Towards Intelligent Ontology Alignment Systems for Question Answering: Challenges and Roadblocks

Abstract

This paper introduces the main challenges and future research directions for the Ontology Alignment problem. To date a good number of ontology alignment solutions have been proposed. These solutions utilise a wide variety of techniques from machine learning to uncertain reasoning. However, none of the approaches have proved to be an integrated solution, which can be used by different communities. Since 2004, the Ontology Alignment Initiative (OAEI) has established an annual evaluation for systems that could be tested using the same datasets. This, of course, has helped to improve the work on ontology alignment, as the ontology community now has a set of common datasets to make comparisons on the performance of different algorithms for ontology alignment. In this paper we discuss the main challenges and roadblocks that need to be addressed in order to built successful mapping frameworks. Finally this paper presents DSSim and our results on the ontology evaluation 2008.