Reconciling ontological differences by assistant agents

In this paper, we propose an alternative solution for approaching the problem of semantic interoperability, based on recent projects to develop software agents and systems that attempt to reconcile ontological differences without explicit formal ontologies. An application capable of reconciling ontologies based on example data is first described. The results from one of the case studies where the proposed solution has been applied are reported. Finally, the design of an agent embedding the solution is sketched. The solution is aimed at the creation of assistant agents helping the users to cope with heterogeneity in both information sources and contexts of their usage