Scholarlink

For the purpose of understanding the impact of a Reinforcement Learning (RL) agent’s decisions on the satisfaction of a given arbitrary predicate, we present a method based on the evaluation of the importance of actions. This highlights to the user the most important action(s) (relative to the predicate) in a history of the agent’s interactions with the environment. Having shown that calculating the importance of an action for a predicate to hold is #W[1]-hard, we propose a time-saving approximation. To do so, we use the most likely transitions in the environment. Experiments confirm the relevance of this approach.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predicate-Based Explanation of a Reinforcement Learning Agent via Action Importance Evaluation

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Léo Saulières,
  • Martin C. Cooper,
  • Florence Dupin de Saint-Cyr

摘要

For the purpose of understanding the impact of a Reinforcement Learning (RL) agent’s decisions on the satisfaction of a given arbitrary predicate, we present a method based on the evaluation of the importance of actions. This highlights to the user the most important action(s) (relative to the predicate) in a history of the agent’s interactions with the environment. Having shown that calculating the importance of an action for a predicate to hold is #W[1]-hard, we propose a time-saving approximation. To do so, we use the most likely transitions in the environment. Experiments confirm the relevance of this approach.