NOISE-AWARE QUANTUM SOFTWARE TESTING - QOIN 方案用于缓解噪声影响

Aug 5, 2026

Upload by Himbo

Quantum Computing promises computational speedups, yet current and near-term quantum devices suffer from hardware noise. This makes quantum software testing unreliable: a test can fail due to noise rather than genuine faults, especially when testing is executed on ideal simulators that ignore noise. The proposed noise-aware method, QOIN, uses machine learning to learn and filter a quantum computer’s noise effects, then assesses test cases using the filtered outputs against an oracle. Experiments on IBM noise models show noise reduction and improved oracle performance.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages20

Reading Time50 min

AccessFree to download

Abstract

1 Introduction

QOIN

Problem with noise on NISQ

Noise-aware testing via machine learning

NOISE-AWARE QUANTUM SOFTWARE TESTING arXiv :2306 . 16992v1 [ cs . SE] 29 Jun 2023 Asmar Muqeet Simula Research Laboratory University of Oslo Oslo [asmar@simula.no](asmar@simula.no) Tao Yue Simula Research Laboratory Oslo [tao@simula.no](tao@simula.no) Shaukat Ali Simula Research Laboratory and Oslo Metropolitan University Oslo [shaukat@simula.no](shaukat@simula.no) Paolo Arcaini National Institute of Informatics Tokyo [arcaini@nii.ac.jp](arcaini@nii.ac.jp) ABSTRACT Quantum Computing (QC) promises computational speedup over classic computing for solving some complex problems. However, noise exists in current and near-term quantum computers. Quantum software testing (for gaining conﬁdence in quantum software's correctness) is inevitably impacted by noise, to the extent that it is impossible to know if a test case failed due to noise or real faults. Existing testing techniques test quantum programs without considering noise, i.e., by executing tests on ideal quantum computer simulators. Consequently, they are not directly applicable to testing quantum software on real QC hardware or noisy simulators. To this end, we propose a noiseaware approach (named QOIN ) to alleviate the noise effect on test results of quantum programs. QOIN employs machine learning techniques (e.g., transfer learning) to learn the noise effect of a quantum computer and ﬁlter it from a quantum program's outputs. Such ﬁltered outputs are then used as the input to perform test case assessments (determining the passing or failing of a test case execution against a test oracle) . We evaluated QOIN on IBM's 23 noise models with nine real-world quantum programs and 1000 artiﬁcial quantum programs. We also generated faulty versions of these programs to check if a failing test case execution can be determined under noise. Results show that QOIN can reduce the noise effect by more than 80% . To check QOIN 's effectiveness for quantum software testing, we used an existing test oracle for quantum software testing. The results showed that the F1-score of the test oracle was improved on average by 82% for six real-world programs and by 75% for 800 artiﬁcial programs, demonstrating that QOIN can effectively learn noise patterns and enable noise-aware quantum software testing. Keywords Software and its engineering 􀀁 Software testing and debugging 􀀁 Computing methodologies 􀀁 Instancebased learning 􀀁 Quantum Computing 􀀁 Machine learning. 1 Introduction There has been an increased interest in quantum software engineering over the past few years, focusing on designing, developing, and testing quantum computing (QC) applications [1, 2, 3, 4, 5, 6, 7, 8] . This growth of interest is due to the computational power promised by quantum computers to solve a particular class of problems more efﬁciently than classic computers [9] . In addition, quantum computers (IBM [10], Google [11], Honeywell [12]), and quantum computer simulators such as QuEST [13], QX [14], and IBM's Qiskit Aer simulator [15] are becoming available. However, quantum computers are susceptible to hardware noise due to immature hardware and environmental factors (e.g., magnetic ﬁelds, radiations) [16, 17] . Noise affects the accuracy of calculations a quantum computer performs, thus resulting in incorrect program outputs. Such computers with inherent noise are known as Noisy IntermediateScale Quantum (NISQ) computers [18] . QOIN Quantum software testing aims to cost-effectively ﬁnd quantum software bugs to achieve a certain level of conﬁdence in their correctness [19, 20, 21, 22, 1] . Testing quantum software is challenging due to the inherent quantum mechanics' features, such as superposition and entanglement [19, 22] . In addition, noise brings another layer of complexity to the challenge. For example, when checking test results, it becomes difﬁcult to conclude whether a test case execution failed due to a faulty quantum program or noise. Existing quantum software testing appr

FAQs

为什么现有量子软件测试方法难以直接用于噪声量子硬件？

现有方法通常在理想、无噪声模拟器上执行，导致测试结果无法反映真实噪声影响；在 NISQ 设备上，噪声会使相同程序的输出随设备而变化，从而难以判断测试失败到底来自噪声还是程序故障。

QOIN 如何缓解噪声对量子程序测试结果的影响？

QOIN 通过监督学习训练模型学习量子计算机的噪声模式，并将噪声从量子程序输出中进行过滤；再用过滤后的输出进行测试用例评估，与测试判定准则（oracle）对照通过/失败。

实验结果表明 QOIN 的有效性如何？

在 IBM 的 23 种噪声模型上，对真实与人工量子程序进行评估；结果显示噪声影响可降低 80% 以上。使用现有测试 oracle 时，6 个真实程序的平均 F1-score 提升 82%，800 个人工程序平均提升 75%。

Aug 5, 2026

Upload by Himbo