SparseFeatureSelectionMakesBatchReinforcementLearningMoreSampleEfficientBotaoHao1YaqiDuan2TorLattimore1CsabaSzepesva´ri13MengdiWang21Abstract1.IntroductionThispaperprovidesastatisticalanalysisofhi...
Run-Sort-ReRun:EscapingBatchSizeLimitationsinSlicedWassersteinGenerativeModelsJose´Lezama1WeiChen2QiangQiu2Abstract2017;Lietal.,2017;Mrouehetal.,2017;Heuseletal.,2017;Deshpandeetal.,2018).However,...
RiskBoundsandRademacherComplexityinBatchReinforcementLearningYaqiDuan1ChiJin2ZhiyuanLi3Abstractalgorithmsincludingsupportvectormachines(Cortes&Vapnik,1995;Suykens&Vandewalle,1999),boosting(Fre-This...
OntheOptimalityofBatchPolicyOptimizationAlgorithmsChenjunXiao12YifanWu3TorLittlemore4BoDai2JinchengMei12LihongLi†5CsabaSzepesvari14DaleSchuurmans12Abstractafixeddatasetofpreviouslycollectedexperie...
ExponentialLowerBoundsforBatchReinforcementLearning:BatchRLcanbeExponentiallyHarderthanOnlineRLAndreaZanette1AbstractweconsidertwoclassicalBatchRLproblems:1)theoff-policyevaluation(OPE)problem,wher...
BatchValue-functionApproximationwithOnlyRealizabilityTengyangXie1NanJiang1Abstractthissubproblem,wecreateapiecewiseconstantfunctionclassofstatisticalcomplexityO(1/2)thatcanexpressbothWemakeprogress...
TASKNORM:RethinkingBatchNormalizationforMeta-LearningJohnBronskill1JonathanGordon1JamesRequeima12SebastianNowozin3RichardE.Turner13Abstractthe-artperformanceinarangeofbenchmarktasks(Finnetal.,2017;...
ReducingSamplingErrorinBatchTemporalDifferenceLearningBrahmaS.Pavse1IshanDurugkar1JosiahP.Hanna23PeterStone14Abstractpolicy(Puterman&Shin,1978;Bertsekas,1987;Konda&Tsitsiklis,2000).Thesealgorithmsr...
PowerNorm:RethinkingBatchNormalizationinTransformersShengShen1ZheweiYao1AmirGholami1MichaelW.Mahoney1KurtKeutzer1Abstract1.IntroductionThestandardnormalizationmethodforneuralNormalizationhasbecomeo...
History-GradientAidedBatchSizeAdaptationforVarianceReducedAlgorithmsKaiyiJi1ZheWang1BowenWeng1YiZhou2WeiZhang3YingbinLiang1AbstracthavebeenproposedtoreducethevarianceofSGD.Suchvariancereductiontech...
BatchStationaryDistributionEstimationJunfengWen1BoDai2LihongLi2DaleSchuurmans12Abstractunderlyingprocess.Nevertheless,onewouldstillliketoestimatetargetpropertiesofthestationarydistribution,suchWeco...
BatchReinforcementLearningwithHyperparameterGradientsByung-JunLee1JongminLee1PeterVrancx2DonghoKim2Kee-EungKim23Abstractrealenvironment.However,thisapproachrequiresalotofhumaneffortincludingdomaine...
Transferabilityvs.Discriminability:BatchSpectralPenalizationforAdversarialDomainAdaptationXinyangChen1SinanWang1MingshengLong1JianminWang1Abstractscaledatasetscanbeleveraged(Torralba&Efros,2011).Ad...
QuantileSteinVariationalGradientDescentforBatchBayesianOptimizationChengyueGong1JianPeng2QiangLiu1AbstractorexpensiveobjectivefunctionasarandomvariableandleverageBayesianinference,typicallywithaGau...
Information-TheoreticConsiderationsinBatchReinforcementLearningJinglinChen1NanJiang1AbstractwhentheyworkiscentraltoourunderstandingofRL.Ex-istingworksthatanalyzeerrorpropagationandfinitesam-Value-f...
BatchPolicyLearningunderConstraintsHoangM.Le1CameronVoloshin1YisongYue1Abstractdeed,manysuchreal-worldapplicationsrequiretheprimaryobjectivefunctionbeaugmentedwithanappropriatesetofWhenlearningpoli...
AsynchronousBatchBayesianOptimisationwithImprovedLocalPenalisationAhsanS.Alvi12BinxinRu1JanCalliess13StephenJ.Roberts123MichaelA.Osborne12Abstractexample,considertheoptimisationofthenumberofunitsin...
AQuantitativeAnalysisoftheEffectofBatchNormalizationonGradientDescentYongqiangCai1QianxiaoLi12ZuoweiShen1AbstracteffectsofBNareattributedtotheso-called“reductionofcovariateshift”.However,itisuncl...
FastVarianceReductionMethodwithStochasticBatchSizeXuanqingLiu1Cho-JuiHsieh12AbstractHereweassumeeachfi(w)isaµ-stronglyconvex,L-smoothfunction,theregularizationtermg(w)isconvexInthispaperwestudyafa...
BayesianUncertaintyEstimationforBatchNormalizedDeepNetworksMattiasTeye12HosseinAzizpour1KevinSmith13Abstractorifthenetworkisconfrontedwithadversarialexamples(Goodfellowetal.,2014).Whenexposedtodata...