An Introduction to Online Convex Optimization

报告题目:An Introduction to Online Convex Optimization

报告人:袁德明 教授 博导

报告时间:2023年12月2日9:00

报告地点:逸夫楼1楼报告厅(原模拟法庭)

报告对象:计算机科学与技术学院研究生、本科生及其他感兴趣师生

报告内容:

In this talk we will present a detailed review of the recent advance in online convex optimization. First, we introduce the problem formulation and the motivations of online convex optimization. Second, we present several state-of-the-art algorithms, such as online gradient descent, follow-the-leader, online mirror descent, etc. In addition, the detailed convergence results are presented and discussions about the parameters are also given. Third, we talk about the bandit setting which is widely studied in the machine learning community. Finally, we point out several possible future research directions。

报告人简介:

袁德明,于2007年7月、2012年6月毕业于南京理工大学,分别获工学学士和工学博士学位;现为南京理工大学自动化学院教授,博士生导师,澳大利亚“奋进”研究学者。主要从事分布式优化、机器学习与控制相关研究。近年来,在IEEE Transactions on Information Theory、IEEE Transactions on Automatic Control、Automatica、SIAM Journal on Control and Optimization等学术刊物发表论文数篇。2020年获国家自然科学基金优秀青年基金资助,2017年获江苏省自然科学基金优秀青年基金资助,同时主持并参与多项国家和江苏省自然科学基金,2022年获中国自动化学会自然科学奖二等奖(排一)。目前担任国际刊物Journal of the Franklin Institute、Transactions of the Institute of Measurement and Control、Franklin Open编委。