Machine learning mapping of lattice correlated data

Physics Letters B, vol. 856, pp. 138894

Abstract

We discuss a machine learning (ML) regression model to reduce the computational cost of disconnected diagrams in lattice QCD calculations. This method creates a mapping between the results of fermionic loops computed at different quark masses and flow times. The ML mapping, trained with just a small fraction of the complete data set, makes use of translational invariance and provides consistent result with comparable uncertainties over the calculation done over the whole ensemble, resulting in a significant computational gain.

Authors 3

1. Jangho Kim corresponding

   Forschungszentrum Jülich

   Affiliation as printed

   Institute for Advanced Simulation (IAS-4) - Forschungszentrum Jülich, Wilhelm-Johnen-Straße, Jülich, 52428, Germany

2. Giovanni Pederiva

   Jülich Supercomputing Centre

   Affiliation as printed

   Jülich Supercomputing Centre (JSC) & Center for Advanced Simulation and Analytics (CASA), Wilhelm-Johnen-Straße, Jülich, 52428, Germany

3. Andrea Shindler Aachen Institute for Theoretical Particle Physics and Cosmology

   RWTH Aachen University · Lawrence Berkeley National Laboratory · University of California, Berkeley

   Affiliation as printed

   Department of Physics, University of California, Berkeley, CA 94720, USA

   Institute for Theoretical Particle Physics and Cosmology, TTK, RWTH Aachen University, Sommerfeldstr. 16, Aachen, 52074, Germany

   Nuclear Science Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, USA