The endless search for better alloys: Machine learning narrows down the enormous search space for functional materials.

The article discusses research by Rao et al. reported in the October 7th Science issue, on the use of machine learning to identify materials with low thermal expansion coefficients within the iron-cobalt-nickel-chromium and iron-cobalt-nickel-chromium-copper composition space.The researchers used an artificial neural network to train a learning algorithm to generate a large number of candidate compositions with low thermal coefficients, leading to the discovery of 17 high-entropy alloys.