UMBC ebiquity
Weaving the Web of Belief into the Semantic Web
May 22, 2004
Collaboration, especially knowledge sharing, enables the advance of science as well as human society. In cyberspace, socializing the traditionally isolated intelligent software agents is the ultimate goal of the emerging Semantic Web activity. When making collaboration decisions, an agent usually needs explicitly represented facts about the agent world, such as ``who knows what" and ``who can do what". However, the limited computation and storage resources forbid an agent to independently maintain rational beliefs on all facts about the agent world. So, the full picture of the agent world has to be distributed in the knowledge-sharing social network of those resident agents. In this paper, we propose a generic representation framework for this distributed knowledge network, which is also called the reminiscent of Quine's web of belief. The framework includes: the RDF-based semantic relation model, which is a cognitive data model for the agent world; a general OWL ontology, which facilitates representing agent properties (such as knowledge and capability) and finer inter-agent trust relations; and the practical issues on maintaining the web of belief for distributed inference.
TechReport
UMBC
Ebiquity
Downloads: 7210 downloads