Sofie Labat

Sofie Labat conducts research on the computational modeling of affect in social interaction, leveraging language technology to analyze and interpret emotional cues. Her work bridges psychology and computational linguistics to better understand how emotions are expressed, regulated, and perceived in communicative contexts. More specifically, her research interests include:

• Detection of emotion trajectories in conversations
• The interplay between emotions, behavioral coping strategies, and interpersonal emotion regulation
• Perspectivism in data collection and emotion modeling
• Social biases in machine learning models and methods for debiasing
• Multimodal affective and social signals across human–human and human–machine interactions

Sofie obtained her PhD in October 2025 with a dissertation titled "Affective Threads: Data-Driven Modeling of Emotional Processes in Social Interaction".  Her research was co-supervised by Prof. Dr. Véronique Hoste and Prof. Dr. Ir. Thomas Demeester. From October 2019 until October 2021, she contributed to the human-like AI research challenge within the Flanders AI Research Program. In November 2021, she was awarded a PhD fellowship for Strategic Basic Research (SB) by the Research Foundation – Flanders (FWO). Her doctoral work focused on modeling emotion trajectories in social interaction, specifically within the domain of customer service. Over time, she broadened her scope to explore the relationship between emotions and coping strategies in the context of online micro-aggressions. Between September and December 2022, Sofie conducted a research visit at the Institute for Natural Language Processing (IMS) of the University of Stuttgart, Germany, where she collaborated with Prof. Dr. Roman Klinger and his team.

As an affiliated member of the Language and Translation Technology Team (LT3), Sofie is currently a visiting scholar at the Digital Emotions Lab, led by Prof. Dr. Amit Goldenberg, within the Digital Data Design Institute (D^3) at Harvard University. During her stay, she is investigating bias in AI systems, with a particular focus on emotion detection and regulation. Her research visit is supported by a postdoctoral fellowship from the Belgian American Educational Foundation (BAEF).