Algorithm and Tool for Automated Ontology Merging and Alignment

Researchers in the ontology-design field have developed the content for ontologies in many domain areas. Recently, ontologies have become increasingly common on the WorldWide Web where they provide semantics for annotations in Web pages. This distributed nature of ontology development has led to a large number of ontologies covering overlapping domains. In order for these ontologies to be reused, they first need to be merged or aligned to one another. The processes of ontology alignment and merging are usually handled manually and often constitute a large and tedious portion of the sharing process. We have developed and implemented PROMPT, an algorithm that provides a semi-automatic approach to ontology merging and alignment. PROMPT performs some tasks automatically and guides the user in performing other tasks for which his intervention is required. PROMPT also determines possible inconsistencies in the state of the ontology, which result from the user’s actions, and suggests ways to remedy these inconsistencies. PROMPT is based on an extremely general knowledge model and therefore can be applied across various platforms. Our formative evaluation showed that a human expert followed 90% of the suggestions that PROMPT generated and that 74% of the total knowledge-base operations invoked by the user were suggested by PROMPT. 1 Ontologies in AI and on the Web Ontologies today are available in many different forms: as artifacts of a tedious knowledge-engineering process, as information that was extracted automatically from informal electronic sources, or as simple “light-weight” ontologies that specify semantic relationships among resources available on the World-Wide Web (Brickley and Guha 1999). But what does a user do when he finds several ontologies that he would like to use but that do not conform to one another? The user must establish correspondences among the source ontologies, and to determine the set of overlapping concepts, concepts that are similar in meaning but have different names or structure, concepts that are unique to each of the sources. This work must be done regardless of whether the ultimate goal is to create a single coherent ontology that includes the information from all the sources (merging) or if the sources must be made consistent and coherent with one another but kept separately (alignment). Currently the work of mapping, merging, or aligning ontologies is performed mostly by hand, without any tools Copyright © 2000, American Association for Artificial Intelligence (www.aaai.org). All rights reserved. to automate the process fully or partially (Fridman Noy and Musen 1999).Our participation in the ontology-alignment effort within DARPA’s High-Performance KnowledgeBases project (Cohen et al. 1999) was a strong motivation for developing semi-automated specialized tools for ontology merging and alignment. Several teams developed ontologies in the domain of military planning, which then needed to be aligned to one another. We found the experience of manually aligning the ontologies to be an extremely tedious and time-consuming process. At the same time we noticed many steps in the process that could be automated, many points where a tool could make reasonable suggestions, and many conflicts and constraint violations for which a tool could check. We developed a formalism-independent algorithm for ontology merging and alignment—PROMPT (formerly SMART)—which automates the process as much as possible. Where an automatic decision is not possible, the algorithm guides the user to the places in the ontology where his intervention is necessary, suggests possible actions, and determines the conflicts in the ontology and proposes solutions for these conflicts. We implemented the algorithm in an interactive tool based on the Protégé-2000 knowledge-modeling environment (Fridman Noy et al. 2000). Protégé-2000 is an ontology-design and knowledgeacquisition tool with an OKBC-compatible (Chaudhri et al. 1998) knowledge model, which allows domain experts (and not necessarily knowledge engineers) to develop ontologies and perform knowledge acquisition. We have evaluated PROMPT, comparing its performance with the humanexpert performance and with the performance of another ontology-merging tool.