The main object of this project is the research, development and validation of high quality automatic translation systems through architectures and processes based on deep learning and artificial neural networks. This global objective is broken down into the following scientific-technological objectives: 1. Research and development of neural translation architectures and algorithms for the significant improvement of translation quality and information preservation. 2. Research and development of data modeling and exploitation methods 3. Creation of advanced neural automatic translation systems with high generic quality and also adapted to specific domains. 4. Creation and preparation of data sets for the generic Basque-Spanish translation and the Spanish-English translation in the health domain. 5. Automatic and human evaluation of the developed systems.

Webpage:
Organization:  Eusko Jaurlaritza
Main researcher: Kepa Sarasola
Participants: 
Eneko Agirre, Iñaki Alegria, Begoña Altuna, Nora Aranberri, Kepa Bengoetxea, Arantza Díaz de Ilarraza, Itziar Gonzalez-Dios, Uxoa Iñurrieta, Mikel Lersundi, Oier López de Lacalle , Kepa Sarasola, Ruben Urizar