End-to-End Neuro-Symbolic Reinforcement Learning with Textual Explanations
This work addresses explainable decision-making in AI, specifically for users needing interpretable policies, though it appears incremental by building on existing neuro-symbolic methods.
The paper tackles the problem of inefficient structured state refinement and high cognitive load in neuro-symbolic reinforcement learning by proposing a framework that distills a vision foundation model into a perception module and uses GPT-4 to generate textual explanations, achieving verification on nine Atari tasks.
Neuro-symbolic reinforcement learning (NS-RL) has emerged as a promising paradigm for explainable decision-making, characterized by the interpretability of symbolic policies. NS-RL entails structured state representations for tasks with visual observations, but previous methods cannot refine the structured states with rewards due to a lack of efficiency. Accessibility also remains an issue, as extensive domain knowledge is required to interpret symbolic policies. In this paper, we present a neuro-symbolic framework for jointly learning structured states and symbolic policies, whose key idea is to distill the vision foundation model into an efficient perception module and refine it during policy learning. Moreover, we design a pipeline to prompt GPT-4 to generate textual explanations for the learned policies and decisions, significantly reducing users' cognitive load to understand the symbolic policies. We verify the efficacy of our approach on nine Atari tasks and present GPT-generated explanations for policies and decisions.