Set-Coupled Guidance: Set-Level Coordination in Diffusion-Based Dataset Distillation

Ziang Gan, Qi Zhu, Libao Zhang
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:32896-32916, 2026.

Abstract

Diffusion models serve as generative priors for dataset distillation, yet existing pipelines rely on per-sample update rules that evolve each synthetic image independently, limiting their ability to optimize collective set-level objectives. We propose Set-Coupled Guidance (SCG), a plug-and-play auxiliary controller that shifts from per-image to group (IPC-at-once) sampling by injecting set-symmetric feedback at each diffusion step. SCG combines spectral set-point regulation, which aligns set-level statistics to real data via empirical characteristic function matching, with cooperative kernel coupling that stabilizes joint trajectories under noisy feedback. All computations operate on lightweight descriptors extracted from predicted clean latents, adding low overhead to the base method. We provide theoretical analysis including Lyapunov descent and input-to-state stability for distributional tracking. Experiments on ImageNette, ImageWoof, ImageNet-100 and ImageNet-1K show consistent accuracy gains across multiple diffusion-based baselines; code is available at https://github.com/tade1s/SCG.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-gan26d, title = {Set-Coupled Guidance: Set-Level Coordination in Diffusion-Based Dataset Distillation}, author = {Gan, Ziang and Zhu, Qi and Zhang, Libao}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {32896--32916}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/gan26d/gan26d.pdf}, url = {https://proceedings.mlr.press/v306/gan26d.html}, abstract = {Diffusion models serve as generative priors for dataset distillation, yet existing pipelines rely on per-sample update rules that evolve each synthetic image independently, limiting their ability to optimize collective set-level objectives. We propose Set-Coupled Guidance (SCG), a plug-and-play auxiliary controller that shifts from per-image to group (IPC-at-once) sampling by injecting set-symmetric feedback at each diffusion step. SCG combines spectral set-point regulation, which aligns set-level statistics to real data via empirical characteristic function matching, with cooperative kernel coupling that stabilizes joint trajectories under noisy feedback. All computations operate on lightweight descriptors extracted from predicted clean latents, adding low overhead to the base method. We provide theoretical analysis including Lyapunov descent and input-to-state stability for distributional tracking. Experiments on ImageNette, ImageWoof, ImageNet-100 and ImageNet-1K show consistent accuracy gains across multiple diffusion-based baselines; code is available at https://github.com/tade1s/SCG.} }
Endnote
%0 Conference Paper %T Set-Coupled Guidance: Set-Level Coordination in Diffusion-Based Dataset Distillation %A Ziang Gan %A Qi Zhu %A Libao Zhang %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-gan26d %I PMLR %P 32896--32916 %U https://proceedings.mlr.press/v306/gan26d.html %V 306 %X Diffusion models serve as generative priors for dataset distillation, yet existing pipelines rely on per-sample update rules that evolve each synthetic image independently, limiting their ability to optimize collective set-level objectives. We propose Set-Coupled Guidance (SCG), a plug-and-play auxiliary controller that shifts from per-image to group (IPC-at-once) sampling by injecting set-symmetric feedback at each diffusion step. SCG combines spectral set-point regulation, which aligns set-level statistics to real data via empirical characteristic function matching, with cooperative kernel coupling that stabilizes joint trajectories under noisy feedback. All computations operate on lightweight descriptors extracted from predicted clean latents, adding low overhead to the base method. We provide theoretical analysis including Lyapunov descent and input-to-state stability for distributional tracking. Experiments on ImageNette, ImageWoof, ImageNet-100 and ImageNet-1K show consistent accuracy gains across multiple diffusion-based baselines; code is available at https://github.com/tade1s/SCG.
APA
Gan, Z., Zhu, Q. & Zhang, L.. (2026). Set-Coupled Guidance: Set-Level Coordination in Diffusion-Based Dataset Distillation. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:32896-32916 Available from https://proceedings.mlr.press/v306/gan26d.html.

Related Material