机器人学一次就会：把经验写进上下文，不写进大脑
机器人学新技能，传统做法是把经验“焊死”在神经网络参数里，换个场景就失灵。这篇提出一个反直觉的思路：让机器人像人一样，看一遍演示，然后在模拟器里自己练几遍，把练出来的心得记在“上下文”里，而不是改大脑。这样它换到真实世界、换台机器人，照样会做。50 个任务、每个只看一次演示、零梯度更新，成功率 83.4%，比最强的传统训练方法高 20 多个点。它不是你明天就能买回家的功能，但指向一个更省事的方向：以后教机器人干活，可能就像教实习生，演示一遍、让它自己琢磨，不用再为每个新场景重新训练。
📄 原文摘要(英文)
Trained action policies write manipulation skill into network weights, tying what is learned to the distribution it is learned from. In-context learning (ICL) avoids this coupling but has not matched trained policies, because a demonstration recorded in another scene carries no feedback from the world the robot currently faces. We observe that what an agent learns in context is semantic, recording task steps and object roles rather than visual appearance, so experience gathered in one world transfers to another. Based on this insight, we propose World In-Context Learning (WICL), a paradigm in which a frozen model starts from a demonstration, rehearses the task in a world such as a simulator, and writes the feedback into its context rather than its weights. We build RoboScript, a modular infrastructure that implements WICL for agentic manipulation. From a single demonstration, RoboScript learns each task through rehearsal in simulation, freezes the learned knowledge as a Task Memory, and deploys the frozen memory on new scenes with data isolation that ensures every deployment input is available on a real robot. On all 50 tasks of RoboTwin 2.0, RoboScript reaches 83.4% mean success across clean and randomized settings with one demonstration per task and no gradient updates, exceeding the strongest trained action policy by over 20 points. The method generalizes across visual distributions, improves through reflection, scales with the foundation model, and transfers to a physical dual-arm robot through cross-embodiment memory adaptation. With a world in the loop, a robot can learn a manipulation task in its context as well as it learns one in its weights. All code is available at https://github.com/ArcheOne-AI/RoboScript