Practical issues in modeling large diagnostic systems with multiply sectioned Bayesian networks

As Bayesian networks become widely accepted as a normative formalism for di agnosis based on probabilistic knowledge, they are applied to increasingly larger problem domains. These large projects demand a systematic approach t o handle the complexity in knowledge engineering. The needs include modular ity in representation, distribution in computation, as well as coherence in inference. Multiply Sectioned Bayesian Networks (MSBNs) provide a distribu ted multiagent framework to address these needs. According to the framework, a large system is partitioned into subsystems a nd rep resented as a set of related Bayesian subnets. To ensure exact infer ence, the partition of a large system into subsystems and the representatio n of subsystems must follow a set of technical constraints. How to satisfy these goals for a given system may not be obvious to a practitioner. In thi s paper, we address three practical modeling issues.