AnHitz

The AnHitz project, the second collaboration experience of the BerbaTek consortium in strategic research into language technologies, for voice and multimedia, was run during 2006-2008.

The aim of the AnHitz project was the research and development of language technologies in Basque required for enabling person-device interaction and the management of knowledge in a natural, intuitive and friendly way with a view to using it in Ambient Intelligence. The advance of the Information and Knowledge Society has turned person-device interaction into a priority channel for communication and into the new paradigm for the acceptance of this new Society.

To achieve greater acceptance of the Information and Knowledge Society paradigms in the Basque Country, it is necessary to develop not only language technology but also the industry required to support the society’s language diversity.

  • Development of language technology in Basque for person-device interaction.
  • The taking of a significant new step forward in the strengthening of the language industry.
  • The setting up of new technology-based firms (NTBFs) and the consolidating of the ones already set up.
  • The generation of knowledge.

The following results of the AnHitz project are worth highlighting:

  • A tool for creating a Basque corpus using the Internet
  • A tool for consulting the Internet as a Basque corpus
  • A Basque corpus with syntactic and semantic tagging.
  • A terminology extraction tool using comparable corpora
  • The expanding of lexical ontologies
  • The improvement of existing analysers
  • Statistical English-Basque machine translation
  • A Basque-language search engine
  • A specialised translingual search engine
  • A question reply system
  • A semantic, similarity-based document relater
  • Multilingual voice synthesis
  • Voice synthesis with emotions
  • Dictionary- and grammar-based voice recognition for Basque
  • A speaker identification system
  • Basque-speaking avatars

Furthermore, at the end of the AnHitz project a demo with a virtual 3D expert in science and technology with voice interaction in Basque was developed. It is capable of responding to questions and conducting a translingual search with subsequent machine translation into Basque.

AnHitz demo

This demo was widely reported in the media, as can be seen in the section on dissemination in the media. What is more, it received the awards for the best presentation and best paper at the international conference on Computational Linguistics - Applications (CLA'09) held in Mrągowo (Poland).