Mapping Chemical Structure–Glass Transition Temperature Relationship through Artificial Intelligence

• DOI: 10.1021/acs.macromol.0c02594
• Publication Date: 2021-02-05T19:58:55Z
• Authors: Luis A. Miccio; Gustavo A. Schwartz
• Keywords: Chemical space; Mathematical structure; Chemical structure

Artificial neural networks (ANNs) have been successfully used in the past to predict different properties of polymers based on their chemical structure and localize quantify intramonomer contributions these properties. In this work, we propose move forward order use mathematical framework ANN for embedding monomers into a high-dimensional abstract space. This approach allows us not only accurately glass transition temperature (Tg) but, even more important, also encode as m-dimensional vectors For aim, employed fully connected network trained with set than 200 atactic acrylates that provide coordinates vectorized structures These data points were then treated hierarchical nonparametric clusterization method automatically group similar clusters alike projected human-readable three-dimensional space using principal component analysis. deal if they entities therefore perform quantitative operations, so far hardly imaginable, being essential both design new materials understanding structure–property relationships.