An overview of diffusion models for generative artificial intelligence

TL;DR

This paper offers a comprehensive, mathematically rigorous overview of diffusion models in generative AI, covering foundational concepts, training, generation, and recent advancements like improved variants and latent diffusion models.

Contribution

It provides a detailed mathematical framework and synthesizes recent extensions and improvements in diffusion models for generative AI.

Findings

Detailed mathematical framework for DDPMs

Review of recent extensions like classifier-free guidance

Comparison of different diffusion model variants

Abstract

This article provides a mathematically rigorous introduction to denoising diffusion probabilistic models (DDPMs), sometimes also referred to as diffusion probabilistic models or diffusion models, for generative artificial intelligence. We provide a detailed basic mathematical framework for DDPMs and explain the main ideas behind training and generation procedures. In this overview article we also review selected extensions and improvements of the basic framework from the literature such as improved DDPMs, denoising diffusion implicit models, classifier-free diffusion guidance models, and latent diffusion models.