Belief Update in CLG Bayesian Networks With Lazy Propagation

TL;DR

This paper introduces an exact belief update architecture for Conditional Linear Gaussian Bayesian Networks that leverages lazy propagation and graph structure to improve efficiency, demonstrated through examples and preliminary performance results.

Contribution

It extends lazy propagation methods to CLG Bayesian Networks, exploiting independence and irrelevance for more efficient belief updating.

Findings

Significant potential shown in preliminary empirical evaluation.

Architecture effectively decomposes potentials into factors.

Utilizes graph structure to optimize belief updates.

Abstract

In recent years Bayesian networks (BNs) with a mixture of continuous and discrete variables have received an increasing level of attention. We present an architecture for exact belief update in Conditional Linear Gaussian BNs (CLG BNs). The architecture is an extension of lazy propagation using operations of Lauritzen & Jensen [6] and Cowell [2]. By decomposing clique and separator potentials into sets of factors, the proposed architecture takes advantage of independence and irrelevance properties induced by the structure of the graph and the evidence. The resulting benefits are illustrated by examples. Results of a preliminary empirical performance evaluation indicate a significant potential of the proposed architecture.