Quantum Error Mitigation at the pre-processing stage

The realization of fault-tolerant quantum computers remains a challenging endeavor, forcing state-of-the-art quantum hardware to rely heavily on noise mitigation techniques. Standard quantum error mitigation is typically based on post-processing strategies. In contrast, the present work explores a pre-processing approach, in which the effects of noise are mitigated before performing a measurement on the output state. The main idea is to find an observable $Y$ such that its expectation value on a noisy quantum state $\mathcal{E(ρ)}$ matches the expectation value of a target observable $X$ on the noiseless quantum state $ρ$. Our method requires the execution of a noisy quantum circuit, followed by the measurement of the surrogate observable $Y$. The main enablers of our method in practical scenarios are Tensor Networks. The proposed method improves over Tensor Error Mitigation (TEM) in terms of average error, circuit depth, and complexity, attaining a measurement overhead that approaches the theoretical lower bound. The improvement in terms of classical computation complexity is in the order of $\sim 10^6$ times when compared to the post-processing computational cost of TEM in practical scenarios. Such gain comes from eliminating the need to perform the set of informationally complete positive operator-valued measurements (IC-POVM) required by TEM, as well as any other tomographic strategy.

💡 Research Summary

The paper introduces a novel pre‑processing quantum error mitigation (QEM) strategy that eliminates the need for post‑processing of measurement data. Traditional QEM techniques such as zero‑noise extrapolation, probabilistic error cancellation, or tensor error mitigation (TEM) rely on collecting many shots from noisy circuits and then applying sophisticated classical post‑processing, often involving informationally complete POVMs or shadow tomography. In contrast, the authors propose to construct a surrogate observable (\hat Y) whose expectation value on the noisy output state (\mathcal{E}(\rho)) exactly matches the expectation value of the target observable (X) on the ideal (noiseless) state (\rho). Formally, they require (\mathcal{E}^\dagger(\hat Y)=X); measuring (\hat Y) on the noisy circuit therefore yields a noise‑free estimate of (\langle X\rangle_\rho) without any additional classical correction.

The direct solution (\hat Y = y_0 I + y^T P) with (y = A^{-T}x) and (y_0 = -c^T A^{-T}x) (where (A) and (c) are the Pauli‑transfer matrix representation of the noise channel) is computationally intractable because (A) is a ((4^n-1)\times(4^n-1)) matrix. To overcome this, the authors adopt two complementary ideas. First, they model realistic hardware noise using a sparse Pauli‑Lindblad description, which makes (A) close to diagonal and dramatically reduces its effective rank. Second, they employ tensor‑network (TN) techniques—specifically Matrix Product Operators (MPO) for channels and Matrix Product States (MPS) for states/observables—to compress (A), (c), and the observable representations. By restricting to one‑dimensional nearest‑neighbor architectures (the connectivity of most superconducting and trapped‑ion devices), the MPO bond dimension (\chi) remains modest for low noise, turning exponential storage into linear (O(n\chi^2)) memory and polynomial (O(n\chi^3)) computational cost.

Because constructing the exact (\hat Y) is still costly, the authors introduce the Dominant Component Approximation (DCA). When the target observable (X) is a single Pauli string, the optimal (\hat Y) can be well approximated by a rescaled version of the same Pauli string, (\hat Y \approx \alpha P_i). They validate this approximation on the Trotterized Ising model, a standard benchmark for QEM. Numerical simulations compare DCA‑based pre‑processing against TEM. The results show: (i) a reduction of average estimation error and variance by roughly a factor of two to three, (ii) measurement overhead approaching the theoretical lower bound (i.e., essentially one observable measurement per circuit run), and (iii) a dramatic decrease in classical computational effort—about six orders of magnitude fewer tensor contractions than required by TEM. Moreover, the authors prove that the estimator based on (\hat Y) saturates the Quantum Cramér‑Rao Bound (QCRB), establishing its statistical optimality.

Key advantages of the proposed method are: (1) elimination of complex post‑processing pipelines, (2) removal of the need for informationally complete POVMs or shadow tomography, thereby reducing experimental overhead, and (3) scalability to tens of qubits thanks to the efficient MPO/MPS representation. Limitations include the reliance on 1‑D nearest‑neighbor connectivity, the assumption of relatively low noise (so that the MPO bond dimension stays small), and the neglect of SPAM (state preparation and measurement) errors, which are set to zero for the analysis. Extending the approach to higher‑dimensional connectivity, stronger non‑unital noise, and realistic SPAM errors remains an open research direction.

In summary, the paper presents a compelling pre‑processing QEM framework that leverages tensor‑network compression to construct a surrogate observable enabling noise‑free expectation‑value estimation directly at measurement time. By achieving near‑optimal statistical performance and orders‑of‑magnitude reductions in classical computational cost, this work offers a practical pathway for improving the accuracy of NISQ‑scale quantum computations without the heavy overhead of traditional post‑processing error mitigation techniques.