Researchers used machine learning to perform accelerated simulations of the physico-chemical properties of molten salt FLiNaK. Their framework can help characterize and screen other molten salts and determine which are ideal to use in an advanced nuclear reactor.
