报告人： 夏志宏教授（美国西北大学）
报告时间： 2024年4月18日（周四）上午 10：00—11:00
报告地点： 数学楼一楼报告厅
报告摘要：We propose a novel machine learning algorithm inspired by complex analysis. Our algorithm has a better mathematical formulation and can much better represent functions. Its universal approximation property can be easily derived. The algorithm can be implemented in two self-learning neural networks: The CauchyNet and the X-Net. The CauchyNet is very efficient for low-dimensional problems such as extropolation, imputation, numerical solutions of PDEs and ODEs. The X-Net works for image and voice recognition, transformer and large language models. As examples, our algorithm is much more efficient than the current popular PINN models for various scientific computations, and for a set of medical image we tested, it can increase accuracy from 88% to 98%.
Our algorithm is new and it is yet to be tested on large complex problems.
邀请人：曹永罗