Normalizing flow sampling with Langevin dynamics in the latent space

Autor: Coeurdoux, Florentin, Dobigeon, Nicolas, Chainais, Pierre

Normalizing flows (NF) use a continuous generator to map a simple latent (e.g. Gaussian) distribution, towards an empirical target distribution associated with a training data set. Once trained by minimizing a variational objective, the learnt map pr

Externí odkaz: http://arxiv.org/abs/2305.12149

Sliced-Wasserstein normalizing flows: beyond maximum likelihood training

Autor: Coeurdoux, Florentin, Dobigeon, Nicolas, Chainais, Pierre

Despite their advantages, normalizing flows generally suffer from several shortcomings including their tendency to generate unrealistic data (e.g., images) and their failing to detect out-of-distribution data. One reason for these deficiencies lies i

Externí odkaz: http://arxiv.org/abs/2207.05468

Tento výsledek nelze pro nepřihlášené uživatele zobrazit.
K zobrazení výsledku je třeba se přihlásit.

Méthode MCMC plug-and-play avec a priori génératif profond

Autor: Coeurdoux, Florentin, Dobigeon, Nicolas, Chainais, Pierre

Publikováno v: GRETSI 2023
XXIXème Colloque Francophone de Traitement du Signal et des Images (GRETSI 2023)
XXIXème Colloque Francophone de Traitement du Signal et des Images (GRETSI 2023), Aug 2023, Grenoble, France. à paraître

International audience; Plug-and-Play (PnP) methods are a class of iterative algorithms for inverse problems solving in which regularization is provided by a generic denoiser. Although producing very good results, these PnP optimization methods produ

Externí odkaz: https://explore.openaire.eu/search/publication?articleId=od______4254::ae621017b70a4d0c8949b3a4b7d71113
https://hal.science/hal-04154783/document

Tento výsledek nelze pro nepřihlášené uživatele zobrazit.
K zobrazení výsledku je třeba se přihlásit.