DeepVerse: 4D Autoregressive Video Generation as a World Model
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866915317766160384 |
|---|---|
| author | Chen, Junyi Zhu, Haoyi He, Xianglong Wang, Yifan Zhou, Jianjun Chang, Wenzheng Zhou, Yang Li, Zizun Fu, Zhoujie Pang, Jiangmiao He, Tong |
| author_facet | Chen, Junyi Zhu, Haoyi He, Xianglong Wang, Yifan Zhou, Jianjun Chang, Wenzheng Zhou, Yang Li, Zizun Fu, Zhoujie Pang, Jiangmiao He, Tong |
| contents | World models serve as essential building blocks toward Artificial General Intelligence (AGI), enabling intelligent agents to predict future states and plan actions by simulating complex physical interactions. However, existing interactive models primarily predict visual observations, thereby neglecting crucial hidden states like geometric structures and spatial coherence. This leads to rapid error accumulation and temporal inconsistency. To address these limitations, we introduce DeepVerse, a novel 4D interactive world model explicitly incorporating geometric predictions from previous timesteps into current predictions conditioned on actions. Experiments demonstrate that by incorporating explicit geometric constraints, DeepVerse captures richer spatio-temporal relationships and underlying physical dynamics. This capability significantly reduces drift and enhances temporal consistency, enabling the model to reliably generate extended future sequences and achieve substantial improvements in prediction accuracy, visual realism, and scene rationality. Furthermore, our method provides an effective solution for geometry-aware memory retrieval, effectively preserving long-term spatial consistency. We validate the effectiveness of DeepVerse across diverse scenarios, establishing its capacity for high-fidelity, long-horizon predictions grounded in geometry-aware dynamics. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_01103 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DeepVerse: 4D Autoregressive Video Generation as a World Model Chen, Junyi Zhu, Haoyi He, Xianglong Wang, Yifan Zhou, Jianjun Chang, Wenzheng Zhou, Yang Li, Zizun Fu, Zhoujie Pang, Jiangmiao He, Tong Computer Vision and Pattern Recognition World models serve as essential building blocks toward Artificial General Intelligence (AGI), enabling intelligent agents to predict future states and plan actions by simulating complex physical interactions. However, existing interactive models primarily predict visual observations, thereby neglecting crucial hidden states like geometric structures and spatial coherence. This leads to rapid error accumulation and temporal inconsistency. To address these limitations, we introduce DeepVerse, a novel 4D interactive world model explicitly incorporating geometric predictions from previous timesteps into current predictions conditioned on actions. Experiments demonstrate that by incorporating explicit geometric constraints, DeepVerse captures richer spatio-temporal relationships and underlying physical dynamics. This capability significantly reduces drift and enhances temporal consistency, enabling the model to reliably generate extended future sequences and achieve substantial improvements in prediction accuracy, visual realism, and scene rationality. Furthermore, our method provides an effective solution for geometry-aware memory retrieval, effectively preserving long-term spatial consistency. We validate the effectiveness of DeepVerse across diverse scenarios, establishing its capacity for high-fidelity, long-horizon predictions grounded in geometry-aware dynamics. |
| title | DeepVerse: 4D Autoregressive Video Generation as a World Model |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2506.01103 |