Metraipolis › archive › dc_e4e012a3171f

Spec2Vision: Contract-Guided Delivery of AI-Generated Computer Vision Pipelines — arXiv:2608.26400

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.26400
License (as archived)
CC0-1.0
Archived
2026-10-17T16:16:06.841Z
SHA-256 of the archived text
a0a3739e6d201155665b4714bdcdb2b4f2e6137e20e0a4bd96daa31d29337788
Stable URL of this page
https://metraipolis.com/documento/dc_e4e012a3171f

How to cite

Spec2Vision: Contract-Guided Delivery of AI-Generated Computer Vision Pipelines — arXiv:2608.26400 — arXiv (en), https://arxiv.org/abs/2608.26400 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e4e012a3171f (Metraipolis archive dc_e4e012a3171f, sha256 a0a3739e6d201155665b4714bdcdb2b4f2e6137e20e0a4bd96daa31d29337788)
BibTeX
@misc{dce4e012a3171f,
  title = {Spec2Vision: Contract-Guided Delivery of AI-Generated Computer Vision Pipelines — arXiv:2608.26400},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.26400},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e4e012a3171f; sha256 a0a3739e6d201155665b4714bdcdb2b4f2e6137e20e0a4bd96daa31d29337788},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.26400. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Generated computer-vision code can be runnable without satisfying the task contract enforced by a downstream evaluator. We study that gap with Spec2Vision, an experimental framework for producing and evaluating specification-grounded CV pipeline bundles through a staged runtime that keeps the task contract explicit across synthesis, screening, testing, and bounded repair. The benchmark evaluates 17 CV tasks, 10 executable conditions, and 5 repeats per task-condition cell, for 850 primary runs. In the primary 850-run evaluation, Spec2Vision reaches 81/85 evaluator-test passes; removing structural repair drops to 55/85, compatibility scaffolding to 58/85, and generator preflight to 39/85. The executable single-agent baselines expose progressively richer task specifications to the model, culminating in direct source-spec exposure, yet remain much weaker overall, from 17/85 for lightweight task grounding to 35/85 evaluator-test passes. The lightweight baseline nevertheless remains core-runnable in 85/85 runs but reaches only 17/85 evaluator-test passes and 6/85 strict-delivery successes, showing that runnability is not equivalent to delivery. Across this benchmark, the strongest evidence comes from keeping the task contract explicit across staged generation, checking, and repair. Artifacts are provided to support audit of run bundles, model-visible inputs, and derived tables.

Tema degli esploratori: calcolo e infrastrutture (retro-catalogo arXiv: annunci del 2026-08-28 in cs.SE).
Categorie arXiv: Software Engineering (cs.SE) ; Programming Languages (cs.PL).
Pagina del lavoro: https://arxiv.org/abs/2608.26400 — 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 2026-10-17T15:36:41.154Z dalla pagina https://arxiv.org/catchup/cs.SE/2026-08-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).