MP-ALOE: An r2SCAN dataset for universal machine learning interatomic potentials
This provides a high-quality public dataset for materials science researchers to develop universal machine learning interatomic potentials, though it represents an incremental improvement through systematic benchmarking of existing methods.
The researchers tackled the challenge of creating accurate machine learning interatomic potentials by developing MP-ALOE, a dataset of nearly 1 million DFT calculations using the r2SCAN method covering 89 elements, which demonstrated strong performance across multiple benchmarks including thermochemical properties and molecular dynamic stability.
We present MP-ALOE, a dataset of nearly 1 million DFT calculations using the accurate r2SCAN meta-generalized gradient approximation. Covering 89 elements, MP-ALOE was created using active learning and primarily consists of off-equilibrium structures. We benchmark a machine learning interatomic potential trained on MP-ALOE, and evaluate its performance on a series of benchmarks, including predicting the thermochemical properties of equilibrium structures; predicting forces of far-from-equilibrium structures; maintaining physical soundness under static extreme deformations; and molecular dynamic stability under extreme temperatures and pressures. MP-ALOE shows strong performance on all of these benchmarks, and is made public for the broader community to utilize.