The project has been partially financed by the Department of Economic Development and competitiveness of the Basque Government through the ELKARTEK program and has also counted with the support of MondragonLingua T&C, EiTB and ARGIA.
The machine translation system developed in the project is based on the automatic neuronal translation, the dominant machine translation paradigm, both academically and commercially.
Throughout the project several automatic translation systems have been developed for Euskara and Castilian and the different approaches used have been compared. The two alternatives to the automatic neuronal translation evaluated have been the statistical machine translation systems developed in the project itself and the rules-based system offered by the Basque Government.
In all objective evaluations carried out in the project, neuronal translation systems offer significantly better results than the rest of the evaluated systems. The members of the consortium believe that there has been an unprecedented qualitative leap in terms of the quality of automatic translation systems for Euskara.
The challenge is now in how to translate these results into the market, a task in which ISEA is working with MondragonLingua T&C and the rectum of members of the consortium.