The quantum computing teams of the BMW group and Capgemini benchmark quantum computing applications with QUARK

Quantum computing has the potential to solve industry-relevant problems in the near future. But what applications are the most promising?

Which optimization algorithm or machine learning model is best-suited for any given use case? And what quantum hardware technology performs best in the most important metrics?

Why QUARK?

The BMW Group and Capgemini have recently joined forces to explore the potential of quantum computing, with a particular focus on benchmarking quantum algorithms using the Quantum Computing Application Benchmarking (QUARK) framework. QUARK is a standardized benchmarking tool that aims to provide an unbiased assessment of quantum computing performance.

One of the key advantages of QUARK is its ability to offer an unbiased framework for evaluating quantum algorithms and quantum machine learning (QML) training models. This is particularly important in the rapidly evolving field of quantum computing, where different hardware platforms and approaches are being developed. QUARK’s open-source nature and neutrality ensures that the benchmarking process is not skewed towards any specific vendor or technology.

QUARK addresses the challenges inherent in benchmarking QML methods and optimization algorithms. These tasks are notoriously difficult to assess, as they often involve complex interactions between the algorithm, the problem instance, and the quantum hardware. QUARK provides a structured framework to tackle these challenges, enabling a more comprehensive and meaningful evaluation of quantum computing capabilities.

Why Maximum Independent Set Problems?

The variety of industry-relevant combinatorial optimization problems is immense. They involve finding the best solution from a large number of possible configurations. Quantum computers, with their ability to explore multiple solutions simultaneously, have the potential to outperform classical computers in solving these complex optimization problems.

Many optimization use cases, such as sensor placement, windmill placement, and traffic optimization, can be modelled as maximum independent set (MIS) problems. Neutral atom based quantum devices benefit from the structure of this problem class and could be a candidate for an early quantum advantage. By leveraging the QUARK framework, the teams can rigorously assess the performance of such devices on these real-world applications, providing valuable insights into the practical applications of quantum computing.

Next Steps and Challenges

As the BMW Group and Capgemini continue to develop and work on the QUARK framework, the next steps will involve further development and refinement of the QUARK framework, particularly in the area of graph coloring problems. This will allow the teams to benchmark quantum algorithms on a wider range of optimization challenges, pushing the boundaries of what is possible with quantum computing.

The methods generated in the coming weeks and months will undoubtedly challenge quantum hardware providers. By using objective benchmarks with open-source tools like QUARK, the providers can transparently and directly compare the performance for different applications across different quantum technologies. We hope the neutral atom quantum computer providers will join in the open benchmarking methodology, as this will help drive the development of more powerful and efficient quantum hardware, ultimately accelerating the adoption of quantum computing in various industries.

The quantum computing teams of the BMW Group and Capgemini are very excited to develop and work with this framework and to answer the question: which device will provide a real quantum advantage first?