David B. Dunson

Arts and Sciences Distinguished Professor
Department of Statistical Science
dunson@stat.duke.edu
Curriculum Vitæ

Research Interests

Bayesian statistical methodology motivated by complex biomedical data and machine learning
applications. Ongoing methodologic research focuses on nonparametric Bayes, latent variable
methods, big data, scalable Bayesian inferences, functional and object data, and dimensionality
reduction. An emphasis is on developing adaptive Bayes approaches for "learning" low-dimensional
structure underlying high-dimensional "objects" ( images, surfaces, shapes, text, array data,
contingency table ). This work involves inter-discplinary thinking at the intersection of
statistics, geometry and computer science. The motivation comes from applications in epidemiology,
environmental health, neurosciences, genetics, and other settings (music, fine arts, humanities, etc).