A smarter approach to designing metamaterials AI-driven framework creates defect-tolerant materials with complex functionality by Marni Ellery

A team led by researchers in UC Berkeley’s Department of Materials Science and Engineering (led by Prof. Rayne Zheng and postdoc Marco Maurizi) has developed a new AI-driven method for the fully automated design of engineering materials with complex physical behaviors. The approach enables the inverse design of metamaterials for advanced functionalities such as energy absorption and sound cancellation. Notably, the AI model also accounts for processing-related defects, offering more robust and manufacturable solutions. This work is also featured on the front cover of Nature Machine Intelligence.