Review of machine learning-assisted multi-property design of high-entropy alloys: phase structure, mechanical, tribological, corrosion, and hydrogen storage properties

High-entropy alloys (HEAs), a novel class of multi-component alloy systems, have attracted much attention due to their unique compositions and outstanding performance advantages. However, the development of high-performance HEAs is severely hindered by their complex composition, high difficulty of design, and the time- and resource-intensive nature of traditional design methods. In recent years, the rapid development of artificial intelligence has led to the widespread adoption of machine learning (ML) as a powerful tool in HEAs research. The core objective of ML in the context of HEAs is to establish a quantitative relationship model linking alloy composition, processing, structure, and properties. This enables accurate phase structure prediction, the design of high-performance HEAs, reduction in the development cycle of new materials, and considerable cost savings in experimental efforts. Accordingly, this paper reviews recent advancements in the application of ML to HEAs research. It outlines the basic workflow, including data collection, data preprocessing, ML algorithm selection, hyperparameter optimization, model evaluation and model interpretability. Subsequently, specific ML applications are discussed in relation to phase structure, hardness, yield strength, modulus of elasticity, tribological properties, corrosion resistance, and hydrogen storage properties. Finally, the paper identifies key challenges in current ML applications for HEAs and offers insights into future research directions.