Guided Policy Search for Sequential Multitask Learning

doi:10.1109/TSMC.2018.2800040

	Guided Policy Search for Sequential Multitask Learning
	Xiong, Fangzhou1,2 ; Sun, Biao 3; Yang, Xu1,2 ; Qiao, Hong1,2,3,4,5 ; Huang, Kaizhu 6; Hussain, Amir 7; Liu, Zhiyong1,2,4,5
刊名	IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
	2019
卷号	49 期号:1 页码:216-226
关键词	Elastic weight consolidation (EWC) guided policy search (GPS) reinforcement learning (RL) sequential multitask learning
ISSN号	2168-2216
DOI	10.1109/TSMC.2018.2800040
通讯作者	Liu, Zhiyong(zhiyong.liu@ia.ac.cn)
英文摘要	Policy search in reinforcement learning (RL) is a practical approach to interact directly with environments in parameter spaces, that often deal with dilemmas of local optima and real-time sample collection. A promising algorithm, known as guided policy search (GPS), is capable of handling the challenge of training samples using trajectory-centric methods. It can also provide asymptotic local convergence guarantees. However, in its current form, the GPS algorithm cannot operate in sequential multitask learning scenarios. This is due to its batch-style training requirement, where all training samples are collectively provided at the start of the learning process. The algorithm's adaptation is thus hindered for real-time applications, where training samples or tasks can arrive randomly. In this paper, the GPS approach is reformulated, by adapting a recently proposed, lifelong-learning method, and elastic weight consolidation. Specifically, Fisher information is incorporated to impart knowledge from previously learned tasks. The proposed algorithm, termed sequential multitask learning-GPS, is able to operate in sequential multitask learning settings and ensuring continuous policy learning, without catastrophic forgetting. Pendulum and robotic manipulation experiments demonstrate the new algorithms efficacy to learn control policies for handling sequentially arriving training samples, delivering comparable performance to the traditional, and batch-based GPS algorithm. In conclusion, the proposed algorithm is posited as a new benchmark for the real-time RL and robotics research community.
资助项目	NSFC[U1613213] ; NSFC[61375005] ; NSFC[61503383] ; NSFC[61210009] ; NSFC[61627808] ; NSFC[91648205] ; NSFC[61702516] ; NSFC[61473236] ; National Key Research and Development Plan of China[2017YFB1300202] ; National Key Research and Development Plan of China[2016YFC0300801] ; MOST[2015BAK35B00] ; MOST[2015BAK35B01] ; Guangdong Science and Technology Department[2016B090910001] ; Suzhou Science and Technology Program[SYG201712] ; Suzhou Science and Technology Program[SZS201613] ; Strategic Priority Research Program of the Chinese Academy of Science[XDB02080003] ; Key Program Special Fund in XJTLU[KSF-A-01] ; U.K. Engineering and Physical Sciences Research Council[EP/M026981/1]
WOS研究方向	Automation & Control Systems ; Computer Science
语种	英语
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号	WOS:000454241100019
资助机构	NSFC ; National Key Research and Development Plan of China ; MOST ; Guangdong Science and Technology Department ; Suzhou Science and Technology Program ; Strategic Priority Research Program of the Chinese Academy of Science ; Key Program Special Fund in XJTLU ; U.K. Engineering and Physical Sciences Research Council
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/25641]
专题	中国科学院自动化研究所
通讯作者	Liu, Zhiyong
作者单位	1.Chinese Acad Sci, Inst Automat, State Key Lab Management & Control Complex Syst, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Sch Comp & Control, Beijing 100049, Peoples R China 3.Univ Sci & Technol Beijing, Sch Automat & Elect Engn, Beijing 100083, Peoples R China 4.CAS Ctr Excellence Brain Sci & Intelligence Techn, Shanghai 200031, Peoples R China 5.Chinese Acad Sci, Cloud Comp Ctr, Dongguan 523808, Peoples R China 6.Xian Jiaotong Liverpool Univ, Dept Elect & Elect Engn, Suzhou 215123, Peoples R China 7.Univ Stirling, Sch Nat Sci, Div Comp Sci & Maths, Stirling FK9 4LA, Scotland
推荐引用方式 GB/T 7714	Xiong, Fangzhou,Sun, Biao,Yang, Xu,et al. Guided Policy Search for Sequential Multitask Learning[J]. IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS,2019,49(1):216-226.
APA	Xiong, Fangzhou.,Sun, Biao.,Yang, Xu.,Qiao, Hong.,Huang, Kaizhu.,...&Liu, Zhiyong.(2019).Guided Policy Search for Sequential Multitask Learning.IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS,49(1),216-226.
MLA	Xiong, Fangzhou,et al."Guided Policy Search for Sequential Multitask Learning".IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS 49.1(2019):216-226.