Tracking calibration drifts in continuous quantum measurement

We investigate how to correct the calibration drifts of a qubit undergoing continuous measurement. Currently, qubit measurement calibrations only remain valid for short measurement durations before naturally drifting due to uncontrolled environmental factors. In order to continuously measure a qubit for a longer duration and correctly interpret its evolution, it is necessary to dynamically update the calibrations while the measurements occur. We explore the use of machine learning models such as recurrent neural networks for predicting the behavior of such calibration drifts over a longer duration. This effort is a step towards autocalibration for the continuous measurement process.