English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Finding the minimal structural assumptions that empower sample-efficientlearning is one of the most important research directions in ReinforcementLearning (RL). This paper advances our understanding of this fundamentalquestion by introducing a new complexity measure -- Bellman Eluder (BE)dimension. We show that the family of RL problems of low BE dimension isremarkably rich, which subsumes a vast majority of existing tractable RLproblems including but not limited to tabular MDPs, linear MDPs, reactivePOMDPs, low Bellman rank problems as well as low Eluder dimension problems.This paper further designs a new optimization-based algorithm -- GOLF, andreanalyzes a hypothesis elimination-based algorithm -- OLIVE (proposed in Jianget al., 2017). We prove that both algorithms learn the near-optimal policies oflow BE dimension problems in a number of samples that is polynomial in allrelevant parameters, but independent of the size of state-action space. Ourregret and sample complexity results match or improve the best existing resultsfor several well-known subclasses of low BE dimension problems.

Bellman Eluder Dimension: New Rich Classes of RL Problems, and Sample-Efficient Algorithms