Metraipolis › archive › dc_1790cbcdb5aa
MGPO: Manifold-Guided Diffusion Alignment for Task-Aware Dataset Distillation — arXiv:2610.05252
- Source
- arXiv
- Original
- https://arxiv.org/abs/2610.05252
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-06T05:26:45.802Z
- SHA-256 of the archived text
b965b3081e38893cdd4ae459393149e647eac2b62a7ee11793fdbb698f935676- Stable URL of this page
https://metraipolis.com/documento/dc_1790cbcdb5aa
How to cite
MGPO: Manifold-Guided Diffusion Alignment for Task-Aware Dataset Distillation — arXiv:2610.05252 — arXiv, https://arxiv.org/abs/2610.05252 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1790cbcdb5aa (Metraipolis archive dc_1790cbcdb5aa, sha256 b965b3081e38893cdd4ae459393149e647eac2b62a7ee11793fdbb698f935676)
BibTeX
@misc{dc1790cbcdb5aa,
title = {MGPO: Manifold-Guided Diffusion Alignment for Task-Aware Dataset Distillation — arXiv:2610.05252},
howpublished = {arXiv},
url = {https://arxiv.org/abs/2610.05252},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1790cbcdb5aa; sha256 b965b3081e38893cdd4ae459393149e647eac2b62a7ee11793fdbb698f935676},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.05252. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Diffusion-based dataset distillation (DD) suffers from a fundamental objective mismatch: likelihood-driven diffusion models prioritize density approximation over the discriminative decision boundaries required for downstream tasks. Beyond semantic mismatch, relying solely on density also leads to geometric coverage loss, where generated samples collapse into a few high-density modes and fail to cover the manifold's structural diversity. We propose Manifold-Guided Policy Optimization (MGPO), which reformulates DD as a multi-objective reinforcement learning problem and achieves Dual-Space Alignment via a pixel-space discriminative reward and a latent-space geometric reward guided by a class-wise Minimum Spanning Tree (MST). The discriminative reward enforces class separability, while the MST-based geometric reward encourages generated latents to cover a sparse geometric skeleton of each class, jointly addressing both failure modes. We further provide an idealized analysis that motivates the MST-based reward, including a Hausdorff approximation bound and a subsampling bound independent of the dataset size. The reward-modular design extends to structured tasks such as object detection and segmentation by substituting the frozen task reward model. Extensive experiments show MGPO consistently outperforms existing methods, including a +8.0% mIoU gain on segmentation under low-budget settings. 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). Pagina del lavoro: https://arxiv.org/abs/2610.05252 — 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 2026-10-06T05:24:05.987Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).