Locality and Error Mitigation of Quantum Circuits

TL;DR

This paper advances error mitigation in quantum circuits by refining PEC and ZNE techniques using light-cone analysis, significantly reducing sampling overhead and providing more accurate error bounds for local observable estimation.

Contribution

It introduces a light-cone-based estimator for PEC with reduced sampling overhead and establishes improved error bounds for ZNE, enhancing local observable accuracy in quantum computations.

Findings

New PEC estimator reduces sampling overhead by orders of magnitude.

Light-cone analysis provides tighter error bounds for ZNE.

Enhanced techniques improve local observable estimation accuracy.

Abstract

In this work, we study and improve two leading error mitigation techniques, namely Probabilistic Error Cancellation (PEC) and Zero-Noise Extrapolation (ZNE), for estimating the expectation value of local observables. For PEC, we introduce a new estimator that takes into account the light cone of the unitary circuit with respect to a target local observable. Given a fixed error tolerance, the sampling overhead for the new estimator can be several orders of magnitude smaller than the standard PEC estimators. For ZNE, we also use light-cone arguments to establish an error bound that closely captures the behavior of the bias that remains after extrapolation.