Quantum device fine-tuning using unsupervised embedding learning

Abstract

Quantum devices with a large number of gate electrodes allow for precise control of device parameters. This capability is hard to fully exploit due to the complex dependence of these parameters on applied gate voltages. We experimentally demonstrate an algorithm capable of fine-tuning several device parameters at once. The algorithm acquires a measurement and assigns it a score using a variational auto-encoder. Gate voltage settings are set to optimize this score in real-time in an unsupervised fashion. We report fine-tuning times of a double quantum dot device within approximately 40 min.

Theoretical Bounds on Data Requirements for the Ray-Based Classification

“…Within the field of quantum dots, presented here as an application of the RBC framework, several strategies for classification and tuning using various machine learning techniques have been implemented. Using variational autoencoders, standard device measurements have been optimized to reduce the total number of measurements required [24] and to automate fine tuning in higher ( N > 2 ) dimensions [25]. Machine learning-based binary classifiers have been used to classify 2D stability diagrams as either good or bad for further experimental use [26].…”

Deep reinforcement learning for efficient measurement of quantum devices

“…A full tuning procedure could consist of a super coarse tuning algorithm, followed by this algorithm to locate bias triangles, and a fine-tuning algorithm such as the one described in ref. 39 . Different pairs of bias triangles could be explored.…”

“…However, quantum dot devices are subject to variability, and many measurements are required to characterise each device and find the conditions for qubit operation. Machine learning has been used to automate the tuning of devices from scratch, known as super coarse tuning [34][35][36] , the identification of single or double quantum dot regimes, known as coarse tuning 37,38 , and the tuning of the inter-dot tunnel couplings and other device parameters, referred to as fine tuning [39][40][41] .…”

A deep-learning approach to realizing functionality in nanoelectronic devices

“…Large systems that consist of many coupled quantum dots have a large parameter space and therefore require more sophisticated tuning approaches. Existing tuning methods are based on 'device-in-the-loop' approaches [32][33][34] in which measurement of the physical device is required during optimization. By contrast, parameter tuning in our approach can be performed entirely off-chip.…”