作者: Roberto Navigli , Stefano Faralli
DOI:
关键词:
摘要: We present a new minimally-supervised framework for performing domain-driven Word Sense Disambiguation (WSD). Glossaries several domains are iteratively acquired from the Web by means of bootstrapping technique. The glosses then used as sense inventory fully-unsupervised domain WSD. Our experiments, on and gold-standard datasets, show that our wide-coverage enables high-performance results dozens at coarse fine-grained level.