Machine learning predictions for local electronic properties of disordered correlated electron systems

Abstract

We present a scalable machine learning (ML) model to predict local electronic properties such as on-site electron number and double occupation for disordered correlated electron systems. Our approach is based on the locality principle, or the nearsightedness nature, of many-electron systems, which means local electronic properties depend mainly on the immediate environment. A ML model is developed to encode this complex dependence of local quantities on the neighborhood. We demonstrate our approach using the square-lattice Anderson-Hubbard model, which is a paradigmatic system for studying the interplay between Mott transition and Anderson localization. We develop a lattice descriptor based on group-theoretical method to represent the on-site random potentials within a finite region. The resultant feature variables are used as input to a multi-layer fully connected neural network, which is trained from datasets of variational Monte Carlo (VMC) simulations on small systems. We show that the ML predictions agree reasonably well with the VMC data. Our work underscores the promising potential of ML methods for multi-scale modeling of correlated electron systems.

arXiv comment12 pages, 6 figures

References in corpus (50)

• Deep Learning in Neural Networks: An Overview
• Gaussian Approximation Potentials: the accuracy of quantum mechanics, without the electrons
• On representing chemical environments
• Machine learning and the physical sciences
• Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning
• Deep Potential Molecular Dynamics: a scalable model with the accuracy of quantum mechanics
• ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost
• Moment Tensor Potentials: a class of systematically improvable interatomic potentials
• The Kernel Polynomial Method
• Big Data of Materials Science - Critical Role of the Descriptor
• Machine learning for molecular simulation
• DScribe: Library of Descriptors for Machine Learning in Materials Science
• Nearsightedness of Electronic Matter
• Unified Representation of Molecules and Crystals for Machine Learning
• Wave function optimization in the variational Monte Carlo method
• Mott--Hubbard transition vs. Anderson localization of correlated, disordered electrons
• Variational description of Mott insulators
• Mean field theory of the Mott-Anderson transition
• Typical medium theory of Anderson localization: A local order parameter approach to strong disorder effects
• Machine learning meets quantum physics
• Machine Learning for Condensed Matter Physics
• Critical behavior at Mott-Anderson transition: a TMT-DMFT perspective
• Inhomogeneous metallic phase upon disordering a two dimensional Mott insulator
• Disorder Screening in Strongly Correlated Systems
• Emergent Novel Metallic State in a Disordered 2D Mott Insulator
• Glassy behavior of electrons near metal-insulator transitions
• Two-dimensional Holstein-Hubbard model: Critical temperature, Ising universality, and bipolaron liquid
• Single-particle excitations under coexisting electron correlation and disorder: a numerical study of the Anderson-Hubbard model
• Electronic Griffiths phase of the d=2 Mott transition
• Disorder-Induced Stabilization of the Pseudogap in Strongly Correlated Systems
• Anderson-Hubbard model with box disorder: Statistical dynamical mean-field theory investigation
• Magnetic Correlations in the Two Dimensional Anderson-Hubbard Model
• Machine learning for molecular dynamics with strongly correlated electrons
• Gradient-based stochastic estimation of the density matrix
• Disorder and Interaction in 2D: Exact diagonalization study of the Anderson-Hubbard-Mott model
• Non Fermi liquid behavior and continuously tunable resistivity exponents in the Anderson-Hubbard model at finite temperature
• Delocalizing effect of the Hubbard repulsion for electrons on a two-dimensional disordered lattice
• Arrested phase separation in double-exchange models: machine-learning enabled large-scale simulation
• The role of strong electronic correlations in the metal-to-insulator transition in disordered LiAl_yTi_(2-y)O_4
• Ground-state properties of the disordered Hubbard model in two dimensions
• Machine learning electron correlation in a disordered medium
• Quantum ripples in strongly correlated metals
• Crystal Structure Representations for Machine Learning Models of Formation Energies
• Collapse of Charge Gap in Random Mott Insulators
• The two-dimensional disordered Mott metal-insulator transition
• Anomalous phase separation dynamics in a correlated electron system: machine-learning enabled large-scale kinetic Monte Carlo simulations
• Local moments and magnetic order in the two-dimensional Anderson-Mott transition
• Local Spectroscopies Reveal Percolative Metal in Disordered Mott Insulators
• Variational Monte Carlo Study of Anderson Localization in the Hubbard Model
• Descriptors for Machine Learning Model of Generalized Force Field in Condensed Matter Systems

Cited by in corpus (8)

• Machine learning nonequilibrium electron forces for adiabatic spin dynamics
• Convolutional neural networks for large-scale dynamical modeling of itinerant magnets
• Machine learning for phase ordering dynamics of charge density waves
• Machine learning for structure-property relationships: Scalability and limitations
• Kinetics of orbital ordering in cooperative Jahn-Teller models: Machine-learning enabled large-scale simulations
• Unsupervised Machine Learning Phase Classification for Falicov-Kimball Model
• Machine-learning force-field models for dynamical simulations of metallic magnets
• Machine learning approach for vibronically renormalized electronic band structures