PolicyInformationCapacity:Information-TheoreticMeasureforTaskComplexityinDeepReinforcementLearningHirokiFuruta1TatsuyaMatsushima1TadashiKozuno2YutakaMatsuo1SergeyLevine3OfirNachum3ShixiangShaneGu3A...
LEEP:ANewMeasuretoEvaluateTransferabilityofLearnedRepresentationsCuongV.Nguyen1TalHassner2MatthiasSeeger1CedricArchambeau1Abstractchoosegoodsourcemodelsforagiventargettask(Achilleetal.,2019;Baoetal...