Conceptual and practical bases for the high accuracy of machine learning interatomic potential

Abstract

Machine learning interatomic potentials (MLIPs) based on a large dataset obtained by density functional theory (DFT) calculation have been developed recently. This study gives both conceptual and practical bases for the high accuracy of MLIPs, although MLIPs have been considered to be simply an accurate black-box description of atomic energy. We also construct the most accurate MLIP of the elemental Ti ever reported using a linearized MLIP framework and many angular-dependent descriptors, which also corresponds to a generalization of the modified embedded atom method (MEAM) potential.

References in corpus (8)

• Gaussian Approximation Potentials: the accuracy of quantum mechanics, without the electrons
• On representing chemical environments
• Interatomic potentials for atomistic simulations of the Ti-Al system
• Interatomic potentials for ionic systems with density functional accuracy based on charge densities obtained by a neural network
• Classical potential describes martensitic phase transformations between the $\alpha$, $\beta$ and $\omega$ titanium phases
• Development of an Interatomic Potential for the Simulation of Defects, Plasticity and Phase Transformations in Titanium
• A sparse representation for potential energy surface
• First-principles interatomic potentials for ten elemental metals via compressed sensing

Cited by in corpus (20)

• Embedded Atom Neural Network Potentials: Efficient and Accurate Machine Learning with a Physically Inspired Representation
• De novo exploration and self-guided learning of potential-energy surfaces
• Extending the Accuracy of the SNAP Interatomic Potential Form
• Ab initio vibrational free energies including anharmonicity for multicomponent alloys
• Physically Motivated Recursively Embedded Atom Neural Networks: Incorporating Local Completeness and Nonlocality
• Application of machine learning potentials to predict grain boundary properties in fcc elemental metals
• Group-theoretical high-order rotational invariants for structural representations: Application to linearized machine learning interatomic potential
• Accurate machine learning force fields via experimental and simulation data fusion
• Drawing Phase Diagrams of Random Quantum Systems by Deep Learning the Wave Functions
• Machine learning potentials for multicomponent systems: The Ti-Al binary system
• Accelerating Atomistic Simulations with Piecewise Machine Learned Ab Initio Potentials at Classical Force Field-like Cost
• Exploring a potential energy surface by machine learning for characterizing atomic transport
• Combining quantum mechanics and machine-learning calculations for anharmonic corrections to vibrational frequencies
• Ab-initio based models for temperature-dependent magneto-chemical interplay in bcc Fe-Mn alloys
• Linearized machine-learning interatomic potentials for non-magnetic elemental metals: Limitation of pairwise descriptors and trend of predictive power
• Predictive power of polynomial machine learning potentials for liquid states in 22 elemental systems
• Phase transitions in titanium with an analytic bond-order potential
• Analysis of Kohn-Sham Eigenfunctions Using a Convolutional Neural Network in Simulations of the Metal-insulator Transition in Doped Semiconductors
• Systematic global structure search of bismuth-based binary systems under pressure using machine learning potentials
• REANN: A PyTorch-based End-to-End Multi-functional Deep Neural Network Package for Molecular, Reactive and Periodic Systems