FastAlgorithmsforStackelbergPredictionGamewithLeastSquaresLossJialiWang1HeChen2RujunJiang1XudongLi1ZihaoLi2Abstractcertainareas,suchascybersecurity,thenatureofapplica-tionsrequireshighrobustnessofM...
EmphaticAlgorithmsforDeepReinforcementLearningRayJiang1TomZahavy1ZhongwenXu1AdamWhite12MatteoHessel1CharlesBlundell1HadovanHasselt1AbstractManyreinforcementlearning(RL)agentslearnoff-policytosomeex...
Benchmarks,Algorithms,andMetricsforHierarchicalDisentanglementAndrewSlavinRoss1FinaleDoshi-Velez1Abstractable(Locatelloetal.,2018),anditseemsunlikelythatcon-tinuous,factorized,fixed-dimensionalrepr...
AcceleratedAlgorithmsforSmoothConvex-ConcaveMinimaxProblemswithO(1/k2)RateonSquaredGradientNormTaeHoYoon1ErnestK.Ryu1Abstractvalueforquantifyingsuboptimality.Moreover,thenotionismeaningfulfordiffer...
ThompsonSamplingAlgorithmsforMean-VarianceBanditsQiuyuZhu1VincentY.F.Tan123AbstractTheprimaryconcernofthisbodyofliteratureistofindalearningalgorithmwhichcanmaximizetheexpectedcu-Themulti-armedbandi...
StructureAdaptiveAlgorithmsforStochasticBanditsRe´myDegenne1HanShao2WouterM.Koolen3Abstractstartingwithasymptoticresultsinthe80sand90s(Lai&Robbins,1985;Graves&Lai,1997)andmovingtothefi-Westudyrewa...
StochasticGauss-NewtonAlgorithmsforNonconvexCompositionalOptimizationQuocTran-Dinh1NhanH.Pham1LamM.Nguyen2Abstractandnpi=1,thenbyintroductingFi(x):=i=1WedeveloptwonewstochasticGauss-Newtonalgorithm...
ProvableSelf-PlayAlgorithmsforCompetitiveReinforcementLearningYuBai1ChiJin2Abstractconflictingrewards(sothattheyessentiallycompetewitheachother)yetcanbetrainedinacentralizedfashion(i.e.Self-play,wh...
OnlineMetricAlgorithmswithUntrustedPredictionsAntoniosAntoniadis1ChristianCoester2asdatacenters(Iranietal.,2003;Linetal.,2013),andareMarekElia´sˇ3AdamPolak4BertrandSimon5alsorelatedtotheexpertspr...
OnThompsonSamplingwithLangevinAlgorithmsEricMazumdar1AldoPacchiano1Yi-AnMa23PeterL.Bartlett14MichaelI.Jordan14Abstractexploitationtradeoffs(Aueretal.,2002;LattimoreandSzepesva´ri,2020),whereinanal...
NewOracle-EfficientAlgorithmsforPrivateSyntheticDataReleaseGiuseppeVietri1GraceTian2MarkBun3ThomasSteinke4StevenWu1AbstractmensiondandadatasetD∈XnconsistingofthedataWepresentthreenewAlgorithmsforc...
Multi-stepGreedyReinforcementLearningAlgorithmsMananTomar1YonathanEfroni2MohammadGhavamzadeh3Abstractestimations(Greensmithetal.,2004)andtohavedifficultiesinhandlingfunctionapproximation(e.g.,Thrun...
ImprovedOptimisticAlgorithmsforLogisticBanditsLouisFaury12MarcAbeille1Cle´mentCalauze`nes1OlivierFercoq2Abstractetal.(2017)andreferencestherein),itspracticalinterestislimitedbythelinearstructureof...
EvaluatingthePerformanceofReinforcementLearningAlgorithmsScottM.Jordan1YashChandak1DanielCohen1MengxueZhang1PhilipS.Thomas1AbstractusabilityofRLAlgorithms,wesuggestthatitshouldhavefourproperties.Fi...
AutoML-Zero:EvolvingMachineLearningAlgorithmsFromScratchEstebanReal1ChenLiang1DavidR.So1QuocV.Le1Abstractfield,rangingfromlearningstrategiestonewarchitectures[Rumelhartetal.,1986;LeCunetal.,1995;Ho...
CustomizingMLPredictionsForOnlineAlgorithmsKeertiAnand1RongGe1DebmalyaPanigrahi1AbstractThekeytothisquestionistheobservationthatunlikeinagenericlearningsetting,wearenotinterestedintraditionalApopul...
Continuous-timeLowerBoundsforGradient-basedAlgorithmsMichaelMuehlebach1MichaelI.Jordan1Abstractindependentlowerboundsthenresultfromanunboundedincreaseintheproblemdimension.Thisestablishes,forThisar...
ApproximationGuaranteesofLocalSearchAlgorithmsviaLocalizabilityofSetFunctionsKaitoFujii1Abstractproblemoffindingan(approximately)optimalsetfromallfeasiblesets.VariousmachinelearningtaskshavebeenThi...
AnewregretanalysisforAdam-typeAlgorithmsAhmetAlacaoglu1YuraMalitsky1PanayotisMertikopoulos23VolkanCevher1AbstractOnecanwonderwhetherthereisaninherentobstacle–intheproposedmethodsorthesetting–whic...
SubmodularMaximizationbeyondNon-negativity:Guarantees,FastAlgorithms,andApplicationsChristopherHarshaw1MoranFeldman2JustinWard3AminKarbasi4Abstract1.IntroductionItisgenerallybelievedthatsubmodularf...