The advent of the quantum computer is creating the need for new approaches to algorithm design. CEA-List is investigating these approaches and evaluating the advantages of quantum for combinatorial analysis and machine learning.
Our research on algorithms addresses two main challenges.
First, we are exploring new computational models suitable for hybrid (quantum and classical) computer architectures.
Second, we are evaluating the advantage of quantum for specific combinatorial optimization and machine learning use cases and developing the necessary algorithms.
CEA-List researchers are developing ways to evaluate quantum processors. The goal is to have objective criteria for assessing a machine’s performance and selecting the right machine for a given problem. This research will also help improve our understanding of the phenomena at work inside quantum machines for more effective programming.
For example, we completed an assessment of the computational capabilities of an analog quantum machine. This study, which focused on the D-Wave quantum processor, revealed the need to achieve denser topologies to increase performance on optimization problems.
Accurately calculating worst-case execution times (WCET) for components integrated into critical systems is a complex problem that CEA-List’s R&D partners frequently bump up against. WCET is a helpful piece of information for determining how an overall system can be optimized without negatively affecting security or reliability.
Calculating WCET is an optimization problem—exactly the kind of problem where quantum computers could provide a decisive advantage over classical machines. Our researchers are looking at how to use quantum computing to calculate WCET. The case studied here, although simple, was somewhat different than those usually studied and it did provide an idea of possible solutions for complex optimization problems and insights into how to determine the size of the machines needed.
To learn more, read the published article
