MANTRA (Multilingual Annotation of Named Entities and Terminology Resources Acquisition) will provide multilingual terminology and semantically annotated multilingual corpora (e.g. patents). This is intended to improve access to scientific information in multilingual documents. The MANTRA project wants to generate translations of words and phrases, especially with the help of parallel corpora and various alignment methods. A large number of such parallel corpora are available for this in the biomedical domain on which MANTRA focuses.
Project partners will tap into these multilingual document collections to generate terms and concepts in different languages. This is to complement the landscape of available biomedical vocabularies (such as Medical Subject Headings (MeSH)).