English

Sig-DEG for Distillation: Making Diffusion Models Faster and Lighter

Machine Learning 2025-08-26 v1 Probability Machine Learning

Abstract

Diffusion models have achieved state-of-the-art results in generative modelling but remain computationally intensive at inference time, often requiring thousands of discretization steps. To this end, we propose Sig-DEG (Signature-based Differential Equation Generator), a novel generator for distilling pre-trained diffusion models, which can universally approximate the backward diffusion process at a coarse temporal resolution. Inspired by high-order approximations of stochastic differential equations (SDEs), Sig-DEG leverages partial signatures to efficiently summarize Brownian motion over sub-intervals and adopts a recurrent structure to enable accurate global approximation of the SDE solution. Distillation is formulated as a supervised learning task, where Sig-DEG is trained to match the outputs of a fine-resolution diffusion model on a coarse time grid. During inference, Sig-DEG enables fast generation, as the partial signature terms can be simulated exactly without requiring fine-grained Brownian paths. Experiments demonstrate that Sig-DEG achieves competitive generation quality while reducing the number of inference steps by an order of magnitude. Our results highlight the effectiveness of signature-based approximations for efficient generative modeling.

Keywords

Cite

@article{arxiv.2508.16939,
  title  = {Sig-DEG for Distillation: Making Diffusion Models Faster and Lighter},
  author = {Lei Jiang and Wen Ge and Niels Cariou-Kotlarek and Mingxuan Yi and Po-Yu Chen and Lingyi Yang and Francois Buet-Golfouse and Gaurav Mittal and Hao Ni},
  journal= {arXiv preprint arXiv:2508.16939},
  year   = {2025}
}
R2 v1 2026-07-01T05:02:43.511Z