Langevin Incremental Mixture Importance Sampling

TL;DR

This paper introduces a new importance sampling method that uses local gradient information to construct efficient mixture components, improving scalability and flexibility for high-dimensional and non-log-concave targets.

Contribution

It develops a novel importance sampler leveraging local approximations via differential equations, enhancing efficiency and applicability over existing methods.

Findings

Performs well on high-dimensional non-Gaussian densities

Handles non-log-concave targets effectively

Demonstrates improved efficiency in Bayesian logistic regression

Abstract

This work proposes a novel method through which local information about the target density can be used to construct an efficient importance sampler. The backbone of the proposed method is the Incremental Mixture Importance Sampling (IMIS) algorithm of Raftery and Bao (2010), which builds a mixture importance distribution incrementally, by positioning new mixture components where the importance density lacks mass, relative to the target. The key innovation proposed here is that the mixture components used by IMIS are local approximations to the target density. In particular, their mean vectors and covariance matrices are constructed by numerically solving certain differential equations, whose solution depends on the gradient field of the target log-density. The new sampler has a number of advantages: a) it provides an extremely parsimonious parametrization of the mixture importance density, whose configuration effectively depends only on the shape of the target and on a single free parameter representing pseudo-time; b) it scales well with the dimensionality of the target; c) it can deal with targets that are not log- concave. The performance of the proposed approach is demonstrated on a synthetic non-Gaussian multimodal density, defined on up to eighty dimensions, and on a Bayesian logistic regression model, using the Sonar data-set. The Julia code implementing the importance sampler proposed here can be found at https:/github.com/mfasiolo/LIMIS.