# Denoising Diffusion Probabilistic Models
**Authors**: Jonathan Ho, Ajay Jain, Pieter Abbeel
**Year**: 2020
**Source**: arXiv:2006.11239
**URL**: https://arxiv.org/abs/2006.11239
## Summary
The paper that refined and popularized diffusion-based image generation, introducing **DDPMs** ([[Denoising Diffusion Probabilistic Models (DDPMs)]]). Ho et al. established the practical training objective and the denoising score matching connection that made diffusion models viable at scale. Their results on CIFAR-10 and LSUN benchmarks demonstrated quality competitive with GANs, sparking the wave of diffusion-based generators like [[list-Tools_&_platforms#Stable Diffusion|Stable Diffusion]] and [[list-Tools_&_platforms#Midjourney|Midjourney]].