From skeptic to believer: The power of models

Community action on FAIR data will fuel a revolution in materials research

“…Such efforts, the rapidly growing number of papers in materials science using machine learning and their citation rates (Figure 4 in References 8 and 9), and the emergence of journal publications focused on scientific data and associated metadata (e.g., Nature Scientific Data) make clear the global importance of data to materials science and engineering. [10][11][12][13][14][15][16][17] Yet despite large investments in materials science and engineering-more than $37B in 2018 by US industry alone 18most data languish in local storage systems or reports and papers. 2,12,13 In contrast, imagine being able to "Google" all materials ever synthesized or predicted, to find organized, annotated, quantitative, referenced, citable, and downloadable data for the subset of materials that have a desired combination of properties and characteristics.…”