Metraipolis › archive › dc_2139cc2cb129

Learning Clean Generative Processes from Noisy Measurements — arXiv:2610.42036

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.42036
License (as archived): CC0-1.0
Archived: 2026-11-03T07:13:30.513Z
SHA-256 of the archived text: bd10b2acda03e5366399494d1ccc0c21a672e7ad57a019b17891eaf8744d13f6
Stable URL of this page: https://metraipolis.com/documento/dc_2139cc2cb129

How to cite

Learning Clean Generative Processes from Noisy Measurements — arXiv:2610.42036 — arXiv (en), https://arxiv.org/abs/2610.42036 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2139cc2cb129 (Metraipolis archive dc_2139cc2cb129, sha256 bd10b2acda03e5366399494d1ccc0c21a672e7ad57a019b17891eaf8744d13f6)

BibTeX

@misc{dc2139cc2cb129,
  title = {Learning Clean Generative Processes from Noisy Measurements — arXiv:2610.42036},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.42036},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2139cc2cb129; sha256 bd10b2acda03e5366399494d1ccc0c21a672e7ad57a019b17891eaf8744d13f6},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.42036. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

We study how to learn a generative model of clean images using only noisy observations. Diffusion processes and their generalizations to non-Gaussian noise require posterior predictions at every noise level. We first show that it is possible to learn these predictions from independent noisy views using Noise2Noise regression, but only at or above the noise level of the training data. To learn the predictions below that level, needed to generate clean images, we impose consistency: a prediction should equal its expectation after continuing the sampling process. We derive conditions under which consistency uniquely determines the continuation to the clean image distribution and verify them for Poisson, binomial, Gamma, and negative-binomial measurements. Experiments on CIFAR-10, photon-counting microscopy, and synthetic aperture radar show that the required posterior predictions can be learned from noisy views alone and that the consistency relation repairs the continuation beyond the noise level of the data, which severely deteriorates without consistency.

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); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.42036 — 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 2026-11-03T07:09:31.891Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).