News & Events
Topic: Large-Scale Word Sense Disambiguation
Posted - 03/10/2000 : 13:48:28
Large-Scale Word Sense Disambiguation
Professor Yorick Wilks
University of Sheffield, UK
City University of Hong Kong
Time: 4:30 - 6:00 pm
Date: 16 October 2000 (Monday)
Venue: CTL Conference Room B7533, CityU
The talk describes a system that attempts to resolve all content words in a text against the semantic sense distinctions of the LDOCE dictionary (a UK Learners' Dictionary for English with about 50K words). The method we use (in joint work with Mark Stevenson) involves a machine learning method using the TiMBL example-based learner) that combines the outputs from a set of separate disambiguation devices working with different forms of lexical information (e.g. definitions, topic codes, parts of speech etc.). The current best results, in the low 90+%, are, we believe, the best currently available, when evaluated over all text words rather than a selected subset.
About the speaker
Professor Yorick Wilks is Head of the Department of Computer Science at the University of Sheffield (UK) and Director of the Institute of Language, Speech and Hearing. Before that (1985-93) he was founding Director of the Computing Research Laboratory in Las Cruces, New Mexico (US), which specialised in natural language processing. Before that he worked at the Universities of Essex, Edinburgh and Stanford and received his PhD from Cambridge in 1968.