PosteriorValueFunctions:HindsightBaselinesforPolicyGradientMethodsChrisNota1BrunoCastrodaSilva1PhilipS.Thomas1Abstractcases,suchinformationcanbeusefulforassessingwhichoutcomeswerelikelytohaveoccurr...
Instance-OptimalCompressedSensingviaPosteriorSamplingAjilJalal1SushrutKarmalkar2AlexandrosG.Dimakis1EricPrice2AbstractforameasurementmatrixA∈Rm×nandnoisevectorWecharacterizethemeasurementcomplexi...
HowGoodistheBayesPosteriorinDeepNeuralNetworksReally?FlorianWenzel1KevinRoth+2BastiaanS.Veeling+31JakubS´wia˛tkowski4+LinhTran5+StephanMandt6+JasperSnoek1TimSalimans1RodolpheJenatton1SebastianNow...
MyopicPosteriorSamplingforAdaptiveGoalOrientedDesignofExperimentsKirthevasanKandasamy1WillieNeiswanger1ReedZhang1AkshayKrishnamurthy2JeffSchneider1Barnaba´sPo´czos1Abstractity.Onadifferentday,she...
AutomaticPosteriorTransformationforLikelihood-freeInferenceDavidS.Greenberg1MarcelNonnenmacher1JakobH.Macke1AbstractForcomplexdata-generatingprocessessuchassimulation-basedmodels,itisoftenimpossibl...
WhyisPosteriorSamplingBetterthanOptimismforReinforcementLearning?IanOsband12BenjaminVanRoy1Abstractmateoffuturevalueandselectstheactionwiththegreatestestimate.Ifaselectedactionisnotnear-optimal,the...
VariationalBoosting:IterativelyRefiningPosteriorApproximationsAndrewC.Miller1NicholasJ.Foti2RyanP.Adams13AbstractchainMonteCarlo(MCMC)methodsthatapproximateatargetdistributionwithsamplesdrawnfromaM...
EvaluatingBayesianModelswithPosteriorDispersionIndicesAlpKucukelbir1YixinWang1DavidM.Blei1Abstractvalidationcapturesanotionofgeneralizationandjustifiesholdingoutdata.Thissimplecombinationhasfueledt...