Expectation maximization as message passing

TL;DR

This paper demonstrates how the expectation maximization (EM) algorithm can be interpreted and implemented as a message passing process within factor graphs, providing a new perspective on EM's operation.

Contribution

It introduces a novel interpretation of EM as message passing in factor graphs, connecting two important concepts in probabilistic inference.

Findings

EM can be formulated as message passing in factor graphs

Provides a unified view of EM and message passing algorithms

Potentially simplifies implementation of EM in graphical models

Abstract

Based on prior work by Eckford, it is shown how expectation maximization (EM) may be viewed, and used, as a message passing algorithm in factor graphs.