LLMs help robots understand vague instructions and focus on key details

Researchers at MIT have developed a system using language models to help robots better understand what people actually want them to do. The approach uses two language models in sequence: the first clarifies vague or ambiguous instructions from users, and the second filters out irrelevant details so the robot focuses on what matters. The technique could make robots more useful in homes and factories, where instructions are often informal and contextual.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton