Litcius/Paper detail

MP-ALOE: an r2SCAN dataset for universal machine learning interatomic potentials

Matthew C. Kuner, Aaron D. Kaplan, Kristin A. Persson, Mark Asta, D. C. Chrzan

2025npj Computational Materials9 citationsDOIOpen Access PDF

Abstract

Abstract We present MP-ALOE, a dataset of nearly 1 million DFT calculations using the accurate r 2 SCAN 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.

Topics & Concepts

Stability (learning theory)SoundnessBenchmark (surveying)Machine learningArtificial intelligenceComputer scienceInteratomic potentialSeries (stratigraphy)Extreme learning machineAlgorithmStatistical physicsActive learning (machine learning)Time seriesTraining setOnline machine learningComputational learning theoryData miningMolecular dynamicsMachine Learning in Materials ScienceModel Reduction and Neural NetworksCrystallography and molecular interactions
MP-ALOE: an r2SCAN dataset for universal machine learning interatomic potentials | Litcius