Metraipolis › archive › dc_0c19f47c9c3b

Velox: Spatial Memory Makes World Models Efficient — arXiv:2610.22311

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.22311
License (as archived): CC0-1.0
Archived: 2026-10-16T10:39:15.526Z
SHA-256 of the archived text: fe6d0f4c9a828995e8b34113c835cc2163dc7dd91a9422f0d818f6d0793e9b8b
Stable URL of this page: https://metraipolis.com/documento/dc_0c19f47c9c3b

How to cite

Velox: Spatial Memory Makes World Models Efficient — arXiv:2610.22311 — arXiv (en), https://arxiv.org/abs/2610.22311 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0c19f47c9c3b (Metraipolis archive dc_0c19f47c9c3b, sha256 fe6d0f4c9a828995e8b34113c835cc2163dc7dd91a9422f0d818f6d0793e9b8b)

BibTeX

@misc{dc0c19f47c9c3b,
  title = {Velox: Spatial Memory Makes World Models Efficient — arXiv:2610.22311},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.22311},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0c19f47c9c3b; sha256 fe6d0f4c9a828995e8b34113c835cc2163dc7dd91a9422f0d818f6d0793e9b8b},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.22311. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Video world models turn actions and camera motion into long visual rollouts, making video a natural interface for interactive simulation. Such interaction requires responsive generation, yet coherent rollouts must retain visual memory of what has already been generated. Chunked video diffusion models usually keep this context in growing KV caches, so later chunks become progressively slower. To address this bottleneck, we introduce Velox, a training-free attention router for region-level context selection using frustum slab correspondences from streaming geometry. Velox represents historical blocks as camera frustum slabs with depth intervals and obtains current-block depth through recency-priority multiwarp. For well-covered regions, frustum slab overlap softly modulates each layer's model-derived block importance. Uncovered regions retain the model-derived route without geometric bias. The fixed token-block granularity matches FlashAttention-style tiled execution and reduces the attention term that grows with rollout length. Frustum slab construction runs asynchronously with denoising on the same H200. At chunk 48, Velox accelerates attention by $4.36\times$ and denoising by $2.43\times$. On the full 158-case WBench navigation split, it remains within 0.15 API-free average points of dense LingBot-Fast-Cam. Please see our project page at this https URL for more details.

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/2610.22311 — 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 2026-10-16T08:52:02.737Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).