Vincent Souveton

I hold a Starting Research Position at Inria, in the Atlantis project-team, where I work on neural operators for Maxwell’s equations.

My main goal is to accelerate traditional numerical simulations using machine learning. I am interested in neural operators, neural generative models and Markov chain Monte Carlo methods applied to a broad range of physics applications.

Before Inria, I was a postdoctoral researcher at CEA, building fast surrogate models for blast-wave propagation with uncertainty quantification. I did my PhD at Laboratoire de Mathématiques Blaise Pascal (Université Clermont Auvergne) on non-reversible and generative sampling algorithms for cosmology, as a member of the Aquila consortium.