Gespeichert in:
| Hauptverfasser: | , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2605.24578 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917528021762048 |
|---|---|
| author | Wang, Zijie Zhang, Wei Zhang, Weiming Zhang, Fanqi Tan, Xiao Qin, Yipeng Li, Guanbin |
| author_facet | Wang, Zijie Zhang, Wei Zhang, Weiming Zhang, Fanqi Tan, Xiao Qin, Yipeng Li, Guanbin |
| contents | Video world models have achieved strong visual realism, but this does not ensure that their dynamics are truly governed by actions. In this work, we argue that action faithfulness should be understood through the compositional structure of actions, which in many embodied settings follows a group structure (e.g., SE(2) for navigation). Based on this insight, we formalize action-conditioned world modeling as realizing a group action on the state space, providing a principled criterion for evaluating dynamics beyond visual quality. To operationalize this framework, we propose a unified approach that enforces identity, inverse, and composition consistency via latent-space regularization with synthesized supervision, avoiding additional data collection. We further introduce two metrics: Group-Action Consistency (GAC) and Group-Action Robustness (GAR), to evaluate structural correctness and rollout stability. Extensive experimental results show that our method consistently improves both GAC and GAR in state-of-the-art video world models without degrading perceptual quality. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_24578 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | World Models as Group Actions Wang, Zijie Zhang, Wei Zhang, Weiming Zhang, Fanqi Tan, Xiao Qin, Yipeng Li, Guanbin Computer Vision and Pattern Recognition Video world models have achieved strong visual realism, but this does not ensure that their dynamics are truly governed by actions. In this work, we argue that action faithfulness should be understood through the compositional structure of actions, which in many embodied settings follows a group structure (e.g., SE(2) for navigation). Based on this insight, we formalize action-conditioned world modeling as realizing a group action on the state space, providing a principled criterion for evaluating dynamics beyond visual quality. To operationalize this framework, we propose a unified approach that enforces identity, inverse, and composition consistency via latent-space regularization with synthesized supervision, avoiding additional data collection. We further introduce two metrics: Group-Action Consistency (GAC) and Group-Action Robustness (GAR), to evaluate structural correctness and rollout stability. Extensive experimental results show that our method consistently improves both GAC and GAR in state-of-the-art video world models without degrading perceptual quality. |
| title | World Models as Group Actions |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.24578 |