Towards A Hybrid Approach To Word-Sense Disambiguation In Machine Translation

The task of word sense disambiguation aims to select the correct sense of a polysemous word in a given context. When applied to machine translation, the correct translation in the target language must be selected for a polysemous lexical item in the source language. In this paper, we present work in progress on a supervised WSD system with a hybrid approach: on the one hand it relies on supervised learning from manually sense-tagged corpora, and on the other hand it has the ability to use information from manually crafted disambiguation rules. We present evaluation and further plans to improve the system.