unimelb: Topic Modelling-based Word Sense Induction
Jey Han Lau, Paul F. Cook, Timothy J. Baldwin · 2013
This paper describes our system for shared task 13 “Word Sense Induction for Graded and Non-Graded Senses ” of SemEval-2013. The task is on word sense induction (WSI), and builds on earlier SemEval WSI tasks in exploring the possibility of multiple senses being compatible to varying degrees with a single contextual instance: participants are asked to grade senses rather than selecting a single sense like most word sense disambiguation (WSD) settings. The evaluation measures are designed to assess how well a system perceives the different senses in a contextual instance. We adopt a previously-proposed WSI methodology for the task, which is based on a Hierarchical Dirichlet Process (HDP), a nonparametric topic model. Our system requires no parameter tuning, uses the English ukWaC as an external resource, and achieves encouraging results over the shared task. 1