Metraipolis › archive › dc_62db2d48ce7a

Mitigating Saliency Collapse: Robust Saliency-Aware Long-Text Image-Text Alignment — arXiv:2610.31005

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.31005
License (as archived)
CC0-1.0
Archived
2026-10-25T20:53:11.084Z
SHA-256 of the archived text
e7e479c4e3d3766b32d5bd1e92ab9f5bfcf2413a82348fbbab909c58bf5df2be
Stable URL of this page
https://metraipolis.com/documento/dc_62db2d48ce7a

How to cite

Mitigating Saliency Collapse: Robust Saliency-Aware Long-Text Image-Text Alignment — arXiv:2610.31005 — arXiv (en), https://arxiv.org/abs/2610.31005 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_62db2d48ce7a (Metraipolis archive dc_62db2d48ce7a, sha256 e7e479c4e3d3766b32d5bd1e92ab9f5bfcf2413a82348fbbab909c58bf5df2be)
BibTeX
@misc{dc62db2d48ce7a,
  title = {Mitigating Saliency Collapse: Robust Saliency-Aware Long-Text Image-Text Alignment — arXiv:2610.31005},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.31005},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_62db2d48ce7a; sha256 e7e479c4e3d3766b32d5bd1e92ab9f5bfcf2413a82348fbbab909c58bf5df2be},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.31005. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Long-text image-text alignment is essential for fine-grained scene understanding, yet existing methods mainly focus on extending context length or improving local alignment. We identify an overlooked failure mode, saliency collapse, where models over-rely on text that features visual prominence while underutilizing non-salient context. Consequently, retrieval performance degrades sharply when salient content is reduced, despite that the non-salient context remains informative. To address this, we propose RoSA, a saliency-aware finetuning framework that decomposes image-text pairs into complementary salient and contextual views, enabling joint learning of prominent and contextual semantics. We propose a modality-independent decomposition strategy to combine object-level visual saliency estimation with LLM-based text decomposition to bridge vision and language. We employ auxiliary objectives to explicitly align these views, ensuring robust representation with both salient and contextual features. RoSA greatly improves long-text retrieval and demonstrates strong robustness to saliency collapse across benchmarks. Moreover, our auxiliary objectives are plug-and-play, bringing consistent gains to existing tuning methods. Code and checkpoints are available at: this https URL

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.31005 — 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 2026-10-25T20:47:56.891Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).