跳到正文
Seven Minute Feed·· 10 天前AI 评分38

MIT and NYU researchers built FLEX, a feature-editing pipeline that uses an LLM to help robots read meaning beyond literal instructions. Told "Don't spill on the laptop!" while carrying coffee, the robot reasons that liquids and devices don't mix and steers clear of all electronics. The work will be presented at the Conference on Robot Learning this month. It shows how LLM common sense can turn one correction into a general rule for motion planning.

MIT's FLEX lets robots infer intent from natural language commands

AI 导读

MIT 与 NYU 研究人员构建 FLEX 管道,利用 LLM 使机器人能理解自然语言指令背后的潜在意图。 FLEX 通过特征编辑让机器人从“别洒到笔记本电脑上”等指令推理出液体与设备互斥,从而规避所有电子设备。该工作将于本月在机器人学习会议上展示,证明 LLM 常识可将单次纠正转化为通用运动规划规则。

来源:Seven Minute Feed · news.mit.edu