Hugo Lebeau
Pour voir un beau Gauss, clique ici.
Research Topics
I am broadly interested in statistical and algorithmic aspects of machine learning with high-dimensional data, notably through the study of random matrix and tensor models. In particular, this includes,
- spiked random matrix and tensor models,
- random tensors and their low-rank approximations,
- theoretical study of complex clustering tasks (multi-view, time-varying),
- learning with memory constraints, compressive learning.
Random matrix theory is a powerful tool to study practical statistical settings of the big data era.
Publications
- JMLR – A Random Matrix Approach to Low-Multilinear-Rank Tensor Approximation, Hugo Lebeau, Florent Chatelain, Romain Couillet [article]
- IEEE Signal Processing Letters – Asymptotic Gaussian Fluctuations of Eigenvectors in Spectral Clustering, Hugo Lebeau, Florent Chatelain, Romain Couillet [article (IEEE)][article (arXiv)][code]
- ICLR 2024 – Performance Gaps in Multi-view Clustering under the Nested Matrix-Tensor Model, Hugo Lebeau, Mohamed El Amine Seddik, José Henrique de Morais Goulart [article][code]
- ICML 2022 – A Random Matrix Analysis of Data Stream Clustering: Coping With Limited Memory Resources, Hugo Lebeau, Romain Couillet, Florent Chatelain [article][code][video]
- GRETSI 2025 – Performance of Rank-One Tensor Approximation on Incomplete Data, Hugo Lebeau [article]
- GRETSI 2023 – HOSVD Tronquée : Analyse d'une Approximation Tensorielle Rapide, Hugo Lebeau, Romain Couillet, Florent Chatelain [article]
- GRETSI 2022 – Une analyse par matrices aléatoires du clustering en ligne : comprendre l’impact des limitations en mémoire, Hugo Lebeau, Romain Couillet, Florent Chatelain [article]
PhD Thesis
I was supervised by Romain Couillet and Florent Chatelain, and defended my PhD in March 2025.
My manuscript Random Matrix and Tensor Models for Large Data Processing is available here.
- Want to know what this is about? Check the abstract on page 3!
- Want to know more? Great! The introduction (16 pages) will give you a broad overview of the main challenges.
- Still interested? Feel free to dive into the chapters. Chapter 2 may be a prerequisite if you want to fully understand the following ones.
Teaching
