The University of Arizona

Low-dimensional couplings for Bayesian computation

Low-dimensional couplings for Bayesian computation

Series: Program in Applied Mathematics Colloquium
Location: Math 402
Presenter: Youssef Marzouk, Aeronautics and Astronautics, MIT

Integration against an intractable probability measure is among the fundamental challenges of statistical inference, particularly in the Bayesian setting. A useful approach to this problem seeks a deterministic coupling of the measure of interest with a tractable "reference" measure (e.g., a standard Gaussian). This coupling is induced by a transport map, and enables direct simulation from the desired measure simply by evaluating the transport map at samples from the reference. Yet characterizing such a map---e.g., representing, constructing, and evaluating it---grows challenging in high dimensions.

We will use the conditional independence structure of the target measure to establish the existence of certain low-dimensional couplings, induced by transport maps that are sparse or decomposable. We also describe conditions, common in Bayesian inverse problems, under which transport maps have a particular low-rank structure. Our analysis not only facilitates the construction of couplings in high-dimensional settings, but also suggests certain inference methodologies. For instance, in the context of nonlinear and non-Gaussian state space models, we will describe new variational algorithms for nonlinear smoothing and sequential parameter estimation. We will also outline a new class of nonlinear filters induced by local couplings, for inference in high-dimensional spatiotemporal processes with chaotic dynamics.

This is joint work with Alessio Spantini and Daniele Bigoni.

Department of Mathematics, The University of Arizona 617 N. Santa Rita Ave. P.O. Box 210089 Tucson, AZ 85721-0089 USA Voice: (520) 621-6892 Fax: (520) 621-8322 Contact Us © Copyright 2018 Arizona Board of Regents All rights reserved