QuantifyingtheBenefitofUsingDifferentiableLearningoverTangentKernelsEranMalach1PritishKamath2EmmanuelAbbe3NathanSrebro2AbstractCanallthesuccessofdeeplearningbeexplainedusingtheNTK?Thiswouldimplytha...
PODS:PolicyOptimizationviaDifferentiableSimulationMiguelZamora1MomchilPeychev1SehoonHa2MartinVechev1StelianCoros1Abstractpotentiallyunsafe.Fortunately,recentyearshaveseenexcit-ingprogressinsimulati...
iDARTS:DifferentiableArchitectureSearchwithStochasticImplicitGradientsMiaoZhang12StevenSu2ShiruiPan1XiaojunChang1EhsanAbbasnejad3RezaHaffari1Abstractplications(Renetal.,2020;Chengetal.,2020;Cheneta...
HardCoRe-NAS:HardConstrainedDifferentiableNeuralArchitectureSearchNivNayman1YonathanAflalo1AsafNoy1LihiZelnik-Manor1Abstract78Oursfromscratch77.5Oursfine-tune:400+15NRealisticuseofneuralnetworksoft...
EfficientDifferentiableSimulationofArticulatedBodiesYi-LingQiao12JunbangLiang1VladlenKoltun2MingC.Lin1Abstract2019).However,theydonothaveaccesstoanalyticderiva-tivesofthearticulateddynamics.Therefo...
DifferentiableSpatialPlanningusingTransformersDevendraSinghChaplot12DeepakPathak2JitendraMalik13Projectwebpage:https://devendrachaplot.github.io/projects/spatial-planning-transformersAbstractFigure...
DifferentiableSortingNetworksforScalableSortingandRankingSupervisionFelixPetersen1ChristianBorgelt2HildeKuehne34OliverDeussen1AbstractRecently,theideaofend-to-endtrainingofneuralnetworkswithsorting...
DifferentiableParticleFilteringviaEntropy-RegularizedOptimalTransportAdrienCorenflos1JamesThornton2GeorgeDeligiannidis2ArnaudDoucet2Abstractprocess(Xt)t≥1andY-valuedobservations(Yt)t≥1satis-fying...
DifferentiableDynamicQuantizationwithMixedPrecisionandAdaptiveResolutionZhangZhaoyang1ShaoWenqi1GuJinwei23WangXiaogang1LuoPing4Abstractdynamicrangeandstepsizearefreezedineachlayer.Sec-ondly,gradien...
DatasetCondensationwithDifferentiableSiameseAugmentationBoZhao1HakanBilen1Abstractmassivedatasets(e.g.(Kuznetsovaetal.,2020;Chenetal.,2020a;Kwiatkowskietal.,2019)).Forexample,ittakesInmanymachinele...
ADifferentiablePointProcesswithItsApplicationtoSpikingNeuralNetworksHiroshiKajino1Abstractraisesthefollowingquestion:howmuchofthecurrentper-formancegapscanbeattributedtothisdifferenceonneuralThispa...
TheDifferentiableCross-EntropyMethodBrandonAmos1DenisYarats12Abstractmodelthatisgeneratingpotentialoutcomesconditionalonxhappeningthatyouwanttooptimizeover.End-to-endlearn-WestudytheCross-EntropyMe...
StabilizingDifferentiableArchitectureSearchviaPerturbation-basedRegularizationXiangningChen1Cho-JuiHsieh1AbstractHowever,thesemethodsusuallyrequiremassivecomputa-tionresources.Recently,avarietyofap...
ScalableDifferentiablePhysicsforLearningandControlYi-LingQiao1JunbangLiang1VladlenKoltun2MingC.Lin1AbstractRecenteffortshavesignificantlyadvancedtheunderstandingofDifferentiablephysicsinmachinelear...
OptimizingDataUsageviaDifferentiableRewardsXinyiWang1HieuPham12PaulMichel1AntoniosAnastasopoulos1JaimeCarbonell1GrahamNeubig1AbstractPreviousworkhasattemptedtocreatestrategiestohandlethissensitivit...
Operation-AwareSoftChannelPruningusingDifferentiableMasksMinsooKang1BohyungHan1AbstractNotUpdated1.00.0UpdatedstrongweakBinaryMasksFeatureMapWeproposeasimplebuteffectivedata-drivenDifferentiableMas...
LP-SparseMAP:DifferentiableRelaxedOptimizationforSparseStructuredPredictionVladNiculae1AndréF.T.Martins123AbstractsleeptheclockaroundBUDGETfactorTREEfactorStructuredpredictorsrequiresolvingacombin...
LinearLowerBoundsandConditioningofDifferentiableGamesAdamIbrahim1WaïssAzizian2GauthierGidel1IoannisMitliagkas1Abstractmind(Meschederetal.,2017),andtomakemattersworse,havebeenoftentunedsuboptimally...
LearningReasoningStrategiesinEnd-to-EndDifferentiableProvingPasqualeMinervini1SebastianRiedel12PontusStenetorp1EdwardGrefenstette12TimRocktäschel12Abstract2018;Johnsonetal.,2017;Sinhaetal.,2019).F...
FastDifferentiableSortingandRankingMathieuBlondel1OlivierTeboul1QuentinBerthet1JosipDjolonga1Abstractarchitecturesforwhichagradientcanbecomputed.Thesortingoperationisoneofthemostcom-Wefocusinthispa...