The machine learning community has high hopes for quantum computers — devices that can store and process quantum data and are expected to perform many computational tasks exponentially faster than classical computers. The variability among different quantum devices however presents challenges for the scalability of semiconductor quantum devices.
In a new Nature paper, researchers from the University of Oxford, DeepMind, University of Basel and Lancaster University propose a novel machine learning (ML) algorithm that can tune quantum devices to optimal performance in a median time of under 70 minutes, faster than a typical tuning process performed by human experts. The proposed algorithm is also approximately 180 times faster than an automated random search of the parameter space, and is capable of dealing with different material systems and device architectures.
“Until this work, coarse tuning required manual input or was restricted to a small gate voltage subspace,” the researchers explain. Many ML techniques and other automated approaches have been proposed for quantum devices tuning, but these solutions tend to be limited to small regions of a device parameter space or require information about device characteristics.
“We believe our work significantly improves the state-of-the-art: our algorithm models the entire parameter space and tunes a device completely automatically without human input,” the researchers say, adding that tuning times could be further improved with the addition of more efficient measurement techniques.
The proposed algorithm comprises a sampling stage that generates candidate locations on the hypersurface and an investigation stage to evaluate transport features. It minimizes tuning times by identifying candidate locations on a hypersurface model updated with each measurement, by prioritizing the most promising of these locations, and by avoiding the acquisition of two-dimensional current maps.
The research team says that over several runs on two different devices or over multiple thermal cycles on the same device, the algorithm successfully found transport features corresponding to double quantum dots, which are promising candidates for scalable quantum computation and simulation.
The researchers suggest other device architectures could use the sampling stage of their algorithm as a first tuning step, and that the investigation stage could be adapted to tune quantum devices into more diverse configurations.
The paper Machine Learning Enables Completely Automatic Tuning of a Quantum Device Faster Than Human Experts is on Nature.
Reporter: Yuan Yuan | Editor: Michael Sarazen
This report offers a look at how China has leveraged artificial intelligence technologies in the battle against COVID-19. It is also available on Amazon Kindle. Along with this report, we also introduced a database covering additional 1428 artificial intelligence solutions from 12 pandemic scenarios.
Click here to find more reports from us.
We know you don’t want to miss any story. Subscribe to our popular Synced Global AI Weekly to get weekly AI updates.