Metraipolis › archive › dc_f599b2fbe9d2

DynThinker: Unifying Static 3D and Dynamic 4D Spatial Reasoning — arXiv:2610.20598

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.20598
License (as archived)
CC0-1.0
Archived
2026-10-15T21:57:30.025Z
SHA-256 of the archived text
7256640bd54cdff02a4b280663e8d95d222b1f07c44d0b009fcc09b949c53a2f
Stable URL of this page
https://metraipolis.com/documento/dc_f599b2fbe9d2

How to cite

DynThinker: Unifying Static 3D and Dynamic 4D Spatial Reasoning — arXiv:2610.20598 — arXiv (en), https://arxiv.org/abs/2610.20598 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f599b2fbe9d2 (Metraipolis archive dc_f599b2fbe9d2, sha256 7256640bd54cdff02a4b280663e8d95d222b1f07c44d0b009fcc09b949c53a2f)
BibTeX
@misc{dcf599b2fbe9d2,
  title = {DynThinker: Unifying Static 3D and Dynamic 4D Spatial Reasoning — arXiv:2610.20598},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.20598},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f599b2fbe9d2; sha256 7256640bd54cdff02a4b280663e8d95d222b1f07c44d0b009fcc09b949c53a2f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.20598. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Achieving general spatial intelligence requires a unified understanding of both static 3D and dynamic 4D. However, we find that existing models struggle to excel at both: 3D post-training can weaken dynamic understanding, while 4D-specialized models often lack strong static 3D reasoning. Our analysis reveals that this spatial--dynamic interference is highly non-uniform across parameter groups, with the balance between static and dynamic capabilities particularly sensitive to changes in late-layer MLPs. To address this interference, we propose DynThinker, a unified framework for jointly strengthening static and dynamic spatial reasoning. Decoupled Exponential Moving Average (D-EMA) applies configurable group-wise averaging to constrain drift in interference-sensitive pretrained parameters while preserving flexibility for capability enhancement. Spatial-Dynamic Fusion (SDF) complements D-EMA by selectively integrating spatial and temporal priors from a dynamic geometry encoder into intermediate VLM representations under semantic guidance. Across ten metrics spanning nine static 3D and dynamic 4D benchmarks, DynThinker achieves the highest average score among all evaluated models. Notably, DynThinker achieves 56.0 on ReVSI, 67.2 on DSR-Bench, and 51.2 and 40.1 under the sample-wise and group-wise evaluation protocols of DSI-Bench, respectively, setting new state-of-the-art results among the compared open-source models across all four evaluations within a unified model.

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/2610.20598 — 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 2026-10-15T21:38:29.551Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).