Aug 2026· Advanced Intelligent Discovery· 0 citations· 19 references
Abstract
Molecular dynamics (MD) is essential for investigating atomic‐scale processes in materials and molecular systems, but the cost of high‐accuracy machine learning force field simulations still limits accessible system sizes and timescales. Here, we propose a practical model‐switching strategy for Deep Potential (DP)‐based MD simulations that alternates between independently trained DP models with different cutoff radii: a standard 6 Å model for higher accuracy and a reduced‐cutoff 4 Å model for faster inference. The method was implemented in LAMMPS/DeePMD and evaluated using solid‐phase anatase TiO
2
and liquid‐phase polyethylene glycol (PEG). For anatase TiO
2
, the 1:3 4–6 Å switching scheme preserved radial distribution function (RDF) correlations of 0.996 or higher relative to the 6 Å baseline while achieving a 1.24‐fold speedup. For PEG, the switching scheme maintained RDF correlations of 0.996 or higher with a 1.18‐fold speedup. Additional optimization using network‐size reduction and mixed‐precision inference achieved a 2.53‐fold speedup with RDF correlations of 0.975–0.988. Constant particle‐number, pressure, and temperature (NPT) simulations remained stable, whereas constant particle‐number, volume, and energy (NVE) simulations revealed system‐dependent energy‐drift behavior, particularly for aggressively optimized models. These results demonstrate that DP model switching provides a simple and practical route for accelerating structural MD simulations while highlighting the need for validation when strict energy conservation is required.
We introduce implicit machine learning force fields (I-MLFFs), which replace explicit stacks of neural network layers with self-consistent fixed-point equations. In molecular simulations, this formulation enables intermediate representations to be reused across successive timesteps, thereby warm-starting force evaluation. The resulting models effectively combine the computational footprint of a shallow, single-layer MLFF with the representational capacity and accuracy of a deep neural network. Our approach unlocks architecture-agnostic efficiency gains that are inaccessible when force prediction and trajectory integration are considered separately. We demonstrate this across three major classes of graph neural networks: invariant, equivariant Cartesian tensor, and SO(3)-equivariant spherical-tensor architectures. Each yields a two- to five-fold reduction in compute and memory footprint. Crucially, these gains are achieved while retaining full atomistic resolution and the original integration timestep, avoiding spatial or temporal coarse graining. Our contribution therefore advances the scaling frontier of quantum-mechanically faithful molecular simulation, enabling longer trajectories and larger atomistic systems within fixed GPU memory and compute budgets, and thereby opening access to new insights across biomolecular and material systems.
J. Maess, Leon Werner, J. Frank et al.· 0 citations
Low‐melting liquid metals, especially Ga–In alloys, are essential for flexible electronics, soft robotics, and adaptive thermal interfaces. Predictive atomistic modeling of their melting behavior is challenging because experiments generally provide limited microscopic insight, whereas first‐principles simulations are computationally expensive. In this work, Neuroevolution Potential (NEP) potentials for the Ga–In system are developed using density functional theory (DFT) datasets of 200–700 atom supercells generated with the local density approximation (LDA), Perdew‐Burke‐Ernzerhof (PBE), and PBE + D3 functionals. Eighteen independent NEP models with different energy/force/virial weights were trained and benchmarked against density, radial distribution function (RDF), self‐diffusion coefficient (SDC), and melting temperature (
T
m
). Radar‐metric analysis shows that PBE + D3‐based models provide the most balanced overall performance for density, liquid structure, and diffusion, whereas LDA‐based models give the best
T
m
predictions for Ga and EGaIn. We further design a local Lindemann parameter for solid–liquid identification in disordered alloys, extending the two‐phase method to compositionally complex liquid‐metal systems. Together, these results provide a benchmarked NEP model set and a scalable framework for phase‐transition and transport simulations of Ga–In liquid metals.
Machine learning interatomic potentials (MLPs) are promising for accelerating the simulation of ion transport in all‐solid‐state battery materials, but their accuracy across diverse material compositions and symmetries remains unquantified. Here, we systematically benchmark six state‐of‐the‐art MLPs, namely CHGNet, EquiformerV2 in two training variants, MatterSim, SevenNet, and MACE, on representative Li‐ and Na‐based superionic conductors. By comparing predicted atomic forces, diffusion coefficients (DCs) from MLP‐driven molecular dynamics simulations, and second‐ and third‐order interatomic force constants (IFCs) against density functional theory (DFT) and ab initio molecular dynamics (AIMD), we assess the predictive reliability of each model across these properties. EquiformerV2 models, especially the variant trained only on the OMAT dataset, exhibit the lowest force prediction errors and yield DCs in the closest agreement with AIMD. Continually, MLPs with more accurate force predictions produce more reliable diffusion metrics. However, while second‐order IFCs are reasonably captured, all models struggle to reproduce third‐order (anharmonic) IFCs, highlighting significant challenges in modeling anharmonic lattice dynamics with current MLPs. This benchmark study highlights the importance of accurately capturing atomic forces in reasonably producing ion transport properties in all‐solid‐state battery materials and provides practical guidance for universal MLP selection for ion transport simulations and future improvement and development.
Coarse‐grained (CG) molecular dynamics (MD) can greatly extend the accessible time and length scales for water, provided that the reduced model captures key structural, dynamical, and thermodynamic properties. Here, we introduce a CG machine learning potential (MLP) for water, named CGNEP‐MB‐pol, which integrates a one‐molecule to one‐bead mapping with the neuroevolution potential (NEP) framework and MB‐pol reference data, thereby aiming to alleviate, within the liquid‐water and ice‐Ih states examined here, the state dependence and limited transferability that often constrain conventional CG models. Without altering the NEP descriptor or network architecture, the model is trained on CG coordinates and labels derived from atomistic configurations through a two‐step dataset refinement strategy, combining multi‐state training based on forces, energies, and virials. CGNEP‐MB‐pol remains compatible with the GPUMD + NEP workflow while delivering a substantial speedup over its all‐atom counterpart. It reproduces mapped reference forces, radial distribution functions of liquid water and ice, and the overall temperature dependence of self‐diffusion and viscosity. Two‐phase simulations capture ice growth, melting, and solid–liquid coexistence, yielding a melting temperature of approximately 274 K. This study establishes a practical route for constructing CG‐MLPs from high‐fidelity atomistic reference models.
Ke Xu, Fuyin Yin, Yue Zhang et al.· Materials Genome Engineering...· 0 citations
Accurate descriptions of interactions between atoms are essential for molecular simulations used to study biology and support drug discovery. Existing force fields often face a trade-off between physical reliability, computational efficiency, and accuracy across unfamiliar molecules. Here we show that Residual Learning Force Field, a hybrid machine learning force field, can reduce this trade-off by combining simple physics-based descriptions of bonded interactions with learned corrections for remaining energetic effects. The two components are trained together through a three-step strategy so that each contributes complementary information. In tests covering drug-like molecules, molecular dimers, torsional energy profiles, energy-minimum structures, and biomolecular simulations, Residual Learning Force Field gives accurate and stable predictions across diverse systems. These results suggest that combining physical constraints with data-driven corrections can provide a practical route toward more reliable and efficient molecular simulation for biological research and drug discovery. Molecular force fields underpin biomolecular simulation but often trade accuracy for generality. This study presents ResFF, a residual-learning model that combines physical rules with neural corrections to improve accuracy while retaining generality.
Abstract Motivation Machine Learning Force Fields (MLFFs) have emerged as promising tools for accelerating molecular dynamics simulations. However, existing approaches often struggle to capture the geometric characteristics of long-range interactions, including distance and direction, remain sensitive to conformational variations, and lack adaptive mechanisms for balancing short- and long-range forces. To address these limitations, we propose GeoNet, a physicochemical-principle-guided framework for modeling dual-range atomic interactions. GeoNet employs geometric attention over atom–fragment bipartite graphs to characterize long-range dependencies, introduces dual-level augmentation to enforce semantic consistency across molecular conformations, and uses an adaptive fusion module to dynamically balance short- and long-range interaction pathways according to local atomic environments. Results Extensive experiments show that GeoNet consistently outperforms ten state-of-the-art baselines across the evaluated benchmarks. Moreover, it achieves the smallest model size and the shortest training time, demonstrating both superior predictive performance and computational efficiency. Availability The source code is publicly available at https://github.com/XMUDM/GeoNet.