COS597C: Bayesian Nonparametrics


Nonparametric Bayesian methods are an increasingly important tool in applied machine learning and statistics. We will focus on the Dirichlet process (DP) and its cousins. Topics include the original development of the DP, the suite of approximate inference techniques that have made it practical for data analysis, and recently developed variants.

Readings are posted on the syllabus.

Time and Location

Monday 1:30PM-4:20PM, location CS 402

Course Staff


David M. Blei
204 CS Building
blei [at]
Office hours: TBA


All students are expected to be comfortable with probability and basic statistics. Please see Prof. Blei if unsure whether you meet the requirements.

Course Grades and Workload

The course consists of readings, lectures, and discussions. The requirements for everyone (including auditors) are to do the reading, participate in class, and scribe for at least one class. For those taking the course for a grade, the requirements include a final project or report.

Mailing list

Please join the course mailing list by visiting here and following the instructions for subscribing.

This list will be used for general announcements and discussion. You can post to the list by sending mail to Note that you can only post to the list using the email address you used to subscribe to it.