Pierre Humbert

Pierre Humbert is a postdoctoral researcher at Sorbonne Université's Laboratoire de Probabilités, Statistique et Modélisation (LPSM), affiliated with the MARS project. He previously held postdoctoral positions at Laboratoire de Mathématiques d'Orsay (LMO) and the INRIA Celeste team. Humbert completed his PhD in 2021 at ENS Paris-Saclay, focusing on multivariate analysis with tensors and graphs in neuroscience under Professors Nicolas Vayatis, Laurent Oudre, and Julien Audiffren.

His research interests span conformal prediction, statistical learning, non-parametric methods, robust statistics, signal processing, and applications in neuroscience. He contributes to federated learning frameworks, graph signal processing, and tensor-based methodologies for analyzing complex data structures.

Humbert's recent work emphasizes federated conformal prediction for privacy-preserving distributed learning and robust statistical techniques. His publications frequently address graph Laplacian estimation, EEG signal processing, and tensor decomposition applications in biomedical contexts. He collaborates actively with teams like INRIA Celeste and maintains open-source implementations of his algorithms.