极速赛车官网

当前位置: 极速赛车官网 > 极速赛车官网 > 学术活动 > 正文

Learning PDEs/SPDEs via neural networks based on regularity structure

发布日期:2024-12-17点击数:

报告人:陈炳光 讲师(福建师范大学)

时间:2024年12月20日 10:00-

地点:极速赛车官网 LD402


摘要:Artificial intelligence has brought technological innovation to various fields and achieved remarkable results, especially in modeling dynamic systems. In this talk, basic knowledge about deep learning will be presented.  We propose neural operator networks based on regularity structure theory to model PDEs/SPDEs. The key step of our network structure is to project the stochastic noise and initial values to the model feature vectors. The experimental results show that our networks significantly improve the accuracy of solutions and speed up inference time. Based on joint work with Qi Meng, Shiqi Gong, Peiyan Hu, etc.


邀请人:杨寰宇


欢迎广大师生积极参与!



关于我们
极速赛车官网 的前身是始建于1929年的重庆大学理极速赛车官网 和1937年建立的重庆大学商极速赛车官网 ,理极速赛车官网 是重庆大学最早设立的三个极速赛车官网 之一,首任院长为数学家何鲁先生。