A neural network potential with self-trained atomic fingerprints: a test with the mW water potential

Abstract

We present a neural network (NN) potential based on a new set of atomic fingerprints built upon two- and three-body contributions that probe distances and local orientational order respectively. Compared to existing NN potentials, the atomic fingerprints depend on a small set of tuneable parameters which are trained together with the neural network weights. To tackle the simultaneous training of the atomic fingerprint parameters and neural network weights we adopt an annealing protocol that progressively cycles the learning rate, significantly improving the accuracy of the NN potential. We test the performance of the network potential against the mW model of water, which is a classical three-body potential that well captures the anomalies of the liquid phase. Trained on just three state points, the NN potential is able to reproduce the mW model in a very wide range of densities and temperatures, from negative pressures to several GPa, capturing the transition from an open random tetrahedral network to a dense interpenetrated network. The NN potential also reproduces very well properties for which it was not explicitly trained, such as dynamical properties and the structure of the stable crystalline phases of mW.

arXiv comment14 pages, 11 figures

References in corpus (33)

• On representing chemical environments
• ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
• E(3)-Equivariant Graph Neural Networks for Data-Efficient and Accurate Interatomic Potentials
• Quantum-Chemical Insights from Deep Tensor Neural Networks
• Machine Learning of Accurate Energy-Conserving Molecular Force Fields
• Structural transformation in supercooled water controls the crystallization rate of ice
• Perspective: How good is DFT for water?
• A Fourth-Generation High-Dimensional Neural Network Potential with Accurate Electrostatics Including Non-local Charge Transfer
• Physics-inspired structural representations for molecules and materials
• Ab initio thermodynamics of liquid and solid water
• The Phase Diagram of a Deep Potential Water Model
• ANI-1: A data set of 20M off-equilibrium DFT calculations for organic molecules
• Automatic Selection of Atomic Fingerprints and Reference Configurations for Machine-Learning Potentials
• 86 PFLOPS Deep Potential Molecular Dynamics simulation of 100 million atoms with ab initio accuracy
• Machine learning force fields and coarse-grained variables in molecular dynamics: application to materials and biological systems
• On the Completeness of Atomic Structure Representations
• Coarse Graining Molecular Dynamics with Graph Neural Networks
• Evidence for supercritical behavior of high-pressure liquid hydrogen
• Comparison of permutationally invariant polynomials, neural networks, and Gaussian approximation potentials in representing water interactions through many-body expansions
• Nature of the anomalies in the supercooled liquid state of the mW model of water
• Water-like anomalies as a function of tetrahedrality
• Recursive evaluation and iterative contraction of $N$-body equivariant features
• Heat transport in liquid water from first-principles and deep-neural-network simulations
• Quantum-mechanical exploration of the phase diagram of water
• Ranking the information content of distance measures
• Ab-initio Structure and Thermodynamics of the RPBE-D3 Water/Vapor Interface by Neural-Network Molecular Dynamics
• The physics of Empty Liquids: from Patchy particles to Water
• Machine learning many-body potentials for colloidal systems
• Structural and topological changes across the liquid-liquid transition in water
• Roles of liquid structural ordering in glass transition, crystallization, and water's anomalies
• Machine-learning effective many-body potentials for anisotropic particles using orientation-dependent symmetry functions
• Measuring transferability issues in machine-learning force fields: The example of Gold-Iron interactions with linearized potentials
• Modeling of many-body interactions between elastic spheres through symmetry functions