Metraipolis › archive › dc_405a031fb7ae
Controlling Polar Exposure to Delay Memorization in Diffusion Models — arXiv:2610.02780
- Source
- arXiv
- Original
- https://arxiv.org/abs/2610.02780
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-05T23:09:52.539Z
- SHA-256 of the archived text
2d5b1438f4bf191f4613130dd52d74228bb410dd70ad90f9a6a1bb972402df58- Stable URL of this page
https://metraipolis.com/documento/dc_405a031fb7ae
How to cite
Controlling Polar Exposure to Delay Memorization in Diffusion Models — arXiv:2610.02780 — arXiv, https://arxiv.org/abs/2610.02780 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_405a031fb7ae (Metraipolis archive dc_405a031fb7ae, sha256 2d5b1438f4bf191f4613130dd52d74228bb410dd70ad90f9a6a1bb972402df58)
BibTeX
@misc{dc405a031fb7ae,
title = {Controlling Polar Exposure to Delay Memorization in Diffusion Models — arXiv:2610.02780},
howpublished = {arXiv},
url = {https://arxiv.org/abs/2610.02780},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_405a031fb7ae; sha256 2d5b1438f4bf191f4613130dd52d74228bb410dd70ad90f9a6a1bb972402df58},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.02780. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Diffusion models can reach useful sample quality before copying training examples, but fast optimization can compress this generalization window by accelerating sample-specific fitting. We investigate this effect through update geometry and propose Quality-Gated De-whitening (QGD), a controller that retains a fast polar-update prefix and progressively restores fixed-gain momentum. Our random-feature analysis separates covariance-controlled, curvature-equalized and amplitude-controlled memorization clocks. Under aligned spectral assumptions, it establishes a finite-exposure condition under which a fixed-gain tail recovers a delay proportional to dataset size. QGD implements this principle with a confirmed quality gate, a bounded decay envelope and causal copy feedback. Immediate switching is the conservative limit; gradual control balances delayed copying against continued quality improvement. We pair QGD with Copy-Budgeted Selection (CBS), which applies simultaneous binomial calibration to a frozen checkpoint family, followed by a fresh evaluation of the released checkpoint. On 2,000-image CIFAR-10 subsets, QGD preserves the polar baseline's quality-arrival time while expanding its useful interval by 8.32x and reducing common-checkpoint copying by 75.9%. With identical calibration and independent quality evaluation, QGD achieves FID 75.56 versus 79.37 for SGD with the same selector. Exposure-matched controls, independent detector audits and transfer to flow matching and dance generation support adaptive exposure control as a practical way to improve the quality-copying tradeoff. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.02780 — 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 2026-10-05T23:01:38.634Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).