Representations of Materials for Machine Learning

TL;DR

This paper reviews how to convert diverse materials data into numerical representations suitable for machine learning, discusses modern techniques for learning and transferring representations, and highlights unresolved questions in the field.

Contribution

It provides a comprehensive overview of strategies for constructing and learning material representations for machine learning applications.

Findings

Various representation strategies enable effective ML modeling of materials.

Modern ML techniques can learn transferable representations from data.

Identifies key unresolved questions in material representation learning.

Abstract

High-throughput data generation methods and machine learning (ML) algorithms have given rise to a new era of computational materials science by learning relationships among composition, structure, and properties and by exploiting such relations for design. However, to build these connections, materials data must be translated into a numerical form, called a representation, that can be processed by a machine learning model. Datasets in materials science vary in format (ranging from images to spectra), size, and fidelity. Predictive models vary in scope and property of interests. Here, we review context-dependent strategies for constructing representations that enable the use of materials as inputs or outputs of machine learning models. Furthermore, we discuss how modern ML techniques can learn representations from data and transfer chemical and physical information between tasks. Finally, we outline high-impact questions that have not been fully resolved and thus, require further investigation.