Language, voice and multimedia technologies in the service of Language Industry

Ber2tek is a strategic research project with a duration of three years (2012-2014). Its consortium is made up of the Elhuyar Foundation, the IXA and Aholab research groups of the UPV/EHU-University of the Basque Country, the Vicomtech technology centre and the Tecnalia Research & Innovation foundation. The project Ber2tek is partly funded by the Departments of Industry and Culture of the Basque Government.

The Ber2tek project is continuation of BerbaTek which was subsidised by the Etortek programme in the period 2009-2011, and carried out by the same consortium. In BerbaTek we began to research and develop language-, voice- and multimedia-technologies for their application in the sector of the languages industry.

Linguistic technologies are a fundamental pillar for the development of the language industry, especially in the field of the impulse of competitiveness. Taking this into account, the Ber2tek project aims to further develop those linguistic technologies that contribute to the sector and promote the competitiveness of the Basque Country as a leader in the field of the linguistic services.

Ber2tek is born, therefore, as a natural continuation of the BerbaTek project, and as a consequence of the strategic reflection of its partners regarding the technological needs identified by the local and European language industries.

Its main goal is to advance in the research and development of cross-media contents analysis technologies, high quality machine translation and natural spoken multimodal interaction. Ber2tek also wants to generate knowledge and training of the qualified critical mass that will allow to face the future of R&D inside this strategic line. To achieve this, we will organise training courses, post-degree masters, skilled seminars and doctorate courses in the field of language technologies.

To achieve the aims set, the partners of the Ber2tek project have a scientific community with the necessary critical mass, internationally recognised, in addition to preferential relations with a lot of companies of the language industry.

After the success of the approach adopted in the previous project, Ber2tek will also be oriented towards applications. Without leaving basic research apart, we will try to present experimental applications as demos that can later be developed and turned into products by companies devoted to translation, language training or content management. In addition to using these demos as evaluators and proofs of feasibility, we intend to obtain the cooperation of companies in the field to participate in the project as beta-testers.

Besides, being the Ber2tek project and its predecessor BerbaTek the first ones to comprise the different areas that integrate the economic sector of the language industry, it will facilitate its structuration. The development of language technologies will also positively affect the process of globalisation of Basque companies, since it will allow them to face the jump to the multilingual stage, in which translation and language training are key factors. Also, the fact that Euskadi is a multilingual community offers an excellent know-how and benchmark for the development of both language technologies and the language industry, and offers a good starting point for positioning Euskadi among the reference countries in language- and voice-technologies.

Vicomtech’s role within Ber2tek is to research on automatic speech recognition technologies, avatar interaction systems, natural language processing, automatic translation, and image analysis, classification, segmentation and automatic annotation.

Duration of the project

01/01/2012 - 31/12/2014