Although the discovery and synthesis of new materials, catalysts, and functional transition metal molecules represents the foremost effort that unifies much of the chemical science community, presently characterized compounds represent a minute fraction of chemical space. This challenge is particularly daunting for the design of transition metal complexes due to the combinatorial challenges associated with their variable spin, oxidation state, and coordination number. Machine learning has emerged as a tool to reveal emergent phenomena and accelerate design of materials and catalysts in this vast space. In this talk, I will describe our development of representations uniquely tailored for transition metal complex properties and how we obtain not just accurate predictions with machine learning models but also develop new chemical insights. I will describe our next steps in developing artificial intelligence engines that not just accelerate simulation but automate the decisions that typically require advanced training to determine also when and how to carry out simulations. Time permitting, I will describe applications of these tools for advancing understanding in materials synthesis, functional spin crossover materials, and in catalysis.

Accelerating discovery in inorganic chemistry with machine learning

2019 Spring National Meeting

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