Mirror Descent with Relative Smoothness in Measure Spaces, with application to Sinkhorn and EM

Many problems in machine learning can be formulated as optimizing a convex functional over a space of measures. This paper studies the convergence of the mirror descent algorithm in this infinite-dimensional setting. Defining Bregman divergences through directional derivatives, we derive the convergence of the scheme for relatively smooth and strongly convex pairs of functionals. Applying our result to joint distributions and the Kullback–Leibler (KL) divergence, we show that Sinkhorn's primal iterations for entropic optimal transport in the continuous setting correspond to a mirror descent, and we obtain a new proof of its (sub)linear convergence. We also show that Expectation Maximization (EM) can always formally be written as a mirror descent, and, when optimizing on the latent distribution while fixing the mixtures, we derive sublinear rates of convergence.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
11/02/2020

Homeomorphic-Invariance of EM: Non-Asymptotic Convergence in KL Divergence for Exponential Families via Mirror Descent

Expectation maximization (EM) is the default algorithm for fitting proba...
research
06/30/2020

Sinkhorn EM: An Expectation-Maximization algorithm based on entropic optimal transport

We study Sinkhorn EM (sEM), a variant of the expectation maximization (E...
research
04/28/2019

Blended Matching Pursuit

Matching pursuit algorithms are an important class of algorithms in sign...
research
05/24/2019

A view of Estimation of Distribution Algorithms through the lens of Expectation-Maximization

We show that under mild conditions, Estimation of Distribution Algorithm...
research
09/07/2022

Conformal Mirror Descent with Logarithmic Divergences

The logarithmic divergence is an extension of the Bregman divergence mot...
research
06/07/2023

A Mirror Descent Perspective on Classical and Quantum Blahut-Arimoto Algorithms

The Blahut-Arimoto algorithm is a well known method to compute classical...

Please sign up or login with your details

Forgot password? Click here to reset