Metraipolis › archive › dc_bc2ffcadb1ff
Do Vision Models Learn Physical Constraints or Rendering Shortcuts? A Counterfactual Benchmark for Grounded Physical Consistency — arXiv:2610.09205
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.09205
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T11:10:40.563Z
- SHA-256 of the archived text
65ab8541edc855ae2211ea3fa5c67dee0fd6dec7b50f808665adc21e77f0c05f- Stable URL of this page
https://metraipolis.com/documento/dc_bc2ffcadb1ff
How to cite
Do Vision Models Learn Physical Constraints or Rendering Shortcuts? A Counterfactual Benchmark for Grounded Physical Consistency — arXiv:2610.09205 — arXiv (en), https://arxiv.org/abs/2610.09205 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bc2ffcadb1ff (Metraipolis archive dc_bc2ffcadb1ff, sha256 65ab8541edc855ae2211ea3fa5c67dee0fd6dec7b50f808665adc21e77f0c05f)
BibTeX
@misc{dcbc2ffcadb1ff,
title = {Do Vision Models Learn Physical Constraints or Rendering Shortcuts? A Counterfactual Benchmark for Grounded Physical Consistency — arXiv:2610.09205},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.09205},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bc2ffcadb1ff; sha256 65ab8541edc855ae2211ea3fa5c67dee0fd6dec7b50f808665adc21e77f0c05f},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.09205. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Modern image editing models can satisfy a text instruction while breaking the physics of the edited scene. A new object may cast no shadow, a mirror may fail to reflect visible geometry, or an object may float above a surface that should support it. We study physical plausibility diagnosis, detecting whether an edited image violates scene physics, naming the violation type, localizing the affected region, and explaining the failure in language. We introduce a counterfactual benchmark whose controlled synthetic component uses Mitsuba~3 to generate 5,500 images from 500 scene families. Each family contains one clean image and ten matched violations involving shadows, reflection, support, surface response, and occlusion. The renderer pipeline provides category labels, affected-region masks and boxes, scene metadata, and explanation targets. We use LLaVA-1.5-7B, Qwen2.5-VL-7B, and InternVL3.5-8B as diagnostic baselines rather than proposed methods. On a 1,650-image synthetic test set, the adapted baselines reach 64.0--67.8\% category macro-F1 on standard held-out scenes. For LLaVA-1.5-7B, category macro-F1 falls from 64.0\% on the standard split to 40.8\% under intervention shift. This gap shows that high in-distribution accuracy partly reflects cues tied to rendering and counterfactual construction. 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/2610.09205 — 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 2026-10-08T10:14:06.057Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).