IIT Home Page CNR Home Page

All-words Word Sense Disambiguation on a Specific Domain

Domain portability and adaptation of NLP components and Word Sense Disambiguation systems present new challenges. The difficulties found by supervised systems to adapt might change the way we assess the strengths and weaknesses of supervised and knowledgebased WSD systems. Unfortunately, all existing evaluation datasets for specific domains are lexical-sample corpora. With this paper we want to motivate the creation of an allwords test dataset for WSD on the environment domain in several languages, and present the overall design of this SemEval task.


Workshop on Semantic Evaluation, Boulder, Colorado, USA, 2009

External authors: Eneko Agirre (UBC (Spain), Oler Lopez de Lacalle (UBC (Spain), Christiane Fellbaum (Department of Computer Science Princeton University (USA), Antonio Toral (ILC-CNR), Plek Vossen (Faculteit der Letteren Vrije Universiteit Amsterdam )
IIT authors:

Type: Article in proceedings of international peer-reviewed conference
Field of reference: Computer Science & Engineering

Activity: Social and Semantic Web
Web of Data