发表机构
North Carolina State University; Indian Institute of Technology Bombay; Tata Research Development and Design Centre; Northeastern University（北卡罗来纳州立大学; 印度理工学院孟买分校; 塔塔研究开发与设计中心; 东北大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文提出一种类似人类认知的策略，将强化学习代理的游戏策略综合为可执行程序，以学习可解释的表示，并在国际象棋和网格环境中验证其有效性。
AI 中文摘要
作为学习复杂游戏的一部分，人类玩家会发展出与游戏规则一致的抽象概念和策略，以提高他们的表现。这些概念被用于解释其他玩家的行为，并指导他们在游戏中的行动。理解其他玩家的策略是这种提升的关键部分，但需要时间和精力。在本文中，我们提出了一种类似于人类认知的策略，用于训练强化学习代理将学习到的策略和政策综合为基于游戏动作序列的可执行程序。我们提出了自动学习这些程序的方法，以玩国际象棋并解决基于网格的环境中的任务。我们表明，学习到的策略产生有效的动作，并且可以从游戏数据中学习。
英文摘要
As part of learning to play complex games, human players develop develop abstractions for concepts and strategies of gameplay consistent with game rules to improve their performance. These concepts are applied to explain other players' actions, and to inform their own actions in-game. Understanding other players' strategies is a crucial part of such improvement, but requires time and effort. In this paper, we propose a strategy similar to human cognition for training RL agents to synthesize learned strategies and policies as executable procedures based on sequences of gameplay actions. We present methods to automatically learn such programs to play chess and to solve tasks in a grid-based environment. We show that the learned strategies produce effective actions, and can be learned from gameplay data.
Journal refProceedings of the Eleventh Annual Conference on Advances in Cognitive Systems 2024
