Non Omniscient Intensional Contextual Reasoning for Query-agents in P2P Systems.
Zoran Majkić · Indian International Conference on Artificial Intelligence · 2007
Given a de-centralized nature of the development of the Semantic Web, there will be an explosion in the number of ontologies in P2P systems. To integrate data from disparate ontologies, we must know the semantic correspondence between their elements. What we argue is the full epistemic independency of peer databases: they can change their ontology and/or extension of their knowledge independently from other peers and without any communication to other peers. We intend to obtain very robust P2P systems, able to answer to user queries also when intended mappings between peers do not feet with the modified ontologies of peers, but also to have the possibility to map naturally P2P database systems into a grid computation. We introduce S5 modal intensional FOL embedding of P2P systems and consider the real-life query-agents, with a non omniscient deductive model, based on the contextualization of P2P systems. We define the bridge rules for such contextual reasoning, and a query-agent with a ’belief’ predicate for intensionally equivalent queries over peers, deductively inferred by agent w.r.t. the given user query, and we show that it is sound weakening of the omniscient S5 modal intensional FOL.