LogarithmicRegretforReinforcementLearningwithLinearFunctionApproximationJiafanHe1DongruoZhou1QuanquanGu1AbstractAcommonapproachtocopewithhigh-dimensionalstateandactionspacesistoutilizefunctionappro...
LogarithmicRegretforAdversarialOnlineControlDylanJ.Foster1MaxSimchowitz2Abstractbyawell-behavedstochasticprocessordrivenbyaworst-caseprocesstowhichthelearnermustremainrobustinWeintroduceanewalgorit...
LogarithmicRegretforLearningLinearQuadraticRegulatorsEfficientlyAsafCassel1AlonCohen2TomerKoren1Abstract√O(T)regretboundforthissettingalbeitwithacomputa-WeconsidertheproblemoflearninginLin-tionall...
ImprovedBoundsonMinimaxRegretunderLogarithmicLossviaSelf-ConcordanceBlairBilodeau123DylanJ.Foster4DanielM.Roy123AbstractTheloglosspenalizestheplayerbasedonhowmuchprob-abilitymasstheyplaceontheactua...
VariantsofRMSPropandAdagradwithLogarithmicRegretBoundsMaheshChandraMukkamala12MatthiasHein1AbstractThegoalofthispaperistwofold.First,weproposeSC-AdagradwhichisavariantofAdagradadaptedtotheAdaptiveg...
LogarithmicTimeOne-Against-SomeHalDaume´III1NikosKarampatziakis2JohnLangford2PaulMineiro2AbstractxLf000(x)L{2,4,12,...}f0(x)Lf00(x)RRWecreateanewonlinereductionofmulticlassf01(x)f001(x)L{1,15,20,....