Updating Ontology Alignment on the Concept Level Based on Ontology Evolution

According to one of the base definitions of an ontology, this representation of knowledge can be understood as a formal specification of conceptualization. In other words - they can be treated as a set of well-defined concepts, which represent classes of objects from the real world, along with relationships that hold between them. In the context of distributed information systems, it cannot be expected that all of the interacting systems can use one, shared ontology. It entails a plethora of difficulties related to maintaining such a large knowledge structure. A solution for this problem is called an ontology alignment, sometimes it is also referred to as an ontology mapping. It is a task of designating similar fragments of ontologies, that represent the same elements of their domain. This allows different components of a distributed infrastructure to preserve its own independent ontology while asserting mutual interoperability. However, when one of the participating ontologies change over time, the designated alignment may become stale and invalid. As easily seen in a plethora of methods found in the literature, aligning ontologies is a complex task. It may become very demanding not only in terms of its computational complexity. Thus relaunching it from the beginning may not be acceptable. In this paper, we propose a set of algorithms capable of updating a pre-designated alignment of ontologies based solely on the analysis of changes applied during their evolution, without the necessity of relaunching the mapping algorithms from scratch.