Seminar| Institute of Mathematical Sciences
Time: Tuesday, June 23th, 2026,16:00-17:00
Location: IMS, R408
Speaker: 罗民兴, 中国科学院院士, 北京计算科学中心讲座教授⛹🏻♂️、主任
Abstract:报告简要回顾AI的发展史，指出由大数据🤧👩🏻🍼、大算力和标度律驱动的大模型正引领人类走向AIGC时代🕷。在科研领域🤏，AI推动数学与理论物理研究从竞赛解题向证明全新定理⏬🦹🏼、从独立证明向AI导航的范式转变👩🏻🎓。面对AI黑箱☂️，报告讨论物理学家的人工智能观👨🏼🚒，尝试用物理工具从微观🔎、介观及宏观(基准测试)层面理解Al🧕🏼。最后，报告展望由“管理者-教师-学生-人工智能”协同共创的未来教育新生态。
The talk will briefly recall the evolution of AI, highlighting how large models driven by big data, massive computing power, and scaling laws are ushering humanity into the era of AIGC (Artificial Intelligence Generated Content). In scientific research, AI is driving a paradigm shift in mathematics and theoretical physics—from solving Olympiad-level competition problems to proving entirely new theorems, and from unassisted human proof to AI-assisted theorem proving. To address the “black box” of AI, the talk explores physicists’ perspectives on the technology, drawing on physical concepts to understand AI at the microscopic, mesoscopic, and macroscopic (benchmarking) levels. Finally, the talk outlines a vision for a future educational ecosystem shaped through the collaboration of policymakers, teachers, students, and AI.
演讲人介绍：罗民兴是北京计算科学研究中心讲座教授和主任💂🏻♀️。他的研究兴趣包括量子场论🚖、高能物理和高能量密度物理🚜。他本科在浙江大学物理系就读并于1983年获学士学位🧑🏻🍼，随后赴宾夕法尼亚大学物理系深造并于1990年获博士学位。他先后在宾夕法尼亚大学和华盛顿大学做博士后工作;1994年至2020年他任职于浙江大学物理系，历任副教授、教授👼🏼、讲座教授和理学部主任;罗民兴是中国科学院院士⏯。