Metraipolis › archive › dc_c003a7e695eb

SolveEdit: Benchmarking Visual Problem Solving in Generative Models — arXiv:2609.35504

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.35504
License (as archived)
CC0-1.0
Archived
2026-10-07T21:03:41.850Z
SHA-256 of the archived text
852927ff2046c72c4e1e6beb001fecffd3b94a83c800ab1b35be95cea7de86ff
Stable URL of this page
https://metraipolis.com/documento/dc_c003a7e695eb

How to cite

SolveEdit: Benchmarking Visual Problem Solving in Generative Models — arXiv:2609.35504 — arXiv (en), https://arxiv.org/abs/2609.35504 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c003a7e695eb (Metraipolis archive dc_c003a7e695eb, sha256 852927ff2046c72c4e1e6beb001fecffd3b94a83c800ab1b35be95cea7de86ff)
BibTeX
@misc{dcc003a7e695eb,
  title = {SolveEdit: Benchmarking Visual Problem Solving in Generative Models — arXiv:2609.35504},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.35504},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c003a7e695eb; sha256 852927ff2046c72c4e1e6beb001fecffd3b94a83c800ab1b35be95cea7de86ff},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.35504. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Machine intelligence is often evaluated through abstract reasoning problems, yet many real-world problems are visual, such as arranging objects, repairing layouts, or tracing routes. Solving these problems requires understanding a scene, inferring what must change to achieve a goal, and realizing that change without disturbing unrelated content. However, existing benchmarks mainly evaluate perception, generation, or explicitly specified transformations, leaving goal-driven visual problem solving underexplored. To bridge this gap, we introduce SolveEpIT, a benchmark for visual problem solving through scene transformation. Given an image and a goal, a model must infer a valid transformation from the request, the scene, or a visually expressed rule, then execute it while preserving unrelated content. SoLvEEDrr contains 2,728 cases. Atomic transition contracts specify required and protected conditions, enabling SoLvEScoRE to measure completion and unintended changes without a single reference output. The strongest evaluated model achieves only57.0% SolvEScore. We further introduce SolveEdiT-PLAN, a two-stage visual planner that instantiates the transition before generation. Under matched single-generation evaluation, it improves SoLvEScoRE by 9.1 points on average across three tested generators, including a gain from 57.0% to 71.6% for GPT-Image-2, without modifying the editor.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.35504 — 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 2026-10-07T20:29:03.377Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).