A new study from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) demonstrates a significant advancement in robotic manipulation. Researchers have developed a system that allows a robot to perform complex, dexterous tasks, such as hanging mugs on a rack, by learning from a single human demonstration. The system, named ‘Diffusion Policy,’ uses a type of AI model called a diffusion model to translate visual data into precise, smooth robotic actions. This approach enables the robot to generalize from the demonstration to handle variations in object placement and orientation, a challenge for previous methods that required extensive, task-specific programming or vast amounts of demonstration data. The research points toward more adaptable and efficient robots capable of learning household and industrial tasks more intuitively. Read the full article at https://technologyreview.com/2024/05/20/1093035/robot-learns-dexterous-tasks-from-one-demonstration/