Computationally accelerated experimental materials characterization -- drawing inspiration from high-throughput simulation workflows

Computational materials science is increasingly benefitting from data management, automation, and algorithm-based decision-making in controlling simulations. Experimental materials science is also undergoing a change and increasingly more `machine learning' is incorporated in materials discovery campaigns. The obvious benefits include automation, reproducibility, data provenance, and reusability of managed data, however, is not widely available. We demonstrate an implementation of a Gaussian Process Regression directly controlling an experimental measurement device in pyiron, a framework designed for high-throughput simulations, as a first step to increasingly combine experimental and simulated data in one framework. With data from both in the same framework, a heretofore untapped and much-needed potential for the acceleration of materials characterization and materials discovery campaigns becomes available.
17 pages, 4 figures, preprint