EXTENDED ABSTRACT: Metamaterials can achieve performance beyond the inherent properties of their constituent materialsthrough structural design. However, metamaterial design faces challenges such as vast design space, heavy reliance on empirical methods, small search spaces, extensive human intervention, and slow computational speed. The development of artificial intelligence technologies has provided new powerful tools for computing designs in metamaterials. This report briefly introduces the software platforms developed by our research group from three aspects: sample space generation, computational acceleration by surrogate models, and optimization schemes. It also presents application examples covering two directions—small AI models and large language AI models.
Keywords：Metamaterial Design; Surrogate Modeling; Intelligent Computing; Application of Large Language Models
SUN, Sheng is graduated from The Hong Kong University of Science and Technology (HKUST), completed his postdoctoral research at HKUST. He is currently a professor and Ph.D. supervisor at the Institute for Materials Genomics Engineering at Shanghai University. His long-term research focuses on computational materials science, computational mechanics, material informatics, and mechanics informatics. He has led three National Natural Science Foundation of China (NSFC) projects as principal investigator, one major project under the National Key R&D Program of China, and participated in multiple NSFC projects and provincial-level programs. Over the past five years, he has published more than 30 academic papers with over 1000 citations added; co-authored a book ranked second among authors (one); obtained five invention patents and registered software copyrights for more than twenty items.