CV

Education

• PhD in applied mathematics and machine learningSorbonne University (LPSM) and Google DeepMind, ParisAdvisors: Gérard Biau, Claire Boyer and Pierre Marion; Quentin Berthet and Romuald Elie at Google DeepMind.
  2024 – 2027 (expected)
• Diploma of the École Normale Supérieure, applied mathematicsÉcole Normale Supérieure, Paris
  2024
• MSc in statistics and machine learningUniversité Paris-Saclay
  2023
• BSc in applied mathematicsUniversité Paris Cité
  2021
• Classe préparatoire (MPSI/MP)Lycée Louis-le-Grand, Paris
  2020

Publications

Conference and journal papers

1. Pierre Marion and Yu-Han Wu (2026). Understanding diffusion models requires rethinking (again) generalization. arXiv:2605.06077. page
2. Quentin Berthet, Yu-Han Wu, Clément Crepy, Romuald Elie, Klaus Greff and Michael E. Sander (2026). MIND: Monge Inception Distance for Generative Models Evaluation. arXiv:2605.06797. page
3. Yu-Han Wu, Quentin Berthet, Gérard Biau, Claire Boyer, Romuald Elie and Pierre Marion (2026). Optimal Stopping in Latent Diffusion Models. In International Conference on Machine Learning (ICML). Oral at the PriGM workshop, EurIPS 2025 page
4. Yu-Han Wu, Pierre Marion, Gérard Biau and Claire Boyer (2025). Taking a Big Step: Large Learning Rates in Denoising Score Matching Prevent Memorization. In Proceedings of the Thirty Eighth Conference on Learning Theory (COLT), PMLR 291:5718–5756. page
5. Pierre Marion, Yu-Han Wu, Michael E. Sander and Gérard Biau (2024). Implicit regularization of deep residual networks towards neural ODEs. In The Twelfth International Conference on Learning Representations (ICLR). Spotlight page
6. David Conlon and Yu-Han Wu (2023). More on lines in Euclidean Ramsey theory. Comptes Rendus. Mathématique, 361(G5), 897–901. page

Preprints

1. Guillaume Couairon, Alexis Jacq, Yu-Han Wu, Renu Singh, Yana Hasson, Quentin Berthet and Romuald Elie (2026). Kastor: An efficient fine-tuning strategy for generative emulation of PDE simulations. arXiv:2608.06107. page
2. DiffusionGemma Team (2026). DiffusionGemma Technical Report. arXiv:2608.00146. page

Talks and posters

• Invited talk at Optimisation et Apprentissage appliqués aux contenus numériques MathSTIC, USPN, France
  Jun 2026
• Poster at Lemanth 2026 EPFL Bernoulli Center, Lausanne, Switzerland
  Apr 2026
• Oral at PriGM Workshop, EurIPS 2025 Copenhagen Bella Center, Copenhagen, Denmark
  Dec 2025
• Talk at COLT 2025 ENS Lyon, Lyon, France
  Jul 2025
• Talk at Journées de Statistique 2025 Université Aix-Marseille, Marseille, France
  Jul 2025

Distinctions

• Oral at the PriGM workshop, EurIPS 2025Optimal Stopping in Latent Diffusion Model
  2025
• Spotlight, ICLR 2024Implicit regularization of deep residual networks towards neural ODEs
  2023

Research experience

• Research internOwkin, ParisFederated learning and invariant risk minimization. Supervisor Ulysse Marteau-Ferey.
  Winter 2023
• Research internSorbonne University (LPSM)Implicit regularization of residual networks. Supervisors Gérard Biau and Pierre Marion.
  Summer 2023
• Research internCalifornia Institute of TechnologyExtremal combinatorics and Euclidean Ramsey theory. Supervisor David Conlon.
  Summer 2022

Service

• Reviewer, ICML
  2026
• Reviewer, NeurIPS
  2026
• Reviewer, PriGM workshop at EurIPS
  2025 – 2026

Skills

• Mathematics probability, statistics, optimization, stochastic analysis
• Programming Python (JAX, PyTorch, NumPy), LaTeX, Git