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

State-of-the-art efficient model-based Reinforcement Learning (RL) algorithmstypically act by iteratively solving empirical models, i.e., by performing\emph{full-planning} on Markov Decision Processes (MDPs) built by the gatheredexperience. In this paper, we focus on model-based RL in the finite-statefinite-horizon MDP setting and establish that exploring with \emph{greedypolicies} -- act by \emph{1-step planning} -- can achieve tight minimaxperformance in terms of regret, $\tilde{\mathcal{O}}(\sqrt{HSAT})$. Thus,full-planning in model-based RL can be avoided altogether without anyperformance degradation, and, by doing so, the computational complexitydecreases by a factor of $S$. The results are based on a novel analysis ofreal-time dynamic programming, then extended to model-based RL. Specifically,we generalize existing algorithms that perform full-planning to such that actby 1-step planning. For these generalizations, we prove regret bounds with thesame rate as their full-planning counterparts.

Tight Regret Bounds for Model-Based Reinforcement Learning with Greedy Policies