Using Machine Learning to Aid in NMR Crystallographic Refienements of Siliceous Zeolites

Maxwell Venetos – @MaxwellVenetos

Happy to share an example of using machine learning to greatly reduce the cost of NMR crystallographic refinements with little to no loss in accuracy. #GlobalnmrTC2023 pic.twitter.com/RNlRVODRmZ

— Max Venetos (@MaxwellVenetos) July 19, 2023

By combining Si-29 2D J spectra with equivariant graph neural networks, we have developed a rapid structure refinement strategy for crystalline solids. We demonstrate the use of this strategy with the refinement of siliceous zeolite ZSM-12.