معرفی
Sofie Labat is a Researcher at Ghent University affiliated with the Language and Translation Technology Team (LT3). She is pursuing her PhD under the supervision of Prof. Véronique Hoste and Prof. Thomas Demeester, focusing on emotion-aware conversational systems. Funded by an FWO PhD fellowship since November 2021, she previously contributed to the Flanders AI Program's 'human-like AI' research challenge (2019-2021) and conducted collaborative research at Stuttgart University's IMS institute in 2022.
Her research centers on emotion trajectory modeling in conversations, examining the interplay between emotions and coping strategies while addressing perspectivism in emotion data collection. She investigates social biases in machine learning systems and develops debiasing methodologies, with emphasis on multimodal emotional/social signals in human-human and human-machine interactions. Her work bridges computational linguistics, affective computing, and human-centered AI design for customer service applications.
Analysis of her 15 most recent publications (2019-2025) reveals a cohesive research trajectory in emotion-aware conversational AI. She pioneered specialized corpora like EmoTwiCS (Dutch customer service dialogues) and EmoProgress (emotion progression in dreams/dialogues), advancing multimodal emotion annotation frameworks. Her work consistently integrates transformer models with domain-specific constraints while addressing bias and perspective issues. Earlier contributions in cognate detection demonstrate methodological versatility across computational linguistics subfields.
Her scientific recognition includes:
- FWO PhD fellowship (strategic basic research) awarded November 2021
As a doctoral researcher, Labat has not supervised students but has secured competitive funding through the FWO fellowship and Flanders AI Program. Her research leverages interdisciplinary collaboration, particularly with LT3's core team and international partners like Stuttgart University's IMS group. Current work extends emotion trajectory modeling to multimodal contexts while refining bias mitigation techniques for real-world deployment.
Labat operates within Ghent University's Language and Translation Technology Team (LT3), a leading NLP research group. Her international collaboration with Prof. Roman Klinger's IMS team at Stuttgart University demonstrates her integration into European AI research networks. Within LT3, she contributes to emotion-aware system development while mentoring junior researchers through project-based involvement in the METRICS PhD project.
