Spectral density analysis of time correlation functions in lattice QCD using the maximum entropy method

TL;DR

This paper explores the use of the maximum entropy method with simulated annealing to extract spectral information from lattice QCD correlation functions, addressing data quality and parameter dependence.

Contribution

It introduces a novel application of simulated annealing to improve spectral density extraction in lattice QCD using MEM.

Findings

Effective spectral mass density functions obtained

Dependence on MEM parameters analyzed

Practical implementation issues discussed

Abstract

We study various aspects of extracting spectral information from time correlation functions of lattice QCD by means of Bayesian inference with an entropic prior, the maximum entropy method (MEM). Correlator functions of a heavy-light meson-meson system serve as a repository for lattice data with diverse statistical quality. Attention is given to spectral mass density functions, inferred from the data, and their dependence on the parameters of the MEM. We propose to employ simulated annealing, or cooling, to solve the Bayesian inference problem, and discuss practical issues of the approach.