Learning controllable elements oriented representations for reinforcement learning

doi:10.1016/j.neucom.2023.126455

	Learning controllable elements oriented representations for reinforcement learning
	Yi, Qi 1,2,3; Zhang, Rui 2,3; Peng, Shaohui 2,3,4; Guo, Jiaming 2,3,4; Hu, Xing 2,3; Du, Zidong 2,3; Guo, Qi 2; Chen, Ruizhi 5; Li, Ling 4,5; Chen, Yunji 2,4
	2023-09-07
发表期刊	NEUROCOMPUTING
ISSN	0925-2312
卷号	549 页码:13
摘要	Deep Reinforcement Learning (deep RL) has been successfully applied to solve various decision-making problems in recent years. However, the observations in many real-world tasks are often high dimensional and include much task-irrelevant information, limiting the applications of RL algorithms. To tackle this problem, we propose LCER, a representation learning method that aims to provide RL algorithms with compact and sufficient descriptions of the original observations. Specifically, LCER trains representations to retain the controllable elements of the environment, which can reflect the action-related environment dynamics and thus are likely to be task-relevant. We demonstrate the strength of LCER on the DMControl Suite, proving that it can achieve state-of-the-art performance. LCER enables the pixel -based SAC to outperform state-based SAC on the DMControl 100 K benchmark, showing that the obtained representations can match the oracle descriptions (i.e. the physical states) of the environment. We also carry out experiments to show that LCER can efficiently filter out various distractions, especially when those distractions are not controllable.& COPY; 2023 Elsevier B.V. All rights reserved.
关键词	Reinforcement learning Representation learning
DOI	10.1016/j.neucom.2023.126455
收录类别	SCI
语种	英语
资助项目	National Key Research and Development Program of China[2017YFA0700900] ; NSF of China[61925208] ; NSF of China[62102399] ; NSF of China[62002338] ; NSF of China[U19B2019] ; NSF of China[61732020] ; Beijing Academy of Artificial Intelligence (BAAI) ; CAS Project for Young Scientists in Basic Research[YSBR-029] ; Youth Innovation Promotion Association CAS and Xplore Prize
WOS研究方向	Computer Science
WOS类目	Computer Science, Artificial Intelligence
WOS记录号	WOS:001035238900001
出版者	ELSEVIER
引用统计	被引频次：4[WOS] [WOS记录] [WOS相关记录]
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/21302
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Yi, Qi
作者单位	1.Univ Sci & Technol China, Hefei, Peoples R China 2.Chinese Acad Sci, Inst Comp Technol, SKL Processors, Beijing, Peoples R China 3.Cambricon Technol, Beijing, Peoples R China 4.Univ Chinese Acad Sci, Beijing, Peoples R China 5.Chinese Acad Sci, Inst Software, Beijing, Peoples R China
推荐引用方式 GB/T 7714	Yi, Qi,Zhang, Rui,Peng, Shaohui,et al. Learning controllable elements oriented representations for reinforcement learning[J]. NEUROCOMPUTING,2023,549:13.
APA	Yi, Qi.,Zhang, Rui.,Peng, Shaohui.,Guo, Jiaming.,Hu, Xing.,...&Chen, Yunji.(2023).Learning controllable elements oriented representations for reinforcement learning.NEUROCOMPUTING,549,13.
MLA	Yi, Qi,et al."Learning controllable elements oriented representations for reinforcement learning".NEUROCOMPUTING 549(2023):13.