In less than a century, polymer consumption has become significant in everyday life and high-technology. Polymer informatics efforts are underway more vigorously than ever before to efficiently and effectively develop, design, and discover new polymers that meet specific application needs. So far, these data-driven efforts have focused largely on homopolymers. Here, we address the property prediction challenge for copolymers, ushering in copolymer informatics beyond homopolymers. Advanced polymer fingerprinting and deep-learning schemes that incorporate multi-task learning and meta-learning are proposed to solve the emerging problems in copolymer informatics. A large data set containing over 18,000 data points of glass transition, melting, and degradation temperatures of homopolymers and copolymers of up to two monomers is used to demonstrate the copolymer prediction efficacy. Our developed models are accurate, fast, flexible, and scalable to other copolymer properties, and copolymers composed of more than two monomers when suitable data become available. Production-ready models are deployed at https://polymergenome.org/.

Copolymer informatics with multi-task deep neural networks

ACS Fall 2021

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.