The trouble with SMT consistency

  1. (PDF, 221 KB)
AuthorSearch for: ; Search for:
Proceedings titleProceedings of the 7th Workshop on Statistical Machine Translation
Conference7th Workshop on Statistical Machine Translation, June 7-8, 2012, Montreal, Quebec
Pages442449; # of pages: 8
AbstractSMT 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.
Publication date
PublisherAssociation for Computational Linguistics
AffiliationInformation and Communication Technologies; National Research Council Canada
Peer reviewedYes
NPARC number20794314
Export citationExport as RIS
Report a correctionReport a correction
Record identifiera7b9f071-84b0-48c8-a215-312739ffe880
Record created2012-10-12
Record modified2016-05-09
Bookmark and share
  • Share this page with Facebook (Opens in a new window)
  • Share this page with Twitter (Opens in a new window)
  • Share this page with Google+ (Opens in a new window)
  • Share this page with Delicious (Opens in a new window)