论文/arXiv 2610.07922·22 天前

OpenWAM：用于可组合世界-动作模型的开放框架

OPENWAM用因果机器人视频预训练与可配置视频-动作交互构建WAM，并在LIBERO套件和真实双臂任务上验证。

AI 整理

• 本体 SDK 与控制
• 数据与采集
• 具身模型

原文链接：OpenWAM: An Open Framework for Composable World-Action Models（arXiv 2610.07922 · 发布日期：2026-10-06）

OpenWAM: An Open Framework for Composable World-Action Models

OPENWAM用因果机器人视频预训练与可配置视频-动作交互构建WAM，并在LIBERO套件和真实双臂任务上验证。

核心要点

• 现有WAM常同时改变骨干、交互、监督和推理，难以比较设计选择。
• 从Wan2.2-5B出发，经因果机器人视频预训练，并以共享Mixture-of-Transformers接入动作专家。
• 在四个LIBERO套件和真实双臂任务上取得高成功率，LIBERO-Long的VTA成功率由68.4%升至97.8%。
• 框架扩展到逆向和正向动力学，用于研究反事实转移相对仅示教数据的改进。

验证范围与局限：验证范围包括四个LIBERO套件、LIBERO-90留出任务、真实双臂任务及动力学评估。

英文摘要 (Abstract)

World-action models (WAMs) couple future prediction with robot control, yet existing systems often vary the video backbone, interaction structure, supervision, and inference procedure simultaneously, making their design choices difficult to compare. We introduce OPENWAM, an open world-action modeling framework built around a common causal robot-video foundation and configurable video-action interaction. Starting from Wan2.2-5B, we perform causal robot-video pretraining on over 10,000 hours of video, then integrate an action expert through a shared Mixture-of-Transformers architecture that supports joint, video-then-action, action-then-video, and decoupled generation. OPENWAM achieves high success rates on four LIBERO suites and real-world bimanual tasks; robot-video training with causal adaptation improves VTA success on LIBERO-Long from 68.4% to 97.8%. The same configurable architecture naturally extends to inverse and forward dynamics, allowing us to study how counterfactual transitions improve independently trained dynamics models beyond demonstrations alone. When only the video predictor is adapted to a new task, a frozen local-context inverse dynamics model trained on counterfactual data and demonstrations achieves 84.0% mean success across four held-out LIBERO-90 tasks, compared with 47.0% for a full-context inverse model and 21.5% for a local-context model trained only on demonstrations. For forward dynamics, counterfactual supervision reduces RGB prediction error by 34.5% and raises outcome identification from 21.1% to 71.3% among 16 same-state outcomes. OPENWAM provides a common testbed for comparing WAM interaction designs and for studying dynamics learning from video data beyond successful demonstrations.

作者：Heng Yu, David D. Yuan, Juze Zhang, Changan Chen, Yao Feng, Michelle Q. Wang Baldonado, Steve Cousins, Li Fei-Fei, Jiajun Wu, Ehsan Adeli

来源

本页事实对应的公开页面，以原文为准。

1. arXiv 2610.07922：OpenWAM: An Open Framework for Composable World-Action Models访问 2026-10-21