Performance Assessment of Universal Machine Learning Interatomic Potentials: Challenges and Directions for Materials' Surfaces

Abstract

Machine learning interatomic potentials (MLIPs) are one of the main techniques in the materials science toolbox, able to bridge ab initio accuracy with the computational efficiency of classical force fields. This allows simulations ranging from atoms, molecules, and biosystems, to solid and bulk materials, surfaces, nanomaterials, and their interfaces and complex interactions. A recent class of advanced MLIPs, which use equivariant representations and deep graph neural networks, is known as universal models. These models are proposed as foundational models suitable for any system, covering most elements from the periodic table. Current universal MLIPs (UIPs) have been trained with the largest consistent dataset available nowadays. However, these are composed mostly of bulk materials' DFT calculations. In this article, we assess the universality of all openly available UIPs, namely MACE, CHGNet, and M3GNet, in a representative task of generalization: calculation of surface energies. We find that the out-of-the-box foundational models have significant shortcomings in this task, with errors correlated to the total energy of surface simulations, having an out-of-domain distance from the training dataset. Our results show that while UIPs are an efficient starting point for fine-tuning specialized models, we envision the potential of increasing the coverage of the materials space towards universal training datasets for MLIPs.

References in corpus (15)

• E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentials
• A Universal Graph Deep Learning Interatomic Potential for the Periodic Table
• Atomistic Line Graph Neural Network for Improved Materials Property Predictions
• Exceptional piezoelectricity, high thermal conductivity and stiffness and promising photocatalysis in two-dimensional MoSi2N4 family confirmed by first-principles
• Evaluation of the MACE Force Field Architecture: from Medicinal Chemistry to Materials Science
• A critical examination of robustness and generalizability of machine learning prediction of materials properties
• How to validate machine-learned interatomic potentials
• Performance Assessment of Universal Machine Learning Interatomic Potentials: Challenges and Directions for Materials' Surfaces
• Ultra-fast interpretable machine-learning potentials
• Machine learning for materials discovery: two-dimensional topological insulators
• Linear Jacobi-Legendre expansion of the charge density for machine learning-accelerated electronic structure calculations
• Lattice reconstruction in MoSe${}_2$-WSe${}_2$ heterobilayers synthesized by chemical vapor deposition
• Domain-dependent surface adhesion in twisted few-layer graphene: Platform for moiré-assisted chemistry
• One-Dimensional Moiré Physics and Chemistry in Heterostrained Bilayer Graphene
• Conformational analysis of tannic acid: environment effects in electronic and reactivity properties

Cited by in corpus (14)

• A practical guide to machine learning interatomic potentials -- Status and future
• Performance Assessment of Universal Machine Learning Interatomic Potentials: Challenges and Directions for Materials' Surfaces
• Machine Learning and Data-Driven Methods in Computational Surface and Interface Science
• Universal machine learning interatomic potentials poised to supplant DFT in modeling general defects in metals and random alloys
• Augmentation of Universal Potentials for Broad Applications
• Application-specific machine-learned interatomic potentials: exploring the trade-off between DFT convergence, MLIP expressivity, and computational cost
• Multi-head committees enable direct uncertainty prediction for atomistic foundation models
• Lifelong Machine Learning Potentials for Chemical Reaction Network Explorations
• RAFFLE: Active learning accelerated interface structure prediction
• Fine-tuning of universal machine-learning interatomic potentials for high-entropy alloys with application to 2D (Mo,Ta,Nb,W,V)S${}_2$
• Active Δ-learning with universal potentials for global structure optimization
• Uncertainty in AI-driven Monte Carlo simulations
• Accelerated Discovery of Topological Conductors for Nanoscale Interconnects
• M-CODE: Materials Categorization via Ontology, Dimensionality and Evolution