Robust Training of Machine Learning Interatomic Potentials with Dimensionality Reduction and Stratified Sampling

2023/07/24 by Ji Qi, Qi, Ji, Tsz Wai Ko +7 · 11 citations
Chemistry · Materials Science · #FOS: Physical sciences #Inorganic Chemistry and Materials #Machine Learning in Materials Science #Materials Science (cond-mat.mtrl-sci) #X-ray Diffraction in Crystallography

paper · pdf · doi:10.48550/arxiv.2307.13710

openalex publication_date 2023/07/24 · openalex created_date 2023/07/28 · openalex updated_date 2026/07/29

Abstract

Machine learning interatomic potentials (MLIPs) enable the accurate simulation of materials at larger sizes and time scales, and play increasingly important roles in the computational understanding and design of materials. However, MLIPs are only as accurate and robust as the data they are trained on. In this work, we present DImensionality-Reduced Encoded Clusters with sTratified (DIRECT) sampling as an approach to select a robust training set of structures from a large and complex configuration space. By applying DIRECT sampling on the Materials Project relaxation trajectories dataset with over one million structures and 89 elements, we develop an improved materials 3-body graph network (M3GNet) universal potential that extrapolate more reliably to unseen structures. We further show that molecular dynamics (MD) simulations with universal potentials such as M3GNet can be used in place of expensive ab initio MD to rapidly create a large configuration space for target materials systems. Combined with DIRECT sampling, we develop a highly reliable moment tensor potential for Ti-H system without the need for iterative optimization. This work paves the way towards robust high throughput development of MLIPs across any compositional complexity.

Cited by

• Beyond Adam: Disentangling Optimizer Effects in the Fine-Tuning of Atomistic Foundation Models
• Platonic representation of foundation machine learning interatomic potentials
• Hydrogen diffusion in TiCr₂H_x Laves phases: A combined ab initio and machine-learning-potential study
• Automated Prediction of Thermodynamic Properties via Bayesian Free-Energy Reconstruction from Molecular Dynamics
• Maximizing Efficiency of Dataset Compression for Machine Learning Potentials With Information Theory
• A Fast, Accurate, and Reactive Equivariant Foundation Potential
• AQCat25: Unlocking spin-aware, high-fidelity machine learning potentials for heterogeneous catalysis
• PASS: Perturbation augmented space group structure sampling for transferable Fe-O machine learning interatomic potential
• Benchmarking CHGNet Universal Machine Learning Interatomic Potential Against DFT and EXAFS: Case of Layered WS2 and MoS2
• Autonomous thermodynamically informed database generation for machine-learned interatomic potentials and application to magnesium
• Model Accuracy and Data Heterogeneity Shape Uncertainty Quantification in Machine Learning Interatomic Potentials

Related

• Universal Machine Learning Potential for Systems with Reduced Dimensionality
• Data-efficient multi-fidelity training for high-fidelity machine learning interatomic potentials
• Machine-learning interatomic potentials from a users perspective: A comparison of accuracy, speed and data efficiency
• Energy & Force Regression on DFT Trajectories is Not Enough for Universal Machine Learning Interatomic Potentials
• Fast and Accurate Foundation Models for Equivariant Machine-Learned Interatomic Potentials