Using a Large Monolingual Corpus to Improve Translation Accuracy

摘要: The existence of a phrase in large monolingual corpus is very useful information, and so its frequency. We introduce an alternative approach to automatic translation phrases/sentences that operationalizes this observation. use statistical machine system produce translations (re)rank these translations. Our results show combination yields better translations, especially when translating out-of-domain phrases/sentences. can be also used automatically construct parallel corpora from resources.