Scholar
Pierre Marion
Google Scholar ID: Q8H5LgIAAAAJ
Inria - Ecole Normale Supérieure
Machine Learning
Deep Learning
Follow
Homepage
↗
Google Scholar
↗
Citations & Impact
All-time
Citations
441
H-index
10
i10-index
11
Publications
18
Co-authors
15
list available
Publications
11 items
Diffusion Removes Langevin's Conditioning Dependence: A Sharp Gaussian Analysis
2026
Cited
0
Edge Flow: A Tractable and Predictive Continuous-Time Model for Gradient Descent at the Edge of Stability
2026
Cited
0
Distilling LLM Feedback for Lean Theorem Proving
2026
Cited
0
Understanding diffusion models requires rethinking (again) generalization
2026
Cited
0
Exponential Convergence of (Stochastic) Gradient Descent for Separable Logistic Regression
2026
Cited
0
Clustering in Deep Stochastic Transformers
2026
Cited
0
Optimal Stopping in Latent Diffusion Models
2025
Cited
0
Phase Diagram of Dropout for Two-Layer Neural Networks in the Mean-Field Regime
2025
Cited
0
Load more
Resume
Academic Achievements
Insufficient information
Research Experience
- Research faculty at INRIA within the SIERRA team (a joint team between CNRS, Ecole Normale Supérieure and INRIA)
- Supervising multiple PhD and MSc students
Education
- Postdoctoral researcher, 2024-25, EPFL, Advisor: Lénaïc Chizat
- PhD student, Sorbonne Université, Advisors: Gérard Biau and Jean-Philippe Vert
Background
Research interests mainly focus on the theory of deep learning and starting to work on AI for maths. Concerned by climate change.
Miscellany
Strives to fly less and build viable alternatives to frequent intercontinental travel for the research community
Co-authors
13 total
Gérard Biau
Professor, Sorbonne University, Paris
Pierre L'Ecuyer
Professor, DIRO, University of Montreal
Adeline Fermanian
Califrais
Claire Boyer
Université Paris-Saclay
Yu-Han Wu
Google DeepMind co-advised PhD student, Sorbonne University
Francesco Piccinno
Google DeepMind
Raphaël Berthier
Inria Sorbonne Université
Peter Bartlett
Professor, EECS and Statistics, UC Berkeley