Metraipolis › archive › dc_120c87f841ba

Learning to Look: A Foundation for Visual Understanding — arXiv:2611.02053

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.02053
License (as archived)
CC0-1.0
Archived
2026-11-03T11:41:49.891Z
SHA-256 of the archived text
f957449d2e74df03d124845133db906c5e40f3a9bb1ae22839908f9fda2c0086
Stable URL of this page
https://metraipolis.com/documento/dc_120c87f841ba

How to cite

Learning to Look: A Foundation for Visual Understanding — arXiv:2611.02053 — arXiv (en), https://arxiv.org/abs/2611.02053 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_120c87f841ba (Metraipolis archive dc_120c87f841ba, sha256 f957449d2e74df03d124845133db906c5e40f3a9bb1ae22839908f9fda2c0086)
BibTeX
@misc{dc120c87f841ba,
  title = {Learning to Look: A Foundation for Visual Understanding — arXiv:2611.02053},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.02053},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_120c87f841ba; sha256 f957449d2e74df03d124845133db906c5e40f3a9bb1ae22839908f9fda2c0086},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.02053. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Vision-language models (VLMs) suffer from visual hallucination, such as saying an object that is not present in an image. We ask a more fundamental question: whether a VLM is even aware of where it is looking before understanding the semantic meaning in the image. We formulate this question as a LOOK task: identifying which parts of its visual input were available to the model. We construct the task in a self-supervised way without manually labeling the available region to the model. Our experiments show that existing VLMs fail to localize the visible region. We further show that fine-tuning VLMs on the proposed LOOK task improves a model's awareness of visual input.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2611.02053 — 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 2026-11-03T11:06:37.540Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).