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 2007.09569cs.AIcs.LG
收藏

理解和缓解优先经验重放的局限性

Understanding and Mitigating the Limitations of Prioritized Experience Replay

University of Alberta(阿尔伯塔大学);Huawei Noah’s Ark Lab(华为诺亚方舟实验室);University of Toronto(多伦多大学);Vector Institute(向量研究所);CIFAR AI Chair(加拿大高等研究院人工智能讲席)

Yangchen Pan, Jincheng Mei, Amir-massoud Farahmand, Martha White, Hengshuai Yao, Mohsen Rohani, Jun Luo

摘要:

优先经验重放(ER)已被经验证明,以提高样本效率在许多领域,并吸引了极大的关注,但是,很少有理论上的理解,为什么这样的优先抽样帮助及其局限性。在这项工作中,我们深入研究优先ER。在一个监督学习设置,我们显示了基于误差的优先级采样方法的均方误差和立方功率损失的均匀采样之间的等价性。然后,我们提供了理论上的见解,为什么它提高了收敛速度均匀采样在早期学习。在此基础上,我们进一步指出了优先ER方法的两个局限性:1)过时的优先级和2)样本空间的覆盖不足。为了减轻这些限制,我们提出了基于模型的随机梯度Langevin动态采样方法。我们表明,我们的方法提供的状态分布接近一个理想的优先抽样分布估计的蛮力法,不受这两个限制。我们进行离散和连续控制问题的实验,以显示我们的方法的有效性,并检查我们的方法在自动驾驶应用程序的实际意义。

英文摘要:

Prioritized Experience Replay (ER) has been empirically shown to improve sample efficiency across many domains and attracted great attention; however, there is little theoretical understanding of why such prioritized sampling helps and its limitations. In this work, we take a deep look at the prioritized ER. In a supervised learning setting, we show the equivalence between the error-based prioritized sampling method for mean squared error and uniform sampling for cubic power loss. We then provide theoretical insight into why it improves convergence rate upon uniform sampling during early learning. Based on the insight, we further point out two limitations of the prioritized ER method: 1) outdated priorities and 2) insufficient coverage of the sample space. To mitigate the limitations, we propose our model-based stochastic gradient Langevin dynamics sampling method. We show that our method does provide states distributed close to an ideal prioritized sampling distribution estimated by the brute-force method, which does not suffer from the two limitations. We conduct experiments on both discrete and continuous control problems to show our approach's efficacy and examine the practical implication of our method in an autonomous driving application.

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。

↑