Universally Accurate or Specifically Inadequate? Stress‐Testing General Purpose Machine Learning Interatomic Potentials
Open Access
- 9 July 2025
- journal article
- Published by Wiley in Advanced Intelligent Discovery
- Vol. 2 (4)
Abstract
Machine learning interatomic potentials (MLIPs) have revolutionized the field of atomistic materials simulation, both due to their remarkable accuracy and their computational efficiency compared to established ab initio methods. Very recently, several general purpose MLIPs have been reported, which are broadly applicable across the periodic table. These represent a fascinating opportunity for materials discovery, provided that they are robust and transferable. In order to stress test current general purpose MLIPs, we evaluate the performance of M3GNet and MACE models in element‐substitution based structure prediction workflows for a diverse range of inorganic, crystalline materials. Importantly, these results are compared with a full density functional based workflow, shifting the focus from merely evaluating single‐point energy and force predictions of MLIPs toward an end‐to‐end perspective. We find that general purpose MLIPs are in general well‐suited to accelerate computational materials discovery and structure prediction, but also display certain systematic biases. To address these, a simple metric to quantify MLIP reliability for materials discovery is introduced. As a by‐product, we also predict novel ground state structures for 15 out of 100 analyzed compositions.Keywords
All Related Versions
Funding Information
- Max-Planck-Gesellschaft
This publication has 77 references indexed in Scilit:
- Efficiency, accuracy, and transferability of machine learning potentials: Application to dislocations and cracks in ironActa Materialia, 2024
- Machine‐Learning‐Assisted Determination of the Global Zero‐Temperature Phase Diagram of MaterialsAdvanced Materials, 2023
- Crystal structure prediction with machine learning-based element substitutionComputational Materials Science, 2022
- New Light on an Old Story: The Crystal Structure of Boron Tetrathiophosphate RevisitedZeitschrift für anorganische und allgemeine Chemie, 2019
- Property prediction of crystalline solids from composition and crystal structureAIChE Journal, 2016
- Room temperature multiferroism in CaTcO3 by interface engineeringComputational Materials Science, 2015
- Python Materials Genomics (pymatgen): A robust, open-source python library for materials analysisComputational Materials Science, 2013
- Gromov–Wasserstein Distances and the Metric Approach to Object MatchingFoundations of Computational Mathematics, 2011
- Investigation into antimony pentafluoride-based catalyst in preparing organo-fluorine compoundsJournal of Molecular Catalysis A: Chemical, 2005
- Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis setComputational Materials Science, 1996