Academic Report of School of Mathematical Sciences [2025] No. 133  

(Series Report for High-Level University Construction No. 1235)  

Title: Robust Multi-task Learning for Clustering and PCA  

Speaker: Associate Professor Haolei Weng (Michigan State University, USA)  

Time: 16:00–17:00, November 30, 2025 (Sunday)  

Location: Classroom 1, Huixing Building, Yuehai Campus, Shenzhen University  

Abstract:  

In this talk, we discuss a general EM-flavored multi-task learning approach to learn mixture models. Our approach not only can effectively utilize unknown similarity between related tasks but is also robust against a fraction of outlier tasks from arbitrary sources. We will focus on Gaussian mixture model to demonstrate our method and theory, and then briefly mention the extension to general mixture models. In the last part of the talk, we also discuss a suboptimal robustness issue for our approach and present our initial efforts to address this issue in the context of PCA.  

Speaker Biography:  

Haolei Weng is currently an Associate Professor at the Department of Statistics and Probability, Michigan State University. Prior to MSU, he completed his Ph.D. in statistics from Columbia University in 2017 and was a postdoctoral researcher at Princeton University in 2018. Before going to Columbia, he received a B.S. in statistics from University of Science and Technology of China. His research interests are broadly in the area of high-dimensional statistics and statistical machine learning.  

All faculty and students are welcome!  

Host: Hu Xianghong  

School of Mathematical Sciences  

November 21, 2025