Co-Development of Experiment Automation and Data Science for Accelerated Materials Discovery

As materials discovery efforts increasingly expand in high-order composition spaces and/or far-from-equilibrium syntheses, efficient exploration requires both automation of experiments and advancement of data science to interpret and plan experiments. We will discuss the automation of experiments for research modalities including exploratory experimentation, validation of theory predictions, and incorporation of artificial intelligence for accelerating both experiment design and data interpretation. In the area of solar fuels materials discovery, these modalities have been prolific for discovery of (photo)electrocatalysts. Furthermore, the interplay of experiment, theory, and data science has inspired the development of new artificial intelligence methods for materials research, including prediction of electronic structure to accelerate computational screening (Mat2Spec) as well as the integration of physical constrains in deep learning models, which enables super-human analytic capabilities for inference of phases in x-ray diffraction patterns via Deep Reasoning Networks (DRNets).