Stanford researchers have developed a machine learning system to create algorithms that can help engineers analyze the properties of thousands of materials to help narrow the list of ideal materials for their application. One test application of a computer generated short-list proved twice as accurate and over a thousand times faster than the human team.