Norwegian text classifier optimisation with active learning

AI & ML interests

The research is carried out as part of a master's thesis for Norwegian University of Science and Technology (NTNU) and Kantega AS. The work is carried out by the students Aabol, Simen Tvete and Dragsten, Marcus Klomsten. Labeling datasets can be an expensive process, and alleviating this workload can enable organisations to make use of their possibly extensive unlabeled datasets. The main focus of this project is to study active learning in relation to text classification tasks. In particular, classification of Norwegian data through pre-trained Norwegian models.