一句话：首次形式化量子程序可测试性，提出信息论度量 Quantum Squeeziness，并通过大规模变异分析验证其有效性。
可测试性度量在经典软件工程中源远流长，但在量子软件领域几乎空白。本文首次为量子程序给出可测试性的形式化定义，将经典概念适配到量子计算语境，并提出 Quantum Squeeziness——一个受经典 squeeziness 启发的信息论度量，用以量化故障如何被掩盖而无法传播到可观测结果。该度量被视为可测试性的逆指标：squeeziness 越高，故障掩盖越强、可检测性越低。作者给出高效计算方法，并通过超过 5 万个量子变异体的大规模变异分析进行验证，观察到 squeeziness 与变异得分之间存在强负相关，表明高 squeeziness 可预测低可测试性。此外，作者提出一种限制扰动注入位置与方式的技术，在保持精度的同时降低成本：其计算比标准误差传播基线快 15%，而优化变体 FastSqueeziness 将平均运行时间从 3200 秒降至 189 秒（16.9 倍加速），精度达 95%。该工作为量子软件可测试性奠定了严谨基础，并为量子纠错与错误缓解提供支持。
Testability—the degree to which faults can be revealed and diagnosed—has long been fundamental in software engineering but remains largely unexplored for quantum software. This work presents the first formalisation of testability for quantum programs, adapting classical concepts to quantum computation. We introduce Quantum Squeeziness, an information-theoretic metric inspired by classical squeeziness, to quantify how faults are masked and prevented from propagating to observable outcomes. We treat it as an inverse measure of testability: higher squeeziness implies stronger fault masking and reduced detectability. We propose an efficient method to compute Quantum Squeeziness and validate it through large-scale quantum mutation analysis. Across more than 50k mutants, we observe a strong negative correlation between squeeziness and mutation score, showing that higher squeeziness predicts lower testability. We further introduce a technique that constrains where and how perturbations are injected, reducing cost while preserving accuracy. The computation is 15% faster than a standard error-propagation baseline, while the optimised variant, FastSqueeziness, reduces mean runtime from 3200s to 189s (16.9 times speedup) with 95% accuracy. By establishing a rigorous foundation of testability for quantum software, this work opens new directions at the intersection of software engineering, quantum information theory, and testing. Beyond testing and debugging, it also supports applications in quantum error mitigation and correction, enabling principled assessment and improvement of program robustness.
@inproceedings{paperbot12194,
title = {Quantum Squeeziness: An Information Theoretical Metric for Quantum Software Testability},
author = {Avner Bensoussan and Hector Menendez and Mohammad Reza Mousavi},
booktitle = {ASE},
year = {2026},
doi = {10.1145/3832783.3834341}
}