English(EN)Can World and Action Modeling Benefit Each Other under Unified Causal Training?
新的机器人世界动作模型提高效率和性能 · 跟踪 6 个来源
作者PulseAugur 编辑部·[6 个来源]·
研究人员开发了机器人世界动作模型（WAMs）的新方法。RAP-WAM 使用预测性排练和动作扰动来改进策略学习，在现实世界任务中取得了很高的成功率。Box-WAM 通过使用定向边界框轨迹而不是完整的视频预测来表示未来状态，提供了一种计算效率更高的方法，在保持性能的同时降低了成本。此外，OneWM 框架表明，在因果训练下统一世界模型和动作模型可以带来互惠互利，从而提高策略学习和世界预测的保真度。
AI
arXiv:2610.16443v1 Announce Type: new Abstract: Recent world action models draw on video generation priors to jointly predict future observations and robot actions. Adapting these priors to control requires determining how future predictions should guide policy learning. RAP-WAM …
arXiv cs.AI
TIER_1English(EN)·Qinxi Yu, Wenzhou Ding, Gloria X. Wang, Albert J. Zhai, Shenlong Wang·
arXiv:2610.17526v1 Announce Type: cross Abstract: World action models (WAMs) are a category of policy models that ex- plicitly predict both actions and future world states. They have been shown to outperform action-only models across various robotic manipulation tasks. How- ever,…
World and action modeling are closely related, yet their reciprocal benefits under unified causal training remain unclear. We introduce OneWM, a simple framework for unified causal training that treats actions as both conditions for world prediction and targets for policy learnin…
OneWM unifies causal training of world and action modeling, showing that learning action-conditioned dynamics improves policies and action learning improves multi-view world prediction.
World action models (WAMs) are a category of policy models that ex- plicitly predict both actions and future world states. They have been shown to outperform action-only models across various robotic manipulation tasks. How- ever, representing the future with full-frame video pre…
arXiv:2610.21095v1 Announce Type: new Abstract: World and action modeling are closely related, yet their reciprocal benefits under unified causal training remain unclear. We introduce OneWM, a simple framework for unified causal training that treats actions as both conditions for…
