Coarse-to-Fine Sequential Monte Carlo for Probabilistic Programs

TL;DR

This paper introduces a method to transform probabilistic programs into coarse-to-fine versions that generate data progressively at increasing levels of detail, improving inference efficiency by leveraging model structure.

Contribution

It presents an algorithm for converting probabilistic programs into coarse-to-fine forms that preserve marginals and enhance inference efficiency.

Findings

Applied to Ising, depth-from-disparity, and factorial HMM models.

Preliminary evidence of improved inference efficiency.

Demonstrated the method's applicability to structured probabilistic models.

Abstract

Many practical techniques for probabilistic inference require a sequence of distributions that interpolate between a tractable distribution and an intractable distribution of interest. Usually, the sequences used are simple, e.g., based on geometric averages between distributions. When models are expressed as probabilistic programs, the models themselves are highly structured objects that can be used to derive annealing sequences that are more sensitive to domain structure. We propose an algorithm for transforming probabilistic programs to coarse-to-fine programs which have the same marginal distribution as the original programs, but generate the data at increasing levels of detail, from coarse to fine. We apply this algorithm to an Ising model, its depth-from-disparity variation, and a factorial hidden Markov model. We show preliminary evidence that the use of coarse-to-fine models can make existing generic inference algorithms more efficient.