An Experiment in Spanish-Portuguese Statistical Machine Translation

摘要

Statistical approaches to machine translation have long been successfully applied to a number of 'distant' language pairs such as English-Arabic and English-Chinese. In this work we describe an experiment in statistical machine translation between two 'related' languages: European Spanish and Brazilian Portuguese. Preliminary results suggest not only that statistical approaches are comparable to a rule-based system, but also that they are more adaptive and take considerably less effort to be developed.