Ji’s work on “Robust training of machine learning interatomic potentials with dimensionality reduction and stratified sampling” is now out in npj Computational Materials! Machine learning interatomic potentials (MLIPs) enable accurate simulations of materials at scales beyond that accessible by ab initio methods. In this work, we present DImensionality-Reduced Encoded Clusters with sTratified (DIRECT) sampling as an approach to select a robust training set of structures from a large and complex configuration space. By applying DIRECT sampling on the Materials Project relaxation trajectories dataset with over one million structures and 89 elements, we develop an improved materials 3-body graph network (M3GNet) universal potential that extrapolates more reliably to unseen structures. We further show that molecular dynamics (MD) simulations with the M3GNet universal potential can be used instead of expensive ab initio MD to rapidly create a large configuration space for target systems. We combined this scheme with DIRECT sampling to develop a reliable moment tensor potential for titanium hydrides without the need for iterative augmentation of training structures.
Check out this work here. If you want to use DIRECT sampling for your work, please check out our implementation available on our MAML repository on Github.