We propose interpolated backoff methods to strike the balance between traditional surface form translation models and factored models that decompose translation into lemma and morphological feature mapping steps. We show that this approach improves translation quality by 0.5 BLEU (German–English) over phrase-based models, due to the better translation of rare nouns and adjectives.
|Title of host publication||In Proceedings of the Tenth Conference of the Association for Machine Translation in the Americas (AMTA)|
|Publisher||Association for Machine Translation in the Americas, AMTA|
|Number of pages||10|
|Publication status||Published - 2012|