On diffusion-based generative models and their error bounds: The log-concave case with full convergence estimates

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Bruno, Stefano, Zhang, Ying, Lim, Dong-Young, Akyildiz, Ömer Deniz, Sabanis, Sotirios
Format: Preprint
Publié: 2023
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866915153221517312
author Bruno, Stefano
Zhang, Ying
Lim, Dong-Young
Akyildiz, Ömer Deniz
Sabanis, Sotirios
author_facet Bruno, Stefano
Zhang, Ying
Lim, Dong-Young
Akyildiz, Ömer Deniz
Sabanis, Sotirios
contents We provide full theoretical guarantees for the convergence behaviour of diffusion-based generative models under the assumption of strongly log-concave data distributions while our approximating class of functions used for score estimation is made of Lipschitz continuous functions avoiding any Lipschitzness assumption on the score function. We demonstrate via a motivating example, sampling from a Gaussian distribution with unknown mean, the powerfulness of our approach. In this case, explicit estimates are provided for the associated optimization problem, i.e. score approximation, while these are combined with the corresponding sampling estimates. As a result, we obtain the best known upper bound estimates in terms of key quantities of interest, such as the dimension and rates of convergence, for the Wasserstein-2 distance between the data distribution (Gaussian with unknown mean) and our sampling algorithm. Beyond the motivating example and in order to allow for the use of a diverse range of stochastic optimizers, we present our results using an $L^2$-accurate score estimation assumption, which crucially is formed under an expectation with respect to the stochastic optimizer and our novel auxiliary process that uses only known information. This approach yields the best known convergence rate for our sampling algorithm.
format Preprint
id arxiv_https___arxiv_org_abs_2311_13584
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle On diffusion-based generative models and their error bounds: The log-concave case with full convergence estimates
Bruno, Stefano
Zhang, Ying
Lim, Dong-Young
Akyildiz, Ömer Deniz
Sabanis, Sotirios
Machine Learning
Optimization and Control
Probability
We provide full theoretical guarantees for the convergence behaviour of diffusion-based generative models under the assumption of strongly log-concave data distributions while our approximating class of functions used for score estimation is made of Lipschitz continuous functions avoiding any Lipschitzness assumption on the score function. We demonstrate via a motivating example, sampling from a Gaussian distribution with unknown mean, the powerfulness of our approach. In this case, explicit estimates are provided for the associated optimization problem, i.e. score approximation, while these are combined with the corresponding sampling estimates. As a result, we obtain the best known upper bound estimates in terms of key quantities of interest, such as the dimension and rates of convergence, for the Wasserstein-2 distance between the data distribution (Gaussian with unknown mean) and our sampling algorithm. Beyond the motivating example and in order to allow for the use of a diverse range of stochastic optimizers, we present our results using an $L^2$-accurate score estimation assumption, which crucially is formed under an expectation with respect to the stochastic optimizer and our novel auxiliary process that uses only known information. This approach yields the best known convergence rate for our sampling algorithm.
title On diffusion-based generative models and their error bounds: The log-concave case with full convergence estimates
topic Machine Learning
Optimization and Control
Probability
url https://arxiv.org/abs/2311.13584