报告题目：Structure-preserving machine learning moment closure models for the radiative transfer equation
报 告 人：黄俊涛 博士 密歇根州立大学
报告时间：2021年11月11日 10:00-11:00
报告地点：腾讯会议 ID：751 765 472
会议链接：https://meeting.tencent.com/dm/p8zOgEmEA4KM
校内联系人：陶詹晶 zjtao@jlu.edu.cn
报告摘要：In this talk, we present our work on structure-preserving machine learning (ML) moment closure models for the radiative transfer equation. Most of the existing ML closure models are not able to guarantee the stability, which directly causes blow up in the long-time simulations. In our work, with carefully designed neural network architectures, the ML closure model can guarantee the stability (or hyperbolicity). Moreover, other mathematical properties, such as physical characteristic speeds, are also discussed. Extensive benchmark tests show the good accuracy, long-time stability, and good generalizability of our ML closure model.
报告人简介：Juntao Huang is currently working as a postdoctoral fellow at Michigan State University, after he obtained Ph.D. degree in applied math in 2018 at Tsinghua University in China. His current research interests focus on the design and analysis of numerical methods for PDEs and, more recently, using machine learning to assist traditional scientific computing tasks. Topics of special interests include adaptive sparse grid discontinuous Galerkin (DG) methods, structure-preserving machine learning moment closures for kinetic models, structure-preserving time discretizations for hyperbolic equations, and boundary schemes for the lattice Boltzmann method.
