Florentin Coeurdoux

2 accepted papers

2026

Probing the Geometry of Diffusion Models with the String Method

Elio Moreau, Florentin Coeurdoux, Grégoire Ferré, Eric Vanden-Eijnden

ICML 2026poster

Understanding the geometry of learned distributions is fundamental to improving and interpreting diffusion models, yet systematic tools for exploring their landscape remain limited. Standard latent-space interpolations fail to respect the structure of the learned distribution, often traversing low-d…

Cited by 0 Source ↗Scholar ↗

2025

Multitask Learning with Stochastic Interpolants

Hugo Negrel, Florentin Coeurdoux, Michael Samuel Albergo, Eric Vanden-Eijnden

NeurIPS 2025spotlight

We propose a framework for learning maps between probability distributions that broadly generalizes the time dynamics of flow and diffusion models. To enable this, we generalize stochastic interpolants by replacing the scalar time variable with vectors, matrices, or linear operators, allowing us to…

Cited by 0 Source ↗Scholar ↗