BECAUSE: Bilinear Causal Representation for Generalizable Offline Model-based Reinforcement Learning

TL;DR

BECAUSE introduces a causal representation approach in offline model-based reinforcement learning to address distribution shift and objective mismatch, leading to improved robustness and sample efficiency across diverse tasks.

Contribution

The paper proposes BECAUSE, a novel causal representation method that reduces objective mismatch in offline MBRL, with theoretical guarantees and superior empirical performance.

Findings

BECAUSE outperforms existing offline RL algorithms on 18 diverse tasks.

It demonstrates robustness with fewer samples and more confounders.

Theoretical analysis confirms error bounds and sample efficiency.

Abstract

Offline model-based reinforcement learning (MBRL) enhances data efficiency by utilizing pre-collected datasets to learn models and policies, especially in scenarios where exploration is costly or infeasible. Nevertheless, its performance often suffers from the objective mismatch between model and policy learning, resulting in inferior performance despite accurate model predictions. This paper first identifies the primary source of this mismatch comes from the underlying confounders present in offline data for MBRL. Subsequently, we introduce \textbf{B}ilin\textbf{E}ar \textbf{CAUS}al r\textbf{E}presentation~(BECAUSE), an algorithm to capture causal representation for both states and actions to reduce the influence of the distribution shift, thus mitigating the objective mismatch problem. Comprehensive evaluations on 18 tasks that vary in data quality and environment context demonstrate the superior performance of BECAUSE over existing offline RL algorithms. We show the generalizability and robustness of BECAUSE under fewer samples or larger numbers of confounders. Additionally, we offer theoretical analysis of BECAUSE to prove its error bound and sample efficiency when integrating causal representation into offline MBRL.