Denoising Diffusion Probabilistic Models, Jonathan Ho, Ajay Jain, Pieter Abbeel, 2020Advances in Neural Information Processing Systems, Vol. 33 (NeurIPS)DOI: 10.55917/cb9f6785 - The original paper that introduced and popularized the Denoising Diffusion Probabilistic Model (DDPM), detailing its forward and reverse processes, training objective, and generative capabilities.
Deep Unsupervised Learning using Nonequilibrium Thermodynamics, Jascha Sohl-Dickstein, Eric Weiss, Niru Maheswaranathan, Surya Ganguli, 2015Proceedings of the 32nd International Conference on Machine Learning, Vol. 37 (PMLR (Proceedings of Machine Learning Research))DOI: 10.5555/3045118.3045233 - The foundational paper introducing the concept of diffusion models by drawing an analogy to non-equilibrium thermodynamics, providing the theoretical basis for later DDPM developments.
Generative Modeling by Estimating Gradients of the Data Distribution, Yang Song, Stefano Ermon, 2019Advances in Neural Information Processing Systems, Vol. 32 (Curran Associates, Inc.)DOI: 10.55917/cb0b8364 - Introduced score-based generative models and score matching, establishing a strong theoretical connection to the simplified training objective used in DDPMs.