Insider Brief
- NPL is collaborating with NVIDIA to integrate AI tools into quantum measurement systems for automated calibration and characterization of quantum devices.
- The approach uses NVIDIA’s Ising AI models to assess qubit stability, reducing reliance on manual calibration and improving performance monitoring.
- The work supports broader efforts to develop benchmarking frameworks and scalable, AI-driven methods for reliable quantum computing systems.
PRESS RELEASE — NPL, the UK’s National Metrology Institute (NMI), plays a central role in providing accurate and trusted measurement across emerging technology. Within its Institute for Quantum Standards and Technology (IQST), the team is developing methods to characterise and calibrate quantum devices, particularly quantum computing.
As part of a new collaboration, NPL is integrating NVIDA’s Ising AI tools into its quantum measurement systems to automate key calibration tasks. This approach will help address one of the major challenges facing quantum computing: in the need to manage large numbers of qubits, each affected by multiple sources of noise and instability.
Qubit performance is commonly assessed using metrics such as the qubit relaxation time, usually referred to as T1 time, which is a metric for the timescale at which a qubit decays from its excited state to the ground state. These values can fluctuate or drift due to interactions with the environment, requiring frequent checks to ensure reliable operation. Traditionally, such checks are carried out manually by experts.
NPL has demonstrated that these assessments can be automated using NVIDIA Ising Calibration – a trained vision language model. The system can determine whether a qubit’s coherence time is stable and identify different types of instability, such as sudden fluctuations or gradual drifts. This information can then be used to inform corrective actions and improve overall system performance.
In a joint paper, a benchmarking suite was developed which evaluates the performance of different AI methods in analysing qubit calibration data. The qubit coherence stability checks are used as one of the benchmarks within the suite. The work builds on earlier research showing that machine-learning techniques can accelerate quantum device characterisation and provide new insights into the physical sources of noise.
This collaboration will contribute to NPL’s wider efforts to develop independent and transparent benchmarking frameworks for quantum computing. Such metrics are seen as strategically important for the UK’s National Quantum Technologies Programme, helping to guide investment decisions and support the development of commercial quantum hardware.
The next phase of the project will focus on demonstrating scalable, AI-driven calibration methods and developing assurance frameworks to ensure the outputs of AI tools used in quantum measurement can be trusted.
