Metraipolis › archive › dc_08efbb3671bf

Universal Local Error and Realized Amplification for the First-Order EDM Predictor — arXiv:2610.10190

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.10190
License (as archived): CC0-1.0
Archived: 2026-10-08T12:37:24.303Z
SHA-256 of the archived text: 0d5a21e1be0ca0f7ccd9b5ccca18c4661d523552ba0c04148f61abd242fad2b8
Stable URL of this page: https://metraipolis.com/documento/dc_08efbb3671bf

How to cite

Universal Local Error and Realized Amplification for the First-Order EDM Predictor — arXiv:2610.10190 — arXiv (en), https://arxiv.org/abs/2610.10190 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_08efbb3671bf (Metraipolis archive dc_08efbb3671bf, sha256 0d5a21e1be0ca0f7ccd9b5ccca18c4661d523552ba0c04148f61abd242fad2b8)

BibTeX

@misc{dc08efbb3671bf,
  title = {Universal Local Error and Realized Amplification for the First-Order EDM Predictor — arXiv:2610.10190},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.10190},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_08efbb3671bf; sha256 0d5a21e1be0ca0f7ccd9b5ccca18c4661d523552ba0c04148f61abd242fad2b8},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.10190. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We analyze the first-order deterministic diffusion sampler of Karras et al. (2022), termed EDM, in 2-Wasserstein distance by separating two sources of error: local discretization error and its amplification by subsequent learned steps. We prove that local error admits a universal bound: for any data distribution with finite second moment, the one-step discretization error is quadratic in the step size, with an explicit constant that does not depend on the data distribution. Error propagation, in contrast, depends on the learned network. At high noise levels, we exploit the network parametrization of EDM to derive an explicit contraction criterion. At low noise levels, we measure propagation through the amplification realized on the distributions transported by the sampler; this realized amplification can be arbitrarily smaller than the worst-case Lipschitz constant. This analysis yields an $O(e^{\Lambda_K}/K)$ global discretization error for $K$ sampling steps, where $\Lambda_K$ is the low-noise log-amplification. Experiments on a one-dimensional Gaussian mixture show how measured amplification accounts for slower error decay on finite sampling grids. Diagnostics on a pretrained CIFAR-10 model illustrate related stability mechanisms without certifying the global assumptions.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Machine Learning (stat.ML) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.10190 — 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 2026-10-08T12:14:15.451Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).