CSE 599: Approximate Counting and Mixing Time of Markov chains

In this course we will study several techniques developed in the last 30 years to sample from sophisticated probability distributions of exponential size. Approximately half of the course will focus on classical techniques and the second half will study more recent developments.

Administrative Information

  • Instructor: Shayan Oveis Gharan
  • Office Hours: By appointment, email me at shayan at cs dot washington dot edu.
    Ziyun's OH: Thursdays 14:30-15:30 at Gates 152
  • Lectures: Monday - Wednesday 1:30 - 2:50 at CSE2 (Gates Building) 271

Assignments

Related Courses

Related Books/Monographs

Disclaimer: Several of these notes are borrowed from Kuikui Liu's course
Lecture Topic Notes Bonus Materials
Lecture 1 (09/25/24) Equivalence of Counting and Sampling pdf
Lecture 2 (09/30/24) Markov chains Intro, Metropolis Rule, Glauber Dynamics and Fundamental Theorem of Markov chains pdf
Lecture 3 (10/02/24) Introduction to Coupling pdf
(10/07/24) No Lecture: Instructor is traveling
Lecture 4 (10/09/24) Graph Coloring pdf
Lecture 5 (10/14/24) Functional Analysis of Markov Chains pdf
Lecture 6 (10/16/24) Canonical Path method and applications pdf
Lecture 7 (10/21/24) Sampling from the Ising Model pdf
Lecture 8 (10/23/24) Lower Bounds on Mixing Time pdf
Lecture 9 (10/28/24) Spectral Independence pdf
Lecture 10 (10/30/24) Spectral Independence Continued
Lecture 11 (11/04/24) Correlation decay and Sampling from the Hardcore Model
Lecture 12 (11/06/24) Entropic Independence and Near Linear time algorithm for the Hardcore Model
(11/11/24) No Lecture: Veternas Day
Lecture 13 (11/13/24) Trickledown Theorems
Lecture 14 (11/18/24) Intro to Real Stable Polyns
Lecture 15 (11/20/24) Matroids and Log Concave Polynomials
Lecture 16 (11/25/24) Stochastic Localization schemes
(11/27/24) No Lecture: Thanksgiving day
Lecture 17 (12/03/24) Cluster Expansion
Lecture 18 (12/05/24)