Linear Bellman Completeness Suffices for Efficient Online Reinforcement Learning with Few Actions.

Noah Golowich, Ankur Moitra

Browse the full COLT paper archive.