Show-Harness：通过轻量语义接口让通用视觉语言模型直接操控机器人

Translation Summary

Alright, let's get this done. I need to deliver a clean, professional translation of this robotics abstract into Simplified Chinese. Here's how I'm thinking about it:

First, understand the task. It's a straight-up translation of an abstract about robot control using VLMs (vision-language models). The target language is Simplified Chinese. Keep it professional, and no extra commentary. Paragraph breaks and URLs are to be preserved if any appear (there are none).

Okay, let's go sentence by sentence. I'll translate the meaning accurately and professionally, making sure the terminology is correct. I have to select the best terminology.

• "Foundation vision-language models (VLMs) exhibit broad intelligence about the world, yet translating this intelligence into robot control remains challenging." - The standard translation would be "视觉语言基座模型（VLM）". But there are other options like "基础视觉语言模型". I think "视觉语言基座模型（VLM）" is standard and concise. The rest should read, "视觉语言基座模型（VLM）展现出对世界的广泛认知，但将这种智能转化为机器人控制能力仍然极具挑战性。"

• "We present Show-Harness, an Embodied Harness that enables VLMs to "play" robots through a compact semantic interface linking intent to action." - I'm proposing that we should translate "Show-Harness" as "我们提出了 Show-Harness". Then, "Embodied Harness" translates to "具身驾驭框架". The phrase "play robots" presents a slight challenge because of the quotation marks, and the inherent gaming metaphor. The closest and best translation is "操控" with the quotation marks, which becomes: "我们提出了 Show-Harness，一个通过连接意图与动作的紧凑语义接口使 VLM 能够“操控”机器人的具身驾驭框架（Embodied Harness）"

• "Show-Harness exposes discrete semantic action units that VLMs can naturally reason over, while embodiment-specific interpreters deterministically ground them into local robot actions, keeping the VLM directly responsible for fine-grained physical decisions." - "exposes discrete semantic action units" gets translated to "提供了 VLM 可以自然进行推理的离散语义动作单元". The phrase "embodiment-specific interpreters deterministically ground them into local robot actions" is handled with "针对特定本体的解释器则确定性地将其映射为机器人的局部动作". Fine-grained is best as "细粒度的". Putting it together becomes: "Show-Harness 提供了 VLM 可以自然进行推理的离散语义动作单元，而针对特定本体的解释器则确定性地将其映射为机器人的局部动作，从而使 VLM 直接负责细粒度的物理决策。"

• "Through the same interface, Show-Harness demonstrates the feasibility of (1) directly unlocking closed-source frontier VLMs for zero-shot robot control, and (2) adapting small-scale open-source VLMs for low-cost deployment with just a few GPU-hours of fine-tuning." - This will translate to: "通过同一接口，Show-Harness 证明了以下方案的可行性：(1) 直接解锁闭源前沿 VLM 以实现零样本机器人控制；(2) 仅需数个 GPU 小时的微调，即可适配小型开源 VLM 以进行低成本部署。" I'm being mindful of terms like "frontier VLMs" (前沿 VLM) and "zero-shot robot control" (零样本机器人控制) as standard terminology.

• "We further develop GUMI (GUI Manipulation Interface), which extends the same semantic action space to GUI-based demonstration collection, allowing humans and agents to "play" robots across embodiments without specialized teleoperation hardware." - "GUMI (GUI Manipulation Interface)" will be "GUMI（图形用户界面操作接口，GUI Manipulation Interface）". The crucial phrase "allowing humans and agents to "play" robots" is again "使人类和智能体无需专用的遥操作硬件即可跨本体“操控”机器人".

• "Extensive experiments show that Show-Harness-equipped VLM agents generalize robustly across tasks, embodiments, and environments, outperforming representative agentic and VLA paradigms." - "Show-Harness-equipped VLM agents" translates directly to "配备 Show-Harness 的 VLM 智能体". Then it continues to: "大量实验表明，配备 Show-Harness 的 VLM 智能体能够在各种任务、本体和环境之间稳健泛化，表现优于代表性的智能体（Agent）范式和 VLA 范式。"

• "These results suggest that the right interface can unlock substantial embodied capability from foundation VLMs, without requiring additional model capacity or costly embodiment-specific pretraining." - The final sentence

视觉语言基座模型（VLM）展现出了对世界的广泛认知，但将这种智能转化为机器人控制能力仍然极具挑战性。我们提出了 Show-Harness，这是一个具身驾驭框架（Embodied Harness），通过连接意图与动作的紧凑语义接口，使 VLM 能够“操控”机器人。Show-Harness 提供了 VLM 可以自然进行推理的离散语义动作单元，而针对特定本体的解释器则确定性地将其映射为机器人的局部动作，从而使 VLM 直接负责细粒度的物理决策。通过同一接口，Show-Harness 证明了以下方案的可行性：(1) 直接解锁闭源前沿 VLM 以实现零样本机器人控制；(2) 仅需数个 GPU 小时的微调，即可适配小型开源 VLM 以进行低成本部署。我们还进一步开发了 GUMI（GUI Manipulation Interface，图形界面操作接口），将相同的语义动作空间拓展到基于 GUI 的示教数据收集，使人类和智能体无需专用的遥操作硬件即可跨本体“操控”机器人。大量实验表明，配备 Show-Harness 的 VLM 智能体能够在不同任务、本体和环境之间实现稳健的泛化，表现优于代表性的智能体（Agent）范式和 VLA（视觉-语言-动作）范式。这些结果表明，一个合适的接口便能从基座 VLM 中激发出巨大的具身能力，而无需增加额外的模型容量或进行昂贵的针对特定本体的预训练。