各位老师同学好, 以下发布一则学术报告通知, 欢迎各位老师同学参加, 谢谢！
报告题目: Data-driven Learning of Generalized Langevin Equations with State-dependent Memory
报告人：葛沛 博士后 伍斯特理工学院 数学学院
报告时间：2025年6月30日下午14:00
报告地点：红瓦楼726
报告内容简介：The generalized Langevin equation (GLE) is a common approach to construct coarse-grained models of molecular systems. However, the model is based on a homogeneous kernel to account for the memory effect, which, however, shows limitations in capturing the heterogeneous energy dissipation arising from the unresolved dynamics. Here, we propose a data-driven approach to construct GLE with a state-dependent non-Markovian memory. The main idea is to seek a generalized representation of the memory in terms of a set of state-dependent features, which can be efficiently learned from the statistics of the CG variables. We demonstrate how to introduce the stochastic noise so that the second fluctuation-dissipation theorem is exactly satisfied. Numerical results on a molecular system show that state-dependent memory plays a profound role in molecule kinetics such as the transition rate.
报告人简介：葛沛，博士，2019年毕业于东南大学数学学院，获学士学位。2025年毕业于美国密歇根州立大学计算数学科学与工程系，获博士学位，导师为Huan Lei教授。即将赴美国伍斯特理工学院担任博士后。主要研究方向为多尺度建模，通过建立从微观到宏观的可解释映射，从而利用机器学习构造可信的保结构的宏观模型。相关研究成果发表在《Physical Review Letters》、《The Journal of Chemical Physics》等期刊。
报告邀请人：方礼冬