Bayesian Off-Policy Evaluation and Learning for Large Action Spaces

TL;DR

This paper introduces a Bayesian framework for off-policy evaluation and learning in large action spaces, effectively capturing action correlations to improve sample efficiency and performance.

Contribution

It proposes a unified Bayesian approach, sDM, that leverages action correlations without sacrificing computational efficiency, and introduces Bayesian metrics for algorithm assessment.

Findings

sDM outperforms existing methods in empirical tests

Leveraging action correlations improves evaluation accuracy

Bayesian metrics provide better performance insights

Abstract

In interactive systems, actions are often correlated, presenting an opportunity for more sample-efficient off-policy evaluation (OPE) and learning (OPL) in large action spaces. We introduce a unified Bayesian framework to capture these correlations through structured and informative priors. In this framework, we propose sDM, a generic Bayesian approach for OPE and OPL, grounded in both algorithmic and theoretical foundations. Notably, sDM leverages action correlations without compromising computational efficiency. Moreover, inspired by online Bayesian bandits, we introduce Bayesian metrics that assess the average performance of algorithms across multiple problem instances, deviating from the conventional worst-case assessments. We analyze sDM in OPE and OPL, highlighting the benefits of leveraging action correlations. Empirical evidence showcases the strong performance of sDM.