ImportanceSamplingPolicyEvaluationwithanEstimatedBehaviorPolicyJosiahP.Hanna1ScottNiekum1PeterStone1Abstractdeterminetheexpectedreturn–sumofrewards–thatanevaluationpolicy,πe,willobtainwhendeploy...
HybridModelswithDeepandInvertibleFeaturesEricNalisnick1AkihiroMatsukawa1YeeWhyeTeh1DilanGorur1BalajiLakshminarayanan1Abstractdetection(Bishop,1994),semi-supervisedlearning(Drucketal.,2007),andinfor...
High-FidelityImageGenerationwithFewerLabelsMarioLucic1MichaelTschannen2MarvinRitter1XiaohuaZhai1OlivierBachem1SylvainGelly1AbstractFigure1.MedianFIDofthebaselinesandtheproposedmethod.Theverticallin...
Guidedevolutionarystrategies:AugmentingrandomsearchwithsurrogategradientsNiruMaheswaranathan1LukeMetz1GeorgeTucker1DamiChoi1JaschaSohl-Dickstein1Abstractmethodssuchasgradientdescentarepopularduetot...
GuaranteesforSpectralClusteringwithFairnessConstraintsMattha¨usKleindessner1SamiraSamadi2PranjalAwasthi1JamieMorgenstern2Abstracttostudyappropriatenotionsoffairnessforunsupervisedlearningsettings(...
GainingFreeorLow-CostTransparencywithInterpretablePartialSubstituteTongWang1Abstractingtwosolutionsforfacilitatinghumanunderstandability.Thisworkaddressesthesituationwhereablack-Thefirstsolutionist...
Flow++:ImprovingFlow-BasedGenerativeModelswithVariationalDequantizationandArchitectureDesignJonathanHo1XiChen12AravindSrinivas1YanDuan2PieterAbbeel12Abstractvideos,toaudiosignalsandnaturallanguage(...
FormalPrivacyforFunctionalDatawithGaussianPerturbationsArdalanMirshani1MatthewReimherr1AleksandraSlavkovic1Abstractdata,withthecorrespondingbranchofstatisticscalledfunc-tionaldataanalysis,FDA.FDAis...
FlatMetricMinimizationwithApplicationsinGenerativeModelingThomasMo¨llenhoff1DanielCremers1Abstractfromlefttorightwevarythelatentcodez1(time)WetakethenovelperspectivetoviewdatanotFigure1.Discoverin...
Finite-TimeAnalysisofDistributedTD(0)withLinearFunctionApproximationforMulti-AgentReinforcementLearningThinhT.Doan12SivaThejaMaguluri1JustinRomberg2Abstractenvironment,oftenmodeledasaMarkovDecision...
FasterAttend-Infer-RepeatwithTractableProbabilisticModelsKarlStelzner1RobertPeharz2KristianKersting13AbstractRecently,deepneuralgenerativemodelssuchasvariationalautoencoders(VAEs)(Kingma&Welling,20...
FastandSimpleNatural-GradientVariationalInferencewithMixtureofExponential-familyApproximationsWuLin1MohammadEmtiyazKhan2MarkSchmidt1AbstractMinimalMixtureofEFExamplesofMCEFsConditionallyCurvedEFFin...
FastAlgorithmforGeneralizedMultinomialModelswithRankingDataJiaqiGu1GuoshengYin1Abstractsharedstructuresinthelog-likelihoodfunction.Wetheoret-icallyshowthatthemaximumlikelihoodestimator(MLE)Wedevelo...
FairnesswithoutHarm:DecoupledClassifierswithPreferenceGuaranteesBerkUstun1YangLiu2DavidC.Parkes1AbstractManyoftheproposedmethodsforfairmachinelearninghaveaimedtobuildmodelsthatpredictorperforminthe...
ExplainingDeepNeuralNetworkswithaPolynomialTimeAlgorithmforShapleyValuesApproximationMarcoAncona1CengizO¨ztireli2MarkusGross12Abstractton,2016),severalworkshavefocusedoninvestigatingmethodsforloca...
EscapingSaddlePointswithAdaptiveGradientMethodsMatthewStaib12SashankReddi3SatyenKale3SanjivKumar3SuvritSra1AbstractAdagradupdatestheparametersinthefollowingmanner:AdaptivemethodssuchasAdamandRMSPro...
EMI:ExplorationwithMutualInformationHyoungseokKim12JaekyeomKim12YeonwooJeong12SergeyLevine3HyunOhSong12Abstract83Reinforcementlearningalgorithmsstrugglewhen177therewardsignalisverysparse.Inthesecas...
EfficientOptimizationofLoopsandLimitswithRandomizedTelescopingSumsAlexBeatson1RyanP.Adams1Abstractovertime,eachlearningsteprequiresloopingovertimesteps.Morebroadly,inmanyscientificandengineeringWec...
EfficientNonconvexRegularizedTensorCompletionwithStructure-awareProximalIterationsQuanmingYao12JamesT.Kwok2BoHan3AbstractInmanyapplications,onlyafewtensorentriesareob-served.Forexample,ofteneachYou...
EfficientDictionaryLearningwithGradientDescentDarGilboa12SamBuchanan32JohnWright32Abstractthesemaybehighlysuboptimal,sincenonconvexoptimiza-tionisingeneralanNP-hardprobem(Bertsekas,1999).Randomlyin...