Time-resolved noise characterization tool to track fluctuating noise effects in superconducting qubits

Superconducting qubits have seen rapid increases in their coherence in the last few decades. However, low-frequency noise present in the qubits still causes non-Markovian errors and qubit instability. Collectively characterising different sources of low-frequency noise can be challenging, and noise sources such as charge parity switching and coupling to thermal fluctuators are typically characterised independently. In order to characterise the combined noise, we develop a tool based on Ramsey tomography that uses few-shot data to detect and diagnose qubit frequency fluctuations. To further disambiguate different sources of frequency fluctuations we develop a Hidden-Markov-Model based time series segmentation software. We demonstrate the tool by computing time and spectrally resolved properties, with which we extract qubit frequency changes and the coupled variation of two level fluctuator strength and charge offset. Our framework for frequency fluctuation detection and disambiguation can be used to thoroughly characterize low-frequency noise in qubits as well as to develop methods to mitigate the noise.