arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2301.04741cs.LG
收藏

使用学习到的动力学模型的高效基于偏好的强化学习

Efficient Preference-Based Reinforcement Learning Using Learned Dynamics Models

UC Berkeley(加州大学伯克利分校);Army Research Lab(陆军研究实验室);University of Utah(犹他大学)

Yi Liu, Gaurav Datta, Ellen Novoseller, Daniel S. Brown

AI总结:

本文研究了在基于偏好的强化学习中使用学习到的动力学模型,证明其能显著减少环境交互、安全合成偏好查询并实现无环境交互的奖励预训练,使机器人更安全高效地学习定制化策略。

摘要:

基于偏好的强化学习(PbRL)能使机器人基于个人偏好学习执行任务,而无需手工设计的奖励函数。然而,现有方法要么假设可访问高保真模拟器或解析模型,要么采用无模型方法,这需要大量且可能不安全的在线环境交互。本文研究了在执行PbRL时使用学习到的动力学模型的益处与挑战。具体而言,我们提供的证据表明,学习到的动力学模型在执行PbRL时提供以下益处:(1) 偏好获取和策略优化所需的环境交互显著少于无模型PbRL;(2) 多样化的偏好查询可作为标准基于模型的RL的副产品被安全高效地合成;(3) 基于次优演示的奖励预训练可在无任何环境交互的情况下进行。本文提供了实证证据,表明学习到的动力学模型能使机器人以比先前偏好学习方法更安全、更具样本效率的方式,基于用户偏好学习定制化策略。补充材料和代码可在https://sites.google.com/berkeley.edu/mop-rl获取。

英文摘要:

Preference-based reinforcement learning (PbRL) can enable robots to learn to perform tasks based on an individual's preferences without requiring a hand-crafted reward function. However, existing approaches either assume access to a high-fidelity simulator or analytic model or take a model-free approach that requires extensive, possibly unsafe online environment interactions. In this paper, we study the benefits and challenges of using a learned dynamics model when performing PbRL. In particular, we provide evidence that a learned dynamics model offers the following benefits when performing PbRL: (1) preference elicitation and policy optimization require significantly fewer environment interactions than model-free PbRL, (2) diverse preference queries can be synthesized safely and efficiently as a byproduct of standard model-based RL, and (3) reward pre-training based on suboptimal demonstrations can be performed without any environmental interaction. Our paper provides empirical evidence that learned dynamics models enable robots to learn customized policies based on user preferences in ways that are safer and more sample efficient than prior preference learning approaches. Supplementary materials and code are available at https://sites.google.com/berkeley.edu/mop-rl.

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。

↑