李世鹏教授：从GPT的人类认知类比中窥见通用人工智能的黎明

【引言】

GPT/ChatGPT的快速发展掀起了人工智能新一轮的浪潮，但其表现出智能的原理依然众说纷纭。一直以来，认知科学的新发现往往可以指导人工智能的研究方向。本次讲座将试着分析GPT和人类认知过程的相似之处，并指出GPT/ChatGPT不仅代表着一个新时代的开始，也代表着通向人工通用智能的道路。

第十期 IEEE TNSE 杰出讲座系列活动，我们有幸邀请到李世鹏教授介绍从人类认知的角度理解GPT的原理，并分享他在这个领域内的相关研究成果与有趣发现。

执行主席

Executive Chair

黄建伟

香港中文大学（深圳）校长讲座教授、协理副校长

AIRS 副院长兼群体智能中心主任

IEEE TNSE主编

IEEE Fellow

AAIA Fellow

报告嘉宾

Speaker

李世鹏

粤港澳大湾区数字经济研究院（IDEA）讲席科学家

中国科技部人工智能产业技术创新战略联盟（AITISA）联合创始人兼联合秘书长

香港中文大学（深圳）和香港科技大学（广州）兼职教授

国际欧亚科学院（IEAS）院士

IEEE Fellow

李世鹏教授是粤港澳大湾区数字经济研究院（IDEA）讲席科学家、SAIL实验室（苏州）常务董事。他曾任深圳市人工智能与机器人研究院（AIRS）首席科学家、执行院长，科大讯飞公司副总裁兼研究院联席院长，科通芯城集团CTO。他也是微软亚洲研究院（MSRA）的创始成员并曾担任副院长、首席研究员及多媒体计算组主任研究员。目前，他任中国电子学会常务理事、中国科技部人工智能产业技术创新战略联盟（AITISA）联合创始人兼联合秘书长。李世鹏教授本科和硕士毕业于中国科学与技术大学，并于1996年在Lehigh大学获得博士学位，在多媒体、物联网及人工智能领域颇具影响力，发表论文330余篇，被引用26,732次（H指数：85），并拥有美国专利205项。他在2011年获评IEEE fellow，并且在2020年被 Guide2Research 评为全球 1000 名顶尖计算机科学家之一。现在还是国际欧亚科学院（IEAS）院士和香港中文大学（深圳）和香港科技大学（广州）的兼职教授。李教授曾任IEEE核心学刊的总编辑和副总编辑；培养出4位MIT TR35创新奖获得者。

The Analogy of Human Cognition in GPT – The Dawn of AGI?

Shipeng Li

Abstract: The rapid development of GPT/ChatGPT, brings a new wave in AI development. But why it can enable intelligence emergence still largely remains as guess work today. Let’s pause for a moment to see if we can gain some hints to better understand the principles of GPT from the human cognition perspective. New discoveries in cognitive science could guide the direction of the research of artificial intelligence. With large-scale models, and large-scale real data collected in the human cognitive process, we may be able to emulate more precisely the human cognitive process. “Start like a baby, learn like a kid” (Credit: Prof. Josh Tenenbaum, MIT) may be a viable approach for us to better understand GPT. This talk will try to analyze the analogy of GPT and human cognition process, and point out GPT/ChatGPT not only represents the start of a new era, but also represents a path to artificial general intelligence.

Bio：

Dr. Shipeng Li is a Chair Scientist of International Digital Economy Academy (IDEA) and Managing Director of SAIL Labs (Suzhou). He was the Chief Scientist and Executive President of Shenzhen Institute of Artificial Intelligence and Robotics for Society (AIRS), Corporate Vice President and Co-President of Research of iFLYTEK, and the CTO of Cogobuy Group. He was a founding member of Microsoft Research Asia (MSRA) and served as the Research Area Manager. Dr. Li was the EiC (2018-2019) of IEEE Transactions on Circuits and Systems for Video Technology. He is quite influential in multimedia, IoT and artificial intelligence areas, and holds 205 US patents and published 330+ papers with 26,732 citations (H-index: 85). He was named as one of the top 1,000 computer scientists in the world by Guide2Research. He has trained 4 MIT TR35 Innovator Award Winners. Dr. Li received B.S. and M.S. from USTC and Ph.D. from Lehigh University. He is a Standing Director for the Chinese Institute of Electronics, the co-founder and Joint Secretary General of AI Industry Technology Innovation Strategic Alliance (AITISA) under China MOST. Dr. Li is an IEEE Fellow and an IEAS (International Eurasian Academy of Science) Academician, and an Adjunct Professor of the Chinese University of Hong Kong (Shenzhen) and Hong Kong University of Science and Technology (Guangzhou).