Application of machine learning for advanced material prediction and design

• Published in: EcoMat (Beijing, China) Vol. 4; no. 4
• Main Authors: Chan, Cheuk Hei, Sun, Mingzi, Huang, Bolong
• Format: Journal Article
• Language: English
• Published: Hoboken, USA John Wiley & Sons, Inc 01.07.2022; Wiley
• Subjects: Accuracy; Algorithms; Artificial intelligence; Computer applications; Data collection; Data mining; Datasets; Efficiency; Experiments; Learning algorithms; Machine learning; Material properties; materials science; new structure design; Optimization; Pattern recognition systems; Predictions; property predictions; Science; Simulation
• ISSN: 2567-3173; 2567-3173
• DOI: 10.1002/eom2.12194

In material science, traditional experimental and computational approaches require investing enormous time and resources, and the experimental conditions limit the experiments. Sometimes, traditional approaches may not yield satisfactory results for the desired purpose. Therefore, it is essential to develop a new approach to accelerate experimental progress and avoid unnecessary wasting of time and resources. As a data‐driven method, machine learning provides reliable and accurate performance to solve problems in material science. This review first outlines the fundamental information of machine learning. It continues with the research concerning the prediction of various properties of materials by machine learning. Then it discusses the methods for the discovery of new materials and the prediction of their structural information. Finally, we summarize other applications of machine learning in material science. This review will be beneficial for future application of machine learning in more material science research. Due to the low computational cost and short development cycle, machine learning has been widely used in material detection, material analysis, and material design. In this review, the background of machine learning has been systematically introduced and their applications in materials science from both macroscopic and microscopic perspectives has been summarized.