AI-Assisted Design of Advanced Polymeric Materials: Challenges and Solutions

Leveraging the artificial intelligence (AI) paradigm for the innovation of advanced polymeric materials is emerging as an exciting frontier in the cross-disciplinary material community. The big data analysis and prediction capabilities of AI accelerate the discovery and development of polymers with tailored properties. The concept of AI-assisted polymer design involves a paradigm shift from traditional trial-and-error experimentation to a more efficient, data-driven methodology. To date, various AI algorithms have been developed for the structural design and composition optimization of advanced polymeric materials. It mainly involves three steps, including material information digitization and database construction, establishment of AI prediction models, and AI-based design and optimization. However, the intrinsic data characteristics and intricate relationships between multiscale structure and polymer property pose tough data and modeling challenges. Advanced approaches are proposed to address these challenges, such as digitalizing polymer multiscale structures, implementing multitask and multimodal learning methods, and inverse design and automatic optimization. This review aims to provide an insightful overview of the current state of AI-assisted polymer design, highlighting the polymer characteristics and corresponding challenges, achievable strategies, and development direction.