UniformConvergence,AdversarialSpheresandaSimpleRemedyGregorBachmann1Seyed-MohsenMoosavi-Dezfooli1ThomasHofmann1AbstractofuniformConvergenceasatooltoboundthecapacityofthemodel(Aroraetal.,2018;Bartle...
TheLimitsofMin-MaxOptimizationAlgorithms:ConvergencetoSpuriousNon-CriticalSetsYa-PingHsieh1PanayotisMertikopoulos23VolkanCevher4AbstractGivenanalgorithmforsolving(SP),itisthennaturaltoComparedtoord...
StabilityandConvergenceofStochasticGradientClipping:BeyondLipschitzContinuityandSmoothnessVienV.Mai1MikaelJohansson1Abstractproblemsareatthecoreofmanymachine-learningappli-cations,andareoftensolved...
OntheConvergenceofHamiltonianMonteCarlowithStochasticGradientsDifanZou1QuanquanGu1AbstracttionssuchasBayesianinference,reinforcementlearning,andcomputervision.Inthepastdecades,manyMCMCHamiltonianMo...
OnCharacterizingGANConvergenceThroughProximalDualityGapSahilSidheekh1AroofAimen1NarayananC.Krishnan1Abstractmostpopularandsuccessfulofinnovationsforlearningdatadistributions.AGANformulatesthegenera...
Model-TargetedPoisoningAttackswithProvableConvergenceFnuSuya1SaeedMahloujifar2AnshumanSuri1DavidEvans1YuanTian1AbstractMostworkonpoisoningattackshasconsideredoneoftwoextremalattackerobjectives:indi...
OntheProofofGlobalConvergenceofGradientDescentforDeepReLUNetworkswithLinearWidthsQuynhNguyen1Abstracttrainingdata,thentheoutputatlayerlisgivenbyWegiveasimpleprooffortheglobalconver-genceofgradien...
LearningWhilePlayinginMean-FieldGames:ConvergenceandOptimalityQiaominXie1ZhuoranYang2ZhaoranWang3AndreeaMinca1Abstractfromthescalabilityissue.Specifically,inamulti-agentsystem,eachagentinteractswit...
GlobalConvergenceofPolicyGradientforLinear-QuadraticMean-FieldControl/GameinContinuousTimeWeichenWang1JiequnHan2ZhuoranYang3ZhaoranWang4Abstractmorerealisticreal-worldproblems,suchasroboticcontrol(...
GeometricConvergenceofEllipticalSliceSamplingViacheslavNatarovskii1DanielRudolf1Bjo¨rnSprungk2AbstractonRdtakestheformForBayesianlearning,givenlikelihoodfunction(x)andGaussianprior,theellipticalsl...
FunctionalSpaceAnalysisofLocalGANConvergenceValentinKhrulkov1ArtemBabenko12IvanOseledets3Abstractousgradientdescent.InNagarajan&Kolter(2017)itwasshownthatundermildassumptions,thismethodislocallyRec...
FromPoincare´RecurrencetoConvergenceinImperfectInformationGames:FindingEquilibriumviaRegularizationJulienPerolat1RemiMunos1Jean-BaptisteLespiau1ShayeganOmidshafiei1MarkRowland1PedroOrtega1NeilBurc...
DoublyRobustOff-PolicyActor-Critic:ConvergenceandOptimalityTengyuXu1ZhuoranYang2ZhaoranWang3YingbinLiang1Abstract(Haarnojaetal.,2018),etc.However,thesesuccessesusu-allyrelyontheaccesstoon-policysam...
BilevelOptimization:ConvergenceAnalysisandEnhancedDesignKaiyiJi1JunjieYang1YingbinLiang1Abstractoptimization(Franceschietal.,2018;Shabanetal.,2019;Feurer&Hutter,2019),reinforcementlearning(Konda&Bi...
UniformConvergenceofRank-weightedLearningJustinKhim1LiuLeqi1AdarshPrasad1PradeepRavikumar1Abstractlearning,wherepercentilebasedriskmeasureshavebeenusedtoquantifythetail-riskofmodels.ArecentlineofTh...
TrainingLinearNeuralNetworks:Non-LocalConvergenceandComplexityResultsArminEftekhari1AbstractbythelinearmapLinearnetworksprovidevaluableinsightsintoRdx→Rdytheworkingsofneuralnetworksingeneral.Thisp...
OntheGlobalConvergenceRatesofSoftmaxPolicyGradientMethodsJinchengMei12ChenjunXiao1CsabaSzepesva´ri31DaleSchuurmans21Abstracttheyguaranteemonotonicimprovementofthevalue.Asec-ondaryappealisthatpolic...
OntheConvergenceofNesterov’sAcceleratedGradientMethodinStochasticSettingsMahmoudAssran123MichaelRabbat23AbstractHowever,thetheoreticalunderstandingofacceleratedmeth-odsremainslimitedwhenusedwithst...
Min-MaxOptimizationwithoutGradients:ConvergenceandApplicationstoBlack-BoxEvasionandPoisoningAttacksSijiaLiu1SongtaoLu2XiangyiChen3YaoFeng4KaidiXu5AbdullahAl-Dujaili6MingyiHong3Una-MayO’Reilly71Abs...
LinearConvergenceofRandomizedPrimal-DualCoordinateMethodforLarge-scaleLinearConstrainedConvexProgrammingDaoliZhu1LeiZhao2Abstractadditivewithrespecttothefollowingspacedecomposition,Linearconstraine...