Toward Determining the Comprehensibility of Machine Translations
NAVAL RESEARCH LAB WASHINGTON DC
Pagination or Media Count:
Economic globalization and the needs of the intelligence community have brought ma-chine translation into the forefront. There are not enough skilled human translators to meet the growing demand for high quality transla-tions or good enough translations that suf-fice only to enable understanding. Much research has been done in creating transla-tion systems to aid human translators and to evaluate the output of these systems. Metrics for the latter have primarily focused on im-proving the overall quality of entire test sets but not on gauging the understanding of in-dividual sentences or paragraphs. Therefore, we have focused on developing a theory of translation effectiveness by isolating a set of translation variables and measuring their ef-fects on the comprehension of translations. In the following study, we focus on investi-gating how certain linguistic permutations, omissions, and insertions affect the under-standing of translated texts.