The Centre for Speech and Language Applications and Technologies (TALP) at the UPC focuses on research and development in the area of human language processing. It is comprised of two research groups: the Speech Processing group and the Natural Language Processing group, and has been a stable research team through all these years. While the Natural Language Processing Group from the Computer Science Department is oriented to textual language, the VEU group, from the Signal Theory and Communication Department, has speech as its main source of data to be processed. Both research groups had already been active since mid 1980s, and after the birth of the TALP Center in 1998, they have been collaborating mostly regarding higher levels of language structure and machine translation technologies. Since its creation in 2002, the centre has been involved in research and development of spoken and written language processing technologies and applications with special attention to English, Catalan, and Spanish.
TALP has a long trajectory on knowledge and technology transfer with several spin-off from our group, contracts with companies for research and development and licensing technologic products and linguistic resources.
The center has started technology transfer regarding deep learning methods, and a large increase of that is expected for the next coming years. After decades of small year-by-year increase in performance, the introduction of deep learning techniques is producing big steps towards human or even super-human performance in tasks as speech recognition, speaker verification, text-to-speech, or machine translation.
Our technology has been applied on Health, Culture and Education, Web Services, Social Networks Analytics, Tourism, Security, Financial sector or Automotive industry.
The group is member of META-SHARE. It aims at providing an open, distributed, secure, and interoperable infrastructure for the Language Technology domain. The group collaborates actively in META-SHARE through linguistic resources in Spanish and Catalan, and providing free software. It is important to mention that our contribution to META-SHARE has made possible to place Catalan at the level of other non-global European languages. In parallel, the group also offer other language resources through the European Language Resources Association under distribution agreements.
We distribute versions of Software for translation, voice conversion as well as an open source language analysis tool suite, Freeling. Our purpose is to follow the Open Access line of research results and contribute to the UPC in the dissemination of research articles Open Access through its publications repository. Most of our recent research publications are also linked to its corresponding open source code for replicability of the results. This is a common practice in the scientific community now, and the TALP is following this practice as well.
On 2017 we agreed with industrial doctorate projectes. These projects is a new way to transfer technology and knowledge to the industry and to increase our research potential. We have found several companies interested and we will continue exploring this new option in the next years.