2025京师调和分析论坛(2) (2025.1.25)
会议安排
Zoom会议号：825 615 69480，入会密码： 654321
时间 |
报告人 |
报告题目 |
单位 |
8:00-8:55 |
Sergey Tikhonov |
Fourier Inequalities in Various Function Spaces |
Centre de Recerca Matemàtica, Barcelona, Spain |
9:00-9:55 |
Han Feng |
An Introduction to Machine Learning and its Mathematical Framework for Error Analysis |
City University of Hong Kong, China |
10:00-10:55 |
Bin Han |
Numerical Methods for Elliptic Interface Problem |
University of Alberta, Canada |
本会议由以下项目支持:
北京师范大学优秀青年创新团队项目“调和分析及其在偏微分方程中的应用”
北京师范大学教育部&科技部“调和分析及其应用创新引智基地”
国家自然科学基金重点项目“相关于球平均、Chirp函数和区域上算子的函数空间公理化实变理论及其应用”
会议组委会：戴峰(本次会议组织者), 杨大春, 袁文, 张阳阳, 周渊
2025年1月21日
报告信息
1. 报告人: Sergey Tikhonov (Centre de Recerca Matemàtica, Barcelona, Spain)
报告题目: Fourier Inequalities in Various Function Spaces
报告摘要: I plan to start by recalling several classical inequalities for Fourier series and transforms. The main goal of the talk will be to discuss new nonclassical Fourier estimates.
2. 报告人: Han Feng (City University of Hong Kong, China)
报告题目: An Introduction to Machine Learning and its Mathematical Framework for Error Analysis
报告摘要: The field of machine learning has revolutionized the way we approach complex problems and make intelligent decisions. In this talk, I am going to first give an introduction to machine learning with a journey through the key concepts, techniques, and methodologies that shape the field. Subsequently, I will focus on the theoretical frameworks that govern error analysis, including bias-variance tradeoffs, overfitting and underfitting, and model complexity. We will talk about the mathematical framework that enables us to understand and quantify the performance of machine learning algorithms. This includes a look at empirical risk minimization, approximation ability, and the role of hypothesis spaces generated by convolutional neural networks and graph convolutional networks.
3. 报告人: Bin Han (University of Alberta, Canada)
报告题目: Numerical Methods for Elliptic Interface Problem
报告摘要:
