QESEM (ZNE), ibm_boston, ZZd2, 51qx26c

[TasneemWtd]

Name

QESEM (ZNE), ibm_boston, ZZd2, 51qx26c

Circuit

floquet_mixed_field_ising_zzd2_51qx26c

Observable value

0.2604

Error bound (low)

0.2569

Error bound (high)

0.2639

Method

QESEM-ZNE, IBM

Method proof

The experiment is executed on ibm_boston using Qedma’s QESEM error-suppression and error-mitigation framework. The protocol uses the same device characterization, Pauli-noise modeling, error-suppression stack, drift handling, and hardware-native fractional-angle gate support as QESEM-unbiased, building on the framework described in Ref. [1]. As in the “QESEM (unbiased), ibm_boston, mag, 51qx16c #167” solution, the Floquet circuit is implemented directly using the device’s native fractional-angle gates, including fractional-angle RZZ gates.

For the later Floquet cycles reported here, the sampling overhead of the unbiased quasiprobabilistic estimator is too large. We therefore use zero-noise extrapolation (ZNE) based on QESEM, following the standard approach of estimating observables at amplified noise levels and extrapolating to zero noise [2]. In this implementation, double-noise data are generated using quasiprobabilistic error amplification from the characterized noise model, and the zero-noise value is inferred using an exponential extrapolation ansatz.

1. D. Aharonov et al., Reliable high-accuracy error mitigation for utility-scale quantum circuits, arXiv:2508.10997, 2025.
2. A. Kandala et al., Error mitigation extends the computational reach of a noisy quantum processor, Nature 567, 491–495 (2019).

Quantum runtime (seconds)

1200

Classical runtime (seconds)

668

Compute resources (quantum)

ibm_boston

Compute resources (classical)

32 CPUS, 32 GB

Notes

No response

Authors

No response

Institutions

Qedma, IBM, Riken, BlueQubit