Automatic Characterization of Fluxonium Superconducting Qubits Parameters with Deep Transfer Learning

Abstract

This work develops a transfer-learning-based workflow for automatic characterization of fluxonium-qubit Hamiltonian parameters from spectroscopic data. The method uses simulated spectra for training and combines machine-learning estimates with automated fitting, including tests with noisy or incomplete spectral inputs.

Citation

@misc{Kung2025FluxoniumTransferLearning,
  author        = {Kung, Huan-Hsuan and Liu, Chen-Yu and Lee, Qian-Rui and Hu, Chiang-Yuan and Chang, Yu-Chi and Chen, Ching-Yeh and Wang, Daw-Wei and Lin, Yen-Hsiang},
  title         = {Automatic Characterization of Fluxonium Superconducting Qubits Parameters with Deep Transfer Learning},
  year          = {2025},
  eprint        = {2503.12099},
  archivePrefix = {arXiv}
}