L.K. Wang, A. Acovic, et al.
MRS Spring Meeting 1993
The use of generative artificial intelligence models for accurate design of new high-performance materials can potentially facilitate significant improvements to and acceleration of research workflows. However, there exists several barriers to realize meaningful generative AI for polymeric materials and catalysts capable of being translated into experimental settings. In this talk, we will highlight the importance of the development of material representation systems and their influence on the development of new foundation models for polymer property prediction and structure generation. In addition, we will discuss how generative foundation models for polymer design can be incorporated into multi-agent workflows capable of facilitating complex design tasks using natural language through a single chat interface. Finally, we will discuss efforts to integrate these generative AI systems into experimental workflows wherein the generated designs can be experimentally validated.