Instance-Dependent Complexity of Contextual Bandits and Reinforcement Learning: A Disagreement-Based Perspective

Submitted by xiar@mit.edu on

TitleInstance-Dependent Complexity of Contextual Bandits and Reinforcement Learning: A Disagreement-Based Perspective
Publication TypeJournal Article
Year of Publication2021
AuthorsFoster D.J, Rakhlin A., Simchi-Levi D., Xu Y.
JournalTo appear in COLT 2021

Type: 

Theory