WAM-OPD：面向世界动作模型后训练的视频-动作联合监督与同策略蒸馏

Authors:Liuhaichen Yang, Zhengyang Zhong, Hanshang Zhu, Ningwei Bai, Qichen Yin, Zhi Han, Jiarui Qin, Zhuang Jiang, Chenchao Sheng, Hanbo Ma, Junkai Liu, Junkai Sun, Dongcheng Lyu, Bo Liu, Yi Dong, Zezhi Tang

View PDF HTML (experimental)

Abstract:World Action Models (WAMs) generate both future video and robot actions, offering two connected outputs for post-training supervision. How can a pretrained WAM learn from a stronger Teacher on the histories it encounters during execution? We present WAM-OPD, which collects Student rollout histories and queries a Teacher for paired video and action targets. The Student learns from both targets while retaining its one-step video and action generation at deployment. Across 12 RoboTwin 2.0 tasks, WAM-OPD improves average success from 33.8% to 65.7%; across four real-robot tasks, it improves average success from 51.4% to 64.6%. With the collected Student histories held fixed, joint video-action supervision achieves the highest observed success on all three ablation tasks, while either modality alone also improves performance. A separate comparison with Teacher-generated histories finds task-dependent differences between the two history sources. These results demonstrate the value of paired video-action supervision for improving WAM policies without increasing their deployed sampling budget.

Subjects:	Artificial Intelligence (cs.AI); Robotics (cs.RO)
Cite as:	arXiv:2608.22364 [cs.AI]
	(or arXiv:2608.22364v2 [cs.AI] for this version)
	https://doi.org/10.48550/arXiv.2608.22364 arXiv-issued DOI via DataCite

Submission history

From: Liuhaichen Yang [view email]
[v1] Sun, 23 Aug 2026 11:06:45 UTC (1,220 KB)
[v2] Thu, 1 Oct 2026 18:05:02 UTC (3,635 KB)