Walk-Sum Interpretation and Analysis of Gaussian Belief Propagation

Abstract

This paper presents a new framework based on walks in a graph for analysis and inference in Gaussian graphical models. The key idea is to decompose correlations between variables as a sum over all walks between those variables in the graph. The weight of each walk is given by a product of edgewise partial correlations. We provide a walk-sum interpretation of Gaussian belief propagation in trees and of the approximate method of loopy belief propagation in graphs with cycles. This perspective leads to a better understanding of Gaussian belief propagation and of its convergence in loopy graphs.

Cite

Text

Malioutov et al. "Walk-Sum Interpretation and Analysis of Gaussian Belief Propagation." Neural Information Processing Systems, 2005.

Markdown

[Malioutov et al. "Walk-Sum Interpretation and Analysis of Gaussian Belief Propagation." Neural Information Processing Systems, 2005.](https://mlanthology.org/neurips/2005/malioutov2005neurips-walksum/)

BibTeX

@inproceedings{malioutov2005neurips-walksum,
  title     = {{Walk-Sum Interpretation and Analysis of Gaussian Belief Propagation}},
  author    = {Malioutov, Dmitry and Willsky, Alan S. and Johnson, Jason K.},
  booktitle = {Neural Information Processing Systems},
  year      = {2005},
  pages     = {579-586},
  url       = {https://mlanthology.org/neurips/2005/malioutov2005neurips-walksum/}
}