Paper: Neighbors Help: Bilingual Unsupervised WSD Using Context

ACL ID P13-2096
Title Neighbors Help: Bilingual Unsupervised WSD Using Context
Venue Annual Meeting of the Association of Computational Linguistics
Session Short Paper
Year 2013

Word Sense Disambiguation (WSD) is one of the toughest problems in NLP, and in WSD, verb disambiguation has proved to be extremely difficult, because of high de- gree of polysemy, too fine grained senses, absence of deep verb hierarchy and low in- ter annotator agreement in verb sense an- notation. Unsupervised WSD has received widespread attention, but has performed poorly, specially on verbs. Recently an unsupervised bilingual EM based algo- rithm has been proposed, which makes use only of the raw counts of the transla- tions in comparable corpora (Marathi and Hindi). But the performance of this ap- proach is poor on verbs with accuracy level at 25-38%. We suggest a modifica- tion to this mentioned formulation, using context and semantic relatedness of neigh- boring words. An improvement...