Automation Tackles Quantum Skill Shortages for Faster Software Development

A thorough review of existing literature reveals automation can sharply improve efficiency in quantum computing and hybrid quantum-classical applications, according to Nazanin Siavash and Armin Moin at University of Colorado Springs. Automated approaches address a key need arising from the limited number of skilled developers and data scientists able to navigate the complexities of quantum technologies. This review specifically details automation techniques to broaden the development and deployment of quantum and AI-enabled software systems, offering a new set of tools for these fields.

Automation’s current status in Quantum Software Engineering and Artificial Intelligence

A systematic literature review underpinned this analysis, employing rigorous protocols to identify, select, and critically appraise all relevant studies. Multiple digital libraries and databases were searched carefully, utilising specific keywords, including ‘quantum software engineering’, ‘quantum artificial intelligence’, ‘automation’, ‘quantum compilation’, and ‘quantum testing’, to unearth studies concerning automation in Quantum Software Engineering and Quantum Artificial Intelligence. The process involved screening thousands of papers, initially by title and abstract, then through full-text review against strict inclusion and exclusion criteria, ensuring only the most pertinent studies were included. These criteria focused on papers detailing (semi-) automated techniques applied to the development lifecycle of quantum software, encompassing areas like code generation, optimisation, verification, and deployment. Unlike previous surveys examining these fields separately, this review thoroughly assesses automation within both disciplines, recognising the limited number of experts possessing both quantum computing and artificial intelligence skills. Automation is therefore key for efficient development and deployment. The scarcity of qualified personnel presents a significant bottleneck to progress, as quantum programming demands expertise in linear algebra, quantum mechanics, and software engineering principles, a combination rarely found in a single individual. This review aims to provide a comprehensive overview of how automation can mitigate this challenge.

Advancing quantum application development through automated resource allocation

Automated techniques in Quantum Software Engineering and Quantum Artificial Intelligence now handle applications previously limited to simulations, representing a substantial leap forward. Prior to this systematic review, productivity gains in quantum and hybrid quantum-classical systems heavily relied on manual intervention and expertise, now enabling broader development and deployment, particularly given the scarcity of qualified professionals. The review details how automation assists in deciding where to deploy application components, either on quantum hardware or conventional platforms, optimising efficiency for both experts and novices. This resource allocation is crucial, as quantum hardware is currently limited in qubit count and coherence times, making it impractical to execute entire applications on quantum processors. Automated tools can analyse application requirements and intelligently partition workloads, executing computationally intensive tasks on classical hardware while leveraging quantum resources for specific algorithms where they offer a demonstrable advantage. This hybrid approach maximises performance and minimises resource consumption.

Practices from conventional computing are increasingly being applied to quantum software and artificial intelligence, extending to these emerging fields. Wang et al. demonstrated reversible implementation of algebraic functions, although circuits required substantial qubits and operations, with validation limited to quantum simulators. This work, while showcasing the potential of reversible computing, highlights the challenges of scaling quantum circuits due to the exponential growth in qubit requirements. Zhou et al. utilised simulated annealing to optimise circuit mapping, yet performance diminished with larger circuits and failed to account for hardware limitations like crosstalk. Crosstalk, the unwanted interaction between qubits, is a significant source of error in quantum computations and must be considered during circuit optimisation. Zhao introduced a preliminary catalogue of code refactoring for quantum programs, but the tool remains under development and lacks empirical validation. Refactoring, the process of improving code structure without changing its functionality, is essential for maintaining code quality and reducing complexity, but its application to quantum programs is still in its infancy. Bugs4Q, a benchmark of 36 verified Qiskit bugs, enhances reproducibility but represents a limited scope of potential errors. While valuable for testing, Bugs4Q only covers a small fraction of the possible errors that can occur in quantum programs, necessitating more comprehensive testing methodologies.

Current validation techniques and the need for hardware-integrated testing

Automating quantum workflows promises to unlock the potential of this nascent technology, easing the strain on a severely limited workforce of qualified experts. The review highlights a current reliance on quantum simulators for validating many automated techniques, a necessary step but ultimately insufficient. Quantum simulators, while enabling rapid prototyping and testing, cannot fully replicate the behaviour of real quantum hardware due to limitations in simulating decoherence, noise, and qubit connectivity. Tools like Bugs4Q offer reproducible testing, yet their limited scope raises concerns about thoroughly identifying and addressing real-world errors in increasingly complex quantum programs. The development of more robust and scalable validation techniques is therefore paramount.

Quantum simulators define a necessary stage for development, enabling reproducible testing and identifying potential errors, building a foundation for validation as quantum hardware improves and becomes more accessible. Automation is aiding quantum software development, addressing the limited number of developers with the required expertise. Mutation analysis and similar techniques offer reproducible testing, establishing a base for validation as quantum hardware advances. Mutation analysis involves introducing small, deliberate errors into the code and verifying that the testing suite can detect them, providing a measure of test coverage. However, applying these techniques to quantum programs requires careful consideration of the unique characteristics of quantum computation.

Further advances will likely unlock more complex applications within the decade. This systematic review establishes a consolidated understanding of automation’s role in both Quantum Software Engineering and Quantum Artificial Intelligence, fields currently hampered by a shortage of skilled professionals. Mapping existing techniques reveals a shift towards adapting conventional software development practices for quantum systems, demanding new approaches to code optimisation and deployment. Identifying this trend clarifies that it isn’t simply a convenience, but a necessity for scaling quantum and hybrid quantum-classical technologies. Future research should focus on developing hardware-integrated testing methodologies, allowing automated techniques to be validated directly on quantum hardware, and exploring the use of machine learning to automate the optimisation of quantum circuits and resource allocation. The integration of automated testing with continuous integration and continuous delivery (CI/CD) pipelines will also be crucial for accelerating the development and deployment of quantum software.

This research identified a growing trend of applying automated techniques from software engineering and artificial intelligence to the development of quantum and hybrid quantum-classical software. Automation is particularly important in this field due to a limited number of developers possessing the necessary expertise in quantum computing. The systematic review highlights the adaptation of existing software practices for quantum systems, which is becoming essential for scaling these technologies. Researchers suggest future work should concentrate on hardware-integrated testing and utilising machine learning to optimise quantum circuits and resource allocation.