Leveraging recurrent phrase structure in large-scale ontology translation

Abstract. This paper presents a process for leveraging structural relationships and reusable phrases when translating large-scale ontologies. Digital libraries are becoming more and more prevalent. An important step in providing universal access to such material is to pro-vide multi-lingual access to the underlying principles of organization via ontologies, thesauri, and controlled vocabularies. Machine translation of these resources requires high accuracy and a deep vocabulary. Human input is often required, but full manual translation can be slow and expensive. We report on a cost-effective approach to ontology translation. We describe our technique of prioritization, our process of collecting aligned translations and generating a new lexicon, and the resulting improvement to translation system output. Our preliminary evaluation indicates that this technique provides significant cost savings for human-assisted translation. The process we developed can be applied to ontologies in other domains and is easily incorporated into other translation systems.