Stanford researchers have developed an innovative computer vision model that recognizes the real-world functions of objects, potentially allowing autonomous robots to select and use tools more ...