Machine learning-assisted design of high-entropy alloys with superior mechanical properties

Abstract

Most recently, high-entropy alloys (HEAs) with 5 or more elements open a new area for materials exploration with substantial mechanical properties. The large composition space and numerous structures of HEAs bring significant difficulties for phase design and determination of mechanical property. Machine learning, one of most rapidly growing scientific and technical field, meets at the intersection of computer science and materials science, and at the center of artificial intelligence. Machine learning provides the opportunity to build up the relationship between multiple physical properties and mechanical properties. The fast changes of this field call for significant practice for materials community to utilize it as a more efficient, accurate and interpretable tool. In this review, we summarize the most promising machine learning models, combined with high-throughput simulation and experimental screening, to predict and fabricate HEAs with desired superb mechanical properties.