Machine Learning Spin Dynamics of Strongly Correlated Electron Systems

In this dissertation, new numerical frameworks based on machine learning potentials to enable large-scale adiabatic quantum Landau-Lifshitz-Gilbert dynamics simulations of itinerant electron magnets are established. Such metallic spin systems are central to novel phenomena such as colossal magnetoresistance and spin-transfer torques. This approach is similar in spirit to the Behler-Parrinello machine learning scheme that has become a cornerstone of large-scale molecular dynamics method with the accuracy of quantum calculation. Based on the principle of locality for electronic systems, the total electronic energy is partitioned into contributions from individual spins which only depend on the local environment. A neural network model is then trained from exact solutions on small systems to approximate the complex dependence of the local energy on the neighborhood spin configuration. We further develop novel descriptors to ensure the spin rotation symmetry as well as the discrete lattice symmetry. This work opens new avenues for using deep learning models to simulate and understand large-scale dynamical phenomena in functional magnetic systems.