arthur.stephanovitch@ensae.fr

I am a postdoctoral researcher at ENSAE-CREST working with Austin Stromme.

Before that, I was a PhD student at DMA (École normale supérieure, Paris) under the supervision of Clément Levrard and Eddie Aamari. Here is the manuscript thesis: Theoretical foundation of generative models.

Publications and preprints:

Wasserstein GANs are Minimax Optimal Distribution Estimators, A. Stéphanovitch, E. Aamari, C. Levrard . The Annals of Statistics (2024). arxiv
Optimal 1-Wasserstein distance for WGANs, A.Stéphanovitch U. Tanielian, B. Cadre, N. Klutchnikoff, G. Biau. Bernoulli (2024). arxiv
Generalization bounds for score-based generative models: a synthetic proof, A. Stéphanovitch, E. Aamari, C. Levrard . Preprint. arxiv
Regularity of the score function in generative models, A. Stéphanovitch . Preprint. arxiv
Smooth transport map via diffusion process, A. Stéphanovitch . Preprint. arxiv
Integral Probability Metrics on submanifolds: interpolation inequalities and optimal inference, A. Stéphanovitch . Preprint. arxiv
Generative model for optimal density estimation on unknown manifold, A. Stéphanovitch . Preprint arxiv
Optimal Transport through a toll, A. Stephanovitch A. Dong, T. Georgiou. European journal of applied mathematics (2024). arxiv
Monge–Kantorovich OT through constrictions and flow-rate constraints, A. Dong, A. Stephanovitch, T. Georgiou. Automatica (2024). arxiv

Teaching:

Deep learning, M2 Modelisation (2022-2025)
Bayesian statistics, M1 mathématiques fondamentales (2023-2025)
Research internship supervision, L3 mathématiques ENS (2023-2024)
Examens oraux, L1 mathématiques (2023-2024)
Group theory and arithmetic, L2 mathématiques et informatique (2022-2023)

Talks:

Page updated

Google Sites

Report abuse