Refining Noise Mitigation in NISQ Hardware Through Qubit Error Probability

Abstract: A key characteristic of the Noisy Intermediate-Scale Quantum (NISQ) era is the accurate assessment and consideration of errors through error mitigation techniques, which play a fundamental role in achieving quantum utility, i.e., quantum results that are at the state of the art of classical calculations. In this paper, we examine the primary sources of errors in contemporary (IBM) quantum computers through a new important metric, the qubit error probability (QEP), which estimates the probability of a qubit to suffer an error. We show that QEP can be used to improve one of the most important error mitigation techniques, the zero noise extrapolation (ZNE). Our method, named zero error probability extrapolation (ZEPE), is based on calibration parameters which enables a good scalability in terms of number of qubits and circuit depth. Besides, we prove that ZEPE performs better than the standard ZNE, specially for a mid-size depth ranges.