Metraipolis › archive › dc_72cdac75d6f6

Unifying Video Tasks via Spatiotemporal Analogy — arXiv:2609.33935

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.33935
License (as archived)
CC0-1.0
Archived
2026-10-07T21:01:56.115Z
SHA-256 of the archived text
29de0ca0851ef46379abc7990bde7b0976b5ae2fa72fa6cccb4dc30445d2cb4b
Stable URL of this page
https://metraipolis.com/documento/dc_72cdac75d6f6

How to cite

Unifying Video Tasks via Spatiotemporal Analogy — arXiv:2609.33935 — arXiv (en), https://arxiv.org/abs/2609.33935 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_72cdac75d6f6 (Metraipolis archive dc_72cdac75d6f6, sha256 29de0ca0851ef46379abc7990bde7b0976b5ae2fa72fa6cccb4dc30445d2cb4b)
BibTeX
@misc{dc72cdac75d6f6,
  title = {Unifying Video Tasks via Spatiotemporal Analogy — arXiv:2609.33935},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.33935},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_72cdac75d6f6; sha256 29de0ca0851ef46379abc7990bde7b0976b5ae2fa72fa6cccb4dc30445d2cb4b},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.33935. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Adapting video models to new tasks typically requires dedicated data curation and fine-tuning. While visual analogy provides a training-free alternative by specifying tasks in-context, it remains restricted to the image domain. To explore whether analogy-based methods can unify diverse video tasks and generalize to out-of-distribution scenarios, we introduce ViGeo, a framework that extends visual in-context learning to the video domain via spatiotemporal canvas completion. Evaluated on a diverse task taxonomy with a strict train-test split, ViGeo generalizes to unseen video manipulations and zero-shot modalities (e.g., event cameras). Finally, we identify task internalization, where a query format associated with a pretrained task overrides the demonstration, and show that this shortcut can be removed with a small amount of task-unrelated data, highlighting the need to decorrelate prompt format from task identity.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.33935 — 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 2026-10-07T20:29:02.792Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).