The Lecture Committee is delighted to announce our first speaker for the inaugural David O. Siegmund Lecture. This special series of seminars was named for Professor Emeritus Siegmund to honor his accomplished career in methodological and applied statistical research, as well as his dedication as a mentor in his field. The Committee is excited to recognize and highlight these same merits in their selection for 2025:
Dr. Brian Reich is the Gertrude M. Cox Distinguished Professor of Statistics at North Carolina State University. His research interests include Bayesian methods, spatial statistics, extreme value analysis, variable selection and dimension reduction and machine learning. In addition to these methodological interests, Brian applies these methods to environmental areas such as ecology, epidemiology, meteorology, and climate. His work is funded by grants from the NIH, NSF and the NSA and he currently serves as Director of the NSF training grant “Uncertainty Quantification for the Life Sciences” and co-director of the NSF REU “Directed Research for Undergraduates in Mathematics and Statistics”. For this body of work Dr. Reich has won numerous awards including the American Statistical Association Outstanding Statistical Application Award and the LeRoy & Elva Martin Teaching Award at NC State.
Day 1: Leveraging deep learning for spatiotemporal interpolation
Gaussian Process (GP) models are the workhorse of spatial statistics. They provide provably optimal prediction and a robust framework for statistical inference. However, fitting GPs usually requires unrealistic assumptions such as stationarity, i.e., the process behaves similarly across the spatial domain, and even with this overly simplistic assumption the computation is not scalable. We consider a nonstationary model based on a latent dimension expansion and show that this model permits an arbitrarily precise spectral approximation. For computation, we place the model in the deep learning architecture and use variational Bayesian deep learning to fit the model to millions of observations in minutes. While the variational Bayesian approach is fast to fit, to ensure valid uncertainty quantification we develop a local conformal inference step. We study the theoretical properties of this approach, empirically compare performance against benchmarks and apply the method to spatiotemporal interpolation of global daily aerosol optical depth.
The David O. Siegmund Lecture Committee for 2025 are:
- Hua Tang, Professor of Genetics, Stanford University
- Paul Switzer, Professor Emeritus of Statistics, Stanford University
- Nancy Zhang, Professor of Statistics and Data Science, Wharton School University of Pennsylvania