Conditional sampling within generative diffusion models

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Zhao, Zheng, Luo, Ziwei, Sjölund, Jens, Schön, Thomas B.
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866929719801282560
author Zhao, Zheng
Luo, Ziwei
Sjölund, Jens
Schön, Thomas B.
author_facet Zhao, Zheng
Luo, Ziwei
Sjölund, Jens
Schön, Thomas B.
contents Generative diffusions are a powerful class of Monte Carlo samplers that leverage bridging Markov processes to approximate complex, high-dimensional distributions, such as those found in image processing and language models. Despite their success in these domains, an important open challenge remains: extending these techniques to sample from conditional distributions, as required in, for example, Bayesian inverse problems. In this paper, we present a comprehensive review of existing computational approaches to conditional sampling within generative diffusion models. Specifically, we highlight key methodologies that either utilise the joint distribution, or rely on (pre-trained) marginal distributions with explicit likelihoods, to construct conditional generative samplers.
format Preprint
id arxiv_https___arxiv_org_abs_2409_09650
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Conditional sampling within generative diffusion models
Zhao, Zheng
Luo, Ziwei
Sjölund, Jens
Schön, Thomas B.
Machine Learning
Generative diffusions are a powerful class of Monte Carlo samplers that leverage bridging Markov processes to approximate complex, high-dimensional distributions, such as those found in image processing and language models. Despite their success in these domains, an important open challenge remains: extending these techniques to sample from conditional distributions, as required in, for example, Bayesian inverse problems. In this paper, we present a comprehensive review of existing computational approaches to conditional sampling within generative diffusion models. Specifically, we highlight key methodologies that either utilise the joint distribution, or rely on (pre-trained) marginal distributions with explicit likelihoods, to construct conditional generative samplers.
title Conditional sampling within generative diffusion models
topic Machine Learning
url https://arxiv.org/abs/2409.09650