Large Vocabulary Word Sense Disambiguation
Mark Stevenson, Yorick Alexander Wilks · 2000
Abstract The methodology and evaluation of word sense disambiguation (WSD) as a distinct task are somewhat different from those of others in NLP, and one can distinguish three aspects of this difference, all of which come down to evaluation problems, as does so much in NLP these days. First, researchers are divided over using a general method (one that attempts to apply WSD to all the content words of texts, the large vocabulary approach taken in this paper) versus one that is applied to only a small trial selection of words (e.g. Schutze 1992; Yarowsky 1995). The latter researchers have obtained very high levels of success: Yarowsky quotes 97 per cent correct disambiguation for the small vocabulary over which his system operates, results close to the figures for other ‘solved’ NLP modules, such as part of speech taggers. The issue is whether these small word sample methods and techniques will transfer to general WSD over a more complete vocabulary. Others, besides ourselves (e.g. Mahesh et al. 1997; Harley and Glennon 1997) have pursued the general option on the grounds that it is the real task and should be tackled directly, even with rather lower success rates. The division between the approaches probably comes down to no more than the availability of gold standard text in sufficient quantities, which is more costly to obtain for WSD than other tasks. In this paper we describe a method we have used for obtaining more test material by transforming one resource into another, an advance we believe is unique and helpful in this impasse.