Quantum
Disentangling Expressibility, Symmetry Protection, and Hardware Noise in Variational Quantum Simulation of the Two-Flavor Schwinger Model
Publication: quant-ph updates on arXiv.org
Author: Karthikeya Machiraju, Krishna Sujith, Kaustav Bhowmick
Source published: 28/09/2026, 04:00:00 UTC
Briefing created: 28/09/2026, 04:21:20 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 28/09/2026, 04:13:25 UTC.
Briefings
This study investigates variational quantum simulation of the two-flavor Schwinger model by expanding lattice size from N=2 to N=6. Key findings indicate hardware noise, not circuit expressibility, limits simulation reach, identifying N=3 as the viable extension for trapped-ion experiments. The energy error collapses to one function of the parameter-to-dimension ratio, decreasing by over two orders of magnitude as this ratio increases. Charge conservation protects trainability, reducing normalized gradient variance by 1/3.56 when qubit count doubles from 4 to 8, compared to 1/13.57 for unconstrained circuits. A fixed-parameter control shows noise degradation depends on the noise model, not the optimizer's location in the noisy loop.
Key points
- The binding constraint for variational quantum simulation of the two-flavor Schwinger model is hardware noise rather than circuit expressibility or trainability, limiting simulation reach to N=3 for trapped-ion experiments.
- Energy error of a charge-conserving ansatz collapses onto a single function of the parameter-to-dimension ratio, decreasing by more than two orders of magnitude as this ratio rises through order unity.
Context and limitations
The study covers lattice sizes from N=2 to N=6 using a single snapshot at arXiv:2609.30496v1. Results are preliminary and subject to verification as the paper is a preprint.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 28/09/2026, 04:21:20 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeigg6h5rwodbb33c5ywi7i7ri6sbatsgopsqiudoit4eziwwsjkxkq