888.3net新浦京游戏机器人与信息自动化研究所 天津市智能机器人技术重点实验室

Institute of Robotics and Automatic Information System

Tianjin Key Laboratory of Intelligent Robotics

2019年春季先进机器人与人工智能系列学术讲座（第138期）

Seminar Series：Advanced Robotics & Artificial Intelligence

报告题目：Autonomous Mental Development: Auto-Programming for General Purposes by Robots for Lifetime Skills Learning

时间：2019年5月27日（周一）14:00-15:30

地点：888.3net新浦京游戏北楼102会议室

报告人：Juyang (John) Weng, Professor

Department of Computer Science and Engineering,

Cognitive Science Program, and Neuroscience Program

Michigan State University

East Lansing, MI 48824 USA

http://www.cse.msu.edu/~weng/
翁老师是最早提出智能发育的学者之一，早在2001年就在Science上发表题为“Autonomous Mental Development by Robots and Animals”的论文。

Abstract:

Since their paper in Science 2001, Dr. Weng with his research group has made major advances in understanding and modeling how human brains learn and work, through theory, algorithms, and experimental demonstrations.  These advances explain how to enable a robot to not only perform autonomously but also learn autonomously; not only for a task but also through its “lifetime” for all open-ended tasks. In this introductory talk, Dr. Weng will outline a new kind of neural network, Developmental Network (DN), that models how a human/robot brain learns vision, audition, natural languages, and many other skills fully autonomously.  The representations inside the “skull” emerge while the human/robot body is interacting with the external physical world using the body’s sensors (e.g., eyes, ears, skins, limbs) and effectors (e.g., muscles and glands).  In the eyes of human teachers the learner becomes increasingly capable, from simple tasks as a “baby” to complex tasks as an “adult”.  Many roboticists have used neural networks for pattern recognition and control, but the DN learns a general-purpose computer (i.e., Emergent Universal Turing Machine, EUTM).  The developmental learning of EUTM is fully autonomous inside the skull and across the entire life without pre-given lifetime tasks. This is significant because such robots can auto-program for general purposes (APFGP).  The talk uses vision, audition, natural languages, and vision-based autonomous navigation as examples.

Short Bio:

Juyang (John) Weng is a professor at the Department of Computer Science and Engineering, the Cognitive Science Program, and the Neuroscience Program, Michigan State University, East Lansing, Michigan, USA. He received his BS degree from Fudan University, China, in 1982, his MS and PhD degrees from University of Illinois at Urbana-Champaign, USA, 1985 and 1989, respectively, all in Computer Science.  From August 2006 to May 2007, he was also a visiting professor at the Department of Brain and Cognitive Science of MIT, USA.   Cresceptron by Drs. J. Weng, N. Ahuja, and T. S. Huang published 1992 was the first deep learning neural network for the 3D natural world.  It seeded the current deep learning wave in artificial intelligence (AI).  His research interests include brain models, computational neuroscience, computational developmental psychology, computer vision, audition, touch, natural languages, and intelligent robots.  He is the author or coauthor of over three hundred research articles. He is an editor-in-chief of International Journal of Humanoid Robotics, the editor-in-chief of the Brain-Mind Magazine, and an associate editor of the new IEEE Transactions on Autonomous Mental Development (now IEEE Transactions on Cognitive and Developmental Systems). He was the Chairman of the Governing Board of the International Conferences on Development and Learning (ICDLs) (2005-2007, http://cogsci.ucsd.edu/~triesch/icdl/), chairman of the Autonomous Mental Development Technical Committee of the IEEE Computational Intelligence Society (2004-2005), an associate editor of IEEE Trans. on Pattern Recognition and Machine Intelligence, and an associate editor of IEEE Trans. on Image Processing.  He is a Fellow of IEEE.