A Unifying Tutorial on Approximate Message Passing

Abstract

Over the last decade or so, Approximate Message Passing (AMP) algorithms have become extremely popular in various structured high-dimensional statistical problems. Although the origins of these techniques can be traced back to notions of belief propagation in the statistical physics literature, our goals in this work are to present the main ideas of AMP from a statistical perspective and to illustrate the power and flexibility of the AMP framework. Along the way, we strengthen and unify many of the results in the existing literature.