Fine-Grained Causal Dynamics Learning with Quantization for Improving Robustness in Reinforcement Learning

TL;DR

This paper introduces a novel causal dynamics learning model with quantization that enhances robustness in reinforcement learning by capturing fine-grained causal structures and context-specific dependencies.

Contribution

It proposes a joint learning approach combining a discrete latent variable with causal structure inference to improve robustness and interpretability in RL dynamics modeling.

Findings

Improved robustness to unseen states and spurious correlations

Effective discovery of fine-grained causal relationships

Superior performance over prior methods in causal reasoning

Abstract

Causal dynamics learning has recently emerged as a promising approach to enhancing robustness in reinforcement learning (RL). Typically, the goal is to build a dynamics model that makes predictions based on the causal relationships among the entities. Despite the fact that causal connections often manifest only under certain contexts, existing approaches overlook such fine-grained relationships and lack a detailed understanding of the dynamics. In this work, we propose a novel dynamics model that infers fine-grained causal structures and employs them for prediction, leading to improved robustness in RL. The key idea is to jointly learn the dynamics model with a discrete latent variable that quantizes the state-action space into subgroups. This leads to recognizing meaningful context that displays sparse dependencies, where causal structures are learned for each subgroup throughout the training. Experimental results demonstrate the robustness of our method to unseen states and locally spurious correlations in downstream tasks where fine-grained causal reasoning is crucial. We further illustrate the effectiveness of our subgroup-based approach with quantization in discovering fine-grained causal relationships compared to prior methods.