Paper: Name Translation in Statistical Machine Translation - Learning When to Transliterate

ACL ID P08-1045
Title Name Translation in Statistical Machine Translation - Learning When to Transliterate
Venue Annual Meeting of the Association of Computational Linguistics
Session Main Conference
Year 2008
Authors

We present a method to transliterate names in the framework of end-to-end statistical machine translation. The system is trained to learn when to transliterate. For Arabic to English MT, we developed and trained a transliterator on a bitext of 7 million sen- tences and Google’s English terabyte ngrams and achieved better name translation accuracy than 3 out of 4 professional translators. The paper also includes a discussion of challenges in name translation evaluation.