Invariant Causal Prediction for Block MDPs

TL;DR

This paper introduces a causal inference-based method for learning state abstractions in block MDPs that generalize across different environments, improving reinforcement learning robustness.

Contribution

It proposes a novel invariant prediction approach to identify model-irrelevant features, with theoretical guarantees and empirical validation in linear and nonlinear settings.

Findings

Method achieves high-probability identification of causal features.

Provides bounds on model and generalization errors.

Demonstrates improved generalization over baselines.

Abstract

Generalization across environments is critical to the successful application of reinforcement learning algorithms to real-world challenges. In this paper, we consider the problem of learning abstractions that generalize in block MDPs, families of environments with a shared latent state space and dynamics structure over that latent space, but varying observations. We leverage tools from causal inference to propose a method of invariant prediction to learn model-irrelevance state abstractions (MISA) that generalize to novel observations in the multi-environment setting. We prove that for certain classes of environments, this approach outputs with high probability a state abstraction corresponding to the causal feature set with respect to the return. We further provide more general bounds on model error and generalization error in the multi-environment setting, in the process showing a connection between causal variable selection and the state abstraction framework for MDPs. We give empirical evidence that our methods work in both linear and nonlinear settings, attaining improved generalization over single- and multi-task baselines.