A repetition-and-majority-vote scheme used extra qubits to suppress measurement errors by several orders of magnitude.
Measurement errors can reduce the reliability of quantum sampling tasks, including random circuit sampling and measurement-based quantum computation. The researchers developed a repetition-based scheme that copies each data qubit across multiple physical qubits before readout and uses the majority result.
Using experimental noise profiles from leading superconducting platforms, they found that the method converted individual physical errors into a higher-order process. The error suppression remained exponential even when the encoding operations themselves had realistic imperfections, while requiring a modest linear qubit overhead.
Evidence and caveats
The study used experimental noise profiles from superconducting quantum platforms to evaluate the proposed method, rather than reporting a direct hardware demonstration in the abstract. The abstract does not specify the platforms, circuit sizes, exact noise levels, error reductions or fidelity gains. Its conclusions therefore support the scheme's performance under the tested noise profiles, but do not by themselves establish how it will perform across all quantum hardware or sampling tasks.
// Source
npj Unconventional Computing · 2026 · DOI: 10.1038/s44335-026-00094-0
Authors: He-Liang Huang
Institutions: State Key Laboratory of Cryptology