Metraipolis › archive › dc_3aafe8087dc0
Sample-Adaptive Latent Rewards for Uncertainty-Guided Diffusion Post-Training — arXiv:2608.06125
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.06125
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-31T10:37:38.421Z
- SHA-256 of the archived text
199b62df9fc3be40ba1f78402eacef56e7d2384ad61feb43c9d8f89fc8033add- Stable URL of this page
https://metraipolis.com/documento/dc_3aafe8087dc0
How to cite
Sample-Adaptive Latent Rewards for Uncertainty-Guided Diffusion Post-Training — arXiv:2608.06125 — arXiv (en), https://arxiv.org/abs/2608.06125 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_3aafe8087dc0 (Metraipolis archive dc_3aafe8087dc0, sha256 199b62df9fc3be40ba1f78402eacef56e7d2384ad61feb43c9d8f89fc8033add)
BibTeX
@misc{dc3aafe8087dc0,
title = {Sample-Adaptive Latent Rewards for Uncertainty-Guided Diffusion Post-Training — arXiv:2608.06125},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.06125},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_3aafe8087dc0; sha256 199b62df9fc3be40ba1f78402eacef56e7d2384ad61feb43c9d8f89fc8033add},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.06125. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Latent reward models can supervise visual diffusion models without decoding intermediate states into pixel space. This makes alignment with human preferences more efficient. However, existing latent reward models output only scalar scores. They do not estimate the uncertainty of each prediction. The generator therefore cannot determine which feedback is reliable. This can drive optimization in the wrong direction and lead to reward hacking. We propose \textsc{SURE}, a unified latent-space framework for image and video diffusion models. It learns reward distributions and directly uses their reliability to guide dense post-training. First, we propose sample-adaptive latent reward model (\textsc{SURE-LRM}). It predicts a Gaussian utility for each noisy latent. Its mean predicts the reward score. Its variance reflect the uncertainty of prediction without human annotation. The learned distribution then guides post-training through uncertainty-guided reward feedback learning (\textsc{SURE-REFL}). This method provides uncertainty-guided dense feedback along the denoising trajectory. At selected transitions, \textsc{SURE-REFL} queries the frozen \textsc{SURE-LRM}. It converts detached variance into reliability weights for samples at the same transition. Each weighted reward is backpropagated only through its local transition. The entire process remains in latent space and requires neither pixel-space decoding nor the full denoising graph. Experiments show that \textsc{SURE-LRM} improves preference prediction over strong baselines. \textsc{SURE-REFL} achieves the sota performance among various metrics and further improves optimization stability. It also achieves the highest VBench quality, semantic, and total scores among the evaluated methods.
Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-08-07 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2608.06125 — 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 2026-10-31T10:23:31.053Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-08-07?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).