Metraipolis › archive › dc_74952cc8281b

Allocate Before You Embed: Adaptive Visual Input Allocation for Video Embeddings — arXiv:2609.01778

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.01778
License (as archived)
CC0-1.0
Archived
2026-10-17T05:49:42.028Z
SHA-256 of the archived text
8e1facb7a1b5e460931c372e0862c3026d5d0398c205923c43b95ce0db799db4
Stable URL of this page
https://metraipolis.com/documento/dc_74952cc8281b

How to cite

Allocate Before You Embed: Adaptive Visual Input Allocation for Video Embeddings — arXiv:2609.01778 — arXiv (en), https://arxiv.org/abs/2609.01778 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_74952cc8281b (Metraipolis archive dc_74952cc8281b, sha256 8e1facb7a1b5e460931c372e0862c3026d5d0398c205923c43b95ce0db799db4)
BibTeX
@misc{dc74952cc8281b,
  title = {Allocate Before You Embed: Adaptive Visual Input Allocation for Video Embeddings — arXiv:2609.01778},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.01778},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_74952cc8281b; sha256 8e1facb7a1b5e460931c372e0862c3026d5d0398c205923c43b95ce0db799db4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.01778. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large-scale video retrieval requires embedding models to encode long and diverse videos under tight visual-input and inference budgets. Existing methods typically sample a small, fixed set of frames at their original resolution, limiting temporal coverage and ignoring frame importance. Our empirical analysis shows that expanding temporal coverage improves retrieval even under a fixed visual-input budget. Gains are larger when the original per-frame resolution is preserved, highlighting the complementary roles of temporal coverage and spatial fidelity. Motivated by this finding, we propose AllocEmbed, an allocate-then-embed framework that reallocates a fixed visual-input budget across more frames. A lightweight allocator uses low-cost previews to assign frame-wise resolutions before the embedding backbone, preserving more detail where it most benefits retrieval while reducing visual cost elsewhere. We further introduce Retrieval-Driven Policy Optimization (RDPO), which learns the allocator directly from retrieval feedback using a rank-validated similarity gap and a confidence-guided efficiency incentive. Operating entirely before the backbone, AllocEmbed integrates with existing retrieval systems without modifying the embedding model or downstream pipeline. Experiments on the MMEB-V2 V-QA and V-RET tasks and our LongRet benchmark show that AllocEmbed achieves the best overall retrieval performance among the evaluated budget-matched methods and transfers across embedding backbones. Our code is publicly available at this https URL .

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-03 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.01778 — 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 2026-10-17T05:25:58.014Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-03?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).