High pressure hydrogen by machine learning and quantum Monte Carlo

Abstract

We have developed a technique combining the accuracy of quantum Monte Carlo in describing the electron correlation with the efficiency of a Machine Learning Potential (MLP). We use kernel regression in combination with SOAP (Smooth Overlap of Atomic Position) features, implemented here in a very efficient way. The key ingredients are: i) a sparsification technique, based on farthest point sampling, ensuring generality and transferability of our MLPs and ii) the so called $\Delta$-learning, allowing a small training data set, a fundamental property for highly accurate but computationally demanding calculations, such as the ones based on quantum Monte Carlo. As the first application we present a benchmark study of the liquid-liquid transition of high-pressure hydrogen and show the quality of our MLP, by emphasizing the importance of high accuracy for this very debated subject, where experiments are difficult in the lab, and theory is still far from being conclusive.

arXiv commentrevised exposition, performed more validation tests. Comments welcome!

References in corpus (21)

• Quantum ESPRESSO: a modular and open-source software project for quantum simulations of materials
• Canonical sampling through velocity-rescaling
• Gaussian Approximation Potentials: the accuracy of quantum mechanics, without the electrons
• Conventional superconductivity at 203 K at high pressures
• On representing chemical environments
• Superconductivity at 250 K in lanthanum hydride under high pressures
• Evidence for superconductivity above 260 K in lanthanum superhydride at megabar pressures
• Big Data meets Quantum Chemistry Approximations: The $\Delta$-Machine Learning Approach
• Hydrogen-Helium Mixtures in the Interiors of Giant Planets
• Evidence for supercritical behavior of high-pressure liquid hydrogen
• A First-order Phase Transition to Metallic Hydrogen
• Quantum Monte Carlo study of the phase diagram of solid molecular hydrogen at extreme pressures
• Benchmark of Exchange-Correlation Functionals for High Pressure Hydrogen using Quantum Monte Carlo
• Phase diagram of hydrogen and a hydrogen-helium mixture at planetary conditions by Quantum Monte Carlo simulations
• Fate of density functional theory in high-pressure solid hydrogen
• Quantum Monte Carlo Simulation of the High-Pressure Molecular-Atomic Crossover in Fluid Hydrogen
• TurboRVB: a many-body toolkit for {\it ab initio} electronic simulations by quantum Monte Carlo
• Conductivity and Dissociation in Metallic Hydrogen: Implications for Planetary Interiors
• Thermodynamic anomalies and three distinct liquid-liquid transitions in warm dense liquid hydrogen
• Fully Consistent Density Functional Theory Determination of the Insulator-Metal Transition Boundary in Warm Dense Hydrogen
• Space-warp coordinate transformation for efficient ionic force calculations in quantum Monte Carlo

Cited by in corpus (5)

• Stable solid molecular hydrogen above 900K from a machine-learned potential trained with diffusion Quantum Monte Carlo
• Deep Variational Free Energy Approach to Dense Hydrogen
• Principal deuterium Hugoniot via Quantum Monte Carlo and $\Delta$-learning
• A structural optimization algorithm with stochastic forces and stresses
• Extending the reach of quantum computing for materials science with machine learning potentials