Benchmarking, Visualization and Hyperparameter Optimization of UF₃ to Enhance Molecular Dynamics Simulations

Ab initio methods offer great promise for materials design, but they come with a hefty computational cost. Recent advances with machine learning potentials (MLPs) have revolutionized molecular dynamic simulations by providing high accuracies approaching those of ab initio models but at much reduced computational cost. Our study evaluates the ultra-fast potential (UF₃), employing linear regression with cubic B-spline basis for learning effective two- and three-body potentials. On benchmarking, UF₃ displays comparable precision to established models like GAP, MTP, NNP(Behler Parrinello), and qSNAP MLPs, yet is significantly faster by two to three orders of magnitude. A distinct feature of UF₃ is its capability to render visual representations of learned two- and three-body potentials, shedding light on potential gaps in the learning model. In refining UF₃'s performance, a comprehensive sweep of the hyperparameter space was undertaken, emphasizing finer granularity in zones indicative of optimal performance. This endeavor aims to provide insights into the smoothness of the UF₃ hyperparameter space, and offer users a foundational default set of hyperparameters as a starting point for optimization. While our current optimizations are concentrated on energies and forces, we are primed to broaden UF₃’s evaluation spectrum, focusing on its applicability in Molecular Dynamics simulations. The outcome of these investigations will not only enhance the predictability and usability of UF₃ but also pave the way for its broader applications in advanced materials discovery and simulations.