A software package for generating robust and accurate potentials using the moment tensor potential framework

We present the Plan for Robust and Accurate Potentials (PRAPs), a software package for training and using moment tensor potentials (MTPs) in concert with the Machine Learned Interatomic Potentials (MLIP) software package. PRAPs provides an automated workflow to train MTPs using active learning procedures, and a variety of utilities to ease and improve workflows when utilizing the MLIP software. PRAPs was originally developed in the context of crystal structure prediction, in which one calculates convex hulls and predicts low energy metastable and thermodynamically stable structures, but the potentials PRAPs develops are not limited to such applications. PRAPs produces two potentials, one capable of rough estimates of the energies, forces and stresses of almost any chemical structure in the specified compositional space – the Robust Potential – and a second potential intended to provide more accurate descriptions of ground state and metastable structures – the Accurate Potential. We also present a Python library, mliputils, designed to assist users in working with the chemical structural files used by the MLIP package. PROGRAM SUMMARY: Program Title: The Plan for Robust and Accurate Potentials (PRAPs) CPC Library link to program files: (to be added by Technical Editor) Developer's repository link: https://github.com/Dryctarth/PRAPs.git Code Ocean capsule: (to be added by Technical Editor) Licensing provisions(please choose one): BSD 3-clause Programming language: Bash, Python Supplementary material: User manual Nature of problem: Keeping track of all the steps involved in training moment tensor potentials across several systems has proven to be a challenge in need of project management. For every large step, like training, there are several small, mundane commands that need to be handled, and these must all be repeated identically across any chemical system users may care about (while tracking variations). Finally, communication must be made between the AFLOW, MLIP, and VASP programs. Solution method: The PRAPs package incorporates a degree of automation, handling the different job submissions and tasks needed to train multiple moment tensor potentials, file management, identifying and removing unphysical chemical structures, and performing some analytical tasks. The package also includes some simple utility functions to allow users to better read, write, and manipulate MLIP's chemical structure file format. Additional comments including restrictions and unusual features: Requires a local installation of Automatic FLOW (AFLOW) v3.10+, the Vienna ab initio Software Package (VASP) v5+, and the Machine Learning for Interatomic Potentials (MLIP) v2+ program packages.

Duke Scholars

Author Stefano Curtarolo Thomas Lord Department of Mechanical Engineering and Materia ...

Author Simon Divilov Thomas Lord Department of Mechanical Engineering and Materia ...