Institute of Formal and Applied Linguistics

Domain adaptation of machine translation deals with the problem when the domain of training data of a machine translation system differs from its target domain. In our work, not only we address adaptation to specific domains (e.g. medical), but also genres (search queries), and applications (cross-lingual information retrieval). The talk reviews our current and previous work in this area which has been conducted within three EU projects: Panacea, Khresmoi, and KConnect and reports the main findings and results. favicongenerator.org