Cookies on this website

We use cookies to ensure that we give you the best experience on our website. If you click 'Accept all cookies' we'll assume that you are happy to receive all cookies and you won't see this message again. If you click 'Reject all non-essential cookies' only necessary cookies providing core functionality such as security, network management, and accessibility will be enabled. Click 'Find out more' for information on how to change your cookie settings.

Funded by Innovate UK, the £6.8M project will apply machine learning techniques to find fast, automated, and scalable ways to calibrate quantum computers. The aim is to build a system capable of controlling hundreds of qubits simultaneously across different types of quantum hardware.

Artist's impression of quantum computer

For quantum computers to be used practically, a large number of qubits, or quantum bits, need to be controlled with absolute precision and without errors. This is an extremely difficult task due to the fragility of qubits, which can collapse and lose information due to even the smallest changes in the environment. As a result, quantum computers require constant supervision by highly skilled physicists. Calibration is currently inefficient and time intensive. The cost, time, and effort involved in this process are currently not scalable and constitute a major bottleneck to progressing quantum computing technology.

The quantum device group led by Prof Ares at University of Oxford, in collaboration with machine-learning expert Prof Osborne, pioneered the use of machine learning techniques for quantum device control in real time. Their success in developing quantum device tuning faster than human experts revealed the potential of machine-learning based approaches for the scaling of quantum circuits. Deltaflow Control, a control system being developed by Riverlane, will now build on these techniques to manipulate several qubits simultaneously and will be portable across multiple types of quantum hardware.

Artificial Intelligence specialists Mind Foundry will develop the bespoke machine learning techniques. Machine-learning based qubit calibration will enable faster, more predictable measurements. Users can focus on running useful experiments and get more out of their qubits, instead of spending time and energy on setting qubits up and keeping them from collapsing.  The consortium includes world-leading quantum hardware suppliers Oxford Ionics and SEEQC each representing different types of quantum hardware: trapped ion and superconducting qubits; the University of Oxford will lead the effort on semiconductor qubit circuits, sharing their expertise in tuning such devices through ML-based techniques. The National Physical Laboratory (NPL) and the University of Edinburgh will set the standards for measurement.

According to Prof Natalia Ares at University of Oxford: ‘Every researcher that has characterised and tuned a quantum device has experienced how much experience it requires and how time consuming it can be. When we started a project in collaboration with Prof Osborne to apply machine learning to this challenge, we were not expecting the extent to which we could delegate these tasks to learning algorithms. These have been transformative for my lab and for others labs around the world with which we collaborate. We are now tremendously excited to unleash the potential of these approaches to allow for the scalability of quantum circuits in a vast range of device architectures.’