Show-Harness：仅靠 VLM 智能体即可操控机器人
原标题：Paper page - Show-Harness: Just a VLM Agent Can Play Robots
AI 摘要
该论文提出 Show-Harness，一种通过离散语义动作单元将视觉语言模型（VLM）与机器人控制连接起来的具身接口，由特定于本体的解释器将语义动作确定性地映射为本地机器人动作。Show-Harness 证明了两点可行性：直接解锁闭源前沿 VLM 实现零样本机器人控制，以及仅用少量 GPU 小时微调小规模开源 VLM 实现低成本部署。作者还开发了 GUMI（GUI 操作接口），将同一语义动作空间扩展到基于 GUI 的演示收集，使人类和智能体无需专用遥操作硬件即可跨本体操控机器人。实验显示，配备 Show-Harness 的 VLM 智能体在任务、本体和环境间泛化稳健，优于代表性的智能体和 VLA 范式。
正文节选
Show-Harness: Just a VLM Agent Can Play Robots Abstract Show-Harness links vision-language models to robot control via discrete semantic actions interpreted by embodiment-specific modules, enabling zero-shot and efficient fine-tuned deployment across robots and GUIs. Foundation vision-language models (VLMs) exhibit broad intelligence about the world, yet translating this intelligence into robot control remains challenging. We present Show-Harness, an Embodied Harness that enables VLMs to "play"