Building a cross-lingual referential knowledge database using dictionaries
Date of Issue2005
Recent Advances in Natural Language Processing (2005 : Borovets, Bulgaria)
School of Humanities and Social Sciences
Referential knowledge is vital for resolving various problems in NLP, such as anaphora resolution. For example, we have the referential knowledge that diagnose is most likely a member of the referential relation ‘doctor diagnose patient’s illness’. Nariyama et al. (2005) presented an inventory of such referents as doctor, collected from Japanese dictionary definition sentences. Such referential information is based on world knowledge and is applicable across languages. This paper describes our work using the inventory to build a crosslingual referential database for multilingual applications.