Machine Translation @ the National Centre for Language Technology

Dublin City University, Ireland

Home People Projects Publications Events Announcements Links
-- current projects --
  EuroMatrix+
  Panacea
  CoSyne
  PLuTO
  T4ME
  Confident MT

-- completed projects --
  Prospect
  Attempt
  Sign language
    translation

  Evaluation
  Transbooster
  DVD subtitling
  LFG-DOP
  EBMT & Marker
    Hypothesis

  DOP & DOT
  Hybrid EBMT-SMT
Title:Hybrid Data Driven MT: Integrating EBMT and SMT
Duration:October 1st 2003 -- September 30th 2006
Funded by:IRCSET's Embark Initiative
People:Declan Groves, Andy Way
Description:Almost all research in MT being carried out today is corpus-based. The two main data-driven approaches are that of Statistical Machine Translation (SMT) and Example-Based Machine Translation (EBMT), with SMT being by far the more dominant of the two.
 SMT provides us with an empirical-based, language independant and robust approach to the problem of translation, and with the advent of more sophisticated phrase-based approaches, it is possible to achieve high-quality translation. However, statistical approaches have inherent limitations due to their lack of linguistic knowledge. Most SMT systems do not encorporate any explicit syntax, unlike EBMT which makes use of syntax at its core. EBMT has always made use of both phrasal and lexical correspondences to produce high-quality translations.
 This work focuses on merging elements of EBMT and SMT with a view to producing a hybrid data-driven EBMT-SMT system which combines the advantages of both systems to produce highly-accurate, syntactically well-formed translations without a significant loss in coverage.
 
Last update: June 16 2007
Related Sites: NCLT | School of Computing | School of Applied Languages and Intercultural Studies | Dublin City University