Winter 2019 STAT 41510

Course: STAT 41510
Title: Bayesian Nonparametrics
Instructor(s): Chao Gao
Class Schedule: Sec 01: MW 9:30 AM–10:50 AM in Jones 226
Office Hours:
Textbook(s): Ghosal and van der Vaart, Fundamentals of Nonparametric Bayesian Inference
Description: Bayesian nonparametric methods are increasingly important tools in machine learning and statistics. We will discuss nonparametric Bayesian approaches to mixture models, latent feature models, hierarchical models, network models and high-dimensional regression models. Topics that will be covered include Dirichlet process, Chinese restaurant process, Pitman-Yor process, Indian buffet process, Gaussian process, and their computational techniques via Gibbs sampling and variational inference. Frequentist evaluations of posterior distributions will also be discussed in nonparametric and high-dimensional settings.

Prerequisite(s): STAT 30200