Applying machine learning methods to prediction problems of lattice observables

Journal Article · 31 May 2022 · SciPost Physics Proceedings

DOI: https://doi.org/10.21468/SciPostPhysProc.6.020 · OSTI ID:1870544

Gerasimeniuk, Nikolai ^[1]; Chernodub, Maxim N. ^[2]; Goy, V. A. ^[2]; Boyda, D. L. ^[3]; Liubimov, S. D. ^[1]; Molochkov, A. V. ^[1]

---

1. Far Eastern Federal University
2. François Rabelais University, Far Eastern Federal University
3. Argonne National Laboratory

We discuss the prediction of critical behavior of lattice observables in SU(2) and SU(3) gauge theories. We show that feed-forward neural network, trained on the lattice configurations of gauge fields as input data, finds correlations with the target observable, which is also true in the critical region where the neural network has not been trained. We have verified that the neural network constructs a gauge-invariant function and this property does not change over the entire range of the parameter space.

• Sponsoring Organization: USDOE
• Grant/Contract Number: AC02-06CH11357
• OSTI ID: 1870544
• Journal Information: SciPost Physics Proceedings, Journal Name: SciPost Physics Proceedings Journal Issue: 6; ISSN 2666-4003
• Publisher: Stichting SciPostCopyright Statement
• Country of Publication: Netherlands
• Language: English