Metraipolis › archive › dc_ea8bfa20aae6

Temporal Specification Guidance for Text-to-Video Generation — arXiv:2610.20822

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.20822
License (as archived)
CC0-1.0
Archived
2026-10-16T08:59:30.233Z
SHA-256 of the archived text
ae6abe34cf84e9831d0c8f03a1df774775ad01a5d37c8f48df7ac14800df6946
Stable URL of this page
https://metraipolis.com/documento/dc_ea8bfa20aae6

How to cite

Temporal Specification Guidance for Text-to-Video Generation — arXiv:2610.20822 — arXiv (en), https://arxiv.org/abs/2610.20822 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ea8bfa20aae6 (Metraipolis archive dc_ea8bfa20aae6, sha256 ae6abe34cf84e9831d0c8f03a1df774775ad01a5d37c8f48df7ac14800df6946)
BibTeX
@misc{dcea8bfa20aae6,
  title = {Temporal Specification Guidance for Text-to-Video Generation — arXiv:2610.20822},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.20822},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ea8bfa20aae6; sha256 ae6abe34cf84e9831d0c8f03a1df774775ad01a5d37c8f48df7ac14800df6946},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.20822. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Text-to-video models often generate the visual concepts requested in natural language prompts but fail to satisfy their intended ordering, timing, and persistence requirements. We propose representing these requirements as specifications in a formal Temporal Logic, and introduce Temporal Specification Guidance (TSG), a training-free inference-time framework that turns temporal specifications into executable objectives for guiding video generation during sampling. TSG scores traces of visual concept evidence using differentiable quantitative semantics and uses the score's gradients to construct a specification-guided denoising direction. We instantiate this framework with video-text attention traces, efficient differentiable scores for Metric Temporal Logic, and gradient-derived attention biases. This instantiation requires neither an external concept classifier nor backpropagation through the video transformer. Its attention-steered forward pass yields a denoising direction that augments classifier-free guidance (CFG). On a benchmark of 220 prompts spanning 11 categories of temporal constraints, including ordered object entry and state transitions, this instantiation improves specification satisfaction accuracy over CFG by 9.5-20 percentage points across five models, with minimal memory overhead. These results demonstrate the potential of temporal specifications and differentiable satisfaction objectives as a flexible interface for controlling video generation.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.20822 — 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 2026-10-16T08:44:17.663Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).