Machine-learning the spectral function of a hole in a quantum antiferromagnet

TL;DR

This paper demonstrates that machine learning techniques, specifically neural networks, can accurately predict spectral functions and infer Hamiltonian parameters in a quantum antiferromagnet model, facilitating rapid analysis of complex quantum systems.

Contribution

It introduces the use of neural networks for predicting spectral functions and inferring model parameters in the $t$-$t'$-$t''$-$J$ model, advancing computational methods in quantum many-body physics.

Findings

Both KNN and FFNN accurately predict spectral functions.

FFNN can infer Hamiltonian parameters from spectra.

KNN cannot reliably predict parameters from density-of-states.

Abstract

Understanding charge motion in a background of interacting quantum spins is a fundamental problem in quantum many-body physics. The most extensively studied model for this problem is the so-called $t$-$t'$-$t''$-$J$ model, where the determination of the parameter $t'$ in the context of cuprate superconductors is challenging. Here we present a theoretical study of the spectral functions of a mobile hole in the $t$-$t'$-$t''$-$J$ model using two machine learning techniques: K-nearest Neighbors regression (KNN) and a feed-forward neural network (FFNN). We employ the self-consistent Born approximation to generate a dataset of about $1.3 \times 10^5$ spectral functions. We show that for the forward problem, both methods allow for the accurate and efficient prediction of spectral functions, allowing for e.g. rapid searches through parameter space. Furthermore, we find that for the inverse problem (inferring Hamiltonian parameters from spectra), the FFNN can, but the KNN cannot, accurately predict the model parameters using merely the density-of-state. Our results suggest that it may be possible to use deep learning methods to predict materials parameters from experimentally measured spectral functions.