Danish Foundation Models - We strengthen the Danish language in the digital age

Danish Foundation Models (DFM) works with machine learning and Natural Language Processing (NLP) with a special focus on Danish. The project is part of the government's strategic AI effort.

Our goal is to develop, maintain and provide open access to large Danish language models. DFM wants to promote innovation and ensure that language technology is accessible and inclusive for all.

The Danish society must develop transparent language models to solve tasks, where understanding Danish norms and values as well as the Danish language is important.

A secure, transparent platform must be established for the development and training of language models, to which Danish companies and authorities must have access.

The effort is part of Danish Foundations Models – a collaboration between Aarhus University, the University of Southern Denmark, the University of Copenhagen and the Alexandra Institute.  See also Strategisk indsats for kunstig intelligens (2024), action 3 (in Danish). The models will be made freely available to companies and authorities, who will thus be able to use and further develop Danish language models.

The Centre for Language Technology contributes with benchmark data for Danish and with post training of language models such as cultural adaption of the models to the Danish society.