AI MIT and Stanford Researchers Developed a Machine-Learning Technique that can Efficiently Learn to Control a Robot, Leading to Better Performance with Less Data
