Approximate Message Passing Algorithms

Explore the fundamentals of Approximate Message Passing (AMP) algorithms in this comprehensive lecture by Cynthia Rush from Columbia University. Delve into the computational complexity of statistical inference, understanding the motivation behind AMP and its applications. Learn about the asymptotic regime, abstract approach, and key concepts such as state evolution and pseudolipschitz functions. Examine the theoretical aspects, including convergence properties and joint distribution analysis. Conclude with an in-depth look at rank 1 matrix estimation, gaining valuable insights into this powerful statistical inference technique.

Intro
Tutorial Article
What is Approximate Message Passing
Why Approximate Message Passing
Asymptotic Regime
Abstract Approach
Outline
State evolution
Pseudolipschitz function
Theory
Convergence
Joint Distribution
Rank 1 Matrix Estimation