arXiv cs.AI· L\'eo Sauli\`eres·· 12 小时前AI 评分31
可解释强化学习综述：目标、方法与需求
A Survey of Explainable Reinforcement Learning: Targets, Methods and Needs
AI 导读
一篇 69 页综述提出基于“What”与“How”两个问题的可解释强化学习（XRL）分类法，前者关注方法解释的目标，后者关注解释的提供方式，并据此回顾了超过 250 篇论文。文章还梳理了与 XRL 相近、值得社区关注的领域，并指出该领域尚待满足的需求。
正文 · 原文
Abstract:The success of recent Artificial Intelligence (AI) models has been accompanied by the opacity of their internal mechanisms, due notably to the use of deep neural networks. In order to understand these internal mechanisms and explain the output of these AI models, a set of methods have been proposed, grouped under the domain of eXplainable AI (XAI). This paper focuses on a sub-domain of XAI, called eXplainable Reinforcement Learning (XRL), which aims to explain the actions of an agent that has learned by reinforcement learning. We propose an intuitive taxonomy based on two questions "What" and "How". The first question focuses on the target that the method explains, while the second relates to the way the explanation is provided. We use this taxonomy to provide a state-of-the-art review of over 250 papers. In addition, we present a set of domains close to XRL, which we believe should get attention from the community. Finally, we identify some needs for the field of XRL.
| Comments: | 69 pages, 19 figures |
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2507.12599 [cs.AI] |
| (or arXiv:2507.12599v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2507.12599 arXiv-issued DOI via DataCite |
|
| Journal reference: | Journal of Artificial Intelligence Research 87, Article 16 (October 2026) |
| Related DOI: | https://doi.org/10.1613/jair.1.20550
DOI(s) linking to related resources |
Submission history
From: Léo Saulières [view email]
[v1]
Wed, 16 Jul 2025 19:41:41 UTC (8,336 KB)
[v2]
Thu, 15 Oct 2026 14:48:40 UTC (5,024 KB)
来源：arXiv cs.AI · arxiv.org