报告人：陈小丽 教授（中国地质大学（武汉））

时间： 2026年05月29日 21:00-

腾讯会议ID: 818 927 330

摘要：In this talk, I will explore the use of machine learning techniques to model stochastic systems and derive reduced-order models for complex, dissipative dynamics. First, I will introduce how to combine Physics-Informed Neural Networks with sample observational data to learn stochastic differential equations driven by Brownian and Lévy noise. I will then present a novel framework, Stochastic OnsagerNet, which leverages the Onsager principle to learn closure relationships in dissipative systems out of equilibrium. This general machine learning approach enables the construction of reduced models for noisy, dissipative dynamics, making it applicable to a wide range of scientific and technological problems. To demonstrate its utility, I will apply the method to model the folding and unfolding behavior of a long polymer chain in an external field—an important problem in polymer rheology. However, the methodology is versatile and can be extended to a broad spectrum of complex dynamical systems across various domains.

邀请人：吴风艳       魏崴

欢迎广大师生积极参与！