Our paper Machine learning enables completely automatic tuning of a quantum device faster than human experts has been published today in Nature Communications.
Semiconductors are among the most promising materials for making a quantum computer. However, semiconductor devices contain defects, which means that every qubit is slightly different. These differences must be cancelled by adjusting the voltages used to control each device. To find the correct voltage settings by hand is time-consuming even for a single device. This paper shows how a neural network can learn the effect of the gate voltages and configure a double quantum dot.