Metraipolis › archive › dc_7abdad7d6858
Reason First, Retrieve What Matters: Hypothesis-Guided Keyframe Selection for Long Video Understanding — arXiv:2611.10609
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.10609
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-10T18:42:39.728Z
- SHA-256 of the archived text
61255e0a27060bd4d05eaf0dc7d19cbc99c406662ae25374b60fc437eda1489d- Stable URL of this page
https://metraipolis.com/documento/dc_7abdad7d6858
How to cite
Reason First, Retrieve What Matters: Hypothesis-Guided Keyframe Selection for Long Video Understanding — arXiv:2611.10609 — arXiv (en), https://arxiv.org/abs/2611.10609 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7abdad7d6858 (Metraipolis archive dc_7abdad7d6858, sha256 61255e0a27060bd4d05eaf0dc7d19cbc99c406662ae25374b60fc437eda1489d)
BibTeX
@misc{dc7abdad7d6858,
title = {Reason First, Retrieve What Matters: Hypothesis-Guided Keyframe Selection for Long Video Understanding — arXiv:2611.10609},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.10609},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7abdad7d6858; sha256 61255e0a27060bd4d05eaf0dc7d19cbc99c406662ae25374b60fc437eda1489d},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.10609. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Long video understanding requires models to identify and reason over sparse, temporally distributed evidence within limited frame budgets. Existing keyframe selection methods typically retrieve query-relevant frames from a densely sampled candidate set before inference. However, global image-text matching incurs unnecessary computation when a small set of uniformly sampled frames already provides sufficient evidence. Query relevance alone may also overlook evidence that distinguishes competing answers, while one-shot selection offers no opportunity to correct earlier selection errors. We propose HypoKFS, a training-free, plug-and-play framework that couples keyframe selection with iterative inference by multimodal large language models (MLLMs). HypoKFS first reasons over sparsely sampled frames to construct an initial hypothesis space. The resulting answer distribution determines whether additional frame selection is needed. When needed, the retained hypotheses guide the selection of discriminative frames, followed by renewed inference and hypothesis refinement. This iterative process directs additional visual computation toward unresolved hypotheses while avoiding unnecessary global frame matching. Experiments on Video-MME, LongVideoBench, and MLVU demonstrate consistent improvements across multiple MLLM backbones. HypoKFS improves accuracy by up to 6.7 percentage points and achieves 72.9% accuracy on Video-MME at only 1.3% of the reference frame-hypothesis matching cost. 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/2611.10609 — 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 2026-11-10T18:37:37.597Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).