The Speech and Natural Language Technologies department researches and develops advanced speech and language processing techniques to transfer innovative applications to the industry and meet multi-sector needs for multilingual multimedia content management and/or human-machine interaction characteristic of the information society.
The department’s main lines of activity are:
- Speech Processing: automatic speech recognition, text-to-speech alignment, capitalisation and punctuation, speaker segmentation and recognition, language identification, voice biometrics, acoustic segmentation and classification, spoken emotion recognition
- Natural Language Processing: core text processing technology, text normalisation, analysis and classification, named entity detection and disambiguation, opinion and sentiment analysis, syntactic and semantic parsing, entity-relation extraction, terminology extraction
- Machine Translation: statistical and neural machine translation, domain adaptation, multilingual resource creation, comparable corpora exploitation and data selection, format-preserving document translation, quality estimation
- Dialogue Systems: spoken language understanding, dialogue management, natural language generation
It carries out fundamental research in collaboration with universities and other research centres, publishing in high-impact conferences and journals, and co-supervising Master’s and PhD thesis.
It also collaborates with established businesses, SMEs and startups, adapting its speech and language processing technology to their specific use-cases and business models.
Its highly technology-oriented nature facilitates research and development of innovative applications in sectors related to the language industry (subtitling, transcription, translation, social media analysis, customer care, business intelligence) as well as more traditional sectors (tourism, security, health, assistive technology, education, manufacturing), in collaboration with the other Vicomtech departments.
The department has participated in and coordinated research and development projects at a regional, national and European level since 2001. It belongs to several alliances related to the promotion of speech and language technologies and the language industry: Langune, The Thematic Network of Speech Technology, META-NET and LT-Innovate.
Team members include doctors from internationally acclaimed universities specialising in the department’s main lines of research, researchers in doctoral training, and software developers.
Engineering & Technology
- Advanced manufacturing