Abstract:

Lightweight alloys, primarily based on magnesium, aluminum, or titanium, are characterized by their low density and high strength. They are widely used in industries operating under extreme conditions, such as aerospace, petrochemical, automotive, marine, as well as military and defense. Traditional design methods for these alloys often suffer from drawbacks such as an over-reliance on empirical intuition, long development cycles, and high costs. While more advanced approaches like integrated computational materials engineering (ICME) offer improvements, they frequently require substantial computational resources and may lack accuracy. Advances in computing power and algorithms have positioned artificial intelligence (AI) as a transformative tool in light alloy design, leveraging its strong data processing capabilities and reliable precision. This article first outlines the fundamental workflow and algorithms of data-driven methodologies. Furthermore, this review highlights key applications of AI in lightweight alloy design, focusing on property prediction and inverse design for conventional mechanical properties, corrosion resistance, high-temperature performance and other critical characteristics. Finally, it addresses current challenges, including incomplete inverse design frameworks, limited extrapolation capability, and difficulties in integrating domain knowledge. The concept of data scarcity is introduced, emphasizing the critical need for alignment between data and algorithms, and prospective research directions are discussed.

Key words: Artificial neural networks, Magnesium alloy, Aluminum alloy, Mechanical property, Phase diagram