Bruno Loureiro

I am a senior researcher (directeur de recherche) at the Computer Science department of École Normale Supérieure in Paris, where I lead a research group working at the intersection of machine learning and statistical mechanics, with a particular interest in developing mathematical theory for modern machine learning models.

My group is based at the interdisciplinary Centre for Data Science, and is part of the DATA team. I am also a Professeur Attaché at Université Paris Sciences et Lettres (PSL), where I teach in undergraduate and graduate programmes.

You can find more about my group and a complete list of my publications.

news

Sep 24, 2026	Our works: How does feature learning change the function space evolution? Optimal scaling laws in learning hierarchical multi-index models Annealing in variational inference mitigates mode collapse: a theoretical study on Gaussian mixtures The balance between feature learning and collapse in generative dynamical systems were accepted for NeurIPS 2026. Come see our posters in Paris!
May 01, 2026	Our works: Sharp description of local minima in the loss landscape of high-dimensional two-layer ReLU neural networks A Noise Sensitivity Exponent Controls Large Statistical-to-Computational Gaps in Single- and Multi-Index Models (Spotlight) A Random Matrix Theory of Masked Self-Supervised Regression On the existence of consistent adversarial attacks in high-dimensional linear classification (Spotlight) were accepted for ICML 2026, two of them as spotlights! Come see the talks and posters in Seoul!
Apr 19, 2026	Excited to co-organise the third edition of the Lausanne Event on Machine Learning and Neural Network Theory (LemanTh) from April 20-24 at the Bernouilli Center (EPFL), together with Florent Krzakala.
Feb 06, 2026	Our works: Scaling Laws and Spectra of Shallow Neural Networks in the Feature Learning Regime (Oral) Fast Escape, Slow Convergence: Learning Dynamics of Phase Retrieval under Power-Law Data (Oral) Statistical Advantage of Softmax Attention: Insights from Single-Location Regression were accepted for ICLR 2026, two of them as orals! Come see the talks and posters in Rio de Janeiro!
Jan 26, 2026	I am co-organising a fourth edition of the PSL Week Statistical Physics and Machine Learning” together with Antoine Maillard. Inscriptions are open!