arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1503.04269cs.LG
收藏

非策略时间差学习问题的一种重点研究方法

An Emphatic Approach to the Problem of Off-policy Temporal-Difference Learning

University of Alberta(阿尔伯塔大学)

Richard S. Sutton, A. Rupam Mahmood, Martha White

摘要:

在本文中,我们介绍了提高性能的参数时间差(TD)学习算法的思想,有选择地强调或不强调他们的更新在不同的时间步长。特别是,我们表明,不同的重点线性TD($λ$)的更新,以一种特定的方式导致其预期的更新,成为稳定下的政策培训。实现这一点的唯一现有无模型TD方法是梯度TD方法族,包括TDC、GTD($λ$)和GQ($λ$),其中每步计算在函数逼近参数的数量上是线性的。与这些方法相比,我们的学习TD($λ$)更简单,更易于使用,它只有一个学习参数向量和一个步长参数。我们的治疗包括一般的状态依赖贴现和自举功能,并指定不同程度的兴趣,准确地评估不同的状态的方式。

英文摘要:

In this paper we introduce the idea of improving the performance of parametric temporal-difference (TD) learning algorithms by selectively emphasizing or de-emphasizing their updates on different time steps. In particular, we show that varying the emphasis of linear TD($λ$)'s updates in a particular way causes its expected update to become stable under off-policy training. The only prior model-free TD methods to achieve this with per-step computation linear in the number of function approximation parameters are the gradient-TD family of methods including TDC, GTD($λ$), and GQ($λ$). Compared to these methods, our _emphatic TD($λ$)_ is simpler and easier to use; it has only one learned parameter vector and one step-size parameter. Our treatment includes general state-dependent discounting and bootstrapping functions, and a way of specifying varying degrees of interest in accurately valuing different states.

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。

↑