Time: Tuesday, April 21, 2026, 3:30-5:00 p.m.

Venue: Guoshun Campus, Starr Building, Room 603

Title: Debiased Bayesian Inference for High-dimensional Regression Models

Speaker: Associate Professor Qihui Chen, School of Economics and Management, CUHK-Shenzhen

Moderator: Associate Professor Xuewen Yu

Abstract: There has been significant progress in Bayesian inference based on sparsity-inducing priors, such as spike-and-slab and horseshoe-type priors, for high-dimensional regression models. However, the resulting posterior distributions generally do not possess desirable frequentist properties, and the corresponding credible sets therefore cannot serve as valid confidence sets, even asymptotically. We introduce a novel debiasing approach that corrects the bias of the entire Bayesian posterior distribution. We establish a new Bernstein-von Mises theorem that guarantees the frequentist validity of the debiased posterior. We demonstrate the practical performance of our proposed method through Monte Carlo simulations and two empirical applications in economics.

Bio: Qihui Chen is an Associate Professor in the School of Economics and Management at CUHK-Shenzhen, and serves as Director of the MSc Programme in Economics at the Shenzhen Finance Institute. He received his PhD in Economics from the University of California, San Diego in 2017. His research interests include econometrics, machine learning, asset pricing, and related fields. His work has appeared in leading economics and statistics journals such as Quantitative Economics, Journal of Econometrics, Econometric Theory, and Journal of the American Statistical Association.