Metraipolis › archive › dc_0985dbefbfeb

Atomic Visual Entailment: Enhancing Zero-Shot Vision-Language Reasoning through Atomic Fact Decomposition and Learned Selection — arXiv:2610.05630

Source: arXiv
Original: https://arxiv.org/abs/2610.05630
License (as archived): CC0-1.0
Archived: 2026-10-06T05:28:23.196Z
SHA-256 of the archived text: 00d233a51276e1ea64ee17c566f4ed79d73a51ec1660e74c9d71d927a6768d27
Stable URL of this page: https://metraipolis.com/documento/dc_0985dbefbfeb

How to cite

Atomic Visual Entailment: Enhancing Zero-Shot Vision-Language Reasoning through Atomic Fact Decomposition and Learned Selection — arXiv:2610.05630 — arXiv, https://arxiv.org/abs/2610.05630 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0985dbefbfeb (Metraipolis archive dc_0985dbefbfeb, sha256 00d233a51276e1ea64ee17c566f4ed79d73a51ec1660e74c9d71d927a6768d27)

BibTeX

@misc{dc0985dbefbfeb,
  title = {Atomic Visual Entailment: Enhancing Zero-Shot Vision-Language Reasoning through Atomic Fact Decomposition and Learned Selection — arXiv:2610.05630},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.05630},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0985dbefbfeb; sha256 00d233a51276e1ea64ee17c566f4ed79d73a51ec1660e74c9d71d927a6768d27},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.05630. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Visual entailment (VE) asks whether an image supports, contradicts, or leaves undecided a textual hypothesis. Strong results come from fine-tuning large vision-language models on labelled data, while zero-shot and hybrid approaches remain far behind. A VE hypothesis often bundles several visual claims, yet existing zero-shot methods reason over it as a single unit. We propose Atomic Visual Entailment (AVE), which decomposes the hypothesis into atomic facts, produces candidate predictions from both the full hypothesis and its facts using frozen vision-language models, and predicts the final label with a lightweight classifier trained only on how those candidates behave. We find that decomposition helps only when the hypothesis context is preserved: judging facts in isolation is worse than not decomposing at all. Full-hypothesis and atomic prediction make complementary errors, and learning which to trust recovers far more of that complementarity than majority voting, reaching 0.803 test accuracy on SNLI-VE without fine-tuning any vision-language model. AVE also localises the visual evidence behind its prediction without region-level supervision. These results suggest that learning which candidate prediction to trust can close much of the gap to fine-tuned systems, offering a practical alternative where fine-tuning a vision-language model directly would need more labelled data or compute than is available.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.05630 — 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 2026-10-06T05:24:20.340Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).