State-of-the-art on mapping maintenance and challenges towards a fully automatic approach

In several domains, software applications have intensively used Knowledge Organization Systems (KOS) like database schemas, ontologies, taxonomies and thesauri and their associated semantic correspondences (i.e., mappings). This underlines the relevance and capabilities of KOS and mappings to manage and integrate vast amounts of data. However, the dynamic nature of domain knowledge forces knowledge engineers to constantly modify KOS, to keep them up to date and useful. In this context, the maintenance of mappings affected by KOS evolution still remains an open research issue. Although this problem appears relevant for many different computer science fields, ranging from database to artificial intelligence, literature has so far only superficially addressed it to enable more flexible, automatic and precise solutions. This article presents, discusses and compares existing approaches for maintaining mappings and describes open research challenges.