TowardsBetterRobustGeneralizationwithShiftConsistencyRegularizationShufeiZhang12ZhuangQian12KaizhuHuang1QiufengWang1RuiZhang3XinpingYi2Abstract2015;Heetal.,2016;Miyatoetal.,2017;Zagoruyko&Komodakis...
Near-OptimalLinearRegressionunderDistributionShiftQiLei1WeiHu1JasonD.Lee1Abstractprovidehelpfulmarginaldistributioninformationaboutthetargetdomain.AlthoughthisapproachisintegraltomanyTransferlearni...
MANDOLINE:ModelEvaluationunderDistributionShiftMayeeChen1KaranGoel1NimitSohoni2FaitPoms1KayvonFatahalian1ChristopherRe´1Abstracttionerstodetermineiftheirmodelswillperformwellwhendeployed.Unfortuna...
ATheoryofLabelPropagationforSubpopulationShiftTianleCai12RuiqiGao12JasonD.Lee1QiLei1Abstracttees.ManyoftheworksarebasedonthecovariateShiftas-sumption(Heckman,1979;Shimodaira,2000),whichstatesOneoft...
TheEffectofNaturalDistributionShiftonQuestionAnsweringModelsJohnMiller1KarlKrauth1BenjaminRecht1LudwigSchmidt1AbstractmerelytoobtainhighscoresontheSQuADleaderboard,butrathertogeneralizetonewexample...
RobustandStableBlackBoxExplanationsHimabinduLakkaraju1NinoArsov2OsbertBastani3Abstractanddecidewhenandhowtoemploythem.However,theproprietarynatureandincreasingcomplexityofmachineAsmachinelearningbl...
LTF:ALabelTransformationFrameworkforCorrectingTargetShiftJiaxianGuo1MingmingGong2TongliangLiu1KunZhang3DachengTao1Abstract1.IntroductionDistributionShiftisamajorobstacletothede-Standardsupervisedle...
DetectingandCorrectingforLabelShiftwithBlackBoxPredictorsZacharyC.Lipton12Yu-XiangWang23AlexanderJ.Smola2Abstractiidassumption.Absentfamiliarguarantees,wewonder:Isfstillaccurate?What’stherealcurre...