arXiv cs.AI· Zixin Jessie Chen, Zhuo Chen, Archer Wang, Jeff Gore, William T. Freeman, Congyue Deng, Marin Solja\v{c}i\'c·· 12 小时前AI 评分38
SKILD：统一生成与连续超分辨率的尺度不变扩散模型
Everything at Every Scale: Scale-Invariant Diffusion with Continuous Super-Resolution
AI 导读
研究者提出 SKILD，一个尺度不变的 K 空间图像学习扩散模型，在单一无条件框架内统一图像生成与连续超分辨率。该模型将尺度作为扩散动力学的显式坐标，仅改变起始时间步即可完成两项任务，无需任务专用架构、条件分支、无分类器引导或按缩放因子重训练。
正文 · 原文
Abstract:Creating images from noise is image generation; reconstructing fine details from coarse inputs is super-resolution. Despite their practical differences, both can be understood as reversing information loss across scales. We introduce $\textbf{SKILD}$, a $\textbf{S}$cale-invariant $\textbf{K}$-Space $\textbf{I}$mage $\textbf{L}$earning $\textbf{D}$iffusion model that unifies generation and continuous super-resolution within a single unconditional framework. Both natural images and critical physical systems exhibit scale invariance, and we leverage it to design a forward process that attenuates image content from fine to coarse scales while injecting spectrum-matched Gaussian noise, making scale an explicit coordinate of the diffusion dynamics. The same trained reverse process performs generation and continuous super-resolution by varying only the starting timestep: $\textit{no task-specific architecture, no conditioning branch, no classifier-free guidance, no retraining per scale factor}$. Empirically, SKILD reaches FID $2.61$ and Inception Score $9.65$ on unconditional CIFAR-10, performs $2\times$--$8\times$ super-resolution on ImageNet from a single unconditional checkpoint while outperforming conditional models across perceptual metrics, and reconstructs critical Ising models whose connected four-point correlations closely track the ground truth.
| Comments: | 39 pages, 18 figures, NeurIPS 2026 |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Statistical Mechanics (cond-mat.stat-mech); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2605.26032 [cs.CV] |
| (or arXiv:2605.26032v2 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2605.26032 arXiv-issued DOI via DataCite |
Submission history
From: Zixin Chen [view email]
[v1]
Mon, 25 May 2026 17:01:21 UTC (22,264 KB)
[v2]
Thu, 5 Nov 2026 18:49:37 UTC (22,422 KB)
来源：arXiv cs.AI · arxiv.org