Output details

Disambiguation of ambiguous biomedical terms using examples generated from the UMLS Metathesaurus

<22> Supervised Word Sense Disambiguation systems require labelled training data; however, this is expensive and time-consuming to create. This paper describes novel algorithms for generating labelled training data using information from a structured knowledge source (UMLS). The research was EPSRC-funded (BioWSD). Follow-on funding from an EPSRC Knowledge Transfer Account explores commercialisation of the approach. The research led to further papers at high-impact international conferences (NAACL-2010, 2012 and 2013) and international collaboration with the University of Minnesota (Bridget McInnes <btmcinnes@gmail.com>) and Universidad Complutense de Madrid (Laura Plaza Morales <lplaza@lsi.uned.es>).