Practical applications of machine-learned flows on gauge fields
Authors:
- Daniel HackettDHDaniel Hackett
- Ryan AbbottRARyan Abbott
- Denis BoydaDBDenis Boyda
- Fernando Romero-LopezFRFernando Romero-Lopez
- Phiala ShanahanPSPhiala Shanahan
- Julian UrbanJUJulian Urban
- Michael AlbergoMAMichael Albergo
Open Access
- 3 May 2024
- conference paper
- Published by Sissa Medialab Srl
- Vol. 453, 011
Abstract
Normalizing flows are machine-learned maps between different lattice theories which can be used as components in exact sampling and inference schemes. Ongoing work yields increasingly expressive flows on gauge fields, but it remains an open question how flows can improve lattice QCD at state-of-the-art scales. We discuss and demonstrate two applications of flows in replica exchange (parallel tempering) sampling, aimed at improving topological mixing, which are viable with iterative improvements upon presently available flows.Keywords
All Related Versions
This publication has 52 references indexed in Scilit:
- Stochastic normalizing flows as non-equilibrium transformationsJournal of High Energy Physics, 2022
- Large-N SU(N) Yang-Mills theories with milder topological freezingJournal of High Energy Physics, 2021
- Lattice QCD without topology barriersJournal of High Energy Physics, 2011
- Critical slowing down and error analysis in lattice QCD simulationsNuclear Physics B, 2011
- Properties and uses of the Wilson flow in lattice QCDJournal of High Energy Physics, 2010
- Trivializing Maps, the Wilson Flow and the HMC AlgorithmCommunications in Mathematical Physics, 2009
- Critical slowing downNuclear Physics B - Proceedings Supplements, 1990
- Hybrid Monte CarloPhysics Letters B, 1987
- Improved heatbath method for Monte Carlo calculations in lattice gauge theoriesPhysics Letters B, 1985
- A new method for updating SU(N) matrices in computer simulations of gauge theoriesPhysics Letters B, 1982