LAUSR

Description Machine learning narrows down the enormous search space for functional materials Researchers and engineers are constantly searching for materials with specific properties to drive the rapid development of various technologies. Because of the practically infinite combinations of materials, these searches need to be strategized. In the case of conventional alloys, they generally consist of a single principal metal element accompanied by other elements. More recently, researchers have ventured into looking for alloys with multiple principal elements (1, 2). This type of alloy, called a high-entropy alloy (HEA), greatly expands the search space of alloys for materials design. On page 78 of this issue, Rao et al. (3) present a physics-informed machine-learning approach to screen alloys with low thermal expansion coefficient within the huge iron-cobalt-nickel-chromium (Fe–Co–Ni–Cr) and iron-cobalt-nickel-chromium-copper (Fe–Co–Ni–Cr–Cu) composition space. These materials, which expand and contract very little with temperature changes, make them valuable for application on precision instruments for which high dimensional stability of the components is required.