Specialized machine learning force fields for materials dynamics
Abstract
Machine learning interatomic potentials (MLIPs) are transforming atomistic simulations by accessing unprecedented length and time scales. While pretrained equivariant graph neural networks achieve robust zero-shot performance for near-equilibrium properties across broad chemical spaces, their translation to complex materials dynamics remains fundamentally challenged by out-of-distribution reactive states, representation biases, and computational scaling limits. In this Review, we examine how physics-driven specialization extends the applicability of MLIPs to complex dynamical systems. We systematically evaluate the structural trade-offs in MLIP design: the undersampling of highly strained configurations, the prohibitive computational overhead of high-order message-passing architectures, and the necessity of nonlocal interactions for open and field-coupled systems. Through four demanding application contexts - electrified interfaces, compositionally fluctuating open systems, multiphase evolution, and large-scale fracture - we establish a framework for observable-specific validation. Highlighting the complementary roles of universal foundation models and task-specific potentials, we emphasize that targeted adaptations must be rigorously benchmarked against intended observables. We conclude with a roadmap for developing physically consistent, hardware-aware force fields that seamlessly connect electronic-structure accuracy to macroscopic materials phenomena.
Related
- Modelling the dynamics of the martensitic transformation in Ni2MnGa with machine-learning force-fields
- Sorption Molecular Dynamics of Point Defects in 2D Materials with a Machine Learning Potential
- Learning materials spectra across modalities and scales with structure-conditioned neural fields
- Autonomous Analysis of Thermoelectric Materials Big Data by Multi-Agentic AI and Construction of a Knowledge Base and Specialized AI
- Benchmarking machine-learning interatomic potentials using classical force fields
BibTeX
@misc{wu2026specialized,
title = {{Specialized machine learning force fields for materials dynamics}},
author = {Yue Wu and Haoyu Wan and Yuan Tian and Deng Pan and Mingwei Chen},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.12151},
eprint = {2610.12151},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.12151},
}
RIS
TY - UNPB
TI - Specialized machine learning force fields for materials dynamics
AU - Yue Wu
AU - Haoyu Wan
AU - Yuan Tian
AU - Deng Pan
AU - Mingwei Chen
PY - 2026
DA - 2026/10/09
PB - arXiv
DO - 10.48550/arxiv.2610.12151
UR - https://arxiv.org/abs/2610.12151
AB - Machine learning interatomic potentials (MLIPs) are transforming atomistic simulations by accessing unprecedented length and time scales. While pretrained equivariant graph neural networks achieve robust zero-shot performance for near-equilibrium properties across broad chemical spaces, their translation to complex materials dynamics remains fundamentally challenged by out-of-distribution reactive states, representation biases, and computational scaling limits. In this Review, we examine how physics-driven specialization extends the applicability of MLIPs to complex dynamical systems. We systematically evaluate the structural trade-offs in MLIP design: the undersampling of highly strained configurations, the prohibitive computational overhead of high-order message-passing architectures, and the necessity of nonlocal interactions for open and field-coupled systems. Through four demanding application contexts - electrified interfaces, compositionally fluctuating open systems, multiphase evolution, and large-scale fracture - we establish a framework for observable-specific validation. Highlighting the complementary roles of universal foundation models and task-specific potentials, we emphasize that targeted adaptations must be rigorously benchmarked against intended observables. We conclude with a roadmap for developing physically consistent, hardware-aware force fields that seamlessly connect electronic-structure accuracy to macroscopic materials phenomena.
ER -