Profile

Quentin Bouniot is an Assistant Professor at Télécom Paris, Institut Polytechnique de Paris. He was previously a Post-Doctoral researcher in the Explainable Machine Learning group at TUM and Helmholtz Munich, led by Prof. Zeynep Akata, and before that at Télécom Paris, with Prof. Florence d'Alché-Buc and Prof. Pavlo Mozharovskyi. He got his PhD from Université Jean-Monnet Saint-Étienne, under the supervision of Prof. Amaury Habrard, and CEA-List, under the supervision of Romaric Audigier and Angélique Loesch. He graduated from CentraleSupélec (France) in 2019.

His main line of research revolves around Deep Representation Learning, and his current research interests include:

• mechanistic interpretability of representations in foundation models across modalities
• alignment of representation between modalities, mainly vision-language
• robustness, distribution shift, uncertainty quantification, learning with limited labeled data