报告人:刘扬 副教授(中国科学院大学工程科学学院)
报告内容:
Harnessing data to model complex physical systems has become a critical scientific problem in many science and engineering areas. The state-of-the-art advances of AI (in particular deep learning, thanks to its rich representations for learning complex nonlinear functions) have great potential to tackle this challenge, but in general (i) rely on a large amount of rich data to train a robust model, (ii) have generalization and extrapolation issues, and (iii) lack of interpretability and explainability, with little physical meaning. To bridge the knowledge gaps between AI and complex physical systems in the sparse/small data regime, this talk will introduce the integration of bottom-up (data-driven) and top-down (physics-based) processes through a new AI-enabled computing paradigm for modeling, simulation and discovery of complex physical systems. This talk will discuss knowledge embedding and discovery in scientific computing, and show examples on data-driven modeling/discovery of nonlinear PDEs that govern the behavior of complex physical systems, e.g., wave propagation, reaction-diffusion processes, fluid flows, etc.
报告人简介:
刘扬,中国科学院大学工程科学学院“长聘教轨副教授、博导”(2022 年 3 月至今),国家自然科学基金委优秀青年基金(海外)获得者。先后取得河海大学土木工程本科学位(2006)、哥伦比亚大学工程力学硕士(2011、2014)与博士学位(2015),随后在麻省理工学院航空航天系从事博士后研究(2015-2017)。近年来致力于人工智能融合科学计算前沿研究,提出了成套物理驱动/编码深度学习法,成功解决在小训练样本下复杂多物理场系统建模、非线性偏微分方程求解、物理定律探索、材料特性表征等交叉领域关键科学问题,实现了深度学习的物理可诠释性和优异的外推与泛化能力。此外,研究成果完善了离散-连续多尺度计算理论和方法,揭示了复杂材料非线性变形和破坏机理。在国际顶级 SCI 期刊(如 Nature Communications、Nature Machine Intelligence、CMAME)和计算机顶级会议(如 IJCAI、ICLR)共发表/录用 26 篇论文、合作专著章节 2 篇,其中第一或通讯文章 18 篇、ESI 高引论文 2 篇,Google Scholar 引用 1250 余次;主持国家高层次人才计划项目、中科院百人计划项目、中央军委某工程重点项目等项目共计 1000 余万元。受邀到哈佛大学、圣母大学等世界名校做学术报告,荣获 USNCCM’2015 最佳论报告奖。
主持人:袁子峰 研究员(北京大学应用物理与技术研究中心)
时 间:2023年3月2日(周四)12:20
地 点:北京大学工学院1号楼210会议室
|