arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2106.09110cs.LGcs.ROcs.SYeess.SY
收藏

使用基于记忆的干预的安全强化学习

Safe Reinforcement Learning Using Advantage-Based Intervention

Institute for Robotics and Intelligent Machines, Georgia Institute of Technology(佐治亚理工学院机器人与智能机器研究所);Paul G. Allen School of Computer Science and Engineering, University of Washington(华盛顿大学保罗·G·艾伦计算机科学与工程学院);Microsoft Research(微软研究院)

Nolan Wagener, Byron Boots, Ching-An Cheng

摘要:

许多序贯决策问题涉及到寻找一个策略,使总回报最大化,同时遵守安全约束。尽管最近的许多研究都集中在开发安全的强化学习(RL)算法,这些算法在训练后产生安全的策略,但确保训练期间的安全仍然是一个悬而未决的问题。一个基本的挑战是执行探索,同时仍然满足未知的马尔可夫决策过程(MDP)的约束。在这项工作中,我们解决这个问题的机会约束设置。我们提出了一个新的算法,SAILR,使用干预机制的基础上优势函数,以保持整个训练的代理安全和优化代理的政策,使用现成的RL算法设计的无约束MDPs。我们的方法在训练和部署期间都有很强的安全保证(即,在训练之后并且没有干预机制)以及与最优安全约束策略相比的策略性能。在我们的实验中,我们表明,SAILR在训练过程中违反的约束远远低于标准的安全RL和受约束的MDP方法,并收敛到一个性能良好的策略,可以安全地部署,而无需干预。我们的代码可在https://github.com/nolanwagener/safe_rl上获得。

英文摘要:

Many sequential decision problems involve finding a policy that maximizes total reward while obeying safety constraints. Although much recent research has focused on the development of safe reinforcement learning (RL) algorithms that produce a safe policy after training, ensuring safety during training as well remains an open problem. A fundamental challenge is performing exploration while still satisfying constraints in an unknown Markov decision process (MDP). In this work, we address this problem for the chance-constrained setting. We propose a new algorithm, SAILR, that uses an intervention mechanism based on advantage functions to keep the agent safe throughout training and optimizes the agent's policy using off-the-shelf RL algorithms designed for unconstrained MDPs. Our method comes with strong guarantees on safety during both training and deployment (i.e., after training and without the intervention mechanism) and policy performance compared to the optimal safety-constrained policy. In our experiments, we show that SAILR violates constraints far less during training than standard safe RL and constrained MDP approaches and converges to a well-performing policy that can be deployed safely without intervention. Our code is available at https://github.com/nolanwagener/safe_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。

↑