The project aims at researching methods and techniques to overcome one of the central problems of machine translation (MT) – the lack of linguistic resources such as training data for under-resourced areas of machine translation. The main goal is to find, analyze and evaluate novel methods that exploit comparable corpora on order to compensate for the shortage of linguistic resources, and ultimately to significantly improve MT quality for under-resourced languages and narrow domains. Models generated from comparable corpora will be compared against baseline models generated from parallel corpora.
Partners
- Tilde, LV (Coordinator)
- University of Sheffield, UK
- University of Leeds, School of Modern Languages and Cultures, Centre for Translation Studies, UK
- Institute for Language and Speech Processing, GR
- University of Zagreb, HR
- German Research Center for Artificial Intelligence, Language Technology Lab, DE
- Research Institute for AI, Romanian Academy, Romania
- Linguatec, Germany
- Zemanta, Slovenia
