Tony Cai


Professor Tony Cai

Tony Cai received his Ph.D. from Cornell University in 1996 and is currently the Dorothy Silberberg Professor of Statistics at the Wharton School, a professor in the Applied Mathematics and Computational Science Grad-uate Group, and an Associate Scholar of the Depart-ment of Biostatistics and Epidemiology in the Perelman School of Medicine at the University of Pennsylvania. He is the recipient of the 2008 COPSS Presidents' Award, a Fellow and a Medallion Lecturer of the Institute of Mathematical Statistics. He is an associate editor of the Journal of the Royal Statistical Society, Series B, and a past Editor of the Annals of Statistics. 

His research interests include high dimensional statistics, large-scale multiple testing, nonparametric function estimation, functional data analysis, and statistical decision theory. 

Address:Department of Statistics The Wharton School University of Pennsylvania Philadelphia,

PA 19104, USA


Title:  An Integrative Framework for Two-Sample Sparse Inference

Abstract:  The conventional approach to 2-sample multiple testing is to first reduce the data matrix to a single vector of p-values and then choose a cutoff along the rankings to adjust for multiplicity. However, this inference framework often leads to suboptimal multiple testing procedures due to the loss of information in the data reduction step. In this talk, we introduce a new framework for two-sample multiple testing by constructing primary and auxiliary variables from the original observations and incorporating both in the inference procedure to improve the power. A data-driven multiple testing procedure is developed by employing a covariate-assisted ranking and screening (CARS) approach that optimally combines the information from both the primary and auxiliary variables.
The proposed CARS procedure is shown to be asymptotic valid with proper control of the false discovery rate (FDR). Numerical results confirm the effectiveness of CARS in FDR control and show that it achieves substantial power gain over existing methods.