Averaged-DQN: Variance Reduction and Stabilization for Deep Reinforcement Learning

最新推荐文章于 2020-11-20 14:58:38 发布

Adam婷

最新推荐文章于 2020-11-20 14:58:38 发布

阅读量2.1k

点赞数 3

CC 4.0 BY-SA版权

分类专栏：算法机器学习深度学习强化学习深度强化学习 DQN ICLR 论文研读

本文链接：https://blog.csdn.net/weixin_41697507/article/details/94402477

机器学习同时被 3 个专栏收录

161 篇文章 ¥19.90 ¥99.00

订阅专栏

超级会员免费看

强化学习

26 篇文章 ¥19.90 ¥99.00

订阅专栏

超级会员免费看

算法

161 篇文章

订阅专栏

Averaged-DQN是一种针对深度强化学习（DQN）的改进方法，通过平均先前学习的Q值估计来减少目标近似误差方差，从而提高训练稳定性和性能。论文分析了DQN中的过高估计问题，并提出Averaged-DQN算法作为解决方案。实验证实在Arcade Learning Environment和Gridworld环境中，Averaged-DQN表现出优于传统DQN和Ensemble DQN的稳定性和性能。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

Averaged-DQN：深度强化学习的方差减少和稳定性

Abstract

Instability and variability of Deep Reinforcement Learning (DRL) algorithms tend to adversely affect their performance. Averaged-DQN is a sim-ple extension to the DQN algorithm, based on averaging previously learned Q-values estimates, which leads to a more stable training procedure and improved performance by reducing approximation error variance in the target values. To understand the effect of the algorithm, we examine the source of value function estimation errors and provide an analytical comparison within a simplified model. We further present experiments on the Arcade Learning Environment benchmark that demonstrate significantly improved stability and performance due to the