Automatic Rao-Blackwellization for Sequential Monte Carlo with Belief Propagation
Machine Learning
2023-12-18 v1 Computation
Abstract
Exact Bayesian inference on state-space models~(SSM) is in general untractable, and unfortunately, basic Sequential Monte Carlo~(SMC) methods do not yield correct approximations for complex models. In this paper, we propose a mixed inference algorithm that computes closed-form solutions using belief propagation as much as possible, and falls back to sampling-based SMC methods when exact computations fail. This algorithm thus implements automatic Rao-Blackwellization and is even exact for Gaussian tree models.
Keywords
Cite
@article{arxiv.2312.09860,
title = {Automatic Rao-Blackwellization for Sequential Monte Carlo with Belief Propagation},
author = {Waïss Azizian and Guillaume Baudart and Marc Lelarge},
journal= {arXiv preprint arXiv:2312.09860},
year = {2023}
}