TaylorExpansionsofDiscountFactorsYunhaoTang1MarkRowland2Re´miMunos3MichalValko3Abstractexample,TcouldbethefirsttimetheMDPgetsintoatermi-Inpracticalreinforcementlearning(RL),thedis-nalstate(e.g.,ar...
SystematicAnalysisofClusterSimilarityIndices:HowtoValidateValidationMeasuresMartijnGo¨sgens1AlexeyTikhonov2LiudmilaProkhorenkova345Abstracttobedenselyinterconnected.Clusteringisusedacrossvariousap...
StabilityandConvergenceofStochasticGradientClipping:BeyondLipschitzContinuityandSmoothnessVienV.Mai1MikaelJohansson1Abstractproblemsareatthecoreofmanymachine-learningappli-cations,andareoftensolved...
StabilityandGeneralizationofStochasticGradientMethodsforMinimaxProblemsYunwenLei1ZhenhuanYang2TianbaoYang3YimingYing2Abstracting(Goodfellowetal.,2014),robustoptimization(Chenetal.,2017;Namkoong&Duc...
SlotMachines:DiscoveringWinningCombinationsofRandomWeightsinNeuralNetworksMaxwellMbabillaAladago1LorenzoTorresani1AbstractLearningtypicallyinvolveseitheroptimizinganetworkfromscratch(Krizhevskyetal...
SGA:ARobustAlgorithmforPartialRecoveryofTree-StructuredGraphicalModelswithNoisySamplesAnshooTandon1AldricJ.Y.Han2VincentY.F.Tan12Abstractnetworks(Lauritzen,1996)andcomputervision(Besag,1986).Forade...
SECANT:Self-ExpertCloningforZero-ShotGeneralizationofVisualPoliciesLinxiFan12GuanzhiWang1De-AnHuang2ZhidingYu2LiFei-Fei1YukeZhu32AnimaAnandkumar42AbstractTrainingZero-shotEvaluationGeneralizationha...
ScalingPropertiesofDeepResidualNetworksAlain–SamCohen1RamaCont2AlainRossier21RenyuanXu2Abstractwhereh(kL)isthehiddenstateatlayerk=0,...,L,h(0L)=x∈Rdtheinput,h(LL)∈Rdtheoutput,σ:R→Risanon-Resid...
ScalableComputationsofWassersteinBarycenterviaInputConvexNeuralNetworksJiaojiaofan1AmirhosseinTaghvaei2YongxinChen1Abstractthepastfewyears,ithasfoundapplicationsinseveralma-chinelearningproblems.Fo...
ScalableEvaluationofMulti-AgentReinforcementLearningwithMeltingPotJoelZ.Leibo1EdgarDue´n˜ez-Guzma´n1AlexanderSashaVezhnevets1JohnP.Agapiou1PeterSunehag1RaphaelKoster1JaydMatyas1CharlesBeattie1Ig...
Sample-OptimalPACLearningofHalfspaceswithMaliciousNoiseJieShen1AbstractGenerallyspeaking,alargebodyofexistingworksstudytheproblemoflearninghalfspacesunderlabelnoise.Thisin-WestudyefficientPAClearni...
SampleComplexityofRobustLinearClassificationonSeparatedDataRobiBhattacharjee1SomeshJha2KamalikaChaudhuri1Abstractthusaimstofindaclassifierthatmaximizesaccuracyonexamplesthataredistancerormorefromth...
RevealingtheStructureofDeepNeuralNetworksviaConvexDualityTolgaErgen1MertPilanci1Abstractwheretwo-layerReLUnetworkswiththeminimumEu-clideannormsolutionandzerotrainingerrorareprovenWestudyregularized...
Re-understandingFinite-StateRepresentationsofRecurrentPolicyNetworksMohamadH.Danesh1AnuragKoul1AlanFern1SaeedKhorram1Abstracttivehumaninterpretationoftheunderlying“strategicrole"oftheattended-toel...
RepresentationalaspectsofdepthandconditioninginnormalizingflowsFredericKoehler1VirajMehta2AndrejRisteski3Abstract1.IntroductionNormalizingflowsareamongthemostpopularDeepgenerativemodelsareoneofthel...
RepresentationMatters:AssessingtheImportanceofSubgroupAllocationsinTrainingDataEstherRolf1TheodoraWorledge1BenjaminRecht1MichaelI.Jordan12AbstractOurworkaimstodevelopageneralunifyingperspectiveonth...
ReinforcementLearningofImplicitandExplicitControlFlowinInstructionsEthanA.Brooks1JanarthananRajendran1RichardL.Lewis2SatinderSingh1Abstracttaskinstructionsthatrequiretheagenttolearncontrolfloweithe...
Zoo-Tuning:AdaptiveTransferfromaZooofModelsYangShu1ZhiKou1ZhangjieCao1JianminWang1MingshengLong1Abstractleveragemodelspretrainedonlarge-scaledatasets(Rus-sakovskyetal.,2015)andfine-tunethemodelonth...
WILDS:ABenchmarkofin-the-WildDistributionShiftsPangWeiKoh1ShioriSagawa1HenrikMarklund1SangMichaelXie1MarvinZhang2AkshayBalsubramani1WeihuaHu1MichihiroYasunaga1RichardLanasPhillips3IrenaGao1TonyLee1...
WhenAllWeNeedisaPieceofthePie:AGenericFrameworkforOptimizingTwo-wayPartialAUCZhiyongYang12QianqianXu3ShilongBao12YuanHe4XiaochunCao12QingmingHuang3567Abstractframework.TheAreaUndertheROCCurve(AUC)i...