Stochastic AUC Maximization Algorithms for Streaming Data

报告题目: Stochastic AUC Maximization Algorithms for Streaming Data

报告人：Professor Yiming Ying（State University of New York (SUNY) at Albany, USA）

时间: 5月24日（周四）下午2:30-4:00

地点：工商楼200-9报告厅

摘要: Stochastic optimization algorithms such as stochastic gradient descent (SGD) update the model sequentially with cheap per-iteration costs, making them amenable for large-scale streaming data analysis. However, most of the existing studies focus on the classification accuracy which can not be directly applied to the important problems of maximizing the Area under the ROC curve (AUC) in imbalanced classification and bipartite ranking. In this talk, I will talk about our recent work on developing novel stochastic optimization algorithms for AUC maximization. Compared with the previous literature which requires high storage and per-iteration costs, our algorithms have both space and per-iteration costs of one datum and achieve optimal convergence rates. 

报告人简介:  Yiming Ying is currently a tenured Associate Professor in the Department of Mathematics and Statistics at the State University of New York (SUNY) at Albany, USA. Before that, he was a Lecturer (Assistant Professor) in Computer Science at the University of Exeter (UK).  He received the PhD degree in Mathematics in 2002 from Zhejiang University.  His research interests center on data science including learning theory and machine learning,  and their applications in big data analysis.  His research has been supported by EPSRC (UK), Simons Foundation (USA) and Department of Energy (USA).

联系人：郭正初（guozhengchu@zju.edu.cn）