当前位置: 首 页 - 科学研究 - 学术报告 - 正文

304am永利集团、所2022年系列学术活动(第060场):邹青松 教授 中山大学

发表于: 2022-07-05   点击: 

报告题目:Two novel deep neural networks methods for high dimensional PDEs

报 告 人:邹青松 教授 中山大学

报告时间:2022年7月8日 星期五 9:00-10:00

报告平台:腾讯会议 ID: 169-740-404

会议联系人:王翔 wxjldx@jlu.edu.cn


报告摘要:In this talk, we will report two new deep NN methods for high order PDEs. The first one is the so-called adaptive neural networks method (ADN). By applying three adaptive techniques: adaptive activation function, adaptive loss and adaptive sampling, to the well-known DNN method PINN, we significantly improve the accuracy of the PINN method. Our second method is the so-called deep temporal difference methods (DTD). With this method, we first transform the deterministic parabolic PDE to a system of forward backward stochastic differential equation. Then by regarding this FBSDE as a Markov rewarding process, we use the Temporal Difference method in the reinforcement learning to train a neural network. Comparing to the deep stochastic method such as deep BSDE in the literature, our method can improve the accuracy and computational speed.


报告人简介:邹青松,中山大学计算机学院教授,博士生导师,数据科学系主任,广东省计算数学学会理事长,期刊International Journal of Numerical Analysis and Modelling编委。长期从事偏微分方程数值解法方面的研究工作,在包括SIAM J Numer Anal, Math Comp, Numer Math等在内的知名国际发表论文60余篇。 主要研究方向包括高阶高精度有限体积法(项目“高次有限体积法的构造和理论分析”获得2020年广东省自然科学奖二等奖),偏微分方程深度学习算法,以及药物设计和金融工程中的人工智能算法等。