MIT Develops AI Method for Robots to Interpret Ambiguous Commands

Researchers at MIT have introduced a new method, 'Masked Inverse Reinforcement Learning,' which utilizes large language models to help robots understand unclear human instructions. This approach allows robots to execute tasks more effectively with less training data. The development aims to reduce the human effort required for robot interaction.