Optimal Community Detection in Degree-Corrected Stochastic Block
主 题: Optimal Community Detection in Degree-Corrected Stochastic Block
报告人: Zongming Ma (University of Pennsylvania)
时 间: 2016-10-19 14:00-15:00
地 点: 理科一号楼1418
In this talk, we first present the minimax risk of community detection in degree-corrected stochastic block models. Guided by the minimax benchmark, we further introduce a polynomial-time algorithm that can be proven to achieve the optimal rates adaptively. This talk is based on joint work with Chao Gao, Anderson Zhang and Harrison Zhou. About the speaker: Zongming Ma is Associate Professor of Statistics at the Wharton School of the University of Pennsylvania. He received his Ph.D. from Stanford University in 2010 and B.S. from Peking University in 2005. His research interests include high-dimensional statistics, machine learning, network data analysis, and nonparametric statistics. He is a recipient of the Sloan Research Fellowship and NSF CAREER Award.