Posted on Dec. 5, 2022
Last Thursday, 2 December 2022, Luna De Bruyne successfully defended her PhD entitled Feeling EmotioNL: Automatically detecting emotions in Dutch texts. She was supervised by Professor Véronique Hoste and Professor Orphée De Clercq.
The past years Luna has developed a framework for Dutch emotion detection. She collected a corpus of Dutch tweets and captions which have been annotated in a bi-representational manner, i.e. with emotional categories (anger, fear, joy, love and sadness) and with the emotional dimensions valence, aroudal and dominance. This dataset was used to investigate emotion detection methodologies using transformer models and led to the first state-of-the-art Dutch emotion detection model. In a next phase, it was assessed how well this model generalizes across different domains, emotion frameworks and languages. In a final step the external validity of the model was tested by applying it in the context of crisis communication.
The actual defence was preceded by the workshop "Feeling EmotioNL".
Congratulations dr. De Bruyne!