Post-édition statistique pour l’adaptation aux domaines de spécialité en traduction automatique (Statistical Post-Editing of Machine Translation for Domain Adaptation) [in French]

Statistical Post-Editing of Machine Translation for Domain Adaptation This paper presents a statistical approach to adapt generic machine translation systems to the medical domain through an unsupervised post-edition step. A statistical post-edition model is built on statistical machine translation outputs aligned with their translation references. Evaluations carried out to translate medical texts from French to English show that a generic machine translation system can be adapted a posteriori to a specific domain. Two systems are studied : a state-of-the-art phrase-based implementation and an online publicly available software. Our experiments also indicate that selecting sentences for post-edition leads to significant improvements of translation quality and that more gains are still possible with respect to an oracle measure.