Metraipolis › archive › dc_2b3129354791
Disen-Forcing: Disentangling Semantic Anchoring from Motion for Autoregressive Video Diffusion — arXiv:2611.02822
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2611.02822
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-03T11:25:27.186Z
- SHA-256 of the archived text
504fb843978a62d2b7e7b74ef8f6ce73775c8fe4ee098dd0091bf9654769cab0- Stable URL of this page
https://metraipolis.com/documento/dc_2b3129354791
How to cite
Disen-Forcing: Disentangling Semantic Anchoring from Motion for Autoregressive Video Diffusion — arXiv:2611.02822 — arXiv (en), https://arxiv.org/abs/2611.02822 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2b3129354791 (Metraipolis archive dc_2b3129354791, sha256 504fb843978a62d2b7e7b74ef8f6ce73775c8fe4ee098dd0091bf9654769cab0)
BibTeX
@misc{dc2b3129354791,
title = {Disen-Forcing: Disentangling Semantic Anchoring from Motion for Autoregressive Video Diffusion — arXiv:2611.02822},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2611.02822},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2b3129354791; sha256 504fb843978a62d2b7e7b74ef8f6ce73775c8fe4ee098dd0091bf9654769cab0},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.02822. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Autoregressive video diffusion models (AR-VDMs) enable real-time, long-horizon video generation but degrade rapidly once the inference horizon outruns the training horizon, due to compounding exposure bias. To stabilize long rollouts, recent methods heuristically adopt frame sink, keeping initial frames as a persistent model context, inspired by attention sinks in LLMs. We revisit this design and find that the mechanism by which initial frames stabilize long-range consistency in AR-VDMs is not an LLM-style attention sink; instead, long-range consistency is anchored through a sparse subset of attention heads, while the majority of heads predominantly capture local temporal dynamics. This observation reveals an over-conditioning pathology in existing frame-sink strategies: by universally exposing every head to the pinned initial-frame context, the model becomes prone to collapsing motion toward the low-dynamic modes of the teacher during model distillation. We propose Disen-Forcing, a distillation-time plug-in for AR-VDMs that disentangles and reinforces this head specialization through disentangled context routing and motion-disentangling regularization. On long-horizon video generation benchmarks, Disen-Forcing simultaneously improves long-range consistency and motion dynamics of base AR-VDMs, a trade-off that sink-based methods fail to balance. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2611.02822 — 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 2026-11-03T11:02:37.886Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).