Deep learning lattice gauge theories

Anuj Apte,Anthony Ashmore,Clay Cordova,Tzu-Chen Huang

2024-05-24

Abstract:Monte Carlo methods have led to profound insights into the strong-coupling behaviour of lattice gauge theories and produced remarkable results such as first-principles computations of hadron masses. Despite tremendous progress over the last four decades, fundamental challenges such as the sign problem and the inability to simulate real-time dynamics remain. Neural network quantum states have emerged as an alternative method that seeks to overcome these challenges. In this work, we use gauge-invariant neural network quantum states to accurately compute the ground state of $\mathbb{Z}_N$ lattice gauge theories in $2+1$ dimensions. Using transfer learning, we study the distinct topological phases and the confinement phase transition of these theories. For $\mathbb{Z}_2$, we identify a continuous transition and compute critical exponents, finding excellent agreement with existing numerics for the expected Ising universality class. In the $\mathbb{Z}_3$ case, we observe a weakly first-order transition and identify the critical coupling. Our findings suggest that neural network quantum states are a promising method for precise studies of lattice gauge theory.

High Energy Physics - Lattice,Disordered Systems and Neural Networks,Strongly Correlated Electrons,Machine Learning,High Energy Physics - Theory

What problem does this paper attempt to address?

This paper discusses the use of deep learning methods to address the challenges of lattice gauge theories in lattice gauge theories, particularly for 2+1-dimensional ZN lattice gauge theories. These theories play a crucial role in understanding strong coupling behaviors, such as quantum chromodynamics (QCD). Despite significant achievements in simulating QCD using Monte Carlo methods, there are still fundamental problems such as sign problems and the inability to simulate real-time dynamics.

Neural network quantum states have been proposed as alternative approaches to overcome these problems. Researchers use symmetry-invariant neural network quantum states to accurately compute the ground state of ZN theories and study different topological phases and phase transitions through transfer learning. In the Z2 theory, they discovered a continuous phase transition and calculated critical exponents consistent with the expected Ising universality class. In the Z3 theory, they observed a weak first-order phase transition and determined the critical coupling.

The main contributions of the paper are:

1. Accurate computation of the ground state energy of ZN lattice gauge theories using neural network quantum states.
2. Study of the phase structure of the Z2 theory through transfer learning, identification of a continuous phase transition, and calculation of critical exponents.
3. Observation of a weak first-order phase transition and determination of the critical coupling in the Z3 theory.
4. Provision of evidence that neural network quantum states are a promising approach for accurate study of lattice gauge theories.

This work provides new tools for understanding phase transitions and topological properties in lattice gauge theories and may contribute to future applications in more complex theories and dimensions.