Equivariant flow-based sampling for lattice gauge theory

INSPIRE: 1785309
arXiv: 2003.06413
DOI: 10.1103/PhysRevLett.125.121601

Authors: Kanwar, Gurtej, Albergo, Michael S., Boyda, Denis, Cranmer, Kyle, Hackett, Daniel C., Racanière, Sébastien, Rezende, Danilo Jimenez, Shanahan, Phiala E.

Submitted: 13 March 2020

Subjects:

• hep-lat
• cond-mat.stat-mech
• cs.LG
• Lattice
• General Physics
• Computing
• Condensed Matter

Journal reference: Phys.Rev.Lett. 125 (12) 121601 (2020)

Citations: 221

Abstract

We define a class of machine-learned flow-based sampling algorithms for lattice gauge theories that are gauge invariant by construction. We demonstrate the application of this framework to U(1) gauge theory in two spacetime dimensions, and find that, at small bare coupling, the approach is orders of magnitude more efficient at sampling topological quantities than more traditional sampling procedures such as hybrid Monte Carlo and heat bath.