Summer School

New tools for optimal mixing of Markov chains: Spectral independence and entropy decay

Monday, August 8 - Friday, August 12, 2022.

Set on the beautiful oceanside UC Santa Barbara campus, this summer school will be an in-depth tutorial on exciting and powerful new tools for establishing optimal mixing time bounds for Markov Chain Monte Carlo (MCMC) algorithms. The focus of the school will be on understanding the spectral independence technique which utilizes ideas from high-dimensional expanders, and related approaches for establishing entropy decay. Notable examples of recent work that will be explained at the summer school include generating random bases of a matroid and sampling from spin systems in the correlation decay region.

The summer school is designed for graduate students with some basic knowledge of Markov chains who are interested in a comprehensive understanding of these techniques and in exploring future directions.

Lecture videos and notes

Detailed program and timetable

Folder of all lecture videos

Supplemental notes


View from hotel

Housing and meals will be provided for students (on campus at the Manzanita Village, next to the ocean cliffs). There is limited travel support available. Faculty and postdocs are welcome to attend; a block of hotel rooms are reserved at the on-campus hotel The Club.

View from Manzanita dorms

Interested? Expired AJO application site

Supported by the University of California Santa Barbara and the NSF.

Organizers: Daniel Stefankovic, Prasad Tetali, and Eric Vigoda.