Communication-efficientAlgorithmsforDistributedStochasticPrincipalComponentAnalysisDanGarber1OhadShamir2NathanSrebro3Abstractchinelearning.fordatathatconsistsofNvectorsinWestudythefundamentalproble...
CombiningModel-BasedandModel-FreeUpdatesforTrajectory-CentricReinforcementLearningYevgenChebotar12KarolHausman1MarvinZhang3GauravSukhatme1StefanSchaal12SergeyLevine3AbstractFigure1.Realrobottasksus...
CombinedGroupandExclusiveSparsityforDeepNeuralNetworksJaehongYoon1SungJuHwang12Abstractnumberofparameters,deepnetworksrequirelargeamountofmemoryandcomputationpowertotrain.Further,largeThenumberofpa...
CoherentProbabilisticforecastsforHierarchicalTimeSeriesSouhaibBenTaieb1JamesW.Taylor2RobJ.Hyndman1Abstractvariouslevelsofaggregationinordertooperatethepowergridefficientlyandsecurely(vanErven&Cugli...
CognitivePsychologyforDeepNeuralNetworks:AShapeBiasCaseStudySamuelRitter1DavidG.T.Barrett1AdamSantoro1MattM.Botvinick1Abstracttothinkofthesemodelsasblackboxes,andtoquestionwhethertheycanbeunderstoo...
CapacityReleasingDiffusionforSpeedandLocalityDiWang1KimonFountoulakis2MonikaHenzinger3MichaelW.Mahoney2SatishRao1Abstract2003;Mahoneyetal.,2012),andPageRank-basedspectralrankinginwebranking(Pageeta...
BidirectionalLearningforTime-seriesModelswithHiddenUnitsTakayukiOsogami1HiroshiKajino1TaroSekiyama1Abstractdenunits,however,makeitdifficulttolearntheparametersofthosemodels.forexample,theBaum-Welch...
BeyondFilters:CompactFeatureMapforPortableDeepModelYunheWang1ChangXu2ChaoXu1DachengTao2Abstractsignificantredundancyinmostofexistingconvolutionalneuralnetworks.forinstance,theResNet-50(Heetal.,Conv...
AxiomaticAttributionforDeepNetworksMukundSundararajan1AnkurTaly1QiqiYan1AbstractShrikumaretal.,2016;Binderetal.,2016;Springenbergetal.,2014).Westudytheproblemofattributingthepre-dictionofadeepnetwo...
Averaged-DQN:VarianceReductionandStabilizationforDeepReinforcementLearningOronAnschel1NirBaram1NahumShimkin1Abstractforproblem-specificstaterepresentation.Theseproblem-specificfeaturesdiminishtheag...
AutomatedCurriculumLearningforNeuralNetworksAlexGraves1MarcG.Bellemare1JacobMenick1Re´miMunos1KorayKavukcuoglu1Abstractmenttothenexttask,alongwithafixedprobabilityofre-turningtoearliertasks,toprev...
AsymmetricTri-trainingforUnsupervisedDomainAdaptationKuniakiSaito1YoshitakaUshiku1TatsuyaHarada12Abstractdataset(Girshicketal.,2014;Longetal.,2015a).ItisimportanttoapplymodelstrainedonalargeOneofth...
AsynchronousDistributedVariationalGaussianProcessforRegressionHaoPeng1ShandianZhe1XiaoZhang1YuanQi2Abstractmodelsistheirexpensivecomputationalcost:exactGPin-ferencerequiresO(n3)timecomplexityandO(n...
AnalogicalInferenceforMulti-relationalEmbeddingsHanxiaoLiu1YuexinWu1YimingYang1AbstractnucleussurroundedbyelectronsdownLarge-scalemulti-relationalembeddingreferstoscaleattractmadeofdownthetaskoflea...
AnEfficient,Sparsity-Preserving,OnlineAlgorithmforLow-RankApproximationDavidAnderson1MingGu1AbstractAdditionally,arecenttrendinmachinelearningistoin-cludeanapproximationofsecondorderinformationforL...
AnAlternativeSoftmaxOperatorforReinforcementLearningKavoshAsadi1MichaelL.Littman1AbstractAnidealsoftmaxoperatorisaparameterizedsetofoperatorsthat:Asoftmaxoperatorappliedtoasetofvaluesactssomewhatli...
AlgorithmsforpLow-RankApproximationFlavioChierichetti1SreenivasGollapudi2RaviKumar2SilvioLattanzi3RinaPanigrahy2DavidP.Woodruff4Abstractlow-rankapproximationinsteadoftheusuallystudied2(i.e.,Frobeni...
AlgebraicVarietyModelsforHigh-RankMatrixCompletionGregOngie1RebeccaWillett2RobertD.Nowak2LauraBalzano1Abstractthematrix.Theassumptionthatthematrixislow-rankisequivalenttoassumingthedatalieson(ornea...
AdversarialFeatureMatchingforTextGenerationYizheZhang1ZheGan1KaiFan1ZhiChen1RicardoHenao1DinghanShen1LawrenceCarin1Abstractdomsamplesfromthislatentspace.However,thisapproachoftenfailstogeneratereal...
AdaptiveSamplingProbabilitiesforNon-SmoothOptimizationHongseokNamkoong1AmanSinha2SteveYadlowsky2JohnC.Duchi23Abstractcertainfeaturesareinfrequent—wewishtoleveragethisduringoptimization.Likewise,wh...