QUT: A Unit Testing Framework for Quantum Subroutines

The gist: QUT introduces a novel unit testing framework for quantum subroutines that combines architectural design principles with a practical prototype implementation to provide automated, context-aware testing protocols. This framework addresses the lack of existing quantum software testing techniques by enabling polymorphic probabilistic assertions and automatic protocol orchestration based on the semantics of quantum subroutines.

Theoretical Background

The unit testing approach is based on the notion of local contextual equivalence of quantum subroutines, which means two pieces of code are considered equivalent if they can be interchanged in any complete program without altering its observable behavior. Verifying the equivalence of subroutines requires understanding their semantics through denotational or mathematical semantics, which formalizes the meaning of a program by mapping its syntax to mathematical models. Many adaptations of semantics from classical computer science for quantum programs are based on category theory, specifically using the category FdHilb as the monoidal category of finite-dimensional Hilbert spaces equipped with the symmetric dagger Frobenius structure. The denotational semantics of a quantum subroutine is defined as a quantum process. This process, denoted as Φθ, maps the tensor product of several Hilbert spaces to another Hilbert space, and it can also have a classical output.

Polymorphic Assertions

The context-dependent semantics of quantum subroutines give rise to a variety of outcomes that can serve as arguments in assertion statements within unit tests. A probabilistic assertion is defined as the conditional probability P (AB), where A denotes the assertion statement and B represents the outputs of quantum measurements. The framework provides a unified specification of equality assertions for all argument types, referred to as polymorphic probabilistic assertions, which adapt their evaluation methods based on the types of arguments supplied. The evaluation becomes context-dependent, with the context inferred from the argument types.

Quantum Unit Tests and Protocols

The framework adopts the arrange-act-assert pattern known from classical unit testing. The denotational semantics of each stage corresponds to a quantum process, allowing the entire unit test program to be modeled as a pipeline of quantum channels. The specific values and combinations of parameters, along with the defined form of the quantum processes Φpre n and Φmeas m fully determine the testing protocol in use. For instance, if n is fixed to a single value, and m spans a range of parameters such that Φmeas m forms an informationally complete POVM, the testing protocol corresponds to quantum state tomography.

Methodology and Artifact Design

The methodology of this work is based on the design science research methodology (DSRM) which focuses on the development and validation of prescriptive knowledge in information science. The key objective of the framework is to facilitate the writing of unit tests for quantum subroutines and to provide methods for evaluating various assertion statements involving different data types. The prioritized quality attributes relevant to the proposed framework include accuracy and usability. The artifact was developed in three iterative steps, culminating in an orchestrator class that enables context detection and automatic selection of testing protocols.

Evaluation and Results

The evaluation demonstrated that QUT improves usability by substantially reducing the amount of code and prior knowledge required from users while maintaining accuracy in detecting faulty subroutines. The results confirm that the framework can reliably distinguish between correct and erroneous implementations under noiseless conditions, as well as in the noisy environment inherent to current NISQ devices. The evaluation highlights the benefits of contextual information, as well as the importance of selecting the appropriate testing protocol based on the nature of the subroutine and available information. The presence of realistic hardware noise reduces the J-metric by roughly 0.1 across all testing protocols, indicating a similar overall sensitivity to noise for each protocol.

Conclusion

QUT provides a unified interface for probabilistic contextdependent assertions, automatically selecting suitable testing protocols based on the semantics of the subroutine under test. The experimental results confirm that the framework can reliably distinguish between correct and erroneous implementations under noiseless conditions, as well as in the noisy environment inherent to current NISQ devices. This work represents a step toward integrating rigorous testing practices into the quantum software development lifecycle. Future work will extend QUT in several directions, including expanding the protocol library and exploring error mitigation strategies.

--- Page 1 ---

QUT: A Unit Testing Framework for Quantum Subroutines M. V. KLYMENKO, CSIRO’s Data61, Australia T. HOANG, CSIRO’s Data61, Australia H. T. NGUYEN, School of Computing and Information Systems, The University of Melbourne, Australia S. A. WILKINSON, CSIRO’s Data61, Australia and School of Science, RMIT University, Australia B. GOLDOZIAN, CSIRO’s Data61, Australia X. ZHENCHANG, CSIRO’s Data61, Australia Q. LU, CSIRO’s Data61, Australia and University of New South Wales, Australia M. USMAN, CSIRO’s Data61, Australia and School of Physics, The University of Melbourne, Australia L. ZHU, We present the architectural design and prototype implementation of QUT (Quantum Unit Testing), a framework for unit testing of quantum subroutines. The framework prioritizes usability and simplicity, making the complex theoretical foundations of quantum unit testing - including statistical tests and quantum tomography - accessible to users with diverse backgrounds. This is achieved through the implementation of polymorphic probabilistic assertions, whose evaluation methods adapt to the data types of the arguments used in assertion statements, which may vary according to the context-dependent semantics of quantum subroutines. These arguments can be represented as qubit measurement outcomes, density matrices, or Choi matrices. For each type, the architecture integrates a specific testing protocol – such as quantum process tomography, quantum state tomography, or Pearson’s chi-squared test. The framework is built on the Qiskit software stack, providing compatibility with a broad range of quantum hardware backends and simulation platforms. Drawing on the reasoning provided by the denotational semantics of quantum subroutines, this work also highlights the key distinctions between quantum unit testing and its classical counterpart.

--- Page 2 ---

M. V. Klymenko, T. Hoang, H. T. Nguyen, S. A. Wilkinson, B. Goldozian, X. Zhenchang, QUT: A Unit Testing Framework for Quantum Subroutines 1 Introduction Quantum computing has rapidly progressed from experimental demonstrations to programmable platforms accessible via the cloud. In the current Noisy Intermediate-Scale Quantum (NISQ) era, hardware is characterized by limited number of qubits, coherence time and fidelity, which constrain the reliable execution of quantum programs [1]. Program behavior is not only affected by static hardware properties but also by temporal variability. For example, fluctuations in device parameters such as T1 relaxation times often exhibit temporal autocorrelation, which can degrade circuit fidelity between different calibration cycles [2]. Hardware noise and temporal variations in hardware properties, combined with traditional human errors, can cause quantum programs to produce unexpected outputs, thereby requiring their routine verification. This motivates the exploration of testing techniques for quantum software in general.

1. Bold header: Automated Protocol Selection

QUT implements an orchestrator class that automatically selects the appropriate testing protocol based on the types of arguments used in the assertions, which allows users to focus on writing test cases rather than manually choosing between protocols like quantum state tomography, or quantum process tomography.

2. Bold header: Context-Dependent Polymorphic Assertions

The framework provides a unified specification of equality assertions for all argument types called polymorphic probabilistic assertions that adapt their evaluation methods based on the types of arguments supplied, making the evaluation context-dependent.

3. Bold header: Enhanced Test Quality Metrics

The framework utilizes a metric like Youden’s J statistic, which is calculated as J = (1/NP ∑j=1 αj - 1/NN ∑j=1 βj), allowing for a quantitative measure of the ability of a test to distinguish between correct and incorrect code.

4. Bold header: Robust Hardware Noise Resilience

The framework's evaluation shows that protocols like process tomography successfully identify all mutants, though it is noted that the presence of realistic hardware noise does not significantly affect the number of mutants detected by tomographic protocols, indicating high reliability in complex verification.

5. Bold header: Scalable and Efficient Verification

The system is designed to handle various quantum resources, with computational complexity scaling linearly for statistical tests while acknowledging the exponential complexity of state tomography and process tomography, allowing for middle of the scale protocols to be relevant in practice.

We present the architectural design and prototype implementation of QUT (Quantum Unit Testing), a framework for unit testing of quantum subroutines. The framework is developed with a focus on usability and simplicity, making the complex theoretical concepts behind quantum unit testing accessible to a wide range of users with diverse backgrounds. This is achieved through the implementation of polymorphic probabilistic assertions, whose evaluation methods adapt to the data types of the arguments used in assertion statements, which may vary according to the context-dependent semantics of quantum subroutines. These arguments can be represented as qubit measurement outcomes, density matrices, or Choi matrices. For each type, the architecture integrates a specific testing protocol - such as quantum process tomography, quantum state tomography, or Pearson's chi-squared test - while remaining flexible enough to incorporate additional protocols in the future. The framework is built on the Qiskit software stack, providing compatibility with a broad range of quantum hardware backends and simulation platforms. Drawing on the reasoning provided by the denotational semantics of quantum subroutines, this work also highlights the key distinctions between quantum unit testing and its classical counterpart.