| title | Rethinking World-Action Model for Compositional and In-Context Robotic Manipulation |
|---|---|
| authors | Shukai Gong, Xuanran Zhai, Yintianrun Zhang, Ruopeng Cui, Ye Huang, Yiyang Fu, Dexuan Lyu, Chaojie Li, Xinyi Song, Peiwen Lin, Chuang Wang, Mingyuan Jia, Yufan Deng, Jiaxin Fang, Bo Liang, Jiaxin Li, Yuxiang Gao, Hao Liu, Daquan Zhou |
| date | 2026-10-01 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2610.02368 |
| affiliation | DAGroup-PKU |
| upvotes | 69 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | Long-horizon compositional manipulation has become increasingly important for real-world robot deployment, where a single task involves multiple coordinated subtasks. Existing world-action models (WAMs) jointly predict short-horizon visual futures and actions, but typically lack explicit subtask-level reasoning. We propose Visual Goal-conditioned Action Reasoning (ViGAR), a hierarchical framework that factorizes manipulation into a visual subgoal planner and a subgoal executor. Given the current observation and global instruction, the subgoal planner predicts a visual subgoal for the next subtask. The subgoal executor then jointly generates future visual trajectories and actions conditioned on the predicted subgoal. Both components share a pretrained world-model representation, enabling task-level planning and action generation to benefit from common physical knowledge. Moreover, our framework naturally supports in-context learning: using a global goal image as context can induce different subtask decompositions and behaviors without parameter updates. On the RoboTwin Clean2Random benchmark, ViGAR achieves 82.00% and 67.02% success rates under the Clean and Random settings, respectively, surpassing the strongest baseline by 12.86 percentage points in average success rate. Real-world robot experiments on five compositional and two in-context learning tasks further confirm the effectiveness of ViGAR. |
cite
@misc{gong2026rethinking,
title={Rethinking World-Action Model for Compositional and In-Context Robotic Manipulation},
author={Shukai Gong and Xuanran Zhai and Yintianrun Zhang and Ruopeng Cui and Ye Huang and Yiyang Fu and Dexuan Lyu and Chaojie Li and Xinyi Song and Peiwen Lin and Chuang Wang and Mingyuan Jia and Yufan Deng and Jiaxin Fang and Bo Liang and Jiaxin Li and Yuxiang Gao and Hao Liu and Daquan Zhou},
year={2026},
eprint={2610.02368},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2610.02368}
}