Abstract: Machine learning (ML) is gaining popularity as a tool for materials scientists to accelerate computation, automate data analysis, and predict materials properties. Here, I present some ways forward that myself and members of the Energy & Materials team at Toyota Research Institute have used to accelerate the design and discovery of new functional materials. Case studies will draw from studies of predicting synthesizability using databases of thermochemical data, the development of new architectures and representations for working with device-level data, and the role of first-principles methods in the process. We present several promising directions for future research: devising representations of varied experimental conditions and observations, the need to find ways to integrate machine learning into laboratory practices, and making multi-scale informatics toolkits to bridge the gaps between atoms, materials, and devices.
