Metraipolis › archive › dc_f23bd428c4fa

History Forcing: Distilling History-Dependent Continuations for Long-Horizon Video World Models — arXiv:2611.02782

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.02782
License (as archived): CC0-1.0
Archived: 2026-11-03T11:42:38.580Z
SHA-256 of the archived text: 2d8dcac56dcea683e67a85535c4b4bff919828fef4bda1a3f7f456a9c77fa113
Stable URL of this page: https://metraipolis.com/documento/dc_f23bd428c4fa

How to cite

History Forcing: Distilling History-Dependent Continuations for Long-Horizon Video World Models — arXiv:2611.02782 — arXiv (en), https://arxiv.org/abs/2611.02782 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f23bd428c4fa (Metraipolis archive dc_f23bd428c4fa, sha256 2d8dcac56dcea683e67a85535c4b4bff919828fef4bda1a3f7f456a9c77fa113)

BibTeX

@misc{dcf23bd428c4fa,
  title = {History Forcing: Distilling History-Dependent Continuations for Long-Horizon Video World Models — arXiv:2611.02782},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.02782},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f23bd428c4fa; sha256 2d8dcac56dcea683e67a85535c4b4bff919828fef4bda1a3f7f456a9c77fa113},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.02782. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Few-step autoregressive video diffusion enables efficient generation, yet maintaining consistency over long rollouts remains challenging, especially when subjects or scenes reappear after going out of sight. Despite progress in fidelity and rollout robustness, existing distillation methods do not fully exploit richer historical information to constrain the student's current prediction. To address this gap, we introduce \textbf{History Forcing}, a unified on-policy distillation framework that strengthens historical supervision for few-step autoregressive video generators. The framework uses the residual between teacher predictions with and without informative historical context to decouple this supervision. Such a historical context can be obtained using earlier generated chunks with less accumulated drift or a fixed-view teacher rollout that keeps the scene visible while the student looks away. The residual is then weighted and added to the score or velocity prediction used for distillation, forming a corrected target. By aligning the student with this corrected target through distillation, History Forcing incorporates long-range historical constraints into autoregressive generation. Our method achieves state-of-the-art performance among 1.3B models, with a VBench quality score of 85.07 for 15s video generation and a VBench-Long aggregate score of 79.82 for 60s generation. For out-of-sight world modeling, evaluations further demonstrate more plausible scene evolution and more consistent states when the camera returns, with Evolution Quality improving from 2.35 to 2.70 on PlayWorld and Object Consistency from 58.43 to 62.18 on MemoBench.

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/2611.02782 — 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 2026-11-03T11:06:37.591Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).