Jonas Schuff, Miguel J. Carballido, Madeleine S. Kotzagiannidis, Juan Calvo, Marco Caselli, J. H. Rawling, David Craig, Barnaby van Straaten, Brandon Severin, Federico Fedele, Simon Svab, Pierre Chevalier Kwon, Rafael S. Eggli, Taras Patlatiuk, Nathan Korda, Dominik M. Zumbühl, Natalia Ares
Abstract The development of large-scale semiconductor quantum circuits is limited by the difficulties involved in efficiently tuning and operating such circuits. Identifying optimal operating conditions for these qubits is, in particular, complex and involves the exploration of vast parameter spaces. Here we report the autonomous tuning of a semiconductor qubit, from a grounded device to Rabi oscillations. Our approach integrates deep learning, Bayesian optimization and computer vision techniques. We demonstrate this automation in a germanium–silicon core–shell nanowire device. To illustrate the potential of full automation, we characterize how the Rabi frequency and g -factor depend on barrier gate voltages for one of the qubits found by the algorithm. We expect our automation algorithm to be applicable to a range of semiconductor qubit devices, allowing for the statistical studies of qubit-quality metrics.