Search papers, labs, and topics across Lattice.
This paper introduces a semiparametric interatomic potential that generalizes the Abell-Tersoff bond-order potential by incorporating explicit high-order many-body correlations and a chemically informed functional form. The model is trained on diverse datasets, including silicon, carbon, water, and small molecules, achieving interpolation accuracy on par with existing machine learning interatomic potential (MLIP) models while significantly enhancing extrapolation capabilities to out-of-distribution configurations, particularly under extreme conditions. The findings highlight the importance of physically motivated constraints in improving the reliability of interatomic potentials for atomistic simulations beyond their training domains.
Extrapolation of interatomic potentials can be dramatically improved by integrating high-order many-body correlations, outperforming existing models in extreme conditions.
Machine-learning interatomic potentials enable accurate and efficient atomistic simulations, but their reliability for out-of-distribution configurations far beyond the training domain remains a significant challenge. Here, to tackle this challenge, we introduce a semiparametric interatomic potential based on a generalization of the Abell--Tersoff bond-order potential, with a chemically informed functional form and explicit high-order many-body correlations. The model is trained and evaluated on various datasets, including silicon, carbon, water, and small molecules, achieving interpolation accuracy comparable to that of existing MLIP models while demonstrating better extrapolation to unseen configurations, including those under high-pressure and high-temperature conditions. These results demonstrate that physically motivated constraints incorporated into the functional form can improve the extrapolation behavior of interatomic potentials, providing insights into the inductive biases that can be used to control extrapolation in machine-learning interatomic potentials.
