Prototypical Context-Aware Dynamics for Generalization in Visual Control With Model-Based Reinforcement Learning

doi:10.1109/TII.2024.3396525

CORC > 自动化研究所 > 中国科学院自动化研究所 > 复杂系统管理与控制国家重点实验室 > 深度强化学习

	Prototypical Context-Aware Dynamics for Generalization in Visual Control With Model-Based Reinforcement Learning
	Wang, Junjie 4,5; Zhang, Qichao 4,5; Mu, Yao 3; Li, Dong 2; Zhao, Dongbin 4,5; Zhuang, Yuzheng 2; Luo, Ping 3; Wang, Bin 2; Hao, Jianye 1,2
刊名	IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
	2024-05-15
页码	11
关键词	Context-aware dynamics generalization model-based reinforcement learning visual control
ISSN号	1551-3203
DOI	10.1109/TII.2024.3396525
通讯作者	Zhang, Qichao(zhangqichao2014@ia.ac.cn)
英文摘要	The latent world model, which efficiently represents high-dimensional observations within a latent space, has shown promise in reinforcement learning-based policies for visual control tasks. Due to a lack of clear environmental context comprehension, its applicability in a variety of contexts with unknown dynamics is constrained. We propose a prototypical context- aware dynamics (ProtoCAD) model to address this issue. This model captures local dynamics using temporally consistent latent contexts and aids generalization in visual control tasks. By grouping prototypes over historical experiences, ProtoCAD collects useful contextual information that improves model-based reinforcement learning dynamics generalization in two ways. First, to guarantee the consistency of prototype assignments for various temporal segments of the same latent trajectory, a temporally consistent prototypes regularizer is used. Then, a context representation is devised to combine the aggregated prototype with the projection embedding of latent states. According to extensive trials, ProtoCAD outperforms competing approaches in terms of dynamics generalization for visual robotic control and autonomous driving applications.
资助项目	National Key Research and Development Program of China
WOS研究方向	Automation & Control Systems ; Computer Science ; Engineering
语种	英语
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号	WOS:001226155200001
资助机构	National Key Research and Development Program of China
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/58455]
专题	复杂系统管理与控制国家重点实验室_深度强化学习
通讯作者	Zhang, Qichao
作者单位	1.Tianjin Univ, Coll Intelligence & Comp, Tianjin 300350, Peoples R China 2.Huawei Technol, Noahs Ark Lab, Beijing 100085, Peoples R China 3.Univ Hong Kong, Hong Kong 100085, Peoples R China 4.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China 5.Chinese Acad Sci, Inst Automat, Key Lab Multimodal Artificial Intelligence Syst, Beijing 100190, Peoples R China
推荐引用方式 GB/T 7714	Wang, Junjie,Zhang, Qichao,Mu, Yao,et al. Prototypical Context-Aware Dynamics for Generalization in Visual Control With Model-Based Reinforcement Learning[J]. IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS,2024:11.
APA	Wang, Junjie.,Zhang, Qichao.,Mu, Yao.,Li, Dong.,Zhao, Dongbin.,...&Hao, Jianye.(2024).Prototypical Context-Aware Dynamics for Generalization in Visual Control With Model-Based Reinforcement Learning.IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS,11.
MLA	Wang, Junjie,et al."Prototypical Context-Aware Dynamics for Generalization in Visual Control With Model-Based Reinforcement Learning".IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS (2024):11.