SparseBERT:RethinkingtheImportanceAnalysisinSelf-attentionHanShi1JiahuiGao2XiaozheRen3HangXu3XiaodanLiang4ZhenguoLi3JamesT.Kwok1AbstractincludetheBERT(Devlinetal.,2019),whichachievesstate-of-the-ar...
RepresentationMatters:AssessingtheImportanceofSubgroupAllocationsinTrainingDataEstherRolf1TheodoraWorledge1BenjaminRecht1MichaelI.Jordan12AbstractOurworkaimstodevelopageneralunifyingperspectiveonth...
MarginalContributionFeatureImportance-anAxiomaticApproachforExplainingDataAmnonCatav1BoyangFu2YazeedZoabi3AhuvaWeiss-Meilik4NoamShomron3JasonErnst256SriramSankararaman257RanGilad-Bachrach8Abstract1...
ADeepReinforcementLearningApproachtoMarginalizedImportanceSamplingwiththeSuccessorRepresentationScottFujimoto1DavidMeger1DoinaPrecup1Abstractapproachcanhavesignificantlylowervariancethantradi-tiona...
ARepresentationLearningPerspectiveontheImportanceofTrain-ValidationSplittinginMeta-LearningNikunjSaunshi1ArushiGupta1WeiHu1Abstractfrommany“train”taskstolearnausefulpriorthatcanhelpsolvenew“test...
ProblemswithShapley-value-basedexplanationsasfeatureImportancemeasuresI.ElizabethKumar1SureshVenkatasubramanian1CarlosScheidegger2SorelleA.Friedler3Abstractbythemodel,andthegameisthepredictionofthe...
FromImportanceSamplingtoDoublyRobustPolicyGradientJiaweiHuang1NanJiang1AbstractSummaryofthePaperWeprovideasimpleandpositiveanswertotheabovequestionintheepisodicRLsetting.InWeshowthaton-policypolicy...
TowardsUnderstandingtheImportanceofNoiseinTrainingNeuralNetworksMoZhou1TianyiLiu2YanLi2DachaoLin1EnluZhou2TuoZhao2AbstractSimplefirstorderalgorithmssuchasStochasticGradientDescent(SGD)anditsvariant...
OptimisticPolicyOptimizationviaMultipleImportanceSamplingMatteoPapini1AlbertoMariaMetelli1LorenzoLupo1MarcelloRestelli1Abstractpeholtetal.,2018).Thisiswellmotivated,asinteractingwithsomeenvironment...
ImportanceSamplingPolicyEvaluationwithanEstimatedBehaviorPolicyJosiahP.Hanna1ScottNiekum1PeterStone1Abstractdeterminetheexpectedreturn–sumofrewards–thatanevaluationpolicy,πe,willobtainwhendeploy...
HierarchicalImportanceWeightedAutoencodersChin-WeiHuang12KrisSankaran1EeshanDhekane1AlexandreLacoste2AaronCourville13Abstractboundswithprogressivelysmallergapusingmultiplei.i.d.samplesfromthevariat...
Dimension-WiseImportanceSamplingWeightClippingforSample-EfficientReinforcementLearningSeungyulHan1YoungchulSung1Abstractsamplesgeneratedbythebehaviorpolicywhichcanbedif-ferentfromthetargetpolicy.Of...
WhatistheEffectofImportanceWeightinginDeepLearning?JonathonByrd1ZacharyC.Lipton1AbstractEq[f(x)],Importancesamplingproducesanunbiasedesti-matebyweightingeachsamplexaccordingtothelikelihoodImportanc...
NonparametricvariableImportanceusinganaugmentedneuralnetworkwithmulti-tasklearningJeanFeng1BrianD.Williamson1MarcoCarone12NoahSimon1Abstractrequiresarigorousdefinitionofanestimablevariableim-portan...
ImportanceWeightedTransferofSamplesinReinforcementLearningAndreaTirinzoni1AndreaSessa1MatteoPirotta2MarcelloRestelli1Abstracttions,parameters,policies,etc.)andinthecriteriausedtoestablishwhethersuc...
IMPALA:ScalableDistributedDeep-RLwithImportanceWeightedActor-LearnerArchitecturesLasseEspeholt1HubertSoyer1RemiMunos1KarenSimonyan1VolodymyrMnih1TomWard1YotamDoron1VladFiroiu1TimHarley1IainDunning1...
IdentifyingBestInterventionsthroughOnlineImportanceSamplingRajatSen1KarthikeyanShanmugam2AlexandrosG.Dimakis1SanjayShakkottai1AbstractHiddenVariablesMotivatedbyapplicationsincomputationalad-User-UI...