arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2601.04441cs.LG
收藏

在大型离散动作空间中通过结构化策略初始化改进和加速离线强化学习

Improving and Accelerating Offline RL in Large Discrete Action Spaces with Structured Policy Initialization

University of Virginia(弗吉尼亚大学);MBZUAI(马克斯·普朗克智能系统研究所)

Matthew Landers, Taylor W. Killian, Thomas Hartvigsen, Afsaneh Doryab

AI总结:

SPIN通过结构化策略初始化方法,在大型离散动作空间中提升离线强化学习性能,显著提高平均回报并加速收敛。

摘要:

在离散组合动作空间中进行强化学习需要在指数级众多的联合动作中搜索,以同时选择多个子动作形成连贯的组合。现有方法要么通过假设子动作之间独立来简化策略学习,这往往导致不连贯或无效的动作,要么尝试联合学习动作结构和控制,但速度慢且不稳定。我们引入了结构化策略初始化(SPIN),一种两阶段框架,首先预训练动作结构模型(ASM)以捕捉有效动作的流形,然后冻结这一表示并训练轻量级的策略头进行控制。在具有挑战性的离散DM控制基准上,SPIN在提升平均回报方面比现有最佳方法高至39%,同时将收敛时间减少至12.8倍。

英文摘要:

Reinforcement learning in discrete combinatorial action spaces requires searching over exponentially many joint actions to simultaneously select multiple sub-actions that form coherent combinations. Existing approaches either simplify policy learning by assuming independence across sub-actions, which often yields incoherent or invalid actions, or attempt to learn action structure and control jointly, which is slow and unstable. We introduce Structured Policy Initialization (SPIN), a two-stage framework that first pre-trains an Action Structure Model (ASM) to capture the manifold of valid actions, then freezes this representation and trains lightweight policy heads for control. On challenging discrete DM Control benchmarks, SPIN improves average return by up to 39% over the state of the art while reducing time to convergence by up to 12.8$\times$.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑