 1.36472717e-01 3.53971332e-01
2.04551995e-01 4.26172316e-01 -2.46315688e-01 7.00452998e-02
4.87925798e-01 -2.67580114e-02 -5.72993755e-01 -2.52225846e-02
5.07497787e-01 -2.10481465e-01 -3.69805157e-01 1.17650259e+00
3.92082721e-01 2.86649168e-01 5.59484601e-01 -1.35524142e+00
-4.73669738e-01 1.81175053e-01 -6.38538718e-01 -1.03652410e-01
1.84334829e-01 4.57975604e-02 5.90705037e-01 -6.57131791e-01
4.86350060e-01 8.66159558e-01 9.45156991e-01 3.41028482e-01
-1.30922210e+00 -7.20400453e-01 2.33747125e-01 2.71069437e-01
-1.79578900e+00 -6.29627049e-01 9.44802225e-01 -4.91352171e-01
7.39939272e-01 -1.33585945e-01 2.17320979e-01 1.24862099e+00
6.07713833e-02 6.22718453e-01 8.55537653e-01 -3.78807098e-01
6.00700617e-01 1.48336783e-01 -3.75112236e-01 6.86715364e-01
4.49543178e-01 3.41292202e-01 -6.37859225e-01 8.95378292e-02
8.51962447e-01 -4.44323495e-02 -2.37034127e-01 -1.22239912e+00
-1.19164705e+00 7.29182303e-01 1.01359701e+00 1.34225562e-01
-2.58394897e-01 5.58304250e-01 -5.54822236e-02 -1.39356777e-01
3.21396142e-01 4.66575444e-01 -2.86849529e-01 -1.69040993e-01
-1.06448197e+00 -8.25041607e-02 5.81282675e-01 1.22471976e+00
1.17403448e+00 -5.53730875e-02 4.51987326e-01 5.98091006e-01
5.03702819e-01 7.25733638e-01 1.30960584e-01 -9.73773301e-01
3.12285423e-01 3.12738538e-01 3.85449767e-01 -1.03557611e+00
-5.51425755e-01 -5.85844100e-01 -5.77443898e-01 3.57069969e-01
3.73332113e-01 -2.30058223e-01 -1.07454801e+00 1.81489217e+00
4.56646472e-01 4.52671140e-01 -1.03234179e-01 9.06464934e-01
2.91997224e-01 2.20009267e-01 -3.11949432e-01 4.75217193e-01
7.54757285e-01 -8.87196958e-01 -3.47614408e-01 -9.61744606e-01
6.70349002e-01 -5.68625987e-01 8.64493787e-01 1.56710938e-01
-2.93381870e-01 -3.97529602e-01 -1.51088154e+00 2.09547520e-01
-3.96509290e-01 1.23926044e-01 6.56550527e-01 5.34833908e-01
-1.00956798e+00 6.25648975e-01 -1.24410903e+00 -6.90367818e-01
3.99348617e-01 4.00390744e-01 -7.13220775e-01 -8.68159607e-02
-9.14116561e-01 1.17013633e+00 4.84101772e-01 1.66054294e-01
-9.73537862e-01 -4.61149931e-01 -1.28930986e+00 -3.43171924e-01
3.65023911e-01 -5.23120582e-01 1.28634560e+00 -6.53814316e-01
-1.35962462e+00 6.99155629e-01 -1.42467842e-01 -5.19585192e-01
5.82460463e-01 -3.17187041e-01 -1.75646633e-01 -1.16816089e-01
3.28420937e-01 6.46519005e-01 6.43467426e-01 -1.59118295e+00
-5.48414946e-01 -4.73061025e-01 -7.12869763e-02 1.93804428e-01
2.41557539e-01 -5.58902740e-01 -5.36100924e-01 -1.67185873e-01
6.49403155e-01 -1.25570786e+00 -3.89016420e-01 2.56637216e-01
-3.26168716e-01 5.10086536e-01 5.57802975e-01 -4.27559108e-01
5.50507247e-01 -2.21333051e+00 -3.47833410e-02 4.85550135e-01
1.52908091e-03 -1.85426965e-01 -1.50160660e-04 3.80394459e-01
1.11376502e-01 -1.62633330e-01 -3.37946504e-01 -4.94707495e-01
1.58628568e-01 3.77544224e-01 -3.46767813e-01 1.19137931e+00
5.25997467e-02 7.26657391e-01 -1.01944315e+00 -1.77180827e-01
5.42823017e-01 5.13251722e-01 -3.21004033e-01 8.25058222e-02
-1.03130005e-01 8.06734681e-01 -4.08565640e-01 7.05239415e-01
8.11303198e-01 -3.19213927e-01 3.99685092e-02 -2.94837542e-02
-7.73606971e-02 1.72033593e-01 -1.25514507e+00 2.22128940e+00
-5.45567393e-01 7.43910491e-01 -4.96826554e-03 -6.73233747e-01
1.03892219e+00 -1.91299230e-01 3.74126345e-01 -5.29871881e-01
1.56906053e-01 2.54437089e-01 -3.23662609e-01 5.91918640e-02
4.84510928e-01 3.53809632e-02 -1.53676733e-01 3.95495668e-02
1.44664913e-01 -3.70498389e-01 -3.04955512e-01 -1.40927270e-01
1.07203245e+00 5.76917529e-01 3.75600487e-01 -5.08852936e-02
3.11053216e-01 2.07671240e-01 7.03685105e-01 7.43189633e-01
-5.08174717e-01 8.09625864e-01 -1.45817371e-02 -3.08308154e-01
-9.52262163e-01 -1.19239414e+00 -2.99719781e-01 6.06137276e-01
6.48330629e-01 -9.59152579e-02 -6.29361510e-01 -6.85615361e-01
2.71693498e-01 6.69315219e-01 -5.51847219e-01 -2.48901606e-01
-2.29308873e-01 -2.97300011e-01 5.72626710e-01 7.69599438e-01
5.22068918e-01 -5.37403464e-01 -5.69969058e-01 6.32959977e-02
-1.66364849e-01 -1.15820110e+00 -1.55072972e-01 4.08132792e-01
-5.59769511e-01 -9.47198570e-01 -5.46966434e-01 -4.45647925e-01
8.61627817e-01 2.92737752e-01 7.43030369e-01 -3.32736462e-01
2.95628812e-02 6.22731745e-01 -3.94728780e-01 -4.29814398e-01
-1.06429033e-01 1.80871680e-01 4.12801027e-01 -9.09404010e-02
3.10686857e-01 -8.94167483e-01 -4.41837370e-01 4.38925296e-01
-5.40190816e-01 -1.46328539e-01 5.15188932e-01 6.66604757e-01
8.16754460e-01 -1.72495767e-01 1.15757391e-01 -4.64436412e-01
9.55142304e-02 -4.15834963e-01 -1.20229530e+00 2.36800522e-01
-6.07626677e-01 3.90899956e-01 2.32022345e-01 -2.52210289e-01
-7.32318461e-01 6.83406830e-01 2.68635154e-02 -7.05997586e-01
-4.41453427e-01 5.63037992e-01 -1.51541397e-01 -8.37269962e-01
1.12059450e+00 1.22347735e-01 -1.86683252e-01 -4.32224154e-01
5.26295483e-01 6.38914347e-01 7.68255115e-01 -6.60593092e-01
1.22899401e+00 7.59148419e-01 1.44312486e-01 -4.23946321e-01
-7.75065005e-01 -7.10060477e-01 -9.18508351e-01 -1.71427414e-01
6.79412007e-01 -1.10661542e+00 -4.76297289e-01 2.72067577e-01
-1.12483943e+00 -4.15251464e-01 3.69775109e-02 5.36508024e-01
-8.41600299e-01 3.99097860e-01 -3.95716205e-02 -8.76019895e-01
7.57649988e-02 -1.04801297e+00 1.24242735e+00 4.27437037e-01
5.60100470e-03 -9.59583223e-01 3.67326647e-01 9.46222395e-02
3.11679810e-01 4.28545743e-01 5.03065921e-02 -6.16047382e-01
-9.29577827e-01 -3.25099349e-01 -7.93128088e-02 3.74832191e-02
1.53653175e-01 -2.50550091e-01 -1.27682519e+00 -3.59638274e-01
-1.42281741e-01 -2.43876234e-01 7.18129933e-01 2.94693261e-01
8.47690582e-01 1.37981966e-01 -5.68114698e-01 1.04470074e+00
1.41002297e+00 3.09371073e-02 5.16843140e-01 8.67544472e-01
7.51342237e-01 3.68902892e-01 8.46369207e-01 3.61876398e-01
6.61563516e-01 7.01159358e-01 7.51856685e-01 -2.99271662e-02
1.52291715e-01 -5.49531996e-01 1.19264610e-01 1.69825643e-01
2.11839631e-01 2.44014058e-02 -1.44204330e+00 7.03374028e-01
-2.28945589e+00 -6.59012377e-01 3.48288864e-01 2.60177040e+00
4.24397856e-01 4.37303772e-03 -2.87065536e-01 -2.24870428e-01
7.20523477e-01 3.25213224e-01 -6.94712043e-01 2.24991530e-01
1.59994215e-01 -2.18703464e-01 1.07762492e+00 6.77371264e-01
-1.46462941e+00 1.19764972e+00 5.67627430e+00 5.24369776e-01
-1.43103778e+00 7.57220015e-02 2.36866280e-01 2.34570891e-01
-9.13003907e-02 2.74010271e-01 -7.48585701e-01 2.06392333e-01
8.49319100e-01 -6.42339736e-02 5.94905555e-01 1.40734494e+00
-1.10146269e-01 -2.85594583e-01 -1.30299246e+00 1.27725387e+00
3.23827378e-02 -1.29532254e+00 -6.88672185e-01 1.08546637e-01
6.10897064e-01 5.56872189e-01 1.06549561e-01 2.59304076e-01
4.31204349e-01 -8.54204535e-01 8.71608257e-01 3.35487157e-01
7.18508363e-01 -6.35075986e-01 6.29664183e-01 4.44607437e-01
-1.15861678e+00 6.24655560e-02 -4.57862556e-01 1.11430138e-01
2.55194873e-01 5.39684474e-01 -1.20925498e+00 7.85731077e-01
6.47140086e-01 8.05722713e-01 -6.54963851e-01 1.53375351e+00
-5.16987026e-01 2.16585070e-01 -7.03919649e-01 2.63055354e-01
1.07276991e-01 2.86987051e-02 5.97218156e-01 7.98147202e-01
4.40764725e-01 -4.31751460e-01 3.09093803e-01 1.08905470e+00
7.99536929e-02 -3.32445890e-01 -1.07152534e+00 1.20224774e-01
7.43856430e-01 1.06970823e+00 -7.23676622e-01 9.00949091e-02
-6.39316365e-02 9.17803645e-01 5.16852021e-01 3.67217362e-01
-1.00467074e+00 -2.89493144e-01 8.19740772e-01 -3.14687416e-02
2.52856404e-01 -6.55790746e-01 -4.56844151e-01 -1.34094155e+00
2.31488466e-01 -4.99728888e-01 -4.06730268e-03 -7.92418182e-01
-1.14350522e+00 5.45400083e-01 -5.13190515e-02 -1.59731066e+00
-4.39361215e-01 -5.26847363e-01 -1.94046095e-01 7.49923825e-01
-1.72994685e+00 -1.29917336e+00 -5.59068978e-01 3.58698308e-01
8.68457183e-02 6.93807006e-02 7.19819605e-01 2.18059912e-01
-3.04021388e-01 5.30015528e-01 2.74886310e-01 2.33407751e-01
9.82660294e-01 -1.12540066e+00 6.07368946e-01 1.29701805e+00
5.02141237e-01 6.17682993e-01 8.64622772e-01 -7.40882635e-01
-1.30291390e+00 -1.19236648e+00 4.36925054e-01 -7.21616268e-01
7.36707926e-01 -5.58460712e-01 -6.80010140e-01 1.05462265e+00
-1.63409457e-01 3.76816005e-01 4.52036172e-01 2.87438363e-01
-5.10587037e-01 -2.37060785e-01 -1.19513249e+00 5.36656737e-01
9.81966197e-01 -8.90382051e-01 -4.95299101e-01 3.48989964e-01
5.20629108e-01 -1.07956839e+00 -4.73261207e-01 6.04279101e-01
3.48469824e-01 -8.54307473e-01 9.19070423e-01 -1.01958424e-01
-1.63490996e-01 -8.56421709e-01 -5.43738604e-01 -1.41663539e+00
-3.49352986e-01 -3.34235311e-01 2.06222069e-02 1.06140399e+00
4.04990375e-01 -8.43084335e-01 9.17839587e-01 7.43416727e-01
-1.51192382e-01 -1.28218263e-01 -1.21459758e+00 -1.14420974e+00
-1.46015942e-01 -7.05228329e-01 5.69729328e-01 9.07066762e-01
-1.98301613e-01 -1.07749715e-01 -4.33592290e-01 1.06320953e+00
9.15778995e-01 -3.66280004e-02 9.74389851e-01 -1.18102789e+00
-2.13096991e-01 -4.29292582e-02 -9.16830778e-01 -9.77561712e-01
3.96523744e-01 -6.08526170e-01 6.45141304e-01 -1.62890685e+00
-2.01487035e-01 -7.38419533e-01 -3.60829860e-01 7.66895950e-01
2.34233513e-01 1.69288382e-01 1.30424827e-01 2.63457537e-01
-6.84977889e-01 6.95017874e-01 5.72539508e-01 -2.35729784e-01
-2.14884520e-01 -5.27500212e-02 -5.61802089e-01 6.83243930e-01
8.00671697e-01 -6.77088797e-01 -3.79655182e-01 -5.92491329e-01
4.24912497e-02 -2.32704490e-01 3.94672811e-01 -1.61447942e+00
5.86191535e-01 -2.14636013e-01 4.67137545e-01 -5.22831559e-01
5.13202012e-01 -1.14106190e+00 3.75360876e-01 1.68081865e-01
4.70399670e-02 -2.22100884e-01 5.11235416e-01 7.05778778e-01
-1.19345061e-01 -1.33302480e-01 7.59054005e-01 1.13794982e-01
-1.11145878e+00 2.84492642e-01 4.40689772e-02 -1.89854741e-01
1.10832047e+00 -1.46911532e-01 -4.19422895e-01 -6.12343252e-01
-4.48989809e-01 2.70157844e-01 1.07422054e+00 4.73528445e-01
5.86511433e-01 -1.45980895e+00 -1.57866523e-01 2.36067325e-01
5.95940173e-01 3.18310380e-01 1.11604370e-01 7.51998842e-01
-8.37684810e-01 3.17485780e-01 -2.60766596e-01 -9.56256509e-01
-7.88085341e-01 5.92284024e-01 6.11565948e-01 2.26512060e-01
-4.05243129e-01 7.54683197e-01 1.26054585e-01 -8.72438133e-01
4.27470833e-01 -2.76721865e-01 3.45065534e-01 -4.66215640e-01
3.98659706e-01 1.13242948e-02 1.60991684e-01 -7.04239130e-01
-8.75455916e-01 6.89323902e-01 1.71503633e-01 -3.57700437e-01
1.26714051e+00 -3.48780990e-01 1.99493036e-01 5.04122615e-01
9.15423393e-01 1.25422776e-01 -1.64254463e+00 -1.80861965e-01
2.46600002e-01 -6.66257262e-01 1.78373694e-01 -6.90574706e-01
-7.34512866e-01 7.65941083e-01 8.63466740e-01 -4.12518442e-01
6.91147685e-01 -6.67150170e-02 2.51156628e-01 8.01615536e-01
1.03528309e+00 -1.00310338e+00 -2.24240035e-01 6.29051447e-01
8.24795783e-01 -1.64878774e+00 -3.52015533e-02 -2.66336441e-01
-6.07405901e-01 8.75578225e-01 5.59428096e-01 -2.71637797e-01
6.23131990e-01 3.21950734e-01 4.26638573e-01 2.51390524e-02
-1.51693553e-01 -3.20373178e-01 4.32142019e-01 1.01190960e+00
8.19315538e-02 5.71615621e-02 5.48892796e-01 2.12307841e-01
-1.14437118e-01 3.12291663e-02 3.90994906e-01 1.09506619e+00
-5.09801209e-01 -1.07541060e+00 -5.70625186e-01 -2.19253153e-01
1.24617584e-01 2.44582035e-02 -3.85813236e-01 8.04585338e-01
-8.56603235e-02 8.48938704e-01 -2.04954013e-01 -6.47445917e-01
2.46027112e-01 -5.12848459e-02 6.04390681e-01 -5.47372580e-01
5.76334745e-02 -6.27551079e-02 1.08172759e-01 -9.37262416e-01
-1.48913443e-01 -5.27871370e-01 -1.30116606e+00 -2.50734597e-01
-1.99291855e-01 -2.47616358e-02 1.05914509e+00 8.78117204e-01
6.44645870e-01 1.88573658e-01 4.30849403e-01 -1.35146642e+00
-4.93123919e-01 -6.37052655e-01 -5.17248154e-01 -3.33778225e-02
5.88863969e-01 -1.07500231e+00 -2.71796346e-01 -2.65385121e-01] | [7.568046569824219, -2.083448648452759] |
4280234b-e56c-495c-8cc9-6445cf313e01 | a-multimodal-perceived-stress-classification | 2206.10846 | null | https://arxiv.org/abs/2206.10846v1 | https://arxiv.org/pdf/2206.10846v1.pdf | A Multimodal Perceived Stress Classification Framework using Wearable Physiological Sensors | Mental stress is a largely prevalent condition known to affect many people and could be a serious health concern. The quality of human life can be significantly improved if mental health is properly managed. Towards this, we propose a robust method for perceived stress classification, which is based on using multimodal data, acquired from forty subjects, including three (electroencephalography (EEG), galvanic skin response (GSR), and photoplethysmography (PPG)) physiological modalities. The data is acquired for three minutes duration in an open eyes condition. A perceived stress scale (PSS) questionnaire is used to record the stress of participants, which is then used to assign stress labels (two- and three classes). Time (four from GSR and PPG signals) and frequency (four from EEG signal) domain features are extracted. Among EEG based features, using a frequency band selection algorithm for selecting the optimum EEG frequency subband, the theta band was selected. Further, a wrapper-based method is used for optimal feature selection. Human stress level classification is performed using three different classifiers, which are fed with a fusion of the selected set of features from three modalities. A significant accuracy (95% for two classes, and 77.5% for three classes) was achieved using the multilayer perceptron classifier. | ['Syed Muhammad Anwar', 'Aamir Arsalan', 'Muhammad Majid'] | 2022-06-22 | null | null | null | null | ['photoplethysmography-ppg'] | ['medical'] | [ 3.48604262e-01 -2.94969827e-01 3.29561122e-02 -4.25702870e-01
-2.89536625e-01 -1.32097498e-01 1.05231171e-02 6.00387812e-01
-5.25017738e-01 8.61168027e-01 1.11782879e-01 2.88897842e-01
-3.44716161e-01 -5.26246727e-01 1.37174368e-01 -8.37100208e-01
-3.42148483e-01 -4.97420162e-01 -1.39503211e-01 -2.85431087e-01
7.61929393e-01 4.14417326e-01 -1.77806342e+00 2.17918083e-01
9.74772930e-01 1.28021359e+00 1.44894600e-01 2.73043454e-01
2.48966917e-01 1.26536399e-01 -8.14119816e-01 1.21205583e-01
-1.62089139e-01 -5.55583119e-01 -5.69897830e-01 4.68162186e-02
-4.40295845e-01 2.82459468e-01 3.24400514e-01 9.85383451e-01
8.11714530e-01 3.34488690e-01 3.99357557e-01 -1.31313288e+00
-6.02322780e-02 1.42020106e-01 -3.24521452e-01 3.92177582e-01
8.79259408e-01 -6.88182786e-02 1.61299273e-01 -1.06591678e+00
-4.18638028e-02 8.53458881e-01 4.95136261e-01 1.96822047e-01
-1.17794228e+00 -7.06538796e-01 -4.18241650e-01 7.58941412e-01
-1.49745619e+00 -2.22629979e-01 1.12321007e+00 -5.16530335e-01
1.13435042e+00 3.09832513e-01 1.17612290e+00 9.98567998e-01
1.05852556e+00 -1.13073893e-01 1.89302647e+00 -5.02133787e-01
7.15670168e-01 6.25101745e-01 5.62096477e-01 1.15336955e-01
4.55326252e-02 -8.38648900e-02 -9.40744281e-01 -5.75065255e-01
8.26606303e-02 2.36138403e-02 -6.32417619e-01 3.27312887e-01
-9.68313992e-01 4.58080024e-01 7.42113516e-02 7.04178333e-01
-8.62979352e-01 -4.96901780e-01 5.04734159e-01 5.01394093e-01
4.08646375e-01 4.58030075e-01 -6.29994810e-01 -3.09070349e-01
-7.37500966e-01 -6.29342273e-02 8.72488320e-01 1.42344847e-01
7.18224108e-01 -8.76000989e-03 -5.15947342e-02 9.07969892e-01
3.72517169e-01 7.19610095e-01 1.02752757e+00 -5.32459140e-01
-1.97418854e-01 8.38410020e-01 -1.13229675e-03 -1.56049383e+00
-9.41884518e-01 -8.49162340e-02 -8.84802699e-01 2.09081769e-01
-1.49180219e-01 -3.89471591e-01 -4.61484551e-01 1.36133528e+00
2.06554696e-01 2.90389895e-01 2.33729184e-02 9.20325518e-01
7.32244730e-01 7.56981730e-01 2.43930042e-01 -9.02782798e-01
1.89100695e+00 1.64833572e-02 -1.09120226e+00 -2.30258390e-01
1.24527842e-01 -5.28885484e-01 8.16847920e-01 8.53603244e-01
-7.89234161e-01 -6.22225285e-01 -1.39706624e+00 5.26469588e-01
-5.62709093e-01 -6.92631975e-02 3.21359158e-01 9.37891901e-01
-9.61221993e-01 6.92662954e-01 -6.91213846e-01 -5.27584195e-01
-1.09876625e-01 5.31108677e-01 -5.95559120e-01 3.71109813e-01
-1.57745171e+00 1.27185869e+00 2.55320519e-01 2.16016054e-01
-1.11640900e-01 -3.41093868e-01 -7.55795538e-01 1.05394661e-01
-2.09781617e-01 -3.03157806e-01 6.00232184e-01 -1.01771653e+00
-1.81764543e+00 7.41404235e-01 -2.37774298e-01 4.56284313e-03
-4.80152756e-01 6.46459833e-02 -9.21173453e-01 6.84526742e-01
-2.46492639e-01 1.23851977e-01 9.32673514e-01 -7.08833158e-01
-8.91626626e-02 -9.05524611e-01 -6.11558497e-01 2.92252392e-01
-5.46963692e-01 3.73721838e-01 6.85193896e-01 -2.02852443e-01
4.93349940e-01 -8.13171387e-01 8.62979293e-02 -6.99669957e-01
-8.22731331e-02 -2.87746936e-01 4.15070742e-01 -1.07759571e+00
1.27601397e+00 -2.24951506e+00 1.49157777e-01 4.83954668e-01
-1.53813779e-01 1.04412377e-01 3.29544127e-01 5.17095625e-01
-4.53896344e-01 -1.50735721e-01 -2.03012154e-01 5.73383346e-02
-9.48957577e-02 -2.23476529e-01 1.56249022e-02 5.30484080e-01
2.00707421e-01 2.16164380e-01 -6.77798271e-01 -7.60593414e-02
3.30256552e-01 6.60105467e-01 3.31553891e-02 3.56195807e-01
7.92129636e-01 4.49724942e-01 -1.88872904e-01 6.15765929e-01
7.45634258e-01 4.33484048e-01 5.78434244e-02 -2.60516554e-01
-1.12168834e-01 -6.40633628e-02 -1.14997303e+00 1.38263929e+00
-1.82588547e-01 3.76327425e-01 -1.31638288e-01 -9.40258443e-01
1.57276583e+00 7.51504838e-01 5.59661806e-01 -8.89647663e-01
4.52394515e-01 1.17021836e-01 -1.56544402e-01 -1.11375022e+00
1.06797963e-01 -3.55478913e-01 -1.76564559e-01 1.15874149e-01
2.65294522e-01 -6.01983629e-02 -1.36183619e-01 -2.06703112e-01
8.15199852e-01 -3.69121581e-01 7.54551291e-01 -6.13692522e-01
6.78684235e-01 -4.94910717e-01 6.71638608e-01 3.79523784e-02
-6.09296262e-01 9.90044251e-02 8.09077799e-01 -2.03329131e-01
-2.85971403e-01 -5.62046170e-01 -4.62409467e-01 6.45882249e-01
7.53637403e-02 -4.01712731e-02 -8.67256403e-01 -3.66421975e-02
-1.72780409e-01 7.79363334e-01 -5.60128510e-01 -5.68074584e-01
5.86866103e-02 -9.68016505e-01 1.41852796e-01 6.10467531e-02
4.78955179e-01 -1.26992571e+00 -9.79476929e-01 2.84909606e-01
-1.26208887e-01 -5.31434476e-01 2.00127199e-01 3.80988210e-01
-8.50515366e-01 -8.20462644e-01 -4.91425037e-01 -4.57934588e-01
2.92701483e-01 2.66026072e-02 6.98520482e-01 -2.03456640e-01
-3.06587517e-01 3.79478812e-01 -4.12554592e-01 -6.08571649e-01
1.55149773e-01 -4.24296141e-01 4.72853333e-01 2.90718377e-01
8.65630746e-01 -9.95726407e-01 -6.81402504e-01 1.54687673e-01
-4.92180228e-01 -4.46266681e-01 4.30949569e-01 4.89925951e-01
2.69272506e-01 1.07014813e-01 1.23126447e+00 -1.81985214e-01
9.56020296e-01 -7.76741147e-01 3.65238786e-02 -1.48871109e-01
-5.28106928e-01 -7.26359129e-01 3.07375878e-01 -3.44544858e-01
-9.79142487e-01 -2.66065598e-01 -5.33390790e-03 6.81538582e-02
-6.81486070e-01 6.22341752e-01 -2.93351620e-01 -2.09565833e-01
9.60438728e-01 1.02252729e-01 2.58164061e-03 -1.67729348e-01
-4.40676242e-01 1.14971352e+00 2.82937050e-01 -1.34780645e-01
-1.31073609e-01 -1.34318590e-01 9.28022116e-02 -1.14068365e+00
-3.77409607e-01 -3.71021330e-01 -5.58581293e-01 -5.75927019e-01
8.86427164e-01 -6.93355560e-01 -9.21217203e-01 6.57927275e-01
-8.87693346e-01 2.34640449e-01 4.45298672e-01 1.03302574e+00
-3.83674443e-01 2.44361162e-01 -4.55733955e-01 -1.32804000e+00
-5.34859538e-01 -9.01932299e-01 5.74805975e-01 5.95686495e-01
-4.78099346e-01 -7.09554493e-01 1.37170002e-01 1.19723074e-01
3.23652327e-01 6.07640564e-01 7.75584817e-01 -6.56759679e-01
4.39847529e-01 -3.37594926e-01 3.23666304e-01 5.01060128e-01
1.47599354e-01 -2.82279700e-01 -1.23723090e+00 -1.37937695e-01
8.29497576e-01 -2.60251641e-01 2.68247724e-01 4.73728329e-01
8.87455702e-01 1.47357851e-03 -1.20315231e-01 1.56627104e-01
1.42601335e+00 8.60949874e-01 9.03621137e-01 3.86234708e-02
7.78604746e-02 7.66422987e-01 5.65319657e-01 5.88018239e-01
2.04124585e-01 3.22188914e-01 1.55665472e-01 1.10071890e-01
5.74780226e-01 4.43811953e-01 5.73948145e-01 9.11217391e-01
-1.01895630e-01 1.20633908e-01 -7.76761949e-01 2.47068867e-01
-1.41424322e+00 -9.54099476e-01 -1.87131360e-01 2.40050578e+00
6.92865849e-01 2.94110272e-03 1.29028007e-01 9.19471979e-01
8.00595462e-01 -3.03801447e-01 -4.00697172e-01 -8.05861592e-01
4.79583675e-03 4.98257190e-01 -9.10343006e-02 9.68714654e-02
-7.67603934e-01 -2.32281350e-02 5.98858166e+00 1.77100122e-01
-1.39970815e+00 -1.41164199e-01 5.17055273e-01 7.51336962e-02
2.64758945e-01 -8.94298479e-02 -2.81237483e-01 9.19138610e-01
1.45370364e+00 -3.13378364e-01 4.52762365e-01 4.99238610e-01
5.97027242e-01 -7.24346459e-01 -7.72772610e-01 1.33975589e+00
2.41404727e-01 -3.71060371e-01 -7.68351197e-01 -1.82315826e-01
1.48394838e-01 -5.15804231e-01 -2.27979288e-01 2.29283825e-01
-1.03706169e+00 -8.53367865e-01 1.99158490e-01 1.20414567e+00
4.31089789e-01 -1.10625708e+00 9.77039456e-01 3.44637126e-01
-8.08783531e-01 -2.29251832e-01 -2.56605178e-01 -5.31289995e-01
-5.57871014e-02 1.10125220e+00 -4.05268759e-01 5.87655842e-01
8.12490761e-01 3.88496518e-01 -3.40009987e-01 1.07510984e+00
-7.80720711e-02 6.96594596e-01 -1.75114766e-01 -2.91204631e-01
-4.79420602e-01 -2.39410281e-01 5.43209791e-01 1.13327515e+00
6.00074649e-01 6.20621026e-01 -3.82662527e-02 6.64430439e-01
4.92425501e-01 3.77904207e-01 -5.26529491e-01 1.66990638e-01
6.47466183e-01 1.66041398e+00 -8.36807787e-01 -1.50184125e-01
-1.85474500e-01 7.07407057e-01 -2.16888815e-01 2.41499558e-01
-4.91910636e-01 -1.05236101e+00 2.52630621e-01 -3.05086434e-01
-6.17359161e-01 8.53999332e-02 -5.58476567e-01 -9.83692527e-01
9.44646634e-03 -6.09496176e-01 2.21553653e-01 -9.95055914e-01
-1.21862078e+00 5.34375429e-01 2.13649288e-01 -1.12696326e+00
7.23068938e-02 -3.77566963e-01 -7.72968829e-01 1.28865635e+00
-1.10547042e+00 -3.23273569e-01 -4.98566031e-01 5.30276656e-01
3.31723578e-02 -3.36066969e-02 1.14888906e+00 3.61063868e-01
-6.86530590e-01 1.90836266e-01 -4.70263332e-01 -5.36138237e-01
7.98704505e-01 -1.20655727e+00 -5.55302143e-01 3.40620071e-01
-7.42490113e-01 5.96195102e-01 6.11822009e-01 -4.87773716e-01
-1.32029474e+00 -5.34490108e-01 1.22338152e+00 8.49648714e-02
2.76911318e-01 -1.18454345e-01 -1.06430960e+00 -5.65815046e-02
4.06407446e-01 -3.78915727e-01 1.17706299e+00 -5.35215549e-02
4.28352982e-01 -2.65338957e-01 -1.65142012e+00 1.65724963e-01
1.18473023e-01 -4.32198137e-01 -8.96542430e-01 3.67045961e-02
3.30966115e-01 -1.73913129e-02 -1.53341794e+00 4.14812595e-01
6.43818915e-01 -1.16065955e+00 4.84244347e-01 -5.07819317e-02
1.35825157e-01 -1.08077221e-01 -9.97721329e-02 -1.73755074e+00
-3.76314193e-01 -4.45766836e-01 5.38511313e-02 1.08238709e+00
-4.23967391e-02 -1.06947339e+00 7.30933920e-02 8.02789032e-01
-1.56035140e-01 -8.58300626e-01 -7.43784845e-01 -2.10303769e-01
-5.84488571e-01 -1.97063535e-01 4.04736698e-01 1.17345500e+00
1.24974000e+00 4.21286047e-01 -1.34818479e-01 1.46511406e-01
3.51186931e-01 -4.22527879e-01 7.50001222e-02 -1.36086273e+00
1.92394331e-01 -1.92734465e-01 -9.63698030e-01 1.52448133e-01
3.52873020e-02 -5.86938322e-01 -1.91166967e-01 -1.27705753e+00
1.16238065e-01 2.44659454e-01 -8.58664215e-01 5.68389177e-01
-2.68278003e-01 2.66219974e-01 -1.14629559e-01 -3.28591853e-01
2.47040316e-01 3.79686415e-01 6.71161175e-01 4.42339987e-01
-6.85260892e-01 -5.50842062e-02 -6.71120286e-01 7.24837422e-01
9.94758368e-01 -4.38293964e-01 -4.76963758e-01 3.76896560e-01
3.07568163e-01 7.29649663e-01 3.38933438e-01 -1.46499979e+00
8.91611055e-02 -2.19739735e-01 7.22294211e-01 -4.01814044e-01
6.40733004e-01 -6.50633276e-01 2.54425973e-01 5.48600793e-01
-1.26683950e-01 -3.24699208e-02 2.98765242e-01 1.38204634e-01
-2.54538625e-01 -2.97814816e-01 8.90424252e-01 1.98947191e-01
-4.58709896e-01 -1.64442182e-01 -6.86323225e-01 -7.86300540e-01
1.14392078e+00 -3.72512341e-01 -1.06047414e-01 -1.22520648e-01
-1.04081440e+00 -4.21314575e-02 -7.64602721e-02 1.40192553e-01
8.12867343e-01 -1.28564394e+00 -3.73736411e-01 4.40487474e-01
9.12782773e-02 -8.79007518e-01 7.86320686e-01 1.40596962e+00
4.22871560e-02 1.41486824e-01 -8.57153714e-01 -4.90551740e-01
-1.20552564e+00 3.00292552e-01 2.68584967e-01 3.63099873e-01
-1.98951006e-01 4.94872808e-01 -6.23996139e-01 1.75391212e-01
8.09300169e-02 -2.44702682e-01 -1.00517857e+00 5.69347024e-01
9.60230350e-01 8.54021251e-01 5.27307570e-01 -5.96125603e-01
-5.60734272e-01 5.07969260e-01 5.09475887e-01 -3.10780525e-01
1.12832224e+00 -1.93420783e-01 -5.42534471e-01 8.98731112e-01
1.07627356e+00 -1.41284138e-01 -5.40767789e-01 2.59193718e-01
-7.75744952e-03 -3.72625768e-01 1.69833586e-01 -1.00075650e+00
-4.92358685e-01 9.05613840e-01 1.16971457e+00 7.54354179e-01
1.74983490e+00 -7.40212977e-01 6.02925181e-01 2.38582402e-01
4.82791841e-01 -1.25784862e+00 -8.27740803e-02 5.09081520e-02
8.18134964e-01 -8.22122872e-01 -1.71638787e-01 -1.64645284e-01
-8.07322443e-01 1.30866325e+00 3.20485950e-01 -3.44913006e-01
9.85215843e-01 4.79428330e-03 -1.20318599e-01 -2.79588193e-01
-8.04763079e-01 2.42814466e-01 2.76213795e-01 5.23583710e-01
5.22326350e-01 2.09906206e-01 -9.91893172e-01 1.17844605e+00
-1.08979382e-01 3.12715024e-01 5.19552350e-01 1.00792742e+00
-1.01102591e+00 -6.66697383e-01 -7.16619968e-01 6.22426033e-01
-5.56850672e-01 3.27145666e-01 -2.18166411e-01 1.10115953e-01
4.76842850e-01 1.50824547e+00 -1.06190361e-01 -6.16200089e-01
6.06177688e-01 6.05356395e-01 4.23179507e-01 -4.21845526e-01
-7.86815763e-01 -3.39926191e-04 -8.49139616e-02 -6.01468682e-01
-6.28238678e-01 -3.66981626e-01 -1.23708332e+00 -6.07009009e-02
-2.19101384e-01 3.19915533e-01 1.12380660e+00 1.03530586e+00
3.75075847e-01 3.91662240e-01 1.08951020e+00 -1.09256995e+00
-6.20834678e-02 -1.30480719e+00 -1.16195869e+00 2.40723550e-01
1.02956384e-01 -8.90784919e-01 -6.13392651e-01 3.55270542e-02] | [13.517129898071289, 3.1150341033935547] |
f90fe6bf-60aa-4acc-8d0b-a40c4bdf9b67 | an-empirical-comparison-of-lm-based-question | 2305.17002 | null | https://arxiv.org/abs/2305.17002v1 | https://arxiv.org/pdf/2305.17002v1.pdf | An Empirical Comparison of LM-based Question and Answer Generation Methods | Question and answer generation (QAG) consists of generating a set of question-answer pairs given a context (e.g. a paragraph). This task has a variety of applications, such as data augmentation for question answering (QA) models, information retrieval and education. In this paper, we establish baselines with three different QAG methodologies that leverage sequence-to-sequence language model (LM) fine-tuning. Experiments show that an end-to-end QAG model, which is computationally light at both training and inference times, is generally robust and outperforms other more convoluted approaches. However, there are differences depending on the underlying generative LM. Finally, our analysis shows that QA models fine-tuned solely on generated question-answer pairs can be competitive when compared to supervised QA models trained on human-labeled data. | ['Jose Camacho-Collados', 'Fernando Alva-Manchego', 'Asahi Ushio'] | 2023-05-26 | null | null | null | null | ['answer-generation'] | ['natural-language-processing'] | [ 5.56629658e-01 4.86957401e-01 2.10861564e-01 -5.34764946e-01
-1.55162668e+00 -9.08856332e-01 9.40562487e-01 1.80507049e-01
-1.95677280e-01 8.13233137e-01 4.05145377e-01 -9.48475957e-01
1.68600157e-01 -9.40228403e-01 -7.93712795e-01 2.47839792e-03
5.76120138e-01 8.24467063e-01 1.47352025e-01 -6.05936170e-01
3.51080120e-01 -1.53039142e-01 -1.50728512e+00 6.84993565e-01
1.27694511e+00 6.31555736e-01 3.44263501e-02 1.12577415e+00
-6.89962089e-01 1.29830003e+00 -8.23916495e-01 -7.59523094e-01
-2.49698713e-01 -9.96136367e-01 -1.38092363e+00 -3.73739153e-01
1.07017756e+00 -8.00254196e-02 -3.49472538e-02 5.98722935e-01
5.23740232e-01 2.65584677e-01 6.13784313e-01 -1.08042634e+00
-1.18090653e+00 4.63704169e-01 3.35593559e-02 1.67957604e-01
8.75453055e-01 2.79471189e-01 1.12621129e+00 -1.08671117e+00
6.45315707e-01 1.50510001e+00 4.09511387e-01 9.43367124e-01
-1.34081340e+00 -3.17046076e-01 -7.32966959e-02 3.43285769e-01
-7.39915133e-01 -3.49579483e-01 4.34415817e-01 -3.52276176e-01
1.00351691e+00 3.94270748e-01 5.28852548e-03 1.03703141e+00
1.91299945e-01 8.42279315e-01 1.09049118e+00 -8.31772208e-01
2.11137429e-01 -1.98489532e-01 4.74004507e-01 6.69238150e-01
-1.33372620e-01 -2.72846520e-01 -3.47122043e-01 -3.55176866e-01
1.92209467e-01 -6.07375801e-01 -2.56345719e-01 2.26362310e-02
-9.36537862e-01 1.30545354e+00 1.97336748e-01 2.80149318e-02
-1.42181560e-01 2.26826578e-01 1.32823750e-01 6.32530749e-01
3.47444892e-01 1.08293390e+00 -5.31720340e-01 -1.59557968e-01
-8.16679478e-01 6.69241965e-01 1.14636207e+00 8.98503721e-01
7.88543880e-01 -1.52139366e-01 -1.06836677e+00 8.90903473e-01
2.70861089e-01 4.87412393e-01 5.38777232e-01 -1.02892888e+00
8.33380640e-01 6.98528588e-01 1.14589691e-01 -5.06995499e-01
-2.46653602e-01 -1.20292813e-01 -5.43076873e-01 -2.84562618e-01
6.36911333e-01 -4.46636796e-01 -1.24489713e+00 1.89338100e+00
3.41499478e-01 1.17512487e-01 1.91154554e-01 4.44059193e-01
1.55634713e+00 7.78285265e-01 4.04518485e-01 1.85206026e-01
1.65983832e+00 -1.31982458e+00 -8.39074790e-01 -6.49218857e-01
7.89889038e-01 -9.42762017e-01 1.45354342e+00 -1.09820932e-01
-1.32599568e+00 -8.09010267e-01 -5.59612989e-01 -6.90530241e-01
-4.33781356e-01 1.36705367e-02 1.77468225e-01 8.47811759e-01
-1.22493458e+00 6.00455478e-02 -2.55539477e-01 -3.42300743e-01
8.58616754e-02 -3.15946452e-02 -2.10883208e-02 -3.98175478e-01
-1.43290603e+00 1.16002369e+00 1.75695613e-01 -1.79126948e-01
-9.09737527e-01 -7.55741298e-01 -1.16255176e+00 1.67326748e-01
1.75697550e-01 -1.37037563e+00 1.86254478e+00 -5.66234767e-01
-1.59498835e+00 7.99343765e-01 -5.92284083e-01 -6.77283645e-01
1.67769119e-01 -3.29073995e-01 -2.47347787e-01 1.50524214e-01
1.53346315e-01 1.01235330e+00 8.18261981e-01 -1.00674856e+00
-2.07132816e-01 -1.24230996e-01 3.75004768e-01 1.25652313e-01
1.74287558e-01 2.85462830e-02 -2.17111900e-01 -4.51268613e-01
-1.38640404e-01 -7.34781981e-01 -3.40963215e-01 -6.49104774e-01
-2.94878006e-01 -6.34449780e-01 5.16473889e-01 -1.13788176e+00
1.38136053e+00 -1.43722367e+00 -2.02770144e-01 -1.09906219e-01
-2.75967896e-01 4.91698354e-01 -6.87059522e-01 9.35915112e-01
-1.67591646e-01 2.65333772e-01 -3.84668887e-01 -3.83841187e-01
1.46936104e-01 1.19001955e-01 -6.58010960e-01 -3.53727400e-01
7.83948064e-01 1.64400268e+00 -1.01191390e+00 -4.34446841e-01
-1.43784568e-01 4.76490036e-02 -6.87200069e-01 6.80824280e-01
-9.27318871e-01 5.55104315e-01 -3.16590726e-01 3.33529532e-01
2.69334495e-01 -4.84893173e-01 -6.29354119e-02 2.80063540e-01
4.51369226e-01 9.65624511e-01 -6.20139956e-01 1.72780919e+00
-7.43370354e-01 5.87455273e-01 -1.76442787e-01 -6.36598587e-01
9.39455032e-01 5.17770469e-01 -5.37117839e-01 -8.54543209e-01
-1.55479595e-01 1.78360030e-01 5.64796291e-02 -7.68371105e-01
7.81781256e-01 -1.03067793e-01 -1.57452404e-01 7.03353226e-01
4.53334987e-01 -6.98627949e-01 6.46665335e-01 4.48854625e-01
1.27387166e+00 2.32055739e-01 2.58051544e-01 -6.00841902e-02
8.38986337e-01 2.00204715e-01 -6.57908320e-02 1.11294043e+00
1.06849112e-01 8.11118722e-01 3.59219909e-01 1.24371439e-01
-8.65692496e-01 -9.84917760e-01 9.68175381e-02 1.23729074e+00
-3.59567791e-01 -3.80690515e-01 -9.60945666e-01 -8.32068563e-01
-1.08127370e-01 1.35683715e+00 -7.20931888e-01 -3.09451818e-01
-7.67298937e-01 -3.34752798e-01 7.24351108e-01 5.70182323e-01
3.19496572e-01 -1.29729092e+00 -1.76183388e-01 3.62834960e-01
-5.27816653e-01 -1.05511999e+00 -4.60812390e-01 -1.75553367e-01
-9.03864741e-01 -8.73239040e-01 -6.55445039e-01 -8.79730761e-01
6.15753353e-01 3.89333181e-02 2.02115893e+00 2.71502852e-01
1.17385477e-01 8.32776725e-01 -4.97590065e-01 -4.68152195e-01
-9.54919696e-01 4.51812238e-01 -6.91269875e-01 -2.92290926e-01
4.10053521e-01 -2.62697846e-01 -6.15824819e-01 3.53355967e-02
-1.13447714e+00 3.60037573e-02 5.37066579e-01 9.93502319e-01
4.20541406e-01 -8.73675287e-01 1.33231616e+00 -1.32044077e+00
1.29930425e+00 -6.06630445e-01 -2.79922128e-01 8.03214669e-01
-5.15345693e-01 4.31775033e-01 4.80367720e-01 -1.56105936e-01
-1.39071453e+00 -4.03443754e-01 -6.98509932e-01 1.68191090e-01
-3.28287452e-01 5.88321447e-01 -1.41276062e-01 2.27835357e-01
9.62702453e-01 2.32974574e-01 -3.38598154e-02 -3.77518028e-01
1.13987982e+00 4.68982518e-01 6.35734379e-01 -7.37805247e-01
8.38849366e-01 -1.80093378e-01 -2.42839858e-01 -3.98169875e-01
-1.24080062e+00 -5.01195073e-01 -4.15312320e-01 6.90473169e-02
1.02767193e+00 -6.94378257e-01 -4.41821367e-01 -2.21300572e-02
-1.38832366e+00 -5.28715312e-01 -4.41652387e-01 -7.24835396e-02
-4.67877060e-01 2.59886324e-01 -7.45358050e-01 -7.00002730e-01
-6.75633907e-01 -7.67697275e-01 1.07939601e+00 5.17999589e-01
-5.17778277e-01 -1.28260970e+00 3.64611089e-01 1.13810730e+00
4.23200250e-01 -4.84793000e-02 1.29673982e+00 -8.92092049e-01
-5.54034591e-01 -9.09950137e-02 -8.27979445e-02 4.46698606e-01
1.94389131e-02 -2.56789804e-01 -1.14209592e+00 -5.74590042e-02
-9.13137421e-02 -8.31709564e-01 8.45245361e-01 4.37437631e-02
1.12355149e+00 -5.13554692e-01 1.51419565e-01 -1.13541327e-01
1.05230677e+00 -2.00666934e-02 8.07703495e-01 5.66184856e-02
5.34228742e-01 9.07955885e-01 5.42251468e-01 -3.83765072e-01
7.25698769e-01 2.82233447e-01 2.40168825e-01 7.97036365e-02
-5.26917160e-01 -5.97868502e-01 2.95289665e-01 7.17466831e-01
6.21666074e-01 -5.11643827e-01 -1.06967366e+00 7.91580379e-01
-1.62122703e+00 -9.56254303e-01 -5.86112261e-01 2.02531481e+00
1.34436393e+00 -2.17646480e-01 -1.18458092e-01 -2.79187471e-01
3.78720492e-01 2.21552048e-02 -4.00292248e-01 -6.92349136e-01
-3.65398452e-02 9.47367072e-01 -7.71935582e-02 9.16436791e-01
-6.28195763e-01 8.83261979e-01 6.72585344e+00 7.23645806e-01
-4.09162968e-01 1.93455130e-01 7.13891983e-01 4.08800453e-01
-8.75686824e-01 6.44444749e-02 -9.10233557e-01 2.30055124e-01
1.42999589e+00 -1.16690859e-01 1.92097113e-01 5.19983351e-01
-6.06661849e-02 -1.45739436e-01 -1.28080261e+00 5.62961698e-01
4.69130337e-01 -1.38299108e+00 5.41658103e-01 -4.59492624e-01
1.09432328e+00 -2.73679584e-01 -9.90263931e-03 8.81847501e-01
7.85075307e-01 -1.32239604e+00 2.87548691e-01 5.77554524e-01
4.48103100e-01 -5.27218699e-01 8.58027518e-01 5.20754457e-01
-7.50910819e-01 4.61188368e-02 -8.84822458e-02 -2.29847431e-01
4.13337886e-01 3.32478195e-01 -1.34416032e+00 7.23465562e-01
2.61902213e-01 -1.40581369e-01 -1.19528675e+00 9.10740912e-01
-8.12788844e-01 1.19052398e+00 7.18344599e-02 -1.59973115e-01
3.16041440e-01 -1.33254737e-01 8.79120305e-02 1.31131017e+00
8.19472894e-02 1.55623227e-01 -1.80297956e-01 1.09553146e+00
-3.62648547e-01 1.32595837e-01 -4.37639475e-01 -4.53520045e-02
4.59983557e-01 1.18338990e+00 -1.43200353e-01 -5.48740029e-01
-4.42419589e-01 8.55355144e-01 4.53900397e-01 5.43797016e-01
-4.41801667e-01 -3.22884321e-01 2.71420240e-01 -6.25684261e-02
1.13682225e-01 -1.15803890e-01 -4.52405512e-01 -1.11622524e+00
-5.23432679e-02 -1.38923895e+00 7.20446706e-01 -1.06180573e+00
-1.50373805e+00 4.09972787e-01 -9.15872678e-02 -7.05360413e-01
-9.38937664e-01 -4.77714777e-01 -7.74981737e-01 1.23844528e+00
-1.67523265e+00 -1.15407062e+00 -3.13312382e-01 3.82732123e-01
6.97506070e-01 2.08703637e-01 9.89041924e-01 1.87929571e-01
-2.29673430e-01 7.46253848e-01 -1.14303887e-01 3.06785166e-01
7.33266532e-01 -1.64109147e+00 1.02623856e+00 9.73895192e-01
4.14245427e-01 7.95971572e-01 5.73908329e-01 -4.96947348e-01
-1.12776887e+00 -1.18591833e+00 1.45086467e+00 -1.23025608e+00
4.53797579e-01 -2.76127994e-01 -1.27715743e+00 7.85623431e-01
7.55897999e-01 -4.88383800e-01 9.51636374e-01 1.64003536e-01
-3.68801862e-01 3.85336727e-01 -8.90161693e-01 6.82069659e-01
5.76424360e-01 -8.91016662e-01 -1.16740906e+00 4.18382794e-01
9.58218515e-01 -5.48693001e-01 -6.30398870e-01 4.73976314e-01
-7.53717788e-04 -6.05066776e-01 9.96011257e-01 -1.19215906e+00
7.52783656e-01 -2.87851423e-01 1.67654723e-01 -1.23300648e+00
-3.04250009e-02 -5.73432446e-01 -4.37587559e-01 1.45912910e+00
7.35441506e-01 -4.45791662e-01 6.00904942e-01 7.73008823e-01
-2.27950975e-01 -6.68685555e-01 -7.95001268e-01 -5.70563555e-01
6.62651896e-01 -3.11055809e-01 6.89051926e-01 7.20905840e-01
-3.03278595e-01 1.24977934e+00 -4.07129005e-02 -5.82555979e-02
1.31507248e-01 3.35102797e-01 9.61813629e-01 -9.48085666e-01
-2.56983638e-01 -3.25730920e-01 3.11706752e-01 -1.48923993e+00
2.15309098e-01 -9.46956635e-01 2.18145192e-01 -2.21345329e+00
-6.44947663e-02 -6.89110756e-02 2.26238653e-01 3.07690412e-01
-1.10347271e+00 5.74814491e-02 8.02088827e-02 -2.92570531e-01
-6.23922467e-01 7.55546629e-01 1.50041294e+00 -4.73249368e-02
5.16464189e-02 4.98793013e-02 -7.49183297e-01 4.74198282e-01
8.06585968e-01 -2.44580150e-01 -5.97482383e-01 -7.10295916e-01
5.20519197e-01 2.34836400e-01 3.40434402e-01 -6.36209369e-01
1.81489348e-01 6.32091686e-02 4.42128591e-02 -5.98283410e-01
2.06107914e-01 -1.15568593e-01 -6.43833637e-01 3.29123765e-01
-7.96045184e-01 2.30591133e-01 3.10159534e-01 5.16892910e-01
-4.60291624e-01 -8.41470242e-01 3.98005664e-01 -2.35482171e-01
-4.09309030e-01 -6.53233901e-02 -3.74819815e-01 8.72590303e-01
4.48209047e-01 5.19913360e-02 -7.62760162e-01 -1.00865066e+00
-4.12353516e-01 7.27082014e-01 -1.96678415e-01 5.40024996e-01
4.51882035e-01 -1.26997650e+00 -1.21722007e+00 -2.80633479e-01
2.71572500e-01 1.10453427e-01 2.42332593e-01 2.81354040e-01
-3.86942595e-01 8.97681653e-01 4.21606958e-01 -4.24497753e-01
-1.22208524e+00 3.19468260e-01 3.73720616e-01 -8.52670789e-01
1.36991113e-01 1.13495636e+00 1.61633447e-01 -1.22491550e+00
-5.48296981e-02 -5.18300891e-01 -4.58905071e-01 6.53601736e-02
6.26400054e-01 2.17329934e-01 2.43158564e-01 -2.18779787e-01
9.07960311e-02 1.92455202e-01 -5.71561977e-02 -3.12936157e-01
7.28688776e-01 -4.92828898e-02 2.14169431e-03 2.18731791e-01
8.65632296e-01 -1.23651348e-01 -7.82801926e-01 -3.30178857e-01
3.75546157e-01 -1.38671130e-01 -5.38900256e-01 -1.16871107e+00
-2.94450581e-01 1.21031106e+00 2.25889340e-01 2.86186934e-01
8.47631752e-01 1.68575004e-01 9.63767469e-01 8.10303271e-01
-5.43269387e-04 -6.37120605e-01 4.94222224e-01 1.00711000e+00
1.05763829e+00 -1.30091619e+00 -4.77441967e-01 -4.02179122e-01
-5.09782493e-01 8.46219897e-01 9.61073220e-01 1.48291409e-01
6.37089461e-02 -3.34724367e-01 3.18752497e-01 -1.32284895e-01
-1.00434101e+00 -4.26475018e-01 6.47993207e-01 5.74944973e-01
8.64345729e-01 -2.70700872e-01 -4.82998908e-01 4.84027803e-01
-6.65318072e-01 -5.47418706e-02 5.04406154e-01 8.68019462e-01
-4.07787651e-01 -1.55799723e+00 -4.48524922e-01 5.87067127e-01
-4.49801624e-01 -6.31443918e-01 -5.49643040e-01 5.49585223e-01
-1.95546374e-01 1.64663875e+00 -9.24925655e-02 4.76470888e-02
3.46294582e-01 7.92265594e-01 5.44126332e-01 -1.24541914e+00
-1.04751992e+00 -7.09681273e-01 6.61806822e-01 -2.02524617e-01
-2.36473083e-01 -4.45920765e-01 -1.10293639e+00 7.05933049e-02
-3.63971323e-01 5.79062879e-01 2.24453017e-01 1.20830083e+00
5.99851727e-01 5.41752517e-01 1.70692414e-01 2.03719646e-01
-7.99130380e-01 -1.39531851e+00 2.18701184e-01 5.65355957e-01
2.67763346e-01 2.40241345e-02 -3.67320441e-02 3.28505665e-01] | [11.462923049926758, 8.217885971069336] |
56414871-6d83-4829-959a-7ab0825b641f | an-augmented-linear-mixing-model-to-address | 1810.12000 | null | http://arxiv.org/abs/1810.12000v1 | http://arxiv.org/pdf/1810.12000v1.pdf | An Augmented Linear Mixing Model to Address Spectral Variability for Hyperspectral Unmixing | Hyperspectral imagery collected from airborne or satellite sources inevitably
suffers from spectral variability, making it difficult for spectral unmixing to
accurately estimate abundance maps. The classical unmixing model, the linear
mixing model (LMM), generally fails to handle this sticky issue effectively. To
this end, we propose a novel spectral mixture model, called the augmented
linear mixing model (ALMM), to address spectral variability by applying a
data-driven learning strategy in inverse problems of hyperspectral unmixing.
The proposed approach models the main spectral variability (i.e., scaling
factors) generated by variations in illumination or typography separately by
means of the endmember dictionary. It then models other spectral variabilities
caused by environmental conditions (e.g., local temperature and humidity,
atmospheric effects) and instrumental configurations (e.g., sensor noise), as
well as material nonlinear mixing effects, by introducing a spectral
variability dictionary. To effectively run the data-driven learning strategy,
we also propose a reasonable prior knowledge for the spectral variability
dictionary, whose atoms are assumed to be low-coherent with spectral signatures
of endmembers, which leads to a well-known low coherence dictionary learning
problem. Thus, a dictionary learning technique is embedded in the framework of
spectral unmixing so that the algorithm can learn the spectral variability
dictionary and estimate the abundance maps simultaneously. Extensive
experiments on synthetic and real datasets are performed to demonstrate the
superiority and effectiveness of the proposed method in comparison with
previous state-of-the-art methods. | ['Xiao Xiang Zhu', 'Naoto Yokoya', 'Danfeng Hong', 'Jocelyn Chanussot'] | 2018-10-29 | null | null | null | null | ['hyperspectral-unmixing'] | ['computer-vision'] | [ 5.41786075e-01 -8.02082062e-01 -3.37260999e-02 8.23360160e-02
-2.58390367e-01 -5.31823277e-01 5.85358739e-01 -1.30999461e-01
-8.76422673e-02 8.03385198e-01 2.89737713e-02 -8.81695375e-02
-2.72265643e-01 -8.46080244e-01 -4.47077751e-01 -1.36420763e+00
2.59790003e-01 2.69822359e-01 -3.48424971e-01 -2.43727699e-01
-1.59604698e-01 2.95560747e-01 -1.73118091e+00 -1.99165180e-01
1.33586133e+00 7.02719331e-01 5.37514448e-01 1.08878814e-01
-1.83370560e-01 3.86536777e-01 -3.44304979e-01 3.03079098e-01
4.01948184e-01 -5.75333953e-01 -7.79479295e-02 7.21675277e-01
3.19376558e-01 -2.07658052e-01 -2.20483214e-01 1.55090368e+00
2.70483226e-01 2.67150164e-01 7.72708833e-01 -9.62938905e-01
-3.39223206e-01 5.31758249e-01 -9.77764070e-01 -1.03224412e-01
-3.72874558e-01 8.50859731e-02 6.75933540e-01 -9.38464463e-01
1.38644472e-01 1.17792976e+00 5.20579457e-01 -4.99738902e-02
-1.46625495e+00 -7.12166548e-01 7.48360753e-02 4.04868871e-02
-1.70765388e+00 -3.20957571e-01 1.08793020e+00 -6.47856951e-01
2.38251582e-01 4.42066431e-01 7.32040823e-01 6.70927465e-01
-1.74679458e-01 3.81826371e-01 1.56862843e+00 -4.79259193e-01
8.67816508e-02 5.09147719e-02 1.44721732e-01 4.12824512e-01
7.52709806e-01 2.94952869e-01 -2.64833927e-01 -3.84595871e-01
6.84101939e-01 2.62449086e-01 -6.94470763e-01 -4.74721193e-01
-1.21398008e+00 7.21594810e-01 2.81050831e-01 1.62796885e-01
-6.60680711e-01 -3.92966926e-01 -9.46045443e-02 1.72152251e-01
6.40193284e-01 -5.21833077e-02 -2.69477665e-01 6.42839611e-01
-1.09831870e+00 3.31110694e-02 4.71993476e-01 7.21240222e-01
1.28300917e+00 5.85913002e-01 3.03981662e-01 1.16211236e+00
8.93266678e-01 1.31979108e+00 5.04285932e-01 -4.51274931e-01
1.55113444e-01 4.66538966e-01 4.16564554e-01 -1.12777078e+00
-2.85986602e-01 -7.89470136e-01 -1.35035884e+00 1.62745506e-01
1.67505831e-01 -1.38087809e-01 -9.26012695e-01 1.77994096e+00
4.77136970e-01 6.81167781e-01 3.32042992e-01 9.90082562e-01
7.58469522e-01 9.19051945e-01 1.03926891e-03 -9.91827011e-01
1.05083060e+00 -6.73032582e-01 -1.02348304e+00 -3.36110950e-01
1.36154711e-01 -8.62364113e-01 6.95962787e-01 3.62682253e-01
-5.24860322e-01 -5.74102342e-01 -1.23965156e+00 5.00967264e-01
-2.26719648e-01 1.72988519e-01 5.49391329e-01 5.86529195e-01
-4.24504131e-01 2.86888272e-01 -8.63702178e-01 -1.33734345e-01
-8.10713843e-02 -6.27371622e-03 -9.49240252e-02 -1.24371365e-01
-1.18401086e+00 7.66121805e-01 6.29468620e-01 4.39028352e-01
-9.09307897e-01 -4.95756358e-01 -7.35462129e-01 -2.02291682e-01
2.72970557e-01 -6.23682380e-01 5.01944780e-01 -1.25130725e+00
-1.33214378e+00 5.10210097e-01 -2.24676624e-01 4.49362444e-03
6.09090589e-02 -1.70795601e-02 -8.13699722e-01 -2.36050580e-02
-1.00919671e-01 -1.12298891e-01 1.27844834e+00 -1.72879004e+00
-2.79765040e-01 -5.05777061e-01 -4.43391144e-01 3.42128456e-01
-4.56780910e-01 -3.03390205e-01 4.63333577e-02 -8.74911249e-01
6.06706500e-01 -1.02794957e+00 -2.65202224e-01 -3.41740936e-01
-1.06379978e-01 5.24591267e-01 8.41114581e-01 -8.45950127e-01
1.28798807e+00 -2.36872387e+00 4.52088296e-01 5.52849531e-01
-3.74185294e-02 3.44833970e-01 -2.99931645e-01 3.52184445e-01
-2.60834873e-01 -2.38165274e-01 -9.88803804e-01 -1.17987350e-01
-3.28198284e-01 4.37724382e-01 -2.81109840e-01 7.97772288e-01
-9.16651264e-02 2.62809664e-01 -9.55646574e-01 -4.04763930e-02
5.10788321e-01 6.54555500e-01 7.31183589e-02 2.27177709e-01
-2.85711795e-01 7.40752995e-01 -1.18829578e-01 7.49742150e-01
1.25012887e+00 -8.34644362e-02 2.52759159e-01 -6.02859557e-01
-3.69902074e-01 -2.78125942e-01 -1.85812104e+00 1.54921997e+00
-3.13935876e-01 1.16593030e-03 6.66127205e-01 -1.09998643e+00
9.55685496e-01 2.87990451e-01 5.02070725e-01 1.98547705e-03
4.48943004e-02 5.68018496e-01 1.44468591e-01 -4.00144637e-01
2.15719178e-01 -5.86124241e-01 5.44088900e-01 1.97501391e-01
5.49319386e-02 -1.58676118e-01 -5.82942963e-02 -2.89361000e-01
1.82554871e-01 1.05492219e-01 5.44453621e-01 -6.61924958e-01
9.38645601e-01 1.16164990e-01 5.96169829e-01 3.94602507e-01
1.75840929e-01 2.18058199e-01 -3.40497881e-01 -2.38252968e-01
-9.15278435e-01 -8.96625459e-01 -5.37445545e-01 6.92961276e-01
3.67160380e-01 -2.21340433e-02 -4.18314755e-01 5.46196960e-02
2.33755279e-02 4.55411941e-01 -2.05163017e-01 -1.84297666e-01
-7.53194913e-02 -1.77115738e+00 1.98243469e-01 -5.27214557e-02
7.13510513e-01 -4.88024384e-01 -5.64037785e-02 4.03959513e-01
-4.37145084e-01 -7.03220129e-01 -8.47023204e-02 1.55854881e-01
-1.01189911e+00 -9.82582390e-01 -6.13666356e-01 -3.88428926e-01
5.08072793e-01 7.30715394e-01 6.57771051e-01 -1.66031405e-01
-4.58735898e-02 1.84936821e-01 -4.34021384e-01 -4.33353722e-01
-4.82248724e-01 -4.45526809e-01 3.98099244e-01 7.28657424e-01
3.07481706e-01 -8.25792134e-01 -4.34846997e-01 3.44855487e-01
-1.34865892e+00 -8.32387246e-03 7.44606912e-01 1.04950452e+00
8.66397440e-01 7.73735642e-01 3.23706299e-01 -7.01605260e-01
3.07906717e-01 -6.45965636e-01 -7.67610013e-01 3.26439947e-01
-6.43235147e-01 -2.21635848e-01 4.04747963e-01 -6.36256158e-01
-1.27298069e+00 1.68748602e-01 3.00727338e-01 -4.89727825e-01
-2.61526316e-01 1.04284739e+00 -6.54682696e-01 -3.52350593e-01
6.55171454e-01 7.33630478e-01 2.17617303e-01 -6.13624692e-01
3.72373223e-01 8.67053449e-01 5.93605816e-01 -5.40719450e-01
1.28156805e+00 7.32754350e-01 1.75968200e-01 -1.44324911e+00
-7.13800490e-01 -6.96729600e-01 -5.54246604e-01 -6.37113005e-02
5.12326956e-01 -1.44854653e+00 8.29602256e-02 9.49518681e-01
-7.51762867e-01 -5.21582924e-02 6.08848408e-02 7.94357777e-01
-1.82092547e-01 7.30118155e-01 -3.04087788e-01 -1.03446555e+00
-2.51675248e-01 -9.95861411e-01 6.95120931e-01 1.62620619e-01
4.01015878e-01 -1.10425019e+00 1.09674320e-01 2.42364541e-01
3.74712348e-01 5.36296308e-01 8.58887732e-01 -1.38744846e-01
-4.44326341e-01 1.89655200e-01 -1.41552627e-01 7.39534259e-01
6.33855939e-01 5.80677129e-02 -8.42277348e-01 -4.63454813e-01
5.17839789e-01 8.50414336e-02 9.93920982e-01 5.01478434e-01
7.31989026e-01 -2.39639625e-01 -1.22697607e-01 8.50071967e-01
1.81619477e+00 9.05215293e-02 5.04433155e-01 1.81691185e-01
1.00391591e+00 5.14168262e-01 6.04577482e-01 7.07815826e-01
-6.01387508e-02 3.14041495e-01 5.44188201e-01 -2.61937588e-01
2.32843727e-01 1.01730190e-01 2.28977993e-01 1.11662996e+00
-2.46859461e-01 -2.16538176e-01 -7.64161348e-01 2.39715576e-01
-1.88780630e+00 -9.37400818e-01 -7.34661758e-01 2.40129781e+00
7.31605649e-01 -5.06584883e-01 -1.98701486e-01 3.11770767e-01
8.46542954e-01 5.96073508e-01 -5.14483571e-01 4.46618110e-01
-7.69284427e-01 5.89330345e-02 7.61410773e-01 5.86606920e-01
-1.29461908e+00 6.64196491e-01 5.47191048e+00 8.23865950e-01
-1.18742609e+00 2.27108151e-01 7.52480254e-02 3.61199588e-01
-5.24545074e-01 6.20610975e-02 -4.85208988e-01 5.24462044e-01
5.40643096e-01 -4.07550149e-02 9.85712111e-01 3.77487391e-01
4.47285235e-01 -7.12832436e-04 -3.55713993e-01 1.26320386e+00
2.14936689e-01 -7.25959897e-01 2.69149274e-01 7.14595020e-02
1.06021154e+00 2.85136290e-02 -2.59982087e-02 -6.01283908e-02
5.70317879e-02 -6.12265289e-01 5.20146668e-01 7.23157704e-01
7.95355797e-01 -5.81162930e-01 6.54587626e-01 6.44963503e-01
-1.29063499e+00 -9.35631543e-02 -6.18391633e-01 -1.53464288e-01
9.14663672e-02 1.21484458e+00 -9.53434967e-03 8.55737746e-01
2.83679903e-01 1.02696800e+00 -1.83183700e-01 9.27871585e-01
-3.61462355e-01 8.15529644e-01 -5.26506484e-01 5.97594500e-01
6.07444607e-02 -1.00423360e+00 7.86445796e-01 9.63849843e-01
6.61657393e-01 3.83477360e-01 3.15268070e-01 9.19279993e-01
2.33654499e-01 3.82152349e-01 -4.26710337e-01 -3.28119457e-01
5.26737928e-01 1.41375470e+00 -4.19169396e-01 -3.82213473e-01
-3.89835447e-01 6.89367890e-01 -4.26907629e-01 7.13028729e-01
-5.27228236e-01 2.46335685e-01 7.17626810e-01 -1.12111695e-01
1.40897959e-01 -3.91525418e-01 -7.16216117e-02 -1.60100234e+00
-1.44203007e-01 -1.24436355e+00 2.63649583e-01 -7.02180326e-01
-1.41564834e+00 2.85511583e-01 7.34982640e-02 -1.64723361e+00
3.50652546e-01 -6.85798883e-01 -4.08194989e-01 1.27786922e+00
-1.87819302e+00 -1.25674140e+00 -7.03696251e-01 7.40146399e-01
1.36425197e-01 -2.34051093e-01 9.36179638e-01 3.24517667e-01
-6.88204885e-01 -2.66958684e-01 8.23874533e-01 -3.82692516e-01
7.25224137e-01 -9.51087773e-01 -5.51803589e-01 1.03971589e+00
1.01397887e-01 7.27652907e-01 8.09766114e-01 -7.03991950e-01
-1.47738230e+00 -1.35478485e+00 2.25127950e-01 1.85141578e-01
9.10452306e-01 6.13706931e-02 -1.12958372e+00 4.38851655e-01
2.21097842e-02 1.26767922e-02 1.13377380e+00 -1.91032514e-01
-3.28135192e-01 -3.84790182e-01 -9.22783673e-01 2.27595016e-01
5.86815476e-01 -4.67431843e-01 -5.48897743e-01 5.28679192e-01
3.28779966e-01 -1.38092235e-01 -8.94825697e-01 7.77727485e-01
3.74472827e-01 -5.66123247e-01 1.04306448e+00 -1.74177662e-01
8.59676227e-02 -9.13799524e-01 -5.39742231e-01 -1.58921278e+00
-6.95674956e-01 -2.82032222e-01 -2.62421429e-01 1.23672664e+00
1.03755631e-01 -7.75582373e-01 8.15499127e-02 1.38299400e-02
-3.68191674e-02 4.95843813e-02 -6.71733201e-01 -8.47682118e-01
-1.57954469e-01 -4.01719399e-02 7.18941748e-01 1.49858594e+00
-4.39734161e-01 3.13669235e-01 -8.39664400e-01 8.20771158e-01
1.08874190e+00 3.08220208e-01 5.70902169e-01 -1.62737763e+00
-4.10777539e-01 -3.71596903e-01 6.96519539e-02 -7.68775165e-01
2.58804291e-01 -6.58751070e-01 7.61395246e-02 -1.38478208e+00
2.18315378e-01 -4.11000311e-01 -5.23045003e-01 1.25634879e-01
-5.96631289e-01 1.57171950e-01 -1.37964085e-01 7.30713129e-01
1.82567373e-01 6.84852779e-01 1.06106949e+00 -5.74780226e-01
-3.38415712e-01 -1.67767078e-01 -6.01326942e-01 7.68545210e-01
7.54816115e-01 -3.00189823e-01 -5.50101757e-01 -4.74505424e-01
4.64573577e-02 9.28539038e-02 3.22597742e-01 -1.01125050e+00
-1.82433829e-01 -5.94738066e-01 9.97866243e-02 -5.01741409e-01
2.13364631e-01 -1.18027997e+00 8.08848023e-01 3.47473711e-01
1.65090576e-01 -5.87866127e-01 2.34283790e-01 6.97589636e-01
-4.35327202e-01 -2.58749217e-01 8.89465511e-01 -1.51501253e-01
-7.18869269e-01 4.44729537e-01 -3.22229803e-01 -5.34863532e-01
6.39990449e-01 1.90680809e-02 -2.06038877e-01 -4.17201594e-02
-6.98813021e-01 6.90385029e-02 5.73127031e-01 4.83236425e-02
3.03844184e-01 -1.38831162e+00 -1.02759874e+00 3.99292499e-01
2.91575462e-01 -4.98126075e-02 6.17020726e-01 9.79077399e-01
-4.33798015e-01 -3.60997627e-03 -7.89359957e-02 -5.82084477e-01
-1.09246564e+00 7.54371524e-01 5.16193509e-01 -1.73682941e-03
-5.37883900e-02 6.16231740e-01 2.37377495e-01 -5.29251575e-01
-2.60721862e-01 -7.56708011e-02 -3.36424381e-01 3.61646235e-01
6.58040047e-01 3.90053213e-01 -6.64941370e-02 -1.11465776e+00
-1.46015853e-01 7.14665294e-01 7.23331690e-01 5.03506400e-02
1.13808012e+00 -3.83952141e-01 -8.11844170e-01 7.10655928e-01
8.81602108e-01 1.64500371e-01 -9.18227971e-01 -7.04418123e-01
-4.07422930e-01 -4.85061556e-01 4.55466121e-01 -6.28379524e-01
-9.25142169e-01 8.08128715e-01 8.18122268e-01 1.72018915e-01
1.44925070e+00 -6.57154202e-01 3.43968898e-01 2.80803233e-01
3.21868658e-01 -8.23539138e-01 -4.60920364e-01 2.48790666e-01
7.34078228e-01 -1.39490485e+00 2.91437089e-01 -5.92817426e-01
-2.36903697e-01 1.01662171e+00 3.68202120e-01 1.10050842e-01
7.97825277e-01 3.75175551e-02 2.71425366e-01 9.49333608e-02
-4.14022766e-02 -4.64300424e-01 3.60544562e-01 4.71145928e-01
2.36230209e-01 4.74959075e-01 -4.34269786e-01 3.81938487e-01
1.98514894e-01 -2.41923735e-01 3.01374435e-01 4.57368135e-01
-7.36033380e-01 -1.07061136e+00 -1.10273886e+00 4.23677176e-01
2.69330014e-03 -2.71707028e-01 -8.50629434e-02 3.16517681e-01
6.01643264e-01 1.18522048e+00 -2.52847731e-01 -3.55983675e-01
8.89119655e-02 1.01682983e-01 1.79714024e-01 -5.90257406e-01
1.61018997e-01 8.49386454e-01 -3.53659600e-01 -9.97155905e-03
-1.22983813e+00 -7.62110353e-01 -7.88540006e-01 -1.21545151e-01
-6.80813372e-01 2.13314325e-01 5.46943367e-01 9.07005131e-01
-1.91857368e-01 3.99622411e-01 7.19861805e-01 -9.48807776e-01
-6.00057602e-01 -1.21418858e+00 -1.28445375e+00 3.66101235e-01
3.87922227e-01 -8.82487118e-01 -7.83094704e-01 1.93714619e-01] | [10.105111122131348, -2.07995867729187] |
59acbfd7-a38b-4aa8-b419-1e28676eb2b7 | deep-shape-matching | 1709.03409 | null | http://arxiv.org/abs/1709.03409v2 | http://arxiv.org/pdf/1709.03409v2.pdf | Deep Shape Matching | We cast shape matching as metric learning with convolutional networks. We
break the end-to-end process of image representation into two parts. Firstly,
well established efficient methods are chosen to turn the images into edge
maps. Secondly, the network is trained with edge maps of landmark images, which
are automatically obtained by a structure-from-motion pipeline. The learned
representation is evaluated on a range of different tasks, providing
improvements on challenging cases of domain generalization, generic
sketch-based image retrieval or its fine-grained counterpart. In contrast to
other methods that learn a different model per task, object category, or
domain, we use the same network throughout all our experiments, achieving
state-of-the-art results in multiple benchmarks. | ['Ondřej Chum', 'Filip Radenović', 'Giorgos Tolias'] | 2017-09-11 | deep-shape-matching-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Filip_Radenovic_Deep_Shape_Matching_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Filip_Radenovic_Deep_Shape_Matching_ECCV_2018_paper.pdf | eccv-2018-9 | ['sketch-based-image-retrieval'] | ['computer-vision'] | [ 1.65093198e-01 -3.47741604e-01 -4.04147804e-01 -4.64002669e-01
-9.76390839e-01 -7.99658000e-01 1.06624460e+00 8.85644332e-02
-5.99131107e-01 2.20445663e-01 2.40083069e-01 1.14026949e-01
-1.82315022e-01 -6.66437745e-01 -7.49578476e-01 -3.37572008e-01
1.39180496e-01 7.36928701e-01 3.35977525e-01 -2.72952616e-01
5.15424371e-01 7.73434460e-01 -1.31881690e+00 5.45425415e-01
3.62231672e-01 1.47937632e+00 -7.33408779e-02 4.57331061e-01
-2.03952819e-01 5.50857365e-01 -1.12287894e-01 -6.02252781e-01
4.49149460e-01 -3.51613923e-03 -9.15429473e-01 1.40782952e-01
9.26692963e-01 -4.70805496e-01 -5.91082275e-01 9.53639805e-01
3.94615889e-01 4.35305238e-02 1.02322388e+00 -1.15567267e+00
-8.10912907e-01 4.23256494e-02 -5.13880551e-01 -1.99632689e-01
2.44706467e-01 8.25088751e-03 1.13367343e+00 -1.43197441e+00
1.01111996e+00 1.48942459e+00 7.11257815e-01 2.92481601e-01
-1.54176295e+00 -4.53632921e-01 1.32582948e-01 2.07462072e-01
-1.47707140e+00 -5.13525248e-01 9.08326864e-01 -4.80066866e-01
9.54395115e-01 -1.18472844e-01 2.19461441e-01 1.14740586e+00
1.43078342e-02 9.42127109e-01 7.60909915e-01 -2.71545410e-01
2.16731459e-01 -2.21951306e-01 -6.79330900e-02 6.02626204e-01
1.04350761e-01 2.02151120e-01 -5.23106456e-01 -1.65019423e-01
8.96354258e-01 1.52814984e-01 6.43712729e-02 -1.11100900e+00
-1.25277054e+00 7.20813513e-01 7.67385721e-01 2.29415178e-01
-3.57149541e-01 4.94411170e-01 5.91128886e-01 4.45327103e-01
5.48945069e-01 2.12559998e-01 -3.09047163e-01 1.50658324e-01
-1.10730648e+00 4.72290248e-01 8.01905394e-01 9.82895911e-01
8.90970588e-01 -3.43668580e-01 -3.77774835e-01 1.00916791e+00
9.48074833e-02 3.80327940e-01 2.77992666e-01 -1.01322722e+00
4.96465683e-01 6.46282852e-01 1.18169270e-01 -1.01899517e+00
-1.98996276e-01 -2.42941365e-01 -7.08192587e-01 4.01235014e-01
5.78784168e-01 4.84324127e-01 -1.16878593e+00 1.53801370e+00
-4.07067360e-03 1.11427382e-01 -3.06031883e-01 8.32655787e-01
8.51648033e-01 1.63563862e-01 1.70076698e-01 7.84248531e-01
1.21382082e+00 -1.29649699e+00 -1.30480632e-01 -3.20655316e-01
3.12236995e-01 -7.56692290e-01 9.45131660e-01 3.34717631e-01
-1.02642012e+00 -7.02031314e-01 -8.89543176e-01 -5.22455573e-01
-7.13255703e-01 2.17932373e-01 3.71524274e-01 1.53818488e-01
-1.14150536e+00 8.77746403e-01 -4.71681178e-01 -3.81240755e-01
8.73929918e-01 8.99554193e-02 -7.21311212e-01 -3.52693260e-01
-7.48903453e-01 7.54991412e-01 2.22291276e-01 -1.76039457e-01
-9.91899669e-01 -8.59631538e-01 -9.77359474e-01 8.15680772e-02
4.63910364e-02 -8.12054515e-01 1.24931335e+00 -1.01069212e+00
-1.24683261e+00 1.49808097e+00 -7.30939433e-02 -3.32951784e-01
8.02151382e-01 -2.18886793e-01 5.30302525e-02 2.20548421e-01
1.09689325e-01 1.01574135e+00 1.22336876e+00 -1.15813303e+00
-6.00942016e-01 -2.48753026e-01 8.40980485e-02 -1.44578665e-01
-2.62811959e-01 -3.06739751e-02 -7.20525384e-01 -9.99367774e-01
2.68846136e-02 -8.98485541e-01 -4.54118475e-02 9.39789534e-01
-5.91565073e-02 -4.46773678e-01 9.00326073e-01 -5.59521079e-01
6.80095792e-01 -2.18497539e+00 3.78098786e-01 1.78118184e-01
1.85060412e-01 1.53327093e-01 -6.30745947e-01 4.73721474e-01
-2.06021503e-01 1.52953416e-02 -4.70672518e-01 -7.91787148e-01
3.68914634e-01 5.36666289e-02 -5.43394268e-01 5.00376105e-01
4.94463742e-01 1.36695397e+00 -9.05919075e-01 -3.12943369e-01
2.68305004e-01 5.20399034e-01 -4.74386364e-01 2.89683789e-01
-2.43788794e-01 2.34607488e-01 -3.25117916e-01 7.38578796e-01
7.78609812e-01 -3.11329007e-01 -9.82449800e-02 -2.66825229e-01
2.18327165e-01 2.77560264e-01 -1.13906670e+00 2.49260640e+00
-5.10692775e-01 7.00367093e-01 -2.73789391e-02 -1.14660978e+00
9.49971199e-01 8.02189782e-02 4.43259150e-01 -9.33725536e-01
-1.47172034e-01 3.71119976e-01 -3.90796214e-01 -8.73144642e-02
3.99550468e-01 1.69910997e-01 -9.27733481e-02 5.33644795e-01
2.36707792e-01 -1.64019555e-01 1.27709955e-01 1.76707506e-01
1.08280218e+00 6.44704282e-01 1.18153132e-01 -3.75212759e-01
4.89352554e-01 -5.48047200e-02 1.53873116e-01 6.67513490e-01
-1.09623574e-01 1.15590525e+00 4.08401966e-01 -8.16313446e-01
-1.25164795e+00 -1.33543062e+00 -1.92282632e-01 1.39087737e+00
3.24724466e-01 -3.33335400e-01 -4.85106170e-01 -8.70979488e-01
4.94129658e-01 1.71751976e-01 -8.65048051e-01 -1.11721672e-01
-6.01474643e-01 -4.71362239e-03 4.81390506e-01 8.48292351e-01
5.51636398e-01 -1.29179037e+00 -4.78274018e-01 2.72324234e-01
1.27094045e-01 -1.18786335e+00 -7.36025453e-01 -1.70825999e-02
-7.30515718e-01 -1.14612412e+00 -1.03454423e+00 -9.18076575e-01
5.69792986e-01 4.68810529e-01 1.60336864e+00 1.44804448e-01
-5.09525299e-01 6.49536073e-01 -2.03198895e-01 -9.74927619e-02
-8.11300054e-03 1.69925332e-01 -4.05889243e-01 1.47759438e-01
3.08600903e-01 -5.49360096e-01 -9.60475206e-01 3.72874737e-01
-9.90187287e-01 -3.15749824e-01 6.92935944e-01 9.97683465e-01
6.06449544e-01 -5.88129818e-01 4.17162985e-01 -6.65761471e-01
4.12326694e-01 -1.68804765e-01 -6.42375350e-01 4.24466193e-01
-3.61615658e-01 4.16953564e-01 4.18114960e-01 -3.72232735e-01
-7.41782546e-01 2.57948905e-01 -1.85159855e-02 -7.23461986e-01
-2.96693057e-01 1.48462221e-01 3.59355994e-02 -3.70380133e-01
6.66798353e-01 2.34208032e-01 7.31296316e-02 -7.60675669e-01
6.84008121e-01 4.53410119e-01 6.43075287e-01 -8.09988797e-01
1.03182554e+00 7.49301732e-01 -8.52117129e-03 -5.56250513e-01
-7.34332204e-01 -5.78881264e-01 -9.32455719e-01 2.06534564e-01
7.33178735e-01 -8.80051196e-01 -4.65830117e-01 5.25409520e-01
-1.36719763e+00 -5.72814524e-01 -1.33200496e-01 1.50225069e-02
-6.71958148e-01 4.06198978e-01 -4.34834033e-01 -2.42334113e-01
-4.91183311e-01 -9.93236661e-01 1.75886393e+00 -2.43414957e-02
-1.74724892e-01 -1.05884242e+00 1.24777123e-01 -9.39729959e-02
6.53131127e-01 3.44922751e-01 9.70921695e-01 -5.68895459e-01
-6.69216454e-01 -2.58753538e-01 -7.36957669e-01 2.87092239e-01
4.22872193e-02 -2.03596696e-01 -1.06890869e+00 -5.06960571e-01
-4.57956254e-01 -6.18482590e-01 1.31750369e+00 4.23874408e-02
1.37425470e+00 1.60979219e-02 -5.14523864e-01 9.05596673e-01
1.43871498e+00 -2.38329858e-01 7.20280945e-01 4.91445273e-01
5.93745947e-01 5.97328484e-01 4.79699939e-01 1.71482608e-01
2.99403101e-01 9.35057640e-01 4.96386856e-01 -3.16538215e-01
-5.62717855e-01 -4.27890122e-01 2.82272287e-02 -5.61393276e-02
2.80909628e-01 6.73882365e-02 -7.96354890e-01 8.62802327e-01
-1.95865512e+00 -7.82112539e-01 4.12806988e-01 2.24191499e+00
5.76593280e-01 1.09589174e-01 2.65216351e-01 -1.52140021e-01
5.24432480e-01 4.38326955e-01 -6.10089004e-01 -1.53768614e-01
-4.92831096e-02 4.87590164e-01 4.18009460e-01 2.95790136e-01
-1.32542002e+00 1.05943799e+00 6.28637934e+00 8.19388509e-01
-1.36123526e+00 -1.46781951e-01 6.37650788e-01 5.41271530e-02
-2.75693983e-01 -4.85316804e-03 -4.79763657e-01 2.32153937e-01
2.03404456e-01 1.01509802e-01 5.33892334e-01 8.59936178e-01
-4.22676057e-01 2.11272240e-01 -1.68128324e+00 1.26621985e+00
-1.36771098e-01 -1.50296164e+00 2.13739350e-01 -1.07817866e-01
7.37146497e-01 1.82447955e-01 1.15733460e-01 2.31152400e-01
3.02277327e-01 -1.17986214e+00 9.53660905e-01 5.37212193e-01
1.20478904e+00 -5.76296031e-01 2.09056586e-01 -3.35320346e-02
-1.42203033e+00 2.39946619e-02 -6.89236939e-01 1.61262721e-01
-1.42920166e-01 3.12607229e-01 -3.84457052e-01 5.84221900e-01
7.47301280e-01 1.00302565e+00 -7.39077985e-01 1.14548612e+00
-1.30217314e-01 9.74746943e-02 -2.66585499e-01 4.00683910e-01
3.88953924e-01 -1.13225564e-01 3.85182023e-01 1.51532543e+00
4.86984197e-03 -3.02493542e-01 1.04234770e-01 1.05295563e+00
-4.09079432e-01 -2.19211295e-01 -9.01226759e-01 9.88736153e-02
5.10938883e-01 1.44249046e+00 -5.94190776e-01 -3.68691355e-01
-4.22711968e-01 1.14772463e+00 6.30173683e-01 5.85775971e-01
-5.71103454e-01 -6.30357981e-01 7.74525285e-01 2.91322500e-01
7.95788765e-01 -2.43550867e-01 -2.56362528e-01 -1.00091517e+00
2.25051567e-01 -8.81183088e-01 4.75185037e-01 -5.73699117e-01
-1.57420373e+00 4.79144335e-01 -3.02292407e-02 -1.20776868e+00
-3.72908831e-01 -9.13402855e-01 -7.68189013e-01 9.17671621e-01
-1.83290184e+00 -1.45513093e+00 -4.94796485e-01 4.99444038e-01
6.01911783e-01 -2.32844993e-01 8.24736774e-01 3.88700992e-01
-6.08653203e-02 7.93976128e-01 1.48265630e-01 5.46485662e-01
1.05690575e+00 -1.16779017e+00 9.37000811e-01 5.18726945e-01
4.90454763e-01 5.56521773e-01 9.43138301e-02 -2.90019155e-01
-1.43022549e+00 -1.11336470e+00 7.05642819e-01 -5.62913597e-01
5.67956746e-01 -6.73043370e-01 -8.92666161e-01 4.82006073e-01
1.29571974e-01 7.48435020e-01 4.33014371e-02 -5.25271967e-02
-1.06376815e+00 -2.23769933e-01 -1.09059787e+00 3.85765940e-01
1.40775776e+00 -9.06085670e-01 -5.85980415e-01 2.56061465e-01
4.18735862e-01 -3.87867004e-01 -8.06172431e-01 2.95298547e-01
9.63552952e-01 -8.54151368e-01 1.43974233e+00 -9.13041830e-01
5.78438759e-01 -1.91352203e-01 -3.33544552e-01 -1.20047545e+00
-5.68584919e-01 -4.42254394e-01 9.36534330e-02 9.55104649e-01
1.63761944e-01 -4.70775694e-01 7.88467944e-01 4.14812446e-01
1.35604948e-01 -8.08579922e-01 -8.52005005e-01 -1.01171899e+00
3.97998899e-01 -1.78503186e-01 5.73933661e-01 7.19558835e-01
-5.10168672e-01 2.17169836e-01 -8.79105479e-02 -3.96978319e-01
9.09816563e-01 5.01242042e-01 9.42150652e-01 -1.48982322e+00
-9.64103267e-02 -9.44294274e-01 -4.71911103e-01 -1.28332746e+00
3.91501129e-01 -1.09347367e+00 -7.26480484e-02 -1.54579020e+00
1.85781479e-01 -3.53836000e-01 -4.40285355e-01 5.38779199e-01
6.94329647e-05 6.23307467e-01 3.60667050e-01 2.81377435e-01
-8.47812057e-01 5.68480492e-01 1.05667210e+00 -5.01325786e-01
1.65795997e-01 -2.91646242e-01 -4.35212880e-01 5.12420058e-01
3.96300286e-01 -5.64740539e-01 -2.34445006e-01 -7.12457180e-01
3.64700034e-02 -1.78947031e-01 7.98958123e-01 -8.05375099e-01
1.40848964e-01 -4.94578015e-03 5.89037895e-01 -5.20201921e-01
3.31431597e-01 -9.24160302e-01 -1.06314659e-01 4.29089069e-01
-4.80387807e-01 -2.73366515e-02 2.74662763e-01 5.65382600e-01
-3.23093921e-01 -1.67411432e-01 8.58674049e-01 3.66167426e-02
-1.02305818e+00 5.95274568e-01 3.59453410e-01 2.12050840e-01
6.35566115e-01 -2.97451943e-01 -2.34900758e-01 -4.10526276e-01
-4.49699759e-01 4.12389413e-02 6.78394198e-01 8.12669277e-01
6.29904747e-01 -1.75813091e+00 -8.11676323e-01 1.65348768e-01
5.55254579e-01 4.46176454e-02 -1.22406498e-01 5.34578919e-01
-4.64596510e-01 4.61644471e-01 -4.55787748e-01 -8.12013447e-01
-8.46408010e-01 7.08612502e-01 4.55972016e-01 -3.45786422e-01
-6.46197319e-01 6.51553929e-01 6.54946685e-01 -5.65202475e-01
4.04562414e-01 -1.37424409e-01 2.29919180e-01 3.33653204e-02
4.80685562e-01 2.91366845e-01 2.14003518e-01 -5.12015760e-01
-5.66136897e-01 9.80236590e-01 -1.69023424e-01 -1.01831868e-01
1.39643145e+00 1.95932001e-01 -8.65749433e-04 -1.10316603e-03
1.64399600e+00 -3.47198457e-01 -1.65845644e+00 -6.06447160e-01
2.68427759e-01 -5.98234653e-01 -1.80694997e-01 -7.60717571e-01
-1.07352793e+00 1.04833376e+00 5.84802568e-01 -2.75869578e-01
8.69875729e-01 -2.67028287e-02 7.82880962e-01 5.22529066e-01
4.75723058e-01 -9.61537421e-01 4.13818300e-01 6.58341765e-01
1.36536670e+00 -1.52432203e+00 -4.38879430e-02 1.34240631e-02
-3.62490714e-01 1.20011866e+00 2.53315449e-01 -6.73134327e-01
7.04119086e-01 4.28083306e-03 -1.31772850e-02 -2.26055071e-01
-5.30459940e-01 -2.43462890e-01 8.55266869e-01 6.77230299e-01
3.86840135e-01 -2.54766941e-01 -5.28226607e-02 2.17255637e-01
4.17359799e-01 1.08245283e-01 -2.00131938e-01 8.80813420e-01
-2.92962939e-01 -1.37729621e+00 -8.71910304e-02 1.81520402e-01
-1.74312234e-01 1.27112577e-02 -7.22349942e-01 8.78496945e-01
-3.04706097e-01 5.14832854e-01 1.84891164e-01 -6.83467165e-02
5.34803569e-01 1.35102764e-01 5.66103280e-01 -3.71504813e-01
-5.11788845e-01 -2.58850187e-01 -1.86263546e-01 -9.93390083e-01
-1.27942473e-01 -3.39483947e-01 -9.43054080e-01 -1.72627300e-01
2.39702895e-01 -3.54258537e-01 6.30619764e-01 7.98527181e-01
5.84291816e-01 1.37938663e-01 5.03211856e-01 -1.29112005e+00
-7.26078629e-01 -8.84764373e-01 -2.70970821e-01 1.03820610e+00
3.29104573e-01 -9.15469706e-01 -2.52667465e-03 -6.03873990e-02] | [11.601428985595703, 0.5544795989990234] |
a0050d3d-93a5-4cda-9555-b1cd26e2eb96 | let-the-chart-spark-embedding-semantic | 2304.14630 | null | https://arxiv.org/abs/2304.14630v2 | https://arxiv.org/pdf/2304.14630v2.pdf | Let the Chart Spark: Embedding Semantic Context into Chart with Text-to-Image Generative Model | Pictorial visualization seamlessly integrates data and semantic context into visual representation, conveying complex information in a manner that is both engaging and informative. Extensive studies have been devoted to developing authoring tools to simplify the creation of pictorial visualizations. However, mainstream works mostly follow a retrieving-and-editing pipeline that heavily relies on retrieved visual elements from a dedicated corpus, which often compromise the data integrity. Text-guided generation methods are emerging, but may have limited applicability due to its predefined recognized entities. In this work, we propose ChartSpark, a novel system that embeds semantic context into chart based on text-to-image generative model. ChartSpark generates pictorial visualizations conditioned on both semantic context conveyed in textual inputs and data information embedded in plain charts. The method is generic for both foreground and background pictorial generation, satisfying the design practices identified from an empirical research into existing pictorial visualizations. We further develop an interactive visual interface that integrates a text analyzer, editing module, and evaluation module to enable users to generate, modify, and assess pictorial visualizations. We experimentally demonstrate the usability of our tool, and conclude with a discussion of the potential of using text-to-image generative model combined with interactive interface for visualization design. | ['Wei Zeng', 'Yilin Ye', 'Yue Lin', 'Suizi Huang', 'Shishi Xiao'] | 2023-04-28 | null | null | null | null | ['text-guided-generation'] | ['computer-vision'] | [ 2.02991202e-01 -3.01560443e-02 2.69121081e-01 -3.31516981e-01
-1.26587972e-01 -7.36562252e-01 1.05888402e+00 4.52741861e-01
6.48527453e-03 4.69483376e-01 6.41179740e-01 -6.31698668e-01
-1.93356292e-03 -8.27298522e-01 -2.29578450e-01 -2.72691876e-01
3.23862791e-01 2.71116998e-02 1.38037741e-01 -2.08339840e-01
7.03384995e-01 4.71649796e-01 -1.87029624e+00 7.94963181e-01
1.03109515e+00 5.64986587e-01 5.85828841e-01 5.13782382e-01
-1.15005267e+00 5.56846380e-01 -1.24258304e+00 -2.44945556e-01
-9.67085883e-02 -6.54797733e-01 -2.62719184e-01 -5.12611009e-02
2.51136810e-01 6.72654212e-02 4.07580912e-01 8.40713263e-01
4.04788435e-01 -2.19969302e-01 5.76971173e-01 -1.34722924e+00
-1.14840150e+00 4.02565032e-01 -6.23693764e-01 -1.32591516e-01
7.89368868e-01 2.24579304e-01 6.05281293e-01 -1.02225721e+00
9.69932079e-01 1.50397909e+00 2.80989319e-01 2.84354657e-01
-1.18510652e+00 -6.25630856e-01 2.81176299e-01 -5.74859157e-02
-9.95826960e-01 7.20075965e-02 9.27135110e-01 -7.58636892e-01
7.00989485e-01 8.50145161e-01 1.24943745e+00 1.09205258e+00
-2.32736710e-02 6.51664555e-01 1.26274014e+00 -7.87488699e-01
1.71679154e-01 6.60387099e-01 1.61531597e-01 4.92087990e-01
2.94743538e-01 -1.56862602e-01 -6.85221910e-01 1.95558116e-01
8.62197816e-01 -2.40085553e-02 -1.14512853e-01 -2.85484970e-01
-1.04822743e+00 5.83192945e-01 2.28769019e-01 5.68768144e-01
-2.62917846e-01 -1.63196743e-01 5.50806940e-01 6.85322136e-02
5.84891737e-01 5.76176882e-01 3.37155938e-01 -2.87611961e-01
-1.09722567e+00 2.35576943e-01 4.61462319e-01 1.03059590e+00
5.41221738e-01 1.31908894e-01 -7.75641561e-01 5.97486794e-01
5.19090354e-01 6.36042655e-01 4.09918994e-01 -2.67980933e-01
3.83480698e-01 1.42345059e+00 -2.26737815e-03 -1.19717157e+00
-3.57190520e-01 -1.42794847e-01 -3.91920984e-01 8.83325517e-01
-1.00563847e-01 2.16608983e-03 -1.07424438e+00 1.12862086e+00
4.97848570e-01 -5.99698663e-01 -1.56740993e-01 8.77934813e-01
1.22973800e+00 7.43368864e-01 5.23234665e-01 6.68215454e-02
1.64441383e+00 -5.47757685e-01 -1.26305890e+00 1.05284594e-01
5.83575368e-01 -1.04133368e+00 1.96110320e+00 8.81210268e-02
-8.91963184e-01 -6.97740912e-01 -1.36282527e+00 -4.03162479e-01
-1.01488090e+00 4.93898749e-01 3.02025706e-01 9.29732382e-01
-8.64942133e-01 2.18882635e-01 -4.68936354e-01 -5.81764817e-01
3.14879090e-01 -5.94197035e-01 -1.79705799e-01 4.75376636e-01
-9.60848510e-01 8.34663570e-01 4.81974363e-01 -8.78819823e-02
-1.72341809e-01 -1.01781905e+00 -7.96539724e-01 7.80899450e-03
1.78930327e-01 -8.64565730e-01 1.02364993e+00 -8.46082747e-01
-1.33649933e+00 6.43962026e-01 -8.71088505e-02 -1.64990928e-02
6.55971408e-01 -5.64494848e-01 -5.19312024e-01 1.35994107e-01
2.93200642e-01 3.53866100e-01 6.57185495e-01 -1.81077504e+00
-4.93364573e-01 -3.16183627e-01 -1.14890307e-01 3.08353871e-01
-4.28431898e-01 2.27284059e-01 -6.41593575e-01 -1.16591024e+00
-2.07378849e-01 -5.18395118e-02 8.03590417e-02 2.47849777e-01
-7.04944730e-01 1.74797580e-01 1.63063002e+00 -8.49799216e-01
1.99250472e+00 -2.19034934e+00 -3.58314186e-01 4.36446577e-01
2.50718027e-01 4.74918276e-01 2.97129691e-01 1.22687328e+00
-1.39913812e-01 5.16042292e-01 -9.37167853e-02 -5.44256389e-01
1.69040382e-01 -1.88459128e-01 -3.61064613e-01 -3.32087994e-01
1.06009051e-01 8.58170092e-01 -8.68570447e-01 -7.86190510e-01
5.81598222e-01 7.03194022e-01 -1.62564516e-01 1.81716293e-01
-3.24603528e-01 3.75576586e-01 -2.21878871e-01 5.44753015e-01
6.93736672e-01 -1.37618393e-01 2.93408096e-01 -4.43255216e-01
-6.54228449e-01 1.69358954e-01 -1.23752367e+00 1.67073894e+00
-4.79166865e-01 1.09926772e+00 -2.21288264e-01 -3.39346170e-01
1.29877460e+00 2.74451338e-02 -9.13182423e-02 -9.57830727e-01
-1.46144494e-01 -2.43270963e-01 -6.13801837e-01 -9.12758648e-01
8.56037796e-01 -2.83559635e-02 1.81143865e-01 7.29719102e-01
-3.78534555e-01 -1.61977634e-01 4.42944676e-01 7.21466184e-01
4.28853244e-01 7.45487094e-01 3.14029306e-01 3.52308229e-02
2.06005901e-01 2.23424107e-01 -2.52364099e-01 5.79227328e-01
7.22668409e-01 5.98645866e-01 7.22092807e-01 -3.02264571e-01
-9.66242373e-01 -1.02685606e+00 -8.82382914e-02 8.95272374e-01
1.30256474e-01 -1.16406107e+00 -8.77048552e-01 -3.91871840e-01
-2.44921803e-01 1.27713859e+00 -9.38775182e-01 1.28398001e-01
-2.02422976e-01 -4.67099041e-01 2.56028771e-01 3.75980705e-01
4.76243347e-01 -1.37051117e+00 -1.42790353e+00 -7.10435137e-02
3.17821056e-01 -4.28548157e-01 -1.50084570e-01 -2.95594335e-01
-8.96269798e-01 -9.76125121e-01 -5.40301919e-01 -4.07611966e-01
8.79148722e-01 2.41509303e-01 1.10416460e+00 -2.31525917e-02
-6.53630793e-01 5.04355788e-01 -8.05702567e-01 -8.00502121e-01
-5.53855658e-01 -2.47714356e-01 -8.49367082e-01 -2.62493372e-01
1.11543030e-01 -4.42405105e-01 -7.23494947e-01 3.71918567e-02
-1.43541121e+00 1.04335487e+00 3.62860352e-01 3.38826537e-01
3.72115910e-01 -4.62464452e-01 3.28162760e-01 -1.27801585e+00
1.24054301e+00 -3.65888625e-01 -3.12759876e-01 6.14903271e-01
-7.89884269e-01 9.64251906e-02 4.56476629e-01 -4.85717207e-02
-1.56122112e+00 -3.31299961e-01 3.29828560e-01 -3.94813269e-01
-2.37232625e-01 7.46138513e-01 -2.69020259e-01 5.43986917e-01
6.88230276e-01 6.76154792e-02 6.30579665e-02 -6.46139860e-01
1.06875396e+00 5.07074177e-01 5.52875757e-01 -4.16392773e-01
7.57962286e-01 2.93859154e-01 -2.73408771e-01 -6.50387526e-01
-1.13002598e-01 -1.36560246e-01 -6.56214356e-01 -7.96624005e-01
7.41565645e-01 -3.86130273e-01 -4.64893758e-01 -1.08634077e-01
-1.11037612e+00 -1.07327893e-01 -6.49759412e-01 4.01588902e-02
-1.37496307e-01 2.89057374e-01 6.55995831e-02 -1.00788057e+00
-5.82590461e-01 -7.64150500e-01 1.19245362e+00 3.90798092e-01
-5.78838527e-01 -1.03657401e+00 8.96867737e-02 -2.02405423e-01
4.37106997e-01 8.23988497e-01 1.13513398e+00 -8.89754668e-02
-4.26032871e-01 -1.95734322e-01 -5.74156523e-01 -1.72863483e-01
4.49725837e-01 6.60798073e-01 -9.21950758e-01 2.94396490e-01
-4.72506642e-01 1.68900222e-01 3.72975022e-01 7.44875614e-03
1.16805208e+00 -4.94694114e-01 -6.15427196e-01 4.10624325e-01
1.44412529e+00 7.14319766e-01 6.90422297e-01 5.04698932e-01
8.31682742e-01 8.56529236e-01 7.27534950e-01 6.29583895e-01
1.52315229e-01 5.82514942e-01 3.25486183e-01 -5.58044314e-01
-3.65806848e-01 -7.78992772e-01 -1.30410790e-01 5.12631714e-01
-5.44235110e-02 -1.57897651e-01 -8.14617634e-01 1.88879848e-01
-1.80097234e+00 -9.67160940e-01 -3.13551277e-01 2.05073285e+00
7.38014340e-01 -2.54473034e-02 1.48588553e-01 1.12364203e-01
5.68943083e-01 2.03353286e-01 -1.25834808e-01 -6.65302813e-01
-1.14156261e-01 3.15961808e-01 -1.23986587e-01 3.02648336e-01
-6.52073324e-01 7.92690694e-01 5.86454391e+00 6.88437819e-01
-1.44316757e+00 -3.97285447e-02 3.12476218e-01 -2.44085655e-01
-9.76055622e-01 4.02857363e-02 -3.73071194e-01 4.90456462e-01
4.85956669e-01 -4.00797606e-01 -1.68423772e-01 8.47627401e-01
7.77469635e-01 -3.82101595e-01 -8.64937901e-01 1.14376652e+00
7.75166303e-02 -1.74080336e+00 7.27352262e-01 -2.89067209e-01
3.69052351e-01 -1.02349949e+00 2.33842731e-01 -1.72049567e-01
9.86481160e-02 -9.83068526e-01 1.25078261e+00 7.78638184e-01
1.21147907e+00 -5.01424670e-01 7.67463669e-02 -1.56837761e-01
-1.21202171e+00 2.80286938e-01 5.86607791e-02 -2.55090836e-02
4.44857389e-01 4.85195309e-01 -1.17132294e+00 7.58065403e-01
7.68361211e-01 6.96736455e-01 -1.10632169e+00 1.18567574e+00
-3.87902081e-01 3.60332847e-01 5.23974746e-02 -4.08984393e-01
-5.17545640e-02 -4.79704767e-01 5.00000715e-01 1.95756948e+00
5.91676950e-01 -2.36727744e-01 -2.52831787e-01 1.39776564e+00
3.99027199e-01 6.99930787e-01 -8.69466782e-01 -2.32728481e-01
7.19032884e-01 1.36766076e+00 -1.08392477e+00 -4.46215123e-01
-1.79592222e-01 7.96447992e-01 -1.35557756e-01 5.45237780e-01
-6.31583035e-01 -7.33244896e-01 8.99454206e-02 4.83333141e-01
-2.66175270e-01 -2.16617525e-01 -1.01006210e+00 -6.32211924e-01
2.19985291e-01 -7.10268676e-01 1.41015157e-01 -1.49870837e+00
-7.69490540e-01 5.72377682e-01 4.80170935e-01 -1.27295280e+00
-8.95415694e-02 -4.12328959e-01 -1.05792582e+00 1.12300539e+00
-6.51329041e-01 -1.47316504e+00 -8.06832016e-01 2.33357027e-01
5.40732980e-01 2.41789699e-01 7.50828207e-01 2.28012919e-01
-4.77561682e-01 1.68714151e-01 -7.11339638e-02 -1.16962440e-01
6.36008322e-01 -1.50613678e+00 6.80780947e-01 8.76371026e-01
3.74011397e-01 8.08948040e-01 8.18031371e-01 -9.93121505e-01
-1.09887862e+00 -7.07340300e-01 4.12477642e-01 -3.91261071e-01
3.29317003e-01 -7.28932202e-01 -8.97648990e-01 4.09025580e-01
6.63973153e-01 -7.03771889e-01 8.68127465e-01 -1.05998643e-01
-3.24455112e-01 1.69781417e-01 -8.30834329e-01 1.13853288e+00
9.86946225e-01 -3.99968207e-01 -7.29403019e-01 -1.62425071e-01
6.53601587e-01 -2.83697873e-01 -3.56541306e-01 -3.07442620e-02
7.90691733e-01 -1.17632759e+00 8.05273235e-01 -4.01474804e-01
3.77278090e-01 -6.36048436e-01 3.38183075e-01 -1.26026356e+00
9.53998882e-03 -7.28502870e-01 2.25135416e-01 1.64103675e+00
5.69650888e-01 -3.18731308e-01 3.27825963e-01 4.59490567e-01
-2.61501104e-01 -3.28595370e-01 -3.94952357e-01 -2.82512307e-01
-5.77474833e-01 -6.52655900e-01 7.86155641e-01 8.75960290e-01
3.75485629e-01 2.09730893e-01 -1.26156077e-01 -2.49910340e-01
6.37509972e-02 1.99755877e-01 1.07931340e+00 -1.22394311e+00
2.18962923e-01 -7.62405396e-01 -1.30814731e-01 -4.58765209e-01
-6.78098202e-01 -7.93103933e-01 -4.81859714e-01 -2.19088817e+00
1.69216450e-02 -2.71787256e-01 2.73340642e-01 3.63738179e-01
-2.76071310e-01 -5.77503722e-03 7.23118365e-01 1.63710192e-01
-2.69080490e-01 4.36206877e-01 1.38404834e+00 1.29986539e-01
-5.60478985e-01 -6.27417505e-01 -8.55719626e-01 5.07751584e-01
6.10313237e-01 -9.94717777e-02 -8.17899585e-01 -1.20529011e-01
5.08307815e-01 -4.49075878e-01 4.41944718e-01 -9.70065057e-01
-6.99288622e-02 -1.86000422e-01 7.82281697e-01 -9.96021986e-01
1.09639160e-01 -7.87304759e-01 5.46140492e-01 2.26310134e-01
-3.43777835e-01 5.00790894e-01 5.79221845e-01 2.94557422e-01
-5.32178171e-02 -1.54311448e-01 2.15640664e-01 -2.97065563e-02
-5.71324170e-01 -4.42467004e-01 -4.00743365e-01 -1.58238679e-01
9.64717507e-01 -7.19565749e-01 -7.02914894e-01 -3.72127712e-01
-6.23409927e-01 -4.68787029e-02 5.06475747e-01 7.00965285e-01
7.82043874e-01 -1.38195539e+00 -7.65471607e-02 3.83509427e-01
2.72042602e-01 -1.60728306e-01 2.05171421e-01 2.34730914e-01
-9.09876943e-01 3.02336782e-01 -5.31663835e-01 -3.95653933e-01
-1.45648777e+00 6.70131743e-01 -2.83331096e-01 4.91542220e-02
-9.82051790e-01 2.22449839e-01 5.73600233e-01 9.16186422e-02
8.38887021e-02 -6.14237189e-01 -6.44682407e-01 4.80790049e-01
8.10506999e-01 5.50714910e-01 -6.94685057e-02 -2.91336507e-01
6.11542203e-02 4.91576701e-01 3.00487399e-01 -5.00658095e-01
9.83573318e-01 -2.97346205e-01 1.73506722e-01 8.26272845e-01
6.55953646e-01 2.49933302e-01 -1.04279900e+00 2.99429119e-01
1.40108109e-01 -6.00118041e-01 -4.05776739e-01 -1.42984056e+00
-5.43658257e-01 1.12592089e+00 6.38715863e-01 7.26294756e-01
1.05163312e+00 -9.80152413e-02 1.15550444e-01 -1.84974343e-01
-6.17562197e-02 -1.06062281e+00 1.43577456e-01 -1.70070231e-01
1.56851137e+00 -8.09735358e-01 1.74586594e-01 -5.67970276e-01
-7.67733991e-01 1.40194929e+00 7.31985569e-01 4.40335363e-01
4.10988957e-01 3.42627108e-01 4.27292347e-01 -6.99551344e-01
-5.55717766e-01 -1.23853087e-01 5.78796208e-01 9.02875006e-01
7.39139915e-01 9.76948291e-02 -5.43124795e-01 5.64055204e-01
-4.39522237e-01 1.55299231e-02 1.78564474e-01 1.06409168e+00
-4.02047276e-01 -1.09761631e+00 -8.05272102e-01 3.14479351e-01
-9.51404199e-02 -2.02968165e-01 -6.99967086e-01 1.07323492e+00
1.66101351e-01 7.57588267e-01 3.12523574e-01 -9.40318629e-02
4.84514147e-01 1.54767588e-01 1.50181681e-01 -6.84563696e-01
-7.32528448e-01 2.00630307e-01 4.59187813e-02 -4.36740488e-01
-3.50224257e-01 -4.02405500e-01 -1.21890199e+00 -1.58050340e-02
2.54970074e-01 1.24939747e-01 1.15959132e+00 4.79710042e-01
7.24810600e-01 8.34659576e-01 -3.55337583e-03 -7.37871587e-01
3.87048244e-01 -1.25058579e+00 -3.04201156e-01 4.91912186e-01
3.27796995e-04 -6.60845995e-01 1.79981127e-01 3.86810720e-01] | [11.2767333984375, 1.8061137199401855] |
8fe59320-9736-47a9-99c3-7fd16f67f859 | hitrans-a-hierarchical-transformer-network | null | null | https://aclanthology.org/2021.findings-emnlp.12 | https://aclanthology.org/2021.findings-emnlp.12.pdf | HiTRANS: A Hierarchical Transformer Network for Nested Named Entity Recognition | Nested Named Entity Recognition (NNER) has been extensively studied, aiming to identify all nested entities from potential spans (i.e., one or more continuous tokens). However, recent studies for NNER either focus on tedious tagging schemas or utilize complex structures, which fail to learn effective span representations from the input sentence with highly nested entities. Intuitively, explicit span representations will contribute to NNER due to the rich context information they contain. In this study, we propose a Hierarchical Transformer (HiTRANS) network for the NNER task, which decomposes the input sentence into multi-grained spans and enhances the representation learning in a hierarchical manner. Specifically, we first utilize a two-phase module to generate span representations by aggregating context information based on a bottom-up and top-down transformer network. Then a label prediction layer is designed to recognize nested entities hierarchically, which naturally explores semantic dependencies among different spans. Experiments on GENIA, ACE-2004, ACE-2005 and NNE datasets demonstrate that our proposed method achieves much better performance than the state-of-the-art approaches. | ['Yi Chang', 'Yunke Zhang', 'Hechang Chen', 'Jing Ma', 'Zhiwei Yang'] | null | null | null | null | findings-emnlp-2021-11 | ['nested-named-entity-recognition'] | ['natural-language-processing'] | [ 9.16050449e-02 2.15022638e-01 -1.84616834e-01 -4.46114510e-01
-7.98827946e-01 -5.47164142e-01 1.83433324e-01 2.67379671e-01
-3.00914288e-01 8.18257630e-01 7.74550736e-01 -1.40778765e-01
1.24068938e-01 -1.01458943e+00 -5.44128478e-01 -2.44577676e-01
-3.84676792e-02 1.84046760e-01 3.20030272e-01 8.07056483e-03
2.95778122e-02 1.40664816e-01 -1.17320502e+00 4.95450765e-01
1.15954459e+00 9.19479489e-01 2.01184582e-02 2.33230427e-01
-7.47452319e-01 1.16748750e+00 -5.46541989e-01 -5.91632307e-01
-2.45811105e-01 -3.55186433e-01 -1.12705076e+00 9.42846909e-02
2.64597565e-01 1.57085285e-01 -2.34368905e-01 9.25057292e-01
2.57683665e-01 3.06531519e-01 5.64353824e-01 -7.86028326e-01
-9.30880427e-01 1.33781791e+00 -4.48794991e-01 7.51543865e-02
1.27859443e-01 -3.41628939e-01 1.49704766e+00 -1.13059008e+00
6.03522182e-01 1.20710921e+00 7.89411366e-01 4.10691112e-01
-1.12379396e+00 -7.33074784e-01 3.47928107e-01 6.29017949e-02
-1.47425210e+00 -2.18486473e-01 8.22256923e-01 -3.11729580e-01
9.07574654e-01 -2.13291822e-03 2.68933326e-01 9.69520867e-01
1.60199814e-02 9.66993153e-01 8.69797766e-01 -4.42397445e-01
7.11111650e-02 -1.31470025e-01 6.09404445e-01 7.37346768e-01
2.41150334e-01 -2.35555813e-01 -2.79742330e-01 -2.15725318e-01
4.96752739e-01 1.85814053e-01 2.58734394e-02 -2.58485556e-01
-1.11806262e+00 8.69330108e-01 7.23315716e-01 7.04302430e-01
-5.77564120e-01 -1.38564721e-01 6.74962878e-01 -7.68227726e-02
4.22887474e-01 5.13953984e-01 -5.26802599e-01 1.42399058e-01
-8.69876742e-01 -8.56014267e-02 7.96931207e-01 1.04900956e+00
8.52893949e-01 -1.04594149e-01 -6.33499682e-01 1.09145069e+00
3.19053799e-01 4.32467721e-02 3.65904063e-01 -4.46111143e-01
9.16214764e-01 1.00443399e+00 1.17632553e-01 -8.44057500e-01
-5.36470473e-01 -6.31082356e-01 -8.57070625e-01 -6.78245664e-01
2.45595891e-02 -3.15731853e-01 -8.43680501e-01 1.95833862e+00
3.17816943e-01 2.11864948e-01 3.45400780e-01 6.09972596e-01
9.57563937e-01 7.64207721e-01 8.46686482e-01 -8.09056088e-02
1.76675057e+00 -1.22060525e+00 -7.24572241e-01 -4.02230442e-01
9.67833996e-01 -4.80143547e-01 8.97047877e-01 -1.32348225e-01
-8.22144628e-01 -5.84480345e-01 -8.22970927e-01 -2.76977539e-01
-5.34827709e-01 4.70026225e-01 4.36229438e-01 2.84839779e-01
-4.85930949e-01 3.63164574e-01 -5.79935610e-01 -2.62258291e-01
2.98895985e-01 2.34588124e-02 -3.87771368e-01 -2.21405342e-01
-1.74605095e+00 8.29798102e-01 9.95635331e-01 2.07863480e-01
-5.88244975e-01 -7.84314036e-01 -1.42653620e+00 5.20594418e-01
5.83772242e-01 -5.95172048e-01 1.09278452e+00 -5.93505204e-01
-9.85408783e-01 7.07640767e-01 -2.73107052e-01 -5.81581950e-01
-1.04098432e-01 -4.24796313e-01 -7.26733208e-01 7.69494772e-02
3.02154124e-01 6.16662920e-01 2.33095437e-01 -1.08924401e+00
-5.62322736e-01 -2.52983302e-01 2.12859198e-01 2.24745721e-01
-5.99131703e-01 2.29003042e-01 -1.37669891e-01 -7.27616012e-01
2.93158628e-02 -7.15722620e-01 -3.85878503e-01 -6.90902829e-01
-5.47964036e-01 -7.43408084e-01 4.14091438e-01 -8.39980602e-01
1.96082366e+00 -2.05151486e+00 1.82778179e-03 -7.75064379e-02
2.66959310e-01 1.99675694e-01 -1.32726669e-01 7.43845522e-01
-1.84958726e-01 3.27605993e-01 -2.08603755e-01 -3.80181998e-01
2.08755255e-01 1.69474125e-01 -3.51571411e-01 -7.46474341e-02
3.48560065e-01 8.32887232e-01 -9.75924194e-01 -9.57319796e-01
-1.03122808e-01 1.50212333e-01 -2.42888942e-01 3.97356987e-01
-8.90057907e-02 -4.33945172e-02 -4.49284494e-01 5.53525507e-01
5.67831635e-01 -3.63451689e-01 4.01650012e-01 -3.52758110e-01
-3.03091049e-01 7.25052178e-01 -1.01332510e+00 1.66674900e+00
-7.34494627e-01 4.31667157e-02 -2.26418272e-01 -8.37837756e-01
1.10458302e+00 4.45645750e-01 2.00901479e-01 -5.21631241e-01
-9.75829214e-02 1.43465966e-01 -2.42883012e-01 -3.93078059e-01
7.31442332e-01 -3.56613606e-01 -7.33939826e-01 9.03692096e-02
2.09259853e-01 7.57860601e-01 4.45271015e-01 2.21781731e-01
1.26710534e+00 2.27038772e-03 6.88264847e-01 -1.90095738e-01
7.17533052e-01 -6.72585815e-02 1.21556640e+00 4.71091688e-01
-9.91566032e-02 2.07657263e-01 5.01384914e-01 -3.55520755e-01
-8.61513078e-01 -1.10113406e+00 1.87133439e-02 1.36460495e+00
1.35169521e-01 -6.41679883e-01 -7.34441459e-01 -1.10251629e+00
-2.25740671e-01 8.43650401e-01 -7.89815903e-01 -9.78571847e-02
-7.64092147e-01 -2.91058719e-01 8.46414983e-01 1.07393467e+00
7.26823509e-01 -1.42848468e+00 -3.03390115e-01 5.40042996e-01
-4.42936450e-01 -1.14624810e+00 -6.76885426e-01 2.33204365e-01
-8.34940434e-01 -7.08147466e-01 -5.05030215e-01 -1.07833457e+00
5.45315027e-01 -1.42224533e-02 1.37170088e+00 -6.12570904e-02
-5.15991673e-02 -2.24827692e-01 -6.38943374e-01 -3.12204659e-02
-2.61089746e-02 6.54507697e-01 -2.41682187e-01 -3.45127583e-02
6.66189373e-01 -3.68290365e-01 -4.41544116e-01 2.04078570e-01
-8.48015904e-01 1.14065088e-01 9.79447067e-01 1.07400429e+00
5.06760478e-01 -6.55678287e-02 9.58226740e-01 -1.49094331e+00
3.38910758e-01 -7.89107442e-01 -1.44094005e-01 7.18216062e-01
-4.02453363e-01 3.75924319e-01 9.21156228e-01 -2.89823860e-01
-1.48298812e+00 7.11628720e-02 -3.11358303e-01 -3.76092345e-01
-3.26407999e-01 8.34513962e-01 -5.08867204e-01 5.82808197e-01
4.32389736e-01 2.26267502e-01 -8.99515152e-01 -5.99965036e-01
5.33365369e-01 6.61399186e-01 4.09318656e-01 -9.74638760e-01
6.66027606e-01 -1.86592247e-02 -3.74933571e-01 -4.68419015e-01
-1.52586091e+00 -6.36057556e-01 -6.43726945e-01 1.93322420e-01
9.90868866e-01 -9.57875133e-01 -2.67053276e-01 6.16677739e-02
-1.12946498e+00 1.28158972e-01 -2.83165395e-01 3.66523445e-01
-7.63240037e-03 2.97748178e-01 -9.54239368e-01 -6.71285629e-01
-6.00394547e-01 -5.64771235e-01 1.03981280e+00 6.53839886e-01
-2.25777522e-01 -9.95830238e-01 1.57037050e-01 2.92019933e-01
1.69031590e-01 6.38836697e-02 1.04527318e+00 -9.72219408e-01
-2.84072578e-01 3.66256870e-02 -4.42570537e-01 8.51776972e-02
1.56748563e-01 -3.91752809e-01 -8.23376954e-01 -3.26514768e-04
-4.28764373e-01 -4.53600675e-01 9.08017933e-01 -1.29014596e-01
1.17517650e+00 -4.92033541e-01 -3.63607556e-01 2.61151105e-01
1.57669938e+00 1.55379802e-01 6.36222720e-01 2.22233325e-01
8.72151434e-01 7.06902206e-01 7.54000127e-01 3.73046488e-01
6.93199456e-01 2.15927511e-01 1.47984430e-01 -7.71998689e-02
1.42488033e-01 -6.70307755e-01 2.27096796e-01 1.11639321e+00
2.69692153e-01 -1.73420638e-01 -8.58847022e-01 8.80545199e-01
-1.79118526e+00 -9.91684377e-01 8.12659934e-02 1.80398154e+00
9.77337658e-01 1.89227238e-01 -1.13791749e-01 -2.11812660e-01
1.00460231e+00 4.46959734e-01 -4.91152316e-01 -3.47889900e-01
8.03372487e-02 8.03777650e-02 2.19202325e-01 -3.19570303e-02
-1.30903327e+00 1.21701610e+00 5.46692324e+00 8.97132993e-01
-6.28129542e-01 9.41933542e-02 4.03255254e-01 5.52741408e-01
-4.56164986e-01 1.55238494e-01 -1.11004603e+00 3.88787836e-01
9.68317449e-01 -2.21860856e-01 -8.97810087e-02 1.09582686e+00
-1.83428153e-01 3.10052812e-01 -9.82545435e-01 6.88125193e-01
-4.32332158e-02 -1.30441403e+00 4.33679335e-02 -1.28567457e-01
7.81494677e-01 -2.01199293e-01 -4.34535235e-01 9.03747976e-01
5.94371319e-01 -7.78662562e-01 6.05042815e-01 5.99446714e-01
5.08123100e-01 -8.71158600e-01 9.50052440e-01 3.98224086e-01
-2.02076745e+00 -1.70445114e-01 -5.49004912e-01 1.58040926e-01
1.94683567e-01 6.92102551e-01 -5.55139780e-01 9.31991577e-01
3.97358596e-01 5.86213946e-01 -6.01592183e-01 9.73957479e-01
-3.98108870e-01 8.86677742e-01 -6.98665455e-02 -5.63416891e-02
4.07573760e-01 5.34051983e-03 4.82012443e-02 1.65298820e+00
2.39336118e-01 2.93912351e-01 5.56754410e-01 8.09626997e-01
-6.01128995e-01 3.88888776e-01 -4.85546142e-01 -1.96708933e-01
9.00981486e-01 1.48965716e+00 -6.29559219e-01 -4.81258303e-01
-6.71549797e-01 7.25543499e-01 1.08804345e+00 1.25530168e-01
-8.48490417e-01 -8.95813644e-01 3.59546959e-01 -2.28580683e-01
6.06605828e-01 -1.64552420e-01 -2.92001963e-01 -1.38061154e+00
-1.15964692e-02 -4.32756752e-01 8.47324073e-01 -5.60575843e-01
-1.64662635e+00 7.39904046e-01 -2.36506730e-01 -1.14258182e+00
-7.46464282e-02 -3.09444398e-01 -7.99097478e-01 7.32543647e-01
-1.44609392e+00 -1.32667220e+00 1.33793894e-02 2.60462195e-01
6.65939331e-01 2.20017433e-01 8.61301005e-01 3.97002846e-01
-8.38278294e-01 6.52617574e-01 -2.12886840e-01 7.68468559e-01
5.37650883e-01 -1.33088183e+00 1.97231159e-01 9.95166838e-01
2.91278452e-01 9.65886116e-01 1.52447358e-01 -7.11445510e-01
-8.14379573e-01 -1.46686149e+00 1.55334198e+00 -1.72357291e-01
7.59689450e-01 -4.42784697e-01 -1.09868526e+00 8.40561569e-01
4.31806833e-01 1.91867322e-01 9.63988185e-01 5.83181500e-01
-6.87573791e-01 -1.78940408e-02 -8.14734459e-01 4.27465707e-01
1.30299437e+00 -6.82878435e-01 -1.31173503e+00 -1.78321391e-01
1.11026609e+00 -1.38448507e-01 -1.22290587e+00 3.88615549e-01
2.12983713e-01 -5.54440320e-01 6.96535408e-01 -7.86482394e-01
5.72976530e-01 -3.27245146e-01 -1.53137490e-01 -1.20561087e+00
-5.45546412e-01 -9.44816843e-02 -2.38166809e-01 2.01203704e+00
6.34192824e-01 -2.84824938e-01 5.95642745e-01 4.63755280e-01
-4.16145891e-01 -9.11852479e-01 -5.35774410e-01 -7.89447784e-01
-3.65674379e-03 -1.47910148e-01 8.72644544e-01 1.04647124e+00
1.93442658e-01 9.50789332e-01 -3.51806313e-01 3.62688750e-01
5.29645860e-01 6.47667348e-01 8.72294381e-02 -1.21684158e+00
2.68966947e-02 -1.15948860e-02 -7.63622150e-02 -1.15721381e+00
5.23464620e-01 -8.36918950e-01 3.18599790e-01 -1.79178238e+00
3.78836721e-01 -6.50590718e-01 -6.70517087e-01 8.11895847e-01
-6.60125136e-01 -1.50838003e-01 1.19439356e-01 8.86036009e-02
-1.17359412e+00 6.92732394e-01 8.41625750e-01 -3.61356325e-02
5.81029169e-02 -3.44822377e-01 -7.90108562e-01 7.37360477e-01
6.89100206e-01 -6.23683989e-01 -2.93365687e-01 -4.63105947e-01
2.94965237e-01 2.49035552e-01 -1.05093569e-02 -1.00791407e+00
2.75007039e-01 -1.57205388e-01 4.22810137e-01 -7.92236626e-01
-1.43289119e-01 -6.84915662e-01 -8.84550735e-02 9.84721854e-02
-7.11025298e-01 -5.35250781e-03 -1.74708247e-01 3.89867693e-01
-5.56848764e-01 -5.59127927e-01 5.22232413e-01 -2.96469837e-01
-9.19919550e-01 3.53330761e-01 -2.10739240e-01 5.59295118e-01
8.43159974e-01 2.87163794e-01 -3.43875945e-01 3.44153374e-01
-9.02085185e-01 5.04137218e-01 -2.05531297e-03 4.12859499e-01
3.83210450e-01 -1.69260561e+00 -6.31678879e-01 -1.05437770e-01
4.40005511e-01 1.37603015e-01 5.84643960e-01 3.58254135e-01
-9.73129123e-02 3.38366956e-01 -2.21363995e-02 -1.38419837e-01
-1.10794258e+00 8.06479037e-01 -2.33214032e-02 -1.18734491e+00
-5.62560797e-01 7.21223712e-01 3.65943700e-01 -5.59663832e-01
8.76535382e-03 -3.08417171e-01 -7.45226800e-01 2.83046037e-01
3.86438370e-01 2.34871551e-01 -1.58980295e-01 -6.26154363e-01
-3.44501585e-01 3.65195364e-01 -3.31440121e-01 2.73418456e-01
1.13090479e+00 -2.77108878e-01 -1.47792578e-01 4.21430647e-01
1.16132486e+00 5.40746450e-02 -9.62079883e-01 -6.60243213e-01
6.89144254e-01 8.38081469e-04 -3.37747991e-01 -4.23684806e-01
-8.31274152e-01 9.48804855e-01 1.46268187e-02 2.64812768e-01
1.07274044e+00 2.56418765e-01 1.18760788e+00 4.21074718e-01
3.54949236e-01 -1.07502007e+00 -2.41292547e-02 8.09504032e-01
5.15256345e-01 -8.74089897e-01 -4.35516894e-01 -5.76303184e-01
-7.74512172e-01 9.99086499e-01 9.87956882e-01 -8.36100057e-02
3.73989820e-01 1.60227746e-01 -2.12984562e-01 9.16115288e-03
-8.46053541e-01 -4.08003002e-01 2.81081408e-01 2.07718849e-01
7.52547562e-01 1.69394970e-01 -5.27762115e-01 1.25069726e+00
2.07417216e-02 1.45413876e-02 -3.39398757e-02 8.13322902e-01
-6.01765931e-01 -1.12584043e+00 -1.45687148e-01 4.06847566e-01
-5.46816289e-01 -4.20477480e-01 -1.31459996e-01 5.24045110e-01
4.58771288e-01 8.04147601e-01 -1.98649019e-02 -3.09546620e-01
4.87850279e-01 5.14287651e-01 1.05019659e-04 -9.43135560e-01
-8.62787902e-01 -1.91432163e-01 4.88782257e-01 -3.83052349e-01
-3.98599625e-01 -5.72103560e-01 -1.56799304e+00 1.36468828e-01
-4.17439044e-01 7.64419556e-01 8.54239836e-02 1.11261475e+00
4.26957577e-01 7.00209737e-01 6.83552384e-01 -2.30988592e-01
-4.78651494e-01 -1.17078078e+00 -6.25331998e-01 5.26059628e-01
-1.53183460e-01 -6.62853241e-01 -2.71085296e-02 3.64509895e-02] | [9.588216781616211, 9.387736320495605] |
41f6f345-5fdf-47fc-93d5-c2b822789f1e | improving-expressivity-of-graph-neural-1 | 2305.19659 | null | https://arxiv.org/abs/2305.19659v1 | https://arxiv.org/pdf/2305.19659v1.pdf | Improving Expressivity of Graph Neural Networks using Localization | In this paper, we propose localized versions of Weisfeiler-Leman (WL) algorithms in an effort to both increase the expressivity, as well as decrease the computational overhead. We focus on the specific problem of subgraph counting and give localized versions of $k-$WL for any $k$. We analyze the power of Local $k-$WL and prove that it is more expressive than $k-$WL and at most as expressive as $(k+1)-$WL. We give a characterization of patterns whose count as a subgraph and induced subgraph are invariant if two graphs are Local $k-$WL equivalent. We also introduce two variants of $k-$WL: Layer $k-$WL and recursive $k-$WL. These methods are more time and space efficient than applying $k-$WL on the whole graph. We also propose a fragmentation technique that guarantees the exact count of all induced subgraphs of size at most 4 using just $1-$WL. The same idea can be extended further for larger patterns using $k>1$. We also compare the expressive power of Local $k-$WL with other GNN hierarchies and show that given a bound on the time-complexity, our methods are more expressive than the ones mentioned in Papp and Wattenhofer[2022a]. | ['Anirban Dasgupta', 'Binita Maity', 'Shubhajit Roy', 'Shrutimoy Das', 'Anant Kumar'] | 2023-05-31 | null | null | null | null | ['subgraph-counting'] | ['graphs'] | [ 8.01444575e-02 3.25002968e-01 -2.38881946e-01 -2.14008093e-02
-5.08001387e-01 -7.94946790e-01 -1.41341746e-01 3.22624534e-01
-5.35492361e-01 7.53200531e-01 -1.47233009e-01 -5.70606232e-01
-5.50417483e-01 -1.49514306e+00 -7.91566610e-01 -4.26603585e-01
-7.45514452e-01 4.37311143e-01 8.89812887e-01 -2.55504251e-01
8.23011845e-02 7.37950742e-01 -1.43502617e+00 1.07651524e-01
2.68463194e-01 6.63212001e-01 -2.30082393e-01 9.18049455e-01
-3.39871794e-01 3.41807902e-01 -1.91671655e-01 -4.48698848e-01
6.45077944e-01 -4.33259368e-01 -1.34309816e+00 2.38668486e-01
5.16192496e-01 3.02344143e-01 -2.73038626e-01 1.03947890e+00
4.12002802e-02 -1.70738116e-01 1.31866187e-01 -1.33045888e+00
3.55188280e-01 1.25894654e+00 -8.51469398e-01 -3.44918631e-02
4.10433352e-01 -2.31776610e-01 1.12417006e+00 -2.42395967e-01
7.01309383e-01 8.19065452e-01 9.18631494e-01 3.36534679e-01
-1.50001299e+00 -7.72864401e-01 2.57101417e-01 -2.15304010e-02
-1.69300222e+00 -4.08067070e-02 5.43873727e-01 8.44510365e-03
1.19962561e+00 7.67063916e-01 6.64131343e-01 -2.36382410e-01
-1.53634027e-01 3.95293862e-01 1.14135432e+00 -8.29229593e-01
-9.55815315e-02 -2.47155890e-01 6.10216260e-01 1.38196719e+00
8.48535001e-01 -3.55532348e-01 -3.48975837e-01 -4.42754686e-01
6.93133533e-01 -1.95246220e-01 -2.67455935e-01 -3.80109310e-01
-8.72797489e-01 7.06207812e-01 2.78513402e-01 8.65843832e-01
3.81097257e-01 6.92254424e-01 3.63107771e-01 5.84848523e-01
1.45950541e-01 4.73912239e-01 -5.85575342e-01 2.18860269e-01
-9.74454165e-01 2.75327682e-01 1.15381086e+00 1.26739848e+00
1.31203938e+00 -2.58327305e-01 4.17284340e-01 4.53007221e-01
-2.31850028e-01 3.41934532e-01 -2.46037081e-01 -8.54625165e-01
5.15993714e-01 9.64056194e-01 -2.32212599e-02 -1.03145742e+00
-6.64388835e-01 -1.63113222e-01 -7.21377611e-01 -6.67892918e-02
5.59625745e-01 2.28027612e-01 -4.37390924e-01 2.14674664e+00
1.74582362e-01 -2.50263512e-01 -5.28175592e-01 1.79488838e-01
8.04831982e-02 6.89877391e-01 -2.71963477e-01 -6.54733062e-01
1.21246004e+00 -8.85781467e-01 -1.71126187e-01 2.59148516e-02
1.25952506e+00 -5.10679245e-01 9.16450739e-01 3.01503897e-01
-1.57635152e+00 -7.29797408e-02 -8.84900510e-01 2.00388640e-01
-5.51091373e-01 -1.63663954e-01 8.70724201e-01 1.00369143e+00
-1.52589750e+00 5.23228467e-01 -5.20088792e-01 -6.13657534e-01
-1.42597139e-01 8.01452279e-01 -5.33104658e-01 -5.79186119e-02
-6.51973784e-01 3.83701354e-01 4.66274232e-01 -3.08430552e-01
-2.26683766e-01 -4.93338287e-01 -8.21721494e-01 1.16200127e-01
7.86971569e-01 -3.02715063e-01 8.92714262e-01 -6.25485122e-01
-8.85171533e-01 9.97804940e-01 -3.67711395e-01 -3.65154237e-01
1.00002997e-01 4.61157918e-01 -1.67404845e-01 1.70863062e-01
-1.27497958e-02 2.19776612e-02 -1.01413250e-01 -1.05399024e+00
-6.60296202e-01 -4.23037827e-01 5.70951045e-01 -2.67447263e-01
-2.97188371e-01 1.50915638e-01 -5.08762538e-01 -2.02988416e-01
3.24648589e-01 -8.47323954e-01 -5.59648812e-01 -1.45397648e-01
-3.73943597e-01 -3.88389349e-01 3.85713309e-01 8.34216923e-02
1.84519839e+00 -1.89570713e+00 -4.48482223e-02 9.89363194e-01
6.11431122e-01 1.26447752e-01 -1.27863467e-01 8.52868140e-01
-5.37295453e-02 4.73437071e-01 -2.46325046e-01 1.86175182e-02
2.02127829e-01 4.17634547e-01 5.55468872e-02 6.00369871e-01
-4.04185027e-01 5.78815460e-01 -5.98101258e-01 -5.93077481e-01
-1.51510030e-01 -5.02813518e-01 -7.43637323e-01 -1.78502172e-01
-3.19203377e-01 -7.31388450e-01 -1.53775215e-01 4.30346906e-01
9.38286841e-01 -3.94086868e-01 7.18447030e-01 -9.75350738e-02
-4.76518750e-01 7.98132643e-02 -1.54175305e+00 1.13547802e+00
-3.57953399e-01 5.32366056e-03 4.36240137e-01 -1.06471491e+00
5.31441689e-01 -8.67692679e-02 3.94473344e-01 -3.45198274e-01
-3.53570818e-03 5.14035642e-01 -3.01617771e-01 -1.76806580e-02
4.98707324e-01 -4.93672758e-01 -6.76416755e-01 7.49455214e-01
-2.01748386e-01 1.63592875e-03 8.26075792e-01 6.62152946e-01
1.61866415e+00 -4.45966721e-01 4.05364841e-01 -8.43152583e-01
5.95099092e-01 -5.06372787e-02 5.67795038e-01 9.33562875e-01
1.15507975e-01 -1.43891394e-01 1.02489150e+00 -2.06913978e-01
-8.58123481e-01 -9.73061979e-01 7.92643055e-02 1.00143158e+00
3.08694631e-01 -1.21522975e+00 -8.55662823e-01 -5.69400489e-01
-2.05707207e-01 3.64543617e-01 -6.67894185e-01 8.84444416e-02
-7.39218354e-01 -3.96317035e-01 7.46939719e-01 4.72934544e-01
6.67718291e-01 -6.56202197e-01 -6.04297757e-01 2.39441335e-01
-8.01012591e-02 -9.71706688e-01 -7.02401400e-01 3.61520112e-01
-8.84467363e-01 -1.42998636e+00 -1.29744962e-01 -9.52245474e-01
1.00434709e+00 3.22329760e-01 9.99735892e-01 4.41186100e-01
-4.15905297e-01 5.02171159e-01 -2.25869268e-01 2.29675353e-01
-9.74261388e-02 2.46138841e-01 -5.66575713e-02 -4.62183446e-01
5.95886707e-02 -5.48404276e-01 -4.27392274e-01 4.14124846e-01
-1.07026684e+00 -1.13038041e-01 3.08771968e-01 4.09871370e-01
7.19458699e-01 6.52386904e-01 -1.49070084e-01 -1.42511618e+00
2.00086907e-01 1.14391651e-02 -9.83330607e-01 1.94360301e-01
-7.63426065e-01 3.32525879e-01 1.06111932e+00 -1.04231238e-01
-4.08569187e-01 1.75129354e-01 -8.35789517e-02 1.91631675e-01
3.13796341e-01 5.74113011e-01 -1.74290016e-01 -5.04442751e-01
2.86580354e-01 -2.99920868e-02 -5.26173353e-01 -3.00743908e-01
3.04726154e-01 9.50178951e-02 9.95271653e-02 -7.31056154e-01
8.52944791e-01 5.48159599e-01 6.71894312e-01 -6.95196271e-01
-3.71698290e-01 -6.40468597e-01 -1.22007243e-01 7.72017315e-02
1.99453652e-01 -1.74493715e-01 -1.01900601e+00 2.32856423e-01
-8.02610457e-01 -6.61273360e-01 -3.48886788e-01 -7.34197348e-02
-5.90633214e-01 8.43443811e-01 -7.54811585e-01 -7.93525755e-01
-1.74891651e-01 -8.48646224e-01 6.37204051e-01 -3.09340209e-01
-2.34642848e-01 -7.28332102e-01 3.02404612e-01 -2.12914512e-01
3.67390215e-01 5.03887594e-01 1.33168793e+00 -4.26789016e-01
-5.60974240e-01 -2.42181212e-01 -5.67705393e-01 -2.01578736e-01
-1.87953129e-01 -2.34626710e-01 -5.71436360e-02 -4.16662633e-01
-4.38097209e-01 -1.79206595e-01 7.36265659e-01 -1.51632249e-03
1.22175908e+00 -8.40333223e-01 -6.72965825e-01 3.24531555e-01
2.19485021e+00 -1.34442896e-01 6.99558198e-01 2.33156517e-01
2.69755036e-01 4.03007507e-01 1.71392813e-01 1.92271173e-01
1.42411157e-01 5.04606366e-01 1.89180553e-01 6.40156418e-02
1.58147722e-01 -1.99535206e-01 1.94751605e-01 1.00015247e+00
-2.26914883e-01 -2.05911055e-01 -6.23516738e-01 7.57389307e-01
-1.69319093e+00 -9.27502096e-01 -3.76996368e-01 2.43783426e+00
1.01685607e+00 3.92579347e-01 6.11091733e-01 5.17642558e-01
7.11751699e-01 1.76423699e-01 8.34495202e-02 -1.06536329e+00
2.29433663e-02 1.04806292e+00 1.16533875e+00 8.30242813e-01
-4.70833093e-01 9.18093443e-01 6.65441799e+00 1.08337641e+00
-5.18301547e-01 1.25297874e-01 3.46226364e-01 -9.88269374e-02
-4.71208215e-01 3.26416135e-01 -7.76449323e-01 2.09692881e-01
1.14923584e+00 -3.28700840e-01 4.68110889e-01 8.54427695e-01
-2.54433692e-01 -4.01197970e-01 -1.00728476e+00 6.25957370e-01
-1.17709741e-01 -1.36186266e+00 -4.15894426e-02 4.32553142e-01
6.02056921e-01 -3.71892959e-01 -4.52683806e-01 1.50112927e-01
5.30576646e-01 -7.34462321e-01 3.54377687e-01 -1.89585164e-02
8.65312636e-01 -1.12930548e+00 5.61839938e-01 3.96925956e-01
-2.04079700e+00 1.39446005e-01 -4.49858546e-01 -1.47713661e-01
9.15115327e-02 5.62592387e-01 -1.13194339e-01 8.20564210e-01
6.09721661e-01 -1.84291348e-01 -2.20587775e-01 7.74463892e-01
2.72784323e-01 4.44585502e-01 -8.75634968e-01 -2.34400600e-01
3.99259567e-01 -1.42263785e-01 1.36601865e-01 1.58243489e+00
4.46676821e-01 5.00916839e-01 2.30444878e-01 3.86437118e-01
-3.45136046e-01 2.59249091e-01 -6.14369333e-01 -6.05948120e-02
2.58069873e-01 9.07668769e-01 -1.08565474e+00 -2.57546246e-01
-4.57376063e-01 6.71531022e-01 3.94523084e-01 -8.99914801e-02
-6.46176875e-01 -9.80982721e-01 2.57470101e-01 5.38800418e-01
4.97201562e-01 -5.80541492e-01 1.89403951e-01 -8.58690381e-01
-8.32218602e-02 -6.70071363e-01 8.01520109e-01 -2.78176218e-01
-8.89850140e-01 5.68130851e-01 2.92070210e-01 -4.94719863e-01
1.83235303e-01 -9.61025119e-01 -2.80399173e-01 4.81350571e-01
-1.16289532e+00 -8.11648965e-01 7.53596127e-02 9.07597303e-01
-1.47168577e-01 6.67282522e-01 8.29555571e-01 3.52075510e-02
-1.63383693e-01 7.92599380e-01 -2.34229267e-01 -1.17350504e-01
1.38286605e-01 -1.26404846e+00 -1.55027330e-01 1.26033616e+00
1.54784136e-02 8.90426755e-01 7.15804815e-01 -4.04136032e-01
-1.60930061e+00 -9.86014605e-01 1.32517576e+00 2.67592847e-01
8.10102701e-01 -4.16755706e-01 -4.16658133e-01 8.38155806e-01
1.30495325e-01 -1.54373556e-01 8.36000144e-01 2.03206539e-01
-5.59691608e-01 -6.36335433e-01 -1.30907655e+00 7.46585310e-01
1.71558952e+00 -4.94235367e-01 -6.88052252e-02 2.31200129e-01
7.00653136e-01 -1.40041769e-01 -9.87817109e-01 4.06088471e-01
3.82738173e-01 -1.22457874e+00 5.67996502e-01 -1.63650155e-01
-2.70444810e-01 -4.24780309e-01 -3.60882312e-01 -4.88160104e-01
-5.47949255e-01 -9.43364620e-01 2.57253528e-01 8.71683002e-01
6.67172849e-01 -9.11215305e-01 8.91295493e-01 2.97336817e-01
1.13437898e-01 -8.18433702e-01 -1.14176714e+00 -1.02303708e+00
1.04330018e-01 -5.26948690e-01 6.62805259e-01 7.48828888e-01
7.69440711e-01 1.00667804e-01 -1.40025318e-01 1.12223923e-01
6.28542364e-01 5.15215099e-01 8.69485438e-01 -7.86602736e-01
-7.12162316e-01 -6.19805038e-01 -4.98553813e-01 -9.72865343e-01
-6.92280009e-02 -1.13468444e+00 -2.93469697e-01 -1.42399609e+00
5.95707476e-01 -5.57434201e-01 -2.26865292e-01 8.57587457e-01
5.63004315e-01 4.80716854e-01 1.15121650e-02 -1.14202455e-01
-8.51341486e-01 -3.01300049e-01 9.51203287e-01 2.92034447e-01
1.36142112e-02 -2.07602993e-01 -6.15627766e-01 5.66161990e-01
7.34157622e-01 -3.87507945e-01 -2.00840101e-01 -7.51480535e-02
7.09413946e-01 1.63239464e-01 -1.44134928e-03 -8.14207911e-01
1.57627210e-01 -1.75040275e-01 -3.98131788e-01 -5.23678839e-01
-7.38729537e-02 -8.22205722e-01 3.91891360e-01 9.65895593e-01
-2.88156420e-01 3.64801913e-01 3.65298748e-01 4.88222748e-01
1.22297853e-02 -4.15834695e-01 8.10472906e-01 -3.75510246e-01
-4.15381044e-01 4.45925444e-01 -3.45101982e-01 -2.43691146e-01
1.04735279e+00 -4.60403591e-01 -5.34892261e-01 -2.33545616e-01
-8.30661714e-01 2.77737409e-01 6.58153832e-01 -6.21924520e-01
2.57237583e-01 -1.01381850e+00 -1.56510711e-01 1.61648523e-02
2.03176379e-01 -1.83554307e-01 2.19857082e-01 9.55459237e-01
-9.92784917e-01 5.55010557e-01 1.48254499e-01 -6.65327460e-02
-1.61645198e+00 8.18883121e-01 1.85883537e-01 -9.63048816e-01
-3.73604983e-01 7.84881234e-01 3.17907594e-02 -2.08861068e-01
1.10533878e-01 -5.15279531e-01 6.79442346e-01 -6.32588446e-01
3.61427516e-01 6.47835791e-01 -1.56654626e-01 -2.52206951e-01
-6.38525307e-01 9.78413403e-01 -8.90090615e-02 -1.46418840e-01
1.21161413e+00 -2.12456644e-01 -1.00584424e+00 9.33903456e-02
1.27427244e+00 6.73481405e-01 -3.99928719e-01 -1.67761937e-01
1.63073644e-01 -2.64653802e-01 -5.01251936e-01 -4.16182965e-01
-1.01605594e+00 3.81419271e-01 1.06093764e-01 7.34174371e-01
1.57992208e+00 4.93646353e-01 7.95174062e-01 3.89696568e-01
1.10561430e+00 -9.42518055e-01 -2.76614994e-01 2.33761296e-01
2.94511437e-01 -1.33094370e-01 1.95959359e-01 -8.86820257e-01
3.12162995e-01 1.25072753e+00 4.99257356e-01 -3.49583030e-01
6.14126027e-01 7.21405268e-01 -7.09968984e-01 -2.19506219e-01
-6.70949697e-01 -5.99128008e-01 -3.39788795e-01 -1.17587999e-01
2.86913365e-01 1.44743890e-01 -1.15841126e+00 5.46480477e-01
-2.20100537e-01 8.80455226e-02 6.27985001e-01 1.03880870e+00
-8.35159063e-01 -1.79498887e+00 -2.77036242e-02 3.29658180e-01
-4.83462095e-01 1.92895215e-02 -5.57626605e-01 1.22489834e+00
2.96804249e-01 8.47569346e-01 -7.34860152e-02 -4.29678440e-01
2.35882059e-01 3.92471720e-03 1.08422577e+00 -6.24391913e-01
-5.22421539e-01 2.72585273e-01 5.40296614e-01 -8.99312437e-01
-6.49224699e-01 -2.07331315e-01 -1.32282376e+00 -1.06359160e+00
-6.12469256e-01 5.99984348e-01 1.31902233e-01 6.64547145e-01
1.15882874e-01 -2.25581434e-02 6.20608866e-01 2.04065517e-02
-1.92160264e-01 -4.30292130e-01 -1.17645085e+00 -1.27740532e-01
-2.25023255e-01 -6.77815080e-02 -7.90532351e-01 -3.83611888e-01] | [6.892168045043945, 5.204005718231201] |
e3a27ea7-8295-48cc-bcff-cfe837b215ef | local-discriminant-hyperalignment-for-multi | 1611.08366 | null | http://arxiv.org/abs/1611.08366v1 | http://arxiv.org/pdf/1611.08366v1.pdf | Local Discriminant Hyperalignment for multi-subject fMRI data alignment | Multivariate Pattern (MVP) classification can map different cognitive states
to the brain tasks. One of the main challenges in MVP analysis is validating
the generated results across subjects. However, analyzing multi-subject fMRI
data requires accurate functional alignments between neuronal activities of
different subjects, which can rapidly increase the performance and robustness
of the final results. Hyperalignment (HA) is one of the most effective
functional alignment methods, which can be mathematically formulated by the
Canonical Correlation Analysis (CCA) methods. Since HA mostly uses the
unsupervised CCA techniques, its solution may not be optimized for MVP
analysis. By incorporating the idea of Local Discriminant Analysis (LDA) into
CCA, this paper proposes Local Discriminant Hyperalignment (LDHA) as a novel
supervised HA method, which can provide better functional alignment for MVP
analysis. Indeed, the locality is defined based on the stimuli categories in
the train-set, where the correlation between all stimuli in the same category
will be maximized and the correlation between distinct categories of stimuli
approaches to near zero. Experimental studies on multi-subject MVP analysis
confirm that the LDHA method achieves superior performance to other
state-of-the-art HA algorithms. | ['Muhammad Yousefnezhad', 'Daoqiang Zhang'] | 2016-11-25 | null | null | null | null | ['multi-subject-fmri-data-alignment'] | ['medical'] | [-2.19100546e-02 -8.02633882e-01 1.15981199e-01 -4.24677432e-01
-4.68482196e-01 -4.61275578e-01 5.88502526e-01 -5.80604114e-02
-4.10742670e-01 6.02836490e-01 2.60195136e-01 1.86096489e-01
-3.97976577e-01 -3.35616410e-01 -3.36555600e-01 -1.23955727e+00
-1.62389681e-01 1.47480145e-01 1.85881242e-01 7.20568895e-02
2.77762979e-01 3.36796314e-01 -1.20968568e+00 4.24683452e-01
8.63475621e-01 8.00894439e-01 4.87707913e-01 2.48637646e-02
4.20973375e-02 1.47785291e-01 -3.19364965e-01 -4.33095098e-02
2.59502739e-01 -7.88085997e-01 -4.31877524e-01 -2.76798368e-01
7.78251514e-02 2.88356274e-01 4.53381240e-02 1.20172668e+00
5.91329396e-01 2.45690659e-01 8.12647164e-01 -1.40962529e+00
-1.93669632e-01 4.09737080e-01 -7.35599518e-01 6.31773472e-01
6.36209548e-02 1.01159588e-01 9.47213531e-01 -1.03464401e+00
3.77843618e-01 1.19350731e+00 1.80488825e-01 1.83272988e-01
-1.63568532e+00 -9.29647267e-01 2.89336219e-02 6.17427647e-01
-1.66222417e+00 -2.50770241e-01 8.00231397e-01 -7.70572603e-01
4.55579281e-01 2.83082724e-01 7.71137536e-01 1.05497158e+00
8.10747206e-01 5.45703173e-01 1.71798062e+00 -2.10839003e-01
4.02951390e-01 -4.98779789e-02 2.06289545e-01 2.46289968e-01
6.82205558e-02 -2.54798472e-01 -6.99411273e-01 -1.01234026e-01
5.71609437e-01 -1.33868203e-01 -3.88190389e-01 -5.86939812e-01
-1.73372221e+00 5.39082825e-01 6.40278161e-01 8.27806115e-01
-6.02294922e-01 -2.92411804e-01 2.57145405e-01 -7.48176202e-02
1.28809273e-01 4.33513165e-01 -1.40693784e-01 2.05681082e-02
-9.68999386e-01 1.75117657e-01 8.54540840e-02 3.09030890e-01
7.37961352e-01 -4.55000140e-02 -3.50440055e-01 9.60698128e-01
4.83559698e-01 5.31859875e-01 9.81006563e-01 -7.25712776e-01
3.58776867e-01 6.31363630e-01 -3.11000049e-01 -1.41753781e+00
-5.98774195e-01 -7.10735381e-01 -8.92701447e-01 -3.52186221e-03
3.87131453e-01 9.24452767e-02 -3.30373347e-01 1.98448050e+00
1.05118558e-01 1.03038192e-01 -8.81359950e-02 9.90047574e-01
4.16408598e-01 5.10434270e-01 2.54181534e-01 -6.92210078e-01
1.33606243e+00 -5.14738560e-01 -6.92695618e-01 -3.80017072e-01
3.14815253e-01 -5.09856641e-01 1.05245876e+00 3.01820785e-01
-5.99549174e-01 -7.85448909e-01 -1.17121494e+00 4.70407128e-01
-1.19268693e-01 1.10794440e-01 5.25361955e-01 3.54138136e-01
-6.98018551e-01 3.47364515e-01 -9.23259676e-01 -3.19006890e-01
3.66579413e-01 2.55503297e-01 -7.46366918e-01 1.40678482e-02
-9.06621397e-01 8.79493654e-01 4.62782145e-01 3.88796836e-01
-9.43922162e-01 -4.03964996e-01 -4.41369265e-01 8.97802487e-02
-6.18602848e-03 -2.13330075e-01 4.85660493e-01 -9.76099253e-01
-1.09014487e+00 6.24113739e-01 -4.52792197e-01 9.04631466e-02
2.91441511e-02 9.91786122e-02 -3.42887551e-01 -2.58270483e-02
4.48573828e-01 5.78692853e-01 7.61229455e-01 -1.07709587e+00
-2.74371430e-02 -6.15380645e-01 -6.08885705e-01 3.18619788e-01
-4.56374913e-01 8.61727670e-02 -9.06659067e-02 -4.27924633e-01
8.05868447e-01 -9.37132716e-01 1.01535292e-02 -4.46216911e-01
-1.57770157e-01 -2.95505017e-01 3.92680317e-01 -7.97927260e-01
1.11505830e+00 -2.37427950e+00 5.88485062e-01 5.86165428e-01
2.27279529e-01 -2.07004428e-01 -3.92259732e-02 5.87660335e-02
-4.39339817e-01 -1.30344227e-01 -3.95207763e-01 8.85940641e-02
-2.00085238e-01 1.27410755e-01 1.35312483e-01 7.89908946e-01
4.19648252e-02 7.92549014e-01 -7.10450053e-01 -6.70877755e-01
1.99697376e-03 1.71542481e-01 -4.72655267e-01 1.96244061e-01
3.27753395e-01 1.02373862e+00 -2.92631835e-01 3.71148765e-01
6.90145850e-01 -2.05293875e-02 6.87646151e-01 -6.49967968e-01
-2.61163235e-01 -7.18250424e-02 -1.15459633e+00 1.57916021e+00
4.68323641e-02 6.30568564e-01 -2.02243641e-01 -1.34816742e+00
1.04382193e+00 3.25825065e-01 7.70629764e-01 -8.59221995e-01
2.23534346e-01 2.31385067e-01 7.37141192e-01 -4.09775436e-01
-2.16760278e-01 -9.55983717e-03 3.46140228e-02 3.88231754e-01
2.56311595e-01 4.45732176e-01 1.90361544e-01 -8.38834792e-02
8.53202641e-01 2.44738515e-02 5.78185499e-01 -9.32557166e-01
8.78209651e-01 -4.57331449e-01 1.02158237e+00 4.40160215e-01
-5.13021350e-01 2.12179080e-01 6.76557839e-01 -1.40548527e-01
-8.46270263e-01 -1.25423086e+00 -4.52173203e-01 7.75375903e-01
-1.66763067e-02 -4.05217737e-01 -8.14475894e-01 -3.11828107e-01
-3.66068244e-01 2.42182970e-01 -5.89783609e-01 -9.94255766e-02
-5.67499995e-01 -1.25483000e+00 -1.76676456e-02 5.14778435e-01
8.12973797e-01 -8.85301650e-01 -3.54887813e-01 9.42380577e-02
-5.84205389e-01 -1.00795949e+00 -1.01713635e-01 5.64390607e-02
-8.23520005e-01 -1.00358784e+00 -5.33769786e-01 -7.64905691e-01
8.26079547e-01 4.26965028e-01 4.57011938e-01 -3.03944618e-01
-5.36528677e-02 7.01047108e-02 -2.85678059e-01 1.52267531e-01
-7.58730322e-02 -1.70631513e-01 4.41469818e-01 5.63533902e-01
4.11128581e-01 -9.31140065e-01 -7.14067221e-01 5.09314716e-01
-5.37527800e-01 1.24934517e-01 8.10629547e-01 8.37282479e-01
8.83615911e-01 8.56295452e-02 4.89053041e-01 -3.28162640e-01
6.53880239e-01 -5.03595948e-01 -2.55978167e-01 3.31492931e-01
-5.48836052e-01 1.13373108e-01 4.68901992e-01 -5.39608717e-01
-9.11149442e-01 1.56796537e-02 2.21700415e-01 -2.58499533e-01
-3.74929532e-02 6.52237594e-01 -5.61870873e-01 -1.42372176e-01
8.09618175e-01 6.79274619e-01 -4.01919940e-03 -3.95734251e-01
7.67441466e-02 4.41790253e-01 4.97814447e-01 -5.37427843e-01
4.92144555e-01 2.64017463e-01 3.27802122e-01 -7.28636682e-01
-4.09243673e-01 -3.78905445e-01 -1.23987639e+00 -6.86380565e-01
1.04637492e+00 -7.44257808e-01 -7.29888976e-01 3.28874588e-01
-7.43314326e-01 -2.30025575e-02 4.03854042e-01 8.93544257e-01
-3.31041425e-01 6.72874928e-01 -1.34135842e-01 -6.12113118e-01
-1.96642458e-01 -1.36325359e+00 5.11644602e-01 1.75288379e-01
-3.06376278e-01 -6.82879686e-01 2.55915999e-01 2.96145260e-01
3.04550439e-01 3.43316346e-01 1.14450121e+00 -7.33069599e-01
-1.17389701e-01 -1.09175228e-01 9.71958712e-02 3.91602606e-01
6.73722923e-02 -2.61856079e-01 -8.65031421e-01 -2.34486505e-01
3.87773484e-01 1.80562913e-01 4.59133774e-01 4.48130816e-01
1.06374025e+00 -1.46053255e-01 -4.42570895e-01 2.00923473e-01
1.22751606e+00 3.60986114e-01 6.73282862e-01 2.95255035e-01
6.58568680e-01 6.68690205e-01 5.47346234e-01 1.05477519e-01
1.31874308e-01 9.04432118e-01 1.13358803e-01 3.11487585e-01
1.93867534e-01 1.07170992e-01 5.70196271e-01 1.11340749e+00
-1.86134532e-01 3.07287484e-01 -1.02380610e+00 3.42755198e-01
-2.01141930e+00 -1.03268409e+00 -4.83683109e-01 2.46729612e+00
7.10161865e-01 -7.07400218e-02 4.98055778e-02 2.39327863e-01
6.92727149e-01 8.48624632e-02 -2.40044907e-01 8.31098184e-02
-2.65769780e-01 1.69283718e-01 -1.22591875e-01 1.25280693e-01
-9.62150633e-01 5.05286276e-01 6.01570654e+00 7.17771113e-01
-1.05055964e+00 5.40844679e-01 3.52188408e-01 5.15373796e-02
9.40565765e-02 9.14995447e-02 -3.94424945e-01 7.41804242e-01
8.40448320e-01 -2.61632591e-01 5.27267396e-01 7.88429141e-01
4.71985847e-01 -2.84892619e-01 -1.05007827e+00 1.20120347e+00
1.78515092e-01 -7.34306097e-01 -2.46071473e-01 7.55345970e-02
6.48772418e-01 -8.85406062e-02 2.84710038e-03 1.15209386e-01
-4.67867315e-01 -6.86112761e-01 4.69401896e-01 7.45445132e-01
1.63555726e-01 -5.40941119e-01 7.74635434e-01 3.87979031e-01
-1.28618050e+00 -3.81991453e-02 -6.63899839e-01 8.49572420e-02
-1.07645795e-01 6.51359141e-01 -4.91389066e-01 3.78045738e-01
7.76902139e-01 8.04819584e-01 -8.66738975e-01 1.00288439e+00
-3.48124087e-01 4.96177405e-01 -6.47882149e-02 5.54304617e-03
-2.21410885e-01 -4.05037254e-01 6.07167602e-01 9.29342508e-01
2.16366947e-01 -4.74230982e-02 2.43230462e-01 9.64594424e-01
3.78582746e-01 6.30852699e-01 -3.57278228e-01 9.60185453e-02
3.68168116e-01 1.51685488e+00 -8.87180150e-01 -8.47745985e-02
-3.33705127e-01 6.32483900e-01 2.65072584e-01 1.88836828e-01
-6.64449275e-01 1.15005322e-01 5.35410702e-01 -1.27420083e-01
-1.82024226e-01 -6.40116811e-01 -2.97257304e-01 -1.17334473e+00
1.28254309e-01 -8.34388018e-01 4.25251842e-01 -8.65236044e-01
-1.45548749e+00 4.72005814e-01 3.17149490e-01 -1.41137075e+00
2.23707338e-03 -5.92678070e-01 -5.95429897e-01 9.75756347e-01
-8.67071211e-01 -8.89403343e-01 -4.85530049e-01 8.74788702e-01
3.25188518e-01 -2.44369492e-01 8.75149250e-01 2.89754808e-01
-7.75894701e-01 3.20180535e-01 2.02269480e-01 3.16637084e-02
7.78529823e-01 -7.74193108e-01 -5.25129735e-01 1.38648832e+00
1.70459718e-01 8.61509800e-01 4.90879059e-01 -6.21548891e-01
-1.13700604e+00 -6.20151043e-01 5.78413606e-01 -2.50174284e-01
6.17452979e-01 -4.46391016e-01 -1.11297166e+00 4.58459437e-01
9.64606479e-02 -1.20814390e-01 1.14229023e+00 5.67186028e-02
-1.57503322e-01 -4.04269248e-01 -7.35544741e-01 6.19104803e-01
7.23964036e-01 -4.74957615e-01 -6.26227796e-01 5.41051745e-01
7.08332611e-03 9.59752649e-02 -1.21591449e+00 3.35708737e-01
6.73965752e-01 -8.54086578e-01 7.82229304e-01 -4.52720940e-01
1.60957322e-01 -6.47261143e-01 -2.72863805e-01 -1.39518785e+00
-9.18420672e-01 1.31402194e-01 5.72587326e-02 1.28937054e+00
2.30909839e-01 -7.09639132e-01 1.40793070e-01 5.82978785e-01
-1.39645580e-02 -3.51834655e-01 -1.13699818e+00 -8.21954489e-01
-4.29148264e-02 -3.24851930e-01 2.53326893e-01 1.10194957e+00
3.73962849e-01 3.49252045e-01 -3.03947657e-01 1.65133998e-01
6.43732607e-01 1.76672280e-01 5.17204344e-01 -1.32790494e+00
-2.66463995e-01 -4.98202503e-01 -6.44951761e-01 -3.71773750e-01
2.15162218e-01 -1.25553417e+00 2.36359581e-01 -1.30681551e+00
6.15345538e-01 -1.33934587e-01 -6.90547466e-01 5.87246537e-01
-2.91789949e-01 9.59291086e-02 4.37263586e-02 7.81928360e-01
-2.74669051e-01 5.46029568e-01 1.10749066e+00 3.73292863e-02
-2.62976348e-01 -2.74637908e-01 -4.80673462e-01 3.75955433e-01
8.98904979e-01 -5.86149573e-01 -3.44121695e-01 -7.22984038e-03
4.95548099e-02 -3.67795855e-01 2.62584567e-01 -1.36255538e+00
1.11724153e-01 -2.53163010e-01 7.03246653e-01 -2.47304812e-01
1.42663062e-01 -7.46442378e-01 4.40798610e-01 5.53019404e-01
-1.34062856e-01 2.24380657e-01 1.88901257e-02 4.34647501e-01
-1.29869699e-01 -7.90310800e-02 8.09597850e-01 -1.42222419e-01
-6.48040891e-01 4.68031988e-02 -4.46926445e-01 -2.37949073e-01
1.07729936e+00 -6.16125800e-02 -2.13976532e-01 1.15927182e-01
-7.33676314e-01 2.89129280e-02 -5.67488670e-02 3.26355785e-01
4.84522194e-01 -1.58026326e+00 -6.32924497e-01 3.07234883e-01
-3.83687615e-02 -5.46684265e-01 4.86261278e-01 1.59244049e+00
-1.20285690e-01 5.30232728e-01 -8.10189486e-01 -1.02121818e+00
-1.20329714e+00 4.27669406e-01 2.54236728e-01 -1.57613128e-01
-4.36739653e-01 3.87831956e-01 5.32419860e-01 4.46680896e-02
-4.19231087e-01 1.34804696e-01 -6.91013813e-01 3.06892246e-01
5.39293408e-01 2.66165048e-01 1.83069389e-02 -8.78600836e-01
-7.48128533e-01 5.32197356e-01 5.58404401e-02 -1.51102975e-01
1.35476184e+00 -1.00628465e-01 -5.83659053e-01 6.98853910e-01
1.06987226e+00 -4.48985621e-02 -9.84440684e-01 -2.08249792e-01
-1.61412477e-01 -5.53088367e-01 7.95320123e-02 -5.45035541e-01
-1.01661515e+00 9.39415514e-01 9.57841635e-01 -9.12207812e-02
1.10729682e+00 -7.33620748e-02 7.81058744e-02 2.00372547e-01
4.63831633e-01 -1.05314243e+00 3.14094663e-01 3.46400104e-02
1.11140418e+00 -1.01296878e+00 1.61782816e-01 -2.07409352e-01
-8.52047086e-01 1.06864440e+00 7.63394296e-01 -1.15349971e-01
7.43079603e-01 -2.96848685e-01 -3.33797574e-01 -2.47113243e-01
-2.37121552e-01 1.19479403e-01 5.23974478e-01 3.91202241e-01
4.35388952e-01 2.24675193e-01 -9.48657990e-01 9.77860570e-01
-1.50679216e-01 -5.41759193e-01 1.38347089e-01 6.81672394e-01
-3.90155852e-01 -1.06238663e+00 -4.12792146e-01 5.05735278e-01
-2.36543998e-01 1.26330703e-01 -7.10108578e-02 6.68363333e-01
2.74257287e-02 7.55364120e-01 -7.95784872e-03 -5.03188670e-01
1.42862290e-01 5.78866422e-01 4.47752237e-01 -4.11286056e-01
-3.48446965e-01 3.59542787e-01 -4.91622716e-01 -6.80720270e-01
-7.04073429e-01 -1.05904174e+00 -1.21954298e+00 -8.90097544e-02
-1.47931695e-01 9.78983194e-02 6.44074678e-01 1.18324721e+00
2.60871559e-01 4.80940670e-01 7.71348715e-01 -7.43125439e-01
2.96168514e-02 -1.11980164e+00 -6.20173872e-01 4.92342651e-01
-3.42480332e-01 -1.20631611e+00 -3.35908145e-01 9.95334610e-02] | [12.658282279968262, 3.3936026096343994] |
58c195a5-d7ad-43c1-82f4-9352c21e8e83 | what-you-see-is-what-you-read-improving-text | 2305.10400 | null | https://arxiv.org/abs/2305.10400v2 | https://arxiv.org/pdf/2305.10400v2.pdf | What You See is What You Read? Improving Text-Image Alignment Evaluation | Automatically determining whether a text and a corresponding image are semantically aligned is a significant challenge for vision-language models, with applications in generative text-to-image and image-to-text tasks. In this work, we study methods for automatic text-image alignment evaluation. We first introduce SeeTRUE: a comprehensive evaluation set, spanning multiple datasets from both text-to-image and image-to-text generation tasks, with human judgements for whether a given text-image pair is semantically aligned. We then describe two automatic methods to determine alignment: the first involving a pipeline based on question generation and visual question answering models, and the second employing an end-to-end classification approach by finetuning multimodal pretrained models. Both methods surpass prior approaches in various text-image alignment tasks, with significant improvements in challenging cases that involve complex composition or unnatural images. Finally, we demonstrate how our approaches can localize specific misalignments between an image and a given text, and how they can be used to automatically re-rank candidates in text-to-image generation. | ['Idan Szpektor', 'Eran Ofek', 'Oran Lang', 'Jonathan Herzig', 'Roee Aharoni', 'Soravit Changpinyo', 'Yonatan Bitton', 'Michal Yarom'] | 2023-05-17 | null | null | null | null | ['visual-reasoning', 'question-generation', 'visual-reasoning'] | ['computer-vision', 'natural-language-processing', 'reasoning'] | [ 8.88431311e-01 1.21363215e-01 2.99969465e-01 -6.25013709e-01
-1.44653881e+00 -8.43823016e-01 1.14616334e+00 1.74513429e-01
-3.69527042e-01 3.21723342e-01 2.88902730e-01 -2.58296549e-01
4.67887133e-01 -3.35369080e-01 -9.18115616e-01 -4.29782182e-01
6.91899717e-01 9.84558344e-01 2.29413211e-01 -1.17002651e-01
3.83105487e-01 7.59482458e-02 -1.56164873e+00 7.66904473e-01
6.10599458e-01 8.76353323e-01 3.60347062e-01 1.03009033e+00
-3.35258037e-01 7.32833028e-01 -5.16706884e-01 -8.06005955e-01
1.92538917e-01 -8.56660247e-01 -1.07469904e+00 5.10043740e-01
1.58844197e+00 -2.15988725e-01 1.98296294e-01 9.78404045e-01
5.34042656e-01 -1.32136092e-01 9.23912108e-01 -1.38792622e+00
-7.32563853e-01 4.06359345e-01 -5.55506051e-01 -5.63766481e-03
5.55566907e-01 3.08726072e-01 1.11851645e+00 -1.34707797e+00
9.12323415e-01 1.28581834e+00 4.81536746e-01 3.89573395e-01
-1.41297567e+00 -2.16217160e-01 -6.02253936e-02 1.21730305e-01
-1.05905616e+00 -8.65217209e-01 4.56120193e-01 -7.93811619e-01
9.31589782e-01 2.01838091e-01 4.94101882e-01 9.45548832e-01
7.71337524e-02 9.25783634e-01 1.12605548e+00 -7.11492181e-01
-1.73715726e-01 1.89955398e-01 -5.31875908e-01 7.95570314e-01
-8.42497125e-02 -5.04162759e-02 -4.74594504e-01 -4.96861786e-02
5.53468704e-01 -5.42939782e-01 -1.29830688e-01 -5.00369251e-01
-1.62116647e+00 7.63756394e-01 2.40541235e-01 3.05784434e-01
-2.76339889e-01 3.49331170e-01 2.22463578e-01 1.24232836e-01
2.94670999e-01 7.15571463e-01 -4.25777920e-02 2.37127706e-01
-1.36164296e+00 5.44948399e-01 5.83270490e-01 1.05001986e+00
7.48649478e-01 4.76181395e-02 -6.85549855e-01 7.22241163e-01
2.54709750e-01 6.74795270e-01 3.90249342e-01 -8.61989796e-01
6.56926155e-01 3.35567504e-01 6.68189079e-02 -8.54497194e-01
-9.13075954e-02 -6.50381949e-03 -4.58193183e-01 1.37433931e-01
5.16507089e-01 1.88285738e-01 -1.12990510e+00 1.59649718e+00
2.57161736e-01 -1.96024954e-01 1.99735269e-01 8.48723531e-01
1.00904322e+00 4.21365350e-01 1.62649482e-01 -3.04304864e-02
1.65626454e+00 -1.33601367e+00 -5.27764976e-01 -7.21975744e-01
3.70840043e-01 -1.53691030e+00 1.17943251e+00 1.25423810e-02
-1.39563119e+00 -6.82535827e-01 -5.94001293e-01 -3.68455231e-01
-2.42478043e-01 5.36188424e-01 -9.74833816e-02 3.48831326e-01
-1.42500746e+00 9.96691212e-02 -3.29764128e-01 -7.72213280e-01
2.90459007e-01 -9.00166258e-02 -3.47805947e-01 -1.36286482e-01
-6.65333033e-01 1.01168776e+00 4.60767716e-01 -2.64023393e-01
-1.09142315e+00 -4.07487333e-01 -9.88359153e-01 -3.06679904e-01
9.61022675e-02 -1.32197785e+00 1.60623813e+00 -1.55065465e+00
-1.04954839e+00 1.63619888e+00 -4.90919024e-01 -4.21081126e-01
6.06607735e-01 1.23283669e-01 -1.07371189e-01 4.47830975e-01
5.94124556e-01 1.38378263e+00 1.46382129e+00 -1.64329255e+00
-7.15107977e-01 -2.83213288e-01 -2.01984614e-01 6.00040972e-01
8.22169632e-02 2.55381733e-01 -7.18435407e-01 -8.12789619e-01
4.22406718e-02 -9.43274558e-01 -2.07792044e-01 7.99521506e-02
-7.25861430e-01 -1.34388760e-01 7.37444282e-01 -8.53176415e-01
5.95774710e-01 -1.47947276e+00 2.14781910e-01 -9.30050984e-02
8.83949772e-02 5.98269794e-03 -5.97910047e-01 5.25936425e-01
-1.98072717e-01 8.14786851e-02 -3.44832540e-01 -7.00864673e-01
-3.89923491e-02 -6.80137100e-03 -5.49128890e-01 8.06624517e-02
5.30379355e-01 1.33072281e+00 -8.02695990e-01 -9.34566557e-01
2.62015879e-01 8.74019712e-02 -3.95096809e-01 5.67292392e-01
-6.22146964e-01 5.36395669e-01 -5.13185374e-02 6.40333474e-01
3.24912965e-01 -3.34714204e-01 -7.14342594e-02 -6.03985369e-01
7.00519234e-02 -2.71046665e-02 -6.21883273e-01 1.57424641e+00
-3.97344857e-01 9.08181906e-01 -2.44799955e-03 -8.16918492e-01
5.92699409e-01 1.88936830e-01 2.48912111e-01 -8.02528083e-01
1.17700301e-01 1.53809965e-01 -3.02991658e-01 -7.72746861e-01
9.05409515e-01 -2.28048012e-01 -1.48445621e-01 7.03575611e-01
2.21234426e-01 -9.42659080e-01 5.91078222e-01 4.51221526e-01
5.04812121e-01 3.36968839e-01 2.27257490e-01 -1.85051002e-02
5.63801110e-01 3.23124588e-01 -3.77970546e-01 9.10973072e-01
8.57601780e-03 1.23167193e+00 2.73981065e-01 -2.64522254e-01
-1.44227505e+00 -9.65983331e-01 1.75930083e-01 1.26573157e+00
9.79418680e-02 -3.54315609e-01 -1.03080475e+00 -6.34300172e-01
-2.42957488e-01 9.79253173e-01 -6.43013537e-01 2.18653142e-01
-3.15340698e-01 -3.99496734e-01 6.35656655e-01 4.44550335e-01
2.37121165e-01 -1.16693020e+00 -4.19915587e-01 -1.40109286e-01
-8.35674345e-01 -1.61191654e+00 -9.52144027e-01 -2.05561593e-01
-5.48476517e-01 -8.94826710e-01 -8.32167685e-01 -1.19814599e+00
9.99479949e-01 3.33184034e-01 1.70587718e+00 1.14564791e-01
-4.70332533e-01 9.55870152e-01 -2.23136052e-01 -4.75125402e-01
-9.32837188e-01 -1.09965697e-01 -3.14685673e-01 8.83804262e-02
-3.52173932e-02 6.94929957e-02 -6.55442357e-01 4.63408172e-01
-1.16354847e+00 6.10494316e-01 6.24474168e-01 8.12147558e-01
8.08201969e-01 -5.42444527e-01 7.94166848e-02 -5.07975519e-01
7.21441746e-01 -4.61306162e-02 -4.04337615e-01 6.75762951e-01
-3.81732076e-01 1.10248722e-01 2.35407993e-01 -2.19353974e-01
-8.76438260e-01 5.01061201e-01 -1.27628669e-01 -5.08247256e-01
-2.86992341e-01 1.63707986e-01 1.90296751e-02 -8.77971351e-02
7.87097394e-01 5.36514044e-01 1.04023322e-01 1.75651520e-01
8.90664160e-01 4.89036828e-01 9.36865151e-01 -6.48010612e-01
9.23047900e-01 5.19844890e-01 -2.68840224e-01 -6.90910995e-01
-1.09711647e+00 -4.75352824e-01 -8.67100298e-01 -3.19614112e-01
1.39589107e+00 -9.41283464e-01 -8.56273472e-02 4.40651089e-01
-1.47861099e+00 -4.10006285e-01 -2.12906241e-01 8.01705196e-02
-9.45135176e-01 4.81377512e-01 -3.02453101e-01 -3.32780182e-01
-6.28194153e-01 -1.40927970e+00 1.81849384e+00 1.64474353e-01
-2.41356701e-01 -1.04703414e+00 -6.36669900e-03 9.97764468e-01
2.80468673e-01 3.07943057e-02 8.10892344e-01 -6.05784774e-01
-6.02870822e-01 -3.41184959e-02 -5.36924422e-01 1.70832649e-01
-6.04328215e-02 8.28021318e-02 -8.75576198e-01 -3.88864160e-01
-3.10184509e-01 -8.16144228e-01 7.28681028e-01 3.64246130e-01
8.63662779e-01 -3.46820295e-01 -1.54601738e-01 3.42398942e-01
1.18805850e+00 -2.53109634e-01 7.96397567e-01 1.44731581e-01
9.32417095e-01 8.56905758e-01 7.54065335e-01 -3.08315642e-02
7.02304602e-01 9.39763069e-01 4.67530221e-01 -6.73478246e-01
-5.90271831e-01 -3.84409517e-01 3.24257076e-01 3.93112749e-01
2.97679186e-01 -5.53369224e-01 -9.43807304e-01 8.07904243e-01
-1.73126674e+00 -1.07077384e+00 -3.83202702e-01 2.15474510e+00
9.99107003e-01 -2.48965889e-01 1.01651154e-01 -5.69023669e-01
8.97737563e-01 3.40488441e-02 -2.94038385e-01 -2.44864568e-01
-2.63891369e-01 5.50566390e-02 2.68499672e-01 6.24827087e-01
-1.22408819e+00 1.48926866e+00 6.97351360e+00 6.92900181e-01
-9.89987493e-01 7.69152567e-02 9.30012465e-01 2.16480866e-01
-3.98242444e-01 1.96406126e-01 -7.09296227e-01 4.80733253e-03
5.21220148e-01 4.99999709e-02 2.24830568e-01 6.69847608e-01
2.24335134e-01 -2.93433964e-01 -1.33854449e+00 1.11710536e+00
8.90022993e-01 -1.35885525e+00 6.60108447e-01 -2.04293251e-01
1.03974545e+00 -3.11398599e-03 7.00747222e-02 -5.76621369e-02
3.84731710e-01 -1.20330715e+00 1.16043735e+00 4.74803805e-01
9.49131787e-01 -1.99498862e-01 5.03728509e-01 -8.18090737e-02
-1.13301933e+00 5.07668197e-01 -1.63800076e-01 5.78473151e-01
4.48453188e-01 2.00128317e-01 -1.47245765e+00 3.73709708e-01
6.05251729e-01 4.04257566e-01 -1.18446088e+00 7.29372323e-01
-3.07521313e-01 2.55083144e-01 2.69043110e-02 7.87730068e-02
2.22371027e-01 -3.93880159e-02 6.23913467e-01 1.33683348e+00
4.00212973e-01 -4.78741765e-01 4.24162835e-01 1.07235730e+00
3.35255824e-02 2.86093950e-01 -6.73883438e-01 -1.78053409e-01
1.12041757e-01 1.55924666e+00 -7.32577980e-01 -5.49836695e-01
-1.53947815e-01 1.52997911e+00 1.10988863e-01 4.02414799e-01
-7.17344940e-01 -1.33883536e-01 1.51588246e-01 2.37827018e-01
3.15040320e-01 -1.09954923e-01 -5.48601560e-02 -1.06996441e+00
-3.80173698e-02 -1.13408685e+00 3.35263103e-01 -1.64989114e+00
-1.29757476e+00 8.48030210e-01 -1.78529182e-03 -1.20250392e+00
-6.90800488e-01 -4.53551292e-01 -5.50660670e-01 7.41167367e-01
-1.29190719e+00 -1.83009934e+00 -6.16545498e-01 7.39368379e-01
7.41071701e-01 -1.76958859e-01 6.34846747e-01 -1.75396763e-02
-2.67499201e-02 5.32841980e-01 -2.14953378e-01 2.09076285e-01
1.04859149e+00 -1.34444571e+00 7.90777802e-01 1.17131305e+00
6.49112940e-01 2.82164961e-01 8.31950963e-01 -5.87642848e-01
-1.07863569e+00 -1.35210073e+00 9.51333404e-01 -8.96333516e-01
5.33676028e-01 -3.88459593e-01 -5.99732459e-01 6.68219864e-01
8.53095472e-01 -1.60669103e-01 3.54053766e-01 -4.38689679e-01
-4.23326194e-01 2.13515535e-01 -7.90587962e-01 9.89606380e-01
8.23883057e-01 -6.46766901e-01 -5.89474797e-01 7.66808867e-01
5.01619875e-01 -6.83854640e-01 -6.30859613e-01 2.24428132e-01
3.91070575e-01 -9.58436072e-01 1.03606367e+00 -5.92347026e-01
8.01592410e-01 -6.04430437e-01 -1.89135641e-01 -1.19384062e+00
-2.91101709e-02 -5.15776396e-01 6.22354448e-01 1.21376514e+00
5.11571229e-01 -5.09346053e-02 4.44927514e-01 2.98495591e-01
-1.08675987e-01 -1.87210828e-01 -7.74879158e-01 -2.97273934e-01
-1.51294461e-02 -2.10847571e-01 1.12716019e-01 8.68313670e-01
-4.63897735e-01 8.95382762e-01 -3.86040002e-01 -9.51344520e-02
5.61104774e-01 3.55874121e-01 1.25975466e+00 -6.77793205e-01
-2.60213852e-01 -7.69421577e-01 -3.63109261e-01 -9.96535778e-01
5.41397221e-02 -1.14986253e+00 3.83315206e-01 -1.80099022e+00
4.27256405e-01 -9.66154188e-02 5.41026473e-01 5.52128375e-01
-4.51505601e-01 7.12923527e-01 2.87399441e-01 2.78084725e-01
-8.26809943e-01 4.48261321e-01 1.30789757e+00 -4.02799189e-01
1.64403290e-01 -3.15095186e-01 -6.76555395e-01 5.64529359e-01
5.47095120e-01 -3.89494836e-01 -3.48395377e-01 -6.95265234e-01
2.16760084e-01 6.14951029e-02 7.52463400e-01 -7.93704331e-01
9.15450379e-02 -1.49184972e-01 4.56106961e-01 -8.14100444e-01
2.78422058e-01 -5.09151995e-01 2.85267904e-02 2.48952419e-01
-6.02961361e-01 4.80333567e-01 1.86294854e-01 4.82659638e-01
-3.08290154e-01 -3.89298975e-01 9.24031675e-01 -1.69164300e-01
-6.54686689e-01 1.27340972e-01 -2.98193365e-01 3.90508801e-01
1.01425099e+00 -2.27967113e-01 -4.59085912e-01 -8.76708746e-01
-5.96239924e-01 2.95410722e-01 7.04915881e-01 6.60566390e-01
8.56681705e-01 -1.35429347e+00 -1.17619562e+00 -5.32530509e-02
7.25329638e-01 -2.30500519e-01 9.04309675e-02 9.01479423e-01
-6.37591124e-01 3.70365381e-01 1.07724080e-02 -1.23896050e+00
-1.79078734e+00 3.09156686e-01 5.87377727e-01 -2.34928757e-01
-1.21618122e-01 8.10486674e-01 5.41143835e-01 -5.67352414e-01
2.41036396e-02 -1.77370280e-01 2.34111361e-02 -2.79835016e-02
4.37976509e-01 -4.52717781e-01 1.44753724e-01 -1.20632434e+00
-1.40354827e-01 9.33266282e-01 -2.17658579e-01 -6.00461125e-01
6.81283057e-01 -3.32586616e-01 -2.71595448e-01 1.21694811e-01
1.04971278e+00 -6.23089820e-02 -9.50952470e-01 -2.94108778e-01
-2.01566666e-01 -3.24827850e-01 -1.96754247e-01 -8.97032142e-01
-9.05517519e-01 9.60212409e-01 6.40689552e-01 -1.39645347e-02
9.73416388e-01 3.90331507e-01 7.14063823e-01 2.89267927e-01
-1.11760885e-01 -1.01550114e+00 8.39474022e-01 4.68167901e-01
1.23852646e+00 -1.56737471e+00 -2.28545256e-02 -1.88502401e-01
-1.16561043e+00 1.07405972e+00 7.39465892e-01 3.38903725e-01
-1.91822369e-02 -6.14557005e-02 4.17900681e-01 -2.55562574e-01
-6.72601044e-01 -5.82182765e-01 8.01732659e-01 6.53351665e-01
5.29613078e-01 -2.07552433e-01 8.21560696e-02 -1.29974797e-01
-3.89885485e-01 -4.09593701e-01 4.65837598e-01 6.53043807e-01
-2.84397572e-01 -1.12464488e+00 -5.92679203e-01 3.81229311e-01
-3.22278351e-01 -5.52791834e-01 -8.80162239e-01 5.20530522e-01
-6.76393583e-02 9.65672910e-01 3.64192516e-01 -1.96826816e-01
9.29258093e-02 4.24154624e-02 8.45262229e-01 -7.03745067e-01
-6.28478765e-01 2.91163355e-01 2.05290601e-01 -3.29929322e-01
-6.68532193e-01 -8.05142283e-01 -1.00081336e+00 3.19179773e-01
-2.94747859e-01 -1.37149453e-01 7.76998341e-01 9.57547665e-01
4.28030878e-01 1.86284289e-01 3.09348315e-01 -1.10939896e+00
-1.48138180e-01 -9.17173803e-01 2.04725236e-01 8.72682691e-01
6.43801689e-02 -1.13300696e-01 -1.28194004e-01 8.99148762e-01] | [10.99161434173584, 1.3196877241134644] |
80864d9c-f6a2-428f-99bc-e2d23c351b70 | atm-fraud-detection-using-streaming-data | 2303.04946 | null | https://arxiv.org/abs/2303.04946v1 | https://arxiv.org/pdf/2303.04946v1.pdf | ATM Fraud Detection using Streaming Data Analytics | Gaining the trust and confidence of customers is the essence of the growth and success of financial institutions and organizations. Of late, the financial industry is significantly impacted by numerous instances of fraudulent activities. Further, owing to the generation of large voluminous datasets, it is highly essential that underlying framework is scalable and meet real time needs. To address this issue, in the study, we proposed ATM fraud detection in static and streaming contexts respectively. In the static context, we investigated a parallel and scalable machine learning algorithms for ATM fraud detection that is built on Spark and trained with a variety of machine learning (ML) models including Naive Bayes (NB), Logistic Regression (LR), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF), Gradient Boosting Tree (GBT), and Multi-layer perceptron (MLP). We also employed several balancing techniques like Synthetic Minority Oversampling Technique (SMOTE) and its variants, Generative Adversarial Networks (GAN), to address the rarity in the dataset. In addition, we proposed a streaming based ATM fraud detection in the streaming context. Our sliding window based method collects ATM transactions that are performed within a specified time interval and then utilizes to train several ML models, including NB, RF, DT, and K-Nearest Neighbour (KNN). We selected these models based on their less model complexity and quicker response time. In both contexts, RF turned out to be the best model. RF obtained the best mean AUC of 0.975 in the static context and mean AUC of 0.910 in the streaming context. RF is also empirically proven to be statistically significant than the next-best performing models. | ['Laveti Ramesh Naidu', 'Abhay Anand Mane', 'Vadlamani Ravi', 'Yelleti Vivek'] | 2023-03-08 | null | null | null | null | ['fraud-detection'] | ['miscellaneous'] | [-6.72903359e-02 -4.07075703e-01 1.04277700e-01 -3.34706694e-01
-4.97751117e-01 -4.59461123e-01 3.12509656e-01 4.14031625e-01
-4.09013361e-01 1.07724881e+00 1.42292948e-02 -5.65189302e-01
-5.73607087e-02 -1.05196023e+00 -4.93806809e-01 -4.50116009e-01
-1.67185869e-02 5.41426003e-01 3.58436882e-01 -2.79557616e-01
5.95697343e-01 6.15708709e-01 -1.25871325e+00 6.95305228e-01
9.78561044e-01 1.24131000e+00 -4.52639550e-01 6.81149304e-01
-1.55918851e-01 1.33262682e+00 -7.30052710e-01 -8.10228109e-01
3.33793759e-01 -6.46880642e-02 -5.14481604e-01 -5.28404117e-01
-2.23857164e-01 -5.65485835e-01 1.37909502e-01 4.06168014e-01
4.52982694e-01 -1.01975687e-01 4.62006509e-01 -1.49668658e+00
-2.31558576e-01 3.33869547e-01 -7.19046116e-01 4.83643502e-01
3.38781059e-01 -1.17005771e-02 4.77978945e-01 -7.60687649e-01
2.44954526e-01 8.03727686e-01 1.14907217e+00 2.46005073e-01
-9.18542027e-01 -8.12923849e-01 -3.62596542e-01 2.04295263e-01
-8.37074876e-01 -8.58753026e-02 6.92572057e-01 -2.98775971e-01
1.21052194e+00 2.47080475e-01 7.45069802e-01 1.00827861e+00
3.80535662e-01 7.33299911e-01 1.29481471e+00 -6.85369551e-01
4.79446769e-01 5.17501235e-01 4.93555330e-02 1.92211509e-01
5.19337475e-01 1.28185432e-02 -5.64900875e-01 -8.28519881e-01
6.36933327e-01 4.29612905e-01 2.65207916e-01 1.55999377e-01
-7.01041043e-01 1.04710698e+00 4.69034761e-02 3.48514169e-01
-6.99447215e-01 -1.12650460e-02 8.80583465e-01 5.04524291e-01
2.46590510e-01 -1.03855121e-03 -5.47584713e-01 -3.98866683e-01
-1.11593127e+00 3.61502945e-01 1.03645384e+00 3.96936715e-01
2.56825417e-01 2.86028802e-01 -3.09916288e-02 7.11197078e-01
2.23836705e-01 1.02609046e-01 9.87212777e-01 -4.53509301e-01
7.51565635e-01 8.24240506e-01 3.27856898e-01 -8.65094841e-01
-1.20306373e-01 -4.49671000e-01 -9.11984265e-01 9.03648585e-02
2.79648095e-01 -4.16204967e-02 -4.92605090e-01 1.14648199e+00
3.37240160e-01 3.21974427e-01 2.56742358e-01 5.66608429e-01
1.76631793e-01 4.91029918e-01 1.47005230e-01 -3.00382137e-01
1.06688476e+00 -6.53014064e-01 -3.65376264e-01 1.69097394e-01
7.05984056e-01 -9.57207263e-01 7.87705004e-01 7.62372196e-01
-8.98641348e-01 -2.76036650e-01 -1.00217819e+00 5.40503740e-01
-3.64628404e-01 -3.21860641e-01 8.20478022e-01 1.12973523e+00
-8.53885114e-01 5.02093196e-01 -7.72459984e-01 -3.07595074e-01
5.48527122e-01 2.66471863e-01 -1.10501997e-01 -1.44574270e-01
-1.14143908e+00 7.05688834e-01 2.39306744e-02 5.38233891e-02
-5.27350664e-01 -2.56510139e-01 -2.77367383e-01 -2.06275582e-01
-1.38589710e-01 -6.54124200e-01 1.12805986e+00 -1.17304802e+00
-1.15994346e+00 7.15672135e-01 -3.05279251e-02 -1.09456789e+00
9.91891921e-01 -1.56063840e-01 -5.83204687e-01 8.20159242e-02
-1.38072595e-01 -1.77012593e-01 7.16757774e-01 -8.47618818e-01
-6.18411601e-01 -5.77681363e-01 -2.56577283e-01 -9.63512212e-02
-2.24418327e-01 3.98266286e-01 6.32561505e-01 -7.75983512e-01
1.89520791e-01 -4.93636966e-01 -4.32373993e-02 -5.85395396e-01
-1.89559713e-01 -8.22916254e-02 1.00531411e+00 -1.09141529e+00
1.31973004e+00 -1.91303980e+00 -7.23372996e-01 4.33800638e-01
-3.78764480e-01 3.51451010e-01 7.65561283e-01 8.65391135e-01
-3.68486978e-02 7.94980675e-02 -2.36465767e-01 -9.63076577e-03
-2.24274814e-01 2.02414796e-01 -6.23506308e-01 2.05774933e-01
1.28201604e-01 6.35625899e-01 -5.03529906e-01 -5.25797307e-01
1.02777489e-01 1.80006057e-01 -4.83531684e-01 3.11200112e-01
1.65803283e-01 2.06272304e-01 -3.66689295e-01 1.23894072e+00
7.01543748e-01 1.13979585e-01 -1.21468000e-01 8.78129080e-02
2.70383477e-01 -2.38798968e-02 -1.45001006e+00 9.36568856e-01
-3.75348121e-01 1.13387562e-01 -2.17921525e-01 -1.32871032e+00
1.46305943e+00 5.92774510e-01 4.42267448e-01 -7.12542415e-01
-3.06129634e-01 6.76815987e-01 -4.17663217e-01 -7.17013359e-01
4.23616886e-01 -4.19125050e-01 8.68953764e-02 6.46387875e-01
-4.33212489e-01 3.31309706e-01 -2.77778715e-01 1.18737034e-01
1.25110090e+00 6.69820160e-02 4.35455948e-01 2.44906932e-01
5.07708251e-01 1.31169960e-01 5.58414042e-01 6.16023064e-01
-5.10738611e-01 4.95339066e-01 6.12207353e-01 -8.65714133e-01
-1.07639456e+00 -9.48526800e-01 -1.01511046e-01 7.35305130e-01
-2.97605008e-01 2.83788472e-01 -3.78905535e-01 -6.00213706e-01
3.65597844e-01 8.44331682e-01 -2.62407869e-01 -9.55312848e-02
-6.24037206e-01 -1.06116557e+00 8.61510336e-01 5.09751141e-01
8.73197496e-01 -1.23513234e+00 -8.39626849e-01 5.52453220e-01
-2.86430176e-02 -8.23796213e-01 1.38971522e-01 1.62889913e-01
-1.37049556e+00 -1.01460731e+00 -4.60184485e-01 -3.37353081e-01
2.57490993e-01 -1.49781495e-01 1.05874383e+00 3.99106964e-02
-4.42820698e-01 -1.71907708e-01 -5.78613460e-01 -6.87225401e-01
-4.73109990e-01 -2.85340101e-01 2.24740747e-02 1.82967126e-01
6.39316022e-01 -7.32536912e-01 -7.88718998e-01 2.57522374e-01
-1.04621565e+00 -3.68337631e-01 4.62340772e-01 9.22773123e-01
2.57185459e-01 1.20610602e-01 1.22809100e+00 -1.03210127e+00
7.57186472e-01 -8.21499050e-01 -2.25969449e-01 1.20403238e-01
-8.72043192e-01 -3.70644957e-01 6.99444592e-01 -3.52381349e-01
-9.75796342e-01 -3.01818460e-01 3.14969681e-02 -1.93086639e-01
-1.78687185e-01 4.67766464e-01 2.87585020e-01 1.41536191e-01
8.15337002e-01 2.27571979e-01 -1.68625005e-02 -3.15432668e-01
-3.26830029e-01 1.26568568e+00 1.09424278e-01 -4.97540265e-01
2.36306384e-01 5.42914510e-01 -2.15677068e-01 -4.04210508e-01
-3.31253976e-01 -6.14490747e-01 -2.51956463e-01 -1.13795370e-01
2.44694024e-01 -8.11459541e-01 -5.51315367e-01 9.47135746e-01
-1.02142382e+00 7.45614842e-02 -1.84431538e-01 4.12613809e-01
-3.69891793e-01 4.37409699e-01 -8.92458797e-01 -1.51287997e+00
-9.35650468e-01 -5.18935680e-01 4.80744779e-01 1.32304326e-01
-2.90931374e-01 -7.85218179e-01 1.15616314e-01 6.77368760e-01
7.21474648e-01 1.00335479e+00 8.36870968e-01 -1.16509509e+00
-3.13054919e-01 -8.28617334e-01 -2.65139103e-01 6.47724569e-01
-6.00234568e-02 2.13794410e-02 -9.89389658e-01 -2.58070529e-01
4.08282608e-01 -3.94949138e-01 5.67159891e-01 2.65104622e-01
8.47080588e-01 -4.80323583e-01 -1.79879367e-02 1.73082933e-01
1.85253191e+00 4.64165717e-01 7.96681583e-01 8.21869135e-01
4.81064767e-02 3.57639641e-01 7.58365214e-01 7.13722408e-01
2.09429413e-01 4.29311275e-01 3.63042206e-01 3.17549556e-02
6.58092678e-01 -2.60108471e-01 4.44076806e-01 3.98657560e-01
-1.87005192e-01 6.80624470e-02 -9.70336020e-01 4.09527630e-01
-1.82272458e+00 -1.17055690e+00 -2.01634914e-01 2.56703210e+00
6.67909920e-01 5.83694100e-01 5.47457695e-01 8.37109566e-01
6.82693243e-01 -4.32083577e-01 -5.02869427e-01 -1.18461812e+00
-1.25594273e-01 3.87568980e-01 6.57493830e-01 -2.12476268e-01
-7.99864471e-01 3.97500873e-01 5.36685371e+00 5.62112987e-01
-9.94712472e-01 2.44821236e-01 1.00325656e+00 -6.92709014e-02
4.28873263e-02 8.45262036e-02 -5.55990279e-01 8.47451925e-01
1.27389193e+00 5.07114716e-02 2.05157965e-01 1.00155342e+00
3.32739979e-01 -3.47220391e-01 -7.01421857e-01 6.45203292e-01
-2.15778887e-01 -1.10847461e+00 -1.11158296e-01 1.40315574e-02
5.80479622e-01 1.64699722e-02 -4.12600219e-01 3.65233630e-01
2.54046410e-01 -1.08851600e+00 5.06710052e-01 6.70444906e-01
3.92494380e-01 -1.03615904e+00 1.39601982e+00 5.35321295e-01
-6.91253424e-01 -4.28492308e-01 -2.58132130e-01 -2.90971965e-01
6.24052212e-02 9.69564199e-01 -8.90029490e-01 5.90075135e-01
1.00585783e+00 1.05815724e-01 -1.53576270e-01 1.04892349e+00
3.66981477e-01 1.09480250e+00 -5.65157712e-01 -8.62179548e-02
9.81009752e-02 -1.87358350e-01 1.76577792e-01 1.20947838e+00
2.94971973e-01 -9.45664570e-02 -1.26580372e-01 4.09239829e-01
2.23263159e-01 4.47587401e-01 -5.61607420e-01 4.19742405e-01
5.15277803e-01 9.24251974e-01 -6.47555053e-01 -2.86387563e-01
-3.97144765e-01 7.23424494e-01 6.99596899e-03 1.26621470e-01
-8.28388810e-01 -4.19746578e-01 -4.53062989e-02 4.31177527e-01
2.34151021e-01 6.99940473e-02 -6.72545075e-01 -8.61432195e-01
4.82780367e-01 -9.00001168e-01 8.11634481e-01 -4.60910529e-01
-1.43320930e+00 4.75336462e-01 -3.08212847e-01 -1.17496884e+00
-2.80400574e-01 -2.76387513e-01 -8.42551231e-01 9.70228493e-01
-1.46305895e+00 -1.07505488e+00 -2.55284369e-01 7.63367057e-01
2.95201778e-01 -4.29423928e-01 7.74420202e-01 3.54048938e-01
-4.39330161e-01 6.21224165e-01 3.65580112e-01 -3.09014730e-02
5.85224211e-01 -9.67843533e-01 8.56196359e-02 5.47759473e-01
-2.16606721e-01 3.11211258e-01 5.76874495e-01 -8.19087923e-01
-9.82927442e-01 -9.83721435e-01 9.52718854e-01 -2.45125264e-01
4.72713292e-01 5.17324246e-02 -1.04368746e+00 4.96977687e-01
-3.59389424e-01 -9.11353901e-02 7.19967663e-01 -2.54799664e-01
-2.61272252e-01 -5.07192850e-01 -1.97281003e+00 -1.46962941e-01
2.82124758e-01 -3.29053223e-01 -4.71088797e-01 3.70371431e-01
3.82968992e-01 -1.60952047e-01 -9.11478996e-01 3.06508631e-01
7.60291278e-01 -1.60777617e+00 8.66093636e-01 -6.25723362e-01
5.46293020e-01 2.13835210e-01 -3.60402018e-01 -5.48856676e-01
2.66249716e-01 -4.63786453e-01 -3.72889757e-01 1.50386131e+00
2.84094840e-01 -1.22629344e+00 1.04803026e+00 7.60003388e-01
3.64676446e-01 -1.18838239e+00 -1.12682390e+00 -7.14451969e-01
-1.49980560e-02 -4.75161701e-01 7.67921269e-01 1.12069094e+00
-7.90049806e-02 -3.31698418e-01 -3.67900461e-01 -1.55744910e-01
5.22828817e-01 2.00253129e-01 6.89939260e-01 -1.10552061e+00
-5.15195608e-01 -8.66629407e-02 -6.97924852e-01 6.09065555e-02
-6.29569411e-01 -4.46661383e-01 -8.32112849e-01 -1.22463346e+00
1.56215087e-01 -7.48858094e-01 -6.76901877e-01 2.33391300e-01
1.10670269e-01 9.21766311e-02 -1.04449570e-01 6.12145364e-01
2.64053121e-02 -2.56119892e-02 4.62184936e-01 2.17250243e-01
-3.13051790e-01 7.33005106e-01 -4.95022357e-01 4.32734638e-01
1.10766065e+00 -5.80191314e-01 -3.70308198e-02 8.89101923e-02
3.03757370e-01 3.77263576e-01 4.16822731e-01 -8.74148548e-01
2.16313377e-01 -2.20928416e-01 5.41001678e-01 -6.97415173e-01
-7.50081567e-03 -6.98027730e-01 2.74297327e-01 8.99689078e-01
6.76917359e-02 4.44398075e-01 5.11142388e-02 5.57231724e-01
-3.39937329e-01 -4.17546749e-01 5.57976544e-01 -2.83776402e-01
-2.12877095e-01 -1.15097612e-01 -5.67162298e-02 -1.86712325e-01
1.19545853e+00 -8.23702097e-01 -1.49292452e-03 -2.87617892e-01
-2.49737188e-01 4.04113978e-02 3.79364043e-01 -4.33213040e-02
6.04072988e-01 -9.79532480e-01 -8.26817691e-01 2.54142463e-01
-1.37724563e-01 7.75356293e-02 6.68202639e-02 9.35340941e-01
-8.88801575e-01 3.53822321e-01 -1.89976230e-01 -3.42299551e-01
-1.07391250e+00 2.65321195e-01 2.32439533e-01 -7.78097272e-01
-4.38976377e-01 6.92423105e-01 -7.70601153e-01 -2.95557648e-01
2.17594281e-01 -2.22827107e-01 -5.73014393e-02 -1.82767108e-01
3.20548683e-01 8.95283520e-01 4.23581421e-01 -2.22359166e-01
-3.40694517e-01 5.32348976e-02 -2.09317982e-01 -1.60732656e-03
1.38640630e+00 3.36718827e-01 -1.93136543e-01 4.91328478e-01
7.98144877e-01 -1.29408792e-01 -7.73364246e-01 -9.74054448e-03
2.88689405e-01 -5.14189899e-01 -1.64157107e-01 -9.75835741e-01
-7.84579635e-01 5.14212012e-01 5.79393208e-01 5.57654202e-01
1.27623761e+00 -6.52399898e-01 1.08828890e+00 1.52904928e-01
6.49201572e-01 -1.29693699e+00 -1.81589752e-01 -8.59091803e-02
5.24908721e-01 -1.17043197e+00 5.04407734e-02 -3.40120122e-02
-7.80867994e-01 1.19941580e+00 3.59744847e-01 -3.65019113e-01
6.31665707e-01 3.15544158e-01 -2.04564855e-02 7.81611875e-02
-8.89748573e-01 7.13504791e-01 -6.26020014e-01 5.46492994e-01
4.18284297e-01 6.32087216e-02 -4.18080896e-01 7.59111464e-01
-2.39812940e-01 6.93886220e-01 4.82464373e-01 1.24873078e+00
-2.51221240e-01 -9.32161987e-01 -6.11591339e-01 9.19482768e-01
-9.57749844e-01 1.20331243e-01 -1.00100830e-01 6.74623132e-01
1.27700865e-01 9.35175717e-01 -6.92404583e-02 -2.00006336e-01
3.68580252e-01 3.42571735e-01 -4.77009714e-02 -4.36091051e-02
-1.12077248e+00 -3.80453408e-01 7.86728933e-02 -2.79997140e-01
-1.64694116e-01 -9.07063484e-01 -1.08438158e+00 -5.80215454e-01
-4.59963381e-01 1.75313920e-01 1.01996756e+00 9.61027741e-01
1.31615788e-01 2.89240968e-03 9.89642859e-01 -1.93768725e-01
-1.04926562e+00 -1.08405840e+00 -6.99379981e-01 5.36052823e-01
-6.14657439e-03 -2.35786796e-01 -5.50250053e-01 -3.73568609e-02] | [7.951199531555176, 5.052997589111328] |
3c3f0b9f-8016-4bef-994b-e804cd79eeb9 | investigating-language-relationships-in | null | null | https://aclanthology.org/2022.cl-3.5 | https://aclanthology.org/2022.cl-3.5.pdf | Investigating Language Relationships in Multilingual Sentence Encoders Through the Lens of Linguistic Typology | Multilingual sentence encoders have seen much success in cross-lingual model transfer for downstream NLP tasks. The success of this transfer is, however, dependent on the model’s ability to encode the patterns of cross-lingual similarity and variation. Yet, we know relatively little about the properties of individual languages or the general patterns of linguistic variation that the models encode. In this article, we investigate these questions by leveraging knowledge from the field of linguistic typology, which studies and documents structural and semantic variation across languages. We propose methods for separating language-specific subspaces within state-of-the-art multilingual sentence encoders (LASER, M-BERT, XLM, and XLM-R) with respect to a range of typological properties pertaining to lexical, morphological, and syntactic structure. Moreover, we investigate how typological information about languages is distributed across all layers of the models. Our results show interesting differences in encoding linguistic variation associated with different pretraining strategies. In addition, we propose a simple method to study how shared typological properties of languages are encoded in two state-of-the-art multilingual models—M-BERT and XLM-R. The results provide insight into their information-sharing mechanisms and suggest that these linguistic properties are encoded jointly across typologically similar languages in these models. | ['Ekaterina Shutova', 'Rochelle Choenni'] | null | null | null | null | cl-acl-2022-9 | ['xlm-r'] | ['natural-language-processing'] | [-3.60517681e-01 -2.21220538e-01 -4.82569367e-01 -5.38952649e-01
-7.05046475e-01 -8.85444701e-01 6.51995242e-01 2.57895559e-01
-6.53767765e-01 6.39946878e-01 7.63460338e-01 -4.29729193e-01
3.05267386e-02 -6.49345458e-01 -1.02512860e+00 -3.09097975e-01
6.11975417e-02 6.61555409e-01 -1.88096762e-01 -4.44448650e-01
2.02315338e-02 -7.55736977e-03 -1.12681532e+00 9.15691257e-01
1.02962160e+00 2.89463967e-01 5.88589013e-01 1.47351176e-01
-3.88536245e-01 4.68515635e-01 -1.99103579e-01 -5.57420254e-01
-1.95998233e-02 -6.75066710e-01 -8.33282709e-01 -3.36906403e-01
9.55746770e-01 2.09379256e-01 -2.09797740e-01 1.07250905e+00
3.06472570e-01 -2.07866430e-01 9.25341666e-01 -5.85677803e-01
-1.31398344e+00 1.29262161e+00 -1.68933019e-01 5.57365716e-01
2.68200666e-01 1.71236962e-01 1.39595366e+00 -8.77999246e-01
1.05848420e+00 1.56273842e+00 8.30444336e-01 5.61817646e-01
-1.69490099e+00 -4.80291098e-01 3.09971154e-01 3.37459475e-01
-1.34579408e+00 -7.50020444e-01 5.65721929e-01 -7.50446975e-01
1.26117897e+00 -2.01498091e-01 4.75122452e-01 1.20368838e+00
5.86574614e-01 7.84257770e-01 1.60766530e+00 -5.17820001e-01
-2.14652702e-01 4.56264406e-01 1.12213649e-01 7.91846633e-01
2.51184106e-01 2.49477148e-01 -9.22262311e-01 3.49879622e-01
5.73348045e-01 -6.81721151e-01 -3.74134839e-01 -1.44844055e-01
-1.44712877e+00 8.36574852e-01 2.83502728e-01 8.69856477e-01
4.80899289e-02 1.31774042e-02 8.12697828e-01 6.22076750e-01
4.34880137e-01 6.65830612e-01 -9.08232272e-01 1.26767620e-01
-6.85101032e-01 -7.80800283e-02 5.87685466e-01 9.68945384e-01
9.01496768e-01 1.42827660e-01 -1.02674477e-01 1.21268988e+00
4.02780920e-01 6.52218163e-01 8.60920429e-01 -4.78917748e-01
7.57220387e-01 1.51288703e-01 -5.37001014e-01 -6.82130635e-01
-3.16369087e-01 -1.78663224e-01 -3.73385400e-01 -3.89021516e-01
3.73938382e-01 -6.02489524e-02 -4.56810802e-01 2.33308411e+00
-2.62451857e-01 -3.60391557e-01 3.14779162e-01 5.46274722e-01
7.16999829e-01 5.88594913e-01 5.53398728e-01 5.71444072e-02
1.53405488e+00 -6.31136179e-01 -3.94359857e-01 -5.65936446e-01
1.14146841e+00 -7.68112957e-01 1.43986869e+00 -1.94920450e-01
-1.20987082e+00 -7.64506578e-01 -7.97531366e-01 -4.73004669e-01
-6.20200694e-01 1.27389506e-01 6.83262229e-01 5.90042055e-01
-1.32377315e+00 5.43764949e-01 -5.24120331e-01 -6.44923389e-01
1.67840019e-01 1.32469967e-01 -5.05982220e-01 -1.31115139e-01
-1.53166664e+00 1.48429692e+00 6.63801253e-01 -2.49770239e-01
-6.86549425e-01 -8.91081512e-01 -1.15681744e+00 4.63863015e-02
-2.77455449e-01 -8.49109828e-01 1.16336095e+00 -1.03830826e+00
-1.21834791e+00 1.26935911e+00 -3.86760622e-01 -5.05821630e-02
-2.33552560e-01 8.48400369e-02 -3.98064017e-01 -1.67001992e-01
3.77629310e-01 7.26564109e-01 1.77687123e-01 -1.15354788e+00
-5.22041857e-01 -4.88441110e-01 -9.58288312e-02 5.16622066e-01
-3.43672216e-01 2.31923133e-01 1.33749872e-01 -6.96829677e-01
-2.27853745e-01 -8.68464112e-01 2.03108415e-01 -4.17593449e-01
-3.38515714e-02 -3.66123229e-01 5.95379882e-02 -7.85830557e-01
9.50019836e-01 -2.09439492e+00 3.95378172e-01 -1.49408862e-01
-4.26196367e-01 -7.22117210e-03 -4.89198208e-01 7.94196129e-01
7.66430646e-02 4.81748402e-01 -1.94040179e-01 -3.21018308e-01
3.20645124e-01 4.79602128e-01 -3.15908849e-01 4.07452822e-01
2.23082334e-01 1.23904455e+00 -9.20591295e-01 -4.82526690e-01
3.20977569e-02 3.41225147e-01 -6.98762476e-01 -2.95797467e-01
-2.89516419e-01 5.77948689e-01 -1.32281007e-02 3.17964494e-01
2.22799182e-01 3.17352146e-01 5.29345095e-01 -1.12511426e-01
-3.48279595e-01 1.16460788e+00 -4.24237788e-01 1.90627468e+00
-9.52272773e-01 6.60242617e-01 -2.03934032e-02 -9.97396946e-01
4.75377142e-01 3.98341715e-01 -9.45123285e-02 -9.04655397e-01
6.25487119e-02 5.53390920e-01 6.48499191e-01 -3.93128335e-01
4.87880141e-01 -6.96486473e-01 -3.86756957e-01 5.76052725e-01
6.23423219e-01 2.71339584e-02 4.35043275e-01 -7.25471526e-02
6.79444373e-01 9.33673382e-02 2.59280562e-01 -8.03584874e-01
4.43382740e-01 4.33262102e-02 5.69857776e-01 5.09935021e-01
-2.13214699e-02 2.24514771e-02 3.06419551e-01 -4.47017401e-02
-1.12754595e+00 -1.23397291e+00 -7.84876287e-01 1.27632487e+00
-3.30242425e-01 -4.37749773e-01 -5.26940525e-01 -5.17726421e-01
1.86479524e-01 1.07104421e+00 -5.40153801e-01 -3.47530544e-01
-8.33337426e-01 -7.19227791e-01 7.37058520e-01 5.50955892e-01
1.49110705e-01 -1.44699883e+00 2.06930339e-02 1.46458879e-01
-2.01086789e-01 -1.13563645e+00 -5.34400225e-01 2.57165730e-01
-9.31341231e-01 -7.03375041e-01 -2.16548964e-01 -1.18423855e+00
4.58529621e-01 -6.36244789e-02 1.54935813e+00 -2.51583397e-01
2.21800417e-01 3.87124330e-01 -2.09780186e-01 -2.22008467e-01
-8.02791417e-01 4.37800229e-01 2.09499091e-01 -4.09385413e-01
6.08658314e-01 -4.02167857e-01 1.20313041e-01 1.15965249e-03
-6.46315336e-01 -1.37232408e-01 6.81440592e-01 6.93982840e-01
5.07511854e-01 -2.52484173e-01 5.07017434e-01 -1.19115508e+00
6.94734812e-01 -6.68690741e-01 -1.62267044e-01 3.29017013e-01
-2.98360676e-01 4.63215500e-01 9.00811017e-01 -2.94083923e-01
-1.05280423e+00 -4.11368221e-01 -1.47202224e-01 1.48119599e-01
-3.37112635e-01 9.71208632e-01 -3.14248741e-01 1.45871595e-01
4.40935552e-01 3.48173857e-01 -1.87964112e-01 -5.79586923e-01
6.42867506e-01 4.51438189e-01 2.83384383e-01 -1.06852710e+00
5.77618480e-01 1.09582141e-01 -4.21436936e-01 -1.12048697e+00
-9.18738961e-01 -1.76648080e-01 -1.01011503e+00 2.87172675e-01
9.92266774e-01 -1.07904553e+00 -2.61486620e-01 3.62764835e-01
-1.32078600e+00 -5.58310330e-01 -3.86485815e-01 6.07034683e-01
-6.69659913e-01 1.45288065e-01 -9.92800593e-01 -6.61584213e-02
-1.08886575e-02 -1.19202971e+00 9.76767063e-01 -5.16987324e-01
-5.98314345e-01 -1.75101900e+00 3.16707283e-01 -7.22434046e-03
4.11276132e-01 -3.87474358e-01 1.61723316e+00 -6.10690594e-01
-3.95144165e-01 5.72913289e-01 -2.68993992e-02 3.39108795e-01
4.04841602e-02 -3.92833352e-01 -6.41002297e-01 -2.73851931e-01
-1.12687282e-01 -4.30340469e-01 1.10051835e+00 4.37240601e-01
5.02746463e-01 -4.07800227e-01 -2.16817096e-01 8.60598326e-01
1.50945699e+00 -7.54465684e-02 2.41502166e-01 3.26905459e-01
7.39906192e-01 8.72836292e-01 1.52194686e-02 -3.74906331e-01
7.95938611e-01 6.25198424e-01 -4.25671637e-01 2.35773385e-01
-4.60344225e-01 -4.79664177e-01 1.00507796e+00 1.70354998e+00
3.73660892e-01 1.92993954e-02 -9.88561153e-01 8.88716698e-01
-1.48470378e+00 -9.79917645e-01 -4.32287455e-02 2.09114099e+00
1.41508949e+00 -1.47955760e-01 -8.96505639e-02 -3.80347580e-01
5.28169632e-01 3.44844908e-01 -8.28009620e-02 -1.00678122e+00
-4.70129788e-01 2.49242336e-01 3.71837169e-01 7.23006129e-01
-7.91467607e-01 1.57923543e+00 6.78665686e+00 6.66173875e-01
-9.87928212e-01 2.59798527e-01 1.97583884e-01 1.32472739e-01
-7.89446354e-01 2.62086689e-02 -1.17878497e+00 5.64560652e-01
1.29381633e+00 -2.38956660e-01 5.40979028e-01 3.69231522e-01
-4.46238294e-02 3.28272656e-02 -1.63373530e+00 3.61795634e-01
3.48691136e-01 -1.16553903e+00 3.55372220e-01 -1.21668652e-01
7.81183958e-01 4.56129223e-01 2.09950820e-01 5.40068269e-01
5.07465541e-01 -1.01538324e+00 9.55187738e-01 3.18756878e-01
9.72909033e-01 -5.99824369e-01 3.83132279e-01 2.21703693e-01
-1.09516001e+00 1.87045068e-01 -6.91636086e-01 -6.63914010e-02
2.50168711e-01 2.44384423e-01 -6.12313509e-01 2.01334059e-01
4.36038435e-01 9.06559289e-01 -7.03716218e-01 4.63533193e-01
-6.32803321e-01 7.84182429e-01 7.78005552e-03 -3.15795699e-03
2.95602053e-01 -8.63672122e-02 4.03071910e-01 1.63393748e+00
1.61959991e-01 -3.21432799e-01 2.30005234e-01 9.35797572e-01
-9.72031206e-02 6.01063311e-01 -9.09253180e-01 -1.15736082e-01
4.97413546e-01 6.53570056e-01 -3.11957955e-01 -2.72956431e-01
-8.11982572e-01 7.97182262e-01 9.97282028e-01 2.46791974e-01
-2.91418612e-01 1.54032856e-01 7.33776629e-01 7.05586672e-02
2.41696671e-01 -6.54624045e-01 -4.39051837e-01 -1.37147653e+00
-3.09643485e-02 -9.27240789e-01 2.14303628e-01 -5.12912452e-01
-1.76642907e+00 5.62581182e-01 4.80329655e-02 -5.99814653e-01
-2.42175311e-01 -9.68592167e-01 -2.83066958e-01 1.23503542e+00
-1.52026796e+00 -1.39423013e+00 6.93100870e-01 6.28888428e-01
5.00780165e-01 -4.77179199e-01 1.01917207e+00 2.40649715e-01
-4.00318027e-01 7.60521889e-01 3.12685251e-01 4.11476791e-01
9.59348440e-01 -1.37739909e+00 6.20330513e-01 6.12597585e-01
7.02658713e-01 9.88858700e-01 3.41965497e-01 -6.55570865e-01
-1.15210199e+00 -1.08848393e+00 1.62822723e+00 -7.84992456e-01
9.62922513e-01 -6.46050632e-01 -1.01141214e+00 1.15878034e+00
6.57169163e-01 -3.44483107e-01 1.01106799e+00 9.25942242e-01
-7.29912221e-01 5.53217456e-02 -6.87949419e-01 7.22199440e-01
1.16772246e+00 -1.12043214e+00 -1.02886391e+00 2.69659489e-01
7.13214040e-01 4.85892296e-02 -9.59347963e-01 1.50552630e-01
3.35207790e-01 -7.95828164e-01 6.48691118e-01 -9.03353453e-01
6.40960097e-01 9.82512832e-02 -4.53974158e-01 -2.15901780e+00
-7.37719357e-01 -5.23325279e-02 5.56955457e-01 1.40355873e+00
7.13882506e-01 -6.73568606e-01 7.45869204e-02 -1.46119036e-02
-5.14389575e-01 -5.69083869e-01 -9.97465014e-01 -1.00932360e+00
1.03145719e+00 -4.38612819e-01 2.65770525e-01 1.35287619e+00
2.39264682e-01 8.96127522e-01 2.38390669e-01 -3.02750152e-02
3.16777170e-01 1.75235383e-02 2.36959994e-01 -1.03581142e+00
-3.40527087e-01 -6.63024306e-01 -3.78326625e-01 -6.63836539e-01
1.03505409e+00 -1.96858978e+00 -8.19396973e-02 -1.42189229e+00
2.84539819e-01 -5.51732063e-01 -3.93679917e-01 5.76203585e-01
-4.07217741e-02 2.78602485e-02 3.33973765e-01 2.13017255e-01
-3.04005802e-01 5.67953110e-01 1.06233346e+00 1.35092959e-01
5.48501173e-03 -6.27533078e-01 -7.48425305e-01 7.96392024e-01
8.74293983e-01 -3.88507575e-01 -2.54630893e-01 -1.28874683e+00
3.99508178e-01 -1.85481101e-01 9.58113745e-02 -7.13838816e-01
-8.90081525e-02 -1.36108235e-01 2.82264113e-01 -1.14309311e-01
-5.28589748e-02 -5.08715570e-01 -4.13743407e-01 2.83021182e-01
-5.70799053e-01 4.62385029e-01 5.10289013e-01 -3.44754010e-02
-4.18854237e-01 -3.06726009e-01 8.15691054e-01 -4.54663783e-01
-8.56222808e-01 1.58081532e-01 -5.95234811e-01 7.85413980e-01
5.31613946e-01 9.23084170e-02 -4.25566256e-01 1.57376423e-01
-5.25588274e-01 3.16379741e-02 6.00841403e-01 7.14324892e-01
4.29931395e-02 -1.58339238e+00 -1.04953170e+00 3.87624383e-01
3.30811411e-01 -6.98048174e-01 1.15048664e-03 7.31816828e-01
-8.87246430e-02 7.93793082e-01 -3.80996704e-01 -4.35516179e-01
-7.06403613e-01 6.97546244e-01 3.64889026e-01 -1.89764738e-01
-2.13172942e-01 8.78553510e-01 4.49164063e-01 -8.63688767e-01
-3.90575409e-01 -3.17533940e-01 -5.85931875e-02 3.21278512e-01
-2.29409225e-02 -1.10693527e-02 -6.33555874e-02 -1.16433954e+00
-4.32295501e-01 8.65416706e-01 -2.87221313e-01 -3.26608658e-01
1.17733848e+00 -1.99157521e-01 -4.69449282e-01 1.04859650e+00
1.33377218e+00 4.37131226e-01 -5.85758924e-01 -4.16870683e-01
2.67281264e-01 -7.89672136e-03 -2.24038988e-01 -8.36180627e-01
-7.83812582e-01 1.01538754e+00 -5.14726304e-02 -4.25772250e-01
6.20424390e-01 2.82801867e-01 7.87006199e-01 1.96620733e-01
4.81278598e-01 -1.19337928e+00 -4.79033530e-01 1.25043249e+00
8.78396511e-01 -1.06398964e+00 -3.65780026e-01 -1.99905053e-01
-7.51310527e-01 6.56535089e-01 5.25945246e-01 -2.84143090e-01
6.71379983e-01 2.99403369e-01 1.85178339e-01 -9.51572880e-02
-9.42716002e-01 -3.80037040e-01 5.43605745e-01 5.52895010e-01
1.21640599e+00 3.20935160e-01 -6.74302101e-01 4.85382408e-01
-8.07521820e-01 -6.42350256e-01 1.84805706e-01 4.41238046e-01
-3.86254966e-01 -1.37709069e+00 -1.04757361e-01 2.99527138e-01
-4.85735804e-01 -7.75639296e-01 -4.86051589e-01 9.15872335e-01
4.63375688e-01 5.23377120e-01 1.98212698e-01 -1.12903744e-01
3.12977195e-01 6.22285068e-01 8.71074140e-01 -1.26767254e+00
-6.78198397e-01 -1.97844848e-01 3.44047755e-01 -2.90915877e-01
-2.66320944e-01 -1.16397095e+00 -1.06016326e+00 -8.76815841e-02
1.95362061e-01 7.70706832e-02 4.55674469e-01 1.03411007e+00
1.98803112e-01 4.16453034e-01 1.62666515e-01 -5.24640024e-01
-5.42466760e-01 -1.11540163e+00 -6.59629345e-01 5.99215806e-01
9.50137302e-02 -5.22734046e-01 -2.40454659e-01 -1.02837533e-02] | [10.854185104370117, 9.91356086730957] |
af1d2f97-9525-4cef-beda-7613b9dc1403 | parallel-computation-of-pdfs-on-big-spatial | 1805.03141 | null | http://arxiv.org/abs/1805.03141v1 | http://arxiv.org/pdf/1805.03141v1.pdf | Parallel Computation of PDFs on Big Spatial Data Using Spark | We consider big spatial data, which is typically produced in scientific areas
such as geological or seismic interpretation. The spatial data can be produced
by observation (e.g. using sensors or soil instrument) or numerical simulation
programs and correspond to points that represent a 3D soil cube area. However,
errors in signal processing and modeling create some uncertainty, and thus a
lack of accuracy in identifying geological or seismic phenomenons. Such
uncertainty must be carefully analyzed. To analyze uncertainty, the main
solution is to compute a Probability Density Function (PDF) of each point in
the spatial cube area. However, computing PDFs on big spatial data can be very
time consuming (from several hours to even months on a parallel computer). In
this paper, we propose a new solution to efficiently compute such PDFs in
parallel using Spark, with three methods: data grouping, machine learning
prediction and sampling. We evaluate our solution by extensive experiments on
different computer clusters using big data ranging from hundreds of GB to
several TB. The experimental results show that our solution scales up very well
and can reduce the execution time by a factor of 33 (in the order of seconds or
minutes) compared with a baseline method. | ['Patrick Valduriez', 'Esther Pacitti', 'Noel Moreno Lemus', 'Ji Liu', 'Fabio Porto'] | 2018-05-08 | null | null | null | null | ['seismic-interpretation'] | ['miscellaneous'] | [-2.04088166e-01 -2.08203822e-01 5.79748750e-01 -1.26395285e-01
-9.92769957e-01 -5.37615597e-01 3.82893652e-01 6.88548326e-01
-3.87956440e-01 9.16681349e-01 2.27327663e-02 -4.78590637e-01
-1.85568556e-01 -1.56076682e+00 -9.65640068e-01 -8.44278336e-01
-4.71235693e-01 7.34224617e-01 7.67965853e-01 1.72834828e-01
2.62053818e-01 6.16438985e-01 -1.56016350e+00 8.35197046e-02
1.14469635e+00 1.32061076e+00 4.06241685e-01 5.82314968e-01
-4.31174099e-01 5.96678078e-01 -9.27740097e-01 2.84371048e-01
6.24576136e-02 2.32177511e-01 -5.61252654e-01 -3.87432992e-01
-4.37280834e-01 -3.35271090e-01 1.99634671e-01 1.13970852e+00
5.70494950e-01 -1.02429919e-01 5.19972146e-01 -1.08584344e+00
6.89634159e-02 9.34625626e-01 -8.39287281e-01 2.45049998e-01
2.60626853e-01 -2.90864080e-01 1.45961016e-01 -8.21909070e-01
1.08171850e-01 1.19031906e+00 7.24727213e-01 -5.79950392e-01
-1.12870336e+00 -6.22403264e-01 -4.08546299e-01 1.80753469e-01
-2.02516222e+00 -7.79682770e-02 2.97290862e-01 -5.11740744e-01
7.01468706e-01 4.40236509e-01 5.36892176e-01 3.45647633e-01
4.09720302e-01 4.76455241e-01 1.17449093e+00 -1.60726458e-01
9.65496838e-01 -2.47257829e-01 1.72673669e-02 -1.11437380e-01
4.69120383e-01 -3.32956724e-02 -3.84008318e-01 -7.72367597e-01
5.75826049e-01 1.93528179e-02 4.02999017e-03 3.04822832e-01
-1.08300400e+00 7.40426183e-01 1.93290263e-01 1.01153746e-01
-7.12779999e-01 3.28927517e-01 2.61111557e-01 1.74572632e-01
6.21772408e-01 -5.56772761e-02 -3.82479519e-01 -3.34708273e-01
-1.03657019e+00 5.53280294e-01 8.03189695e-01 6.38627827e-01
8.66370499e-01 -6.43971562e-02 1.97608188e-01 4.49939907e-01
1.77193627e-01 1.14997423e+00 8.43942165e-02 -8.36999297e-01
4.45940614e-01 2.07531735e-01 4.78641093e-01 -1.26813900e+00
-6.17343605e-01 -7.45153593e-05 -1.22231805e+00 5.85385002e-02
5.94477117e-01 -3.17754716e-01 -6.48670971e-01 1.03919709e+00
6.82931364e-01 4.02207166e-01 -8.76903236e-02 7.77584314e-01
5.74474096e-01 1.21105826e+00 -1.13357544e-01 -1.58813387e-01
1.44245255e+00 -6.36255443e-02 -5.79612792e-01 1.85724869e-01
6.29254580e-01 -6.69545114e-01 9.34725881e-01 7.06253409e-01
-8.31223965e-01 -2.04340428e-01 -7.62227654e-01 5.83918810e-01
-4.13131475e-01 -2.77354568e-01 5.60037971e-01 5.56039810e-01
-8.59410942e-01 8.10330927e-01 -1.31112564e+00 -9.29233879e-02
8.06622580e-02 2.51583084e-02 1.90289374e-02 -7.85603002e-03
-1.26767087e+00 3.71072322e-01 3.78937483e-01 1.59689099e-01
-6.24608457e-01 -9.92580771e-01 -5.91659606e-01 2.88679868e-01
3.09272081e-01 -2.52363861e-01 7.05020964e-01 9.20696557e-03
-1.11752701e+00 1.82887465e-01 -6.24728426e-02 -4.36890423e-01
2.52729803e-01 -1.14402503e-01 -3.23933482e-01 -4.93098199e-02
2.18894482e-01 -1.47173315e-01 3.65041107e-01 -1.06897330e+00
-6.39657438e-01 -6.29204452e-01 -4.18759048e-01 -3.31137180e-01
-9.81735587e-02 4.05003726e-01 -1.94971323e-01 -6.23247862e-01
4.41083252e-01 -8.41502249e-01 -7.40337849e-01 -5.15187562e-01
-4.80577141e-01 -6.19737245e-02 4.53355432e-01 -8.44853163e-01
1.31106424e+00 -2.05697203e+00 -3.50988567e-01 8.64295661e-01
7.48890117e-02 -2.90241212e-01 4.60302025e-01 5.71443617e-01
3.13358068e-01 1.90401807e-01 -4.90550756e-01 2.20760301e-01
-2.22940534e-01 1.21159002e-01 -5.40942192e-01 5.25931954e-01
-2.90392160e-01 1.98669583e-01 -6.96838200e-01 -5.24923027e-01
1.64276302e-01 3.93290758e-01 -1.63044959e-01 7.12341815e-02
-9.75155458e-02 5.38120329e-01 -8.01882088e-01 5.36154866e-01
1.27993655e+00 -3.10066909e-01 -3.86207551e-02 1.40344083e-01
-4.30666000e-01 2.47758418e-01 -1.94365740e+00 1.25930023e+00
-3.98690641e-01 1.82425439e-01 1.41470253e-01 -1.23445070e+00
1.23294234e+00 2.13097766e-01 7.44750738e-01 -3.79510999e-01
-1.52428657e-01 5.39192259e-01 -2.78993428e-01 -3.08304608e-01
5.18051207e-01 2.25379333e-01 -2.52397716e-01 8.08236480e-01
-6.80318713e-01 -3.87277395e-01 1.49639845e-01 6.05941564e-02
1.42093921e+00 -4.24017012e-01 7.11334944e-02 -5.78597963e-01
1.50689542e-01 2.50921011e-01 6.60046935e-01 6.87276602e-01
3.22614700e-01 4.50840920e-01 7.70407140e-01 -6.04753077e-01
-1.03945267e+00 -9.54693913e-01 -5.55887103e-01 6.30459070e-01
6.78533986e-02 -5.92655718e-01 -6.07259214e-01 2.79083222e-01
3.15008491e-01 5.47839761e-01 -2.55868286e-01 3.58049363e-01
-4.89886314e-01 -1.33231175e+00 4.57293630e-01 6.34544432e-01
6.57377362e-01 -7.66541183e-01 -8.18629265e-01 4.93063927e-01
-6.57289699e-02 -9.50516820e-01 3.87817711e-01 2.18813166e-01
-1.04136646e+00 -5.64554453e-01 -4.96260077e-01 6.09783223e-03
3.36214811e-01 1.12013310e-01 1.09699309e+00 -4.69896980e-02
-2.31849089e-01 -2.99529642e-01 -4.00771976e-01 -5.39334416e-01
-9.23903473e-03 -3.18524867e-01 1.19194932e-01 -1.74320668e-01
3.58119875e-01 -9.89828527e-01 -4.30270165e-01 3.59843940e-01
-9.50776696e-01 -2.37572923e-01 2.66957074e-01 4.12084341e-01
8.63441944e-01 9.03265536e-01 4.27149504e-01 -5.30519545e-01
6.19456112e-01 -1.06110358e+00 -1.14257920e+00 9.37547758e-02
-2.61366218e-01 9.70380753e-02 5.37669480e-01 1.36244800e-02
-8.54757071e-01 1.50622562e-01 -1.10361785e-01 -8.20244849e-02
-2.32320443e-01 1.11189055e+00 -7.50118271e-02 8.67402181e-02
5.87412536e-01 1.90239534e-01 -2.49742940e-01 -6.78949416e-01
7.37291649e-02 1.04381430e+00 5.24679601e-01 -1.02044928e+00
5.46478271e-01 6.17090940e-01 1.47312880e-01 -1.06340885e+00
-1.44524515e-01 -3.80435199e-01 -2.41580337e-01 -9.18855071e-02
3.54233831e-01 -9.40704346e-01 -8.77828300e-01 5.34943163e-01
-1.14992595e+00 -2.73587853e-01 -1.45681232e-01 6.33741021e-01
-8.40295777e-02 3.28288108e-01 -4.91569489e-01 -9.73902822e-01
-2.03953028e-01 -9.16363597e-01 1.20686364e+00 1.21465623e-01
-2.06835661e-03 -5.52939713e-01 1.98868677e-01 -1.00238711e-01
5.73501289e-01 6.70735121e-01 4.72567856e-01 -1.54409871e-01
-7.25983262e-01 -3.06839049e-01 -4.43087667e-01 -2.39530459e-01
-1.36197150e-01 2.40882561e-01 -8.45292985e-01 -4.84856628e-02
1.40983835e-01 -9.55799744e-02 7.69114614e-01 8.58253658e-01
1.76998281e+00 -1.18729621e-01 -5.66105366e-01 4.60869461e-01
1.48939431e+00 6.49546683e-02 9.21815753e-01 2.93437660e-01
3.81161451e-01 6.69236183e-01 8.66169751e-01 1.18777740e+00
3.71186256e-01 5.56643069e-01 3.52017999e-01 1.82627439e-01
5.49678028e-01 8.09202045e-02 -4.93430793e-02 6.48510516e-01
-4.53806311e-01 -1.67719781e-01 -1.53956103e+00 6.70751750e-01
-1.93027627e+00 -9.53715622e-01 -6.19506836e-01 2.54963207e+00
7.65552402e-01 -7.54276514e-02 7.91177228e-02 6.30151331e-01
7.22781301e-01 -5.64334616e-02 -2.54941463e-01 -1.21320039e-01
-1.16962768e-01 1.38811588e-01 9.91068661e-01 4.21715707e-01
-8.28081667e-01 5.28997660e-01 6.18309879e+00 1.18569958e+00
-1.21449685e+00 1.16018228e-01 5.42127371e-01 3.44757996e-02
-5.05503118e-01 -3.17313373e-02 -6.05139196e-01 9.14697468e-01
1.23337638e+00 -2.56319106e-01 2.89933860e-01 8.64181817e-01
4.93616551e-01 -8.08299184e-01 -4.89553988e-01 1.14089870e+00
-6.66970849e-01 -1.38450134e+00 -5.54490387e-01 2.99608886e-01
5.82087696e-01 2.33379900e-01 -3.98985773e-01 -1.57765687e-01
7.28717923e-01 -1.13922870e+00 6.03285730e-01 4.93838429e-01
5.56940019e-01 -9.30398345e-01 8.60793173e-01 5.76378882e-01
-1.43699908e+00 2.13224918e-01 -8.23041916e-01 -3.77529860e-01
4.54327285e-01 1.73958611e+00 -5.76264739e-01 5.55614412e-01
1.51958835e+00 1.32772520e-01 -1.92897782e-01 1.20292974e+00
9.04721767e-03 9.05758202e-01 -1.29102099e+00 -3.38953376e-01
1.38953105e-02 -2.54011840e-01 3.73494506e-01 9.57649589e-01
1.05994487e+00 5.15984178e-01 1.54960275e-01 8.05791914e-01
3.70557517e-01 1.42985314e-01 -6.01249635e-01 2.95548320e-01
1.00919092e+00 8.99783194e-01 -8.44955742e-01 -4.73269820e-01
-1.22155815e-01 2.45209306e-01 -6.35609999e-02 1.89118952e-01
-7.77869999e-01 -7.30952561e-01 3.27367008e-01 2.97895163e-01
2.43737325e-01 -4.14776266e-01 -8.61350775e-01 -7.53212631e-01
3.00014824e-01 -5.30517519e-01 2.03595579e-01 -6.44788265e-01
-9.77660537e-01 4.73460436e-01 1.39747843e-01 -1.11216605e+00
-2.72921771e-01 -3.72798324e-01 -7.53852189e-01 9.47676897e-01
-9.30763841e-01 -5.03514469e-01 -5.44637918e-01 6.74378097e-01
-2.60903865e-01 8.32062960e-02 6.91274881e-01 5.48483431e-01
-2.90531635e-01 -1.83348835e-01 8.07970822e-01 -3.00980002e-01
2.44079649e-01 -1.21898770e+00 4.97504443e-01 6.53132975e-01
-2.51394212e-01 1.53558895e-01 1.09733021e+00 -9.25232053e-01
-1.36216688e+00 -8.89002860e-01 8.58807385e-01 -1.47871152e-01
8.42027843e-01 -2.84826040e-01 -1.15584588e+00 1.56011820e-01
-2.61355519e-01 4.06622976e-01 5.15665948e-01 1.52733371e-01
3.76148559e-02 -4.06655550e-01 -1.29553664e+00 1.18105724e-01
5.52025974e-01 -1.63892522e-01 -7.38471597e-02 5.18801808e-01
5.66937089e-01 -3.69543463e-01 -1.33308339e+00 4.99816895e-01
2.95605481e-01 -9.21305418e-01 1.00101030e+00 -6.35277480e-02
2.32769966e-01 -6.67175055e-01 -4.55185443e-01 -1.34153140e+00
8.73117596e-02 -4.43086058e-01 1.36553317e-01 1.13823855e+00
1.07902847e-01 -7.32954562e-01 6.56924367e-01 7.05720663e-01
1.31458908e-01 -4.65618402e-01 -1.20812631e+00 -1.04824376e+00
1.43985391e-01 -1.02032471e+00 1.41088820e+00 7.93252051e-01
1.18736058e-01 -2.49573797e-01 -2.43742675e-01 6.68400943e-01
9.55437541e-01 3.37081403e-01 8.97376060e-01 -1.42647934e+00
-1.80543005e-01 -9.35627595e-02 -5.39163888e-01 -7.50424564e-01
-4.05533254e-01 -1.72394082e-01 8.34720284e-02 -1.37306511e+00
-9.51902643e-02 -1.18339241e+00 3.86741497e-02 3.36439103e-01
5.62762395e-02 -8.34665745e-02 -3.72636437e-01 3.10879678e-01
-1.71128035e-01 3.72884750e-01 5.99953771e-01 -6.77965358e-02
-2.46806607e-01 4.59806696e-02 -2.44548053e-01 8.62556875e-01
8.67046714e-01 -7.32468486e-01 5.89688420e-02 -5.51062882e-01
7.48822331e-01 5.88366270e-01 3.57675642e-01 -1.26022220e+00
3.43503267e-01 -4.21669841e-01 1.90754324e-01 -1.20823550e+00
-4.26813848e-02 -7.71402359e-01 6.80001438e-01 5.04602253e-01
3.37688655e-01 -1.28864110e-01 2.32396722e-01 4.86001194e-01
-5.38699389e-01 -1.00358196e-01 4.24848348e-01 1.20806374e-01
-4.13781404e-01 4.91648167e-02 -4.33575720e-01 -2.54198998e-01
9.96670246e-01 1.88727081e-01 -3.86751771e-01 -3.37851256e-01
-5.03900528e-01 4.11575645e-01 3.00151974e-01 -2.21462592e-01
4.79916215e-01 -1.34455442e+00 -9.42132771e-01 -6.50425581e-03
-1.25902519e-01 5.22035539e-01 3.50102305e-01 8.37869644e-01
-1.00936997e+00 3.83159190e-01 8.10812265e-02 -9.40567672e-01
-7.47332454e-01 4.93023872e-01 -9.76854265e-02 -2.92876065e-01
-6.32889688e-01 6.49667859e-01 -2.34577894e-01 -4.80702400e-01
1.00929439e-01 -6.62907302e-01 -2.27837097e-02 -1.17608018e-01
8.48906696e-01 9.00682032e-01 3.10112149e-01 -1.38368338e-01
-5.69742739e-01 7.65406489e-01 6.97480917e-01 -2.89060235e-01
1.60033464e+00 -3.42166312e-02 -4.55858439e-01 5.31452298e-01
7.59266913e-01 2.38858193e-01 -7.39930332e-01 -2.53726065e-01
1.89739242e-01 -7.01821864e-01 1.40436724e-01 -2.69570649e-01
-1.01613533e+00 9.48886871e-01 4.62166876e-01 7.29860604e-01
1.06657493e+00 2.26869896e-01 7.96951771e-01 5.46295464e-01
1.00577176e+00 -1.29497445e+00 -5.85283756e-01 3.24732035e-01
1.03369129e+00 -1.13201201e+00 8.26985165e-02 -4.89749134e-01
-3.04112881e-01 9.84202564e-01 1.43240578e-03 -1.53880194e-01
1.27488887e+00 9.53935027e-01 -3.42702329e-01 -1.31985947e-01
-3.75170410e-01 8.71450529e-02 -3.55789512e-01 5.09120166e-01
2.29198247e-01 5.48270226e-01 -2.74138182e-01 7.65849173e-01
-5.08647084e-01 1.64056897e-01 5.53586185e-01 1.03759050e+00
-9.19253349e-01 -8.59447360e-01 -1.25058675e+00 7.14593172e-01
-4.52062756e-01 8.01402554e-02 3.84528816e-01 1.46180287e-01
-1.35333896e-01 1.01749432e+00 3.88741970e-01 -3.44257832e-01
2.19497122e-02 -4.51985151e-01 -9.13545117e-02 -3.30204368e-01
-2.86223255e-02 1.79520205e-01 1.88860707e-02 -7.51240134e-01
-1.07919961e-01 -8.13318431e-01 -1.65402102e+00 -9.35586154e-01
-1.25223100e-01 4.24252659e-01 1.05240488e+00 7.34340250e-01
3.15563619e-01 2.14792311e-01 6.27941966e-01 -8.06016743e-01
-5.20496845e-01 -1.08877635e+00 -1.29876101e+00 -1.45102322e-01
-1.08486868e-01 -6.31539643e-01 -4.62189227e-01 -1.67040855e-01] | [6.993375778198242, 3.9449167251586914] |
5eccdbab-eff3-4b92-bb61-9b107bbb02b7 | multi-fidelity-active-learning-with-gflownets | 2306.11715 | null | https://arxiv.org/abs/2306.11715v1 | https://arxiv.org/pdf/2306.11715v1.pdf | Multi-Fidelity Active Learning with GFlowNets | In the last decades, the capacity to generate large amounts of data in science and engineering applications has been growing steadily. Meanwhile, the progress in machine learning has turned it into a suitable tool to process and utilise the available data. Nonetheless, many relevant scientific and engineering problems present challenges where current machine learning methods cannot yet efficiently leverage the available data and resources. For example, in scientific discovery, we are often faced with the problem of exploring very large, high-dimensional spaces, where querying a high fidelity, black-box objective function is very expensive. Progress in machine learning methods that can efficiently tackle such problems would help accelerate currently crucial areas such as drug and materials discovery. In this paper, we propose the use of GFlowNets for multi-fidelity active learning, where multiple approximations of the black-box function are available at lower fidelity and cost. GFlowNets are recently proposed methods for amortised probabilistic inference that have proven efficient for exploring large, high-dimensional spaces and can hence be practical in the multi-fidelity setting too. Here, we describe our algorithm for multi-fidelity active learning with GFlowNets and evaluate its performance in both well-studied synthetic tasks and practically relevant applications of molecular discovery. Our results show that multi-fidelity active learning with GFlowNets can efficiently leverage the availability of multiple oracles with different costs and fidelities to accelerate scientific discovery and engineering design. | ['Yoshua Bengio', 'Cheng-Hao Liu', 'Moksh Jain', 'Nikita Saxena', 'Alex Hernandez-Garcia'] | 2023-06-20 | null | null | null | null | ['active-learning', 'active-learning'] | ['methodology', 'natural-language-processing'] | [ 3.41733336e-01 -1.05275497e-01 -3.86103094e-01 4.63311467e-03
-1.40820527e+00 -6.27323031e-01 6.82234824e-01 6.46133125e-01
-8.19852471e-01 1.36804712e+00 -3.73846769e-01 -4.22801763e-01
-6.49691343e-01 -9.23043668e-01 -1.10625732e+00 -1.10910463e+00
-2.93923408e-01 1.05996442e+00 9.16180164e-02 3.03920984e-01
3.19562435e-01 7.03302264e-01 -1.52058113e+00 4.02496420e-02
9.96427834e-01 8.67435694e-01 2.81695634e-01 7.13123739e-01
-3.40510868e-02 2.43763238e-01 -2.94937909e-01 -1.94084629e-01
1.55959338e-01 1.23278117e-02 -8.93302202e-01 -4.06956762e-01
1.39989972e-01 1.23501554e-01 9.18366015e-02 6.18170321e-01
8.28024566e-01 4.71054673e-01 4.44298893e-01 -8.91678572e-01
-5.51981628e-02 3.84495080e-01 -2.72286117e-01 7.31063634e-02
3.09158146e-01 5.28994501e-01 1.02517331e+00 -1.12454748e+00
7.32655168e-01 1.14889371e+00 3.85353029e-01 1.39776647e-01
-1.61338425e+00 -5.97757339e-01 -1.16749525e-01 3.66029441e-01
-1.36758232e+00 -6.19634569e-01 4.83753026e-01 -2.55445719e-01
9.11195278e-01 3.80429626e-01 6.94172740e-01 1.17499304e+00
1.22938208e-01 7.65431404e-01 1.03741086e+00 -3.24150085e-01
8.02210391e-01 -7.94715732e-02 -3.43978465e-01 7.07216561e-01
2.84722269e-01 2.12338030e-01 -7.68274307e-01 -7.60974526e-01
5.99250615e-01 2.60776550e-01 -7.52275735e-02 -5.61503172e-01
-1.33054197e+00 1.04024839e+00 1.78483143e-01 1.09797440e-01
-4.54977602e-01 2.70853996e-01 1.47626355e-01 1.58749837e-02
4.04582173e-01 1.07700622e+00 -6.11376882e-01 -4.36711639e-01
-1.05331397e+00 2.83879995e-01 9.52304244e-01 4.70555723e-01
8.34089577e-01 -3.68377239e-01 1.03695162e-01 5.63838124e-01
3.09293538e-01 2.28126898e-01 -3.56554538e-02 -1.10093439e+00
1.77829817e-01 4.08650249e-01 2.69129395e-01 -6.91805840e-01
-1.68287233e-01 -2.82899141e-01 -9.76219475e-01 2.81702608e-01
4.42629844e-01 9.01766270e-02 -8.50392044e-01 1.51759219e+00
8.37886631e-01 3.21312249e-01 -1.62806526e-01 5.57522357e-01
3.90807569e-01 8.20359170e-01 1.94798976e-01 -5.17421424e-01
9.92704570e-01 -6.00606561e-01 -4.46770519e-01 1.93297088e-01
7.18408465e-01 -7.06951261e-01 9.42348301e-01 7.79649973e-01
-1.25353909e+00 -1.93945602e-01 -8.70822191e-01 -8.60529393e-02
-5.40628076e-01 -4.00003076e-01 1.18716455e+00 4.71054524e-01
-6.79225922e-01 7.87915349e-01 -9.29623246e-01 1.36505291e-01
9.95050967e-01 6.84766412e-01 -5.37931085e-01 -3.39070380e-01
-1.15694714e+00 5.78293264e-01 5.27579606e-01 1.15508489e-01
-1.31927407e+00 -1.15490007e+00 -4.78801876e-01 4.89601353e-03
8.58730316e-01 -8.27553809e-01 6.65916264e-01 -3.09598237e-01
-1.48704731e+00 4.35883194e-01 1.54766619e-01 -3.91328812e-01
4.71549273e-01 -3.11340746e-02 7.08008856e-02 5.86224422e-02
-2.15730295e-01 5.91522336e-01 5.14883101e-01 -6.97155237e-01
-2.55667061e-01 -5.01787901e-01 5.35314940e-02 -9.02265459e-02
-1.50383234e-01 -2.98169792e-01 2.83025708e-02 -4.51703042e-01
6.47698063e-03 -9.10016000e-01 -7.42039144e-01 4.08374757e-01
-3.84232879e-01 -4.81698930e-01 6.43596888e-01 -1.90656427e-02
8.62718701e-01 -1.69444275e+00 5.13502300e-01 2.84755677e-01
3.62861842e-01 2.89022446e-01 2.91751977e-02 4.93951291e-01
2.71477878e-01 2.34500811e-01 -7.65064180e-01 -1.00982718e-01
-1.75304860e-01 3.60389382e-01 -1.03017896e-01 4.37660009e-01
2.00868025e-01 9.94523525e-01 -1.21512449e+00 -4.71242875e-01
7.69836679e-02 6.07863188e-01 -6.08995855e-01 2.73592293e-01
-7.23642766e-01 6.28742993e-01 -5.58800459e-01 6.33259177e-01
4.24567193e-01 -6.40784919e-01 1.61309510e-01 1.70781881e-01
-1.94795951e-01 1.38192594e-01 -1.24346292e+00 2.04486632e+00
-6.31582797e-01 2.09676400e-01 -8.81166235e-02 -1.29683542e+00
7.43854821e-01 4.36455518e-01 8.52856815e-01 -4.49434042e-01
-3.07222456e-01 5.49903512e-01 -1.08865641e-01 -2.67548263e-01
1.58053427e-03 -2.33300716e-01 -4.69406471e-02 4.86500949e-01
3.20731364e-02 -1.77957267e-01 3.24031681e-01 1.12385243e-01
1.40684378e+00 1.65011033e-01 3.35097581e-01 -1.62989616e-01
4.99761492e-01 -4.11740057e-02 3.71386886e-01 9.90709007e-01
1.44289821e-01 9.17997733e-02 3.56407285e-01 -5.24876356e-01
-1.20537531e+00 -1.06302881e+00 -4.25499976e-01 1.02338207e+00
-2.81326681e-01 -3.94297004e-01 -2.68673211e-01 -6.81986928e-01
-3.36660212e-03 2.97195584e-01 -2.80989885e-01 -7.95954019e-02
-4.33896065e-01 -1.13604176e+00 4.56831366e-01 1.46065295e-01
3.89422365e-02 -9.52483058e-01 -5.66858828e-01 4.98469561e-01
3.26341599e-01 -7.38741159e-01 7.26200342e-02 5.06554544e-01
-9.35144544e-01 -1.14295673e+00 -7.57531583e-01 -3.60413700e-01
4.43295568e-01 -1.19650438e-01 1.00876641e+00 -1.10558897e-01
-8.81598413e-01 -3.28999050e-02 8.41167662e-03 -3.91192138e-01
-1.65742293e-01 2.99430847e-01 -4.54258397e-02 -1.74451664e-01
-3.45935151e-02 -8.16896677e-01 -6.06251359e-01 5.60108088e-02
-1.16395223e+00 -9.99590829e-02 7.06529140e-01 1.14289474e+00
1.17189777e+00 2.94929883e-03 7.29541779e-01 -1.18151772e+00
2.00921968e-01 -7.91974366e-01 -9.02566850e-01 2.69224554e-01
-6.46308362e-01 5.67600906e-01 7.02659249e-01 -5.19329786e-01
-7.01169729e-01 2.56896973e-01 -3.99300218e-01 -1.90764636e-01
5.94172105e-02 6.53317392e-01 -2.90842563e-01 -2.07736269e-01
7.27129340e-01 -1.25270337e-01 -2.58580983e-01 -4.92230773e-01
4.16735113e-01 4.46550220e-01 6.78704381e-02 -9.55639243e-01
5.78367352e-01 4.40824956e-01 6.79149091e-01 -7.76318729e-01
-9.18836355e-01 -5.97129941e-01 -3.43298197e-01 2.18248248e-01
4.55576569e-01 -6.12079799e-01 -1.14982307e+00 1.69930726e-01
-8.77564430e-01 -1.98207751e-01 -6.39636338e-01 5.34911692e-01
-8.08335423e-01 2.39219129e-01 -1.22861445e-01 -9.16830480e-01
-3.61022532e-01 -1.44591582e+00 1.09303296e+00 3.95189552e-03
6.91002468e-03 -1.15294147e+00 5.15560210e-01 4.20241952e-01
4.81903285e-01 5.44586599e-01 1.14375401e+00 -6.44073486e-01
-1.08926046e+00 -3.70753929e-02 2.52151966e-01 -1.00397512e-01
-4.21833955e-02 -1.51713610e-01 -9.06227767e-01 -2.66200125e-01
-2.49903530e-01 -7.13185728e-01 8.75145674e-01 3.00966561e-01
1.68375123e+00 -5.45560181e-01 -3.57212722e-01 6.64256155e-01
1.34205258e+00 3.41623239e-02 5.20839989e-01 -4.48065959e-02
5.22568405e-01 3.85880589e-01 5.51620722e-01 5.61639309e-01
-3.33275609e-02 7.88974524e-01 3.21678430e-01 -2.71024197e-01
3.92613173e-01 -2.65640646e-01 2.94704307e-02 6.52140498e-01
-1.09227024e-01 -3.26429367e-01 -9.65490997e-01 2.72225857e-01
-1.85197353e+00 -9.99075353e-01 7.44473115e-02 2.56013966e+00
1.49672627e+00 2.69814610e-01 -6.46492187e-03 3.97017628e-01
2.66315132e-01 -1.24158181e-01 -1.01566708e+00 -2.14532748e-01
-2.23905206e-01 8.81513119e-01 3.86187017e-01 4.51661110e-01
-9.05628324e-01 5.01104534e-01 5.98431253e+00 1.30762649e+00
-1.04102182e+00 2.20772550e-01 9.51781571e-01 -2.13193789e-01
-4.09950137e-01 2.89587677e-01 -7.32286274e-01 5.01409054e-01
1.31215382e+00 -1.04447067e-01 4.04979676e-01 7.96047330e-01
1.39930755e-01 -2.80488729e-01 -1.26066470e+00 1.02488852e+00
-5.93069971e-01 -1.89024794e+00 -9.66937616e-02 3.11764568e-01
8.66812170e-01 1.00644557e-02 7.44423866e-02 2.71256566e-02
3.79568428e-01 -1.35950089e+00 1.64273828e-01 8.10979128e-01
8.55462670e-01 -8.74232054e-01 5.13121247e-01 6.38607621e-01
-8.13567340e-01 4.48117293e-02 -4.10597026e-01 2.59324878e-01
2.24686656e-02 1.10978472e+00 -1.07264066e+00 4.00510967e-01
3.61969948e-01 5.11068583e-01 -1.57476306e-01 1.36656737e+00
1.35237709e-01 5.87213635e-01 -7.36980557e-01 -1.72885224e-01
1.64296180e-01 -2.76133865e-01 5.72971344e-01 7.89939284e-01
3.36781859e-01 -1.33207202e-01 3.55585903e-01 9.60516095e-01
-3.34712684e-01 7.60477930e-02 -5.26935458e-01 -5.01034915e-01
5.08577108e-01 1.06420481e+00 -6.23290718e-01 -2.01439410e-01
9.00962204e-02 4.12894875e-01 2.43143886e-01 9.31342989e-02
-6.32295728e-01 -2.19241604e-01 4.26782906e-01 2.35947311e-01
1.54623538e-01 -3.67843032e-01 -7.90899023e-02 -8.01518798e-01
-2.48098657e-01 -9.14290190e-01 5.28893948e-01 -3.58664811e-01
-1.10573113e+00 1.08949952e-01 -2.26824544e-02 -4.91680592e-01
-1.10451028e-01 -4.76347864e-01 -1.03449434e-01 7.47536004e-01
-1.45757616e+00 -8.04830432e-01 2.22586155e-01 3.49204242e-01
4.89753246e-01 -3.15358303e-02 1.05502772e+00 2.96057314e-01
-5.36558032e-01 3.91180098e-01 6.69706941e-01 -5.79695225e-01
5.00648499e-01 -1.17167127e+00 1.87544197e-01 2.86762655e-01
3.89353335e-01 8.10052812e-01 5.52477002e-01 -4.95662868e-01
-2.01905823e+00 -8.34497988e-01 6.12300098e-01 -3.90925735e-01
5.82473338e-01 -6.72489226e-01 -9.31407392e-01 -6.87703199e-04
-5.76569676e-01 4.32764322e-01 8.50543022e-01 2.95488060e-01
-3.23070973e-01 -1.37087628e-01 -1.34969270e+00 1.68446615e-01
9.79196131e-01 -5.61203122e-01 8.39598402e-02 8.29860270e-01
4.91313696e-01 2.87302583e-03 -1.24540913e+00 4.90748584e-01
3.84035707e-01 -4.60748881e-01 1.19994283e+00 -8.89014304e-01
1.70659795e-01 -1.86912805e-01 -3.33192125e-02 -9.49942410e-01
-8.30018427e-03 -1.17781973e+00 -7.37260878e-01 5.54782212e-01
6.41476810e-01 -5.36363184e-01 8.59770954e-01 3.63411576e-01
1.08799390e-01 -1.44221592e+00 -1.27780008e+00 -5.18052638e-01
1.22674547e-01 -2.33621046e-01 4.80766445e-01 7.37264931e-01
-3.04725647e-01 3.72917980e-01 -2.43812531e-01 -3.03839386e-01
7.84030557e-01 7.85022601e-02 5.71757793e-01 -1.62283087e+00
-6.81410134e-01 -1.64868936e-01 -3.68413329e-01 -8.68229330e-01
-4.65711281e-02 -7.88045704e-01 -4.46900316e-02 -1.29033649e+00
3.37589175e-01 -1.13792527e+00 -3.36367160e-01 4.15437996e-01
-2.81699389e-01 1.39713958e-01 -2.66582489e-01 2.35970572e-01
-8.44508708e-01 6.23632848e-01 1.17373037e+00 -2.11957902e-01
-1.00719168e-04 1.58280693e-02 -3.59736651e-01 2.84614176e-01
5.11900544e-01 -6.83165491e-01 -3.35658073e-01 9.50377062e-02
8.47394824e-01 1.33391574e-01 3.21942270e-01 -9.17378306e-01
4.25310999e-01 -5.07069826e-01 4.25009936e-01 -4.55561638e-01
5.21680295e-01 -5.54058433e-01 6.14253819e-01 3.68112355e-01
-4.42582279e-01 -3.06110680e-01 -1.33523107e-01 8.66072714e-01
5.18155247e-02 -2.06623748e-01 8.94994378e-01 -4.13370132e-01
-9.86926034e-02 8.58217299e-01 -1.44365326e-01 1.46249622e-01
1.07592249e+00 2.69114599e-02 -6.01255670e-02 1.21542029e-01
-7.50237405e-01 4.16719913e-02 3.98997247e-01 5.63741140e-02
5.68545520e-01 -9.35892642e-01 -6.83417797e-01 7.98211843e-02
-2.77806018e-02 4.32298243e-01 1.96986690e-01 8.62854600e-01
-3.73147398e-01 6.05610847e-01 -4.93886974e-03 -7.28741884e-01
-8.08727741e-01 6.20503068e-01 9.36367586e-02 -5.83809495e-01
-2.43295819e-01 9.02085900e-01 6.76240996e-02 -2.85002947e-01
1.79855287e-01 1.14789113e-01 1.97491795e-01 6.23654090e-02
4.43480849e-01 5.06611705e-01 3.96290511e-01 -3.09442561e-02
-1.35698229e-01 2.10319579e-01 -1.85781270e-01 6.99239001e-02
1.83862495e+00 4.50467020e-01 -4.72977087e-02 6.23865426e-01
1.14438283e+00 -9.23360065e-02 -1.21093488e+00 -2.53684253e-01
2.27438912e-01 -3.66156280e-01 2.67901748e-01 -7.16621220e-01
-7.08273709e-01 9.09306526e-01 5.38641751e-01 7.10499063e-02
8.07550609e-01 3.14829238e-02 7.49025702e-01 8.34716499e-01
7.41076350e-01 -1.12031984e+00 1.86330661e-01 2.39460319e-01
5.14584839e-01 -1.18056822e+00 3.66364121e-01 -1.37857884e-01
-3.91308293e-02 1.01997793e+00 5.03527895e-02 2.46016964e-01
5.89399993e-01 1.96508482e-01 -6.70674622e-01 -3.06657463e-01
-1.08845055e+00 4.61164154e-02 1.78393170e-01 3.49841624e-01
3.62076372e-01 -9.51195427e-04 -9.11230296e-02 -8.87532905e-02
2.61668712e-01 -1.29592074e-02 1.77771196e-01 1.12480247e+00
-5.51183999e-01 -1.63528657e+00 -3.08529645e-01 6.20492816e-01
-4.34949160e-01 -1.50595590e-01 -9.43415761e-02 3.54547083e-01
2.53701508e-01 6.29503071e-01 -2.84147888e-01 3.32708746e-01
-8.19180831e-02 8.23375769e-03 7.42224753e-01 -7.20708013e-01
-4.56888735e-01 -2.00292006e-01 2.20828168e-02 -6.15817845e-01
-4.48176622e-01 -4.85581845e-01 -1.18332267e+00 -3.17162931e-01
-4.77352589e-01 5.64552009e-01 9.11161184e-01 1.02849579e+00
6.32779896e-01 1.02818735e-01 6.54503465e-01 -6.13252103e-01
-6.60233498e-01 -5.66516399e-01 -1.91832229e-01 -8.30874313e-03
3.17878515e-01 -7.74138629e-01 -2.29513809e-01 -1.36795238e-01] | [5.2215423583984375, 5.278721332550049] |
a4ee5041-960c-478a-8412-820e011471fc | ngram-lstm-open-rate-prediction-model-nlorp | 2302.00651 | null | https://arxiv.org/abs/2302.00651v2 | https://arxiv.org/pdf/2302.00651v2.pdf | Ngram-LSTM Open Rate Prediction Model (NLORP) and Error_accuracy@C metric: Simple effective, and easy to implement approach to predict open rates for marketing email | Our generation has seen an exponential increase in digital tools adoption. One of the unique areas where digital tools have made an exponential foray is in the sphere of digital marketing, where goods and services have been extensively promoted through the use of digital advertisements. Following this growth, multiple companies have leveraged multiple apps and channels to display their brand identities to a significantly larger user base. This has resulted in products, worth billions of dollars to be sold online. Emails and push notifications have become critical channels to publish advertisement content, to proactively engage with their contacts. Several marketing tools provide a user interface for marketers to design Email and Push messages for digital marketing campaigns. Marketers are also given a predicted open rate for the entered subject line. For enabling marketers generate targeted subject lines, multiple machine learning techniques have been used in the recent past. In particular, deep learning techniques that have established good effectiveness and efficiency. However, these techniques require a sizable amount of labelled training data in order to get good results. The creation of such datasets, particularly those with subject lines that have a specific theme, is a challenging and time-consuming task. In this paper, we propose a novel Ngram and LSTM-based modeling approach (NLORPM) to predict open rates of entered subject lines that is easier to implement, has low prediction latency, and performs extremely well for sparse data. To assess the performance of this model, we also devise a new metric called 'Error_accuracy@C' which is simple to grasp and fully comprehensible to marketers. | ['Indradumna Banerjee', 'Shubham Joshi'] | 2023-01-25 | null | null | null | null | ['marketing'] | ['miscellaneous'] | [ 2.75185704e-01 1.46252820e-02 -7.03386843e-01 -5.32763004e-01
-6.14658594e-01 -4.81107116e-01 7.57971883e-01 3.24598074e-01
-3.13137889e-01 4.75168824e-01 -2.05542683e-03 -4.34452325e-01
4.97478154e-03 -1.11226451e+00 -6.62969530e-01 -3.86680178e-02
4.63707857e-02 6.29093409e-01 2.48216927e-01 -3.37514997e-01
2.91855812e-01 3.00998956e-01 -1.54531586e+00 4.26041991e-01
7.49285996e-01 1.50024819e+00 3.37296456e-01 2.26613939e-01
-7.83980191e-01 5.05639255e-01 -5.80984414e-01 -8.64329875e-01
2.90003061e-01 -7.13249892e-02 -4.14294302e-01 -2.07777664e-01
3.27588350e-01 -4.76537794e-01 -1.84669554e-01 8.76648128e-01
3.64201844e-01 -2.07659096e-01 2.77568787e-01 -1.03391564e+00
-5.84446549e-01 1.13968456e+00 -7.41259634e-01 4.14305478e-02
4.32877213e-01 -2.36302063e-01 1.21084619e+00 -4.90315706e-01
5.50930560e-01 1.15687668e+00 5.81917584e-01 2.01279029e-01
-1.13616657e+00 -1.16239631e+00 -7.21791834e-02 3.96634154e-02
-9.79679048e-01 -2.74491571e-02 6.92686677e-01 -4.36300397e-01
6.60505295e-01 2.97017187e-01 7.43525445e-01 1.44910800e+00
3.95228773e-01 7.54633307e-01 1.23441243e+00 -2.72295505e-01
1.53521955e-01 6.25516355e-01 6.63667619e-02 2.07413554e-01
1.13065273e-01 -2.17618197e-02 -6.12645924e-01 1.68650411e-02
7.27445900e-01 3.26602817e-01 2.97820061e-01 1.95668533e-01
-7.68574476e-01 1.36631978e+00 4.35457349e-01 6.31150126e-01
-5.06857336e-01 3.60840112e-02 3.02685708e-01 2.06008002e-01
4.55485135e-01 5.59770584e-01 -5.11500061e-01 -6.98737442e-01
-8.86976302e-01 3.30582052e-01 1.14983070e+00 8.33578348e-01
4.10299420e-01 -2.65392691e-01 2.25539342e-01 9.95185792e-01
2.66856372e-01 5.24190664e-01 6.58473969e-01 -4.34603125e-01
4.94810969e-01 1.03320730e+00 -4.75247391e-02 -1.41503882e+00
-3.21561515e-01 -6.53088272e-01 -8.34715664e-01 -3.69963884e-01
2.16031939e-01 -8.84979405e-03 -6.74255967e-01 1.32662392e+00
1.28941059e-01 -2.10424572e-01 -4.78852749e-01 5.78485310e-01
5.96270025e-01 8.93424571e-01 5.95003180e-02 -3.09761986e-02
1.33455873e+00 -5.36946714e-01 -7.62793183e-01 -3.70977640e-01
3.75024557e-01 -1.10163319e+00 1.33132362e+00 6.31322026e-01
-9.36135948e-01 -6.02136374e-01 -9.56572354e-01 1.71258897e-01
-6.85362935e-01 -1.29871711e-01 1.17806494e+00 8.76087904e-01
-5.16884685e-01 7.70324707e-01 -5.10627627e-01 -3.46729934e-01
6.15397513e-01 6.31973922e-01 -4.68208194e-02 1.17595993e-01
-1.28341985e+00 7.32940018e-01 4.20676842e-02 -1.12136722e-01
-3.43177557e-01 -6.18998885e-01 -3.56849849e-01 3.86659522e-03
4.61999863e-01 -9.15217847e-02 1.38778043e+00 -9.86089289e-01
-1.41717625e+00 6.63214684e-01 2.08290815e-01 -6.12671971e-01
4.19753164e-01 -2.72159249e-01 -7.40629733e-01 -3.29730302e-01
9.05930102e-02 6.90018654e-01 8.24845135e-01 -9.07618701e-01
-9.47988451e-01 -4.11976665e-01 1.27207693e-02 -2.88298994e-01
-8.51052701e-01 2.31279746e-01 -3.00816059e-01 -4.76782173e-01
-4.85256538e-02 -7.88050234e-01 -7.93700665e-02 -5.53935707e-01
-4.08299863e-01 -2.86117554e-01 6.94068193e-01 -7.70267248e-01
1.52722001e+00 -2.03558397e+00 -2.48590916e-01 4.69386309e-01
2.60006875e-01 2.80279994e-01 2.04323128e-01 6.69325531e-01
5.33944666e-01 4.11119580e-01 2.52942979e-01 -4.55471948e-02
2.66438127e-01 9.15286317e-02 -4.13905829e-01 -2.91805595e-01
-9.92593095e-02 8.66997778e-01 -5.02701759e-01 -2.48790100e-01
-4.05343510e-02 4.65492874e-01 -4.21838045e-01 9.14369226e-02
-4.55128998e-01 1.33698687e-01 -5.02711058e-01 8.15618157e-01
4.81441379e-01 -5.54348946e-01 1.65017411e-01 2.61034239e-02
-7.78741688e-02 3.71512681e-01 -9.72533047e-01 1.11058939e+00
-7.40802765e-01 4.93415207e-01 -4.78632748e-02 -5.09664714e-01
1.25843477e+00 -1.52253479e-01 5.74125767e-01 -1.10905528e+00
5.23819566e-01 5.60438931e-01 7.98926950e-02 -2.86399901e-01
7.70861089e-01 4.13075648e-02 -2.42392659e-01 3.08331966e-01
-2.14064538e-01 2.21584409e-01 1.67416334e-01 8.46950710e-02
1.08120751e+00 -1.30888164e-01 3.60712335e-02 2.92460293e-01
1.91105425e-01 -5.35592139e-02 3.43189985e-01 6.44640744e-01
1.44357428e-01 -4.10284773e-02 2.97882169e-01 -4.71581787e-01
-1.05925632e+00 -6.78645670e-01 -1.13195397e-01 1.15106189e+00
-2.82900222e-02 -2.82302618e-01 -6.17439866e-01 -5.23724258e-01
1.63245454e-01 5.58424950e-01 -2.50669867e-01 5.60700037e-02
-6.16744161e-01 -3.52543533e-01 3.08147222e-01 4.98062819e-01
6.46221399e-01 -1.07383394e+00 -2.95007825e-01 6.65570021e-01
1.45370131e-02 -1.29688859e+00 -3.15954775e-01 1.23510204e-01
-8.86474490e-01 -6.20698929e-01 -6.57321393e-01 -7.37144232e-01
2.03152344e-01 -5.07887118e-02 8.82268727e-01 -3.31803977e-01
-1.54642776e-01 -1.21012226e-01 -4.04481173e-01 -6.99501634e-01
-4.71461743e-01 7.17180014e-01 -3.11639346e-02 2.37313733e-01
8.12901616e-01 -5.34087479e-01 -4.80205595e-01 6.67706430e-01
-9.67105210e-01 -1.11130133e-01 9.35332596e-01 4.49807227e-01
4.72227097e-01 -5.56221083e-02 9.53245938e-01 -1.06824148e+00
8.86345625e-01 -5.92523336e-01 -7.00665832e-01 -2.04866186e-01
-7.92564332e-01 -4.00119573e-01 5.58019221e-01 -7.99527824e-01
-6.60892308e-01 -4.99974936e-02 -4.80880558e-01 3.10353190e-02
1.04964703e-01 8.07614505e-01 2.97364313e-02 -2.24850401e-02
5.88807821e-01 -3.45193446e-02 1.03462569e-01 -8.01527500e-01
3.32573116e-01 1.27828944e+00 7.68429637e-02 -1.19443789e-01
6.79358423e-01 1.73691049e-01 -3.34432751e-01 -8.38117123e-01
-5.72775960e-01 -4.28898692e-01 -1.51447922e-01 -3.81281346e-01
5.65558076e-01 -5.87110519e-01 -1.06567895e+00 3.83213788e-01
-9.50613737e-01 5.50355054e-02 -9.80552584e-02 4.40244704e-01
-1.45186037e-01 -3.71251325e-03 -6.68430746e-01 -7.01200247e-01
-4.12808120e-01 -1.07937992e+00 8.18821251e-01 4.02145296e-01
-6.71679318e-01 -7.36856103e-01 -3.18679780e-01 5.99816859e-01
8.33807886e-01 1.16147429e-01 9.90836859e-01 -1.09984076e+00
-7.90447474e-01 -8.77387762e-01 -3.89636427e-01 3.91148984e-01
9.87229869e-02 -2.73383468e-01 -7.32581496e-01 3.88206579e-02
-1.42444253e-01 -9.26394910e-02 3.41239244e-01 4.54965174e-01
1.21659923e+00 -4.18410093e-01 -5.57642460e-01 1.44745335e-01
1.14870715e+00 6.30520523e-01 5.87475002e-01 5.40132821e-01
4.41265225e-01 6.85115874e-01 6.53934062e-01 3.59301835e-01
3.89082253e-01 8.03207874e-01 3.57194364e-01 -9.53142866e-02
3.33483428e-01 -6.65986180e-01 3.43076706e-01 1.09186578e+00
1.26286581e-01 -1.13785788e-01 -5.67994595e-01 1.41059771e-01
-1.54205847e+00 -6.96133614e-01 -1.25963360e-01 2.16389275e+00
9.09355521e-01 6.64547384e-01 3.66724283e-01 2.12788299e-01
6.61753356e-01 -2.72496250e-02 -4.51802582e-01 -7.11111009e-01
1.78779915e-01 3.97116244e-01 7.85158277e-01 8.35551396e-02
-8.31660151e-01 9.57892239e-01 5.77469730e+00 1.10573089e+00
-1.22081196e+00 1.46134079e-01 7.46985257e-01 5.51400445e-02
-3.47884595e-01 -1.83823407e-01 -1.35016286e+00 9.43743348e-01
1.30147743e+00 9.19794664e-02 4.20183927e-01 1.11784053e+00
2.56968170e-01 -6.41373545e-02 -9.04336631e-01 1.08353007e+00
-2.48858184e-01 -1.71045959e+00 -8.70333835e-02 5.88186502e-01
4.90862489e-01 -9.06595066e-02 2.91909933e-01 4.77392435e-01
5.70464171e-02 -8.86509657e-01 4.32791650e-01 2.77527481e-01
5.88644922e-01 -6.82529807e-01 8.10563564e-01 2.59340078e-01
-8.05066645e-01 -3.48806024e-01 -2.03950301e-01 -1.47175282e-01
4.44561213e-01 6.95276797e-01 -1.00046194e+00 5.13377264e-02
6.87013865e-01 4.10906583e-01 -2.92413205e-01 9.90126967e-01
2.22120658e-01 7.21190572e-01 -4.49824691e-01 -8.06843281e-01
4.05411780e-01 -2.84209639e-01 1.72978282e-01 9.42021191e-01
4.60021883e-01 -4.83277321e-01 1.36745974e-01 7.77650952e-01
-4.66731101e-01 5.02931058e-01 -5.89025676e-01 -8.48618686e-01
4.78597134e-01 1.43063676e+00 -8.78733516e-01 -4.07599472e-02
-5.22987247e-01 5.29855788e-01 -2.62880385e-01 -6.89625889e-02
-9.49728906e-01 -5.54465652e-01 4.16603506e-01 8.42329860e-01
5.21219708e-02 -2.66239166e-01 -2.43868411e-01 -5.15204906e-01
6.68765381e-02 -1.20859671e+00 -2.01701112e-02 -3.26116741e-01
-1.39849269e+00 6.48793936e-01 -2.51450002e-01 -1.03714895e+00
-2.95780301e-01 -6.07941449e-01 -8.72818306e-02 6.28297687e-01
-1.22225296e+00 -1.13881969e+00 -1.98195785e-01 1.70470670e-01
6.32445931e-01 -3.46462876e-01 5.82217455e-01 9.40555394e-01
-2.53517091e-01 6.44880950e-01 -4.30982932e-02 -7.89270489e-05
4.19067204e-01 -9.09258664e-01 5.43136239e-01 1.39756277e-01
2.95802146e-01 7.13951826e-01 4.98094559e-01 -7.84962952e-01
-1.53079069e+00 -7.95838296e-01 1.19079816e+00 -3.69546026e-01
8.75188887e-01 -7.27395236e-01 -5.53951502e-01 6.27376139e-01
-4.41174395e-02 -9.22490358e-01 8.47397685e-01 4.01937723e-01
8.38206708e-03 -5.29727697e-01 -9.44306552e-01 4.05381739e-01
8.02834570e-01 -3.10754061e-01 -1.69970393e-01 4.48362380e-01
7.91222453e-01 -2.11046502e-01 -1.03331065e+00 1.87058523e-01
6.58061981e-01 -8.12367857e-01 6.26937270e-01 -2.39459634e-01
6.04558289e-01 2.98362553e-01 -1.54134497e-01 -8.49023879e-01
-7.85615947e-03 -8.34288001e-01 -1.56433135e-01 1.58051944e+00
8.37324977e-01 -9.25984979e-01 9.80243623e-01 5.89026570e-01
1.81453422e-01 -1.10636389e+00 -5.91865301e-01 -6.60521805e-01
-3.93596619e-01 -6.78398430e-01 8.65982711e-01 6.97090030e-01
-6.93947449e-02 5.59945285e-01 -5.66101909e-01 -4.16554570e-01
1.21854529e-01 1.03225395e-01 8.14692080e-01 -1.51824927e+00
-4.36654419e-01 -5.11305332e-01 -3.83923262e-01 -1.29690254e+00
-4.30780172e-01 -7.94009924e-01 -5.78767478e-01 -1.45701027e+00
-8.70942250e-02 -6.49982393e-01 7.15514123e-02 2.75550902e-01
5.91747284e-01 2.63923496e-01 1.88877303e-02 3.16174209e-01
-1.79320112e-01 6.74186498e-02 1.23716879e+00 -1.32152364e-01
-5.36770821e-01 5.22239268e-01 -1.03018975e+00 5.39989650e-01
8.27556014e-01 -2.82721847e-01 -1.68830857e-01 2.22832672e-02
9.01357710e-01 -1.45413771e-01 -1.49042606e-01 -8.49694967e-01
4.76157963e-02 1.93342030e-01 3.86845708e-01 -6.18652821e-01
4.62259263e-01 -8.85668457e-01 1.19690493e-01 3.36499244e-01
-4.33563471e-01 1.63889453e-01 -4.02241610e-02 4.94647712e-01
-3.37082595e-01 -1.53007597e-01 4.19859022e-01 1.06299393e-01
-5.31422496e-01 3.40712875e-01 -3.82168859e-01 -2.63941318e-01
9.19503629e-01 -4.13746566e-01 -9.77883637e-02 -6.30532086e-01
-5.24257064e-01 -3.52596603e-02 7.94565007e-02 8.81300449e-01
3.77772987e-01 -1.07371926e+00 -1.88760787e-01 2.83099204e-01
-1.41118824e-01 -4.58983511e-01 1.34666795e-02 5.87888539e-01
-4.51342881e-01 5.22461414e-01 -1.77555829e-01 -4.47082609e-01
-1.01537812e+00 2.58505702e-01 -1.29074007e-01 -2.48746291e-01
-4.49259430e-01 8.37210894e-01 -4.53959435e-01 -2.87436277e-01
5.15975654e-01 -5.87533534e-01 -3.40059072e-01 4.83580589e-01
6.77114546e-01 3.63023341e-01 8.65806639e-02 -3.97896439e-01
-4.87182476e-02 2.43054613e-01 -4.91276383e-01 1.16853518e-02
1.43236363e+00 1.52210787e-01 9.88788903e-02 3.79611522e-01
1.23016191e+00 3.75786796e-02 -8.98718655e-01 -1.37789980e-01
1.73100129e-01 -6.98019087e-01 1.49239264e-02 -1.13261402e+00
-1.05414677e+00 8.53286147e-01 7.02203751e-01 9.22890067e-01
7.54545867e-01 7.49254748e-02 1.58128583e+00 -1.16444789e-01
6.27741933e-01 -1.25416148e+00 1.78306609e-01 2.07790375e-01
5.40275693e-01 -1.20608747e+00 -3.20420712e-01 -4.43003714e-01
-5.14635265e-01 8.10331702e-01 3.97128850e-01 3.46669674e-01
6.08463526e-01 3.40688109e-01 -2.84119733e-02 -1.90656185e-01
-2.41532937e-01 1.26906082e-01 1.17373727e-01 4.12995636e-01
4.45109904e-01 1.85873866e-01 -4.95735228e-01 9.22576547e-01
-4.74970490e-01 1.77821487e-01 2.53128529e-01 7.28667736e-01
-5.85030913e-01 -1.59539282e+00 2.96513774e-02 1.08694530e+00
-8.71334612e-01 4.51115845e-03 -4.80439335e-01 7.85174787e-01
-1.18518911e-01 9.59909320e-01 -8.67892504e-02 -7.95380712e-01
4.08639193e-01 -6.40192553e-02 3.13256867e-02 -5.07539690e-01
-8.34263563e-01 1.67866737e-01 2.77255416e-01 -4.81303334e-01
-5.10860011e-02 -3.71945590e-01 -1.04256546e+00 -7.91600466e-01
-5.71611524e-01 8.97903070e-02 1.35940504e+00 7.62927115e-01
5.08832574e-01 4.39606667e-01 7.07190275e-01 -6.93142593e-01
-7.51918495e-01 -1.00262082e+00 -6.31134331e-01 2.03955546e-01
-4.96746004e-01 -6.51476920e-01 -7.53196850e-02 -4.44186032e-01] | [9.923748016357422, 5.943975925445557] |
37e44ec1-a5f7-4f60-8930-177905ca0a81 | learning-parameters-for-balanced-index | 2012.08067 | null | https://arxiv.org/abs/2012.08067v1 | https://arxiv.org/pdf/2012.08067v1.pdf | Learning Parameters for Balanced Index Influence Maximization | Influence maximization is the task of finding the smallest set of nodes whose activation in a social network can trigger an activation cascade that reaches the targeted network coverage, where threshold rules determine the outcome of influence. This problem is NP-hard and it has generated a significant amount of recent research on finding efficient heuristics. We focus on a {\it Balance Index} algorithm that relies on three parameters to tune its performance to the given network structure. We propose using a supervised machine-learning approach for such tuning. We select the most influential graph features for the parameter tuning. Then, using random-walk-based graph-sampling, we create small snapshots from the given synthetic and large-scale real-world networks. Using exhaustive search, we find for these snapshots the high accuracy values of BI parameters to use as a ground truth. Then, we train our machine-learning model on the snapshots and apply this model to the real-word network to find the best BI parameters. We apply these parameters to the sampled real-world network to measure the quality of the sets of initiators found this way. We use various real-world networks to validate our approach against other heuristic. | ['Boleslaw K. Szymanski', 'Gyorgy Korniss', 'Manqing Ma'] | 2020-12-15 | null | null | null | null | ['graph-sampling'] | ['graphs'] | [ 2.79869556e-01 4.58844990e-01 -5.65929174e-01 -8.79744962e-02
-4.12729353e-01 -7.76215792e-01 7.56656766e-01 9.54421088e-02
-2.32094169e-01 8.64202797e-01 -1.14412848e-02 -3.35984945e-01
-5.92636228e-01 -1.22943866e+00 -6.22233748e-01 -7.07344115e-01
-6.53115034e-01 1.01732874e+00 4.82391238e-01 -2.10872248e-01
5.30355930e-01 3.22930992e-01 -1.02263010e+00 -2.91844215e-02
8.38966668e-01 4.86678421e-01 -2.04251409e-01 7.44705081e-01
9.08552408e-02 5.13218462e-01 -8.71758878e-01 -3.32438767e-01
4.48174119e-01 -6.20973229e-01 -8.61429989e-01 1.73191428e-01
-4.35093015e-01 1.92969233e-01 -2.54971236e-01 8.58743370e-01
3.84088039e-01 -2.04311265e-03 7.04685330e-01 -1.34719706e+00
2.03119457e-01 1.18218517e+00 -6.40757680e-01 4.55000192e-01
4.88840014e-01 2.20057786e-01 1.29275966e+00 -7.29644150e-02
1.01567256e+00 1.20688081e+00 2.56432176e-01 7.31920749e-02
-1.46538210e+00 -5.20196438e-01 7.79450387e-02 -1.98462848e-02
-1.19515777e+00 -3.14715773e-01 9.76262271e-01 -2.21171945e-01
4.08540547e-01 2.46488690e-01 1.10465133e+00 8.92580211e-01
9.08438023e-03 3.55989516e-01 1.08831954e+00 -4.88517702e-01
4.32747215e-01 5.01488522e-02 1.13146426e-03 7.79121459e-01
4.98857468e-01 2.05039456e-02 -4.74814534e-01 -6.41794860e-01
4.85445976e-01 -3.80363613e-01 -2.85654306e-01 -1.02218367e-01
-1.05926597e+00 9.17044640e-01 6.04277432e-01 4.96925324e-01
-4.57756162e-01 3.06888312e-01 1.84008833e-02 6.13922238e-01
5.47117174e-01 8.87227714e-01 -4.31921780e-01 1.64104309e-02
-9.35235739e-01 7.08148815e-03 1.43834913e+00 4.18694109e-01
1.06911016e+00 -4.70749110e-01 -1.43454418e-01 7.23076820e-01
3.39681119e-01 3.80000293e-01 -2.65717328e-01 -8.40898156e-01
3.71205151e-01 9.86167610e-01 7.34834746e-02 -1.30079007e+00
-2.89177656e-01 -5.90542376e-01 -6.35274649e-01 -1.59938142e-01
6.11477196e-01 -6.18871987e-01 -7.68911064e-01 1.60143101e+00
6.45629704e-01 2.55078793e-01 -4.72445369e-01 6.68017387e-01
1.90852247e-02 8.09569418e-01 -3.38714480e-01 -5.88649571e-01
6.98999047e-01 -7.32707024e-01 -1.51095599e-01 -1.28058285e-01
5.16661108e-01 -4.90694672e-01 6.73008204e-01 3.28183025e-01
-8.38522613e-01 2.05536559e-01 -1.08771861e+00 9.36549008e-01
-1.72692850e-01 -3.81683290e-01 6.13057971e-01 6.84478760e-01
-1.12810266e+00 8.40284705e-01 -4.50443864e-01 -5.76273859e-01
2.33465642e-01 5.52098870e-01 1.09917544e-01 -8.00268501e-02
-1.31539106e+00 6.20481193e-01 3.44455451e-01 -6.08784333e-02
-1.13128829e+00 -3.59962553e-01 -1.52592719e-01 -9.93374959e-02
1.15495849e+00 -7.93949544e-01 7.53412724e-01 -1.08906853e+00
-1.34305394e+00 6.48695052e-01 -6.97715208e-02 -2.63745159e-01
3.69494498e-01 4.71522093e-01 -1.12216406e-01 2.58023471e-01
1.29605412e-01 1.52982786e-01 8.95773590e-01 -1.36642087e+00
-4.91847515e-01 -1.69527784e-01 3.48375291e-01 5.39825335e-02
-2.60871172e-01 -4.33737598e-02 -7.16691434e-01 -6.96494356e-02
7.74962083e-02 -1.19408453e+00 -7.34992266e-01 -7.42047489e-01
-9.51101720e-01 -3.40244979e-01 1.62462771e-01 -1.41229369e-02
1.61974597e+00 -1.49806023e+00 3.27877522e-01 1.40967762e+00
4.28959519e-01 -1.38481110e-01 -2.64315933e-01 7.29759634e-01
3.80982488e-01 6.50235057e-01 9.96259786e-03 2.23809108e-01
-3.85898679e-01 8.36650282e-02 1.68702051e-01 4.87971455e-01
-1.12695962e-01 6.41807735e-01 -1.06208432e+00 -5.19911647e-01
-1.86299190e-01 -8.38338435e-02 -5.49378991e-01 2.36802503e-01
-5.44464171e-01 2.82509059e-01 -9.20736969e-01 4.85454798e-01
1.75737053e-01 -4.85723794e-01 7.80093968e-01 2.01730132e-01
1.37152910e-01 8.71821493e-02 -1.31766605e+00 7.33868837e-01
-1.07534654e-01 2.55341887e-01 1.81865245e-01 -8.81488740e-01
8.79563987e-01 -9.96993482e-02 8.08323324e-01 -2.51059383e-01
3.83711994e-01 2.22236305e-01 4.51579124e-01 -4.08200383e-01
-9.22291577e-02 2.68034339e-01 -9.84711125e-02 9.78084624e-01
-2.17199594e-01 2.86728609e-02 7.67344534e-01 6.98727310e-01
1.80458462e+00 -7.40703344e-01 2.92326182e-01 -3.37003380e-01
4.27510262e-01 1.37991786e-01 2.84548640e-01 9.55631018e-01
1.12143785e-01 -1.47881405e-02 1.26637375e+00 -3.95742476e-01
-1.02718949e+00 -5.16228318e-01 3.73516828e-01 1.12470424e+00
1.26908541e-01 -4.23891187e-01 -1.04190445e+00 -8.00367057e-01
3.48993167e-02 3.45247567e-01 -8.83395493e-01 -4.56077978e-02
-5.55508316e-01 -1.00777578e+00 2.15527892e-01 -4.95178998e-01
1.74625590e-01 -1.33473742e+00 -6.52141869e-02 4.61888164e-01
-1.49081200e-01 -7.88615942e-01 -2.91546762e-01 3.47165274e-03
-7.19075024e-01 -1.74619079e+00 -1.12066463e-01 -4.54981029e-01
9.90948915e-01 4.45261337e-02 1.34422755e+00 6.63974464e-01
-3.01652569e-02 2.24917918e-01 -2.60492235e-01 -8.26546177e-02
-6.26365125e-01 7.11853147e-01 -1.60327643e-01 3.76379073e-01
-4.10575047e-02 -8.86435091e-01 -5.51779866e-01 5.72453499e-01
-6.73162580e-01 -2.85424113e-01 6.81512177e-01 3.48228186e-01
4.35274988e-01 3.42164934e-01 7.81755805e-01 -1.38353479e+00
1.30134571e+00 -8.23803246e-01 -7.30884314e-01 3.64338607e-01
-9.91772056e-01 2.77696133e-01 4.38898802e-01 -4.52350706e-01
-4.78791773e-01 -2.32779473e-01 4.50921357e-01 3.74840060e-03
3.89075071e-01 8.88109803e-01 -9.93935019e-02 -1.12093963e-01
8.78082514e-01 -1.72408432e-01 -5.39173931e-02 2.50048693e-02
4.40931261e-01 4.23093408e-01 -1.72643840e-01 -6.59469903e-01
8.83835614e-01 3.36731791e-01 2.73724675e-01 -6.92969799e-01
-8.84971142e-01 -3.79428506e-01 -3.76247495e-01 -6.96135163e-01
9.70953852e-02 -1.39595568e-01 -7.47642457e-01 2.07469836e-01
-8.73598933e-01 -4.00147319e-01 6.05998971e-02 -8.35193768e-02
-3.27995777e-01 -8.36743563e-02 -3.14402878e-01 -8.69336128e-01
-4.93573040e-01 -7.96123147e-01 5.10505140e-01 6.34796545e-02
-2.75062263e-01 -9.91184652e-01 5.62247515e-01 1.36666745e-01
2.87122846e-01 6.19874477e-01 7.82407880e-01 -8.49880397e-01
-9.32960927e-01 -3.23005229e-01 2.64144223e-02 -2.79101968e-01
1.22421626e-02 4.71721619e-01 -3.42508167e-01 -3.19144666e-01
-5.01020312e-01 6.67531788e-02 8.34819376e-01 5.46189010e-01
8.22476387e-01 -6.32013738e-01 -9.46639478e-01 3.87222886e-01
1.43828976e+00 8.70482102e-02 4.46318507e-01 3.44954610e-01
2.82040656e-01 4.79796231e-01 6.08284175e-01 3.88499826e-01
1.03564948e-01 4.65978503e-01 5.40409207e-01 2.62769610e-02
3.76090229e-01 -4.19481903e-01 1.59547687e-01 6.02090418e-01
-3.12422931e-01 -6.27939641e-01 -9.34536934e-01 6.09317899e-01
-1.88890326e+00 -8.44742060e-01 7.37657323e-02 2.23588729e+00
9.46768999e-01 6.73272014e-01 5.97435176e-01 2.37860471e-01
9.49651062e-01 2.38434061e-01 -4.82869029e-01 -1.39171079e-01
1.55793622e-01 3.65708768e-02 5.66656768e-01 7.99456596e-01
-7.66907275e-01 1.02599072e+00 6.11296606e+00 7.18231082e-01
-8.64063203e-01 -1.69241443e-01 8.68311346e-01 -3.18993390e-01
-3.91870290e-01 3.97160977e-01 -6.02801263e-01 3.57850283e-01
1.15071905e+00 -5.45432210e-01 1.05321884e+00 4.30953950e-01
6.35376751e-01 -2.82272577e-01 -9.06650960e-01 3.04083198e-01
-3.17739725e-01 -1.42724395e+00 -1.99834749e-01 2.97968477e-01
1.08191824e+00 1.81190848e-01 -5.13286173e-01 -1.03829496e-01
8.35662961e-01 -8.63275707e-01 1.06669866e-01 5.54000735e-01
4.65218127e-01 -9.09677863e-01 4.65028733e-01 5.79598665e-01
-8.30880702e-01 -2.17967793e-01 -1.39780954e-01 1.27627745e-01
1.09475836e-01 1.36983860e+00 -1.44728875e+00 1.73127234e-01
2.69410610e-01 4.78880793e-01 -4.10211027e-01 1.16859603e+00
-5.07079124e-01 1.13826728e+00 -6.85301960e-01 -7.06278086e-01
1.90890193e-01 -4.04209763e-01 9.64559793e-01 8.73888195e-01
-7.22315833e-02 -3.53696048e-02 2.70761281e-01 7.08862066e-01
-4.11011845e-01 1.52219698e-01 -7.42527843e-01 -2.70357400e-01
8.61696839e-01 1.56107652e+00 -1.15673721e+00 -1.42459795e-01
3.55286986e-01 4.28913444e-01 5.41872740e-01 4.61416394e-01
-5.25731862e-01 -3.00557554e-01 3.69246185e-01 5.94464958e-01
1.06258571e-01 -1.58603564e-01 -8.11342746e-02 -6.83022738e-01
-2.63409823e-01 -9.52366829e-01 5.59118152e-01 -1.87809289e-01
-1.06120205e+00 6.25953972e-01 3.91808040e-02 -6.22439265e-01
-4.21659738e-01 -3.23506892e-02 -1.03587663e+00 3.65675122e-01
-1.02403593e+00 -6.69968247e-01 -9.83323082e-02 3.21263045e-01
5.28890714e-02 -2.16018856e-02 2.32445225e-01 -6.11627996e-02
-6.90032184e-01 1.95535198e-01 -4.84985821e-02 3.46488063e-03
3.36737394e-01 -1.03773415e+00 1.92037791e-01 7.45104432e-01
1.32599071e-01 4.44342852e-01 8.49947393e-01 -9.25105929e-01
-1.32883239e+00 -9.29372430e-01 8.59662712e-01 -3.02829742e-01
8.31138790e-01 -3.68530303e-01 -3.07898670e-01 5.32823384e-01
1.42017886e-01 -2.70485371e-01 2.47027516e-01 3.91957849e-01
1.23356417e-01 -2.65405089e-01 -1.22144651e+00 7.83210576e-01
1.31614912e+00 2.37157673e-01 1.39556319e-01 5.57274103e-01
6.87208533e-01 1.27792299e-01 -6.44986928e-01 1.49199620e-01
2.18747154e-01 -7.73365259e-01 6.14808023e-01 -6.86911583e-01
3.74054760e-01 -2.29475379e-01 5.07213354e-01 -1.85459387e+00
-3.83351535e-01 -1.27840853e+00 2.38837935e-02 1.16070795e+00
9.79663908e-01 -7.52433777e-01 8.84916246e-01 1.81662083e-01
7.95256317e-01 -1.25246823e+00 -7.47219622e-01 -1.99527770e-01
-4.48376119e-01 -8.59146640e-02 7.12052643e-01 8.04830432e-01
4.45333757e-02 8.08084846e-01 -1.84825674e-01 2.30684988e-02
9.23237801e-01 2.55940616e-01 1.01124883e+00 -1.49325001e+00
-2.41465107e-01 -4.29305702e-01 -2.37326063e-02 -6.43957675e-01
9.44099948e-02 -7.99385667e-01 -2.34745964e-01 -1.53461194e+00
3.67442966e-01 -7.95149863e-01 -1.44828102e-02 2.93743283e-01
-1.40222013e-01 -1.13923073e-01 -1.59918547e-01 1.49862468e-01
-8.34632993e-01 8.85364264e-02 1.40833199e+00 -7.15747401e-02
-6.51690125e-01 5.11258841e-01 -8.28030944e-01 4.39496368e-01
9.57390666e-01 -9.07983780e-01 -4.83590215e-01 2.63321400e-01
8.27770650e-01 2.24483252e-01 -1.86332107e-01 -5.78193843e-01
3.01403850e-01 -5.99461079e-01 -2.58332323e-02 -2.32108131e-01
-2.06453636e-01 -7.03183830e-01 1.29252464e-01 5.17959535e-01
-4.72353309e-01 -1.58017278e-01 -4.84103143e-01 7.66473532e-01
3.61600310e-01 -2.45432630e-01 5.79572976e-01 -1.24056041e-01
3.26436348e-02 4.65223402e-01 -2.47762799e-01 3.48661244e-01
1.08249056e+00 -6.20958097e-02 -1.87236205e-01 -8.89314055e-01
-6.88586295e-01 5.75410545e-01 3.96278113e-01 7.48705342e-02
3.32248569e-01 -8.48175943e-01 -9.14255142e-01 -1.71977729e-02
-2.83664852e-01 -1.90573990e-01 -5.53502440e-01 8.43332291e-01
-2.92732775e-01 5.05410023e-02 1.69359714e-01 -4.46918488e-01
-1.09253335e+00 3.97597998e-01 5.38546324e-01 -8.16102326e-01
-5.14053926e-02 4.44403529e-01 -6.75688922e-01 -3.36201161e-01
-2.87609845e-02 1.69002011e-01 -2.86524087e-01 5.24875596e-02
1.12229615e-01 5.95460653e-01 -2.65875995e-01 -7.77369961e-02
-2.60575801e-01 2.20148087e-01 2.75780886e-01 -3.27907622e-01
1.48317206e+00 -1.00851893e-01 -5.11305571e-01 7.95832872e-02
9.93983626e-01 3.20356876e-01 -7.55149543e-01 -2.08244696e-01
3.32414418e-01 -4.19986874e-01 -9.40536112e-02 -6.59831405e-01
-1.32241535e+00 3.05567458e-02 -1.93332344e-01 1.04887235e+00
8.67497146e-01 2.66618729e-01 1.73683584e-01 5.01149178e-01
5.55074453e-01 -1.12030804e+00 1.55531138e-01 4.53890324e-01
6.28116548e-01 -8.57432187e-01 2.64702082e-01 -6.38342202e-01
-2.47670010e-01 8.84562194e-01 5.28831720e-01 -2.97225684e-01
9.80184317e-01 1.80056900e-01 -2.42264479e-01 -5.72914422e-01
-1.13422644e+00 -2.63035685e-01 -4.49301414e-02 2.56462008e-01
-4.08130437e-02 2.50727475e-01 -5.87228358e-01 6.15386814e-02
-3.14633101e-01 -6.41801283e-02 5.62336445e-01 4.82433766e-01
-8.17739129e-01 -1.24377453e+00 -2.52961457e-01 9.57318902e-01
-2.65162766e-01 1.55099675e-01 -1.15519047e+00 5.79585969e-01
-2.95972407e-01 1.26454878e+00 -2.89225578e-01 -4.95386064e-01
2.79869854e-01 -3.24763596e-01 4.07049894e-01 -5.00220358e-01
-7.29701400e-01 -7.79580325e-02 6.64673090e-01 -5.19196510e-01
-2.94393092e-01 -5.31246781e-01 -9.39229012e-01 -6.52554989e-01
-5.76244116e-01 5.11773705e-01 4.70784694e-01 9.18770134e-01
4.00678247e-01 3.03800464e-01 1.28596222e+00 -6.46627545e-01
-4.56501603e-01 -1.11257482e+00 -4.54193473e-01 2.79965401e-01
-8.33102018e-02 -5.42466164e-01 -9.17125344e-01 -5.79540789e-01] | [6.927389621734619, 5.40385627746582] |
76724824-14e6-4f89-bc72-44bf856c1545 | about-evaluation-of-f1-score-for-recent | 2305.09410 | null | https://arxiv.org/abs/2305.09410v1 | https://arxiv.org/pdf/2305.09410v1.pdf | About Evaluation of F1 Score for RECENT Relation Extraction System | This document contains a discussion of the F1 score evaluation used in the article 'Relation Classification with Entity Type Restriction' by Shengfei Lyu, Huanhuan Chen published on Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021. The authors created a system named RECENT and claim it achieves (then) a new state-of-the-art result 75.2 (previous 74.8) on the TACRED dataset, while after correcting errors and reevaluation the final result is 65.16 | ['Michał Olek'] | 2023-05-16 | null | null | null | null | ['relation-extraction', 'relation-classification'] | ['natural-language-processing', 'natural-language-processing'] | [-1.39064878e-01 6.66560948e-01 -6.53646052e-01 -4.29754555e-01
-6.37491405e-01 -2.75032282e-01 6.51694596e-01 5.92722833e-01
-9.18724179e-01 1.14646375e+00 4.33206260e-01 -4.48623627e-01
-4.05014008e-01 -5.23236990e-01 -4.81864512e-01 -8.09855089e-02
3.52645181e-02 5.83863974e-01 2.35249132e-01 -4.08532798e-01
4.45317477e-01 2.38538548e-01 -1.67787862e+00 5.08714736e-01
1.01368523e+00 7.08559096e-01 -2.22137228e-01 5.97821832e-01
-2.20745504e-02 8.83372724e-01 -5.47245502e-01 -1.03473592e+00
-5.25870733e-02 -3.79707776e-02 -1.67809868e+00 -6.84877872e-01
4.80868071e-01 2.73523271e-01 -3.19587857e-01 1.00525630e+00
4.39127624e-01 1.51346475e-01 5.77041864e-01 -1.20147884e+00
-9.78689015e-01 1.28473485e+00 -3.30889285e-01 5.71765065e-01
7.33223796e-01 -6.55036271e-01 1.28696489e+00 -9.52182651e-01
1.26836896e+00 8.74146640e-01 9.41283226e-01 4.95576084e-01
-1.10489047e+00 -7.85097063e-01 -1.16479002e-01 4.53106254e-01
-1.64367855e+00 -6.96899414e-01 2.00665846e-01 -2.94574469e-01
1.58914459e+00 4.73598242e-01 3.88280571e-01 8.98068488e-01
-9.42449719e-02 7.21917272e-01 1.14311075e+00 -8.61535311e-01
-7.51530901e-02 5.09399809e-02 5.32958865e-01 3.50072443e-01
8.10640395e-01 -4.47671488e-02 -4.59964424e-01 -2.14967251e-01
2.28936430e-02 -8.84955287e-01 4.09711245e-03 2.31693864e-01
-9.47386861e-01 6.67599618e-01 2.62186915e-01 8.57404053e-01
-4.39201385e-01 -4.50872451e-01 7.32978821e-01 5.99741518e-01
7.55820751e-01 5.16262114e-01 -8.94016027e-01 -3.05414259e-01
-7.17784524e-01 5.42885065e-01 1.06411612e+00 1.21790731e+00
1.98556378e-01 -4.72356707e-01 4.63395789e-02 9.74828482e-01
3.41743529e-01 1.26485407e-01 4.40003932e-01 -7.48478472e-01
6.82876885e-01 7.02878892e-01 -5.21699106e-03 -6.74398839e-01
-5.84492326e-01 -7.47844130e-02 -5.50404072e-01 -3.75110447e-01
4.65724945e-01 -3.10550153e-01 -4.96413052e-01 1.44208956e+00
1.53557241e-01 -3.98548186e-01 6.50035203e-01 4.18745399e-01
1.32702529e+00 3.15321594e-01 8.34013402e-01 -5.89914799e-01
1.38728380e+00 -7.73631930e-01 -1.23068821e+00 2.07823738e-02
1.21106887e+00 -9.36449528e-01 6.00714147e-01 2.23857179e-01
-1.05192912e+00 -5.18904805e-01 -1.07472706e+00 -2.37184614e-01
-8.00566792e-01 3.09916049e-01 1.19116473e+00 7.17406929e-01
-7.73272991e-01 4.65564609e-01 -5.13348103e-01 -8.18236470e-01
3.15541118e-01 3.30448151e-01 -7.12589145e-01 4.81933579e-02
-1.57519901e+00 1.48515451e+00 9.41751122e-01 1.26178339e-01
3.21782202e-01 -6.52473986e-01 -8.41887891e-01 -3.13943118e-01
5.51898241e-01 -1.70068353e-01 1.14080596e+00 -9.08590481e-02
-1.04556155e+00 1.49116528e+00 -1.97127491e-01 -6.95094824e-01
2.08927184e-01 -5.90109706e-01 -1.10544705e+00 -1.83807552e-01
4.97711539e-01 4.42937940e-01 -5.17282128e-01 -8.00090909e-01
-1.09653616e+00 -1.95597857e-01 1.69046875e-02 2.13782057e-01
3.77827697e-02 5.74986577e-01 -1.37918875e-01 -6.86448514e-01
1.45279784e-02 -7.81315565e-01 1.04716562e-01 -7.80823469e-01
-4.02766585e-01 -1.08124352e+00 5.05606830e-01 -8.15177560e-01
1.87958717e+00 -1.89732337e+00 -2.85700113e-01 2.24764094e-01
1.79835811e-01 5.82027614e-01 2.85001040e-01 4.39887375e-01
-5.41498780e-01 6.73712432e-01 5.57086021e-02 -1.38115346e-01
-2.19310690e-02 2.88887415e-02 2.67539442e-01 3.66544396e-01
-2.16751724e-01 1.01237869e+00 -9.50196028e-01 -6.69600487e-01
-1.03035741e-01 1.61374569e-01 -9.95198786e-02 -9.24192593e-02
4.65177625e-01 -2.00634420e-01 -3.83931994e-01 6.74949408e-01
5.58714926e-01 1.45687312e-02 5.27151644e-01 -3.49840581e-01
-3.88033450e-01 6.33990288e-01 -9.85193372e-01 1.56837761e+00
5.33516631e-02 5.62813163e-01 -3.35595787e-01 -8.33197236e-01
1.03348124e+00 6.77974224e-01 5.07100821e-01 -3.53756160e-01
2.90281981e-01 6.92002416e-01 3.31827015e-01 -6.28397822e-01
7.39319921e-01 -2.74352897e-02 -3.81698400e-01 4.24630828e-02
2.02807225e-02 3.81139964e-02 6.14927351e-01 2.84616500e-01
7.92238355e-01 3.76094788e-01 7.37899244e-01 -6.76113009e-01
7.67172813e-01 2.05892935e-01 6.62849188e-01 4.95956540e-01
-3.18943411e-01 1.20699383e-01 3.32367808e-01 -5.38399637e-01
-1.11448681e+00 -5.18702924e-01 -6.24684691e-01 8.28950882e-01
-1.93085417e-01 -7.98614085e-01 -6.84556365e-01 -1.00434756e+00
3.95325460e-02 1.12402999e+00 -9.70647156e-01 -9.52856690e-02
-6.37377381e-01 -6.84992909e-01 8.36680710e-01 4.40278083e-01
6.16320670e-01 -1.21135211e+00 -4.05608982e-01 2.72251815e-01
-5.71145773e-01 -1.31197166e+00 -1.62995681e-01 9.05995816e-02
-3.40903878e-01 -1.20454836e+00 -1.37590721e-01 -1.01725745e+00
5.28418496e-02 -3.75556469e-01 1.30559123e+00 2.48679981e-01
9.51709375e-02 -2.95884252e-01 -8.49018216e-01 -4.88879889e-01
-3.22798401e-01 7.46934235e-01 -4.74240072e-02 -7.02084720e-01
1.25630736e+00 -1.82472229e-01 -1.37559157e-02 -2.52426267e-01
-3.81986469e-01 -2.55231649e-01 4.28275377e-01 7.96884716e-01
2.35709772e-01 1.57104433e-01 9.40467000e-01 -1.57394886e+00
6.98162615e-01 -3.78271133e-01 -1.09269150e-01 5.83712459e-01
-1.26187479e+00 -1.81680888e-01 1.59022808e-01 -4.25366908e-02
-1.29112673e+00 -3.03841680e-01 -4.68159795e-01 4.17979836e-01
-2.53822088e-01 7.99942851e-01 -1.91047847e-01 1.61419138e-01
7.06670344e-01 -5.11289597e-01 -3.04028153e-01 -4.50670749e-01
8.49889517e-02 8.92612576e-01 7.31182158e-01 -5.43901205e-01
2.02706203e-01 -2.62206882e-01 -1.02201603e-01 -7.60055244e-01
-1.10554016e+00 -4.68805850e-01 -1.02694142e+00 -4.04692553e-02
8.97850037e-01 -6.85539186e-01 -8.48778188e-01 2.62299061e-01
-1.08840239e+00 1.61019593e-01 -3.21385473e-01 7.91139781e-01
-1.49317294e-01 1.65432066e-01 -7.94203460e-01 -6.27749741e-01
-4.42639410e-01 -4.77397263e-01 4.52938050e-01 1.88137621e-01
-8.69729996e-01 -1.09597337e+00 1.83512002e-01 5.63466132e-01
9.39976647e-02 3.11026603e-01 7.99364507e-01 -1.30911756e+00
6.12972021e-01 -4.68670845e-01 -2.42453560e-01 -1.02095596e-01
-3.00058667e-02 -1.51959490e-02 -8.83893788e-01 -1.60714891e-02
-3.63217771e-01 -1.93459347e-01 4.23337489e-01 2.59627044e-01
5.54779768e-01 -2.57501364e-01 -7.01664627e-01 -1.38546443e-02
1.44472539e+00 4.66232300e-01 8.72513115e-01 7.05951750e-01
6.15580142e-01 8.05290163e-01 9.35847282e-01 -1.06733665e-01
6.87734962e-01 6.56251669e-01 -3.24861974e-01 2.77702510e-01
-1.68703482e-01 -3.72689873e-01 -3.37324291e-01 6.50170445e-01
-5.46073258e-01 -2.33983070e-01 -9.78980958e-01 6.88749373e-01
-1.79085720e+00 -1.01597416e+00 -7.09925592e-01 1.85488474e+00
1.00602973e+00 2.87056267e-01 1.08805589e-01 2.82747388e-01
7.80167758e-01 -9.18535888e-02 2.22245172e-01 -6.97096467e-01
-4.56555188e-01 4.64097619e-01 5.96144795e-01 7.05354035e-01
-1.23651695e+00 1.45608532e+00 7.02314281e+00 7.92506158e-01
-4.18762624e-01 1.23180240e-01 4.96587187e-01 4.79227424e-01
-1.18988395e-01 1.16829127e-01 -1.03695297e+00 2.29682803e-01
1.46367574e+00 -5.39961517e-01 -3.71423870e-04 5.20668566e-01
-4.23457921e-01 -1.69255197e-01 -8.12944055e-01 6.48125470e-01
2.01290697e-01 -1.16629970e+00 -2.84456819e-01 3.83217484e-02
4.89235252e-01 -2.35705346e-01 -4.36786115e-01 7.02254534e-01
2.53560275e-01 -1.06178522e+00 4.22544152e-01 3.92952889e-01
8.55897605e-01 -1.04200768e+00 1.43464518e+00 5.88341430e-02
-1.14330983e+00 2.31594652e-01 -2.60040343e-01 -4.57790717e-02
7.46393204e-02 3.97717506e-01 -5.91734767e-01 8.09725404e-01
7.31848955e-01 6.08602703e-01 -7.13229001e-01 6.58675849e-01
-2.66833276e-01 5.74513912e-01 -1.45450592e-01 -3.23798984e-01
-1.97645456e-01 2.80407727e-01 5.17744720e-01 1.45445240e+00
-1.96035221e-01 6.91296637e-01 -1.24572724e-01 2.10351050e-01
-7.61139914e-02 4.43034649e-01 -6.20794713e-01 1.14771025e-02
9.16802227e-01 1.01421821e+00 -8.54386270e-01 -4.60339874e-01
-4.44478512e-01 4.87616152e-01 6.35718644e-01 -2.54282244e-02
-4.63817120e-01 -7.80281305e-01 6.27294034e-02 -2.60668367e-01
1.54378921e-01 1.70975938e-01 -5.56925356e-01 -9.47611213e-01
-1.12504646e-01 -5.94386101e-01 7.08627105e-01 -4.44296390e-01
-1.34041405e+00 8.79892707e-01 5.43350101e-01 -7.56219625e-01
-1.90912619e-01 -7.37448752e-01 6.22603931e-02 1.02657270e+00
-8.33492517e-01 -1.38006806e+00 2.88668424e-01 -4.79471125e-03
2.18523249e-01 -3.33495677e-01 1.20378578e+00 7.91676104e-01
-3.84224921e-01 1.07275522e+00 -2.83334196e-01 2.75042653e-01
8.39042723e-01 -1.31593275e+00 3.59775573e-01 7.50602067e-01
-1.21168628e-01 6.73305571e-01 8.82578552e-01 -9.53158438e-01
-6.96509480e-01 -5.44880271e-01 2.29340410e+00 -6.25444889e-01
6.09435976e-01 8.38540941e-02 -8.90233099e-01 9.75952864e-01
6.26264155e-01 -3.47129792e-01 8.53233099e-01 4.81009066e-01
-1.42468005e-01 2.83543795e-01 -1.55423594e+00 3.59904289e-01
1.13730681e+00 -2.50271529e-01 -1.12217021e+00 2.76147574e-01
5.95628440e-01 -6.35081291e-01 -1.54838133e+00 7.90314496e-01
8.04143131e-01 -6.48831487e-01 6.27168000e-01 -8.95664334e-01
3.04588556e-01 5.11345565e-02 -2.51209438e-01 -8.88931334e-01
-4.56784427e-01 -3.90959769e-01 4.62483801e-02 1.42366242e+00
9.96848404e-01 -6.74030364e-01 7.36133814e-01 8.89293075e-01
-2.81578928e-01 -9.10186887e-01 -9.19044018e-01 -4.99490887e-01
3.11483532e-01 -4.10536915e-01 6.43181324e-01 1.47704828e+00
7.13834941e-01 7.09262490e-01 8.27142745e-02 -1.44684657e-01
4.20140952e-01 -2.64551580e-01 2.98742384e-01 -1.34859776e+00
3.03270310e-01 -6.25736356e-01 -3.81891966e-01 -5.24857879e-01
5.51187575e-01 -1.02306831e+00 -1.70912266e-01 -1.44534445e+00
1.84204981e-01 -5.98071337e-01 -1.50570020e-01 6.87862277e-01
-3.57710391e-01 2.16028720e-01 1.32557480e-02 2.47788519e-01
-4.64968473e-01 7.88810179e-02 8.85640264e-01 3.44738126e-01
-4.16484550e-02 1.37909008e-02 -1.17937362e+00 4.05775219e-01
6.50833428e-01 -5.12631774e-01 -1.09962456e-01 -5.70492335e-02
3.69306862e-01 -2.12621078e-01 -1.94202095e-01 -5.98282814e-01
1.69522211e-01 -8.53286386e-02 2.14022368e-01 -7.24836588e-01
-1.14319615e-01 -6.49522960e-01 3.15050095e-01 5.11262417e-01
-6.68515265e-01 3.32531184e-01 4.36633945e-01 -4.49352711e-02
-2.31957376e-01 -5.06332159e-01 3.70205581e-01 1.31023034e-01
-8.50503623e-01 -1.70895219e-01 -4.05952305e-01 2.99758971e-01
1.06140029e+00 -1.47443652e-01 -7.09473789e-01 7.74730742e-02
-8.33862603e-01 1.70598000e-01 -4.63955477e-02 5.77159584e-01
3.16811442e-01 -1.44728875e+00 -1.19812477e+00 -1.24030359e-01
1.84489936e-01 -3.08541000e-01 1.78296771e-02 8.69203329e-01
-5.15291274e-01 7.80968785e-01 -9.18771774e-02 1.81543872e-01
-1.37909555e+00 2.80117571e-01 -6.30425364e-02 -5.18377125e-01
-7.04471827e-01 1.02753830e+00 -7.64469385e-01 -5.98943710e-01
-2.34723330e-01 -1.11404518e-02 -9.89411950e-01 2.10140839e-01
2.66140729e-01 6.90947533e-01 4.66893196e-01 -1.03885472e+00
-6.15193844e-01 1.75538674e-01 -4.05739248e-01 -2.13963851e-01
1.32536101e+00 -8.74549448e-02 -2.65340745e-01 6.81953371e-01
1.24662566e+00 3.70429724e-01 1.37495622e-01 -1.81737483e-01
7.27779627e-01 -2.27813616e-01 -9.54001248e-02 -1.11193669e+00
-6.86683595e-01 -6.73194975e-02 4.63675022e-01 4.12253886e-01
7.23256290e-01 6.06946088e-02 6.87813997e-01 2.54954487e-01
1.98493719e-01 -1.19456816e+00 -8.98425698e-01 9.44568336e-01
7.60591984e-01 -1.01673388e+00 5.39678931e-01 -8.58417213e-01
-8.03364336e-01 9.56153870e-01 6.18591011e-01 -7.38884583e-02
1.01785862e+00 2.01572701e-01 4.16713022e-02 -3.79824907e-01
-7.63635039e-01 -1.91997126e-01 4.66081053e-01 6.17772698e-01
1.20173895e+00 2.02713475e-01 -1.55169487e+00 8.55330884e-01
-7.84461200e-01 -1.52866125e-01 2.07050845e-01 1.02963471e+00
1.32020146e-01 -1.22612321e+00 1.53206751e-01 6.10727429e-01
-9.91443336e-01 -2.88315117e-01 -4.55452889e-01 1.30130172e+00
3.15525711e-01 1.00346518e+00 -4.01031673e-02 -5.59651256e-01
7.14335322e-01 2.18438402e-01 4.87321138e-01 -7.38483787e-01
-8.27702940e-01 -2.71247536e-01 1.19866073e+00 -2.87353322e-02
-9.79889154e-01 -1.24372458e+00 -1.27512181e+00 -5.56245446e-01
-6.87314212e-01 8.68722141e-01 4.16660517e-01 1.17857456e+00
2.34444350e-01 5.31788588e-01 9.26639810e-02 -2.66890507e-02
4.22607511e-02 -1.59166193e+00 -7.08127439e-01 4.19267207e-01
-1.79540798e-01 -9.42416430e-01 -2.78265566e-01 2.77924407e-02] | [9.323450088500977, 8.891791343688965] |
8e373a9f-f5e3-41d8-bc3b-e4e96d43cc4f | 3d-graph-embedding-learning-with-a-structure | 1902.05247 | null | http://arxiv.org/abs/1902.05247v1 | http://arxiv.org/pdf/1902.05247v1.pdf | 3D Graph Embedding Learning with a Structure-aware Loss Function for Point Cloud Semantic Instance Segmentation | This paper introduces a novel approach for 3D semantic instance segmentation
on point clouds. A 3D convolutional neural network called submanifold sparse
convolutional network is used to generate semantic predictions and instance
embeddings simultaneously. To obtain discriminative embeddings for each 3D
instance, a structure-aware loss function is proposed which considers both the
structure information and the embedding information. To get more consistent
embeddings for each 3D instance, attention-based k nearest neighbour (KNN) is
proposed to assign different weights for different neighbours. Based on the
attention-based KNN, we add a graph convolutional network after the sparse
convolutional network to get refined embeddings. Our network can be trained
end-to-end. A simple mean-shift algorithm is utilized to cluster refined
embeddings to get final instance predictions. As a result, our framework can
output both the semantic prediction and the instance prediction. Experiments
show that our approach outperforms all state-of-art methods on ScanNet
benchmark and NYUv2 dataset. | ['Ming Yang', 'Chunxiang Wang', 'Zhidong Liang'] | 2019-02-14 | null | null | null | null | ['3d-instance-segmentation-1', '3d-semantic-instance-segmentation'] | ['computer-vision', 'computer-vision'] | [-5.00177816e-02 3.01784933e-01 -1.91398934e-01 -6.68707252e-01
-4.50053573e-01 -5.59396595e-02 3.47303271e-01 3.79706062e-02
-7.95062408e-02 6.05958626e-02 1.53640389e-01 -9.62562039e-02
-2.22692892e-01 -8.86342525e-01 -9.07268465e-01 -3.55215073e-01
-8.45839083e-03 6.67260826e-01 4.52665627e-01 1.40801281e-01
5.20096183e-01 7.38551319e-01 -1.45385170e+00 6.99403929e-04
9.18732285e-01 1.13262856e+00 6.43313348e-01 1.82724893e-01
-5.89489758e-01 3.46434057e-01 -2.44127568e-02 8.76568481e-02
3.89103800e-01 -9.29848254e-02 -8.01940441e-01 4.57714617e-01
3.41065228e-01 -1.32248864e-01 -4.51271862e-01 1.10273337e+00
3.87045652e-01 3.63516897e-01 8.41291487e-01 -1.03719783e+00
-9.13856030e-01 2.87836820e-01 -7.90096819e-01 -3.81130725e-02
2.10505381e-01 8.76282081e-02 1.04694653e+00 -1.16522288e+00
4.29456204e-01 1.38790858e+00 4.40504164e-01 4.23383117e-01
-1.04246366e+00 -5.69117069e-01 5.43033659e-01 2.80119330e-01
-1.24824190e+00 1.13455988e-01 1.23158407e+00 -3.67737979e-01
1.08937991e+00 -5.80170527e-02 8.94259930e-01 5.72619975e-01
-7.91168436e-02 8.69731843e-01 5.79617262e-01 -9.59577337e-02
2.22309306e-01 4.73154569e-03 -4.14938182e-02 9.48998094e-01
1.24492921e-01 -2.75751919e-01 -8.56103301e-02 -3.04772072e-02
9.37783539e-01 4.97731268e-01 -1.48862645e-01 -7.30913281e-01
-1.21975899e+00 8.60585749e-01 1.25824034e+00 8.26754048e-02
-5.77802360e-01 1.91051871e-01 1.73079789e-01 -1.50302559e-01
6.32077694e-01 3.96775454e-01 -6.29402876e-01 4.00023103e-01
-6.68009102e-01 2.55227864e-01 4.76672351e-01 1.03164709e+00
1.36019742e+00 -1.67013437e-01 -1.50019988e-01 9.47275341e-01
6.87084258e-01 2.47379512e-01 3.33522886e-01 -8.07772577e-01
5.56165397e-01 1.26347876e+00 -1.31237924e-01 -1.20370686e+00
-3.93616647e-01 -4.72002536e-01 -7.96855509e-01 1.26150355e-01
-2.48220369e-01 2.87879080e-01 -1.43843901e+00 1.08409512e+00
7.17015505e-01 9.03830469e-01 -4.74605374e-02 1.44038248e+00
8.69465470e-01 8.41512978e-01 -1.50865158e-02 5.00169396e-01
1.09268785e+00 -1.19675255e+00 -6.98512495e-02 -2.85272390e-01
6.02825999e-01 -5.52358568e-01 8.37391913e-01 -2.63352811e-01
-6.96619868e-01 -6.12431049e-01 -9.86995280e-01 -2.49640644e-01
-1.43713757e-01 -7.27085173e-02 4.01323974e-01 -1.33457839e-01
-6.54147387e-01 5.76579213e-01 -1.05344021e+00 -3.04120362e-01
8.78052711e-01 2.80104816e-01 -3.27267557e-01 -4.06074792e-01
-7.86338747e-01 5.40885270e-01 4.63310540e-01 1.80801004e-01
-8.03885698e-01 -7.34946728e-01 -1.22341502e+00 3.64792645e-02
5.92962541e-02 -8.51322353e-01 7.86857009e-01 -6.51941359e-01
-1.29745960e+00 8.73879254e-01 -1.69604674e-01 -3.27477396e-01
4.74879704e-02 2.37146579e-03 3.61858383e-02 2.75529712e-01
3.02037597e-01 8.47288489e-01 8.80834162e-01 -1.32193053e+00
-5.13788283e-01 -5.55163503e-01 -5.67479432e-02 4.94160175e-01
-3.70655805e-02 -5.37307501e-01 -6.72185957e-01 -4.12583023e-01
8.76624763e-01 -9.47898388e-01 -6.68005705e-01 -5.05838878e-02
-6.77626371e-01 -4.80941474e-01 9.09286678e-01 -4.27707553e-01
7.62978554e-01 -2.06677175e+00 4.40315396e-01 4.20604706e-01
4.45558757e-01 1.03218101e-01 -2.98746824e-01 2.90782209e-02
-3.63522060e-02 1.13428213e-01 -5.17378807e-01 -3.83271813e-01
6.37113973e-02 2.91862756e-01 1.47224152e-02 4.22093540e-01
6.32389665e-01 9.98250246e-01 -8.71987641e-01 -4.11287308e-01
6.95734918e-01 6.50964320e-01 -8.38826299e-01 3.72479409e-01
-4.72582966e-01 4.30833399e-01 -9.33678508e-01 5.60674727e-01
1.01396680e+00 -4.71638352e-01 -2.67901778e-01 -3.16329151e-01
1.90393422e-02 2.33117208e-01 -9.49121654e-01 2.25665402e+00
-6.78839624e-01 1.54261023e-01 -3.70769203e-01 -1.40772200e+00
1.38099420e+00 -1.58246294e-01 3.59616339e-01 -4.77925718e-01
3.94896865e-02 2.52064258e-01 -4.02020216e-01 -5.36238313e-01
2.78592587e-01 7.18358010e-02 1.75095093e-03 1.63368404e-01
1.55549556e-01 -3.06015790e-01 -5.80951750e-01 1.11043684e-01
8.81298482e-01 2.59070992e-01 -2.42123663e-01 -2.24185035e-01
6.98038697e-01 2.41605520e-01 6.38147414e-01 1.72725603e-01
7.69348070e-02 8.73248279e-01 3.41529310e-01 -8.33461285e-01
-1.09250319e+00 -8.99356902e-01 -2.36994345e-02 4.88512427e-01
7.76569843e-01 -1.97791979e-01 -5.89774489e-01 -9.82781231e-01
3.01845461e-01 5.01664996e-01 -7.08756566e-01 -2.44074970e-01
-4.31016326e-01 -1.82731926e-01 -2.15589985e-01 5.99323094e-01
5.49125671e-01 -9.54227209e-01 -2.90385008e-01 3.04623961e-01
1.51167348e-01 -1.03420782e+00 -4.88855869e-01 -4.16408516e-02
-1.20288038e+00 -1.02816916e+00 -7.21355498e-01 -1.14164782e+00
1.05385506e+00 4.40778852e-01 7.15268970e-01 2.94395179e-01
-1.59363985e-01 2.50095904e-01 -4.07778889e-01 -7.35574737e-02
2.19484255e-01 3.93661350e-01 -2.59578079e-01 6.81089833e-02
6.68949127e-01 -6.93327308e-01 -9.53697383e-01 1.97858885e-01
-6.79104388e-01 6.13510236e-02 8.26128900e-01 7.04634070e-01
1.12539077e+00 -1.51533887e-01 2.32775673e-01 -7.43666649e-01
1.50236338e-01 -6.51009321e-01 -5.89380741e-01 -7.03452975e-02
-3.61330748e-01 4.70277667e-01 4.84642804e-01 -2.06079245e-01
-6.70820355e-01 4.39008325e-01 -2.29588687e-01 -1.07732272e+00
-3.74224097e-01 5.82734466e-01 -1.69746414e-01 -1.30262226e-01
3.10205311e-01 3.01308274e-01 -2.78417487e-02 -7.59682477e-01
3.97785097e-01 7.51455128e-01 1.47268698e-01 -3.28163266e-01
9.86398041e-01 2.22074211e-01 -3.33918594e-02 -9.33761775e-01
-1.07500184e+00 -5.12872338e-01 -7.99653828e-01 3.07117850e-02
1.24924266e+00 -1.12081838e+00 -2.55969912e-01 2.17030451e-01
-1.26104367e+00 -1.30400419e-01 -1.14238888e-01 6.36214912e-01
-5.26323259e-01 8.51569250e-02 -2.54963666e-01 -5.18545449e-01
-3.53147626e-01 -1.36957431e+00 1.38179839e+00 4.00017649e-01
2.16274157e-01 -8.78337383e-01 -4.85567153e-02 3.45809221e-01
7.95212016e-02 3.90867203e-01 8.47796559e-01 -6.31788969e-01
-1.05081904e+00 -3.18890721e-01 -6.27521753e-01 2.92071044e-01
-2.14355215e-02 -2.31473714e-01 -6.87651157e-01 -3.69235687e-03
-2.29802236e-01 -4.53991443e-02 1.18090510e+00 4.05782461e-01
1.64228225e+00 -2.15099350e-01 -5.88068902e-01 8.92925322e-01
1.61921573e+00 -2.30689675e-01 5.91561437e-01 2.69044340e-01
1.32284176e+00 3.77809495e-01 6.56026781e-01 2.76640862e-01
7.82357335e-01 3.21047217e-01 8.15078437e-01 4.60669957e-02
3.94992456e-02 -6.22629523e-01 -2.16540694e-01 7.77106285e-01
-7.76899094e-03 9.32990760e-02 -8.99327338e-01 7.48034596e-01
-1.86466432e+00 -5.91667771e-01 -5.60879596e-02 1.90151536e+00
1.13611996e-01 2.02594876e-01 -1.89559191e-01 -7.85565078e-02
8.39060128e-01 4.15183812e-01 -7.16564476e-01 -2.59279966e-01
4.54673767e-01 3.26822162e-01 3.81835341e-01 4.50656116e-01
-9.83005822e-01 1.17349851e+00 4.27701187e+00 6.92458510e-01
-1.18397439e+00 -7.87316530e-05 4.79395807e-01 1.86210945e-01
-7.01455414e-01 4.39736880e-02 -6.13019884e-01 5.04750252e-01
2.92726487e-01 1.47123247e-01 2.34109297e-01 1.14230669e+00
8.09681490e-02 3.06516677e-01 -6.80265903e-01 1.06007636e+00
3.33596244e-02 -1.51070929e+00 4.15301740e-01 3.91092151e-02
7.53329217e-01 2.63974160e-01 -1.27674356e-01 1.69176221e-01
-2.28556804e-03 -9.59080279e-01 5.54258287e-01 6.52454436e-01
5.22894740e-01 -1.11333799e+00 6.02969706e-01 3.58985156e-01
-1.48945773e+00 1.96544267e-02 -9.52852666e-01 1.24001637e-01
4.18308936e-02 7.46335387e-01 -7.05966532e-01 6.90910935e-01
6.88858032e-01 1.28530896e+00 -2.76921034e-01 1.17304063e+00
-4.79258627e-01 2.85946876e-01 -2.71328241e-01 -1.45048529e-01
6.56884074e-01 -3.89833957e-01 6.38106465e-01 8.24530125e-01
4.55542028e-01 3.11817765e-01 4.79940444e-01 1.09193218e+00
-1.09516770e-01 6.83613569e-02 -6.08658254e-01 1.31694868e-01
5.86763859e-01 1.33824039e+00 -7.51205504e-01 -1.71108678e-01
-3.99730146e-01 1.34313202e+00 6.42852187e-01 2.21992061e-01
-5.63155234e-01 -6.01646841e-01 1.00164950e+00 1.73363641e-01
7.76053905e-01 -3.63125622e-01 -2.23909184e-01 -1.11255085e+00
-6.33275509e-02 -6.14907928e-02 1.55924797e-01 -8.06833684e-01
-1.40735924e+00 5.41023314e-01 -3.00004989e-01 -1.46989882e+00
2.89612442e-01 -4.81882244e-01 -8.61931324e-01 1.07789052e+00
-1.91967762e+00 -1.29325068e+00 -5.58290303e-01 3.28965157e-01
5.49911559e-01 1.06631599e-01 5.29038966e-01 1.42447287e-02
-5.81570148e-01 2.34996408e-01 -4.26162124e-01 1.79527476e-01
1.73176870e-01 -1.24628747e+00 6.33635402e-01 4.16356921e-01
3.00157070e-02 3.52274239e-01 1.82161465e-01 -6.80888951e-01
-1.37948954e+00 -1.77238059e+00 7.35576093e-01 -2.75095642e-01
3.78320992e-01 -2.57521570e-01 -1.17545640e+00 6.66180849e-01
-2.75568366e-01 4.28573132e-01 4.47169602e-01 -7.06251040e-02
-3.14869314e-01 -7.78567269e-02 -1.00681043e+00 3.54718536e-01
1.38840902e+00 -3.84187967e-01 -7.55392551e-01 3.53440702e-01
1.36058819e+00 -6.38881266e-01 -9.37889278e-01 4.69662428e-01
1.01994112e-01 -7.09733307e-01 1.09223568e+00 -5.92728972e-01
7.47143269e-01 -5.97205818e-01 -1.71393096e-01 -1.35865414e+00
-6.05008483e-01 1.41967805e-02 1.72043461e-02 8.58138561e-01
6.33172035e-01 -5.01378179e-01 9.46609199e-01 3.73074949e-01
-5.96878350e-01 -1.17048931e+00 -7.17486858e-01 -5.19377887e-01
-3.86761278e-02 -5.60850561e-01 1.06874120e+00 9.57162976e-01
-4.85569775e-01 3.69792640e-01 -5.56144342e-02 7.53094733e-01
8.88026893e-01 5.42502344e-01 7.83892274e-01 -1.39073062e+00
2.17498511e-01 -2.06676468e-01 -1.04225326e+00 -1.45054674e+00
4.88943756e-01 -1.37303209e+00 -9.97447595e-02 -1.94577634e+00
-3.72662507e-02 -7.69134045e-01 -5.15956938e-01 3.91418993e-01
-2.92869151e-01 2.63553470e-01 6.78760633e-02 6.17328361e-02
-6.77732050e-01 9.66897607e-01 1.55486500e+00 -1.92931160e-01
-3.85216385e-01 -1.21326387e-01 -5.49613178e-01 5.40309846e-01
8.00743878e-01 -5.13835311e-01 -3.45861971e-01 -7.27753162e-01
-1.48345202e-01 -1.38538122e-01 5.76888204e-01 -1.09146559e+00
2.51925647e-01 -4.83346023e-02 6.63674951e-01 -9.53802049e-01
2.66635269e-01 -1.02997959e+00 -1.03417374e-01 3.09255898e-01
-1.91028282e-01 -4.69083697e-01 1.27478670e-02 8.36820900e-01
-1.88763246e-01 -1.22294258e-02 7.63501167e-01 -2.44302318e-01
-8.56971979e-01 1.17842877e+00 4.40680742e-01 -5.99729083e-02
1.12612092e+00 -3.03267241e-01 1.54739454e-01 2.03277275e-01
-7.54655004e-01 6.73562348e-01 6.70897961e-01 5.84228635e-01
1.05944407e+00 -1.66171658e+00 -5.96305430e-01 4.06598777e-01
3.65401447e-01 7.67580926e-01 5.40879011e-01 5.86120009e-01
-6.42216682e-01 3.08555588e-02 -7.33618066e-02 -1.09139371e+00
-6.92923784e-01 5.14066815e-01 3.91216785e-01 1.65494740e-01
-9.17848885e-01 1.05839610e+00 8.55903924e-02 -9.04879808e-01
-8.63453560e-03 -4.15339768e-01 -1.91544667e-01 -3.14367294e-01
3.51286083e-01 -1.52915567e-02 -1.96842507e-01 -5.62186599e-01
-5.81358552e-01 1.03393924e+00 -3.94264199e-02 3.68377984e-01
1.73299730e+00 -1.44197186e-02 -9.63394120e-02 2.21781716e-01
1.59158480e+00 -5.71603298e-01 -1.38141727e+00 -3.21960717e-01
-2.25525811e-01 -6.78918421e-01 3.07672530e-01 -1.53533414e-01
-1.39688313e+00 1.04872012e+00 4.33379024e-01 -1.42999664e-02
9.05357659e-01 3.90647382e-01 1.06804252e+00 1.94631785e-01
1.05973318e-01 -7.48282909e-01 5.29532135e-02 5.92840314e-01
8.83721650e-01 -1.27335608e+00 -1.19905561e-01 -3.77598912e-01
-5.52587032e-01 1.07152975e+00 7.96139240e-01 -8.70168209e-01
9.18728352e-01 -4.56477731e-01 -1.65797830e-01 -5.22721291e-01
-3.46165419e-01 -3.17709088e-01 5.12832463e-01 5.86332619e-01
2.09745113e-02 4.65966836e-02 1.04636364e-02 6.12486720e-01
-7.03328997e-02 -2.79496610e-01 -1.03859755e-03 5.97834945e-01
-8.08695912e-01 -7.68874705e-01 -1.05460554e-01 7.44314551e-01
2.27020279e-01 1.56495899e-01 -1.73108116e-01 4.97381181e-01
1.19184755e-01 4.57049310e-01 4.38397229e-01 -7.47201800e-01
6.00719929e-01 -1.92203224e-02 3.47483873e-01 -9.37717140e-01
-2.84140892e-02 -2.30655357e-01 -4.19737637e-01 -8.68334651e-01
-4.46943372e-01 -3.95847231e-01 -1.70415950e+00 -1.58135951e-01
-2.77423590e-01 1.23742163e-01 7.57087290e-01 9.30233002e-01
6.81098998e-01 5.47839701e-01 1.02435100e+00 -1.32961988e+00
-1.05506428e-01 -8.55812848e-01 -5.75916529e-01 3.29974353e-01
1.85530901e-01 -7.41847157e-01 -4.43735629e-01 -3.31289500e-01] | [7.978462219238281, -3.208561897277832] |
6a11376a-fbcd-407f-9804-34e9ce2ec92d | learning-to-detect-adversarial-examples-based | 2107.04435 | null | https://arxiv.org/abs/2107.04435v1 | https://arxiv.org/pdf/2107.04435v1.pdf | Learning to Detect Adversarial Examples Based on Class Scores | Given the increasing threat of adversarial attacks on deep neural networks (DNNs), research on efficient detection methods is more important than ever. In this work, we take a closer look at adversarial attack detection based on the class scores of an already trained classification model. We propose to train a support vector machine (SVM) on the class scores to detect adversarial examples. Our method is able to detect adversarial examples generated by various attacks, and can be easily adopted to a plethora of deep classification models. We show that our approach yields an improved detection rate compared to an existing method, whilst being easy to implement. We perform an extensive empirical analysis on different deep classification models, investigating various state-of-the-art adversarial attacks. Moreover, we observe that our proposed method is better at detecting a combination of adversarial attacks. This work indicates the potential of detecting various adversarial attacks simply by using the class scores of an already trained classification model. | ['Oliver De Candido', 'Felix Michels', 'Tobias Uelwer'] | 2021-07-09 | null | null | null | null | ['adversarial-attack-detection', 'adversarial-attack-detection'] | ['computer-vision', 'knowledge-base'] | [ 4.43537891e-01 -2.75127012e-02 3.40569079e-01 -1.05478592e-01
-5.50511062e-01 -1.14573216e+00 9.74375010e-01 2.15784147e-01
-5.61392128e-01 5.72800756e-01 -3.21085632e-01 -6.01609707e-01
1.07257627e-01 -9.79106545e-01 -8.11880231e-01 -7.32404292e-01
-2.76600510e-01 1.60890639e-01 6.00633979e-01 -4.03564990e-01
1.55681536e-01 1.02031612e+00 -1.30285192e+00 2.93513387e-01
3.56954485e-01 9.56840575e-01 -3.76127809e-01 9.62165892e-01
3.39297950e-01 7.71044612e-01 -1.25423801e+00 -9.59740758e-01
5.43711126e-01 -2.73470640e-01 -6.63138986e-01 -3.87852043e-01
8.36014628e-01 -4.44025666e-01 -6.49238706e-01 1.61712360e+00
5.52592099e-01 -9.95297953e-02 5.78238368e-01 -1.42528689e+00
-3.38995486e-01 6.44781172e-01 7.16163367e-02 5.81416607e-01
1.79288030e-01 2.16770232e-01 7.21806169e-01 -4.56267059e-01
3.57510775e-01 1.18122935e+00 6.42489552e-01 8.09370399e-01
-9.99439895e-01 -1.10058224e+00 -4.81417961e-02 3.26725364e-01
-9.46139634e-01 -3.48251075e-01 8.74432862e-01 -4.18232679e-01
6.51935577e-01 4.24409360e-01 2.14061156e-01 1.54110265e+00
3.18994522e-01 5.61450899e-01 1.04123223e+00 -4.77144361e-01
4.80393946e-01 2.90145665e-01 -8.62455592e-02 6.45475626e-01
2.26321787e-01 6.01353109e-01 5.91271818e-02 -5.01295805e-01
5.38150370e-01 5.02085127e-02 -1.12111412e-01 -2.59426892e-01
-5.17488718e-01 1.14755809e+00 5.89769900e-01 3.87228191e-01
-3.48391868e-02 2.00739682e-01 8.07851315e-01 6.17072046e-01
4.31297958e-01 6.98441803e-01 -5.36438525e-01 1.99011087e-01
-5.77655196e-01 2.90978700e-01 9.15759206e-01 5.45213044e-01
3.52393389e-01 3.46178889e-01 2.29568675e-01 6.86346889e-01
1.65432468e-02 2.75909156e-01 3.63120914e-01 -5.50022244e-01
4.76642787e-01 2.47555524e-01 -4.25319970e-01 -9.57929850e-01
-1.63149953e-01 -4.20748591e-01 -8.62460852e-01 9.67583418e-01
5.69992542e-01 -3.82927597e-01 -8.11783850e-01 1.50941491e+00
9.73819271e-02 3.93856645e-01 3.16144854e-01 4.01699334e-01
4.84834224e-01 3.60582799e-01 1.60961002e-01 1.19277537e-01
1.00958610e+00 -5.96577585e-01 -3.31691951e-01 -1.31506234e-01
7.13835418e-01 -7.69925475e-01 3.91472846e-01 6.35894060e-01
-7.35358596e-01 -3.80236477e-01 -1.29151392e+00 5.72825432e-01
-8.40346456e-01 -3.04291427e-01 4.85333592e-01 1.52092743e+00
-7.69110560e-01 9.00216937e-01 -7.39441991e-01 2.24264283e-02
6.84271574e-01 4.93534654e-01 -5.25333464e-01 -6.42054081e-02
-1.51005220e+00 1.07483602e+00 4.27847445e-01 -2.47721449e-01
-1.18010962e+00 -4.58452553e-01 -8.07569623e-01 -1.20684035e-01
2.58459337e-02 -5.64453052e-03 1.13113070e+00 -1.14790726e+00
-1.29093313e+00 8.41204345e-01 5.94088435e-01 -8.86936069e-01
7.74679780e-01 -2.08617896e-01 -6.86597288e-01 3.22481275e-01
-4.63009536e-01 3.49941850e-02 9.83617067e-01 -1.19254434e+00
-4.02666807e-01 -3.24586630e-01 3.58904004e-01 -3.71137172e-01
-8.46530378e-01 5.62126279e-01 3.47717702e-01 -9.83355105e-01
-3.59261930e-01 -9.42955136e-01 -3.98660660e-01 7.59290233e-02
-4.66085881e-01 -5.02523519e-02 9.81703043e-01 -3.65243375e-01
1.06946933e+00 -2.16002488e+00 -1.53889254e-01 3.91536355e-01
3.15523893e-01 1.13529050e+00 -1.25547007e-01 4.80819851e-01
-4.20096010e-01 1.58149615e-01 -3.75567079e-01 -1.86158419e-01
1.03694133e-01 3.49210441e-01 -7.52769291e-01 6.76765382e-01
4.65617418e-01 6.57877445e-01 -8.78377497e-01 -5.41140810e-02
3.57333153e-01 3.54134440e-01 -4.93645340e-01 2.46582583e-01
3.77951488e-02 -2.72963359e-03 -3.36755753e-01 5.62034547e-01
7.80658841e-01 4.98419613e-01 1.43957520e-02 1.38069957e-01
4.62285906e-01 6.79355264e-02 -1.14095509e+00 7.10004389e-01
-2.21963704e-01 8.87184203e-01 -1.45065546e-01 -1.38908505e+00
1.08021176e+00 4.69485760e-01 -5.01761809e-02 -3.03750098e-01
4.73864973e-01 3.78719002e-01 4.06569451e-01 -2.09421173e-01
-2.69721523e-02 -1.38142139e-01 -1.92610949e-01 1.86809868e-01
4.44584191e-01 2.07008630e-01 -6.34098127e-02 1.21745788e-01
1.70707262e+00 -5.56773007e-01 4.01927054e-01 -5.37978671e-02
8.26424003e-01 -4.00016367e-01 5.17698824e-02 1.13250625e+00
-4.35258418e-01 2.37353861e-01 5.73522270e-01 -7.61447668e-01
-9.33557928e-01 -1.23226976e+00 -2.56102592e-01 1.08114386e+00
-2.25803733e-01 -2.87762910e-01 -8.94406617e-01 -1.31489670e+00
2.53687445e-02 4.46095854e-01 -8.06705356e-01 -5.38382530e-01
-6.55722678e-01 -7.21643090e-01 1.34325075e+00 7.01995373e-01
3.63845944e-01 -1.26062906e+00 -4.33312476e-01 1.36567101e-01
6.45085573e-01 -1.15323186e+00 1.03815012e-01 4.20207053e-01
-7.71509051e-01 -1.06685317e+00 -4.69522417e-01 -8.34720552e-01
6.08743429e-01 -3.52691859e-01 9.37391698e-01 1.21925354e-01
-4.19473708e-01 1.67710021e-01 -5.13235807e-01 -7.85187244e-01
-1.12133133e+00 -6.29169196e-02 3.61621499e-01 8.76277511e-04
3.11677814e-01 -6.68661654e-01 -9.21994597e-02 3.08658838e-01
-1.24657691e+00 -8.97883117e-01 4.04790401e-01 6.36139870e-01
-2.52199266e-02 4.13458675e-01 6.62177324e-01 -1.14934361e+00
6.17199659e-01 -5.68137467e-01 -7.39336371e-01 -4.98384051e-02
-2.89616853e-01 -2.52947938e-02 1.29472041e+00 -9.03029442e-01
-4.10560578e-01 1.53527945e-01 -8.62889588e-01 -6.84672952e-01
-6.46129310e-01 3.65853980e-02 -3.09754282e-01 -7.49269783e-01
9.80729401e-01 6.36778027e-02 -2.83314526e-01 -4.36110735e-01
2.00037763e-01 4.70373154e-01 5.76658726e-01 -2.79692620e-01
1.36853588e+00 1.31447509e-01 2.66961306e-01 -7.26040781e-01
-6.10562384e-01 -1.10486209e-01 -6.47892892e-01 -2.23760962e-01
3.51780415e-01 -4.77830827e-01 -4.82083052e-01 9.23814952e-01
-1.10448253e+00 -1.30791888e-01 8.72066021e-02 1.84447661e-01
-1.86880648e-01 4.62669015e-01 -7.14341938e-01 -8.45647335e-01
-2.08012670e-01 -1.13747740e+00 5.79615533e-01 -1.18668824e-01
-2.91545913e-02 -1.35045350e+00 1.81855768e-01 7.35390037e-02
4.95152622e-01 7.25673079e-01 6.61352634e-01 -1.53318310e+00
-1.38464078e-01 -9.23697054e-01 6.25840053e-02 8.49799514e-01
3.48657183e-02 1.36004731e-01 -1.30170870e+00 -4.43704069e-01
1.02755852e-01 -4.62659538e-01 9.21807885e-01 -3.09666276e-01
1.28143179e+00 -6.71609581e-01 -6.85188919e-02 5.56571901e-01
1.51309991e+00 2.94306606e-01 8.26656878e-01 6.93267584e-01
7.63492823e-01 4.00871485e-01 1.61322504e-01 2.54924357e-01
-4.80798572e-01 5.53720415e-01 1.03132701e+00 -5.47530055e-02
2.54741400e-01 1.18171453e-01 5.14585376e-01 8.97770524e-02
8.38914216e-02 -2.81696856e-01 -9.24056470e-01 2.16834754e-01
-1.29092813e+00 -1.14748228e+00 -2.59252684e-03 1.98652315e+00
7.49951780e-01 7.16541171e-01 1.47126168e-01 8.17959607e-01
6.45855069e-01 1.33128956e-01 -3.75956774e-01 -8.70802641e-01
-5.16626425e-02 8.90563786e-01 5.51082134e-01 2.91514575e-01
-1.60673940e+00 8.60504627e-01 6.79393005e+00 8.54011953e-01
-1.14472806e+00 -7.66644776e-02 3.99599761e-01 3.98166329e-02
2.73424983e-01 -5.70320964e-01 -6.85319543e-01 4.47629809e-01
1.01690376e+00 -1.28479619e-02 1.74135253e-01 1.13203943e+00
-5.57875335e-01 5.17597675e-01 -1.15320492e+00 5.29331982e-01
2.16863230e-01 -1.28197074e+00 1.61662638e-01 7.32995868e-02
5.94438314e-01 5.33002056e-02 2.04671338e-01 3.58737171e-01
6.66409433e-01 -1.04927754e+00 4.78317499e-01 -9.78007764e-02
3.56644213e-01 -1.19138455e+00 1.14941597e+00 4.29058760e-01
-9.30069387e-01 -2.93666512e-01 -3.54725838e-01 -3.94420363e-02
-4.13678110e-01 1.98151931e-01 -8.80702555e-01 3.78185689e-01
5.94934583e-01 3.95352960e-01 -7.55947948e-01 9.13429618e-01
-3.00452828e-01 9.21673417e-01 -2.09736034e-01 -1.49167836e-01
4.33171391e-01 3.68656248e-01 5.51209152e-01 1.39730394e+00
-8.92381743e-02 -2.15360820e-01 -1.29640149e-02 3.05012077e-01
-2.87871957e-01 -5.83043657e-02 -9.90028560e-01 -1.81608833e-02
2.97556460e-01 1.15222907e+00 -5.84667087e-01 -2.88941771e-01
-2.68984586e-01 1.06775820e+00 3.55473369e-01 -6.29355535e-02
-7.89654076e-01 -6.86542630e-01 8.73727024e-01 -2.42763147e-01
3.51737320e-01 -9.70693713e-04 -4.22597751e-02 -8.19260478e-01
-1.35881668e-02 -1.20382714e+00 5.20839453e-01 -2.52257526e-01
-1.58456278e+00 9.42368567e-01 -1.33853599e-01 -1.23874390e+00
-1.80423304e-01 -1.21274996e+00 -1.06589401e+00 7.23955691e-01
-1.16557682e+00 -7.44972527e-01 1.23163790e-01 7.40598381e-01
2.93885529e-01 -6.67182803e-01 1.10573924e+00 4.01269287e-01
-5.95816135e-01 1.16582465e+00 1.54647663e-01 8.63998711e-01
5.42928398e-01 -1.31183040e+00 8.16560090e-01 1.20001483e+00
4.17040557e-01 4.07497138e-01 8.93294871e-01 -3.74450594e-01
-9.16046560e-01 -1.21625698e+00 5.22566795e-01 -6.98842764e-01
1.15586340e+00 -5.02613664e-01 -1.06975234e+00 6.76799893e-01
-3.16530727e-02 2.34100536e-01 9.26831663e-01 -2.80514926e-01
-8.01015973e-01 3.88203561e-02 -1.44341731e+00 5.57053566e-01
8.21777642e-01 -5.78829706e-01 -7.87459910e-01 3.64188254e-01
4.85787570e-01 -3.89351994e-01 -7.10150480e-01 4.92398709e-01
4.16548848e-01 -9.79818702e-01 1.26956189e+00 -9.54257905e-01
5.85036159e-01 6.54991567e-02 -2.48323992e-01 -1.36867297e+00
-1.25563532e-01 -5.55636287e-01 -3.79663557e-01 1.00784588e+00
2.22629786e-01 -9.90792036e-01 7.36296952e-01 2.66792625e-01
1.18272714e-01 -7.76246905e-01 -1.12160492e+00 -1.03889692e+00
3.52524310e-01 -4.90311682e-01 3.98951173e-01 1.19373906e+00
-2.85496533e-01 -1.81446254e-01 -3.93186122e-01 7.36623287e-01
7.70795286e-01 -6.17290080e-01 6.77729309e-01 -1.28643310e+00
-4.96782601e-01 -6.16777718e-01 -1.24729908e+00 -2.65171468e-01
2.83491939e-01 -7.91375577e-01 -2.71066636e-01 -6.94500744e-01
-3.85311425e-01 -3.09775978e-01 -7.69887209e-01 6.77347898e-01
-2.36211672e-01 9.26700115e-01 1.68411568e-01 -2.27198124e-01
-2.55372703e-01 7.26029724e-02 7.42263079e-01 -1.34900942e-01
3.80056798e-01 4.40507382e-01 -5.00502110e-01 9.87041950e-01
1.01371765e+00 -9.10329163e-01 -8.89395177e-02 -8.18700269e-02
7.24954382e-02 -3.92772883e-01 6.02305353e-01 -1.37037015e+00
1.07748464e-01 2.18713269e-01 5.78466594e-01 -1.14292480e-01
1.87048569e-01 -8.20405483e-01 -1.95157871e-01 9.05661106e-01
-3.63291502e-01 3.87463719e-02 5.69072247e-01 5.01683652e-01
-1.15698382e-01 -5.56124032e-01 1.11668003e+00 -4.25247401e-02
-4.62992966e-01 3.42346430e-01 -5.22261918e-01 -1.01830550e-01
1.16500282e+00 2.06093863e-02 -2.94344366e-01 -6.33785315e-03
-7.49044001e-01 -2.71061927e-01 2.64905691e-01 5.25990188e-01
6.37154639e-01 -1.23058200e+00 -6.22263730e-01 4.83451098e-01
9.75216180e-02 -6.82531059e-01 -1.68654189e-01 -4.24922928e-02
-7.19771028e-01 -2.86010355e-02 -5.82748175e-01 -2.12012231e-01
-1.81596267e+00 8.70614886e-01 4.48701054e-01 -4.20305103e-01
-3.96637321e-01 1.03703332e+00 -6.52287751e-02 -4.26983654e-01
4.46329683e-01 2.00336024e-01 -3.19884986e-01 -9.49646309e-02
7.78351784e-01 2.94654101e-01 1.85530171e-01 -6.26044571e-01
-4.72521663e-01 1.49496377e-01 -2.89892018e-01 2.30642214e-01
1.23176110e+00 7.47004867e-01 8.71396363e-02 4.00709845e-02
1.48994327e+00 1.03493735e-01 -9.76323307e-01 -1.80018961e-01
-7.95646310e-02 -4.53400612e-01 -1.58704728e-01 -6.93032265e-01
-1.16074574e+00 9.87803936e-01 6.73430383e-01 8.40838313e-01
1.25397718e+00 -1.60951406e-01 6.36804640e-01 5.93684018e-01
2.87839919e-01 -3.96281034e-01 2.28235662e-01 5.09073436e-01
6.00080132e-01 -1.14484811e+00 -1.57247022e-01 -3.09673220e-01
-2.29318768e-01 1.29105794e+00 4.48438972e-01 -7.46208131e-01
6.19425476e-01 4.65344936e-01 2.65364677e-01 9.24475417e-02
-4.64861751e-01 1.51210219e-01 2.96991318e-01 7.60530293e-01
-9.33277011e-02 -1.29563157e-02 9.28504765e-02 2.42815569e-01
-4.64350551e-01 -5.96469820e-01 5.58699608e-01 8.55319738e-01
-4.13460046e-01 -1.46174753e+00 -5.22366583e-01 3.99681062e-01
-1.06881475e+00 -2.06447184e-01 -7.39973068e-01 6.28114879e-01
1.68323830e-01 8.73149872e-01 -2.31432736e-01 -6.47317827e-01
4.32718903e-01 8.39061067e-02 5.13497472e-01 -6.10899091e-01
-9.21896458e-01 -6.10054851e-01 5.15864883e-03 -5.05746245e-01
-7.16916099e-02 -5.24996936e-01 -5.58917880e-01 -2.64638156e-01
-4.03873622e-01 -6.95180595e-02 7.60407567e-01 1.08000362e+00
-2.05281854e-01 5.52709043e-01 1.07897639e+00 -9.27270770e-01
-9.70516264e-01 -6.94290400e-01 -4.98282433e-01 6.28490508e-01
6.72770858e-01 -5.99543273e-01 -9.19007361e-01 -1.62657008e-01] | [5.640776634216309, 7.772590637207031] |
45bd795e-e359-4d91-87c0-170ee6a9fefe | variational-quantum-regression-algorithm-with | 2307.03334 | null | https://arxiv.org/abs/2307.03334v1 | https://arxiv.org/pdf/2307.03334v1.pdf | Variational quantum regression algorithm with encoded data structure | Variational quantum algorithms (VQAs) prevail to solve practical problems such as combinatorial optimization, quantum chemistry simulation, quantum machine learning, and quantum error correction on noisy quantum computers. For variational quantum machine learning, a variational algorithm with model interpretability built into the algorithm is yet to be exploited. In this paper, we construct a quantum regression algorithm and identify the direct relation of variational parameters to learned regression coefficients, while employing a circuit that directly encodes the data in quantum amplitudes reflecting the structure of the classical data table. The algorithm is particularly suitable for well-connected qubits. With compressed encoding and digital-analog gate operation, the run time complexity is logarithmically more advantageous than that for digital 2-local gate native hardware with the number of data entries encoded, a decent improvement in noisy intermediate-scale quantum computers and a minor improvement for large-scale quantum computing Our suggested method of compressed binary encoding offers a remarkable reduction in the number of physical qubits needed when compared to the traditional one-hot-encoding technique with the same input data. The algorithm inherently performs linear regression but can also be used easily for nonlinear regression by building nonlinear features into the training data. In terms of measured cost function which distinguishes a good model from a poor one for model training, it will be effective only when the number of features is much less than the number of records for the encoded data structure to be observable. To echo this finding and mitigate hardware noise in practice, the ensemble model training from the quantum regression model learning with important feature selection from regularization is incorporated and illustrated numerically. | ['Ryan S. Bennink', 'C. -C. Joseph Wang'] | 2023-07-07 | null | null | null | null | ['combinatorial-optimization'] | ['methodology'] | [ 5.56207478e-01 1.65829390e-01 -5.93712479e-02 -2.09847689e-01
-9.31814194e-01 -3.97824049e-01 4.32218015e-01 3.51234138e-01
-7.06174850e-01 9.73607957e-01 -5.45617521e-01 -4.77615446e-01
-1.95508540e-01 -1.15896952e+00 -7.56844997e-01 -1.27115989e+00
4.86897901e-02 5.87530255e-01 -1.51741326e-01 -6.50118589e-01
5.60243547e-01 3.45761061e-01 -1.46780336e+00 -7.43934065e-02
8.15021574e-01 7.52895772e-01 9.48870331e-02 6.72737896e-01
-6.44157967e-03 3.64186078e-01 -1.63826525e-01 -2.79926538e-01
3.59284312e-01 -8.69133115e-01 -5.31917572e-01 -4.01504874e-01
-6.68757409e-02 1.01883478e-01 -8.23681414e-01 1.29610920e+00
4.68436271e-01 2.69104987e-01 9.29092824e-01 -5.43988347e-01
-4.83582616e-01 5.43159962e-01 1.62746727e-01 1.48409322e-01
7.82710463e-02 3.80473077e-01 1.17568517e+00 -3.82307798e-01
6.87105656e-01 7.75872052e-01 2.33333975e-01 6.62800312e-01
-1.79002428e+00 -3.74803245e-01 -8.98863196e-01 4.22883928e-01
-1.68324113e+00 -3.43620628e-01 6.67703807e-01 -1.46690905e-01
1.23774529e+00 3.92614722e-01 6.41424000e-01 6.93798780e-01
4.30242926e-01 -4.87495139e-02 1.22017145e+00 -7.92184889e-01
5.85329950e-01 4.30508316e-01 2.55845904e-01 9.33597445e-01
2.34834731e-01 7.72216678e-01 -5.80918252e-01 -2.59038031e-01
5.19656777e-01 -2.50619769e-01 -1.81741923e-01 -3.58489603e-01
-9.58885550e-01 1.19421041e+00 5.53731859e-01 3.05174321e-01
-2.66942352e-01 4.79417443e-01 3.62202287e-01 5.96613884e-01
-1.10177010e-01 8.59935403e-01 -3.90394598e-01 -1.09578319e-01
-8.02873135e-01 3.29958588e-01 7.77639687e-01 5.48514307e-01
1.27102661e+00 3.68584171e-02 5.15600741e-02 1.55134901e-01
9.42043737e-02 7.09752202e-01 3.97135168e-01 -8.59971046e-01
1.34341463e-01 4.16891545e-01 1.08197434e-02 -3.60931516e-01
-2.90872425e-01 -1.91130713e-01 -9.04077888e-01 1.91808537e-01
3.80883217e-01 1.23741657e-01 -7.99201310e-01 1.48370492e+00
2.45833918e-01 -9.13355574e-02 3.93612802e-01 8.52094531e-01
5.55157423e-01 8.51644456e-01 -3.44893634e-01 -6.13523543e-01
1.20224082e+00 -2.68561631e-01 -6.48739219e-01 2.52143860e-01
1.08477879e+00 -4.13446277e-01 7.13817358e-01 4.49083269e-01
-7.70777583e-01 -3.46919030e-01 -1.39778733e+00 -2.96869665e-01
-5.51843166e-01 -3.76438081e-01 9.79983985e-01 1.11193633e+00
-7.09508836e-01 1.40045774e+00 -7.47544408e-01 3.29669677e-02
-1.46576285e-01 9.48278368e-01 -5.03301203e-01 1.70705229e-01
-1.39503634e+00 9.46848571e-01 7.23953366e-01 2.47650370e-01
-6.17995203e-01 -2.54758984e-01 -6.14083827e-01 3.17427441e-02
4.79739346e-02 -6.62951648e-01 6.82887614e-01 -4.23416346e-01
-1.85619199e+00 6.99286044e-01 -3.82875055e-01 -5.93047202e-01
1.25834405e-01 5.99575460e-01 -2.54716963e-01 3.12556848e-02
-3.44955862e-01 1.69692278e-01 9.60678756e-01 -5.36674440e-01
1.30334049e-01 -5.43459356e-01 -2.84900330e-02 -2.34792158e-01
8.18325803e-02 -4.07153279e-01 2.41496995e-01 2.33875245e-01
5.35574317e-01 -1.05671060e+00 -6.39511168e-01 -4.23008829e-01
-1.60264030e-01 -1.33340687e-01 2.86642730e-01 -8.90360847e-02
9.49331105e-01 -2.02950835e+00 5.46601593e-01 6.36539519e-01
1.45782083e-01 1.01664148e-01 1.95255596e-02 5.34025609e-01
-1.19211547e-01 5.63530773e-02 -3.09719115e-01 1.09000161e-01
-1.29603792e-03 4.03747946e-01 -1.54808283e-01 8.16505671e-01
2.12724656e-01 9.48408365e-01 -7.12746561e-01 -3.18399966e-01
1.91965818e-01 3.76657009e-01 -9.31731105e-01 1.36264774e-03
-1.28313974e-01 7.58990169e-01 -7.46420264e-01 1.98039114e-01
6.13084316e-01 -1.84195623e-01 2.38351300e-01 -2.54579663e-01
-2.41965711e-01 5.36875546e-01 -1.26304066e+00 1.66826010e+00
-3.84657949e-01 4.65746492e-01 -1.48549274e-01 -1.18753946e+00
9.44507062e-01 2.95250267e-01 2.87208557e-01 -1.02576709e+00
2.91445553e-01 6.42088294e-01 3.05486083e-01 -4.71536756e-01
7.74033010e-01 -5.59118450e-01 -1.49246380e-01 2.68503964e-01
2.96021104e-01 -6.13292873e-01 1.58643320e-01 1.34806246e-01
9.58526611e-01 1.06593877e-01 4.34102237e-01 -3.73956800e-01
6.86078131e-01 6.92282990e-02 2.78080314e-01 9.52449739e-01
-1.58431709e-01 3.16690385e-01 6.06924534e-01 -4.15080458e-01
-1.39967203e+00 -6.82945907e-01 -6.84029341e-01 6.27553701e-01
3.85712653e-01 -7.34462500e-01 -7.15752125e-01 8.38800296e-02
-1.34802520e-01 6.67366385e-01 -5.17865121e-01 -5.59466422e-01
-4.53336060e-01 -1.20031190e+00 5.08972943e-01 -2.25519389e-01
4.08976004e-02 -7.76830435e-01 -3.16269934e-01 2.53267109e-01
1.43757761e-01 -8.55151534e-01 3.08450133e-01 1.00253904e+00
-1.17823875e+00 -8.43354642e-01 1.97650924e-01 -1.28478467e-01
5.55496693e-01 -1.93892509e-01 5.61385512e-01 1.14745293e-02
-4.53435808e-01 -2.02624425e-01 -2.28339642e-01 -1.26893464e-02
-8.77502620e-01 -1.01537751e-02 4.05052900e-01 -2.59785324e-01
5.66012919e-01 -5.61816156e-01 -4.41716760e-01 -2.12341234e-01
-8.29146445e-01 -2.32837111e-01 5.51480234e-01 1.46339309e+00
6.73191667e-01 -8.66467506e-02 -2.81238090e-03 -8.36163342e-01
1.51003748e-01 -1.44661352e-01 -8.86316359e-01 1.09384395e-02
-7.96564758e-01 1.01533985e+00 9.11043227e-01 4.45391983e-02
-5.18576443e-01 1.34042904e-01 -2.82680303e-01 1.17555730e-01
1.81588605e-01 4.21089470e-01 1.84693113e-01 -6.12234414e-01
1.15628564e+00 4.62270558e-01 7.19041973e-02 -2.48878092e-01
4.51598793e-01 7.52535224e-01 1.07466884e-01 -6.58740997e-01
9.06545162e-01 2.65056968e-01 8.76839578e-01 -1.10214841e+00
-3.68968755e-01 -1.72723606e-01 -8.92926931e-01 3.02776694e-01
7.93125153e-01 -6.40701592e-01 -1.14918530e+00 -1.87356323e-01
-1.04070246e+00 1.19504750e-01 -4.50369954e-01 7.65807331e-01
-6.93754435e-01 4.91129875e-01 -5.96695364e-01 -1.15684831e+00
-1.75605819e-01 -1.48771715e+00 8.44808221e-01 5.82396276e-02
2.29884103e-01 -7.62050688e-01 8.17880780e-02 3.04582089e-01
1.55911624e-01 4.14665714e-02 1.11474001e+00 -5.13650477e-01
-1.04941821e+00 -4.49213117e-01 8.48591104e-02 2.88913935e-01
-4.80624825e-01 -6.02414608e-02 -1.11241603e+00 -3.51988256e-01
2.39908889e-01 -5.19532502e-01 1.05671310e+00 1.24464996e-01
9.29870009e-01 2.20015511e-01 -1.21185012e-01 7.64682472e-01
1.59939814e+00 2.26999205e-02 7.00433731e-01 -1.21796420e-02
5.81208885e-01 1.78490788e-01 2.70572633e-01 2.46121153e-01
-2.60236502e-01 7.03164935e-01 4.25236166e-01 4.42255020e-01
4.80423570e-01 -1.10959604e-01 4.45116282e-01 1.22636652e+00
-5.05549908e-01 2.67350376e-01 -3.88829529e-01 -9.06294286e-02
-1.47223294e+00 -9.99121785e-01 -5.90315223e-01 2.71629024e+00
7.21815407e-01 2.05713049e-01 -3.79468769e-01 2.96496361e-01
3.43384236e-01 6.45099357e-02 -3.15794826e-01 -9.25944209e-01
-9.18474495e-02 8.57745886e-01 8.64788294e-01 6.66491389e-01
-7.59838641e-01 8.47795010e-01 6.14115143e+00 1.08913803e+00
-1.14749229e+00 3.04246843e-01 1.68917418e-01 2.82598399e-02
-4.09301162e-01 6.22465312e-01 -6.61797404e-01 2.68614590e-01
1.57204795e+00 -3.16028036e-02 1.03613091e+00 5.75959682e-01
2.03250200e-01 -2.20966145e-01 -1.33590925e+00 1.24843001e+00
-4.13675874e-01 -1.38100016e+00 -2.32894979e-02 4.29410279e-01
5.88487744e-01 1.50804311e-01 -2.70146113e-02 3.52789372e-01
-5.53573251e-01 -1.21632218e+00 3.94046575e-01 4.32972163e-01
8.93314600e-01 -6.57115698e-01 7.40609884e-01 4.85842466e-01
-7.71013319e-01 2.65956391e-03 -8.29518497e-01 -4.41906840e-01
-1.30409479e-01 2.50218570e-01 -4.99395877e-01 6.25226438e-01
4.65292521e-02 2.31712878e-01 -3.11735690e-01 5.24691403e-01
1.15326464e-01 5.42548895e-01 -6.20957375e-01 -5.92241347e-01
3.62615019e-01 -1.05125082e+00 5.22666514e-01 8.18032324e-01
3.12627077e-01 3.87039930e-01 -3.83011013e-01 1.12071991e+00
1.77270621e-02 2.99361855e-01 -8.06504428e-01 -4.22193676e-01
1.80230051e-01 9.81858373e-01 -4.44474906e-01 -2.04529345e-01
-2.06009403e-01 9.86046195e-01 2.04246551e-01 1.68682829e-01
-6.39432311e-01 -3.95163834e-01 3.72442365e-01 -2.25314319e-01
2.04597250e-01 -3.79188061e-01 -3.94581437e-01 -1.33528662e+00
-2.91155249e-01 -4.39574569e-01 -2.54156291e-01 -2.02530414e-01
-7.82529771e-01 4.10159081e-01 -4.28707004e-01 -1.03320861e+00
-4.13878530e-01 -9.73711908e-01 -2.75660515e-01 1.12478137e+00
-1.25251722e+00 -4.56135750e-01 3.14165533e-01 4.08717752e-01
-2.90587008e-01 -1.00846954e-01 1.47878230e+00 9.01104286e-02
-6.23236775e-01 5.62085748e-01 1.00314736e+00 -2.13005111e-01
3.23050827e-01 -1.29088748e+00 9.66202468e-02 5.58975279e-01
4.21207458e-01 9.22130406e-01 1.13851476e+00 -2.37400323e-01
-2.11194062e+00 -4.55332428e-01 8.83167922e-01 -5.73312461e-01
7.34325171e-01 -5.78984559e-01 -8.21787715e-01 1.94815457e-01
-1.93456009e-01 1.03943869e-01 6.30903840e-01 1.81058392e-01
-2.03536689e-01 -1.43071249e-01 -1.12780225e+00 1.99271858e-01
6.27411127e-01 -1.12739801e+00 -4.17892367e-01 6.60356581e-01
4.70476776e-01 -4.19051439e-01 -9.94201183e-01 -3.28506082e-02
4.94457692e-01 -9.21597838e-01 7.69693196e-01 -7.12722123e-01
3.09584498e-01 -2.96691447e-01 -3.91593218e-01 -9.50799167e-01
-2.02877000e-01 -9.89107370e-01 -7.09425285e-03 3.24735761e-01
5.77268541e-01 -7.70508409e-01 7.74187684e-01 5.91738403e-01
1.23591028e-01 -4.40717369e-01 -1.42603314e+00 -5.49401164e-01
2.93367386e-01 -3.26643556e-01 1.48566052e-01 7.86109090e-01
2.29253143e-01 5.13119757e-01 -3.85579288e-01 7.49358013e-02
7.76209056e-01 1.65200621e-01 6.49873197e-01 -1.16043806e+00
-7.60395527e-01 -2.23051786e-01 -8.24952900e-01 -8.60790491e-01
4.30705249e-02 -1.29373622e+00 -1.10869952e-01 -5.99923611e-01
2.40633339e-01 -4.64209378e-01 -4.83063340e-01 -1.61200389e-01
2.51835566e-02 2.65469640e-01 -1.10408794e-02 8.93677324e-02
-2.61785239e-01 6.48832500e-01 1.17431724e+00 -2.16513157e-01
-2.52092630e-01 -2.81159896e-02 -1.78637698e-01 1.46362409e-01
3.80937338e-01 -8.62416089e-01 -3.20800655e-02 2.26138949e-01
6.39839113e-01 3.96063417e-01 3.01419199e-01 -8.04207444e-01
2.58542448e-01 7.28798285e-02 8.00176486e-02 -8.65574479e-02
6.40458345e-01 -5.53958893e-01 1.67739898e-01 7.40555704e-01
-2.19024315e-01 -4.25040483e-01 -2.97201753e-01 6.96667433e-01
-1.41544297e-01 -8.09325635e-01 8.88366699e-01 -1.65992931e-01
-5.00847697e-01 1.05118454e-01 -2.46826857e-01 -3.96120638e-01
6.56173050e-01 -2.11350501e-01 -8.11847672e-02 -1.28953694e-03
-9.52136278e-01 -2.47339696e-01 5.86843789e-01 -2.82354027e-01
3.24781567e-01 -9.41022515e-01 -4.50433910e-01 5.24683356e-01
1.63974732e-01 -3.22578490e-01 3.42491597e-01 9.61104691e-01
-8.54629219e-01 7.83273637e-01 -7.59817287e-02 -6.22540891e-01
-7.52764106e-01 5.47958493e-01 4.43563461e-01 -1.51405796e-01
-4.07062054e-01 6.83919430e-01 -3.63813609e-01 -3.81733447e-01
-3.24983329e-01 -3.18527609e-01 1.93902597e-01 -2.16861799e-01
2.99858719e-01 2.84561634e-01 2.57373661e-01 -7.82803953e-01
-1.59061596e-01 5.44582129e-01 1.25641599e-01 -1.74777620e-02
1.17903793e+00 1.38521031e-01 -5.37304223e-01 5.97066581e-01
1.45096910e+00 -1.00355186e-01 -8.50384355e-01 -2.16173232e-01
-1.43845171e-01 -6.22678436e-02 3.89455110e-01 -1.90428644e-01
-4.21634555e-01 1.19246364e+00 8.78825068e-01 3.39804322e-01
5.55611014e-01 -1.61837056e-01 4.78205740e-01 1.17263865e+00
9.04804230e-01 -1.09823418e+00 -4.77312565e-01 4.97569680e-01
2.21406072e-01 -1.41073585e+00 1.79182127e-01 -1.81531116e-01
-1.69713363e-01 1.41859269e+00 -1.03247002e-01 -4.04098839e-01
5.16014278e-01 -1.85779750e-01 -4.72959131e-01 -2.30315313e-01
-5.37039995e-01 -2.79319853e-01 1.16609134e-01 2.07429871e-01
3.12612623e-01 4.43650275e-01 -7.18564510e-01 2.88113952e-01
-3.23942333e-01 -2.64336348e-01 8.44440222e-01 7.05762744e-01
-5.68316340e-01 -1.56687546e+00 -3.36392343e-01 4.04854417e-01
-1.82844654e-01 -3.86849761e-01 -5.37390634e-02 5.69090307e-01
1.19674779e-01 7.83614218e-01 -1.41357988e-01 -4.68450755e-01
1.00885652e-01 4.03596967e-01 1.03729701e+00 -4.64515984e-01
-4.82310683e-01 -1.41104028e-01 -1.27257615e-01 -6.75425887e-01
-3.34377110e-01 -4.90834534e-01 -1.56692743e+00 -6.25792980e-01
-7.95702100e-01 6.76069260e-01 1.07256377e+00 1.23721623e+00
5.22556305e-02 3.05958837e-01 7.23864019e-01 -8.80411506e-01
-1.08145642e+00 -8.15947890e-01 -1.04963148e+00 9.01177973e-02
4.43094432e-01 -6.46769583e-01 -4.57807750e-01 -3.75326306e-01] | [5.567449569702148, 4.952250957489014] |
989b73e7-14d4-4a4d-94f9-dd23036b15b9 | transfer-learning-from-high-resource-to-low | 2103.11764 | null | https://arxiv.org/abs/2103.11764v1 | https://arxiv.org/pdf/2103.11764v1.pdf | Transfer learning from High-Resource to Low-Resource Language Improves Speech Affect Recognition Classification Accuracy | Speech Affect Recognition is a problem of extracting emotional affects from audio data. Low resource languages corpora are rear and affect recognition is a difficult task in cross-corpus settings. We present an approach in which the model is trained on high resource language and fine-tune to recognize affects in low resource language. We train the model in same corpus setting on SAVEE, EMOVO, Urdu, and IEMOCAP by achieving baseline accuracy of 60.45, 68.05, 80.34, and 56.58 percent respectively. For capturing the diversity of affects in languages cross-corpus evaluations are discussed in detail. We find that accuracy improves by adding the domain target data into the training data. Finally, we show that performance is improved for low resource language speech affect recognition by achieving the UAR OF 69.32 and 68.2 for Urdu and Italian speech affects. | ['Umair Arshad', 'Sara Durrani'] | 2021-03-04 | null | null | null | null | ['cross-corpus'] | ['computer-vision'] | [-2.72240430e-01 4.96494919e-02 8.74871686e-02 -4.91112828e-01
-1.19833958e+00 -6.94645166e-01 3.87832046e-01 -2.78986961e-01
-5.39265573e-01 8.21447730e-01 6.02074444e-01 1.82001680e-01
7.03333139e-01 -1.44544825e-01 -1.20525032e-01 -3.13430756e-01
1.71872690e-01 -2.73681898e-03 -4.33730572e-01 -4.37355161e-01
-3.70408922e-01 3.05349052e-01 -1.39414048e+00 5.66638947e-01
3.95905405e-01 1.04176009e+00 -4.37551588e-01 9.91965890e-01
-1.87981173e-01 1.26801467e+00 -1.15877736e+00 -9.25421774e-01
-3.83065417e-02 -1.05116911e-01 -9.77518082e-01 -1.76105961e-01
1.51963100e-01 -4.09001373e-02 -1.60134971e-01 7.21933126e-01
1.10991788e+00 1.97606623e-01 6.97144032e-01 -1.07118452e+00
-8.17051113e-01 7.63187408e-01 -5.82642853e-01 4.16960269e-01
5.08636117e-01 -1.20192915e-01 9.65221345e-01 -7.75842071e-01
3.99638742e-01 1.76118672e+00 5.00438035e-01 9.04547274e-01
-8.94646585e-01 -7.74548709e-01 1.76449791e-01 3.16084214e-02
-9.93568242e-01 -8.90297174e-01 6.52390063e-01 -3.17710608e-01
1.75140345e+00 5.24366140e-01 2.00466111e-01 1.51535106e+00
-1.86966896e-01 1.06254053e+00 1.12562907e+00 -3.10815364e-01
-6.11910336e-02 3.30180466e-01 2.76372463e-01 2.61494994e-01
-6.65560007e-01 -2.15448707e-01 -1.00664604e+00 -2.24405289e-01
1.09653240e-02 -8.43775332e-01 -1.79014474e-01 9.50148344e-01
-8.72770071e-01 7.53511786e-01 -1.88593477e-01 6.30311608e-01
-7.28579462e-01 1.99379295e-01 1.20469677e+00 8.85306120e-01
1.07427347e+00 4.48193431e-01 -9.79813516e-01 -1.01451063e+00
-3.92772675e-01 -4.26304460e-01 8.27577233e-01 6.60715282e-01
3.27689677e-01 6.13502860e-01 -1.88048109e-01 1.59513664e+00
3.96360010e-02 7.84593642e-01 7.77088344e-01 -6.94942951e-01
3.83524686e-01 1.32274389e-01 -1.08737508e-02 -5.76932549e-01
-3.68779838e-01 -9.88261625e-02 -4.69664216e-01 -2.19007313e-01
1.17238715e-01 -1.00696230e+00 -7.72295356e-01 2.03100419e+00
3.03088099e-01 -2.99048543e-01 5.76144755e-01 5.51171124e-01
1.05041361e+00 1.01981747e+00 5.96034408e-01 -3.56077462e-01
1.53727543e+00 -7.81651914e-01 -1.20610046e+00 -4.97170150e-01
5.74016631e-01 -1.13690388e+00 1.57870781e+00 4.03027803e-01
-1.20114696e+00 -4.57257360e-01 -7.73773909e-01 -8.12352374e-02
-3.61672580e-01 4.00933862e-01 6.18553877e-01 9.87141430e-01
-1.07419050e+00 -7.63363913e-02 -4.18611079e-01 -3.47555280e-01
1.99638709e-01 2.72906125e-01 -4.14330572e-01 5.44160247e-01
-1.47678840e+00 1.02667284e+00 1.77055765e-02 -7.17238843e-01
-4.75243062e-01 -7.33200312e-01 -7.52770185e-01 -2.38798633e-01
-2.86586106e-01 -7.53841400e-02 1.51929474e+00 -1.66974378e+00
-1.97906291e+00 1.20566225e+00 -3.06916654e-01 -3.14386159e-01
4.38885726e-02 -4.94845212e-01 -9.63824034e-01 2.44207472e-01
-7.44537711e-02 6.69811666e-01 8.06729138e-01 -7.88558781e-01
-5.95626771e-01 -4.02817369e-01 -4.80977558e-02 2.99940556e-01
-5.24791479e-01 1.02855682e+00 -2.69189745e-01 -5.46999693e-01
-7.28891015e-01 -7.80379891e-01 3.57988238e-01 -5.83767533e-01
-1.82087377e-01 -4.34821308e-01 8.12355042e-01 -1.15063238e+00
1.28669369e+00 -2.51542139e+00 -2.12154731e-01 -1.30030066e-01
-3.27992260e-01 1.83266208e-01 -2.15922311e-01 1.71270192e-01
-3.86802703e-01 1.83247134e-01 3.03318113e-01 -4.89887744e-01
2.60132819e-01 8.62023905e-02 -6.39420822e-02 1.89482555e-01
5.95317841e-01 7.28351235e-01 -5.89296520e-01 -3.68004292e-01
1.18228979e-01 6.71653986e-01 -2.85367906e-01 3.99650067e-01
2.52236038e-01 1.87361851e-01 -3.72100085e-01 6.02063715e-01
3.76459867e-01 5.83157957e-01 -1.52424574e-01 -1.18334174e-01
-1.80416793e-01 7.31064737e-01 -7.83062041e-01 1.30262864e+00
-1.16208661e+00 8.92062247e-01 3.27480286e-01 -6.25802934e-01
1.11920631e+00 9.66684341e-01 2.69678563e-01 -8.87095094e-01
4.67691153e-01 5.14431484e-02 1.03532769e-01 -7.01213658e-01
3.49654406e-01 -7.52596438e-01 -6.18632615e-01 4.62453336e-01
2.58855999e-01 -2.07635745e-01 -7.38758743e-02 2.95707863e-03
1.23068810e+00 -5.06831884e-01 6.08896315e-01 -8.12542513e-02
6.64214075e-01 -3.37544292e-01 2.85339326e-01 1.73921973e-01
-9.01892066e-01 2.63818622e-01 5.37395835e-01 8.69807042e-03
-6.40058935e-01 -7.05916405e-01 -1.43974066e-01 1.64100301e+00
-8.16013336e-01 -2.80536354e-01 -7.70680785e-01 -6.26854897e-01
-2.75918454e-01 9.52558935e-01 -7.18690872e-01 -2.65406489e-01
-2.17687115e-01 -7.51515210e-01 1.17534077e+00 4.76819217e-01
4.80467767e-01 -1.33835137e+00 -4.03202891e-01 1.64923012e-01
-6.23481929e-01 -1.26378965e+00 -2.47671172e-01 3.22444916e-01
-9.24855992e-02 -4.88536358e-01 -4.23752099e-01 -6.64833248e-01
-3.75800580e-01 -3.51068407e-01 1.55082893e+00 -3.21469426e-01
-4.85025086e-02 6.51138246e-01 -4.02452648e-01 -9.18990850e-01
-6.20630205e-01 -6.95937946e-02 2.72603393e-01 8.78868103e-02
8.44981790e-01 -4.47963029e-01 7.78786689e-02 -8.73824134e-02
-3.76811028e-01 -4.48334008e-01 3.10554415e-01 5.91969669e-01
2.05756903e-01 -3.45772147e-01 1.33387327e+00 -4.44830894e-01
7.87692308e-01 -6.59634769e-01 3.78559858e-01 -1.53211847e-01
1.56764567e-01 -2.23469123e-01 5.80349505e-01 -6.46147311e-01
-1.65223241e+00 -5.08680455e-02 -7.46766806e-01 -1.15782976e-01
-5.58137476e-01 2.38722757e-01 -4.65005189e-01 4.31463242e-01
7.00614452e-01 -3.76916617e-01 -3.14696282e-01 -2.78162688e-01
3.80265057e-01 1.58018565e+00 4.82722014e-01 -3.94349486e-01
5.90135679e-02 2.58263379e-01 -8.47339451e-01 -1.07103527e+00
-8.60902250e-01 -5.94780564e-01 -1.71286449e-01 -3.00450057e-01
9.29437280e-01 -1.31539965e+00 -5.20681739e-01 4.17547911e-01
-1.23299932e+00 -3.71965051e-01 -2.41117865e-01 4.57703114e-01
-4.35623348e-01 3.64878960e-02 -1.11380827e+00 -1.20579541e+00
-9.28364575e-01 -6.66698813e-01 1.11346042e+00 -1.04690917e-01
-1.01518047e+00 -8.41385603e-01 1.82104960e-01 3.93022895e-01
3.28286141e-01 1.16383485e-01 5.91506362e-01 -8.68780494e-01
9.36527491e-01 -1.23128921e-01 -8.93184021e-02 5.85814059e-01
4.31470007e-01 1.98459566e-01 -1.63917589e+00 3.05938840e-01
3.50317359e-02 -1.02091801e+00 2.37932920e-01 9.50898826e-02
6.75320089e-01 -2.07495883e-01 3.36999923e-01 -1.06861547e-01
8.52775216e-01 5.43155968e-01 5.24009466e-01 5.13377739e-03
4.22522545e-01 7.76015878e-01 7.36072004e-01 7.06970751e-01
2.05084547e-01 5.30163050e-01 -9.98598039e-02 -2.07542907e-02
-2.75086850e-01 2.06163883e-01 1.03564441e+00 1.11181664e+00
3.63194585e-01 -2.51240760e-01 -7.60540366e-01 9.25426841e-01
-1.45362043e+00 -9.82244849e-01 -2.15207055e-01 1.57622457e+00
1.43002272e+00 -2.28823885e-01 4.23595935e-01 2.02292830e-01
5.78130960e-01 1.08837731e-01 -3.71223807e-01 -1.27887297e+00
-3.20341676e-01 5.21867156e-01 -1.21369861e-01 6.42701924e-01
-1.16264844e+00 1.25746202e+00 6.06111002e+00 1.01182210e+00
-1.31456327e+00 5.44237494e-01 9.88049328e-01 -5.75372875e-01
1.37768477e-01 -8.60524774e-01 -3.02858055e-01 3.58648747e-01
1.68033147e+00 -2.18058765e-01 3.55822295e-01 8.19241107e-01
4.10283536e-01 1.58430278e-01 -8.84896040e-01 1.47784722e+00
1.19642355e-01 -3.02656263e-01 -3.78517509e-01 -5.54689050e-01
4.88883108e-01 3.77114028e-01 8.47382620e-02 7.75727093e-01
3.11684042e-01 -1.02342856e+00 7.09397972e-01 -4.64094132e-02
1.00546432e+00 -1.30580437e+00 7.17071533e-01 -6.08708225e-02
-8.57677817e-01 1.75999627e-01 -3.88854116e-01 -3.45998287e-01
1.34384036e-01 5.75039327e-01 -7.43958056e-01 3.09153576e-03
9.59972322e-01 6.31416559e-01 -1.91937163e-01 -3.42606269e-02
-2.62366712e-01 1.10783815e+00 -3.62113178e-01 -3.41833323e-01
1.96759000e-01 2.16138065e-01 4.87877548e-01 1.90181172e+00
1.41082227e-01 1.00296862e-01 -2.47642606e-01 3.19664508e-01
-3.28893781e-01 7.09102154e-01 -6.32852077e-01 -2.15091959e-01
3.31027269e-01 1.56646585e+00 -6.46562735e-03 -4.92199540e-01
-4.91997898e-01 1.22565281e+00 6.24855399e-01 1.35426760e-01
-1.06040907e+00 -3.17893654e-01 1.40205836e+00 -7.73859084e-01
-4.38004620e-02 2.12428093e-01 -3.46760303e-01 -9.42556024e-01
-2.18955517e-01 -1.35150385e+00 2.92845756e-01 -9.59288657e-01
-1.44014692e+00 9.59267676e-01 -6.23243332e-01 -6.49409652e-01
-4.72998887e-01 -7.73172557e-01 -5.95341265e-01 9.07859981e-01
-1.11826742e+00 -9.49959695e-01 1.35920480e-01 6.53647602e-01
8.77863824e-01 -3.05999726e-01 1.05873489e+00 6.64042711e-01
-6.90266669e-01 8.51892531e-01 -4.56310630e-01 1.60155296e-01
1.24559200e+00 -1.32153773e+00 2.59841233e-01 5.42718649e-01
8.13653544e-02 1.72429740e-01 8.03565085e-01 -4.51329917e-01
-1.02360868e+00 -1.10258937e+00 1.17810059e+00 -5.93851328e-01
1.07022464e+00 -3.26993644e-01 -7.53801942e-01 6.76913023e-01
9.41366792e-01 -4.30301458e-01 1.23528051e+00 4.71178412e-01
-4.94101554e-01 1.84773222e-01 -1.42004991e+00 7.01113641e-01
7.77970314e-01 -9.64298010e-01 -4.36633557e-01 1.18423007e-01
8.55827510e-01 -1.47914350e-01 -1.00638199e+00 -9.68098864e-02
4.51907426e-01 -7.48997688e-01 6.29660964e-01 -9.32998836e-01
3.93023938e-01 1.78960472e-01 -6.04825556e-01 -1.81231451e+00
-1.39653198e-02 -8.79406512e-01 1.38782874e-01 1.94901729e+00
7.69137800e-01 -5.68409681e-01 -3.78976427e-02 7.09082484e-01
-1.99740320e-01 -5.25568366e-01 -1.05908883e+00 -3.23460907e-01
3.01384568e-01 -1.04332554e+00 4.27695006e-01 1.06018078e+00
5.04360795e-01 1.14352906e+00 -4.02072817e-01 2.05476442e-03
-1.99546665e-01 -4.41321254e-01 6.54780924e-01 -6.91449225e-01
-3.36662859e-01 -2.15908855e-01 -3.63531321e-01 -4.55702424e-01
8.67242217e-01 -4.62231994e-01 -1.38789788e-01 -9.87757325e-01
8.60454515e-02 5.67030907e-02 -2.09046930e-01 7.27898180e-01
-4.21094805e-01 4.91010159e-01 1.63481042e-01 -4.05876905e-01
-4.23963636e-01 7.34503508e-01 6.82983220e-01 -5.50196730e-02
-1.41818464e-01 -2.46992052e-01 -8.73578310e-01 8.63781869e-01
1.42877626e+00 -2.59017676e-01 -2.93008983e-01 -1.38492420e-01
8.39541256e-02 -9.93635207e-02 -1.95804104e-01 -6.91953838e-01
-6.11190259e-01 1.17701247e-01 2.04345852e-01 -2.21693888e-01
7.09138811e-01 -5.28450370e-01 -2.93366045e-01 -6.72343522e-02
-3.34091961e-01 2.01730981e-01 7.62663424e-01 -1.13241784e-01
-4.18800950e-01 -5.25809638e-02 9.57953453e-01 1.27855152e-01
-6.27320051e-01 -1.14364535e-01 -9.47264314e-01 4.23054248e-01
6.85268819e-01 4.37926561e-01 -2.35130504e-01 -9.70272005e-01
-7.00161994e-01 -1.26827136e-01 -1.05811216e-01 7.35432565e-01
1.67169824e-01 -1.38887739e+00 -7.86316812e-01 -2.24651203e-01
2.19931513e-01 -9.79687452e-01 2.07404867e-01 8.12630832e-01
1.70988023e-01 2.39222437e-01 -4.87542711e-02 2.02845149e-02
-1.91271245e+00 1.63555235e-01 6.35444701e-01 -1.81660503e-01
1.98425636e-01 1.00365627e+00 1.85343176e-01 -7.38022387e-01
6.27677292e-02 -2.45469078e-01 -2.70487994e-01 3.45170051e-01
7.11304307e-01 3.64499450e-01 1.83710083e-01 -1.02608180e+00
-4.86262143e-01 1.96350291e-02 2.31907025e-01 -5.74151993e-01
1.12029433e+00 -5.28969765e-01 -1.39873877e-01 9.68755722e-01
1.53144085e+00 5.67520201e-01 -6.89180911e-01 6.12813868e-02
-3.50698680e-02 2.15704635e-01 2.00189024e-01 -1.30543160e+00
-1.05027425e+00 9.00644124e-01 6.43267274e-01 1.48146495e-01
1.28834903e+00 2.05534503e-01 7.85432100e-01 3.17612112e-01
-7.39244297e-02 -1.71039009e+00 1.71356052e-01 8.89615834e-01
1.21695030e+00 -1.19710588e+00 -6.78028345e-01 -3.68361831e-01
-1.24616826e+00 7.25298524e-01 8.02509129e-01 1.02447838e-01
5.17167628e-01 7.40960360e-01 8.28202665e-01 -2.95708299e-01
-1.19793272e+00 -4.04700845e-01 1.53571844e-01 5.90915918e-01
1.22091484e+00 3.84781480e-01 -2.82977790e-01 1.10597384e+00
-6.13798738e-01 -5.40281534e-01 3.20228606e-01 5.23067534e-01
-9.14109424e-02 -9.00330603e-01 -4.39465523e-01 3.05332363e-01
-1.55478632e+00 -3.85796905e-01 -1.06067193e+00 7.55744576e-01
-1.63164541e-01 1.56892478e+00 1.62684441e-01 -3.54995459e-01
5.24485230e-01 7.90258169e-01 8.55587348e-02 -4.95476514e-01
-1.05402696e+00 2.92006671e-01 1.11358750e+00 -4.27592248e-01
-4.80134279e-01 -7.55143881e-01 -1.48343742e+00 -1.66478157e-01
3.15059982e-02 1.77560166e-01 6.04718208e-01 9.62537944e-01
5.70326567e-01 6.79109514e-01 6.97935641e-01 -4.11650747e-01
-1.69995740e-01 -1.43205881e+00 -5.56909621e-01 5.87072551e-01
3.61059159e-01 -3.36666167e-01 -5.77047646e-01 5.92384450e-02] | [13.567543029785156, 5.8486223220825195] |
fd2f669f-c76b-4f96-9f51-a131b1505e09 | meet-spinky-an-open-source-spindle-and-k | null | null | https://doi.org/10.3389/fninf.2017.00015 | https://www.frontiersin.org/articles/10.3389/fninf.2017.00015/pdf | Meet Spinky: An Open-Source Spindle and K-Complex Detection Toolbox Validated on the Open-Access Montreal Archive of Sleep Studies (MASS). | Sleep spindles and K-complexes are among the most prominent micro-events observed in electroencephalographic (EEG) recordings during sleep. These EEG microstructures are thought to be hallmarks of sleep-related cognitive processes. Although tedious and time-consuming, their identification and quantification is important for sleep studies in both healthy subjects and patients with sleep disorders. Therefore, procedures for automatic detection of spindles and K-complexes could provide valuable assistance to researchers and clinicians in the field. Recently, we proposed a framework for joint spindle and K-complex detection (Lajnef et al., 2015a) based on a Tunable Q-factor Wavelet Transform (TQWT; Selesnick, 2011a) and morphological component analysis (MCA). Using a wide range of performance metrics, the present article provides critical validation and benchmarking of the proposed approach by applying it to open-access EEG data from the Montreal Archive of Sleep Studies (MASS; O'Reilly et al., 2014). Importantly, the obtained scores were compared to alternative methods that were previously tested on the same database. With respect to spindle detection, our method achieved higher performance than most of the alternative methods. This was corroborated with statistic tests that took into account both sensitivity and precision (i.e., Matthew's coefficient of correlation (MCC), F1, Cohen κ). Our proposed method has been made available to the community via an open-source tool named Spinky (for spindle and K-complex detection). Thanks to a GUI implementation and access to Matlab and Python resources, Spinky is expected to contribute to an open-science approach that will enhance replicability and reliable comparisons of classifier performances for the detection of sleep EEG microstructure in both healthy and patient populations. | ['Sonia Frenette', 'Pierre-Emanuel Aguera', 'Sahbi Ch1aibi', "Christian O'Reilly", 'Karim Jerbi', 'Jb Eichenlaub', 'Etienne Combrisson', 'Abdennaceur Kachouri', 'Perrine Ruby', 'Mounir Samet', 'Julie Carrier', 'Tarek Lajnef'] | 2017-03-02 | null | null | null | frontiers-in-neuroinformatics-2017-3 | ['k-complex-detection', 'spindle-detection'] | ['medical', 'medical'] | [ 7.13720694e-02 -2.41593316e-01 1.94738984e-01 -1.40299246e-01
-5.62826693e-01 -5.90131283e-01 3.77244383e-01 5.96587598e-01
-7.34214723e-01 9.14307535e-01 -5.50489016e-02 -1.20127246e-01
-5.59018552e-01 -3.82850587e-01 -1.52986282e-02 -7.77373672e-01
-4.18398082e-01 1.31719783e-01 2.79358298e-01 5.02869263e-02
6.66268826e-01 5.01548350e-01 -1.71254098e+00 -1.41410172e-01
1.15319490e+00 1.08634663e+00 5.69713712e-01 3.14702094e-01
6.03028715e-01 8.92923251e-02 -6.34049177e-01 -3.49971682e-01
-1.70036495e-01 -4.42678839e-01 -4.87450570e-01 -1.09171771e-01
-1.91084072e-01 2.87809819e-01 3.25066924e-01 1.06231391e+00
6.72358036e-01 2.08418801e-01 5.49405992e-01 -8.77865195e-01
-1.80775419e-01 7.56441653e-02 -2.23411113e-01 1.04690123e+00
5.63430607e-01 2.78989077e-01 8.02131832e-01 -8.16339254e-01
3.38787317e-01 3.48478377e-01 5.83989799e-01 1.12203844e-01
-1.21803784e+00 -8.85633409e-01 -5.08586049e-01 8.74432981e-01
-1.48597157e+00 -6.74191415e-01 4.20725793e-01 -4.77034569e-01
9.98460293e-01 2.88448274e-01 1.29677665e+00 9.51510787e-01
6.85236573e-01 1.57089382e-01 1.71785522e+00 -1.37659341e-01
6.49336517e-01 1.85460031e-01 2.65780747e-01 3.00704569e-01
4.51046646e-01 -1.19519390e-01 -7.34092712e-01 -8.87830481e-02
3.94482672e-01 -2.12056637e-01 -4.07365113e-01 1.58560351e-01
-1.16667819e+00 5.97232521e-01 2.68595591e-02 6.36745453e-01
-7.31467605e-01 -4.38904762e-01 4.05905217e-01 1.27194330e-01
5.28669238e-01 5.00454307e-01 -2.25946397e-01 -7.32092083e-01
-1.49144399e+00 5.94474413e-02 4.65113580e-01 4.73182410e-01
4.54808950e-01 -1.21377692e-01 -1.15796290e-01 7.32107162e-01
2.21220687e-01 4.31755215e-01 8.88976693e-01 -5.63953102e-01
1.39958888e-01 6.30020261e-01 8.52577109e-03 -9.63125348e-01
-1.05803990e+00 -6.80440784e-01 -8.26173246e-01 -5.90239130e-02
1.81073457e-01 1.70712814e-01 -2.47372508e-01 1.47077155e+00
-8.39316249e-02 1.77982852e-01 -1.88084453e-01 9.24968421e-01
7.36036777e-01 1.96943693e-02 5.99573413e-03 -6.25211716e-01
1.86860621e+00 -2.46915713e-01 -7.78681517e-01 -1.65101096e-01
2.06695989e-01 -7.53799677e-01 1.16304862e+00 6.21045589e-01
-1.16521692e+00 -4.15611535e-01 -1.17827785e+00 3.20566356e-01
-2.44078904e-01 3.31517100e-01 3.48426610e-01 8.98004711e-01
-1.23577762e+00 3.70128334e-01 -1.23950422e+00 -8.63707602e-01
5.79144418e-01 5.82408071e-01 -5.33271790e-01 3.59015316e-01
-8.12719405e-01 1.21556926e+00 5.78701459e-02 2.05891460e-01
-6.42817020e-01 -3.63501787e-01 -4.24010724e-01 2.97447480e-02
5.47837093e-02 -6.35737181e-01 6.90892100e-01 -5.71831644e-01
-1.27411091e+00 9.64302063e-01 -3.06692570e-01 -4.34895188e-01
-4.03556190e-02 -6.29991814e-02 -5.38402975e-01 4.30612355e-01
2.33252466e-01 7.23306462e-02 6.88802063e-01 -4.65166390e-01
-3.11401248e-01 -7.19631016e-01 -1.74510777e-01 -2.20847111e-02
-1.97619304e-01 2.78879493e-01 2.57209957e-01 -6.42417431e-01
1.04410946e-01 -6.95012748e-01 3.34631652e-01 -5.10379553e-01
-9.92604122e-02 -2.19534308e-01 1.23036921e-01 -8.03028822e-01
1.28959656e+00 -2.04291272e+00 1.94405720e-01 2.68904895e-01
4.32981491e-01 1.76084116e-01 2.91171253e-01 4.59351093e-01
-8.64418298e-02 -1.11229494e-01 -3.51316005e-01 -2.69971818e-01
-6.71909600e-02 -2.92981356e-01 2.38715366e-01 1.13625968e+00
1.24079198e-01 6.23706818e-01 -7.91516721e-01 -1.71517938e-01
2.70767748e-01 3.81224275e-01 -1.96868554e-01 3.06673590e-02
4.70891535e-01 5.23133636e-01 1.15098171e-01 5.82754612e-01
4.57611740e-01 -4.99134213e-02 -1.09524779e-01 -1.75716624e-01
-5.48138559e-01 5.36394715e-01 -8.67783129e-01 1.52382290e+00
-2.42177248e-01 6.70422673e-01 -1.44868806e-01 -8.69458914e-01
8.89864802e-01 4.97280836e-01 9.81590301e-02 -7.49421835e-01
2.90828884e-01 3.10863197e-01 4.99175727e-01 -4.98940617e-01
2.60516852e-01 -2.94522554e-01 2.92239159e-01 4.87986237e-01
2.84365982e-01 -1.89463735e-01 4.61084455e-01 -4.03467417e-02
1.18328023e+00 -2.06740946e-01 7.26168096e-01 -7.52639234e-01
6.37800455e-01 -4.64170724e-01 2.49251410e-01 4.02251154e-01
-4.31431592e-01 3.32810134e-01 4.93971944e-01 1.18654054e-02
-5.27712345e-01 -1.04218256e+00 -6.28160715e-01 5.25198460e-01
-1.37209386e-01 -5.81238329e-01 -8.25946569e-01 -1.67798642e-02
-3.95646900e-01 6.66631997e-01 -6.87040091e-01 -3.76116455e-01
1.23212822e-01 -1.26883495e+00 4.22791511e-01 1.93841115e-01
1.88711345e-01 -1.23445070e+00 -1.25598502e+00 9.52329114e-02
-1.14038266e-01 -1.11777341e+00 -1.97068974e-02 4.67610568e-01
-9.44973886e-01 -1.31619716e+00 -5.13191044e-01 -3.32625836e-01
3.82202417e-01 1.15615197e-01 7.94567883e-01 -2.31626872e-02
-4.68002707e-01 3.91444296e-01 -5.81441879e-01 -3.08360994e-01
1.23573095e-02 7.52014369e-02 3.80155295e-01 -6.08622506e-02
5.29083669e-01 -1.09568572e+00 -9.75651920e-01 2.63862312e-01
-7.25492239e-01 -3.10911745e-01 7.70590305e-01 4.64064211e-01
3.92485619e-01 -8.38970765e-02 9.74269509e-01 -2.05128610e-01
9.33133662e-01 -6.99782431e-01 -6.68654919e-01 -2.41869405e-01
-9.06476498e-01 -4.54805404e-01 6.09762371e-01 -5.66942468e-02
-5.13998866e-01 -5.64439535e-01 -2.22871885e-01 1.81740895e-01
-4.26062912e-01 4.80985910e-01 4.68027405e-02 -2.15501353e-01
4.26818579e-01 4.76927757e-01 -1.13764465e-01 -2.06765622e-01
-3.39661509e-01 7.04180598e-01 4.75268066e-01 -1.03144445e-01
3.76181453e-01 2.88120598e-01 3.71472612e-02 -1.13267958e+00
-4.41661626e-01 -7.33663857e-01 -6.63240075e-01 -2.13235617e-01
9.38760459e-01 -8.91731262e-01 -9.09875453e-01 4.07520294e-01
-7.42227793e-01 -1.22444831e-01 2.36262590e-01 8.38815808e-01
-5.16216695e-01 4.88897234e-01 -2.26868019e-01 -9.44966435e-01
-6.10891163e-01 -1.13581026e+00 7.39564300e-01 4.34226483e-01
-6.63940012e-01 -8.78588021e-01 2.71274835e-01 2.24692822e-01
4.61331189e-01 6.32317215e-02 8.30454528e-01 -8.31370771e-01
-3.53817791e-02 -1.28686324e-01 1.45002320e-01 2.16168419e-01
2.86803424e-01 -2.10144818e-01 -9.62132871e-01 -2.14072809e-01
6.00870788e-01 -5.60843274e-02 3.55695635e-01 4.53060150e-01
7.38006651e-01 5.92641644e-02 5.39261699e-02 3.81178975e-01
1.19871581e+00 3.47179353e-01 7.56250381e-01 7.38631725e-01
-1.53339252e-01 3.97399396e-01 3.17763627e-01 5.84844589e-01
4.62599725e-01 7.87977338e-01 3.28700930e-01 2.35571623e-01
7.66730681e-02 5.66511750e-01 4.85749364e-01 9.07484770e-01
-4.72372681e-01 1.98782876e-01 -7.63446033e-01 5.51173806e-01
-1.36119521e+00 -9.64336455e-01 -4.59967852e-01 2.40329981e+00
5.52110493e-01 3.35422546e-01 4.08622712e-01 4.83768135e-01
4.22828466e-01 -2.89395869e-01 -3.25951219e-01 -4.02447402e-01
-2.25492865e-01 8.20651889e-01 5.14060780e-02 -6.23583160e-02
-6.58959508e-01 4.09677923e-01 5.70597219e+00 6.33900642e-01
-1.04381943e+00 4.41484749e-01 1.10013567e-01 -3.48085910e-01
2.29599759e-01 -2.50056595e-01 -3.60338271e-01 8.87193084e-01
1.49328613e+00 -4.13167149e-01 8.19349051e-01 2.87424028e-01
9.14362848e-01 -8.99962842e-01 -8.23480129e-01 1.03557801e+00
3.40925306e-02 -1.08667636e+00 -8.37814391e-01 -4.17821482e-02
2.78135002e-01 1.61294609e-01 -6.46090731e-02 -4.79750969e-02
-7.03802466e-01 -8.95603955e-01 7.62755632e-01 8.35504174e-01
7.81676710e-01 -7.19802916e-01 1.17481399e+00 1.44083530e-01
-1.05435729e+00 1.52643479e-03 -1.17099144e-01 -3.05873185e-01
-2.34463420e-02 8.48710120e-01 -4.88784879e-01 4.57124323e-01
9.32996869e-01 7.18855023e-01 -9.33198154e-01 1.44104457e+00
-2.52010256e-01 9.13193226e-01 -1.31927043e-01 -1.38135105e-01
-1.72811583e-01 -4.61470723e-01 6.32575691e-01 9.81773496e-01
5.81725061e-01 -5.18543944e-02 -7.13744879e-01 1.20148814e+00
3.64833921e-01 1.41553462e-01 -1.14949860e-01 -7.70311952e-02
5.43237746e-01 1.71677065e+00 -1.48393404e+00 -1.59716867e-02
-4.62042660e-01 6.51452243e-01 8.69077221e-02 8.16152841e-02
-5.20818114e-01 -4.68123227e-01 4.15254325e-01 9.53781977e-02
5.30421771e-02 -9.95244384e-02 -5.06383121e-01 -1.03655708e+00
2.29669049e-01 -8.02644908e-01 6.03026785e-02 -7.11130917e-01
-1.06646550e+00 7.35481262e-01 2.04779103e-01 -1.05511832e+00
5.15550673e-02 -4.34471071e-01 -7.82850862e-01 1.02796972e+00
-1.17279577e+00 -6.26946747e-01 -3.69656473e-01 5.98563075e-01
3.10261667e-01 -1.81481559e-02 1.05956531e+00 3.47999781e-01
-6.80731416e-01 3.24408621e-01 9.16099474e-02 -5.03955185e-01
5.15920043e-01 -1.20191729e+00 1.20032057e-02 7.94836283e-01
-1.45869069e-02 9.62330997e-01 8.42005730e-01 -4.01508808e-01
-1.12092650e+00 -5.61815321e-01 7.81356156e-01 -2.07591638e-01
9.76706028e-01 -4.29964840e-01 -6.54711068e-01 3.64475757e-01
2.08188221e-01 -5.14941692e-01 1.19524741e+00 -8.40881467e-03
4.51036751e-01 -1.56856820e-01 -1.11584878e+00 1.16596654e-01
5.37718415e-01 -5.00670850e-01 -9.04518723e-01 1.81259200e-01
-1.71690509e-01 -2.47965660e-03 -1.14633322e+00 1.27784163e-01
6.07605517e-01 -1.63154602e+00 4.71395791e-01 1.82649285e-01
4.26828377e-02 -2.77509123e-01 2.33426213e-01 -1.37402236e+00
-2.13228092e-01 -5.08025765e-01 4.50312048e-02 9.82399046e-01
4.08577872e-03 -8.24553549e-01 1.61087528e-01 1.63169369e-01
-5.81950188e-01 -8.07740986e-01 -1.24820781e+00 -6.91666186e-01
-2.54685849e-01 -5.88182211e-01 2.85140783e-01 3.39553863e-01
4.58370298e-01 1.26315430e-01 1.78008646e-01 8.38561803e-02
3.78746927e-01 -1.61504015e-01 1.88296184e-01 -1.42234230e+00
9.07103792e-02 -6.36274397e-01 -7.88832188e-01 -3.10303308e-02
6.69369400e-02 -1.08474183e+00 -2.18294740e-01 -1.43557858e+00
4.12840843e-01 -6.93129376e-02 -4.56667066e-01 3.02929252e-01
-8.10246095e-02 7.10608363e-01 -8.57792422e-02 3.00267458e-01
-4.60926086e-01 4.91195172e-01 6.65574670e-01 4.30920899e-01
-2.66359031e-01 1.25666633e-01 -7.27199733e-01 6.75392509e-01
9.77150738e-01 -5.69399476e-01 -3.69045615e-01 3.06525767e-01
3.95746261e-01 -6.41781613e-02 5.72680354e-01 -1.33249009e+00
1.83152929e-01 3.06511372e-01 3.79759878e-01 -3.97831261e-01
3.98454785e-01 -5.52443266e-01 2.45390266e-01 4.88006115e-01
2.11126566e-01 4.17630553e-01 3.93290699e-01 1.80986628e-01
1.78091340e-02 -2.96551198e-01 6.91268682e-01 6.92032576e-02
-2.88022995e-01 -1.22341380e-01 -9.22705472e-01 -1.62542313e-01
1.04523039e+00 -4.61647928e-01 -4.78478611e-01 -2.46641248e-01
-8.57619584e-01 -1.77738681e-01 5.31464338e-01 7.53686130e-02
6.94414496e-01 -8.65232646e-01 -4.20756906e-01 3.99532765e-01
1.48645461e-01 -5.69662929e-01 3.55451286e-01 2.01158452e+00
-3.33745509e-01 5.86494565e-01 -7.22506762e-01 -5.03430963e-01
-1.07967782e+00 2.05525234e-01 2.11931318e-02 -4.20844257e-02
-6.11782372e-01 5.27230382e-01 -2.75483936e-01 3.30189437e-01
-3.62117961e-02 -4.38481539e-01 -4.70782906e-01 4.17454451e-01
6.41921937e-01 7.51750827e-01 6.55108750e-01 -7.42594838e-01
-7.71195114e-01 3.13187480e-01 3.49363148e-01 4.50587459e-03
1.60030544e+00 -3.56581748e-01 -7.05343187e-01 9.34146702e-01
6.82900846e-01 2.46986836e-01 -5.74508548e-01 5.96615374e-01
2.66948283e-01 -7.84448460e-02 2.86679044e-02 -8.26167047e-01
-7.87121296e-01 8.01465690e-01 9.26187515e-01 4.71862763e-01
1.33042395e+00 -1.64543297e-02 4.41662461e-01 5.28684929e-02
5.35185695e-01 -8.64154756e-01 -1.44083023e-01 -2.48095188e-02
6.90121949e-01 -7.58455873e-01 2.01081187e-01 -2.12517884e-02
-3.40522647e-01 1.18393183e+00 1.26451135e-01 -1.50911570e-01
5.84648669e-01 7.96022341e-02 -2.52467275e-01 -4.26859468e-01
-6.97066188e-01 -3.23316187e-01 4.09740210e-01 5.66236258e-01
6.24454558e-01 1.20492235e-01 -1.09040272e+00 1.19443738e+00
-6.31277084e-01 1.74323425e-01 7.32785821e-01 5.38854241e-01
-3.85640681e-01 -7.82090068e-01 -5.40976286e-01 1.01934779e+00
-6.20645821e-01 -2.49366894e-01 -6.41357852e-03 7.39319205e-01
2.56775647e-01 1.38916814e+00 -2.08876990e-02 -3.65634680e-01
1.41278535e-01 2.02576295e-01 5.59362292e-01 -7.78054357e-01
-7.88763523e-01 -7.35816956e-02 -1.46583140e-01 -5.35439551e-01
-8.88339698e-01 -9.69314516e-01 -1.00262296e+00 4.15961333e-02
-3.36284727e-01 3.22207212e-01 6.23031139e-01 1.21093130e+00
4.06850934e-01 5.84123790e-01 1.94307491e-01 -8.59549999e-01
1.39166098e-02 -1.42507672e+00 -9.81629789e-01 3.55147868e-02
1.27167165e-01 -1.21153057e+00 -5.87671757e-01 6.15857802e-02] | [13.407393455505371, 3.443270444869995] |
20778b43-335f-4731-8b35-4d0687ef5fbf | cross-modal-data-discovery-over-structured | 2306.00932 | null | https://arxiv.org/abs/2306.00932v2 | https://arxiv.org/pdf/2306.00932v2.pdf | Cross Modal Data Discovery over Structured and Unstructured Data Lakes | Organizations are collecting increasingly large amounts of data for data driven decision making. These data are often dumped into a centralized repository, e.g., a data lake, consisting of thousands of structured and unstructured datasets. Perversely, such mixture of datasets makes the problem of discovering elements (e.g., tables or documents) that are relevant to a user's query or an analytical task very challenging. Despite the recent efforts in data discovery, the problem remains widely open especially in the two fronts of (1) discovering relationships and relatedness across structured and unstructured datasets where existing techniques suffer from either scalability, being customized for a specific problem type (e.g., entity matching or data integration), or demolishing the structural properties on its way, and (2) developing a holistic system for integrating various similarity measurements and sketches in an effective way to boost the discovery accuracy. In this paper, we propose a new data discovery system, named CMDL, for addressing these two limitations. CMDL supports the data discovery process over both structured and unstructured data while retaining the structural properties of tables. | ['Mohammad Shahmeer Ahmad', 'Ahmed Elmagarmid', 'Mayuresh Kunjir', 'Mohamed Y. Eltabakh'] | 2023-06-01 | null | null | null | null | ['data-integration'] | ['knowledge-base'] | [-1.47475049e-01 -2.14283884e-01 -2.32526705e-01 -3.18454534e-01
-4.93855059e-01 -5.82525253e-01 3.87572289e-01 1.08984506e+00
-1.84190914e-01 6.20803773e-01 4.64898556e-01 -1.79380029e-01
-7.02231050e-01 -1.20375049e+00 -1.43510669e-01 -3.65789622e-01
-6.66950345e-02 7.19018340e-01 3.77791107e-01 -1.70606703e-01
5.25080740e-01 5.05138934e-01 -1.97396183e+00 4.17209059e-01
1.04572093e+00 1.14871204e+00 2.11726546e-01 1.03475824e-01
-8.88211668e-01 7.57484972e-01 -3.94882977e-01 -5.58878422e-01
2.64894396e-01 -9.84902680e-03 -9.07964706e-01 -2.52354354e-01
1.45049840e-01 -8.28804150e-02 6.77677467e-02 1.19731319e+00
3.80620718e-01 -2.03489751e-01 1.44042388e-01 -1.40912974e+00
-5.17576933e-01 7.55142570e-01 -6.23528838e-01 1.36170983e-01
4.60339308e-01 -3.75415653e-01 1.02850938e+00 -1.00901592e+00
8.25257301e-01 1.00140309e+00 6.55275464e-01 -5.34258597e-02
-8.65119636e-01 -5.51743031e-01 -4.83223498e-02 1.78976029e-01
-1.62204003e+00 -5.44667602e-01 5.99553466e-01 -6.20649636e-01
6.72883153e-01 5.74716687e-01 2.38836229e-01 4.99640703e-01
-1.96877435e-01 2.31517032e-01 6.75324023e-01 -2.40094215e-01
3.96826684e-01 5.41778147e-01 4.82893586e-01 6.70752823e-02
1.12528062e+00 -6.55423582e-01 -4.62052613e-01 -5.60172856e-01
2.96342701e-01 3.43110830e-01 1.12585552e-01 -5.36869407e-01
-1.19138432e+00 4.55312431e-01 -4.74918354e-03 5.64137042e-01
-5.86630702e-01 -7.03652084e-01 6.30751431e-01 3.52650255e-01
5.63380569e-02 6.37019813e-01 -4.13291514e-01 -1.41003996e-01
-6.21454120e-01 4.56773967e-01 1.05325294e+00 1.31582451e+00
9.46107090e-01 -5.59689045e-01 1.17865369e-01 7.10674226e-01
2.31988490e-01 4.49354062e-03 5.42382717e-01 -4.73001212e-01
8.73728931e-01 1.52535748e+00 2.57269263e-01 -1.45194197e+00
-4.58124608e-01 -2.59108305e-01 -9.28848565e-01 -3.40001255e-01
3.12315673e-01 1.91350609e-01 -3.08840990e-01 1.34552324e+00
6.83788657e-01 -5.24652243e-01 1.16399862e-01 6.78523600e-01
9.93747652e-01 3.52930784e-01 9.29728523e-02 -3.91054273e-01
1.49179447e+00 -3.78704399e-01 -9.38463807e-01 -1.22679174e-02
6.78306758e-01 -7.88171649e-01 7.46246874e-01 2.21848071e-01
-9.79101062e-01 -4.16482687e-01 -7.81301975e-01 -2.20771164e-01
-8.29128981e-01 -2.73447365e-01 4.17704731e-01 2.96911627e-01
-8.06879878e-01 4.24074769e-01 -3.95598084e-01 -6.68232143e-01
8.59678686e-02 3.73047054e-01 -4.58726466e-01 -1.68228999e-01
-1.14224577e+00 6.14133418e-01 6.68815792e-01 -2.42023692e-01
5.72452461e-03 -6.96493745e-01 -4.28130299e-01 2.61678874e-01
8.26299489e-01 -5.47571540e-01 7.92321503e-01 -4.52157021e-01
-3.44968528e-01 7.42036581e-01 1.24512315e-02 -2.28255719e-01
2.19799459e-01 -5.55479936e-02 -8.41886044e-01 -3.11365098e-01
4.09034222e-01 -1.76669046e-01 1.35841206e-01 -1.23308253e+00
-9.83573854e-01 -1.04790986e+00 -2.01135445e-02 1.27257362e-01
-6.82543695e-01 3.19903702e-01 -5.67963719e-01 -5.03511310e-01
5.22409737e-01 -3.69043589e-01 -1.75381094e-01 -1.97779238e-01
-5.84032953e-01 -3.88613254e-01 9.28114295e-01 -6.26371682e-01
1.93763161e+00 -2.01451612e+00 2.84945220e-02 3.87021124e-01
6.27453387e-01 2.93665528e-01 2.55592883e-01 1.16904402e+00
8.07461739e-02 5.07883787e-01 -1.31901264e-01 9.90148410e-02
-1.07917793e-01 1.21777900e-01 -2.54866153e-01 -7.65722767e-02
-2.33049802e-02 5.52909315e-01 -8.10746014e-01 -7.97541142e-01
-2.02106267e-01 -2.82616373e-02 -1.96682215e-01 2.51111984e-01
6.63707703e-02 2.27043927e-01 -7.49869466e-01 1.04417765e+00
7.27664471e-01 -3.74701142e-01 4.50798482e-01 -2.67191231e-01
-4.69208777e-01 3.81890267e-01 -1.67767656e+00 1.29949820e+00
-4.98465188e-02 8.66373628e-02 2.03102410e-01 -1.02437890e+00
1.21566975e+00 3.02553624e-01 8.96505415e-01 -6.94215059e-01
-1.12716913e-01 5.68564713e-01 3.46999094e-02 -7.62487948e-01
7.81276584e-01 1.79571941e-01 -1.02830231e-01 5.72144747e-01
-4.26447302e-01 6.76087379e-01 4.72619683e-01 2.55624264e-01
1.28635049e+00 -3.93110901e-01 6.84651375e-01 -2.76852101e-01
5.03131628e-01 4.09177005e-01 7.54140258e-01 5.17771900e-01
1.65477768e-01 2.96043813e-01 4.21697170e-01 -5.62684178e-01
-1.26639771e+00 -6.96343839e-01 -3.04647297e-01 8.10724974e-01
3.89263600e-01 -9.65639293e-01 -3.21742654e-01 -4.20011461e-01
5.13254941e-01 4.14480418e-01 -3.32360536e-01 4.11031060e-02
-2.79034197e-01 -5.07722378e-01 3.85597080e-01 4.15129662e-01
4.56469923e-01 -8.02887321e-01 -5.23138165e-01 4.98106748e-01
-2.51709908e-01 -9.69611526e-01 -1.72740091e-02 3.73843499e-02
-8.28942239e-01 -1.26459610e+00 1.29987091e-01 -5.10988891e-01
5.67497730e-01 6.10243797e-01 1.26938641e+00 2.75695890e-01
-2.73023248e-01 -6.57020286e-02 -6.98633432e-01 -5.01810670e-01
-2.11951017e-01 1.14456035e-01 2.08388656e-01 -5.25472760e-02
7.73940563e-01 -7.06807196e-01 -4.74161804e-01 6.48212194e-01
-1.17163742e+00 -7.04264939e-02 6.17703736e-01 3.57182831e-01
6.08176112e-01 4.43049282e-01 7.81580627e-01 -1.07271385e+00
8.98776352e-01 -1.05731571e+00 -3.92493486e-01 7.00900853e-01
-1.10299444e+00 4.77140285e-02 8.01777363e-01 -8.60557631e-02
-7.62553096e-01 -1.74846247e-01 7.58320764e-02 -3.61634456e-02
-1.30262747e-01 9.34664607e-01 -4.57308829e-01 4.26443189e-01
4.23964977e-01 1.00228384e-01 -9.42861289e-02 -9.67453122e-01
8.58206823e-02 1.14194930e+00 5.73018909e-01 -6.12049818e-01
8.96806240e-01 2.71395296e-01 -2.21447617e-01 -4.02099818e-01
-6.05930090e-01 -8.79247069e-01 -8.85614812e-01 1.04148984e-01
5.41619778e-01 -7.31616914e-01 -5.51492631e-01 -1.73511207e-02
-8.25663090e-01 7.37721980e-01 -3.84661108e-01 2.66138613e-01
2.74449103e-02 3.07790637e-01 -1.19358553e-02 -6.25164866e-01
-3.51937175e-01 -7.89220691e-01 6.45819485e-01 2.11341411e-01
-4.29566652e-01 -5.77893496e-01 2.35791534e-01 4.85836864e-01
4.63169008e-01 4.17070061e-01 1.19897056e+00 -1.15499783e+00
-6.04464114e-01 -5.17399907e-01 -4.75326836e-01 -1.50835440e-01
5.98899782e-01 6.58807606e-02 -5.29038846e-01 -1.64390713e-01
-9.11486894e-02 -4.13421206e-02 8.76674205e-02 -4.05794531e-01
1.17102206e+00 -5.88261724e-01 -5.37053168e-01 3.31767648e-01
1.37285316e+00 5.58145225e-01 5.41241407e-01 7.12112784e-01
6.11326337e-01 1.03425193e+00 8.64365160e-01 8.61219049e-01
5.86698651e-01 7.60082066e-01 2.84266770e-01 1.70534834e-01
3.42618555e-01 -3.89704317e-01 -4.39759105e-01 9.23288941e-01
-1.43136736e-02 -1.74074709e-01 -1.27035296e+00 6.15150094e-01
-2.14484000e+00 -1.02365828e+00 -3.86241287e-01 2.40319204e+00
8.93715799e-01 -2.77251247e-02 8.24226886e-02 3.19634885e-01
7.43497968e-01 -2.90346742e-01 -6.77935481e-01 -1.06843159e-01
-3.35765593e-02 -3.61685693e-01 3.30806613e-01 -3.25889826e-01
-6.54220939e-01 3.08159113e-01 5.27743196e+00 4.18490261e-01
-8.42423081e-01 -2.41329893e-01 3.49947333e-01 2.31512517e-01
-4.72469807e-01 1.88232437e-01 -1.04875791e+00 6.36962235e-01
8.29881549e-01 -7.18889236e-01 1.54296085e-01 9.20499682e-01
1.53741658e-01 -8.74017328e-02 -1.13718104e+00 9.78842378e-01
-2.35363156e-01 -1.42639351e+00 1.86090529e-01 3.66820276e-01
3.84545118e-01 -3.16254050e-01 -3.82486075e-01 5.76727837e-02
2.72403300e-01 -9.20399308e-01 5.06452858e-01 4.55694318e-01
5.86947441e-01 -5.92648864e-01 7.96892405e-01 5.43335855e-01
-1.28512537e+00 -1.90444306e-01 -3.57911915e-01 -6.95188250e-03
-2.08149523e-01 7.13598788e-01 -6.18985474e-01 9.24628794e-01
1.07121992e+00 6.81682825e-01 -6.98885798e-01 1.18572485e+00
4.92445678e-01 1.92613155e-01 -1.39995426e-01 1.65593967e-01
4.98171076e-02 -3.81487221e-01 2.70502478e-01 8.77170146e-01
4.68815207e-01 1.47048041e-01 2.20662028e-01 6.45384848e-01
-1.36578321e-01 5.24099410e-01 -8.31937253e-01 -2.71013469e-01
1.08687901e+00 1.33749378e+00 -6.55937016e-01 -3.29076111e-01
-6.60101891e-01 2.13183329e-01 4.06693488e-01 -5.44031523e-02
-3.24862063e-01 -6.92351341e-01 8.38325143e-01 6.03139758e-01
1.67677775e-02 -1.67686194e-01 -5.02801716e-01 -1.02312827e+00
4.14449245e-01 -1.04295504e+00 6.25014961e-01 -4.26895410e-01
-1.40184081e+00 5.53651690e-01 1.44972801e-02 -1.51070535e+00
-1.67205885e-01 -1.31550640e-01 -3.64554971e-01 8.58527422e-01
-1.03533828e+00 -6.95514321e-01 -6.20900869e-01 5.80414951e-01
1.07252285e-01 -1.16015263e-01 6.06012106e-01 7.59460568e-01
-6.45648181e-01 2.81086653e-01 4.39752102e-01 1.28934264e-01
7.65561998e-01 -1.09861052e+00 3.45701307e-01 5.98201275e-01
-6.35277927e-02 1.20944607e+00 5.07372856e-01 -8.00084174e-01
-1.77010190e+00 -9.52381849e-01 1.29454100e+00 -4.98759836e-01
8.14769447e-01 -3.75898153e-01 -1.44763517e+00 3.66769880e-01
-1.68687850e-01 -1.27662584e-01 1.00385761e+00 4.32409137e-01
-5.39282799e-01 -4.82438624e-01 -1.39056313e+00 3.39148819e-01
1.19804215e+00 -2.64278799e-01 -6.94429636e-01 8.21337178e-02
5.82360208e-01 5.35092689e-02 -1.29297686e+00 3.90483558e-01
5.13095737e-01 -9.95999157e-01 9.33392346e-01 -9.57499981e-01
3.64118725e-01 -5.17123759e-01 -3.99802715e-01 -8.43624651e-01
-3.43063831e-01 -5.02717137e-01 -2.32258558e-01 1.78847146e+00
2.49244153e-01 -5.33339322e-01 5.44382095e-01 9.44749057e-01
1.61896810e-01 -7.65914500e-01 -6.84824646e-01 -6.56360269e-01
-4.61802661e-01 -1.10615470e-01 1.13198173e+00 1.30715358e+00
1.65812552e-01 4.64069694e-01 -2.14411214e-01 1.97350323e-01
5.47085166e-01 5.06502450e-01 9.91169930e-01 -1.81474090e+00
3.05753201e-01 -4.12042469e-01 -4.63133842e-01 -3.87878090e-01
-6.85855567e-01 -7.98883855e-01 -4.34078902e-01 -1.94926488e+00
3.01509082e-01 -1.04133201e+00 -3.19144994e-01 4.12024051e-01
-1.28064960e-01 -4.66580987e-01 -5.23537435e-02 8.63170385e-01
-6.13865077e-01 -2.85418984e-02 8.04351628e-01 1.62180196e-02
-3.05818468e-01 -1.40167803e-01 -1.26293850e+00 5.22676289e-01
7.21306801e-01 -5.94738901e-01 -5.53580403e-01 -3.26865405e-01
5.46825290e-01 1.36900663e-01 -2.34848186e-02 -9.01623666e-01
7.37699509e-01 -4.02120620e-01 4.28687856e-02 -9.17413414e-01
-2.61121631e-01 -1.06028032e+00 7.72678316e-01 2.46858507e-01
-2.67100811e-01 4.48811859e-01 -9.91979986e-02 5.34564018e-01
-6.13905430e-01 -1.78440988e-01 3.93614471e-01 -2.21142456e-01
-8.31223965e-01 2.92569727e-01 2.82824874e-01 3.04640889e-01
1.09681690e+00 -4.45429355e-01 -6.00091815e-01 8.03523660e-02
-3.87688160e-01 5.31128347e-01 6.23592377e-01 6.30819023e-01
3.53994220e-01 -1.26874208e+00 -7.15957344e-01 1.72026634e-01
6.17480338e-01 2.68454731e-01 -9.47391521e-03 7.60843098e-01
-2.93892711e-01 4.47582930e-01 -3.59035879e-01 -1.77404583e-01
-1.32220614e+00 7.28694499e-01 -3.01435351e-01 -5.49802899e-01
-6.34510517e-01 2.94036210e-01 -4.66578454e-02 -4.07671750e-01
3.44188422e-01 -1.35268092e-01 -3.38461936e-01 6.22755110e-01
7.71115720e-01 4.78540957e-01 4.01864350e-01 -5.12593269e-01
-5.47774673e-01 3.72638494e-01 -2.80378759e-01 5.68785906e-01
1.50859213e+00 -3.60148430e-01 -6.10346556e-01 4.21525508e-01
9.22635615e-01 3.06678861e-01 -5.34407496e-01 -6.30354345e-01
8.11882079e-01 -7.25023031e-01 -3.53372812e-01 -6.81349576e-01
-8.02449942e-01 3.62107068e-01 1.74956188e-01 6.76588416e-01
1.08123612e+00 1.17581695e-01 6.61078691e-01 5.87840796e-01
5.95131576e-01 -9.97564554e-01 -3.27416927e-01 1.77103683e-01
6.91751242e-01 -1.33609164e+00 1.70404419e-01 -4.08693373e-01
-4.60143507e-01 1.07177556e+00 5.99609435e-01 4.80536759e-01
5.83600283e-01 4.44486290e-01 -1.38014928e-01 -4.94633734e-01
-8.30532491e-01 -1.62035957e-01 2.15579391e-01 3.79745960e-01
5.05944729e-01 -5.13110496e-02 -6.51469707e-01 7.22574592e-01
-5.15520871e-02 2.09035277e-02 2.99411744e-01 1.31460643e+00
-7.04268932e-01 -1.27752626e+00 -4.61395800e-01 8.40760946e-01
-4.27461833e-01 6.29418194e-02 -7.23961711e-01 7.75590181e-01
3.62481892e-01 8.59196365e-01 2.01049149e-01 -5.25817752e-01
6.19497001e-01 5.42334579e-02 -4.13924783e-01 -6.74951434e-01
-6.15886867e-01 -3.26192558e-01 7.46264830e-02 -3.86444002e-01
-2.42653951e-01 -6.85572922e-01 -1.29927683e+00 -4.74831194e-01
-8.27551559e-02 5.64953268e-01 7.49442756e-01 6.32862628e-01
7.51276433e-01 2.24921122e-01 6.63920045e-01 1.19890414e-01
-5.39180279e-01 -7.30600536e-01 -6.68500781e-01 7.05489755e-01
2.27589849e-02 -6.22902155e-01 1.50282353e-01 -1.28062382e-01] | [9.199874877929688, 7.899125576019287] |
af490730-29d0-4988-8949-aa867af63d1f | visual-question-answering-based-on-local | 2101.08978 | null | https://arxiv.org/abs/2101.08978v1 | https://arxiv.org/pdf/2101.08978v1.pdf | Visual Question Answering based on Local-Scene-Aware Referring Expression Generation | Visual question answering requires a deep understanding of both images and natural language. However, most methods mainly focus on visual concept; such as the relationships between various objects. The limited use of object categories combined with their relationships or simple question embedding is insufficient for representing complex scenes and explaining decisions. To address this limitation, we propose the use of text expressions generated for images, because such expressions have few structural constraints and can provide richer descriptions of images. The generated expressions can be incorporated with visual features and question embedding to obtain the question-relevant answer. A joint-embedding multi-head attention network is also proposed to model three different information modalities with co-attention. We quantitatively and qualitatively evaluated the proposed method on the VQA v2 dataset and compared it with state-of-the-art methods in terms of answer prediction. The quality of the generated expressions was also evaluated on the RefCOCO, RefCOCO+, and RefCOCOg datasets. Experimental results demonstrate the effectiveness of the proposed method and reveal that it outperformed all of the competing methods in terms of both quantitative and qualitative results. | ['Seong-Whan Lee', 'Hong-Gyu Jung', 'Jialin Wu', 'Dong-Gyu Lee', 'Jung-Jun Kim'] | 2021-01-22 | null | null | null | null | ['referring-expression-generation'] | ['computer-vision'] | [-2.09263489e-01 1.50327617e-02 5.23123294e-02 -6.03404403e-01
-5.51811278e-01 -4.67385203e-01 7.22360969e-01 1.68556303e-01
-4.29725736e-01 5.20014048e-01 4.47675407e-01 -1.26537085e-01
2.02021748e-01 -5.50005138e-01 -6.73461616e-01 -4.03823018e-01
5.14779508e-01 2.73748428e-01 2.74005979e-01 -2.00698435e-01
2.53753364e-01 1.41575590e-01 -1.76944482e+00 7.33707905e-01
8.07942271e-01 1.25612032e+00 2.10262328e-01 4.97706473e-01
-4.95002508e-01 1.48913276e+00 -5.32723725e-01 -7.92948306e-01
-1.79292232e-01 -6.47485852e-01 -9.22152638e-01 2.43975714e-01
6.57990873e-01 -4.73667532e-01 -3.67944181e-01 8.62557828e-01
3.73943001e-01 2.11943358e-01 7.68203795e-01 -1.37077975e+00
-1.26506376e+00 -1.16380841e-01 -3.87139320e-01 1.52719617e-01
7.63568461e-01 2.89704621e-01 1.38602638e+00 -1.25190568e+00
7.43190289e-01 1.49265850e+00 7.89672509e-02 6.05455637e-01
-1.08522677e+00 -4.72231209e-01 3.94970238e-01 8.74928832e-01
-1.17909348e+00 -4.18937862e-01 9.30626392e-01 -4.17304695e-01
9.00578856e-01 2.63419449e-01 4.51388925e-01 1.01224887e+00
-3.51354070e-02 1.09916544e+00 1.33616042e+00 -4.04979676e-01
8.11553895e-02 4.67781514e-01 2.80924797e-01 1.05984998e+00
-3.18735301e-01 -2.28194088e-01 -5.16523123e-01 2.24653110e-02
3.98610651e-01 6.05840981e-02 -5.28433621e-01 -3.38342786e-01
-1.26770461e+00 9.92278874e-01 1.08677578e+00 2.52821743e-01
-3.69174600e-01 2.25584418e-01 2.69536853e-01 5.91019318e-02
3.52380663e-01 4.00755256e-01 6.47854386e-03 3.12308013e-01
-4.64258820e-01 2.74419248e-01 5.34796834e-01 9.36816335e-01
8.90279531e-01 -9.08698142e-02 -6.75698936e-01 7.78843641e-01
4.63617951e-01 4.45242554e-01 1.82114780e-01 -1.13564718e+00
6.49017334e-01 9.84470963e-01 3.06239128e-01 -1.50157583e+00
-3.40691745e-01 -1.32970333e-01 -7.69917369e-01 9.64865014e-02
3.84254843e-01 2.46981040e-01 -8.98711264e-01 1.70078218e+00
2.77113974e-01 -3.69711608e-01 1.08732171e-01 1.39035547e+00
1.57353878e+00 8.47699463e-01 2.73774654e-01 8.86640921e-02
1.68239677e+00 -1.46034122e+00 -1.22119689e+00 -3.64250362e-01
3.42647821e-01 -4.97111320e-01 1.34292841e+00 -2.11686179e-01
-1.11790216e+00 -8.91732097e-01 -8.04486156e-01 -6.57138765e-01
-4.23806697e-01 2.57600367e-01 3.96111399e-01 1.75449386e-01
-1.03412032e+00 -2.14577779e-01 -2.37122551e-01 -2.22940132e-01
5.99464178e-01 -3.41331176e-02 -4.68634278e-01 -3.72384459e-01
-1.21520972e+00 9.55797315e-01 2.01478556e-01 2.98709065e-01
-7.80775368e-01 -3.27583760e-01 -1.06320906e+00 2.88363069e-01
2.83953816e-01 -1.02041960e+00 1.01650906e+00 -1.10262370e+00
-1.14764285e+00 9.92804229e-01 -5.76716542e-01 -1.95544243e-01
1.59294277e-01 -1.01897441e-01 -2.30246454e-01 8.00190449e-01
1.04170486e-01 1.14221013e+00 7.57407367e-01 -1.55052412e+00
-2.84792840e-01 -4.23952669e-01 6.69493318e-01 2.86761820e-01
-1.74300849e-01 3.78337763e-02 -6.06567740e-01 -2.70677567e-01
-2.83546802e-02 -7.02792525e-01 9.39196348e-02 3.98118556e-01
-2.73983002e-01 -4.35621709e-01 8.83500457e-01 -9.17807162e-01
9.13464427e-01 -1.90475118e+00 1.77611440e-01 -2.90987730e-01
5.19725800e-01 1.17738277e-01 -2.71863580e-01 5.01769483e-01
2.63941027e-02 4.84976284e-02 -2.37719435e-02 -4.23578590e-01
1.51830316e-01 2.12279275e-01 -2.64108390e-01 3.17608029e-01
4.08405393e-01 1.42719972e+00 -7.69196749e-01 -8.21378112e-01
2.33943433e-01 5.60385644e-01 -5.46496511e-01 5.44526994e-01
-5.48640013e-01 4.50872928e-01 -5.02652407e-01 7.44124889e-01
5.40830672e-01 -7.96987116e-01 -1.47279486e-01 -5.57882786e-01
3.44792813e-01 -6.11926727e-02 -4.93529320e-01 1.36372781e+00
-6.06240153e-01 8.39994848e-01 -6.56682774e-02 -8.65685344e-01
1.00835025e+00 3.18147570e-01 -1.26606792e-01 -1.08043623e+00
1.55199438e-01 -8.84754658e-02 -5.05657159e-02 -1.11533749e+00
2.83888400e-01 -2.03355491e-01 9.60844904e-02 3.25801671e-01
1.74101263e-01 -9.60072279e-02 1.28826901e-01 4.90139157e-01
6.02783918e-01 1.95227131e-01 2.86469102e-01 8.34642202e-02
8.06376994e-01 -1.42190769e-01 2.46043041e-01 5.26141405e-01
-4.20415163e-01 7.34849393e-01 6.58924997e-01 -2.54124016e-01
-8.28758776e-01 -8.25287700e-01 2.11790577e-01 1.06162393e+00
5.26570916e-01 -3.21945131e-01 -5.47691047e-01 -8.30815375e-01
-3.64335626e-02 8.88725638e-01 -9.53779459e-01 8.83193389e-02
-2.58781761e-01 -1.65743038e-01 1.59509510e-01 6.71183884e-01
7.51515150e-01 -1.49390340e+00 -6.73858106e-01 -2.58997083e-01
-6.64914012e-01 -1.45798981e+00 -3.76109898e-01 -3.74162853e-01
-3.88359576e-01 -1.27850628e+00 -7.78825879e-01 -9.74848270e-01
8.16610277e-01 3.30690175e-01 1.27072060e+00 3.60441208e-01
-2.49015633e-02 6.81116879e-01 -5.13455093e-01 -2.14304551e-01
-9.05665755e-02 -4.15399432e-01 -5.78351557e-01 4.92697179e-01
3.15435350e-01 -2.18371332e-01 -8.75653088e-01 2.10105270e-01
-9.37247038e-01 2.73095191e-01 5.34455955e-01 1.01951063e+00
3.27255815e-01 -6.12878025e-01 7.77320504e-01 -7.20875442e-01
7.02661872e-01 -3.89579773e-01 -2.01722458e-01 6.02709889e-01
-2.30614915e-01 1.85854480e-01 4.63391811e-01 -1.88387945e-01
-1.38548708e+00 -2.45311379e-01 -1.15020357e-01 -5.67150176e-01
-1.40065536e-01 4.26504642e-01 -2.45320231e-01 -1.11212749e-02
3.67517263e-01 1.06615819e-01 -5.70808761e-02 -2.97376007e-01
7.52218783e-01 5.61994195e-01 4.64196950e-01 -2.70083278e-01
3.97411704e-01 6.12703204e-01 -1.56563982e-01 -7.05319822e-01
-1.07894135e+00 -4.83143806e-01 -4.72929329e-01 -5.08194745e-01
1.33648241e+00 -7.64560759e-01 -1.06257868e+00 1.24942549e-01
-1.70037282e+00 1.10427514e-01 7.20759109e-02 1.13955021e-01
-5.26752174e-01 4.67248231e-01 -6.13396943e-01 -9.57141042e-01
-2.26096585e-01 -1.13440323e+00 1.08790970e+00 3.42762053e-01
-1.70887075e-02 -1.11003399e+00 -2.40654171e-01 1.04622722e+00
3.49571407e-01 4.57370907e-01 1.36340153e+00 -3.66960824e-01
-7.24180996e-01 2.11822111e-02 -8.61089468e-01 2.65250593e-01
-1.82361752e-01 -2.84669578e-01 -1.07380748e+00 2.68482454e-02
-3.64707746e-02 -8.35718572e-01 9.99547780e-01 -5.09477332e-02
1.41060102e+00 -4.22378361e-01 -1.06036380e-01 2.06640393e-01
1.48216581e+00 5.27524315e-02 6.69556081e-01 8.38144943e-02
7.70733297e-01 1.17422652e+00 5.89860260e-01 1.00585438e-01
8.62642169e-01 7.46452153e-01 9.06832635e-01 -2.81866401e-01
-3.19297880e-01 -2.30166748e-01 7.92943090e-02 8.42620015e-01
1.99985281e-02 -4.42873091e-01 -9.03592944e-01 7.78560579e-01
-2.00725031e+00 -9.52560723e-01 -3.86922836e-01 1.51023233e+00
6.05726540e-01 -3.35760206e-01 -2.06470639e-01 -1.59964994e-01
5.74687183e-01 3.48449260e-01 -4.82455432e-01 -3.86892945e-01
-3.32837075e-01 -3.15192007e-02 -2.12782919e-01 3.81273985e-01
-8.95683944e-01 8.04889500e-01 5.71909380e+00 3.90297204e-01
-7.02803552e-01 2.87568629e-01 6.91148460e-01 1.43216565e-01
-5.06537676e-01 -3.70732471e-02 -3.54523987e-01 1.75827011e-01
4.79966611e-01 1.59290090e-01 1.05950348e-01 5.00184536e-01
-1.57216731e-02 -2.13456407e-01 -1.10403919e+00 1.17805672e+00
6.01607323e-01 -1.34476197e+00 2.54672825e-01 -4.04826880e-01
7.12626338e-01 -4.83194411e-01 1.13179378e-01 4.80882257e-01
-1.87095761e-01 -1.20701087e+00 8.39049757e-01 9.27476585e-01
4.67740566e-01 -4.80615824e-01 9.76970851e-01 2.74691582e-01
-1.03340578e+00 -1.58706456e-01 -2.84833670e-01 -2.09307104e-01
3.15265626e-01 -1.94766298e-02 -5.70030749e-01 6.12432241e-01
8.24863613e-01 7.24206030e-01 -1.13002503e+00 6.28679454e-01
-5.67590237e-01 4.61826086e-01 2.56648481e-01 -4.05617774e-01
3.24972510e-01 -6.52189031e-02 1.57159105e-01 9.92560387e-01
-6.42726272e-02 3.42052817e-01 -8.89829025e-02 1.24782419e+00
-1.62673011e-01 3.22257400e-01 -4.04940009e-01 -1.62607636e-02
7.04859868e-02 1.25545597e+00 -3.73591900e-01 -4.94688779e-01
-8.01548719e-01 9.75202739e-01 5.93585670e-01 8.97770464e-01
-8.43272209e-01 -3.62385601e-01 2.41742790e-01 8.09456855e-02
3.64229977e-01 6.87019005e-02 -3.40203084e-02 -1.36122370e+00
2.15236291e-01 -9.30852950e-01 4.62900668e-01 -1.55355191e+00
-1.47377777e+00 8.35539162e-01 -1.71383079e-02 -1.01171350e+00
-1.64584711e-01 -7.34304786e-01 -4.10591841e-01 9.15004730e-01
-1.71708584e+00 -1.51777792e+00 -8.81014466e-01 6.17984712e-01
5.56676269e-01 6.58846498e-02 8.03827643e-01 1.74870491e-01
-2.96195835e-01 4.73062843e-01 -3.62866789e-01 1.38451904e-01
6.29113078e-01 -1.14500523e+00 -1.04916237e-01 6.03385448e-01
2.65800178e-01 4.43531841e-01 6.37051404e-01 -1.65322110e-01
-1.19776440e+00 -9.00899768e-01 9.68944430e-01 -4.94420528e-01
4.58094209e-01 -4.68714595e-01 -1.17994714e+00 5.53860545e-01
7.81443894e-01 2.22048968e-01 6.10941052e-01 -1.03409976e-01
-5.61508954e-01 5.00076786e-02 -9.59082603e-01 6.44435942e-01
7.46297836e-01 -8.19057524e-01 -8.29839289e-01 1.89501211e-01
9.80944991e-01 -1.08389251e-01 -6.40797734e-01 3.05048108e-01
4.39062476e-01 -1.22997963e+00 9.41979587e-01 -8.84315670e-01
8.28939676e-01 -3.16612512e-01 -3.54780108e-01 -1.00517857e+00
-3.13852936e-01 1.55521035e-01 -1.91457406e-01 1.33031714e+00
3.14269960e-01 -3.62490773e-01 5.32131493e-01 6.53654575e-01
2.18132257e-01 -8.68329287e-01 -6.61500692e-01 -1.65794149e-01
-2.17126384e-01 -1.48485273e-01 3.47827435e-01 7.94366777e-01
-1.95072100e-01 8.06142271e-01 -4.82397467e-01 2.76904870e-02
3.64331633e-01 5.56778491e-01 7.55494595e-01 -7.67605603e-01
-2.19014227e-01 -2.62284487e-01 -5.89512706e-01 -1.17576063e+00
5.22946060e-01 -8.83039355e-01 -1.22421771e-01 -2.06571937e+00
4.94363338e-01 9.38514620e-02 -2.47505844e-01 3.27735454e-01
-5.27135909e-01 3.27293307e-01 5.60464919e-01 8.33766460e-02
-1.10428524e+00 1.03004098e+00 1.64765871e+00 -3.70484293e-01
3.33153158e-01 -4.75583702e-01 -5.95159173e-01 7.24632859e-01
5.46540558e-01 -7.30221793e-02 -3.85006338e-01 -6.31852388e-01
3.18164557e-01 2.72276938e-01 8.96672010e-01 -6.56778276e-01
8.97103921e-02 3.31601351e-02 4.39393878e-01 -7.25322247e-01
6.44708335e-01 -8.09976757e-01 -4.13454324e-01 1.59201339e-01
-6.39366090e-01 1.94804356e-01 7.58987293e-02 7.85591245e-01
-6.63093686e-01 -1.74633354e-01 5.30188143e-01 -4.13754582e-02
-9.74535108e-01 2.32589208e-02 5.31174149e-03 3.07384342e-01
9.06862140e-01 -6.26417026e-02 -5.69204152e-01 -1.05314469e+00
-6.34383380e-01 6.10937476e-01 2.30060622e-01 6.17757559e-01
1.16792428e+00 -1.61582875e+00 -7.23630190e-01 -1.46460190e-01
7.52153218e-01 -2.67525882e-01 6.03616238e-01 8.97317231e-01
-5.63124180e-01 7.30272651e-01 -2.27133006e-01 -6.53905392e-01
-1.31813538e+00 8.57169509e-01 4.14720982e-01 -1.21905893e-01
-2.85260797e-01 7.76337802e-01 8.75067413e-01 -4.36961174e-01
2.03460574e-01 -3.79575849e-01 -6.78979039e-01 6.46884292e-02
5.45137525e-01 -2.86102924e-03 -4.15865660e-01 -9.66449738e-01
-4.56136584e-01 6.84428513e-01 1.63538381e-01 -1.09403826e-01
9.93241251e-01 -3.93441200e-01 -1.67904139e-01 5.20658731e-01
1.40982676e+00 -3.28491956e-01 -1.15484071e+00 -4.10118610e-01
-1.51212007e-01 -5.87857187e-01 -1.45775914e-01 -8.23519409e-01
-1.03988421e+00 1.53386843e+00 5.16426563e-01 1.29292607e-02
1.20075047e+00 3.33047777e-01 5.99726319e-01 3.65177542e-01
1.28845587e-01 -5.88101983e-01 7.57804513e-01 2.13020951e-01
1.35145164e+00 -1.70256770e+00 -1.98883384e-01 -4.44580436e-01
-1.01434410e+00 9.79489207e-01 8.80614519e-01 9.44408104e-02
3.65450442e-01 -6.54571414e-01 2.53657371e-01 -4.89210665e-01
-1.02210748e+00 -4.85854208e-01 7.45123327e-01 4.53855336e-01
4.86078262e-01 -1.75718695e-01 -2.47850135e-01 5.92052042e-01
2.34800726e-01 -2.67315000e-01 2.84732193e-01 6.00734949e-01
-2.89208114e-01 -6.72458231e-01 -2.40724206e-01 2.36752301e-01
-3.69133025e-01 -1.07905651e-02 -5.90908825e-01 8.95566344e-01
4.48338427e-02 1.29802310e+00 1.28731038e-02 -1.91758156e-01
3.06520730e-01 3.67588587e-02 5.29194951e-01 -4.07440215e-01
-4.62822974e-01 -3.39691103e-01 1.44176453e-01 -6.21252120e-01
-8.29654932e-01 -2.88827032e-01 -1.13585019e+00 8.00603479e-02
-2.69625992e-01 6.21289350e-02 4.65516299e-01 9.85367000e-01
3.34458023e-01 6.26582086e-01 3.39862943e-01 -4.21467602e-01
-1.91086054e-01 -9.48167205e-01 -2.23418728e-01 9.25068378e-01
3.85424525e-01 -7.70378590e-01 -3.72550517e-01 1.99281201e-01] | [10.715779304504395, 1.71048903465271] |
6974f377-aaf8-4197-8173-5d238560f9a6 | classify-or-select-neural-architectures-for | 1611.04244 | null | http://arxiv.org/abs/1611.04244v1 | http://arxiv.org/pdf/1611.04244v1.pdf | Classify or Select: Neural Architectures for Extractive Document Summarization | We present two novel and contrasting Recurrent Neural Network (RNN) based
architectures for extractive summarization of documents. The Classifier based
architecture sequentially accepts or rejects each sentence in the original
document order for its membership in the final summary. The Selector
architecture, on the other hand, is free to pick one sentence at a time in any
arbitrary order to piece together the summary. Our models under both
architectures jointly capture the notions of salience and redundancy of
sentences. In addition, these models have the advantage of being very
interpretable, since they allow visualization of their predictions broken up by
abstract features such as information content, salience and redundancy. We show
that our models reach or outperform state-of-the-art supervised models on two
different corpora. We also recommend the conditions under which one
architecture is superior to the other based on experimental evidence. | ['Bo-Wen Zhou', 'Ramesh Nallapati', 'Mingbo Ma'] | 2016-11-14 | null | null | null | null | ['extractive-document-summarization'] | ['natural-language-processing'] | [ 5.91771305e-01 4.62680876e-01 -2.84253478e-01 -3.60096574e-01
-5.91584027e-01 -5.84789872e-01 8.18893790e-01 6.77154839e-01
-4.02101845e-01 5.89600265e-01 9.20494378e-01 -4.93887812e-01
-2.57545233e-01 -4.67584461e-01 -3.24942708e-01 -5.35987079e-01
-1.27958506e-01 4.69469130e-01 -1.17621897e-02 -2.96140552e-01
7.31171131e-01 3.92916858e-01 -1.59433770e+00 6.51062012e-01
7.92081475e-01 5.71589589e-01 3.88869345e-01 7.69360781e-01
-3.51349473e-01 9.77306783e-01 -8.53610933e-01 -3.72475743e-01
-9.03304741e-02 -6.14594340e-01 -9.13853884e-01 -8.11766088e-02
4.52867389e-01 -2.43200868e-01 -3.02890867e-01 8.84836376e-01
3.84602219e-01 1.18277043e-01 8.61763656e-01 -7.68719494e-01
-7.00400233e-01 1.20712984e+00 -2.74454057e-01 3.64110440e-01
5.09131908e-01 -2.75737166e-01 1.49799025e+00 -8.89538586e-01
8.27622533e-01 9.57623541e-01 3.79483402e-01 5.56498706e-01
-1.25321436e+00 -6.69132099e-02 5.42741120e-01 4.35329564e-02
-7.80081153e-01 -7.97726274e-01 8.26903403e-01 -2.11147740e-01
1.32503569e+00 7.36199915e-01 5.94441473e-01 9.63787496e-01
5.37726223e-01 1.01590836e+00 5.90208232e-01 -6.08141303e-01
2.24555343e-01 -1.25356972e-01 5.60146749e-01 4.38208729e-01
5.29653966e-01 -2.89944381e-01 -5.94952464e-01 -1.69635743e-01
3.38061392e-01 -1.08296722e-01 -4.01710153e-01 -9.20365453e-02
-1.15083146e+00 6.61830604e-01 1.87637955e-01 4.22260702e-01
-4.98267144e-01 2.10357457e-02 5.68990827e-01 2.66432643e-01
5.64492106e-01 8.33294332e-01 -2.78885037e-01 2.13940740e-02
-1.37932467e+00 3.64340901e-01 8.37433040e-01 7.79783905e-01
2.48404220e-01 1.18740171e-01 -5.42135239e-01 7.08950877e-01
6.40438497e-02 -3.93065140e-02 6.95814848e-01 -8.69564652e-01
6.80118144e-01 5.38869679e-01 1.28679842e-01 -9.63517249e-01
-5.76482952e-01 -6.56034052e-01 -9.35335517e-01 7.51836151e-02
5.85838221e-03 2.52965055e-02 -7.84025848e-01 1.54764688e+00
-4.52569723e-01 -2.37075180e-01 3.54812175e-01 6.67073488e-01
1.22797585e+00 7.56644189e-01 -9.61148664e-02 -6.96470559e-01
1.29613256e+00 -8.17413628e-01 -9.03045475e-01 -2.21779808e-01
5.04224062e-01 -5.62708855e-01 7.96059132e-01 4.08717692e-01
-1.51174867e+00 -2.84492731e-01 -1.37933159e+00 -2.72178978e-01
-1.60521850e-01 3.73757303e-01 4.63461041e-01 2.23235279e-01
-1.21828008e+00 9.64970827e-01 -7.72720158e-01 -2.65701622e-01
8.95555839e-02 3.26558292e-01 -4.88902777e-02 6.15593255e-01
-1.05679607e+00 1.09357083e+00 5.26873708e-01 2.90232718e-01
-4.51803535e-01 -1.09470986e-01 -7.47439444e-01 5.96148431e-01
1.29965529e-01 -9.02334154e-01 1.48490775e+00 -1.02825296e+00
-1.40100598e+00 7.28968620e-01 -3.89231324e-01 -7.88230419e-01
3.26249957e-01 -3.40398461e-01 -2.65846521e-01 2.99964011e-01
2.31129248e-02 5.14928937e-01 6.29987299e-01 -1.12922645e+00
-5.16543508e-01 -2.25934416e-01 7.73325041e-02 4.56128478e-01
-2.98220158e-01 2.35322699e-01 -1.55747324e-01 -8.17551732e-01
2.67199248e-01 -6.52403057e-01 -2.42396861e-01 -6.22152805e-01
-8.54493201e-01 -2.84011155e-01 6.26021147e-01 -6.65020466e-01
1.65749633e+00 -1.72222435e+00 4.38979506e-01 -1.73320204e-01
2.59925932e-01 1.08737059e-01 -1.62998721e-01 7.52927244e-01
-2.48728544e-01 5.61248124e-01 -1.82372376e-01 -5.76800704e-01
-6.92924485e-02 -4.40469980e-02 -6.85787797e-01 2.12803960e-01
3.04639667e-01 8.29099238e-01 -8.94273996e-01 -4.61402953e-01
-9.87314358e-02 2.10662875e-02 -3.46835405e-01 7.63631463e-02
-3.15098256e-01 7.31636584e-02 -3.97389054e-01 3.12600248e-02
2.55526692e-01 -2.40437895e-01 4.89756048e-01 -7.72214234e-02
-2.23458052e-01 1.07505620e+00 -9.63421643e-01 1.47229433e+00
-1.30013585e-01 1.02903795e+00 -1.74001306e-01 -9.44436133e-01
8.79145026e-01 5.30137300e-01 -7.74384439e-02 -3.11170995e-01
2.51786392e-02 3.57462093e-02 1.44269139e-01 -3.11456531e-01
1.36740220e+00 -5.55302911e-02 -1.75360933e-01 9.10962462e-01
4.24368531e-02 1.42273739e-01 5.73358417e-01 6.93020761e-01
1.00703108e+00 5.47819659e-02 5.78993380e-01 -3.09683621e-01
3.91120136e-01 -3.30129206e-01 4.82016921e-01 1.11693490e+00
2.39315614e-01 8.10309827e-01 8.63318503e-01 -3.66992265e-01
-1.11286008e+00 -9.50434446e-01 2.30965018e-01 1.21598649e+00
-1.13644406e-01 -7.92015910e-01 -5.61531544e-01 -5.61225712e-01
-4.20396239e-01 1.21514583e+00 -6.39377892e-01 2.51894388e-02
-7.78857946e-01 -2.64462203e-01 4.39091772e-01 6.32783592e-01
-3.08903549e-02 -1.34806514e+00 -9.64242578e-01 1.80121258e-01
-4.29520518e-01 -5.77748358e-01 -3.22785199e-01 4.80063945e-01
-1.08794248e+00 -7.25224137e-01 -4.76898283e-01 -6.83232665e-01
8.02419126e-01 2.93802947e-01 1.14987743e+00 2.10329846e-01
3.40872318e-01 5.81462122e-02 -4.16828752e-01 -4.17443842e-01
-6.98465765e-01 4.79336858e-01 8.87266695e-02 -3.45098883e-01
5.27285412e-02 -5.46246588e-01 -4.04007047e-01 -2.40255281e-01
-1.07160771e+00 3.37944567e-01 5.11272550e-01 8.35297525e-01
2.73506403e-01 -7.59325773e-02 6.75725639e-01 -1.05748951e+00
1.30996180e+00 -2.31073514e-01 5.06301709e-02 3.70474637e-01
-4.87165838e-01 3.73570532e-01 8.01491261e-01 -8.24263096e-02
-1.10755622e+00 -2.01565593e-01 1.20567881e-01 -3.59283686e-02
4.68599098e-03 9.98671234e-01 1.11723728e-01 8.68492007e-01
6.18117750e-01 5.05366802e-01 -1.82449594e-01 -5.64559698e-01
3.89143109e-01 5.92852354e-01 7.39075541e-01 -2.80914545e-01
3.59152347e-01 1.81580842e-01 -3.17298383e-01 -8.94654512e-01
-8.55943739e-01 -2.66818315e-01 -8.66587102e-01 -7.42495880e-02
5.70942879e-01 -6.14491165e-01 -3.80621821e-01 -2.72836462e-02
-1.64416075e+00 2.90852070e-01 -3.72514218e-01 2.46915117e-01
-5.94132304e-01 5.55147350e-01 -7.45247900e-01 -9.96532679e-01
-6.84186876e-01 -8.09104443e-01 7.89966643e-01 2.25861758e-01
-8.49166155e-01 -8.21948349e-01 -1.69026107e-01 -6.11000322e-03
1.92981228e-01 8.90812129e-02 1.20412719e+00 -1.27127612e+00
-2.06864998e-01 -1.50962994e-01 1.14761867e-01 9.71121043e-02
6.37138262e-02 3.08306962e-01 -6.23339832e-01 -2.44356856e-01
1.68148398e-01 -1.31088883e-01 1.40447235e+00 5.09451747e-01
9.34385300e-01 -7.63249099e-01 -2.39529490e-01 -1.53574941e-03
9.52776194e-01 5.48608676e-02 6.90956593e-01 3.17001671e-01
2.92378545e-01 7.48233557e-01 2.87608922e-01 3.32801133e-01
1.52469158e-01 3.63090307e-01 3.69883299e-01 1.60481393e-01
9.89778638e-02 -3.94048005e-01 4.97219622e-01 1.07977235e+00
-9.34542045e-02 -6.36562228e-01 -5.87670743e-01 5.49756646e-01
-2.17058992e+00 -1.34716308e+00 -1.29629061e-01 2.14784074e+00
7.20055640e-01 6.60330296e-01 1.07899293e-01 1.66769683e-01
7.24578381e-01 6.27856255e-01 -2.61630863e-01 -9.13603485e-01
-3.16090405e-01 -7.83754662e-02 9.83806849e-02 3.81894469e-01
-9.46948767e-01 7.91633785e-01 7.09866667e+00 5.18557489e-01
-9.25613225e-01 -3.53413194e-01 7.19774306e-01 -4.11061555e-01
-6.64952278e-01 1.33121297e-01 -7.28507876e-01 2.75264293e-01
1.08793986e+00 -2.58948326e-01 8.54448527e-02 5.83687127e-01
4.17610526e-01 -2.34321415e-01 -1.34421086e+00 4.47671562e-01
3.30512136e-01 -1.80909121e+00 5.40514350e-01 -2.00257808e-01
4.02088076e-01 -1.38685197e-01 -7.75580257e-02 2.30701943e-03
3.82135332e-01 -1.03404582e+00 9.89711642e-01 6.95059180e-01
5.29478073e-01 -7.88846195e-01 7.02332735e-01 5.80988884e-01
-7.25709915e-01 9.31360014e-03 -2.92422146e-01 -3.90896887e-01
1.50895014e-01 2.48687297e-01 -6.37534022e-01 7.94055343e-01
2.91319221e-01 7.26480842e-01 -6.57985985e-01 7.28039265e-01
-4.01063085e-01 5.34356356e-01 1.06405750e-01 -3.87865126e-01
3.02924842e-01 -1.08550496e-01 9.75311816e-01 1.58562064e+00
2.06651464e-01 5.42378128e-02 -6.78750649e-02 7.81777501e-01
-8.63729417e-02 2.80046556e-02 -6.52325869e-01 -4.10801508e-02
6.24933481e-01 1.09279239e+00 -8.92505944e-01 -5.77213168e-01
5.78467920e-02 7.72926033e-01 4.47861165e-01 3.65350693e-01
-2.54000664e-01 -6.25871241e-01 1.73177749e-01 -1.06119916e-01
5.48766255e-01 -2.05842838e-01 -5.16418159e-01 -1.32653308e+00
2.07268566e-01 -6.81977749e-01 3.54899049e-01 -7.66376078e-01
-9.68758702e-01 8.53113234e-01 -5.94345555e-02 -1.09320700e+00
-6.76700175e-01 -2.79589742e-01 -9.48144913e-01 7.07237184e-01
-1.07789969e+00 -7.81067431e-01 3.21161181e-01 -1.49754465e-01
7.78298259e-01 -2.16927201e-01 9.83384550e-01 -5.29771507e-01
-7.10209846e-01 2.88371176e-01 3.02653939e-01 1.54027015e-01
3.40092599e-01 -1.32261610e+00 6.10778511e-01 8.91302466e-01
3.89697075e-01 1.07605612e+00 1.20399666e+00 -5.71899056e-01
-1.00997138e+00 -7.23091304e-01 1.36391032e+00 -3.58581811e-01
3.89255017e-01 -4.66465771e-01 -8.93492997e-01 8.42443049e-01
6.62178636e-01 -8.01938772e-01 7.67350972e-01 4.54486310e-01
-2.48930633e-01 1.50178224e-01 -6.21980011e-01 9.83485758e-01
7.31644988e-01 -3.04658115e-01 -1.31487119e+00 1.45308524e-01
8.19363117e-01 -2.22359106e-01 -3.48945498e-01 2.56467104e-01
5.99525452e-01 -1.27098322e+00 5.39252222e-01 -9.01000142e-01
1.03686023e+00 -2.34552726e-01 7.44073763e-02 -1.41413450e+00
-5.94702244e-01 -8.40014100e-01 -3.28802466e-01 1.19970059e+00
7.81347394e-01 -3.83609772e-01 6.89916253e-01 4.05034453e-01
-2.45151699e-01 -9.04480100e-01 -8.54369044e-01 -5.45257092e-01
-1.64642930e-02 -3.44719827e-01 4.26149786e-01 5.23154914e-01
6.03301942e-01 9.39095557e-01 -3.50865841e-01 -3.73095125e-01
3.67369018e-02 4.67292428e-01 4.30247575e-01 -1.34684300e+00
-1.16972059e-01 -9.35561121e-01 -1.38263926e-02 -1.27591503e+00
2.31183663e-01 -9.97102976e-01 -4.67613302e-02 -1.99385786e+00
3.78934711e-01 1.79738089e-01 -4.44121242e-01 3.92927140e-01
-1.58591613e-01 -1.93947136e-01 4.40260500e-01 5.47660232e-01
-7.88550138e-01 5.47448575e-01 7.24451005e-01 -1.75054073e-01
-5.08538723e-01 1.00940987e-01 -1.25049365e+00 9.84302640e-01
8.90554965e-01 -4.21369493e-01 -3.05315852e-01 -3.94957811e-01
3.40263963e-01 2.67551810e-01 6.96632415e-02 -7.03093827e-01
3.87444347e-01 -1.11559797e-02 5.09878337e-01 -1.02236986e+00
2.22650871e-01 -3.18026811e-01 -1.53517589e-01 4.29466486e-01
-1.15933537e+00 2.41672426e-01 1.23386934e-01 6.06770515e-01
-1.64758638e-02 -6.26699209e-01 3.40960860e-01 -2.66951978e-01
-1.67503998e-01 -3.79644245e-01 -8.22171390e-01 -1.77352890e-01
2.94931978e-01 -2.99429148e-01 -4.38306212e-01 -7.11491764e-01
-5.53220451e-01 -8.88625334e-04 2.69628644e-01 4.49708015e-01
7.41132319e-01 -9.86902416e-01 -9.37009275e-01 -1.07696623e-01
-2.78785564e-02 -2.56245106e-01 -9.00494307e-02 4.81192559e-01
-3.24551344e-01 7.79434323e-01 -6.66430295e-02 -2.66620666e-01
-1.41605723e+00 5.25693119e-01 -5.21336496e-02 -6.68487012e-01
-7.01157272e-01 3.13526034e-01 1.04304710e-02 -2.42371224e-02
2.03261048e-01 -4.46843237e-01 -5.94373167e-01 2.35333741e-01
7.07671642e-01 3.00982088e-01 1.64280817e-01 -5.74581385e-01
-1.94674253e-01 3.31534049e-03 -5.99784732e-01 -2.70449162e-01
1.46098685e+00 -1.52412727e-01 -3.37181151e-01 7.46317148e-01
7.49723613e-01 1.09000161e-01 -8.55895460e-01 -1.84918046e-01
4.82341260e-01 -9.19560844e-04 -2.54736900e-01 -9.37503219e-01
-4.51039135e-01 6.52940392e-01 -4.22641248e-01 6.78306878e-01
1.05634773e+00 -1.13914073e-01 3.98165673e-01 6.57480061e-01
-2.28049845e-01 -1.13922060e+00 -2.77524199e-02 7.75092244e-01
1.22626555e+00 -7.14418530e-01 4.62247550e-01 -9.86531898e-02
-8.17637205e-01 1.48232126e+00 2.55464673e-01 -2.61665165e-01
1.69900653e-03 8.13086927e-02 -1.78883478e-01 -2.03113869e-01
-1.39182556e+00 1.07617117e-01 3.82989645e-01 1.84924722e-01
8.54775310e-01 3.19293626e-02 -7.48018086e-01 6.69470191e-01
-5.37418306e-01 -4.29510772e-01 1.01539874e+00 9.41950262e-01
-7.53098071e-01 -8.56499910e-01 -1.95713162e-01 8.04548383e-01
-6.17553592e-01 -2.63413996e-01 -8.01943600e-01 5.35809755e-01
-4.86200362e-01 1.13487220e+00 2.16156140e-01 -3.12519312e-01
1.74278960e-01 2.44463891e-01 2.70990580e-01 -8.52984250e-01
-8.83334339e-01 1.97282359e-01 5.23404181e-01 7.99297318e-02
-3.39275420e-01 -7.15929031e-01 -1.25700605e+00 -7.63144121e-02
-4.52156216e-01 4.09040749e-01 5.57379067e-01 9.36935246e-01
6.08307123e-01 6.16609633e-01 4.43545491e-01 -1.00513947e+00
-7.65545547e-01 -1.19257474e+00 -4.35043424e-01 1.20423049e-01
4.75939661e-01 -1.33692786e-01 -2.57159650e-01 6.61655739e-02] | [12.447826385498047, 9.473484992980957] |
87914cb4-af05-4938-a7b2-dc64f4c51296 | semantics-enhanced-task-oriented-dialogue | null | null | https://aclanthology.org/I17-3009 | https://aclanthology.org/I17-3009.pdf | Semantics-Enhanced Task-Oriented Dialogue Translation: A Case Study on Hotel Booking | We showcase TODAY, a semantics-enhanced task-oriented dialogue translation system, whose novelties are: (i) task-oriented named entity (NE) definition and a hybrid strategy for NE recognition and translation; and (ii) a novel grounded semantic method for dialogue understanding and task-order management. TODAY is a case-study demo which can efficiently and accurately assist customers and agents in different languages to reach an agreement in a dialogue for the hotel booking. | ['Long-Yue Wang', 'Qun Liu', 'Zhaopeng Tu', 'Liangyou Li', 'Andy Way', 'Jinhua Du'] | 2017-11-01 | semantics-enhanced-task-oriented-dialogue-1 | https://aclanthology.org/I17-3009 | https://aclanthology.org/I17-3009.pdf | ijcnlp-2017-11 | ['dialogue-understanding'] | ['natural-language-processing'] | [-4.01828066e-02 6.11314058e-01 2.66772121e-01 -6.52357101e-01
-8.69695485e-01 -8.10051441e-01 1.04857266e+00 6.22956865e-02
-5.79778612e-01 1.41869009e+00 5.91950297e-01 -4.10315394e-01
-1.09412640e-01 -3.37473661e-01 2.26273537e-01 1.33150578e-01
2.76304960e-01 1.61931169e+00 1.56178847e-01 -1.33285964e+00
3.86034220e-01 4.88946527e-01 -9.66282904e-01 4.88916844e-01
1.15848005e+00 8.12392473e-01 3.83775741e-01 7.50460804e-01
-9.34695959e-01 1.10731125e+00 -9.00739491e-01 -6.16378546e-01
-1.02034369e-02 -5.48476279e-01 -1.78594398e+00 1.24476589e-01
-3.95846844e-01 -5.82851619e-02 2.19132960e-01 9.09427881e-01
4.41282660e-01 3.25623900e-01 4.94295567e-01 -1.36573553e+00
-3.85276824e-01 4.22563404e-01 7.12062299e-01 -2.94099301e-01
1.16154873e+00 -5.57353310e-02 6.34078681e-01 -8.02585363e-01
1.32828772e+00 1.23173189e+00 4.59164739e-01 8.79061401e-01
-9.64976847e-01 -1.14497311e-01 -3.96397203e-01 6.89142719e-02
-1.23203254e+00 -8.70315313e-01 2.80571461e-01 -1.09700382e-01
1.54707861e+00 6.10150278e-01 5.72981238e-01 7.98867881e-01
1.00945689e-01 4.74695444e-01 1.50729752e+00 -6.54449046e-01
2.39490271e-01 8.75615597e-01 -4.18064557e-02 4.63024855e-01
-3.97826314e-01 -2.50564605e-01 -9.23609555e-01 -3.15199733e-01
5.01241446e-01 -7.32036650e-01 9.71733108e-02 -8.90327916e-02
-1.43923628e+00 9.84340370e-01 -3.59072924e-01 7.23611653e-01
-5.96348703e-01 -5.18170297e-01 9.91215646e-01 8.34816933e-01
5.71308076e-01 1.09094965e+00 -9.21878338e-01 -8.07311296e-01
-3.56412381e-01 4.00071770e-01 2.20598912e+00 1.50551403e+00
4.66638416e-01 -3.65931839e-02 -3.65867943e-01 1.08635104e+00
2.66299933e-01 4.26179558e-01 3.47957700e-01 -1.07590675e+00
4.51549768e-01 7.83113658e-01 8.74407768e-01 -6.28888547e-01
-5.90191185e-01 3.10126334e-01 -2.23584369e-01 -3.32995862e-01
3.97522658e-01 -5.23755252e-01 -3.52312416e-01 1.27127862e+00
2.73361564e-01 -7.68963337e-01 7.82958984e-01 9.19606924e-01
1.11446393e+00 3.10700536e-01 3.43689770e-01 -2.85765231e-01
1.80685949e+00 -1.20349157e+00 -1.31517959e+00 -2.86355346e-01
1.21360111e+00 -9.76322174e-01 7.08586514e-01 -3.86546738e-02
-1.17657554e+00 -5.65136559e-02 -5.51057160e-01 -3.67833644e-01
-9.01232183e-01 -1.53404847e-01 9.01260316e-01 5.73619127e-01
-1.34553826e+00 3.42672914e-02 1.10913955e-01 -1.10871089e+00
-5.31959534e-01 4.84168679e-01 -5.46889722e-01 4.53669667e-01
-1.71914947e+00 1.51346934e+00 6.97031260e-01 1.23449534e-01
-3.82549524e-01 -2.80789405e-01 -9.52166855e-01 -1.19143523e-01
5.08488655e-01 -8.86844575e-01 1.80370867e+00 -9.69383001e-01
-2.26077127e+00 1.43135023e+00 -2.11344108e-01 -3.73487800e-01
5.77524245e-01 -1.16136111e-01 -9.47207153e-01 -1.09902574e-02
6.66671097e-01 5.90901971e-01 1.38859302e-01 -1.20834172e+00
-1.19656527e+00 -4.44630295e-01 -2.35498324e-02 8.51256847e-01
2.19663501e-01 8.33290398e-01 5.58349676e-02 -2.64712065e-01
-1.71000361e-01 -5.91248810e-01 -2.28927612e-01 -4.46965307e-01
-4.83414173e-01 -5.00810802e-01 2.88712353e-01 -1.14898264e+00
9.19390023e-01 -1.48833668e+00 2.22128615e-01 -1.02708161e-01
2.62142986e-01 4.55640703e-01 -6.83602020e-02 1.28012085e+00
4.80016708e-01 2.11244941e-01 -1.97564508e-03 -2.22447529e-01
4.98065233e-01 4.03878361e-01 -1.61627874e-01 -3.53311777e-01
-7.94122741e-02 1.12696970e+00 -1.06687605e+00 -7.41071761e-01
3.82149249e-01 -1.08375937e-01 8.16875324e-02 4.97228235e-01
-2.53356546e-01 4.06233579e-01 -6.77726448e-01 3.58718634e-01
1.87505364e-01 1.94450319e-01 5.89205444e-01 7.07413852e-02
-5.96191227e-01 6.55094087e-01 -8.14430416e-01 2.03381896e+00
-8.81098747e-01 5.52012324e-01 6.79471314e-01 -3.91713947e-01
1.28348672e+00 7.89140522e-01 2.08434775e-01 -1.04656732e+00
1.38986304e-01 6.89631045e-01 -4.60556358e-01 -6.92771494e-01
1.16928065e+00 -2.56985635e-01 -5.62853575e-01 3.69502217e-01
7.28293806e-02 -4.85375762e-01 2.36189231e-01 1.51276350e-01
6.33555353e-01 7.97946081e-02 4.98086840e-01 -6.29307985e-01
8.32100630e-01 9.98195469e-01 2.27060810e-01 5.77525616e-01
-3.64948541e-01 -2.31209144e-01 2.50931263e-01 -4.15104419e-01
-1.08097291e+00 -4.96967793e-01 2.07258523e-01 1.49792099e+00
8.04383457e-02 -2.90586829e-01 -9.98315334e-01 -7.66576469e-01
-5.67084730e-01 1.44729984e+00 -1.51716983e-02 2.95773804e-01
-5.11558056e-01 2.35838428e-01 8.15458477e-01 2.21297309e-01
9.04892504e-01 -1.14548743e+00 -4.53837633e-01 7.07553208e-01
-8.03001106e-01 -1.30286646e+00 -6.11124098e-01 1.59350768e-01
-5.62602937e-01 -8.77949238e-01 -5.46376467e-01 -1.21525514e+00
2.93741196e-01 -2.00731277e-01 1.22571480e+00 -2.98552960e-01
2.81110317e-01 6.47295892e-01 -4.36311483e-01 -5.26718378e-01
-1.18628526e+00 3.67153883e-01 -8.31831023e-02 -4.86018896e-01
7.08048582e-01 2.93907225e-02 1.26403958e-01 7.24117100e-01
-3.15267265e-01 3.96859616e-01 6.34877384e-02 8.99228752e-01
-4.06338543e-01 -6.18242025e-01 8.40033829e-01 -8.38874400e-01
1.48766840e+00 -9.74830091e-02 -2.51020074e-01 7.00717628e-01
-7.29459763e-01 -6.30615130e-02 5.08565366e-01 2.29268149e-01
-1.65711677e+00 -3.39701027e-01 -3.80597234e-01 6.35683537e-01
-5.44936359e-01 3.98094177e-01 -1.26488701e-01 -3.56530212e-02
5.64621329e-01 5.87592900e-01 2.91645020e-01 -4.99039859e-01
4.02577043e-01 1.13182497e+00 4.90631014e-01 -6.60934508e-01
2.89212465e-01 3.48269008e-02 -3.99979591e-01 -8.78236592e-01
-2.05698237e-01 -7.87817180e-01 -8.58482063e-01 -2.45658621e-01
9.71371949e-01 -7.96010315e-01 -9.84966218e-01 2.72727638e-01
-1.69670641e+00 -5.05601943e-01 -5.89007318e-01 3.19014996e-01
-9.97585952e-01 -3.16053219e-02 -7.24836588e-01 -9.71109688e-01
-5.51459253e-01 -6.78777874e-01 1.00800955e+00 1.11180164e-01
-6.41690552e-01 -1.38663578e+00 1.38593301e-01 7.10386217e-01
5.77002466e-01 -2.52873749e-01 9.80764687e-01 -1.68620169e+00
3.78083467e-01 1.49245486e-01 -4.45444733e-01 4.65163998e-02
6.27324283e-02 -6.40093625e-01 -4.65435356e-01 1.87399983e-01
1.26617730e-01 -4.66921479e-01 -4.33795452e-01 -4.21108723e-01
-5.41847885e-01 -6.82661712e-01 -1.86515391e-01 -7.26043731e-02
8.55926275e-01 8.23479950e-01 5.73838711e-01 7.01579094e-01
2.21754357e-01 1.35365617e+00 1.06669462e+00 4.11108166e-01
8.65685463e-01 1.20395339e+00 -4.74728823e-01 -2.73191184e-01
6.89458773e-02 -2.90059775e-01 2.76945978e-01 1.07090008e+00
-7.54833892e-02 -1.10263117e-01 -1.10253358e+00 4.72231001e-01
-2.10138106e+00 -5.59301615e-01 -1.78980991e-01 1.74980736e+00
8.49727333e-01 -4.35142040e-01 1.16473727e-01 -6.97097242e-01
9.41619575e-01 -1.10128172e-01 -1.36150541e-02 -1.19442356e+00
-1.12406649e-01 -1.28922701e-01 3.91822577e-01 5.29371083e-01
-5.93290329e-01 1.62243068e+00 6.82038498e+00 8.25318158e-01
-6.28390610e-01 4.37817723e-01 1.38698578e-01 7.54770458e-01
-1.08994484e-01 -2.71287169e-02 -7.53979862e-01 -9.73004103e-02
1.14718163e+00 -7.73172855e-01 6.73595130e-01 7.08328545e-01
2.44442597e-01 -2.10579652e-02 -1.02047527e+00 8.40008557e-01
1.62324294e-01 -1.35014403e+00 -8.82408246e-02 -3.75152826e-02
3.00246477e-01 -1.47103012e-01 -7.47141063e-01 6.28615439e-01
6.79268837e-01 -8.36393654e-01 8.26852143e-01 5.83148301e-01
7.49274015e-01 -6.32953405e-01 1.05993211e+00 5.10461867e-01
-9.89370048e-01 1.00610681e-01 -1.88196544e-02 6.58761039e-02
7.33625948e-01 -4.47515845e-01 -1.29210424e+00 1.07126606e+00
1.74419403e-01 1.39942005e-01 -5.88218756e-02 4.58408028e-01
-1.73669100e-01 -1.26329973e-01 -7.35938177e-02 -5.10038435e-01
4.91490841e-01 -5.93848109e-01 1.03778720e+00 1.32603276e+00
1.36549650e-02 3.67985517e-01 3.03343952e-01 7.00668991e-01
1.06228210e-01 5.60703099e-01 -6.71327889e-01 -3.15956384e-01
4.42922145e-01 1.16778648e+00 -7.07170963e-01 -4.30272967e-01
-1.29641876e-01 1.47310889e+00 -1.77504078e-01 2.87683696e-01
-3.03196520e-01 -6.66775167e-01 5.52982748e-01 -2.46441245e-01
-1.85337231e-01 -3.48760515e-01 -3.11124474e-02 -9.29285526e-01
-2.84598678e-01 -1.10767388e+00 1.69871524e-01 -8.70795250e-01
-1.05291450e+00 9.83091652e-01 1.03836032e-02 -8.62308323e-01
-9.31907475e-01 -5.40651917e-01 -1.64250180e-01 9.00910676e-01
-1.14851320e+00 -1.59787524e+00 1.36739209e-01 7.29846656e-01
9.31553543e-01 -4.88785148e-01 1.48632085e+00 3.32875818e-01
-1.79693982e-01 8.74186587e-03 8.50400925e-02 1.43124923e-01
7.97077417e-01 -1.21718228e+00 3.38455588e-01 1.71341654e-02
-3.35311174e-01 5.75626791e-01 7.24295795e-01 -9.44406509e-01
-1.57333279e+00 -7.45375216e-01 1.92364538e+00 -3.66294861e-01
6.69883728e-01 -5.07601321e-01 -3.87468845e-01 6.01231873e-01
5.80579937e-01 -1.06131852e+00 6.92737103e-01 1.87136963e-01
3.01774710e-01 3.29037845e-01 -1.59232080e+00 7.16905296e-01
1.20966935e+00 -8.89953554e-01 -1.13506639e+00 8.30283701e-01
8.22356045e-01 -5.61596334e-01 -1.04175794e+00 -1.09440602e-01
4.06485170e-01 -7.01017082e-01 4.96524692e-01 -1.01902258e+00
-3.71179819e-01 2.92873532e-02 -2.13649154e-01 -1.23887801e+00
1.92776009e-01 -1.29541707e+00 4.54791784e-01 1.38046658e+00
5.90699971e-01 -1.19380093e+00 1.75248876e-01 1.39575195e+00
-4.77082878e-01 -5.16213961e-02 -1.12713277e+00 -7.33312130e-01
-4.54170853e-01 -3.72392893e-01 6.91349030e-01 9.46309745e-01
4.85918671e-01 8.56915712e-01 -2.35348105e-01 -2.66236395e-01
-1.62567228e-01 -4.52505440e-01 6.69498444e-01 -1.20153821e+00
3.76908392e-01 -3.08122247e-01 -1.59111425e-01 -9.87681568e-01
3.86447102e-01 -6.56475723e-01 2.41212353e-01 -1.93349898e+00
-3.35857630e-01 -3.33997846e-01 5.31110108e-01 3.94556522e-01
6.36741221e-01 -2.34665990e-01 2.57136375e-01 4.78376895e-01
-8.81398916e-01 6.31471813e-01 9.96767581e-01 2.36649707e-01
-5.34311235e-01 -3.44220400e-02 -6.25743389e-01 4.88174438e-01
5.20825207e-01 -5.42361915e-01 -1.37191787e-01 -6.53284192e-02
4.61189955e-01 7.08727658e-01 3.63565356e-01 -3.26743543e-01
6.93901837e-01 -3.37457746e-01 -3.46351296e-01 -2.40457475e-01
2.65106529e-01 -8.31218719e-01 3.03584188e-01 4.92398113e-01
-5.27280331e-01 1.36503175e-01 1.78910613e-01 3.11531816e-02
-4.27192450e-01 -5.29276013e-01 1.98071316e-01 -2.30638042e-01
-1.34800100e+00 -3.68853182e-01 -1.08548701e+00 2.39391550e-01
1.06590295e+00 -5.35031378e-01 -3.78620774e-01 -6.71217620e-01
-7.41160572e-01 3.94242853e-01 4.80121166e-01 5.34248412e-01
3.74954581e-01 -1.06284285e+00 -8.52594852e-01 -4.37906347e-02
2.56705731e-01 -5.85842013e-01 -1.77278277e-02 7.65061438e-01
-6.74009740e-01 1.11521900e+00 -5.20344019e-01 2.92862300e-02
-1.04519117e+00 -5.52798696e-02 4.19212669e-01 -6.27810538e-01
-2.74717838e-01 4.22270149e-01 -2.97594905e-01 -1.20752299e+00
-1.44491911e-01 2.33326629e-01 -5.70469081e-01 2.22396970e-01
4.16774839e-01 6.88614547e-01 2.39023212e-02 -1.05009699e+00
-4.84119654e-01 -7.83027932e-02 6.65362552e-02 -8.86076272e-01
9.00827825e-01 -8.98596287e-01 -4.90674734e-01 6.21960402e-01
6.81462109e-01 -2.04466164e-01 -2.66684860e-01 -1.91676021e-01
7.03363240e-01 -1.99531242e-01 -3.19938391e-01 -1.45919514e+00
-1.13892384e-01 2.74667621e-01 4.53540117e-01 6.42315686e-01
7.38635063e-01 -3.18972617e-02 1.08566964e+00 1.10335696e+00
9.04583395e-01 -1.80208397e+00 -4.97536331e-01 1.20482337e+00
1.15710151e+00 -1.11108661e+00 -6.41414583e-01 -5.10516405e-01
-1.40611279e+00 1.27248812e+00 2.99099475e-01 8.44923198e-01
-3.44695188e-02 1.97129324e-01 5.96855879e-01 -5.64634264e-01
-7.89802372e-01 -3.49914223e-01 4.79423702e-02 1.09798276e+00
4.15436596e-01 1.10663258e-01 -8.36756051e-01 5.94452083e-01
-1.71930730e-01 1.38338506e-01 1.26497909e-01 1.11868083e+00
-3.34567755e-01 -1.21564496e+00 -9.74068269e-02 1.39904425e-01
-1.38176754e-01 -2.23343536e-01 -1.05511510e+00 1.10321486e+00
-3.18198711e-01 1.57529116e+00 -1.30437478e-01 -2.11501196e-01
1.02930462e+00 6.58821881e-01 4.14421642e-03 -9.69893873e-01
-1.50232995e+00 -3.48068148e-01 1.57344532e+00 -5.05147219e-01
-5.37227094e-01 -3.06870431e-01 -1.46440208e+00 -4.10320014e-01
-3.79338652e-01 9.47294235e-01 1.06563938e+00 1.08924997e+00
6.48678243e-01 -8.62737074e-02 7.08738983e-01 -4.50237989e-01
-5.63455701e-01 -1.11142981e+00 -8.17905605e-01 3.37297678e-01
-3.64518732e-01 -1.83607891e-01 -5.40995710e-02 -1.42290339e-01] | [12.719876289367676, 8.088706970214844] |
a1baa2d5-7d0f-4c27-a90b-00a79de577ca | a-dataset-for-starcraft-ai-an-example-of | 1211.4552 | null | http://arxiv.org/abs/1211.4552v1 | http://arxiv.org/pdf/1211.4552v1.pdf | A Dataset for StarCraft AI \& an Example of Armies Clustering | This paper advocates the exploration of the full state of recorded real-time
strategy (RTS) games, by human or robotic players, to discover how to reason
about tactics and strategy. We present a dataset of StarCraft games
encompassing the most of the games' state (not only player's orders). We
explain one of the possible usages of this dataset by clustering armies on
their compositions. This reduction of armies compositions to mixtures of
Gaussian allow for strategic reasoning at the level of the components. We
evaluated this clustering method by predicting the outcomes of battles based on
armies compositions' mixtures components | ['Gabriel Synnaeve', 'Pierre Bessiere'] | 2012-11-19 | null | null | null | null | ['real-time-strategy-games'] | ['playing-games'] | [-2.65560776e-01 2.85437554e-01 -3.07723135e-02 -1.38046354e-01
-6.47919029e-02 -1.42296004e+00 1.07577765e+00 -2.83114374e-01
-2.50595421e-01 4.06961590e-01 5.71120620e-01 -6.72775507e-01
-8.24126601e-01 -5.80453277e-01 -7.05237165e-02 -5.76799691e-01
-2.59075135e-01 1.42282140e+00 6.19163871e-01 -9.20239806e-01
6.16311252e-01 7.69105077e-01 -1.49299932e+00 7.43320227e-01
2.07706898e-01 8.54214132e-01 5.09716868e-02 9.41062748e-01
-2.99316049e-01 1.86592042e+00 -1.03254080e+00 -6.91078007e-01
7.36013532e-01 -2.46329665e-01 -1.03615665e+00 -9.08150747e-02
-5.22680163e-01 -1.39621869e-01 -4.51229066e-01 9.22320247e-01
-2.09351741e-02 2.28256598e-01 7.57756412e-01 -1.32993674e+00
4.01180506e-01 1.40647876e+00 -1.44541413e-01 2.14173362e-01
3.89102221e-01 3.21260840e-01 8.28190386e-01 2.38038033e-01
8.61303449e-01 1.35402155e+00 3.76565814e-01 4.33347285e-01
-9.63817060e-01 -5.10730386e-01 -9.20782536e-02 5.25910258e-01
-1.04634047e+00 -1.71208784e-01 7.39050388e-01 -8.08281898e-01
7.13780284e-01 5.21796644e-01 9.17459488e-01 1.16611814e+00
1.14729419e-01 8.83342087e-01 1.31390333e+00 -1.72060683e-01
6.12382948e-01 -1.56920388e-01 2.02458858e-01 3.40822846e-01
-8.84288773e-02 2.96169609e-01 -3.15845937e-01 -5.40855587e-01
7.68644989e-01 -3.54021639e-01 4.17158037e-01 -4.28423941e-01
-8.53931665e-01 8.31057012e-01 -9.28665996e-02 2.91137815e-01
-4.74958003e-01 8.36327597e-02 2.83681929e-01 2.27394164e-01
4.07750122e-02 1.11466324e+00 -4.80913758e-01 -1.00445938e+00
-6.60670340e-01 8.35033000e-01 1.27446175e+00 6.49969697e-01
3.02269727e-01 -4.07250263e-02 2.22534090e-02 3.62303585e-01
-8.97067934e-02 -6.69118240e-02 4.18542564e-01 -1.51883256e+00
2.14288309e-01 7.26119399e-01 1.71925321e-01 -7.92085826e-01
-7.54504323e-01 -3.34058374e-01 3.76260579e-02 6.11013889e-01
4.92213249e-01 -2.02675387e-01 -5.36576450e-01 1.37769747e+00
4.53167409e-02 -9.52321216e-02 1.36440679e-01 8.58326495e-01
5.63601136e-01 6.87685013e-02 -1.55528471e-01 8.20974857e-02
1.34266031e+00 -5.65331042e-01 -4.43206668e-01 -4.75214906e-02
2.78044283e-01 -1.84352219e-01 5.54217398e-01 9.09344256e-01
-1.08209479e+00 -2.38293007e-01 -6.65047884e-01 5.94407380e-01
-1.15754113e-01 -1.54361933e-01 1.15261102e+00 5.76359451e-01
-5.77340603e-01 8.95126700e-01 -1.02440727e+00 1.04608528e-01
1.16120800e-01 4.06338573e-01 -6.89622685e-02 6.25296712e-01
-9.77060556e-01 1.12156463e+00 6.15941286e-01 -2.06625655e-01
-1.19615996e+00 -7.74128735e-02 -3.92878830e-01 1.03485614e-01
1.02342522e+00 1.16333757e-02 1.52693558e+00 -3.86377960e-01
-1.81869328e+00 8.00394177e-01 7.04179823e-01 -5.94349205e-01
6.56477690e-01 8.98644254e-02 -1.66560233e-01 6.82990905e-03
-1.67184114e-01 9.27744210e-02 4.03519034e-01 -1.09397697e+00
-8.41067493e-01 -2.78392762e-01 7.59133935e-01 3.44329149e-01
4.45734590e-01 2.86592633e-01 -8.24701041e-02 -1.21289678e-01
6.08533770e-02 -1.21879458e+00 -6.62383199e-01 -1.10694766e+00
-4.56493497e-01 -2.68076211e-01 3.63117158e-02 -5.55930614e-01
1.21077621e+00 -2.00662494e+00 9.51465011e-01 3.54557365e-01
6.18691921e-01 -1.29229501e-01 1.49535477e-01 8.87392342e-01
7.44050881e-03 2.03158259e-02 3.25406134e-01 -2.68114567e-01
4.87257063e-01 5.03792822e-01 -5.82078755e-01 4.23377037e-01
-6.05783343e-01 7.63177395e-01 -6.70927227e-01 -3.11483681e-01
3.87404621e-01 -3.10470074e-01 -5.37164569e-01 2.91042536e-01
-4.58214611e-01 5.88157594e-01 -6.94608629e-01 3.98683310e-01
1.63462564e-01 4.82083887e-01 5.80088317e-01 1.68254420e-01
-3.85331154e-01 5.31712472e-01 -1.00311220e+00 1.40688479e+00
2.76175261e-01 5.11432230e-01 1.89820170e-01 -8.74521554e-01
7.93272436e-01 -1.54002495e-02 7.00048029e-01 -2.21421555e-01
7.14124978e-01 3.06788441e-02 7.73682892e-01 -3.12368721e-01
6.55577064e-01 -3.24392736e-01 -9.40181851e-01 6.62649035e-01
1.93451703e-01 -6.32934749e-01 3.89791191e-01 2.51912981e-01
1.44222641e+00 3.45320165e-01 3.67502600e-01 -3.53474319e-01
1.50731858e-02 6.25357270e-01 3.70024830e-01 1.18007457e+00
-1.62843689e-01 1.97303861e-01 1.13308966e+00 -6.52220368e-01
-7.84575045e-01 -1.09624422e+00 4.73322451e-01 1.26305687e+00
3.14024724e-02 -7.87515581e-01 -7.06897557e-01 -3.28105837e-01
-3.33189845e-01 1.06892383e+00 -9.99002874e-01 -1.54200882e-01
-4.30165172e-01 -3.26827019e-01 8.48335564e-01 2.08295718e-01
9.46831480e-02 -1.31314516e+00 -1.36274850e+00 2.41416439e-01
-9.04161260e-02 -9.80064213e-01 3.52892846e-01 5.40201426e-01
-3.72558415e-01 -1.57105207e+00 2.71699101e-01 3.24678242e-01
-2.72767663e-01 -3.86231124e-01 1.05348110e+00 -2.28549227e-01
-2.22827286e-01 6.00522697e-01 -7.84283400e-01 -7.53173232e-01
-7.21491337e-01 -7.60805607e-03 1.47055611e-01 -4.34135079e-01
1.58251777e-01 -7.78660357e-01 -7.13446438e-02 4.06758755e-01
-6.01855278e-01 2.16131479e-01 4.02324468e-01 2.92923480e-01
-1.77553996e-01 6.59124255e-01 -7.09801793e-01 -7.26857305e-01
1.32149982e+00 -6.37278974e-01 -5.71935654e-01 2.18013406e-01
1.57962546e-01 2.01878279e-01 3.99027199e-01 -6.79198921e-01
-7.95491636e-01 -3.34620714e-01 3.05336118e-01 -6.08925402e-01
-3.35036248e-01 5.02551615e-01 2.58577149e-02 2.89553434e-01
7.36683846e-01 1.70194089e-01 1.98590755e-01 -5.47431648e-01
4.72652584e-01 5.16364396e-01 8.72744918e-01 -1.22568882e+00
4.55372155e-01 2.83373356e-01 1.09631538e-01 -4.44105268e-01
-3.07535499e-01 -2.41986603e-01 -6.20441079e-01 -7.20507920e-01
7.10277379e-01 -3.52175206e-01 -1.73646879e+00 4.45878059e-01
-9.09745157e-01 -5.47311902e-01 -5.45148671e-01 4.41950232e-01
-1.12558353e+00 9.98828337e-02 -6.97987378e-01 -1.32375348e+00
3.13710988e-01 -1.29964495e+00 5.65249622e-01 2.02698857e-01
-7.43620455e-01 -6.53306007e-01 3.68795812e-01 3.21210444e-01
3.08839142e-01 2.95579374e-01 8.52462471e-01 -1.44595194e+00
7.22660261e-05 -1.14956588e-01 5.75601161e-01 -2.74757385e-01
-1.08634733e-01 -2.01165341e-02 -3.57183605e-01 3.44984651e-01
4.97011632e-01 -1.18204229e-01 1.27886653e-01 1.68941528e-01
7.39903271e-01 -1.64723739e-01 -1.35438159e-01 1.98385343e-01
7.67636180e-01 7.52527416e-01 7.70643651e-01 6.13975406e-01
3.21270853e-01 1.12591207e+00 5.56992292e-01 7.01992273e-01
2.73365229e-01 1.00403559e+00 7.37927198e-01 5.91049731e-01
2.76950628e-01 -4.39175874e-01 3.06147933e-01 2.37909555e-01
-9.05030787e-01 -1.96052920e-02 -1.08290160e+00 6.34817705e-02
-2.11993647e+00 -1.39112282e+00 5.34808412e-02 1.65515220e+00
3.51102889e-01 5.15189409e-01 8.63633096e-01 2.74368048e-01
4.80386168e-01 9.85777751e-02 -3.29033822e-01 -6.94116712e-01
1.75398625e-02 2.94009954e-01 7.03274906e-01 5.06794214e-01
-7.88897216e-01 1.26191878e+00 7.58095837e+00 1.27098405e+00
-4.21034127e-01 -1.25062972e-01 6.14343351e-03 -3.62159908e-01
-7.28614107e-02 4.57017839e-01 -3.39872956e-01 5.06231725e-01
7.25525260e-01 7.53644109e-02 1.13123453e+00 8.23013186e-01
-5.88088892e-02 -3.44817638e-01 -7.33614385e-01 6.40814364e-01
-1.75595418e-01 -1.37720323e+00 -2.23288432e-01 4.79767710e-01
-2.52876766e-02 -2.45317712e-01 -2.48764277e-01 7.65363991e-01
1.64743686e+00 -1.16565752e+00 1.53117871e+00 5.51105022e-01
2.45099753e-01 -6.67116761e-01 4.98049676e-01 8.40439498e-01
-8.60388517e-01 -7.63292730e-01 -6.49218028e-03 -7.47512877e-01
1.13292210e-01 -4.46470588e-01 -8.88305187e-01 6.05790436e-01
5.55181205e-01 2.99220324e-01 -4.60871249e-01 4.16050673e-01
-2.54093081e-01 5.21137893e-01 -4.55045551e-01 -2.80063212e-01
4.81254697e-01 -6.84971750e-01 8.93510878e-01 6.79995000e-01
-2.39123330e-02 7.65967071e-01 3.96440566e-01 7.87597716e-01
1.04256332e+00 -5.14213026e-01 -2.47526556e-01 -4.47196662e-01
5.24476588e-01 1.01338410e+00 -1.08572340e+00 -1.21191800e-01
4.83979464e-01 4.78611141e-01 8.54401067e-02 -1.91880047e-01
-9.07419443e-01 2.61153519e-01 9.44750130e-01 2.88158983e-01
6.19242489e-02 -4.39055026e-01 -2.25185484e-01 -9.17587280e-01
-6.85020745e-01 -1.38990414e+00 4.33338404e-01 -9.21511531e-01
-1.00251591e+00 7.14622974e-01 7.87469268e-01 -1.10905862e+00
-7.88355649e-01 -8.19824696e-01 -7.09209025e-01 3.83095533e-01
-1.05109304e-01 -1.04949641e+00 1.57717079e-01 6.43019974e-01
3.60011160e-01 -9.28223670e-01 6.61822200e-01 -5.45557201e-01
-9.52195972e-02 -2.54030466e-01 -3.86532038e-01 2.09027052e-01
-1.92307845e-01 -1.33419895e+00 1.97518140e-01 1.95832789e-01
4.13455993e-01 4.14596468e-01 1.48541880e+00 -6.40051126e-01
-1.18270099e+00 2.86231488e-01 -1.34770572e-01 -8.15397322e-01
1.18204033e+00 -3.37699443e-01 -1.78288385e-01 7.54212022e-01
1.46758119e-02 -9.84661698e-01 6.32506490e-01 4.43616927e-01
-1.67300552e-01 4.94721323e-01 -8.67223501e-01 9.95494187e-01
1.34787059e+00 -4.25802857e-01 -1.13078141e+00 -3.04339100e-02
2.13560790e-01 -5.51785350e-01 -4.67088312e-01 3.57222617e-01
7.73526669e-01 -1.41131353e+00 8.83611381e-01 -1.19782329e+00
4.53579038e-01 -1.88923717e-01 -4.20594305e-01 -1.30298567e+00
-4.50959474e-01 -8.13588321e-01 -4.30235043e-02 5.49958289e-01
-1.76544547e-01 -4.02416408e-01 8.63021731e-01 8.69270921e-01
-2.59118110e-01 -3.97105336e-01 -1.01008463e+00 -5.94987929e-01
2.04516158e-01 -8.46277475e-01 7.19543576e-01 5.28620958e-01
5.96441329e-01 1.49836093e-01 -4.09049630e-01 -4.05046582e-01
5.29512227e-01 2.96765327e-01 1.10059118e+00 -1.35694706e+00
-9.49694991e-01 -1.06137168e+00 -7.72212863e-01 -7.72662163e-01
3.01783592e-01 -6.06407642e-01 -2.21444275e-02 -1.08892310e+00
2.74871737e-02 -3.30987751e-01 2.40484372e-01 3.65426898e-01
4.58778441e-01 -2.62956828e-01 6.79668128e-01 5.50220549e-01
-7.71562874e-01 1.76468957e-02 7.70741045e-01 8.41735955e-03
-1.17675327e-01 3.41860890e-01 -6.91169918e-01 1.11726630e+00
7.38002598e-01 -4.69285846e-01 -3.62772375e-01 8.11687335e-02
5.03235340e-01 6.48070812e-01 3.96848544e-02 -1.02859187e+00
4.17343020e-01 -7.94166505e-01 -3.19459081e-01 -4.25697803e-01
5.47920644e-01 -7.34108627e-01 9.97786880e-01 6.59414113e-01
-4.57710922e-01 -4.52982821e-02 1.59745261e-01 2.64344484e-01
-7.39192590e-02 -3.31886649e-01 9.08807144e-02 -3.79412264e-01
-7.35912442e-01 -2.66611129e-01 -1.10042834e+00 -1.70224681e-01
1.22163093e+00 -4.44374710e-01 -3.00576150e-01 -7.56516695e-01
-1.17375433e+00 9.52645615e-02 3.53706867e-01 3.23915392e-01
8.70975107e-02 -7.07613766e-01 -5.54414809e-01 -8.88736248e-02
-3.96822184e-01 -5.80724955e-01 3.94429952e-01 6.14292204e-01
-7.92916834e-01 1.77457467e-01 -7.65269279e-01 -1.96728304e-01
-1.28162420e+00 6.61978364e-01 6.60688460e-01 -7.09042549e-01
-5.28729618e-01 5.70800602e-01 1.78004559e-02 -5.24220288e-01
-5.11454828e-02 9.83923525e-02 -7.33952284e-01 -5.96067961e-03
2.98025757e-01 4.13056910e-01 -5.17974615e-01 -6.26631975e-01
-4.45673555e-01 1.96594253e-01 2.09807858e-01 -8.32320750e-01
1.38975930e+00 2.89229661e-01 -1.95812091e-01 6.66400552e-01
7.06285089e-02 4.09815282e-01 -1.16788805e+00 2.40736097e-01
1.93625942e-01 -3.90119404e-01 -6.48981214e-01 -8.94529760e-01
-5.78844011e-01 4.28521723e-01 -1.81307122e-01 6.49937212e-01
6.62014961e-01 4.54006612e-01 -5.09749092e-02 3.59895349e-01
1.22616184e+00 -8.46681058e-01 -1.28814176e-01 8.15750062e-01
7.73195446e-01 -4.70432341e-01 7.92428628e-02 -2.27313504e-01
-1.15534782e+00 1.18369925e+00 2.17412829e-01 -5.20722687e-01
5.70188761e-01 6.48460865e-01 2.78520435e-02 -6.97187006e-01
-9.66923535e-01 -4.41274732e-01 6.22268617e-02 6.09903574e-01
-4.04951453e-01 6.11373425e-01 -5.21404624e-01 1.52370536e+00
-1.14063287e+00 -4.31094468e-01 8.70803952e-01 9.22170818e-01
-4.43393379e-01 -1.03551722e+00 -5.32323718e-01 3.90710145e-01
-4.08815980e-01 2.96472579e-01 -1.18353570e+00 7.48071253e-01
3.35140258e-01 1.16660655e+00 1.24479741e-01 -1.13251531e+00
4.65147287e-01 -2.74802715e-01 7.31005788e-01 -6.28499269e-01
-1.00968468e+00 -1.90504476e-01 4.64216322e-01 -7.95506716e-01
-1.85354590e-01 -8.48903239e-01 -9.39334810e-01 -8.51509035e-01
3.17131765e-02 6.42559767e-01 7.07899988e-01 1.11959922e+00
-1.89803213e-01 4.98912007e-01 3.38698149e-01 -1.19049573e+00
-7.94535697e-01 -1.13077331e+00 -9.57509518e-01 4.98690277e-01
-5.07695317e-01 -1.07829416e+00 -5.13351262e-01 -4.57216084e-01] | [3.4836955070495605, 1.4696345329284668] |
5a1b5279-cd54-4b76-886a-7e01aef57f79 | deepfit-3d-surface-fitting-via-neural-network | 2003.10826 | null | https://arxiv.org/abs/2003.10826v1 | https://arxiv.org/pdf/2003.10826v1.pdf | DeepFit: 3D Surface Fitting via Neural Network Weighted Least Squares | We propose a surface fitting method for unstructured 3D point clouds. This method, called DeepFit, incorporates a neural network to learn point-wise weights for weighted least squares polynomial surface fitting. The learned weights act as a soft selection for the neighborhood of surface points thus avoiding the scale selection required of previous methods. To train the network we propose a novel surface consistency loss that improves point weight estimation. The method enables extracting normal vectors and other geometrical properties, such as principal curvatures, the latter were not presented as ground truth during training. We achieve state-of-the-art results on a benchmark normal and curvature estimation dataset, demonstrate robustness to noise, outliers and density variations, and show its application on noise removal. | ['Yizhak Ben-Shabat', 'Stephen Gould'] | 2020-03-23 | null | https://www.ecva.net/papers/eccv_2020/papers_ECCV/html/283_ECCV_2020_paper.php | https://www.ecva.net/papers/eccv_2020/papers_ECCV/papers/123460018.pdf | eccv-2020-8 | ['surface-normals-estimation'] | ['computer-vision'] | [ 1.11117221e-01 -2.17761821e-03 1.42467529e-01 -2.92755067e-01
-7.64051437e-01 -1.78744838e-01 3.28398705e-01 7.54431114e-02
-3.58733505e-01 1.70646384e-01 -3.84959430e-01 -5.85673861e-02
-5.67016006e-02 -8.07628989e-01 -9.80238497e-01 -5.73181868e-01
-2.47529224e-01 6.55875742e-01 3.87448519e-01 -2.48729438e-01
5.03577411e-01 1.19754171e+00 -1.38860691e+00 -2.61325300e-01
9.60587204e-01 1.12763059e+00 -5.10665476e-02 3.34504753e-01
-1.73261717e-01 -2.55725473e-01 -1.41789392e-01 -3.07703823e-01
5.48395216e-01 6.25029385e-01 -2.48803750e-01 1.45921722e-01
8.92731786e-01 -3.24203700e-01 -3.31631228e-02 1.11089706e+00
3.68779808e-01 7.97059909e-02 9.11919713e-01 -8.80800247e-01
-4.85211164e-01 -2.81195909e-01 -7.87822485e-01 -5.54481387e-01
9.39183533e-02 4.90698554e-02 9.16631639e-01 -1.56860459e+00
7.03479588e-01 1.13168955e+00 1.33364069e+00 4.47745949e-01
-1.37608230e+00 -2.83442855e-01 -1.13021836e-01 -2.87517458e-01
-1.45967066e+00 -3.07788193e-01 1.28125942e+00 -5.09332836e-01
9.07753706e-01 3.74299914e-01 5.92265844e-01 5.95425665e-01
3.21463309e-02 5.96722543e-01 4.86317039e-01 -4.08838093e-01
2.24908650e-01 -1.05372339e-01 -1.92667022e-01 7.36769557e-01
3.54519367e-01 -1.24789320e-01 -6.18955456e-02 -5.33389449e-01
1.37300742e+00 -1.76226310e-02 -3.83434325e-01 -1.00557733e+00
-1.08301473e+00 7.62288153e-01 5.20071864e-01 -3.44878547e-02
-5.05444229e-01 1.54334605e-01 2.26667538e-01 1.02554254e-01
9.70746815e-01 3.54153454e-01 -7.12735116e-01 1.57399535e-01
-7.99805641e-01 4.44049984e-01 5.68554521e-01 9.20830846e-01
1.23722970e+00 -1.10308768e-03 1.01474501e-01 1.06462538e+00
6.00336075e-01 8.45474243e-01 2.93379426e-02 -1.25544059e+00
4.79670793e-01 1.03564572e+00 2.73160338e-01 -1.44191337e+00
-6.55401707e-01 -3.23005095e-02 -1.05638289e+00 6.68111384e-01
2.48170853e-01 8.09140131e-02 -9.48314428e-01 1.14868295e+00
5.71792364e-01 4.04879689e-01 -4.70122516e-01 8.72808218e-01
6.34986997e-01 3.32857579e-01 -5.84107399e-01 9.52845439e-03
7.09693789e-01 -5.69074929e-01 -3.44449639e-01 -1.89289346e-01
3.82508844e-01 -7.87546992e-01 1.05914938e+00 3.90600353e-01
-1.44370174e+00 -2.34964862e-01 -9.20977652e-01 -2.02207595e-01
-5.30529879e-02 2.13247165e-01 4.08549130e-01 2.06401989e-01
-1.05835581e+00 1.18548727e+00 -9.85118687e-01 1.61293209e-01
5.90572655e-01 5.95337510e-01 -3.69647294e-01 3.65631640e-01
-5.89905024e-01 6.96899176e-01 -2.75483817e-01 1.96114972e-01
-5.30220158e-02 -8.75755012e-01 -9.10700738e-01 -2.17073262e-02
2.73604244e-01 -7.60291755e-01 9.94955838e-01 -3.67339551e-01
-1.77910793e+00 1.01670194e+00 -4.73631144e-01 -1.58685893e-01
6.92272782e-01 -3.71283501e-01 8.65742788e-02 -2.11613886e-02
-1.52002707e-01 3.78171206e-01 1.09235013e+00 -1.64083910e+00
-9.49770287e-02 -5.26521623e-01 -7.11625934e-01 -1.12143427e-01
-2.06602514e-01 -2.24279135e-01 -7.10788369e-01 -6.24223292e-01
7.32755899e-01 -7.59852588e-01 -6.52247250e-01 5.30325711e-01
-4.58945185e-01 -4.03897285e-01 8.82445216e-01 -4.64598775e-01
9.47505057e-01 -2.17805767e+00 -2.15538070e-02 1.10698652e+00
4.82265949e-01 1.67218402e-01 -1.66210175e-01 9.07277539e-02
1.00269593e-01 1.47823468e-01 -5.91792166e-01 -7.44275272e-01
1.01118691e-01 6.02679793e-03 -1.53429046e-01 6.77159488e-01
4.90731329e-01 8.28821123e-01 -6.59434021e-01 -1.26595750e-01
4.04165149e-01 7.15296745e-01 -5.66926301e-01 1.84281185e-01
-1.99534118e-01 1.11659728e-01 -4.07786936e-01 6.75624251e-01
1.16598296e+00 -3.38320762e-01 -5.10533392e-01 -4.81972098e-01
1.33695761e-02 4.95811477e-02 -1.44455087e+00 1.63218462e+00
-3.46736372e-01 2.27851480e-01 3.59383374e-01 -6.36375964e-01
1.51011384e+00 6.55696914e-02 8.68629694e-01 -1.23759724e-01
1.57011762e-01 6.00830078e-01 -4.89137053e-01 -2.52213061e-01
4.07632023e-01 5.11877425e-02 4.17693436e-01 -8.01120028e-02
-4.34902191e-01 -8.52943659e-01 -4.84991401e-01 -2.47144878e-01
8.29484701e-01 2.17091635e-01 6.07116073e-02 -2.85215974e-01
5.53611040e-01 -1.01606764e-01 5.67200661e-01 2.54533738e-01
1.08633421e-01 9.23992515e-01 2.53840297e-01 -6.77305698e-01
-1.36585414e+00 -9.31114495e-01 -4.68006194e-01 4.69723642e-01
1.78600252e-01 -1.82660073e-01 -6.49195552e-01 -3.71333867e-01
5.50137639e-01 3.77662927e-01 -5.48628092e-01 5.62380105e-02
-8.29591691e-01 -3.44847500e-01 1.40358210e-01 5.11037707e-01
-9.43300128e-03 -7.14048088e-01 -3.15808654e-01 6.50548488e-02
4.14669931e-01 -1.17591655e+00 -5.80314994e-01 3.71094942e-02
-1.41397285e+00 -1.06321979e+00 -6.88333988e-01 -7.06631780e-01
1.03370321e+00 2.01483760e-02 1.30434060e+00 3.23248327e-01
-8.53836536e-02 3.16436589e-01 2.63654888e-01 -4.65931565e-01
-2.12347105e-01 -7.93792158e-02 7.99566582e-02 1.08576640e-02
4.81741905e-01 -6.49925292e-01 -6.57405555e-01 4.86116379e-01
-5.02664745e-01 -6.39641732e-02 2.11271569e-01 7.71453798e-01
1.09059644e+00 -4.93126750e-01 3.47954743e-02 -7.96027303e-01
4.99700874e-01 -1.24224000e-01 -8.94064784e-01 -1.04071610e-01
-4.19404000e-01 -4.20647524e-02 3.80209297e-01 -3.38178664e-01
-5.47766566e-01 3.64577472e-01 -4.73927110e-01 -8.63522053e-01
-1.44468993e-01 2.52529740e-01 1.51492193e-01 -9.00132298e-01
7.58406758e-01 -1.16107941e-01 2.87897706e-01 -6.39020920e-01
1.85116410e-01 3.53850335e-01 4.87827659e-01 -5.07469237e-01
1.31666958e+00 8.60080540e-01 4.90024447e-01 -1.04427636e+00
-3.31193298e-01 -7.23691583e-01 -9.57364440e-01 1.61407385e-02
4.54765320e-01 -7.22088456e-01 -8.14444780e-01 5.15726388e-01
-1.44375861e+00 -2.95468420e-01 -4.32372302e-01 1.47634342e-01
-8.73506784e-01 6.46049976e-01 -4.11582172e-01 -7.87232339e-01
-5.40943384e-01 -1.09914124e+00 1.53740454e+00 -8.37364271e-02
-9.68957320e-02 -1.03722620e+00 9.36486498e-02 -2.22854167e-01
4.28353101e-01 5.47666430e-01 7.65552580e-01 -2.51890749e-01
-7.02425659e-01 -5.41607618e-01 -2.97608256e-01 4.01515454e-01
1.15306616e-01 4.29571360e-01 -8.03981125e-01 -2.87124306e-01
4.84155901e-02 1.79499332e-02 7.60445058e-01 5.41841149e-01
1.47414017e+00 -1.38578683e-01 -4.25216317e-01 1.13150752e+00
1.53896511e+00 -4.14044201e-01 7.26241529e-01 3.78838539e-01
1.02776241e+00 2.73176312e-01 1.87940747e-01 3.94696176e-01
1.87038630e-01 7.78634250e-01 8.03514898e-01 -2.90539831e-01
2.20483959e-01 -5.86433932e-02 -1.44078985e-01 8.89076233e-01
-5.12360156e-01 4.71715689e-01 -1.05985713e+00 5.10352254e-01
-1.80131006e+00 -3.67507398e-01 -5.15476882e-01 2.42245960e+00
6.66704595e-01 1.28477022e-01 -4.93950285e-02 1.04905091e-01
5.03129721e-01 -1.53996557e-01 -7.25021541e-01 -2.78150380e-01
-2.20670730e-01 2.72483915e-01 7.99989700e-01 5.20839751e-01
-1.13085532e+00 9.36478317e-01 6.66961002e+00 7.52517462e-01
-1.09575057e+00 -2.80533433e-01 2.85293877e-01 3.22300375e-01
-6.76837742e-01 -3.65429103e-01 -5.97519755e-01 1.07856832e-01
3.28975558e-01 2.86427617e-01 3.46713871e-01 1.03640890e+00
2.63110787e-01 3.81736100e-01 -6.56593323e-01 1.05481768e+00
-1.62907332e-01 -1.53290725e+00 -8.97148699e-02 4.23994027e-02
8.42710614e-01 5.79997897e-01 -1.66566983e-01 -1.89606935e-01
1.08845733e-01 -9.50687170e-01 3.71955723e-01 8.67328346e-01
9.53824699e-01 -7.37594366e-01 6.72015786e-01 3.16841632e-01
-1.03372431e+00 2.50725031e-01 -6.81603551e-01 2.65124440e-01
3.97040993e-01 9.15149450e-01 -7.47429311e-01 3.11083764e-01
6.57189250e-01 7.82681406e-01 -2.94660032e-01 1.35216105e+00
-2.84772571e-02 3.60172004e-01 -8.69156480e-01 2.82166619e-02
3.58691216e-02 -6.40781403e-01 9.51036870e-01 9.61016715e-01
4.21554893e-01 -9.23385993e-02 1.71150163e-01 1.02478492e+00
-2.16802731e-01 4.44713533e-01 -7.15472579e-01 5.28117776e-01
4.51838553e-01 1.14781725e+00 -5.27528644e-01 2.42481716e-02
-3.56558949e-01 7.20693767e-01 4.85718399e-01 4.58370537e-01
-2.19806612e-01 -3.99815619e-01 7.63999224e-01 5.84920049e-01
2.93313026e-01 -6.54941261e-01 -9.91960943e-01 -1.02312708e+00
2.76575834e-01 -4.26237911e-01 -3.73656124e-01 -6.13515913e-01
-1.62469757e+00 2.65055448e-01 -1.96821794e-01 -1.25111508e+00
2.80472964e-01 -9.54566240e-01 -9.59049284e-01 9.35806513e-01
-1.69193304e+00 -9.58435714e-01 -3.57547879e-01 2.99608946e-01
2.06065267e-01 -1.39521360e-01 7.02958286e-01 1.14633419e-01
-2.14647219e-01 4.47466016e-01 2.48749256e-01 1.78171191e-02
5.13032496e-01 -1.33163583e+00 9.76277173e-01 3.56249332e-01
-2.65462905e-01 7.89712191e-01 5.00036716e-01 -7.80622900e-01
-1.44146943e+00 -9.68794107e-01 7.80342758e-01 -3.61093402e-01
5.34332812e-01 -3.38372737e-01 -1.50183904e+00 4.52891886e-01
-4.31288838e-01 3.68601561e-01 1.53188437e-01 2.58372054e-02
-1.57439679e-01 1.29057735e-01 -1.41152322e+00 5.30247152e-01
1.03110635e+00 -1.64504290e-01 -3.31997037e-01 2.42579520e-01
6.93640172e-01 -8.34373236e-01 -1.09976637e+00 1.03404343e+00
4.40094262e-01 -8.06540370e-01 1.32159960e+00 -3.00713271e-01
1.19019836e-01 -3.90470415e-01 3.87338363e-02 -1.24378300e+00
-5.86282194e-01 -7.73600459e-01 -4.53447104e-01 6.46264374e-01
4.21397835e-01 -5.92271626e-01 1.14966428e+00 7.54985690e-01
-3.31401944e-01 -1.21749103e+00 -1.15703702e+00 -6.34916246e-01
1.56479463e-01 -6.40232444e-01 7.55000472e-01 1.00271273e+00
-4.50651497e-01 -1.30689725e-01 -1.18958235e-01 2.41354957e-01
9.35013473e-01 -2.35900179e-01 9.95289862e-01 -1.87577093e+00
3.69889975e-01 -7.56341517e-01 -5.15047133e-01 -1.21000338e+00
1.96533233e-01 -7.11272895e-01 2.84617599e-02 -1.57281661e+00
-5.02278507e-01 -6.63219452e-01 8.61610547e-02 4.67171609e-01
-1.61543489e-01 3.08250457e-01 -4.61996533e-02 3.58579963e-01
-2.96560507e-02 8.14437389e-01 1.51236272e+00 3.08400448e-02
-4.64560062e-01 3.56372774e-01 -2.58278757e-01 1.13836658e+00
7.39861667e-01 -1.92190170e-01 1.70065552e-01 -5.26304901e-01
5.10322571e-01 -4.21504855e-01 3.06180149e-01 -8.41676354e-01
2.49422774e-01 -3.16996515e-01 4.71909940e-01 -9.19237852e-01
4.29197103e-01 -1.29335999e+00 -1.59597680e-01 2.03526877e-02
7.17966855e-02 -1.71602368e-02 3.46486717e-01 5.32123446e-01
1.58039689e-01 -2.36664131e-01 7.88109303e-01 1.42642498e-01
-1.62344664e-01 7.64890254e-01 2.51814067e-01 -2.66386896e-01
7.36738026e-01 -5.84396660e-01 2.58860379e-01 -3.78106013e-02
-5.35510540e-01 2.24143878e-01 9.11497116e-01 1.68814287e-01
1.11415350e+00 -1.55736172e+00 -7.95350373e-01 5.91713250e-01
1.11729898e-01 7.71739542e-01 -5.30233532e-02 7.06013620e-01
-7.81987727e-01 -2.18873635e-01 2.07916573e-01 -9.47983623e-01
-9.78045762e-01 2.18945771e-01 5.54074526e-01 8.39609951e-02
-9.23954606e-01 8.61597478e-01 -3.19372088e-01 -9.69801068e-01
2.02705964e-01 -6.60523057e-01 -1.27558023e-01 -4.07861859e-01
1.47183165e-01 6.07735097e-01 3.43086809e-01 -6.27678871e-01
-2.67486989e-01 1.37883294e+00 3.23467642e-01 3.43772948e-01
1.61741447e+00 1.91757511e-02 -3.46785307e-01 4.70752984e-01
1.33076131e+00 2.17903554e-01 -1.47040331e+00 -4.38537925e-01
1.28562555e-01 -6.71479344e-01 1.87710494e-01 -2.50807822e-01
-1.07034385e+00 7.76052475e-01 5.15966356e-01 1.89156309e-01
6.12480223e-01 -2.82502621e-01 1.09762764e+00 5.88760376e-01
2.37579405e-01 -1.15165329e+00 -2.47416511e-01 8.35802853e-01
1.09007955e+00 -1.37109590e+00 9.72756520e-02 -6.83048725e-01
-2.05588147e-01 1.47498524e+00 4.15696234e-01 -7.16674507e-01
9.60410774e-01 2.76393443e-01 1.00276999e-01 -4.76677984e-01
-2.24272996e-01 -1.76462792e-02 9.13835704e-01 6.04297578e-01
2.95126468e-01 -7.92514831e-02 -2.33975351e-02 1.64799765e-01
-1.23318218e-01 -2.50173882e-02 6.60879090e-02 5.61474979e-01
-6.00775182e-01 -8.81632984e-01 -5.64734995e-01 6.25713170e-01
-1.48735493e-01 1.62452623e-01 -2.38600567e-01 4.64153230e-01
-1.89745471e-01 1.97900608e-01 2.84483969e-01 -2.25371301e-01
8.20251524e-01 -8.73180777e-02 2.37047464e-01 -4.33623433e-01
-4.19693142e-01 1.85945436e-01 -2.28222370e-01 -6.08594000e-01
-2.23807707e-01 -5.69432735e-01 -1.35567772e+00 -1.73773661e-01
-4.67224538e-01 -1.52117893e-01 1.00303626e+00 5.85262716e-01
5.96362293e-01 6.20994233e-02 7.87597358e-01 -1.61547649e+00
-6.70282304e-01 -7.89814174e-01 -3.43033612e-01 8.52630615e-01
5.27758896e-01 -7.16392756e-01 -6.11642718e-01 -2.35702619e-01] | [8.201544761657715, -3.4246022701263428] |
83f72902-e811-4d7f-b33d-fd08114810ae | context-encoding-for-semantic-segmentation | 1803.08904 | null | http://arxiv.org/abs/1803.08904v1 | http://arxiv.org/pdf/1803.08904v1.pdf | Context Encoding for Semantic Segmentation | Recent work has made significant progress in improving spatial resolution for
pixelwise labeling with Fully Convolutional Network (FCN) framework by
employing Dilated/Atrous convolution, utilizing multi-scale features and
refining boundaries. In this paper, we explore the impact of global contextual
information in semantic segmentation by introducing the Context Encoding
Module, which captures the semantic context of scenes and selectively
highlights class-dependent featuremaps. The proposed Context Encoding Module
significantly improves semantic segmentation results with only marginal extra
computation cost over FCN. Our approach has achieved new state-of-the-art
results 51.7% mIoU on PASCAL-Context, 85.9% mIoU on PASCAL VOC 2012. Our single
model achieves a final score of 0.5567 on ADE20K test set, which surpass the
winning entry of COCO-Place Challenge in 2017. In addition, we also explore how
the Context Encoding Module can improve the feature representation of
relatively shallow networks for the image classification on CIFAR-10 dataset.
Our 14 layer network has achieved an error rate of 3.45%, which is comparable
with state-of-the-art approaches with over 10 times more layers. The source
code for the complete system are publicly available. | ['Ambrish Tyagi', 'Hang Zhang', 'Amit Agrawal', 'Zhongyue Zhang', 'Kristin Dana', 'Xiaogang Wang', 'Jianping Shi'] | 2018-03-23 | context-encoding-for-semantic-segmentation-1 | http://openaccess.thecvf.com/content_cvpr_2018/html/Zhang_Context_Encoding_for_CVPR_2018_paper.html | http://openaccess.thecvf.com/content_cvpr_2018/papers/Zhang_Context_Encoding_for_CVPR_2018_paper.pdf | cvpr-2018-6 | ['thermal-image-segmentation'] | ['computer-vision'] | [ 3.45696479e-01 5.51463850e-02 -6.94340467e-02 -5.60115635e-01
-7.78652847e-01 -5.06212533e-01 3.51594359e-01 -1.42628282e-01
-9.32139099e-01 6.53702319e-01 -6.46589249e-02 -1.48476645e-01
1.18826397e-01 -6.92807794e-01 -8.98729682e-01 -5.43498576e-01
-7.23755583e-02 6.94702193e-02 5.11819601e-01 -8.29190761e-02
2.20072493e-01 3.23768795e-01 -1.36193895e+00 7.89093852e-01
7.66347349e-01 1.45335352e+00 3.67580026e-01 7.42996514e-01
-2.17549816e-01 6.51802540e-01 -7.35365093e-01 -2.51050115e-01
1.91139445e-01 -1.80090785e-01 -1.26402259e+00 -2.04217479e-01
1.21775770e+00 -7.43623003e-02 -1.57934129e-01 1.03716433e+00
3.94697040e-01 2.59938259e-02 3.00503224e-01 -8.77196550e-01
-8.13722491e-01 6.16520107e-01 -5.54797828e-01 6.00091279e-01
-2.19760150e-01 -3.10670603e-02 1.13432014e+00 -9.36595976e-01
5.97136199e-01 1.13133836e+00 8.05128634e-01 5.87738931e-01
-1.04253089e+00 -7.54898250e-01 3.87200475e-01 4.19573009e-01
-1.50149834e+00 2.47318689e-02 5.25628984e-01 -2.45901436e-01
1.12043214e+00 1.82006389e-01 6.45408511e-01 9.41699564e-01
-3.73427831e-02 1.04298067e+00 1.31674767e+00 -1.85340360e-01
9.26164985e-02 -2.38498837e-01 3.30235451e-01 6.14288092e-01
1.68771431e-01 -1.09265417e-01 -4.57764238e-01 3.45365912e-01
7.21040010e-01 -1.22021392e-01 -9.57508832e-02 -3.41782905e-02
-1.04178238e+00 7.17826009e-01 1.08978760e+00 2.93561041e-01
-1.51679382e-01 4.83672470e-01 3.42367977e-01 -6.79258481e-02
4.75678205e-01 5.55649340e-01 -8.79582882e-01 1.16620593e-01
-1.12835491e+00 3.86673570e-01 4.15853411e-01 8.87331426e-01
7.23624825e-01 1.04299821e-02 -4.55718219e-01 9.05716896e-01
-2.75576822e-02 2.94416994e-01 2.41790280e-01 -1.16438138e+00
4.74440575e-01 6.00574434e-01 -1.17533505e-01 -4.67291832e-01
-5.31059563e-01 -9.32420015e-01 -5.57238758e-01 1.87147632e-01
4.57776487e-01 -1.67427853e-01 -1.54138613e+00 1.59229672e+00
9.41669270e-02 6.24941409e-01 -4.05887216e-02 9.84699786e-01
1.16309857e+00 4.52339560e-01 3.69879901e-01 6.35089815e-01
1.47877562e+00 -1.47029233e+00 -4.60822314e-01 -4.66234446e-01
5.85748732e-01 -7.20005035e-01 1.13325441e+00 3.58003289e-01
-8.34457099e-01 -6.33351624e-01 -1.20022666e+00 -3.56231302e-01
-6.81361914e-01 3.82095903e-01 8.36522937e-01 5.62834322e-01
-1.36495662e+00 6.54388487e-01 -6.59448683e-01 -3.24658126e-01
1.18215609e+00 3.81517231e-01 -2.26220742e-01 -2.36553788e-01
-1.15285349e+00 5.54094076e-01 4.52571660e-01 2.02877611e-01
-9.69103158e-01 -1.05401027e+00 -8.38236749e-01 9.29256994e-03
3.77059817e-01 -3.80563557e-01 1.22504008e+00 -1.02593112e+00
-1.14804780e+00 9.80271399e-01 -1.02138519e-02 -8.36960137e-01
5.19120634e-01 -6.01348162e-01 -3.18922848e-01 4.03097868e-01
1.59104183e-01 1.46486974e+00 2.83053517e-01 -1.05866539e+00
-9.45283532e-01 -2.08226740e-01 1.40617698e-01 1.65896654e-01
1.38666723e-02 -7.76180103e-02 -6.55629694e-01 -7.81363487e-01
3.39340302e-03 -8.69191229e-01 -3.00325453e-01 2.25809645e-02
-3.24792713e-01 -3.32174361e-01 1.11494291e+00 -6.37110233e-01
9.29384768e-01 -2.01428604e+00 -2.68816650e-01 -2.30989188e-01
-1.50261015e-01 6.87721968e-01 -4.02433455e-01 -1.58796102e-01
-8.54510888e-02 3.19994956e-01 -7.25806832e-01 -4.91723865e-01
-2.19722539e-01 9.40796137e-02 -2.19251603e-01 2.58162320e-01
5.85105419e-01 1.32284129e+00 -6.63951993e-01 -2.33576745e-01
3.13354641e-01 6.92304611e-01 -7.48696864e-01 -2.39915431e-01
-2.97083229e-01 4.58034277e-01 -1.25757664e-01 7.43977487e-01
8.48401666e-01 -2.11391166e-01 -2.30877951e-01 -1.88250497e-01
-1.63748860e-01 3.02131265e-01 -9.16074872e-01 2.28356624e+00
-3.62113595e-01 9.99639571e-01 1.32134119e-02 -1.07398391e+00
7.17424810e-01 -4.35688794e-02 2.02638999e-01 -9.15590227e-01
1.72437489e-01 8.57557729e-02 -2.65799463e-02 -1.79940835e-01
4.79311466e-01 3.69633317e-01 -1.22442462e-01 -2.25355387e-01
3.22111547e-01 2.21327245e-02 2.77821243e-01 6.12677708e-02
8.96868229e-01 4.29274440e-01 -2.51588672e-01 -6.38526499e-01
4.45930809e-01 1.26372248e-01 5.12879133e-01 8.92521262e-01
-5.24044812e-01 1.06428134e+00 4.95280802e-01 -5.03751397e-01
-7.89089382e-01 -9.94203985e-01 -4.72449273e-01 1.10670555e+00
2.58766860e-01 -2.95343637e-01 -1.17694020e+00 -8.11555088e-01
-1.04467377e-01 6.13251567e-01 -1.04492164e+00 1.14347145e-01
-7.57212102e-01 -8.69653404e-01 7.95688629e-01 9.57283676e-01
1.29135656e+00 -1.11914909e+00 -6.30977213e-01 1.38880312e-01
-7.20607936e-02 -1.61871815e+00 -2.66952932e-01 3.22615653e-01
-8.69468927e-01 -1.03293824e+00 -7.99599409e-01 -1.01353514e+00
3.93097669e-01 1.01628058e-01 1.11047852e+00 -7.49947131e-02
-6.80457294e-01 -3.89690436e-02 -3.56172770e-01 -2.84224063e-01
3.88203830e-01 4.56269383e-01 -8.49737406e-01 -2.89860487e-01
4.69856441e-01 -2.11823747e-01 -1.05472600e+00 3.46471578e-01
-8.45399737e-01 2.27989510e-01 5.31488657e-01 7.42874384e-01
9.44020808e-01 -1.82620957e-01 6.09440267e-01 -1.15961039e+00
-3.52271944e-02 -2.63990104e-01 -4.48035985e-01 1.01339854e-01
-4.63698179e-01 -7.33748451e-02 5.27721465e-01 9.35243517e-02
-1.17850840e+00 1.42613366e-01 -4.05770063e-01 -1.07104689e-01
-4.94262457e-01 8.93388838e-02 -3.56747955e-02 -1.24345168e-01
5.62236249e-01 -1.74775556e-01 -7.48711288e-01 -5.84272206e-01
4.79900956e-01 5.16205609e-01 9.08412218e-01 -5.58444023e-01
2.18007848e-01 6.38105333e-01 -5.83011135e-02 -5.32927573e-01
-1.31147778e+00 -6.54213428e-01 -7.51017988e-01 2.11702496e-01
1.29023623e+00 -1.16312420e+00 -2.96956748e-01 8.12999666e-01
-9.53720987e-01 -9.12205279e-01 -2.04061329e-01 1.44965410e-01
-3.12818259e-01 -1.71826169e-01 -6.83877587e-01 -1.41777709e-01
-3.42329890e-01 -1.22292781e+00 1.31570494e+00 5.19110978e-01
6.51510283e-02 -8.51805031e-01 -3.72410417e-01 7.17516422e-01
4.92796928e-01 3.88019174e-01 5.09342492e-01 -3.76523703e-01
-6.93663418e-01 5.20890281e-02 -9.06123698e-01 4.97304589e-01
-2.55684018e-01 -2.80170858e-01 -1.48059309e+00 -1.10672072e-01
-4.99321938e-01 -3.31587195e-01 1.68298709e+00 6.53453052e-01
1.91112018e+00 2.48516753e-01 -3.04915547e-01 1.08913851e+00
1.58979058e+00 4.00128886e-02 6.79959595e-01 4.86872077e-01
1.02498555e+00 4.52991039e-01 6.86027646e-01 3.83675210e-02
4.88778472e-01 4.56569076e-01 6.30040765e-01 -2.86851466e-01
-8.63712966e-01 -4.98178564e-02 -2.31808305e-01 5.57211526e-02
6.45359680e-02 -9.23657939e-02 -1.01958537e+00 9.10813391e-01
-1.66809618e+00 -4.72550958e-01 -2.04327479e-01 1.66992354e+00
8.19691002e-01 2.19923839e-01 -1.25313565e-01 -9.17027146e-02
6.75517023e-01 3.18486273e-01 -6.48450136e-01 -5.88960767e-01
-2.82725662e-01 7.41562426e-01 1.03651571e+00 5.51594138e-01
-1.74307489e+00 1.58283377e+00 6.09212732e+00 1.00875151e+00
-1.21073210e+00 2.81749547e-01 1.03509212e+00 -1.22510634e-01
-4.56146151e-02 -1.82185441e-01 -1.02530527e+00 3.12513441e-01
9.24447417e-01 6.02199078e-01 2.08204880e-01 8.88855100e-01
-8.35693255e-02 -3.66480529e-01 -7.94259071e-01 8.76368046e-01
3.27512212e-02 -1.58454525e+00 -9.95630175e-02 -1.06776610e-01
1.26564956e+00 6.04137123e-01 3.57257426e-01 3.71235311e-01
1.51827678e-01 -1.41975009e+00 8.05647790e-01 1.62942365e-01
9.89253402e-01 -8.56931031e-01 8.62964153e-01 -1.10367052e-01
-1.30749321e+00 -1.14507198e-01 -3.70526344e-01 -1.01097964e-01
-1.08817399e-01 5.11873722e-01 -6.33050382e-01 3.55996192e-01
1.17801785e+00 8.56217384e-01 -9.82123435e-01 1.28186798e+00
-4.74076241e-01 8.83398294e-01 -2.96653539e-01 2.69510299e-01
8.53709161e-01 2.21007764e-01 1.16392523e-01 1.71079814e+00
1.88843943e-02 -6.38518929e-02 6.75091445e-02 9.02600527e-01
-3.84111762e-01 -1.45886019e-01 -6.34055287e-02 2.40531147e-01
2.51558334e-01 1.32107067e+00 -1.13036203e+00 -3.47906649e-01
-3.28247488e-01 1.15444374e+00 4.38021570e-01 6.53240442e-01
-9.75625217e-01 -5.76991081e-01 7.05525279e-01 -2.93609649e-01
7.47599602e-01 -2.18112096e-01 -9.06187594e-01 -8.35116208e-01
-1.52575925e-01 -3.85065913e-01 3.85520935e-01 -5.93895137e-01
-8.23335648e-01 7.25749433e-01 -3.97530347e-02 -5.45050323e-01
4.61144149e-01 -1.00064981e+00 -5.46028733e-01 8.86763215e-01
-2.03125572e+00 -1.27038813e+00 -5.17594218e-01 3.92034233e-01
7.85001576e-01 7.56681338e-02 6.18448377e-01 5.48191428e-01
-6.82555854e-01 7.35885739e-01 -1.02598317e-01 3.22873473e-01
6.46273375e-01 -1.55231917e+00 7.56109595e-01 9.60978091e-01
1.48037806e-01 3.48209918e-01 2.36584842e-01 -4.51151788e-01
-6.40827239e-01 -1.57800543e+00 5.57868063e-01 -4.52423841e-01
4.31491345e-01 -6.08243763e-01 -7.70295143e-01 5.92884004e-01
2.44320616e-01 6.30199492e-01 4.16093379e-01 -8.15509539e-03
-4.18338686e-01 -8.73941481e-02 -1.23756194e+00 4.94791180e-01
1.35985100e+00 -4.80357707e-01 -4.69180793e-01 4.05144803e-02
1.12764060e+00 -6.13652050e-01 -5.55145383e-01 6.76292956e-01
4.08202022e-01 -9.06311154e-01 1.06995213e+00 -2.73256361e-01
5.51459670e-01 -3.68506402e-01 -2.49510333e-01 -1.04699898e+00
-2.68842936e-01 -2.31265072e-02 2.49936730e-01 1.10268390e+00
4.36054915e-01 -5.03700316e-01 1.05979860e+00 1.70061722e-01
-5.22123039e-01 -1.16023135e+00 -8.60017896e-01 -6.25092387e-01
3.94327015e-01 -6.39370978e-01 5.34277856e-01 7.38663137e-01
-9.69129324e-01 8.63856170e-03 1.15859203e-01 4.41744402e-02
4.88441139e-01 1.02179646e-01 2.13282019e-01 -8.67711604e-01
-4.44827192e-02 -5.96883297e-01 -3.53928596e-01 -1.02262211e+00
2.70548373e-01 -1.10417295e+00 -3.65911163e-02 -1.75580907e+00
1.38110101e-01 -4.92505491e-01 -7.03645229e-01 6.96886539e-01
-3.78354102e-01 9.72267926e-01 2.22073317e-01 -2.42716819e-01
-9.01138365e-01 3.32979679e-01 1.43294001e+00 -1.96612105e-01
1.22791044e-01 -3.99589926e-01 -7.93297589e-01 8.00548613e-01
9.85255420e-01 -2.53015757e-01 -3.97273362e-01 -7.86419630e-01
-2.16185525e-01 -8.20195615e-01 6.69758737e-01 -1.29230118e+00
-7.51451850e-02 1.47896975e-01 7.70362437e-01 -7.95709133e-01
1.62867725e-01 -6.99274659e-01 -3.77578765e-01 3.49622786e-01
-4.38215196e-01 -3.97300988e-01 6.88056409e-01 4.85743165e-01
-3.90085191e-01 -1.15474708e-01 8.65942061e-01 -1.55438080e-01
-1.40395451e+00 3.93354356e-01 1.80787928e-02 3.62749279e-01
8.54828119e-01 -1.78593948e-01 -6.15720212e-01 2.77024418e-01
-7.17647970e-01 4.25310761e-01 2.26331234e-01 4.89774436e-01
3.02404821e-01 -1.04268539e+00 -5.65667033e-01 6.77050874e-02
1.27136416e-03 5.11499286e-01 4.02931660e-01 5.81834674e-01
-8.98769557e-01 6.68017924e-01 -1.97083905e-01 -8.93010378e-01
-1.03823400e+00 -5.91578260e-02 5.21767497e-01 -1.07650816e-01
-6.21651113e-01 1.57913971e+00 5.88899076e-01 -3.90440494e-01
2.23329902e-01 -6.65589631e-01 -2.60886371e-01 -7.05070347e-02
6.19156301e-01 2.48209611e-01 2.37747267e-01 -5.30058324e-01
-6.06488824e-01 7.58226752e-01 -2.39933252e-01 6.83343560e-02
1.30003858e+00 -6.36870787e-03 1.28379032e-01 7.65931532e-02
1.39269316e+00 -4.56425250e-01 -1.93430555e+00 -1.59347132e-01
4.97411899e-02 -3.34585726e-01 4.20073420e-01 -1.46476340e+00
-1.36584079e+00 1.12819493e+00 9.73472655e-01 -3.95837605e-01
1.17816055e+00 1.30529478e-01 1.05549395e+00 1.64514899e-01
1.51047885e-01 -1.13943541e+00 -5.12451380e-02 7.79832959e-01
7.48930156e-01 -1.39638638e+00 -1.62824750e-01 -6.04856312e-01
-5.69443643e-01 8.37093353e-01 9.18313861e-01 -5.39130867e-01
4.88683313e-01 1.98569357e-01 1.07701816e-01 -1.87000647e-01
-3.77108663e-01 -4.68388677e-01 6.14319205e-01 5.29957056e-01
6.11442029e-01 3.35365236e-01 -1.41973764e-01 7.03159034e-01
-3.46878558e-01 -1.18019626e-01 2.13062868e-01 6.48057938e-01
-4.58033353e-01 -9.00412261e-01 -6.74830973e-02 1.67448193e-01
-8.86005700e-01 -4.55477864e-01 -3.63277376e-01 8.01878154e-01
5.80941021e-01 8.07855189e-01 3.53635579e-01 2.40594633e-02
2.74200499e-01 9.32179987e-02 5.03119469e-01 -7.21748948e-01
-5.82695901e-01 1.01259584e-02 8.47554505e-02 -8.73748302e-01
-4.10774082e-01 -5.06265223e-01 -1.64253044e+00 1.44957572e-01
-1.18884891e-02 -2.91718751e-01 7.79518485e-01 9.12061989e-01
4.02659893e-01 9.82676327e-01 3.77707630e-02 -7.86449015e-01
2.27775201e-01 -8.76981914e-01 -3.33566755e-01 3.25705230e-01
3.62049401e-01 -5.70221484e-01 -3.54770757e-02 -9.01630819e-02] | [9.545992851257324, 0.28491559624671936] |
8e63f22d-b7ca-4e9b-b01b-dbc871aace51 | cd-fsod-a-benchmark-for-cross-domain-few-shot | 2210.05311 | null | https://arxiv.org/abs/2210.05311v3 | https://arxiv.org/pdf/2210.05311v3.pdf | CD-FSOD: A Benchmark for Cross-domain Few-shot Object Detection | In this paper, we propose a study of the cross-domain few-shot object detection (CD-FSOD) benchmark, consisting of image data from a diverse data domain. On the proposed benchmark, we evaluate state-of-art FSOD approaches, including meta-learning FSOD approaches and fine-tuning FSOD approaches. The results show that these methods tend to fall, and even underperform the naive fine-tuning model. We analyze the reasons for their failure and introduce a strong baseline that uses a mutually-beneficial manner to alleviate the overfitting problem. Our approach is remarkably superior to existing approaches by significant margins (2.0\% on average) on the proposed benchmark. Our code is available at \url{https://github.com/FSOD/CD-FSOD}. | ['Wuti Xiong'] | 2022-10-11 | null | null | null | null | ['cross-domain-few-shot'] | ['computer-vision'] | [ 1.52821332e-01 -2.43008822e-01 -3.26053768e-01 -1.51993826e-01
-1.08999348e+00 -4.83466089e-01 7.94231892e-01 -4.09705400e-01
-4.04988140e-01 6.12604141e-01 2.03566760e-01 1.30069092e-01
-1.66515335e-02 -5.47379375e-01 -5.51463902e-01 -4.71043915e-01
2.93006748e-01 3.51441830e-01 9.82623875e-01 -1.75555393e-01
3.82645369e-01 2.80047804e-01 -1.88345373e+00 4.17333752e-01
6.26139104e-01 1.02020025e+00 2.94029266e-01 6.40035331e-01
-6.04442358e-02 8.66406381e-01 -6.01999342e-01 -5.85830808e-01
5.52511394e-01 -4.08066511e-01 -7.93533444e-01 3.34845372e-02
8.93123269e-01 -5.07969975e-01 -7.55001307e-01 1.24168921e+00
6.74472094e-01 3.21526647e-01 7.74953187e-01 -1.37716126e+00
-7.74980664e-01 3.10278207e-01 -7.10380852e-01 6.84469402e-01
-6.06114510e-03 3.48341882e-01 1.17520070e+00 -1.11689603e+00
8.67293060e-01 1.27733445e+00 6.80742264e-01 8.14773738e-01
-1.25053120e+00 -6.16808414e-01 -3.67885679e-02 3.21696907e-01
-1.50569534e+00 -6.81584358e-01 3.70504022e-01 -5.17028928e-01
1.14185393e+00 6.48462474e-02 2.14650497e-01 1.35444009e+00
-3.52657363e-02 1.14347255e+00 1.06820166e+00 -5.05800843e-01
3.35280061e-01 6.88958541e-02 6.07020080e-01 7.52798915e-01
4.93227899e-01 3.09813231e-01 -5.70876181e-01 -2.42559046e-01
3.72478396e-01 3.62835154e-02 -4.41121534e-02 -3.13531846e-01
-9.27085400e-01 9.91315722e-01 2.81091630e-01 1.92790821e-01
-1.33304164e-01 1.84478790e-01 6.33980691e-01 1.39604127e-02
5.65211535e-01 4.19193983e-01 -3.45599174e-01 -2.49485806e-01
-9.27173734e-01 4.48815018e-01 7.27386355e-01 1.02839637e+00
6.10928178e-01 -6.75956979e-02 -6.83936596e-01 1.02229106e+00
-1.40678763e-01 4.31342989e-01 4.21398312e-01 -1.22486782e+00
1.53871939e-01 3.73118252e-01 2.76337951e-01 -6.37407005e-01
-2.27025107e-01 -1.36353955e-01 -3.59532893e-01 2.46575758e-01
2.96929806e-01 1.01638306e-02 -1.18848693e+00 1.66342449e+00
2.16115773e-01 4.53542352e-01 -2.50346273e-01 8.44582796e-01
1.07155442e+00 6.37621224e-01 2.13236809e-01 -1.98416367e-01
1.20413101e+00 -1.52285957e+00 -6.73026383e-01 -3.72541100e-01
4.87321675e-01 -6.74267232e-01 1.52168894e+00 3.09938639e-01
-8.99449110e-01 -6.19326055e-01 -1.04263341e+00 3.06803472e-02
-3.45680565e-01 1.70228794e-01 3.94378036e-01 4.71016169e-01
-8.75598192e-01 6.52244866e-01 -7.55615413e-01 -7.87026286e-01
8.24450374e-01 -2.31221080e-01 5.37544712e-02 -1.94380790e-01
-9.84490454e-01 9.26378548e-01 5.80480397e-01 -7.54045248e-01
-1.41251230e+00 -7.99499214e-01 -5.36128640e-01 -1.68354943e-01
8.30382645e-01 -4.72835958e-01 1.66647208e+00 -4.86664712e-01
-1.08619022e+00 1.15104353e+00 -1.68627053e-02 -5.08372784e-01
5.48375666e-01 -6.97215319e-01 -5.27200341e-01 2.23184064e-01
3.06743354e-01 6.29461408e-01 8.68345141e-01 -1.17036533e+00
-8.26610863e-01 -1.56110555e-01 1.68263644e-01 -1.92446202e-01
-3.45546305e-01 1.57416180e-01 -7.74520338e-01 -7.78580189e-01
-6.45273745e-01 -9.41017330e-01 -2.04563886e-01 1.40322760e-01
-5.16207814e-01 -2.36293316e-01 8.92217159e-01 -1.46921992e-01
1.59669256e+00 -2.24511313e+00 -2.70966768e-01 -5.17379642e-01
2.03929022e-01 7.30136931e-01 -3.65628809e-01 4.00611073e-01
2.58104682e-01 8.03297944e-03 -2.62115121e-01 -4.95274872e-01
1.56724244e-01 1.84808876e-02 -2.64634311e-01 4.46759552e-01
1.55286387e-01 8.66581082e-01 -8.85836720e-01 -5.18804669e-01
2.51391172e-01 7.14857057e-02 -3.21049511e-01 1.88533530e-01
-5.02159357e-01 -3.33606720e-01 -3.08075309e-01 7.58798063e-01
7.58012235e-01 -5.26318669e-01 -1.72894135e-01 -2.61747211e-01
-8.63139927e-02 1.29175097e-01 -1.36486495e+00 1.88574874e+00
1.78829744e-01 5.66448510e-01 -3.14568937e-01 -8.17806244e-01
6.82356894e-01 -1.31019279e-01 2.57047415e-01 -7.42644787e-01
2.56580114e-01 1.47499532e-01 -3.24116647e-01 -4.42222416e-01
5.37842214e-01 -1.25724018e-01 -2.72696931e-02 4.59489912e-01
5.43987334e-01 2.23083720e-01 5.60977817e-01 4.26170528e-01
1.48063207e+00 6.28252402e-02 6.32055104e-01 -4.74785864e-01
2.85885066e-01 4.10283953e-01 5.35324514e-01 1.20885944e+00
-7.81131268e-01 7.34123707e-01 3.11031759e-01 -2.92971194e-01
-9.73484576e-01 -1.20465660e+00 -1.39290348e-01 1.52551746e+00
4.85438526e-01 -6.14757895e-01 -7.85244524e-01 -9.61601079e-01
1.97492242e-01 8.83679867e-01 -8.47191215e-01 -3.08739513e-01
-1.12538084e-01 -8.13452303e-01 7.37908304e-01 6.26228034e-01
5.35929441e-01 -1.07523787e+00 -7.70110965e-01 4.45783362e-02
5.93910739e-03 -1.13321292e+00 -4.83640760e-01 2.23736718e-01
-5.94149828e-01 -1.12507272e+00 -7.43351996e-01 -4.53021675e-01
9.41798314e-02 8.77692103e-01 1.29278016e+00 -1.19360285e-02
-7.37708747e-01 3.24714929e-01 -4.93586779e-01 -5.85233510e-01
-2.82836974e-01 6.24821484e-02 1.35783643e-01 -3.12707305e-01
8.26832354e-01 -1.77815035e-01 -6.88186944e-01 4.42606747e-01
-8.76264095e-01 -2.76465774e-01 3.33147317e-01 7.85276890e-01
6.72139108e-01 -3.19857955e-01 7.11561322e-01 -1.07953525e+00
4.18173254e-01 -6.67129874e-01 -7.27687299e-01 3.87892306e-01
-8.45899045e-01 -1.61815956e-01 1.88932225e-01 -4.01094168e-01
-1.17802787e+00 -3.83913741e-02 8.19498226e-02 -8.53420734e-01
-3.04469049e-01 -7.05260262e-02 1.10423282e-01 2.30789185e-04
1.11966348e+00 1.20263465e-01 -2.20587656e-01 -7.43601143e-01
4.60167199e-01 5.98694384e-01 5.61845958e-01 -3.35074335e-01
7.00248361e-01 7.42011666e-01 -4.40000534e-01 -5.96544564e-01
-1.33458233e+00 -8.30760419e-01 -4.77584422e-01 -4.82725091e-02
7.35697627e-01 -1.01889360e+00 4.44613993e-02 5.69765449e-01
-9.63487923e-01 -4.50510204e-01 -4.15137619e-01 2.05555320e-01
-6.38307035e-01 2.93518245e-01 -6.78402424e-01 -5.56983829e-01
-4.72249895e-01 -8.57231140e-01 1.13989365e+00 2.59602547e-01
-1.56151906e-01 -6.31477833e-01 4.39536333e-01 2.92871118e-01
3.45189273e-01 1.21650681e-01 3.47150862e-01 -8.81826937e-01
-4.13328171e-01 1.19362853e-01 -4.60366488e-01 3.65743488e-01
9.97478738e-02 1.83537275e-01 -1.29984021e+00 -3.56296629e-01
-2.99843729e-01 -7.39105105e-01 1.48057342e+00 5.23104548e-01
9.71871674e-01 7.96982720e-02 -5.24256527e-01 5.55967987e-01
1.70198035e+00 3.43535356e-02 6.12436056e-01 5.41225612e-01
3.00095320e-01 1.97402313e-01 1.01223135e+00 8.21021140e-01
2.19788209e-01 7.76200593e-01 2.84940988e-01 1.69667944e-01
-5.86076081e-01 -1.35134026e-01 3.66848111e-01 1.35432228e-01
2.55504213e-02 -5.49770355e-01 -8.50062609e-01 7.92464852e-01
-1.99064636e+00 -1.25718021e+00 -1.42576054e-01 1.96581781e+00
6.28866494e-01 3.65152955e-01 6.10060930e-01 -1.51906520e-01
8.86152446e-01 3.99800628e-01 -6.86164379e-01 -1.27222434e-01
-9.55997854e-02 1.78286180e-01 3.63904059e-01 2.14469999e-01
-1.49822855e+00 1.20042896e+00 6.83196354e+00 1.36121202e+00
-7.29558051e-01 5.81516027e-01 2.30299786e-01 -5.18245339e-01
1.92538142e-01 -2.81380564e-01 -1.20817721e+00 4.84255791e-01
8.19938838e-01 -4.92211848e-01 3.91442239e-01 1.04746509e+00
-1.99333325e-01 -2.42899731e-01 -8.72803271e-01 8.93302917e-01
2.68699110e-01 -1.61057413e+00 -4.47265878e-02 -2.66642034e-01
9.74745095e-01 6.17997706e-01 -1.68915048e-01 6.42323971e-01
5.41669369e-01 -6.23740554e-01 7.67320752e-01 3.06669474e-01
7.47607708e-01 -4.36846465e-01 3.73499125e-01 2.92429656e-01
-1.20332074e+00 -2.24153876e-01 -5.66745400e-01 8.87673721e-02
1.46287560e-01 6.00258052e-01 -4.78812128e-01 2.97872305e-01
1.18028915e+00 7.12746143e-01 -8.10762882e-01 1.20247126e+00
2.77401302e-02 6.91472888e-01 -1.08573638e-01 1.00295343e-01
2.21000537e-01 3.04338932e-01 6.79526806e-01 1.52928996e+00
2.64867157e-01 1.25473827e-01 1.98148295e-01 9.78459060e-01
-1.64792746e-01 -2.97341406e-01 -6.89526439e-01 -1.46983936e-01
5.94950259e-01 1.27967000e+00 -7.32009590e-01 -6.18320763e-01
-7.58186042e-01 9.27094519e-01 5.18472433e-01 1.68936923e-01
-1.20652068e+00 -4.58265930e-01 8.60501885e-01 2.86469143e-02
7.87132442e-01 2.62976736e-01 -1.91808178e-03 -1.10883391e+00
-6.56500161e-02 -7.01773345e-01 8.95326436e-01 -6.21689320e-01
-1.62817419e+00 5.46818316e-01 5.87711781e-02 -1.29733253e+00
1.10599041e-01 -5.71618497e-01 -6.07120872e-01 2.45561153e-01
-1.63795435e+00 -9.28104579e-01 -5.59487045e-01 7.15443134e-01
8.07565391e-01 -2.59362519e-01 5.80688894e-01 4.84962821e-01
-6.33353889e-01 6.47230864e-01 3.53276372e-01 -5.68781532e-02
9.98196542e-01 -1.11155987e+00 6.76923394e-01 8.90342593e-01
3.31370771e-01 2.55543381e-01 6.26580656e-01 -7.07572281e-01
-1.01189518e+00 -1.34033120e+00 4.46390063e-01 -6.69740677e-01
7.51372039e-01 -3.47347498e-01 -9.15199280e-01 6.33760750e-01
1.83021992e-01 4.49294776e-01 5.75362742e-01 -5.72312064e-02
-5.22968948e-01 -1.25935022e-02 -1.28341353e+00 3.73993099e-01
1.57363474e+00 -4.17453766e-01 -8.31071377e-01 2.82723963e-01
8.01331937e-01 -1.16906464e-01 -5.59946716e-01 4.26849514e-01
4.11401898e-01 -1.28344095e+00 1.06106365e+00 -7.39206314e-01
5.55872679e-01 -2.33969897e-01 -4.46937203e-01 -1.09931004e+00
-6.90224946e-01 -4.89013433e-01 -7.21335888e-01 1.31451476e+00
1.67161301e-01 -3.01282734e-01 6.66726410e-01 5.16423322e-02
-9.87140313e-02 -6.10266566e-01 -8.40208590e-01 -1.43729043e+00
-1.26925379e-01 -3.92851770e-01 3.38820845e-01 7.34368920e-01
-2.14965418e-01 2.93555677e-01 -4.43243831e-01 4.00140584e-02
9.25912201e-01 2.78323114e-01 7.97813952e-01 -1.09076953e+00
-3.57538283e-01 -4.48486060e-01 -3.84577990e-01 -8.15099418e-01
-8.80280361e-02 -6.96035028e-01 1.63495809e-01 -1.41852784e+00
8.43041003e-01 -3.24378535e-02 -7.06838310e-01 5.85431457e-01
-4.95372534e-01 5.70121229e-01 4.54452157e-01 3.52800518e-01
-1.26219153e+00 5.17229617e-01 9.26309705e-01 -6.98515028e-02
4.94451448e-02 -1.84992000e-01 -7.38395989e-01 7.72894382e-01
8.23329091e-01 -6.96347296e-01 -3.97702575e-01 -3.38275194e-01
-4.28295493e-01 -3.94012958e-01 5.56452990e-01 -1.14715087e+00
6.03485554e-02 -2.25360662e-01 1.23906195e-01 -5.76726615e-01
3.04935426e-01 -3.66010010e-01 -1.48079187e-01 5.56513667e-01
-2.74836540e-01 -2.74666429e-01 2.99115032e-01 7.08450019e-01
-7.04488456e-02 -4.05578524e-01 1.18404078e+00 -2.84535766e-01
-1.40491259e+00 3.78651470e-01 -1.78411677e-01 7.10887313e-01
1.41174805e+00 -1.37452215e-01 -8.78774345e-01 -5.13138212e-02
-4.95194495e-01 1.00496486e-01 4.38075334e-01 6.08805716e-01
4.96716172e-01 -1.45703161e+00 -6.80409908e-01 -2.09916934e-01
6.70852840e-01 -5.48542798e-01 2.70748407e-01 8.12177360e-01
-1.80385202e-01 2.90344834e-01 -2.42588997e-01 -5.15157044e-01
-1.17745864e+00 7.71429777e-01 2.93805569e-01 -2.28609309e-01
-6.78984165e-01 1.07636356e+00 3.30809534e-01 -2.79404193e-01
4.64680612e-01 3.39572988e-02 1.12817012e-01 1.77926153e-01
8.80484343e-01 8.72345090e-01 4.80488362e-03 -3.36533308e-01
-5.75561762e-01 3.83260667e-01 -2.34091505e-01 2.21974105e-01
1.39299834e+00 -3.81183401e-02 4.81680185e-01 3.57715607e-01
1.04899526e+00 -3.71219784e-01 -1.38612688e+00 -6.43734813e-01
1.21041119e-01 -7.92766035e-01 1.10413827e-01 -1.02674782e+00
-1.03354776e+00 6.52362943e-01 7.68581212e-01 1.13838352e-01
1.11735451e+00 4.37118948e-01 7.55113125e-01 4.05456007e-01
4.62674379e-01 -1.22535789e+00 3.83587420e-01 5.14636636e-01
6.57982111e-01 -1.52174330e+00 1.98295504e-01 -2.74528563e-01
-9.04092908e-01 8.07967484e-01 9.74558711e-01 -3.35435241e-01
6.30843997e-01 4.36511278e-01 -7.38441497e-02 -3.61035198e-01
-9.23101127e-01 -6.58276260e-01 2.03679591e-01 4.94185030e-01
-1.24491509e-02 -1.25726432e-01 -2.25155652e-01 7.24860132e-01
4.43817586e-01 4.97559994e-01 3.24288309e-01 1.02130616e+00
-8.28763783e-01 -6.80702925e-01 -1.50446728e-01 6.42602623e-01
-3.65714252e-01 -7.37196133e-02 -5.10794401e-01 8.73809159e-01
1.21179685e-01 8.41521323e-01 -5.52064590e-02 -4.00594175e-01
4.96585846e-01 -7.16068521e-02 4.72827017e-01 -6.79297924e-01
-2.92537630e-01 -4.66330573e-02 2.06292197e-01 -9.26024735e-01
-4.82176632e-01 -7.02941597e-01 -1.05036747e+00 -3.58182847e-01
-4.05635744e-01 -2.64343917e-01 -8.19672644e-02 5.93040287e-01
5.14328539e-01 2.92784065e-01 4.41453040e-01 -7.91797817e-01
-9.20413971e-01 -9.55024540e-01 -7.68459141e-01 5.77117682e-01
2.77528316e-01 -1.02817440e+00 -6.05737746e-01 -1.35075056e-03] | [9.852492332458496, 2.246732234954834] |
363f7a70-5929-4aa0-992f-6e3c29d7dca9 | codexglue-a-machine-learning-benchmark | 2102.04664 | null | https://arxiv.org/abs/2102.04664v2 | https://arxiv.org/pdf/2102.04664v2.pdf | CodeXGLUE: A Machine Learning Benchmark Dataset for Code Understanding and Generation | Benchmark datasets have a significant impact on accelerating research in programming language tasks. In this paper, we introduce CodeXGLUE, a benchmark dataset to foster machine learning research for program understanding and generation. CodeXGLUE includes a collection of 10 tasks across 14 datasets and a platform for model evaluation and comparison. CodeXGLUE also features three baseline systems, including the BERT-style, GPT-style, and Encoder-Decoder models, to make it easy for researchers to use the platform. The availability of such data and baselines can help the development and validation of new methods that can be applied to various program understanding and generation problems. | ['Shujie Liu', 'Shengyu Fu', 'Shao Kun Deng', 'Neel Sundaresan', 'Nan Duan', 'Ming Zhou', 'Ming Gong', 'Michele Tufano', 'Long Zhou', 'Linjun Shou', 'Lidong Zhou', 'Ge Li', 'Duyu Tang', 'Daxin Jiang', 'Dawn Drain', 'Colin Clement', 'Ambrosio Blanco', 'Alexey Svyatkovskiy', 'JunJie Huang', 'Shuo Ren', 'Daya Guo', 'Shuai Lu'] | 2021-02-09 | null | null | null | null | ['code-translation', 'text-to-code-generation', 'code-search', 'code-search', 'cloze-test'] | ['computer-code', 'computer-code', 'computer-code', 'computer-vision', 'natural-language-processing'] | [-6.68363869e-02 -2.65296161e-01 -6.54597759e-01 -7.74810314e-01
-7.07821906e-01 -5.14096320e-01 4.90598351e-01 2.04189032e-01
-2.09246632e-02 3.38040888e-01 1.71631038e-01 -9.27461505e-01
6.74214363e-01 -1.01591969e+00 -9.47685361e-01 9.86658931e-02
-7.97975361e-02 1.48413882e-01 1.09532125e-01 -4.46398854e-01
5.38981318e-01 -1.86010942e-01 -1.43345845e+00 8.65957081e-01
8.68085146e-01 2.94943511e-01 3.69809747e-01 8.98435116e-01
-1.30041242e-01 9.94470894e-01 -6.17527783e-01 -5.49859226e-01
-7.64045194e-02 5.80854248e-03 -1.06290448e+00 -7.10377991e-01
4.23267990e-01 -3.28408390e-01 -1.52701154e-01 8.36899042e-01
4.54908937e-01 -1.25238866e-01 2.14842603e-01 -1.46429551e+00
-9.57248509e-01 9.93075430e-01 -6.69975340e-01 2.89102167e-01
4.76392955e-01 2.89469719e-01 9.99137640e-01 -8.07027400e-01
6.91218853e-01 1.23313892e+00 7.50416756e-01 7.68817723e-01
-1.22557747e+00 -8.17460656e-01 -3.45643997e-01 -8.02701190e-02
-1.10657012e+00 -4.35781866e-01 3.71920347e-01 -6.60222292e-01
1.63638866e+00 1.56956799e-02 4.22714889e-01 1.26742768e+00
2.71761745e-01 1.02361619e+00 8.34801197e-01 -5.71750939e-01
-5.90241961e-02 9.90254581e-02 7.17149377e-01 1.00826287e+00
1.30740911e-01 3.37019295e-01 -3.27403009e-01 -3.74904394e-01
4.05617028e-01 -4.01324987e-01 -2.27508053e-01 -2.72873621e-02
-1.32637000e+00 9.12261844e-01 4.05644506e-01 2.25921750e-01
5.82779199e-02 6.28122091e-01 8.33124161e-01 3.11597139e-01
1.86633959e-01 8.36534619e-01 -4.03597087e-01 -8.37718904e-01
-8.69428217e-01 7.23866761e-01 9.48968053e-01 1.61891711e+00
5.91218174e-01 2.93489248e-01 -3.14261794e-01 7.16887951e-01
2.21953809e-01 2.62191445e-01 4.90028560e-01 -8.89285266e-01
1.00336838e+00 7.19605327e-01 -1.15286000e-01 -7.85931170e-01
-1.41580880e-01 -1.46102116e-01 -2.11401910e-01 -1.51074231e-02
1.52301177e-01 -2.08842039e-01 -5.30508697e-01 1.63218200e+00
-3.55197638e-01 3.86647582e-02 9.94813368e-02 4.05724734e-01
1.16299331e+00 8.61310720e-01 6.66988268e-02 6.51957750e-01
1.06187367e+00 -1.44019306e+00 -2.69769579e-01 -5.15400231e-01
1.42081130e+00 -9.56722915e-01 1.59060407e+00 2.14302003e-01
-1.19553828e+00 -7.17107773e-01 -1.10036719e+00 -3.42055798e-01
-4.75913048e-01 3.11349422e-01 1.01895332e+00 8.20575356e-01
-1.44796896e+00 4.67342615e-01 -9.71533000e-01 -4.82102066e-01
4.34953779e-01 -3.35804336e-02 -3.68868820e-02 -1.74987644e-01
-7.96681941e-01 8.63728404e-01 6.50598288e-01 -2.90911317e-01
-9.73765433e-01 -1.11385143e+00 -9.73993301e-01 4.04280014e-02
-1.64486721e-01 -6.72536790e-01 1.82504547e+00 -5.51679611e-01
-1.09512186e+00 1.05329132e+00 -1.81811258e-01 -4.42794561e-01
1.33170858e-01 -1.46654844e-01 -1.85909003e-01 -4.51610178e-01
4.87011187e-02 7.71024287e-01 2.59580314e-01 -8.87259722e-01
-4.63636786e-01 8.18387121e-02 2.33303040e-01 -2.00592622e-01
-3.83487433e-01 4.83304948e-01 -4.72308457e-01 -5.83640814e-01
-7.07005680e-01 -8.94618809e-01 -1.93488595e-04 -3.49537253e-01
-4.48153794e-01 -3.20649087e-01 6.69281602e-01 -8.82962406e-01
1.38907218e+00 -2.09343266e+00 8.90582427e-02 -2.23004937e-01
2.18497172e-01 2.67359376e-01 -5.57328343e-01 5.61781108e-01
-2.96217442e-01 5.92539012e-01 -1.21402927e-03 -1.36002108e-01
2.60768026e-01 9.89160389e-02 -6.76234066e-01 -1.45473748e-01
3.02376986e-01 1.06859553e+00 -1.00090945e+00 -4.14863169e-01
-2.77731344e-02 5.27000949e-02 -1.16463268e+00 4.04586554e-01
-6.23200178e-01 1.02576129e-02 -4.67756301e-01 6.71311855e-01
4.62818235e-01 -3.00228357e-01 -2.31759131e-01 2.07044631e-01
-2.26359904e-01 6.76545978e-01 -4.99455392e-01 2.03281283e+00
-9.99209344e-01 1.02231753e+00 -2.20310703e-01 -5.86091101e-01
1.04085910e+00 1.35084063e-01 -2.45171234e-01 -6.39897883e-01
-2.27651685e-01 2.34357715e-01 4.61312979e-02 -5.90335608e-01
1.02964401e+00 5.64095199e-01 -3.86198610e-01 7.52818584e-01
-5.33116274e-02 -5.06034553e-01 7.34340370e-01 5.33490241e-01
1.24877012e+00 2.71108449e-01 -3.31140496e-02 -5.28644323e-01
2.33745396e-01 2.90674835e-01 1.70057356e-01 9.80857313e-01
1.61055356e-01 2.43289202e-01 5.77630162e-01 -6.13905668e-01
-1.28463292e+00 -7.57544875e-01 -1.60125554e-01 1.56960201e+00
-4.36848462e-01 -1.04640388e+00 -8.19368184e-01 -4.51123893e-01
-4.38948572e-02 1.15705121e+00 -5.05717874e-01 -3.67850035e-01
-7.47415781e-01 -9.19340372e-01 1.10270357e+00 9.77735341e-01
6.71249449e-01 -8.37773740e-01 -5.54189146e-01 7.84168988e-02
-3.85357291e-01 -7.25181162e-01 -5.59078872e-01 5.02209067e-02
-7.76137829e-01 -9.25027907e-01 -2.52133697e-01 -9.06840980e-01
4.05598640e-01 1.05365559e-01 1.98200166e+00 6.74742222e-01
-2.42266759e-01 2.82716423e-01 -4.27370191e-01 -4.89438534e-01
-1.00485396e+00 4.47296977e-01 -6.95817292e-01 -1.20076740e+00
4.68133926e-01 -2.51525402e-01 -9.55482945e-02 1.27928331e-01
-7.01093793e-01 4.89641875e-01 2.72076637e-01 8.71629894e-01
-2.27731708e-02 -1.68696910e-01 2.82986164e-01 -1.02302480e+00
8.86001587e-01 -6.84569001e-01 -8.51311028e-01 3.74454796e-01
-7.00122058e-01 1.78025961e-01 4.94499028e-01 -4.32050601e-02
-1.14465356e+00 -4.25138861e-01 -3.44561547e-01 1.53309345e-01
1.81139663e-01 9.08998013e-01 4.36091609e-02 -1.84154451e-01
1.08108270e+00 2.09430695e-01 -3.43381017e-01 -4.62940931e-01
5.27879357e-01 6.03524685e-01 5.24884403e-01 -1.25544405e+00
5.16628742e-01 -3.07204962e-01 -5.06402254e-01 -5.77532649e-01
-3.12581748e-01 -6.66251853e-02 -1.94778681e-01 2.74754524e-01
5.59792876e-01 -9.80880141e-01 -1.37426704e-01 5.95844090e-01
-1.49351633e+00 -9.78305399e-01 2.04902381e-01 -9.19793472e-02
-5.17300904e-01 7.18082488e-02 -9.42978740e-01 -1.46204978e-01
-4.08096552e-01 -1.74166071e+00 1.12351489e+00 1.46247625e-01
-5.45201242e-01 -1.24030578e+00 2.54449248e-01 3.48103106e-01
8.20090592e-01 5.06989621e-02 1.50031376e+00 -3.60414594e-01
-7.66315877e-01 -4.27861288e-02 -3.54534060e-01 2.57625163e-01
-4.66671109e-01 6.49940133e-01 -6.98442459e-01 -2.98216075e-01
-4.62493658e-01 -7.46958792e-01 7.16698825e-01 1.20886803e-01
1.68402195e+00 -1.81705534e-01 -6.26502812e-01 1.00911820e+00
1.32567692e+00 2.29940265e-01 8.67971957e-01 4.14060086e-01
6.36329412e-01 1.96479738e-01 2.93945938e-01 2.31951386e-01
8.32552016e-01 7.35546291e-01 1.77698448e-01 -9.71632451e-02
-1.32212386e-01 -4.50524271e-01 3.96641552e-01 9.44677293e-01
1.54331416e-01 -3.21856469e-01 -1.63962269e+00 5.98944068e-01
-1.60656357e+00 -6.91434264e-01 -5.94523072e-01 1.93756080e+00
1.17121279e+00 -3.16025759e-03 -4.75817546e-02 -4.12580341e-01
4.96019691e-01 1.69298738e-01 -5.74985705e-02 -9.72115993e-01
2.78568655e-01 5.60023129e-01 3.17271471e-01 2.67585516e-01
-8.09590757e-01 1.07664442e+00 7.56111574e+00 6.82243109e-01
-1.15710998e+00 5.54745458e-02 6.02622628e-01 2.72624165e-01
-5.32246530e-01 1.85068712e-01 -9.50455368e-01 4.84436572e-01
1.28164458e+00 -8.99396360e-01 5.28113842e-01 1.35899138e+00
-1.92518085e-02 -8.17185640e-02 -1.64717710e+00 6.00292981e-01
-1.55608609e-01 -1.60412526e+00 -1.23835988e-01 -4.00319189e-01
8.65034461e-01 4.11962837e-01 1.07459165e-01 9.63371277e-01
8.87774706e-01 -1.13885665e+00 6.19853795e-01 1.25067145e-01
7.28035092e-01 -4.43202138e-01 6.24103248e-01 1.89209327e-01
-1.05580926e+00 4.46144827e-02 -4.45439458e-01 -3.65382612e-01
-2.45648801e-01 3.62494260e-01 -1.02956355e+00 2.01151118e-01
7.02112138e-01 9.79270577e-01 -1.15755546e+00 9.71442461e-01
-2.71875083e-01 8.18859518e-01 2.21687630e-01 -2.45718077e-01
1.46231158e-02 2.04254240e-01 1.49368867e-01 1.52397823e+00
4.02938783e-01 -2.89141417e-01 2.50511497e-01 1.52322888e+00
-2.92924672e-01 2.28782501e-02 -8.57821643e-01 -5.57357311e-01
6.76285207e-01 1.04241443e+00 -4.58894163e-01 -5.24689496e-01
-7.65659988e-01 4.91390705e-01 4.16790485e-01 1.99866503e-01
-1.05671835e+00 -6.79819643e-01 5.58033645e-01 -7.36566558e-02
-1.84559330e-01 -4.56917763e-01 -5.48056841e-01 -1.26054847e+00
-6.48051724e-02 -1.42419457e+00 2.83379257e-01 -9.75872159e-01
-7.24829435e-01 6.82279050e-01 3.63404274e-01 -5.75373709e-01
-4.85422432e-01 -6.48723245e-01 -1.13267171e+00 1.07150710e+00
-1.32462335e+00 -1.01461518e+00 -5.12982607e-01 1.06010310e-01
4.35572356e-01 -3.17984283e-01 9.27536130e-01 3.99525464e-01
-8.38282108e-01 8.83851051e-01 2.71335337e-02 2.52810150e-01
7.44287074e-01 -1.45015752e+00 1.27241731e+00 9.86401141e-01
-3.03014040e-01 1.15975416e+00 6.24832749e-01 -5.25067985e-01
-1.71433973e+00 -1.39708412e+00 7.38647163e-01 -9.20009017e-01
9.03414071e-01 -4.89860654e-01 -9.57079232e-01 1.05266249e+00
3.31548899e-01 -3.07061076e-01 5.92930734e-01 3.05437237e-01
-6.63064420e-01 1.97458774e-01 -6.05950773e-01 5.32321274e-01
8.72033954e-01 -6.39734685e-01 -6.01011455e-01 5.31048357e-01
8.24268520e-01 -8.04703653e-01 -9.47353959e-01 2.64282376e-01
4.76272583e-01 -8.18204999e-01 7.62484133e-01 -7.39983439e-01
1.41752326e+00 1.65204987e-01 -1.07123651e-01 -1.46620762e+00
-7.51055824e-03 -4.33418930e-01 3.58514711e-02 1.31281435e+00
8.73710036e-01 -4.69971359e-01 6.84952796e-01 8.33767235e-01
-5.28954804e-01 -5.80284238e-01 -1.04840286e-01 -5.99569499e-01
7.05703616e-01 -6.19778156e-01 1.02266216e+00 9.04844284e-01
2.78391212e-01 1.40013680e-01 -1.32484827e-02 -2.36153618e-01
1.86269790e-01 4.17883605e-01 1.26561451e+00 -6.45553410e-01
-6.57423735e-01 -7.33938694e-01 -1.42832190e-01 -1.11222792e+00
3.86921704e-01 -1.52513504e+00 -6.36093616e-02 -1.18903077e+00
4.32800233e-01 -7.09091067e-01 2.64024079e-01 8.39189887e-01
-3.63482505e-01 -9.05288309e-02 5.32488003e-02 6.71353936e-02
-3.65360379e-01 3.69035631e-01 8.35700035e-01 -3.93343925e-01
-1.53047979e-01 -7.72787556e-02 -8.69650483e-01 5.54916024e-01
8.10468912e-01 -4.64632064e-01 -4.17386651e-01 -1.25791991e+00
3.76601756e-01 1.89840525e-01 2.96832502e-01 -1.05041814e+00
2.58090384e-02 -1.15964144e-01 -1.27488628e-01 -4.39126581e-01
-9.21843573e-02 2.99592745e-02 -1.37057574e-02 5.03648520e-01
-7.20320702e-01 7.21423209e-01 6.64340079e-01 -1.35647848e-01
-2.24805713e-01 -5.19519150e-01 5.53008556e-01 -2.73798048e-01
-1.09309220e+00 7.06082806e-02 -2.33025312e-01 2.94465691e-01
8.43106151e-01 1.79537430e-01 -1.01819932e+00 -2.01157838e-01
5.95899858e-02 5.01417100e-01 8.70977819e-01 8.01912606e-01
3.98507804e-01 -1.25099349e+00 -8.48640561e-01 2.21570149e-01
6.40176773e-01 -2.03281790e-01 -6.36169985e-02 4.85006183e-01
-9.38953578e-01 5.74896693e-01 -3.40410471e-01 -5.60238779e-01
-1.27287507e+00 4.43726510e-01 2.15828702e-01 -3.52720201e-01
-2.86716282e-01 1.05829346e+00 1.14955552e-01 -8.02943647e-01
-3.03079635e-02 -6.54002011e-01 2.15021282e-01 -7.37485111e-01
7.53310204e-01 3.13999414e-01 5.98599575e-02 -1.47132292e-01
-8.64331648e-02 -1.60894334e-01 -3.39123160e-01 2.62400359e-01
1.34234762e+00 5.66934347e-01 -5.75780809e-01 2.99460918e-01
1.25503695e+00 -2.34820262e-01 -6.85688317e-01 -9.86925736e-02
3.67381454e-01 -5.05330682e-01 2.26498380e-01 -9.65251684e-01
-9.39255655e-01 1.16053414e+00 3.80864650e-01 3.18161435e-02
8.61779571e-01 -2.50714719e-01 7.62766182e-01 4.98885512e-01
6.31676733e-01 -6.30021632e-01 1.58293188e-01 8.22667837e-01
8.29599559e-01 -1.18837762e+00 -3.18687797e-01 -3.08308393e-01
-1.95641324e-01 1.32542109e+00 1.08554852e+00 1.80093840e-01
3.40942800e-01 8.07406068e-01 -2.36826435e-01 -1.10359810e-01
-1.21910429e+00 3.81649107e-01 1.93158045e-01 6.77776158e-01
1.28964722e+00 1.16688050e-01 -2.51627833e-01 7.45059788e-01
-5.71780384e-01 4.33235914e-01 8.82903457e-01 1.18957305e+00
-4.33779182e-03 -1.65567720e+00 -2.20718026e-01 7.02407479e-01
-2.28873417e-01 -5.87762058e-01 -1.71330228e-01 9.90222394e-01
-2.13889822e-01 8.25442255e-01 -1.15924612e-01 -4.81331885e-01
1.29544482e-01 2.12269928e-02 6.05390251e-01 -1.03694975e+00
-7.04253197e-01 -6.45358026e-01 4.73758489e-01 -6.70445263e-01
1.06421068e-01 -2.70105451e-01 -8.85093331e-01 -7.98532784e-01
-2.73788333e-01 1.70409933e-01 6.30943179e-01 4.21502471e-01
6.28852725e-01 6.14368260e-01 2.21739784e-01 -5.93510151e-01
-6.20943248e-01 -8.82113874e-01 3.53977233e-02 2.76040316e-01
-2.02096432e-01 -3.30053657e-01 1.90669388e-01 3.23090881e-01] | [7.79059362411499, 7.813659191131592] |
762c9a35-caf5-416d-8134-0edda1460960 | openapmax-abnormal-patterns-based-model-for | 2307.00936 | null | https://arxiv.org/abs/2307.00936v1 | https://arxiv.org/pdf/2307.00936v1.pdf | OpenAPMax: Abnormal Patterns-based Model for Real-World Alzheimer's Disease Diagnosis | Alzheimer's disease (AD) cannot be reversed, but early diagnosis will significantly benefit patients' medical treatment and care. In recent works, AD diagnosis has the primary assumption that all categories are known a prior -- a closed-set classification problem, which contrasts with the open-set recognition problem. This assumption hinders the application of the model in natural clinical settings. Although many open-set recognition technologies have been proposed in other fields, they are challenging to use for AD diagnosis directly since 1) AD is a degenerative disease of the nervous system with similar symptoms at each stage, and it is difficult to distinguish from its pre-state, and 2) diversified strategies for AD diagnosis are challenging to model uniformly. In this work, inspired by the concerns of clinicians during diagnosis, we propose an open-set recognition model, OpenAPMax, based on the anomaly pattern to address AD diagnosis in real-world settings. OpenAPMax first obtains the abnormal pattern of each patient relative to each known category through statistics or a literature search, clusters the patients' abnormal pattern, and finally, uses extreme value theory (EVT) to model the distance between each patient's abnormal pattern and the center of their category and modify the classification probability. We evaluate the performance of the proposed method with recent open-set recognition, where we obtain state-of-the-art results. | ['Jianfeng Zhan', 'Zhifei Zhang', 'Suqin Tang', 'Li Ma', 'Wenjing Liu', 'Jiyue Xie', 'Xiuxia Miao', 'Xiaoshuang Liang', 'Xiangjiang Lu', 'Xianglong Guan', 'Yunyou Huang'] | 2023-07-03 | null | null | null | null | ['open-set-learning'] | ['miscellaneous'] | [ 2.90688187e-01 1.93712890e-01 -1.86420888e-01 -5.01806021e-01
-5.08971214e-01 -3.29555839e-01 1.47139147e-01 1.10800564e-01
-4.35847044e-02 8.84253204e-01 -1.05587818e-01 -8.28423202e-02
-5.60884178e-01 -6.24691725e-01 -2.73394346e-01 -7.78933227e-01
-2.02928230e-01 9.87723649e-01 3.13942991e-02 2.08366126e-01
1.34674594e-01 2.93614358e-01 -1.86537588e+00 4.09174860e-01
1.33555818e+00 1.39048851e+00 1.58082590e-01 2.35376075e-01
-3.32252622e-01 3.23327184e-01 -5.47991514e-01 -2.81895071e-01
3.09421957e-01 -2.62643576e-01 -8.48668337e-01 2.12870955e-01
4.52834159e-01 -2.22934440e-01 -2.51369268e-01 1.25112391e+00
5.43057561e-01 -2.57924974e-01 1.10979676e+00 -1.28904700e+00
-8.32687199e-01 3.31919082e-03 -2.57767141e-01 3.86326969e-01
2.68267035e-01 1.15924887e-01 8.92047584e-01 -5.60296178e-01
4.89028662e-01 1.02840185e+00 6.04084313e-01 8.25516820e-01
-1.10935092e+00 -3.53092670e-01 3.43480796e-01 8.74165177e-01
-1.32497954e+00 -2.82401145e-01 4.79354471e-01 -6.41288519e-01
5.08643866e-01 4.21562880e-01 1.04330230e+00 1.35658312e+00
2.95573980e-01 7.14479327e-01 1.26170063e+00 -1.72592610e-01
6.81539357e-01 -6.32229149e-02 5.41097045e-01 3.93152624e-01
3.29224050e-01 2.06321198e-02 1.33432671e-01 -5.32839119e-01
5.46554744e-01 6.73676729e-01 -5.37351310e-01 -2.78916627e-01
-1.12808859e+00 5.48103333e-01 3.99361610e-01 2.80495018e-01
-5.88485777e-01 -5.53536952e-01 1.13490134e-01 3.53420079e-01
3.76054734e-01 1.93150848e-01 -5.22431195e-01 3.66431065e-02
-1.02450442e+00 2.34764948e-01 8.99732530e-01 6.61467552e-01
1.34157300e-01 -4.98870641e-01 -2.39800900e-01 1.05898440e+00
5.66474676e-01 5.66359341e-01 9.53280807e-01 -6.12321854e-01
1.23985797e-01 9.92833316e-01 -1.38688505e-01 -6.79593027e-01
-2.91293919e-01 -6.74355030e-01 -1.15771353e+00 1.58806667e-01
6.13915205e-01 7.16313422e-02 -1.10738313e+00 1.36181021e+00
1.44205600e-01 3.97917360e-01 1.21491939e-01 8.40294600e-01
4.79707837e-01 2.44961068e-01 -3.12382102e-01 -6.63421333e-01
1.40690577e+00 -4.02462959e-01 -7.40367293e-01 -2.47960061e-01
4.55351353e-01 -1.16446733e-01 8.67382586e-01 6.46945059e-01
-5.86348355e-01 -5.36563694e-02 -9.10100698e-01 3.92831683e-01
-3.30507964e-01 -1.13506727e-01 3.06231797e-01 6.01949036e-01
-7.66235888e-01 3.83899629e-01 -9.34024513e-01 -5.01213193e-01
1.04992759e+00 3.16275030e-01 -1.77829504e-01 -4.67306972e-01
-9.93764579e-01 9.17071939e-01 5.10067726e-03 1.24871746e-01
-6.82476521e-01 -8.77720773e-01 -4.36872959e-01 -3.11931521e-02
3.39880168e-01 -1.01428664e+00 9.40421104e-01 -8.08882892e-01
-1.04642403e+00 9.07695353e-01 -2.17259377e-01 -5.42021096e-01
6.60211205e-01 -1.41982257e-01 -7.15581119e-01 3.99606079e-02
4.87286597e-02 4.61769283e-01 8.24963450e-01 -1.01437604e+00
-7.06789076e-01 -1.05456877e+00 -4.15722609e-01 -9.88120586e-02
-2.97476470e-01 -2.26660520e-01 2.27088511e-01 -6.68008864e-01
5.09604752e-01 -8.64986479e-01 -4.94872391e-01 5.29526114e-01
-5.02306819e-01 -3.25983346e-01 8.71508121e-01 -4.90386128e-01
1.17199826e+00 -2.12837172e+00 -9.13593993e-02 3.32847118e-01
7.55798578e-01 3.61606926e-01 3.70342314e-01 -3.37624550e-01
-2.34516650e-01 -4.45181206e-02 -7.13893116e-01 1.15397185e-01
-6.43163919e-02 5.72085261e-01 -3.56195897e-01 5.30460656e-01
4.81824428e-02 6.19977474e-01 -6.29119456e-01 -4.43333149e-01
3.07609811e-02 2.58360147e-01 -7.04528391e-01 1.69840291e-01
-3.54171246e-01 4.26448882e-01 -5.12021244e-01 8.83388937e-01
6.95538580e-01 -5.12347579e-01 1.38059869e-01 -5.55300042e-02
1.67655498e-01 -2.16548696e-01 -1.23723555e+00 9.91199613e-01
1.81045428e-01 2.17179120e-01 -3.85034382e-01 -1.40555215e+00
9.93354440e-01 3.26009929e-01 7.63038516e-01 -4.77931947e-01
2.36234561e-01 3.22843313e-01 3.82255346e-01 -7.08141863e-01
-4.52650040e-01 7.47422278e-02 1.76199242e-01 2.98822135e-01
-1.96153462e-01 4.78890985e-01 7.12693781e-02 -3.12997997e-01
1.25024557e+00 -4.62417096e-01 5.48950195e-01 -1.40646473e-01
3.35763037e-01 -1.22645153e-02 9.47884798e-01 6.13595545e-01
-6.38122022e-01 7.81784534e-01 4.98437047e-01 -5.31434417e-01
-7.44532704e-01 -1.42147636e+00 -8.57002258e-01 2.94364512e-01
-7.56517192e-03 3.45269218e-02 -7.40310192e-01 -8.52276385e-01
8.85356814e-02 5.25540233e-01 -6.66626096e-01 -3.80411357e-01
-1.81828335e-01 -1.09718907e+00 4.74536538e-01 4.88684595e-01
5.18394470e-01 -6.66908205e-01 -5.63419819e-01 -1.35055778e-03
-2.87038535e-01 -6.53695881e-01 -2.93298453e-01 1.14144586e-01
-1.06862462e+00 -1.51618183e+00 -1.06667435e+00 -9.28655207e-01
8.71055305e-01 -2.88782209e-01 8.70803952e-01 -2.25477219e-01
-5.94905496e-01 2.61792183e-01 -3.47573787e-01 -4.55608219e-01
-2.66618848e-01 -4.75558609e-01 4.16928440e-01 5.15381336e-01
6.06737018e-01 -7.60858297e-01 -8.53439271e-01 6.06115341e-01
-5.84143639e-01 -4.71835643e-01 8.12084675e-01 7.87922800e-01
1.01718938e+00 2.17570066e-01 8.40538740e-01 -5.77798963e-01
3.78799260e-01 -7.03825235e-01 -1.69107631e-01 4.80296433e-01
-9.79753733e-01 -8.32615346e-02 4.79289591e-01 -7.68314600e-01
-6.84924960e-01 3.83369364e-02 -3.08454912e-02 -5.95460534e-01
-6.82588875e-01 3.31636757e-01 -5.58181226e-01 3.53035837e-01
6.16248369e-01 4.55021292e-01 3.96742016e-01 -6.38910174e-01
5.90102188e-02 1.22660315e+00 6.52066231e-01 -2.24724337e-01
2.05162704e-01 4.54160959e-01 -1.55558750e-01 -7.97114611e-01
-9.97421324e-01 -4.11395520e-01 -7.16937840e-01 -4.31151390e-02
7.48795927e-01 -5.79324126e-01 -4.42896843e-01 6.34277999e-01
-8.34442794e-01 2.98141483e-02 -4.89396751e-01 5.05618751e-01
-5.02940357e-01 3.96961600e-01 -1.16527803e-01 -5.45613647e-01
-4.59834427e-01 -1.10705268e+00 6.75533712e-01 1.55296504e-01
-4.08646911e-01 -9.45809722e-01 1.06687874e-01 3.56705606e-01
6.04341701e-02 3.98277380e-02 1.20843375e+00 -1.27572906e+00
-1.83659136e-01 -2.25032941e-01 -2.49532267e-01 5.54713070e-01
3.65093917e-01 -4.52608228e-01 -7.17017531e-01 -3.14584449e-02
3.69020104e-01 8.67689252e-02 6.04837239e-01 6.35411799e-01
1.27723777e+00 -3.87611151e-01 -7.32907355e-01 3.91754895e-01
8.67630899e-01 3.86956990e-01 9.04943347e-01 2.88765043e-01
4.12711740e-01 3.05039614e-01 3.83803874e-01 7.49230683e-01
3.45711291e-01 5.49096584e-01 4.35635507e-01 1.70896307e-01
-1.50651872e-01 4.20469671e-01 2.36534178e-01 4.10425574e-01
3.19574215e-02 -3.72748561e-02 -9.02200997e-01 4.15171206e-01
-1.79463923e+00 -1.01445627e+00 -2.82634854e-01 2.37665081e+00
8.60384643e-01 1.22824743e-01 1.31586358e-01 3.97115260e-01
9.15544033e-01 -5.17095625e-01 -1.05402231e+00 1.66023195e-01
-2.35081151e-01 1.07460216e-01 1.04347467e-02 -2.45295644e-01
-8.59797060e-01 1.17677100e-01 6.88564825e+00 5.08234322e-01
-1.12309241e+00 1.39241442e-01 7.14521587e-01 -2.20481738e-01
6.52403980e-02 -3.11034232e-01 -7.96612322e-01 9.94402170e-01
7.42982745e-01 -9.19931009e-02 1.51948065e-01 9.09345925e-01
1.41384201e-02 2.44878173e-01 -1.68820465e+00 1.23076653e+00
1.55605391e-01 -9.33322728e-01 2.02295482e-01 4.02742982e-01
3.61949176e-01 -1.46761000e-01 1.53371319e-01 2.80840099e-01
-1.76312327e-01 -8.85018170e-01 3.52161407e-01 1.00144494e+00
7.56499887e-01 -4.13164228e-01 7.63776779e-01 2.36501783e-01
-6.66981697e-01 -4.69914734e-01 -1.96858853e-01 9.81855169e-02
-2.43252926e-02 9.53551471e-01 -8.44427764e-01 2.06295222e-01
7.69560575e-01 8.29972029e-01 -6.21239305e-01 1.58347976e+00
9.04674307e-02 7.34562814e-01 -3.84255558e-01 2.84771055e-01
-1.71699479e-01 -2.34460935e-01 7.97206640e-01 6.76226556e-01
5.61625302e-01 1.23853222e-01 3.81730050e-01 1.06529391e+00
3.25736821e-01 -4.76563573e-02 -4.11853015e-01 -4.76045720e-02
3.25514436e-01 6.32983446e-01 -5.33478200e-01 -2.81367093e-01
-4.01270270e-01 9.36383724e-01 -6.89919293e-02 1.10248037e-01
-5.51830947e-01 -3.36589456e-01 8.65107894e-01 3.38869750e-01
2.34729856e-01 3.27160954e-01 -4.81530190e-01 -1.16437364e+00
2.37393096e-01 -8.32141101e-01 9.70204711e-01 -6.06719196e-01
-1.88434303e+00 5.07534325e-01 -2.10560307e-01 -1.58700871e+00
-1.39891177e-01 -8.43317449e-01 -4.87231523e-01 5.25759578e-01
-1.13933837e+00 -7.69340873e-01 -1.25899822e-01 8.30605388e-01
3.63853037e-01 -3.62854272e-01 9.48601365e-01 5.19693553e-01
-7.70027339e-01 6.76716030e-01 5.13917208e-01 2.72197407e-02
7.62213051e-01 -1.21956170e+00 -1.87115058e-01 3.99549693e-01
-2.61580527e-01 3.54991674e-01 5.12750804e-01 -9.15696323e-01
-1.05189800e+00 -1.27434313e+00 6.79541707e-01 -5.06786823e-01
5.74507535e-01 1.81281175e-02 -1.23211360e+00 5.92317283e-01
-4.91086990e-01 5.28284848e-01 9.36856508e-01 5.93468435e-02
-2.15303436e-01 -1.67912394e-01 -1.63863468e+00 4.82707441e-01
1.25332391e+00 -6.51830509e-02 -1.19827938e+00 3.91669661e-01
2.72620738e-01 -2.43550558e-02 -1.10376811e+00 4.18730795e-01
6.38921857e-01 -5.70158839e-01 1.04346824e+00 -7.59421766e-01
2.17141971e-01 -1.47081539e-01 -4.87585187e-01 -1.46173894e+00
-3.50291103e-01 -7.89024904e-02 -4.53839481e-01 1.01524997e+00
1.34507909e-01 -1.04065645e+00 6.66563511e-01 5.99590182e-01
-3.17280948e-01 -1.08835864e+00 -1.19740212e+00 -1.01889575e+00
6.31672665e-02 -3.00545990e-01 4.71532822e-01 9.14538085e-01
1.22262642e-01 7.80204907e-02 1.87598526e-01 4.73135263e-01
7.95299053e-01 7.06114098e-02 -1.02419883e-01 -2.09642458e+00
-8.87040272e-02 -5.55416763e-01 -9.43686008e-01 -6.30274832e-01
6.79144030e-03 -1.12045169e+00 -6.55478686e-02 -1.54853725e+00
3.28174621e-01 -6.69689834e-01 -5.09436429e-01 4.96096015e-01
-1.52215511e-02 3.99838127e-02 -2.99269319e-01 5.89174926e-01
-5.34687817e-01 5.94354808e-01 1.08740151e+00 -4.26560879e-01
-3.59586000e-01 3.04387957e-01 -8.73809576e-01 9.24691677e-01
4.77219552e-01 -4.76203382e-01 -3.93183380e-01 2.89559364e-01
-1.21488072e-01 -1.72876820e-01 6.23416305e-01 -1.21220624e+00
-1.61800925e-02 -1.32568091e-01 5.73220611e-01 -5.55022776e-01
1.55764699e-01 -9.43158209e-01 1.90833788e-02 7.10449874e-01
-6.23264648e-02 -4.69008029e-01 -2.94510543e-01 7.94734478e-01
5.05522154e-02 -2.69563168e-01 8.11126292e-01 -4.93471101e-02
-5.55493832e-01 6.50307536e-01 -5.50302565e-01 2.61230737e-01
1.25458431e+00 -5.54131985e-01 -4.15702641e-01 1.80886373e-01
-1.28181875e+00 2.59961993e-01 2.44164914e-01 3.04724753e-01
8.11354041e-01 -1.24117243e+00 -5.98165333e-01 3.07817429e-01
4.16556269e-01 8.03571641e-02 2.88472533e-01 1.20341516e+00
-3.66495587e-02 1.15860388e-01 -2.81975210e-01 -9.91624296e-01
-1.17810893e+00 5.96270382e-01 5.38284957e-01 -9.51347426e-02
-1.01581013e+00 2.30996639e-01 3.04795355e-01 -4.05642211e-01
4.62991923e-01 -3.36167812e-01 -4.37381178e-01 1.72002524e-01
9.57603812e-01 4.81858224e-01 2.92529553e-01 -3.31370831e-01
-4.70008045e-01 3.34443390e-01 -4.73008692e-01 3.77459705e-01
1.42264211e+00 -1.63821623e-01 -7.67827928e-02 5.34631431e-01
9.11154628e-01 -7.70622075e-01 -8.94385934e-01 -2.48751476e-01
-1.38629347e-01 -3.67018610e-01 1.44703060e-01 -1.08625674e+00
-9.29240227e-01 5.78995526e-01 1.47270668e+00 2.02240482e-01
8.84367287e-01 3.04737419e-01 9.51564848e-01 5.99350870e-01
3.08160484e-01 -8.85792077e-01 2.27453597e-02 3.39686990e-01
8.99345398e-01 -1.09521568e+00 -2.97191381e-01 -3.27366412e-01
-4.21434075e-01 9.50720191e-01 4.99614388e-01 4.27167192e-02
9.64677274e-01 7.19835013e-02 1.11309096e-01 -1.62744850e-01
-6.00939453e-01 -8.93217251e-02 2.56999314e-01 1.09476364e+00
-1.34987429e-01 2.95437485e-01 -1.16547018e-01 1.34981346e+00
-6.23561293e-02 3.07787210e-01 3.32651228e-01 9.63577092e-01
-5.98697603e-01 -9.98786330e-01 -7.34042823e-01 1.32928813e+00
-2.58993208e-01 1.48249730e-01 -4.21080112e-01 1.64716169e-01
4.96319503e-01 7.72524655e-01 4.56167430e-01 -2.74715036e-01
3.50369513e-01 4.11048651e-01 6.80797517e-01 -5.64798713e-01
2.20527783e-01 -2.80797482e-01 -1.18775763e-01 -4.26094592e-01
-6.67042509e-02 -1.11143279e+00 -1.29113507e+00 8.86481330e-02
-1.70088589e-01 -1.02788173e-01 3.32632303e-01 1.33582091e+00
7.29744792e-01 4.92639869e-01 4.66138393e-01 -3.67577404e-01
-8.06918263e-01 -7.90819585e-01 -8.41862857e-01 4.06114876e-01
2.62486696e-01 -1.20354497e+00 -3.10717046e-01 2.54547298e-01] | [12.541380882263184, 3.0207161903381348] |
f05a8572-b13e-48f3-b058-e13766a8d6fc | refined-an-efficient-zero-shot-capable-1 | 2207.04108 | null | https://arxiv.org/abs/2207.04108v1 | https://arxiv.org/pdf/2207.04108v1.pdf | ReFinED: An Efficient Zero-shot-capable Approach to End-to-End Entity Linking | We introduce ReFinED, an efficient end-to-end entity linking model which uses fine-grained entity types and entity descriptions to perform linking. The model performs mention detection, fine-grained entity typing, and entity disambiguation for all mentions within a document in a single forward pass, making it more than 60 times faster than competitive existing approaches. ReFinED also surpasses state-of-the-art performance on standard entity linking datasets by an average of 3.7 F1. The model is capable of generalising to large-scale knowledge bases such as Wikidata (which has 15 times more entities than Wikipedia) and of zero-shot entity linking. The combination of speed, accuracy and scale makes ReFinED an effective and cost-efficient system for extracting entities from web-scale datasets, for which the model has been successfully deployed. Our code and pre-trained models are available at https://github.com/alexa/ReFinED | ['Andrea Pierleoni', 'Christos Christodoulopoulos', 'Joseph Fisher', 'Shubhi Tyagi', 'Tom Ayoola'] | 2022-07-08 | refined-an-efficient-zero-shot-capable | https://aclanthology.org/2022.naacl-industry.24 | https://aclanthology.org/2022.naacl-industry.24.pdf | naacl-acl-2022-7 | ['entity-typing', 'entity-disambiguation'] | ['natural-language-processing', 'natural-language-processing'] | [-7.15319574e-01 4.32393700e-01 -4.36113864e-01 -6.03582002e-02
-1.15466964e+00 -1.03705466e+00 7.90631115e-01 7.63036668e-01
-7.87100613e-01 1.15600240e+00 3.21537942e-01 -5.01795635e-02
3.90879102e-02 -9.64360774e-01 -7.50749171e-01 1.35411128e-01
-2.55736530e-01 9.55938935e-01 5.73692203e-01 -3.33010048e-01
5.36167212e-02 1.04606293e-01 -1.31588292e+00 -8.48459899e-02
1.07622993e+00 3.40815306e-01 -3.64495292e-02 6.06682479e-01
-4.90581930e-01 4.88978684e-01 -4.72709149e-01 -1.07503259e+00
-1.04333028e-01 3.60933781e-01 -1.32221174e+00 -8.15452456e-01
9.68042612e-01 -1.05746463e-03 -3.60987574e-01 8.70201707e-01
7.47995853e-01 -2.27977075e-02 4.67260927e-01 -1.07852793e+00
-1.01330984e+00 1.03471494e+00 -2.80886918e-01 2.70938843e-01
6.90763474e-01 -4.74365771e-01 1.43946815e+00 -1.14045703e+00
1.23259902e+00 7.81073868e-01 1.04296529e+00 4.75355953e-01
-9.31914508e-01 -6.50138676e-01 -1.85525849e-01 4.93724644e-02
-1.67074001e+00 -5.96316278e-01 -3.44868511e-01 -3.48428965e-01
1.63256538e+00 6.66509122e-02 1.81264028e-01 4.85618174e-01
-3.55820358e-01 4.76590604e-01 5.05655289e-01 -4.37901527e-01
-2.56651253e-01 1.27407953e-01 3.82363617e-01 8.76380324e-01
1.05207241e+00 -3.97283852e-01 -3.61055046e-01 -3.56370062e-01
2.89946109e-01 -5.67381740e-01 -1.41899779e-01 -1.82879046e-01
-1.41949475e+00 3.49575460e-01 5.55597544e-01 3.21118057e-01
-4.53931451e-01 1.96534976e-01 6.01767540e-01 1.98163856e-02
3.63761485e-01 9.72856045e-01 -7.95881867e-01 -2.03077346e-01
-8.19103539e-01 4.16587085e-01 1.29145002e+00 1.56598735e+00
8.67288470e-01 -6.11379743e-01 -6.43669143e-02 8.50082874e-01
1.36557594e-01 4.99729633e-01 1.79276884e-01 -8.74487996e-01
7.63729870e-01 7.76013494e-01 5.06078839e-01 -6.76707447e-01
-6.16374314e-01 -3.67956311e-01 -2.02142805e-01 -3.72561991e-01
4.34319317e-01 -2.48535037e-01 -7.15571105e-01 1.64797413e+00
6.68010354e-01 2.35317022e-01 1.11352578e-01 5.35562456e-01
1.42941773e+00 3.76021087e-01 6.25286341e-01 1.62643954e-01
1.89516425e+00 -8.81952226e-01 -7.00142264e-01 -1.91279873e-01
1.14676619e+00 -7.49035954e-01 5.61303377e-01 -4.55916494e-01
-1.27747273e+00 -2.04823241e-01 -8.65275860e-01 -7.07143068e-01
-1.05339956e+00 -6.29207166e-03 7.78711736e-01 3.01175386e-01
-9.91721511e-01 6.56425118e-01 -7.21587121e-01 -8.49335670e-01
2.55160928e-01 3.13253254e-01 -9.74989891e-01 6.76534744e-03
-1.64719737e+00 1.27077556e+00 8.25231910e-01 -4.01367426e-01
-1.93407014e-01 -1.10803938e+00 -1.07467341e+00 7.64162093e-02
4.10964638e-01 -9.99453008e-01 1.31435466e+00 1.03839271e-01
-5.38757205e-01 1.22015631e+00 -2.42053851e-01 -3.22690606e-01
2.64658984e-02 -5.10341048e-01 -7.67184556e-01 1.40298992e-01
6.92988336e-01 6.77726865e-01 -8.94425623e-03 -9.47359204e-01
-1.01922071e+00 1.42674632e-02 2.53251374e-01 1.15695946e-01
-3.37997228e-01 2.26697072e-01 -9.44459021e-01 -4.10568029e-01
-3.42612684e-01 -8.43898118e-01 4.21751887e-02 -3.85552853e-01
-5.25798857e-01 -5.94698310e-01 3.19311917e-01 -1.04922712e+00
1.70891130e+00 -1.73311543e+00 -1.09403424e-01 3.02050239e-03
4.74763989e-01 6.32121146e-01 -8.59816819e-02 8.65900934e-01
-7.26462156e-03 4.40970629e-01 -2.51547862e-02 -2.28373021e-01
3.35409701e-01 -6.20024838e-02 2.16344237e-01 1.49183586e-01
1.92344353e-01 1.35013199e+00 -1.36650980e+00 -8.55151355e-01
-3.62014055e-01 4.08903569e-01 -3.12121749e-01 1.18553981e-01
-1.19210973e-01 -2.67028391e-01 -4.63031173e-01 7.61817098e-01
4.92580950e-01 -4.39292938e-01 4.04250950e-01 -4.38368499e-01
-2.17969224e-01 7.69021690e-01 -1.29513133e+00 1.80057073e+00
-2.83131599e-01 4.58793670e-01 -1.01674870e-01 -1.89321503e-01
6.22313797e-01 5.98882914e-01 1.85037434e-01 -4.82989579e-01
-2.29976594e-01 6.73886359e-01 -5.65670252e-01 -4.95554775e-01
1.06350720e+00 4.24932629e-01 -5.40703475e-01 3.15394193e-01
5.36367416e-01 3.14177513e-01 9.49340045e-01 8.27982843e-01
1.40820825e+00 2.15370283e-01 5.84760368e-01 -2.30287343e-01
3.44238609e-01 3.07920456e-01 5.78503728e-01 6.07841074e-01
1.48593113e-01 3.14988196e-02 7.23195374e-02 -1.75816312e-01
-1.28248692e+00 -1.01011884e+00 -1.16490796e-01 1.13301671e+00
1.66519731e-01 -9.07014489e-01 -7.53545702e-01 -9.60222125e-01
4.62326288e-01 6.12460375e-01 -4.38325971e-01 3.04703325e-01
-6.67953014e-01 -3.78299892e-01 1.04702961e+00 6.17369533e-01
4.42078978e-01 -9.46303785e-01 -1.26957530e-02 3.90715212e-01
-5.23454905e-01 -1.35909295e+00 -4.47863221e-01 -4.49727140e-02
-3.01253945e-01 -1.20783222e+00 -6.00093842e-01 -9.91836727e-01
6.43068671e-01 9.30714980e-03 1.73898554e+00 3.83524179e-01
-3.47212493e-01 3.40713829e-01 -3.57137591e-01 -8.65055770e-02
-1.01442128e-01 9.01153803e-01 5.06563224e-02 -9.25630629e-01
7.34691739e-01 -2.32544646e-01 -3.80463511e-01 1.73000038e-01
-5.83216250e-01 -4.00911540e-01 4.74575102e-01 6.73921406e-01
4.07757610e-01 -2.22249225e-01 8.59669685e-01 -1.43017983e+00
4.38588619e-01 -8.91084671e-01 -4.67459679e-01 6.04673803e-01
-8.48915517e-01 1.35847628e-01 2.84782082e-01 2.71880925e-02
-1.16722989e+00 -8.18081349e-02 -3.20816696e-01 4.28062439e-01
-1.64848372e-01 8.03142667e-01 3.64495702e-02 -6.79413229e-02
6.75829768e-01 -5.22665977e-01 -6.12701833e-01 -7.29436994e-01
8.25731039e-01 6.23576999e-01 9.75359976e-01 -7.82115161e-01
1.11126196e+00 3.34492582e-03 -3.85080338e-01 -2.27872998e-01
-9.23024237e-01 -8.78134787e-01 -9.85513687e-01 1.93910569e-01
7.83724189e-01 -1.42289042e+00 -6.48845375e-01 2.89890856e-01
-1.03209257e+00 -1.05331980e-01 -1.42152563e-01 2.81703293e-01
-2.81119868e-02 2.26280540e-01 -9.58229244e-01 -1.49349004e-01
-6.58455431e-01 -1.99098572e-01 1.01390088e+00 4.88612086e-01
-4.18794364e-01 -1.19216406e+00 3.26341510e-01 3.53863806e-01
2.80157387e-01 1.95597455e-01 7.03888714e-01 -1.00925350e+00
-5.45293033e-01 -4.91957754e-01 -5.52625120e-01 -3.92762482e-01
-1.45032227e-01 2.24412069e-01 -4.07163382e-01 -2.22677574e-01
-1.36968899e+00 -3.04528505e-01 6.93810880e-01 -2.78451145e-01
1.48763850e-01 -3.66459846e-01 -9.27877367e-01 5.23080587e-01
1.61324084e+00 -2.86254525e-01 5.50348043e-01 9.50177729e-01
7.86246836e-01 3.41938913e-01 9.80131805e-01 2.02779606e-01
1.18958259e+00 7.74252415e-01 1.75939590e-01 -1.73405334e-01
-3.53241950e-01 -4.35728490e-01 -2.01312959e-01 8.70274305e-01
-1.55048952e-01 -3.30202162e-01 -1.15397084e+00 1.28259015e+00
-1.89177167e+00 -1.21850228e+00 -4.54728335e-01 1.98214376e+00
1.30834341e+00 -1.10665858e-02 1.59095496e-01 -5.53725362e-01
9.16334331e-01 -1.95759624e-01 -2.67013758e-01 -1.02897055e-01
-8.39142352e-02 3.45368832e-01 9.62059140e-01 5.90379834e-01
-1.37369192e+00 1.35654390e+00 6.01916838e+00 7.36537993e-01
-4.87927169e-01 2.40530357e-01 -2.29341298e-01 1.59790277e-01
-3.02109808e-01 1.50048569e-01 -1.54664052e+00 4.19302225e-01
1.05071938e+00 -6.56469047e-01 1.00209929e-01 7.59104550e-01
-5.89930058e-01 1.06914520e-01 -9.19613898e-01 3.60852897e-01
-2.95085341e-01 -1.68010879e+00 -3.09467405e-01 -7.37161338e-02
7.19318151e-01 4.97417718e-01 -6.39578700e-01 7.41256714e-01
9.57060993e-01 -8.10580969e-01 5.67803144e-01 4.22590703e-01
1.15086138e+00 -6.71620846e-01 1.02767909e+00 9.53250602e-02
-1.56910741e+00 2.59566098e-01 -1.76118895e-01 3.00775737e-01
4.97162700e-01 6.08468890e-01 -9.84769225e-01 9.34403241e-01
7.64212132e-01 6.10764384e-01 -7.06351340e-01 1.22608137e+00
-4.75622952e-01 2.55951345e-01 -4.31546569e-01 5.81970811e-02
7.56811798e-02 4.54252273e-01 3.74305695e-01 1.89917827e+00
3.47188145e-01 2.56062031e-01 -1.44068012e-02 2.87870347e-01
-7.50716746e-01 1.20806292e-01 -4.16204125e-01 -1.49854422e-01
1.51241112e+00 1.77509201e+00 -3.92307997e-01 -6.49610639e-01
-5.01687884e-01 8.15877438e-01 1.02151251e+00 9.40567777e-02
-8.76809299e-01 -1.20184481e+00 6.77211642e-01 -1.27024492e-02
5.63108027e-01 -2.45360896e-01 1.61791265e-01 -1.24058616e+00
2.84615830e-02 -4.90157694e-01 8.94055784e-01 -6.41466022e-01
-1.24627173e+00 5.84585667e-01 4.13384056e-03 -8.11735630e-01
-2.51458824e-01 -3.48525494e-01 -1.92146108e-01 8.62804055e-01
-1.83234930e+00 -1.29106832e+00 -3.44427794e-01 3.22207481e-01
-1.19132191e-01 1.64746001e-01 1.19546759e+00 9.78326857e-01
-6.08529329e-01 8.95156384e-01 8.96251574e-02 7.57538259e-01
1.17577267e+00 -1.63008440e+00 1.22959292e+00 9.14522588e-01
1.41600028e-01 1.11142576e+00 5.97238958e-01 -8.93097401e-01
-1.21251190e+00 -1.25422060e+00 1.93182909e+00 -8.54929447e-01
1.02425158e+00 -2.21764624e-01 -9.57654834e-01 9.84365165e-01
3.50511253e-01 8.57176259e-02 7.34795272e-01 6.32195055e-01
-8.94276917e-01 2.60832071e-01 -1.24961174e+00 3.41174245e-01
1.32544291e+00 -5.10391653e-01 -9.25745964e-01 3.38218480e-01
7.14588642e-01 -7.22299039e-01 -1.69610500e+00 3.52793366e-01
5.18956482e-01 -1.44677699e-01 1.19375944e+00 -6.96913600e-01
1.95112586e-01 -5.24046421e-01 1.55519471e-01 -1.32644641e+00
-7.67039299e-01 -5.10686934e-01 -6.92081809e-01 1.88236427e+00
1.01542377e+00 -7.62139559e-01 3.80230516e-01 7.83084989e-01
-4.28519920e-02 -3.84475231e-01 -5.90182900e-01 -9.21245217e-01
-8.74199495e-02 2.94600397e-01 7.35527933e-01 1.43191671e+00
4.96331871e-01 5.39760590e-01 4.71999012e-02 5.52846611e-01
7.07737982e-01 3.68387476e-02 6.95980132e-01 -1.49338782e+00
-4.51622996e-03 -1.55101061e-01 -3.56243104e-01 -7.82081425e-01
2.91045815e-01 -1.13071752e+00 -1.18554108e-01 -2.10427213e+00
1.84569016e-01 -9.17446017e-01 -1.78220600e-01 1.09068775e+00
-5.92372000e-01 5.69172084e-01 6.91809552e-03 2.79628754e-01
-1.07988942e+00 -1.32255733e-01 4.91992474e-01 1.44881517e-01
2.04303786e-01 -6.28899217e-01 -9.15948749e-01 3.90290558e-01
4.61797625e-01 -6.67914927e-01 2.89058089e-01 -5.55728495e-01
6.33639753e-01 -2.81231105e-01 -3.40696163e-02 -7.66939282e-01
6.14415944e-01 2.58789271e-01 1.52255639e-01 -3.78032982e-01
8.15594718e-02 -4.77226138e-01 3.04678321e-01 2.54704375e-02
-2.30089396e-01 1.37781292e-01 3.31856400e-01 1.93101034e-01
-2.62794614e-01 -4.15979385e-01 3.58079612e-01 -2.34470367e-01
-1.35809112e+00 4.11487281e-01 1.60445064e-01 7.88169026e-01
1.03761721e+00 2.11294279e-01 -1.09403348e+00 1.43176123e-01
-6.02327168e-01 5.84976256e-01 7.11718261e-01 4.34326798e-01
-1.14558727e-01 -1.41161764e+00 -8.26409459e-01 -5.01670539e-01
4.84121084e-01 -1.52149871e-01 2.65101297e-03 5.84409714e-01
-6.97281539e-01 5.32468736e-01 -9.45876315e-02 4.89989892e-02
-1.44929194e+00 4.67746884e-01 -4.87792045e-02 -4.86692697e-01
-5.90807498e-01 1.04236007e+00 -5.81833899e-01 -7.62525380e-01
-6.42215759e-02 3.27771336e-01 -2.35769108e-01 2.61047453e-01
6.41230643e-01 5.14293492e-01 2.96031266e-01 -6.77792907e-01
-7.69665837e-01 5.34364820e-01 -4.11273807e-01 1.55304432e-01
1.37914443e+00 -3.08920652e-01 -3.28622550e-01 -3.76459658e-02
1.01227021e+00 4.78463858e-01 -5.71691632e-01 -4.31678504e-01
6.41706586e-01 -1.10709339e-01 -2.04179436e-01 -1.17809117e+00
-6.28553033e-01 1.87540725e-01 -1.18559606e-01 3.48099262e-01
6.13071084e-01 3.94073159e-01 1.14661551e+00 6.25828803e-01
6.61179602e-01 -1.14908743e+00 -7.85111964e-01 7.75935531e-01
2.75605023e-01 -1.07256484e+00 1.91614360e-01 -6.85433030e-01
-4.65335160e-01 8.17099214e-01 6.38223171e-01 1.83502713e-03
3.88314754e-01 5.79327226e-01 -4.35978320e-04 -4.11363602e-01
-8.05809736e-01 -6.30803049e-01 3.78104508e-01 6.54252946e-01
7.72857428e-01 9.21477303e-02 -4.30183202e-01 3.98720324e-01
-1.07907474e-01 5.34017794e-02 2.69336611e-01 8.89771938e-01
-4.85303134e-01 -1.29420733e+00 2.07225040e-01 3.31445068e-01
-7.67609119e-01 -5.23115635e-01 -1.89742804e-01 1.06931829e+00
1.95603482e-02 6.89325511e-01 -6.02319278e-02 6.03387272e-03
6.01049006e-01 1.85228392e-01 3.08621287e-01 -8.84823143e-01
-8.69222403e-01 -7.09487498e-01 1.10083055e+00 -5.50448596e-01
-3.72139722e-01 -6.62827790e-01 -1.71406245e+00 -7.66296208e-01
-6.03860855e-01 5.40226996e-01 5.88148057e-01 7.11599171e-01
9.37751174e-01 1.45917803e-01 -7.71965133e-04 -5.11547863e-01
3.70616019e-02 -8.23279977e-01 -3.10790956e-01 5.34432113e-01
-1.69393364e-02 -7.55618036e-01 3.09186056e-02 6.91961721e-02] | [9.467817306518555, 8.896150588989258] |
e55923aa-6a3b-469f-ba10-eaf1f96033f6 | multiclass-mri-brain-tumor-segmentation-using | 2305.06203 | null | https://arxiv.org/abs/2305.06203v1 | https://arxiv.org/pdf/2305.06203v1.pdf | Multiclass MRI Brain Tumor Segmentation using 3D Attention-based U-Net | This paper proposes a 3D attention-based U-Net architecture for multi-region segmentation of brain tumors using a single stacked multi-modal volume created by combining three non-native MRI volumes. The attention mechanism added to the decoder side of the U-Net helps to improve segmentation accuracy by de-emphasizing healthy tissues and accentuating malignant tissues, resulting in better generalization power and reduced computational resources. The method is trained and evaluated on the BraTS 2021 Task 1 dataset, and demonstrates improvement of accuracy over other approaches. My findings suggest that the proposed approach has potential to enhance brain tumor segmentation using multi-modal MRI data, contributing to better understanding and diagnosis of brain diseases. This work highlights the importance of combining multiple imaging modalities and incorporating attention mechanisms for improved accuracy in brain tumor segmentation. | ['Maryann M. Gitonga'] | 2023-05-10 | null | null | null | null | ['tumor-segmentation', 'brain-tumor-segmentation'] | ['computer-vision', 'medical'] | [ 4.37069505e-01 4.59383696e-01 -2.38447800e-01 -3.05903673e-01
-1.05832374e+00 -5.03252372e-02 3.11332762e-01 1.43168885e-02
-3.88516873e-01 5.84498644e-01 4.64734823e-01 -3.68874133e-01
2.66905546e-01 -5.80905855e-01 -4.46881145e-01 -6.77210987e-01
-3.21768411e-02 4.06853825e-01 2.66849220e-01 -3.53675857e-02
4.74216370e-03 3.68602335e-01 -8.43073428e-01 3.60426039e-01
8.28560114e-01 9.95494545e-01 4.28395867e-01 4.87830311e-01
-1.27014279e-01 8.63394499e-01 -4.51931179e-01 -1.75457194e-01
7.45808557e-02 -2.06446379e-01 -1.18560469e+00 2.10057840e-01
7.60500371e-01 -6.15112603e-01 -3.09550405e-01 1.17854011e+00
6.22713506e-01 6.66715577e-02 7.12117314e-01 -7.19162762e-01
-9.24986184e-01 7.85021603e-01 -1.06414926e+00 9.71142352e-01
-2.29876220e-01 1.62931412e-01 8.31070602e-01 -4.19167757e-01
6.78726375e-01 7.96348333e-01 6.98159158e-01 8.52720916e-01
-1.12117124e+00 -5.72787523e-01 -4.16748673e-02 1.15643941e-01
-8.97569180e-01 -2.94991769e-02 5.16168535e-01 -4.89707321e-01
1.22935653e+00 3.03351581e-01 6.24101341e-01 1.06989145e+00
8.50330293e-01 1.20349312e+00 9.93242741e-01 -7.66337365e-02
-1.49256229e-01 -2.37785488e-01 6.39505148e-01 7.01503515e-01
1.87984556e-01 8.87517643e-04 1.15352929e-01 -3.30227241e-02
9.52157915e-01 4.18569930e-02 -3.73436660e-01 -1.61155909e-01
-1.38992655e+00 8.41160238e-01 1.10372865e+00 7.77015448e-01
-5.89574635e-01 3.34464818e-01 7.97414839e-01 -2.63057977e-01
9.10213053e-01 4.84330863e-01 -2.09179237e-01 3.62061441e-01
-1.15299594e+00 -1.69122040e-01 -3.04783612e-01 4.89274025e-01
-5.81354573e-02 2.99137175e-01 -5.43739676e-01 9.50716496e-01
1.70563117e-01 6.14737757e-02 1.24276161e+00 -5.65192580e-01
1.31673723e-01 5.46684325e-01 -6.84377313e-01 -3.19231004e-01
-9.60139632e-01 -8.80183756e-01 -9.24104691e-01 3.12593877e-01
2.56995648e-01 -3.56158555e-01 -1.70113468e+00 1.61738789e+00
2.29096681e-01 1.30871594e-01 -1.53243363e-01 9.78171468e-01
1.12690294e+00 1.85405836e-01 3.23458672e-01 1.91524163e-01
1.40379107e+00 -1.20465803e+00 -6.82035744e-01 -2.60134369e-01
9.09788668e-01 -2.04887912e-01 7.38569915e-01 -8.56334195e-02
-1.15275455e+00 -2.08039463e-01 -8.72824252e-01 -2.68364578e-01
-2.74886370e-01 -1.32949829e-01 8.09251904e-01 6.52905524e-01
-1.44824243e+00 1.86733529e-01 -1.02925670e+00 -1.15876459e-01
1.08735144e+00 5.44194996e-01 -3.35394025e-01 9.40196291e-02
-1.13120663e+00 1.14833343e+00 5.25874197e-01 -1.82258800e-01
-9.78199244e-01 -1.14017069e+00 -8.97242904e-01 5.71834259e-02
2.35556290e-02 -6.76695347e-01 1.30098379e+00 -1.05577326e+00
-9.09978032e-01 9.25083756e-01 6.54174481e-03 -8.00826073e-01
4.46348369e-01 2.10343242e-01 -1.83053195e-01 4.37629312e-01
1.31911218e-01 1.20493758e+00 5.23220122e-01 -1.08285439e+00
-4.53522891e-01 -8.19416940e-01 -2.67748475e-01 3.47799063e-01
-9.24717560e-02 -1.07574992e-01 -1.38408571e-01 -8.08514655e-01
1.19091369e-01 -6.77965105e-01 -5.06744683e-01 -3.74280304e-01
-3.67417008e-01 5.32542132e-02 1.05807161e+00 -1.33909309e+00
6.90762401e-01 -1.77960575e+00 1.79167271e-01 1.49436146e-01
7.70988405e-01 1.41780153e-01 -6.99965432e-02 -8.90296817e-01
-3.95529747e-01 3.50601345e-01 -5.11210501e-01 -1.08592510e-01
-5.39430439e-01 -1.04130395e-01 3.70386511e-01 4.20094699e-01
1.97812736e-01 1.39636087e+00 -7.38403261e-01 -5.20802259e-01
3.30380052e-01 6.02692723e-01 -5.57757795e-01 -3.16832691e-01
2.68503409e-02 5.99167883e-01 -2.45115772e-01 8.79742682e-01
4.91413325e-01 -2.60346204e-01 -1.09651208e-01 -2.01584727e-01
3.41439724e-01 -1.53708652e-01 -2.96299428e-01 1.73563063e+00
-3.23498458e-01 7.74662137e-01 3.17368031e-01 -1.07866669e+00
5.59183717e-01 5.41955054e-01 1.11027706e+00 -9.62948561e-01
7.43468583e-01 -1.09581560e-01 6.07535958e-01 -4.36750442e-01
3.32410008e-01 -2.38126665e-01 1.78923860e-01 4.23136681e-01
2.34673321e-01 -1.87477380e-01 1.24561645e-01 2.24121094e-01
1.11761677e+00 -3.46520334e-01 1.11085750e-01 -6.01992905e-01
4.52540189e-01 -5.80032319e-02 2.20148310e-01 5.46002865e-01
-8.90878916e-01 4.23680156e-01 3.30718309e-01 -2.08094999e-01
-9.55476582e-01 -9.34289336e-01 -5.13133526e-01 8.78622830e-01
-2.76590466e-01 3.00104350e-01 -9.52614427e-01 -8.94837976e-01
1.69071034e-02 6.45542741e-01 -1.30033040e+00 -2.47966722e-01
-4.08353299e-01 -1.21049178e+00 4.73751247e-01 9.23935533e-01
6.13670707e-01 -1.02484500e+00 -7.43887961e-01 1.49708256e-01
-3.80141109e-01 -8.55116546e-01 -7.25492954e-01 4.15438294e-01
-1.34662151e+00 -1.00617290e+00 -1.43536472e+00 -1.08100438e+00
9.81061041e-01 -9.90904402e-04 1.01255834e+00 2.59375628e-02
-5.85487008e-01 4.00734872e-01 -8.89233500e-02 -3.98688167e-01
-3.87650877e-01 2.15730101e-01 -3.20199192e-01 -4.63034481e-01
1.54579595e-01 -1.79300681e-01 -3.75740767e-01 -1.19848158e-02
-1.05168879e+00 3.43214929e-01 7.78892219e-01 1.14080811e+00
6.93313241e-01 -3.02188247e-01 5.62640011e-01 -8.73246193e-01
7.72026420e-01 -4.76408631e-01 -1.41296968e-01 1.85274303e-01
-3.35677385e-01 -1.75847318e-02 1.77610159e-01 -3.54850322e-01
-1.07843077e+00 -6.46767542e-02 -3.19445342e-01 -3.63163918e-01
-3.90324071e-02 3.74601245e-01 2.13540390e-01 -5.00208616e-01
5.07416844e-01 1.20215096e-01 4.29760158e-01 9.54376087e-02
1.72382161e-01 4.08862501e-01 6.21813953e-01 -1.63354725e-01
-3.73677351e-02 4.49516445e-01 -8.42369273e-02 -6.94063604e-01
-5.75463533e-01 -3.18578839e-01 -8.11976969e-01 -4.22980219e-01
1.23018444e+00 -5.27560830e-01 -2.56292850e-01 5.03282726e-01
-6.92415178e-01 -5.61895430e-01 -2.03052118e-01 4.61538255e-01
-3.78835529e-01 1.39615282e-01 -1.13583398e+00 -1.46080449e-01
-9.04287875e-01 -1.80676496e+00 9.39434886e-01 3.39388967e-01
-2.51615673e-01 -1.11009884e+00 -1.92857876e-01 8.30928385e-01
6.55219674e-01 1.93146691e-01 1.00109458e+00 -6.51756704e-01
-2.06222802e-01 -1.28540888e-01 -4.93644476e-01 2.20210984e-01
1.72938585e-01 -2.69274056e-01 -9.21436429e-01 -2.27242962e-01
-2.44494140e-01 -2.85254091e-01 1.20075154e+00 1.20340109e+00
1.56999600e+00 2.33534619e-01 -4.70344692e-01 6.23568654e-01
1.19466627e+00 4.86349195e-01 4.43309188e-01 4.95921485e-02
8.36788356e-01 2.65132159e-01 -1.44223571e-01 -1.35268569e-01
3.68699521e-01 1.62509441e-01 6.99007452e-01 -5.75100780e-01
-5.37678242e-01 5.75064600e-01 -9.14268866e-02 6.44561291e-01
5.54693043e-02 2.05660522e-01 -1.22923613e+00 1.03791249e+00
-1.11977589e+00 -9.23296571e-01 -4.09164801e-02 1.79791188e+00
6.95608079e-01 3.62653211e-02 7.01636523e-02 -1.21364184e-01
7.19241381e-01 1.06426768e-01 -8.46391320e-01 -3.78327578e-01
-7.36185461e-02 2.60908484e-01 7.29895532e-01 6.34658635e-01
-1.26043582e+00 7.58161902e-01 7.51344013e+00 6.77767336e-01
-1.45900130e+00 5.99523485e-01 1.32342911e+00 -2.10321456e-01
-2.00671598e-01 -8.10952365e-01 -4.41778213e-01 2.88754582e-01
7.84318626e-01 -2.75131147e-02 -3.75068150e-02 6.44522488e-01
-5.34989405e-03 -3.69564176e-01 -6.62141502e-01 6.30477250e-01
4.03295696e-01 -1.53451288e+00 -4.35231514e-02 2.43805170e-01
7.70822406e-01 6.24126673e-01 2.32978657e-01 3.52061480e-01
3.34709078e-01 -1.02239084e+00 3.88522565e-01 2.91209131e-01
7.85684049e-01 -6.71712160e-01 9.61896956e-01 -1.50894254e-01
-8.54034245e-01 6.74859658e-02 3.63993198e-02 4.64629084e-01
7.01565668e-02 4.04670417e-01 -1.26216781e+00 2.58440554e-01
6.77491248e-01 4.10998255e-01 -5.91244042e-01 1.16393375e+00
3.37778360e-01 5.26791215e-01 -1.00737587e-01 1.80065528e-01
4.84702110e-01 1.12538092e-01 5.26469767e-01 1.18789363e+00
8.36938545e-02 1.38533711e-01 1.38459101e-01 8.29934776e-01
-2.36696377e-01 2.00147271e-01 -3.86737555e-01 1.12319805e-01
-2.68176228e-01 1.36113811e+00 -1.21310484e+00 -6.93989277e-01
-4.22263145e-01 8.61668408e-01 1.99324787e-01 1.37235984e-01
-1.00062883e+00 -3.08857616e-02 4.65643018e-01 -6.85031563e-02
-7.11415112e-02 -2.42654011e-02 -1.10961306e+00 -9.67206478e-01
-6.86789572e-01 -5.33642173e-01 4.81083214e-01 -8.28067601e-01
-9.65346932e-01 8.31128597e-01 -2.33319789e-01 -4.59876865e-01
1.34046271e-01 -6.15041673e-01 -6.63860559e-01 8.06153595e-01
-1.37540269e+00 -1.11299920e+00 -2.30553657e-01 5.28279722e-01
5.30376911e-01 4.49801572e-02 5.70708930e-01 3.15956354e-01
-7.99150109e-01 5.51361322e-01 8.51199497e-03 3.41280222e-01
5.35632193e-01 -1.38714290e+00 1.89173058e-01 8.60503852e-01
-2.73568928e-01 2.28039026e-01 3.92305911e-01 -1.02872014e+00
-6.71468318e-01 -1.09817541e+00 2.42460236e-01 -2.76499659e-01
6.53413236e-01 2.05808580e-01 -9.29113925e-01 9.07884061e-01
5.95848978e-01 7.45071694e-02 8.67277324e-01 -2.32838079e-01
1.62183657e-01 3.62379104e-01 -1.72316408e+00 5.18397808e-01
5.41003704e-01 -2.59021699e-01 -5.76068819e-01 3.98165494e-01
7.52779365e-01 -8.32735240e-01 -1.29854155e+00 5.28737783e-01
1.76206067e-01 -3.69886160e-01 1.04862249e+00 -5.46110034e-01
7.44289458e-01 3.57770681e-01 1.37202933e-01 -1.70655465e+00
-6.79966986e-01 3.83342862e-01 5.74133061e-02 5.49779475e-01
6.89088345e-01 -4.72101718e-01 9.11147594e-01 9.14238572e-01
-6.00553334e-01 -1.16150224e+00 -9.52297270e-01 -1.72799960e-01
5.90900898e-01 -4.46130574e-01 3.33478510e-01 1.02181852e+00
-5.41881099e-02 -2.73300875e-02 -4.16451357e-02 -1.56385243e-01
7.21127927e-01 -4.20058340e-01 -1.59657940e-01 -9.69166398e-01
1.20204516e-01 -1.10894871e+00 -3.37766677e-01 -3.47146988e-01
2.01762691e-01 -1.36122417e+00 -3.17312121e-01 -1.88150847e+00
3.87353182e-01 -9.48277414e-02 -6.20081723e-01 7.89326608e-01
-3.01441610e-01 7.06759095e-01 1.03949025e-01 -1.21745594e-01
-1.38163731e-01 4.26345110e-01 1.93085086e+00 -6.50296986e-01
5.38305901e-02 -3.18057597e-01 -8.35959196e-01 4.83605713e-01
8.64236116e-01 -3.11002135e-02 -2.60832727e-01 -5.76182246e-01
-8.92753184e-01 5.09726144e-02 2.57847935e-01 -1.05185437e+00
1.53875183e-02 1.75221652e-01 9.21000183e-01 -6.22119725e-01
2.05057397e-01 -7.63661563e-01 -5.16808867e-01 7.98025906e-01
-3.46211016e-01 1.05893694e-01 6.92511022e-01 8.70760754e-02
-9.20121521e-02 -3.63450535e-02 1.11854088e+00 -4.67779875e-01
-8.25722575e-01 5.19644856e-01 -7.01975107e-01 -1.21895425e-01
1.43576300e+00 -2.54185885e-01 -2.36013904e-01 -3.77744958e-02
-1.21835649e+00 3.69615138e-01 4.53643426e-02 3.19657296e-01
5.31837702e-01 -1.20743072e+00 -6.97748482e-01 2.28216320e-01
-2.80405283e-01 -1.31361663e-01 7.73152292e-01 1.37998092e+00
-6.87466800e-01 6.81204081e-01 -7.50581026e-01 -7.44523883e-01
-1.31684732e+00 1.89116061e-01 9.18019593e-01 -6.47486746e-01
-9.18379009e-01 1.39872277e+00 2.87871152e-01 -3.59696358e-01
1.92975864e-01 -7.47343481e-01 -4.32096034e-01 1.32128611e-01
6.04947150e-01 7.46892393e-02 3.04545313e-01 -8.34681869e-01
-3.58025998e-01 6.93222415e-03 -6.06107652e-01 -5.97199798e-02
1.24548566e+00 -7.74740707e-03 -1.19499937e-01 -1.28455713e-01
1.16731632e+00 -5.53265691e-01 -1.06261206e+00 -3.20422679e-01
-7.96336904e-02 2.88274325e-02 9.90642250e-01 -1.29632199e+00
-2.03163815e+00 7.50732303e-01 1.03106272e+00 -1.69769242e-01
1.04923105e+00 -4.21044938e-02 1.12798607e+00 -1.96308866e-01
3.42384130e-02 -9.20384228e-01 -5.99143915e-02 4.65069383e-01
7.31858134e-01 -1.58918345e+00 -3.16931993e-01 4.31041494e-02
-9.19197321e-01 1.07476437e+00 1.13844752e+00 -1.27034485e-01
5.78433990e-01 6.47659421e-01 1.59547895e-01 -4.12828326e-01
-3.31190586e-01 -1.18745759e-01 4.94158179e-01 6.41050518e-01
8.58452082e-01 4.17589575e-01 -1.86138287e-01 6.39934599e-01
-2.65742783e-02 -8.23770985e-02 6.59423828e-01 6.73311174e-01
-4.22729164e-01 -6.78042054e-01 -5.32442987e-01 1.36605477e+00
-7.90435076e-01 -2.43576199e-01 -3.81936789e-01 6.51148796e-01
4.73256186e-02 3.86661708e-01 1.99578792e-01 -1.82936758e-01
-1.71472296e-01 3.46355498e-01 6.38656974e-01 -4.02255714e-01
-9.59815145e-01 1.69665083e-01 -2.41871476e-01 -4.26215798e-01
-1.32036939e-01 -7.04711735e-01 -1.39155030e+00 -3.60227264e-02
-2.49815211e-01 -1.98642686e-01 6.02754891e-01 8.65501761e-01
3.17772150e-01 1.42880416e+00 2.76578572e-02 -1.00271666e+00
-1.41110674e-01 -1.08834851e+00 -4.72382456e-01 2.87570059e-01
3.76560181e-01 -7.94366598e-01 9.04147252e-02 -3.22977692e-01] | [14.580185890197754, -2.4225728511810303] |
17341240-4984-463b-a1a0-9c853e032580 | semi-supervised-medical-image-segmentation-1 | 2009.04448 | null | https://arxiv.org/abs/2009.04448v3 | https://arxiv.org/pdf/2009.04448v3.pdf | Semi-supervised Medical Image Segmentation through Dual-task Consistency | Deep learning-based semi-supervised learning (SSL) algorithms have led to promising results in medical images segmentation and can alleviate doctors' expensive annotations by leveraging unlabeled data. However, most of the existing SSL algorithms in literature tend to regularize the model training by perturbing networks and/or data. Observing that multi/dual-task learning attends to various levels of information which have inherent prediction perturbation, we ask the question in this work: can we explicitly build task-level regularization rather than implicitly constructing networks- and/or data-level perturbation-and-transformation for SSL? To answer this question, we propose a novel dual-task-consistency semi-supervised framework for the first time. Concretely, we use a dual-task deep network that jointly predicts a pixel-wise segmentation map and a geometry-aware level set representation of the target. The level set representation is converted to an approximated segmentation map through a differentiable task transform layer. Simultaneously, we introduce a dual-task consistency regularization between the level set-derived segmentation maps and directly predicted segmentation maps for both labeled and unlabeled data. Extensive experiments on two public datasets show that our method can largely improve the performance by incorporating the unlabeled data. Meanwhile, our framework outperforms the state-of-the-art semi-supervised medical image segmentation methods. Code is available at: https://github.com/Luoxd1996/DTC | ['Shaoting Zhang', 'Guotai Wang', 'Yinan Chen', 'Jieneng Chen', 'Xiangde Luo', 'Tao Song'] | 2020-09-09 | null | null | null | null | ['semi-supervised-medical-image-segmentation'] | ['computer-vision'] | [ 4.10638213e-01 6.70753777e-01 -4.41986531e-01 -7.43207276e-01
-1.27523983e+00 -2.75453746e-01 1.45716339e-01 -9.47383791e-03
-2.75796384e-01 4.55901563e-01 -3.60232405e-02 -2.52811372e-01
1.16764292e-01 -6.08569264e-01 -9.63786066e-01 -8.22410524e-01
3.11356783e-01 6.01310432e-01 3.49223882e-01 -1.16000235e-01
-1.10335104e-01 2.56160796e-01 -9.32993948e-01 2.30895489e-01
1.27448320e+00 1.00713682e+00 1.40929058e-01 3.16956609e-01
-1.16958255e-02 7.52519965e-01 -9.98395234e-02 -1.24974892e-01
4.66924608e-01 -4.39246297e-01 -1.08599329e+00 5.17155051e-01
4.78360444e-01 -1.30348682e-01 -1.53933227e-01 1.25280738e+00
4.94196534e-01 -1.00456238e-01 5.67788839e-01 -9.68654752e-01
-7.63136208e-01 4.48147476e-01 -9.67301190e-01 -6.49646223e-02
-1.33794114e-01 2.56672829e-01 5.63720226e-01 -6.67717218e-01
5.57247818e-01 9.01202440e-01 8.29594493e-01 5.30614018e-01
-1.44213533e+00 -4.30927098e-01 1.72140673e-01 -2.70903438e-01
-1.47917914e+00 -9.08396691e-02 1.18846607e+00 -5.83653748e-01
2.05462188e-01 9.65949818e-02 4.69349265e-01 8.14934671e-01
-3.21126431e-02 1.13881981e+00 1.29309833e+00 -3.14304143e-01
2.08749324e-02 1.01218395e-01 7.38641024e-02 1.21104777e+00
3.83171439e-02 -4.90955301e-02 -3.06725148e-02 6.71295589e-03
9.88335609e-01 -9.43907276e-02 -2.70611346e-01 -5.52050471e-01
-1.23068988e+00 6.69418991e-01 6.94557667e-01 2.44843006e-01
-2.09801853e-01 9.74557251e-02 4.55316782e-01 3.32907513e-02
8.76524270e-01 2.18320161e-01 -5.70181966e-01 4.38177407e-01
-1.15729392e+00 6.14417829e-02 3.27027023e-01 8.91441047e-01
1.00539351e+00 -1.61003932e-01 -3.31235260e-01 8.49392891e-01
3.47538143e-01 2.97753602e-01 4.36414778e-01 -1.00628662e+00
3.27547967e-01 8.48920286e-01 -1.45807803e-01 -7.57980347e-01
-5.15237153e-01 -5.21653652e-01 -1.11359429e+00 6.07688054e-02
5.48873842e-01 -2.54723847e-01 -1.22768617e+00 1.77168310e+00
7.92984247e-01 3.94577444e-01 -1.83595672e-01 9.62695777e-01
8.60699534e-01 3.28075230e-01 3.45073789e-02 -1.69641361e-01
1.19273913e+00 -1.35991395e+00 -6.13410056e-01 -1.37310907e-01
1.05930543e+00 -5.93570709e-01 1.37403357e+00 2.41925254e-01
-1.16614461e+00 -6.80761635e-01 -8.60141516e-01 -3.12380046e-01
-8.48439261e-02 2.83170849e-01 4.38788593e-01 4.76479560e-01
-1.01374912e+00 5.76083958e-01 -1.21330357e+00 1.53003857e-01
8.31854880e-01 3.70961547e-01 -2.25537401e-02 2.65112147e-02
-1.04508269e+00 5.86208880e-01 2.67338991e-01 3.14670682e-01
-6.96248829e-01 -1.02468419e+00 -9.98300076e-01 -4.45332497e-01
5.46839118e-01 -6.66134655e-01 1.07489526e+00 -1.15764153e+00
-1.29503882e+00 1.44004786e+00 -4.14432473e-02 -2.57415831e-01
8.98680806e-01 2.60362402e-02 -1.38647646e-01 2.46485010e-01
4.87807691e-01 8.47549915e-01 6.44458711e-01 -1.43237162e+00
-4.62705463e-01 -4.41478252e-01 -1.54314667e-01 7.90939853e-02
-2.01081663e-01 -4.28134024e-01 -7.42269158e-01 -9.10265207e-01
2.84003288e-01 -1.00294542e+00 -7.17708051e-01 5.30846834e-01
-8.39881659e-01 -1.43093631e-01 6.41605020e-01 -6.54646516e-01
1.07408953e+00 -2.03501940e+00 1.84021994e-01 1.37885407e-01
3.31572413e-01 3.17071050e-01 -8.65994301e-03 -2.94546425e-01
-9.47990920e-03 1.56223312e-01 -8.92330408e-01 -6.27992928e-01
-2.63592541e-01 3.25849950e-01 1.01636434e-02 7.68649101e-01
5.70900142e-02 1.24875855e+00 -9.76626277e-01 -1.10463512e+00
4.70332175e-01 5.55691957e-01 -5.92314601e-01 1.69472054e-01
-3.85929883e-01 1.12517715e+00 -7.61985481e-01 7.02573121e-01
7.66688287e-01 -6.53082848e-01 4.72787693e-02 -3.77630144e-01
-2.09832918e-02 -2.97424421e-02 -8.30522835e-01 2.16344523e+00
-3.71990204e-01 2.45810568e-01 1.36134326e-01 -1.42026758e+00
6.76345587e-01 3.41133922e-01 9.04630661e-01 -5.85893631e-01
2.44514391e-01 3.37349296e-01 -3.85017008e-01 -5.93673706e-01
-1.04413845e-01 -2.09894121e-01 -7.30312839e-02 2.88423926e-01
-1.45301893e-01 -3.61728340e-01 -9.44485441e-02 1.28840795e-02
5.53448319e-01 4.03236777e-01 9.38285142e-02 -4.46849585e-01
7.36408293e-01 1.94440842e-01 8.22066426e-01 5.03748000e-01
-3.77019852e-01 7.71373332e-01 4.93705153e-01 -6.17276311e-01
-8.85355413e-01 -9.73083377e-01 -5.01866877e-01 8.49807799e-01
3.97517681e-01 4.76428717e-02 -1.09052145e+00 -9.91351068e-01
-1.04124881e-01 3.41086119e-01 -8.83567512e-01 -1.02301449e-01
-7.19099641e-01 -8.19475710e-01 6.29697323e-01 5.48798442e-01
6.36858463e-01 -9.03674185e-01 -2.88400561e-01 2.10251614e-01
-3.34143519e-01 -1.21670902e+00 -8.82840812e-01 1.72496438e-01
-1.03104627e+00 -1.05968225e+00 -9.96492147e-01 -1.15445447e+00
1.19294035e+00 -7.82318488e-02 1.02495372e+00 2.08491400e-01
-3.77889812e-01 7.94041250e-03 -1.75113425e-01 -1.70673937e-01
-5.04756391e-01 2.67546773e-01 -2.47694716e-01 3.15022692e-02
-6.43455908e-02 -6.08013093e-01 -8.86215568e-01 4.02606517e-01
-1.00349629e+00 5.41804254e-01 5.73917389e-01 8.45140874e-01
1.27933002e+00 -3.12376440e-01 4.91235971e-01 -1.44129980e+00
9.12215561e-02 -3.18166226e-01 -5.97084522e-01 4.70146149e-01
-7.03578830e-01 2.06081107e-01 5.83969712e-01 -3.36580783e-01
-1.04965293e+00 5.41691899e-01 -3.67685765e-01 -4.75235462e-01
-2.68754780e-01 4.18148339e-01 -8.07986781e-02 -3.33190590e-01
6.46771908e-01 2.30351746e-01 8.35211575e-02 -3.43469173e-01
4.39831316e-01 4.44931597e-01 6.95212364e-01 -7.79616833e-01
7.86112309e-01 8.83040667e-01 5.68702258e-02 -4.44594532e-01
-1.51976657e+00 -4.66008693e-01 -1.02638936e+00 -1.92815140e-01
1.17363739e+00 -9.03862238e-01 -3.80221546e-01 5.31546950e-01
-8.90894473e-01 -9.30969000e-01 -4.96977806e-01 1.25074610e-01
-7.05191553e-01 6.47540867e-01 -6.95070922e-01 -2.39083633e-01
-4.44028854e-01 -1.51115167e+00 1.48241854e+00 2.58678179e-02
9.10735503e-02 -1.33409858e+00 3.72454748e-02 7.82369792e-01
1.17888510e-01 7.83544123e-01 7.63102710e-01 -4.49487925e-01
-4.77844030e-01 2.38391366e-02 -2.87462413e-01 5.07678449e-01
5.31666756e-01 -3.47686499e-01 -9.02902782e-01 -1.99241072e-01
1.40436590e-01 -6.06208146e-01 8.30105126e-01 9.55364585e-01
1.80032253e+00 -2.06251353e-01 -4.24889416e-01 1.27430570e+00
1.26505291e+00 -3.01588655e-01 2.62089461e-01 7.52409473e-02
1.28685045e+00 5.93702078e-01 6.51690304e-01 1.98642492e-01
5.42440295e-01 5.43138206e-01 2.68115997e-01 -8.14293683e-01
-2.31232449e-01 -3.29436570e-01 -1.91115409e-01 7.70598531e-01
1.74264520e-01 8.42317343e-02 -1.01587713e+00 4.72502559e-01
-1.96610963e+00 -2.50128448e-01 -3.62453550e-01 1.87047958e+00
1.27361763e+00 1.53813452e-01 9.78408754e-02 -1.74807861e-01
6.55503452e-01 2.14969754e-01 -9.25361633e-01 8.65773782e-02
6.47030547e-02 3.89021449e-02 7.33379543e-01 6.77421570e-01
-1.46332586e+00 1.05717456e+00 5.38080215e+00 9.95146334e-01
-1.34567630e+00 2.82616913e-01 1.38712335e+00 3.93786468e-03
-3.60071301e-01 -3.14537495e-01 -4.02186006e-01 4.40716684e-01
5.41073203e-01 6.50515556e-02 -6.83324877e-03 7.99003661e-01
4.24905419e-01 1.33050725e-01 -1.08546376e+00 8.35180044e-01
-3.88947539e-02 -1.53949845e+00 5.81638282e-03 8.60111639e-02
1.07760870e+00 9.27716419e-02 1.78276941e-01 5.08288927e-02
2.57906437e-01 -9.81717408e-01 6.39605582e-01 3.74460101e-01
1.15372479e+00 -4.35315788e-01 5.49260676e-01 4.27951276e-01
-1.26485920e+00 3.98365080e-01 -1.51140355e-02 3.19031745e-01
3.52642834e-01 6.12343490e-01 -5.42145252e-01 5.53444982e-01
5.66947341e-01 9.39628005e-01 -5.47766268e-01 6.67968154e-01
-2.82569289e-01 7.21716881e-01 -2.26539791e-01 5.38266480e-01
4.75683987e-01 -3.61426204e-01 1.93523601e-01 1.09805727e+00
-1.56100124e-01 9.84587520e-02 5.60171247e-01 1.18224728e+00
-2.25614712e-01 7.26784095e-02 -2.06245497e-01 4.62090254e-01
9.35029164e-02 1.25322509e+00 -1.10775125e+00 -4.45384622e-01
-3.16560119e-01 1.07893324e+00 2.42056668e-01 3.49647164e-01
-9.32793856e-01 1.58621356e-01 1.56333849e-01 3.09195220e-01
-3.64098474e-02 5.66664450e-02 -7.97791660e-01 -1.09269404e+00
1.40154526e-01 -5.27396202e-01 4.08280492e-01 -4.43467498e-01
-1.28654253e+00 4.99272317e-01 -8.20135102e-02 -1.22817361e+00
7.76237398e-02 -3.54077399e-01 -5.73451459e-01 7.43551314e-01
-1.80984569e+00 -1.50064838e+00 -2.92523563e-01 6.16080642e-01
5.95676422e-01 1.69807896e-01 4.96488005e-01 4.13900882e-01
-6.96207941e-01 7.40311503e-01 -3.59497010e-03 4.03625339e-01
7.01690793e-01 -1.46284819e+00 3.46476614e-01 6.43799126e-01
-1.16839662e-01 2.45296046e-01 3.79097909e-01 -6.09848261e-01
-9.45558429e-01 -1.55991673e+00 4.10022348e-01 -4.84054625e-01
4.10946399e-01 -2.49434903e-01 -1.06493306e+00 8.35969090e-01
-1.85581312e-01 8.11894059e-01 5.54857910e-01 -4.47135687e-01
8.30735639e-02 2.13643350e-02 -1.29945111e+00 4.22503412e-01
1.00790215e+00 -4.92164582e-01 -3.17557901e-01 7.22697198e-01
9.79103088e-01 -8.62596035e-01 -9.68311429e-01 5.90027571e-01
1.83666974e-01 -7.65449345e-01 1.12033308e+00 -3.74456823e-01
5.71833193e-01 -2.99068451e-01 1.63818300e-01 -1.01312482e+00
-7.79405385e-02 -5.37203372e-01 5.53610027e-02 9.60760713e-01
4.29067373e-01 -4.69215959e-01 1.16709936e+00 7.35557079e-01
-5.77235639e-01 -1.40510154e+00 -9.39819574e-01 -5.45748293e-01
5.18287182e-01 -3.82304758e-01 1.95781648e-01 1.12592983e+00
-3.77069622e-01 -4.33922894e-02 -3.18191767e-01 2.83391207e-01
9.53844309e-01 2.29059607e-01 4.50084060e-01 -1.05194914e+00
-1.64249256e-01 -3.49825084e-01 3.00196279e-02 -1.28113484e+00
3.31899911e-01 -1.17520106e+00 1.86953798e-01 -1.57153893e+00
3.64815742e-02 -7.56400108e-01 -1.73053265e-01 7.18749464e-01
-3.40720326e-01 4.62731481e-01 -3.86638880e-01 3.59637469e-01
-6.66211069e-01 5.63634276e-01 1.86132300e+00 -2.23776951e-01
-3.17922533e-01 4.71198000e-02 -7.97365487e-01 1.00520515e+00
7.00531602e-01 -4.77554470e-01 -4.25830573e-01 -4.56124961e-01
-2.57173292e-02 1.85142189e-01 5.44406474e-01 -6.45208418e-01
2.73819566e-01 -2.48540595e-01 1.35382950e-01 -5.19358575e-01
-1.55456178e-02 -6.73123181e-01 -2.11182907e-01 6.15409374e-01
-6.04902565e-01 -6.26903713e-01 8.96651745e-02 3.97128105e-01
-2.12105602e-01 7.02972934e-02 1.11413836e+00 -2.10251912e-01
-4.49581414e-01 7.81737089e-01 4.90379594e-02 3.71459872e-01
9.63478684e-01 -2.45021492e-01 1.80699736e-01 3.01620811e-02
-8.82704556e-01 6.09365165e-01 4.21643585e-01 7.52539709e-02
4.21634972e-01 -1.14626753e+00 -7.07282364e-01 4.28748950e-02
6.58010021e-02 9.94103849e-01 4.39593315e-01 1.12743819e+00
-6.78027272e-01 1.74150854e-01 9.80450287e-02 -9.31423545e-01
-8.19794714e-01 4.80560720e-01 7.16913700e-01 -3.93276781e-01
-8.21906090e-01 1.00455177e+00 5.68569481e-01 -8.07399631e-01
2.00766996e-01 -6.40889883e-01 2.14495167e-01 -3.69814694e-01
3.16340849e-02 1.01954870e-01 -9.57239941e-02 -7.08234966e-01
-3.61355782e-01 8.73837471e-01 -1.22888654e-01 3.08996052e-01
1.32455707e+00 -1.86102927e-01 4.70449170e-03 3.55388075e-01
1.40729964e+00 -4.09804910e-01 -1.68486917e+00 -5.30133128e-01
9.09213442e-03 -1.19466878e-01 1.85904399e-01 -6.00939810e-01
-1.47986066e+00 7.85977602e-01 7.43939877e-01 -1.57876387e-01
1.03919470e+00 2.76027262e-01 1.16804934e+00 1.26890615e-01
9.79012772e-02 -1.16396296e+00 2.25334644e-01 1.51267846e-03
6.85363233e-01 -1.69508207e+00 6.67114137e-03 -8.46127808e-01
-7.25477397e-01 7.46207118e-01 7.22073317e-01 -2.93771982e-01
9.12976563e-01 2.79282719e-01 2.90661097e-01 -3.88233542e-01
-1.56677619e-01 -2.42212079e-02 6.00840449e-01 5.46635568e-01
5.50847888e-01 1.23415194e-01 -2.33232081e-02 5.42714238e-01
-9.32739079e-02 1.82568327e-01 1.27209544e-01 7.45504200e-01
-1.71295971e-01 -1.04757261e+00 -1.88091710e-01 5.25781572e-01
-4.26206708e-01 -5.59331626e-02 -3.87446582e-02 6.68724775e-01
4.19098228e-01 4.77062076e-01 -1.10314198e-01 -5.98182678e-02
3.16738129e-01 -3.07678640e-01 2.98182666e-01 -7.78907120e-01
-5.09185851e-01 3.00443143e-01 -4.62624729e-01 -6.13366961e-01
-5.45371056e-01 -6.25718772e-01 -1.65985906e+00 1.81419656e-01
-7.95393139e-02 -3.55758891e-02 1.49284676e-01 1.12375569e+00
1.42776936e-01 6.94706619e-01 6.58756793e-01 -8.82738709e-01
-3.05702567e-01 -5.98247170e-01 -4.98250782e-01 6.38038814e-01
5.18530905e-01 -4.89028990e-01 -2.74088621e-01 3.91313910e-01] | [14.660948753356934, -2.0838422775268555] |
738dc22c-ff72-4558-a420-ad3619b244fb | implicit-functions-in-feature-space-for-3d | 2003.01456 | null | https://arxiv.org/abs/2003.01456v2 | https://arxiv.org/pdf/2003.01456v2.pdf | Implicit Functions in Feature Space for 3D Shape Reconstruction and Completion | While many works focus on 3D reconstruction from images, in this paper, we focus on 3D shape reconstruction and completion from a variety of 3D inputs, which are deficient in some respect: low and high resolution voxels, sparse and dense point clouds, complete or incomplete. Processing of such 3D inputs is an increasingly important problem as they are the output of 3D scanners, which are becoming more accessible, and are the intermediate output of 3D computer vision algorithms. Recently, learned implicit functions have shown great promise as they produce continuous reconstructions. However, we identified two limitations in reconstruction from 3D inputs: 1) details present in the input data are not retained, and 2) poor reconstruction of articulated humans. To solve this, we propose Implicit Feature Networks (IF-Nets), which deliver continuous outputs, can handle multiple topologies, and complete shapes for missing or sparse input data retaining the nice properties of recent learned implicit functions, but critically they can also retain detail when it is present in the input data, and can reconstruct articulated humans. Our work differs from prior work in two crucial aspects. First, instead of using a single vector to encode a 3D shape, we extract a learnable 3-dimensional multi-scale tensor of deep features, which is aligned with the original Euclidean space embedding the shape. Second, instead of classifying x-y-z point coordinates directly, we classify deep features extracted from the tensor at a continuous query point. We show that this forces our model to make decisions based on global and local shape structure, as opposed to point coordinates, which are arbitrary under Euclidean transformations. Experiments demonstrate that IF-Nets clearly outperform prior work in 3D object reconstruction in ShapeNet, and obtain significantly more accurate 3D human reconstructions. | ['Gerard Pons-Moll', 'Julian Chibane', 'Thiemo Alldieck'] | 2020-03-03 | implicit-functions-in-feature-space-for-3d-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Chibane_Implicit_Functions_in_Feature_Space_for_3D_Shape_Reconstruction_and_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Chibane_Implicit_Functions_in_Feature_Space_for_3D_Shape_Reconstruction_and_CVPR_2020_paper.pdf | cvpr-2020-6 | ['3d-object-reconstruction'] | ['computer-vision'] | [-9.44928452e-02 4.05323990e-02 1.30188555e-01 -3.12788278e-01
-3.97062838e-01 -6.72503293e-01 6.73839688e-01 -1.75581157e-01
-1.60553038e-01 4.75719124e-01 3.68099272e-01 -9.77953598e-02
-3.14427257e-01 -9.50417042e-01 -9.30231452e-01 -5.65895796e-01
2.93316948e-03 1.00929761e+00 4.19272147e-02 -1.63055196e-01
2.08808810e-01 1.21964455e+00 -1.44225836e+00 1.83207870e-01
5.54438412e-01 9.91958022e-01 4.04589325e-02 5.16953588e-01
-3.26310724e-01 2.53146201e-01 -9.14760232e-02 -1.30824924e-01
4.08843517e-01 -1.14257902e-01 -7.66525030e-01 2.52504468e-01
5.45613348e-01 -4.35295761e-01 -3.45152646e-01 8.44441414e-01
2.50186294e-01 -1.36196926e-01 9.99959469e-01 -1.07850945e+00
-9.92036641e-01 -2.89387722e-02 -2.81941772e-01 -5.04765511e-01
4.00327533e-01 2.86528945e-01 9.49212015e-01 -1.27532816e+00
8.56754720e-01 1.45385885e+00 9.93182421e-01 4.71791297e-01
-1.48396015e+00 -2.59489626e-01 -7.61031434e-02 -2.95729309e-01
-1.18641162e+00 -4.05554920e-01 9.26494777e-01 -6.21644616e-01
8.85060191e-01 3.11774909e-01 9.41073895e-01 1.00941622e+00
3.47327441e-02 5.90336919e-01 1.08160722e+00 -1.48191437e-01
1.55878693e-01 -2.40610868e-01 -1.05756633e-01 6.57605410e-01
8.22999701e-02 2.79959083e-01 -2.60814309e-01 -2.36577049e-01
1.66031301e+00 4.59755301e-01 -3.18063647e-01 -9.17094171e-01
-1.53545368e+00 7.54925489e-01 6.91616237e-01 2.90843040e-01
-4.66798246e-01 2.58894205e-01 -4.53295559e-02 2.73589164e-01
3.58966589e-01 3.96619409e-01 -5.36112905e-01 -2.63348874e-02
-6.96394324e-01 3.90130252e-01 7.61064410e-01 9.64518905e-01
9.87212777e-01 1.10160425e-01 2.17733845e-01 6.03165686e-01
3.38273525e-01 8.06926131e-01 6.92302063e-02 -1.16190338e+00
2.55035311e-01 7.61145115e-01 7.25613311e-02 -1.18677986e+00
-4.51182395e-01 -4.47609752e-01 -1.28840649e+00 5.11712134e-01
3.38284135e-01 4.34736043e-01 -1.11136615e+00 1.60304248e+00
2.98991561e-01 -1.65750962e-02 -1.98245421e-01 1.16325784e+00
7.70832121e-01 4.85348344e-01 -4.49571311e-01 1.99677676e-01
1.02565408e+00 -3.39863718e-01 -2.88410574e-01 2.71049619e-01
1.49800032e-01 -5.65091252e-01 9.20197248e-01 2.40322679e-01
-1.36000454e+00 -4.92608100e-01 -8.44281375e-01 -4.76081461e-01
-3.33956569e-01 -1.95260823e-01 5.69989502e-01 -1.02881873e-02
-1.27347875e+00 7.54703999e-01 -8.91264677e-01 -1.32659823e-01
4.94410574e-01 4.18048590e-01 -8.14294219e-01 -1.73323035e-01
-7.14574456e-01 7.73940802e-01 -1.77879825e-01 1.69750348e-01
-6.64657295e-01 -8.26050162e-01 -9.95183468e-01 -1.14792427e-02
1.88679710e-01 -1.16419661e+00 8.80156279e-01 -6.52824163e-01
-1.31723440e+00 1.02273047e+00 4.37828340e-02 -8.36103857e-02
5.94889343e-01 -1.19910225e-01 1.97930500e-01 1.47849500e-01
2.52074655e-02 6.75828636e-01 9.44349885e-01 -1.48897660e+00
4.84730676e-02 -7.59248793e-01 1.73383847e-01 1.13481931e-01
1.96026281e-01 -5.70866764e-01 -1.50859788e-01 -5.74022412e-01
7.92867780e-01 -7.34010875e-01 -4.10727710e-01 7.00821221e-01
-2.73105145e-01 -1.10981464e-01 8.49252105e-01 -4.87349331e-01
2.50831187e-01 -2.18464231e+00 5.67188084e-01 3.55201423e-01
6.02081537e-01 -4.27701063e-02 -1.84669510e-01 3.24105173e-01
1.20192833e-01 3.24140996e-01 -6.58441603e-01 -4.50690627e-01
1.87790126e-01 6.48497403e-01 -4.26464498e-01 5.80603659e-01
5.57398260e-01 1.16393077e+00 -7.72283494e-01 -1.57849759e-01
2.74905413e-01 8.14010203e-01 -7.61083126e-01 2.50700623e-01
-2.35449523e-01 7.84149349e-01 -6.74168766e-01 7.34061301e-01
7.49606550e-01 -1.60820156e-01 -2.82732487e-01 -3.92392874e-01
-2.00547606e-01 3.13374072e-01 -1.16272271e+00 2.15898633e+00
-3.42752784e-01 2.24361545e-03 4.27051097e-01 -1.17283285e+00
1.01057136e+00 4.39934552e-01 7.85763323e-01 -6.06873810e-01
4.40539867e-02 3.51299495e-01 -2.59971946e-01 -1.29099831e-01
1.96401089e-01 -3.36789519e-01 3.18749957e-02 4.99570340e-01
4.98852693e-02 -7.63264000e-01 -4.56989437e-01 6.30142763e-02
1.07145011e+00 3.68033081e-01 -2.06883568e-02 -1.79802850e-01
3.36597115e-01 -3.62846740e-02 5.46789110e-01 3.90655160e-01
2.42582455e-01 1.16461527e+00 4.79863584e-01 -7.46108770e-01
-1.45007741e+00 -1.40644348e+00 -3.34408730e-01 2.14049041e-01
1.80956591e-02 -9.17927176e-02 -2.94961482e-01 -4.38417971e-01
4.05454606e-01 2.30843842e-01 -6.52431190e-01 7.96366334e-02
-8.67528856e-01 -1.43430710e-01 2.61377305e-01 5.34997821e-01
1.29314408e-01 -1.15234363e+00 -6.05111599e-01 2.06548125e-01
8.97905007e-02 -1.04640031e+00 -4.40612435e-01 1.91844746e-01
-1.33563805e+00 -1.14470208e+00 -8.07185471e-01 -6.68449819e-01
9.43694055e-01 1.76112771e-01 1.18600142e+00 3.41084182e-01
-3.40914816e-01 7.60747015e-01 -2.43061170e-01 -1.90534350e-02
-1.18390441e-01 -2.18181625e-01 1.94948956e-01 7.76804909e-02
-9.81250256e-02 -1.29518855e+00 -4.32060063e-01 3.12544018e-01
-1.07532668e+00 2.13688821e-01 7.20369399e-01 9.66794729e-01
1.05322158e+00 -3.38800818e-01 3.86172742e-01 -7.54139960e-01
2.80146182e-01 -3.38243335e-01 -2.72553146e-01 -4.32048962e-02
-1.49683684e-01 2.55610496e-01 5.34245312e-01 -3.85652959e-01
-5.64260483e-01 2.57459819e-01 -5.11083484e-01 -9.58090723e-01
-2.77884185e-01 2.50901520e-01 -1.29232541e-01 -5.21360375e-02
4.86672878e-01 2.73580074e-01 3.50170076e-01 -9.96016741e-01
6.01563454e-01 1.85764760e-01 5.92928410e-01 -8.89085412e-01
1.09509122e+00 6.58698380e-01 3.04281682e-01 -7.96592236e-01
-4.26347345e-01 -2.66852349e-01 -1.33846188e+00 1.91423357e-01
7.67688990e-01 -7.10426748e-01 -7.39573002e-01 3.44993144e-01
-1.35459292e+00 -1.09220304e-01 -8.02744746e-01 3.60075533e-01
-9.11811113e-01 3.16389769e-01 -6.79960251e-01 -5.48828542e-01
-2.47471392e-01 -1.23991358e+00 1.36854088e+00 -3.78557146e-01
-1.85860544e-01 -8.84703636e-01 -9.68148932e-02 -5.45683503e-02
4.03952748e-01 6.66210473e-01 1.25520861e+00 -1.20115273e-01
-9.57837462e-01 -1.17011547e-01 -2.00771660e-01 2.47117743e-01
1.63386106e-01 -1.98082611e-01 -7.01662421e-01 -2.74261385e-01
1.43818632e-01 -3.67805302e-01 6.04928672e-01 3.16358328e-01
1.28626847e+00 -4.23431575e-01 -7.85941333e-02 1.07501686e+00
1.24239111e+00 -4.06112760e-01 4.92690861e-01 -3.98444720e-02
8.25066805e-01 5.86118281e-01 8.60658102e-03 2.49504179e-01
4.17995393e-01 5.68915427e-01 8.45328808e-01 -2.74965912e-01
-3.79678577e-01 -5.24751306e-01 1.98646218e-01 1.22400427e+00
-5.28491676e-01 4.75848794e-01 -7.18101382e-01 4.78856325e-01
-1.67472422e+00 -6.15646124e-01 -1.18880257e-01 2.24484205e+00
7.70474195e-01 -6.35644943e-02 1.36744985e-02 2.48889908e-01
2.42608577e-01 -9.58250370e-03 -8.03290427e-01 -1.71692371e-01
-2.98754901e-01 5.99258959e-01 1.46747246e-01 4.38058615e-01
-6.11880362e-01 6.21655107e-01 5.99114752e+00 3.07530910e-01
-1.21093941e+00 1.75494701e-02 1.86726972e-01 -5.32814190e-02
-7.83954322e-01 -2.51704548e-02 -2.94729233e-01 8.96225944e-02
3.36859792e-01 1.31189808e-01 4.78028983e-01 7.13553071e-01
-1.23847261e-01 4.81488258e-01 -1.55584025e+00 1.26895356e+00
-1.05256326e-01 -1.48792660e+00 4.19885397e-01 4.60365653e-01
5.81880629e-01 -2.20291829e-03 5.47498912e-02 9.09315199e-02
2.83702970e-01 -1.30694139e+00 1.02616060e+00 8.29226017e-01
1.16694772e+00 -4.67051715e-01 2.68019348e-01 6.40944242e-01
-1.08065903e+00 2.61238843e-01 -6.67979598e-01 -4.80076522e-02
2.62092412e-01 6.99852645e-01 -4.78920460e-01 6.31420672e-01
7.17795014e-01 9.91148531e-01 -6.66097626e-02 9.23943102e-01
-1.02260523e-01 1.10077441e-01 -4.54832435e-01 3.50980818e-01
2.15483904e-01 -4.41804022e-01 7.49033332e-01 8.22187185e-01
4.51986194e-01 4.62340951e-01 2.65455127e-01 1.42186034e+00
-1.21675059e-01 -1.69709757e-01 -9.97675836e-01 1.33441970e-01
-3.53517532e-02 1.15853345e+00 -5.03378987e-01 -1.16372243e-01
-2.04824612e-01 8.35657001e-01 4.46732223e-01 4.44228530e-01
-3.00427526e-01 1.10337131e-01 8.80635917e-01 4.84185696e-01
3.11533064e-01 -8.96824718e-01 -5.09382963e-01 -1.25522208e+00
2.08619580e-01 -5.89506447e-01 -1.49306610e-01 -8.42158675e-01
-1.55214095e+00 5.11519969e-01 -1.75848916e-01 -1.24146712e+00
-4.53038514e-02 -8.27991009e-01 -3.05240095e-01 9.62195158e-01
-1.43533874e+00 -1.25912166e+00 -7.21505731e-02 7.39420593e-01
3.72114241e-01 2.43829042e-01 8.61206770e-01 2.83676326e-01
3.30358952e-01 1.79710984e-01 -1.93389848e-01 2.98624188e-01
2.04549417e-01 -1.17476332e+00 5.44117212e-01 2.51078516e-01
4.20676708e-01 7.64564276e-01 1.38762131e-01 -5.81769764e-01
-2.11926508e+00 -7.59024501e-01 5.74718654e-01 -7.68756330e-01
2.71002471e-01 -4.62441087e-01 -1.07728946e+00 8.23260009e-01
-4.06966686e-01 3.73770863e-01 1.56292930e-01 4.89504747e-02
-5.48572004e-01 1.61766052e-01 -1.31063962e+00 3.35385352e-01
1.64706469e+00 -5.60908675e-01 -7.93104887e-01 1.19637325e-01
8.16549718e-01 -5.94594777e-01 -1.04529285e+00 5.76126575e-01
7.23988771e-01 -1.03487241e+00 1.38822556e+00 -7.36733496e-01
4.98425037e-01 -3.09617162e-01 -4.23706353e-01 -1.30844688e+00
-5.12521267e-01 -1.81372479e-01 -1.91812202e-01 7.36562252e-01
5.64056113e-02 -6.51473165e-01 6.79578602e-01 5.54149389e-01
-4.24933881e-01 -9.42351341e-01 -1.06986892e+00 -6.67054892e-01
3.68469059e-01 -4.15783644e-01 9.86369252e-01 1.00192738e+00
-7.01091170e-01 1.42887637e-01 -2.71670401e-01 3.29894051e-02
7.89817870e-01 4.64865655e-01 7.95882404e-01 -1.66811180e+00
-7.78837875e-02 -5.31475544e-01 -5.82183659e-01 -1.59552896e+00
1.59110595e-02 -1.14579785e+00 -2.40330882e-02 -1.62982595e+00
-1.03300087e-01 -9.60084140e-01 4.37359586e-02 6.05917156e-01
5.47547877e-01 3.46790761e-01 2.22658753e-01 4.74962533e-01
-9.51302126e-02 1.04694724e+00 1.77200150e+00 -2.49587104e-01
2.22699139e-02 -3.80160324e-02 -4.58017945e-01 7.67699838e-01
2.49831587e-01 -2.86522001e-01 -7.39997029e-02 -9.57131326e-01
2.05549851e-01 2.59173900e-01 7.13450730e-01 -7.14919031e-01
1.17666624e-01 -2.00509280e-01 7.66854823e-01 -8.51375580e-01
7.02224374e-01 -1.22205949e+00 4.55143601e-01 2.09314376e-01
3.52198462e-04 1.04226217e-01 -2.06868559e-01 4.58388627e-01
-1.23707809e-01 8.66227746e-02 5.04007697e-01 -6.15290642e-01
-2.07657263e-01 9.41571832e-01 1.91542581e-01 -1.40875578e-01
5.01298785e-01 -4.98971790e-01 1.86013848e-01 -1.60635069e-01
-8.74250829e-01 -9.05643180e-02 9.61927831e-01 4.16443050e-01
1.06132424e+00 -1.71151781e+00 -7.92521000e-01 6.17270708e-01
-9.58908349e-02 8.16336393e-01 9.16742533e-02 7.39915490e-01
-5.52193105e-01 2.99440563e-01 -4.36461031e-01 -1.07357979e+00
-6.75983489e-01 3.58364254e-01 2.98248500e-01 3.43486667e-02
-1.27010298e+00 4.33753759e-01 5.47140241e-01 -1.10795927e+00
7.67820403e-02 -7.17103243e-01 1.10472530e-01 -3.48357797e-01
1.52754351e-01 -2.36196462e-02 1.52606875e-01 -9.22638178e-01
-1.50667951e-01 1.20828092e+00 5.09061515e-01 -5.01781527e-04
1.84748828e+00 2.00225651e-01 -3.96171212e-01 5.25498033e-01
1.20729792e+00 -4.67496924e-02 -1.29622912e+00 -5.27187288e-01
-3.18598747e-01 -6.75889075e-01 -2.01098725e-01 -5.25158405e-01
-1.18506432e+00 1.16111434e+00 5.56783676e-02 1.57934785e-01
7.54684031e-01 1.95262671e-01 8.61665010e-01 2.58520126e-01
7.02309966e-01 -4.30110008e-01 7.48495013e-02 7.67342627e-01
1.52559733e+00 -9.93384659e-01 1.09013915e-01 -3.96225929e-01
-9.58041847e-02 1.18978882e+00 2.92976797e-01 -4.05635595e-01
7.73457825e-01 7.25241080e-02 -2.45941311e-01 -5.92474222e-01
-4.64659363e-01 -5.54912202e-02 4.79191929e-01 7.86395073e-01
1.41173601e-01 1.76619709e-01 9.39315408e-02 4.74603355e-01
-4.23027605e-01 -3.13421875e-01 1.66210443e-01 8.19574237e-01
-3.06770116e-01 -1.16935432e+00 -5.36097527e-01 4.58229333e-01
1.82810854e-02 2.38786086e-01 -4.51376110e-01 6.95490122e-01
9.68328267e-02 2.12684333e-01 1.66307777e-01 -2.21123993e-01
6.17748618e-01 -4.97586578e-02 7.67119765e-01 -6.97787344e-01
-2.39931509e-01 -9.46440324e-02 -3.52435082e-01 -6.34880781e-01
-4.02375221e-01 -5.46004474e-01 -1.32893419e+00 -3.08318824e-01
1.28841639e-01 -1.22255394e-02 9.74958181e-01 7.77010620e-01
4.45870519e-01 1.72500759e-01 5.78440428e-01 -1.50922561e+00
-6.66395307e-01 -5.90416551e-01 -6.79551721e-01 8.61106932e-01
6.31546974e-01 -9.65486407e-01 -4.51120764e-01 1.41101303e-02] | [8.652941703796387, -3.659486770629883] |
c37dd893-471d-464c-b32a-1d780f07c105 | guiding-teacher-forcing-with-seer-forcing-for | 2106.06751 | null | https://arxiv.org/abs/2106.06751v1 | https://arxiv.org/pdf/2106.06751v1.pdf | Guiding Teacher Forcing with Seer Forcing for Neural Machine Translation | Although teacher forcing has become the main training paradigm for neural machine translation, it usually makes predictions only conditioned on past information, and hence lacks global planning for the future. To address this problem, we introduce another decoder, called seer decoder, into the encoder-decoder framework during training, which involves future information in target predictions. Meanwhile, we force the conventional decoder to simulate the behaviors of the seer decoder via knowledge distillation. In this way, at test the conventional decoder can perform like the seer decoder without the attendance of it. Experiment results on the Chinese-English, English-German and English-Romanian translation tasks show our method can outperform competitive baselines significantly and achieves greater improvements on the bigger data sets. Besides, the experiments also prove knowledge distillation the best way to transfer knowledge from the seer decoder to the conventional decoder compared to adversarial learning and L2 regularization. | ['Chenze Shao', 'Zhengxin Yang', 'Dengji Guo', 'Shuhao Gu', 'Yang Feng'] | 2021-06-12 | null | https://aclanthology.org/2021.acl-long.223 | https://aclanthology.org/2021.acl-long.223.pdf | acl-2021-5 | ['l2-regularization'] | ['methodology'] | [ 2.41335258e-01 6.68419063e-01 -5.40648878e-01 -1.79645225e-01
-1.05893528e+00 -6.76863968e-01 8.01756203e-01 -6.44346774e-01
-2.46820301e-01 1.16887379e+00 5.80237091e-01 -7.76017964e-01
7.44160950e-01 -5.74778080e-01 -1.31911409e+00 -4.62466389e-01
3.76586407e-01 6.12706602e-01 -1.50271147e-01 -4.20824736e-01
-3.21697742e-01 -5.68385869e-02 -4.90151823e-01 4.56165403e-01
8.86099756e-01 4.18992758e-01 4.20732349e-01 3.65916610e-01
3.46135885e-01 9.49377179e-01 -3.68654221e-01 -8.64556491e-01
3.80057454e-01 -6.89605236e-01 -8.59006286e-01 -3.03348780e-01
1.51072489e-02 -6.48945630e-01 -9.26987946e-01 1.05549884e+00
4.55712318e-01 -1.16397720e-02 5.13059080e-01 -5.97844303e-01
-1.31132686e+00 1.08675575e+00 -1.91833645e-01 3.27609815e-02
3.72612327e-01 3.98344934e-01 8.45277369e-01 -1.00018227e+00
7.73175657e-01 1.33551478e+00 3.92945588e-01 1.07357848e+00
-1.13922584e+00 -7.18451977e-01 3.26362520e-01 1.83591381e-01
-1.11745799e+00 -6.45643771e-01 6.20878041e-01 -1.14295185e-01
1.15482891e+00 -3.79320718e-02 2.67807245e-01 1.64438713e+00
3.64023685e-01 1.09754264e+00 1.13008964e+00 -2.60655463e-01
-1.65416673e-01 2.11387485e-01 -4.29818481e-01 5.56310952e-01
-3.34858924e-01 7.06176937e-01 -6.25243366e-01 2.37271935e-01
5.54814279e-01 -2.64883071e-01 -6.15055501e-01 3.23026896e-01
-1.54401684e+00 5.55755794e-01 2.60871112e-01 3.92266512e-02
-1.33490622e-01 3.97581667e-01 2.77200848e-01 8.86724234e-01
6.34182930e-01 5.00754654e-01 -8.66639018e-01 -2.34438062e-01
-7.55572021e-01 7.92112760e-03 7.94007838e-01 1.30459511e+00
6.84988499e-01 2.71463275e-01 -3.62593025e-01 3.62678498e-01
2.00110286e-01 8.59502733e-01 7.00417578e-01 -7.50359356e-01
1.08039975e+00 6.13142028e-02 8.80373865e-02 -4.67682004e-01
1.69091105e-01 -6.06550395e-01 -8.09431076e-01 -2.39742130e-01
3.32840860e-01 -6.70333862e-01 -1.12297964e+00 2.01329160e+00
-1.41530916e-01 4.59159553e-01 5.02147734e-01 8.41373026e-01
4.66651529e-01 1.04081404e+00 -2.00134709e-01 -3.64825487e-01
7.16842473e-01 -1.34294295e+00 -8.87165606e-01 -6.28522158e-01
8.42311859e-01 -6.31911218e-01 9.51834083e-01 1.39629066e-01
-1.16440248e+00 -5.15270352e-01 -9.56437588e-01 -2.13384911e-01
-2.63557851e-01 8.01432133e-02 4.24196720e-01 1.48251638e-01
-1.14517462e+00 7.98042476e-01 -1.22244513e+00 1.54281467e-01
2.38549620e-01 4.14081335e-01 -3.55940402e-01 -1.18034892e-01
-1.65182042e+00 1.44805944e+00 5.72996795e-01 1.70264006e-01
-1.50386751e+00 -6.11546755e-01 -9.69220996e-01 -5.61624542e-02
2.27429122e-01 -7.40931034e-01 1.60070205e+00 -1.24417222e+00
-1.91770208e+00 5.98628700e-01 -4.12969947e-01 -8.38943124e-01
7.15896368e-01 -4.52937573e-01 -3.28228474e-01 -2.98316509e-01
-2.36339509e-01 7.13211119e-01 7.84371316e-01 -1.15970623e+00
-1.39911219e-01 1.86876990e-02 2.23648071e-01 4.81768847e-01
-5.70788160e-02 -1.37012228e-01 -3.52229595e-01 -5.23578346e-01
-1.51208907e-01 -1.01695395e+00 -2.62002617e-01 -4.36348259e-01
-5.64356565e-01 -1.11065790e-01 6.63553476e-01 -1.09397566e+00
1.10031354e+00 -2.08495092e+00 4.97907460e-01 -2.16141164e-01
-8.96534249e-02 3.82992357e-01 -3.43788743e-01 5.02344668e-01
-1.59455150e-01 3.53864506e-02 -2.12358028e-01 -3.97187352e-01
-1.47990957e-01 4.80370075e-01 -9.41149712e-01 4.56002951e-01
3.07954103e-01 1.49062538e+00 -1.13686466e+00 -1.74522951e-01
-1.17974602e-01 3.83636147e-01 -5.98555624e-01 3.46838266e-01
-7.33026564e-01 1.00729275e+00 -6.36254132e-01 3.61037701e-01
3.38286132e-01 -1.44213498e-01 2.39334926e-01 4.03177112e-01
2.44929850e-01 8.27055871e-01 -3.81722629e-01 2.04620171e+00
-7.06582844e-01 6.19761169e-01 -2.42146671e-01 -1.00886083e+00
8.68129551e-01 8.44267786e-01 -1.58518806e-01 -8.17284763e-01
4.10429755e-04 5.13370275e-01 2.25754663e-01 -2.50730604e-01
3.15345377e-01 -4.00935620e-01 -6.09160773e-03 4.12063390e-01
2.50290096e-01 -4.76494990e-02 -3.10727954e-01 1.33840725e-01
1.00965405e+00 6.59199178e-01 -8.46210048e-02 7.62559026e-02
4.22541976e-01 -8.35796539e-03 8.14843953e-01 5.78815341e-01
-2.53933426e-02 7.05638468e-01 3.26049924e-01 -3.09434146e-01
-9.94073093e-01 -1.10034001e+00 2.78897345e-01 8.70702207e-01
1.21359073e-01 -3.06230515e-01 -6.55207455e-01 -9.55055177e-01
-2.72831798e-01 1.22739720e+00 -5.93274474e-01 -5.30250013e-01
-1.00902367e+00 -5.07406950e-01 6.32710516e-01 4.60160971e-01
4.14660096e-01 -1.02743936e+00 6.49470836e-02 3.88075233e-01
-5.56857526e-01 -1.09528458e+00 -6.87998056e-01 2.28908166e-01
-1.00530016e+00 -5.76208055e-01 -9.61607635e-01 -9.32518482e-01
5.62917531e-01 -1.62341297e-01 1.34719491e+00 -1.93435088e-01
6.96477652e-01 -3.30717117e-02 -2.60498971e-01 -3.38251382e-01
-1.03732002e+00 3.63650173e-01 6.71409220e-02 -2.80253589e-01
2.50995606e-01 -7.41844654e-01 -4.30712551e-01 2.78909355e-01
-4.43656832e-01 6.43101931e-01 6.47374988e-01 1.10853565e+00
6.56877339e-01 -3.60743910e-01 8.03847373e-01 -1.06235790e+00
4.51714903e-01 -5.53247452e-01 -4.12866175e-01 3.48588109e-01
-6.28477573e-01 4.02189523e-01 1.25057471e+00 -5.51794529e-01
-1.21888721e+00 -3.33053097e-02 -2.85168678e-01 -4.67956573e-01
1.16677620e-01 4.88599986e-01 -3.15888166e-01 2.65069127e-01
5.86792648e-01 7.76400566e-01 -1.99666321e-01 -6.50626957e-01
5.98394811e-01 6.54292583e-01 6.66576326e-01 -6.85704052e-01
1.01617765e+00 1.78416893e-01 -5.20758390e-01 -8.77922343e-04
-9.99466121e-01 1.92736894e-01 -5.94001591e-01 1.40392736e-01
7.35932410e-01 -1.35921693e+00 -2.26609647e-01 2.68663645e-01
-1.70358157e+00 -6.97057784e-01 -2.81743586e-01 5.46036720e-01
-7.80613244e-01 -4.19671135e-03 -7.51026154e-01 -3.81174088e-01
-3.80408943e-01 -1.38763142e+00 9.07031119e-01 4.37668040e-02
5.97943217e-02 -1.26521289e+00 1.31490886e-01 3.23059261e-01
4.44832802e-01 1.52288228e-02 8.61072481e-01 -8.33104372e-01
-8.02029550e-01 4.10241932e-02 9.55833271e-02 5.77671170e-01
-1.15773737e-01 -5.58586895e-01 -8.48915815e-01 -3.92300546e-01
1.17682502e-01 -4.49089170e-01 9.99232650e-01 -2.23536462e-01
1.00679052e+00 -6.72334433e-01 -4.46455330e-01 7.61724651e-01
1.31769228e+00 2.87950248e-01 8.31826985e-01 1.77393079e-01
7.06819892e-01 1.74852327e-01 3.81015301e-01 -1.58968642e-01
5.66029549e-01 5.53485155e-01 3.82815301e-01 -5.57579705e-03
-4.22250479e-01 -9.22613859e-01 1.14439106e+00 1.58830905e+00
-7.45363310e-02 -3.09939772e-01 -9.00280416e-01 5.52916586e-01
-1.78727949e+00 -6.89472377e-01 1.65801018e-01 1.94601405e+00
1.51966631e+00 4.07604612e-02 -5.99969387e-01 -5.06502628e-01
4.55075204e-01 2.03758672e-01 -6.69073403e-01 -4.78967041e-01
-7.74701089e-02 1.62114292e-01 6.30839467e-01 7.13781297e-01
-6.99286282e-01 1.55571461e+00 6.69650507e+00 7.92558849e-01
-1.15593100e+00 5.10477364e-01 6.42276943e-01 1.05721354e-01
-5.83506227e-01 2.91075498e-01 -9.15146768e-01 5.31706572e-01
1.30428123e+00 -3.59148860e-01 9.30206299e-01 6.51336491e-01
1.80164143e-01 4.07928884e-01 -1.54571569e+00 5.58192790e-01
-1.19484961e-01 -1.14350402e+00 9.37781855e-02 -8.93358737e-02
1.15269852e+00 4.72458720e-01 1.73624590e-01 7.79220223e-01
6.96354449e-01 -1.26102781e+00 6.52948201e-01 4.61353362e-01
1.13601267e+00 -5.91763198e-01 6.74322307e-01 9.45381641e-01
-5.32424986e-01 1.35025218e-01 -5.71115613e-01 -3.21924299e-01
3.50059718e-01 2.00962126e-01 -1.14968085e+00 6.57226324e-01
-3.06390729e-02 9.29824889e-01 -2.59819835e-01 3.82233709e-01
-8.22540283e-01 1.04333997e+00 -3.65964249e-02 2.50834197e-01
3.31954569e-01 -1.73732564e-01 5.53065538e-01 1.12642252e+00
2.63394952e-01 -1.25286251e-01 9.71246213e-02 7.91078746e-01
-6.56956971e-01 -1.77184105e-01 -8.29579651e-01 -1.40823975e-01
2.80262917e-01 4.77558315e-01 8.83212686e-02 -3.71917635e-01
-5.76219141e-01 1.22016919e+00 6.16305113e-01 7.31332719e-01
-9.92533863e-01 -1.10771647e-02 4.87650275e-01 -6.81015775e-02
1.60370857e-01 -2.95171112e-01 -3.01434666e-01 -1.64195681e+00
1.53915519e-02 -1.28794038e+00 -2.23158285e-01 -7.51300514e-01
-9.35277820e-01 5.96764803e-01 -2.39651501e-01 -1.12650669e+00
-5.46640277e-01 -5.45888662e-01 -6.11183167e-01 1.12112963e+00
-1.85658908e+00 -1.16990042e+00 6.69356287e-01 5.73452055e-01
7.20997274e-01 -3.21653485e-01 1.04004359e+00 1.28084645e-01
-4.28463131e-01 7.88742661e-01 4.98936981e-01 4.26217139e-01
7.23281622e-01 -1.19367242e+00 8.52633059e-01 9.94407952e-01
3.02493125e-01 5.69745839e-01 6.57796741e-01 -7.26604998e-01
-1.64210868e+00 -1.28425694e+00 1.19369721e+00 -7.60468006e-01
8.35835755e-01 -4.82732713e-01 -9.33854878e-01 1.31365800e+00
6.36839569e-01 -2.93306291e-01 3.29129547e-01 3.71687748e-02
-3.80130053e-01 1.94954038e-01 -6.64575934e-01 7.24779546e-01
1.09797013e+00 -7.65345156e-01 -9.92752850e-01 6.38799012e-01
1.44945288e+00 -9.46884930e-01 -7.87068844e-01 3.72678578e-01
2.04961866e-01 -3.31796825e-01 8.00551176e-01 -9.86201406e-01
9.33713078e-01 -1.14050150e-01 -8.47351700e-02 -1.77242792e+00
-5.54357329e-03 -1.11546659e+00 -6.28576219e-01 1.12715840e+00
9.88031685e-01 -6.83354974e-01 6.46920443e-01 3.84861767e-01
-3.11292857e-01 -8.94497037e-01 -1.13025498e+00 -8.99700642e-01
8.07244658e-01 -4.37651664e-01 5.73539376e-01 8.46203506e-01
-3.33271846e-02 5.58468342e-01 -8.75271320e-01 8.68957788e-02
1.71121478e-01 9.61312354e-02 6.52648985e-01 -4.62635130e-01
-8.44928503e-01 -8.48048925e-02 9.51912701e-02 -1.72588742e+00
4.97026443e-01 -1.46873236e+00 3.07852536e-01 -1.57688928e+00
1.77859783e-01 -1.87395498e-01 -3.33837301e-01 6.16939306e-01
-3.45568597e-01 -7.41733387e-02 2.61456817e-01 2.98727363e-01
-4.25546944e-01 8.54922116e-01 2.01173282e+00 -2.06675455e-01
6.25608675e-03 1.28054544e-01 -7.34384835e-01 5.62686443e-01
7.21888781e-01 -6.40893877e-01 -4.21571821e-01 -1.05650592e+00
3.14017117e-01 5.29151320e-01 -4.05295677e-02 -5.31490028e-01
1.80429503e-01 -1.34373382e-01 1.59406021e-01 -2.30611250e-01
2.54523158e-01 -6.33044958e-01 -4.09721658e-02 3.36119801e-01
-5.91381311e-01 6.49445457e-03 4.41990532e-02 6.98455215e-01
-3.30274165e-01 6.97416365e-02 5.47964871e-01 -2.27315679e-01
-3.62737179e-01 3.99074942e-01 -2.61751264e-01 3.48683000e-01
6.13910556e-01 3.40867251e-01 -3.54654610e-01 -6.21110201e-01
-7.32861936e-01 5.49519539e-01 3.09813589e-01 5.12204587e-01
4.98364478e-01 -1.24690866e+00 -1.01049590e+00 1.04072891e-01
-2.48305976e-01 1.80003896e-01 -2.33157441e-01 8.62081170e-01
-2.26192683e-01 5.58723807e-01 3.26445818e-01 -2.36665115e-01
-6.05750144e-01 6.80908561e-01 5.15838504e-01 -6.72747195e-01
-7.68768787e-01 7.99164414e-01 4.88194019e-01 -7.75263786e-01
1.83176070e-01 -6.10127114e-02 1.30262941e-01 -5.29610276e-01
4.09409583e-01 -8.58363882e-02 -8.96737874e-02 -4.65342402e-01
-2.72847004e-02 6.02448173e-02 -3.01922739e-01 -3.90033245e-01
1.09634948e+00 -2.30750829e-01 9.61264521e-02 3.81820142e-01
1.20125675e+00 -7.68275261e-02 -1.45965874e+00 -5.57741165e-01
-8.41119140e-02 -6.38110563e-02 -1.43440232e-01 -1.16832733e+00
-9.42578018e-01 1.34351718e+00 -2.73711979e-02 -3.64766091e-01
1.02090371e+00 -1.29447311e-01 1.15772235e+00 7.29186237e-01
5.18898547e-01 -8.49955738e-01 -2.82613963e-01 1.09461713e+00
9.62344289e-01 -1.25283408e+00 -5.27758241e-01 -4.92356159e-02
-8.26307297e-01 1.06916666e+00 6.15360320e-01 -1.49473578e-01
5.00733733e-01 3.13845873e-01 2.60042667e-01 4.31678534e-01
-1.32514989e+00 7.88573697e-02 1.76763579e-01 5.29334664e-01
3.74175817e-01 1.93174869e-01 -2.09078088e-01 7.28384912e-01
-4.27532554e-01 1.97770208e-01 3.67916048e-01 4.27149385e-01
-1.18824713e-01 -1.43176401e+00 -1.72897766e-03 1.25188977e-01
-7.41386235e-01 -5.07766724e-01 -3.41546834e-01 6.13276541e-01
-5.35667241e-02 7.43626833e-01 -3.34324181e-01 -6.45808816e-01
1.80667922e-01 3.89297813e-01 3.61923128e-01 -7.29687452e-01
-6.74684048e-01 -6.31714612e-02 1.67408854e-01 -4.02955174e-01
6.72046915e-02 -3.45654428e-01 -1.18401241e+00 -2.56677598e-01
-4.21529055e-01 2.80547649e-01 5.00701249e-01 1.25423610e+00
3.91789377e-01 5.91233611e-01 7.03983665e-01 -4.18466538e-01
-1.13219059e+00 -1.10935020e+00 -1.13759539e-03 2.36936748e-01
4.37842607e-01 -1.75743371e-01 -2.27263197e-01 3.65354478e-01] | [11.705164909362793, 10.077173233032227] |
c664645d-5b11-4da3-bc1d-49fab28a2aea | joint-learning-of-intrinsic-images-and | 1807.11857 | null | http://arxiv.org/abs/1807.11857v1 | http://arxiv.org/pdf/1807.11857v1.pdf | Joint Learning of Intrinsic Images and Semantic Segmentation | Semantic segmentation of outdoor scenes is problematic when there are
variations in imaging conditions. It is known that albedo (reflectance) is
invariant to all kinds of illumination effects. Thus, using reflectance images
for semantic segmentation task can be favorable. Additionally, not only
segmentation may benefit from reflectance, but also segmentation may be useful
for reflectance computation. Therefore, in this paper, the tasks of semantic
segmentation and intrinsic image decomposition are considered as a combined
process by exploring their mutual relationship in a joint fashion. To that end,
we propose a supervised end-to-end CNN architecture to jointly learn intrinsic
image decomposition and semantic segmentation. We analyze the gains of
addressing those two problems jointly. Moreover, new cascade CNN architectures
for intrinsic-for-segmentation and segmentation-for-intrinsic are proposed as
single tasks. Furthermore, a dataset of 35K synthetic images of natural
environments is created with corresponding albedo and shading (intrinsics), as
well as semantic labels (segmentation) assigned to each object/scene. The
experiments show that joint learning of intrinsic image decomposition and
semantic segmentation is beneficial for both tasks for natural scenes. Dataset
and models are available at: https://ivi.fnwi.uva.nl/cv/intrinseg | ['Hoang-An Le', 'Thomas T. Groenestege', 'Theo Gevers', 'Sezer Karaoglu', 'Anil S. Baslamisli', 'Partha Das'] | 2018-07-31 | joint-learning-of-intrinsic-images-and-1 | http://openaccess.thecvf.com/content_ECCV_2018/html/Anil_Baslamisli_Joint_Learning_of_ECCV_2018_paper.html | http://openaccess.thecvf.com/content_ECCV_2018/papers/Anil_Baslamisli_Joint_Learning_of_ECCV_2018_paper.pdf | eccv-2018-9 | ['intrinsic-image-decomposition'] | ['computer-vision'] | [ 5.94442129e-01 -1.19596057e-01 2.31879309e-01 -7.44607806e-01
-4.32307422e-01 -3.84885997e-01 1.92856148e-01 -2.68681616e-01
-2.65095472e-01 4.37951297e-01 -5.92327528e-02 -1.08237348e-01
2.23172113e-01 -1.03350759e+00 -8.54108751e-01 -1.02558041e+00
4.54850197e-01 8.02816376e-02 6.00120761e-02 -1.66601911e-01
-3.28085050e-02 3.42843384e-01 -1.58030093e+00 1.96612224e-01
1.17574370e+00 1.11539292e+00 5.86829603e-01 6.78845465e-01
-3.11472625e-01 5.24743080e-01 -4.82002228e-01 -5.97903244e-02
5.69155633e-01 -4.33014959e-01 -8.02748859e-01 4.66985911e-01
3.97372335e-01 -3.66529286e-01 8.34669024e-02 1.20959961e+00
2.97579676e-01 3.17598164e-01 5.68384647e-01 -1.10541236e+00
-6.15002632e-01 2.94963062e-01 -8.01791489e-01 -2.57484347e-01
-5.34685962e-02 3.02058995e-01 9.62317050e-01 -7.26220191e-01
1.33703023e-01 1.19847131e+00 2.58658946e-01 3.03998500e-01
-1.18475187e+00 -4.49568897e-01 3.50455076e-01 -1.16709381e-01
-1.13602102e+00 -4.13169950e-01 1.08389759e+00 -4.12846982e-01
5.54110289e-01 3.03418994e-01 7.71515608e-01 8.69743884e-01
-3.16434622e-01 1.03528702e+00 1.32190168e+00 -3.52747530e-01
8.30696225e-02 2.38680065e-01 2.86650538e-01 5.41639209e-01
1.80360138e-01 -1.78291008e-01 -3.99306357e-01 4.39814359e-01
7.03449309e-01 9.09731090e-02 -3.98832500e-01 -8.79884884e-02
-1.00870454e+00 5.76818645e-01 6.25518084e-01 -2.16815084e-01
-3.59171927e-01 1.91777721e-01 8.31116363e-02 1.62055552e-01
6.69647157e-01 2.30138510e-01 -6.84414685e-01 4.41146076e-01
-6.13489389e-01 1.67356834e-01 5.66515684e-01 9.49618280e-01
1.30574441e+00 1.92361325e-01 4.51974608e-02 1.24072516e+00
6.13413751e-01 7.24600673e-01 8.70951638e-02 -1.16698921e+00
1.94308817e-01 6.97539806e-01 1.83246464e-01 -8.79187167e-01
-5.08681893e-01 -4.52403367e-01 -8.01974118e-01 1.15942217e-01
4.42028612e-01 -5.40895574e-02 -1.20466828e+00 1.80096161e+00
5.93197167e-01 5.34057915e-02 2.63237983e-01 1.15196753e+00
9.72624362e-01 6.38780475e-01 8.29047859e-02 -6.00282215e-02
1.36752307e+00 -1.30332851e+00 -5.60059071e-01 -4.40409392e-01
4.25188035e-01 -9.14083064e-01 1.38935745e+00 3.65153521e-01
-9.23387527e-01 -6.78760290e-01 -8.51129413e-01 -4.05873388e-01
-4.23672587e-01 3.90193313e-01 8.21302712e-01 7.33690321e-01
-8.82886410e-01 2.19828010e-01 -8.36925626e-01 -3.15976650e-01
4.08643931e-01 1.66907102e-01 2.43179530e-01 -7.84964636e-02
-1.13427687e+00 3.01633090e-01 3.35320294e-01 5.89558184e-01
-8.63743126e-01 -3.91318709e-01 -6.28314793e-01 -2.11924165e-01
4.24690723e-01 -6.82274461e-01 1.07388687e+00 -1.40968812e+00
-1.50343370e+00 9.39642966e-01 -1.27269298e-01 -1.05426222e-01
4.57332492e-01 -4.44956094e-01 -7.43836537e-02 2.42437080e-01
1.27511829e-01 7.37584352e-01 9.07218635e-01 -1.70085895e+00
-4.35951650e-01 -5.21477044e-01 3.09337020e-01 6.35167062e-01
-2.41359368e-01 -2.03941524e-01 -6.12203181e-01 -2.96354920e-01
3.60271603e-01 -6.53772295e-01 -3.50242347e-01 5.63864037e-02
-7.17360616e-01 1.20537721e-01 6.36152089e-01 -8.61320913e-01
4.82293874e-01 -2.03145695e+00 -1.94022357e-02 -4.93855327e-02
-8.64355417e-04 8.55386183e-02 -1.98169798e-01 -1.03813095e-03
2.46456843e-02 1.54020160e-01 -5.82851887e-01 -3.07885766e-01
-1.71824485e-01 1.78350180e-01 -1.25201240e-01 4.85278875e-01
2.07511231e-01 8.45875382e-01 -7.41649806e-01 -4.09165889e-01
4.43097591e-01 6.52510285e-01 -2.88034648e-01 3.59952658e-01
-5.00058115e-01 9.93887544e-01 -6.17760241e-01 7.34868407e-01
9.97048020e-01 -5.91865219e-02 -2.34410726e-02 -3.44894469e-01
9.32521448e-02 -3.73957269e-02 -1.06845427e+00 1.63737464e+00
-6.25733554e-01 3.96963149e-01 3.75425726e-01 -1.13407707e+00
1.17465913e+00 -8.66926908e-02 5.50359309e-01 -7.40851343e-01
1.64063632e-01 1.70107827e-01 -3.49029958e-01 -4.97772247e-01
5.08005679e-01 -4.89898175e-02 2.90303648e-01 3.00435811e-01
-1.99725300e-01 -3.68601620e-01 1.09987877e-01 -4.09804918e-02
3.30982953e-01 5.63356459e-01 -1.08560435e-01 -1.55183941e-01
6.67084217e-01 6.31472990e-02 7.02242911e-01 5.07907569e-01
-3.17155756e-02 7.10179746e-01 3.84744793e-01 -3.63426119e-01
-9.31607842e-01 -9.78288531e-01 -1.37009367e-01 9.53034043e-01
6.73706055e-01 1.73495278e-01 -1.04046178e+00 -1.97372928e-01
-3.01030159e-01 6.13453329e-01 -3.05756509e-01 1.06679365e-01
-4.93206531e-01 -1.05851912e+00 2.34660685e-01 4.41618890e-01
1.09189248e+00 -1.10442948e+00 -7.22508311e-01 -1.15164429e-01
-4.39446807e-01 -1.43639326e+00 -1.40285134e-01 1.08359203e-01
-8.46031606e-01 -1.18999410e+00 -6.17591977e-01 -6.13498807e-01
7.29836345e-01 8.88316929e-01 1.03162313e+00 2.06939265e-01
-4.15677965e-01 4.74063665e-01 -4.62235600e-01 -4.14460480e-01
-1.77323297e-01 1.06441379e-02 -3.13269079e-01 3.45328778e-01
1.77898943e-01 -4.82340604e-01 -9.24116433e-01 3.83515507e-01
-1.06286156e+00 6.21369898e-01 3.84041429e-01 6.18111014e-01
8.33553016e-01 -1.47750713e-02 2.27192372e-01 -1.01379740e+00
1.68305144e-01 -1.41604811e-01 -7.25880623e-01 3.31775576e-01
-3.81140411e-01 -4.71553236e-01 6.13244176e-01 6.97729886e-02
-1.67636418e+00 3.50763291e-01 -1.08314529e-01 -2.59793192e-01
-6.88841641e-01 2.49409720e-01 -6.25156045e-01 1.25537694e-01
4.91581053e-01 2.61256248e-01 -7.43020177e-02 -4.51740712e-01
5.07140875e-01 7.36617744e-01 2.50691146e-01 -7.25978017e-01
7.30149984e-01 7.78195739e-01 -5.76993413e-02 -1.36628199e+00
-1.14121044e+00 -5.40229619e-01 -7.42068768e-01 -3.34043920e-01
1.25771523e+00 -1.23182595e+00 -6.22764409e-01 1.01606762e+00
-1.07044065e+00 -7.90810406e-01 -2.75281042e-01 2.89672226e-01
-5.32774508e-01 4.78128552e-01 -3.63372982e-01 -8.99267137e-01
-3.14996272e-01 -1.48358500e+00 1.23108935e+00 5.61270177e-01
5.06828904e-01 -9.84568179e-01 -3.83161604e-01 9.86831129e-01
1.63717151e-01 3.06065500e-01 5.50521076e-01 -1.26097217e-01
-9.39391971e-01 2.61570275e-01 -7.59913087e-01 8.79469335e-01
1.56802624e-01 4.27074730e-02 -1.36752999e+00 -6.81406930e-02
2.67519057e-01 -4.24389154e-01 1.16549528e+00 5.94171464e-01
1.44056654e+00 -9.45480987e-02 1.50090486e-01 1.04270959e+00
1.69815445e+00 -8.73770192e-02 6.36747003e-01 3.33106101e-01
1.22699249e+00 1.15470970e+00 7.43465006e-01 2.60841429e-01
4.01339561e-01 4.54528838e-01 6.60026073e-01 -7.37517595e-01
-2.47858793e-01 1.91466630e-01 3.17267179e-01 6.34145319e-01
-1.02990806e-01 -3.39339972e-01 -9.27295983e-01 4.41571236e-01
-1.72724462e+00 -4.72463936e-01 -7.04262495e-01 1.95082271e+00
7.29845881e-01 -2.93258041e-01 -1.21795870e-01 -1.15901932e-01
7.14399219e-01 4.36094075e-01 -7.07157671e-01 -2.08232313e-01
-3.59626681e-01 2.67047919e-02 7.40579128e-01 5.30110300e-01
-1.11484802e+00 1.17870510e+00 4.96613216e+00 6.88049376e-01
-1.10481179e+00 1.36864617e-01 1.10887277e+00 2.25644097e-01
-4.84961629e-01 9.54706967e-02 -6.64352298e-01 2.10720420e-01
4.99605328e-01 5.66759825e-01 6.51303768e-01 5.49977422e-01
4.41578448e-01 -5.11082470e-01 -5.43135166e-01 8.29194844e-01
-1.10403687e-01 -7.38105476e-01 2.53369194e-02 -1.32529169e-01
8.60872686e-01 7.71075711e-02 4.13040891e-02 -2.06981003e-01
2.26746500e-01 -9.75690365e-01 6.07774258e-01 5.65615594e-01
5.74564755e-01 -4.64845866e-01 4.89215165e-01 2.19783187e-01
-1.18071640e+00 1.15290232e-01 -5.42422771e-01 8.18798244e-02
9.96323675e-02 9.30532813e-01 -3.77961695e-01 8.11763585e-01
7.65238643e-01 8.30624342e-01 -5.34515738e-01 5.99551380e-01
-6.50167108e-01 6.63853884e-01 -2.88594335e-01 3.69759738e-01
2.10708916e-01 -8.24312747e-01 2.67351806e-01 1.03826022e+00
6.62730890e-04 2.36885417e-02 4.90253061e-01 1.32473254e+00
-6.45015389e-02 2.52072781e-01 -3.06176066e-01 5.50088100e-02
-1.39992498e-02 1.50202656e+00 -1.03994167e+00 -7.58014023e-02
-3.99259478e-01 1.01935887e+00 2.02440023e-02 1.01410830e+00
-8.50748777e-01 -3.02347153e-01 7.27287233e-01 -1.20821021e-01
-6.86089545e-02 -3.17329794e-01 -5.43588817e-01 -1.21638584e+00
2.55645271e-02 -5.22461057e-01 3.42927314e-02 -9.22163904e-01
-9.80608642e-01 1.74882606e-01 -1.75031573e-01 -9.78429914e-01
5.60962856e-01 -7.23619103e-01 -6.03776038e-01 1.14327765e+00
-2.16282248e+00 -1.45723891e+00 -9.46308553e-01 6.68371081e-01
8.30712616e-01 2.81144857e-01 5.37837863e-01 1.44838601e-01
-7.23507702e-01 3.40862311e-02 1.86578363e-01 2.63321787e-01
5.23391426e-01 -1.29651892e+00 1.68375954e-01 9.62836206e-01
-6.27468750e-02 3.28567952e-01 4.97021884e-01 -4.79315907e-01
-1.32910502e+00 -1.41529250e+00 8.19835439e-02 8.09635967e-04
3.02004486e-01 -3.60818446e-01 -7.86318004e-01 3.70586812e-01
9.23778862e-02 -1.46074861e-01 5.78090847e-01 -7.91449659e-03
-3.03448349e-01 -3.69771302e-01 -9.37910438e-01 6.33316934e-01
9.40620363e-01 -4.76206720e-01 1.24258384e-01 7.98895597e-01
1.13618565e+00 -4.01728809e-01 -7.15663195e-01 3.55735689e-01
3.88959795e-01 -1.20978296e+00 1.17125094e+00 -8.94650072e-02
6.36479974e-01 -4.74827558e-01 -4.51987952e-01 -1.12451267e+00
1.97854742e-01 -2.94432372e-01 3.17415446e-01 1.21876550e+00
3.50646853e-01 -7.18022525e-01 8.07186246e-01 4.70008194e-01
-2.98168391e-01 -5.04784822e-01 -3.88949811e-01 -6.10809565e-01
2.31925547e-02 -5.89628637e-01 5.88512540e-01 8.59500766e-01
-9.87863064e-01 4.23832059e-01 -3.94668460e-01 3.51482213e-01
8.34862590e-01 4.34308261e-01 9.84655023e-01 -1.05807436e+00
-3.04042190e-01 -4.27936465e-01 1.70002550e-01 -1.21137428e+00
1.57797545e-01 -8.61333609e-01 3.11737180e-01 -1.86574769e+00
1.64477065e-01 -4.95465249e-01 -1.25054389e-01 4.32775706e-01
-2.77304500e-01 3.05888653e-01 6.43444061e-03 3.94868404e-01
-3.43457401e-01 7.35968649e-01 1.65731728e+00 -2.65687317e-01
-4.08436656e-01 1.41887972e-02 -6.76709890e-01 7.82694817e-01
1.19184792e+00 -1.12898722e-01 -6.03711605e-01 -8.05744529e-01
1.31024927e-01 -2.67250799e-02 5.57756424e-01 -7.61822939e-01
-1.53870746e-01 -4.23961788e-01 2.52540946e-01 -6.18253410e-01
5.65632641e-01 -6.01527691e-01 -1.40759781e-01 1.34780720e-01
-2.00072095e-01 -7.55133331e-01 -1.29324764e-01 5.62554002e-01
-2.21633971e-01 -1.85695559e-01 9.50274765e-01 -4.49791431e-01
-7.60491788e-01 3.37393284e-01 -2.65754443e-02 -1.63448974e-02
7.64430165e-01 -5.65730691e-01 -3.28009397e-01 -2.37667844e-01
-5.55956662e-01 4.11525637e-01 5.22802174e-01 1.26059160e-01
5.67093194e-01 -6.80158019e-01 -6.15012765e-01 1.66658256e-02
3.97845134e-02 6.55121744e-01 5.82401514e-01 9.69730496e-01
-7.63355911e-01 2.23835302e-03 -2.01693580e-01 -7.42126524e-01
-1.12240100e+00 1.72385722e-01 4.84213322e-01 5.73549904e-02
-3.86374474e-01 9.55686331e-01 8.27762604e-01 -8.53259325e-01
1.41555697e-01 -2.96454549e-01 -1.93249837e-01 2.70925350e-02
1.23143427e-01 2.78050661e-01 -1.34220377e-01 -7.78113782e-01
6.88361051e-03 7.49650598e-01 2.42496118e-01 1.19409949e-01
1.42981470e+00 -5.10146558e-01 -4.94406074e-01 4.61497635e-01
1.16665018e+00 -3.60672504e-01 -1.45273888e+00 -3.15787405e-01
-3.35260749e-01 -6.23171389e-01 7.79079944e-02 -7.35501528e-01
-1.55457950e+00 1.29849529e+00 4.88294274e-01 3.37443389e-02
1.55871880e+00 -2.89137691e-01 8.85259748e-01 3.44880760e-01
1.47987843e-01 -1.24157333e+00 1.79702580e-01 3.97721857e-01
6.28062069e-01 -1.47737539e+00 3.86213847e-02 -8.71320248e-01
-6.70911491e-01 1.09779561e+00 7.82983959e-01 1.12280831e-01
4.11559373e-01 -1.97229832e-01 3.42531323e-01 -2.31285036e-01
-1.87534809e-01 -5.74614882e-01 2.27450788e-01 6.23032808e-01
4.88469094e-01 3.26862574e-01 -1.13085352e-01 2.97112018e-01
3.51596326e-02 -5.01455426e-01 6.32484674e-01 6.25832558e-01
-5.07985532e-01 -9.35580075e-01 -4.51052964e-01 3.57522130e-01
-2.97760934e-01 -1.02594614e-01 -3.08302045e-01 3.90074015e-01
1.79388255e-01 1.10431969e+00 -1.99677765e-01 -2.48047367e-01
7.65795559e-02 -1.50404960e-01 1.68302000e-01 -6.08819187e-01
-4.31488961e-01 4.65941072e-01 -1.09366244e-02 -5.70724070e-01
-8.20713103e-01 -4.61244524e-01 -1.33221984e+00 6.27490878e-02
-4.23429698e-01 -2.35639676e-01 9.90171075e-01 7.34458864e-01
-1.65277958e-01 7.69086838e-01 8.32298100e-01 -8.52179945e-01
-1.51654407e-01 -6.87056184e-01 -8.21635187e-01 4.21896577e-01
1.78467467e-01 -4.60617363e-01 -3.98657739e-01 2.96780318e-01] | [9.969810485839844, -2.478238582611084] |
df9273f6-84f0-4eb5-b8bc-eee25679dfaa | ugent-t2k-at-the-2nd-dialdoc-shared-task-a | null | null | https://aclanthology.org/2022.dialdoc-1.12 | https://aclanthology.org/2022.dialdoc-1.12.pdf | UGent-T2K at the 2nd DialDoc Shared Task: A Retrieval-Focused Dialog System Grounded in Multiple Documents | This work presents the contribution from the Text-to-Knowledge team of Ghent University (UGent-T2K) to the MultiDoc2Dial shared task on modeling dialogs grounded in multiple documents. We propose a pipeline system, comprising (1) document retrieval, (2) passage retrieval, and (3) response generation. We engineered these individual components mainly by, for (1)-(2), combining multiple ranking models and adding a final LambdaMART reranker, and, for (3), by adopting a Fusion-in-Decoder (FiD) model. We thus significantly boost the baseline system’s performance (over +10 points for both F1 and SacreBLEU). Further, error analysis reveals two major failure cases, to be addressed in future work: (i) in case of topic shift within the dialog, retrieval often fails to select the correct grounding document(s), and (ii) generation sometimes fails to use the correctly retrieved grounding passage. Our code is released at this link. | ['Chris Develder', 'Thomas Demeester', 'Johannes Deleu', 'Amir Hadifar', 'Yiwei Jiang'] | null | null | null | null | dialdoc-acl-2022-5 | ['passage-retrieval'] | ['natural-language-processing'] | [ 1.63514331e-01 4.47156668e-01 3.67577851e-01 -3.86337340e-01
-1.48458362e+00 -7.55381525e-01 1.12954855e+00 4.51778501e-01
-4.98713821e-01 9.70854163e-01 7.57108748e-01 -3.24135244e-01
-1.90467816e-02 -3.49447727e-01 -7.07138717e-01 -2.30786115e-01
2.15834275e-01 1.11619377e+00 6.27967954e-01 -5.56533456e-01
5.79386652e-01 -4.66805473e-02 -1.22648489e+00 5.93541682e-01
9.06898022e-01 4.31552500e-01 3.59033644e-01 1.09619415e+00
-3.92525107e-01 9.23909009e-01 -7.14313507e-01 -5.04622102e-01
-1.38389304e-01 -5.84058762e-01 -1.30309939e+00 -6.05384186e-02
4.43845928e-01 -4.79946107e-01 5.21006733e-02 6.91764712e-01
5.23219228e-01 4.10208434e-01 7.21171975e-01 -8.46475601e-01
-2.57944614e-01 8.54661524e-01 -1.38870612e-01 -1.05516136e-01
7.30790615e-01 -1.75378732e-02 1.10238373e+00 -9.16621506e-01
7.98085332e-01 1.54049182e+00 3.66408646e-01 3.61729860e-01
-1.26289392e+00 -2.02570558e-01 1.31020725e-01 -2.05547348e-01
-1.17614388e+00 -7.99823880e-01 1.50129050e-01 -4.40571487e-01
1.24068952e+00 2.88463324e-01 2.13958055e-01 1.02634120e+00
-6.92072511e-02 8.48176897e-01 9.45324659e-01 -7.18576431e-01
1.03385076e-01 5.31261086e-01 4.02524352e-01 4.33860809e-01
-2.37142444e-01 -1.57594711e-01 -6.62061512e-01 -4.82034385e-01
1.93438068e-01 -4.10922498e-01 -2.62431294e-01 3.20340395e-01
-1.11240137e+00 9.07281458e-01 8.57747719e-02 3.38291824e-01
-4.32989836e-01 -2.95549799e-02 1.75748691e-01 4.01054204e-01
5.25229216e-01 6.18156850e-01 -5.34023881e-01 -3.80187780e-01
-1.16866636e+00 9.97307301e-01 1.39454734e+00 1.12339377e+00
6.85381591e-01 -6.12045586e-01 -7.36724734e-01 1.09945238e+00
6.77109718e-01 1.83494374e-01 4.81480122e-01 -7.06003547e-01
6.50785923e-01 4.95321274e-01 6.23174489e-01 -6.25886023e-01
-5.20600200e-01 -1.73037931e-01 -1.09188095e-01 -3.24673355e-01
5.56460083e-01 -4.21288788e-01 -6.45283580e-01 1.56008565e+00
2.65528589e-01 -3.24063808e-01 2.76999176e-01 8.49714696e-01
9.95396852e-01 8.08335364e-01 2.00079873e-01 -1.77607849e-01
1.64807439e+00 -1.17990589e+00 -7.42334425e-01 -3.38898122e-01
8.21110189e-01 -1.30452120e+00 8.22670400e-01 9.24945772e-02
-1.32321668e+00 -3.60786557e-01 -8.91811311e-01 -3.79045188e-01
-4.03834313e-01 1.32442757e-01 2.66822398e-01 3.10986012e-01
-1.48277962e+00 4.34651554e-01 -4.51744586e-01 -7.94330001e-01
-6.04166090e-01 2.34556720e-01 -2.32903212e-02 -7.96793401e-02
-1.62833846e+00 1.03122318e+00 2.85555512e-01 -1.30010009e-01
-7.78088212e-01 -5.35767436e-01 -5.66969693e-01 -6.61660135e-02
4.44811910e-01 -8.39598954e-01 1.60578799e+00 -2.77410388e-01
-1.68920982e+00 7.89487302e-01 -2.96299577e-01 -4.00241464e-01
7.60906577e-01 -5.60871601e-01 -8.95266682e-02 -4.98019904e-02
1.58123016e-01 8.41979861e-01 3.48045945e-01 -1.24522924e+00
-8.58666658e-01 -2.73570240e-01 2.35100254e-01 6.31455600e-01
2.45777108e-02 2.90744841e-01 -6.89656019e-01 -4.19920504e-01
3.15142050e-02 -1.15293956e+00 -3.99594121e-02 -6.97806358e-01
-6.68001354e-01 -5.34795821e-01 2.94636875e-01 -9.65694785e-01
1.39249086e+00 -1.74728453e+00 3.81039470e-01 -6.02024645e-02
-1.66485775e-02 5.30871004e-02 -2.59676054e-02 9.93034959e-01
3.72840524e-01 1.19479850e-01 5.19155860e-02 -6.90481842e-01
8.97663981e-02 -3.03685516e-01 -5.11153877e-01 -1.34489402e-01
8.78156945e-02 6.75001264e-01 -9.34471667e-01 -3.59590560e-01
-1.05025120e-01 2.17841849e-01 -4.52572972e-01 1.97294384e-01
-6.68507099e-01 3.06374341e-01 -5.14487982e-01 1.32777274e-01
2.61499882e-01 -1.65747955e-01 1.45107716e-01 -3.60142887e-02
-4.40474093e-01 8.31194460e-01 -1.15963089e+00 1.65984762e+00
-4.01817143e-01 3.89740914e-01 2.57507920e-01 1.53422151e-02
7.44325578e-01 5.06028533e-01 6.91770390e-02 -4.35819894e-01
-1.46848157e-01 2.84341276e-01 -2.34775528e-01 -3.72484654e-01
1.30628335e+00 3.17883976e-02 -1.00993790e-01 7.92813420e-01
2.36711606e-01 -1.78049773e-01 4.10109252e-01 7.80849636e-01
9.03250635e-01 1.16511099e-01 -1.22467361e-01 -3.43834102e-01
4.13614959e-01 1.82881430e-01 -5.03860079e-02 9.65932906e-01
8.24317560e-02 5.74795067e-01 6.57528043e-01 1.90538973e-01
-7.56418943e-01 -5.19093990e-01 1.01325689e-02 1.47678590e+00
-1.12140402e-01 -5.56530356e-01 -8.58472109e-01 -5.95663249e-01
-7.25999847e-03 1.21341991e+00 -3.60958844e-01 5.51727600e-02
-5.77396393e-01 -5.88722706e-01 7.03796506e-01 2.84298211e-01
3.56108516e-01 -9.01061177e-01 -1.10934995e-01 4.20914978e-01
-7.37684071e-01 -1.00121546e+00 -5.56991100e-01 2.27375746e-01
-6.77864313e-01 -6.32680714e-01 -9.43193316e-01 -4.20838326e-01
1.17458776e-01 3.39722067e-01 1.14986122e+00 4.17633712e-01
3.50519687e-01 4.31012154e-01 -5.14802516e-01 -4.05923128e-01
-7.49884963e-01 3.23421896e-01 -1.73943967e-01 -3.34157050e-01
2.93349981e-01 1.11506984e-01 -5.19403279e-01 3.61387044e-01
-7.07834005e-01 7.67080560e-02 3.51426005e-01 7.81512737e-01
1.38330385e-01 -3.68285477e-01 6.68665588e-01 -1.00007713e+00
1.08348262e+00 -5.08811533e-01 -5.99914968e-01 3.67438972e-01
-5.26398301e-01 1.15037933e-01 6.64574802e-02 2.79398989e-02
-1.30915105e+00 -2.67212957e-01 -3.16176295e-01 2.04199046e-01
-1.01582378e-01 6.86799824e-01 1.82265550e-01 4.22965854e-01
8.46982360e-01 1.30962089e-01 -6.47039637e-02 -6.43796802e-01
5.98554552e-01 9.65124846e-01 1.93642661e-01 -6.36085451e-01
2.95189500e-01 -5.81782646e-02 -5.56007326e-01 -8.68215144e-01
-7.24348783e-01 -6.48257971e-01 -5.23738682e-01 -3.09640378e-01
8.85636568e-01 -1.16588163e+00 -3.67627770e-01 3.17074448e-01
-1.36377299e+00 -4.68646079e-01 1.46621361e-01 4.89283383e-01
-3.08732092e-01 2.40571365e-01 -8.60507786e-01 -1.04400373e+00
-3.94587696e-01 -1.17759836e+00 1.30330920e+00 4.30810958e-01
-6.34040773e-01 -1.03135574e+00 3.04686397e-01 7.86734283e-01
4.96361494e-01 -3.72384310e-01 8.82451475e-01 -1.20162380e+00
-4.98628795e-01 -2.11222455e-01 -6.26550838e-02 -1.47655919e-01
-2.27290913e-01 3.47758899e-03 -1.17961073e+00 -2.95622051e-01
-2.88783371e-01 -5.11091709e-01 8.32844913e-01 1.78138196e-01
1.37643278e-01 -1.54241070e-01 -4.35025066e-01 -2.24123925e-01
8.88612509e-01 1.49261519e-01 5.42076468e-01 3.24248493e-01
2.45402202e-01 9.49397206e-01 6.05740130e-01 3.42711270e-01
8.25297058e-01 1.12313259e+00 -9.04899910e-02 1.53080896e-01
-2.16615111e-01 -3.18043172e-01 5.66942632e-01 8.28668058e-01
1.53595090e-01 -6.76272511e-01 -7.28348017e-01 6.73383594e-01
-2.01387572e+00 -7.10085273e-01 -3.91157657e-01 2.39583111e+00
1.00401890e+00 1.30628437e-01 3.01734537e-01 -5.32193780e-01
5.42968810e-01 2.32103113e-02 -1.45194098e-01 -3.90484244e-01
7.08675534e-02 -2.08138481e-01 2.23000288e-01 1.15723372e+00
-7.48737097e-01 1.30249429e+00 6.01146984e+00 5.63306093e-01
-7.35880613e-01 1.34152904e-01 3.91160548e-01 -8.22575539e-02
-3.79765838e-01 3.60182494e-01 -1.45647860e+00 3.47146928e-01
1.13563848e+00 -2.67545760e-01 3.67153615e-01 4.00171310e-01
1.00663610e-01 -3.41868430e-01 -9.91240740e-01 1.14339799e-01
9.08863544e-02 -9.53394115e-01 7.88996518e-02 1.74747612e-02
3.72247219e-01 2.10122466e-01 -4.20918435e-01 7.81814456e-01
7.81455696e-01 -6.77678406e-01 6.91081822e-01 5.99342406e-01
2.99145132e-01 -4.32531595e-01 7.84748971e-01 4.73668844e-01
-6.63771510e-01 2.73262978e-01 -3.66192877e-01 3.62476110e-01
2.99897522e-01 5.04304886e-01 -1.27948260e+00 7.35723436e-01
3.62565845e-01 2.09166408e-01 -4.15370166e-01 6.97704613e-01
-3.47465128e-01 5.77748001e-01 -3.10074359e-01 -2.95964658e-01
3.78976315e-01 3.40986811e-03 8.59826446e-01 1.56983352e+00
1.44330651e-01 1.58594832e-01 3.31684202e-01 7.47800291e-01
-1.02714494e-01 2.59466738e-01 -1.48247853e-01 -8.33701342e-02
6.12462461e-01 1.21925521e+00 -6.60497725e-01 -4.22785401e-01
-2.01404989e-01 1.06675804e+00 2.60774761e-01 4.17261481e-01
-3.64873111e-01 -5.02375782e-01 2.46241450e-01 -4.24881727e-02
1.28214970e-01 -1.35020763e-02 8.58050026e-03 -8.98020148e-01
-1.94895893e-01 -8.26680362e-01 4.93137628e-01 -7.46809840e-01
-1.05123925e+00 7.10534513e-01 1.59867808e-01 -6.61702812e-01
-8.24563801e-01 -1.88358933e-01 -2.05398053e-01 1.29332769e+00
-1.33647597e+00 -1.00903344e+00 2.17583776e-02 3.79215121e-01
7.25089967e-01 5.07982746e-02 1.03513122e+00 3.00182551e-01
-2.46574074e-01 4.32458073e-01 4.67107184e-02 -1.72074974e-01
1.22722328e+00 -1.48309016e+00 3.55857611e-01 4.08101767e-01
-2.27098055e-02 8.36976409e-01 9.90639031e-01 -9.03528273e-01
-1.17699850e+00 -8.97333145e-01 1.73702955e+00 -6.28001153e-01
5.46200335e-01 -2.90528059e-01 -1.05527401e+00 6.45277381e-01
5.78392267e-01 -1.06269467e+00 6.94296837e-01 5.54889798e-01
-1.04121946e-01 3.50849450e-01 -1.01018167e+00 7.89619327e-01
3.96243811e-01 -3.06319177e-01 -7.24269092e-01 8.78966510e-01
8.05422068e-01 -7.53769636e-01 -7.98119903e-01 -3.76404114e-02
5.48593402e-01 -7.81580150e-01 6.74742281e-01 -4.52872843e-01
4.56409067e-01 -3.74804586e-01 -2.92032748e-01 -1.39741766e+00
-2.45344028e-01 -7.74530113e-01 -5.75254411e-02 1.44577551e+00
7.61186481e-01 -3.22710752e-01 1.88613206e-01 7.41367102e-01
-2.45188132e-01 -4.63786155e-01 -6.92198932e-01 -3.76209378e-01
1.64874941e-01 -1.39352202e-01 3.89228851e-01 6.39315009e-01
2.30643526e-01 9.17476177e-01 -4.07677025e-01 7.54525661e-02
1.19654782e-01 -1.24919005e-01 8.62565160e-01 -1.13766110e+00
-3.97600561e-01 -5.63177347e-01 2.87478089e-01 -1.50417268e+00
-2.12195382e-01 -7.85871744e-01 5.32184780e-01 -1.81494081e+00
1.79027528e-01 -2.76389062e-01 4.02918495e-02 3.63639355e-01
-4.32937980e-01 -1.27840832e-01 2.78629243e-01 4.72671032e-01
-7.96997488e-01 4.41235870e-01 8.84613633e-01 2.10489407e-01
-5.14414310e-01 4.35209870e-02 -9.70787585e-01 3.81153226e-01
3.83651048e-01 -4.49554145e-01 -2.87528604e-01 -5.58511794e-01
2.97342151e-01 5.70268035e-01 1.80160612e-01 -6.74458683e-01
5.19713461e-01 1.34572998e-01 1.40206516e-01 -6.65668607e-01
4.81579065e-01 -2.76469350e-01 -1.15556389e-01 2.70302296e-01
-8.74619186e-01 8.54653716e-02 4.06295598e-01 4.39372241e-01
-6.55473769e-02 -5.27567327e-01 4.10871118e-01 -2.38786265e-01
-2.87292659e-01 -3.42707306e-01 -7.52116501e-01 1.12509392e-01
4.29392934e-01 2.90427566e-01 -6.86356306e-01 -8.53252709e-01
-6.54867530e-01 7.99590766e-01 2.77801126e-01 5.07096112e-01
9.04616266e-02 -8.34739208e-01 -1.03806794e+00 -1.28813475e-01
1.45802841e-01 -1.47453159e-01 2.29628220e-01 8.32413793e-01
-2.86674082e-01 7.94695258e-01 4.63770062e-01 -4.97528374e-01
-1.24448526e+00 -2.74220467e-01 1.04905300e-01 -8.23497415e-01
-4.12242830e-01 1.00032365e+00 -1.85296819e-01 -5.82485020e-01
3.56073081e-01 6.75372556e-02 -5.02102196e-01 4.78413403e-01
8.17642331e-01 3.16211998e-01 4.31805462e-01 -5.92199504e-01
-2.28669032e-01 6.99228197e-02 -7.53068209e-01 -8.60325754e-01
9.50352609e-01 -3.66011918e-01 -1.15948776e-02 5.16769826e-01
8.92830133e-01 5.96227720e-02 -7.88968563e-01 -3.56480062e-01
3.07816565e-01 3.17726731e-02 1.20693348e-01 -1.34982944e+00
-1.77047789e-01 6.80838645e-01 2.69708186e-01 3.94235969e-01
5.63807070e-01 -4.53828387e-02 6.73289120e-01 5.68199575e-01
2.23684296e-01 -1.18579137e+00 -1.15267828e-01 1.04617715e+00
1.06841195e+00 -1.06812382e+00 8.54306296e-02 -8.80303755e-02
-9.78287339e-01 1.15622139e+00 4.91207212e-01 4.16804254e-01
3.41465414e-01 -1.53795471e-02 3.30950320e-01 -3.47310692e-01
-1.22612178e+00 -1.83132648e-01 2.99265802e-01 -2.75675599e-02
9.64623034e-01 -1.34382337e-01 -5.25058389e-01 4.66162324e-01
-3.62728506e-01 -5.67926988e-02 4.09734458e-01 9.89457309e-01
-6.57870293e-01 -1.10886514e+00 -3.32704365e-01 3.62664580e-01
-5.02849996e-01 -3.30228955e-01 -7.48404205e-01 6.24108255e-01
-5.55631101e-01 1.50916851e+00 1.34009616e-02 -3.67139935e-01
4.85343784e-01 4.75663871e-01 8.82194340e-02 -8.30891073e-01
-1.13616288e+00 4.67253506e-01 5.71785212e-01 -1.71048269e-01
-1.61033869e-01 -7.26940036e-01 -1.04355550e+00 -1.00881442e-01
-5.73371410e-01 5.73421657e-01 9.02698278e-01 1.12596369e+00
5.82661569e-01 6.14212871e-01 2.39836156e-01 -6.61312938e-01
-6.52158558e-01 -1.54121244e+00 -3.24660391e-01 7.26487786e-02
7.13258088e-02 -5.24915218e-01 -3.22637051e-01 -1.84603482e-02] | [12.351924896240234, 8.033326148986816] |
77ffa945-fea7-4438-89ab-74b1c0c7b5c2 | trapacc-and-trapaccs-at-parseme-shared-task | null | null | https://aclanthology.org/W18-4930 | https://aclanthology.org/W18-4930.pdf | TRAPACC and TRAPACCS at PARSEME Shared Task 2018: Neural Transition Tagging of Verbal Multiword Expressions | We describe the TRAPACC system and its variant TRAPACCS that participated in the closed track of the PARSEME Shared Task 2018 on labeling verbal multiword expressions (VMWEs). TRAPACC is a modified arc-standard transition system based on Constant and Nivre{'}s (2016) model of joint syntactic and lexical analysis in which the oracle is approximated using a classifier. For TRAPACC, the classifier consists of a data-independent dimension reduction and a convolutional neural network (CNN) for learning and labelling transitions. TRAPACCS extends TRAPACC by replacing the softmax layer of the CNN with a support vector machine (SVM). We report the results obtained for 19 languages, for 8 of which our system yields the best results compared to other participating systems in the closed-track of the shared task. | ['Behrang Qasemizadeh', 'Regina Stodden', 'Laura Kallmeyer'] | 2018-08-01 | null | null | null | coling-2018-8 | ['lexical-analysis'] | ['natural-language-processing'] | [-2.22179756e-01 3.76923084e-01 -6.34783685e-01 -6.61150694e-01
-9.77160573e-01 -7.38762617e-01 5.69462478e-01 3.75640929e-01
-8.65193784e-01 7.53525794e-01 3.01481605e-01 -8.00339460e-01
4.26637143e-01 -3.56326282e-01 -8.00539732e-01 -2.94195920e-01
-1.40160903e-01 6.65483236e-01 2.70367172e-02 1.78036522e-02
-6.23440921e-01 8.66781473e-02 -8.11450481e-01 6.30792737e-01
4.90014970e-01 1.15438843e+00 -2.43994668e-01 7.68417895e-01
-3.69025707e-01 1.18056858e+00 -2.70673096e-01 -8.46288621e-01
-6.35266602e-02 1.17820092e-01 -1.45428729e+00 -5.48271477e-01
3.94518584e-01 1.12420082e-01 -2.70770669e-01 1.05104160e+00
1.02449752e-01 2.74124414e-01 7.16514885e-01 -1.12322605e+00
-6.58837378e-01 1.53522062e+00 -3.93462181e-03 1.10304669e-01
2.35735267e-01 3.12833041e-02 1.29292274e+00 -1.10639894e+00
8.23926866e-01 1.49823487e+00 9.52522993e-01 5.24298906e-01
-1.28110385e+00 -6.64515257e-01 2.37280846e-01 3.45461965e-01
-1.33253336e+00 -5.43859541e-01 4.36690807e-01 -6.40605628e-01
1.82783103e+00 1.35091975e-01 4.26511705e-01 1.63349509e+00
1.25262067e-02 1.07323813e+00 9.83865857e-01 -5.93412101e-01
2.59942174e-01 7.43627399e-02 7.09683180e-01 1.03027785e+00
-1.95401222e-01 -1.34274900e-01 -2.02504829e-01 -3.22723985e-01
2.16979161e-02 -8.74019504e-01 2.11604610e-01 -6.54862002e-02
-7.49552786e-01 1.08719707e+00 1.33078590e-01 2.97181100e-01
-1.23493567e-01 4.38638210e-01 9.90375996e-01 2.99983948e-01
7.10775793e-01 3.47714186e-01 -1.02916145e+00 -3.47756743e-01
-8.27177346e-01 2.67623633e-01 9.02301490e-01 9.67716336e-01
5.27196527e-01 2.46055424e-01 -6.05196118e-01 9.01023865e-01
2.28742629e-01 9.84962285e-02 6.28423572e-01 -1.06141448e+00
6.32799685e-01 4.05132771e-01 -2.16206372e-01 -2.32815996e-01
-6.78568125e-01 -3.23723733e-01 -2.85240233e-01 -2.91254759e-01
6.46306753e-01 -7.02404022e-01 -7.79526830e-01 2.12751365e+00
9.89155322e-02 7.87029788e-02 1.17421232e-01 3.62107247e-01
9.18776155e-01 7.37701476e-01 8.12150180e-01 -1.78596988e-01
1.36927235e+00 -1.24425745e+00 -1.01151085e+00 -3.39478642e-01
1.43377280e+00 -4.50502545e-01 9.32861507e-01 3.72092009e-01
-1.19020092e+00 -3.82761329e-01 -1.02769208e+00 -3.97034049e-01
-8.73663664e-01 2.47139901e-01 9.58637118e-01 5.80280423e-01
-1.06824517e+00 2.48127729e-01 -9.71227109e-01 -4.29609232e-02
1.50517195e-01 2.36638144e-01 -6.86429560e-01 2.17885435e-01
-1.88699007e+00 1.36282969e+00 9.01463389e-01 1.49821177e-01
-4.64550018e-01 -5.45467019e-01 -1.46756971e+00 -8.40722993e-02
4.00392383e-01 -1.28055230e-01 1.63520181e+00 -1.02397561e+00
-1.82899117e+00 1.24418104e+00 -1.61226019e-01 -9.04002309e-01
4.71912563e-01 -2.89659262e-01 -4.44139540e-01 -2.94017076e-01
6.95601404e-02 7.18361855e-01 3.77786875e-01 -4.95807469e-01
-4.35525388e-01 7.55233765e-02 -5.89105450e-02 -2.34708548e-01
-1.69711083e-01 6.05463386e-01 -3.12930435e-01 -3.59825790e-01
-5.37247837e-01 -8.85645449e-01 -5.80313019e-02 -6.23149693e-01
-5.69197595e-01 -1.01507080e+00 5.74820936e-01 -1.01084161e+00
1.45574689e+00 -2.09385324e+00 4.04485643e-01 -1.21425256e-01
-4.92007136e-02 4.48395759e-01 -5.23872711e-02 2.14636222e-01
-5.16239285e-01 3.12622935e-01 -4.21517342e-01 -9.56908047e-01
4.39806670e-01 3.68334711e-01 -2.04197809e-01 5.77799797e-01
3.22087139e-01 1.32895422e+00 -8.86724889e-01 -5.04472494e-01
5.81576899e-02 1.99191213e-01 -3.06629300e-01 3.18949372e-02
-6.23006105e-01 8.37228820e-02 1.03159696e-01 3.71625423e-01
5.54008663e-01 5.82294352e-02 4.96469438e-01 -1.91721749e-02
-2.51917779e-01 5.49167216e-01 -8.83714080e-01 1.70061386e+00
-7.17644453e-01 6.35712028e-01 3.40209305e-02 -1.06565702e+00
4.88840044e-01 5.05850911e-01 3.76672059e-01 -3.41207236e-01
4.71692145e-01 4.55912322e-01 -1.40405610e-01 -4.24968958e-01
5.60293436e-01 -5.98056940e-03 -7.52892256e-01 1.72123730e-01
7.30705559e-01 1.10819116e-01 1.99213937e-01 2.45640919e-01
9.52741861e-01 5.43586373e-01 5.27787924e-01 -5.26921153e-01
6.80009961e-01 7.51506863e-03 4.46308881e-01 4.89291787e-01
-3.77047062e-01 -9.23239067e-03 8.12161684e-01 -4.75032985e-01
-8.06936383e-01 -9.90680873e-01 -4.31270689e-01 1.48666489e+00
-6.48016214e-01 -7.40161717e-01 -8.85681868e-01 -9.52621818e-01
-1.24007076e-01 1.01464999e+00 -8.95757377e-01 -9.46721733e-02
-7.17579603e-01 -6.87017620e-01 1.40891540e+00 6.61725461e-01
3.65265071e-01 -1.19014668e+00 -2.50184536e-01 3.94241929e-01
-3.44772279e-01 -1.74036705e+00 -3.02290350e-01 9.13084805e-01
-6.72972202e-02 -8.87359917e-01 -8.43125284e-02 -9.33401525e-01
-1.14754617e-01 -1.02608526e+00 9.45225775e-01 -4.84206587e-01
-7.12953061e-02 -1.15915912e-03 -2.88728029e-01 -4.06246752e-01
-5.09051204e-01 6.07451260e-01 -3.38745974e-02 -8.10579583e-02
6.21575713e-01 -6.11629672e-02 3.63649964e-01 -5.35122752e-01
-5.12667000e-01 -1.94582250e-02 -4.07536440e-02 8.21967125e-01
3.90409619e-01 -5.35809517e-01 4.51562822e-01 -1.10936904e+00
6.10272706e-01 -7.80401349e-01 -8.23998690e-01 1.72139078e-01
-3.20075512e-01 2.92899191e-01 6.72903299e-01 -3.19560140e-01
-9.92088854e-01 3.12063307e-01 -5.43209434e-01 -2.10314035e-01
-1.54705852e-01 7.45833993e-01 -2.06497476e-01 3.38286340e-01
4.74278599e-01 -1.17073208e-01 -5.97050190e-01 -6.22253597e-01
7.76412129e-01 8.08093667e-01 8.25805008e-01 -1.02095306e+00
2.03535870e-01 -1.81565106e-01 -1.08757891e-01 -6.07360959e-01
-1.22158098e+00 -3.17542523e-01 -8.33828628e-01 -1.36646196e-01
1.52276945e+00 -9.09856260e-01 -7.13832736e-01 6.39142752e-01
-1.56566346e+00 -9.18719292e-01 -1.65285513e-01 4.87954736e-01
-4.18522716e-01 1.37695432e-01 -1.11378622e+00 -7.65883863e-01
-2.11616918e-01 -1.06919169e+00 9.75026011e-01 -2.09385335e-01
-4.24043417e-01 -1.45791996e+00 1.69930264e-01 3.60976532e-02
4.08377498e-01 3.89876634e-01 1.20356727e+00 -1.09484637e+00
2.41514832e-01 5.82087375e-02 1.10098273e-02 5.95395803e-01
-2.76318669e-01 4.33505848e-02 -1.28525519e+00 -5.43025546e-02
-4.30174112e-01 -7.93035448e-01 1.03861046e+00 3.72523218e-01
1.18794215e+00 -4.17664260e-01 -2.41274908e-01 8.27737093e-01
1.18204820e+00 -1.24368295e-01 1.62245259e-01 3.47026557e-01
6.66771054e-01 6.77216947e-01 2.60209560e-01 9.69219729e-02
7.83669889e-01 8.92820060e-01 4.11888391e-01 9.03642550e-02
-2.22807154e-02 -1.59581870e-01 8.73232782e-01 8.88310611e-01
2.78146774e-01 -2.99085468e-01 -1.19985425e+00 6.39103055e-01
-1.97603476e+00 -4.45181787e-01 -5.69086552e-01 1.68726027e+00
1.20044577e+00 9.54934806e-02 9.63677093e-02 -1.56783938e-01
4.71250117e-01 3.02418083e-01 -2.13011131e-02 -1.31511831e+00
-4.49020803e-01 7.70817459e-01 5.05797684e-01 1.03588164e+00
-1.50700772e+00 1.76570618e+00 6.52927589e+00 9.19333398e-01
-1.12477386e+00 6.42499387e-01 4.78092194e-01 2.93696336e-02
1.91895336e-01 9.98278521e-03 -1.03831291e+00 3.31022620e-01
1.71930528e+00 3.04683037e-02 3.96301121e-01 8.15040708e-01
-4.32715803e-01 3.27534266e-02 -1.20459127e+00 6.35398209e-01
2.98785721e-03 -9.91300881e-01 -3.13177168e-01 -2.91347980e-01
4.46515620e-01 4.71905559e-01 -7.09477141e-02 1.04481125e+00
8.64549041e-01 -1.28676188e+00 1.06753564e+00 2.93678850e-01
9.29773271e-01 -8.95459354e-01 9.29520667e-01 2.00794101e-01
-1.34920895e+00 2.90005114e-02 -9.31916088e-02 -5.45476340e-02
2.99990445e-01 2.96623707e-01 -5.67123175e-01 5.99307835e-01
4.69034702e-01 5.27394593e-01 -5.88744581e-01 3.09134603e-01
-6.32985830e-01 1.15231013e+00 -1.73012570e-01 5.13699874e-02
7.04930544e-01 3.67643759e-02 6.24519825e-01 2.14180183e+00
-3.34987909e-01 -3.13695490e-01 4.65628088e-01 6.72414899e-01
-3.36611807e-01 4.12743866e-01 -3.06022227e-01 2.20859572e-02
3.56520116e-01 1.37539315e+00 -3.29447210e-01 -5.61518669e-01
-3.31690192e-01 8.93792331e-01 1.15655720e+00 1.27759993e-01
-9.43918943e-01 -3.39420199e-01 4.35503095e-01 -4.26652730e-01
3.55083972e-01 -2.56582797e-01 -1.60235122e-01 -1.06425333e+00
-3.11609179e-01 -5.03078163e-01 7.63340473e-01 -6.25495076e-01
-1.11290026e+00 9.33384478e-01 3.64300519e-01 -2.31743574e-01
-5.12342155e-01 -1.14059377e+00 -8.41907859e-02 9.60857689e-01
-1.41430962e+00 -1.45707560e+00 3.88891578e-01 4.65850025e-01
3.53495657e-01 -7.29706138e-02 1.20181811e+00 3.09259504e-01
-8.17752182e-01 9.04042423e-01 -1.89437404e-01 5.31094432e-01
7.13888049e-01 -1.55737340e+00 5.70284247e-01 7.82030404e-01
7.67014697e-02 3.60703021e-01 5.85124493e-01 -3.91677916e-01
-9.83392298e-01 -1.45376539e+00 1.43181574e+00 -5.66044688e-01
1.22270525e+00 -9.45813060e-01 -1.01792479e+00 1.54031074e+00
4.95503962e-01 2.92646408e-01 3.50572020e-01 2.28619084e-01
-6.09167874e-01 4.56532359e-01 -8.83221686e-01 3.45214903e-01
8.33062708e-01 -8.64358485e-01 -7.65431643e-01 4.14481878e-01
9.59904015e-01 -7.08048761e-01 -9.78285611e-01 2.09716663e-01
4.67935801e-01 -2.48200119e-01 4.39231694e-01 -1.18114412e+00
3.38783085e-01 3.11548412e-01 -4.62126851e-01 -1.51860154e+00
-3.26336384e-01 -5.82772434e-01 -4.11014289e-01 1.25549638e+00
7.00520158e-01 -7.33657837e-01 2.05416933e-01 3.92073750e-01
-4.62990373e-01 -7.45856702e-01 -1.40650237e+00 -8.96910489e-01
7.71408021e-01 -9.91980374e-01 4.47151095e-01 1.11487687e+00
4.25218344e-01 6.69317961e-01 -2.00005278e-01 -1.23380125e-01
3.31408173e-01 -5.98596811e-01 1.84312358e-01 -1.05009949e+00
-3.69358093e-01 -3.98309857e-01 -2.10699946e-01 -6.91519439e-01
1.35865772e+00 -1.61626601e+00 4.34508361e-03 -1.11384678e+00
3.75410281e-02 -2.43589684e-01 -2.81369328e-01 1.09077311e+00
3.83946598e-02 5.66904470e-02 8.27451348e-02 -4.47655618e-01
-7.85903215e-01 4.22920883e-01 4.17999744e-01 -9.07972753e-02
-1.18169992e-03 -3.90755445e-01 -3.55576724e-01 7.48042345e-01
6.84672654e-01 -5.40932238e-01 1.77324533e-01 -5.81856370e-01
5.79114616e-01 -8.77685323e-02 3.43240090e-02 -5.92780352e-01
1.03108220e-01 1.02715328e-01 -6.40433282e-02 -4.24674809e-01
3.39629710e-01 -4.39084202e-01 -2.62924194e-01 4.56547618e-01
-4.50186670e-01 2.13583857e-01 6.29617155e-01 -8.74174386e-03
-1.44565821e-01 -2.46879205e-01 7.54040062e-01 1.81299150e-02
-6.35576308e-01 1.44260272e-01 -7.57344961e-01 3.14334393e-01
8.80056620e-01 4.07372624e-01 -4.28922236e-01 -1.22939982e-01
-1.10213792e+00 3.08817565e-01 -2.19619572e-01 5.51230669e-01
-6.59489036e-02 -1.35294366e+00 -8.23562801e-01 -5.24297450e-03
2.25731477e-01 -1.66778341e-01 -1.31604388e-01 8.97696078e-01
-2.45486185e-01 7.61092007e-01 6.20314889e-02 -3.29354197e-01
-1.05105650e+00 4.66242254e-01 7.22287536e-01 -7.21905470e-01
-3.06169271e-01 1.02595115e+00 -3.25172931e-01 -9.18951154e-01
1.27864212e-01 -7.16737330e-01 -3.78147632e-01 1.23951867e-01
2.56542563e-01 3.17811221e-01 3.55011195e-01 -9.10019100e-01
-5.51846325e-01 -9.24386382e-02 -1.02901287e-01 -3.96170557e-01
1.30901766e+00 2.46113703e-01 -2.47700766e-01 9.14037764e-01
1.41094863e+00 6.47847503e-02 -8.33755195e-01 -5.00302970e-01
3.82035881e-01 6.32068455e-01 1.84433684e-01 -9.34743583e-01
-6.76988721e-01 8.05492282e-01 2.63620734e-01 -1.01979218e-01
3.85148585e-01 1.21069841e-01 7.14231849e-01 3.85669798e-01
2.06989527e-01 -1.29006696e+00 -5.02185702e-01 1.39099944e+00
5.72421134e-01 -9.31624234e-01 -4.66659278e-01 -3.43058169e-01
-6.25859022e-01 1.07827091e+00 6.22007906e-01 -8.05446655e-02
7.96153188e-01 6.47145927e-01 -6.90893307e-02 -1.57874361e-01
-1.12492990e+00 1.31425653e-02 1.47711501e-01 3.31122279e-01
6.24736369e-01 5.42528689e-01 -5.14885247e-01 1.32200086e+00
-4.46654588e-01 -4.10320014e-01 2.01219797e-01 6.45456195e-01
-1.07642524e-01 -1.13139665e+00 7.75391906e-02 1.64298058e-01
-7.98764884e-01 -5.77180564e-01 -4.77425933e-01 9.99808609e-01
2.89786458e-01 7.48974085e-01 9.92109627e-02 -3.05723190e-01
2.62693256e-01 9.15857732e-01 5.03435910e-01 -7.26328731e-01
-9.41622496e-01 -3.12874287e-01 7.45755136e-01 -6.74973607e-01
-1.43795937e-01 -7.94622719e-01 -1.43870294e+00 -1.58595815e-02
-2.79007182e-02 2.41725922e-01 7.39110827e-01 1.38972294e+00
-1.56111315e-01 6.77601993e-01 1.03934683e-01 -3.35950255e-01
-5.43837905e-01 -1.43394136e+00 -3.75449330e-01 1.53479129e-01
7.82325789e-02 -7.25968719e-01 -1.78117931e-01 -1.19130045e-01] | [10.516159057617188, 10.044373512268066] |
2006fa64-dafe-4635-a2d6-29679eb27017 | deep-denerative-models-for-drug-design-and | 2109.06469 | null | https://arxiv.org/abs/2109.06469v1 | https://arxiv.org/pdf/2109.06469v1.pdf | Deep Denerative Models for Drug Design and Response | Designing new chemical compounds with desired pharmaceutical properties is a challenging task and takes years of development and testing. Still, a majority of new drugs fail to prove efficient. Recent success of deep generative modeling holds promises of generation and optimization of new molecules. In this review paper, we provide an overview of the current generative models, and describe necessary biological and chemical terminology, including molecular representations needed to understand the field of drug design and drug response. We present commonly used chemical and biological databases, and tools for generative modeling. Finally, we summarize the current state of generative modeling for drug design and drug response prediction, highlighting the state-of-art approaches and limitations the field is currently facing. | ['Lada Nuzhna', 'Karina Zadorozhny'] | 2021-09-14 | null | null | null | null | ['drug-response-prediction'] | ['medical'] | [ 3.42793316e-01 -2.25282907e-01 -5.12867808e-01 -9.49959829e-02
-7.95125723e-01 -7.90640056e-01 4.31364805e-01 3.13248158e-01
1.33515924e-01 1.47296786e+00 5.76106308e-04 -5.76287389e-01
-1.43718615e-01 -6.32931709e-01 -5.89031816e-01 -1.08926237e+00
7.65009820e-02 7.99767733e-01 -2.60647446e-01 -1.18715942e-01
3.38051349e-01 8.49910736e-01 -9.61835623e-01 3.66910428e-01
1.00140190e+00 4.66209501e-01 2.16485530e-01 5.69558263e-01
6.58453852e-02 5.33638895e-01 -8.25827658e-01 -2.26325899e-01
-3.95690978e-01 -9.46356535e-01 -7.33640671e-01 -2.97447205e-01
1.47117302e-01 5.18932752e-02 -2.48309284e-01 5.05796194e-01
1.20898211e+00 1.50563329e-01 1.12108016e+00 -5.69692373e-01
-1.29180610e+00 1.56650975e-01 2.32229277e-01 1.36144772e-01
5.37871420e-01 3.93769979e-01 7.69131482e-01 -1.14095664e+00
9.15200949e-01 8.65789652e-01 4.26357478e-01 1.13675356e+00
-1.52338290e+00 -6.82431340e-01 -1.01846538e-01 1.18243821e-01
-1.48626220e+00 -4.14403975e-01 4.71073359e-01 -7.65715897e-01
1.65907776e+00 1.84495404e-01 7.61360288e-01 1.42792070e+00
9.26078141e-01 3.80153567e-01 7.42582023e-01 2.01554090e-01
6.53192699e-01 -4.20455495e-03 -2.10934222e-01 5.81362545e-01
3.10809880e-01 2.48148680e-01 -7.10713983e-01 -6.31099403e-01
5.38704216e-01 1.47746742e-01 -1.24727413e-01 -2.93356985e-01
-5.12246251e-01 1.37125456e+00 2.50534385e-01 3.30179363e-01
-6.08473599e-01 1.48797140e-01 -7.92926028e-02 -2.62602389e-01
3.26123595e-01 8.87939990e-01 -5.52568853e-01 6.75002337e-02
-9.98994946e-01 4.93012697e-01 6.92427576e-01 7.94782758e-01
2.91180283e-01 4.45218325e-01 -5.05619198e-02 5.92707932e-01
5.05244136e-01 2.43862569e-01 2.26172253e-01 -2.31602892e-01
-4.37582970e-01 2.46038646e-01 -6.81855083e-02 -5.38736522e-01
-6.58130050e-01 -8.00328255e-01 -6.31259441e-01 -6.65702745e-02
-1.38444915e-01 -1.42204225e-01 -1.19310594e+00 1.29777801e+00
2.50403464e-01 -4.47363965e-02 1.38157591e-01 3.33407372e-01
1.43364286e+00 8.23407769e-01 8.79810333e-01 -5.58593750e-01
9.88344848e-01 -4.49270368e-01 -7.35066533e-01 6.71234652e-02
5.06215215e-01 -9.79781568e-01 2.09857523e-01 3.07422161e-01
-1.13492584e+00 -2.83637375e-01 -1.01523876e+00 -1.18933015e-01
-6.48980141e-01 -1.17407642e-01 1.13990951e+00 9.67273772e-01
-5.67591846e-01 9.42817867e-01 -8.30187917e-01 -2.85157394e-02
1.02987051e+00 8.66675258e-01 -3.17212462e-01 -1.08140461e-01
-1.20047534e+00 1.09793055e+00 2.59132177e-01 -7.96981007e-02
-1.58033800e+00 -1.29382300e+00 -6.14768565e-01 -2.69599825e-01
-8.82705525e-02 -1.48716640e+00 9.68524575e-01 5.64332269e-02
-1.74626315e+00 5.76146245e-01 -3.73440266e-01 -2.33092040e-01
-1.74044203e-02 4.04971316e-02 -2.46546060e-01 -2.45923176e-01
-2.59957910e-01 6.48202062e-01 8.49086419e-02 -9.05932903e-01
-1.84381023e-01 -5.32818675e-01 -3.50947082e-01 1.20147318e-01
4.19103116e-01 2.38545872e-02 1.68232769e-01 -4.76383746e-01
-3.88341755e-01 -9.03595328e-01 -7.54935741e-01 -5.24215460e-01
-5.89808643e-01 -1.91481397e-01 4.66245502e-01 -2.93596804e-01
1.30870426e+00 -1.28469968e+00 4.85103697e-01 4.68004122e-02
4.33707356e-01 3.98260415e-01 -5.87123409e-02 1.09653211e+00
-5.99051714e-01 4.39635038e-01 -5.20237833e-02 5.65827042e-02
-4.01396036e-01 -1.13600589e-01 -2.60119021e-01 6.77041173e-01
2.99528360e-01 1.34413540e+00 -9.26060498e-01 1.38730943e-01
3.39311987e-01 1.01237524e+00 -8.13335598e-01 9.15256962e-02
-7.01754391e-01 1.02836943e+00 -9.31969762e-01 1.07717979e+00
3.26904923e-01 -3.91302526e-01 4.17738646e-01 -6.82167858e-02
-5.38209155e-02 5.90029061e-01 -3.59443992e-01 1.55395722e+00
-7.60137215e-02 1.08622782e-01 -9.70172763e-01 -5.75363517e-01
9.81590986e-01 5.20592153e-01 8.38782191e-01 -4.25989389e-01
1.43606588e-01 3.73955846e-01 2.55677730e-01 -1.10952519e-01
-2.45447338e-01 -5.99313974e-01 3.52658123e-01 1.12787083e-01
2.31982589e-01 -2.75623351e-01 6.08205460e-02 -1.26765534e-01
9.02585268e-01 3.46817672e-01 6.90035284e-01 3.37294936e-02
4.51267809e-01 1.54206634e-01 3.89715195e-01 2.71707118e-01
1.92969605e-01 2.59087503e-01 3.90745819e-01 -7.05671310e-01
-9.39534187e-01 -6.94900155e-01 -4.85949397e-01 7.77724206e-01
-4.37515467e-01 -7.88557172e-01 -6.49847925e-01 -4.25267637e-01
-2.06517816e-01 6.43724740e-01 -9.27063823e-01 -3.86431456e-01
-3.37848276e-01 -1.57110536e+00 4.47640121e-01 1.68952197e-01
-2.56621867e-01 -9.42906737e-01 1.60007343e-01 7.45097935e-01
3.85592788e-01 -4.51618075e-01 -1.30637586e-01 4.98578876e-01
-9.61648226e-01 -1.26460052e+00 -8.60636353e-01 -8.32856238e-01
4.36884016e-01 -1.08883895e-01 1.10111260e+00 -2.03050300e-01
-7.56514966e-01 -3.36827040e-01 5.66466451e-02 -7.68626809e-01
-6.11675978e-01 -3.03197633e-02 9.96043347e-03 -7.34998703e-01
3.70889813e-01 -6.66453898e-01 -9.70959961e-01 1.76661626e-01
-7.10508287e-01 -2.91680098e-01 5.09164631e-01 9.65507567e-01
1.23458862e+00 -2.83548295e-01 7.60135293e-01 -1.23347592e+00
8.27982306e-01 -5.70213974e-01 -4.99708086e-01 1.19174361e-01
-8.38126898e-01 4.59820926e-02 4.50410008e-01 -4.13319796e-01
-6.23722851e-01 3.54099661e-01 -8.60519826e-01 2.49890015e-01
-5.23460191e-03 7.23073423e-01 -3.45918715e-01 -4.61867273e-01
9.66233611e-01 3.12924862e-01 -2.15996400e-01 -6.26053095e-01
3.62952232e-01 2.46688902e-01 -2.79131025e-01 -2.57191747e-01
1.34704590e-01 -5.04942685e-02 4.79293436e-01 -7.88125455e-01
-6.08197927e-01 -2.04742942e-02 -2.75364161e-01 3.24458599e-01
7.99334645e-01 -6.09013677e-01 -1.05954230e+00 6.06327206e-02
-1.14092696e+00 -1.21066228e-01 -1.16145387e-01 3.65139663e-01
-8.39130104e-01 1.07112437e-01 -5.21938145e-01 -3.59021723e-01
-7.17425466e-01 -1.81290495e+00 8.46618414e-01 2.73382634e-01
-5.55013478e-01 -1.19573891e+00 7.49948740e-01 2.34233215e-01
3.03784668e-01 7.24568844e-01 1.31413031e+00 -7.14792728e-01
-7.22776771e-01 -3.04658532e-01 4.85358119e-01 -1.56427845e-01
4.03293341e-01 2.60552973e-01 -9.01387274e-01 -1.08162872e-02
-2.46945456e-01 -2.60750711e-01 7.66340852e-01 1.18877411e+00
1.08849931e+00 -2.01545998e-01 -8.86172473e-01 7.70692945e-01
1.29985547e+00 1.16538239e+00 8.67724061e-01 -1.63767681e-01
6.19336665e-01 2.03642324e-01 3.24696332e-01 4.64322954e-01
-9.89784449e-02 6.20433331e-01 2.53487796e-01 -2.77245075e-01
-1.29087642e-01 -4.84472781e-01 2.51155384e-02 1.53954595e-01
-3.87245893e-01 -6.47100210e-01 -6.38874590e-01 -6.50504977e-02
-1.35666275e+00 -9.79740322e-01 -6.50719851e-02 2.08452010e+00
1.29016829e+00 -3.51324081e-01 3.23110521e-01 -4.71775591e-01
2.05140188e-01 -4.07982409e-01 -1.08283496e+00 -4.34297115e-01
-4.07803118e-01 1.06598473e+00 3.26698244e-01 6.39553666e-01
-9.19720590e-01 1.27440119e+00 8.84401226e+00 1.00770795e+00
-1.08517265e+00 -2.65862167e-01 1.15298831e+00 6.27948120e-02
-5.08435249e-01 4.89893369e-02 -1.19120467e+00 3.07961345e-01
1.33123600e+00 -4.11822677e-01 9.70507562e-02 6.71201050e-01
3.65268737e-01 2.50773251e-01 -1.36583042e+00 9.78973389e-01
-1.17372334e-01 -2.24195933e+00 4.70340133e-01 3.77240688e-01
8.28728557e-01 -1.41739845e-01 4.82896388e-01 -8.01667422e-02
2.62862116e-01 -1.80714059e+00 -1.72146745e-02 4.77921098e-01
8.59229207e-01 -9.19287145e-01 3.79710972e-01 -1.26690805e-01
-6.24472320e-01 3.33772928e-01 -4.29478556e-01 4.60197568e-01
4.33491841e-02 3.85464311e-01 -9.92587209e-01 4.30967689e-01
-6.32647797e-02 1.07712817e+00 -3.31225991e-01 1.27627397e+00
-2.78526932e-01 2.85225838e-01 1.71653688e-01 -3.94437581e-01
2.11371228e-01 -3.01634669e-01 1.69545725e-01 1.15132380e+00
2.30216175e-01 2.45936096e-01 2.14518636e-01 1.13547087e+00
-3.82744968e-02 2.91418076e-01 -5.34226120e-01 -7.64092326e-01
2.73138791e-01 6.93424821e-01 -3.33971053e-01 -2.20380560e-01
-3.36996140e-03 7.63637066e-01 -2.21866384e-01 4.99453574e-01
-1.00968504e+00 -9.48437080e-02 1.06895685e+00 4.06116135e-02
-1.47515342e-01 -1.43548399e-01 -1.30121529e-01 -5.72621822e-01
-8.38692963e-01 -1.12797666e+00 3.13655227e-01 -5.54639161e-01
-1.04414642e+00 5.69930375e-01 -2.81202495e-01 -8.59632134e-01
-1.15205288e-01 -8.49675179e-01 -1.22673161e-01 1.11573780e+00
-9.55844641e-01 -9.68205333e-01 4.06047463e-01 1.51642025e-01
6.52885020e-01 -2.57575035e-01 1.45351267e+00 4.81108963e-01
-6.40966594e-01 3.78188878e-01 4.50171322e-01 -8.14884901e-01
6.28822029e-01 -9.31959391e-01 5.39215505e-01 -1.76714763e-01
-7.06377402e-02 1.24384868e+00 1.09511054e+00 -1.00386429e+00
-1.51500714e+00 -8.73746634e-01 7.64791787e-01 -6.64310098e-01
3.86180848e-01 -2.81715184e-01 -5.97701132e-01 3.24809968e-01
1.25797400e-02 -5.50032437e-01 1.80413008e+00 -1.83409844e-02
2.09452603e-02 4.46416378e-01 -1.12274349e+00 6.32667363e-01
9.97142732e-01 -2.35458538e-01 1.96174532e-01 1.05462241e+00
5.54822922e-01 -5.71111739e-01 -1.09999299e+00 4.06773716e-01
6.81689739e-01 -5.53043365e-01 1.35178196e+00 -1.37833762e+00
2.97840774e-01 -4.27515000e-01 1.28678530e-01 -1.35846519e+00
-8.01394284e-01 -9.20196116e-01 -1.54069707e-01 3.91928732e-01
1.01054645e+00 -4.29810315e-01 1.08340931e+00 5.48419237e-01
-3.59852761e-01 -1.35983741e+00 -6.88046694e-01 -4.24558938e-01
4.92371559e-01 -3.20293829e-02 4.37740415e-01 6.59104168e-01
3.68441492e-02 9.71282601e-01 -4.70396042e-01 -4.31874961e-01
1.43546462e-01 1.26072899e-01 4.26038235e-01 -1.01452708e+00
-3.77192318e-01 -6.41072154e-01 -5.47387183e-01 -8.40619802e-01
-1.89802021e-01 -9.03807640e-01 -6.94653690e-01 -1.92011881e+00
4.55820113e-01 1.09816119e-02 3.95365022e-02 2.75171369e-01
-1.12809144e-01 1.72862008e-01 -4.88074183e-01 -3.45285460e-02
-7.24294335e-02 7.38450229e-01 1.54662430e+00 -4.70106244e-01
-5.37307203e-01 2.01860994e-01 -1.05487621e+00 1.99374795e-01
7.89223313e-01 -5.12623310e-01 -5.44052362e-01 3.18068683e-01
4.54746455e-01 -1.59061745e-01 -1.85548112e-01 -3.69368166e-01
-1.15081146e-01 -6.08327687e-01 9.23392177e-01 -7.30869472e-01
5.09890795e-01 -1.95269898e-01 9.43151891e-01 8.04410338e-01
-8.23845938e-02 -1.94208279e-01 3.97272855e-01 5.38880646e-01
-6.71084411e-03 -2.02072244e-02 9.79431808e-01 -1.74288318e-01
-1.57882422e-01 8.79366577e-01 -7.68636763e-01 -4.77327615e-01
9.99495685e-01 -4.48496580e-01 -1.54328138e-01 -1.19177476e-01
-1.16924322e+00 -1.75493896e-01 1.86247945e-01 2.08459213e-01
9.90037918e-01 -1.10065496e+00 -5.96903324e-01 -1.92387149e-01
1.90493509e-01 -4.86923426e-01 4.57521111e-01 5.38217664e-01
-7.65818000e-01 1.00644207e+00 2.94200480e-02 -1.67456031e-01
-1.42576504e+00 8.58115494e-01 7.74443030e-01 -2.34501302e-01
7.46390373e-02 9.10094380e-01 7.42717683e-01 1.28181681e-01
-7.31197149e-02 1.14694662e-01 -3.58177751e-01 -1.36144236e-01
5.42884767e-01 5.57430238e-02 3.44686866e-01 -3.32599521e-01
-6.26653969e-01 5.52623391e-01 -2.07408726e-01 5.69281042e-01
1.63828111e+00 5.23868561e-01 8.32533166e-02 -1.12020664e-01
1.09069014e+00 -3.62300098e-01 -5.70954025e-01 3.05623919e-01
-3.26290607e-01 -1.03027858e-01 -2.06607822e-02 -1.33777249e+00
-5.69203973e-01 8.50196779e-01 6.47952437e-01 -5.19214451e-01
6.72710121e-01 1.85565487e-01 6.20750010e-01 2.28630096e-01
2.15351418e-01 -6.73676193e-01 -3.95891868e-04 4.51750934e-01
1.00963628e+00 -8.75729144e-01 4.85368609e-01 -5.57480514e-01
-3.41503114e-01 9.36836720e-01 8.26255456e-02 9.77378488e-02
6.53890371e-01 -1.26027204e-02 -4.01220083e-01 -7.30744064e-01
-8.28762710e-01 -1.20975971e-01 5.19096196e-01 9.32084680e-01
1.27355182e+00 -4.72660316e-03 -5.93859494e-01 4.36121076e-01
1.25671327e-01 3.23305950e-02 -2.51403339e-02 1.00532985e+00
-5.02357781e-01 -1.96218097e+00 1.19746318e-02 9.80307832e-02
-6.81876600e-01 -6.19543552e-01 -1.19378388e+00 4.83120590e-01
1.64935648e-01 8.55906010e-01 -6.13281131e-01 -1.95165098e-01
3.91533077e-01 7.31214881e-02 1.03232121e+00 -9.39011872e-01
-5.72577178e-01 6.32071197e-01 -9.42169130e-03 -2.10828662e-01
-2.59668827e-01 -3.70742589e-01 -1.26145744e+00 -4.50909674e-01
-4.34799314e-01 2.91272998e-01 7.55651772e-01 5.44583559e-01
8.75723183e-01 7.32681990e-01 4.53844398e-01 -4.87294674e-01
3.78506742e-02 -6.90288901e-01 -4.98365641e-01 -4.53105301e-01
6.65231720e-02 -7.62118876e-01 2.13944688e-01 3.59552175e-01] | [4.966722011566162, 5.826615810394287] |
d4ec7839-d6b3-473c-86b4-cd4cc2f54707 | feature-learning-for-chord-recognition-the | 1612.05065 | null | http://arxiv.org/abs/1612.05065v1 | http://arxiv.org/pdf/1612.05065v1.pdf | Feature Learning for Chord Recognition: The Deep Chroma Extractor | We explore frame-level audio feature learning for chord recognition using
artificial neural networks. We present the argument that chroma vectors
potentially hold enough information to model harmonic content of audio for
chord recognition, but that standard chroma extractors compute too noisy
features. This leads us to propose a learned chroma feature extractor based on
artificial neural networks. It is trained to compute chroma features that
encode harmonic information important for chord recognition, while being robust
to irrelevant interferences. We achieve this by feeding the network an audio
spectrum with context instead of a single frame as input. This way, the network
can learn to selectively compensate noise and resolve harmonic ambiguities.
We compare the resulting features to hand-crafted ones by using a simple
linear frame-wise classifier for chord recognition on various data sets. The
results show that the learned feature extractor produces superior chroma
vectors for chord recognition. | ['Filip Korzeniowski', 'Gerhard Widmer'] | 2016-12-15 | null | null | null | null | ['chord-recognition'] | ['audio'] | [ 4.47472095e-01 3.20176296e-02 3.11941542e-02 -2.29607657e-01
-1.03266811e+00 -6.73427284e-01 3.66502315e-01 -2.84253955e-02
-4.39306200e-01 5.00122905e-01 3.97086829e-01 2.14438647e-01
-3.90503317e-01 -7.69288838e-01 -3.89433712e-01 -8.17889690e-01
-2.63124049e-01 -1.01629846e-01 1.39843374e-01 -5.14882922e-01
3.76982540e-01 5.87696016e-01 -2.10191607e+00 7.53398776e-01
7.40357637e-02 1.37897027e+00 -1.90367782e-03 1.33109546e+00
9.00379196e-02 9.94253993e-01 -1.19798613e+00 1.09792762e-01
3.87622952e-01 -8.80580306e-01 -8.76908302e-01 -3.21297556e-01
4.71641034e-01 1.11489534e-01 -2.69850075e-01 8.07260275e-01
7.63795495e-01 2.80239969e-01 6.47427261e-01 -1.04503357e+00
-2.90441066e-01 9.35336232e-01 4.24480766e-01 1.99209198e-01
6.48053646e-01 -5.66584393e-02 1.50013447e+00 -8.94409359e-01
6.12056971e-01 8.39129388e-01 1.05973613e+00 4.27306354e-01
-1.18083298e+00 -4.05524313e-01 -4.35856640e-01 6.30050123e-01
-1.34405637e+00 -6.69045269e-01 1.33537245e+00 -1.91952407e-01
1.12803197e+00 8.99941921e-01 1.15491116e+00 1.03133929e+00
1.01072736e-01 7.15541899e-01 7.26565480e-01 -8.01661074e-01
2.70333320e-01 -3.54861140e-01 -1.86708808e-01 2.24929973e-01
-6.58969283e-01 3.74219924e-01 -1.10318315e+00 9.39244404e-03
6.19028628e-01 -4.88344163e-01 -5.24615765e-01 -1.68050230e-02
-1.05423343e+00 6.98226571e-01 2.70652384e-01 5.46653390e-01
-2.30153903e-01 2.28407025e-01 5.88234067e-01 8.97631228e-01
7.41698295e-02 1.06660628e+00 -6.42555773e-01 -4.32276636e-01
-1.01095545e+00 4.22524214e-01 8.05311263e-01 1.92714930e-01
5.41956782e-01 7.92078853e-01 -1.99609384e-01 9.42514181e-01
-1.96854830e-01 3.36188316e-01 8.12497377e-01 -1.41380513e+00
-2.08096206e-01 -8.84717926e-02 -3.27084661e-01 -1.13773298e+00
-6.95937216e-01 -4.23419118e-01 -6.09855413e-01 5.94561279e-01
6.84669554e-01 -9.37524736e-02 -5.07899225e-01 1.75404644e+00
-4.81962934e-02 1.90724418e-01 1.28514856e-01 9.87799764e-01
6.30622625e-01 6.34538174e-01 -6.85392141e-01 -4.38718408e-01
1.09126151e+00 -5.86838782e-01 -1.03254390e+00 3.59730184e-01
2.34219849e-01 -1.07159853e+00 1.24259067e+00 1.04901886e+00
-1.22029757e+00 -1.01567721e+00 -1.37300873e+00 -9.37149823e-02
-3.49646747e-01 -1.00373328e-01 5.26225090e-01 6.76486611e-01
-9.64751005e-01 1.17083991e+00 -3.71580064e-01 3.14512491e-01
-2.35883325e-01 3.16156626e-01 -1.76185846e-01 7.89324045e-01
-1.42556477e+00 4.39909488e-01 6.20243192e-01 -6.03206307e-02
-7.96330988e-01 -6.54324114e-01 -7.48578966e-01 2.26203069e-01
1.83950454e-01 -1.83615893e-01 1.61101222e+00 -1.29238355e+00
-2.00330973e+00 4.11262602e-01 1.75097525e-01 -6.01855993e-01
1.28225535e-01 -2.26780340e-01 -6.19130671e-01 4.65372771e-01
-3.71274263e-01 4.14588958e-01 1.24923551e+00 -9.37322199e-01
-6.88088119e-01 7.32841417e-02 -1.12008527e-01 -8.10338333e-02
-3.03222090e-01 -6.51592836e-02 1.29903629e-01 -1.14080799e+00
2.93963164e-01 -7.66866267e-01 2.83543646e-01 -2.38616392e-01
-3.17386389e-01 -1.44917607e-01 8.47164810e-01 -5.11427701e-01
1.28108513e+00 -2.26395798e+00 9.37282220e-02 5.23190796e-01
-7.54329786e-02 1.52971745e-01 -3.10463309e-01 4.03357565e-01
-2.29924917e-01 -2.05286533e-01 5.08272089e-02 5.84032424e-02
-3.38957049e-02 7.63334632e-02 -5.11455238e-01 1.79650158e-01
2.98749030e-01 6.51180267e-01 -6.77450776e-01 -2.71228045e-01
4.83237691e-02 7.36504674e-01 -8.51753950e-01 2.77177393e-01
-1.64359584e-01 2.27896407e-01 -9.80648957e-03 5.55130839e-01
1.94308102e-01 3.80723238e-01 1.00953706e-01 -4.23370779e-01
-8.48982260e-02 6.10288799e-01 -1.57216704e+00 1.75014317e+00
-3.26452941e-01 1.11500967e+00 -1.35401919e-01 -9.46959019e-01
1.14704251e+00 7.42824554e-01 5.48298657e-01 -7.11170912e-01
2.60940373e-01 2.53444523e-01 1.96514711e-01 -3.60524267e-01
6.47342861e-01 -2.94060022e-01 -2.43057340e-01 2.10758999e-01
5.53780556e-01 -4.06378508e-01 -3.57469954e-02 -3.36802244e-01
1.17827260e+00 7.39412382e-02 7.85027072e-02 6.77217767e-02
5.10063708e-01 -3.36513102e-01 6.40233755e-01 9.82812107e-01
-1.28577650e-01 1.01820564e+00 3.60217035e-01 -7.05091596e-01
-1.00281060e+00 -1.02913570e+00 1.86453871e-02 1.43767786e+00
-3.69508445e-01 -8.13868284e-01 -7.48372316e-01 -9.49410349e-02
-3.57611448e-01 2.30479017e-01 -5.58488667e-01 -4.29679155e-01
-7.68263340e-01 -2.84348279e-02 1.09883666e+00 4.57042336e-01
-5.54748140e-02 -1.47357070e+00 -8.03660274e-01 4.30451125e-01
-1.80650100e-01 -6.62941933e-01 -3.52891892e-01 8.74261022e-01
-4.49732274e-01 -1.08069098e+00 -4.21083331e-01 -8.33030760e-01
-3.09792489e-01 -1.01816766e-01 9.95649099e-01 9.55627412e-02
-5.05498648e-01 2.49358624e-01 -5.89207470e-01 -3.60884100e-01
-4.76680309e-01 6.21541776e-02 3.89441371e-01 -7.27124289e-02
6.85285777e-02 -1.08204389e+00 -3.47135723e-01 -4.42687459e-02
-8.70076597e-01 -1.60023138e-01 1.98002025e-01 1.06910288e+00
6.13790035e-01 2.11564645e-01 8.31557214e-01 -3.95025253e-01
8.95904720e-01 2.49491021e-01 -4.22480077e-01 -1.98421046e-01
-3.44750017e-01 1.00000583e-01 1.06520939e+00 -6.74131036e-01
-5.89994967e-01 2.71326602e-01 -5.78681469e-01 -3.39479506e-01
-9.89148468e-02 3.98070186e-01 -1.46044388e-01 -1.96879536e-01
1.04995358e+00 1.57936350e-01 -1.84040472e-01 -5.35995901e-01
3.05618018e-01 5.98046541e-01 1.13866591e+00 -5.42557418e-01
6.80445313e-01 5.22375964e-02 -1.37438271e-02 -9.50096428e-01
-7.30442345e-01 -2.58412868e-01 -6.72948897e-01 -4.62157696e-01
6.65242493e-01 -5.68505466e-01 -9.49931979e-01 2.98889130e-01
-1.02354467e+00 -1.67673022e-01 -9.17798579e-01 6.09933913e-01
-1.13715911e+00 -2.89090145e-02 -9.37926412e-01 -9.45520997e-01
-2.70294011e-01 -9.14136648e-01 8.10798287e-01 1.14227198e-01
-3.87060165e-01 -4.78668541e-01 4.98787612e-01 -2.42418125e-01
2.70257056e-01 3.54402244e-01 8.26712132e-01 -7.98545778e-01
-3.13288033e-01 -2.47411862e-01 5.13573945e-01 4.48672920e-01
7.24974051e-02 1.50652736e-01 -1.75804365e+00 -2.84059830e-02
1.35383204e-01 -5.11025548e-01 8.69340599e-01 2.76992917e-01
1.32454622e+00 -6.07673287e-01 4.48481560e-01 8.13619792e-01
1.10426903e+00 4.16461706e-01 6.30551159e-01 3.25858772e-01
4.07428890e-01 2.54592419e-01 3.65651220e-01 6.51689947e-01
-3.11602831e-01 8.78597140e-01 2.59857953e-01 2.12208375e-01
-3.14659685e-01 -2.81800836e-01 3.38888317e-01 1.31844151e+00
-3.03836763e-01 1.48906499e-01 -6.03621602e-01 4.12995607e-01
-1.46363485e+00 -1.34354007e+00 2.60084778e-01 1.95565689e+00
1.17965508e+00 4.06297833e-01 4.74297285e-01 1.40619206e+00
2.44189560e-01 1.94132850e-01 -1.21081196e-01 -6.03356361e-01
-4.31979805e-01 8.15176487e-01 -2.84294993e-01 5.51304579e-01
-1.02869594e+00 7.45202005e-01 6.96162033e+00 1.04302514e+00
-1.32001913e+00 -2.10983187e-01 3.16117182e-02 -1.41958028e-01
-2.17133760e-01 -1.04140500e-02 -1.53448865e-01 -1.01492174e-01
1.19902861e+00 -1.28387004e-01 8.10457289e-01 8.62272978e-01
4.91497628e-02 4.57633495e-01 -1.16419518e+00 1.04430866e+00
-2.70077735e-02 -1.66517186e+00 9.74308401e-02 -3.35980356e-01
4.16654497e-01 -3.59561682e-01 7.76324123e-02 2.89321899e-01
-3.16962510e-01 -1.11792779e+00 9.20185089e-01 7.23706961e-01
5.76812923e-01 -1.14415276e+00 5.53090215e-01 9.04612318e-02
-1.33948314e+00 -2.04459414e-01 -3.22868764e-01 -5.43407500e-01
-1.26566321e-01 2.34484926e-01 -9.98813748e-01 2.21685201e-01
5.48875332e-01 4.04223442e-01 -5.08497417e-01 1.11543715e+00
-2.95715276e-02 6.74011827e-01 -2.97072351e-01 6.40709326e-02
-2.38683075e-02 3.17692846e-01 5.30337572e-01 1.38087916e+00
3.68559092e-01 -1.28061354e-01 1.18264526e-01 3.89098495e-01
-3.33970524e-02 2.77234584e-01 -5.98982990e-01 1.92989931e-02
5.12740254e-01 1.08419275e+00 -6.94578290e-01 -1.87315956e-01
2.49723252e-02 7.95487881e-01 4.20442522e-02 2.33903050e-01
-4.37037289e-01 -1.06369030e+00 7.05099106e-01 -1.42661288e-01
3.65772098e-01 5.90554718e-03 -1.11629978e-01 -8.66860926e-01
-2.08861545e-01 -1.22949827e+00 3.48252267e-01 -8.46161723e-01
-9.55781579e-01 8.43851447e-01 -4.03769761e-01 -1.38285089e+00
-9.22480404e-01 -6.94169581e-01 -7.23716378e-01 4.78732884e-01
-1.12022066e+00 -7.13257074e-01 1.76446103e-02 7.95361638e-01
5.88045359e-01 -5.00945628e-01 1.40363073e+00 1.35806844e-01
3.18315923e-01 7.58588254e-01 -4.33036536e-02 4.03627247e-01
6.43750131e-01 -1.54873896e+00 1.99242681e-01 4.29163337e-01
1.04254222e+00 2.87753940e-01 9.44529891e-01 -1.99348684e-02
-1.34382641e+00 -4.80464309e-01 9.40372646e-01 -2.45359257e-01
4.77593184e-01 -4.47781026e-01 -9.90292430e-01 2.00573578e-01
1.19236290e-01 9.52951983e-02 9.15067196e-01 1.97965264e-01
-6.64292753e-01 -3.90854686e-01 -6.99903488e-01 3.41339499e-01
7.10497022e-01 -1.35109949e+00 -9.20415759e-01 9.37228501e-02
9.26164150e-01 -3.20712090e-01 -9.87349868e-01 3.74030739e-01
1.02598870e+00 -1.22039962e+00 1.20059860e+00 -7.04872370e-01
2.74956673e-01 -4.43797112e-01 -3.73585314e-01 -1.40067232e+00
-3.14307570e-01 -8.99273098e-01 -1.86318174e-01 8.55554998e-01
3.80753249e-01 1.27519920e-01 6.84456766e-01 -1.56453907e-01
-3.55217069e-01 -2.50001788e-01 -1.20888889e+00 -7.58319080e-01
-4.35533933e-02 -9.03436482e-01 4.09390569e-01 1.00802124e+00
4.18849200e-01 3.58965605e-01 -6.68746352e-01 -1.37930602e-01
3.03408265e-01 1.86312482e-01 5.99979043e-01 -1.41788447e+00
-7.74381161e-01 -5.17605603e-01 -8.28717530e-01 -4.05833483e-01
1.60219327e-01 -7.93481052e-01 2.46564388e-01 -5.74377298e-01
-5.38935065e-01 -1.00448169e-02 -6.89133048e-01 4.88182187e-01
2.18830988e-01 8.12292337e-01 6.05872273e-01 -8.28206390e-02
-3.66161495e-01 3.20416152e-01 9.67422724e-01 -3.00337434e-01
-4.69059438e-01 1.87246710e-01 -2.65347689e-01 1.01999426e+00
9.16880846e-01 -4.54529226e-01 -3.30987185e-01 1.76637635e-01
4.40520674e-01 3.40616524e-01 2.10937157e-01 -1.55098855e+00
2.61783421e-01 -1.70164404e-03 8.45067859e-01 -3.93289238e-01
7.39732087e-01 -6.85241520e-01 1.74859926e-01 3.46760154e-01
-8.37360740e-01 -2.81541832e-02 1.49882913e-01 2.59594560e-01
-6.90290451e-01 -3.99269491e-01 5.03789067e-01 -9.95948166e-02
-5.12732327e-01 -2.98981249e-01 -6.28797472e-01 6.54539764e-02
1.13936163e-01 -2.20870361e-01 9.93490145e-02 -8.14331651e-01
-1.20539689e+00 -7.75884688e-01 -3.43963355e-02 3.27743500e-01
7.03266621e-01 -1.60794783e+00 -5.71449816e-01 4.95384037e-01
-7.17356652e-02 -5.86206198e-01 7.16822743e-02 1.63864031e-01
-2.87318021e-01 1.49621248e-01 -5.24044394e-01 -5.41003704e-01
-1.41468287e+00 4.61389840e-01 4.13360864e-01 1.70453683e-01
-7.80045331e-01 7.62625933e-01 -2.86412328e-01 -8.28296319e-02
5.92987597e-01 -5.98646224e-01 -2.65801936e-01 3.79078448e-01
9.78702664e-01 1.32672004e-02 2.23408207e-01 -5.06457806e-01
-1.29225895e-01 6.03837132e-01 3.54043871e-01 -4.47189003e-01
1.36545861e+00 1.25066295e-01 2.41364285e-01 9.73756254e-01
1.33482194e+00 4.35320377e-01 -9.97441232e-01 -1.09318927e-01
1.78941593e-01 -3.79266590e-01 3.08993943e-02 -7.22022355e-01
-9.69039321e-01 8.16144526e-01 8.23723137e-01 6.84951007e-01
1.51762486e+00 -4.16586220e-01 5.75641334e-01 9.28157270e-01
1.89646542e-01 -1.40113533e+00 4.09906060e-01 7.72527099e-01
9.78456736e-01 -6.59276545e-01 -2.86516726e-01 1.58902571e-01
-2.47530803e-01 1.64099753e+00 8.85216296e-02 -4.34909552e-01
5.73967040e-01 5.25006831e-01 5.11004686e-01 1.05363250e-01
-7.66599178e-01 -3.20480615e-01 4.92218316e-01 7.76429772e-01
4.26313251e-01 -1.05021991e-01 7.01401336e-03 8.81824017e-01
-9.64145780e-01 -1.66254207e-01 6.86637402e-01 7.39830673e-01
-6.94470286e-01 -1.15261316e+00 -5.91518760e-01 6.11533374e-02
-7.19275594e-01 -8.29219818e-02 -5.94991744e-01 6.71312034e-01
2.90338337e-01 7.28497446e-01 3.50938812e-02 -8.56041789e-01
3.61167192e-01 6.19614363e-01 5.47332168e-01 -1.32726878e-01
-1.10171974e+00 5.52384257e-01 1.30112931e-01 -5.67855656e-01
-4.48129416e-01 -2.35058114e-01 -1.16957319e+00 1.41218096e-01
-3.04716408e-01 4.64184374e-01 5.42494535e-01 7.32878983e-01
-1.10246591e-01 9.15546000e-01 8.29418063e-01 -1.21377790e+00
-3.36624920e-01 -9.14421201e-01 -7.34436691e-01 4.33716089e-01
9.04822350e-01 -3.10858727e-01 -4.19625908e-01 4.78750050e-01] | [15.810012817382812, 5.299386024475098] |
a63a5038-e2c3-4b4a-b44b-3da6a684a215 | weakly-supervised-cross-lingual-named-entity | 1707.02483 | null | http://arxiv.org/abs/1707.02483v1 | http://arxiv.org/pdf/1707.02483v1.pdf | Weakly Supervised Cross-Lingual Named Entity Recognition via Effective Annotation and Representation Projection | The state-of-the-art named entity recognition (NER) systems are supervised
machine learning models that require large amounts of manually annotated data
to achieve high accuracy. However, annotating NER data by human is expensive
and time-consuming, and can be quite difficult for a new language. In this
paper, we present two weakly supervised approaches for cross-lingual NER with
no human annotation in a target language. The first approach is to create
automatically labeled NER data for a target language via annotation projection
on comparable corpora, where we develop a heuristic scheme that effectively
selects good-quality projection-labeled data from noisy data. The second
approach is to project distributed representations of words (word embeddings)
from a target language to a source language, so that the source-language NER
system can be applied to the target language without re-training. We also
design two co-decoding schemes that effectively combine the outputs of the two
projection-based approaches. We evaluate the performance of the proposed
approaches on both in-house and open NER data for several target languages. The
results show that the combined systems outperform three other weakly supervised
approaches on the CoNLL data. | ['Jian Ni', 'Georgiana Dinu', 'Radu Florian'] | 2017-07-08 | weakly-supervised-cross-lingual-named-entity-1 | https://aclanthology.org/P17-1135 | https://aclanthology.org/P17-1135.pdf | acl-2017-7 | ['cross-lingual-ner'] | ['natural-language-processing'] | [-1.17656216e-01 1.40371576e-01 6.89059421e-02 -6.50537014e-01
-1.28939486e+00 -7.54418433e-01 4.64571923e-01 1.32569253e-01
-1.01874149e+00 7.79500067e-01 5.59844434e-01 -1.95241198e-01
4.36450690e-01 -7.30762959e-01 -3.78647685e-01 -5.74209690e-01
4.17478502e-01 7.00367212e-01 2.13833824e-01 -2.19481781e-01
-6.25559911e-02 3.12184185e-01 -1.03089643e+00 1.64387062e-01
8.70938540e-01 3.22450131e-01 4.84065562e-01 3.82379860e-01
-7.68132448e-01 3.48097891e-01 -5.11714697e-01 -8.15348327e-01
3.06693733e-01 -4.07885998e-01 -1.01477313e+00 -2.62535274e-01
1.00596569e-01 3.09206754e-01 9.51495618e-02 1.31746328e+00
7.86469817e-01 1.79433614e-01 5.35876691e-01 -6.62472606e-01
-8.45987380e-01 9.79583263e-01 -1.31441951e-01 -6.50968775e-02
1.84003100e-01 -3.90086085e-01 9.59437788e-01 -1.17306888e+00
6.61368787e-01 1.06453395e+00 8.48357737e-01 7.47242451e-01
-1.00590396e+00 -6.23968601e-01 -2.07733959e-01 -2.57897586e-01
-1.60079014e+00 -5.82528532e-01 6.33275032e-01 -2.12683767e-01
9.47433949e-01 -9.29939970e-02 -9.57813859e-02 8.35641682e-01
-2.74479479e-01 4.03135449e-01 1.12224293e+00 -8.06817055e-01
3.20391834e-01 4.51947600e-01 3.02385241e-01 6.20768309e-01
2.56088972e-01 -1.86927348e-01 -3.55527550e-01 -4.52577174e-01
2.09076390e-01 -2.31003508e-01 -2.04904228e-01 -7.36585706e-02
-1.14105427e+00 8.27523649e-01 9.31757838e-02 6.92457438e-01
-4.96972412e-01 -3.80072296e-01 5.86621463e-01 1.09696738e-01
5.65979123e-01 3.78242135e-01 -9.92020905e-01 -3.81764993e-02
-9.14029062e-01 -4.40189093e-01 1.13893449e+00 1.02364755e+00
8.19398820e-01 -1.01944715e-01 1.44996971e-01 1.14687824e+00
4.25397307e-01 5.75571060e-01 8.54879320e-01 -2.15398699e-01
7.60609806e-01 5.88654220e-01 1.85059950e-01 -5.05384386e-01
-2.73249000e-01 -5.38476296e-02 -6.26377940e-01 -2.10730001e-01
2.32274473e-01 -6.87858820e-01 -8.04516971e-01 1.77989268e+00
4.18282092e-01 -9.54383239e-02 9.15468931e-01 5.63651145e-01
1.00384843e+00 9.74121690e-01 5.85211933e-01 -2.49143422e-01
1.31561172e+00 -1.18889701e+00 -7.95620918e-01 -2.18575001e-01
1.10586035e+00 -9.94084358e-01 1.05111122e+00 -7.06045702e-02
-5.98762929e-01 -4.59306717e-01 -7.24902272e-01 -1.45183980e-01
-7.49692500e-01 7.69301176e-01 2.72797376e-01 7.35627711e-01
-8.14314008e-01 4.84631151e-01 -7.24194825e-01 -8.81420970e-01
-1.51125751e-02 4.06652093e-01 -7.63709128e-01 -1.74821988e-01
-1.23364043e+00 1.15124381e+00 8.80300105e-01 1.22025050e-01
-5.98371983e-01 -2.56892145e-01 -1.00536489e+00 6.19187392e-02
6.24444298e-02 3.94501630e-03 1.15726435e+00 -8.11079681e-01
-1.49576581e+00 9.29865956e-01 -3.45810413e-01 -1.74932882e-01
-1.23820513e-01 -4.03724104e-01 -7.70715117e-01 -2.43618816e-01
3.04198742e-01 4.56976563e-01 1.11833826e-01 -1.14211822e+00
-6.90103531e-01 -2.25411385e-01 -3.80837142e-01 4.65193301e-01
-7.72179127e-01 6.31176710e-01 -3.50031376e-01 -4.20478076e-01
-4.76966389e-02 -6.88479960e-01 -4.97934848e-01 -5.84914327e-01
-3.36787909e-01 -5.44615865e-01 5.69047093e-01 -8.15784156e-01
1.18853462e+00 -2.12714434e+00 -3.96754593e-02 -2.10688263e-02
-4.24314827e-01 5.46760142e-01 -3.30386788e-01 5.90198755e-01
2.98545063e-02 2.75548011e-01 -3.20037127e-01 -3.65672529e-01
-9.65095758e-02 3.94979119e-01 -1.51889324e-01 1.98480055e-01
1.34477958e-01 5.61565161e-01 -9.51108515e-01 -7.03653693e-01
-8.73485878e-02 5.20476639e-01 -7.32942075e-02 5.55896461e-01
2.23430455e-01 2.56863117e-01 -4.04602170e-01 4.56537604e-01
4.46517944e-01 2.31015727e-01 4.93538558e-01 -3.45035106e-01
-3.45433235e-01 6.11641467e-01 -1.36019218e+00 1.74151564e+00
-7.22858250e-01 1.73718750e-01 -1.53094336e-01 -8.30526590e-01
1.34478676e+00 8.81843150e-01 3.41369696e-02 -7.03296661e-02
1.44494638e-01 5.63956261e-01 -2.48924509e-01 -4.60213572e-01
4.57575589e-01 -4.96743262e-01 -5.78052342e-01 5.51455081e-01
6.08961701e-01 1.60674468e-01 2.54567057e-01 -2.69066066e-01
9.03107226e-01 1.68041170e-01 6.69522345e-01 -2.66187400e-01
7.75724292e-01 3.81837517e-01 1.05053234e+00 2.46385157e-01
-1.61795616e-01 2.33404115e-01 -4.16233987e-02 -4.21042979e-01
-1.10566032e+00 -7.65833735e-01 -8.36784244e-02 1.40433443e+00
-1.28214940e-01 -2.94250876e-01 -9.97470319e-01 -1.12514055e+00
-6.97604239e-01 8.58094752e-01 -1.67158931e-01 3.01521212e-01
-6.74897432e-01 -7.99974561e-01 1.19377553e+00 5.13226569e-01
5.04099190e-01 -1.29692423e+00 9.44530293e-02 3.82702380e-01
-3.61434370e-01 -1.11171925e+00 -5.51572680e-01 5.74456811e-01
-7.58081615e-01 -7.21847892e-01 -6.84145451e-01 -1.57285500e+00
1.01717317e+00 -1.31959602e-01 8.47656429e-01 -1.66900009e-01
3.41758698e-01 -6.22346476e-02 -5.09710789e-01 -1.04357421e-01
-6.03182971e-01 4.45175856e-01 4.02471036e-01 -7.67600685e-02
1.08356178e+00 -1.86691120e-01 2.26904541e-01 4.71619636e-01
-8.17777753e-01 -2.07571000e-01 6.87401175e-01 9.34326708e-01
8.24614823e-01 8.37552622e-02 5.13412356e-01 -1.14728761e+00
6.67637527e-01 -5.36444962e-01 -4.76167917e-01 5.47715306e-01
-3.11766654e-01 4.34920013e-01 1.02251351e+00 -5.97263277e-01
-1.54745555e+00 7.14336157e-01 -5.26110351e-01 1.85817808e-01
-5.92602074e-01 6.61687970e-01 -6.74274385e-01 1.32468402e-01
8.78077865e-01 6.15021214e-02 -8.03862929e-01 -1.05344045e+00
5.88503957e-01 1.27843153e+00 6.69038713e-01 -6.75401151e-01
8.27241004e-01 9.39386897e-03 -6.03450418e-01 -7.65171289e-01
-1.09868801e+00 -7.61510253e-01 -1.15924549e+00 3.90280187e-01
1.04181409e+00 -9.38950062e-01 2.67949551e-01 1.12133838e-01
-1.48259294e+00 2.22259015e-02 -2.50778675e-01 8.28190565e-01
-2.33120359e-02 2.89273351e-01 -5.88978946e-01 -5.89248180e-01
-7.16121793e-01 -8.85991633e-01 7.98395216e-01 4.16144103e-01
-8.99027511e-02 -1.22709334e+00 7.65567243e-01 1.29562587e-01
1.03858039e-01 -2.42696851e-01 7.06358790e-01 -1.65626395e+00
3.98829997e-01 -4.37046528e-01 -7.43624568e-02 6.02574050e-01
1.64133549e-01 -4.15930092e-01 -9.85900521e-01 -4.22189683e-02
-6.16336986e-02 -5.04190862e-01 2.42290482e-01 -3.17099005e-01
2.41456017e-01 -2.21349955e-01 -3.79214466e-01 3.58863413e-01
1.63307011e+00 -4.05684207e-03 3.85411322e-01 2.39939004e-01
8.14768195e-01 7.32519925e-01 6.44866943e-01 -1.48621306e-01
3.89830291e-01 3.12424183e-01 -3.39956641e-01 -2.85278589e-01
1.27634212e-01 -5.48559666e-01 6.17898285e-01 1.58871853e+00
1.58018321e-02 -1.89561486e-01 -1.15072727e+00 8.43757689e-01
-1.61806178e+00 -7.25570202e-01 -1.70013949e-01 2.14261794e+00
1.14120650e+00 -2.49936119e-01 -3.11121196e-01 -3.70757520e-01
1.15608370e+00 -2.31670037e-01 -7.65559599e-02 -5.63234806e-01
-1.10821441e-01 3.49803895e-01 6.62191272e-01 4.30034220e-01
-1.21743023e+00 1.33732831e+00 6.07724667e+00 7.24260330e-01
-8.53540599e-01 6.34740770e-01 2.21467122e-01 5.17998755e-01
-1.70081928e-01 3.08269590e-01 -1.43648994e+00 2.29768261e-01
1.49166059e+00 -2.71507144e-01 9.60913971e-02 1.15344429e+00
2.48827972e-04 4.05727506e-01 -1.10701823e+00 7.28532851e-01
2.67106056e-01 -9.14374828e-01 -4.41185087e-02 -1.08637266e-01
7.59279847e-01 4.41307783e-01 -8.09806645e-01 5.92747569e-01
7.79469013e-01 -7.77384102e-01 3.09212357e-01 2.64903814e-01
8.18309486e-01 -8.05471003e-01 1.41958690e+00 7.04763532e-01
-1.43880928e+00 3.32255840e-01 -8.39827240e-01 2.95831472e-01
5.25703311e-01 5.53119957e-01 -7.08149552e-01 4.51705217e-01
6.13901198e-01 1.52477905e-01 -3.00881237e-01 9.72135723e-01
-7.30005026e-01 7.67578661e-01 -5.02980709e-01 -2.43494406e-01
1.40622824e-01 -1.30242705e-01 2.20044389e-01 1.69067860e+00
5.26629984e-01 1.51331320e-01 4.76843208e-01 5.69940448e-01
-4.65428472e-01 8.50583136e-01 -7.11642861e-01 -3.07534903e-01
6.08977020e-01 1.61936843e+00 -5.24285913e-01 -5.22302032e-01
-5.90024590e-01 1.18098187e+00 7.37590253e-01 8.88180882e-02
-4.60151404e-01 -8.12732577e-01 1.80643395e-01 -4.20302987e-01
3.37379932e-01 -2.59097248e-01 9.03322082e-03 -1.33483875e+00
-7.78124407e-02 -6.19326890e-01 5.85819662e-01 -4.93119925e-01
-1.68521416e+00 1.21801603e+00 -3.61326844e-01 -1.06450367e+00
-1.95570707e-01 -6.20829701e-01 -6.02103114e-01 1.08928192e+00
-1.44473886e+00 -1.23084450e+00 1.48254290e-01 5.12540579e-01
5.29664516e-01 -3.00560623e-01 1.48546517e+00 5.31036079e-01
-5.00333071e-01 4.78882492e-01 2.48448402e-01 8.31891298e-01
9.18555856e-01 -1.16856778e+00 4.92210567e-01 1.10481870e+00
4.46853250e-01 7.17832267e-01 3.00361603e-01 -8.23626637e-01
-9.72106814e-01 -1.28503740e+00 1.84183836e+00 -3.52951974e-01
5.96869767e-01 -4.18637961e-01 -9.62439179e-01 7.43347406e-01
3.26121926e-01 2.60750443e-01 1.36066616e+00 2.96309114e-01
-3.71412843e-01 1.97455138e-01 -1.35748529e+00 2.98093051e-01
5.64686716e-01 -6.90911412e-01 -1.29926932e+00 3.69782746e-01
7.47101903e-01 -4.82203364e-02 -1.00959003e+00 -3.19301784e-02
9.49185342e-02 -2.23399535e-01 4.82755780e-01 -7.63090611e-01
-3.93150374e-02 -7.07948685e-01 -4.23768997e-01 -1.47135472e+00
-1.41710371e-01 -3.84814024e-01 5.45919061e-01 1.87772620e+00
8.58915329e-01 -4.64447826e-01 3.76747847e-01 6.63434684e-01
-1.29675478e-01 -1.56518847e-01 -7.84622431e-01 -8.38763416e-01
2.34950051e-01 -3.62167031e-01 5.87408423e-01 1.18652809e+00
3.29291463e-01 9.02055681e-01 -3.72332722e-01 5.85853875e-01
4.66742635e-01 -2.63184279e-01 5.37939012e-01 -1.29843235e+00
-7.41716325e-02 3.03954780e-01 -2.82649457e-01 -8.41245592e-01
4.22227353e-01 -1.09008408e+00 6.62923396e-01 -1.53115082e+00
1.36193052e-01 -6.02172196e-01 -3.82215440e-01 9.41166759e-01
-1.17738128e-01 2.99978822e-01 -6.52329773e-02 3.77593756e-01
-5.21787345e-01 4.05799955e-01 4.19221938e-01 2.62068003e-01
-3.14484835e-01 -3.52621436e-01 -6.12659514e-01 1.00191438e+00
8.15258265e-01 -1.10180807e+00 -3.86014283e-02 -7.40230680e-01
3.99679737e-03 -2.14192107e-01 -3.30373049e-01 -8.86510193e-01
4.09492373e-01 -7.83353373e-02 1.88593283e-01 -3.61739725e-01
-1.36855125e-01 -8.80311906e-01 -4.34053019e-02 4.58626226e-02
-2.87410647e-01 1.17156632e-01 3.03429645e-02 3.33379090e-01
-3.11348289e-01 -8.99775028e-01 7.98214972e-01 -1.63219288e-01
-8.11764181e-01 9.73758847e-02 -2.75121689e-01 2.50278533e-01
8.80442739e-01 3.00790817e-01 4.51156832e-02 1.51252255e-01
-7.07209051e-01 -6.72561079e-02 3.90387923e-01 2.84874976e-01
2.02735499e-01 -1.49050879e+00 -8.43825340e-01 7.01243356e-02
2.14150205e-01 -2.32700631e-01 -2.91322678e-01 2.87861139e-01
-4.28835571e-01 2.92057306e-01 -4.59692478e-02 -8.60559270e-02
-1.20142853e+00 4.67148215e-01 1.36815384e-01 -5.33062518e-01
-4.62689936e-01 6.80162549e-01 -3.12165648e-01 -1.42497122e+00
1.57403663e-01 1.43020749e-01 -5.15942454e-01 -1.65036902e-01
5.46881497e-01 1.43297866e-01 1.36572435e-01 -1.17189407e+00
-4.07662630e-01 4.76271659e-01 -6.06709439e-03 -4.80512351e-01
1.47745562e+00 -1.81615815e-01 -1.83330417e-01 4.22191888e-01
1.23408222e+00 3.22269201e-01 -3.95741045e-01 -5.87480664e-01
5.28315723e-01 -1.71721019e-02 3.14589925e-02 -6.62956715e-01
-6.51397347e-01 8.55802476e-01 5.82476676e-01 -1.47041082e-01
8.50089729e-01 3.56245134e-03 1.07738924e+00 8.97730947e-01
4.75474209e-01 -1.45053780e+00 -6.97040379e-01 7.68582702e-01
2.22675070e-01 -1.22998679e+00 -2.77168572e-01 -3.00954849e-01
-8.17175150e-01 1.19115996e+00 4.77669746e-01 -5.39256930e-02
7.41664946e-01 3.15712631e-01 5.10977745e-01 -1.15130886e-01
-4.32116479e-01 -3.48524898e-01 2.55180269e-01 7.11158574e-01
8.78754795e-01 1.12557299e-01 -6.94148421e-01 1.09849811e+00
-2.22287998e-01 -1.09108254e-01 3.79344702e-01 8.99446428e-01
-5.75338602e-01 -1.62570786e+00 -3.74362290e-01 3.37119363e-02
-7.13032603e-01 -4.95539427e-01 -4.26140189e-01 5.01030266e-01
3.65093589e-01 1.01621580e+00 -1.93621203e-01 -2.43216008e-01
4.15209085e-01 6.75458133e-01 -2.82563657e-01 -1.16593575e+00
-6.90425217e-01 5.33567443e-02 5.23087919e-01 -4.66063470e-02
-5.92678070e-01 -4.14714277e-01 -1.59368920e+00 1.37027025e-01
-6.86466813e-01 8.65287721e-01 9.14394915e-01 1.16494155e+00
2.23325193e-01 -1.54111683e-01 5.94766080e-01 -6.74431741e-01
-6.08131588e-01 -1.09547293e+00 -4.92387116e-01 3.92548531e-01
-3.96167964e-01 -2.31328294e-01 -4.52159256e-01 4.77792382e-01] | [9.946303367614746, 9.728140830993652] |
be7b5d4f-7acc-4e56-969b-d545b9fc544d | improving-adversarial-text-generation-by | 2005.01279 | null | https://arxiv.org/abs/2005.01279v1 | https://arxiv.org/pdf/2005.01279v1.pdf | Improving Adversarial Text Generation by Modeling the Distant Future | Auto-regressive text generation models usually focus on local fluency, and may cause inconsistent semantic meaning in long text generation. Further, automatically generating words with similar semantics is challenging, and hand-crafted linguistic rules are difficult to apply. We consider a text planning scheme and present a model-based imitation-learning approach to alleviate the aforementioned issues. Specifically, we propose a novel guider network to focus on the generative process over a longer horizon, which can assist next-word prediction and provide intermediate rewards for generator optimization. Extensive experiments demonstrate that the proposed method leads to improved performance. | ['Lawrence Carin', 'Dinghan Shen', 'Wenlin Wang', 'Changyou Chen', 'Zheng Wen', 'Zhe Gan', 'Ruiyi Zhang', 'Guoyin Wang'] | 2020-05-04 | improving-adversarial-text-generation-by-1 | https://aclanthology.org/2020.acl-main.227 | https://aclanthology.org/2020.acl-main.227.pdf | acl-2020-6 | ['adversarial-text'] | ['adversarial'] | [ 8.61499831e-02 4.16856229e-01 -3.20135742e-01 -1.25991836e-01
-6.90843225e-01 -3.79180849e-01 8.41621101e-01 -8.12747553e-02
-6.65276274e-02 1.12453854e+00 4.81783360e-01 -1.89878896e-01
1.13012798e-01 -8.68365347e-01 -5.39259553e-01 -4.22360599e-01
3.75010788e-01 6.13469779e-01 -2.20336139e-01 -4.61702883e-01
1.75759882e-01 4.77828383e-02 -1.10455728e+00 -4.85532172e-02
1.50084710e+00 5.23514330e-01 8.20058465e-01 4.56032068e-01
-2.67550349e-01 1.03896046e+00 -7.94016421e-01 -3.58051658e-01
-9.35152471e-02 -1.11126935e+00 -8.29865813e-01 2.48496294e-01
-4.26380485e-01 -4.49712515e-01 -3.56637388e-01 9.62951422e-01
5.42976558e-01 4.99084532e-01 7.84060538e-01 -1.04196846e+00
-1.06124091e+00 9.96618986e-01 -2.32779488e-01 -1.17055960e-01
5.39336443e-01 3.79390717e-01 1.02167022e+00 -7.13635385e-01
6.12604737e-01 1.48856854e+00 1.90923318e-01 9.14651692e-01
-9.93294716e-01 -4.38734144e-01 5.32287538e-01 1.08417705e-01
-1.25757992e+00 -3.13792050e-01 9.76661682e-01 -3.03204376e-02
8.93474281e-01 -8.66094157e-02 9.00461197e-01 1.55656803e+00
4.95389849e-01 1.01119387e+00 9.45296109e-01 -4.31954384e-01
3.06225747e-01 -5.74931502e-02 -6.51714027e-01 7.17776120e-01
3.88308987e-02 9.07633528e-02 -5.10441899e-01 9.42809582e-02
9.33396399e-01 -1.12059973e-01 -1.24369904e-01 4.54020407e-03
-1.33100998e+00 1.15666020e+00 3.50079715e-01 1.83291197e-01
-6.32874489e-01 2.10036337e-01 3.10776711e-01 1.82864353e-01
5.17059565e-01 7.64593601e-01 -1.24392649e-02 -2.71897584e-01
-8.29703093e-01 6.64002478e-01 6.74313247e-01 1.12071550e+00
4.85231906e-01 4.40766543e-01 -6.00361407e-01 8.82381499e-01
3.18953067e-01 6.33504570e-01 9.00375903e-01 -7.64428794e-01
5.62243700e-01 2.07910240e-01 3.40292960e-01 -9.12239552e-01
-2.48284966e-01 -2.62909681e-01 -8.10075164e-01 -1.99181601e-01
6.06617108e-02 -5.25738478e-01 -9.10517931e-01 1.85028613e+00
1.81742907e-01 7.95469880e-02 2.21667334e-01 1.03058064e+00
5.39025128e-01 8.04195225e-01 1.95173860e-01 -2.82638580e-01
1.03797448e+00 -1.26462293e+00 -9.95865762e-01 -4.79566336e-01
4.98658180e-01 -4.78630930e-01 1.23463058e+00 1.27542481e-01
-1.27978241e+00 -3.90201807e-01 -8.67969990e-01 9.12544802e-02
-2.87832487e-02 1.77065358e-01 7.00336397e-01 1.05591834e-01
-9.93099689e-01 7.66763985e-01 -7.81688571e-01 -2.41438612e-01
2.01499432e-01 1.89256102e-01 3.01710874e-01 1.92942888e-01
-1.59179759e+00 1.12951398e+00 7.83871055e-01 2.55218804e-01
-1.08114433e+00 -1.09993763e-01 -8.57246995e-01 -1.29832746e-02
3.96640003e-01 -9.84247327e-01 1.52176690e+00 -9.55566466e-01
-2.30973291e+00 1.46911368e-01 -2.58629233e-01 -3.98116231e-01
5.36941767e-01 -5.50736822e-02 -3.60625207e-01 1.22519089e-02
1.94862917e-01 9.77787793e-01 1.04389226e+00 -1.01604390e+00
-4.44760114e-01 1.45452231e-01 4.07039113e-02 7.04549372e-01
-3.14432859e-01 -2.92448223e-01 -1.11999400e-01 -1.06668663e+00
-2.66303360e-01 -8.58609498e-01 -5.88968635e-01 -5.35692096e-01
-5.88190794e-01 -4.59462047e-01 2.49789461e-01 -7.76410103e-01
1.12034750e+00 -1.49717569e+00 4.61519152e-01 -6.90394789e-02
2.49997005e-02 1.61998451e-01 -3.52930695e-01 5.84333122e-01
4.27792639e-01 4.08258736e-02 -1.75012186e-01 -2.18907624e-01
8.19061995e-02 1.79757923e-01 -6.37980640e-01 -1.97419748e-01
2.66012341e-01 1.22515857e+00 -1.29646778e+00 -5.87815166e-01
1.03381991e-01 2.27196187e-01 -5.16608953e-01 4.96769428e-01
-8.59095395e-01 6.03983939e-01 -1.05440068e+00 3.35253298e-01
9.44488943e-02 -3.61672729e-01 2.54723489e-01 2.73832828e-01
8.76555741e-02 5.36963522e-01 -6.38398290e-01 1.95253670e+00
-8.61180902e-01 2.08119765e-01 -3.07226032e-01 -8.40519607e-01
1.05584097e+00 2.52986014e-01 -4.54729237e-02 -6.69908941e-01
3.12362820e-01 1.17852621e-01 7.17406571e-02 -3.23072404e-01
7.71273851e-01 -2.27627724e-01 -1.37240663e-01 7.75502205e-01
-3.85940969e-02 -6.45750642e-01 2.06160843e-01 1.59929767e-01
7.43986905e-01 4.57499176e-01 3.74110699e-01 -1.66404899e-02
5.29173672e-01 5.78924716e-02 5.00404477e-01 7.56166220e-01
1.72171071e-02 3.99988443e-01 3.42557788e-01 -1.14311300e-01
-1.01724255e+00 -8.38395774e-01 5.14673889e-01 9.54697967e-01
2.45995790e-01 -3.50393921e-01 -9.82371390e-01 -6.65485442e-01
-3.56776625e-01 1.24799132e+00 -3.49544048e-01 -5.08183897e-01
-6.23073339e-01 -5.22670448e-01 3.87336999e-01 6.11466587e-01
3.81695896e-01 -1.57600439e+00 -4.15723771e-01 6.69268429e-01
-4.68216836e-01 -8.44423294e-01 -7.50815213e-01 -3.31554592e-01
-9.36450362e-01 -3.93092185e-01 -1.06710470e+00 -8.44376385e-01
6.89140379e-01 2.56952513e-02 9.68527555e-01 3.96909043e-02
1.47442639e-01 -8.92560631e-02 -4.38637286e-01 -4.16270673e-01
-7.58258283e-01 4.18172061e-01 3.74788269e-02 -2.76584774e-01
2.34025121e-02 -4.45958614e-01 -6.00055873e-01 5.36605865e-02
-6.17881298e-01 5.69127262e-01 7.67222643e-01 1.27105093e+00
4.42254752e-01 -1.60058185e-01 1.21569705e+00 -5.97406030e-01
1.37894523e+00 -4.35451120e-01 -5.59024930e-01 4.36638147e-01
-9.37983572e-01 3.30995828e-01 1.16682446e+00 -6.21613085e-01
-1.47897327e+00 -1.02324292e-01 -1.27916839e-02 -4.66029644e-01
8.66557062e-02 4.74725366e-01 -1.66882977e-01 2.86138654e-01
4.26091850e-01 6.82656527e-01 1.44407794e-01 -1.52744353e-01
7.77962685e-01 6.71649873e-01 3.48985314e-01 -6.83161318e-01
6.56089127e-01 3.71378064e-02 -2.73662180e-01 -4.05674815e-01
-8.68868649e-01 3.76798034e-01 -3.65328759e-01 -9.28229019e-02
8.14265192e-01 -8.76772583e-01 -5.00720263e-01 3.02631110e-01
-1.44435394e+00 -7.26407588e-01 -2.63001770e-01 4.57149595e-01
-1.11772799e+00 2.68975437e-01 -6.90426469e-01 -7.74897099e-01
-6.78620040e-01 -1.12401140e+00 1.04866076e+00 4.78868604e-01
-4.73854601e-01 -1.15783226e+00 1.10707521e-01 1.47142231e-01
4.00120825e-01 -9.90501046e-02 8.44157696e-01 -5.84084451e-01
-5.84929526e-01 7.02681020e-02 1.54601410e-02 -4.13553528e-02
1.66063800e-01 -4.17426735e-01 -3.99419636e-01 -1.21385045e-01
-1.38065308e-01 -5.91252148e-01 5.45386493e-01 2.16283649e-01
1.17867827e+00 -6.96413934e-01 -3.34320337e-01 3.05418521e-01
7.98076868e-01 2.14972198e-01 2.96307802e-01 1.09125972e-01
7.89361238e-01 6.48918629e-01 9.25127625e-01 6.59403563e-01
7.01039791e-01 7.28351355e-01 7.26039931e-02 1.49780780e-01
-6.50114715e-02 -9.10048306e-01 4.68277574e-01 1.00649023e+00
4.96722665e-03 -5.13276577e-01 -5.98291695e-01 5.52493572e-01
-2.10803294e+00 -8.24368417e-01 3.73297870e-01 1.68422008e+00
1.14400351e+00 2.68824436e-02 -7.34913424e-02 -4.53559428e-01
7.63853550e-01 3.09232861e-01 -7.54281819e-01 -3.71963263e-01
2.61212885e-02 1.71662465e-01 1.18618861e-01 5.79289675e-01
-6.49736404e-01 1.63121521e+00 6.93855190e+00 1.15806723e+00
-1.02120590e+00 4.41476032e-02 4.83002275e-01 -1.16200969e-01
-7.76676416e-01 -3.78275365e-02 -5.71564496e-01 5.87742925e-01
9.25754428e-01 -7.35103965e-01 8.34246576e-01 7.75781035e-01
6.24354362e-01 1.60972103e-01 -9.80242252e-01 8.00709069e-01
1.43563882e-01 -1.25469100e+00 3.77432317e-01 -1.71555847e-01
9.52583849e-01 -4.38224554e-01 -9.46968943e-02 4.88286883e-01
6.33770883e-01 -1.16080320e+00 9.47528958e-01 6.31237447e-01
6.02935493e-01 -9.97386754e-01 3.47882509e-01 6.03008866e-01
-1.01134264e+00 1.35390181e-02 -5.18548846e-01 -1.47696614e-01
5.11655688e-01 2.92122006e-01 -9.63273644e-01 4.38978344e-01
-2.02533051e-01 6.68422341e-01 -2.39187419e-01 4.60431904e-01
-8.43567014e-01 4.51048285e-01 1.37673929e-01 -7.51489460e-01
3.75478089e-01 -3.11494112e-01 5.25929272e-01 8.04725528e-01
5.26130557e-01 2.09412456e-01 4.73400503e-01 1.30887210e+00
-8.12176690e-02 2.54299432e-01 -6.37893558e-01 -4.12629962e-01
5.30465245e-01 1.13120496e+00 -7.81006336e-01 -4.16587591e-01
-9.65874642e-02 1.40565395e+00 5.54894865e-01 4.71182883e-01
-8.91932666e-01 -4.20456916e-01 2.82314420e-01 -1.34999588e-01
2.21520200e-01 -2.30999917e-01 -1.83533102e-01 -1.27492988e+00
-2.90009320e-01 -9.95180726e-01 1.65626581e-03 -9.01822746e-01
-1.16683590e+00 5.93799531e-01 -9.07228366e-02 -9.28960204e-01
-1.06932521e+00 -1.88201040e-01 -8.78022909e-01 9.61756289e-01
-1.31189764e+00 -1.10505688e+00 1.75061151e-02 4.59118336e-01
9.16353643e-01 -3.28881353e-01 7.93605924e-01 -2.03629538e-01
-5.58359325e-01 6.61465585e-01 -6.90675750e-02 -1.80026457e-01
3.92837524e-01 -1.12243748e+00 5.91054320e-01 7.07718670e-01
1.69599503e-02 5.64778507e-01 7.16849089e-01 -9.88925815e-01
-1.24932575e+00 -1.29630947e+00 8.53513122e-01 1.40662510e-02
7.01627076e-01 -3.86351824e-01 -7.01224685e-01 5.53072274e-01
3.42743248e-01 -6.32141590e-01 2.10801005e-01 -1.45656317e-01
2.72581458e-01 2.93090582e-01 -8.66432250e-01 1.11379731e+00
1.19327641e+00 -4.29401904e-01 -6.39605224e-01 6.78005338e-01
9.71248269e-01 -6.49677694e-01 -5.20823479e-01 -2.97157317e-02
2.71291077e-01 -3.34655374e-01 6.64419949e-01 -5.53856909e-01
6.49143517e-01 -3.64799462e-02 4.16720390e-01 -1.96313810e+00
-2.96758801e-01 -9.95948732e-01 -3.78245562e-01 1.05712426e+00
5.44772625e-01 -6.31928682e-01 6.16125822e-01 3.35146219e-01
-2.16335684e-01 -8.37770581e-01 -6.29422605e-01 -9.00051713e-01
1.43963188e-01 -6.34043664e-02 8.63428056e-01 6.50736332e-01
2.93366283e-01 7.68278003e-01 -7.52224088e-01 -3.54309976e-01
3.16860437e-01 2.52185732e-01 5.39550781e-01 -8.23599041e-01
-4.38698471e-01 -5.27409077e-01 1.62152946e-01 -1.33069670e+00
7.29427576e-01 -9.44188893e-01 5.85122228e-01 -1.70809126e+00
7.42822215e-02 -3.75204772e-01 6.92072585e-02 3.84468585e-01
-5.84940732e-01 -3.69677961e-01 1.22266024e-01 2.71345496e-01
-4.64029700e-01 1.26068616e+00 1.82662344e+00 -7.94553906e-02
-2.48593673e-01 -9.36061377e-04 -8.43038440e-01 5.23032367e-01
8.80326867e-01 -2.60097861e-01 -8.87075603e-01 -4.15808290e-01
1.42458141e-01 4.65443313e-01 1.04645029e-01 -5.64298093e-01
1.54461667e-01 -6.17381334e-01 2.80021012e-01 -3.60398501e-01
1.87068686e-01 -1.75994083e-01 -1.59390420e-01 5.03520548e-01
-7.47045457e-01 1.30772859e-01 -1.99374646e-01 7.17490852e-01
-9.76195112e-02 -3.37225258e-01 3.97240192e-01 -3.18450242e-01
-3.33117127e-01 3.43530923e-01 -5.06937504e-01 2.54955053e-01
1.00993586e+00 2.04092890e-01 1.15617178e-01 -9.15306032e-01
-4.81474280e-01 5.03327131e-01 2.84791738e-01 7.17229068e-01
6.29091561e-01 -1.57369220e+00 -7.31619477e-01 -1.66218579e-01
-1.04791641e-01 3.17289196e-02 1.71329916e-01 4.23731476e-01
-2.23367885e-01 5.77252686e-01 -5.38479462e-02 -1.03138097e-01
-4.99510616e-01 5.30500352e-01 3.17261070e-01 -5.65448582e-01
-6.70842290e-01 6.55324459e-01 1.85895011e-01 -3.60319585e-01
-7.10559189e-02 -1.31487653e-01 -2.41286024e-01 -7.25359097e-02
4.91812557e-01 3.24872673e-01 -2.76939511e-01 -3.30915898e-01
7.69415498e-02 9.39915180e-02 -1.95241883e-01 -5.40073812e-01
9.39142048e-01 -3.11522782e-01 2.12933138e-01 3.28188092e-01
5.75487673e-01 -2.91610748e-01 -1.34159338e+00 -1.88094899e-01
-2.19886631e-01 -1.75215825e-01 -8.22900701e-03 -8.51160228e-01
-7.80478835e-01 7.48246729e-01 3.12609747e-02 -8.40068888e-03
8.49427283e-01 -2.21396014e-01 1.19760084e+00 5.73232710e-01
3.68514448e-01 -1.41875923e+00 3.69875520e-01 5.87144077e-01
1.03109241e+00 -9.76500690e-01 -2.49528602e-01 -1.38170376e-01
-1.08393133e+00 9.78119850e-01 9.01399374e-01 -7.79829770e-02
1.36902794e-01 -1.21841751e-01 -6.03714958e-02 1.15885392e-01
-1.01785457e+00 -1.18470542e-01 1.86339274e-01 6.05189323e-01
5.01866579e-01 1.66368946e-01 -6.75452888e-01 7.01608479e-01
-4.59025204e-01 1.07964925e-01 5.60661376e-01 7.15164304e-01
-3.56964260e-01 -1.30340135e+00 -1.72600091e-01 4.96569276e-01
-1.88312203e-01 -3.21459353e-01 -3.85314852e-01 3.11328739e-01
-3.05230886e-01 1.00485885e+00 -1.44047588e-01 -1.77816167e-01
1.08199380e-01 1.60252005e-01 6.07606769e-01 -7.44417310e-01
-3.28051090e-01 1.88812315e-01 1.63301200e-01 -4.18874472e-01
2.68985014e-02 -5.29508412e-01 -1.55782557e+00 -2.39587143e-01
-4.10960227e-01 2.69718975e-01 4.15192693e-01 1.11481202e+00
5.41894794e-01 6.92280293e-01 8.27531517e-01 -8.46633792e-01
-9.20962930e-01 -1.19936490e+00 -4.94218856e-01 2.37923160e-01
-3.35198715e-02 -6.61770463e-01 -4.59789112e-02 -1.23048812e-01] | [11.89515209197998, 9.130379676818848] |
18f5f108-39c8-4209-a098-86ab4beff4b7 | permutation-invariance-and-uncertainty-in | 2105.12409 | null | https://arxiv.org/abs/2105.12409v1 | https://arxiv.org/pdf/2105.12409v1.pdf | Permutation invariance and uncertainty in multitemporal image super-resolution | Recent advances have shown how deep neural networks can be extremely effective at super-resolving remote sensing imagery, starting from a multitemporal collection of low-resolution images. However, existing models have neglected the issue of temporal permutation, whereby the temporal ordering of the input images does not carry any relevant information for the super-resolution task and causes such models to be inefficient with the, often scarce, ground truth data that available for training. Thus, models ought not to learn feature extractors that rely on temporal ordering. In this paper, we show how building a model that is fully invariant to temporal permutation significantly improves performance and data efficiency. Moreover, we study how to quantify the uncertainty of the super-resolved image so that the final user is informed on the local quality of the product. We show how uncertainty correlates with temporal variation in the series, and how quantifying it further improves model performance. Experiments on the Proba-V challenge dataset show significant improvements over the state of the art without the need for self-ensembling, as well as improved data efficiency, reaching the performance of the challenge winner with just 25% of the training data. | ['Enrico Magli', 'Diego Valsesia'] | 2021-05-26 | null | null | null | null | ['multi-frame-super-resolution'] | ['computer-vision'] | [ 2.93439090e-01 -2.69519180e-01 1.65646181e-01 -4.72538918e-01
-9.11309958e-01 -5.99242449e-01 6.06589854e-01 -1.54402152e-01
-4.90367234e-01 7.87506759e-01 1.07854694e-01 -8.26553181e-02
-6.10584915e-01 -9.12263036e-01 -8.16171348e-01 -7.85257220e-01
-3.79315078e-01 2.92219400e-01 2.47051015e-01 -3.92542213e-01
-2.41658792e-01 5.59878290e-01 -1.75668657e+00 2.95723677e-01
7.56581604e-01 1.05134058e+00 3.31806302e-01 5.43135405e-01
1.94559827e-01 5.97014248e-01 -2.42341459e-01 1.99831165e-02
6.14360988e-01 -3.59625399e-01 -8.19320738e-01 -1.22470386e-01
6.39263093e-01 -5.79777241e-01 -2.35333398e-01 1.10539591e+00
3.83940786e-01 2.40029886e-01 4.23648030e-01 -5.22318780e-01
-6.04146481e-01 6.66987777e-01 -6.68617070e-01 5.02819180e-01
-3.86611372e-01 4.80657555e-02 1.17466331e+00 -7.10188270e-01
5.50556660e-01 1.00985312e+00 8.77515554e-01 1.88168272e-01
-1.51734424e+00 -3.88175905e-01 1.24426320e-01 2.51657277e-01
-1.40059102e+00 -5.58066249e-01 4.15882021e-01 -5.31738460e-01
7.82087743e-01 2.46281445e-01 3.53984177e-01 5.98089516e-01
-1.29295737e-01 3.15517426e-01 1.14823544e+00 -2.50901729e-02
1.63375407e-01 -4.59281281e-02 1.51474804e-01 2.80364841e-01
1.25069216e-01 4.78266478e-01 -4.91586238e-01 1.84345543e-01
7.52103567e-01 -2.60993600e-01 -4.62444365e-01 -1.95543751e-01
-1.03972960e+00 7.40391254e-01 8.54657233e-01 4.39724296e-01
-6.15060687e-01 -5.48165403e-02 1.03909105e-01 2.86066830e-01
7.85854101e-01 6.74620569e-01 -6.29564464e-01 1.57157809e-01
-1.27894330e+00 1.17638998e-01 2.79887229e-01 4.11403924e-01
8.21312129e-01 1.02905251e-01 1.19874598e-02 7.91308463e-01
-2.52974212e-01 3.51992160e-01 1.71216980e-01 -1.35380161e+00
2.37206012e-01 2.25712195e-01 5.92627704e-01 -8.91359627e-01
-5.00893891e-01 -6.68488204e-01 -9.50680912e-01 2.58918881e-01
3.58496904e-01 -1.85022652e-01 -9.67961073e-01 1.93734527e+00
1.01598771e-02 1.69782370e-01 -1.03469174e-02 1.02802527e+00
5.42030454e-01 7.44692802e-01 -2.14464322e-01 -3.03471506e-01
1.14711130e+00 -5.25595725e-01 -5.08105457e-01 -4.23312426e-01
1.07904479e-01 -3.38067979e-01 9.41029310e-01 3.49883646e-01
-9.66934562e-01 -6.98274910e-01 -1.11893916e+00 1.16896756e-01
-2.94298470e-01 7.48862922e-02 7.17677951e-01 1.95311278e-01
-1.31133521e+00 1.30855334e+00 -9.35086727e-01 -1.61113203e-01
6.40028059e-01 1.46473080e-01 -4.03747171e-01 -5.13967406e-03
-1.32178378e+00 9.95688438e-01 5.49716830e-01 3.58138025e-01
-5.78542769e-01 -1.13916361e+00 -5.80099821e-01 3.52336407e-01
3.31752896e-01 -4.74324465e-01 1.00409675e+00 -9.43517268e-01
-1.07219660e+00 5.68912327e-01 -2.87568979e-02 -6.44234121e-01
6.85645401e-01 -2.45617777e-01 -3.18389893e-01 1.61707908e-01
2.02335313e-01 7.42086828e-01 8.43395233e-01 -1.20532250e+00
-8.42620611e-01 -4.29586440e-01 4.31429110e-02 8.01866949e-02
-6.26749471e-02 -2.00053170e-01 -8.18198025e-02 -1.87870815e-01
3.11322778e-01 -8.33024919e-01 -3.72483969e-01 -2.97606201e-03
9.85542610e-02 9.82195511e-02 4.48784947e-01 -7.91528285e-01
7.47622967e-01 -2.31894517e+00 -9.91600379e-02 -9.51317772e-02
1.25578508e-01 1.86520740e-01 -3.99172664e-01 1.41557828e-01
-1.91542581e-01 4.29541260e-01 -4.53350663e-01 -1.30746186e-01
-1.22425131e-01 2.69451410e-01 -4.51979250e-01 4.41985607e-01
5.88370085e-01 5.85754514e-01 -8.18147779e-01 -6.17328919e-02
1.17251098e-01 6.58002377e-01 -3.48470896e-01 8.93061459e-02
-2.76099861e-01 6.79353535e-01 -1.85354039e-01 -1.61393639e-02
9.87048984e-01 -5.18767893e-01 1.89443350e-01 -3.31994176e-01
-3.79609615e-01 5.02084613e-01 -1.21128297e+00 1.43383169e+00
-5.06901860e-01 6.55743182e-01 1.21469267e-01 -8.21456254e-01
7.50940263e-01 6.27720580e-02 5.86410642e-01 -9.38737929e-01
-3.09700936e-01 9.81922448e-02 6.91427216e-02 -3.60561848e-01
6.97098017e-01 -2.91868001e-01 4.17071193e-01 3.89343560e-01
-1.63020223e-01 -5.45487739e-02 2.63587683e-01 -7.56626874e-02
8.46223116e-01 2.66718417e-01 5.46878316e-02 -2.69595444e-01
5.77110201e-02 1.45970225e-01 5.58388770e-01 9.58043456e-01
8.84754285e-02 7.90010333e-01 3.51734459e-01 -5.66798985e-01
-1.23595655e+00 -9.31848049e-01 -4.63138819e-01 9.90937769e-01
-5.37972897e-02 6.83740154e-02 -3.87391597e-01 -2.03492567e-01
-1.22093767e-01 6.76846981e-01 -8.49636137e-01 1.58434302e-01
-3.59813899e-01 -1.21301150e+00 1.69770390e-01 5.26799738e-01
7.38427699e-01 -6.98614419e-01 -7.60845125e-01 2.31538370e-01
-3.54439795e-01 -1.31145060e+00 1.39683619e-01 2.23428577e-01
-1.11888373e+00 -6.25598311e-01 -3.57653916e-01 -6.37425017e-03
2.02997237e-01 5.17025769e-01 1.22895336e+00 -1.94749400e-01
-1.91110969e-01 -1.71082824e-01 -2.76782393e-01 -6.08433634e-02
-1.18387684e-01 1.19228847e-01 -1.84786648e-01 -1.17142592e-02
1.99009806e-01 -9.37386632e-01 -6.00479543e-01 1.24161839e-01
-9.09245610e-01 1.59031883e-01 4.81804818e-01 9.15673018e-01
5.94007015e-01 5.13267279e-01 4.40777421e-01 -6.16858363e-01
2.03037634e-01 -4.33702379e-01 -8.08905244e-01 1.21442646e-01
-5.86658895e-01 4.64487642e-01 5.09248376e-01 -9.72806364e-02
-1.14439940e+00 5.56839928e-02 6.52302355e-02 -3.58048826e-01
-7.85849392e-02 9.08227861e-01 1.62243158e-01 9.75167304e-02
9.58115160e-01 1.76867738e-01 2.39920113e-02 -8.22963357e-01
4.29662794e-01 3.07691187e-01 6.03059351e-01 -3.07666779e-01
7.30746567e-01 8.40249419e-01 2.75180757e-01 -7.91502953e-01
-1.18311632e+00 -3.76085788e-01 -7.85628498e-01 8.62838328e-02
6.35087371e-01 -1.11485803e+00 -4.16072845e-01 2.52043337e-01
-8.81757677e-01 -4.18228149e-01 -4.38638180e-01 5.16158462e-01
-3.44408929e-01 2.07982421e-01 -3.83242100e-01 -8.09911251e-01
-2.11737365e-01 -8.88350546e-01 9.07116950e-01 7.64151439e-02
1.87958345e-01 -6.57345533e-01 -8.12511425e-03 1.08289674e-01
7.54711568e-01 2.74661511e-01 7.58173168e-01 -1.79700688e-01
-8.88470650e-01 -3.27765010e-02 -6.80956244e-01 5.09782910e-01
1.79523185e-01 -1.28193516e-02 -1.27164447e+00 -1.13372311e-01
1.49347022e-01 -2.30356753e-01 1.37730551e+00 6.06242776e-01
1.06059289e+00 -2.22682253e-01 1.52163595e-01 8.50694537e-01
1.57756341e+00 -3.01118404e-01 6.73528016e-01 3.76017809e-01
5.30500174e-01 8.49400997e-01 4.90976036e-01 3.92787427e-01
2.23629311e-01 5.69441795e-01 7.34854698e-01 -4.97385822e-02
-1.56402197e-02 1.18380502e-01 -2.16883048e-02 2.37950906e-01
-5.59171975e-01 6.66579306e-02 -8.12443256e-01 7.30713904e-01
-1.84626091e+00 -1.22174072e+00 -1.70996949e-01 2.32636714e+00
8.90551090e-01 -2.34813690e-01 -1.45960435e-01 -1.76953021e-02
4.20665652e-01 5.67812085e-01 -6.88156247e-01 1.96487859e-01
-3.79229367e-01 3.10426265e-01 8.91457021e-01 6.43770158e-01
-1.37873375e+00 7.71159947e-01 6.40414572e+00 5.38800716e-01
-1.31557965e+00 1.34575948e-01 5.08823574e-01 -1.56902984e-01
-3.16937745e-01 -1.02649458e-01 -5.35275161e-01 2.45982602e-01
1.06809163e+00 5.94666126e-05 6.52684808e-01 6.17420912e-01
4.02829826e-01 -3.09475005e-01 -1.08902931e+00 6.42850041e-01
-3.22824419e-01 -1.33873320e+00 -9.31852907e-02 1.23864867e-01
7.53086686e-01 5.52454770e-01 2.16330275e-01 5.81957139e-02
4.97276038e-01 -1.20153379e+00 6.19493127e-01 7.42354989e-01
7.41153717e-01 -6.57133758e-01 8.07658672e-01 3.86457443e-01
-1.00485241e+00 -4.28658016e-02 -6.34724557e-01 -1.29197240e-01
-1.47778383e-02 8.11303496e-01 -7.17779458e-01 8.24094653e-01
9.07328308e-01 8.38999033e-01 -6.82527244e-01 7.18733490e-01
-8.26877579e-02 6.23279691e-01 -6.21757448e-01 5.33634543e-01
1.94958374e-01 -2.70517856e-01 4.32924300e-01 8.06375384e-01
5.39402723e-01 3.34434986e-01 -6.88292310e-02 1.08310235e+00
6.17022105e-02 -3.39462101e-01 -4.79079843e-01 2.69982442e-02
2.89136797e-01 1.08838856e+00 -4.08938557e-01 -1.35726243e-01
-1.75465137e-01 6.77559137e-01 2.84176528e-01 3.46524179e-01
-5.40652871e-01 2.57440746e-01 7.78585196e-01 1.75402209e-01
6.49925292e-01 -3.42016011e-01 -4.14888859e-01 -1.01572680e+00
1.93236887e-01 -8.08327734e-01 3.68299037e-01 -1.02691352e+00
-1.24809766e+00 7.02439845e-01 -9.09499303e-02 -1.15133882e+00
-3.34949434e-01 -4.88301873e-01 -1.41748264e-01 1.17872334e+00
-1.81704581e+00 -1.02203000e+00 -4.49307114e-01 2.32113853e-01
1.87410682e-01 3.24654251e-01 6.27758741e-01 1.72294185e-01
-2.41752326e-01 -6.52730791e-03 4.12026733e-01 -2.60931905e-02
5.22002101e-01 -1.16715753e+00 3.96575451e-01 1.13908076e+00
1.07024178e-01 5.09907305e-01 8.89576852e-01 -2.81808347e-01
-1.15966284e+00 -1.21222579e+00 7.07856297e-01 -4.67316806e-01
8.26355934e-01 8.09810534e-02 -1.28010905e+00 5.11794984e-01
-1.09055668e-01 3.15196604e-01 3.05186689e-01 2.74115711e-01
-5.94320953e-01 -4.93475199e-01 -1.01863003e+00 2.88252681e-01
1.06266248e+00 -6.68568552e-01 -4.28084671e-01 2.17089295e-01
6.81121588e-01 -1.55090824e-01 -9.05078232e-01 6.54665232e-01
5.34053743e-01 -1.32327402e+00 1.06438899e+00 -6.09510720e-01
7.63052821e-01 -3.83812040e-01 -4.57658798e-01 -1.35302603e+00
-6.58207953e-01 -2.03355141e-02 2.91387796e-01 8.80521655e-01
5.09443879e-01 -5.03502011e-01 3.06374609e-01 5.94596982e-01
8.51612687e-02 -1.64081439e-01 -9.07122135e-01 -9.16804135e-01
1.10545009e-01 -2.49367937e-01 8.23735297e-01 1.08563006e+00
-6.44025028e-01 2.95580894e-01 -4.82604057e-01 6.75378203e-01
7.46387720e-01 4.45737809e-01 3.63225579e-01 -1.51542985e+00
-3.33061755e-01 -4.47359532e-01 -7.67361298e-02 -7.83429980e-01
-1.76739275e-01 -6.15802705e-01 2.86178380e-01 -1.46020508e+00
2.59727091e-01 -4.59232718e-01 -3.72607857e-01 5.59994459e-01
-3.20510834e-01 2.85153240e-01 1.20762527e-01 3.68111730e-01
-3.80216569e-01 4.45057929e-01 1.06426096e+00 -6.45532385e-02
-1.57433361e-01 -1.29313916e-01 -6.60673797e-01 4.62918311e-01
7.82687008e-01 -3.94411922e-01 -3.72777313e-01 -8.90774488e-01
4.85110253e-01 2.99101714e-02 6.72117293e-01 -9.45079565e-01
-3.36486399e-02 -2.28117839e-01 3.89124781e-01 -5.49651086e-01
4.23815012e-01 -8.49367380e-01 5.92888474e-01 5.37996665e-02
-3.93372118e-01 -3.12497586e-01 3.38294506e-01 3.66410106e-01
-1.58310771e-01 -1.44811004e-01 9.22365427e-01 -3.05166215e-01
-8.28000307e-01 4.46997732e-01 5.19659072e-02 -1.86010867e-01
3.31924498e-01 1.45542383e-01 -4.42921788e-01 -3.98519039e-01
-8.45168293e-01 1.06241457e-01 4.97965991e-01 2.29742855e-01
7.11260587e-02 -9.72642243e-01 -1.12369967e+00 1.01673983e-01
-8.10393319e-02 2.04531848e-01 6.90454066e-01 6.83573186e-01
-2.65368819e-01 2.39113182e-01 -3.04415107e-01 -8.23504388e-01
-8.55100691e-01 3.95540565e-01 6.38638973e-01 -3.68109643e-01
-6.37949228e-01 7.09393501e-01 8.46768767e-02 -2.53285140e-01
-3.57356161e-01 -4.25641149e-01 -2.02479213e-01 4.54366893e-01
6.55557930e-01 1.11273207e-01 3.20612043e-01 -4.53850985e-01
-2.86666036e-01 5.02108932e-01 1.22326970e-01 -2.73087114e-01
1.93483841e+00 -1.98078349e-01 -2.33600318e-01 5.73776782e-01
8.76373291e-01 -3.75009447e-01 -1.67865980e+00 -5.66527188e-01
-1.24382362e-01 -6.75900042e-01 5.59194446e-01 -8.19968641e-01
-1.24937749e+00 1.09155798e+00 7.13615894e-01 4.34521526e-01
1.24564564e+00 -2.25835428e-01 3.20864111e-01 5.57000995e-01
4.59231794e-01 -8.35379064e-01 -4.27139014e-01 6.50637269e-01
9.52778995e-01 -1.54745936e+00 2.32487783e-01 3.91684920e-02
-5.07509291e-01 8.75132620e-01 8.19604918e-02 1.24979364e-02
7.39986420e-01 1.57243535e-01 -4.24980698e-03 -2.02014521e-01
-7.05957532e-01 -5.74320734e-01 2.34950855e-01 5.28863549e-01
1.86193332e-01 2.12517515e-01 6.08148091e-02 4.01423961e-01
-2.84810185e-01 1.74662352e-01 4.59686399e-01 4.76598233e-01
-5.25528908e-01 -6.25919700e-01 -2.55160570e-01 4.82168823e-01
-4.30981874e-01 -3.52157354e-01 2.18146728e-04 6.47551596e-01
8.27524588e-02 6.97380841e-01 2.76637137e-01 -2.46043831e-01
1.39877066e-01 -8.60297829e-02 2.94965029e-01 -4.88700807e-01
-1.87025398e-01 -4.92185391e-02 2.09806442e-01 -5.89329720e-01
-6.75301909e-01 -7.75832415e-01 -7.52948999e-01 -4.30920571e-01
-7.97788650e-02 2.18579412e-01 6.34012341e-01 9.48650837e-01
4.95188057e-01 5.20448923e-01 7.98414290e-01 -1.11166799e+00
-8.24872017e-01 -1.12174320e+00 -6.72257960e-01 1.37595996e-01
8.60235095e-01 -4.85519141e-01 -4.82379943e-01 1.03240564e-01] | [9.845919609069824, -1.6535167694091797] |
771f2b52-d37e-4848-baa5-03fb223da2b2 | an-effective-entropy-assisted-mind-wandering | 2005.12076 | null | https://arxiv.org/abs/2005.12076v2 | https://arxiv.org/pdf/2005.12076v2.pdf | An Effective Entropy-assisted Mind-wandering Detection System with EEG Signals based on MM-SART Database | Mind-wandering (MW), which usually defined as a lapse of attention, occurs between 20%-40% of the time, has negative effects on our daily life. Therefore, detecting when MW occurs can prevent us from those negative outcomes resulting from MW, such as failing to keep track of course during learning. In this work, we first collect a multi-modal Sustained Attention to Response Task (MM-SART) database for detecting MW. Eighty-two participants' data are collected in our experiments. For each participant, we collect measures of 32-channels electroencephalogram (EEG) signals, photoplethysmography (PPG) signals, galvanic skin response (GSR) signals, eye tracker signals, and several questionnaires for detailed analyses. Then, we propose an effective MW detection system based on the collected EEG signals. To explore the non-linear characteristics of EEG signals, we utilize the entropy-based features in time, frequency, and wavelet domains. The experimental results show that we can reach 0.712 AUC score by using the random forest (RF) classifier with the leave-one-subject-out cross-validation. Moreover, to lower the overall computational complexity of the MW detection system, we apply techniques of channel selection and feature selection. By using the only two most significant EEG channels, we can reduce the training time of the classifier by 44.16%. By performing correlation importance feature elimination (CIFE) on the feature set, we can further improve the AUC score to 0.725 but with only 14.6% of the selection time compared with the recursive feature elimination (RFE) method. By proposing the MW detection engine, current work can be applied to educational scenarios, especially in the era of remote learning nowadays. | ['An-Yeu Wu', 'Su-Ling Yeh', 'Win-Ken Beh', 'Zih-Ling Chen', 'Ching-Yen Shih', 'Hsing-Hao Lee', 'Yi-Ta Chen'] | 2020-05-25 | null | null | null | null | ['photoplethysmography-ppg'] | ['medical'] | [ 1.23706073e-01 -2.63166755e-01 1.72384173e-01 -1.74495041e-01
-3.74074399e-01 -1.30590990e-01 -1.05289489e-01 8.67696106e-02
-5.53936005e-01 9.30643916e-01 -6.96124360e-02 -2.36904457e-01
-4.50489849e-01 -6.70551121e-01 -1.42815694e-01 -7.16043174e-01
-1.74957514e-01 -5.65825045e-01 2.27356348e-02 -6.85168579e-02
7.19054520e-01 3.44086349e-01 -1.63913977e+00 7.04972027e-03
1.41018248e+00 1.20609653e+00 2.19844431e-01 1.74984515e-01
1.28197804e-01 4.96539950e-01 -9.04095471e-01 -1.16827460e-02
-2.27254048e-01 -5.86610675e-01 -4.69452232e-01 -2.41749316e-01
-2.39291891e-01 -1.38263762e-01 -1.14316180e-01 1.16365445e+00
9.07088995e-01 2.61741012e-01 2.87210733e-01 -1.30624175e+00
-3.01248550e-01 2.72131056e-01 -6.77338600e-01 5.06725967e-01
6.32352293e-01 -1.32221393e-02 2.23902017e-01 -7.61632204e-01
-1.37244128e-02 7.16080368e-01 5.51881075e-01 5.18302262e-01
-9.94041800e-01 -1.28735340e+00 -1.16999753e-01 1.00167155e+00
-1.75321722e+00 -2.86140174e-01 9.47238982e-01 -3.77662867e-01
7.90055335e-01 5.79031587e-01 1.02827406e+00 9.69676018e-01
6.80872440e-01 4.23925936e-01 1.57261050e+00 -3.76183063e-01
2.80926287e-01 2.85896271e-01 5.52862942e-01 4.45251465e-01
1.84579596e-01 -2.12733120e-01 -7.20883489e-01 -1.76645875e-01
3.82966608e-01 -1.41175129e-02 -7.72393823e-01 2.63348669e-01
-9.23641205e-01 4.60038066e-01 1.28407767e-02 3.57675433e-01
-4.28397119e-01 -4.84091014e-01 1.75573871e-01 4.84910846e-01
4.23259735e-01 5.03235281e-01 -4.57434475e-01 -5.24823666e-01
-7.90888727e-01 -3.18676084e-01 6.41026914e-01 5.03861725e-01
5.57218850e-01 -1.28176689e-01 -3.51278722e-01 8.48140776e-01
-6.90785050e-02 4.68863338e-01 1.13684571e+00 -3.94656926e-01
7.66446516e-02 6.97949946e-01 -1.63465440e-01 -1.00318301e+00
-6.04132652e-01 -4.70277309e-01 -8.95431459e-01 -2.09860891e-01
-7.74230883e-02 -3.48392904e-01 -4.53176022e-01 1.62768185e+00
1.50341734e-01 3.16725165e-01 -2.57885486e-01 7.11423934e-01
7.44270325e-01 3.47011566e-01 4.99029681e-02 -9.05373871e-01
1.58240402e+00 -3.25141579e-01 -1.09310424e+00 7.80022820e-04
3.80198747e-01 -5.00507593e-01 1.21331847e+00 7.86708772e-01
-6.55584395e-01 -4.48768198e-01 -1.08198822e+00 4.77087826e-01
-2.92038262e-01 1.41253948e-01 4.77270097e-01 8.65988672e-01
-7.09605396e-01 6.30020976e-01 -7.22482800e-01 -2.14339018e-01
3.90136212e-01 5.44012725e-01 -3.20172042e-01 3.12872261e-01
-1.40063369e+00 1.04491937e+00 1.41381333e-02 -1.96668133e-03
-4.71885175e-01 -7.02342808e-01 -4.59805369e-01 2.85729736e-01
1.05689958e-01 -2.79223323e-01 7.19011605e-01 -8.41294706e-01
-1.58752334e+00 3.23993295e-01 -3.69815141e-01 -4.11508642e-02
1.82698339e-01 -1.51712433e-01 -7.77218640e-01 1.49348423e-01
-1.54389217e-01 -1.39739609e-03 8.24120820e-01 -3.70153695e-01
-2.75786281e-01 -5.26511550e-01 -5.59102654e-01 7.72475749e-02
-8.42725456e-01 2.00802401e-01 9.11881328e-02 -3.45207661e-01
1.58478320e-01 -8.25174630e-01 1.59399629e-01 -6.13630056e-01
-3.47319901e-01 -3.23816627e-01 6.10870779e-01 -8.82389009e-01
1.65992820e+00 -2.37288880e+00 -2.41038337e-01 4.39551383e-01
2.20013052e-01 1.15055591e-01 9.44198817e-02 2.65485160e-02
-4.37636793e-01 -5.05488515e-02 -3.57742421e-02 1.70115337e-01
-3.46055210e-01 -4.43538785e-01 -4.50183339e-02 6.14627421e-01
-4.19359878e-02 4.99243885e-01 -4.70081002e-01 -3.61012101e-01
2.46955916e-01 3.02321970e-01 -2.14130357e-01 2.61773884e-01
8.23751152e-01 4.00353372e-01 -2.77061105e-01 4.67473000e-01
6.91331327e-01 1.80160731e-01 -1.73512220e-01 3.37500051e-02
-2.34165445e-01 2.75011390e-01 -1.01712286e+00 1.43851149e+00
-3.81843567e-01 8.45260441e-01 -2.78291672e-01 -8.36626172e-01
1.06017017e+00 4.88942623e-01 4.21099424e-01 -9.10936892e-01
3.45371574e-01 5.01045957e-02 2.74053097e-01 -8.98750544e-01
-1.41807973e-01 7.62422383e-02 1.76355153e-01 3.08441043e-01
3.14421058e-02 2.64943808e-01 -7.31350109e-02 -1.39269650e-01
1.19096708e+00 -3.32241416e-01 5.47512293e-01 -4.61376697e-01
8.29304695e-01 -5.44807434e-01 5.58746397e-01 4.02899861e-01
-2.73195595e-01 1.06408924e-01 4.66042846e-01 -1.01209089e-01
-9.41832364e-02 -5.41241765e-01 -5.42514801e-01 7.59500980e-01
1.10621326e-01 -3.01559240e-01 -9.74419057e-01 -3.96313310e-01
-1.18721515e-01 1.06270123e+00 -4.83807683e-01 -7.43245125e-01
-4.92445827e-02 -1.00525665e+00 3.09206277e-01 1.54515356e-01
5.51918328e-01 -1.22920084e+00 -7.25940526e-01 1.15157850e-01
-2.72782058e-01 -6.56871617e-01 -2.59804547e-01 2.05684200e-01
-8.96171749e-01 -1.02171755e+00 -5.11913538e-01 -3.59800130e-01
5.57433605e-01 2.86607534e-01 4.74328578e-01 -9.29685831e-02
-4.65367764e-01 2.73527712e-01 -2.55299389e-01 -6.34215355e-01
2.27675930e-01 -6.15905672e-02 3.77156734e-01 8.82944763e-02
8.41502607e-01 -9.46132243e-01 -7.59871602e-01 2.41843879e-01
-2.89711118e-01 -1.43748969e-01 6.26225889e-01 5.21975935e-01
1.65018797e-01 1.64300427e-01 9.86647546e-01 -3.21183413e-01
1.05248559e+00 -5.44403493e-01 -2.60565192e-01 2.07757860e-01
-1.00007188e+00 -3.53686005e-01 4.84453410e-01 -8.81295800e-01
-9.10879374e-01 -3.18757027e-01 2.15156339e-02 -2.44137570e-01
-3.02312076e-01 3.40074003e-01 -1.82203248e-01 -3.09782863e-01
6.12238944e-01 5.40160716e-01 -9.13268849e-02 -3.63720417e-01
-4.45514679e-01 1.22679627e+00 2.30784312e-01 -1.44997835e-01
2.84625024e-01 -1.12575509e-01 -8.82583559e-02 -9.94121552e-01
-5.52436352e-01 -3.58758807e-01 -3.15334886e-01 -3.54487896e-01
5.94164908e-01 -9.17973816e-01 -1.28598070e+00 3.76814306e-01
-9.41418946e-01 2.61952609e-01 2.30925769e-01 1.12004983e+00
-2.57907867e-01 3.54991078e-01 -3.62572640e-01 -1.22471726e+00
-8.05914283e-01 -8.69277537e-01 1.89177006e-01 4.94849682e-01
-4.88796413e-01 -3.95079643e-01 1.06370850e-02 2.12423414e-01
4.17952657e-01 -6.47826344e-02 8.29832196e-01 -7.25426912e-01
1.25721321e-01 -2.63598859e-01 2.04629004e-01 4.12626565e-01
1.13384776e-01 -3.12093109e-01 -1.23882985e+00 -1.76489666e-01
7.04147220e-01 2.82138400e-02 4.95713979e-01 3.91436726e-01
1.69744539e+00 -8.84775594e-02 -9.56562683e-02 5.64319432e-01
1.05354679e+00 8.28300118e-01 8.13266337e-01 3.27733517e-01
3.44279230e-01 5.31445622e-01 4.55431670e-01 5.40915370e-01
1.66719958e-01 2.21007913e-01 1.00561261e-01 1.34249225e-01
2.21422151e-01 4.66280207e-02 6.07321084e-01 1.01972842e+00
-2.80302167e-01 1.20598115e-01 -7.49983907e-01 3.10871750e-01
-1.49917293e+00 -8.67141664e-01 -4.79334354e-01 2.52513790e+00
9.08529580e-01 -4.43873554e-02 -1.75202250e-01 5.17112494e-01
7.91266382e-01 -4.28807408e-01 -5.94928324e-01 -3.26341301e-01
2.45462507e-01 6.09914720e-01 1.95128545e-01 -1.03601426e-01
-7.11427748e-01 1.91931576e-01 5.38918686e+00 8.56235445e-01
-1.35842884e+00 1.10470362e-01 3.94373327e-01 -3.06254923e-01
2.88521070e-02 -1.82485744e-01 -4.35254931e-01 9.66279864e-01
1.25020516e+00 -5.46163499e-01 6.72015309e-01 7.01969504e-01
5.15115321e-01 -4.87806022e-01 -8.31108868e-01 1.50539362e+00
1.83054507e-01 -4.84120697e-01 -5.56647182e-01 -1.04116246e-01
1.49763301e-01 -4.11350667e-01 -1.29189551e-01 4.38703030e-01
-6.43586457e-01 -9.83354509e-01 2.98208743e-01 8.08826864e-01
7.98468173e-01 -9.48697090e-01 7.27860034e-01 6.39895320e-01
-8.29949439e-01 -2.31632695e-01 -2.56057620e-01 -3.57743233e-01
-4.94194210e-01 8.67860436e-01 -4.04504806e-01 4.47275996e-01
8.11065316e-01 3.90420407e-01 -6.52575910e-01 1.23305726e+00
-3.13910455e-01 7.35063493e-01 -1.06536269e-01 -4.09835488e-01
-5.77297568e-01 -2.04963520e-01 4.94411260e-01 8.20590138e-01
7.89432287e-01 5.53977370e-01 -5.22490203e-01 8.26957941e-01
2.09687382e-01 3.80814075e-01 -3.57965767e-01 3.35606515e-01
6.10386848e-01 1.49303246e+00 -4.43894058e-01 -1.18315503e-01
-3.70629728e-01 8.43343794e-01 3.12951766e-02 3.26529026e-01
-7.62298405e-01 -9.94193017e-01 3.59112978e-01 -4.90294173e-02
-3.85484815e-01 9.92925465e-02 -4.76530999e-01 -1.29371727e+00
5.15721254e-02 -6.88595057e-01 1.19554870e-01 -8.20626140e-01
-1.07576013e+00 4.60298747e-01 -1.03593454e-01 -1.19230783e+00
1.83387607e-01 -1.73617765e-01 -9.75840271e-01 1.17602944e+00
-1.20894170e+00 -7.51514435e-02 -6.37733221e-01 9.51301336e-01
2.92174011e-01 5.61401956e-02 1.09638906e+00 4.09276932e-01
-1.01313901e+00 6.51425958e-01 -1.86017141e-01 -1.69750765e-01
8.61511171e-01 -8.73132944e-01 -4.29631680e-01 5.82895517e-01
-3.44200402e-01 8.23876619e-01 5.02909720e-01 -3.86944830e-01
-1.25168300e+00 -6.34920955e-01 8.80564272e-01 6.37265444e-02
3.94685060e-01 -2.57087767e-01 -1.01172948e+00 1.26203597e-01
1.64484575e-01 -3.70010197e-01 1.07366765e+00 2.32538864e-01
2.99342960e-01 -3.23531330e-01 -1.21041787e+00 4.29199100e-01
6.32861733e-01 -4.60390627e-01 -7.86326170e-01 6.37269914e-02
2.57757157e-01 7.30713218e-05 -1.04999602e+00 3.19898337e-01
7.26543367e-01 -8.94878328e-01 4.83051687e-01 -1.62095338e-01
1.11560218e-01 4.43766173e-03 3.04640025e-01 -1.56109667e+00
-4.30875391e-01 -6.80479586e-01 -1.65868804e-01 1.22541296e+00
9.75605547e-02 -9.36371684e-01 3.21307749e-01 7.87138343e-01
-3.56583274e-03 -7.85529554e-01 -8.84407878e-01 -5.74916780e-01
-2.92401791e-01 -4.37977016e-01 6.06411517e-01 1.04095852e+00
8.11853945e-01 3.65594238e-01 -1.55646846e-01 -2.06748415e-02
5.11103034e-01 -1.92791931e-02 2.89042354e-01 -1.44665897e+00
7.36796260e-02 -2.94074297e-01 -4.24740374e-01 -3.66723686e-01
9.66810957e-02 -6.22480154e-01 -1.69604361e-01 -1.18625772e+00
5.31224549e-01 7.38504231e-02 -8.55707228e-01 6.23276293e-01
-3.67261350e-01 9.14735813e-03 -1.59321517e-01 -1.15434313e-02
-2.01192468e-01 7.18683004e-01 1.00361753e+00 1.80937827e-01
-5.06200552e-01 1.89129844e-01 -6.99482739e-01 7.30424643e-01
1.01940870e+00 -5.22204101e-01 -4.60349560e-01 2.49836817e-01
-8.16090405e-03 3.28617275e-01 6.74628541e-02 -1.06715214e+00
2.51518935e-01 -2.00131223e-01 8.00479293e-01 -2.84160078e-01
1.18787110e-01 -6.71325803e-01 3.11983153e-02 5.62273145e-01
-4.30914871e-02 -9.52160284e-02 3.22534233e-01 1.34255826e-01
-4.02666666e-02 -2.87523031e-01 6.83893442e-01 2.16739863e-01
-4.38839376e-01 -2.65807789e-02 -6.56215191e-01 -3.15846771e-01
1.30992115e+00 -2.83025324e-01 -3.21394652e-01 -2.59807169e-01
-7.40062654e-01 1.41415462e-01 -1.25728205e-01 2.42164522e-01
7.66161919e-01 -1.18354762e+00 -3.27130944e-01 6.73871517e-01
-3.52534652e-02 -5.46065569e-01 6.26423538e-01 1.54079831e+00
3.18952948e-02 2.57371455e-01 -5.59518754e-01 -2.92491078e-01
-1.52855563e+00 3.41016114e-01 9.02253017e-02 2.40487963e-01
-5.15377760e-01 6.56013846e-01 -1.28865108e-01 2.88394660e-01
2.88124859e-01 -1.08311146e-01 -7.54766822e-01 4.02715296e-01
7.08751261e-01 7.82249093e-01 3.75715971e-01 -1.25568286e-01
-6.89160168e-01 3.62941235e-01 1.61662981e-01 2.86765844e-02
1.25221860e+00 -1.39071330e-01 -4.13909227e-01 5.41849852e-01
1.03053832e+00 2.23685160e-01 -6.12744212e-01 4.16964769e-01
-2.03819737e-01 -4.08203542e-01 3.87305081e-01 -7.75857389e-01
-8.76262248e-01 1.02805352e+00 1.04698384e+00 2.87548840e-01
1.62811089e+00 -5.35730243e-01 5.62279582e-01 2.56373823e-01
5.55224836e-01 -1.05530548e+00 -2.24383622e-01 7.83987120e-02
6.18733048e-01 -7.77786434e-01 7.56048113e-02 -1.30307406e-01
-6.09514058e-01 1.13451755e+00 6.51416004e-01 -1.11072529e-02
8.73134434e-01 1.22464456e-01 -2.59437591e-01 -3.17571759e-02
-8.94878447e-01 1.65032446e-02 3.58569801e-01 3.56095344e-01
5.05768359e-01 1.41865656e-01 -8.51946831e-01 1.23359227e+00
-3.63417685e-01 1.83164179e-01 6.18833065e-01 6.04991257e-01
-6.42162740e-01 -6.47170544e-01 -4.48054343e-01 7.76951432e-01
-5.86960673e-01 -2.21147284e-01 -2.66123563e-01 4.43448544e-01
1.91185847e-01 1.27885258e+00 5.58104292e-02 -8.03423762e-01
4.92427230e-01 4.46329534e-01 4.29899693e-01 -3.58837962e-01
-4.94184583e-01 1.60183534e-02 -2.99421906e-01 -6.30409837e-01
-2.03107834e-01 -5.09723127e-01 -1.16115415e+00 -3.87245506e-01
-6.78932428e-01 4.16418493e-01 7.77756393e-01 8.84662032e-01
4.12593544e-01 7.89388537e-01 8.44420731e-01 -3.06617826e-01
-4.24222976e-01 -1.45191443e+00 -9.14101839e-01 4.54771407e-02
1.93793759e-01 -7.96193242e-01 -6.93952441e-01 -3.94895524e-01] | [13.304965019226074, 3.3495094776153564] |
e829c409-7a1a-4367-801e-fa1eb77476d0 | i-dont-know-where-he-is-not-does-deception | null | null | https://aclanthology.org/W12-0402 | https://aclanthology.org/W12-0402.pdf | ``I Don't Know Where He is Not'': Does Deception Research yet Offer a Basis for Deception Detectives? | null | ['Lee Gillam', 'Anna Vartapetiance'] | 2012-04-01 | null | null | null | ws-2012-4 | ['deception-detection'] | ['miscellaneous'] | [-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] | [-7.2529401779174805, 3.779604911804199] |
c7c69d0f-9d86-40d3-a868-e9306e9463b1 | 2305-14405 | 2305.14405 | null | https://arxiv.org/abs/2305.14405v1 | https://arxiv.org/pdf/2305.14405v1.pdf | NeuralMatrix: Moving Entire Neural Networks to General Matrix Multiplication for Efficient Inference | In this study, we introduce NeuralMatrix, a novel framework that enables the computation of versatile deep neural networks (DNNs) on a single general matrix multiplication (GEMM) accelerator. The proposed approach overcomes the specificity limitations of ASIC-based accelerators while achieving application-specific acceleration levels compared to general-purpose processors such as CPUs and GPUs. We address the challenges of mapping both linear and nonlinear operations in DNN computation to general matrix multiplications and the impact of using a GEMM accelerator on DNN inference accuracy. Extensive experiments are conducted on various DNN models from three popular categories (i.e., CNN, Transformers, and GNN) as illustrative backbone models. Our results demonstrate that DNNs suffer only up to a 2.02% accuracy loss after being converted to general matrix multiplication, while achieving 113x to 19.44x improvements in throughput per power compared to CPUs and GPUs. | ['An Zou', 'Yiran Li', 'Xin He', 'Jie Zhao', 'Ruiqi Sun'] | 2023-05-23 | null | null | null | null | ['specificity'] | ['natural-language-processing'] | [-1.25549987e-01 -3.61858070e-01 2.29093060e-01 -4.82985556e-01
-5.25305159e-02 -4.64476138e-01 6.39734149e-01 -9.47149470e-03
-8.84499550e-01 5.96016943e-01 -3.95971864e-01 -9.52829003e-01
3.65461171e-01 -9.79281604e-01 -9.05926108e-01 -7.68705070e-01
2.18734860e-01 2.31727242e-01 2.15590626e-01 -1.56467661e-01
1.11423455e-01 8.90668511e-01 -1.95721173e+00 9.59318429e-02
6.24637246e-01 1.17778432e+00 2.01508328e-01 7.68064976e-01
-5.81264086e-02 9.22305167e-01 -7.96633899e-01 -6.46168947e-01
3.04925412e-01 1.52161181e-01 -2.74586082e-01 -7.20227778e-01
8.71036887e-01 -7.38665640e-01 -5.79988122e-01 1.21486044e+00
6.81306243e-01 1.02437213e-01 3.40151578e-01 -1.48427892e+00
-1.66036412e-01 5.40017486e-01 -4.84496266e-01 7.00559556e-01
-3.51553380e-01 2.56774932e-01 6.46239936e-01 -8.30025673e-01
1.70948580e-01 1.31793296e+00 1.01559770e+00 6.20301187e-01
-9.50857341e-01 -1.02160180e+00 -1.78219989e-01 6.74210191e-02
-1.27818823e+00 -6.33211672e-01 6.89843222e-02 -3.02494407e-01
1.60654140e+00 2.13653982e-01 7.58702934e-01 1.09655428e+00
6.57529652e-01 4.00679886e-01 4.80015516e-01 -1.42960018e-02
5.06618619e-01 -2.30244130e-01 6.94251299e-01 8.23704779e-01
8.41055751e-01 5.33014275e-02 -6.26307964e-01 -1.32000044e-01
9.90174472e-01 -1.10608153e-01 2.48644605e-01 4.54520911e-01
-1.04876876e+00 8.06832433e-01 5.60032904e-01 -1.09629981e-01
-5.22597790e-01 6.67897284e-01 1.03171790e+00 4.56370413e-02
1.73086777e-01 -8.25260729e-02 -4.73205537e-01 -4.09430057e-01
-9.38119531e-01 3.93090755e-01 7.35472977e-01 1.20822251e+00
4.64632094e-01 5.92935443e-01 7.49559840e-03 4.67495829e-01
3.51924628e-01 5.14852762e-01 5.04941940e-01 -7.78916538e-01
6.95937693e-01 4.37088728e-01 -4.52142209e-01 -6.54560924e-01
-7.16886401e-01 -5.25743186e-01 -1.40679944e+00 2.38495082e-01
1.84266508e-01 -5.02618790e-01 -9.55292583e-01 1.66126132e+00
2.69613326e-01 1.76776052e-01 2.82112241e-01 6.68549299e-01
1.23790145e+00 9.33350563e-01 4.47655320e-01 3.61837149e-01
1.67148232e+00 -1.32514668e+00 -4.37642902e-01 -2.37223774e-01
8.55707169e-01 -5.28737783e-01 9.19425189e-01 2.20322266e-01
-1.24269307e+00 -7.28289664e-01 -1.25716996e+00 -7.28722692e-01
-3.63162488e-01 3.90201718e-01 1.07681334e+00 9.35455322e-01
-1.49319625e+00 5.94061494e-01 -1.45006609e+00 -2.79710859e-01
5.71745992e-01 6.87010586e-01 5.46818897e-02 1.90728769e-01
-8.13995540e-01 8.03397357e-01 6.97161853e-01 4.57535684e-02
-6.97414517e-01 -1.35960186e+00 -4.66548085e-01 3.39380085e-01
-5.76370895e-01 -1.01442432e+00 1.27461159e+00 -4.25414383e-01
-1.50545084e+00 5.72051227e-01 8.06023926e-03 -8.15373719e-01
3.54112417e-01 -1.26257047e-01 -3.99920136e-01 -1.34585306e-01
-2.18922079e-01 9.70105231e-01 3.96262586e-01 -1.04178548e-01
-8.38435590e-01 -3.29197705e-01 1.24198049e-01 -2.77461205e-02
-5.40127635e-01 2.33812407e-01 -2.63165712e-01 -5.14839292e-01
1.18737713e-01 -1.08703017e+00 -2.61623919e-01 3.65128905e-01
-3.39004278e-01 -1.73309118e-01 8.88022959e-01 -5.12154162e-01
8.46095502e-01 -2.06840754e+00 -2.64749676e-01 7.52605870e-02
3.78874600e-01 5.89801431e-01 7.91400820e-02 -1.69119462e-01
1.07252777e-01 -3.97678912e-01 3.19720477e-01 -4.95878398e-01
-6.13818988e-02 6.15039408e-01 -3.24233890e-01 2.20698580e-01
-9.14439261e-02 7.83818007e-01 -6.03455663e-01 -1.30873233e-01
1.87881887e-01 9.34377670e-01 -8.41517508e-01 -7.14920163e-02
1.45943537e-01 1.24605998e-01 -4.90187295e-02 7.07369864e-01
1.02673709e+00 -2.77574927e-01 2.68663019e-01 -4.95359182e-01
-3.04438949e-01 6.08331740e-01 -7.31099904e-01 1.41790962e+00
-6.28680527e-01 9.29399848e-01 7.50543401e-02 -7.10481822e-01
7.53204167e-01 2.68351853e-01 -1.72282621e-01 -6.06892228e-01
3.87419283e-01 3.27011645e-01 3.68494391e-01 3.15633565e-01
7.34739780e-01 2.59095520e-01 2.42217571e-01 4.36632559e-02
4.81925189e-01 4.49450493e-01 3.83669466e-01 -1.40176699e-01
1.09487212e+00 -4.12636638e-01 4.63658012e-03 -9.02234733e-01
3.32678914e-01 -1.29302949e-01 4.12973762e-01 8.00709069e-01
8.56111478e-03 -9.75613147e-02 5.46653450e-01 -8.43833745e-01
-1.16150093e+00 -1.17368293e+00 -4.68523860e-01 1.32264316e+00
-3.55864674e-01 -6.37524664e-01 -8.83080065e-01 8.93318504e-02
-1.25158429e-01 5.78700006e-01 -1.20510183e-01 1.83012672e-02
-6.21642888e-01 -1.56535637e+00 1.13635516e+00 9.39035177e-01
9.50254321e-01 -6.31991148e-01 -1.11887574e+00 1.88401118e-01
4.70418572e-01 -1.45918977e+00 1.09955780e-01 4.93328005e-01
-1.45667875e+00 -4.01404500e-01 -2.87053198e-01 -6.50081992e-01
5.81754804e-01 1.24554344e-01 1.24620771e+00 -5.31259030e-02
-2.65040547e-01 -3.75931203e-01 3.91614169e-01 -3.82734835e-01
-3.59404922e-01 4.24315959e-01 4.82906997e-01 -5.48912466e-01
2.94562787e-01 -1.04328334e+00 -6.00321352e-01 -9.51971412e-02
-5.49413502e-01 4.36390013e-01 5.84500432e-01 7.27756977e-01
4.14669305e-01 7.26196868e-03 6.26385733e-02 -7.16693997e-01
4.62833792e-01 -4.40428078e-01 -1.11963558e+00 -1.05297253e-01
-5.96359491e-01 6.25400543e-02 9.22336817e-01 -4.56214309e-01
-7.92567313e-01 4.74171862e-02 -7.90772617e-01 -3.62977892e-01
1.61499307e-01 2.95290917e-01 -1.30638018e-01 -4.67022985e-01
6.38114691e-01 1.84352532e-01 -1.58949479e-01 -4.07814026e-01
3.68959531e-02 1.77531496e-01 8.01942348e-01 -7.55798757e-01
2.47918710e-01 2.15291560e-01 3.91509205e-01 -9.72960353e-01
-4.74505305e-01 3.98415700e-02 -2.22830757e-01 1.28300264e-01
7.88462639e-01 -1.43407464e+00 -1.11597645e+00 7.85616934e-01
-1.43958843e+00 -2.26698652e-01 8.63563865e-02 5.41029453e-01
2.91392603e-03 -7.26003274e-02 -1.28649676e+00 -4.21447754e-01
-1.11288226e+00 -1.32225096e+00 7.97010362e-01 4.26874191e-01
5.05449995e-02 -9.84318376e-01 -3.86239111e-01 2.51729898e-02
7.73469210e-01 3.71714309e-02 9.40671742e-01 -3.62131357e-01
-7.04847038e-01 4.53738309e-02 -7.21450388e-01 2.79918939e-01
-4.84333009e-01 2.52082586e-01 -1.07827401e+00 -4.49813366e-01
6.65597618e-02 -8.92499015e-02 7.18784392e-01 3.35315794e-01
1.18710613e+00 -4.55920398e-01 -2.48500317e-01 1.32795775e+00
1.65617836e+00 3.35695416e-01 7.86799848e-01 2.27259204e-01
1.02881777e+00 -2.35537231e-01 -1.48965210e-01 5.17497718e-01
2.22454742e-01 5.57956100e-01 4.86095577e-01 2.14563757e-02
8.95686075e-02 8.37795320e-04 2.52043992e-01 1.26292884e+00
-3.55235189e-01 -1.93222806e-01 -9.68424797e-01 1.06777556e-01
-1.59858227e+00 -5.04101396e-01 -5.37715673e-01 1.62707829e+00
6.05671823e-01 3.59234750e-01 -1.74814418e-01 1.10236615e-01
4.93605971e-01 -1.58496760e-02 -7.21028507e-01 -7.91834891e-01
-3.98653038e-02 5.56921482e-01 8.55010152e-01 1.43662840e-01
-9.39088881e-01 9.06227887e-01 6.81438398e+00 9.15155172e-01
-1.33456492e+00 4.56727862e-01 6.17844760e-01 -3.74368817e-01
3.55353951e-01 -4.07023877e-01 -1.47589970e+00 7.23273873e-01
1.64906693e+00 2.38840319e-02 2.80690879e-01 1.29409158e+00
-2.84589648e-01 1.95902705e-01 -1.15485573e+00 1.37639713e+00
-4.36687261e-01 -1.55612624e+00 3.83202970e-01 1.08622454e-01
4.65403676e-01 6.39126420e-01 4.67693284e-02 3.99321616e-01
2.15225145e-01 -8.80650759e-01 7.13188708e-01 -8.24781880e-02
8.28445196e-01 -1.05162847e+00 8.87711465e-01 1.24550916e-01
-9.79016840e-01 1.27104387e-01 -8.30264151e-01 -5.87685168e-01
-5.70968762e-02 6.86895132e-01 -5.60884356e-01 1.16530955e-02
8.96933556e-01 3.51892143e-01 -4.18256968e-01 5.20185590e-01
4.24610637e-02 6.23248875e-01 -6.00927591e-01 -1.87979162e-01
3.92325431e-01 -2.28022292e-01 -2.50476673e-02 1.36650229e+00
5.26487410e-01 1.32596523e-01 -5.54295838e-01 7.52656221e-01
-2.13335365e-01 -2.34529287e-01 -7.79284537e-02 1.78192064e-01
5.22538841e-01 1.59079194e+00 -9.33646619e-01 -6.91249847e-01
-3.73401999e-01 9.66335535e-01 4.31327820e-01 -1.13855533e-01
-1.30209112e+00 -4.20543045e-01 1.35886300e+00 -1.66102216e-01
1.40870005e-01 -3.24833900e-01 -7.56998718e-01 -1.16721749e+00
-5.19635947e-03 -5.38114905e-01 4.57617044e-02 -6.30492091e-01
-9.35547113e-01 9.56235707e-01 -3.61576706e-01 -6.90612614e-01
2.49213316e-02 -1.05649877e+00 -5.57189882e-01 8.69324446e-01
-1.19404900e+00 -8.63367617e-01 -5.68890512e-01 4.82070208e-01
2.74432153e-02 -4.18472558e-01 9.29736257e-01 9.63527679e-01
-9.43185568e-01 1.01882839e+00 3.15087765e-01 1.11205146e-01
-8.24087635e-02 -7.46482909e-01 1.27529764e+00 9.90953863e-01
-1.93616211e-01 8.72152805e-01 5.99146068e-01 -1.19032614e-01
-1.61724591e+00 -1.31322360e+00 9.03391361e-01 -1.11251853e-01
6.53975487e-01 -7.69414723e-01 -6.67965770e-01 8.87806058e-01
-4.55534533e-02 3.19750279e-01 4.80727345e-01 -2.48382743e-02
-3.46124232e-01 -2.87113786e-01 -1.22692776e+00 7.68409014e-01
1.26771712e+00 -3.97004783e-01 6.52747750e-02 5.23320138e-01
5.78267992e-01 -9.49882686e-01 -8.16642344e-01 2.49210656e-01
6.24905229e-01 -1.07387817e+00 1.18850887e+00 -4.62206960e-01
5.17662883e-01 -1.04707286e-01 -3.29595745e-01 -8.22145343e-01
-2.60588706e-01 -3.82396281e-01 -5.43268204e-01 9.46416616e-01
2.61509586e-02 -9.04886067e-01 9.69771624e-01 6.97075427e-01
-3.92050475e-01 -6.59219384e-01 -1.19400263e+00 -7.60227323e-01
-6.78352043e-02 -4.73448724e-01 7.22550571e-01 7.13383973e-01
-4.91504073e-01 5.18829525e-01 -3.74681354e-02 2.67873943e-01
7.07899213e-01 -3.81210774e-01 5.63260734e-01 -1.20219874e+00
-8.72817263e-02 -4.72990096e-01 -9.39183950e-01 -1.15870142e+00
3.52264009e-02 -9.18277979e-01 -4.21864599e-01 -1.11219120e+00
8.68026316e-02 -4.47657883e-01 -1.91981852e-01 5.63178003e-01
2.08309665e-01 4.91545498e-01 1.47694588e-01 -4.37937095e-04
-2.73600191e-01 2.23969474e-01 7.14933991e-01 2.34612480e-01
1.30549803e-01 -1.59329504e-01 -4.75792170e-01 8.14353645e-01
8.95863831e-01 -3.07463408e-01 -5.25829911e-01 -1.08528483e+00
1.46347642e-01 -1.82679296e-01 6.09622121e-01 -1.70596397e+00
5.78367710e-01 4.28177118e-01 5.43316245e-01 -8.21566939e-01
4.90604579e-01 -4.84799892e-01 3.74303669e-01 8.27812791e-01
1.25454217e-01 7.95480132e-01 8.20064247e-01 -2.21153279e-03
1.42936874e-02 -1.29418094e-02 9.79614198e-01 1.95856974e-01
-8.18076491e-01 3.36348772e-01 -5.72025776e-01 -9.68699977e-02
6.15388393e-01 -4.21772189e-02 -7.17285335e-01 2.00100005e-01
-2.78389573e-01 -2.27103695e-01 5.06319059e-03 1.19470902e-01
3.90870184e-01 -1.37314177e+00 -4.83953804e-01 4.29326862e-01
-4.66843367e-01 8.65155458e-02 4.61036950e-01 6.86754346e-01
-1.35874522e+00 8.36586416e-01 -8.48249555e-01 -6.98876143e-01
-1.34591937e+00 2.84660369e-01 3.45580399e-01 -1.57222420e-01
-8.29969645e-01 1.23040032e+00 2.46425465e-01 -3.48921508e-01
2.60849774e-01 -9.86771345e-01 2.71508425e-01 -2.00629428e-01
6.75130188e-01 6.44057512e-01 7.16540933e-01 -3.49370092e-01
-5.84624708e-01 1.09909505e-01 -7.96092004e-02 2.60882288e-01
1.18964732e+00 4.27883357e-01 -4.05523717e-01 -7.50996396e-02
1.39146757e+00 -7.32987404e-01 -1.06312990e+00 9.87157151e-02
-2.75292039e-01 1.14567623e-01 3.98726285e-01 -4.45016354e-01
-1.59612703e+00 1.04094326e+00 8.40329289e-01 -3.82907659e-01
1.15233159e+00 -5.18762648e-01 1.10658443e+00 7.34441042e-01
3.78606170e-01 -8.50269616e-01 -6.17007732e-01 9.95695710e-01
2.12722898e-01 -8.20259213e-01 9.08517167e-02 -2.71779895e-01
4.01678830e-02 1.21646130e+00 8.99399102e-01 -2.96179533e-01
7.90155232e-01 8.73185039e-01 -2.54928619e-01 -4.27492335e-02
-1.01974976e+00 4.70796824e-01 -1.83961973e-01 3.30754787e-01
3.79439831e-01 1.20326065e-01 -1.34146929e-01 5.48022687e-01
-8.15681338e-01 1.30069945e-02 3.52947533e-01 8.97078574e-01
1.06575228e-02 -5.23978114e-01 -2.40191184e-02 6.83823526e-01
-6.17845416e-01 -6.05210006e-01 4.95508194e-01 7.09095776e-01
2.14304283e-01 3.63504082e-01 8.49261403e-01 -5.31568408e-01
2.88712289e-02 -1.76513553e-01 5.14901400e-01 -1.82323381e-01
-9.82772887e-01 -5.70891738e-01 6.63103908e-02 -4.22917396e-01
6.69958517e-02 -3.38450402e-01 -1.28279078e+00 -1.31615031e+00
1.28374279e-01 -3.16424817e-01 1.22157085e+00 5.45452476e-01
6.64632916e-01 9.12363946e-01 -2.18417048e-01 -1.04800963e+00
-5.33363700e-01 -6.75371766e-01 -4.73070890e-01 -2.62628436e-01
1.22261323e-01 -4.94810790e-01 -2.12848708e-01 -2.40773857e-01] | [8.432574272155762, 2.944458246231079] |
673958f9-667c-4c0b-a2f8-ff674c1de7c9 | i-2-sdf-intrinsic-indoor-scene-reconstruction | 2303.07634 | null | https://arxiv.org/abs/2303.07634v2 | https://arxiv.org/pdf/2303.07634v2.pdf | I$^2$-SDF: Intrinsic Indoor Scene Reconstruction and Editing via Raytracing in Neural SDFs | In this work, we present I$^2$-SDF, a new method for intrinsic indoor scene reconstruction and editing using differentiable Monte Carlo raytracing on neural signed distance fields (SDFs). Our holistic neural SDF-based framework jointly recovers the underlying shapes, incident radiance and materials from multi-view images. We introduce a novel bubble loss for fine-grained small objects and error-guided adaptive sampling scheme to largely improve the reconstruction quality on large-scale indoor scenes. Further, we propose to decompose the neural radiance field into spatially-varying material of the scene as a neural field through surface-based, differentiable Monte Carlo raytracing and emitter semantic segmentations, which enables physically based and photorealistic scene relighting and editing applications. Through a number of qualitative and quantitative experiments, we demonstrate the superior quality of our method on indoor scene reconstruction, novel view synthesis, and scene editing compared to state-of-the-art baselines. | ['Rui Wang', 'Hujun Bao', 'Wei Hua', 'Rui Tang', 'Lisha Wang', 'Dianbing Xi', 'Jifan Li', 'Fujun Luan', 'Qi Ye', 'Yuchi Huo', 'Jingsen Zhu'] | 2023-03-14 | null | null | null | null | ['indoor-scene-reconstruction'] | ['computer-vision'] | [ 6.30288005e-01 -3.00564200e-01 7.17350304e-01 -7.00846612e-01
-6.69413030e-01 -6.72957420e-01 5.80779076e-01 -3.18295687e-01
4.67678503e-04 1.00980520e+00 3.24492633e-01 1.00643970e-01
-8.24834555e-02 -1.34712708e+00 -1.22106838e+00 -4.20492768e-01
3.73683572e-01 3.02061588e-01 -4.75084223e-02 -2.02846482e-01
1.16329454e-01 7.47545242e-01 -1.52380157e+00 3.00475568e-01
1.26813650e+00 8.47818851e-01 3.30327541e-01 8.55897844e-01
-1.57618940e-01 9.55615282e-01 -1.77363738e-01 -3.49237263e-01
4.22289610e-01 -4.46235180e-01 -6.72889352e-01 -1.14213333e-04
1.03902185e+00 -6.41388655e-01 -4.70532387e-01 9.23524797e-01
4.17324245e-01 5.56847334e-01 7.31223524e-01 -5.55979192e-01
-6.85910940e-01 -2.27296457e-01 -5.00344694e-01 -3.15849811e-01
8.36866260e-01 8.08188245e-02 5.64650536e-01 -1.19705164e+00
8.38022113e-01 1.34995711e+00 9.57912743e-01 3.51245463e-01
-1.42651355e+00 -4.49521929e-01 2.54886776e-01 -4.87646967e-01
-1.26027083e+00 -4.02434915e-01 1.17291057e+00 -5.38077533e-01
7.97896028e-01 5.32640338e-01 8.39000285e-01 1.02047765e+00
3.06836635e-01 3.86765420e-01 1.36225009e+00 -2.44390160e-01
4.42120165e-01 -1.34468421e-01 -4.49896932e-01 1.18638098e+00
-1.38076469e-01 5.00416398e-01 -6.20022058e-01 -1.48161218e-01
1.41558838e+00 2.65207440e-01 -4.47623432e-01 -5.49718618e-01
-1.38884807e+00 4.59662199e-01 7.67013252e-01 -2.89789766e-01
-2.43613139e-01 7.06636071e-01 -1.48549363e-01 -1.98126301e-01
1.04972243e+00 2.59430200e-01 -3.81505728e-01 2.31909126e-01
-9.74405766e-01 4.73473132e-01 5.16585886e-01 1.08312094e+00
9.32003498e-01 2.75043964e-01 -8.77068043e-02 8.16763520e-01
5.08356333e-01 1.04111648e+00 -6.41044676e-01 -1.66126478e+00
1.69047192e-01 2.21253380e-01 2.93286294e-01 -9.30911005e-01
1.02054551e-01 -3.00953299e-01 -8.84813249e-01 4.79667693e-01
7.43212476e-02 1.21125422e-01 -1.09537327e+00 1.46391201e+00
6.22780502e-01 4.72113073e-01 -2.12437198e-01 6.87430620e-01
1.05044389e+00 8.59217584e-01 -4.20901746e-01 5.82285039e-02
9.84828651e-01 -1.07539141e+00 -5.15912354e-01 -6.24414161e-03
-1.83852799e-02 -6.31881535e-01 1.20683157e+00 3.81326735e-01
-1.46480370e+00 -3.83167624e-01 -8.00516665e-01 -5.04037559e-01
5.23222750e-03 -2.11234629e-01 8.51603806e-01 5.04756868e-01
-9.54469085e-01 6.83941603e-01 -8.23919594e-01 6.24314025e-02
7.72476792e-01 -5.38071059e-02 -2.28248187e-03 -5.94870746e-01
-6.86851978e-01 1.91247821e-01 -3.86518776e-01 -9.04564112e-02
-1.27553785e+00 -1.51461387e+00 -1.06789577e+00 -2.99895436e-01
1.20702051e-01 -1.59275782e+00 9.85300839e-01 -6.76375866e-01
-1.59193802e+00 8.28497529e-01 -3.66050839e-01 9.20465514e-02
7.60158837e-01 -1.60948142e-01 8.07066038e-02 2.83166375e-02
2.78296798e-01 5.71002662e-01 6.81102395e-01 -2.02971268e+00
-2.78250247e-01 -3.29397172e-01 2.65921056e-01 4.25067723e-01
5.20055473e-01 -4.96825159e-01 -2.50716895e-01 -9.55051124e-01
2.13966444e-01 -4.92223442e-01 -4.69811141e-01 6.37016475e-01
-4.69291389e-01 6.53599262e-01 4.28852618e-01 -7.48168111e-01
5.80589235e-01 -1.95752943e+00 2.98117369e-01 2.16197848e-01
9.44163427e-02 -6.37661219e-01 5.77295460e-02 2.17395008e-01
2.82995850e-01 -7.30400383e-02 -8.75223398e-01 -7.93028653e-01
-5.46179600e-02 1.92007616e-01 -4.75046337e-01 5.13654649e-01
-3.55216444e-01 8.86097789e-01 -1.15486610e+00 -1.24973416e-01
6.62642837e-01 1.16611338e+00 -1.01387429e+00 1.99307069e-01
-5.01091838e-01 1.17787755e+00 -4.13785368e-01 6.91148996e-01
1.08984661e+00 -9.48432386e-02 -3.42257649e-01 -3.78160775e-01
-3.40404332e-01 4.96239141e-02 -1.21360636e+00 2.68244767e+00
-1.03775179e+00 4.60810810e-01 4.28221732e-01 -1.37500957e-01
5.60406446e-01 -1.68407679e-01 5.38280547e-01 -8.04369628e-01
-1.08403832e-01 1.66172534e-01 -1.18389678e+00 4.60417494e-02
7.04400897e-01 -2.79987782e-01 2.14871943e-01 2.70938128e-01
-2.75452644e-01 -1.10588014e+00 -4.09529567e-01 2.37413228e-01
7.76260853e-01 7.65321970e-01 -1.86076239e-01 -4.16256368e-01
4.37487602e-01 -1.61800712e-01 2.92810202e-01 4.99536008e-01
5.00863969e-01 1.12482870e+00 -4.31174010e-01 -3.44379038e-01
-1.15678310e+00 -1.68524909e+00 -3.30555528e-01 7.28374898e-01
3.51708144e-01 -2.07368582e-02 -1.06446636e+00 -3.78092915e-01
2.68309303e-02 1.04694676e+00 -5.08288622e-01 3.26306492e-01
-6.63201749e-01 -3.83623570e-01 5.39372489e-02 4.22579139e-01
8.49226832e-01 -7.70916939e-01 -6.11765444e-01 1.57661587e-01
-2.58009017e-01 -1.02271318e+00 -7.07137585e-01 -2.05976501e-01
-8.40978324e-01 -9.64902639e-01 -8.36851716e-01 -5.64501822e-01
1.01636195e+00 4.70106721e-01 1.44930232e+00 -7.62196481e-02
-5.44160604e-01 8.91665161e-01 2.50260502e-01 -7.21438080e-02
-2.76373684e-01 -7.01869607e-01 -2.64747441e-01 -8.79314318e-02
-9.59778070e-01 -9.97059762e-01 -1.05756569e+00 2.77462572e-01
-9.67287958e-01 6.64678395e-01 -2.30080366e-01 5.18270075e-01
1.06893337e+00 -9.12393928e-02 -2.68000603e-01 -1.00901389e+00
1.30262613e-01 -6.31751716e-02 -7.87009239e-01 2.04122722e-01
-2.05985904e-01 1.43089825e-02 5.84842503e-01 1.57260731e-01
-1.79833984e+00 -8.59933998e-03 -2.92564213e-01 -3.52866054e-01
-2.35657439e-01 -2.40429536e-01 -1.99974716e-01 -5.05255699e-01
6.26984894e-01 3.28692615e-01 -6.49960756e-01 -3.50969255e-01
7.53114104e-01 -6.25416562e-02 6.71224296e-01 -9.93648052e-01
8.12082231e-01 1.38359320e+00 2.01740354e-01 -8.91452909e-01
-1.07804251e+00 -9.61094275e-02 -5.67745030e-01 -3.80300283e-01
9.79641080e-01 -1.14571762e+00 -5.98431528e-01 7.82369435e-01
-1.41160095e+00 -8.79382670e-01 -7.91551232e-01 9.84708220e-02
-9.06888723e-01 1.62359178e-01 -5.43623507e-01 -7.92997420e-01
-1.37711406e-01 -1.13537848e+00 1.63774681e+00 -2.37895865e-02
1.65829748e-01 -1.24716234e+00 3.25058669e-01 6.16831005e-01
3.14469308e-01 8.27065527e-01 6.68899179e-01 9.21140552e-01
-1.29718125e+00 6.96133256e-01 -3.63812149e-01 3.08694631e-01
2.32678503e-01 5.46325743e-03 -1.40364933e+00 -1.81051075e-01
3.85520682e-02 -1.36564886e-02 8.15114498e-01 7.68363118e-01
1.42379820e+00 -1.58242300e-01 -2.89957732e-01 1.55088472e+00
1.75434566e+00 1.54086232e-01 7.76714742e-01 -2.08785310e-01
1.25145006e+00 6.04630053e-01 2.77282596e-01 5.15644550e-01
6.95433497e-01 4.50283408e-01 5.97982824e-01 -2.89148450e-01
-5.81082642e-01 -5.77525496e-01 1.10905252e-01 7.67458916e-01
-2.82150239e-01 -3.38202894e-01 -5.19818485e-01 3.34512204e-01
-1.31598115e+00 -8.72627974e-01 -4.32102382e-01 2.15973020e+00
6.16643131e-01 -3.69706035e-01 -5.69477022e-01 -2.21744731e-01
1.78864762e-01 4.15576428e-01 -6.35818005e-01 -1.03630111e-01
-1.91200644e-01 4.69698578e-01 5.68076551e-01 1.19725204e+00
-7.20261216e-01 7.83899784e-01 6.23988056e+00 7.33833075e-01
-7.61943638e-01 2.77543455e-01 8.44618678e-01 -1.32265046e-01
-1.46245134e+00 -2.02462047e-01 -3.90622526e-01 7.81848505e-02
3.73809516e-01 3.10493320e-01 9.76815403e-01 4.83870834e-01
3.84736270e-01 -3.13866049e-01 -1.05553782e+00 1.19561756e+00
1.11031331e-01 -1.90028799e+00 7.89738595e-02 -7.75424903e-03
1.55046237e+00 8.32178257e-03 2.28838827e-02 -2.95267999e-01
7.73064673e-01 -1.00235558e+00 1.19067609e+00 9.92901683e-01
1.28918922e+00 -7.81676650e-01 -2.36759514e-01 1.55788898e-01
-1.32092237e+00 3.49148154e-01 3.36637311e-02 2.63269603e-01
5.20630240e-01 1.00294542e+00 -9.50009152e-02 9.40988123e-01
7.68174887e-01 8.68093848e-01 -1.05732605e-02 5.60203016e-01
-1.06679961e-01 1.53786376e-01 -2.92478621e-01 5.33250391e-01
-1.67134598e-01 -6.20970964e-01 5.06174564e-01 9.19189632e-01
3.74659389e-01 4.12177205e-01 -4.99887094e-02 1.65493774e+00
-3.18980455e-01 -2.51717418e-01 -6.87167823e-01 4.79139149e-01
1.45629585e-01 1.01362991e+00 -7.17235982e-01 -2.00199932e-01
-3.57504338e-02 1.55811191e+00 1.19117975e-01 7.60032058e-01
-8.96374345e-01 -9.53363925e-02 6.76318228e-01 4.25023466e-01
1.00692630e-01 -5.06821394e-01 -7.15457618e-01 -1.40938365e+00
3.28453295e-02 -2.66006708e-01 -2.94151336e-01 -1.27729857e+00
-1.24077880e+00 3.57274264e-01 3.75809823e-03 -8.11960101e-01
2.22635403e-01 -3.23845923e-01 -4.86069530e-01 7.96493709e-01
-1.79768336e+00 -1.33952844e+00 -6.58792734e-01 7.44836688e-01
6.48492515e-01 5.23988485e-01 6.85801864e-01 3.16565335e-01
6.23503374e-03 1.50524052e-02 5.06476879e-01 -2.91570723e-01
3.24962258e-01 -1.22462189e+00 7.75402188e-01 8.24900329e-01
-2.33549401e-02 4.91289943e-01 4.61862534e-01 -7.79632032e-01
-1.51506400e+00 -1.43613899e+00 5.27667776e-02 -5.56266844e-01
-9.90035906e-02 -6.20889366e-01 -5.22229135e-01 5.25027573e-01
-1.39538040e-02 2.98617661e-01 3.43763411e-01 -5.51962078e-01
-2.77083218e-01 -6.84800297e-02 -1.84038532e+00 5.67073762e-01
1.82547188e+00 -6.81359351e-01 2.90921330e-02 4.79486555e-01
9.75882649e-01 -8.42097104e-01 -7.50534892e-01 4.26226825e-01
5.02039194e-01 -1.38131440e+00 1.63980055e+00 8.31714198e-02
7.00159073e-01 -4.09803301e-01 -6.52126253e-01 -1.47633350e+00
-5.89011014e-02 -7.77260423e-01 -2.81962771e-02 1.01474035e+00
-1.26576141e-01 -5.31916916e-01 6.92484081e-01 4.97559369e-01
-6.23756588e-01 -6.34223819e-01 -7.74642169e-01 -2.86960751e-01
-5.36123700e-02 -7.59397388e-01 7.95323253e-01 1.00158620e+00
-1.23362482e+00 -2.50771374e-01 -1.12496868e-01 2.43974596e-01
1.35893846e+00 2.91643709e-01 7.44329572e-01 -9.59500372e-01
-3.69452029e-01 2.77549550e-02 3.76938552e-01 -1.44211555e+00
4.59765606e-02 -8.33080113e-01 2.34504849e-01 -2.11683941e+00
4.46649408e-03 -5.57164848e-01 4.31999058e-01 -3.01443458e-01
2.08262399e-01 3.83297116e-01 -2.23825544e-01 -2.32145578e-01
-2.68096924e-01 9.81932878e-01 1.81893659e+00 -9.15399417e-02
-9.35343504e-02 -2.75452882e-01 -3.94405782e-01 9.56164062e-01
1.15066141e-01 -2.70516217e-01 -5.87633073e-01 -9.05014873e-01
5.52424669e-01 9.39564183e-02 8.39944720e-01 -9.47310090e-01
-1.37047738e-01 -4.76519644e-01 6.02347434e-01 -7.06347227e-01
8.21286976e-01 -7.87671804e-01 5.67085862e-01 8.78518000e-02
-1.40165940e-01 -2.97307462e-01 1.41605958e-01 7.92113721e-01
3.27867359e-01 3.92038763e-01 9.65747237e-01 -5.75538635e-01
-3.05547655e-01 6.59887016e-01 1.63507462e-01 3.17506701e-01
7.53659010e-01 -4.80462343e-01 3.48013826e-03 -3.07790309e-01
-4.07989889e-01 -1.90142989e-01 1.07739830e+00 -1.55103341e-01
1.03702235e+00 -1.52397633e+00 -5.07772744e-01 4.78463054e-01
-1.39722824e-01 8.83489609e-01 7.43961990e-01 2.35267371e-01
-1.13080275e+00 -5.30935153e-02 1.82326332e-01 -5.85063636e-01
-9.93042946e-01 -1.18409030e-01 6.70829833e-01 4.68476191e-02
-7.95910299e-01 1.29208004e+00 8.05481255e-01 -1.10848093e+00
-5.98951317e-02 -8.19315135e-01 6.11825347e-01 -6.08871162e-01
2.57963479e-01 7.40273833e-01 -1.71156991e-02 -5.47631681e-01
-2.65516996e-01 1.03219616e+00 5.93849182e-01 -3.91182333e-01
1.57019746e+00 -4.03866321e-01 -2.44768515e-01 4.99420941e-01
1.07367945e+00 5.12822807e-01 -1.89423990e+00 5.05809672e-03
-1.04848826e+00 -9.89143610e-01 3.64053398e-01 -8.85822952e-01
-1.06645691e+00 8.34028900e-01 3.29426706e-01 -3.91544223e-01
8.15762699e-01 -2.00534046e-01 1.07894027e+00 2.59270500e-02
8.23945761e-01 -8.12623441e-01 3.58882360e-02 4.76911217e-01
1.10660756e+00 -9.46646333e-01 3.03673953e-01 -8.40420663e-01
-1.49923056e-01 9.02818263e-01 2.38046065e-01 -3.41921955e-01
8.54966164e-01 4.49273258e-01 -3.82552952e-01 -5.23859322e-01
-7.20011368e-02 4.60295081e-01 5.41334510e-01 6.16392970e-01
3.16419959e-01 7.85911083e-02 5.47748923e-01 1.23053836e-02
-3.44203293e-01 1.54821919e-02 3.52272689e-01 9.17680979e-01
-1.12935998e-01 -6.58250928e-01 -4.71189648e-01 1.80737421e-01
7.90754631e-02 -3.72017056e-01 1.16470689e-02 1.62180603e-01
2.31566012e-01 6.72550917e-01 1.55869216e-01 9.10540968e-02
5.82107067e-01 -5.33652186e-01 1.11668468e+00 -6.95467532e-01
-4.60416168e-01 -4.30853963e-02 -1.53024374e-02 -1.03137600e+00
-5.41196346e-01 -5.65291405e-01 -1.46890664e+00 -5.16938090e-01
9.45961699e-02 -4.25937682e-01 8.80363107e-01 5.95317006e-01
3.71652842e-01 8.96583915e-01 7.58989990e-01 -1.34301114e+00
2.93338299e-01 -2.16596261e-01 -5.68103015e-01 3.01577181e-01
5.36278486e-01 -6.52468622e-01 -3.72592181e-01 2.18296662e-01] | [9.604225158691406, -3.1153600215911865] |
d7f91bd7-abde-4e06-8b7c-cd30ba1be3d8 | homogcl-rethinking-homophily-in-graph | 2306.09614 | null | https://arxiv.org/abs/2306.09614v1 | https://arxiv.org/pdf/2306.09614v1.pdf | HomoGCL: Rethinking Homophily in Graph Contrastive Learning | Contrastive learning (CL) has become the de-facto learning paradigm in self-supervised learning on graphs, which generally follows the "augmenting-contrasting" learning scheme. However, we observe that unlike CL in computer vision domain, CL in graph domain performs decently even without augmentation. We conduct a systematic analysis of this phenomenon and argue that homophily, i.e., the principle that "like attracts like", plays a key role in the success of graph CL. Inspired to leverage this property explicitly, we propose HomoGCL, a model-agnostic framework to expand the positive set using neighbor nodes with neighbor-specific significances. Theoretically, HomoGCL introduces a stricter lower bound of the mutual information between raw node features and node embeddings in augmented views. Furthermore, HomoGCL can be combined with existing graph CL models in a plug-and-play way with light extra computational overhead. Extensive experiments demonstrate that HomoGCL yields multiple state-of-the-art results across six public datasets and consistently brings notable performance improvements when applied to various graph CL methods. Code is avilable at https://github.com/wenzhilics/HomoGCL. | ['Jian-Huang Lai', 'Hui Xiong', 'Chang-Dong Wang', 'Wen-Zhi Li'] | 2023-06-16 | null | null | null | null | ['contrastive-learning', 'contrastive-learning'] | ['computer-vision', 'methodology'] | [-2.20209107e-01 4.26888555e-01 -6.35157108e-01 -2.31831849e-01
-1.55987427e-01 -6.23376608e-01 7.93422580e-01 4.14028138e-01
3.97841968e-02 5.30802965e-01 1.14881195e-01 -5.29556811e-01
-1.19576402e-01 -9.08035934e-01 -8.39441240e-01 -8.49507630e-01
-2.97494769e-01 2.13032410e-01 4.83956747e-02 -2.12600365e-01
-1.17945187e-01 1.96978286e-01 -1.01537776e+00 -1.52489571e-02
6.35459840e-01 4.78692204e-01 -1.45585045e-01 4.45982128e-01
6.31875843e-02 8.34465146e-01 2.96222344e-02 -7.43285775e-01
4.32421803e-01 -3.67218137e-01 -8.08419585e-01 4.52651456e-02
6.79601073e-01 7.48039410e-02 -6.55795634e-01 1.19813097e+00
2.76724368e-01 -2.27365419e-01 5.94566762e-01 -1.86528754e+00
-9.23517108e-01 9.48171735e-01 -9.17126119e-01 1.05371073e-01
2.41446018e-01 1.83040842e-01 1.70517719e+00 -9.16931927e-01
7.19078600e-01 1.24174547e+00 8.18824828e-01 3.20030957e-01
-1.39711022e+00 -7.13821411e-01 3.34637582e-01 1.27475619e-01
-1.31122708e+00 -7.81975687e-02 1.21786010e+00 -2.41685063e-01
4.99084413e-01 2.83664495e-01 7.86302626e-01 1.11966360e+00
5.30213229e-02 1.09422076e+00 9.89644289e-01 -4.53137577e-01
1.87992640e-02 3.32321584e-01 1.88038200e-01 1.15768921e+00
6.62765920e-01 4.36267555e-02 -5.80819190e-01 -2.32529253e-01
6.33864880e-01 5.90365939e-02 -2.30823442e-01 -8.98899436e-01
-9.48423564e-01 1.05003977e+00 1.02012038e+00 2.14840427e-01
2.22774353e-02 3.76994550e-01 4.91518468e-01 4.29801375e-01
7.66807854e-01 3.52278024e-01 -3.73114526e-01 3.53565931e-01
-4.14749682e-01 -1.72492281e-01 8.76489878e-01 8.76888454e-01
1.04879570e+00 -1.10461608e-01 2.52302915e-01 5.67733049e-01
3.69182527e-01 1.15566440e-01 1.42632321e-01 -6.30311370e-01
1.93475008e-01 1.01622581e+00 -6.69205606e-01 -1.26113963e+00
-2.88108438e-01 -9.63230073e-01 -1.07857609e+00 -6.95746616e-02
1.60723656e-01 -4.82025072e-02 -4.70126212e-01 1.92630315e+00
5.75660110e-01 5.71583927e-01 -4.30175290e-02 5.17318726e-01
1.12567234e+00 2.64703661e-01 1.41118979e-02 -6.27057925e-02
9.77479815e-01 -1.15516770e+00 -3.84952694e-01 -3.10895383e-01
1.06265092e+00 -2.85478711e-01 1.34801507e+00 2.85767987e-02
-5.82073927e-01 -1.80049628e-01 -9.96710062e-01 1.45452783e-01
-4.84748155e-01 -1.83949396e-01 1.16271830e+00 6.99365258e-01
-1.30139434e+00 5.76222062e-01 -5.21920621e-01 -5.85311294e-01
5.64966083e-01 1.14654213e-01 -6.97589219e-01 -5.41247725e-02
-9.17999983e-01 4.95084882e-01 3.68629038e-01 -2.03954220e-01
-5.34248233e-01 -8.86746466e-01 -1.07622874e+00 -4.36770171e-03
6.49426997e-01 -7.45519340e-01 5.78208625e-01 -8.88235629e-01
-8.86826992e-01 1.05418825e+00 2.07387879e-02 -5.34621239e-01
5.67887664e-01 -7.64593557e-02 -1.92719623e-01 5.50671946e-03
1.50176704e-01 6.25443518e-01 6.59013271e-01 -1.53581595e+00
-1.18598513e-01 -4.82375860e-01 1.61795631e-01 1.53825283e-01
-7.63534784e-01 -6.13676012e-01 -5.46111703e-01 -5.16408980e-01
1.51006415e-01 -1.09690332e+00 -1.97838292e-01 1.58965543e-01
-6.89967096e-01 -5.64723432e-01 1.02763724e+00 -1.08657211e-01
1.26984453e+00 -2.16511822e+00 2.82262340e-02 4.08862829e-01
9.85384166e-01 9.12209153e-02 -2.96433687e-01 6.90098822e-01
-2.17332244e-01 3.65368962e-01 -1.19686887e-01 -3.71627569e-01
-6.95022419e-02 1.71874702e-01 -2.41581835e-02 8.09685230e-01
7.43268570e-03 1.32333994e+00 -1.32125151e+00 -6.35299504e-01
2.48202845e-01 2.97504842e-01 -6.09960020e-01 -5.30990995e-02
1.31814862e-02 1.61721349e-01 -2.93776810e-01 7.39966452e-01
6.22826934e-01 -9.76952732e-01 6.58305168e-01 -3.06030214e-01
3.84219229e-01 1.62237510e-02 -8.37458670e-01 1.40657854e+00
-1.33405909e-01 6.57968462e-01 -7.46561736e-02 -1.02394450e+00
8.42162848e-01 -1.51823610e-01 3.03518683e-01 -4.26173359e-01
1.00898482e-02 8.75369981e-02 -1.40570905e-02 -1.31722629e-01
1.81100249e-01 1.78554520e-01 1.21709995e-01 4.66791809e-01
1.55516595e-01 1.01267606e-01 2.07728937e-01 9.44467545e-01
1.25262332e+00 3.13389003e-02 6.63570940e-01 -5.14620364e-01
3.46947968e-01 -1.44638449e-01 3.33129138e-01 7.17389703e-01
-2.38886133e-01 1.91807717e-01 7.91054666e-01 -3.72145742e-01
-6.75761104e-01 -9.87461805e-01 1.68141752e-01 1.20026445e+00
3.02853703e-01 -8.00635874e-01 -3.91504407e-01 -1.21573341e+00
3.26712161e-01 4.56832409e-01 -9.04058456e-01 -3.33389640e-01
-3.26387197e-01 -7.81222165e-01 3.24407279e-01 4.07630652e-01
5.59270322e-01 -7.27271616e-01 2.60448098e-01 -2.90074140e-01
1.48092896e-01 -1.12029684e+00 -3.59025091e-01 4.92444932e-02
-7.68791378e-01 -1.34816551e+00 -1.98042601e-01 -8.94803524e-01
9.13161397e-01 7.17652440e-01 1.44502437e+00 6.24829173e-01
-2.46163934e-01 5.98953605e-01 -4.44691300e-01 5.17190667e-03
-3.30662906e-01 3.58514190e-01 -3.09778098e-02 -4.05370668e-02
2.70336211e-01 -9.92180824e-01 -6.76500976e-01 -1.22078052e-02
-6.96910143e-01 1.62301242e-01 5.33759177e-01 8.84955049e-01
2.91256160e-01 -3.83790955e-02 6.36355162e-01 -1.50207007e+00
6.14401221e-01 -7.59723067e-01 -3.85567069e-01 2.85268426e-01
-1.01039326e+00 4.71884239e-04 6.44464850e-01 -2.46223718e-01
-4.49786216e-01 -7.20553249e-02 1.34959817e-01 -5.21859407e-01
2.98194587e-01 6.42635167e-01 -1.79109931e-01 -4.19710010e-01
5.71535766e-01 3.20468456e-01 7.63549656e-02 -1.74880773e-01
7.55675256e-01 1.74467891e-01 4.18270379e-01 -4.09206897e-01
1.21858239e+00 6.95113063e-01 2.50167131e-01 -8.34719598e-01
-8.27330410e-01 -5.65472662e-01 -5.95592380e-01 -3.52249831e-01
2.61716664e-01 -1.00768042e+00 -7.87241936e-01 1.44142643e-01
-6.15113795e-01 -4.28582639e-01 -1.66809574e-01 1.58370450e-01
-1.30891815e-01 6.48744643e-01 -4.82481420e-01 -5.76143146e-01
-3.11648041e-01 -5.87822855e-01 6.88186109e-01 1.50635734e-01
-6.81940690e-02 -1.67564726e+00 1.12352222e-01 2.57893264e-01
7.81993940e-02 3.88347328e-01 1.00610411e+00 -9.31540251e-01
-5.54537594e-01 -2.42098510e-01 -4.48424488e-01 2.36337528e-01
3.56330007e-01 -5.61633483e-02 -8.34810555e-01 -6.72694564e-01
-5.60972095e-01 -3.30286801e-01 1.00685966e+00 1.24726519e-01
1.02001834e+00 -4.97818440e-01 -6.01012349e-01 6.89383388e-01
1.67068720e+00 -5.72564900e-01 3.77783269e-01 1.25863375e-02
1.25623524e+00 4.41872835e-01 2.95757324e-01 2.77361751e-01
8.31935763e-01 2.48314783e-01 7.08889782e-01 -3.80460650e-01
-3.74719441e-01 -7.92829812e-01 2.37295315e-01 8.06917012e-01
2.73347408e-01 -3.64635557e-01 -9.54781115e-01 5.92238903e-01
-1.94563270e+00 -6.89428210e-01 -4.52577531e-01 1.92220724e+00
6.63277686e-01 1.36480868e-01 2.04430997e-01 -1.17463998e-01
7.76744902e-01 5.73625267e-01 -6.46037340e-01 8.61778706e-02
-3.26116383e-01 -1.39684498e-01 6.05913103e-01 5.49677253e-01
-1.11969507e+00 1.27292776e+00 5.68746185e+00 7.93311298e-01
-1.04888010e+00 1.19357534e-01 7.97040641e-01 2.15826645e-01
-7.41276860e-01 3.27788800e-01 -4.66347039e-01 2.41490692e-01
4.13144261e-01 -3.43864858e-01 2.50589848e-01 1.03079820e+00
-2.24199086e-01 2.38199517e-01 -1.25623977e+00 8.80512476e-01
1.24272659e-01 -1.64659059e+00 1.45261377e-01 2.11279154e-01
8.90771091e-01 2.75116056e-01 1.57785147e-01 3.00297797e-01
7.28823423e-01 -9.20841813e-01 1.76023766e-01 -1.20109521e-01
5.68524182e-01 -4.77800250e-01 5.13817132e-01 3.39619994e-01
-1.47311008e+00 1.85709804e-01 -2.51644999e-01 -7.98981339e-02
-3.18092644e-01 7.99971640e-01 -1.04158950e+00 7.00076282e-01
6.69903457e-01 1.29120135e+00 -1.00269639e+00 7.03346193e-01
-4.62772250e-01 7.94995189e-01 -1.47799984e-01 -1.61841497e-01
2.32084662e-01 -2.17642605e-01 5.89562833e-01 1.10672092e+00
-1.68994635e-01 -2.72978187e-01 2.81928331e-01 7.30793417e-01
-4.87871110e-01 2.56190121e-01 -1.02692497e+00 -2.34341457e-01
6.42148554e-01 1.62639916e+00 -8.57699931e-01 -3.14479649e-01
-5.15823603e-01 8.17884147e-01 6.79058790e-01 2.96172768e-01
-6.48335993e-01 1.75784305e-02 3.15792352e-01 2.99903005e-01
1.80129662e-01 -7.96745904e-03 -1.70698583e-01 -1.17634106e+00
-1.76107399e-02 -7.30028093e-01 5.62954605e-01 -4.33198333e-01
-1.78122830e+00 4.17001694e-01 -2.73548573e-01 -1.25831318e+00
5.21981083e-02 -4.85182285e-01 -8.20455968e-01 2.18434870e-01
-1.62279010e+00 -1.59600019e+00 -5.05185544e-01 6.16881847e-01
1.79108366e-01 -9.18693393e-02 5.99649191e-01 2.39137337e-02
-6.25989199e-01 8.12304199e-01 -6.97946474e-02 2.81254083e-01
6.65453136e-01 -1.47161949e+00 5.51642537e-01 7.61651397e-01
5.16698360e-01 5.62306643e-01 5.99933445e-01 -8.21118414e-01
-1.57092381e+00 -1.24633861e+00 6.95245624e-01 -4.27114427e-01
1.11224639e+00 -4.98965502e-01 -9.13797617e-01 9.68136728e-01
4.39372659e-01 4.93079543e-01 7.98353016e-01 4.66384709e-01
-9.06042993e-01 -1.34666368e-01 -9.80034530e-01 7.42690444e-01
1.48721147e+00 -5.88070095e-01 -1.13010280e-01 5.90127945e-01
8.99244487e-01 -2.38452610e-02 -7.56140590e-01 4.66000497e-01
3.30085009e-01 -1.13801646e+00 9.04057384e-01 -6.62199557e-01
4.77635652e-01 -9.18389708e-02 -9.45947543e-02 -1.36969674e+00
-3.39928031e-01 -7.50150502e-01 -3.13054144e-01 1.15149665e+00
5.09684384e-01 -1.09811056e+00 1.04648519e+00 1.20503761e-01
1.54417112e-01 -9.57285166e-01 -5.79130709e-01 -9.13007736e-01
1.43055335e-01 -8.79507139e-02 4.28602248e-01 1.42789721e+00
1.66256994e-01 5.43571174e-01 -3.52853596e-01 2.05492154e-01
1.05535829e+00 1.05881602e-01 1.03835416e+00 -1.54804063e+00
-2.71766067e-01 -4.12796140e-01 -4.66345608e-01 -8.34740341e-01
5.48524618e-01 -1.46992958e+00 -4.29455787e-01 -1.46902692e+00
6.37895942e-01 -5.94763339e-01 -2.28027314e-01 5.77508211e-01
-4.12921429e-01 3.69531512e-01 2.30850682e-01 3.10896486e-01
-8.04097176e-01 5.00231922e-01 1.18356371e+00 -1.01687022e-01
-7.95120373e-02 -1.13506950e-01 -1.04032910e+00 7.40092635e-01
9.31022286e-01 -3.40721905e-01 -6.64801776e-01 -7.25895166e-02
6.30916357e-01 -3.57825756e-01 5.92788100e-01 -5.94112813e-01
4.19463545e-01 6.04927838e-02 4.93013375e-02 -3.06605965e-01
6.66353181e-02 -6.43143415e-01 -6.40755445e-02 5.57518065e-01
-4.00777698e-01 9.21305865e-02 -1.73443943e-01 9.23651218e-01
9.87335015e-03 2.23611966e-01 5.37573576e-01 -1.87218696e-01
-7.34983981e-01 6.00311875e-01 3.44987005e-01 3.18511426e-01
9.21598554e-01 -6.39869645e-02 -6.21164322e-01 -5.34233630e-01
-3.41196805e-01 3.20939034e-01 5.41317999e-01 3.80033821e-01
4.42213893e-01 -1.35555017e+00 -8.23676527e-01 1.21346422e-01
4.27569419e-01 -1.97012886e-01 1.55379757e-01 9.53703642e-01
-2.18342200e-01 -8.82447418e-03 2.05286190e-01 -7.13844001e-01
-1.39622223e+00 5.21424055e-01 2.28744164e-01 -4.11934376e-01
-7.58020043e-01 1.05916011e+00 4.33739543e-01 -7.14262962e-01
9.76209715e-02 1.86562791e-01 1.78209711e-02 -9.29348767e-02
4.80726920e-03 2.91062683e-01 -2.25478396e-01 -3.76877308e-01
-5.80250025e-01 3.00759941e-01 -2.33412206e-01 3.28006178e-01
1.37499511e+00 -1.01356484e-01 -3.86606634e-01 3.80164415e-01
1.41128850e+00 2.80610412e-01 -1.08645141e+00 -5.43363929e-01
3.94338518e-02 -4.35052067e-01 1.32439956e-02 -4.67253655e-01
-1.40975952e+00 5.53016126e-01 1.94280520e-01 5.54721177e-01
8.58782470e-01 5.46787381e-01 3.40544134e-01 3.24870169e-01
3.27426463e-01 -6.60913646e-01 4.45608586e-01 8.55753273e-02
7.59944797e-01 -1.62447190e+00 3.91210645e-01 -8.99372399e-01
-6.60448611e-01 6.95079446e-01 8.26079309e-01 -3.90239626e-01
9.75990117e-01 1.47062048e-01 -1.70211032e-01 -4.57779914e-01
-9.53554809e-01 -2.19020978e-01 1.90559834e-01 5.72396100e-01
4.72952843e-01 3.90662342e-01 -1.93906188e-01 1.27685413e-01
-2.51130879e-01 -4.59861428e-01 4.65956599e-01 6.70336127e-01
-8.48006010e-02 -1.16796410e+00 2.51131058e-01 4.68695492e-01
-1.11728096e-02 -3.59744161e-01 -7.65081167e-01 1.21676719e+00
-1.92393735e-01 7.23451078e-01 -1.42003551e-01 -6.48725688e-01
-2.36923099e-01 -3.08336377e-01 3.37598681e-01 -5.38185716e-01
-4.96962458e-01 -1.72827840e-01 5.11240587e-02 -4.88988101e-01
-4.06599611e-01 -2.47760922e-01 -1.08503342e+00 -6.26262844e-01
-4.00761664e-01 6.52125850e-02 2.44292840e-01 7.15062141e-01
5.27191460e-01 2.59501308e-01 8.49880278e-01 -3.83638114e-01
-3.37026387e-01 -7.36387968e-01 -5.32958925e-01 3.77630055e-01
1.80058435e-01 -4.55433547e-01 -8.09171915e-01 -3.64129394e-01] | [7.163616180419922, 6.203781604766846] |
b1b29463-eb62-4c52-818e-dac6aa6ac4d4 | shifted-chunk-transformer-for-spatio-temporal | 2108.11575 | null | https://arxiv.org/abs/2108.11575v5 | https://arxiv.org/pdf/2108.11575v5.pdf | Shifted Chunk Transformer for Spatio-Temporal Representational Learning | Spatio-temporal representational learning has been widely adopted in various fields such as action recognition, video object segmentation, and action anticipation. Previous spatio-temporal representational learning approaches primarily employ ConvNets or sequential models,e.g., LSTM, to learn the intra-frame and inter-frame features. Recently, Transformer models have successfully dominated the study of natural language processing (NLP), image classification, etc. However, the pure-Transformer based spatio-temporal learning can be prohibitively costly on memory and computation to extract fine-grained features from a tiny patch. To tackle the training difficulty and enhance the spatio-temporal learning, we construct a shifted chunk Transformer with pure self-attention blocks. Leveraging the recent efficient Transformer design in NLP, this shifted chunk Transformer can learn hierarchical spatio-temporal features from a local tiny patch to a global video clip. Our shifted self-attention can also effectively model complicated inter-frame variances. Furthermore, we build a clip encoder based on Transformer to model long-term temporal dependencies. We conduct thorough ablation studies to validate each component and hyper-parameters in our shifted chunk Transformer, and it outperforms previous state-of-the-art approaches on Kinetics-400, Kinetics-600, UCF101, and HMDB51. | ['Ji Liu', 'Sen yang', 'Tingxun Lv', 'Wentao Zhu', 'Xuefan Zha'] | 2021-08-26 | null | http://proceedings.neurips.cc/paper/2021/hash/5edc4f7dce28c711afc6265b4f99bf57-Abstract.html | http://proceedings.neurips.cc/paper/2021/file/5edc4f7dce28c711afc6265b4f99bf57-Paper.pdf | neurips-2021-12 | ['action-anticipation'] | ['computer-vision'] | [ 9.93682742e-02 -5.03362477e-01 -5.29056609e-01 -3.46162409e-01
-8.94935608e-01 -1.86040953e-01 5.55939257e-01 -8.06656629e-02
-5.07803738e-01 5.51519990e-01 4.21778053e-01 -2.69340375e-03
-1.99936509e-01 -4.93921071e-01 -9.05394912e-01 -7.84614205e-01
-2.23856121e-01 -1.36895860e-02 6.63155258e-01 6.78408816e-02
4.18044999e-02 1.92024082e-01 -1.46818852e+00 8.97757947e-01
5.12431324e-01 1.31840801e+00 2.15498850e-01 6.57759249e-01
-1.69078737e-01 1.43502700e+00 -4.01173532e-01 1.44356443e-02
-6.12664111e-02 -4.65340972e-01 -8.01038623e-01 4.03290465e-02
4.54997979e-02 -3.38369191e-01 -9.02200818e-01 5.20913601e-01
4.38031822e-01 2.86892384e-01 3.87686282e-01 -1.26437533e+00
-7.67147958e-01 4.23664540e-01 -7.36838877e-01 6.97492599e-01
2.95859635e-01 4.55259591e-01 7.89838016e-01 -7.31326759e-01
5.68528891e-01 1.25426126e+00 6.59078598e-01 4.01835769e-01
-8.53552222e-01 -7.25180268e-01 5.50587237e-01 8.31133068e-01
-1.36321843e+00 -3.25178027e-01 5.87897658e-01 -3.51605892e-01
1.27734375e+00 -1.73634440e-02 8.78205538e-01 1.32147682e+00
4.81796563e-01 1.40599477e+00 8.86481166e-01 -2.13251710e-02
2.00076471e-03 -6.99909151e-01 -2.14630499e-01 5.87974787e-01
-7.18607664e-01 1.32717550e-01 -8.57974708e-01 2.59551048e-01
9.47011411e-01 4.00038421e-01 -9.29276347e-02 -1.13326348e-01
-1.26221430e+00 6.33503318e-01 5.64732015e-01 5.13325870e-01
-4.86168593e-01 6.00117981e-01 9.07281280e-01 1.58417419e-01
6.28009617e-01 -1.23591892e-01 -7.23390877e-01 -6.24491274e-01
-8.66115987e-01 8.73232186e-02 9.72873196e-02 9.40433443e-01
4.05470520e-01 -1.91061955e-03 -6.39172018e-01 7.76428998e-01
5.80108911e-02 1.84127122e-01 9.42082286e-01 -8.36407781e-01
5.75746417e-01 6.44506335e-01 -1.63888693e-01 -7.18997061e-01
-4.45424855e-01 -1.75595790e-01 -1.02289450e+00 -3.32713515e-01
1.36560321e-01 1.83997259e-01 -1.18366444e+00 1.68054223e+00
2.92407990e-01 6.93833888e-01 -2.20357671e-01 9.09823239e-01
6.74132228e-01 9.97060478e-01 4.65605974e-01 -3.77925515e-01
1.21666920e+00 -1.26140237e+00 -6.40376210e-01 -8.58530961e-03
7.24982858e-01 -4.50388342e-01 1.05446351e+00 1.93387508e-01
-9.93532062e-01 -6.99740231e-01 -5.92816710e-01 -4.51997310e-01
-2.65578568e-01 2.19734311e-01 6.53411269e-01 -1.68263331e-01
-6.87299073e-01 6.19653583e-01 -1.22053313e+00 -2.05536932e-01
7.60627389e-01 3.49838465e-01 -3.67384583e-01 -1.72300860e-01
-1.32783699e+00 4.85191375e-01 3.92383754e-01 1.33037299e-01
-1.19514251e+00 -7.86037922e-01 -9.38763618e-01 1.39302850e-01
4.60320473e-01 -5.54961920e-01 1.28327775e+00 -9.43789840e-01
-1.68162751e+00 5.59354782e-01 -4.40100521e-01 -7.81773388e-01
4.28061038e-01 -3.44533592e-01 -3.25337797e-01 2.24770591e-01
7.78343976e-02 8.29552054e-01 9.99435782e-01 -4.10921991e-01
-7.97369957e-01 -2.34112263e-01 9.42464620e-02 2.02338174e-01
-3.03035676e-01 1.88997239e-01 -7.64857113e-01 -9.53801036e-01
-1.02817871e-01 -7.70897865e-01 -1.92182645e-01 -8.72386098e-02
-1.34426013e-01 -5.99166989e-01 9.50869024e-01 -5.07697463e-01
1.39832187e+00 -2.30514383e+00 1.53103888e-01 -3.93365115e-01
5.89461811e-02 4.07239586e-01 -2.46434718e-01 3.20767671e-01
-1.65070549e-01 1.22355847e-02 2.39151400e-02 -2.28484929e-01
-6.78668395e-02 2.48844326e-01 -4.02031600e-01 4.60785747e-01
2.07161710e-01 1.29626679e+00 -8.74799967e-01 -6.70286536e-01
3.84080380e-01 7.00767517e-01 -4.34697777e-01 1.50892466e-01
-4.38630968e-01 5.04292727e-01 -7.18077421e-01 5.10913551e-01
1.87736377e-01 -4.63485330e-01 -1.88076258e-01 -2.26970449e-01
-2.62199134e-01 3.78176242e-01 -4.82811928e-01 2.17996526e+00
-5.23532927e-01 6.12526298e-01 -5.56148231e-01 -1.18782222e+00
5.56640923e-01 4.00640815e-01 9.29685354e-01 -1.26527619e+00
-2.04352699e-02 -1.33451760e-01 -2.01258704e-01 -7.87086904e-01
2.02354461e-01 -7.21772090e-02 2.42367480e-02 3.80979627e-01
8.13012943e-02 5.56405663e-01 2.48790741e-01 1.32443774e-02
1.38102937e+00 6.09076798e-01 1.61124274e-01 1.72698304e-01
6.07300878e-01 -1.88715532e-01 8.92186999e-01 2.98636198e-01
-4.13278431e-01 5.99719822e-01 4.59461540e-01 -8.54751110e-01
-8.94773722e-01 -7.73601413e-01 2.35464066e-01 1.55587077e+00
1.25086876e-02 -6.54646873e-01 -4.49327022e-01 -7.40645170e-01
-1.77542970e-01 2.41840228e-01 -8.11486363e-01 -4.03429449e-01
-8.11890841e-01 -6.88010693e-01 5.13672709e-01 9.77219045e-01
7.82075047e-01 -1.35020089e+00 -7.98268497e-01 4.29954976e-01
-2.91316748e-01 -1.20815659e+00 -7.30488122e-01 9.73272100e-02
-7.93135881e-01 -9.54436481e-01 -7.85707474e-01 -6.61230087e-01
2.27137282e-01 2.70319015e-01 9.15985465e-01 -1.14099719e-01
-3.54167491e-01 2.23212481e-01 -7.01844454e-01 -6.15238175e-02
3.34513813e-01 1.98417589e-01 -1.79240212e-01 1.05406806e-01
3.79081190e-01 -7.08253920e-01 -8.10659647e-01 3.82116824e-01
-1.01635861e+00 2.56708920e-01 5.71023226e-01 9.11742866e-01
8.70474577e-01 1.67846322e-01 4.41901594e-01 -6.10604823e-01
2.28335544e-01 -4.54103261e-01 -3.08543354e-01 3.57247293e-01
-1.25468876e-02 3.92826460e-02 7.72192597e-01 -7.44375408e-01
-1.04022920e+00 9.65972096e-02 -1.17080748e-01 -9.76671875e-01
-9.74528259e-04 7.34828472e-01 -4.29391712e-02 2.18667403e-01
3.14617068e-01 7.06315160e-01 -3.35533828e-01 -5.04052997e-01
2.70130426e-01 2.97162503e-01 4.12094057e-01 -4.72487718e-01
2.92750776e-01 3.96516234e-01 -1.92390919e-01 -5.27258873e-01
-1.13140213e+00 -4.97371972e-01 -6.63585544e-01 7.25175440e-03
1.05099607e+00 -1.09269202e+00 -8.34380805e-01 6.83701515e-01
-1.19199812e+00 -7.79285610e-01 -3.70537639e-01 5.04657924e-01
-8.37224007e-01 1.42401889e-01 -8.95655394e-01 -4.91897255e-01
-1.18479684e-01 -1.05944264e+00 1.38561070e+00 9.56242606e-02
5.15456796e-02 -8.35554242e-01 8.15472454e-02 3.19888622e-01
2.30974525e-01 1.96802437e-01 7.50987589e-01 -1.84788749e-01
-6.37452126e-01 -1.54558038e-02 -2.77522266e-01 1.54755637e-01
1.39962986e-01 -2.02537049e-02 -7.38436997e-01 -1.71659946e-01
-2.10928261e-01 -5.15803754e-01 1.18880451e+00 6.26379371e-01
1.73996902e+00 -2.97756612e-01 -4.20021236e-01 7.81606138e-01
1.00528586e+00 3.38242501e-01 8.71278882e-01 3.30921233e-01
8.67563665e-01 3.04689258e-01 8.57183039e-01 6.20567441e-01
4.75994945e-01 8.17105830e-01 1.78128347e-01 5.87422326e-02
2.28640232e-02 -3.06300908e-01 5.40208995e-01 7.52785683e-01
-1.58658400e-01 -2.64028519e-01 -8.39939773e-01 5.63010335e-01
-2.27525091e+00 -1.15931225e+00 3.54894370e-01 1.77353728e+00
7.75638521e-01 7.90991411e-02 2.29009882e-01 8.36938340e-03
5.11698484e-01 4.98735756e-01 -7.35681653e-01 2.10998673e-02
-3.42464671e-02 2.55275220e-02 2.93692976e-01 1.10020868e-01
-1.37797403e+00 1.26923990e+00 5.71260309e+00 1.24931228e+00
-1.39698267e+00 3.32863152e-01 8.26149106e-01 -4.38793778e-01
6.78541288e-02 -2.08937526e-01 -5.80415905e-01 7.87648380e-01
1.02668643e+00 -1.37600824e-01 1.90963492e-01 7.62305021e-01
4.88132060e-01 -2.79124044e-02 -1.08416629e+00 1.24722147e+00
-5.25023043e-02 -1.47573125e+00 6.85715303e-02 -7.30227828e-02
5.93461990e-01 1.54693305e-01 1.32660761e-01 6.93522036e-01
5.42616881e-02 -1.30332243e+00 5.64696312e-01 6.08107448e-01
8.59654069e-01 -7.56349444e-01 6.03380978e-01 3.45603108e-01
-1.66239524e+00 -3.08210522e-01 -2.85807341e-01 -7.82350078e-02
3.11537176e-01 3.76865685e-01 -1.53314039e-01 4.92240667e-01
9.51902747e-01 1.32812703e+00 -3.97573978e-01 8.77789974e-01
-1.30081654e-01 6.18792593e-01 -2.26145372e-01 1.88842595e-01
6.01958036e-01 1.22487783e-01 2.61379667e-02 1.07402670e+00
2.37632275e-01 5.05206645e-01 3.21071923e-01 4.34009790e-01
-3.29552405e-02 -7.29785636e-02 -2.31880009e-01 -2.98195690e-01
1.10404722e-01 9.24717724e-01 -5.42101085e-01 -4.74033028e-01
-5.72902381e-01 1.06348455e+00 4.42920208e-01 4.37154740e-01
-1.31443202e+00 -3.91282663e-02 7.44822681e-01 1.08956702e-01
6.29453182e-01 -3.47945094e-01 1.57681063e-01 -1.25507545e+00
1.41860709e-01 -8.86648655e-01 6.00872219e-01 -8.03684115e-01
-1.08301091e+00 6.51179612e-01 2.56019756e-02 -1.38735628e+00
-1.93094432e-01 -4.11113054e-01 -5.98921835e-01 4.14364547e-01
-1.42139959e+00 -1.41390705e+00 -1.45347118e-01 9.62111533e-01
8.98921311e-01 -1.65745132e-02 4.62654650e-01 4.74574447e-01
-9.07809556e-01 4.20306176e-01 -6.19494952e-02 1.47056967e-01
6.07154965e-01 -9.06014442e-01 3.53947341e-01 6.69084549e-01
1.29929066e-01 4.00126189e-01 1.94586009e-01 -4.75963145e-01
-1.24998343e+00 -1.46300137e+00 7.25173235e-01 -2.30356768e-01
7.52112269e-01 -2.67362267e-01 -8.87646496e-01 8.75855923e-01
6.73192143e-02 7.12319374e-01 4.18510139e-01 -5.69436587e-02
-3.72106254e-01 -2.87646651e-01 -5.51405609e-01 4.68253195e-01
1.34343922e+00 -7.54486442e-01 -3.56972694e-01 4.60856050e-01
9.77880359e-01 -4.91590530e-01 -9.02688026e-01 4.60381716e-01
5.98335207e-01 -9.24874485e-01 1.00533426e+00 -9.66178894e-01
7.00426519e-01 -2.52732813e-01 3.75305414e-02 -9.66159642e-01
-6.15766525e-01 -7.08294988e-01 -4.12506163e-01 1.11584687e+00
1.28417341e-02 -1.98234469e-01 9.73591208e-01 3.13365936e-01
-3.16318989e-01 -1.26709712e+00 -1.11251140e+00 -8.29279780e-01
1.33273785e-03 -6.30147874e-01 5.06032288e-01 6.26787066e-01
-8.90153199e-02 4.46135670e-01 -7.03185439e-01 -2.61029571e-01
1.02168277e-01 3.05147588e-01 6.10474706e-01 -6.70556128e-01
-3.11328828e-01 -3.32870126e-01 -5.36092520e-01 -1.56190872e+00
2.66933620e-01 -4.49072421e-01 1.53058842e-01 -1.33471322e+00
3.83616596e-01 -3.12875956e-01 -6.69569671e-01 7.72722065e-01
-1.77134261e-01 2.14773059e-01 1.52339712e-01 2.25004256e-01
-1.24064541e+00 1.09151709e+00 1.51874435e+00 -2.89956093e-01
-1.25947773e-01 -2.07091659e-01 -1.50078386e-01 6.22588634e-01
5.80669343e-01 -5.43054461e-01 -6.43908322e-01 -5.88708520e-01
1.17500067e-01 2.08214656e-01 4.24213290e-01 -9.90401983e-01
4.49889839e-01 -4.29208785e-01 5.04325211e-01 -6.76647425e-01
4.94358897e-01 -7.05522954e-01 -1.18360579e-01 3.30718279e-01
-4.53076750e-01 8.68492052e-02 1.63273215e-01 7.26901770e-01
-6.25248134e-01 4.37439173e-01 5.08903742e-01 -2.19971463e-01
-9.97794986e-01 9.86464858e-01 -4.54156071e-01 1.50838614e-01
1.20854890e+00 -1.42148420e-01 -2.76947647e-01 -2.74721473e-01
-7.84570098e-01 3.41784835e-01 1.29943848e-01 5.70840538e-01
6.89806044e-01 -1.53956938e+00 -4.53791171e-01 9.19857714e-03
5.53867780e-02 1.61140442e-01 8.47896039e-01 1.29727626e+00
-3.13874602e-01 6.58563018e-01 -3.12198967e-01 -8.27192605e-01
-1.18109739e+00 8.08467090e-01 2.03511581e-01 -6.10523760e-01
-8.23723316e-01 1.01180851e+00 6.62848115e-01 1.30937099e-01
1.96816191e-01 -5.34365058e-01 -1.29436523e-01 -8.81031901e-02
6.58705950e-01 1.44902214e-01 -3.35102439e-01 -6.01780355e-01
-4.78538990e-01 6.46763682e-01 -2.33006313e-01 1.32264569e-01
1.45656013e+00 -1.65922090e-01 6.22650497e-02 5.72987735e-01
1.36587775e+00 -5.95423400e-01 -1.76742864e+00 -3.64004940e-01
-1.20984994e-01 -4.60685879e-01 -1.07748422e-03 -3.79255265e-01
-1.27771902e+00 1.10003340e+00 4.27816242e-01 -8.54706392e-02
1.43233299e+00 -8.39614645e-02 1.20465159e+00 2.01827168e-01
2.98299283e-01 -1.08091342e+00 4.61280912e-01 7.67955303e-01
8.50257635e-01 -1.14987326e+00 -8.05975795e-02 -2.65744388e-01
-7.08952248e-01 9.98417556e-01 7.32307196e-01 -1.13396131e-01
6.99717462e-01 1.69766620e-01 -1.93242788e-01 -1.24868572e-01
-1.32602239e+00 -2.52605379e-01 2.86750138e-01 2.98275113e-01
5.78142583e-01 -2.09248692e-01 -9.71347243e-02 6.48671806e-01
3.15664858e-01 4.53384012e-01 -3.79266813e-02 1.03486919e+00
-1.32844567e-01 -1.02170014e+00 1.76959664e-01 3.97459537e-01
-4.40870315e-01 -1.19281679e-01 -3.64025123e-02 7.03533709e-01
3.65262091e-01 5.26204526e-01 2.31781587e-01 -5.19208133e-01
1.24567851e-01 -2.98981369e-02 5.05609930e-01 -4.37729061e-01
-4.88950580e-01 3.22236210e-01 -2.10338011e-01 -1.11926043e+00
-7.46460736e-01 -7.36523211e-01 -1.36368084e+00 -1.88963115e-01
5.96626066e-02 -3.65276560e-02 2.74631903e-02 1.25666177e+00
5.95361650e-01 9.02795494e-01 4.40992981e-01 -8.27048302e-01
-1.14669390e-01 -9.89961028e-01 -3.49584430e-01 3.79299879e-01
2.87118882e-01 -7.43024647e-01 1.29058659e-01 3.00243676e-01] | [8.688278198242188, 0.4486660659313202] |
1464e741-9b7f-46b1-981c-edfbbc9d6b5d | data-refinement-for-fully-unsupervised-visual | 2202.12759 | null | https://arxiv.org/abs/2202.12759v1 | https://arxiv.org/pdf/2202.12759v1.pdf | Data refinement for fully unsupervised visual inspection using pre-trained networks | Anomaly detection has recently seen great progress in the field of visual inspection. More specifically, the use of classical outlier detection techniques on features extracted by deep pre-trained neural networks have been shown to deliver remarkable performances on the MVTec Anomaly Detection (MVTec AD) dataset. However, like most other anomaly detection strategies, these pre-trained methods assume all training data to be normal. As a consequence, they cannot be considered as fully unsupervised. There exists to our knowledge no work studying these pre-trained methods under fully unsupervised setting. In this work, we first assess the robustness of these pre-trained methods to fully unsupervised context, using polluted training sets (i.e. containing defective samples), and show that these methods are more robust to pollution compared to methods such as CutPaste. We then propose SROC, a Simple Refinement strategy for One Class classification. SROC enables to remove most of the polluted images from the training set, and to recover some of the lost AUC. We further show that our simple heuristic competes with, and even outperforms much more complex strategies from the existing literature. | ['Pierre Gutierrez', 'Benjamin Missaoui', 'Antoine Cordier'] | 2022-02-25 | null | null | null | null | ['one-class-classification'] | ['miscellaneous'] | [ 3.52606565e-01 8.16078708e-02 4.83574510e-01 -1.66205894e-02
-4.69253749e-01 -5.36337614e-01 6.46937549e-01 6.08664036e-01
-4.17463243e-01 3.41212511e-01 -2.61886597e-01 -2.69982159e-01
-4.13755298e-01 -7.25065649e-01 -8.05633664e-01 -1.04012167e+00
-1.82211846e-01 2.78553426e-01 5.85566044e-01 -1.03886768e-01
2.62425065e-01 7.41909564e-01 -2.06794739e+00 2.31811851e-01
1.00639892e+00 1.03275383e+00 -3.37447703e-01 7.45926380e-01
1.90501720e-01 6.04242206e-01 -7.89844453e-01 -2.65155673e-01
4.30860996e-01 -3.88704836e-01 -7.26696968e-01 3.09826732e-01
6.46348834e-01 -3.22260916e-01 -1.82998762e-01 1.14355934e+00
4.25756156e-01 2.24659190e-01 7.47593820e-01 -1.12749887e+00
-4.01451856e-01 2.76076376e-01 -5.35538077e-01 4.10701722e-01
1.44185469e-01 1.85399070e-01 8.37449551e-01 -8.82949233e-01
5.12398779e-01 8.01538646e-01 7.36697972e-01 3.15841168e-01
-1.19660485e+00 7.76371919e-03 1.45950794e-01 4.69391048e-01
-1.02087760e+00 -3.14450204e-01 7.10310876e-01 -4.02899027e-01
9.47866142e-01 2.18745321e-01 4.92868811e-01 1.14820457e+00
1.68522820e-01 8.61961663e-01 1.02527857e+00 -6.50692523e-01
3.72856081e-01 -2.42035892e-02 2.36688673e-01 5.87045789e-01
7.52882361e-01 2.11074129e-02 -2.95604736e-01 -2.19959036e-01
8.37099776e-02 1.30853325e-01 -3.56761485e-01 -6.96251035e-01
-6.44026577e-01 7.87101865e-01 2.51670718e-01 4.15511191e-01
-5.05209625e-01 -8.75515416e-02 6.70796156e-01 6.32068455e-01
6.28043711e-01 5.91821194e-01 -4.43434238e-01 1.29238233e-01
-1.00260282e+00 1.62627026e-01 6.21624947e-01 5.22032619e-01
6.86382830e-01 1.60920292e-01 -1.68837219e-01 7.54530728e-01
1.92931861e-01 4.47407424e-01 2.45181337e-01 -8.72308373e-01
1.81961939e-01 7.13505685e-01 9.42306966e-02 -1.06242812e+00
-4.71497834e-01 -5.94833195e-01 -7.99721420e-01 6.06486797e-01
7.87848592e-01 8.81245956e-02 -1.11021245e+00 1.19844830e+00
2.42728427e-01 8.18058252e-02 1.55208826e-01 5.43623805e-01
1.88534871e-01 3.26965034e-01 -3.78033459e-01 -1.56375363e-01
8.45708668e-01 -6.66848600e-01 -5.86544991e-01 8.38347536e-04
7.42797256e-01 -4.43836331e-01 9.50040579e-01 1.15137124e+00
-7.24833846e-01 -2.19261616e-01 -1.15790546e+00 3.54933262e-01
-7.10224330e-01 2.41510086e-02 4.14608032e-01 8.34738493e-01
-1.04417109e+00 1.04400706e+00 -1.09676719e+00 -5.92716038e-01
7.04061091e-01 1.97125986e-01 -5.43353081e-01 -3.92732531e-01
-8.23933542e-01 7.30918646e-01 3.58834565e-01 3.82799536e-01
-1.08438313e+00 -4.28494096e-01 -8.88318419e-01 4.23046872e-02
7.69914389e-01 -3.24662365e-02 1.01140928e+00 -1.02922821e+00
-1.05057859e+00 8.08468640e-01 2.43833121e-02 -8.49971592e-01
5.90277493e-01 -6.49183154e-01 -4.37905788e-01 3.62614691e-01
-3.11348382e-02 6.85050711e-02 9.67268109e-01 -1.43440402e+00
-7.25612938e-01 -3.71518433e-01 -5.04015312e-02 -3.83377105e-01
-5.10780454e-01 2.10935846e-02 -1.86327174e-01 -4.56242055e-01
7.10700974e-02 -6.24247074e-01 -2.62314290e-01 7.44479373e-02
-5.94198108e-01 -2.42923394e-01 9.69417751e-01 -5.81390917e-01
1.06203008e+00 -2.30016327e+00 -2.24263072e-01 7.01609135e-01
2.56433815e-01 6.77417397e-01 -1.28218025e-01 5.35052240e-01
-2.03840435e-01 -7.41821378e-02 -8.87120247e-01 -3.36200297e-01
1.86947361e-01 4.98331428e-01 -2.80204773e-01 9.71748590e-01
6.08421087e-01 4.38019872e-01 -9.92909610e-01 -2.30808005e-01
4.31655198e-01 1.55768901e-01 -4.44144756e-01 1.43784434e-01
-1.21913195e-01 1.63451776e-01 -1.23976171e-02 7.81277418e-01
8.68402600e-01 2.58246779e-01 -2.30133384e-01 1.16411828e-01
-7.00911507e-02 -2.12311655e-01 -1.09923863e+00 1.33618653e+00
-5.55890203e-02 7.27579415e-01 -9.30262506e-02 -1.51439619e+00
7.38953590e-01 2.01097578e-01 4.71664250e-01 -7.59420276e-01
-1.21818352e-02 4.76029187e-01 1.66963011e-01 -7.61948049e-01
3.68198514e-01 2.39778340e-01 2.22414851e-01 2.30032548e-01
3.19972813e-01 3.06522936e-01 5.66871762e-01 1.03673331e-01
1.50684464e+00 1.54248968e-01 2.21069828e-01 -2.29873270e-01
5.53770304e-01 3.33862454e-02 4.63818967e-01 1.17221594e+00
-3.40138674e-01 8.97498667e-01 8.56650591e-01 -2.74683356e-01
-1.07538235e+00 -1.07721782e+00 -2.15801224e-01 7.45682657e-01
-9.33186486e-02 -1.75352454e-01 -8.63233507e-01 -9.96518552e-01
4.74156253e-03 5.72218776e-01 -7.90303767e-01 -3.16690475e-01
-5.08933842e-01 -1.11117470e+00 6.69094145e-01 4.96188641e-01
3.34794104e-01 -1.21145415e+00 -6.81981742e-01 1.68616995e-01
3.06615055e-01 -9.21081901e-01 3.42188835e-01 5.11657119e-01
-7.54341364e-01 -1.52646160e+00 -4.25166547e-01 -2.63628334e-01
7.11125374e-01 -6.36365116e-02 1.02852929e+00 4.70879644e-01
-5.22136748e-01 4.54386264e-01 -7.23465741e-01 -5.17535269e-01
-5.05859137e-01 -1.08646326e-01 5.59007339e-02 3.51225108e-01
5.09418011e-01 -5.47007382e-01 -4.31193590e-01 1.51215225e-01
-1.40233159e+00 -8.58285666e-01 7.20580697e-01 9.02949333e-01
5.70348799e-01 3.80066574e-01 4.35474783e-01 -1.12293851e+00
3.26804429e-01 -5.59769392e-01 -5.90870380e-01 5.64329773e-02
-7.43947506e-01 -2.97536757e-02 7.91104019e-01 -1.13023870e-01
-8.43616784e-01 5.35101667e-02 -4.62807983e-01 -4.07441437e-01
-7.83347189e-01 3.63307983e-01 -2.02429742e-01 -2.14504600e-02
9.03094769e-01 1.24961175e-01 -1.13979027e-01 -6.94660366e-01
7.03778639e-02 4.86405402e-01 8.79936278e-01 -3.89520913e-01
1.13521433e+00 7.11470127e-01 1.61124602e-01 -1.01468360e+00
-9.08843338e-01 -7.72611380e-01 -9.25060213e-01 -9.15740877e-02
7.09683359e-01 -3.94470274e-01 -2.30092704e-01 8.50441694e-01
-8.26714754e-01 -3.59205753e-01 -6.99966013e-01 2.27241710e-01
-3.39333713e-01 8.61157537e-01 -3.87267768e-01 -1.12200129e+00
-2.45244596e-02 -7.84353733e-01 9.62627411e-01 5.15115857e-02
1.54001251e-01 -1.04444480e+00 2.38461033e-01 -1.09582178e-01
3.38300467e-01 7.36435294e-01 6.83346689e-01 -1.24397910e+00
-2.57734179e-01 -5.04103124e-01 2.67651659e-02 9.58854854e-01
1.07771575e-01 3.95361960e-01 -1.30424309e+00 -4.60164875e-01
4.83854301e-02 -6.91991940e-04 1.24238658e+00 1.06890582e-01
1.28311360e+00 -9.67339426e-02 -1.96358085e-01 4.94484544e-01
1.51082623e+00 3.74397868e-03 9.82364118e-01 7.20197082e-01
6.09251440e-01 6.90626144e-01 5.94650209e-01 3.47350866e-01
-2.49068066e-01 3.83127779e-01 1.07613444e+00 -2.37503558e-01
1.36079714e-01 2.24684104e-01 4.50345606e-01 3.67835581e-01
-2.03518674e-01 -5.13172865e-01 -9.88679349e-01 9.09013689e-01
-1.81224871e+00 -1.06841350e+00 -5.63569307e-01 2.21183825e+00
1.64571583e-01 3.14239979e-01 1.23310089e-01 8.52320850e-01
5.37391782e-01 -7.85917267e-02 -4.75419581e-01 -4.00539696e-01
-4.16621298e-01 4.26918745e-01 4.01368231e-01 -9.50060959e-04
-1.50494182e+00 4.12998855e-01 5.92313194e+00 5.80180407e-01
-7.00315773e-01 5.89041039e-02 3.18460494e-01 4.68392223e-02
4.97334711e-02 -1.02239139e-01 -2.80216366e-01 4.17567581e-01
9.82324481e-01 4.95755762e-01 -3.60053517e-02 8.72253478e-01
6.31132349e-02 -3.22129786e-01 -1.04948735e+00 4.66224432e-01
3.47409815e-01 -6.78801894e-01 -5.13651259e-02 1.56825319e-01
8.11990440e-01 9.06167701e-02 -7.45249614e-02 2.29415998e-01
1.23108812e-01 -9.24678564e-01 5.47526598e-01 7.36082017e-01
1.28704548e-01 -9.17388380e-01 1.38349497e+00 1.35470003e-01
-6.94637001e-01 -4.85466510e-01 -6.16765141e-01 1.56449482e-01
-1.09978639e-01 1.08053780e+00 -3.70387971e-01 1.02773988e+00
1.14465058e+00 7.10766315e-01 -1.14006412e+00 1.54823327e+00
-2.21948519e-01 1.03165567e+00 -5.11241317e-01 5.17967165e-01
4.59650755e-01 -7.71196187e-02 7.57950306e-01 1.18980873e+00
3.38576823e-01 -4.97272193e-01 -5.66225266e-03 5.08210242e-01
1.81759611e-01 1.33276805e-01 -9.34448600e-01 1.64119497e-01
-1.73188105e-01 1.00251234e+00 -8.72312009e-01 -2.15704903e-01
-5.66058040e-01 9.60222006e-01 1.59768671e-01 2.96550155e-01
-4.65765536e-01 -6.70003414e-01 5.46297967e-01 -1.69222392e-02
5.99453390e-01 6.67123720e-02 8.90676689e-04 -1.14484096e+00
3.61803800e-01 -8.66179585e-01 6.30255818e-01 -4.71443474e-01
-1.47845054e+00 5.26273131e-01 2.58271154e-02 -1.32934570e+00
-2.36863777e-01 -1.11418462e+00 -8.33364546e-01 3.25558960e-01
-1.54511607e+00 -6.75343454e-01 -3.44508916e-01 4.94169503e-01
2.71545500e-01 -1.61349624e-01 6.01488352e-01 2.86332458e-01
-7.56093144e-01 4.94484276e-01 5.63014567e-01 3.37024063e-01
5.74566185e-01 -1.62551117e+00 2.76700109e-01 1.71634877e+00
2.96711892e-01 2.75146693e-01 8.95992994e-01 -4.84038651e-01
-9.00337517e-01 -1.16222739e+00 3.69697243e-01 -7.38503397e-01
6.41147316e-01 -3.98663104e-01 -1.58988047e+00 4.22036052e-01
2.44230434e-01 2.84317255e-01 3.68571222e-01 -9.51445326e-02
-2.11198837e-01 -1.53355747e-01 -1.11558914e+00 3.30374539e-01
8.95844877e-01 -2.77895749e-01 -6.49523258e-01 1.79787934e-01
3.58738631e-01 -1.34886086e-01 -6.19880795e-01 5.64565778e-01
7.48056248e-02 -1.39548695e+00 8.13838184e-01 -5.58294713e-01
2.86023706e-01 -5.42024612e-01 -4.93711755e-02 -1.27228701e+00
7.40146488e-02 -4.69488919e-01 -3.98326963e-01 1.33717465e+00
2.75215328e-01 -7.97950566e-01 5.21842062e-01 -9.12403166e-02
-6.12181425e-01 -5.49649239e-01 -9.29598391e-01 -1.06835377e+00
4.82154600e-02 -6.35207951e-01 5.23029208e-01 7.83801496e-01
-4.37269151e-01 -3.68540406e-01 -3.52170557e-01 5.09490192e-01
6.36214077e-01 -1.20438583e-01 6.34874344e-01 -1.67960119e+00
-1.35393232e-01 -4.42182273e-01 -7.89174080e-01 -3.05331260e-01
-1.17508911e-01 -6.74049914e-01 3.79319727e-01 -1.43598211e+00
-1.85003355e-01 -1.41050473e-01 -5.47611535e-01 7.20565975e-01
-2.75398016e-01 6.08067751e-01 -3.97269040e-01 -4.40578386e-02
-7.17142045e-01 4.59456444e-01 6.22554362e-01 -1.44558817e-01
-4.99276184e-02 3.13821025e-02 -4.30758387e-01 8.69892657e-01
9.08551216e-01 -6.99233949e-01 -1.60744295e-01 -1.67024523e-01
4.05383289e-01 -8.21862817e-01 6.53075635e-01 -1.40603185e+00
-7.39522129e-02 1.96809143e-01 4.03957248e-01 -5.93146741e-01
-2.06852198e-01 -9.55574274e-01 -3.51422936e-01 4.51045305e-01
1.89574983e-03 8.40500891e-02 2.12692738e-01 8.09439838e-01
-2.58326292e-01 -5.79557717e-01 7.62998104e-01 6.75371066e-02
-7.03491092e-01 1.05502643e-01 -7.49866664e-01 -6.01252168e-02
9.19234693e-01 -3.52415413e-01 -2.97199845e-01 -1.22086406e-02
-8.36613357e-01 1.60048649e-01 6.09121680e-01 2.76925951e-01
4.74644005e-01 -9.77956414e-01 -6.66223824e-01 2.74906337e-01
4.45267379e-01 2.65391976e-01 2.23207369e-01 1.11662674e+00
-7.14070499e-01 9.13906842e-02 -2.34023780e-01 -9.15209889e-01
-1.11795259e+00 9.30793345e-01 4.88559693e-01 -2.71797657e-01
-8.93644869e-01 2.52539515e-01 -1.01295166e-01 -3.07691723e-01
1.72719002e-01 -3.44373316e-01 -4.15119231e-01 1.11405879e-01
4.55867648e-01 5.21048665e-01 6.53687477e-01 -4.79641825e-01
-3.17832440e-01 3.43438834e-01 -2.46254560e-02 3.47064495e-01
1.60200560e+00 6.74583688e-02 -2.31056243e-01 4.04834360e-01
8.91144872e-01 6.90203011e-02 -1.20324516e+00 -1.57050982e-01
5.40847063e-01 -4.92116064e-01 -2.66170442e-01 -5.28430641e-01
-1.26543474e+00 1.06334245e+00 8.27597260e-01 7.05811977e-01
1.43230593e+00 -1.32408231e-01 3.27483833e-01 8.27539921e-01
-1.57379843e-02 -1.22959828e+00 2.51922637e-01 3.87607545e-01
5.47632039e-01 -1.19277239e+00 -4.63393405e-02 -1.94519997e-01
-1.56323820e-01 1.19449484e+00 5.96402884e-01 -3.94173056e-01
3.47626954e-01 8.39132071e-02 6.57281429e-02 -4.10369337e-01
-5.34993112e-01 -5.60297906e-01 2.75070786e-01 9.83755827e-01
-1.11373715e-01 -4.34043825e-01 -8.44719820e-03 2.42942572e-01
1.78632259e-01 -4.48078603e-01 9.26910520e-01 1.00196469e+00
-5.93007803e-01 -8.36452067e-01 -6.17341995e-01 8.48581016e-01
-7.37085581e-01 8.96131396e-02 -4.50707972e-01 9.98332739e-01
3.36623251e-01 1.09001946e+00 1.19041756e-01 -3.65581624e-02
6.69123709e-01 3.49336118e-01 1.40688360e-01 -5.77354074e-01
-6.20503604e-01 -2.31590256e-01 -9.78800431e-02 -7.49709666e-01
-5.18874586e-01 -8.30768287e-01 -7.65688181e-01 -1.19794803e-02
-5.70394039e-01 1.68093249e-01 2.98854291e-01 1.14630330e+00
7.17814490e-02 6.57140017e-01 4.92879242e-01 -7.82373190e-01
-4.04517114e-01 -9.50784504e-01 -7.35903978e-01 6.90743685e-01
7.64327824e-01 -7.54661739e-01 -8.30190301e-01 -1.02973260e-01] | [7.671454906463623, 2.1881444454193115] |
889f2107-27d3-4ec1-b5a6-cd386502e0e7 | sequence-to-sequence-pre-training-with-data | 1909.06002 | null | https://arxiv.org/abs/1909.06002v2 | https://arxiv.org/pdf/1909.06002v2.pdf | Sequence-to-sequence Pre-training with Data Augmentation for Sentence Rewriting | We study sequence-to-sequence (seq2seq) pre-training with data augmentation for sentence rewriting. Instead of training a seq2seq model with gold training data and augmented data simultaneously, we separate them to train in different phases: pre-training with the augmented data and fine-tuning with the gold data. We also introduce multiple data augmentation methods to help model pre-training for sentence rewriting. We evaluate our approach in two typical well-defined sentence rewriting tasks: Grammatical Error Correction (GEC) and Formality Style Transfer (FST). Experiments demonstrate our approach can better utilize augmented data without hurting the model's trust in gold data and further improve the model's performance with our proposed data augmentation methods. Our approach substantially advances the state-of-the-art results in well-recognized sentence rewriting benchmarks over both GEC and FST. Specifically, it pushes the CoNLL-2014 benchmark's $F_{0.5}$ score and JFLEG Test GLEU score to 62.61 and 63.54 in the restricted training setting, 66.77 and 65.22 respectively in the unrestricted setting, and advances GYAFC benchmark's BLEU to 74.24 (2.23 absolute improvement) in E&M domain and 77.97 (2.64 absolute improvement) in F&R domain. | ['Xu sun', 'Furu Wei', 'Tao Ge', 'Ming Zhou', 'Yi Zhang'] | 2019-09-13 | null | null | null | null | ['formality-style-transfer'] | ['natural-language-processing'] | [ 6.53158128e-01 2.93433338e-01 2.11829796e-01 -5.70878029e-01
-1.18769419e+00 -8.07557285e-01 4.64794636e-01 9.61293653e-03
-7.50960708e-01 9.92676198e-01 2.00051606e-01 -7.30295360e-01
4.93909866e-01 -5.55296540e-01 -8.23319554e-01 -2.54797310e-01
3.87943536e-01 3.37287545e-01 -1.02814063e-01 -9.98834789e-01
1.96856946e-01 -9.03988406e-02 -1.11696482e+00 4.04113591e-01
1.55502260e+00 5.17555833e-01 2.29485139e-01 1.07033670e+00
-4.67825048e-02 5.56796432e-01 -9.82501090e-01 -8.47167194e-01
1.94430351e-01 -5.30904531e-01 -8.98776829e-01 -3.80127132e-01
5.91316462e-01 3.01297456e-02 -8.17426518e-02 1.07678401e+00
6.44363582e-01 2.09387794e-01 2.34989598e-01 -6.01831615e-01
-1.11438584e+00 7.36500442e-01 -3.84712458e-01 2.17764601e-01
4.40272480e-01 2.16225445e-01 9.52203691e-01 -7.96201110e-01
7.37863541e-01 1.23928738e+00 6.14094675e-01 1.11171937e+00
-8.87973547e-01 -5.01831472e-01 2.32507616e-01 -1.50080934e-01
-9.84343946e-01 -5.02350092e-01 2.25345373e-01 -1.64044891e-02
1.61816168e+00 5.15134633e-01 -3.08797006e-02 1.16300070e+00
1.21203281e-01 5.83714128e-01 1.00294304e+00 -7.04289198e-01
-1.33873999e-01 -2.40569398e-01 1.80022776e-01 7.71685362e-01
-2.07776160e-04 2.82154888e-01 -2.77635276e-01 2.53904611e-01
1.43090934e-01 -2.80125856e-01 -2.43662000e-01 6.15134060e-01
-1.06396365e+00 5.47114313e-01 3.35925311e-01 2.06404716e-01
-7.15440437e-02 1.73552603e-01 7.42797315e-01 7.32975602e-01
4.27605271e-01 9.08446908e-01 -7.64610887e-01 -6.19814992e-01
-6.53211772e-01 3.57626230e-01 7.53479004e-01 1.28736567e+00
3.86153221e-01 2.72262961e-01 -7.76974499e-01 9.90593493e-01
-1.14456594e-01 9.49686050e-01 4.55816656e-01 -7.37790585e-01
1.13328493e+00 5.80633342e-01 1.31002646e-02 -4.28505599e-01
-6.65663183e-02 -7.07589865e-01 -9.24978554e-01 -2.72482574e-01
1.74422443e-01 -2.69896954e-01 -1.07680690e+00 2.09440827e+00
-1.95401758e-01 -2.48712495e-01 4.66959119e-01 6.08290017e-01
1.01592863e+00 5.89752614e-01 1.51129201e-01 -2.11510971e-01
1.24502277e+00 -1.33683348e+00 -6.62643552e-01 -3.88838470e-01
1.26333034e+00 -9.06749904e-01 1.67711592e+00 2.36132145e-01
-1.28381526e+00 -7.88104832e-01 -1.01361001e+00 -4.34354872e-01
-1.77018121e-01 2.97639757e-01 2.01323673e-01 5.19224107e-01
-1.07521152e+00 7.58522093e-01 -5.76582015e-01 -3.30826521e-01
2.62424827e-01 3.88604730e-01 -2.34749675e-01 -3.04012805e-01
-1.53420103e+00 1.12175655e+00 2.60910481e-01 -6.59895465e-02
-5.94447017e-01 -8.65993321e-01 -8.57746482e-01 6.03195950e-02
2.92251080e-01 -6.92656696e-01 1.49028254e+00 -7.96051443e-01
-1.57977819e+00 9.05681372e-01 -2.84229398e-01 -4.70346123e-01
3.03878784e-01 -5.07107377e-01 -6.52572691e-01 -4.37218964e-01
6.57617748e-02 7.20696151e-01 2.63003230e-01 -9.03401017e-01
-5.71672559e-01 -2.83059627e-02 3.93292427e-01 2.46641070e-01
-1.65538833e-01 2.46602118e-01 -3.27694356e-01 -7.94538617e-01
-5.33445358e-01 -9.82886612e-01 -2.05678403e-01 -7.31787860e-01
-4.82335210e-01 -3.37671787e-01 3.10782969e-01 -9.09763694e-01
1.64569890e+00 -1.77655804e+00 2.15271398e-01 -2.67355859e-01
-3.67468670e-02 1.00650835e+00 -7.41986215e-01 5.28071761e-01
-1.07689135e-01 3.63536447e-01 -5.85941672e-01 -7.52452135e-01
-1.27228647e-01 3.38760823e-01 -1.84469625e-01 -2.19821140e-01
4.86840159e-01 1.33400548e+00 -1.16723454e+00 -1.74553037e-01
-1.91286117e-01 5.25521152e-02 -7.37610340e-01 2.54881710e-01
-2.26213321e-01 4.03867126e-01 -1.94469571e-01 4.98765767e-01
4.96445030e-01 2.22562701e-02 1.24092743e-01 1.63852319e-01
-7.48866126e-02 6.39637828e-01 -5.43841302e-01 2.07935381e+00
-6.93601727e-01 3.56069386e-01 -4.21191454e-01 -5.90082586e-01
1.11592746e+00 1.72980621e-01 -2.58490384e-01 -1.08162618e+00
3.79416309e-02 3.30310822e-01 2.60660172e-01 -4.37321782e-01
9.32249963e-01 -1.40540376e-01 -4.24459547e-01 3.73803586e-01
1.15284465e-01 -6.04982674e-02 4.85020190e-01 3.18778932e-01
1.50799501e+00 1.17020510e-01 1.54361844e-01 -3.01228166e-01
8.34205687e-01 -2.03136206e-01 7.48012841e-01 7.82119095e-01
3.50631550e-02 6.21006846e-01 4.72091109e-01 -1.26716495e-01
-1.20211792e+00 -7.33354867e-01 3.45010012e-01 1.31883252e+00
-2.57714063e-01 -7.64000177e-01 -9.52414691e-01 -1.26356065e+00
-2.15424970e-01 1.00162113e+00 -5.53232551e-01 -4.75915641e-01
-1.01026273e+00 -5.32786846e-01 9.71560299e-01 6.53574646e-01
4.98246014e-01 -1.24861217e+00 1.54552897e-02 1.45136029e-01
-4.17868197e-01 -1.13096714e+00 -9.60769713e-01 -7.14788139e-02
-5.40617883e-01 -8.79673958e-01 -2.81278789e-01 -7.09012866e-01
7.97301292e-01 5.77615686e-02 1.44867051e+00 5.26684165e-01
8.33780318e-03 -1.84230044e-01 -8.59024465e-01 -4.35999542e-01
-9.35616791e-01 5.62351823e-01 5.61924130e-02 -5.98097563e-01
5.46730310e-02 -3.22626621e-01 -2.87960798e-01 6.37761950e-02
-8.03805768e-01 1.75895542e-02 5.89506447e-01 1.08089328e+00
3.33853483e-01 -9.23681200e-01 8.91505241e-01 -1.35823047e+00
8.27125013e-01 -1.78840324e-01 -3.35523516e-01 7.25714445e-01
-8.91220510e-01 1.37452662e-01 9.88327742e-01 1.55432317e-02
-1.04877484e+00 -4.12210703e-01 -4.17885512e-01 -1.60782516e-01
1.17856391e-01 5.21423519e-01 -1.61206529e-01 1.55430079e-01
9.62431848e-01 -3.23129073e-02 -7.52825066e-02 -4.75281209e-01
3.18223834e-01 7.64659166e-01 8.29733133e-01 -6.60902381e-01
8.78802180e-01 -4.08077627e-01 -2.54857689e-01 -1.40590712e-01
-1.17979002e+00 -1.17910236e-01 -5.47782898e-01 2.73485303e-01
7.07840621e-01 -8.40352654e-01 -5.23473501e-01 2.62707025e-01
-1.43594611e+00 -3.23474109e-01 -3.68740350e-01 -4.59422171e-02
-2.24765673e-01 5.82902133e-01 -4.70449775e-01 -4.72041398e-01
-1.07745790e+00 -1.14674437e+00 1.02587295e+00 1.38016716e-01
-3.18117112e-01 -8.90660167e-01 2.73643911e-01 4.79125321e-01
5.91609001e-01 3.36575449e-01 8.68907273e-01 -8.22241962e-01
-2.20795553e-02 -1.63334101e-01 -1.16503157e-01 9.34300959e-01
1.95761144e-01 -7.74210766e-02 -7.80882537e-01 -4.93382514e-01
-2.15262771e-01 -4.29362357e-01 7.79751003e-01 -2.93964505e-01
1.09183621e+00 -2.89949387e-01 1.26430720e-01 6.33932352e-01
1.18468726e+00 1.74753591e-01 7.99390793e-01 1.85651347e-01
6.66325927e-01 2.11215049e-01 9.64195132e-01 1.62902132e-01
4.63492960e-01 8.02334845e-01 3.93754601e-01 -8.42548460e-02
-3.14915925e-01 -2.88615793e-01 6.95440531e-01 1.24833405e+00
-1.27816573e-01 -5.76271057e-01 -9.87214565e-01 4.09157485e-01
-1.84407938e+00 -6.32791519e-01 -3.04156452e-01 1.97906697e+00
1.24304736e+00 1.85254648e-01 -2.61312604e-01 -6.88543990e-02
7.02669442e-01 -4.46121320e-02 -3.58459264e-01 -1.15562773e+00
-3.53819758e-01 9.27861094e-01 3.08622152e-01 8.52085710e-01
-9.47605550e-01 1.46051025e+00 5.48488665e+00 7.69575953e-01
-8.74634445e-01 1.42273247e-01 4.67177451e-01 -1.45506844e-01
-4.83952314e-01 7.84138218e-02 -1.01399398e+00 4.71714526e-01
1.36797762e+00 -2.54489779e-01 5.58216274e-01 5.16503036e-01
1.64412323e-03 2.22708657e-01 -1.13541925e+00 6.55957460e-01
1.58694252e-01 -1.26683772e+00 2.92935744e-02 -2.29833141e-01
1.04593146e+00 1.26334041e-01 -1.37638092e-01 1.04953396e+00
4.91103083e-01 -1.31112731e+00 4.63580281e-01 5.41439235e-01
1.21956086e+00 -7.87242949e-01 1.08664680e+00 3.91568780e-01
-7.29645431e-01 1.12702362e-01 -4.00457561e-01 -3.66562545e-01
2.81747282e-01 4.62213337e-01 -8.27163160e-01 9.64794040e-01
3.87968361e-01 7.22446203e-01 -8.07490110e-01 2.36012548e-01
-6.20829940e-01 7.07043946e-01 -1.16445079e-01 -1.49985328e-01
3.07299197e-01 -1.17838094e-02 5.13733029e-01 1.35530221e+00
2.06221208e-01 2.38718405e-01 4.53101844e-02 5.80209136e-01
-6.12431645e-01 1.43090803e-02 -3.09936970e-01 9.27676558e-02
7.24985778e-01 1.20718372e+00 2.57596165e-01 -4.68878955e-01
-1.51863962e-01 1.12215543e+00 7.80780196e-01 1.77363008e-01
-9.77091968e-01 -7.02979863e-01 6.44112766e-01 -3.65970016e-01
1.72375709e-01 -6.43820688e-02 -3.17626804e-01 -1.23197579e+00
2.96187341e-01 -1.34771168e+00 1.87951177e-01 -5.38441122e-01
-1.12638760e+00 9.82404768e-01 -3.10954422e-01 -1.05022502e+00
-2.46189803e-01 -3.98507714e-01 -8.14822614e-01 1.30581737e+00
-1.68363893e+00 -1.15959227e+00 -1.27442017e-01 1.76328883e-01
7.44774580e-01 -3.57664943e-01 1.11407435e+00 3.39570582e-01
-7.06548095e-01 1.43243277e+00 2.24674359e-01 1.18328303e-01
9.62259531e-01 -1.48082757e+00 1.20376468e+00 1.14931500e+00
-7.73348212e-02 7.86238492e-01 4.55772817e-01 -6.00319684e-01
-1.31626689e+00 -1.35994649e+00 1.47156513e+00 -7.27519393e-01
4.68447208e-01 -4.87680942e-01 -9.69642878e-01 5.80353677e-01
2.91124642e-01 -1.18859030e-01 4.67016280e-01 3.02325934e-01
-2.84432501e-01 -9.61086378e-02 -1.07887340e+00 5.65883100e-01
1.29266298e+00 -4.45345849e-01 -6.29535794e-01 3.51028055e-01
1.10528123e+00 -8.86584044e-01 -1.09105778e+00 5.96849799e-01
9.60460901e-02 -4.18002933e-01 5.70780158e-01 -1.26332247e+00
9.49702144e-01 -5.10090329e-02 -3.39412332e-01 -1.56537724e+00
-2.99928397e-01 -9.08762455e-01 -2.94501963e-03 1.54549599e+00
7.68941879e-01 -4.59784001e-01 4.55466092e-01 4.19417888e-01
-8.77123535e-01 -1.03939450e+00 -8.33006382e-01 -8.76035571e-01
4.48174775e-01 -1.22954242e-01 7.33369470e-01 8.56809318e-01
-2.82757785e-02 6.19592965e-01 -2.83796608e-01 -1.47962168e-01
1.13001138e-01 -2.61758447e-01 8.45649600e-01 -6.93450689e-01
-3.49225849e-01 -1.08580939e-01 1.57657757e-01 -1.07086706e+00
3.77167195e-01 -1.28844774e+00 1.34703502e-01 -1.36886096e+00
1.08308494e-01 -4.08821970e-01 -2.59690732e-01 7.54713237e-01
-8.54720294e-01 1.86404418e-02 4.85613912e-01 -1.70114025e-01
-7.68294871e-01 7.26130843e-01 1.33848488e+00 1.33208901e-01
4.18248549e-02 -2.55464971e-01 -9.88948405e-01 2.08262101e-01
1.00982332e+00 -4.35421348e-01 -3.48372400e-01 -9.93464828e-01
2.55527794e-01 -3.71481590e-02 -2.48585306e-02 -7.66698539e-01
-1.00473925e-01 3.07599362e-02 -2.12940902e-01 -2.53176719e-01
-1.92637816e-01 -8.98112208e-02 -3.81562084e-01 4.47638512e-01
-6.08139992e-01 6.05577171e-01 4.18324143e-01 3.53167444e-01
-9.67865586e-02 -3.43467981e-01 6.87281609e-01 -3.61898988e-02
-2.78263092e-01 1.21335844e-02 8.63387366e-04 5.88195026e-01
6.22262716e-01 1.42928049e-01 -8.17562819e-01 -1.50649279e-01
-5.90145946e-01 6.11073613e-01 3.52221400e-01 7.91248977e-01
4.43999082e-01 -1.22566867e+00 -1.12956882e+00 1.37576103e-01
1.61730707e-01 3.86165231e-02 7.97267258e-02 6.72986269e-01
-3.73560190e-01 5.82052588e-01 5.87438531e-02 -4.02741879e-01
-1.49059081e+00 2.14645490e-01 2.46770859e-01 -9.36908960e-01
-1.92146987e-01 1.03480053e+00 -1.07574783e-01 -9.31671500e-01
1.35039538e-02 -5.48796713e-01 4.81215976e-02 -6.19989097e-01
2.78977513e-01 3.95135939e-01 5.36472559e-01 -3.47192854e-01
-3.17539006e-01 3.40335697e-01 -6.19778037e-01 9.03252512e-02
1.19931889e+00 1.03491098e-01 -1.82285756e-01 -3.84568274e-02
1.22329521e+00 -2.89448560e-03 -7.39187598e-01 -4.06924129e-01
2.55039576e-02 -1.31996825e-01 -3.68478924e-01 -1.56913376e+00
-7.23612487e-01 9.63214099e-01 1.86519355e-01 -1.73201323e-01
1.17698681e+00 -3.28261614e-01 1.04611063e+00 6.97421491e-01
1.18808605e-01 -9.70549107e-01 2.59912629e-02 1.34679937e+00
1.19354010e+00 -1.24523890e+00 -2.45957643e-01 -3.81577462e-01
-9.86373425e-01 8.65316510e-01 9.14644122e-01 -1.93102717e-01
4.32806537e-02 1.51645452e-01 8.23200122e-02 1.02044158e-02
-1.14274323e+00 4.66833264e-02 4.06030983e-01 3.54244322e-01
7.81183064e-01 8.12853053e-02 -6.78470075e-01 8.43675137e-01
-5.60608804e-01 -1.19525075e-01 4.99638110e-01 9.40130413e-01
-3.33755285e-01 -1.80852652e+00 1.79742068e-01 4.93023336e-01
-5.52336514e-01 -6.45292580e-01 -5.14575601e-01 7.22841501e-01
-1.08289924e-02 9.95790899e-01 -1.76712036e-01 -5.56802750e-01
9.31280851e-01 1.07199840e-01 5.92530608e-01 -1.05796969e+00
-1.38649631e+00 -3.81349474e-01 7.53915489e-01 -3.89302254e-01
-3.15620266e-02 -4.99818563e-01 -1.38505530e+00 -4.28575158e-01
-3.37887168e-01 3.38453203e-01 5.35680830e-01 1.05504620e+00
6.29411519e-01 8.08683872e-01 7.44318724e-01 -1.52957425e-01
-9.91030514e-01 -1.33957934e+00 2.28584796e-01 5.08075237e-01
1.15411349e-01 -8.86119753e-02 2.27327365e-02 9.81965102e-03] | [11.34938907623291, 10.252737045288086] |
c11fd423-d249-4a63-8f0c-a2be2478ed2d | one-step-and-two-step-classification-for | 1706.01206 | null | http://arxiv.org/abs/1706.01206v1 | http://arxiv.org/pdf/1706.01206v1.pdf | One-step and Two-step Classification for Abusive Language Detection on Twitter | Automatic abusive language detection is a difficult but important task for
online social media. Our research explores a two-step approach of performing
classification on abusive language and then classifying into specific types and
compares it with one-step approach of doing one multi-class classification for
detecting sexist and racist languages. With a public English Twitter corpus of
20 thousand tweets in the type of sexism and racism, our approach shows a
promising performance of 0.827 F-measure by using HybridCNN in one-step and
0.824 F-measure by using logistic regression in two-steps. | ['Pascale Fung', 'Ji Ho Park'] | 2017-06-05 | one-step-and-two-step-classification-for-1 | https://aclanthology.org/W17-3006 | https://aclanthology.org/W17-3006.pdf | ws-2017-8 | ['abuse-detection'] | ['natural-language-processing'] | [-8.94070193e-02 -1.57835316e-02 -7.06542373e-01 -4.30547088e-01
-4.26967710e-01 -5.54386735e-01 1.09987962e+00 4.15415317e-01
-9.94619548e-01 9.47680175e-01 3.31427783e-01 -4.41568404e-01
3.12341362e-01 -8.41444790e-01 1.50559932e-01 -3.39186221e-01
1.72005147e-01 5.32466710e-01 1.00479671e-03 -6.97042465e-01
7.60849714e-01 6.26897216e-01 -1.07782543e+00 3.08603793e-01
6.36439383e-01 4.79130417e-01 -1.00899243e+00 5.96592367e-01
-7.36923337e-01 1.12253582e+00 -7.85091639e-01 -9.53410447e-01
1.69518441e-01 -6.59468770e-01 -1.11294949e+00 -1.74232885e-01
4.70233321e-01 -1.64018586e-01 2.11701803e-02 1.26172495e+00
2.70688504e-01 -2.60903779e-02 8.51447642e-01 -1.19550431e+00
-3.61604132e-02 9.70452905e-01 -9.85139787e-01 7.70981014e-01
6.44547880e-01 8.38550106e-02 7.94947684e-01 -5.67077696e-01
7.41362870e-01 1.90861821e+00 6.70572102e-01 6.16185069e-01
-1.40221763e+00 -1.57809138e+00 -2.18144521e-01 -1.90586269e-01
-1.35280097e+00 -2.96317875e-01 7.75887728e-01 -8.40341568e-01
9.23097670e-01 3.23961705e-01 6.82492435e-01 1.41492450e+00
1.50023922e-01 2.37715125e-01 1.67959881e+00 -7.73504078e-02
-1.71139225e-01 7.68201530e-01 6.98155224e-01 5.57159841e-01
6.86127782e-01 -2.44049594e-01 -4.84895289e-01 -7.34302163e-01
1.12905808e-01 -1.42221734e-01 7.87499726e-01 7.47680426e-01
-6.78909123e-01 1.73757398e+00 9.21237618e-02 7.68582582e-01
-1.92935560e-02 -1.98719576e-01 1.06634331e+00 4.54675496e-01
9.11323130e-01 6.99528992e-01 3.64310920e-01 -1.51659086e-01
-1.20198441e+00 2.70114183e-01 9.39373970e-01 3.53603691e-01
4.19077069e-01 1.49524838e-01 9.59680155e-02 1.10363460e+00
1.70123652e-01 4.81511623e-01 9.76877093e-01 -3.79056782e-01
4.37686026e-01 7.28926122e-01 -2.56904900e-01 -1.37183464e+00
-6.29584074e-01 -4.78673503e-02 -5.75727999e-01 -4.75477614e-02
5.96503556e-01 -2.26228565e-01 -6.15161479e-01 1.40278327e+00
4.69691813e-01 -2.88684666e-01 -2.31388718e-01 4.58458006e-01
8.52054834e-01 6.39884412e-01 4.87431347e-01 -5.34629941e-01
1.29381633e+00 -5.79584956e-01 -7.57013798e-01 -2.50566661e-01
7.10717976e-01 -7.22576737e-01 8.33072126e-01 6.63637042e-01
-9.24149215e-01 -1.23793907e-01 -9.35380518e-01 -1.62717581e-01
-8.33434165e-01 -6.43177688e-01 7.42557168e-01 1.37753594e+00
-5.24866104e-01 3.23533893e-01 -1.10184483e-01 -5.93330443e-01
4.05254334e-01 2.18994856e-01 -3.58446062e-01 3.04391682e-01
-1.11288750e+00 9.71251547e-01 2.34966367e-01 -5.62861383e-01
-5.80149353e-01 -3.84392828e-01 -7.38903642e-01 -5.86015284e-01
2.03065097e-01 2.89833277e-01 7.23984122e-01 -1.21933830e+00
-1.22117949e+00 1.62336206e+00 -8.20572078e-02 -4.33019489e-01
8.39491010e-01 -1.55167893e-01 -7.86887467e-01 1.34054050e-01
5.55973113e-01 3.13757986e-01 9.00863290e-01 -1.07781065e+00
-6.40080273e-01 -6.01781607e-01 2.04693805e-02 -1.04788154e-01
-5.43547869e-01 8.49537730e-01 7.17790544e-01 -6.09048963e-01
9.88682657e-02 -7.44611144e-01 1.74826290e-02 -4.29979950e-01
-6.75953269e-01 -4.34722155e-01 6.65348470e-01 -9.45116997e-01
1.64778042e+00 -2.10800219e+00 -1.77455068e-01 5.48341215e-01
6.20284975e-01 2.09167451e-01 4.62193131e-01 3.25298727e-01
-7.76192918e-02 7.38565683e-01 -4.06964449e-03 -3.81239474e-01
-2.03815252e-01 -5.00238407e-03 -2.77417094e-01 8.48441005e-01
2.43701473e-01 1.78321034e-01 -1.08038151e+00 -9.20318305e-01
-9.38226804e-02 2.68510491e-01 -4.04878944e-01 -8.09784308e-02
5.90241134e-01 1.14600435e-01 -1.71087161e-01 7.66059577e-01
6.29920363e-01 5.61021090e-01 2.89829314e-01 4.50180173e-01
-3.85055065e-01 5.97546220e-01 -8.48920763e-01 5.75070083e-01
-3.16833049e-01 1.00354803e+00 2.55565733e-01 -6.46773160e-01
1.27256668e+00 -2.04043955e-01 2.23429009e-01 -5.65025985e-01
6.94173574e-01 4.20895636e-01 3.45422506e-01 -3.84932578e-01
5.63778758e-01 -8.11066210e-01 -6.23614073e-01 5.32516837e-01
-1.05173411e-02 -1.56129360e-01 4.16322291e-01 3.56277823e-01
9.05762136e-01 -5.93932092e-01 7.06483245e-01 -7.07041204e-01
9.17581856e-01 2.61865258e-02 3.08605999e-01 7.17804670e-01
-8.41366053e-01 4.54068091e-03 9.72609997e-01 -6.12467825e-01
-8.99703443e-01 -7.37690270e-01 -2.26393536e-01 1.31508410e+00
-1.69509456e-01 -2.13053301e-01 -5.60513616e-01 -8.22260022e-01
2.65739352e-01 1.00340402e+00 -8.32326055e-01 -2.45391816e-01
-6.09873056e-01 -1.13478911e+00 9.71609235e-01 -1.99838281e-01
5.24591684e-01 -1.05308533e+00 -1.58318341e-01 -4.88985889e-02
-8.24538805e-03 -8.74334514e-01 3.56574655e-02 7.53378198e-02
-3.71462911e-01 -1.01137125e+00 -7.13086650e-02 -4.77903694e-01
2.07653329e-01 1.70822740e-01 1.00640917e+00 3.59380960e-01
-3.61328006e-01 -4.51007009e-01 -2.49587357e-01 -4.60698873e-01
-8.02292347e-01 2.93974936e-01 2.00699076e-01 2.56232113e-01
9.68066573e-01 -5.68162799e-01 -1.50995627e-01 -2.33569562e-01
-5.95235050e-01 -4.06021625e-01 8.11437592e-02 3.52452219e-01
-3.81244659e-01 -3.22538793e-01 4.53048408e-01 -1.36693788e+00
9.93733704e-01 -1.24479938e+00 -1.78999886e-01 -5.39388657e-01
-6.61535203e-01 -6.11227810e-01 5.79922795e-01 -7.82969654e-01
-6.91657901e-01 -2.34910458e-01 -4.38381553e-01 2.01387405e-01
-2.52635866e-01 5.87604344e-02 4.90383089e-01 1.00006655e-01
9.04715002e-01 -9.14377570e-02 1.07409880e-01 -2.10433751e-01
-3.03457677e-01 1.09359241e+00 -2.03408226e-02 1.34134954e-02
9.53729033e-01 5.92342496e-01 -1.19342178e-01 -1.17504334e+00
-9.23929393e-01 -9.22549427e-01 -7.93776989e-01 -4.07675207e-01
1.00405812e+00 -7.90756106e-01 -9.42925930e-01 7.13028491e-01
-7.88981736e-01 5.50856590e-02 3.81644338e-01 1.81086093e-01
1.04887165e-01 3.13213497e-01 -8.85737777e-01 -1.31352043e+00
-5.02628744e-01 -5.95438540e-01 4.58523393e-01 -1.16719328e-01
-8.52145791e-01 -9.77247894e-01 3.01403016e-01 6.02972031e-01
3.46031457e-01 6.06927335e-01 6.69948697e-01 -1.41841340e+00
8.27849805e-01 -4.39936817e-01 -2.46551543e-01 6.53654784e-02
7.87203237e-02 2.62169212e-01 -8.67526948e-01 -1.11420140e-01
-4.06850800e-02 -6.27638161e-01 7.71549344e-01 -2.55632371e-01
6.17893398e-01 -5.99638164e-01 -2.32490852e-01 1.28156215e-01
1.44271004e+00 1.74611405e-01 4.48032558e-01 5.54221749e-01
7.71387339e-01 7.73288190e-01 5.30521810e-01 4.11685497e-01
1.33385420e-01 2.24973604e-01 1.82836697e-01 3.75710614e-02
2.97297537e-01 -1.46397263e-01 5.77156425e-01 3.39187056e-01
2.49376938e-01 2.30778977e-01 -1.22756135e+00 5.67640007e-01
-1.34755301e+00 -1.52130127e+00 -5.12081504e-01 1.72825027e+00
9.46336448e-01 6.63081825e-01 1.14024520e+00 5.60165286e-01
9.81488228e-01 2.18641326e-01 -1.30667061e-01 -1.44066441e+00
-7.93800279e-02 1.26404271e-01 5.17789006e-01 1.11901844e+00
-1.06215072e+00 1.19393480e+00 6.98783445e+00 9.71380830e-01
-1.46264040e+00 3.16636115e-01 8.06025863e-01 -3.24594617e-01
1.69056267e-01 -3.76420528e-01 -1.11396110e+00 6.87211990e-01
1.21356821e+00 -2.10192934e-01 3.28360140e-01 6.82559431e-01
2.74451166e-01 -4.36510921e-01 -4.41493511e-01 8.22931826e-01
5.66286802e-01 -7.99908459e-01 1.88836902e-01 2.39136353e-01
2.51503825e-01 -1.20775655e-01 -1.57806501e-01 5.66160262e-01
3.15047950e-01 -1.14563978e+00 8.55873048e-01 -4.18756418e-02
5.37081182e-01 -1.07076848e+00 7.61782587e-01 5.08610547e-01
-3.74639124e-01 -3.80171567e-01 -4.15815786e-02 -7.02817380e-01
-1.08509265e-01 4.68757123e-01 -8.65127087e-01 -3.63457382e-01
9.45855558e-01 6.06291115e-01 -6.47558331e-01 4.49463874e-01
1.13346968e-02 1.04697800e+00 -2.91721374e-01 -5.69722116e-01
5.95727205e-01 -4.06761736e-01 5.74604452e-01 1.89277685e+00
-2.38093868e-01 2.46159248e-02 8.41325596e-02 4.87071097e-01
1.54477686e-01 5.52984715e-01 -8.38724911e-01 -2.33818144e-01
2.89064616e-01 1.47090018e+00 -1.08069456e+00 -6.04511559e-01
-1.20833129e-01 6.09004796e-01 4.79071379e-01 -4.11665201e-01
-7.17214465e-01 -6.15189791e-01 5.17240167e-01 6.07010484e-01
-5.02683878e-01 -7.80817494e-02 -5.91843367e-01 -8.17709684e-01
-5.92473269e-01 -9.57133651e-01 8.57523203e-01 7.30462186e-03
-1.52196908e+00 5.92743814e-01 2.49927744e-01 -7.33894587e-01
-2.25964233e-01 -6.45970404e-01 -5.82570314e-01 5.58622301e-01
-9.82161582e-01 -1.27182376e+00 -1.10549524e-01 4.37187850e-01
5.00833750e-01 -2.32580274e-01 5.27475595e-01 2.86077321e-01
-8.65368247e-01 6.59294188e-01 -2.90394127e-01 5.57617366e-01
8.12133670e-01 -9.59136724e-01 -2.38084137e-01 6.46972299e-01
-5.58987319e-01 5.94940186e-01 1.01188266e+00 -8.62824976e-01
-5.36478162e-01 -7.98194706e-01 1.45445073e+00 -3.67782861e-01
1.38575280e+00 -8.16450775e-01 -4.98198450e-01 6.51541293e-01
4.13769811e-01 -6.36019588e-01 1.25897193e+00 2.87987381e-01
-7.51890600e-01 2.69955158e-01 -1.88768137e+00 5.38720965e-01
1.11977506e+00 -4.33402956e-01 -6.32330179e-01 6.39158309e-01
2.12785646e-01 1.73435390e-01 -7.47353673e-01 -2.04930127e-01
5.71558118e-01 -1.08457983e+00 5.56437075e-01 -8.41160119e-01
8.04578722e-01 3.71020764e-01 -1.30075784e-02 -7.79422402e-01
-2.18434095e-01 -6.27037644e-01 5.28761923e-01 1.29357708e+00
5.35157144e-01 -1.03094053e+00 4.37294930e-01 1.77794084e-01
3.68187517e-01 -1.90994307e-01 -1.08017099e+00 -5.76745033e-01
7.75453508e-01 -2.62479335e-01 2.79259328e-02 1.63619626e+00
3.25648993e-01 6.30979359e-01 -4.26954627e-01 -4.78225499e-01
4.49001163e-01 -3.21329683e-01 9.04528260e-01 -1.22923195e+00
3.07779878e-01 -9.14408565e-01 -6.23068333e-01 1.17703810e-01
5.09396732e-01 -7.90131986e-01 -6.16669178e-01 -6.66141927e-01
6.84212744e-01 -1.17655173e-01 2.20145464e-01 3.41256142e-01
3.59901367e-03 1.17019951e+00 7.01369047e-02 9.44256634e-02
-3.86273444e-01 -1.64320603e-01 5.33974946e-01 -2.97585994e-01
-4.15587634e-01 -2.85294771e-01 -9.03045058e-01 8.73910427e-01
9.44287181e-01 -8.65876079e-01 1.34717207e-02 4.28287506e-01
5.55320680e-01 -5.05422533e-01 1.22911267e-01 -8.23851883e-01
-4.11363363e-01 -6.58808827e-01 2.46455491e-01 -4.04939681e-01
4.56007987e-01 -3.45846772e-01 -3.20335299e-01 1.07962286e+00
-4.57047880e-01 2.22445264e-01 1.68367654e-01 7.29897320e-02
-1.68896109e-01 -4.74299610e-01 1.16034758e+00 -3.12720448e-01
-1.96396589e-01 -2.22753823e-01 -9.15259480e-01 1.39897943e-01
1.27607048e+00 -4.16959912e-01 -3.46863240e-01 -4.42759961e-01
-8.67984056e-01 -1.18286759e-01 2.31372893e-01 3.78382564e-01
1.01639196e-01 -9.64303493e-01 -9.34551775e-01 7.85861164e-02
2.65280213e-02 -1.07801378e+00 -4.65253323e-01 7.76729167e-01
-7.03936875e-01 -2.29524598e-02 -5.87700427e-01 -2.33873472e-01
-1.69446325e+00 3.48433107e-01 2.77622670e-01 -3.34895641e-01
1.81093533e-02 6.32867157e-01 -4.13253248e-01 -4.24781919e-01
-3.02436769e-01 5.73674858e-01 -7.81490624e-01 7.76747048e-01
7.04717577e-01 7.47741699e-01 -3.10824007e-01 -1.52067816e+00
-4.70084637e-01 2.31642857e-01 -3.97020459e-01 -4.71193343e-01
1.05474949e+00 -4.43020090e-02 -6.79187179e-01 9.28402305e-01
1.46607983e+00 6.49877489e-01 4.97145802e-02 2.02219158e-01
2.15665683e-01 -7.24301755e-01 -1.29822209e-01 -6.27596736e-01
-5.17383337e-01 6.32961810e-01 1.37859017e-01 8.67915928e-01
5.72586298e-01 -1.21301733e-01 3.62454802e-01 2.18852167e-03
4.11803946e-02 -1.43341208e+00 2.00526699e-01 6.12057567e-01
7.72413790e-01 -1.39279473e+00 3.19309562e-01 -8.81934822e-01
-5.48118412e-01 9.74872887e-01 7.69682705e-01 -5.64269900e-01
7.45897830e-01 -7.71386027e-02 3.26659709e-01 -3.88867736e-01
-2.32357547e-01 5.29979579e-02 -1.86177511e-02 2.43594691e-01
8.59279990e-01 2.18114272e-01 -1.32278407e+00 2.75979638e-01
-9.27572191e-01 -8.20290148e-01 6.96008921e-01 8.84743392e-01
-5.36478877e-01 -6.42529964e-01 -5.27162731e-01 8.08299482e-01
-1.16148245e+00 5.52087240e-02 -1.43901932e+00 1.33731377e+00
1.45298839e-01 1.23200822e+00 4.05085504e-01 -6.11287236e-01
-1.75576329e-01 2.26679668e-01 1.06269665e-01 -6.63147688e-01
-1.41620529e+00 -1.56288370e-01 7.41819561e-01 -4.32151318e-01
-4.71199661e-01 -9.29345012e-01 -9.55247819e-01 -1.10826039e+00
-1.18894376e-01 1.00498229e-01 7.15482891e-01 1.01120353e+00
-2.81569809e-01 -3.24436337e-01 7.93244839e-01 -6.45398378e-01
-3.82677883e-01 -1.29693413e+00 -5.85411668e-01 8.63083363e-01
4.35317397e-01 -5.68449736e-01 -8.43423069e-01 -6.03067815e-01] | [8.747007369995117, 10.516603469848633] |
f28fe92e-840d-4427-9435-0ab1b64afdf1 | claws-contrastive-learning-with-hard | 2112.00847 | null | https://arxiv.org/abs/2112.00847v2 | https://arxiv.org/pdf/2112.00847v2.pdf | CLAWS: Contrastive Learning with hard Attention and Weak Supervision | Learning effective visual representations without human supervision is a long-standing problem in computer vision. Recent advances in self-supervised learning algorithms have utilized contrastive learning, with methods such as SimCLR, which applies a composition of augmentations to an image, and minimizes a contrastive loss between the two augmented images. In this paper, we present CLAWS, an annotation-efficient learning framework, addressing the problem of manually labeling large-scale agricultural datasets along with potential applications such as anomaly detection and plant growth analytics. CLAWS uses a network backbone inspired by SimCLR and weak supervision to investigate the effect of contrastive learning within class clusters. In addition, we inject a hard attention mask to the cropped input image before maximizing agreement between the image pairs using a contrastive loss function. This mask forces the network to focus on pertinent object features and ignore background features. We compare results between a supervised SimCLR and CLAWS using an agricultural dataset with 227,060 samples consisting of 11 different crop classes. Our experiments and extensive evaluations show that CLAWS achieves a competitive NMI score of 0.7325. Furthermore, CLAWS engenders the creation of low dimensional representations of very large datasets with minimal parameter tuning and forming well-defined clusters, which lends themselves to using efficient, transparent, and highly interpretable clustering methods such as Gaussian Mixture Models. | ['Ali Jannesari', 'Matthew Darr', 'John Just', 'Ramakrishnan Sundareswaran', 'Jansel Herrera-Gerena'] | 2021-12-01 | null | null | null | null | ['hard-attention'] | ['methodology'] | [ 2.86742628e-01 1.70242295e-01 -1.78374887e-01 -3.02475095e-01
-1.02906249e-01 -6.11540794e-01 4.20110554e-01 4.82252300e-01
-2.26595700e-01 1.22161783e-01 -3.18045467e-01 -2.40377113e-01
-2.78770290e-02 -6.75305426e-01 -8.27297866e-01 -8.79049122e-01
-2.94608355e-01 4.18771029e-01 7.22000003e-02 -4.31309529e-02
-1.42507672e-01 6.32182300e-01 -1.80221570e+00 2.13193253e-01
1.05765748e+00 9.98404384e-01 3.70878398e-01 5.08349538e-01
-1.42149657e-01 6.27070546e-01 -4.40471470e-01 -3.03306431e-01
4.24437106e-01 -2.04183087e-01 -6.18322134e-01 6.66903198e-01
6.13742292e-01 -1.52837308e-02 3.24392095e-02 1.13103759e+00
9.15193409e-02 2.81654373e-02 9.61215615e-01 -1.66229832e+00
-8.46264780e-01 5.32145679e-01 -1.13050854e+00 2.13454925e-02
-3.55432004e-01 3.01323265e-01 8.47708464e-01 -8.71358275e-01
2.88103759e-01 1.19124055e+00 7.00443804e-01 2.88897812e-01
-1.67215097e+00 -5.50663292e-01 4.45326775e-01 2.46289819e-02
-1.24655342e+00 -2.48713076e-01 6.71990395e-01 -6.11767650e-01
5.04354298e-01 1.32802516e-01 4.92592692e-01 8.42145026e-01
-1.84142068e-01 9.73786950e-01 7.95341134e-01 -6.85512900e-01
3.52514029e-01 9.39488411e-03 2.48292297e-01 9.42390859e-01
4.59902644e-01 -1.30239069e-01 -1.11153901e-01 -2.21582294e-01
7.07776666e-01 2.73352116e-01 -3.92742306e-02 -9.27990615e-01
-1.08883679e+00 8.96064460e-01 6.19902432e-01 1.03054382e-01
-3.27506185e-01 -2.91552097e-01 4.41024691e-01 1.96702853e-01
6.54208064e-01 6.11378789e-01 -4.26918864e-01 6.78789556e-01
-8.77123237e-01 -2.79166162e-01 6.01186097e-01 8.18227530e-01
7.94792414e-01 2.38729596e-01 1.40946209e-02 9.74087596e-01
3.24660301e-01 5.67242801e-01 3.07557136e-01 -1.21409357e+00
-7.64364824e-02 8.65948975e-01 -1.99480444e-01 -1.23592317e+00
-3.64848584e-01 -3.23327094e-01 -1.30340266e+00 6.31852746e-01
4.80111003e-01 7.53518716e-02 -1.29291284e+00 1.75431180e+00
2.58061647e-01 1.40494317e-01 1.80647910e-01 6.54410481e-01
5.79821765e-01 5.70670485e-01 1.34213641e-01 -2.19814807e-01
1.05405092e+00 -1.08542383e+00 -6.09259367e-01 -4.23798054e-01
6.78639054e-01 -6.49716973e-01 1.23842525e+00 5.29704809e-01
-8.12854171e-01 -5.05423546e-01 -1.14899528e+00 1.88209981e-01
-6.00099206e-01 5.72211385e-01 7.56877244e-01 3.24114412e-01
-9.45581138e-01 6.02655590e-01 -1.00749397e+00 -4.64292347e-01
8.64720225e-01 2.62545466e-01 -5.82664073e-01 -3.00276369e-01
-4.01328146e-01 6.28998816e-01 6.77791297e-01 2.68694580e-01
-8.95950735e-01 -6.49056554e-01 -1.08646250e+00 4.01315056e-02
5.31662762e-01 -1.38449550e-01 7.69158363e-01 -1.38716006e+00
-1.20037329e+00 1.28260791e+00 -1.68156382e-02 -5.33117712e-01
1.97432801e-01 -3.16814274e-01 -2.30502233e-01 7.84297660e-02
2.76347786e-01 8.86548102e-01 9.95107412e-01 -1.53801501e+00
-4.94923502e-01 -5.97900748e-01 -3.31027836e-01 1.47314340e-01
-5.72007537e-01 -1.84425637e-01 -3.94785613e-01 -7.33729005e-01
2.85131365e-01 -1.00586092e+00 -4.89654094e-01 5.29929280e-01
-2.77118802e-01 1.52171375e-02 9.84574914e-01 -6.77556455e-01
8.32602441e-01 -2.49593735e+00 9.05202627e-02 3.53612691e-01
4.39815253e-01 5.55159390e-01 -6.27047122e-01 -1.56498566e-01
-2.79952437e-01 -2.62799580e-02 -7.66024172e-01 -2.35557169e-01
-3.07489663e-01 4.06754613e-01 -1.34204030e-01 4.86516297e-01
5.76269329e-01 6.64424121e-01 -9.22594965e-01 -4.31746840e-01
3.31517667e-01 2.00588465e-01 -3.30817550e-01 4.00371879e-01
-3.67130756e-01 -9.74327102e-02 3.71851064e-02 8.17882657e-01
9.10838246e-01 -4.32996422e-01 1.33037806e-01 -2.63175637e-01
4.36705118e-03 -6.16737902e-01 -1.02579224e+00 1.63899004e+00
-3.96389998e-02 6.15960002e-01 3.46135527e-01 -1.40745449e+00
1.15625942e+00 -1.41286314e-01 4.51064944e-01 -2.39255413e-01
-5.46177775e-02 -2.32765935e-02 -4.51773666e-02 -2.67865747e-01
6.15220815e-02 4.89664018e-01 2.56258219e-01 2.82712102e-01
3.16876084e-01 -3.16040143e-02 4.61402088e-01 3.82305473e-01
1.00038528e+00 2.38784820e-01 4.87339109e-01 -3.86085451e-01
2.53363073e-01 2.22772479e-01 5.65682530e-01 6.58015251e-01
-2.73691773e-01 6.40471995e-01 4.75391746e-01 -5.16767144e-01
-9.87775266e-01 -1.18165421e+00 1.04936875e-01 1.23756087e+00
4.79924493e-02 -2.79194504e-01 -6.56943560e-01 -9.55308259e-01
1.66843921e-01 5.65029621e-01 -7.06605196e-01 -3.63198340e-01
-3.23447794e-01 -1.17672908e+00 4.37804222e-01 6.27163589e-01
5.86901844e-01 -1.14100385e+00 -3.51284713e-01 -1.14155702e-01
-1.04981668e-01 -1.05998576e+00 -2.00988218e-01 7.70396292e-01
-8.50751877e-01 -1.17723274e+00 -4.87487793e-01 -1.05920351e+00
9.84774411e-01 5.62182665e-01 1.21009123e+00 8.28969330e-02
-7.96353221e-01 1.70054168e-01 -3.81471097e-01 -5.47154665e-01
-4.65299398e-01 1.25021800e-01 -2.07726285e-03 -4.74463701e-02
6.08216465e-01 -5.26522934e-01 -2.34356672e-01 2.64756411e-01
-9.16758895e-01 -1.87253207e-01 7.73534536e-01 9.42858160e-01
8.04666638e-01 -7.83087835e-02 3.40419292e-01 -8.51306617e-01
1.34570926e-01 -5.42913496e-01 -8.66527796e-01 4.76712346e-01
-6.66805863e-01 1.25141814e-01 5.37003815e-01 -5.56740522e-01
-8.74420822e-01 5.82405567e-01 3.41488630e-01 -8.03040981e-01
-5.17621875e-01 2.73239911e-01 -3.35066736e-01 4.79212962e-02
9.75912035e-01 -9.96928010e-03 5.41442573e-01 -2.44281128e-01
6.58698797e-01 5.29121518e-01 9.32061374e-01 -3.89202327e-01
9.60616529e-01 4.71588343e-01 -7.03567192e-02 -1.13018990e+00
-1.01569784e+00 -6.21674836e-01 -9.37453449e-01 -1.52416469e-04
8.04365933e-01 -9.24308598e-01 -5.03119767e-01 5.45644522e-01
-7.44290233e-01 -5.29791951e-01 -6.62005901e-01 3.14265490e-01
-3.69177014e-01 4.66790318e-01 -4.75899369e-01 -6.61261082e-01
-2.65202820e-01 -7.33680606e-01 8.75933230e-01 1.49962410e-01
-4.33041975e-02 -8.49226236e-01 -2.64407903e-01 1.92994565e-01
1.62279844e-01 5.07001340e-01 1.00958693e+00 -7.65063286e-01
-2.16840759e-01 -2.72829141e-02 -4.61354882e-01 7.60400474e-01
4.86436397e-01 2.84983426e-01 -1.14479005e+00 -4.17910278e-01
-3.60396892e-01 -6.31350458e-01 1.15703857e+00 5.67315519e-01
1.51821315e+00 -1.37914002e-01 -4.62216675e-01 7.50571668e-01
1.14031184e+00 7.55699202e-02 5.07180095e-01 1.40309244e-01
9.61694419e-01 7.18209088e-01 6.22698069e-01 2.70231694e-01
7.16870651e-02 1.44629121e-01 8.60377133e-01 -6.62970722e-01
4.17618789e-02 7.60217980e-02 2.95152575e-01 4.77720380e-01
4.35848869e-02 -2.51867861e-01 -8.71020973e-01 5.44133902e-01
-2.03015447e+00 -8.99270892e-01 1.00772120e-01 2.22001171e+00
5.86274564e-01 -7.88435712e-02 -7.72334030e-03 2.32979625e-01
8.78120959e-01 5.71030341e-02 -8.35188270e-01 -1.03131592e-01
-4.61162984e-01 8.03917125e-02 5.06169796e-01 1.79307699e-01
-1.66144753e+00 1.10863578e+00 6.02520275e+00 7.61049271e-01
-8.50816429e-01 -3.06167781e-01 8.55192661e-01 1.00221366e-01
2.91694015e-01 -2.69263893e-01 -4.60519731e-01 2.29839787e-01
5.10568678e-01 1.85625568e-01 3.49390179e-01 1.06812286e+00
1.96058094e-03 -5.65577112e-02 -1.06901443e+00 7.79601932e-01
3.05954576e-01 -1.06712842e+00 1.39711082e-01 -1.26169669e-02
6.42390430e-01 -5.22852549e-03 9.76769850e-02 1.21556543e-01
8.14880729e-01 -1.08546579e+00 2.54602462e-01 8.52797180e-02
5.42349279e-01 -7.46641397e-01 6.98447704e-01 2.07791939e-01
-1.17819548e+00 -2.27339044e-01 -5.65199196e-01 1.72010601e-01
-4.78538722e-01 7.84645915e-01 -7.48081028e-01 4.81950849e-01
1.02905691e+00 9.26905096e-01 -9.80091393e-01 8.91384244e-01
6.54215962e-02 7.29741096e-01 -4.47174132e-01 4.65092719e-01
1.77462921e-01 -4.27443832e-01 2.39858866e-01 1.21957064e+00
3.97070497e-02 -2.50243157e-01 6.41096652e-01 9.00125980e-01
-1.73571687e-02 8.10498893e-02 -8.66861761e-01 -2.22968698e-01
4.04914707e-01 1.57834554e+00 -1.07100856e+00 -1.72125205e-01
-2.16051206e-01 1.18452764e+00 4.34500098e-01 1.89248323e-01
-4.47425336e-01 -2.50720114e-01 6.10428631e-01 8.19076970e-03
3.28051537e-01 -3.26737165e-02 -1.05719455e-01 -9.29169893e-01
-3.94616984e-02 -9.91429210e-01 5.73114634e-01 -6.24579132e-01
-1.64436042e+00 4.61550444e-01 -8.76171514e-02 -1.10645092e+00
-3.41011435e-02 -9.11601603e-01 -5.13754129e-01 4.76194441e-01
-1.42573249e+00 -1.37936437e+00 -7.11803257e-01 5.55348098e-01
5.80418885e-01 -4.70519692e-01 1.06718493e+00 -1.30136665e-02
-8.09036911e-01 5.54938197e-01 1.88946217e-01 2.43631020e-01
8.44893515e-01 -1.47336638e+00 1.65073037e-01 1.00084817e+00
1.92696080e-01 2.15419933e-01 4.18355465e-01 -5.58401942e-01
-7.74376392e-01 -1.60074675e+00 2.26644248e-01 -2.82381266e-01
4.84209836e-01 -3.45920354e-01 -1.27655542e+00 6.63323641e-01
2.23002583e-01 6.00888669e-01 7.24383891e-01 -7.12945610e-02
-5.58072090e-01 -8.26630592e-02 -1.24610603e+00 4.15492952e-01
1.10594761e+00 -1.00925960e-01 -2.08061308e-01 4.35212970e-01
7.57936776e-01 -3.84256206e-02 -5.80444753e-01 5.76551795e-01
2.06087425e-01 -6.95412636e-01 9.31823850e-01 -6.35645390e-01
3.90977561e-01 -4.87840533e-01 -2.52797753e-01 -1.47314501e+00
-5.63368738e-01 -2.58154929e-01 -1.40457988e-01 1.30809546e+00
3.28672767e-01 -2.92494833e-01 7.86658406e-01 1.46969706e-01
-1.52031898e-01 -4.22069848e-01 -8.05661231e-02 -8.21129680e-01
5.64413443e-02 6.55649155e-02 1.82254732e-01 1.29264843e+00
-2.86554158e-01 2.75011897e-01 -1.37907237e-01 5.08478880e-01
1.01211798e+00 3.46253365e-02 6.61262989e-01 -1.82867944e+00
3.66727002e-02 -3.02197069e-01 -4.14684892e-01 -4.44737643e-01
4.73990679e-01 -1.02275848e+00 2.28876770e-01 -1.21213877e+00
3.22747350e-01 -4.38876688e-01 -1.32537618e-01 8.93617153e-01
-2.82100141e-01 2.11847469e-01 1.61730200e-02 2.00500488e-01
-5.84321916e-01 4.60449964e-01 7.46965468e-01 -5.00675619e-01
-3.72226566e-01 -6.81803897e-02 -6.04301691e-01 1.13729453e+00
1.07234669e+00 -2.97299623e-01 -4.32460487e-01 -3.02476555e-01
-3.31375390e-01 -7.81140804e-01 4.73640621e-01 -8.59679937e-01
-8.24782327e-02 -1.30934611e-01 8.38003814e-01 -4.67344731e-01
6.59240186e-02 -8.65126193e-01 -2.95364588e-01 4.26795870e-01
-3.99368346e-01 1.34131700e-01 3.63312423e-01 6.42761350e-01
-1.48236096e-01 -1.73251420e-01 1.06703508e+00 -2.98654139e-01
-8.86053801e-01 2.51352489e-01 -2.90837556e-01 4.70924526e-02
1.24907041e+00 -9.99575332e-02 -2.14205220e-01 6.99851811e-02
-8.50429356e-01 4.59181130e-01 4.69098330e-01 4.87247676e-01
5.12191415e-01 -1.10108328e+00 -7.85223544e-01 4.71695393e-01
4.08916235e-01 4.14757401e-01 -1.27284348e-01 3.39573592e-01
-4.90404010e-01 -1.54574066e-01 -4.70239580e-01 -1.01105392e+00
-1.48863649e+00 7.19846964e-01 2.08815426e-01 -1.86418399e-01
-6.57887816e-01 6.49536669e-01 4.15433049e-01 -7.01856554e-01
5.03854215e-01 -8.20216015e-02 -3.99962276e-01 1.67159170e-01
5.38948894e-01 3.27555180e-01 -2.49104109e-02 -3.71283203e-01
-2.29875848e-01 2.77863950e-01 -2.86130965e-01 3.38048488e-01
1.36185718e+00 3.57172787e-02 -1.92272604e-01 4.05541509e-01
9.09001172e-01 -5.39685667e-01 -1.41977155e+00 -4.27756310e-01
2.31940717e-01 -2.86815107e-01 1.13355152e-01 -6.42534733e-01
-1.43725026e+00 7.55217731e-01 1.10027134e+00 1.98319703e-01
1.24973726e+00 6.99135587e-02 1.44107535e-01 7.35107005e-01
-1.67768016e-01 -1.08501053e+00 5.74718833e-01 3.15401942e-01
7.10488737e-01 -1.77920306e+00 -8.96284878e-02 -5.51019132e-01
-6.80749357e-01 9.11701381e-01 1.08122993e+00 -1.39158025e-01
5.22864699e-01 5.95145583e-01 3.19823325e-01 -2.35159665e-01
-4.53362584e-01 -4.26461667e-01 4.83795069e-02 1.11258364e+00
2.89183080e-01 2.46173237e-02 3.82345557e-01 2.21247435e-01
2.13977009e-01 -2.16993436e-01 4.60210770e-01 9.82971728e-01
-4.85316753e-01 -6.88839138e-01 -3.82952899e-01 6.07094646e-01
-1.39822021e-01 8.98777694e-02 -4.58396882e-01 6.18085742e-01
2.52184302e-01 7.29702830e-01 3.77105534e-01 -1.19222559e-01
1.86090991e-01 1.40611917e-01 3.30216318e-01 -6.44394696e-01
-2.40038291e-01 -8.57689977e-02 -5.06040454e-01 -5.22319257e-01
-6.14883840e-01 -6.02221012e-01 -1.09362793e+00 -2.73661930e-02
-5.80562949e-01 -4.61905673e-02 4.48603630e-01 6.68981016e-01
4.42511797e-01 4.64466065e-01 6.70788705e-01 -1.01603484e+00
-4.72290486e-01 -7.76834846e-01 -6.45697057e-01 5.12644172e-01
3.72498661e-01 -5.16472578e-01 -4.91684854e-01 5.06657958e-01] | [9.566758155822754, 1.8660024404525757] |
efdae121-bc97-41df-b90c-5e2d42f55e71 | kevin-a-knowledge-enhanced-validity-and | null | null | https://aclanthology.org/2022.argmining-1.9 | https://aclanthology.org/2022.argmining-1.9.pdf | KEViN: A Knowledge Enhanced Validity and Novelty Classifier for Arguments | The ArgMining 2022 Shared Task is concerned with predicting the validity and novelty of an inference for a given premise and conclusion pair. We propose two feed-forward network based models (KEViN1 and KEViN2), which combine features generated from several pretrained transformers and the WikiData knowledge graph. The transformers are used to predict entailment and semantic similarity, while WikiData is used to provide a semantic measure between concepts in the premise-conclusion pair. Our proposed models show significant improvement over RoBERTa, with KEViN1 outperforming KEViN2 and obtaining second rank on both subtasks (A and B) of the ArgMining 2022 Shared Task. | ['Nadin Kökciyan', 'Jeff Z. Pan', 'Björn Ross', 'Vaishak Belle', 'Sandrine Chausson', 'Xue Li', 'Ameer Saadat-Yazdi'] | null | null | null | null | argmining-acl-2022-10 | ['valnov'] | ['natural-language-processing'] | [-1.43748429e-02 5.69381893e-01 -2.60269552e-01 -6.65090084e-01
-3.73572558e-01 -3.50225061e-01 8.72981787e-01 4.03859943e-01
-3.94663930e-01 7.24746943e-01 4.06991631e-01 -3.58096719e-01
-8.92292082e-01 -1.07343388e+00 -7.43086457e-01 -2.04747152e-02
-6.00926913e-02 5.27463078e-01 2.66484231e-01 -2.32608333e-01
4.16112721e-01 1.27297327e-01 -1.40179086e+00 6.38214469e-01
8.86748195e-01 1.39887702e+00 -1.49611933e-02 4.22802180e-01
-2.66102672e-01 1.34338331e+00 -2.75486171e-01 -1.09529471e+00
1.07643083e-01 -3.38661015e-01 -1.41077840e+00 -7.83312559e-01
6.51112020e-01 -3.51271108e-02 -3.22505176e-01 1.06004715e+00
4.33230288e-02 4.11258519e-01 7.96575129e-01 -1.54629493e+00
-6.83485746e-01 1.18516171e+00 3.06699984e-02 4.82897311e-01
3.48232388e-01 -2.34991878e-01 1.80210555e+00 -1.31897521e+00
6.08480215e-01 1.14750528e+00 8.20821106e-01 6.61271289e-02
-6.77356899e-01 -5.81020296e-01 2.99232751e-01 1.06880939e+00
-1.29206645e+00 -5.27402818e-01 5.88176131e-01 -2.43172571e-01
1.55564427e+00 8.53565708e-02 6.24861240e-01 8.07336688e-01
-6.94520697e-02 7.85125911e-01 8.70956540e-01 -2.31945172e-01
1.76450521e-01 3.31013680e-01 4.26259071e-01 8.93246889e-01
4.07109231e-01 -4.18277159e-02 -8.98295641e-01 2.77104173e-02
3.23508769e-01 -6.95109889e-02 -7.67872706e-02 -1.27144068e-01
-1.19076455e+00 7.34418094e-01 6.68730617e-01 3.29517871e-01
-6.41587079e-01 1.74212009e-01 3.68160963e-01 5.83439469e-01
4.20191973e-01 9.77968514e-01 -7.48048723e-01 -1.59498304e-01
-8.92481327e-01 5.08943677e-01 9.11404252e-01 9.63500559e-01
8.73760223e-01 -3.36988151e-01 -3.45725924e-01 6.12313807e-01
3.16638798e-01 4.12725545e-02 5.88950574e-01 -1.03802085e+00
5.65278232e-01 9.03007865e-01 -8.37301388e-02 -1.30401111e+00
-4.41925168e-01 -6.77026629e-01 -2.93826967e-01 -3.26868862e-01
3.71414781e-01 -5.85400984e-02 -5.52034438e-01 1.62507761e+00
1.10222973e-01 3.62894356e-01 2.55772859e-01 4.72617567e-01
1.25675738e+00 5.20817518e-01 -1.95360426e-02 5.51746190e-02
1.18729782e+00 -1.23002255e+00 -6.09938860e-01 -3.72793019e-01
8.32297444e-01 -5.47520399e-01 7.46947825e-01 3.90213549e-01
-1.31195140e+00 -3.61700863e-01 -1.09254241e+00 -2.63203710e-01
-7.17527092e-01 -1.25680327e-01 5.99628925e-01 -4.55696173e-02
-1.02242494e+00 1.04908299e+00 -1.74505219e-01 -3.44828457e-01
4.90566880e-01 1.10372171e-01 -4.82534617e-01 -4.13179338e-01
-1.74300647e+00 1.61155975e+00 7.18433857e-01 8.46882612e-02
-8.51151168e-01 -1.03402710e+00 -1.12470293e+00 4.29804534e-01
7.37584412e-01 -7.99774826e-01 1.14330482e+00 -4.74099427e-01
-1.17331278e+00 8.06769609e-01 -8.78935084e-02 -7.71767378e-01
2.23950729e-01 -3.04533988e-01 -6.90626621e-01 4.41172905e-02
3.30873877e-01 5.04431546e-01 4.33358014e-01 -8.76241207e-01
-9.28238273e-01 -2.58574575e-01 3.27720553e-01 3.91020417e-01
-3.78776252e-01 -1.73172683e-01 -1.62499830e-01 -3.20282042e-01
1.45342186e-01 -2.51259387e-01 1.48672715e-01 -1.76261932e-01
-6.62115693e-01 -9.71472204e-01 6.24573112e-01 -6.86534405e-01
1.17542577e+00 -1.55389059e+00 -3.08805034e-02 5.73769093e-01
4.70405638e-01 3.44245285e-01 -1.71634600e-01 3.99572045e-01
-2.28215322e-01 1.94243214e-03 2.38928646e-01 -1.18239723e-01
9.37220082e-03 1.54998958e-01 -1.77964628e-01 -6.80892318e-02
2.06882894e-01 1.10420310e+00 -1.23644149e+00 -3.80589426e-01
2.77295485e-02 -3.07614058e-01 -2.88896292e-01 4.48106229e-01
-3.07608664e-01 -4.40354824e-01 -1.72411680e-01 2.81394035e-01
2.19707906e-01 -4.42902118e-01 2.74127275e-01 -6.88586235e-01
1.67613715e-01 1.07908082e+00 -8.80418539e-01 1.56683981e+00
-6.61541343e-01 4.99162257e-01 -4.72143531e-01 -1.37909722e+00
9.91007507e-01 1.45668432e-01 1.87881961e-02 -9.88890707e-01
1.17450438e-01 2.07511872e-01 8.20827484e-02 -7.50107348e-01
5.05277634e-01 -3.34681600e-01 1.37313962e-01 4.39692378e-01
5.24347305e-01 -2.51836866e-01 3.51880193e-01 7.74671853e-01
1.16770792e+00 1.17003940e-01 3.69571447e-01 -4.02958632e-01
4.83864754e-01 -5.31974584e-02 5.93315244e-01 8.37497711e-01
2.87121702e-02 1.09666467e-01 7.76546836e-01 -6.01833880e-01
-6.40895367e-01 -1.02029061e+00 2.83124685e-01 9.42825556e-01
1.07861690e-01 -6.16172850e-01 -2.24817708e-01 -1.17951608e+00
2.66882747e-01 1.32365906e+00 -9.77137625e-01 -6.05584919e-01
-3.69290054e-01 6.37122914e-02 7.75608420e-01 8.87845278e-01
6.83776617e-01 -8.01992774e-01 -2.39547953e-01 -5.09726144e-02
-4.80326146e-01 -1.11852670e+00 -1.95664808e-01 3.04527014e-01
-6.19086921e-01 -1.51063478e+00 -8.60296115e-02 -8.23848903e-01
2.13269576e-01 -4.21129279e-02 1.52456975e+00 1.13248667e-02
5.38928388e-03 1.99527070e-01 -1.92577288e-01 -4.69453812e-01
2.67032348e-02 -8.32797959e-02 -2.41306555e-02 -2.09344253e-01
6.47956848e-01 -5.60404480e-01 -3.56300145e-01 1.43894643e-01
-4.58968431e-01 9.09769088e-02 4.41417217e-01 7.48231173e-01
1.41992614e-01 2.04289258e-01 9.20713484e-01 -1.09011078e+00
7.90194809e-01 -8.39135826e-01 -1.17058061e-01 4.75560367e-01
-1.32758927e+00 1.23987392e-01 7.52169192e-01 7.60449246e-02
-1.21539772e+00 -6.99086010e-01 -2.46061850e-02 -2.95966238e-01
4.20251310e-01 1.16857004e+00 -1.31785095e-01 4.55033809e-01
7.07290053e-01 -7.34148920e-02 -4.79042381e-02 -3.44899833e-01
6.40277267e-01 1.40304819e-01 7.00922430e-01 -5.47366738e-01
7.13258445e-01 3.65838334e-02 -9.22229216e-02 -3.83384407e-01
-1.63599396e+00 -5.02824485e-01 -2.89892435e-01 -6.44066408e-02
4.07193810e-01 -8.67633820e-01 -7.72130966e-01 1.77263692e-01
-1.05340362e+00 -7.79921710e-02 -6.39564216e-01 6.56995833e-01
-3.18191379e-01 3.95154096e-02 -3.07877928e-01 -3.18800777e-01
-4.44394380e-01 -4.15828615e-01 1.95559472e-01 -3.61313927e-03
-5.35458028e-01 -1.38112104e+00 -6.13517761e-02 4.27824169e-01
4.48945075e-01 -1.55695662e-01 1.35522521e+00 -1.40362310e+00
-1.61255047e-01 -4.08545494e-01 -5.04807711e-01 5.21680653e-01
5.71740717e-02 -2.02907130e-01 -7.63558865e-01 1.48318678e-01
-8.77152383e-02 -7.20885456e-01 1.06372273e+00 2.33593866e-01
1.16514277e+00 -6.58792734e-01 -2.04518855e-01 2.05903932e-01
1.16460848e+00 -2.10492089e-01 6.30039632e-01 1.73447832e-01
5.05284190e-01 8.12159956e-01 5.64631402e-01 1.87425762e-01
8.59190702e-01 4.39317822e-01 4.01347846e-01 2.66632408e-01
-2.74036467e-01 -6.22781873e-01 1.42429903e-01 7.55139112e-01
1.68284718e-02 -1.42929047e-01 -8.15763235e-01 1.19770455e+00
-2.01463366e+00 -9.49684083e-01 -5.01299575e-02 1.92653918e+00
9.28614318e-01 3.50903422e-01 -6.67278767e-02 7.21934736e-02
5.32535553e-01 7.03289881e-02 -5.55557370e-01 -3.84184629e-01
-9.75576192e-02 3.37477833e-01 -9.27893147e-02 5.14131725e-01
-5.62177479e-01 9.29589570e-01 6.43443108e+00 9.18684900e-01
-4.64687526e-01 -6.99507520e-02 2.96374708e-01 -1.46367058e-01
-6.63981974e-01 3.28947932e-01 -5.60888827e-01 7.88137838e-02
9.33757544e-01 -5.23344278e-01 1.34156778e-01 7.40497291e-01
-2.00118780e-01 -1.16018966e-01 -1.57088649e+00 5.83099782e-01
2.50228941e-01 -1.75435543e+00 3.01633775e-01 -4.55519021e-01
6.52055323e-01 5.68272993e-02 -3.30055766e-02 6.58980191e-01
4.91928488e-01 -1.13121998e+00 5.75262904e-01 6.31089389e-01
4.10618037e-01 -8.37186635e-01 9.92454231e-01 1.88407347e-01
-1.01127422e+00 6.74809655e-03 -2.01288521e-01 -3.56823325e-01
-9.41728279e-02 7.48592675e-01 -9.84876692e-01 6.69263124e-01
6.90922260e-01 1.24540985e+00 -4.18883055e-01 7.41035938e-01
-8.68729055e-01 6.46366358e-01 1.58289112e-02 -2.09370941e-01
3.65388393e-01 8.33032355e-02 4.57182616e-01 1.14369547e+00
9.08964574e-02 -1.01719640e-01 -1.30060673e-01 1.27826869e+00
-6.03578985e-01 -6.55258670e-02 -5.30420780e-01 -2.49437839e-01
8.48286092e-01 1.34182274e+00 -1.91864833e-01 -8.04431915e-01
-2.68110961e-01 6.81904614e-01 7.30183065e-01 1.86708361e-01
-5.96466601e-01 -7.81150699e-01 5.95879436e-01 -4.55505289e-02
3.72218937e-01 4.03504729e-01 -3.32180858e-01 -1.05040693e+00
-1.44204181e-02 -4.29108977e-01 6.85077727e-01 -9.25608695e-01
-1.59893906e+00 2.75481999e-01 1.17159754e-01 -5.72616696e-01
-2.58974761e-01 -6.50600076e-01 -1.10310984e+00 8.93358052e-01
-1.75642323e+00 -1.14477277e+00 -2.65330404e-01 7.80303419e-01
3.03011119e-01 -2.83469081e-01 8.22010517e-01 1.32825360e-01
-3.69954705e-01 6.72213554e-01 -1.99855074e-01 1.25268489e-01
5.19784033e-01 -1.37931883e+00 4.14108276e-01 6.56576335e-01
8.59248731e-03 9.99866486e-01 5.51379979e-01 -7.69035935e-01
-9.45275724e-01 -1.24932075e+00 1.57880282e+00 -2.53627270e-01
7.99075365e-01 1.51849106e-01 -7.28700936e-01 9.05312419e-01
6.43797293e-02 -1.33426666e-01 8.09657097e-01 7.79628873e-01
-7.27255166e-01 -1.36874557e-01 -1.22356248e+00 3.27875584e-01
1.11188364e+00 -6.85491681e-01 -1.24530840e+00 4.02308732e-01
8.32247257e-01 -5.70855066e-02 -1.17517138e+00 2.41348162e-01
5.60101092e-01 -8.03329885e-01 8.72695386e-01 -1.07154667e+00
9.12023723e-01 -7.72547396e-03 -8.22608173e-02 -1.47536588e+00
-5.64380348e-01 -2.23038360e-01 -5.24020135e-01 9.26918626e-01
1.01657748e+00 -5.32528102e-01 6.81412220e-01 6.33825064e-01
-2.42260993e-01 -8.50107431e-01 -6.76913023e-01 -9.67503369e-01
1.77553063e-03 -4.02236611e-01 6.61866307e-01 1.22201800e+00
5.90966582e-01 8.93051028e-01 -1.72821045e-01 -8.81081820e-02
4.10984129e-01 2.15458348e-01 1.88394487e-01 -1.32827079e+00
-2.57862695e-02 -3.37573379e-01 -3.79204631e-01 -9.71222341e-01
2.67060041e-01 -1.37251556e+00 -1.39907748e-01 -1.99729466e+00
3.96942407e-01 -3.39577228e-01 -5.28392136e-01 7.74603307e-01
-3.49224627e-01 7.54489154e-02 -1.53202966e-01 -1.47672696e-02
-9.11303997e-01 6.13922000e-01 8.32931995e-01 -1.13065414e-01
1.39808849e-01 -5.53072542e-02 -1.00601470e+00 9.31657076e-01
6.54763997e-01 -3.16615105e-01 -6.92061067e-01 -5.22078037e-01
8.76653850e-01 -2.54945576e-01 5.26591539e-01 -7.60659099e-01
5.30636370e-01 -7.09930584e-02 3.15221012e-01 -5.31327307e-01
2.86302328e-01 -5.58536172e-01 -3.44678819e-01 3.55575532e-01
-8.91988575e-01 1.72747448e-01 -6.27909824e-02 6.09904468e-01
-2.58910596e-01 -3.98612589e-01 3.76426518e-01 -2.23046646e-01
-7.98681736e-01 1.82872742e-01 -1.71912059e-01 4.93933648e-01
7.41832733e-01 -6.12786599e-02 -6.02219045e-01 -3.85811567e-01
-7.79480338e-01 4.29961383e-01 -1.58832818e-01 4.99396354e-01
1.15778410e+00 -1.34602034e+00 -8.58221471e-01 -6.05557673e-02
3.55274379e-01 8.55139717e-02 7.40745068e-02 9.37062740e-01
1.33620286e-02 6.15660727e-01 -8.52695033e-02 7.58715943e-02
-9.49439943e-01 3.47418219e-01 3.51083875e-01 -5.93310177e-01
-3.75705808e-01 1.43503523e+00 -1.04030557e-01 -6.78757608e-01
1.54794499e-01 -3.26218635e-01 -4.02952790e-01 -4.86802198e-02
4.51748580e-01 7.91366160e-01 4.58279371e-01 -1.93361700e-01
-5.37756741e-01 -1.81966051e-01 -5.32664239e-01 1.16290830e-01
1.44903100e+00 1.43927515e-01 -4.27616626e-01 3.43057156e-01
1.26068449e+00 -4.05655026e-01 -4.20939296e-01 -8.29866886e-01
3.73251826e-01 -1.92041323e-01 2.10754782e-01 -1.41189778e+00
-7.99928486e-01 6.33248925e-01 -1.31782904e-01 6.77161738e-02
5.66183329e-01 9.42049474e-02 8.97372842e-01 8.39080274e-01
5.55782057e-02 -1.33663476e+00 2.07660899e-01 7.57563651e-01
9.61834133e-01 -1.04673517e+00 8.38373974e-02 -3.01232815e-01
-7.07792580e-01 9.93617415e-01 8.62283409e-01 -1.19316317e-02
5.89865386e-01 -1.87218949e-01 -4.28328097e-01 -7.66555369e-01
-1.09438145e+00 -7.63390884e-02 6.80525720e-01 4.27860349e-01
3.81696343e-01 -2.91190315e-02 -2.93391347e-01 8.40032339e-01
-3.36916387e-01 -5.92714474e-02 1.63726702e-01 7.90136039e-01
-4.82986063e-01 -6.28330946e-01 3.08054298e-01 1.08813810e+00
1.10207191e-02 -4.54005748e-01 -7.12592781e-01 5.07432878e-01
1.02168866e-01 1.18668699e+00 -8.91115293e-02 -7.05594718e-01
3.77890050e-01 1.90029725e-01 4.87345189e-01 -6.56818569e-01
-7.36657202e-01 -8.45943511e-01 8.07403028e-01 -7.11728692e-01
-1.92703098e-01 -2.13925868e-01 -1.10771596e+00 -4.01607215e-01
-6.48522437e-01 4.29918587e-01 4.86459523e-01 1.49551642e+00
4.78380769e-01 6.76890135e-01 3.59791428e-01 5.39908074e-02
-7.20089912e-01 -1.15784383e+00 -6.69281244e-01 5.78441024e-01
-5.45643307e-02 -6.55653059e-01 -2.58409947e-01 -3.48367393e-01] | [9.392732620239258, 8.145943641662598] |
2d9fe227-af9c-4363-8fc0-4dd89d877e6b | adaptive-loose-optimization-for-robust | 2305.03971 | null | https://arxiv.org/abs/2305.03971v2 | https://arxiv.org/pdf/2305.03971v2.pdf | Adaptive loose optimization for robust question answering | Question answering methods are well-known for leveraging data bias, such as the language prior in visual question answering and the position bias in machine reading comprehension (extractive question answering). Current debiasing methods often come at the cost of significant in-distribution performance to achieve favorable out-of-distribution generalizability, while non-debiasing methods sacrifice a considerable amount of out-of-distribution performance in order to obtain high in-distribution performance. Therefore, it is challenging for them to deal with the complicated changing real-world situations. In this paper, we propose a simple yet effective novel loss function with adaptive loose optimization, which seeks to make the best of both worlds for question answering. Our main technical contribution is to reduce the loss adaptively according to the ratio between the previous and current optimization state on mini-batch training data. This loose optimization can be used to prevent non-debiasing methods from overlearning data bias while enabling debiasing methods to maintain slight bias learning. Experiments on the visual question answering datasets, including VQA v2, VQA-CP v1, VQA-CP v2, GQA-OOD, and the extractive question answering dataset SQuAD demonstrate that our approach enables QA methods to obtain state-of-the-art in- and out-of-distribution performance in most cases. The source code has been released publicly in \url{https://github.com/reml-group/ALO}. | ['Jun Liu', 'Ting Han', 'Min Hu', 'Dechen Kong', 'Zewei Wang', 'Pinghui Wang', 'Jie Ma'] | 2023-05-06 | null | null | null | null | ['reading-comprehension', 'machine-reading-comprehension'] | ['natural-language-processing', 'natural-language-processing'] | [ 9.41124279e-03 1.66173995e-01 -9.95929465e-02 -5.12993991e-01
-1.29950607e+00 -7.06253469e-01 3.04032207e-01 1.84306696e-01
-4.20931399e-01 5.38122833e-01 9.77280065e-02 -6.49030805e-01
-1.02446266e-01 -6.26050472e-01 -8.22118223e-01 -6.28773928e-01
4.98532504e-01 4.25455123e-01 3.72758299e-01 -4.22214240e-01
3.00271392e-01 2.57020723e-03 -1.35995781e+00 3.55806887e-01
1.31543386e+00 1.10857344e+00 1.14575818e-01 9.00151074e-01
-4.38961118e-01 1.08174181e+00 -7.83834100e-01 -6.72713518e-01
9.95590463e-02 -5.92133939e-01 -9.93589342e-01 -3.74398232e-01
1.15092754e+00 -4.75801557e-01 -3.38967413e-01 1.06602454e+00
8.39932501e-01 2.49976590e-01 5.89500010e-01 -1.11907005e+00
-1.11792886e+00 2.90187150e-01 -5.96339881e-01 5.56293368e-01
2.47280791e-01 3.08060527e-01 1.22214139e+00 -6.94049418e-01
3.26922834e-01 1.25826871e+00 3.33528370e-01 8.51786315e-01
-1.24228787e+00 -4.99650657e-01 2.20525458e-01 6.66830122e-01
-1.02688539e+00 -3.23513448e-01 8.17623019e-01 -2.98360646e-01
6.72385931e-01 4.57715601e-01 2.04863697e-01 1.08729649e+00
-3.93114649e-02 1.25787175e+00 1.17415667e+00 -2.67061472e-01
2.74113268e-01 -3.41345333e-02 6.14725173e-01 7.73164928e-01
-1.28089532e-01 -3.52175921e-01 -7.18186796e-01 -1.70210093e-01
2.83595115e-01 -3.85009855e-01 -6.45279884e-01 -4.05616224e-01
-7.97506571e-01 8.05756092e-01 7.49726117e-01 1.54593782e-02
-1.89039707e-01 2.26116151e-01 3.39886844e-01 6.00810587e-01
4.93430436e-01 4.56307203e-01 -4.62912530e-01 -3.64368200e-01
-7.68297791e-01 5.98218262e-01 6.12820983e-01 7.45094836e-01
7.35611439e-01 -2.60759175e-01 -7.81960011e-01 9.91413414e-01
-8.62480141e-03 6.14104033e-01 5.41304708e-01 -1.00236440e+00
7.90303349e-01 5.78591526e-01 9.05398726e-02 -8.33010733e-01
-2.17976794e-01 -3.83186519e-01 -7.90368974e-01 1.26434162e-01
1.01622093e+00 3.00220475e-02 -1.08665979e+00 1.96305692e+00
5.19351304e-01 -3.89223427e-01 -1.45246863e-01 9.98744905e-01
8.49984169e-01 6.38918936e-01 7.34345466e-02 1.94773450e-01
1.75082338e+00 -1.12958324e+00 -7.01089025e-01 -3.86593997e-01
5.82686007e-01 -5.94565988e-01 1.87000573e+00 3.91578764e-01
-1.11930072e+00 -5.88563383e-01 -8.90939772e-01 -7.59354591e-01
-1.24162860e-01 -9.73509625e-02 1.77235976e-01 6.45476520e-01
-7.03691602e-01 1.77521229e-01 -6.24500453e-01 6.78581893e-02
7.72550106e-01 -2.08245497e-02 -1.61774561e-01 -3.32867533e-01
-1.09673059e+00 9.16737974e-01 2.59670336e-02 -1.45476118e-01
-8.57764363e-01 -1.13508391e+00 -5.35139501e-01 1.93617359e-01
7.45911062e-01 -9.19645905e-01 1.51184344e+00 -9.97955739e-01
-1.39792836e+00 8.97592902e-01 -2.90083945e-01 -3.83704454e-01
6.89321220e-01 -6.89315677e-01 1.39442012e-01 3.04900885e-01
5.02198301e-02 6.73367441e-01 1.04903650e+00 -1.05603480e+00
-2.46659070e-01 -7.31640697e-01 4.40541744e-01 3.97738099e-01
-4.02423888e-01 -4.38973516e-01 -4.56330091e-01 -5.77835619e-01
-1.14933416e-01 -6.22355402e-01 1.70187533e-01 2.20730707e-01
-3.09038460e-01 -5.44268608e-01 6.28360271e-01 -1.02649760e+00
1.21176600e+00 -2.14350724e+00 2.12301522e-01 -2.15097368e-01
4.10989463e-01 4.10118192e-01 -2.94646919e-01 1.81344122e-01
1.96465358e-01 -1.02319352e-01 -2.82315284e-01 -3.04519773e-01
1.77275911e-01 1.95587531e-01 -5.28395116e-01 2.87182987e-01
2.08628535e-01 1.08336282e+00 -9.14929867e-01 -5.35272717e-01
-1.04806535e-01 1.77058056e-01 -6.72039330e-01 6.16797090e-01
-6.30597055e-01 3.58508378e-01 -3.77030581e-01 5.38445055e-01
7.82967508e-01 -4.27243501e-01 -1.86622009e-01 -1.19450226e-01
4.81881291e-01 4.71857697e-01 -7.75582850e-01 1.63215959e+00
-4.22879785e-01 5.27751029e-01 2.78796464e-01 -8.67878437e-01
6.09297037e-01 6.93668798e-02 -1.56575471e-01 -1.08794403e+00
6.90286309e-02 1.20413455e-03 -2.49993689e-02 -6.83319092e-01
6.06456816e-01 -1.67217225e-01 1.47773087e-01 2.93986708e-01
1.09461866e-01 -3.28930527e-01 1.33391380e-01 5.21471977e-01
9.44745660e-01 -7.35074878e-02 1.81473434e-01 -2.97619343e-01
4.75410551e-01 -1.67075366e-01 1.36021808e-01 9.27264273e-01
-4.62698996e-01 7.53973961e-01 8.49984348e-01 -4.44139875e-02
-9.67813730e-01 -1.16477537e+00 -3.36469524e-02 1.52595973e+00
1.24784105e-01 -1.62697792e-01 -9.29591954e-01 -1.10482919e+00
6.67028502e-02 1.01502347e+00 -7.79281199e-01 -2.82836765e-01
-5.37875891e-01 -5.67757010e-01 6.45372987e-01 4.71562207e-01
5.18390656e-01 -9.41914558e-01 -4.60685521e-01 -5.28849140e-02
-4.39831167e-01 -6.93455994e-01 -6.93856657e-01 7.21750362e-03
-6.37194633e-01 -1.00053787e+00 -1.12736654e+00 -5.66765487e-01
5.32120705e-01 3.14313412e-01 1.39936614e+00 2.25593612e-01
5.78201823e-02 4.89157587e-01 -4.08780426e-01 -4.68826830e-01
-2.79147804e-01 4.49719101e-01 -5.67027986e-01 -1.11185439e-01
4.12273586e-01 -3.49210709e-01 -9.08448458e-01 3.64207208e-01
-1.18351126e+00 -1.67939946e-01 1.69814035e-01 1.12760174e+00
4.32334155e-01 -5.50865889e-01 8.25993836e-01 -8.52967680e-01
8.07587266e-01 -5.78550279e-01 -5.33515453e-01 4.83796149e-01
-6.24603570e-01 2.33567134e-01 5.25314033e-01 -5.32528222e-01
-1.11019480e+00 -7.10953832e-01 -4.38560039e-01 -3.23012173e-01
7.95808807e-02 3.02948445e-01 -3.70917678e-01 2.47498080e-01
9.50116754e-01 1.42730042e-01 1.19065180e-01 -5.83705306e-01
8.31610680e-01 5.98196268e-01 6.25353873e-01 -7.99234271e-01
5.85393488e-01 4.12306249e-01 -3.86795163e-01 -5.84598005e-01
-1.32745159e+00 -4.86506641e-01 -4.98548821e-02 -7.47232139e-02
8.40972900e-01 -6.21556640e-01 -8.11058998e-01 6.58872724e-01
-9.50973511e-01 -5.38541913e-01 -3.94955754e-01 -8.63474533e-02
-3.57563496e-01 6.12630725e-01 -4.36732769e-01 -7.37402439e-01
-6.47700369e-01 -9.46534395e-01 8.97500336e-01 3.47506851e-01
-1.35879472e-01 -7.51564801e-01 1.98800176e-01 1.02449989e+00
5.67900538e-01 -2.87296087e-01 1.20353687e+00 -6.39561117e-01
-5.89505136e-01 2.07432002e-01 -4.32849795e-01 7.19485044e-01
-1.16067223e-01 -4.22215074e-01 -1.09143686e+00 -4.92463470e-01
8.98148790e-02 -8.45845520e-01 1.09458935e+00 2.40821540e-01
1.48491454e+00 -2.88576275e-01 2.97926396e-01 5.03793776e-01
1.10818028e+00 -1.72465995e-01 8.32723320e-01 3.23587805e-01
8.36157143e-01 7.04942465e-01 5.41060090e-01 2.09085882e-01
6.10881925e-01 5.99635422e-01 6.72958851e-01 1.88293457e-01
-3.86315137e-01 -3.81319731e-01 2.65044898e-01 6.71165287e-01
2.85662919e-01 -5.26901424e-01 -9.13277209e-01 8.49606574e-01
-1.72345531e+00 -7.77491510e-01 -1.26674831e-01 2.13099980e+00
1.20879114e+00 -7.79076740e-02 1.08268663e-01 -7.05655813e-02
2.94606090e-01 4.16127920e-01 -7.76229143e-01 -2.43473202e-01
-8.97180364e-02 2.75901884e-01 1.19569726e-01 6.17490411e-01
-8.09284031e-01 8.22716117e-01 5.03988647e+00 1.24778414e+00
-8.20282578e-01 3.01792920e-01 7.58835018e-01 -2.00999171e-01
-5.99838674e-01 -7.16909096e-02 -6.31459534e-01 5.44016898e-01
7.56810427e-01 1.34206653e-01 4.24014568e-01 7.70415068e-01
-4.37046826e-01 -2.22405925e-01 -9.61036623e-01 9.83756244e-01
2.05935135e-01 -1.11670601e+00 1.24623470e-01 -3.42928946e-01
5.00397086e-01 5.18919760e-03 3.44550043e-01 5.48666060e-01
2.46699713e-02 -1.04817164e+00 8.42495084e-01 4.27925259e-01
6.46151423e-01 -5.66486657e-01 6.03050590e-01 4.28721011e-01
-5.23101926e-01 1.35230506e-02 -4.15939689e-01 1.15391672e-01
2.56111234e-01 7.49975741e-01 -5.22237241e-01 4.68183488e-01
9.57613051e-01 7.12597966e-02 -8.11606824e-01 7.75605381e-01
-4.66778129e-01 9.51776028e-01 -2.14556009e-01 -2.48103440e-01
1.68260723e-01 1.75269565e-03 6.17381752e-01 8.70742500e-01
2.90473457e-02 5.20930998e-02 -2.93343633e-01 8.17380846e-01
-4.54588860e-01 2.28387967e-01 -4.69913781e-02 1.68449879e-01
3.65201175e-01 9.44551229e-01 -4.00140472e-02 -4.54638898e-01
-2.71000773e-01 9.76917326e-01 8.68239641e-01 4.19980407e-01
-8.44955921e-01 -4.50645685e-01 5.38299084e-01 1.52168959e-01
5.55544496e-01 -3.69317420e-02 -3.28198940e-01 -1.36248326e+00
4.20955747e-01 -1.47067654e+00 6.82141662e-01 -8.53470564e-01
-1.70976067e+00 4.48952973e-01 4.54339683e-02 -6.18810713e-01
-3.67167294e-02 -7.44951546e-01 -2.46826723e-01 9.42310691e-01
-1.58395660e+00 -9.69195426e-01 -4.75130618e-01 6.00789309e-01
4.29578751e-01 -8.55049305e-03 3.65033716e-01 3.79440606e-01
-4.29247379e-01 1.00539041e+00 1.27216890e-01 -6.54217601e-03
9.89186823e-01 -1.54528093e+00 2.03101322e-01 6.55038655e-01
1.48381427e-01 4.63870376e-01 7.68958390e-01 -2.64866054e-01
-1.19097173e+00 -5.89114547e-01 7.10019886e-01 -8.85717630e-01
5.72665453e-01 -3.84864658e-01 -1.51506329e+00 5.54362476e-01
4.12218809e-01 -1.07451260e-01 5.01076221e-01 2.44143665e-01
-9.21093702e-01 -2.43071243e-01 -1.10380280e+00 6.44984245e-01
7.57415831e-01 -7.91153431e-01 -8.95545125e-01 2.15330556e-01
7.08769083e-01 -6.23510301e-01 -6.34960651e-01 2.97364354e-01
3.29189271e-01 -1.04949737e+00 9.42349195e-01 -7.80300081e-01
5.42612135e-01 -2.11515337e-01 -2.26790532e-01 -1.21144414e+00
-4.77152206e-02 -5.50224602e-01 -3.78115892e-01 1.33786368e+00
2.49913856e-01 -7.30490863e-01 8.66806924e-01 5.93131542e-01
4.92803715e-02 -9.81376886e-01 -1.23027134e+00 -5.38399577e-01
5.43157041e-01 -3.48209590e-01 5.39364755e-01 7.27831781e-01
-3.99050683e-01 7.01709867e-01 -3.77473831e-01 6.83031157e-02
6.17574871e-01 7.66167492e-02 1.05164754e+00 -5.92721641e-01
-5.58109581e-01 -4.33463901e-01 3.26343067e-02 -1.84789240e+00
-1.49751334e-02 -8.46697032e-01 3.26371081e-02 -1.44331169e+00
3.49516392e-01 -3.02300125e-01 -2.01113552e-01 4.71093118e-01
-8.60185385e-01 1.57068297e-02 3.30879688e-01 1.05999440e-01
-7.15972722e-01 9.20288265e-01 1.55485535e+00 -2.83327490e-01
3.62674817e-02 -1.53149456e-01 -8.71816993e-01 3.99131387e-01
6.98950768e-01 -4.65975344e-01 -6.36411369e-01 -7.51283228e-01
3.48618656e-01 -8.60315859e-02 6.70335054e-01 -5.94722271e-01
-2.84218732e-02 1.32683679e-01 8.51305947e-02 -6.00376308e-01
3.08899760e-01 -4.07298714e-01 -6.62851930e-01 1.46034747e-01
-4.35211778e-01 1.20991826e-01 1.68759391e-01 5.34036398e-01
-3.38821262e-01 -4.12625372e-01 8.62529457e-01 1.19351558e-01
-5.06220222e-01 1.82129711e-01 3.83939855e-02 9.32416558e-01
4.89774406e-01 2.71690696e-01 -9.95033622e-01 -7.67250717e-01
-2.91769475e-01 5.78628361e-01 2.65493363e-01 4.05590892e-01
4.01409596e-01 -1.04001141e+00 -8.84125352e-01 -5.43410666e-02
3.35537195e-01 2.05499426e-01 8.79917324e-01 8.32823813e-01
-5.70939481e-01 1.45465583e-01 -2.99449712e-02 -6.22852743e-01
-1.28147340e+00 6.07292891e-01 4.22664464e-01 -4.69976783e-01
-3.61778051e-01 1.22863472e+00 4.60217535e-01 -5.96765280e-01
2.93979079e-01 -3.12404454e-01 1.13698110e-01 5.61983325e-02
6.81080163e-01 4.59681183e-01 1.95436344e-01 -2.25262657e-01
-1.90675467e-01 3.34520578e-01 -3.17091703e-01 -1.00381464e-01
8.10520053e-01 -1.87951475e-01 -2.62087677e-02 1.75881445e-01
1.27838099e+00 1.94845632e-01 -1.37033784e+00 -2.94887275e-01
-2.24998966e-01 -5.34861147e-01 -4.17975709e-02 -1.01754439e+00
-8.12666237e-01 1.24910605e+00 6.23482525e-01 2.97334373e-01
1.14750457e+00 3.25219989e-01 9.69524026e-01 3.88384342e-01
6.07246086e-02 -7.92572141e-01 3.97586197e-01 4.50718164e-01
1.12374997e+00 -1.30072677e+00 -1.40897349e-01 -9.23825130e-02
-7.63875425e-01 5.68699181e-01 8.65727663e-01 9.67434142e-03
3.11754525e-01 -3.37317407e-01 4.89138424e-01 -2.48798445e-01
-7.37416983e-01 -1.06654897e-01 6.57768905e-01 4.86559391e-01
2.63089001e-01 -1.19224144e-02 -1.52881354e-01 6.03930414e-01
-3.65202963e-01 -3.08286965e-01 9.09122601e-02 7.38276839e-01
-2.87437737e-01 -1.03021383e+00 -3.97838563e-01 3.96228641e-01
-6.61878407e-01 -2.36722007e-01 -2.57457405e-01 8.65333736e-01
-1.89560354e-01 8.29723954e-01 -6.36071106e-03 2.98952740e-02
5.50805092e-01 2.49451339e-01 6.67777002e-01 -2.14052320e-01
-5.91277659e-01 -4.74295378e-01 1.59436867e-01 -5.36981940e-01
-1.02560848e-01 -3.09795141e-01 -9.73372519e-01 -3.93325418e-01
-4.03888524e-01 5.89657314e-02 3.06274503e-01 7.60391414e-01
4.96302992e-01 4.06807840e-01 7.56936905e-04 -1.53133020e-01
-1.20232117e+00 -1.12122595e+00 -3.25975925e-01 8.36426079e-01
6.42194271e-01 -4.44799840e-01 -5.32975256e-01 -2.78716385e-01] | [11.243130683898926, 8.071991920471191] |
dd047223-9163-4fdd-bb1a-7aa112cb2f84 | cloudbrain-reconai-an-online-platform-for-mri | 2212.01878 | null | https://arxiv.org/abs/2212.01878v1 | https://arxiv.org/pdf/2212.01878v1.pdf | CloudBrain-ReconAI: An Online Platform for MRI Reconstruction and Image Quality Evaluation | Efficient collaboration between engineers and radiologists is important for image reconstruction algorithm development and image quality evaluation in magnetic resonance imaging (MRI). Here, we develop CloudBrain-ReconAI, an online cloud computing platform, for algorithm deployment, fast and blind reader study. This platform supports online image reconstruction using state-of-the-art artificial intelligence and compressed sensing algorithms with applications to fast imaging and high-resolution diffusion imaging. Through visiting the website, radiologists can easily score and mark the images. Then, automatic statistical analysis will be provided. CloudBrain-ReconAI is now open accessed at https://csrc.xmu.edu.cn/CloudBrain.html and will be continually improved to serve the MRI research community. | ['Xiaobo Qu', 'Di Guo', 'Jiyang Dong', 'Qing Hong', 'Jianzhong Lin', 'Taishan Kang', 'Jianjun Zhou', 'Liuhong Zhu', 'Biao Qu', 'Yu Hu', 'Zi Wang', 'Jiayu Li', 'Chen Qian', 'Yirong Zhou'] | 2022-12-04 | null | null | null | null | ['mri-reconstruction'] | ['computer-vision'] | [ 1.11206017e-01 -5.98107755e-01 -1.17887288e-01 -2.10128069e-01
-8.95737648e-01 -2.71203220e-01 -1.97355032e-01 -1.22919239e-01
-3.83738458e-01 4.04081076e-01 9.89628658e-02 -5.45121133e-01
-2.12642521e-01 -3.81238669e-01 -1.02461122e-01 -8.70025814e-01
-2.94630885e-01 6.20727122e-01 3.74533311e-02 4.32166755e-01
3.27025384e-01 6.50842130e-01 -1.05827200e+00 4.18236017e-01
9.47701335e-01 9.69230771e-01 1.02618408e+00 6.97324157e-01
6.23943377e-03 8.45409393e-01 -4.77657020e-02 -9.14385393e-02
2.07673490e-01 -2.65921324e-01 -8.58304918e-01 -8.09731930e-02
-1.62415892e-01 -3.85228992e-01 -4.82551485e-01 1.57675421e+00
1.05288029e+00 -3.24774355e-01 3.50396246e-01 -7.37892687e-01
-9.47177470e-01 5.19948721e-01 -7.55619764e-01 9.74258244e-01
1.84702575e-01 2.85977602e-01 2.80356109e-01 -1.15004694e+00
8.28980267e-01 3.87001812e-01 4.59897637e-01 2.76893407e-01
-5.57442665e-01 -7.66815603e-01 -6.27974808e-01 9.98208344e-01
-1.42374957e+00 -3.65687817e-01 8.07901323e-01 -5.11206567e-01
5.24504364e-01 6.87592864e-01 8.67785454e-01 5.70811987e-01
3.92940938e-01 1.01574671e+00 1.49478590e+00 -2.56536126e-01
2.86124647e-01 -3.82520080e-01 1.98101088e-01 6.19042039e-01
1.16143867e-01 -8.45607519e-02 -3.33075941e-01 -1.34009257e-01
9.62654412e-01 4.44026113e-01 -4.49151695e-01 8.94857496e-02
-1.60131729e+00 5.46699882e-01 6.20225847e-01 6.37735844e-01
-9.42012906e-01 -2.16345370e-01 2.64501363e-01 1.70956299e-01
2.82513916e-01 1.11998484e-01 1.08944178e-02 -9.94754881e-02
-1.01592207e+00 -4.22501154e-02 1.01819292e-01 6.25588059e-01
1.69650018e-01 -1.09808199e-01 7.22664073e-02 1.09129155e+00
1.15650885e-01 8.08924615e-01 1.11106229e+00 -1.34670222e+00
-2.69128442e-01 -5.22397608e-02 -3.26540500e-01 -1.03555560e+00
-5.21895230e-01 -7.31207013e-01 -1.17811251e+00 6.47076964e-02
-2.12788999e-01 4.69762646e-02 -6.52504027e-01 7.58843124e-01
2.40156129e-01 3.00219625e-01 -3.89377028e-01 1.39435542e+00
1.03335917e+00 2.13031456e-01 -3.36425215e-01 -4.42545295e-01
1.80886877e+00 -1.10784352e+00 -1.08900571e+00 -2.18411684e-02
3.46218109e-01 -7.63911784e-01 9.82078016e-01 6.37124956e-01
-1.29770303e+00 -1.33987874e-01 -8.39153588e-01 1.48844436e-01
1.32163420e-01 2.46636674e-01 7.84642994e-01 3.60595435e-01
-1.10644734e+00 2.39662096e-01 -1.33538473e+00 8.08075964e-02
9.07398045e-01 2.22788274e-01 -3.28603864e-01 -7.76663780e-01
-9.78236854e-01 9.49020922e-01 -1.46666139e-01 6.65736869e-02
-6.51284218e-01 -8.12984109e-01 -4.26618457e-01 -5.34553051e-01
-3.05257291e-02 -6.94489658e-01 1.24189472e+00 -6.80319428e-01
-1.08172631e+00 1.17624640e+00 -1.27156362e-01 -3.80247295e-01
3.97257715e-01 1.23214796e-01 -6.97658956e-01 8.63048017e-01
2.79975235e-01 3.17238420e-01 6.27723455e-01 -7.46649981e-01
-3.21591683e-02 -9.14791226e-01 -7.76400566e-01 1.42595246e-01
-7.33055919e-02 6.33068681e-01 -5.55151641e-01 -7.43463933e-01
6.04608357e-01 -9.66877401e-01 -3.36311370e-01 3.34511966e-01
-2.54776359e-01 1.77158386e-01 6.05793536e-01 -1.35046816e+00
9.86648142e-01 -2.09148264e+00 -2.64933288e-01 2.78444588e-01
6.42360151e-01 2.27499500e-01 3.52199189e-02 -1.77948847e-01
-4.91172880e-01 -1.74837679e-01 -2.96770573e-01 1.31880581e-01
-6.61728442e-01 -2.25485131e-01 4.42014128e-01 8.98542881e-01
-4.82886970e-01 9.82520878e-01 -7.33255446e-01 -7.35876739e-01
1.50841251e-01 3.46405149e-01 -3.07998717e-01 1.64682969e-01
7.38705218e-01 8.69444847e-01 -4.99643236e-01 9.50167060e-01
9.73692834e-01 -7.05651581e-01 1.18961833e-01 -2.41677523e-01
-6.38510734e-02 -2.10525379e-01 -9.89729583e-01 1.99559975e+00
-2.43054882e-01 5.33076942e-01 4.12746072e-01 -8.93904328e-01
5.46738923e-01 4.82692659e-01 1.08074188e+00 -7.74019480e-01
1.16269872e-01 3.74024063e-01 3.49289477e-01 -1.02898836e+00
-1.62081383e-02 8.36302415e-02 6.89470053e-01 9.29850817e-01
-3.94170463e-01 1.70230344e-01 4.77631576e-02 4.13311511e-01
1.14744794e+00 -6.18988335e-01 1.32170588e-01 -1.68298498e-01
3.08219612e-01 2.70790029e-02 3.58720362e-01 7.46297657e-01
-8.08198214e-01 8.49644721e-01 -2.62599826e-01 -3.53887647e-01
-1.13088572e+00 -1.18074369e+00 -7.16332614e-01 6.74882412e-01
-7.18075261e-02 -2.43846830e-02 -9.58159804e-01 -1.64749473e-01
-3.18444997e-01 1.44419342e-01 -2.44601533e-01 2.20515490e-01
-3.22595179e-01 -8.46366286e-01 1.29110962e-01 2.18372494e-01
4.49373901e-01 -1.11030161e+00 -5.10757029e-01 1.92513108e-01
-6.32346272e-01 -9.58637536e-01 -7.25055516e-01 -4.01797622e-01
-9.75739956e-01 -9.47192550e-01 -1.31715477e+00 -8.03572595e-01
7.50125766e-01 7.52789736e-01 8.68418694e-01 3.97660881e-01
-1.04758775e+00 2.95954227e-01 -4.47251201e-01 -2.64251918e-01
-8.50753039e-02 -3.55683625e-01 9.22545195e-02 -1.69670075e-01
3.08313593e-02 -7.37103820e-01 -1.31620824e+00 1.44857019e-01
-8.53353441e-01 3.48438621e-01 7.20121741e-01 7.75523841e-01
1.17947781e+00 -2.95405388e-02 3.55572820e-01 -9.04006779e-01
5.74906826e-01 -4.99210060e-01 -3.13338012e-01 2.47929886e-01
-6.39191628e-01 -3.38706970e-01 2.35916808e-01 -2.28137448e-01
-4.84986603e-01 3.58500198e-04 -4.28098112e-01 -4.42610055e-01
-2.87334751e-02 6.89652920e-01 2.13927761e-01 -3.89715910e-01
5.91378570e-01 5.35910666e-01 3.17519456e-01 -4.57907379e-01
1.76244497e-01 1.11062598e+00 9.84594524e-01 9.33515280e-02
1.32962689e-01 5.87012291e-01 -5.35262346e-01 -6.80096090e-01
-3.79242212e-01 -6.91440165e-01 -2.84537435e-01 -5.43038249e-01
8.26093733e-01 -8.61415803e-01 -3.57260704e-01 5.65649509e-01
-8.21137607e-01 1.35000870e-02 7.69905299e-02 8.87130558e-01
-4.84939873e-01 3.78281683e-01 -7.77276397e-01 -3.96899283e-01
-9.08856571e-01 -1.50234735e+00 6.61448896e-01 2.01423019e-01
7.58203939e-02 -6.83991849e-01 -4.46453243e-02 7.73691773e-01
7.73438931e-01 -6.63462132e-02 3.71858090e-01 -4.52395111e-01
-5.83019137e-01 -3.39328200e-01 -3.79178911e-01 1.20251589e-01
-5.60885295e-02 -2.41553277e-01 -6.40342593e-01 -3.18561077e-01
4.75295931e-01 -4.98550534e-02 5.26183188e-01 9.96450663e-01
2.04286814e+00 -1.72825068e-01 -2.43410319e-01 8.70732307e-01
1.26613307e+00 2.93199956e-01 9.08412933e-01 3.88749242e-01
4.21633601e-01 1.60477191e-01 1.71831146e-01 5.39470851e-01
4.01299179e-01 5.27720690e-01 1.07151493e-01 -2.83777386e-01
-4.12917107e-01 3.04941833e-01 -7.02849180e-02 1.69951761e+00
-1.01304919e-01 7.05818772e-01 -1.45089400e+00 3.85096788e-01
-1.39825606e+00 -9.52608287e-01 -4.14495051e-01 1.65568137e+00
8.43916893e-01 -4.03329581e-01 7.44387135e-02 1.65285528e-01
8.58812392e-01 -1.92029923e-01 -7.27562785e-01 1.58461183e-01
-1.11432128e-01 2.91682988e-01 3.95511508e-01 2.35475436e-01
-7.18343914e-01 3.60710353e-01 6.29460907e+00 1.08761847e+00
-1.33297241e+00 1.13965809e+00 8.60694408e-01 -2.12566197e-01
-3.52619588e-01 -3.08862120e-01 1.94570929e-01 7.68332899e-01
8.07415962e-01 -3.88264775e-01 9.73482728e-01 8.24493289e-01
4.92197126e-01 -1.96948007e-01 -7.38210529e-02 1.46698761e+00
-6.63501024e-03 -1.86382830e+00 -4.65457976e-01 1.48301544e-02
5.87877333e-01 7.48744905e-01 2.79840618e-01 -4.93077964e-01
-1.52736098e-01 -8.91513288e-01 2.77649969e-01 8.75206590e-01
1.16673434e+00 -5.57355523e-01 7.14536309e-01 2.86604583e-01
-7.96908140e-01 1.28560275e-01 -3.20611298e-01 3.43889743e-01
-2.81583872e-02 1.06829655e+00 -4.67576563e-01 4.59153652e-01
1.01057518e+00 6.84480727e-01 -6.68073475e-01 1.53556943e+00
-2.10244656e-02 5.74054062e-01 8.78314003e-02 1.81047127e-01
-2.48432860e-01 -1.95519596e-01 5.30724883e-01 9.87570882e-01
2.37531841e-01 5.97260237e-01 1.18296675e-01 6.62830353e-01
1.80348784e-01 2.94925183e-01 -3.95301670e-01 5.87080941e-02
6.61864698e-01 1.50316787e+00 -1.00968349e+00 -3.17794204e-01
-2.81976372e-01 9.24251854e-01 2.80499589e-02 2.46750653e-01
-5.38925707e-01 -3.80820036e-01 1.34210691e-01 1.95536852e-01
2.15263311e-02 -2.79115915e-01 -7.74189234e-01 -1.28559816e+00
-1.48004681e-01 -1.13161433e+00 2.77848303e-01 -1.39671588e+00
-1.24759364e+00 8.77546966e-01 -4.11500394e-01 -1.08193517e+00
1.09145761e-01 -4.57135886e-01 -5.34332693e-01 7.53042877e-01
-1.36662984e+00 -8.17047477e-01 -3.05761069e-01 9.93200898e-01
1.35974690e-01 -4.97526318e-01 8.02769959e-01 6.78378761e-01
-6.16806984e-01 4.06340003e-01 6.60657287e-01 1.84341386e-01
3.26707125e-01 -9.52132583e-01 -2.19522223e-01 9.23795223e-01
1.07182637e-01 6.21934474e-01 4.19671834e-01 -6.28328860e-01
-1.66535807e+00 -8.63162696e-01 4.46108133e-01 5.98961487e-02
7.82000780e-01 1.70179218e-01 -7.94434726e-01 1.79449901e-01
2.21555322e-01 5.33893704e-01 1.05195880e+00 -4.83519256e-01
2.50633031e-01 -1.67725518e-01 -1.36733937e+00 2.54470348e-01
7.67777085e-01 -8.27699065e-01 -2.93114513e-01 9.54065979e-01
3.80422562e-01 -4.86097842e-01 -1.06035125e+00 -2.96351388e-02
4.06262696e-01 -8.71860921e-01 1.00066662e+00 -5.16186990e-02
4.42760617e-01 -3.02135378e-01 7.21483305e-02 -9.65883493e-01
-5.52250206e-01 -3.49195808e-01 -7.49522820e-02 3.12133312e-01
4.01606373e-02 -6.40016079e-01 8.54464889e-01 4.78017896e-01
-2.30607778e-01 -9.05430555e-01 -1.15256274e+00 -4.11833405e-01
-1.18165173e-01 -8.30929518e-01 4.42232281e-01 9.05681431e-01
-6.47396222e-02 -4.56809700e-01 -1.20140165e-01 7.79267251e-02
1.02156830e+00 2.83469945e-01 -8.24738070e-02 -6.69380307e-01
-3.47227007e-01 -4.75887150e-01 -4.49797720e-01 -4.54811305e-01
-1.44234508e-01 -1.47198141e+00 -3.06987703e-01 -1.56563556e+00
6.92448318e-01 -4.56710041e-01 -4.84206945e-01 3.61247241e-01
-2.41316646e-01 7.93892086e-01 1.95443943e-01 1.09647489e+00
-6.54407442e-01 2.37103313e-01 1.80753899e+00 -1.06121629e-01
4.34410363e-01 2.07516458e-02 -8.79867554e-01 4.37930286e-01
9.08676326e-01 -6.95218861e-01 -7.28190504e-03 -5.48672557e-01
-2.85071611e-01 3.31650078e-01 5.16004264e-01 -1.15316427e+00
5.39604664e-01 6.26376122e-02 4.25888211e-01 -6.91250384e-01
-9.25187767e-03 -7.50332892e-01 4.98886317e-01 9.94325161e-01
-2.32287481e-01 4.32785034e-01 -3.07905048e-01 2.34284289e-02
-2.10085064e-01 -4.15736288e-01 8.01065266e-01 -4.72081065e-01
-5.52437723e-01 7.28765726e-01 -4.71425653e-01 -1.17643848e-01
9.02816653e-01 -9.82130989e-02 -2.50302434e-01 -3.02411526e-01
-1.08017612e+00 7.95888379e-02 2.36897394e-01 -1.89899728e-01
1.21458375e+00 -1.57193077e+00 -9.37434196e-01 2.76651591e-01
7.68809288e-04 -4.52016205e-01 9.64565456e-01 1.66126835e+00
-1.06812632e+00 1.78478062e-01 -4.04495090e-01 -7.26013005e-01
-1.16020095e+00 4.70149010e-01 2.10855871e-01 -1.74076363e-01
-1.03355706e+00 6.80120409e-01 -4.53999132e-01 -3.16250086e-01
-4.30424325e-02 1.75838292e-01 -5.45718856e-02 -4.51067388e-01
1.33646047e+00 3.90770733e-01 3.58377457e-01 -5.25590301e-01
-5.68572998e-01 2.12139949e-01 -1.73277751e-01 -2.65903175e-01
1.50089633e+00 -3.25243682e-01 -4.84971076e-01 2.22511157e-01
1.30138195e+00 -2.76458383e-01 -5.81272244e-01 -2.09940463e-01
-1.65819824e-01 -6.49406314e-01 7.58738577e-01 -1.00315678e+00
-1.60969198e+00 8.70525777e-01 1.38717949e+00 -9.33795422e-02
1.37056708e+00 5.24287857e-02 1.05130422e+00 7.81133175e-02
5.95762849e-01 -7.76740432e-01 2.98567656e-02 8.80369991e-02
1.01428795e+00 -1.29360545e+00 3.34600538e-01 -1.71999894e-02
-8.88481438e-01 9.14556861e-01 -4.74539176e-02 -1.01680756e-01
9.59308028e-01 6.79346681e-01 3.58244926e-01 -3.63851070e-01
-3.77637565e-01 1.97365638e-02 2.28923544e-01 8.67079020e-01
4.95178431e-01 4.49858844e-01 -5.51759124e-01 7.86723912e-01
-2.13533714e-01 4.97390032e-01 4.51118112e-01 7.94723392e-01
-2.49887213e-01 -9.41207707e-01 -6.92315698e-01 8.65884423e-01
-7.03056633e-01 -5.55140793e-01 2.72744894e-01 -3.16787548e-02
-1.84546169e-02 7.15702653e-01 -1.53696194e-01 -1.08591489e-01
-2.38679469e-01 -3.93291205e-01 4.31446135e-01 -3.43586802e-01
-4.05216366e-01 2.91653246e-01 -3.42063844e-01 -6.07220829e-01
-3.62726688e-01 -8.99670899e-01 -1.44192278e+00 -3.34986657e-01
-1.86463773e-01 2.70226777e-01 1.21980000e+00 5.33883810e-01
5.74204445e-01 5.93824863e-01 7.30669260e-01 -6.89991832e-01
-1.37470067e-01 -8.14426780e-01 -8.05025578e-01 1.11896411e-01
-5.31725287e-02 -4.78355467e-01 1.85231060e-01 -6.59371614e-02] | [13.745709419250488, -2.3208060264587402] |
54913555-3938-4a61-b378-9bcf60bb124f | recurrent-segmentation-meets-block-models-in | 2205.09862 | null | https://arxiv.org/abs/2205.09862v1 | https://arxiv.org/pdf/2205.09862v1.pdf | Recurrent segmentation meets block models in temporal networks | A popular approach to model interactions is to represent them as a network with nodes being the agents and the interactions being the edges. Interactions are often timestamped, which leads to having timestamped edges. Many real-world temporal networks have a recurrent or possibly cyclic behaviour. For example, social network activity may be heightened during certain hours of day. In this paper, our main interest is to model recurrent activity in such temporal networks. As a starting point we use stochastic block model, a popular choice for modelling static networks, where nodes are split into $R$ groups. We extend this model to temporal networks by modelling the edges with a Poisson process. We make the parameters of the process dependent on time by segmenting the time line into $K$ segments. To enforce the recurring activity we require that only $H < K$ different set of parameters can be used, that is, several, not necessarily consecutive, segments must share their parameters. We prove that the searching for optimal blocks and segmentation is an NP-hard problem. Consequently, we split the problem into 3 subproblems where we optimize blocks, model parameters, and segmentation in turn while keeping the remaining structures fixed. We propose an iterative algorithm that requires $O(KHm + Rn + R^2H)$ time per iteration, where $n$ and $m$ are the number of nodes and edges in the network. We demonstrate experimentally that the number of required iterations is typically low, the algorithm is able to discover the ground truth from synthetic datasets, and show that certain real-world networks exhibit recurrent behaviour as the likelihood does not deteriorate when $H$ is lowered. | ['Nikolaj Tatti', '{Chamalee Wickrama Arachchi'] | 2022-05-19 | null | null | null | null | ['stochastic-block-model'] | ['graphs'] | [ 0.09965593 0.25420046 -0.13893837 -0.01050542 0.09151454 -0.6790573
0.5584532 0.10223176 -0.600989 0.8124717 -0.39039457 -0.26035225
-0.4505368 -1.116752 -0.7513039 -0.81184053 -0.82838935 0.90485924
0.54677933 0.04734632 0.08677974 0.42411488 -1.0582666 -0.4527895
0.50216323 0.67180985 -0.0327613 0.8069053 -0.11142977 0.78408766
-0.6436942 -0.28422835 0.4294113 -0.58758354 -0.8527632 0.4308092
-0.76492023 0.13621508 -0.40796706 0.8084474 -0.03440227 0.08779449
0.5365758 -1.7953874 0.3725752 1.1063354 -0.9688093 0.2614682
0.2100604 0.10785175 0.990832 -0.00830541 0.67641526 1.0336097
0.59303415 0.14577416 -1.3502995 -0.7119077 0.41863257 -0.04844575
-1.5921187 -0.07815971 0.79303145 -0.31887996 0.6990333 0.39219996
0.9560735 0.7171748 -0.0192278 0.50984323 0.90339 -0.29300982
0.33062845 -0.15803799 0.07063589 0.52985096 0.31793016 -0.21733695
-0.36209655 -0.46714672 0.8209673 0.2340393 -0.14737698 -0.10949939
-1.1135806 0.7921246 0.23444349 0.48839393 -0.3847793 0.7129132
0.1376231 0.44611895 0.3043292 0.10783784 -0.43528822 -0.30188236
-0.86585796 0.26512113 1.0736185 0.9591782 0.8348645 -0.46209726
0.39172336 0.5423092 0.10340825 0.2809173 -0.11613388 -1.0390965
0.36737105 0.48663247 0.39426243 -1.1499184 -0.38731137 -0.374067
-1.0836961 -0.40080345 0.7017068 -0.24237403 -0.7353193 2.0305665
0.17660594 0.5184764 -0.36133227 0.21138763 -0.14617912 0.85721725
-0.19209619 -0.8613561 1.0878185 -0.6582728 -0.59873796 -0.09153016
0.7379995 -0.37384298 0.49587783 0.37740394 -1.3834118 0.10275935
-0.58564353 0.75445443 -0.19213828 -0.6048516 0.52227396 0.52370477
-1.0556479 0.51664174 -1.0221343 -0.36901912 -0.00928753 0.5204714
-0.01710742 -0.07011405 -1.1759386 0.38635975 0.29887095 0.23949781
-0.84077805 -0.24887255 -0.5640222 -0.02423791 0.79538894 -0.5106391
0.97626525 -0.7060735 -0.9255632 0.66262347 -0.4785763 -0.72593415
0.79585195 0.53295624 -0.18885075 0.17374268 0.05977364 0.22658199
0.5878106 -1.3210981 -0.6898094 -0.11129441 0.32676774 0.09117957
-0.09451086 0.12828101 -0.8407111 -0.3648985 0.24270317 -1.3151214
-0.7522624 -0.32443422 -0.5441305 -0.35345906 0.6792807 -0.15376417
1.4833806 -2.0836492 0.1568103 0.9557649 0.41679215 -0.34788358
0.17724225 0.7857334 0.14989752 0.67169476 -0.48766115 -0.4061118
-0.09734431 0.603511 0.0430586 0.6100625 -0.44810587 0.5922469
-0.8216946 -0.42488253 -0.18559994 0.01779443 -0.27358544 -0.01016623
-0.42533332 0.26820087 -0.33754042 0.1391929 0.49870455 -0.6270808
0.55846316 0.41279685 -0.12769105 0.02228673 -1.6282213 0.9284962
-0.26455548 0.32070312 0.3561425 -1.0246453 0.51755905 0.41492638
0.88672686 -0.21982133 0.1537161 0.19512324 0.11560962 -0.12105165
0.13616441 -0.09891538 -0.23357308 0.94389254 -0.68673676 -0.0394062
0.6389253 0.39128605 1.7001237 -0.66517675 0.07694928 -0.01564574
0.21419923 -0.03200909 0.8614126 0.8140571 0.0123897 0.01492796
1.2094679 -0.1596542 -1.0409418 -0.7773341 0.2901224 0.54350775
0.34526658 -0.44361565 -0.789632 -0.41953084 -0.23280153 0.2685774
-0.8315204 0.16417347 -0.8079385 -1.0036522 0.20071514 0.22076629
0.4700377 -1.0599339 -0.5022115 0.54288375 -0.2628998 -1.0849599
-0.5041402 0.24897954 -0.8057234 -1.2677759 -0.50026715 -0.6634173
1.041645 0.3682356 1.0721008 0.3475642 -0.06593432 0.36105165
-0.2838931 -0.10141959 -0.2358615 0.17626695 -0.16297626 0.05318435
-0.06894165 -0.8802423 -0.69392645 0.5527967 -1.2195061 -0.10868043
0.3185731 0.47799087 0.40498316 0.86731565 0.3216485 -1.1599239
0.5111926 -0.7154888 -0.80133516 0.06076829 -0.39135548 -0.24847706
0.64542454 -0.561379 -0.4109926 -0.01083346 0.41429394 -0.139269
0.21911353 0.8327453 0.0878775 0.10699247 0.19084854 0.1084341
-0.12806392 0.08048735 0.15694503 0.23937328 -0.16487177 -0.45159546
0.86703575 0.70582616 0.24686386 -0.89517546 -0.23338728 -0.46380928
-0.2197455 -0.36650613 0.3124102 -0.50408643 -1.1749867 0.6597004
-1.0469704 -0.74434793 -0.19958283 0.1268547 -0.56371194 0.3838943
-0.86539835 -1.2108457 0.19299746 -0.83938986 0.33624804 -0.07718617
-0.43228525 -1.0298344 0.05121684 -0.08004752 0.16746293 0.42850503
0.8880619 -0.5204011 -1.0249037 -0.22667207 -0.04608933 -0.18389338
0.13402362 0.1838865 -0.2000443 -0.271917 -0.05383429 0.35287058
0.47446105 0.5630926 1.1481526 -0.6410961 -0.6680898 0.18188806
1.3526754 0.6171549 0.6451203 0.13989219 0.4746861 0.7208659
0.4765229 0.6261019 0.36942264 0.4519714 0.60278255 -0.07617848
0.8001243 0.10338953 0.19480532 0.71592855 -0.08069043 -0.6957037
-1.1250159 0.914685 -2.0003831 -1.0953624 -0.23794106 2.3410983
0.7871282 0.61524355 0.45789042 0.45163116 0.8395393 0.26927274
-0.3568928 -0.07597905 0.07852726 0.04498981 0.85701346 0.68526125
-0.56013626 0.6986527 6.3678656 0.73215973 -0.7132527 -0.13437869
0.8623691 -0.15100285 -0.2895275 0.22490375 -0.5778949 0.7416424
1.1092479 -0.33125722 0.6704373 0.32539335 0.6952184 -0.40853518
-0.94183624 0.8075789 -0.51280963 -1.044254 -0.43733823 0.46122485
0.69805974 -0.16651402 -0.21369548 -0.2233362 0.8799343 -1.0058705
0.47923556 0.30463466 0.31162056 -0.9882175 0.39813215 0.75803685
-1.5920587 -0.04421959 0.18313038 -0.02523776 0.6100169 0.74628055
-0.6437448 0.40241316 0.66403466 0.61106116 0.11286189 1.0853275
0.07331681 0.65234846 -0.9699597 -0.14941007 0.3703013 -0.61106396
0.52760726 0.90840167 0.35379982 0.2903617 0.30681068 0.49172905
-0.17936051 -0.29298547 -0.6856134 -0.19584757 0.72221607 0.9401167
-1.3098186 -0.28776872 -0.22207302 0.8520326 0.01648128 0.57991505
-1.0987774 -0.21379237 0.45350495 0.61814004 0.23212217 -0.58385503
0.06689981 -0.73840094 0.0728806 -0.64434445 0.40468442 -0.39690334
-1.0802771 0.48112082 0.18064883 -0.8689411 -0.3191025 -0.0400128
-0.5579247 0.6015603 -1.1414199 -0.68955415 0.00872733 0.6430327
0.25747266 0.58265376 0.19585352 0.1187023 -0.6866546 0.08940713
0.01743185 0.18417838 0.09204598 -1.13434 0.31559438 0.8972607
-0.04452009 0.61862373 1.0803648 -0.71521187 -0.9654738 -1.0288761
1.0551485 0.01526788 0.915147 -0.51805687 -0.69936216 0.92506546
0.09623767 -0.1239118 0.49872443 -0.03041582 0.04221893 -0.11495218
-1.0802585 0.849689 1.2564647 -0.25692055 0.19744758 0.38423902
0.5747341 -0.05949616 -0.75989515 0.2927116 0.14512944 -0.75981617
0.5269489 -0.17454372 -0.05344433 -0.43944108 0.35214755 -1.1193587
-0.03264501 -1.1256069 -0.14128858 1.0874916 0.66189426 -0.8187232
1.1704346 0.5948072 0.5690476 -0.746271 -1.0979316 -0.8691398
-0.39126632 -0.37833735 0.5192886 1.101887 0.09129931 0.14499566
-0.44057566 0.15673447 0.6511904 -0.05202417 0.76838255 -1.3460785
-0.52812034 -0.44895577 0.0076504 -1.2333472 -0.05976876 -0.27873877
-0.02562776 -1.3499666 0.30873623 -0.9524401 0.11183685 0.15374735
0.28216675 -0.05565583 0.02039273 0.22789903 -0.6719028 0.22262536
0.9872181 0.01902655 -0.49587354 0.3925494 -0.25721273 0.6871639
0.8413579 -0.613317 -0.56878155 -0.00643932 0.72795206 0.6144457
0.10289084 -0.67445886 0.43536252 -0.3537429 -0.3244824 -0.6763837
0.46367314 -1.1014774 0.9656513 0.6619286 -0.1958249 0.28525722
-0.19457757 0.94824934 -0.10770152 -0.19129418 0.59113324 -0.36111194
0.04221782 0.70746416 -0.6489239 -0.147221 1.3753903 -0.41317785
-0.05068877 -0.72119635 -1.1385558 0.7170974 0.61552 -0.03506445
0.18297108 -0.91506207 -0.23616116 -0.29378295 -0.3400104 0.37338945
0.18975863 1.0372776 -0.53455794 0.26343757 0.38997513 -0.63312376
-1.3692002 0.5661429 0.44292593 -0.7363613 -0.18759485 0.5806593
-0.13222425 0.06472205 0.24237001 -0.2401978 -0.08148265 0.25056145
0.14253828 0.55325997 -0.41961145 -0.38178906 -0.12062316 0.4746818
-0.1737939 -0.6275985 1.3515012 -0.44327712 -0.56203765 0.6820204
0.89664036 -0.03301148 -1.059874 -0.35324252 0.20607755 -0.19136657
-0.5107798 0.01269896 -1.208629 0.2864137 -0.14049889 1.2113854
1.142153 0.2517813 0.7015246 0.0431155 0.807897 -0.9903529
-0.09045072 0.65392613 0.33896884 -0.5412207 -0.23122445 -0.88084704
-0.13625813 0.6266645 0.3353584 -0.16110091 0.89936644 0.23275757
-0.4110679 -0.2679812 -0.8487225 -0.08969787 -0.65525836 0.1198882
0.01810927 0.09164541 -0.69753164 0.24360025 -0.1860674 -0.18501794
0.7892597 1.0980062 -0.40725166 -1.4309278 -0.28947854 0.51720625
-0.70317996 0.03272248 -0.28162912 0.7877258 -0.05917095 1.0369459
0.20262536 -0.0690254 0.22277664 -0.20554963 0.20329814 -0.50276697
-0.5813798 0.2778375 0.12400324 -0.35099337 -0.58779246 -0.788331
-1.2383248 -0.90499926 -0.36192918 0.58795315 0.28289258 0.89982384
0.07472412 0.41885325 1.0360404 -0.386517 0.07468876 -0.8014035
-0.9352909 0.06212768 0.25385395 -0.4905404 -0.8514184 0.08484729] | [6.849013805389404, 5.173060417175293] |
74cd2084-9bfb-41b6-b358-5beae370511a | rurebus-a-case-study-of-joint-named-entity | 2010.15939 | null | https://arxiv.org/abs/2010.15939v1 | https://arxiv.org/pdf/2010.15939v1.pdf | RuREBus: a Case Study of Joint Named Entity Recognition and Relation Extraction from e-Government Domain | We show-case an application of information extraction methods, such as named entity recognition (NER) and relation extraction (RE) to a novel corpus, consisting of documents, issued by a state agency. The main challenges of this corpus are: 1) the annotation scheme differs greatly from the one used for the general domain corpora, and 2) the documents are written in a language other than English. Unlike expectations, the state-of-the-art transformer-based models show modest performance for both tasks, either when approached sequentially, or in an end-to-end fashion. Our experiments have demonstrated that fine-tuning on a large unlabeled corpora does not automatically yield significant improvement and thus we may conclude that more sophisticated strategies of leveraging unlabelled texts are demanded. In this paper, we describe the whole developed pipeline, starting from text annotation, baseline development, and designing a shared task in hopes of improving the baseline. Eventually, we realize that the current NER and RE technologies are far from being mature and do not overcome so far challenges like ours. | ['Ivan Smurov', 'Elena Tutubalina', 'Veronika Sarkisyan', 'Vladimir Ivanov', 'Tatiana Batura', 'Ekaterina Artemova', 'Vitaly Ivanin'] | 2020-10-29 | null | null | null | null | ['text-annotation'] | ['natural-language-processing'] | [ 1.78032249e-01 6.03389978e-01 -5.88477664e-02 -3.81529331e-01
-1.04322910e+00 -7.84632266e-01 9.96601224e-01 2.67724037e-01
-7.91210115e-01 1.13677716e+00 3.92655015e-01 -5.30443370e-01
1.34245783e-01 -4.81679142e-01 -3.19179267e-01 -1.94109842e-01
1.70960240e-02 9.19204950e-01 5.33761263e-01 -4.25936490e-01
1.96855783e-01 5.22102594e-01 -1.07719874e+00 4.13130939e-01
6.43721759e-01 5.49806714e-01 -3.82818803e-02 6.24092519e-01
-5.36053717e-01 9.61559176e-01 -6.46357656e-01 -7.84970939e-01
5.23928888e-02 -2.04295203e-01 -1.54650652e+00 -1.65871873e-01
1.13635369e-01 2.71315761e-02 -1.49064824e-01 7.14780927e-01
3.58906597e-01 7.78982788e-02 6.06281281e-01 -7.78912127e-01
-3.90102267e-01 1.02093744e+00 -2.09945858e-01 2.01047257e-01
3.29419672e-01 -4.46240634e-01 9.71631706e-01 -8.36713254e-01
1.07236779e+00 7.38043249e-01 8.32816184e-01 7.45022774e-01
-9.63061213e-01 -3.19106340e-01 8.25735629e-02 -4.93145287e-02
-1.35756934e+00 -8.46296787e-01 4.76703763e-01 -3.61527413e-01
1.55919552e+00 2.54332781e-01 2.83593684e-02 1.20108318e+00
-1.70118481e-01 8.00506473e-01 9.81416225e-01 -8.95729899e-01
-2.71126684e-02 5.40995479e-01 1.29539102e-01 3.95494193e-01
1.96090803e-01 -3.15326959e-01 -2.23123863e-01 -2.89158612e-01
3.28412652e-01 -4.84185308e-01 -1.95276678e-01 -2.29821578e-01
-1.23276830e+00 5.55626154e-01 -1.94785252e-01 1.09869838e+00
-2.90867597e-01 -2.85421371e-01 7.45024621e-01 3.18029046e-01
6.85266197e-01 6.39348865e-01 -1.07613218e+00 -3.29553366e-01
-1.14286518e+00 6.19616807e-02 1.32390070e+00 1.19653416e+00
5.60925782e-01 -3.54532629e-01 -9.79843065e-02 7.31077611e-01
-1.84186902e-02 -6.24580383e-02 5.39833724e-01 -5.54542661e-01
7.98571646e-01 6.14642024e-01 4.39194381e-01 -5.38837910e-01
-5.56063294e-01 -1.78116381e-01 -4.72437054e-01 -2.74576306e-01
6.41098678e-01 -5.29459774e-01 -8.23688388e-01 1.44858849e+00
2.67534345e-01 -1.12010397e-01 4.16520178e-01 4.05622214e-01
8.63139033e-01 3.49117875e-01 3.27449173e-01 -2.82125115e-01
1.55072510e+00 -8.16864669e-01 -1.04890203e+00 -2.89094061e-01
1.07953727e+00 -9.73218441e-01 6.41143203e-01 2.84692407e-01
-1.08562136e+00 -3.12425703e-01 -8.67421448e-01 -1.98481247e-01
-9.51326489e-01 2.86146313e-01 6.23382568e-01 8.08670282e-01
-9.71474826e-01 6.34449542e-01 -9.30542588e-01 -9.33754385e-01
2.89627519e-02 3.48497987e-01 -6.93369806e-01 2.75009632e-01
-1.26539922e+00 1.16742218e+00 7.73682356e-01 6.23967871e-02
-3.82748812e-01 -3.42829436e-01 -8.01556110e-01 -2.21430305e-02
6.32810473e-01 -3.11904192e-01 1.55104542e+00 -3.97420615e-01
-1.40730548e+00 1.06491101e+00 -2.96081334e-01 -4.21769470e-01
4.79436070e-01 -4.61021125e-01 -8.18122268e-01 -1.34460792e-01
2.09898576e-01 1.84851974e-01 2.72271514e-01 -1.06079423e+00
-9.10701990e-01 -4.13166523e-01 -1.68085665e-01 3.37026492e-02
-4.19333100e-01 7.28222132e-01 -3.90036643e-01 -4.55897868e-01
-2.27909699e-01 -7.29451358e-01 -4.35293287e-01 -8.15567493e-01
-6.06307566e-01 -4.47992593e-01 7.63318181e-01 -7.19667077e-01
1.27434742e+00 -1.93566680e+00 -2.36864373e-01 -8.67166817e-02
-5.59993740e-03 5.34758449e-01 8.70737061e-02 9.32081401e-01
-2.88205713e-01 4.61370200e-01 -1.27470165e-01 -3.60153973e-01
-1.45247709e-02 2.16496825e-01 -3.18418831e-01 1.21865660e-01
4.95686948e-01 7.81864226e-01 -1.01220441e+00 -5.94299793e-01
-1.07004613e-01 3.23995948e-01 3.71516906e-02 1.72668457e-01
-1.19817808e-01 3.87306362e-01 -6.92179561e-01 3.13156009e-01
3.48665148e-01 -1.08314037e-01 5.13336658e-01 -7.18076825e-02
-5.05168617e-01 7.93240130e-01 -1.32027328e+00 1.82453716e+00
-4.80468065e-01 6.36871696e-01 -1.86915398e-02 -1.05934083e+00
9.23499763e-01 8.44309747e-01 2.26855576e-01 -2.30240032e-01
1.24883220e-01 5.05490243e-01 -2.59563148e-01 -6.12645090e-01
8.05160999e-01 -6.13975972e-02 -3.25607330e-01 3.75539064e-01
4.53448504e-01 8.58292356e-02 5.81848323e-01 9.98108760e-02
1.39011204e+00 3.29473734e-01 6.68642402e-01 -1.77993760e-01
5.34517050e-01 4.46431756e-01 4.97595847e-01 6.05421543e-01
-6.28927499e-02 4.71621633e-01 4.54640716e-01 -2.82010257e-01
-1.17510998e+00 -3.74622673e-01 -2.32873082e-01 1.08951330e+00
-4.43530023e-01 -6.90736294e-01 -7.95044005e-01 -1.13352847e+00
-6.32234216e-01 8.62159133e-01 -3.88579249e-01 4.64817256e-01
-8.23152065e-01 -7.16391563e-01 1.00591493e+00 5.51521838e-01
3.52350175e-01 -1.02828908e+00 -3.90864819e-01 5.95822930e-01
-2.30469108e-01 -1.59480941e+00 -5.73165081e-02 6.45813286e-01
-7.37242520e-01 -9.00654256e-01 -7.04165578e-01 -8.97257626e-01
3.53535116e-01 -2.62513638e-01 1.26790965e+00 -1.91517323e-01
5.32028340e-02 7.23188147e-02 -6.09644055e-01 -4.59275216e-01
-5.70654571e-01 6.50894225e-01 -2.40333319e-01 -3.71535331e-01
8.18102598e-01 -3.75177771e-01 -1.48538262e-01 1.21421337e-01
-7.26750314e-01 -2.77008235e-01 7.37247229e-01 7.64901042e-01
9.01899710e-02 1.84104145e-01 5.88944316e-01 -1.58438158e+00
5.15077174e-01 -3.08659434e-01 -2.10849106e-01 5.01074076e-01
-6.35527730e-01 1.73011988e-01 6.42589033e-01 -1.32890731e-01
-1.50241613e+00 2.31522053e-01 -4.92184848e-01 3.33685070e-01
-7.25948572e-01 5.17941117e-01 -2.29054987e-01 3.57872844e-01
7.75667489e-01 -4.35463637e-02 -6.54706657e-01 -9.10261691e-01
4.66141403e-01 1.11605644e+00 6.18054748e-01 -5.04574537e-01
5.64733922e-01 2.57202208e-01 -3.81830931e-01 -7.02302754e-01
-1.00821412e+00 -8.55255961e-01 -9.99166787e-01 3.59203219e-01
8.71258378e-01 -8.53959799e-01 -2.83658355e-01 1.79923952e-01
-1.40286744e+00 -1.87878266e-01 -4.56649631e-01 3.98038179e-01
-2.68515825e-01 3.76818091e-01 -9.42260623e-01 -8.17224920e-01
-3.82973671e-01 -6.92372441e-01 9.97658908e-01 1.86518893e-01
-3.73882324e-01 -1.07270658e+00 3.43223751e-01 4.84451145e-01
2.99243897e-01 4.81875837e-02 6.64954007e-01 -1.57900691e+00
1.06699891e-01 -3.60637397e-01 -9.84285176e-02 1.14769585e-01
1.17499933e-01 2.77960543e-02 -1.12874997e+00 -5.90963522e-03
-5.01270480e-02 -3.73774648e-01 4.46858227e-01 -3.31221610e-01
5.27253926e-01 -1.17432140e-01 -6.94457591e-01 -4.89233956e-02
1.22307992e+00 1.93745255e-01 6.44149542e-01 6.53635204e-01
4.61114168e-01 8.11654091e-01 6.43302381e-01 2.27094948e-01
5.25442004e-01 5.63742638e-01 -1.36872277e-01 -2.53493339e-01
2.92392299e-02 -2.72914439e-01 2.61356264e-01 8.02630723e-01
-1.73969030e-01 -3.69670779e-01 -1.19452298e+00 9.43760753e-01
-1.87681019e+00 -8.30693126e-01 -4.57019538e-01 1.97171962e+00
1.12599099e+00 3.36537957e-01 6.50665462e-02 2.04512780e-03
6.50867581e-01 -1.62909254e-01 -3.62871811e-02 -5.23544967e-01
-5.84337413e-02 3.54088008e-01 5.81087232e-01 1.89399555e-01
-1.29414129e+00 1.14266336e+00 6.98260069e+00 5.79201221e-01
-7.96957374e-01 2.79832721e-01 3.18176687e-01 4.91239130e-01
1.83368754e-02 4.12330657e-01 -1.28308618e+00 3.55136320e-02
1.50309193e+00 -1.91991986e-03 -5.26673812e-03 7.49123991e-01
-1.16731897e-01 2.04103321e-01 -1.16801584e+00 3.96601409e-01
-1.00942450e-02 -1.10362375e+00 -3.85966986e-01 -4.73344214e-02
3.17177922e-01 4.33432221e-01 -6.90653503e-01 6.53436840e-01
6.82713330e-01 -7.40746558e-01 4.77478236e-01 3.09014469e-01
5.95381916e-01 -5.12240171e-01 1.06837285e+00 5.21715343e-01
-1.11977410e+00 3.31227839e-01 -2.68579215e-01 1.34337887e-01
3.61128747e-01 5.33030093e-01 -1.05365264e+00 1.05617929e+00
5.85226297e-01 3.57802212e-01 -5.97311437e-01 8.04123878e-01
-3.82688642e-01 6.55990124e-01 -4.24561322e-01 -3.59859690e-02
2.12906390e-01 1.68355346e-01 4.60687160e-01 1.82748425e+00
1.10252693e-01 9.51866508e-02 1.37818724e-01 4.41082090e-01
-1.28093436e-01 4.33142662e-01 -6.92414641e-01 -2.81131208e-01
3.56017232e-01 1.63299477e+00 -7.98040330e-01 -4.82437342e-01
-7.86263525e-01 9.76717532e-01 5.71550608e-01 1.30635470e-01
-3.21148396e-01 -7.65196204e-01 1.04454355e-02 4.29770462e-02
6.17098510e-01 -6.36499524e-02 3.71527374e-02 -1.41092396e+00
1.38113936e-02 -7.35912025e-01 6.20574534e-01 -5.64309537e-01
-1.29187608e+00 1.12518251e+00 1.74137037e-02 -8.35627854e-01
-6.18904293e-01 -6.81998909e-01 -4.61580813e-01 7.69962370e-01
-1.52350783e+00 -1.25654197e+00 1.74512342e-01 3.93011630e-01
4.04243350e-01 -6.17304966e-02 1.17666948e+00 5.45919418e-01
-6.37097120e-01 5.00397801e-01 5.72802275e-02 6.12494111e-01
9.18513715e-01 -1.41400409e+00 5.64225256e-01 9.34730470e-01
5.02208948e-01 6.51350975e-01 7.79304206e-01 -5.03934264e-01
-8.55936706e-01 -8.59704614e-01 1.81114542e+00 -7.56659269e-01
1.09187710e+00 -6.28196061e-01 -9.29007828e-01 9.17429805e-01
6.11363649e-01 -6.63444176e-02 8.78040433e-01 6.52723730e-01
-1.86743617e-01 3.72077644e-01 -1.13809633e+00 3.39934230e-01
9.54455078e-01 -5.20937145e-01 -1.07793605e+00 4.58615363e-01
5.70634663e-01 -3.50893259e-01 -1.22274268e+00 4.45679843e-01
1.74813747e-01 -6.11553788e-01 4.62278634e-01 -9.68307197e-01
7.61035830e-02 -2.13102087e-01 -6.13669865e-02 -1.07240844e+00
-8.28508660e-02 -8.31065297e-01 1.42915174e-02 2.04387093e+00
9.20165181e-01 -4.33878511e-01 7.06874192e-01 7.71034479e-01
-1.79241866e-01 -4.14327234e-01 -8.39574039e-01 -7.54039407e-01
1.92199677e-01 -4.07174557e-01 5.24392426e-01 1.16360676e+00
4.23828542e-01 9.84332800e-01 -2.48144984e-01 -5.42651862e-03
1.01078197e-01 -1.81337699e-01 6.45737171e-01 -1.30044341e+00
-1.94888040e-01 -9.49299037e-02 -2.00588107e-01 -8.70577812e-01
2.22075284e-01 -6.43736720e-01 2.43425950e-01 -1.56863391e+00
2.42493793e-01 -5.36221504e-01 -2.32100666e-01 8.40656996e-01
-7.46017694e-02 8.84942412e-02 -6.60702065e-02 2.61650503e-01
-6.26736760e-01 1.96571857e-01 6.41962051e-01 9.13182199e-02
-1.21183567e-01 -1.55920684e-01 -8.26317132e-01 7.60699511e-01
6.24545753e-01 -7.21998930e-01 -5.45637459e-02 -3.14607203e-01
1.88603252e-01 1.80266928e-02 -2.40048796e-01 -6.05150580e-01
3.81785959e-01 1.32325277e-01 2.36042753e-01 -5.62229156e-01
-5.48616275e-02 -7.94072390e-01 -3.64082269e-02 -8.28898326e-03
-2.61654794e-01 1.71250757e-02 1.65086314e-01 2.51105338e-01
-4.16260600e-01 -6.94880784e-01 3.77041072e-01 -2.33496144e-01
-7.85924971e-01 -1.16192803e-01 -5.40400982e-01 3.23362827e-01
9.06892002e-01 -2.68504699e-03 -3.06304723e-01 -9.98023525e-03
-8.61730874e-01 -9.45784897e-02 3.51520956e-01 3.72850448e-01
-1.74629226e-01 -8.57652664e-01 -9.01015282e-01 -2.10729674e-01
3.06399912e-01 -9.78491902e-02 -2.61981636e-01 6.86855853e-01
-2.59306461e-01 7.20387757e-01 1.42010987e-01 -1.51786283e-01
-1.15608096e+00 5.41078687e-01 5.56655936e-02 -1.10108638e+00
-6.46715641e-01 4.93621051e-01 -2.74146467e-01 -7.69819617e-01
-7.63147883e-03 -8.54080841e-02 -7.06062853e-01 1.36381805e-01
5.54123223e-01 1.70890361e-01 6.24846399e-01 -6.25906050e-01
-4.90332305e-01 1.12531431e-01 -4.76292849e-01 -2.42983952e-01
1.59091020e+00 -1.20029718e-01 -2.79638972e-02 4.40290421e-01
9.10313070e-01 3.30800742e-01 -4.57477361e-01 -2.48105481e-01
7.86495030e-01 5.80020510e-02 -9.36700106e-02 -9.10964012e-01
-7.05475688e-01 5.46118021e-01 1.55569911e-01 4.95566785e-01
1.02704597e+00 2.56838262e-01 8.26209486e-01 7.36487806e-01
4.46489364e-01 -1.19024885e+00 -6.10217750e-01 8.04570258e-01
4.66863722e-01 -1.16933692e+00 -1.13065299e-02 -5.51464200e-01
-6.89547896e-01 1.11730719e+00 2.66518861e-01 1.83488265e-01
5.92373967e-01 6.45225346e-01 1.72494233e-01 -2.90130198e-01
-8.44671249e-01 -4.97865319e-01 2.08530694e-01 5.52085638e-01
1.05960143e+00 -2.97006279e-01 -5.46184480e-01 5.80454707e-01
-4.62755412e-02 1.16309904e-01 4.96662259e-01 1.23999381e+00
-1.42931655e-01 -1.64052963e+00 -6.97732568e-02 2.61856943e-01
-1.06446230e+00 -2.46013477e-01 -7.23248005e-01 1.10612488e+00
-1.28385872e-02 1.11334491e+00 -2.20820516e-01 -1.11705124e-01
7.37014711e-01 4.31927860e-01 3.00351381e-01 -1.04158962e+00
-9.06441569e-01 7.02160671e-02 9.40922797e-01 -2.14309588e-01
-6.42116070e-01 -7.69693375e-01 -1.30023313e+00 4.88388874e-02
-7.56768465e-01 7.36264348e-01 6.78854942e-01 1.23503137e+00
3.13685745e-01 2.18905985e-01 2.81543225e-01 -4.70418066e-01
-4.35178578e-01 -1.26176262e+00 -4.85944808e-01 2.71967620e-01
-1.33538067e-01 -4.17660177e-01 -1.57412305e-01 2.25263134e-01] | [9.705061912536621, 9.315640449523926] |
1c6092cb-0957-43b2-9fbd-98504ea3ab54 | gesgpt-speech-gesture-synthesis-with-text | 2303.13013 | null | https://arxiv.org/abs/2303.13013v1 | https://arxiv.org/pdf/2303.13013v1.pdf | GesGPT: Speech Gesture Synthesis With Text Parsing from GPT | Gesture synthesis has gained significant attention as a critical research area, focusing on producing contextually appropriate and natural gestures corresponding to speech or textual input. Although deep learning-based approaches have achieved remarkable progress, they often overlook the rich semantic information present in the text, leading to less expressive and meaningful gestures. We propose GesGPT, a novel approach to gesture generation that leverages the semantic analysis capabilities of Large Language Models (LLMs), such as GPT. By capitalizing on the strengths of LLMs for text analysis, we design prompts to extract gesture-related information from textual input. Our method entails developing prompt principles that transform gesture generation into an intention classification problem based on GPT, and utilizing a curated gesture library and integration module to produce semantically rich co-speech gestures. Experimental results demonstrate that GesGPT effectively generates contextually appropriate and expressive gestures, offering a new perspective on semantic co-speech gesture generation. | ['Dongdong Weng', 'Shuwu Zhang', 'Zhi Zeng', 'Zeyu Zhao', 'Nan Gao'] | 2023-03-23 | null | null | null | null | ['gesture-generation'] | ['robots'] | [ 5.59817016e-01 3.56481830e-03 -2.58831024e-01 -4.05239642e-01
-8.41614246e-01 -5.64813435e-01 8.55079114e-01 -3.36308688e-01
-1.74374491e-01 1.86519444e-01 1.05353320e+00 -1.75438508e-01
-4.71013924e-03 -7.55071998e-01 -3.31155181e-01 -3.62458438e-01
2.17307523e-01 5.08396029e-01 4.13472876e-02 -2.73184747e-01
2.64512956e-01 3.80279958e-01 -1.68452895e+00 5.73431134e-01
5.05478740e-01 6.61584675e-01 1.36619613e-01 8.23800087e-01
-8.46507072e-01 7.47120440e-01 -7.34197736e-01 -1.35218427e-01
2.23773252e-02 -7.88770080e-01 -9.79985654e-01 -1.77464262e-01
7.95298144e-02 -6.23689294e-01 -1.57572225e-01 4.97346252e-01
4.69160318e-01 4.78237808e-01 5.34104586e-01 -1.42082572e+00
-4.01006460e-01 1.04381561e+00 6.17625862e-02 -5.05612910e-01
9.03583825e-01 3.22241843e-01 1.17182159e+00 -8.15142512e-01
8.19545507e-01 1.59533751e+00 5.32954633e-01 1.14572501e+00
-7.58017182e-01 -6.50420427e-01 2.48163357e-01 -6.18597828e-02
-1.03719938e+00 -4.34921414e-01 7.86813617e-01 -1.24993049e-01
1.27697194e+00 8.57246339e-01 8.85735095e-01 1.48640478e+00
-4.39249545e-01 1.44491887e+00 6.58087432e-01 -8.91443014e-01
2.51118362e-01 -5.46142101e-01 3.67203206e-02 4.45823640e-01
-3.56006235e-01 1.42568842e-01 -9.97771084e-01 1.86044276e-01
9.33005035e-01 1.17823355e-01 5.76324165e-02 1.25956804e-01
-1.50165963e+00 4.98653889e-01 3.13237369e-01 6.14840209e-01
-3.01849902e-01 4.89324182e-01 2.61380911e-01 1.67769060e-01
1.83026060e-01 4.30438399e-01 -2.30743617e-01 -7.27470279e-01
-7.58420527e-01 4.67417777e-01 8.94994974e-01 1.28313863e+00
3.09861392e-01 -2.24705599e-02 -5.59489727e-01 5.72189152e-01
6.17533445e-01 7.45298684e-01 6.40945613e-01 -8.34595382e-01
5.91197908e-01 1.10108614e+00 6.88763857e-02 -6.21105373e-01
-3.17191809e-01 3.90012711e-01 -1.25209600e-01 5.55601865e-02
4.79608774e-01 -1.18270695e-01 -1.10221326e+00 1.59714854e+00
2.70243257e-01 -5.45805134e-02 2.74331331e-01 1.11067033e+00
1.13994634e+00 5.01285493e-01 7.38756835e-01 2.53321826e-01
1.30827725e+00 -9.45299923e-01 -7.03542054e-01 -1.83148786e-01
6.87723815e-01 -6.58748925e-01 1.78044045e+00 5.10557257e-02
-7.33923197e-01 -1.58861145e-01 -7.17509627e-01 -3.61721635e-01
-4.22035635e-01 -4.17254753e-02 7.83566892e-01 5.64198256e-01
-7.95739889e-01 3.55772942e-01 -1.06380260e+00 -7.17256308e-01
4.49897528e-01 2.32591555e-01 3.96908857e-02 1.25328317e-01
-6.43543243e-01 5.47015905e-01 5.09089649e-01 -6.36084750e-02
-3.96025002e-01 -4.55969363e-01 -8.88368368e-01 -6.36464506e-02
6.82937086e-01 -5.74081242e-01 1.69774294e+00 -1.02180374e+00
-1.96237969e+00 4.16171432e-01 -3.53881031e-01 -1.13794021e-02
5.07498562e-01 -5.06977856e-01 -2.93487906e-01 1.19633101e-01
-9.66094285e-02 1.01149476e+00 7.72766411e-01 -1.10313892e+00
-7.15265512e-01 -3.47821623e-01 -9.76420417e-02 6.04562879e-01
-4.07408893e-01 5.45685649e-01 -4.10663575e-01 -9.52621996e-01
2.03397334e-01 -8.59402120e-01 -2.09474549e-01 -2.31106311e-01
-5.93481541e-01 -3.75271618e-01 1.14064944e+00 -4.71540153e-01
1.37923598e+00 -1.88304794e+00 1.66123942e-01 3.44974726e-01
-3.19179446e-02 2.51633584e-01 -3.06196451e-01 8.49090099e-01
5.58929145e-01 1.49115458e-01 -1.58932224e-01 -3.14680964e-01
5.83266437e-01 5.21265984e-01 -6.06968880e-01 -5.78316510e-01
1.05845943e-01 1.67654335e+00 -1.22049737e+00 -5.07684052e-01
6.20317101e-01 4.74789262e-01 -5.89242816e-01 4.03295279e-01
-8.83129001e-01 8.03220212e-01 -9.40355718e-01 7.44319916e-01
-1.81995690e-01 -1.20287046e-01 3.47395569e-01 7.42878020e-02
-7.61421770e-02 5.25225580e-01 -8.19809496e-01 2.01347780e+00
-5.19102037e-01 5.68777800e-01 -3.33369434e-01 -3.72823209e-01
7.71481693e-01 4.85095471e-01 5.02961814e-01 -6.01616383e-01
2.24024817e-01 2.63413310e-01 -3.28178972e-01 -9.87094104e-01
6.09413326e-01 -2.16705605e-01 -3.51419359e-01 1.05228961e+00
-1.81868419e-01 -4.28587019e-01 -9.86367315e-02 1.08769774e-01
1.01985955e+00 8.34112525e-01 3.10946941e-01 3.99645150e-01
3.96496290e-03 3.47557843e-01 -9.85978842e-02 6.01823509e-01
2.12977290e-01 5.24495900e-01 7.24773705e-02 -4.22636479e-01
-7.05622435e-01 -8.94005775e-01 6.24222338e-01 1.58030415e+00
-2.12957039e-02 -7.45126724e-01 -1.01678753e+00 -7.36659646e-01
-3.83767277e-01 7.52232790e-01 -2.79597610e-01 2.17916474e-01
-1.03761411e+00 -4.11644995e-01 9.21870470e-01 1.03623164e+00
4.05265927e-01 -1.81835985e+00 -1.26205087e+00 2.34706938e-01
-3.47609818e-01 -1.04597700e+00 -4.59277689e-01 -1.29693180e-01
-8.47735226e-01 -9.56672966e-01 -7.27127552e-01 -7.96277344e-01
6.73643529e-01 8.12919661e-02 9.55106854e-01 1.79485306e-01
-2.29615733e-01 7.10934043e-01 -9.99699652e-01 -5.01679897e-01
-7.06358016e-01 1.63203835e-01 -4.42659765e-01 -2.47420356e-01
8.40507686e-01 -3.90851170e-01 -3.20109338e-01 1.34384200e-01
-1.11242259e+00 5.36870778e-01 5.40977895e-01 4.56865489e-01
2.31448621e-01 -6.34216487e-01 2.54398853e-01 -2.82577306e-01
9.28863466e-01 9.26087275e-02 -5.39433807e-02 3.21047395e-01
-2.15381891e-01 3.42269748e-01 2.52387404e-01 -5.57837248e-01
-1.30057847e+00 2.30701938e-01 -3.96958113e-01 -1.95918486e-01
-5.51072955e-01 5.04159093e-01 -3.86446506e-01 2.47391939e-01
4.08206552e-01 3.22118878e-01 -5.34378402e-02 -6.24088168e-01
8.61250758e-01 8.27235162e-01 5.38152218e-01 -1.00077522e+00
5.95854938e-01 3.68374646e-01 -2.21017540e-01 -9.66418564e-01
-2.77592272e-01 -3.10414404e-01 -8.05901587e-01 -4.12306368e-01
8.69602084e-01 -2.88807273e-01 -8.94213617e-01 4.88356441e-01
-1.09963202e+00 -8.69687378e-01 -6.58798873e-01 5.02175570e-01
-7.10644364e-01 1.15170233e-01 -4.49270457e-01 -8.77025127e-01
-4.83336002e-01 -1.08117437e+00 1.65680218e+00 1.67310923e-01
-1.09280229e+00 -7.72216618e-01 -1.36504829e-01 2.84580827e-01
3.39312851e-01 2.49561146e-01 9.84247386e-01 -9.28083122e-01
-6.68768167e-01 -1.72139600e-01 -1.64057165e-01 -2.64489591e-01
2.83730268e-01 -1.54735968e-01 -8.67062509e-01 3.35370958e-01
-5.58460474e-01 -3.54614168e-01 3.07783574e-01 1.42268285e-01
7.27624953e-01 -6.94145799e-01 -5.36520720e-01 4.98699456e-01
9.77566361e-01 5.94801664e-01 4.42326486e-01 2.61090636e-01
1.04981005e+00 7.68821895e-01 4.91107076e-01 3.47087353e-01
4.39872861e-01 8.55020165e-01 9.99533534e-02 2.03371987e-01
-5.90298772e-01 -9.10645843e-01 3.18151861e-01 6.83924019e-01
-3.04442465e-01 -2.45387241e-01 -1.08889437e+00 4.31157231e-01
-1.99919426e+00 -9.01717901e-01 -7.33986571e-02 1.57411182e+00
8.60142529e-01 -4.66819882e-01 2.01741412e-01 2.23219618e-01
1.59366608e-01 1.95949480e-01 -3.63762826e-01 -1.85861573e-01
1.71069995e-01 6.54003561e-01 -5.62166236e-02 4.30698782e-01
-8.43617260e-01 1.57062936e+00 6.69501543e+00 7.01845109e-01
-1.27992773e+00 -1.16222844e-01 -1.45626396e-01 -3.01906645e-01
-5.64346790e-01 -1.31731004e-01 -6.09527111e-01 1.18259743e-01
5.13136625e-01 1.19389109e-02 4.80287552e-01 8.18021655e-01
4.40063059e-01 1.61651686e-01 -1.32445884e+00 9.23452079e-01
8.39967951e-02 -1.25491750e+00 6.11527920e-01 -9.71448496e-02
4.52809423e-01 -3.64616603e-01 -1.43320784e-01 2.52094865e-01
5.85096002e-01 -1.26756406e+00 9.70981359e-01 2.56626040e-01
1.01888490e+00 -3.34278464e-01 2.17577964e-01 3.58933300e-01
-1.44610858e+00 -3.70158069e-02 4.81406152e-01 -3.89820129e-01
3.26125413e-01 -4.30427283e-01 -1.22933674e+00 2.09430337e-01
4.42513078e-01 6.06266260e-01 -1.52599603e-01 4.41163123e-01
-6.78547025e-01 5.91255963e-01 -4.73627716e-01 -6.90441966e-01
3.75585288e-01 1.12386614e-01 7.34269381e-01 1.43477225e+00
4.86463726e-01 4.21985209e-01 2.42012322e-01 9.13394034e-01
4.78230044e-02 3.20894748e-01 -6.23311996e-01 -4.93584007e-01
6.57154441e-01 8.01457226e-01 -1.02958035e+00 -5.03691077e-01
-2.66548604e-01 1.06078994e+00 -2.34083712e-01 5.54890811e-01
-4.64254677e-01 -2.95637608e-01 7.36927271e-01 -2.28843078e-01
1.86300778e-04 -3.40311527e-01 -6.35288537e-01 -1.01388156e+00
1.72665834e-01 -1.03283954e+00 1.77490190e-01 -8.04233670e-01
-8.09903562e-01 4.02954757e-01 2.68850476e-01 -1.11871111e+00
-7.90514112e-01 -5.99087000e-01 -5.88399768e-01 5.96304178e-01
-8.32364678e-01 -1.78296030e+00 -5.04949927e-01 5.92361331e-01
1.03147030e+00 -1.81448478e-02 1.10940087e+00 -1.06263116e-01
-2.09997416e-01 4.78612542e-01 -5.10014713e-01 5.29263318e-01
1.34575486e-01 -9.96079564e-01 8.01729262e-01 7.29698896e-01
4.98874933e-01 9.18212950e-01 4.72003788e-01 -8.78104866e-01
-1.62865698e+00 -8.84210408e-01 1.01987112e+00 -7.65446424e-01
5.43264508e-01 -3.94202918e-01 -5.06164014e-01 9.39691126e-01
-3.56314518e-02 -5.79522669e-01 7.31248975e-01 -8.02170485e-02
-2.87377179e-01 6.23592138e-01 -7.77411222e-01 1.22537804e+00
1.78942287e+00 -6.74661100e-01 -9.84469235e-01 -4.42397855e-02
9.19332027e-01 -6.11216068e-01 -4.74969119e-01 3.02448362e-01
1.04529142e+00 -5.10824561e-01 1.03068900e+00 -8.99204791e-01
4.91079986e-01 -8.30765665e-02 -2.64542758e-01 -1.24159813e+00
4.03192639e-01 -8.82437587e-01 -1.61657408e-01 1.03047264e+00
2.48957291e-01 -1.44420728e-01 9.81912434e-01 9.27111447e-01
-3.53486091e-01 -2.22722694e-01 -6.81871355e-01 -7.55432606e-01
-1.78383887e-01 -1.09950674e+00 1.29419339e+00 7.22246766e-01
6.08970582e-01 2.15263829e-01 -1.20986886e-01 -3.21759790e-01
1.26997352e-01 2.86015570e-01 1.16314662e+00 -1.13531542e+00
-9.86109152e-02 -8.98209095e-01 -4.86658961e-02 -1.34927583e+00
1.88952386e-01 -1.09363341e+00 5.39599001e-01 -1.82296205e+00
-2.27440119e-01 -5.59111595e-01 2.40502506e-01 1.01805091e+00
-2.29503792e-02 2.68282384e-01 4.77923870e-01 4.20684159e-01
-5.16942680e-01 3.32952350e-01 1.41263628e+00 -6.97862953e-02
-7.49044001e-01 -2.47003242e-01 -4.98908281e-01 8.20209801e-01
5.93227148e-01 -2.75549948e-01 -4.86456245e-01 -6.60791159e-01
6.00186400e-02 -2.40188800e-02 5.00469863e-01 -8.49892914e-01
1.75800100e-01 -6.47114754e-01 4.75328527e-02 -4.19120252e-01
3.68151009e-01 -7.40438282e-01 9.54245701e-02 2.34781250e-01
-7.54134893e-01 -2.66002327e-01 1.11626215e-01 1.79474771e-01
-1.43565938e-01 6.66988641e-02 -1.81622252e-01 -2.42334232e-01
-1.15305126e+00 -3.64588946e-02 -4.47834969e-01 -1.23280033e-01
7.71786511e-01 -5.23869455e-01 -1.79089233e-01 -4.02595162e-01
-8.24399531e-01 -8.17484334e-02 3.94393206e-01 9.10278320e-01
8.42680454e-01 -1.35681319e+00 -3.85542601e-01 4.62421715e-01
3.92176539e-01 1.00042336e-01 -2.71791130e-01 3.84483457e-01
-4.29297239e-01 5.85341096e-01 -1.30287051e-01 -4.01223570e-01
-1.35817194e+00 8.05737004e-02 -1.15414940e-01 7.60257691e-02
-1.13068891e+00 8.10608864e-01 -1.68861017e-01 -5.48292220e-01
3.89474690e-01 -9.22023237e-01 -5.36947809e-02 -1.36572540e-01
7.10368991e-01 3.21911007e-01 -1.85373202e-01 -6.88659012e-01
-2.03321010e-01 6.18854642e-01 5.22444010e-01 -8.24516594e-01
1.17628419e+00 1.26273587e-01 2.06320792e-01 2.94425815e-01
8.69932950e-01 3.51500772e-02 -1.32058752e+00 -1.74575523e-01
6.06195271e-01 -4.20629680e-01 -3.86096001e-01 -1.17096305e+00
-5.83130956e-01 9.06344593e-01 5.87863289e-03 -6.17200956e-02
9.37678754e-01 1.52582705e-01 1.32720256e+00 6.49911046e-01
5.62371373e-01 -1.04994392e+00 4.24505204e-01 7.47007847e-01
9.41052794e-01 -9.11229551e-01 -5.21534264e-01 -2.64513761e-01
-7.27168679e-01 1.28990746e+00 3.38571727e-01 2.73690641e-01
1.34626657e-01 6.17705524e-01 4.64546233e-01 -3.50635737e-01
-3.27532440e-01 -3.92069012e-01 4.35285449e-01 9.34063554e-01
5.92160761e-01 3.98566395e-01 -2.07226381e-01 6.99400008e-01
-5.35922527e-01 3.35794717e-01 -1.91921219e-01 1.34663796e+00
-4.08300340e-01 -1.43088949e+00 -5.08988738e-01 1.75854176e-01
-1.01133008e-02 -3.50498021e-01 -9.27013755e-01 7.97025800e-01
-2.51126617e-01 9.62946355e-01 -6.17300570e-02 -6.34945571e-01
3.01734954e-01 6.65383458e-01 6.15043461e-01 -7.84627259e-01
-7.68381596e-01 1.59835622e-01 2.10130602e-01 -8.36013258e-01
-5.28846562e-01 -6.43163383e-01 -1.93162775e+00 -4.54381527e-03
2.23545969e-01 -1.07611038e-01 7.15825796e-01 1.38887918e+00
2.29711726e-01 4.99030799e-01 -2.21771061e-01 -1.20141411e+00
-2.09439650e-01 -9.82181311e-01 8.75538494e-03 6.26655281e-01
8.83485675e-02 -3.15692663e-01 5.36510311e-02 3.50336790e-01] | [5.623302459716797, -0.1225379928946495] |
578e6da6-82af-4d63-97a6-d8d039db7ccd | structured-state-space-models-for-multiple | 2306.15789 | null | https://arxiv.org/abs/2306.15789v1 | https://arxiv.org/pdf/2306.15789v1.pdf | Structured State Space Models for Multiple Instance Learning in Digital Pathology | Multiple instance learning is an ideal mode of analysis for histopathology data, where vast whole slide images are typically annotated with a single global label. In such cases, a whole slide image is modelled as a collection of tissue patches to be aggregated and classified. Common models for performing this classification include recurrent neural networks and transformers. Although powerful compression algorithms, such as deep pre-trained neural networks, are used to reduce the dimensionality of each patch, the sequences arising from whole slide images remain excessively long, routinely containing tens of thousands of patches. Structured state space models are an emerging alternative for sequence modelling, specifically designed for the efficient modelling of long sequences. These models invoke an optimal projection of an input sequence into memory units that compress the entire sequence. In this paper, we propose the use of state space models as a multiple instance learner to a variety of problems in digital pathology. Across experiments in metastasis detection, cancer subtyping, mutation classification, and multitask learning, we demonstrate the competitiveness of this new class of models with existing state of the art approaches. Our code is available at https://github.com/MICS-Lab/s4_digital_pathology. | ['Stergios Christodoulidis', 'Paul-Henry Cournède', 'Maria Vakalopoulou', 'Joseph Boyd', 'Leo Fillioux'] | 2023-06-27 | null | null | null | null | ['whole-slide-images', 'multiple-instance-learning'] | ['computer-vision', 'methodology'] | [ 8.53731573e-01 -9.42621380e-02 -5.49787283e-01 -2.13515341e-01
-1.35135305e+00 -3.79522473e-01 3.04832637e-01 3.58300179e-01
-4.58182186e-01 7.58664966e-01 -4.52328362e-02 -4.46292639e-01
-1.40158027e-01 -4.84151363e-01 -6.95359111e-01 -1.31557393e+00
5.08639142e-02 6.27672136e-01 6.68705702e-02 8.84771496e-02
2.25241184e-01 3.90946984e-01 -1.39474297e+00 7.95185268e-01
3.88291627e-01 8.72807801e-01 3.64758551e-01 1.19134235e+00
-4.16712224e-01 7.42391527e-01 -5.69944739e-01 -1.21562384e-01
-1.42704234e-01 -3.97643834e-01 -8.64785075e-01 7.68263862e-02
1.97988659e-01 7.82539845e-02 -2.56922573e-01 9.55487907e-01
6.88785315e-01 -2.35628873e-01 5.86880147e-01 -8.53991449e-01
3.27984273e-01 4.69407409e-01 -4.55168575e-01 5.04833519e-01
2.97095248e-04 7.51065686e-02 8.15066278e-01 -4.65082049e-01
8.84948075e-01 6.87488914e-01 6.28952980e-01 7.50047386e-01
-1.43931568e+00 -4.70504314e-01 -1.29588634e-01 2.41656452e-01
-1.24068844e+00 -4.81445491e-01 5.65594614e-01 -4.58488256e-01
1.05759275e+00 6.33611977e-01 7.13650048e-01 1.29315650e+00
8.15567970e-01 9.46656883e-01 9.69886959e-01 -3.48079532e-01
2.89359123e-01 -4.17252257e-03 3.84021014e-01 7.38449633e-01
-2.52928287e-01 -2.40960121e-01 -5.44927418e-01 -6.21448457e-01
3.98591548e-01 7.00142503e-01 -1.66832700e-01 -1.13919944e-01
-1.19501483e+00 6.95700169e-01 8.61959755e-02 4.09788072e-01
-1.31395489e-01 2.91260749e-01 8.81111681e-01 2.81780481e-01
6.07295513e-01 1.96487889e-01 -3.72037500e-01 -1.50366843e-01
-1.19088042e+00 1.00028068e-02 7.71409810e-01 4.70322222e-01
4.52816427e-01 -5.31211913e-01 -2.71949112e-01 6.72453403e-01
-7.16100633e-02 -6.40789047e-02 1.02188599e+00 -5.86088300e-01
4.84831333e-02 7.43862510e-01 -3.10232759e-01 -6.07347310e-01
-6.49267018e-01 -6.71400428e-01 -1.39021158e+00 -1.56314194e-01
2.33786762e-01 1.22304171e-01 -1.13940406e+00 1.54487944e+00
3.79667580e-01 6.38197660e-01 1.93750486e-02 3.85584205e-01
4.20464128e-01 7.76939988e-01 1.50682330e-01 -4.58779275e-01
1.38897753e+00 -8.63264799e-01 -7.30136991e-01 -6.98864534e-02
1.19375110e+00 -4.80536640e-01 6.23497784e-01 4.48564410e-01
-1.07097697e+00 -1.56614050e-01 -7.03012884e-01 2.20161788e-02
-4.61899132e-01 -8.69209692e-02 4.69647259e-01 2.20681936e-01
-1.21266246e+00 6.28332734e-01 -1.33754814e+00 -2.66761035e-01
6.81959450e-01 5.07144153e-01 -2.75411993e-01 -8.58765319e-02
-9.57716703e-01 7.06019878e-01 1.99809194e-01 2.09562201e-03
-9.61597204e-01 -9.54640269e-01 -6.13014996e-01 9.48054865e-02
-2.67932452e-02 -5.91567576e-01 1.15739536e+00 -7.63162255e-01
-1.13551044e+00 1.14386523e+00 -4.84627366e-01 -6.27602994e-01
3.35412592e-01 3.88507187e-01 -6.78966045e-02 1.03648201e-01
-2.61050582e-01 3.62486631e-01 8.17965329e-01 -6.14629328e-01
-5.46408176e-01 -4.36942518e-01 -5.97514153e-01 2.51954682e-02
-3.99476171e-01 -1.97862405e-02 -2.18518183e-01 -5.49360573e-01
-5.88972569e-02 -1.14696372e+00 -8.12560499e-01 -1.10171847e-01
-4.89677966e-01 -1.63559124e-01 7.27448046e-01 -6.84877157e-01
1.22594345e+00 -2.27172065e+00 3.47832650e-01 2.52738446e-01
4.10129905e-01 2.04662502e-01 -1.34276405e-01 4.28529620e-01
-2.67817169e-01 1.27151161e-01 -3.44981819e-01 -4.48763281e-01
-3.56627882e-01 1.44957632e-01 -8.57297480e-02 7.08201289e-01
9.02532414e-02 9.42147732e-01 -7.07240164e-01 -6.80569947e-01
-1.20965250e-01 4.52448666e-01 -1.87659830e-01 1.27574533e-01
-4.38139915e-01 2.91777551e-01 -3.31549555e-01 6.21978045e-01
2.71402150e-01 -9.29071784e-01 5.23303866e-01 4.98024002e-02
2.91903913e-01 1.29713997e-01 -7.58377969e-01 2.02288413e+00
-2.43297860e-01 4.55582052e-01 -1.00147650e-01 -1.32645738e+00
5.38265646e-01 6.09672964e-01 8.80152941e-01 -3.15164208e-01
7.79715553e-02 2.57404238e-01 1.36340737e-01 -5.12360573e-01
3.81879434e-02 -3.11259836e-01 -8.66621360e-02 7.57178783e-01
3.64686027e-02 1.56442255e-01 4.87035304e-01 2.59519070e-01
1.75674665e+00 -3.87856275e-01 5.41141868e-01 3.22686881e-03
4.54742283e-01 1.33235008e-01 6.25589788e-01 5.16441703e-01
-2.20590875e-01 5.37247121e-01 7.95534313e-01 -5.65181017e-01
-1.09509540e+00 -6.48087740e-01 -3.27876508e-01 9.86278832e-01
-3.57807875e-01 -3.56460869e-01 -5.52692711e-01 -6.41432226e-01
-1.83080435e-01 3.28239314e-02 -8.09104502e-01 -1.70760676e-01
-5.84728241e-01 -1.34923744e+00 6.26750171e-01 2.77233869e-01
-2.53199458e-01 -1.02578032e+00 -7.45039403e-01 4.54731524e-01
-1.52489737e-01 -5.68979025e-01 -3.68910462e-01 6.84212923e-01
-1.04128456e+00 -1.04317939e+00 -7.03387678e-01 -8.15042913e-01
8.66226375e-01 -2.26125047e-02 9.23501313e-01 4.43350554e-01
-1.02301860e+00 3.30748074e-02 -1.31720543e-01 -4.69714880e-01
-7.91212797e-01 3.01663190e-01 -2.56316572e-01 1.22130997e-01
2.84830868e-01 -5.75129807e-01 -5.93281150e-01 -1.99797302e-01
-1.32281494e+00 2.17290044e-01 1.04137206e+00 1.34091938e+00
1.14733732e+00 -9.80577543e-02 3.75310063e-01 -1.33778167e+00
3.07009786e-01 -6.23087525e-01 -2.91417778e-01 4.05576080e-01
-2.14549720e-01 3.69723886e-02 6.33157790e-01 -5.81941783e-01
-5.56517005e-01 3.93587708e-01 -9.27210525e-02 -4.53994155e-01
-1.72597557e-01 8.54941845e-01 3.64966601e-01 -1.22968636e-01
5.51926613e-01 6.63590729e-01 6.87805295e-01 -1.97722301e-01
-2.82286167e-01 5.97926974e-01 1.95008010e-01 2.06413064e-02
2.24037796e-01 5.60704291e-01 3.16556275e-01 -6.79024935e-01
-8.36242557e-01 -7.42096126e-01 -3.85804027e-01 -9.09955949e-02
5.67156136e-01 -6.98534250e-01 -6.67124510e-01 5.64711571e-01
-7.89554596e-01 -7.08947599e-01 -2.55023360e-01 1.56613246e-01
-6.78558230e-01 2.42250830e-01 -1.27780855e+00 -4.43402529e-01
-5.22426546e-01 -1.29313636e+00 1.26561224e+00 2.29421221e-02
-3.84743631e-01 -1.09755003e+00 4.68074381e-01 1.90835297e-01
3.87935936e-01 3.17788154e-01 1.14776015e+00 -8.38647187e-01
-3.97683471e-01 -5.94645202e-01 2.47298837e-01 -3.97455469e-02
4.03145701e-02 8.01843554e-02 -8.41130197e-01 -5.51864922e-01
1.14118345e-01 -3.89752388e-01 1.08066010e+00 5.83283901e-01
1.70278132e+00 -2.73737937e-01 -8.86898100e-01 5.58384061e-01
1.54780793e+00 8.23897123e-02 5.86527169e-01 1.92352235e-01
2.90752172e-01 5.43942392e-01 3.34871680e-01 4.45049405e-01
4.79998766e-03 3.91711175e-01 3.31789672e-01 -1.90050885e-01
2.56674625e-02 2.66297221e-01 1.50658011e-01 9.46124971e-01
3.90248805e-01 -3.58222187e-01 -1.18825948e+00 4.95653301e-01
-1.74717474e+00 -1.01604712e+00 2.33737096e-01 1.90417421e+00
1.06170285e+00 -6.82263896e-02 -5.09349927e-02 3.73174429e-01
5.75085700e-01 1.23601772e-01 -7.42046535e-01 -1.38037890e-01
-4.05852683e-03 2.26933554e-01 3.62362623e-01 2.48106092e-01
-1.02009726e+00 4.24972892e-01 5.98002338e+00 1.18365610e+00
-1.33269536e+00 1.66798294e-01 1.23030198e+00 -5.77241004e-01
-2.43406203e-02 -2.19859034e-01 -8.67750406e-01 6.24058068e-01
1.51247394e+00 -2.04531401e-01 -4.85053770e-02 5.07751167e-01
6.73073158e-02 -7.84952939e-02 -1.16937697e+00 9.86732066e-01
3.05155087e-02 -1.83941221e+00 -1.06577806e-01 3.89812440e-01
7.70069301e-01 1.75719351e-01 2.68961251e-01 -6.83419348e-04
9.60869715e-02 -1.16421270e+00 8.25594515e-02 5.60858548e-01
8.35683286e-01 -5.35559356e-01 8.14188719e-01 5.65147877e-01
-7.28262424e-01 -1.35658681e-01 -4.16545689e-01 3.76358479e-01
5.07681966e-02 8.13185751e-01 -9.10992205e-01 3.71589661e-01
4.49228317e-01 8.75294328e-01 -5.47429800e-01 9.57275450e-01
5.16984999e-01 6.32572174e-01 -2.61935651e-01 -1.44799873e-01
2.09401533e-01 1.68792054e-01 2.47439682e-01 1.43318617e+00
3.73266429e-01 4.89964895e-02 1.97506547e-01 2.02586919e-01
4.75007184e-02 4.22997139e-02 -4.47819889e-01 -9.35938507e-02
1.37950808e-01 1.51430595e+00 -9.08736587e-01 -4.58017319e-01
-2.46099368e-01 7.16447353e-01 5.31430602e-01 1.49529248e-01
-6.71062887e-01 -1.63410231e-01 5.07480264e-01 5.89360297e-02
1.60243928e-01 3.02465647e-01 -1.57182366e-01 -8.85972142e-01
-1.65656880e-01 -1.04062033e+00 8.38879704e-01 -2.17532516e-01
-1.17074549e+00 4.26381201e-01 -5.72163522e-01 -1.46207881e+00
-3.29670221e-01 -5.10619938e-01 -5.65600514e-01 5.12838125e-01
-1.39262438e+00 -8.71901453e-01 -6.78935573e-02 4.07950819e-01
6.26470983e-01 -7.89075717e-02 1.09048855e+00 1.12348564e-01
-7.07494497e-01 5.80792010e-01 4.01223689e-01 -3.61271165e-02
5.92418015e-01 -1.13793278e+00 -3.79840322e-02 5.48785210e-01
1.25118140e-02 3.81847978e-01 6.59807384e-01 -5.70885777e-01
-1.63759899e+00 -1.31006682e+00 7.55842805e-01 -2.45364606e-01
7.15988874e-01 -3.59603614e-01 -1.15485942e+00 6.80820763e-01
-2.38082670e-02 2.50968516e-01 1.40041995e+00 -1.75190285e-01
1.37956990e-02 -4.19232436e-02 -9.89273787e-01 3.76067042e-01
6.65637016e-01 -4.86082822e-01 1.88504115e-01 8.46939623e-01
4.60529655e-01 -5.44921637e-01 -1.16080129e+00 1.24916092e-01
4.96486902e-01 -6.85861528e-01 7.71138489e-01 -7.36923277e-01
5.91529965e-01 -6.23274036e-03 9.87775177e-02 -1.23197949e+00
-3.24126869e-01 -5.73505044e-01 -3.65287930e-01 5.34944415e-01
6.42983615e-01 -5.51066279e-01 1.33510518e+00 2.79421270e-01
-1.50991976e-01 -1.28069758e+00 -1.10995138e+00 -4.59761798e-01
-1.41588496e-02 -5.46873175e-03 2.51139432e-01 7.52815008e-01
3.52692425e-01 7.28837252e-02 -1.45330504e-01 -5.91786131e-02
7.19565213e-01 2.00928375e-01 3.28439891e-01 -9.72372532e-01
-4.49556619e-01 -5.40041804e-01 -6.69845402e-01 -4.50313300e-01
2.91405231e-01 -1.22022450e+00 2.92530823e-02 -1.25686705e+00
6.55329764e-01 -3.99597287e-01 -6.74007893e-01 4.00642604e-01
-2.32933402e-01 1.92310184e-01 -2.06164867e-01 2.83887863e-01
-8.05428267e-01 5.54772420e-03 1.03242326e+00 -4.73594218e-01
-5.82654253e-02 1.21984571e-01 -3.78161460e-01 2.26588041e-01
9.28054750e-01 -7.15741634e-01 -8.05716589e-02 -1.26033649e-01
2.83166796e-01 4.60227370e-01 2.39141032e-01 -9.25241530e-01
7.61809528e-01 -7.63059184e-02 4.09469903e-01 -6.46980047e-01
4.66114461e-01 -5.70798099e-01 6.59909189e-01 1.00581479e+00
-7.05784500e-01 -1.53936058e-01 3.70935760e-02 7.24732995e-01
-4.41382319e-01 -2.35500056e-02 7.83473313e-01 -4.14448351e-01
-3.49522144e-01 6.61373258e-01 -5.19061029e-01 -4.04634982e-01
1.09575748e+00 -1.57519296e-01 -3.72921288e-01 6.57635853e-02
-9.57146943e-01 1.73436508e-01 3.32537413e-01 -8.80990624e-02
6.39398396e-01 -9.42899823e-01 -8.73664737e-01 1.85854644e-01
1.10213824e-01 1.67974189e-01 8.87579143e-01 1.20619512e+00
-3.49593669e-01 5.27293563e-01 -8.67216140e-02 -8.66382718e-01
-1.58340275e+00 5.06220460e-01 3.26301724e-01 -1.08170593e+00
-5.90845287e-01 9.17222798e-01 1.71977699e-01 -3.06918085e-01
1.69687405e-01 -6.44895136e-02 -5.44307493e-02 1.67335153e-01
7.81440914e-01 9.58592892e-02 3.23024988e-01 -1.94441468e-01
-2.23760545e-01 2.58865774e-01 -6.02748871e-01 1.98711351e-01
1.53503370e+00 2.25160956e-01 -3.50506455e-01 6.88233852e-01
1.52053535e+00 -7.31009662e-01 -9.76412475e-01 -2.01299146e-01
1.28726095e-01 4.50788885e-02 -1.10530496e-01 -4.07779276e-01
-9.72085893e-01 9.26620722e-01 6.03150904e-01 2.25033686e-01
1.17346418e+00 7.10283890e-02 6.91977859e-01 3.21535736e-01
1.85253695e-01 -7.24594295e-01 9.36195478e-02 2.45168433e-01
4.41951424e-01 -9.97738123e-01 -9.59320739e-02 -1.39915451e-01
-2.64498204e-01 1.06938398e+00 1.80722609e-01 -3.30805667e-02
6.65262640e-01 7.73950279e-01 3.13782692e-03 -2.78721422e-01
-1.55868220e+00 2.77112424e-01 -2.61292100e-01 1.94906369e-01
5.55120766e-01 1.16546735e-01 -5.41162863e-02 3.44917417e-01
9.72444378e-03 2.23252565e-01 6.28085732e-01 1.26149023e+00
-2.96082914e-01 -1.19104302e+00 -1.49659172e-01 1.19832373e+00
-8.25755954e-01 1.49274180e-02 3.50877978e-02 2.72037089e-01
-4.73507136e-01 2.52228498e-01 1.77051276e-01 -1.87662646e-01
-2.37853527e-01 3.78879666e-01 3.86221975e-01 -7.14541674e-01
-7.05167532e-01 1.77290559e-01 -3.34193021e-01 -5.59011281e-01
-3.46975088e-01 -8.87590468e-01 -1.04671896e+00 -1.61305983e-02
-9.17076394e-02 1.99860483e-02 6.59399688e-01 7.79555023e-01
4.16615725e-01 7.41211355e-01 5.59441626e-01 -7.38646030e-01
-6.41256273e-01 -9.37324762e-01 -7.32203484e-01 1.65248767e-01
6.65227413e-01 -2.23125920e-01 -4.37782377e-01 3.83260310e-01] | [15.117100715637207, -2.8752505779266357] |
5e0bff08-4359-4f1c-863a-a556c28844f4 | lattice-protein-design-using-bayesian | 2003.06601 | null | https://arxiv.org/abs/2003.06601v5 | https://arxiv.org/pdf/2003.06601v5.pdf | Lattice protein design using Bayesian learning | Protein design is the inverse approach of the three-dimensional (3D) structure prediction for elucidating the relationship between the 3D structures and amino acid sequences. In general, the computation of the protein design involves a double loop: a loop for amino acid sequence changes and a loop for an exhaustive conformational search for each amino acid sequence. Herein, we propose a novel statistical mechanical design method using Bayesian learning, which can design lattice proteins without the exhaustive conformational search. We consider a thermodynamic hypothesis of the evolution of proteins and apply it to the prior distribution of amino acid sequences. Furthermore, we take the water effect into account in view of the grand canonical picture. As a result, on applying the 2D lattice hydrophobic-polar (HP) model, our design method successfully finds an amino acid sequence for which the target conformation has a unique ground state. However, the performance was not as good for the 3D lattice HP models compared to the 2D models. The performance of the 3D model improves on using a 20-letter lattice proteins. Furthermore, we find a strong linearity between the chemical potential of water and the number of surface residues, thereby revealing the relationship between protein structure and the effect of water molecules. The advantage of our method is that it greatly reduces computation time, because it does not require long calculations for the partition function corresponding to an exhaustive conformational search. As our method uses a general form of Bayesian learning and statistical mechanics and is not limited to lattice proteins, the results presented here elucidate some heuristics used successfully in previous protein design methods. | ['Tomoei Takahashi', 'Kei Tokita', 'George Chikenji'] | 2020-03-14 | null | null | null | null | ['protein-design'] | ['medical'] | [ 1.74135908e-01 6.07540831e-02 -1.32506654e-01 -7.50074387e-02
-1.97227746e-01 -4.66038048e-01 4.38080460e-01 7.35755265e-02
-4.93058443e-01 1.12833285e+00 -1.22539975e-01 -7.11889505e-01
9.08511803e-02 -7.41053760e-01 -9.64338601e-01 -1.31103444e+00
-1.08174115e-01 7.08317876e-01 5.61766803e-01 -2.52505898e-01
4.54658866e-01 7.38354146e-01 -1.24981105e+00 3.02264523e-02
7.92448878e-01 5.79796672e-01 6.24952078e-01 2.94419318e-01
-2.22717524e-01 2.68393569e-02 -6.59992769e-02 5.14243804e-02
1.45325512e-01 -6.01323247e-01 -6.49567425e-01 -5.97331785e-02
-5.40586375e-02 1.61134854e-01 8.94541889e-02 8.41938257e-01
6.25667095e-01 4.61729318e-02 1.02174473e+00 -3.00579637e-01
-4.55597073e-01 2.75340769e-02 -4.41739231e-01 -3.43156785e-01
5.14059305e-01 4.19371814e-01 1.04826999e+00 -9.41937327e-01
8.54057610e-01 1.11582601e+00 5.32779336e-01 3.81564617e-01
-1.72363865e+00 -2.93048710e-01 -3.78790163e-02 3.40231210e-01
-1.37471604e+00 -1.00490288e-03 7.24935949e-01 -6.65311515e-01
1.19119394e+00 7.41816983e-02 9.41052020e-01 7.92214274e-01
6.44507349e-01 2.62875527e-01 1.25591123e+00 -8.39926541e-01
7.78334260e-01 -3.18114609e-01 1.83053657e-01 7.56907940e-01
3.09670299e-01 2.71284342e-01 -3.02799910e-01 -7.36587167e-01
2.68233865e-01 -1.10088140e-01 -2.33928934e-01 -9.81000781e-01
-8.07736874e-01 8.77031624e-01 1.33719770e-02 2.25631952e-01
-4.43409532e-01 -1.36105835e-01 6.63300902e-02 -2.74783999e-01
-9.10060015e-03 6.83626533e-01 -7.68509626e-01 1.29105374e-01
-7.51761198e-01 5.58539987e-01 9.89275873e-01 4.69179451e-01
1.02030671e+00 -4.27053630e-01 3.24304998e-01 3.55774790e-01
6.06185675e-01 6.16299391e-01 1.05380543e-01 -6.84993804e-01
-1.14684448e-01 4.00914520e-01 3.68676960e-01 -6.89444304e-01
-4.13201541e-01 1.43183842e-01 -5.98149717e-01 4.42683488e-01
5.37418962e-01 2.02682838e-01 -6.79829955e-01 1.59451497e+00
4.38522100e-01 -2.73939341e-01 1.68772683e-01 6.26518428e-01
2.93884814e-01 6.55010343e-01 -3.68736088e-02 -9.91890192e-01
1.48348737e+00 -3.38212609e-01 -5.74317276e-01 2.41843313e-01
5.71738362e-01 -7.88191319e-01 6.71320200e-01 6.15490615e-01
-8.30212176e-01 -4.05881554e-01 -1.23972666e+00 3.19263160e-01
-8.93274620e-02 -1.78173393e-01 5.15575826e-01 5.60570836e-01
-7.01761842e-01 8.08109462e-01 -8.29179466e-01 -4.12730068e-01
-2.76702017e-01 5.41586995e-01 -3.08060646e-01 2.67697304e-01
-1.27636933e+00 1.24847209e+00 7.68582582e-01 -5.55465929e-02
-3.26763362e-01 -2.53669381e-01 -3.92411053e-01 -2.16877639e-01
2.75712103e-01 -6.37343228e-01 8.44186246e-01 -5.28130114e-01
-1.78557158e+00 7.05927134e-01 -6.12199724e-01 -2.66120046e-01
1.07118517e-01 1.32929757e-01 1.04411334e-01 8.25813785e-02
-4.15830255e-01 5.09984016e-01 3.81582767e-01 -1.11825979e+00
5.61311357e-02 -3.94067377e-01 -4.05312091e-01 7.47032315e-02
4.29979324e-01 -2.44387433e-01 5.26775718e-02 -2.31344000e-01
4.89964277e-01 -1.20110154e+00 -7.11790323e-01 -2.67309427e-01
-2.78656751e-01 -3.44796419e-01 3.38751942e-01 -4.21698034e-01
1.17388666e+00 -1.86940396e+00 5.95079720e-01 7.18925536e-01
1.12832844e-01 1.93594411e-01 3.15559059e-01 8.06923270e-01
-4.16619122e-01 7.99425468e-02 -2.68464416e-01 5.60207069e-01
-2.13957757e-01 2.31262997e-01 -1.79293394e-01 5.02859235e-01
-4.19125557e-02 6.11302435e-01 -5.45095980e-01 -2.81810433e-01
2.82427240e-02 3.71933460e-01 -8.35611343e-01 7.97506198e-02
-6.37625813e-01 4.85930115e-01 -6.70200765e-01 2.04265490e-01
8.65424633e-01 -1.72451049e-01 1.03154278e+00 -4.04685140e-01
-5.58681786e-01 3.62947345e-01 -9.33500409e-01 1.44314885e+00
3.12494606e-01 -1.44647956e-01 -3.36509079e-01 -7.31247663e-01
1.34534013e+00 2.19465151e-01 6.98195517e-01 -5.40993869e-01
-1.91593915e-01 4.95171368e-01 3.86821628e-01 -4.53284860e-01
-1.17406383e-01 -4.91754949e-01 1.67651191e-01 4.35700744e-01
-2.02423722e-01 -7.58945942e-02 1.44995883e-01 -2.16517851e-01
7.90560901e-01 7.96157300e-01 7.93630242e-01 -5.56945562e-01
6.59331024e-01 -1.42097846e-02 7.70115077e-01 3.70061547e-01
1.35352716e-01 2.66697615e-01 8.37872744e-01 -7.93577433e-01
-1.59660757e+00 -7.77215302e-01 -3.80696177e-01 5.82206249e-01
1.84640840e-01 -4.91618425e-01 -9.67159748e-01 -2.35636801e-01
7.56432535e-03 5.09288728e-01 -2.75075465e-01 -1.64705679e-01
-7.63158739e-01 -1.26943994e+00 9.23081636e-02 -2.69672692e-01
-1.04404807e-01 -1.07502460e+00 -7.59142399e-01 4.91719633e-01
6.15738286e-03 -4.96526152e-01 -2.06882238e-01 6.15315795e-01
-1.05058289e+00 -1.27697670e+00 -4.45960939e-01 -5.60420692e-01
5.00455201e-01 -9.38460082e-02 6.83834553e-01 -9.59227979e-02
-3.21244746e-01 -2.26409465e-01 -2.02997312e-01 -1.45890236e-01
-6.76862717e-01 -6.39016703e-02 4.56229299e-01 -3.18779975e-01
8.38563383e-01 -8.50765586e-01 -6.91661596e-01 3.11598003e-01
-5.72807193e-01 1.47213563e-01 7.23400533e-01 9.25145805e-01
9.29408848e-01 -2.53555365e-02 2.03088373e-01 -5.20694375e-01
2.99888819e-01 -1.16049923e-01 -7.08605707e-01 4.17185605e-01
-8.71056676e-01 7.24858582e-01 2.91970164e-01 -2.22234234e-01
-8.84060502e-01 6.59322739e-01 -2.54246891e-01 1.97658896e-01
-1.87136054e-01 3.28252107e-01 -4.24132526e-01 -6.21076487e-02
6.99066639e-01 4.34435815e-01 4.59092379e-01 -7.81255960e-01
6.07421435e-02 3.63978088e-01 -1.66787088e-01 -9.99767244e-01
4.15015370e-01 1.28573060e-01 4.54961538e-01 -9.60785210e-01
-1.93299070e-01 -5.79804964e-02 -1.10874104e+00 1.59270726e-02
9.71622944e-01 -3.60610187e-01 -1.29621792e+00 2.40553662e-01
-1.47220373e+00 1.09948955e-01 2.02504084e-01 7.50383198e-01
-9.32336271e-01 1.17882514e+00 -1.83708608e-01 -8.69084597e-01
-7.78360367e-02 -1.68304956e+00 7.39223778e-01 -2.88679123e-01
-5.11292100e-01 -6.99607193e-01 4.73136485e-01 1.86608180e-01
-1.53184056e-01 2.89728284e-01 1.70221460e+00 -4.77617085e-01
-8.03377211e-01 2.40004435e-01 3.94363761e-01 -7.59490505e-02
4.31050733e-02 1.76037654e-01 -4.28456992e-01 -3.14105272e-01
7.48021826e-02 -9.50391442e-02 8.06206703e-01 4.61661786e-01
6.54892802e-01 -1.89391851e-01 -4.56150353e-01 2.20125198e-01
1.48549974e+00 8.50340009e-01 7.30753064e-01 4.00623888e-01
3.90730202e-01 7.46596158e-01 6.36726677e-01 3.80591542e-01
-1.26494899e-01 1.07216084e+00 3.33407760e-01 9.33163334e-03
3.74272496e-01 -2.36094654e-01 3.97993684e-01 6.50508761e-01
-4.27837908e-01 1.14263833e-01 -1.11136734e+00 -2.08056778e-01
-1.92582524e+00 -9.72846985e-01 -3.76672000e-01 2.53687191e+00
1.35890627e+00 1.28654078e-01 1.81432933e-01 -1.30408555e-01
4.88280177e-01 -2.74624586e-01 -8.09174180e-01 -5.07453382e-01
-2.23227724e-01 2.98538268e-01 5.79375863e-01 7.84611583e-01
-7.64187992e-01 8.04713428e-01 7.32620811e+00 8.25852454e-01
-8.31624627e-01 -3.46311212e-01 2.14725211e-01 1.86178088e-01
-4.21480924e-01 5.26063442e-01 -1.06033802e+00 4.70878571e-01
7.52438188e-01 1.97862014e-01 2.55770475e-01 6.25168383e-01
5.62295854e-01 -4.65653092e-01 -9.93535876e-01 5.89415491e-01
-4.31235641e-01 -1.40925467e+00 2.04556271e-01 4.78407830e-01
5.01900434e-01 -3.48629385e-01 -4.05724585e-01 -4.39544588e-01
-2.93292254e-02 -9.15416479e-01 4.07148778e-01 6.86797023e-01
6.17998660e-01 -8.46875072e-01 5.59022546e-01 5.32208025e-01
-7.62343824e-01 3.49268466e-01 -5.28321862e-01 -1.06079079e-01
1.53772637e-01 6.42644763e-01 -9.52671468e-01 1.74092308e-01
3.23841751e-01 1.87856570e-01 -7.93932155e-02 7.11592197e-01
-4.72581722e-02 3.19022655e-01 -3.94862980e-01 -4.00853485e-01
9.24331322e-02 -8.95863414e-01 5.78248560e-01 8.78863752e-01
1.77851483e-01 3.83374661e-01 3.78350824e-01 8.01733017e-01
6.17848456e-01 3.63573521e-01 -5.31370401e-01 2.13864353e-02
2.82654285e-01 4.05079275e-01 -6.06972039e-01 7.98022971e-02
-2.23441273e-01 5.05752981e-01 7.88904503e-02 4.56936806e-01
-5.39955735e-01 -1.27484530e-01 7.12023497e-01 2.03224212e-01
6.65660441e-01 -4.23776418e-01 1.41902305e-02 -8.09693635e-01
-9.09268260e-02 -9.47131991e-01 -3.06795299e-01 -4.13898528e-01
-9.35191929e-01 2.17540637e-01 4.64959629e-03 -7.11810470e-01
-3.47328603e-01 -8.78687024e-01 -1.20988168e-01 1.15506172e+00
-1.15618563e+00 -5.14678419e-01 6.88778698e-01 3.86844203e-02
9.36294273e-02 -2.51217037e-01 1.14063144e+00 -3.19401056e-01
-2.72921979e-01 1.67355046e-01 6.39857531e-01 -6.74286842e-01
7.53126264e-01 -1.19867980e+00 2.16229036e-01 3.87174994e-01
-4.10009444e-01 9.55164015e-01 1.24620044e+00 -1.08202362e+00
-1.44771254e+00 -3.22660148e-01 9.87613022e-01 -2.84331501e-01
4.32864577e-01 -3.38133961e-01 -1.05009890e+00 -9.61852167e-03
1.29303196e-02 -7.52131820e-01 1.05744171e+00 1.94870740e-01
-1.91335186e-01 2.73767650e-01 -9.59054232e-01 5.49806535e-01
8.41009498e-01 -2.80003279e-01 -8.02640140e-01 4.53811020e-01
6.88741326e-01 1.24235684e-02 -9.81324613e-01 5.00939131e-01
8.29892457e-01 -1.02476656e+00 1.01603520e+00 -6.71866953e-01
3.35740438e-03 -6.82991743e-01 -2.72809505e-01 -7.29465485e-01
-6.68301582e-01 -6.79153085e-01 -5.72546273e-02 3.88151467e-01
6.10604942e-01 -5.60389578e-01 7.77265310e-01 5.25890946e-01
1.58532977e-01 -1.05426967e+00 -1.00186074e+00 -6.91965997e-01
2.75910109e-01 9.61832702e-02 4.09428835e-01 5.81325531e-01
1.78383872e-01 4.30125505e-01 -4.23767447e-01 4.69815684e-04
6.90600336e-01 3.56781006e-01 4.91694003e-01 -1.37006557e+00
-7.38588691e-01 -1.52920321e-01 -1.47585049e-01 -1.17644095e+00
1.23384260e-01 -8.01455200e-01 -1.32936135e-01 -9.45203006e-01
4.31032270e-01 -2.73698777e-01 -1.06555119e-01 2.82749444e-01
1.04906604e-01 -4.26908493e-01 -6.34204596e-02 4.46185827e-01
-1.62460789e-01 5.87953210e-01 1.24948108e+00 1.54159740e-01
-4.47011173e-01 3.85857560e-03 -2.73987502e-01 6.19433343e-01
6.62244260e-01 -5.42646229e-01 -1.61617428e-01 4.81384277e-01
5.77620089e-01 -2.21956745e-02 -2.09156051e-02 -4.63285834e-01
-2.00623684e-02 -4.14560288e-01 2.28165612e-01 -9.28504109e-01
2.17967391e-01 -9.46521997e-01 5.75556397e-01 9.97010052e-01
5.56816868e-02 -1.62776977e-01 -1.00573830e-01 7.16809630e-01
1.84933975e-01 -6.01505578e-01 8.46746683e-01 -2.88527697e-01
-4.54875112e-01 -5.05335070e-03 -7.94286549e-01 -6.78142190e-01
1.08162642e+00 -5.48177540e-01 2.75052458e-01 2.97661096e-01
-1.07494807e+00 -8.82803649e-02 9.78956044e-01 -1.93978637e-01
4.47530955e-01 -1.08470678e+00 -2.88471192e-01 3.36347878e-01
-1.32181346e-01 -2.58943349e-01 3.38939615e-02 6.55899286e-01
-9.49021935e-01 8.20056677e-01 -2.31518373e-01 -7.29482114e-01
-1.36242640e+00 8.68156910e-01 4.29167241e-01 -2.36937255e-01
-4.23602790e-01 2.79366314e-01 2.33285636e-01 -4.45250779e-01
-5.62699996e-02 -4.73744906e-02 -1.38319746e-01 2.58492287e-02
3.10848713e-01 1.38181195e-01 -7.45789409e-02 -3.88295174e-01
-5.02412915e-01 1.04012549e+00 -1.69908732e-01 1.00513510e-01
1.35415733e+00 1.94366071e-02 -4.41512674e-01 3.28638345e-01
9.05296326e-01 -9.49392989e-02 -1.28685272e+00 -1.96061134e-01
2.23728120e-01 -4.69527878e-02 -3.95291060e-01 -5.76235771e-01
-9.99311283e-02 6.23077571e-01 6.80733025e-01 -2.25940526e-01
7.48359680e-01 -7.87061732e-03 5.18865168e-01 8.80102813e-01
6.80165470e-01 -9.47081923e-01 -2.10542098e-01 6.62756503e-01
7.54034638e-01 -8.48723650e-01 3.32641304e-01 -3.31942976e-01
-2.40564331e-01 1.42346072e+00 1.79655999e-01 1.24438405e-01
5.92481971e-01 8.49781092e-03 -3.04161370e-01 -1.89863175e-01
-7.64330685e-01 -7.14037120e-02 1.12025000e-01 4.13475901e-01
5.53258538e-01 1.47522494e-01 -1.03138626e+00 1.65647149e-01
-5.18672317e-02 -7.66506493e-02 2.42818728e-01 8.71639073e-01
-8.55440617e-01 -1.88476801e+00 -4.75419402e-01 -3.23658258e-01
-1.51970983e-01 -1.50904000e-01 -5.87850332e-01 5.70588291e-01
2.36680940e-01 5.30038595e-01 -2.65660197e-01 -1.32470310e-01
2.59229273e-01 6.88832819e-01 7.99439251e-01 -3.98304343e-01
-8.93697068e-02 4.09247458e-01 -3.51314358e-02 -3.16359103e-01
-5.08324206e-01 -7.97958076e-01 -1.44024050e+00 -3.36523980e-01
-5.09464681e-01 7.02152848e-01 7.27446735e-01 9.77869928e-01
4.34791625e-01 1.32056419e-02 6.50214612e-01 -8.26766312e-01
-5.90091705e-01 -5.94603300e-01 -5.99744081e-01 3.26247364e-02
3.87918875e-02 -8.22210848e-01 -1.63640186e-01 -7.80046433e-02] | [4.801727294921875, 5.311549663543701] |
78b9602c-fb25-4079-84f1-d29bd2d7a269 | distilled-semantics-for-comprehensive-scene | 2003.14030 | null | https://arxiv.org/abs/2003.14030v1 | https://arxiv.org/pdf/2003.14030v1.pdf | Distilled Semantics for Comprehensive Scene Understanding from Videos | Whole understanding of the surroundings is paramount to autonomous systems. Recent works have shown that deep neural networks can learn geometry (depth) and motion (optical flow) from a monocular video without any explicit supervision from ground truth annotations, particularly hard to source for these two tasks. In this paper, we take an additional step toward holistic scene understanding with monocular cameras by learning depth and motion alongside with semantics, with supervision for the latter provided by a pre-trained network distilling proxy ground truth images. We address the three tasks jointly by a) a novel training protocol based on knowledge distillation and self-supervision and b) a compact network architecture which enables efficient scene understanding on both power hungry GPUs and low-power embedded platforms. We thoroughly assess the performance of our framework and show that it yields state-of-the-art results for monocular depth estimation, optical flow and motion segmentation. | ['Luigi Di Stefano', 'Pierluigi Zama Ramirez', 'Filippo Aleotti', 'Stefano Mattoccia', 'Samuele Salti', 'Matteo Poggi', 'Fabio Tosi'] | 2020-03-31 | distilled-semantics-for-comprehensive-scene-1 | http://openaccess.thecvf.com/content_CVPR_2020/html/Tosi_Distilled_Semantics_for_Comprehensive_Scene_Understanding_from_Videos_CVPR_2020_paper.html | http://openaccess.thecvf.com/content_CVPR_2020/papers/Tosi_Distilled_Semantics_for_Comprehensive_Scene_Understanding_from_Videos_CVPR_2020_paper.pdf | cvpr-2020-6 | ['motion-segmentation'] | ['computer-vision'] | [ 2.24149629e-01 6.93852827e-02 -1.26496151e-01 -4.29923564e-01
-2.26096362e-01 -7.44317114e-01 6.95180833e-01 -3.38075846e-01
-6.33034468e-01 8.80270064e-01 3.30788791e-02 -2.47468174e-01
2.86711425e-01 -7.45021105e-01 -9.89057541e-01 -5.50491750e-01
4.44524698e-02 3.12424779e-01 3.70163083e-01 9.94769335e-02
3.01485747e-01 5.12388110e-01 -1.79011738e+00 3.16857137e-02
6.60559893e-01 9.74556863e-01 3.25550526e-01 1.30147278e+00
3.12809020e-01 1.28873742e+00 -8.09242949e-02 1.83321480e-02
4.74037290e-01 -1.61336452e-01 -1.19671202e+00 1.03425257e-01
9.42267358e-01 -9.68080461e-01 -7.61745453e-01 8.53808224e-01
2.25756705e-01 1.21325001e-01 3.33375663e-01 -1.11194277e+00
-2.77449600e-02 1.72124139e-03 -2.50987917e-01 3.07866216e-01
4.40701842e-01 5.83542764e-01 5.67215323e-01 -4.96622920e-01
8.42765689e-01 1.05189812e+00 5.09554982e-01 5.83021343e-01
-7.74275124e-01 -2.26674750e-01 2.86231697e-01 3.13041478e-01
-8.84445429e-01 -4.93674189e-01 5.32575071e-01 -5.27213752e-01
1.23732424e+00 -2.62881577e-01 8.54018807e-01 8.36186469e-01
4.22945572e-03 8.44550788e-01 8.52494180e-01 -2.75203705e-01
3.65685850e-01 -1.03736825e-01 -1.30159318e-01 9.41845655e-01
3.32819253e-01 3.38985533e-01 -6.19421601e-01 3.68059754e-01
1.09087861e+00 -1.95735350e-01 -5.27176917e-01 -6.12321019e-01
-1.16594756e+00 5.90817750e-01 5.69502771e-01 -6.57639280e-02
-2.00772315e-01 7.82261491e-01 4.06100243e-01 8.22715387e-02
3.58965397e-01 2.36380696e-01 -7.21803784e-01 -3.14613074e-01
-9.42207277e-01 8.50470066e-02 9.47318852e-01 9.67273712e-01
1.23908496e+00 9.62029994e-02 4.01730895e-01 6.69528469e-02
1.79684043e-01 4.47512865e-01 2.60093182e-01 -1.63883913e+00
3.46124113e-01 3.53742748e-01 2.49290794e-01 -7.28539944e-01
-4.90891427e-01 -1.27981946e-01 -6.18150949e-01 4.23697680e-01
7.41635084e-01 -4.06453907e-01 -8.54466200e-01 1.59790158e+00
5.68049252e-01 5.30724466e-01 1.26430064e-01 1.14724565e+00
7.74733722e-01 3.34341913e-01 -2.27098241e-01 2.83873111e-01
1.00581336e+00 -1.32585704e+00 -2.75456548e-01 -6.26987159e-01
8.86949301e-01 -4.45274144e-01 7.41937518e-01 5.16454875e-01
-1.03780603e+00 -5.69939554e-01 -1.28947937e+00 -6.43245339e-01
-2.90059239e-01 7.19456300e-02 1.05098403e+00 6.72190070e-01
-1.51103103e+00 7.28459418e-01 -1.28763247e+00 -3.50457489e-01
5.48352540e-01 5.51983774e-01 -4.84946907e-01 -3.59858274e-01
-8.17964554e-01 6.30496562e-01 4.66863424e-01 -1.58254826e-03
-1.04645598e+00 -7.36815751e-01 -1.30623233e+00 -2.61872709e-01
4.30805117e-01 -1.36741865e+00 1.13200915e+00 -9.88347650e-01
-1.89450192e+00 9.85443473e-01 -1.21606231e-01 -7.55201161e-01
8.26201737e-01 -7.20921159e-01 4.90536958e-01 6.76735640e-01
-1.39983548e-02 1.18141234e+00 7.26193428e-01 -1.13323021e+00
-8.03154886e-01 -4.74438816e-01 4.43347871e-01 4.13550407e-01
2.57496238e-01 -5.20015359e-01 -4.39434230e-01 1.30435511e-01
4.02479023e-02 -9.08438265e-01 -2.23449752e-01 3.92601669e-01
-3.78616124e-01 2.90405482e-01 9.79404986e-01 -6.46328866e-01
3.81348401e-01 -1.64863682e+00 1.58858538e-01 -2.84199148e-01
2.53487468e-01 2.45078295e-01 1.54603288e-01 -1.01298548e-01
2.17893526e-01 -4.04589325e-01 -2.14230150e-01 -5.33822954e-01
-1.66033566e-01 5.15977442e-01 -1.85643807e-01 7.26448596e-01
2.12558687e-01 1.13903940e+00 -1.23961413e+00 -2.97906518e-01
8.65842223e-01 6.58456385e-01 -8.82288039e-01 2.94165164e-01
-3.32023770e-01 8.93034995e-01 -4.85161953e-02 5.31794727e-01
6.75042510e-01 -3.40803772e-01 1.81514621e-01 -1.55168533e-01
-2.42962033e-01 3.67933929e-01 -1.05657887e+00 2.22877169e+00
-6.41901970e-01 9.95542169e-01 2.62626052e-01 -8.45775843e-01
4.86753404e-01 1.94586301e-03 4.58856881e-01 -4.73257840e-01
4.17032540e-01 1.94715723e-01 -4.51006651e-01 -5.13524950e-01
7.04171777e-01 6.06457219e-02 2.92353511e-01 3.72950703e-01
1.79052845e-01 -6.15668952e-01 5.18142134e-02 8.22550952e-02
1.14992332e+00 8.39312613e-01 2.23940566e-01 -1.34533823e-01
5.60432076e-01 1.88293993e-01 3.12586486e-01 6.72664285e-01
-3.10232788e-01 6.59179628e-01 2.23288119e-01 -7.25349545e-01
-1.05069959e+00 -8.40703905e-01 2.05228031e-01 6.55780911e-01
5.44321597e-01 -3.00498214e-02 -8.36492062e-01 -4.44343776e-01
-1.74963191e-01 2.46581286e-01 -6.22525394e-01 1.25081152e-01
-6.90686524e-01 -5.02399743e-01 3.04493010e-01 8.28813732e-01
9.63810980e-01 -7.71415591e-01 -1.47263443e+00 6.78282455e-02
-8.19680467e-02 -2.02951598e+00 -1.58909187e-01 2.54291266e-01
-1.03387523e+00 -1.28826189e+00 -5.61225891e-01 -5.87421119e-01
4.83971894e-01 5.20537972e-01 1.25609839e+00 -2.63863597e-02
-3.73667598e-01 6.59989357e-01 -1.76435970e-02 -6.40556589e-02
5.30939922e-02 2.12534711e-01 -1.65627494e-01 -9.68237594e-02
6.42512068e-02 -7.78397620e-01 -1.08970380e+00 5.14870621e-02
-8.19541752e-01 4.06788558e-01 4.11437780e-01 5.61159730e-01
4.20312256e-01 -4.07202393e-01 -1.58954397e-01 -7.43040502e-01
-2.59009659e-01 -3.70623738e-01 -9.84927833e-01 -3.47098291e-01
-1.86154187e-01 2.38218103e-02 5.81964374e-01 1.03556521e-01
-1.08146262e+00 6.33755445e-01 5.60448021e-02 -5.04223943e-01
-4.86309111e-01 -5.37067689e-02 -6.37715459e-02 -4.67204779e-01
5.88746250e-01 2.84677714e-01 -8.00802857e-02 -5.54669499e-02
7.23076820e-01 2.10997075e-01 1.15113485e+00 -5.43857038e-01
5.96671999e-01 1.26477981e+00 2.63999403e-01 -9.89409804e-01
-1.06933928e+00 -6.24371529e-01 -1.34115636e+00 -1.25227854e-01
1.24590719e+00 -1.38193011e+00 -9.29729700e-01 6.41173780e-01
-1.30380476e+00 -9.03563797e-01 -1.88414216e-01 5.60864329e-01
-1.05220711e+00 6.34645224e-01 -6.16540015e-01 -5.96772492e-01
-2.83620298e-01 -9.46874321e-01 1.39946306e+00 3.13245803e-01
-2.80536320e-02 -1.29314160e+00 1.00388758e-01 6.04454875e-01
7.22215548e-02 5.34478188e-01 3.54588151e-01 2.04869419e-01
-1.37504196e+00 1.41613960e-01 -6.46471798e-01 3.49437416e-01
-8.22402313e-02 -2.56129086e-01 -1.43445957e+00 -1.92317218e-01
1.74977109e-02 -4.86225903e-01 1.02230430e+00 6.49318159e-01
9.17218626e-01 -1.80737078e-02 -2.18711734e-01 1.38422561e+00
1.66959536e+00 -4.26362045e-02 6.57596946e-01 5.11664152e-01
1.22983551e+00 5.24329185e-01 3.09599340e-01 3.35909396e-01
7.56329954e-01 4.54925299e-01 7.40290880e-01 -1.31830931e-01
-2.55540431e-01 -1.21200152e-01 1.23116225e-01 4.80962276e-01
-1.38958871e-01 -3.13659199e-02 -8.94661486e-01 8.43232512e-01
-2.00034046e+00 -6.59355640e-01 -1.26612425e-01 1.95029020e+00
5.42622983e-01 -7.20411018e-02 -2.83708517e-02 1.44079596e-01
1.40064746e-01 1.51440859e-01 -8.21819544e-01 -1.78479701e-01
-2.15575501e-01 1.51213169e-01 7.27162957e-01 9.19482946e-01
-1.31242967e+00 1.31912827e+00 6.35092878e+00 7.67206103e-02
-1.26088035e+00 9.34775546e-02 6.49874032e-01 -2.58472204e-01
-6.02713116e-02 1.25550672e-01 -6.86015666e-01 -7.83803593e-03
7.00415850e-01 3.54631484e-01 3.97654921e-01 8.69875729e-01
2.10552856e-01 -7.80592620e-01 -1.29020345e+00 1.26014590e+00
3.94716822e-02 -1.40472543e+00 -1.39970064e-01 5.68944588e-02
1.15441406e+00 5.44278026e-01 -3.22502285e-01 -2.81489849e-01
4.89653736e-01 -1.04879105e+00 5.55086911e-01 3.43632847e-01
6.55096173e-01 -5.50922096e-01 8.36648762e-01 3.19286168e-01
-1.15703452e+00 1.24476083e-01 -2.29366526e-01 -6.21661901e-01
2.71754861e-01 5.19027412e-01 -5.79028666e-01 6.80039227e-01
5.82302809e-01 1.29706633e+00 -2.57248878e-01 8.10637116e-01
-3.65203530e-01 1.74924344e-01 -5.36828279e-01 3.49294156e-01
5.24515629e-01 -7.74262026e-02 2.68113256e-01 1.03954005e+00
-4.29870375e-02 1.71294555e-01 1.06703900e-01 8.16899359e-01
1.15080606e-02 -3.88933897e-01 -6.78587794e-01 3.68751347e-01
1.38582168e-02 1.07702255e+00 -8.30181122e-01 -4.23874736e-01
-4.29717362e-01 1.21917045e+00 3.72505426e-01 2.72230566e-01
-5.77054203e-01 -1.81002930e-01 9.48679268e-01 -3.25992401e-03
3.82204980e-01 -6.94474995e-01 -6.06087327e-01 -1.54034722e+00
-4.73687053e-03 -1.54305100e-01 9.63388234e-02 -1.08032656e+00
-7.09157050e-01 4.96357620e-01 -1.61640897e-01 -8.74856234e-01
-5.03328145e-01 -9.25352812e-01 -3.32504898e-01 6.47190511e-01
-2.29225326e+00 -9.67416584e-01 -1.05388391e+00 8.08338106e-01
3.61398757e-01 2.30865598e-01 4.33788478e-01 1.78539529e-01
-2.36832798e-01 7.28875846e-02 -2.92410165e-01 1.43585101e-01
3.22956115e-01 -1.36561644e+00 6.18435800e-01 1.03311265e+00
-2.20515747e-02 2.11874336e-01 5.66531837e-01 -3.87429088e-01
-1.51528358e+00 -1.08206475e+00 4.50297028e-01 -8.10734093e-01
5.21676481e-01 -1.61344200e-01 -5.52394748e-01 7.58469820e-01
1.50078356e-01 3.60592544e-01 1.56678498e-01 -4.11609501e-01
-1.44313604e-01 2.24099964e-01 -9.39382851e-01 4.42496717e-01
1.42048621e+00 -6.76272631e-01 -2.00936943e-01 2.87086695e-01
5.89352071e-01 -7.73711324e-01 -4.89450425e-01 2.35142916e-01
4.89518255e-01 -1.51515222e+00 1.10480857e+00 -2.00682223e-01
7.03630149e-01 -3.89051467e-01 -2.13150814e-01 -8.44343483e-01
4.27911729e-01 -8.68387878e-01 -3.39120626e-01 4.14935201e-01
-2.04027459e-01 -4.32864249e-01 1.42279720e+00 5.08147717e-01
-2.05966130e-01 -5.45351624e-01 -7.90198088e-01 -5.60043991e-01
-1.29786462e-01 -7.05286801e-01 2.14914992e-01 7.83994675e-01
-3.26516837e-01 2.87969649e-01 -4.42388743e-01 2.58723438e-01
7.46977508e-01 6.33100942e-02 1.18367362e+00 -9.48089242e-01
-3.76021326e-01 -1.27612665e-01 -7.45213568e-01 -1.79813075e+00
2.36062750e-01 -5.46226680e-01 1.09374948e-01 -1.54115176e+00
2.08121315e-02 -1.05091659e-02 4.77503508e-01 4.32958037e-01
1.54917771e-02 5.15657544e-01 -1.22298645e-02 9.41061080e-02
-7.06163764e-01 6.08680725e-01 1.40189898e+00 1.77563950e-01
-2.07359120e-01 -3.52564961e-01 -3.55973870e-01 9.81598139e-01
5.12735903e-01 -1.27360418e-01 -6.87905550e-01 -8.21111441e-01
1.86812043e-01 1.51200682e-01 8.37440193e-01 -1.28514695e+00
5.05993366e-01 6.73337504e-02 3.78685713e-01 -5.08108914e-01
6.06070220e-01 -6.11645162e-01 -2.44474664e-01 5.94329774e-01
1.87852621e-01 -1.99869335e-01 3.05144638e-01 5.59448600e-01
-1.79367021e-01 8.75659809e-02 7.64367938e-01 -4.36427891e-01
-1.24338484e+00 3.52741927e-01 -1.06526501e-01 3.01218122e-01
8.19384873e-01 -6.27706826e-01 -4.12905544e-01 -5.96732914e-01
-7.18546391e-01 1.78474054e-01 6.72539175e-01 1.15656137e-01
6.03852630e-01 -8.79355431e-01 -4.10387546e-01 3.22555333e-01
-1.48548394e-01 5.97392619e-01 1.70725152e-01 6.08984649e-01
-1.36582124e+00 8.13708544e-01 -3.83338392e-01 -9.58526969e-01
-9.85760570e-01 1.55223414e-01 6.86441898e-01 -5.59992231e-02
-6.70985401e-01 9.16314483e-01 5.33495545e-01 -4.64128077e-01
4.57433537e-02 -6.91570818e-01 1.71937615e-01 -3.91518265e-01
4.33490723e-01 4.95647669e-01 -8.67936462e-02 -5.15694022e-01
-3.35957438e-01 9.53656495e-01 2.43399024e-01 -9.83259082e-02
1.18861425e+00 -4.73525584e-01 3.28818597e-02 1.89537004e-01
1.21563399e+00 -5.85706174e-01 -2.14051104e+00 -5.25175519e-02
-2.89061666e-01 -6.10560417e-01 4.41510230e-01 -5.82134247e-01
-1.13302517e+00 1.11179769e+00 3.31521034e-01 -3.62463772e-01
1.07661068e+00 -2.26464421e-01 9.61939931e-01 4.37131494e-01
6.70534849e-01 -8.96028221e-01 1.38283521e-01 8.47005904e-01
2.54510999e-01 -1.50497591e+00 2.92401500e-02 -7.35609353e-01
-1.82799712e-01 1.24301708e+00 6.27214968e-01 -3.71975958e-01
6.59802377e-01 2.08959818e-01 2.40638152e-01 -1.12538844e-01
-6.36547208e-01 -3.70835185e-01 2.01847598e-01 7.52228796e-01
1.12553976e-01 -2.29338184e-01 4.54743028e-01 -9.74971801e-02
-3.62737924e-01 3.14794600e-01 7.07078278e-01 9.88694668e-01
-3.28719497e-01 -7.23506212e-01 -3.52219716e-02 -1.76356420e-01
-2.32943431e-01 -9.31945890e-02 -3.22101563e-01 9.02971268e-01
1.91380009e-01 8.42659473e-01 1.75668269e-01 -2.21377537e-01
-4.09367085e-02 -2.50645131e-01 9.67478454e-01 -6.18702769e-01
-3.22434813e-01 -2.98439234e-01 8.47999826e-02 -1.01659906e+00
-7.05762625e-01 -6.20464027e-01 -1.20344198e+00 -4.28827018e-01
1.19065240e-01 -4.55003262e-01 8.93354297e-01 1.31448102e+00
4.05847937e-01 4.83528882e-01 1.48444980e-01 -1.55878007e+00
1.09187350e-01 -4.75441724e-01 -2.74699122e-01 1.78877845e-01
5.80006421e-01 -4.74143267e-01 -3.57923090e-01 3.04235458e-01] | [8.574851036071777, -2.1358020305633545] |
f7036e1b-9ed7-4f07-bfa7-428a31834e03 | noise-aware-unsupervised-deep-lidar-stereo | 1904.03868 | null | http://arxiv.org/abs/1904.03868v1 | http://arxiv.org/pdf/1904.03868v1.pdf | Noise-Aware Unsupervised Deep Lidar-Stereo Fusion | In this paper, we present LidarStereoNet, the first unsupervised Lidar-stereo
fusion network, which can be trained in an end-to-end manner without the need
of ground truth depth maps. By introducing a novel "Feedback Loop'' to connect
the network input with output, LidarStereoNet could tackle both noisy Lidar
points and misalignment between sensors that have been ignored in existing
Lidar-stereo fusion studies. Besides, we propose to incorporate a piecewise
planar model into network learning to further constrain depths to conform to
the underlying 3D geometry. Extensive quantitative and qualitative evaluations
on both real and synthetic datasets demonstrate the superiority of our method,
which outperforms state-of-the-art stereo matching, depth completion and
Lidar-Stereo fusion approaches significantly. | ['Pan Ji', 'Yiran Zhong', 'Yuchao Dao', 'Xuelian Cheng', 'Hongdong Li'] | 2019-04-08 | noise-aware-unsupervised-deep-lidar-stereo-1 | http://openaccess.thecvf.com/content_CVPR_2019/html/Cheng_Noise-Aware_Unsupervised_Deep_Lidar-Stereo_Fusion_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Cheng_Noise-Aware_Unsupervised_Deep_Lidar-Stereo_Fusion_CVPR_2019_paper.pdf | cvpr-2019-6 | ['stereo-matching'] | ['computer-vision'] | [ 3.18112552e-01 1.59306660e-01 1.07271463e-01 -7.27221310e-01
-8.12271237e-01 -5.05502641e-01 5.89679658e-01 7.41338581e-02
-5.64516544e-01 6.43114388e-01 7.04768747e-02 -1.51812598e-01
6.29945938e-03 -9.25364733e-01 -9.12293673e-01 -1.72243431e-01
3.39611232e-01 6.36469007e-01 3.59000117e-01 -8.90807584e-02
2.37417772e-01 7.50091851e-01 -1.65623856e+00 3.68657857e-02
9.89015937e-01 1.14240468e+00 1.95358455e-01 3.64480108e-01
-5.79469055e-02 2.24234864e-01 7.27838576e-02 -4.43397760e-01
7.30466783e-01 3.82839173e-01 -3.48364055e-01 2.38412432e-02
1.33616793e+00 -6.33257866e-01 -7.24747956e-01 9.19273913e-01
6.58087969e-01 -7.08123595e-02 5.40277600e-01 -1.19994450e+00
-1.79306671e-01 1.42350614e-01 -7.08689928e-01 -3.35968703e-01
2.88624883e-01 2.53769100e-01 8.10550690e-01 -1.19068646e+00
5.83946645e-01 1.46636355e+00 1.15369391e+00 3.08874846e-01
-1.56119084e+00 -9.82570410e-01 6.01185039e-02 -1.85473368e-01
-1.49419475e+00 -4.91755873e-01 9.80176985e-01 -5.48472166e-01
8.56631100e-01 -2.13449985e-01 6.12808645e-01 7.98883915e-01
-4.83092964e-02 5.70405602e-01 8.58275354e-01 -1.87688783e-01
1.14797704e-01 -3.20929766e-01 -3.54443103e-01 7.33968675e-01
2.06554100e-01 5.66984892e-01 -7.68236518e-01 6.05128072e-02
1.06286287e+00 3.07405405e-02 -5.20836972e-02 -9.45636213e-01
-1.05230045e+00 7.56039023e-01 8.52170587e-01 -2.75567085e-01
-2.80188024e-01 3.62374574e-01 1.17749378e-01 -4.62878831e-02
6.67156219e-01 1.15449429e-01 -2.33819529e-01 3.00853014e-01
-1.11132717e+00 3.77643347e-01 2.99192667e-01 1.08676755e+00
1.33944976e+00 9.12370160e-02 1.87162280e-01 5.47624946e-01
4.01081204e-01 7.61605024e-01 -1.38784438e-01 -1.65098035e+00
7.91432858e-01 7.08687842e-01 -5.84301502e-02 -8.03456903e-01
-2.74016708e-01 -8.60770196e-02 -9.27232325e-01 4.15831298e-01
2.68216908e-01 -1.65303454e-01 -1.07739484e+00 1.48521292e+00
3.94041330e-01 4.00129288e-01 2.50350276e-04 8.43219995e-01
9.39788818e-01 1.70782745e-01 -3.78219485e-01 3.48009855e-01
6.17211342e-01 -5.64914346e-01 -2.35273495e-01 -5.68929136e-01
4.57505643e-01 -6.91173077e-01 7.83559680e-01 2.91731209e-01
-1.12025225e+00 -6.00514174e-01 -9.90859091e-01 -3.59114617e-01
-2.37697408e-01 -1.70405731e-01 7.91499078e-01 3.08719158e-01
-1.26480711e+00 7.79629588e-01 -8.13438475e-01 -2.67970562e-01
5.85643053e-01 5.08695304e-01 -5.35271466e-01 -4.69193280e-01
-7.50250161e-01 5.70755601e-01 3.93291444e-01 1.73240274e-01
-7.81208575e-01 -9.48852956e-01 -1.24195838e+00 -3.22468668e-01
4.88839775e-01 -1.16713202e+00 1.38609421e+00 -3.80259961e-01
-1.18332815e+00 7.87818611e-01 -4.50458556e-01 -4.04174358e-01
7.70560384e-01 -3.99388134e-01 1.94801152e-01 7.16730431e-02
3.33818108e-01 1.51079476e+00 2.22547367e-01 -1.55577075e+00
-7.83501565e-01 -5.89874148e-01 2.93407906e-02 2.20117554e-01
5.65766245e-02 -7.06172645e-01 -6.25320435e-01 -3.01670492e-01
5.89718640e-01 -7.00573325e-01 -5.20253658e-01 7.00702012e-01
-3.97740752e-01 1.29255373e-02 9.76665080e-01 -1.26627550e-01
5.13977408e-01 -1.98521566e+00 -1.19318485e-01 2.85868019e-01
2.77863294e-01 7.50216916e-02 -1.44781932e-01 4.41799223e-01
6.93875737e-03 -3.14823464e-02 -4.13067728e-01 -1.03816593e+00
-5.00641651e-02 4.39755678e-01 -3.07041228e-01 5.68987131e-01
1.37298673e-01 8.64775598e-01 -9.01568413e-01 -6.15056872e-01
8.24416041e-01 7.79081464e-01 -7.13503897e-01 1.79007098e-01
-2.02441350e-01 6.19580448e-01 -8.98998380e-02 8.84407878e-01
9.61042821e-01 2.31598780e-01 -2.53218025e-01 -3.46787721e-01
-3.98848146e-01 2.90833026e-01 -1.31448972e+00 2.27137184e+00
-5.53206205e-01 6.85418189e-01 1.93599701e-01 -4.24092531e-01
9.81392026e-01 6.61909729e-02 5.86997569e-01 -8.28932464e-01
1.40419956e-02 3.13963413e-01 -3.47889572e-01 1.38378562e-02
6.13259375e-01 -2.29592547e-01 1.21957529e-02 -9.06203911e-02
-4.27614339e-02 -7.45633125e-01 -3.93108279e-02 1.51624516e-01
6.64057016e-01 3.70409817e-01 -1.84291825e-01 -6.12884015e-02
4.73622024e-01 -5.15583605e-02 7.38855302e-01 6.16296768e-01
1.75097317e-01 1.09716523e+00 -5.67190722e-03 -3.57312351e-01
-1.15033937e+00 -1.38296926e+00 -2.95459241e-01 3.95499825e-01
3.37598979e-01 -2.72866070e-01 -5.60201883e-01 -2.88538188e-01
5.25379658e-01 4.58613396e-01 -3.28572243e-01 1.54735386e-01
-5.14137387e-01 1.72109500e-01 6.49438143e-01 9.39163327e-01
8.12908232e-01 -5.67979515e-01 -5.27210176e-01 1.71843171e-01
-1.05152659e-01 -1.52348351e+00 -4.02054667e-01 2.15208143e-01
-1.19506836e+00 -1.06382573e+00 -3.19901139e-01 -4.84207004e-01
5.86033225e-01 4.90414679e-01 1.00801969e+00 -1.10380322e-01
-9.48514044e-02 2.92218119e-01 2.14078501e-01 -3.39785844e-01
1.07057609e-01 9.13522616e-02 -2.38361349e-03 -4.22255933e-01
2.00788677e-01 -1.07914472e+00 -7.94156671e-01 2.78320730e-01
-8.74661028e-01 2.08656266e-01 3.01757812e-01 4.50757563e-01
7.34590352e-01 -2.29357243e-01 -6.25873581e-02 -5.76418400e-01
2.97556520e-01 -1.34552587e-02 -1.03140354e+00 -3.27891976e-01
-6.25418365e-01 7.56147057e-02 3.71489584e-01 -2.49577090e-02
-8.31896961e-01 5.66905141e-01 -2.33739078e-01 -9.81199026e-01
-2.92520195e-01 3.65515113e-01 -2.90657222e-01 -3.61838251e-01
4.71279204e-01 -7.34034851e-02 -6.38541207e-02 -5.33693492e-01
5.10816097e-01 4.79642510e-01 1.15025866e+00 -6.78359389e-01
1.18102276e+00 9.83431697e-01 3.64346504e-01 -5.61485291e-01
-8.56333673e-01 -6.51822448e-01 -1.26107764e+00 -2.51209229e-01
6.33129895e-01 -1.33498442e+00 -7.06890106e-01 6.21299922e-01
-1.51390517e+00 -2.46929243e-01 -3.21666926e-01 3.80195230e-01
-6.60733104e-01 5.08671701e-01 -3.16536874e-01 -7.36476004e-01
-1.82738662e-01 -1.12971342e+00 1.66932511e+00 1.46645457e-01
-6.13022633e-02 -9.27934825e-01 1.49299845e-01 3.75096679e-01
8.47324263e-03 4.84358162e-01 3.56683731e-01 1.04769997e-01
-1.01185644e+00 -6.85004964e-02 -5.00303447e-01 3.57100606e-01
1.14817232e-01 8.79202560e-02 -1.19248879e+00 -2.91534126e-01
-4.36925679e-01 -4.56271470e-01 1.00209248e+00 3.11662197e-01
1.00820947e+00 2.67094433e-01 -3.39702666e-01 1.17952394e+00
1.61900890e+00 -8.40903074e-02 6.50863230e-01 2.38512844e-01
1.06467938e+00 6.72829092e-01 6.09339356e-01 3.01880866e-01
8.53085816e-01 4.90964144e-01 9.84396458e-01 -4.15882051e-01
-1.53188109e-01 -8.24121833e-01 5.77028329e-03 4.46182698e-01
5.89797124e-02 -3.53911296e-02 -1.13099337e+00 5.61802149e-01
-1.89145565e+00 -5.66793084e-01 -2.96328992e-01 2.34969687e+00
4.39506918e-01 3.75840843e-01 -1.92907214e-01 8.25976804e-02
5.38323462e-01 2.66046375e-01 -7.18733847e-01 -1.34154409e-01
-1.72954798e-01 3.41469646e-01 8.97763968e-01 8.28261435e-01
-1.02421677e+00 1.21614528e+00 6.43571615e+00 4.61370409e-01
-1.16850936e+00 -1.19984046e-01 1.12839580e-01 3.80946100e-02
-5.61559856e-01 2.49013051e-01 -8.18464816e-01 8.93744901e-02
5.37194014e-01 1.82792053e-01 2.57461756e-01 6.67607367e-01
3.39747757e-01 -3.70551318e-01 -1.31127119e+00 1.08351958e+00
-2.12459296e-01 -1.46320558e+00 -9.67859998e-02 3.17154229e-01
8.98680866e-01 4.80977714e-01 -1.87328681e-01 2.53639407e-02
6.40824914e-01 -9.13294315e-01 8.13479543e-01 5.83269835e-01
1.02195120e+00 -8.63395154e-01 5.50937533e-01 6.08001053e-01
-1.59775305e+00 1.60560086e-01 -3.26405078e-01 -1.90927818e-01
4.63979512e-01 7.33494580e-01 -8.02836597e-01 7.10610151e-01
6.49446547e-01 8.61657798e-01 -5.15695274e-01 1.19192350e+00
-2.76238829e-01 -4.97881137e-02 -7.02772498e-01 7.06573904e-01
3.73046249e-01 -1.47282124e-01 4.39678401e-01 1.00789344e+00
4.37532872e-01 -1.16521247e-01 3.31557751e-01 1.10726631e+00
-1.83517069e-01 -3.18274379e-01 -1.05631614e+00 3.50544095e-01
7.60128438e-01 1.00996768e+00 -4.58401889e-01 -1.96927100e-01
-2.11452261e-01 7.08987951e-01 1.81287259e-01 1.58990026e-01
-5.52394569e-01 -1.71446085e-01 8.20916653e-01 4.67604816e-01
2.67105877e-01 -5.14671266e-01 -6.35917306e-01 -9.27460194e-01
2.41193950e-01 -1.94477066e-01 -9.86966565e-02 -1.06155181e+00
-1.29398680e+00 2.03589842e-01 1.99399933e-01 -1.24997699e+00
-3.38097394e-01 -4.82527554e-01 -4.78145182e-01 1.03497422e+00
-1.85224724e+00 -1.29950881e+00 -6.56243086e-01 7.79785216e-01
2.59691387e-01 2.24097446e-01 3.11300099e-01 4.59103256e-01
-2.06731066e-01 3.30247343e-01 -8.77918378e-02 4.58851233e-02
6.99111581e-01 -1.20867193e+00 7.62525916e-01 8.01482141e-01
-1.99681655e-01 3.77412438e-01 5.16810834e-01 -8.16177607e-01
-1.14589512e+00 -1.31768811e+00 8.89930189e-01 -2.78229952e-01
4.60722297e-01 -5.11906266e-01 -7.34012902e-01 6.33603692e-01
6.73261061e-02 2.53108770e-01 1.99267454e-02 -9.06744823e-02
-5.23000836e-01 -3.90176773e-01 -1.18239510e+00 4.67714161e-01
1.47135139e+00 -8.33390951e-01 -4.30989653e-01 7.06259236e-02
8.74845266e-01 -8.04624915e-01 -7.37306893e-01 9.35218453e-01
5.62212944e-01 -1.39694238e+00 1.24463952e+00 7.48977736e-02
4.42679226e-01 -3.47833395e-01 -4.35873032e-01 -1.16905081e+00
8.33405256e-02 -3.92330766e-01 1.18045636e-01 1.25012302e+00
-1.04141096e-02 -5.35134673e-01 1.26982260e+00 4.47314948e-01
-5.11877120e-01 -5.84548354e-01 -1.29032028e+00 -6.41102374e-01
1.48764402e-01 -8.26504230e-01 6.52816057e-01 7.04527795e-01
-6.19251192e-01 2.10151479e-01 -1.90690950e-01 4.11016315e-01
9.95056570e-01 -9.31518599e-02 1.12409568e+00 -1.67578912e+00
1.07839674e-01 -3.70298326e-01 -4.67427790e-01 -1.52807009e+00
2.25164130e-01 -7.79838443e-01 3.47601414e-01 -1.74034643e+00
-1.58025861e-01 -4.01670098e-01 1.83077276e-01 4.39747512e-01
1.82051450e-01 4.21140224e-01 9.92640853e-02 1.18920378e-01
-1.48711741e-01 8.44553590e-01 1.28225291e+00 -1.47991568e-01
-8.74758214e-02 -1.29924670e-01 -4.86403674e-01 9.51334298e-01
4.64474916e-01 -4.42266881e-01 -5.21632493e-01 -8.00608754e-01
8.35182816e-02 2.29531959e-01 6.73280060e-01 -1.33435452e+00
4.78089571e-01 -1.56728566e-01 3.16153049e-01 -1.34524047e+00
8.64144802e-01 -1.10584950e+00 5.67511544e-02 9.26896036e-02
9.84486639e-02 2.99604297e-01 3.73422563e-01 4.72538501e-01
-2.41624549e-01 1.58562437e-01 7.57245183e-01 -4.47205789e-02
-5.88080049e-01 6.02710545e-01 9.14429054e-02 -1.52745545e-01
7.04978883e-01 -7.54373968e-01 -1.21271983e-01 -3.60810101e-01
-3.10110182e-01 5.45105577e-01 8.07642698e-01 2.54456997e-01
9.48023856e-01 -1.52918577e+00 -4.75975633e-01 3.24666262e-01
1.97494507e-01 1.09981298e+00 9.30808932e-02 5.13633072e-01
-5.47483504e-01 4.96681064e-01 6.09127106e-03 -1.17666078e+00
-1.05943096e+00 1.26521498e-01 6.04073703e-01 -1.16833560e-02
-6.28188372e-01 6.09836996e-01 1.93062529e-01 -1.14832091e+00
4.81149822e-01 -5.75348020e-01 4.52821016e-01 -1.57310903e-01
-9.22085270e-02 2.18872324e-01 1.04888685e-01 -7.70412385e-01
-2.75737107e-01 9.01622772e-01 2.20748067e-01 -3.19212228e-01
1.29557514e+00 -1.85782254e-01 8.55427980e-02 3.86220217e-01
1.07057357e+00 -1.55679226e-01 -1.71520793e+00 -4.58933681e-01
-2.24679917e-01 -7.15527356e-01 2.67079413e-01 -5.05486965e-01
-1.27925682e+00 1.26632869e+00 6.29858494e-01 -5.05601287e-01
7.17449188e-01 -1.73353866e-01 9.30961728e-01 5.54329157e-01
5.17765880e-01 -8.64958167e-01 6.22834675e-02 6.86643124e-01
6.28347278e-01 -1.28869832e+00 -2.55455077e-03 -5.61601698e-01
-8.25814754e-02 1.02612495e+00 8.91635835e-01 -2.52740771e-01
5.99601626e-01 4.35675353e-01 1.22429267e-01 -2.05739364e-01
-3.48623931e-01 -3.53736818e-01 2.90814131e-01 7.33714461e-01
1.76642239e-01 -3.34318578e-01 3.47112089e-01 -2.14584008e-01
-4.01544183e-01 2.64609337e-01 1.63007632e-01 1.05790484e+00
-4.47682440e-01 -1.24229717e+00 -5.06465375e-01 4.19289023e-02
3.01727414e-01 -1.13376595e-01 -4.84272927e-01 1.02863908e+00
5.25953948e-01 6.89541698e-01 4.66405511e-01 -6.20563149e-01
7.67525077e-01 -2.26714715e-01 5.88002503e-01 -6.63022578e-01
-3.31946135e-01 2.51419872e-01 -6.14294559e-02 -8.47647190e-01
-5.11874616e-01 -7.21331537e-01 -1.31947005e+00 -5.19638896e-01
-1.49740323e-01 -3.45157146e-01 7.23465562e-01 9.64935899e-01
3.68725181e-01 2.58770138e-01 5.17192006e-01 -1.54144776e+00
-2.63437718e-01 -7.50408471e-01 -3.00155878e-01 1.49837628e-01
4.82416987e-01 -7.90279686e-01 -2.82086283e-01 -3.08278620e-01] | [8.22388744354248, -2.8222124576568604] |
3b646418-fa2c-42e8-8a67-16cdff051c0a | linking-graph-entities-with-multiplicity-and | 1908.04464 | null | https://arxiv.org/abs/1908.04464v2 | https://arxiv.org/pdf/1908.04464v2.pdf | Linking Graph Entities with Multiplicity and Provenance | Entity linking and resolution is a fundamental database problem with applications in data integration, data cleansing, information retrieval, knowledge fusion, and knowledge-base population. It is the task of accurately identifying multiple, differing, and possibly contradicting representations of the same real-world entity in data. In this work, we propose an entity linking and resolution system capable of linking entities across different databases and mentioned-entities extracted from text data. Our entity linking/resolution solution, called Certus, uses a graph model to represent the profiles of entities. The graph model is versatile, thus, it is capable of handling multiple values for an attribute or a relationship, as well as the provenance descriptions of the values. Provenance descriptions of a value provide the settings of the value, such as validity periods, sources, security requirements, etc. This paper presents the architecture for the entity linking system, the logical, physical, and indexing models used in the system, and the general linking process. Furthermore, we demonstrate the performance of update operations of the physical storage models when the system is implemented in two state-of-the-art database management systems, HBase and Postgres. | ['Michael Bewong', 'Selasi Kwashie', 'Lin Liu', 'Jiuyong Li', 'Jixue Liu'] | 2019-08-13 | null | null | null | null | ['knowledge-base-population'] | ['natural-language-processing'] | [-5.07630289e-01 1.06923141e-01 -2.56893128e-01 -1.97883561e-01
-2.54441410e-01 -5.12304068e-01 6.51295781e-01 1.25539911e+00
-2.53414840e-01 9.74759042e-01 6.89457655e-02 1.29108071e-01
-5.27442515e-01 -1.25849390e+00 -3.23436052e-01 -2.23669276e-01
-2.86231041e-01 5.86118519e-01 9.36332405e-01 -1.61722094e-01
2.71581501e-01 6.42049193e-01 -1.82598853e+00 1.58484876e-01
5.33067286e-01 1.10023916e+00 -3.16947728e-01 1.62328646e-01
-7.33615160e-01 8.91475976e-01 -8.75734866e-01 -4.54377919e-01
-8.08707029e-02 7.65052140e-02 -7.69562066e-01 -7.39223421e-01
2.08629295e-01 -1.16810750e-03 5.37337698e-02 1.08874512e+00
4.35013533e-01 -3.51195484e-01 1.93952262e-01 -1.86563277e+00
-6.74849868e-01 6.74985051e-01 -2.55737782e-01 -8.68623611e-03
3.20702523e-01 -5.12564123e-01 4.03323144e-01 -5.91260433e-01
1.27967238e+00 7.71906972e-01 5.11498868e-01 6.59491420e-02
-5.63946724e-01 -5.08776724e-01 -3.75115007e-01 5.88849068e-01
-1.67931581e+00 -7.33148158e-01 1.52121246e-01 -4.14457887e-01
9.04463768e-01 5.91053128e-01 4.06790823e-01 -6.86012506e-02
3.82710069e-01 -2.23984972e-01 6.89494312e-01 -5.27528584e-01
3.20137054e-01 4.86580729e-01 6.77395165e-01 3.75534683e-01
1.14722383e+00 -3.82085532e-01 -9.09300327e-01 -7.68939078e-01
1.42868653e-01 -2.18262613e-01 -1.78927481e-01 -6.57633543e-01
-8.99497569e-01 -9.56518203e-03 5.34936264e-02 1.90507904e-01
-5.22885442e-01 1.18147910e-01 5.78559279e-01 7.88414478e-02
-1.51121169e-01 3.53371166e-02 -4.79553342e-01 1.21388339e-01
-4.72742707e-01 3.03711146e-01 1.15098643e+00 1.62276840e+00
7.21702933e-01 -3.67820591e-01 9.50681642e-02 3.75230685e-02
7.93454587e-01 2.91289985e-01 2.66784012e-01 -7.67776906e-01
3.06085289e-01 1.02010143e+00 7.25786626e-01 -1.00048566e+00
-3.96784931e-01 1.21362887e-01 -4.16763425e-01 2.58196115e-01
-2.53552616e-01 3.60008597e-01 -5.72818637e-01 1.50927246e+00
9.01944935e-01 -5.40117547e-02 5.40196240e-01 2.76696324e-01
1.10937154e+00 1.58638000e-01 2.79345244e-01 -7.86832035e-01
1.76315415e+00 -8.00924748e-02 -1.52817547e+00 4.10008758e-01
3.54346424e-01 -8.23120058e-01 -1.79022595e-01 3.35453637e-02
-1.23514974e+00 -2.36537248e-01 -1.12804604e+00 -1.24243155e-01
-1.07036102e+00 -4.78560895e-01 3.33414733e-01 3.15377086e-01
-8.18682969e-01 2.39265740e-01 -6.56462610e-01 -5.89531064e-01
-7.03822821e-02 9.51176286e-02 -5.97722471e-01 2.42243379e-01
-1.44744062e+00 1.11459291e+00 1.15535438e+00 -7.67630711e-02
-1.90684840e-01 -7.60372281e-01 -5.81300378e-01 2.57424742e-01
4.51070875e-01 -7.47269392e-01 8.04553092e-01 8.22917372e-02
-3.92846674e-01 6.87654555e-01 6.24589808e-02 -4.93947327e-01
6.25115931e-02 -8.48932490e-02 -1.46406353e+00 -3.11010480e-01
2.19421029e-01 -3.21730018e-01 1.49440557e-01 -1.38720620e+00
-9.17650223e-01 -6.54291689e-01 -2.83590108e-01 -2.19083712e-01
8.69920105e-02 3.35771471e-01 -6.01999998e-01 -1.51059821e-01
1.45436868e-01 -2.22767904e-01 2.71383971e-01 1.97191723e-02
-2.05394819e-01 -7.00847730e-02 6.80022180e-01 -7.97817528e-01
1.97380972e+00 -2.25704336e+00 -1.13247015e-01 5.52638650e-01
7.09201917e-02 4.32173938e-01 6.44137979e-01 1.10710990e+00
1.87653735e-01 5.27147055e-01 -3.22066285e-02 3.20274651e-01
1.40681192e-01 4.10557598e-01 -3.48668784e-01 1.12082139e-01
-1.61953151e-01 3.01221251e-01 -4.69937295e-01 -9.42161858e-01
-1.69312716e-01 2.56042510e-01 1.63391948e-01 1.79428354e-01
-4.10329968e-01 -2.56217092e-01 -3.82977158e-01 8.16675901e-01
9.23324049e-01 -1.17185051e-02 9.14049685e-01 -5.16449571e-01
-5.33114672e-01 1.09504469e-01 -2.15362668e+00 1.39304149e+00
2.50780225e-01 -1.69265375e-01 2.98965365e-01 -1.01786898e-02
9.87724423e-01 6.92342579e-01 4.29293901e-01 -7.66046643e-01
-3.27821136e-01 4.48564112e-01 -5.97951174e-01 -7.73058295e-01
1.05696070e+00 5.70113540e-01 -1.26503244e-01 4.89678502e-01
-1.96479574e-01 3.63498628e-01 5.41387796e-01 3.56951118e-01
1.02286494e+00 3.21034104e-01 8.43512595e-01 9.33512896e-02
7.62534559e-01 1.38556495e-01 9.68646765e-01 2.19425485e-01
2.82133698e-01 -2.40589708e-01 5.57713985e-01 -3.67291778e-01
-9.64310825e-01 -1.12134707e+00 -2.67202139e-01 3.61283839e-01
6.46163821e-01 -9.93749678e-01 -2.63540447e-01 -3.06188881e-01
4.81796861e-01 8.15115809e-01 -2.64896750e-01 -8.59486610e-02
-6.47889614e-01 -3.71339262e-01 5.59096813e-01 4.39915568e-01
3.89675528e-01 -1.05745769e+00 -8.30551326e-01 3.24215502e-01
-7.05063045e-02 -7.09359467e-01 1.94689289e-01 -2.40948394e-01
-4.58785892e-01 -1.54092121e+00 6.83616459e-01 -3.82261366e-01
3.06160659e-01 -2.60719419e-01 1.31406105e+00 6.13596141e-01
-1.61069348e-01 2.49838129e-01 -3.16238135e-01 -7.02196240e-01
-5.22576332e-01 -1.64154634e-01 -1.28307343e-01 -2.77223229e-01
2.90806085e-01 -3.34433049e-01 -1.21475510e-01 3.67483526e-01
-1.34284282e+00 -1.75984591e-01 -5.56132533e-02 2.19427615e-01
8.99713635e-01 3.00956994e-01 6.99688971e-01 -1.07099545e+00
8.22168469e-01 -1.00540650e+00 -8.84778798e-01 1.20509732e+00
-1.32204926e+00 2.91276395e-01 2.34544978e-01 1.98725566e-01
-1.19102609e+00 -4.68108088e-01 4.18066472e-01 1.20704912e-01
6.34029508e-02 8.93158078e-01 -5.81880271e-01 3.25305283e-01
3.21571410e-01 -2.41372868e-01 -1.21583968e-01 -8.90903294e-01
5.79822063e-01 6.74797177e-01 8.85216892e-01 -7.01966524e-01
5.29829800e-01 4.60969180e-01 1.90872684e-01 -7.94658363e-02
-6.13586903e-02 -4.52586561e-01 -6.01938307e-01 7.29409903e-02
4.47245657e-01 -7.07118273e-01 -8.02004397e-01 3.66280556e-01
-1.47917581e+00 6.39203668e-01 -5.66554189e-01 5.16354963e-02
-1.63769990e-01 1.07585207e-01 -1.34605780e-01 -7.11135924e-01
-6.63588703e-01 -5.22147655e-01 5.23591638e-01 3.07145357e-01
-8.18768591e-02 -4.45003629e-01 4.07317162e-01 -5.31520583e-02
5.54677427e-01 7.16915309e-01 1.20232224e+00 -7.60816634e-01
-9.11384463e-01 -3.05163920e-01 -2.43357554e-01 -3.51094961e-01
1.00943305e-01 6.01627707e-01 -2.30096132e-01 -7.72809908e-02
-4.54087853e-01 2.87676990e-01 -2.37056628e-01 -5.48345566e-01
5.63093841e-01 -4.11290228e-01 -6.40748620e-01 4.24377739e-01
1.73251212e+00 5.79232872e-01 9.04017925e-01 6.23950958e-01
3.94389927e-01 5.98892927e-01 9.41387653e-01 6.77667260e-01
6.10688508e-01 9.48796749e-01 6.25631273e-01 4.30024505e-01
-1.96108028e-01 -1.34829074e-01 -3.26922536e-01 6.20752454e-01
1.70876473e-01 -3.49275440e-01 -8.39846671e-01 4.42141682e-01
-2.32431793e+00 -1.27574766e+00 -6.02954984e-01 2.56405950e+00
1.01326823e+00 -1.88393250e-01 -2.90819079e-01 4.40320112e-02
1.00540316e+00 -1.34220824e-01 -5.67259192e-01 -3.98567885e-01
2.02210043e-02 -1.41310245e-02 6.82932615e-01 3.27603221e-01
-3.89645875e-01 5.36130428e-01 5.79961252e+00 2.94397660e-02
-7.33757854e-01 2.34481126e-01 -6.26588106e-01 1.52192041e-01
-3.58548701e-01 4.22051162e-01 -1.19246852e+00 5.91161191e-01
1.28776932e+00 -9.34815466e-01 2.84283847e-01 5.08746505e-01
1.05221116e-03 -2.33719423e-01 -8.63092184e-01 4.79714602e-01
1.90281346e-02 -1.72604752e+00 5.07597513e-02 1.43452175e-02
1.63181260e-01 -1.04170866e-01 -7.09688663e-01 -2.37247035e-01
2.82863140e-01 -4.03572887e-01 1.15364373e+00 1.24599266e+00
6.46188796e-01 -6.69253528e-01 9.22360301e-01 -7.05506802e-02
-1.13848996e+00 1.61323324e-01 -2.10664216e-02 5.33945739e-01
3.79301846e-01 4.51097488e-01 -6.06037557e-01 1.35189486e+00
1.04523814e+00 2.44979877e-02 -6.62529767e-01 1.18261957e+00
5.89796640e-02 -3.13054264e-01 -4.70327437e-01 3.37996632e-01
-9.64936852e-01 4.69607338e-02 1.52840316e-01 8.44890058e-01
5.53256929e-01 1.87544763e-01 -2.47145802e-01 3.93331140e-01
-3.45946610e-01 1.98798716e-01 -2.06430271e-01 2.25782972e-02
1.49882782e+00 9.79003906e-01 -1.12436069e-02 -6.96198225e-01
-2.37952501e-01 -1.33637890e-01 2.05501348e-01 2.07566008e-01
-6.77087903e-01 -7.30792701e-01 7.67409325e-01 4.19280320e-01
2.38564506e-01 -1.86585411e-01 -1.34594902e-01 -9.23819542e-01
5.93392491e-01 -7.77069688e-01 7.31483757e-01 -1.10375988e+00
-8.54583442e-01 5.48748374e-01 3.73795599e-01 -9.39380527e-01
-3.70292693e-01 1.05829246e-01 -1.82714030e-01 1.13244379e+00
-1.40601313e+00 -1.04730797e+00 -6.94240332e-01 7.28918910e-01
-7.22312629e-01 1.82185024e-02 1.06026018e+00 6.90283060e-01
-5.67494869e-01 1.40023187e-01 2.74134457e-01 -5.50090782e-02
9.89548206e-01 -9.20516133e-01 2.46831208e-01 1.10527098e+00
-2.60463774e-01 1.09929800e+00 6.75126791e-01 -1.11402059e+00
-1.50936282e+00 -8.77287328e-01 1.09693670e+00 -1.15325376e-01
6.65599167e-01 -3.18347327e-02 -1.35855889e+00 8.52964103e-01
2.91281670e-01 4.00405109e-01 6.36931062e-01 -1.94730669e-01
-6.76702857e-01 -7.31525540e-01 -1.75101447e+00 5.83362095e-02
7.20670402e-01 -3.04567575e-01 -6.34533048e-01 1.14011839e-01
7.03865469e-01 -4.61209267e-01 -1.78617299e+00 3.20344269e-01
6.52314365e-01 -7.73372293e-01 8.66826653e-01 -7.15237141e-01
-4.01694119e-01 -1.20375764e+00 -2.38741159e-01 -6.14699364e-01
-1.15368560e-01 -4.03847754e-01 -8.27557087e-01 2.20741105e+00
2.17863411e-01 -8.77505839e-01 1.28113776e-01 1.14142442e+00
8.51873159e-02 -1.19747825e-01 -1.07376432e+00 -5.90978086e-01
-7.95924187e-01 1.77958369e-01 1.62321043e+00 1.19546032e+00
8.59959349e-02 -6.35215119e-02 -3.52102332e-02 9.31611300e-01
6.68214321e-01 2.03858852e-01 8.44205558e-01 -1.78250790e+00
1.39144287e-01 1.13613799e-01 -6.02993309e-01 3.94023865e-01
-4.79210973e-01 -5.53726137e-01 -5.31395376e-01 -1.99590981e+00
-3.84096265e-01 -9.62938130e-01 -3.12267512e-01 5.57077110e-01
3.27495039e-01 -6.02120340e-01 1.70230433e-01 9.33718443e-01
-4.39360350e-01 -1.71570837e-01 1.95614800e-01 6.30036518e-02
-1.21544875e-01 -3.78068209e-01 -4.61360514e-01 2.02519611e-01
5.79746783e-01 -8.89272630e-01 -2.26397723e-01 -4.79113251e-01
6.88302219e-01 2.30219662e-01 2.06349507e-01 -9.62752104e-01
8.63822281e-01 -6.35819435e-01 -6.11589802e-03 -9.21956062e-01
1.46663845e-01 -1.25760710e+00 1.66272426e+00 4.04386610e-01
-3.70420851e-02 6.11349285e-01 5.23259416e-02 3.88410270e-01
-3.72728765e-01 -5.55088401e-01 4.09818321e-01 -1.10341042e-01
-1.01638043e+00 3.32381427e-01 4.78590876e-01 5.35577312e-02
1.41067529e+00 1.18059635e-01 -1.47351635e+00 5.55222392e-01
-5.20925522e-01 4.60055858e-01 9.78562534e-01 4.66786683e-01
4.05070275e-01 -1.38889420e+00 -4.76176083e-01 1.96972564e-01
4.91289765e-01 1.50829524e-01 -1.54885858e-01 3.12302142e-01
-7.88781464e-01 9.95416790e-02 -4.60546553e-01 -3.82677317e-02
-1.49132204e+00 1.02777839e+00 2.34492153e-01 -1.14474371e-01
-1.08071566e-01 -2.99676001e-01 -9.40481901e-01 -1.80452894e-02
3.19163203e-02 1.60057470e-01 -3.66316944e-01 4.27992165e-01
7.90348113e-01 7.72141218e-01 5.77225864e-01 -6.39830232e-01
-8.35544288e-01 4.62566674e-01 9.28634778e-02 -1.03801481e-01
1.17746377e+00 -3.28676343e-01 -8.96704853e-01 4.63803113e-01
4.25471395e-01 4.14165884e-01 -2.54983753e-01 -3.35255533e-01
5.57380855e-01 -4.46535587e-01 -3.16170216e-01 -1.06996417e+00
-7.32933521e-01 1.08564109e-01 5.47086656e-01 7.19294786e-01
8.13527703e-01 -5.57870902e-02 4.39122677e-01 1.74420238e-01
7.53452122e-01 -9.73048329e-01 -1.04456532e+00 -3.10032349e-02
7.98696280e-01 -6.40556276e-01 5.71508288e-01 -4.22609210e-01
-1.63323358e-01 9.56527293e-01 6.25335038e-01 7.10729659e-01
8.06710422e-01 4.87506330e-01 -4.26243655e-02 -6.04938924e-01
-9.07115459e-01 -1.18304446e-01 -2.98289210e-01 7.95111239e-01
5.48132509e-02 1.16733490e-02 -9.72980320e-01 3.58575255e-01
3.05750549e-01 6.87219024e-01 7.19260752e-01 1.60916567e+00
-4.79492903e-01 -1.39937305e+00 -5.75134933e-01 1.24652714e-01
-4.98956263e-01 1.07328571e-01 -1.96347982e-01 9.60030794e-01
6.98191166e-01 6.52040660e-01 4.19928759e-01 -4.87242006e-02
9.85194564e-01 2.70220965e-01 -1.32798031e-03 -4.00078416e-01
-7.91118622e-01 -6.16959035e-01 5.48339427e-01 -3.29774678e-01
-3.26285571e-01 -6.22540355e-01 -1.98922622e+00 -5.12987912e-01
-3.75621587e-01 6.60288692e-01 1.21758258e+00 4.56247598e-01
1.07241666e+00 3.15638900e-01 1.91113248e-01 3.99443358e-01
-1.75644875e-01 -4.56868887e-01 -6.90944016e-01 5.10705411e-01
-3.87523025e-02 -6.18827939e-01 4.50336635e-02 2.75711775e-01] | [9.155673027038574, 7.846449851989746] |
054af165-ccd7-4b64-9e3e-c198d6fc3f86 | improving-description-based-person-re | 1906.09610 | null | https://arxiv.org/abs/1906.09610v1 | https://arxiv.org/pdf/1906.09610v1.pdf | Improving Description-based Person Re-identification by Multi-granularity Image-text Alignments | Description-based person re-identification (Re-id) is an important task in video surveillance that requires discriminative cross-modal representations to distinguish different people. It is difficult to directly measure the similarity between images and descriptions due to the modality heterogeneity (the cross-modal problem). And all samples belonging to a single category (the fine-grained problem) makes this task even harder than the conventional image-description matching task. In this paper, we propose a Multi-granularity Image-text Alignments (MIA) model to alleviate the cross-modal fine-grained problem for better similarity evaluation in description-based person Re-id. Specifically, three different granularities, i.e., global-global, global-local and local-local alignments are carried out hierarchically. Firstly, the global-global alignment in the Global Contrast (GC) module is for matching the global contexts of images and descriptions. Secondly, the global-local alignment employs the potential relations between local components and global contexts to highlight the distinguishable components while eliminating the uninvolved ones adaptively in the Relation-guided Global-local Alignment (RGA) module. Thirdly, as for the local-local alignment, we match visual human parts with noun phrases in the Bi-directional Fine-grained Matching (BFM) module. The whole network combining multiple granularities can be end-to-end trained without complex pre-processing. To address the difficulties in training the combination of multiple granularities, an effective step training strategy is proposed to train these granularities step-by-step. Extensive experiments and analysis have shown that our method obtains the state-of-the-art performance on the CUHK-PEDES dataset and outperforms the previous methods by a significant margin. | ['Kai Niu', 'Liang Wang', 'Yan Huang', 'Wanli Ouyang'] | 2019-06-23 | null | null | null | null | ['nlp-based-person-retrival'] | ['computer-vision'] | [ 1.17546305e-01 -4.03501272e-01 -6.12399876e-02 -4.14370865e-01
-8.73715281e-01 -3.39392185e-01 9.08526361e-01 1.00434855e-01
-5.74997187e-01 4.26200122e-01 4.07750189e-01 1.94005355e-01
-2.00147986e-01 -6.16976380e-01 -4.05174762e-01 -8.88051569e-01
1.58264980e-01 5.51587045e-01 4.12375689e-01 -1.63505152e-01
8.54272693e-02 4.48119074e-01 -1.82254076e+00 6.59547925e-01
6.87700272e-01 7.79746294e-01 1.26250163e-01 4.89614308e-01
-7.12042451e-02 3.19334269e-01 -4.06185955e-01 -6.72395885e-01
3.29552531e-01 -5.12119830e-01 -6.44226074e-01 5.23826241e-01
1.02501535e+00 -3.17917436e-01 -2.56462067e-01 1.42499471e+00
6.62357450e-01 2.10950151e-01 6.22825503e-01 -1.25611055e+00
-5.10113776e-01 6.41416609e-02 -8.40417385e-01 3.10446024e-01
5.02296031e-01 2.82149702e-01 8.10156584e-01 -8.31445456e-01
4.90603924e-01 1.62394619e+00 7.76648819e-01 7.43851066e-01
-1.16208625e+00 -7.34326661e-01 3.40906054e-01 4.95457262e-01
-1.61291003e+00 -3.17662060e-01 6.32915974e-01 -6.26465857e-01
7.06365168e-01 3.46141845e-01 5.19037187e-01 1.15726030e+00
-1.39498673e-02 4.60799366e-01 1.15266669e+00 -2.85657585e-01
-2.74926364e-01 2.23159060e-01 1.37194768e-01 7.23684072e-01
3.62868875e-01 2.18888164e-01 -3.04572284e-01 -9.05557424e-02
6.42946720e-01 3.88952345e-01 -2.47456595e-01 -3.17205906e-01
-1.30418408e+00 6.62896216e-01 4.00589377e-01 4.66316730e-01
-2.70069093e-01 -2.21416414e-01 5.12024760e-01 1.25620171e-01
1.24337636e-01 1.98267847e-01 -6.12687394e-02 2.81853527e-01
-9.54741657e-01 4.02837276e-01 2.24417418e-01 8.63860667e-01
8.75939071e-01 -3.83559197e-01 -7.64575720e-01 9.22335148e-01
1.72892362e-01 5.27005970e-01 6.14622414e-01 -4.47693378e-01
7.40971804e-01 6.25206769e-01 3.70594971e-02 -1.24954677e+00
-3.88069272e-01 -4.45132703e-01 -1.33822250e+00 7.36154392e-02
5.77353835e-01 1.55069828e-01 -9.28009391e-01 1.84721112e+00
5.02859533e-01 1.48111612e-01 -1.33797094e-01 1.18241107e+00
1.02804351e+00 5.23495018e-01 5.38291156e-01 -8.96772817e-02
1.95709395e+00 -1.06074405e+00 -5.37307680e-01 -1.99524015e-01
2.68487424e-01 -6.58933163e-01 8.16209197e-01 -2.32210919e-01
-8.90574455e-01 -9.93147552e-01 -8.20972562e-01 5.55157624e-02
-5.88770032e-01 1.84784666e-01 1.42191663e-01 3.31041098e-01
-1.07026267e+00 1.32758036e-01 -3.26240301e-01 -7.14762509e-01
7.43591934e-02 2.83680499e-01 -8.67490470e-01 -1.97984487e-01
-1.27562463e+00 7.93404102e-01 6.74795687e-01 7.12469220e-02
-7.04074442e-01 -2.58812189e-01 -9.13399100e-01 1.60701394e-01
3.45055252e-01 -7.99785078e-01 5.90913951e-01 -1.00022948e+00
-9.26745534e-01 1.31709015e+00 -2.87717164e-01 -1.80136815e-01
3.89121652e-01 2.79011518e-01 -5.45647621e-01 2.64412075e-01
3.89687121e-01 7.89180100e-01 8.86772573e-01 -1.28722632e+00
-1.09132123e+00 -4.27237809e-01 5.12827635e-02 3.48902285e-01
-2.40740836e-01 2.84782052e-01 -9.40842152e-01 -9.62795615e-01
1.45369759e-02 -8.71502578e-01 4.27409410e-02 -1.13108113e-01
-2.25764871e-01 -4.14364487e-01 6.29125118e-01 -9.74234462e-01
1.21863461e+00 -2.23182535e+00 1.74175248e-01 2.01604143e-01
2.71812111e-01 3.44666332e-01 -3.22709858e-01 2.71673113e-01
-1.33636370e-01 -1.48036256e-01 -7.01370910e-02 -5.41564107e-01
4.39032130e-02 -5.88075034e-02 1.96813449e-01 4.44487423e-01
2.45268464e-01 8.51705372e-01 -8.54469478e-01 -9.03619647e-01
2.46979281e-01 3.82643849e-01 -3.82868260e-01 4.84920353e-01
3.03457469e-01 4.93424028e-01 -2.19527036e-01 7.77337909e-01
7.96661794e-01 -1.64319098e-01 7.52286389e-02 -7.06276119e-01
-1.19808167e-01 -1.73627466e-01 -1.18255889e+00 1.32240319e+00
-1.09833613e-01 3.15837860e-01 -6.04387186e-02 -1.00993860e+00
7.07147539e-01 3.16025823e-01 3.05159569e-01 -1.01982951e+00
-1.46205381e-01 1.01801559e-01 -2.88006067e-02 -5.71808994e-01
4.59937125e-01 -1.12245448e-01 -3.27421755e-01 8.05912167e-02
1.35661751e-01 6.81265652e-01 3.90882283e-01 1.72569174e-02
5.56031406e-01 -8.13730732e-02 5.44140041e-01 -2.37161919e-01
1.01435423e+00 -9.40911174e-02 7.80937493e-01 7.94866800e-01
-5.09034753e-01 8.90603662e-01 1.07495464e-01 -5.76745927e-01
-1.09876776e+00 -9.71553504e-01 4.70195338e-02 1.16078651e+00
5.67712426e-01 -3.75639409e-01 -8.40061903e-01 -7.14411080e-01
-9.69934613e-02 1.52819455e-01 -6.73873663e-01 5.15400432e-02
-6.10865057e-01 -8.17109168e-01 5.71442425e-01 4.88790810e-01
1.09312534e+00 -9.67483222e-01 -3.32947075e-01 -3.39369774e-02
-5.49744248e-01 -1.29276228e+00 -9.12876725e-01 -3.09601516e-01
-4.04732496e-01 -1.03205085e+00 -9.35849130e-01 -9.94686306e-01
7.56855607e-01 5.27366817e-01 9.38963652e-01 1.62044972e-01
-2.21397653e-01 3.53906155e-01 -3.59165639e-01 2.92377234e-01
-1.45770609e-01 -1.69126436e-01 1.20977007e-01 3.94945085e-01
5.68156660e-01 -1.22769199e-01 -6.75732493e-01 5.66396952e-01
-7.20591605e-01 2.22375453e-01 7.70835340e-01 1.10680091e+00
6.78357780e-01 1.86139524e-01 3.67291898e-01 -4.45919186e-01
4.58384335e-01 -1.47907570e-01 -3.49302351e-01 6.77026808e-01
-2.33711049e-01 -5.92849664e-02 4.93962854e-01 -6.58910215e-01
-1.06633079e+00 1.94195658e-02 -1.14532962e-01 -3.39528590e-01
-5.02391219e-01 2.09050566e-01 -5.73679745e-01 -9.05117169e-02
2.54419774e-01 4.52627271e-01 -1.70149714e-01 -3.41567338e-01
7.94046074e-02 6.92854464e-01 8.78836274e-01 -4.54088509e-01
1.02098882e+00 4.28943336e-01 -2.00419813e-01 -5.53601444e-01
-7.08208203e-01 -8.14839125e-01 -7.34169006e-01 -2.66379416e-01
1.36381328e+00 -1.08034182e+00 -5.53029358e-01 6.05520844e-01
-1.10964143e+00 8.76594335e-02 -3.23254392e-02 4.01795000e-01
-1.69825062e-01 5.76955795e-01 -4.57376689e-01 -5.03505886e-01
-3.05595517e-01 -1.22702789e+00 1.23167706e+00 5.25171578e-01
4.68576998e-02 -7.70874023e-01 -9.23720673e-02 4.95471448e-01
3.29549700e-01 1.65791199e-01 7.18688905e-01 -7.54346073e-01
-3.99811208e-01 -2.05125108e-01 -7.29224324e-01 9.03197825e-02
2.40913540e-01 -3.62378716e-01 -8.50816488e-01 -6.19687080e-01
-1.37109861e-01 -5.76787293e-02 7.90309489e-01 4.94111665e-02
9.58382249e-01 -4.50277746e-01 -4.10027206e-01 6.67444527e-01
1.24989903e+00 5.74002229e-02 6.70287490e-01 6.49970770e-01
9.52173352e-01 7.53899038e-01 7.69403160e-01 2.48873681e-01
5.89728773e-01 1.11669934e+00 6.96378648e-02 -4.30401772e-01
-4.08987939e-01 -1.42374724e-01 3.36580127e-01 3.33570033e-01
-2.17342302e-01 -9.26053748e-02 -5.96839011e-01 5.46153247e-01
-2.04276061e+00 -1.54521275e+00 6.12189285e-02 2.36178327e+00
6.64034724e-01 -1.68966264e-01 3.77992481e-01 -1.73119411e-01
1.29410946e+00 1.20521478e-01 -6.93222806e-02 4.16199589e-04
-2.80173004e-01 -4.92715418e-01 3.55776042e-01 4.72777337e-01
-1.61170411e+00 8.39166343e-01 4.73053360e+00 1.30236959e+00
-9.25082207e-01 1.74765900e-01 7.08385766e-01 3.58670503e-01
-6.75419904e-03 -3.33739549e-01 -1.24617660e+00 7.44510114e-01
3.76558363e-01 3.47173631e-01 2.40466401e-01 5.16126156e-01
2.30730884e-03 -8.32293369e-03 -1.01802850e+00 1.47805059e+00
3.90599042e-01 -1.07378066e+00 3.73123467e-01 -4.58723307e-02
6.27641559e-01 -4.50345665e-01 1.32287443e-02 2.28396922e-01
-1.54971257e-01 -8.51995051e-01 8.51689041e-01 6.39775813e-01
9.60761070e-01 -7.89257288e-01 1.04743135e+00 1.92418575e-01
-1.77650166e+00 -1.99572578e-01 -4.39645529e-01 4.85141635e-01
1.88375637e-01 1.91711470e-01 -3.92750651e-01 6.78546309e-01
9.71844554e-01 4.86579448e-01 -9.19901788e-01 1.01405215e+00
1.02746107e-01 1.40794320e-02 -2.35188305e-02 3.55249614e-01
3.30394655e-01 -2.27668956e-01 6.24473691e-01 1.70816565e+00
2.10644916e-01 6.01214841e-02 3.62971872e-01 6.96999073e-01
1.96952716e-01 -2.67073780e-01 -1.91650689e-01 2.70792216e-01
2.39792466e-01 1.30988300e+00 -5.64940453e-01 -5.71871102e-01
-6.14067674e-01 1.29139185e+00 2.90579647e-01 4.01113987e-01
-9.01975334e-01 -2.65387446e-01 5.03346980e-01 2.18843147e-01
3.03658545e-01 -1.35819539e-02 1.67281359e-01 -1.24527049e+00
1.21558078e-01 -1.18075633e+00 9.17366982e-01 -4.63992864e-01
-1.81488621e+00 9.18975890e-01 4.50111955e-01 -1.48168588e+00
-4.01130795e-01 -4.44389522e-01 -4.34257239e-01 1.01792753e+00
-1.44393086e+00 -1.62679768e+00 -7.38984883e-01 1.06693506e+00
5.66933632e-01 -2.86551386e-01 6.49586558e-01 6.64188683e-01
-5.87685585e-01 1.05586517e+00 -1.90588251e-01 4.28933114e-01
9.12625670e-01 -9.88408029e-01 1.71161935e-01 1.26111460e+00
-9.59271491e-02 6.18294358e-01 4.74498838e-01 -7.47928858e-01
-8.59980345e-01 -1.25360858e+00 1.24903464e+00 -1.28017873e-01
3.58433932e-01 -3.68834347e-01 -9.69655395e-01 3.78348380e-01
4.88341115e-02 -5.35937250e-02 4.75496501e-01 -1.16859436e-01
-5.78820467e-01 -3.40703756e-01 -1.19978583e+00 6.56105876e-01
1.00988638e+00 -7.08693147e-01 -6.38677835e-01 1.33318871e-01
3.96080583e-01 -1.17744476e-01 -6.60516083e-01 4.85738933e-01
5.78966856e-01 -1.07736313e+00 1.22992718e+00 -3.94795150e-01
-4.83345836e-02 -8.08364391e-01 -2.60534644e-01 -8.57178152e-01
-8.72408748e-01 -2.53156751e-01 1.66066542e-01 1.70510912e+00
-2.42135655e-02 -6.02582991e-01 3.24423015e-01 3.96279424e-01
6.57684803e-02 -3.85503471e-01 -1.01881897e+00 -6.93021774e-01
-3.88099700e-01 1.97466776e-01 6.12786889e-01 9.24213171e-01
-1.85715541e-01 1.71081275e-01 -6.20337427e-01 4.19817597e-01
7.80217886e-01 1.85806647e-01 7.24392831e-01 -1.08117664e+00
-4.49534327e-01 -5.28276324e-01 -6.62060022e-01 -1.02782154e+00
-1.21525051e-02 -7.44120419e-01 1.10684097e-01 -1.41887164e+00
9.96152878e-01 -3.06678504e-01 -3.42436165e-01 4.39891517e-01
-6.06822968e-01 5.49368799e-01 4.70072657e-01 5.27929664e-01
-9.41027999e-01 3.04526359e-01 1.03693914e+00 -4.42727268e-01
-6.98596379e-03 -1.13839261e-01 -5.74045539e-01 5.38143337e-01
5.53016067e-01 -2.97878563e-01 -7.42505416e-02 -1.92183435e-01
-3.12648892e-01 -1.99428633e-01 7.60715723e-01 -1.21297812e+00
4.63612229e-01 -1.01062939e-01 6.60391033e-01 -6.91946626e-01
2.99106300e-01 -8.30083966e-01 2.49538749e-01 3.80506009e-01
-2.89584279e-01 3.03352922e-01 3.69049013e-02 5.09448469e-01
-4.32393759e-01 -6.80435076e-02 1.02778065e+00 -1.66719899e-01
-1.03011000e+00 5.46174586e-01 -1.25613406e-01 -5.98414354e-02
9.37285185e-01 -4.85967249e-01 -4.96367335e-01 -1.69171572e-01
-5.79505622e-01 1.82318181e-01 5.50376594e-01 4.59642082e-01
6.09296858e-01 -1.66330743e+00 -9.77337360e-01 3.57558638e-01
4.38560158e-01 -2.44339585e-01 7.21165180e-01 9.88999546e-01
-2.23727077e-01 3.50661784e-01 -5.39734185e-01 -7.15088606e-01
-1.68556893e+00 7.65749574e-01 6.09194398e-01 -6.88363612e-01
-4.91017073e-01 7.32381403e-01 8.24744165e-01 -1.51111349e-01
3.37020010e-01 1.95806503e-01 -4.92990404e-01 8.90465304e-02
8.85704875e-01 2.49306828e-01 -2.58522689e-01 -1.38757443e+00
-5.97090244e-01 1.06361914e+00 -1.63142651e-01 1.37851730e-01
8.56215119e-01 -4.48269516e-01 -2.58803695e-01 -4.73656841e-02
1.21342528e+00 -1.98661059e-01 -1.23334527e+00 -3.69476974e-01
-2.61390895e-01 -5.64870715e-01 -4.97520328e-01 -6.25420094e-01
-9.79530692e-01 7.85881698e-01 1.02107131e+00 8.22485238e-02
1.24422956e+00 9.37038809e-02 6.49905205e-01 -2.18136366e-02
2.25917250e-01 -1.05317390e+00 2.08901510e-01 3.53950351e-01
8.04575980e-01 -1.36710739e+00 -5.09366691e-02 -2.85773277e-01
-8.07266176e-01 9.71103787e-01 7.31002450e-01 1.01619966e-01
1.59631342e-01 -2.60229498e-01 -1.00523658e-01 -1.92550912e-01
-2.28203744e-01 -6.91265643e-01 9.49213922e-01 8.01493049e-01
1.64644226e-01 5.46429269e-02 -1.83334455e-01 4.84942287e-01
2.78343856e-01 -5.92144370e-01 -1.54878452e-01 4.10056889e-01
-4.03157711e-01 -1.05306542e+00 -7.72979140e-01 9.24368650e-02
-3.23196769e-01 -1.21568426e-01 -2.60139346e-01 8.86610508e-01
7.21770287e-01 1.10752988e+00 2.74344921e-01 -5.67077935e-01
2.88797617e-01 -1.70135602e-01 3.34061950e-01 -1.53945297e-01
-7.37851620e-01 2.03078151e-01 2.25222334e-02 -5.85028589e-01
-7.28336990e-01 -6.25184894e-01 -8.69586945e-01 -3.43123853e-01
-5.63272275e-02 -7.04347119e-02 1.19034134e-01 9.90339279e-01
2.95901716e-01 2.90123373e-01 4.63260174e-01 -1.10483086e+00
-2.96755254e-01 -8.55045915e-01 -3.59900117e-01 1.11715496e+00
3.48705500e-01 -7.36018300e-01 -3.34788263e-01 1.30865961e-01] | [14.699137687683105, 0.8866704106330872] |
64af894b-a514-4678-a30c-29ddd8251792 | learning-to-detect-3d-reflection-symmetry-for | 2006.10042 | null | https://arxiv.org/abs/2006.10042v1 | https://arxiv.org/pdf/2006.10042v1.pdf | Learning to Detect 3D Reflection Symmetry for Single-View Reconstruction | 3D reconstruction from a single RGB image is a challenging problem in computer vision. Previous methods are usually solely data-driven, which lead to inaccurate 3D shape recovery and limited generalization capability. In this work, we focus on object-level 3D reconstruction and present a geometry-based end-to-end deep learning framework that first detects the mirror plane of reflection symmetry that commonly exists in man-made objects and then predicts depth maps by finding the intra-image pixel-wise correspondence of the symmetry. Our method fully utilizes the geometric cues from symmetry during the test time by building plane-sweep cost volumes, a powerful tool that has been used in multi-view stereopsis. To our knowledge, this is the first work that uses the concept of cost volumes in the setting of single-image 3D reconstruction. We conduct extensive experiments on the ShapeNet dataset and find that our reconstruction method significantly outperforms the previous state-of-the-art single-view 3D reconstruction networks in term of the accuracy of camera poses and depth maps, without requiring objects being completely symmetric. Code is available at https://github.com/zhou13/symmetrynet. | ['Yi Ma', 'Yichao Zhou', 'Shichen Liu'] | 2020-06-17 | null | null | null | null | ['single-view-3d-reconstruction'] | ['computer-vision'] | [ 1.25529006e-01 -1.58375710e-01 1.16619788e-01 -4.01533753e-01
-7.05414772e-01 -6.21528327e-01 4.94003296e-01 -2.96192378e-01
-1.60019815e-01 2.94491053e-02 1.02427393e-01 -3.18816245e-01
7.00915754e-02 -6.79688096e-01 -1.00118566e+00 -3.14160645e-01
4.47177410e-01 7.78717339e-01 2.08591193e-01 -5.25762178e-02
5.00255585e-01 7.80298948e-01 -1.46569753e+00 1.53357595e-01
3.46088648e-01 1.17205691e+00 2.23175243e-01 4.21717584e-01
1.55099079e-01 3.38804066e-01 7.93078765e-02 -2.64343292e-01
8.34804773e-01 -1.20070152e-01 -5.95247805e-01 3.96095335e-01
1.00186837e+00 -8.05208385e-01 -4.07380044e-01 8.07215452e-01
7.51894176e-01 -1.47687539e-01 3.71747375e-01 -9.39086914e-01
-3.92495781e-01 -1.47844270e-01 -6.97819293e-01 -6.19955957e-02
5.70483923e-01 1.43215746e-01 9.40646529e-01 -1.22813010e+00
8.63547444e-01 1.05888236e+00 7.17121899e-01 3.72364402e-01
-1.19528425e+00 -6.37274742e-01 5.41274808e-02 -1.99026410e-02
-1.37274802e+00 -4.29233491e-01 1.17230070e+00 -5.76647937e-01
1.26179695e+00 -9.61755514e-02 6.79401636e-01 8.50839257e-01
-2.77333148e-02 6.13851666e-01 1.29301834e+00 -3.45042616e-01
9.20320675e-02 -3.28839839e-01 -3.57380927e-01 8.44998598e-01
-3.97375040e-02 2.45206788e-01 -6.19109571e-01 2.74854779e-01
1.37084687e+00 2.58975625e-01 -4.53986883e-01 -1.14669025e+00
-1.18646502e+00 6.01050496e-01 6.61146224e-01 -1.78201929e-01
-3.61922115e-01 1.81327894e-01 -1.08667742e-02 1.27281800e-01
6.70605898e-01 2.76755899e-01 -5.85267544e-01 -1.34325385e-01
-7.40461648e-01 3.87813836e-01 4.98015910e-01 9.67211366e-01
7.50925720e-01 -1.39134079e-01 5.70719004e-01 5.59445560e-01
2.76550323e-01 5.95910549e-01 1.19745359e-01 -1.20042503e+00
5.92436850e-01 7.57803977e-01 5.54967970e-02 -7.42397070e-01
-5.92392206e-01 -5.24819374e-01 -4.34474289e-01 3.95490885e-01
4.77480382e-01 3.82572412e-01 -9.53903317e-01 1.37725389e+00
5.16847968e-01 8.52683112e-02 -2.58480072e-01 1.25196028e+00
9.65661049e-01 1.66079417e-01 -8.73623252e-01 2.91477799e-01
9.21506107e-01 -7.08329916e-01 1.08339697e-01 -2.62283176e-01
2.56079078e-01 -8.28628480e-01 1.03496611e+00 5.26775777e-01
-1.23395061e+00 -2.09799767e-01 -8.93828452e-01 -4.36563283e-01
3.55817489e-02 5.02496026e-03 5.83997250e-01 4.40306276e-01
-9.50558066e-01 5.39222658e-01 -8.32259953e-01 -3.42272848e-01
5.79930067e-01 2.53249019e-01 -6.37129247e-01 -5.61059117e-01
-4.41406488e-01 8.09699595e-01 -3.30141671e-02 -8.07805583e-02
-7.65846848e-01 -9.92146015e-01 -9.37151492e-01 -2.75218844e-01
5.31230450e-01 -9.73575711e-01 1.37253344e+00 -5.62516510e-01
-1.30237198e+00 1.41869366e+00 -1.78969115e-01 -8.68758038e-02
7.51846790e-01 -2.17852309e-01 4.36659098e-01 1.81468993e-01
3.76898199e-02 5.97631633e-01 6.00478649e-01 -1.41455853e+00
-2.32255921e-01 -9.78635430e-01 1.95936307e-01 4.60239083e-01
2.75980532e-01 -3.70459646e-01 -6.73338294e-01 -1.97593778e-01
8.34227085e-01 -9.11980689e-01 -7.92699531e-02 4.87606525e-01
-4.25418913e-01 8.05349573e-02 5.15499175e-01 -4.73533154e-01
3.04588556e-01 -1.92042351e+00 3.30147892e-01 1.87569261e-01
2.57159740e-01 -1.57117590e-01 -3.97148216e-03 2.92083651e-01
6.60580620e-02 -1.11191265e-01 -1.98533848e-01 -6.81172967e-01
-1.51815876e-01 -6.44566342e-02 -2.26744577e-01 8.73323083e-01
-7.08768517e-02 8.36002409e-01 -6.21795595e-01 -1.81783885e-02
6.19731128e-01 6.45131111e-01 -6.11342907e-01 1.84026182e-01
-2.27073997e-01 6.30182922e-01 -2.49663264e-01 9.65551078e-01
8.73107255e-01 -3.31951469e-01 -1.59973726e-01 -3.59509230e-01
-2.83860117e-01 4.68309194e-01 -1.04333746e+00 2.30786896e+00
-7.18113720e-01 5.17638326e-01 5.92899024e-02 -6.53336823e-01
9.35755491e-01 -3.47048081e-02 7.00158238e-01 -8.65035295e-01
1.62459120e-01 3.34797949e-01 -2.62801349e-01 -1.47909597e-01
1.78451926e-01 -1.65002495e-01 2.33770952e-01 4.16686356e-01
-1.47727296e-01 -8.12255502e-01 -3.66166234e-01 -1.50680408e-01
8.55701149e-01 5.11369646e-01 3.28470469e-01 1.73825830e-01
1.47596046e-01 -1.09828785e-01 3.86232376e-01 3.46541315e-01
1.60354763e-01 1.31205165e+00 2.25655362e-01 -7.04471052e-01
-1.48326623e+00 -1.28009939e+00 -2.76692867e-01 3.38536888e-01
3.77371758e-01 -2.62083381e-01 -3.47800851e-01 -2.75691628e-01
2.54208237e-01 4.27115232e-01 -3.28472316e-01 2.30010986e-01
-5.33446133e-01 -2.38786817e-01 -3.57561894e-02 5.68034887e-01
5.76243579e-01 -7.37665355e-01 -9.81893957e-01 -7.85136670e-02
-1.12390943e-01 -1.38727844e+00 -3.89223516e-01 6.08072728e-02
-1.18587911e+00 -1.26606655e+00 -6.85230434e-01 -5.19923449e-01
6.94072545e-01 7.33257413e-01 1.15520751e+00 -5.83895855e-02
-3.99306297e-01 5.64758062e-01 -1.53659761e-01 -2.54321426e-01
1.06531657e-01 -1.16382197e-01 -7.57675096e-02 -2.37564012e-01
1.81552574e-01 -1.01195920e+00 -9.01724756e-01 5.30993640e-01
-5.76785982e-01 5.24332285e-01 5.75949609e-01 5.22954285e-01
1.08699048e+00 -2.32794777e-01 -1.85963631e-01 -5.41895032e-01
-5.76763302e-02 -1.03372067e-01 -9.53962386e-01 -2.01390147e-01
-5.13544619e-01 -8.95453170e-02 3.64580154e-01 5.76603971e-03
-6.42366409e-01 5.81896007e-01 -1.00562505e-01 -9.61090922e-01
-3.29404175e-01 1.85407013e-01 -1.33667916e-01 -2.46643513e-01
4.97839749e-01 3.93117011e-01 5.71490452e-02 -6.00440800e-01
1.23053573e-01 2.94325441e-01 4.56094205e-01 -3.42587501e-01
7.30816543e-01 9.98804748e-01 4.24830794e-01 -8.05721998e-01
-9.42857265e-01 -7.78103292e-01 -9.90571976e-01 -3.00333858e-01
7.08850741e-01 -1.19534302e+00 -6.34637952e-01 6.75780594e-01
-1.24868584e+00 -4.23840225e-01 1.48193669e-02 5.03479481e-01
-9.36770141e-01 3.45666587e-01 -2.53134906e-01 -5.76622188e-01
-3.17899972e-01 -1.19674385e+00 1.58041346e+00 -1.20351851e-01
2.37041607e-01 -5.68493664e-01 -4.27211681e-03 8.32871914e-01
1.08495884e-01 4.33927208e-01 6.88805699e-01 3.76725709e-03
-1.22469079e+00 -8.34405944e-02 -2.25919113e-01 1.37612909e-01
-1.11200891e-01 -3.50663126e-01 -1.13571930e+00 -2.41435677e-01
2.72296034e-02 -4.66074675e-01 8.79539013e-01 5.63649893e-01
1.21675777e+00 1.81559697e-01 4.15111370e-02 1.24430513e+00
1.71266079e+00 -1.86490163e-01 6.58553600e-01 4.65221107e-01
1.00741088e+00 6.23732269e-01 4.44779783e-01 5.19499004e-01
6.39548123e-01 8.86393964e-01 1.01301873e+00 -1.09076776e-01
-1.22476995e-01 -5.04736900e-01 2.47659218e-02 3.72293741e-01
-2.89564669e-01 -8.42252672e-02 -1.02479172e+00 4.05733973e-01
-1.56012297e+00 -7.02458143e-01 -1.18010946e-01 2.39049196e+00
3.77825528e-01 1.42042875e-01 2.26542912e-02 9.95013565e-02
3.24260920e-01 6.28914982e-02 -8.75274599e-01 -2.00704038e-01
5.03874458e-02 2.31568635e-01 7.32574284e-01 5.55430353e-01
-9.73451257e-01 8.64715576e-01 4.85349894e+00 2.86715686e-01
-1.24745762e+00 -5.24284914e-02 3.36065471e-01 -4.74461764e-01
-3.88757765e-01 -4.15424593e-02 -6.47405446e-01 -1.93075594e-02
1.08402520e-01 3.68598133e-01 5.51477492e-01 9.12762225e-01
1.42838523e-01 -2.23108664e-01 -1.36015534e+00 1.51262212e+00
4.01018679e-01 -1.21436012e+00 -9.85370278e-02 4.33467805e-01
7.95760632e-01 5.70467532e-01 8.37048814e-02 -3.67300957e-01
-1.09559827e-01 -1.00660813e+00 8.87346208e-01 3.51904362e-01
8.44102442e-01 -7.27629244e-01 4.59196419e-01 4.86107618e-01
-1.01580918e+00 9.35309678e-02 -3.18743259e-01 -5.55120073e-02
2.65818119e-01 6.94779217e-01 -9.96381700e-01 5.07449448e-01
8.47181261e-01 8.73293102e-01 -4.32057500e-01 1.06116307e+00
-2.39148229e-01 2.43817419e-02 -4.96211827e-01 4.12319899e-01
8.12920630e-02 -2.66089261e-01 6.48445606e-01 5.63249946e-01
5.10675192e-01 7.67315999e-02 7.83527792e-02 1.04584563e+00
-2.06559032e-01 -1.17749378e-01 -9.36081648e-01 4.33527440e-01
1.97346345e-01 1.04908383e+00 -8.15901995e-01 2.89246351e-01
-5.62650025e-01 1.11469781e+00 2.56115377e-01 2.94842254e-02
-4.93192703e-01 2.08240166e-01 6.20337367e-01 6.55196667e-01
6.63318634e-01 -6.11504555e-01 -5.74870050e-01 -1.24649966e+00
4.85133260e-01 -4.85738218e-01 -3.97855714e-02 -1.12576282e+00
-1.10661459e+00 3.36861461e-01 -6.98871538e-02 -1.48017049e+00
-1.30280986e-01 -8.47907662e-01 -3.41833860e-01 7.66359568e-01
-1.63354766e+00 -1.29851842e+00 -6.29729867e-01 6.96373940e-01
6.25275612e-01 9.44512710e-03 6.16151452e-01 2.25542225e-02
9.21656117e-02 3.25716466e-01 -1.63572445e-03 -3.65026928e-02
4.98729646e-01 -1.06858802e+00 5.38284183e-01 7.02802181e-01
2.85773605e-01 3.83878291e-01 4.08046603e-01 -4.49621469e-01
-1.88763392e+00 -7.64824986e-01 5.58753073e-01 -6.88931763e-01
1.30140677e-01 -6.66041553e-01 -5.43629110e-01 5.73509634e-01
-3.42207074e-01 2.47786716e-01 2.69072920e-01 -3.32283042e-02
-6.41073465e-01 -1.42623425e-01 -1.16145682e+00 3.49335909e-01
1.52355468e+00 -6.09807789e-01 -2.50207692e-01 2.84820408e-01
5.37293673e-01 -8.99226844e-01 -7.28703737e-01 4.48875010e-01
8.17489803e-01 -1.48436165e+00 1.32879281e+00 -7.38415644e-02
7.50204086e-01 -3.16280305e-01 -4.57032591e-01 -1.15895259e+00
-7.28656277e-02 -2.21676409e-01 1.40938997e-01 6.94634736e-01
2.15889111e-01 -5.19838989e-01 1.03476465e+00 4.52096939e-01
-3.53252947e-01 -9.95268166e-01 -1.09496295e+00 -7.52471328e-01
-2.71817073e-02 -6.14883542e-01 3.68023068e-01 6.98799551e-01
-6.60860658e-01 1.16239652e-01 -1.52356818e-01 2.94605613e-01
8.26990306e-01 6.88180864e-01 1.07143998e+00 -1.30641222e+00
-3.05720448e-01 -4.15411383e-01 -7.34402120e-01 -1.35577762e+00
-7.40599483e-02 -9.58059311e-01 -9.95666757e-02 -1.53114355e+00
1.78742588e-01 -3.31489056e-01 2.24788100e-01 1.69428632e-01
5.52441120e-01 4.39365178e-01 1.60341382e-01 2.06452951e-01
-3.75032544e-01 5.72823584e-01 1.49319065e+00 2.97287494e-01
-1.89352140e-01 -3.84744555e-02 -5.43159842e-01 8.74588490e-01
7.96733916e-01 -2.34565645e-01 -4.03594911e-01 -7.99419165e-01
3.57096225e-01 6.79376349e-02 7.58365154e-01 -7.87181735e-01
1.89081609e-01 -8.26318040e-02 6.15382254e-01 -1.06108892e+00
7.73407221e-01 -9.57684994e-01 9.64292660e-02 2.40711987e-01
-4.14338447e-02 1.73859581e-01 6.78968877e-02 4.31545794e-01
1.46674588e-01 7.81222135e-02 7.38275468e-01 -5.49205303e-01
-7.28463888e-01 6.79168046e-01 3.16794366e-01 1.17666274e-01
7.94322968e-01 -5.01093090e-01 2.78902743e-02 -3.95017415e-01
-2.37798125e-01 -5.88718764e-02 1.17680907e+00 2.74054170e-01
1.24184585e+00 -1.16470277e+00 -6.56932294e-01 3.92560154e-01
2.78209746e-01 6.60655081e-01 1.58683255e-01 7.33725846e-01
-8.51436555e-01 3.97394538e-01 -2.66760826e-01 -1.10089397e+00
-1.32882297e+00 2.97949821e-01 6.08988762e-01 1.23207472e-01
-1.03535020e+00 7.28343606e-01 3.66718411e-01 -7.91898310e-01
1.58076033e-01 -4.78484005e-01 3.76454800e-01 -4.54772830e-01
2.16142103e-01 1.96316615e-01 4.06666368e-01 -7.16597915e-01
-4.15649831e-01 1.30401945e+00 9.60261822e-02 -2.80945867e-01
1.68343699e+00 -8.32601041e-02 7.37531260e-02 2.49748290e-01
1.36052358e+00 -7.47430697e-02 -1.57105517e+00 -3.01044315e-01
-5.85197985e-01 -1.04457235e+00 3.64214063e-01 -7.11777329e-01
-1.21354735e+00 1.05174077e+00 6.48571312e-01 -4.73379552e-01
1.02876282e+00 2.93944985e-01 7.93429673e-01 4.12475914e-01
7.04910934e-01 -6.99817359e-01 1.80875912e-01 5.11352301e-01
1.22648180e+00 -1.51844871e+00 2.38595277e-01 -4.98441964e-01
-2.64579177e-01 1.13784003e+00 7.64986873e-01 -4.14594054e-01
6.09516799e-01 5.26704639e-02 -1.77088097e-01 -5.18185079e-01
-3.92814010e-01 -1.35809630e-01 3.39054853e-01 6.61884069e-01
1.97944731e-01 -7.52990693e-02 2.33138040e-01 9.11416784e-02
-4.61937755e-01 -1.51286840e-01 4.44488138e-01 9.21436012e-01
-2.17007056e-01 -8.39737117e-01 -3.57332289e-01 1.09929360e-01
-2.28251174e-01 -4.34111133e-02 -6.20342791e-01 7.39673853e-01
3.55928838e-02 5.80482662e-01 2.49346673e-01 -4.09936428e-01
6.69850409e-01 -2.21152186e-01 1.15605712e+00 -6.32584870e-01
-9.10041928e-02 -2.03832667e-02 -8.71143565e-02 -9.87546444e-01
-5.33901215e-01 -7.72273004e-01 -1.18158758e+00 -3.34106684e-01
3.27890702e-02 -8.24179173e-01 1.06436479e+00 6.48104370e-01
3.54210705e-01 -1.15019649e-01 7.56635785e-01 -1.42971170e+00
-3.52903992e-01 -5.72579920e-01 -4.23667490e-01 4.67265517e-01
4.90129203e-01 -8.41233075e-01 -3.82874191e-01 -3.24819684e-01] | [8.550950050354004, -2.872652292251587] |
d5a237d9-d118-446f-9523-5cbc4a028a58 | differentially-private-decision-trees-with | 2305.15394 | null | https://arxiv.org/abs/2305.15394v1 | https://arxiv.org/pdf/2305.15394v1.pdf | Differentially-Private Decision Trees with Probabilistic Robustness to Data Poisoning | Decision trees are interpretable models that are well-suited to non-linear learning problems. Much work has been done on extending decision tree learning algorithms with differential privacy, a system that guarantees the privacy of samples within the training data. However, current state-of-the-art algorithms for this purpose sacrifice much utility for a small privacy benefit. These solutions create random decision nodes that reduce decision tree accuracy or spend an excessive share of the privacy budget on labeling leaves. Moreover, many works do not support or leak information about feature values when data is continuous. We propose a new method called PrivaTree based on private histograms that chooses good splits while consuming a small privacy budget. The resulting trees provide a significantly better privacy-utility trade-off and accept mixed numerical and categorical data without leaking additional information. Finally, while it is notoriously hard to give robustness guarantees against data poisoning attacks, we prove bounds for the expected success rates of backdoor attacks against differentially-private learners. Our experimental results show that PrivaTree consistently outperforms previous works on predictive accuracy and significantly improves robustness against backdoor attacks compared to regular decision trees. | ['Sicco Verwer', 'Zekeriya Erkin', 'Tianyu Li', 'Jelle Vos', 'Daniël Vos'] | 2023-05-24 | null | null | null | null | ['data-poisoning'] | ['adversarial'] | [ 1.84288114e-01 3.86888057e-01 -4.82814938e-01 -7.87062883e-01
-1.13576114e+00 -9.05276477e-01 2.05170020e-01 4.72366571e-01
-4.01426554e-01 7.96226561e-01 -2.48032883e-01 -8.64292860e-01
-1.05429284e-01 -1.22894478e+00 -6.19075298e-01 -1.14886796e+00
-1.83486074e-01 4.78579044e-01 7.36711323e-02 2.83457458e-01
3.17188948e-02 6.27523363e-01 -1.47912788e+00 4.42528158e-01
7.43696332e-01 9.45230961e-01 -1.03761423e+00 5.05011797e-01
7.73475841e-02 5.79878509e-01 -5.82432926e-01 -1.18665147e+00
7.77554154e-01 -2.98196822e-01 -8.53733659e-01 -3.90568882e-01
7.08357751e-01 -5.37093103e-01 -3.35398018e-01 1.09975421e+00
5.86627722e-01 -4.53512609e-01 6.00561559e-01 -1.82689500e+00
-4.59064633e-01 9.81643736e-01 -5.68980813e-01 -3.05135876e-01
-6.61649555e-02 -8.31699818e-02 1.09657633e+00 -2.05249816e-01
2.54753977e-01 1.16352856e+00 8.85074258e-01 6.72787726e-01
-1.68208039e+00 -1.27620590e+00 -2.22773492e-01 2.25395262e-01
-1.37351596e+00 -3.87580335e-01 4.46676731e-01 -2.49023750e-01
1.55210838e-01 1.02554739e+00 3.91906500e-01 1.20951676e+00
2.92315274e-01 9.08329606e-01 1.32488823e+00 -2.47713953e-01
5.93435287e-01 6.75732076e-01 3.60751510e-01 6.58024073e-01
8.21905792e-01 2.31537059e-01 -5.27476907e-01 -9.73024070e-01
1.04265906e-01 3.96411836e-01 -2.45565519e-01 -1.09164691e+00
-4.82072115e-01 1.27300274e+00 3.03844154e-01 -1.53726384e-01
3.04921627e-01 1.18521035e-01 5.56483209e-01 6.81265056e-01
4.57893997e-01 -1.51872425e-03 -5.33765912e-01 2.87331581e-01
-7.00236917e-01 5.61286092e-01 1.22215521e+00 9.60208654e-01
5.93115151e-01 -5.89120507e-01 -2.26232901e-01 2.61214972e-01
1.42595783e-01 2.69595236e-01 2.98282593e-01 -7.44939923e-01
5.05629182e-01 2.54864454e-01 1.95348132e-02 -9.88012731e-01
5.34153245e-02 -5.02042651e-01 -7.75376379e-01 4.96071368e-01
7.69562542e-01 -2.57732719e-01 -4.21831101e-01 1.79106569e+00
6.78143144e-01 -5.09730637e-01 2.04400256e-01 4.07607764e-01
6.06804602e-02 2.93303251e-01 9.13191363e-02 -7.20313489e-02
1.49046087e+00 -4.39337850e-01 -5.76023877e-01 8.49349871e-02
1.27064991e+00 -2.07147285e-01 9.15757239e-01 5.29188097e-01
-7.75985897e-01 3.47851008e-01 -1.10787416e+00 -2.76466638e-01
-4.70707804e-01 -4.59534228e-01 1.08167195e+00 1.67675984e+00
-5.26491821e-01 5.68640292e-01 -8.92644644e-01 1.90733485e-02
1.05584347e+00 6.91518962e-01 -6.43504262e-01 -1.58882558e-01
-9.21737492e-01 5.72124243e-01 6.22737259e-02 -4.89262372e-01
-5.54553032e-01 -8.76918554e-01 -6.83673382e-01 1.69827193e-01
3.91835064e-01 -5.42042077e-01 1.32788193e+00 -5.41573584e-01
-1.12515557e+00 1.13946795e+00 1.63913354e-01 -1.25845611e+00
1.10463035e+00 -4.82253963e-03 1.29543200e-01 -2.36165747e-01
-1.88676581e-01 4.18795645e-01 6.14892125e-01 -9.95062768e-01
-9.84554887e-01 -1.12807679e+00 -1.68475956e-01 -3.18910144e-02
-8.01162899e-01 -3.46446931e-01 3.53942126e-01 -4.93690431e-01
4.05969173e-02 -1.03609264e+00 -3.94266844e-01 6.22434258e-01
-5.71080387e-01 -1.76125571e-01 1.31991482e+00 -2.52824545e-01
1.16637778e+00 -2.30444169e+00 -4.86929089e-01 2.90493190e-01
4.24983859e-01 1.25995368e-01 3.92063648e-01 2.33081430e-01
1.67448089e-01 5.59114575e-01 -5.13195813e-01 -4.47178125e-01
3.75904262e-01 3.04181367e-01 -7.45237947e-01 8.56639862e-01
-6.04368091e-01 5.44302344e-01 -4.16772157e-01 -4.41647202e-01
-1.57196641e-01 2.40642503e-01 -7.85339773e-01 -1.77507568e-02
-5.37695969e-03 -4.56957556e-02 -4.71967012e-01 5.16615629e-01
1.12719738e+00 1.07319593e-01 2.35620290e-01 4.10951376e-01
3.36064577e-01 3.51189613e-01 -9.91183162e-01 1.02819383e+00
-1.40529156e-01 3.87592852e-01 2.12298527e-01 -8.42921257e-01
7.68599391e-01 1.37176737e-01 3.37936878e-01 4.57236767e-02
1.93281889e-01 1.49475992e-01 -1.67700127e-01 -2.20433027e-02
3.98561619e-02 -3.26273292e-01 -3.60637724e-01 6.21577084e-01
-3.99099916e-01 4.93529290e-02 -6.86671078e-01 2.43605915e-02
1.26132464e+00 -4.20424521e-01 3.38436484e-01 -2.46743903e-01
2.97714561e-01 -2.07736745e-01 7.62582183e-01 9.00582075e-01
-2.79204577e-01 3.18230897e-01 7.87457407e-01 -5.80068648e-01
-6.71854496e-01 -8.44375491e-01 -5.24952292e-01 1.13874567e+00
-1.86184838e-01 -5.09316742e-01 -1.01998460e+00 -1.25316453e+00
7.15147734e-01 9.74613488e-01 -7.94054031e-01 -3.25510830e-01
1.04327649e-01 -8.15244853e-01 1.04246175e+00 2.25064725e-01
6.79469585e-01 -2.59552479e-01 -6.96118474e-01 -1.13507830e-01
1.56320527e-01 -6.22369111e-01 -3.61910105e-01 8.03349018e-01
-9.65462565e-01 -1.01786268e+00 -2.18292043e-01 -3.77478033e-01
6.24013066e-01 2.79659569e-01 6.23914778e-01 -1.98740929e-01
-2.95661539e-01 1.49229020e-01 9.30400193e-02 -6.98913515e-01
-5.05890250e-01 2.94691980e-01 -1.39049679e-01 1.59151554e-01
9.85184789e-01 -6.37331784e-01 -3.49416584e-01 3.02744150e-01
-8.77797246e-01 -2.12093636e-01 1.70009926e-01 8.34571660e-01
6.49616539e-01 2.56288379e-01 1.53423354e-01 -1.57253397e+00
3.64036649e-01 -5.37919402e-01 -8.07324469e-01 2.88247526e-01
-9.60701644e-01 3.06267142e-01 8.27416420e-01 -4.25544769e-01
-7.36642957e-01 3.53638738e-01 3.70745845e-02 -2.45985955e-01
-9.47548151e-02 -2.30790466e-01 -5.34662426e-01 -3.56961161e-01
8.13741744e-01 3.10760364e-02 8.18964392e-02 -4.59531039e-01
3.49556416e-01 9.85330999e-01 2.77920812e-01 -5.65420628e-01
7.39378810e-01 6.35363758e-01 4.55542147e-01 -3.56182337e-01
-8.44192982e-01 -5.52954758e-03 -3.76730770e-01 7.20741212e-01
4.81607795e-01 -6.51375711e-01 -9.47054207e-01 4.66575444e-01
-4.88668710e-01 -3.38365324e-02 -6.44697964e-01 1.25058174e-01
-4.95197237e-01 4.40576375e-01 -3.96785021e-01 -1.01880562e+00
-4.77654397e-01 -8.05666924e-01 5.76215982e-01 -2.38844588e-01
-1.13481782e-01 -8.15916359e-01 -2.77050316e-01 4.95763630e-01
1.39170036e-01 5.54904699e-01 9.94963646e-01 -1.23193371e+00
-6.35183752e-01 -6.41309202e-01 2.43106425e-01 1.32329166e-01
-9.27413553e-02 -2.66497642e-01 -1.33095396e+00 -4.96208102e-01
4.23117250e-01 -5.05841434e-01 9.42208529e-01 2.64198408e-02
1.75652599e+00 -1.13190341e+00 -5.01626074e-01 9.76918697e-01
1.11882401e+00 4.40083407e-02 3.96773756e-01 1.78168848e-01
4.28363472e-01 8.06431592e-01 4.45924222e-01 7.52593577e-01
2.44952261e-01 5.25969088e-01 3.58397961e-01 1.80809423e-02
5.87683141e-01 -6.85819447e-01 2.45117247e-01 -2.18070880e-01
7.76588202e-01 -1.66837722e-01 -6.18631780e-01 3.56957167e-01
-1.73330486e+00 -8.56629193e-01 -1.59187421e-01 2.87060094e+00
1.28117692e+00 6.70056716e-02 1.62711486e-01 4.90784019e-01
4.26003188e-01 -2.23796085e-01 -7.96121538e-01 -6.90616250e-01
-1.14039294e-01 2.57135004e-01 1.27288902e+00 3.59230608e-01
-1.35216272e+00 5.79732180e-01 5.92392731e+00 9.56821740e-01
-8.04280818e-01 5.33891879e-02 1.09124267e+00 -3.84064168e-01
-5.16116202e-01 1.82623342e-01 -9.66030598e-01 3.25152069e-01
1.04214501e+00 -6.06479049e-01 7.97171891e-02 1.41441059e+00
-5.98180234e-01 1.50935039e-01 -1.61328471e+00 9.18880939e-01
-3.55256259e-01 -1.10806012e+00 -2.37719670e-01 7.56500244e-01
3.28343749e-01 -4.18998331e-01 5.28852046e-01 1.86275661e-01
9.74230349e-01 -1.08302832e+00 4.74305928e-01 -1.12580828e-01
7.43151128e-01 -1.23369682e+00 5.98362207e-01 6.33258283e-01
-5.30902445e-01 -4.81880069e-01 -6.24989808e-01 1.35007054e-01
-7.16765702e-01 6.20446444e-01 -8.39240432e-01 1.84359878e-01
7.82691836e-01 1.33348852e-01 -6.93170428e-01 8.49821806e-01
-2.46085525e-01 9.09108043e-01 -8.15923989e-01 9.93375760e-03
4.12537530e-02 -1.34330140e-02 2.96215415e-01 1.08825409e+00
1.71782315e-01 6.48117214e-02 -8.93565733e-03 6.35907292e-01
-3.72304171e-01 2.13348791e-01 -1.05702257e+00 3.54538977e-01
8.86851609e-01 9.04052258e-01 -2.16701165e-01 -1.21494876e-02
-8.31656158e-02 9.81557965e-01 2.61380166e-01 -1.97636962e-01
-4.75855857e-01 -3.96805048e-01 9.59916830e-01 1.48238510e-01
3.71408701e-01 2.54250228e-01 -7.95286894e-01 -1.05474985e+00
-2.88998969e-02 -1.04478157e+00 1.07589281e+00 1.82696432e-01
-1.43801975e+00 -4.83166613e-02 -1.56901509e-01 -9.62059617e-01
-2.77492106e-01 -2.78637916e-01 -1.27406120e-01 6.21341765e-01
-9.31731820e-01 -1.10543025e+00 2.73992419e-01 6.72712684e-01
-2.15736166e-01 -4.83529493e-02 1.35094857e+00 -1.17657907e-01
-5.59432864e-01 1.56899679e+00 5.86257577e-01 3.14914100e-02
6.32156849e-01 -1.41324663e+00 2.74035871e-01 6.92614377e-01
2.05027208e-01 6.04761839e-01 6.66144669e-01 -2.50857323e-01
-1.45270896e+00 -1.02880669e+00 1.00529122e+00 -5.58007598e-01
2.46404991e-01 -8.51235986e-01 -9.71556306e-01 9.57938373e-01
-2.90906608e-01 3.29612285e-01 1.38874245e+00 3.16934586e-01
-9.47899818e-01 -5.34682035e-01 -2.05663013e+00 5.35831094e-01
8.01541269e-01 -4.78040218e-01 -1.66675135e-01 3.23956758e-01
5.55044293e-01 -2.74267316e-01 -8.34734678e-01 5.09518236e-02
7.60109544e-01 -1.20095062e+00 8.00053954e-01 -8.05568516e-01
-1.01131506e-01 -3.60464118e-03 -3.17489684e-01 -7.20299482e-01
-1.57885568e-03 -9.72526729e-01 -2.61328548e-01 1.33881891e+00
3.05709958e-01 -1.05370045e+00 1.37042367e+00 1.34546101e+00
7.73995459e-01 -6.71551764e-01 -1.30600333e+00 -8.41387331e-01
7.42980301e-01 -2.20772386e-01 9.37195182e-01 1.14547420e+00
4.63534370e-02 -8.00984818e-03 -4.43721980e-01 1.29639149e-01
1.29038537e+00 1.86671078e-01 1.09196103e+00 -1.41279340e+00
-2.31505945e-01 -1.64646298e-01 -6.15164518e-01 -6.73502564e-01
2.11552277e-01 -1.14068973e+00 -5.51513910e-01 -7.34865069e-01
1.91287473e-01 -7.81172156e-01 -1.82657272e-01 1.04549825e+00
6.03812523e-02 -7.05397502e-02 6.73021227e-02 3.08565069e-02
-1.57814503e-01 5.00491858e-01 5.92892706e-01 -7.68089965e-02
-5.17280772e-02 6.63279712e-01 -1.16360438e+00 5.58747411e-01
8.01751733e-01 -1.09466243e+00 -4.93056059e-01 1.08029954e-01
-1.31874874e-01 5.75906858e-02 3.69287074e-01 -9.02190804e-01
3.41269612e-01 -3.35099697e-02 3.06670398e-01 -6.47881210e-01
-7.97001868e-02 -1.31502581e+00 3.60163391e-01 9.14667666e-01
-8.22055161e-01 -4.20231253e-01 -1.86441645e-01 8.55390787e-01
2.52780497e-01 -2.81274319e-01 9.21819806e-01 2.60038108e-01
2.31598333e-01 5.95599353e-01 -3.43947828e-01 -2.95454636e-02
1.43139434e+00 -3.09848070e-01 -2.13561445e-01 -4.16747540e-01
-4.25423294e-01 2.81386197e-01 7.26665258e-01 -7.51176616e-03
1.73980907e-01 -1.02461255e+00 -6.22056425e-01 4.57033306e-01
1.17872424e-01 4.58071344e-02 -6.39939681e-02 3.38999927e-01
-1.77365422e-01 2.72700310e-01 -1.10428138e-02 -2.27635145e-01
-2.00037146e+00 7.77987659e-01 3.69122773e-01 -3.39220285e-01
-8.09647620e-01 1.06547201e+00 1.46564811e-01 -6.06122196e-01
8.16201687e-01 -1.17336176e-01 4.78924960e-01 1.01197444e-01
5.44845939e-01 4.34151292e-01 5.15755191e-02 5.59047004e-03
-3.13748479e-01 -8.32156688e-02 -4.25210088e-01 -9.78974849e-02
1.10349143e+00 1.76456898e-01 3.74292508e-02 7.17907399e-02
1.44761777e+00 3.12944829e-01 -1.18738914e+00 -3.88537496e-01
5.62370159e-02 -8.18094373e-01 1.66867245e-02 -6.96536541e-01
-9.35136676e-01 1.12223268e+00 8.97173762e-01 4.01317537e-01
1.07352221e+00 -3.54319334e-01 8.15131903e-01 5.66175818e-01
6.58685029e-01 -6.90495908e-01 -8.30949962e-01 -8.48951042e-02
3.41718197e-01 -9.97160375e-01 3.58557731e-01 -6.68257356e-01
-5.76735973e-01 9.16899264e-01 3.22439015e-01 2.85736024e-01
7.64479995e-01 6.66938305e-01 -9.60557088e-02 4.30796087e-01
-1.01971757e+00 3.95495206e-01 -3.58800620e-01 8.65271151e-01
-5.44991307e-02 4.31403637e-01 -3.68861258e-01 9.82452393e-01
-6.11182868e-01 -9.11321118e-02 3.33733082e-01 1.08568335e+00
-2.93452531e-01 -1.57418466e+00 -5.19270957e-01 4.34076220e-01
-1.04999387e+00 1.04297295e-01 -8.15524817e-01 6.96165621e-01
-9.27725155e-03 6.56315804e-01 -1.37630880e-01 -4.82442677e-01
3.36025134e-02 2.76724994e-01 2.35353291e-01 -2.27911174e-01
-9.63299692e-01 -5.33733845e-01 -2.29791962e-02 -3.68253887e-01
3.69520515e-01 -8.54490578e-01 -1.01062155e+00 -9.74411368e-01
-2.98301905e-01 3.37849379e-01 5.06029665e-01 3.83921295e-01
5.05515218e-01 -3.79040718e-01 1.00402486e+00 1.89562902e-01
-1.49360383e+00 -3.36798221e-01 -8.64001989e-01 3.34817797e-01
3.30139101e-01 -1.27064571e-01 -6.44747913e-01 -3.41902286e-01] | [5.936432838439941, 6.887723445892334] |
0d8d6927-f8c5-4eee-8a8d-40ae1179b2f2 | emotion-cause-pair-extraction-in-customer | 2112.03984 | null | https://arxiv.org/abs/2112.03984v1 | https://arxiv.org/pdf/2112.03984v1.pdf | Emotion-Cause Pair Extraction in Customer Reviews | Emotion-Cause Pair Extraction (ECPE) is a complex yet popular area in Natural Language Processing due to its importance and potential applications in various domains. In this report , we aim to present our work in ECPE in the domain of online reviews. With a manually annotated dataset, we explore an algorithm to extract emotion cause pairs using a neural network. In addition, we propose a model using previous reference materials and combining emotion-cause pair extraction with research in the domain of emotion-aware word embeddings, where we send these embeddings into a Bi-LSTM layer which gives us the emotionally relevant clauses. With the constraint of a limited dataset, we achieved . The overall scope of our report comprises of a comprehensive literature review, implementation of referenced methods for dataset construction and initial model training, and modifying previous work in ECPE by proposing an improvement to the pipeline, as well as algorithm development and implementation for the specific domain of reviews. | ['Wyatt Pease', 'Nathan Johns', 'Aishwarya Kaliki', 'Jeel Tejaskumar Vaishnav', 'Arpit Mittal'] | 2021-12-07 | null | null | null | null | ['emotion-cause-pair-extraction'] | ['natural-language-processing'] | [ 1.28190488e-01 4.25040781e-01 -2.82306254e-01 -7.42411435e-01
-6.05415642e-01 -4.48861480e-01 4.30166841e-01 4.38005090e-01
-7.51413226e-01 6.25505030e-01 4.09480691e-01 -2.55497266e-02
3.60586122e-02 -4.66791600e-01 -3.28082561e-01 -1.68814704e-01
-1.04979709e-01 1.73209980e-01 -6.21440768e-01 -1.38138562e-01
4.92094189e-01 4.19665545e-01 -1.56984615e+00 4.97051090e-01
5.42000651e-01 1.07577562e+00 -2.29820326e-01 6.92944467e-01
-5.20381868e-01 9.79208410e-01 -7.08813131e-01 -7.93967366e-01
-9.19938385e-02 -3.27465862e-01 -1.06496203e+00 -1.49953485e-01
5.43665886e-02 -8.69950205e-02 2.45744586e-01 6.54152274e-01
5.56376457e-01 3.97105962e-01 6.20320559e-01 -1.48182797e+00
-7.76986659e-01 7.91598082e-01 -2.61113703e-01 -1.38296634e-01
3.52063537e-01 -6.05968297e-01 1.31025064e+00 -1.15712500e+00
9.75559950e-01 8.56957197e-01 6.56331241e-01 9.00938809e-01
-5.87253153e-01 -5.39587855e-01 2.63600260e-01 5.15317500e-01
-9.59399462e-01 -2.20076784e-01 1.13863993e+00 -1.26030862e-01
1.81203699e+00 5.22956206e-03 8.60944867e-01 1.17828107e+00
1.95293352e-01 1.07466030e+00 9.44127321e-01 -8.10169160e-01
4.08852160e-01 5.92776597e-01 4.12445098e-01 3.04571360e-01
-1.13393866e-01 -3.97739947e-01 -4.64187533e-01 3.15228067e-02
1.55327946e-01 -4.02006000e-01 6.45909905e-02 -7.65466988e-02
-6.52502775e-01 1.14290929e+00 2.35682383e-01 6.99249446e-01
-4.45881099e-01 7.15919882e-02 8.97718847e-01 3.13217282e-01
8.24812531e-01 8.08448315e-01 -9.92356837e-01 -3.74360591e-01
-1.02393377e+00 3.25032830e-01 1.11979938e+00 8.78151238e-01
3.90459925e-01 -7.53270313e-02 1.19316936e-01 1.05237865e+00
2.36462206e-01 -2.24454552e-01 6.70891285e-01 -7.52499700e-01
2.38207921e-01 7.85295129e-01 -8.99848565e-02 -1.15741587e+00
-5.64186752e-01 -2.22856864e-01 -3.80617827e-01 -1.57729194e-01
-3.42531562e-01 -6.36390090e-01 -5.78671992e-01 1.59154141e+00
2.39167765e-01 -9.31004584e-02 4.00453031e-01 6.88806415e-01
1.42722452e+00 7.45057642e-01 2.83456057e-01 -3.54547799e-01
1.53398252e+00 -1.07891023e+00 -1.33062696e+00 -4.13307965e-01
9.71274376e-01 -7.30738521e-01 8.67714107e-01 6.92525744e-01
-9.83505666e-01 -2.56265998e-01 -1.19240916e+00 -6.49020910e-01
-1.26893878e+00 4.39594388e-01 8.47162545e-01 4.42999005e-01
-1.04745531e+00 4.57850367e-01 -3.07890296e-01 -4.60949838e-01
3.38404447e-01 3.93252909e-01 -4.96311665e-01 2.24860772e-01
-1.58651829e+00 1.36095190e+00 5.95962107e-01 1.75508842e-01
2.41482556e-02 -5.01862228e-01 -1.34546888e+00 -6.16729818e-02
1.57430708e-01 -1.87164649e-01 1.32555401e+00 -1.21105325e+00
-1.53006661e+00 1.01665235e+00 -3.52421701e-01 -5.01709700e-01
-3.17918271e-01 -6.32535756e-01 -7.83918262e-01 2.92726029e-02
-2.26833925e-01 1.05526805e+00 4.65502441e-01 -9.63949800e-01
-4.43676859e-01 -7.98699185e-02 -3.93501595e-02 2.31140211e-01
-7.23459959e-01 5.39779067e-01 -3.00930053e-01 -5.83366394e-01
-7.89405763e-01 -5.94681740e-01 -3.98901403e-01 -2.95061707e-01
-1.05453469e-01 -6.94644988e-01 8.46432447e-01 -7.01249957e-01
1.55764103e+00 -2.07722402e+00 -7.07347458e-03 -8.40172470e-02
1.17981352e-01 3.03020835e-01 -3.35841179e-01 5.66682279e-01
-6.37142599e-01 4.42970365e-01 -2.27427244e-01 -6.30178034e-01
3.04436982e-01 2.98882633e-01 -3.04416716e-01 -1.15908245e-02
9.39337611e-01 1.04261267e+00 -7.69418418e-01 -6.10143065e-01
2.13191912e-01 6.59170866e-01 -4.78645325e-01 3.34656447e-01
-9.62017924e-02 -5.13663292e-01 -1.89723164e-01 2.90269196e-01
6.86414003e-01 2.87082374e-01 5.33259427e-03 -2.71848828e-01
-2.96792358e-01 4.87862378e-01 -9.69613731e-01 1.62891507e+00
-8.85363460e-01 9.11989152e-01 3.60899530e-02 -9.58880186e-01
1.22482848e+00 5.88648140e-01 4.04810756e-01 -5.35985827e-01
5.63930929e-01 3.43093842e-01 -9.03993472e-02 -8.60694408e-01
9.81252074e-01 -2.41191342e-01 -2.84569114e-01 5.15933096e-01
5.48430383e-01 -3.95983100e-01 4.88356203e-01 1.81888178e-01
9.00004625e-01 2.59852380e-01 7.14298189e-01 1.36384785e-01
4.89624768e-01 2.39292085e-02 3.31886262e-01 -2.32467446e-02
-5.14621854e-01 3.40395004e-01 6.69200361e-01 -5.44558883e-01
-9.16290998e-01 -2.52784133e-01 2.14078277e-02 9.35677648e-01
-5.68861306e-01 -7.89745390e-01 -7.92232215e-01 -8.19873750e-01
-4.61569607e-01 1.20054913e+00 -1.00801826e+00 -1.19210161e-01
-5.88723123e-01 -8.45111609e-01 3.74457270e-01 4.74649280e-01
1.33931562e-01 -1.59727526e+00 -8.23354840e-01 3.77813786e-01
-4.80581410e-02 -1.23908758e+00 1.61749616e-01 7.83005059e-01
-4.72305775e-01 -8.92907679e-01 -3.28625381e-01 -1.00404596e+00
3.27437997e-01 -3.98941696e-01 1.46903217e+00 -2.28501543e-01
-4.45177406e-01 2.43474692e-01 -7.22941875e-01 -1.02572608e+00
-7.35040009e-02 1.78356066e-01 -2.57217288e-01 -4.01384562e-01
1.36260009e+00 -1.89226106e-01 -2.31574938e-01 -5.87811589e-01
-1.02947509e+00 -2.38122538e-01 4.62183028e-01 5.55970550e-01
2.52757937e-01 -3.77767086e-02 8.81026924e-01 -8.46586168e-01
1.29641604e+00 -7.09624290e-01 3.11118849e-02 7.02903047e-03
-5.77633858e-01 -1.09750517e-01 6.84117854e-01 -1.64385632e-01
-1.08152056e+00 9.56906378e-02 -5.23510814e-01 -5.61492331e-02
-4.74397212e-01 8.18735421e-01 -9.46527272e-02 3.94848555e-01
3.43843430e-01 -4.13779050e-01 -2.57908612e-01 -2.12754905e-01
8.40454102e-01 8.88080657e-01 8.59689116e-02 -3.74430567e-01
4.36753966e-02 1.33330645e-02 -2.97235131e-01 -5.29596269e-01
-9.16883826e-01 -3.72189492e-01 -7.31283307e-01 -2.60293454e-01
8.84158015e-01 -6.72071457e-01 -3.40760887e-01 -2.47348268e-02
-1.49637842e+00 1.00382879e-01 -4.64535564e-01 4.65357393e-01
-1.35341868e-01 2.88349520e-02 -6.42058134e-01 -8.59473765e-01
-6.97378457e-01 -8.62485111e-01 9.78288591e-01 3.60627115e-01
-9.55593586e-01 -1.41114032e+00 2.84977168e-01 -5.54398410e-02
3.83018821e-01 4.10447389e-01 8.73308063e-01 -1.21309412e+00
5.08274019e-01 -3.72052222e-01 -1.20423518e-01 6.19036078e-01
1.04190847e-02 4.24984723e-01 -1.23465669e+00 4.11402553e-01
2.75751352e-01 -7.97689021e-01 6.29150867e-01 3.58605683e-01
1.04998899e+00 -1.67280048e-01 -2.69764308e-02 1.60379991e-01
1.43898129e+00 3.85563225e-01 5.21907747e-01 5.45915306e-01
3.89355242e-01 1.18517482e+00 8.25550497e-01 4.79509711e-01
4.69777256e-01 2.45690614e-01 2.62500107e-01 -3.57218832e-01
2.04919413e-01 1.68739483e-01 4.73851562e-01 1.27683210e+00
3.66483688e-01 -2.90830791e-01 -6.24554157e-01 9.82874691e-01
-1.64141262e+00 -6.37129188e-01 -2.77278870e-01 1.25332940e+00
9.87500191e-01 -1.32018477e-01 -1.49134368e-01 2.00053513e-01
4.32516098e-01 2.51456350e-01 -2.11812854e-01 -1.86087489e+00
2.47617229e-03 7.46325731e-01 1.61245223e-02 4.59714204e-01
-1.22795272e+00 1.04493165e+00 6.53135824e+00 7.49439240e-01
-1.14267206e+00 3.10591478e-02 6.53994203e-01 -2.95679629e-01
-3.14017326e-01 -2.66635388e-01 -5.45541883e-01 -4.05864604e-02
1.30271673e+00 -9.82247591e-02 1.32031426e-01 1.09093320e+00
2.93927222e-01 -6.16578106e-03 -1.27335298e+00 8.58574569e-01
5.25658727e-01 -9.61851835e-01 -2.62618363e-01 -4.02737707e-01
5.03579617e-01 -7.97473863e-02 -1.46445543e-01 5.69644153e-01
1.25381082e-01 -1.11937106e+00 2.47313425e-01 1.71740904e-01
5.03985703e-01 -1.14708006e+00 1.37832010e+00 -1.59065530e-01
-9.55626309e-01 1.49351612e-01 -2.16860950e-01 -4.16285723e-01
2.08311409e-01 8.71077538e-01 -8.24125350e-01 5.90947747e-01
7.49047637e-01 1.16109633e+00 -4.61184233e-01 4.61780310e-01
-5.83505213e-01 4.12687510e-01 -1.45807475e-01 -4.91882086e-01
3.79512370e-01 -9.62926820e-03 4.17542532e-02 1.90854561e+00
2.38267168e-01 -1.23263095e-02 -1.69592142e-01 9.56452906e-01
-2.54485488e-01 5.37092328e-01 -6.64291859e-01 -2.68469572e-01
3.20362598e-01 1.93836808e+00 -6.41740024e-01 -4.16341782e-01
-3.37932944e-01 8.51975441e-01 6.46191120e-01 2.92362757e-02
-8.01735103e-01 -1.04417479e+00 3.75838101e-01 -7.93543398e-01
2.51407593e-01 1.57583892e-01 -4.93474215e-01 -1.05877459e+00
2.21535504e-01 -8.19349408e-01 2.40778878e-01 -8.65712047e-01
-1.44321322e+00 8.80331337e-01 -2.51468946e-03 -9.34275031e-01
-5.14239669e-01 -9.91535485e-01 -7.76519954e-01 6.22367084e-01
-1.74153686e+00 -1.07768285e+00 6.38424903e-02 1.03559040e-01
6.82760060e-01 5.89498170e-02 1.27445161e+00 3.41910839e-01
-6.56239867e-01 3.76526564e-01 -5.04871428e-01 4.20580842e-02
9.03107405e-01 -1.43317103e+00 2.96572953e-01 7.34851360e-01
2.81581599e-02 8.19639385e-01 6.58350170e-01 -4.93030459e-01
-1.08266652e+00 -9.21667159e-01 1.77301347e+00 -3.00456315e-01
9.25954938e-01 -6.41377509e-01 -7.55163968e-01 4.72897708e-01
1.13410032e+00 -1.91077083e-01 1.25767052e+00 2.76966393e-01
-9.89235789e-02 2.44122848e-01 -1.32910967e+00 5.15584350e-01
4.43299502e-01 -5.48522711e-01 -1.01692724e+00 1.23699449e-01
7.86929369e-01 -1.91372693e-01 -1.05960441e+00 1.86340228e-01
4.72274780e-01 -5.35949707e-01 3.74684751e-01 -6.19313419e-01
1.21569300e+00 6.34396300e-02 1.15681097e-01 -1.56553483e+00
-4.71963584e-02 -3.20926875e-01 -1.46104693e-01 1.76199603e+00
6.47584498e-01 -1.60373122e-01 4.36862200e-01 9.40991044e-01
-2.51712024e-01 -1.24544585e+00 -6.65234447e-01 -1.20204888e-01
2.94853628e-01 -1.00945365e+00 3.46789151e-01 1.09287524e+00
6.20433092e-01 6.03110611e-01 -2.71586984e-01 -4.74256158e-01
-1.88870728e-01 -1.23854391e-01 3.72360915e-01 -9.93078530e-01
6.48517385e-02 -4.80809063e-01 8.38755295e-02 -3.05482060e-01
7.08803892e-01 -8.42166245e-01 2.62350529e-01 -2.00670910e+00
5.06798252e-02 5.85660115e-02 -3.57957065e-01 7.28655219e-01
-1.31280720e-01 1.21494599e-01 8.70837942e-02 -6.24174237e-01
-6.31101668e-01 5.36311269e-01 5.88525593e-01 1.16914563e-01
-2.08828986e-01 -7.08170950e-01 -8.51410151e-01 8.51901293e-01
9.27441061e-01 -6.18461132e-01 -3.59418005e-01 -2.75790811e-01
9.13657665e-01 -7.60943949e-01 5.46573177e-02 -4.13211495e-01
-1.54689522e-02 2.58215010e-01 4.30832893e-01 -6.09604776e-01
2.48218521e-01 -1.13086915e+00 -3.96076977e-01 -9.52164605e-02
-7.08878219e-01 3.22444588e-01 5.82082033e-01 -2.00510308e-01
-6.21078014e-01 -8.16807389e-01 3.21653157e-01 -1.01057440e-01
-7.77600050e-01 -2.92109430e-01 -7.61680305e-01 -2.65673734e-02
1.00323045e+00 1.01458184e-01 -1.46719858e-01 -3.68892103e-01
-5.07671833e-01 2.32792810e-01 1.21477388e-01 8.30799043e-01
7.44363964e-01 -1.21161819e+00 -4.39168572e-01 -1.18014917e-01
1.55050278e-01 -2.75174379e-01 -1.59701481e-02 4.29634959e-01
-9.29295570e-02 5.60380101e-01 -1.62271038e-01 2.42405713e-01
-1.43280518e+00 6.38373256e-01 1.03693485e-01 -6.23849928e-01
-1.42661557e-01 8.35902870e-01 -4.91296649e-01 -7.51004338e-01
1.82928026e-01 -6.65892065e-01 -1.08219099e+00 5.53066492e-01
5.73395193e-01 1.46416416e-02 3.59299839e-01 -6.03636563e-01
-5.71514189e-01 4.98766482e-01 -1.72903717e-01 -1.46435216e-01
1.73105276e+00 6.34924769e-02 -5.82154930e-01 7.20228314e-01
1.71256959e+00 5.69038577e-02 -3.47024560e-01 2.79769242e-01
3.26135576e-01 2.27973938e-01 2.24567086e-01 -1.01447153e+00
-1.00363898e+00 1.03633559e+00 4.02383685e-01 2.13383377e-01
1.28079093e+00 -1.11562751e-01 9.18588519e-01 3.83947819e-01
-3.24263871e-01 -1.68930197e+00 -2.61717319e-01 8.41156662e-01
9.38164055e-01 -1.15320826e+00 2.38063738e-01 -2.64227837e-01
-8.08586419e-01 1.50956368e+00 6.65240526e-01 -2.47261286e-01
7.73269176e-01 6.83867574e-01 3.16998839e-01 -5.14582217e-01
-1.15232718e+00 -7.81872403e-03 3.47567916e-01 4.34532017e-01
1.16625094e+00 -1.78491920e-01 -8.83044958e-01 1.22740316e+00
-3.25361609e-01 3.49255413e-01 5.93757570e-01 1.02923083e+00
-1.85680818e-02 -1.42661989e+00 7.81199485e-02 3.68338078e-01
-9.55773354e-01 -4.25030887e-01 -1.19399369e+00 9.29937899e-01
4.94295150e-01 1.12601316e+00 1.97789237e-01 -2.50696331e-01
3.37613910e-01 3.72446835e-01 2.08387710e-02 -5.86210072e-01
-1.10885179e+00 -5.49672022e-02 6.87268674e-01 -4.58944678e-01
-1.01397717e+00 -4.67026055e-01 -1.53054321e+00 2.68629998e-01
-2.71396637e-01 3.29709679e-01 1.03812468e+00 1.23743010e+00
6.11959994e-01 7.57621527e-01 3.60460967e-01 -8.32348943e-01
4.04132307e-01 -1.16104472e+00 -3.00581604e-01 3.10132772e-01
9.80669186e-02 -3.80563945e-01 -4.82018322e-01 9.67661142e-02] | [12.694828033447266, 6.2060699462890625] |
d409ce8f-9a36-48eb-a3c9-8c882277417b | iseeu2-visually-interpretable-icu-mortality | 2005.09284 | null | https://arxiv.org/abs/2005.09284v1 | https://arxiv.org/pdf/2005.09284v1.pdf | ISeeU2: Visually Interpretable ICU mortality prediction using deep learning and free-text medical notes | Accurate mortality prediction allows Intensive Care Units (ICUs) to adequately benchmark clinical practice and identify patients with unexpected outcomes. Traditionally, simple statistical models have been used to assess patient death risk, many times with sub-optimal performance. On the other hand deep learning holds promise to positively impact clinical practice by leveraging medical data to assist diagnosis and prediction, including mortality prediction. However, as the question of whether powerful Deep Learning models attend correlations backed by sound medical knowledge when generating predictions remains open, additional interpretability tools are needed to foster trust and encourage the use of AI by clinicians. In this work we show a Deep Learning model trained on MIMIC-III to predict mortality using raw nursing notes, together with visual explanations for word importance. Our model reaches a ROC of 0.8629 (+/-0.0058), outperforming the traditional SAPS-II score and providing enhanced interpretability when compared with similar Deep Learning approaches. | ['William Caicedo-Torres', 'Jairo Gutierrez'] | 2020-05-19 | null | null | null | null | ['icu-mortality'] | ['medical'] | [-1.50919169e-01 4.15058911e-01 -5.41086681e-02 -3.78613651e-01
-5.98744869e-01 -1.77642301e-01 9.36288312e-02 8.14236283e-01
-2.66930759e-01 8.54495466e-01 7.66702116e-01 -8.26965034e-01
-3.00429851e-01 -5.94013035e-01 -1.79879963e-02 -4.06793773e-01
-3.27071130e-01 8.22270274e-01 -5.31668305e-01 5.77423647e-02
1.50035441e-01 4.27777499e-01 -8.35050404e-01 7.36999810e-01
7.99682319e-01 8.68715167e-01 -2.68943757e-01 8.27877581e-01
-3.20414044e-02 1.48434317e+00 -5.90287209e-01 -4.88846630e-01
-5.03854193e-02 -6.54011309e-01 -6.47802413e-01 -6.11583054e-01
7.51426667e-02 -7.50814795e-01 -3.19049269e-01 2.41425961e-01
7.81002402e-01 -3.52526575e-01 9.81147528e-01 -8.97137523e-01
-6.77548528e-01 6.52470589e-01 -1.64776236e-01 4.15051311e-01
1.81103677e-01 4.28252399e-01 9.25665855e-01 -6.57107770e-01
7.21814483e-02 9.91068602e-01 1.17796731e+00 6.96908116e-01
-1.19826639e+00 -6.02784574e-01 -2.58142650e-01 3.86512518e-01
-9.30173039e-01 -1.97754949e-01 2.75125831e-01 -6.05300367e-01
1.02231634e+00 2.09122866e-01 6.81027770e-01 1.32703662e+00
5.05848110e-01 3.63897234e-01 9.76337790e-01 -1.57829270e-01
1.28521353e-01 3.56247579e-03 7.53448606e-02 8.04304242e-01
5.94705760e-01 9.48680043e-02 -3.48241061e-01 -3.09664905e-01
7.19690979e-01 6.65535629e-01 -3.56232882e-01 5.46941273e-02
-1.41042256e+00 9.26765025e-01 5.97143769e-01 6.75331950e-02
-6.46764755e-01 1.20356411e-01 5.70028663e-01 8.45055431e-02
2.21552417e-01 6.69058681e-01 -7.35728443e-01 -3.76519769e-01
-9.37687278e-01 4.47770171e-02 8.25616956e-01 1.17518157e-01
-7.59360045e-02 1.61475375e-01 -3.88864934e-01 6.22666836e-01
5.25300391e-02 3.76837403e-01 5.07728517e-01 -9.24424291e-01
2.34041840e-01 7.59598434e-01 9.63252112e-02 -7.77734041e-01
-8.37410510e-01 -6.84531271e-01 -1.30388677e+00 4.67779487e-01
5.38454056e-01 -3.49488914e-01 -6.20744228e-01 1.27199221e+00
-3.60433549e-01 -5.57443909e-02 2.06907496e-01 7.30157137e-01
9.46439862e-01 1.65394858e-01 6.18442178e-01 -1.14795286e-03
1.30617964e+00 -4.68105435e-01 -5.77654898e-01 -1.09778583e-01
1.18153906e+00 -3.13374013e-01 1.02063465e+00 5.65286517e-01
-8.75891924e-01 -1.72943309e-01 -6.64402187e-01 7.75918886e-02
4.69153188e-03 1.05102219e-01 3.74380171e-01 3.15997392e-01
-9.92984951e-01 9.25075352e-01 -8.28392029e-01 -3.82549137e-01
8.77931476e-01 3.62270802e-01 -3.93357813e-01 1.25336391e-03
-1.06342602e+00 1.25044954e+00 3.68999094e-01 -1.65412605e-01
-6.57790244e-01 -8.91838491e-01 -5.87407827e-01 3.06232303e-01
-9.98174846e-02 -1.51604712e+00 1.06989574e+00 -7.92440474e-01
-9.04490530e-01 6.96976662e-01 6.40001521e-02 -8.77733946e-01
7.06434608e-01 -5.80241084e-01 -2.31378660e-01 3.46835077e-01
-2.31996432e-01 5.16236961e-01 3.45066607e-01 -1.03143644e+00
-4.42180902e-01 -4.56179887e-01 -7.25302249e-02 -4.25266363e-02
-3.94596696e-01 -6.58387877e-03 4.63438690e-01 -7.51459599e-01
-4.39890653e-01 -5.98055422e-01 -4.62657392e-01 1.97930261e-01
-3.31643671e-01 4.77839168e-03 4.38190132e-01 -1.03630435e+00
1.55475330e+00 -1.93514884e+00 -9.76388603e-02 -6.01609573e-02
8.94881546e-01 3.25594515e-01 2.04392478e-01 3.92471313e-01
-3.20458144e-01 2.66368598e-01 -3.62468451e-01 -2.45027274e-01
-3.10716480e-01 1.46428138e-01 -2.44655490e-01 2.31494382e-01
4.97059792e-01 1.16770840e+00 -1.03510106e+00 -6.93642557e-01
4.73575950e-01 5.98525047e-01 -6.92671120e-01 6.36785567e-01
2.62857169e-01 5.13101697e-01 -2.31566504e-01 3.69647384e-01
-3.70964818e-02 -8.14685524e-01 3.82131487e-01 1.23302899e-01
3.23316783e-01 2.12075397e-01 -2.73528785e-01 9.11486983e-01
-3.83357435e-01 6.67838335e-01 -2.87287533e-01 -9.33626354e-01
7.86001265e-01 4.15291220e-01 4.88872409e-01 -3.16178411e-01
2.46380970e-01 7.00058863e-02 2.97550976e-01 -7.41668522e-01
-3.28632802e-01 -7.96777844e-01 2.74427116e-01 4.09931093e-01
-3.58121842e-01 2.21884683e-01 -4.21886414e-01 1.99337289e-01
1.37653875e+00 -3.62812206e-02 8.20935190e-01 -1.51782766e-01
2.05621541e-01 4.60154489e-02 3.10528010e-01 6.84299409e-01
-2.60646641e-01 7.94465303e-01 6.89724982e-01 -1.18210185e+00
-8.65905225e-01 -1.03457701e+00 -2.90676951e-01 6.63581610e-01
-4.95892584e-01 -3.42954427e-01 -6.77647471e-01 -9.02361751e-01
1.31593198e-01 1.26414216e+00 -9.59557295e-01 -4.68126088e-01
-3.79082292e-01 -8.74479115e-01 6.37014449e-01 1.08455932e+00
-3.47033441e-02 -1.16584051e+00 -1.18949485e+00 3.89182776e-01
-2.28980273e-01 -5.97141564e-01 1.27866283e-01 3.57329279e-01
-1.28288722e+00 -1.17186975e+00 -8.60977113e-01 -2.65986055e-01
3.86069685e-01 -1.88093051e-01 1.59747732e+00 4.99351412e-01
-5.16104519e-01 2.42513090e-01 -2.84291029e-01 -6.36213779e-01
-6.45505309e-01 -8.16030726e-02 1.31971031e-01 -5.12389660e-01
6.57164037e-01 -5.88226318e-01 -1.20121086e+00 -4.79194298e-02
-7.88641691e-01 3.14064384e-01 8.82921636e-01 1.04605627e+00
6.35470152e-02 -6.37564838e-01 8.41674149e-01 -7.15840042e-01
6.02358580e-01 -8.66117001e-01 2.47571290e-01 6.22128025e-02
-1.16390264e+00 1.65968806e-01 1.02426732e+00 -5.35173416e-02
-6.00054324e-01 -3.72187734e-01 3.11198719e-02 -3.10564637e-01
-4.71874267e-01 4.66765314e-01 4.73077625e-01 5.05447626e-01
8.20044994e-01 -1.45597070e-01 1.76025033e-01 -5.73290110e-01
-4.60081287e-02 6.22778237e-01 4.35594887e-01 -1.21924460e-01
4.06568199e-01 2.29144529e-01 2.18107149e-01 -3.61484170e-01
-1.16818845e+00 -9.88259837e-02 -6.22935355e-01 5.05918153e-02
1.09973347e+00 -8.47867548e-01 -9.26814139e-01 -1.47243172e-01
-1.01096535e+00 -4.57670927e-01 -2.64578074e-01 5.31654656e-01
-4.11345720e-01 3.75526100e-01 -6.44702315e-01 -6.98542118e-01
-6.39623225e-01 -7.55162299e-01 8.28402221e-01 -1.86631307e-01
-1.07167661e+00 -1.28972828e+00 1.91012084e-01 4.15931433e-01
4.45447773e-01 6.28867090e-01 1.43697286e+00 -1.05605757e+00
-2.10387968e-02 -2.38595381e-01 -4.81724560e-01 2.79798239e-01
2.14577973e-01 -1.85238004e-01 -9.31737959e-01 7.21068531e-02
-2.77668744e-01 -3.64462107e-01 8.64097416e-01 5.36457717e-01
1.59277940e+00 -6.06666982e-01 -2.02688545e-01 6.65994525e-01
1.12531495e+00 2.17281938e-01 4.98087138e-01 4.50969100e-01
7.71963358e-01 6.62492394e-01 -3.22185121e-02 8.31631005e-01
6.17111146e-01 1.61206111e-01 4.40520763e-01 -3.81539673e-01
1.96776073e-02 -2.43043751e-02 7.09921718e-02 4.82628942e-01
-2.93568462e-01 -1.88684359e-01 -1.50609386e+00 1.62817031e-01
-1.69799352e+00 -8.95041049e-01 -3.51960599e-01 2.05113101e+00
6.51039720e-01 2.51373112e-01 8.85056481e-02 1.67644233e-01
7.79651105e-02 -3.03540230e-01 -5.61392546e-01 -5.83071351e-01
-8.45018923e-02 2.03716919e-01 1.23698555e-01 2.44437024e-01
-7.52324164e-01 3.23905468e-01 6.76603603e+00 1.89300906e-02
-9.70109820e-01 -2.12268025e-01 1.30923200e+00 -3.01310241e-01
-1.79493234e-01 -4.48493689e-01 -7.39432573e-02 3.67699921e-01
1.37251973e+00 -7.10642189e-02 1.40366986e-01 6.74052119e-01
5.60252070e-01 1.88904211e-01 -1.39119852e+00 9.87113059e-01
-2.00542491e-02 -1.26711333e+00 2.95646340e-01 -1.09613873e-02
3.57980430e-01 -2.16665447e-01 1.19346902e-01 2.57674128e-01
4.42167014e-01 -1.68616402e+00 3.65251869e-01 9.13408697e-01
1.03870273e+00 -6.36451960e-01 1.23764515e+00 6.09459840e-02
-5.07572353e-01 -4.13283646e-01 3.58034484e-02 -3.49619448e-01
-3.30868512e-02 4.73715574e-01 -1.45549858e+00 1.85494035e-01
7.30532825e-01 6.05897903e-01 -5.57249367e-01 9.12999451e-01
3.36458571e-02 9.91797209e-01 2.29387701e-01 1.15245163e-01
2.10012376e-01 2.18106881e-01 9.45824087e-02 1.40029061e+00
5.04758477e-01 4.05357122e-01 -2.20194280e-01 6.07955515e-01
4.69966754e-02 2.83780634e-01 -6.77695155e-01 -1.37408152e-01
-4.39715832e-02 1.03938019e+00 -3.76213163e-01 -6.90591455e-01
-4.78626370e-01 8.31455886e-01 3.36324483e-01 1.84895128e-01
-6.11041903e-01 -1.19133778e-01 6.88811362e-01 5.61235130e-01
-1.84302524e-01 -2.22082902e-03 -1.08079529e+00 -9.95907903e-01
-5.99766076e-01 -9.22427833e-01 8.24303210e-01 -9.06734586e-01
-1.34904122e+00 7.17957497e-01 -3.34110022e-01 -1.22755349e+00
-4.87973213e-01 -9.29830313e-01 -6.34842634e-01 9.21881437e-01
-1.31797516e+00 -7.71912396e-01 -5.87474883e-01 1.74563900e-01
3.51504266e-01 -2.07816109e-01 1.32796025e+00 -1.65935755e-01
-5.56462109e-01 4.26233500e-01 1.43049687e-01 2.97272980e-01
9.47154939e-01 -1.56575394e+00 4.31698896e-02 1.07306935e-01
-3.90833229e-01 6.13208592e-01 7.56682456e-01 -6.14869714e-01
-7.53361344e-01 -1.18826914e+00 9.63204741e-01 -1.06183493e+00
5.15713513e-01 4.15527254e-01 -1.15799773e+00 5.02938986e-01
8.81937146e-02 -2.11800665e-01 1.48880064e+00 -2.93896720e-02
-3.48872662e-01 3.43636051e-02 -1.01149833e+00 7.26848245e-01
7.15021253e-01 -1.85448170e-01 -9.51869369e-01 5.10600545e-02
4.36182767e-01 1.90956995e-01 -1.01599014e+00 5.78033149e-01
8.00081909e-01 -1.20202482e+00 1.05057704e+00 -1.34335017e+00
1.24467993e+00 1.65273488e-01 1.34753153e-01 -1.24581552e+00
-6.48641348e-01 -2.44864777e-01 -3.86005193e-01 3.70165825e-01
4.39542323e-01 -4.42815602e-01 6.17587745e-01 8.09644938e-01
-1.90058108e-02 -1.12469864e+00 -4.60469782e-01 -1.16458610e-01
3.36070240e-01 -3.90248358e-01 4.89910573e-01 1.05571711e+00
2.27818385e-01 3.65958959e-01 -3.67567122e-01 -7.00666979e-02
6.16320252e-01 -1.61790445e-01 3.53687823e-01 -1.60510480e+00
-2.89052874e-01 -8.53372216e-01 -3.27694178e-01 -2.26148844e-01
-7.83834830e-02 -8.50923896e-01 -4.33515459e-01 -2.12442088e+00
5.37780762e-01 -1.83251232e-01 -9.66242433e-01 7.61044860e-01
-7.05521047e-01 2.47400329e-01 -5.34194708e-02 1.59872949e-01
-3.70741725e-01 3.31649452e-01 1.00005054e+00 1.42833322e-01
-5.00835106e-02 -1.68229491e-01 -1.09514856e+00 9.68920648e-01
1.06957924e+00 -3.73116463e-01 -7.26920441e-02 -3.34884197e-01
1.51061326e-01 3.72935683e-01 5.64344287e-01 -9.41161215e-01
-3.50229681e-01 -2.92768210e-01 9.85314310e-01 -2.49362603e-01
-1.21725192e-02 -8.14617097e-01 2.41969861e-02 1.00972021e+00
-7.19188511e-01 4.57267910e-01 2.77132362e-01 4.45696533e-01
1.68339834e-02 1.59993216e-01 6.28823996e-01 -2.85159856e-01
-2.56623477e-01 2.46427208e-01 -5.93745947e-01 2.39006341e-01
9.32450294e-01 -7.65647665e-02 -1.68538511e-01 -6.53418124e-01
-8.48982453e-01 1.38574898e-01 2.91674376e-01 1.20623097e-01
7.83472776e-01 -1.03498232e+00 -1.02496564e+00 -6.38750196e-02
3.45404863e-01 -1.92371726e-01 3.38385880e-01 1.06134760e+00
-1.02236986e+00 4.46476191e-01 -2.90608764e-01 -2.35919610e-01
-1.14973629e+00 5.82316577e-01 3.55455548e-01 -3.36747974e-01
-1.02904999e+00 6.21613145e-01 5.39330900e-01 2.38947779e-01
3.37365657e-01 -3.88406903e-01 -4.17699516e-01 -6.14990033e-02
8.68139148e-01 5.68080723e-01 -1.67497352e-01 -2.03014612e-01
-4.31455404e-01 6.45938441e-02 2.02210724e-01 2.98989594e-01
1.55587900e+00 1.55212238e-01 1.53972926e-02 4.85806048e-01
8.05689573e-01 -3.12805086e-01 -9.66735184e-01 1.81987882e-01
2.28535384e-01 -2.19220698e-01 -5.75414449e-02 -1.44634688e+00
-6.22359574e-01 1.58237755e+00 6.08801126e-01 5.50292172e-02
1.04816628e+00 -2.02596843e-01 7.06037104e-01 4.15891379e-01
-1.82597876e-01 -4.86815155e-01 2.50340283e-01 1.97259739e-01
8.89412999e-01 -1.46837258e+00 -2.71556675e-01 3.84486914e-01
-1.02589631e+00 1.60460782e+00 3.85833830e-01 1.15950122e-01
4.29076672e-01 2.74550438e-01 4.69452143e-01 -5.17013334e-02
-9.79491055e-01 2.65321910e-01 3.74789834e-01 5.26109457e-01
9.56802845e-01 5.13329089e-01 -5.12808487e-02 8.87245417e-01
-1.69434056e-01 3.42905641e-01 3.56581748e-01 4.86531079e-01
-4.84377414e-01 -7.73584783e-01 -3.79535645e-01 1.12088013e+00
-1.01798165e+00 -5.46960950e-01 -4.09449369e-01 4.12846446e-01
-1.59318298e-02 7.98034847e-01 -3.71829048e-02 -3.90798271e-01
1.38940021e-01 5.07736266e-01 9.84709710e-02 -5.50296664e-01
-6.91048980e-01 -4.40285444e-01 2.76313126e-01 -3.95125568e-01
1.63615242e-01 -2.97959268e-01 -1.21184576e+00 -3.44573319e-01
4.13641930e-01 8.23020712e-02 1.11049451e-01 7.53156781e-01
5.87457120e-01 8.85741293e-01 2.53533483e-01 -3.28771174e-01
-8.81581187e-01 -1.00397921e+00 -2.48702362e-01 3.95714462e-01
7.12908804e-01 -4.07398671e-01 -4.45146888e-01 1.86018318e-01] | [8.085128784179688, 6.0986762046813965] |
d70f9bf0-6305-406e-9276-e2fb4e574b75 | dcil-deep-contextual-internal-learning-for | 1912.04229 | null | https://arxiv.org/abs/1912.04229v1 | https://arxiv.org/pdf/1912.04229v1.pdf | DCIL: Deep Contextual Internal Learning for Image Restoration and Image Retargeting | Recently, there is a vast interest in developing methods which are independent of the training samples such as deep image prior, zero-shot learning, and internal learning. The methods above are based on the common goal of maximizing image features learning from a single image despite inherent technical diversity. In this work, we bridge the gap between the various unsupervised approaches above and propose a general framework for image restoration and image retargeting. We use contextual feature learning and internal learning to improvise the structure similarity between the source and the target images. We perform image resize application in the following setups: classical image resize using super-resolution, a challenging image resize where the low-resolution image contains noise, and content-aware image resize using image retargeting. We also provide comparisons to the relevant state-of-the-art methods. | ['Shanmuganathan Raman', 'Indra Deep Mastan'] | 2019-12-09 | null | null | null | null | ['image-retargeting'] | ['computer-vision'] | [ 8.68606985e-01 -4.44028936e-02 -1.35691449e-01 -1.11234456e-01
-9.71963465e-01 -2.18754664e-01 7.15763807e-01 -1.99558198e-01
-3.28224033e-01 6.27873778e-01 5.27753234e-01 2.13804826e-01
-2.13993028e-01 -7.03506768e-01 -8.37701559e-01 -8.50662589e-01
4.99340177e-01 -1.43484131e-01 2.95236588e-01 -2.79388785e-01
6.78586900e-01 3.60442549e-01 -1.98122633e+00 2.85904497e-01
7.32939363e-01 7.29182720e-01 6.11917377e-01 7.45883763e-01
-2.02318460e-01 8.37874472e-01 -2.87407219e-01 2.05417633e-01
4.23390418e-01 -5.53520739e-01 -1.05833137e+00 6.46145582e-01
6.29047990e-01 -2.51800418e-01 -4.38674927e-01 1.30387533e+00
7.73270726e-01 3.81852090e-01 5.44231534e-01 -8.05809379e-01
-1.21043384e+00 4.06077981e-01 -6.37765050e-01 5.54764211e-01
4.63009506e-01 6.72874451e-02 6.58355117e-01 -1.08552003e+00
8.82884979e-01 1.10219729e+00 4.61862266e-01 6.27896547e-01
-1.66705680e+00 -3.50139350e-01 -3.76614928e-02 5.53229392e-01
-1.27285922e+00 -6.95945740e-01 7.48323679e-01 -6.71189249e-01
7.39590585e-01 1.42891929e-01 2.60792166e-01 1.01225376e+00
1.99962661e-01 4.27951634e-01 1.46016228e+00 -9.09991622e-01
3.73062342e-01 1.03638716e-01 -1.85415506e-01 3.36912036e-01
-1.83328405e-01 4.59483653e-01 -5.05236566e-01 4.78945561e-02
1.01371229e+00 1.78125098e-01 -5.12889922e-01 -5.23463786e-01
-1.09979534e+00 7.26939440e-01 4.34642345e-01 3.71105313e-01
-3.71634394e-01 -8.97953510e-02 1.03961244e-01 2.58481503e-01
4.08260971e-01 6.62040472e-01 -3.84209156e-01 3.75321627e-01
-1.10320508e+00 4.59241383e-02 2.29891792e-01 7.96024561e-01
1.26510882e+00 3.21853250e-01 -2.51061320e-01 9.77843523e-01
-6.32816479e-02 1.35283142e-01 8.94162357e-01 -1.21574426e+00
-2.31298938e-01 1.12357073e-01 1.52336851e-01 -7.89448857e-01
1.05491728e-02 -2.56167203e-01 -8.00078988e-01 5.94811440e-01
-7.78834000e-02 2.25818366e-01 -1.32591164e+00 1.49958479e+00
3.64419281e-01 6.95616722e-01 3.25469404e-01 7.17451513e-01
9.93974388e-01 5.16706765e-01 -6.06167465e-02 -4.47815299e-01
1.10785198e+00 -1.02183890e+00 -8.08859885e-01 -1.63426116e-01
1.16704531e-01 -1.08963382e+00 1.25677788e+00 3.41718793e-02
-1.17244875e+00 -1.00844812e+00 -1.03386831e+00 -1.01043701e-01
-3.46369088e-01 -2.17230394e-01 1.29353017e-01 5.43584347e-01
-1.23496366e+00 8.48586440e-01 -3.85637552e-01 -5.59818864e-01
4.49474037e-01 1.47425428e-01 -5.19810975e-01 -3.07101637e-01
-1.03269994e+00 8.32226932e-01 5.80683589e-01 -3.86192769e-01
-1.11145270e+00 -1.02723610e+00 -9.10203636e-01 -1.47948533e-01
3.71183723e-01 -7.71882772e-01 8.61588597e-01 -1.34620202e+00
-1.72315907e+00 1.16690457e+00 -8.76156837e-02 -3.25159073e-01
2.77522743e-01 -1.09522134e-01 -3.78000915e-01 1.26978457e-01
1.82511553e-01 6.07338130e-01 1.50492823e+00 -1.76349092e+00
-5.70873976e-01 -1.13968365e-01 -8.06641057e-02 2.18760386e-01
-7.18075559e-02 -5.62723391e-02 -3.00408959e-01 -8.07803929e-01
4.67307307e-02 -7.67166853e-01 -4.65765655e-01 -3.33224982e-01
-3.97205740e-01 4.03783500e-01 8.56191576e-01 -5.22742450e-01
8.66325915e-01 -2.14483213e+00 3.09624970e-01 -1.83887795e-01
7.13873506e-02 2.89860904e-01 -5.04544258e-01 3.19401860e-01
-5.30017495e-01 1.60915822e-01 -2.95721352e-01 -1.14305697e-01
-5.12817979e-01 2.21734777e-01 -4.59810138e-01 6.26132190e-01
-7.51787424e-03 8.73476803e-01 -8.47014248e-01 -5.33400297e-01
6.89683318e-01 5.24684191e-01 -4.75992411e-01 3.17955017e-01
-1.55344037e-02 8.58783185e-01 -6.23763986e-02 4.49630022e-01
8.19291770e-01 -3.96901339e-01 -2.75042690e-02 -6.32293820e-01
-3.64095330e-01 -3.35030675e-01 -1.17137098e+00 1.75725627e+00
-3.80562752e-01 4.81655002e-01 -1.71668470e-01 -1.05096805e+00
6.58666134e-01 1.98649049e-01 6.43924654e-01 -8.27644527e-01
5.04729263e-02 -9.83223692e-02 -4.13475782e-01 -6.96909189e-01
6.00131273e-01 -3.21825266e-01 2.78467268e-01 4.26503658e-01
4.52594936e-01 -1.78954944e-01 4.96311896e-02 1.26268014e-01
8.15738261e-01 1.96411178e-01 6.63124025e-01 -3.11855435e-01
5.32980978e-01 -1.82908505e-01 4.02343869e-01 8.96543086e-01
-1.43852949e-01 1.28526568e+00 -1.61753014e-01 -3.89822662e-01
-1.48001766e+00 -1.27217638e+00 -3.12557846e-01 1.04929030e+00
5.31966507e-01 -3.03285152e-01 -7.25836396e-01 -3.23865235e-01
-4.81367946e-01 6.02629125e-01 -9.22919869e-01 -1.75522625e-01
-3.47632468e-01 -6.90092027e-01 -1.79033369e-01 2.18995735e-01
5.50380349e-01 -1.06237841e+00 -4.06927735e-01 4.80281152e-02
-4.13743407e-01 -1.19689465e+00 -4.32283401e-01 -1.01330010e-02
-7.47578382e-01 -1.01673710e+00 -8.55811536e-01 -8.65353227e-01
7.59057939e-01 7.62086391e-01 1.13790810e+00 4.34314646e-02
-5.50668597e-01 8.49603176e-01 -6.30689085e-01 1.79658115e-01
-5.82949996e-01 -2.13655829e-01 7.13936538e-02 3.34150702e-01
-9.55357775e-02 -9.38147962e-01 -8.24674189e-01 1.90049812e-01
-1.40546715e+00 2.17096820e-01 7.11348414e-01 9.55014884e-01
1.03112411e+00 2.67110080e-01 3.91124606e-01 -9.76999760e-01
2.68162757e-01 -3.80583286e-01 -2.72390395e-01 3.46964598e-01
-6.59448564e-01 1.96071625e-01 4.33053881e-01 -5.40154755e-01
-1.33770561e+00 3.22243601e-01 -7.87405670e-02 -5.96982360e-01
-2.87182778e-01 1.06956325e-01 -1.45867348e-01 -5.70564210e-01
1.17543185e+00 6.54134989e-01 6.28780648e-02 -5.62216759e-01
8.79272699e-01 3.84080678e-01 8.31753731e-01 -2.88082749e-01
9.94949877e-01 7.83089101e-01 -1.18230335e-01 -1.05363834e+00
-1.07412636e+00 -4.20501053e-01 -8.45472634e-01 -1.46399930e-01
8.98674786e-01 -9.60142791e-01 -2.88975239e-02 3.49189013e-01
-7.77001917e-01 -3.68097514e-01 -9.04984117e-01 2.34383300e-01
-1.02213931e+00 6.68432534e-01 -2.82124281e-01 -2.33544067e-01
-1.75079659e-01 -1.27132785e+00 8.94504189e-01 5.58273852e-01
1.04386561e-01 -8.79863143e-01 2.39444554e-01 2.39522845e-01
6.38839602e-01 2.69113332e-01 7.79411912e-01 -2.39028990e-01
-6.38253033e-01 1.76001176e-01 -3.94242674e-01 5.04928589e-01
4.80609506e-01 -2.88063020e-01 -1.16194236e+00 -4.57852840e-01
5.01442552e-02 -3.97452950e-01 9.99516487e-01 7.58891463e-01
1.03890979e+00 -1.79391548e-01 -1.54705808e-01 8.05361569e-01
1.86481285e+00 -1.23292454e-01 1.28816640e+00 6.15770400e-01
5.74322164e-01 5.20311773e-01 4.94338989e-01 4.25975621e-01
2.51505766e-02 6.06610596e-01 4.28018928e-01 -2.86185592e-01
-6.71170413e-01 -1.65266946e-01 1.82505563e-01 4.17281747e-01
-7.02250227e-02 1.43407837e-01 -3.97513330e-01 7.48876870e-01
-1.78511763e+00 -1.24586844e+00 2.88469493e-01 2.25247574e+00
9.67337489e-01 -3.62476259e-01 -2.41686180e-01 -1.53421953e-01
1.00830340e+00 3.13744187e-01 -6.07128203e-01 1.79266512e-01
-3.23409677e-01 1.91068470e-01 4.77509707e-01 6.91174388e-01
-1.26125646e+00 1.23181474e+00 7.41622019e+00 1.03945208e+00
-1.03858864e+00 2.70829350e-01 7.39621401e-01 9.89167318e-02
-2.42295623e-01 2.57204980e-01 -6.93979979e-01 2.49584377e-01
6.24612451e-01 -3.77465189e-01 6.39063001e-01 7.26698518e-01
2.79853344e-02 -2.05451548e-01 -8.42179298e-01 1.17074776e+00
4.38361675e-01 -1.65678811e+00 1.44842759e-01 -8.95430297e-02
1.30956602e+00 2.50195200e-03 2.63832450e-01 2.86618769e-02
3.75612974e-01 -9.59599495e-01 4.45788175e-01 7.54018307e-01
9.69526410e-01 -4.37328309e-01 3.42157483e-01 -1.47910258e-02
-1.02620220e+00 -1.05110325e-01 -6.05940640e-01 2.09061563e-01
2.14863956e-01 5.04889846e-01 -1.96115136e-01 4.91468459e-01
8.80518258e-01 7.99857736e-01 -7.09159076e-01 9.52483237e-01
-1.21006973e-01 9.73829031e-02 1.93562731e-01 9.05053794e-01
-2.82743722e-01 -8.91636908e-02 6.68920219e-01 9.26403999e-01
3.02042633e-01 1.38740942e-01 3.73969644e-01 8.07149291e-01
4.63922545e-02 1.65937796e-01 -5.94196796e-01 2.70139605e-01
1.57968894e-01 1.29132712e+00 -7.15958416e-01 -3.84594083e-01
-5.00420392e-01 1.29322970e+00 1.38871908e-01 5.37091434e-01
-4.24617738e-01 -1.34924740e-01 5.03328741e-01 2.19762340e-01
4.57364589e-01 8.81512314e-02 -7.39258528e-02 -1.35888672e+00
-4.74969894e-01 -8.18817139e-01 1.35094121e-01 -9.39719975e-01
-1.40967667e+00 7.11634696e-01 7.18565658e-02 -1.34767544e+00
-2.83179581e-01 -2.33768284e-01 -4.97862667e-01 7.79807806e-01
-1.82324111e+00 -1.10794389e+00 -5.20466745e-01 7.95325339e-01
8.63182783e-01 -3.48152757e-01 7.11979330e-01 3.16563882e-02
-2.21276298e-01 2.68834859e-01 3.86106521e-01 -2.34949663e-01
9.09421623e-01 -9.69975650e-01 2.68744051e-01 9.81771052e-01
8.77077058e-02 5.47896504e-01 1.04724383e+00 -6.78674400e-01
-1.19877851e+00 -1.19858849e+00 3.18778604e-01 -3.00852597e-01
4.08104599e-01 9.77436081e-03 -1.00169361e+00 6.12179399e-01
4.21676815e-01 3.96950126e-01 7.26316512e-01 -2.78890043e-01
-3.80537063e-01 7.87875280e-02 -1.20102525e+00 6.60068750e-01
8.47369015e-01 -5.55090308e-01 -6.08889818e-01 2.71133125e-01
7.10303307e-01 -1.49784163e-01 -7.86081612e-01 3.36907864e-01
3.31979722e-01 -1.18743145e+00 1.54528689e+00 -5.34931600e-01
5.64314663e-01 -3.78156215e-01 -4.44838583e-01 -1.42330396e+00
-8.31789970e-01 -5.40471792e-01 1.41839832e-01 1.22274375e+00
-6.15233369e-02 -3.83993804e-01 5.25130987e-01 4.72026110e-01
7.45672882e-02 -1.90770656e-01 -5.94691992e-01 -6.74071133e-01
-4.79143560e-02 4.33071554e-02 2.26353839e-01 1.01360428e+00
-4.23579633e-01 3.50722522e-01 -7.74060607e-01 1.13204211e-01
1.05938148e+00 2.80443311e-01 8.42140913e-01 -1.08384693e+00
-3.29512537e-01 -1.76122516e-01 -4.00942773e-01 -6.86073363e-01
2.84501046e-01 -7.47973084e-01 1.48686498e-01 -1.56261015e+00
6.06950223e-01 8.91348906e-03 -4.23149943e-01 2.60946631e-01
-1.84673980e-01 6.56988144e-01 6.25204593e-02 5.03039181e-01
-6.05947375e-01 6.66146815e-01 1.28009760e+00 -1.19681492e-01
-2.82646388e-01 -3.24323863e-01 -9.80767369e-01 7.53883719e-01
7.94444799e-01 -4.59547967e-01 -4.92555737e-01 -1.26700684e-01
-1.32819135e-02 -2.33007029e-01 3.77320230e-01 -9.64516461e-01
2.64914721e-01 -3.90951246e-01 5.11140406e-01 -3.05184841e-01
1.63861126e-01 -6.82673454e-01 1.61333740e-01 1.28109872e-01
-4.82473403e-01 -4.36774820e-01 4.42803763e-02 9.13736939e-01
-1.25762582e-01 -4.26668286e-01 1.54508281e+00 -5.42157471e-01
-1.27857924e+00 3.87223184e-01 -3.21924955e-01 1.13725349e-01
9.73442256e-01 -4.33946788e-01 -2.59483308e-01 -3.78551573e-01
-1.00965083e+00 -4.67380166e-01 7.30826020e-01 5.67682683e-01
9.11969662e-01 -1.28850079e+00 -7.87804246e-01 3.73665780e-01
2.38854140e-01 -4.36056167e-01 9.52633321e-01 5.05576909e-01
-2.44786501e-01 -1.49233922e-01 -6.04311407e-01 -5.53203523e-01
-1.23643410e+00 1.07716358e+00 3.05855900e-01 1.77791994e-02
-9.27208841e-01 5.97591877e-01 6.47710204e-01 -2.70042121e-01
-1.78510949e-01 5.55577815e-01 -3.74094099e-01 -1.97735235e-01
1.07713819e+00 3.39476228e-01 -2.09869206e-01 -9.77967918e-01
1.37259483e-01 9.96669829e-01 -3.92971158e-01 -1.77864671e-01
1.34276652e+00 -7.34202564e-01 -1.70385316e-01 2.68471837e-01
1.19897306e+00 -9.72271264e-02 -1.44772077e+00 -7.69195378e-01
-2.12367475e-01 -1.03464973e+00 4.35637534e-01 -3.95840943e-01
-1.09977245e+00 5.65224826e-01 1.22377694e+00 1.77760944e-01
1.44596362e+00 1.48149431e-01 4.12152082e-01 3.72789167e-02
1.81648493e-01 -1.18604600e+00 6.23578608e-01 3.20751846e-01
9.58609760e-01 -1.44168723e+00 3.43412638e-01 -2.76696175e-01
-5.67828953e-01 9.73220825e-01 4.56700981e-01 -4.98111725e-01
7.74668336e-01 2.18436480e-01 5.13320528e-02 6.61482364e-02
-4.13148135e-01 -6.78175986e-01 1.58106744e-01 1.03815901e+00
1.38946533e-01 -4.28002208e-01 1.45008471e-02 2.12138191e-01
1.28072992e-01 1.58209920e-01 7.71659613e-01 8.29013228e-01
-6.87471569e-01 -1.21337903e+00 -4.98040408e-01 1.18537791e-01
-5.76912165e-01 -3.30407560e-01 -1.43842138e-02 3.59196424e-01
-1.83530841e-02 8.38568568e-01 -1.35651976e-01 -4.11038458e-01
1.56831890e-01 -1.67546913e-01 7.32557595e-01 -6.86950386e-01
-9.03680921e-02 2.97372639e-01 -5.91374040e-01 -6.44340277e-01
-8.23041320e-01 -4.66526836e-01 -8.63426626e-01 2.31421571e-02
-3.06083381e-01 -3.13692421e-01 4.39647287e-01 9.64908421e-01
5.31933784e-01 5.37312150e-01 7.67119765e-01 -1.54066074e+00
-2.28907436e-01 -7.93170393e-01 -7.19893098e-01 6.13892555e-01
5.03036261e-01 -6.23111725e-01 -4.21113133e-01 6.65159822e-01] | [11.201550483703613, -1.9918873310089111] |
b3f4852c-cea5-4544-96f7-6760ac21bb36 | toward-face-biometric-de-identification-using | 2302.03657 | null | https://arxiv.org/abs/2302.03657v1 | https://arxiv.org/pdf/2302.03657v1.pdf | Toward Face Biometric De-identification using Adversarial Examples | The remarkable success of face recognition (FR) has endangered the privacy of internet users particularly in social media. Recently, researchers turned to use adversarial examples as a countermeasure. In this paper, we assess the effectiveness of using two widely known adversarial methods (BIM and ILLC) for de-identifying personal images. We discovered, unlike previous claims in the literature, that it is not easy to get a high protection success rate (suppressing identification rate) with imperceptible adversarial perturbation to the human visual system. Finally, we found out that the transferability of adversarial examples is highly affected by the training parameters of the network with which they are generated. | ['Raymond Veldhuis', 'Zohra Rezgui', 'Aythami Morales', 'Ruben Vera-Rodriguez', 'Luis Felipe Gomez', 'Julian Fierrez', 'Mahdi Ghafourian'] | 2023-02-07 | null | null | null | null | ['de-identification'] | ['natural-language-processing'] | [ 5.50812244e-01 4.21467960e-01 4.67050254e-01 -1.70751780e-01
-3.93071622e-01 -1.15158081e+00 8.31027448e-01 -4.28616464e-01
-3.12453657e-01 8.10245991e-01 -1.99633658e-01 -4.65067267e-01
-3.27219963e-02 -7.36517012e-01 -7.84771860e-01 -6.26663327e-01
-2.43921980e-01 -2.15675637e-01 -2.73933917e-01 -1.52305320e-01
2.64082402e-01 8.96020532e-01 -1.39140749e+00 2.35145450e-01
6.64757311e-01 1.00371063e+00 -7.15350926e-01 7.31139302e-01
3.23909670e-01 7.45245874e-01 -1.08722615e+00 -1.32668924e+00
6.95873141e-01 -3.54797423e-01 -7.62957454e-01 -2.35123411e-01
7.49482393e-01 -5.43631256e-01 -6.49352968e-01 1.49064243e+00
6.24617517e-01 -2.60874629e-01 6.47149503e-01 -1.42542958e+00
-1.12287593e+00 2.38464192e-01 -2.96175241e-01 1.48033500e-01
5.97270489e-01 2.44808257e-01 2.80959308e-01 -5.56531668e-01
4.45999980e-01 1.42164326e+00 7.41785586e-01 1.11666524e+00
-1.09930313e+00 -1.14460695e+00 -3.33053917e-01 -2.64702011e-02
-1.64790022e+00 -7.81886399e-01 6.46525681e-01 -3.32753420e-01
2.10582614e-01 6.53101981e-01 2.08290741e-01 1.59013903e+00
1.59467250e-01 1.42850280e-01 1.32516432e+00 -5.01852274e-01
4.91190888e-03 5.82594454e-01 -4.04879123e-01 5.56610763e-01
4.08180237e-01 7.21952736e-01 -1.56764954e-01 -6.33695304e-01
7.51416147e-01 -2.91858286e-01 -3.02902400e-01 3.85674797e-02
-5.50699949e-01 6.88883841e-01 3.10805649e-01 3.83778334e-01
-1.73607603e-01 1.16438180e-01 2.85001367e-01 6.16111219e-01
2.84756422e-01 6.22184336e-01 -1.19159445e-01 2.71816671e-01
-3.92110109e-01 6.48082718e-02 7.91015387e-01 7.17489541e-01
2.30629012e-01 1.89687639e-01 -8.59485865e-02 4.78264302e-01
1.14066169e-01 7.40989566e-01 3.18217188e-01 -9.31047976e-01
1.77155137e-01 1.95862949e-01 1.50551453e-01 -1.51914108e+00
2.55679637e-01 -1.61293164e-01 -8.63062263e-01 5.77145755e-01
6.77122772e-01 -4.10751730e-01 -7.29138434e-01 1.88877964e+00
1.29760161e-01 -7.36372964e-03 1.81208953e-01 4.50940371e-01
4.18159932e-01 2.76073903e-01 3.76021653e-01 -6.85339421e-02
1.02805555e+00 -1.38931394e-01 -7.12340772e-01 -9.81269181e-02
-4.86992523e-02 -9.52558041e-01 8.18026364e-01 4.13213104e-01
-8.91728699e-01 -5.52185416e-01 -1.09951556e+00 4.60147709e-01
-5.77031374e-01 -4.08081055e-01 6.96789622e-01 1.71400166e+00
-9.31735754e-01 8.24824274e-01 -1.94711655e-01 -3.14111590e-01
9.00893629e-01 7.76512861e-01 -8.07730079e-01 -5.82120009e-02
-1.36423206e+00 9.55653787e-01 1.87571183e-01 2.03496777e-02
-1.00442612e+00 -5.78068912e-01 -3.49961042e-01 -2.55710840e-01
1.40192702e-01 -4.39872146e-01 8.12060893e-01 -1.54292583e+00
-1.26532745e+00 1.21862698e+00 2.78294057e-01 -5.16053915e-01
9.34437931e-01 -1.68323383e-01 -1.10594714e+00 2.04718754e-01
-3.24256837e-01 5.00179589e-01 1.57466984e+00 -1.34785390e+00
-1.70798227e-01 -5.60575664e-01 1.55944347e-01 -3.24694723e-01
-5.11198938e-01 2.70670503e-01 2.27018103e-01 -8.62581372e-01
-5.11464417e-01 -8.28056514e-01 7.12295324e-02 1.97924897e-01
-5.19094586e-01 2.43428573e-01 1.07205534e+00 -7.96874464e-01
9.45203900e-01 -2.60566974e+00 -4.17210281e-01 4.21123266e-01
1.66903779e-01 9.32207048e-01 -1.31132320e-01 2.61813790e-01
-4.56527442e-01 7.56394863e-01 -1.12310806e-02 9.86890271e-02
-2.23179068e-02 1.33756995e-01 -5.84237754e-01 7.71006525e-01
1.92974538e-01 8.81574094e-01 -6.98777854e-01 -3.37035298e-01
7.03670159e-02 8.66740704e-01 -3.28884959e-01 3.85511875e-01
2.89484829e-01 5.22207618e-01 -4.48465258e-01 5.84601820e-01
1.04092908e+00 1.91394493e-01 4.75453399e-02 9.87006724e-02
5.49041390e-01 -3.72522742e-01 -7.95172811e-01 6.90059721e-01
-5.29648885e-02 6.48356497e-01 -1.34945130e-02 -5.93777180e-01
6.83708012e-01 5.26423156e-01 3.93316895e-02 -3.49956602e-01
3.37703228e-01 3.70988697e-02 7.31038004e-02 -5.08067489e-01
1.62185788e-01 -1.61118761e-01 3.58908102e-02 1.56055480e-01
5.93417957e-02 1.61854640e-01 -5.96749067e-01 -1.67706367e-02
9.35702205e-01 -3.21303546e-01 2.29036361e-01 -2.32412785e-01
7.90458262e-01 -5.12206376e-01 2.17191041e-01 9.05675650e-01
-5.44505835e-01 4.67642337e-01 3.31763595e-01 -3.26245993e-01
-9.81989682e-01 -1.28513384e+00 -2.77543187e-01 6.17058218e-01
-1.89829066e-01 -5.27759679e-02 -1.17779517e+00 -1.09527659e+00
3.55492324e-01 5.83401442e-01 -8.60252023e-01 -5.97612917e-01
-4.33214515e-01 -4.64318097e-01 1.34564030e+00 2.75808182e-02
6.82710350e-01 -9.38619912e-01 1.29091525e-02 -1.73671022e-01
2.68436760e-01 -1.24207759e+00 -4.42237616e-01 -4.21669126e-01
-4.20809627e-01 -1.29315650e+00 -8.14539075e-01 -4.97011572e-01
9.09379900e-01 -4.03847918e-02 8.58965397e-01 3.72707129e-01
-5.27423441e-01 7.35542774e-01 -1.84427097e-01 -7.49082446e-01
-9.90119040e-01 -3.32996547e-01 2.92565823e-01 4.15398866e-01
3.87434512e-01 -5.66401303e-01 -5.03995001e-01 4.52107877e-01
-1.05539167e+00 -7.93274283e-01 2.87941098e-01 4.35326964e-01
-1.80359975e-01 2.56161332e-01 7.46212780e-01 -1.13953698e+00
7.20494688e-01 -2.94372112e-01 -3.99756432e-01 3.20453405e-01
-4.88216639e-01 -2.64911890e-01 8.24216604e-01 -7.48025119e-01
-8.59153450e-01 -2.74658557e-02 -3.29146534e-01 -4.11369264e-01
-4.01451379e-01 -3.61921251e-01 -3.39673936e-01 -1.09756958e+00
8.03075314e-01 1.00865476e-01 2.07069114e-01 -1.78611279e-01
3.02164078e-01 7.26470172e-01 7.19083905e-01 -3.58631760e-01
1.38659692e+00 4.42994803e-01 1.06585108e-01 -1.02240050e+00
-4.71240222e-01 2.92210370e-01 -3.00723583e-01 -4.74108756e-01
7.17600882e-01 -5.23554742e-01 -1.10995901e+00 8.52995694e-01
-1.03980362e+00 1.99763164e-01 -4.15860303e-02 8.67902488e-02
-2.87801027e-01 6.46558940e-01 -3.23684931e-01 -1.10179007e+00
-1.89462095e-01 -5.85087180e-01 4.72597361e-01 1.33655548e-01
-1.32448763e-01 -9.51988876e-01 -3.23700190e-01 4.42810297e-01
6.95101798e-01 8.34799826e-01 5.49797475e-01 -6.31857574e-01
-3.19361418e-01 -7.17312217e-01 -5.66802882e-02 8.35973740e-01
3.01719218e-01 1.02622017e-01 -1.49161887e+00 -4.40841347e-01
3.36264759e-01 -1.66166082e-01 4.02447373e-01 -1.38280526e-01
1.30083263e+00 -9.66431558e-01 -1.31543323e-01 6.74589217e-01
1.38180816e+00 3.48950356e-01 1.31343710e+00 6.56974316e-02
5.30684888e-01 8.06507647e-01 7.69593790e-02 1.08409576e-01
-5.68586051e-01 4.98224258e-01 5.44550717e-01 -1.63053237e-02
5.95345087e-02 -4.57847953e-01 4.26050484e-01 -2.37649940e-02
-3.69345129e-01 -4.04984057e-01 -5.30210972e-01 4.46389914e-02
-1.21547210e+00 -1.29131258e+00 1.96654379e-01 2.48077941e+00
6.09856367e-01 6.73232749e-02 5.51627949e-02 1.69202700e-01
1.03187668e+00 4.74278070e-02 -3.60333204e-01 -4.70319420e-01
-2.66125768e-01 2.60223657e-01 7.37091541e-01 3.99077922e-01
-1.23825979e+00 8.19234312e-01 7.74132204e+00 7.61267960e-01
-8.50893795e-01 1.00327365e-01 9.31584597e-01 1.43708095e-01
-1.40153328e-02 -3.66538554e-01 -5.17791033e-01 6.44349396e-01
9.86971021e-01 -2.99203992e-01 6.51379645e-01 7.32389271e-01
-2.83637166e-01 3.68740082e-01 -1.01380730e+00 8.77921045e-01
2.54861712e-01 -9.75682199e-01 3.20020527e-01 3.74612093e-01
4.62235063e-01 -7.79590070e-01 6.34361386e-01 6.60629794e-02
4.04371500e-01 -1.52397871e+00 3.75666589e-01 4.18085873e-01
9.08903658e-01 -1.02093613e+00 5.84199429e-01 7.66736865e-02
-5.53467453e-01 -1.69295847e-01 -5.20011902e-01 1.00494340e-01
-2.31875584e-01 2.17023909e-01 -6.83381021e-01 2.68006921e-01
4.59500879e-01 -5.00955880e-02 -7.75572121e-01 5.21740139e-01
-2.95282930e-01 5.06626725e-01 -1.35412261e-01 2.03537509e-01
-7.83100054e-02 8.71420726e-02 6.55145407e-01 7.83678174e-01
9.21474099e-02 1.63708597e-01 -5.38983822e-01 8.00004482e-01
-3.02785218e-01 -2.27241397e-01 -1.24741936e+00 -3.71258020e-01
4.42891985e-01 9.39922273e-01 -2.42015481e-01 1.74986303e-01
-9.12002772e-02 1.35266447e+00 -3.18770744e-02 3.26996416e-01
-7.14770913e-01 -2.58140564e-01 7.93131232e-01 1.39281228e-01
5.04366122e-02 -8.85289162e-03 1.56273991e-02 -8.54617059e-01
5.78617826e-02 -1.21509731e+00 3.13223660e-01 -4.88742113e-01
-1.67822671e+00 7.28541076e-01 -4.19233233e-01 -1.06182790e+00
-2.13211536e-01 -6.17858112e-01 -4.34598655e-01 1.01810908e+00
-9.66573834e-01 -1.07062197e+00 6.72171190e-02 9.21677947e-01
-2.49977559e-01 -6.31216586e-01 1.13638997e+00 3.07334393e-01
-4.00911540e-01 1.33951235e+00 -1.16685927e-02 5.40246725e-01
5.63495517e-01 -8.75669241e-01 5.99874854e-01 8.92550409e-01
-5.51341753e-03 7.81884789e-01 8.41975212e-01 -4.64915156e-01
-1.37527239e+00 -9.33222592e-01 6.54261410e-01 -7.63748407e-01
4.73765194e-01 -4.85529959e-01 -7.96322107e-01 6.83118284e-01
2.02751175e-01 1.23196572e-01 7.76484847e-01 -4.02312011e-01
-6.15150988e-01 -7.54741058e-02 -2.00823259e+00 6.12339497e-01
9.28416193e-01 -1.01256883e+00 -4.19138521e-01 2.65684366e-01
6.18368685e-01 1.45216048e-01 -8.18824112e-01 3.06527913e-01
7.57175326e-01 -1.03471732e+00 1.51008332e+00 -8.82635236e-01
2.00674787e-01 5.57310730e-02 -2.62375444e-01 -8.76791775e-01
-1.65371016e-01 -1.02697754e+00 -1.03902437e-01 1.52724671e+00
1.61337495e-01 -9.79928613e-01 8.00930917e-01 8.78237307e-01
6.60135627e-01 -9.24014226e-02 -1.01188040e+00 -1.07245851e+00
2.12807342e-01 -1.55814886e-01 6.82762802e-01 1.22521317e+00
-1.96513772e-01 -1.10776074e-01 -7.15909481e-01 5.72016656e-01
8.94971013e-01 -7.79098332e-01 7.63817310e-01 -1.14778554e+00
-2.11816519e-01 -9.57914814e-02 -8.67768228e-01 -2.21908852e-01
2.24564195e-01 -5.79226136e-01 -5.58858633e-01 -3.26245725e-01
-1.93000540e-01 -2.16824412e-01 -3.19821298e-01 2.88770437e-01
-1.53376415e-01 7.55028129e-01 3.02348614e-01 1.66518971e-01
8.79882351e-02 1.37350678e-01 1.12275934e+00 -1.19369313e-01
2.87884653e-01 2.08520830e-01 -1.05786443e+00 7.72014499e-01
8.70024502e-01 -6.84497178e-01 -2.40779489e-01 2.19414644e-02
1.68912053e-01 -1.50469616e-01 7.17096508e-01 -1.13827634e+00
-1.33692399e-01 2.95518786e-02 6.88224554e-01 1.43367082e-01
2.36693293e-01 -1.21407533e+00 3.50472689e-01 5.61381578e-01
-2.94847846e-01 -9.95991230e-02 3.51693392e-01 5.24524808e-01
5.32463342e-02 -3.43950838e-01 1.18264055e+00 -2.65141338e-01
-4.52295303e-01 2.93160170e-01 -2.25905687e-01 -9.01698321e-03
1.27464294e+00 -2.48306364e-01 -4.84112680e-01 -4.88737673e-01
-6.65617466e-01 -3.42402369e-01 5.35166860e-01 3.85129601e-01
5.63922882e-01 -1.30432689e+00 -6.82424903e-01 4.17254716e-01
-1.36928037e-01 -9.30223703e-01 2.19216883e-01 1.25293419e-01
-4.76187646e-01 1.14618160e-01 -5.15871644e-01 2.39275973e-02
-1.55067778e+00 1.12519526e+00 5.93040943e-01 2.76504964e-01
-2.30692640e-01 8.16411972e-01 6.03389516e-02 -1.75946906e-01
2.79409945e-01 7.94292808e-01 -6.74344227e-02 -3.14466596e-01
8.45190585e-01 3.81932586e-01 -1.01103462e-01 -7.90050030e-01
-4.66301471e-01 3.13784659e-01 -1.47639751e-01 1.09895654e-01
9.32194233e-01 1.33711368e-01 -1.59355283e-01 -1.69396713e-01
1.41691232e+00 2.81019121e-01 -9.39672291e-01 2.12786824e-01
-2.64697641e-01 -1.06719458e+00 -2.61008948e-01 -7.67172694e-01
-1.03633559e+00 5.51558793e-01 1.01374555e+00 8.41121793e-01
1.07211936e+00 -3.57197016e-01 4.19261396e-01 2.98029989e-01
4.37743008e-01 -7.73589611e-01 -2.98923925e-02 3.78868058e-02
9.41108882e-01 -1.20884907e+00 -2.15043463e-02 -6.82408333e-01
-4.15575773e-01 8.63201022e-01 3.15635949e-01 -2.17516139e-01
7.53454745e-01 1.17028370e-01 2.08000451e-01 1.78046674e-02
-1.21573508e-01 3.10547322e-01 3.46257865e-01 1.26599395e+00
9.21414196e-02 3.05073299e-02 -1.74349934e-01 2.27789029e-01
-4.18384820e-01 -7.37846941e-02 5.18875659e-01 8.25527668e-01
-9.28924531e-02 -1.31712162e+00 -8.31496298e-01 2.75348902e-01
-1.15510082e+00 1.13861509e-01 -9.85208154e-01 8.35175097e-01
2.20913991e-01 1.16697764e+00 -2.64789641e-01 -6.39088809e-01
2.18031526e-01 6.45944104e-02 5.17096162e-01 -1.15202554e-01
-8.79178762e-01 -7.95530498e-01 5.20794056e-02 -5.63531518e-01
-4.45986599e-01 -4.52102542e-01 -2.76639313e-01 -8.81765842e-01
-1.26669794e-01 9.40315705e-03 5.42189300e-01 7.34117985e-01
3.54143471e-01 7.83221498e-02 1.11931610e+00 -4.71069455e-01
-8.56113791e-01 -6.53317750e-01 -7.95009732e-01 5.98509967e-01
4.34376806e-01 -3.43532085e-01 -7.92523742e-01 4.39904816e-02] | [12.874418258666992, 1.0355737209320068] |
ad20f423-f714-40dd-8c57-3aabc68e77a4 | exploration-based-language-learning-for-text-1 | 2001.08868 | null | https://arxiv.org/abs/2001.08868v2 | https://arxiv.org/pdf/2001.08868v2.pdf | Exploration Based Language Learning for Text-Based Games | This work presents an exploration and imitation-learning-based agent capable of state-of-the-art performance in playing text-based computer games. Text-based computer games describe their world to the player through natural language and expect the player to interact with the game using text. These games are of interest as they can be seen as a testbed for language understanding, problem-solving, and language generation by artificial agents. Moreover, they provide a learning environment in which these skills can be acquired through interactions with an environment rather than using fixed corpora. One aspect that makes these games particularly challenging for learning agents is the combinatorially large action space. Existing methods for solving text-based games are limited to games that are either very simple or have an action space restricted to a predetermined set of admissible actions. In this work, we propose to use the exploration approach of Go-Explore for solving text-based games. More specifically, in an initial exploration phase, we first extract trajectories with high rewards, after which we train a policy to solve the game by imitating these trajectories. Our experiments show that this approach outperforms existing solutions in solving text-based games, and it is more sample efficient in terms of the number of interactions with the environment. Moreover, we show that the learned policy can generalize better than existing solutions to unseen games without using any restriction on the action space. | ['Adrien Ecoffet', 'Piero Molino', 'Mahdi Namazifar', 'Alexandros Papangelis', 'Joost Huizinga', 'Huaixiu Zheng', 'Dian Yu', 'Andrea Madotto', 'Gokhan Tur', 'Chandra Khatri'] | 2020-01-24 | null | https://openreview.net/forum?id=BygSXCNFDB | https://openreview.net/pdf?id=BygSXCNFDB | null | ['text-based-games'] | ['playing-games'] | [ 2.86782961e-02 2.35354707e-01 2.20652111e-02 2.70096809e-01
-4.51945215e-01 -7.91182041e-01 8.62698734e-01 -4.34601940e-02
-7.52399743e-01 8.43598843e-01 -1.93695843e-01 -5.41162074e-01
-2.38161251e-01 -1.20889568e+00 -7.33066738e-01 -4.66477722e-01
-9.32980403e-02 1.12105572e+00 4.48745191e-01 -7.98231125e-01
2.25303262e-01 2.67779320e-01 -1.53197408e+00 -4.59479401e-03
1.06969106e+00 3.64818901e-01 7.87949920e-01 8.33044529e-01
-3.39413404e-01 8.72312963e-01 -7.12885618e-01 -1.48325160e-01
4.51136410e-01 -7.86610723e-01 -9.67831135e-01 3.02749634e-01
-6.34444833e-01 -2.51760662e-01 -2.59706974e-01 1.10807717e+00
1.34669125e-01 7.09050298e-01 3.39974612e-01 -1.21384752e+00
-1.46668836e-01 5.77966213e-01 9.41207074e-03 -1.88373700e-01
6.56026959e-01 4.86945719e-01 8.50658774e-01 -4.47671860e-01
8.78816307e-01 1.12262821e+00 -6.82047103e-03 1.12586868e+00
-9.54110205e-01 -3.40927333e-01 2.81042546e-01 1.42753229e-01
-1.08241260e+00 2.11621802e-02 5.48092961e-01 -2.55576491e-01
1.19605005e+00 2.76098400e-01 1.16845548e+00 1.02074134e+00
9.22927186e-02 9.89932299e-01 1.00477183e+00 -6.29650056e-01
6.66488171e-01 5.24485894e-02 -5.98771274e-01 7.57334411e-01
-1.73464328e-01 4.12104607e-01 -4.36644197e-01 -8.94635767e-02
9.51007903e-01 -1.58818334e-01 2.15352699e-02 -6.12519681e-01
-9.72369432e-01 1.18710899e+00 1.07727207e-01 2.51180023e-01
-4.52940732e-01 2.42337927e-01 2.85160124e-01 5.98020673e-01
2.42567584e-01 1.05492413e+00 -2.85631239e-01 -6.45588160e-01
-5.91980755e-01 7.77721703e-01 1.10273695e+00 8.70513916e-01
6.16630077e-01 6.17979988e-02 1.50458768e-01 6.38930976e-01
2.32463479e-01 3.69957536e-01 6.66753471e-01 -9.70682323e-01
4.78656054e-01 6.06676817e-01 3.32546324e-01 -4.53773886e-01
-3.22579145e-01 -8.88588093e-03 -2.80690014e-01 5.82319081e-01
5.32415211e-01 -4.17380333e-01 -8.14250231e-01 1.60291827e+00
2.80121982e-01 3.30554396e-01 4.69958097e-01 7.93155193e-01
4.78679270e-01 8.60485196e-01 -1.04448952e-01 -2.32846797e-01
1.09246945e+00 -1.19707429e+00 -3.02532017e-01 -6.12625897e-01
7.91542709e-01 -1.53285265e-01 1.22665024e+00 5.11217356e-01
-1.24279273e+00 -1.10229254e-01 -7.01666176e-01 4.65747893e-01
-4.58784014e-01 -3.77905577e-01 6.95570767e-01 5.93030930e-01
-1.11488485e+00 6.81414962e-01 -8.87806952e-01 -4.09256130e-01
1.24958893e-02 5.63936889e-01 -7.18333200e-02 5.16337436e-03
-1.18561697e+00 9.24883664e-01 9.39050317e-01 -1.90792561e-01
-1.07826340e+00 -1.20659634e-01 -1.02628398e+00 4.86804359e-02
1.15769005e+00 -4.69173819e-01 1.49208927e+00 -8.56551230e-01
-2.07832146e+00 5.86233497e-01 1.55224264e-01 -5.94585121e-01
7.74639726e-01 1.72212988e-01 1.57612815e-01 2.06261296e-02
-2.20636483e-02 7.27891684e-01 5.81779122e-01 -1.09426439e+00
-7.95437932e-01 2.62078680e-02 8.29658270e-01 6.27409339e-01
-2.11982816e-01 4.67304997e-02 -6.29282415e-01 -3.50106269e-01
-2.35704944e-01 -1.17733443e+00 -8.41604292e-01 -3.43480319e-01
-2.15635851e-01 -3.11314642e-01 5.24603963e-01 -1.40411407e-01
8.52451086e-01 -1.76539671e+00 5.71557224e-01 3.09644520e-01
2.61745244e-01 4.92686957e-01 -3.58206928e-01 8.37160945e-01
4.43448931e-01 3.04752052e-01 -2.12643612e-02 -1.81962013e-01
1.75146624e-01 6.34868622e-01 -1.60966739e-01 -2.00309277e-01
-7.28834467e-03 1.09772813e+00 -1.32642078e+00 -2.20483422e-01
2.99245745e-01 -2.00871408e-01 -8.48687053e-01 3.70292991e-01
-9.63863730e-01 6.78644657e-01 -9.13885891e-01 -3.47373262e-02
9.33332741e-02 -5.45484498e-02 4.13482219e-01 1.10551977e+00
-1.05072401e-01 2.39527106e-01 -1.16613841e+00 1.68548298e+00
-6.14404202e-01 4.24359530e-01 -9.58364308e-02 -1.07530940e+00
8.96975756e-01 1.77937388e-01 3.60429734e-01 -4.86409515e-01
1.11392833e-01 1.08659342e-01 1.33855969e-01 -4.83302027e-01
6.05543017e-01 -1.20766461e-01 -2.27741748e-01 8.62392724e-01
-9.28834006e-02 -7.88819671e-01 7.54624009e-01 1.81902096e-01
1.09174860e+00 1.82773292e-01 3.28326523e-01 9.47612301e-02
4.35803622e-01 2.77114838e-01 3.68612379e-01 1.25425458e+00
2.28558719e-01 1.41034663e-01 6.70220613e-01 -4.95826513e-01
-1.04821706e+00 -7.71323025e-01 6.95939541e-01 1.13664424e+00
2.25344837e-01 -7.02120245e-01 -9.15202737e-01 -5.80850899e-01
-5.51511109e-01 8.53055954e-01 -4.14742798e-01 -1.32828087e-01
-7.44185627e-01 -2.42627457e-01 3.11879635e-01 2.90075481e-01
5.44804990e-01 -1.85217166e+00 -1.10449541e+00 4.56208378e-01
-1.76631048e-01 -1.08833623e+00 -3.38695407e-01 1.40900254e-01
-7.23744154e-01 -1.07624245e+00 -5.63480675e-01 -8.14898074e-01
5.54555237e-01 -8.94772336e-02 9.00947809e-01 3.83258313e-01
-1.20355912e-01 5.81471384e-01 -6.41163349e-01 -7.31749177e-01
-8.56345713e-01 1.90776125e-01 -1.00914322e-01 -4.57937658e-01
2.44140327e-01 -4.61566895e-01 -1.20214127e-01 1.61875904e-01
-1.02134216e+00 4.60908562e-01 2.48101890e-01 1.05172002e+00
2.08418399e-01 3.34887266e-01 2.81955242e-01 -7.20528841e-01
1.28009582e+00 -2.84699857e-01 -9.98718262e-01 2.32935250e-01
-2.10033640e-01 2.74444401e-01 8.92727017e-01 -6.82571590e-01
-8.19896221e-01 1.37552455e-01 -1.58734918e-02 -1.89900190e-01
-2.24931926e-01 6.55219436e-01 7.97989815e-02 -5.08725718e-02
7.03869164e-01 6.36582077e-01 6.93034455e-02 -3.29822749e-02
2.70430207e-01 4.28117037e-01 2.02057883e-01 -9.81775284e-01
8.56690884e-01 4.41739000e-02 -1.08496733e-01 -8.07120502e-01
-3.39481950e-01 -3.40668052e-01 -3.98706973e-01 -3.04571331e-01
7.03286290e-01 -3.82279217e-01 -1.06666183e+00 4.89008904e-01
-9.38046455e-01 -9.16653812e-01 -4.61982399e-01 4.54312712e-01
-1.17329192e+00 2.27808803e-01 -4.36038971e-01 -1.16118121e+00
1.83319151e-02 -1.40430903e+00 7.73516238e-01 4.80177552e-01
-1.62661925e-01 -1.13068485e+00 1.67383641e-01 1.62387446e-01
4.13113534e-01 1.10394932e-01 7.54851937e-01 -8.68301749e-01
-6.21661246e-01 1.34565637e-01 4.22818661e-01 -2.19143704e-01
-2.99442653e-02 -4.37201560e-01 -3.70424807e-01 -3.65202665e-01
-1.09763019e-01 -6.59165382e-01 3.54929715e-01 1.74862891e-01
1.08009410e+00 -3.34566027e-01 -3.23027045e-01 3.22012037e-01
1.12917936e+00 5.97458780e-01 5.23363411e-01 7.88853228e-01
1.56026840e-01 6.32224143e-01 7.11500645e-01 5.97661614e-01
2.24768326e-01 7.99794495e-01 5.91702700e-01 1.58637613e-01
6.35857403e-01 -4.51366574e-01 3.10878575e-01 3.02214235e-01
-2.84268528e-01 -4.01445895e-01 -1.10192978e+00 6.15746975e-01
-2.19226646e+00 -9.29770112e-01 4.21032071e-01 2.10905695e+00
8.37427557e-01 2.59089261e-01 4.18677330e-01 -5.28570935e-02
4.19590205e-01 -1.45052716e-01 -6.29925787e-01 -6.74157321e-01
2.21546337e-01 5.37002206e-01 2.23192409e-01 7.00247586e-01
-7.99014449e-01 1.64305449e+00 5.85341930e+00 8.63888800e-01
-9.41217184e-01 -2.41487056e-01 1.96053028e-01 -1.82507634e-01
-1.55647337e-01 -4.31736037e-02 -4.09364849e-01 1.62460402e-01
8.36988389e-01 -6.24751627e-01 1.10690951e+00 7.49210477e-01
4.72624362e-01 -1.93491861e-01 -1.03487432e+00 6.38696849e-01
-3.24947596e-01 -1.39825439e+00 3.82088013e-02 2.74057299e-01
5.92934191e-01 -1.58750847e-01 -1.02408007e-02 5.54021418e-01
9.52351093e-01 -1.29853201e+00 6.76064670e-01 1.10697180e-01
4.65747118e-01 -1.00643170e+00 5.06559253e-01 1.19938636e+00
-1.09008372e+00 -3.61214578e-01 -4.25322413e-01 -4.24443245e-01
1.06538057e-01 -3.77822012e-01 -9.68800962e-01 4.90955025e-01
3.48953307e-01 1.94901481e-01 -8.74584094e-02 1.17287517e+00
-5.56589961e-01 3.97041678e-01 -3.40183973e-01 -8.27118039e-01
8.04373384e-01 -5.09081483e-01 7.09239960e-01 7.23442972e-01
3.84719789e-01 5.45565784e-01 7.51606584e-01 9.50038075e-01
2.61759341e-01 1.27076998e-01 -9.58943009e-01 -3.02951634e-01
2.92449355e-01 7.85901487e-01 -8.73732090e-01 -2.53651977e-01
-2.04521254e-01 9.67118859e-01 4.96740311e-01 3.66290152e-01
-5.50849259e-01 -3.19017112e-01 4.88048941e-01 -6.35930896e-02
2.69177616e-01 -4.06355709e-01 2.55710661e-01 -1.15838885e+00
-1.33537844e-01 -1.24010086e+00 8.90657753e-02 -7.25385606e-01
-6.93154514e-01 7.36559689e-01 2.40704194e-01 -9.31688309e-01
-1.06978929e+00 -6.30109131e-01 -7.50767708e-01 7.45269239e-01
-1.03004503e+00 -8.33177030e-01 1.88980214e-02 7.58120954e-01
9.60643709e-01 -5.26460648e-01 9.46194410e-01 -5.18803775e-01
-2.04857543e-01 2.16727614e-01 1.02897406e-01 1.17309444e-01
-1.10700257e-01 -1.46789372e+00 6.30965292e-01 6.08943343e-01
5.04334390e-01 3.93400639e-01 8.85952294e-01 -5.81386745e-01
-1.38746512e+00 -7.37308025e-01 4.26931858e-01 -1.62625939e-01
7.59179354e-01 -6.42005503e-01 -6.89453661e-01 6.05178237e-01
7.16237575e-02 -4.11275566e-01 1.92284510e-01 -8.33190531e-02
4.36335623e-01 5.66813946e-01 -9.82707918e-01 1.09912598e+00
1.14639711e+00 -2.82580018e-01 -6.00034177e-01 5.40454686e-01
7.07394361e-01 -9.35832679e-01 -2.57518589e-01 -1.95328549e-01
1.51934579e-01 -6.88425004e-01 7.18630075e-01 -9.84016776e-01
3.83876413e-01 -1.43220067e-01 3.79011899e-01 -1.76108313e+00
9.05519724e-02 -1.09056509e+00 1.10563211e-01 5.98635375e-01
4.92944032e-01 -7.78340578e-01 1.22177517e+00 6.25500202e-01
1.52611643e-01 -8.76855314e-01 -7.54370868e-01 -1.01349032e+00
4.48581368e-01 -5.62726557e-01 6.18447304e-01 5.39705217e-01
4.38086838e-01 6.37410656e-02 -3.77927631e-01 -1.55954793e-01
2.19819143e-01 4.35010940e-02 9.84185457e-01 -1.15704584e+00
-7.38281250e-01 -5.14492989e-01 -1.11908406e-01 -1.33915842e+00
4.68878150e-01 -6.55352175e-01 3.23017806e-01 -1.50134385e+00
-2.34647058e-02 -4.93248314e-01 2.52872378e-01 4.06850427e-01
8.28551129e-03 -3.57115746e-01 6.01978660e-01 -5.38915535e-03
-6.62704408e-01 5.32398880e-01 1.58244765e+00 -1.61538482e-01
-6.60240531e-01 3.29280734e-01 -3.84132832e-01 7.62554169e-01
9.46543276e-01 -3.71438056e-01 -7.95324087e-01 -2.15221226e-01
4.07226562e-01 5.03506899e-01 4.36819419e-02 -7.92460978e-01
5.27819455e-01 -7.81521022e-01 -3.19527388e-01 -1.94918998e-02
5.66440761e-01 -5.90890527e-01 -1.12311266e-01 7.15169907e-01
-4.76763010e-01 2.53241360e-02 3.48721296e-01 4.68202025e-01
-1.51197925e-01 -8.04138064e-01 3.66105467e-01 -6.47452950e-01
-8.48348737e-01 2.27942675e-01 -1.15660024e+00 2.56915152e-01
1.42074513e+00 -3.61219555e-01 1.10186428e-01 -9.67850745e-01
-9.11292255e-01 5.98787487e-01 5.32883763e-01 3.62762123e-01
6.80573106e-01 -9.70656455e-01 -6.70619905e-01 -3.96215543e-03
-9.05469526e-03 1.78348571e-01 -2.37774834e-01 6.71607703e-02
-6.22561812e-01 3.44769776e-01 -2.96755016e-01 -2.28418425e-01
-1.32259440e+00 5.15299857e-01 6.15209222e-01 -8.46351206e-01
-7.81962872e-01 6.58151805e-01 2.73205221e-01 -7.97996402e-01
1.43293351e-01 -2.63704836e-01 -3.72337192e-01 -5.34274697e-01
4.97661859e-01 -9.87004954e-04 -3.69975775e-01 -1.90016285e-01
-3.41500156e-02 3.45415086e-01 -8.89795125e-02 -6.34624958e-01
1.40924037e+00 3.18562686e-01 1.00656189e-01 6.26273379e-02
3.37247461e-01 -2.06269115e-01 -1.11003578e+00 -2.95829445e-01
1.12242937e-01 -4.45923895e-01 -3.23880732e-01 -5.91644049e-01
-6.81591332e-01 6.39236212e-01 -6.12232722e-02 5.83693326e-01
1.00764954e+00 5.18613681e-02 5.69297373e-01 9.41795886e-01
7.72099614e-01 -1.16329730e+00 5.33127964e-01 1.08687973e+00
8.11098754e-01 -9.38167274e-01 -4.53567743e-01 -2.75598854e-01
-8.14905286e-01 1.13073277e+00 8.85013998e-01 -1.69393450e-01
1.08983435e-01 3.12594414e-01 -9.97482315e-02 -2.00621516e-01
-1.05459261e+00 -4.73703563e-01 -3.83756161e-02 9.76043046e-01
-1.24722742e-01 -3.70353349e-02 -3.43635798e-01 3.14132005e-01
-6.52069092e-01 -2.61925012e-02 8.95954013e-01 1.08385396e+00
-6.53077841e-01 -1.62377548e+00 -3.09137642e-01 2.31830046e-01
-1.99896723e-01 5.78269316e-03 -7.29277134e-01 9.61105883e-01
-2.08879441e-01 1.23009408e+00 -1.88554525e-01 -1.93009838e-01
2.62780130e-01 -1.44290730e-01 6.02048337e-01 -1.05415726e+00
-7.20110536e-01 -1.63118109e-01 2.60676861e-01 -6.39148533e-01
1.29789915e-02 -5.29514015e-01 -1.50587559e+00 -3.26661885e-01
-2.76302367e-01 6.40402794e-01 5.04495144e-01 1.35033977e+00
-7.89713487e-02 3.08166027e-01 4.85768855e-01 -9.20312285e-01
-5.69122255e-01 -6.79143548e-01 -4.92683977e-01 1.34887174e-01
-4.90690395e-02 -5.55130780e-01 2.13526227e-02 -3.38826060e-01] | [3.8117687702178955, 1.480571985244751] |
e189fad4-262c-4355-a793-fabdbff19b4e | egocentric-video-task-translation | 2212.06301 | null | https://arxiv.org/abs/2212.06301v2 | https://arxiv.org/pdf/2212.06301v2.pdf | Egocentric Video Task Translation | Different video understanding tasks are typically treated in isolation, and even with distinct types of curated data (e.g., classifying sports in one dataset, tracking animals in another). However, in wearable cameras, the immersive egocentric perspective of a person engaging with the world around them presents an interconnected web of video understanding tasks -- hand-object manipulations, navigation in the space, or human-human interactions -- that unfold continuously, driven by the person's goals. We argue that this calls for a much more unified approach. We propose EgoTask Translation (EgoT2), which takes a collection of models optimized on separate tasks and learns to translate their outputs for improved performance on any or all of them at once. Unlike traditional transfer or multi-task learning, EgoT2's flipped design entails separate task-specific backbones and a task translator shared across all tasks, which captures synergies between even heterogeneous tasks and mitigates task competition. Demonstrating our model on a wide array of video tasks from Ego4D, we show its advantages over existing transfer paradigms and achieve top-ranked results on four of the Ego4D 2022 benchmark challenges. | ['Lorenzo Torresani', 'Kristen Grauman', 'Yale Song', 'Zihui Xue'] | 2022-12-13 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Xue_Egocentric_Video_Task_Translation_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Xue_Egocentric_Video_Task_Translation_CVPR_2023_paper.pdf | cvpr-2023-1 | ['video-understanding'] | ['computer-vision'] | [ 2.70645946e-01 -1.23218998e-01 -3.69525328e-02 -2.39762694e-01
-6.54995561e-01 -9.08211827e-01 7.32551873e-01 -3.98757160e-01
-3.37457716e-01 4.37788129e-01 6.36211038e-01 -3.08426414e-02
-3.66492532e-02 -3.60145211e-01 -1.14498889e+00 -2.84168839e-01
1.51394457e-02 3.98868322e-01 1.10413238e-01 -9.35521871e-02
3.84673513e-02 -1.95126727e-01 -1.51405883e+00 5.38016677e-01
3.70989352e-01 7.28750706e-01 2.92916626e-01 9.17335629e-01
3.49413246e-01 1.08087361e+00 -2.22278833e-01 -4.40670371e-01
2.57878006e-01 -8.17632899e-02 -9.74239826e-01 1.92238986e-01
9.14061606e-01 -6.04842961e-01 -8.16160440e-01 7.59130895e-01
3.79793227e-01 3.17187190e-01 6.54146314e-01 -1.44503069e+00
-9.42508817e-01 2.29810968e-01 -3.49450856e-01 1.93166539e-01
7.03924060e-01 4.24302191e-01 7.48406649e-01 -8.29987228e-01
8.74109685e-01 1.20399463e+00 6.76794410e-01 6.52826667e-01
-1.25078523e+00 -5.86143374e-01 3.53793681e-01 4.98257756e-01
-8.06646705e-01 -5.31114161e-01 4.57044125e-01 -7.51849532e-01
1.02806437e+00 -3.44462208e-02 7.58830965e-01 1.95849025e+00
2.61806190e-01 9.21707571e-01 7.64893115e-01 2.05761135e-01
-1.56281311e-02 -1.09267160e-01 2.02902690e-01 4.13630873e-01
2.33596742e-01 -1.13150828e-01 -1.18788719e+00 3.41174275e-01
7.37171233e-01 1.03706099e-01 -3.21926206e-01 -8.06249440e-01
-1.68360341e+00 3.09331954e-01 2.15222761e-01 -1.06578529e-01
-3.02793205e-01 6.14290297e-01 8.08138430e-01 4.20312405e-01
5.18860817e-01 2.84859985e-01 -4.75454956e-01 -6.32639468e-01
-4.27166849e-01 5.78781903e-01 6.74079061e-01 1.36072862e+00
6.75539315e-01 -2.26252824e-01 -2.82017857e-01 2.84030408e-01
-1.87393665e-01 3.56638014e-01 9.01645496e-02 -1.18276119e+00
7.65393376e-01 4.10799474e-01 2.64921695e-01 -7.88182497e-01
-3.95299107e-01 -1.73911378e-01 -4.66054887e-01 -6.34319484e-02
6.06838048e-01 -2.77723312e-01 -8.00915301e-01 2.12796330e+00
2.67958730e-01 5.78134716e-01 -8.23027045e-02 1.10328555e+00
8.74230146e-01 5.04883111e-01 3.14571977e-01 4.56150889e-01
1.57570148e+00 -1.26074684e+00 -5.99370241e-01 -6.30798042e-01
8.23964417e-01 -2.57286549e-01 1.23783147e+00 4.29317266e-01
-1.18669069e+00 -7.43784368e-01 -8.29081655e-01 -6.72180831e-01
-5.91005921e-01 -9.75012407e-02 6.99986279e-01 3.25206906e-01
-1.17974186e+00 4.03643757e-01 -7.91690648e-01 -6.85779929e-01
5.24385750e-01 1.52245462e-01 -9.06841457e-01 -1.77125201e-01
-1.05218697e+00 8.82448733e-01 4.56323996e-02 -5.47017455e-02
-1.47210515e+00 -1.15876043e+00 -9.64756131e-01 9.62416604e-02
5.60230076e-01 -1.30383313e+00 1.49042118e+00 -8.13964844e-01
-1.14421237e+00 1.23575115e+00 6.48202375e-02 -3.62129122e-01
7.88943708e-01 -9.27685738e-01 9.19184089e-03 5.70680723e-02
2.91514218e-01 8.64839315e-01 8.29424202e-01 -8.66751313e-01
-4.66541231e-01 -5.18370152e-01 6.59499943e-01 6.53517485e-01
-4.76089150e-01 -7.58353174e-02 -6.49966240e-01 -4.69034880e-01
-2.80461103e-01 -1.10416889e+00 1.89875931e-01 2.43018582e-01
-5.22317924e-02 -2.16018677e-01 1.08665693e+00 -6.24517083e-01
8.08055341e-01 -2.18106031e+00 9.10984516e-01 -6.42915726e-01
7.27951229e-01 -1.56481322e-02 -1.34390935e-01 4.44444031e-01
-2.99616933e-01 6.54219091e-02 2.15683818e-01 -6.89066350e-01
9.56822187e-02 2.32972670e-03 -4.78830710e-02 4.24205214e-01
-2.52298173e-02 1.23907983e+00 -1.25517702e+00 -1.62430227e-01
3.47238094e-01 5.21560431e-01 -7.26896524e-01 1.10624008e-01
6.41821418e-03 7.07445860e-01 -5.42357504e-01 4.33530927e-01
3.29904348e-01 -4.95806396e-01 -1.02764191e-02 -4.45325553e-01
2.32170299e-01 1.71355128e-01 -8.64197910e-01 2.50101900e+00
-6.18424416e-01 9.64215040e-01 1.59921587e-01 -1.14865589e+00
2.49296799e-01 5.75834334e-01 5.90535402e-01 -6.09631598e-01
-9.27854553e-02 -1.91468820e-01 -2.44239256e-01 -9.92642760e-01
4.22023445e-01 1.32420689e-01 -4.08793420e-01 6.52600706e-01
6.18056059e-01 3.36299203e-02 8.15456659e-02 4.39225078e-01
1.37667274e+00 7.98727989e-01 1.61109298e-01 -2.32446954e-01
3.09337918e-02 8.13118443e-02 4.23976183e-01 9.12560523e-01
-6.55095994e-01 6.58179224e-01 5.45824170e-01 -9.06803489e-01
-1.09572351e+00 -1.24353671e+00 2.90382653e-01 1.54985011e+00
3.84236395e-01 -4.75841135e-01 -8.18149388e-01 -7.83704281e-01
9.86349955e-02 5.56874156e-01 -1.03079915e+00 -3.60990584e-01
-6.01287484e-01 -2.90975869e-01 4.86605078e-01 6.02703452e-01
5.93857229e-01 -9.76500928e-01 -8.51010740e-01 -3.58648114e-02
-6.28315151e-01 -1.56380713e+00 -6.60813391e-01 4.10929322e-02
-6.63929462e-01 -1.24659836e+00 -6.30359232e-01 -6.42912388e-01
3.65422875e-01 8.37140083e-01 1.53922391e+00 -5.39159536e-01
-1.96451098e-01 1.06589949e+00 -4.50312734e-01 -3.00531149e-01
1.45301998e-01 1.92636922e-01 2.48444557e-01 4.87956852e-02
6.42142653e-01 -8.42172265e-01 -6.61615789e-01 3.37395370e-01
-5.33237934e-01 5.22896349e-01 2.82074243e-01 5.29143512e-01
3.95552590e-02 -6.61733568e-01 4.61431354e-01 -8.95421684e-01
3.95280391e-01 -7.82749116e-01 1.60304736e-02 2.15172768e-01
5.51626161e-02 -3.32505643e-01 3.25885862e-01 -5.77025831e-01
-1.04392481e+00 1.80203896e-02 5.50954580e-01 -8.95969689e-01
-2.68646687e-01 7.48473778e-02 -2.41217390e-01 1.28149018e-01
7.80397475e-01 3.33896399e-01 -1.14831567e-01 -4.10729200e-01
5.92947841e-01 2.32205600e-01 8.08505774e-01 -8.29343677e-01
8.34480464e-01 6.14593685e-01 -1.85366377e-01 -6.47414446e-01
-9.81401742e-01 -3.66969049e-01 -7.90762842e-01 -7.09607661e-01
1.38588762e+00 -1.46080315e+00 -1.11693954e+00 5.46537757e-01
-1.27110541e+00 -6.18722320e-01 -2.34508336e-01 4.46746886e-01
-9.81717527e-01 9.08985212e-02 -4.11653727e-01 -2.15695754e-01
5.08156884e-03 -1.27058744e+00 1.47116399e+00 1.78799838e-01
-4.34767663e-01 -1.07979453e+00 1.08528873e-02 7.92914033e-01
3.17937911e-01 1.92861900e-01 5.89189172e-01 -6.04103804e-01
-7.77883768e-01 -1.51519492e-01 -4.02986109e-01 2.01599285e-01
1.32457510e-01 -6.87674165e-01 -9.09349918e-01 -3.09470862e-01
5.23696654e-02 -7.43686557e-01 7.16771066e-01 3.18186551e-01
1.32553887e+00 5.01163229e-02 -4.26535100e-01 9.15248692e-01
9.61493433e-01 -1.96073517e-01 5.98794341e-01 2.63254821e-01
9.99787807e-01 6.21773601e-01 3.38850319e-01 1.03391960e-01
8.25966179e-01 8.07068765e-01 4.83601391e-01 1.83379173e-01
-1.62955731e-01 -4.13462490e-01 6.54444873e-01 4.80876625e-01
-4.20733690e-01 -2.68457502e-01 -8.93306375e-01 6.24687910e-01
-2.16610432e+00 -1.08707154e+00 -2.84018014e-02 1.99869502e+00
3.84364933e-01 7.18498975e-02 1.81579351e-01 -5.04407823e-01
4.83037144e-01 3.68196636e-01 -8.92396808e-01 -1.38999015e-01
8.56809318e-02 -2.54791915e-01 4.09778118e-01 1.52607352e-01
-1.35071337e+00 9.99336898e-01 6.89921761e+00 4.23854500e-01
-7.56513536e-01 4.69475895e-01 3.33823323e-01 -6.17657423e-01
1.20747127e-01 -3.13999429e-02 -4.41396952e-01 3.59107256e-01
6.68095946e-01 -2.82988608e-01 5.35015106e-01 8.32508683e-01
2.04541877e-01 1.85930021e-02 -1.76542222e+00 1.24556398e+00
2.92242646e-01 -1.26209605e+00 1.85642034e-01 1.12519071e-01
6.40759408e-01 8.29457492e-02 9.95792225e-02 6.33656144e-01
3.18790257e-01 -9.59517658e-01 8.86922240e-01 4.99834895e-01
7.83050001e-01 -6.57031164e-02 1.73274249e-01 3.24409962e-01
-1.25471652e+00 -1.24512762e-01 1.33881375e-01 -4.04354900e-01
4.17420387e-01 -1.04118161e-01 -1.70139685e-01 4.38715845e-01
1.04499292e+00 1.23380268e+00 -3.71499538e-01 5.82074285e-01
-5.64340316e-02 2.08711490e-01 -4.81600687e-02 3.44120800e-01
1.96266234e-01 -3.24337244e-01 7.59788692e-01 1.26272774e+00
2.41650522e-01 8.96540880e-02 5.89373447e-02 6.23913944e-01
-3.51916969e-01 -2.75477827e-01 -1.15300739e+00 1.19947441e-01
1.64850190e-01 1.24714530e+00 -2.41901100e-01 -4.22461271e-01
-7.73963630e-01 1.21555805e+00 5.16897857e-01 6.10856533e-01
-1.19512403e+00 3.35945538e-03 1.23095143e+00 2.78032780e-01
1.20244317e-01 -4.13862079e-01 -2.53283620e-01 -1.64285064e+00
2.64229298e-01 -9.19797659e-01 2.41862431e-01 -1.23785150e+00
-1.12134516e+00 1.59216180e-01 2.77720571e-01 -1.41767323e+00
-1.22515626e-01 -7.94665933e-01 -6.15865767e-01 5.28237462e-01
-1.21503949e+00 -1.56566870e+00 -8.54974091e-01 7.32704520e-01
1.14009953e+00 -3.74069624e-02 5.30793071e-01 4.90913868e-01
-5.09647727e-01 3.90720576e-01 -3.58320504e-01 1.06454015e-01
1.00458360e+00 -1.29108405e+00 7.36059368e-01 5.96398413e-01
8.26783553e-02 4.06199038e-01 6.25357449e-01 -4.99539614e-01
-1.82008779e+00 -9.37129676e-01 5.81779540e-01 -1.28512788e+00
7.52757430e-01 -8.78801167e-01 -6.74856365e-01 1.49156058e+00
2.33624801e-01 -7.30996728e-02 6.69491708e-01 4.94021177e-01
-6.79845691e-01 5.15426509e-02 -5.58859706e-01 7.94458330e-01
1.92623460e+00 -8.33451390e-01 -7.34439731e-01 7.39700615e-01
8.54906976e-01 -5.68952382e-01 -7.82652140e-01 1.12318896e-01
8.66631150e-01 -9.94376898e-01 1.31045866e+00 -1.35750210e+00
9.25332665e-01 -1.34931400e-01 -1.72433317e-01 -1.31630814e+00
-5.06126165e-01 -7.76247919e-01 -2.97229350e-01 7.88326859e-01
5.07152826e-02 -2.93009520e-01 7.97496080e-01 6.52541280e-01
-3.88558537e-01 -3.26353580e-01 -7.74329484e-01 -6.06254399e-01
-6.18998185e-02 -5.22643685e-01 1.68747947e-01 9.87877607e-01
1.83422625e-01 5.05632043e-01 -8.74595225e-01 -8.32472071e-02
7.99269795e-01 -3.36810559e-01 1.28096628e+00 -1.07157826e+00
-2.54914850e-01 -2.08967134e-01 -4.97847289e-01 -1.38024616e+00
1.48502290e-01 -8.69709611e-01 -2.15956390e-01 -1.38617480e+00
5.58389604e-01 7.17480406e-02 -7.26564676e-02 4.38132495e-01
-1.51492923e-01 2.64143169e-01 3.01308602e-01 2.25035086e-01
-1.14547861e+00 4.28117454e-01 1.35012555e+00 -3.05509958e-02
-1.73513982e-02 -2.74167508e-01 -8.26407075e-01 9.54096854e-01
2.91563451e-01 -3.24962914e-01 -6.55123591e-01 -1.22626460e+00
4.03930485e-01 9.57255885e-02 7.79790878e-01 -1.27618027e+00
4.83828992e-01 -3.19296867e-02 3.09751272e-01 -2.34496757e-01
7.12832510e-01 -6.81100607e-01 2.67857462e-01 1.34864822e-01
-3.05860251e-01 1.36928231e-01 5.38338311e-02 9.52270746e-01
-9.87151265e-02 4.03871506e-01 2.63848156e-01 -3.20387274e-01
-1.03740013e+00 4.27760512e-01 -2.66800433e-01 4.06692594e-01
1.28688526e+00 -5.33712447e-01 -6.08216763e-01 -6.29211545e-01
-1.05005097e+00 5.87689877e-01 4.17837560e-01 1.04669392e+00
3.41300339e-01 -1.19368005e+00 -7.98672616e-01 -6.27379864e-02
4.05317813e-01 1.12084262e-01 7.37540007e-01 7.67964721e-01
-3.70827138e-01 3.93077552e-01 -6.25328898e-01 -7.87350297e-01
-1.01439404e+00 5.31298161e-01 4.31040108e-01 -2.13705301e-01
-7.30580211e-01 7.35031843e-01 9.21868384e-01 -4.89343733e-01
1.23489842e-01 -5.21096438e-02 -1.31509332e-02 1.28047913e-01
4.10011083e-01 3.67633700e-01 -1.65704384e-01 -5.38814604e-01
-2.07856089e-01 7.20368207e-01 -2.16064453e-01 7.18432665e-02
1.36861074e+00 -2.37041801e-01 2.84200728e-01 6.14630282e-01
1.11486769e+00 -7.06887305e-01 -1.77583802e+00 -3.49054575e-01
-3.44176918e-01 -5.26107371e-01 -1.37135774e-01 -5.98389626e-01
-7.51106799e-01 1.00034475e+00 2.31297195e-01 -6.58271015e-02
1.05915105e+00 1.54999062e-01 7.30945945e-01 5.96476912e-01
5.59504926e-01 -1.11496007e+00 5.27597547e-01 4.61049229e-01
8.83525014e-01 -1.43356359e+00 -1.01653136e-01 -3.49811763e-01
-8.38798463e-01 8.22977722e-01 9.94311452e-01 -1.55321106e-01
5.65989375e-01 -2.59461496e-02 -3.47889245e-01 -2.86839575e-01
-1.01677179e+00 8.18163007e-02 6.86880797e-02 7.84432292e-01
1.52767211e-01 -7.18172342e-02 2.97457635e-01 6.03733182e-01
9.81375128e-02 3.54319990e-01 4.22446430e-01 9.13117945e-01
-4.18692641e-02 -7.18622863e-01 -4.81651351e-02 3.76573712e-01
-7.66263157e-02 2.60935873e-01 -1.33649617e-01 8.58090162e-01
1.60887763e-01 6.14489615e-01 2.67030358e-01 -5.26623607e-01
4.96161580e-01 2.56400675e-01 5.87947965e-01 -6.26737714e-01
-5.76931715e-01 -4.12388355e-01 1.29317075e-01 -1.07972157e+00
-4.19836879e-01 -9.52837527e-01 -4.09489065e-01 -4.40943718e-01
3.38729024e-01 -3.49482298e-01 5.44981360e-01 9.84145701e-01
6.42660081e-01 6.39378488e-01 2.09916994e-01 -1.39028502e+00
-3.48080039e-01 -9.28558350e-01 -2.89744556e-01 8.25491190e-01
2.15127170e-01 -9.85648155e-01 -2.18829468e-01 5.54703355e-01] | [8.648500442504883, 0.5409883856773376] |
ba33bcc0-c72a-46e2-b2a6-ca385756e8ae | multi-view-class-incremental-learning | 2306.09675 | null | https://arxiv.org/abs/2306.09675v1 | https://arxiv.org/pdf/2306.09675v1.pdf | Multi-View Class Incremental Learning | Multi-view learning (MVL) has gained great success in integrating information from multiple perspectives of a dataset to improve downstream task performance. To make MVL methods more practical in an open-ended environment, this paper investigates a novel paradigm called multi-view class incremental learning (MVCIL), where a single model incrementally classifies new classes from a continual stream of views, requiring no access to earlier views of data. However, MVCIL is challenged by the catastrophic forgetting of old information and the interference with learning new concepts. To address this, we first develop a randomization-based representation learning technique serving for feature extraction to guarantee their separate view-optimal working states, during which multiple views belonging to a class are presented sequentially; Then, we integrate them one by one in the orthogonality fusion subspace spanned by the extracted features; Finally, we introduce selective weight consolidation for learning-without-forgetting decision-making while encountering new classes. Extensive experiments on synthetic and real-world datasets validate the effectiveness of our approach. | ['Zhigang Zeng', 'Cheng Lian', 'Kenji Kawaguchi', 'Junwei Chen', 'Tianqi Wang', 'Depeng Li'] | 2023-06-16 | null | null | null | null | ['class-incremental-learning', 'multi-view-learning', 'incremental-learning'] | ['computer-vision', 'computer-vision', 'methodology'] | [ 3.35663259e-01 -2.14414999e-01 -4.23644900e-01 -2.26170167e-01
-5.83191037e-01 -5.83172798e-01 5.08742571e-01 1.75546408e-01
-1.57508090e-01 6.52478635e-01 1.98024958e-01 8.00258741e-02
-3.52587849e-01 -5.81657767e-01 -4.77496386e-01 -9.09840286e-01
-3.70391421e-02 2.24547192e-01 3.03465605e-01 1.97231919e-01
3.72418672e-01 3.90174985e-01 -1.99073851e+00 7.18464494e-01
5.97883999e-01 6.66833997e-01 2.40682602e-01 3.75738531e-01
-8.39577541e-02 7.89501429e-01 -3.02290201e-01 -1.96802109e-01
3.30277324e-01 -2.55172879e-01 -5.91713428e-01 3.92300010e-01
5.40973604e-01 -2.13766590e-01 -9.29012448e-02 8.71175945e-01
6.31020904e-01 2.57396013e-01 4.61527020e-01 -1.43826437e+00
-3.57085973e-01 3.47906739e-01 -6.90374315e-01 3.79896849e-01
1.85226992e-01 -3.27770784e-02 7.74521708e-01 -1.54385173e+00
9.22335505e-01 1.03778172e+00 3.16600233e-01 7.22010314e-01
-1.32544768e+00 -6.92563057e-01 8.69019866e-01 4.98103291e-01
-1.11532068e+00 -6.14776373e-01 9.40922558e-01 -4.10561174e-01
5.28669119e-01 1.43268019e-01 6.62495077e-01 1.26071489e+00
4.30706561e-01 1.01115048e+00 1.22874403e+00 -2.89707720e-01
4.42145735e-01 2.69932538e-01 3.49385709e-01 6.05112314e-01
5.70213974e-01 8.74178261e-02 -1.12124288e+00 -1.17597148e-01
2.70191014e-01 5.48524082e-01 -2.80999243e-01 -1.14945149e+00
-1.37260723e+00 5.08752584e-01 2.57300641e-02 2.90200263e-01
-4.53084379e-01 -4.21327978e-01 4.48682040e-01 7.45826423e-01
4.81933743e-01 5.59012592e-02 -5.48985958e-01 1.49072155e-01
-7.63753057e-01 -6.02097884e-02 2.90113270e-01 7.61114717e-01
6.71680927e-01 4.24586348e-02 -1.17922835e-01 7.01810420e-01
4.15855553e-03 4.23588216e-01 8.51005554e-01 -8.71591091e-01
6.10131741e-01 7.97945082e-01 -1.60877809e-01 -8.01576734e-01
-2.59314328e-01 -7.37671316e-01 -9.17191207e-01 4.30590928e-01
-7.72105306e-02 6.81023765e-03 -6.94538355e-01 1.83936608e+00
5.56645930e-01 1.37717649e-01 3.72438908e-01 4.12328333e-01
4.44653422e-01 4.18186247e-01 -1.68719530e-01 -9.01856422e-01
1.01392925e+00 -9.17026818e-01 -5.16700089e-01 -1.73038557e-01
3.60765576e-01 -3.42482358e-01 6.57627761e-01 8.48417759e-01
-8.34237516e-01 -9.30759132e-01 -1.19560099e+00 4.93048966e-01
-2.81962812e-01 -1.09965287e-01 4.19974715e-01 4.33511496e-01
-6.68288052e-01 4.71353590e-01 -8.34114790e-01 -7.13427886e-02
5.48738539e-01 1.89890742e-01 -7.54027188e-01 -3.52822095e-01
-6.75626099e-01 4.68986362e-01 3.88759464e-01 -1.00726791e-01
-1.10374606e+00 -6.46273196e-01 -5.22907257e-01 -4.71619889e-02
6.97878420e-01 -7.32376099e-01 7.73302197e-01 -9.30484414e-01
-1.07450020e+00 4.80585039e-01 -3.00975561e-01 -1.83601230e-01
4.10875320e-01 -4.49923903e-01 -5.97013414e-01 1.09150194e-01
1.44734040e-01 3.70433331e-01 1.46097958e+00 -1.63313437e+00
-9.44576740e-01 -8.49437833e-01 -4.75134030e-02 5.86040080e-01
-5.36135554e-01 -6.59633577e-01 -1.41581103e-01 -6.36427224e-01
7.88094401e-01 -8.71692002e-01 -8.23261216e-02 -1.08339652e-01
5.54542020e-02 -7.66722783e-02 9.78680015e-01 -2.35034063e-01
1.41363192e+00 -2.27090478e+00 5.84240913e-01 -1.43857179e-02
4.26624656e-01 2.29586795e-01 -1.19821563e-01 3.81903112e-01
-4.61991906e-01 -2.59906292e-01 4.20942530e-02 -3.75725031e-01
-7.07918286e-01 1.59190193e-01 -5.22281051e-01 2.72775114e-01
-1.94725364e-01 5.15758872e-01 -1.04348326e+00 -3.17830056e-01
3.87008116e-02 5.20435907e-02 -4.92168486e-01 2.44140342e-01
3.62611204e-01 5.44624507e-01 -3.69417012e-01 7.36456454e-01
7.17939973e-01 -1.77065000e-01 3.01240385e-01 -3.32035482e-01
-8.03012587e-03 -4.69043374e-01 -1.57317495e+00 1.80487835e+00
-3.75157207e-01 8.80547613e-02 -1.53629109e-01 -1.14692700e+00
5.20772338e-01 3.25069010e-01 5.98572433e-01 -5.04377961e-01
-2.86362588e-01 -9.93249193e-02 -1.56475529e-01 -5.06159008e-01
3.81417270e-03 -2.79299527e-01 5.85917532e-02 6.75526738e-01
4.29014832e-01 4.49927479e-01 1.12000279e-01 3.17834705e-01
9.45886016e-01 1.12552628e-01 6.06845617e-01 4.80893068e-02
7.16036797e-01 -4.29453790e-01 9.24753845e-01 8.82017314e-01
-4.33120340e-01 4.59811598e-01 4.70254347e-02 -7.20683098e-01
-4.75156039e-01 -1.39447427e+00 1.39453650e-01 1.27332056e+00
1.30016148e-01 -3.31100851e-01 -1.84834480e-01 -1.22192097e+00
-1.58546399e-02 6.66467488e-01 -6.72314346e-01 -6.30039692e-01
-3.90888542e-01 -6.84682012e-01 -2.94434935e-01 4.34114695e-01
3.56257290e-01 -8.39176118e-01 -1.00646687e+00 3.11263412e-01
-2.03252882e-01 -6.63667798e-01 -2.30715409e-01 4.89061117e-01
-1.31294084e+00 -1.21361959e+00 -5.78853190e-01 -5.34292519e-01
9.12313521e-01 9.63563859e-01 8.34119499e-01 -3.60810488e-01
-1.11866482e-01 8.33458543e-01 -3.88493598e-01 -2.82930732e-01
-2.01242089e-01 2.84724813e-02 4.77746934e-01 5.27293861e-01
8.27001482e-02 -7.55996585e-01 -4.86551434e-01 2.05334395e-01
-9.21906352e-01 2.44531885e-01 6.82744324e-01 9.65215564e-01
7.93253481e-01 1.70974538e-01 8.19328010e-01 -1.12068105e+00
3.68060410e-01 -5.11369705e-01 -2.49900788e-01 6.24395788e-01
-9.79721487e-01 1.64305881e-01 5.46413183e-01 -6.64237976e-01
-1.25045431e+00 9.37771201e-02 5.04936099e-01 -7.88581073e-01
8.98902789e-02 4.17010218e-01 -2.75515795e-01 2.54703343e-01
4.95704770e-01 5.16121209e-01 5.81616946e-02 -3.00841451e-01
5.32248735e-01 3.39400411e-01 3.17396581e-01 -3.86877388e-01
8.85705829e-01 7.10729063e-01 -7.69658387e-02 -3.47617060e-01
-1.10300601e+00 -5.46621978e-01 -9.96487021e-01 -5.56451678e-01
5.11237383e-01 -1.13235962e+00 -3.76374573e-01 5.26152909e-01
-5.55328488e-01 3.03911269e-01 -6.33068681e-01 4.52980340e-01
-3.66968393e-01 3.70044917e-01 9.28105935e-02 -7.55006492e-01
-2.10085675e-01 -8.42562139e-01 7.54684567e-01 1.67719036e-01
1.00457110e-01 -6.25736892e-01 2.48704016e-01 3.49334687e-01
1.72764853e-01 -6.33787811e-02 8.93885791e-01 -7.14587927e-01
-4.49350536e-01 -2.10779011e-01 2.32383907e-01 4.05761689e-01
4.26287055e-01 -4.94873375e-01 -1.17967844e+00 -9.57826853e-01
4.67586309e-01 -3.60100448e-01 1.07168770e+00 -4.97907139e-02
1.08731449e+00 -1.67579189e-01 -4.94360864e-01 3.06544065e-01
1.36880362e+00 4.83464152e-01 -5.88085176e-03 1.98179960e-01
6.15355492e-01 5.87245107e-01 7.03517079e-01 6.77296758e-01
2.75375783e-01 2.88696498e-01 4.40352112e-01 2.77020514e-01
-1.97689235e-01 -2.22549364e-01 4.04869020e-01 1.21351695e+00
-9.81058478e-02 -1.90626234e-01 -4.55000669e-01 4.03467208e-01
-1.96369076e+00 -1.27737820e+00 5.50727904e-01 2.71444988e+00
3.99906307e-01 3.90607446e-01 -2.18731701e-01 2.85998166e-01
5.72793961e-01 2.80920744e-01 -1.00699472e+00 1.65034920e-01
6.77601108e-03 -1.28146723e-01 4.93805250e-03 1.69436261e-01
-1.13839209e+00 4.87018585e-01 5.62647963e+00 5.33132970e-01
-9.88999128e-01 2.91763633e-01 6.18965864e-01 -4.72241193e-01
-2.72813350e-01 1.32030576e-01 -7.83222437e-01 1.83372259e-01
5.98198593e-01 -2.71087646e-01 3.96982312e-01 8.32091630e-01
-2.74894655e-01 -2.38221109e-01 -1.28018963e+00 1.17767775e+00
5.04553080e-01 -1.02141380e+00 5.56568921e-01 -1.01099975e-01
6.57151699e-01 -1.35420039e-01 2.29517445e-01 6.41424358e-01
1.84448157e-02 -1.67624205e-01 5.78071654e-01 7.58737504e-01
6.42989814e-01 -7.82960594e-01 4.08665091e-01 6.92129076e-01
-1.27971137e+00 -6.90336406e-01 -2.44937494e-01 1.85608685e-01
-1.26270860e-01 5.67455590e-01 -4.47120339e-01 9.96888220e-01
5.50147891e-01 8.51262450e-01 -1.04625666e+00 6.71368301e-01
7.37940669e-02 4.28656459e-01 1.72442988e-01 6.90476894e-01
-2.24077836e-01 6.31379755e-03 6.46671772e-01 4.48782206e-01
3.89400870e-01 -4.24141623e-02 2.53775328e-01 6.13060482e-02
1.90456420e-01 -1.26416534e-01 -9.22285914e-01 3.81150156e-01
5.16315758e-01 1.15655208e+00 -7.18376994e-01 -4.92520660e-01
-5.89687765e-01 1.18341911e+00 5.08846223e-01 3.77107501e-01
-5.02849162e-01 5.01135923e-02 2.81880736e-01 -1.83205366e-01
4.52343553e-01 -7.38049224e-02 8.25871378e-02 -1.68627453e+00
1.27003640e-01 -9.25194502e-01 7.61755288e-01 -5.38483500e-01
-1.34757042e+00 4.95630383e-01 1.15065180e-01 -1.79696488e+00
-1.24911211e-01 -2.00878769e-01 -2.50939190e-01 2.67982274e-01
-1.38793445e+00 -9.92177427e-01 -2.40728587e-01 6.52213275e-01
1.08571565e+00 -7.37126470e-01 7.56724179e-01 1.92432255e-01
-4.90729779e-01 4.50178564e-01 1.76061884e-01 -4.46833640e-01
8.97106051e-01 -8.96650016e-01 -7.64681175e-02 7.73303330e-01
6.12230420e-01 5.74467301e-01 2.71497786e-01 -6.52388692e-01
-1.53456891e+00 -1.12335491e+00 5.65787733e-01 -5.28993070e-01
1.78191125e-01 -3.46663356e-01 -1.01124036e+00 5.97528398e-01
-4.74573532e-03 1.91752359e-01 1.02284431e+00 1.70768484e-01
-5.67183971e-01 -5.36154985e-01 -9.51182187e-01 4.72286791e-01
1.20104575e+00 -4.09888536e-01 -6.97224140e-01 -4.45607305e-02
6.00171745e-01 1.98468920e-02 -4.73364860e-01 6.85789406e-01
7.81971335e-01 -1.13658512e+00 1.01821470e+00 -7.75849700e-01
7.69601688e-02 -4.24657583e-01 -2.90047020e-01 -1.34321547e+00
-6.72228992e-01 -1.04707450e-01 -4.44851398e-01 1.19308937e+00
9.72270966e-02 -6.35708332e-01 5.95059037e-01 -2.24330071e-02
3.37589011e-02 -8.47628474e-01 -1.24156535e+00 -6.85251296e-01
-2.49688342e-01 -2.52139747e-01 3.16333205e-01 1.00676727e+00
-1.69748679e-01 5.99611044e-01 -4.97345328e-01 3.04552436e-01
8.67048085e-01 5.77514052e-01 6.73672676e-01 -1.46768415e+00
-4.36217904e-01 -9.27375481e-02 -2.83094794e-01 -5.49136639e-01
-1.22458622e-01 -1.00813830e+00 -2.82445729e-01 -1.24447751e+00
6.14245117e-01 -1.28727719e-01 -1.00356662e+00 4.47243243e-01
-4.98900145e-01 -1.59059778e-01 3.57892483e-01 5.72624683e-01
-9.80457902e-01 7.15735972e-01 1.05727386e+00 -2.62239963e-01
-2.59197295e-01 2.75559485e-01 -7.33981252e-01 7.46141315e-01
3.76498073e-01 -6.00920320e-01 -1.09446180e+00 -2.96351135e-01
2.40312845e-01 2.69208938e-01 -4.04268689e-02 -1.42204130e+00
3.28349590e-01 -1.38851956e-01 6.86325729e-01 -9.05593634e-01
1.74851432e-01 -1.05511916e+00 1.06769793e-01 5.45719564e-01
-2.62738168e-01 1.76245794e-01 -5.70926517e-02 1.18936753e+00
-1.08548701e-01 -1.13294400e-01 6.58355713e-01 -3.07279348e-01
-8.69955719e-01 4.12337780e-01 -3.98517251e-01 -9.38135013e-03
1.34650767e+00 -1.93273187e-01 -1.18212469e-01 -2.36090776e-02
-1.20124984e+00 2.41276398e-01 2.03359172e-01 6.91113770e-01
8.60280395e-01 -1.44006205e+00 -4.45831239e-01 7.51800478e-01
5.20882964e-01 -1.02030024e-01 5.47688186e-01 4.46806192e-01
3.01643699e-01 4.11628410e-02 -3.41150939e-01 -6.75910652e-01
-1.41474652e+00 1.16221690e+00 -1.48180202e-01 -5.07410824e-01
-5.67490757e-01 7.65602231e-01 4.28554684e-01 -2.75413543e-01
2.43681327e-01 1.61704585e-01 -5.78921795e-01 7.24816203e-01
6.59424186e-01 4.85109687e-01 1.74900338e-01 -3.13306481e-01
-2.60555953e-01 4.52318668e-01 -5.22881031e-01 -1.93545744e-01
1.35011566e+00 -5.25275469e-01 2.25516424e-01 1.29044461e+00
8.38267028e-01 -1.96830053e-02 -1.41295874e+00 -5.69970846e-01
-6.05244264e-02 -4.71808583e-01 -3.04072015e-02 -7.03084528e-01
-1.14304686e+00 8.29529881e-01 1.13421381e+00 -1.84037402e-01
1.28235769e+00 -2.56894261e-01 4.05532628e-01 6.42283678e-01
7.96287239e-01 -1.10240960e+00 5.63077509e-01 2.78435439e-01
7.44773149e-01 -1.20597255e+00 3.40148568e-01 -4.34187241e-02
-7.87881076e-01 1.03653800e+00 8.01214099e-01 3.23431492e-01
8.75513732e-01 -1.45815164e-01 -6.41403124e-02 -1.72580361e-01
-1.29926813e+00 5.58135472e-02 6.87973201e-02 6.08102560e-01
-1.05169982e-01 -2.26955339e-01 1.13716863e-01 4.63329017e-01
4.36260402e-01 3.90786678e-02 3.77368718e-01 1.43059468e+00
-4.22672123e-01 -1.14020956e+00 -1.46497667e-01 5.43524027e-01
5.15550142e-03 9.94676650e-02 -1.00288978e-02 3.85495752e-01
4.78147000e-01 6.40908420e-01 -7.20742345e-02 -4.63812709e-01
3.43123108e-01 5.95102191e-01 5.17506421e-01 -8.75535846e-01
-1.77095234e-01 1.78223595e-01 -4.65927869e-01 -5.06866395e-01
-5.89155853e-01 -1.03461587e+00 -7.81643689e-01 4.60082620e-01
-2.14373469e-01 7.93357939e-02 2.17217326e-01 8.40949178e-01
6.20884001e-01 7.51474619e-01 1.19057655e+00 -6.81509972e-01
-7.73897290e-01 -5.61668098e-01 -5.48342228e-01 3.62089694e-01
4.62624878e-01 -1.02461708e+00 -4.45510656e-01 2.59195715e-01] | [9.81478214263916, 3.42516827583313] |
a02c39ab-79d5-4875-80eb-bcd9da2ab0d6 | omnidirectional-dso-direct-sparse-odometry | 1808.02775 | null | http://arxiv.org/abs/1808.02775v1 | http://arxiv.org/pdf/1808.02775v1.pdf | Omnidirectional DSO: Direct Sparse Odometry with Fisheye Cameras | We propose a novel real-time direct monocular visual odometry for
omnidirectional cameras. Our method extends direct sparse odometry (DSO) by
using the unified omnidirectional model as a projection function, which can be
applied to fisheye cameras with a field-of-view (FoV) well above 180 degrees.
This formulation allows for using the full area of the input image even with
strong distortion, while most existing visual odometry methods can only use a
rectified and cropped part of it. Model parameters within an active keyframe
window are jointly optimized, including the intrinsic/extrinsic camera
parameters, 3D position of points, and affine brightness parameters. Thanks to
the wide FoV, image overlap between frames becomes bigger and points are more
spatially distributed. Our results demonstrate that our method provides
increased accuracy and robustness over state-of-the-art visual odometry
algorithms. | ['Daniel Cremers', 'Jörg Stückler', 'Hidenobu Matsuki', 'Lukas von Stumberg', 'Vladyslav Usenko'] | 2018-08-08 | null | null | null | null | ['monocular-visual-odometry'] | ['robots'] | [-2.67415851e-01 5.42229749e-02 -1.48188815e-01 -1.56400546e-01
5.94480410e-02 -6.07122481e-01 7.35973120e-01 -5.62337160e-01
-6.74781919e-01 5.59692621e-01 2.17235968e-01 -2.17846021e-01
1.54694393e-01 -5.79094470e-01 -6.58257365e-01 -5.60973287e-01
4.86062258e-01 5.90693533e-01 3.83099914e-01 -1.88852817e-01
3.32611740e-01 7.20654666e-01 -1.43929040e+00 -7.67582953e-01
7.10163414e-01 5.17873406e-01 5.85734844e-01 6.27444208e-01
3.83409739e-01 6.65905237e-01 -3.00864965e-01 -7.26886094e-02
6.66437328e-01 1.47163153e-01 -3.82712066e-01 4.79900330e-01
8.78854871e-01 -6.65853143e-01 -1.00079370e+00 1.35910475e+00
4.22437906e-01 2.26065293e-01 2.41750255e-01 -9.23287928e-01
-5.40822446e-01 -4.49394196e-01 -5.67409098e-01 -1.40860438e-01
8.18378866e-01 2.91661590e-01 6.58454359e-01 -1.14366949e+00
1.20203006e+00 1.02890253e+00 4.36594427e-01 2.78910756e-01
-1.09510100e+00 -2.93225944e-01 -5.58231249e-02 1.17090173e-01
-1.64862704e+00 -4.50247049e-01 5.33509970e-01 -4.62118298e-01
9.52268779e-01 -2.10021570e-01 8.56879771e-01 4.95369345e-01
2.33109981e-01 2.26729035e-01 8.06747198e-01 -4.62470204e-01
2.68678553e-02 9.70253348e-02 -1.60015419e-01 9.17203069e-01
7.25716352e-01 2.12694183e-01 -4.81522471e-01 1.60371527e-01
1.22271812e+00 4.55496132e-01 -6.96648777e-01 -1.39665055e+00
-1.41878271e+00 6.06773078e-01 4.50550914e-01 -3.65693510e-01
-1.83041334e-01 -8.03124011e-02 -2.19586298e-01 1.39620349e-01
6.00824989e-02 4.78822798e-01 -9.32698548e-02 -2.24846408e-01
-5.54087341e-01 1.13357328e-01 5.32539487e-01 1.40408683e+00
1.21675098e+00 1.51893049e-01 6.63298845e-01 7.09403753e-01
3.78066987e-01 1.09528935e+00 2.79814929e-01 -1.27459705e+00
4.49609458e-01 6.43683374e-01 4.31023031e-01 -8.56628656e-01
-4.67563689e-01 -1.44948155e-01 -5.92726171e-01 6.29413605e-01
2.82401770e-01 -2.08394602e-01 -9.09735203e-01 1.30740929e+00
4.57704633e-01 -1.97575927e-01 1.49651125e-01 1.30601859e+00
6.56438589e-01 4.11783844e-01 -9.89096701e-01 2.06646211e-02
1.17346823e+00 -8.78803849e-01 -5.85882723e-01 -6.54648066e-01
3.18065763e-01 -7.54702449e-01 7.37544179e-01 4.53668058e-01
-9.34687376e-01 -3.79276931e-01 -1.47105956e+00 -4.39922869e-01
1.15929022e-02 9.79456529e-02 2.83882022e-01 4.74368513e-01
-1.17403054e+00 -1.68903232e-01 -7.09521472e-01 -4.17037189e-01
-2.45378718e-01 5.04904509e-01 -7.98600554e-01 -5.48871040e-01
-7.19730556e-01 1.32663608e+00 1.37607967e-02 -2.57195562e-01
-7.11278915e-01 -3.37918431e-01 -1.41846812e+00 -1.59824505e-01
3.76713395e-01 -9.21556532e-01 1.01362848e+00 -2.65401185e-01
-1.86591184e+00 1.01936257e+00 -4.65789378e-01 -2.76312113e-01
4.17709708e-01 -3.40288788e-01 -8.18514451e-02 1.86763778e-01
-2.83092186e-02 6.22116089e-01 6.10734582e-01 -1.39709973e+00
-5.11398375e-01 -6.93513572e-01 3.48223001e-01 8.81204665e-01
2.22395077e-01 -4.44405586e-01 -8.09235573e-01 2.47800797e-02
7.73046196e-01 -1.36133289e+00 -3.01821858e-01 2.99051583e-01
-2.00270340e-01 6.09190524e-01 1.02909529e+00 -3.50080878e-01
7.53344595e-01 -2.04767466e+00 3.06070685e-01 4.73726392e-02
3.59367013e-01 2.16192171e-01 3.84520888e-01 1.09719768e-01
3.28894615e-01 -7.78842211e-01 3.39716434e-01 -5.60088277e-01
-2.53585994e-01 5.88231087e-01 -1.66375116e-01 1.09928620e+00
-6.63904309e-01 5.44987202e-01 -1.05126536e+00 -3.35783437e-02
8.13623607e-01 5.55146575e-01 -5.08352876e-01 2.48730049e-01
4.51407999e-01 5.26498735e-01 1.59051508e-01 4.96376157e-01
1.06323397e+00 -1.29784837e-01 2.24391013e-01 -3.00318673e-02
-4.93246555e-01 4.69515145e-01 -1.57335687e+00 1.90191138e+00
-4.50575978e-01 8.14999580e-01 2.24748641e-01 -3.55315655e-01
1.17218065e+00 -8.60199779e-02 2.78319269e-01 -7.21638381e-01
-9.52744782e-02 2.28820652e-01 -3.58844221e-01 -1.52275339e-01
1.05303097e+00 1.13940492e-01 2.28948474e-01 8.48113969e-02
1.15885913e-01 -7.58008718e-01 4.95403409e-02 1.60713181e-01
6.89772546e-01 2.55647480e-01 8.92088234e-01 -2.32747003e-01
4.56198901e-01 -4.17535603e-02 5.45998216e-01 4.38008487e-01
-3.07341162e-02 9.63073015e-01 -2.20153984e-02 -5.06466031e-01
-1.31696725e+00 -1.30024350e+00 -3.13593209e-01 1.00070134e-01
1.02129447e+00 -2.37044513e-01 -3.04224908e-01 -2.52815157e-01
9.86895561e-02 1.40642062e-01 -1.83950350e-01 6.03506975e-02
-5.70314527e-01 -2.50381589e-01 7.24438354e-02 1.95078015e-01
5.92374206e-01 -2.30287820e-01 -7.98941731e-01 -6.11735210e-02
-1.60537079e-01 -1.41035557e+00 -6.08743370e-01 6.75232112e-02
-9.99306619e-01 -1.15997124e+00 -8.08461189e-01 -6.20605707e-01
9.79555130e-01 1.22845912e+00 9.65181530e-01 -4.26019520e-01
1.45167217e-01 4.51975405e-01 1.08790986e-01 -1.07061870e-01
1.01238929e-01 -3.10917556e-01 4.89080697e-01 -1.93799034e-01
3.44829053e-01 -4.54783022e-01 -6.99187398e-01 7.65468001e-01
-5.26440561e-01 2.53929347e-01 2.00372562e-01 8.41088772e-01
3.92242610e-01 -8.02678347e-01 -5.12840688e-01 -1.55227661e-01
-5.72194234e-02 -1.90541148e-01 -1.18057144e+00 -2.97016174e-01
-6.48011565e-01 1.53249102e-02 2.80759186e-01 -2.13665962e-01
-9.38509822e-01 2.49447122e-01 3.56220514e-01 -5.45112014e-01
1.48808599e-01 4.09489032e-03 -2.31086388e-01 -5.17517447e-01
6.86739206e-01 2.23528460e-01 3.99525970e-01 -3.05973411e-01
3.42280984e-01 6.60427392e-01 7.70706773e-01 2.09414549e-02
8.16604972e-01 1.27150774e+00 4.07739162e-01 -1.20523143e+00
-3.54567051e-01 -1.09778881e+00 -9.87644017e-01 -1.24738939e-01
6.09182596e-01 -1.37741351e+00 -6.90620482e-01 4.13461924e-01
-1.12097633e+00 -6.08930103e-02 -1.64192483e-01 1.08999336e+00
-5.75270832e-01 7.80145347e-01 -2.61152446e-01 -5.91822982e-01
1.39615297e-01 -1.52566886e+00 1.08553886e+00 2.57231981e-01
1.28432671e-02 -9.83267426e-01 4.51301366e-01 4.67982106e-02
-1.19631425e-01 -1.92667469e-01 2.57955701e-03 3.65941375e-01
-9.73849714e-01 -2.44594477e-02 -2.50662327e-01 1.32712558e-01
9.32115987e-02 -1.68390945e-01 -7.48215556e-01 -4.41007227e-01
-5.74253201e-02 4.95203994e-02 5.83266616e-01 4.95159507e-01
9.07155946e-02 -1.33150846e-01 -3.52126539e-01 1.33862162e+00
1.59713137e+00 2.13459194e-01 9.57691789e-01 7.25396276e-01
9.04350162e-01 2.32669309e-01 6.80540979e-01 5.01283228e-01
8.99792314e-01 1.35796797e+00 1.00807858e+00 -1.84356734e-01
4.24670540e-02 -3.58425945e-01 4.60222244e-01 5.71778655e-01
-3.38469267e-01 -3.15641961e-03 -7.60255456e-01 7.17125058e-01
-1.66684937e+00 -7.73841023e-01 -1.59594923e-01 2.65671659e+00
3.21843207e-01 -2.71528780e-01 -1.40176907e-01 -2.88354661e-02
5.14038563e-01 3.82859141e-01 -3.66792887e-01 -1.89182177e-01
-2.83035845e-01 -4.34000999e-01 1.26944065e+00 1.06637812e+00
-7.60990918e-01 7.27892876e-01 6.43757868e+00 -1.68192998e-01
-1.29914021e+00 -1.84245005e-01 -4.91335332e-01 -2.34562069e-01
-2.05082476e-01 2.68097758e-01 -1.17071342e+00 2.34409437e-01
3.91704708e-01 3.36951278e-02 6.50833249e-01 1.06793356e+00
1.18427925e-01 -5.07536471e-01 -6.58728302e-01 1.48898971e+00
3.65042895e-01 -1.22317314e+00 -2.26191714e-01 7.11493731e-01
1.10732043e+00 6.08612537e-01 -4.11851332e-02 -2.33620644e-01
6.31123185e-01 -5.85423529e-01 7.03501701e-01 3.04848522e-01
9.87880409e-01 -3.73212516e-01 5.29885650e-01 4.45059478e-01
-1.11237276e+00 -1.43883005e-01 -8.63006771e-01 -3.21005821e-01
5.67884624e-01 6.49572790e-01 -1.05400288e+00 5.92665493e-01
6.12888992e-01 9.25186396e-01 -3.09694767e-01 1.06894839e+00
-3.56579721e-01 -3.39004904e-01 -6.25976980e-01 4.06742066e-01
1.56464890e-01 -6.42945170e-01 8.15060794e-01 7.79615521e-01
6.92871928e-01 -3.59707139e-02 4.01194505e-02 2.45422453e-01
6.34061918e-02 -3.31860632e-01 -1.20697749e+00 9.37578022e-01
4.59821373e-01 8.40649784e-01 6.28929771e-03 -4.41546947e-01
-7.56106615e-01 9.94437337e-01 9.81039256e-02 4.35864121e-01
-4.30930525e-01 -2.48067439e-01 1.21906805e+00 2.61920273e-01
3.71333659e-01 -8.26197386e-01 -3.50472838e-01 -1.91547132e+00
1.50894076e-01 -5.58548808e-01 -3.09245922e-02 -1.14092672e+00
-2.09152624e-01 2.48140931e-01 -8.29137042e-02 -1.85886502e+00
-4.11975622e-01 -8.17462623e-01 -6.06178716e-02 8.14787686e-01
-1.66197348e+00 -8.85953248e-01 -9.17039096e-01 7.17884123e-01
3.51280570e-01 1.53893065e-02 5.12428761e-01 9.40202624e-02
2.67627567e-01 -3.95188183e-02 4.41983491e-01 9.73790884e-03
8.42624545e-01 -1.24777687e+00 4.24105972e-01 1.20359075e+00
2.13784337e-01 8.94261777e-01 8.29034150e-01 -3.10235292e-01
-1.69832242e+00 -6.14086747e-01 9.06140506e-01 -6.65202558e-01
5.18743694e-02 -2.61524260e-01 -4.26210999e-01 9.94459927e-01
1.01669366e-03 3.75750989e-01 -1.87133923e-01 -2.20510304e-01
-3.60163361e-01 -1.83834374e-01 -1.07881463e+00 7.15072632e-01
1.21884513e+00 -5.43580472e-01 -5.81735849e-01 -1.68139398e-01
3.26328993e-01 -1.12255836e+00 -5.12542725e-01 1.57984391e-01
9.15879488e-01 -1.21004319e+00 1.27836275e+00 2.71844417e-01
-2.19414532e-01 -8.60346556e-01 -2.77365714e-01 -1.47061205e+00
-2.35148191e-01 -7.50434756e-01 -1.86988831e-01 3.78837079e-01
-2.16423422e-01 -9.12581444e-01 6.48734987e-01 1.73621461e-01
-2.14460924e-01 -1.43903820e-02 -1.01878035e+00 -6.65388346e-01
-7.21465170e-01 -1.01372458e-01 4.16720718e-01 7.96522319e-01
3.42855752e-01 2.94319779e-01 -7.70897448e-01 6.28498256e-01
8.05540800e-01 -1.11638822e-01 1.58480942e+00 -1.31087160e+00
-1.11931451e-01 1.74990058e-01 -1.01911712e+00 -2.27300310e+00
-2.54765451e-01 -3.29966962e-01 -1.75126158e-02 -1.39942360e+00
-1.23032540e-01 -1.05084263e-01 4.30075616e-01 -1.17538109e-01
1.65413067e-01 4.39240187e-01 2.76910901e-01 4.84569579e-01
-3.13590199e-01 3.85437101e-01 1.43035531e+00 1.21018980e-02
-2.49530777e-01 -6.51091114e-02 -1.99113190e-01 8.54727268e-01
1.83572799e-01 -2.70436138e-01 -4.54110086e-01 -6.69769585e-01
2.23631307e-01 2.33314887e-01 3.92619163e-01 -9.72965240e-01
4.02486086e-01 -3.62716556e-01 1.96615815e-01 -7.13620305e-01
7.14620173e-01 -9.78900194e-01 3.73830765e-01 4.40422028e-01
4.74027544e-01 4.00352001e-01 -2.68057734e-01 4.57111120e-01
-2.64724821e-01 -1.31432012e-01 9.98926103e-01 -2.48560593e-01
-9.26633477e-01 3.60817432e-01 -1.83973834e-01 -2.84140825e-01
8.26446593e-01 -7.70012379e-01 -4.79876369e-01 -7.87263691e-01
-3.55838656e-01 -1.13730013e-01 1.53046548e+00 3.29692662e-01
7.95907259e-01 -1.22930562e+00 -1.41346738e-01 7.05974042e-01
4.77327585e-01 3.98153782e-01 7.35020414e-02 1.01062655e+00
-9.85188782e-01 8.07364404e-01 -3.50223571e-01 -1.07037199e+00
-1.38271165e+00 5.43721318e-01 5.22936702e-01 -4.50983122e-02
-9.87428844e-01 2.32598349e-01 5.50232768e-01 -7.22609878e-01
-1.62051946e-01 -4.47808087e-01 -6.54600784e-02 -5.67292631e-01
5.77195048e-01 5.14948249e-01 -1.92197021e-02 -1.08597898e+00
-3.97358447e-01 1.08355784e+00 4.19648409e-01 -4.44251537e-01
9.79983747e-01 -9.23688293e-01 1.98400944e-01 7.62600824e-02
1.02689266e+00 6.16093099e-01 -1.78481746e+00 -3.95902693e-01
-5.18217027e-01 -1.02213824e+00 9.96558443e-02 -5.46983406e-02
-7.42566884e-01 8.43643546e-01 3.85620326e-01 -3.13690931e-01
7.33933508e-01 -2.95577258e-01 3.48239183e-01 5.87017655e-01
8.99300516e-01 -8.65911126e-01 -2.64146328e-01 1.03764701e+00
5.46792805e-01 -1.20417643e+00 4.74258035e-01 -3.60544145e-01
-5.51806629e-01 1.23040891e+00 4.03837740e-01 -2.30236724e-01
3.55039477e-01 1.77316532e-01 3.65321606e-01 -4.30839509e-02
-3.28784585e-01 -2.31363028e-01 1.89497992e-01 9.94777977e-01
-1.29025072e-01 -2.73497015e-01 4.11834158e-02 -5.42842388e-01
-9.13388282e-02 -5.32139242e-02 1.07826805e+00 8.67628217e-01
-6.47426724e-01 -7.89980114e-01 -7.26802051e-01 -3.24717164e-01
2.90525425e-02 1.17242955e-01 1.13011718e-01 8.62265229e-01
-1.80707484e-01 6.18529677e-01 2.82813162e-01 -2.20587477e-01
4.22523856e-01 -3.91152382e-01 7.21072078e-01 -5.10175109e-01
2.04720184e-01 -1.67280555e-01 7.99549147e-02 -6.85091674e-01
-4.16592628e-01 -5.64785182e-01 -1.11115730e+00 -4.75069642e-01
-2.03547373e-01 -2.03647524e-01 1.07937050e+00 7.25251973e-01
3.62066656e-01 -3.00249189e-01 5.77007473e-01 -1.24441719e+00
-3.83308679e-01 -7.77743995e-01 -6.94672465e-01 2.16846868e-01
9.79284406e-01 -1.03610563e+00 -5.28718174e-01 8.70663971e-02] | [7.710915565490723, -2.2315902709960938] |
b7838695-3659-4e37-a900-42738092c51e | improving-block-based-compensated-wavelet | 2212.04330 | null | https://arxiv.org/abs/2212.04330v1 | https://arxiv.org/pdf/2212.04330v1.pdf | Improving block-based compensated wavelet lifting by reconstructing unconnected pixels | This paper presents a new approach for improving the visual quality of the lowpass band of a compensated wavelet transform. A high quality of the lowpass band is very important as it can then be used as a downscaled version of the original signal. To adapt the transform to the signal, compensation methods can be implemented directly into the transform. We propose an improved inversion of the block-based motion compensation by processing unconnected pixels by a reconstruction method. We obtain a better subjective visual quality while furthermore saving up to 2.6% of bits for lossless coding. | ['André Kaup', 'Jürgen Seiler', 'Wolfgang Schnurrer'] | 2022-12-08 | null | null | null | null | ['motion-compensation'] | ['computer-vision'] | [ 6.20199561e-01 -7.75867179e-02 -1.44972861e-01 -2.84283608e-01
-5.10414004e-01 -1.74693361e-01 1.02658831e-02 9.99012496e-03
-5.88642180e-01 8.76215100e-01 3.10145885e-01 -1.81064382e-01
2.81647861e-01 -9.10655856e-01 -6.00282311e-01 -5.80463171e-01
1.39014497e-01 -4.67049956e-01 8.71240258e-01 -3.12238991e-01
2.33267099e-01 6.50159299e-01 -1.43145478e+00 6.75357997e-01
6.14348471e-01 1.07785380e+00 3.96787435e-01 1.03382218e+00
1.07461073e-01 6.05325103e-01 -5.51783264e-01 -1.18587628e-01
3.88263434e-01 -8.76803994e-01 -6.11921072e-01 2.64019996e-01
3.53724986e-01 -7.44390309e-01 -3.74918044e-01 1.08394003e+00
3.97815406e-01 -1.02848977e-01 2.54650205e-01 -6.99746907e-01
-9.07137915e-02 2.53468245e-01 -5.61385870e-01 1.43893436e-01
2.78149933e-01 -1.80575967e-01 6.58923328e-01 -7.83675611e-01
8.46804798e-01 1.04798138e+00 6.52124047e-01 2.02813640e-01
-1.48772216e+00 -3.00053239e-01 -5.48043966e-01 7.50051558e-01
-1.19729090e+00 -7.48382568e-01 1.01155043e+00 -5.63202761e-02
8.02767694e-01 5.34995794e-01 1.03198767e+00 2.11502805e-01
4.86409754e-01 3.00270587e-01 9.69311655e-01 -8.15231323e-01
2.03632176e-01 -1.81633055e-01 -2.59279490e-01 4.44399595e-01
3.07869613e-01 1.43825263e-01 -4.18023735e-01 1.98243290e-01
9.38328326e-01 -4.50820513e-02 -8.02308857e-01 -2.71610528e-01
-1.01183915e+00 4.61524516e-01 6.50853932e-01 6.20688021e-01
-4.50737983e-01 5.66873908e-01 4.14707839e-01 5.02706647e-01
5.62587261e-01 -1.98210534e-02 -2.19588131e-01 4.84774038e-02
-1.48069286e+00 -1.11466542e-01 4.68222201e-01 5.36259532e-01
9.13912058e-01 2.46537849e-01 -9.32752714e-03 7.32659459e-01
1.94572285e-01 4.46914285e-01 4.27769661e-01 -1.22979617e+00
2.32911795e-01 1.04728960e-01 1.95937961e-01 -1.03037560e+00
-2.89698094e-02 -2.91246533e-01 -9.33718204e-01 9.10712659e-01
2.26877064e-01 2.60803014e-01 -6.60251737e-01 1.21302068e+00
9.55544636e-02 -1.12479843e-01 5.90262152e-02 7.11415172e-01
3.40620011e-01 1.10104978e+00 -3.16290855e-01 -5.83378613e-01
1.19088435e+00 -8.52398634e-01 -1.10219407e+00 -9.42376927e-02
2.13425666e-01 -1.27645695e+00 7.58874834e-01 6.34732068e-01
-1.56934381e+00 -8.44829381e-01 -1.37088323e+00 -3.01324189e-01
2.05160946e-01 1.96647290e-02 -1.32499307e-01 5.47473073e-01
-1.36507034e+00 9.72616136e-01 -6.76876962e-01 5.29903881e-02
1.52053446e-01 1.90381333e-01 -4.77067292e-01 -2.50969440e-01
-8.63221407e-01 8.32209289e-01 4.35217828e-01 -1.16336234e-01
-1.31962717e-01 -5.68049669e-01 -7.12233245e-01 3.87470663e-01
-1.19056016e-01 -5.20632744e-01 9.00718033e-01 -1.27627802e+00
-1.49286437e+00 7.10108817e-01 -4.71448392e-01 -5.52923143e-01
7.04999089e-01 -9.71044004e-02 -3.75817031e-01 8.57900262e-01
-3.51535112e-01 2.66747624e-01 1.50544631e+00 -1.10166943e+00
-6.22705638e-01 8.48725587e-02 -2.93766737e-01 -6.88406900e-02
-3.74642998e-01 -2.22011775e-01 -2.36330509e-01 -1.26567340e+00
4.56787109e-01 -3.67194742e-01 -3.02773993e-02 8.54305029e-01
1.55401662e-01 7.26118684e-01 1.15897906e+00 -1.23261261e+00
1.33214986e+00 -2.50934553e+00 7.22987158e-03 1.68028995e-02
4.22833301e-02 4.31564510e-01 -1.53910473e-01 3.40016663e-01
-2.63003379e-01 -5.07043302e-02 -4.10391480e-01 -2.62173116e-01
-6.16403580e-01 1.70463189e-01 -2.86958188e-01 5.09247839e-01
-8.46020207e-02 4.54261899e-01 -6.94356263e-01 -4.82841700e-01
3.50584298e-01 8.67588162e-01 -7.58057952e-01 1.07675746e-01
3.90873194e-01 1.81318164e-01 1.41896844e-01 2.20466107e-01
1.13750565e+00 2.44554162e-01 6.78774016e-03 -7.79500127e-01
-2.41213739e-01 6.59875646e-02 -9.98661160e-01 1.39731836e+00
-5.58987856e-01 1.03077328e+00 4.04540807e-01 -8.31799448e-01
1.03356159e+00 4.71091807e-01 4.77333367e-01 -5.89112401e-01
-1.81931108e-01 4.25568014e-01 -3.18795383e-01 -3.83122742e-01
6.26193941e-01 -6.03832424e-01 4.53749657e-01 -4.63571288e-02
-2.75996238e-01 -4.80029613e-01 5.00419289e-02 -2.55149305e-01
1.06451821e+00 3.87214348e-02 7.59505808e-01 -2.04871476e-01
8.13985229e-01 -1.26422152e-01 6.23376608e-01 1.99857414e-01
-2.05393359e-01 8.70183110e-01 9.25951302e-02 -3.82934242e-01
-1.44686663e+00 -9.04765308e-01 -4.78801392e-02 4.34189826e-01
1.09884068e-02 -2.49371201e-01 -6.37364149e-01 -1.21826485e-01
-3.98352981e-01 5.95271826e-01 -2.30205636e-02 -2.66196370e-01
-9.70290244e-01 -2.49322563e-01 2.55353659e-01 3.54704767e-01
9.89400268e-01 -5.89354634e-01 -9.17405605e-01 5.39530814e-01
-4.81870651e-01 -9.80594516e-01 -4.96854037e-01 1.23888478e-01
-1.51810622e+00 -7.59890437e-01 -1.06497610e+00 -8.88501525e-01
6.50227427e-01 4.72083330e-01 8.93292010e-01 5.53953528e-01
-5.72738722e-02 2.26671947e-03 -5.43143570e-01 3.05958003e-01
-8.52173567e-01 -5.44568956e-01 -5.56128085e-01 7.50435814e-02
-3.40264738e-01 -7.17501819e-01 -8.78696263e-01 2.35328779e-01
-1.12514913e+00 1.34830877e-01 4.91577238e-01 9.71167743e-01
6.32031024e-01 4.89322394e-01 1.09355435e-01 -4.37184781e-01
2.34317213e-01 3.12679410e-01 -7.43160367e-01 -1.69706628e-01
-6.35867655e-01 5.57956845e-02 1.03523779e+00 -3.69007230e-01
-1.10021567e+00 3.00334185e-01 -5.45194387e-01 -2.43975937e-01
2.10713848e-01 1.47528928e-02 -3.18322629e-02 -4.99328405e-01
4.13659126e-01 2.57797390e-01 6.09942572e-03 -7.47952044e-01
2.21411735e-01 4.93855089e-01 6.94217682e-01 1.74949393e-01
9.07787800e-01 8.44309032e-01 3.63483191e-01 -1.03514755e+00
7.96603039e-02 -3.60384256e-01 -6.75701141e-01 -2.88075805e-01
6.06531501e-01 -6.59964323e-01 -2.01653615e-01 4.12344933e-01
-1.34827054e+00 -3.04784089e-01 -7.22791612e-01 5.86575150e-01
-7.15904653e-01 8.46350193e-01 -8.08023095e-01 -3.92344713e-01
-2.95329779e-01 -9.92457151e-01 6.71505868e-01 -2.75011748e-01
-1.34235442e-01 -9.18553770e-01 -6.17084205e-02 -9.70590189e-02
6.34507656e-01 1.30797297e-01 1.00487435e+00 3.14439178e-01
-4.68230069e-01 -1.83194846e-01 -1.62038505e-01 7.17831135e-01
2.71735877e-01 -4.30290140e-02 -9.05658066e-01 -5.20340502e-01
5.03524780e-01 3.03703636e-01 1.17373192e+00 5.33320487e-01
7.94912696e-01 -4.62003917e-01 -1.80221200e-01 8.82842302e-01
1.87775016e+00 1.90715790e-01 1.17964852e+00 3.35860223e-01
1.00614861e-01 2.33842283e-01 5.81679940e-01 4.58873659e-01
-3.50682765e-01 8.85813594e-01 2.95546353e-01 -3.41933876e-01
-8.52973759e-01 -2.12796047e-01 4.76365179e-01 9.01175916e-01
-1.57406181e-01 -1.73238963e-01 -3.21829557e-01 4.94604617e-01
-1.42121303e+00 -1.24627614e+00 -4.28892136e-01 2.35888624e+00
9.17685270e-01 3.82168852e-02 -3.04241717e-01 1.06989682e+00
7.91555583e-01 3.51731658e-01 -1.28703430e-01 -5.77547431e-01
-8.11595246e-02 2.84185469e-01 8.01702440e-01 1.06921470e+00
-6.93610609e-01 5.10408580e-01 7.68975687e+00 9.63425040e-01
-1.30414724e+00 1.19567849e-01 4.07349348e-01 2.48627886e-01
-2.77710348e-01 2.00356841e-01 -3.01865518e-01 4.53972906e-01
6.43758535e-01 -2.47507289e-01 2.78787434e-01 5.19028723e-01
6.96724474e-01 -5.05668283e-01 -6.62085831e-01 9.65171695e-01
-5.61389178e-02 -1.15670431e+00 8.51951316e-02 2.08677771e-03
5.30312300e-01 -7.51484275e-01 -1.01336353e-01 -5.27862608e-01
-4.86883819e-01 -4.66820776e-01 9.25526440e-01 4.74556416e-01
1.08867919e+00 -8.00160229e-01 6.68838382e-01 1.55183956e-01
-1.47510576e+00 1.31840855e-01 -6.37488902e-01 -7.30880350e-02
5.33924758e-01 1.03849649e+00 -5.92750072e-01 3.22700232e-01
6.71884537e-01 7.43352413e-01 -2.32813448e-01 1.30359697e+00
-3.06639701e-01 6.33977950e-01 -2.47421965e-01 6.08967245e-01
-2.37346783e-01 -6.11043014e-02 8.05964887e-01 1.25414538e+00
7.89451063e-01 1.54829785e-01 -3.38759929e-01 4.13429826e-01
1.40738755e-01 1.05139583e-01 -4.51422900e-01 4.71375644e-01
1.01222336e-01 8.36111188e-01 -7.66014218e-01 -6.38425112e-01
-2.97166049e-01 1.63718677e+00 -4.52977002e-01 3.55914742e-01
-4.59226549e-01 -7.56682932e-01 3.69981170e-01 3.37302238e-01
7.91601777e-01 -2.11292595e-01 -1.84824377e-01 -1.04695868e+00
8.15320536e-02 -6.05314314e-01 1.36445612e-01 -9.67895746e-01
-6.58321023e-01 5.97104132e-01 -1.70983911e-01 -1.76826084e+00
-3.04411799e-01 -3.19304228e-01 -3.02891374e-01 9.85568106e-01
-1.79834974e+00 -5.80773473e-01 -2.95450091e-01 6.94664478e-01
5.72365880e-01 2.61004955e-01 7.01375604e-01 4.59375113e-01
3.05270106e-01 2.41586506e-01 3.71232629e-01 -2.79004186e-01
7.58642375e-01 -7.48260260e-01 1.09694727e-01 1.29723287e+00
-3.46724838e-01 -9.53891501e-02 1.03899944e+00 -5.56323767e-01
-9.40564096e-01 -8.45650852e-01 1.19559741e+00 7.09287763e-01
1.70231253e-01 2.70390064e-01 -1.05004728e+00 3.83680671e-01
3.74485791e-01 1.22160248e-01 3.05804610e-01 -1.06169522e+00
-3.87603454e-02 -5.88280320e-01 -1.68820870e+00 4.06440407e-01
6.14633799e-01 -3.67001772e-01 -5.98494411e-01 -1.39521211e-01
6.47325873e-01 -1.43397331e-01 -6.84140146e-01 1.69685796e-01
5.67477763e-01 -1.46744275e+00 1.27244782e+00 3.72094393e-01
4.96091932e-01 -4.72317249e-01 -3.22826833e-01 -1.30006361e+00
-4.92079437e-01 -6.82923615e-01 3.83024365e-02 8.93253267e-01
-7.50107318e-02 -5.60454309e-01 6.75741971e-01 -8.20103437e-02
1.63081154e-01 -2.03153715e-01 -1.21745288e+00 -8.08810472e-01
-3.98718327e-01 -2.49096975e-01 1.83244437e-01 5.07405639e-01
1.31193295e-01 -1.74782962e-01 -5.68813801e-01 3.79157178e-02
8.11239421e-01 -3.36527526e-02 2.62180924e-01 -9.20149028e-01
-4.00414348e-01 -1.72298864e-01 -7.00760126e-01 -1.40615499e+00
-3.46685857e-01 -6.02312028e-01 8.66174772e-02 -1.50311971e+00
-2.89010674e-01 9.68998075e-02 -4.77130972e-02 2.39292443e-01
1.46628276e-01 9.68109250e-01 4.82958198e-01 3.36310774e-01
4.08030480e-01 3.56107116e-01 1.37504864e+00 -2.21555054e-01
-1.91613615e-01 -3.30798998e-02 2.55938955e-02 8.29805315e-01
7.17076421e-01 -6.13920331e-01 -2.33658120e-01 -3.90380532e-01
4.68402430e-02 3.71786326e-01 3.37857157e-01 -1.33924127e+00
-7.17047229e-02 3.93031351e-02 5.27517855e-01 -6.60171866e-01
6.89511776e-01 -1.47666419e+00 3.59111726e-01 1.17720008e+00
-1.89896092e-01 -8.97301547e-03 1.66334614e-01 4.45582628e-01
-5.70234358e-01 -7.66191840e-01 1.33681118e+00 -1.35674611e-01
-6.92817748e-01 -4.03203517e-01 -7.42535055e-01 -5.47957897e-01
7.50095546e-01 -7.36525178e-01 8.52072164e-02 -7.09787011e-01
-9.36920404e-01 -6.89838409e-01 8.86656821e-01 -3.69244546e-01
1.16369176e+00 -1.30414414e+00 -5.51444590e-01 4.52399015e-01
-4.81774628e-01 -5.96562624e-01 1.60687640e-01 6.84979796e-01
-1.28506160e+00 -5.40044643e-02 -7.52581000e-01 -3.62454176e-01
-1.72902966e+00 6.02962494e-01 2.76627511e-01 -2.13206023e-01
-9.18186963e-01 3.74615699e-01 -4.32191133e-01 8.50733638e-01
-8.89800936e-02 -4.45769101e-01 -8.48653093e-02 -5.87237850e-02
9.03742492e-01 6.94843173e-01 1.37368462e-03 -7.24487841e-01
-1.78978533e-01 1.03724408e+00 4.94755059e-01 -5.58112383e-01
1.40888929e+00 -5.70328414e-01 -3.35122228e-01 2.51380146e-01
1.51126468e+00 4.68678653e-01 -1.25677907e+00 1.84599888e-02
-4.44450974e-01 -1.07235944e+00 4.92414445e-01 -2.95922995e-01
-1.29243255e+00 9.55197811e-01 7.73840010e-01 2.69442886e-01
1.91620505e+00 -5.91376066e-01 1.01122272e+00 -5.55012114e-02
3.40528041e-01 -9.05021489e-01 7.19983503e-02 3.54156829e-03
9.83738601e-01 -6.14039481e-01 4.24745023e-01 -7.75184572e-01
7.90471807e-02 1.83393514e+00 -4.09012467e-01 -1.83155909e-01
7.01200843e-01 4.76591378e-01 -9.08403285e-03 5.33106625e-01
-3.77076834e-01 -5.90108708e-02 1.83341622e-01 6.43308997e-01
4.35225457e-01 -4.46538806e-01 -8.33554626e-01 -2.88272083e-01
1.03616312e-01 3.94676775e-01 7.37969458e-01 1.02384770e+00
-9.19417918e-01 -1.22440350e+00 -8.98423672e-01 -1.54950172e-01
-4.92486030e-01 -1.68504462e-01 1.37509197e-01 3.71783882e-01
3.17945331e-02 9.77679491e-01 -7.36285225e-02 -1.47865236e-01
3.02501023e-01 -1.26279797e-02 5.49911618e-01 -7.18075410e-02
-2.95393258e-01 5.02771974e-01 -7.72512779e-02 -7.98370242e-01
-6.27812743e-01 -1.30107105e-01 -1.02876461e+00 -5.68397164e-01
-6.17501475e-02 -1.44267142e-01 5.52066147e-01 3.43506277e-01
-3.38918180e-04 5.80527604e-01 7.17034638e-01 -1.02830136e+00
-2.15010017e-01 -5.44087648e-01 -6.11473083e-01 4.44702148e-01
9.19162750e-01 -5.93108032e-03 -5.97790360e-01 7.71860182e-01] | [11.404601097106934, -2.3172309398651123] |
3e483472-b5d0-48b3-acc7-f5f711ced728 | adamae-adaptive-masking-for-efficient | 2211.09120 | null | https://arxiv.org/abs/2211.09120v1 | https://arxiv.org/pdf/2211.09120v1.pdf | AdaMAE: Adaptive Masking for Efficient Spatiotemporal Learning with Masked Autoencoders | Masked Autoencoders (MAEs) learn generalizable representations for image, text, audio, video, etc., by reconstructing masked input data from tokens of the visible data. Current MAE approaches for videos rely on random patch, tube, or frame-based masking strategies to select these tokens. This paper proposes AdaMAE, an adaptive masking strategy for MAEs that is end-to-end trainable. Our adaptive masking strategy samples visible tokens based on the semantic context using an auxiliary sampling network. This network estimates a categorical distribution over spacetime-patch tokens. The tokens that increase the expected reconstruction error are rewarded and selected as visible tokens, motivated by the policy gradient algorithm in reinforcement learning. We show that AdaMAE samples more tokens from the high spatiotemporal information regions, thereby allowing us to mask 95% of tokens, resulting in lower memory requirements and faster pre-training. We conduct ablation studies on the Something-Something v2 (SSv2) dataset to demonstrate the efficacy of our adaptive sampling approach and report state-of-the-art results of 70.0% and 81.7% in top-1 accuracy on SSv2 and Kinetics-400 action classification datasets with a ViT-Base backbone and 800 pre-training epochs. | ['Vishal M. Patel', 'Motilal Agrawal', 'Mehdi Nikkhah', 'Ali Gholami', 'Naman Patel', 'Wele Gedara Chaminda Bandara'] | 2022-11-16 | null | http://openaccess.thecvf.com//content/CVPR2023/html/Bandara_AdaMAE_Adaptive_Masking_for_Efficient_Spatiotemporal_Learning_With_Masked_Autoencoders_CVPR_2023_paper.html | http://openaccess.thecvf.com//content/CVPR2023/papers/Bandara_AdaMAE_Adaptive_Masking_for_Efficient_Spatiotemporal_Learning_With_Masked_Autoencoders_CVPR_2023_paper.pdf | cvpr-2023-1 | ['action-classification'] | ['computer-vision'] | [ 3.53368223e-01 2.19771892e-01 -3.59888017e-01 -7.36729056e-02
-9.81343865e-01 -1.86433703e-01 5.75270474e-01 -4.13118005e-01
-6.37614250e-01 6.98086500e-01 2.68098533e-01 6.20713644e-02
4.09464896e-01 -4.60020810e-01 -1.13541996e+00 -9.00177062e-01
-3.16340685e-01 5.04331887e-02 2.86121279e-01 -2.43034363e-02
8.77925232e-02 2.05202445e-01 -1.80848849e+00 1.14399493e+00
2.76074350e-01 1.21775866e+00 5.01022339e-01 7.51592100e-01
-3.57249975e-02 1.24846017e+00 -9.02822435e-01 -1.40288249e-01
3.94882262e-01 -7.41443872e-01 -5.26136398e-01 9.50063858e-03
5.57161629e-01 -5.14732420e-01 -5.14058232e-01 6.33200645e-01
6.05615497e-01 1.74372584e-01 6.29310548e-01 -1.26735759e+00
-5.36563337e-01 9.27626610e-01 -4.87197042e-01 3.80033046e-01
3.77502255e-02 5.57647109e-01 6.96356416e-01 -1.20625055e+00
7.47888505e-01 1.11421502e+00 5.49153388e-01 1.25089276e+00
-1.25827146e+00 -6.48858309e-01 3.36125791e-01 4.76308256e-01
-1.27466154e+00 -6.96073592e-01 6.19935870e-01 -3.54041338e-01
1.24149728e+00 2.37819284e-01 7.74919391e-01 1.72655618e+00
2.76453346e-01 1.22467351e+00 1.04669249e+00 -3.14124286e-01
5.70881903e-01 1.07510693e-01 -5.58856010e-01 7.27408946e-01
-2.81392574e-01 3.62438411e-01 -1.20560896e+00 3.77681553e-02
8.70327711e-01 1.76020954e-02 -1.44664824e-01 -2.09192321e-01
-1.26019263e+00 7.83783197e-01 3.12196374e-01 -1.98860783e-02
-6.68904662e-01 7.54158080e-01 3.60875219e-01 2.94324696e-01
3.50077391e-01 1.49480738e-02 -2.77144909e-01 -3.69616717e-01
-1.16716254e+00 2.27956101e-01 1.65750906e-01 6.97806537e-01
6.65826261e-01 8.35348427e-01 -7.02682018e-01 6.92779422e-01
1.20741867e-01 4.91376728e-01 6.92833602e-01 -1.15081644e+00
4.71939147e-01 9.83153284e-02 2.69661486e-01 -3.14868033e-01
-3.70022468e-02 -2.73292243e-01 -6.00147009e-01 8.81498098e-01
2.20589459e-01 -1.41958743e-01 -1.34102011e+00 2.04968452e+00
2.55684197e-01 4.31121707e-01 1.61318317e-01 1.08741629e+00
6.67166531e-01 5.60829937e-01 1.89044058e-01 8.39421377e-02
1.23503911e+00 -8.30715358e-01 -6.82434678e-01 -3.22014332e-01
3.97704780e-01 -3.25413108e-01 1.23004270e+00 6.09021246e-01
-1.32228899e+00 -7.77563035e-01 -1.06176555e+00 1.10748380e-01
-1.34512201e-01 4.80286032e-01 3.60886395e-01 4.61801976e-01
-1.12380445e+00 9.22586381e-01 -1.10028231e+00 3.60691458e-01
9.16720629e-01 3.31467807e-01 -9.54050124e-02 2.75915056e-01
-1.10903716e+00 6.05988979e-01 1.77175894e-01 -4.07469608e-02
-1.80325150e+00 -6.24420404e-01 -8.19899201e-01 2.13744611e-01
2.22548798e-01 -3.17988127e-01 1.15158761e+00 -1.54118955e+00
-1.64888835e+00 6.46216512e-01 -2.77726710e-01 -9.18491125e-01
6.79805219e-01 -1.46032557e-01 -2.48181686e-01 5.94698906e-01
2.55488139e-02 1.41408241e+00 1.56764150e+00 -1.07872355e+00
-5.31798720e-01 6.49395287e-02 -1.89149141e-01 2.24792823e-01
-2.17251539e-01 -2.99152583e-01 6.52906373e-02 -9.50158477e-01
-2.39678115e-01 -7.99820542e-01 -6.10316731e-02 1.10628881e-01
-4.89884540e-02 -9.72823799e-02 7.40747631e-01 -6.47223413e-01
7.79620051e-01 -2.38778710e+00 1.04874656e-01 -1.68825775e-01
2.95809805e-01 2.11777613e-02 -3.91355306e-01 2.49643460e-01
-2.33143568e-01 -1.67038798e-01 -1.77632168e-01 -6.88540816e-01
3.57531663e-03 6.18952140e-02 -7.03197777e-01 4.68946397e-01
4.26979631e-01 8.74226630e-01 -7.55122602e-01 -3.53853762e-01
3.54554713e-01 5.43242574e-01 -6.47928715e-01 1.15514994e-01
-4.98547822e-01 3.91561419e-01 4.38407622e-02 5.82857966e-01
3.70536655e-01 1.58947781e-02 -1.77270293e-01 1.03295080e-01
1.95618913e-01 2.58041650e-01 -1.09128618e+00 1.90601051e+00
-3.16363364e-01 9.83802676e-01 1.24373525e-01 -1.00538707e+00
7.36087918e-01 5.94823420e-01 6.47273362e-01 -6.98430955e-01
6.34128228e-02 -6.90036416e-02 -2.45891958e-01 -4.02534902e-01
6.18157327e-01 6.82330355e-02 1.33472979e-01 2.85650373e-01
1.73416555e-01 3.38495791e-01 -1.99393526e-01 3.69202167e-01
1.15908813e+00 5.58160365e-01 -8.44482034e-02 -3.10018826e-02
-8.79760832e-02 -3.44786122e-02 5.29745877e-01 9.95413482e-01
-3.56824696e-01 7.36565888e-01 4.27355021e-01 -4.51673746e-01
-9.67271924e-01 -1.25142121e+00 2.37529442e-01 1.09106028e+00
-7.87492767e-02 -3.32180649e-01 -9.78322208e-01 -6.41096711e-01
-1.07718922e-01 7.59647846e-01 -9.23042715e-01 -4.30122077e-01
-5.19654572e-01 -1.67691335e-01 6.29696310e-01 6.95087790e-01
6.10651433e-01 -1.62810540e+00 -1.12293303e+00 3.93932551e-01
-3.00626308e-01 -1.13801515e+00 -3.53271872e-01 4.32588726e-01
-8.31779182e-01 -6.93041265e-01 -6.16173804e-01 -5.62539816e-01
5.66638470e-01 1.89311445e-01 1.01473022e+00 -1.16060108e-01
-2.97493935e-01 3.45159143e-01 -3.62299591e-01 -3.06774229e-01
-4.82862830e-01 -2.47686267e-01 1.69451699e-01 1.81537509e-01
3.17902058e-01 -4.55239117e-01 -7.68647432e-01 2.76118070e-01
-9.11761642e-01 1.06811062e-01 5.08373797e-01 1.01240730e+00
7.07502842e-01 -5.17088175e-01 6.06135130e-01 -3.92995387e-01
1.12501688e-01 -4.98509020e-01 -4.60874707e-01 -1.72140703e-01
-1.83536589e-01 1.39752641e-01 6.73969030e-01 -1.06847060e+00
-7.94885874e-01 3.53400260e-01 -8.04120973e-02 -1.24066448e+00
-1.56744406e-01 -2.92908326e-02 1.27567291e-01 1.32593557e-01
1.01336873e+00 5.55703223e-01 -1.68013796e-02 -1.67176738e-01
3.42297465e-01 5.83874464e-01 6.33096635e-01 -6.10688686e-01
3.52021962e-01 7.73992658e-01 -3.13267857e-01 -6.02266669e-01
-3.91117007e-01 1.66392289e-02 -1.76894516e-01 -4.49001104e-01
1.12360239e+00 -1.18233192e+00 -7.01279879e-01 3.30758989e-01
-8.56726587e-01 -9.54324663e-01 -8.06133211e-01 6.04357481e-01
-9.70894873e-01 2.96912771e-02 -5.85970283e-01 -8.48142505e-01
-2.53237277e-01 -1.31097317e+00 1.13247108e+00 -9.07317474e-02
-3.09641927e-01 -3.81098002e-01 -2.10397303e-01 1.83941096e-01
3.39911789e-01 1.35200173e-01 3.41820717e-01 -3.88999850e-01
-8.36016655e-01 1.37173146e-01 3.82773548e-01 3.81935745e-01
-1.06705554e-01 -2.35784933e-01 -1.46143401e+00 -4.32127923e-01
-3.56157832e-02 -6.65737092e-01 1.29544044e+00 7.60411441e-01
1.64169955e+00 -4.15110856e-01 -1.58905014e-01 7.08868086e-01
1.14129388e+00 1.10388435e-01 8.58626127e-01 2.77420342e-01
3.98933679e-01 2.38519892e-01 4.57953572e-01 7.42526412e-01
-1.14088625e-01 5.26139319e-01 8.25599492e-01 6.90680817e-02
-3.72479409e-01 -4.21282113e-01 9.02685463e-01 1.72537014e-01
1.69140205e-01 -1.83418289e-01 -3.39558810e-01 8.71321678e-01
-1.76625800e+00 -1.41607225e+00 4.68153894e-01 2.06567812e+00
9.80053067e-01 3.39637399e-01 2.78533429e-01 1.26314148e-01
7.22653210e-01 2.67441899e-01 -7.72567809e-01 -4.49069172e-01
-1.58611342e-01 4.49719042e-01 5.01413643e-01 3.80340397e-01
-1.05071080e+00 1.08090174e+00 6.42513227e+00 1.26239347e+00
-1.19620574e+00 3.13366324e-01 7.31387198e-01 -7.28004754e-01
-1.42674297e-01 -2.31426358e-01 -8.79834294e-01 6.75263882e-01
1.19324625e+00 6.41602635e-01 8.06122422e-01 9.57047880e-01
3.88465792e-01 -1.31215498e-01 -1.15413725e+00 1.03101134e+00
-9.97397974e-02 -1.80111206e+00 -3.08405440e-02 -1.06817447e-01
6.80280328e-01 1.91965446e-01 4.50016499e-01 4.73257303e-01
4.26721275e-01 -1.20676708e+00 1.35878503e+00 2.18241721e-01
1.13849962e+00 -4.39056426e-01 1.01562515e-01 2.60391235e-01
-1.05053389e+00 -2.28700057e-01 -4.57673490e-01 2.42929459e-02
-2.37996712e-01 2.53479958e-01 -8.81579697e-01 -2.78872326e-02
9.24016654e-01 5.96273780e-01 -2.01326177e-01 5.60716510e-01
-2.69671440e-01 1.00734031e+00 -3.03789914e-01 -1.39450073e-01
3.99285436e-01 2.85276264e-01 4.90455538e-01 1.14466012e+00
4.05381322e-02 -6.94718808e-02 -2.32550744e-02 1.13051331e+00
-1.43951982e-01 -5.08279681e-01 -4.64371383e-01 2.73425162e-01
5.90478063e-01 9.83748496e-01 -6.84273601e-01 -6.58196211e-01
-2.30795369e-01 1.17649138e+00 3.29667687e-01 5.30091047e-01
-1.17490315e+00 -2.22656876e-01 6.61986351e-01 2.67685112e-03
9.12216544e-01 3.78845609e-03 -1.97749257e-01 -1.09701347e+00
6.77393153e-02 -1.11941421e+00 2.48783261e-01 -9.09526050e-01
-8.78421664e-01 5.45228779e-01 1.17410704e-01 -1.47044456e+00
-4.87969011e-01 -4.51908588e-01 -4.76926357e-01 5.84305584e-01
-1.18476498e+00 -6.90139890e-01 -2.17669114e-01 8.48116159e-01
1.01072681e+00 -4.48924035e-01 7.24921346e-01 2.01468766e-02
-3.34947556e-01 7.44712770e-01 1.97387606e-01 2.99091637e-01
3.50692451e-01 -1.10313213e+00 4.73021477e-01 7.82017529e-01
4.98557568e-01 1.61501750e-01 5.52901983e-01 -5.21634638e-01
-1.13236320e+00 -1.11960411e+00 1.30977884e-01 -4.25164223e-01
2.40770549e-01 -7.29009569e-01 -7.40384758e-01 8.78150225e-01
3.94428521e-01 3.29445243e-01 2.96076685e-01 -5.45098603e-01
-3.95794898e-01 -9.45391506e-02 -1.25030184e+00 7.85348415e-01
1.03147161e+00 -5.98216116e-01 -4.48020935e-01 1.48727953e-01
5.56555450e-01 -5.48225939e-01 -5.80973327e-01 -4.97778691e-02
4.65770125e-01 -1.06432509e+00 1.00207186e+00 -6.01336837e-01
5.92486024e-01 -1.48156360e-01 -3.09237748e-01 -1.35566032e+00
-7.03025311e-02 -7.78562963e-01 -5.61144114e-01 6.65057063e-01
2.65274346e-01 -3.33228856e-01 1.25744712e+00 2.08296046e-01
-2.36806437e-01 -8.20795357e-01 -1.38078487e+00 -8.01626027e-01
-1.05664939e-01 -5.05957603e-01 4.03500378e-01 6.74935400e-01
-1.12070777e-01 -7.49657750e-02 -6.47922814e-01 3.75458859e-02
6.70227230e-01 -1.79834351e-01 5.66063821e-01 -4.21982080e-01
-5.13098300e-01 -3.68351549e-01 -2.89938509e-01 -1.23653090e+00
1.39892504e-01 -6.73107505e-01 2.34345317e-01 -1.19469321e+00
-1.05740666e-01 -1.75153121e-01 -3.31504464e-01 9.02493954e-01
-9.82145220e-02 3.36297452e-01 2.03308061e-01 1.96280658e-01
-4.94679451e-01 9.27624464e-01 9.62344050e-01 -4.09543842e-01
-2.16597721e-01 -2.25490883e-01 -2.29512230e-01 4.46394444e-01
9.15865064e-01 -7.84808517e-01 -5.23606300e-01 -4.76116270e-01
-1.29310176e-01 2.19898269e-01 6.73501313e-01 -1.15092492e+00
-2.11270675e-01 -2.09871665e-01 8.67808640e-01 -6.73413038e-01
8.23778629e-01 -6.51042640e-01 -4.51428294e-02 6.81191146e-01
-7.76878297e-01 -3.00985165e-02 4.08097446e-01 6.38787448e-01
3.91002521e-02 -2.90716767e-01 8.18246484e-01 -3.11998993e-01
-7.71302700e-01 1.52838826e-01 -8.85734737e-01 7.18929544e-02
9.63826895e-01 -5.09992898e-01 -1.69561073e-01 -4.78680640e-01
-9.60420310e-01 -1.33375421e-01 1.86621845e-01 3.53591263e-01
1.10260391e+00 -1.38394272e+00 -8.81123245e-01 3.53648812e-01
-1.60475940e-01 -9.58224535e-02 2.66514122e-01 5.64264774e-01
-2.53010064e-01 -6.78245211e-03 -3.95089090e-01 -8.85650277e-01
-1.15542316e+00 5.44763625e-01 5.37928522e-01 5.78412265e-02
-9.49645102e-01 1.03179634e+00 7.14243278e-02 1.69066265e-01
4.78155851e-01 -4.47568029e-01 8.93806145e-02 -2.34595746e-01
6.39852703e-01 1.92635119e-01 -1.41126681e-02 -2.40301415e-01
-3.39122087e-01 1.45609155e-01 9.52934176e-02 -5.94529033e-01
1.14841115e+00 1.80313215e-01 6.68350816e-01 3.34770918e-01
1.00551891e+00 -5.21758720e-02 -2.09842706e+00 -1.39943644e-01
-4.75333512e-01 -4.21208799e-01 9.20028239e-02 -6.97051942e-01
-1.28084862e+00 1.02962589e+00 8.79741192e-01 -4.65330891e-02
1.05275440e+00 -1.12027779e-01 6.55531049e-01 2.24151835e-01
1.36482418e-01 -1.19740105e+00 4.51481462e-01 -6.95076725e-03
8.21264327e-01 -1.01096106e+00 -1.70305029e-01 2.59664148e-01
-1.14554906e+00 9.14792717e-01 7.57533312e-01 -3.80171537e-01
2.53680050e-01 3.99932802e-01 -3.01879784e-03 1.43385632e-02
-1.21193135e+00 -6.74612597e-02 -7.04536662e-02 7.89344311e-01
-9.19285193e-02 -1.08543821e-01 2.10609943e-01 4.38998729e-01
-1.65576469e-02 3.15339938e-02 5.61475515e-01 1.02414310e+00
-4.89632189e-01 -5.28599620e-01 -4.53354567e-01 3.81044209e-01
-5.86486876e-01 -2.44672701e-01 -1.34648263e-01 6.23089850e-01
6.18160814e-02 8.55837524e-01 2.68341929e-01 -5.63167810e-01
-2.28232980e-01 2.42191553e-01 6.61422849e-01 -2.69609511e-01
-7.52863944e-01 1.11623943e-01 -9.02783126e-02 -8.29892278e-01
-3.90532017e-01 -6.39656484e-01 -1.56472945e+00 -1.88391805e-01
-1.15794763e-01 1.08048826e-01 6.19416714e-01 7.11363614e-01
6.71029508e-01 7.45997965e-01 6.61724210e-01 -1.09912825e+00
-7.69340336e-01 -9.87252891e-01 -2.58335203e-01 4.22854513e-01
6.35715663e-01 -8.09936762e-01 -3.76198947e-01 1.90470725e-01] | [9.305830955505371, 1.0508308410644531] |
f5374584-3192-49fe-b563-29bf6d8812d1 | video-summarization-through-human-detection | 1901.10713 | null | https://arxiv.org/abs/1901.10713v2 | https://arxiv.org/pdf/1901.10713v2.pdf | A Mobile Robot Generating Video Summaries of Seniors' Indoor Activities | We develop a system which generates summaries from seniors' indoor-activity videos captured by a social robot to help remote family members know their seniors' daily activities at home. Unlike the traditional video summarization datasets, indoor videos captured from a moving robot poses additional challenges, namely, (i) the video sequences are very long (ii) a significant number of video-frames contain no-subject or with subjects at ill-posed locations and scales (iii) most of the well-posed frames contain highly redundant information. To address this problem, we propose to \hl{exploit} pose estimation \hl{for detecting} people in frames\hl{. This guides the robot} to follow the user and capture effective videos. We use person identification to distinguish a target senior from other people. We \hl{also make use of} action recognition to analyze seniors' major activities at different moments, and develop a video summarization method to select diverse and representative keyframes as summaries. | ['Jane Yung-jen Hsu', 'Chih-Yuan Yang', 'Heeseung Yun', 'Srenavis Varadaraj'] | 2019-01-30 | null | null | null | null | ['person-identification'] | ['computer-vision'] | [ 3.54664356e-01 2.03262180e-01 -2.33162656e-01 -2.23585427e-01
-8.73618186e-01 -4.46634054e-01 1.59100816e-01 5.46932518e-02
-2.08359435e-01 1.02609575e+00 7.98172891e-01 7.26576388e-01
2.89415959e-02 -2.57967591e-01 -6.00393355e-01 -5.70326328e-01
-6.35949150e-02 4.19970632e-01 2.94179946e-01 6.92165596e-03
1.76271856e-01 5.32792032e-01 -1.87885940e+00 2.80406684e-01
8.56855214e-01 6.28907979e-01 4.79202509e-01 9.97597933e-01
3.76129061e-01 1.28330112e+00 -9.35173869e-01 4.64380756e-02
6.15965687e-02 -4.95614290e-01 -7.21675336e-01 7.11792886e-01
6.94404781e-01 -9.57828224e-01 -7.87930131e-01 1.11056721e+00
6.30997479e-01 5.95826745e-01 5.84205806e-01 -1.44147491e+00
-3.32706481e-01 3.91916305e-01 -5.42094171e-01 4.62926418e-01
1.58407009e+00 6.60420358e-02 4.49597329e-01 -5.96581042e-01
7.49834359e-01 1.41292155e+00 8.79127204e-01 6.25361085e-01
-8.01141024e-01 -5.40445089e-01 4.88473207e-01 3.93912286e-01
-1.51890719e+00 -7.84694850e-01 6.10211253e-01 -3.25208217e-01
6.26676619e-01 5.54934084e-01 1.12818682e+00 1.33993042e+00
-6.85104206e-02 1.22765517e+00 3.12454998e-01 -2.74571419e-01
1.48336411e-01 -1.32958055e-01 3.46053541e-01 7.44813025e-01
3.35300207e-01 -8.00600886e-01 -6.36351168e-01 -1.70527846e-01
6.50938392e-01 6.43598497e-01 -5.38832426e-01 2.11425364e-01
-1.47911310e+00 4.05771554e-01 -4.98669036e-02 4.08622742e-01
-6.92023754e-01 2.44447634e-01 1.77124113e-01 1.02854185e-02
3.16121817e-01 4.58335392e-02 -8.73277262e-02 -3.91292483e-01
-7.97201395e-01 7.58791804e-01 6.69048309e-01 1.57378936e+00
6.75886631e-01 -2.11116493e-01 -3.05254251e-01 6.35472059e-01
2.34013438e-01 7.97756016e-01 5.30796289e-01 -1.66818511e+00
6.96227849e-01 7.26889670e-01 5.70255339e-01 -1.29286623e+00
-4.90000129e-01 3.78692538e-01 -6.99099898e-01 -4.07239676e-01
1.88339502e-01 -3.99584353e-01 -4.32896882e-01 1.39108133e+00
3.23069572e-01 1.17974184e-01 3.83086200e-03 6.72138274e-01
1.11864185e+00 7.36583114e-01 -1.02746889e-01 -8.34153891e-01
1.44357884e+00 -7.99596429e-01 -1.00887752e+00 -4.11867052e-01
3.69062543e-01 -3.83658975e-01 7.57943928e-01 2.49286920e-01
-1.48062670e+00 -7.44547248e-01 -9.03880298e-01 1.96051300e-01
2.07631037e-01 4.19765860e-01 1.93334624e-01 3.73141825e-01
-1.19449902e+00 4.86995608e-01 -7.40685105e-01 -1.01505244e+00
2.24760190e-01 4.72257763e-01 -4.22613800e-01 -3.11290205e-01
-6.53645337e-01 4.10422146e-01 2.58058667e-01 -9.68549326e-02
-6.71809971e-01 -3.54619682e-01 -9.97386932e-01 -2.11888701e-01
1.88709676e-01 -5.82586586e-01 1.43383265e+00 -1.08025289e+00
-1.09632015e+00 5.15676200e-01 -3.21201861e-01 -1.09174661e-01
6.46804810e-01 -7.66861558e-01 -2.43592948e-01 8.82012188e-01
9.54079211e-01 6.30205572e-01 6.46407902e-01 -9.03650463e-01
-1.04642224e+00 -6.72186434e-01 -1.40508741e-01 7.48842597e-01
-4.97120887e-01 4.17187840e-01 -6.86573207e-01 -8.81600797e-01
3.55204701e-01 -1.00200522e+00 -3.10497791e-01 -1.53987527e-01
-4.79567498e-01 -2.12734222e-01 8.56884181e-01 -9.56769407e-01
1.27841198e+00 -1.85218501e+00 3.15280795e-01 -1.52204841e-01
1.94827467e-01 1.04733638e-01 2.85146058e-01 3.13533485e-01
1.76537663e-01 -4.74574447e-01 2.12705731e-01 -4.51404542e-01
-1.41401112e-01 1.48134112e-01 8.34523588e-02 5.62036335e-01
-3.99752438e-01 3.02188605e-01 -1.14363110e+00 -9.37975287e-01
4.35814142e-01 1.58360764e-01 -5.22017717e-01 4.58286971e-01
1.61262512e-01 7.35856473e-01 -8.77255261e-01 8.65591586e-01
3.27704012e-01 9.58938673e-02 3.65589857e-02 -1.65856425e-02
2.46005040e-02 -3.12488526e-01 -1.41300309e+00 1.60647738e+00
4.99886394e-01 6.78049088e-01 6.42043650e-02 -8.20152581e-01
5.25443256e-01 4.73362267e-01 1.06846023e+00 -6.46635145e-02
1.16466634e-01 -1.53129861e-01 -1.02703130e+00 -1.04419720e+00
7.27630973e-01 5.36290109e-01 -3.76497507e-01 3.71222109e-01
-1.95207372e-01 3.56930614e-01 5.20484388e-01 1.95680469e-01
1.58506775e+00 2.44348109e-01 4.47397918e-01 -8.46537277e-02
4.08701599e-01 2.21937317e-02 8.47903073e-01 9.65591133e-01
-5.87820292e-01 7.92269588e-01 1.12278290e-01 -5.22214413e-01
-5.14174223e-01 -8.63181472e-01 6.56758964e-01 1.24976683e+00
3.34955901e-01 -5.37555456e-01 -9.74908113e-01 -3.83788019e-01
-5.50348461e-01 4.40614462e-01 -1.52950957e-01 -6.29392499e-03
-1.03933156e+00 -5.17649233e-01 4.36617196e-01 7.01564610e-01
9.97000098e-01 -1.19415128e+00 -1.05351043e+00 3.42359453e-01
-1.03046811e+00 -1.14616156e+00 -7.55434394e-01 -3.86402905e-01
-5.62937498e-01 -1.15591264e+00 -1.38743687e+00 -1.13386059e+00
8.86812389e-01 7.98829019e-01 8.20847154e-01 -4.74128634e-01
-2.71871120e-01 1.13060093e+00 -6.15336835e-01 -1.91336334e-01
4.67488021e-02 -1.80434301e-01 5.64789712e-01 -2.16207087e-01
4.19882864e-01 -7.37399518e-01 -7.92188287e-01 3.13529581e-01
-3.17766845e-01 -4.18164618e-02 2.86829442e-01 4.30742726e-02
3.69016737e-01 4.34965283e-01 1.34423852e-01 -2.34160095e-01
4.03097808e-01 -2.71572828e-01 2.00593799e-01 3.44520062e-01
5.37939429e-01 -6.33990705e-01 2.72453070e-01 -5.98227859e-01
-9.11161661e-01 2.77704090e-01 2.21070096e-01 -4.61207211e-01
-4.11894441e-01 -3.22798610e-01 -5.38228631e-01 4.48256224e-01
5.24117768e-01 2.76205987e-01 -3.43987286e-01 -2.40064904e-01
-9.81728137e-02 6.63948119e-01 8.42440784e-01 -3.43928516e-01
2.61773944e-01 5.35740077e-01 -4.37035441e-01 -1.56594193e+00
-5.66415191e-01 -6.56049669e-01 -8.44070613e-01 -7.61361718e-01
1.10530901e+00 -1.14387751e+00 -1.08081257e+00 5.73965073e-01
-1.32250118e+00 -6.67456686e-02 -2.67262548e-01 7.39183247e-01
-9.17384624e-01 8.35508883e-01 -6.88286126e-01 -1.08396864e+00
-2.22801387e-01 -1.05130398e+00 1.24026990e+00 6.22601867e-01
-7.29466856e-01 -4.94061023e-01 -2.01698765e-01 6.95535600e-01
-4.13708508e-01 4.96840179e-01 6.38016015e-02 -4.01088297e-01
-4.96576995e-01 -3.90884846e-01 4.47001681e-02 7.27351159e-02
2.72499382e-01 -1.38619468e-01 -5.12334526e-01 -3.17566514e-01
-1.05321191e-01 1.02961443e-01 2.45101541e-01 9.52157438e-01
6.69018865e-01 -7.70360768e-01 -6.47938251e-01 1.00569159e-01
5.28685153e-01 5.12434542e-01 5.10549188e-01 1.94115832e-01
9.88339245e-01 6.91880584e-01 9.69577491e-01 8.52795243e-01
8.11017632e-01 7.00919330e-01 -1.02465436e-01 5.40196836e-01
1.29795164e-01 4.54805568e-02 1.06778598e+00 4.20785367e-01
-6.48213446e-01 -1.96895733e-01 -5.59612572e-01 3.93980145e-01
-2.57472563e+00 -1.55193543e+00 -2.59741284e-02 1.84298122e+00
3.03947210e-01 -2.16441169e-01 6.18400991e-01 1.61758497e-01
1.06771886e+00 1.35499313e-01 -4.38558429e-01 4.85427707e-01
2.38734439e-01 -5.85801601e-01 4.23272491e-01 1.16763934e-01
-1.29610384e+00 7.73491383e-01 6.22089481e+00 3.15785170e-01
-2.53918827e-01 -2.39871174e-01 4.16450620e-01 -2.13643700e-01
2.77082831e-01 -5.27598858e-01 -1.08169031e+00 6.09534383e-01
4.89570528e-01 1.35381028e-01 -2.74351556e-02 1.02459908e+00
4.55156922e-01 -4.87850070e-01 -1.18816781e+00 1.43197966e+00
7.76403964e-01 -8.41359258e-01 -8.09805319e-02 -1.73006937e-01
5.54954588e-01 -5.15049875e-01 -4.15663123e-01 3.92576337e-01
3.40024829e-01 -4.43818897e-01 9.43162560e-01 1.01918185e+00
4.17048961e-01 -7.81006038e-01 2.85479397e-01 4.43503022e-01
-1.49564254e+00 -3.76804501e-01 -4.04540777e-01 -4.98843677e-02
3.71018291e-01 6.56469464e-02 -6.47492826e-01 1.70008659e-01
1.31747890e+00 1.08119750e+00 -6.30345523e-01 8.12244713e-01
1.83901249e-03 -7.63049489e-03 -2.09512368e-01 -1.91124871e-01
-1.36252478e-01 -2.74957389e-01 8.70260715e-01 1.11119652e+00
7.25705504e-01 6.91200852e-01 8.76737237e-01 -7.87393153e-02
2.82649010e-01 -1.45730540e-01 -7.37572610e-01 1.85203254e-01
3.68858427e-01 6.32702231e-01 -9.63152707e-01 -5.72905302e-01
-4.02653337e-01 1.58455610e+00 -3.25618058e-01 4.20089811e-01
-9.15852368e-01 -3.38433981e-01 5.64359426e-01 4.30215120e-01
7.11021060e-03 -2.32977211e-01 4.16573405e-01 -1.37678158e+00
2.73739696e-01 -7.83185899e-01 5.48333645e-01 -1.26532340e+00
-5.51957488e-01 2.90615380e-01 4.06925052e-01 -1.51966858e+00
-6.37455523e-01 -7.18055516e-02 -4.76862907e-01 -1.58545688e-01
-5.43107152e-01 -1.05027425e+00 -7.45917678e-01 9.25706387e-01
1.30015075e+00 -3.59054327e-01 4.74501878e-01 3.92787546e-01
-8.81541431e-01 2.49198616e-01 -1.06100328e-01 -2.28233216e-03
6.33908331e-01 -1.06761730e+00 -1.09563395e-01 9.42366004e-01
-4.74930406e-01 4.88556296e-01 1.11321867e+00 -1.09328604e+00
-1.27251780e+00 -1.12844133e+00 9.50964689e-01 -5.28212965e-01
1.17645718e-01 9.79743451e-02 -2.30561599e-01 1.17966092e+00
-7.26556033e-02 -3.15104485e-01 6.97028935e-01 -4.65104103e-01
6.03846788e-01 -1.47929728e-01 -1.20121551e+00 8.89459670e-01
1.51909876e+00 -3.79560813e-02 -8.70814562e-01 6.97138906e-01
6.31690979e-01 -1.23379990e-01 -4.52262789e-01 1.73046514e-01
6.33285463e-01 -1.14565897e+00 1.24327826e+00 -9.51285064e-02
1.73429444e-01 -2.41735056e-01 -4.56114262e-01 -7.16394782e-01
-2.47769967e-01 -9.87685978e-01 -2.22412378e-01 1.43856883e+00
-4.82834160e-01 1.58930440e-02 9.32059050e-01 1.00559628e+00
-2.92291254e-01 1.49628907e-01 -6.81967497e-01 -2.82261848e-01
-1.02064657e+00 -1.29828095e-01 1.92171514e-01 5.60039759e-01
3.40369999e-01 1.25068009e-01 -7.58309603e-01 3.29351127e-01
7.30640411e-01 -3.96157920e-01 1.05460525e+00 -1.30677569e+00
-5.26323803e-02 7.35827610e-02 -7.20945120e-01 -1.18566239e+00
4.09652144e-02 4.07294184e-02 3.28818142e-01 -1.84800351e+00
4.78378028e-01 2.88132876e-01 3.21754813e-01 2.31676981e-01
-1.53607190e-01 1.92963272e-01 2.73654275e-02 5.68523765e-01
-1.37324762e+00 4.48503524e-01 7.20079482e-01 -1.09188534e-01
-6.60647571e-01 5.62701344e-01 -5.55557489e-01 1.37517810e+00
4.23028290e-01 -1.04719505e-01 -4.67031330e-01 -1.63167462e-01
-1.97965086e-01 6.59479380e-01 2.99219221e-01 -1.62084925e+00
2.32648313e-01 -2.14561418e-01 8.69753420e-01 -1.14392459e+00
5.70781946e-01 -8.23659956e-01 2.41573334e-01 6.58729017e-01
-3.05509996e-02 2.09015653e-01 -3.23708624e-01 6.77303016e-01
1.78850859e-01 -1.90057218e-01 5.30630291e-01 -6.25295818e-01
-9.27690148e-01 2.22902834e-01 -9.90998089e-01 -2.78915226e-01
1.43672121e+00 -6.54031813e-01 -4.22818288e-02 -7.65480816e-01
-9.22377467e-01 3.58399957e-01 4.72511679e-01 2.68187612e-01
5.56788266e-01 -1.30862391e+00 -3.92743260e-01 -1.14092559e-01
1.34571465e-02 2.91822582e-01 5.44371724e-01 6.35355473e-01
-8.84345233e-01 -5.40467612e-02 -2.53119022e-01 -5.44914305e-01
-1.78401935e+00 2.04941750e-01 -1.00065857e-01 2.07802087e-01
-1.01205254e+00 7.67256796e-01 1.58108115e-01 -4.99568656e-02
6.50566936e-01 -3.99478883e-01 -7.54120171e-01 6.20601416e-01
1.09427488e+00 1.20267427e+00 -6.72668457e-01 -1.24705279e+00
-3.30082029e-01 5.90845287e-01 2.31391527e-02 -6.23862557e-02
1.44922698e+00 -7.69435525e-01 1.44020259e-01 3.21730196e-01
9.18371737e-01 -3.02407295e-01 -1.56275630e+00 -2.47121796e-01
-9.80536267e-02 -4.55626518e-01 -6.41173840e-01 2.07048759e-01
-5.40218890e-01 1.38722792e-01 5.35807371e-01 -6.32627979e-02
1.13003612e+00 1.21463738e-01 8.25186312e-01 7.80650795e-01
5.96405625e-01 -1.46480501e+00 8.18820179e-01 2.16005892e-01
9.91134822e-01 -9.89669204e-01 4.06332999e-01 -4.11813498e-01
-9.64807868e-01 1.04625797e+00 5.73844314e-01 -2.43119299e-01
2.75998384e-01 -1.71284944e-01 -3.54775816e-01 -1.24055713e-01
4.54756990e-02 -2.05176353e-01 -2.33122289e-01 8.04985821e-01
-1.98092595e-01 -1.15286835e-01 5.71168773e-02 7.14396656e-01
-3.92714232e-01 6.49955571e-02 7.62069106e-01 1.43527865e+00
-1.01007688e+00 -3.57406229e-01 -8.78041625e-01 3.08902711e-01
-3.31382871e-01 6.15248024e-01 -2.32137904e-01 4.05950457e-01
2.46692926e-01 1.25469565e+00 -1.73310358e-02 -2.95417726e-01
6.00707114e-01 5.26747713e-03 4.55599546e-01 -6.62893116e-01
-1.51808724e-01 3.39709699e-01 6.14265762e-02 -6.86988235e-01
-7.69042373e-01 -1.48961353e+00 -1.02168858e+00 -3.04151982e-01
2.77263373e-01 -5.13798296e-02 -2.10775569e-01 8.11874390e-01
2.16408819e-02 3.94106895e-01 5.41472852e-01 -1.78781712e+00
5.08572049e-02 -8.31081688e-01 -8.22792113e-01 2.74033636e-01
4.03570771e-01 -5.00735283e-01 -9.02145803e-02 5.29206872e-01] | [8.152048110961914, 0.3767518699169159] |
7b1d5248-dc1b-4e82-a14e-f74778566e62 | spatial-gradient-consistency-for-unsupervised | 2302.10927 | null | https://arxiv.org/abs/2302.10927v1 | https://arxiv.org/pdf/2302.10927v1.pdf | Spatial gradient consistency for unsupervised learning of hyperspectral demosaicking: Application to surgical imaging | Hyperspectral imaging has the potential to improve intraoperative decision making if tissue characterisation is performed in real-time and with high-resolution. Hyperspectral snapshot mosaic sensors offer a promising approach due to their fast acquisition speed and compact size. However, a demosaicking algorithm is required to fully recover the spatial and spectral information of the snapshot images. Most state-of-the-art demosaicking algorithms require ground-truth training data with paired snapshot and high-resolution hyperspectral images, but such imagery pairs with the exact same scene are physically impossible to acquire in intraoperative settings. In this work, we present a fully unsupervised hyperspectral image demosaicking algorithm which only requires exemplar snapshot images for training purposes. We regard hyperspectral demosaicking as an ill-posed linear inverse problem which we solve using a deep neural network. We take advantage of the spectral correlation occurring in natural scenes to design a novel inter spectral band regularisation term based on spatial gradient consistency. By combining our proposed term with standard regularisation techniques and exploiting a standard data fidelity term, we obtain an unsupervised loss function for training deep neural networks, which allows us to achieve real-time hyperspectral image demosaicking. Quantitative results on hyperspetral image datasets show that our unsupervised demosaicking approach can achieve similar performance to its supervised counter-part, and significantly outperform linear demosaicking. A qualitative user study on real snapshot hyperspectral surgical images confirms the results from the quantitative analysis. Our results suggest that the proposed unsupervised algorithm can achieve promising hyperspectral demosaicking in real-time thus advancing the suitability of the modality for intraoperative use. | ['Tom Vercauteren', 'Jonathan Shapey', 'Oscar MacCormac', 'Conor Horgan', 'Muhammad Asad', 'Peichao Li'] | 2023-02-21 | null | null | null | null | ['demosaicking'] | ['computer-vision'] | [ 8.07823777e-01 -8.26525241e-02 -3.94524708e-02 -1.33622572e-01
-8.46947849e-01 -1.57861292e-01 1.04166470e-01 -5.61714359e-02
-5.97882926e-01 6.86121941e-01 4.59743068e-02 -3.59230727e-01
-6.14273787e-01 -5.82461894e-01 -5.65418959e-01 -1.29960406e+00
-6.67875335e-02 1.67187318e-01 -6.40864074e-01 -1.87279388e-01
-5.30809127e-02 6.33453667e-01 -1.51893878e+00 3.41547996e-01
1.34411466e+00 1.20043230e+00 5.98449588e-01 3.47935885e-01
2.90023208e-01 4.22869593e-01 -5.23145013e-02 2.33145654e-01
8.21432531e-01 -3.75702083e-01 -4.71038550e-01 2.56373048e-01
6.42612576e-01 -3.42998385e-01 -4.04035181e-01 1.38012528e+00
7.02340961e-01 1.20395999e-02 3.73048335e-01 -5.91166079e-01
-3.34779292e-01 4.68218207e-01 -5.05166650e-01 -3.83880511e-02
-4.62849401e-02 1.94674402e-01 4.35915560e-01 -6.76893950e-01
3.08814079e-01 2.63498366e-01 8.36868644e-01 3.58898073e-01
-1.16539586e+00 -5.36700249e-01 -3.39245945e-01 2.09691659e-01
-1.37891054e+00 -2.46515691e-01 1.04622543e+00 -3.95823181e-01
3.88296515e-01 5.79116464e-01 1.11629832e+00 6.27849936e-01
4.41111028e-01 4.17278945e-01 1.45782447e+00 -4.79405284e-01
1.17250204e-01 -4.21029329e-02 -3.55802506e-01 8.13634098e-01
3.99939716e-01 3.83234411e-01 -4.04985547e-01 1.01206698e-01
9.04070377e-01 4.67469543e-01 -9.15844738e-01 -5.30904710e-01
-1.33520496e+00 3.75893950e-01 1.00660026e+00 6.13691986e-01
-9.23408866e-01 3.17355655e-02 3.88208032e-02 1.17032193e-01
4.26816314e-01 7.00391531e-01 1.91580847e-01 3.01857620e-01
-1.12907720e+00 -4.69133109e-01 2.84801006e-01 4.29832518e-01
7.68807232e-01 5.70929423e-02 2.10124463e-01 9.47175026e-01
1.56005859e-01 5.71513593e-01 7.74561465e-01 -8.21817696e-01
1.73070267e-01 3.60008299e-01 -9.36516300e-02 -7.15157449e-01
-7.43914783e-01 -9.01570559e-01 -1.19834805e+00 5.38042963e-01
6.24027438e-02 -1.46001205e-01 -1.04275298e+00 1.18534768e+00
1.52250394e-01 2.98130363e-01 3.26338232e-01 1.45045829e+00
6.57626390e-01 3.94877076e-01 -2.27562532e-01 -5.54169118e-01
9.58484292e-01 -5.62798619e-01 -7.40605295e-01 -2.68527001e-01
6.11575723e-01 -5.07040679e-01 8.51757407e-01 7.28583336e-01
-8.23757887e-01 -8.69738776e-03 -1.43747544e+00 3.81617606e-01
1.11377247e-01 3.00643682e-01 9.08661902e-01 7.51839936e-01
-1.06252778e+00 6.42080367e-01 -1.00948489e+00 -1.40778825e-01
3.89252663e-01 5.13312042e-01 -5.82196593e-01 -3.19318622e-01
-9.30154562e-01 6.57162368e-01 4.37963486e-01 5.50048947e-01
-7.12475955e-01 -8.72857928e-01 -8.97320271e-01 -1.35622501e-01
1.00500733e-01 -6.43025100e-01 6.46928072e-01 -1.36380112e+00
-1.58113599e+00 1.00217223e+00 4.49517250e-01 -2.33361110e-01
3.74793291e-01 2.51055323e-02 -4.08250064e-01 6.54480994e-01
-1.74610540e-01 1.64809123e-01 7.07452297e-01 -1.47758174e+00
-1.60520092e-01 -6.36743844e-01 -1.59769267e-01 6.24777615e-01
-6.40862823e-01 -6.05786026e-01 1.90330446e-02 -3.91886115e-01
8.12431812e-01 -1.03037798e+00 -4.94063616e-01 1.90242261e-01
-4.38636750e-01 9.63854671e-01 2.59109110e-01 -7.48817801e-01
7.11130142e-01 -2.29332352e+00 -1.69144273e-02 4.10849005e-01
1.55388921e-01 1.50997907e-01 -5.37907071e-02 3.11250985e-01
-6.07734203e-01 -4.65440571e-01 -8.02176118e-01 -7.78210089e-02
-6.28058136e-01 4.14189734e-02 2.12091669e-01 1.11965346e+00
-3.58909458e-01 6.38412416e-01 -1.10508728e+00 -3.25211108e-01
4.32658494e-01 4.64454830e-01 -3.77841383e-01 1.12500433e-02
9.50515494e-02 6.54321730e-01 -1.81098416e-01 6.07425511e-01
9.40003395e-01 -3.00966233e-01 4.58337307e-01 -6.08097970e-01
-3.48351151e-01 -2.79652983e-01 -7.84313142e-01 2.34035826e+00
-6.82884216e-01 5.49834847e-01 5.54737568e-01 -1.24252522e+00
5.82242250e-01 3.69802237e-01 1.15461671e+00 -8.62038314e-01
6.15232550e-02 6.00634992e-01 -1.57285616e-01 -7.52512872e-01
3.40990037e-01 -1.06251216e+00 5.83043993e-01 -1.02930367e-01
-2.64071137e-01 -4.04477030e-01 -3.63682330e-01 -1.51905239e-01
8.19625139e-01 -1.77992746e-01 1.77032366e-01 -5.58068097e-01
3.00816238e-01 3.75763506e-01 3.40034157e-01 5.03713071e-01
9.43927318e-02 8.29106271e-01 -6.49434850e-02 -2.52507925e-01
-6.93911314e-01 -8.71349454e-01 -5.70100069e-01 6.94162697e-02
5.13159275e-01 3.67083281e-01 -6.65660024e-01 -2.21944347e-01
-2.10571051e-01 4.04126257e-01 -6.86342120e-01 -1.65733010e-01
-3.50316465e-01 -1.19195545e+00 -2.26903912e-02 1.99366778e-01
4.26245958e-01 -6.55758083e-01 -6.96390092e-01 1.98326841e-01
-3.40793669e-01 -9.03065681e-01 -9.77671221e-02 3.67966980e-01
-1.24113107e+00 -1.10328710e+00 -9.58522558e-01 -6.93058729e-01
8.49825919e-01 6.91605508e-01 5.29653966e-01 6.39293343e-02
-6.31145597e-01 3.45074415e-01 -3.30754936e-01 -3.19819152e-01
-3.33278984e-01 -3.87334466e-01 7.28130043e-02 3.88561368e-01
-2.55995125e-01 -8.72355878e-01 -9.66102123e-01 -2.52742041e-02
-1.40564227e+00 4.63094711e-01 1.07538521e+00 1.08394492e+00
6.81276083e-01 2.08301365e-01 1.43217444e-01 -9.59451139e-01
3.38055670e-01 -3.33904088e-01 -5.16177654e-01 9.14672017e-02
-7.59497941e-01 -4.26999852e-02 6.22800112e-01 -4.54471111e-02
-1.01014137e+00 2.46247888e-01 -9.84375104e-02 -7.18518078e-01
1.87400319e-02 1.15957093e+00 8.80324244e-02 -5.17854214e-01
1.11934781e+00 4.26523060e-01 7.36573398e-01 -8.13901499e-02
-1.03887610e-01 5.82669318e-01 4.88907516e-01 -1.45983741e-01
7.56389797e-01 1.03669417e+00 2.21624121e-01 -1.16782284e+00
-7.06155539e-01 -8.54796350e-01 -3.55301142e-01 -4.89058852e-01
8.34456682e-01 -1.10972154e+00 -6.33346677e-01 6.36436284e-01
-5.20576656e-01 -6.64041579e-01 -2.40903512e-01 1.00301683e+00
-7.41284370e-01 6.77150309e-01 -5.58196187e-01 -6.18592143e-01
-4.01667714e-01 -1.09738433e+00 9.10716295e-01 -5.13882451e-02
5.80094039e-01 -1.00011659e+00 9.28464010e-02 7.18653560e-01
3.89514357e-01 5.18456697e-01 7.04946280e-01 1.39305532e-01
-6.85557008e-01 -3.23312521e-01 -1.90383777e-01 5.07808864e-01
3.65294278e-01 -5.17164052e-01 -1.08575988e+00 -4.71405685e-01
4.53374952e-01 -1.70536131e-01 9.23887491e-01 8.96884084e-01
1.35683596e+00 -6.34036139e-02 -2.40564272e-01 1.24154031e+00
2.06290722e+00 7.20851868e-02 7.38163471e-01 4.63728696e-01
7.12467551e-01 5.25609314e-01 7.80506492e-01 3.82276148e-01
-2.13666707e-01 3.32661003e-01 9.61110294e-01 -6.14851892e-01
2.38134377e-02 1.47636995e-01 -7.53640831e-02 6.41159415e-01
-3.59338343e-01 1.40769839e-01 -1.00109613e+00 5.34585953e-01
-1.64557183e+00 -9.05782938e-01 -1.72289550e-01 2.53743243e+00
7.15468228e-01 -4.83345747e-01 -4.56250191e-01 1.58021048e-01
4.94690686e-01 1.72999218e-01 -5.05971014e-01 4.25823212e-01
-9.91003439e-02 3.06213886e-01 8.69641304e-01 3.59425426e-01
-8.84286821e-01 2.88897246e-01 5.04036427e+00 5.08105457e-01
-1.73685789e+00 1.04562238e-01 5.06960630e-01 -3.23080033e-01
-3.32046241e-01 -3.41449022e-01 1.79213330e-01 6.73928782e-02
6.39216244e-01 -1.00224890e-01 5.26062548e-01 3.98613006e-01
5.07219911e-01 -2.37195551e-01 -7.42221236e-01 1.36797285e+00
1.73648745e-01 -1.45803416e+00 -2.14576557e-01 3.51358742e-01
7.24818885e-01 2.59893715e-01 2.07891166e-01 -4.79930550e-01
-2.77242303e-01 -1.12245989e+00 3.97175193e-01 6.41079664e-01
1.12051415e+00 -5.18383980e-01 6.72564685e-01 2.63583034e-01
-7.13221848e-01 -2.38619760e-01 -1.68122470e-01 3.02131534e-01
3.31801325e-02 1.09090972e+00 -8.64810884e-01 9.73957717e-01
4.35602099e-01 9.80215430e-01 8.29796046e-02 1.37334561e+00
7.78877586e-02 2.23757342e-01 -2.58487552e-01 2.27514371e-01
2.64760524e-01 -6.20201111e-01 5.19313991e-01 7.85395384e-01
6.17842138e-01 4.93962497e-01 -5.82581162e-02 7.19602823e-01
1.93038195e-01 1.64509445e-01 -5.36885142e-01 -6.38505518e-02
-9.95443463e-02 1.37857652e+00 -5.81809461e-01 -5.61445616e-02
-3.08846027e-01 9.56068516e-01 -1.28859654e-01 5.76289833e-01
-5.31446815e-01 -7.36626163e-02 7.04093695e-01 3.76125216e-01
-3.04093361e-01 -2.12821111e-01 -4.78121907e-01 -1.36388099e+00
3.42303254e-02 -8.26498449e-01 3.13495517e-01 -1.00864398e+00
-1.02667749e+00 6.86590552e-01 -2.46551841e-01 -1.84416199e+00
3.35909456e-01 -8.28585505e-01 -2.55101562e-01 8.81040394e-01
-1.85522664e+00 -1.00616884e+00 -1.00594735e+00 6.81849003e-01
-4.40789908e-02 2.18560398e-01 8.41247439e-01 3.06553334e-01
-4.14686561e-01 5.97771630e-02 4.49059129e-01 -1.33074671e-01
5.49702466e-01 -1.15094769e+00 -6.67825818e-01 8.26950133e-01
-3.64075422e-01 5.14349937e-01 5.74066579e-01 -5.74264884e-01
-1.75476265e+00 -8.93788099e-01 -1.11242250e-01 4.66709644e-01
5.38443327e-01 2.60925353e-01 -6.28935874e-01 4.05711025e-01
1.24590054e-01 1.55490965e-01 1.06580234e+00 -1.62965581e-01
2.43417308e-01 -3.95022362e-01 -1.12600660e+00 3.99845153e-01
8.57034445e-01 -4.89439219e-01 -8.92085135e-02 8.13006282e-01
3.06868792e-01 -6.45203531e-01 -1.06889927e+00 7.38797009e-01
4.78976220e-01 -1.24969089e+00 9.29544866e-01 1.68756787e-02
6.56853199e-01 -5.83492927e-02 5.51057383e-02 -1.63759184e+00
-2.38470525e-01 -5.05341053e-01 7.03029335e-01 1.74715117e-01
4.94828582e-01 -9.04263616e-01 1.10466683e+00 5.51766098e-01
-7.99967110e-01 -5.76215267e-01 -9.30384040e-01 -7.30218112e-01
-6.59153685e-02 -3.04659516e-01 3.52196395e-02 1.37106943e+00
4.28823888e-01 -3.77663702e-01 -1.15261100e-01 6.54164076e-01
1.01527345e+00 4.27923501e-01 5.33894710e-02 -1.05360234e+00
-3.07858676e-01 -2.53147691e-01 -5.10422289e-01 -6.28868103e-01
8.79446045e-02 -1.08324015e+00 3.69730473e-01 -1.79183030e+00
2.40326539e-01 -7.54925728e-01 -4.83455956e-01 3.58446270e-01
-1.81097910e-01 4.20850933e-01 -1.23341016e-01 3.53663415e-01
1.89520001e-01 5.66845000e-01 1.55439079e+00 -3.29140007e-01
-3.12637061e-01 -1.56347886e-01 -5.62134981e-01 3.33503544e-01
6.28222585e-01 -2.29784071e-01 -5.42637587e-01 -5.65357924e-01
2.36030087e-01 5.80480278e-01 3.76244307e-01 -1.37008262e+00
4.31651175e-01 -2.90189415e-01 1.03958294e-01 1.10515304e-01
5.95131218e-01 -1.31798053e+00 5.19387126e-01 6.54362023e-01
9.35578998e-03 -7.97411382e-01 2.53581673e-01 4.30885583e-01
-4.69909757e-01 -2.31628135e-01 1.00954235e+00 -2.27184594e-01
-6.41323149e-01 6.00039184e-01 -2.70306557e-01 -4.76002842e-01
9.16917622e-01 -6.98051870e-01 -6.83441162e-02 -2.40678176e-01
-8.39394510e-01 -2.60935754e-01 8.12536597e-01 -2.94657111e-01
9.04698372e-01 -1.14278698e+00 -6.55427217e-01 4.46323067e-01
3.12810004e-01 4.08377610e-02 7.90412962e-01 1.48864675e+00
-1.20013809e+00 1.23672754e-01 -4.25474703e-01 -6.80223644e-01
-1.05932713e+00 4.46622789e-01 9.61985826e-01 -6.68379515e-02
-7.78925836e-01 7.89256334e-01 3.28624845e-01 -2.62409627e-01
-1.92554444e-01 -3.41894776e-01 -2.87635457e-02 -1.87090740e-01
3.58684182e-01 -1.31069213e-01 6.17861629e-01 -3.52656156e-01
-2.56208211e-01 5.48939109e-01 4.00473595e-01 -9.23889279e-02
1.56597745e+00 7.77272461e-03 -2.21854910e-01 5.84209561e-02
1.21219218e+00 -1.47064850e-01 -1.33749664e+00 -1.26313075e-01
-6.38559222e-01 -6.12666786e-01 7.84019470e-01 -7.38256335e-01
-1.44942105e+00 8.29799414e-01 8.54635358e-01 -6.40885299e-03
1.72842062e+00 -4.58576530e-01 5.48642099e-01 2.54592389e-01
5.06697595e-01 -7.68424928e-01 -5.10498211e-02 -9.17938203e-02
8.42351913e-01 -1.30488276e+00 3.22493404e-01 -7.21283734e-01
-3.79417807e-01 1.15343451e+00 2.07735837e-01 1.23602211e-01
6.01777911e-01 1.51512280e-01 4.62376237e-01 -4.17399198e-01
-6.51089326e-02 -1.31238759e-01 3.19389224e-01 3.26680005e-01
2.98612386e-01 2.42433831e-01 -1.87527224e-01 -7.77424574e-02
-4.66119014e-02 1.46841377e-01 6.47898972e-01 9.01645899e-01
-3.58412713e-01 -6.51816726e-01 -3.48390102e-01 6.12594008e-01
-2.20208362e-01 -4.13184077e-01 1.22192748e-01 4.05809700e-01
-2.11292863e-01 6.70182168e-01 -2.24156886e-01 -2.51331776e-01
2.79408723e-01 -2.15428308e-01 6.79282904e-01 -6.22725844e-01
-4.08529133e-01 1.25820622e-01 1.42568341e-02 -7.30069041e-01
-7.95185566e-01 -5.21958113e-01 -1.24470460e+00 7.23168850e-02
-3.08284551e-01 2.23948032e-01 1.13573897e+00 8.02610517e-01
8.52145329e-02 6.14328444e-01 6.68451726e-01 -1.18596494e+00
-3.44375610e-01 -6.76056385e-01 -1.26075566e+00 3.98413658e-01
6.65316045e-01 -2.68525511e-01 -4.53721464e-01 2.45346176e-03] | [10.267462730407715, -2.12937593460083] |
73f77f28-78f1-477a-894f-8a0fc2a237df | neural-collapse-inspired-feature-classifier-1 | 2302.03004 | null | https://arxiv.org/abs/2302.03004v1 | https://arxiv.org/pdf/2302.03004v1.pdf | Neural Collapse Inspired Feature-Classifier Alignment for Few-Shot Class Incremental Learning | Few-shot class-incremental learning (FSCIL) has been a challenging problem as only a few training samples are accessible for each novel class in the new sessions. Finetuning the backbone or adjusting the classifier prototypes trained in the prior sessions would inevitably cause a misalignment between the feature and classifier of old classes, which explains the well-known catastrophic forgetting problem. In this paper, we deal with this misalignment dilemma in FSCIL inspired by the recently discovered phenomenon named neural collapse, which reveals that the last-layer features of the same class will collapse into a vertex, and the vertices of all classes are aligned with the classifier prototypes, which are formed as a simplex equiangular tight frame (ETF). It corresponds to an optimal geometric structure for classification due to the maximized Fisher Discriminant Ratio. We propose a neural collapse inspired framework for FSCIL. A group of classifier prototypes are pre-assigned as a simplex ETF for the whole label space, including the base session and all the incremental sessions. During training, the classifier prototypes are not learnable, and we adopt a novel loss function that drives the features into their corresponding prototypes. Theoretical analysis shows that our method holds the neural collapse optimality and does not break the feature-classifier alignment in an incremental fashion. Experiments on the miniImageNet, CUB-200, and CIFAR-100 datasets demonstrate that our proposed framework outperforms the state-of-the-art performances. Code address: https://github.com/NeuralCollapseApplications/FSCIL | ['DaCheng Tao', 'Philip Torr', 'Zhouchen Lin', 'Xiangtai Li', 'Haobo Yuan', 'Yibo Yang'] | 2023-02-06 | null | null | null | null | ['class-incremental-learning', 'few-shot-class-incremental-learning'] | ['computer-vision', 'methodology'] | [ 5.22307269e-02 7.59617463e-02 -1.21471487e-01 -3.75509262e-01
-2.50544876e-01 -3.73940796e-01 4.24734384e-01 3.96880247e-02
-3.97309095e-01 8.71213317e-01 -2.26229697e-01 3.42713565e-01
-1.78961411e-01 -6.02726698e-01 -1.00263917e+00 -1.19022954e+00
5.30048199e-02 5.67084670e-01 5.29315531e-01 8.79813880e-02
1.94101676e-01 3.83710057e-01 -2.03815079e+00 2.12471545e-01
7.56294608e-01 9.66906786e-01 3.18735003e-01 1.40075728e-01
-1.16271175e-01 2.29887858e-01 -4.04393554e-01 -4.91697669e-01
2.54713476e-01 -4.03371990e-01 -5.55714369e-01 1.78521484e-01
6.53385699e-01 -7.93239325e-02 -3.80729765e-01 1.13823462e+00
3.91825557e-01 3.17362189e-01 6.19047344e-01 -1.60199332e+00
-6.22618675e-01 5.23724854e-01 -7.20195413e-01 9.07397270e-02
-1.06327213e-01 -8.46851319e-02 7.09647715e-01 -1.38348889e+00
8.16267550e-01 1.05528295e+00 7.33013034e-01 8.02421629e-01
-1.16220808e+00 -8.49919379e-01 4.64037359e-01 6.48674965e-01
-1.65612984e+00 -5.29529214e-01 7.03800738e-01 -4.33994293e-01
4.28879946e-01 1.36576027e-01 9.41999435e-01 1.06171846e+00
3.10880452e-01 6.95507824e-01 9.70762730e-01 -3.48500490e-01
5.02834857e-01 3.56588811e-01 5.97891927e-01 8.49391282e-01
3.58059376e-01 7.84890950e-02 -7.07640231e-01 4.49523143e-03
4.88498867e-01 3.14529002e-01 -2.87181288e-01 -8.41650248e-01
-9.51614082e-01 7.08243906e-01 6.28843606e-01 2.52079725e-01
-1.34102061e-01 -1.26414001e-01 2.12192297e-01 3.50165635e-01
2.24831104e-01 -4.46047448e-03 -3.56865227e-01 2.82315642e-01
-7.54742622e-01 1.85098037e-01 4.69261259e-01 1.00497484e+00
1.00045371e+00 -2.36874297e-01 -2.11827338e-01 8.14516664e-01
-1.64116751e-02 2.51506567e-01 6.98130965e-01 -8.29599917e-01
7.32313469e-02 5.16001523e-01 -2.86694288e-01 -9.61556911e-01
-3.89005244e-01 -9.01281416e-01 -9.89631534e-01 1.60503030e-01
3.78853559e-01 1.27815783e-01 -9.46717083e-01 1.99531126e+00
5.71241856e-01 5.59369266e-01 -4.85786013e-02 5.96185982e-01
7.56605923e-01 5.67377627e-01 -2.34104946e-01 -6.55524075e-01
1.09743392e+00 -9.38561082e-01 -4.81797278e-01 -7.52104893e-02
2.83068419e-01 -3.53039354e-01 1.17178106e+00 3.93645287e-01
-8.23046148e-01 -7.70757616e-01 -1.19430268e+00 2.51073509e-01
-2.88612515e-01 -1.31338596e-01 3.48648787e-01 3.39300483e-01
-7.47624338e-01 6.80471182e-01 -8.31177890e-01 -3.32952857e-01
7.07849264e-01 2.69709438e-01 -2.99785852e-01 -6.96574599e-02
-7.77159691e-01 5.55836082e-01 4.75556731e-01 5.63520975e-02
-8.46499085e-01 -8.55352879e-01 -4.05097663e-01 1.00524031e-01
3.35844398e-01 -5.24245441e-01 1.09164143e+00 -1.07276309e+00
-1.18826926e+00 7.60269344e-01 -2.96846330e-01 -3.15695494e-01
5.21579087e-01 1.43266261e-01 -8.72220472e-02 -1.26535177e-01
8.66208002e-02 8.60736430e-01 1.17095673e+00 -1.29217863e+00
-8.81917417e-01 -5.32582939e-01 -1.26863837e-01 3.94986540e-01
-5.68937361e-01 -6.47027612e-01 -4.09172833e-01 -5.16673386e-01
4.78396893e-01 -1.12608922e+00 1.43368706e-01 9.62113962e-02
-2.23283157e-01 -4.38067079e-01 8.08877647e-01 4.24876250e-03
1.23415887e+00 -2.33303452e+00 3.29387665e-01 3.15131657e-02
3.70185852e-01 4.66996282e-02 8.85776505e-02 -6.54663332e-03
-2.02438638e-01 -2.82115132e-01 -1.92568064e-01 -3.44992965e-01
-1.57650307e-01 2.81864107e-01 -4.48865294e-01 5.51359653e-01
-3.33031207e-01 6.16507530e-01 -9.52185333e-01 -4.02350247e-01
6.68722764e-02 2.22813666e-01 -5.79493582e-01 -5.95874563e-02
4.33934554e-02 4.12305385e-01 -4.34325375e-02 5.38750887e-01
9.11675572e-01 -2.61030495e-01 -3.58966924e-02 -2.56382018e-01
-2.38233507e-02 -5.54640472e-01 -1.11911905e+00 1.76510036e+00
-1.55056622e-02 3.11511070e-01 -4.62249666e-01 -1.05632198e+00
7.63155997e-01 4.07609716e-02 3.47604215e-01 -5.06304681e-01
1.34067750e-02 1.51002944e-01 -5.63239083e-02 -2.79447198e-01
4.35062982e-02 -1.86876014e-01 -6.51946738e-02 1.79575160e-01
5.74356556e-01 2.52648532e-01 2.58709371e-01 1.44372389e-01
7.86863565e-01 -3.31938341e-02 1.70174420e-01 -1.85079545e-01
5.02999008e-01 -2.84521371e-01 9.52356458e-01 1.00009871e+00
-3.78775775e-01 7.31166720e-01 4.01852489e-01 -6.39297605e-01
-9.15853679e-01 -1.32100165e+00 -5.32272756e-01 1.09899926e+00
2.86682785e-01 -1.34779260e-01 -9.58238125e-01 -7.48646855e-01
-1.00619905e-01 8.17238331e-01 -7.81436563e-01 -5.93130767e-01
-3.40055138e-01 -7.55996108e-01 5.18586487e-02 1.24781355e-01
5.54960012e-01 -1.10829008e+00 -6.46176994e-01 3.38819355e-01
-2.93150544e-02 -7.02902138e-01 -4.23603147e-01 2.36022219e-01
-1.01802099e+00 -1.19681752e+00 -7.68842936e-01 -1.03318644e+00
9.40004170e-01 3.92606914e-01 7.10133851e-01 -8.70551690e-02
-3.82503957e-01 1.84400767e-01 -2.53758520e-01 -2.88729697e-01
-8.22535753e-02 1.39148325e-01 4.39543664e-01 3.93741578e-01
3.05749923e-01 -7.38483369e-01 -6.59397483e-01 3.45764786e-01
-6.63401902e-01 2.16088384e-01 2.96163201e-01 9.73227859e-01
8.88528466e-01 6.58221915e-02 8.60691607e-01 -8.14222515e-01
2.76263952e-01 -3.87517273e-01 -5.25619388e-01 4.58906740e-01
-7.47105539e-01 -9.58916619e-02 5.99826276e-01 -7.30420530e-01
-9.34368491e-01 1.67766795e-01 2.26191416e-01 -7.02820361e-01
3.68035585e-03 2.90935546e-01 -8.36372599e-02 -8.59117731e-02
7.20228791e-01 2.59583920e-01 -4.47527058e-02 -3.42754662e-01
3.17829192e-01 4.40585077e-01 6.84859276e-01 -4.61221039e-01
7.93913245e-01 5.21393776e-01 -8.60895589e-02 -6.99880064e-01
-1.20099485e+00 -2.13223353e-01 -8.89426529e-01 -5.07714391e-01
4.99735832e-01 -6.87634647e-01 -5.73490381e-01 8.41757298e-01
-1.06399035e+00 -5.47951870e-02 -7.17445016e-01 4.07564223e-01
-5.27087986e-01 1.51540205e-01 -2.62164831e-01 -4.31877911e-01
-2.34955266e-01 -1.06056988e+00 5.31055808e-01 7.31725097e-01
1.04970217e-01 -6.02685630e-01 2.17264101e-01 -8.26491490e-02
8.32581818e-02 9.54020321e-02 9.93208766e-01 -4.97442216e-01
-4.42061126e-01 -1.42889574e-01 -2.59597991e-02 2.81958371e-01
3.14271115e-02 -1.48149773e-01 -1.07424712e+00 -7.53596365e-01
1.62587136e-01 -2.70817578e-01 1.04724061e+00 3.99189621e-01
1.31169116e+00 -1.97347179e-01 -5.11571884e-01 7.96962023e-01
1.33049595e+00 4.36433613e-01 5.21269500e-01 1.79053754e-01
5.55794835e-01 3.65289509e-01 4.74022686e-01 4.46785539e-01
1.67126432e-01 5.27982771e-01 2.53090233e-01 3.46361846e-01
-3.01781267e-01 -2.32509002e-01 1.67407647e-01 1.05442548e+00
5.41635640e-02 -2.58230437e-02 -6.56409085e-01 4.21062618e-01
-2.09984303e+00 -9.87796068e-01 2.08791390e-01 2.57646704e+00
8.68330061e-01 2.03160420e-01 -3.20552081e-01 4.76900414e-02
9.30501223e-01 -1.55686826e-01 -1.01452506e+00 -5.73580787e-02
-1.12862282e-01 1.92072704e-01 2.09386274e-01 4.00151074e-01
-1.01147091e+00 8.68624330e-01 5.36367846e+00 9.20419633e-01
-1.00831819e+00 4.24096227e-01 6.63982749e-01 -3.81479591e-01
9.88380164e-02 8.22776631e-02 -1.09431779e+00 5.38558006e-01
6.36680424e-01 -5.22083402e-01 5.10071933e-01 7.55538106e-01
-2.65149802e-01 -4.30068485e-02 -1.19704759e+00 1.20525312e+00
3.21551651e-01 -1.44488490e+00 7.25499988e-02 -2.50548512e-01
8.20532322e-01 6.71390723e-03 3.04082215e-01 5.42526126e-01
-2.48808324e-01 -5.45535028e-01 8.63139987e-01 9.23003018e-01
9.22510266e-01 -8.21956754e-01 4.87598956e-01 3.65323842e-01
-1.09706640e+00 -4.05837625e-01 -6.74963772e-01 1.42939895e-01
-2.07391158e-01 5.55475652e-01 -6.41991913e-01 1.76461414e-01
9.56432223e-01 7.60025680e-01 -8.33203375e-01 1.42704177e+00
4.22052667e-02 3.22723508e-01 -2.24067256e-01 2.33070567e-01
3.38026532e-03 -2.21365288e-01 7.34661639e-01 6.36223137e-01
3.85387480e-01 9.56760570e-02 1.21263333e-01 6.47646546e-01
-2.50272989e-01 -7.04543665e-02 -4.53136265e-01 4.13090080e-01
6.44165039e-01 1.07610095e+00 -9.37996030e-01 -4.14633989e-01
-2.17462569e-01 1.07619178e+00 5.16692996e-01 3.53903770e-01
-8.91594410e-01 -4.37644869e-01 4.14991319e-01 4.48660320e-03
3.23696613e-01 2.44634017e-01 -1.71463981e-01 -1.33662033e+00
1.70142353e-01 -4.65556264e-01 3.64512742e-01 -5.53855896e-01
-1.31227362e+00 7.39782870e-01 1.33463042e-02 -1.22538042e+00
7.48175457e-02 -2.14461610e-01 -5.61777532e-01 2.16187820e-01
-1.04954660e+00 -8.76003802e-01 -5.74929178e-01 7.32114494e-01
7.24909902e-01 -4.10069227e-01 7.81418025e-01 2.78274715e-01
-7.37045228e-01 9.80439305e-01 5.40665746e-01 -3.22166771e-01
6.93817317e-01 -8.45992863e-01 -2.73114815e-03 6.54787481e-01
2.28813216e-01 4.90807742e-01 5.97680390e-01 -6.06628895e-01
-1.01769888e+00 -1.22097409e+00 6.48750007e-01 -1.83791503e-01
4.30623323e-01 -6.11097574e-01 -1.22840357e+00 5.49469113e-01
-1.66709319e-01 2.46648952e-01 4.55670804e-01 -6.28936291e-02
-2.50956923e-01 -5.17486930e-01 -1.13220835e+00 5.01201093e-01
1.32486057e+00 -2.49121502e-01 -6.51139677e-01 6.07694864e-01
7.53566325e-01 -2.29329944e-01 -3.72619778e-01 4.20655996e-01
7.05717325e-01 -9.59372520e-01 7.98921764e-01 -4.66636837e-01
-5.87427653e-02 -3.53968352e-01 -1.27233341e-01 -1.25204754e+00
-5.69042265e-01 -4.06996042e-01 -2.54430950e-01 1.22250235e+00
2.12503389e-01 -7.61688113e-01 7.91958928e-01 2.72495359e-01
-2.49184951e-01 -9.11299884e-01 -1.43059969e+00 -8.21277499e-01
6.26213402e-02 -9.44089293e-02 4.44664955e-01 8.96981061e-01
-2.83457160e-01 4.14447099e-01 -1.25218689e-01 3.97512019e-02
1.06819713e+00 1.15697064e-01 4.52600121e-01 -1.54026854e+00
-6.34150803e-02 -4.34445292e-01 -4.72934544e-01 -5.53233624e-01
1.02629796e-01 -1.14979005e+00 -4.46450114e-02 -1.15033793e+00
5.77260494e-01 -8.11835170e-01 -6.11226141e-01 7.30614185e-01
5.11808284e-02 1.69513091e-01 3.04731697e-01 4.52763945e-01
-7.09177971e-01 7.78198242e-01 1.05806148e+00 -5.53586520e-02
-4.28218514e-01 1.45339116e-01 -4.88934100e-01 7.84079611e-01
7.09741414e-01 -7.50752687e-01 -5.65140069e-01 -1.51508749e-01
5.96055612e-02 -3.03577989e-01 3.76981527e-01 -1.38021672e+00
5.55072248e-01 4.08519544e-02 6.27232552e-01 -6.75643206e-01
2.49954924e-01 -8.64537299e-01 3.41385096e-01 5.51111400e-01
-2.14462623e-01 -2.89785206e-01 3.18309404e-02 7.69829392e-01
2.12576926e-01 -4.65517551e-01 1.11277437e+00 -2.76654065e-02
-5.42323112e-01 7.01980770e-01 -1.21140435e-01 -4.50165971e-04
1.42572331e+00 -3.77028257e-01 -3.31356883e-01 1.87673986e-01
-1.04687166e+00 1.96998060e-01 4.40795124e-01 5.81375122e-01
8.30157876e-01 -1.57315135e+00 -5.16583681e-01 5.16494393e-01
1.75421029e-01 1.09122388e-01 5.04101992e-01 7.87078202e-01
-1.29342198e-01 -2.63946410e-02 -4.71476406e-01 -8.76864493e-01
-1.15748096e+00 7.02442408e-01 3.52626413e-01 7.98648149e-02
-7.86014676e-01 1.05696416e+00 5.84187925e-01 -4.42093968e-01
5.62536240e-01 -2.79702805e-02 -2.25371301e-01 4.47392374e-01
6.36465132e-01 2.97576725e-01 8.99402425e-02 -4.39303398e-01
-2.26229176e-01 5.85740089e-01 -5.61414599e-01 1.28057361e-01
1.45189393e+00 -3.11261863e-01 -1.45605281e-01 8.71131420e-01
1.22454393e+00 -5.37404001e-01 -1.46125388e+00 -6.08561635e-01
-2.67747492e-01 -3.83974284e-01 -8.82147402e-02 -5.31610370e-01
-1.12650585e+00 7.10115731e-01 1.17427552e+00 -9.44966152e-02
1.06300080e+00 9.45223272e-02 6.59583867e-01 3.25015962e-01
6.11229300e-01 -1.13983333e+00 1.93585694e-01 5.92282236e-01
9.41296935e-01 -7.34237492e-01 -1.51998341e-01 -1.84023790e-02
-5.75957358e-01 1.18654466e+00 7.98204482e-01 -2.23603114e-01
8.29103708e-01 -2.48454779e-01 -4.46859390e-01 -1.36503726e-01
-7.90039659e-01 1.78766847e-01 4.25667688e-02 4.48065996e-01
-2.03054741e-01 6.15253672e-02 -1.00438684e-01 7.42471576e-01
-1.80404499e-01 1.72534026e-02 5.42237222e-01 8.13739717e-01
-7.12248862e-01 -9.05507922e-01 -2.31310785e-01 5.92203319e-01
5.40133007e-02 1.27383530e-01 6.46598563e-02 5.61570883e-01
5.78216016e-01 5.07112861e-01 3.77563596e-01 -4.67044353e-01
2.79816926e-01 2.82681525e-01 6.83130264e-01 -5.88719130e-01
-1.53999358e-01 -6.67101070e-02 -7.25700498e-01 -3.90677303e-01
-1.29820481e-01 -7.93507636e-01 -1.40726161e+00 -2.78842628e-01
-3.26751053e-01 1.93250835e-01 4.41087395e-01 8.15757871e-01
3.58963311e-01 4.67202783e-01 9.31423903e-01 -8.97449136e-01
-4.25844938e-01 -7.55575716e-01 -5.46043992e-01 2.46374547e-01
3.45269978e-01 -1.05768573e+00 -5.80989122e-01 1.06373109e-01] | [9.78442096710205, 3.3367199897766113] |
5d15cb4b-cf78-4746-863e-81d49be9dd66 | cell-detection-in-microscopy-images-with-deep | 1708.03307 | null | http://arxiv.org/abs/1708.03307v3 | http://arxiv.org/pdf/1708.03307v3.pdf | Cell Detection in Microscopy Images with Deep Convolutional Neural Network and Compressed Sensing | The ability to automatically detect certain types of cells or cellular
subunits in microscopy images is of significant interest to a wide range of
biomedical research and clinical practices. Cell detection methods have evolved
from employing hand-crafted features to deep learning-based techniques. The
essential idea of these methods is that their cell classifiers or detectors are
trained in the pixel space, where the locations of target cells are labeled. In
this paper, we seek a different route and propose a convolutional neural
network (CNN)-based cell detection method that uses encoding of the output
pixel space. For the cell detection problem, the output space is the sparsely
labeled pixel locations indicating cell centers. We employ random projections
to encode the output space to a compressed vector of fixed dimension. Then, CNN
regresses this compressed vector from the input pixels. Furthermore, it is
possible to stably recover sparse cell locations on the output pixel space from
the predicted compressed vector using $L_1$-norm optimization. In the past,
output space encoding using compressed sensing (CS) has been used in
conjunction with linear and non-linear predictors. To the best of our
knowledge, this is the first successful use of CNN with CS-based output space
encoding. We made substantial experiments on several benchmark datasets, where
the proposed CNN + CS framework (referred to as CNNCS) achieved the highest or
at least top-3 performance in terms of F1-score, compared with other
state-of-the-art methods. | ['Nilanjan Ray', 'Yao Xue'] | 2017-08-10 | null | null | null | null | ['cell-detection'] | ['computer-vision'] | [ 6.67689979e-01 -1.33036882e-01 -9.79464352e-02 -9.18133184e-02
-6.78423524e-01 -2.41920322e-01 4.86658037e-01 2.56954461e-01
-5.26783109e-01 7.80090749e-01 -2.17204422e-01 -2.22456399e-02
2.32811213e-01 -6.53303742e-01 -6.32103086e-01 -1.21924782e+00
-2.31387746e-02 1.06566608e-01 1.69818446e-01 2.12960139e-01
6.25098944e-01 8.05378258e-01 -1.31740570e+00 5.32827556e-01
3.90671879e-01 1.45592749e+00 3.84642154e-01 7.11415648e-01
-5.43685332e-02 9.04573083e-01 -2.91372716e-01 1.71133146e-01
1.04600161e-01 -4.82054591e-01 -5.32909453e-01 -6.04210161e-02
2.78299332e-01 7.09770024e-02 -4.62311149e-01 1.00455678e+00
5.26048660e-01 -2.90070951e-01 8.30497861e-01 -5.82876801e-01
-5.21081924e-01 9.08581540e-02 -4.86814022e-01 5.67806721e-01
2.02860013e-01 1.02659866e-01 6.10390306e-01 -1.16371298e+00
9.09555197e-01 6.98042333e-01 6.41877353e-01 4.18960273e-01
-1.30907953e+00 -3.99028778e-01 -2.80495614e-01 -7.83348307e-02
-1.74026370e+00 -4.90292102e-01 6.75011814e-01 -4.32039559e-01
1.07034874e+00 2.50503719e-01 7.56401896e-01 6.50861800e-01
4.20540899e-01 7.89309263e-01 1.10159481e+00 -5.63697994e-01
4.10019845e-01 1.48098812e-01 -3.01899195e-01 9.75806177e-01
-8.78748018e-03 -1.03432991e-01 -4.82897162e-01 -4.74123992e-02
8.88801754e-01 5.93048692e-01 -5.88425636e-01 -1.16034456e-01
-1.57584476e+00 8.25985074e-01 7.17106164e-01 6.59973979e-01
-2.49142542e-01 1.34327650e-01 2.63316453e-01 -1.70345917e-01
2.65332133e-01 4.18247521e-01 -3.64191115e-01 2.96392024e-01
-1.28448820e+00 1.40625328e-01 6.88412249e-01 4.87974823e-01
5.05930841e-01 -8.47590193e-02 -3.06154728e-01 7.64429510e-01
1.50176480e-01 3.28101963e-01 8.11351240e-01 -8.84556770e-01
-6.01430191e-03 8.13155174e-01 -3.02225560e-01 -1.20561016e+00
-4.47889268e-01 -7.52319872e-01 -1.21498668e+00 3.83143350e-02
5.33978105e-01 -6.82582706e-02 -7.99479842e-01 1.47728026e+00
1.30265474e-01 5.17968059e-01 1.23276502e-01 9.07127976e-01
5.63562453e-01 7.33928382e-01 -3.30855072e-01 -3.75551999e-01
1.14107144e+00 -6.32922947e-01 -4.76994276e-01 -9.71170701e-03
8.74045312e-01 -5.72732151e-01 5.74648261e-01 3.37084562e-01
-7.31711805e-01 -3.00027817e-01 -9.81637299e-01 1.03202328e-01
-5.17444909e-01 3.88779193e-01 5.26927471e-01 1.73760206e-01
-1.06354856e+00 5.98359942e-01 -8.33238602e-01 -4.63148773e-01
9.01919365e-01 4.27988857e-01 -4.28646415e-01 -3.11503500e-01
-6.34155691e-01 4.70587969e-01 3.64613198e-02 3.00858673e-02
-8.22817743e-01 -4.94907081e-01 -8.00944328e-01 5.77846989e-02
-4.58178930e-02 -5.27111173e-01 6.50075555e-01 -5.81983685e-01
-1.07636058e+00 1.20529139e+00 -4.27047372e-01 -6.04891181e-01
6.83341026e-02 3.31968665e-01 -5.50692901e-02 3.99751902e-01
1.54508069e-01 8.08430791e-01 6.08294845e-01 -1.09431863e+00
-7.58662045e-01 -4.35385048e-01 -5.84809124e-01 -1.76261365e-01
-4.87073332e-01 -1.61822230e-01 -3.40680480e-01 -7.11813748e-01
4.70365435e-01 -9.50928986e-01 -3.13774616e-01 3.44353586e-01
-5.47632396e-01 4.83089350e-02 8.85972321e-01 -4.13281173e-01
1.06645632e+00 -2.26777387e+00 1.93324789e-01 1.31945476e-01
3.20723563e-01 2.79119253e-01 1.13347404e-01 3.28295469e-01
1.09504666e-02 -1.42421350e-01 -4.12356675e-01 -4.99526471e-01
-5.84181309e-01 -9.48989093e-02 -2.79670924e-01 9.38652873e-01
1.36125684e-01 8.94064128e-01 -7.34137952e-01 -7.41030097e-01
1.18235432e-01 7.94130027e-01 -3.89251918e-01 2.68523935e-02
9.72755775e-02 4.79118496e-01 -2.78050274e-01 1.12229836e+00
3.94000798e-01 -7.04539299e-01 1.30505711e-01 -2.87901968e-01
-1.55765146e-01 -2.80710876e-01 -8.12367499e-01 1.62167013e+00
1.23538457e-01 8.50380540e-01 -4.74646762e-02 -1.33622372e+00
1.02762973e+00 3.26212049e-01 8.87473464e-01 -4.20565128e-01
1.19444303e-01 4.59445328e-01 -1.60709530e-01 -2.72785097e-01
1.06200270e-01 -1.43943608e-01 1.74536854e-01 8.50293711e-02
3.88188176e-02 1.87196687e-01 -1.18687116e-01 1.38309330e-01
1.39550221e+00 -2.63042688e-01 4.99829978e-01 -1.75784498e-01
7.75266051e-01 1.17722757e-01 6.24134660e-01 4.09239709e-01
-2.91953474e-01 9.57700968e-01 4.89039451e-01 -5.72832227e-01
-8.88751388e-01 -5.46077669e-01 -3.86310279e-01 6.35057628e-01
-1.21595509e-01 1.94797311e-02 -7.08863616e-01 -3.88005227e-01
5.02111167e-02 -2.56607425e-03 -6.96048915e-01 7.21398089e-03
-7.02613831e-01 -6.80421174e-01 4.70566273e-01 3.89087647e-01
2.69316018e-01 -1.21156669e+00 -7.84806371e-01 2.74866611e-01
-2.61165705e-02 -1.12559032e+00 -3.15350205e-01 6.28768504e-01
-9.60029185e-01 -9.78271663e-01 -9.58509684e-01 -1.39488959e+00
1.00241041e+00 2.51755476e-01 5.56766450e-01 3.16874027e-01
-7.42037356e-01 -1.78704858e-01 -2.25187451e-01 -2.17415348e-01
1.85775012e-01 -4.40270975e-02 -5.15689664e-02 2.86157608e-01
3.70457530e-01 -3.66305262e-01 -1.02638364e+00 5.82036329e-03
-8.36219132e-01 -6.38120919e-02 6.90430224e-01 1.12396562e+00
1.21776521e+00 -4.70917486e-02 4.45142388e-01 -1.05971944e+00
2.12143093e-01 -4.06221688e-01 -3.56510907e-01 -4.22460865e-03
-3.39040041e-01 -1.25310645e-01 8.93417180e-01 -3.51293504e-01
-2.89778054e-01 5.85682034e-01 -8.42855424e-02 -6.49719954e-01
-2.68021256e-01 6.29211009e-01 2.62986004e-01 -4.00206506e-01
5.84020376e-01 7.82322347e-01 1.82103768e-01 -1.59434214e-01
-1.94107652e-01 6.71596825e-01 7.27432787e-01 -2.74556000e-02
1.53816164e-01 1.01589596e+00 3.66031915e-01 -6.94482148e-01
-6.86694145e-01 -6.58090949e-01 -4.39415395e-01 -1.95593201e-02
7.15586960e-01 -8.58111024e-01 -5.64224124e-01 6.16303802e-01
-1.00853693e+00 -2.26608992e-01 -2.25914255e-01 2.86411464e-01
-6.78507924e-01 2.48158380e-01 -9.83652711e-01 -6.27544641e-01
-3.18412006e-01 -1.18749738e+00 1.12003171e+00 1.90655753e-01
-7.15951174e-02 -8.72345984e-01 2.94141825e-02 1.22513488e-01
4.15473998e-01 5.70079982e-01 8.88817132e-01 -6.47577643e-01
-5.85685909e-01 -6.54048622e-01 -1.28684759e-01 4.69205342e-02
1.05120465e-01 -2.36481041e-01 -1.02143157e+00 -5.28360724e-01
4.79583666e-02 -2.04655215e-01 1.16788924e+00 8.57271552e-01
1.60929263e+00 -2.04686940e-01 -9.12977457e-01 1.05892909e+00
1.70109701e+00 1.64005086e-01 7.04388976e-01 1.52303040e-01
4.73703474e-01 3.35353494e-01 3.14246655e-01 4.80368584e-01
-1.06662968e-02 5.54404080e-01 3.91753495e-01 -3.41620475e-01
-1.34240806e-01 -6.39354661e-02 1.50777651e-02 6.68109238e-01
1.96977228e-01 -2.81395376e-01 -7.93092251e-01 6.68668509e-01
-1.72455609e+00 -8.65220845e-01 1.09254681e-01 2.05485249e+00
6.90949321e-01 -2.56037023e-02 -1.94940329e-01 3.77543569e-01
5.75733662e-01 -9.95517671e-02 -7.29320943e-01 8.12135823e-03
-3.19765180e-01 1.97427541e-01 4.47414905e-01 2.24790648e-01
-1.05464196e+00 7.56149113e-01 5.91820002e+00 9.96474206e-01
-1.56633306e+00 -1.19664013e-01 1.20220208e+00 -1.20298669e-01
2.37914234e-01 -1.98358744e-01 -9.30719018e-01 7.45091796e-01
6.01191938e-01 4.36986208e-01 3.60587776e-01 5.77521324e-01
3.05597913e-02 -1.08470224e-01 -1.07160473e+00 1.28447127e+00
2.44220868e-01 -1.84295654e+00 -1.65860295e-01 5.00730693e-01
7.76575685e-01 2.75155790e-02 2.39847422e-01 -8.16523209e-02
-3.33904684e-01 -1.39438009e+00 2.55918771e-01 7.33782172e-01
1.31424260e+00 -5.52255630e-01 9.84606445e-01 6.49722457e-01
-1.02868283e+00 -2.25786030e-01 -4.40280288e-01 1.80688892e-02
1.58619240e-03 6.82097137e-01 -8.98774505e-01 -2.62550652e-01
5.29411793e-01 1.02591419e+00 -3.10746014e-01 1.06428218e+00
3.91310930e-01 6.69541657e-01 -1.93134964e-01 -1.67679399e-01
1.58367500e-01 -2.23494731e-02 3.08620363e-01 1.50225031e+00
7.08061337e-01 2.42174476e-01 5.28926253e-02 8.27481091e-01
-2.22524747e-01 1.72952227e-02 -8.66706729e-01 -1.21787429e-01
4.66382056e-01 1.43539989e+00 -1.17415643e+00 -4.10665244e-01
-3.16926122e-01 9.43306565e-01 4.48094666e-01 2.46446759e-01
-3.63023460e-01 -4.69635427e-01 4.39511061e-01 4.51988935e-01
5.66169202e-01 2.24600099e-02 -4.73184109e-01 -7.95051932e-01
-3.28642815e-01 -6.27678514e-01 2.24319890e-01 -4.40594435e-01
-1.23172796e+00 5.01291752e-01 -7.31609941e-01 -1.26819384e+00
7.82788172e-03 -7.33321190e-01 -4.09482628e-01 7.90551364e-01
-1.61695695e+00 -8.57320607e-01 -2.31510162e-01 5.11178613e-01
4.90505844e-01 -4.82334048e-01 1.00754583e+00 8.51618126e-02
-5.63878953e-01 4.30444092e-01 3.63850832e-01 3.01177323e-01
4.20937926e-01 -1.15254641e+00 -1.92222342e-01 6.05848670e-01
1.31974369e-01 6.60792351e-01 4.30119514e-01 -4.82256413e-01
-1.42618680e+00 -1.12835205e+00 1.06759834e+00 -6.71190098e-02
2.44432449e-01 -2.36055732e-01 -7.86058366e-01 3.71397614e-01
-5.76645620e-02 7.96022356e-01 9.03189361e-01 -5.90615988e-01
3.95863131e-02 -2.11603567e-01 -1.34775662e+00 3.16653371e-01
5.98693550e-01 -5.70007503e-01 -3.26489680e-03 5.51519752e-01
3.92847776e-01 -3.55958104e-01 -7.52317250e-01 3.08361143e-01
4.33853775e-01 -8.85388970e-01 1.03604400e+00 -1.09113783e-01
6.10947669e-01 -5.06142259e-01 -4.28561062e-01 -9.29823577e-01
-4.83474225e-01 -1.07499644e-01 -3.74175817e-01 5.27132750e-01
2.85216987e-01 -4.42922503e-01 1.19421124e+00 2.52647325e-02
-3.04293513e-01 -1.52091897e+00 -9.19371128e-01 -1.99225858e-01
-5.75572141e-02 6.85085356e-02 2.63766408e-01 8.48610878e-01
1.50891632e-01 1.49407461e-01 -1.85013205e-01 5.74683100e-02
5.86058319e-01 3.65376711e-01 3.46170574e-01 -9.57591295e-01
-2.45135725e-01 -5.04083455e-01 -7.42912889e-01 -1.27627063e+00
-5.16789258e-02 -1.10847723e+00 1.09390646e-01 -1.34117389e+00
3.92237633e-01 -4.38177258e-01 -6.96051896e-01 4.41867560e-01
5.47341220e-02 6.58957779e-01 2.38830652e-02 5.10230780e-01
-7.86332965e-01 1.08988300e-01 1.19921613e+00 -2.84974843e-01
-1.25512788e-02 -2.97631651e-01 -7.24656880e-01 4.83454317e-01
7.91662037e-01 -4.78551596e-01 8.00772086e-02 -2.07614690e-01
-5.23167625e-02 2.31991336e-01 4.06514823e-01 -1.39734852e+00
7.41681933e-01 -2.13448554e-02 8.86247814e-01 -6.53084695e-01
4.73360002e-01 -5.99662900e-01 6.15817867e-02 7.70907998e-01
-4.43528354e-01 -3.11944634e-01 -2.57726520e-01 6.53721869e-01
-4.19818521e-01 -1.45063028e-02 9.52131510e-01 -4.55041736e-01
-4.18717504e-01 3.41348171e-01 -5.23617685e-01 -3.53908032e-01
1.14200437e+00 -7.62991130e-01 -1.92932963e-01 -4.30238955e-02
-3.95383894e-01 -1.35836139e-01 5.82259655e-01 -4.23326075e-01
1.11232555e+00 -1.11918414e+00 -4.87167448e-01 4.73275483e-01
9.07068700e-02 7.07645714e-02 1.58531487e-01 1.03439867e+00
-8.36820364e-01 8.64341915e-01 -1.55395612e-01 -9.19392586e-01
-1.05259979e+00 5.85637450e-01 3.72644246e-01 -2.48476654e-01
-6.44588351e-01 1.18516219e+00 1.44107386e-01 -1.51691869e-01
2.63183385e-01 -3.26820314e-01 -3.52450281e-01 -3.79897416e-01
6.66129231e-01 6.69678152e-02 7.62622729e-02 -8.04250956e-01
-4.59007114e-01 7.22074687e-01 -7.78367668e-02 3.48397434e-01
1.43840480e+00 1.56195283e-01 -2.98974842e-01 3.75948757e-01
1.62623334e+00 -1.74205095e-01 -1.30278707e+00 -3.64249974e-01
-2.15005711e-01 -4.01443332e-01 2.83558488e-01 -6.07419491e-01
-1.41090858e+00 8.57319057e-01 7.49845982e-01 -1.06253056e-02
1.05357254e+00 1.05596803e-01 9.57686782e-01 2.72802681e-01
3.77358317e-01 -8.45883906e-01 8.44330043e-02 3.78421605e-01
4.19656098e-01 -1.28104210e+00 2.34397165e-02 -3.72813135e-01
-3.91434491e-01 1.04685700e+00 4.32890415e-01 -2.43576676e-01
8.38416636e-01 4.24302995e-01 -4.05694433e-02 -4.86627430e-01
-8.03621829e-01 -3.37310359e-02 -1.16635956e-01 4.80992913e-01
6.32119060e-01 8.56027976e-02 -4.61782962e-02 3.24614644e-01
1.34567752e-01 3.23714375e-01 1.08683817e-01 1.00784397e+00
-7.72458434e-01 -6.09322608e-01 -3.25212985e-01 9.07793105e-01
-6.18158162e-01 -4.83479910e-02 -3.09436411e-01 1.66850448e-01
3.18106532e-01 6.00633979e-01 2.06693649e-01 -3.23605776e-01
-2.35126168e-01 -1.90106630e-01 2.32151553e-01 -6.48059905e-01
-2.79484361e-01 2.71469235e-01 -8.11459064e-01 -5.34460247e-01
-3.67801100e-01 -7.93194234e-01 -1.53620124e+00 -5.89367002e-02
-2.03381166e-01 -1.37358904e-01 5.09907722e-01 8.01242292e-01
3.76779914e-01 2.87119389e-01 5.91290414e-01 -8.89152825e-01
-2.10093394e-01 -7.65694439e-01 -8.34305763e-01 3.11096400e-01
5.59876442e-01 -3.53738308e-01 -3.00420284e-01 2.97035187e-01] | [14.72002124786377, -3.1635472774505615] |
65ac809e-c35a-4daf-8a45-4a927e0d6eee | towards-low-latency-energy-efficient-deep | 2107.12445 | null | https://arxiv.org/abs/2107.12445v1 | https://arxiv.org/pdf/2107.12445v1.pdf | Towards Low-Latency Energy-Efficient Deep SNNs via Attention-Guided Compression | Deep spiking neural networks (SNNs) have emerged as a potential alternative to traditional deep learning frameworks, due to their promise to provide increased compute efficiency on event-driven neuromorphic hardware. However, to perform well on complex vision applications, most SNN training frameworks yield large inference latency which translates to increased spike activity and reduced energy efficiency. Hence,minimizing average spike activity while preserving accuracy indeep SNNs remains a significant challenge and opportunity.This paper presents a non-iterative SNN training technique thatachieves ultra-high compression with reduced spiking activitywhile maintaining high inference accuracy. In particular, our framework first uses the attention-maps of an un compressed meta-model to yield compressed ANNs. This step can be tuned to support both irregular and structured channel pruning to leverage computational benefits over a broad range of platforms. The framework then performs sparse-learning-based supervised SNN training using direct inputs. During the training, it jointly optimizes the SNN weight, threshold, and leak parameters to drastically minimize the number of time steps required while retaining compression. To evaluate the merits of our approach, we performed experiments with variants of VGG and ResNet, on both CIFAR-10 and CIFAR-100, and VGG16 on Tiny-ImageNet.The SNN models generated through the proposed technique yield SOTA compression ratios of up to 33.4x with no significant drops in accuracy compared to baseline unpruned counterparts. Compared to existing SNN pruning methods, we achieve up to 8.3x higher compression with improved accuracy. | ['Peter A. Beerel', 'Massoud Pedram', 'Gourav Datta', 'Souvik Kundu'] | 2021-07-16 | null | null | null | null | ['sparse-learning'] | ['methodology'] | [ 4.48988020e-01 -1.38274193e-01 8.49957764e-02 -1.98267415e-01
-5.22304296e-01 -2.38930523e-01 3.45780730e-01 1.61602676e-01
-8.41157734e-01 8.64363194e-01 -2.55151689e-01 -2.31828213e-01
-1.55843586e-01 -7.74884284e-01 -9.85628724e-01 -7.85554290e-01
1.63403839e-01 9.62374881e-02 4.61571604e-01 2.23382935e-01
5.31663537e-01 5.41982472e-01 -1.76639497e+00 3.53825659e-01
7.04802692e-01 1.38194430e+00 5.14492571e-01 5.64454556e-01
-1.06756771e-02 7.33112574e-01 -5.39829493e-01 -3.08125794e-01
4.16166753e-01 -2.59729475e-01 -1.14727557e-01 -5.18470168e-01
4.94015336e-01 -3.00344884e-01 -7.27605402e-01 1.04771924e+00
5.81196368e-01 -1.71309084e-01 2.32599318e-01 -9.48573589e-01
-1.85985610e-01 9.59491313e-01 -3.71942490e-01 6.62996352e-01
-4.78618681e-01 2.78425336e-01 6.96145535e-01 -7.08167553e-01
3.24686229e-01 7.12666154e-01 7.73373544e-01 5.62659860e-01
-1.53462934e+00 -1.08648622e+00 -2.61936545e-01 1.58633038e-01
-1.16066110e+00 -9.08852756e-01 3.46974283e-01 2.14458391e-01
1.71676302e+00 -8.74976143e-02 1.11899805e+00 9.80505407e-01
4.79797959e-01 3.88648152e-01 9.93643105e-01 -9.76462886e-02
8.36520851e-01 -4.46237504e-01 2.06859156e-01 4.08930749e-01
8.88729632e-01 -1.53945452e-02 -1.03531909e+00 -3.48731540e-02
8.01783442e-01 4.43534464e-01 -7.99144804e-02 1.60021365e-01
-7.32294679e-01 4.21530455e-01 5.62434912e-01 8.13110471e-02
-4.19162333e-01 9.09368038e-01 3.97020578e-01 3.82133089e-02
-2.15570077e-01 4.02112037e-01 -3.06587040e-01 -4.49176848e-01
-1.36019135e+00 4.60017443e-01 7.31727839e-01 7.09542513e-01
7.72381425e-01 6.42066061e-01 -2.95706429e-02 8.67847025e-01
2.46475916e-02 6.72609150e-01 6.87151790e-01 -1.09554601e+00
3.99636179e-01 6.77365720e-01 -6.00020945e-01 -7.16339231e-01
-2.58547425e-01 -7.23975003e-01 -9.62301433e-01 2.44539276e-01
1.29269242e-01 9.17061344e-02 -1.28040421e+00 1.76361644e+00
-3.56311440e-01 4.46747661e-01 7.63202906e-02 5.45230985e-01
5.59120059e-01 5.42674005e-01 6.87002167e-02 -3.80208641e-02
1.12959838e+00 -5.47260821e-01 -5.17869651e-01 -7.26128340e-01
4.41443264e-01 -3.14978391e-01 6.42766953e-01 3.41210186e-01
-1.39400172e+00 -2.07377344e-01 -1.42421162e+00 -1.94846451e-01
-1.25814274e-01 -1.58459291e-01 6.64040267e-01 4.33055460e-01
-1.21840513e+00 9.80683267e-01 -1.34209752e+00 -1.90615594e-01
1.24263024e+00 9.04278219e-01 9.72518399e-02 -3.21547850e-03
-5.13804913e-01 5.69765031e-01 5.34271300e-01 -2.34035268e-01
-1.06733537e+00 -9.22114015e-01 -4.92743194e-01 6.16097033e-01
-7.37772435e-02 -5.17002761e-01 1.23906553e+00 -4.81694907e-01
-1.26232719e+00 6.41124964e-01 -2.93285191e-01 -1.49232018e+00
-1.78666115e-01 -9.98360887e-02 -5.47565222e-02 2.45361522e-01
-1.93675548e-01 1.00588262e+00 4.52714384e-01 -5.64177275e-01
-4.75180864e-01 -5.39362311e-01 -2.64802396e-01 -8.82154256e-02
-5.84909618e-01 -2.72090256e-01 -3.26705277e-01 -7.81366348e-01
3.71384889e-01 -7.96797097e-01 -2.05840677e-01 7.66753033e-02
-1.51289552e-01 2.24503189e-01 9.05813813e-01 -2.14535370e-01
9.89930570e-01 -2.16743755e+00 -1.77178383e-01 7.57896677e-02
4.06246662e-01 5.25704622e-01 -4.50825915e-02 -1.99033588e-01
3.68633717e-01 -4.44747657e-02 -6.11966848e-01 -3.40334564e-01
-3.99466217e-01 4.28830087e-01 -3.97348255e-01 2.64424592e-01
2.75081784e-01 1.18323398e+00 -4.57952499e-01 -2.12889612e-01
7.70223513e-02 5.51333010e-01 -9.07049298e-01 -1.37529939e-01
-2.55429804e-01 -9.91353169e-02 -5.50674163e-02 8.71795237e-01
5.59338391e-01 -3.36764097e-01 1.87051341e-01 -1.65214434e-01
-1.56810641e-01 4.63080376e-01 -5.52991450e-01 1.84577465e+00
-2.76284963e-01 8.12086642e-01 -5.68161123e-02 -1.19623566e+00
1.10256612e+00 -1.37132660e-01 4.01269555e-01 -1.20220363e+00
4.77104515e-01 7.13330388e-01 3.10535759e-01 2.50104517e-01
1.97819933e-01 1.53200209e-01 1.75551802e-01 3.26025784e-01
5.34317195e-01 2.83523500e-01 1.52290121e-01 1.80900395e-01
1.58698201e+00 -2.22020969e-01 -2.36384608e-02 -4.34036463e-01
-7.04819113e-02 -2.05697730e-01 5.59460402e-01 7.99101949e-01
-1.17087819e-01 2.53580511e-01 3.07677120e-01 -2.90505946e-01
-1.13242948e+00 -1.00559580e+00 -3.79207820e-01 8.05941463e-01
7.28235692e-02 -3.74329656e-01 -8.87178242e-01 2.73691148e-01
-1.42537653e-01 4.82440501e-01 -2.77341664e-01 -4.72506344e-01
-7.08696425e-01 -9.55143034e-01 1.14528584e+00 5.89760900e-01
9.19499993e-01 -1.12866986e+00 -1.51745450e+00 4.38930780e-01
2.88166225e-01 -1.35230184e+00 1.06949411e-01 1.00855291e+00
-1.48909473e+00 -6.45778775e-01 -2.89958954e-01 -8.90035033e-01
4.59228605e-01 4.35243510e-02 8.96083772e-01 4.38392982e-02
-5.63255727e-01 -3.26553017e-01 2.82942653e-02 -5.95617771e-01
2.23333642e-01 2.52068013e-01 -6.81514144e-02 -4.45860267e-01
6.50458634e-01 -1.35022426e+00 -8.00320625e-01 -2.33651027e-01
-9.04643774e-01 8.20567608e-02 8.38618457e-01 8.31283271e-01
1.27398968e+00 -3.30138087e-01 6.97640002e-01 -8.03400993e-01
1.89165130e-01 -3.03027600e-01 -7.85734236e-01 -1.18658707e-01
-8.10887098e-01 3.62754256e-01 9.76563871e-01 -4.60565060e-01
-5.85536420e-01 7.87395909e-02 -1.69681445e-01 -5.22289813e-01
1.06086522e-01 2.61769861e-01 4.86714244e-02 -4.71284419e-01
7.49590158e-01 5.08853376e-01 -2.08601519e-01 -2.50146657e-01
-2.25330859e-01 2.27309838e-01 8.10901344e-01 -2.71993250e-01
3.30941528e-01 6.16094530e-01 1.49245650e-01 -5.55753410e-01
-3.79411399e-01 9.77892801e-02 -1.91747758e-03 1.94831997e-01
3.63803685e-01 -8.87040913e-01 -6.73691094e-01 5.99995971e-01
-1.05404341e+00 -4.92753714e-01 -4.09352690e-01 3.62535208e-01
-4.38550532e-01 -2.10479736e-01 -8.81478369e-01 -6.50250673e-01
-9.78916645e-01 -8.66721392e-01 6.26514912e-01 5.51785648e-01
1.11183763e-01 -4.16221291e-01 -1.25679314e-01 -2.36224055e-01
8.13136101e-01 4.68444414e-02 1.01552272e+00 -7.70857990e-01
-9.78246987e-01 1.59651637e-02 -4.45285678e-01 9.40786824e-02
-3.52652848e-01 -3.82043570e-01 -1.17773318e+00 -2.89791316e-01
1.00025393e-01 -6.22894406e-01 1.35899627e+00 7.23169923e-01
1.63372278e+00 -3.70778084e-01 -3.56209368e-01 1.31102479e+00
1.92061567e+00 3.34986001e-01 8.90388846e-01 1.95859715e-01
3.68739724e-01 -2.06906319e-01 -1.58434585e-01 6.67397738e-01
-8.80287439e-02 3.13960940e-01 7.40310550e-01 4.52272415e-01
-3.22903007e-01 -2.29259565e-01 1.87460005e-01 9.86889422e-01
8.56986046e-02 -1.07822798e-01 -8.46181273e-01 5.71033597e-01
-1.58062959e+00 -8.52769971e-01 3.06769073e-01 2.26036954e+00
8.85417163e-01 4.63521093e-01 -3.20576459e-01 3.33806992e-01
3.93405199e-01 5.63361421e-02 -1.33047163e+00 -4.72494572e-01
-4.26429302e-01 1.02476728e+00 8.71552765e-01 -1.20882906e-01
-5.28344214e-01 8.37488174e-01 5.50609446e+00 7.06730187e-01
-1.30236077e+00 1.03791229e-01 6.56977654e-01 -8.32502246e-01
-1.77177683e-01 -1.69189349e-02 -1.26311350e+00 5.92488706e-01
1.63565087e+00 -2.94950247e-01 9.30444121e-01 7.58633852e-01
-1.52245909e-01 -9.40245837e-02 -9.22645926e-01 1.38037360e+00
-1.93416685e-01 -1.93403172e+00 5.90153001e-02 1.54967904e-01
5.89772284e-01 5.50719738e-01 5.07563092e-02 2.82157183e-01
5.50668575e-02 -1.19960928e+00 6.78933918e-01 2.29837254e-01
9.58283365e-01 -9.38327253e-01 5.93510866e-01 3.18549901e-01
-1.02117336e+00 -3.96703184e-01 -8.15335333e-01 -1.51676074e-01
-2.01302748e-02 7.84189820e-01 -4.72101927e-01 -2.67694086e-01
1.11219966e+00 7.33078480e-01 -4.87833291e-01 1.35348928e+00
2.99521148e-01 6.65671170e-01 -6.97633564e-01 -2.41189510e-01
-1.13376358e-03 1.59205675e-01 2.40944102e-01 1.10398471e+00
6.02670968e-01 2.30571985e-01 -3.87849182e-01 1.04542696e+00
-7.81537414e-01 -3.61123830e-01 -6.35242701e-01 -2.35240966e-01
1.11414897e+00 1.03499758e+00 -9.82296526e-01 -1.88581169e-01
-1.42430559e-01 7.65406728e-01 5.38004518e-01 2.13638276e-01
-8.11762154e-01 -6.29450321e-01 5.69198251e-01 2.73030967e-01
5.38313806e-01 -9.20457765e-02 -1.08419716e+00 -8.38887334e-01
2.91942116e-02 -6.47674084e-01 -1.33094296e-01 -3.38903993e-01
-5.27151167e-01 8.17680478e-01 -5.39328337e-01 -7.18366086e-01
-9.54954922e-02 -5.42935908e-01 -5.06359458e-01 5.62762797e-01
-1.51144326e+00 -5.93701124e-01 -4.49026167e-01 4.20621336e-01
5.31080246e-01 -2.58391500e-01 8.80375087e-01 4.82214272e-01
-6.55191362e-01 8.68472755e-01 5.32232560e-02 -1.55543312e-01
1.21011771e-01 -6.34925723e-01 5.39714456e-01 1.06721234e+00
2.11608201e-01 6.44906521e-01 4.29851502e-01 -4.94209826e-01
-1.60072064e+00 -1.41583776e+00 4.45037931e-01 2.98973709e-01
3.74083370e-01 -5.64189255e-01 -1.12893879e+00 4.96705562e-01
1.21789910e-01 2.80160218e-01 5.16558945e-01 -4.40752447e-01
-5.11094332e-01 -4.67390418e-01 -1.15860415e+00 5.59520304e-01
1.33869648e+00 -3.40198696e-01 -2.62808055e-01 -1.60372630e-02
6.85370803e-01 -2.49723464e-01 -4.89663333e-01 4.47312921e-01
4.92519796e-01 -1.05254269e+00 7.95394599e-01 -1.60694093e-01
5.42787611e-01 3.13865137e-04 -4.20320600e-01 -7.87694097e-01
-1.91688299e-01 -5.22471905e-01 -8.10451329e-01 8.76454473e-01
2.53165483e-01 -5.77540934e-01 1.29333210e+00 4.16376352e-01
-4.27916020e-01 -1.18684590e+00 -1.14473462e+00 -7.25090146e-01
-1.57098010e-01 -4.06202614e-01 3.75460476e-01 2.34879479e-02
-2.80929863e-01 1.13583535e-01 -4.23943670e-03 -1.23224678e-02
8.68016541e-01 -1.08157918e-01 2.52410859e-01 -1.15164578e+00
-3.59412283e-01 -6.87246919e-01 -6.08825505e-01 -8.97912264e-01
-1.48878247e-02 -9.89970624e-01 1.33649945e-01 -1.27060187e+00
2.60564625e-01 -5.09707451e-01 -6.26491845e-01 8.31035435e-01
3.44321072e-01 6.39538348e-01 3.00190896e-02 2.53036082e-01
-5.55513501e-01 4.08729613e-01 4.65786934e-01 1.23552240e-01
-1.26094386e-01 -4.63199288e-01 -6.80381119e-01 6.55564964e-01
9.32457805e-01 -8.10752988e-01 -5.66660166e-01 -8.57783735e-01
1.18085921e-01 -1.60892263e-01 5.12602031e-01 -1.80540550e+00
8.33106160e-01 2.65676081e-01 5.09848118e-01 -5.37579954e-01
4.51144993e-01 -4.66122806e-01 1.14226192e-01 8.75866473e-01
-2.76767522e-01 3.78732271e-02 6.08594835e-01 6.34575188e-01
-8.83699656e-02 -1.02479391e-01 1.10380757e+00 -2.69369222e-03
-4.81734753e-01 4.84457225e-01 -3.62313598e-01 1.37660310e-01
7.94749558e-01 -5.76250255e-01 -5.23613632e-01 1.45162180e-01
-1.89093322e-01 -1.69133842e-01 4.33789641e-01 -1.86268613e-01
1.11208403e+00 -1.06797552e+00 -2.99887270e-01 4.89084333e-01
-2.78775901e-01 1.99689448e-01 2.50887841e-01 7.28958011e-01
-6.89481258e-01 6.10218942e-01 -7.57469296e-01 -6.45987749e-01
-9.02358830e-01 1.08581521e-01 3.88642877e-01 -1.78202674e-01
-6.81884348e-01 1.13100135e+00 -1.59357637e-01 2.43602678e-01
5.51858425e-01 -4.37415749e-01 1.64225802e-01 -5.02409637e-01
7.19964623e-01 3.57604593e-01 4.94589746e-01 9.57004167e-03
-4.33166206e-01 1.77350163e-01 -2.62043059e-01 4.55745570e-02
1.77294803e+00 3.04506421e-01 -1.22550502e-01 1.53794661e-01
1.20067811e+00 -4.81673419e-01 -1.59374714e+00 -1.84953317e-01
1.28259331e-01 -1.37421936e-01 3.67388636e-01 -6.14868820e-01
-1.48443842e+00 1.02799380e+00 8.33797991e-01 -4.05086905e-01
1.46305954e+00 -2.45122001e-01 1.34035397e+00 7.19375849e-01
6.20526731e-01 -9.31295037e-01 1.55039504e-01 6.06513321e-01
3.26342821e-01 -7.79108703e-01 -9.73218903e-02 6.97927326e-02
-9.76717249e-02 1.01409245e+00 8.35070372e-01 -6.45352840e-01
5.55200577e-01 9.73367512e-01 -5.97974479e-01 -1.63950056e-01
-1.07866299e+00 2.07717389e-01 -3.76968980e-01 3.37326527e-01
1.70618016e-02 -8.74382704e-02 -3.84354033e-02 8.35230649e-01
-2.14634359e-01 3.76200229e-01 3.80205244e-01 1.07034326e+00
-8.25037062e-01 -7.85320938e-01 1.78628668e-01 1.28022110e+00
-6.48846984e-01 -5.28735399e-01 1.54235408e-01 2.22314775e-01
-1.49214584e-02 5.28721213e-01 2.96413481e-01 -6.43664002e-01
1.51518837e-01 -8.35983828e-02 5.31017780e-01 -5.73454916e-01
-7.06673682e-01 -3.70532623e-03 -4.74849522e-01 -6.47235096e-01
-6.37103692e-02 -2.97279686e-01 -1.78616250e+00 -5.70840061e-01
-8.92902985e-02 -3.00252110e-01 8.54414046e-01 7.97696471e-01
8.27401936e-01 6.78482056e-01 4.53046896e-02 -8.98214459e-01
-4.98690516e-01 -6.39885306e-01 -4.66675282e-01 -5.77597087e-03
1.23253196e-01 -3.94275874e-01 -2.42773458e-01 -1.19001545e-01] | [8.234722137451172, 2.503411054611206] |
40f27020-d5ef-4cf3-b23c-9ce8c69c39cd | multi-task-identification-of-entities | 1808.09602 | null | http://arxiv.org/abs/1808.09602v1 | http://arxiv.org/pdf/1808.09602v1.pdf | Multi-Task Identification of Entities, Relations, and Coreference for Scientific Knowledge Graph Construction | We introduce a multi-task setup of identifying and classifying entities,
relations, and coreference clusters in scientific articles. We create SciERC, a
dataset that includes annotations for all three tasks and develop a unified
framework called Scientific Information Extractor (SciIE) for with shared span
representations. The multi-task setup reduces cascading errors between tasks
and leverages cross-sentence relations through coreference links. Experiments
show that our multi-task model outperforms previous models in scientific
information extraction without using any domain-specific features. We further
show that the framework supports construction of a scientific knowledge graph,
which we use to analyze information in scientific literature. | ['Hannaneh Hajishirzi', 'Mari Ostendorf', 'Yi Luan', 'Luheng He'] | 2018-08-29 | multi-task-identification-of-entities-1 | https://aclanthology.org/D18-1360 | https://aclanthology.org/D18-1360.pdf | emnlp-2018-10 | ['joint-entity-and-relation-extraction'] | ['natural-language-processing'] | [-7.33574033e-02 4.28540945e-01 -7.04280198e-01 -2.16841057e-01
-1.28225827e+00 -9.99311566e-01 6.49787962e-01 6.24418020e-01
-2.21081167e-01 9.76662219e-01 5.17393708e-01 -4.32828158e-01
-5.66852152e-01 -5.46387672e-01 -9.38371778e-01 -2.49386489e-01
9.08569470e-02 7.16925561e-01 -7.62116686e-02 2.48502538e-01
4.38553512e-01 3.78658801e-01 -7.88723171e-01 6.95629478e-01
1.04556668e+00 2.85044849e-01 1.31109104e-01 5.11813283e-01
-3.79216254e-01 8.04336548e-01 -6.74290121e-01 -8.81928742e-01
-2.69764334e-01 5.36159752e-03 -1.48005188e+00 -8.13585997e-01
5.30036986e-01 6.34675264e-01 -2.84213424e-01 7.16335297e-01
5.40166259e-01 -3.34727485e-03 5.98024845e-01 -1.16830862e+00
-6.74981415e-01 1.41964364e+00 -6.14577055e-01 4.73508328e-01
6.45730078e-01 -5.55305421e-01 1.58431149e+00 -9.63147640e-01
1.50621223e+00 1.30426455e+00 8.57611954e-01 2.96833068e-01
-1.39268768e+00 -1.01964700e+00 -8.02400932e-02 5.17483592e-01
-1.20352018e+00 -6.70835197e-01 5.34221590e-01 -4.48725253e-01
1.24322343e+00 1.91922471e-01 7.39604756e-02 1.59462607e+00
3.43805432e-01 5.55171609e-01 8.77038062e-01 -3.95185053e-01
-2.70121783e-01 -8.74521956e-02 9.50650692e-01 7.17081726e-01
7.16242254e-01 -4.42558825e-01 -1.08423674e+00 -4.01427150e-01
1.14394180e-01 -5.06410718e-01 -4.02845681e-01 7.51833916e-02
-1.38759458e+00 4.58010346e-01 8.81173238e-02 4.59393114e-01
-7.78453574e-02 -2.34507933e-01 6.36630297e-01 3.68409812e-01
4.36633646e-01 1.09471798e+00 -6.46513641e-01 1.77810878e-01
-7.73354471e-01 1.49038434e-01 1.43002844e+00 1.27753508e+00
3.60882342e-01 -6.99703217e-01 -6.55058920e-01 7.63778806e-01
-9.23783630e-02 1.89130783e-01 6.49552867e-02 -1.18434012e+00
4.92309511e-01 5.98723352e-01 -1.16306014e-01 -8.76678705e-01
-8.32552254e-01 -7.28672922e-01 -5.72469413e-01 -4.97590363e-01
2.91886866e-01 -3.01608741e-01 -4.26678807e-01 1.80991030e+00
1.37930736e-01 2.90456206e-01 2.59876996e-01 4.68880534e-01
1.71025825e+00 2.02137426e-01 4.74296868e-01 -2.36298218e-01
1.76222026e+00 -9.33629334e-01 -1.29899371e+00 -3.45372781e-02
8.97982955e-01 -9.14869905e-01 3.62795800e-01 1.80642918e-01
-1.00166011e+00 -2.54907310e-01 -7.88137913e-01 -6.38045490e-01
-7.89927900e-01 3.83695334e-01 8.46653998e-01 -6.90954998e-02
-8.82622898e-01 7.29542196e-01 -3.75379741e-01 -2.90229499e-01
3.79190356e-01 1.15126275e-01 -7.11550117e-01 1.72599405e-01
-1.34011066e+00 1.29483175e+00 3.93751889e-01 -2.45758161e-01
-4.82023716e-01 -1.50367141e+00 -9.05769706e-01 3.34054321e-01
4.99333709e-01 -1.03258657e+00 7.47834623e-01 5.29985800e-02
-9.75627184e-01 1.36539721e+00 -5.48571408e-01 -3.97191912e-01
-2.68987179e-01 -2.80946493e-01 -5.03044188e-01 2.77994908e-02
4.41032529e-01 3.71969312e-01 -1.24282084e-01 -1.17459762e+00
-3.11244309e-01 -3.69241536e-01 -1.59286618e-01 -1.22584626e-01
-1.37530416e-01 5.29955685e-01 -6.36793196e-01 -5.09571075e-01
-5.35200357e-01 -8.04128587e-01 1.40924305e-01 -5.47419310e-01
-1.26071048e+00 -8.00846756e-01 5.14143944e-01 -7.12574422e-01
1.19646895e+00 -1.63393164e+00 5.46855927e-01 1.54377982e-01
6.87778711e-01 -2.96936333e-02 -3.99028271e-01 4.47564542e-01
-4.54476655e-01 4.94107932e-01 7.76278749e-02 -5.72354198e-01
-1.68884218e-01 4.48453426e-02 -2.49523222e-01 7.62694851e-02
1.09340668e-01 1.00627041e+00 -1.10050523e+00 -8.13178360e-01
-5.46681702e-01 1.43331394e-01 -7.50216767e-02 1.61151752e-01
-2.18706235e-01 5.04655242e-01 -8.25578213e-01 6.01782441e-01
5.24398685e-01 -7.76471317e-01 5.47967196e-01 -5.94079912e-01
-1.63223907e-01 9.21182454e-01 -6.96330130e-01 2.11275005e+00
-4.18528944e-01 6.02845430e-01 2.15201259e-01 -1.19032252e+00
8.41066778e-01 4.13001746e-01 5.96914589e-01 -1.37267277e-01
-3.06751102e-01 1.26767933e-01 -1.59275699e-02 -7.28775561e-01
9.95477065e-02 4.64852989e-01 -2.70406514e-01 5.10436475e-01
7.20363259e-01 2.42522284e-01 6.22021914e-01 8.83003533e-01
1.52211571e+00 3.38323385e-01 3.05498749e-01 -6.51903391e-01
6.41440392e-01 2.96924502e-01 8.91203046e-01 8.24794114e-01
4.66402501e-01 3.78389768e-02 8.50683153e-01 -2.10611388e-01
-5.63564718e-01 -7.54325271e-01 -5.36426067e-01 1.08639503e+00
-2.94774681e-01 -9.63251829e-01 -1.66805357e-01 -1.05663633e+00
3.64058524e-01 9.03213382e-01 -7.43200839e-01 6.40280396e-02
-5.05496979e-01 -7.60044098e-01 9.88568902e-01 2.18624040e-01
-6.11951202e-02 -7.96047211e-01 8.59555230e-02 -6.73947409e-02
-5.94133496e-01 -1.51371062e+00 -3.99956763e-01 3.47886205e-01
-4.92573410e-01 -1.68697226e+00 -2.56026030e-01 -6.61696076e-01
2.67599404e-01 -2.94076018e-02 1.72179925e+00 -3.87026161e-01
-3.92631114e-01 3.75301689e-01 -4.55992557e-02 -4.81560767e-01
-2.48650923e-01 5.64850211e-01 -2.05837026e-01 -7.81670988e-01
7.73085892e-01 -4.08093452e-01 -1.28733695e-01 -2.79068559e-01
-2.62086153e-01 1.02669194e-01 6.05202377e-01 8.05527449e-01
5.17237425e-01 -6.67332172e-01 1.00674760e+00 -1.57790518e+00
5.90612710e-01 -8.84665906e-01 -2.31727898e-01 9.01643276e-01
-6.71439946e-01 3.88068348e-01 4.47485030e-01 1.07053727e-01
-1.36589956e+00 -3.40029985e-01 1.01206943e-01 -2.87734389e-01
-2.51776259e-02 9.54844654e-01 -2.01016411e-01 -1.94124237e-01
5.08811533e-01 -4.38312680e-01 -3.46971095e-01 -8.69008422e-01
6.11605823e-01 5.23392200e-01 8.43501806e-01 -9.30780351e-01
4.54113275e-01 9.54971164e-02 4.60289299e-01 -2.02764288e-01
-1.74365473e+00 -6.37157559e-01 -8.19545150e-01 2.20898554e-01
8.68966401e-01 -1.19567394e+00 -1.25544357e+00 -2.91048378e-01
-1.73675334e+00 2.96491146e-01 1.93180367e-02 4.79963303e-01
3.93422805e-02 4.17889982e-01 -9.29574609e-01 -2.54498273e-01
-8.22054446e-01 -7.50876546e-01 1.04886556e+00 9.32352543e-02
-4.99773234e-01 -1.35739279e+00 3.02968591e-01 6.45045578e-01
-1.89033031e-01 3.97325665e-01 1.18517256e+00 -1.29381216e+00
-2.43435115e-01 3.03050518e-01 -4.79762942e-01 -5.52465558e-01
-8.99137184e-02 1.11484565e-01 -8.30922008e-01 5.10181002e-02
-5.67924798e-01 -5.72152317e-01 1.25165927e+00 1.57996327e-01
1.48146963e+00 -3.03915173e-01 -1.11361074e+00 8.65087271e-01
1.06117320e+00 -1.06618501e-01 2.30058506e-01 4.66260076e-01
9.75351989e-01 8.42342734e-01 1.77038923e-01 9.49400961e-02
7.19164371e-01 5.90175867e-01 -3.13640475e-01 -6.17688000e-02
-3.38048607e-01 -6.93236366e-02 5.66190332e-02 1.01299286e+00
-2.22817883e-01 -2.23146230e-01 -1.11393034e+00 6.35684192e-01
-1.81895018e+00 -8.99133503e-01 -6.66025341e-01 1.69992566e+00
1.71818101e+00 -3.82891774e-01 -3.17031264e-01 -6.98679030e-01
5.23110271e-01 -2.07383409e-01 -4.51564223e-01 -2.92605013e-01
-5.49599826e-01 5.84667802e-01 5.85231006e-01 4.78582680e-01
-1.07567394e+00 1.00146365e+00 7.13946342e+00 8.56124043e-01
-3.61902893e-01 3.12366098e-01 2.87164778e-01 -1.84617250e-03
-6.56118929e-01 1.22769974e-01 -1.16286385e+00 1.66592360e-01
1.09813809e+00 -7.04745829e-01 -1.73102781e-01 4.33897883e-01
-9.04786736e-02 3.35972458e-01 -1.51795578e+00 5.06919444e-01
1.12188391e-01 -1.89243078e+00 2.21063584e-01 -1.38800651e-01
4.01531786e-01 2.23213747e-01 -5.19955277e-01 2.60049015e-01
1.13331974e+00 -1.10715556e+00 -2.35065036e-02 8.24948370e-01
8.35936069e-01 -3.73119652e-01 7.90775597e-01 -1.81481559e-02
-8.10221732e-01 3.34349334e-01 -2.93337673e-01 5.59127450e-01
3.65707725e-02 1.13296914e+00 -3.06676775e-01 1.28548229e+00
7.47541189e-01 1.21834254e+00 -6.41317844e-01 8.60477448e-01
-4.11522180e-01 6.20912313e-01 -2.24161819e-02 3.72900188e-01
-2.33207092e-01 -2.50019014e-01 8.84558260e-01 1.92393768e+00
2.28146851e-01 1.07158028e-01 9.81341675e-02 1.23319387e+00
-8.15523744e-01 7.98880830e-02 -7.76928425e-01 -1.88877493e-01
8.58992696e-01 1.69668984e+00 -1.31318256e-01 -4.03954148e-01
-4.08933848e-01 5.92438877e-01 9.28213179e-01 4.37763989e-01
-3.70859891e-01 -7.10687041e-01 4.44248199e-01 -6.62331522e-01
-1.24910381e-02 4.66018580e-02 -6.85309649e-01 -1.50953662e+00
-4.66545314e-01 -6.51958287e-01 9.80158031e-01 -5.63167512e-01
-1.60621238e+00 2.54630953e-01 -8.41071978e-02 -4.54081506e-01
-1.32297009e-01 -4.04156834e-01 -8.07623744e-01 8.87576878e-01
-1.65610504e+00 -1.49053288e+00 3.16436850e-02 4.47688311e-01
1.84620768e-01 -2.18709260e-01 1.20579588e+00 2.62377858e-01
-1.02026868e+00 7.38071620e-01 -4.96725179e-02 2.88755417e-01
1.37862444e+00 -1.48233747e+00 2.27129608e-01 4.46753561e-01
1.07860342e-01 1.21367717e+00 5.49211383e-01 -8.87432218e-01
-1.62411904e+00 -1.00645268e+00 1.58633208e+00 -8.16318095e-01
1.06449449e+00 -5.12079716e-01 -1.13177979e+00 9.65454102e-01
6.16886795e-01 -2.87806839e-01 1.40040195e+00 1.10584831e+00
-8.82860363e-01 1.98384076e-01 -7.53164709e-01 1.17425650e-01
1.49084675e+00 -4.80421990e-01 -1.00344276e+00 8.78018737e-01
7.67699182e-01 -4.60930884e-01 -1.72575247e+00 5.28221905e-01
2.12026641e-01 -4.14270647e-02 1.16124046e+00 -1.07249784e+00
7.70806074e-01 1.39226109e-01 7.06044495e-01 -1.40711963e+00
-7.98003554e-01 -6.41702831e-01 -4.37369615e-01 1.61112511e+00
9.79432583e-01 -2.81489521e-01 2.11001396e-01 4.54661399e-01
-3.10587525e-01 -3.31215858e-01 -8.46306741e-01 -5.78370631e-01
3.68954450e-01 1.26157522e-01 8.28481317e-02 1.67889750e+00
7.66387463e-01 1.12636256e+00 2.63020699e-03 5.94387576e-02
9.29476857e-01 7.46684611e-01 2.95303673e-01 -1.66015995e+00
-8.77249092e-02 -3.45582843e-01 3.36796433e-01 -4.31478739e-01
1.11864460e+00 -1.52655983e+00 -4.46898788e-01 -1.83935618e+00
7.46491313e-01 -4.47255850e-01 -4.80398238e-01 1.04292929e+00
-5.24219811e-01 -2.06097335e-01 -1.96770713e-01 5.50829411e-01
-1.05683243e+00 2.10739613e-01 9.72616255e-01 -4.91994582e-02
3.01067293e-01 -5.46413064e-01 -1.08389449e+00 5.16601086e-01
3.76608610e-01 -6.92530215e-01 7.83416927e-02 -4.43643391e-01
5.04805028e-01 -1.77491903e-02 -2.81095365e-03 -4.27670926e-01
7.60098100e-01 4.00310718e-02 4.36075687e-01 -3.95293057e-01
-1.85358927e-01 5.53359743e-03 5.58910519e-02 -2.57019848e-02
-9.09825027e-01 -1.57771289e-01 4.64395672e-01 5.60446084e-01
-5.89636303e-02 -2.13142931e-01 2.55637169e-01 -1.08350024e-01
-2.63244212e-01 -7.57286176e-02 -7.65004903e-02 3.81284833e-01
8.07048142e-01 8.14748585e-01 -1.13114047e+00 1.68987513e-01
-8.32796216e-01 8.52551818e-01 -1.32883862e-01 4.52344567e-01
1.73700511e-01 -9.44104433e-01 -1.09920597e+00 -6.37721539e-01
1.67317748e-01 -2.63074577e-01 6.22268952e-02 8.30631554e-01
2.36398384e-01 7.96549857e-01 1.85233746e-02 -3.47748734e-02
-1.49624681e+00 6.42531514e-01 -3.16072367e-02 -8.85572731e-01
-6.76884234e-01 8.50028694e-01 -1.77922025e-02 -5.52827477e-01
1.20909646e-01 1.94419980e-01 -8.14410925e-01 4.15487468e-01
2.67810702e-01 2.56697297e-01 1.72666356e-01 -2.25215971e-01
-6.41642988e-01 4.60874110e-01 -4.81351465e-01 3.27988446e-01
1.51548326e+00 5.94225377e-02 -8.06625545e-01 5.05679727e-01
1.01841068e+00 4.19053137e-01 -2.06860498e-01 -5.30621171e-01
7.00802684e-01 1.90609276e-01 1.45926431e-01 -1.28626215e+00
-9.33867097e-01 5.07952809e-01 -3.78637761e-01 -2.09620357e-01
4.84331578e-01 4.12456959e-01 6.18026733e-01 7.91630566e-01
4.15438134e-03 -9.32206392e-01 -5.08849442e-01 6.44198537e-01
9.37186837e-01 -1.02373230e+00 3.21867704e-01 -9.52697635e-01
-5.80712140e-01 1.20876336e+00 5.71548700e-01 1.77163973e-01
6.63627803e-01 7.92012453e-01 -2.48546630e-01 -8.19331825e-01
-1.20276451e+00 -2.86194593e-01 7.51568675e-01 4.81740981e-01
1.07033372e+00 -2.19597921e-01 -7.64491379e-01 1.18093526e+00
-5.67540452e-02 1.79514036e-01 3.44353825e-01 4.67972517e-01
1.58397526e-01 -1.34566176e+00 2.67894063e-02 5.51699042e-01
-9.61680233e-01 -4.59472358e-01 -8.16990495e-01 4.73600149e-01
5.67883402e-02 9.14761424e-01 3.37625407e-02 -1.95958719e-01
1.74916834e-01 3.64443183e-01 3.43589067e-01 -7.27353692e-01
-9.58544731e-01 -2.74629682e-01 9.03829336e-01 -5.97776413e-01
-6.38745070e-01 -5.23230493e-01 -1.14438975e+00 -2.82541931e-01
-6.43536747e-02 6.71480715e-01 4.13396895e-01 9.64585423e-01
1.30397964e+00 9.30499792e-01 1.15098983e-01 -1.99537933e-01
-5.47091961e-02 -1.11234844e+00 -1.16345525e-01 5.61717808e-01
-3.98291089e-02 -7.21699595e-01 -1.78089887e-01 9.92224365e-02] | [9.035056114196777, 8.4945707321167] |
92dc5b33-8235-44e8-ae97-7d387f669013 | nettailor-tuning-the-architecture-not-just-1 | 1907.00274 | null | https://arxiv.org/abs/1907.00274v1 | https://arxiv.org/pdf/1907.00274v1.pdf | NetTailor: Tuning the Architecture, Not Just the Weights | Real-world applications of object recognition often require the solution of multiple tasks in a single platform. Under the standard paradigm of network fine-tuning, an entirely new CNN is learned per task, and the final network size is independent of task complexity. This is wasteful, since simple tasks require smaller networks than more complex tasks, and limits the number of tasks that can be solved simultaneously. To address these problems, we propose a transfer learning procedure, denoted NetTailor, in which layers of a pre-trained CNN are used as universal blocks that can be combined with small task-specific layers to generate new networks. Besides minimizing classification error, the new network is trained to mimic the internal activations of a strong unconstrained CNN, and minimize its complexity by the combination of 1) a soft-attention mechanism over blocks and 2) complexity regularization constraints. In this way, NetTailor can adapt the network architecture, not just its weights, to the target task. Experiments show that networks adapted to simple tasks, such as character or traffic sign recognition, become significantly smaller than those adapted to hard tasks, such as fine-grained recognition. More importantly, due to the modular nature of the procedure, this reduction in network complexity is achieved without compromise of either parameter sharing across tasks, or classification accuracy. | ['Pedro Morgado', 'Nuno Vasconcelos'] | 2019-06-29 | nettailor-tuning-the-architecture-not-just | http://openaccess.thecvf.com/content_CVPR_2019/html/Morgado_NetTailor_Tuning_the_Architecture_Not_Just_the_Weights_CVPR_2019_paper.html | http://openaccess.thecvf.com/content_CVPR_2019/papers/Morgado_NetTailor_Tuning_the_Architecture_Not_Just_the_Weights_CVPR_2019_paper.pdf | cvpr-2019-6 | ['traffic-sign-recognition'] | ['computer-vision'] | [ 4.22336847e-01 9.43192318e-02 4.70647663e-02 -4.02861565e-01
-1.06786653e-01 -4.77184117e-01 3.45586747e-01 -2.77135700e-01
-6.87467396e-01 7.31684923e-01 -2.80025840e-01 -7.92194977e-02
-8.72144997e-02 -6.76102519e-01 -9.37373817e-01 -7.02858210e-01
4.31505948e-01 2.70969063e-01 5.14993012e-01 2.17121877e-02
1.04518726e-01 6.55169129e-01 -1.48652041e+00 2.80968159e-01
9.51490521e-01 1.24225652e+00 5.21690845e-01 4.96061414e-01
-8.88910890e-02 7.87108183e-01 -6.74570382e-01 -4.09712791e-01
1.99311852e-01 -2.69157350e-01 -6.78847373e-01 2.65379101e-01
4.42284167e-01 -3.51920202e-02 -2.72334516e-01 9.70460355e-01
3.18914026e-01 3.00541133e-01 4.61989224e-01 -1.11343980e+00
-5.26934147e-01 5.48931360e-01 -2.31535897e-01 1.10090459e-02
-4.88380641e-01 2.54547596e-01 8.83160174e-01 -7.16769397e-01
2.18536317e-01 1.15346587e+00 5.31707942e-01 8.88377845e-01
-1.48017979e+00 -7.62028277e-01 5.68902731e-01 4.43468317e-02
-1.21740127e+00 -5.97241700e-01 6.16910338e-01 -3.96613628e-01
8.52105618e-01 1.34164542e-01 5.33579886e-01 1.02027869e+00
3.46791930e-02 4.98878002e-01 7.06093490e-01 -2.74155498e-01
3.19467068e-01 2.04786241e-01 -8.28281697e-03 5.78741789e-01
4.87236410e-01 -3.48669678e-01 -2.76273936e-01 7.46141374e-02
8.49918008e-01 5.22614717e-02 -2.37984657e-01 -6.07396960e-01
-1.10901046e+00 6.49562061e-01 6.58503473e-01 4.17077154e-01
-2.84789175e-01 2.63115644e-01 4.36861515e-01 4.58252281e-01
2.47406244e-01 7.70631850e-01 -7.73982823e-01 2.25910038e-01
-6.89562023e-01 -1.60231106e-02 7.83380747e-01 6.66294098e-01
1.06858289e+00 2.70026237e-01 -8.74020234e-02 1.02029204e+00
1.95107892e-01 1.97111875e-01 6.86982274e-01 -8.04753125e-01
4.85995293e-01 7.64342248e-01 -2.64161795e-01 -7.30362117e-01
-3.83612007e-01 -7.00232387e-01 -1.04950988e+00 3.16068649e-01
6.13091528e-01 -3.01415920e-01 -9.96720076e-01 1.97464502e+00
1.16667069e-01 -1.28278121e-01 -1.87519804e-01 7.77743042e-01
3.75423163e-01 4.65029478e-01 6.74792975e-02 1.18459322e-01
1.26828945e+00 -1.21276736e+00 -2.97148466e-01 -8.43419135e-01
5.62103271e-01 -4.62013036e-01 1.13319016e+00 3.00235093e-01
-1.05366349e+00 -6.35648429e-01 -9.89589512e-01 8.02114885e-03
-4.06852990e-01 2.33930036e-01 5.40742874e-01 4.33707178e-01
-9.04691935e-01 5.32575190e-01 -5.42391837e-01 -2.01900810e-01
7.27352202e-01 7.05182433e-01 -2.49497756e-01 -5.66750392e-03
-8.52237165e-01 9.32140291e-01 7.85860956e-01 3.45674187e-01
-8.64055932e-01 -6.00278676e-01 -7.38638759e-01 6.05962217e-01
5.75101554e-01 -7.77010441e-01 1.15370667e+00 -1.50049996e+00
-1.69860935e+00 6.64222896e-01 -6.18124520e-03 -2.65946150e-01
2.98159450e-01 2.63090227e-02 -7.83560947e-02 -1.71083570e-01
-2.36083865e-01 6.50542140e-01 1.35744417e+00 -1.09466434e+00
-5.90821028e-01 -1.19635694e-01 2.74329424e-01 9.09237936e-02
-7.73523092e-01 -2.14043595e-02 -6.86741948e-01 -6.36510372e-01
-1.54601663e-01 -1.08153152e+00 -4.52063620e-01 2.28581563e-01
-1.71318471e-01 -1.86017871e-01 7.46468306e-01 -2.11036339e-01
1.07599795e+00 -2.48799896e+00 4.40372527e-01 3.48384142e-01
4.48054641e-01 5.51488698e-01 -5.05547762e-01 -1.62009776e-01
-1.44140914e-01 1.03601627e-01 -1.84439585e-01 -3.08237642e-01
-1.08178578e-01 3.52478266e-01 -2.74736546e-02 1.16781555e-01
4.94520456e-01 1.08253789e+00 -6.54151678e-01 -1.61640108e-01
-3.23332511e-02 1.81483746e-01 -5.84096432e-01 1.30173609e-01
-3.98803473e-01 3.45856637e-01 -5.43472826e-01 2.26096556e-01
5.12686014e-01 -6.37691319e-01 3.70764256e-01 -2.71446705e-01
1.74188137e-01 1.90051883e-01 -1.07655787e+00 1.37343180e+00
-8.02789330e-01 6.03792965e-01 3.08261901e-01 -1.45642376e+00
9.63005364e-01 1.51831806e-01 5.96755706e-02 -6.48514271e-01
1.46810636e-01 2.98170775e-01 3.06824148e-01 -2.99791127e-01
-9.20804963e-02 5.41757345e-02 -4.02183533e-02 5.48593044e-01
1.20660067e-01 1.54672163e-02 3.11598480e-01 -1.77002102e-01
1.02877474e+00 -1.44575149e-01 1.43555492e-01 -1.31396949e-01
5.66950381e-01 -2.29109153e-01 7.55151093e-01 7.09729671e-01
-1.46777434e-02 2.59870768e-01 5.54598451e-01 -5.72583199e-01
-1.14207458e+00 -6.83778465e-01 8.23260248e-02 1.47359657e+00
-2.37075895e-01 -9.04458761e-02 -8.02210629e-01 -7.62473464e-01
2.40556210e-01 4.37451564e-02 -7.36714721e-01 -4.20374662e-01
-9.92181063e-01 -6.71698928e-01 5.22121727e-01 5.40131211e-01
5.09224534e-01 -1.38298404e+00 -5.61257124e-01 3.06977361e-01
2.05560431e-01 -1.27911747e+00 -6.96877718e-01 5.29542804e-01
-1.08652854e+00 -9.77719545e-01 -7.54823208e-01 -1.02262282e+00
1.00191581e+00 2.85023779e-01 9.04805779e-01 2.76455581e-01
-1.57115266e-01 -1.39651582e-01 -1.18831456e-01 -1.82762921e-01
-2.25430176e-01 5.08594573e-01 -1.15603074e-01 5.73036909e-01
1.26398116e-01 -6.23072922e-01 -4.22152936e-01 4.23463821e-01
-9.61842418e-01 4.92949039e-02 1.02033710e+00 1.08846521e+00
2.59365916e-01 -1.01983674e-01 6.90912902e-01 -8.63591373e-01
6.14445031e-01 -1.72064468e-01 -7.96495974e-01 3.78181934e-01
-4.36615855e-01 2.86220998e-01 9.48889494e-01 -9.70166445e-01
-9.61644709e-01 3.26432496e-01 2.13884577e-01 -4.97310340e-01
7.68936723e-02 4.72143799e-01 -2.34519809e-01 -4.92576659e-01
7.37900257e-01 2.53094912e-01 1.82991311e-01 -5.25548041e-01
1.63420632e-01 4.70419347e-01 2.54767090e-01 -4.24674630e-01
8.80437791e-01 1.69223130e-01 -1.16965286e-01 -6.29945159e-01
-9.40683901e-01 9.65562242e-04 -6.96185172e-01 -1.65500883e-02
8.45294058e-01 -7.73360491e-01 -9.21198308e-01 6.64781988e-01
-1.15269732e+00 -7.62239516e-01 -3.85707021e-01 4.01276469e-01
-3.16881448e-01 6.36145798e-03 -4.26906317e-01 -3.36260945e-01
-1.73007756e-01 -1.11058915e+00 6.32351279e-01 2.93064117e-01
-1.10642239e-01 -1.03776443e+00 -4.40697998e-01 2.60662347e-01
7.88347125e-01 -1.50224328e-01 1.02057910e+00 -5.58672607e-01
-7.01080024e-01 -2.19471321e-01 -5.78198373e-01 8.83568406e-01
2.14264080e-01 -2.20459223e-01 -9.58511055e-01 -4.87261236e-01
1.63300615e-02 -4.61143076e-01 9.69466686e-01 1.98094115e-01
1.57076585e+00 -3.99602175e-01 -2.88174093e-01 7.63115108e-01
1.06819105e+00 1.02949396e-01 5.59198141e-01 4.30733860e-01
8.20134282e-01 5.01922488e-01 -4.52696830e-02 2.68664390e-01
1.81856036e-01 7.47213721e-01 3.27741951e-01 -2.05005527e-01
-2.51135856e-01 2.25514378e-02 3.65456283e-01 7.55418539e-01
-1.25827730e-01 2.17584614e-02 -8.33728611e-01 2.87561864e-01
-1.97024441e+00 -7.08061934e-01 3.94208163e-01 2.10906720e+00
8.36441278e-01 2.91890353e-01 -5.65864667e-02 -1.90696064e-02
7.72395909e-01 1.11125767e-01 -8.57780218e-01 -4.61217135e-01
-8.03904012e-02 2.41451353e-01 3.93720597e-01 2.55177706e-01
-8.53324413e-01 1.03075039e+00 6.46125269e+00 8.60422254e-01
-1.52488518e+00 1.39381468e-01 5.05024135e-01 -3.02862734e-01
1.14026824e-02 -1.29642785e-01 -7.13774741e-01 5.34130096e-01
7.57051110e-01 -9.37731639e-02 7.91623056e-01 8.80446374e-01
5.45742735e-02 2.36297280e-01 -1.19912815e+00 8.03230286e-01
-1.22377109e-02 -1.35618639e+00 1.39992744e-01 4.43697609e-02
5.67566991e-01 1.01193637e-01 -6.55719340e-02 4.92659628e-01
1.47223905e-01 -9.46315587e-01 6.79607213e-01 4.62591022e-01
7.43891716e-01 -5.35137057e-01 5.10606587e-01 5.32072365e-01
-1.07123780e+00 -4.57774609e-01 -4.98869896e-01 -8.48790705e-02
-9.88470092e-02 7.06515372e-01 -4.90142405e-01 -8.78930837e-03
7.32873917e-01 4.67387408e-01 -5.73794484e-01 1.10514581e+00
-2.35401839e-01 3.71946990e-01 -2.22453490e-01 -7.29339272e-02
2.33718723e-01 2.46968921e-02 2.09211528e-01 1.22907293e+00
1.14761546e-01 -1.68584377e-01 1.47094637e-01 8.07280958e-01
-3.88538986e-01 -8.50235894e-02 -4.41266060e-01 -5.35628982e-02
4.90712523e-01 1.41497910e+00 -4.88529831e-01 -4.37330872e-01
-3.72264147e-01 8.97188187e-01 8.80651474e-01 4.79826719e-01
-7.13348567e-01 -5.96970975e-01 6.99885190e-01 -1.97833478e-02
6.83343887e-01 -1.65816620e-01 -4.74825472e-01 -1.33675694e+00
3.48993778e-01 -1.06290817e+00 8.98443833e-02 -4.90435719e-01
-1.21528089e+00 8.02969933e-01 -3.96043718e-01 -1.06368673e+00
7.24963248e-02 -8.06002676e-01 -7.35672534e-01 9.04109895e-01
-1.57569253e+00 -7.90931106e-01 -3.83494973e-01 7.48945117e-01
4.79576766e-01 -2.82312065e-01 6.70739174e-01 5.19828677e-01
-9.48671877e-01 8.43241513e-01 -2.80092172e-02 1.28676578e-01
6.83034480e-01 -8.74047756e-01 3.48355681e-01 6.31375253e-01
-2.21150830e-01 4.41294849e-01 1.08389825e-01 -2.42756844e-01
-1.14573383e+00 -1.25712168e+00 9.65731263e-01 -7.00790510e-02
7.73005664e-01 -7.65480459e-01 -1.10072649e+00 5.60610294e-01
-2.13037491e-01 3.88069332e-01 4.44254428e-01 2.80535817e-01
-6.46638036e-01 -5.54010808e-01 -9.14083838e-01 5.78438044e-01
9.30056036e-01 -5.48997283e-01 -3.75336796e-01 2.99182594e-01
6.82156622e-01 -1.29500046e-01 -7.30510056e-01 2.53036022e-01
6.36729360e-01 -5.24210513e-01 8.55719566e-01 -7.78556228e-01
2.60190308e-01 -2.20615059e-01 1.29525617e-01 -1.34789920e+00
-6.87077165e-01 -5.12178719e-01 -2.97051162e-01 9.08326507e-01
7.67321885e-01 -8.64383638e-01 8.67413819e-01 7.01084554e-01
-1.75055578e-01 -6.93015218e-01 -9.10332203e-01 -1.00607061e+00
2.38251351e-02 -1.25840142e-01 6.25794291e-01 8.27769458e-01
-2.82806158e-01 6.33072674e-01 -3.78837347e-01 -7.54937530e-02
4.33344990e-01 -5.04295491e-02 7.43347883e-01 -1.51305008e+00
-5.54436266e-01 -7.27874041e-01 -2.05170900e-01 -1.14229453e+00
1.71893060e-01 -8.42396975e-01 1.46712095e-01 -1.16073561e+00
1.39365911e-01 -7.57656991e-01 -4.05021012e-01 9.15822744e-01
-1.83641165e-01 3.67442906e-01 5.78112721e-01 4.06109899e-01
-5.63469887e-01 4.98201519e-01 1.52466130e+00 -2.23687917e-01
-2.66705394e-01 1.70371488e-01 -8.22433233e-01 8.25706959e-01
8.27448964e-01 -6.29420936e-01 -3.05276304e-01 -7.81119466e-01
2.72777140e-01 -1.68686584e-01 2.72519261e-01 -9.53956008e-01
2.80819476e-01 -2.48260736e-01 4.43135947e-01 2.60556310e-01
3.07415456e-01 -9.14118230e-01 -1.20725602e-01 5.95062017e-01
-3.45284283e-01 -1.45115629e-01 2.82908231e-01 3.35943550e-01
-2.23970652e-01 -5.05744576e-01 9.44032431e-01 -1.08048439e-01
-5.36502421e-01 4.26385731e-01 -4.86872733e-01 -1.59022301e-01
1.10895646e+00 -3.37945461e-01 -4.14085001e-01 -3.51535082e-02
-9.03057516e-01 2.75135279e-01 1.10676445e-01 4.52604264e-01
3.89287323e-01 -1.32824576e+00 -6.36097372e-01 2.59312034e-01
-2.08478589e-02 3.17269057e-01 1.26938984e-01 7.59669363e-01
-2.15415597e-01 3.99137765e-01 -4.49506730e-01 -4.07753885e-01
-1.05115163e+00 5.28305888e-01 4.65819240e-01 -4.10214990e-01
-4.88705277e-01 7.74780691e-01 6.61141992e-01 -5.20369768e-01
3.64534676e-01 -2.25900933e-01 -1.36727422e-01 8.82147625e-02
4.66620862e-01 2.96951290e-02 1.56038865e-01 -1.13253638e-01
-2.91886896e-01 5.60830057e-01 -5.10221541e-01 3.21589261e-01
1.40530062e+00 2.71692649e-02 -1.94434851e-01 1.37408242e-01
1.02916574e+00 -4.29618537e-01 -1.66418445e+00 -5.18044174e-01
3.83592211e-02 -2.11085528e-01 -1.51263326e-02 -7.00801671e-01
-1.41136754e+00 7.58896708e-01 2.25159720e-01 2.39648148e-01
1.32946825e+00 -6.76099584e-02 5.53502619e-01 7.93633223e-01
1.61535233e-01 -1.06573808e+00 3.96569878e-01 9.29561019e-01
8.62938344e-01 -1.07338703e+00 -3.65101159e-01 -2.15570018e-01
-4.91370648e-01 1.12660897e+00 9.85604167e-01 -3.00435573e-01
5.11334300e-01 3.14500839e-01 -1.64136216e-01 -7.51125254e-03
-9.83570993e-01 -1.02370374e-01 3.63792896e-01 4.95144069e-01
1.66498631e-01 -1.74142092e-01 5.65516017e-02 5.68329155e-01
1.35498151e-01 2.16251403e-01 3.17409188e-01 7.59846747e-01
-5.72163403e-01 -1.08938146e+00 -1.43384844e-01 6.44396961e-01
-2.07937509e-01 -3.39132547e-02 -2.40008444e-01 5.23665428e-01
2.97272712e-01 4.87077862e-01 1.67986885e-01 -2.58198351e-01
4.64277357e-01 -2.15211269e-02 4.38424498e-01 -8.21860552e-01
-7.39625096e-01 -2.51642913e-01 -1.11724451e-01 -3.97286922e-01
-1.87699676e-01 -3.01645726e-01 -8.32423747e-01 -3.13230217e-01
-4.44932789e-01 -1.16777830e-02 6.18658185e-01 1.05361032e+00
5.01398861e-01 8.16342771e-01 5.84378660e-01 -1.11605525e+00
-7.71713436e-01 -9.11149919e-01 -3.14614296e-01 2.63612419e-01
3.31098467e-01 -5.42072177e-01 -3.14231843e-01 5.26966713e-02] | [9.08887004852295, 3.028538703918457] |
92eb48dd-a32a-43ea-9a22-8e634bba03b5 | localised-generative-flows | null | null | https://openreview.net/forum?id=SyegvgHtwr | https://openreview.net/pdf?id=SyegvgHtwr | Localised Generative Flows | We argue that flow-based density models based on continuous bijections are limited in their ability to learn target distributions with complicated topologies, and propose localised generative flows (LGFs) to address this problem. LGFs are composed of stacked continuous mixtures of bijections, which enables each bijection to learn a local region of the target rather than its entirety. Our method is a generalisation of existing flow-based methods, which can be used without modification as the basis for an LGF model. Unlike normalising flows, LGFs do not permit exact computation of log likelihoods, but we propose a simple variational scheme that performs well in practice. We show empirically that LGFs yield improved performance across a variety of common density estimation tasks. | ['Arnaud Doucet', 'George Deligiannidis', 'Anthony Caterini', 'Rob Cornish'] | 2019-09-25 | null | null | null | null | ['normalising-flows'] | ['methodology'] | [-3.57069910e-01 3.29184420e-02 -4.23554510e-01 -2.37193421e-01
-5.91474295e-01 -6.94375277e-01 1.21686625e+00 -4.28728580e-01
-7.97397271e-02 1.13368976e+00 3.82555485e-01 -4.07413125e-01
-3.21969211e-01 -1.01481366e+00 -7.86067009e-01 -6.25007927e-01
-2.53395647e-01 8.95319641e-01 4.46093649e-01 2.27326840e-01
7.50916898e-02 5.47005594e-01 -9.37372744e-01 -2.11361334e-01
1.04139054e+00 4.35261935e-01 -2.15263721e-02 9.56202984e-01
-3.66390228e-01 9.33976650e-01 -5.53048968e-01 -7.00713873e-01
1.08206859e-02 -5.87609529e-01 -7.40241408e-01 8.15207958e-02
5.80000460e-01 -5.71499050e-01 -4.96428847e-01 7.63230383e-01
8.55795108e-03 2.07902595e-01 1.63849294e+00 -1.54448318e+00
-7.40266085e-01 3.15236747e-01 -4.12952989e-01 5.15216410e-01
-4.41595800e-02 9.19564441e-02 9.75234866e-01 -9.25435483e-01
5.67238450e-01 1.66814184e+00 8.03896546e-01 4.06362742e-01
-1.77549684e+00 -3.65887642e-01 2.08798721e-01 -1.32608384e-01
-1.32918906e+00 -5.20782650e-01 4.86680806e-01 -4.30068314e-01
7.71860957e-01 7.65293911e-02 7.90640831e-01 1.18070924e+00
3.49266022e-01 1.01096296e+00 8.49956632e-01 -1.44533247e-01
3.95317912e-01 1.93574518e-01 -3.92108649e-01 7.27853060e-01
2.68569171e-01 -8.14724490e-02 -4.75554764e-01 -5.68278670e-01
1.50044763e+00 -2.17141822e-01 -2.67793030e-01 -9.97011900e-01
-9.70866799e-01 1.42550051e+00 3.63704264e-01 -6.06691986e-02
-1.30288765e-01 5.64478636e-01 1.02474138e-01 -1.22188339e-02
7.95641661e-01 7.14057358e-03 -9.47327316e-02 -4.26793098e-01
-1.13091779e+00 4.86607730e-01 1.37602341e+00 1.07288742e+00
8.88985217e-01 2.26215616e-01 -6.10525757e-02 8.31465662e-01
7.38418579e-01 4.63593751e-01 -1.60057291e-01 -1.51634943e+00
1.61571920e-01 -1.23195127e-02 1.90445349e-01 -6.71580791e-01
1.98732018e-01 -2.47253880e-01 -6.24952853e-01 6.78299442e-02
6.71366751e-01 -2.22685143e-01 -9.73972082e-01 1.86462235e+00
2.61499226e-01 5.68628728e-01 -1.87199175e-01 4.30558205e-01
1.48882270e-01 9.95052576e-01 1.53894827e-01 -9.20406729e-02
7.03636348e-01 -8.17258477e-01 -4.21286583e-01 -1.35301977e-01
1.93842575e-01 -6.21620357e-01 7.93772161e-01 1.86095402e-01
-1.24761879e+00 -1.39458850e-01 -6.06243789e-01 4.05079760e-02
-1.07284747e-01 -4.19127792e-01 9.25723791e-01 7.57694185e-01
-1.62576067e+00 4.62987274e-01 -1.15177858e+00 -4.29161072e-01
7.38683522e-01 -3.35077435e-04 -5.09903356e-02 -8.32810849e-02
-7.60095775e-01 8.77110183e-01 7.07149282e-02 -1.01139933e-01
-1.11925638e+00 -1.12023199e+00 -1.05246198e+00 2.01907024e-01
1.15172334e-01 -1.00338018e+00 1.26585996e+00 -3.82362455e-01
-1.52761614e+00 3.14152479e-01 -5.15856862e-01 -5.01280189e-01
8.49129140e-01 -1.05347916e-01 7.61326104e-02 3.01930338e-01
8.25811997e-02 1.10741317e+00 1.03661239e+00 -1.34467208e+00
-4.72721964e-01 3.04396808e-01 -1.13734484e-01 1.65503606e-01
-9.01648700e-02 -4.15229201e-01 -3.59056622e-01 -6.10466897e-01
-4.18832242e-01 -5.41465759e-01 -5.80035597e-02 3.18152934e-01
-3.83993447e-01 -3.72001380e-01 9.42922771e-01 -2.39888325e-01
1.11959875e+00 -1.71211874e+00 2.17280298e-01 1.52094543e-01
3.47753644e-01 1.05244726e-01 9.26480591e-02 5.57800889e-01
5.79038918e-01 3.52418631e-01 -6.66181028e-01 -6.47350132e-01
2.23541513e-01 6.59185112e-01 -5.22914827e-01 4.92260426e-01
5.56440830e-01 1.07316947e+00 -9.81600940e-01 -7.27133095e-01
3.67355078e-01 8.67660105e-01 -7.64016151e-01 1.59861282e-01
-3.03352892e-01 1.64007753e-01 -2.40259305e-01 1.98598430e-01
8.51999938e-01 -2.47115433e-01 -7.80646130e-02 2.34117731e-01
1.95592701e-01 3.76310229e-01 -1.03153861e+00 1.46551573e+00
-4.31834519e-01 7.12587357e-01 1.83426231e-01 -9.75611746e-01
8.39272678e-01 1.90373465e-01 4.63879198e-01 1.07909277e-01
-2.12357327e-01 8.29200596e-02 -1.67097315e-01 1.58077851e-02
2.30423227e-01 -4.74541217e-01 1.43171877e-01 7.77316868e-01
6.29067242e-01 -4.50316250e-01 4.07122910e-01 7.63810694e-01
9.62680042e-01 4.40046638e-01 3.33600603e-02 -4.89909768e-01
1.80267379e-01 -1.92201227e-01 3.57938647e-01 1.09260678e+00
-3.19784284e-01 7.27460682e-01 8.20527196e-01 -1.07560068e-01
-1.26064146e+00 -2.10933399e+00 -5.00309289e-01 6.28565788e-01
1.96234378e-05 -4.61230963e-01 -5.96780241e-01 -9.94572341e-01
1.70675814e-01 6.91550016e-01 -6.13213897e-01 9.62185636e-02
-3.94148201e-01 -9.77984905e-01 5.94715655e-01 5.69729745e-01
4.27222490e-01 -9.58976924e-01 -2.00239837e-01 2.83065587e-01
-3.32151987e-02 -8.24994504e-01 -6.63175881e-01 -2.77453661e-03
-1.02677751e+00 -1.04768968e+00 -9.60166395e-01 -6.02591753e-01
4.16636616e-01 2.99719553e-02 1.40868926e+00 -3.77586067e-01
-2.43974128e-03 7.21646011e-01 1.32638142e-01 -2.03818560e-01
-6.91469908e-01 3.01555037e-01 -1.62872657e-01 -1.88494846e-01
3.77244651e-01 -8.51232111e-01 -5.18345654e-01 2.97679633e-01
-8.83505464e-01 -2.51626015e-01 2.41590455e-01 7.33112872e-01
2.30949327e-01 -2.44600430e-01 8.61501694e-01 -1.01904988e+00
8.99956882e-01 -8.69556487e-01 -4.41606760e-01 9.92505550e-02
-4.29262400e-01 6.82721958e-02 3.56957346e-01 -5.22995889e-01
-1.24063683e+00 -4.64709371e-01 -9.86353755e-02 -6.77441299e-01
-2.54459649e-01 3.03575456e-01 2.88678437e-01 8.86498168e-02
4.92478728e-01 2.41810709e-01 3.91775042e-01 -4.11544889e-01
7.19777405e-01 2.60084122e-01 5.44476092e-01 -7.26115048e-01
7.91035116e-01 5.21470010e-01 6.37400970e-02 -8.50011408e-01
-4.41212475e-01 -3.00247759e-01 -5.60006618e-01 -2.25045368e-01
6.13947153e-01 -8.67744982e-01 -4.27236378e-01 3.60761791e-01
-9.58959222e-01 -6.10561550e-01 -5.48855066e-01 4.72839415e-01
-9.12138879e-01 2.90096253e-01 -9.70106423e-01 -1.03239262e+00
1.53922707e-01 -8.64592075e-01 1.07628131e+00 1.91807374e-01
-3.03761452e-01 -1.87923133e+00 4.30528402e-01 -3.09335887e-01
7.12555468e-01 -1.62080735e-01 9.17528391e-01 -2.67481327e-01
-7.40902007e-01 2.37390846e-01 -2.40190789e-01 3.93905491e-01
1.77909464e-01 3.85139525e-01 -8.31254244e-01 -2.15392917e-01
7.11933753e-05 -4.11870331e-01 1.12991714e+00 9.35500085e-01
7.40058601e-01 -2.04257488e-01 -5.04215717e-01 6.10737801e-01
1.29600036e+00 -1.52482226e-01 8.80660117e-01 -9.75442678e-02
8.73398840e-01 3.91854554e-01 -1.25278592e-01 2.48437777e-01
6.06482923e-01 2.92810529e-01 1.57383144e-01 4.42736037e-02
-1.94428831e-01 -6.75118446e-01 4.52698469e-01 7.62164593e-01
1.75593466e-01 -4.71792817e-01 -6.35475695e-01 8.09751689e-01
-1.77652955e+00 -9.80143428e-01 -1.30899489e-01 1.94172955e+00
7.63663650e-01 6.88501373e-02 6.17583215e-01 -3.75857383e-01
5.81741571e-01 3.55375201e-01 -5.53780794e-01 -3.66032720e-01
1.48314893e-01 3.93253177e-01 2.99549401e-01 8.93835783e-01
-1.00725567e+00 8.94082665e-01 8.49146366e+00 1.04657161e+00
-5.49676895e-01 1.30391642e-01 6.18635178e-01 -3.53236571e-02
-7.97385633e-01 1.38287604e-01 -7.10286558e-01 4.87604767e-01
9.06885564e-01 -5.35151362e-02 3.53308022e-01 4.95458871e-01
-5.02719246e-02 -3.04472804e-01 -9.73926187e-01 6.00902617e-01
-1.10311836e-01 -1.42049754e+00 2.36200079e-01 3.88092667e-01
7.23296344e-01 1.43230949e-02 1.18853033e-01 3.53557259e-01
9.82958198e-01 -1.19621456e+00 5.63862026e-01 6.48465157e-01
5.96241891e-01 -8.62203956e-01 2.91796058e-01 4.36966747e-01
-1.11432409e+00 3.82505596e-01 -5.95622540e-01 1.27962112e-01
6.33646429e-01 5.49190879e-01 -8.94400239e-01 2.34402001e-01
5.26663542e-01 9.08319831e-01 -1.96605146e-01 1.24348164e+00
-2.09585652e-01 1.08252335e+00 -6.15322351e-01 9.15425569e-02
3.32885921e-01 -5.98677695e-01 6.82527483e-01 1.31638718e+00
3.16974610e-01 -5.54341078e-01 1.04389235e-01 1.54072189e+00
5.40943146e-02 -2.84353793e-01 -8.11958849e-01 -3.57262790e-02
6.20770693e-01 1.01404464e+00 -8.23091626e-01 -2.80432701e-01
-4.00978535e-01 5.92057824e-01 5.82160950e-01 7.63374627e-01
-8.48481894e-01 -1.68372124e-01 7.76312172e-01 2.43772224e-01
7.42350996e-01 -4.29840297e-01 2.44095139e-02 -1.46114802e+00
-3.61495763e-01 -3.06803286e-01 3.63288462e-01 -5.30548811e-01
-1.85701799e+00 2.07927093e-01 5.41519463e-01 -8.13903809e-01
-7.15334773e-01 -4.96492356e-01 -1.03493834e+00 9.70284820e-01
-1.32929683e+00 -1.10485554e+00 1.46395996e-01 5.33110261e-01
4.49661165e-01 7.39297867e-02 7.15508223e-01 1.92363728e-02
-1.65965959e-01 1.60019100e-01 7.98331648e-02 -5.16813360e-02
4.76744831e-01 -1.71004844e+00 6.50321960e-01 7.86523759e-01
6.36510789e-01 5.76214612e-01 5.40538609e-01 -7.88590014e-01
-8.54124963e-01 -9.42460954e-01 5.15423536e-01 -6.56561077e-01
7.50936270e-01 -6.55204177e-01 -1.08290362e+00 1.00328982e+00
2.88499385e-01 8.14849958e-02 4.77959543e-01 2.43615463e-01
-2.51246274e-01 2.98497558e-01 -1.09932804e+00 3.70751649e-01
1.03444862e+00 -5.04053354e-01 -3.36518168e-01 1.14478961e-01
3.66523057e-01 8.37780628e-03 -8.54295373e-01 1.51496246e-01
3.33936840e-01 -1.15502501e+00 1.12572694e+00 -3.14710647e-01
2.88475662e-01 -1.15307681e-01 4.45290580e-02 -1.60214412e+00
-3.36146593e-01 -6.12985551e-01 -7.03866243e-01 1.30713582e+00
2.30859429e-01 -8.29119742e-01 9.14285302e-01 2.30767593e-01
1.05954394e-01 -7.18816459e-01 -9.29829299e-01 -8.66094172e-01
7.57638693e-01 -4.53785390e-01 3.04409325e-01 6.68702900e-01
-1.98142707e-01 3.02443534e-01 -3.63480449e-01 -2.81170487e-01
9.50981855e-01 -3.61082971e-01 7.96544909e-01 -1.33184648e+00
-4.33587700e-01 -6.79597676e-01 -3.61788839e-01 -1.51017940e+00
3.79267931e-01 -8.31819177e-01 6.86171511e-03 -1.73869324e+00
2.48500064e-01 -7.05705166e-01 1.70040563e-01 -4.95734140e-02
-8.58949497e-02 3.62575740e-01 6.45307153e-02 2.66081840e-01
-3.96432817e-01 8.85208249e-01 1.18451047e+00 3.36939730e-02
-6.81212693e-02 8.25516284e-02 -5.70497870e-01 6.13117933e-01
5.28945565e-01 -3.42710853e-01 -1.04877448e+00 -2.66561955e-01
-1.00058965e-01 -9.89653394e-02 5.27911961e-01 -7.01718748e-01
4.55956385e-02 -1.74614877e-01 6.10103130e-01 -4.03084099e-01
3.43021601e-01 -2.46745765e-01 2.55160213e-01 1.20819129e-01
-1.39420897e-01 -2.74056971e-01 6.88346326e-02 8.38742793e-01
-1.35491654e-01 -2.76654005e-01 6.75086141e-01 -1.42776296e-01
-2.41650209e-01 4.88611817e-01 -7.03749597e-01 4.18807745e-01
8.47879589e-01 -9.25428793e-02 -3.26006800e-01 -9.30625319e-01
-7.24581778e-01 7.65096992e-02 5.35333097e-01 5.91197461e-02
5.10000348e-01 -1.39645600e+00 -7.00979590e-01 1.82867706e-01
-4.53371167e-01 4.00208116e-01 -1.67266712e-01 7.13984549e-01
-6.48076236e-01 2.43115008e-01 7.43775070e-02 -8.75067651e-01
-3.51523995e-01 3.25313330e-01 5.85273981e-01 -3.02282810e-01
-7.00080991e-01 9.18586135e-01 6.44928992e-01 -5.40084004e-01
2.60194466e-02 -7.72917569e-02 7.85470232e-02 -1.06063090e-01
4.02388245e-01 4.78738874e-01 -5.42505383e-01 -4.93685246e-01
-1.02583177e-01 2.68165946e-01 -8.16339478e-02 -5.69702446e-01
1.23068035e+00 -3.32573384e-01 4.66720611e-02 7.00174689e-01
1.22754359e+00 -1.85628071e-01 -1.92587173e+00 -8.49985927e-02
-3.53679687e-01 -7.23248422e-01 -2.35445816e-02 -4.08812404e-01
-1.02412319e+00 9.40815449e-01 -5.44737615e-02 4.92018729e-01
5.90555906e-01 3.63574326e-01 5.31761706e-01 -1.73879057e-01
3.19067955e-01 -3.97268027e-01 1.66653588e-01 5.31467676e-01
5.53751707e-01 -9.24315214e-01 -1.59466043e-01 -5.80251992e-01
-5.42561412e-01 9.78100002e-01 4.52874750e-01 -4.97582078e-01
1.02115870e+00 5.03228247e-01 -2.06570148e-01 -2.09370390e-01
-8.17107677e-01 -1.02902040e-01 3.59766722e-01 1.08892906e+00
2.42295966e-01 -2.02997059e-01 1.52388960e-01 -1.55169249e-01
-2.12597817e-01 -3.84306088e-02 4.87066716e-01 9.10680354e-01
-4.46421057e-01 -1.05015504e+00 -1.60651699e-01 8.03282440e-01
-2.83682764e-01 1.89868845e-02 -4.50536385e-02 9.55664277e-01
-2.32084662e-01 7.28635848e-01 7.39955723e-01 1.91286594e-01
-2.74517715e-01 2.96996176e-01 8.42314243e-01 -4.52336639e-01
1.64289564e-01 4.09976184e-01 -1.22448221e-01 -3.62569332e-01
-5.76693296e-01 -8.08966577e-01 -6.94779634e-01 -7.34907627e-01
-1.83475524e-01 2.34214738e-01 1.88070342e-01 9.62398350e-01
8.48442316e-02 2.08269656e-01 3.37564588e-01 -9.40468550e-01
-4.55406398e-01 -1.10316336e+00 -9.06593144e-01 2.70055592e-01
4.46314991e-01 -1.00206196e+00 -6.53150320e-01 8.31122845e-02] | [7.071974754333496, 3.904134511947632] |
0ffd7819-a257-4463-a08e-6913d447abd9 | a-modulation-domain-loss-for-neural-network-1 | 2102.07330 | null | https://arxiv.org/abs/2102.07330v1 | https://arxiv.org/pdf/2102.07330v1.pdf | A Modulation-Domain Loss for Neural-Network-based Real-time Speech Enhancement | We describe a modulation-domain loss function for deep-learning-based speech enhancement systems. Learnable spectro-temporal receptive fields (STRFs) were adapted to optimize for a speaker identification task. The learned STRFs were then used to calculate a weighted mean-squared error (MSE) in the modulation domain for training a speech enhancement system. Experiments showed that adding the modulation-domain MSE to the MSE in the spectro-temporal domain substantially improved the objective prediction of speech quality and intelligibility for real-time speech enhancement systems without incurring additional computation during inference. | [] | 2021-02-15 | a-modulation-domain-loss-for-neural-network | https://arxiv.org/pdf/2102.07330.pdf | https://arxiv.org/pdf/2102.07330.pdf | null | ['speaker-identification', 'speech-denoising'] | ['speech', 'speech'] | [ 4.49688196e-01 3.95537689e-02 3.15529346e-01 -9.30178821e-01
-1.14095640e+00 4.95285401e-03 3.55035216e-01 -1.66019663e-01
-5.20519376e-01 6.08419597e-01 3.99261415e-01 -3.45647782e-01
-2.34560445e-01 -4.30944115e-01 -4.91265148e-01 -7.84083247e-01
-4.40895140e-01 -5.32773495e-01 -2.38494836e-02 -3.53984565e-01
4.46901545e-02 4.86682862e-01 -1.54551506e+00 5.12039185e-01
8.08747113e-01 1.28457594e+00 7.55616486e-01 1.04904914e+00
4.32089031e-01 6.58156931e-01 -1.08846951e+00 -9.91083682e-02
7.19053224e-02 -4.07158643e-01 -4.38057959e-01 1.47160918e-01
3.33730072e-01 -3.54422808e-01 -9.51349616e-01 1.18467689e+00
1.01202607e+00 9.65995252e-01 3.85928631e-01 -6.33411407e-01
-6.90471828e-01 4.13543791e-01 1.64095968e-01 6.15876555e-01
-2.04898603e-02 4.99554276e-02 7.84164190e-01 -1.05919540e+00
1.29684702e-01 1.47029924e+00 6.40618861e-01 7.17780590e-01
-9.78727221e-01 -3.99720699e-01 5.32453228e-03 3.44130188e-01
-1.09800994e+00 -1.02380705e+00 4.89630342e-01 1.48006544e-01
1.54333806e+00 4.64813292e-01 1.89111009e-01 7.91161776e-01
3.99524391e-01 5.97940385e-01 1.01679575e+00 -4.41777855e-01
3.44025940e-01 6.78139031e-02 -3.22878212e-01 5.70231795e-01
-5.81862211e-01 8.83332610e-01 -7.17730224e-01 3.30554307e-01
6.63032651e-01 -7.48875022e-01 -3.23736757e-01 6.60700500e-01
-5.07898569e-01 5.58228731e-01 5.79242647e-01 3.93443227e-01
-4.05585855e-01 3.86680961e-02 3.44247490e-01 6.64517105e-01
7.53093660e-01 4.36267734e-01 -4.98648554e-01 -3.46736252e-01
-1.10756981e+00 -1.99664429e-01 4.04023290e-01 3.67993385e-01
3.76852781e-01 8.30584586e-01 -5.08580804e-01 1.42431724e+00
3.62400919e-01 6.88901126e-01 5.04426003e-01 -1.23373818e+00
4.06386942e-01 -3.66465598e-01 -1.44633397e-01 -6.81011975e-01
-2.56922603e-01 -7.64672339e-01 -7.07234621e-01 4.56010520e-01
-1.03482015e-01 -3.63227844e-01 -1.09102321e+00 1.80795026e+00
7.06207901e-02 9.27447751e-02 2.40958333e-01 1.06904471e+00
5.53989410e-01 1.08415198e+00 2.00551134e-02 -4.68048424e-01
9.87072110e-01 -9.71527874e-01 -1.14056849e+00 -3.54563296e-01
-3.03463694e-02 -8.78405094e-01 9.26511943e-01 2.67618477e-01
-1.52318239e+00 -8.61843228e-01 -1.33725381e+00 2.56292615e-02
-1.82308733e-01 -2.95768641e-02 2.32381165e-01 7.81270742e-01
-1.54634762e+00 1.08619666e+00 -7.67412841e-01 2.73558050e-01
1.51298717e-01 6.21580005e-01 -6.00737110e-02 5.30371726e-01
-1.54924893e+00 1.00122273e+00 7.77788162e-02 -5.35230041e-02
-1.20859301e+00 -6.56292021e-01 -9.03337181e-01 3.61133486e-01
-7.51276836e-02 -3.75194669e-01 1.57274508e+00 -8.82998943e-01
-2.19567466e+00 4.22459364e-01 -3.42095882e-01 -6.92660391e-01
-1.84339762e-01 8.32616538e-02 -1.07205188e+00 9.56437513e-02
-5.25611937e-01 6.29555166e-01 1.57053471e+00 -8.83124530e-01
-5.99108398e-01 -9.00272354e-02 -5.89754283e-02 4.29021716e-01
-6.31986141e-01 1.64824799e-01 -2.82445848e-02 -1.18171990e+00
9.79658589e-02 -4.44391310e-01 -8.84562135e-02 -6.30834922e-02
4.29588407e-02 -1.14875827e-02 1.02908957e+00 -1.38669193e+00
1.40922725e+00 -2.37777853e+00 -3.54143995e-04 2.58937255e-02
-1.00070655e-01 7.69251108e-01 -5.56285262e-01 -6.95555955e-02
1.59955230e-02 -1.74198151e-01 -3.18046480e-01 -5.41747987e-01
8.67848173e-02 2.44223773e-02 -2.69140899e-01 1.81695238e-01
4.45527673e-01 4.13933545e-01 -7.63900757e-01 -1.69473011e-02
5.04411340e-01 1.14596355e+00 -6.00086570e-01 5.76313496e-01
3.74387428e-02 2.63472378e-01 2.25030407e-01 1.51331276e-01
5.72378755e-01 2.86447018e-01 -1.72276810e-01 -5.83440289e-02
5.55262435e-03 1.04338920e+00 -7.87298262e-01 1.35605884e+00
-1.02215922e+00 1.23226416e+00 6.23917818e-01 -8.19833934e-01
1.05628335e+00 6.59193814e-01 1.96350619e-01 -1.06206441e+00
2.00208556e-02 1.33732960e-01 4.18531820e-02 -2.22247750e-01
4.31848019e-01 -4.64603692e-01 4.94453341e-01 2.03886613e-01
4.52130288e-01 -3.54068488e-01 -2.88155943e-01 -2.64540225e-01
9.62717891e-01 -4.71679211e-01 -1.93175375e-01 -1.79056674e-01
7.22715497e-01 -8.55918407e-01 4.11152333e-01 5.65276861e-01
-4.23279703e-01 4.04127419e-01 -2.50517815e-01 -1.02534024e-02
-1.15407693e+00 -1.42382622e+00 -3.96527171e-01 1.45891070e+00
-3.77528876e-01 -9.98101830e-02 -1.08714867e+00 -3.44482094e-01
-3.56175542e-01 8.39342415e-01 -9.13697854e-02 -5.50912380e-01
-5.27787328e-01 -3.35344404e-01 7.63498783e-01 5.59443116e-01
4.51217324e-01 -1.03333151e+00 -1.88323513e-01 4.34380621e-01
-1.72320247e-01 -9.73880231e-01 -1.17789173e+00 5.44516802e-01
-8.05186868e-01 -1.35989413e-01 -8.04874539e-01 -1.04510736e+00
4.02421445e-01 -7.91102648e-02 7.27554381e-01 -2.24948734e-01
-1.63977340e-01 6.29480034e-02 6.54246360e-02 -3.00240189e-01
-7.84492970e-01 -4.74972904e-01 3.93302947e-01 -1.52760386e-01
3.60971428e-02 -5.47077119e-01 -4.37886417e-01 2.84301698e-01
-6.56076312e-01 -4.58688498e-01 2.51954138e-01 1.20042574e+00
4.93761152e-01 4.50749785e-01 9.76482153e-01 1.25399008e-01
9.32541788e-01 2.50576943e-01 -7.19785571e-01 6.24194629e-02
-4.44434911e-01 2.93955624e-01 4.23749566e-01 -6.71779275e-01
-1.48495162e+00 -4.11525100e-01 -7.57310688e-01 -2.97645569e-01
1.19333200e-01 5.62174618e-01 -1.44683808e-01 -1.93800628e-01
7.44120598e-01 2.70546585e-01 -1.57371797e-02 -4.48178262e-01
1.09706447e-02 1.19624293e+00 7.30481744e-01 -1.80973724e-01
3.66030037e-01 -1.41278952e-01 -2.02219710e-01 -1.14699960e+00
-6.83178008e-01 -2.63330817e-01 -5.57579771e-02 -1.27586186e-01
6.43141627e-01 -8.71190608e-01 -4.97788817e-01 5.57164729e-01
-1.12817037e+00 -5.86001873e-01 -5.07307649e-01 8.05389524e-01
-8.22003543e-01 2.22730473e-01 -8.93666983e-01 -1.21557701e+00
-6.47464156e-01 -1.04971170e+00 9.28703427e-01 2.13877708e-01
1.11802652e-01 -1.14272988e+00 -2.38833338e-01 1.83228374e-01
9.78376806e-01 -5.94533980e-01 8.33854437e-01 -1.09169811e-01
-5.74991368e-02 4.89727817e-02 -9.30640996e-02 1.20252538e+00
3.14331889e-01 -4.10000235e-01 -1.67437267e+00 -4.57483888e-01
5.24694681e-01 7.81750306e-03 7.72259116e-01 9.19569731e-01
1.45766592e+00 -5.41178465e-01 3.16423446e-01 8.68567765e-01
8.55569363e-01 7.28457153e-01 7.02406228e-01 -1.01545252e-01
4.17056121e-02 3.75618517e-01 3.36264312e-01 4.98945951e-01
-3.16557854e-01 9.51839030e-01 -1.66271940e-01 -2.00035185e-01
-7.17009842e-01 -7.02297613e-02 8.53581727e-01 1.17534411e+00
1.33988485e-01 -2.54658937e-01 -4.66683298e-01 3.73488545e-01
-1.16891050e+00 -1.17545998e+00 8.05893481e-01 2.20477533e+00
1.22755384e+00 1.71997368e-01 -6.90750778e-02 3.39997500e-01
9.19262409e-01 2.82388449e-01 -6.21161997e-01 -8.92858148e-01
-2.35093102e-01 6.81211054e-01 2.57956713e-01 1.09747565e+00
-1.16763985e+00 6.78686976e-01 7.63941383e+00 1.07312965e+00
-1.34228361e+00 1.85009018e-01 6.90059185e-01 -1.13201059e-01
-1.50185943e-01 -6.25500619e-01 -5.02989531e-01 1.40594617e-01
1.72662592e+00 -3.37394893e-01 1.09798729e+00 6.11186266e-01
5.23911715e-01 4.40452248e-01 -7.46167421e-01 8.69532466e-01
-1.47245988e-01 -9.22988892e-01 -4.06705827e-01 -8.42845961e-02
7.17432678e-01 8.32198635e-02 6.85334027e-01 3.89272183e-01
-4.43195812e-02 -1.18379772e+00 5.62949479e-01 3.99122834e-01
1.10134101e+00 -7.65529871e-01 4.23279732e-01 1.48422346e-01
-9.85181987e-01 -2.81064153e-01 -6.22396052e-01 8.15443993e-02
2.46267319e-01 6.93106592e-01 -1.17567301e+00 -1.54928774e-01
5.25585234e-01 2.10332379e-01 1.45182252e-01 8.71810973e-01
-1.65230632e-01 6.99754298e-01 -2.72910058e-01 -7.32555389e-02
2.19254643e-01 2.09292099e-01 9.63302016e-01 1.51995611e+00
4.96072948e-01 1.10363461e-01 -4.62574333e-01 7.33799219e-01
-2.23071605e-01 -4.38402414e-01 -7.57019967e-02 -2.91097343e-01
5.48376441e-01 8.18339586e-01 1.54709760e-02 -1.51743039e-01
7.22253509e-03 9.68814194e-01 -1.26854680e-03 6.03096545e-01
-7.24937975e-01 -9.04323339e-01 1.09487009e+00 2.43724268e-02
5.44062316e-01 -1.32583931e-01 -3.41022432e-01 -6.79458499e-01
-1.18004434e-01 -9.05081451e-01 -6.27427548e-02 -8.30334723e-01
-1.18036854e+00 1.06529844e+00 -2.72753894e-01 -9.49695766e-01
-3.85742992e-01 -8.25401545e-01 -3.63548368e-01 1.32147348e+00
-1.41812134e+00 -4.90262687e-01 2.42587104e-01 5.35135329e-01
7.55858243e-01 -5.21272302e-01 1.05421531e+00 5.25652707e-01
-3.25069368e-01 1.07577372e+00 2.81743169e-01 -7.52961487e-02
4.42280829e-01 -1.31063688e+00 7.18236566e-01 8.28819215e-01
-1.86141506e-02 4.02105927e-01 7.96950161e-01 -2.12704435e-01
-8.33947420e-01 -1.16193414e+00 1.10098779e+00 1.08509533e-01
2.77169198e-01 -2.68118709e-01 -9.50473428e-01 9.89264101e-02
3.18197697e-01 1.16302259e-01 3.87260497e-01 6.84270263e-02
-3.20606738e-01 -3.86816263e-01 -1.44310367e+00 4.72661406e-01
9.80352759e-01 -1.17494833e+00 -5.24532437e-01 2.32202306e-01
9.80819702e-01 -6.01821005e-01 -1.03335869e+00 4.05128449e-01
2.62538642e-01 -5.17700315e-01 1.14744210e+00 -6.29957080e-01
1.32113159e-01 -2.80601811e-02 -4.94632393e-01 -1.89850605e+00
-3.67000103e-01 -7.86524415e-01 -4.16735083e-01 8.61712515e-01
6.32479489e-01 -3.72742027e-01 3.03398460e-01 3.12180847e-01
-7.93424368e-01 -5.09677351e-01 -1.37390339e+00 -1.07804835e+00
-5.52807143e-03 -6.00021482e-01 4.86018002e-01 3.03976446e-01
2.68678457e-01 8.52821320e-02 -2.97994137e-01 4.81826782e-01
4.14667040e-01 -7.18970537e-01 -2.71077812e-01 -4.86998379e-01
-7.40795732e-01 -4.89200354e-01 -5.06238699e-01 -1.16186309e+00
9.63283330e-02 -6.15702689e-01 7.23900616e-01 -1.29424167e+00
-4.40106243e-01 -2.86584459e-02 -8.44827354e-01 2.46016145e-01
-3.03580701e-01 -1.29541323e-01 1.49549022e-01 -4.41125602e-01
-1.73656926e-01 8.51489961e-01 1.14359069e+00 -4.91632551e-01
-3.57891798e-01 3.50945055e-01 -2.72644490e-01 4.08564329e-01
7.18274593e-01 -3.65552694e-01 -5.18535972e-01 -5.13439119e-01
-5.99112928e-01 3.25078726e-01 7.21878558e-02 -1.41181314e+00
2.47222379e-01 -1.23490607e-02 3.02831352e-01 -2.94565380e-01
9.53293622e-01 -4.39254582e-01 -3.25762779e-01 3.38893652e-01
-8.08738589e-01 -3.61170560e-01 6.12122357e-01 3.39270502e-01
-5.75234652e-01 -1.22386858e-01 1.26381934e+00 3.66903275e-01
-5.55879831e-01 -2.18960131e-03 -6.43596172e-01 -1.88648328e-01
4.83693451e-01 9.91368517e-02 -6.55537993e-02 -8.29820573e-01
-1.07339013e+00 -4.96699959e-01 -3.22950721e-01 4.10524040e-01
1.10304224e+00 -1.25556314e+00 -8.73252630e-01 6.99945331e-01
-5.00467360e-01 -7.26284385e-01 4.67826813e-01 3.41553211e-01
-6.27277493e-02 5.49056649e-01 -1.62774235e-01 -4.02911693e-01
-1.24995983e+00 2.47190073e-01 9.78469431e-01 -1.03017122e-01
-2.23202735e-01 1.47636759e+00 4.92841639e-02 -4.31615919e-01
7.44724214e-01 -3.48244488e-01 1.62418395e-01 -6.21921182e-01
1.03639674e+00 3.67141068e-01 5.68795443e-01 -5.12793005e-01
-3.08809102e-01 -6.40779808e-02 1.68210179e-01 -8.41922581e-01
1.26952910e+00 -2.36836150e-01 4.29989398e-01 1.37107879e-01
1.42522657e+00 -3.37971807e-01 -1.56781220e+00 -3.47131342e-01
-2.77948290e-01 -4.28421795e-01 1.02856040e+00 -1.34023941e+00
-1.09817219e+00 1.20036972e+00 1.18776882e+00 1.06953688e-01
1.70738840e+00 -4.66188073e-01 9.51882899e-01 3.39569479e-01
1.18717805e-01 -1.55655205e+00 1.63140073e-01 8.93280447e-01
1.21096778e+00 -1.02781749e+00 -4.16362822e-01 -1.76027402e-01
-6.48882627e-01 1.11280870e+00 3.49179506e-01 2.94430822e-01
1.03892374e+00 7.68616498e-01 1.57871693e-01 4.22531843e-01
-8.42293978e-01 -5.01229465e-02 6.72748089e-01 1.09375989e+00
5.94616115e-01 8.77440795e-02 1.51556343e-01 4.62779760e-01
-3.66974354e-01 -2.95334518e-01 -2.00014383e-01 5.39921463e-01
-8.34648907e-01 -8.18183959e-01 -4.02105391e-01 2.97641158e-01
-5.16915381e-01 -4.34373558e-01 1.75750196e-01 -1.62362158e-01
1.32634873e-02 1.54270291e+00 3.16314697e-01 -6.86495781e-01
3.58432204e-01 -4.52219211e-02 5.36600292e-01 -4.53715533e-01
-5.77494681e-01 2.97282577e-01 1.83704183e-01 -2.03403175e-01
-6.29538968e-02 -4.29467708e-01 -1.22900426e+00 -2.96999037e-01
-4.88241017e-01 7.98367187e-02 1.05180836e+00 8.04838598e-01
2.56061345e-01 9.83838141e-01 1.14814174e+00 -6.74566031e-01
-1.14055061e+00 -1.18838143e+00 -6.28506601e-01 1.71568722e-01
9.75245535e-01 -2.56097496e-01 -5.67939341e-01 1.58515871e-01] | [14.986944198608398, 5.932658672332764] |
9c25e410-7f82-4b50-b674-542fbdb75fbf | ru-net-regularized-unrolling-network-for | 2205.01297 | null | https://arxiv.org/abs/2205.01297v1 | https://arxiv.org/pdf/2205.01297v1.pdf | RU-Net: Regularized Unrolling Network for Scene Graph Generation | Scene graph generation (SGG) aims to detect objects and predict the relationships between each pair of objects. Existing SGG methods usually suffer from several issues, including 1) ambiguous object representations, as graph neural network-based message passing (GMP) modules are typically sensitive to spurious inter-node correlations, and 2) low diversity in relationship predictions due to severe class imbalance and a large number of missing annotations. To address both problems, in this paper, we propose a regularized unrolling network (RU-Net). We first study the relation between GMP and graph Laplacian denoising (GLD) from the perspective of the unrolling technique, determining that GMP can be formulated as a solver for GLD. Based on this observation, we propose an unrolled message passing module and introduce an $\ell_p$-based graph regularization to suppress spurious connections between nodes. Second, we propose a group diversity enhancement module that promotes the prediction diversity of relationships via rank maximization. Systematic experiments demonstrate that RU-Net is effective under a variety of settings and metrics. Furthermore, RU-Net achieves new state-of-the-arts on three popular databases: VG, VRD, and OI. Code is available at https://github.com/siml3/RU-Net. | ['DaCheng Tao', 'Yibing Zhan', 'Jing Zhang', 'Changxing Ding', 'Xin Lin'] | 2022-05-03 | null | http://openaccess.thecvf.com//content/CVPR2022/html/Lin_RU-Net_Regularized_Unrolling_Network_for_Scene_Graph_Generation_CVPR_2022_paper.html | http://openaccess.thecvf.com//content/CVPR2022/papers/Lin_RU-Net_Regularized_Unrolling_Network_for_Scene_Graph_Generation_CVPR_2022_paper.pdf | cvpr-2022-1 | ['scene-graph-generation'] | ['computer-vision'] | [ 2.87067235e-01 1.10837400e-01 -1.87549844e-01 -3.91448736e-01
-3.25311452e-01 -7.92082027e-02 3.93231362e-01 1.85932219e-01
1.47084132e-01 4.99904245e-01 1.07335486e-01 -7.93836825e-03
-3.26786757e-01 -9.44907248e-01 -7.10734010e-01 -5.35975993e-01
-1.43796146e-01 2.23483101e-01 1.91980138e-01 -1.37758121e-01
1.20448627e-01 4.37536240e-01 -1.28605151e+00 8.87030438e-02
1.23119271e+00 9.54271376e-01 2.19603419e-01 2.92801231e-01
-8.79765153e-02 1.00075698e+00 -3.02425295e-01 -4.38913643e-01
3.39842260e-01 -4.98521656e-01 -5.61798453e-01 3.74198645e-01
6.54480517e-01 2.34611835e-02 -8.04400325e-01 1.21105886e+00
4.96570885e-01 1.96950480e-01 6.17673874e-01 -1.66762733e+00
-7.72286594e-01 5.88707626e-01 -1.05706882e+00 1.14918850e-01
9.93178338e-02 9.30620581e-02 1.25871181e+00 -1.04841459e+00
6.38928771e-01 1.45770192e+00 7.04099953e-01 1.79618612e-01
-1.52349234e+00 -7.17577219e-01 4.36035573e-01 3.14751983e-01
-1.70682228e+00 -1.40064076e-01 9.85732377e-01 -6.04718924e-01
5.73424637e-01 3.06640327e-01 4.97339457e-01 8.17813814e-01
2.25269552e-02 6.37561142e-01 6.25465393e-01 -6.04606532e-02
1.28268242e-01 -1.39127627e-01 3.39092374e-01 8.99021089e-01
5.66711664e-01 -2.94695318e-01 -7.88129628e-01 -1.55676901e-01
8.22899103e-01 -7.54639208e-02 -4.67349261e-01 -5.22556484e-01
-1.06945348e+00 7.72036135e-01 8.06426823e-01 -7.70938396e-02
-3.55009615e-01 1.89489797e-01 1.72520116e-01 2.44033977e-01
5.96783221e-01 4.63180214e-01 -1.45984784e-01 5.99352539e-01
-5.58328867e-01 1.14687078e-01 7.53103971e-01 1.05387795e+00
8.45525146e-01 -2.13708282e-02 -3.46249729e-01 1.07940090e+00
5.27194917e-01 1.18867636e-01 -1.38153974e-02 -7.93592513e-01
6.86398983e-01 1.03624856e+00 -4.89647359e-01 -1.79083991e+00
-5.07286966e-01 -7.95757115e-01 -1.35267496e+00 -1.40576720e-01
7.22534582e-02 6.65619597e-02 -8.59216630e-01 1.66687632e+00
4.58247215e-01 4.30642813e-01 -2.22413480e-01 1.00530589e+00
1.31710768e+00 6.28305614e-01 -6.26217574e-02 -2.07376257e-01
1.10024154e+00 -1.01205730e+00 -5.99384367e-01 -3.65623415e-01
5.56366622e-01 -4.82106507e-01 9.29425359e-01 1.58428892e-01
-8.19900990e-01 -4.03155237e-01 -9.21698868e-01 -5.19553050e-02
-7.11539686e-02 1.74351856e-01 6.84465110e-01 1.44851908e-01
-9.25047696e-01 5.53091347e-01 -6.14358902e-01 -2.53105640e-01
6.62076712e-01 3.59601229e-01 -2.71152586e-01 -3.76179129e-01
-8.76759112e-01 3.76469702e-01 2.73524314e-01 4.74176109e-01
-5.91239512e-01 -5.93197286e-01 -8.91704381e-01 4.68823239e-02
7.24538922e-01 -8.25373530e-01 4.48547602e-01 -7.63428748e-01
-8.64186466e-01 7.66347587e-01 -1.07894175e-01 -1.92582607e-01
4.28487867e-01 -5.00385696e-03 -3.12375069e-01 9.86815616e-02
2.28232220e-01 6.07679367e-01 7.67420292e-01 -1.50468183e+00
-3.15225124e-01 -5.86943567e-01 -5.85872531e-02 3.46849561e-01
-3.97263944e-01 -2.77024239e-01 -8.46624970e-01 -9.27546561e-01
7.03276813e-01 -8.79975915e-01 -4.62709188e-01 1.27055466e-01
-8.64628494e-01 -1.36269093e-01 5.46259940e-01 -5.92416346e-01
1.37133300e+00 -2.16282725e+00 2.43264407e-01 4.77023780e-01
7.08425224e-01 9.67185348e-02 -4.03603256e-01 3.17273736e-01
-1.49789050e-01 3.95190865e-02 -4.51917112e-01 -3.32207590e-01
-1.85644776e-01 3.72760028e-01 3.72540355e-02 4.96534765e-01
2.71855175e-01 8.47728550e-01 -9.01852310e-01 -4.35800076e-01
-4.01824974e-02 5.50778270e-01 -6.06743455e-01 8.90597999e-02
-1.88117385e-01 6.73374116e-01 -4.95011955e-01 7.29715943e-01
9.05168533e-01 -6.96748972e-01 2.92175412e-01 -5.50654650e-01
1.84684798e-01 7.85103738e-02 -1.50917339e+00 1.47792459e+00
1.09297775e-01 4.70074803e-01 4.77096140e-02 -1.06750047e+00
1.08849132e+00 -1.78312451e-01 4.37147528e-01 -4.83786106e-01
1.19673856e-01 9.25230607e-02 -1.12752356e-01 -4.05268371e-01
3.96207452e-01 2.85514295e-01 3.42099369e-01 3.83033138e-03
-2.91260495e-03 2.42052332e-01 3.24627489e-01 5.92346072e-01
1.33964992e+00 -1.93911746e-01 3.05574089e-02 -2.13115364e-01
4.12223905e-01 -1.87721878e-01 9.70434964e-01 8.46206188e-01
5.43897040e-02 7.73207366e-01 8.81018341e-01 -1.26150653e-01
-4.86865610e-01 -9.22627449e-01 8.42421055e-02 9.66650426e-01
6.01066291e-01 -6.14872515e-01 -2.62317955e-01 -6.18887842e-01
1.32722810e-01 5.73295474e-01 -5.28258801e-01 -3.41134548e-01
-4.28063214e-01 -1.15419602e+00 1.86645091e-01 3.53567183e-01
5.45258939e-01 -7.64361620e-01 2.85866648e-01 9.21851769e-02
-2.89815813e-01 -1.19264102e+00 -5.11274517e-01 -6.40075952e-02
-8.73942733e-01 -1.21259522e+00 -3.72948945e-01 -8.93177629e-01
1.15758729e+00 6.34073377e-01 1.24747360e+00 4.03215766e-01
-2.10917965e-01 2.33616143e-01 -3.16349834e-01 -5.85115282e-03
-1.88528225e-01 1.59468159e-01 -1.63405865e-01 3.56616646e-01
6.16044514e-02 -8.63495350e-01 -6.83638752e-01 4.74629611e-01
-8.43563855e-01 4.14854228e-01 5.11456072e-01 6.62358522e-01
8.67408276e-01 1.17776677e-01 5.24935722e-01 -1.05508661e+00
5.25244892e-01 -6.51873767e-01 -6.24010861e-01 2.96931654e-01
-6.92196846e-01 -6.60152063e-02 3.48541886e-01 -3.42224956e-01
-7.74544001e-01 1.38506249e-01 9.67816263e-02 -5.58926284e-01
2.40110710e-01 6.64133012e-01 -3.14967364e-01 -1.79175735e-01
5.24175644e-01 2.84240954e-02 -8.58558714e-02 -4.66066092e-01
3.32376480e-01 1.85456097e-01 4.61977363e-01 -4.14818645e-01
1.00183904e+00 4.52132523e-01 3.00692379e-01 -8.39813352e-01
-9.71567869e-01 -4.59903985e-01 -4.37257111e-01 -3.17115277e-01
6.18133903e-01 -1.14475381e+00 -6.06512487e-01 5.11785984e-01
-9.70505655e-01 -3.09145719e-01 -7.07854182e-02 3.63067299e-01
-1.14442088e-01 4.32776928e-01 -6.52081788e-01 -4.57696617e-01
-2.86230147e-01 -8.84220064e-01 7.26696551e-01 2.61054188e-01
7.36282989e-02 -8.45276296e-01 -2.20409691e-01 4.65561152e-01
5.20255044e-02 4.19989228e-01 9.01417315e-01 -4.62560117e-01
-8.50847602e-01 -1.10889718e-01 -6.69885457e-01 3.49338621e-01
8.92436653e-02 -1.62572712e-01 -6.23111546e-01 -3.28484803e-01
-2.22031251e-01 -9.30797532e-02 1.12606764e+00 4.04879153e-01
1.27902997e+00 -2.79509842e-01 -3.18282396e-01 9.02241409e-01
1.37187421e+00 -2.23279193e-01 5.56756735e-01 3.42473611e-02
1.44366455e+00 5.88804007e-01 4.97652084e-01 5.76715887e-01
6.93526089e-01 5.45655787e-01 6.72281384e-01 -1.52214512e-01
-3.10133874e-01 -3.40079278e-01 4.44855094e-02 1.05357492e+00
3.20417397e-02 -4.26604301e-01 -9.40574884e-01 4.10319775e-01
-2.25141788e+00 -4.01995271e-01 -6.76985085e-01 2.05823445e+00
5.52024603e-01 7.05310181e-02 -6.82216957e-02 -1.95998639e-01
1.04500496e+00 2.99785614e-01 -6.10272765e-01 3.30980510e-01
-3.68037850e-01 -3.38236958e-01 5.05409658e-01 4.41996276e-01
-9.01449263e-01 8.81509125e-01 4.72912693e+00 8.22776139e-01
-8.68306458e-01 -5.70048904e-03 8.40805948e-01 -2.72561964e-02
-4.04842794e-01 4.23009209e-02 -7.79932261e-01 4.00838733e-01
7.16825351e-02 -8.91886875e-02 4.43397850e-01 5.54760933e-01
1.92535028e-01 1.10691022e-02 -9.70520496e-01 1.09904313e+00
2.88979262e-01 -1.26227880e+00 2.26257414e-01 -5.63639728e-03
7.93549597e-01 4.38398831e-02 -1.69914484e-01 1.25056356e-01
1.51819065e-01 -8.36745679e-01 4.59857136e-01 3.53640735e-01
4.13848519e-01 -5.60071528e-01 5.07508576e-01 1.21841885e-01
-1.43966198e+00 8.40445384e-02 -5.05313158e-01 7.00806230e-02
6.36757389e-02 1.07189119e+00 -6.50673926e-01 8.86695206e-01
6.22669876e-01 1.00499344e+00 -8.35118651e-01 1.12147391e+00
-3.71762902e-01 7.27959573e-01 -3.20089430e-01 2.57620960e-01
-5.46894558e-02 -3.93522769e-01 8.77967358e-01 9.37380552e-01
1.05014317e-01 1.63999721e-01 3.54851961e-01 1.05395770e+00
-3.68129969e-01 1.10498980e-01 -4.27644819e-01 7.88039342e-02
4.63720471e-01 1.33544290e+00 -9.51371253e-01 -3.48423310e-02
-3.56438130e-01 8.90287697e-01 5.47118247e-01 5.85384309e-01
-8.05098236e-01 -1.22873381e-01 5.21275520e-01 2.20244825e-01
2.79698759e-01 -2.73546875e-01 -4.58733678e-01 -1.31127453e+00
2.47645766e-01 -7.37416327e-01 5.77938616e-01 -7.37712324e-01
-1.55417073e+00 4.49569374e-01 -1.11638956e-01 -1.23137903e+00
5.07697880e-01 -2.72640556e-01 -5.22147775e-01 5.21360695e-01
-1.52548409e+00 -9.49794888e-01 -7.14895248e-01 5.39766490e-01
2.63101995e-01 -1.81484576e-02 1.73846886e-01 7.37279415e-01
-1.02091551e+00 4.75332171e-01 -1.60653591e-01 1.13943435e-01
5.61632693e-01 -1.02111363e+00 3.00257891e-01 9.32782650e-01
1.11459345e-01 3.76382798e-01 5.35052717e-01 -9.45135593e-01
-1.40232968e+00 -1.53685510e+00 7.36900568e-01 2.24466287e-02
5.83891749e-01 -4.23677534e-01 -1.15622604e+00 6.93877876e-01
-3.19310099e-01 3.38588327e-01 3.59968960e-01 1.16885141e-01
-3.40976149e-01 -2.60168433e-01 -8.35743606e-01 7.50626266e-01
1.52114749e+00 -2.01478451e-01 6.01046272e-02 5.69411993e-01
8.12554240e-01 -3.62267137e-01 -8.26572895e-01 6.76274955e-01
1.55432895e-01 -8.29165459e-01 1.07583797e+00 -2.75832951e-01
5.88896930e-01 -4.63093668e-01 -1.93332016e-01 -1.15882266e+00
-5.79178154e-01 -4.73903835e-01 -2.56204337e-01 1.43460977e+00
4.26922977e-01 -7.46097505e-01 7.91439414e-01 4.45036322e-01
-1.28050730e-01 -9.79255319e-01 -5.86704552e-01 -7.80683815e-01
-4.90490139e-01 -3.53521287e-01 3.67598444e-01 1.22972012e+00
-4.74294633e-01 7.10308790e-01 -5.23544431e-01 5.60353994e-01
8.92973542e-01 8.80523399e-02 8.16075683e-01 -1.24863577e+00
-3.93971503e-01 -3.31818163e-01 -5.40986717e-01 -1.12685061e+00
4.21237946e-02 -1.10565567e+00 -1.82181410e-02 -1.79066527e+00
2.48320788e-01 -4.89949346e-01 -2.71786034e-01 5.33315122e-01
-5.28682768e-01 3.70088130e-01 2.82917529e-01 3.17096144e-01
-7.74976313e-01 8.49472702e-01 1.36903834e+00 -2.19438359e-01
-3.79420519e-01 2.55722106e-02 -8.89066696e-01 8.24693620e-01
7.73933232e-01 -6.51434302e-01 -4.64917809e-01 -4.24969107e-01
5.65144837e-01 -1.26540229e-01 5.83468854e-01 -8.34419131e-01
2.60672212e-01 -3.60056311e-02 2.01151565e-01 -5.92908919e-01
1.77626446e-01 -5.34898996e-01 2.70406544e-01 2.78038919e-01
-2.79909521e-01 -1.97220564e-01 -7.99335539e-02 7.60130525e-01
-1.76419124e-01 5.46264276e-02 6.35558128e-01 -7.43453996e-03
-7.62086391e-01 5.62133491e-01 7.22744241e-02 1.94795296e-01
7.26278484e-01 -2.45268747e-01 -4.15906072e-01 -3.50807816e-01
-5.87617874e-01 7.00206220e-01 2.55983949e-01 2.93223798e-01
7.10235357e-01 -1.46319234e+00 -8.76727462e-01 1.01343445e-01
1.98573351e-01 3.43154490e-01 3.02424848e-01 1.12158608e+00
-3.78864974e-01 -2.55652010e-01 2.23098129e-01 -6.74801528e-01
-1.35631180e+00 2.61138588e-01 2.42511526e-01 -3.42863500e-01
-8.06436062e-01 1.08565247e+00 5.77709138e-01 -5.78463972e-01
3.94079179e-01 -3.50702256e-02 -2.41806373e-01 8.63994062e-02
1.23033024e-01 5.13809800e-01 -1.12786599e-01 -6.14304960e-01
-4.35642064e-01 3.94213736e-01 -4.07537781e-02 5.46976209e-01
1.36557424e+00 -2.91211814e-01 -4.41817313e-01 3.17242891e-01
1.13399398e+00 -2.52919763e-01 -1.29167497e+00 -4.63339478e-01
-6.47130236e-02 -6.08204484e-01 2.75788195e-02 -3.88085276e-01
-1.48486006e+00 3.35163981e-01 3.24383438e-01 2.05294222e-01
1.20336461e+00 2.47271359e-01 6.63720667e-01 2.87775457e-01
3.18905376e-02 -8.64061475e-01 2.87621200e-01 4.12080348e-01
1.16006529e+00 -1.23572052e+00 4.06435549e-01 -1.32032752e+00
-5.81018806e-01 7.19741225e-01 8.57597649e-01 -1.90865055e-01
7.30375230e-01 -1.31402537e-01 -1.22068174e-01 -5.94100595e-01
-6.14426970e-01 -2.37754717e-01 5.34091711e-01 3.47688556e-01
2.37801328e-01 -5.57101220e-02 -3.78530174e-01 4.23848689e-01
5.68040907e-02 -1.77373722e-01 3.74416620e-01 7.70167351e-01
-2.05456465e-01 -7.35769212e-01 -2.98351109e-01 7.82558560e-01
-1.58174679e-01 -1.38951972e-01 -4.93490189e-01 4.99705315e-01
1.15393050e-01 9.78548408e-01 -5.71654662e-02 -5.06067693e-01
4.12442535e-01 -5.34541368e-01 3.12367707e-01 -5.89266777e-01
-3.59717578e-01 9.75615382e-02 1.86236545e-01 -6.84367180e-01
-4.03778255e-01 -3.76364827e-01 -1.15137172e+00 -2.18267560e-01
-5.48457146e-01 -2.40938559e-01 1.94977492e-01 6.34327590e-01
6.61446869e-01 8.25463593e-01 6.16915226e-01 -7.35837460e-01
-2.77016371e-01 -7.01851785e-01 -8.86252105e-01 5.37331283e-01
2.79315989e-02 -6.91699088e-01 -5.17317057e-01 -1.66375205e-01] | [7.461837291717529, 6.009706020355225] |
1946cd7d-a1e2-41e3-a686-6ad4f3083a88 | thy-friend-is-my-friend-iterative | null | null | http://papers.nips.cc/paper/7057-thy-friend-is-my-friend-iterative-collaborative-filtering-for-sparse-matrix-estimation | http://papers.nips.cc/paper/7057-thy-friend-is-my-friend-iterative-collaborative-filtering-for-sparse-matrix-estimation.pdf | Thy Friend is My Friend: Iterative Collaborative Filtering for Sparse Matrix Estimation | The sparse matrix estimation problem consists of estimating the distribution of an $n\times n$ matrix $Y$, from a sparsely observed single instance of this matrix where the entries of $Y$ are independent random variables. This captures a wide array of problems; special instances include matrix completion in the context of recommendation systems, graphon estimation, and community detection in (mixed membership) stochastic block models. Inspired by classical collaborative filtering for recommendation systems, we propose a novel iterative, collaborative filtering-style algorithm for matrix estimation in this generic setting. We show that the mean squared error (MSE) of our estimator converges to $0$ at the rate of $O(d^2 (pn)^{-2/5})$ as long as $\omega(d^5 n)$ random entries from a total of $n^2$ entries of $Y$ are observed (uniformly sampled), $\E[Y]$ has rank $d$, and the entries of $Y$ have bounded support. The maximum squared error across all entries converges to $0$ with high probability as long as we observe a little more, $\Omega(d^5 n \ln^5(n))$ entries. Our results are the best known sample complexity results in this generality. | ['Christina E. Lee', 'Jennifer Chayes', 'Devavrat Shah', 'Christian Borgs'] | 2017-12-01 | null | null | null | neurips-2017-12 | ['graphon-estimation'] | ['graphs'] | [ 1.47600874e-01 1.23769656e-01 -4.78080735e-02 -4.76881489e-02
-9.80757713e-01 -6.01818204e-01 -1.39283642e-01 8.92224070e-03
-4.92033541e-01 5.74055493e-01 -1.59835339e-01 -3.95109832e-01
-7.86010981e-01 -1.00091660e+00 -8.77859831e-01 -8.59873354e-01
-9.09501553e-01 6.03220642e-01 -2.65366137e-01 -1.92413688e-01
1.33547887e-01 -5.19699156e-02 -1.14345419e+00 -4.61126506e-01
6.18564487e-01 1.20301688e+00 -1.43520176e-01 8.44674110e-01
-7.56213814e-03 5.95364511e-01 -4.54722643e-01 -7.64997184e-01
7.10595191e-01 -4.14115787e-01 -4.37676579e-01 3.71227264e-01
2.11785480e-01 8.26573148e-02 -5.98672926e-01 1.59097672e+00
1.78132445e-01 3.47226202e-01 5.32337904e-01 -1.03316271e+00
-3.11571956e-01 9.80862856e-01 -8.76926959e-01 4.66257818e-02
4.41248804e-01 -1.68739825e-01 1.10656679e+00 -9.27078903e-01
5.98958790e-01 9.96055126e-01 7.55685270e-01 -1.14605956e-01
-1.50245082e+00 -1.17031085e+00 3.42254639e-01 -2.95991719e-01
-1.82690585e+00 -4.21277612e-01 5.29403389e-01 -4.92912024e-01
3.11288685e-01 2.72728235e-01 3.27691674e-01 5.12324095e-01
-4.20903355e-01 5.77635944e-01 8.64502549e-01 -4.27202493e-01
3.75097215e-01 5.22572473e-02 1.15073323e-02 8.45623434e-01
7.31416643e-01 -2.18890503e-01 -4.96676683e-01 -4.75895375e-01
7.51107633e-01 3.20629090e-01 -2.12059930e-01 -3.42387408e-01
-9.19132829e-01 1.01022220e+00 4.81751896e-02 1.45196185e-01
-3.93964946e-01 5.32220066e-01 4.95010894e-03 8.12033415e-01
1.15848951e-01 9.77611169e-02 -3.86318833e-01 -7.54118860e-02
-8.62482011e-01 1.46654740e-01 1.13331592e+00 1.27028728e+00
9.48962331e-01 1.79179490e-01 2.93566972e-01 7.07498014e-01
4.08720613e-01 1.36804199e+00 -3.22271138e-01 -1.35515487e+00
8.24729323e-01 4.38513726e-01 4.60922390e-01 -1.23237765e+00
8.16147700e-02 -6.11471951e-01 -1.24295533e+00 1.52303483e-02
7.95467973e-01 -4.94465351e-01 -2.19329342e-01 1.97813177e+00
1.40261397e-01 -2.97541991e-02 -3.58800560e-01 4.85871404e-01
2.49593277e-02 4.60000277e-01 -4.12085265e-01 -6.43744528e-01
1.02210855e+00 -1.73796698e-01 -3.82109821e-01 -3.70017916e-01
2.67063946e-01 -7.30024219e-01 6.28787637e-01 5.56515574e-01
-1.52511847e+00 -1.23626351e-01 -6.53136373e-01 7.51666903e-01
3.32764804e-01 -4.44430448e-02 6.97383344e-01 9.52054381e-01
-1.08987045e+00 4.50174689e-01 -4.91416723e-01 9.10564512e-02
2.78280675e-01 6.07295573e-01 -3.43792289e-01 -5.91625571e-01
-7.63362885e-01 3.62881981e-02 -4.77610886e-01 4.24898490e-02
-7.20465064e-01 -3.43464851e-01 -6.34501576e-01 2.85788346e-02
7.85378695e-01 -4.83323514e-01 7.88474560e-01 -5.59375584e-01
-9.26745713e-01 5.40094852e-01 -3.63537967e-01 -4.23210621e-01
2.74959803e-01 2.89132521e-02 -2.76686132e-01 1.99230865e-01
3.36422741e-01 -3.67013067e-01 9.46148813e-01 -8.60813439e-01
-5.52939475e-01 -6.68889582e-01 -4.93274443e-02 -8.44520703e-02
-1.14171088e-01 1.80848241e-02 -6.44958973e-01 -3.85138601e-01
5.03497481e-01 -1.06466746e+00 -8.49358976e-01 -8.41227174e-02
-2.03383788e-01 1.17090248e-01 -4.36099209e-02 -2.10878611e-01
1.45870912e+00 -2.25995493e+00 1.75962448e-01 1.10622013e+00
6.57154560e-01 -2.63624221e-01 -1.53514385e-01 8.10581267e-01
1.80578053e-01 5.13706915e-02 -3.68274003e-01 -3.12781125e-01
1.78183898e-01 -1.55710563e-01 1.80096645e-02 8.76448691e-01
-7.02017069e-01 1.13390341e-01 -7.20542371e-01 8.05652887e-02
-7.77839497e-02 1.11847907e-01 -7.53762662e-01 -1.68480605e-01
-1.55890167e-01 1.09558232e-01 -5.10812104e-01 5.00554562e-01
9.18905795e-01 -8.72904420e-01 7.71045268e-01 3.36160660e-01
2.09233895e-01 -2.13561580e-01 -2.13487029e+00 1.50495136e+00
-2.14853004e-01 3.50844264e-01 8.80874634e-01 -1.30089593e+00
6.32727265e-01 3.94004405e-01 8.19752276e-01 -2.38109380e-01
3.20713043e-01 3.54965150e-01 1.03447489e-01 -6.76925853e-02
1.95746601e-01 -2.69735545e-01 -3.57971698e-01 8.94164443e-01
-9.51651707e-02 1.79677531e-01 3.80492389e-01 7.16163635e-01
1.80020976e+00 -8.62146318e-01 2.18456715e-01 -2.13317066e-01
2.63555735e-01 -3.64065170e-01 6.48453653e-01 1.43655312e+00
-1.14661396e-01 3.69603038e-02 6.90300286e-01 1.26899526e-01
-9.58609045e-01 -1.13633108e+00 7.56483078e-02 1.04026830e+00
1.77375615e-01 -6.35277927e-01 -5.62101126e-01 -3.17341149e-01
2.08589554e-01 1.80358261e-01 -6.18707240e-01 3.27932775e-01
-3.79839212e-01 -8.41045022e-01 -4.49727066e-02 2.66614199e-01
4.76026356e-01 -6.04678690e-01 3.01723480e-01 3.70057344e-01
7.94402361e-02 -8.89338732e-01 -6.84156239e-01 1.95150718e-01
-8.36552501e-01 -1.29888320e+00 -5.21803856e-01 -6.02048755e-01
9.22791064e-01 5.57324290e-01 9.83988225e-01 -2.98007093e-02
-1.85594216e-01 7.09078312e-01 -1.94155648e-01 8.47446695e-02
-6.66785091e-02 -4.22731847e-01 3.09183806e-01 1.20407835e-01
4.29788530e-01 -8.98824990e-01 -9.12939787e-01 4.29075152e-01
-7.45714188e-01 -9.04165983e-01 6.73541665e-01 6.07632279e-01
6.20266378e-01 3.69482219e-01 2.21994996e-01 -1.12784100e+00
4.47933584e-01 -7.03825355e-01 -9.45547521e-01 -2.64461599e-02
-7.69870818e-01 -8.38508904e-02 8.24677527e-01 -4.95867729e-01
-3.13933522e-01 -1.56514257e-01 -5.39680198e-03 -5.27610064e-01
2.65564233e-01 7.75057375e-01 -8.20794702e-02 -5.59291206e-02
7.18605995e-01 3.19243759e-01 1.97371036e-01 -5.37214041e-01
5.66637218e-01 3.56525779e-01 2.29521617e-01 -5.40059268e-01
8.42845321e-01 7.33289242e-01 3.08902442e-01 -7.65430450e-01
-4.45476532e-01 -5.95082045e-01 5.50094694e-02 1.80681750e-01
4.56268303e-02 -1.16478789e+00 -1.40348744e+00 8.29052851e-02
-2.66974062e-01 -3.23906749e-01 -3.39877397e-01 7.11762369e-01
-3.96031380e-01 5.61665058e-01 -4.68058676e-01 -1.36798346e+00
-9.32242647e-02 -8.01764905e-01 4.10813987e-01 -1.91139400e-01
-5.05528525e-02 -7.50369072e-01 -7.34645128e-02 2.19605431e-01
3.99035335e-01 -4.93294448e-02 7.33396411e-01 -4.48781550e-01
-9.10749376e-01 -6.88157380e-01 -2.20131800e-01 3.03629577e-01
-1.21249028e-01 -3.94099832e-01 -2.68602502e-02 -5.96711457e-01
8.44549015e-02 8.94236639e-02 5.65697312e-01 6.99145198e-01
9.84060824e-01 -6.77947283e-01 -3.69018197e-01 3.15742433e-01
1.59624279e+00 1.52951255e-01 3.82557899e-01 -3.18842411e-01
1.93941265e-01 3.76122817e-02 3.43844056e-01 1.10877252e+00
9.49919596e-02 2.31865302e-01 2.76714534e-01 2.98732013e-01
4.55639988e-01 4.50320765e-02 3.31847459e-01 8.40386271e-01
2.17498258e-01 -2.54489243e-01 -6.28248990e-01 6.09221876e-01
-1.73585010e+00 -1.12543344e+00 -1.39732808e-01 2.65309763e+00
7.08442807e-01 3.02017897e-01 3.84927571e-01 2.00927556e-01
8.07326853e-01 -2.74719112e-02 -5.32311141e-01 6.47482276e-02
-1.83752865e-01 9.93424177e-01 9.97620881e-01 5.95052183e-01
-7.33914733e-01 4.16567981e-01 6.32934284e+00 1.05451965e+00
-4.12926704e-01 1.56231159e-02 3.99994940e-01 -5.24698734e-01
-4.11693603e-01 2.20662624e-01 -8.15580010e-01 6.76261365e-01
9.54638660e-01 -6.47833794e-02 6.81129694e-01 1.07429647e+00
7.92098492e-02 -3.86404902e-01 -9.01157200e-01 1.29702055e+00
3.15016955e-02 -1.38248646e+00 -5.06373286e-01 5.74128985e-01
9.81723845e-01 5.54047013e-03 3.64945889e-01 1.25295341e-01
1.23684561e+00 -7.90249586e-01 2.52575874e-01 2.62152195e-01
9.30304110e-01 -7.79089451e-01 3.21056157e-01 4.62591738e-01
-1.32126808e+00 -5.08796632e-01 -4.41160470e-01 -1.52933493e-01
3.01101923e-01 1.11220908e+00 -3.55543554e-01 1.07890166e-01
7.74119377e-01 4.28348213e-01 2.14754447e-01 9.16850924e-01
1.19765155e-01 7.71719813e-01 -8.04568887e-01 -1.89447433e-01
7.94862434e-02 -7.12918043e-01 6.61941469e-01 6.63763046e-01
6.44258857e-01 5.45316458e-01 3.55547398e-01 3.20392370e-01
-6.16835654e-01 2.86370814e-01 -6.03402674e-01 -9.04928595e-02
9.68066931e-01 7.62543201e-01 -6.11659229e-01 -2.97699481e-01
-4.51656371e-01 7.59565592e-01 1.23972021e-01 3.56279105e-01
-5.06335497e-01 -5.48099220e-01 8.62611890e-01 3.24169159e-01
9.25793946e-01 -5.78461051e-01 -1.50690749e-01 -1.15076077e+00
-1.27568152e-02 -1.00380981e+00 5.23899496e-01 -1.94008559e-01
-1.49809432e+00 2.82630146e-01 -3.33502501e-01 -1.07427084e+00
-1.68128967e-01 -5.01109421e-01 -3.00336629e-01 7.49832988e-01
-6.07779324e-01 -1.26441836e-01 1.84842452e-01 9.49892998e-01
-2.19097465e-01 -3.83113384e-01 7.54044652e-01 4.34413165e-01
-5.88692784e-01 6.75800443e-01 8.89927506e-01 2.61723310e-01
4.38295275e-01 -1.07052398e+00 -7.07261730e-03 9.84980881e-01
2.83563793e-01 9.61487293e-01 7.87080109e-01 -5.81312776e-01
-1.71462429e+00 -6.33773267e-01 6.65252626e-01 -2.52416134e-01
9.07299817e-01 -3.55841756e-01 -3.29580158e-01 7.67241716e-01
-1.19286381e-01 2.32589673e-02 1.24123359e+00 4.65441197e-01
-5.23370922e-01 -2.91284382e-01 -1.25557888e+00 5.76463461e-01
1.30434370e+00 -4.75707144e-01 2.69213736e-01 3.95636499e-01
2.39038348e-01 -9.33635831e-02 -1.13010669e+00 9.42459255e-02
4.40862715e-01 -8.89300406e-01 1.05063856e+00 -3.50553453e-01
-1.19559541e-01 -2.26398677e-01 -8.39981139e-01 -7.11143553e-01
-4.08922166e-01 -1.05248523e+00 -2.77640164e-01 4.80472863e-01
6.07335150e-01 -4.78476465e-01 1.33562279e+00 5.71048617e-01
5.35294592e-01 -3.84690017e-01 -9.73342240e-01 -5.35975099e-01
-2.79518422e-02 -5.60078442e-01 2.85867244e-01 7.06298828e-01
-2.71524340e-02 1.77151963e-01 -4.94908363e-01 1.77890226e-01
1.07068217e+00 3.02731723e-01 1.01777077e+00 -1.06053078e+00
-1.01853406e+00 -3.17580372e-01 -1.41007990e-01 -1.55345559e+00
-2.51171619e-01 -7.01694310e-01 -2.59586632e-01 -9.54440951e-01
3.79353046e-01 -1.08441556e+00 -1.65815264e-01 -1.04412407e-01
4.75689135e-02 1.93297446e-01 7.47850090e-02 2.10212871e-01
-1.00033033e+00 2.09882021e-01 7.90318549e-01 5.54611273e-02
3.00777778e-02 4.71811026e-01 -1.04314411e+00 6.40561759e-01
2.95071393e-01 -5.31521201e-01 -3.04321736e-01 -3.00911009e-01
7.98364282e-01 9.45373178e-01 -1.72324508e-01 -7.71597743e-01
2.65329868e-01 -7.74582550e-02 5.63844331e-02 -5.37899971e-01
3.87976527e-01 -7.23712683e-01 3.98923337e-01 3.71582538e-01
-2.14890867e-01 -1.46761620e-02 -4.18576688e-01 1.14809072e+00
5.19991033e-02 -2.97343224e-01 6.57538295e-01 -3.40377867e-01
-1.31268250e-02 7.23851562e-01 -4.98276144e-01 2.86481768e-01
6.80188298e-01 -1.51083991e-01 8.75081569e-02 -1.00320673e+00
-1.02527833e+00 1.48941010e-01 3.23182136e-01 -5.16105115e-01
3.02249223e-01 -9.90279198e-01 -6.97285295e-01 1.67677999e-01
-8.45699832e-02 6.11335598e-02 6.71237111e-01 1.03653085e+00
-1.10654913e-01 4.90223728e-02 4.91445929e-01 -6.03696585e-01
-8.51885378e-01 4.07423675e-01 1.70820660e-03 -3.98378819e-01
-1.38372540e-01 1.44776940e+00 -2.55878359e-01 -6.22014813e-02
3.70992661e-01 1.41114101e-01 4.73062426e-01 -1.52442396e-01
5.79866886e-01 4.90228951e-01 -2.59777278e-01 -2.58445323e-01
-2.25699916e-01 2.68486917e-01 -1.30128264e-01 -5.73300898e-01
1.29435527e+00 -4.78217304e-01 -2.76047945e-01 1.17260478e-01
1.32777250e+00 7.85014927e-01 -1.13678527e+00 -8.18785906e-01
-3.68513972e-01 -6.03830516e-01 -5.45479655e-01 -2.60939270e-01
-1.23746061e+00 4.20333058e-01 3.75609785e-01 4.41159248e-01
8.87223423e-01 1.28717735e-01 4.85255033e-01 4.78810698e-01
7.42044985e-01 -9.65490818e-01 1.85695291e-01 4.74073440e-01
2.21993744e-01 -1.06304765e+00 -5.11067063e-02 -3.19568962e-01
-2.87746042e-01 6.15533173e-01 -1.50858415e-02 -6.80268705e-01
1.25537884e+00 1.73260733e-01 -4.58632469e-01 -2.83007622e-01
-5.59456348e-01 -3.06091234e-02 -2.47944579e-01 3.38422000e-01
2.93155342e-01 1.17671743e-01 -2.75369108e-01 7.92281091e-01
-2.15992734e-01 -1.78719595e-01 5.27087510e-01 8.74271214e-01
-7.65141070e-01 -1.45739841e+00 -4.84644622e-01 9.29929614e-01
-6.60017729e-01 -3.32612514e-01 2.23061725e-01 4.89452392e-01
-1.51677296e-01 1.18736160e+00 -1.03608876e-01 -2.76596844e-01
9.37283188e-02 -1.79392025e-01 5.75788975e-01 -5.64823210e-01
-3.04310471e-01 3.65344286e-01 -1.26674458e-01 -4.32224005e-01
-1.79678097e-01 -9.52100635e-01 -9.08988774e-01 -1.05111527e+00
-1.81228042e-01 6.80468738e-01 5.83839774e-01 7.22762227e-01
3.94193411e-01 -3.07924122e-01 1.09989393e+00 -1.56479985e-01
-7.00985312e-01 -8.84987891e-01 -1.46288085e+00 3.67003888e-01
-8.44589397e-02 -4.70996350e-01 -8.20956111e-01 -1.99078470e-01] | [6.717716217041016, 4.7385430335998535] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.