Towards Generalizable Reinforcement Learning via Causality-Guided Self-Adaptive Representations

TL;DR

This paper introduces CSR, a causality-guided self-adaptive representation method that enables reinforcement learning agents to generalize across environments with changing dynamics and structures, improving adaptability with minimal samples.

Contribution

The paper presents a novel causality-based approach for RL that characterizes latent causal variables, allowing agents to adapt to diverse environment changes more effectively than existing methods.

Findings

CSR outperforms state-of-the-art baselines in various environments.

It adapts efficiently with only a few samples.

The method generalizes across different tasks with evolving dynamics.

Abstract

General intelligence requires quick adaption across tasks. While existing reinforcement learning (RL) methods have made progress in generalization, they typically assume only distribution changes between source and target domains. In this paper, we explore a wider range of scenarios where not only the distribution but also the environment spaces may change. For example, in the CoinRun environment, we train agents from easy levels and generalize them to difficulty levels where there could be new enemies that have never occurred before. To address this challenging setting, we introduce a causality-guided self-adaptive representation-based approach, called CSR, that equips the agent to generalize effectively across tasks with evolving dynamics. Specifically, we employ causal representation learning to characterize the latent causal variables within the RL system. Such compact causal representations uncover the structural relationships among variables, enabling the agent to autonomously determine whether changes in the environment stem from distribution shifts or variations in space, and to precisely locate these changes. We then devise a three-step strategy to fine-tune the causal model under different scenarios accordingly. Empirical experiments show that CSR efficiently adapts to the target domains with only a few samples and outperforms state-of-the-art baselines on a wide range of scenarios, including our simulated environments, CartPole, CoinRun and Atari games.