Learning from the Harvard Clean Energy Project: The Use of Neural Networks to Accelerate Materials Discovery

• Published in: Advanced functional materials 2015-11, Vol.25 (41), p.6495-6502
• Main Authors: Pyzer-Knapp, Edward O., Li, Kewei, Aspuru-Guzik, Alan
• Format: Article
• Language: English
• Subjects: Algorithms; big data; Clean energy; Computation; Libraries; machine learning; materials genomes; Mathematical analysis; Neural networks; organic materials screening; organic photovoltaics; Quantum chemistry; Screening
• ISSN: 1616-301X; 1616-3028
• DOI: 10.1002/adfm.201501919

Here, the employment of multilayer perceptrons, a type of artificial neural network, is proposed as part of a computational funneling procedure for high‐throughput organic materials design. Through the use of state of the art algorithms and a large amount of data extracted from the Harvard Clean Energy Project, it is demonstrated that these methods allow a great reduction in the fraction of the screening library that is actually calculated. Neural networks can reproduce the results of quantum‐chemical calculations with a large level of accuracy. The proposed approach allows to carry out large‐scale molecular screening projects with less computational time. This, in turn, allows for the exploration of increasingly large and diverse libraries. The utility of including neural networks as a highly accurate screening function is demonstrated for molecules from the Harvard Clean Energy Project. The neural network described can predict power conversion efficiencies of molecules with an error of 0.12%. By using this network as a screen for generated molecules, the scope of high‐throughput virtual screening is expanded by several orders of magnitude.