Metraipolis › archive › dc_b668d5129a29

Leveraging Inference-Time Compute for Diffusion Models via Global Scheduling of Denoising Trajectories — arXiv:2609.22867

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.22867
License (as archived)
CC0-1.0
Archived
2026-10-09T07:06:53.536Z
SHA-256 of the archived text
15284f90d14564985550458e312b128e8dbd32e99bc2a2a163b980f869679f14
Stable URL of this page
https://metraipolis.com/documento/dc_b668d5129a29

How to cite

Leveraging Inference-Time Compute for Diffusion Models via Global Scheduling of Denoising Trajectories — arXiv:2609.22867 — arXiv (en), https://arxiv.org/abs/2609.22867 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b668d5129a29 (Metraipolis archive dc_b668d5129a29, sha256 15284f90d14564985550458e312b128e8dbd32e99bc2a2a163b980f869679f14)
BibTeX
@misc{dcb668d5129a29,
  title = {Leveraging Inference-Time Compute for Diffusion Models via Global Scheduling of Denoising Trajectories — arXiv:2609.22867},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.22867},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b668d5129a29; sha256 15284f90d14564985550458e312b128e8dbd32e99bc2a2a163b980f869679f14},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.22867. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Diffusion models generate a sample by traversing a denoising trajectory, a sequence of stochastic noise-reduction steps that transforms pure noise into a draw from a target distribution. At deployment time, additional computation can improve sample quality without retraining: at each step, the sampler draws several candidate noise samples, scores the resulting predictions with a quality criterion called the verifier, and retains the best candidate at the cost of one network evaluation per candidate. This raises a resource allocation question: given a fixed budget of function evaluations, how should search effort be distributed across the steps of the denoising trajectory? We formulate this as a computational budget allocation problem. First, we show that, to leading order in the step size, the expected gain from evaluating $K$ candidates at a step factorizes into an endogenous, step-specific sensitivity parameter times a universal sample-size factor equal to the expected best of $K$ standard-normal draws. Second, for a fixed sensitivity profile, the optimal allocation solves a separable concave integer program with water-filling structure; at fixed total sensitivity, its advantage over uniform allocation increases with sensitivity dispersion in the majorization order. Third, we prove that when sensitivities vary across instances, no adaptive policy can avoid worst-case regret that grows linearly in the trajectory length, which motivates a design that anchors the allocation offline and adapts online only to recover instance-specific slack. We extend the analysis from independent random search to a broader family of local search operators, and instantiate it as an implementable algorithm. Experiments on three families of diffusion samplers show that the proposed allocation attains the quality of the uniform benchmark with 20 to 50 percent fewer function evaluations.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-22 in cs.LG).
Categorie arXiv: Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.22867 — 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 2026-10-09T06:58:09.237Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-09-22?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).