Saved in:
Bibliographic Details
Main Authors: Lee, Harin, Oh, Min-hwan
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2503.00810
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912503515054080
author Lee, Harin
Oh, Min-hwan
author_facet Lee, Harin
Oh, Min-hwan
contents In our quest for a reinforcement learning (RL) algorithm that is both practical and provably optimal, we introduce EQO (Exploration via Quasi-Optimism). Unlike existing minimax optimal approaches, EQO avoids reliance on empirical variances and employs a simple bonus term proportional to the inverse of the state-action visit count. Central to EQO is the concept of quasi-optimism, where estimated values need not be fully optimistic, allowing for a simpler yet effective exploration strategy. The algorithm achieves the sharpest known regret bound for tabular RL under the mildest assumptions, proving that fast convergence can be attained with a practical and computationally efficient approach. Empirical evaluations demonstrate that EQO consistently outperforms existing algorithms in both regret performance and computational efficiency, providing the best of both theoretical soundness and practical effectiveness.
format Preprint
id arxiv_https___arxiv_org_abs_2503_00810
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Minimax Optimal Reinforcement Learning with Quasi-Optimism
Lee, Harin
Oh, Min-hwan
Machine Learning
In our quest for a reinforcement learning (RL) algorithm that is both practical and provably optimal, we introduce EQO (Exploration via Quasi-Optimism). Unlike existing minimax optimal approaches, EQO avoids reliance on empirical variances and employs a simple bonus term proportional to the inverse of the state-action visit count. Central to EQO is the concept of quasi-optimism, where estimated values need not be fully optimistic, allowing for a simpler yet effective exploration strategy. The algorithm achieves the sharpest known regret bound for tabular RL under the mildest assumptions, proving that fast convergence can be attained with a practical and computationally efficient approach. Empirical evaluations demonstrate that EQO consistently outperforms existing algorithms in both regret performance and computational efficiency, providing the best of both theoretical soundness and practical effectiveness.
title Minimax Optimal Reinforcement Learning with Quasi-Optimism
topic Machine Learning
url https://arxiv.org/abs/2503.00810