OntoBayes: An Ontology-Driven Uncertainty Model

Abstract

This paper describes an ontology-driven model, which integrates Bayesian Networks (BN) into the Ontology Web Language (OWL) to preserve the advantages of both. This model makes use of probability and dependency-annotated OWL to represent uncertain information in BN structures. These extensions enhance knowledge representation in OWL and enable agents to act under uncertainty and complex structured open environments at the same time. This paper presents the underlying principles and scratches the surface of the decision theoretic agent system design based on "OntoBayes".