CONFERENCE PROCEEDINGS
Advanced Search
Home > Proceedings > Volume 12253 > Article
Paper
27 June 2022 Improved the sample efficiency of episodic reinforcement learning by forcing state representations
Ruiyuan Zhang, William Zhu
Author Affiliations +
Proceedings Volume 12253, International Conference on Automation Control, Algorithm, and Intelligent Bionics (ACAIB 2022); 122531I (2022) https://doi.org/10.1117/12.2639467
Event: Second International Conference on Automation Control, Algorithm, and Intelligent Bionics (ACAIB 2022), 2022, Qingdao, China
Abstract
Episodic reinforcement learning (ERL) is a class of algorithms that use episodic memory for improving the performance and the sample efficiency of reinforcement learning (RL). Although it has achieved some success, existing ERL algorithms have to interact with the environment for many rounds to gain satisfying performance. In this paper, we propose the algorithm episodic memory by forcing state representations (EMSR) to improve the performance and sample efficiency of ERL. Specifically, our algorithm uses a transition model to predict the hidden state representations of the agent’s multiple future steps for augmenting reward maximization, which can help the agent learn quickly. In this way, our method can achieve better performance and higher sample efficiency than previous state-of-the-art algorithms. Experimental results demonstrate the superiority of our method.
© (2022) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Citation Download Citation
Ruiyuan Zhang and William Zhu "Improved the sample efficiency of episodic reinforcement learning by forcing state representations", Proc. SPIE 12253, International Conference on Automation Control, Algorithm, and Intelligent Bionics (ACAIB 2022), 122531I (27 June 2022); https://doi.org/10.1117/12.2639467
ACCESS THE FULL ARTICLE
ORGANIZATIONAL
Sign in with credentials provided by your organization.
INSTITUTIONAL
Select your institution to access the SPIE Digital Library.
SELECT YOUR INSTITUTION
PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.
PERSONAL SIGN IN
No SPIE Account? Create one
PURCHASE THIS CONTENT
SUBSCRIBE TO DIGITAL LIBRARY
50 downloads per 1-year subscription
Members: $195
Non-members: $335 ADD TO CART
25 downloads per 1 - year subscription
Members: $145
Non-members: $250 ADD TO CART
PURCHASE SINGLE ARTICLE
Includes PDF, HTML & Video, when available
Members: $17.00
Non-members: $21.00 ADD TO CART
PROCEEDINGS
 6 PAGES
DOWNLOAD PAPER SAVE TO MY LIBRARY
GET CITATION
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Computer programming
Neural networks
Stochastic processes
Machine learning
RELATED CONTENT
Subscribe to Digital Library
Receive Erratum Email Alert
Back to Top