Metraipolis › archive › dc_1d76da54409b

Scalable Minimal-Change Learning for Controllable Image Editing — arXiv:2610.06021

Source
arXiv
Original
https://arxiv.org/abs/2610.06021
License (as archived)
CC0-1.0
Archived
2026-10-06T05:27:22.040Z
SHA-256 of the archived text
f89b1e1c3c458478cca776418c486e5615f8bf0e9c25de4b2b0e8efd71f8719c
Stable URL of this page
https://metraipolis.com/documento/dc_1d76da54409b

How to cite

Scalable Minimal-Change Learning for Controllable Image Editing — arXiv:2610.06021 — arXiv, https://arxiv.org/abs/2610.06021 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1d76da54409b (Metraipolis archive dc_1d76da54409b, sha256 f89b1e1c3c458478cca776418c486e5615f8bf0e9c25de4b2b0e8efd71f8719c)
BibTeX
@misc{dc1d76da54409b,
  title = {Scalable Minimal-Change Learning for Controllable Image Editing — arXiv:2610.06021},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.06021},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1d76da54409b; sha256 f89b1e1c3c458478cca776418c486e5615f8bf0e9c25de4b2b0e8efd71f8719c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.06021. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Image editing should change only the attributes specified by an instruction while preserving everything else, yet current methods often make unintended changes. We treat this minimal-change principle as an optimization objective for instruction-based editing. Latent L1 regularization is a poor proxy for output locality in modern nonlinear generators and often requires supervision unavailable at scale. We instead optimize edit outcomes with reinforcement learning. An agentic vision-language reward model audits each source image, instruction, and edited image for two failure types: unimplemented requested changes and unintended changes. A group-level rubric merges and verifies these issues to provide consistent rewards across candidate edits without per-instruction human annotations. On FLUX.1 Kontext-dev, ARRO raises average EditScore from 5.21 to 5.88 across MinEval, MagicBrush, AnyBench, and Emu-Edit. On 600 evaluation examples, it reduces off-target pixel change by 8.4% relative to the base editor. Reward and SFT controls, blinded human evaluations, and transfer to OmniGen2 provide complementary evidence. Code: this https URL

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/2610.06021 — 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 2026-10-06T05:24:06.104Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).