当前位置:首页>学术报告
Solving PDEs by deep neural networks

 
Title:
Solving PDEs by deep neural networks
Speaker:
陈景润 教授,中国科学技术大学数学科学学院/苏州高等研究院
Inviter: 刘晓东 副研究员
Time & Venue:

2022.5.10 10:00 腾讯会议号:563-747-902

Abstract:

Solving partial differential equations (PDEs) by deep neural networks has attracted significant attentions in recent years. In this presentation, I will discuss two problems related to this topic: (1) how to capture shock waves with (high-dimensional) random inputs? (2) how to find the global optimizer of highly nonconvex objective functions arising from machine learning tasks? These are joint works with Shi Jing (SJTU) and Liyao Lyu (MSU).

Affiliation:  

学术报告中国科学院数学与系统科学研究院应用数学研究所
地址 北京市海淀区中关村东路55号 思源楼6-7层 南楼5-6、8层 邮编:100190 电子邮箱:iam@amss.ac.cn
@2000-2022 京ICP备05058656号-1