arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2006.09436cs.LGcs.AIcs.ROstat.ML
收藏

SAMBA:安全的基于模型的主动强化学习

SAMBA: Safe Model-Based & Active Reinforcement Learning

Huawei R&D UK(华为英国研发部);Technical University Darmstadt(达姆施塔特工业大学);University College London(伦敦大学学院)

Alexander I. Cowen-Rivers, Daniel Palenicek, Vincent Moens, Mohammed Abdullah, Aivar Sootla, Jun Wang, Haitham Ammar

摘要:

在本文中,我们提出了SAMBA,一个新的安全强化学习框架,结合了概率建模,信息论和统计学方面。我们的方法建立在PILCO的基础上,通过支持条件风险价值约束的多目标问题,使用新的(半)样本外高斯过程评估指标进行积极探索。我们评估我们的算法在各种安全的动力系统基准,涉及低维和高维状态表示。我们的研究结果显示,与最先进的方法相比,样本和违规行为的数量级减少。最后,我们通过对我们的主动指标和安全约束的详细分析,提供了关于框架有效性的直觉。

英文摘要:

In this paper, we propose SAMBA, a novel framework for safe reinforcement learning that combines aspects from probabilistic modelling, information theory, and statistics. Our method builds upon PILCO to enable active exploration using novel(semi-)metrics for out-of-sample Gaussian process evaluation optimised through a multi-objective problem that supports conditional-value-at-risk constraints. We evaluate our algorithm on a variety of safe dynamical system benchmarks involving both low and high-dimensional state representations. Our results show orders of magnitude reductions in samples and violations compared to state-of-the-art methods. Lastly, we provide intuition as to the effectiveness of the framework by a detailed analysis of our active metrics and safety constraints.

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。

↑