30 likes | 123 Views
Explore the current state of Machine Translation paradigms at the 2009 Translingual Europe event in Prague, discussing challenges, advancements, and the potential of RBMT systems. Discover the ideal translation world and the strengths of RBMT with a focus on quality, terminology management, translation memory, and translation management systems.
E N D
Translingual Europe Prague May 13, 2009 Daniel Grasmick Panel Discussion Competition of MT ParadigmsWhere is RBMT? Translingual Europe 2009 1
Translingual Europe 20092 • And Machine Translation?… as perfect as all modules and technologies below … The Ideal Translation World • Source text • Good quality of language and format • Properly structured and tagged (CMS…) • Terminology • Source terms maintained at early stage • All target equivalents maintained before translation starts • Translation Memory • All previous versions properly translated and maintained • Content adapted to up-to-date terminology • Able to linguistically analyze source and target • Translation Management Systems • All processes under control, reusable and adaptable • Access for all parties and transparency at all stages • Human factor (cost-efficient, always available) • Only subject matter and language experts involved
RBMT – A Strong Case • Numerous solutions in place • Integration in translation production environment • See among others SDL, LionBridge; increasing interest from LSPs • Integration in gisting / communication solutions • Lucy LT alone translated some 5 billion words on free sites in 2008 • Technology vendors in place • Long and extensive experience of vendors in place • Systran, Linguatec, Promt, Siemens/Comprendium/Lucy … • Technology • Much better than it‘s reputation • Eas(ier) to integrate, adapt and control, flexible • Efforts to develop but ROI proven in numerous solutions • Potential currently starting to be better exploited • Improving integration in TMS and TM • Making better usage of linguistic assets maintained by customers • Enhancing results by adding statistical approaches • Developing hybrid solutions based on extensive linguistic experience accumulated over the years Translingual Europe 20093