arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1804.04577cs.LGstat.ML
收藏

基于知识的聚合和深度强化学习:综述和一些新的实现

Feature-Based Aggregation and Deep Reinforcement Learning: A Survey and Some New Implementations

Dimitri P. Bertsekas

摘要:

在本文中,我们讨论了有限状态贴现马尔可夫决策问题的近似解的策略迭代方法,重点是基于特征的聚合方法及其与深度强化学习方案的联系。我们引入的原始问题的状态的功能,我们制定了一个较小的“聚合”马尔可夫决策问题,其状态相关的功能。我们讨论这种类型的聚合的属性和可能的实现,包括一种新的方法来近似的政策迭代。在该方法中,策略改进操作将基于特征的聚合与使用深度神经网络或其他计算的特征构造相结合。我们认为,与基于神经网络的强化学习提供的特征的线性函数相比,通过聚合提供的特征的非线性函数可以更准确地近似策略的成本函数,从而可能导致更有效的策略改进。

英文摘要:

In this paper we discuss policy iteration methods for approximate solution of a finite-state discounted Markov decision problem, with a focus on feature-based aggregation methods and their connection with deep reinforcement learning schemes. We introduce features of the states of the original problem, and we formulate a smaller "aggregate" Markov decision problem, whose states relate to the features. We discuss properties and possible implementations of this type of aggregation, including a new approach to approximate policy iteration. In this approach the policy improvement operation combines feature-based aggregation with feature construction using deep neural networks or other calculations. We argue that the cost function of a policy may be approximated much more accurately by the nonlinear function of the features provided by aggregation, than by the linear function of the features provided by neural network-based reinforcement learning, thereby potentially leading to more effective policy improvement.

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。

↑