Lexical substitution for the medical domain

Abstract

In this paper we examine the lexical substitution task for the medical domain. We adapt the current best system from the open domain, which trains a single classifier for all instances using delexicalized features. We show significant improvements over a strong baseline coming from a distributional thesaurus (DT). Whereas in the open domain system, features derived from WordNet show only slight improvements, we show that its counterpart for the medical domain (UMLS) shows a significant additional benefit when used for feature generation.