Metraipolis › archive › dc_7d1fb3afe58b

Learning to Retrieve: Internalizing Memory Retrieval for Video World Models — arXiv:2610.11444

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.11444
License (as archived)
CC0-1.0
Archived
2026-10-09T06:30:33.225Z
SHA-256 of the archived text
e12ee09e7a1da91e40da3d5453e3eac934dde6022cea66b90036380853cf636c
Stable URL of this page
https://metraipolis.com/documento/dc_7d1fb3afe58b

How to cite

Learning to Retrieve: Internalizing Memory Retrieval for Video World Models — arXiv:2610.11444 — arXiv (en), https://arxiv.org/abs/2610.11444 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7d1fb3afe58b (Metraipolis archive dc_7d1fb3afe58b, sha256 e12ee09e7a1da91e40da3d5453e3eac934dde6022cea66b90036380853cf636c)
BibTeX
@misc{dc7d1fb3afe58b,
  title = {Learning to Retrieve: Internalizing Memory Retrieval for Video World Models — arXiv:2610.11444},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.11444},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7d1fb3afe58b; sha256 e12ee09e7a1da91e40da3d5453e3eac934dde6022cea66b90036380853cf636c},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.11444. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Video world models aim to generate explorable, 3D-consistent scene videos conditioned on camera trajectories. Existing approaches often rely on external memory systems that explicitly retrieve previously observed content to mitigate scene drift during long-horizon generation. However, these auxiliary memory pathways operate outside the model's internal generative dynamics, preventing the model from intrinsically learning when and what historical information should be retrieved. We propose to internalize memory retrieval into the generation process, allowing retrieval to emerge as an intrinsic behavior of the video world model rather than relying on an external memory system. Based on this principle, we introduce \textbf{Learning-to-Retrieve (L2R)}, which repurposes the model's persistent internal state as a memory for historical context. A camera-conditioned retrieval gate selectively accesses relevant historical information from this state, determining \textit{what to retrieve}, while a retrieval trigger determines \textit{when to retrieve}. We further supervise the trigger with a 3D re-visibility signal, activating retrieval when previously observed content re-enters the current view while otherwise preserving the existing context. Together, these components enable the model to intrinsically acquire memory retrieval behavior and incorporate relevant historical observations into generation without a separate retrieval pathway. Across multiple base models and camera-revisit benchmarks, L2R improves long-term scene consistency while eliminating the need for an external memory bank or 3D conditions. this https URL

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/2610.11444 — 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 2026-10-09T05:55:07.397Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).