English

Duality for Continuous Graphical Models

Methodology 2021-11-04 v1 Information Theory math.IT Computation Machine Learning

Abstract

The dual normal factor graph and the factor graph duality theorem have been considered for discrete graphical models. In this paper, we show an application of the factor graph duality theorem to continuous graphical models. Specifically, we propose a method to solve exactly the Gaussian graphical models defined on the ladder graph if certain conditions on the local covariance matrices are satisfied. Unlike the conventional approaches, the efficiency of the method depends on the position of the zeros in the local covariance matrices. The method and details of the dualization are illustrated on two toy examples.

Keywords

Cite

@article{arxiv.2111.01938,
  title  = {Duality for Continuous Graphical Models},
  author = {Mehdi Molkaraie},
  journal= {arXiv preprint arXiv:2111.01938},
  year   = {2021}
}

Comments

Proc. of the 2021 IEEE Information Theory Workshop (ITW2021), Kanazawa, Japan