UCBMomentumQ-learning:CorrectingtheBiaswithoutforgettingPierreMénard1OmarDarwicheDomingues2XuedongShang23MichalValko234Abstractbalancetheexplorationoftheenvironmentandexploitationofthecurrentknowl...
TheImplicitRegularizationforAdaptiveOptimizationAlgorithmsonHomogeneousNeuralNetworksBohanWang1QiMeng1WeiChen1Tie-YanLiu1Abstractprocessing(Youngetal.,2018).Inpractice,deepneuralnetworks(DNN)learne...
QuantifyingandReducingBiasinMaximumLikelihoodEstimationofStructuredAnomaliesUthsavChitra1KimberlyDing1JasperC.H.Lee2BenjaminJ.Raphael1Abstract1.IntroductionAnomalyestimation,ortheproblemoffindingAn...
OntheImplicitBiasofInitializationShape:BeyondInfinitesimalMirrorDescentShaharAzulay1EdwardMoroshko2MorShpigelNacson2BlakeWoodworth3NathanSrebro3AmirGloberson1DanielSoudry2Abstractparameterizedmodel...
LIME:LearningInductiveBiasforPrimitivesofMathematicalReasoningYuhuaiWu12MarkusRabe3WendaLi4JimmyBa12RogerGrosse12ChristianSzegedy3Abstractmathematicalreasoning.Attemptstodesignelaboratear-chitectur...
FairnessandBiasinOnlineSelectionJose´Correa1Andre´sCristi1PaulDu¨tting2AshkanNorouzi-Fard2Abstractalsoitnaturallyencompassesmanyreal-worlddecisionmak-ingprocesseswhereBiasedevaluationsshouldbeav...
DisentanglingLabelingandSamplingBiasforLearninginLarge-outputSpacesAnkitSinghRawat1AdityaKrishnaMenon1WittawatJitkrittum1SadeepJayasumana1FelixX.Yu1SashankJ.Reddi1SanjivKumar1Abstractcloselyascheme...
DirectionalBiasAmplificationAngelinaWangandOlgaRussakovskyPrincetonUniversityAbstractcorrespondingly,aplethoraofnewalgorithmsandmetricsarebeingproposed(seee.g.,Mehrabietal.(2019)forasur-Mitigatingb...
CorrectingExposureBiasforLinkRecommendationShantanuGupta12HaoWang3ZacharyLipton2YuyangWang4AbstractconcernsaboutexposureBias—manymissinglinksarefalsenegatives,anddidnotformduetolackofexposurerathe...
AsymmetricHeavyTailsandImplicitBiasinGaussianNoiseInjectionsAlexanderCamuto1XiaoyuWang2LingjiongZhu2MertGu¨rbu¨zbalaban3ChrisHolmes1UmutS¸ims¸ekli4AbstractFigure1.IllustrationoftheeffectofGNIsa...
IdentifyingStatisticalBiasinDatasetReplicationLoganEngstrom1AndrewIlyas1ShibaniSanturkar1DimitrisTsipras1JacobSteinhardt2AleksanderMa˛dry1Abstractaprocessknownasdatasetreplication.Thoughnotyetwide...
OnconditionalversusmarginalBiasinmulti-armedbanditsJaehyeokShin1AlessandroRinaldo1AadityaRamdas12AbstractThedataarecollectedsequentiallyinstages,duringwhichtheanalystdrawsasamplefromoneamongfinitel...
LowBiasLowVarianceGradientEstimatesforHierarchicalBooleanStochasticNetworksAdeelPervez1TacoCohen2EfstratiosGavves1Abstractgradients,onecansamplegradientsfromsomedistribu-tion.Thesampleestimatescanb...
FrequencyBiasinNeuralNetworksforInputofNon-UniformDensityRonenBasri1MeiravGalun1AmnonGeifman1DavidJacobs2YoniKasten1ShiraKritchman1AbstractEpoch=3Epoch=100Recentworkshavepartlyattributedthegenerali...
ExplorationThroughRewardBiasing:Reward-BiasedMaximumLikelihoodEstimationforStochasticMulti-ArmedBanditsXiLiu1Ping-ChunHsieh2Yu-HengHung2AnirbanBhattacharya3P.R.Kumar1Abstractandthenappliestheaction...
DoubleTroubleinDoubleDescent:BiasandVariance(s)intheLazyRegimeSte´phaned’Ascoli1MariaRefinetti1GiulioBiroli1FlorentKrzakala1Abstractrecognition(Hintonetal.,2012),andautomatictransla-tion(Sutskeve...
DatapreprocessingtomitigateBias:AmaximumentropybasedapproachL.ElisaCelis1VijayKeswani1NisheethK.Vishnoi2AbstracttodeBiasdatastrivetoensurethateither1)therepresenta-tionofsalientsocialgroupsinthedat...
ControllingOverestimationBiaswithTruncatedMixtureofContinuousDistributionalQuantileCriticsArseniiKuznetsov1PavelShvechikov12AlexanderGrishin13DmitryVetrov13AbstractThrun&Schwartz(1993)elucidatetheo...
ConQUR:MitigatingDelusionalBiasinDeepQ-learningDiJia(Andy)Su12JaydenOoi1TylerLu1DaleSchuurmans13CraigBoutilier1Abstract&Smart,2004;Melo&Ribeiro,2007;Maeietal.,2010;Munosetal.,2016);butitremainsdif...