The Information-Flow Approach to Ontology-Based Semantic Alignment
Speaker(s) : Marco SCHORLEMMER (IIIA - CSIC, Barcelone)
Ontologies are commonly defined as specifications of the conceptualisations of particular domains, and they have been widely adopted as a key technology that may favour knowledge sharing in distributed environments, such as multi-agent systems, federated databases, or the semantic web. But the proliferation of many diverse ontologies caused by different conceptualisations of even the same domain has highlighted the need of ontology matching techniques. Currently, most ontology matching mechanisms developed so far have taken a classical functional approach to the semantic heterogeneity problem, in which ontology matching is seen as a process taking two or more ontologies as input and producing a semantic alignment of ontological entities as output. Furthermore, matching often has been carried out at design-time, before integrating knowledge-based systems or making them interoperate. But, multi-agent communication, peer-to-peer information sharing, and web-service composition are all of a decentralised, dynamic, and open-ended nature, and they require ontology matching to be locally performed during run-time. In this talk I shall describe a formal model that formalises progressive semantic alignment as a sequence of information-channel refinements in the sense of Barwise and Seligman s theory of information flow. This theory is particularly useful for our endeavour because it models the flow of information occurring in distributed systems due to the shared entities---or tokens---that carry information. We claim that semantic alignment of ontological terminology is often also relative to the particular situation in which the alignment is carried out, and that this situation should be made explicit and brought into the alignment mechanism, too.
Javier.Diaz (at) nulllip6.fr