Metraipolis › archive › dc_7d696197fa22

SV-TAD: Native Sparse Convs for Efficient Temporal Action Detection — arXiv:2610.11579

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.11579
License (as archived)
CC0-1.0
Archived
2026-10-09T06:30:35.625Z
SHA-256 of the archived text
cdf5e5845f53c7f5741f75959f3137aed51995c4c209ccce5d5da64d0d548e1a
Stable URL of this page
https://metraipolis.com/documento/dc_7d696197fa22

How to cite

SV-TAD: Native Sparse Convs for Efficient Temporal Action Detection — arXiv:2610.11579 — arXiv (en), https://arxiv.org/abs/2610.11579 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7d696197fa22 (Metraipolis archive dc_7d696197fa22, sha256 cdf5e5845f53c7f5741f75959f3137aed51995c4c209ccce5d5da64d0d548e1a)
BibTeX
@misc{dc7d696197fa22,
  title = {SV-TAD: Native Sparse Convs for Efficient Temporal Action Detection — arXiv:2610.11579},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.11579},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7d696197fa22; sha256 cdf5e5845f53c7f5741f75959f3137aed51995c4c209ccce5d5da64d0d548e1a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.11579. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

To adapt billion-parameter Vision Transformers for long-video understanding, recent methods freeze the backbone and train lightweight convolutional modules. While effective for parameter-efficient training, existing adapters do not reduce inference-time computation, leaving scalability with respect to video length largely unaddressed. Token selection can reduce attention cost by pruning redundant tokens, but it breaks the spatial grid structure required by convolutional adapters. This forces an expensive dense reconstruction, nullifying much of the potential speedup. We address this by introducing native sparse 2D convolutions, a primitive that allows these adapters, for the first time, to operate directly and efficiently on dynamically pruned token sets. We integrate this primitive into SV-TAD, an adapter framework for temporal action detection, reducing VideoMAEv2-L computation by up to 64% and achieving 2.2x faster inference, while maintaining state-of-the-art accuracy on THUMOS-14 and ActivityNet-1.3. When scaled to InternVideoNext-L, our approach surpasses the previous state of the art at roughly half its computational cost. Moreover, the sparse formulation naturally supports auxiliary task tokens, which improves fine-grained assembly detection on ATTACH.

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/2610.11579 — 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 2026-10-09T05:55:07.413Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).