arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2302.00727cs.LGcs.AIstat.ML
收藏

基于核的Q学习的样本复杂度

Sample Complexity of Kernel-Based Q-Learning

MediaTek Research(联发科研究院);National Taiwan University(台湾大学);Academia Sinica(中央研究院)

Sing-Yuan Yeh, Fu-Chieh Chang, Chang-Wei Yueh, Pei-Yuan Wu, Alberto Bernacchia, Sattar Vakili

摘要:

现代强化学习(RL)往往面临着巨大的状态-动作空间。现有的分析结果通常是设置与少量的状态动作,或简单的模型,如线性建模的Q函数。为了得到统计上有效的RL策略,处理大的状态-动作空间,更一般的Q函数,最近的一些作品已经考虑使用核岭回归的非线性函数逼近。在这项工作中,我们推导出基于核的Q-学习的样本复杂性时,生成模型存在。我们提出了一个非参数Q学习算法,它可以在任意大规模的折扣MDP中找到一个$ε$ -最优策略。所提出的算法的样本复杂度是相对于$ε$和核的复杂度(在其信息增益方面)的阶最优。据我们所知,这是第一个结果显示有限样本的复杂性下,这样一个一般的模型。

英文摘要:

Modern reinforcement learning (RL) often faces an enormous state-action space. Existing analytical results are typically for settings with a small number of state-actions, or simple models such as linearly modeled Q-functions. To derive statistically efficient RL policies handling large state-action spaces, with more general Q-functions, some recent works have considered nonlinear function approximation using kernel ridge regression. In this work, we derive sample complexities for kernel based Q-learning when a generative model exists. We propose a nonparametric Q-learning algorithm which finds an $ε$-optimal policy in an arbitrarily large scale discounted MDP. The sample complexity of the proposed algorithm is order optimal with respect to $ε$ and the complexity of the kernel (in terms of its information gain). To the best of our knowledge, this is the first result showing a finite sample complexity under such a general model.

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。

↑