Virtual Knowledge Communities for Semantic Agents

International audienceVirtual Knowledge Communities are a well suited paradigm for decentralized knowldege exchanges and they have been applied in several domains. In this paper we investigate the implementation of virtual knowledge communities with se- mantic agents. Using the SAM (Semantic Agent Modeling) approach, we show that agents can exchange community re- lated concepts (in OWL) and behavior (in SWRL). Agents can then learn and adapt new community-related behavior, which is usefull when changing the role or entering into a new environment. For this purpose, we formalize Virtual Knowledge Communities in a set-theoretic way and we im- plement this formalization in an OWL ontology. Some ex- amples of community representation using our formalization are presented in this paper