Machine learning enables completely automatic tuning of a quantum device faster than human experts

Area of Science:

• Quantum Computing
• Materials Science
• Machine Learning

Background:

• Semiconductor quantum devices face scalability challenges due to significant variability.
• Optimizing these devices involves navigating a large parameter space with a narrow operating range.

Purpose of the Study:

• To develop a machine learning algorithm for efficiently tuning gate-defined quantum dot devices.
• To address and quantify device variability in semiconductor quantum technologies.

Main Methods:

• A statistical tuning algorithm employing machine learning was utilized.
• The algorithm searches an eight-dimensional gate voltage space for optimal electron transport features.

Main Results:

• The machine learning algorithm achieved optimal device performance in a median time of under 70 minutes.
• This method demonstrated approximately 180 times greater speed compared to automated random search.
• Quantitative measurements of device variability were obtained, including inter-device and post-thermal cycling variations.

Conclusions:

• The developed machine learning algorithm offers a scalable solution for tuning quantum devices.
• The approach is adaptable to various material systems and device architectures, paving the way for broader quantum technology development.