Gauge-equivariant neural networks as preconditioners in lattice QCD

2023/01/01 by Christoph Lehner, Lehner, Christoph, Tilo Wettig +1 · 4 citations
Medicine · Physics and Astronomy · #Advanced Neuroimaging Techniques and Applications #Atomic and Subatomic Physics Research #FOS: Computer and information sciences #FOS: Mathematics #FOS: Physical sciences #High Energy Physics - Lattice (hep-lat) #Machine Learning (cs.LG) #Numerical Analysis (math.NA) #Quantum Chromodynamics and Particle Interactions

paper · pdf · doi:10.48550/arxiv.2302.05419

openalex publication_date 2023/01/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We demonstrate that a state-of-the-art multigrid preconditioner can be learned efficiently by gauge-equivariant neural networks. We show that the models require minimal retraining on different gauge configurations of the same gauge ensemble and to a large extent remain efficient under modest modifications of ensemble parameters. We also demonstrate that important paradigms such as communication avoidance are straightforward to implement in this framework.

Cited by

• Reinforcement Learning techniques for the flavor problem in particle physics
• Matrix-free Neural Preconditioner for the Dirac Operator in Lattice Gauge Theory
• Graph Lineages and Skeletal Graph Products
• Accelerating multigrid with streaming chiral SVD for Wilson fermions in lattice QCD

Related

• Gauge-equivariant pooling layers for preconditioners in lattice QCD
• Matrix-free Neural Preconditioner for the Dirac Operator in Lattice Gauge Theory
• Lattice gauge symmetry in neural networks
• Symmetry-preserving neural networks in lattice field theories
• Gauge Equivariant Neural Networks for 2+1D U(1) Gauge Theory Simulations in Hamiltonian Formulation