Policy generation network for zero-shot policy learning

doi:10.1111/coin.12591

	Policy generation network for zero-shot policy learning
	Qian, Yiming 1,2; Zhang, Fengyi1,2 ; Liu, Zhiyong1,2,3
发表期刊	COMPUTATIONAL INTELLIGENCE
ISSN	0824-7935
	2023-07-04
页码	27
通讯作者	Liu, Zhiyong(zhiyong.liu@ia.ac.cn)
摘要	Lifelong reinforcement learning is able to continually accumulate shared knowledge by estimating the inter-task relationships based on training data for the learned tasks in order to accelerate learning for new tasks by knowledge reuse. The existing methods employ a linear model to represent the inter-task relationships by incorporating task features in order to accomplish a new task without any learning. But these methods may be ineffective for general scenarios, where linear models build inter-task relationships from low-dimensional task features to high-dimensional policy parameters space. Also, the deficiency of calculating errors from objective function may arise in the lifelong reinforcement learning process when some errors of policy parameters restrain others due to inter-parameter correlation. In this paper, we develop a policy generation network that nonlinearly models the inter-task relationships by mapping low-dimensional task features to the high-dimensional policy parameters, in order to represent the shared knowledge more effectively. At the same time, we propose a novel objective function of lifelong reinforcement learning to relieve the deficiency of calculating errors by adding weight constraints for errors. We empirically demonstrate that our method improves the zero-shot policy performance across a variety of dynamical systems.
关键词	knowledge representation lifelong reinforcement learning zero-shot policy generation
DOI	10.1111/coin.12591
收录类别	SCI
语种	英语
资助项目	National Key Research and Development Plan of China[2020AAA0108902] ; Strategic Priority Research Program of Chinese Academy of Science[XDB32050100] ; National Natural Science Foundation of China[61627808] ; Dongguan Core Technology Research Frontier Project[2019622101001]
项目资助者	National Key Research and Development Plan of China ; Strategic Priority Research Program of Chinese Academy of Science ; National Natural Science Foundation of China ; Dongguan Core Technology Research Frontier Project
WOS研究方向	Computer Science
WOS类目	Computer Science, Artificial Intelligence
WOS记录号	WOS:001022798600001
出版者	WILEY
引用统计
文献类型	期刊论文
条目标识符	http://ir.ia.ac.cn/handle/173211/53664
专题	多模态人工智能系统全国重点实验室
通讯作者	Liu, Zhiyong
作者单位	1.Univ Chinese Acad Sci UCAS, Sch Artificial Intelligence, Beijing 100049, Peoples R China 2.Chinese Acad Sci, Inst Automat, State Key Lab Management & Control Complex Syst, Beijing, Peoples R China 3.Chinese Acad Sci, Cloud Comp Ctr, Dongguan, Guangdong, Peoples R China
第一作者单位	中国科学院自动化研究所
通讯作者单位	中国科学院自动化研究所
推荐引用方式 GB/T 7714	Qian, Yiming,Zhang, Fengyi,Liu, Zhiyong. Policy generation network for zero-shot policy learning[J]. COMPUTATIONAL INTELLIGENCE,2023:27.
APA	Qian, Yiming,Zhang, Fengyi,&Liu, Zhiyong.(2023).Policy generation network for zero-shot policy learning.COMPUTATIONAL INTELLIGENCE,27.
MLA	Qian, Yiming,et al."Policy generation network for zero-shot policy learning".COMPUTATIONAL INTELLIGENCE (2023):27.