Modelos Generativos basados en Mecanismos de Difusi\'on

TL;DR

Diffusion-based generative models mimic physical diffusion processes to generate new images and signals, using neural networks trained to reverse progressive corruption, with theoretical foundations and applications explained.

Contribution

This paper presents the theoretical basis and applications of diffusion models for generative tasks, emphasizing their reversible corruption process and relevance to Spanish-speaking researchers.

Findings

Neural networks can effectively reverse progressive pixel corruption.

Diffusion models can generate images from random noise.

Theoretical foundations support diverse applications.

Abstract

Diffusion-based generative models are a design framework that allows generating new images from processes analogous to those found in non-equilibrium thermodynamics. These models model the reversal of a physical diffusion process in which two miscible liquids of different colors progressively mix until they form a homogeneous mixture. Diffusion models can be applied to signals of a different nature, such as audio and image signals. In the image case, a progressive pixel corruption process is carried out by applying random noise, and a neural network is trained to revert each one of the corruption steps. For the reconstruction process to be reversible, it is necessary to carry out the corruption very progressively. If the training of the neural network is successful, it will be possible to generate an image from random noise by chaining a number of steps similar to those used for image deconstruction at training time. In this article we present the theoretical foundations on which this method is based as well as some of its applications. This article is in Spanish to facilitate the arrival of this scientific knowledge to the Spanish-speaking community.