Neuro-symbolic Action Masking for Deep Reinforcement Learning

TL;DR

This paper introduces NSAM, a framework that automatically learns symbolic models during deep reinforcement learning to improve sample efficiency and reduce infeasible actions, integrating symbolic reasoning with policy learning.

Contribution

NSAM is a novel approach that automatically learns symbolic models from high-dimensional states during DRL, enhancing action feasibility and learning efficiency.

Findings

NSAM significantly improves sample efficiency in DRL tasks.

NSAM reduces constraint violations during learning.

NSAM outperforms baseline methods in multiple domains.

Abstract

Deep reinforcement learning (DRL) may explore infeasible actions during training and execution. Existing approaches assume a symbol grounding function that maps high-dimensional states to consistent symbolic representations and a manually specified action masking techniques to constrain actions. In this paper, we propose Neuro-symbolic Action Masking (NSAM), a novel framework that automatically learn symbolic models, which are consistent with given domain constraints of high-dimensional states, in a minimally supervised manner during the DRL process. Based on the learned symbolic model of states, NSAM learns action masks that rules out infeasible actions. NSAM enables end-to-end integration of symbolic reasoning and deep policy optimization, where improvements in symbolic grounding and policy learning mutually reinforce each other. We evaluate NSAM on multiple domains with constraints, and experimental results demonstrate that NSAM significantly improves sample efficiency of DRL agent while substantially reducing constraint violations.