On April 2nd, Dávid í Lág started a new PhD research position at the University of the Faroe Islands. Dávid holds a degree in computer science with a specialization in data science and machine learning from the IT University of Copenhagen.
The PhD project aims to research automatic speech recognition (ASR) for smaller languages, focusing on the Faroese language. In recent years, ASR has become a part of the technology we use daily, but not to the same extent for Faroese. The reason for this is that a good speech recognizer has, until now, required thousands of hours of recordings with corresponding texts before it performs well.
In the summer of 2022 the BLARK by the Ravnur Project was released. BLARK is an acronym for Basic Language Resource Kit, and this dataset was created to develop Faroese speech recognition. The dataset is a collection of high-quality recordings with speakers from various dialectal regions and age groups. By using modern AI techniques in machine learning and employing multilingual ASR models that include over 100 languages, it is now possible to create a Faroese speech recognizer that performs well, even with a relatively small data foundation.
The PhD project will also explore the possibilities of generating new data from existing datasets, producing audio from text, and using knowledge transfer between languages.
The principal supervisor of the PhD project is Jón Guðnason, a professor at the University of Reykjavík with a PhD in signal processing. Barbara Scalvini at the Faculty of Science and Technology is the co-supervisor. Additionally, Iben Nyholm Debess at the Faculty of Faroese Language and Literature and Annika Simonsen at the University of Iceland are also associated with the project as linguists.
