The trouble with SMT consistency
Affiliation:
Information and Communication Technologies; National Research Council Canada
Type:
Conference publication
Conference:
7th Workshop on Statistical Machine Translation, June 7-8, 2012, Montreal, Quebec
Title:
Proceedings of the 7th Workshop on Statistical Machine Translation
Abstract:
SMT typically models translation at the sentence
level, ignoring wider document context.
Does this hurt the consistency of translated
documents? Using a phrase-based SMT system
in various data conditions, we show that
SMT translates documents remarkably consistently,
even without document knowledge.
Nevertheless, translation inconsistencies often
indicate translation errors. However, unlike in
human translation, these errors are rarely due
to terminology inconsistency. They are more
often symptoms of deeper issues with SMT
models instead.