analiza aiqbench
This research proposes an energy-adaptive noise scheduling mechanism for spectral diffusion models, aligning the forward diffusion process with the inherent spectral-energy distribution of natural images. By integrating global spectral whitening with energy-conditioned noise allocation, the method achieves Gaussian transitions compatible with standard DDPM and DDIM architectures without requiring architectural modifications. Experimental results on CIFAR-10 show a significant reduction in Fréchet Inception Distance, demonstrating improved generation quality through better modeling of frequency-domain statistics.
Energy-Conditioned Noise Schedule and Whitening for Spectral Diffusion
arXiv:2610.07206v1 Announce Type: new Abstract: This paper introduces an energy-adaptive noise scheduling and whitening strategy for transform-domain diffusion models. Existing spectral diffusion methods account for the non-uniform statistics of transform coefficients through coefficient scaling, normalization, or frequency prioritization, while the forward diffusion noise schedule remains largely independent of the underlying spectral-energy distribution. We investigate whether the temporal evolution of the forward diffusion process should also follow the spectral organization of natural images. The proposed formulation combines global spectral whitening with energy-conditioned noise allocation that jointly modulates the injected noise according to the energy of individual transform coefficients and an image-dependent energy path over diffusion time. The resulting forward process preserves Gaussian transitions with closed-form marginals and remains compatible with standard DDPM and DDIM procedures without modifying the diffusion architecture. Experiments on CIFAR-10 demonstrate the contribution of the proposed energy-conditioned noise schedule and spectral whitening, reducing Fr\'echet Inception Distance from 142.48 for a compact DCTdiff U-Net variant to 100.45.
Abstract:This paper introduces an energy-adaptive noise scheduling and whitening strategy for transform-domain diffusion models. Existing spectral diffusion methods account for the non-uniform statistics of transform coefficients through coefficient scaling, normalization, or frequency prioritization, while the forward diffusion noise schedule remains largely independent of the underlying spectral-energy distribution. We investigate whether the temporal evolution of the forward diffusion process should also follow the spectral organization of natural images. The proposed formulation combines global spectral whitening with energy-conditioned noise allocation that jointly modulates the injected noise according to the energy of individual transform coefficients and an image-dependent energy path over diffusion time. The resulting forward process preserves Gaussian transitions with closed-form marginals and remains compatible with standard DDPM and DDIM procedures without modifying the diffusion architecture. Experiments on CIFAR-10 demonstrate the contribution of the proposed energy-conditioned noise schedule and spectral whitening, reducing Fréchet Inception Distance from 142.48 for a compact DCTdiff U-Net variant to 100.45.
Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV); Emerging Technologies (cs.ET)
arXiv:2610.07206 [cs.AI]
(or arXiv:2610.07206v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2610.07206 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
https://doi.org/10.2139/ssrn.7557215 Focus to learn more DOI(s) linking to related resources
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 266 cuvinte
Subiecte conexe
Graph Residual Conjugate Diffusion: SNR-Equalized Heat Flow for Graph Signals
arXiv · 1 oct. 2026
Energy Time-Series Imputation with Differentially Private Diffusion Models via Clipping-Aware Objective Conditioning
arXiv · 2 oct. 2026
Multiscale Hypersonic Boundary Layer Reconstruction via Spectral Binning and Subdomain-wise Conditional Diffusion
arXiv · 16 iun. 2026
Turn Your Latest Observations Into Timely Weather Decisions With NVIDIA Earth-2
NVIDIA · 21 sept. 2026