Abstract
The discovery of advanced materials is the cornerstone of human technologicaldevelopment and progress. The structures of materials and their correspondingproperties are essentially the result of a complex interplay of multipledegrees of freedom such as lattice, charge, spin, symmetry, and topology. Thisposes significant challenges for the inverse design methods of materials.Humans have long explored new materials through a large number of experimentsand proposed corresponding theoretical systems to predict new materialproperties and structures. With the improvement of computational power,researchers have gradually developed various electronic structure calculationmethods, such as the density functional theory and high-throughputcomputational methods. Recently, the rapid development of artificialintelligence technology in the field of computer science has enabled theeffective characterization of the implicit association between materialproperties and structures, thus opening up an efficient paradigm for theinverse design of functional materials. A significant progress has been made ininverse design of materials based on generative and discriminative models,attracting widespread attention from researchers. Considering this rapidtechnological progress, in this survey, we look back on the latest advancementsin AI-driven inverse design of materials by introducing the background, keyfindings, and mainstream technological development routes. In addition, wesummarize the remaining issues for future directions. This survey provides thelatest overview of AI-driven inverse design of materials, which can serve as auseful resource for researchers.