The ability to predict materials properties from atomistic simulations is essential for modern materials design. Machine learning interatomic potentials (MLIPs), trained on data from electronic ...

We present a method to model interatomic interactions such as energy and forces in a computationally efficient way. The proposed model approximates the energy/forces using a linear combination of ...

Machine learning is transforming many scientific fields, including computational materials science. For about two decades, scientists have been using it to make accurate yet inexpensive calculations ...

illustrating the comprehensive zero-shot benchmark of 19 universal machine learning interatomic potentials and the dominant impact of training data composition for surface energy prediction. A ...