Adaptive reward-free exploration

Émilie Kaufmann, Pierre Ménard, Omar Darwiche Domingues, Anders Jonsson, Édouard Leurent, Michal Valko

2021 · in (ALT 2021) and (ICML 2020 - RL Theory)

Abstract

Reward-free exploration is a reinforcement learning setting recently studied by Jin et al., who address it by running several algorithms with regret guarantees in parallel. In our work, we instead propose a more adaptive approach for reward-free exploration which directly reduces upper bounds on the maximum MDP estimation error. We show that, interestingly, our reward-free UCRL algorithm can be seen as a variant of an algorithm of Fiechter from 1994, originally proposed for a different objective that we call best-policy identification. We prove that RF-UCRL needs O((SAH4/$\epsilon$2)ln(1/$\delta$)) episodes to output, with probability 1−$\delta$, an $\epsilon$-approximation of the optimal policy for any reward function. We empirically compare it to oracle strategies using a generative model.

PDF · arXiv preprint · video 1 · video 2 · Algorithmic Learning Theory · bibtex · DOI