As quantum technologies scale up in size, the testing and evaluation metrology for systems needs to be largely automated. For example, when the qubit number in quantum computers reaches thousands and beyond as expected in the next five years, existing processes will no longer be practical. NPL has developed methods to characterise and model the noise sources in quantum systems such as qubits, which are required as part of automated characterisation and calibration processes. They can then be combined with machine learning (ML) based approaches to automate individual tasks in the characterisation and calibration chain. We have developed a number of quantum computing algorithms that use quantum computers for practical application in materials science, chemistry and machine learning.
Using machine learning within automated processes typically relies on neural networks, which is a ‘black box’ type approach with unspecified error bars and uncertainties. This breaks the traceability chain. NPL is developing methods to overcome this critical problem and determine the trustworthiness and uncertainty of the machine learning components in the automated systems. This would have the benefit of providing error bars within the automated characterisation, allowing the estimation of the uncertainty propagation and maintaining the traceability chain in the quantum hardware testing and characterisation.
Our capability on trustworthy machine learning for characterisation and calibration will enable the UK quantum industry to develop faster, scalable systems that can be trusted and give a significant competitive advantage to the UK’s emerging quantum technologies sector. This capability will benefit quantum technology software companies, hardware companies, hardware component manufacturers, as well as end-users.
NPL will provide uncertainty quantification allowing estimation of the reliability of results of quantum computations and of wider quantum technology such as single electron pumps, and their dependence on individual hardware components as well as used algorithms. This is key for user trust and adoption of the emerging quantum computing technology.
Find out more about NPL’s research into Computational modelling.
Would you like to speak to our quantum team?
Contact our quantum team