English

Gauge-equivariant neural networks as preconditioners in lattice QCD

High Energy Physics - Lattice 2023-02-13 v1 Machine Learning Numerical Analysis Numerical Analysis

Abstract

We demonstrate that a state-of-the art multi-grid preconditioner can be learned efficiently by gauge-equivariant neural networks. We show that the models require minimal re-training 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.

Keywords

Cite

@article{arxiv.2302.05419,
  title  = {Gauge-equivariant neural networks as preconditioners in lattice QCD},
  author = {Christoph Lehner and Tilo Wettig},
  journal= {arXiv preprint arXiv:2302.05419},
  year   = {2023}
}

Comments

12 pages, 12 figures