The Superposition of Diffusion Models Using the Itô Density Estimator

• URL: http://arxiv.org/abs/2412.17762v2
• Date: Fri, 28 Feb 2025 23:00:23 GMT
• Title: The Superposition of Diffusion Models Using the Itô Density Estimator
• Authors: Marta Skreta, Lazar Atanackovic, Avishek Joey Bose, Alexander Tong, Kirill Neklyudov,
• Abstract summary: We show that SuperDiff is scalable to large pre-trained diffusion models.<n>We also show that SuperDiff is efficient during inference time, and mimics traditional composition operators.<n>We empirically demonstrate the utility of using SuperDiff for generating more diverse images on CIFAR-10.
• Score: 46.03684204456143
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: The Cambrian explosion of easily accessible pre-trained diffusion models suggests a demand for methods that combine multiple different pre-trained diffusion models without incurring the significant computational burden of re-training a larger combined model. In this paper, we cast the problem of combining multiple pre-trained diffusion models at the generation stage under a novel proposed framework termed superposition. Theoretically, we derive superposition from rigorous first principles stemming from the celebrated continuity equation and design two novel algorithms tailor-made for combining diffusion models in SuperDiff. SuperDiff leverages a new scalable It\^o density estimator for the log likelihood of the diffusion SDE which incurs no additional overhead compared to the well-known Hutchinson's estimator needed for divergence calculations. We demonstrate that SuperDiff is scalable to large pre-trained diffusion models as superposition is performed solely through composition during inference, and also enjoys painless implementation as it combines different pre-trained vector fields through an automated re-weighting scheme. Notably, we show that SuperDiff is efficient during inference time, and mimics traditional composition operators such as the logical OR and the logical AND. We empirically demonstrate the utility of using SuperDiff for generating more diverse images on CIFAR-10, more faithful prompt conditioned image editing using Stable Diffusion, as well as improved conditional molecule generation and unconditional de novo structure design of proteins. https://github.com/necludov/super-diffusion

Related papers

• Heavy-Tailed Diffusion Models [38.713884992630675]
  We show that traditional diffusion and flow-matching models fail to capture heavy-tailed behavior. We address this by repurposing the diffusion framework for heavy-tail estimation. We introduce t-EDM and t-Flow, extensions of existing diffusion and flow models that employ a Student-t prior.
  arXiv  Detail & Related papers  (2024-10-18T04:29:46Z)
• Broadening Target Distributions for Accelerated Diffusion Models via a Novel Analysis Approach [49.97755400231656]
  We show that a novel accelerated DDPM sampler achieves accelerated performance for three broad distribution classes not considered before. Our results show an improved dependency on the data dimension $d$ among accelerated DDPM type samplers.
  arXiv  Detail & Related papers  (2024-02-21T16:11:47Z)
• Guided Diffusion from Self-Supervised Diffusion Features [49.78673164423208]
  Guidance serves as a key concept in diffusion models, yet its effectiveness is often limited by the need for extra data annotation or pretraining. We propose a framework to extract guidance from, and specifically for, diffusion models.
  arXiv  Detail & Related papers  (2023-12-14T11:19:11Z)
• Diff-Instruct: A Universal Approach for Transferring Knowledge From Pre-trained Diffusion Models [77.83923746319498]
  We propose a framework called Diff-Instruct to instruct the training of arbitrary generative models. We show that Diff-Instruct results in state-of-the-art single-step diffusion-based models. Experiments on refining GAN models show that the Diff-Instruct can consistently improve the pre-trained generators of GAN models.
  arXiv  Detail & Related papers  (2023-05-29T04:22:57Z)
• Structural Pruning for Diffusion Models [65.02607075556742]
  We present Diff-Pruning, an efficient compression method tailored for learning lightweight diffusion models from pre-existing ones. Our empirical assessment, undertaken across several datasets highlights two primary benefits of our proposed method.
  arXiv  Detail & Related papers  (2023-05-18T12:38:21Z)
• Unifying Diffusion Models' Latent Space, with Applications to CycleDiffusion and Guidance [95.12230117950232]
  We show that a common latent space emerges from two diffusion models trained independently on related domains. Applying CycleDiffusion to text-to-image diffusion models, we show that large-scale text-to-image diffusion models can be used as zero-shot image-to-image editors.
  arXiv  Detail & Related papers  (2022-10-11T15:53:52Z)
• Non-Uniform Diffusion Models [0.8602553195689513]
  We show that non-uniform diffusion leads to multi-scale diffusion models which have similar structure to this of multi-scale normalizing flows. We experimentally find that in the same or less training time, the multi-scale diffusion model achieves better FID score than the standard uniform diffusion model. We also show that non-uniform diffusion leads to a novel estimator for the conditional score function which achieves on par performance with the state-of-the-art conditional denoising estimator.
  arXiv  Detail & Related papers  (2022-07-20T09:59:28Z)
• How Much is Enough? A Study on Diffusion Times in Score-based Generative Models [76.76860707897413]
  Current best practice advocates for a large T to ensure that the forward dynamics brings the diffusion sufficiently close to a known and simple noise distribution. We show how an auxiliary model can be used to bridge the gap between the ideal and the simulated forward dynamics, followed by a standard reverse diffusion process.
  arXiv  Detail & Related papers  (2022-06-10T15:09:46Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.