Monte Carlo guided Diffusion for Bayesian linear inverse problems | Vidéo | Carmin.tv

00:00:00 / 00:00:00

Monte Carlo guided Diffusion for Bayesian linear inverse problems

De Sylvain Le Corff

Apparaît dans la collection : Autumn school in Bayesian Statistics / École d'automne en statistique bayésienne

Ill-posed linear inverse problems that combine knowledge of the forward measurement model with prior models arise frequently in various applications, from computational photography to medical imaging. Recent research has focused on solving these problems with score-based generative models (SGMs) that produce perceptually plausible images, especially in inpainting problems. In this study, we exploit the particular structure of the prior defined in the SGM to formulate recovery in a Bayesian framework as a Feynman–Kac model adapted from the forward diffusion model used to construct score-based diffusion. To solve this Feynman–Kac problem, we propose the use of Sequential Monte Carlo methods. The proposed algorithm, MCGdiff, is shown to be theoretically grounded and we provide numerical simulations showing that it outperforms competing baselines when dealing with ill-posed inverse problems.

Informations sur la vidéo

Date de captation 31/10/2023
Date de publication 27/11/2023
Institut CIRM
Licence CC BY NC ND
Langue Anglais
Audience Chercheurs, Doctorants
Réalisateur(s) Luca Recanzone
Format MP4

Données de citation

DOI 10.24350/CIRM.V.20107003
Citer cette vidéo Le Corff, Sylvain (31/10/2023). Monte Carlo guided Diffusion for Bayesian linear inverse problems. CIRM. Audiovisual resource. DOI: 10.24350/CIRM.V.20107003
URL https://dx.doi.org/10.24350/CIRM.V.20107003

Domaine(s)

Bibliographie

CARDOSO, Gabriel, IDRISSI, Yazid Janati El, CORFF, Sylvain Le, et al. Monte Carlo guided Diffusion for Bayesian linear inverse problems. arXiv preprint arXiv:2308.07983, 2023. - https://arxiv.org/abs/2308.07983

Codes MSC

Dernières questions liées sur MathOverflow

Pour poser une question, votre compte Carmin.tv doit être connecté à mathoverflow

Poser une question sur MathOverflow

Copyright Carmin.tv 2024

Donner son avis