Exploring Localization in Bayesian Networks for Large Expert Systems

TL;DR

This paper introduces multiply sectioned Bayesian networks that allow localized reasoning in large expert systems, improving efficiency by updating only relevant subdomains instead of the entire network.

Contribution

It proposes a novel structure of multiply sectioned Bayesian networks with separate subnets and junction trees for efficient localized inference in large domains.

Findings

Probabilities match those from homogeneous networks

Computational load depends on the size of a single junction tree

Supports shifting attention between subdomains

Abstract

Current Bayesian net representations do not consider structure in the domain and include all variables in a homogeneous network. At any time, a human reasoner in a large domain may direct his attention to only one of a number of natural subdomains, i.e., there is ?localization' of queries and evidence. In such a case, propagating evidence through a homogeneous network is inefficient since the entire network has to be updated each time. This paper presents multiply sectioned Bayesian networks that enable a (localization preserving) representation of natural subdomains by separate Bayesian subnets. The subnets are transformed into a set of permanent junction trees such that evidential reasoning takes place at only one of them at a time. Probabilities obtained are identical to those that would be obtained from the homogeneous network. We discuss attention shift to a different junction tree and propagation of previously acquired evidence. Although the overall system can be large, computational requirements are governed by the size of only one junction tree.