GM-DDPM: Denoising diffusion probabilistic models with Gaussian Mixture Noise
Keywords: Diffusion model, Image generation
Abstract: Denoising diffusion probabilistic models (DDPM) have shown impressive performance in various domains as a class of deep generative models. In this paper, we introduce the Gaussian Mixture noise-based DDPM (GM-DDPM), which considers the Markov diffusion posterior as a Gaussian mixture model. Specifically, GM-DDPM randomly selects a Gaussian component and then add the chosen Gaussian noise, which can be demonstrated as more efficient way to perturb the signals into a simple known distribution. We further define the reverse probabilistic model as a parameterized Gaussian mixture kernel. Due to the intractability in calculating the KL divergence between Gaussian mixture models, we derive a variational bound to maximize the likelihood, offering a concise formulation for optimizing the denoising model and valuable insights for designing the sampling strategies. Our theoretical derivation highlights that GM-DDPM only requires the inclusion of a random offset in both the diffusion and reverse processes, which can be efficiently implemented with just several lines of code. Furthermore, we present three streamlined sampling strategies that interface with diverse fast dedicated solvers for diffusion ordinary differential equations, boosting the efficacy of image representation in the sampling phase and alleviating the issue of slow generation speed, thereby enhancing both efficiency and accuracy. Extensive experiments on benchmark datasets demonstrate the effectiveness of GM-DDPM and its superiority over the original DDPM.
Supplementary Material: zip
Primary Area: generative models
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2024/AuthorGuide.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors' identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 1360
Loading