Agnes Harness: A Framework for Natural Language Task Execution and Multi-Device Coordination
一句话指挥机器人组团打工,还能越干越聪明?这波物理AI有点夯
Agnes Harness is a system that translates natural language commands into coordinated multi-device tasks, enabling robots and devices to perform complex actions autonomously. It integrates a full-modal model with a task execution framework, allowing for real-time decision-making, learning from experience, and deployment across various physical and simulated environments. The system supports both simulated and real-world applications, including robotics, gaming, and enterprise workflows.
Source: Machine Heart — Robotics on WeChat · Read original article ↗
English translation unavailable. Choose the original article to read it.
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
技术报告显示,在针对游戏内置 Bot 的测试中,1v1 最佳策略的胜率从约 3% 提升至 90%;另一组独立的 3v3 实验经过三轮 RSI,胜率达到约 60%。
Open source S10
而现场大屏显示, 5v5 实验经过五轮 RSI,胜率已经达到 100%。
Open source S11
Source:Machine Heart — Robotics on WeChat · mp.weixin.qq.com