Metraipolis › archive › dc_d63bae5a6dbe

PEEK: One-Step-Look-Ahead Exposure Bias Correction for Diffusion Sampling — arXiv:2611.05407

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.05407
License (as archived)
CC0-1.0
Archived
2026-11-05T18:17:18.978Z
SHA-256 of the archived text
7761243f67c27cc9ab10b110595f2adaed071271c8c9d986b3c5c226708f8d7c
Stable URL of this page
https://metraipolis.com/documento/dc_d63bae5a6dbe

How to cite

PEEK: One-Step-Look-Ahead Exposure Bias Correction for Diffusion Sampling — arXiv:2611.05407 — arXiv (en), https://arxiv.org/abs/2611.05407 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d63bae5a6dbe (Metraipolis archive dc_d63bae5a6dbe, sha256 7761243f67c27cc9ab10b110595f2adaed071271c8c9d986b3c5c226708f8d7c)
BibTeX
@misc{dcd63bae5a6dbe,
  title = {PEEK: One-Step-Look-Ahead Exposure Bias Correction for Diffusion Sampling — arXiv:2611.05407},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.05407},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d63bae5a6dbe; sha256 7761243f67c27cc9ab10b110595f2adaed071271c8c9d986b3c5c226708f8d7c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.05407. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Diffusion image generators are successful due to accurate denoising networks and strong numerical samplers. Yet the two components are not fully matched. During sampling, the denoiser predicts for the next step based on flawed input that differ much from the real data. This train-test gap, or exposure bias, caps image generation quality (i.e., blurry, lacking in details or awkward in shape). To mitigate the problem, we propose PEEK, a novel training-free deterministic sampler plugin. At each eligible denoising step, PEEK first performs a provisional one-step update, evaluates the denoiser once at the look-ahead state, and interpolates it with the one-step-later prediction to form the next noisy state. By taking the additional step, we obtain a generally more-image-like prediction from the higher-signal noisy state, mitigating the exposure bias. To optimize the interpolation positions, PEEK searches for an interpolation schedule search made efficient by the prefix and baseline noise-schedule-rebasing caches. On 8 pixel-space and latent-space diffusion models with unconditional, label-guided, and text-guided settings across various output resolutions, PEEK consistently improves over NFE-matched baselines, with FID improvements up to 49.3\% at minimal extra computation cost (e.g., the one-time-per-model calibration takes at most 2.2 minutes across our experiments.). Project website is at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.05407 — 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 2026-11-05T18:12:49.024Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).