Research

Research

Google Scholar has an up-to-date list of my papers.

My research focuses on Bayesian statistics, scalable computation, and interdisciplinary applications. I have worked on foundational problems in Bayesian inference, including shrinkage priors, sparse inference, and multiple testing, and I was a co-inventor of the horseshoe prior, now widely used in Bayesian statistics and machine learning. I also co-developed the Pólya–Gamma data augmentation framework for fully Bayesian logistic regression. My applied work spans real-time disease forecasting, maternal and child health, telemedicine access, and public health and policy. Current projects focus on the statistical reliability of machine-learning methods, especially generative models and neural posterior estimation.

Bayesian statistics and computation

My methodological work includes Bayesian shrinkage, sparse inference, multiple testing, high-dimensional regression, and scalable posterior computation. A recurring theme is the development of Bayesian methods that preserve uncertainty while remaining usable in realistic data-analysis settings.

Statistical machine learning

Recent work studies the reliability of machine-learning methods, especially generative models, simulation-based inference, neural posterior estimation, and methods whose behavior depends on the fit between a statistical model and the real data-generating process.

Applied work

I collaborate on projects in health care, infectious disease, maternal and child health, telemedicine access, policy evaluation, security, neuroscience, finance, management, astronomy, molecular biology, political science, and linguistics. These projects typically involve observational data, uncertainty quantification, causal or policy-relevant claims, and careful communication of empirical evidence.

Selected topics

  • Bayesian shrinkage priors and sparse inference
  • Multiple testing and false-discovery-rate methods
  • Pólya–Gamma data augmentation and Bayesian logistic regression
  • Statistical reliability of generative models and simulation-based inference
  • Public-health applications, including disease forecasting and telemedicine access
  • Data science in business, policy, and interdisciplinary research

Ph.D. students and alumni

I’ve had the privilege to work with some very smart, dedicated, and wonderful Ph.D. students over the years. Current advisees and alumni include: