Prof.Liu Hailiang（刘海亮教授）
爱荷华州立大学
Prof Liu Hailiang is a Professor of Mathematics and Computer Science at the lowa State University (ISU).
Abstract 摘要
We will present a partial differential equation framework for deep residual neural networksand for the associated learning problem. This is done by carrying out the continuum limitsof neural nefworks with respect to width and depth We study the well-posedness of theforward problem,and establish several optimal conditions for the inverse deep learningproblem. This talk concerns several mathematical aspects of deep learning and the uise ofoptimal control tools. This presentation is based on a joint work with Peter Markowich(KAUST)
Speaker Biography 讲者简介
Prof Liu Hailiang is a Professor of Mathematics and Computer Science at the lowa State University (ISU). He earned his Master degree in Applied Mathemnatics from TsinghuaUniversity (in 1988) and Ph.D. degree from the Chinese Academy (in 1995)He hasreceived many awards and honors,including the Alexander von Humboldt-ResearchFellowship (1996)and the inaugural Holl Chair in Applied Mathematics (2002-2012).Hisresearch interests include analysis of applied partial differential equations thedevelopmnent of novel high order algorithins for the approximate solution of theseproblems, and the interplay between analytical theory and coputational aspects of suchalgorithms with various applications. He has published more than 160 research papers.Hisrecent work has focused on data-driven optial control and learning problems.