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.
@article{arxiv.2507.05559,
title = {MP-ALOE: An r2SCAN dataset for universal machine learning interatomic potentials},
author = {Matthew C. Kuner and Aaron D. Kaplan and Kristin A. Persson and Mark Asta and Daryl C. Chrzan},
journal= {arXiv preprint arXiv:2507.05559},
year = {2025}
}
Comments
To download the dataset and associated files, see https://doi.org/10.6084/m9.figshare.29452190