A Probabilistic Approach to Hierarchical Model-based Diagnosis

TL;DR

This paper introduces a probabilistic, hierarchical Bayesian network approach to model-based diagnosis, improving computational efficiency in fault isolation by leveraging hierarchical structures.

Contribution

It extends traditional model-based diagnosis to a probabilistic framework using Bayesian networks with hierarchical support, enabling more efficient inference.

Findings

Hierarchical Bayesian networks improve diagnostic inference efficiency.

The approach effectively models complex systems with hierarchical structures.

Probabilistic diagnosis outperforms traditional methods in accuracy and speed.

Abstract

Model-based diagnosis reasons backwards from a functional schematic of a system to isolate faults given observations of anomalous behavior. We develop a fully probabilistic approach to model based diagnosis and extend it to support hierarchical models. Our scheme translates the functional schematic into a Bayesian network and diagnostic inference takes place in the Bayesian network. A Bayesian network diagnostic inference algorithm is modified to take advantage of the hierarchy to give computational gains.