arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2008.00511cs.LGstat.ML
收藏

具有递进功能的课程学习

Curriculum Learning with a Progression Function

University of York(约克大学);Phantasma Labs GmbH(Phantasma实验室有限公司);King’s College London(伦敦国王学院);L3S research center(L3S研究中心);Leibniz Universität Hannover(汉诺威莱布尼茨大学)

Andrea Bassich, Francesco Foglino, Matteo Leonetti, Daniel Kudenko

摘要:

强化学习的课程学习是一种越来越流行的技术,它涉及在一系列中间任务上训练代理,称为课程,以提高代理的性能和学习速度。本文介绍了一种基于级数和映射函数的课程生成新范式。虽然级数函数指定了环境在任何给定时间的复杂度,但映射函数生成特定复杂度的环境。不同的进展功能,包括自主的在线任务进展的基础上代理的性能。我们的方法的好处和广泛的适用性是通过经验比较其性能的两个国家的最先进的课程学习算法上的六个领域。

英文摘要:

Curriculum Learning for Reinforcement Learning is an increasingly popular technique that involves training an agent on a sequence of intermediate tasks, called a Curriculum, to increase the agent's performance and learning speed. This paper introduces a novel paradigm for curriculum generation based on progression and mapping functions. While progression functions specify the complexity of the environment at any given time, mapping functions generate environments of a specific complexity. Different progression functions are introduced, including an autonomous online task progression based on the agent's performance. Our approach's benefits and wide applicability are shown by empirically comparing its performance to two state-of-the-art Curriculum Learning algorithms on six domains.

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。

↑