Machine learning assisted canonical sampling (MLACS)

The acceleration of material property calculations while maintaining ab initio accuracy (1 meV/atom) is one of the major challenges in computational physics. In this paper, we introduce a Python package enhancing the computation of (finite temperature) material properties at the ab initio level using machine learning interatomic potentials (MLIP). The Machine Learning Assisted Canonical Sampling (MLACS) method, grounded in a self-consistent variational approach, iteratively trains a MLIP using an active learning strategy in order to significantly reduce the computational cost of ab initio simulations. MLACS offers a modular and user-friendly interface that seamlessly integrates Density Functional Theory (DFT) codes, MLIP potentials, and molecular dynamics packages, enabling a wide range of applications, while maintaining a near-DFT accuracy. These include sampling the canonical ensemble of a system, performing free energy calculations, transition path sampling, and geometry optimization, all by utilizing surrogate MLIP potentials, in place of ab initio calculations. This paper provides a comprehensive overview of the theoretical foundations and implementation of the MLACS method. We also demonstrate its accuracy and efficiency through various examples, showcasing the capabilities of the MLACS package. Program summary: Program title: MLACS CPC Library link to program files: https://doi.org/10.17632/vtfzjnc6cr.1 Licensing provisions: GNU General Public License, version 3 Programming language: Python Nature of problem: Numerous material properties, whether related to the ground state or finite temperature thermodynamic quantities, cannot be deduced from classical simulations and require accurate but highly demanding ab initio calculations. Enhancing the efficiency of these simulations while preserving a near-ab initio accuracy is one of the biggest challenges in modern computational physics. Solution method: The emergence of MLIP potentials enables us to tackle this issue. The method implemented in MLACS allows for the acceleration of ab initio calculations by training a MLIP potential on the fly. At the end of the simulation, MLACS produces an optimal local surrogate potential, a database that includes a sample of representative atomic configurations with their statistical weights, as well as information on convergence control and thermodynamic quantities. Additional comments: The seminal version is defined in [1]. The new version [2], MLACS v1.0.2, works on various architectures and includes several new features. References: [1] A. Castellano, F. Bottin, J. Bouchet, A. Levitt, G. Stoltz, Ab initio canonical sampling based on variational inference, Phys. Rev. B 106 (2022) L161110. [2] MLACS github repository, first production version v1.0.2 (2024). https://github.com/mlacs-developers/mlacs/tree/main