Controlling Quantum Device Measurement using Deep Reinforcement Learning
share
Qubits based on semiconductor quantum dot devices are promising building blocks for the realisation of quantum computers. However, measuring and characterising these quantum dot devices can be challenging and laborious for the experimentalists. In this paper, we develop an elegant application using deep reinforcement learning for controlling the measurement of quantum dot devices. Specifically, we present a computer-automated algorithm that measures a map of current flowing through a double quantum dot device for different settings of its gate electrodes. The algorithm seeks particular features called bias-triangles indicating the device is in the right operating regime of realising a qubit. Our approach requires no human intervention and significantly reduces the measurement time. This work alleviates the user effort required to measure multiple quantum dot devices, each with multiple gate electrodes.
READ FULL TEXT