NEPMaker: Active learning of neuroevolution machine learning potential for large cells

TL;DR

NEPMaker introduces an active learning framework for neuroevolution potentials that enhances large-scale materials simulations by reducing extrapolation errors and improving model robustness.

Contribution

It develops a D-optimality-driven active learning method integrated into GPUMD, enabling scalable, on-the-fly identification of atomic environments for large systems.

Findings

Reduces extrapolation errors in large-scale simulations.

Improves model robustness and transferability.

Enables construction of reliable ML potentials for complex materials.

Abstract

Machine learning potentials (MLPs) achieve near first-principles accuracy but often fail for atomic environments outside the training distribution. Active learning can mitigate this limitation; however, its application to large-scale simulations is hindered by the prohibitive cost of labeling entire configurations. Here, we develop a D-optimality-driven active learning framework for the neuroevolution potential (NEP) implemented within the GPUMD package, named NEPMaker. Extrapolative atomic environments are identified on-the-fly and embedded into locally periodic structures, where boundary atoms are optimized to remain close to the training distribution. This strategy enables large-scale simulations to directly contribute to dataset construction, significantly reducing extrapolation errors while improving model robustness and transferability. The proposed framework provides a scalable route for constructing reliable machine learning potentials in complex materials systems, including those involving defects, interfaces, and phase transitions.