勷勤数学•专家报告

题      目：Machine Learning for Engineering Computations–Data-driven Approach and Neural Network Solvers

报  告  人：张庆辉 教授  (邀请人：钟柳强)

                                   暨南大学

时      间：1月7日  10：00-11：00

地     点：数科院西楼111报告厅

报告人简介:

       张庆辉，哈尔滨工业大学（深圳）教授，博士生导师。中山大学博士毕业，美国雪城大学博士联合培养，曾在香港大学工学院从事博士后研究工作。研究方向为广义有限元法，无网格方法，机器学习等。在Numerische mathematik,CMAME等杂志发表多项研究成果。主持国家基金面上、青年项目、广东省自然科学杰出青年基金项目、国家基金重大研究计划项目集成项目子课题，联合主持国家基金海外与港澳学者合作研究项目等。

摘      要：

       We report research developments on machine Learning (ML) methods for Engineering Computations. Recently, the ML techniques have gained extensive attention for sciences and engineering computations because of their abilities to deal with nonlinearities, complex geometries, and high-dimensional problems. Our research efforts about the ML techniques are concentrated on three aspects: data driven surrogate models (model reductions), neural network solvers for PDEs, and coupling the ML methods with traditional numerical techniques. The concrete examples include deep learning-based surrogate model for flight load analysis, neural network methods for solving singularity problems and high-dimensional problems, (generalized/eXtended) FEM coupled with neural networks for interface problems.

          欢迎老师、同学们参加、交流！