An Evaluation of Graded Sense Disambiguation using Word Sense Induction
David Jurgens · 2012
Word Sense Disambiguation aims to label the sense of a word that best applies in a given context. Graded word sense disambiguation relaxes the single label assumption, allowing for multiple sense labels with varying degrees of applicability. Training multi-label classifiers for such a task requires substantial amounts of annotated data, which is currently not available. We consider an alternate method of annotating graded senses using Word Sense Induction, which automatically learns the senses and their features from corpus properties. Our work proposes three objective to evaluate performance on the graded sense annotation task, and two new methods for mapping between sense inventories using parallel graded sense annotations. We demonstrate that sense induction offers significant promise for accurate graded sense annotation. 1