Speaker
Description
By combining Lattice QCD data with physics-informed machine learning (PIML), we investigate the behavior of the coupling constant and the quark anomalous magnetic moment (AMM) within the Nambu–Jona-Lasinio (NJL) model under a magnetic field. We train a neural network using Lattice QCD data and the analytical expression of the thermodynamic potential in the NJL model as core constraints. This model not only successfully reproduces the dependence of the chiral phase transition temperature on the magnetic field as revealed by Lattice QCD, but also precisely predicts the specific running forms of the coupling constant G and the AMM parameter $v_{2}$ over a wide range of magnetic fields. On this basis, we incorporate these magnetic-field-dependent parameters into the calculation of meson polarization functions to systematically investigate the behavior of the masses of the neutral scalar σ meson and the neutral ${\pi }^{0}$ meson as functions of the magnetic field. Consequently, this study reveals the properties of the meson spectrum in the magnetized QCD vacuum within a unified framework that includes running coupling and anomalous magnetic moment effects.
