Convergence of the Expectation Maximization algorithm revisited

by Dominikus Noll (IMT)

Tuesday Sep 8, 2026, 11:15 AM → 12:15 PM Europe/Paris

Salle K. Johnson (1R3, 1er étage)

The EM-algorithm assures monotone increase of the incomplete data likelihood, but does not in general guarantee convergence of the parameter estimates. We take a fresh look at the situation and present novel sufficient conditions for convergence that are conveniently verifiable in practical situations. Illustrations with typical applications of the EM-algorithm are given.

Key words: Kullback-Leibler divergence, Fisher information, exponential family, proximal point method, definability, information geometry