Accurate and Efficient Parsing of Biomedical Text

The aim of this research is to take an existing wide-coverage parser of English, currently tuned for newspaper text, and adapt it to biomedical text. How to adapt domain-dependent parsers to new domains is one of the key questions facing researchers in statistical parsing. A further aim is to adapt an existing named entity recogniser to recognise biological entities such as genes and proteins, and to integrate the named entity recogniser into the parser. The creation of a language processing tool which can infer the grammatical relations between words and simultaneously identify biological entities has the potential to be of huge benefit to the biomedical research community, which is struggling to manage the vast amount of research output being produced.