Deep dive into machine learning density functional theory for materials science and chemistry

With the growth of computational resources, the scope of electronic structure\nsimulations has increased greatly. Artificial intelligence and robust data\nanalysis hold the promise to accelerate large-scale simulations and their\nanalysis to hitherto unattainable scales. Machine learning is a rapidly growing\nfield for the processing of such complex datasets. It has recently gained\ntraction in the domain of electronic structure simulations, where density\nfunctional theory takes the prominent role of the most widely used electronic\nstructure method. Thus, DFT calculations represent one of the largest loads on\nacademic high-performance computing systems across the world. Accelerating\nthese with machine learning can reduce the resources required and enables\nsimulations of larger systems. Hence, the combination of density functional\ntheory and machine learning has the potential to rapidly advance electronic\nstructure applications such as in-silico materials discovery and the search for\nnew chemical reaction pathways. We provide the theoretical background of both\ndensity functional theory and machine learning on a generally accessible level.\nThis serves as the basis of our comprehensive review including research\narticles up to December 2020 in chemistry and materials science that employ\nmachine-learning techniques. In our analysis, we categorize the body of\nresearch into main threads and extract impactful results. We conclude our\nreview with an outlook on exciting research directions in terms of a citation\nanalysis.\n