PlanningtoExploreviaSelf-SupervisedWorldModelsRamananSekar1OlehRybkin1KostasDaniilidis1PieterAbbeel2DanijarHafner34DeepakPathak56AbstractModelLearningReinforcementlearningallowssolvingcomplexTaskAt...
Self-SupervisedExplorationviaDisagreementDeepakPathak1DhirajGandhi2AbhinavGupta23Abstracttotheagentaresparse.Thecommonapproachtoexplo-rationhasbeentogenerate“intrinsic”rewards,i.e.,rewardsEfficie...
HowdoesDisagreementHelpGeneralizationagainstLabelCorruption?XingruiYu1BoHan2JiangchaoYao3GangNiu2IvorW.Tsang1MasashiSugiyama24Abstractthesecomplexmodelscanfullymemorizenoisylabels(Zhangetal.,2017;A...
ActiveLearningwithDisagreementGraphsCorinnaCortes1GiuliaDeSalvo1ClaudioGentile1MehryarMohri12NingshanZhang3Abstractheinteractivelyselectspointstolabel.Intheon-linesetting,thelearnerreceivesasequenc...