Exact inference on Gaussian graphical models of arbitrary topology using path-sums

MaRDI QIDQ2810882zbMATH OpenFDO

Authors P.-L. Giscard, Z. Choo, S. J. Thwaite, Dieter Jaksch

Publication date 6 June 2016

Published in Journal of Machine Learning Research (JMLR) (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1410.7165, http://jmlr.csail.mit.edu/papers/v17/14-445.html

zbMATH Keywords

belief propagationblock matricescovariance matrixGaussian graphical modelsgraphs of arbitrary topologypath-sumwalk-sum

Mathematics Subject Classification ID

Multivariate analysis (62H99)

Abstract: We present the path-sum formulation for exact statistical inference of marginals on Gaussian graphical models of arbitrary topology. The path-sum formulation gives the covariance between each pair of variables as a branched continued fraction of finite depth and breadth. Our method originates from the closed-form resummation of infinite families of terms of the walk-sum representation of the covariance matrix. We prove that the path-sum formulation always exists for models whose covariance matrix is positive definite: i.e.~it is valid for both walk-summable and non-walk-summable graphical models of arbitrary topology. We show that for graphical models on trees the path-sum formulation is equivalent to Gaussian belief propagation. We also recover, as a corollary, an existing result that uses determinants to calculate the covariance matrix. We show that the path-sum formulation formulation is valid for arbitrary partitions of the inverse covariance matrix. We give detailed examples demonstrating our results.

Recommendations

• scientific article; zbMATH DE number 5957391
• Gaussian Markov distributions over finite graphs
• Graphical models, exponential families, and variational inference
• Closed-form Bayesian inference of graphical model structures by averaging over trees
• Positivity for Gaussian graphical models

Cited in

(6)

• Degenerate Gaussian factors for probabilistic inference
• scientific article; zbMATH DE number 5957391 (Why is no real title available?)
• Convergence analysis of distributed inference with vector-valued Gaussian belief propagation
• Elementary integral series for Heun functions: Application to black-hole perturbation theory
• Inference With Aggregate Data in Probabilistic Graphical Models: An Optimal Transport Approach
• Efficient distributed transfer learning for large-scale Gaussian graphic models

Describes a project that uses

Uses Software

• GMRFLib

This page was built for publication: Exact inference on Gaussian graphical models of arbitrary topology using path-sums

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2810882)