arXiv:2610.13038 · IEEE Access
Error Mitigation by Zero-Noise Extrapolation (ZNE) in Indirect-Control System
Arijit Das, Masaki Owari
Abstract
The indirect-control paradigm offers a scalable approach to quantum computation by restricting external operations to a small subset of qubits, whereas the entire system evolves under its intrinsic many-body Hamiltonian. This framework reduces hardware complexity and limits the noise introduced through control channels. Recently, Anan et al. proposed an implementation of the variational quantum eigensolver (VQE) within this indirect-control architecture. However, the impact and integration of quantum error mitigation techniques in such a setting remain largely unexplored. In particular, zero-noise extrapolation (ZNE), a widely used error mitigation strategy, has not yet been systematically adapted to indirect-control VQE. A key requirement of ZNE is the controlled amplification of noise, which is commonly achieved through circuit-level techniques such as unitary folding. In the indirect-control framework, however, quantum operations are largely governed by continuous time-evolution under the system Hamiltonian, making the construction of physically realizable inverse operations non-trivial. In this study, we investigate the feasibility of implementing ZNE within the indirect-controlVQEframework.We address the challenges associated with unitary folding in the presence of time-evolution operators and, with a minimal relaxation of the indirect-control constraint, we propose a ZNE-based error-mitigation strategy that exploits the symmetry of the one-dimensional XY spin chain.
AI Summary
The paper asks how to implement zero-noise extrapolation (ZNE) for indirect-control VQE, where gates act only on a small control subsystem and the remaining qubits evolve under a fixed many-body Hamiltonian. The obstacle to unitary folding is realizing inverse drift evolution without reversing the Hamiltonian. For a field-free chain with $H_S=\frac12\sum_k(X_kX_{k+1}+Z_kZ_{k+1})$, applying $Y$ to every odd site gives $PH_SP=-H_S$, hence $Pe^{-iH_St}P=e^{iH_St}$. Allowing these additional pulses outside the control subsystem therefore enables folding. Numerical tests use density-matrix simulations of 30-layer VQE circuits, primarily targeting transverse-field Ising energies, with local depolarizing, dephasing, and amplitude-damping noise. For seven qubits with noise only during drift evolution and decoherence rate $\gamma=10^{-3}$, second-order Richardson extrapolation gives $-8.44\pm0.06$, compared with $-7.66\pm0.10$ unmitigated and $-8.41\pm0.07$ for noiseless VQE. With all gates noisy on eight qubits, third-order multivariate extrapolation gives $-9.52\pm0.08$, versus $-7.03\pm0.16$ unmitigated and $-9.66\pm0.05$ noiseless, using 20 circuit configurations and an estimated equal-allocation sampling overhead of $4.35\times10^3$. The reported uncertainties reflect random VQE initialization, not finite-shot noise. Local noise does not generally scale proportionally under folding, even to first order. The construction requires additional odd-site control and relies on the drift Hamiltonian’s symmetry; extension to general interacting systems remains unresolved. Mitigation deteriorates at stronger noise or longer evolution times, sampling overhead grows rapidly with extrapolation order, and experimental validation and modeling of correlated pulse errors, crosstalk, leakage, and coherent errors remain open.
Error found: For the simulated gate-plus-local-noise model, extrapolation in the folding scale generally retains a first-order noise term, so the reported bias reductions do not establish convergence to the true zero-noise expectation. The final Trotter-error bound is smaller by a factor of two than follows from the two inequalities explicitly used to derive it. The runtime and ZNE computational-cost formulas overcount resources by the number of extrapolation circuits because they apply the total shot budget to every circuit in the sum.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- DOI
- 10.1109/ACCESS.2026.3741203
- Comments
- 28 pages, 21 figures, 5 tables. Accepted for publication in IEEE Access