arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2302.05372cs.LGstat.ML
收藏

迈向基于模型的鲁棒强化学习的极小极大最优性

Towards Minimax Optimality of Model-based Robust Reinforcement Learning

CMAP, CNRS, Ecole Polytechnique, Institut Polytechnique de Paris(CMAP、法国国家科学研究中心、巴黎综合理工学院、巴黎理工学院);INRIA Paris, HeKA(法国国家信息与自动化研究所巴黎分部、HeKA);Centre de Recherche des Cordeliers, INSERM, Universite de Paris, Sorbonne Universite(科尔德里耶研究所、法国国家健康与医学研究院、巴黎大学、索邦大学);Cohere(科here公司)

Pierre Clavier, Erwan Le Pennec, Matthieu Geist

AI总结:

本文研究鲁棒折扣MDP在生成模型下的样本复杂度,针对Lp球不确定性集合,将sa-和s-矩形情形的样本复杂度统一改进至H^4|S||A|/ε^2,并在不确定性足够小时恢复非鲁棒下界。

摘要:

我们研究在仅能访问名义核的生成模型的情况下,获得鲁棒折扣马尔可夫决策过程(RMDPs)中ε-最优策略的样本复杂度。该问题在非鲁棒情形下已被广泛研究,已知对使用$\tilde{\mathcal{O}}(\frac{H^3 \mid S \mid\mid A \mid}{ε^2})$个样本估计的经验MDP应用任何规划方法都能提供ε-最优策略,并且这是极小极大最优的。鲁棒情形下的结果则少得多。对于$sa$-(分别为$s$-)矩形不确定性集合,在不确定性集合基于总变差(TV)、KL或卡方散度时,特定算法的最佳已知样本复杂度为$\tilde{\mathcal{O}}(\frac{H^4 \mid S \mid^2\mid A \mid}{ε^2})$(分别为$\tilde{\mathcal{O}}(\frac{H^4 \mid S \mid^2\mid A \mid^2}{ε^2})$)。在本文中,我们考虑由$L_p$-球定义的不确定性集合(可恢复TV情形),并研究应用于使用生成模型估计的经验RMDP的任何规划算法(对解具有高精度保证)的样本复杂度。在一般情况下,我们证明了$sa$-和$s$-矩形情形的样本复杂度均为$\tilde{\mathcal{O}}(\frac{H^4 \mid S \mid\mid A \mid}{ε^2})$(分别改进了$\mid S \mid$和$\mid S \mid\mid A \mid$)。当不确定性的大小足够小时,我们将样本复杂度改进为$\tilde{\mathcal{O}}(\frac{H^3 \mid S \mid\mid A \mid }{ε^2})$,首次恢复了非鲁棒情形的下界,并在不确定性大小足够小时恢复了鲁棒下界。

英文摘要:

We study the sample complexity of obtaining an $ε$-optimal policy in \emph{Robust} discounted Markov Decision Processes (RMDPs), given only access to a generative model of the nominal kernel. This problem is widely studied in the non-robust case, and it is known that any planning approach applied to an empirical MDP estimated with $\tilde{\mathcal{O}}(\frac{H^3 \mid S \mid\mid A \mid}{ε^2})$ samples provides an $ε$-optimal policy, which is minimax optimal. Results in the robust case are much more scarce. For $sa$- (resp $s$-)rectangular uncertainty sets, the best known sample complexity is $\tilde{\mathcal{O}}(\frac{H^4 \mid S \mid^2\mid A \mid}{ε^2})$ (resp. $\tilde{\mathcal{O}}(\frac{H^4 \mid S \mid^2\mid A \mid^2}{ε^2})$), for specific algorithms and when the uncertainty set is based on the total variation (TV), the KL or the Chi-square divergences. In this paper, we consider uncertainty sets defined with an $L_p$-ball (recovering the TV case), and study the sample complexity of \emph{any} planning algorithm (with high accuracy guarantee on the solution) applied to an empirical RMDP estimated using the generative model. In the general case, we prove a sample complexity of $\tilde{\mathcal{O}}(\frac{H^4 \mid S \mid\mid A \mid}{ε^2})$ for both the $sa$- and $s$-rectangular cases (improvements of $\mid S \mid$ and $\mid S \mid\mid A \mid$ respectively). When the size of the uncertainty is small enough, we improve the sample complexity to $\tilde{\mathcal{O}}(\frac{H^3 \mid S \mid\mid A \mid }{ε^2})$, recovering the lower-bound for the non-robust case for the first time and a robust lower-bound when the size of the uncertainty is small enough.

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。

↑