Paper: Lightly-Supervised Word Sense Translation Error Detection for an Interactive Conversational Spoken Language Translation System

ACL ID E14-4011
Title Lightly-Supervised Word Sense Translation Error Detection for an Interactive Conversational Spoken Language Translation System
Venue Annual Meeting of The European Chapter of The Association of Computational Linguistics
Session Main Conference
Year 2014
Authors

Lexical ambiguity can lead to concept transfer failure in conversational spo- ken language translation (CSLT) systems. This paper presents a novel, classification- based approach to accurately detecting word sense translation errors (WSTEs) of ambiguous source words. The approach requires minimal human annotation effort, and can be easily scaled to new language pairs and domains, with only a word- aligned parallel corpus and a small set of manual translation judgments. We show that this approach is highly precise in de- tecting WSTEs, even in highly skewed data, making it practical for use in an in- teractive CSLT system.