Accelerated characterization of two-level systems in superconducting qubits via machine learning

We introduce a data-driven approach for extracting two-level system (TLS) parameters-frequency $\omega_{TLS}$, coupling strength $g$, dissipation time $T_{TLS, 1}$, and the pure dephasing time $T^{\phi}_{TLS, 2}$, labelled as a 4-component vector $\vec{q}$, directly from simulated spectroscopy data generated for a single TLS by a form of two-tone spectroscopy. Specifically, we demonstrate that a custom convolutional neural network model(CNN) can simultaneously predict $\omega_{TLS}$, $g$, $T_{TLS, 1}$ and $T^{\phi}_{TLS, 2}$ from the spectroscopy data presented in the form of images. Our results show that the model achieves superior performance to perturbation theory methods in successfully extracting the TLS parameters. Although the model, initially trained on noise-free data, exhibits a decline in accuracy when evaluated on noisy images, retraining it on a noisy dataset leads to a substantial performance improvement, achieving results comparable to those obtained under noise-free conditions. Furthermore, the model exhibits higher predictive accuracy for parameters $\omega_{TLS}$ and $g$ in comparison to $T_{TLS, 1}$ and $T^{\phi}_{TLS, 2}$.