Pessimistic Model Selection for Offline Deep Reinforcement Learning

TL;DR

This paper introduces a pessimistic model selection method for offline deep reinforcement learning that provides theoretical guarantees and improves policy generalization in real-world applications.

Contribution

It proposes a novel pessimistic model selection framework with theoretical backing, addressing the challenge of model overfitting in offline DRL.

Findings

Outperforms existing model selection methods in numerical studies

Provides theoretical guarantees for policy optimality

Addresses bias in DRL model identification

Abstract

Deep Reinforcement Learning (DRL) has demonstrated great potentials in solving sequential decision making problems in many applications. Despite its promising performance, practical gaps exist when deploying DRL in real-world scenarios. One main barrier is the over-fitting issue that leads to poor generalizability of the policy learned by DRL. In particular, for offline DRL with observational data, model selection is a challenging task as there is no ground truth available for performance demonstration, in contrast with the online setting with simulated environments. In this work, we propose a pessimistic model selection (PMS) approach for offline DRL with a theoretical guarantee, which features a provably effective framework for finding the best policy among a set of candidate models. Two refined approaches are also proposed to address the potential bias of DRL model in identifying the optimal policy. Numerical studies demonstrated the superior performance of our approach over existing methods.