arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2006.04761cs.LGmath.OCstat.ML
收藏

时序差分与Q学习能否学到表示?一种平均场理论

Can Temporal-Difference and Q-Learning Learn Representation? A Mean-Field Theory

Northwestern University(西北大学);Princeton University(普林斯顿大学);Georgia Institute of Technology(佐治亚理工学院)

Yufeng Zhang, Qi Cai, Zhuoran Yang, Yongxin Chen, Zhaoran Wang

AI总结:

本文利用平均场理论证明,在过参数化两层神经网络下,时序差分和Q学习能以次线性速率全局最小化均方投影贝尔曼误差,且特征表示收敛到最优表示,推广了神经正切核机制下的既有结果。

摘要:

时序差分和Q学习在深度强化学习中扮演着关键角色,它们借助神经网络等具有表达能力的非线性函数逼近器而得到增强。其经验成功的核心在于学习到的特征表示,该表示将丰富的观测(例如图像和文本)嵌入到编码语义结构的潜在空间中。同时,这种特征表示的演化对于时序差分和Q学习的收敛至关重要。特别地,当函数逼近器在特征表示上是线性的,且该表示在整个学习过程中保持固定时,时序差分学习收敛;否则可能发散。我们旨在回答以下问题:当函数逼近器是神经网络时,相关的特征表示如何演化?如果它收敛,是否收敛到最优表示?我们证明,利用过参数化的两层神经网络,时序差分和Q学习以次线性速率全局最小化均方投影贝尔曼误差。此外,相关的特征表示收敛到最优表示,推广了Cai等人(2019)在神经正切核机制下的先前分析,在该机制中相关特征表示稳定在初始表示。我们分析的关键是一种平均场视角,它将有限维参数的演化与其在无限维Wasserstein空间上的极限对应物联系起来。我们的分析推广到soft Q-learning,并进一步与策略梯度建立联系。

英文摘要:

Temporal-difference and Q-learning play a key role in deep reinforcement learning, where they are empowered by expressive nonlinear function approximators such as neural networks. At the core of their empirical successes is the learned feature representation, which embeds rich observations, e.g., images and texts, into the latent space that encodes semantic structures. Meanwhile, the evolution of such a feature representation is crucial to the convergence of temporal-difference and Q-learning. In particular, temporal-difference learning converges when the function approximator is linear in a feature representation, which is fixed throughout learning, and possibly diverges otherwise. We aim to answer the following questions: When the function approximator is a neural network, how does the associated feature representation evolve? If it converges, does it converge to the optimal one? We prove that, utilizing an overparameterized two-layer neural network, temporal-difference and Q-learning globally minimize the mean-squared projected Bellman error at a sublinear rate. Moreover, the associated feature representation converges to the optimal one, generalizing the previous analysis of Cai et al. (2019) in the neural tangent kernel regime, where the associated feature representation stabilizes at the initial one. The key to our analysis is a mean-field perspective, which connects the evolution of a finite-dimensional parameter to its limiting counterpart over an infinite-dimensional Wasserstein space. Our analysis generalizes to soft Q-learning, which is further connected to policy gradient.

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。

↑