Knowledge web: heterogeneity work package

Business integration and Web Services collaboration (and more generally the semantic web) require the use of heterogeneous ontologies, which, in turn, describe heterogeneous data. There are many sources of heterogeneity: differences of terminology (the same concept can be given different names in two ontologies), differences of modelling (the same concept can be defined in different ways), or differences of knowledge representation languages (a concept cannot be defined as precisely in a language as in another). Semantic interoperability can be grounded in ontology reconciliation: finding relationships between concepts belonging to different ontologies (alignment) and implementing the operator (transformation or merge).

The Heterogeneity work package (2.2) of the Knowledge web European network of excellence, contributes to improve solving heterogeneity problems through matching and exploiting the resulting alignments following the process depicted in the figure below:

It has already produced a number of results among which:

• State of the art comparison (D2.2.3)
• Several general characterization of alignments (D2.2.1 and 2.2.5);
• An alignment format (D2.2.6);
• A methodology for evaluating matching algorithms (D2.2.2 and D2.2.4) that has been experimented.

These results are available in the forementionned deliverable as well as in publications. They also seeded some independently sustainable efforts:

• An Alignment API and its implementation that allows alignment algorithms and other efforts to cooperate.
• The Ontology Alignment Evaluation Initiative: an effort for evaluating alignment algorithms providing benchmark tests and results.
• The Reference site on alignment.

Resources:

• The Knowledge web site which provide access to all deliverables.
• The Knowledge web work package 2.2 mailing list in which discussion occur.

Contact: Jerome:Euzenat#inrialpes:fr