Machine Learning in Materials Science

This chapter addresses the role that data-driven approaches, especially machine learning methods, are expected to play in materials research in the immediate future. Machine learning, an important part of artificial intelligence, has made monumental contributions to areas outside materials science, ranging from commerce to gaming to search engines to drug design. Machine learning algorithms can be separated into two broad classes: supervised and unsupervised learning. The chapter first provides the necessary mathematical background to allow a materials researcher entering this field to use these methods most effectively. It then presents an assortment of examples of recent machine learning applications within materials science. The chapter also discusses a range of emerging efforts, including high-throughput phase diagram and crystal structure determination methods, accelerated prediction of materials properties, development of interatomic potentials and functionals for accelerating materials simulations, and efficient and low-cost methods for materials process control.

# Machine Learning in Materials Science # Materials Science # Advanced Chemical Physics Studies # Physics and Astronomy # Computational Drug Discovery Methods # Computer Science

Tim Mueller, Aaron Gilad Kusne, Rampi Ramprasad (2016). "Machine Learning in Materials Science". Reviews in computational chemistry. https://doi.org/10.1002/9781119148739.ch4