Coarse “split and lump” bilingual language models for richer source information in SMT

  1. (PDF, 661 KB)
AuthorSearch for: ; Search for: ; Search for: ; Search for:
Proceedings titleProceedings of the Eleventh Conference of the Association for Machine Translation in the Americas (AMTA), 2014
Conference11th Conference of the Association for Machine Translation in the Americas (AMTA), October 22-26, 2014, Vancouver, BC, Canada
SubjectAluminum; Computational linguistics; Computer aided language translation; Speech transmission; Syntactics Automatically generated; Bilingual language model; Coarse models; Contextual information; Language pairs; Parts of speech; Statistical machine translation; Word clustering
AbstractRecently, there has been interest in automatically generated word classes for improving sta- tistical machine translation (SMT) quality: e.g, (Wuebker et al, 2013). We create new mod- els by replacing words with word classes in features applied during decoding; we call these “coarse models”. We find that coarse versions of the bilingual language models (biLMs) of (Niehues et al, 2011) yield larger BLEU gains than the original biLMs. BiLMs provide phrase-based systems with rich contextual information from the source sentence; because they have a large number of types, they suffer from data sparsity. Niehues et al (2011) miti- gated this problem by replacing source or target words with parts of speech (POSs). We vary their approach in two ways: by clustering words on the source or target side over a range of granularities (word clustering), and by clustering the bilingual units that make up biLMs (bitoken clustering). We find that loglinear combinations of the resulting coarse biLMs with each other and with coarse LMs (LMs based on word classes) yield even higher scores than single coarse models. When we add an appealing “generic” coarse configuration chosen on English > French devtest data to four language pairs (keeping the structure fixed, but providing language-pair-specific models for each pair), BLEU gains on blind test data against strong baselines averaged over 5 runs are +0.80 for English > French, +0.35 for French > English, +1.0 for Arabic > English, and +0.6 for Chinese > English.
Publication date
PublisherAssociation for Machine Translation in the Americas
AffiliationNational Research Council Canada; Information and Communication Technologies
Peer reviewedYes
NPARC number23001442
Export citationExport as RIS
Report a correctionReport a correction
Record identifierb9054aec-0086-41ab-b1b8-11dddc2cca51
Record created2017-02-08
Record modified2017-02-08
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)
Date modified: