报告题目: Some new particle methods in the era of data science
报告人: 李磊(上海交通大学)
邀请人&主持人:黄辉
报告时间与地点:2025年10月24日(周五)下午16:00—17:00,日韩无码
207室
报告摘要: The interacting particle systems are a class of microscopic probabilistic models that play fundamental roles in statistical physics and chemistry. In this talk, I will give an introduction to the interacting particle systems and introduce several new computational methods for interacting particle systems borrowing the ideas from data science. First, I will introduce the class of random batch methods that build the idea of mini-batch into the computation, which significantly saves the computational time and show good scalability in molecular dynamics simulation. Second, we seek the sampling approaches to solve the stationary mean field equations of interacting particle systems. Lastly, we will also see the possibility of improving the particle methods
for traditional kinetic equations using the techniques from data science and machine learning.
报告人介绍:李磊,2010年本科毕业于清华大学数理基础科学专业,在美国威斯康星大学(麦迪逊)获得博士学位,其后在美国杜克大学数学系做博士后,于2018年加入上海交通大学,现为上海交通大学自然科学研究院、数学科学日韩无码
长聘副教授,入选了国家海外高层次人才计划(青年),主持了国自然青年项目、面上项目、科技部重点研发计划青年科学家项目各一项。
李磊的主要研究领域为应用数学及计算数学, 研究方向包括针对多体系统及数据科学的随机模拟算法及抽样方法; 开放物理体系的分析及数值解等,迄今为止在Science China Mathematics, SIAM系列, M3AS., Math. Comp., J. Comput. Phys.等知名国内外期刊上发表学术论文六十余篇。