arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1907.09765cs.LGstat.ML
收藏

行为者评价方法中的方差缩减(ACM)

Variance Reduction in Actor Critic Methods (ACM)

A.I Square Connect;Lamsade PSL(巴黎文理研究大学Lamsade)

Eric Benhamou

摘要:

在介绍了Actor Critic方法(ACM)之后,我们证明了ACM是控制变量估计。使用投影定理,我们证明了Q和优势Actor Critic(A2 C)方法是最优的意义下的$L^2$范数的控制变量估计由当前状态和动作的条件下的功能所跨越。毕达哥拉斯定理的这种直接应用提供了QAC和AAC在深度策略梯度方法中最常被称为A2C方法的强大性能的理论证明。这使我们能够推导出一个新的公式,具有较低的方差,并改善了传统的A2C方法的优势演员批评的方法。

英文摘要:

After presenting Actor Critic Methods (ACM), we show ACM are control variate estimators. Using the projection theorem, we prove that the Q and Advantage Actor Critic (A2C) methods are optimal in the sense of the $L^2$ norm for the control variate estimators spanned by functions conditioned by the current state and action. This straightforward application of Pythagoras theorem provides a theoretical justification of the strong performance of QAC and AAC most often referred to as A2C methods in deep policy gradient methods. This enables us to derive a new formulation for Advantage Actor Critic methods that has lower variance and improves the traditional A2C method.

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。

↑