Metraipolis › archive › dc_6be86f170f43

Spectral-Guided Diffusion: Accelerating Inference via Static Spectral Layer Scheduling — arXiv:2609.29505

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.29505
License (as archived)
CC0-1.0
Archived
2026-10-08T05:26:02.783Z
SHA-256 of the archived text
62496c25065dd98498c84a10ea70d2c00b9925a2bf36ba207942f14a45df5543
Stable URL of this page
https://metraipolis.com/documento/dc_6be86f170f43

How to cite

Spectral-Guided Diffusion: Accelerating Inference via Static Spectral Layer Scheduling — arXiv:2609.29505 — arXiv (en), https://arxiv.org/abs/2609.29505 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6be86f170f43 (Metraipolis archive dc_6be86f170f43, sha256 62496c25065dd98498c84a10ea70d2c00b9925a2bf36ba207942f14a45df5543)
BibTeX
@misc{dc6be86f170f43,
  title = {Spectral-Guided Diffusion: Accelerating Inference via Static Spectral Layer Scheduling — arXiv:2609.29505},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.29505},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6be86f170f43; sha256 62496c25065dd98498c84a10ea70d2c00b9925a2bf36ba207942f14a45df5543},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.29505. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Diffusion inference repeatedly evaluates the same large network. We ask whether pretrained weights alone can identify residual branches that need not be recomputed throughout the trajectory. Our \textbf{Spectral Concentration Ratio (SCR)} measures leading-versus-tail singular-value energy. Combined with Frobenius magnitude, it yields an offline sensitivity proxy and a deterministic lifetime for each scheduled unit. A frozen unit reuses its cached residual-branch update while the current residual stream and all external conditioning continue to propagate. The method needs no router, calibration prompts, or input-dependent search. At matched layer-step budgets, SCR/Frobenius preserves quality better than random, depth, norm, stable-rank, and Frobenius--stable-rank schedules on LLaDA-8B, DiT-XL/2, U-ViT-L, and SDXL. Broader LLaDA tests cover retrieval, reasoning, code, summarization, and open-ended generation; matched-horizon controls retain the ranking down to ten denoising steps. The complete captured-graph system reaches $2.8\times$--$3.0\times$ wall-clock speedup over eager inference. This is a systems-level number: on LLaDA, padded graph execution already gives $2.7\times$, while eliminating inactive branch work raises it to $3.0\times$. The perturbation analysis motivates pre-norm attention and MLP components under explicit local assumptions; results on AdaLN, U-shaped, convolutional, and cross-attention blocks are empirical transfer, not certified guarantees.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-25 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.29505 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-08T04:50:39.434Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-25?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).