Using corpus statistics and WordNet relations for sense identification

Corpus-based approaches to word sense identification have flexibility and generality but suffer from a knowledge acquisition bottleneck. We show how knowledge-based techniques can be used to open t...