Metraipolis › archive › dc_4dc25d016cb2
WEIRDO: WEak resIdual Regularized DOob's h-transform diffusion alignment — arXiv:2609.39531
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.39531
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T13:47:39.449Z
- SHA-256 of the archived text
667cd0b3123410e5afa1ff9e10efd9833f941fac0cbdd98d4f5799314f5a4375- Stable URL of this page
https://metraipolis.com/documento/dc_4dc25d016cb2
How to cite
WEIRDO: WEak resIdual Regularized DOob's h-transform diffusion alignment — arXiv:2609.39531 — arXiv (en), https://arxiv.org/abs/2609.39531 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4dc25d016cb2 (Metraipolis archive dc_4dc25d016cb2, sha256 667cd0b3123410e5afa1ff9e10efd9833f941fac0cbdd98d4f5799314f5a4375)
BibTeX
@misc{dc4dc25d016cb2,
title = {WEIRDO: WEak resIdual Regularized DOob's h-transform diffusion alignment — arXiv:2609.39531},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.39531},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4dc25d016cb2; sha256 667cd0b3123410e5afa1ff9e10efd9833f941fac0cbdd98d4f5799314f5a4375},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.39531. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We study the problem of estimating the guidance that steers the distribution learned by a diffusion generative model toward a tilted target $q_0 \propto w\,p_0$ at inference time. Relying on the stochastic optimal control approach, we observe that the exact drift correction is the gradient of the logarithm of Doob's $h$-function, and we study the problem of estimating it from a sample. In the present paper, we assume that the score of the pretrained model is available, that the tilting weight is bounded and positive, and that the reference distribution has a bounded support, no smoothness of the weight is required. Introducing a penalized least-squares risk in which the penalty is the residual of the space-time harmonicity equation satisfied by the $h$-function, measured in a dual Sobolev norm, we derive high-probability bounds on the squared error of the resulting guidance estimate. Since the penalty vanishes at the target, the estimator is free of regularization bias, and in favourable scenarios its rate of convergence is faster than the minimax rate of estimating first-order derivatives of a smooth regression function. Assuming that $w$ is bounded and positive with $\mathbb{E}_{p_0}[w^{-\mathrm{s}}] < \infty$ for some $\mathrm{s} \in (0,\infty]$, and that the reference data are compactly supported, we prove that the guidance is estimable in squared $L^2$ at rate $\varepsilon_n^{\mathrm{s}/(\mathrm{s}+4)}$, where $\varepsilon_n = n^{-2(\beta-1)/(2(\beta-1)+d)}.$ We also transfer the obtained bounds to the total variation distance between the marginals of the estimated and the exactly guided samplers, and illustrate the performance of the suggested approach with numerical experiments.
Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.LG).
Categorie arXiv: Statistics Theory (math.ST) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2609.39531 — 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 2026-10-07T13:18:10.653Z dalla pagina https://arxiv.org/catchup/cs.LG/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).