- Accepted Paper
Neural-network-enhanced quantum simulation of the Schwinger model
Phys. Rev. A - Accepted 16 November, 2026
DOI: https://doi.org/10.1103/x74f-gk22
Phys. Rev. A - Accepted 16 November, 2026
DOI: https://doi.org/10.1103/x74f-gk22
The Schwinger Model, as a D Quantum Electrodynamics~(QED) theory, has long served as a valuable simplified model for exploring key physical phenomena in Quantum Chromodynamics~(QCD)—a non-Abelian gauge theory rich in fundamental physics but substantially more complex. While explicit representation learning succeeds in some well confined simulation studies on the Schwinger Model, the need for tedious adjustment whenever physical setup changes remains a main issue of the technique. Here, we propose a model that we refer to as the Neural Network Facilitated Implicit Quantum Simulation~(NN-IQS) model as a solution to that issue. After training on limited discrete phase data from the Schwinger Model, the NN-IQS model learns an implicit functional mapping and allows quick generation of extra sample points over a continuous parameter domain without any retraining. The model further shows a potential to generalize beyond its training range, maintaining robust performance in previously unexplored parameter-domains and system sizes.
If the author has provided any supplemental materials with this article they will be available upon publication of the version of record.