监督幸存信息：基于几何引导的合成机器人视频VLA适配
Supervise What Survives: Geometry-Guided VLA Adaptation from Synthetic Robot Videos
AI总结：
提出GRA方法，从合成视频中提取几何信息（2D末端执行器路径点）监督视觉表征，仅用真实演示训练动作头，在真实机器人任务中优于伪动作基线。
摘要：
视觉-语言-动作（VLA）模型需要大规模视频-动作对，但真实遥操作数据仍然稀缺。虽然生成的机器人视频提供了一种可扩展的替代方案，现有方法通过从合成像素中恢复伪动作将其视为真实机器人数据。我们认为从生成视觉中推导低级控制是一种不匹配的抽象。视频仅捕捉几何：表示任务“在哪里”的空间轨迹。真实演示捕捉控制：表示“如何做”的精确电机指令。从人到机器人的视频生成不均等地保留了这些信息：可见几何在生成过程中幸存，而底层控制信号丢失。这种**非对称保留原则**规定了一个清晰的规则：幸存几何应仅监督视觉感知，将控制留给真实演示。遵循这一原则，我们提出**GRA**（**G**eometry-guided **R**epresentation **A**lignment），它提取几何内容作为未来的2D末端执行器路径点，通过姿态估计、重定向、仿真和标定投影从源人类视频中计算得到，并通过辅助2D头将其路由到VLA视觉骨干。动作头仅在真实演示上训练。在微调过程中，路径点损失作为一个**空间表征锚点**持续存在，防止骨干丢失其几何基础。在真实机器人任务中，GRA在匹配数据预算下优于伪动作基线，并缩小了与使用更多真实演示训练的策略之间的差距，表明正确路由的几何比恢复的动作更可靠地将生成视频与机器人策略连接起来。
英文摘要：
Vision-Language-Action (VLA) models require large-scale video-action pairs, yet real teleoperation remains scarce. While generated robot videos offer a scalable alternative, existing methods treat them as real robot data by recovering pseudo-actions from synthesized pixels. We argue that deriving low-level control from generated visuals is a mismatched abstraction. A video captures only \emph{geometry}: the spatial trajectory representing the \emph{where} of a task. A real demonstration captures \emph{control}: the exact motor commands representing the \emph{how}. Human-to-robot video generation preserves these unequally: the visible geometry survives the generation process, while the underlying control signals are lost. This \textbf{Asymmetric Preservation Principle} dictates a clean rule: this surviving geometry should solely supervise visual perception, leaving control to real demonstrations. Following this principle, we propose \textbf{GRA} (\textbf{G}eometry-guided \textbf{R}epresentation \textbf{A}lignment), which extracts the geometric content as future 2D end-effector waypoints, computed from the source human video through pose estimation, retargeting, simulation, and calibrated projection, and routes them to the VLA vision backbone via an auxiliary 2D head. The action head is trained on real demonstrations only. During fine-tuning, the waypoint loss persists as a \textbf{spatial representation anchor} that prevents the backbone from losing its geometric grounding. On real-robot tasks, GRA outperforms pseudo-action baselines under matched data budgets and narrows the gap to policies trained with substantially more real demonstrations, suggesting that correctly routed geometry bridges generated videos to robot policies more reliably than recovered actions.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
