Regularity of the score and convergence rates of generative diffusion models
| Quand ? |
Le 24/03/2026, de 13:00 à 14:00 |
|---|---|
| Où ? | M7 101 |
| Participants |
Arthur Stéphanovitch |
| Ajouter un événement au calendrier |
vCal
iCal
|
Arthur Stéphanovitch
Title: Regularity of the score and convergence rates of generative diffusion models
Abstract: We show that diffusion-based generative models adapt to the smoothness of the target distribution: the learned score function inherits the target’s regularity. Leveraging this adaptivity, we obtain a concise proof that diffusion models achieve minimax-optimal rates for density estimation.
Bio: Arthur Stéphanovitch is a postdoctoral researcher at ENSAE working with Austin Stromme on the implicit biases induced by training neural networks. He completed his PhD at the ENS under the supervision of Clément Levrard and Eddie Aamari on generative models.
Website: https://sites.google.com/view/arthur-stephanovitch/
In Room M7 101, 1st floor, Monod campus, ENSL (to be confirmed).