Metraipolis › archive › dc_fbabe53de7c8

GenDSR: Transferring Representations from Video Generation Models for Dynamic Spatial Reasoning — arXiv:2610.41314

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.41314
License (as archived): CC0-1.0
Archived: 2026-11-02T07:30:59.628Z
SHA-256 of the archived text: cc62e5324a70fe00fd1f310e0fbecfc8cb93a43f4ae8bf2f5f70d0333fe74787
Stable URL of this page: https://metraipolis.com/documento/dc_fbabe53de7c8

How to cite

GenDSR: Transferring Representations from Video Generation Models for Dynamic Spatial Reasoning — arXiv:2610.41314 — arXiv (en), https://arxiv.org/abs/2610.41314 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fbabe53de7c8 (Metraipolis archive dc_fbabe53de7c8, sha256 cc62e5324a70fe00fd1f310e0fbecfc8cb93a43f4ae8bf2f5f70d0333fe74787)

BibTeX

@misc{dcfbabe53de7c8,
  title = {GenDSR: Transferring Representations from Video Generation Models for Dynamic Spatial Reasoning — arXiv:2610.41314},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.41314},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fbabe53de7c8; sha256 cc62e5324a70fe00fd1f310e0fbecfc8cb93a43f4ae8bf2f5f70d0333fe74787},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41314. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Dynamic spatial reasoning is essential for connecting visual intelligence to thephysical world, yet remains challenging for vision-language models (VLMs), which must preserve scene state while tracking how spatial relations evolve over time. Meanwhile, video generation models (VGMs), trained to synthesize temporally coherent world dynamics, offer a rich source of spatiotemporal representations. We present GenDSR, a framework that transfers intermediate representations from VGM into VLM for dynamic spatial reasoning. Rather than treating a VGM feature as a monolithic prior, GenDSR constructs two complementary branches: a raw branch preserving holistic generative context and a delta branch emphasizing temporal variation. Task-conditioned tokenwise gates then fuse both signals and inject the result into the native visual representation. On DSR-Bench, GenDSR achieves 67.2%, surpassing the previous best result and establishing a new state of the art.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.41314 — 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 2026-11-02T07:26:57.805Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).