World-Ego Modeling for Long-Horizon Evolution in Hybrid Embodied Tasks
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917511292780544 |
|---|---|
| author | Lin, Zuyao Zhang, Jianhui Jia, Peidong Zhao, Xiaoguang Zhang, Shanghang Chen, Xingyu |
| author_facet | Lin, Zuyao Zhang, Jianhui Jia, Peidong Zhao, Xiaoguang Zhang, Shanghang Chen, Xingyu |
| contents | World models are widely explored in embodied intelligence, yet they typically predict distinct evolutions of the world and the ego within a single stream, where the world captures persistent instruction-agnostic scene regularities and the ego captures robot-centric instruction-conditioned dynamics. This world-ego entanglement leads to a degradation in long-horizon embodied scenarios, particularly in hybrid tasks with interleaved navigation and manipulation behaviors. In this paper, we introduce \emph{World-Ego Modeling}, a new conceptual paradigm that decomposes future evolution into world and ego components. We define the world-ego boundary from three perspectives, i.e., motion-, semantic-, and intention-based views, and analyze three disentanglement strategies with post-, pre-, and full disentanglement. Further, we instantiate this paradigm as the World-Ego Model (WEM), a unified embodied world model that couples an implicit separate world-ego planner with a cascade-parallel mixture-of-experts (CP-MoE) diffusion generator. To enable rigorous evaluation, we further construct HTEWorld, the first benchmark for long-horizon world modeling with hybrid navigation-manipulation tasks, providing 125K video clips (over 4.5M frames) with fine-grained action annotations and 300 multi-turn evaluation trajectories (over 2K instructions). Extensive experiments show that WEM achieves state-of-the-art performance on HTEWorld while remaining competitive on existing manipulation-only benchmarks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_19957 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | World-Ego Modeling for Long-Horizon Evolution in Hybrid Embodied Tasks Lin, Zuyao Zhang, Jianhui Jia, Peidong Zhao, Xiaoguang Zhang, Shanghang Chen, Xingyu Computer Vision and Pattern Recognition Artificial Intelligence Robotics World models are widely explored in embodied intelligence, yet they typically predict distinct evolutions of the world and the ego within a single stream, where the world captures persistent instruction-agnostic scene regularities and the ego captures robot-centric instruction-conditioned dynamics. This world-ego entanglement leads to a degradation in long-horizon embodied scenarios, particularly in hybrid tasks with interleaved navigation and manipulation behaviors. In this paper, we introduce \emph{World-Ego Modeling}, a new conceptual paradigm that decomposes future evolution into world and ego components. We define the world-ego boundary from three perspectives, i.e., motion-, semantic-, and intention-based views, and analyze three disentanglement strategies with post-, pre-, and full disentanglement. Further, we instantiate this paradigm as the World-Ego Model (WEM), a unified embodied world model that couples an implicit separate world-ego planner with a cascade-parallel mixture-of-experts (CP-MoE) diffusion generator. To enable rigorous evaluation, we further construct HTEWorld, the first benchmark for long-horizon world modeling with hybrid navigation-manipulation tasks, providing 125K video clips (over 4.5M frames) with fine-grained action annotations and 300 multi-turn evaluation trajectories (over 2K instructions). Extensive experiments show that WEM achieves state-of-the-art performance on HTEWorld while remaining competitive on existing manipulation-only benchmarks. |
| title | World-Ego Modeling for Long-Horizon Evolution in Hybrid Embodied Tasks |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Robotics |
| url | https://arxiv.org/abs/2605.19957 |