Benchmarking Universal Machine Learning Force Fields for Crystal Structure Prediction of High-Energy Molecular Systems
Abstract
Recent developments of universal machine learning interatomic potentials (UMLIPs) offer a fast route for screening molecular crystals based on geometry relaxation and energy ranking, but their reliability across chemically diverse energetic materials remains elusive. In particular, it is unclear whether or not these UMLIPs are over-sensitive to break the desired molecular connectivity for relaxing the periodic crystals. Herein we tested the hypothesis that classical force-field pre-relaxation can provide a more suitable starting geometry for subsequent UMLIP relaxation on a large database of high energy molecular crystals. Three models (MACE, MACE-OFF and UMA) in conjunction with the General Amber Force Field (GAFF) were applied to test this hypothesis. Among them, direct MACE-OFF and UMA showed very high relaxation success and preserved the reference geometries most closely, but they still exhibit failures for some rare cases. Using GAFF pre-relaxation can systematically reduce the number of failed relaxations with lower computational costs. Our comparative failure and robustness analyses revealed distinct trade-offs among the evaluated models. Among them, MACE-OFF achieves a better compromise between potential energy surface smoothness, structural fidelity, and stress convergence, serving as a good choice to provide a reliable foundation for automated structural optimization.
Related
- Riemannian Flow Models with Reinforcement Learning for Molecular Crystal Structure Prediction
- Benchmarking Universal Machine-Learning Interatomic Potentials for Temperature-Dependent Elasticity of Binary and High-Entropy Refractory Carbides
- Physical structure as an organizing principle for universal Hamiltonian learning
- Molecular Crystal Structure Prediction from Conditional Flow on the Unit Cells
- Equivariant Deep Learning in Direct-Product Space for Electronic-Structure Hamiltonian Prediction
BibTeX
@misc{mukta2026benchmarking,
title = {{Benchmarking Universal Machine Learning Force Fields for Crystal Structure Prediction of High-Energy Molecular Systems}},
author = {Musiha Mahfuza Mukta and Osman Goni Ridwan and Romain Perriot and Qiang Zhu},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.07477},
eprint = {2609.07477},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.07477},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Benchmarking Universal Machine Learning Force Fields for Crystal Structure Prediction of High-Energy Molecular Systems
AU - Musiha Mahfuza Mukta
AU - Osman Goni Ridwan
AU - Romain Perriot
AU - Qiang Zhu
PY - 2026
DA - 2026/09/07
PB - arXiv
DO - 10.48550/arxiv.2609.07477
UR - https://arxiv.org/abs/2609.07477
AB - Recent developments of universal machine learning interatomic potentials (UMLIPs) offer a fast route for screening molecular crystals based on geometry relaxation and energy ranking, but their reliability across chemically diverse energetic materials remains elusive. In particular, it is unclear whether or not these UMLIPs are over-sensitive to break the desired molecular connectivity for relaxing the periodic crystals. Herein we tested the hypothesis that classical force-field pre-relaxation can provide a more suitable starting geometry for subsequent UMLIP relaxation on a large database of high energy molecular crystals. Three models (MACE, MACE-OFF and UMA) in conjunction with the General Amber Force Field (GAFF) were applied to test this hypothesis. Among them, direct MACE-OFF and UMA showed very high relaxation success and preserved the reference geometries most closely, but they still exhibit failures for some rare cases. Using GAFF pre-relaxation can systematically reduce the number of failed relaxations with lower computational costs. Our comparative failure and robustness analyses revealed distinct trade-offs among the evaluated models. Among them, MACE-OFF achieves a better compromise between potential energy surface smoothness, structural fidelity, and stress convergence, serving as a good choice to provide a reliable foundation for automated structural optimization.
ER -