Invited lecture of ICONIP 2017
Learning-based Control: Opportunities and Challenges
With the recent development of deep learning and hardware computing technologies, scientists and engineers will hopefully find efficient ways to design brain-like intelligent systems that are highly robust, adaptive, scalable, and fault-tolerant to uncertain and unstructured environments. Yet, developing such truly intelligent systems requires significant research on both fundamental understanding of brain intelligence as well as complex engineering design.
In this talk, I will present a new reinforcement learning (RL) and adaptive dynamic programing (ADP) framework for improved decision-making and control capability, named goalrepresentation ADP (GrADP). The two key questions addressed by this new type of GrADP include: (1) where does the reinforcement signal comes from; and (2) how to develop an internal goal representation. Compared to the existing methods with a manual or “handcrafted” reinforcement signal design, this GrADP framework can automatically and adaptively develop the internal goal representation over time. Under this framework, I will present numerous applications ranging from smart grid control to human-robot interaction to demonstrate its broader and far-reaching applications.
- Paper Submission Deadline
June 15, 2017
June 30, 2017 - Notification of Acceptance
July 31, 2017 - Final Paper Submission
August 20, 2017 - Congress Date
November 14–18, 2017
Contact Us
Address: 95 Zhongguancun East Road,
Beijing 100190, China
Email: iconip2017@foxmail.com
Fax: +86-10-8254-4799